The Minnesota Pylos Project, 1990-98 9781407315348, 9781407344867

Citation preview

________ Frederick A. Cooper (1936–2011) was Morse Alumni Distinguished Professor of Art History at the University of Minnesota (1971–2010) and Andrew F. Mellon Professor of Classical Studies at the American School of Classical Studies at Athens (1982–85). He directed the Minnesota Archaeological Researches in the Western Peloponnese (MARWP) programme. Diane Fortenberry is Senior Editor at Phaidon Press Ltd. She received her PhD in Bronze Age Aegean Archaeology from the University of Cincinnati and acted at various times as Assistant Director for MARWP projects.

BAR  S2856  2017   COOPER & FORTENBERRY (Eds)   THE MINNESOTA PYLOS PROJECT, 1990–98

In 1990 the University of Minnesota carried out an architectural survey of the standing remains of the Bronze Age Palace of Nestor, discovered by Carl Blegen in 1939 and excavated from 1952 to 1966. While the first stone-by-stone state plan of the building was being created, it became clear that some of the architectural assumptions about the structure and its history could not be correct. Over the next eight years the Blegenperiod backfill covering the site was systematically removed so that a complete architectural plan could be prepared. The work was carried out using the protocols of an archaeological excavation. Although only backfill was removed, numerous unexpected finds were recovered, ranging from discarded Linear B tablets and wall painting fragments to roof tiles and pottery; in addition, a detailed study of the architecture revealed evidence for startling new conclusions about the structure of the palace and the history of the site.

The Minnesota Pylos Project, 1990–98 Edited by

Frederick A. Cooper Diane Fortenberry

BAR International Series 2856 9 781407 315348

B A R

2017

The Minnesota Pylos Project, 1990–98 Edited by

Frederick A. Cooper Diane Fortenberry Part I New Studies at the Palace of Nestor with contributions by Todd M. Brenningmeyer, Frederick A. Cooper, Joshua N. Distler, Caitlin Downey, Anne B. Hollond, Eleni M. Konstantinidi-Syvridi, George Otto Marquardt, Shawn A. Ross

Part II The Architecture of the Palace of Nestor by Michael C. Nelson

BAR International Series 2856 2017

Published in by BAR Publishing, Oxford BAR International Series The Minnesota Pylos Project, 1990–98 © The editors and contributors serverally The Authors’ moral rights under the UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reser ved. No par t of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any for m of digital for mat or transmitted in any for m digitally, without the written per mission of the Publisher.

ISBN 9781407315348 paperback ISBN 9781407344867 e-format DOI https://doi.org/10.30861/9781407315348 A catalogue record for this book is available from the British Library

BAR titles are available from: BAR Publishing Banbury Rd, Oxford, [email protected] + ( ) + ( ) www.barpublishing.com

,

for FAC and HBF

CONTENTS

Figures Tables, Graphs and Maps Abbreviations Contributors

7 13 14 15

Preface and Acknowledgements – Frederick A. Cooper Editorial Note – Diane Fortenberry Abstract

17 23 25

PART I: NEW STUDIES AT THE PALACE OF NESTOR 1. The Minnesota Pylos Project: Investigations and Results, 1990–98 – Frederick A. Cooper Excavation Protocol The Palace Traverse The Architectural Study of the Palace and Ancillary Buildings Finds Scientific Analyses Topographical Survey and Digitised Elevation Model of the Englianos Ridge Aerial Photographs 2. Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces – Frederick A. Cooper Knossian Parallels Pylian Drains as Evidence for Early Floor Levels The Middle Helladic Palace A The Late Helladic I–II Palace B The Late Helladic IIIB–IIIC Palace 3. Modelling Movement and Use Patterns within the Palace of Nestor: A GIS/Space Syntax Approach – Todd M. Brenningmeyer Space Syntax Theory Architectural Analysis of the Palace Social Interpretations of Architectural Change Conclusions 4. Enclosed Gardens in Courts 42 and 47 – Anne B. Hollond The Architecture of Courts 42 and 47 The Courts as Garden Spaces Precedents and Conclusions 5. Bronze Age Quarrying: A Provenance Study – Joshua N. Distler The Geology of the Southwest Peloponnese The Quarries The Problem of Transportation 5

29 29 31 34 82 86 87 115 135 135 137 139 144 150 155 156 156 160 163 165 165 167 169 171 171 172 179

The Minnesota Pylos Project, 1990–98 Lithic Analysis Major Element Geochemistry Isotope Analysis Conclusions 6. Discarded Chipped Stone from the Palace of Nestor – George Otto Marquardt Catalogue 7. Small Finds from the 1990–98 Excavations – Eleni M. Konstantinidi-Syvridi Metal Objects Stone Objects Fired Clay Objects Organic Remains General Remarks Catalogue 8. Post-Bronze Age Architecture and Stratigraphy – Todd M. Brenningmeyer Post-Bronze Age Stratigraphy Post-Bronze Age Architecture Summary 9. Post-Bronze Age Pottery – Shawn A. Ross Dark Age Wares (Nichoria I–III) Late Dark Age through Early Archaic Wares Late Archaic, Classical or Hellenistic Black-slipped Wares Hellenistic or Roman Red-slipped Wares Late Roman Plain Ridged Ware Medieval Pottery Conclusions Catalogue 10. Post-Bronze Age Industrial Waste and Bronze Casting – Caitlin Downey Description of the Terracotta Fragments Possible Interpretations Dating the Fragments

181 186 188 191 195 196 213 213 215 216 218 219 220 225 225 240 253 257 258 260 260 260 261 261 262 262 275 275 276 279

PART II: THE ARCHITECTURE OF THE PALACE OF NESTOR – Michael C. Nelson 1. The Palace at Englianos A Brief Description of the LH IIIB Palace The Published Chronology 2. Building Materials Stone Mudbrick Wood 3. Building Methods Foundations Block Masonry Construction Pier-wall Construction Rubble Construction 4. Building History Masonry Development at Englianos Comparisons with Minoan Masonry The Development of the Englianos Hilltop Appendix A: Terracotta Building Elements Appendix B: Clamp Cuttings in Ashlar Blocks Appendix C: Analysis of Earth Material Bibliography State Plans Index

283 284 289 293 293 294 294 297 297 303 329 345 349 349 351 352 367 371 373 379 385 419

6

FIGURES

PART I 1. The Minnesota Pylos Project: Investigations and Results, 1990–98 1.1 Grid plan showing areas of excavation, 1991–98. 1.2 MARWP survey of the palace remains. 1.3 MARWP wall designations. 1.4 Northeast Area, showing Phase 1 features in black. 1.5 Wall 104 beneath Wine Magazine vestibule. 1.6 Area 103–106, Phase 1: plaster f loor and western column base of proto-palatial portico. 1.7 Northeast Area, showing Phase 2 features in black. 1.8 Room 102, the cistern, from above. 1.9 Plan of the Shaft Grave, Room 97. 1.10 Northeast Area, with Phase 3 features in black. 1.11–1.12 Two views of the rosso antico stone embedded in the foundation of Building 97/101. 1.13 Northeast Area, with Phase 4 features in black. 1.14 Northeast Area, with phases 5, 6 and 7 in black. 1.15 State plan of Ramp 91. 1.16 Surface of Ramp 91 adjacent to the northwest (outer) face of Court 42. 1.17 Surface of Ramp 91 overlapping the drain. 1.18 Northeast Area, showing phases 10, 11 and postBronze Age in black. 1.19 Antae and sill at entrance to Room 93. 1.20 Painted ‘altar’ in front of Room 93. 1.21 Western section of the aqueduct. 1.22 The Northeast Gateway, looking northwest. 1.23 The kiln. 1.24 Low point of the aqueduct, with pair of larnakes. 1.25 State plan of the Northwest Area. 1.26 Schematic plan showing Kittredge’s trenches in the Northwest Area (designated WK1-10), overlaid with Papathanasopoulos’ trenches to the southwest. 1.27 State plan of the Northwest Area; phase 1 features in black. 1.28 Wall b, below wall O, below wall E; wall B is at bottom right.

1.29 Sections A–A’ and B–B’ as in plan 1.25, showing the sloping gradient of the terrain in the Northwest Area. 1.30 State plan of the Northwest Area; phase 2 features in black. 1.31 State plan of the Northwest Area; phase 3 features in black. 1.32 State plan of the Northwest Area; phases 4 and 5 features in black. 1.33 Wall B where it is crossed by drain b, looking southwest. Reused ashlar blocks can be seen at the top of the photograph, lining drain b. 1.34 State plan of the Northwest Area; phase 6 features in black. 1.35 State plan of the Northwest Area; phase 7 features in black. 1.36 State plan of the Northwest Area; phases 8, 9, 10 and 11 features in black. 1.37 Walls F (left) and D (right) at their juncture with wall E, looking northwest. Note the triangular stone at the near right corner of wall F as it overlaps wall E. 1.38 Vertical views of walls D, F, H/I and M/L. The elevations show both faces of the walls, in mirrored positions. 1.39 Vertical views of walls B, E and K/N, showing both faces of walls B and K/N in mirrored positions. 1.40 View of the Northwest Area looking southeast. In the centre foreground is walls K (Phase 8), with wall N (Phase 10) to its right and wall V (Phase 2) running diagonally at a lower level. 1.41 Circular Structure 87, looking southwest along wall B. 1.42 State plan of the Northwest Area indicating the features of phases 12, 13 and 14 and proposed outlines of temples 1 and 2. 1.43 Isometric rendering of the Northwest Area, showing Temple 1 remains with column bases b and c, and wall ZZ.

7

The Minnesota Pylos Project, 1990–98 1.44 Aerial view of Room 82 and buildings to the southwest, with lines indicating the proposed southeast f lank wall of Temple, crosswall 83/83, and rear wall 83/84/86. 1.45 Terracotta larnax alongside wall ZZ. 1.46 Aerial mosaic showing features outside the shed roof: the orientation of walls for Temples 1 and 2 and temenos wall ZZ as it crosses Court 88 just outside the lower corner of the shed roof. 1.47 The Southwestern Building and Southwest Quadrant. 1.48 Open space southeast of rooms 64 and 65, showing truncated walls and the square structure at centre right. 1.49 Reconstruction of the second-phase rebuilding of the Southwest Gateway staircase. 1.50 Remains of the staircase from the south, at the level of the bottom step. 1.51 From the top of the staircase looking down onto the plaster ramp. 1.52 The staircase from the south. 1.53 LH IIIB stairway at the south side of Building X. 1.54 Rooms 89 and 90 looking west. 1.55 The excavated cavity of Blegen’s ‘chasm’, looking northwest. 1.56 State plan of the chasm. 1.57. Elevation drawings of scarps on either side of the chasm. 1.58 Floor-level view of the chasm looking northwest. 1.59 View of the northeast scarp of the chasm, showing part of an angular stone below the LH IIIB plaster floor. 1.60 State plan of rooms 54–57. 1.61 Wall of squared poros blocks (wall y) at the southeast end of Room 57, looking southwest. 1.62 Later Building phases 4 and 5 in rooms 54–57. 1.63 Staircase in Room 54, rising into Room 55. 1.64 Linear B tablet fragments recovered in 1991; Xn 1481 at left. 1.65 Probable Linear B tablet fragments recovered in 1992 (a) and 1993 (b and c). 1.66 Linear B tablet fragments recovered from the Blegen dump. 1.67 Fresco fragment from the Griffin and Lion fresco, Room 46. 1.68-1.69 Fresco fragments compacted to form a f loor in Area 106 (top); detail of fragments on the f loor surface (bottom). 1.70 Raw digital terrain model (DTM) of the Englianos ridge with contours at 1 metre intervals. 1.71 Exhibition model of the Englianos ridge with aerial photograph and CAD plan of the palace draped over surface TIN depicting the local topography. 1.72 Coarse surface model of the Englianos ridge and surrounding topography generated using HAGS 1:50,000 map 20-metre contours and details digitised from the palace survey. 1.73 Northeastern section of the aqueduct.

1.74 Northwest Area and Southwestern Building, showing wall ZZ in Court 88. 1.75 Ramp 91 and the Northeast Building. 1.76 Northwest end of the Northeast Building, showing sections B and C of the aqueduct and Cistern 102. 1.77 Area 103 in the Northeast Area. 1.78 Southwest Quadrant. 1.79 Southwest Quadrant, showing Building X. 1.80 Southwest Quadrant, with ‘altar’ to the southwest of Court 63. 1.81–1.82 Two views of the Southwest Quadrant. 1.83 Composite image of the Southwest Quadrant. 1.84 Composite image of the southwest side of the site. 1.85 Composite image of the Southwestern Building. 1.86 Composite image of the Northwest Area and the Southwestern Building. 1.87 Composite image of rooms 89–90 and the Northwest Area. 1.88–1.89 Two views of the Northeast Area, showing the Northeast Building and Area 103. 1.90 View of the Northeast Building and Southwest Quadrant from above the Main Building. 1.91 High-level image of the palace and the Englianos ridge among the surrounding olive groves. 1.92–1.95 Oblique views of the southwest side of the site. 1.96 Oblique view of the Northwest and Southwest areas. 1.97 Oblique view of Area 103. 2. Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LHIIIB Palaces 2.1. Schmatic plan of the drains beneath the Palace of Nestor, showing direction of water f low. 2.2 Plan of the drainage system in the Domestic Quarter at Knossos. 2.3 Northwest-to-southeast cross-section showing top and bottom elevations at the intercepts of feeder drains and intakes. 2.4 Pierced inlet to the Main Drain, south of Room 7. 2.5 Visible end of pressure pipe in northeast wall of Court 42. 2.6 Trenches MZ40 and MZ36, north of the Wine Magazine. Other features in black are Phase 1. 2.7 Proposed Middle Helladic Palace A, showing drains, central court, staircases to higher levels, possible lightwells along the southwest side, and positions of column bases in Area 103. 2.8 Marion Rawson’s sounding in Corridor 26, looking north; the Middle Helladic wall can be seen lying diagonally at the centre top of the photograph; the two f lat stones visible here are those labelled by Rawson H1 and H2 . 2.9 Superimposition of Palaces A and B on the plan of the later LH IIIB palace. 2.10 Proposed Late Helladic I–II Palace B, showing drains, central court, magazines in the later rooms 55–57, and outer courts.

8

Figures 2.11 Ashlars below the later LH IIIB wall in Corridor 26. 2.13 Plan of the northern part of the Southwestern Building, including Room 76 and drains 76/82 and 78/82. 2.14 Aerial view of drains within the corridor between Building 82 and Rooms 76 and 78 of the Southwest Building. 2.15 Aerial view of the drain and cover stones outside the southwest corner of Building 82. 2.16 The magazine-like spaces of the Northwest Area, looking southwest. 2.17 State plan of Room 65 of the Southwest Building, showing two phases of construction on the northwest wall. 2.18 Drains in the final phase of the LH IIIB palace.

5.3 Quarry Ia from northwest. 5.4 Quarry Ib from south. 5.5 Platform above Quarry 1a, from north. 5.6 ‘Staircase’ in Group I, from west. 5.7 Group II quarries, from southwest. 5.8 Workings at Group II quarries, from northwest. 5.9 Quarry IIc, from north. 5.10 Benching in Quarry IIa, from southeast. 5.11 Benching in Quarry IId, from southwest. 5.12 Ledge in Quarry IIIc, from northwest. 5.13 Workings in Quarry IIIc: ceiling and rear wall. 5.14–15 Niches in Quarry IIIc, from west. 5.16 Platform at Group III quarries, from southeast. 5.17 Cut block from Group III quarries, in situ. 6. Discarded Chipped Stone from the Palace of Nestor Obsidian 1877. Crested blade with usewear. 2348. Mesial segment of crested blade. 2351. Distal blade fragment of crested blade. 2383. Crested blade with usewear. 2405. Retouched distal segment of crested blade. 2392. Retouched distal segment of crested blade. 1861. Proximal segment of prismatic blade. 2378. Mesial segment of prismatic blade. 2407. Mesial segment of prismatic blade. 2338. Biface with usewear. 2386. Biface. 2388. Biface. 2403. Biface. 2346. Retouched blade with usewear. 1849. Pressure flake. 1851. Percussion flake. 1862. Percussion flake. 1864. Percussion flake. 1867. Percussion flake. 1869. Retouched percussion flake. 1876. Chip. 1878. Retouched flake (bifacial). 2339. Retouched flake. 2340. Pressure flake. 2345. Pressure flake. 2354. Chip. 2360. Pressure flake. 2362. Chip. 2382. Pressure flake. 2396. Percussion flake. 2399. Pressure flake with possible retouch. 1857. Prismatic core. 2367. Core. 2356. Cobble. 2400. Cobble. 1872. Burin. Cherts 2373. Crested blade with retouch. 2387. Crested blade. 2391. Crested blade. 2395. Proximal segment of a crested blade. 2357. Prismatic blade.

3. Modelling Movement and Use Patterns within the Palace: A GIS/SPACE Syntax Approach 3.1 Schematic plan of the Palace of Nestor at Englianos, showing 1-metre contours. 3.2 Example rooms showing network maps and justified access maps. 3.3 Schematic plan of the Palace of Nestor showing Late Helladic IIIA to IIIB periods. 3.4 Late Helladic IIIB network. 3.5 Late Helladic IIIB distribution of controlling spaces. 3.6 Late Helladic IIIB network rings and RA distributions. 3.7 Late Helladic IIIB relative ringiness distributions. 3.8 Later Late Helladic IIIB network rings and RA distributions 3.9 Later Late Helladic IIIB relative ringiness distributions. 3.10 Later Late Helladic IIIB distribution of controlling spaces. 4. Enclosed Gardens in Courts 42 and 47 4.1 Plan showing courts 42 and 47 and their relationship to the main palace building. 4.2 Schmatic plan of the courts, showing the position of holes. 4.3 Row of four holes in Court 47, looking northeast. 4.4 Hole 47-2. 4.5 Hole 47-4. 4.6 Plans and elevations of ten of the holes in courts 42 and 47. 4.7. Courtyard 47 looking northeast, with modern planting in the Bronze Age holes. 4.8. Court 47 with modern f lowerpots, looking northeast. 4.9. Terracotta pipe in the northeast wall of Court 42. 4.10 Fresco fragments showing possible f lora. 5. Bronze Age Quarrying: A Provenance Study 5.1 Plate tectonic scheme for the Aegean region (after Underhill 1989); cross-section of subduction scheme for the Aegean region. 5.2 Angularia at the Palace of Nestor.

9

The Minnesota Pylos Project, 1990–98 2372. Prismatic blade. 2379. Prismatic blade. 2401. Distal segment of prismatic blade. 2341. Sickle element. 2364. Sickle element. 2366. Sickle element. 2369. Sickle element. 2370. Sickle element. 2377. Sickle element. 2380. Sickle element. 2389. Sickle element. 2404. Sickle element. 2408. Sickle element. 1868. Sickle element or saw blade(?). 2350. Biface. 3449. Biface. 1866. Projectile point pre-form (?). 2344. Barbed projectile point. 2352. Barbed projectile point. 2358. Barbed projectile point. 1871. Retouched and polished blade. 1850. Percussion flake. 1853. Percussion flake. 1856. Percussion flake. 1859. Percussion flake. 1863. Flake with usewear. 2349. Percussion flake or chisel bit (?) 2355. Percussion flake. 2359. Percussion flake. 2368. Flake with retouch and usewear. 2371. Conchoidally detached flake. 2374. Percussion flake. 2402. Retouched flake. 2343. Prismatic core. 2361. Core. 2376. Core. 2390. Core. 2394. Prismatic core. 2398. Prismatic core. 3445. Prismatic core. 2406. Corticular cobble flake. Andesite 1858. Retouched flake. Jasper 2353. Sickle element. Chalcedony (?) 2365. Pressure flake. Unknown stone 1873. Retouched flake. 2363. Flake.

7.5 Heads of a female figurines: a) 4959; b) 4916. 7.6 Animal figurine fragments: a) 4911; b) 4912; c) 4914. 7.7 Fired clay whorls and spools: a) 4909; b) 4921; c) 4935; d) 4936; e) 4910; f) 4907. 7.8 Possible themiaterion, 4905. 7.9 Organic remains: a) Cerithium vulgatum 4978; b) Pinna nobilis 4977; c) Monodonta turbinata 4943; d) boar’s tusk 4949; e) astragaloi (knuckle bones) 4913, 4923, 4919. 8. Post-Bronze Age Architecture and Stratigraphy 8.1. Evidence of post-Bronze Age activity across the site. 8.2. Post-Bronze Age finds identified in the Main Building and on the western side of the site. 8.3. Late material from the east side of the palace. 8.4. Plan of the Northwest Area showing Kittredge and Papathanasopoulos trenches. 8.5. Detailed plan of Papathanosopoulos trenches. 8.6. Northwest Area finds and charcoal concentrations. 8.7. Plan of Papathanasopoulos’ trenches, showing spread of stones, tiles and other terracotta remains in Trench 11 (1958). 8.8 Lakonian pan tile embedded in later wall, Trench 11. 8.9. Plan of Kittredge’s 1962 trenches, showing charcoal concentrations. 8.10. Plan of Kittredge’s 1962 trenches, showing concentrations of mudbrick. 8.11 Structure 87, from the southwest. 8.12. Elevations of Circular Structure 87. 8.13. Remains of the early Archaic temple and Circular Structure 87. 8.14. Elevations of wall 85/87. 8.15 Archaic Corinthian pan tile fragments. 8.16 Reconstruction of complete Archaic Corinthian pan tile. 8.17 Fragmentary Archaic Corinthian cover tiles for a hipped roof (right), and reconstructions (left) showing their positions on the roof. 8.18 Fragmentary cover tile for a hipped roof (ID 03823). 8.19 Fragments of Archaic Lakonian roof tiles. 8.20 Fragment of an Archaic Lakonian lateral antefix (ID 05529). 8.21 Reconstruction showing position of fragmentary antefix. 8.22 Acroteria fragment with torus moulding (02262). 8.23 Acroteria fragment with shallow moulding in concentric rings (ID 00191). 8.24. Northwest Area building phases. 8.25 Reconstruction showing shallow relief. 8.26 Fragments of Archaic Corinthian-style rooftiles. 8.27 Bottom of medieval tile showing pitting. 8.28 Reconstruction of Type I medieval pan and cover tiles. 8.29 Type I tiles showing finger-stripe impression.

7. Small Finds from the 1990–98 Excavations 7.1 Bronze small finds: a) nail 4967; b) nail head 4917; c) pin 4908; d) awl? 4915. 7.2 Iron small finds: a) chisel 4926; b) arrowhead 4941; c) ring 4971; d) nail 4906. 7.3 Lead small finds: possible fishing weight 4927. 7.4 Stone small finds: a) button or conulus 4901; b) mould fragment 4980; c) weight 4922.

10

Figures 8.30 Cover tile showing finger-stripe impression. 8.31 Reconstruction of cover tile with ogive profile. 8.32 Type II medieval tiles. 8.33 Reconstruction of a Type II medieval tile.

3.8 Southwest facade of Southwest Building (wall 17L), section 10, from the southeast. 3.9 Plaster column ring, Main Building, Room 1. 3.10 Column base 64NW, Southwestern Building, Room 64. 3.11 Column base 106SW. 3.12 Measuring and defining ashlar blocks. 3.13 Orthostate block from east side of court to the south of the megaron, Tiryns. 3.14 Elevation of the Treasury of Atreus, Mycenae, showing north dromos wall, south face (after Wace 1949, fig. 5). 3.15 Elevation of the Tomb of the Genii, Mycenae, west dromos wall, east face (after Wace 1921–23, pl. LX). 3.16 Random-range ashlar masonry. 3.17 Ashlar block 33, northeast facade of Main Building (wall 11L). 3.18 Ashlar block 35, northeast facade of Main Building (wall 11L). 3.19 Ashlar block 49, northeast facade (wall 11L) of Main Building. 3.20 Doorway 4–5, rooms 4 and 5, Main Building (after PN I, fig. 430). 3.21 Anta 5D.a, northeast anta of doorway 1–2, rooms 1 and 2, Main Building, from the southwest. 3.22 Anta 5D.b, southwest anta of doorway 1–2, rooms 1 and 2, Main Building, from the northeast. 3.23 Southwest facade (wall 18L) of Southwest Building, from the southeast. 3.24 Elevation, southwest facade of Building X, Southwest Quadrant. 3.25 Southwest facade of Building X, Southwest Quadrant, from the southwest. 3.26 Plan and elevation (from northeast), orthostate wall beneath Room 7, Main Building (detail). 3.27 Early orthostate wall beneath Room 7, Main Building, from northeast. 3.28 Early orthostate block, section 7, southwest facade of Southwest Building, from northeast. 3.29 Another early orthostate block from the same section of wall. 3.30 Plan and elevation of orthostate wall northeast of Room 34, Main Building (detail). 3.31 Anta 11L-20 in Court 42, northeast facade of Main Building (wall 11L), from the northeast. 3.32–3.35 Plan and elevation of the northeast facade of the main building (wall 11L), from Room 46 to Room 32. 3.36 So-called ‘weeper’ outside Room 32, northeast facade (wall 11L) of Main Building, from the northeast. 3.37 Plan and elevation of southwest wall (wall 5E) of Court 3, Main Building. 3.38 Ashlar block 5E-5, southwest wall of Court 3 (wall 5E), Main Building. 3.39 Plan of the northeast facade (wall 18L) of the Southwestern Building. 3.40 Plan of southwest facade (wall 17L) of Southwestern Building, sections 9 and 10.

9. Post-Bronze Age Pottery 9.1 Dark Age I sherds. 9.2 Dark Age II sherds. 9.3 Dark Age III sherds. 9.4 Dark Age through early Archaic sherds. 9.5 Late Archaic, Classical or Hellenistic blackslip sherds. 9.6 Late Classical or Hellenistic lamp. 9.7 Hellenistic or Roman sherds. 9.8 Medieval sherds. 10. Post-Bronze Age Industrial Waste and Bronze Casting 10.1 Triple-layer terracotta fragments, showing the fine, slip-like layer. 10.2 Double-layer terracotta fragment. 10.3 Double-layer terracotta fragment showing the porous, spongy main layer. 10.4 Pithos fragment exhibiting blackening to surface and interior fabric. 10.5–7 Fragments forming possible forearm mould, showing the shape of two fingers and possible drapery mould. 10.8–9 Examples of other joining terracotta fragments. PART II 1. The Palace at Englianos 1.1 Mycenae palace plan (after Wace 1921–23, pl. II). 1.2 Tiryns palace plan (after Muller 1930, vol. III, pl. 4). 1.3 Circulation in the main building. 1.4 Doorway of Room 41 showing column or pier base and stone threshold. 1.5 Plan of courts 42 and 47, showing drainage holes. 1.6 Kiln: plan (PN III, fig. 308); central tongue, south face; central tongue, north face. 1.7 Mudbricks of central tongue of kiln, north face. 3. Building Methods 3.1 Plan and elevation, wall 10C, Main Building (detail). 3.2 Elevation, southwest facade of Southwest Building (wall 17L), sections 1, 2 and 3. 3.3 Elevation, northwest facade of Southwest Building (wall SW39). 3.4 Elevation, southwest facade of Southwest Building (wall SW52). 3.5 Elevation, southwest facade of Southwest Building (wall 17L), section 9. 3.6 Southwest facade of Southwest Building (wall 17L), section 9, from the southeast. 3.7 Elevation, southwest facade of Southwest Building (wall 17L), section 10.

11

The Minnesota Pylos Project, 1990–98 3.41 Northeast facade (wall 11L) at Room 32 of Main Building, from the southeast. 3.42 Reconstructed xylodesia building system (after PN I, 78–79). 3.43 Reconstruction of fresco decoration, northeast wall of Room 64 (after PN II, ID64, 39–38C64, 22H64 and 1F2). 3.44 Chase 36, northeast wall (wall 8L) of Room 6, Main Building (southwest face at norwest end), from the southwest. 3.45 Schematic pier and chase plan, Main Building. 3.46 Chases 38 and 39, and pier 37, in northwest wall (wall 5A) of Room 6, Main Building, from the southeast. 3.47 Northwest wall (wall 5A) of Room 6, Main Building, from the southwest. 3.48 Pre-excavation topography over the Main Building (contour elevations shown as metres above sea level). 3.49 Southwest face of southwest wall (wall 4L) of Room 6, Main Building, from the southwest. 3.50 Southwest face of southwest wall (wall 4L) of Room 6 at Room 18, from the southwest. 3.51 Southwest face of the northeast wall (wall 8L) of Room 6, Main Building, from the southwest. 3.52 Northeast section of northwest wall (wall 5C.1) of Room 4, Main Building, from the northeast. 3.53 Northwest wall (wall 5A) of Room 6, Main Building, from the southwest. 3.54 Pier-wall construction method. 3.55 Wall and pier designations, Room 60. 3.56 Room 60, from the northwest. 3.57 Northeast face of southwest wall of Room 60, at southeast end, from the northeast. 3.58 West wall of ramp at Mycenae (after Wace 1949, fig. 24b). 3.59 Wall of South House at Mycenae (after Wace 1949, fig. 24a). 3.60 Stairway 36 in 1991, from the southwest. 3.61 Southwest face of northeast wall of Court 42. 3.62 Southwest face of northeast wall of Court 47.

A.3 Branch C1 of the acqueduct, showing terracotta tile T3, from the northeast. A.4 Profi les of terracotta tiles of branch C1 of the acqueduct at the Northeast Building. Appendix B: Clamp Cuttings in Ashlar Blocks B.1 Dovetail clamp cuttings at Englianos. B.2 Ashlar block with dovetail clamp cutting in the Southwest Quadrant. State Plans I Topographic plan of the palace and ridge at Englianos. II Key Plan: Room and area sesignations. III Key Plan: Wall labels and designations. IV Key Plan: Antae labels. V Key Play: Column bases. VI Main Building VII Southwest side of the Main Building with Room 90 and wall ZZ. VIII Southeast side of the Main Building, including the entrance and Archive Complex. IX East-northeast side of the Main Building, including rooms 54–56. X North-northeast side of the Main Building, including Corridor 26. XI Main Building with detail of the Throne Room. XII The chasm below rooms 7 and 8. XIII Rooms 54–57. XIV Southwestern Building and Southwest Quadrant. XV Southwest Quadrant and Building X. XVI Southwest Quadrant. XVII Southwest Quadrant, including Court 63 and Room 60. XVIII Southwestern Building, including Room 65. XIX Southwestern Building, with courts 64 and 88. XX Northwest section of the Southwestern Building. XXI Northwest Area. XXII Northeast Area. XXIII Northeast Area, including the Wine Magazine and areas 103 and 106. XXIV Northeast Area, including areas 103 and 106, and Cistern 102. XXV Northeast Area, including the Wine Magazine and areas 103 and 106. XXVI Northeast Building. XXVII Northeast Building and Courts 42 and 47. XXVIII Northeast Area, including the Wine Magazine and the MZ trenches in Area 106. XXIX Northeast Area, including the aqueduct, Northeast Gateway and Belevedere area. XXX Western section of the aqueduct, including drains in areas 103 and 101. XXXI Eastern section of the aqueduct. XXXII Northeast Gateway and the kiln. XXXIII Northeast Gateway. XXXIV Belevedere Area.

4. Building History 4.1 Early Late Helladic I phase plan. 4.2 Late Helladic I–II phase plan. 4.3 Early orthostate walls and rubble walls beneath rooms 55–57, Main Building. 4.4 Late Helladic IIIA phase plan. 4.5 Reconstruction of stairway in Southwest Quadrant. 4.6 Buildings and central court in proposed reconstruction of Palace of Nestor by K. Kilian (1985). 4.7 Early Late Helladic IIIB phase plan. 4.8 Late Late Helladic IIIB phase plan. Appendix A: Terracotta Building Elements A.1 Plan of branch C1 of the acqueduct at the Northeast Building, showing position of terracotta tiles. A.2 Branch CI of the acqueduct, showing terracotta tiles, from the north.

12

TABLES, GRAPHS AND MAPS

PART I

Graph 5.3 Mean values for silicon and standard deviations for each sample. Graph 5.4 Delta-13C vs. d-18O plot for limestone samples; note separate fields for samples 1–7 and 8–12. Table 5.4 Stable isotope data for 12 collected limestone samples per mil, relative to pbd.

1. The Minnesota Pylos Project: Investigations and Results, 1990–98 Table 1.1 List of stratigraphic sections, soil colours and diagnostic pottery as recorded by William Kittredge. Table 1.2 Correlation of architectural features in the Northwest Area with MARWP phasing, Blegen’s strata, and original Kittredge and Papathanasopoulos trenches. Table 1.3 Proposed building phases in the area of rooms 1, 2, 7 and 8. Table 1.4 Proposed building phases in the area of rooms 1, 2, 7 and 8. Table 1.5 Results of an examination of terrace hedgerows on the Englianos ridge. Table 1.6 Results of scientific analyses carried out on materials collected by Blegen and MARWP.

PART II 1. The Palace at Englianos Table 1.1 Published dates for the buildings at Englianos. 2. Building Materials Table 2.1 Analysed wood samples from the palace at Englianos. 3. Building Methods Table 3.1 Column base dimensions. Table 3.2 Bronze chisels and chisel fragments recovered in the original excavations. Table 3.3 Anta blocks. Table 3.4 Details of mortises, northeast facade (wall 11L) of Main Building. Table 3.5 Details of mortises, southwest wall (wall 5E) of court 3, Main Building. Table 3.6 Details of mortises, northeast facade (wall 18L) of Southwestern Building. Table 3.7 Dimensions of ashlar blocks in wall 17L, sections 9 (above) and 10 (below). Table 3.8 Pier and chase elevations, wall 8L, Room 6, Main Building. Table 3.9 Pier and chase elevations, wall 5B, Room 6, Main Building. Table 3.10 Analysis of soil samples from Tiryns (Müller 1930, 178–80). Table 3.11 Pier dimensions of wall 4L. Table 3.12 Pier dimensions of wall 5B. Table 3.13 Pier dimensions of wall 5A.

5. Bronze Age Quarrying: A Provenance Study Map 5.1 Englianos area. Portion of Filiatra sheet, IGME 1:50,000, 1980. Map 5.2 Portion of Filiatra sheet, IGME 1:50,000, 1980. Map 5.3 Region around Gargaliani, after Hellenic Geographical Survey, portion of Filiatra sheet, 1977. Map 5.4 Paleographic reconstruction of the region surrounding the Bay of Navarino from the Mesolithic to modern times (after Kraft et al. 1980); shoreline and drainage system of the Helladic period. Map 5.5 Petrographic sample locations. Table 5.1 Results of petrographic study. Table 5.2 Results of paleontological study. Table 5.3 Standard microfaces assignments. Graph 5.1 Mean values for calcium and standard deviation for each sample. Graph 5.2 Mean values for magnesium and standard deviation for each sample.

13

The Minnesota Pylos Project, 1990–98 ABBREVIATIONS

Table 3.14 Pier dimensions of wall 8L. Table 3.15 Pier dimensions of wall 10J. Table 3.16 Pier lengths of wall 9L at stoa 44. Table 3.17 Pier lengths of wall 9L at rooms 30 and 32. Table 3.18 Pier lengths of wall SW42 between rooms 76 and 78. Table 3.19 Pier lengths of wall SW43 between rooms 71 and 76. Table 3.20 Pier lengths of wall SW46 between rooms 71 and 73. Table 3.21 Pier lengths of walls P1–P4, Room 60.

AJA = American Journal of Archaeology ASAtene = Annuario delle scuola archeologica di Atene et delle missioni italiani in Oriente ASCSA = American School of Classical Studies at Athens BSA = British School of Archaeology MARWP = Minnesota Archaeological Researches in the Western Peloponnese

Appendix A: Terracotta Building Elements Table A.5 Dimensions of terracotta drain tiles, branch C1 of the acqueduct.

masl = metres above sea level

Appendix C: Analysis of Earth Material Table C.1 Earth material samples. Table C.2 Approximate calcium carbonate content of earth material samples. Table C.3 Multi-element data results from earth material samples: Batch 74, with amounts shown at mg/kg (ppm w/w). Table C.4 Multi-element data results from earth material samples: Batch 124, with amounts shown as mg/kg (ppm w/w).

mm, cm, m = millimetres, centimetres, metres PN I = Blegen, C.W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia I: The Buildings and their Contents. Princeton University Press, Princeton, NJ. PN II = Lang, M. 1969. The Palace of Nestor at Pylos in Western Messenia II: The Frescoes. Princeton University Press, Princeton, NJ. PN III = Blegen, C.W., Rawson, M., Taylour, L.W., and Donovan, W.P. 1973. The Palace of Nestor at Pylos in Western Messenia, III: Acropolis and Lower Town, Tholoi, Grave Circle, and Chamber Tombs, Discoveries outside the Citadel. Princeton University Press, Princeton, NJ. PRAP = Pylos Regional Archaeological Project UTM = Universal Transverse Mercatur

14

CONTRIBUTORS

Anne B. Hollond received her MA in Classical and Near Eastern Studies from the University of Minnesota.

Todd M. Brenningmeyer is Associate Professor of Art History and Director of the Art History Program at Maryville University, St Louis, Missouri.

Eleni M. Konstantinidi-Syvridi is Curator in the Department of Prehistoric, Egyptian, Cypriot and Eastern Antiquities at the National Archaeological Museum, Athens. She received her PhD in Mycenaean archaeology from the University of Birmingham, UK.

Frederick A. Cooper (1936–2011) was Director of Minnesota Archaeological Researches in the Western Peloponnese. He was the Morse Alumni Distinguished Professor of Art History at the University of Minnesota and served as the Andrew F. Mellon Professor of Classical Studies at the American School of Classical Studies at Athens, among other posts.

George Otto Marquardt received his BA in Anthropology from the University of Minnesota. Michael C. Nelson is Associate Professor of Art History at Queens College, City University of New York, and Co-Director and Architect for excavations at Omrit in Israel.

Joshua N. Distler is Project Architect at Mithun Architecture and Planning, Seattle, Washington. He received his MS in Geology from the University of Montana and his M.Arch from the University of Washington.

Shawn A. Ross is Associate Professor of History and Archaeology at Macquarie University, Sydney. He also serves as Director of the Field Acquired Information Management Systems Project, as Co-Supervisor of the Tundzha Regional Archaeology Project in Bulgaria, and as Deputy Director of the Big History Institute at Macquarie University.

Caitlin Downey received her PhD in Classical and Near Eastern Studies from the University of Minnesota. Diane Fortenberry is Senior Editor at Phaidon Press Ltd, London. She holds a PhD in Bronze Age Aegean Archaeology from the University of Cincinnati.

15

PREFACE AND ACKNOWLEDGEMENTS

the palace and the outlying buildings also suggests a more complicated chronology than Blegen understood. Deservedly, Blegen accorded special acknowledgement to Demetrias Theocharis, who in 1952 laid the groundwork for season-by-season plans thereafter. As explained in further detail below (pp.31–35), Theocharis’ plan, schematic but correctly drawn at scale and in detail, became our working plan for the master survey, facilitating a plot of our survey traverse lines, the underlay of walls for a plot of the UTM north–south, 5-metre grid over the site, and the positioning for the numbering of zones keyed to large-scale plots of walls. Demetrias Theocharis deserves a special note of recognition.

‘Wednesday July 4. Fine cool morning ... Visitors: Fritz Cooper of Yale and Pittsburgh’. In 1962 Carl Blegen jotted this into his day book (1962, p.103) and conducted me on a memorable tour of the excavations of the palace at Englianos, for which I continue to be grateful. Blegen’s interpretations were complemented by two new features on the site: a protective shed roof that recreated somewhat the size and volumes of the original palace had recently been erected, and A Guide to the Palace of Nestor (University of Cincinnati, 1962) was newly available, with a numbered plan and room-by-room descriptions; note-taking was unnecessary. A strong geometry to the published layout suggested that construction of the palace followed a single-period, Bronze Age master plan, a basic tenet to Blegen’s conception of the architectural history of the building.1

In 1960 John Travlos set a plug in the restored southwest corner of the palace, off Room 7. The benchmark was fi xed at an assumed datum of 13.00 m. Travlos’ site plan for that year is the only one with spot elevations, six within the confines of the Main Building; most are pairs at top and bottom of slope, taken around the ring of the Englianos acropolis. A pre-excavation topographic survey of the Englianos hill would have been an invaluable document. Our own actual state plans supply top and bottom wall elevations; only in this way does the fact emerge with mathematical certainty that preserved wall heights decline at a graduated slope in all directions, with a highpoint at 193.68 m on top of the common wall dividing rooms 4 and 35. It is reasonable to assume that the final reduction of the walls of the major buildings on Englianos to their present height was caused by natural or geotechnical action; that is, not by human activity, but by erosion.

Four years after my first visit to Pylos, the first volume of The Palace of Nestor was published (Princeton University Press, Princeton, NJ, 1966), the Key Plan folded into the back of Part I, the text fascicle. At a scale of 1:309, it is an enlarged and revised version of the earlier Guide plan, which was at 1:840. This undimensioned Key Plan has an accuracy no greater than 25 centimetres. A scaled measurement across the thickness of a wall on this plan thus yields an approximate dimension of +/- 0.25 m: a wall of 0.85 m has a scaled tolerance anywhere from 0.65 m to 0.90 m, reducing differing wall widths on the plan to an approximately equal dimension, which they are not. A heterogeneous construction technique throughout Rhys Carpenter, Mellon Professor at the University of Pittsburgh in 1961–62, encouraged that 1962 venture and extended many kindnesses. Two years later I went to Greece as a PhD candidate to create a state plan of the temple at Bassai and to record all dismembered blocks in a comprehensive and objective aggregate of dimensioned drawings that permitted an indelible reconstruction (F. A. Cooper, The Temple at Apollo Basitas, vols I and III (Princeton University Press, Princeton, NJ, 1996)).

Blegen left undone an integrated and comprehensive actual state plan; such an undertaking was simply not folded into his archaeological programme. That said, two members of Blegen’s excavation team did draw masterful actual state plans of features in their respective trenches: Georgos Papathanasopoulis (Room 27; The Palace of Nestor (hereafter PN) I, fig. 421) and David French (Wine

17

Cooper – Preface and Acknowledgements For the architectural and cultural historian, it is essential to understand the conceptual design of any building. This master plan ‘serves as the guide and philosophy for the remainder of the development of the project or for phasing, should the project be constructed in various phases or of different components’ (ibid., 2.10).

Magazine; PN I, fig. 428). Papathanasopoulis also drew a correct, rock-by-rock plan of the so-called chasm, never published. Hero Athanasiades and Piet de Jong produced ‘simplified’ state plans; de Jong’s drawings of doorways is discussed below. The omission of an actual state plan continues to be common practice among archaeologists, partly due to circumscribed resources and the dearth of engineering expertise on excavation staffs. Jesse Fant, in Excavations at Nichoria in Southwest Greece I (eds G. Rapp and S. E. Aschenbrenner (University of Minnesota Press, Minneapolis, 1978), 140–41), explains the tactic taken by the Minnesota Messenia Expedition (MME): ‘The common archaeological practice of assigning to an “architect” the measurement and drawing necessary to complete such an undertaking seemed to us too timeconsuming and interruptive of on-site operations.’

At a meeting with his eventual clients during the architectural competition phase, Philip Johnson sketched a design for the skyscraper that would become the IDS Building in Minneapolis. Once awarded the commission, Johnson’s firm undertook the next phase, the Schematic Design phase, and Johnson discarded his scribbled conceptual design (personal communication, September 1975). In the early 1980s, Frank Ghery laid out his ideas for the University of Minnesota Weisman Museum at a committee luncheon. He supplemented an oral presentation with a ballpoint sketch on a paper napkin and won the approval of the committee. In this case the napkin and other ephemeral sketches were preserved, providing an inkling of the essence of a conceptual design by a leading twentieth-century architect.

MME was not alone among archaeological projects in the use of photogrammetry using single-field and stereoscopic cameras, sometimes mounted on bi-pods, for surveying and mapping trenches and site activity. Elevations are interpolated and are often insufficiently reliable to extrapolate wall heights or trench depths. Photogrammetry has much to recommend it – but as a supplement, not a replacement, for the drawn archaeological record. In a wall built of fieldstones, for example, the plan outline of an irregularly shaped rock is correct only at nadir, becoming increasingly distorted towards the edge of the field of view. The convolutions can be normalised with specialised lenses and soft ware, but even then, ground truth requires a return to the wall for a visual confirmation of the accuracy. This post-processing far outstrips in time and specialisation any expenditure by a practiced draftsperson in the field. Moreover, crucial particulars of construction – wall joints, for example, which are either bonded or butted – are too often masked by the camera.2 The speed, efficiency and accuracy of a large-scale, stone-by-stone plot of a wall or detail of a feature has no substitute.

The design for the palace at Englianos was perhaps, in a similar way, a manifestation of a geometrical conception. The beauty in the design of the palace can only be imagined from the remains, but its architectural harmony was instilled by geometry. The processes of the design phases can be reconstructed by working backwards from the last phase, a supposition as true for ancient monuments as for modern ones. For the palace at Englianos, three features in particular must reflect a conceptual design. First, the walls of the megaron block form a rectangular parallelogram. This geometry is obtainable only by the knowledge (not necessarily the theorem or the proof) that inscribed equal diagonals through a parallelogram (opposite sides are equal) make all inside vertices right angles. The outline of the LH IIIB palace megaron was surely staked out by direct straight-line measurements beginning at a fi xed station. An adjustable station diagonally opposite is moved by trial and error until both straight-line measurements of the two diagonals are equal within a very small tolerance. Generally speaking, in a final construction, the difference is imperceptible between an eye-balled approximate 90 degrees at wall corners, as in modern Greek house construction, and true right angles. A constituent in the conceptual design at Pylos was a compelling requirement for the perpendicular or the gnomon (in both the sense of a carpenter’s square and of one who knows), the idea perhaps coming from Egypt. This determination as geometric fact is made possible today by a scientific, closed traverse of the wall axes at first order accuracy.

Nevertheless, as scientific excavation becomes increasingly interdisciplinary and increasingly expensive, the question arises: why bother to create an overall actual state plan? An answer may be found in established protocol for modern construction project development. A standard publication, Building Design and Construction Handbook, 6th edn (eds F. S. Merritt and J. T. Ricketts (McGraw-Hill, New York 2001), 2.10–2.11), defines project phases: 1. Conceptual Design 2. Schematic Design 3. Design Development 4. Construction Documents 5. Post-construction Services: record drawings, or ‘as-builts’.

A second element in the concept design of the palace at Pylos is that it was conceived as earth-hugging. The main floor is not only the piano nobile but, exceptionally in terms of Late Bronze Age Aegean architecture, lies flat on

2

Since the mid-1970s, the Greek Archaeological Service has required an actual state plan for any project that involves anastylosis or site development.

18

Acknowledgements the ground, with no shifting of floor levels at the ground or upper floors. The three staircases in the Main Building all rise to a calculated equal height of 3.05 m from bottom step to top.

Doorways emerge earlier, in Minoan neo-palatial architecture, as the ubiquitous polythyron design, and a similar arrangement seems to occur on the mainland at Gla. I am convinced the mystery can be solved by a Hercule Poirot assessment of clues and the as-built facts that reflect the Design Development phase of the Main and Southwest buildings. For such an undertaking, the simplified state plans of the doorways prepared by Piet de Jong for the final publication are of primary value, as the de Jong door sill drawings are easier to read than the our actual state plan.

A third conceptual feature of the palace involves doorways as a leitmotif: as Nelson and Brenningmeyer both make clear, someone – architect or client or both – had an obsession with doorways. The main building contains 38 doorways in approximately 1500 square metres. The megaron block, approximately 300 sq. m, has only two, leaving a further 36 doorways, one for every 32 sq. m of space. There are approximately 125 running metres of corridors and surrounding rooms, with one doorway for every 2.4 m, or one doorway every three or four steps. The labyrinthine set of passageways of the Main Building does not carry over to the design of the Southwest Building; there, a profusion of thirteen doorways clusters within the northwest quadrant, rooms 66–68.

I owe Diane Fortenberry more than this note of gratitude for bringing to my attention the challenge of an explanation for the problem of how doorways and corridors were secured, and the potential significance if many of them do, indeed, lock from the inside. The problem is just one example of many derived from the actual state plan that need to be studied in further depth.

Spans to the doorways all fall into a universal standard width of 0.8–1.0 m; cuttings in the sills include doorstops and single or double pivot holes. Four sills in the Southwest Building have no pivot holes, but this might be due to their poor condition. Single-leaf doorways mix with doubleleaf doors. Nowhere are there vertical slots midway along preserved sills for a drop-bar locking system, and jambs are not preserved to a sufficient height to show the presence or absence of cuttings for a horizontal locking bar. There is thus no way to know for certain whether the doors had locks, but doorstop cuttings behind the pivot hole or holes do indicate a horizontal bar lock, if lock there was. The leaf or leaves swing away from the doorstop, folding into a recessed casement, and an assumed horizontal bar lock would have been let into this casement.

Acknowledgements drafting and excavation A trained archaeological ‘architect’, as Minnesota Archaeological Researches in the Western Peloponnese (MARWP) had at Pylos, records particulars along a baseline and makes a visual assessment of each object: position by increment and offset, shape and size. The draftsperson works along the baseline recording the likeness of each detail, looking downwards for a plan or perpendicular for the elevation. The role of the accomplished archaeological draftsperson is that of an artist: training includes a lifeclass in drawing not nudes, but rocks and cut stones, and capturing their personalities.

The door sills in the Main Building and the Southwest Building are arranged so that some doors opened outwards from rooms and passageways, others inwards. Take, for example, the two outer doorways into and through oil magazines 23 and 24, leaving aside for the moment the necessity of the intervening doorway. Both doorways open inwards; to lock both would have required that the person doing so be locked inside. The entrance to Room 23 from Hall 22 can be locked from the inside, and the intervening door from 23 to 24 can be locked from the inside of Room 24, sealing off the one oil magazine, Room 23; but that leaves the other, Room 24, accessible from the outside through its outer doorway.

For an artful and exacting record of the constructions and features at Pylos, credit goes to these archaeological artists: Molly Cole, Caitlin Downey, Helen Foster, Charles Griebel and Michael Nelson. For those pitching in part-time or serving a single season: Phoebe Acheson, Sabine Albersmeier, Pieter Broucke, Eric Brulotte, Heidi Eernise, Jared Erickson, Candis Griggs, Rebecca Holte, Jennifer Hooper, Bronwyn Jackson, Karen Johnson, Heidi Luchsinger, Ted MacLoed, Jill McFarland, Nancy Miller, Katherine Neebe, Vanessa Rousseau, Marissa Schlesinger, David Stewart, Marc Walton, Cindy Weber, Penny Xereas and Racquel Yerbury. Michael Nelson was responsible for the year-by-year creation of the state plan, which was revised in 2009 by Hendrik Hendrickx of the Vrije Universiteit, Brussels. Since 1972 Hendrickx has given unsparingly of his time and expertise on all my projects, as his beautiful drawings testify. His contribution to the Pylos project has been invaluable; a generation of students and I are extremely grateful.

Blegen pauses twice to ‘wonder why there should be so many doorways’ that are ‘mysterious and puzzling’ (PN I, 217 and 276). Nelson offers a single and straightforward explanation: ‘a planned circulation pattern’ (Part II, p.286). Brenningmeyer, in chapter 3 below, tackles the question from the point of view of post-construction modifications to entrances or access points reflecting re-configurations of public and private spaces. Nonetheless, the mixture of egresses and ingresses continued to the last phase.

Over 100 excavators participated in the MARWP Pylos expedition, and each deserves recognition and thanks,

19

Cooper – Preface and Acknowledgements more than I can give here. Michael Nelson served as field director from beginning to end, 1990–98. Principal staff for one or more seasons include Todd Brenningmeyer, Jane Carter, Diane Fortenberry, Charles Griebel, Anne Hollister, Eleni Konstantinidi-Syvridi and Elizabeth Swain, all of whom kept activities running efficiently and effectively.

topographic plan of the englianos hill The physical landscape surrounding the site of Pylos and, more particularly, the physical lay of the land and its resources, played a significant though still-to-be determined role in the architectural history of the site, from its probable Neolithic beginnings through the Frankish period. Blegen published a topographic sketch of the Englianos hill prepared by Jesse Fant and Joseph Shaw, probably from the one readily available source at the time, British Air Force WWII plans. The sketch’s usefulness is limited by an absence of both contour elevations and contour intervals. The Hellenic Army Geodetic Survey (HAGS) 1:5000 contour maps were available to us beginning in 1990; they provided a basis for a wide-ranging stadia and GPS survey of the wider area (pp.87–89, figs 1.70–1.72). MARWP is indebted to the Greek Ministry of Defense for these maps and the later and improved 1:50,000 plans.

Digging is hard work, the drudgery only rewarded by the unexpected solution to a question or problem, or the discovery of an artefact. It is a pleasure to mention all excavators by name. First, those from the University of Minnesota: Judy Allen, Thomas Braun, Brenda Cullen, Bethany Faith, Deborah Fitzl, Wendy Forrester, Joseph Fox, Andrew Hoel, Cheryl Houdek, Peter Hunt, Deborah Jackson, Amy Kakissis, Kathryn Kiffer, Elizabeth Knutson, Karla Kromer, Karla LeTendre, Patricia Libby, George Marquardt, Linda Mierswa-Jacobson, Chris Mulloy, Shawne Osborne, Andrea Plowman, Domini Popielski, Shawn Ross, Corbin Sanft, Edward Silver, Mona Stalekar, Rachel Thompson and Joanna Veith. From other institutions came Thomas Asher, Evangelos Barbatseas, Andre Bekerman, Brendan Burke, Sharon Clapp, Lucia Cosmetico, Lauren Cummings, Joshua Distler, Elizabeth Fauber, Sarah Freeman, Helen Galimanis, Paul Galiotos, Karen Gallucci, Eric Hauxstausen, Adrian Hearn, Judson Heartsill, Andrew Hicks, Matthew Kiskis, Bryon Koukaras, Jonathan Lane, Kyriakos Litsas, Elise Marzluff, John McCarron, Steven Neetens, Colleen Oates, Kristen Pasnak, Victoria Peebles, Katherine Ruddick, Lawrence Smith, Karen Stern, Elizabeth Traglia, Elliott Van Ness, Michael Vito and Tara Walsh.

After the daily closure of the site, our research continued to dusk outside the fence on the acropolis and vicinity, collecting survey data for an eventual digital elevation model (DEM) of the area. All staff participated at one time or another; towards the close of fieldwork, a dedicated team consisting of Jennifer Robeson, Victoria Benjamin and Oliver Benjamin rounded out the taking of spot elevations using GPS equipment and differential correction soft ware read from a base station. funding The multi-year expedition to Pylos operated on a shoestring budget, often sharing resources and personnel with other MARWP projects. The University of Minnesota supplemented grants that came from several different units. On behalf of all Pylos participants, I acknowledge the University of Minnesota Graduate School for a nearannual Grant-in-Aid of Research. Two corollary grants from the Graduate School underwrote the purchase of a first-generation professional Trimble Pathfinder® GPS system and, two years later, the next generation Pathfinder® GPS that included a base station and postprocessing soft ware, which greatly increased accuracy by differential correction. We took satisfaction in the fact that gps mapping, now commonplace in everyday life as well as on scientific expeditions, was then a leap into the future. I extend thanks to the University of Minnesota Graduate School and to those students who had the initiative to learn the technicalities of its operation.

conservation In 1994 C. K. Williams III, Director of the Corinth Excavations, for ten days released Stella Bouzaki from her conservation duties at ancient Corinth. This was high season at Corinth, and there is therefore all the more reason to highlight our gratitude to Charles Williams for his generosity. Stella Bouzaki set up a provisional but professional laboratory at Chora, equipped with microscopes, chemicals and other conservation materials. Heartfelt thanks goes to Bouzaki who, first of all, tackled problematic and urgent objects in need of conservation and, secondly, trained Ginny Gliata and Angela Spyopoulou, who had joined us several years previously as Chora student volunteers. Gliata stayed on after the close of fieldwork and, besides conservation, acquired a host of additional skills and technologies, including database entry, artefact photography, micrometry and drawing. She continued both the conservation and inventory of more than 5000 fresco fragments recovered by the MARWP project. We owe Gliata a special word of appreciation.

An equal expression of recognition goes to the University of Minnesota College of Liberal Arts (CLA), which has an admirable policy of providing faculty with computers. In this case, CLA went beyond the norm and granted, for purposes of the MARWP projects, a succession of two Silicon Graphics® UNIX workstations, which allowed for the high-end post-processing of our gps maps, satellite imagery and gis data. Separately, the college advanced

20

Acknowledgements funds for the purchase of Landsat 5® imagery and a subscription to ERDAS/IMAGINE for image processing and GIS reconnaissance by remote sensing of potential sites and resources. CLA also provided, starting in 1997, an annual subvention to help cover the costs of undergraduate students. My gratitude goes to several associate deans of CLA – Eugene Borgida, Russ Menard, Peter Reed and Barbara Reid – for their roles in securing and maintaining these funds.

The Institute for Aegean Prehistory (INSTAP) awarded six grants between 1990 and 1996. MARWP Pylos Project participants have benefited greatly from the opportunity, experience and results provided by INSTAP funds. A symposium entitled New Contributions to the Art and Archaeology of Bronze Age Classical Greece: The Minnesota Pylos Project launched the final stage of the Pylos project. The conference, held in Minneapolis in March 2001, drew together contributors to the final publication, participants giving papers on related subjects, and respondents. The authors benefited from the exchange of ideas and material and the astute assessments offered by all participants. In particular, I offer special thanks to Diane Fortenberry, Jeremy Rutter, Joseph Shaw and Maria Shaw, who contributed as respondents. The event was made possible with grants from a number of agencies and institutions, to whom we are deeply grateful: the Archaeological Institute of America’s Regional Symposium Grants, the Minnesota Humanities Commission, the Minnesota Historical Society and, at the University of Minnesota, the CLA Scholarly Events Fund, the Office of International Programs and the Nash Gallery.

The McKnight Foundation has endowed a fund to support the University of Minnesota humanities faculty for original research. I was fortunate to receive a three-year grant in 2001 for expenses related to MARWP projects, including Pylos. It enriched my researches, and I am grateful. The Undergraduate Research Opportunities Program (UROP) at the University of Minnesota provides stipends and expenses to undergraduates for original research; awards are made on the basis of a strict proposal format, fashioned after that for faculty project proposals to granting agencies. Twenty Minnesota undergraduate field participants benefited from the UROPs, and several research topics attained a quality to be presented at national conferences or for publication.

other personnel and colleagues

The Classical Civilization Program at the University of Minnesota offers its undergraduate students a vast breadth of disciplines. Archaeology is an underlying component of this programme and thus one of the instigations for MARWP and the appropriate thrust for the Pylos Project. Tom Clayton has been long-standing head of the Classical Civilization Program and has availed the Pylos Project of everything from highly recommended students to funds from a restricted budget. For these I owe Clayton an extra word of thanks.

At the opening stage of the project, Michael Sage, Head of the Department of Classics at the University of Cincinnati, granted access to the Pylos archive held at the Blegen Library there, along with permission to use documents pertinent to our researches (personal communication, 22 April 1991). During an industrious weekend at Cincinnati, Tucker Blackman – an archaeologist, archivist and librarian of unequalled competence and generosity – guided Michael Nelson and me through excavation documents, mostly microfi lms and copies of photographs that have served us through the years of the project. To Sage and Blackman we owe heartfelt thanks.

Within the structure of limited departmental budgets, the Department of Classical and Near Eastern Studies (CNES) contributed in small but significant ways. cnes also had a modest endowment, the Hahn Travel Fund, specifically earmarked for departmental undergraduate travel to excavations abroad. The project gained from the assistance of many student recipients and, on their behalf, acknowledges the former administrator of the Hahn Endowment, Jerry Erickson. Judy Scullin served as Administrative Assistant to CNES until early 2002, and generously stayed at the heart of this archaeological programme, keeping it functioning in all ways and under all conditions. Her friendly gestures touched all of us, and we express a large thank you.

In Greece, much of the success of the Pylos Project is owed to the late William Coulson, Director of the American School of Classical Studies at Athens (ASCSA) during the term of our project. Maria Pilali, the Senior Administrator at the ASCSA, monitored our annual permit submissions and year-end reports, greatly facilitating our work. Deserving many words of gratitude are Zene Arapayianne, Director, and Gia Chatze, Epemilite, of the 6th Prehistoric and Classical Ephoreia at Olympia. From the outset, these two scholars recognised the underlying importance of our programme and advanced the Pylos Project in innumerable ways. The project was fortunate to have a succession of skilled, helpful and cheerful site representatives from the Olympia Ephoreia: Maria Phoutze, Popi Kalogerliou and Litza Malapani. Sakis Manolis is owed thanks for his frequent, skilled and expeditious repairs to the vintage Volkswagen van, new in 1968, which was the soul of the MARWP projects and without which much less would have been achieved.

The University of Minnesota granted me sabbaticals for the years 1995–96 and 2001–02. For the latter period, the College of Liberal Arts awarded a Sabbatical Supplement. The two sabbaticals allowed me to devote uninterrupted time to MARWP projects, in particular the latest stages of research on the Pylos Project and the necessary effort leading to this publication.

21

Cooper – Preface and Acknowledgements Visitors are frequent at digs like this one, and on several occasions we roped in these individuals to take on the odd task. Joseph Alchermes and J. J. Coulton happened by and were enlisted to help build an inverted-bow truss roof over the Northwest Area, a construction mandated by the Archaeological Service. Unfortunately, we were directed to remove this structure when the excavation ended in 1998, and to backfi ll this interesting quadrant, the deep trenches of the area being considered too dangerous to remain open.

us during our work at the site, for all of which we cannot acknowledge him enough. *** Michael Nelson was the principle force behind the foundation of the Pylos project, which evolved from an undergraduate course paper devoted to the physical character of the architecture of the Main Building. His part of this publication bears the fruit of that innocent beginning. Throughout, Nelson has made innumerable invaluable contributions; the project would never have happened but for him, and I add this shout of praise.

At the season’s end in 1994, actual state plans of walls in the Northwest Area had been drawn, but state elevations – a nightmarish complexity of stones – were not. I thank Robin Rhodes for his invaluable service to the project in this respect, and for his irreverent humor, which made the time go quickly and enjoyably.

From the beginning to the very end of the Pylos project, Diane Fortenberry has continued as an active contributor. She has been responsible for bringing this volume to published reality, and her succinct criticisms, often disturbing, have reset points of fact, logic and clarity. Everywhere her reading and analysis has improved this monograph, and we are all indebted to her.

Over 20 tons of excavated artefacts dumped in the Northwest Area during the Blegen years were recovered in 1994 and 1995. Joanna Murphy and Jan Verstraete, graduate students at the University of Cincinnati specialising in Mycenaean pottery, led in the processing of this material during the 1995 season. They more than deserve this word of appreciation.

*** For too long scholars have sought to situate the Bronze Age palace and its grounds within a constructed typology of Bronze Age architecture. Too often, this kind of analysis excludes a richer appreciation of the creative processes inherent in the design of the palace at Pylos. To paraphrase Proust: this archaeological site stands for a civilisation.

The preeminent Linear B scholar, Emmett Bennett Jr., joined us in 1994, 40 years after his first excavation season under Blegen. Bennett returned for two additional seasons; he offered seminars, pitched in with the fieldwork and consistently provided good cheer. He kept the crew ‘from the dangers of thinking anachronistically’ by having them listen to music ‘not from central England or the southern United States, but from the earliest recordings of the folk music of the archaeological region’. With undiminished esteem for Blegen’s accomplishments at Englianos, Bennett rightly questioned many parts of Blegen’s story, and I came to realise that this quietly-spoken scholar is a heretic of Bronze Age orthodoxy, fearless in asking the remains: ‘what if?’. The spirit behind our researches at Pylos reflects Bennett’s dialectic on Bronze Age orthodoxy, and I thank him on behalf of the nearly 60 personnel who enjoyed his company and benefited from his expertise during the three years.

Frederick A. Cooper June 2011

An archaeological field school was not a requisite part of the MARWP projects; instead, all Minnesota students enrolled in a springtime course, Practicum in Archaeological Field Techniques. Robert Clouse, head archaeologist of the Minnesota Historical Society (MHS), co-taught the course. After sessions on engineering survey, drawing, stratigraphic digging and recording, computer applications and other appropriate techniques, the final weeks were spent at prehistoric and historic Minnesota locations under the jurisdiction of the mhs. Over the years, Clouse and I progressively experimented with and developed a protocol appropriate to excavation in both the New and Old worlds. Clouse not only made very tangible contributions to the Pylos project but also consulted with

22

Editorial Note Editorial Note

Helladic period. The finds within the new tomb of the Griffin Warrior provide further evidence for the closest of connections between the builders of this mainland palace and the culture of Bronze Age Crete.

This publication of the MARWP excavations was originally planned for 2004. In the course of preparing the final texts and drawings in the years following the Minnesota Pylos Project symposium held at the University of Minnesota in 2001 – preparations that included reviewing all of the old field notebooks from the Blegen excavations – a number of completely unexpected, very exciting and potentially controversial details came to light. This originally unplanned post-excavation research has enabled us to present here, among other revelations, evidence for the remains of a previously undiscovered Archaic sanctuary with one or two temples in the Northwest Area; Geometric cult areas in the Northwest Area and elsewhere around the Bronze Age building; detailed phasing of the previously un-analysed Northeast and Northwest Areas and the socalled chasm; and – perhaps most interesting – evidence for the proposal of two pre-LH IIIB palaces on the site, based on an analysis of the building’s hydraulics and early walls.

*** Frederick Cooper died on 25 September 2011, in the midst – literally – of preparations for the next in his lifelong series of rigorously conducted and uniquely revealing, if occasionally iconoclastic, archaeological projects. He had read and endorsed each of the chapters in this publication, and he made the last corrections to his own contributions only a few weeks before his death. His seminal and ongoing involvement in the work presented here by former undergraduate and graduate students will be – as he intended – largely invisible to readers; nevertheless, his creative intellect, challenging debate, deep understanding of and love for ancient Greece, and his enormous personal and professional generosity have inspired each of these studies.

The repercussions of this unexpected research and the discoveries it engendered inevitably affected chapters beyond those that focused specifically on these developments, resulting in further deferrals of the publication, but it is hoped that the end result justifies the delay. For the delays caused by my own commitments unrelated to this project, I can only apologise.

One of Fred’s most lasting legacies to those who worked with him was his example of open-minded receptivity to new ideas and approaches, accompanied always by a scholar’s demand for rigorous evidence and analysis. His inspirational ability to ignore obstacles and his unfailing generosity to students, colleagues and friends will long outlive him.

Some contributors have been able to update their chapters in the years since they were originally submitted and others not; the year of submission or final revision is noted at the end of each chapter’s bibliography. Some chapters were written several years before or after others, and although every effort has been made to reconcile the evidence presented and conclusions drawn by different authors, some inconsistencies may remain. In one or two cases there remains disagreement about some interpretations; cross-references within the text direct the reader to alternate views.

Diane Fortenberry August 2016

The remains of and evidence for a decade of excavation and study at Englianos is vast, comprising physical artefacts stored at the Archaeological Museum in Chora, and paper and digital evidence in the form of photographs, drawings, databases, drafts and notes. This material has been deposited in the archives of the American School of Classical Studies at Athens, where it will be made available to future researchers. In 2015 a team from the University of Cincinnati discovered an undisturbed shaft grave near the Palace of Nestor – the tomb of a warrior or ruler whose rich burial goods reveal close cultural connections with Minoan Crete. Several of the chapters that follow set out evidence for Minoan influence on the architecture, engineering and ornamentation of what are proposed as three sequential palaces at Englianos, the earliest dating to the Middle

23

ABSTRACT

The following chapters offer evidence for the existence of two palace buildings before the LH IIIB complex published by Blegen, based on a study of the underlying hydraulic system and construction techniques. The suggestion is made that both of these show striking similarities to Minoan architecture. The presence of courtyard gardens is proposed. Evidence for a different means of construction of the LH IIIB palace than what was proposed by Blegen is put forth. Proof is provided for the occupation of the Englianos ridge after the destruction of the LH IIIB palace, including architectural evidence from the Geometric and Archaic periods, and again during the medieval era. Evidence from pottery and other finds recovered from Blegen’s backfi ll indicates activities of various sorts, from possible cult activity to bronze casting, covering the centuries from the Dark Ages through the thirteenth century ad. In addition, topographic field survey, aligned with geographical information system technology, resulted in the rediscovery of several quarries from which the stone used in the palace buildings may have come, and in the creation of detailed digital models of the palace and the surrounding landscape.

In 1990 the University of Minnesota carried out an architectural survey of the standing remains of the Bronze Age Palace of Nestor at Pylos in southern Greece – a world-renowned site discovered by Carl Blegen in 1939 and excavated from 1952 to 1966. In the course of creating the first stone-by-stone state plan of the building, it became clear that some of the architectural assumptions about the structure and its history could not be correct. Over the next eight years the Blegen-period backfi ll covering the site was systematically removed, using the protocols of a stratigraphic excavation, so that a complete architectural state plan could be prepared. Although only backfi ll was removed, numerous unexpected finds were recovered, ranging from Linear B tablet and wall painting fragments to roof tiles and pottery. In addition, a detailed study of the architecture revealed evidence for startling new conclusions about the structure of the palace and the history of the site. Our study of the palace included its excavation history, in which effort the original field notebooks and Blegen’s unpublished draft of what would become The Palace of Nestor were examined in detail, leading to further unexpected discoveries.

25

THE MINNESOTA PYLOS PROJECT, 1990–98

PART I NEW STUDIES AT THE PALACE OF NESTOR

1 THE MINNESOTA PYLOS PROJECT: INVESTIGATIONS AND RESULTS, 1990–98 frederick a. cooper

the remains, drawing and studying every wall and floor excavated by Blegen and in the process recovering any significant finds from his backfi ll. This work proved to be more complex, and more revealing, than anyone had expected, occupying seven seasons, from 1991 to 1998.

The Minnesota Pylos Project, one of several projects undertaken in the western Peloponnese under the aegis of the Minnesota Archaeological Researches in the Western Peloponnese (MARWP) programme, began in 1990 when a small team from the University of Minnesota received permission to survey and study the exposed architecture of the Bronze Age palace at Englianos. The site was discovered and originally excavated by Carl Blegen on behalf of the University of Cincinnati from 1952 to 1966. Blegen’s excavation was careful, thorough and hugely productive, but in the final publication, the four-volume Palace of Nestor (1966–73), the structural remains were treated as little more than containers for the frescoes, ceramics, sealings, inscriptions and other artefacts that emerged. Despite careful descriptions of the architecture, the walls were never fully documented with state plans, and for the most part only schematic line drawings of rooms, corridors, magazines, courts and other spaces were ever published. Our goal in 1990 was to create a stone-by-stone plan of the main palace building (the area exposed beneath a protective shed roof erected in 1956), thus enabling a long-term study of the palace and its place in the history of Bronze Age architecture. This detailed analysis of the main palace was carried out by Michael Nelson and appears as Part II of this publication.

Excavation Protocol This chapter summarises the work undertaken beneath the shed roof and outside it, starting with the Northeast Building and adjacent areas and moving roughly counterclockwise around the Main Building, as the excavations progressed (fig. 1.1). A 5 x 5-m grid was imposed on the site, orientated to the cardinal points, and walls and floors were cleared of earth down to the lowest course of stone or to the level reached by the original excavators; at no point did we excavate new levels, limiting ourselves exclusively to the removal of Blegen’s backfi ll. Once a trench was cleared, detailed plans were drawn and walls recorded, photographs and spot elevations were taken, and trenches were re-backfi lled with sifted earth. Sifting Blegen’s removed backfill was a time-consuming process that resulted in the recovery of thousands of fragments of fresco and pottery, several Linear B tablet fragments, lithics, terracotta figurines and various other artefacts. Recovered material was treated and processed with the care and protocol of a stratigraphic excavation. After careful consideration, we decided to discard undecorated palace-period (LH IIIB) pottery, although records were made of quantities by artefact type and body part (see pp.82–83 for details of the method used for recording the tons of pottery recovered from the Northwest Area and elsewhere across the site). All decorated pottery was retained, as was all non-lh iiib pottery.

In 1989 a decision had been made by the Olympia Ephoreia, under whose jurisdiction the palace falls, to deepen appreciably the backfill covering the architecture outside of the protective roof, following concerns about the deterioration of the remains across the rest of the site. Investigations in the main palace building in 1990 had revealed a different building method than the ‘halftimber’ system described by Blegen, and this and other anomalies in regard to the published details of the palace made it clear that the whole of Blegen’s excavations needed re-examination and proper structural documentation. It was agreed that prior to the re-burial of the larger site, the Minnesota Pylos Project would systematically expose

In addition to ceramics, we recovered about 7000 pieces of fresco and painted stucco. Some 3600 of these fragments

29

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.1 Grid plan showing areas of excavation, 1991–98.

30

The Palace Traverse triangulation initiated in 1952 was extended and expanded to include new details, even after Theocharis’ departure. This original survey continued in use throughout Blegen and Rawson’s excavation, with new details, walls, floors and other features added as they appeared.

– battered, undistinguished pieces assessed to be beyond future diagnosis, most no more than bits of plaster with flakes of white paint – were lotted and discarded. The 3535 inventoried examples met one or more criteria for catalogue entry. Many had distinguishing features, including multiple colours and patterns. Nearly all are small, averaging 0.03 x 0.03 m, and no adjoining pieces were found in situ. When originally used as paving material in the Bronze Age (see pp.40 and 84–85, figs 1.68–1.69), the fresco fragments were simply layered into place without a binder. Those recovered had been dislodged and became part of the Blegen excavation backfi ll.

Many vertical elevations supplied by Blegen and Rawson are in reference to a benchmark that was assigned an arbitrary elevation of 13.00 m. It is marked with a bronze peg sunk into the southeast facade of Room 7; this wall was reconstructed in 1956. We calculated the geoid height above mean sea level as 192.317 m for the Blegen benchmark. Other vertical measurements recorded during the Blegen excavations were made from the existing ground level prior to excavation, which unfortunately is not preserved today.

At the close of the 1998 field season, almost the entire site was cleared of backfi ll that had been re-deposited from 1992 to 1996, and aerial photography by means of helicopter overflight was carried out (pp.115–34, figs 1.73–1.97). The site was then comprehensively backfi lled again, with an additional 50 cubic metres of earth brought to the site for further protection of the ruins and to fi ll the trenches in the Northwest Area, replacing materials taken from the Blegen dump.

Although Theocharis’ original survey was completed with a high order of accuracy, it was impossible to recreate when MARWP took to the field in 1990. Many of the points used by him could not be recovered, and the triangulation method of field survey would not accommodate the rigorous recording of details required. A new survey was thus instituted (fig. 1.2). The palace and its ancillary buildings are generally rectangular structures that consist of long walls running approximately northwest–southeast and short, transverse walls running between the long ones. Walls usually follow straight courses and return or intersect at right angles. The survey laid out by MARWP plotted the course of each wall and the relationship of one wall to another in terms of parallel alignment and angle of intersection.

Carried out during the final decade of the twentieth century, the MARWP Pylos Project coincided with a rapid growth in digital technology, and from the outset the intent was to take advantage of its long-term benefits. Field sheets were designed to facilitate the transfer of recorded information to a Dbase III database program, and Hewlett-Packard handheld computers, running a customised database programme, permitted on-site data entry by multiple staff members simultaneously. Barcode labels ensured the assignment of unique numbers to each object and their associated field sheets. At the end of the fieldwork, the accumulated data was almost entirely in digital format.

For the Main Building and the Northeast Building, the initial step in the survey was to determine the bearing of each long wall, which was numbered and given the suffi x ‘L’ or ‘LD’ (fig. 1.3). The bearings were established by a series of short offsets, the distances read from the face of the wall by transit. The bearings of the long walls were tied into two control lines, labelled the Front and Rear Baseline. The Front Baseline was set by turning 90°00’00” from the course of wall 8L to run N 43°00’00” E – a bearing not only parallel to the facade of the palace but also very close to the azimuth as determined by GPS. The Rear Baseline was staked parallel to the Front Baseline, though offset 54.94 m. to the northwest (measured along the course of wall 4L). In the next step, the long wall courses (L and LD) were intersected with the Front and Rear baselines. The intersection angles were then turned and the distances measured between the intersecting points to compute traverses. Plotting and mapping of the traverses was completed by coordinate geometry. Most of the walls of the Southwestern Building, the Southwest Quadrant, rooms 60 and 62, the Wine Magazine, areas 103 and 106, the Belvedere Area and the Northeast Gateway are neither sufficient in length nor well enough preserved to establish wall courses. Instead, a set of parallel control lines was staked for recording and measuring. Where appropriate, these lines intersected with the Front and Rear baselines,

The Palace Traverse Demetrios Theocharis first surveyed the Englianos hilltop and its remains using a system of triangulation based on measurements between any three given points and their subsequent calculated angles. In 1952 he completed a comprehensive plan of the remains excavated to that date. I have not come across Theocharis’ field notes for his measured survey, nor a resultant working plan. Judging from the progression of plans published with the annual reports, Theocharis corrected alignments and spacings of walls with each successive opening of trenches. He probably used as his survey baseline the inside face of the southwest lateral wall of the megaron, which has a true alignment and which remained fi xed on the plans. Also probably, he fi xed points along this baseline for a triangulation system of plotting points along walls being mapped. As the triangulation method is prone to error propagation, Theocharis must have practised some form of systematic error correction. As the excavation proceeded and more walls and floors were uncovered, the system of

31

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.2 MARWP survey of the palace remains.

32

The Palace Traverse

1.3 MARWP wall designations.

33

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Significant erosion has occurred at Englianos since Blegen and Rawson left the site in 1966. Combined with heavy and sometimes unmonitored tourist traffic, a few of the more fragile walls and features have disappeared or suffered damage that has rendered them nearly un-interpretable. The Southwest Quadrant has suffered most, since it has been exposed for more than 30 years.

and traverses were computed. Finally, tapes were run along the wall courses or control lines, with their zero ends at either the Front or Rear Baseline. Wall and stone drawing commenced along the tapes by a system of right-angle offsets and increments from the zero end of the tape. All horizontal distances were referenced to either the Front or Rear baselines or both. In 1992, using a Trimble Pathfinder differential base station and a mobile unit of the same manufacture, a Universal Transverse Mercator (UTM) coordinate was established for the intersection point of the Front Baseline with wall 14L (the southwestern facade of the Northeast Building). This point is N 4,098,110.35 E 561,828.57 (latitude 37.03N, longitude 21.70E; decimal degrees).

The Architectural Study of the Palace and Ancillary Buildings The initial season at Englianos, in 1990, resulted in the creation of the first detailed state plan of the main palace building beneath the protective roof. Following the institution of the engineering traverse, spot elevations were taken, and the walls were drawn in plan and elevation at a scale of 1:50. This work revealed that the palace builders laid out the courses of walls with great precision. In the megaron, for instance, the preserved socles fall along extremely straight lines, deviating no more than two millimetres. Angles at wall intersections are within a second of a degree of being true right angles.

No significant discrepancies in measurements of the walls and other features exist between the MARWP survey and the original Theocharis excavation survey, though the MARWP state plans do not record the condition and details of the plaster floors of the Main Building, the Southwestern Building and courts 63 and 88; the extent of the plaster floor of Area 58; the pithoi lining the walls of Room 27; the pithoi of the Wine Magazine; and the Lower Town trenches. All of these areas and features except for the Lower Town trenches are very fragile, and it was deemed best to leave the protective earth fi ll in place, even for the brief period of time needed to record them. The Lower Town trenches were beyond the limits of MARWP’s cleaning permit.

Among our initial concerns during the first season were techniques of construction. Blegen described the halftimber construction of the palace as comprised of ‘short sections of heavy rubble fitted into a framework of large, upright, horizontal and transverse wooden beams’ (PN I, 37). Our centimetre-by-centimetre inspection of the extant walls and sampling of building materials indicated

1.4 Northeast Area, showing Phase 1 features in black.

34

The Architectural Study of the Palace and Ancillary Buildings minimum preserved height for this wall. Sherds recovered from the 1958 sondage that exposed the wall date to the Middle Helladic period.

a different building method, that of a modular system of construction composed of alternating rubble-and-mortar piers between smaller slots for temporary retaining members. The construction materials and building methods used by the palace builders, and the chronological sequence of building phases exhibited in the palace, are discussed in detail by Michael Nelson in Part II.

Another stub of a wall, in this case a wall return (wall MZ36), lies underneath and to the southeast of wall Z17. The east–west arm of Z17 runs along the projected line of wall 104 and within 0.03 m coincides with the top of wall MZ36 at an elevation of 192.42 m. These two sections of wall (wall 104 and wall MZ36) align with the equally massive Middle Helladic wall b in the Northwest Area (see pp.53–54, figs 1.27–1.28), suggesting three sections of the base to a transverse wall that ran along what was probably the top of the slope, descending at a rate of about 6–8 per cent towards the northwest. Drain c, running northwest– southeast in the northwest part of the Northwest Area, has the same orientation and elevation as wall b (c.189.40 m). On the opposite or uphill end of drain c in Blegen’s trench MZ40, there lies an associated corner of a wall, undrawn but noted by Rawson. Altogether these features probably demarcate the northwest extent of the Middle Helladic acropolis and, as is proposed in chapter 2 below, the north facade of an early palace. A massive wall in Corridor 26 runs approximately perpendicular to the axis of wall 104. There is also an early drain of Phase 1 embedded in the later wine Magazine, also discussed in chapter 2.

northeast building, areas 101–103 and 106, and the wine magazine Clearing outside of the protective shed began in 1991, when we uncovered the walls of the Northeast Building (rooms 92–100). Cleaning of the complex immediately north (Area 103), between the Northeast Building and the Wine Magazine, also began that year and was completed in 1992. Our procedure in 1991 involved clearing only the tops and elevations of walls in the Northeast Workshop and the area immediately to its north, as Blegen reported only simple dirt floors here. A test clearing, however, revealed a handsome flagstone floor in Area 92/94; another in Area 102 exposed a maze of walls and levels much more intricate than Blegen indicated. As a result, it was decided to clear floors as well as walls in following seasons; the presence of Blegen’s backfi ll still defined in every case the extent of our own cleaning operations.

Other remains of the first phase of building in this area comprise a proto-palatial portico found in Area 103/106, adjacent to the south side of Room 105 of the Wine Magazine (fig. 1.6). The portico consists of a pair of column bases and a plaster floor with painted design. The column bases are set 3.50 m apart and aligned east–west at a 45-degree angle to the overall axes of the palace and Wine Magazine. Blegen recognised the western base (PN III, 61–62, 309) but not the eastern one, half of which is still preserved and forms an obvious respond to the western base. A plaster floor, painted blue, has two surviving white stripes aligned approximately east–west. The white plaster overlaps both bases, no doubt laid in conjunction with the patterns on the floor. This large pillared hall had a minimum north–south width of 7 m and an elevation of

At the far northeast corner of the Northeast Building, widespread remains of plaster were found along the top of the surviving walls. The removal of earth in Area 103 revealed extensively plastered floors not noted by Blegen, although the field notes of Marion Rawson, who excavated the area in 1958, provide some indication of their presence. This area and that to the north (Area 106 and the Wine Magazine) also contain a clear sequence of walls and floors from pre-palatial through post-palatial times, to which Blegen appears to have paid scant attention: ‘Too little has survived of these houses antedating the Wine Magazine to give an adequate idea of the general plan and character of the prepalatial settlement in this quarter or to fi x its date exactly’ (PN I, 336–37). Phase 1 1 (fig. 1.4) The earliest of the remains is a heavy wall (wall 104) below Room 104 (fig. 1.5), which has an exposed length of around one metre, a minimum width of 1.45 m, and is skewed slightly to the southwest of the long central axis of the Wine Magazine; it is one of the most substantial constructions on the acropolis. Approximately 0.70 m to the southeast of the wall below Room 104 stands a complex of stones and stereo that steps upwards towards the north. The base of this stone construction is bedded on the massive wall at an elevation of 191.75 m, and rises to a height of 192.45 m. This stack probably represents the 1

Th is phase corresponds to that of Palace A, proposed in chapter 2. Further details of this phase are published in E.B. French, ed., Archaeology in Greece: 1991–92, 27–28; 1992–93, 32–33; 1993–94, 29–30.

1.5 Wall 104 beneath Wine Magazine vestibule (Room 104).

35

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.6 Area 103–106, Phase 1: plaster f loor and western column base of proto-palatial portico (looking northeast).

this piece as ‘lying beside the water channel’. If this was its original position, the horns of consecration, a very Minoan emblem, stood somewhere in close proximity to this also very Middle Minoan-style terracotta water pipe (see p.144). Other remnants of drains (MZ40 and branch A), and a wall, Z1/2, are also described in chapter 2.

192.97 m; it aligns with the Middle Helladic constructions mentioned above. The terracotta pressure pipe embedded in the outer wall of Court 42 also belongs to this earliest phase, having an elevation comparable to that of the pillared hall. This branch B of the aqueduct is discussed further below. In association with it, Blegen discovered two fragments of a large stone horns of consecration (PN I, 328, figs 238–39), an exceptional find for a Mycenaean site. A tip of one horn was reused in a scrappy wall of Phase 3, and another larger, lower portion was found at the junction of branches A and B of the aqueduct in this area. Blegen described

Phase 2 (fig. 1.7)2 Beginning with Phase 2, the previous north–south orientation of construction rotates to the northwest– southeast. The most prominent remains of this phase are those of Room 102 (fig. 1.8), which Blegen identified as

1.7 Northeast Area, showing Phase 2 features in black.

36

The Architectural Study of the Palace and Ancillary Buildings

1.9 Plan of the Shaft Grave, Room 97.

Phase 3, earlier than Phases 5 and 6, during which the LH IIIB palace was built.3 Patches of white plaster, probably belonging to different phases around the site, appear here and there underneath later floor levels in the Northeast Area. One such patch is a white stucco remnant cut through by the dogleg angle at the north corner of the wall to Court 42. An elevation of 192.93 m puts it level or nearly level with other plaster patches of this phase in areas 94 and 100 and underneath the east corner of Court 47.

1.8 Room 102, the cistern, from above.

a cistern, based on the presence of a plastered floor and base panels of waterproof cement (PN I, 336) and on the chamber’s size of 2 x 2 m, too small to be a serviceable room. The southwest wall of the chamber was later incorporated into the exterior wall of Court 42, just northwest of the terracotta pressure pipe, branch B of the aqueduct. (For details of the construction of this wall, see Part II, pp.345–46) The southeast wall of Room 102 is aligned with the terracotta pressure pipe, with its top at 193.48 m – some 0.30 m higher than that of the pipe at 193.18 m. Associated with Room 102 by orientation and elevation is a 5-metre stretch of wall 98/103, about 5 m to the northeast and located some 0.5 m outside the later (Phase 9) northwest wall of Room 98. This wall appears to run under wall 97/103, but no further; the northeast end terminates with a pair of large fieldstones, perhaps quoins. It is not a wall return for a building.

Phase 3 (fig. 1.10) At this time, the rubble walls that zig-zag at right angles to each other in Area 106, to the southeast of the Wine Magazine, were cut into the blue plaster floor of Phase 1. These walls form what could be construed as one-half of a megaron-like plan and follow a distinctly different orientation from that of the pillared portico in Phase 1. Walls Z17 and Z18 bond at an acute angle, at elevation 193.18 m; wall Z17, like the northwest half of the ‘zig-zag megaron’, was demolished for construction of the Wine Magazine. Wall stub Z19/Z20 dates to the post-Bronze Age, Phase 10 (pp.41–43).

Blegen found a stone-lined shaft grave (1.53 x 1.05 m and c.1.2 m deep) 0.20 m below the floor of Room 97 of the Northeast Building. The MARWP plan (fig. 1.9) supplements the published photographs (PN I, figs 229–31). According to Blegen (PN I, 314), the date of the few ceramic finds from the grave was ambiguous; he concluded that the shaft grave ‘antedates the surviving palace’. Its top course falls at 192.80 m in elevation, or 0.20 m below the bottom level of the northwest wall of Room 97; this wall was built in

Dating to the same phase as the zig-zag walls are vestiges of a building to the south, in Area 101, composed of stretches of rubble wall not unlike those already mentioned. The first (wall 97/101) of three extant walls was reused in Phase 9 as the oblique northwest wall return of Room 97. A parallel wall, c.5.0 m. to the northwest (wall 101/192), is connected to the first by a cross-wall (97/103) that laps over the earlier 5-m stretch of wall of Phase 2. The northwest parallel wall, in turn, is incorporated as part of the southeast wall of Room 103d of Phase 11. Five oblong blocks (steps 101) embedded in the plaster floor at the northwest of Ramp 91 of Phase 8 (p.41) make a right angle. By elevation and orientation, though not in style of construction, they can be associated with the H-shaped plan of walls just to

2

Th is phase and the following two, phases 2–4, correspond to Palace B as proposed in chapter 2.

3

The grave’s top elevation of 192.80 m lies a few steps below the level (193.20 m) of a posited Northeast Quarter of a second early palace on the site (Palace B) and several steps above the posited Southeast Court of that building (see pp.147–48, fig. 2.10).

37

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.10 Northeast Area, showing Phase 3 features in black.

When I was briefly assisting at Ano Englianos in 1954 there appeared in the earth-floored room I was working in two squarish blocks of purple stone. The workmen immediately called them ‘polythrones’ and imagined

the northeast. Blegen identifies as a threshold the pair of blocks that run somewhat transversely to Ramp 91 (PN I, 326). Although there is no apparent and direct connection, the right-angled set of squared blocks probably compose some sort of unified construction with the H-shaped plan of walls. As is discussed in detail in chapter 2 below (pp.147–48), these remains are assumed to form part of an early propylon-like entrance to a pre-LH IIIB palace (Palace B), one with a different orientation to that of the Middle Helladic palace of Phase 1. Built into the steps 101 was a piece of the upper tip of a horns of consecration, a larger piece being found alongside branch B of the aqueduct c.4.5 m to the northeast in Area 101 (PN I, 326), as noted above (Phase 1). The horns of consecration must have been broken prior to Phase 3. • An Early Revetment Slab Not far from the find spot of the horns of consecration fragment, an equally distinctive architectural piece was discovered by MARWP: a red-purple marble or rosso antico plaque, 0.46 x 0.33 x 0.035 m (figs 1.11–1.12). The Bronze Age piece in the Northeast Area is partly broken along a long side, but otherwise the edges are preserved. It is lodged underneath a nondescript fieldstone in the foundations for Building 97/101, just described. The object appears to be reused in the fabric of this Phase 3 building, making its original period, like the horns of consecration, either Phase 1 or Phase 2. In an anecdote, written years later but surely trustworthy, Emmett Bennett reports (1999, 101):

1.11–1.12 Two views of the rosso antico stone embedded in the foundation of Building 97/101.

38

The Architectural Study of the Palace and Ancillary Buildings It is likely that veining across the stone pieces was variegated, as in the Tainaron coloured marble outcrops, and this early stone revetment at Pylos anticipates its later simulation in wall painting, in particular in dados from the LH IIIB palace. Mabel Lang catalogued 26 examples of this kind of painted decoration and astutely observes (PN II, 164):

the king and queen sitting on them and sipping their tea, though they are definitely uncomfortable as seats of any sort. Blegen found them to be simple bases for the posts holding up the timbers of the second floor. In his excavation notebook for Room 62 and above the Main Drain, Bennett enters for his trench S2/A (Bennett 1954, 62): ‘one small thin piece of bright red stone [inlay?]’; a short time afterwards in the same trench (ibid., 65): ‘taking out some of the inlays’; and (ibid., 76): ‘Piece #37 of table joins together the first and last batches.’ In the notebook’s table of contents, Bennett lists these cited pages under ‘VIII Stone, Table - Tr I’. The ‘bright red’ stone inlays or table are not cited in The Palace of Nestor nor in Blegen’s draft manuscript, nor did MARWP recover any pieces of this ‘table’ from Blegen’s backfi ll. While it is improbable that a square-shaped ‘purple stone’ ever served as a base for a column, Bennett’s descriptions in both 1954 and 1999 point to the existence of revetment of reddish-purple marble comparable to the example underneath the wall of Building 97/101.

Fresco dado may be roughly defined for the present as that painted decoration which appeared on the lowermost portion of the wall. Its location is assured both by much painted plaster of this kind that is still in situ and by the nature of the designs, which are more schematic and conventionalised than those from any other part of the wall and most like the decoration found on the painted floors. The arc dado appears everywhere at Pylos, both inside the palace and out. Because its basic form remains remarkably constant, only a few examples have been catalogued (I-4 D, 7-8 D, II D, 26 D), notably I D 64, which in its ten metres of length runs the gamut of allowable variations. The arc dado has panels ca. 0.60 m. wide marked off by red upright bands; arcs of various colours start in the lower left and upper right corners of each panel and meet in various ways, back to back, in the middle, often leaving a spandrel effect at the right; the arcs are marked both at and just inside their edges with black or red lines of various kinds. These punctuating lines, which may be simple, rippled, scalloped or elaborated with semi-circles, vary from panel to panel but remain the same within each. The whole is obviously a stylized imitation of ca. 0.60 m. plaques of cut stone which are imagined to be various in color and marked with darker veins [emphasis mine].

At a glance, and in advance of archaeometric testing, the red-purple and fine-grained crystalline marble plate (Munsell Light SRP 6/2, red-purple; Pantone S138-4) suggests an extraction from the rosso antico quarries on Cape Tainaron.4 Coloured stone was a popular choice for the production of vases in Minoan times, and a fluted rhyton from Kato Zakro is carved from Tainaron rosso antico (Lazzarini 2007, 91, fig. 6). There are some instances of the architectonic use of coloured limestone and marble on the mainland during the Mycenaean period: rosso antico reliefs from the Tomb of Atreus at Mycenae, the antica verde decorative facade of this same tholos, and the antica verde ceiling of the tholos at Orchomenos. However enigmatic may be the original purpose of the red-purple marble plate under Building 97/101 and the pieces found by Bennett in Room 62, the recovered example represents a significant remnant from the period of the pre-palatial palace(s) at Pylos, for the 0.035-m thickness of the rectangular plate found by MARWP points to wall revetment. Assuming the ‘inlays’ or ‘table’ found by Bennett were also part of a coloured stone wall revetment, each with the same dimensions as the piece found in 1993, these indicate a course of revetment nearly 1.5 m long and 0.35 m high.

Arc dadoes appear in many places over a long period of time. At Hagia Triada the dado which accompanied the Dolphin Floor is divided into panels somewhat over 0.64 m. in width by upright red bands. The panels are divided into arcs of color which start in lower right and upper left corners and are punctuated by red and black scallop-lines. The height of this dado is ca. 0.45 m.; the chief difference from the Pylian specimens is that the arcs run horizontally rather than vertically. Also at Hagia Triada on the Sarcophagus a simplified form of the arc dado is used on the base of the double axe standard, just as it is used in Pylos on the “altar” (26 D 92). Small but perfectly recognizable fragments of arc dado occur at Tiryns also (Tiryns II, pl. III, nos. 11–13) and at Thebes (Frauenfries, 12f., no. 42).

4

I encountered these quarries during my quest for the source of the white marble used in the fi ft h century bc at Bassai (F. A. Cooper 1996, 107–19; F. A. Cooper 1986; F. A. Cooper 1988). A team of Italian geologists has since intensively surveyed the lower Mani quarries and continue publication (Attanasio et al. 2006, 108–14) of archaeometric analyses of these white, variegated and coloured marbles and microcrystaline limestones begun by Herz, Cooper and Wenner (1982). Lazzarini (2007, 71–96, figs 1–53) publishes the results of his analyses and those of others to date; colour-correct plates accompany Lazzarini’s volume. The term rosso antico was coined in the Renaissance/Baroque period and has come to refer to red-purple and sometimes variegated marble.

Painted simulation of polychrome stone revetment probably begins in Middle Minoan III and becomes an increasingly popular type of dado decoration through Late Helladic IIIB (Immerwahr 1989, 145–46). In such an earlier phase on Crete, probably proto-palatial, fine dado decoration consisted of imported coloured marbles in an opus sectile design. Unfortunately, there are few extant

39

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.13 Northeast Area, showing Phase 4 features in black.

examples; I am aware of two on Crete and, now, the redpurple slabs from Englianos. This luxurious decoration eventually yielded to the more practical, painted simulation. Nelson (2007), using the evidence of setting lines and cuttings for attachments, discusses placement of carved stone revetment in Minoan and Mycenaean architecture, including a piece found at Pylos (Hofstra 2000, 205–16, fig. 30). The polychromed plaque in Area 101, however, probably dates to the earliest phases at Pylos.

for a floor. The floor is composed of at least two layers of wall fresco chips, lying painted side up for the most part. The fragments appear to come from walls decorated with geometric patterns, including terracotta-coloured spirals on an ocher background and red-and-white stripes; there is some indication of floral designs and perhaps some miniature figures. The original frescoed walls from which the fragments must logically date to before walls Z24, Z26 and Z3, probably date before the ‘zig-zag megaron’, and may relate to the pre-palatial portico of Phase 1. The fresco-chip floor extends southwest into a small courtyard between walls Z21 and Z24 and abuts the southwest face of wall Z3, the southeast face of wall Z21, and walls Z25 and Z26.

Phase 4 (fig. 1.13) Some 20 metres to the north, at the northern part of Area 103, there is ‘a veritable maze of walls difficult to disentangle and to understand, but obviously representing at least two, if not three phases’ (Blegen 1959, 123, pl. 32, fig. 5). In the field notebooks for Area MZ, trenches and walls are given MZ and Z numbers respectively, and as found; many are not mapped or clearly labeled. Walls Z24, Z26 and Z3 (probably houses and only partially excavated by Blegen) run at oblique angles to the earlier ‘zig-zag megaron’ of Phase 3 and may represent constructions unaligned but perhaps contemporary with those in Area 101, because both sets of walls are chronologically sandwiched between comparable earlier and later phases. The bases of walls Z27 and Z3 lie 0.07 m above the shaved top of the wall of the ‘zig-zag megaron’, at 192.87 m. In addition, a packed layer made of broken painted plaster fragments (see figs 1.65–1.66 below) runs up against the walls of both the ‘zig-zag megaron’ and Blegen’s Oblique Houses.5 At 13.04 m, this layer lies c.0.20 m below the top of the bounding walls but above the ‘zig-zag megaron’. The nature of this layer is unclear, but it seems to consist of wall fresco fragments compacted to make a metalling

Phases 5, 6 and 7 (fig. 1.14) These phases correspond to the LH IIIB palace. Phase 5 saw the construction of the southeast portion of the northeast wing of the palace; that is, the facade extending from the anta for Portico 41 to the southeast corner. The wall is built in the ashlar style, according to Nelson’s criteria (see Part II, pp.318–24). The northwestern portion of the facade is in ashlar shell style with reused blocks and so-called weeper holes, and falls within the next phase, Phase 6. The top, horizontal levelling course of the northwestern facade rises, at 193.26 m, some 0.43 m above that of the preceding southeast section at 192.83 m. The Wine Magazine was 5

Rawson (1958, I, 170–71) noticed this layer and drew a cross-section, briefly describing the trenches MZ24, MZ42, MZ52, MZ36 and MZ44. A bulk of painted plaster fragments from the floor was ‘tossed’ (1958, II, 107–15), probably numbering among the quantities recovered by MARWP in our excavation of the Blegen Dump.

40

The Architectural Study of the Palace and Ancillary Buildingsl

1.14 Northeast Area, showing phases 5, 6 and 7 in black.

built at the time of phases 5 and 6; it is dated by sealings found in it to Late Helladic IIIB. Since it cuts through the ‘zig-zag megaron’, the previous phases must date to an earlier period.

also overlaps the bracket-shaped drain opposite (fig. 1.17), meaning that it was installed after, perhaps immediately after, the construction of courts 42 and 47 and branch C of the aqueduct, all Phase 7. This Phase 8 floor is the latest of a series of floors in this area. At several places, where portions of the upper layer have broken away, a lower level of plaster is exposed, assigned to Phase 2. As noted above, several patches of this lower plaster lie beneath the southeast corner of Court 47.

Sometime after the completion of the northeast wing of the palace, a metre-thick rubble wall was built to create a pair of courtyards, rooms 42 and 47 – Phase 7. The southwest wall of Room 102 (Phase 2), with a width of 0.65 m, was incorporated into this broader enclosing wall (see fig. 1.7 above). In order to accommodate the differences in width, a short segment of rubble wall was inserted alongside the pre-existing southwest wall of Cistern 102. The plastered floor on which basis Blegen identified Room 102 as a cistern passes beneath this added segment of rubble wall.

The rubble wall that forms the southwest flank of the Northeastern Building respects the aqueduct; nevertheless, its base rests on top of the northeast edge of the bracket-shaped stone sections of water channel. The Northeastern Building, therefore, represents Phase 9; its oblique northwest wall is reused from Phase 3. The floor and walls of the Northeastern Building slope downwards from a floor-level high of 193.03 m at the northwest to a low of 192.44 m at the southeast.

As a coherent part of the construction programme of Phase 7, the water channel was diverted, branch C angling towards the southwest corner of the skewed northwest wall of Room 97. Here, at a crudely constructed settling basin, the branch then turns southeast straight down the slope of Ramp 91, running parallel to and 3 metres away from the wall of Court 47. Branch C has a different form of construction from the rest of the aqueduct, consisting of a series of stone blocks cut in bracket shapes and capped by fieldstones, much like those in Room 78 (see Part II, pp.371–72).

Phases 10, 11 and post-Bronze Age (fig. 1.18) Evidence for post-palatial occupation of this area of the site comes from the walls of Area 103, in particular walls Z4, Z13, Z10 and the detached Z15 (overlying Z23 of Phase 3); from walls located immediately northeast of the Wine Magazine (Z16 and wall corner Z19 and Z20); and from elements of rooms 92–93 in the Northeast Building. (Room 103c is later modified in Phase 11). The foundations of all these walls were built over an earth deposition approximately 0.10–0.20 m thick. The deposit has a top elevation of 191.98 to 193.14 m (sloping slightly from north to south) and lies on the demolished tops of earlier walls with elevations of 192.88 to 193.18 m. (The bottom of wall

Phases 8 and 9 The 3 metre-wide corridor, Ramp 91, received an impressive plastered pavement: smooth, polished and glistening white (figs 1.15–1.16). The flooring not only laps the bottom edge of the northeast walls of courts 42 and 47 but 41

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

42

The Architectural Study of the Palace and Ancillary Buildings

1.18 Northeast Area, showing phases 10, 11 and post-Bronze Age in black.

c.1.20 to 1.50 m, and they are domed in mudbrick (Griebel 1992). This construction has an overall diameter of c.1.5 m, and in this and adjacent trenches Rawson found chunks of brick, some baked hard and reddish. (The presence of mudbrick, however, is not decisive, as it was found in deposits over the site.) Blegen mentions fragments found here of what he calls a ‘griddle’ for waffles (PN I, 340, fig. 348:11 and 12): a circular or oval dish with raised edges and shallow perforations. He could suggest ‘no close parallels from other sites’ (PN I, 341). The ‘griddle’ may well be a baking paddle, and this and the bake oven suggest postBronze Age baking activity adjacent to the Iron Age houses of Area 103. An Iron Age date is also assumed for the repositioning here in stereo of the Bronze Age larnax.

Z20, here interpreted as Geometric in date, is separated by this 0.10–0.20 m deposit from the top of wall Z21, here associated with the ‘zig-zag megaron’ of Phase 3.) Rawson recovered an iron blade from the northwest end of wall Z16 and an iron nail from the southern end of wall Z10 (the western wall of Room 103c). These objects, omitted from Blegen’s catalogue of finds from this area (PN I, 340– 41), strongly suggest an Iron Age or later date for these buildings. For further analysis of the post-palatial remains on the site, see pp.66–73, 76–80 and chapters 8–10. Two curious features from the post-palatial phases – one found in the Northeast Area just south of wall Z24 of the zigzag megaron and the other in trench MZ58 – are a terracotta larnax or basin and a large circular construction. The former is described in detail by Blegen (PN I, 339–40); the latter is not, though it does appear on plans of the Northeast Area (PN I, figs 303 and 309). The circular platform consists of a single, large triangular stone with curving sides (c.0.50 m in diameter) that occupies a central position and is surrounded by smaller flat fieldstones (0.35–0.40 m in diameter and 0.15 m thick), with still smaller stones in the interstices. Three large stones and a single small one make a short (c.0.80-m) extension on the north side, c.0.70 m wide. The construction, size and design suggest a bake oven.

Room 93, a 3-metre square chamber at the southeast end of the southwest half of the Northeast Building, preserves the bottom courses to a facade of some sort. A pair of large, reused anta blocks lie to either side of a narrow sill (fig. 1.19). A cubical block bearing painted stucco (fig. 1.20) fronts this entrance c.3 m to the southeast and somewhat off-axis. A third reused anta block was inserted into the juncture of walls at the north corner of Room 93. Blegen interpreted Room 93 and the cubical block as a shrine and altar of the Mycenaean period (PN I, 301–02). He recognised that the pair of antae blocks was propped by a thin layer of dirt and small stones above a supporting course below.

Bake ovens from Neolithic through Roman and early modern times fall within a limited range of diameters,

The three antae, the sill and the ‘altar’ are undeniably Mycenaean in date. The antae blocks, however, have been reused to make the shrine. The 0.10–0.20-m deposit of intervening soil also separates the walls of Area 103 and wall Z20 from Bronze Age levels, as described above. The pillar shrine ‘altar’ was inserted through a square cutting

1.15 (far left) State plan of Ramp 91. 1.16 (near left, top) Surface of Ramp 91 adjacent to the northwest (outer) face of Court 42. 1.17 (near left, bottom) Surface of Ramp 91 overlapping the drain.

43

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 northeast side of the acropolis: the northeast gateway, belvedere, kiln and aqueduct Exploration of the broad stretch of acropolis that lies to the northeast of the palace was facilitated by a magnetometer survey conducted in 1993 by Peter Jenson of the University of Tübingen, in cooperation with the Pylos Regional Archaeological Project (PRAP). At first, the results appeared to reveal the existence of two and possibly three unexcavated structures in the central, northwest area of the hilltop. Two of the configurations were later identified as echoes of two trenches dug in 1952 and 1954. The third was the outline of a building, P13, excavated in 1964 (PN III, 28, figs 58 and 302). Of more use, the magnetometer survey enabled us to locate the outlying trenches in which Blegen uncovered the kiln and the aqueduct. Until their rediscovery these trenches were not tied by survey into permanent points, and their outlines did not appear on the ground surface; no plan of the aqueduct was ever published, and it appears unlikely that a working plan was ever made (PN I, 332–36).

1.19 Antae and sill at entrance to Room 93.

Clearing and re-examination of Blegen’s aqueduct (fig. 1.21) took place in 1994 (see also Part II, pp.368–69, figs A.1–A.4). The long water channel ends at the Northeast Gateway (fig. 1.22) and begins in Area 101 (branch A), where it divides in a Y-conjunction, branch B passing into Court 42, branch C turning south and running in the open stone channel along the base of the exterior southwest wall of the Northeast Building (see fig. 1.4). The Y-conjunction in Area 101 is scrambled with an overlay of stones, and the actual fitting has probably been destroyed. The spot has not been excavated, so the nature of the junction is unknown, but it can be said that branches A, B and C are of different materials and design. The junction had to connect a terracotta pressure pipe line from Court 42 (branch B) and a bracket-shaped terracotta drain (branch C) of various profi les (see Part II, p.369, fig. A.4) with a sometimes stone-lined (if on a single side only, then always on the northwest side: PN I, 334), dirtbottomed channel (branch A) that heads in a serpentine path towards the Northeast Gateway. The line of pressure pipe and thus the initial cut for the aqueduct channel are both assigned to Phase 1 of the Northeast Area (see fig. 1.4). Branch C cuts into the lower, Phase 2 level of plaster floor in Area 101 and is an integral part of the construction of Phase 8 (the plastered Ramp 91), preceding the erection of the southwest flank wall of the Northeast Building.

1.20 Painted ‘altar’ in front of Room 93.

into the stucco floors of phases 2 and 7: ‘the southeast and southwest faces are almost wholly exposed above the pavement’ (PN I, 302). To be sure, there is disagreement in our own field observations as to whether or not a plaster layer continues unbroken from the Bronze Age floor and up the northwestern and northeastern sides. On the other two sides, however, the bottom of the block is visible through the broken paving (PN I, 320), which means that, whatever its period, it was placed on a surface. Stratigraphical evidence suggests this remodelling is post-Bronze Age in date. The telltale black, oily layer with Geometric sherds, which defines the re-occupation of the site after the LH IIIB destruction of the palace, covered this area (see p.227, fig. 8.1). Popham quite correctly recognises a Geometric date for a horse made of gray terracotta that comes from the adjacent Corridor 95 (Popham 1991, 324; PN I, fig. 310). Moreover, as Blegen noted in his preparatory manuscript (n.d.), lots My1–My4 contain a few late sherds with Geometric mixed in. Although a Dark Age date for this area remains tentative, a parallel exists at Minoan Kommos, where a sacred compound of refashioned Minoan materials dates to the Geometric to Archaic periods. At the Cretan site, the Minoan-period offering table held later votive dedications (Shaw and Shaw 2006, 20–24, 35–36, 57–59, pls 144–49).

At the northeast shoulder of the acropolis, above the Northeast Gateway, the stone lining of the main channel steps and then forks into several branches as it wends its way down through the gate. Blegen, in a schematic plan (PN III, fig. 304), shows the bifurcation more clearly than the actual state plan of MARWP.

44

The Architectural Study of the Palace and Ancillary Buildings

1.21 Western section of the aqueduct, looking northwest.

1.22 The Northeast Gateway, looking northwest.

1.23 The kiln.

45

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 direction of the subsidiary water channels that descend to the aqueduct at an angle but in the direction of the northeasterly flow in the aqueduct. Blegen was puzzled by the fact that the rim of Cistern 102 (193.48 m) is raised above the adjacent terracotta pressure pipe in Area 101 by 0.30 m; the difference in elevation precludes an assumption that the aqueduct fed water upwards and over the rim of Cistern 102 (see further pp.139–40). It is far more probable that overflow from the cistern spilled into the northeasterly flow of the aqueduct.

The aqueduct had a very long life, extending from Middle Helladic times to the end of the Dark Ages. Cheryl Kittredge recovered an iron nail from a trench located along the aqueduct (trench III in PN III, fig. 302; MARWP grid intersection N 8159 E 860). The central portion of the aqueduct as it passes across the acropolis contained pottery of Middle Helladic to Late Helladic IIIB date and a fired Linear B tablet (PN I, 335). Mrs Kittredge’s trenches followed the meandering course of the aqueduct; she exposed the entirety along the top side, but it appears from her notes that she dug to the bottom only intermittently. Only at the southwest and northeast extremities did we clean to the bottom of the watercourse; in hindsight, we should have re-opened Mrs Kittredge’s spot trenches.

At several points in his text, Blegen repeats his hypothesis that a bridged aqueduct originated at ‘the spring called Rouvelli, 1,360 m. to the northeast of Epano Englianos’ and joined the actual cut for the acropolis aqueduct part way along its course (PN I, 23, 181, 333). This supposition deserves a brief look. A straight-line plot of such a bridged aqueduct would begin alongside the paved road from the south to Chora. The elevation at the spring (215 m) is indeed above the aqueduct’s elevation of 193.0 m on the Engliainos ridge. The topography, however, speaks against a bridge. Along a descending course about 250 m from the spring, there is an interceding ridge c.10 m above the level of such an aqueduct. Further along, c.350 m from the acropolis, lies a deep valley, about 55 m below the interpolated 200-m height of an aqueduct at this point. This difference in elevation would have required a trestle more than fifteen storeys high.

Blegen determined (PN I, 332–33) that the bottom of the channel rises from both directions to a maximum height of 14.15 m (from an assumed elevation of 13.00 m, interpolated as 193.63 m) at a point 66 m northeast of Court 42. The low point occurs a third of the way along the course; a pair of larnaki was inserted here (fig. 1.24); the first lies in the path of the aqueduct and is broken. Rawson excavated the south portion of the aqueduct in her MY trenches of 1958. Her rod readings up to and including the pair of emplaced larnaki correspond with what we obtained in 1993 and 1994. As noted above, we exposed the top side and took elevations at intervals on top and below the stonework. Where Cheryl Kittredge’s and our rod readings fall on the same stones, the results show her values to be consistently too high – from 0.18 to 0.22 m, or an average of c.0.20 m. When this error is subtracted from her elevations at the bottom of the aqueduct (in the 13.00 m datum), there is no gradient to speak of; instead, the aqueduct has a level flow of 193.03 m. A couple of computed elevations reach 193.12 m. There are ten such spot elevations at fairly regular intervals of between 3–4 m and 7 m along a 40-m stretch, beginning at 30 m from the Y-junction in Area 101.

A wholesale reconsideration of the direction of the aqueduct water flow began with the exposure of the several segments of tapered terracotta pipe in Court 42. In this design, the narrower end, or nozzle, fits securely into the wider. In this manner, water can be put under pressure, and the pipe thus serves as the critical section in an inverted siphon hydraulic system, in which water flows downwards from a header tank to a lower elevation and then under pressure is carried back upwards to a receiving tank at a somewhat lower elevation than the header tank (Wilson 2007, 295–97, fig. 11.2). Such systems are rarely found before Roman times except in Middle Minoan Crete, when they are fairly common, with some examples found in Late Minoan contexts (see J.W. Shaw 1973, 198–201). The lay of the pipe indicates that the water flow was meant to go from Court 42 towards the northeast and Area 101, that is, in the opposite direction to that supposed by Blegen.

With these revised elevations, the water could have flowed in either direction, until it reached the downward fall at the northeast rim of the acropolis.6 A direction for the water flow is determined by several constituents: the northeasterly drop through the dividing channels passing through the Northeast Gateway; the northeasterly direction of the terracotta pressure pipe from Court 42 to the Y-junction in Area 101; and the southeasterly

As noted, the severed section of pressure pipe (branch B) in situ within the exterior wall of Court 42 (pp.138–39, fig. 2.5) represents the preserved south-westernmost end of the water main. The base of this pipe (193.03 m) and the base at 6

There are two additional considerations. First is the likelihood that tectonic thrusting has contributed to the slight disparities (personal communication, Dr Michael Timpson, August 1994). Second, there is reason to suppose that the actual base of the drain, at least in some places, was not reached during the 1962 excavation of the aqueduct. Silting from backwater flow creates a pseudo-‘stereo’ earth fi ll, a level at which Blegen typically ceased digging.

1.24 Low point of the aqueduct, with pair of larnakes.

46

The Architectural Study of the Palace and Ancillary Buildings northwest area

the opposite or northeast extremity of the aqueduct (193.18 m), just at a point where it drops through the Northeast Gateway, are comparable in elevation. The bottom of the main channel runs level at 193.03 m between the pressure pipe in the wall of Court 42 and the larnax embedded within and at the bottom of the aqueduct some 28 m to the northeast.

The area to the northwest of the palace (fig. 1.25) was excavated by George Papathanasopoulos in 1958–59 and by William Kittredge in 1962; Blegen discussed the area in both volume I of The Palace of Nestor (pp.289–98) and volume III (pp.43–46, fig. 311). Kittredge kept remarkably coherent field notes, paginated and often cross-indexed, with notated elevations, plans and pottery lots (though not instrument height or rod readings). He characterised strata by simple colours and density descriptions; only on occasion did he note particle compositions of the soil. He excavated several trenches simultaneously, here designated schematically WK1–WK10 (fig. 1.26). His scarp elevations plot strata by colour, and he usually accompanied diagnostic pottery finds with a brief description. In his notes, Kittredge sometimes hazards a guess as to a stratum that might pass from one trench to another, that is, from one cavity between walls to the next cavity. In one instance, he broaches an important observation about his strata: he notes two thin layers, a heavy pink above a yellow, which separates a squatters’ period from the palace-period levels below (see also p.240).

Creation of the state plan of walls over the site revealed that the axis and returns of the Northeast Gateway are congruent with the lines of the LH IIIB palace. Blegen’s date of Late Helladic I for the gateway may well indicate an association with the eastern ashlar walls of the palace and an earlier palace reconstructed herein as Palace B (see pp.144–50). In the Belvedere area, which Blegen interpreted as houses along either side of a narrow street, the original trenches exposed only a portion of the remains: the walls evidently extend farther to the east and west into unexcavated soil. The so-called kiln (fig. 1.23) is situated in this general area. Pin-pointed by the magnetometer survey of Peter Jensen, the kiln was cleaned of backfi ll in 1993, drawn and photographed (see Part II, p.289, figs 1.6–1.7). This feature is well described by Blegen (PN III, 19, fig. 308; see also a synopsis in Mountjoy 1993, 121), although we observed little surface evidence of fi ring on the mudbrick lining.

Not unreasonably, Kittredge and Blegen expected to be able to reconstruct and correlate strata and pottery deposits and sequences after the fact, but this turned out not to be the case. In his final publication Blegen made little attempt to interpret any of the approximately 40 walls in the Northwest Area, except for the rectangular foundations

1.25 State plan of the Northwest Area.

47

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 of Circular Structure 87; both the published and working plans are rendered as a schema of confused structures. Indeed, the maze of walls presents a bewildering crazy quilt even to the most tutored of observers.7 In regard to stratigraphy, Blegen simply remarked (PN III, 45–46):

walls and strata of the area. My goal has been to assemble an unambiguous stratigraphy for the Northwest Area, dating from the Middle Helladic to the Classical period, and from that for the Englianos acropolis as a whole – a proposition not realised by Blegen.

At the bottom of one trench fragments of an alabastron of Late Helladic II appeared; and on stereo itself was found a sherd of Mycenaean III B (fig. 150, bottom row, second from left), a startling intruder from the higher strata, where it once probably formed part of the rim of a krater-bowl. Perhaps it had fallen from a top level into the pit, or been dislodged from the side of the upper part of the trench.

One of the first facts to emerge from the MARWP investigations was that levels of excavation were originally measured downwards from the top of the trenches’ preexcavation ground surface. That surface sloped from east to west at a grade of about 10 per cent. In terms of absolute elevation, each top of trench had a different datum, and thus a level stratum passing through two or more trenches had differing elevation values. During the 1962 season Kittredge converted his top-of-trench elevation readings to the new site datum of 13.00 m. He appears to have done this mathematically, rather than by direct instrument readings; the assumed datum of 13.00 was set on a permanent benchmark fi xed on top of a metal plug in the reconstructed wall at the west corner of the palace. The MARWP datum is based on a UTM-based elevation fi xed on this benchmark at 192.317 m above sea level, or a difference of 179.32 m from Blegen’s benchmark. Ordinarily, this should allow for calculated conversion of elevations recorded during Blegen’s post-1962 excavations. Unfortunately, this is not the case in the Northwest Area, for in comparing elevations taken at a number of fi xed points, such as corners at the top of walls, it turns out that when converted to 179.32, Blegen’s assumed datum at 13.00 m has an error correction at the Northwest Area of 178.90, or 0.42 m lower in elevation. The discrepancy can be determined from spot elevations on Kittredge’s pencil working drawings. A mean of differentials between Kittredge’s elevations (all in the range of 10.10 m to 13.30 m assumed datum) calculates at a mean of 178.97 m with a standard deviation of 0.39 m. A slightly wider spread of 0.42 m fits both wall and strata elevations.

Investigations in the Northwest Area and the Kittredge Notebooks The MARWP reopening of the Northwest Area resulted in the first state plans and elevations of this part of the site, and provided as well the opportunity to revisit and compare the physical remains with the notes of Papathanasopoulos and, more particularly, Kittredge. A full re-analysis of the post-Bronze Age levels in the Northwest Area was carried out by Todd Brenningmeyer and appears as chapter 8. The autopsy below summarises his research and attempts to explicate in detail the earlier

The errors, though inconvenient (as was the revelation that few earth baulks were left behind, making Kittredge’s observations difficult to verify), are consistent, and the correction factor of 0.42 m may be taken as reliable. The result is that the LH IIIB sherd cited by Blegen and quoted above comes from within drain c in trench WK7 (table 1.1, no. 27). Another unmentioned but recorded sherd of Middle Helladic date comes from the same drain c in trench WK7 (table 1.1, no. 5). It also turns out that a Late Helladic II alabastron, thought by Blegen to come from the lowest stratum, since it was from the bottom of a trench, was actually found on top of a wall and at an elevation of 190.67 m – a position appropriate to the LH II period (table 1.1, no. 24). Table 1.1 lists Kittredge’s section drawings by fieldbook page number, by assigned trench number and, when 1.26 Schematic plan showing Kittredge’s trenches in the Northwest Area (designated WK1-10), overlaid with Papathanasopoulos’ trenches to the southwest.

7

Deepest gratitude is offered to Dr Robin Rhodes, who spent several days assisting in a concerted effort to arrive at an initial understanding of the Northwest Area and to prepare preliminary sections and plans.

48

The Architectural Study of the Palace and Ancillary Buildings other is difficult to determine with any certainty from his notes or the published black-and-white photographs. A future re-examination of the kept pottery lots may further clarify the stratigraphic chronology proposed here for the Northwest Area.

he supplied the information, by sondage area within a trench. A column within this table shows the various soil colours as Kittredge recorded them from the tops of the trenches downwards, and in sequence as I understand Kittredge meant them to be. For the most part he gave no spot elevation readings for the strata, and there are very few obvious correlations across his drawn sections. In his daily entries, however, he sometimes noted the thickness of a deposit as the soil changed from level to level.

The purpose for all the Bronze Age constructions in the Northwest Area mystifies: are the uppermost levels of walls somehow related to the function of those underneath and, if so, what was that function? Hydraulics may have been involved – not just for storm run-off, but in an engineered system of water management (see further chapter 2).

A dark (black) stratum, 0.15 to 0.30 m thick, passes across the Northwest Area, sloping downwards from east to west with a top elevation of c.190.80 m at the higher or east side and falling to 190.70 m at the west. The respective bottom elevations are approximately 190.63 and 190.40 m. A shallow gradient of 2 percent for the top means that this black stratum levelled the slope of the ground below, but for all practical purposes the deposition in the Northwest Area was level. Tops and bottoms to transverse walls also are more or less level within and below this black stratum. Walls built above the black stratum, in particular the bottom courses of wall F and the post-Bronze Age set of transverse walls, drop along the somewhat steeper gradient of about 6 per cent over the same short distance of 3–4 m. This slope to strata and walls represents a minor distortion of stratigraphic levels, but in the face of the complicated factor of differing gradients for the various deposits and walls, Kittredge deserves credit for systematic notetaking, a fact that Blegen recognises (PN III, 45). Kittredge followed his excavation with photographs, but we have not seen these; only end-of-season 2 x 2-in photographs (probably taken by Alison Frantz) have been examined.

Wall Designations The alphabetic lettering system used in the original excavations for walls in the Northwest Area departs from the numbering of room areas used elsewhere in and around the palace, including at the uppermost levels of the Northwest Area, areas 82 through 87. Walls were instead designated by the letters A through Z and reduplicate AB through AN. The single letters follow no discernible topographic sequence, probably being assigned as walls were encountered in the course of excavations; the sequence nevertheless runs more or less from northwest to southwest. Walls that went unnoted during the Blegen excavations have been given herein a lower-case letter designation taken from a nearby recognised wall. This includes most notably wall b, but also walls j1, j2 and wall y1. The Papathanasopoulos-Kittredge plan outlines and labels some walls that cannot now be detected in any form, in particular walls AC–AH, all of which ran along the scarp and have probably been washed away. A couple of walls appear now as doubtful constructions (wall P and wall G).

An analysis of the pottery from the Northwest Area is here limited to sherds to which Kittredge assigns a chronological tag, records an elevation and notes the colour of the stratum. Unfortunately, the list is short. Two examples, numbers 14 and 17, were dated by Blegen and Kittredge to the chronological span ‘Helladic I / II’. George Papathanasopoulos identified a handful of incised Middle Helladic sherds from deep in his Trench 1 (see fig. 1.26); they are included as number 7 in Table 1. The table reveals a consistent correlation between estimated sherd date and level. The exceptions, numbers 19 and 27 – that is, the Late Helladic examples found in the otherwise unmixed uppermost Middle Helladic stratum – were, in fact, found inside the bodies of drains c and d, which are of Middle Helladic construction (see table 1.2) but were in use until at least Late Helladic IIIB.

The walls that frame rooms 83–86 have been assigned a label consisting of the numbers to either side of the wall. This has been necessitated by descriptions devoted to each and because of new associations of certain walls to different buildings. The new labels take the form of ligatures drawn from the contiguous room numbers. For instance, the designation ‘wall 85/86’ refers to the stretch of wall that heads in a southwestward direction from the corner opposite Circular Structure 87 and separated Room 85 from Room 86. New to the legend are patches of pavements; they are given letters drawn from nearby walls or areas: pav b, d, f, m, and pav 82, 83 and 84. This analysis recognises a distinction between a building wall and a retaining wall. A building wall is normally freestanding, whereas a retaining wall fronts and supports some sort of earthwork behind. In practice, a retaining wall may also serve as one or more sides of a building. In the Northwest Area, most of the walls discussed functioned as retaining walls; only a few may have served the double purpose. Besides a stratigraphic correlation between relative elevations and relative date, walls abutting, passing above or cutting through other walls

All Middle Helladic sherds come from levels within the deep ‘hard yellow’ deposit underneath the black stratum; only examples identified as Late Helladic I come from the black stratum or slightly below, while those assigned a composite date of LH I or LH II come from above the black stratum. The LH IIIA sherd at an elevation of 190.85 m (table 1.1, no. 25) falls within a hard yellow soil, as do the LH II examples (Table 1.1, nos 22–24). How these ‘dark strata’ in Kittredge’s sequence correlate one to the

49

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 no.

Kitteredge Trench

Notebook Page no.

Lot no.

Soil Colour

Description

Location in Relation to Dark Layer

Blegen Datum

MARWP Adjusted Datum (m)

Location

UTM

west of wall below base wall north near wall E

N122.0 E750.0

Middle Helladic 1 WK9

145

309

yellow

white slip with dark bands

below

11.45

190.35

2

WK7sw

85

169

yellow

good matt painted

below

11.43

190.28

3

WK7sw

87

174

2 matt painted

below

11.40

190.35

wall N middle

N125.0 E754.0

4

WK7nw

103

203

yellow below dark streaky yellow

conical cups?

below

10.90

189.80

wall N

5

WK7nw

131

270

MH

below

10.15

189.00

6

WK8

143

303

clear yellow yellow

MH

below

10.95

189.85

7

GP9

58

0

below

-1.70

190.10

8

WK3

59

7

yellow below black

deposit MH sherds MH dish

below

11.60

190.50

opening drain c below drain d at wall R opposite room 24 between drain b and wall P, near wall J

N125.0 E753.5 N125.5 E753.5 N121.0 E753.0

9

WK7a

108

203

hard yellow

MH deposit, bowls, cups

below

11.05

189.95

fi lls 2 sq.m at W near W

N127.0 E755.5

Late Helladic I 10 WK3

73

143

LH I pot

below

10.80

189.70

WK3–5

67

139

running spiral centre with eyes

below

10.90

189.80

12

WK3–5

73

143

LH I pot

at

10.80

189.70

13

WK7ne

75

146

LH I pot CWB

above

11.55

190.40

below wall G at O WK3–5–6 wall J floor, N corner WK3–5 below walls O and P wall NE

N130.0 E765.5

11

reddish dark on yellow yellow flecked with white dark red at yellow

14

WK7ne

79

159

LH I or LH II

above

11.25

190.10

dark, not like over yellow dark yellow with white

NE of wall K

Table 1.1 List of stratigraphic sections, soil colours and diagnostic pottery as recorded by William Kittredge.

50

N124.5 E755.5

N127.0 E773.0 N131.0 E764.0

N132.0 E764.5

N130.0 E762.0

N126.0 E755.5

N126.0 E756.0

The Architectural Study of the Palace and Ancillary Buildings 15

WK7se

87

176

dark yellow

LH I CWB

above

11.30

190.15

wall N

N126.0 E753.0

16

WK7sw

91

197

LH I Vapheio

at

11.10

189.83

wall N

N126.5 E753.5

17

WK8se

101

194

dark yellow with white reddish

LH I–II

above

11.70

190.68

wall

18

WK6

147

324

yellow below thin dark line

LH I Vapheio spiral dots

at

11.40

190.25

between walls B and E

N121.0 E753.0 N128.0 E764.0

19

WK8

149

323

yellow

LH I spiral ripple with white

at

10.10

189.42

N123.5 E749.5

20

WK7

78

165

190.40

WK7

71

133

conical cup PN III shape 10 handleless cup as no. 19

below

21

yellow below dark top of yellow

below

190.10

drain d below walls T and Y wall K near wall E wall K near wall E

Late Helladic II 22 WK6ne

43

73

LH II painted rim

above

12.00

190.85

at wall B-30

N129.5 E767.0

23

WK7ne

75

152

LH II

above

11.58

190.51

24

WK7ne

112

227

yellow with dark streaks dark yellow yellow, see no. 42

alabastron rim, PN I, 45

above

11.35

190.25

corner walls M and J at drain A

Late Helladic IIIA 25 WK7ne

113

224

dark yellow with white

12.00

190.85

walls M and J at drain a

Late Helladic IIIB 26 WK9e

135

273

dark

27

131

270

yellow

Post-Bronze Age 28 WK7

58

138

29

WK1

12

2

30

WK3

33

41

slightly dark greyblack white veins granular yellow

31

WK3

33

22

WK7nw

granular yellow

N126.0 E754.5 N127.0 E754.0

N128.0 E757.5 N128.5 E756.5

N128.0 E755.5

LH IIIB palace period LH IIIB in drain C

above

12.00

190.90

below

10.10

189.00

in drain c

miniature kylix iron knife

above

190.76

above

13.28

191.48

near surface K above walls B and C

N126.0 E756.0 N129.5 E766.5

miniature kylix

above

12.06

190.96

N129.0 E764.0

base with incised letters

above

12.58

191.48

in ashlar dowel cutting above wall G

51

N118.0 E751.5 N125.5 E753.5

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Phase 1

CWB Stratum 8b

1

Wall

Period b

MH

8b

drain c

MH

1

8b

drain d

MH

1

8b

drain e

MH

2

8a

AJ

MH

2

8a

C

MH

2

8a

j1 (= W)

MH

2

8a

O

MH

2

8a

P

MH

2

8a

V

MH

2

8a

W

MH

2

8a

y1

MH

3

7b

A

MH

3

7b

AB

MH

3

7b

H

MH

3

7b

j2

MH

3

7b

L

MH

3

7b

pavement d

MH

3

7b

steps h

MH

4

7a

B

MH

4

7a

pavement m

MH

5

6b

drain a

LH I

5

6b

drain b

LH I

6

6a

AG

LH I–IIIA

6

6a

AI

LH I–IIIA

6

6a

AL

LH I–IIIA

6

6a

G

LH I–IIIA

UTM

Azimuth

N128.0 E763.5 N123.0 E750.0 N122.5 E751.0 N131.0 E764.0 N122.0 E749.5 N131.0 E765.0 N130.0 E758.0 N126.5 E761.0 N130.5 E760.0 N123.5 E754.5 N125.5 E751.5 N122.5 E748.0 N129.0 E768.5 N122.5 E751.0 N128.0 E757.5 N129.5 E758.0 N126.0 E755.5 N130.0 E761.0 N128.5 E759.0 N126.0 E763.0 N123.0 E760.5 N127.5 E756.5 N130.5 E764.0 N128.0 E752.5 N125.0 E747.0 N119.5 E745.5 N129.5 E764.0

120

Top Elevation 189.44

Bottom Elevation 189.15

Trench

60

189.20

189.00

WK 7/8

160

189.76

189.56

WK 1/3

150

189.65

189.44

WK 3/6/1

170

189.97

189.77

WK 8b

152

190.98

189.90

WK 1

54

189.78

189.50

WK 5

50

190.06

189.30

WK 3/4

40

189.58

189.35

WK 5

162

189.90

189.30

WK 7

54

189.74

189.36

WK 7

34

190.50

189.31

WK 8b

135

191.30

190.80

WK 1

165

190.42

190.17

WK 8b/c

136

190.53

189.50

WK 4/5

57

190.43

190.00

WK 5

136

190.34

189.73

WK 7

146

190.22

190.08

WK 3

139

190.13

189.57

WK 4/5

49

191.23

189.26

GP 19/20 WK 6

132

190.78

190.66

GP 15

134

191.07

190.58

GP 16/20 WK 1

151

191.20

190.73

WK 4-May

46

190.60

190.43

WK 7

43

190.92

190.70

WK 8b/c

0

190.98

190.48

WK 8a

60

190.92

190.55

WK 3

GP 20

Table 1.2 Correlation of architectural features in the Northwest Area with MARWP phasing, Blegen’s strata, and original Kittredge and Papathanasopoulos trenches.

52

The Architectural Study of the Palace and Ancillary Buildings Phase 6

CWB Stratum 6a

6

Wall

Period

UTM

Azimuth

I

LH I–IIIA

N129.0 E757.5

135

Top Elevation 190.92

6a

M

LH I–IIIA

135

190.81

189.78

WK 7

6

6a

pavement b

LH I–IIIA

135

191.14

190.95

GP 15

7

5d

AK

LH IIIA

0

190.55

190.30

WK 8b

7

5d

E

LH IIIA

50

191.09

190.40

GP 17 WK 4/5

7

5d

pavement f

LH IIIA

145

190.56

190.40

WK 5

7

5d

platform Y

LH IIIA

0

191.47

191.22

WK 8a/b/c

7

5d

T

LH IIIA

82

191.18

190.36

WK 8a

7

5d

U

LH IIIA

172

191.17

190.05

WK 8b

8

5c

D

LH IIIA

135

191.07

190.40

WK 3/5

8

5c

F

LH IIIA

146

191.04

190.60

WK 3/4/5

8

5c

K

LH IIIA

135

190.88

190.60

WK 7

8

5c

X

LH IIIA

51

190.85

190.43

WK 7

9

5b

J

LH IIIB

55

190.78

190.43

WK 5

10

5a

N

LH IIIB

135

190.90

190.40

WK 7

11

4

S

LH IIIB

132

191.80

190.75

WK 8c

11

4

Building 82

LH IIIB

40

191.79

191.48

WK 9

12

3

Building 87

Geometric

0

191.32

191.06

GP 14/16

13

2

83/84

Temple 1

142

191.60

191.40

GP 18

13

2

84/85

Temple 1

142

191.62

191.41

GP 15/18

13

2

84/86

Temple 1

142

191.51

191.30

GP 12/15

13

2

85/87

Temple 1

142

191.49

191.25

GP 12/16

13

2

base b

Temple 1

142

191.91

191.70

13

2

base c

Temple 1

142

191.96

191.75

WK 9

13

2

pavement 82

Temple 1

52

191.62

191.40

WK 9

13

2

pavement 83

Temple 1

52

191.47

191.32

13

2

pavement 84

Temple 1

52

191.44

191.28

13

1

ZZ

Temple 2

N127.0 E755.0 N125.0 E762.5 N120.0 E 750.0 N125.5 E759.0 N129.0 E759.0 N120.5 E746.0 N117.5 E749.0 N121.0 E750.0 N130.5 E761.0 N130.0 E759.0 N126.0 E753.5 N125.5 E750.5 N130.5 E757.5 N126.0 E752.5 N122.0 E752.0 N113.0 E751.0 N126.0 E765.0 N122.5 E757.0 N123.0 E759.5 N120.0 E760.5 N124.5 E763.0 N117.5 E753.0 N114.5 E751.0 N119.5 E753.0 N120.0 E755.5 N120.5 E759.5 N121.5 E768.5

28

192.74

192.30

53

Bottom Elevation 190.35

Trench WK 4/5

GP 15

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 phase

cwb stratum

wall

period

13

1

ZZ

Temple 2

14

1

72/78

Palace B

14

1

82/83

Temple 2

14

1

83/84/86

Temple 2

14

1

Q

Temple 2

14

1

R

Temple 2

utm

azimuth

N109.5 E761.0 N107.5 E752.5 N113.5 E754.5 N118.5 E759.5 N119.5 E751.5 N121.0 E752.0

57

top elevation 192.65

bottom elevation 192.50

trench

52

191.88

191.27

142

191.91

191.70

WK 9

142

191.67

191.44

GP 13/18

52

191.90

191.62

WK 8b/c

52

191.79

191.30

WK 8c

Table 1.2 cont. Correlation of architectural features in the Northwest Area.

architectural stratigraphy. In terms of ceramic evidence, Kittredge was able to distinguish seven strata in a scarp located at the far west corner of the Northwest Area, but none of these could be associated with either specific walls or with building periods. Instead, the pottery from these strata was organised into five general assemblages. The deepest levels produced Middle Helladic wares and ‘some abundance’ of Late Helladic I, i.e. the transition from Middle Helladic to Late Helladic I (PN III, 45–46). Blegen speaks variously of seven strata and seven groups or assemblages of pottery; a column in Table 1.2 is my estimation of the strata according to our determination of the successive phases of wall construction. In nearly all cases, there are more phases of wall construction than Blegen’s seven strata and five pottery assemblages. Not unexpectedly, Blegen assigns the uppermost level of architecture to the final phase of the palace.

are later in date. Walls that lie at the same elevation and which run at true right angles or parallel to one another are considered as probably, but not necessarily, related. Figures 1.38–1.39 show seven vertical views, including the two major walls B and E and the five wide walls that run at cross-axes to these two, namely walls D, F, H/I, L/M, and K/N. The sections plot both faces of the chosen walls, shown in mirrored positions. The drawings are plotted with a vertical scale twice the horizontal, an engineering convention that exaggerates differences in elevation in order to make them more pronounced without losing accuracy. It will be shown below that the constructions of phases 4–10 in this area display a concerted effort on the part of the builders to enlarge a terrace, first behind the southwest sides of wall B and then behind wall E; towards the final phase, the terrace probably expanded all the way to wall J. A relative chronology with three main architectural stratifications can be established for the area. The earliest, dated to the Middle Helladic period (see also pp.140–45 below), comprises wall b, the Middle Helladic wall beneath Vestibule 104, structures in the Blegen MZ trenches, the Middle Helladic wall below Corridor 26, and well as drains c and d (Phase 1). The second stratigraphic phase includes drains a and b, associated with structures that lie beneath the Main Building, particularly Room 27 of Late Helladic I (Phase 5). The remaining phases follow.

At this juncture, it may be appropriate to introduce several observations about an architectural stratification overlooked by Blegen. Drains a and b emerge from the Main Building, passing underneath the exterior walls of rooms 22 and 27 respectively (Phase 5; see fig. 1.32). The two water lines have no apparent beginning points at drains or catch basins within the approximate quadrants of the Main Building, and by all appearances belong to a building period prior to that of the Main Building. The top of drain a (191.70 m) lies half a metre below the corresponding bottom of the exterior wall of Room 22 (192.25 m). Underneath the floor of Room 22, as well as in Room 21, Blegen exposed a set of walls two courses high (c.191.80–192.10 m). The concordance of elevations between drain a and the earlier walls in rooms 21 and 22 strongly suggests a common, chronologically synchronous and widespread building complex of walls and drains. The pottery recovered was not distinctive (PN III, 35), but the pithos-lined floor underlying the floor of Corridor 26 (see pp.142–43 and fig. 2.8) was assumed by Blegen to date to ‘an antecedent phase’; he also remarks on ‘considerable remains of earlier phases and periods’ here (PN I, 145–46).

A number of walls have been exposed by erosion at the extreme northwest edge of the area; these are not discussed below, but are clearly earlier than the Southwestern Building and appear, by orientation and elevation, to be related to the constructions of phases 3–5. Some Corrections to the Published Chronology The field notebooks of George Papathanasopoulos and William Kittredge have been a critical supplement to Blegen’s summary. The Papathanasopoulos notebooks include a few section drawings showing soil depositions, but although these are invaluable for stratigraphic analysis of ceramic sequences, they are less useful for evaluating

The top of drain b (c.191.80 m) passes 0.30 m below the bottom of the northwest wall of Room 27 (192.10 m). A

54

The Architectural Study of the Palace and Ancillary Buildings trench within Corridor 26 revealed another stretch of earlier wall, 1.85 m high and lying between 190.08 and 192.05 m in elevation (p.142–43 and fig. 2.8). Like drain a, drain b is clearly earlier than the northwest exterior wall of the main building. Blegen describes the stratigraphy in this trench as Late Helladic I in the upper stratum and exclusively Middle Helladic in the lower (PN III, 33); in other words, a Middle Helladic stratum that rises from c. -2.34 to c. -1.00 m below Corridor 26 floor level. It should be noted, therefore, that a transitional date between Middle Helladic and Early Mycenaean corresponds to what Kittredge determined to be the deep stratum of drain b in the Northwest Area. • Phase 1: Wall b and Drains c, d and e (fig. 1.27) Mycenaean earthworks along the northwest edge of the acropolis considerably altered the original topography in this area. The re-opening of the Blegen trenches permitted the taking of elevations and assessing the relative positions of the earliest walls here. The vertical sections in figure 1.27 show that footings to the lowest set of walls conform to a dropping terrain, southeast to northwest, more or less in the same direction as the existing acropolis slope; at present, however, there is a terraced slope of 80 per cent as opposed to a far more gradual gradient of the original ground line at 10 per cent, based on the evidence of the footings of the archaeological remains.

1.28 Wall b, below wall O, below wall E; wall B is at bottom right.

narrow space between the later walls O and B (fig. 1.28). The lower courses of wall b were not excavated, and it is exposed to a height of only some 0.25 m; but the wall clearly goes deeper. It runs in an approximate east–west direction, being at a bias to the grillage of overlying walls by approximately 45 degrees to the northwest. Its width, depth, relative position and orientation make a convincing case for wall b being a continuation of the Middle Helladic wall found beneath Room 104 of the Wine Magazine (see fig. 1.5 above). This wall has a width of 1.45 m and a height of 0.40–0.70 m in six courses, with a top elevation of 192.45 m lying at an elevation above stereo of 191.75 m; it has a bearing of approximately north 67 degrees west. A

Wall b does not appear on the PapathanasopoulosKittredge plan (PN III, fig. 311). It is, however, of special interest, being the lowest wall in the Northwest Area (top elev. 189.44 m) and the broadest (c.1.3 m wide). Only a metre of wall b is visible at the bottom and within the

1.27 State plan of the Northwest Area; phase 1 features in black.

55

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.29 Sections A-A’ and B-B’ as in plan 1.25, showing the sloping gradient of the terrain in the Northwest Area.

188.80. The short length of drain d as it heads underneath the southwest side of wall S has a top and bottom elevation of 189.76 and 189.36 m, which is to say the bottom of drain d just overshoots the top of drain c where their directions intersect. An actual juncture is obscured by the baulk underneath wall S; there is no way of telling whether drain d emptied into drain c.

wide stub, wall MZ36, located just outside the northeast end wall of the Wine Magazine, takes the same bearing and elevation, 192.43 m. These data correspond with those given above for wall b, which also takes a bearing of approximately north 67 degrees west. More significant is the fact that a scaled projection of the Middle Helladic wall under Room 104 brings it within one to two metres of the location of wall b. The difference in elevations between the tops of the Room 104 wall and wall b reflect a drop in slope of 2.06 m; in a distance of 28 m, this calculates at a grade of 8 per cent. As already noted, this evidence points to these stretches of wall probably having stood at the edge of the Middle Helladic acropolis, located approximately 6 m to the southeast of the present edge. The acropolis edge at this time also sloped downwards towards the southeast; it is two metres lower within the Northwest Area than at a comparable point under Room 104. Later earthworks, probably beginning in the Middle Helladic period and ending in Middle Helladic IIIB, raised and levelled this part of the acropolis and extended the edge by c.6 m.

The massive wall b appears to follow (more or less) the path of an original contour line as it rounded the acropolis, descending in the 8 per cent grade noted above. Drain c chases the slope as it drops from northeast to southwest. Drains d and e are roughly parallel, and they probably ran approximately perpendicular to the original contour. • Phase 2: Walls AJ, AN, O, P, V, W and yi (fig. 1.30) The drains continue in use to the LH IIIB palace period (phases 9–11), but their various orientations do not appear to dictate the orientations of walls built during Phase 2. Relative bottom elevations (189.30 m for wall V and 189.90 m for wall AN) and architectural stratigraphy place these walls in early phases. Elevations along footers to these walls make it possible to hypothesise the existence of a gully, perhaps a deep one, which passed underneath Circular Structure 87, extending towards the west and below wall V. Drains c and d (Phase 1) run under wall V. To the west of wall S, wall AJ appears to be the earliest feature.

Drains c, d and e lie at the lowest level of construction at the site: their bases lie at 189.00, 189.56 and 189.44 m, respectively. All are relatively substantial constructions of about 0.40 m square in section and have cover slabs of flat fieldstones. Elevations for drain c were obtainable at two positions on either side of wall S, into which is set a modern steel and cement drain. The top and bottom of the stone-built drain on the northeast side of wall S at the upside are 189.40 and 189.00 m, respectively; the same top and bottom on the opposite side of wall S are 189.20 and

The most distinguished wall of Phase 2 is wall O (see fig. 1.28), which does not appear on the published

56

The Architectural Study of the Palace and Ancillary Buildings of the two walls also differ by a short 0.08 m, making it probable that j1 and W are part of the same wall.

Papathanasopoulos-Kittredge plan (PN III, fig. 311). The base of wall O lies at 189.30 m and the wall is 0.76 m high, rising to cross over the top of wall b at 189.44 m, rather than cutting through it. The depth of wall b lies deeper than the base of wall O, making it probable that the latter is built on earth accumulated to either side of the former. The top course of wall O is relatively level at 190.06 m. A deposit of over 0.30 m separates the top of wall O from the bottom of the next construction in this elevation, wall E (Phase 7). There is no evidence that the top of wall O was partially demolished to make way for this intrusion.

Wall y1 is a handsomely coursed fieldstone wall not noted by Blegen but visible in the face of the scarp underneath platform Y (Phase 7) and wall Y; it underlies by 0.20 m the top of wall Y. Wall y1 is 1.19 m in height (190.50 m), and its base, at 189.31 m, is aligned with and at virtually the same elevation as a scrappy fragment (bottom elevation 189.36 m) exposed two metres further to the northeast. In his notebook (Kittredge 1962, 32 and 60) and on a working plan, Kittredge draws a wall, P, that starts just northeast of wall I and about midway between walls E and J. He shows this wall as single stone wide and a single course high and records an elevation of 10.39 m (189.29 m). Just before the line of drain b above (Phase 5), Kittredge’s wall P ends, and another bit of wall (Kittredge elevation 10.46 m, converted to 189.36 m), takes off to the east and passes underneath drain b. We did not encounter either of these walls, except for a patch of stones that is here designated wall P. The location falls along Kittredge’s location for wall P and occurs between walls D and F.

Wall AN follows the orientation of the earlier drain e, but its base (189.90 m) is 0.70 m above drain e. The wall is 1.08 m high, and the double-curtain construction contains an abundance of large fieldstones; this and its girth of 0.9 m suggest that it functioned as a retaining wall. The south end was severed by the construction of wall B (Phase 4), but it does not appear that the original went much farther to the south. The north end has dropped away with the erosion of the scarp on this side of the acropolis. Wall W, at the bottom of the cavity between walls M and K in the western corner, exposed by a partial collapse of wall X, appears to be a cross-wall related to wall L: the elevations differ, and a 0.60-m long section of wall L has been robbed away where it meets wall W, erasing evidence for an actual bond between the two walls. Two other stretches of wall, designated as j1, are overlaid by two later constructions, one on top of the other: j2 (Phase 3) and J (Phase 9). The two sections appear in the space between walls D/F and F/I. Wall j1 falls along the line of wall W further to the southwest. A base elevation of 189.50 m is a mere 0.14 m higher than that of W, and the overall height

• Phase 3: Walls A, AB, H, j2, L, steps h and paving d (fig. 1.31) Phase 3 saw the construction of the twin walls H and L. These are equal in width (0.85–0.89 m) and are built on slopes with footers inclined to the grade of the terrain: base elevations at the southeast (up-slope) ends were 190.17 m for wall H and 189.73 for wall L; at the northwest (downslope) ends the elevations were 189.50 m for H and 189.52 for L. Walls H and L are not only parallel to each other

1.30 State plan of the Northwest Area; phase 2 features in black.

57

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Adding to the puzzle of wall O is the status of its outside or northwest side. A baulk underneath wall E conceals this face; but just to the northwest, a large, oblong stone serves as what appears to be the first of two stepping stones, steps h. A top elevation of 190.13 m places the top step level with the top of wall O; the lower stone situated at the toe of the tread braces the slope between the two stones. The base of the lower step (189.57 m) comes at a floor level, which is level and arraigned above the gradient of drain c. The steps may have led to a patch of pavement (d) two metres to the northeast, later interrupted by wall D.

(north 44°10’ west) but are also perpendicular (at 90°5’) to the direction of the already existing wall O (north 45°55’ east). While this evidence might seem to associate the three walls, difficulties arise with a consideration of the elevations. The footer to wall H (190.17 m) lies approximately 0.10 m above the top of wall O (Phase 2; 190.06 m). Earth baulks that underpin the later walls B, E and I obscure the physical conjunction, if any, between walls O and H. Moreover, trenches farther to the southwest and between walls B and E were excavated only to a level equal to the base of drain a (190.58 m). This leaves in question the progress of wall O towards the southwest and the possibility of any physical bonding further along with wall L, which has a lower and appropriate base elevation of 189.73 m. In other words, the strict rectilinear shape to the plan formed by walls O, H and L suggests a large building, whereas walls O and H have, in fact, no common bond; this speaks against identifying them as a coherent or integrated construction.

Wall j2, impacted between the northeast–southwest succession of walls underneath wall J, is only visible between walls D and F and F and I. Wall j2 is not superimposed directly onto the top of wall j1 (= W, Phase 2). Instead, a 0.20-m layer of earth intervenes; in addition, the face of wall j2 is set back about the same 0.20 m from the face of j1. The evident purpose for the raising of this transverse wall j2 was to keep free the space for steps h and paving d and the activity taking place there.

In a later building period, Phase 5 (see fig. 1.32), walls H and L served as retaining walls carrying drain a. Wall O, on the other hand, must have been a freestanding structure. On its uphill side, the bottom of wall B (Phase 4) rises from the same level as wall O, namely the top surface of the Middle Helladic wall b. This means that the space behind (to the southeast) of wall O was not backfi lled. Furthermore, wall B failed sometime during its early history, due to the lateral pressures of soil and water drainage behind it; the upper levels of wall B were thrust forward, and a portion of its base, just above wall b and behind wall O, shifted outwards. The structural failure at the bottom of wall B is another indication that it was exposed and not shored up by pre-existing backfi ll behind wall O.

The construction of wall H and walls O and W, built earlier, resulted in a hollow, cavea-like space with wall H serving as a flank analemma. Wall O was the backing for an artificial slope that faces a low enclosing wall, W. The constructions of later phases preserve the essential features of this cavea-like space. The function of wall AN (Phase 2, above) presumably is that of a retaining wall supporting earth along its southeastern back side, above drain e (Phase 1). Higher in elevation than wall AN by 0.32 m is wall A, a substantial construction that starts at 190.80 m and rises three courses

1.31 State plan of the Northwest Area; phase 3 features in black.

58

The Architectural Study of the Palace and Ancillary Buildings to an elevation of 191.30 m. Only a sliver of this was was excavated by Kittredge in his trench 1; given this limited exposure, it appears to be either a replacement for wall C or a simple raising of the ground level behind. A column of packing stones fi lls the interstices between it and the later wall B (Phase 4) that is built against it.

a precipitous 67 per cent slope to the top of the Middle Helladic wall b. It seems likely that wall AN functioned as a retaining wall, and a deposit of earth formed a tip against its exposed side. The base to wall B follows the terrain, both the natural one and that shaped by previous earthworks. On the southwest side of wall b (Phase 1), the base of wall B quickly rises and passes over the top of wall H (Phase 3). In sum, wall B cuts through wall AN and continues over wall b, above and behind drain e, over wall H and, finally, underneath wall 84/85 (Phase 13), where it disappears from view.

In the tight, deep and triangular-shaped trench enclosed by the later walls Q, S and U are exposed not only the earlier drain d (Phase 1) but also a section of wall AB (base 190.17), which runs parallel and 0.25 m to the west of, but 0.20 m above drain d. It is built of squared and fitted stones three courses high at the south and two at the north. The function of AB is unclear; it follows that orientation established by drain d and followed by walls of subsequent building periods.

A width of 1.07–1.10 m makes wall B a substantial construction. Its uppermost 0.60 m consists of flat fieldstones with trimmed faces, laid in courses with alternating joints; ungraded rocks in a rubble construction make up the lower portion. A reused ashlar block, 1.0 x 0.25 m, adds a conspicuous piece of evidence for a previous and handsome structure somewhere on the acropolis. As described above, the girth of the wall turned out to be insufficient to restrain the lateral pressures of earth and groundwater behind it. Even though built as a retaining wall, it may also have served as one side of a structure: it makes use of the earlier wall A at right angles to it. A patch of flagged pavement m, below and visible between the later walls 84/85 and 85/86, is located at an appropriate elevation of 190.78 m and served as a floor inside a building formed by walls B and A.

• Phase 4: Wall B and paving m (fig. 1.32) Wall B not only represents the major structure of Phase 4, but also displaces or adapts previous constructions in the Northwest Area. Wall B abuts wall A (Phase 3), but the top elevation of wall B (191.23 m) is slightly below that of wall A (191.30 m). A careful fit of wall B where it meets wall A suggests that the two were intended to function together. On the northwest side, the footing for wall B at the juncture with wall A lies at an elevation of 190.55 m, but on the opposite or southeast side, the same footing has an elevation greater by 0.25 m (190.80 m); this implies that wall B was built through or parallel to a steep grade.

• Phase 5: Drains a and b (fig. 1.32)

Less than a metre to the southwest, wall B cuts through the earlier wall AN (Phase 2); two metres further along, just beyond the limits of wall AN, wall B drops along

The pair of covered water channels, drains a and b, cut through wall B at points 6.5 m apart; both incisions were fully incorporated by careful mends. The exposed bottom

1.32 State plan of the Northwest Area; phases 4 and 5 features in black.

59

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 of drain b shows that stone linings pass underneath the top of wall B and through its fabric. Reused ashlar blocks line either side of drain b, having been fitted as a repair into the partially dismantled gap left within wall B (fig. 1.33). Drain b originates somewhere underneath the northeast quarter of the main palace. A sondage sunk in Corridor 26 exposed a short portion of an early wall there (pp.142–43, fig. 2.8); it passes underneath Room 27 and emerges beneath the southwest wall of that room and the space between wall ZZ (Phase 13) and the northwest or back wall of the palace. With a right-angle elbow off Room 23, it heads to the northwest to run along the foot of wall C at the crown of the steep slope at a point where it has washed away. 1.33 Wall B where it is crossed by drain b, looking southwest. Reused ashlar blocks can be seen at the top of the photograph, lining drain b.

The intrusion of drain a through wall B prompted a careful reconstruction. Drain a also cuts through paving m and adopts the sub-course packing of that pavement as bedding for the base of the channel. Drain a descends from Room 22 of the main building for 4.5 m before disappearing beneath unexcavated baulk. The northwest branch rides on a deposit of earth that fi lls a cavity between walls H and L (Phase 3).

of the cavea and along its flanks. It is difficult to envision what happened to the two water channels once they reached the line formed by the present northeast face of the acropolis. The outer or northwest face to wall J (Phase 9), its predecessors and successors, became badly eroded over time. At present, drain a is at least one metre above the adjacent ground line and over two metres above the bottom of the slope just to the northwest. Construction of drains a and b through the Northwest Area provides a chronological link with pre-palatial building activity on the site. Phase 5 probably dates to Late Helladic I, that is, the period assigned in chapter 2 to Palace B; in any case, it occurred no later than Late Helladic IIIA.

The design and construction of these two water channels indicate serious water management engineering for the acropolis. The water channels begin somewhere beneath the Main Building, unquestionably in connection with earlier, underlying structures. The watercourses converge but do not merge. Their base elevations fall to within 0.15 m of each other (190.58 m for drain a, 190.73 for drain b). Instead of spilling into the pre-existing cavea created by earlier walls AN and H or into the natural gully at the foot of wall L, the two drains divert the water to either side

1.34 State plan of the Northwest Area; phase 6 features in black.

60

The Architectural Study of the Palace and Ancillary Buildings Y but at a level of wall tops, in general, of this Phase 6. A row of several ashlar blocks, three courses high, called wall AL, lies some four metres due south of wall AI. Wall AL appears to run not only parallel with AI but also in a line with wall AG, 10 m to the northeast. Wall AL is also parallel to wall j1 (= W; Phase 2) – not above, but rather a metre or so the southeast. The base of wall AL coincides with the top of wall y1 (190.50 m; also Phase 2); the former was probably a higher replacement for the latter.

• Phase 6: Walls AG, AI, AL, I, M, and paving b (fig. 1.34) The construction of walls I and M predate that of wall E, built in the following Phase 7, for wall E cuts through walls I and M. The later wall D (Phase 8) parallels wall I (within 0°10’), whereas wall M diverges from wall I by nearly 2 degrees. These walls hold in common elevations, profi les and features of construction, which point to contemporary building programmes. Wall I extended to the pre-existing wall B before being cut later by wall E; wall M may have extended to wall B, but its continuation to the southwest beyond wall E is obscured by baulks. These observations do not demand a chronological link, but they do require that walls I and M were built in relation to wall B. This building programme further alters the shape and limits of the cavea to the northwest of wall B. Moreover, the two walls rest atop a deposit of earth that lay against wall B.

A picture of the Northwest Area becomes more vivid with Phase 6. The earlier wall B (Phase 4) established the outer, northwest, limits of a terrace or, less likely, a superstructure in mudbrick. Below it, to the northwest, a cavea already existed, with steps h leading to the lower level. Afterwards, drains cut through wall B, drain a being supported by low retaining walls. In Phase 6, the analemma to either side of the cavea are raised by c.0.50 m, probably to cover over the top pair of water mains. This causeway between walls H/I and L/M, approximately 1.25 m deep, was covered by flagging behind wall B, either as a floor for a building or as paving for a terrace.

Walls I and M make use of pre-existing construction. Wall I, 0.6 m wide, lies on top of the wider wall H (0.85 m wide) at the northwest, but then mounts the slope of earth that accumulated against wall B. The bearings of H and I diverge by nearly a degree, indicating wall I’s use of H only where it was exposed and convenient. Their common footing descends in elevations following the gradient at the time of construction, from a height of 190.92 m (wall I) and 190.81 (wall M) to the bottom of the slope at elevations 190.35 (I) and 189.78 (M). Wall M has its footing sunk c.0.05–0.10 m behind and below the top inside edge of wall L (Phase 3). This means that both walls I and M added an upward extension by c.0.50–0.65 m to walls H and L.

• Phase 7: Walls AK, E, T, U, platform Y and paving f (fig. 1.35) Wall E, like the earlier wall B (Phase 4), transversely crosses the Northwest Area. Although wall E parallels B, it is clearly a successor. Wall E begins under the later wall D (Phase 8), but it cuts through walls I and M and overlies walls H and L. The northeast end of wall E begins four metres to the southwest of the corresponding terminus of wall B; the preserved southwest end of wall E, however, extends beyond the corresponding end preserved for wall B. At the southeast end and below wall D, wall E has been propped up by a single-course footer of thick, flat stones that protrude from the southeast end of the wall. Wall E probably ended at a point underneath the obscuring east corner of wall D.

Walls I and M traverse a downward slope to the northwest. Wall M is one course high (c.0.05 m) at the southeast corner, and five or six courses (c.0.90 m) at the northwest. This pair of walls in Phase 6 reaches beyond the present limits of wall J (Phase 9). The far eastern side of wall I preserves a short bit of a return, the barely detectable remnant of wall j2, which passes underneath wall F.

Throughout most of its length, wall E is 0.60 m high, except for those places where the footing passes over earlier structures. The top course has a consistent elevation of c.191.00 m until a point alongside the later wall K (Phase 8), where both the top and base of wall E rise by 0.30 m at its preserved end. It is clear that the Bronze Age builders carefully incorporated drain a within the body of wall E. It appears that wall E simply extended the retaining function of wall B to the northwest by two metres, a meager distance considering the effort involved. Two patches of stones, paving f, later covered over by wall F (Phase 8), consist of only a single course. While not flagstones, they nonetheless appear to be what is left of a stone floor. The top of these stones (190.56 m) coincides with the top of wall j2 (190.43 m; Phase 3), which survives below the later wall J (Phase 9). If these associations are correct, then the function of walls E and j2 was to raise the floor of the cavea between walls D and I by approximately 0.60 m and to provide the space with a solid floor. At this point, steps

At its northwest end (190.70 m), wall M extends behind and below the later wall J and returns to the northwest, forming wall AG (top elevation 190.43 m). One purpose of walls I and M is clear: they raise the heights of walls H and L to allow the coping stones of drain a to establish a new floor level. This was also the purpose of wall D (Phase 8): to sustain backfi ll for a leveled terrace covering drain b. Flagstone paving b is associated with this phase: part of it abuts wall B in the area of Circular Structure 87. Two additional remnants of pavement are found below Room 85 near and at its eastern corner. This flagstone paving inclines towards the area of the main building by a gentle 6 percent. Its overall limits have yet to be established. The outrigging for wall AI runs for around three metres behind the later platform Y (Phase 7); an elevation of 190.92 m puts it about 0.5 m below the level of platform

61

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.35 State plan of the Northwest Area; phase 7 features in black.

scarp at the northwest, whereas the northerly extension of wall U was cut by the later construction of wall S (Phase 11). Building U/T, therefore, appears to be related in date to wall E, also cut through by wall S. The footer to both walls steps downwards from a wall depth of 0.74 m at the intersection of T and U to an average wall height of 1.17 m for wall T at its exposed west end, and 0.87 m for wall U at its present north end. The single course is part of the 0.25m thick wall AK, which protrudes at right angles from the fabric of wall U. The preserved top of wall AK (190.55 m) is 0.63 m below the preserved top of wall U (191.37 m); as such, it must be a remnant of the wall U/T construction.

h and wall H go out of use. The opposite, southwest, end of wall E probably shored up the top of the slope to inhibit erosion. Early in his excavation of trench 3, Kittredge encountered a matrix of rubble. With further excavation he defined wall G as a distinct wall with splayed sides lodged against walls E and D. Wall G was re-examined, since Kittredge had excavated it, but we regard it as an earthen baulk between drain b and the juncture of walls E and D, and it does not appear on our state plans. Several major discoveries, however, were uncovered by Kittredge from strata above his wall G. In a stratum just below undisturbed surface (-0.70 m), Kittredge found a small wheel-made base, 0.043 m in diameter, with part of the stem preserved (table 1.2, no. 31). He records the presence of ‘base incisions – that may be two characters... A V = visible incisions.’ Although speculation is premature, one possible reading may be AUGE, an epithet of Athena. The form of the alpha cannot represent an accidental incision. The script for the alpha and upsilon are characteristic of the Archaic alphabet for Elis and Arcadia, as laid out by Jeff rey (1961, 206 and fig. 40). The base diameter falls squarely into Blegen’s pottery class no. 26: Kylix: Diminutive (PN I, 366). It will be argued below that this is a kylix simulacrum, a type of votive object used in the Archaic through Classical periods. In a lower stratum and resting directly on top of wall/baulk G, was found ‘part of a reused ashlar block with a possible dowel hole in its top [which] contained a kylix profi le’ (table 1.2, no. 30).

Kittredge’s published plan (PN III, fig. 311) and his working plan schematise the outline of platform Y as a rectilinear trapezoid. The state plan shows the actual configuration to be an arc of a probable pit of some sort, about two metres in diameter. Kittredge (1962, 89, 94, 105) recorded a fi ll of ‘loose stones’ above the level (190.50 m) where he found a spread of ‘many sherds’ belonging to a large vessel. Sometime in the past, a large tree, perhaps an aged olive, may have been uprooted from this spot. Kittredge also distinguished between an overlapping wall Y and the ‘platform’ that projects from it to the southwest. The two appear to be the same construction, platform Y being a surviving portion of the packed-stone base for a plaster floor in this area. Platform Y was constructed on an earth and rubble fill varying in depth from 0.50 to 0.70 m, as well as over the top of wall y1 (Phase 2) and the ashlar wall AL (Phase 6). In its present condition, the uneven top surface of platform Y varies by 0.26 m, between an elevation of 191.22 and 191.47 m. The top of the plaster floor

Walls T and U appear to be a corner of a single building with north–south orientation. Wall T extends into the

62

The Architectural Study of the Palace and Ancillary Buildings at the northeast at 191.38 m probably represents the original surface elevation. This level registers platform Y within or just below Kittredge’s Stratum 4; that is, 191.22 to 191.48 m and within the Late Helladic IIIA period. The plaster floor suggests the presence of a ‘refined’ building in this area, but it is unclear how it relates to other pre-existing or contemporary constructions of the time. Platform Y was not only cut through by the pit mentioned above, but also by the deep wall S (Phase 11). • Phase 8: Walls D, F, K and X (fig. 1.36) In its function as a retaining wall, the outer, northwest face of wall E attacted a deposit of fill. Footers to walls F and K were built to climb this slope; towards their outside, northwestern ends, their bases share an elevation of 190.60 m. Their top courses abut wall E; in the case of walls D and F, several of their uppermost stones overlap the upper surface of wall E (fig. 1.37). Walls F and K are shallow throughout their lengths, ranging from 0.15 m to 0.30 m in depth. They differ in construction technique: unsorted rocks with a deep thickness characterise wall K, whereas wall F consists of uniformly large plates of thick stones; the largest is 0.36 m thick.

1.37 Walls F (left) and D (right) at their juncture with wall E, looking northwest. Note the triangular stone at the near right corner of wall F as it overlaps wall E.

3)). The preserved exterior return to wall X aligns with the outside edge of wall J (Phase 9). Walls X and K bear comparable elevations: at the top c.190.85 m (X) and 190.88 m (K); at the bottom 190.43 m (X) and 190.60 m (K). Wall J was built in segments, closing the gaps between walls F and I, and I and M. Wall J is later than wall F, which therefore makes wall K/X later as well. The situation with wall D is somewhat clearer: two flat stones within the space of a projected extension of wall D to the southwest rest on top of the uppermost construction level of wall B. Moreover, a thin layer of reddish earth or mudbrick, uncharacteristic of the matrix within wall B, separates the two stones and the top of wall B. Wall D on

Other evidence also indicates a different programme for the respective constructions, wall K being the later of the two. At the northwest, wall K returns to form wall X (top elevation 190.85 m). This return covers a portion of wall AG, which belongs to wall M (Phase 6). (Wall X appears to have collapsed, exposing a portion of wall W (Phase

1.36 State plan of the Northwest Area; phases 8, 9, 10 and 11 features in black.

63

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.38 Vertical views of walls D, F, H/I and M/L. The elevations show both faces of the walls, in mirrored positions. Drawings are plotted with a vertical scale twice the horizontal, exaggerating differences in elevation for clarity without losing accuracy.

at the back between walls F and D in figure 1.37 provide a more realistic picture of the condition. As elsewhere, a succession of walls appears at the northwest in the intervals between the series of walls D, F, H/I, L/M and K. The lower walls – W and j2 – visible between walls D and F and walls F and H, have been mentioned above. The top level, seen as wall J between walls D and F, abuts wall F.

the southeast end is only two courses in height (c.0.20 m), but on the opposite side it rises to approximately six or seven courses (c.0.75 m). • Phase 9: Wall J (fig. 1.36) As noted occasionally above, visible remnants of early walls still appear below walls J and X. The outside or northwest face of walls j1 (= W, Phase 2) and j2 (Phase 3) has been fronted by a modern retaining wall. Walls AC, AD, AE, AF, AG and AH, recorded by Kittredge and plotted on his plan (PN III, fig. 311) are no longer visible. Even before the Kittredge excavation, this side of the Northwest Area had became badly eroded, making it difficult to distinguish the various walls and relative building periods at the upper levels. This is particularly true at the western corner, where walls X and W mingle in no definable way. To a lesser extent, the same may be said for the condition at the north corner of the site, where the walls, as sorted on the Papathanasopoulos-Kittredge plan, need further clarification. As noted above (Phase 2) the fugitive wall P either never existed or consists only of an edging of stones seen between walls D and F. The letter on wall J and the plan for the major cross-wall at the northwest part of the Northwest Area must therefore be regarded as no more than a generic designation. The masses of stone shown

1.40 View of the Northwest Area looking southeast. In the centre foreground is wall K (Phase 8), with wall N (Phase 10) to its right and wall V (Phase 2) running diagonally at a lower level.

64

The Architectural Study of the Palace and Ancillary Buildings

1.39 Vertical views of walls B, E and K/N, showing both faces of walls B and K/N in mirrored positions. Drawings are plotted with a vertical scale twice the horizontal.

65

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 • Phase 10: Wall N (fig. 1.36) Wall N (top elevation 190.40 m, bottom 190.90 m) succeeds wall K (Phase 8; fig. 1.40) and was perhaps built to shore up the exposed portions of supporting fi ll for its predecessor. Mounted over an accumulation of earth above wall V (Phase 2) some 0.50 m high, wall N begins 1.20 m from wall E at the toe of a slope represented by the point where wall K descends from wall E and then levels out. Carefully laid reused ashlars (1.0 x 0.25 x 0.40 m) comprise the first course of the wall (bottom elevation 190.40 m; top 190.60 m). Above this level, the construction becomes indifferent rubble, placing the uppermost point (190.90 m) level with wall K (190.88 m). • Phase 11: Wall S and Building 82 (fig. 1.36) The deep wall S (c.1.25 m high) has been obscured by the modern terracotta drain and its cement bed. Both the top and bottom of wall S decline at an 8 per cent gradient from southeast to northeast. Likewise, the bottom of the northwest wall of Building 82 is lower than the southeast side by nearly 0.50 m, or the same 8 per cent declination. It appears that the foundation trench for wall S cut through wall U, platform Y and possibly the coursed wall y1 (underneath platform Y). Building 82 and wall S share a top surface elevation of 191.72 m +/-0.05 m, and an apparent bonded wall return at their juncture at the common inside corner. A depth of 0.5 m for the northwest wall of Building 82 makes it considerably less than the 1.25 m of wall S.

1.41 Circular Structure 87, looking southwest along wall B.

A thin, yellow or yellowish layer of earth, 0.05–0.073 m thick, accumulated over the site after the Bronze Age. This deposit was found in the Northwest Area below the walls of Circular Structure 87 and those of the rectangular Building 83/84/85. Beneath the walls of 83/84/85, but not below Circular Structure 87, was an intervening pink layer 0.13–0.23 m thick (see p.240 below). Thus some period of time elapsed between the construction of Circular Structure 87 and Building 83/84/85. Neither the yellow nor the pink stratum was countenanced by Blegen in setting forth his seven stratigraphic phases.

The wall of Building 82 at the west corner has disintegrated, effacing evidence for either a free-standing return for a building corner or for a wall bonded into the corner at the northwest side and continuing further to the northwest. Given the extension of wall S to the northwest, this hypothetical flank wall probably continued as well to the northwest, as recognised by Blegen (PN I, 290). This means that the northwest wall under discussion may actually be a cross-wall to an annex for a larger building at least 14 m long and 11.5 m wide; it has a southeast-to-northwest orientation. The stratigraphic position at elevation 191.72 m of Building 82 and wall S places the construction(s) within Kittredge’s Stratum 3; that is, in the LH IIIB period, the chronological designation as published by Blegen (PN III, 43–46).

Circular Structure 87 (fig. 1.41) represents a curiosity by any account. The function of this type of construction is a matter for speculation wherever it occurs, especially in cases such as this one, where context is uncertain. Blegen (PN I, 293) devotes a thin paragraph to the building: ... an overall diameter of 3.10 m. The wall, 0.35 m. to 0.40 m. thick, is built on two courses of small rough fieldstones laid on earth. ... a transverse partition, made of a single line of small platted stones set on edge in earth, runs across from southwest to northeast, dividing the space inside into not exactly equal sections. No proper floor was found; nothing of any consequence was recovered within the circle to indicate its character or the purpose for which it was built.

• Phase 12: Circular Structure 87 (fig. 1.42) The post-Bronze Age building activity in the Northwest Area falls into three phases. The first two, here designated phases 12 and 13, are distinguished one from the other by soil colour differentiation in the stratigraphic notes kept by Kittredge, as well as the earlier excavation notes of Papathanasopoulos. Todd Brenningmeyer treats the total evidence for phases 12–14 in detail in chapter 8; I summarise here.

The investigations of 1994 expanded somewhat on Blegen’s description of the physical properties of Circular

66

The Architectural Study of the Palace and Ancillary Buildings Structure 87. Besides rough fieldstones, huge chunks of baked mudbrick are also incorporated into the fabric of the wall. The circular wall is not horizontal but tilts quite conspicuously, approximately 0.50 m downwards from a high point (191.81 m) at the southeast to that at the northwest (191.32 m). The interior face of the wall lies 0.24 m deeper in the ground than does the exterior face. The interior face at the north rises to a height of 0.55 m, in contrast to a height of 0.26 m for the exterior. The floor begins at elevation 190.77 m. The central interior dividing wall marks an upward step between two sunken floors, the upper being 0.22 m above the lower at the northwest.

Double-curtain constructions, existing heights of two courses (c.0.15 m), common widths of 0.80 m, and bonded joints all characterise these walls: long portions of crosswalls 85/87, 84/85, 83/84 (underneath 83/84/86 of Phase 14) and the southeast flank wall, 84/85/86. This set of walls makes a coherent plan, and all are well preserved, from the bonded east corner through the bonded cross-walls 84/85 and 83/84 to a point slightly beyond the limits of wall 83/84. A stub, made up of a couple of protruding stones underneath the overlying and later cross-wall 83/84, shows that the southeast flank wall 84/85/86 continues into Room 83.

An ashlar block lies just to the east of Circular Structure 87 (see p.241, fig. 8.11). This block is bedded on the same earth deposit that separates the constructions of phases 12–14 from the Late Helladic levels. Blegen, however, states that ‘a squared block of poros from the rear wall of the main building was found lying where it fell, leaning against the eastern side the circle’ (PN I, 293). In fact, this ashlar lies 0.15 m east of the outside wall of the circular structure, and according to the Papathanasopoulos’ excavation notebook (Papathanasopoulos 1958, 62–64 and sketch p.63) it remains in situ to this day. In any event, this block did not drop from the palace wall, but was one of many similar blocks that came from a stockpile that remained available from the time of the construction of the Middle Helladic wall B to the Geometric constructions in rooms 89, 90 and 103 (see pp.41–43, 76, 240–42).

The line of stones abutting the inside of wall 85/86 and within Room 84 is all that remains of a stone floor: paving 83, one course thick. This construction is matched by a triangular patch of paving 84, which adjoins the southwestern edge of wall 83/84. For the most part, the stones are flat plates in the approximate form of cobbles or flagstones, but they are ungraded in terms of thickness; moreover, whole or partial stacked mudbricks are included in this pavement. The northwest side of the pavement forms a straight line demarcating the edge where it abutted the inside face of an enclosing wall. This wall was robbed when the overlying wall 83/84/85 was constructed. A second trace of this flank wall appears at the northwest end of cross-wall 84/85, but the north, rear corner of Building 83/84/85 has completely disappeared. In a brief summary, Blegen (PN I, 291) conflates rooms 83, 84, 85 and 86 into ‘walls of a later house’; he does not explain why such a house was built next to a Mycenaean palace.

• Phase 13: Building 83/84/85 (Temple 1), with bases a, b and c, pavements 82, 83 and 84, and wall ZZ (fig. 1.42)

An analysis of the physical remains from phases 12 and 13 leads to a coherent picture of two buildings. Circular Structure 87 does not require a hypothetical restoration

Four extant walls form a multi-chambered building consisting of spaces occupied by rooms 83, 84 and 85.

1.42 State plan of the Northwest Area indicating the features of phases 12, 13 and 14 and proposed outlines of temples 1 (black and extended lines) and 2 (black outline and extended dotted lines).

67

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 of the already-extant Circular Structure 87, which blocks passage at the northeast end, and the fact that the adjacent wall 85/87 shows no sign of an emplacement for a sill. A western orientation defies the basic rule of an eastern orientation for a typical Greek temple, but in Arcadia, Triphylia and the Argolid there are examples of north-facing temples. In the wider Greek world there are examples of temples facing to the south (Mycenae and Athena Pronaia at Delphi) and to the west (Sardis).

of plan, but that of the rectilinear building does need coaxing; nevertheless, a multi-chambered megaron design can be discerned in the tangle of under- and overlying walls. The northwest flank wall of Building 83/84/85 can be readily restored from the evidence of the robber’s trench mentioned above; the southwest end is not preserved. In any case, the dimensions of the building measure 16.97 x 4.64 m. There is a succession of three chambers, the back two of which are 1.60 m and 3.21 m deep. Two column bases, b and c, survive in Room 82. A third, base a, lies just outside the northwest flank wall at midway; its associations are unclear. Blegen describes and illustrates the locations of the bases (PN I, 289–91, and note 1; PN III, fig. 311). There are discrepancies, however, between Blegen’s dimensioned references to the walls of Building 82 and scaled distances taken from the plan. Between bases b and c, Blegen cites a distance of 2.92 m from the northeast wall of Building 82; the plan scales at 1.90 m, and we obtained a distance of 2.80 m. The dimensions from centre to centre between the bases also disagree by corresponding values; we measured a distance of 3.70 m between the two bases. Blegen (op. cit.) observed in passing that the alignment of bases b and c runs askew to the axis of Building 82.

For Wall ZZ, also dated to Phase 13, see pp.70–73 below. • Phase 14: Building 83/82/81 with base e (Temple 2), walls Q, R and 82/83 (fig. 1.42) More than two weathered and reused ashlars (the Bronze Age angularia) form a remnant of wall R (elevation 191.79 m), which adjoins wall Q at a right angle and forms a low step (0.11 m high) below it. The wall R ashlars lie on a rubble base that extends 0.5 m to a depth of 191.30 m. Not enough of this spur is extant to make sense of it, but to hazard a guess, the exposed southwest side may have been a wall face made to leave visible the column base a, from Phase 13.

It is clear now that column bases b and c ran along the centreline of a building, parallel to and 0.87 m inside the southeast flank wall; they are here reconstructed as standing at the front of a monostyle in-antis temple. A spacing of 3.99 m from the front edge of base c (in line with the antae) to the centre of base b is nearly the same as that across the rear chamber (4.01 m). The two ends of the temple are designed for a double-hipped roof: the inner wall and the inner column would have supported a ridge beam, from which splayed pairs of rafters would have dropped to the corners of the building. This design was meant to receive proto-Corinthian hipped roof tiles, fragments of which have been recovered and are discussed in chapter 8 below. An isometric drawing of the Northwest Area by Hendrik Hendrickx (fig. 1.43) gives perhaps a more readily understood picture of the temple remains.

About a third of the 9.5-m length of wall 83/84/86, orientated southeast–northwest, perches on top of the earlier cross-wall 83/84 of Building 83/84/85. The southeast terminus of wall 83/84/86 was carefully assembled, having the appearance of a quoined wall end. This portion of wall, two courses thick, is bedded on earth fi ll until it approaches the corner of Building 83/84/85. For the next 3.5 m, a single course of wall 83/84/86 rests on the earlier cross-wall 83/84, respecting the 0.80-m width of its support. When it reaches the opposite corner, the shallow depression left by the robbed flank wall of Building 83/84/85, the wall again becomes two courses in height. In this latter short stretch, wall 83/84/86 gradually splays to a width of c.0.95 m. Unlike the appearance at the opposite end of the wall, this end shows that wall 83/84/86 continued towards the northwest.

Arguments for placing the entrance and the facade at the southwest end of the building include the close proximity

Wall 83/84/86 not only took its orientation from the earlier cross-wall of Building 83/84/85 but also lapped onto the flank wall, 85/86. This length of wall is double-curtain but differs from the earlier construction: it contained a matrix of small stones laid within a disproportionately large amount of mud mortar. At first, the purpose of wall 83/84/86 was unclear, for it appears as an isolated feature that does not join with any other wall to form a recognisable unit. The fashioned southeast end points to a building corner or another wall that it abuts. The space towards the southeast is interrupted by what may be a stub of wall ZZ. To the west is part of wall Q (3.5 m in length, elevation 191.90 m), which has an orientation perpendicular to that of wall 83/84/85. Wall Q does not appear on the published Blegen plans, but is so labelled in the notebooks. It is slightly skewed to the course of the Bronze Age wall E to the northeast (Phase 7) and is higher

1.43 Isometric rendering of the Northwest Area, showing Temple 1 remains with column bases b and c, and wall ZZ.

68

The Architectural Study of the Palace and Ancillary Buildings

In a cross-section of his southwest baulk in trench 8c, Kittredge (1962, 92) noticed quite different stratigraphic deposits that surround and cover the walls of Building 82 and wall Q. Building 82 is covered by a spread of ‘dark veined’ earth, whereas wall Q was encased on either flank by a three-layer deposit of (bottom to top) yellow, red and black soil. This evidence confirms the sequence of later building phases associated with Circular Structure 87 and the walls of the temples. It also distinguishes the palaceperiod Building 82 and its annex, represented by wall S (Blegen’s ‘additional chamber’; PN I, 290), from a second post-Bronze Age temple.

Building 82, here called wall 82/83. Its southeast end aligns with the quoined end of wall 83/84/86. While a width of 1.25 m makes the wall thicker than the rest of the preserved circuit of a hypothetical temple, it nonetheless runs parallel to wall 83/84/86; the southeast portion falls within the restored thickness of this flank wall, and the top elevation at 191.90 m equals that of the other wall fragments. It is suggested that this is a cross-wall within a replacement temple, Temple 2, built on top of the front portion of its predecessor, Building 83/84/85, Temple 1. There are no other apparent walls to associate with the second building, and this proposal is spectulative, but quite possibly the temple reached to the area of Room 78 of the Southwestern Building, where the top of wall elevations come to about 191.70 m.

Probably belonging to the same building is a short stretch of unnumbered wall that overlies the southeast wall of

A few observations are in order. The two sets of foundations for the successive temples are clearly, if not precisely,

in elevation by 0.60 m. Wall S and the modern drain built on top of it obscure the eastern portion of wall Q.

1.44 Aerial view of Room 82 and buildings to the southwest, with lines indicating the proposed southeast f lank wall of Temple 2, crosswall 83/83, and rear wall 83/84/86 (in the shadow of the shed roof).

69

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 (Isthmia) and 14 (Mycenae)). That said, the origin of the ashlar blocks at Pylos may be Mycenaean, as assumed by Nelson (see Part II, pp.371–72), and if so, the 0.80-m width of the foundations of Building 83/84/85 was determined by the size of these ashlars.

aligned: they stand along the same axis, although the centreline axes are shifted by 1.38 m. A contributing factor may have been the need to avoid the steep slope just to the northwest of the second building. In addition, the rear of Temple 2 has been pushed forward, freeing space around Circular Structure 87. The fact that the second temple is not concentric with the first eliminates the possibility of reconstructing a peristyle added to the original plain sekos. The shift, however, resulted in the columns of the earlier building falling along an axis inside the second temple and 3.10 m from the centreline of its northwest flank wall, to make a subdivision into thirds. It is possible, therefore, that the successor temple reused the first building’s column bases and further subdivided its own interior space into three aisles by two rows of columns, the bases of which are now missing. Obvious in aerial photographs (fig. 1.44), but not on the state plan, are the series of intrusions that occur along a hypothetical extension of the southwest flank wall to the southwest as it cut through the northwest-to-southeast interior walls of rooms 78, 77 and 81. The facade of Temple 2 probably extended no further than the edge of the retaining wall at the western corner of Room 81. The width of the building, c.9.8 m, doubles that of its predecessor at 4.7 m. At a guess, a length of c.28 m was also double that of the predecessor: a megaron plan enlarged to a hekatompedon.8

• Wall ZZ (fig. 1.42) A conspicuous construction separates the Northwest Area from the remainder of the acropolis. The long stretch of wall begins at the west corner of Room 27 and runs in three legs obliquely to the line of the rear wall of the Main Building for a distance of about 21 m. Due northwest of the axis of the southwest flank of the palace wall, at a point approximately two metres from the palace, this wall turns at a slightly oblique angle and joins the west corner of Room 21 at its northwest side. The wall then recommences around the west corner of Room 21 on its southwest side, where it crosses Court 88 and abuts the Southwestern Building at the corner projection in the northeast flank wall outside the southeast wall of Room 76. For convenience, in the following description, I assign the letters ZZ to all of this wall, as it can be traced in a continuous construction from the northeast corner of the main palace to the Southwestern Building at Room 76. The 21-m stretch behind the main palace building falls underneath the protective shed and has been a long unnoticed feature of the site. In his final publication, Blegen does not mention its existence, although the wall appears on his Key Plan at the back of volume I of The Palace of Nestor. MARWP, likewise, all but overlooked wall ZZ, although it crisscrossed the workspace for the various activities at the Northwest Area. We used it as a convenient bench but, unfortunately, did not uncover and investigate the base of the wall underneath the protective shed, nor that of the larnax (fig. 1.45) that lies next to it. We did take photographs from various viewpoints, and cleaned the stretch across Court 88 and the legs to the northwest and west, and we referenced dimensions. The following discussion is drawn from Blegen’s roomby-room descriptions and published photographs in The Palace of Nestor I, as well as from the occasional textual reference in his annual reports and from sketch plans in

Two fragments of roof tiles remain embedded in the matrix of wall 78/82, and a piece of Lakonian pan tile of reddish clay juts out from the wall. Todd Brenningmeyer (p.237, fig. 8.8) associates this tile fragment with surviving remains of a Lakonian roof with fictile revetment, dated to the Archaic period. In any event, this piece provides solid, in situ evidence for continued building activity on this part of the acropolis into the Archaic and Classical periods. There are two storage piles of dressed limestone architectural blocks on the acropolis: a large heap near the main entrance and above the asphalt road, and another at the south edge of the site against the cyclone fence. The origins of these architectural blocks are unknown. Between the two piles there are at least three ashlar blocks, 0.80–1.00 m long x 0.35 m high x 0.40–0.50 m wide, which contain very large and deep swallowtail cuttings, 0.28 x 0.25 x 0.13 m deep. The size of the cuttings is uncharacteristic of typical Late Archaic and Hellenistic clamp cuttings. Swallowtail clamps of this magnitude occur in Roman architecture, for instance in the Odeon of Corinth. In this case the blocks are huge, but the examples of swallowtail cuttings at Pylos occupy a disproportionate part of the top of the blocks. The closest parallels for the broad, deep swallowtails on smaller blocks occur on dressed building stones belonging to the Early Archaic temples at Isthmia and on the summit at Mycenae (Rhodes 1983, 32, figs 17 8

It is worth remarking that a deposit of ash and burnt bones was discovered at the hypothetical southwest anta of Temple 2 (Stocker and Davis 2004, fig. 1).

1.45 Terracotta larnax alongside wall ZZ.

70

The Architectural Study of the Palace and Ancillary Buildings stones fi lled the space between the uppermost surviving portion of wall ZZ and the rear wall of the palace. The stones appear to have overlapped the robbed-out palace wall (shown reconstructed in PN I, fig. 24). Blegen associates the rubble with a collapsed wall of the Main Building (PN I, 46), but this assumption does not accord with the evidence that the rear wall of the palace had been robbed to the foundations, or the fact that the ‘spread’ corresponds to the top of wall ZZ, seen to the right in the published photograph (ibid.). In the preliminary report for the 1955 season (Blegen 1956, 99), he remarks on signs of later construction for wall ZZ: ‘Trenching outside the palace toward the northwest revealed two more walls which have still to be examined in detail.’ A few sentences earlier he notes, ‘There are some puzzling features in this area, but they can probably best be explained as indicating changes and modifications in successive phases’. In the final publication (PN I, 130) Blegen elaborates:

various Blegen-period field notebooks. Equally crucial are the unpublished working plans drawn by Theochares and Papathanasopoulos. Taken together, this evidence provides a clear idea of the appearance, purpose and probable period of wall ZZ. Strips of wall ZZ began to appear in test trenches dug by Rolfe Hubbe in 1955 (PN I, fig. 407). The stretch of wall at the northwest end of Court 88 was exposed by Elizabeth Blegen the next year (PN I, fig. 408). In 1958 George Papathanasopoulos (in his trenches 1 and 8) cleared about 3.4 m of the wall, from its juncture with the west corner of Room 27. By 1959, wall ZZ was cleaned except for two baulks left between it and the reconstructed rear wall of the palace. The back wall of the Main Building, behind Rooms 21 and 22, had been robbed to its 1.20-m wide foundations (Blegen 1957, pl. 41, fig. 5) and was rebuilt by the Greek Archaeological Service, along with robbed sections elsewhere, making it 0.90 m wide and 0.65 m high (PN I, 46, nn. 4, 5). The original width of the northwest wall of the palace is thus unknown. In contrast, the bottom portion of wall ZZ is very much extant. Two doglegs break an otherwise straight course of wall that runs from Room 27, dividing it into approximately three equal segments of 7.0 m. The elbow that returns to meet the corner of Room 21 is the fourth segment and the leg across Court 88 is the fift h.

All that is preserved of the Northwest wall is a stump of 0.35 m. high (Fig. 99). This section of the exterior wall of the building, alongside the northwest sides of Rooms 21 and 22, seems to belong to reconstruction of the western corner in one of the late phases of the palace, and it diverges somewhat from the original line. Blegen does not raise the possibility, as illustrated in the Key Plan, that this later wall construction returns at the west corner of Room 21 in a westerly direction across Court 88 to join the projecting exterior wall of Room 76 in the Southwestern Building. In the discussion devoted to Court 88 (PN I, 293–95), he emphasises Rooms 89 and 90 as intrusive and notes the presence of Geometric pottery and Corinthian roof tiles spread across the court (see also p.229 below); wall ZZ is not mentioned. In a later discussion of ‘walls of later houses’, rooms 83, 84 and 85, Blegen (PN I, 291) associates this length of wall with these ‘houses’ and considers it to be an enclosing wall for them.

Papathanasopoulos’s plans (Papathanasopoulos 1958, 12 and 65) and his short descriptions comprise the single notations for this wall. In the same trenches 1 and 8, Papathanasopoulos found a larnax; it appears on the Key Plan (PN I) but otherwise is not mentioned by Blegen. The rim is situated some 0.30 m northwest of the first leg of wall ZZ and aligned with it, and about 1.2 m from the west corner of Room 27. The larnax measures 0.78 x 0.53 x 0.46 m deep (ibid., 13, plan p. 12). Burned bricks in reddishyellow earth ran around it, and a floor of ‘clearly arranged small stones’ extended to the west (ibid., 62–63, Trench 2B), probably edging along the base to wall ZZ.

In the preliminary report for 1956 (Blegen 1957, 130), Blegen recounts in greater detail the results of the excavation of the northwest portion of Court 88: ‘immediately below the surface, a compact layer of smallish stones closely packed in blackish earth that seemed to have a greasy texture.’ He compares this layer with the same stratum that extended over other portions of the site, which has since been proven to be of Geometric and later date. He continues (ibid., 131):

The bottom of wall ZZ and the adjacent larnax fall within Papathanasopoulos’ fourth stratigraphic layer (0.80 to 1.00 m) below the surface. This he describes as composed of thick earth, yellow-reddish with abundant pottery, traces of charcoal and a quantity of burnt brick. There was a trace layer of gray with ashes (ibid., 15–16, with a stratigraphic cross-section). Remains of bones were recovered in and above his layers 4 and 5. The two strata, for all intents and purposes, correspond in soil colour and elevation (191.20 m) to the pink-yellowish and pink strata that subtends the post-Bronze Age buildings of Phases 12–14 discussed here and in chapter 8.

How the court [88] terminated toward the northwest is not yet clear. In a late phase it seems to have been shut in by an oblique stone wall, differing in orientation both from the old and the new wings of the palace, and perhaps belonging to a shed-like storeroom or stable that was erected farther to the northwest. Because of the extreme shallowness of the earth covering these remains and much disturbance in modern times it is unlikely that useful evidence for close dating has

Wall ZZ stands to a height of approximately 0.40–0.50 m (top c.192.75 m, bottom 192.30 m) in the 21-m stretch at the back of the Main Building. The portion across Court 88 cuts the stucco pavement and has been reduced to a single course approximately 0.15 m high (192.65 m). A spread of

71

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 survived, but the oblique building must presumably be attributed to the final stage of occupation of the palace.

almost exclusively by its foundations in which one large squared block and several stones appear among the rubble.

In his final publication, Blegen devotes only a couple of sentences to the southwest and northwest walls of Room 27 (PN I, 147):

An actual state plan of Room 27, measured and drawn by Papathanasopoulos (PN I, fig. 421), illustrates Blegen’s description.

What can be seen of the foundations is built of rubble with an occasional large stone inserted. The northwestern wall, 1.20 m. thick, is likewise represented

The Court 88 stretch of wall ZZ parallels the longitudinal axis of the proposed successive temples 1 and 2 and runs parallel and approximately four metres from the southwest

1.46 Aerial mosaic showing features outside the shed roof: the orientation of walls for Temple 2 and temenos wall ZZ (dotted lines) as it crosses Court 88.

72

The Architectural Study of the Palace and Ancillary Buildings the northern side of an entranceway or propylon, though there is no indication of an underpinning for a doorsill along the single wall course across Court 88. Blegen has already noted the possibility of an entrance here (PN I, 291).

of the reconstructed flank wall of the last temple. Wall ZZ borders but does not encroach on those walls in the Northwest Area that are associated with the post-Bronze Age occupation; it skirts and otherwise avoids all of the walls within what we assume to be a sanctuary of postBronze Age date. Wall ZZ thus has all the fundamentals of a temenos wall enclosing a sacred space, with terminating points at the western corner of Room 27 and the northwest– southeast orientated bulkhead at Room 76. Th is bulkhead responds to a return opposite, at the corner of Room 21. A mosaic of two aerial photographs (fig. 1.46) shows clearly the features that lie outside the shed roof: the orientation of the walls for Temples 1 and 2 and temenos wall ZZ as it crosses Court 88 (see also fig. 1.44). A scruff y wall, levelled to its bottom course, departs from the end of the wall ZZ bulkhead at Room 76 and fades out to the northwest as it approaches wall 78/82. This wall might be an extension of wall ZZ, but if so it abuts the flank of the presumed Temple 2; it is assumed here not to relate to wall ZZ.

A date for the temenos wall depends on a surmise drawn from relative building phases. The oblique heading taken by wall ZZ passes hard by Circular Building 87 and the corner of the early temple, separating them from the rear wall of the Main Building, which was probably torn out by the Geometric/Archaic phase. The portion across Court 88, as noted (see fig. 1.46), runs parallel to the axis of the first temple. These alignments suggest that the temenos wall is contemporary with Temple 1, dated here to the beginning of the early Archaic period. Since a later temple is assumed here to have enveloped its predecessor, the temenos wall parallels both constructions. southwestern building and southwest quadrant

As a temenos wall, wall ZZ is a typical feature of the Greek sanctuary or sacred enclosure. The identification of this function for the wall was a realisation that came, unfortunately, after the backfi lling of the walls with protective earth. Recognition of wall ZZ as a major architectural feature of the post-Bronze Age occupation means that it deserves further investigation, consideration and discussion beyond what can be afforded here. The following remarks are therefore advanced not only as tentative, but also with the realisation that this is not the place for an extensive digression on the subject of the sacred enclosure or temenos as a typical feature of the Greek sanctuary. Tomlinson (1976, 17) supposes that the temenos was meant to set aside the religious sanctuary from the secular world; here, it takes its plan from the space available.

As explained above (p.29), fieldwork for MARWP’s study of the remains on the Englianos ridge proceeded counterclockwise. There was archaeological logic to this sequence, but it postponed to the last the southwest area (fig. 1.47): a battlefield littered with stones, where the corpses of superstructures had long ago been hauled away. This scene is particularly vivid at the southern section of the Southwest Quadrant, where ribbons of ground-hugging walls weave in and out in no clearly discernable way. Even the articulated plan of the Southwestern Building poses a challenge when it comes to an understanding of the underlying constructions, some substantial. Blegen called the area a ‘maze’ (1959, 123; 1961, 154, figs 3–7; 1965, 96 and caption to fig. 7), referring to ‘masses of wreckage and debris of buildings evidently dumped from higher ground’ (PN III, 41).

A defined boundary setting off a sacred area has a precedence in the Bronze Age in the form of a sacred grove (see Birge in Miller, Kraynak and Birge 1992). By the early Archaic period, a physically bounded area separated the sacred space from the secular (Bergquist 1967), but this demarcation could be a cliff edge, a hedgerow or a built wall. Seldom was the built wall conceived as a fortification. These walls are usually, but not always, built to human height or not much higher; a determined intruder in the Archaic period could easily scale a temenos wall, as would have been the case at Pylos. The shape of a Greek sanctuary depended on available land and the local topography; rectangles and grid arrangements come with Hellenistic design. The zigzag or the indented trace in the diagonal stretch from Room 27 to Room 21 has no obvious explanation, but it does have comparisons elsewhere, for instance at the sanctuary of Apollo at Delphi.

The state of the ruins made a full autopsy of the area a challenge that time did not permit us to meet, though we continued diligently to record, stone by stone. Our draftspeople came to understand their subjects as they drew. One only has to compare Blegen’s Key Plan (PN I, figs 416–17) or even his detailed plan (PN III, fig. 306) with the MARWP state plans to see how much has been overlooked in this important quarter of the hill. Our state plans, here as well as elsewhere across the site, are documents for future analysis. For detailed discussion of the walls of the southwest area, see Michael Nelson’s study in Part II of this publication. A brief analysis of the Southwestern Building and its agglomeration of attached chambers, staircases and the like is offered here. Also remarked on is the quadrangle of nested, multi-purpose walls to the southeast of rooms 64 and 65, some of which were removed in the Bronze Age clearing exercise that opened up a central 5 x 5-m area, stripped in order to accommodate a square structure

A Papathanasopoulos plan of the Northwest Area shows a platform of small fieldstones defining a rectangle approximately 4 x 2 m at the return of wall ZZ just to the northwest of Room 21. This could be the remains of

73

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 (fig. 1.48). Finally, the muddle of walls that covers an area of roughly 10 x 15 m to the west and southwest of Building 60/62, and running down the slope, contains several curious features: a bifurcation in a water main where a lock diverts water from one duct to another; a beautiful plaster ramp at the very bottom of the slope; lines of ashlars peeking out from the rubble; a sequence of modifications that appear to show attentive improvements over time before going out of use, leading to a question of relocation. Exposure of the Southwestern Building in 1996–97, especially in the area of rooms 63–65, allows for a reassessment of building periods beyond that of the primary Late Helladic IIIB constructions. Some walls, but not all, for rooms 64–81 diverge by 3.5 degrees from the overall axes of the Main Building and the Southwestern Building; the implications of this are discussed further in chapter 2. Most conspicuous are the flank walls to Court 88, but this variant alignment also occurs, for instance, in Stairwell 69, an obvious addition built against Room 65. Other examples of later modifications can be seen in the northwest flank of Room 65, which was reinforced by a secondary construction abutting the earlier one on the outside. The northeast end of this reinforcing wall aligns with the cross-axis of the front pair of columns inside the hall. On the opposite side of Room 65 there appears a series of walls that run perpendicular to the flank wall and towards the southeast; all of these postdate Room 65. The northwest ends of one set form a quadrangle outside the flank wall of Room 65. The room has no obvious entrance, though it does boast a 2-m wide terrace with a view that runs along the southwest flank. Subsequently, an agglomeration of walls created additional cells around and to the southeast of this quadrangle. The last built wall in this area, SW6, supersedes the earlier constructions, including Room 65, which had been levelled to the present elevation: wall SW6 is built over the top of the northeast end of the southeast wall of Room 65. This makes the wall post-palatial, probably a post-Bronze Age construction

1.48 Open space southeast of rooms 64 and 65, showing truncated walls and the square structure at centre right.

comparable in date to a number of other features found at the site during these investigations. More enigmatic is the rectangular construction (SW1) in this area, in size 2.0 x 1.5 m (fig. 1.48). It consists of ashlar blocks of the type noticed in numerous walls at Pylos dating from pre-palatial to post-palatial times. Presently this base stands to a height of c.0.40 m; it appears to have collapsed somewhat from a more tightly built construction. As is evident from the plan, this installation involved the tearing out of all earlier walls in the vicinity (e.g. SW3, SW4, SW19, SW20) to form an oval clearing c.5–6 m in size. This freestanding, quadrangular, low and loosely built construction has no obvious structural purpose, and it is placed centrally within an area stripped of preexisting pre-palatial walls. It has the earmarks of an altar situated within a cult area; if so, it has a close parallel with an altar at Kommos (Shaw and Shaw 2006, 20–24, 35–36, 57–59, pls 144–49). The Southwest Quadrant also comprises rooms 60, 62, Building X, the Main Drain and miscellaneous walls left nameless by Blegen. These walls proved to be no less

1.47 The Southwestern Building and Southwest Quadrant.

74

The Architectural Study of the Palace and Ancillary Buildings complex than the structures investigated in the Northwest Area and in areas 103 and 106 of the Northeast Area; Blegen’s plans for the most part do not record the assortment of walls representing diverse phases of construction. A monumental entranceway, a major feature of the Bronze Age site, emerged from a decipherment and interpretation of the hodge-podge of cobble and fieldstones that fi ll the south quadrant, beyond the limits of the Southwestern Building (see Part II, pp.357–59, fig. 4.5). This Southwest Gateway lies on axis with the Northeast Gateway but on the opposite side of the Englianos ridge. In its first phase, the approach consisted of a set of flagstone steps laid in convex arcs with low risers (0.10 m high), and broad treads (0.50 m wide); the details are not readily discernable on the plan. The lowermost portion of this staircase probably extended some eight metres to the south, and the uppermost six steps were preserved in the rebuilding of phase of the structure. The purpose of this second phase was to contract the sprawling staircase by covering the lowermost five metres with a plaster ramp and the middle portion with four steps made of flagstones that rose at a steeper (risers at 0.20 m) and a more even gradient than those of Phase 1 (figs 1.49–1.52). This second-phase gateway was bounded by two wings formed by the facades of Building A and Building X; both serve as retaining walls and vertical faces to the monumental facade. The left or northwest wing returns to the outside to form a two-sided courtyard for the outer flight of steps. The right-hand wing returns to the inside. For some reason the Southwest Gateway was abandoned, perhaps when Room 60 was constructed at the crest of the slope (PN I, 240). It was replaced by a new monumental staircase – a bold, heft y entrance built against the southeast end of Building X (fig. 1.53). A mixture of large stone slabs and reused ashlars mixed in with boulders give this construction an ungainly appearance, not helped by the fact that sections have been robbed away. An architectural veneer of some sort probably dressed the otherwise roughlooking gateway. Basic features speak for a staircase: four risers, each c.0.20 m high, with 1.5-m deep landings over twin parastades,

1.49–1.52 Reconstruction of the second-phase rebuilding of the Southwest Gateway staircase; photographs from top to bottom show the remains of the structure from the south, at the level of the bottom step; from the top of the staircase looking down onto the plaster ramp; and from the south at a higher level.

1.53 LH IIIB stairway at the south side of Building X.

75

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 following the destruction of the Late Helladic IIIB palace, we decided in 1992 to clear for study and drawing rooms 89 and 90 (fig. 1.54), an area where Blegen did record the presence of iron artefacts, Geometric pottery and scrappy walls built of reused material (PN I, 294–98).

project outwards (to the south) from the second bottom step; at the top is a 6-m long ramp with a 9 per cent grade. Including the parastades, the staircase is 3.7 m wide; the width of the steps in between is 2 m. The preserved east portion of the ramp diverges from the line of the east wall of Building X, which established the basic orientation. The orientation of the ramp aims more towards the south corner of the Main Building, as did the earlier Southwest Gateway. Although the upper terminus of the ramp has been destroyed, another 2–3 metres would bring it to the crest of the acropolis at the 192 m contour line, where the climb to the palace entrance levels out.

Blegen argued that these rooms served as an olive press, and interpreted as a ‘firebox’ what must have been a hearth; he did not reconcile this object with the olive press function. In all probability these rooms comprise a modest house that was expanded and remodelled in the Geometric period. Room 89, a 5 m-square house, appears to have been the first construction: blocks taken from the palace were reused in the northwest, northeast and southeast walls, with a stretch of the northeast wall of Court 64 used for the fourth side of the room. The entrance was probably in the northern corner. Later, Room 90 doubled the size of the house by extending the northwest and southeast walls to the flank wall of the palace, outside rooms 15 and 17. In a remodelling of the combined house 89/90, the door that originally opened into Room 90 from Room 89 was blocked, the dividing wall demolished, a flagstone floor of rough, flat stones laid in Room 90, and a new door opened at the southeast corner, as indicated by a threshold block. For further details of this room, see p.229 below.

This later staircase is to be associated with the Late Helladic IIIB palace, whereas its elegant predecessor served the Early Mycenean building. The first-phase stairway is securely dated to Late Helladic I, based on material found in a trench sunk by John Pedley just behind a footer for the shallow-stepped staircase and one metre west of the southwest corner of Building X. Blegen (1965, 96) reports ‘a thick deposit of earth containing early Mycenaean pottery of Late Helladic I with a slight admixture of Middle Helladic wares’. Both phases of the Southwest Gateway and the Late Helladic IIIB staircase adhere to Building X (at opposite ends), and this structure probably served a common purpose to the three building programmes, perhaps as some sort of reception hall. Emissaries and visitors from the sea may have approached the palace from this direction.

the chasm The exterior walls of rooms 7 and 8 of the main building (the so-called Archive) had been robbed away when the area was excavated by Blegen in 1938 and 1952–54. In 1958 the Greek Archaeological Service rebuilt these walls according to a theoretical scheme based, in part, on Blegen’s identification of an adjacent raised dais as a sentry stand, and on the assumption that the cavity

rooms 89 and 90 As a result of the scholarly dissent surrounding Blegen’s interpretation of the acropolis as having been abandoned and, with the exception of a few rooms, not reoccupied

1.54 Rooms 89 and 90 looking west.

76

The Architectural Study of the Palace and Ancillary Buildings dividing the Propylon into rooms 1 and 2 with the flank wall of Archive rooms 7 and 8 (fig. 1.56). In the northeast elevation (fig. 1.57) can be seen a footer of brecciated aggregate, which is laid into a foundation trench cut approximately 0.20 m into a deposit of fi ll accumulated above an earlier floor level (floor a). A thinner layer (c.0.10 m) of the same aggregate spreads to either side of the wall between rooms 1 and 2 and serves as a base for the LH IIIB plaster floor in Room 2 and a (currently) clay floor in Room 1. The chasm has an uneven perimeter and a breadth that varies between 0.95 and 2.40 m. The cut starts at the northwest, where an elbow passes through the original fabric of the northeast flank wall of Room 9. This passageway was later blocked with a rebuilding of this wall (PN I, 102, fig. 81). At the outside corner of Archive Room 8, the chasm extends towards the front of the main building and steps downwards for a depth of about a metre. Details of construction for this set of steps are obscured by the Room 1

Room 2

Room 7

Room 8

1.55 The excavated cavity of Blegen’s ‘chasm’, looking northwest. reconstructed wall (1L)

followed the exterior perimeter of the original palace walls at the southeast and southwest. He named the presumed robbers’ trench beneath the northeast wall the ‘chasm’ because of its great depth and as a way to locate a number of Linear B tablets; a coin of Venetian date was also found here.9 Prior to the rebuilding of these walls, George Papathanasopoulos made a plan of the cavity, which was never published; some features in his plan do not appear in the Blegen photographs. The position of the actual entrance into the Archive rooms has become a matter of some concern (e.g. Palaima and Wright 1985, 251–62). In 1995 MARWP received permission to disassemble the 1958 northeast wall in order to clean the floor in this area and draw a physical state plan. Our re-excavation of the chasm indicates that rather than a robbers’ trench, it was some sort of subterranean construction of post-palatial date.

5 m.

0

NE NW

Main Building, Chasm

SE SW

wall between rooms 7 and 8

reconstructed southeast facade of room 7

shed pier

floor a floor b

concrete pad floor c

cross-wall B

1m

cross-wall A

Chasm: southwest scarp

Chasm: northeast scarp

0 0

1m

wall between rooms 2 and 1

The cavity (fig. 1.55) is much deeper (c.1 m) than would have been necessary to remove the pre-existing wall and its footer. In fact, the chasm penetrates a configuration of earlier walls that appear in cross-section within the trench scarps. The structural design of walls in this section of the Main Building can be extrapolated from the cross-section left when the chasm cut away the juncture of the wall

floor a

floor c

cross-wall B cross-wall A

1.56–1.57. State plan of the chasm, and elevation drawings of each scarp.

9

Blegen makes only oblique references to the chasm in PN I, 97 and 99–100; see also Blegen 1953, 62–63.

77

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.59 View of the northeast scarp of the chasm, showing part of an angular stone below the LH IIIB plaster floor.

part of the design in this area. If so, this stone could be telltale evidence for a base construction underneath a threshold block in a doorway leading from Room 2 into Room 8. The stone, located in the northeast scarp opposite the Dexeion pier, places the hypothetical entrance to the northwest of a central position within the northeast wall of the room. That is to say, a restored portion for the doorway comes adjacent to the edge of the bench in Room 8.10 Re-excavation of the chasm yielded other rewards, as the two long scarps of the trench revealed an assortment of walls, floors and other less decipherable features in addition to the ashlar wall underneath Room 7, which was published by Blegen (PN I, 94). Occurring at multiple levels, these events are not seen in the state plan, which shows the last event in the chronological sequence here.

1.58 Floor-level view of the chasm looking northwest.

concrete pad for the Dexeion pier supporting the modern shed roof. Nonetheless, it is clear that at least one of the steps was fashioned from the native marl. The bottom of the chasm slopes downwards from northwest to southeast and consists of a mixture of large, rounded boulders and flat field stones (brought to the acropolis from elsewhere) and fragments of reused ashlar blocks. Here and there this pavement returns upwards along the vertical walls of the chasm (fig. 1.58). None of the paving extends beyond the confines of the cut and must have been installed after the cut was made. A wall made of the same sort of stones lines a recess cut flush with the extant half of a threshold between rooms 7 and 8. This indicates that the floor and the vertical lining adjacent to the threshold are late additions, not remnants of a pre-existing but robbed original footer and wall to Archive rooms 7 and 8. Moreover, the vertical section of wall or liner at the threshold rises c.0.10 m above the level of the palace-period floors in these rooms, as well as above the level of the missing half of the threshold block, providing further evidence that the chasm is a postpalatial feature.

One of the two surviving orthostates of cross-wall A (noted by Blegen and discussed in more detail by Michael Nelson in Part II, pp.314–15), bears the well-known inscribed double axe (see Part II, figs 3.26–3.27). The wall is missing except for two blocks of orthostates and several socle blocks that extend to either side of the orthostates. A 10

During his first season with us in 1994, Emmett Bennett strolled through the Main Building and voiced a question: what if the entrance into the Archives Complex were through Room 8, inside the Main Building, rather than outside into Room 7, as Blegen argued (PN I, 93) and as is shown on his plans. Blegen wrote: One piece of evidence offers corroboration of this conjecture about the position of the doorway, namely that the stand for a sentry was placed on the left-hand side of the doorway of the Propylon instead of the right. In all the other three instances in the palace in which sentry stands were found they lay at the right. From this stand the sentry could control those entering what may have been a treasury office as well as those going straight on into the palace. Several pages later (PN I, 97), he offered Bennett’s alternate reasoning: Or alternatively, as believed by Professor Bennett, may there have been a door leading out from the Archives V Storeroom into the inner porch of the Propylon through which the clay documents could have been precipitated? It was Bennett’s suggestion to reopen the chasm. In his spirit of speculation, we wondered meanwhile if the sentry stands might instead be plinths for something like a set of horns of consecration.

The ancient excavation that produced the chasm eliminated all surface traces of the entrance into the Archive rooms. However, there is some tentative evidence for such a passageway. A portion of an angular stone, in situ and perhaps from some sort of footer construction, underlies the LH IIIB plaster floor of Room 2 (fig. 1.59). The aggregate base for the floor is built against this stone, an indication that the stone may have been an intentional

78

The Architectural Study of the Palace and Ancillary Buildings large reused ashlar was placed into a position 0.40 m to the southeast of the socle to make a sill or a step down into a continuation of the cobbled pavement of the chasm, which carries to the southeast. This ashlar block and the dressed socle of cross-wall A do not continue into the northeast scarp of the chasm; instead, the socle here changes to large angular fieldstones, indicating that the dressed face of wall A either returned or ended at this juncture. A second cross-wall (not noted by Blegen) runs approximately parallel to wall A and 1.20 m to its northwest. It consists of large rubble stones, some of which can be seen in both scarps. The lower courses of both walls A and B were reused for portions of the pavement in the chasm.

Phase

Cut through by the sinking of the chasm are three prepalatial floors (labelled a–c in fig. 1.57), as well as the later LH IIIB floors of rooms 7, 8, 1 and 2. Floor a is a thin layer of plaster c.0.03 m thick on a 0.05–0.08-m base of aggregated earth. Floor b, made of small stone chips and pebbles, begins at the top and back of the orthostate courses and extends to cross-wall B, where it stops. At this position in plan and elevation, floor b, which paved an interior room, is about 0.60 m above the ground level at the base of the socle for wall A. Floor c is a thin plaster stratum that can be seen in section within the scarps to the southeast of cross-wall A; its elevation corresponds to a height midway up the orthostate course of cross-wall A.

Period

Event

1

pre-palatial

cross-wall A

2

pre-palatial

cross-wall B and f loors b and c

3

pre-palatial

f loor a

4

LH IIIB

Main Building rooms 1, 2, 7 and 8

5

post-palatial

Chasm with its paving

6

post-palatial

blocking of entranceway to chasm through northeast wall of Room 9

Table 1.3 Proposed building phases in the area of rooms 1, 2, 7 and 8.

An opening cut through the original northeast wall of Room 9 served as the starting point for the passageway of the chasm. This opening was later blocked (Phase 6). This relative chronology is summarised in Table 1.3. As mentioned above, the chasm is an intentionally built construction post-dating the Late Helladic IIIB palace; a curved but crude flight of steps downwards from Room 9 provides access to the depth of the chamber with its floor of flat and rough stones. There is no sign that walls lined the top rim of the chasm other than the small portion at the line of the rooms 7/8 entrance cross-wall. The narrow outline of this sunken chamber may have persisted over time as erosion deposit fi lled the cavity (PN I, 99–100). This earth deposit contained Linear B tablets and a mixture of other finds from Middle Helladic to Venetian periods, including five or six votive kylikes.

The following recapitulation is a tentative reconstruction of the building phases here. Building began with the construction of orthostate cross-wall A (Phase 1).11 A hypothetical exterior (southeast) floor level presumably reached the socle to this wall (elevation 191.15 m), with an interior floor level slightly above its backing stones (191.50 m); no evidence for these floors survives. Next (Phase 2), cross-wall B was built a metre to the northwest of crosswall A and parallel to it. The interior floor level between cross-walls A and B (floor b) was laid at elevation 191.85 m (0.35 m above the earlier hypothetical floor here) and made even with the top of the orthostate course of cross-wall A; a packing of stones reinforced wall A on its northwest side, thickening the wall beneath floor b from c.0.90 m to c.1.70 m. Floor c (southeast of Room 7) to the exterior was raised above the previous (hypothetical) floor by the same 0.35 m to elevation 191.45 m. Floor a appeared next (Phase 3); its associations are unknown. During the period of the LH IIIB palace (Phase 4), wall B was demolished and replaced by another wall to the interior or northwest, at the partition or doorway between rooms 7 and 8. The Archive Room (Room 7) floor lies at 0.32 m above the earlier floor b. The floor of the Main Building (also Phase 4) lies at an elevation of 192.17 m. A lining wall of the chasm rises above this to an elevation of 192.25 m, as seen in the southwest scarp. This fact and other evidence make the chasm (Phase 5) later in date than the Main Building.

The date and function of the sunken chamber remain questions that deserve consideration beyond a few thoughts given here. While it is true that the widespread Dark Age stratum of a black, oily soil above a yellowish or light-brown layer was not recorded for this particular location, it should also be noted that soil composition was not as assiduously noted in the early years of excavation as in later years. As for function, one possibility is a subterranean but uncovered storage magazine. Alternatively, the chasm construction may perhaps be associated with evidence for post-Bronze Age cult activity across the site. In a detailed description of Room 7, Blegen mentions, besides Linear B tablets, the presence of a huge corded pithos (Rutter 2005, 36, 37, 57, 59) and a considerable heap of burned animal bones (PN I, 93). These were found in the western corner of the room; close beside them, near the northwest wall, were discovered 11 diminutive kylikes, probably votive offerings. Blegen was baffled: ‘What these apparent remains of sacrifices and dedication vessels had to do in the tax collector’s office raises an unsolved problem’. Although Blegen was somewhat flummoxed by the presence of ‘votives or playthings for children’, he does mention that ‘votives of the same kind have been found by Professor Marinatos in chamber tombs at Volymidia’

11

Blegen remarks (PN I, 94) that ‘no decisive evidence has been found to fi x the date of the earlier wall’. Nelson (Part II, p.350) places this forerunner in the Late Helladic IIIA period.

79

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 rooms 54–57

(PN III, 366). In the intervening years the post-Bronze Age ‘hero cult’ at Volymidia has been widely acknowledged.12

The simplified palace plan published by Blegen (PN I, fig. 424) obscures the complexity of walls within the confines of these rooms at the northeast corner of the main palace building. A clearing of the protective fi ll in 1998 enabled an analysis of the multiplicity of walls crowded together and overlying one another in this area (fig. 1.60), and led to the conclusion that four stages of construction preceded the erection of the walls for the fift h phase, that of the construction of the main Late Helladic IIIB building.

Stocker and Davis (2004) have reviewed comprehensively the evidence for ritual dining at Englianos, including deposits of ash and bone and the enigmatic votive cups. I agree with all the fundamentals of their study, with the exception of date: the activity here is evidence for Archaic-period cult practice, not Bronze Age. In view of the preponderance of evidence for Dark Age, Geometric and Archaic religious activity at the palace (see chapter 8 below; see also Griebel and Nelson 1993 and Popham 1991), I suggest that the presence of bones and the kylix simulacra in Room 7 should be taken as evidence for later cult practice in this area of the post-Bronze Age palace, paralleling evidence found elsewhere on the site, especially in the Northwest Area (see chapter 8 below). The diminutive vessels at Pylos also have parallels at other Archaic and Classical sites in Messenia, among them the sanctuary at Pamisos (Valmin 1938, 454–62, figs 91–96, pls XXXVI–XXXVII), at Messene (P. Themeles, personal communication) and at the sanctuary of Poseidon at Akrovitika (M. Kiderlen, personal communication).

The floors in this suite of chambers lie approximately 0.50–0.90 m below the floor level of the Main Building. MARWP excavations revealed a kind of window through the last construction and onto four earlier buildings or phases of construction. Most important is the realisation that the line of large, squared poros blocks at the southeast of Room 57 (fig. 1.60, wall y, and fig. 1.61) belongs to a monumental earlier building. An elevation of 191.55 m matches the top of the equivalent socle underneath the orthostate course in Archive Room 7, which includes the block with the double axe sign (see above, pp.78–79). An earlier floor (floor b) found underneath that of the present Room 7 (elevation 191.85 m) lies level with the top of this orthostate course.

courts 42 and 47 A cut channel passes underneath the partition wall between Room 43 (where the clay bathtub stands) and leads to a hole within Court 42 (‘a’ in PN I, fig. 426). The floor surrounding the hole slopes downwards to the opening, giving it the appearance of a drainage area. In Court 47 there are 14 sinkholes, half of which appear to be randomly placed. Two (at ‘b’ and ‘c’ in PN I, fig. 426) align with the hole in Court 42. Blegen’s plan connects two sinkholes with the drain in Court 42, and on his plan there is a hatched line connecting the three holes and leading to an elbow outside of the palace facade and at the beginning of the Main Drain.

Returning to Room 57, the socle succeeds an earlier rubble wall two metres to the northwest, of which two segments survive (Phase 1; fig. 1.60, wall m beneath the later wall c).

Clearing of the earth in these rooms in 1998 revealed that the holes in the floors were uniformly cusped- or flanked-shaped, flaring outwards from top to bottom. All are lined with a packing of pebbles, including three openings that Blegen connected to make a drain; there is no sign of a channel passing through these holes, however. The openings have a roughly circular shape, except for one that is triangular, with diameters of 0.12–0.17 m; the holes are 0.15–0.42 m deep. In the excavation notebooks for courts 42 and 47 Marian Rawson recorded a quantity of unpublished pottery that she labelled ‘flower pot ware’: bright red, of coarse fabric, and straight-sided. The possibility that these holes once did contain flowerpots and that these courtyards functioned as enclosed gardens is explored by Anne Hollond in chapter 4 below.

12

Stocker and Davis (2004) quote Blegen’s observation as a point of departure for their study. See also Isaakidou et al. 2002, 194.

1.60 State plan of rooms 54–57.

80

The Architectural Study of the Palace and Ancillary Buildings A packing of stones and earth fi lled the space between the newer orthostate course and the earlier rubble wall. The top of the early wall m lies at a level coincident with the floor level of Phase 2, characterised by the orthostate wall. Although slightly skewed to the line of the later wall, this earlier fragment was reused and expanded to create a sill for an entrance into a set of magazines or galleries. The narrow corridors contained open pithoi sunk into the floor (fig. 1.60, P1, P2, P3 etc.). In a third phase, wall c was built behind the socle/orthostate courses; it was later severed at either end, perhaps during Phase 5 construction. Phase 4 saw major modifications in this area (fig. 1.62). A staircase (fig. 1.63) rose from the area of the later Room 54 and wrapped around the four sides of an atrium-like chamber that was remodelled in Phase 5 to make the present Room 56 in the Late Helladic IIIB palace. The Phase 4 construction consists of double casement walls in a rectangle (fig. 1.62, walls f, q and g). The double curtain walls supported a staircase at the northeast and southwest flanks and probably landings at either end, one halfway up the staircase at the southeast and the other at the upper floor. The segment of surviving staircase at Room 54 was reused in the next phase to raise the floor level to 192.90 m, or 0.45 m above the still extant pilings for the earlier staircase. This segment shows that the original staircase rose in modules of three steps followed by a short landing of 0.65 m.13

1.62 Later building phases 4 and 5 in rooms 54–57.

The double-casement construction is an unusual feature; Blegen understood it as double casement for a tower integral to the Late Helladic IIIB palace (here, Phase 5). To 13

Blegen (PN I, 221, fig. 468) proposed that a single fl ight of steps returned into the corridor, the space above corridors 45–52. Nelson (Part II, p.356) agrees with Blegen’s reconstruction for a later phase stair here.

1.61 Wall of squared poros blocks (wall y) at the southeast end of Room 57, looking southwest.

1.63 Staircase in Room 54, rising into Room 55.

81

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 say the least, a double-shell tower with 0.90-m thick walls separated by 1.20 m of airspace represents an exceptional design for any period in the history of fortification architecture. A more probable reconstruction is that of a rectangular staircase that winds around four sides of a lightwell, rising from the corner of Court 44 to a suite of rooms in the upper floor of the earlier palace Palace B (see chapter 2 below).

for processing. In addition, one out of every ten bags was chosen by computer-generated random number; each selected bag was fully classified by pottery type and body part or by other classification according to an established system. This procedure not only allowed for statistical analyses of finds from the site but also provided a control over the conveyor belt method of inspection and retrieval.

Finds

An estimated 2,250,000 artefacts were contained in the 365 sacks of material removed in 1994, with each sack averaging 42 kg in weight. An estimated 10 percent of the total was retained for inventory and museum storage. An analysis of the contents of these 365 sacks was carried out in 1995 by Joanne Murphy and Jan Verstracte of the University of Cincinnati, with the assistance of Emmett Bennett and Kalliope Kaloyerakou; their results are summarised in Table 1.4. For the pottery, the figures indicate the number of sacks in which pottery of a given period was found, without any attempt to quantify the total number of individual pieces. This synthesis nevertheless produces a striking picture of vast quantities of Middle Helladic pottery on the site, and lesser but still substantial amounts of post-Bronze Age material; of particular interest is the quantity of possible Archaic pottery from the 7th century bc, the date assigned to the first temple construction in the Northwest Area (pp.67–68 above).

As explained above (pp.29–31), a decision was made at the outset of our excavations to sift all removed dirt. We expected to recover a modicum of overlooked finds in the Blegen backfi ll, but it became clear at the end of the first season of excavations in 1991 that the quantity and importance of finds would far exceed our expectations. In 1994 we discovered that Blegen used the three metredeep cavities within the grillage of walls at the Northwest Area as a dump for washed but discarded finds from his years of excavation. 365 olive sacks of material were removed in that year and a further 50 in 1995, totaling some 20 tons of pottery, along with roof tiles, stone tools and weapons, painted fresco fragments, Linear B tablet fragments, terracotta figurines and other artefacts. Although this material lacks all context, it does provide an index of the range of typical artefacts and their relative quantities and gives an idea of the chronological span of activity on the site. A study of the small finds from the excavations is published as chapter 7 in this volume; for lithics, see chapter 6.

A circular bin built of fieldstones in dry wall construction was erected about 20 metres south of the south corner of Room 7 to serve as a storage depot for re-discarded material from the Blegen dump. The bin was designed to be unobtrusive and barely visible, blending in with the stonework of the palace and located below the level of the main buildings. The construction is solid but easily and fully reversible.

A hand-operated conveyor belt was used to inspect in passing every artefact recovered from the dump. Every object of interest was pulled from the conveyor Material Industrial wasters Wasters Rooft iles Plaster Fresco Pottery

Period

Individual Objects

Certain Sacks

Possible Sacks

Out of 365 Sacks

Percentage

192 2 Byzantine

Neolithic EH MH LH I–II Post-Bronze Age Early Iron Age/Dark Age/Geometric Classical/Hellenistic / Roman 7th century bc 8th century bc

38 682 2174 3712

8

1 23 57 58 137 84

1 23 354 91 137 92

less than 1% 0–6% 82–97% 9–25% 0–38% 2–25%

46

142

188

13–52%

111 1

111 1

0–30% less than 1%

3

3

0–1%

297 33

Byzantine

Table 1.4 Analysis of the contents of 365 olive sacks of Blegen discard, collected in 1994.

82

Finds pottery The recovered pottery shows activity on the acropolis beginning in the Late/Final Neolithic and continuing through Late Helladic IIIB and into the Dark Age. Several Bronze Age ceramic specialists, including Jeremy Rutter, Cynthia Shelmerdine and Yannis Lolos, greatly assisted us in the identification of pottery. Andre Beckermann, a graduate student at the University of Toronto at the time of his premature death, was preparing a study of the incised ware for the site; this now awaits further work. There is, in addition, representative pottery covering the chronological spectrum from Geometric, Archaic, Classical, Hellenistic and Roman periods; there was also a Medieval phase on the acropolis. Shawn Ross examines a selection of the post-Bronze Age pottery in chapter 9 below.

a

b

1.64 Linear B tablet fragments recovered in 1991; Xn 1481 at left.

linear b tablets In 1991 and 1992 three certain and several possible fragments of Linear B tablets were recovered from Blegen’s backfi ll. Of two certain fragments found in 1991 (fig. 1.64a and b), the larger is Xn 1481. It has no obvious joins to known pieces and seems to have contained a list of names, starting somewhere above the first extant name of qo-wiro. Shelmerdine and Bennet (1995) publish specifics.

a

b c

Xn 1481 .0 vestiga .1 qo wi ro [ .2 o mặ infra mutila

1.65 Probable Linear B tablet fragments recovered in 1992 (a) and 1993 (b and c).

A fragment, 0.02 x 0.015 m in size, was recovered in 1992; it carried a single character (fig. 1.65a). From the same area, around Room 100, came five further small pieces of fired clay in 1993. These were identified as probable Linear B tablets; two bore incisions (figs 1.65b and c), the remainder had no writing. Two tablet fragments (figs 1.66a and b) were recovered from the Blegen dump. The following transcriptions are provided by J. L. Melena (2000–01): 1995-2 (a) Un 1177 + fr. supra mutila .1 OVIS 6 [ .2 BOS 1 [ .3 ME v 4 [[ ]] AREPA v 1 [ .4 ] 1 tu-ro2 [ . .. .5 deest .l OVIS over [[ ]]

a

1995-3 (b) Un 1482 .1 deest .2 ke-ra-e-we *189[ .3 ka-tu-re-wi-ja 4 [[ ]] ta-ra-[ .4 de-mi-ni-ja , a-ke-re-wi-ja [ infra mutila

b 1.66 Linear B tablet fragments recovered from the Blegen dump.

83

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Professor Bennett, using archive photographs, joined 1995-2 with the extant tablet UN 1777; the first two lines of the transcription are on the original find. The text of Un 1482 led Melena to provide an extensive commentary, particularly on the meaning of ke-ra-e-we *189 in line 2 as meaning either ‘horn’ or ‘gift of honor’ (ibid., 382).

an artistic programme but probably varied from project to project. Thus an analysis of the fragment cross-sections is of crucial assistance in assigning pieces to individual scenes on walls (Pappalalardo 1998–2000; Meggiolargo, Molin, Pappalardo and Vergerio 1997). Interestingly, fresco fragments were found not only within the backfi ll of the main palace building (whence Mabel Lang recovered almost all of the fragments that she catalogues

frescos Some 7000 pieces of fresco and painted stucco were recovered in the MARWP excavations, of which 3535 were inventoried and kept (see above, pp.29–31). Many had distinguishing features of colour or pattern, though all are quite small, averaging 0.03 x 0.03 m. The most important fragment recovered forms part of the famous Griffin and Lion fresco from the Queen’s Hall, Room 46 (PN II, 111–14, cat. no. 21 C 46, pls 54–57, F and P). This fragment is now on display in the Chora Museum with other pieces of the scene. Feathered lines on the newly found fragment suggest a position along one of the animals’ tails. Another fragment (fig. 1.67) may be from a nautilus freize (PN II, 147–50, pls 82–83, R). A conservation laboratory was established in 1994, devoted to the cataloguing and conservation of painted fresco fragments, pottery, lithics and other small finds.14 An inventory procedure was devised to allow off-site analysis of the fresco fragments and to facilitate computer digitising of fragments in order to locate joins and correlate similar pieces. Each fragment was recorded in detail: size, shape, colour according to the Pantone colour scale, pattern and other features. Cross-sections were meticulously measured, and the plaster composition and thickness of different layers of stroses recorded. This last is an important diagnostic feature because the painted walls of the palace were prepared by applying from two to four preliminary layers of plaster. Thicknesses of successive layers and the precise composition of the plaster may be consistent within 14

Th is was carried out under the direction of S. Bouzaki, conservator of the Corinth Excavations of the American School of Classical Studies at Athens. We thank C. K. Williams, then Director of the excavations, for making this possible.

1.67 Fresco fragment showing a nautilus tentacle and part of a second (left edge).

1.68–1.69 Fresco fragments compacted to form a f loor in Area 106 (top); detail of fragments on the f loor surface (bottom).

84

Finds In total, 10 sacks of this material are preserved. The theory that these are the residue of moulds used in a bronzecasting operation is explored in chapter 10 below.

in The Palace of Nestor, vol. II), but from throughout the surrounding rooms and ancillary buildings. Blegen characterised these structures as workshops, armories and the like, but in all likelihood they were also decorated with painted walls. Although frescos occurred across the site, most of our finds came from two contiguous locii in a single MARWP unit within the Northwest Area cavities. In addition, many fragments remain in situ in the pre-palatial compacted floor located in Area 106 to the north of the Main Building (figs 1.68–1.69; see also p.40 above).

roof tiles In The Palace of Nestor, Blegen makes reference only to ‘medieval’ roof tile finds, probably those of the Frankish period in the area of the Northwest Gateway (PN III, 6). The excavation notebooks from numerous trenches throughout the site mention recovery of tiles, particularly in some quantity in the lots of trenches MY6, MY9 (northeast of the Wine Magazine) and in the Throne Room (Room 6). None were inventoried or published. The Blegen dump yielded a total of 470 kg of roof tiles, representing at least five roofs, including an early Archaic Corinthian hipped roof, an Archaic Lakonian gabled roof, another Corinthian roof and at least two Frankish roofs.

In the belief that a widely accessible (i.e. digital) method of documenting, querying and examining the fragments of the Pylos wall paintings would be of scholarly interest, we extracted details as mentioned above from the inventory sheets; taken together, these geometric and attribute data were perfectly suited to a GIS application. Scaled drawings were digitised using AutoCAD and exported to Arc/Info, where the geometry of each piece was tied to its related attributes and topology. As originally conceived and developed within Arc/Info, this GIS application enabled MARWP researchers quickly to query fragments based on attributes of interest, and to display fragments with comparable dimensions for comparison on-screen using uniform display scales.

Representative fragments of the earliest roof include Corinthian pan and cover tiles and an example of a cover tile made to fold over the crease of a hipped roof (see p.247, figs 8.17–8.18); this feature associates this roof with the early Archaic temples of Apollo and Poseidon at Corinth and Isthmia (Rhodes forthcoming). The early Archaic Lakonian roof preserves several fragments of decorative revetment: an akroterion disk (or disks; see p.248, figs 8.22–8.23) and a portion of a painted lateral antefi x (ibid., figs 8.20–8.21), which date the roof to the Lakonian IV pottery period, or the second quarter of the 6th century bc. Pan tiles from a second Corinthian roof bear impressed finger swirls and crosses (pp.249–51, figs 8.25–8.26).

Our basic understanding of figural painting at Pylos was established by Mabel Lang (PN II), who with great acumen synthesised disparate fragments of painted plaster into coherent scenes from specific walls. Later scholars have adapted her groundwork for defined studies on different topics (M. Shaw 1993) or to address the extent of Aegean wall painting (Immerwahr 1989). Lucinda McCullum (1987) plucked from the saved excavated lots held in the museum at Chora painted fragments pertinent to her iconographical study of the megaron frescoes, and a team from the University of Cincinnati has undertaken an archaeometric analysis of the stored pieces. The systematic correlation of painted fragments from the Blegen excavations, as well as those recovered by MARWP, will take prolonged effort, even with the aid of new and future technologies, but the Pylian material is surely worth the patient study.

Frankish tiles can easily be distinguished from anything produced in ancient or modern Greece by fabric alone. These handmade tiles are formed from a clay containing large, coarse inclusions and grit, and they were fired unevenly. Exposed sides have rough surfaces finished with little smoothing; indifferently impressed loops are another characteristic feature. The undersides are especially distinctive, being underfired and crusty. This crumbly underside surface was probably caused by a bedding of straw or daub that separated individual tiles in the kiln; a few examples retain impressions and even bits of straw. Another peculiar design feature is the sharply trapezoidal shape of the pan and cover tiles, diminishing in width from top to bottom (see pp.251–53, figs 8.28–8.33). The span of the cover tile equals that of the pan tile; the result is a wide gap between flanking pans, rather than the usual butting. The typical vertical profi le of the cover tile consists of a high-flung segmented arc that converges at its peak to make a pointed arch. This ogive profi le is segmented further in the cover tiles of one of the Frankish roofs; this distinctive type of pan and cover tile is found in the Frankish period excavations at Corinth (c.1220–1300) and in examples noted during the MARWP survey of the Frankish Morea (F. A. Cooper 2002, 23–25). The tiles of all periods are further analysed in chapter 8 below.

terracotta industrial waste A quantity of fired clay pieces was recovered from the Blegen dump as well as from backfi ll sifted from other areas of the acropolis. These pieces are quite porous, light and friable to the touch, especially on the exterior surface, and they differ radically from pottery: the high porosity yields a specific gravity of 0.8, or approximately one-third that typical of terracotta. In other words, this material floats. The pieces are thin, mostly 0.015–0.025 m, and have no flat surfaces: nearly all curve in concave/convex planes. The interior surface often bears a smooth surface, many with a cream-coloured slip. There are numerous small fragments, but some pieces are as large as 0.20 x 0.30 m.

85

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Scientific Analyses

Dendrochronology and Wood Identification

Characteristic vision and foresight led Carl Blegen to collect specimens of a variety of materials during his excavations at Pylos between 1952 and 1963. These fall into six categories: earth (64 samples), metal and miscellaneous (11 samples), shell (20 samples), bone and teeth (50 samples), and wood and charcoal (71 samples). In 1967 Blegen gave this suite of samples to George Rapp, confident that the emerging scholar of geology and archaeology would study them properly and fully from a scientific point of view. Shortly thereafter Rapp instituted at the Duluth campus of the University of Minnesota the Archaeometry Laboratory, which eventually developed into one of the major archaeometry laboratories in the world. The samples were preserved in paper boxes or envelopes with an inserted label, usually a piece of cardboard torn from Greek cigarette cartons. For the most part the penciled labels remain legible, providing trench identifications and a brief description of the stored item or items. Inscribed later in ink were catalogue (‘C’) numbers for all categories except wood and charcoal. Chunks of plaster were attached to many samples from all categories, probably used as a means of extracting the specimens intact from the soil, as was Blegen’s practice for fresco fragments. In 1977 John Gifford, then a graduate research assistant at Duluth, repackaged the samples, placing them in plastic ziplock bags and further wrapping the wood and charcoal in aluminium foil. The lot was organised in closed trays and placed in climate-controlled storage at the Archaeometry Laboratory.

Twenty-one wood samples (WC 2, 20, 22, 25, 26, 36, 45, 46, 50–53, 56, 58, 59 and 62–67) were sent to Peter Kuniholm at the Malcolm and Carol Wiener Laboratory for Aegean and Near Eastern Dendrochronology at Cornell University. Laboratory assistants identified the wood species and made a count of rings. Unfortunately, the small size of the wood fragments meant that dendrochronological dating of the samples by tree-ring counts was not possible. These samples are also discussed by Michael Nelson in relation to the use of wood in the palace; see Part II, pp.294–96. Soil Analyses Six of Blegen’s 64 ‘earth’ samples (E 3, 16, 23, 40, 55 and 59) were sent to Bill Zanner at the Soil Science Laboratory at the St. Paul campus of the University of Minnesota. The results yielded no significant information except that the ‘predominant clay mineral is montmorillonite’ (Rapp, personal communication). A sub-category of earth samples consists of soils extracted from levels that the excavators characterised as ‘oily black, mixed with small stones’, represented by sample numbers E 29, 30, 33, 36, 37, 50, 62. Blegen interpreted these as residue from an olive press of the Geometric period. Two-thirds of the ‘earth’ specimens come from pithoi or other types of storage vessels. Blegen’s expectation probably was that these soils might reveal evidence of the organic contents originally stored in these vessels and assimilated into the black earth. Four samples (E 28, 32, 35 and 36) from this lot were sent for a determination of organic material using gas chromatography/mass spectrometry to researchers at the Australian National University at Canberra. The work was carried out by Barry Fankhauser of the Division of Archaeology and Natural History, Research School of Pacific and Asian Studies, and Charles Hocart of the Plant Cell Biology Group, Research School of Biological Sciences.

In 1992 I compiled a database inventory of the samples, and Rapp inspected and evaluated each specimen. At the same time the samples were weighed, Munsell readings were taken, cross-references to excavation notebooks were recorded, and samples of special interest or value were flagged. In 1993 the Institute of Aegean Prehistory (INSTAP) provided funding for scientific analysis of a selection of samples of various types. The following discussion and Table 1.6 summarise the findings of the various scientific laboratories. We are very grateful to Professor Rapp for providing the samples and facilitating their analysis.

3-methoxy-4-hydroxybenzaldehyde (common name: vanillin) was found in appreciable quantities in all four samples, especially in sample E 28, found with pithos fragments. The most likely source for the chemical in this sample is resin used for sealing the inside of clay vessels meant to hold liquids such as olive oil or wine. Also found in all four samples were fatty acids, specifically hexadecanoic acid (16:0), which is a major component of olive oil. Sample E 36 contained the largest amount of these fatty acids relative to 16:0. Collected from area ME, or the northwest corner of Court 47, sample E 36 is of particular interest for the presence of benzaldehyde found in it exclusively; benzaldehyde was not detected in the other three samples. As the investigators noted,

ecofact and miscellaneous analyses Seppo Valppu, Assistant Scientist at the Archaeometry Laboratory, conducted a thorough inspection of 70 selected samples of ecofacts, making optical identifications of seeds, fossils, bones and teeth (sample codes S and BT in table 1.6). Rapp and Valppu also examined and wrote brief comments on the 11 samples of ‘metals and miscellaneous’ material (sample code MM), in reality a category of undiagnostic objects usually tagged with the query, ‘what is this?’.

Benzaldehyde occurs naturally in quite a number of essential oils, for example, cinnamon bark and leaf oil, cherry laurel, jasmine, narcissus, neroli, patchouly,

86

Topographical Survey and Digitised Elevation Model of the Englianos Ridge 1300-m area. Pastures, but not olive groves, were plotted, along with modern buildings, paths and roads. Using an altimeter, Fant adopted a datum based on a local sea level reading and set a benchmark of 146.00 m at the site, a value short of the geoid elevation of 192.32 m above sea level (Ellipsoid WGS 1984). He did not tie elevations into a Greek datum, nor was he able to geo-reference the plan to the available coordinate system of the Hellenic Army Geographical Service (HAGS). Despite these limitations, Fant’s map remains a valuable resource that was published in abbreviated form as a location plan for tombs found in the area (PN III, fig. 301).

cajuput, niaouli oil, acacia flower oil etc. The major sources of benzaldehyde are the kernel oils of bitter almond, peach and apricot. Benzaldehyde is widely used in flavours, perfumes, lotions, cosmetics and soaps. The presence of this chemical in soil collected from Court 47, along with the relative preponderance of fatty acids possibly from olive oil, lends weight to the proposal herein (chapter 4) that courts 42 and 47 on the northeast side of the palace may have been used as gardens, as well as reinforcing the evidence of the Linear B tablets from Pylos that suggest an active perfume oil industry at the palace (Shelmerdine 1985).

In the post-Blegen period, HAGS made available the 1:5000 sheets in sectors 7444:6 and portions of 7444:7 and 7444:8 of the Filiatra quadrant, and by the 1990s the newer, aerial photogrammetric photographs and a geodetic map at a scale of 1:50,000 for the Filiatra quadrant. At this time the Institute of Geology and Mining Engineering (IGME) also issued a geological map as an overlay on the 1:50,000 Filiatra sheet. At the end of the 1997 season, MARWP commissioned Dr Bobby Ionnides of the Athens Polytechnion to take a set of low-altitude photographs of the site by helicopter (figs 1.73–1.97). These are largescale, high-resolution, panchromatic photographs that cover both the ridge and the acropolis, though the Main Building of the palace was obscured by its shed roof. A second helicopter run took low-altitude vertical colour negatives with a mounted Haselblad; 35 mm obliques were taken on a third round.

Control soil samples were gathered systematically by the MARWP team from locations on top of the acropolis, from fi ll behind terraces, and from Pliocene depositions of marl and silt in the vicinity. Additionally, small cores were taken from clay mortars found in walls of the Main Building and ancillary buildings as they were cleaned. This sampling was part of a pilot study to determine whether the construction clay mortars might contain evidence for structural additives. This material was analysed at the Soil Testing and Research Analytical Laboratories at the University of Minnesota, using a technique called Inductively Coupled Plarma (ICP), which analyses quantities and percentages of 20 different elements, principally the heavy metals. The results indicated a high percentage of calcium carbonate in the walls, evidence that the palace builders used slaked limestone mixed with soil to make a grout, thereby dramatically increasing the strength of the mortar used in the walls. For details of testing methodology, results and conclusions, see Part II, pp.373–77, tables C.1–C.4.

The MARWP topographic survey of the Englianos ridge commenced in 1991 as a subsidiary project to the archaeological re-examination of the physical remains of the palace and acropolis. From 1992 to 1997 the survey team worked late afternoons and evenings, most often in a cooperative venture with PRAP. Integral to our work was the Trimble Pathfinder GPS, capable of an accuracy of 2–5 m, though this degenerated to ±100 m under certain conditions. Our reliable GPS waypoints had a horizontal accuracy (latitude and longitude) of 0.5–5.0 m, but the elevations (altitude) were randomly in error 2–25 m; elevations are based on the geoid datum, an equipotential surface that best fits mean sea level. In 1997 MARWP acquired a sensitive Pollard altimeter with an accuracy to 0.25 m; a digitised altimeter accurate to 3 m provided backup. In 1993 MARWP installed at our residence in Chora a Trimble base station for differential correction, which enabled the correction of distorted signals, allowing for much greater accuracy in the coordinates gathered by the rover GPS units in the field.

Topographical Survey and Digitised Elevation Model of the Englianos Ridge In addition to the re-excavation of the palace, MARWP undertook a topographical survey of the surrounding area, intended to present the various topographical and environmental relationships between the palace and the nearby tombs and other features. From 1992 it was agreed that MARWP would have responsibility for the Englianos acropolis and for survey of the surrounding ridge; survey, study and publication of the larger Pylian hinterland would be the responsibility of the Pylos Regional Archaeological Project (PRAP). Blegen devoted a minimum of resources to mapping the physical remains of the palace and its ancillary buildings; no effort at all went towards a topographic plan of the Englianos ridge. In the 1960s, Jesse Fant, a member of the Minnesota Messenia Expedition (MME) staff, prepared a contour plan from what appears to be a composite of several sheets of the British Air Force topographic maps made during World War II. Fant’s plan covers a 2000 x

Fieldwork for the topographic survey thus eventually adopted a hybrid combination of GPS with stadia readings taken by transit and rod. The approximately 700 individual stadia shots taken each season were reduced to three-point coordinates and read into AutoCAD for processing into a digital elevation topographic plan. Waypoints were converted using Trimble soft ware into

87

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Dbase III fi les, with descriptive identification appended from handheld computers; the resulting data transfer fi les were loaded into AutoCAD. Using a triangular irregular network (TIN), the waypoint coordinates were converted to a three-dimensional mesh; CivilCAD, a topographic module that runs within AutoCAD, next produced a set of topographic contours at five-metre intervals; AutoCAD then generated a finished plan. The geo-reference satellite spectral data, the high resolution panchromatic aerial photographs, and a diverse set of geodetic, geographical, statistical, historical and land-use maps has allowed not only GIS (geographical information systems) analysis and interpretation of the landscape around Pylos, but also a modelling by Todd Breningmeyer of three-dimensional views of the palace of Nestor and the Englianos ridge. Figure 1.70 shows a detailed and site-specific digital terrain or digital elevation model (DTM/DEM) of the Englianos ridge, with contours at one-metre intervals and with agricultural terraces, trees, paths and other features indicated; from this was made an exhibition model of the Englianos ridge (fig. 1.71), which provides perspectives of the acropolis with terraces and restored staircases from the north, west and other difficult viewpoints. A second DEM plan was created by digitising other details of the palace and the 20-metre contours from the HAGS Filiatra 1:50,000 map (fig. 1.72). 1.70 Raw digital terrain model (DTM) of the Englianos ridge with contours at 1 metre intervals.

The dating of terrace retaining walls continues to be a vexing problem, addressed by, among others, Wayne

1.71 Exhibition model of the Englianos ridge with aerial photograph and CAD plan of the palace draped over surface TIN depicting the local topography.

88

Topographical Survey and Digitised Elevation Model of the Englianos Ridge engineering. Pollard, Hooper and Moore (1979) earlier developed a method of dating hedgerows that might have relevance to the dating of terraces, in which the age of a hedgerow is calculated as being equal to the product of 110 and the number of individual species found in a roughly 30-metre length of hedge, plus 30. The terrace walls along the Englianos slopes were examined in 1991 by Thomas Braun as a phytoarchaeological experiment; without any claim to its scientific basis, we summarise Braun’s results in Table 1.5. Remote sensing technology used by MARWP includes photogrammatic overlays of the region furnished by HAGS, and another set flown in March 1965 by the Royal Hellenic Air Force and provided by Jesse Fant of the MME project (Fant and Loy 1972). Additionally, Landsat5, Landsat7 and Quickbird geo-referenced images provide satellite spectral data. Using GIS soft ware Imagine (Erdas) and ArcInfo (ESRI), it has been possible to generate and to enhance three-dimensional models of the palace, the Englianos ridge and the wider area of Bronze Age Pylos. Taking advantage of our access to the classification utility of the GIS programmes of the University of Minnesota’s Remote Sensing Laboratory, we undertook a reconnaissance for undiscovered ancient quarries as a pilot GIS project in the early 1990s. Ground control points for operating, abandoned and known ancient quarries were acquired by GPS field location, followed by the obtaining of a valid satellite spectral signature characteristic of limestone quarries, then by supervised classification of the Landsat5 image. Pixels on the classified scene were highlighted and their UTM coordinates loaded into the Trimble Pathfinder, which navigated us to 11 hypothetical sites; ground inspection revealed that five of these locations were indeed ancient quarries. In 1992 and 1993 Joshua Distler conducted a geological provenance study of limestone samples taken from ashlar blocks used in palace construction, and from the newly located ancient as well as other local and modern quarries. His research is published here as chapter 5. Preceeding our GIS activity, a team of Greek specialists mapped and ground-truthed a valuable land-use map of the Western Peloponnesos (Toulios, Yassoglou and Moutsoulas 1990), research apparently overlooked by Zangger, Timpson, Yazvenko, Kuhnke and Knauss (1997).

1.72 Coarse surface model of the Englianos ridge and surrounding topography generated using HAGS 1:50,000 map 20metre contours and details digitised from the palace survey.

Lee (2001) and by Simon Price and Lucia Nixon (2005). James Wright (1980) included the handsome retaining wall that supports the Southwestern Building among his Bronze Age terrace walls, and the Price-Nixon criteria for assuming antiquity of terrace walls are found in the walls that retain the precipitous cliffs to the northwest and southwest of the Englianos ridge. If these are Bronze Age, they represent a compelling example of early geotechnical Calculated date (from the year 2000)

Period

Terrace

Number of species

1

17

94 bc–ad 306

Hellenistic/ Roman

2

24

864–464 bc

Archaic/ Classical

3

34

1964–1564 bc

Middle Helladic

4

44

3064–2664 bc

Early Helladic I

5

28

1304–904 bc

Late Helladic

6

41

2734–2334 bc

Early Helladic II

7

36

2184–1784 bc

Early Helladic III

Other potential GIS applications might include the documentation of naturally occurring springs in the region, springs that may have possessed religious significance in the Bronze Age as the locations of openair sanctuaries, and the mapping of Bronze Age tombs in southwestern Messenia, which were generally carved out of deposits of calcitic marl – softer than limestone and easier to excavate, but still resistant to erosion. An understanding of the spatial distribution of springs and tombs may provide a means of interpreting patterns of land use associated with specific cultural activities. The creation of an aspect DEM that highlights natural and eroded gullies dropping from the perimeter of the

Table 1.5 Results of an examination of terrace hedgerows on the Englianos ridge.

89

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 acropolis at Englianos, and comparing the collected data with concentrations of survey finds below the site, would also be of potential interest.

Blegen, C. W., Rawson, M., Taylour, L. W., and Donovan, W. P. 1973. The Palace of Nestor at Pylos in Western Messenia, III: Acropolis and Lower Town, Tholoi, Grave Circle, and Chamber Tombs, Discoveries Outside the Citadel. Princeton University Press, Princeton, NJ.

The Englianos landscape embraces a rich assortment of archaeological enterprises, in which emerging technologies and sciences can be brought into play. Our research at Pylos and the other MARWP projects – as with MME before us – have endeavoured to accumulate the raw material needed for these new analyses, leading to fresh understandings of the western Messenian past and present.

Cooper, F. A. 1986. ‘The Stones of Bassai’, in Science in Archaeology. Fitch Laboratory Occasional Papers 2, ed. R. E. Jones and H. W. Catling, 21-38. BSA, London. Cooper, F. A. 1988. ‘The Quarries of Mount Tayegetos in the Peloponnesos, Greece’, in Classical Marble: Geochemistry, Technology, Trade, ed. N. Herz and M. Waelkens. Proceedings of the NATO Advanced Research Workshop on Marble in Ancient Greece and Rome, Il Ciocco, Lucca, May 1988, 65–76. Kluwer Academic Publishers, Dordrecht and Boston.

Bibliography Attanasio, D., Brilli, M., and Ogle, N. 2006. ‘The Isotopic Signature of Classical Marbles’, Studia Archaeologica 145, 108–14.

Cooper, F. A. 1996. The Temple of Apollo Bassitas, I: The Architecture. ASCSA, Princeton, NJ.

Bennett, E. 1954. Unpublished Pylos excavation notebook. ASCSA, Athens, and University of Cincinnati, OH.

Cooper, F. A. 2002. Houses of the Morea / Σπιτια του Μορεα, Melissa Pubishing, Athens.

Bennett, E. 1999. ‘Postscript’, in Rethinking Mycenaean Palaces, ed. M. Galaty and W. Parkinson. Cotsen Institute Monograph 41, 100-01. Los Angeles.

Cooper, F. A., and Brenningmeyer, T. 2005. ‘Computer-Assisted System for the Reconstruction of Shattered Objects’, in Proceedings of the International Society for Optical Engineering: Sensors, and Command, Control, Communications, and Intelligence (C3I) Technologies for Homeland Security and Homeland Defense IV, vol. 5778, 644–56. SPIE, Los Angeles.

Bergquist, B. 1967. The Archaic Greek Temenos: A Study of Structure and Function. Gleerup, Lund.

Cooper, F. A., and Brenningmeyer, T. 2009. ‘A Digital Approach for Practical Reconstructions of Fragmented Murals’, in Layers of Perception, ed. A. Posluschny, K. Lambers and I. Herzog. Proceedings of the 35th Computer Applications and Quantitative Methods in Archaeology (CAA), Deutsches Archäologisches Institut, Berlin, April 2–6, 2007. Habelt, Bonn.

Blegen, C. W. n.d. Unpublished manuscript for The Palace of Nestor of Pylos. ASCSA, Athens, and University of Cincinnati, OH. Blegen, C. W. 1956. ‘Palace of Nestor: Excavations of 1955’, AJA 60, 95–101. Blegen, C. W. 1957. ‘Palace of Nestor, Excavations of 1956’, AJA 61, 129–35.

Fant, J. E., and Loy, W. G. 1972. ‘Surveying and Mapping’, in The Minnesota Messenia Expedition: Reconstructing a Bronze Age Regional Environment, ed. W. A. McDonald and G. Rapp, 18–35. University of Minnesota Press, Minneapolis.

Blegen, C. W. 1959. ‘Palace of Nestor: Excavations of 1958, Part I’, AJA 63, 121–27.

Griebel, C., and Nelson, N. 1993. ‘Post-Mycenaean Occupation at the Palace of Nestor’. Abstract, 94th Annual Meeting of the Archaeological Institute of America, AJA 97, 331.

Blegen, C. W. 1961. ‘Palace of Nestor: Excavations of 1960, Part I’, AJA 65, 153–63. Blegen, C. W. 1965. ‘Palace of Nestor: Excavations of 1964, Part I’, AJA 69, 95–101.

Griebel, H. B. 1992. ‘Greek Bake Ovens: Technological Aspects’. Abstract, 93rd Annual Meeting of the Archaeological Institute of America, AJA 96, 59–60.

Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and Their Contents. Princeton University Press, Princeton, NJ.

Griebel, H. B. 1997. ‘Greek Bake Ovens’, in The Encyclopedia of Vernacular Architecture of the World, ed. P. Oliver, vol. X. Cambridge University Press, Cambridge.

Blegen, C. W., and Rawson, M. 1967. A Guide to the Palace of Nestor. University of Cincinnati, OH.

Herz, N., Cooper, F. A., and Wenner, D. 1982. ‘The Mani Quarries: Marble Source for the Bassai Temple in the Peloponnesos’. Abstract, 83rd Annual Meeting of the Archaeological Institute of America, AJA 86, 270.

90

Bibliography Nelson, M. C. 2007. ‘Minoan and Mycenaean Stone Revetment’, in Krinoi kai Limenes: Studies in Honor of Joseph and Maria Shaw, ed. P. Betancourt, M. C. Nelson and H. Williams, 17–22. INSTAP Academic Press, Philadelphia, PA.

Hofstra, S. 2000. Small Things Considered: The Finds from LH IIIB Pylos in Context. Unpublished Ph.D. dissertation, University of Texas at Austin. Immerwahr, S. 1989. Aegean Painting in the Bronze Age. Penn State University Press, University Park, PA.

Palaima, T. G., and Wright, J. C. 1985. ‘Ins and Outs of the Archive Rooms at Pylos: Form and Function in a Mycenaean Palace’, AJA 89, 251–62.

Isaakidou, V., Halstead, P., and Davis, J. 2002. ‘Burnt Animal Sacrifice at the Mycenaean Palace of Nestor, Pylos’. Antiquity 76, 86-92.

Pappalalardo, U. 1998–2000. ‘Pittura romana a Corinto: il South-East Building’, ASAtene 76–78, n.s. 60–62, 315–60.

Jeffery, L. 1961. Local Scripts of Archaic Greece. Oxford University Press, Oxford.

Papathanasopoulos, G. 1958. Unpublished Pylos excavation notebooks. ASCSA, Athens, and University of Cincinnati, OH.

Kittredge, W. 1962. Unpublished Pylos excavation notebook. ASCSA, Athens, and University of Cincinnati, OH.

Pollard, E., Hooper, M. D., and Moore, N. W. 1979. Hedges. Collins, London.

Lang, M. 1969. The Palace of Nestor At Pylos in Western Messenia, II: The Frescoes. Princeton University Press, Princeton, NJ.

Popham, M. 1991. ‘Pylos: Reflections on the Date of its Destruction and on its Iron Age Reoccupation’, Oxford Journal of Archaeology 10, 315–24.

Lazzarini, L. 2007. ‘Il Rosso antica Tenario’, in Poikiloi Lithoi, Versicvlores Macvlae: I marmi colorati della Grecia Antica. Edizioni Dell’Ateneo, Pisa and Rome, 71–96.

Price, S., and Nixon, L. 2005. ‘Ancient Agricultural Terraces: Evidence from Texts and Archaeological Survey’ AJA 109: 665– 94.

Lee, W. 2001. ‘The Pylos Regional Archaeological Project, Part IV: Change and the Human Landscape in a Modern Greek Village in Messenia’, Hesperia 70, 49–98.

Rawson, M. 1958. Unpublished Pylos excavation notebook: ‘Pylos 1958: Areas MZ–MY,’ Vols I and II. ASCSA, Athens, and University of Cincinnati, OH.

McCullum, L. R. 1987. Decorative Programs in the Mycenaean Palace at Pylos: the Megaron Frescoes. Unpublished PhD dissertation, University of Pennsylvania, Philadelphia.

Reusch, H. 1956. Die Zeichnerische Rekonstruktion des Frauenfrieses im boötischen Theben. Akademie Verlag, Berlin.

McDonald, W. A. and Rapp, G. eds, 1972. The Minnesota Messenia Expedition: Reconstructing a Bronze Age Regional Environment. University of Minnesota Press, Minneapolis.

Rhodes, R. 1983. The Beginnings of Monumental Architecture in the Corinthia. Unpublished PhD dissertation, University of North Carolina, Chapel Hill.

Meggiolargo, V., Molin, G. M., Pappalardo, U., and Vergerio, P. P. 1997. ‘Contribution to Studies on Roman Wall Painting Materials and Techniques in Greece: Corinth, the Southeast Building’, in Roman Wall Painting: Materials, Techniques, Analysis and Conservation, ed. H. Be’arat, M. Fuchs, M. Maggetti and D. Pannier, 105–118. Institute of Mineralogy and Petrography, Fribourg.

Rhodes, R. forthcoming. The Reconstruction and Interpretation of the Early Temple of Apollo at Corinth. Corinth Publication Series. Princeton, NJ. Rodenwaldt, G. 1912. Tiryns II: Die Fresken des Palastes. Eleft heroudakis and Barth, Athens.

Melena, J. L. 2000–01. ‘54 Joins and Quasi-joins of Fragments in the Linear B Tablets from Pylos’, Minos, n.s. 35–36, 376 and 380–84.

Rutter, J. B. 2005. ‘Southern Triangles Revisited: Laconia, Messenia and Crete in the 14th–12th Centuries bc’, in Ariadne’s Threads: Connections Between Crete and the Greek Mainland in Late Minoan III (LM IIIA2 to LM IIIC), ed. A. L. D’Agata and J. Moody, 17–50. ASAtene, Athens.

Miller, S. G., Kraynk, L. H., and Birge, D. E. 1992. Excavations at Nemea, I. Topographical and Architectural Studies: The Sacred Square, the Xenon and the Bath. University of California Press, Berkeley.

Shaw, J. W. 1973. Minoan Architecture: Materials and Techniques. ASAtene 49, n.s.33, Rome.

Mountjoy, P. A. 1993. Mycenaean Pottery: An Introduction. Oxford University Committee for Archaeology , Oxford.

Shaw, J. W., and Shaw, M. C. 2006. Kommos: An Excavation on the South Coast of Crete, IV: The Greek Sanctuary. Princeton University Press, Princeton, NJ. Shaw, M. C. 1993. ‘The Aegean Garden’, AJA 97, 661–85.

91

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98 Shelmerdine, C. W. 1984. The Perfumed Oil Industry at Pylos. Studies in Mediterranean Archaeology 34. Paul Åström Förlag, Göteborg. Shelmerdine, C., and Bennet, J. 1995. ‘Documents from Bronze Age Pylos’, Kadmos 34, 133–36. Stocker, S. R., and Davis, J. L. 2004. ‘Animal Sacrifice, Archives, and Feasting at the Palace of Nestor’, Hesperia 73, 179–95. Tomlinson, R. 1976. Greek Sanctuaries. Paul Elek, London. Toulios, L. G., Yassoglou, N. J., and Moutsoulas, M. 1990. ‘Landuse Mapping in West Messenia, Greece, Using Satellite Imagery’, Remote Sensing 11, 1645–61. Valmin, M. S., 1938. The Swedish Messenia Expedition: Excavations at Malthi-Dorion. Gleerup, Lund. Wilson, A. 2008. ‘Hydraulic Engineering and Water Supply’, in The Handbook of Engineering and Technology in the Classical World, ed. J. P. Oleson, 285–318. Oxford University Press, Oxford. Wright, J., 1980. ‘Mycenaean Palatial Terraces’, Atenische Mitteilungen 95, 59–86. Zangger, E., Timpson, M. E., Yazvenko, S., Kuhnke, F., and Knauss, J. 1997. ‘The Pylos Regional Archaeological Project, Part II: Landscape Evolution and Site Preservation’, Hesperia 66, 549– 641 (with bibliogaphy on Mycenaean hydraulic engineering).

Revised 2011

92

Sample code

E

E

E

E

E

E

E

E

E

E

Record no.

1

2

3

4

5

6

7

93

8

9

Table 1.6 Results of scientific analyses carried out on materials collected by Blegen and MARWP.

10

10

9

8

7

6

5

4

3

2

1

Sample no.

St. Paul

Testing Lab

Results

Room 23, earth from east corner where Jar 3 stood Room 23, in pithos 7, at 0.65 below top of bench Room 23, pith 14, earth & char. from near bottom Room 23, jar 3, earth from bottom of jar

Room 23, pithos 7

Magazine 23 pithos 16

2.55 frm ins NW w. meg. 0.30 frm e. w. tr. on fl 0.70 deep N corner near hole under pots Magazine C (b) / 23 pithos 8 Magazine 23, pithos 9

Collection location

black earth

black earth from bottom

20

23

RHu

RHu

23

23

RHu

RHu

earth

23

RHu

MR

MR

MR

RHu

Excavator initials

23

earth at very bottom

black earth from bottom

23

23

earth from bottom

black earth

24

23

Context description

Room no.

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

53.149

55.020

Notebook pg. no.

1/7/55

11/7/55

22/7/55

1/7/55

22/6/55

28/5/56

28/5/56

28/5/56

1/1/53

1/1/55

Notebook date

5YR5/2

5YR5/2

7.5YR6/

7.5YR6/

7.5YR7/

10YR5/2

5Y7/2

5Y6/2

10YR4/1

10YR6/2

root material, minor lime plaster

abundant large charcoal

charcoal, up to 7mm across

possibly some lime plaster, little or no charcoal very minor charcoal & lime plaster abundant decaying root material plant impressions, no visible charcoal

some very fine charcoal

very abundant charcoal

small chunks plaster material(?), very white

Munsel no. Inclusions

Scientific Analyses - Table 1.6

94

E

E

20

E

16

19

E

15

E

E

14

18

E

13

E

E

12

17

E

11

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

20

19

18

17

16

15

14

13

12

11

St. Paul

Floor room 9, ca. under smaller channeled stone Magazine 24, pithos 10, published 2 Area M, room 32, from upper stirrup base 34 Area M, room 32, inside pithos 5

Magazine 9, pithos old 10, new 2 Magazine 9, pithos old 9, published 3 Magazine 24, pithos old 11, published 1

Magazine 24, pithos old 8, published 4

Room 23, pithos 1, bottom Room 23, pithos 17

RHu

earth from bottom

black earth w/ MR white veins gray powdery earth

24

32

32

MR

RHu

24

24

54.177

54.175

0.000

55.120

0.000

21/7/54

20/7/54

28/5/56

2/7/55

28/5/56

28/5/56

0.000

24

28/5/56

0.000

black earth from upper part of fi ll black earth from near bottom

24

29/5/56

22/7/55

26/5/56

0.000

0.000

0.000

earth from bottom

RHu

24

23

23

10YR6/1

10YR3/1

5Y7/1

7.5YR6/

7.5YR6/

10YR4/2

10YR3/1

5YR3/1

7.5YR6/

5YR6/4

chert fragment, root material, very minor charcoal

root material

minor charcoal, lime plaster & root material very minor charcoal, bone material, decaying root material plant impressions, probably not identifiable

minor lime plaster, some root material modern root material, very minor lime plaster & charcoal very abundant fine grain charcoal, some lime plaster & root material possibly none

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

95

E

E

E

30

31

E

26

29

E

25

E

E

24

28

E

23

E

E

22

27

E

21

30

29

28

28

27

26

25

24

23

22

21

Canberra

St. Paul

see discussion

Area ME 4&2, rm 46,top heap (angonaria) frt of dr. into corr. Area ME, room 46, above NW door Area ME, room 46, above NW door jamb Area ME6, with pithos fragments Area ME2 & 6, with pithos fragments Area ME4, rm 46, 0.35-0.70 from wall b, from tp & bot Area ME8 & 2, 1.95-2.35 from wall A

Area M, room 32, contents of pithos 8 Area M, room 34, earth from inside big jar Area M, corr. 35, earth agnst face dressed stones in corr. Area ME #4, room 46

brownish black earth

hard gray earth black earth, oily? from black stuff

light gray earth

46

46

46

46

black earth

46

46

46

MR

MR

MR

MR

MR

MR

MR

56.058

56.058

56.050

56.050

56.068

56.068

56.056

56.056

black earth w/ chimney fragment light gray earth

MR

46

54.064

MR

35

same kind of earth elsewhere, M

54.099

MR

34

54.175

MR

32

8/6/56

8/6/56

2/6/56

2/6/56

13/6/56

13/6/56

7/6/56

7/6/56

15/6/54

14/7/54

20/7/54

7.5YR5/

5YR5/1

5YR3/2

10YR7/1

5YR4/2

7.5YR3/

10YR7/2

5YR3/1

7.5YR7/

2.5YR4/

7.5YR7/

lime plaster, charcoal & root material lime plaster & modern root material

lime plaster & root material

root material, very minor lime plaster

root material, lime plaster & rock fragments

minor lime plaster

Scientific Analyses - Table 1.6

96

E

39

E

36

E

E

35

38

E

34

E

E

33

37

E

32

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

38

37

36

35

34

33

32

31

Canberra

Canberra

Canberra

see discussion

see discussion

see discussion

Area ME, room 46, west of hearth Area ME, rm 46, 2.35-2.55 fr wall W, 0.951.15 fr wall A Area M, room 27, north magazine, pithos in south corner Area M, north magazine, pithos in south corner, room 27 Area ME, black earth w/ fallen stones, refuse fr olive press EBW, NW cor Court H, sample of earth from dark stone layer Area M2, magazine D, room 105, pithos central left, SQU #6 Area M2, magazine D, room 105, central pithos left #4 soil sample

soil sample

105

105

black earth w/ MR fallen stones

47

earth from dark stone layer

gray earth inside

27

88

accretion on inside

27

DHF

DHF

EPB

MR

MR

MR

black earth

46

MR

blackish brown

46

58.133

58.136

56.103

56.061

54.183

54.183

56.059

56.058

28/7/58

29/7/58

5/6/56

9/6/56

27/7/54

27/7/54

8/6/56

7/6/56

5YR4/1

5YR4/1

5YR4/3

5YR4/2

10YR6/4

2.5YR5/

5YR5/2

7.5YR5/

root material & minor lime plaster

very minor charcoal, root material

root & shell material, shell looks like E35

abundance of shell fragments, some root material

root material & lime plaster

very minor charcoal

some charcoal

root material & lime plaster

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

E

E

E

46

47

E

43

45

E

42

E

E

41

44

E

40

97

46

45

44

43

42

41

40

39

St Paul

Area M2, magazine D, room 105, pithos central left #7 Area M2, magazine D, room 105, pithos central left #3 Area M2, magazine D, room 105, central pithos right #7 Area M2, magazine D, room 105, central right row pithos #1 Area M2, magazine D, pithos side left #2 Area M2, magazine D, room 105, pithos side right #1 Area M2, Magazine D, room 105, pithos central left #6 Area M2, magazine D, room 105, pithos side right #5 soil sample

soil smaple

soil smaple

105

105

soil sample

105

105

soil sample

105

soil sample

soil sample

105

105

soil sample

105

DHF

DHF

DHF

DHF

DHF

DHF

DHF

DHF

58.139

58.135

58.138

58.137

58.125

58.136

58.130

56.134

29/7/58

28/7/58

29/7/58

29/7/58

23/7/58

29/7/58

27/7/58

28/7/56

10YR4/2

10YR4/2

7.5YR5/

5YR4/2

5YR3/1

5YR4/1

7.5YR5/

5YR5/3

shell fragments, minor root material

root material, shell fragment

rounded quartz grains, root material, minor lime plaster root material, very minor lime plaster, very minor charcoal

root material, very minor lime plaster

root material, minor lime plaster, very minor charcoal

root material & very minor lime plaster

minor lime plaster, very minor charcoal, root material

Scientific Analyses - Table 1.6

E

E

E

54

55

E

51

53

E

50

E

E

49

52

E

48

98

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

54

53

52

51

50

49

48

47

Area M2, magazine D, room 105, central right row, pithos #2 Area M, NE of door from SENW corridor to NE-SW corridor Anas. tr 3, rm 3 (SE), just above stratum of fallen pavement Area M2, magazine D, room 105, pithos central right #4 Area M2, magazine D, room 105, pithos #3 central right Area M2, magazine D, room 105, central right pithos #5 Area M2, magazine D, room 105, pithos central right #6 Floor sample of burnt earth

soil sample

105

0

soil sample

soil sample

105

105

sample of burned earth

0

soil sample

sample of contents of pithos

26

105

soil sample

105

RHu

DHF

DHF

DHF

DHF

55.082

58.134

58.133

58.130

58.131

60.081

60.054

MR

MR

58.126

DHF

1/1/55

29/7/58

28/7/58

26/7/58

26/7/58

27/5/60

21/5/60

24/7/58

10YR3/1

10YR4/2

10YR3/1

10YR4/2

7.5YR4/

10YR3/1

10YR3/1

10YR5/4

root material & minor lime plaster

shell fragments & root materials

root material & lime plaster

root material & shell fragments

root fragments

minor charcoal

plant & root fragments, minor charcoal

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

99

E

63

E

60

E

E

59

62

E

58

E

E

57

61

E

56

61

60

59

58

57

57

56

55

St .Paul

St. Paul

Belvedere B30, dark gray dumpish earth btwn walls H &J Blv B30, dumpish grn earth, w.H (cf dk gry frm btwn walls H&J) NE stairway, earth from dowel hole in N corner of SE block Remains of inner column of propylon Olive refuse in hollow deep in roots of ol. tree #(?) 31m from NW corner of fence at level of bottom of fence

NE stairway, from black deposit from SE anta blcok From bottom of pithos

MR

61.039

0.000

0

0

0.000

0.000

59.154

59.154

58.027

61.062

0

black stuff w/ many sherds of 1 pot

dumpish MR greenish earth

0

0

MR

0

MR

P

blue gray

0

0

1/6/61

8/6/61

1/1/54

26/6/59

26/6/59

20/6/58

22/7/61

7.5YR5/

7.5YR4/

5YR5/2

5Y5/3

10YR5/3

7.5YR4/

7.5YR5/

very minor root material & charcoal

very minor root material & shell fragments

no inclusions

very minor root material

very minor root material & lime plaster very minor root material, lime plaster & shell fragments

minor root material

Scientific Analyses - Table 1.6

E

MM

MM

MM

MM

MM

MM

MM

MM

MM

MM

MM

64

65

66

67

68

69

70

71

100

72

73

74

75

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

11

10

9

8

7

6

5

4

3

2

1

62

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

reddish painted lime plaster chunks

combo of furnace & rock fragments (most), 1 s

furn. products: flint, quartz, 4 cm 3 slag, mag & non-mag

concretions

iron concretion fragments

concretion fragments

iron concretion

4-5 rock fragments, 1 quartz fragment, burnt mudbrick(?)

fragment of quartzite

concretion

2 concretion, 4 shell fragments

53

0

Area ME, tower NE of propylon, contents of pithos

Area MNE, room 100

Area M

Area ME

Room 39

Room 46

0

stuff w/ ivory stone ore?

MR

MR

60.152

60.052

0.000

MR

98

100

54.082

56.099

0.000

0.000

MR various coloured chunks & stones

10R3/4

28/6/60

20/5/60

21/7/54

2/7/56

10R3/2

10R3/2

10R4/6

7/6/57

28/5/56

10R4/8

57.061

56.034

MR

26/6/56

10YR4/1

10R5/6

56.091

MR

22/7/61

0.000

61.057

MR

what is this? MR metal? colour?

what is this?

what is this?

what is this? metal? red colour stone? cylindrical? what is this?

charcoal gray

30

0

39

46

0 Area ME #3, stairway 54, above steps #77 Rm 27, sample 27 of red stone of bases in the north room 97 Area MNE #7, bits of red colour found with red brick

NE stairway, from black deposit from SE anta block Area ME

furnace products, flint, quartz

minor root material

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

101

S

S

82

S

79

81

S

78

S

S

77

80

S

76

7

6

5

5

4

2

1

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

(1)-Fragment of plaster. W: 2.2 cm, L: 2.9 cm, Th: 0.6 cm. Dendrites on both sides. (1)-Pelecypod shell fragment embedded in mud. Possibly a Pecten sp. W: 3.5 cm, L: 2.5 cm, Th: 1.7 cm.

(1)-Pelecypod shell fragment from a large species, nacreous. L: 6 cm, W: 2.7 cm, Th: 1.5 cm. (1)-Thick, smooth shell fragment (Gastropod?). W: 46 cm, th: 1 cm+. (5+)-Smaller, nacreous, thin fragments of Pelecypod shells. (1)-Fragment of Gastropod siphonal canal process, L: 1.8 cm. (1)-Fragment of plaster. W: 4.3 cm, L: 4.8 cm, Th: 0.6 cm. Dendrites on both sides. (1)-Gastropod shell, whole, possibly a Ficus sp. W: 1.8 cm, L: 2.9 cm. Same bag as above.

(1)-Gastropod shell, dextral, siphonal canal process missing, possibly a Ficus sp. L: 6 cm, W: 4 cm. 18

18

20

Area R, room 18

Room 18

Area R, room 20

Area MNE, Leonodis

92

Area R, court 85 85 From washing 0 sherds

0

Area ME, from in front of NE anta of Propylon

flint & obsidian, it’s a shell!

fossil

fossil

MR

MR

pieces of shell MR

piece of shell

MR

0.000

0.000

53.013

0.000

0.000

0.000

0.000

25/5/53

20/5/53

5/5/53

4/7/56

Scientific Analyses - Table 1.6

S

S

S

S

S

S

S

83

84

85

86

87

88

102

89

13

12

11

3

10

9

8

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

(1)-Pelecypod shell fragment. W: 5.3 cm, L: 7.5 cm, Th: 1.2 cm. (1)-Gastropod shell, dextral, whole. Possibly a Ficus sp. W: 1.9 cm, L: 2.8 cm. (1)-Gastropod shell, apical tip and aperture end are missing. W: 1.6, L: 3.5 cm. (1)-Large piece of marl with impression of a Pelecypod shell, possibly a Pecten sp. W: 11 cm, L: 15 cm, Th: 7 cm. (8)-Large Echinoid half intact, 7 smaller fragments of the same species. W: 9-9.5 cm. (1)-A piece of marl with embedded Pelecypod shell fragment, possibly a Pecten sp. W: 6 cm, L: 9.5 cm, Th: 5.cm. (1)-A piece of marl embedded with Pelecypod shell fragments, possibly a Pecten sp., and various Gastropod shells. W: 5.5 cm, L: 10.5 cm, Th: 3.8 cm. (7)-Pieces of marl embedded with Pelecypod shell fragments, possibly a Pecten sp. W: 3 cm, L: 3.5 cm, Th: 0.5 cm. Rest are smaller and some are thicker. 18

88

Area R, room 18

Section H, 0.15-0.40

shell (2 fossilised shells)

petrified shells MR

58 Area MNE, court 58, SE of SE wall

RHo

MR

RHo

MR

33 Area M, room 33, from burned brick fi ll

sea shell

fossil, big piece of marl

88

Section H, 0.15-0.30

2 shells

shell

97

Rooms 97 & 97 97

Area MNE, with sherd lot 64, ploughed earth

30/6/56

30/6/54

0.000

0.000

1/5/53

25/5/53

7/6/57

53.015

0.027

53.014

0.000

57.062

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

103

S

BT

BT

96

97

S

93

95

S

92

S

S

91

94

S

90

2

1

21

19

18

16

15

14

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

(3 larger pieces)-Fragments of a Pelecypod shell. The large shell is possibly a costa of a large Pecten sp. or a fragment of a razor clam, nacreous. W: 2.2 cm, L: 6.1 cm, Th: 2+ cm. (1)-Pelecypod shell, whole, articulated, fossilised, possibly a Dosinia sp. W: 3.5 cm, L: 3.7 cm, Th: 1.5 cm. (9)-Thin fragments of a quite large Pelecypod shell, nacreous. W: 2.2 cm, L: 6.4 cm, Th: 0.1 cm (largest). Some fragments slightly thicker. (1)-Pelecypod shell, whole, articulated, fossilised, possibly a Dosinia sp. W: 3 cm, L: 4.2 cm, Th: 1.7 cm. (1)-Thick (4.5 cm) Pelecypod shell, fossilised, articulated. Possibly an Arca sp. W: 6.5 cm, L: 5.4 cm. (1)-One valve, spines present, possibly a. Laevicardlum sp.? W: 8.5 cm, L: 8.7 cm, Th: 2+ cm. (1)-Burned mammal long bone fragment , light blue pigment on one edge (possibly painted). L: 3.1 cm, W: 1.6 cm, Th: 0.5 cm. (13)-Fragments of mammal long bone, appear to be burned. shell

0

6

11

Area M

Throne room, SW section of

Room 11, undisturbed

10 bones

painted & burned bone

petrified shell

0 Belvedere (w13) b34, west of wall R

RJB

GM

MR

petrified clam MR shell

0

Area M

MR

petrified clam MR? shell

MR

shell

58

Area ME, court 58

shell

Belvedere, B21, 0 in drain

0

Area M

54.147

0.000

58.142

0.000

58.142

59.137

0.000

54.180

1/1/54

1/1/52

1/7/58

29/6/59

1/7/58

20/6/59

26/7/54

Scientific Analyses - Table 1.6

BT

BT

BT

BT

BT

BT

BT

BT

BT

98

99

100

101

102

103

104

104

105

106

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

12

11

10

9

8

7

6

5

4

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

(2)-Fragments of Pelecypod shell, nacreous. (1)-Fragment of herbivore molar. (6)-Fragments of mammal long bone.

(4)-Larger fragments of long bones from medium size mammal. (20)-Smaller fragments of bone. (3)-Larger fragments of long bone from a large mammal. (19)-Smaller fragments, many fossilised (two probably not bone). (3)-Fragments of herbivore molars from a large animal. (1)-Smaller bone possibly a phalanx or metatarsal/metacarpal. L: 4.1 cm, Th: 0.5-1 cm. (24)-Fragments of large mammal fossilised long bone and other bone. (1)-Pelecypod shell fragment, nacreous. (2)-Fragments of fossilised mammal long bone.

(11)-Fragments of burned mammal bone. Including one long bone fragment and a proximal end of rib from a large mammal. (4)-Fragments of fossilised bone. (18)-Fragments of burned bone, some long bone.

lots of bone fragments

bits of bone

Area BB, room 15 15

18 Area R, NE end and room 18 Rooms 18-20, 18 doorway

53.018

0.000

pieces of shell, MR bone & wood

54.065

54.000

54.130

54.045

54.107

54.067

54.057

MR

RJB

teeth from fi ll RJB

Area BB, ext.2, 0 vestibule entrance way

RJB

bone fragments

RJB

RJB

16 bone fragments

bone & ivory

RJB

RJB

ivory?

9 bone fragments

Area BB, NW- 0 SE intrusion over HN, from intrusion

Area BB, room 13 13 Area BB, frag 0 from stairway 15 fi ll Area BB, room 16 12, + doorway

Area BB, room 13 13

4/6/53

6/6/53

1/1/54

24/6/54

1/1/54

1/1/54

5/7/54

22/6/54

1/1/54

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

105

BT

111

BT

BT

110

113

BT

109

BT

BT

108

112

BT

107

18

17

16

15

14

13

12

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

(9)-Mammal long bone fragments, fossilised.

(1)-Phalanx or metatarsal/ metacarpal bone. (17)-Fragments of mediumlarge size mammal long bone (two fragments very dark). (1)-Boar tusk (Sus sp.), 5 cm long. (1)-Mandible/maxilla fagment. (13)-Fragments of medium-large size mammal long bones and other bones. (11)-Fragments of mammal long bones. (5)-Fragments of fossilised mammal long bones. (31)-Fragments of long bones (mostly), some fossilised. (2)-Tooth fragments (one herbivore molar). (1)-Proximal end of ulna (small mammal). (2)-Pelecypod shell fragments. (1)-Gastropod shell fragment. (1)-Fragment of distal end of humerus, small mammal, fits perfectly with ulna in previous (BT 16) sample. (19)-Fragments of long bone, some fossilised. (1)-Fragment of possible tooth enamel. (1)-Phalanx or metatarsal/ metacarpal bone.

18

Room 39

39

33

32

Room 32

Room 33

25

25

Area M, corridor W25 Corridor 25

21 Area R, outside of NW wall of room 21

Area R, room 18

MR

8 small bone & teeth fragments

20 bones

lost of small bones

bones

bone

MR

bones & boar’s MR tusk

15 bones

0.000

0.000

0.000

0.000

53.106

53.135

53.048

30/6/53

24/6/53

4/6/53

Scientific Analyses - Table 1.6

106

BT

BT

BT

BT

120

121

122

BT

117

119

BT

116

BT

BT

115

118

BT

114

27

26

25

24

23

22

21

20

19

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

MR

bone fragments bones

RJB

MR

MR

burned burn bone

bits of bone, ivory?, petri astraga

Area M, room 0 58, north corner (4)-Fragments of fossilised Area BB, court 88 mammal long bones. 88 Corridor 95 0 (1)-Small Pelecypod shell fragment, nacreous. (1)-Fragment of skull bone. (4)-Teeth, three possible canines.

45

Area ME #5, corridor 45

shells & bones

bone, burned bone

44

Stoa 44

MR

Area M, room 57 57

42

Area M, PE, court 42, NE of wall W

5 fragments from single bone bones & bits of carbon

20 small bone MR fragments

40

Room 40

Area M, room 54 54

(10)-Fragments of long bones, two from a large mammal. (5)-Fragments of long bones from a medium size mammal. (2)-Burned long bone fragments. (8)-Pieces of charcoal. (12)-Fragments of Gastropod and Pelecypod shells. (22)-Fragments of long bone and other bone. (1)-Distal end of metacarpus/metatarsus, possibly a Cervus sp. (1)-Astragalus, possibly a Cervus sp. (1)-Crown of a molar (five cuspid). (7)-Smaller bone fragments, some long bone. (1)-Fragment of a tooth. (14)-Fossilised mammal bone fragments, some long bone. (10)-Fragments of burned bone, mostly long bone, smaller mammal. A lot of burned bone, charcoal, ashes, and sand. 17/7/54 22/7/57

0.000

24/6/54

30/6/54

30/6/54

26/5/56

8/6/55

54.121

54.091

54.110

54.108

56.028

0.000

55.058

0.000

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

107

BT

BT

129

BT

126

128

BT

125

BT

BT

124

127

BT

123

34

33

32

31

29

30

28

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

horn

piece of bone or horn

100

bones

100

0

Area WB

MR

MR

MR

RJB

teeth & bones ELB

tip of horn w/ ivory

0

North room and Corridor

petrified clam MR shell

petrified fishbone?

100

0

Area M

Area MNE, room 100, box 2 Area MNE, room 100

0

Area M, from dump

(15+)-Fragments of long bone, larger mammal. Many smaller fragments of bone with clay and sand. (1)-Fragment of a possible Area MNE, scapula (spine/ridge) from room 100 a medium-large animal.

(1)-Possibly a distal fragment of a metapodial/ metacarpal bone (fibula/ radius) of an Equus sp. or a Cervus sp. (4.8 cm long). (1)-Whole (articulated) Pelecypod shell, fossilised, possibly Dosinia sp. (16)-Fragments of stone, some very dark. (9)-Fragments of bone, some long bone. (1)-A piece of charred wood. (6)-Fragments of fossilised long bone. (1)-Fragment of a herbivore molar, fossilised. (2)-Fragments of fossilised bone. (2)-Fragments of Pelecypod shell. (1)-A piece of flint (reddish). (4)-Fragments of dark fossilised bone. (1)-A piece of stone. (19)-Fragments of bone, mostly long bone. (1)-Tooth, canine. (1)-Distal end of humerus, smaller mammal. (1)-A charred piece of horn?

60.047

60.047

60.045

54.165

0.075

54.137

0.000

19/5/60

19/5/60

19/5/60

1/1/54

17/6/57

7/7/54

Scientific Analyses - Table 1.6

108

BT

BT

139

BT

135

138

BT

134

BT

BT

133

137

BT

132

BT

BT

131

136

BT

130

46

45

44

43

42

40

39

38

37

35

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Duluth

Area M, outside wall 1

91

(1)-Possibly a metapodial bone. 3.8 cm long.

27 Room 27, from pithos in south corner (2)-Fragments of small Area M, room 40 long bone. 40 0 (2)-Fragments of herbivore Area M molars. (6)-Fragments of bone (two fossilised long bone fragments). (1)-Fragment of a Gastropod shell. Area R, room 5 (1)-Distal end of tibia. 5 Large mammal possibly Bos sp. 0 Section H4, (3)-Bone fragments, possibly proximal ends of a from hole -0.50-(-1.10) large mammal ribs. Area H53, 0 (1)-A piece of obsidian (1.8 cm long, 0.5 cm wide). 0.45-0.60 (9+)-Fragments of bone, possibly teeth (molar). (5)-Fragments of a large Section H, 0 herbivore molars. -0.4-(-0.50) 0 (1)-Boar tusk (Sus sp.), 6.3 Section H, basket #2, cm long. (4)-Fragments -0.15-(-0.40) of medium size mammal stone level long bone. (2)-Fragments of bone. Trench Y, 0 (2)-Fragments of bone. 0-0.5 (1)-Large axis (vertebra), possibly Bos sp.

(38+)-Fragments of large mammal long bone. Many smaller fragments of bone and dust present.

bone

bone fragments bone fragments

bone fragments

0.000

0.000

RHo

GM

0.000

53.077

RHo

RHo

53.037

bone fragments

54.142

0.000

MR

bones

54.053

54.183

55.157

big bone

MR

MR

MR

burned bone

bone

horn from pavement

1/1/52

21/6/53

24/6/53

12/6/53

4/6/53

30/5/53

9/7/54

4/6/54

27/7/54

22/7/55

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

109

BT

BT

BT

142

143

BT

141

140

51

50

49

48

Duluth

Duluth

Duluth

Duluth

Area MNE

EBW, corner on outside of wall S

Great drain

(2)-Fragments of Pelecypod From drain, sherds S2, #1 shells. (3)-Fragments of scapula. (27)-Fragments of long bones possibly Bos sp. and Ovis sp. (6)-Fragments of mandible, one with a tooth. (1)-Tooth.

(2)-Gastropod shell fragments. (16)-Limestone plaster. (11)-Fragments of teeth, one boar tusk (Sus sp.), herbivore molars. (22)-Fragments of long bones, some from a large mammal. (10)-Fragments of mandible from a medium size mammal. (2)-Fragments of vertebra from a medium size mammal. (1)-Distal end of femur possibly Bos sp. (two fragments). (1)-Proximal end of tibia possibly Bos sp. (matches femur). (1)-Distal end of metatarsus possibly Bos sp. (1)-Distal end of humerus possibly Ovis sp. (22+)-Small fragments of bones. (16)-Fragments of bone, one phalanx or metatarsal/ metacarpal bone. (1)-Burned bone fragment. (3)-Fragments of cranium. 0

0

0

0

fragments of skull?

bones

teeth and bone

CWB

MR

EPB

52.043

58.099

55.085

0.000

13/6/52

17/6/58

22/7/55

1/1/54

Scientific Analyses - Table 1.6

BT

BT

WC

WC

WC

WC

WC

WC

WC

WC

WC

WC

144

145

146

147

148

149

110

150

151

152

153

154

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

155

10

9

8

7

6

5

4

3

2

1

53

52

Cornell

Duluth

Duluth

Trench 2

Area BB, room 2

Area R, room 20 Area R, room 20 Area R, room 20, in pot in north corner Area BB, section BB2, room 21 Area BB, section BB4

(1)-Fragment of Gastropod Trench 2 (L) siphonal canal process. (12)-Teeth and fragments of teeth. (2)-Fragments of mandible/ maxilla, one with a tooth. Vestibule, ashy deposit on stucco floor Quercus sp.; rings=10 Vestibule, SW doorway, burned wood, about 1m below surface Room 12, Area EB Area BB, room 12, S corner

(1)-Fragment of mandible with two teeth. (1)-Herbivore molar. (1)-Fragment of bone. (2)-Incisors.

0

0

0

20

20

20

0

carbon fragments from column carbon fragments

carbonised wood

carbonised wood bits of wood

54.121

54.039

RJB

RJB

0.000

53.149

RJB

MR

0.000

53.131

54.053

RJB wood fragments from S corner petrified wood MR

53.075

54.037

CWB

9/7/54

14/6/54

9/6/54

2/7/53

23/6/53

16/6/54

12/6/54

26/6/53

6/7/63

20/5/58

0.000

0.000

20/5/58

0.000

RJB

burned wood

0

carbon

ashy deposit

0

12

bones and teeth

teeth

0

0

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

WC

WC

WC

WC

168

169

170

WC

163

167

WC

162

WC

WC

161

166

WC

160

WC

WC

159

165

WC

158

WC

WC

157

164

WC

156

111

25

24

23

22

21

20

19

18

17

16

15

14

13

12

11

Cornell

Cornell

Cornell

Coniferus sp.; rings=30 (sampled)

Quercus sp.; rings=10

Area M, stoa 44

Area M, stoa 44, old NE portico Area M, NE stoa 44 Area M, NE stoa 44

Area M, NE stoa, west end

44

44

44

44

44

Area m, room 38 38, from door jamb Area M, room 40 40 Stoa 44 44

Room 27, from 27 pithos in S corner Area M, room 31 53, near wall W Area M, room 32 32, under lid 35

Room 23, SW 23 wall intrusion 23 Room 23, pithos 7, ca. 0.50 below top of bench Corridor 25 25

wood fragment C wood fragment D, carbonised wood fragment G, carbonised

sample of carbonised wood carbonised wood fragments of carbonised wood sample of ashy earth and carbon wood fragments

burned stuff on pavement

bits of woody stuff

carbonised wood carbonised wood

carbon material

55.079

MR

55.077 55.077

55.077

MR MR

MR

55.077

55.112

MR

MR

0.000

55.017

54.175

53.023

54.183

MR

MR

MR

MR

MR

54.106

0.000

RHu

MR

54.014

RJB

16/6/55

16/6/55

16/6/55

16/6/55

16/6/55

29/6/55

4/6/56

24/5/55

20/7/54

1/6/53

27/7/54

30/6/54

22/7/55

1/1/54

seeds (?)

Scientific Analyses - Table 1.6

WC

WC

WC

WC

WC

WC

173

174

175

176

177

178

112

WC

WC

WC

WC

WC

WC

WC

181

182

183

184

185

186

187

WC WC

WC

172

179 180

WC

171

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

42

41

40

39

38

37

36

34 35

33

32

31

30

29

28

27

26

Cornell

Cornell

Quercus sp.; rings=10

Coniferus sp.; rings=20

piece of carbonised wood or bone

0

0

0

0

0

32

0

0 0

0

68

carbonised bark?

carbonised wood burned wood

piece of carbonised wood carbonised wood burned wood

carbonised wood carbonised wood from S wall carbonised wood burned wood carbonised wood burned wood

burned wood

62

MR

MR

MR

MR

MR

MR

MR

MR MR

57.082

54.182

54.182

54.072

0.000

0.000

54.142

54.147 0.056

56.105

0.000

EBW

MR

54.022

54.088

MR

EPB

54.000

55.081

55.033

55.082

MR

sample of MR burned wood carbon w/ pot MR 7 on pavement

carbonised wood

0

47

0

41

92 Area MNE, court 92, mixed with small stones & black earth

Area M

Area M, btwn stones Area M

Area M, room 32, near surface Area M

Room 68, from South wall Area M, NE of wall W Area M Area M, from doorway Area M

Area M, room 45, NE arch, portal 41 Area M, room 47, NW strip Area ME, court 47, old court C Area M, room 56, west corner Area M, room 58, from just above pavement Room 62

18/6/57

26/7/54

27/7/54

10/7/54

14/6/54

9/7/54

9/7/54 11/6/55

30/6/56

15/7/55

15/6/54

23/6/54

30/6/54

21/6/55

30/5/55

16/6/55

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

113

WC

Wc

WC

WC

198

199

200

201

WC

194

WC

WC

193

197

WC

192

WC

WC

191

196

WC

190

WC

WC

189

195

WC

188

56

55

54

53

52

51

50

49

48

47

46

45

44

43

Cornell

Cornell

Cornell

Cornell

Cornell

Cornell

Cornell

Area ME, room 46, old 70, from SW jamb of door

Area M, in drain Area M

Tr I, N room, E of pithos, W wall, 3rd pithos fr S, trngl 2, box 6 Area M

mostly plaster

mostly plaster

Pinus sp.; rings=25 (sampled)

Court (south of portico of megaron) Court (S of portico of megaron) Pylos 52, SW section hearth room, above floor Tr Zb, S edge, 40cm bw surf, 4.5m lth, 4.5m fr S. edge tr. Room w/ hearth, SW corner

Pinus sp.; rings=30 (sampled) (near column base)

mostly plaster

branch rings=10

Quercus sp.; rings=10

6

6

6

0

0

0

0

0

0

0

0

Area M, 25 corridor 25 23 2.35 fr inside NW wall of meg. mid of 1 tr on floor -0.70 Btwn M and N 0

fragments of wood

charred matter

carbonised wood

burned wood, Minoan doubleaxe carbonised wood, 2 bags

carbonised wood carbonised wood fragments of burned wood charcoal

fragments of carbonised wood charcoal

53.148

WAM

GM

GM

GM

MR

MR

MR

MR

52.047

52.075

0.000

0.000

0.000

0.000

0.000

54.000

54.173

54.155

57.075

55.020

RHu

ELB

0.000

MR

7/7/52

4/6/52

11/7/52

30/6/54

26/6/54

20/7/54

13/7/54

23/6/57

7/7/53

30/5/55

27/6/54

Scientific Analyses - Table 1.6

114

WC

WC

211

216

WC

210

WC

WC

209

215

WC

208

WC

WC

207

214

WC

206

WC

WC

205

213

WC

204

WC

WC

203

212

WC

202

71

70

69

68

67

66

65

64

63

62

61

60

59

58

57

Cornell

Cornell

Cornell

Cornell

Cornell

Cornell

Cornell

Cornell

Quercus sp.; rings=25-30 (sampled)

Coniferus sp.; rings=25 (sampled)

Coniferus sp.; rings=27 (sampled)

(black earth); rings=10

Quercus sp.; rings=15

Quercus sp.; rings=10

Coniferus sp.; rings=10

Coniferus sp.; rings=25 (sampled)

6

0

0

0

0

0

0

63

0

0

Trench 1/8, P.NW

0

Area MZ3-11, 0 removing baulk, red burned brick and rubble Area MZ, under 0 room b, below black layer

Scarp

South of portico 0 of megaron

Area D.2, south of portico of megaron

Area P.2, south portico of megaron

Above floor middle of trench Hearth room, from floor of hearth Tr Z, 4th m str, 18cm blw grass, 18th m, 36cm below grass ?? trench 57 o rZ7 ??

Area EB, room 63, section B

North extension 0

Room of hearth

carbonised wood

carbonised wood

plastic bag #6

partly burnt deposit

carbonised wood

Minoan doubleax, charcoal

Minoan doubleax, black earth

illegible

burnt wood

carbonised and charred wood burned matter

charcoal

samples of wood

burned wood

MR

MR

GM

GM

GM

ELB

GM

GM

0.000

58.027

58.159

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

54.000

0.000

0.000

1/7/58

23/5/58

7/7/59

4/6/63

4/6/63

14/7/53

11/7/53

11/7/53

1/1/52

2/6/52

14/6/52

2/6/52

24/6/54

11/6/52

17/6/52

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

Table 1.6 cont. Results of scientific analyses carried out on materials collected by Blegen and MARWP.

AERIAL PHOTOGRAPHS

At the end of the 1997 season, MARWP commissioned Dr Bobby Ionnides of the Athens Polytechnion to take a set of low-altitude photographs of the site by helicopter. These are large-scale, high-resolution, panchromatic photographs that cover both the ridge and the acropolis, though the Main Building of the palace was obscured by its shed roof and parts of the northeast side of the site by trees. A second helicopter run took low-altitude vertical colour negatives with a mounted Haselblad; 35 mm obliques were taken on a third round.

115

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.73 Northeastern section of the aqueduct.

116

Aerial Photographs

1.74 Northwest Area and Southwestern Building, showing wall ZZ in Court 88.

117

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.75 Ramp 91 and the Northeast Building.

118

Aerial Photographs

1.76 Northwest end of the Northeast Building, showing sections B and C of the aqueduct and Cistern 102.

119

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.77 Area 103 in the Northeast Area.

120

Aerial Photographs

1.78 Southwest Quadrant.

121

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.79 Southwest Quadrant, showing Building X.

122

Aerial Photographs

1.80 Southwest Quadrant, with the ‘altar’ to the southwest of Court 63.

123

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.81-1.82 Two views of the Southwest Quadrant.

124

Aerial Photographs

1.83 Composite image of the Southwest Quadrant.

125

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.84 Composite image of the southwest side of the site.

126

Aerial Photographs

1.85 Composite image of the Southwestern Building.

127

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.86 Composite image of the Northwest Area and the Southwestern Building.

128

Aerial Photographs

1.87 Composite image of rooms 89–90 and the Northwest Area.

129

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.88–1.89 Two views of the Northeast Area, showing the Northeast Building and Area 103.

130

Aerial Photographs

1.90 View of the Northeast Building and Southwest Quadrant from above the Main Building. 1.91 High-level image of the palace and the Englianos ridge among the surrounding olive groves.

131

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.92–1.93 Oblique views of the southwest side of the site.

132

Aerial Photographs

1.94–1.95 Oblique views of the southwest side of the site.

133

Cooper – The Minnesota Pylos Project: Investigations and Results, 1990–98

1.96 Oblique view of the Northwest and Southwest areas. 1.97 Oblique view of Area 103 in the Northeast Area.

134

2 HYDRAULIC ENGINEERING ON THE ENGLIANOS RIDGE: EVIDENCE FOR PRE-LH IIIB PALACES frederick a. cooper

the Northwest Area; see pp.55–56 above) and may have included floors; it certainly included painted walls (with fragments surviving from the ‘mosaic-chip’ floor in Area 106; see p.40, and p.84, figs 1.68–1.69). Most useful for its reconstruction is the extensive hydraulic system analysed here, which leaves an imprint of the early palace in the organisation of its water inlets.

The dense tangle of walls in the Northwest Area at Pylos (see above, pp.47–73) and Blegen’s ‘wall maze’ in the Southwest Quadrant (pp. 73–76) embrace a complex system of built drains that also extend around the southeast and northeast sectors of the palace and across the Englianos hilltop. While Blegen appreciated the drainage system and, judging from his draft of The Palace of Nestor, intended to devote attention to the network, in his published version he relegated the hydraulics to short sections on the Main Drain, the so-called aqueduct, and the cistern Room 102 (PN I, 230–35; 329–36). The impressive Bronze Age network of tributary drains in the southeastern area of the palace went into footnotes, and a single map plotted most, but not all, of the network located in this part of the site (PN I, nn. 75–83, fig. 426). The Key Plan in volume I of The Palace of Nestor maps a few additional drains in other parts of the acropolis, but Blegen gave no prominence in his text to the drainage system.

Knossian Parallels Parallels for the following reconstruction of water intakes and outflows may be found in the Domestic Quarter at Knossos, as modelled in a study by Colin MacDonald and Jan Driessen (1988, 235–58). As will be shown, the circuits of piping and conduits at Knossos and at Englianos are kindred, each lying underneath multi-phased architecture, with the effluent discharging beyond building walls. At Knossos, the drains reach the brims of eroded slopes just outside the walls, being destroyed beyond that; at Englianos, all known drainage lines extend from collector intakes that carry precipitation in channels down the slopes to the foot of the acropolis or into impounding reservoirs in the Northwest Area.

A study of these Bronze Age hydraulics produces not merely an appreciation of the engineering expertise involved in Mycenaean water control systems, but also provides useful chronological evidence for the many building sequences that took place at the palace, revealing the presence not only of an Early Mycenaean predecessor to the Late Helladic IIIB palace, but a Middle Helladic complex as well. Superimposed walls uncovered in various 1- to 2-metre deep trenches, especially in the chasm, the Northwest Area, and in soundings made elsewhere during the Blegen excavations, establish multiple phases of construction at Englianos. The abundance of Middle Helladic pottery discovered by Blegen indicates that some of these buildings date from that period, and the evidence for at least three pavement levels beneath the final Late Helladic IIIB floors provide further evidence of earlier buildings. As will be shown, the first building complex (Palace A) probably had a north-south axis, orientated approximately 45 degrees to its successors. This protopalace may have been characterised by the use of ashlars (a reused block shows up in Middle Helladic wall b in

Many short stretches of the drains within the palace were exposed by tunnelling and trenching here and there beneath the palace floors, but evidence for the extent of the network is highly imperfect. The size and poor condition of the Main Drain and its tributaries yield a less complete picture of Pylian hydraulics than that obtained for Knossos, where it is even possible to enter and traverse the interiors of many drains. Nevertheless, although at Englianos a large portion of the waterworks must remain unknown, much can be hypothetically reconstructed based on in situ evidence there and on comparisons with contemporary Bronze Age drainage systems on Crete. In his article re-examining the question of a Minoan thalassocracy, Malcolm Wiener (1990, 128–60) 135

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces

2.1. Schmatic plan of the drains beneath the Palace of Nestor, showing direction of water f low.

emphasised the important evidence of ‘drainage systems showing a high degree of engineering skill and hydraulic sophistication [that] are of course a feature of major Minoan sites in the Proto- and Neo-palatial periods.’ Wiener goes on: ‘By way of contrast, Mycenaean sites even in the Late Bronze Age do not utilise such systems’ (ibid., 140). The palace at Englianos, however, is a notable exception to general Mycenaean practice when it comes to environment-irrigation engineering.1 The installation of impounding reservoirs, the extensive and complicated network of the water collection system, and the multidirectional distribution of collected rainwater runoff, as seen in a schematic composite in figure 2.1, suggest that this is a unique mainland example of effective groundwater management and, as noted by Nelson in Part II, pp.351–52,

not the only element of building at Englianos that has reflections on Crete.

1

This study began with the observation that in Blegen’s partial plan of the drains clustered at the southeastern section of the Late Helladic IIIB palace (PN I, fig. 426), the most striking elements are the simplicity of the apparent maze of random channels and, significantly, the similarity of that maze to the circuit of channels beneath

Of significance is the fact that water diversions take long runs in opposing directions to selected positions around the acropolis, rather than simple, short-run extensions to the nearest point on the acropolis slope. The Middle Helladic drain c, for example, probably tacked cross-ways along the slope rather than taking the shortest route straight down, which would be expected were this simply a drain to carry off storm water. It would appear that among other uses, the hydraulic system at Englianos was created to collect water systematically and continuously in reservoirs, possibly for irrigation of agricultural terraces around the periphery of the acropolis.2

Drainage systems have been found elsewhere in the Mycenaean world – at Thebes, Gla, Tiryns and other sites – but these are not as extensive as at Englianos. A hydraulics system possibly on the scale of that discussed here may exist at the site of Iklaina (https://iklaina.wordpress.com): extensive built drains and a system of pressure piping have been found, with associations still to be established.

136

Pylian Drains as Evidence for Early Floor Levels

2.2 Plan of the drainage system in the Domestic Quarter at Knossos (Evans 1921, 226, fig. 171a).

the Domestic Quarter at Knossos (fig. 2.2). Again, the following autopsy of the hydraulic system at Englianos, and proposals for what it could tell of the earliest palace plan there, owes much to the work of Macdonald and Driessen at Knossos (1988). Their original purpose was an epigraphical survey of the masons’ marks on blocks used in the construction of the drain (ibid., 254), which led to a comprehensive study and description detailing what Evans’ romantic account does not. Macdonald and Driessen are not responsible, of course, for my interpretations of three of their points in respect to the Englianos waterworks: 1) that drains were built first immediately below pavement level and were meant to serve the original plan of rooms and areas; 2) that the drainage enceinte for the most part skirts the perimeter walls and had feeders leading from multiple light wells; and 3) that when room arrangements were later reconfigured, the plan of the drainage system remained essentially the same.

Pylian Drains as Evidence for Early Floor Levels A schematic, northwest-to-southeast cross-section of the palace at Englianos, showing top and bottom elevations at the intercepts of feeder drains and intakes, was prepared in order to compare the palace drainage network (fig. 2.3) with that at Knossos. This revealed a number of features

2

Another piece of evidence in favour of a regulated irrigation system at Englianos in its late stages is the bifurcated and gated faucet beneath the Southeast Gateway. Th is simple valve was part of the fi rst, Late Helladic I gateway, and continued to function when the staircase was plastered over with a ramp in its second phase (see p.75, fig. 1.49). Such a gated valve allows for a diversion of water from one field or terrace to another as needed.

3

Evans put the highpoint of the Domestic Quarter loop at the northwest extremity. Diagonally opposite, there are two lines – the East Exit and the North Tributary – that open to the outside; Macdonald and Driessen follow this lead. It is not clear, however, how a flow of water coming from the north intake can go anywhere except out the East Exit, and certainly not to a higher position in the circuit. Neither Evans nor Macdonald and Driessen use relative or absolute elevations, but taking as correct their identification of high and low points, I must assume that the northward branch, as the eastward one, was a bifurcated outtake for distribution of effluent to different quadrants outside the building in a design like that at Englianos. A widely held and erroneous assumption is that the Knossian water drainage system and others like it, such as the

Both the Englianos and Knossos systems have some straight lines and angular conjunctions, but both are also curvilinear – ‘parabolic’, as Evans (Evans 1930, 240–46) characterised the design. At Knossos, the water main was laid out in a loop with short feeder or influx channels stemming from light wells. One of these feeder channels is a shaft dropping from the southeast corner of the central courtyard, dating to Middle Minoan IIA. In the next two structural and chronological phases (Middle Minoan IIIA and Late Minoan I) the configuration of rooms changes; some rooms no longer function as lightwells, but the drainage feeder system and loop remain basically the same. Later modifications can be detected both in the Domestic Quarter at Knossos and at Englianos, and that section of the Cretan palace provides a guide to reconstructions of the earliest Pylian plans.3

one I propose at Englianos, functioned as a water-based sanitation or sewage system, which would have required flushing in order to function hygienically (see Macdonald and Driessen 1988, 248 n. 5, and PN I, 11 and 220). Macdonald and Driessen are unnecessarily circumspect but follow Walter Graham’s lead (Graham 1962, 220–21) in arguing that the role of lightwells, beyond the aesthetic, was part of an engineered watercollection system. This, too, is my position.

137

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces

2.3 Northwest-to-southeast cross-section showing top and bottom elevations at the intercepts of feeder drains and intakes.

relevant to building phases prior to the Late Helladic IIIB palace at Englianos.4

level with the top or floor level of the Main Drain and the top of the socle of the orthostate wall A beneath Room 7, at 191.16 m (see pp.78–79 above, and Part II, pp.314–15; for the significance of these two features, see also below, pp.145–51, Palace B.) No physical evidence in the area of the Southwestern Building exists for an inlet to drain d, but the drain itself continued in use from the Middle Bronze Age to at least Late Helladic I – that is, during the lifetimes of palaces A and B. Third, the short stretch of drain c in the Northwest Area, when extended in the same way, would have headed due east and climbed to a direct elevation of 193.00 m at the drain in trench MZ40 (discussed below, pp.140–43).

Pierced capstones, 0.35 m thick (fig. 2.4), were placed above improvised shafts dropping into the various drains at inlet holes that penetrated the floors of the Late Helladic IIIB palace. The inlets for which there is evidence today are in the Main Drain in Court 58 at Room 7 and drain 1 at the outside jog between rooms 8 and 9; drain 8 in Court 63; drain 2 in Court 3; and drain 5 in Room 10. The latest floor levels lie at one or another of stepped levels ranging from 0.15 to 1.0 m above the capping stones to the various drains. The top stone covers of the original drains probably served as part of earlier floor levels, as will be discussed below in respect to the Main Drain off Room 7 (PN I, figs 16–17). Evidence for these floor levels comes from drains in the Northwest and Northeast areas.

blegen’s aqueduct and the pressure pipe in court 42 The fourth piece of evidence for early floor levels based on the presence of drains in the palace comes from branch B at the southwest end of Blegen’s aqueduct, which terminates in the northeast wall of Court 42 and is 0.40 m above the latest floor. Suspended in mid-air, so to speak, it does not appear to head anywhere obvious. The terracotta piping is designed to socket together, sealing the joint to make a pressure pipe, with water heading outwards from the courtyard to the northeast (fig. 2.5). This pipe is crucial evidence for the fact that water flowed away from the building and not towards it, as Blegen originally argued (PN I, 332–35). This type of drain pipe is not common in Bronze Age Greece so far as I am aware, and this is its only occurrence at Pylos; but it is found in Middle Minoan drainage systems at Knossos (e.g. Evans 1921, 143, fig. 104, and Evans 1930, 253, fig. 173) and elsewhere on Minoan Crete.5

First, drain e in the Northwest Area, when extended to an assumed inlet within the later Room 23, yields an interpolated elevation of 191.20 m there. Second, if the short segment of the Middle Helladic drain d is extended along a continuous upward gradient and along the same line, the drain would reach a similar elevation of 191.25 m at a point near the early walls underlying Room 65 of the Southwestern Building. This would bring drain d nearly 4

The tributary drains outside and underneath rooms 7–10 were not reentered in the MARWP excavations. Lord William Taylour conducted the original excavation of this area in 1955; his field notes are richly descriptive, and he took a number of photographs. Blegen published nearly verbatim Taylour’s text and dimensions, along with a selection of his photographs (PN I, 232–34 nn. 75–82, figs 181–87). Taylour did not take levels to tops and bottoms of the drains, so in order to visualize how the drains related one to the other, it was necessary to interpolate his dimensions to arrive at absolute elevations.

In Area 101 at the northeast side of the palace, a confusion of terracotta pipes and troughs and stone-cut and stonelined conduits is all that is left of a number of building 138

The Middle Helladic Palace A phases. The upper elevation of the terracotta pressure pipe in the northeast wall of Court 42 stands at 193.18 m and the bottom level at 193.01 m, all but matching the level, some 20 metres to the north (192.97 m), of the bluestriped painted floor in Area 106, dated to the first phase of building in the Northeast Area (see pp.35–36 above). The east–west orientation of the surviving stripes on the plaster floor, and the same orientation of the pair of column bases fronting that floor, help to fi x the axis for this early palace; the floor level for the purposes of this discussion was c.193.2 m. The bottom flow level of the terracotta pipe (193.01 m) is maintained, as far as can be deduced from the interpolation of spot elevations described next. The highest and lowest elevations of the drains in the northeast part of the site, and the conical terracotta pressure pipe in the northeast wall of Court 42, show that water flowed away from the palace here; this means that water collection necessarily occurred somewhere within the body of the palace. As already noted, the conical fitted terracotta pipe in the northeast wall of Court 42 is the only one of its kind at Englianos; all other drains at the site are either stone-built or bracket-shaped terracotta (see Part II, pp.367–69). A pressure pipe of this sort can function as part of a siphon in a hydrostatic design, though it need not necessarily do so. Flow of water from a higher level can drop below the level of the intake and loop back upwards to a spigot just slightly lower than the intake (Wilson 2008, 293–96, fig. 11.2). When installed, the pressure pipe in Court 42 may have functioned in this way as a siphon, one of the many elements of the north–south hydrostatic water system of Phase 1 in the Northeast Area (pp.35–36 above). If so, the pipe must have brought water from a tank somewhat above 193.18 m (the present top of the pressure pipe) to a point along branch B at c.193.0 m. In any case, the pressure pipe would have extended along its axis, probably originating somewhere within an area later occupied by the Late Helladic IIIB Portico 41 or Vestibule 38. At a later date, the area of Court 42 was excavated to a depth of c.0.80 m below the c.193.18-m top of the pressure pipe.

2.4 Pierced inlet to the Main Drain, south of Room 7.

2.5 Visible end of pressure pipe in northeast wall of Court 42.

At present, the terracotta pipe pressure system poses an enigma, for although the evidence of the tapered and fitted pipe in Court 42 shows that water flowed in the opposite direction to what Blegen supposed, from inside the palace to outside the walls, the fact remains that there is no surviving evidence for either an intake or an outflow basin in the form of a collection tank or cistern. Nevertheless, the conical pipe was not simply cut off by the intruding path of the later enclosure wall of Court 42, but was carefully incorporated within it, implying the presence of some now missing functional element.

That the pressure pipe continued to operate into the last building phases of the Late Helladic IIIB palace, as did the system as a whole, is suggested by the careful encasing of the pipe in a protective surround of flat stones within the northeast wall of Court 42, the construction of which dates to the last phase of the LH IIIB palace. In Phase 7 of the Northeast Area (see pp.40–41 above), when branch C was installed, a lower reach of the siphon may have run underneath the floor of Court 42, although any evidence for this is now covered over by the flooring in this area.

The Middle Helladic Palace A The staggering of elevations of drain intakes suggests that there were two distinct levels of drains in a single system of late Middle Helladic date at Englianos. The positions of drain tops and, therefore, of the accompanying floor levels that would have been laid directly above the drains are arranged split-level, with a rise of about a metre going from south to north, requiring about five steps from one level to the other. The lowest level has a hypothetical inlet at the end of drain 4, at 190.80 m; the next floor level corresponds to c.191.3 m: the top levels of the Main Drain at the southeast of the palace and the earliest inlets to drains 6, 1 and 3, as stepped downwards north to south, are 191.40, 191.16 and 191.00 m respectively. Drain e in

5

Examples of Minoan tapered terracotta pipe are described further by Joseph Shaw (1973, 198–201, figs. 231–32; Appendix J, 235). Another terracotta pipe, from Middle Minoan I–II Ialysos, is published by Toula Marketou (1988, pp.27–28, fig. 3). Besides the plain type of conical terracotta pipe at Englianos, examples of a handled variation have been found at Knossos (Evans 1921, 142–43, figs 103–04).

139

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces field notebook plots, the constructions found in MZ40 nevertheless appear on the annually updated site plan and were published with the 1958 report (Blegen and Lang 1959, pl.25, fig. 1) and on a schematic of Area Z and MZ published in The Palace of Nestor III (fig. 309). As was the practice during the Pylos excavations of the 1950s, plans and scarp profi les were infrequent, and measurements for levels were taken downwards from assumed zero benchmarks at the surface edge of the trench.6 As Blegen’s published plan shows, MZ40 was only partially excavated, leaving the constructions – walls and a drain – to disappear into the surrounding ground. In his draft, Blegen writes that he abandoned an expansion of MZ40 because at the time ‘it did not seem practical’. The same held true for MZ36, excavated on either side of wall Z18. Our knowledge of the features in both trenches depend on brief descriptions in Rawson’s notebook and synopses of these by Blegen in his working draft; this record is much too incomplete for a firm reconstruction. In most cases, Rawson’s measured depths to walls have been converted to UTM geodetic heights above sea level (see p.31). In the course of our topographic survey, we took a ground-level reading at the north, top corner of trench of MZ40: 193.38 m. Lengths, widths, heights, locations and other details, when provided, are taken from Rawson’s two notebooks for 1958 as well as Blegen’s summations of her records in his working draft.

the Northwest Area is at 191.20 m, as noted above. The uppermost floor level stands at c.193.2 m – the level of the blue-striped floor in Area 106, the top of the pressure pipe, and the calculated upper reach of drain c. Abstracting from Taylor’s 1955 field report (Blegen n.d., 295–97), Blegen (PN I, 232–33) summarises that the finger drains – 1, and 3 through 8 – that tap into the tributary drain 2, which in turn empties into the Main Drain, show signs that some are earlier and some later. Specifically, the stone vaulting to drain 4 supports capping stones for the Main Drain. Drains 1, 3 and 6 predate drain 2. That portion of the Main Drain that lies between drains 4 and 6 also must be part of this earlier system, belonging to Palace A. Drains 3, 4 and 6 went out of use when they were covered over by the construction of Palace B (see below, pp.144–47). Drain 1, by contrast, continued in use during the time of Palace B and through the life of the Late Helladic IIIB palace. A top elevation of 190.80 m for drain 4 puts it 2.5 m below the drains in the Northeast Area, discussed in detail below. A short (c.5- to 6-m) section of the collection drain, Main Drain A, running in a westerly and downward-sloping direction and served as a collector for drains 1, 3, 4 and 6, also coincides with this top elevation of 190.80 m. A probable location for the inlet to drain 4 is at or somewhat beyond its preserved north end and allows for a reconstruction of a court, perhaps an exterior and irregular court with a south edge following the 191-metre contour. (Here and elsewhere I refer to topographic contours in only the most general terms, as possible evidence for courts and other areas relating to the body of the palace under discussion.) I give in figure 2.7 a reconstructed, hypothetical plan in which a staircase connects two floor levels rising from south to north, at 190.80 m to 191.30 m. The uppermost level of Palace A occupies the Northeast Area, at 193.20 m; direct access there is through an entrance on the east flank. The locations of the first-phase drains and assorted architectural features delineate the exterior and interior lines of walls. I indicate in outline other features such as light wells, doorways and stairs.

6

Regrettably, trenches MZ40 and MZ36 escaped attention at the time of MARWP’s 1990s investigations at Englianos.

trenches mz40 and mz36 An all but forgotten 1958 trial trench, MZ40 (1.5 x 7.4 m north–south) and a sondage, MZ36 (between the northeast end wall of the Wine Magazine and wall Z18 at wall Z17; fig. 2.6), supplement significantly what can be said about the hydraulic system at Englianos, and provide a further definition of the footprint for Palace A. In 1958, David French completed the excavation of the Wine Magazine begun earlier by Marion Rawson, with the fields in the northeast area designated as Z and MZ. Rawson continued to excavate these sectors around the Wine Magazine, including the trial trench MZ40. She did not map trenches beyond wall Z3, including trench MZ40, nor did she sketch features of any constructions underlying those at the top levels. While there are no

2.6 Trenches MZ40 and MZ36, north of the Wine Magazine. Other features shown in black are Phase 1 features (see pp.35–36).

140

The Middle Helladic Palace A

2.7 Proposed Middle Helladic Palace A, showing drains, central court, staircases to higher levels, possible lightwells along the southwest side, and positions of column bases in Area 103.

141

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces Rawson notes in passing (Rawson 1958, 45) a 0.15-m square hole for a drain that penetrates by 0.20 m the northeast end wall of the Wine Magazine. She gives a dimension of 3.17 m northwest from the ‘east corner’ – presumably the east corner of the Wine Magazine. In plan, this places the drain within the Wine Magazine wall just to the northwest and alongside the 1.22-m wall MZ36 in trench MZ36. The bottom of the drain, at 192.55 m, must approach within centimetres the bottom edge of the base to the wall of the Wine Magazine, c.192.50 m. The location of this drain hole puts it directly on the line of azimuth, 67 degrees, and at the computed slope of drain c.7 Extending the azimuth and slope farther to the northeast, the drain arrives at the section of drain in trench MZ40. The opening to the Wine Magazine drain is smaller than the cross-section of the channel elsewhere in drain c, possibly because the end wall of the Wine Magazine was built around and above the pre-existing drain.

Two distinct features show up in MZ40: a stone-lined drain mz40, that runs diagonally across the width of the trench in a more or less east-to-west direction, and two well-built walls joining at right angles in the northwest end of the trench. The drain channel is 0.45 m deep in four courses and 0.15–0.25 m wide; the overall width is 0.85 m. In this respect, the drain is similar to other stone-sided drains across the site – in particular, drains c, d and e in the Middle Helladic and lowermost levels of the Northwest Area. The top elevation of this drain is close to 192.90 m. Of the right-angled walls, the southeast wall, Z1, was exposed to a width of 0.60 m, while surface width of the southwest wall, Z2, was partially dug to 0.50 m. Unfortunately, the length of neither wall was determined. Evidently the bedding course for the corner of walls Z1 and Z2 is irregular, so that the wall has a variable height of 0.78–0.90 m in up to six courses. A pithos, open in crosssection through the body, is embedded within the crux at the wall return. Rawson did not dig the vessel to its bottom. She observed that wall Z1 ran roughly parallel but shifted from wall Z17, distant some 5.20 m to the southwest. Rawson’s observation helps to situate an important wall in MZ36. In this trench is a wall with a corner underlying walls Z17 and Z18. The deep wall has the same alignment as Z17 and therefore of Z1.

Another sounding within MZ36, between walls Z18 and Z3, reached a level of c.192.20 m – not ‘stereo’, but a point where the available space made it difficult to dig further. This deep level is another indication of the probable Middle Helladic stratum at Pylos. Trench MZ36 yielded a quantity of reddish mudbrick. In both MZ36 and MZ40, Rawson lotted a quantity of potsherds, keeping some but discarding most. Painted ware seemed to have predominated: two rim profiles were labelled LH I, along with a piece of ‘oatmeal ware’ and ‘gray ware’. In his draft, Blegen remarks that three sherds near the bottom of the excavation next to wall Z1 appear ‘earlier than the other painted shards’. He does not specify which ‘other painted shards’, but presumably he refers to those listed by Rawson in her summary of the pottery. A deposit reaching 0.95 m below the drain in MZ40 at its southwest side also included a quantity of ‘early, painted shards’.

The constructions in MZ36 and MZ40 probably belong to a single building program. Rawson reports that the walls and drain lay immediately underneath the plough zone, 0.25 m deep, for an interpolated elevation of 193.13 m. She also notes that the orientation of the walls in trench MZ40 aligned with wall Z17 in trench MZ36 nearby the Wine Magazine. That is to say, whatever other purpose this corner construction in MZ40 turns out to have, its orientation and top elevation of 193.07 m conform to the reconstruction azimuth, 67 degrees, of Palace A, as derived additionally from a number of factors presented here. In a sounding within trench MZ36, labelled ‘Pit’, Rawson encountered two separate and different features: like those in MZ40, a wall corner and a drain. Again, these add significantly to a reconstruction of Palace A. An early wall built of ‘largish stones, regularly laid’, 1.22 m long, runs southwest to northeast, with the southwest end close by at the northeast end wall of the Wine Magazine. At the northeast end, a corner with a stub of wall returns to the southeast. The calculated top to this wall MZ36, at 192.42 m, underlies the bottoms of walls Z17 at 192.50 m and Z18 at 192.70 m. The top of the wall remnant MZ36 accords almost precisely with the top elevation of the Middle Helladic ‘massive wall’ under the floor of the vestibule to the Wine Magazine. The estimated bottom elevation of the Wine Magazine wall is 191.75 m, and that, in turn, is above the bottom of the MZ36 wall, 191.56 m. This is to be expected: a section of the same wall, much further along the facade of Palace A and within the Northwest Area, has its base much lower in level, conforming to the topography of the site.

palace a footprint Having used the drains and their inlets to create an initial footprint of our Middle Helladic Palace A, we may now add several more elements – in situ and hypothetical – to build up a picture of the building, working clockwise from the north. The massive east–west orientated wall uncovered in trench MZ36 (top elevation 192.43 m), that in Wine Magazine Room 104 (top elevation 192.45 m) and wall b at the foot of wall B in the Northwest Area (top elevation 189.44 m; (see pp.55–56) – three of the earliest walls on the site – together are excellent candidates for the base footer of the north facade wall of Palace A; a fourth wall, in Corridor 26, is also part of the palace footprint and is discussed below. As already noted, an eastward directional and gradient projection of drain c places 7

The Toreador fragment, dated to the earliest phase of Pylian frescoes (Immerwahr 1990, p.196), came from a ‘pit’ beneath the floor of the Wine Magazine 105 (PN II, 49–50, 36H105). The origin was probably this drain.

142

The Middle Helladic Palace A and drawings. The MARWP plan and elevations were drawn independently of the earlier ones by Rawson, but the two plans and two southeast elevations correspond in most respects; Rawson further labelled a variety of features such as pithoi, rough stones and ‘angonaria’ (ashlars). MARWP documentation adds the northeast elevation that Rawson did not draw, the 0.50-m wide exposed vertical face of the Middle Helladic wall. This multi-period matrix of walls deserves far more study and analysis than afforded either by Blegen’s expedition or by ours.

it immediately alongside this north facade; the drain is assumed here to terminate at a hypothetical inlet at 193.07 m, close to the partially excavated drain in trench MZ40, at the same elevation as the painted striped floor in Area 106 (pp.35–36). The east–west orientation of the blue stripes on the plaster floor in Area 106 and the parallel alignment of the pair of column bases reaffirm the axis of Palace A. The three exposed walls listed above fi x the line of the north facade and the northeast corner. The south and west lines of interior walls probably were marked by the shift of 1.7 m in floor levels, as determined by elevations in the hydraulic system – in particular, the assumed elevation of 191.20 m for an inlet to drain e. Looking at floor levels, the central portion, i.e. the inlet to drain e, is lower by 1.7 m than that of the Northeast Quarter at 193.2 m. An area at the southwest maintains the elevation at 191.3 m. A terrace with a sweeping vista to the north surrounds the northeast corner of Palace A; the porticoed entrance that was present in Phase I of the Northeast Area opens on to it. Seven metres to the east lie the assumed inlet for the Middle Helladic drain c and, close by, the enigmatic wall corner Z1/Z2.

Blegen assumes that the Middle Helladic wall ran southeast to northwest, coincident with the Late Helladic IIIB palace (ibid.). The exposed portion of wall is quite narrow, and available photographs do not expose it to a depth for any sort of assessment of direction. However, stones from its upper course are visible, and Rawson labels them H1 and H2. Top elevations are 191.94 m and 191.93 m respectively. They are visible in figure 2.8. Despite the irregular outline of their shapes, it is clear that they are built in a direction oblique to that of the axis of Corridor 26; in fact, their azimuth runs north–south, within 10 degrees of the axis of the restored Palace A. This Middle Helladic wall is therefore perpendicular to the short stretches of exposed Middle Helladic walls in the Northwest Area assumed to represent the north facade of Palace A. This Corridor 26 Middle Helladic wall thus probably served as a retaining wall for the downward shift in levels between the highest, at 193.2 m in the Northeast Quarter, and the lowest, at 191.3 m in the assumed Central Court.

Another snippet of a major construction embedded in Middle Bronze strata runs more or less north to south underneath the later Corridor 26. In three separate passages (PN I, 145, 146 and PN III, 32–33) Blegen describes this wall from notes recorded by Rawson in 1960. Rawson’s trench was taken to an overall depth of 0.40 m, where she encountered an earlier floor. A sounding of approximately 1.0 m by 0.50 m was sunk against the southwest wall of Corridor 26. This pit went to ‘stereo’ at a depth of 2.34 m, or 190.08 m above sea level. A 0.50-m wide strip of one face of a wall was exposed at its intercept with a later wall reused as a foundation wall for the Late Helladic IIIB southeast wall of Corridor 26 (discussed below in reference to Palace B). Rawson’s field notes include a plan and two elevations: one of the northwest flank wall and the other of the southeast. She did not draw the exposed vertical strip of Middle Helladic wall. These drawings were not published; instead, Blegen summarises the various features in the sounding, noting that the massive wall was about 1.25 m wide (Rawson gives 1.35 m) and 1.85 m in height, lodged on ‘stereo’ and embedded in stratified fi ll (PN III, 33):

I hypothesise another entrance to Palace A at the end of a path leading from the Northeast Gateway. This entranceway, centred along the east facade, finds some confirmation in the find spot of the horns of consecration in Area 103. Large stone horns of consecration (PN I, 328, figs 238–39) are an exceptional find for a Mycenaean site. As noted above (p.36), a tip was reused here in a scrappy wall, and another larger, lower portion was found nearby at the end of branch C of the aqueduct in this area; Blegen described this piece as ‘lying beside the water channel’. If this was its original position, the horns of consecration stood alongside what may have been a processional way leading from the Northeast Entrance. The Minoan ‘horns of consecration’ emblem takes a prominent role at Englianos insofar as horns of consecration appear not only as a large stuccoed stone, but also in two different wall painting fragments – 8A3 and 9A24 from Court 3 and Room 24 (PN II, 139–40, pls 76, I, and R)8 – and on a carnelian seal from an unpublished

In the earth that had accumulated against the southwestern face of the wall, a considerable amount of pottery was recovered. It seemed to lie in stratified order, with a very few Late Mycenaean sherds just under the floor of the corridor, a good many Early Mycenaean fragments, Late Helladic I in the next lower level, and in the greatest depths almost exclusively fragments of Middle Helladic wares, comprising Matt-painted, coarse incised, Argive and Gray Minyan types (Fig. 135).

8

The larger of the two Pylian fresco fragments is remarkably similar to a Knossos counterpart, the representation of horns of consecration parapet that crowns the tripartite shrine in the miniature Grandstand fresco (PM III, 46, pl.XVI). The scale of the Knossian horns is about 80 percent of the Pylian one; the two are so close, not in specific detail but in style, composition and symbolism, that it is easy to image an actual influence one to the other. The usual assumption is that artistic and religious influence flowed from Crete to the mainland, but as a general

In 1996 we reopened Rawson’s trench as a part of our routine coverage for the preparation of actual state plans 143

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces To access the lower level of the proto-palace at Englianos, there are any number of possibilities. Next to a possible basin or light well supplying water to branch B of the aqueduct, I place a landing and a ramp heading from the East Court’s exterior ground level of c.193.2 m downwards to a Central Court at 191.3 m. An inlet for drain e would have taken precipitation to the north and underneath chambers along this side. Drain d would have moved water to the north from hypothetical lightwells situated within a bank of chambers facing the court along its west side.

tholos tomb at Rutsi near the site (Marinatos 1960, 173, fig. 208). Examples of horns of consecration begin to appear in the Old Palace period, Middle Minoan II, at Knossos and continue into the Post-palatial period, Late Minoan III. They appear in various media – stone and terracotta pieces in architectural space, on house models, in headdresses of Late Minoan figurines, on wall paintings, sarcophagi and seals (D’Agata 1992, 247–50) – and there is enough in the way of symbolic context to merit a Panofskianlike iconographic analysis. At the very least, horns of consecration signify a consecrated space: a sprinkling of symbols occur in peak sanctuaries throughout Crete (Kyriakidis 2005, 93–95, 164–65). Knossos is where the most prominent examples are found; the most wellknown monumental example, some two metres in height, stood somewhere above the Stepped Portico at the Northwest Entrance to the palace. Other large stone examples fell from above to the pavement in the Central Court near the Throne Room, another was found in the East Hall of the Domestic Quarter, and still another had fallen from somewhere above the southwest corner of the palace (Evans 1921, 159–62, fig. 81; Evans 1930, 524–25, fig. 367). Something like this spatial setting must pertain to the find spot for the horns of consecration at Englianos, which perhaps fell from a parapet above the entranceways of palaces A and B (after Late Helladic I) or were mounted just outside the entrance at ground level.

The plan at the Southwest Quadrant is complicated by the crowding of tributary drains. As noted above, inlets – both actual and supposed – to drains 1, 3, 6 and 4 step downwards from north to south at elevations of 191.40, 191.16, 191.00 and 190.80 m respectively; in other words, at intervals of staircase risers of approximately 0.20 m in height. Moreover, the four inlets fall along a north–south line that intersects the south facade midway. The drains diverge to the west side to take even gradients rather than the stepping of the staircase. This situation convincingly argues for a monumental entranceway centred along the south facade of Palace A. The lowermost and hypothetical inlet of drain 4 (191.80 m) falls within an irregular court shaped by the 191-metre contour on one side and the south facade on the other.

The Late Helladic I–II Palace B rule, copies are reductions of originals. Lang (PN II, 140) attributes ‘the excellent condition of the piece(s)’ at Pylos and their ‘uniqueness’ to a ‘kinship with what seems to be earlier material found outside the palace’.

The basic configuration of the first phase of the hydraulic system was maintained in its expansion for the palace that succeeded the Middle Helladic one. There are several

2.8 Marion Rawson’s sounding in Corridor 26, looking north; the Middle Helladic wall can be seen lying diagonally at the centre top of the photograph; the two f lat stones visible here are those labelled by Rawson H1 and H2 .

144

The Late Helladic I–II Palace B I–II (see pp.78–79 above, and Part II, pp.316–17), at 191.16 m. As noted already, the base of the orthostate wall was apparently built to accommodate the top stones of the Main Drain. The paving atop the Main Drain establishes a height of 191.16 m for the external courtyard along the south facade of Palace A. Assuming that the edge of the slope ran along the 191-metre contour at this period, this line fi xes the southern extent of the south court, normalised at an elevation of 191.30 m. Sloping downwards from the northwest, the Main Drain passes underneath the court with inlets midway, at the so-called ‘rostrum’ or ‘loggia’ outside Room 53, and, at the southwest end, at its juncture with the existing Main Drain A. Its distant lower preserved end, where it passes beyond the Southeast Gateway, dates to the Late Helladic I period (see p.145). The elevations of drains a and b of the Northwest Area, probably dating to Late Helladic I (see pp.59–60), are all but identical – 192.10 m and 192.06 m – and their inlets, which have not been recovered underneath the later pavements, presumably lie within the later rooms 23 and 32 along the same approximate southwest–northeast line. These two areas are assumed here to have once contained lightwells in which rainwater was collected.

compelling reasons to think that the plan for Palace B was rotated to the northwest c.23 degrees (from N23W to N45W) rather than aligning with its predecessor (fig. 2.9). First, the northeast–southwest orientated arm of a loop to the path of drain b lies to the southeast of what is assumed here to be the north wall of Palace A and takes an alignment that is also seen in the axis of cistern Room 102 (discussed below) and that of the Late Helladic I–II orthostate wall A underneath and in front of the later Room 7 (pp.78–79). The orthostate wall also accommodates the direction of the Main Drain and re-uses the pavement of the Palace A courtyard. Construction of the Main Drain divides into two periods: section A and its extensions, principally Main Drain B. The Main Drain functioned over centuries, from Middle Helladic through Late Helladic IIIB, making it very likely that repairs and alterations occurred during this time. As already noted, the earlier section A runs in a westerly and downward-sloping direction and served as a collector for Palace A drains 1, 3, 4 and 6. The Main Drain then elbows at either end of this section. At the west it abruptly turns downhill; at the east, Main Drain B turns to the northeast to run approximately parallel with the overall orientation of Palace B and that of its final replacement, Blegen’s Late Helladic IIIB building. The top (the level of the coping stones) of the Main Drain to the southeast of the palace is level with or just a bit below the top of the socle of the ashlar orthostate wall A in Room 7, dated to Late Helladic

Room 102 was built at this time (pp.36–37, fig. 1.8). For good reason Blegen identified this room as a cistern: the plaster floor and base panels are evidently made of waterproof cement, and the unually small size of the chamber – 2.0 x 2.3 m – seems undeniably ill-fitted for

2.9 Superimposition of Palaces A and B on the plan of the later LH IIIB palace.

145

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces

2.10 Proposed Late Helladic I–II Palace B, showing drains, central court, magazines in the later rooms 55–57, and outer courts.

146

The Late Helladic I–II Palace B living space.9 The southwest wall of the chamber was later incorporated into the exterior wall of Court 42, just northwest of the terracotta pressure pipe (branch B of the aqueduct); the southeast wall of the cistern respects the line of this branch. Portions of the rim to the cistern are preserved at 193.48 m, approximately 0.47 m above the pressure pipe. Blegen could not reconcile how his presumed southwesterly inflow through this pipe could rise above the rim of the cistern and so suggested an earlier and lower-rimmed cistern in the same area (PN I, 328–32); the problem disappears with a northeasterly flow, in so far as the overflow in Cistern 102 spills into the aqueduct and not vice-versa. However, the functional relationship between branch B and Cistern 102 must remain unclear, due in part to disturbance in this spot.

sunk c.0.80 m below the original ground level to around 193.00 m to make an entrance court that fronted a new entranceway located at or near the later Portico 41. This earthwork would have brought the ground level in front of the new facade of Palace B even with the socle of the new southeast orthostate wall below Room 7, not only leaving the Northeast Quarter intact but also allowing its expansion. palace b footprint The arrangement of spaces for the new Palace B reused some of the areas of its predecessor: part of the Northeast Area at 193.20 m, and a slice of the Southwest Quadrant at 191.30 m. The rest was altered not only by a re-orientation of the major axis but also by raising the Central Court by 0.90 m to 192.20 m. The northwest limits of the new Central Court (25 x 30 m) are fi xed by endpoints to drain b (192.06 m) and drain a (c.192.10 m). At the south corner of this open courtyard is the preserved inlet to drain 5, at 192.00 m. Drain 2 (192.10 m) and drain 8 (192.05 m) I locate within hypothetical lightwells situated within a body of rooms enveloping the Central Court. The terracotta pressure pipe and its collecting drain lie at the southeast corner of this open area. This set of drains continues to function for the Late Helladic IIIB palace but is later raised to meet rebuilt inlets at higher floor levels. The inlet to drain 1, at c.191.40 m, remained in place, and the hypothetical inlet to drain 7 (191.38 m) was extended upwards.

The axis of these multiple cisterns follows the reorientation of the palace footprint when the north–south orientated proto-palace was replaced by its Late Helladic successor. Blegen reports without further comment that the floor of the cistern slopes to the south and northwest; any evidence for outlets is obscured by the rubble along this bottom edge. Taking the physical evidence one step further, the floor, sloping down from elevation 192.66 m at the eastern corner, indicates an outlet or outlets at the south corner. If an outlet here is indeed the case, it means collected rainwater debouched not only to the northeast along the aqueduct, but also towards the southwest, emptying into drain 42/47 and eventually into the Main Drain.

Outside this hypothetical Central Court and flanking rooms, three areas at demonstrably distinct levels are here proposed: one to the south (Southeast Quad), one to the west (Southwest Quad) and a third, the South Court, following the 191-metre contour of the slope, a re-use of a portion of Palace A’s South Terrace. Along the southwest flank, drain d functioned into the Late Helladic I period (see above pp.50–51, table 1.1, no. 19). A set of spur walls attached to a larger southeast–northwest wall underlies the later Court 63. The southwesterly portions of this line of early walls were ripped out along with the later, overlying plaster floor for installation of the square construction presumed to be an altar of the post-Bronze Age period (see pp.43–44 and fig. 1.20). The 1.5- to 2.0-metre interval between the stubs of preserved walls precludes rooms and instead suggests an array of magazines, which I associate with Palace B.

The inlet to drain c of the Northwest Area continued to function into Late Helladic IIIB: a sherd of this period was recovered at the lowermost end of the drain in the Northwest Area (p.51, table 1.1, no. 27). This means that a court of sorts continued to exist off the north corner of Palace B. In addition, a newly laid drain running in front of the ashlar facade at the east, underneath the later courts 42 and 47, helps to fi x the line of the northeast facade of Palace B; both drains continue in use in the Late Helladic IIIB building.10 In preparation for the new plan and orientation for Palace B, a 5-m by 20-m rectangle was 9

Blegen also noticed the evidence for an ‘annex’ or second chamber attached to the first on its northwest side, and perhaps a third next to that (PN I, 330–31).

10

The drain that runs from Room 43 and underlies by 0.10 m the floors to courtyards 42 and 47 is small in cross-section, 0.10 m high by 0.15 m wide. If it passes through holes 47-8 and 47-3 (see p.166, figs 4.2 and 4.6), as supposed by Blegen, it then slightly meanders back and forth on its way to the Main Drain. The drain has a gradient of about 5 percent, and at an intersection of the projection of this line with an extension of the Main Drain to the northeast, the calculated bottom elevation is 191.72 m, or approximately 0.37 m above the top of the Main Drain (191.35 m). Th is section of the Main Drain was explored by Blegen for only 1.4 m beyond the so-called ‘rostrum’, or the catch basin that fronts rooms 52 and 53. How this courtyard drain spilled into the lower basin down through the Main Drain remains unknown. What can be said is that it is associated with a building phase for Palace B, predecessor to the Late Helladic IIIB palace.

In Corridor 26, discussed above, the preserved ‘angonaria’, or ashlars, in a pre-Late Helladic IIIB context puzzled Rawson (they are not mentioned by Blegen), and she opined that ‘the stone seems like dumpish fill and [I] think it must be from the 1939 trench I’. The broken ashlars actually belong to the antecedent floor, along with fractured pithoi. Large ashlars can also be seen in a foundation course subtending the southeast wall of Corridor 26 (fig. 2.11); this wall also survived from an antecedent phase, belonging to the plan of Palace B. This spot appears to be the remains of a magazine of pithoi next to a lightwell.

147

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces table 1.1, no. 19). Such a stratigraphic relationship occurs in the case of drains a and b as well; these are above, and probable replacements for, the Middle Helladic drain d and its twin, drain e. Several different components make up the remains of the hydraulic system at the northwest, in and around the Southwestern Building: two different types of drains and one, possibly two lightwells. The more secure of the two lightwells, Room 76, consists of an oblong, fully enclosed chamber (fig. 2.12); there is no opening or any place in the circuit of walls for a doorway. Blegen identifies Room 76 as a lightwell (PN I, 272) ‘somewhat in the Cretan manner’; he cites parallels at Knossos (Evnas 1930, 328–73, figs 218, 219, 248, 249, 250), and I follow his identification.

2.11 Ashlars below the later LH IIIB wall in Corridor 26.

An elevation of c.191.19 m corresponds with the level and orientation of an enlarged Southeast Quad. A bit of the southeast facade of Palace B probably can be seen in the orthostate wall A below Room 7 (PN I, figs. 16–17). The network of drains discovered by Blegen in his Southwest Quadrant was probably built to service lightwells in this hypothetical Southwest Quad, some of the drains re-used, others abandoned.

The drains in this area are of two different types: first, a bracket- or pi-shaped channel cut into a pair of poros blocks (drains 78-1 and 78-2); second, a long drain built of field stones for its sides and covers (drains 76/82 and 78/82). The channelled poros blocks are embedded perpendicularly into the northwest wall of Room 78; both are broken. The better preserved drain 78-1 lies to the southwest and is 0.64 m long and 0.21 m wide; enough of the two blocks is preserved to say that they are identical in type to the string of pi-shaped poros blocks making up drain c at the toe of the outside wall for the southwest face to the Northeast Building (Part II, pp.271–72). The poros drain blocks at Room 78 have a gutter 0.11–0.135 m wide and a depth of 0.10 m. These dimensions fall within the gutter sizes of the drain at the Northeast Building, with its width of 0.08 to 0.15 m and depth of 0.08 to 0.15 m. Similarities of material, size and gutter design argue for a common source and time of manufacture that may coincide with Phase 7 of the Northeast Area (pp.40–41), when the channel was installed alongside the Northeast Building.

The top or highest level of Palace B lies to the north at c.193.2 m (the elevation of the blue-striped floor) and would have comprised a sort of ‘piano nobile’. A small 3-metre square area of Minoan-like flagged paving in Court 100 (elevation 192.27 m; see Part II, p.412, plan xxvi) lies at the southeast corner of the Southeastern Court. The adjacent South Court lies at a lower level, 191.30 m, and steps may have connected the two. The first, Late Helladic I phase of the Southwest Gateway staircase (see p.75) tops the acropolis midway along the South Court; there a column, represented by an extant base at elevation 191.35 m, greeted visitors. An entrance may have been at an opening just to the northeast end of the orthostate wall below Room 7, as detected from the cross-section of the chasm (pp.77–79). A combined stair tower and lightwell is reconstructed as an early building phase in rooms 54–57 (pp.80–82); any evidence for an input drain connected to the Main Drain in this area has been obscured by modern reconstruction, but then there is drain 53, located just 2.5 m away. As at Knossos, lightwells may well have clustered together at the south of this palace at a median elevation of c.191.40 m.

Two drains fi ll the narrow corridor between Building 82 and Rooms 76 and 78 in the Southwest Building: drain 76/82 intersects with drain 78/82. The dogleg drain 76/82 (fig. 2.13) is preserved for a total length of eight metres. In 1960, only 6.6 m of this drain was uncovered; it was only briefly noted by Blegen in his draft description of Room 78 and mapped on the annual site plan. Otherwise, so far as I can find, the drain is not discussed in the final publication. Following the 1960 season, Blegen says the drain is ‘lost in unexcavated earth’ (Blegen and Lang 1961, 154). Evidently the 76/82 stretch was uncovered in a later season.

The Northwest Area and the Southwestern Building Drain 76/82, sloping downwards and westerly for some eight metres, intersects drain 78/82 just off the corner of Building 82. Drain 78/82 starts within Room 76 and heads towards the northwest. The section of drain 78/82 at its juncture with drain 76/82 is mostly in a ruinous state. The section from its end back to where it emerges from the northwest wall of Room 76, just off the inside return of the walls to rooms 76 and 78, is likewise in a ruinous state. A projecting poros block, broken but perhaps channelled as the two just described, marks the place where drain 78/82 exited from Room 76.

The stratigraphic history of the Northwest Area is analysed above in chapter 1 (pp.47–73). Over the years and centuries of successive building phases in this part of the site, later drains were laid not only above but also close by and parallel to earlier ones, most of which continued to function. Drain d, for example, lies approximately 0.10 m above a line of discoloured and distinctive soil; Kittredge concluded that the soil streak represented a predecessor to drain d, which itself is in a Middle Helladic stratum and contained a Late Helladic I sherd inside (see pp.50–51, 148

The Late Helladic I–II Palace B below the bottom of the pair of channelled poros blocks in the northwest wall. This means that the two drains emptying from Room 78, when hypothetically extended, pass over what must be a construction already levelled to accommodate the drains; the drains from Room 78 were installed after the outer and lower wall went out of use. Contrary to what I argue, Blegen (PN I, 275) was led ‘to wonder if Room 78 was a bathroom from which water had to be drained away’. An entrance through the southwest interior wall of Room 78, outfitted with a stone and stuccoed threshold, differs in this respect from the fully enclosed lightwell Room 76. Another, more imaginative explanation for the water channels is the possibility that Room 78 served as a floral conservatory, along the lines of courts 42 and 47 of building Phase 7 in the Northeast Area (pp.40–41 above). Whatever the purpose of Room 78, it

The two branches, drains 76/82 and 78/82, are of a kind: an outside width of 0.70 m, cover stones c.0.5 to 0.6 m in width (fig. 2.14), a channel 0.20 to 0.30 m in width, an approximately uniform depth of 0.45 m, and a flow at a downward slope of 5–6 per cent to the northwest (see fig. 2.1). At the preserved eastern end of the dogleg branch, drain 76/82 has a top elevation (top of the channel, below the cover slab) of 192.15 m, with a calculated bottom of 191.70 m. At its junction with drain 78/82, the top elevation is 191.73 m and the bottom 191.27 m. The bottom and top elevations at the preserved northwest extremity of drain 78/82, at the crest of the acropolis, are 191.40 and 190.95 m, respectively. At the opposite, southeastern end of drain 78/82, where it meets the northwestern wall of lightwell 76, the top and bottom elevations are 191.92 and 191.47 m. The conjunction of walls at the north corner of Room 78 and the purpose of the wall built 0.75 m outside and parallel to the northwest wall of Room 78 are unclear. Blegen considered this outside wall to be earlier in date of construction than its neighbour to the southwest, which borders Rooms 77 and 78, but the possibility cannot be excluded that the two are contemporary. In that event we might perhaps have here another light-well flanking those two rooms. (PN I, 277) This 7.5-metre long wall lies 0.57 m below the level of Room 78 (ibid.). More important for an assessment of relative phases of construction, this wall lies c.0.45 m

2.13 Aerial view of drains within the corridor between Building 82 and Rooms 76 and 78 of the Southwest Building.

2.12 Plan of the northern part of the Southwestern Building, including Room 76 and drains 76/82 and 78/82.

2.14 Aerial view of drain 76/82 and cover stones outside the southwest corner of Building 82.

149

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces testing that had much the same pedology as samples taken from within the main palace and elsewhere on the site; the results were a composition high enough in calcite to assume the use of grout of some sort (see Part II, pp.336– 38). Any water seepage through the walls would have saturated the core, thus stabilising water loss. The later LH I drains a and b may have spilled into an additional set of reservoirs built further to the northwest on the other side of the line of walls X and J at the top of the slope (see pp.63–64 above), an area now completely eroded away.

appears to be a later appendage at the north inside corner of the Southwestern Building, attached to the outside walls of rooms 76, 80 and 81. Room 76 must be a leftover from Palace B inherited by the Late Helladic IIIB Southwestern Building. The southeast end of the dogleg drain 76/82 is overlain by the post-Bronze Age temenos wall ZZ (pp.70–73 above). It is likely that this drain continued to the east by several metres, ending with an inlet to drain off precipitation from a second lightwell, a companion to lightwell Room 76. The west corner of this latter room also marks the position of the western corner of Palace B. The organisation of drains along the northwest facade of Palace B indicates an indented trace, perhaps reflected in the later plan of the LH IIIB palace. As noted above, the northeast–southwest orientation of drain b looks in plan much like a striking snake as the drain wends its way from a light well at the northeast interior of Palace B to the fields beyond.

There is no evidence for water flowing into any of the successive palaces; rather, water was channelled out of them in all three cases. There are few signs that rainwater was collected in gutters along eaves at the edge of the roofs; instead, the spacing and configuration of the inlets at the southeast argue for lightwells. Precipitation was gathered via basins and drain inlets within lightwells or at corners of open courts spread throughout the early palaces and across the acropolis, its collection and distribution remaining essentially the same in successive palaces.

The multitudinous walls of the Northwest Area and their building history defy interpretation as house or retaining walls. They form a series of magazine-like spaces (fig. 2.15) but are otherwise quite unsuitable as storage magazines comparable to those found at Minoan palaces and villas. Rather, underlying and earlier drainage systems and replacement or reuse at higher elevations points unambiguously to water reservoirs. Patches of plaster/ cement appear here and there on the wall elevations. Earth in-fi ll from the walls was sampled and subjected to field

The Late Helladic IIIB–IIIC Palace There are compelling, but not conclusive reasons to think that the Late Helladic IIIB palace was rotated by 2–3 degrees to the northeast (more precisely, by survey traverse 2°36’) from that of its antecedent, Palace B. Weighing against this supposition is the apparent congruency of the ashlar

2.15 The magazine-like spaces of the Northwest Area, looking southwest.

150

The Late Helladic IIIB–IIIC Palace 5 by c.0.33 m; those for drain 1 and inlet 58 by c.0.53 m; and that for drain 53 by c.1.0 m. These adjustments reflect the lower but staggered floor and paving levels of Palace A, followed by Palace B. Arguably, the plan of the LH IIIB palace in many places was shaped to accommodate the continued functioning of these earlier drains. For instance, the offset in plan and the exterior jog in the west corner of Room 8 were probably designed to wrap around the intake into drain 1. Drain inlets at the north (drain c) and east (branches A and B) continued to operate unaltered by upward extensions, as did drain 42/47. Cistern 102 continued to hold water through the last phases of the LH IIIB palace, coming from some sort of collection device mounted atop the outside walls of courts 42 and 47. The actual source of water for branch B, which must be somewhere within the body of the last palace, must continue as an unsolved mystery.

orthostate wall A of Palace B preserved underneath Room 7, and the coincident southeast length of facade wall for rooms 43 to 50. In support, however, at the other side of the Main Building, are whole sets of walls that are slightly skewed by c.2.5 degrees to the major axes of the Main Building. This shift in orientation is made conspicuous by the trapezoidal shape of Court 88, caused by the 2.5-degree convergence of the outermost exterior walls to the Southwestern Building and the Main Building. Drain 8 and the light well 76 take the same orientation, and the plan of walls at the northwest of the Southwestern Building (PN I, Key Plan) clearly shows that walls of some, but not all, rooms follow this subtle shift in orientation (rooms 67, 69, 70, 72, 76, 78, 80 and 81). The remaining walls are congruent with the Main Building. This 2.5-degree divergence is also noticeable in the twoperiod construction of the wall dividing Room 65 and Staircase 69 (fig. 2.16). This wall has an offset that Blegen (PN I, 255, 264) interpreted as extra support for wooden beams supporting an upper flight. This well may be the case, but as seen on the plan, the two sides of this wall diverge northeast to southeast. The Staircase 69 side of the wall conforms to the overall orientation of the Main Building, whereas the single northwest wall of Room 65 runs parallel to the presumed antecedent orientation. In sum, it is apparent that there are two arrays of walls in the Southwestern Building, differing by 2.5 degrees in orientation, the earlier matrix of walls reused in the later, slightly shifted rebuilding. Other architectural features that conform to the antecedent axis of Palace B include the lines of blocks just outside Court 42. This set of blocks is here taken to be a flight of steps descending to a small entrance court for Palace B.

Of all these intakes, an explanation for the source of water is easiest for those found in courts 3, 58, 59 and 63, all of which were open to the sky. The bathtub in Room 43 may have been emptied by scooping water into the inlet nearby. Rooms 10 and 53, enveloped by partition walls, do not appear to qualify as lightwells, although ‘plaster of good quality’ (PN I, 218) coated both the floor and walls of Room 53, one of the reasons cited by Blegen for his identification of the chamber as a ‘water closet’ (PN I, 220). The pair of channelled poros blocks emptying from Room 78 in the Southwestern Building gave Blegen reason to suggest two different functions: bathroom or lightwell. The choice of lightwell comes from the fact that the contiguous Room 76 is a determined lightwell (see above, p.148). While never seen as commonplace, at least so far, lightwells did exist elsewhere in Mycenaean architecture, perhaps as a physical, stylistic or iconographic carryover from an earlier time. For instance, the so-called ‘bathroom’ within the palace at Tiryns, with its stone slab floor and drainage system, is probably a lightwell.

The extensive hydraulic network just described expanded in the next building period by a single reach, branch C of the aqueduct, which runs alongside the Northeast Building (fig. 2.17). A hodgepodge of drain types and shapes in stone and terracotta, probably all reused, make up the latest run of branch B from the Y-juncture at branches A and B to Area 100 (see Part II, pp.367–69). As noted elsewhere, the overall floor elevation of this last palace was raised to a height of 192.45 +/- 0.10 m, with higher values at the north, lower values at the south. Intermediate spaces between the inlets of earlier and later floors were connected by shafts and capstones. A typical adjustment is described by Blegen for Room 53 (PN I, p. 219): … possible traces of a small drain leading ... to a stone [that] covered an opening into a proper underground drain (fig. 148). Beneath some small fallen stones under the cover lay a broken slab of poros in which part of a neatly cut circular hole (ca. 0.10 in. in diameter) going through the block is preserved. This led into a channel running southeastward and bordered on each side by a well-cut block of poros…’. At the south of the palace, actual intakes for drains 8, 9 and inlet 58 were raised by c.0.15 m; those for drains 2 and

2.16 State plan of Room 65 of the Southwest Building, showing two phases of construction on the northwest wall.

151

Cooper – Hydraulic Engineering on the Englianos Ridge: Evidence for Pre-LH IIIB Palaces

2.17 Drains in the final phase of the LH IIIB palace.

potential history of the palaces here that was previously unsuspected: the visualisation of the plan and elevation data resulting in no fewer than three individual palaces and a discernable pattern of four separate floor levels, each being given shape and definition by the elaborate water system beneath them.

The details and rationale for the two Minoanising plans of pre-LH IIIB palaces at Pylos need refinement beyond what can be treated here, for while Cretan complexes have much in common with this reconstruction of early plans at Pylos, no two Minoan palaces are even nearly identical, and there are no mainland MH/LH I parallels to abet this reconstruction. Klaus Killian (1987, 203–17, fig. 12) proposed a Minoanising plan at Pylos some twenty years ago, but only the magazines in the Southwest Quadrant apply to his version of Palace B. Michael Nelson addresses the topic of Minoanising architecture at Pylos in Part II of this publication. Jeremy Rutter elaborates by providing further and convincing evidence from ceramics, wall painting etc. for a strong Minoan-Pylian connection (Rutter 2005, 17–64). In a review of Blegen’s preserved excavation material in the Chora Museum, Jack Davis, Sheryl Stocker and Gerald Cadogan ‘identified Cretan and Minoan ceramics from MH levels, some at least as early as the Old Palace period’ (Davis 2008, p. 41.) The evidence of the drainage system and the cardinal significance of the orientation of the painted floor and column bases in Area 106 – features underplayed or overlooked by Blegen – were unavailable to Killian, Nelson and Rutter but provide further evidence for Cretan connections at Englianos.

Bibliography Blegen, C. W. n.d. Unpublished manuscript for The Palace of Nestor of Pylos. ASCSA, Athens, and University of Cincinnati, OH. Blegen, C. W., and Lang, M. 1959. ‘The Palace of Nestor Excavations of 1958’, AJA 63, 121–37. Blegen, C. W., and Lang, M. 1961. ‘The Palace of Nestor Excavations of 1960, Part I’, AJA 65, 153–58. Blegen, C. W. and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and Their Contents. Princeton University Press, Princeton, NJ. D’Agata, A. L. 1992. ‘Late Minoan Crete and Horns of Consecration: A Symbol in Action’, in R. Laffineur and J. L. Crowley, eds, Aegean Bronze Age Iconography: Shaping a Methodology. Aegaeum 8, Liège, 247–56.

The hydrostatic circuit of feeding drains underneath the Domestic Quarter at Knossos serves to illuminate the significance of that at Pylos. Further, this analysis of the hydraulics across the Englianos ridge manifests a 152

Bibliography Wilson, A. I. 2008. ‘Hydraulic Engineering and Water Supply’, in The Oxford Handbook of Engineering and Technology in the Classical World, ed. J. Oleson. Oxford University Press, Oxford, 285–318.

Davis, J. 2008. ‘Pylos, Palace of Nestor’, Archaeological Reports 54, 41. Evans, A. 1921. The Palace of Minos, I: The Neolithic and Early and Middle Minoan Ages. Macmillan, London.

Revised 2011

Evans, A. 1928. The Palace of Minos II, Part I: Fresh Lights on Origins and External Relations. Macmillan, London. Evans, A. 1930. Palace of Minos III: The Great Transitional Age in the Northern and Eastern Sections of the Palace. Macmillan, London. Graham, J. W. 1962. The Palaces of Crete. Princeton University Press, Princeton, NJ. Immerwahr, S. 1990. Aegean Painting in the Bronze Age. Pennsylvania State University Press, State College, PA. Kilian, K. 1987. ‘L’Architecture des Résidences Mycéniennes: Origins et Extension d’une Structure du Pouvoir Politique Pendant l’Âge du Bronze Récent’, in Le Systeme Palatial en Orient, en Grece et à Rome, ed. E. Levy. Strasburg, 203–17. Kyriakidis, E. 2005. Ritual in the Bronze Age Aegean: The Minoan Peak Sanctuaries. Duckworth, London. Lang, M. 1969. The Palace of Nestor at Pylos in Western Messenia, II: The Frescoes. Princeton University Press, Princeton, NJ. MacDonald, C. F., and Driessen, J. 1988. ‘The Drainage System of the Domestic Quarter in the Palace at Knossos’, Annual of the BSA 83, 235–58. Marinatos, S. 1960. Crete and Mycenae. Thames & Hudson, London. Marketou. T. 1988. ‘New Evidence on the Topography and Site History of Prehistoric Ialysos’, in Archaeology of the Dodecanese, ed. S. Dietz and I. Papachristodoulou. National Museum of Denmark, Copenhagen, 27–33. Rawson, M. 1958. Unpublished Pylos excavation notebook: ‘Pylos 1958: Areas MZ–MY, vol. 1’. ASCSA, Athens, and University of Cincinnati, OH. Rutter, J. B. 2005. ‘Southern Triangles Revisited: Lakonia, Messenia, and Crete in the 14th–12th Centuries bc’, in Ariadne’s Threads: Connections between Crete and the Greek Mainland in Late Minoan III (LM IIIA2 to LM IIIC), ed. A. L. D’Agata and J. Moody. ASAtene 3, 17–64. Shaw, J. W. 1973. Minoan Architecture: Materials and Techniques. ASAtene 49, n.s.33, Rome. Wiener, M. H. 1990. ‘The Isles of Crete? The Minoan Thalassocracy Revisited’, in Thera and the Aegean World III, 1: Archaeology, ed. D. A. Hardy, C. G. Doumas, J. A. Sakellarakis and P. M. Warren. London, 128–61.

153

3 MODELLING MOVEMENT AND USE PATTERNS WITHIN THE PALACE OF NESTOR: A GIS/SPACE SYNTAX APPROACH Todd M. Brenningmeyer

The line drawing of the palace that appears in volume 1 of The Palace of Nestor (1966) probably represents the most frequently referenced view of the building (an adaptation of Blegen’s plan with contour intervals is shown as figure 3.1). The plan provides a snapshot of the structure as it appeared following numerous architectural modifications extending from the Bronze Age through the Archaic period (for the post-Bronze Age occupation of the palace see chapter 8 below). While the walls illustrated in this

0

5

10

plan present a sometimes confusing patchwork of structural change, the architectural remains themselves preserve in situ evidence for the use of the building as it evolved over time. Each modification changed the way space was organised and altered the way rooms were used and interpreted by individuals travelling within the structure and entering from outside. It is difficult to understand how the Late Helladic IIIB renovations altered the use of space within the palace simply by examining

20m

3.1 Schematic plan of the Palace of Nestor at Englianos, showing 1-metre contours.

155

Brenningmeyer – Modelling Movement and Use Patterns within the Palace of Nestor the ground plan; nevertheless, the effect of these changes can be calculated statistically. Space syntax techniques and geographical information systems (GIS) provide an avenue for modelling the effect of architectural change on potential social interaction. This chapter presents the results of an analysis of the LH IIIB modifications in the palace that relies on this methodology.

-APPED!RCHITECTURAL0LAN 3YMMETRIC.ETWORK

!SSYMETRIC.ETWORK



 







 





Space Syntax Theory Space syntax developed during the early 1980s at the Unit for Architectural Studies in the Bartlett School of Architecture and Planning, University College London (see primarily Hillier and Hanson 1984; see also Brown 1985, Plimpton and Hassan 1987, Hillier 1996, Hanson 1998 and Bafna 2003; for discussion of archaeological applications see Grahame 1997 and 2000, Cutting 2003 and Regev 2009). The approach provides a method for interpreting, representing and quantifying spatial relationships within complex structures. Within this theory, spatial relationships are abstracted to a graph model wherein rooms are symbolised as nodes, and paths between rooms are drawn as connecting arcs. The resulting graphs describe building form as a function of potential social interaction. An example of a building graphed in this manner is shown in figure 3.2, which illustrates two different spatial configurations. The building diagrammed on the right has a symmetrical plan, in which rooms are separated by few interviening spaces. The plan on the left is considered asymmetric, having rooms segregated from the exterior by several intervening spaces. The difference between these two demonstrates one of the main ideas behind spatial syntax: nodes (or rooms) with highly asymmetric graphs are more segregated within a network than those with symmetrical configurations. The more rooms that must be crossed to reach a destination node, the less accessible, more private the space.



*USTIFIED!CCESS-APSFORTHE"UILDINGS

 





 







3.2 Example rooms showing network maps (top) and justified access maps (bottom).

Relative asymmetry (RA) values describe how segregated a room is in relation to all other rooms under condiseration.2 Control provides a local value of connectedness, indicating the extent to which each room controls access to adjacent rooms.3 Relative ringiness (RRING) gives another local measure of connectedness, providing a relative measure of the choice that an individual has in the route to and from a particular room.4 In this study, these measures were calculated using topological relationships built within the GIS model described below.

Architectural Analysis of the Palace early lh iiib In the first remodelling phase of the Late Helladic IIIB period, the Main Building and several subsidiary structures were erected near the pre-existing LH IIIA structure (fig. 3.3). The configuration of the LH IIIB structure was relatively open to the exterior (fig. 3.4). Two entrances to the outside were provided on the northeast side of the building, opening onto rooms 25 and 41. Two additional entrances were located at the southern and

Graph representations of buildings, like those shown in figure 3.2, are termed justified network models.1 Such models allow the extraction of measures of spatial configuration for each space within a building. The methodology relies on the understanding that each node (or room) in a network occupies an ordinal position relative to a common starting point. In other words, all spaces that are one path removed from the starting point are one level above the starting point. All spaces that are two paths away occupy a second level above the starting point. This division of space takes all nodes in a network into consideration, and graphs developed through this articulalation of space present a visual representation of a building as a network of connecting nodes and arcs. Measures of spatial configuration derived from this abstract representaion of architecture provide information about how a space is configured and potentially used.

2

RA values are calculated using the following formula: RA= 2(MD – 1)/k-2; where k = total number of nodes in the network and MD = mean depth = ∑ ki x i/ ∑ ki- 1, where i = the depth of nodes (ki) within the network. 3 The control value is calculated by identifying the number of nodes connected to each adjacent space, taking the inverse of this value and summing these inverse values. Nodes with control values of 1.0 and above will be plotted in figures. 4 RRING = R/p-1, where R is the number of rings connected to a node and p is the total number of nodes in the network (the maximum number of possible rings will always be one less than the total number of nodes).

1

Terminology for each value discussed follows that described in Hillier and Hanson 1984.

156

Architectural Analysis of the Palace southeastern ends of the building (rooms 1 and 59), and two terminal links with the exterior were created through rooms 57 and 81.

the two main courtyards held some of the lowest asymmetry values in the complex, with an average value of 0.012. The one exception to this is Room 64, which is connected to the pre-existing LH IIIA complex and several subsidiary rooms dating to LH IIIB. The spaces that comprise this Southwest Building are among the most segregated in the entire complex, containing several values above 0.02. In addition to its high asymmetry values, none of the passages in the Southwest Building contain rings that would allow choice when travelling through the structure. The combination of high RA values and the absence of rings suggest an architectural configuration that stressed privacy through controlled access and architecturally restricted movement between rooms and the exterior.

As can be seen in the control map in figure 3.5, the two major courtyards in the complex (rooms 3 and 63) are located just two spaces below the root (exterior) node. Access to these courtyards from the exterior is controlled at each entrance by at least two other spaces, at least one of which holds a control value greater than 1, marking it as a space that significantly controls accessibility to adjacent rooms 1, 25, 38 and 59. Although some variation in the control value of these four rooms can be identified, the values are, on average, just above one. The exception to this is Room 38, which carries a value that is nearly twice that of any of the other three controlling spaces.

Three additional areas of the palace also present relatively high asymmetry values. The southwest area of the Main Building (rooms 21, 20, 19 and 17) has RA values that, on average, are slightly less than those of the Southwest Building, averaging between 0.016 and 0.019. The rooms in the northeast (rooms 32, 30, 31, 29, 33 and 34) and

The relative segregation of rooms within the complex can be interpreted from the normalised RA value map. The higher the RA value the more segregated the rooms are within the network; spaces with RA values approaching zero tend to be more public. Rooms directly connected to

0

5

10

20m

3.3 Schematic plan of the Palace of Nestor showing Late Helladic IIIA to IIIB periods.

157

Brenningmeyer – Modelling Movement and Use Patterns within the Palace of Nestor

Figure 4 !

81

!

! 75

!

!

!

79

25

!

!

!

!

!

22

21

!

72

71

73

!

74

Map of Late Helladic IIIB Network

78

77 80

23

!

24

68

70

!

!

20

32

18

!

88

!

!

19

30

6

!

29

!

16 28

33

!

!

!

17

31

!

!

!

64

!

!

65

!

67

!

69

66

!

13

36 39

40

!

!

!

12

!

34

!

!

63

35

!

!

!

14

5

!

4

! !

!

3

!

!

37

11

!

10

41

38

44

! ! 43

!

9

!

59

54

!

!

61

!

!

!

45

!

!

!

56

7

46

!

55

2

8

!

60

48

51

57

!

!

!

!

1

53 49

50

!

! !

!

!

52

!

0

5

10

20m Exterior

3.4 Late Helladic IIIB network.

Figure 5 81

Late Helladic IIIB Period: Distribution of Controlling Spaces

78

77

#

80

75

79 #

25

72

71

23

#

24

68

70 #

67

20

88

32

18

#

66

#

#

69

22

21

#

73

#

74

19

30

6

16 28

#

17

31

29

#

#

64

#

65

33

#

13

36

#

63

35

#

#

14

5

39

40 #

#

12

34

4

37

11

#

#

#

10

3

44

41

38

# #

59

43

9 54

#

61

55

#

2

56

7

46 45

#

8

#

60

48

51

57

#

#

#

1

53

50 #

52 #

49

#

0

5

10

20m

Exterior

3.5 Late Helladic IIIB distribution of controlling spaces.

158

Control Value #

1.0-1.4

#

1.4-2.0

#

2.0-3.5

Architectural Analysis of the Palace

Figure 6 #

81

#

# # #

75

79

25

72

#

71

#

#

23

24

68

70

20

88

32

18

#

67

#

66

#

#

#

#

#

69

19

30

6 29

#

16 28

33

#

#

#

17

#

64

31

#

#

#

#

#

65

#

#

22

21

#

#

73

#

74

Late Helladic IIIB Period: Network Rings and RA Distributions

78

77 80

#

13

35

36

#

#

63

34

#

#

#

#

14

5

39

40

#

#

#

12 #

4 #

37 3

#

#

11

#

#

#

10

41

38

44 # #

#

9

#

59

54

#

#

61

43

#

#

#

45

#

#

#

56

7

46

#

55

2

8

#

60

48

51

57

#

#

#

#

1

53

50

49

#

#

#

#

#

52

Relative Assymetry (RA) Values #

Exterior

0

5

10

20m

#

0.009 - 0.012

#

0.011 - 0.015

#

0.015 - 0.018

#

0.018 - 0.022

#

0.022 - 0.03 Network Rings

3.6 Late Helladic IIIB network rings and RA distributions.

Figure 7

#

81 78

77

#

75

79 #

#

25 #

#

24

#

#

20

88

32

18

#

67

#

66

#

#

#

19

30

6

31

#

#

#

#

#

65

#

#

#

23

68

70 69

#

22

21

#

72

71

73

#

74

Late Helladic IIIB Period: Relative Ringiness Distributions

#

#

80

29

16 28

#

17

#

#

64 #

33

#

#

13

5

36 39

40 #

#

#

12

#

34

#

#

63

35

#

#

#

14

#

4

# 44

#

#

3

#

#

37

11

#

10

41

38

# 54

#

#

#

#

#

45 #

56

7

46

#

55

#

2

8

#

60

48

51

#

#

#

#

1

57

50

# Exterior

5

10

20m

3.7 Late Helladic IIIB relative ringiness distributions.

159

#

#

#

#

#

53

52

49

0

43

#

9 #

59 #

61

Relative Ringiness Values #

0

#

0.001 - 0.015

#

0.015-0.029

#

0.03 - 0.056

Brenningmeyer – Modelling Movement and Use Patterns within the Palace of Nestor In addition to the change in potential circulation, there was a related change in the way rooms were integrated within many parts of the complex (fig. 3.9). The relative asymmetry values for the rooms in the southwest decreased as greater access to the rest of the palace was created through the passages cut into rooms 21 and 20. As a result of these changes, rooms 18, 20, 21 and 22 acquired values averaging about 0.013, approximately 0.003 below their previous values. Smaller changes can be identified in rooms located elsewhere in the palace. Room 46 witnessed a slight decrease in its RA value from the previous period. Likewise, the RA value of the exterior node decreased slightly as a result of closing off entrances on the east side of the palace.

southeast (rooms 46, 48, 49, 50, 51, 52 and 53) sections of the Main Building also hold high asymmetry values, ranging between 0.016 and 0.02. Despite their similarly high asymmetry values, the eastern rooms are configured differently than the rooms to the southwest, which do not lie on rings. The southwest area instead is composed of spaces that form terminal links in a network. In other words, individuals entering these rooms must exit using a previously taken path. This is not the case with the two areas at the eastern end of the palace, where both the northeast and southeast sections lie on rings that offer several possible routes of circulation. Although spaces in both the northeast and southeast areas lie relatively deep within the building, once an individual travels to either location, there are several paths that can be taken to move through each complex of rooms.

The architectural changes also affected the way certain spaces controlled access to adjacent rooms (fig. 3.10). The control value for Court 63 increased during the late LH IIIB period. The court provided access to rooms 12, 61 and 64, as was the case in the earlier period, but also led to the rooms in the southwest area of the palace through the smaller Court 88; this change raised the control value from 1.16 to 1.50. Rooms 41 and 46 also acquired greater control values when rooms 42 and 47 were added to the palace. Although changes can be detected in several areas, not every room witnessed a significant alteration in its control or RA value. For example, Court 3 maintained its control value through both periods, and its RA value remained relatively unchanged through each phase.

The ringed configuration of the northeast and southeast areas is indicative of the spatial configuration of the eastern side of the palace during the LH IIIB period. The polygons in figure 3.6 outline all rings contained in the palace during this period. The largest number of circulation routes fall in the eastern section of the building, forming links between rooms and between different entrances to the palace. Figure 3.7 provides an illustration of the RRING values for all rooms in the building. Rooms with high RRING values tend to cluster around Court 3 and entrances 41 and 1. In contrast to Court 3, Court 63 lies on a single ring that connects the western area of the palace with the main building. late lh iiib

Social Interpretations of Architectural Change

The second building phase of the LH IIIB period significantly altered the spatial configuration that defined the architecture of the previous period (fig. 3.8). Rooms 42 and 47 were added to the southeast end of the palace, closing access to the exterior from Room 41. Rooms 26 and 27 also closed access to the exterior, and a partition was added to the passage between rooms 33 and 31. These architectural alterations removed many of the network rings that were described above (compare figs 3.6 and 3.8). New routes of circulation, however, were created in other areas of the site. Doorways were added to rooms 21 and 20, linking the southwest section of the Main Building with Court 63. Corridor 16 was also partitioned at this time, creating two new rooms (rooms 18 and 22) in the southwest area.

The space syntax values that are described above express changes in potential social interaction as a factor of the building’s architectural layout. In order to understand how these values correspond to actual room use, however, it is necessary to establish criteria for correlating spatial configuration with social activity. Using studies of the House of the Faun at Pompeii as a model, social interaction within the palace will be divided into two broad categories of social encounters that describe different scales of interaction: ‘occasions’ and ‘gatherings’ (Grahame 1997, 155; Graham 2000, 9). Occasions are defined as social encounters that involve multiple individuals and may be associated with specific social practices or rituals. Areas used primarily for occasions may require additional spaces arranged around a central room; these rooms serve as loci for additional and related activities or as storage places for equipment needed for the occasion. Spaces directly connected to multiple rooms in this way will tend to have control values greater than one. Gatherings are less formal meetings and may involve brief encounters in hallways or other informal interactions within a space. Gatherings may occur in a range of spaces, but in general occur in areas that are relatively accessible. In both types of encounters, rooms that are more segregated within a complex will tend to facilitate more private types of interaction than rooms that are relatively accessible.

These architectural modifications effectively closed access to the exterior from the eastern section of the palace and promoted accessibility and movement in the western part of the site. Rooms in the west became more closely connected through additional passageways that introduced new circuits into this section of the main palace building. Figure 3.9 illustrates the new distribution of relative ringiness throughout the complex. The largest concentration of rings is in the western end of the palace, primarily in the area that connects courts 3 and 63.

160

Social Interpretations of Architectural Change Later Late Helladic IIIB Period: Network Rings and RA Distributions #

81 78

77

#

#

80

#

#

#

#

25

72

20

32

18

#

88

#

#

#

#

#

19

30

6 29

#

16

28

#

17

#

64

31

#

#

#

#

#

65

26

#

67

24

#

#

#

66

#

#

#

69

23

#

22

21

#

68

#

70

#

#

71

73

#

74

27

75

79

33

#

#

#

13

5

35

36

34

#

63

#

#

#

#

#

14

39

40

#

#

#

12 #

4

42

#

37

41

38

44

#

3

#

#

11

#

#

#

10

# #

#

54

#

0.022 - 0.03

#

0.018 - 0.022

50

# # #

#

#

#

0.015 - 0.018

#

#

#

0.011 - 0.015

#

53

52

49

#

0.009 - 0.012

#

#

#

#

#

47 48

51

1

57

Relative Assymetry (RA) Values

#

#

#

#

45 56

7

46

#

55

2

8

#

60

#

43

#

9

59 #

61

Exterior

0

Network Rings

5

10

20m

3.8 Later Late Helladic IIIB network rings and RA distributions. Figure 9

#

Later Late Helladic IIIB Period: Relative Ringiness Distributions

81 78

77

#

#

80

#

#

#

#

25

72

#

26

#

#

20

88

32

18

#

67

#

66

#

#

#

#

#

19

30

6

31

#

#

#

#

#

65

#

#

#

24

#

#

#

23

68

70 69

#

22

21

#

71

73

#

74

27

75

79

29

16

28

#

#

17

#

#

64

33

#

#

13

36 39

40 #

#

#

12

#

34

#

#

63

35

#

#

#

14

5

#

4

#

42 41

38

44 #

#

43

#

9 #

54

46

#

55

#

47 #

#

56

#

#

#

#

45

#

2

8

7

48

51

#

#

#

#

1

57

53

50

#

0 0.001 - 0.015

#

0.015-0.029

#

0.03 - 0.056

#

Relative Ringiness Values

#

#

#

#

#

52

49

#

#

#

3

#

60

#

#

59 #

61

#

37

11

#

10

Exterior

0

3.9 Later Late Helladic IIIB relative ringiness distributions.

161

5

10

20m

Brenningmeyer – Modelling Movement and Use Patterns within the Palace of Nestor Figure 10 Later Late Helladic IIIB Period: Distribution of Controlling Spaces

81 78

77

#

80

27

75

79 #

25

72

71

22 #

23

#

24

26

68

70 #

67

20

88

32

18

#

66

#

#

#

#

69

#

21

#

73

#

74

19

30

6

31 #

65

29

16

28

#

17

#

#

64

33 #

13

5

36

#

63

35

#

#

14

39

34 40

#

#

12 4

42 37

3

44

#

11

#

#

#

10

41

38

# #

59

54

55

#

#

56

7

46

#

2

8

#

60

43

9

#

61

45

47 48

51

57

#

#

#

1

53

50 #

52 #

49

#

1.0-1.4

#

1.4-2.0

#

2.0-3.5

#

Control Values

Exterior

0

5

10

20m

3.10 Later Late Helladic IIIB distribution of controlling spaces.

EARLY lh IIIB

was denied or approved after an individual arrived in Court 3.

Figures 3.5 and 3.10 illustrate the distribution of controlling spaces across the two periods under discussion. During the first LH IIIB remodelling, three sections of the palace carried most of the controlling spaces: the Southwest Building, the rooms around Court 3 and the rooms in the eastern end of the building. Each area held at least one room with a control value greater than 1.4. Of these areas, the complex of rooms surrounding Court 3 holds particular interest. Court 3 lies on four different rings and has the highest RRING value in the complex. The court also has a low asymmetry value and is separated from the exterior by just two spaces (rooms 2 and 1). When viewed using the criteria described above, it suggests a space used for public occasions. The spaces located around the court correspond with the pattern mentioned above, as access to additional rooms with differing levels of accessibility is provided from a central ‘public’ space. The rooms surrounding this central space would provide loci for different types of social interaction that might develop from initial contact in Court 3. For instance, visitors could enter the court through spaces 2 or 38 (the entrance chosen would probably depend on the context of the social interaction). After the visitor enters Court 3, movement to several different parts of the palace, each with different RA values and probably different degrees of privacy, is possible. A range of potential factors would likely determine how access to other spaces

The two primary entry points to Court 3, rooms 38 and 2, are configured differently within the first LH IIIB remodelling. Both rooms were located two levels above the exterior node and held controlling positions in the building relative to the exterior. Both spaces also controlled access to the megaron and its forecourt (Court 3) and were probably used as primary entrances, though for different types of individuals or activities. Room 2 is situated on a series of arcs that intersect at Court 3, connecting the southern section of the palace through three intersecting routes. Room 2 opens directly onto the most public space in the building but provides direct access to a limited number of rooms, each containing relatively low RA values; Room 2, therefore, primarily provides access to relatively public spaces within the building. Room 38, in contrast, provides access to six different rings that connect with Court 3. The room also allows access to some of the more private areas of the palace (e.g. rooms 30, 31, 32 and 33) without passing through a ‘public’ courtyard. The configuration of these two entrances suggests a functional division between rooms 2 and 38, whereby the majority of individuals entered through the public access point at Room 2, and select groups entered through Room 38, where access to relatively ‘private’ areas was a primary concern.

162

Bibliography This spatial configuration may be compared to that of Court 63, the second largest court in the complex. This court controls access to the rooms in the Southwest Building, which in terms of potential social interaction are in the most remote part of the palace. The court lies on a single ring that runs between Court 3 and the exterior. Although Court 63 has a relatively low RA value, the combination of low control (at least in relation to other controlling spaces) and ringiness values suggests a space that is used less frequently or for different types of interaction than was the case with Court 3.

Conclusions The picture that begins to emerge is one in which certain spaces were architecturally modified to accommodate a more public and commercial pattern of use. During the early LH IIIB period, Court 63 perhaps held a function closely connected to the Southwest Building. Although the space was accessible from other parts of the palace, it was not well integrated with the other rooms and was positioned on a single circulation route, connecting the socially remote rooms in the Southwest Building with the rest of the palace complex. During the later LH IIIB period, this changed. As access between the storage areas and the exterior was closed through the addition of rooms 26 and 27, Court 63 became the primary route for the transfer of goods. This path avoided the circulation routes within the main building, enabling goods to pass to and from the palace without interrupting activities in other parts of the building. The results of this analysis suggest an intentional division in the functionality of the palace that changed over the course of two remodelling sequences through the re-arrangement of space. While the western side of the building acquired a public character, the eastern side became a more private space.

late lh iiib The social roles that the two courts played changed radically during the late LH IIIB period. During this period, new rooms were built in the southwest part of the Main Building, and new points of access to Court 63 were created through the addition of doorways in rooms 20 and 21. The spatial configuration of this court and the entire west end of the palace took on a much more public character. The RRING and control values for the court increased during the late LH IIIB period. These modifications served to connect Court 63 with several additional rooms in a manner similar to the centres for ‘occasions’ described above and illustrated by Court 3 during the earlier LH IIIB period. Room 59 became one of only two entrances to the palace when rooms 38 and 25 were closed to the exterior.

The interpretation presented here describes the architectural modifications of the LH IIIB and late LH IIIB periods within the context of evolving architectural use and social interaction. The effects of the late LH IIIB modifications are not immediately evident when examining Blegen’s familiar line drawing, but become apparent when the architecture is abstracted to a linear model. As a topological approach, this methodology is well suited to GIS-based analysis of architectural remains, allowing an understanding of the use of space through multiple phases of construction using fragmentary and sometimes incomprehensible remains.

The spatial configuration that has been described is interesting when placed within the context of the available archaeological evidence. Archive rooms 7 and 8 (for the analysis of which see Blegen 1966, 92–100, and Palaima and Wright 1985) were directly connected with Room 2, being situated just to the south of this room near the exterior of the palace. The rooms’ position near what is here interpreted as the primary ‘public’ entrance may relate to the record-keeping activities that have been associated with these spaces.

Bibliography

Linear B tablets recovered from rooms 7 and 8 describe an active palace economy during the last phase of Bronze Age occupation. Numerous items enumerated in the tablets flowed into and out of the palace as commerce during the late LH IIIB period; among these are ceramic vessels, wine and olive oil, all of which are known to have been stored in the palace. Vessels containing oil and other materials were stored in rooms 23 and 24 during the last phase of the palace, and large concentrations of pottery from this period were recovered from rooms 18, 19, 20, 21 and 22, which probably served as storage areas. Located in Room 60 was another store of nearly 800 vessels, which also date to the last phase of the palace and may represent items intended for redistribution or trade. The late LH IIIB date of these remains is the same period that witnessed architectural renovations connecting each of these storage areas to Court 63. The structural modifications were probably closely tied to the function of the rooms.

Bafna, S. 2003. ‘Space Syntax: A Brief Introduction to Its Logic and Analytical Techniques’, Environment and Behavior 35:1, 17– 29. Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and their Contents. Princeton University Press, Princeton, NJ. Brown, M. G. 1985. ‘Objective Measures of Spatial Organization in Architecture’, Research and Design 85: Architectural Applications of Design and Technology Research, General Proceedings, 403–08. Cutting, M. 2003. ‘The Use of Spatial Analysis to Study Prehistoric Settlement Architecture’, Oxford Journal of Archaeology 22:1, 1–21.

163

Brenningmeyer – Modelling Movement and Use Patterns within the Palace of Nestor Grahame, M. 1997. ‘Public and Private in the Roman House: The Casa del Fauno’, in Domestic Space in the Roman World: Pompeii and Beyond, ed. R. Laurence and A. Wallace-Hadrill. Journal of Roman Archaeology Supplemental Series 22, 137–64. Grahame, M. 2000. Reading Space: Social Interaction and Identity in the Houses of Roman Pompeii: A Syntactical Approach to the Analysis and Interpretation of Built Space. BAR International Series 886. British Archaeological Reports, Oxford. Hanson, J. 1998. Decoding Homes and Houses. Cambridge University Press, Cambridge. Hillier, B. 1996. Space is the Machine: A Configurational Theory of Architecture. Cambridge University Press, Cambridge. Hillier, B., and Hanson, J. 1984. The Social Logic of Space. Cambridge University Press, Cambridge. Palaima, T. G., and Wright, J. C. 1985. ‘Ins and Outs of the Archives Rooms at Pylos: Form and Function in a Mycenaean Palace’, AJA 89, 251–62. Plimpton, C. L., and Hassan, F. A. 1987. ‘Social Space: A Determinant of House Architecture’, Environment and Planning B: Planning and Design 14, 439–49. Regev, E. 2009. ‘Access Analysis of Khirbet Qumran: Reading Spatial Organization and Social Boundaries’, Bulletin of the American Schools of Oriental Research 355, 85–99.

Revised 2009

164

4 ENCLOSED GARDENS IN COURTS 42 AND 47 anne b. hollond

The topic of gardens in the Aegean Bronze Age has long been of interest to scholars and excavators. James Walter Graham (1987, figs 6 and 58), for example, offers suggestions for the location of gardens at Knossos, Phaistos and Mallia; Maria Shaw (1993) does the same for Phaistos, demonstrating, moreover, that gardens, rather than natural landscapes, are depicted in a number of frescoes. This interest in the Aegean garden has largely been restricted to Crete and Thera, however; the possibility of gardens – either actual or represented in art – at the mainland Mycenaean palaces has not received much support. Evidence at the Palace of Nestor urges a reconsideration of this possibility.

The Architecture of Courts 42 and 47 Courts 42 and 47 are adjacent enclosed open areas, each approximately the same size (55 x 12 m), located on the northeastern flank of the Main Building (fig. 4.1). They were added during the final phases of occupation in the Late Helladic IIIB period (late 13th century bc) as part of a series of architectural and functional changes to the palace (see Part II, p.364). That they were added at the same time is indicated by the continuous rubble-built wall that encloses both courts and abuts the northeast facade of the Main Building. The wall separating the courts bonds neither with the Main Building nor with the enclosing rubble wall, and must have been added later. Court 42 is irregular in plan and was probably built with respect

4.1 Plan showing courts 42 and 47 and their relationship to the main palace building.

165

Hollond – Enclosed Gardens in Courts 42 and 47 to the presence of surrounding structures. Room 102 was already standing and in use when Court 42 was added, for the southwest wall of that room was incorporated into the northeast wall of the court. The two courts can only be accessed from within the palace, each through a single opening. The Northeast Gateway 41 forms the entrance to Court 42, and a small doorway connects Hall 46 with Court 47. Both courts were paved with stucco, and the interior face of the long northeast wall (wall 13L) was coated with plaster, as were both faces of the cross-wall. Court 42 was fitted with a terracotta pipe, one end of section B of the aqueduct, which pierced the court’s northeast wall; the direction of the pipe, with its narrow end to the northeast and fitting into the wider end of the next pipe in the sequence, appears to indicate that water was channelled away from the court, not into it (see pp.139–40 above). The floor surface of the court slopes from northwest to southeast, facilitating the drainage of runoff to a single hole at the southern end of the southeastern wall (see PN I, fig. 426, where the hole is labelled ‘a’); this hole apparently also served as a drainage hole for Room 43, which held a bathtub, as a cut channel leads from Room 43 under the wall between the two rooms to the hole in Court 42. The stucco floor in Court 47 is both punctured and punctuated by 14 holes (fig. 4.2; see also PN I, figs 41 and 42). Four holes lie in a somewhat regular northeast-to-

4.2 (left) Schmatic plan of the courts, showing the position of holes. 4.3 (above) Row of four holes in Court 47, looking northeast.

166

The Courts as Garden Spaces southwest alignment on the northwest side of the court (fig. 4.3); the remainder follow no discernible pattern. Reexcavation and cleaning of these holes in 1998 revealed that they are conical in shape, flaring outwards from top to bottom (figs 4.4–4.6); all were lined with a packing of pebbles, including the two along the southwest side that Blegen incorrectly assumed were connected to the line of the Main Drain (PN I, fig. 426). The diameters of the roughly circular holes (one is triangular in shape) vary between 12 and 17 cm. Three holes punctuate the floor: one has a depth of 5 cm, the other two a depth of 9 cm. The other holes puncture the floor to a depth of 15–20 cm.

The Courts as Garden Spaces It has been suggested that these two areas served as private patios or courts for royal individuals (PN I, 181), or as storage and work areas for the manufacture of perfumed olive oil in the final days of the palace’s existence (Shelmerdine 1984). An interpretation of these courts as enclosed gardens is proposed here, based on the construction of the holes and their irregular placement in Court 47, the system of hydraulics in Court 42, the enclosed but probably unroofed configuration of the courts, their location at the northeast side of the Main Building, and the pottery found within them. Analysis of soils from the two courts also suggests the presence of flowering plants, and possibly perfume manufacture, in this area (see pp.86–87 above).

4.4 Hole 47-2.

The holes were obviously meant to receive something. Blegen and Rawson suggested that the four aligned holes running southwest to northeast near the centre of Court 47 may have anchored posts of a loom or a now missing structure, such as a shade or partition (PN I, 208). The first 47-2

47-8

4.5 Hole 47-4.

47-6

47-3

47-4

47-5

47-7

47-13

47-14

42-1

4.6 Plans and elevations of ten of the holes in courts 42 and 47.

167

Hollond – Enclosed Gardens in Courts 42 and 47 suggestion is unlikely, for warp-weighted looms such as the one identified in room 206 in Troy IIg (Blegen et al. 1950, 350, fig. 461) require two postholes aligned parallel to and near a wall (see also Barber 1991, 93). It is also unlikely that the holes anchored shelving posts (Shelmerdine 1985, 62). Shelving postholes, which were sunk into the floors of rooms 19, 20 and 60, are positioned in a much more regular layout – typically parallel to and near the walls (see Part II, plan xvii; see also PN I, fig. 189 ). Furthermore, the conical shape and pebble lining of the holes that pierce the stucco floor are inappropriate for architectural elements. In holes dug to receive columns or substantial posts, one would expect large, flat stones at the bottom of the cavity to provide a solid foundation; earth and stucco would be packed and poured closely around the vertical member to prevent it from moving, as occurred with the columns of the Throne Room (PN I, 80, fig. 67; see also Part II, p.302).

had flaring sides. Though it cannot be said for certain, these pots could have functioned as flowerpots. The theory that the holes in the floors of the courts were planting holes was tested by inserting into them flowerpots of the approximate size and shape of those mentioned by Rawson. The opening at the top of the hole cradled the pot, while the hollow space beneath and the pebble lining allowed – indeed, facilitated – drainage into the earth below. The random pattern of the holes in Court 47, with their flowerpots, can be understood as an aesthetic arrangement in a garden space (fig. 4.7), and the pots could be removed quickly and easily from the holes for whatever reason, such as replanting or seasonal arrangements. The other holes in the court that seem to follow a more regular pattern might have been used in a similar manner (fig. 4.8), and it is possible that they also functioned at times as postholes for a sun shelter and/or partition. Such structures would not be out of place in an enclosed garden.

If the holes were not built to receive a permanent loadbearing device, then something movable and temporary must have been meant to sit within them. Marion Rawson, who supervised the original excavation of these courts, makes numerous references in the notebooks to ‘flowerpot fabric’ and ‘flowerpot ware’ in both areas, especially Court 47. This material is, in fact, rarely attested on the site outside of these courts. The references are unclear, unfortunately, and there is no discussion of this material in the final publication, but Rawson’s descriptions of the sherds reveal some important aspects of the ware: it was reddish in colour, coarse, and common shapes were handles and flat bases; vessels were large and sometimes

The pipe in the external wall of Court 42 (fig. 4.9) carried water away from the court, towards the northeast through the aqueduct; its position 0.40 m above the floor suggests that a catch basin of some sort – whether a permanent feature or simply a large open vessel – probably stood below it, the pipe serving as an overflow regulator (see pp.139–40 above). The drainage hole in the south corner of Court 42 would have helped to eliminate any standing water that might have accumulated on the floor. The average elevation of the floor across Court 42 is 192.60 m

4.7 Courtyard 47 looking northeast, with modern planting in the Bronze Age holes.

168

Precedents and Conclusions (wall 13L) argues against any sort of roofing. Assuming the courts were open to the sky, their location on the northeast side of the Main Building gives them about four hours of morning sunlight throughout the year. Conversely, there is plenty of shade on this side of the building during hot summer afternoons, making the two courts cool and desirable spaces at that time of year. The enclosing walls serve to define the garden spaces architecturally. The walls restrict both the egress of their contents (which might have included small animals as well as plants) and the ingress of things that might disturb or threaten these contents (e.g. other animals and the frequent and intense winds that blow across the hilltop). The walls also created two very private spaces. The only entrance into Court 47 is through the Main Building’s northeast suite of rooms, and this corner of the building was itself entered through a single, closeable door (see Part II, p.285, fig. 1.3). Indeed, this room is perhaps the most remote room in the Main Building (see p.161, fig. 3.8), and as such it must have been a quiet and peaceful place, separated from the rest of the complex.

Precedents and Conclusions A Mycenaean interest in plants, particularly flowers and aromatic oils, is suggested by Linear B texts. A number of tablets in the Fr1 and Un2 series from Pylos pertain to the production, supply and use of perfumed olive oil at the palace. Given the importance of this commodity, it is surely possible that some of the raw aromatic plant materials listed on the tablets – coriander, cyperus, sage, roses, alkanet and henna – were also considered important. These aromatic plants, and others such as those named in the Ge tablet series from Mycenae,3 could have been cultivated in courts 42 and 47, especially since some – henna, cumin, sesame – were not native and thus would have had to be obtained through trade and/or specially cultivated, lending them additional value and requiring protected surroundings.

4.8 Court 47 with modern f lowerpots, looking northeast.

above sea level, whereas the lip of the drainage hole lies at 192.30 masl, a drop of some 0.30 m. (It should be kept in mind that in the initial design, courts 42 and 47 were probably one large enclosed area. The non-bonding crosswall separating the two spaces was added later, perhaps when the garden function ceased in Court 47, and Court 42 became either a storage area or manufacturing facility for perfume.) Both courts received adequate sun for growing plants. The long span from the northeast facade wall of the Main Building (wall 11L) to the northeast wall of the courts

Depictions of natural elements in Mycenaean wall paintings are relatively few, especially in comparison with earlier Minoan frescoes, and those that do survive are assumed to be derivative and decorative rather than realistic. Mabel Lang remarked on the landscape frescoes recovered at the palace that ‘Pylos has little to show of nature for nature’s sake’ (1969, 25). Nevertheless, there is evidence for the naturalistic depiction of plants, both in published and unpublished painted plaster fragments. In addition to flowers picked and carried by women, there are paintings showing both large-scale vegetation with animals and smaller plants, including flowers that appear to be 1

Fr 1216, 1220, 1222, 1224,1226, 1228, 1233, 1235, 1236, 1246 (Shelmerdine 1985, 71–72). 2 E.g. Un08 and 09 (Ventris and Chadwick 1956, 223–25). 3 E.g. Ge602, 603, 604 (ibid., 225–31).

4.9 Terracotta pipe in the northeast wall of Court 42.

169

Hollond – Enclosed Gardens in Courts 42 and 47 frescoes from that site show flowers in containers sitting in the windows of houses. The features identified by Shaw and Evans are those found in courts 42 and 47 at the Palace of Nestor. Although it cannot be said for certain, these spaces at least offered the possibility of cultivating plants, whether flowers or herbs and spices, in a controlled and protected environment.

Bibliography 4.10 Fresco fragments showing possible f lora. Barber, E. J. W. Prehistoric Textiles: The Development of Cloth in the Neolithic and Bronze Ages, with Special Reference to the Aegean. Princeton University Press, Princeton, NJ.

growing, perhaps in the context of miniature landscapes. Despite criticisms of their artistic shortcomings in comparison with the flora of Minoan frescoes, Lang (ibid., 125) concedes that at least some of the published fragments ‘depict vegetation in its natural state’, noting in particular that ‘perhaps the crocus (?) of 9 N 47 and the anemones of 15 N sw come closest to Minoan flowers and may have served as natural scenery. The delicate plants of 8 N 32 [possibly Matthiola longipetala, a form of fragrant stock native to the Mediterranean], accompanied as they are by at least one fish, may have been part of a miniature marine frieze’. The MARWP excavations produced some 20 pieces with possible floral decoration (fig. 4.10) to augment the 28 fragments with floral decoration catalogued by Lang (ibid., 124).

Blegen, C. W., Caskey, J. L., Rawson, M., and Sperling, J., eds. 1950. Troy, I: General Introduction and the First and Second Settlements. Princeton University Press, Princeton, NJ. Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and their Contents. Princeton University Press, Princeton, NJ. Evans, A. 1930. Evans, A. 1930. Palace of Minos III: The Great Transitional Age in the Northern and Eastern Sections of the Palace. Macmillan, London. Graham, J. W. 1987. The Palaces of Crete. Princeton University Press, Princeton, NJ.

Architectural evidence of built gardens in the Mycenaean world is almost nonexistent, and it is therefore difficult to relate the two courts proposed here as such spaces to contemporary and similar structures. Other Mycenaean palaces are not nearly as well preserved as that at Englianos, and it may be that garden areas have simply failed to survive, or failed to attract the attention of archaeologists. Nevertheless, comparisons may be made to gardens in the early Late Bronze Age palace culture of Minoan Crete. Maria Shaw successfully demonstrated a Minoan taste for ‘limited and discrete human tampering with the natural landscape’, particularly as depicted in art (Shaw 1993, 664; she also discusses earlier precedents for Aegean gardens). Shaw identified a potential garden space at the palace of Phaistos, which consisted of an L-shaped portico partially surrounding a plot of exposed bedrock. Irregularly placed small holes, into which plants may have been inserted, were carved into the bedrock. The garden space was secluded within the larger palace complex and overlooked the Mesara plain.

Lang, M. 1969. The Palace of Nestor at Pylos in Western Messenia, II: The Frescoes. Princeton University Press, Princeton, NJ. Shaw, M. C. 1993. ‘The Aegean Garden’, AJA 97, 661–85. Shelmerdine, C. W. 1984. ‘The Perfumed Oil Industry at Pylos’, in Pylos Comes Alive: Industry and Administration in a Mycenaean Palace, ed. C. W. Shelmerdine, 81–95. New York Society of the AIA and Fordham University, New York. Ventris, M. and Chadwick, J. 1956. Documents in Mycenaean Greek. Cambridge University Press, Cambridge.

Submitted 2002

At Knossos, Arthur Evans (1930, 277–79) suggested that light wells could have been adorned with potted plants, as they provided sunlight, air and moisture whenever it rained; many lightwells were drained, which would have prevented standing water from accumulating in these spaces. Evans found direct evidence for vases with pierced bases (ibid, fig. 186) and proposed that these functioned as flowerpots, positioned in lightwells and around the courts of Knossos. A flowerpot was recovered from Thera, and

170

5 BRONZE AGE QUARRYING: A PROVENANCE STUDY joshua n. distler

The site of the Bronze Age palace at Englianos and the surrounding region of Messenia have been the object of intense study by the University of Cincinnati, the Greek Archaeological Service and the University of Minnesota since the excavation of the site by Carl Blegen beginning in 1952. The Minnesota Messenia Expedition (MME) conducted interdisciplinary studies from the late 1950s to the late 1960s, including economic analyses of hunting, agriculture and trade, material analyses of metals, ceramics and pollen, and the mapping of topography, geomorphology and demography (McDonald and Rapp 1972); these studies were renewed and extended in the 1990s by the Pylos Regional Archaeological Project (PRAP). The issue of stone provenance was not broached by either group, however, and the quarries discussed below, identified about 5 km northwest of the Palace of Nestor in the summer of 1992, were not even mapped by the MME research team, despite their proximity to the palace, their visual prominence and their size. The location of the quarries, the methods of extraction and the nature of the stone all make these potential candidates for the source of the blocks at Pylos. The dearth of recorded information on Messenian quarries leaves a real possibility that other quarries exist, or once existed, in the region.

the preponderance of this stone in later Classical antiquity, studies such as Wenner and Herz (1992) have shown that the technique is readily applicable to limestone artefacts as well. Limestone, in addition, has the advantage of being unmetamorphosed and therefore retaining its constituent fossils and sedimentary structures unaltered. This opens the option of a more robust, multidisciplinary approach to provenance determination, including paleontological and sedimentological examinations, procedures not possible for marbles.

The Geology of the Southwest Peloponnese The geology of southern and central mainland Greece and the Aegean islands is in large part the result of tectonic activity along the Hellenic Trench, about 100 km off the west coast of the Peloponnese; there, the Ionian plate is subducting below the Aegean plate (fig. 5.1). One effect of this subduction is that marine sediments such as limestones once deposited on the floor (the Ionian plate) were scraped off by the lip of the Aegean plate and thrust east in a series of partly-overlapping sheets, or nappes. Four distinct structural levels of the West Hellenic Nappe System have been identified in the Peloponnese. In the region of Messenia in the southwest quarter of the Peloponnese, the geology (apart from Holocene alluvial sediments of the last 11,000 years) is part of the third of this series, the diverse Tripolitza nappe (igme geologic map, Koroni-Pylos-Skhiza sheet, 1980). It consists of Middle Triassic (240–230 million years ago) to Upper Eocene (44–37 mya) limestones covered by a layer of Oligocene (37–24 mya) flysch (a ‘dirty’ orogenic sandstone of mixed composition).

A host of methods has been developed to characterise lithic attributes, ‘fingerprints’ that allow stones to be correlated with the quarries from which they were excavated. These include major- and trace-element analysis, as well as isotopic, petrographic and paleontological analysis. Stable isotope analysis, pioneered by Craig and Craig (1972), has proved very successful in determining provenance of marbles. By this method, concentrations of stable isotopes of carbon and oxygen present in the rock are measured and plotted relative to a standard. The resulting fields serve as characteristic signatures of a given quarry location. While the significant number of provenance studies utilising stable isotopic signatures have involved marble artefacts, monuments, and their associated source quarries owing to

In the vicinity of the modern city of Pylos and to the south are a succession of sedimentary deposits that include (west to east) thin-bedded Late Cretaceous limestones, EocenePaleocene (37–67 mya) limestones, very recent Holocene 171

Distler – Bronze Age Quarrying: A Provenance Study occur only in a finite number of locations. These include the bluffs just to the east of modern Pylos, a long strip running along the coast of the Messenian peninsula, the island of Sphacteria (and two small islands south of Methoni), the three small bluffs just north of Navarino Bay, the island of Proti across from Gargaliani and, finally, the long bluff running north of the same city. All of these outcrops are Eocene age limestones, with the exception of the Cretaceous (144–67 mya) limestone along the coast west of modern Pylos.

The Quarries As noted, ancient Pylos is situated in the midst of several square kilometres of flysch, a stone that does not lend itself to building purposes. While the upper portions of walls appear to have been primarily composed of local materials such as timber, rubble fill and clay brick, the impressive foundations and lower walls of the main palace and the Southwestern Building were of natural or worked slabs and ashlar blocks of poros limestone (see Part II, pp.298–301). Several disorderly piles of these cut blocks, or angularia, were discovered by Blegen just south of the palace walls, evidence of later stone-robbing attempts (fig. 5.2). Samples for this study were removed from such ex situ angularia. Since the significant quantity of limestone found at the site is not indigenous to Englianos, the task of transporting it must have been at least as time-consuming as that of extracting and shaping it. The quarries under investigation are located in a long limestone outcrop that rises from a plain of sandstone and flysch near the modern Greek town of Gargaliani (map 5.3). The physiography of the outcrop is a moderately steep bluff for most of its length, exposed virtually continuously

5.1 Top: plate tectonic scheme for the Aegean region (after Underhill 1989); bottom: cross-section of subduction scheme for the Aegean region.

(less than 1 mya) alluvial deposits, and Eocene-Oligocene flysch. Between Pylos and the town of Gargaliani to the north, some of the Eocene-Oligocene flysch continues, but the region is primarily covered by Pliocene (5.3–1.5 mya) marls, sandstones and conglomerates. The site of Nestor’s Palace sits in the midst of several square kilometres of this sandstone and flysch (map 5.1). North of the palace site in the region of Gargaliani and north to Filiatra, the sediments are almost exclusively sandstone, with some flysch deposits to the east. However, running northnorthwest to south-southeast is a long belt of Eocene limestone, the southern tip of which ends approximately 2 km northwest of the Palace of Nestor (map 5.2). It is in outcrops of this formation that the ancient quarries were cut. The limestone and flysch are associated with the IonianTripolitza nappe, but the greater volume of sediments in the region (sandstones) date to the Neogene (less than 24 mya) and are the result of post-thrust deposition. Thus, while much of the southwestern Peloponnese is underlain by carbonate sediments, outcrops of limestone

Map 5.1 Englianos area. Portion of Filiatra sheet, IGME 1:50,000, 1980.

172

The Quarries the series of quarries is situated in the upper 20 m of the bluff, the steepest portion of the outcrop. To the south of the town, the quarries are also cut into the seaward face of the outcrop, but the rock exposure is on a smaller scale and more easily approachable owing to a gradual inclination. The nearest recorded cutting is about 16 km from Englianos along the modern highway, or 5 km to the northwest as the crow flies. Erosion has significantly pocked and karsted the face of the entire outcrop, and exposed level features as well as the interiors of the quarries have suffered weathering sufficient to obliterate most tool traces. Iron oxide stains running down the outcrop are concentrated in the quarries themselves, highlighting the fact that the roofed cuttings are not entirely protected from the leaching and weathering effects of precipitation. The soft, friable poros limestone that comprises most of the length of the outcrop is especially susceptible to weathering by meteoric water (rainfall, groundwater), and the high acidity of precipitation in the Aegean due to recent urban and industrial pollution compounds this effect. In addition to natural weathering processes, the quarries have probably been subject to centuries of habitation by animals and humans, which undoubtedly acted as catalysts in the cycle of erosion that defaced the original quarrying traces.

Map 5.2 Portion of Filiatra sheet, IGME 1:50,000, 1980.

with the exception of a couple of interrupting drainages, extending north-northwest to south-southeast from approximately 5 km south of Gargaliani to 12 km north of the village. The base of the cliff, at 200 m above sea level, is separated from the sea by 4 km of Pleistocene (1.6 million–11,000 years ago) coastal terrace. Over a horizontal distance of 200 m, the outcrop ascends another 80 m, with a shallower incline near the base grading to vertical in the topmost 20 m. At its crest, the exposure levels off and gently rises inland.

I have divided the quarries into three groups based solely on their north-south location along the bluff. Groups I and II, the northernmost ones, in particular are virtually continuous and separated only by a minor promontory. Group III consists of a more widely-spaced string of quarries south of Gargaliani in less precipitous terrain. It should be noted that the grouping distinction is primarily for sample-collection purposes and does not imply any genetic differences between the subsets. In fact, as will be shown, it is quite possible that several or all of the quarries were excavated coevally and should be treated as a chronological unit.

The quarry excavations are distributed in a number of clusters along the length of the outcrop, spanning about 2 km south of Gargaliani and 1.5 km in the opposite direction, following the seaward side of the rise. In every case, the cuttings are horizontal cavities, with flat floors and ceilings and slightly concave walls. To the north,

group i Group I, the cluster of cuttings farthest north, displays many attributes that suggest an ancient (and possibly Bronze Age) history. The scatter of quarries is moderately spaced, with 10–30 m between cavities, and it tends to follow a single stratigraphic layer of limestone, maintaining a steady elevation. As in all cases, these quarries are located in the steepest (topmost) portion of the outcrop, where talus and vegetation do not obscure the bare rock. The two best preserved cuttings in Group I are found on the northwest side of the promontory, Quarry Ia (fig. 5.3) being the longest in the group (about 15 m wide and half as high), while Quarry Ib (fig. 5.4), near the tip of the point, is more square (approximately 5 m in height and width). The series of quarries consist of a number of smaller cuttings (5–10 m wide) with rounded openings, as well as longer examples. Similar to several of the longer grottos in the two other groups, quarries more than 15 or 20 m long are

5.2 Angularia at the Palace of Nestor.

173

Distler – Bronze Age Quarrying: A Provenance Study

Map 5.3 Region around Gargaliani, following Hellenic Geographical Survey, portion of Filiatra sheet, 1977.

often supported by interspersed piers — vertical columns of unexcavated native rock attached to the back wall of the quarry, extending from floor to ceiling. The quarry floors and ceilings are consistently flat, even and parallel to the nearly horizontal bedding layers of the outcrop. The ceilings of the Group I quarries gently slope down in the rear of the grottos to meet shorter back walls, where smaller-scale niches or benches were frequently cut. If any tool marks were once evident in these quarries,

they have been weathered away from the soft poros limestone. The excavations are cut straight back into the rock face, with flat benches or platforms of various dimensions at their entrances; these might once have served as loading or storage decks. Before the mouth of Quarry Ia lies a sizeable platform of this sort (about 15–20 m long and 10 m wide), and the stratigraphy above the same quarry has

174

The Quarries been removed to create another similar platform on top of it (fig. 5.5). These two major platforms, which define the two main tiers of quarrying activity in the Group I site, are connected at the northern boundary of the complex by an irregular ‘staircase’ of low, cut benches (fig. 5.6). The majority of quarrying evidence at the Group I site is found between the set of bench steps at the north and Quarry Ib at the southern end of this concave section of outcrop. Nevertheless, a number of smaller, more isolated cuttings were identified along a stretch several kilometres north of Group I. group ii Group II, located just south and on the opposite side of the small rock promontory, contains the largest and bestpreserved assemblage of excavations (fig. 5.7). This series is most evident from the road below (about 1.5 km northwest of Gargaliani), though only accessible by foot (a brief but arduous climb of about 80 m). Again, the excavations were carried out in the upper 20 m of outcrop in two distinct tiers of differing elevations, separated by a horizontal stratum of rock (fig. 5.8). There exist eight or nine major grottos in the Group II sequence, some with rounded apertures and others longer, with rectangular openings. These cuttings tend to be slightly deeper, averaging 5–10 m deep, and closer to one another. The quarries here are on a similar scale to their counterparts just to the north, though it appears that slightly more stone was extracted at

5.3 Quarry Ia from northwest.

5.4 Quarry Ib from south.

5.5 Platform above Quarry 1a, from north.

5.6 ‘Staircase’ in Group I, from west.

175

Distler – Bronze Age Quarrying: A Provenance Study

5.7 Group II quarries, from southwest.

Group II sites.1 All have remarkably flat floors and ceilings, which curve gently downward to meet the wall in the rear of the caves. In almost each instance there are smaller niches and benches recessed in the back walls of these quarries, possibly evidence for smaller extraction needs at a later date. Quarry IIc is a prime example of parallel cuttings and niches (fig. 5.9).

exhibit only narrow platforms (more akin to catwalks) connecting the entrance of one to another. Some of the terraces show signs of low benching as in the Group I ‘staircase’ (figs 5.10–5.11), though here it is occasionally obscured owing to surface weathering.

As in Group I sites, the quarries in Group II all have level platforms before their entrances, ranging from a couple of metres wide to a more significant expanse of 10–15 m (see fig. 5.8). The largest platforms are located at the base of the lowest cuttings, while grottos on the upper tier frequently

Group III sites are located a short walk east of the modern road (about 3 km south of Groups I and II) up a gentle rise planted with olive trees.2 They have notably similar morphologies to those in the preceding two groups, exhibiting round to long horizontal cuttings with flat floors and descending ceilings. Indeed, all of the quarries examined are surprising in their similarity, suggesting coeval excavation. Their size, location on the outcrop and morphology are relatively invariant. Evidence of benching, recessed niches, piers and platforms are all found at these sites as well, with the important addition of two other examples of quarrying traces: a straight ledge in the floor

group iii

1

Due to constraints of time and equipment, an estimate of the total stone removed from any of the quarries was not obtained. The ancient quarries at Gargaliani are certainly on the scale of Minoan analogues such as Mochlos, which supplied the palace of Gournia with an estimated 350 cubic metres of ammoudha blocks (Soles 1983), or possibly Ta Skaria, which supposedly produced 1300 cubic metres in eolianite and ammoudha blocks for the Middle/Late Minoan town of Palaikastro (Driessen 1984). In many Minoan sites, negative impressions from the quarries (block dimensions and total output of stone) have been successfully employed to aid in provenance determination (Soles and others). Simple dimensional analysis of the worked areas in relation to the quantity of limestone employed at Nestor’s Palace would provide fruitful material for future research.

2

The full extent of the Group III quarries was not scrutinised as completely as in Groups I and II. Quarry IIIa has been entirely incorporated into a modern domestic enclosure, and many of the sites at IIIb and IIIc were significantly obscured by thick bushes, vines and creepers; these hid features both directly and by their shadows, and sheltered at least one viper.

5.8 Workings at Group II quarries, from northwest.

176

The Quarries

5.9 Quarry IIc, from north.

and tool traces in the ceiling of a relatively long grotto, Quarry IIIc. The ledge (fig. 5.12) of Quarry IIIc extends almost the length of the cutting, parallel to the entrance, and is consistently a few centimetres in height. Its remarkable straightness suggests that it was a major horizontal splitting plane for the base of a layer of blocks as well as the terminus for a vertical separation trench. The tool traces (fig. 5.13) have fairly shallow to medium depth and are of two varieties. Both examples are located in a few square metres of low ceiling in one of the rectangular niches in the longest quarry of the group. At the outer lip are straight, parallel or sub-parallel incisions almost a metre long. The relatively long cuts and smooth strokes indicate a tool with a long handle, and the smoothness and narrowness of the furrows give evidence for a sharp point. Occasional stray strokes indicate a single point, not a multi-tined rake or chisel. These traces were probably made by a bronze pick removing the overburden above the stratum of limestone to be cut for blocks. The second type of trace occurs farther back on the same ceiling, where the overhang meets a fairly vertical wall at the rear of the grotto. These traces are only a few centimetres in length, sometimes only shallow holes, and with a generally downward direction. The short strokes and irregular, pocked texture indicate a short-handled punch, probably a more efficient tool in the cramped space at the rear of the niche. These traces are terminated by a smooth

5.10 Benching in Quarry IIa, from southeast.

5.11 Benching in Quarry IId, from southwest.

177

Distler – Bronze Age Quarrying: A Provenance Study Group III contains evidence of a small platform of flat outcrop near the quarry entrances of sites IIIb and IIIc (fig. 5.16). No tool traces are visible, but a group of limestone boulders are clustered in the vicinity. Among these boulders an unexpected block of angularia was discovered (fig 5.17). The block, measuring approximately 1 x 0.75 x 1.5 m, was finished on at least five of its six sides and heavily pocked by weathering. Smaller rectilinear embayments were recessed in two sides. Though the block lies in close proximity to the IIIb and IIIc quarries, it appeared to be a slightly harder, more crystalline limestone than the outcrop. This example represents the only piece of worked stone found at any of the quarries under examination. The country rock of the outcrop remains the same soft, yellow-white, calcareous limestone (poros) as at the more northerly sites, and the cuttings follow the same thick to massive beds in much the same horizontal sequences. In these more southern excavations, however, the topographic relief of the exposed rock face is less pronounced, and there is only one vertical tier of grottos, though the individual cuttings are similar in size. In addition, the quarries tend to be clustered in smaller series with greater distances separating each quarry site, implying less efficient removal of blocks.

initial assessment 5.12 Ledge in Quarry IIIc, from northwest.

There are a number of factors that initially drew attention to the Gargaliani sites as potential Bronze Age quarries associated with the Palace of Nestor. The most obvious is a complete lack of other potential ancient quarries among which to choose. The only identified ancient quarries located in the region of the southwest Peloponnese are those along the Cape of Mani. The Mani quarrying sites are a considerable distance from Pylos, and the great majority of them (Mt Taygetos, Pyrgos Dirou, Cape Lamares and others) produced white and coloured marbles. The oldest examples among them reveal evidence of activity only from the Archaic period. The existence of the sites at Gargaliani is all the more compelling because of this scarcity, and their morphology supports ancient origins.

horizontal groove running along the top of the low back wall, undoubtedly the top of the channel used to separate the blocks from the rear wall. The floor at the base of the rear wall reveals a shallow complementary groove that probably represents the bottom of the separation channel. This particular niche (figs 5.14–5.15) shows a morphology similar to some of the cuttings in each group of quarries, yet it appears to have suffered less weathering damage owing to protective foliage and an especially low overhang. Its exceptionally flat floor and walls, clear piers and rectangular shape, and well-preserved tool traces are very possibly indicative of what the other quarries once exhibited.

5.13 Workings in Quarry IIIc: ceiling (above) and rear wall (right).

178

The Quarries

5.14–15 Niches in Quarry IIIc, from west.

The horizontal sequence of flat-floored caves along an outcrop face is a pattern closely reproduced at a series of Middle Minoan III/Late Minoan I ammoudha/eolianite quarries at Ta Skaria on Crete (Papageorgakis et al. 1992, fig. 2). The platforms for loading and storage of blocks after excavation are common at most Bronze Age sites, and tool traces at the Gargaliani quarries match the types of marks found at other Bronze Age quarries. In addition, there is unmistakable use of piers at the Gargaliani sites, which is seen at Ta Skaria and a number of other Minoan sites, and the benches and separation ledges at Gargaliani have direct analogues at many ammoudha quarries on Crete, such as Mochlos (Soles 1983, figs 5–7, 16), Ta Skaria, Spilia, Hagio Pneuma (Waelkins 1992, figs 15, 19, 22), Malamoures and Pelekita (Papageorgakis et al. 1992, figs 4–9).

yellow colour, though markedly grey to black in highly exposed and leached areas. In contrast, the Cretaceous limestone on the west coast of the Messenian peninsula is described as thinly bedded, and outcrops to the east of modern Pylos are of a slightly different hardness and texture. Even samples collected from the modern quarries near Gargaliani, just hundreds of metres east of the string of ancient quarries, show a strikingly different character. Though they do not seem to have suffered a significant degree of metamorphism, since fossils and other primary structures in the rock are intact, the stone is harder with a denser, more crystalline matrix. In addition, the colours are more varied, with grey, black and white bands and streaks. The limestone between outcrops of similarly aged rock seems to vary within relatively short distances in Messenia, and the outcrop that contains the ancient quarries appears to be fairly unique. The lithologies and fossil content of the samples will be discussed in more detail below.

Finally, the block dimensions delineated by the separation grooves in the niche at Quarry IIIc are very close to some of the ashlar block dimensions seen at ancient Pylos, and the stone type of the blocks at Englianos matches fairly well with that found at the Gargaliani quarries. The covered palace blocks are light yellow to cream-coloured (more grey for exposed angularia) sandy-grained poros limestone. The blocks are massive; that is, they do not include marked bedding planes and so must have been excavated from an outcrop with equally wide layers. The quarries sampled show similar attributes, being generally a dirty

The Problem of Transportation No evidence of a finished road for transporting blocks down the outcrop is visible today, and there is no means for a vehicle to approach the quarries from below. Quarries from Hellenistic times onward would have probably

5.16 Platform at Group III quarries, from southeast.

5.17 Cut block from Group III quarries, in situ.

179

Distler – Bronze Age Quarrying: A Provenance Study occasioned a more efficient and permanent approach road, while the cuttings in question must have been served by a simple skid path.

aided. From there, blocks could be loaded onto barges and shipped south to the mouth of either the Koryfasion River or Oxbelly Bay.

The care and skill revealed by the finely finished ashlar courses at Englianos in conjunction with the effort required to transport the stone from the nearest possible outcrop of native limestone suggest a proficiency on the part of Late Helladic Messenians at both quarrying and transportation procedures. As the nearest source of limestone, the Gargaliani outcrop must have been appealing to the ancient builders. The string of sampled quarries ranges from 5–8 km from Englianos, while the bluffs at Oxbelly Bay are at least 7–8 km away, and the highlands east of modern Pylos are an intimidating 12 km distant. The fact that the proposed ancient quarries at Gargaliani are in the closest outcrop does not preclude farther sources, since it is not easy to obtain an accurate gauge of Mycenaean stone transport capabilities. In addition, stone quarried relatively close to the ocean could have been loaded on barges and moved greater distances than by overland routes. This seems to have been the case at many coastal Minoan quarries, such as Mochlos, which is 14 km from the palace of Gournia, the supposed destination of its blocks (Soles 1983). Nevertheless, the relatively small distance from the sampled quarries is a favorable attribute.

Paleogeographic reconstructions of the region using sediment cores show that in the Late Helladic period, neotectonics and sea-level changes would have produced a transgressive (landward-moving) shoreline, and Oxbelly Bay would have communicated with the Bay of Navarino (map 5.4). The modern sandbar that now isolates the lagoon probably did not form until Late Helladic to Classical times, and the Late Helladic shoreline for the north end of Navarino Bay would have been up to a kilometre farther north than its present position (Kraft and Aschenbrenner 1980). The reconstruction suggests that the Helladic drainage system for the area north of the bay would have produced a river that emptied into the bay at this northern shore. The river as projected would have wound from the northeast, within 1–2 km of Englianos, to a southwestern mouth in the Bay of Navarino. Today it empties into the Ionian Sea farther north, owing to geomorphic alterations ascribed to tectonic changes and very possibly human intervention. Analysis of aerial photographs and microtopography by PRAP has provided additional evidence that supports the theory of artificial diversion of the river. The presence of anomalously large Holocene alluvial deposits along this hypothesised path necessitates the past action of a sizeable stream and supports the concept.

Importantly, the stone cargo from these excavations could have reached Englianos by two possible routes. The first is the direct overland approach, transporting the stones by horses, carts, skid roads and the like. While this route has the advantage of minimising the distance covered, obstacles of terrain and vegetation present themselves. The modern terrain between Gargaliani and ancient Pylos probably closely resembles the 12th century bc analogue, since it is shaped by slow erosional and hydrological forces that have not varied much.3 Blegen describes the topography just to the north of the site as ‘a succession of rugged ridges and deep ravines’, and to the west, ‘waves of abrupt crests and hollows and, beyond low bordering hills, the blue Ionian Sea’ (Blegen and Rawson 1966, 30; hereafter PN I). By some path over this daunting terrain, the heavy stone must have travelled if it came overland. The limestone outcrops around the Bay of Navarino to the south are surrounded by low topography, but there are several kilometres of very hilly terrain separating ancient Pylos from those as well. The possibility of taking advantage of waterways for stone transport would seem a favorable alternative to difficult overland portaging. The Gargaliani quarries are 4 km from the ocean, but it is all down-sloping or relatively flat coastal terrace, making the construction of skid or cart roads fairly easy and gravity-

Paleo-reconstructions of the Pamisos River delta, the drainage system located just east of the Navarino system, support the idea as well. Evidence (Kelletat et al. 1978) suggests that neotectonic rise and subsidence, a critical factor in coastal geography and river morphology, varied locally: whereas the area of Kalamata rose during the Pliocene, for instance, the Pamisos delta sank. Nevertheless, it has been shown that both the Bay of Navarino (Kraft et al. 1985) and the Pamisos River (Kraft et al. 1975) experienced similar uplift since the Holocene (encompassing archaeological time). Today, the Pamisos is drained and cultivated with a sandbar at its mouth, and in the recent past it was recorded as ‘widespread coastal malarial swamps in addition to better drained areas of natural levee’ (ibid). Paleogeographic reconstructions for the Late Helladic period, however, project that the coast lay some 1 km inland from its present position; indeed, Pausanias (4.34.1) remarked in the second century ad that ‘the Pasimos flows through plowed land, the water is pure, and ships sail up it from the sea about a mile and a quarter inland; sea living fish run up it as well’ (trans. Levi 1971). This evidence further supports the possibility of a navigable river in the region north of the Bay of Navarino during the Late Helladic (Kraft et al. 1975; Kraft, personal communication). If Mycenaean workmen could have negotiated some or all of this river with a stone barge, possibly pulled upstream by men and oxen on the riverbank, the amount of land transport would be greatly reduced. The total overland portaging by the sea route

3

While neotectonic and erosional processes have altered the geomorphology of Messenia since the Middle Helladic on a scale of metres, this effect is especially pronounced in coastal areas. ‘Erosion of the uplands .... is quite slow. With the exception of a few raw gullies, little damaging erosion is occurring’ (Loy 1967, 60).

180

Lithic Analysis not destroyed by subsequent diagenesis (alteration due to increased temperature and/or pressure), as is the case for marbles. Despite the fact that hand-sample and petrographic evaluations of limestones and marbles may not be conclusive provenance identifiers, a megascopic and mineralogical characterisation of the rocks is important, since significant differences can exist on this scale.

would equal the relatively easy 4 km of coastal terrace between the quarry sites and the ocean and the 1–2 km from the last negotiable point on the river to the palace, as opposed to the direct overland route of 6–8 km over terrain riddled by deep valleys and steep hills.

Lithic Analysis Samples for lithic, paleontological and geochemical analyses were collected from 12 locations (map 5.5 and tables 5.1–5.2). An effort was made to select at least two discrete samples from each rock formation or archaeological site. All the samples were collected from within a 15-km radius and are Eocene limestones. Samples from the ancient quarries, the modern quarries and the bluff near Pylos all represent outcrop (bedrock); samples from the palace, Koryfasion and possibly the cut block represent uncertain sources.

While an archaeological study of the palace blocks and their proposed quarries is elucidating, it is not sufficient to assign provenance. It is therefore important to describe stone artefacts that may be visually similar by other characteristics, such as chemical and isotopic signatures. Both a major element study and an isotopic analysis were conducted on the samples, and results from these analyses will be discussed below. In addition to the chemical and isotopic techniques that have been applied to marble artefacts with success, application of paleontological analysis is possible for limestones, since the fossils are

Map 5.4 Paleographic reconstruction of the region surrounding the Bay of Navarino from the Mesolithic to modern times (after Kraft et al. 1980); note shoreline and drainage system of the Helladic period.

181

Distler – Bronze Age Quarrying: A Provenance Study There is clear motivation for sampling the palace blocks and the proposed ancient quarries. The bluff near modern Pylos was selected because it represents a large Eocene limestone outcrop within possible transport distance. No suspected ancient quarries have been located in this outcrop, but this does not discount their possible past existence. Samples from this site also serve as a control, to show whether the similarities between other samples are significant. The modern quarries were chosen because they are potential ancient quarry sites. They are close to the palace and so would have allowed easy transport;

modern quarrying activity suggests high-quality stone; and modern workings could easily have obliterated ancient evidence. Moreover, these sites are fairly close to the suspected ancient quarries and would provide another, geographically closer control against which to compare variation among samples. The Koryfasion samples represented limestone found as surface ‘float’ on a hill about 4 km south of the palace, in the sandstone/flysch environment that also surrounds Englianos. The outcrop that holds the ancient quarries is

Map 5.5 Petrographic sample locations. 182

Lithic Analysis only slightly farther from the Koryfasion site than from Englianos, and the fragments must have been transported from such a limestone source by natural or human forces. Sherd evidence for Mycenaean activity on and around this sample location

hill has been found by the Pylos Regional Archaeological Project.

modes

description

1. Limestone fragments scattered on a hill near Koryphasion

30% bioclastic, 50% lithic, 20% mud

Fairly homogenous microgranular to mud matrix with some small brecciated bioclasts, very little pore space; low FeO; wellsorted lithics; yellow/cream coloured; slightly more indurate than 3, 4, 4a from palace.

2. Limestone fragments scattered on a hill near Koryphasion

20% bioclastic, 60% lithic, 20% mud

Fairly homogenous microgranular to mud matrix with some small brecciated bioclasts, very little pore space; low FeO; wellsorted lithics; yellow/cream coloured; slightly more indurate than 3, 4, 4a from palace.

3. Block of angularia (fallen cut stone) at the Palace of Nestor

10% bioclastic, 30% lithic, 40% mud, 20% algal

Large algal bodies (>1mm) engulfing bio./lith./mud-spar matrix similar to 1 and 2; medium pore space; high FeO shown in red colouration; clasts of brecciated bioclastic material; sparry veins; soft and friable.

4. Block of angularia at the Palace of Nestor

10% bioclastic, 30% lithic, 50% mud/calcite matrix, 10% algal

Slightly less algae than 3; bio./lith./mud matrix in interstices similar to 3; medium pore space; high FeO; large sparry calcite crystals filling some pore spaces, other pores with layer of isopagous calcite needles.

4a. Block of angularia at the Palace of Nestor (same stone as 4)

30% bioclastic, 40% lithic, 30% mud

Microgranular to mud matrix, little pore space; no algae; low FeO; well-sorted lithics; pale yellow colouration with friable, homogenous matrix, similar to many blocks in main building of palace.

5. Limestone formation at site of proposed ancient quarries near Gargaliani (Quarry I)

30% bioclastic, 20% lithic, 50% mud/calcite matrix

Medium amount of pore space, with layer of isopagous calcite needles, some with sparry grains; brecciated bioclastic clasts similar to 3 and 4, matrix similar to 4a; high FeO in portion of sample and pore rims; yellow/cream coloured; heterogeneous matrix.

6. From limestone formation at proposed ancient quarries near Gargaliani (Quarry II)

30% bioclastic, 20% lithic, 50% mud

Similar to 5, though FeO in pore rims only; slightly more indurate than 5.

7. Limestone formation at proposed ancient quarries near Gargaliani (Quarry III)

30% bioclastic, 20% lithic, 50% mud

Similar to 6, though slightly smaller clasts; small lithic grains; highly friable.

8. Anomalous cut block near entrance to Quarry IIIb

30% bioclastic, 50% lithic, 20% mud

No pore space; rare terrigenous lithic grains; low FeO; no algae; much more indurate than samples 1–7 and possibly crystalline, with greyish-white colouration; matrix mostly homogeneous.

30% bioclastic, 10% 9. Modern limestone quarry SW of Gargaliani, c. 1 km from lithic, 50% mud/calcite matrix, 10% algal ancient quarries

Large sparry calcite crystals in veins and filled pores; large (>1mm) algal bodies; low FeO; colour and hardness similar to 8.

10. Second modern limestone quarry SW of Gargaliani, c. 1 km from ancient quarries

40% bioclastic, 55% mud, 5% algal

Calcite veining similar to 9; same matrix as 9; less algae; no lithics; low FeO; colour and hardness similar to 8.

11. Bluff just E of modern Pylos

60% bioclastic, 40% mud/calcite matrix

Thin calcite veins and filled pore space; mud peloids, no lithic grains; colour and hardness similar to 8.

12. Bluff just E of modern Pylos

Medium to thick veins of diagenetic calcite (more than any 25% bioclastic, 10% lithic, 35% mud/diage- other sample); few-to-no lithics; colour and hardness similar to 8. netic calcite

Table 5.1 Results of petrographic study. 183

Distler – Bronze Age Quarrying: A Provenance Study petrographic study

paleontological study

rocks, since the temperatures and pressures common in metamorphism tend to destroy or significantly alter the shells, or tests. For this reason, paleontological characterisation techniques have not been applied to marble provenancing, and traits such as crystal size are substituted. In contrast, in largely unaltered limestones, the calcium carbonate exoskeletons of a broad range of marine organisms are commonly preserved; test morphology is often characteristic of a given species and can provide an abundant and varied set of assemblages by which to distinguish samples.

Fossil evidence can be used to correlate rock strata with other outcrops or with particular time sequences. Generally, the technique is only possible in sedimentary

Ideally, the tests are separated from the rock and examined individually in three dimensions. Attempts to separate the fossils from the rock samples at the University of

The 12 samples were somewhat similar in hand sample, though there was variation in density, friability and colour (table 5.1). In all the samples, the presence of microbrecciated bioclasts (biomicritic fragments that have been redeposited and lithified), hematite (iron) staining, as well as cement composition, terrigenous lithic content and pore space, were used to judge depositional setting.

sample

description

1. Koryfasion

Foraminifera: high globigerinerid (P) content; significant amount of coiled and keeled forms (P?). Other: bryozoans. Fairly homogenous, slightly layered microgranular-mud matrix; minimal pore space.

2. Koryfasion

Foraminifera: lower fossil content but many of coiled and keeled forms (P), heterostegina (B). Other: ostrocods. More iron oxide than 1.

3. Palace of Nestor angularia

Foraminifera: few forams, coiled and keeled (prob. P); some globigerinerids (P). Other: high algal content, bryozoans. High iron oxide content; sparry (crystalline) calcite veins interspersed in fairly porous matrix; biomicritic material similar to 1 and 2.

4. Palace of Nestor angularia

Foraminifera: coiled and keeled forms with hyaline and porcellaneous (B?) tests; small amount of globigerinerid forms (P). Other: less algae than 3; bryozoans. Very similar to 3, though the alga bodies are slightly smaller.

4a. Palace of Nestor angularia

Foraminifera: significant amount globigerinerid (P) and coiled (P?) forms; no algae.

5. Ancient quarry I

Foraminifera: globigerinerids (P), coiled, keeled (P?), miliolid (B), small fragments of discocyclinid (B) and nummulites (B). Other: high diversity detritus of algae, bryozoans, molluscs, bivalves, echinoderms. Poorly sorted terrigenous lithics and sparry areas in a miciritic/sparry cement with moderate pore space.

6. Ancient quarry II

Foraminifera: globigerinerids (P), keeled and coiled (P), miliolid (B), uni- and bi-serial (B); small fragments of discocyclinid (B) and nummulites (B). Other: bryozoan, algal, mollusc, echinoderm detritus. Similar content to 5, but with less iron.

7. Ancient quarry III Foraminifera: globigerinerids (P), coiled (P) forms. Other: diverse detritus of algae, molluscs and echinoderm. Similar to 5 and 6, but with smaller clast size and lower concentration of bioclasts. 8. Cut block from quarry IIIb

Foraminifera: large nummulitids (B) dominate, large discocyclinid (B) fragments, many coiled and keeled forms. No pore space, low iron oxide and high organic diversity.

9. Modern quarry

Foraminifera: many large discocyclinid (B) fragments, some keeled forms and nummulites (B) fragments. Other: large algal bodies. No pose space and high organic diversity.

10. Modern quarry

Foraminifera: large discocyclinids (B) common; keeled and miliolid forms; rare globigerinerids (P); porcellanous forms. Other: algae, bryozoans. No pore space, homogenous micritic cement.

11. Modern Pylos

Foraminifera: agglutinated uni- and bi-serials (B) dominate; many coiled and keeled forms. Entirely foraminiferal bioclasts, lowest density. Low diversity sample, in a cement matrix similar to 9 and 10.

12. Modern Pylos

Foraminifera: low amount, low diversity; some coiled and keeled forms; fragments of altered nummulitids and discocyclinids. Other: bryozoans. Similar to 11; high concentration of sparry veins and in-fills.

Table 5.2 Results of paleontological study.

184

Lithic Analysis Massachusetts micropaleontology laboratory were not successful, and the fossils had to be identified in thin section (30 micron-thick slices of rock) under a petrographic microscope. Because this method produces slices through fossils rather than views of the whole, fossil genus and species were not commonly identifiable. However, recording organic diversity at the level of phylum and separating of the foraminifera into categories of planktonic (those living in the water column) and benthic (bottom-dwelling) proved more successful. The bioclastic content of the sampled limestones, fossil proportions, and quality of fossil preservation were used to judge depositional environment.

knowledge of modern facies can allow one to link rocks deposited as paleo-sediments to depositional facies that exist today. Table 5.3 shows the assignments of samples to standard microfacies (SMF). According to Walther’s Law of Facies Succession (Tucker and Wright 1990), lithological facies that were laterally continuous (adjacent) when deposited as sediment will be vertically superimposed in the rock record. This occurs because relative sea level, primarily responsible for the location of facies, changes relative to the shore; as it does, the facies shift along with it. Since the facies represented by these collected samples form a laterally continuous spectrum, it is not implausible that they could be found quite close to one another in a limestone outcrop. The variation among the proposed ancient quarries themselves shows that this must have occurred to some degree, and the vast difference between samples 3 or 4 and sample 4a proves that more than one facies was quarried. More complete sampling would be necessary to gauge how great the variation is within the vertical rock sequence that was excavated.

The predominant fossils composing the limestones in most samples were calcareous algae and foraminiferae, or single-celled marine protists that secrete diagnostic calcareous shells. In addition, fragmentary shells of bryozoans (small chambered organisms), molluscs (e.g. snails), bivalves (e.g. clams) and echinoderms (e.g. sea urchins) were found in many of the samples (table 5.2). Microfacies assignments

Diagenetic grouping Compilation of hand sample, petrographic and paleontological evidence for all samples was employed to form a hypothetical depositional environment, or facies, for each. For carbonates, these facies are highly dependent on temperature and water depth and can be used fairly accurately to record changes in these factors. The facies that occur in marine carbonate settings range from nearshore ‘restricted platforms’ (lagoons and tidal bars), to reefs, foreslopes, open shelves and finally deep water basins (Wilson 1970; 1975). Each facies has a set of characteristic lithological and paleontological traits that has remained remarkably similar throughout Phanerozoic time; thus,

The hand sample, paleontological and petrographic evidence suggest another segregation scheme for the samples as well. While the facies assignments themselves do not preclude any of the samples from appearing in one outcrop with the others, examination of the lithic evidence points to quite disparate diagenetic histories for some of the samples. Differing matrix compositions suggest the separate grouping of samples 1–7 and samples 8–12. The matrices of the fi rst eight samples are more or less heterogeneous, with the greatest variation found in the boundstones (samples 3 and 4), and progressively

sample

facies

1. Koryfasion

SMF 3. Toe of slope (planktonic/benthic packed biomicrite)

2. Koryfasion

SMF 3. Toe of slope (planktonic/benthic packed biomicrite)

3. Palace of Nestor angularia

SMF 7. Reef (shallow water algal boundstone)

4. Palace of Nestor angularia

SMF 7. Reef (shallow water algal boundstone)

4a. Palace of Nestor angularia

SMF 3. Toe of slope (planktonic/benthic packed biomicrite)

5. Ancient quarry I

SMF 5. Reef f lank (poorly sorted, brecciated upslope-biomicrite assemblage)

6. Ancient quarry II

SMF 5. Reef f lank (poorly sorted, brecciated upslope-biomicrite assemblage)

7. Ancient quarry III

SMF 4. Foreslope (well sorted, brecciated upslope-biomicrite assemblage)

8. Cut block from quarry IIIb

SMF 8. Back-reef (dominantly fusilinid biomicrite)

9. Modern quarry

SMF 7–8. Reef or back-reef (biotically diverse, dominantly benthic biomicrite)

10. Modern quarry

SMF 7–8. Reef or back-reef (biotically diverse, dominantly benthic biomicrite)

11. Modern Pylos

SMF 8. Back-reef (miliolid lime mud)

12. Modern Pylos

SMF 8. Back-reef (miliolid lime mud)

Table 5.3 Standard microfaces assignments.

185

Distler – Bronze Age Quarrying: A Provenance Study less for samples 1, 2, 5–7 and 4a. These matrices are composed of micritic cement surrounding bio- and lithoclasts and engulfed by later sparry cements. They tend to have iron oxide concentrations, if only on pore rims, and contain medium to small amounts of pore space. Samples 8–12 consistently have no pore space at all and very homogeneous micritic cements. Though samples 9, 10, 11 and 12 have thin or medium-sized veins of crystalline calcite (relative to the other samples), there is no sparry component to the cements of any of the last five samples.

Major Element Geochemistry The measure of a feasible ‘fingerprinting’ technique for determining stone provenance is based on consistency within a quarry outcrop but distinctiveness relative to a larger area. The careful compilation of paleontological and petrographic observations for a marble artefact can potentially produce such a characterisation, but it is not infallible. Because of this, geoarchaeologists have applied a number of geochemical techniques to provenance studies in an effort to refine lithic characterisations and provide supporting data from various analytical methodologies. Indeed, many provenance studies (e.g. Moens et al. 1988; Germann et al. 1988) have emphasised the necessity of such a multi-variate approach for producing a reliably distinctive data set. For marbles, the auxiliary tests include such analyses as grain size, major and trace element abundances,4 stable isotope ratios, and pore space studies. Limestones tend to have greater variations than marbles for many of these attributes even over quite short distances (N. Herz, personal communication), since the rock is relatively uncrystallised and chemically unequilibrated. This trait makes determining provenance for limestones more difficult than marbles, but the disparity is tempered by the option of paleontological analyses for the former.

Abundances of terrigenous sediments such as quartz grains partition the samples in the same way. Samples 1–7 all exhibit significant lithoclastic contents, while samples 8–12 are conspicuously lacking in terrigenous lithic clasts. Samples 1–7 showed both inter- and intragranular porosity, and several cementing events. The intergranular cements vary from micrite (probably primary algal or chemically precipitated) to spar to isopagous calcite crystal layers. The isopagous crystals rimming pores indicate shallow-marine diagenesis, while the larger, equant sparry grains fi lling veins and pores are probably the result of later diagenesis (neomorphism) involving meteoric (fresh) water. Groundwater tends to have higher concentrations of iron, so the higher proportions of iron oxide seen in samples 1–7 would suggest a significant degree of diagenesis in fresh water. It is possible that samples 1–7 have experienced shallow-seawater or marine phreatic diagenesis as well as neomorphism in a freshwater sub-water table, or meteoric phreatic environment. Samples 8–12 would appear to have experienced a less diverse series of diagenetic conditions, though the increased hardness and compaction of pore space would imply a significant amount of total diagenetic alteration. They lack the complex cementation and pore space that are integral to the matrices of the first seven samples.

The method employed in this study for major element analysis involved mounting carbon-coated thin section slides from samples used for the previous lithic analysis in a scanning electron microscope (SEM), and collecting major element analyses from calcite grains throughout the slide. Relatively large grains of at least 5 μm in diameter were selected for elemental analysis, and 20–25 trials were conducted on each sample slide to minimise the statistical significance of anomalous measurements. The SEM records the abundances of major elements to an accuracy of about 1000 parts per million (ppm) using energy-dispersion X-ray spectroscopy, sometimes referred to as electron microprobe analysis. The energy-dispersion spectrometre measured abundances (atoms/formula unit) for 12 of the major elements: silicon (Si), aluminum (Al), titanium (Ti), magnesium (Mg), iron (Fe), manganese (Mn), calcium (Ca), sodium (Na), potassium (K), zinc (Zn), chlorine (Cl), and oxygen (O). Since the limestone was assumed to be nearly pure carbonate, the formula unit chosen was Ca2(CO3)2. In a pure limestone, only calcium would be measured by the spectrometer, but in most cases impurities migrate into the crystal structure of carbonate

Finally, the samples can be divided into the same two basic groups based on hand sample examination. Shaw’s categories of Minoan limestones note the differences best: ‘One of them is a soft , even-grained stone, called ‘poros’ or ‘porolithos’, with colour ranging from yellowish white to light grey… Most ashlar walls were built with blocks of poros limestone, and occur frequently at various sites such as Hagia Triadha, Phaistos, Arkhanes, Vathypetro, Nirou Khani and Knossos… Another type of limestone, quite different from ‘poros’, ranges from greyish to dark-blue or even black in colour, and is much harder… Locally this stone is called ‘sidheropetra’ (ironstone) because of its hardness… its more usual designation being ‘asbestolithos’, a term used by villagers as well since the stone, when burnt in a kiln, makes excellent lime… for building’ (Shaw 1973, 13). Not only do the stone descriptions match precisely, but the ashlar blocks at Pylos seem much closer to Shaw’s description of poros than to sidheropetra. In addition, typographic maps of the Pylos region show three lime kilns not far from the collection site of samples 11 and 12.

4

The distinction between major and trace elements is usually determined by an arbitrary abundance threshold of more than 100 parts per million for major elements. However, in some instances the abundance of an element refers to its global abundance, while in others the term is used to refer to its abundance in a particular sample. Thus, while iron (Fe) is considered a major element in a global frame of reference, in some studies it is termed a ‘trace’ element since it occurs in very small quantities in those particular samples. For convenience, this study uses global standards.

186

Major Element Geochemistry grains. These impurities commonly include magnesium and strontium, as well as clay minerals (which can contain concentrations of silicon, sodium, magnesium etc.). In the measured samples, there were occasional trace amounts of aluminium, sodium, iron and potassium, but these occurrences were not widespread enough to allow meaningful inter-sample comparisons. Calcium, of course, consistently formed an extremely high proportion of the elemental composition, and these data were plotted for each sample to gauge the variation. Magnesium and silicon also were recorded in consistently measurable quantities, and these three constituents provided the major element data set for these samples. Graphs 5.1–5.3 illustrate the variation in abundances (atoms/formula unit) for calcium, magnesium and silicon versus the standard deviation (the variation within one sample).

Graph 5.1 Mean values for calcium and standard deviation for each sample.

Not surprisingly, variation of calcium abundances between samples is quite small, within 0.25 atoms per formula unit; the range of values within one sample, represented by the standard deviation, is also slight. Magnesium presents a more varied spectrum of values. Sample 3 has anomalously low magnesium values: samples 1–2 are about 0.005 atoms per formula unit more and are similar in abundance, whereas 6, 9, 8, 11 and 10 have gradually increasing abundances, with 8 and 9 especially similar. Samples 5 and 4 show still greater increases, and there is a gap of ~0.01 atoms per formula unit between these and the most magnesium-enriched samples, 7 and 12. Standard deviations are large relative to the measured mean values in all cases, indicating varying values within a sample; the range in values is significant, c. 0.01–0.32. Finally, silicon exhibits a trend that is more constant and, to a significant degree, consistent with petrographic evidence. Samples 1–5 all exhibit significantly higher silicon mean values and standard deviations than the other samples, with peak values for samples 2 and 5. Samples 6–12 all have lower signatures, though 6 and 7 show the highest values in this group, both above 0.0005 atoms per formula unit. Samples 9 and 10 are quite close to these levels of abundance, however, and the distinction is difficult to prove.

Graph 5.2 Mean values for magnesium and standard deviation for each sample.

The ratios of calcite to magnesium in calcite grains for all the samples imply that there is very little magnesium substitution in the crystal structure. This fact, in conjunction with very close correlations in absolute calcium abundance, makes the sample set appear quite homogeneous. Relative magnesium abundances do vary significantly between samples, however. It was hoped that magnesium abundances would be useful in identifying signatures for a locality, but there are too many genetic variables in the sampled limestones for this factor to prove reliable. It is quite possible for rocks with slightly different depositional histories but similar diagenetic histories or vice versa to exhibit differing magnesium abundances.

Graph 5.3 Mean values for silicon and standard deviations for each sample.

The silicon content of the samples represents an externally derived impurity in the carbonates (it is not precipitated

187

Distler – Bronze Age Quarrying: A Provenance Study from sea water, nor do the shallow water organisms observed secrete it). The silicon signatures of the samples, whatever their cause, indicate differing depositional environments for the two groups previously identified. With the exception of sample 9, samples 8–12 all exhibit quite low clay contents, whereas samples 1–7 are more enriched (samples 1–5 significantly so). The low silicon contents for samples 6 and 7 do not match the trend as well as they were expected to do, which is especially worrisome since sample 5 shows a remarkably high silicon signature (samples 5, 6 and 7 are virtually adjacent.). Nevertheless, silicon signatures seem to be the most helpful geochemical signature studied in terms of distinguishing the sample groups, though as previous studies have shown, major/ trace element analysis is insufficient as a sole guide to provenance assignment.

each isotope by causing either a lighter or heavier isotope to be preferentially assumed into the solid phase. The cumulative effect of these fractionation processes ultimately determine the isotopic signature for a particular limestone. The partitioning of oxygen isotopes in marine limestone is controlled by several fractionation factors, including the concentration of isotopes in seawater during precipitation; compositions of later, pore waters; degree of diagenesis; temperature of depositional conditions; the volumetric contribution of biological processes (e.g. shell production) and salinity (Faure 1986, 460–67, 496–98). The heavier oxygen isotope will be preferentially precipitated in conditions of high salinity, high chemical precipitation rates (versus organic production), small degrees of diagenesis and low temperatures. Because of these fractionation trends, deeper marine environments will produce carbonates with the highest 18O/16O ratio. Warmer, fresh water (meteoric) environments and greater degrees of diagenesis will produce limestones with significantly lighter oxygen δ-values. For carbon found in marine limestones, the variation is not as great as oxygen. Most significant is the preferential fractionation of the heavy (13C) carbon isotope in seawater relative to fresh water. δ-13C is also enriched in organically produced deposits (especially when foraminifera-rich), and so carbonate produced by organic processes such as shell secretion has a higher δ-13C value than chemically precipitated limestone. ‘Sedimentary’ carbonate, or carbonate formed by chemical precipitation out of seawater, exhibits lower δ-13C values due to enrichment in the lighter carbon isotope from dissolved atmospheric CO2 (Hoefs 1987, 132– 33, 141–42, 170–79).

Isotope Analysis Isotope analysis was conducted as a means of characterising the samples relative to one another and other limestones. Before engaging in a discussion of isotopic analysis results, it may be useful to explain isotopes and the factors that control their abundances. the science of isotope analysis Each element on the periodic table has a unique atomic structure, differing from others in the number of protons in its central nucleus and the number of electrons orbiting this nucleus. Differences in the number of protons in the atomic nucleus will result in a different element; differences in the number of neutrons in combination with a steady number of protons result in different isotopes of a single element. The various isotopes of a given element have identical electrical charges and very similar atomic radii, and because of this each will act chemically like the other isotopes of that element, for the most part. The atomic structure of calcium carbonate (CaCO3), for instance, can contain both the heavy oxygen isotope that has two extra neutrons (18O) as well as a lighter isotope (16O) and two isotopes of carbon, 13C and 12C. It is the relative abundances of these isotopes in limestone (calcium carbonate) that figure in provenance studies.

Diagenesis is an extremely variable process, occurring in both marine and fresh water zones, and isotopic variation in diagenetic carbonates reflects this setting. Cystalline vein-fi ll tends to show lighter values for both carbon and oxygen isotopes, and late-formed calcite grains show lighter values for both as well (Dickson and Coleman 1980). Such vein-fi ll is probably the result of meteoric diagenesis (Tucker and Wright, 1990). Diagenesis involving meteoric water will greatly reduce δ-18O and δ-13C values, about 4 for carbon and 2 for oxygen (Tucker and Wright 1990). For marine diagenesis, original carbon isotope values are retained, but oxygen isotope signatures become lighter. However, if there are concentrations of fibrous calcite or calcified aragonite crystals, the most common cement in Phanerozoic reefs, a covariant trend is not uncommon. This trend, recorded by Given and Lohmann (1985) in cathodoluminescence studies of Permian calcite cements, shows increasing depletion (recorded on a δ-plot as a positive slope) of both heavy carbon and oxygen isotopes as luminescence increases (more recent crystal growth). Other studies showed late cements and zoned, iron-rich calcite spars have more negative δ-18O and δ-13C values (Tucker and Wright, 1990).

The abundances of these stable isotopes are plotted in per mil (‰) relative to a standard, which in the case of carbonate provenance studies is traditionally the Pee Dee Belemnite (PDB), a fossilised carbonaceous marine mollusc from South Carolina. Positive δ-values indicate isotopic enrichment relative to PDB, while negative values indicate depletion. While isotopes of a given element react very similarly in some ways, there are processes, termed ‘fractionation factors’, which do affect isotopic abundances in certain situations (i.e. carbonate precipitation and diagenesis). These factors partition the relative concentrations of

188

Isotope Analysis the study

The effect of these sensitivities is that marine limestones that formed at slightly different times and under just slightly different conditions may exhibit different ratios of isotopes reflecting their genetic and diagenetic variation. The ratio of the measured oxygen isotope to the measured carbon isotope for a given site has been shown to produce a particular and reproducible ‘fingerprint’ of the rock. Extensive research (Craig and Craig 1972; Herz 1987, 1988; Manfra et al. 1975) has proven that these isotopic characterisations are localised to the distance covered by an average quarry outcrop but remain fairly constant within that outcrop, and so have an extremely convenient scale.

Isotopic measurements for this study were conducted at Dartmouth College in January 1993, after the technique of J.M. McCrea (1950). Approximately 10–15 mg of rock powder was drilled from the centre of fresh faces cut on each of the 12 samples, with attention to avoiding megascopic impurities and weathered surfaces. For each sample, the powder was then placed in one leg of a two-legged ‘y’-beaker, and three eyedroppers of 100% phosphoric acid was poured into the opposite leg. The beakers were sealed with valved caps and tilted to allow the acid to react with the carbonate. The reaction was placed in a 25°C water bath for 12 hours to insure complete reaction. Sample beakers were then attached to an evacuated carbonate line and their valves opened. The CO2 gas alone was frozen by means of a liquid nitrogen bath, isolated, then released into the mass spectromet for analysis. This machine accelerates the gas molecules along a curved tube by means of electromagnets acting on the particles’ ionic charges. The mass of the particles, which can be directly interpreted as their isotopic number, affects how much their individual trajectories are curved (heavier isotopes will not be as strongly curved,

Since early work by Craig and Craig (1972), determination of marble provenance using stable isotopes of carbon and oxygen has come to be regarded as one of the most reliable methodologies. Limestones, on the other hand, tend to have higher amounts of clay impurities, possibly producing more heterogeneous isotope values. However, the possibility of determining limestone provenance using isotopic analysis has been explored by Werner and Herz (1992), with positive results. In both cases, isotopic abundances are measured in the same way.

Graph 5.4 Delta-13C vs. -18O plot for limestone samples; note separate fields for samples 1–7 and 8–12.

189

Distler – Bronze Age Quarrying: A Provenance Study sample

carbon

do not affect bulk composition greatly). Thus, isotopic signatures should reflect depositional and later diagenetic processes, not local compositional differences. Indeed, the tight grouping of separate trials from a single sample and all three geographically-separated ancient quarry samples show that the chemical impurity of limestone relative to marble has not affected these results especially. Based on this, the general observation that δ-values remain fairly constant for marbles within one quarry site will be applied to limestones here for the purpose of delineating grouping fields. Comparison of these data with results published by Wenner and Herz (1992) shows these δ-values to fall within the observed range for Tertiary limestones; moreover, distances that are deemed statistically sufficient by the same authors to determine separate fields are easily met by the present data (Wenner and Herz divide fields separated by differences in δ-values of approximately 2, versus fields separated by distances of 3 for these data).

oxygen

1

-4.80

-2.91

2

-6.21

-4.85

3

-4.60

-3.59

3A

-5.80

-3.42

4

-8.19

-5.29

5

-4.06

-3.41

6

-3.56

-2.59

7

-3.46

-1.77

8

0.20

-1.93

9

0.11

-4.46

10

0.45

-0.75

11

1.68

12

0.73

3.24

The upper, high δ-13C field, including those samples from the modern quarries and Pylos, shows little variation in carbon isotope signatures from that of seawater, but a range of normal to depleted oxygen values. It is not uncommon for normal marine carbonates to display a range in oxygen values and fairly constant carbon isotope ratios, depending on the sea water equilibrium (Steve Dunn, personal communication). The trend for my highcarbon field matches the field observed by Anderson and Arthur (1983) for ‘average marine limestones’, displaying a range of values in oxygen from about 0 to -10 and a considerably smaller range (0 to +2) for carbon values. The covariant positively-sloped lower carbon field, including the palace and ancient quarry samples, probably reflects the observed trends in fibrous calcite, ferrroan spar, and late cements that are present in these samples but not in the previous group. The covariant trend observed by Given and Lohmann (1985) closely matches that seen in the data presented here, and late forming cements such as those listed above are implicated. In these observations, the isotope results support the petrographic data.

-2.93

Table 5.4 Stable isotope data for 12 collected limestone samples per mil, relative to PBD.

lighter isotopes will be more curved). Strategically placed collectors at the tube’s end record the number of particles that follow each trajectory, and these numbers represent the number of atoms of each isotope. Graph 5.4 and Table 5.4 illustrate the findings from the isotope analyses. The δ-values plot in two distinct fields. One field exhibits high and relatively unvaried δ-13C values and widely variant δ-18O values, while the other field shows an approximately one-to-one positively-sloped correlation between oxygen and carbon values. In addition, while the first field holds members with greatly diverse δ-18O values, the low δ-13C field is significantly more closely packed and remains clustered in the negative oxygen values. The samples from the bluff east of modern Pylos, those from the modern quarries at Gargaliani, and the sample collected from the shaped block near the ancient quarries all fall into the first, high-carbon, variant-oxygen field. The remaining samples, including those from the palace, the ancient quarries and Koryfasion all fall into the second field. The counting statistic error for δ-13C values ranges from 0.001 to 0.007; for δ-18O values, the error ranges from 0.003 to 0.024. For both, the error would appear smaller than the symbols on the plot.

Isotope values from samples 1–7 all plotted within the low-carbon, covariant field, while samples 8–12 all plotted in the high-carbon, univariant field. Data from both samples from the palace plot within one field, and all the suspected ancient quarry outcrops (aside from the anomalous loose block found at Quarry III) plotted in the same field, remarkably close to Samples 3 and 3a from the palace. The two samples collected near Koryfasion also definitely plotted within this field, especially sample 1, which is nestled adjacent to samples 3 (palace) and 5 (ancient quarry). Thus, the isotope data supports the hypothesis that the palace blocks were quarried at the proposed quarry sites. The reason that the Koryfasion limestones plotted in this field is uncertain. The origin of the Koryfasion limestone is not known, but as there is no defined limestone outcrop at Koryfasion, transport from a distant source cannot be ruled out.

interpretations Judging from the experimental error and the close grouping of two trials run from a single sample (3 and 3a), the results seem to be quite precise. Results from geochemical analyses and petrographic studies showed the carbonate compositions for each sampled site to be fairly homogeneous and pure calcite (some are slightly more marly than others, but trace element differences

190

Conclusions It is noteworthy that all samples collected at single or adjacent sites (e.g. the two from modern Pylos, the two from the modern quarries at Gargaliani, the three from the ancient quarries, the two from Koryfasion, and the two from the palace) all plotted in the same field as their respective partners. The mysterious cut block found near the ancient quarries alone does not match this trend. Interestingly, it matches much more closely with the modern quarries, geographically not far away and petrographically more similar, thus revealing that the isotope fields are not geographically determined. The modern quarries at Gargaliani, within a kilometre of the closest of the ancient quarries and 5 km from the palace, show δ-values of about 0 and extremely varied oxygen, more similar to the values of the bluff near modern Pylos, about 18 km to the south.

iron technology for quarrying between the 8th and 6th centuries bc. The utilisation of such vast amounts of poros limestone is therefore unlikely after that point, especially with the more indurate crystalline limestones nearby, just west of Gargaliani (samples 9 and 10). The relative proximity of the proposed quarries to the palace makes them equal, if not preferred, candidates for the palace’s stone source relative to all other local limestone outcrops. Ancient and modern geographical evidence has denoted no stone workings in the vicinity of Sphacteria, the outcrops near Oxbelly Bay, or the bluffs east of modern Pylos. Since these settings provide the only other reasonably proximal limestone outcrops of sufficient scale, the Gargaliani outcrop is again a prime candidate. The distance for stone transport is still a worrisome consideration, especially in light of the precipitous terrain that separates Gargaliani from Englianos, but it appears that paleogeographic reconstructions can provide a possible solution. Estimates of Late Helladic shorelines in the Bay of Navarino and associated drainage systems suggest a possible water route for stone transport, at least for a good portion of the distance from the quarries.

Conclusions Based on archaeological, petrographic and geochemical studies, the proposed ancient quarries just east of the modern town of Gargaliani seem to be highly probable sources for the limestone ashlar blocks at the Palace of Nestor. Not only does the evidence compiled for this project suggest such a conclusion independently, but many of the observations I have made also have close correlations with observations from similar stone provenancing studies.

Lithic analyses further support this hypothesis by providing evidence of correlations between stone samples from the palace and the proposed ancient quarries. Paleontological evidence link the samples from ancient quarries and Englianos more closely than with the other outcrops, and hand sample study supports this. The Palace of Nestor and the quarries are both composed of soft, yellow poros, whereas samples 8–12, from the modern quarries and modern Pylos, are harder, crystalline, greywhite sidheropetra. Major elemental analysis, especially for silicon abundances, emphasised many of these observations by dividing the samples into high- and lowclay content groups. While many of the same similarities were seen in geochemical studies, some samples produced anomalous silicon signatures, and the major element analysis would not appear to be a reliable provenance technique when used alone.

From an archaeological standpoint, the quarries are strong candidates for Bronze Age workings. The tiered, horizontal excavations, replete with piers, recessed niches and loading platforms in succession along an outcrop have direct analogues in some of the Minoan ammoudha quarries such as Ta Skaria. Tool traces match the repertoire of traces produced by Bronze Age quarrying equipment, notably punches and pointed picks. The lack of any surviving approach road implies that the quarries have not been active for a significant length of time; moreover, the more ephemeral construction of early (Bronze Age and Archaic) transportation systems relative to the more elaborate (occasionally even paved) Hellenistic and Roman approach roads may support a Mycenaean origin for the cuttings. Estimates of the total stone production for these quarries are easily on the same scale as Minoan quarries, as well as being sufficient to supply the needs of the palace. Finally, the height of blocks excavated from the quarries, as interpreted from channeling marks, matches the average block dimension at the palace.

Stable isotope analyses produced the most distinctive sample groupings. Isotope ranges matched ranges observed by Herz and others, and variation from normal marine carbonate appeared by be caused by differing depositional and diagenetic factors. Samples plotted into two clearly distinct groups, one of which (samples 1–7) showed close correlation between palace and quarry samples. The other (samples 8–12) showed greater oxygen isotope variation and less diagenetic influence.

The rock type itself is evidence for Bronze Age or slightly later quarry origin. The friable, cream-yellow limestone that composes the outcrop is one of the softest and most easily worked (with relatively soft bronze tools) stones that is still capable of maintaining its form fairly well under the forces of weathering and compression stresses to which architectural elements are subjected. Its friability reduces its desirability, however, and for this reason the use of poros rapidly decreased after the advent of

Besides similarities between the palace and the proposed ancient quarries, samples from Koryfasion commonly were grouped together in many of the studies. This, compounded with the fact that Koryfasion samples were collected as loose float, not in outcrop, leads to the conclusion that these stones could represent foreign rock. Especially in view of recent archaeological finds at

191

Distler – Bronze Age Quarrying: A Provenance Study Koryfasion, it seems possible that these fragments were transported from the Gargaliani quarries, just as the limestone from Englianos. The cut block, which I had thought to be connected with the ancient quarries where it was found, matches petrographically, paleontologically and geochemically almost exactly with sample 9, from a nearby modern quarry.

Kraft, J. C., Rapp, G. R., Jr., and Aschenbrenner, S. E. 1975. ‘Late Holocene Paleogeography of the Coastal Plain of the Gulf of Messenia, Greece, and its Relationships to Archaeological Settings and to Coastal Change’, Geological Society of America Bulletin 86, 1191–208. Kraft, J. C., and Aschenbrenner, S. E. 1980. ‘Late Holocene Paleogeomorphic Reconstructions in the Area of the Bay of Navarino: Sandy Pylos’, Journal of Archaeological Science 7, 187–210.

Bibliography

Levi, P., trans. 1971. Pausanias: Guide to Greece. Penguin, Harmondsworth.

Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I. The Buildings and their Contents. Princeton University Press, Princeton, NJ. Craig, H., and Craig, V. 1972. ‘Greek Marbles: Determination of Provenance by Isotopic Analysis’. Science 176, 401–03.

Loy, W. G. 1967. The Land of Nestor: The Physical Geography of the Southwestern Peloponnese, Report 34. Washington, DC: Foreign Field Research Program, National Research Council.

Dickson, J. A. D., and Coleman, M. L. 1980. ‘Changes in Carbon and Oxygen Isotope Composition during Limestone Diagenesis’. Sedimentology 27, 107–18.

Manfra, L., et al. 1975. ‘Carbon and Oxygen Isotope Ratios of Marbles from Some Ancient Quarries of Western Anatolia and their Archaeological Significance’, Archaeometry 17, 215–21.

Driessen, J. 1984. ‘Notes on Building Materials and Quarries’, in ‘An Archaeological Survey of the Roussolakos Area at Palaikastro’, ed. J. A. MacGillivray and L. H. Sackett et al., Annual of the BSA 79, 143–49.

McCrea, J. M. 1950. ‘On the Isotopic Chemistry of Carbonates and a Paleotemperature Scale’, Journal of Chemical Physics 18, 849–57. McDonald, W. A., and Rapp, G. R., Jr. 1972. The Minnesota Messenia Expedition: Reconstructing a Bronze Age Regional Environment. University of Minnesota Press, Minneapolis.

Faure, G. 1986. Principles of Isotope Geology, 2nd edn. Wiley & Sons, New York.

Moens, L., et al. 1988. ‘A Multi-method Approach to the Identification of White Marbles Used in Antique Artifacts’, in Classical Marble: Geochemistry, Technology, Trade, ed. N. Herz and M. Waelkens. NATO ASI Series. Series E: Applied Sciences, vol. 153. Dordrecht, London, Boston.

Geologic Maps of Greece: Filiatra and Koroni–Pylos–Skhiza sheets. 1980. Prepared by the Institute of Geology and Mineral Exploration, Athens. Germann, K., et al. 1988. ‘Provenance Characteristics of Cycladic (Paros and Naxos) Marbles: A Multivariate Geological Approach’, in Classical Marble: Geochemistry, Technology, Trade, ed. N. Herz and M. Waelkens. NATO ASI Series. Series E: Applied Sciences, vol. 153. Dordrecht, London, Boston.

Papageorgakis, J., et al. 1992. ‘Bronze Age Quarries on the Eastern Coastal Zone of Crete (Greece)’, in Ancient Stones: Quarrying, Trade and Provenance. Interdisciplinary Studies on Stones and Stone Technology in Europe and the Near East from the Prehistoric to the Early Christian Period, ed. M. Waelkens, N. Herz and L. Moens. ACTA Archaeologica Lovaniensia 4. Leuven.

Given, R. K., and Lohmann, K. C. 1985. ‘Derivation of the Original Composition of Permian Marine Carbonates’, Journal of Sedimentary Petrology 55, 430–39.

Scholle, P. A., et al. 1983. Carbonate Depositional Environments’. American Association of Petroleum Geologists, Tulsa, OK.

Herz, N. 1987. ‘Carbon and Oxygen Isotopic Ratios: A Database for Classical Greek and Roman Marble’, Archaeometry 29, 35– 43.

Shaw, J. W. 1973. Minoan Architecture: Materials and Techniques. ASAtene 49, n.s.33, Rome.

Herz, N. 1988. ‘The Oxygen and Carbon Isotopic Data Base for Classical Marble’, in Classical Marble: Geochemistry, Technology, Trade, ed. N. Herz and M. Waelkens. NATO ASI Series. Series E: Applied Sciences, vol. 153. Dordrecht, London, Boston.

Soles, J. S. 1983. ‘A Bronze Age Quarry in Eastern Crete’, Journal of Field Archaeology 10, 33–46.

Hoefs, J. 1987. Stable Isotope Geochemistry, 3rd edn. Wiley & Sons, New York.

Topographic maps of Greece: Filiatra and Pylos sheets, 1977. Prepared by the Hellenic Army Geographical Service.

Kelletat, D., et al. 1978. ‘Neotectonics in the Peloponnesian Coastal Regions’, in Alps, Apennines, Hellenides, ed. H. Cloos, D. Roeder and K. Schmidt. Schweizerbart, Stuttgart, 512–18.

Tucker, M. E., and Wright, V. P. 1990. Carbonate Sedimentology. Wiley & Sons, Oxford.

192

Bibliography Waelkens, M. 1992. ‘Bronze Age Quarries and Quarrying Techniques in the Eastern Mediterranean and the Near East’, in Ancient Stones: Quarrying, Trade and Provenance. Interdisciplinary Studies on Stones and Stone Technology in Europe and the Near East from the Prehistoric to the Early Christian Period, ed. M. Waelkens, N. Herz and L. Moens. ACTA Archaeologica Lovaniensia 4. Leuven. Werner, D. B., and Herz, N. 1992. ‘Provenance Signatures for Classical Limestone’, in Ancient Stones: Quarrying, Trade and Provenance. Interdisciplinary Studies on Stones and Stone Technology in Europe and the Near East from the Prehistoric to the Early Christian Period, ed. M. Waelkens, N. Herz and L. Moens. ACTA Archaeologica Lovaniensia 4. Leuven. Wilson, J. L. 1970. ‘Depositional Facies across Carbonate Shelf Margins’, Gulf Coast Association of Geological Societies Transactions 20, 229–33. Wilson, J. L. 1975. Carbonate Facies in Geologic History. Springer Verlag, New York.

Revised 2013

193

6 DISCARDED CHIPPED STONE FROM THE PALACE OF NESTOR george otto marquardt

Most of the more than 100 fragments of chipped stone recovered by MARWP were unearthed in 1994 and 1995 from the deep pits within the walls of the Northwest Area, where Blegen’s team discarded tons of pottery (see p.82 above).1 They include blades, cores, waste and retouched flakes, burins, spokeshave/notches, sickle elements, bifaces, end- and sidescrapers and projectile points. The raw materials vary from a multitude of colourful chert/siliceous quartzes, to at least two and probably four varieties of obsidian, jasper, chalcedony and rock crystal.

longitudinal axis are a readily identifiable diagnostic attribute of such pieces. Prismatic blade-making leaves very characteristic negative attributes – flake scars – upon the cores from which the implements are removed. Sickle elements (also called scythe denticulates), used for harvesting grasses or the like, are the single best represented typological form of stone implement found by MARWP, and their frequency of occurrence suggests on-site manufacture. A Bronze Age date is suggested both by Blegen’s preservation of a number of identical artefacts (e.g. PN I, pls 308:2, 319:6, 320:12) and by their presence in Bronze Age layers at nearby Malthi Dorion, Peristeria, Nichoria, Tragana, Arcadian Asea and elsewhere (as displayed in the museums at Pylos and Chora).

The identifications of tool type offered here are, where possible, morphologically based and represent no attempt at strict typological seriation, since to do so would be to impose a false cohesion on artefacts drawn from a secondary context. Nevertheless, certain physical characteristics are used to differentiate the individual lithics within their material categories, particularly the distinction between pressure and percussion manufacture of blades.

Prismatic cores in brown-maroon, yellow-orange, greygreen and red cherts also occur frequently enough in the MARWP assemblage to posit a Late Helladic date for them. The brownish-maroon chert is the most common variety found by MARWP, with 18 worked pieces having been catalogued and numerous unworked examples recovered. The wide variety of chert types and as many as four demonstrably distinct varieties of obsidian may suggest some variation in date for the lithics, even those of closely parallel forms and material. In addition to the most common Melian type of obsidian – jet black and highly glossy, with white streaking that can be seen on blades with semi-translucent edges – other examples include a greyish streaked obsidian, extant in only a few examples (e.g. 1851); a more frequent, but still relatively poorly represented dull black, unstreaked material (e.g. 1870); and an obsidian with an unusual red patch present in the otherwise streakless, jet surface of high gloss, which appears uniquely in the small cobble 2356.

The term ‘crested blade’ denotes an irregularly edged, retouched or unretouched blade of chert or obsidian, triangular in cross-section, produced by direct percussion and/or splinter fracturing. In all cases at least part of a percussion bulb is present. Such blades are usually dated to the Late Neolithic. The term ‘prismatic blade’ refers to a parallel edged blade having a trapezoidal cross-section resulting from an anvil or pressure manufacturing technique. The flat and concave surfaces along the

1

Th is study is dedicated to the memory of Andre Bekermann. For assistance and guidance in identifying and conserving the lithics I thank Edward Silver and Marc Walton. Gratitude is also extended to Curtis Runnels and Nikos Kardulias for advice on documentation and attribute identification. For help in thin-sectioning, I thank Paul Weiblen of the University of Minnesota Geology Department.

The lithics catalogued below are representative of the finds recovered by MARWP, not comprehensive. All are reproduced in the drawings at 100 percent original size. 195

Marquardt – Discarded Chipped Stone from the Palace of Nestor having no cortex. There is a large bulb of force visible on the ventral face. In an elongate fashion the dorsal crest divides into two crests that extend to the terminus, which occurs in an unbroken distal tip. The cortex consists of a thin, whitish layer. There is no usewear evidenced on the edges of the segment, but dulling and some crushing of the distal tip suggests possible reuse as a burin or punch. The opposite terminus is a clean snap. Numerous concentric, rippled striations are present on the exposed ventral side, some of which are deeply grooved, despite the piece’s small size.

Catalogue Obsidian Crested blades 1870. Crested blade. Munsell: unmatched. Pantone: S-327-2. H. 26 mm, w. 8 mm, th. 7 mm. Mohs hardness: untested. A microblade manufactured by percussive detachment, as is shown by the presence of a small percussive bulbar scar at the proximal end. Triangular in cross-section, the object is of jet black obsidian, rather dull in appearance, and bears traces of usewear consisting of directional incisions perpendicular to the lengthwise axis of the blade.

2383. Crested blade with usewear. Munsell: N-3 Medium Dark Grey. Pantone: S-327-2. H. 20 mm, w. 9 mm, th. 2 mm. Mohs hardness: 6.5.

1877. Crested blade with usewear. Munsell: unrecorded. Pantone: S-325-2. H. 34 mm, w. 14 mm, th. 7 mm. Mohs hardness: untested. An intact blade divided longitudinally on its dorsal face by an axial crest. There is a considerable amount of usewear on one of the blade edges. Triangular in cross-section. The material is the jet black variety, having whitish streaks and semi-translucent edges. The gloss has been dulled by use. Manufactured by pressure, despite its crested type, which is evidenced by the remnant of the core’s striking platform still present on the proximal end and the lack of any percussion scar, bulbar or otherwise.

A blade of the crested, percussion-made variety, having directionally incised parallel scratches perpendicular to the piece’s long axis, consistent with usewear. It is divided on the dorsal by an axial crest and is triangular in crosssection. There is also a good deal of cortex remaining on the unused edge, opposite that one bearing the incised usewear.

2405. Retouched distal segment of crested blade. Munsell: N-3 Dark Grey. Pantone: S-327/6-1. H. 24 mm, w. 18 mm, th. 5 mm. Mohs hardness: untested.

2348. Mesial segment of crested blade. Munsell N-2 Greyish Black. Pantone: S-325-1. H. 11 mm, w. 8.5 mm, th. 1 mm. Mohs hardness: untested. An irregularly edged blade, with the hinged longitudinal crest typical of such pieces extending its axis. The edges are roughly parallel, and draw to a lunate terminus at the distal end. Any breakage towards the (now absent) proximal end has been obscured by crushing. This is quite possibly post depositional, however, and there is no clear indication of usewear. The central hinge extends straight to the midpoint of the distal terminus, without any curvature. There is knapped retouch present, consisting of serration by means of flake removal, evidenced by flake scars in series, conventionally termed channel scars, that extend to the distal tip down the length of the artefact in an arc. It bears traces of a bulb of force, produced likely by conchoidal intentional fracture at the pre-form stage, before use as a blade and subsequent retouch. Parallel concavities characterise the crest-to-edge span of the artefact symmetrically in cross-section. The material is glossy, jet black streaked obsidian, probably Melian.

Triangular in section, of jet black, glossy material. Whitish streaking towards the edges indicates a probable Melian provenance. The streaks extend outward (perpendicular to the vertical axis). 2351. Distal blade fragment of crested blade. Munsell: N-2 Greyish Black. Pantone: S-326-3. H. 14 mm, w. 11 mm, th. 14 mm. Mohs hardness: untested.

Cortex is largely intact on the dorsal face. Parallel-sided, probably tip to a crested blade. The obverse is smooth,

196

Catalogue 2392. Retouched distal segment of crested blade. Munsell: N-3 Dark Grey. Pantone: S-327-1. H. 20 mm, w. 12 mm, th. 1–2 mm var. Mohs hardness: 7.

be considered evidence of a blade–core industry. The material is a glossy jet black, probably Melian. Widely paralleled locally at tomb and settlement sites. 2407. Mesial segment of prismatic blade. Munsell: N-3 Medium Dark Grey. Pantone: S-327-1. H. 20 mm, w. 17 mm, th. 9 mm. Mohs hardness: 3.5.

A smooth-backed, thin segment of a percussion-produced blade, having later been worked by percussion into a lunate shape at the distal terminus, which is intact. Retouch is visible to the eye over 180 degrees of the distal perimeter. There is no remnant of a tabular platform anywhere towards the now absent proximal terminus. A skewed crest/hinge runs the length of the piece; cross-section is triangular. Hammer blows are indicated by the lack of any indication of a platform for pressure working. Concentric striations radiate outwards over the entire piece. The obsidian is a lustrous dark grey; no white streaking visible, nor any usewear in evidence.

Trapezoidal in cross-section, having three planes defined by longitudinal fractures on the exterior (dorsal), and a smooth ventral reverse. Both ends are truncated at oblique angles to the main fracture planes. There is cortex present along one of the prismatic planes. Bifaces

Prismatic blades

2338. Biface with usewear. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 26 mm, w. 16 mm, th. 8 mm. Mohs hardness: untested.

1861. Proximal segment of prismatic blade. Munsell: unrecorded. Pantone: S-325-2. H. 19 mm, w. 11 mm, th. 3 mm. Mohs hardness: 7.

Intact proximal segment of a prismatic blade, having a complete bulb of force created by pressure/anvil technique on the dorsal side. Semi-translucent, whitish streaking in the glossy, jet black material extends parallel to the long axis. There is a uniformly straight, even, concave platform extending the surviving length. Trapezoidal in crosssection.

Oblong in shape, with closely parallel edges, it has few percussion scars. Both faces are smooth, one bearing a bulb of force. There are small bits of cortex on the bulb and along one blade edge. Directional usewear is present along one side: parallel diagonal incisions consistent with use as a scraping instrument. These are numerous enough to form a sheen, or patina, which has blunted the affected edge. Possibly reused as a sickle element?

2378. Mesial segment of prismatic blade. Munsell N-3 Dark Grey. Pantone: unrecorded. H. 22 mm, w. 17 mm, th. 5.5 mm. Mohs hardness: untested.

2386. Biface. Munsell: N-2 Greyish Black. Pantone: S-325-. H. 22 mm, w. 17 mm, th. 6 mm. Mohs hardness: untested.

Smooth backed and unevenly truncated, with a trapezoidal cross-section, preserving a 9-mm wide concave platform along the longitudinal axis, which tapers steeply to the edges. No cortex in evidence. A small convex percussion scar swells the dorsal side. Typologically prismatic, may

Not percussively flaked to the extent of 2388 or 2403 below, it is also largely scarred over its entire surface. No evident bulb of force. It is of jet material, having no whitish streaks, and one of the two broad faces of the artefact has an area free of scarring, which forms a flat central platform from which pressure and percussion scars radiate. Flakes were removed outwardly from this platform, and inward towards it from the edge as well.

197

Marquardt – Discarded Chipped Stone from the Palace of Nestor The material is not stressed or crazed, despite the evidence for heavy abuse. Pitted cortex is present in patches. The blade bears an oblique sharp cutting edge, with scores of semi-parallel linear scratches up to 10 mm in length, forming a patina not unlike that present on flake 2340 below. The original blade edge shears in from the flake’s thicker central body to form a straightedge. This area has all usewear present upon the find. Despite this usewear, the edge is sharper than most blades studied here. This can be explained by the flat straightedge, clearly formed during cutting until it widened to the point that a new flake was selected for the same purpose. While the implement retained sharpness, the stoneworker may have found it unsuited to its intended task.

2388. Biface. Munsell: N-3 Dark Grey. Pantone: S-327-1. H. 33 mm, w. 21 mm, th. 11 mm. Mohs hardness: untested.

A heavily percussion-flaked piece of the greyish black variety, shaped by microlithic working into a well-defined rectangle, nearly 360 degrees in the round. Perhaps partially worked in hope of forming a sickle element. No whitish streaking or usewear in evidence.

Flakes and chips 1849. Pressure flake. Munsell: unmatched. Pantone: S-328-1. H. 19 mm, w. 17 mm, th. 6 mm. Mohs hardness: untested.

2403. Biface. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 28 mm, w. 18 mm, th. 10 mm. Mohs hardness: untested.

A piece of obsidian with a hinge fracture present on its dorsal face, and a smooth ventral reverse. The presence of striations radiating outward from a platform at right angles to the flake’s shovel-shaped main surface at its narrower end determines this piece to have been removed by the pressure (anvil) technique.

An oblong, jet piece of obsidian, conchoidally flaked by percussion 360 degrees in the round. It was intentionally thinned, probably for use as a sickle element or as a pre-form for a different implement. On one face a flat corticular surface is framed by small percussion flake or ‘knap’ scars worked on a microlithic scale as 2388 above; it bears as well as many of the indisputable sickle elements documented below. All the scars are striated. There is no evident bulb of force, giving an idea of the extent of the stoneworking after the piece was obtained from a larger specimen.

1851. Percussion flake. Munsell: unmatched. Pantone: S-325-1. H. 17 mm, w. 16 mm, th. 6 mm. Mohs hardness: untested.

A flake of the less well represented grey obsidian, having a high natural gloss. No evidence of retouch or usewear. There is a small bulb of percussion present. May be regarded as manufacturing waste.

Undefined blade 2346. Retouched blade with usewear. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 30 mm, w. 19 mm, th. 11 mm. Mohs hardness: untested.

1862. Percussion flake. Munsell: unmatched. Pantone: S-325-1. H. 21 mm, w. 26 mm, th. 5 mm. Mohs hardness: untested.

An unretouched flake with a pronounced convex fracture scar and a flat, smooth obverse surface. There is no evident usewear. Probably waste from manufacture.

Typologically distinct in form from a prismatic blade, it bears also a prepared edge with usewear retouched upon the original blade. A large area has been scarred by repeated blows and shows several percussion cone scars.

198

Catalogue 1864. Percussion flake. Munsell: unmatched. Pantone: unrecorded. H. 20 mm, w. 14 mm, th. 5 mm. Mohs hardness: untested.

The greyish, less well represented variety of obsidian, with no apparent streaking and a low lustre. The flake bears a pronounced bulb of percussion.

Clearly retouched by many percussion blows on both faces, and intentionally shaped into the rough form of a thumbnail scraper, also somewhat notched on one side. Perhaps a pre-form or blank of some type. The material is of the jet black, glossy variety.

1867. Percussion flake. Munsell: unmatched. Pantone: S-327-1. H. 16 mm, w. 16 mm, th. 5 mm. Mohs hardness: untested.

2339. Retouched flake. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 25 mm, w. 13 mm, th. 10 mm. Mohs hardness: 4.

A simple, unretouched piece of manufacturing waste. 1869. Retouched percussion flake. Munsell: unmatched. Pantone: S-327-1. H. 25 mm, w. 14 mm, th. 3 mm. Mohs hardness: untested.

A piece of glossy, jet black obsidian with light white streaking, probably from Melos. The flake has been retouched by percussion on both faces, being trimmed into a parallel edged form with a rounded distal end. The opposite, proximal end has been evenly snapped off, probably deliberately. There is no cortex present, nor is there any evidence of usewear. Perhaps intended as a preform for further retouch, or simply abandoned. 2340. Pressure flake. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 30 mm, w. 21.5 mm, th. 3.5 mm. Mohs hardness: untested.

Another example of grey obsidian, evidently percussion flaked as indicated by the bulbar scar, and trimmed down to a rough triangular/chevron form by knapping the edges inward. Possibly a pre-form, as there is no indication of usewear. 1876. Chip. Munsell: unmatched. Pantone: S-327-1. H. 19 mm, w. 18 mm, th. 7 mm. Mohs hardness: untested.

A pressure flake with no clear axis created by forcible working, and a deeply concave structure. The convex bulb of force is visible opposite the hollow concavity that characterises the piece, with the corollary convex comprising the dorsal portion of the artefact. One of the triangular flake’s three points has clear evidence for use in the form of numerous abraded linear scratches of varying depth, consistent with the rotation of the flake in use as a punch or burin. Identification as such is not possible, however, due to further directional usewear also found on the edge opposite the worn corner of the overall triangular form. Numerous linear scratches of variable depth lie perpendicular to the edge, forming an overall patina from continual or repeated use. The other two edges are free of any such wear, remaining glossy. The material is of the

Simple waste from percussion fracturing of a larger piece. The material is jet black with grayish-white streaking. 1878. Retouched flake (bifacial). Munsell: unmatched. Pantone: S-327-1. H. 24 mm, w. 22 mm, th. 8 mm. Mohs hardness: untested.

199

Marquardt – Discarded Chipped Stone from the Palace of Nestor jet variety typical for the site, with much visible streaking and semi-translucent edges; probably from Melos.

Glossy black obsidian with whitish streaking in the glass and semi-translucent thinner edges, this example is a flake produced by pressure in an attempt to make blades by the anvil technique. The result is that this thin, short fragment has attributes of a prismatic blade, such as a trapezoidal cross-section and concave edges, but must have been disregarded by its maker as unsuitable for a blade or sickle element. A platform is present at right angles to the main body of the flake, and a two hinge fractures running the length of the flake. There is no wear present.

2345. Pressure flake. Munsell: N-2 Greyish Black. Pantone: unrecorded. H. 16 mm, w. 13 mm, th. 5 mm. Mohs hardness: untested.

An example of an obsidian flake with no scarring from percussive retouch.

2396. Percussion flake. Munsell: N-1 Black. Pantone: S-327-1. H. 18 mm, w. 16 mm, th. 5.5 mm. Mohs hardness: untested.

2354. Chip. Munsell: N-2 Greyish Black. Pantone: S-325-2. H. 14 mm, w. 10 mm, th. 7 mm. Mohs hardness: untested.

Another unretouched piece of the glossy black, whitestreaked variety. A simple piece of manufacturing waste, in the jet black, glossy variety of the material.

2399. Pressure flake with possible retouch. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 23 mm, w. 19 mm, th. 4.5 mm. Mohs hardness: untested.

2360. Pressure flake. Munsell: N-1 Black. Pantone: S-329-1. H. 25 mm, w. 16 mm, th. 7 mm. Mohs hardness: untested.

Roughly triangular, jet-black, glossy flake of obsidian. There is no remaining evidence of a bulb of force. One side is smooth, the opposite is scarred and uneven. No apparent pattern to chip scarring, other than its presence at the narrowest of the flake’s three points. Striations visible in the flake scars and on the piece overall.

The determination that this flake was removed by the pressure (anvil/billet) technique rests on the lack of a percussive bulb, and the presence of a remnant of a platform at right angles to the main body of the flake. The surface of its main area bears no concentric rippling or striations. The flake is of the jet black, glossy material, without any whitish streaking.

Cores 1857. Prismatic core. Munsell: unrecorded. Pantone: S-325-2. H. 45 mm, w. 29 mm, th. 14 mm. Mohs hardness: untested.

2362. Chip. Munsell: N-2 Greyish Black. Pantone: S-325-4. H. 14.5 mm, w. 12 mm, th. 7.5 mm. Mohs hardness: untested.

2382. Pressure flake. Munsell: N-2 Greyish Black. Pantone: unmatched. H. 24 mm, w. 11 mm, th. 2 mm. Mohs hardness: untested.

Numerous flake scars radiate from a platform over the surface of this anvil/pressure-worked core. The core was clearly exhausted for blades and usable flakes, and

200

Catalogue subsequently fractured to obtain more pieces, probably for retouch into sickle elements. Of the typical glossy, jet variety. No cortex present. 2367. Core. Munsell: N-3 Dark Grey. Pantone: unrecorded. H. 50 mm, w. 43 mm, th.25 mm. Mohs hardness: untested.

Consistently homogeneous example of a medium-sized cobble of obsidian. Roughly the shape of a broad-based narrowing triangle, it has cortex on two faces; the third is smooth, a grey-black, less glossy obsidian, streaked but nowhere translucent. Along the parting planes can be seen highly reflective seams and small flake scars. The piece is apparently intact, with no sign of deliberate working. A cobble of almost completely unworked obsidian, laden with a thin layer of cortex. It bears a single pressureremoved scar, 40 mm long. The overall unreduced condition secures the identification as a cobble, as well as the proportion of surface cortex. Trapezoidal when viewed three dimensionally, each flat surface facet offers an unreduced striking platform.

Burin 1872. Burin. Munsell: unrecorded. Pantone: S-329-1. H. 29 mm, w. 13 mm, th. 29 mm. Mohs hardness: untested.

Cobbles 2356. Cobble. Munsell: N-2 Greyish Black. Pantone: S-327-1. H. 21 mm, w. 20 mm, th. 10 mm. Mohs hardness: untested. The percussion scar is elongated on the longitudinal axis of the lithic, consistent with having been worked by the billet/anvil technique or another pressure fractureinducing technique. The edges taper outward from the distal end uniformly and nearly symmetrically, and come to a distal point. The thicker proximal bears traces of originally having been part of a broader platform. The material is a streakless jet with a very high gloss. There is no evident usewear on the edges, but crushing and a crazed area towards the distal point indicate the use of this snout.

An ovoid spherule, sliced to expose a cleanly cut smooth cross-section created by an excavator’s pick. As recovered, the exterior is completely corticular, a smooth layer of grayish-white about 1 mm thick, but hard and unchalky. The obsidian is jet, with barely perceptible streaking, but this is not evidence of Melian provenience in this case, due to the presence of a bright red spot in the glass, which has not been recognised in any known Melian variety of obsidian (P. N. Kardulias, personal communication).

cherts Crested blades

2400. Cobble. Munsell: N-2 Greyish Black. Pantone: S-325-1. H. 41 mm, w. 63 mm, th. 23 mm. Mohs hardness: untested.

2373. Crested blade with retouch. Munsell: N-6 Medium Light Grey. Pantone: S-325-2. H. 27 mm, w. 11 mm, th. 10 mm. Mohs hardness: untested.

201

Marquardt – Discarded Chipped Stone from the Palace of Nestor Proximal segment of a fracture-made blade intact, with a surviving remnant of the core platform preserved. The edges taper roughly in uneven concavities out from the axial crest on one face to the blade edges. Cross-section is triangular. The obverse is a crazed convex having retouch strike fracture-made flake scars on both edges of the ventral. The pearly, even, light grey material is very hard and smooth overall and is unique among MARWP’s worked finds.

The proximal end of this blade is intact, retaining its percussive bulbar scar but lacking any trace of a core platform. It is bisected by an axial crest and is truncated towards the former distal terminus. One of the parallel longitudinal edges has several percussion fracture scars, indicating its having been intentionally trimmed. The material is a dark burgundy chert with greyish-black banding. No usewear in evidence. Prismatic blades

2387. Crested blade. Munsell: 5 R 3/4. Pantone: S-318-2. H. 32 mm, w. 12.5 mm, th. 8.5 mm. Mohs hardness: untested.

2357. Prismatic blade. Munsell: 10 YR 5/4. Pantone: S-28-2. H. 32 mm, w. 21.5 mm, th. 6 mm. Mohs hardness: 7.5.

A crested blade of the same maroonish-brown chert as the sickle element 2364. The obverse is rippled and smooth. There is a straight axial crest dividing the dorsal face, with a few small flaking scars indenting one edge. There is no distinct percussion scar. The proximal retains a portion of the core’s striking platform; the distal is truncated. Pressure-made with knapped retouch.

An unusual piece among MARWP lithics: although it is manufactured from the yellow-orange chert that is the second most well represented variety of worked stone from the backfi ll, no cream-coloured inclusions are present in this example, as with some of the cores described below. The longitudinal striations as well as the lack of a bulb of percussion and presence of a remnant of the platform from its parent core all point to its having been pressuremade. The actual blade is not present on the parallel edges, which are rounded and flare outward from the platform remnant to the wider opposite terminus, where the blade edge is located. The blade edge is wide, and slightly hemispherically curved, of a type often termed a ‘shovel’ form. There is no usewear present.

2391. Crested blade. Munsell: 5 YR 4/4. Pantone: S-316-1. H. 32 mm, w. 15 mm, th. 6 mm. Mohs hardness: untested.

2372. Prismatic blade. Munsell 10 R 4/6. Pantone: S-76-4. H. 28 mm, w. 14 mm, th. 10 mm. Mohs hardness: untested. Manufactured by percussion and showing an intact percussion bulb, this dark brownish-grey chert has a smooth, rippled reverse. Its dorsal is bisected by an axial crest, typical of such blades. There is evidence of intentional chipping in order to achieve its parallel edges. No evidence of usewear. 2395. Proximal segment of a crested blade. Munsell: 10 R 4/4. Pantone: S-319-2. H. 23 mm, w. 19 mm, th. 7 mm. Mohs hardness: 7.5. The microcrystalline structure of this red chert, of which we have several worked examples, is tabular, consisting of almost microscopically thin overlapping plates. It is extremely smooth however, and of such a hardness that despite the structural appearance, identification as shale or slate is not viable. There are dark brown bubbles throughout the stone, fi lled in and surrounded by the red chert, as

202

Catalogue well as tiny flecks of mica visible under the microscope and, occasionally, to the naked eye. A chert of remarkably similar appearance comes from Nichoria (Blizter 1992, pls 5 and 11). Trade between the Five Rivers area to the east and Pylos to the west has been suggested (McDonald and Rapp 1972). The pressure-flaked artefact has an even-sided platform or axial crest extending longitudinally from proximal to dorsal, and two concavities dipping outward in perfect symmetry from this flat platform. There is no evidence of usewear. A slight convexity, very smooth, defines the obverse.

Sickle blades

2379. Prismatic blade. Munsell: 10 R 4/6. Pantone: S-80-4. H. 41 mm, w. 8 mm, th. 8 mm. Mohs hardness: untested.

2341. Sickle element. Munsell: 5 PB 3/2 and 5 Y 6/4. Pantone: S-325-4. H. 26 mm, w. 13.5 mm, th. 5 mm. Mohs hardness: 5.5.

1874. Sickle element. Munsell: 5 YR 3/4. Pantone: S-320-2. H. 39 mm, w. 16 mm, th. 10 mm. Mohs hardness: 5.5. An atypically formed example of the denticulate serrate form of sickle element. Although lacking the lunate back opposite the serration, denticulation is present, as is sickle gloss. The maroon-brown chert is one of the most common materials at the site. (My thanks to P. N. Kardulias for this identification.)

A typical example of the most common type of chert sickle element found in Middle and Late Helladic contexts. It has the distinctive lunate hemispheric backing opposite a serrated-denticulated blade edge. Clearly manufactured by percussion flaking, with visible sickle gloss or patina on the serrated teeth. This chert is unique among MARWP finds, being a bluish-grey with whitish patches of cortex and small areas of black. It is fairly soft, almost chalky to the touch, but does not crumble or score. Sickle blades are one of the most common lithics found in the backfi ll, and are probably of LH IIIB date.

Another pressure or billet technique manufactured prismatic blade, of the variety of chert described in 2372 above. It is about twice the length of 2372 and is defined by four hinge fractures extending from the proximal to the dorsal, giving it a trapezoidal cross-section. The proximal terminus retains a portion of the core platform, upon which downward pressure was applied to produce the blade; it appears as a thumbnail-shaped overhang. There is no bulbar scar in evidence, and no evident usewear. Probably Late Bronze Age in date, based on the frequency of this type of chert among the finds.

2342. Sickle element. Munsell: 10 YR 5/4. Pantone: S-40-3. H. 16 mm, w. 19 mm, th. 2 mm. Mohs hardness: 7.5. One of only four examples of a distinct type of sickle blade recovered by MARWP. Unlike the majority of the sickle elements from Pylos, which have serrated, retouched blades, this and the other three examples (2366, 2369, 2370) all have smoothly curving hemispheric blade edges, are of smaller size, and are retouched over their entire surface. This specimen was retouched by percussion and channel flaking to fit into a handle, and the opposite side reduced to a single facet blade extending the length of the piece’s 16-mm edge. The narrower, thicker butt end intended for insertion is heavily scarred by percussive trimming.

2401. Distal segment of prismatic blade. Munsell: 10 YR 5/4. Pantone: S-39-3. H. 36 mm, w. 27.5 mm, th. 11 mm. Mohs hardness: untested.

2364. Sickle element. Munsell: 5 YR 3/2. Pantone: S-318-1. H. 51 mm, w. 23 mm, th. 15 mm. Mohs hardness: untested.

The distal portion of a very wide prismatic blade in a light whitish-yellow chert is unique among MARWP lithics. The reverse is smooth, and the dorsal has a concave semioval hollow from which the edges taper outward to the parallel blades. Trapezoidal in cross-section. No evidence of usewear. Small patches of cortex remain.

203

Marquardt – Discarded Chipped Stone from the Palace of Nestor 2369. Sickle element. Munsell: 10 YR 6/2. Pantone: S-316-7. H. 29 mm, w. 22 mm, th. 8 mm. Mohs hardness: untested.

Like 2366 above, this piece has the thicker, narrow butt for insertion into a handle, and a wider, thinner, hemispheric cutting edge. Its surface is covered with percussion flake scars removed during the shaping and thinning of the artefact, and it lacks the serration or denticulation found on the more common Late Bronze Age variety.

An elongate, rectangular, relatively flat blade, conchoidally worked, of very hard brownish-maroon chert, the most common variety of chert found by MARWP. Along with 2353 in jasper, this is the rarest type of sickle blade among the MARWP finds. There are areas of cortex on both faces, artificially thinned by channel flaking. There are no welldefined proximal or distally indicative attributes present. Neither terminus is intact: one was snapped off at an oblique angle, the other notched and thinned. The scars on both edges are channel flake scars, running parallel to each other along each longitudinal edge. The basal notch may have been used for hafting as a knife, or even for use as a spokeshave/notch. The identification of this piece as a sickle element is thanks to P. N. Kardulias, who noted that this basal notching may have been intended to facilitate insertion into a wood or bone handle for use as a sickle element. Similar specimens occur at Kephala on Kea (Davis 1986), on Lesbos (Lamb 1936), at Malthi in northern Messenia (Valmin 1938), and among Blegen’s finds at Pylos, where they have been classified as knives generally, or as ‘backed’ knives. They occur more often in obsidian, although not in MARWP’s collection.

2370. Sickle element. Munsell: N-2 Greyish Black. Pantone: S-326-1. H. 25 mm, w. 15 mm, th. 6 mm. Mohs hardness: untested.

Like 2369 above, this piece lacks the serration usually found on sickle blades. It has a thicker, narrower butt for handle insertion, and a thin hemispheric blade edge, wider than the butt, reduced by percussion. A dull, lustreless brown-black chert. 2377. Sickle element. Munsell: 10 YR 5/2. Pantone: S-320-1. H. 30 mm, w. 16 mm, th. 5 mm. Mohs hardness: 7.5.

2366. Sickle element. Munsell: 5 Y 6/4. Pantone: S-11-6 and S-316-1. H. 32 mm, w. 22 mm, th. 6 mm. Mohs hardness: 7.5.

One of the most common lunate backed, denticulateserrated blade edged variety, of maroon-brown chert. This piece is particularly cortex laden, and was clearly manufactured by percussion flaking. The serration itself was probably created by notching, which may be considered a form of pressure flaking, after the overall shape was arrived at by means of percussion. There is no sickle gloss present, and the piece may never have been used.

Another example of the non-denticulated chert sickle element (with 2342, 2369, 2370). There is some sickle gloss present, and as with the other three such examples that we have recovered, there is a narrow (in this case, pointed) thicker butt end present for insertion into a handle of wood, antler or bone, with the curved hemispheric blade edge opposite. The chert is a yellow, almost golden colour, with inclusions of a brown hue, unique among MARWP lithics.

2380. Sickle element. Munsell: 5 Y7/1. Pantone: unmatched. H. 35 mm, w. 22 mm, th. 8.5 mm. Mohs hardness: 8.5.

204

Catalogue MARWP finds, with a striking platform remnant intact at the proximal terminus. A ragged, deeply indented serration extends along both longitudinal edges, bearing some usewear. The sides narrow towards the former distal terminus, ending in a cleanly snapped truncation. The pitted and worn edges have a patina consistent with use in harvesting cereal grains, and with other sickle elements of Bronze Age date. 2408. Sickle element. Munsell: 5 YR 2/4. Pantone: S-317-2. H. 40 mm, w. 27 mm, th. 5 mm. Mohs hardness: 7.5.

A sickle element of the denticulate or serrated edged variety. The piece is completely covered by silicate cereal patina gloss, and was so heavily used that the serration has been almost completely worn down. Many flake scars are still visible around the implement’s perimeter, however, indicating its manufacture by percussion flaking, combined with pressure notching of the serrate itself. The chert is unique among MARWP lithics, being a slightly translucent pine green colour. (Thanks to P. N. Kardulias for the identification of this piece and its heavy use.) 2389. Sickle element. Munsell: N-4 Medium Grey. Pantone: S-327-3. H. 42 mm, w. 31 mm, th. 7 mm. Mohs hardness: untested.

Another maroon-brown chert sickle element of the serrate variety, lunate backed and modified for insertion into a wooden or bone handle. Very worn and laden with patina gloss over most of its surface, which slightly alters the colour of the parent material. Sickle or saw blade 1868. Sickle element or saw blade(?). Munsell: unmatched. Pantone: S-127-5. H. 27 mm, w. 18 mm, th. 5 mm. Mohs hardness: untested.

A sizeable flake retouched into a sickle element. Produced by strike fracture. Of light brownish-green-grey chert, with a lustre near the blade flake scars probably consisting of cereal patina, from its use in cutting grain. Scars from the removal of small flakes are concentrated on one edge parallel to the flake’s long axis, knapped on both sides. One end of the flake is much thicker than the other and appears unworked. There are rough and uneven unworked zones on both the dorsal and ventral faces. The very greasy texture suggests chalcedony rather than chert, but it is in any case a siliceous quartz.

One of only two artefacts in a light pink, semi-translucent chert. Clearly a retouched flake, it bears three parallel pressure-made channel flake scars, and has a longitudinal semi-serrated edge. The flake is hinged, having been produced by percussion prior to its retouch. The small serrations were probably not produced by pressure notching, but rather by simple knapped percussive retouch. The channel flakes are perpendicular to the lengthwise hinge and terminate together at this line. There is no sickle gloss or usewear in evidence.

2404. Sickle element. Munsell: 5YR 3/2. Pantone: S-318-1. H. 19 mm, w. 15 mm, th. 3 mm. Mohs hardness: 7.5.

Bifaces 2350. Biface. Munsell: N-3 Medium Dark Grey. Pantone: S-327-2. H. 29 mm, w. 25 mm, th. 8 mm. Mohs hardness: 7.5.

A curved, crested, irregularly edged blade. Its smooth, rippled ventral has pronounced grooves. The material is a semi-translucent orange tinted pink chert, unique among

205

Marquardt – Discarded Chipped Stone from the Palace of Nestor notch at the wider base, as with the other pieces tentatively classed as projectile point pre-forms, demonstrating that it was at a less complete stage of reduction when it was abandoned. 2344. Barbed projectile point. Munsell: 5 Y 6/1. Pantone: unmatched. H. 23.5 mm, w. 11 mm, th. 4 mm. Mohs hardness: 5.

A roughly triangular piece of carefully percussion retouched chert, worked on both faces. The material is dull black, retaining a portion of the percussive bulbar scar from its original detachment, partially removed by flaking retouch.

Manufactured by a delicate process of micro-percussion fracture, similar to many found by Blegen. Typologically, the form is well known and well represented at numerous mainland sites as well as in the Cyclades. This piece has slightly inward curving barbs in a grey-green chert, with small flecks of mica present.

3449. Biface. Munsell: 10 R 2/4. Pantone: S-316-2. H. 41 mm, w. 30 mm, th. 14 mm. Mohs hardness: 7.5.

2352. Barbed projectile point. Munsell: 5 YR 3/4. Pantone: S-320-2. H. 34 mm, w. 16 mm, th. 11 mm. Mohs hardness: 7.5.

An example of red chert with mottled grey-black inclusions, having a tabular microcrystalline structure present in two prismatic blades (2372 and 2379) described above. The possibility that this chert was at some time imported from the Five Rivers region around Nichoria has already been discussed (see 2372). Labeled a biface here for lack of a better term, as it has been bifacially worked, the piece has been worked by percussion fracture exclusively, and bears numerous resulting scars. Its symmetry, as well as a strong formal resemblance to several other lithics discussed below, suggests that this is a pre-form intended for further manufacture into a barbed projectile point.

One of the most interesting lithics discovered by MARWP, consisting of maroon-brown chert worked into the form of a barbed projectile point with both barbs completed (one snapped off but preserved); the upper portion narrows to a symmetrical but incompletely thinned point, and the object was abandoned unfinished. The type is identical to that of 2344 above.

Projectile points 1866. Projectile point pre-form (?). Munsell: unmatched. Pantone: S-45-4. H. 22 mm, w. 20 mm, th. 4 mm. Mohs hardness: untested.

2358. Barbed projectile point. Munsell: 10 R 3/4. Pantone: S-85-3 and S-79-5. H. 12 mm, w. 12 mm, th. 6 mm. Mohs hardness: untested.

The lower half of a projectile point of the same form as 2344 and 2352 above, comprising the intact barbs. The piece is a little too thick to have been completely finished at the time of its original deposition.

A fragment of chert detached by means of percussion, as indicated by the bulbar scar at the wider end of the roughly triangular flake. Chipping by percussive fracture has left scars around the perimeter, indicating that it was intentionally fashioned into its present form. There is no

206

Catalogue Undefined blades

1856. Percussion flake. Munsell: unmatched. Pantone: S-325-6/ S-325-2. H. 24 mm, w. 13 mm, th. 4 mm. Mohs hardness: untested.

1871. Retouched and polished blade. Munsell: unmatched, Pantone: S-322-8. H. 22 mm, w. 14 mm, th. 5 mm. Mohs hardness: untested.

A simple percussion flake of a dishomogeneous brown chert, with some bits of attached cortex. A curved hinge fracture divides it longitudinally, and the percussive bulb is absent due to a truncation at the narrow flake’s wider terminus.

A uniquely coloured golden chert, more brightly yellow than the ‘honey chert’ of 1863. There is also a small zone of dark brown hue on one side, adjoining the mainly semitranslucent yellow body of the flake. A bulb of percussion and numerous concentric striations radiating from it dominate the lower halves, both dorsal and ventral. The material is carefully worked around in a half-lentoid form or ‘thumbnail’ scraper shape. The percussion scar is opposite this thinned and trimmed area, along an otherwise even base. No evidence of use, but the proximal end has a series of small percussion scars along its border.

1859. Percussion flake. Munsell: unmatched. Pantone: S-328-1. H. 29 mm, w. 25 mm, th. 4 mm. Mohs hardness: 7.5.

Flakes and chips 1850. Percussion flake. Munsell: unmatched. Pantone: S-329-2. H. 24 mm, w. 15 mm, th. 6 mm. Mohs hardness: untested.

In the common maroon-brown chert with a small area of cream coloured inclusion. The piece retains a sizeable bulbar percussion scar, but no evidence of retouch or usewear. 1863. Flake with usewear. Munsell: unmatched. Pantone: S-34-5. H. 45 mm, w. 37 mm, th. 5 mm. Mohs hardness: 8.5.

An unretouched percussion flake, with no evidence of usewear, in a brown-grey chert. Manufacturing waste. 1852. Percussion flake. Munsell: unmatched. Pantone: S-325-4. H. 36 mm, w. 37 mm, th. 11 mm. Mohs hardness: 7.5. An unretouched piece of manufacturing waste in a dark brown chert. Bears a large bulbar percussion scar and small flecks of cortex. 1853. Percussion flake. Munsell: unmatched. Pantone: S-66-2. H. 23 mm, w. 23 mm, th. 8.5 mm. Mohs hardness: untested.

One of the most important lithic finds. The flake has directional usewear, indicating that it was used as a scraping or hide-cleaning tool; there is no appearance of intentional retouch, however. Its interest stems from its material, an extremely rare variety of chert not native to Greece. It is a variety termed elsewhere ‘honey chert’: golden and translucent, and extremely hard.

A piece of manufacturing waste, bearing a bulbar percussion scar covering one of its faces, and three concave scars on the reverse; small flecks of cortex present. The chert is a light greyish-black.

2349. Percussion flake or chisel bit (?) Munsell: 5 Y 7/4. Pantone: S-31-5. H. 31 mm, w. 23 mm, th. 13.5 mm. Mohs hardness: untested.

207

Marquardt – Discarded Chipped Stone from the Palace of Nestor

A piece of yellow-green chert, having parallel edges and coming to a thin distal edge; it could have been used for scraping or cutting. The opposite, thicker end bears a thin layer of cortex. There is directional usewear present on the thin blade edge consistent with use as a scraping implement.

An interesting flake of the yellow-orange dishomogeneous chert, but without any cream-coloured inclusions. The dorsal of this roughly triangular flake bears percussionmade knapping scars along its longest edge, which also retains a great deal of heavy usewear consisting of overlapping directional incised grooves. Probably used as a scraping and/or cutting tool. The ventral bears a large bulb of percussion from its original detachment, prior to retouch and use.

2355. Percussion flake. Munsell: 10 YR 8/4. Pantone: S-25-5. H. 23 mm, w. 13 mm, th. 6 mm. Mohs hardness: untested.

2371. Conchoidally detached flake. Munsell: 10 R 5/4. Pantone: S-318-6. H. 57 mm, Wd.27 mm, th. 11 mm. Mohs hardness: 7.5. Another example of the dishomogeneous yellow-orange chert with cream-coloured inclusions, bearing a bulb of percussion. It has a hinge fracture bisecting the dorsal face, and the ventral is smooth and convex. It is truncated at the wider end, the opposite coming to a point. May be regarded as waste from core reduction. 2359. Percussion flake. Munsell: 5 YR 4/1. Pantone: S-327-3. H. 34 mm, w. 21 mm, th. 8 mm. Mohs hardness: untested.

An excellent example of a conchoidally detached flake, in a pink, dishomogeneous chert with silvery-grey inclusions. The proximal retains a large percussive bulbar scar as well as a portion of cortex from the core, but lacks any tabular platform remnant. Divided along the longitudinal dorsal axis by a long hinge fracture, which itself divides towards both the distal and proximal termini. The flake does not appear to have been modified or retouched in any way, nor are there traces of usewear. This is noteworthy inasmuch as many of the flakes we have recovered and documented are far less suitable for blade-implement manufacture than this fine, elongate piece. One difficulty with using this as a blade is its thinness, which would not have been able to sustain a great degree of pressure.

A percussion flake in a light yellow chert, with small flecks of cortex, a bulbar percussion scar on the ventral, and no evidence of retouch or usewear. May be regarded as waste from manufacturing. 2368. Flake with retouch and usewear. Munsell: 10 YR 8/4. Pantone: S-25-5. H. 33 mm, w. 24 mm, th. 7 mm. Mohs hardness: 7.5.

2374. Percussion flake. Munsell: 5 YR 3/4. Pantone: unmatched. H. 35 mm, w. 27 mm, th. 3 mm. Mohs hardness: 7.5.

208

Catalogue

Another roughly triangular flake of the commonly represented maroon-brown chert, with a small area of cream-coloured inclusion. No evidence of retouch or usewear, bearing a small percussive bulbar scar. May be regarded as manufacturing waste.

A third example of a prismatic core in the dishomogeneous yellow chert with cream-coloured inclusions. It also has the typical conical form common in prismatic cores, and bears a couple of elongate pressure flake removal scars. Much of the platform upon which pressure was applied in order to detach the blades is still intact and, as it lacks cortex, seems to have been prepared for further reduction. This core is far from exhausted, and may have been intended to produce more blades. It seems reasonable that had the palace not been destroyed by warfare and or fire, the core’s user would have continued utilising this valuable source of blades, from which many microblades could yet have been detached. The frequency of this type of chert is only surpassed by the maroon-brown material.

2402. Retouched flake. Munsell: 10 R 2/4. Pantone: S-319-1. H. 34 mm, w. 32 mm, th. 10 mm. Mohs hardness: untested.

2361. Core. Munsell: 5 YR 3/4. Pantone: S-320-2. H. 55 mm, w. 82 mm, th. 23 mm (max.). Mohs hardness: untested.

Perhaps intended as a scraping tool, this piece is convex and smooth on the reverse, while its dorsal face has been knapped by percussion fracture into a three-quarter circle, leaving numerous indented diagnostic scars. No evidence of usewear. The chert is light reddish-brown. 3453. Retouched corticular flake. Munsell: 5 Y 6/4. Pantone: S-13-4. H. 37 mm, w. 30 mm, th.8 mm. Mohs hardness: 7.5. A piece of olive-coloured chert retouched into a rectangular shape, probably as a blank or pre-form for later completion. A pronounced percussion scar covers half of one of its faces, and the reverse is about one-third cortex laden. Detached by percussion from a larger core, there are numerous small flake scars around the flake’s perimeter. Cores 2343. Prismatic core. Munsell: 10 YR 7/6. Pantone: S-43-2. H. 43 mm, w. 35 mm, th. 27 mm. Mohs hardness: 8.5.

The finest example of a chert core found by MARWP. Its entire surface is covered with a multitude of small pressure flaking and percussion scars, many of which are rippled and striated. There are two large cream-coloured dishomogeneities present in the material, and no cortex whatsoever. 2376. Core. Munsell: 5 R 5/6. Pantone: S-79-3. H. 33 mm, w. 29 mm, th. 14 mm. Mohs hardness: untested.

209

Marquardt – Discarded Chipped Stone from the Palace of Nestor 2398. Prismatic core. Munsell: 10 YR 6/6. Pantone: S-35-6. H. 45 mm, w. 30 mm, th. 17 mm. Mohs hardness: 7.5.

A small, highly reduced chunk of smooth, hard, red tabular chert. It bears numerous percussion scars, as well as a several pressure flaking scars consisting of long, narrow concave facets. The strike fracture percussion scars greatly outnumber these larger pressure scars and are concentrated around the edges, probably the result of core preparation prior to initiation of pressure technique in blade reduction. For other examples of this chert, see 2372, 2379 and 3449 above.

A prime example of a prismatic core, bearing several pressure flake scars resulting from pressure actuated blade removal. The raw material is the yellow dishomogeneous chert with cream-coloured inclusions. Its biconical shape is typical of prismatic cores, though it is generally represented in obsidian and uncommon, but not unknown, in siliceous quartzes.

2390. Core. Munsell: 10 YR 6/6. Pantone: S-38-4. H. 34 mm, w. 32 mm, th. 23 mm. Mohs hardness: 7.5.

3445. Prismatic core. Munsell: 5 R 5/4. Pantone: S-320-2. H. 40 mm, w. 28 mm, th. 30 mm. Mohs hardness: 7.5.

Another core in the dishomogeneous yellow chert with cream-coloured inclusions. The surface is covered with numerous flake scars, produced by percussion in order to obtain flakes for retouch. There is a large facet retaining a considerable portion of cortex. No pressure platform is present, and there are no evident blade scars.

A piece of the tabular plated red chert like that described in the prismatic blade 2372 above. Long, wide concavities indicate longitudinally removed pressure flakes, probably in blade manufacture. There is no cortex or inclusion of any kind, and no well-defined striking platform remains, indicating that the core was fully exploited and discarded in prehistoric times. No evidence of strike fracture percussion.

2394. Prismatic core. Munsell: 5 Y 7/4. Pantone: S-31-5. H. 29 mm, w. 27 mm, th. 7 mm. Mohs hardness: 7.5.

Cobble 2406. Corticular cobble flake. Munsell: 10 YR 5/4. Pantone: S-35-4. H. 35.5 mm, w. 47 mm, th. 14 mm. Mohs hardness: 7.5.

Probably detached in an effort to revive its parent core for further blade reduction, this piece of chert is smooth on its ventral, or plane of fracture, and bears three longitudinal pressure flake scars on the dorsal. The chert itself is a light mottled green, containing numerous pockmarked inclusions and phenocrysts. There appears to be a remnant of the original core’s platform still present on the fragment, upon which downward pressure would have been applied in order to detach prismatic blades.

210

Catalogue Chalcedony (?)

An orange chert with heavy cortex on roughly one half of the hinge fracture caused by its initial removal by direct percussion from a larger cobble of the material. There is no certain evidence of further reduction, and it cannot therefore be termed a core. (My thanks to P. N. Kardulias for the identification.)

2365. Pressure flake. Munsell: 10 R 3/4. Pantone: unmatched (purplishbrown). H. 24 mm, w. 14 mm, th. 4.5 mm. Mohs hardness: untested.

Andesite 1858. Retouched flake. Munsell: unmatched. Pantone: S-285-5. H. 42 mm, w. 27 mm, th. 6 mm. Mohs hardness: untested. Possibly chalcedony, on the basis of its greasy texture, this flake shows no evidence of retouch or usewear and was detached by pressure, as no percussive bulb is present. It probably remained unused because of its very small size. Unknown stone 1873. Retouched flake. Munsell: unmatched. Pantone: S-85-3. H. 31 mm, w. 23 mm, th. 6 mm. Mohs hardness: untested. A retouched flake of green andesite, trimmed into a rectangular form with slightly flared longitudinal edges. The functional purpose of this artefact is indeterminate, but it could have been used as a hide-scraper or for polishing leather. It is a very hard material, despite its somewhat grainy texture. The nearest known sources of andesite are located in Arcadia and on the island of Aegina in the Saronic Gulf. Fairly carefully trimmed around the edges, with a truncated base, and percussion flaking scars over most of its surface, its purpose and material remain obscure. The rock itself is a light reddish, large grained, non-siliceous material, crazed by yellow veins.

Jasper 2353. Sickle element. Munsell: 10 R 3/4. Pantone: S-319-2. H. 35 mm, w. 25 mm, th. 12.5 mm. Mohs hardness: untested.

2363. Flake. Munsell: N-3 Medium Dark Grey. Pantone: unmatched.H. 12 mm, w. 14 mm, th. 2 mm. Mohs hardness: untested.

A small flake retouched over both faces from a small oval platform extending outward from a somewhat truncated, formerly even base. The material has no lustre, is very lightweight and does not present a very sharp cutting edge. Certainly not a chert or chalcedony, it nevertheless does have a greasy texture. Possibly andesite?

The second example of the rarest type of sickle element present in the backfi ll (see also the sickle blade 2364 above). It has a thicker butt end, which was intentionally truncated, whereas the blade end is wider and much thinner, coming to a straight blade edge. There is a significant amount of vertical and horizontal usewear on and around the blade edge area, which is slightly wider and much thinner than the butt end. The material is a dark red jasper, shot through with black veins.

211

Marquardt – Discarded Chipped Stone from the Palace of Nestor Bibliography Blegen, C. W. and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I. The Buildings and their Contents. Princeton University Press, Princeton, NJ. Blitzer, H. 1992. ‘The Chipped Stone, Ground and Worked Bone Industries’, in Excavations at Nichoria in Southwest Greece, II: The Bronze Age Occupation, ed. W. A. McDonald and N. C. Wilkie, 712-56. University of Minnesota Press, Minneapolis. Davis, J. L. 1986. Keos, V. Results of The Excavations Conducted by The University of Cinncinnati Under the Auspices of The American School For Classical Studies at Athens. Philipp von Zabern, Mainz on Rhine. Lamb, W. 1936. Excavations at Thermi in Lesbos. Cambridge University Press, Cambridge. McDonald, W. A. and Rapp, G. 1972. The Minnesota Messenia Expedition: Reconstructing a Bronze Age Regional Environment. University of Minnesota Press, Minneapolis. Valmin, M. N. 1938. The Swedish Messenia Expedition: Excavations at Malthi-Dorion. Gleerup, Lund.

Submitted 1999

212

7 SMALL FINDS FROM THE 1990–98 EXCAVATIONS eleni M. konstantinidi-Syvridi

The miscellaneous finds discovered in the investigations at the Palace of Nestor between 1990 and 1998 lack all context, but they nevertheless reflect the multiplicity of activities that took place at the site over a long period, as evidenced by the materials and the forms used. The majority of the finds comprise bronzes, including nails, pins and various nondescript fragments. The presence of iron in the tip of a chisel, an arrowhead, a ring and a building nail reinforces the opinion previously expressed by Mervin Popham and others (Popham 1991, 315–24; Griebel and Nelson 1993), and strengthened by other studies published in this volume, that there was Geometric reoccupation of the site. A late Classical or Hellenistic lamp (see p.270, fig. 9.6) is further evidence for occupation of the site in the post-Bronze Age period.

in various parts of the palace (PN I, 70 and 297 (knives), 76 (blade), 94 (sword and spearhead)). One of the few arrowheads sufficiently well preserved for categorisation is 4933. It belongs to the barbed type with curved sides that dates from LH I to LH IIIB (Buchholz 1962, 25–26 (Type V); see also Avila 1983, 85–86 (Klasse Ia) and 110 (Klasse 2d)). This type of bronze arrowhead, stamped from a sheet of bronze, is the most numerous, widely distributed and long-lived in Greece. Bronze arrowheads are abundant in the palace.1 More than 500 diminutive bronze arrowheads and many ivory bits were found in Room 100, which Blegen concluded had served as a ‘special section in the Workshop’ for the making of delicate objects in bronze and ivory (PN I, 325). A smaller quantity was found in the Main Drain under Room 60, in Court 92 and in Room 99, assumed by Blegen to be the palace workshop (ibid., 321). The requisitioning of scrap bronze for the fabrication of weapons is suggested by Linear B tablet Jn 829 from Pylos, which refers to contributions of bronze from all over the kingdom for the manufacture, among other things, of a large number of javelin- and spear-points (Chadwick 1987, 33–34, fig. 17).

Among the most significant finds are the heads of two female figurines of known Mycenaean type, as well as fragments of three animal figurines that should date to a later period than the Bronze Age. The rest of the finds consist of clay items related to spinning and weaving (a spool, whorls and loomweights), a lead fishing-net weight and some organic material comprising various shells, knuckle bones and a boar’s tusk; there are as well a few clay and stone items, some of unknown use, which seem to have Minoan and Cypriot affinities. The finds are presented under the broad heading of their material. The catalogue of objects is selective, including only those items sufficiently preserved for discussion.

One nail, 4967 (fig. 7.1a), and the heads of two more, 4917 (fig. 7.1b) and 4965, were probably used in architecture. Nail 4967 is of rectangular section and has a slightly biconical head. Nail head 4965 has two deep incisions at the point of the neck. Blegen’s excavations encountered nails in various parts of the palace, most of them with a shank of rectangular section and round or rounded heads; lengths vary between 0.032 and 0.054 m (PN I, 90, 109, 246, 285 and 336). Parallels are to be found at several sites, such as the Treasury of Atreus at Mycenae (Wace 1921–23, 348, fig. 76), Midea (Persson 1942, 11) and Gla (Iakovides 1989, 250, T 10a, 28b), dated from LH IIB to LH IIIB 2.

Metal Objects Bronze Most of the bronzes consist of small pieces of scrap metal, either solid or sheet. Some of them, such as 4932, which still preserves traces of rivets, would originally have been part of weapons of the types discovered by Blegen

1

For a typology of the palace specimens and a catalogue of sites that have produced arrowheads, see Avila 1983, 83–114.

213

Konstantinidis-Syvridi – Small Finds from the 1990–98 Excavations for writing on clay tablets and sealings. Examples made of bronze or bone are reported from Thebes and Midea (Godart 1988, esp. 248–51; Demakopoulou 1974, 167). a

b Iron Iron, although represented by only a few items, provides useful information for the site’s period of occupation. Among the iron fragments discovered is the blade of a chisel, 4926 (fig. 7.2a), and the arrowhead 4941 (fig. 7.2b). The latter, though much corroded, is significant, as evidence for Dark Age arrowheads is scarce: there are only a few examples of iron arrowheads, from Athens, Corinth and Lefkandi (Popham and Sackett 1980, 256). The biggest concentration was at Lefkandi, where a ‘quiverful’ was found in Tomba Tomb 26, together with a hydria representing archers and possible remains of a composite bow (ibid., 182–83). The Englianos specimen is a flat arrow plate similar to those from Lefkandi.

c d 0

1

2

3

4

5cm

7.1 Bronze small finds: a) nail 4967; b) nail head 4917; c) pin 4908; d) awl? 4915.

It has been suggested that large nails were used to attach the metal coating of wooden antae or used in doorjambs. At Gla, the excavator found a couple of such nails by the gate, still adhering to the bronze plate that they once served to keep in place (ibid.). At Pylos, Blegen found evidence in Room 104 that the door casing had burned, evidenced by carbonised wood and black earth at the southwestern side of the doorway (PN I, 343). The smaller nails found at the palace may derive from wooden boxes with hinged covers (ibid., 98, fig. 274:1), which were used for storing tablets in the manner noted also at Knossos by Evans (1935, 668).

The ring 4971 (fig. 7.2c) belongs to the undecorated type of a simple hoop with a perfect circular section and no trace of a joint, which indicates that it was cast in a mould. Such rings are known from various periods – Geometric to Roman, Medieval and Byzantine (e.g. Immerwahr 1971, 106, pl. 40:29–30 (Athenian Agora) and Davidson 1952, 230, nos 1903–13 (Corinth)) – and their sizes vary; larger ones may have been used in some other way than as finger rings. Another iron ring, with an attached hook, was found by Blegen in Court 42. It is thicker and has a slightly smaller diameter than our example, and it was found together with fragments of iron spikes (PN I, 184, fig. 287:5). Although he found some late Geometric pottery in association with this ring, Blegen could not certainly date the iron to the Geometric period or a subsequent one. An iron nail, 4906 (fig. 7.2d), is folded now, with a rectangular head and one end flattened. The leaf-shaped fragment 4934, of fine plate, could be modern.

Pin 4908 (fig. 7.1c), from Blegen’s dump, has a rectangular shank tapering at one end and a discoid head. KilianDirlmeier does not date the type, although she does mention one parallel in a Proto-Geometric tomb in Argos; there, however, the disc is of bone (Kilian-Dirlmeier 1984, fig. 10:273). The length of our specimen – 0.085 m – is rather short in comparison to that of the Sub-Mycenaean/ Proto-Geometric pins, and no decisive conclusion on the date of the pin is possible. Blegen reported a few more possible fragments of pins, and there may be more that have not been properly identified. The head of a pin came from Court 63 (PN I, 246), found together with a few more items of personal use, namely a bead, a steatite button and ivory bits. The spherical head of a pin was found in Room 97 (ibid., 311), and the shank of a pin, now bent, came from Room 98 (ibid., 316, fig. 314:1).

Lead Lead has been found in various places at the palace, mostly in the form of shapeless lumps that may have been used for the mending of pots or to solder in place bronze

Fragments 4915 (fig. 7.1d) and 4958 are not well enough preserved to be safely identified as pins, and their rectangular section suggests that they could belong to small awls, like the ones discovered at three points in the palace (ibid., 102, fig. 278:1; 285, fig. 302:7); these were found in association with a chisel, nails and fragments of bronze indicating workshop activities (Tripathi 1988, no. 9; and PN I, 297, fig. 299:3). Bronze awls make their first appearance in the Early Helladic settlement at Lerna (Banks 1967, 33) and continue to be among the commonest tools throughout the Bronze Age. The small diameter of 4915 and 4958 in comparison to the other awls of the palace suggest that they may have functioned as styluses

a

c

d b 0

1

2

3

4

5cm

7.2 Iron small finds: a) chisel 4926; b) arrowhead 4941; c) ring 4971; d) nail 4906.

214

Stone Objects

0

1

2

3

4

The two cylindrical beads, 4928 and 4928b, are of serpentine (dark green variety) and steatite respectively. Stone beads, though only occasionally found in the palace, were manufactured from a variety of materials: banded agate (PN I, 59 and 99), carnelian (ibid., 159 and 347), amethyst (ibid., 202), a dark blue stone, possibly hematite (ibid., 246), rock-crystal (ibid., 322), marble (ibid., 329) and steatite (ibid., 347). The shaft grave under Room 97 also produced some beads of amethyst, carnelian and amber (ibid., 314), probably grave offerings. For some of those stones the inhabitants of the Aegean depended only partly on imports; rock-crystal, for example, was found in large quantities in Crete but was also imported from Cyprus, Egypt and India. Other stones, however, were always imported: amethyst and steatite from Syria, and hematite from western Asia via Egypt (Ogden 1982, 89ff ). The distribution of semi-precious stones provides evidence for the prosperity and commercial outreach of a region; judging from the semi-precious stones found in the palace and the surrounding area of Englianos (the latter mostly in burial contexts in the form of beads and pendants; PN III, passim), Messenia appears to have been one of the richest areas of mainland Greece.

5cm

7.3 Lead small finds: possible fishing weight 4927.

accessories for doors and windows; in one case it was in the form of a cylindrical pellet (ibid., 259, fig. 300:15). The cylindrical lead object 4927 (fig. 7.3) could be a type of fishing-net weight, like the examples discovered in several Aegean sites – Perati, Brauron, Gla, Naxos, Astypalaia, Phaistos, the Chelidonian Cave and Enkomi (Iakovides 1969–70, 249, pl. 36a – usually in a LH IIIB–IIIC context; most of these, however, are smaller and are bent to form a rectangle or are almost square to allow the passing through of a rope. Very similar weights are still used today in the islands, especially on Crete, for coastal fishing.

The stone weight 4922 (fig. 7.4c) is rectangular in section with rounded edges, a conical piercing half-way up from the base and a circular piercing from front to back. There exists a whole class of similar but not identical items, generally called ‘weights’, the real purpose of which remains obscure. Blegen’s excavations produced seven rectangular objects and fragments of others made of unbaked clay, irregularly shaped but almost square in section, with at least two perforations, one at each side (PN I, 348); our object resembles especially one of them (no. 4). A similar item is reported from the settlement of Trianda on Rhodes, dated to LH IIIA–IIIB (PapazoglouManioudaki 1990, 178); it is described as a hard river stone of cylindrical shape with a flat base, enabling it to stand upright. The other side is ellipsoid, perforated with two oblique holes that communicate with each other, and reach the middle of the height. This allows for the insertion of a rope or suspension of the tool by the fingers, or even for

Stone Objects Fragment 4901 (fig. 7.4a) is part of a button or conulus of a common Mycenaean type. These objects, which occur at almost every variety of Mycenaean site, are usually reported as being made of steatite. Nevertheless, serpentine or chlorite (‘green steatite’) are also common materials for the making of buttons, as both stones, especially serpentine, are found in abundance on the Greek mainland, in a large range of colours (Onasoglou 1979, 18, n. 9). Both the hardness of the material and the well-polished surface suggest that our specimen is serpentine. Three types of stone buttons are known from the palace: the shanked, to which our example belongs (PN I, e.g. 91, 98, 234, 286, 295, 297, 301, 311, 322, 336), the conical (ibid., 155, 234, 237, 246, 259, 284, 297, 301, 303, 306, 322, 336) and the disc-shaped (ibid., 194, 322). Stone buttons are generally thought to succeed the clay examples, which first appear in LH I and II contexts. Clay specimens were also found at the palace; they come mainly from the earlier Southwestern Building (e.g. in Hall 65: ibid., 259) and the exterior wall of that building (ibid., 284 and 286). Stone buttons came into use sometime in LH III, represented first by the conical shape (LH IIIA1), then by the shanked variety (LH IIIA2); which evolved to the flat T-shaped or disc-shaped form (LH IIIB). All types continue in use up to the very end of LH IIIC (Iakovides 1977, 388). They have been variously interpreted as spindle whorls, dress weights or dress accessories, buttons or beads; depending on the circumstances, all four uses are possible.

a

b

0

1

2

3

4

5cm

c 7.4 Stone small finds: a) button or conulus 4901; b) mould fragment 4980; c) weight 4922.

215

Konstantinidis-Syvridi – Small Finds from the 1990–98 Excavations its positioning onto a stable point, in a similar way to our specimen.

It may belong to a naturalistic or Proto-Phi figurine, as this type usually has a straight-sided base.

The mould fragment 4980 (fig. 7.4b) is similar to a fragment found by Blegen by the northwestern wall bordering Rooms 77 and 78 (PN I, 284, fig. 300:4). The excavator interpreted it as a micaceous schist fragment of mould for a spearhead with a midrib. Our example has a narrower width and is more likely an arrowhead mould, but it is very similar in form and material to Blegen’s find. Object 4951 is probably also part of a rounded mould or part of a whetstone like the ones discovered in the same area as Blegen’s find. Both items are related to workshop activities, as is a crucible reported by Blegen to have been found in association with his moulds (ibid., 336).

Blegen’s excavations revealed at least 13 fragments of Mycenaean female figurines (PN I, 149, 234, 246, 284, 291, 311, 322, 336 and 340). Most of the identifiable fragments belong to the Psi type (Blegen’s ‘crescent type’), with the exception of a fragment from the Aqueduct (ibid., 340, fig. 322:4), which should belong to a Phi figurine (Blegen’s ‘disc’ shape). Heads are both polos-type and bare, with the majority belonging to the former class. Of the three animal figurine fragments (fig. 7.6), 4911 and 4914 belong to the usual bovine type. The thin triangular fragment 4912 could belong to a horse figurine of postMycenaean date, like the one from Corridor 95 (see below); a thin horse of larger dimensions, pierced twice, was found at Rhytion in Crete and is dated by its context to the Geometric period (Alexiou 1972, 622). Fragment 4911 consists of the hindquarters of a bovine with no indication of decoration. Blegen reported a fragment found in the proximity of Court 63, consisting of the forepart of a bovine (ibid., 247, fig. 297:3). It has not yet been traced, but one wonders if these two fragments might be part of the same figurine. Blegen’s excavations produced a large number of fragments belonging to animal figurines (ibid., 141, 234, 247, 284 and 336). In Corridor 95 in the Northeast Building, a figurine probably of a horse pierced through the neck (ibid., 308, fig. 310) was re-dated by Popham to Geometric times (French 1971, 159).

Fired Clay Objects Figurines Two female figurine heads of known Mycenaean types – 4959 (fig. 7.5a) and 4916 (fig. 7.5b) – were found in the MARWP investigations, and there are three more possible figurine fragments. Head 4959, with its large nose, relatively long neck and vertical band decoration, probably belongs to a Phi-type figurine, dated from LH IIIA 2 to LH IIIB (French 1971, 132–33). There is a noticeable absence of the typical ‘necklace’, as decoration seems to consist only of vertical bands starting just below the jaw. The polos head 4916 belongs certainly to a Psi-type figurine. The birdlike face and the decoration of crossed lines just below the nose indicate that it should be dated to LH IIIB. Of the other figurine fragments, 4930 may have formed the greater part of the torso of a female figurine, with both ends missing. Unfortunately, no trace of painting survives. 4924 is a miniature cylinder that could be part of a small figurine, and 4920 is probably the lower part of the torso of another female figurine, with typical decoration of vertical bands.

Fragments 4912 and 4914 are pierced. This piercing of figurines, both animal and anthropomorphic, mostly dates to the Geometric period and later. An animal figurine (NM 1340) from the Tholos of Clytaemnystra at Mycenae has holes in the side to aid in firing. From Rhytion on Crete came a bull figurine from a group of three with a hole under the neck for the firing, and a horse figurine with two holes; both objects are dated to the Geometric period (Alexiou 1972, 622). At the sanctuary of Athena Pronaia at Delphi, many of the female Psi figurines are pierced in the upper part of the body, to facilitate hanging by a thread (Demangel 1926, 24–25, figs 28–29). They might have been

a a

b

b

c 0

1

2

3

4

5cm

0

7.5 Heads of a female figurines: a) 4959; b) 4916.

1

2

3

4

5cm

7.6 Animal figurine fragments: a) 4911; b) 4912; c) 4914.

216

Fired Clay Objects suspended from the branches of trees in a sacred grove or worn as amulets, much as the Archaic-period Boeotian figurine in the Louvre (NMB 1313) wears what appears to be a Phi figurine. None of the female figurines from the palace at Englianos appears to have been pierced, at least as far as their state of preservation allows us to judge; piercing is restricted to a few of the animal figurines.

other sites (e.g. Prosymna (Blegen 1937, 313), Mycenae (Wace 1932, 217 pl. XXXV, Tomb 517), Korakou (Blegen 1921, 109), Monemvasia (Demakopoulou 1968, 145ff ), Melathria (Demakopoulou 1977, 33, 38, 42, 48, pl 11a) and Kythera (Coldstream and Huxley 1972, 209)) and dated by context from LH I–II to LH IIIA 2. Spindle whorls can be distinguished from common buttons both by their dimensions (buttons do not exceed 0.03 m in diameter and 0.021 m in height) and their shape, as they are flatter with broad undersides, made to hold the spool of thread as it grew thicker around the stem of the spindle.

Whorls and spools Among the clay finds related to textile production are whorls and spools. Whorls 4909 (fig. 7.7a) and 4921 (fig. 7.7b) belong to the asymmetrical biconical type with truncated top (Nichoria type 7; see Smith 1992, 680 and 682), dated to LH II–IIIA2. Plenty of whorls were found in Blegen’s excavations, most of them belonging to the Mycenaean period and either of biconical or of truncated biconical shape. 4909 is of a type also found at several

a

Three sherds, 4935, 4936 and 4910 (fig 7.7c–e), may have had a secondary use as whorls. A discoid ‘weight’ reported by Blegen (PN I, 107 fig. 278:9), made from the raised base of a small pot, may have been used likewise. In practical terms, almost anything could have been used as a loomweight. Such use is apparent in several Mycenaean sites, such as Nichoria, where sherds from kylikes and stirrup jars could have had the same use (Wilkie 1992, 292). Whorls made out of sherds are rare but appear in all periods, probably because the flat shape enabled the act of spinning.

b

One spool, 4907 (fig. 7.7f), was found in the MARWP investigations. Spools of various materials (stone, clay and bone) are known from as early as the Early Bronze Age at Zygouries (Αρχαιολογικs 1899, 100 and 196), Malthi (ibid., 356) and Aghios Kosmas (ibid., 142–43 fig. 166), and continued in use throughout the Mycenaean period at Prosymna (Blegen 1937, 286, fig. 285:19), Dorion (Valmin 1938, 338, fig. 46), Nichoria (Wilkie 1992, no. 358), Perati (Iakovides 1969–70, 353) and Kephallenia (Metaxata) (Marinatos 1933, 80 and 93, fig. 38); they must have been used in the same way as they are today, for the wrapping of thread.

c A possible themiaterion and other clay objects The function of the tubular object 4905 (fig. 7.8) is enigmatic. On the upper half it bears a collar or rimchannel, now broken. Decoration consisting of a brown circle surrounded by smaller concentric circles in red covers the base. The collar, which was probably visible when in use, was also painted. The drop-shaped protrusion is left undecorated but bears incisions. The collar recalls the Minoan kerniskoi or the Cycladic ‘fruit-stands’, but the shape itself finds parallels in Cyprus, in the so-called miniature themiateria (incense-burners) published by Karagiorgis (1996, pls XLV:5–7); one of them is safely dated to the Cypro-Archaic II/Cypro-Classical I period, 475–400 bc. The complete Cypriot examples consist of two parts: a solid cylindrical stem splaying out towards the base, the top terminating in a shallow bowl for the incense, and a lid with an opening at the top. They are decorated with black and purple-red vertical and horizontal bands on the stem and around the exterior of the bowl; the height varies from 0.074 to 0.125 m.

d

f

e

0

1

2

3

4

5cm

7.7 Fired clay whorls and spools: a) 4909; b) 4921; c) 4935; d) 4936; e) 4910; f) 4907.

217

Konstantinidis-Syvridi – Small Finds from the 1990–98 Excavations an interesting insight into life at the palace. The existence of paint on, or the remodelling of the shape of some of the shells, their presence in graves, and the quantities in which they were found all suggest that they were used in a number of ways, some of which are reflected in our examples.

0

1

2

3

4

The two almost identical Pinna nobilis shells (Greek pinna; 4977; fig. 7.9b) bore holes, probably natural. They may have been jewellery items (earrings or pendants), like examples excavated at several burial sites, especially on Crete (Sackett and Popham 1965, 303–04, fig. 20:19).

5cm

7.8 Possible themiaterion, 4905.

Actual themiateria, mainly in bronze, have come to light, especially in Spain as Phoenician imports; they are dated mainly to the 7th century bc. In Cyprus, themiateria appear in terracotta, bronze and limestone and may be associated with Phoenician influences in religious rituals, as seen at sites such as the archaic sanctuary at Meniko. The clay examples continue to be in use down to the Classical period, and although they vary in details, the typology is more or less the same. Nevertheless, a comparison of our example and the Cypriot ones leaves some unanswered questions: what is the function of the drop-shaped protrusion with no opening (does it represent a lid?), and why does the object have a pierced base (to be placed on a stand?). The general concept is nonetheless the same, though its small size may indicate not so much the practical use of the object as the symbolic.

The small shell 4978 (fig. 7.9a) belongs to Cerithium vulgatum (horn shell, Greek keratios); it is found along rocky shorelines to a depth of 10 m, and its length ranges from 0.05 to 0.06 m. It is edible, and has been found at many sites from all periods (e.g. Mycenae, Lerna, Asine, Nichoria; see Reese 1992, 772). Our example is waterworn, meaning that it was collected dead; it bears a hole to the apex, probably natural, and may have been used as a pendant or other ornament. Conoid shells, usually of smaller size, are well known from other Bronze Age sites, in particular Prosymna, where dozens were found; most were pierced at the side and/or at the spiral end (Blegen 1937, 465). Sometimes they were even fi lled with lead, presumably to be used as toys. The Spondylus gaederopus 4979c, a kind of thorny oyster, is common along the shores of the southern Peloponnese, as shown by the finds at Malthi, Asine and Midea (Reese 1998, 279). The Chama gryphoides, 4979a, is quite rarely found on excavations. Shell 4943 (fig. 7.9c) belongs to a topshell Monodonta turbinata (marine gastropod). This shell reaches a maximum of 0.035 m in diameter and 0.06 m in height and was commonly eaten in the ancient Mediterranean basin, being easily collected on rocky coasts, chiefly from the splash zone. Rings or pendants from topshells are known from Neolithic levels at Franchthi (Jacobsen 1973, 258 pl. 48d, second from left, second row) and in later periods from Aghios Stephanos, Mycenae, Koukounaries, the Unexplored Mansion at Knossos, Chania, Kommos and Enkomi (for Enkomi see Dikaios 1969, 717 pl. 168:48 and 108; for the other sites see Reese 1984, 237). In a closed late Mycenaean context they are artefacts (usually finger rings) rather than organic debris.

The clay disc 4981 is also noteworthy. Similar clay discs, ranging in diameter from 0.026 to 0.09 m and sometimes pierced, have appeared in large quantities at several sites on Crete and the mainland, such as Phaistos (Levi 1961– 62, 410), Kavousi (Boyd 1901, 141) and the Athenian Agora (Burr 1933, 602–03; Lalonde 1968, 131), dating from the Middle Minoan/Middle Helladic II period to Geometric times. They have been interpreted as stoppers, or counters for games or other purposes (Mammina et al. 1990, 5–48, for Mycenaean and earlier periods; see also Cucuzza 1998 for more recent bibliography), while the possibility that at least some of them were used as weights should not be excluded. Item 6010, a clay piece preserving a thumbprint, resembles a seal; these were normally formed around a string that leaves a hole running through the clay, however, which is not the case here. At Akroteri on Thera, such items may have been used as counters, as they sometimes bear incised symbols (Tzachilli 1992, 144).

Two Charonia examples of different varieties were also found, as well as a small piece of Dentalium dentalis (4963). Dentalium shells appear on Crete (e.g. Kommos and the Unexplored Mansion at Knossos; see respectively Reese 1992, 772, and Evely 1993, pl. 222:9, bottom right) and on the mainland (as at Nichoria; Reese 1992, 771), mainly in the form of necklace beads. Not many shells are reported by Blegen: one burnt Dentalium (inedible) came

Organic Remains Organic material from the MARWP excavations is represented by shells, a land snail, a boar’s tusk and several knuckle bones (astragaloi)2. The presence of shells at Pylos, combined with evidence for fishing in the form of the lead fishing weight described above (4927), provides

2

My warmest thanks to Dr David Reese for his help with the identification of the organic material.

218

General Remarks General Remarks

a

The small finds recovered by MARWP at the palace at Englianos are only a very minor percentage of the total, but these are nevertheless indicative of the variety that once existed. Their importance lies in their reflection of the everyday activities of the people of Pylos, especially in regard to workshop products. Small-scale bronzecasting is suggested by the numerous arrowheads and scrap metal fragments; textile production by the spindle whorls, spools and weights. To a lesser extent, the small finds provide evidence for the diet of the inhabitants (fishing-net weight and shells) and personal ornamentation (pins, beads and buttons).

b

c d

Not all the finds date necessarily to the Mycenaean period, as evidence for the use of at least some of them extends over a long time span. As all the objects come from Blegen’s backfi ll, contextual dating is impossible, and dating by means of comparanda, from the original excavations at Pylos and from other sites, is therefore crucial to an understanding of the objects; this has led, in some cases, to a reconsideration of the published material. As noted above, the presence of iron in the palace has troubled some scholars (Popham 1991). The character of certain parts of the building, for example rooms 89–90 and 82–83 (ibid., 320–21), does not sit comfortably with the date offered by Blegen, as there are now indications of a Dark Age reoccupation of the site (Griebel and Nelson 1993) in rooms 82 and 83, which overlie the pre-palatial constructions of the Northwest area (see chapter 8 below). Evidence from the small finds confirm the possibility of a post-palatial date for these walls in the form of the hoard of straight-sided arrowheads found by Blegen, which should not be earlier than Late Helladic IIIC, and some of the animal figurine fragments – especially the ones bearing perforations, which were not common in Mycenaean times. Later material includes the Classical/ Hellenistic lamp, which along with Archaic, Roman and medieval pottery identified in chapter 9, and architectural remains discussed in chapter 8, provide further evidence for activity on the hilltop long after the destruction of the Bronze Age palace.

e

0

1

2

3

4

5cm

7.9 Organic remains: a) Cerithium vulgatum 4978; b) Pinna nobilis 4977; c) Monodonta turbinata 4943; d) boar’s tusk 4949; e) astragaloi (knuckle bones) 4913, 4923, 4919.

from Corridor 26, and examples of Arca noae L. (Greek kalognomi) from Room 27 (PN I, 149, fig. 283:12). The boar’s tusk 4949 (fig. 7.9d), a lower canine, is not modified and is probably food debris. Blegen’s excavations produced several tusks, mostly fragmentary (ibid., 91, 149, 314, 329). It is possible that at least some of them were used for making warrior’s helmets of the type that appears in Middle Helladic III and continues as part of Mycenaean armour until LH IIIB (Borchhardt 1972, 57ff; Varvarigos 1981).

The archaeological evidence has shown that Messenia was among the first regions to have used iron. Two sites close to the palace, Tragana and Kokevi, produced material dated from Sub-Mycenaean and Protogeometric times (12th–10th centuries bc) to the later Dark Age, mostly personal ornaments (pin, ring) and tools (knife). The relative wealth of finds from this period is an indication of the commercial contacts of Messenia, for iron, along with other materials, must have reached Messenia either overland from the Argolid through Arcadia, or by sea from some other region, perhaps Attica or even Cyprus. It is also evidence that occupation continued or resumed in Messenia at a time when there is no indication of similar activity on sites such as Mycenae, Thebes, Tiryns or Midea, sites that played an important role in the political

There are also three knuckle bones, or astragaloi (fig. 7.9e): 4913 and 4923 come from cattle, 4919 from a pig. They seem to be natural food debris, as there is no visible trace of work. Several similar objects were found by Blegen (PN I, 196, 234, 240 and 266), and astragaloi are very commonly reported from archaeological sites across the Mediterranean basin and Near East. Worked or unmodified astragaloi are frequently found in domestic contexts, sanctuaries (as in Kition on Cyprus) and graves, from the Early Bronze Age to the present day. They may have been used as weights or as gaming pieces, as they are still used today in some parts of Greece in the game called kotsia (Reese 1985, 384; Broneer 1947, 241).

219

Konstantinidis-Syvridi – Small Finds from the 1990–98 Excavations and economic life of the Aegean until the end of the Mycenaean period.

4934. Leaf-shaped plate. L. 0.070 m, w. 0.007 m. One sharp point and a hook at the other.

Catalogue

4941. Arrowhead tip. L. 0.032 m, w. of top 0.018 m.

Metal Objects 4971. Ring. Th. 0.003 m, diam. 0.025 m. Plain circle; could be modern.

Bronze 4908. Pin. L. 0.085 m, diam. of head 0.005 m. Intact, with flat discoid head. It has a shank of rectangular section tapering toward the end; two-thirds of the way down its length it narrows and becomes cylindrical.

Lead 4927. Tubular sheet, possibly a fishing-net weight. L. 0.041 m, outer diam. 0.014 m, th. 0.030–0.040 m.

4915. Pin (?). L. 0.044 m. max. th. 0.003 m. Fragment of the shank. Of rectangular section, widens around the middle.

Stone Objects 2384. Loomweight. H. 0.29 m, w. 0.29 m, th. 0.06 m. The preserved portion of a limestone loomweight, with central hole; one quarter of its original radius survives.

4917. Nail head. L. of shank 0.012 m, th. 0.006 m, diam. of head 0.027 m. Preserves the top of the cylindrical shank, now bent. Flat discoid head, much worn.

2385. Arrowshaft straightener (?). H. 0.41 m, w. 0.20 m, th. 0.28 m. Tubular limestone object, intentionally drilled and polished.

4932. Bronze plate. 0.017 x 0.025 m. Elongated piece of plate with traces of two rivets.

4901. Button (conulus). H. 0.01 m, diam. of base 0.018 m. Almost half of it preserved, possibly of serpentine.

4933. Barbed arrowhead. 0.024 x 0.0105 m, th. 0.004 m.

4922. Possible weight. L. 0.073 m, w. 0.045 m, diam. of hole 0.015 m. Elongated object of a dark reddish stone; it has a cylindrical section with rounded edges and a conical piercing from the base to the middle; another piercing runs from front to back.

4958. Pin or nail. L. 0.02 m, th. 0.003 m. Fragment of pin or nail shank, of rectangular section. 4965. Nail head. Diam. 0.012 m. Flattened, with two circular grooves at the top.

4928. Bead. H. 0.012 m, w. 0.01 m, th. 0.004 m. Flat top, cylindrical, of a greenish-brown stone.

4967. Nail. L. (max., with head) 0.04 m, w. 0.017 m, diam. of head (circle not complete) 0.015 m, th. of outline 0.002 m, th. of section of shank 0.005 m. Intact. The head is thick, of solid metal and with a somewhat biconical profi le on the lower side; shank is of rectangular section with pointed end.

4928b. Unknown object. L. 0.04 m, w. 0.029 m, th. 0.023 m, diam. of hole 0.005 m. Steatite item of cylindrical shape, perforated. 4951. Mould fragment. 0.035 x 0.025 m.

Iron 4906. Nail. L. c. 0.032 m (with head), w. of head 0.010 m. Folded shank. One end of the head is rectangular, the other flattened.

4980. Mould fragment. L (extant): 0.035 m, w. 0.018 m. Possibly of schist.

4926. Tool blade (chisel). L. 0.063 m, w. 0.038 m. Triangular shape; surface much corroded.

220

Catalogue Clay Objects

Whorls and spools

Figurines

4907. Spool. 0.042 x 0.02 m. One end (now broken at the edge) wider than the other. 4909. Biconical spindle whorl. H. 0.031m. w. 0.027 m. Surface bears slip.

4911. Animal figurine. H. 0.05 m, l. of body 0.03 m, w. 0.016 m. Fragment of animal figurine of medium clay; only hind quarters are preserved and part of the torso. Left leg a little longer than right (0.06 m). No trace of painting.

4910. Sherd. Max. diam. 0.03 m. Fragment of mouth of a closed vase; no paint preserved. Probably second use as a loomweight, as indicated by the edges, which have been smoothed and rounded.

4912. Animal figurine. H. 0.034 m, l. of back 0.05 m, w. 0.05 m. Fragment of animal figurine of fine clay. Almost triangular shape, with two holes (one is cut off, the other intact) pierced for suspension.

4921. Biconical spindle whorl (?). L. 0.019 m, w. 0.032 m. The top of the right side is chipped off. Red clay.

4914. Animal figurine. L. (beginning of neck) 0.032 m, (end of neck) 0.028 m. Fragment of the torso of an animal figurine, of fine orange clay. It has a large cutting at one side, where it curves slightly upwards, and bears a horizontal hole for suspension around the middle of its length. No trace of decoration.

4935. Sherd. Diam. 0.072 m. Sherd of the mouth of a closed vase; traces of paint. Probably second use as a loomweight, as indicated by the edges, which have been smoothed and rounded.

4916. Female figurine. H. 0.018 m, w.0.019 m, th. 0.007 m, diam. of top 0.012 m, diam. of base 0.012 m. Head of female figurine of fine clay; polos is missing. Broken off below neck (which is relatively long), protruding nose, eyes indicated by applied blobs that bear traces of red paint. Oblique stripes surround the eyes and are finally joined at the bottom of the nose.

4936. Sherd. Diam. 0.075 m. Sherd of the mouth of a closed vase; no trace of paint. Probably second use as a loomweight, as indicated by the edges, which have been smoothed and rounded. Other clay objects

4920. Base of a stand(?). H. 0.018 m, l. of base 0.033 m. Flat base, rectangular section with rounded edges, decoration with four painted vertical bands both on the front and on the base.

3160 Lamp. Late Classical or Roman lamp. See p.270, fig. 9.6. 4905. Possible thermiaterion. H. (with protrusion) 0.047 m, diam. of base 0.028 m, h. of protrusion 0.012 m. Object of tubular shape with flat concave base. The top has a high drop-shaped protrusion surrounded by a conical cup. The item is pierced through the base, up to the middle of the body.

4924. Miniature cylinder. L. 0.011 m, w. 0.012 m. Bears painted decoration of five red lines around the periphery. Purpose unknown; could form part of a figurine or some terracotta composition.

4981. Disc. Diam. 0.027 m. Traces of painted decoration; fine clay.

4930. Female figurine (?). L. (max) 0.052 m, w. at top 0.021 m, at lower end 0.028 m. Fragment of pyramidal shape, hollow. Could be the foot of a vase or part of a female figurine, the lower half of the torso without the base. Traces of dark brown painted decoration of horizontal bands and attached trace of plaster.

6010. Unknown object. L. (max) 0.022 m, w. 0.012 m, th. 0.01 m. Fragment (burnt) with thumbprint. Broken into three pieces and mended. It must have been held between thumb and another finger and then somehow burnt.

4959. Female figurine. H. 0.031 m, w. 0.016 m, th. 0.019 m. Bare head of female figurine, broken off below neck. Decoration of painted stripes has faded from the right eye except from a very small edge, but the outline of the left eye is still visible.

Organic remains 4913. Knuckle bone (cow). L. 0.034 m, w.. 0.055 m, th. 0.032 m.

221

Konstantinidis-Syvridi – Small Finds from the 1990–98 Excavations 4919. Knuckle bone (pig). L. 0.02 m, w. 0.033 m, th. 0.018 m.

Borchhardt, J. 1972. Homerische Helme. Helmformen der Agais in ihren Beziehungen zu orientalischen und europaischen Helmen in der Bronze- und fruhen Eisenzeit. Philip von Zabern, Mainz.

4923. Knuckle bone (cow). L. 0.023 m, w. 0.02 m.

Boyd, H. A. 1901. ‘Excavations at Kavousi, Crete, in 1900’, AJA 5, 25–57.

4943. Topshell Monodonta turbinata (marine gastropod). L. 0.02 m.

Broneer, O. 1947. ‘Investigations at Corinth, 1946–1947’, Hesperia 16, 233–47.

4949. Boar’s tusk. L. 0.105 m, w. 0.017 m. Probably food debris.

Buchholz, H. G. 1962. ‘Der Pfeilglatter aus dem VI Schlachtgrab von Mykene und die helladischen Pfeilspitzen’, Jahrbuch des Deutschen Archäologischen Instituts 77, 1–58.

4963. Dentalium dentalis shell. L. 0.025 m, w. 0.008 m, th. 0.005 m.

Burr, D. 1933. ‘A Geometric House and a Proto-Attic Votive Deposit’, Hesperia 2, 542–640.

4977. Two Pinna nobilis shells. L. (max) 0.038 and 0.034 m, w. (max) 0.016 and 0.015 m.

Chadwick, J. 1987. Linear B and Related Scripts. University of California Press, Berkeley.

4978. Horn shell (Cerithium vulgatum). L. 0.042 m.

Coldstream, J. N., and Huxley, G. L., eds. 1972. Kythera: Excavations and Studies Conducted by the University of Pennsylvania Museum and the British School at Athens. Faber and Faber, London.

4979. Three large shells: a) broken conk shell, b) whole mollusc of murex type, and c) Spondylus gaederopus (thorny oyster). L. a) 0.11 m, b) 0.082 m, c) 0.09 m, and a small fragment 0.024 m.

Cucuzza, N. 1998. ‘Geometric Phaistos: A Survey’, in PostMinoan Crete: Proceedings of the First Colloquium, ed. W. G. Cavanaugh et al. BSA Studies 2, 62–68. BSA, Athens.

Bibliography

Davidson, G. R. 1952. Corinth XII: The Minor Objects. ASCSA, Princeton, NJ.

Alexiou, S. 1972. ‘Rhytion’, Αρχαιολογικόν Δελτίον 27, B2 Chronika, 622.

Demakopoulou, K. 1968. ‘Μυκηναϊκή Κεραμεική από θαλαμωτούϚ τάφους στην περιοχή Αγ. Ιωάννου, Μονεμβασία’ Αρχαιολογικόν Δελτίον 23A, 145–94.

Avila, R. A. J. 1983. Bronzene Lanzen- und Pfeilspitzen der griechischen Spatbronzezeit (Prähistorische Bronzefunde V:1). Beck, Munich.

Demakopoulou, K. 1974. ‘Thebes’. Αρχαιολογικά Ανάλεκτα εξ’ Αθηνών 7, 167.

Banks, E. C. 1967. The Early and Middle Helladic Small Objects from Lerna. Unpublished PhD dissertation, University of Cincinnati, OH.

Demakopoulou, K. 1977. ‘Μυκηναϊκή Ταφή στη Μελαθριά, Λακωνία', Αρχαιολογική Εφημερίς, 29–60.

Blegen, C. W. 1921. Korakou: A Prehistoric Settlement near Corinth. Princeton University Press, Princeton, NJ.

Demangel, R. 1926. Fouilles de Delphes, II:5. Le Sanctuaire d’Athena Pronoia: Topographie du sanctuaire. E. de Boccard, Paris.

Blegen, C. W. 1937. Prosymna: The Helladic Settlement Preceding the Argive Heraeum. Cambridge University Press, Cambridge.

Evans, A. 1935. The Palace of Minos at Knossos, IV. Macmillan, London.

Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and their Contents. Princeton University Press, Princeton, NJ.

Evely, R. D. G. 1993. Minoan Crafts, Tools and Techniques: An Introduction. Studies in Mediterranean Archaeology 92. Paul Åströms Förlag, Göteborg.

Blegen, C. W., Rawson, M., Taylour, W. D. E., and Donovan, W., 1973. The Palace of Nestor at Pylos in Western Messenia, III: Acropolis and Lower Town; Tholoi, Grave Circle, and Chamber Tombs; Discoveries Outside the Citadel. Princeton University Press, Princeton, NJ.

French, E. 1971. ‘The Development of Mycenaean Terracotta Figurines’, Annual of the BSA 66, 101–84. Godart, L. 1988. ‘Autour des textes en Linéaire B de Tirynthe. Ausgrabungen in Tiryns 1982/83’, Archäologischer Anzeiger, 245–51.

222

Bibliography Reese, D. S. 1992. ‘Appendix I. Recent and Fossil Invertebrates’, in Excavations at Nichoria in Southwest Greece, II: The Bronze Age Occupation, ed. W. A. McDonald and N. C. Wilkie. University of Minnesota Press, Minneapolis, 770–78.

Griebel, C. and Nelson, M. 1993. ‘Post-Mycenaean Occupation at the Palace of Nestor’, AJA 97, 331. Immerwahr, S. A. 1971. The Neolithic and Bronze Ages. The Athenian Agora, XIII. ASCSA, Princeton, NJ.

Reese, D. S. 1998. 'The Faunal Remains', in Pseira III: The Plateia Building, ed. C. Floyd. University of Pennsylvania Museum, Philadelphia, 131–44.

Iakovides, S. 1969–70. Περατή. Το Νεκροταφέιον. Β. Γενικαί Παρατηρήσεις. Athens. Iakovides, S. 1977. ‘On the Use of Mycenaean Buttons’, Annual of the BSA 72, 113–19.

Sackett, L. H., Popham, M. R., Warren, P. M., and Engstrand, L. 1965. ‘Excavations at Palaikastro, VI’, Annual of the BSA 60, 248–315.

Iakovides, S. 1989. Γλας Ι, Η Ανασκαφή 1955–1961. Athens. Schliemann, H. 1880. Mycenae: A Narrative of Researches and Discoveries at Mycenae and Tiryns. Charles Scribner's Sons, New York.

Karagiorgis, V. 1996. The Coroplastic Art of Ancient Cyprus, VI. The Cypro-Archaic Period: Monsters, Animals and Miscellanea. University of Cyprus, Nicosia.

Smith, J. C. 1992. 'Spinning and Weaving Equipment', in Excavations at Nichoria in Southwest Greece, II: The Bronze Age Occupation, ed. W. A. Mcdonald and N. C Wilkie. University of Minnesota Press, Minneapolis, 673–711.

Kilian-Dirlmeier, I. 1984. Nadeln der fruhhelladischen bis archaischen Zeit von Peloponnese. C.H. Beck, Munich. Lalonde, G. V. 1968. ‘A Fift h-century Hieron Southwest of the Athenian Agora’, Hesperia 37, 123–33.

Tzachilli, I. 1992. ‘Μικροαντικείμενα της ανασκαφής’, in Ακρωτήρι Θερας είκοσι χρονια έρευνας  (1967–1987), ed. C. Doumas. Athens 139–45.

Levi, D. 1961–62. ‘Gli Scavi à Festos negli anni 1958–60’, ASAtene 39, 377–504.

Tripathi, D. N. 1988. Bronzework of Mainland Greece from c. 2600 bc to c. 1450 bc. Studies in Mediterranean Archaeology 69. Paul Åströms Forlag, Göteborg.

Mammina, G., Marazzi, M., and Tusa, S. 1990. ‘Espedienti di computo: Il caso Vivara’, Dialoghi di Archeologia 8, 2: 5–48.

Valmin, N. M. 1938. The Swedish Messenia Expedition: Excavations at Malthi-Dorion. Gleerup, Lund.

Marinatos, S. 1933. Αρχαιολογική Εφημερίς, 79–95. Ogden, J. 1982. Jewellery of the Ancient World. British Museum Press, London.

Varvarigos, A. P. 1981. Το οδοντόφρακτον κράνος (ως προς την τεχνική κατασκευής του). Διατριβή, Πανεπιστήμιο Αθηνών Athens.

Onasoglou, A. 1979. ‘Ένας νέος μυκηναϊκός θαλαμωτός τάφος στο Κουκάκι’Αρχαιολογικόν Δελτίον34, A’: 15–42.

Wace, A. J. B., et.al. 1921–23. ‘Excavations at Mycenae’, Annual of the BSA 25, 103–402.

Papazoglou-Manoudaki, L. 1990. ‘Μινωικός οικισμός στα Τφιάντα της Ροδου’Αρχαιολογικόν Δελτίον 37, A’: 139–87.

Wace, A. J. B. 1932. Chamber Tombs at Mycenae. Archaeologia 82. Society of Antiquaries, Oxford.

Persson, A. 1942. New Tombs at Dendra near Midea. Gleerup, Lund.

Wilkie, N. C., with Dickinson, O. T. P. K. 1992. 'The MME Tholos Tomb', in Excavations at Nichoria in Southwest Greece, II: The Bronze Age Occupation, ed. W. A. Mcdonald and N. C Wilkie. University of Minnesota Press, Minneapolis, 231–344.

Popham, M. 1991. ‘Pylos: Reflections on the Date of its Destruction and on its Iron Age Reoccupation’, Oxford Journal of Archaeology 10/3, 315–24. Popham, M. R., and Sackett, L. H., eds. 1980. Lefkandi I: The Iron Age: The Settlement, The Cemeteries. Thames & Hudson, London.

Submitted 2002

Reese, D. S. 1984. ‘Topshell Rings in the Aegean Bronze Age’, Annual of the BSA 79, 237–38. Reese, D. S. 1985. ‘Appendix VIII (C). The Kition Astragali’, in Excavations at Kition, V/II, ed. V. Karageorghis. Department of Antiquities, Nicosia, 382–91.

223

8 POST-BRONZE AGE ARCHITECTURE AND STRATIGRAPHY Todd M. Brenningmeyer

Excavations conducted by Carl Blegen and the University of Cincinnati between 1952 and 1966 at Englianos provided the first glimpse into the history of the hilltop – an historical record that preserved evidence of activity extending from the Bronze Age through the modern period. References to later material appear in the original excavation publication,1 but these phases were beyond the chronological scope of the original study and received only passing mention.2 MARWP’s study of the architecture and finds at Englianos revealed evidence of post-Bronze Age construction on the hilltop, suggesting a widespread though sporadic continuation of architectural activity after the end of the Bronze Age. Rooms 87, 89 and 90 preserve evidence of Dark Age construction separated stratigraphically from the Bronze Age remains. The walls of rooms 83/84/85 in the Northwest Area form a building of megaron type also situated above Bronze Age strata; fragments of Archaic-period architectural terracottas derive from this building, interpreted as a small Archaic temple. It was succeeded by a later temple, building 83/82/81. Although a precise stratigraphic provenience cannot be reconstructed for much of the re-excavated post-Bronze Age material, a basic sequence has been established, using archived data from the original excavations and evidence uncovered during MARWP investigations.

volume of The Palace of Nestor (hereafter PN I; Blegen and Rawson 1966) as a thin yellow deposit that covered the Bronze Age architecture to an average depth of 0.05 to 0.25 m. This layer rests below the level of all post-Bronze Age architecture encountered at the site. A second, very distinctive layer of soil lay above the yellow stratum at several locations, described in The Palace of Nestor and the excavation notebooks as a dark, black, oily soil with numerous inclusions of small stones. It was encountered under the plough zone, 0.15–0.25 m below the surface, and extended to a depth of 0.40–0.48 m. This black stratum represents a stratigraphic horizon that consistently marks post-Bronze Age activity. Blegen attributed this black layer to early Archaic (c.600 bc) industrial activity (PN I, 296–97), suggesting that the dark black colour and oily texture of the soil was the result of seepage from an olive oil press located near rooms 89 and 90 (ibid., 177). There is no physical evidence for the proposed oil press, however, and the extent of the black soil (spreading unevenly across almost 20 per cent of the site) and the type of pottery recovered from it are not consistent with Blegen’s interpretation. The date of the black oily stratum is uncertain. The description of this layer is very similar to one identified at Nichoria in the Dark Age and medieval levels, described as ‘very dark, with a characteristic “stickiness”’ (Stein and Rapp 1978, 250). Analysis of soil from the Nichoria levels revealed a high clay content that is common to argillic and cambic horizons that formed through natural geologic processes after the level was covered by later strata (Yassoglou and Haidouti 1978, 34). Such stratigraphic horizons have been identified in geologically young surfaces excavated throughout the Mediterranean. The dark colour and sticky nature of the soil are characteristic of a high clay content in localised geological layers that result from ‘natural soil-forming processes’ rather than from human occupation (ibid., 34–35). The post-Bronze

Post-Bronze Age Stratigraphy After the LH IIIB destruction, a thin layer of soil covered the site. This stratum is described throughout the first 1

Areas said to contain Geometric pottery include rooms 39, 40, 41, 42, 47, 58, 63, 91, 92, 94, 101 and 102, as well as ‘surface’ deposits near the megaron. Rooms 42, 58 and 63 are also associated with iron objects. 2 For re-assessments of Blegen’s interpretation of post-Bronze Age phases at the palace more recent to the time of writing this study, see Popham 1991 and Griebel and Nelson 1998; see also Mountjoy 1997 for a reassessment of Popham’s interpretation.

225

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy Age date of the black stratum is confirmed by the layer’s position above the yellow stratum that covered Bronze Age levels throughout much of the site. The soil’s black, oily character may derive from the illuviation of clay into strata located below the plow zone, or from chemical changes that altered subsurface layers. In either case, the soil colour and character reflect natural geological changes that occurred after the deposition of the soil, and are consistent with similar layers found at Nichoria; it is possible that the black stratum at Englianos represents a similar geological formation. While the presence of Geometric pottery is a recurring theme in Blegen’s descriptions of this stratum, the chronological span may extend beyond the Dark Age.

trench, the feature appears to represent an intentionally built construction of post-palatial date. (For a detailed description of the chasm and its stratigraphy, see pp.76–80 above.) Another layer of black soil with numerous small stones was encountered in the Throne Room (Room 6). According to Blegen’s published account, the black soil was identified only in the central section of the room, lying above a hard yellowish deposit that covered the central hearth and extended slightly beyond the four column bases (ibid., 89). The black stratum was bounded on all sides by a layer of burned and dissolved brick that extended from the plow zone to the Mycenaean floor. Blegen’s description of the stratigraphy suggests that the black stratum formed after the deposition of the burned and dissolved brick that bounds this layer on all sides. Post-Bronze Age material in this room was not specifically associated with the black stratum, but fragments of Geometric pottery were recovered from ‘surface deposits’ presumably located above the throne room (fig. 8.2 no. 45).

main megaron area The original excavators encountered the post-Bronze Age levels described above at several locations throughout the site (fig. 8.1). In Court 3, a dark, black, oily stratum was encountered within a ‘depression’ (PN I, 70) or ‘hollow’ (ibid., 64) that stretched across much of the court and into the adjoining Room 4, where the deposit measured up to 0.50 m thick. This black layer contained ‘some scattered fragments of Late Geometric ware’ (fig. 8.2 no. 32; ibid.; see also Popham 1991, 317; Coulson 1986, 67, dates this pottery to Dark Age III, roughly the second half of the 8th century bc) and was bounded at the northeast and southwest by a deposit of ash and rubble, which in turn was bounded by the extant palace walls. Beneath the black stratum and at each end of the depression was a layer of dry, ashy, sandy soil with patches of dissolved red brick (Blegen and Rawson 1966, 70). This deposit of dry soil rested on a layer of ‘firm light brown’ earth that lay above the Bronze Age floor. The light brown layer is probably the same light brown or yellow layer that covers Bronze Age remains throughout much of the site.

southeast and northeast of the main building In Area 58, outside the southeast wall of the Main Building, the oily black stratum surrounded several large ashlars and fragmented blocks that fell from the outer wall of the palace. Late Geometric sherds and two fragments of iron spikes (fig. 8.2 nos 29–31) were recovered from this layer, which rested above a thin level of ‘whitish limey’ soil that covered the stucco pavement (PN I, 229–30). Post-Bronze Age material was also recovered from an ‘intrusive deposit of Late Geometric times’ composed of black soil with small stones that extended from the southeast corner of Room 39 into Room 40, the northwest half of Room 41, and courts 42 and 47. This black stratum lay above and between fallen ashlars from the palace walls; in no case was the black layer found below worked blocks (ibid., 175). A layer of yellowish soil was identified in the northwest part of Court 42, but much of the Bronze Age surface within the courts was covered with building debris below the level of the black soil. Blegen recovered four complete pots of the Late Geometric period from the black stratum here, comprising two one-handled cups, a two-handled jar and a pedestal base (fig. 8.3 nos 14–17; ibid., 184, pl. 347:615–17; Coulson 1986, 67–68, dates these vessels to Dark Age III). Four iron spikes, an iron ring with a hook attached, and a flat piece of iron were also found in Court 42 (fig. 8.3 nos 8–13); the black stratum in Court 47 also contained Late Geometric pottery (fig. 8.3 no. 7; PN I, 209). Additional material from these areas, possibly post-dating the Bronze Age, is described in Blegen’s unpublished draft of The Palace of Nestor. In this manuscript, Blegen mentions a concentration of sherds found in Court 42, 0.50 m below the surface near the northeast wall, just opposite the southeast anta of Room 41 (fig. 8.3 no. 18; Blegen, n.d., 35 (‘Area M’)). These fragments are described as ‘large pieces of good fabric but strangely flat and of even thickness (tiles?)’. Roof tiles from Archaic and later periods were

At the southwest corner of Court 3 the excavators encountered the edge of an irregular cut that Blegen named the ‘chasm’ because of its great depth (Blegen 1953, 63). The southeast walls of rooms 7 and 8 (the Archive) were removed during the ancient excavation of this cavity, prompting Blegen’s interpretation of the feature as a robber’s trench (PN I, 93 and 96). The dark black oily stratum found within the ‘hollow’ of Court 3 did not continue into this location. The chasm instead was fi lled with loose earth and small stones as well as a mixture of material from Middle Helladic to Venetian periods (ibid. 99–100). Blegen believed the fill was deposited when the looters refi lled the robbing trench after the extraction of the walls (ibid., 96), but the character of the chasm’s construction does not indicate a trench rapidly cut for the extraction of architectural material. The feature extends downward (c.1 m) from Room 9 in a series of crude steps to a floor composed of rounded boulders, flat fieldstones, and fragments of reused ashlar blocks. The depth of this cavity is much greater than would be necessary to remove the walls of the Archive. Rather than a robber’s

226

65

73

77

70

71

78

8.1 Evidence of post-Bronze Age activity across the site.

227

64

66

67

68

72

61

82

60

63

89

90

88

59

83

12

9

10

14

15

17

19

20

21

84 85

7

8

11

16

22

86

0

87

1

2

23

10

²

Medieval deposits

Areas of post Bronze Age activity

Black Stratum

Archaic Period Remains

74

80

81

3

4

5

6

20

58

57

56

55

24

30

53

48

41

50

46

43

40

39 38

34

33

32

36

29

44

25

32

26

27

91

40 Meters

47

42

102

104

92

93

96

97

101

103d

103c

103b

94

95

103a

105

100

99

98

Post-Bronze Age Stratigraphy

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy 39

38

41

# # # ### # 40

#

37

43 44

42

45

#

36 35

# $

# 34

# 33

# 32

# 26 # # 27 28

! 31

Finds Identified in Palace and Western Side of Site

² 0

2

4

8 Meters

26. LHIIIC or Geometric neck handled amphora (Popham 1991, 317) 27. LHIIIC or Geometric bowl (Popham, 1991, 317) 28. Geometric pottery 29. Geometric pottery 30. Iron Spike 31. Iron Spike 32. Geometric oenochoe 33. Geometric pottery 34. Geometric pottery

8.2 Post-Bronze Age finds identified in the Main Building and on the western side of the site.

228

# !

29

30

35. Geometric pottery 36. Corinthian Tile 37. Geometric pottery 38. Geometric krater-bowl 39. Geometric krater-bowl 40. Geometric krater-bowl 41. Geometric krater-bowl 42. Geometric krater-bowl 43. Geometric krater-bowl 44. Geometric krater-bowl 45. Geometric pottery sherds

Post-Bronze Age Stratigraphy recovered during the MARWP excavations of Blegen’s backfi ll (see pp.245–53 below), and Blegen’s fragments must belong to one of these roofs.

1953, 23; Blegen n.d., 5 (‘SE End of Court H (Bathroom)’); PN I, 295). Room 90 was added later, doubling the size of Room 89 through the extension of the northwest and southwest walls of that room to the flank wall of the palace. Sometime later, the door that originally opened into Room 90 from Room 89 was blocked and the dividing wall demolished; a floor of rough flagstones was laid in Room 90, and a new doorway was opened at the southeast corner of the room, as indicated by a threshold block (Greibel and Nelson 1998, 98). The excavators noted a thin yellowish stratum approximately 0.10 m in depth resting above this flagstone floor and above the floor of Room 89. The excavation notebooks describe the recovery of stone and fragments of Corinthian tile from this yellowish stratum (fig. 8.2 no. 36); the number and dimensions of the tiles are not recorded (Hope and MacDonald 1953, 27, 29, 31). An oily black stratum rested above the yellow layer, extending over the walls and across much of Court 88 to a depth of approximately 0.20 m. This stratum contained what were described as Geometric sherds (fig. 8.2 no. 35), as well as pieces of larger worked blocks believed to have fallen from the outer wall of the palace adjacent to Room 17 (PN I, 6 and 298).

The layer of dark black soil and small stones encountered in courts 42 and 47 extended into Ramp 91, Court 92 and Colonnade 94. The stratigraphy of Ramp 91 follows that described above, with a layer of black soil lying immediately below the plow zone, containing many glazed sherds of Late Geometric style (PN I, 300–01). More fragments of Geometric pottery were recovered from Court 92 and Colonnade 94 (fig. 8.2 nos 4, 6); the soil that lies beneath the black stratum in Court 92 is described simply as a destruction level. A small terracotta horse (ibid., 307–08, fig. 310), uncovered within Corridor 95 and dated by Popham (1991, 324) to the Geometric period, indicates a continuation of post-Bronze Age levels to the northeast of Colonnade 94. It is unclear whether the thin black stratum, described as resting above the clay earth floor of Corridor 95, represents a continuation of the black soil identified in Colonnade 94. (For further discussion and a proposed interpretation of post-Bronze Age activity in this area, see pp. 43–44 above.) Blegen provides little stratigraphic information for rooms 101–103 but does note the presence of Geometric sherds in the upper layers of rooms 101 and 102 (fig. 8.3 nos 19, 20; ibid., 328–32). Further evidence of post-Bronze Age activity in Room 103 is provided in the field notes of Marion Rawson (1958, 98 and 112), who describes an iron nail and an iron blade (not recorded in Blegen’s catalogue of finds) that were recovered from the fill around the western wall of rooms 103c and d (fig. 8.3 nos 21, 22). Several underfired pots, brick-red in colour, including belly-handled jars that Popham notes may belong to a later occupation, were uncovered within rooms 103b, c and d (fig. 8.3 nos 23–25; Popham 1991, 320; PN I, 338, 340; see also pp.41–43 above).

The stratigraphic position of the black layer that seals rooms 89 and 90 is peculiar, given the presence of Geometric sherds in its matrix and ‘Corinthian’ tile fragments in the underlying level. Terracotta roof tiles do not appear between the Bronze Age and the 7th century bc (N. Cooper 1983; Wilkander 1990, 288), and the tile fragments found within rooms 89 and 90 should therefore post-date the Geometric sherds that Blegen found within the black stratum. The stratigraphic relationship of the tiles to the overlying black stratum suggests reverse stratigraphy, resulting from later activity in the area. It is possible that the black stratum and the associated Geometric sherds found above the tiles were re-deposited from adjoining Court 88 (where other fragments of Geometric pottery were recovered: fig. 8.2 nos 37–44) during a late restructuring of this area, which used architectural debris and some of the soil from Court 88 to level off the surface above rooms 89 and 90. On this assumption, a terminus post quem for the building sometime during the Late Geometric period is proposed. The character of its construction is similar to that of small rectangular buildings uncovered in Dark Age settlements throughout much of Greece (Ainian 1997; Drerup 1969), and the interior hearth and small size of the two rooms suggest a domestic structure of the late Dark Age.

southwest of the main building The yellowish stratum found above Bronze Age remains throughout the site extended across Court 88 and below rooms 89 and 90. The construction of these two rooms connected the standing external walls of Room 64 and the partially standing southwestern facade of the palace. Room 89, about 5 metres square, was erected first. Reused ashlar blocks from the palace were set directly on the yellowish stratum, forming the northwest, southeast and northeast walls, with the northeastern wall of the Southwestern Building closing the room on the fourth side. The yellowish layer was removed from the interior of the building, which reused the Bronze Age pavement as the floor. The entrance to the room was probably in the north wall (Griebel and Nelson 1998, 97). A small rectangular feature constructed of upright slabs and probably representing a hearth3 lay upon the plaster floor, abutting the northwest wall, which must therefore be earlier than this feature (Hope and MacDonald

3

The fi ll of this feature consisted of ash and charcoal deposits; no further details of the contents are provided in the excavation notes. Blegen was uncertain as to the function of rooms 89 and 90. The structures are referred to as a bathroom area in the original manuscript ‘SE End of Court H (Bathroom)’, p. 8, but in The Palace of Nestor Blegen states that the buildings may represent the site of an olive press from as late as the 7th century bc (PN I, 296).

229

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

* 25 #

9 10

12 24

* # !! !! ! !

8 11

22 21

! !

13

* #

19

# ## #

23

#

17 16 15 14

20

#

18

$

#

#

2

7

#

6

#

#

3

5

#

4

1

²

#

0

2.5

5

10

15

1. Geometric vase 2. LHIIIC or Geometric handmade jug (Popham 1991, 320) 3. LHIIIC or Geometric handmade jug (Popham 1991, 320) 4. Geometric pottery 5. Geometric pottery 6. Geometric pottery 7. Geometric pottery

8.3 Late material from the east side of the palace.

230

20 Meters

8. Iron 9. Iron Spike 10. Iron Spike 11. Iron Spike 12. Iron ring 13. Iron 14. Geometric skyphos

Late Material from East Side of Palace

15. Geometric cup 16. Geometric cup 17. Geometric pedastal base 18. Tiles 19. Geometric pottery 20. Geometric pottery 21. Iron blade 22. Iron nail 23. Underfired pots (Popham 1991, 320) 24. Underfired pots (Popham 1991, 320) 25. Underfired pots (Popham 1991, 320)

Post-Bronze Age Stratigraphy areas surrounding the palace

the northwest area

Post-Bronze Age strata were also encountered in trenches excavated beyond the main palace buildings. To the southeast and near the modern road (fig. 8.1), the excavators uncovered a structure with a stone socle and crude brick superstructure (Blegen et al. 1973, 61–62). A thick surface layer covered this building. Below this was a layer of very black earth containing stones and chunks of crude brick with Geometric pottery. A whitish fill, believed to have eroded from the hilltop, rested below this level and above a deep layer of reddish dissolved crude brick. A thin black stratum was encountered above the structure’s stucco floor.

The Northwest Area preserves the most extensive evidence of post-Bronze Age architectural activity at the site. Although The Palace of Nestor provides a brief overview of this area in volume I (PN I, 291–93) and a somewhat more extensive description in volume III (PN III, 43–47), the published discussion is abbreviated and at times excludes information about post-Bronze Age stratigraphy, architecture and finds that were documented in Blegen’s unpublished excavation notes. The following stratigraphic analysis draws from information in the original excavation notebooks, supplemented with new data recovered through MARWP’s study of the architecture.

A deposit of dark black soil, like that found in courts 3, 42 and 88, was identified in a trench excavated east of Room 98 (fig. 8.1; Blegen et al. 1973, 26–27). The deposit and several Geometric sherds were encountered in what is described as a ‘hollow in stereo’ (ibid.).

In 1958, George Papathanasopoulos began excavation of the Northwest Area. During that season he supervised the excavation of trenches along the northwest wall of the Main Building, uncovering rooms 27, 82, 83, 84, 85, 86 and 87 (figs 8.4 and 8.5). Trenches 9, 10 (each 1.5 x 5 m) and 11 (approximately 1.5 x 7 m) lay immediately to the southeast of Circular Structure 87 and rooms 85, 84 and 83, providing the first glimpse of strata and architectural debris associated with these buildings. These trenches and trenches 12, 14, 15, 16 and 17 immediately to the northwest each exposed architectural remains resting above the yellow stratum that covers Bronze Age floors in the northwest area and elsewhere. In 1962, William Kittredge expanded excavations in the northwest area, uncovering a series of walls to the north and northwest of Papathanasopoulos’ 1958 excavations in his trenches 1, 3, 4 and 6 (fig. 8.4). Kittredge’s notes include stratigraphic sections and descriptions of finds encountered above the walls at the northwest extent of Room 85 and in the area to the north of Structure 87.

Post-Bronze Age material was also uncovered from the fill of the gateway located at the southeast edge of the plateau. The stratigraphy in this area was not well defined, but the upper part of the fi ll held a considerable mixture of pottery that included sherds of Geometric and Hellenistic or Classical date (ibid., 6). Another deposit of the sticky black stratum was encountered in an exploratory trench excavated just beyond the northeast edge of the plateau. The black earth is described as ‘filling a hollow’ in a disturbed layer approximately 0.65 m deep that rested below the surface soil (ibid., 63). The fi ll held tile fragments, a bronze coin, crude brick and stones as well as fragments of Geometric and Mycenaean pottery.

Trenches 9 and 10 The spatial distribution of the post-Bronze Age levels and finds as described in The Palace of Nestor suggests a resumption of activity across certain areas of the hilltop after a partial destruction of the palace (see fig. 8.1). In general, this activity clusters around the exterior of the palace walls and within the open courtyards, where Bronze Age building material was readily available for reuse, and there were fewer obstructions created by debris from the LH IIIB destruction. The exceptions to this are the halls of the megaron and the series of spaces around Room 41. Within the Throne Room, the black post-Bronze Age stratum was confined to the centre of the room, bounded on all sides by debris that extended from the plow zone to the Mycenaean floor. It is possible that the absence of a second floor directly above the centre of the Throne Room reduced the overlying destruction debris in this area, making it an attractive location for later activity. The fact that the post-Bronze Age levels are confined within the Bronze Age walls may further suggest that the walls were still standing to some height into the Dark Age and Archaic periods, when they continued to define areas of habitation.

In trenches 9 and 10 (fig. 8.5) the plow zone reached to 0.35–0.40 m below the surface. Beneath this layer, extending from 0.35 to 0.55 m below the surface, the soil was loose and light yellow in colour and contained sherds of unspecified type, as well as fragments of plaster and bronze. Below 0.55 m the soil was harder but maintained the colour of the overlying stratum. This harder-packed level extended to an average depth of 0.90–0.95 m below the surface in the southwest corner of Trench 9 and to an average depth of 0.90–1.10 m at the northern end of Trench 10, near Circular Structure 87. In the northern section of Trench 10, this layer of hard yellow soil was broken at a depth of 0.70 m by a floor composed of the yellow soil with many small stones. A block from the palace rested on top of this floor, just 0.15 m from the eastern edge of Circular Structure 87. Although Blegen describes the block as resting ‘against the eastern side of the circle [Structure 87]’ (PN I, 293), sketches from Papathanasopoulos’ notes place the block in the location where it remains in situ to this day, bedded on the same post-Bronze Age deposit as Circular Structure 87 (Papathanasopoulos 1958, 62, 63).

231

232

Wall ZZ

8.4 Plan of the Northwest Area showing Kittredge and Papathanasopoulos trenches. 0

Trench 21

Trench 13

83

1

2

4

Trench 11

Trench 19

85

6

86

Trench 15

Trench 12

Trench 17

Trench 18

8 Meters

Trench 10

Trench 14

87

Trench 16

Trench 20

Trench 9

Trench 1

Papathanasopoulos' Trenches

Kittredge's Trenches

Trench 8

Kittredge Trench 1

Kittredge Trench 6

Kittredge Trench 3

Kittredge Trench 4

Kittredge Trench 5

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

8.5 Detailed plan of Papathanosopoulos trenches.

233

Trench 21

Trench 13

83

0 0.5 1

2

3

Trench 11

86

Trench 15

Trench 12

Trench 17

Trench 18

4 Meters

85

Trench 19

Trench 10

Trench 14

87

Trench 16

Trench 20

Trench 9

Papathanasopoulos' Trenches

Trench 8

Trench 1

Post-Bronze Age Stratigraphy

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy Papathanasopoulos encountered the periphery of Circular Structure 87 in the northeast extent of Trench 10, near the intersection of trenches 9 and 14. The base of the building rested at a depth of 0.80 m below the surface, above a shallow layer of yellow soil (0.05–0.07 m thick). The yellow soil with stone inclusions underneath Circular Structure 87 was a continuation of that located below the floor in Trench 10 and is similar to descriptions of the stratigraphy found in the southwest end of Trench 9. Finds from this stratum (fig. 8.6 no. 56) include several sherds, some of which held painted decoration; unfortunately, the notes do not provide illustrations of these vessels. Within this same stratum of hard yellow soil, at the point where trenches 9 and 10 meet Trench 14, several fragments of what Papathanasopoulos describes as an offering table with painted plaster were uncovered (fig. 8.6 no. 47). The offering table fragments were recovered c.0.10 m below the lowest course of Structure 87, at a depth of 0.90 m below the surface (ibid., 99); they were not illustrated by Blegen. At the same depth, bones and large fragments of at least three vases with traces of fire around them were uncovered (fig. 8.6 no. 48). Additional fragments of the offering table, together with burned bones and sherds belonging to at least ten vases, mostly kylikes, were recovered at a comparable depth (0.90–0.95 m) in the southeast extent of Trench 9 (fig. 8.6 nos 49, 50, 51). The soil in this section of Trench 9 (nearest Circular Structure 87) is described as black, possibly reflecting the deposition of burned material, including the burned bone fragments just described.

In Trench 16, a layer of loose dark brown soil extended 0.40–0.60 m below the ground surface. Fragments of bronze and pottery (fig. 8.6 no. 52) were excavated from this stratum, which rested above Circular Structure 87; the top of the building was identified 0.60 m below the ground surface. More pieces of charcoal and brick (fig. 8.6 no. 54) were recovered in this trench, as in Trench 14, from the interior of the structure. Additional remains of burned brick as well as several bones were uncovered to the southwest of the structure at a depth of 0.60–0.80 m (ibid., 90–91). The soil below 0.60 m was yellow in colour, like that identified in trenches 9, 10 and 14. At 0.80 m below the surface the excavators uncovered animal bones, bronze, charcoal and many sherds, including three kylikes (fig. 8.6 nos 61–64). Papathanasopoulos believed that a floor existed at this depth (0.80 m), which roughly corresponds to the floor identified on the southeast side of Circular Structure 87 in Trench 10. Trench 11 Trench 11 (fig. 8.5) was excavated to the southeast of rooms 85, 84 and 83. A disturbed surface layer extended to a depth of 0.25–0.30 m. Sherds, masonry stones and a few fragments of bronze, charcoal and plaster were recovered from the disturbed layer. A stratum of hard-packed, light yellow soil lay below this layer, extending to a depth of approximately 0.95 m below the surface. Below this level, the soil was a greyish yellow colour. This stratum extended to approximately 1.27 m and rested above a layer with a more whitish colour. Material remains recovered from the yellowish layer included a large number of sherds, including many fragments of kylikes and pieces of pithoi as well as tiles (Papathanasopoulos 1958, 55, 56, 58, 61, 64). Numerous bones and traces of fire also were uncovered from this stratum (fig. 8.6 nos 65–68), including a burned skeleton of an animal, uncovered at a depth of 0.70–0.80 m below the surface towards the middle of the trench, just south of Room 85 (ibid., 61, 64, 113).

Trenches 14 and 16 and Circular Structure 87 Trenches 14 and 16 (fig. 8.5) uncovered the majority of Circular Structure 87 as well as the strata to its immediate southeast. Papathanasopoulos’ notes for these trenches indicate stratigraphy similar to that encountered in trenches 9 and 10. The plow zone extended to an average depth of 0.40 m below the surface (Papathanasopoulos 1958, 79, 82–83). In Trench 14, a harder yellowish stratum was encountered at 0.60 m. Large numbers of sherds, animal bones and charcoal were recovered near Circular Structure 87 at a depth of 0.6–0.70 m (fig. 8.6 nos 59– 60). Below this level, 0.70–1.10 m below the surface, the excavators also uncovered animal bones and teeth and many sherds, especially small cups and kylikes. Just 0.55 m east of Circular Structure 87, at a depth of 0.90–1.30 m, were uncovered several tiny metal plates of bronze and very small pieces of gold foil (fig. 8.6 no. 57). Immediately to the east, at a depth of 0.75–1.35 m, sherds – including many small kylikes – were uncovered while excavating drain b (fig. 8.6 no. 58). Additional sherds, plaster and the leg of a tripod vase (fig. 8.6 no. 55) were uncovered above the vertical stones that divide Circular Structure 87 near the boundary of trenches 14 and 16. The interior of the structure held remains of burned brick, sherds and charcoal (fig. 8.6 no. 53; ibid., 91–94).

Papathanasopoulos illustrated the spread of stones, tile and other terracotta remains in his notes (fig. 8.7). The deposit runs along the southeast wall of the building comprising rooms 83, 84 and 85 (ibid., 56). This orientation and concentration of material to the southeast of Building 83/84/85 may reflect the fall of roof tiles. The size of the material illustrated in a scaled drawing by Papathanasopoulos is consistent with fragments of Lakonian tile recovered during MARWP re-excavations. A Lakonian pan tile (fig. 8.8), with the same fabric as examples recovered by MARWP, remains in situ, embedded in a later wall (p.237) that overlies the southeast wall of Building 83/84/85, effectively sealing the underlying remains. The Lakonian tile in the wall that closed Building 83/84/85 may be a remnant of fallen roof material from that building. Roof tiles remain one of the most commonly recycled materials in building construction in Greece, and the incorporation of available material from a preexisting structure would seem likely. It is quite possible

234

76

8.6 Northwest Area finds and charcoal concentrations.

235 Charcoal Concentrations

Northwest Area Finds and Charcoal Concentrations

(Papathanasopoulos 1958

0

2.5

5

46. Geometric Sherds 47. Offering table fragments 48. Large sherds of vase 49. Kylikes 50. Burnt bones 51. Bones 52. Pot sherds 53. Burned brick and pottery 54. Burned brick 55. Sherds and leg of tripodic vase 56. Sherds, some painted 57. Bronze plate and gold foil

46

67

66 65

64 63 62 61

10 Meters

58. Sherds and small kylikes 59. Pot sherds 60. Animal bones 61. Kylikes 62. Pot sherds 63. Burned brick 64. Animal bones 65. Sherds, bases and bodies of kylikes 66. Bones 67. Animal skeleton 68. Animal bones

68

77

56

72 73

52 53 54 55

58 57

51

49 47 48 50

70 69 71

69. Iron knife blade 70. Kylix stems and coarse ware 71. Blackened bone or ivory 72. Animal bones (bovine?) 73. Pot sherds 74. Animal bones (bovine?) 75. Sherds and coarse ware pot 76. Burned animal bones (Stocker and Davis 2004, fig. 1) 77. Burned animal bones

60 59

74 75

Post-Bronze Age Stratigraphy

236

Trench 21

Trench 13

0

Trench 15

Trench 19

1

2

4

Trench 20

6

Papathanasopoulos' 1958 trench and Charcoal Concentrations

8 Meters

Trench 10

Trench 14

Trench 16

Scattering Illustrated by Papathanasopoulos

Trench 11

Trench 12

Trench 17

Trench 18

Trench 9

Trench 1

Papathanasopoulos' Trenches

Traces of Tiles

Charcoal Concentrations

Trench 8

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

8.7 Plan of Papathanasopoulos’ trenches, showing spread of stones, tiles and other terracotta remains in Trench 11 (1958).

Post-Bronze Age Stratigraphy hard-packed yellowish stratum described in 1958. Several dark layers identified within this level were interpreted by the pickmen as evidence of small fires. These men noted, however, that earth excavated from these layers was not baked and did not present evidence that is characteristic of a heavy fire (Kittredge 1962, 13). The absence of baked earth may indicate that the dark layers formed through the re-deposition of burned material rather than from the small fires suggested by the pickmen. An iron knife blade with bronze rivets was recovered from this level, indicating a post-Bronze Age date for the stratum (ibid.). Additional finds recorded by Kittredge as having been recovered from this layer include several kylix stems and fragments of polished, blackened bone or ivory (fig. 8.6 nos 69–71). The light brown level covered the top of wall A and was identified along its perimeter to a depth just below 1.0 m.

8.8 Lakonian pan tile embedded in later wall, Trench 11.

Trenches 3 and 4 that the scattering illustrated by Papathanasopoulos was in fact the remnants of the roof tiles that once covered Building 83/84/85.

In trenches 3 and 4, each approximately 3.6 x 2.45 m in size (fig. 8.9), a level of grey-black soil was uncovered beneath the level of the plow zone, extending to approximately 0.90–1.0 m below the surface. Below this layer was a thin (c.0.10 m) level of darker soil overlying a yellow stratum. Numerous red particles (fig. 8.10) thought to be brick were encountered above the yellow stratum within a layer of reddish earth that extended along the northwest edge of Building 83/84/85 and continued to the northeast baulk of Trench 3 (Kittredge 1962, 37, 51).

Trenches 12, 15, 17 and 18 Trench 12 was excavated parallel to and immediately northwest of Trench 11. The excavators encountered the edge of Room 84 and the southeast flank wall of Building 83/84/85 in the northwest end of this trench. The stratigraphy of Trench 12 follows that found in Trench 11. A disturbed surface layer extended to approximately 0.30 m below ground level. Below this was a yellowish layer with fragments of pottery and clay items found throughout. A similar light yellow layer was encountered at a depth of 0.20 m in trenches 15 and 17 during the excavation of Room 84 (Papathanasopoulos 1958, 84). Papathanasopoulos also describe traces of tiles, which were found within the yellowish layer of trenches 12, 15 and 17 (ibid.; fig. 8.7).

Kittredge describes a particularly large concentration of hard reddened lumps, characteristic of decomposed mudbrick, within this layer, near the junction of trenches 3 and 4. The location of this material near the northwest edge of Building 83/84/85 suggests that mudbrick originally formed part of the superstructure of the building. The concentration of mudbrick described by Kittredge was perhaps removed to the northeast of Building 83/84/85 during the construction of a later structure, which is now preserved solely by walls 83/84/86, wall Q, and a short section of unlabeled cross wall (discussed below, pp.242–45).

The presence of tiles within these trenches indicates that the proposed roof fall first encountered in Trench 11 also spread across part of Room 84. Trench 18 continued the investigation of Room 84. The excavators encountered wall E in the northeast end of the trench. Northwest of this wall were uncovered the remains of fire in a circle with a diameter of 0.30 m. Burned animal bones representing a single animal were recovered from the fill of this fire pit (fig. 8.6 no. 77; Papathanasopoulos 1958, 95–96).

A concentration of ash, charred animal bones (including large pieces believed to be bovine) and fragments of a large, coarse pot were recovered from a shallow depression (fig. 8.6 nos 74–75), 0.90 m in diameter and 0.20 m deep; this depression was located at the top of the yellowish stratum and just above wall E (some 0.90 m from the surface), 0.5 m from the south corner of Trench 3 and extending into Trench 4 (Kittredge 1962, 30–32, 37). Kittredge interpreted this feature as a firepit, but notes that red particles found at the top of the yellow layer, which were initially thought to be fire-reddened brick, were probably brought to this location, as there was no evidence for a conflagration here. Additional bones, charcoal, sherds and plaster fragments were uncovered in a concentration of ash and charcoal to

Trench 1 In Trench 1 (c.5.0 x 2.0 m; fig. 8.9), excavated by William Kittredge in 1962, the strata encountered were similar to what Papathanasopoulos described in 1958. The plow zone extended 0.50 m below the surface. Below this was a lighter coloured stratum with grey-black veins that reached to 0.60 m below the surface. A stratum described as light brown lay below this layer and probably corresponds to the

237

Kittredge's 1962 Trenches

Trench 1

8.9 Plan of Kittredge’s 1962 trenches, showing charcoal concentrations.

238

0 0.5 1

83

2

3

84

4 Meters

86

85

Trench 4

Trench 5

Trench 6

87

Trench 3

Charcoal Concentrations

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

0 0.5 1

83

2

3

239 4 Meters

84

86

85

Trench 4

Trench 6

Trench 5

87

Trench 1

8.10 Plan of Kittredge’s 1962 trenches, showing concentrations of mudbrick. Mud Brick Concentration

Kittredge's 1962 Trenches

Trench 3

Post-Bronze Age Stratigraphy

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy states that the pink soil is probably post-Bronze Age, making Building 83/84/85, which rests above this layer, also post-Bronze Age (ibid., 42). Unfortunately, no specific datable finds from the pink stratum are described in the excavation notes.

the northeast (fig. 8.6 nos 72–73), also located above wall E (c.0.90 m below the surface; Kittredge 1962, 27, 161).4 The proximity of this deposit of charred bone, ash and pottery to Circular Structure 87 and its position at the top of the same yellow stratum that underlies that structure may indicate a function associated with it. The feature is one of several similar concentrations identified in the upper levels of the yellow stratum near the structure. These features, like the deposits uncovered in Trench 1, may indicate multiple depositions of burned remains rather than independent fire pits.5 Several pots, including many kylix fragments, were found within the yellow stratum at 1.3–1.5 m below the surface (ibid., 27, 33).

Kittredge’s description of the stratigraphy below and above Room 85 sheds light on the stratigraphic relationship of Building 83/84/85 to Circular Structure 87. The latter rests on the same stratum of yellow soil that covers Bronze Age floors throughout the Northwest Area and much of the hilltop. The yellow soil extended below the pink stratum that was under the floor of Room 85, and that was below the structure’s foundations. Excavations by MARWP verified the stratigraphy that Kittredge and Papathanosopolous described below the foundations of the two buildings. The presence of the pink stratum below the foundations of Building 83/84/85 and above the yellow level suggests that a period of time elapsed between the constructions of the two structures, during which a thin layer of soil accumulated above the surface; this accumulation is assumed to be contemporary with Circular Structure 87. If this is correct, then Building 83/84/85, constructed sometime after this accumulation, must post-date Circular Structure 87. Neither the precise date of the two buildings nor the period of time that elapsed between their respective erections can be reconstructed from Kittredge’s excavation notes.

Trench 6 Trench 6 (7.5 x 1.6 m) continued the excavations of trenches 19 and 20 of 1958, which ended 0.25 m above the yellow stratum. The excavations of 1962 (fig. 8.9) uncovered the northwest side of Room 85 as well as walls below the foundation of Building 83/84/85. Kittredge provided a detailed discussion of the stratigraphy of building 83/84/85, expanding on the descriptions recorded in Papathanasopoulos’ 1958 notebooks. The stratigraphy Kittredge described at the southwest end of Trench 6 differs in some respects from that encountered southeast of building 83/84/85 in Papathanasopoulos’ Trench 11. A dark stratum with a pinkish or reddish cast rested directly above the floor of Building 83/84/85. Below this level Kittredge identified a second stratum, described as pinkish in colour. His descriptions occasionally conflate this stratum with the yellowish layer, and it appears that the separation between the two strata was sometimes difficult to distinguish. The pink stratum lay below the floor and the foundations of Building 83/84/85. In Room 85 it was above the yellow stratum that covers the Bronze Age floor levels here and elsewhere. In his notes, Kittredge

To summarise, the descriptions in the 1958 and 1962 notebooks document a layer of soil resting above the Bronze Age floors and several of the Bronze Age walls in the Northwest Area. The accumulation of this yellowish stratum suggests that the Northwest Area, like the rest of the hilltop, went out of use for a period of time following the destruction of the palace. Descriptions of this soil are similar to those that document the yellowish stratum in The Palace of Nestor, and indicate a continuation of the stratum into the Northwest Area.

4

The bones comprise part of a collection found in a wooden box and described in Stocker and Davis 2004, 183–84, n. 12. The excavation labels reproduced by Stocker and Davis record six groups of fi nds: ‘1) S2 1954. W Chasm. 10-12.7.54; 2) Room 7 bone: S3 Room of pithos heap of bones; S3 NW extension; 3) WK4. SW wall E. 17.5.62; 4) WK6. Fire on top of wall E. p. 161. 7.7.62. WK354; 5) EBW. 3.6.61. GPA. Σ. Tμήμα 2, 11.1.10-1.30. sel. 26. The label on the sixth group is damaged and incomplete, but seems to read ‘PNW’…’. The labels for groups 3 and 4 correlate with Kittredge’s notebook entries for Trench 4 and Trench 6. The bones from Group 3 were found within the shallow depression located above Wall E (Kittredge 1962, 31, 37). The bones from Group 4 were recovered from a second charcoal concentration found above Wall E (ibid., 27, 161). Stocker and Davis also note that the bones from Group 5 were found in the Northwest area, perhaps near Room 85 (Stocker and Davis 2004, 183, fig. 1). Group 6 probably derives from another charcoal and bone concentration found in the Northwest Area. 5 Stocker and Davis (2004) also suggest that the bones uncovered in the Northwest Area represent discarded sacrifices, though of Bronze Age date.

Sometime after the deposition of the yellowish layer, new construction began on the hilltop. Rooms 89 and 90, Circular Structure 87 and perhaps the building defined by the walls of Room 103 represent the only surviving architectural remains lying directly above the yellowish post-Bronze Age deposit. Building 83/84/85 rests upon a deposit of pinkish soil that overlies the yellowish deposit on its northeast end; this building therefore post-dates rooms 89 and 90 and Circular Structure 87.

Post-Bronze Age Architecture Three stratigraphically distinct building phases can be identified in the patchwork of walls overlying post-palatial levels in the Northwest Area. Assigned to Phase A is Circular Structure 87, probably a Late Geometric building that shares many similarities with circular structures found elsewhere in Greece. (Phase A appears as Phase 12

240

Post-Bronze Age Architecture in Cooper’s reconstruction of the overall sequence in the Northwest Area; see pp.66–67 above.) Phase B is represented by Building 83/84/85 (Cooper’s Phase 13, pp.67–68 above). Four extant walls, two partially preserved sections of stone-paved floors and a single column base comprise the surviving evidence for the temple plan of this phase. Two series of roof tiles – one Lakonian and one Corinthian – are assumed to have come from this building, both dating to the early Archaic period. To Phase C is attributed a single wall that extends perpendicularly across Building 83/84/85, overlying the earlier structure and sealing the underlying remains, and two associated wall fragments, wall Q and wall 82/83 overlying Building 82 (Cooper’s Phase 14, pp.68–69 above). A second set of Corinthian tiles is assumed to have been part of this building. Outside of the Northwest Area, a fourth phase of building, Phase D, is exhibited only in numerous medieval roof tiles found by Blegen in a deposit to the northwest of the Northeast Gateway (Blegen et al. 1973, 6–7).

on a thin layer of yellowish earth. In general, its width is formed by two or three fieldstones set side by side; in several places, however, larger fieldstones were set into the wall, spanning the entire width. A central cross wall composed of flat plates of stones set on edge divides the interior into two halves of slightly different size and, as preserved, rises to the level below the upper course of the surviving exterior walls (fig. 8.12). At the northeast, the interior wall face extends 0.24 m deeper in the ground than does the exterior. The base of the central cross wall also rests below the base of the structure’s exterior walls. The interior therefore was originally excavated slightly below the contemporary surface prior to laying the cross wall and the rubble stonework of the perimeter walls. The rubble walls held an internal earthen fi ll that probably extended to the top of the walls, providing a surface along the top of the structure and at the same time supporting the central dividing wall of stone plates. Elements of this core were encountered during the original excavations: Papathanasopoulos’ notes (1958, 11) indicate that the interior was fi lled with earth, charcoal, pottery and burnt mudbrick when excavated. It is possible that this material derives from objects set on the surface of the structure, perhaps within the context of ritual dining. Over time this material became embedded and preserved within the structure’s earthen fill.

phase a: late geometric period Circular Structure 87 (fig. 8.11) marks the first phase of post-Bronze Age construction in the Northwest Area. In plan, it forms an approximate circle measuring 3.10 m in diameter. Its circumference is defined by a double wall composed primarily of rough fieldstones c.0.10–0.40 m in maximum length, with several large chunks of baked mudbrick also built into its fabric (F. A. Cooper 1994, 12). The wall (max. thickness 0.35–0.40 m) is bedded directly

The function and architectural form of this structure are a curiosity. Circular buildings often hold special

8.11 Structure 87, from the southwest.

241

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

NE Face 191.81

WALL D

87

191.39 191.06 190.77

B

SW Face

0 89.44

1 m. 0

5 m.

8.12 Elevations of Circular Structure 87.

significance in Greek architecture. One class of round buildings, associated with ritual dining, is of particular interest for the interpretation of Circular Structure 87 (Cooper and Morris 1990). Robin Hägg (1983) provided the first treatment of this structural type, arguing that the platforms were used in ritual dining connected with ancestor cult (see also Ainian 1997, 122–23; Antonaccio 1995, 199–207; Lambrinoudakis 1988; for similar structures at Lefkandi, see Drerup 1969, 67, fig. 54). Burned material is frequently found on top of or in proximity to the circles, and animal bones and drinking vessels are sometimes found nearby. The stratigraphy and associated finds described in the notes of Papathanasopoulos and Kittredge share similarities with the circular dining platforms described by Hägg and others, and may suggest a comparable ritual dining function for Circular Structure 87.

back section remains of the building’s southern flank wall, wall 84/85/86. (For an explanation of the logic behind the naming of walls and rooms not labeled by Blegen, see pp.49–50 above.) The northern flank wall is represented only by a shallow robbing trench, the northern terminus of wall 84/85 and remnants of stone flooring that define the inside edge of the northwest flank wall: the triangularshaped pavement 83 and pavement 82. The outer edge of the ovoid column base B, located at the western edge of Room 83, represents the westernmost extent of the structure. The surviving walls of Building 83/84/85 are the remains of the structure’s socle. The layer of decomposed mudbrick that Kittredge identified near the building’s now robbed northwest wall (see above, p.237) suggests the presence of a superstructure of that material above the level of the socle. Three ashlars, preserving large swallowtail clamp cuttings, were found in secondary contexts at the site and may preserve part of the building’s stone cornice (see Part II, pp.271–72). The first-phase roof was of Corinthian type and is dated to the late 7th or early 6th century bc; this roof was replaced sometime during the early to mid-6th century bc by a roof of Lakonian style.

phase b: early archaic period Building 83/84/85 (c.17 x 4.6 m) represents the second construction phase in the Northwest Area. By all indications, this building was an impressive structure of some importance, with a reconstructed length-to-width ratio (1:3) that corresponds with those of 7th- and early 6th-century bc temple plans. The building’s Corinthianand Lakonian-style roofs distinguish it from other buildings of post-Bronze Age date found on the hilltop.

The Walls of Building 83/84/85 Wall 85/87 lies above a shallow deposit of pinkish soil that is some 0.12 m deep at the wall’s northwest edge. The uppermost course of the adjacent Circular Structure 87 is actually higher than the preserved elevation of wall 85/87 (fig. 8.14). Given the minimal stratigraphic separation of the two buildings it seems probable that Circular Structure 87 was visible and perhaps still in use when

Four partial walls establish the outline of the temple plan (fig. 8.13). Wall 85/87 defines the northeast terminus of the building and stands just 0.6 m from Circular Structure 87. Two additional cross walls (84/85 and 83/84) divide the interior of the building into three chambers. Only the

242

0

1

2

4

6

8 Meters

Not Preserved

Extant Architecture

Temple Architecture

Wall ZZ

Post-Bronze Age Architecture

8.13 Remains of the early Archaic temple and Circular Structure 87.

243

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy the later wall was erected. The narrow passage between the two structures (only 0.6 m wide) would have limited access to Building 83/84/85 at this end, and the entrance is assumed to have been at the southwest end of the building. The northwest–southeast orientation may relate to the topographical constraints of the Northwest Area and a desire to site the building near Circular Structure 87, which, as noted above, may have marked a locus of earlier religious activity and ritual dining.6

from an initial elevation of c.191.4 m at the northwest to 191.58 m at the junction with wall 84/85/86. Flank wall 85/86 bonds with wall 85/87 at its northeast end; only part of its southwest end is preserved (fig. 8.13). The wall, like 85/87, is of double curtain construction and measures between 0.86 and 0.89 m in width. It preserves a length of approximately 7.5 m, with the southwest edge running below wall 83/84/86. A few stones protrude from below the overlying wall into Room 83. The flank wall must originally have extended to a point adjacent to column B, which lies approximately 6.15 m to the southwest. Like wall 85/87, the outer and interior facing is carefully constructed of fieldstones that are larger than the stones that comprise the wall’s core. The northeast end of the wall is preserved to an elevation of 191.51 m and rises to an elevation of approximately 191.71 m at its southwest extent.

Wall 85/87 varies in width, ranging from 0.80 to 0.85 m, as do other walls of Building 83/84/85. The northwest edge of wall 85/87 originally extended approximately 1.0 m further to the northwest, where it joined the now robbed northwest flank wall. The building technique employs double curtain construction throughout: a single line of smaller stones (averaging 0.25 m in length or less) set between the outer facings forms the wall’s core. Only two courses of wall 85/87 survive, with a combined height of c.0.27 m. The elevation of the wall varies across its length

Cross wall 84/85 is located four metres from the back end of the building (fig. 8.13). The wall parallels wall 85/87 and abuts wall 84/85/86 at its southeast end, where it lies above a layer of pinkish soil approximately 0.20 m thick. The wall is preserved to c.3.83 m of its original length, with a width of c.0.89 m, and preserves part of the northwest edge of the robbed northwest flank wall. It exhibits the same double curtain construction as the two previously discussed walls and preserves two courses totaling a little over 0.20 m in height. This cross wall creates in the rear of the building a small rectangular chamber approximately 3.1 x 2.6 m in size.

6

North–south orientations can be identified in several early Archaic temples. The Archaic temple of Apollo at Bassai (F.A. Cooper 1996, 85), the nearby temple of Artemis Kotilon (ibid.), the temple of Apollo at Halieis (N. Cooper 1990, 77) and the Archaic temple at Mycenae (Klein 1997, 256) were oriented along a roughly north–south axis, with the Kotilon and Halieis temples opening at their south ends. The Pamisos temple near Ayios Floros has a northeast–southwest orientation like that at Englianos, however, the entrance was positioned at the northwest end (Valmin 1938, 426).

ashlar

NW Face

WALL B

WALL B

SE

84/85 Drain b

85/87 Pav b

ashlar

90.73 90.58

Drain

89.44

89.90

ashlar

0

1 m. 0

5 m.

8.14 Elevations of wall 85/87.

244

Post-Bronze Age Architecture Cross wall 83/84 is located below a later wall, 83/84/86 (obscured in fig. 8.13), and joins the southeast flank wall c.6.4 m from the back, east corner of the building. As with the other walls of the building, the construction technique was one of double curtain construction with large fieldstones forming the outer and interior facing; the wall is preserved to two courses. The northwest end of the wall abuts the robbing trench where the northwest flank wall once ran.

measuring approximately 0.03 m at the top to 0.023 m at the lower end. The catalogued fragments are pinkish white in colour with a Munsell value of 7.5 YR 8/2. They are well crafted from medium clay with some small inclusions that give a gritty texture to the surface. The pan tiles were hard fired and, overall, exhibit relatively little weathering. Although they share common dimensions, slight variations in shape and workmanship can be detected in the preserved fragments, deriving from individual adjustments that were made before the tiles dried completely. In general, the surfaces are undecorated, though some fragments preserve a single finger stripe that runs lengthwise along the upper surface of the tile. Both rounded and angular corners are preserved. The well-preserved undersides are relatively uniform in composition, exhibiting the same rough texture as the upper surface.

This cross wall defines the western side of a second small chamber (84) that measures approximately 1.6 x 3.2 m in size. A small patch of stones abutting the northwest face of wall 84/85/86 is all that remains of the stone pavement that once covered Room 84. The pavement is composed of a single course of flat stones (0.20 m in average thickness) that extends in a line from the southwest face of wall 84/85 to the southeast face of wall 83/84. A second remnant of pavement, triangular in shape, is preserved in the northeast corner of Room 83. This floor rests above a thin (c.0.10 m) layer of pinkish soil. The line of the northwest edge of the pavement is preserved where it originally abutted the robbed northwest flank wall. This patch of pavement measures 2.25 m on its southern side and 1.8 m near the northwest edge of wall 83/84. It is composed of a single course of flat stone plates that take the form of a flagging but are ungraded in terms of thickness (F. A. Cooper 1994, 12). Several whole and partial stacked mudbricks are embedded within this pavement. The upper surface of the pavement rests at 191.49 m at the northwest edge, sloping slightly down at the southeast to an elevation of 191.47 m (ibid.).

In profi le, the pan tiles have a slight curvature along the edge but are flat at the base and lie level when placed on a horizontal surface. At the upper end of the tile (nearest the ridge beam), the upward curvature at the outer edge angles c.0.015 m over a distance of 0.015 m. At the lower end of the tile (down-slope, or farthest from the ridge beam) this curvature is more gradual, rising 0.015 m over a distance of some 0.053 m. A lip was moulded into the upper surface of the pan tiles, extending along the lateral sides to prevent water from seeping over the edges. The curvature along the outer edge and the height of the lip vary slightly throughout the catalogued examples. The lateral sides of several examples are slightly bevelled along the tile edge; this concave edge extends along much of the length of the tile. The function of this bevelled edge

The Archaic Corinthian Roof Fragments of 131 pan tiles and two cover tiles comprise the earliest Corinthian-style roof that can be reconstructed at the site. The Englianos tiles must have belonged to a significant structure, although there is no firm evidence to indicate which building on the hilltop was roofed in this way. The total weight of the pan tiles recovered by MARWP from Blegen’s backfi ll is 55.25 kg and represents only a small portion (4–5 sq.m) of the tiles needed to cover a roof of any size. The original combined weight of the pan and cover tiles would have required the support of a well-built timber roofing system and sturdy walls. In the original excavations at Englianos, Blegen’s excavations recovered several Corinthian tiles from the northern end of Court 42 (fig. 8.3 no. 18) and above the floor debris of rooms 89 and 90 (fig. 8.2 no. 36) (Hope and MacDonald 1953, 27, 29, 31). Rooms 89 and 90 were small structures that were poorly suited to the grand Corinthian tiles recovered by MARWP; the temple that is defined by rooms 83/84/85 represents their most likely destination.

8.15 Archaic Corinthian pan tile fragments.

Several fragments of pan tiles join (figs 8.15 and 8.16) to reconstruct an original length of c.0.72 m, with a width that tapers from 0.36 m at the top to 0.26 m at the bottom. The tiles’ average thickness also varies from top to bottom,

8.16 Reconstruction of complete Archaic Corinthian pan tile.

245

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy is uncertain, but it typically corresponds with sections of the tiles that exhibit an elevated lip.

1983, 25–32) and the temple to Apollo at Corinth (Roebuck 1955, 156–57; Robinson 1976, 231–35; Rhodes idem; Winter 1994, 14–15). Traditional reconstructions of the hip tiles from Corinth suggest that they were constructed in two separate pieces: a roughly square piece that covered the joint between the cover on the hip and lateral roof slope, and a modified pan tile that served as the pan tile elements on either side of the ridge of the hip. Research by Philip Sapirstein (personal communication), however, suggests that the Corinth tiles, like those from Englianos, were constructed as a single unit. The tiles were formed using a simple wooden mould. The articulation of the pan and cover elements derived from the profi le of the form and the technique of the tiles’ manufacture. The hipped tiles from Englianos probably were constructed in a manner similar to that proposed by Sapirstein. The differences in the articulation of the pan and cover elements simply represents a difference in the form used in their creation. The surviving fragments of hipped cover tiles indicate that

The two fragmentary cover tiles exhibit a fabric and composition similar to that of the pan tiles. The bestpreserved example (fig. 8.17 top) measures 0.20 m in width and 0.18 m in length, with a thickness of 0.02 m. Two edges are preserved, exhibiting a complex, three-dimensional configuration: from the top, a slight fold downwards runs diagonally along the cover tile, with a bend part-way along the axis converging at the centre of the tile. The peculiar shape of the piece fits at the intersection of the diagonal pitch of a hipped roof. A larger tile fragment exhibiting similar curvature but lacking preserved edges is illustrated in figures 8.17 (bottom) and 8.18. Tiles of this type are known from the early Archaic temple to Poseidon at Isthmia (Broneer 1971; Rhodes 1984, 98– 102; Rhodes 1987, 477; Hemans 1989, 251–66; N. Cooper

8.17 Fragmentary Archaic Corinthian cover tiles for a hipped roof (right), and reconstructions (left) showing their positions on the roof.

246

Post-Bronze Age Architecture

8.18 Fragmentary cover tile for a hipped roof (ID 03823).

8.19 Fragments of Archaic Lakonian roof tiles.

the Corinthian-style roof carried a hip on at least one end. Hipped roof systems such as those found on the Corinth and Isthmia temples were used most widely from the midto later 7th century bc (Wikander 1988, 216; Mallwitz 1981 and 1983). The use of a hipped roof and similarities of these tiles with the Corinthian tiles from Corinth suggest a date for the Englianos roof (late 7th–early 6th century bc) that is comparable to the Isthmia and Corinth temples.

The Archaic Lakonian Decorative Revetment A comprehensive selection of fragments from the Lakonian decorative revetment was recovered in the MARWP investigations. These include a fragment of a painted semi-circular lateral antefi x and three fragments of an acroterian disk. Generally, the decorative revetment is better preserved than the tiles, to a degree in which most details of design can be established.

The Archaic Lakonian Roof Tiles A single fragment of a lateral antefi x was recovered from Blegen’s backfi ll. The object has a maximum preserved height of 0.057 m, width of 0.097 m, and maximum thickness of 0.012 m. It is impossible to reconstruct the overall dimensions of the original piece, as no edges are preserved. The fabric is a reddish-yellow colour comparable to that of the pan and cover tiles described above. The piece carries a black matte painted decoration of concentric rings that vary in thickness across the radius of the band. The black bands were drawn over a beige slip that covers the entire face of the fragment (figs 8.20 and 8.21). A raised lip, 0.05 m in width, curves just below the innermost black band. The rear side of the fragment is unfinished, though it shows the arcs of the face moulding. The interior raised lip has a short radius of 0.04 m (measured from the lower central section of the antefi x), suggesting that the fragment preserves the lower central portion of the original antefi x plate.

A second, later Archaic roof of Lakonian type replaced the Corinthian roof on Building 83/84/85. (Papathanasopoulos’ notebook contains a description and sketch of the Lakonian tiles as distributed along the southeast wall of the temple in Trench 11; see pp. 234–37 above.) Thirty-nine fragments of pan and cover tiles (fig. 8.19), as well as decorative revetment, have been recovered. The fabric of the tiles has a reddish-yellow colour (5 YR 7/6 Munsell value) and is composed of a fine clay with few inclusions, resulting in a soft texture on the outer surface of the tiles. Remnants of a thin reddish slip can be identified on a few of the fragments, though in most cases weathering has effaced the slip. Collectively, the tiles are well enough preserved to allow a good approximation of size and design. The lateral edges of the cover tiles angle slightly to account for the curve of the upper surface of the pan tiles beneath. The cover tiles measure c.0.025 m in thickness at this edge, decreasing in thickness toward the upper curvature of the tile. The pan tile fragments have a shallower curvature and preserve an angled edge on their lateral and upper ends that is similar to that identified on the cover tiles. A slightly raised lip is preserved at the edge of several pan tiles just before the angled outer edge, serving as a drip to prevent water from seeping over the sides of the tiles. The thickness of the pan tiles is comparable to that of the cover tiles at the lateral edge, thinning slightly towards the centre of the tile.

The shape of this fragment is similar in design to semicircular antefi xes from the early 6th-century bc temple to Artemis Ortheia in Sparta but differs in its painted decoration. The outer antefi x plates of the Ortheia examples replicated the semicircular shape of the cover tile (Winter 1994, 98); they were tied to the covers at their outer edge and hung free at the bottom to cover the void created by the upward curve of the eaves tiles. A similar situation may be reconstructed at Englianos. The decoration suggests that the antefi x plate had a curvature similar to that reconstructed for the Lakonian cover tiles, 247

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy decoration. The upper edge of the fragment is badly damaged but seems to preserve part of the disk’s outer edge, which was moulded by hand and angles slightly along the interior side of the fragment. An approximate overall radius of 0.5–0.6 m can be reconstructed from the shallow decoration. A third fragment, identified during the early phase of investigations and noted by F.A. Cooper, was not among the inventoried examples. Descriptions by Cooper (personal communication) and a field photograph of the profile suggest an acroterion fragment with cavetto moulding that likely formed part of the disk’s interior.

8.20 Fragment of an Archaic Lakonian lateral antefix (ID 05529).

The evidence thus indicates a disk consisting of a series of cavetto mouldings towards a centre comprising a flat panel or disk; a wide torus follows, with several rings of smaller torus mouldings along the outer edge. The survival of beige slip on fragment two makes it probable that all of the moulded decoration was painted. The four components of the Archaic Lakonian roof – pan and cover tile, antefi x and akroterion disk – form a set. The stylistic date falls in the mid-6th century bc, approximately 50 years after the Corinthian roof. The use of decorative revetment in Phase B is particularly diagnostic in terms of its function, as it precludes a domestic use. Archaic Lakonian roof revetment like that found at Englianos was used almost exclusively on temples or other buildings within a sanctuary. The association of these fragments with Building 83/84/85 points to a similar function there and supports the reconstruction of a replacement roof on a small monostyle antae temple or shrine.

8.21 Reconstruction showing position of fragmentary antefix.

and the overall appearance of the antefi x would have been similar to those from the Orthia temple. The decoration of our antefi x is a variant of the Laconian Antefi x Type I, variant 3 described by Nancy Winter and dated to the late 7th–early 6th century bc (ibid., 107). The decoration of the Englianos example deviates from the painted tongue decoration described by Winter for the Laconian Type I Antefi x but may represent a local and later variation of the type (ibid., 106); a similar decoration is preserved in a fragment of disk acroteria recovered near the Asklepeion at Messene (Themelis 1994, 150, fig. 7, pl. 49f). This acroteria dates to the 6th century bc and carries alternating bands of raised, matte painted decoration similar in design to the Englianos antefi x. One of the fragments of acroteria formed part of the outer edge, and the second came from the interior of the disk. The fabric is reddish-yellow, with a beige slip on the exterior. The first fragment (fig. 8.22), measures approximately 0.082 m in length and 0.04 m in height, and exhibits a torus moulding 0.016 high and 0.017 m deep. The shallow curvature of the moulded decoration indicates a radius of approximately 0.367 m. The fragment is damaged on all sides and did not form part of the rim but would have occupied a position interior to the outer edge.

8.22 Acroteria fragment with torus moulding (ID 02262).

The second fragment preserves concentric rings of moulded decoration that are relatively shallow when compared to the torus moulding just described. The fragment in figure 8.23 contains three concentric rings with a very shallow convex moulding; except for traces of beige slip, not much more can be said about painted

8.23 Acroteria fragment with shallow moulding in concentric rings (ID 00191).

248

Post-Bronze Age Architecture Wall ZZ

width makes this wall thicker than the other two walls but is appropriate for a cross wall in what is here called Building 83/82/81. This Phase C building was probably a replacement temple built on top of the front portion of its predecessor, the Phase B early Archaic temple. As suggested by Cooper, the width of the later temple (c.9.74 m), is roughly double that of its predecessor (4.6 m); a length of c.30.1 m, extending to the area of Room 78 of the Southwest Building, where the wall elevations are levelled to a height of c.191.9 m, would double that of the earlier temple – in effect, enlarging the original plan to a hekatompedon.7

A long stretch of wall extends south from the northwest corner of Room 27, turning at an oblique angle to meet the northwest corner of Room 21; another section of the wall crosses Courtyard 88. At its northern end the structure is associated with a larnax lying adjacent to it and having the same orientation. This wall was not mentioned by Blegen in The Palace of Nestor at Pylos, though it and the position of the larnax appear on the Key Plan in volume I. It was also, unfortunately, all but overlooked by MARWP during investigations of the Northwest Area. Cooper discusses the wall in detail in chapter 1 (pp.70–73 above), taking into account the information in the notebooks of George Papathanasopoulos and such dimensions and photographs as were recorded by MARWP. The position of the wall, skirting the Geometric Circular Structure 87 and closely aligned with the southeast wall of Building 83/84/85 of this phase, as well as the restored southeast wall of its Phase C successor temple (see below), suggests that it may be a temenos wall separating the religious structures to the northwest from secular areas to the southeast.

The Second Corinthian-style Roof A second, later set of 47 Corinthian-style rooft iles are a later version of the Archaic Corinthian roof tiles of Phase B. Like those, these tiles must have covered a well-built structure of some size. Continuing Cooper’s interpretation that the evidence argues for a replacement of the Archaic temple, the new structure, Building 83, is where the tiles must have been used. The fabric of the tiles is composed of a very pale brown (10 YR 8/3 Munsel value) medium clay that was fired to produce a hard outer surface. The fragments preserve large sections of both the upper and lower ends, allowing a reconstruction of overall widths (it was not possible to reconstruct overall lengths): the tiles were 0.32 m wide at the upper ends, tapering to 0.28 m at the bottoms (figs 8.27 and 8.28). The thickness of the tiles was not uniform throughout: the lower ends measure approximately 0.030 m in thickness, increasing to 0.033 m. A raised lip rose above the surface of the tiles at the upper end, and a corresponding convex bulge on underside of the lower end fit onto the raised lip of the underlying tiles, forming a tight joint where the pan tiles overlapped.

phase c: later archaic period Phase C of the post-Bronze Age architectural activity at Pylos is preserved most obviously in Wall 83/84/86, which measures approximately 0.85 m in width and has a total length of 9.5 m. Nearly one third of this later construction was built on top of wall 83/84, the cross wall of the Archaic temple (fig. 8.24). The northwest end of the wall extends beyond the earlier temple’s robbed northwest flank wall, and the southeast end similarly overlies and extends beyond the southeast flank wall of the building, effectively sealing the remains of the underlying structure. The wall was carefully built, with the southeast terminus having the appearance of a quoined end, suggesting a building corner; its matrix consists of small stones in mud mortar.

In section, the pans are flat across the bottom, with edges that curve upwards slightly. Along the lower end the upward curve rises a total of 0.025 m along the outer 0.09 m edge of the tile; along the top the curve is slightly greater, rising 0.03 m over a distance of 0.09 m at the tile’s outer edge. The upper surface is relatively smooth and preserves two types of shallow relief. A depressed line running along the surface of several fragments divides the edge of the tile from the slightly lower central portion. The lines are approximately 0.03 m from the lateral edge and 0.05 m from the lower end; they meet at the corners to form a sharply defined corner. The impression may have served a functional purpose, for the edge of the sunken central portion of the tile would have funneled water down the slope of the roof and away from the edge, thereby preventing leaks along the joint between the cover and pan tiles.

The space to the southeast of wall 83/84/86 is interrupted by the wall stubs of Bronze Age palace constructions. To the northwest is a wall fragment (Wall Q, 3.5 m in length) that has an orientation perpendicular to that of wall 83/84/86 and a similar width (0.85 m) and elevation (191.90 m). Wall Q does not appear in Blegen’s published plans of the Northwest Area, but is so labeled in notebooks. Although it looks at first glance to be a continuation of wall E immediately to its northeast, it is slightly skewed to the course of that wall and is higher in elevation by 0.60 m. Wall S and the modern drain built on top of it obscure the southeast portion of wall Q. Probably belonging to the same building as Wall 83/84/86 and wall Q is wall 82/83, which overlies obliquely the southeast wall of Building 82. It runs parallel to wall 83/84/86; the southeast portion falls within the thickness of a restored flank wall running perpendicular to it and parallel to wall E and has a top elevation similar to that of the other two wall fragments, at 191.90. Its 1.25-m

7

The proposed southwest end of this temple is situated just northeast of a significant concentration of burned animal bones illustrated by Stocker and Davis (2004, fig. 1).

249

0

Concentration of burned bone noted by Stocker and Davis 2004, fig. 1

1.5

3

6

9

8.24 Northwest Area building phases.

250 12 Meters

NW Area Building Phases

Wall Q

Wall ZZ

Not Preserved

Phase C - Extant

Phase B - Extant

Phase A - Extant

Building Phases NW Area

Wall 83/84/86

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy

Post-Bronze Age Architecture A second design element, consisting of a meandering finger impression that runs along the width of the tile, is present along the upper and lower ends of most of the surviving fragments of this type at Englianos (figs 8.25 and 8.26). It is confined to the central depressed surface of the pans at the upper edge but extends beyond this central area at the lower end of the tile. The decoration would have been obscured at the upper edge of the pan by the overlying tile. The depressed surface and finger impression at the upper edge of the tile worked together, preventing water from travelling underneath the overlying pan tiles and slowing the force of water as it flowed down along the centre of the pan. At the lower edge, the finger grooves functioned as a hydraulic ‘spoiler’, which reduced the force of the water as it funneled down the slope of the roof; at the same time the impressions channeled the flow of water away from the outer edge of the tile, thereby minimising leaks along the tile edge. Similar finger grooves occur on a number of pan tiles recovered at Corinth.8 Unfortunately, the tiles from Corinth, like those recovered at Englianos, were found in secondary deposits, and little information regarding their original context or date is provided in the museum’s inventory cards.9

phase d: medieval period Tiles of the medieval period are briefly mentioned in The Palace of Nestor. Blegen’s unpublished manuscript presents a slightly more detailed discussion, providing descriptions of both the tiles and their associated architectural remains. The tiles were recovered in 1958 and 1959 from an exploratory trench excavated between the acropolis and the tholos tomb. Blegen’s manuscript describes two distinct tile shapes from this trench. In both cases the tile dimensions are very close to those of medieval types I and II described below, making it likely that the tiles recovered by MARWP derive from the same location. The larger of the two types is described as broad and almost flat, with a slight curve. The second type was narrow and nearly hemispherical in section. Blegen attributed the tiles to a small building located northwest of the Mycenaean gateway; associated pottery was dated to the 11th and 12th centuries by Alison Frantz (Blegen et al. 1973, 6 and fig. 121; Blegen n.d., 3, 9 and 10 (‘Wall Hunt 1958–1959’)). According to Blegen’s excavation notes, the larger, flatter tiles and several stone slabs from a pre-existing road formed a pavement for the medieval building; the smaller hemispherical tiles were believed to be the roofing material for the structure (Blegen n.d. (‘Wall Hunt’), 9). Davis and Stocker (2013) recently re-dated this material to the late 12th or early 13th century, noting the presence of glazed sherds from a ‘post-Byzantine use of the area’. It is more likely, however, that these tiles either represent multiple 8

The tiles are located in the museum storeroom under the following catalogue numbers: FP 91,153,168,183,184,186,190 and 217. 9 My thanks to Robin Rhodes for his assistance in retrieving the original inventory information from the museum archives. The inventory cards indicate that the tiles were recovered from a variety of secondary contexts. They were found within grave #6 on the road to Acrocorinth and near or within several tombs of the Roman period. Other examples derive from unspecified contexts.

8.25 Reconstruction showing shallow relief.

8.26 Fragments of Archaic Corinthian-style rooftiles.

251

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy top and 0.30 m at the bottom, with an average length of 0.70 m. The tiles’ thickness ranges from 0.025 m at the top to 0.05 m at the bottom. Finger-striped impressions are present on many of the pan tiles, in each case consisting of parallel lines that run nearly the width of the tile at the upper and lower ends (fig. 8.29). A similar impression was moulded into the upper and lower ends of the surface of the cover tiles (fig. 8.30). Like the impressions found on the pan tiles, the finger stripes on the cover tiles consist of parallel lines approximately 0.01 m wide that run nearly the width of the tile but do not extend to the very edge. The indentation appears to have served a functional purpose, for at the lower end of the cover tiles, the finger marks directed water down the outside of the tile and away from the end where the cover tiles overlapped.

8.27 Bottom of medieval tile showing pitting.

medieval-period building phases or (more probably) indicate the existence of more than one medieval building at the site.

The cover tiles are peculiar in design, carrying a vertical profi le that consists of a high segmented arch that converges at the peak to form an ogive pointed arch (fig. 8.31). The cover and pan tiles share common widths. The crude grooves at the edges of the pans receive the lateral edges of the covers. The interior lateral edges of the covers abut and lock against a low raised lip at the lateral edges of the pans. In this distinctive system, the pans do not touch; the gap between them is defined by the width of

The tiles recovered at Englianos are very similar to Frankish roof tiles found elsewhere. As do the Englianos examples, many of the pan tiles uncovered at Corinth contain finger impressions that extend horizontally across the concave surface of the upper end (Williams and Zervos 1992, 142– 45, pl. 33; Williams, Snyder, Barnes and Zervos 1998, 240– 41). Davis and Stocker suggest that the dimensions and decorative swirls found on the medieval tiles recovered by Blegen are ‘characteristic of tiles of the Middle Byzantine and Frankish periods’ (Davis and Stocker 2013, 684). The smaller pans have a similar treatment that runs for part of the length of the convex side of the cover, in a manner very similar to the type II tiles recovered from Pylos. Many of the Frankish tiles catalogued in MARWP’s survey of the Morea carried the same finger striping as the Englianos examples, along with severely pock-marked undersides and heavily weathered exteriors (F. A. Cooper 2002, 23–25). The severe pitting identified in the tiles from Englianos, Corinth and the Morea survey may be taken as a characteristic feature of medieval tiles (fig. 8.27), deriving from the method of manufacture. The fabric of the tiles contained large, coarse inclusions that often broke away during or after firing. A bedding of straw separated individual tiles in the kiln, and impressions of this straw bedding are preserved on the undersides of several of the fragments. The combination of the coarse inclusions and straw bedding gives a rough and pocked character to the tile surfaces that is particularly evident on these undersides.

8.28 Reconstruction of Type I medieval pan and cover tiles.

Type I Medieval Tile The better preserved of the two styles of medieval roof types from Englianos is represented by 316 fragments of pan and cover tiles (fig. 8.28). The catalogued examples are composed of a medium clay in three variations of Munsell colours: very pale brown (10 YR 8/3 Munsell), pale yellow (5 YR 8/3), and red (2.5 YR 5/8). Despite differences in colour, the tiles have similar dimensions and construction details. The pan tiles, on average, are 0.38 m wide at the

8.29 Type I tiles showing finger-stripe impression.

252

Summary are segments that measure 0.07 m at the lower end of the tile widening to 0.11 m at the top. The final two segments measure approximately 0.03–0.04 m at the bottom and 0.10 m at the top along the upper end. The cover tiles do not have the pronounced ogive arch found on the thicker tiles described above. The tiles preserve pocked undersides and a coarse, gritty texture similar to the type I tiles. Only one fragment preserves evidence of a finger-striped impression, running along the upper surface of the cover. The parallel impressions found on the type I roof are not present in the examples catalogued for type II. 8.30 Cover tile showing finger-stripe impression.

Summary The evidence for renewed construction on the Pylos hilltop after the end of the Bronze Age is extensive. Tiles of the early Archaic through the medieval period attest to a long, though sporadic pattern of site use. There is clear evidence for new construction and the re-use of existing architecture across much of the site in the late Dark Ages. In general, this construction took the form of small domestic buildings built within, above and near the remains of the palace. The buildings were simple rubble structures that were sited opportunistically, typically occupying positions above the open courtyards and halls of the palace, where destruction debris was least extensive.

8.31 Reconstruction of cover tile showing ogive profile.

the overlying cover tile. This differs from typical tile roof systems, in which the lateral edges of the pan tiles abut and a narrower cover tile spans a width necessary to close the joint between the pans (F. A. Cooper 2002, 24). Cooper identified similar tiles at Frankish sites in the Morea and on extant Romanesque church roofs such as at Vézeley in Burgundy (ibid., fig. 1). The finger impression may have been necessary to prevent water from seeping under this high arch and behind the pan tile below. Like the 4thcentury bc Corinthian pan tiles, the finger impression on the upper end of the medieval pan tiles also prevented water from seeping over the rear edge of the tile. The finger patterns along the lower end serve as ‘hydraulic spoilers’, functioning in a manner similar to that described for the later Corinthian tiles discussed above. Type II Medieval Tile Tiles from the second type of medieval roof are smaller but maintain several of the characteristics identified in the thick type I tiles just described. As reconstructed, the tiles were 0.26 m wide at the top and 0.16 m at the bottom, with a reconstructed length of approximately 0.48 m (figs 8.32 and 8.33). In profi le, the fragments vary in thickness from side to centre and from top to bottom. At the lower lateral edge, the fragments have a thickness of 0.035 m that tapers toward the arched centre of the tile to 0.02 m. At the lateral edge of the upper end of the tile the thickness measures 0.04 m, tapering to 0.03 m at the centre. The surface of the cover tiles is segmented lengthwise into five sections, possibly preserving the width of the boards used in the construction of the tile moulds. The thickest segment is located along the centre and measures 0.10 m in width at the lower end and 0.15 m at the top. On either side

8.32 Type II medieval tiles.

8.33 Reconstruction of a Type II medieval tile.

253

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy A different situation can be identified in the Northwest Area. Unlike the open courtyards to the south and east, there is no evidence to suggest that domestic buildings were erected at this location. Instead, the Northwest Area was reserved for a distinct purpose early in the post-Bronze Age period. By Late Geometric times the location appears to have been the site of ritual dining activity centred on Circular Structure 87. Numerous concentrations of charcoal, bone and drinking vessels that were uncovered in shallow pits excavated into the surface contemporary with that structure attest to this activity. The deposit situated just northeast of the ashlar block shown in figure 6 is particularly remarkable for the concentrations of material it contained, as well as for its proximity to the circular structure. Additional sherds, including the leg of a tripod vase, were uncovered within the fi ll of Circular Structure 87, and it is tempting to see this material as the remains of objects used in dining rituals.

our understanding of the region and the post-Bronze Age history of one of the most thoroughly studied sites in Greece. The material evidence uncovered during MARWP’s investigation provides a reappraisal of the later history and architectural development of Englianos, as well as its correspondence with architectural developments elsewhere in Greece.

The construction of temple 83/84/85 during the 7th century bc further distinguished the Northwest Area as a sacred location, an appellation that may have been reinforced by the erection of temenos wall, wall ZZ. The building constitutes a structure of some importance, and its location near the remains of the Bronze Age palace places it among other early Archaic temples situated above or near Bronze Age palatial remains. The construction of the temple represents an intensification of earlier cult and ritual dining activity associated with Circular Structure 87. That structure may have been incorporated into the new sanctuary, where it perhaps functioned as an altar associated with the early Archaic temple. The concentration of bones uncovered above wall E (fig. 8.6 nos 72, 74) may derive from sacrifice associated with the temple; bones and an animal skeleton uncovered in Papathanasopoulos’ Trench 11 represent additional evidence for animal sacrifice. Lakonian tiles replaced the earlier Corinthian-style roof sometime during the early to mid-6th century bc; it is not clear how long the temple remained in use after that replacement of the roof.

Blegen, C. W. n.d. Unpublished manuscript for The Palace of Nestor of Pylos. ASCSA, Athens, and University of Cincinnati, OH.

Bibliography Ainian, A. M. 1997. From Rulers’ Dwelling to Temples. Architecture, Religion and Society in Early Iron Age Greece (1100– 700 bc). Studies in Mediterranean Archaeology 121. Paul Åströms Förlag, Göteborg. Antonaccio, C. M. 1995. An Archaeology of Ancestors. Tomb Cult and Hero Cult in Early Greece. Routledge, London.

Blegen, C. W. 1953. ‘The Palace of Nestor Excavations at Pylos, 1952’, AJA 57, 59–64. Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I. The Buildings and their Contents. Princeton University Press, Princeton, NJ. Blegen, C. W., Rawson, M., Taylour, W., Donovan, W. P. 1973. The Palace of Nestor at Pylos in Western Messenia, III. Acropolis and Lower Town, Tholoi, Grave Circle, and Chamber Tombs, Discoveries Outside the Citadel. Princeton University Press, Princeton, NJ. Broneer, O. 1971. Isthmia I: Temple of Poseidon. ASCSA, Princeton, NJ. Cooper, F. A. 1994. Unpublished ephoreia report.

Architectural activity resumed on the hilltop sometime later, with the construction of temple 83/82/81, erected over the southwest end of the earlier Archaic temple. The final phase of building activity is marked by the fragments of medieval tile recovered from Blegen’s deposits, which indicate new construction at the site sometime during the late 12th or 13th centuries ad.

Cooper, F. A. 1996. The Temple of Apollo Bassitas, I. The Architecture. ASCSA, Princeton, NJ. Cooper, F. A. 2002. Houses of the Morea / Σπιτια του Μορεα, Melissa Pubishing, Athens. Cooper, F. A., and Morris, S. 1990. ‘Dining in Round Buildings’, in Sympotica: A Symposium on the Symposion, ed. O. Murray, 66–85. Clarendon Press, Oxford.

The picture that this material presents is one of sporadic activity, not dissimilar to that on other Mycenaean palatial sites. The development of cult and religious architecture that characterises the palaces at Athens, Mycenae, Tiryns and Sparta is also present at Englianos. The site differs from those centres, however, in the lack of historical documents to describe the significance and nature of the religious architecture identified in the Northwest Area. The peculiar historical situation that defined Messenia after the Spartan conquest has in many ways obscured

Cooper, N. K. 1983. The Development of Roof Revetment in the Peloponnese. Studies in Mediterranean Archaeology 88. Paul Åströms Förlag, Göteborg. Cooper, N. K. 1990. ‘Archaic Architectural Terracottas from Halieis and Bassai’, Hesperia 59, 65–93.

254

Bibliography Cooper, N. K. 1995. ‘The Archaic Temple of Apollo at Bassai: Correspondences to the Classical Temple’, Hesperia 64, 227–77.

Papathanasopoulos, G. 1958. Unpublished Pylos excavation notebooks. ASCSA, Athens, and University of Cincinnati, OH.

Coulson, W. D. E. 1986. The Dark Age Pottery of Messenia. Studies in Mediterranean Archaeology 43. Paul Åströms Förlag, Göteborg.

Popham, M. 1991. ‘Pylos: Reflections on the Date of its Destruction and on its Iron Age Reoccupation’, Oxford Journal of Archaeology 10, 315–24.

Davis, J. D., and Stocker, S. R. 2013. ‘The Medieval Deposit from the Northeast Gateway at the Palace of Nestor’, Hesperia 82, 673–731.

Rawson, M. 1958. Unpublished Pylos excavation notebook: ‘Pylos 1958: Areas MZ–MY, vol. 1’. ASCSA, Athens, and University of Cincinnati, OH.

Drerup, H. 1969. Griechische Baukunst in geometrischer Zeit. Archaeologia Homerica, Band II, Kapitel O. Vandenhoeck & Ruprecht, Göttingen.

Rhodes, R. F. 1984. The Beginning of Monumental Architecture in the Corinthia. Unpublished PhD dissertation. University of North Carolina, Chapel Hill.

Griebel, C. G., and Nelson, M. C. 1998. ‘The Ano Englianos Hilltop after the Palace’, in Sandy Pylos. An Archaeological History from Nestor to Navarino, ed. J. L. Davis, 97–100. University of Texas Press, Austin.

Rhodes, R. F. 1987. ‘Early Corinthian Architecture and the Origins of the Doric Order,’ AJA 91, 477–480. Robinson, H. S. 1976. ‘Excavations at Corinth. Temple Hill, 1968–1972’, Hesperia 45, 203–39.

Hägg, R. ‘Funerary Meals in the Geometric Necropolis at Asine?’, in The Greek Renaissance of the Eighth Century bc. Tradition and Innovation, ed. R. Hägg, 27–31. Skrifter Utgivna av Svenska Institutet i Athen, Stockholm.

Roebuck, M. C. 1955. ‘Excavation at Corinth: 1954’, Hesperia 24, 147–57. Skoog, V. N. 1998. The Laconian-style Roof. Development, Distribution and Technology. Unpublished PhD dissertation. University of Missouri, Columbia.

Hemans, F. P. 1989. ‘The Archaic Roof Tiles at Isthmia. A Reexamination’, Hesperia 58, 251–66.

Stein, J. and Rapp, G. 1978. ‘Archaeological Geology of Site’, in Excavation at Nichoria in Southwest Greece, I: Site, Environs and Techniques, ed. G. Rapp and S. E. Aschenbrenner, 234–57. University of Minnesota Press, Minneapolis.

Hope, R. and McDonald, W. A. 1953. Unpublished Pylos excavation notebook. ASCSA, Athens, and University of Cincinnati, OH. Kittredge, W. 1962. Unpublished Pylos excavation notebook. ASCSA, Athens, and University of Cincinnati, OH.

Stocker, S. R. and Davis, J. L. 2004. ‘Animal Sacrifice, Archives and Feasting at the Palace of Nestor’, Hesperia 73, 179–95.

Klein, N. 1997. ‘Excavations of the Greek Temples at Mycenae by the British School at Athens’, Annual of the BSA 92, 247–332.

Themelis, P. 1994. ‘Hellenistic Architectural Terracottas from Messene’, in Proceedings of the International Conference on Greek Architectural Terracottas of the Classical and Hellenistic Periods, 12–15 December 1991, 141–69. Athens.

Lambrinoudakis, V. K. 1988. ‘Veneration of Ancestors in Geometric Naxos’, in Early Greek Cult Practice, ed. R. Hägg, N. Marinatos and G. C. Nordquist. Svenska Institutet i Athen, Stockholm.

Wikander, O. 1988. ‘Ancient Roof-tiles. Use and Function’, Opuscula Atheniensia 17, 203–16.

Le Roy, C. 1967. Les terres cuites architecturales. De Boccard, Paris.

Wikander, O. 1990. ‘Archaic Roof Tiles. The First Generation’, Hesperia 59, 285–90.

Mallwitz, A. 1959–60. ‘Die Ausgrabung beim Athena-Tempel in Milet 1957’, Istanbuler Mitteilungen 9–10, 67–85.

Williams, C. K., Snyder, L. M., Barnes, E. and Zervos, O. 1998. ‘Frankish Corinth: 1997 (Excavation Results)’, Hesperia 67, 223– 81.

Mallwitz, A. 1981. ‘Kritisches zur Architektur Griechenlands im 8. und 7. Jahrhundert’, Archäologischer Anzeiger, 599–642.

Williams, C. K. and Zervos, O. 1992. ‘Frankish Corinth’, Hesperia 61, 133–91.

Mallwitz, A. 1983. ‘Osservazioni sull’Architettura nella Grecia dei Secoli VIII e VII a.c.’, ASAtene 59, 1–16.

Winter, N. A. 1994. Greek Architectural Terracottas from the Prehistoric to the End of the Archaic Period. Clarendon Press, Oxford.

Mountjoy, P. A. 1997. ‘The Destruction of the Palace at Pylos Reconsidered’, Annual of the BSA 29, 109–37.

255

Brenningmeyer – Post-Bronze Age Architecture and Stratigraphy Yassoglou, N. J. and Haidouti, C. F. 1978. ‘Soil Formation’, in Excavation at Nichoria in Southwest Greece, I: Site, Environs and Techniques, ed. G. Rapp and S. E. Aschenbrenner, 234–57. University of Minnesota Press, Minneapolis.

Revised 2016

256

9 POST-BRONZE AGE POTTERY SHAWN A. ROSS

Carl Blegen and Marion Rawson offer a detailed and thorough treatment of the palatial (Late Helladic) material from Englianos in The Palace of Nestor (PN) publication of 1966. They also deal extensively with the Middle Helladic pottery from the site. Pottery analysis conducted by the Minnesota Archeological Researches in the Western Peloponese (MARWP) therefore focused on the postpalatial ceramics. MARWP identified several types of post-Bronze Age pottery in Blegen’s backfi ll, which he did not describe, or described only briefly, in The Palace of Nestor. Dark Age, Archaic, Classical, Late Roman and Medieval ceramics all attest to continued habitation at the site after end of the Bronze Age.

Protogeometric sherds near the Palace of Nestor, but concluded that it was ‘impossible on our evidence to show continuity of occupation … from the Bronze Age to the Geometric period’ (Davis, Alcock et al., 1997, 452). The MARWP excavations have changed this picture. Employing comparanda from Nichoria and elsewhere, a significant body of Dark Age pottery dating from the 11th through the 8th centuries bc has been identified, erasing most of the period of abandonment and securing a place for Englianos in the roster of Messenian sites occupied during this period. Study of the pottery contained in backfi ll from Blegen’s excavations became a significant part of the MARWP Pylos Project following the discovery in 1994 of Blegen’s pottery dump in the Northwest Area of the site. Material from this dump and elsewhere was recovered and sorted, and diagnostic, non-Palace period pottery was catalogued and photographed (see pp.82–83 and table 1.4 above). Since the project was strictly limited to a re-excavation of Blegen’s earlier work at the site, however, MARWP dealt only with material from his backfi ll. All material was out of context and generally limited to individual sherds – often small and poorly preserved. The nature of pottery analysis for this project more closely resembles that of an archaeological surface survey than a traditional excavation, with our main task being to identify and date individual sherds based upon their intrinsic characteristics rather than from context.

The most significant material identified dates to the Greek Dark Age. Blegen usually assigns pottery that post-dates the palace’s Late Bronze Age destruction to the ‘late Geometric’ period; in one instance he specifically refers to the late 7th century bc (PN I, 91; cf. McDonald and Coulson 1983, 319–21, and Popham 1991, esp. 316–21). Only once, when discussing material from outside the palace, does he identify a Dark Age presence at the site: in reference to the ‘Protogeometric tholos’, dated by V. Desborough to the 10th century bc (PN III, 237–42). Since the publication of the final volume of The Palace of Nestor in 1973, however, excavations at sites such as Nichoria and Lefkandi have revolutionised understanding of material culture during the Dark Age, allowing re-examination of ceramics from Englianos. Mervin Popham (1991, 316–22), in perhaps the most significant revision to date, re-assesses some of the Late Helladic IIIC and ‘Geometric’ pottery recorded by Blegen, arguing for re-occupation of the site of the palace shortly after its destruction, with continuing habitation for perhaps another century. Popham’s re-assessment shortens the gap between the abandonment of the palace and the 8th-century bc activity noted by Blegen, but still leaves a hiatus of some 300 years (c.1100–800 bc). The Pylos Regional Archaeological Project (PRAP) recovered, tantalisingly, a number of Submycenaean through

Furthermore, the sheer volume of ceramic material provided its own challenges. Blegen’s backfi ll contained many tons of ceramics, allowing only a hurried search for non-Palace period pottery; most of the ceramics, by necessity, were soon reburied. As a result, the sherds catalogued in no way represent an exhaustive inventory of post-Palatial pottery from the site of the Palace of Nestor, only a somewhat haphazard sampling 257

Ross – Post-Bronze Age Pottery of what is certainly a more varied and extensive deposit of material.

in PN III, 6, 26–27, 61–63, 237–42; cf. McDonald, Coulson and Rosser 1983, 76, 101, 110, 319–21; Coulson 1986, 67–68; Popham 1991, 316–21; Griebel and Nelson 1998, 97, 99–100). Such observations culminate with an admission that ‘it is thus evident that there was fairly widespread activity on the site in Geometric times’ (PN I, 294). On the other hand, in the conclusion to the publication, Blegen not only claims that ‘no evidence whatever of a reoccupation in Mycenaean times after the catastrophe has been observed’, but states definitively that the catastrophe ‘was the end of human occupation of the site’ (ibid., 422, 424).

Only the classes of pottery identified with some degree of certainty, using convincing comparanda, are discussed in this chapter. Some types have been omitted due to ambiguity in their identification and dating, including, for example, what appear to be Roman and medieval plain wares. In general, the lack of context, small size and poor condition of the vast majority of sherds examined meant that in most cases only decorated pottery could be identified with any certainty. Such a conservative approach to the material yielded a relatively small number of sherds that could be securely identified. These include approximately 100 Nichoria-type Dark Age sherds, a few dozen Dark Age sherds with parallels other than Nichoria, a few dozen Archaic or Classical black-slipped sherds, perhaps a dozen red-slipped sherds (Hellenistic or terra sigillata), a few dozen distinctive late Roman sherds, and several dozen medieval (Byzantine or Frankish) sherds. Due to the lack of a large and systematically collected sample, quantitative analysis of the pottery is avoided; an attempt was made to identify and date only a sampling of representative post-Bronze Age ceramics from Englianos, which, however, attests to the ebb and flow of occupation at the site.

Reconsideration of the situation began with two of Blegen’s own collaborators, McDonald and Coulson, who observed that ‘a few miscellaneous vases from Englianos are similar [to Nichoria DAIII ceramics], but they occur in undated contexts’ (McDonald, Coulson and Rosser 1983, 319; see also 76, 101, 110, 320–21; cf. Coulson 1986, 11, 67–68). Popham carries the revision further. Drawing on his knowledge of the material from Lefkandi, he analyses the nature and extent of the ‘Geometric Ware’ noted in The Palace of Nestor and re-dates some Late Helladic pottery (Popham 1991, 316–21). He ultimately argues for a Late Helladic IIIC re-occupation of Englianos, perhaps lasting into the early Dark Age, but cautions that this dating requires further investigation (cf. Davis, Alcock, et al., 1997, 451–52). As noted above, a few sherds retrieved in the vicinity of the palace by the PRAP archaeological surface survey also suggest Dark Age activity, but evidence was too meager and scattered for PRAP to draw any firm conclusions (Davis, Alcock et al., 1997, 452).

Comparative material from a variety of sites in the Peloponnese and beyond has been of the greatest importance in this task. Ceramics from Nichoria for the Dark Age and medieval periods proved invaluable, both because the sites lie in the same topographical region of Messenia (southwest of the Pamisos River) and because the importance of Nichoria in establishing a Dark Age (and, to a lesser extent, medieval) pottery typology and chronology for Messenia cannot be overstated (see Coulson 1986, 9–11, esp. 11, where he notes how closely Dark Age pottery from across Messenia follows the sequence at Nichoria). Coulson’s The Dark Age Pottery of Messenia (1986) extended Messenian parallels beyond Nichoria. Published material from Pylos in Elis, Sparta, Tegea, Corinth, Isthmia, Lefkandi and the Athenian Agora, as well as unpublished pottery from museum collections in Athens and Istanbul, supplemented that from Messenia and proved particularly valuable where no Messenian comparanda exist.

Although our material is out of context, it considerably expands the body of evidence for the Dark Age from Englianos beyond the ceramics catalogued in The Palace of Nestor or the Geometric vases mentioned in Excavations at Nichoria. The MARWP excavations uncovered a broad range of sherds dating from the 11th century bc through the 8th. As noted above, Nichoria and other sites in Messenia provide the most useful comparative material from this period, especially considering that the relative isolation and localism of the Dark Age increases the value of nearby sites. Dark Age pottery found at Englianos is analogous to that from Nichoria and other Messenian sites in type and size of vessels and quality and thickness of fabric, as well as in paint style, colour and quality of decoration. Material from Nichoria provided the basis for identifying and dating most Dark Age material recovered by MARWP, supplemented by parallels from other Messenian sites and comparanda from further afield, particularly from Corinth (McDonald, Coulson and Rosser 1983, 61–259, esp. 110–11; Weinberg 1943).

Dark Age Wares (Nichoria I–III) In The Palace of Nestor, Blegen seems to be of two minds concerning post-Bronze Age habitation at Englianos. On the one hand, he makes repeated references to ‘Geometric’ or ‘Late Geometric’ wares – mostly from a deposit of ‘black oily earth’ that covered a significant portion of the hilltop (see p.227, fig. 8.1 above) – which he dates variously to the late 8th century bc through the 5th (PN I, 64–65, 91, 177, 180–81, 184–85, 203, 209, 229, 291, 294–95, 296–98, 300– 301, 303, 307, 329; other ‘Geometric’ pottery is discussed

dark age i (c.1075–975) Nichoria-type DAI material found at Englianos is identified by its manufacture from dirty white or pale brown clay with a soft and crumbly texture. Prominent wheel marks on the interior (and sometimes the exterior)

258

Dark Age Wares (Nichoria I–III) are common. Open shapes predominate, particularly skyphoi and deep bowls. Fabric thickness is typically at least 4–5 mm, and as thick as 8 mm. Rim profi les tend to be simple: straight to slightly everted. The interior of open shapes is usually painted monochrome, whereas the exterior may be monochrome or include simple decoration, such as a single wavy line. Paint tends to be badly worn and streaky or washy (i.e., transparent, light and worn, especially from the ridges of fine wheel marks), more so than in later periods. Paint colours include reddishbrown (more common than in later periods), brown or black. As at Nichoria, these sherds tend to be poorly preserved and difficult to identify. Our identification of pottery representing the DAI period is tentative, based primarily on the colour and quality of the paint, the presence of prominent wheel-riding, and the simple rim profi les (McDonald, Coulson and Rosser 1983, 19, 61–72, esp. 63–68 (deep bowls); see also 110; cf. Coulson 1986, 13). Identification is, however, reinforced by comparisons with pre-Geometric skyphoi from Corinth dated to the second half of the 11th century bc. These skyphoi resemble our DAI sherds, displaying light-coloured clay and a fine redbrown glossy slip that wears poorly (Weinberg 1943, 3–8, plate 1, esp. nos. 1–3).

found at Rizes, Malthi, Antheia, Kaphirio and elsewhere (see Coulson 1986, Chapter 3, esp. 28, 56). Beyond Messenia, comparanda for DAII pottery also appear at Pylos in Elis – an important site for comparative material, reflecting the emergence of a west Greek ceramic koiné during DAI and DAII (cf. Coulson 1986, 18; 55–56; 71). Further east, early Geometric pottery from Corinth (late 10th–late 9th century bc), includes numerous examples that correspond to Nichoria-type DAII material from Englianos. Corinthian fabrics and surface decoration are similar; light-coloured fabrics and red-brown, brown or black decoration predominate (as at Nichoria, red-brown paint gives way to black over time, and often the paint wears badly). Skyphoi of comparable size, decoration (monochrome or simple), handle type (loop) and rim profi le also occur (Weinberg 1943, 9–24, esp. 21–22, and plates 2–11, esp. nos. 23, 29, 37, 38–45, 48, 59–62, 72). Somewhat later Geometric material (equivalent to Nichoria DAII/ III transitional) also finds parallels at Corinth. Skyphoi from the North Cemetery dated by the excavators to the second half of the 9th century bc are similar in diameter, fabric and decoration. The excavators describe them as constructed of ‘buff Corinthian clay, slightly pinkish; black glaze, in places thin and streaky, and somewhat metallic’, a description that corresponds well with several of our DAII or III sherds (Blegen, Palmer and Young 1964, 21–22, 25, 40–41, and plates 6–7, esp. nos. 14–2, 15–2, 17–4; cf. Stillwell and Benson 1984, 16, nos. 10, 12–14, and plates 1–2; see also Coleman 1986, 18–30, esp. 19–23 (kraters) and 26–27 (kantharoi)). Overall, considering comparanda from Nichoria, other sites in Messenia, Pylos in Elis, and Corinth, our identification of DAII pottery at Englianos is more secure than that of DAI.

dark age ii (c.975–850, including dark age ii/iii transitional, c.850–800) Like DAI material, Nichoria-type DAII pieces from Englianos tend to display a chalky, fine fabric; fabric colour is typically light reddish-yellow or, less often, tan or pale brown. Wheel marks remain common but are less prominent than those found in DAI. Shapes are similar to DAI, but a greater variety of rim forms is attested, while body thickness tends to be thinner and vessels smaller. Solid or streaky black paint, thicker than DAI, predominates. Paint colour varies from a black to dark brown, and may be variegated, but it is never solid brown as in DAI, while reddish-brown paint is less common. On some pieces it appears that the paint has worn away, leaving a reddish-brown stain on the pale fabric. Again, all sherds from open shapes have monochrome interiors. Exteriors may be monochrome or display simple decoration, with bands and parallel lines being the most common.

dark age iii (c.800–750) The fabric of DAIII pottery remains soft and crumbly; fabric colour is pale, usually light reddish-yellow or creamcoloured, although pale brown, light yellowish-brown and gray also occur. Body thickness of the open shapes drops another 1 mm compared to DAII sherds, to 2–3 mm, and the average size of vessels continues to decrease, while rim forms continue to grow more varied. Most vases are simply monochrome coated on both exterior and interior. Although dark brown paint still appears, black becomes standard; it is further distinguished by its thickness, bluish tinge and metallic sheen, although it remains streaky (McDonald, Coulson and Rosser 1983, 96–109; cf. Coulson 1986, 66). DAIII is the best attested and most securely identified of the DA pottery periods at Englianos.

DAII/III transitional pottery from Nichoria, although catalogued separately by the excavators, is very similar to Nichoria DAII material. Since the former is identifiable primarily through ‘stratigraphical and architectural association’, it is grouped here with DAII material; no attempt has been made to differentiate a separate DAII/ III transitional phase of pottery at Englianos (McDonald, Coulson and Rosser 1983, 72–90, esp. 90 (DAII); 90–96 (DAII/III Transitional); cf. Coulson 1986, 10, 55; Davis, Alcock et al., 1997, 452).

dark age summary Determination of the specific period (DAI, II or III) presents some problems, especially considering that local stylistic variation may easily obscure the chronological details of stylistic evolution (cf. Coulson 1986, 78). Furthermore, poorly preserved Archaic and Classical sherds may also

DAII was a period of extensive and relatively homogenous ceramic production across Messenia; in addition to Nichoria, comparanda for the Englianos material can be

259

Ross – Post-Bronze Age Pottery feature chalky fabric with streaky red-brown or faded black paint (A. Berlin, personal communication). Based on comparison with ceramics from Nichoria, however, it appears that Dark Age pottery does occur at Englianos, and that this material spans a considerable length of time. At least two and perhaps all three Nichoria periods are represented, with the most uncertainty concerning DAI. Pottery from Englianos reflects both the homogeneity of Dark Age Messenian pottery and the larger West Greek ceramic koiné described by Coulson (1986, 18, 55, 66; Chapter 5, esp. 71, 77–78).

the mid-7th century bc (McDonald, Coulson and Rosser 1983, 318–22).

Englianos appears to have been reoccupied beginning perhaps as early as the 11th century bc, more certainly in the 10th century bc. The ceramic evidence further suggests that habitation at Englianos then paralleled that at Nichoria until the destruction of the latter and, as will be argued below, probably continued thereafter. If Popham is correct in his assessment that Englianos was occupied in LH IIIC and the early Dark Age, any gap in occupation at the site after the destruction of the palace may have been very short indeed.

Blegen reports finding a ‘stray Hellenistic or Classical Black Glazed sherd’ in the Northeastern Gateway (PN III, 6). The MARWP excavations also encountered much plain, black-slipped pottery dating to the Classical and/ or Hellenistic periods. Most of these sherds are small and non-diagnostic and were identified solely on the basis of their surface treatment. Typically, these black-slipped pieces are of a fine ware constructed from an evenlyfired pinkish-buff to reddish-brown fabric with a glossy black slip readily distinguished from the paint found on other material from the site. Most sherds display a slip on both the interior and the exterior; slip colour ranges from a dark purplish-black to a somewhat lighter but still dark grayish-black. The sherds, for the most part, appear to come from small vessels, including perhaps a cup or kantharos with the stump of a thin, vertical strap handle (1A05915; a strap handle fragment, 1A04519, is also preserved). Few rim sherds were recovered. These pieces resemble Classical pottery dating to 5th and 4th centuries bc from the Tholos at Nichoria – although precise dating is difficult, as pottery with a high-quality black gloss was produced from c.600 to 100 bc (McDonald, Coulson and Rosser 1983, 334–35; see also Sparkes 1991, 103–04, and 1970, 1, note 2).

Late Archaic, Classical or Hellenistic Black-slipped Wares Although Dark Age pottery from Englianos marks perhaps the most noteworthy addition to our knowledge of settlement at the site, later ceramics also warrant discussion.

Late Dark Age through Early Archaic Wares Although Blegen and the excavators working with him identified no Dark Age material, he made several references in the final publication to ‘Geometric’ or ‘Late Geometric’ ware from a deposit of ‘black oily earth’ that covered a significant portion of the hilltop (PN I, 64–65, 91, 177, 180– 81, 184–85, 203, 209, 229, 291, 294–95, 296–98, 300–301, 303, 307, 329; other ‘Geometric’ pottery is discussed in PN III, 6, 26–27, 61–63, 237–42). The MARWP excavations recovered a number of sherds that appear to be of one type briefly remarked on by Blegen. The pottery Blegen describes as Geometric, and which he dates variously to the late 8th through the 5th centuries bc, may instead belong to the late Dark Age or early Archaic period and date to the 9th– 6th centuries bc.

Hellenistic or Roman Red-slipped Wares A very few pieces of Hellenistic red-slipped ware or terra sigillata were found. These are constructed from a thin fabric exhibiting noticeable wheel marks. The red slip is lustrous but tends to wear poorly. These sherds are small and badly preserved, making their identification difficult. They could date to the Hellenistic period but no earlier, as solid red slip is not introduced until then (Sparkes 1991, 26). One piece in particular, 1A04800, has a thicker, better-wearing slip that is also somewhat darker in colour than the others, and is more likely to be terra sigillata. Identification of Hellenistic red-slipped ware and terra sigillata rests on ceramics viewed by the author in the Agora Museum, Athens, and the Aphrodisias Museum, Aphrodisias, Turkey.1 Red-slipped pottery of any type

These sherds are generally of a higher quality with respect to fineness of fabric, firing and decoration. One example is 1A03949, made from a hard, tan clay. This piece is particularly noteworthy for its neat, compassdrawn concentric circles and, based on comparanda from Lakonia, Arcadia and elsewhere, seems to date from the late 8th century. Most of the sherds of this type appear to be of a similar date (8th–7th centuries bc), although sherds such as 1A05876, a large section of a black-coated amphora neck with a simple profi le, may be as early as the 9th century bc. One piece, 1A01911, appears to be later (early 6th century bc), based on comparanda from Corinth and Elis. Although the earliest pottery considered in this section may overlap chronologically with Nichoria-type sherds (perhaps suggesting contact with other sites in the Peloponnese or beyond during the Dark Age), most appear to be later, suggesting that Englianos continued to be inhabited after the abandonment of Nichoria in

1

Due to the limited number, small size and poor preservation of the sherds of this type recovered, I chose to rely for the identification of these sherds primarily on comparanda that I personally inspected rather than on photographs from publications.

260

Medieval Pottery is poorly represented compared to the other classes of ceramics recovered by MARWP: only a handful of such sherds were found.

tions of more widely distributed pottery types. Ceramics viewed by the author at the Byzantine Museum, the Agora Museum and the Monastiraki Mosque, all in Athens, and in the Istanbul Museum, also contributed to identification and dating of the medieval pottery. Significantly, the period of manufacture of medieval ceramics from Englianos appears to reflect an era of revived ceramic production at Nichoria and other sites in the eastern Mediterranean in the late 10th through 13th centuries ad (McDonald, Coulson and Rosser 1983, 379–83, esp. Area IV Glazed Wares, Area IV Sgraffito Wares and Area IV Painted Wares).

Late Roman Plain Ridged Ware A distinctive class of pottery, marked by narrow parallel ridges around the body of the vase (0.5–2.0 mm wide, 0.5 mm high, and separated by 0.5–1.0 mm wide indentations), closely resembles Late Roman pottery from the Athenian Agora. The fabric of this pottery is sandy (sometimes micaceous), and varies in colour from light brown to dark brown, red-brown or gray. No decoration other than the ridging survives, although this may be a result of the sandy nature of the clay and the sherds’ poor state of preservation. Vessel walls are thin (3–5 mm), but the diameter of the vessels appears to be rather large (120–260 mm), probably representing bowls, jugs or cooking vessels, although diameter estimation proved difficult due to the small size of the sherds. Despite their poor preservation, the prominent ridging makes identification relatively secure. Comparanda for this class of pottery comes from the Athenian Agora, where numerous ridged vessels of various types dated to the 2nd–6th centuries ad occur, with the closest parallels among the MARWP material dating to the 3rd–5th centuries ad (Robinson 1959; see groups G, J, K, L and M (2nd–6th centuries), esp. groups K, L and M (3rd–5th centuries)).2 That several dozen sherds, varying considerably in colour, thickness and ridging, were found may indicate a larger and more diverse deposit of Roman material, at least in the later Roman Empire, than that indicated by the handful of (possible) terra sigillata sherds.

glazed, decorated bases The Pylos Project identified two glazed and decorated bases similar to those found among the Painted Wares from Nichoria. Both are constructed of fine, soft, chalky, pink fabric, comparable to material from Nichoria and elsewhere. The first (1A04797) displays a flaky white interior slip with a clear covering glaze and a curvilinear design in thick green paint, which bleeds into the covering glaze. The exterior of this base has a trace of yellow glaze. Such white glazes or slips and green paint are common on Byzantine ceramics from Nichoria, and similar pieces are on display at the Agora Museum in Athens and the Istanbul Museum, where they are dated to the 10th–13th centuries (for Nichoria, see McDonald, Coulson and Rosser 1983, 381–83; comparanda from Athens and Istanbul viewed by author). Similar green and brown painted ware from Corinth also dates to the 10th–13th centuries, with later material from the 12th and 13th centuries offering the closest analogies (Morgan 1942, plates XIX–XXIII, esp. nos. 440, 447, 453, 476). ‘Glazed painted ware’ from Sparta, dated by the excavators to the 13th century, offers another body of comparative material (Sanders, et. al. 1993, 258–60). The second base (1AO4801) has a worn, mottled/crackled brown interior glaze, with a very dark brown line and curve pattern. The exterior is bare, which could be a result of the sherd’s poor state of preservation. One instance of a painted brown oval on the interior surface of a base is found at Nichoria, although brown or dark brown paint occurs more frequently (McDonald, Coulson and Rosser 1983, 381–83). Similar base fragments from the Istanbul Museum are dated from the 10th through the 14th centuries.

Medieval Pottery In Volume III of The Palace of Nestor, Blegen and his collaborators very briefly report finding ‘glazed ware of medieval times’, in what he called an ‘extensive intrusion of the eleventh and twelft h centuries ad’ (PN III, 6, 8, fig. 121 nos. 1–7 (Plain Ribbed Ware) and 8–18 (Glazed Ware)). A handful of these sherds are on display in the Chora Museum. During MARWP’s analysis of the pottery, however, we discovered a wider array of glazed ware than that mentioned by Blegen, as well as examples of ‘spongy ware’ similar to examples unearthed at Nichoria and Sparta. As was the case with Dark Age pottery, medieval pottery from Nichoria and Englianos share many features, and Nichoria provided comparanda of first resort for the identification of medieval pottery recovered in the MARWP excavations. Unfortunately, as at Nichoria, the quality and preservation of most of the medieval pottery from Englianos is poor (Rosser 1983, 381). Medieval pottery from Englianos, like that from Nichoria, consists of local varia-

yellow glazed ware The fabric for a group of yellow-glazed sherds tends to be harder, pinker and darker than the bases discussed above, and is consistently dark pink to reddish-yellow in colour. Wheel marks or evenly spaced ridges are often visible on both interior and exterior. Similar to the yellow-brown glazed ware, glaze on this material is thin, crackled or flaky, and generally more worn. One or both sides may be coated with yellow-brown or mustard-coloured glaze, or one side may be coated with a thicker, rougher, offwhite slip (again similar to that found on the yellow-

2

My thanks to A. Berlin for assisting with the identification of this class of pottery.

261

Ross – Post-Bronze Age Pottery brown glazed ware). Nichoria yielded pale yellow, yellow, or yellowish-brown glaze pieces, sometimes described as mottled or crackled, dating to the 12th or 13th century (McDonald, Coulson and Rosser 1983, 381–82, nos. P1708, 1709, 1717). Comparable examples of yellow-glazed pottery include a one-handled water jug with a crackled and flecked but otherwise very even mustard-yellow glaze from Pylos in Messenia (Monasteraki Mosque, Athens, Ceramic Collection no. MEΛΤ 15403), and a vessel glazed in yellow with an iridescent sheen from Athens (Byzantine Museum, Athens, no. 159). Another example in the Istanbul Museum dates to the 11th–13th centuries ad. Vessels with a paler glaze, which are most similar to pieces recovered by the MARWP excavations, consistently fall later in this range (see also Sanders, et. al. 1993, 266–67, nos. 27–28, dated to the late 12th century).

suggests that Englianos was in contact with settlements beyond Nichoria during the later Dark Age at least. Archaic pottery (Late Geometric, proto-Attic, protoCorinthian and Corinthian), dating to the late 8th through 6th centuries bc, argues for continued activity on the hill following the end of the Dark Age, as does black-slipped pottery dating to the 6th century bc or later. The presence of Archaic pottery marks a divergence in settlement pattern from Nichoria, which was abandoned after 750 bc. A single handmade late Classical or Hellenistic lamp and a few sherds of Hellenistic red slip ware or terra sigillata vessels may not indicate an extensive level of activity at the site during the Classical, Hellenistic or Roman eras, but at a minimum they provide evidence that the area was not entirely unvisited. A more diverse collection of 3rd– 5th century ad Roman plain ridged ware sherds indicate renewed or continued activity at the site.

byzantine ‘spongy’ ware Following the Late Antique period the site produces no identifiable ceramic material until late in the Byzantine era. The medieval pottery most closely resembles the 11th–13th century ad material found in Area IV at Nichoria, dates confirmed by comparisons from further afield. The presence of Byzantine material indicates a reoccupation after a long hiatus, a pattern common at other contemporary sites in the eastern Mediterranean. Thus the pattern of habitation reflected in the post-Bronze Age pottery from the Palace of Nestor is similar, but not identical, to that at nearby Nichoria: after the destruction of the Bronze Age palace, the site was occupied, perhaps continuously or with a relatively brief hiatus, from LH IIIC, through the Dark Age and well into the Archaic period; thereafter it witnessed intermittent activity, particularly during the Classical, Late Roman and medieval eras.

The characteristic feature of Byzantine spongy ware is ‘pinholing’ over the surface of the fabric. This fabric is immediately recognisable, making it one of the few coarse wares the Pylos Project attempted to identify and date. This piece (1A04589), typical of our finds and similar in most respects to spongy ware from Nichoria, is constructed from a medium fabric tempered with coarse sand, displaying an uneven exterior surface colour and often a blackened interior, and of course having distinctive pinholes in the fabric (Rosser 1983, 378–97, esp. 379–80 and plate 10-4). Spongy ware has also been found in Lakonia at Ayios Stephanos and Sparta, and on the island of Kythera. All examples of spongy ware date from the 11th–14th centuries (see Sanders, et. al. 1993, 255 (Fabric 7), 279, nos. 62–73; Coldstream and Huxley 1972, 176, nos. 16–18, plate 50).

Catalogue3 Conclusions Dark Age (Nichoria I–III) It is hoped that the MARWP Pylos Project has at least partially fulfi lled a desire expressed in 1992 that ‘a chronology for [the] post-palatial settlement at Pylos may be clarified by future restudy of retained pottery’ (Griebel and Nelson 2001). If Popham is correct in his reassessment of the LH IIIC ceramics, habitation resumed shortly after the destruction of the Bronze Age Palace in the 13th century, and continued until c.1100 bc. Any subsequent abandonment of the site may therefore have been very brief indeed, if it occurred at all. The pottery recovered by MARWP from Blegen’s backfi ll indicates reoccupation no later than Nichoria DAII (975–800 bc), and perhaps as early as DAI (1075–975 bc). Thereafter, habitation continued through DAIII (800–750 bc), during which Englianos appears to have avoided a trend of declining ceramic production seen at other sites in Messenia. DAII and DAIII habitation at the site helps to explain the presence of nearby burials (Coulson 1986, 7475). Moreover, the existence of Geometric pottery with parallels from Corinth, Lefkandi, Athens and elsewhere

Dark Age I (c.1075–975 bc) 1A03961. Rim sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.036. Th. 0.006. Rim D. (est.) 0.100. Evenly-fired, chalky light pink-tan fabric (Pantone 53-5). Monochrome streaky, red-brown paint (Pantone S812; badly worn) interior and exterior. Prominent wheel ridges, especially interior. (Surface pattern appears to be the result of uneven paint deterioration, not intentional). 1A04811. Rim sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.027. Th. 0.004. Rim D. (est.) 0.180. Evenly fired, fine, chalky light tan fabric (Pantone S57-4) with reddish-brown streaky paint interior (Pantone S3172) and red to black streaky paint exterior (predominantly Pantone S59-3; both interior and exterior badly worn). Noticeable wheel marks. Simple rim shape.

262

Catalogue Dark Age II (c.975–850 bc) 1A04084. Rim sherd; open shape (skyphos or deep bowl). H. (pres.) 0.051. Th. 0.005. Rim D. (est.) 0.160. Soft, gray fabric (Pantone of fabric colour not recorded). Monochrome coating in streaky black paint interior (Pantone S325-2); reddish-brown, variegated, streaky paint exterior (predominantly Pantone S36-1). 1A04557. Rim sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.030. Th. 0.004. Rim D. (est.) 0.100. Evenly-fired light tan fabric (Pantone S58-6). Monochrome coating in reddish-brown, variegated, streaky paint interior and exterior (predominantly Pantone S316-2 to S317-1). 1A04602. Body sherd; open shape. H. (pres.) 0.042. Th. 0.005. D. (est. from max. pres.) 0.160. Evenly-fired, soft pinkish-tan fabric (Pantone S67-6). Cream-coloured slip exterior (Pantone 23-6). Horizontal band and edge of an arc (spiral or circle?) in reddish-brown, variegated, streaky paint on exterior (predominantly Pantone S68-2). Traces of monochrome coating in similar reddish-brown streaky paint on interior. Noticeable wheel marks. Cf. Nichoria nos. P607, P781, P845.

9.1 Dark Age I sherds.

1A05850. Body sherd; open shape (?). H. (pres.) 0.050. Th. 0.007. D. (est. from max. pres.) 0.200. Flaky, tan fabric (Pantone S49-5). Monochrome dark brown paint interior and exterior. Prominent wheel ridges. 1A05874. Body sherd; open shape (?). H. (pres.) 0.037. Th. 0.007. D. (est. from max. pres.) 0.120. Somewhat harder tan fabric (Pantone S59-5). Single wavy line in brown paint below a solid band in streaky brown paint (Pantone S316-1; Nichoria Motif 3a). Cf. Nichoria nos. P99, P111, and P126.

1A04610. Body sherd; closed shape(?). H. (pres.) 0.025. Th. 0.006. D. (est. from max. pres.) 0.160. Sandy pinkish-tan fabric (Pantone S29-6). Horizontal band intersected by the ends of diagonal or circular (concentric circles or spiral?) lines in variegated red-brown to black, streaky paint (Pantone S36-7 to S319-4). Interior appears unpainted but is badly damaged. For diagonal lines compare Nichoria P922, P1591. For concentric circles see Nichoria Motif 7a; for spirals see Motif 7b. For similar examples on closed shapes, see Nichoria Motif 3. Compare Nichoria nos. P699, P851,and P1605 (concentric circles); no. P917 (spirals).

1A05869. Body sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.038. Th. 0.004. D. (est. from max. pres.) 0.140. Chalky light pinkish-tan to orange-tan fabric (Pantone S90-7). Monochrome coating in reddish-brown to black streaky paint (predominantly Pantone S319-3) interior and exterior.

1A04612. Body sherd; closed shape (oinochoe?). H. (pres.) 0.073. Th. 0.007. D. (est. from max. pres.) 0.180. Cream-coloured to light pink sandy fabric (Pantone S387). No interior decoration; horizontal band intersected by the ends of circular lines (concentric circles?) in worn, black, streaky paint (Pantone S325-2) on exterior. Compare Nichoria nos. P1586, P1588, P1587, P1605, P1604 and motif 3.

3

Th is catalogue is representative, not exhaustive, of the pottery identified by the MARWP Pylos Project at Englianos. DA Dark Age (refers to Nichoria typology and dating) EG Early Geometric EC Early Corinthian EPC Early Proto-Corinthian EPG Early Protogeometric G Geometric LG Late Geometric LPG Late Protogeometric MC Middle Corinthian MG Middle Geometric MPG Middle Protogeometric PA Proto-Attic PC Proto-Corinthian PG Protogeometric SPG Sub-Protogeometric (refers to Lefk andi typology and dating)

1A04816. Body sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.027. Th. 0.005. D. (est. from max. pres.) 0.140. Chalky light pinkish-tan fabric (Pantone S57-5). Monochrome coating in reddish-brown to black variegated, streaky paint on interior and exterior (predominantly Pantone S59-3). 1A04819. Handle attached to body fragment of an open shape (skyphos or deep bowl?). H. (pres.) 0.030; Th. 0.003. D. (est. from max. pres.) 0.220.

263

Ross – Post-Bronze Age Pottery

9.2 Dark Age II sherds.

Evenly-fired light pinkish-tan fabric (Pantone S57-5). Monochrome coating in reddish-brown, variegated, streaky paint on interior and exterior (predominantly Pantone S59-3).

1A05853. Body sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.047. Th. 0.006. D. (est. from max. pres.) 0.100. Cream-coloured to light pink chalky fabric (Pantone S386). Monochrome coating in brownish-black streaky paint on interior and exterior (Pantone S321-1 to S322-8).

264

Catalogue 1A05860. Body sherd near rim; open shape. H. (pres.) 0.025. Th. 0.003. D. (est. from max. pres.) 0.110. Evenly-fired cream-coloured to light tan soft fabric (Pantone S32-7). Band around outer rim with parallel vertical lines below, executed in dark brown to black paint (predominantly Pantone S318-1 to S352-2). Trace of monochrome paint interior(?). Compare Nichoria nos. P629, P707, P975, P991, P1583; also Lefkandi no. 144 (SPG I-II, c.900–825 bc): Popham and Sackett 1979–80), pl. 15.

1A01846. Body sherd; open shape (krater?). H. (pres.) 0.042. Th. 0.006. D. (est. from max. pres.) 0.100. Soft cream-coloured fabric (Pantone S26-8). Wide black band over narrow black band executed in lustrous but streaky and crackled black paint on exterior (Pantone S325-1), intersected by parallel lines visible only as stains left by paint that has since worn away. Interior badly worn. Although the design is similar to DAII sherds (see esp. 1A05865) and common on DAII material from Nichoria, it has been categorised as DAIII due to the colour and quality of its paint. Compare Nichoria no. P1245, a DAIII piece that displays a similar motif, although its design incorporates ridging as well as paint.

1A05863. High handle; cup. H. (pres.) 0.032. Th. 0.006. Cream-coloured chalky fabric (Pantone S56-8). Brown streaky paint on the exterior of handle (Pantone S316-4 to S317-1; Nichoria Handle Type A1). McDonald, Coulson and Rosser 1983, fig. 3–20.

1A02285. Body sherd; open shape (krater?). H. (pres.) 0.047. Th. 0.007. D. (est. from max. pres.) 0.160. Chalky cream-coloured fabric (Pantone S51-5) with noticeable wheel marks. Monochrome coating in lustrous but streaky and crackled black paint interior and exterior (Pantone S325-1; very badly worn on interior; categorised as DAIII based on the colour and quality of its paint). Although thicker than typical of DAIII wares, this falls within the size range and fabric thickness of DAIII Kraters from Nichoria. Compare Nichoria nos. P1185 and P1186.

1A05864. Base sherd; open shape. H. (pres.) 0.010. Th. 0.004. D. (est. from max. pres.) 0.090. Evenly fired pinkish-tan fabric (Pantone S37-7). Flat base with bump at centre. Badly worn dark brown streaky paint interior and exterior. (predominantly Pantone S853). Monochrome interior in similar colour. Compare Nichoria nos. P705, P847 and P848; cf. shapes 1–3. 1A05865. Rim sherd; open shape (cup?). H. (pres.) 0.032. Th. 0.004. D. (est. from max. pres.) 0.080. Cream-coloured to light tan soft fabric (Pantone S32-7). Band around outer rim with parallel vertical lines below, executed in variegated dark brown to black paint (Pantone S318-1 to S352-2; Motif 9). Monochrome interior in similar colour. Compare Nichoria nos. P629, P707, P975, P991 and P1583.

1A04495. Rim sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.030. Th. 0.003. Rim D. (est.) 0.100. Harder-fired reddish-yellow fabric (Pantone S44-5) with noticeable wheel marks. Monochrome black lustrous but streaky paint interior and exterior (Pantone 327-2). 1A04640. Rim sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.019. Th. 0.004. Rim D. (est.) 0.060. Soft pale tan fabric (Pantone S65-7). Monochrome streaky brown-black to black paint interior and exterior (predominantly Pantone 325-2).

1A05866. Body sherd; open shape (skyphos or deep bowl?). H. (pres.) 0.032. Th. 0.004. D. (est. from max. pres.) 0.140. Evenly-fired, light pinkish-orange chalky fabric (Pantone S50-5). Monochrome coating in reddish-brown to black variegated, streaky paint on interior (predominantly Pantone S318-2) and exterior (Pantone S48-1 to 325-2).

1A04823. Rim sherd; closed shape. H. (pres.) 0.019. Th. 0.003. Rim D. (est.) 0.160. Harder-fired light gray fabric (Pantone S31-5). Monochrome brown-black paint exterior, painted band around rim interior (Pantone S325-2).

1A05873. Body sherd; thickening toward base of open vessel (?). H. (pres.) 0.037. Th. 0.003. D. (est. from max. pres.) 0.055. Chalky light tan fabric (Pantone S70-6). Monochrome coating in reddish brown to black, variegated, streaky paint on interior and exterior (predominantly Pantone S69-2).

1A05872. Rim sherd; open shape (krater?). H. (pres.) 0.070. Th. 0.005. Rim D. (est.) 0.240. Chalky but evenly fired pale reddish-yellow fabric (Pantone S70-6). Monochrome, variegated, brown to (lustrous) blueblack paint interior and exterior (predominantly Pantone S69-2). Although thicker than typical of DAIII wares, this piece falls within the size range and fabric thickness of DAIII kraters from Nichoria, while its rim form resembles type 1A. This ware is difficult to distinguish from Late Geometric (c.750–700 bc) pottery from cemeteries around Nichoria. I have categorised these pieces as DAIII due to the combination of their large size and dark brown to black monochrome paint (similar Late Geometric monochrome pieces are rarer than in DAIII and tend to be smaller vessels such as cups). Compare Nichoria nos. P1576, P1594,

Dark Age III (c.800–750 bc) 1A01838. Body sherd; open shape (krater?). H. (pres.) 0.021. Th. 0.003. D. (est. from max. pres.) 0.080. Soft light tan fabric (Pantone S38-8). Monochrome streaky and crackled black paint on interior and exterior (Pantone S318-7).

265

Ross – Post-Bronze Age Pottery

9.3 Dark Age III sherds.

and P1596. See McDonald, Coulson, and Rosser 1983: table 3-26 and figure 3-48.

late dark age through early archaic (c.900– 550 bc)

1A05877. Rim sherd; open (skyphos or deep bowl?). H. (pres.) 0.046. Th. 0.003. Rim D. (est.) 0.140. Green-gray fabric (Pantone S41-8). Monochrome black to blue-black, lustrous, streaky and crackled paint interior and exterior (predominantly Pantone S325-3).

1A01911. Body sherd; open shape. Early 6th century bc (?) H. (pres.) 0.028. Th. 0.004. D. (est. from max. pres.) 0.080. Chalky light tan fabric (Pantone S53-8). Exterior decoration displays a broad horizontal band painted in lustrous but somewhat streaky blue-black paint (Pantone S325-2), separated from remnants of another band in reddishbrown paint (Pantone S84-5) by a narrow unpainted area. Badly worn interior surface displays traces of blue-black paint similar to the exterior band. Noticeable wheel marks in interior. Vessels constructed of fine, pale clay decorated with blue-black, purple, and reddish-brown bands are common in EC (late 7th century bc) and MC (early 6th century bc) pottery (see Weinberg 1943, chapters V and VI; nos. 188, 212, 244–48, 278, 279, 283, 342, 365, 373). Beginning in the EC period, an increasing amount of Corinthian pottery was exported, exposing a wider area to its influence (Weinberg 1943, 55). Locally manufactured copies of Corinthian wares are found at other sites in the

1A05910. Body sherd; closed shape. H. (pres.) 0.033. Th. 0.004. D. (est. from max. pres.) 0.080. Flaky light tan fabric (Pantone S70-8). Cream-coloured slip exterior (Pantone S33-7). Exterior decoration displaces a horizontal band intersected by the ends of circular lines (two abutting spirals or sets of concentric circles?) in thick, lustrous, crackled black paint fading to dark-brown (predominantly Pantone S425-1; categorised as DAIII based on the colour and quality of its paint). Unpainted interior. For abutting sets of concentric circles, compare Nichoria P699 Motif 7A. For abutting spirals see Motif 7b and Nichoria P957; for an example appearing on a closed shape, see Motif 3 and P917.

266

Catalogue western Peloponnese such as Pylos at Elis, where they are dated to the first half of the 6th century bc (Coleman 1968, nos. C47, C26, C28 and C102; illustrations 6, 7, 10–12; my thanks to J. Rutter for help in identifying this sherd).

Corinth, may both display the ‘hollow’ pattern of concentric circles depicted on this sherd (in which a number of concentric circles, in this case three, surround a blank centre). I cautiously advance a later date based on the Corinthian examples, described as being of extremely fine quality, with black paint on buff, polished clay (particularly Stillwell and Benson 1984, no. 165).

1A03190. Body sherd; closed shape (?). Late 9th to early 7th century bc H. (pres.) 0.081. Th. 0.012. D. (est. from max. pres.) 0.320. Medium fabric. Buff slip exterior (Pantone S23-8), decorated with nine narrow, parallel, diagonal, carefullydrawn, brown painted bands sloping downward and to the right into a narrow band of the same colour (fabric Pantone colour not recorded). Interior undecorated. A very similar, carefully drawn pattern appears at Lefkandi, dating to SPGIII, c.825–750 bc. See also a similar pattern on Athenian shallow bowls (late 8th–early 7th century bc), as well as ceramics from Isthmia (including a MGI skyphos, 835/25–800 bc; a MGII oinochoe, 800–750 bc; an EPC oinochoe, 725–700 bc; and an EPA krater, 710–680 bc). Similar material from Lakonia dates to the first half of the 8th century bc. From Corinth, MGII–EPC (800–700 bc) pieces are the most similar, especially those that display neatly drawn diagonals used to fill meander patterns. For Lefkandi, see Popham and Sackett 1979–80, plate 20, no. 445; Brann 1962, plate 6, no. 106. From Isthmia, see Morgan 1999, no. 257 (MGI skyphos), no. 268 (MGII oinochoe), no. 413 (EPC oinochoe), no. 431 (EPA krater), and plates 54 and 56. From Lakonia, compare Margreiter 1988, Tafel 11, no. 133, which is, admittedly, much smaller in size than this piece. From Corinth, compare Stillwell and Benson 1984, plate 2, nos. 18 and 21, and plate 4, no. 57, as well as other MGII–EPC pieces.

1A04654. Body sherd; closed shape (?). 8th–7th century bc H. (pres.) 0.067. Th. 0.007. D. (est. from max. pres.) 0.180. Hard-fired tan fabric. Light tan slip exterior (Pantone S52-1), decorated with four narrow, parallel, dark-brown to black painted bands bounded above and below by wider bands in the same colour. Paint leaves an orangebrown stain where worn. (Fabric and paint Pantone colours not recorded.) Interior appears undecorated but is badly encrusted, with a hard deposit accreted from the surrounding fi ll. Comparanda displaying similar banded patterns from the 8th century bc include Attica (LGIb–Iib, 750–700 bc), Corinthia (MGII–EPC, 800–700 bc), the Argolid (LGII, 730–690 bc), Euboia (LG, 750–690 bc), Arcadia to (LG– EPC, 750–700 bc), Lakonia (8th century bc), and Isthmia (720–700 bc). Extensive comparanda from Athens include 8th–7th century bc amphorae, 8th–7th century bc banded jugs, later 8th century bc oinochoai, and 8th century bc olphai. See Coldstream 1968, Attica (plates 8, 11, 12, 13, 14, 15: LGIb–IIb), Corinthia (plates 18, 19: MGII–EPC), the Argolid (plates 25–29: LG II) and Euboia (plate 41: LG). For further comparanda from Corinth, see Weinberg 1943, kotylai nos. 135, 112 and 146 (EPC). For examples from Arcadia, compare Voyatzis 1990, nos. P62, P73 and P64 (LG); P69 (EPC). From Lakonia, compare Margreiter 1988, Tafel 14, no. 157 and Tafel 15, no. 158. Finally, for an example from Isthmia, see Morgan 1971, plate I.52 no. 406. For further comparanda from Athens, see Brann 1962, plate 1 no. 8, esp. no. 39 (8th–7th century bc amphorae); plate 2, no. 23 and plate 4, no. 58 (both are later 8th century bc vessels, the first an amphora, the second an oinochoe); plate 5 (8th–7th century bc banded jugs); plate 21 (8th century bc olphai). It is also possible that the sherd under consideration is somewhat later, as sub-Geometric linear styles continue at Corinth throughout the 7th and into the 6th century bc. For later pieces from Corinth, see Stillwell and Benson 1984, 9–10, plate 61 no. 1476. Absolute chronology is derived from the specific publications cited, with reference where necessary to Coldstream 1968, 330.

1A03949. Body sherd; closed shape. Late 8th century bc (?) H. (pres.) 0.020. Th. 0.005. D. (est. from max. pres.) 0.100. Hard- and evenly-fired very fine tan fabric (Pantone S36-7). Exterior decoration displays three well-executed, compass-drawn concentric circles, the outermost of which is tangent to a forth, all in lustrous black paint (Pantone S325-2). Interior is undecorated. High quality of both fabric and decoration distinguish this piece. Simple, compassdrawn concentric circles have a long history from LPG through G to EPC at sites such as Tegea (LPG, late 10th century bc or later), Lefkandi (PG–SPGIII, 1050–750 bc), Lakonia (9th–8th century bc), and Corinth (last quarter of the 8th century bc). For LPG examples from Tegea, see Voyatzis 1990, nos. P5 and P8. From Lefkandi, see Popham and Sackett 1980, plate 16, no. 184 (SPGI–II); plate 18, no. 275 (SPGIII); plate 25, nos. 653, 654 and 656 (PG–SPG). For later examples from Lakonia, see Coldstream 1968, plate 46 (Lakonian LG) and Margreiter 1988, Tafel 6, nos. 52 and 53; Tafel 8, no. 77; Tafel 11, no. 127; Tafel 14 no. 157; Tafel 15, nos. 164; Tafel 16, nos. 176, 182, 186; Tafel 18, nos. 205, 206; Abbildung 8, no. 37 (9th–8th century bc). From Corinth, see Stillwell and Benson 1984, plate 9, nos. 165 and 167.

1A05876. Neck and Rim sherd; amphora. 10th–8th century bc (?) H. (pres.) 0.084. Th. 0.010. Rim D. 0.120. Hard- and evenly-fired light pinkish orange-tan fabric (Pantone S51-4). Wheel marks on interior and exterior. Exterior coated in thick but streaky matt black paint (Pantone S325-3). Interior undecorated. Judging from Attic parallels, the simple neck/rim profile and the matt black paint may indicate a LPG (950–900 bc) or EG (900–850

Dating this piece is particularly difficult, since early examples from Lefkandi, as well as late examples from

267

Ross – Post-Bronze Age Pottery

9.4 Dark Age through early Archaic sherds.

bc) date; LPG paints were usually, but not always, more lustrous than the surface displayed by this sherd. (A LPG oinochoe very similar in fabric and glaze to this piece comes from Corinth, but similar EG pieces are also attested; see Weinburg 1943, no. 7, which is LPG, but compare nos. 22, 26 and 28, all of which are EG.) LPG and EG black painted

surfaces display a great deal of similarity, and LPG black paint tends to be applied unevenly and wear poorly, producing a surface similar to that found on this sherd. By EG the sheen becomes duller and the paint more prone to peel (as on this piece), but by MGI (850–800 bc) large areas are no longer coated in black. At Corinth, streaky,

268

Catalogue sold black paint that wears poorly, covering large areas of the vessel, persists from EG through LG (c.900–725 bc), but again the simple rim profi le may argue for an early date (EG, 9th century bc). See Cook 1997, 9; Coldstream 1968, 9, 18; Desborough 1952, 119–20.

1A02789. Body sherd; open shape. H. (pres.) 0.022. Th. 0.006. D. (est. from max. pres.) 0.100. Evenly-fired, reddish-tan fabric (Pantone S56-3). Black slip interior and exterior (Pantone S325-2). 1A02828. Base sherd; open shape. Th. 0.008. Base D. 0.049. Chalky, fine, pinkish-tan fabric (Pantone S49-6). Thick but worn and somewhat dull black slip interior (Pantone S325-2). Exterior plain or slip completely worn off.

1A05899. Body sherd; closed shape (?). 8th century bc (?) H. (pres.) 0.031. Th. 0.007. D. (est. from max. pres.) 0.120. Hard- and evenly-fired fine light orangish-tan fabric (Pantone S22-4). Pattern on exterior consists of three complementary wavy lines neatly painted in black (Pantone S317-1). Interior undecorated. Chevrons and parallel wavy lines are found at Corinth (EG–MG, 900–750 bc), Athens (EG–LG, 900–700 bc), and the Argolid (EG–MG, 900–750 bc). (See Coldstream 1968, 17 and plates 16–18 (Corinth); 15 and plates 2–5, 10–11 (Athens); plates 22–23 (Argolid).) The closest analogies, however, come from Athens, where neatly drawn parallel wavy lines are most common in the Dipylon Period of the LG (730–700 bc). (Cf. Brann 1962, plate 15, nos. 254, 262, 267, 269 and 273; plates 18 and 19, nos. 319 and 325; plate 22, no. 384.)

1A03066. Handle. L. (pres.) 0.041. Handle D. 0.010. Hard-fired, orangish-tan fabric (Pantone S50-6). Black slip (Pantone S325-2). Well-made piece. 1A04083. Body sherd; open shape. H. (pres.) 0.022. Th. 0.004. D. (est. from max. pres.) 0.080. Evenly fired, fine, reddish tan fabric (Pantone colour

1A05900. Body sherd; closed shape (?). 8th–7th century bc H. (pres.) 0.054. Th. 0.007. D. (est. from max. pres.) 0.280. Light orange fabric (Pantone S50-6). Decorated with three narrow, parallel, dark red-orange painted bands bounded above and below by wider bands of similar colour (Pantone 71-2). Paint leaves a lighter red-orange stain where worn (S51-2). Interior appears undecorated but is badly encrusted with a hard deposit accreted from the surrounding fi ll. Corinth provides the best parallels with orange fabric and red-orange paint (cf. Stillwell and Benson 1984, 16, no. 13). See 1A04654. 1A05911. Body sherd; closed shape (?). Second half 8th or early 7th century bc (?) H. (pres.) 0.027. Th. 0.003. D. (est. from max. pres.) 0.040. Thin, hard-fired, fine greenish-tan fabric (Pantone S145). Pattern on exterior consists of five parallel chevrons or wavy lines neatly painted in black (Pantone S325-2). Interior undecorated. Corinth provides examples with analogous patterns (both horizontal and vertical) dating to the LG (750–725 bc); similar pieces have also been found in Lakonia (late 8th to early 7th century bc). For Corinth, see Stillwell and Benson 1984, plate 5, nos. 76, 78 and 80; plate 8, no. 153. For Lakonia, see Margreiter 1988, Tafel 44, no. 518. See also 1A05899. late archaic, classical or hellenistic black slip (600–100 bc; probably 500–300 bc) 1A01915. Rim sherd; open shape (cup or kantharos?). H. (pres.) 0.035. Th. 0.003. Rim D. (est. from max. pres.) 0.140. Dark pinkish-tan fabric (Pantone S51-5). Black slip interior and exterior (Pantone S325-1). My thanks to J. Rutter for helping to identify this sherd.

9.5 Late Archaic, Classical or Hellenistic black slip sherds.

269

Ross – Post-Bronze Age Pottery not recorded). Black slip on both interior and exterior (Pantone S325-1). 1A04519. Handle. L. (pres.) 0.019. Th. 0.003. Hard-fired, reddish-tan fabric (Pantone S44-6). Fragment of a strap handle. Black slip (Pantone S325-2). 1A04611. Body sherd; open shape. H. (pres.) 0.024. Th. 0.003. D. (est. from max. pres.) 0.080. Hard-fired, reddish-tan fabric (Pantone S58-4). Fine wheel marks. Lustrous blue-black slip interior and exterior (Pantone S325-1). 1A04660. Body sherd; closed shape (?). H. (pres.) 0.018. Th. 0.006. D. (est. from max. pres.) 0.160. Chalky, reddish-tan fabric (Pantone S52-4). Black slip exterior only (Pantone S327-1). 1A05830. Rim sherd; open shape. H. (pres.) 0.029. Th. 0.004. Rim D. (est. from max. pres.) 0.140. Hard-fired, fine, reddish-tan fabric (Pantone S71-5). Fine wheel marks interior. Lustrous black slip interior and exterior (Pantone S327-1). 1A05834. Rim sherd; open shape. H. (pres.) 0.019. Th. 0.004. Rim D. (est. from max. pres.) 0.100. Light brown fabric (Pantone S55-8). Badly worn black slip interior and exterior (Pantone S325-2). 1A05841. Lid sherd (?). Th. 0.005. Lid D. (est. from max. pres.) 0.160. Fine, pinkish-tan fabric (Pantone S57-6). Very flat piece, extending to preserved rim. Thick but somewhat dull black slip interior and exterior (Pantone S325-2). 1A05915. Rim sherd; open shape. H. (pres.) 0.028. Th. 0.005. Rim D. (est. from max. pres.) 0.320. Dark pink fabric (Pantone S89-8). Black slip interior and exterior (Pantone S325-2).

9.6 Late Classical or Hellenistic lamp.

Athenian Agora, types 25A, 25B (mid-4th through early 3rd century bc) and 28A (late 4th through mid-3rd century bc). Cf. Isthmia types VII–X (mid-4th through early 2nd century bc).

late classical or hellenistic lamp (c.500–150 bc; probably 350–250 bc)

hellenistic or roman red slip (c.300 bc– ad 200)

1A03160. Handmade lamp. H. (pres.) 0.040. L. (pres.) 0.095. W (pres.) 0.065. Handmade. Coarse, dark pink (Pantone S60-4) fabric with many inclusions. Body wider below than above. Edge of the fi lling-hole, although damaged, appears raised, creating a short neck. Handleless; convex rim; slightly concave base. Robust, elongated nozzle with small wick hole. Reminiscent of 5th- through early 2nd-century bc wheel-made lamps from Athens, Corinth and Isthmia. The convex top and small wick-hole might suggest a late Classical or early Hellenistic date; see Corinth types VII– XII (late 5th through the early 2nd century bc). Compare

1A01912. Body sherd; plate. H. (pres.) 0.050. Th. 0.007. D. (est. from max. pres.) 0.200. Fine orange-red fabric raised marks. Evenly fired; very consistent colour (Pantone 47-8). 1A04800. Rim sherd; open shape. H. (pres.) 0.023. Th. 0.003. Rim D. (est. from max. pres.) 0.100. Evenly fired, fine, but chalky pinkish-orange fabric (Pantone S60-6). Worn brownish-red slip interior and exterior (Pantone S59-2).

270

Catalogue 1A04802. Body sherd; open vessel. H. (pres.) 0.042. Th. 0.004. D. (est. from max. pres.) 0.120. Very fine fabric, with red-brown cracked slip on both sides (Pantone colours not recorded). Very even tempering and firing.

(Pantone S77-6). Worn dark pinkish-red slip interior and exterior (Pantone S62-5). 1A05892. Rim sherd; open vessel. H. (pres.) 0.031. Th. 0.009. Rim D. (est.) 0.160. Fine, soft, chalky, pinkish fabric; evenly fired. Red to redorange slip.

1A04804. Rim sherd; open shape. H. (pres.) 0.025. Th. 0.005. Rim D. (est. from max. pres.) 0.140. Evenly fired, fine, but chalky dark pinkish-tan fabric

1A05913. Rim sherd; open vessel. H. (pres.) 0.029. Th. 0.009. Rim D. (est.) 0.200.

9.7 Hellenistic or Roman sherds.

271

Ross – Post-Bronze Age Pottery Some inclusions, hard fabric orange-red in colour. Evenly and well fired.

Somewhat chalky, pinkish fabric. Mottled brown glaze interior (predominantly Pantone S45-2 to S320-5), with dark brown line and wave pattern (Pantone S327-1); crackled glaze, worn through in places. No glaze exterior; pronounced wheel marks exterior. Similar examples come from Nichoria (no. P1711, as well as other Area IV Glazed Wares) and the Istanbul Museum.

late roman plain ridged ware (c. ad 100–600; probably ad 200–500) 1A04553. Body sherd. H. (pres.) 0.028. Th. 0.003. Small sherd size precludes diameter estimate. Pale brown sandy fabric (Pantone S43-5). Parallel bands end approx. 12 mm from top of sherd.

1A05842. Body sherd; open vessel. 10th–14th century H. (pres.) 0.044. Th. 0.006. D. (est. from max. pres.) 0.240. Chalky, pinkish fabric (Pantone S58-4); distinct wheel marks interior. Yellowish tan glaze on interior (Pantone S58-7), reminiscent of yellow glazed ware (see below). Exterior coated with thick, off-white slip (Pantone S63); single dark green-black band in thick paint exterior (Pantone S314-1).

1A04554. Body sherd. H. (pres.) 0.031. Th. 0.005. D. (est. from max. pres.) 0.160. Pale red-brown sandy fabric (Pantone S43-6). Parallel bands end approx. 10 mm from top of sherd; the edge of a raised band is preserved at the top edge, perhaps indicating a sherd near the rim of the vessel.

Yellow Glazed Wares 1A04590. Body sherd. H. (pres.) 0.030. Th. 0.003. D. (est. from max. pres.) 0.220. Red-brown sandy fabric (Pantone S65-3). Parallel bands end approx. 10 mm from rim.

1A04798. Rim sherd; open vessel (plate or shallow bowl?). 12th–13th century Th. (pres.) 0.005. Rim D. (est.) 0.440. Hard pink fabric (Pantone S38-7) with wheel marks. Thick pale yellow glaze exterior (Pantone S13-3), continues over rim where it thins, lightens, and takes on a greenish tinge. Interior within rim has no paint or glaze, only a clear covering glaze.

1A05848. Body sherd. H. (pres.) 0.026. Th. 0.003. D. (est. from max. pres.) 0.260. Dark brown exterior (Pantone S324-7) with red-brown interior (Pantone S79-3); sandy fabric. 1A05975. Rim sherd; open vessel (?). H. (pres.) 0.027. Th. 0.003. Rim D. (est.) 0.120. Dark gray-brown sandy fabric (Pantone S327-5). Simple rim shape (see Athenian Agora M353 in Robinson 1959). The small size of this sherd makes identification of the shape difficult. The curvature of the sherd and the large estimated rim diameter argue for an open shape, although some closed shapes have similar rim profiles. Compare Robinson 1959, nos. L37, M151, M176, M219 and M266, all closed shapes with simple rims.

1A04799. Body sherd; closed vessel. 12th–13th century H. (pres.) 0.036. Th. 0.006. D. (est. from max. pres.) 0.160. Somewhat chalky pink fabric (Pantone S38-7) with prominent wheel ridges. Flaky mustard-coloured glaze interior with a trace remaining exterior (Pantone 13-3). 1A04814. Body sherd; closed vessel (?). 12th–13th century H. (pres.) 0.027. Th. 0.007. D. (est. from max. pres.) 0.170. Softer, somewhat chalky pink fabric (Pantone S38-7) with prominent wheel ridges. Flaky mustard-coloured glaze interior only (Pantone 13-3).

medieval (c. 1000–1400) Glazed Wares

Byzantine Spongy Ware 1A04797. Base sherd; open vessel. 12th–13th century (?) H. (pres.) 0.027. Th. 0.011. Base D. (est.) 0.075. Fine, soft, chalky, pinkish fabric (Pantone S38-7); evenly fired; does not hold glaze well. Trace of yellow glaze exterior (Pantone S13.3), similar to yellow glazed ware (see below). Flaky white glaze interior (Pantone S6.7) with green curvilinear design (Pantone S261-1); thick green paint bleeds into white. Nichoria (no. P1710, as well as other Area IV Glazed Wares), and Athens (Athens Byzantine Museum #174) offer comparanda.

1A04589. Body sherd. 11th–14th century H. (pres.) 0.044. Th. 0.015. D. (est. from max. pres.) 0.028. Brown fabric (Pantone 65-5), tempered with coarse sand. Uneven exterior surface colour; blackened interior. Distinctive pinholes in fabric. Cf. Nichoria (nos. P1680– 89) and Sparta (Sanders et al. 1993, nos. 62–73) .

1A04801. Base; open vessel. 10th–14th century H. (pres.) 0.019. Th. 0.010. Base D. 0.071.

272

Catalogue

9.8 Medieval sherds.

Bibliography Blegen, C. W., Palmer, H. and Young, R. S. 1964. Corinth, The North Cemetery. ASCSA, Princeton, NJ.

Brann, E. T. H. 1962. The Athenian Agora, VIII: Late Geometric and Protoattic Pottery. ASCSA, Princeton, NJ. XIII: Broneer, O. 1930. Corinth, IV/ II: Terracotta Lamps. ASCSA, Cambridge, MA.

Blegen, C. W., and Rawson, M. 1966. The Palace of Nestor at Pylos in Western Messenia, I: The Buildings and Their Contents. Princeton University Press, Princeton, NJ.

Broneer, O. 1977. Isthmia, III: Terracotta Lamps. Princeton University Press, Princeton, NJ. Coldstream, J. N., and Huxley, G. L., eds. 1972. Kythera: Excavations and Studies Conducted by the University of Pennsylvania Museum and the British School at Athens. Faber & Faber, London.

Blegen, C. W., Rawson, M., Taylour, W., and Donovan, W. P. 1973. The Palace of Nestor at Pylos in Western Messenia, III: Acropolis and Lower Town; Tholoi, Grave Circle, and Chaber Tombs; Discoveries Outside the Citadel. Princeton University Press, Princeton, NJ.

273

Ross – Post-Bronze Age Pottery Rosser, J. 1983. ‘The [Byzantine] Pottery’, in Excavations at Nichoria in Southwest Greece, III: Dark Age and Byzantine Occupation, ed. W. D. McDonald, W. D. E. Coulson and J. Rosser, 378–97. University of Minnesota Press, Minneapolis.

Coldstream, J. N. 1968. Greek Geometric Pottery: A Survey of Ten Local Styles and Their Chronology. Methuen, London. Coleman, J. E. 1986. Excavations at Pylos in Elis. ASCSA, Princeton, NJ.

Sparkes, B. A. 1970. Athenian Agora, XII: Black and Plain Pottery of the 6th, 5th, and 4th Centuries BC. ASCSA, Princeton, NJ.

Cook, R. M. 1997. Greek Painted Pottery, 3rd edn. Routledge, New York.

Sparkes, B. A. 1991. Greek Pottery: An Introduction. Manchester University Press, Manchester and New York.

Coulson, W. D. E. 1983. ‘The Dark Age: The Pottery’, in Excavations at Nichoria in Southwest Greece, III: Dark Age and Byzantine Occupation, ed. W. A. McDonald, W. D. E. Coulson and J. Rosser, 61–259. University of Minnesota Press, Minneapolis.

Stillwell, A. N., and Benson, J. L. 1984. Corinth, XV/3: The Potters’ Quarter. The Pottery. ASCSA, Princeton, NJ.

Desborough, V. R. d’A. 1952. Protogeometric Pottery. Oxford University Press, Oxford.

Voyatzis, M. E. 1990. The Early Sanctuary of Athena Alea at Tegea and Other Archaic Sanctuaries in Arcadia. Studies in Mediterranean Archaeology 97. Paul Åströms Förlag, Göteborg.

Griebel, C., and Nelson, M. C. 2001. ‘Post-Mycenaean Occupation at the Palace of Nestor’, unpublished paper, http://clvl.cla.umn. edu/marwp/PYLOS/aia1992.html.

Waywell, G. B., Wilkes, J. J., Baily, D. M., Sanders, G. D. R. 1993. ‘Excavations at Sparta: The Roman Stoa, 1988–91: Preliminary Report, Part I’, Annual of the BSA 88, 219–86.

Howland, R. H. 1958. The Athenian Agora, IV: Greek Lamps and their Survivals. ASCSA, Princeton, NJ.

Weinberg, S. S. 1943. Corinth, VII/I: The Geometric and Orientalizing Pottery. ASCSA, Cambridge, MA.

Margreiter, I. 1988. Frühe lakonische Keramik der geometrischen bis archaischen (10. bis 6. Jahrhundert v.Chr.). Stift land-Verlag, Waldsassen-Bayern.

Revised 2013

McDonald, W. A. and Coulson, W. D. E. 1983. ‘The Dark Age at Nichoria: A Perspective’, in Excavations at Nichoria in Southwest Greece, III: Dark Age and Byzantine Occupation, ed. W. A. McDonald, W. D. E. Coulson and J. Rosser, 316–29. University of Minnesota Press, Minneapolis. McDonald, W. A., Coulson, W. D. E., and Rosser, J., eds. 1983. Excavations at Nichoria in Southwest Greece, III: Dark Age and Byzantine Occupation. University of Minnesota Press, Minneapolis. Morgan, C. 1999. Isthmia, VIII: The Late Bronze Age Settlement and Early Iron Age Sanctuary. ASCSA, Princeton, NJ. Morgan, C. H. 1942. Corinth, XI: The Byzantine Pottery. ASCSA, Cambridge, MA. Popham, M. R. 1991. ‘Pylos: Reflections on the Date of its Destruction and on its Iron Age Reoccupation’, Oxford Journal of Archaeology 10/3, 315–24. Popham, M. R., and Sackett, L. H. 1979–80. Lefkandi I: The Iron Age. Thames & Hudson, London. Robinson, H. S. 1959. The Athenian Agora, V: Pottery of the Roman Period. Chronology. ASCSA, Princeton, NJ.

274

10 POST-BRONZE AGE INDUSTRIAL WASTE AND BRONZE CASTING caitlin downey

During the 1990–98 seasons of MARWP’s investigations at Pylos, an assortment of porous and semi-vitrified terracotta fragments were uncovered (fig. 10.1). These pieces were originally interpreted as debris from pottery production, either pottery ‘wasters’ (burned and malformed pieces produced by misfires or kiln accidents) or fragments of kiln walls. On closer examination, however, it became evident that they have traits not typically found in ceramic industry debris. Most of the pieces are much lighter than typical ceramics, having a specific gravity of 0.8 – approximately one-third that normally found in ceramics. Nearly all preserve smooth convex and concave planes on one side, while the opposite side appears lumpy and formless. Furthermore, many of the pieces show heavy vitrification on this lumpy surface, whereas the smooth side typically retains a fine, creamcoloured slip. These features, when taken in combination, are more indicative of metalworking than ceramic or other craft activity; indeed, the fragments appear to be the broken remnants of bronze casting moulds. Based on the joins made between several of the pieces, is seems clear that bronze casting was being conducted at Englianos in the historical period.

heat, beyond that to which ceramics would be exposed during kiln firing. Some are scorched on their surfaces, and many have patches of vitrification on the surface or running through the fabric of the pieces. While such characteristics may be found in pottery wasters, they are not typical of ordinary potsherds. The third feature involves the structure and composition of the pieces, which fall into three groups: triple-layer fragments, double-layer fragments and double-layer spongy fragments. In triple-layer fragments (fig. 10.1) the first or inner layer is quite thin, normally measuring 1 mm or less. It is composed of extremely fine clay with no inclusions and resembles a coating of fine terracotta slip. The surface of this layer is sometimes charred black and gray, and it is often badly cracked. Also visible at places are fine, parallel marks that resemble brush or rasp strokes. This layer is covered by a second one, usually 2–3 mm thick, which is composed of a fine clay with few inclusions. The final layer consists of a much coarser clay with medium-sized (1–4 mm) inclusions. This layer varies in thickness, typically between 2–3 cm. The exterior of this outer layer appears to have been smoothed by hand. Double-layer fragments (fig. 10.2) are identical to the triple-layer fragments except that they lack the innermost, fine layer of clay. Double-layer spongy fragments (fig. 10.3) also have two layers, but unlike the simple doublelayer fragments, the inner layer of these is made of a thin application of fine clay similar to that on the triple-layer type. The second layer, however, is quite different than anything in the other two types, being highly porous with a spongy appearance. This layer is often thicker than those of the other two categories, measuring 2–6 cm. Unlike the outer layers of the previous two types, however, it is somewhat formless and mirrors the inner shape of the piece much less. When vitrification occurs, it invariably affects the exterior of this porous surface.

Description of the Terracotta Fragments The first step in determining what these strange Pylian fragments originally were was to determine what they were not. Three features in particular distinguished the Englianos pieces from ordinary potsherds. The first is the shape of the pieces themselves. As noted, all the fragments preserve, on one side, a distinct convex and/ or concave surface. The exterior surface of these pieces sometimes roughly mirrors that of the interior surface, but the two surfaces are never perfectly parallel to each other, unlike the form of typical body sherds. Second, the Englianos pieces show evidence of exposure to extreme 275

Downey – Post-Bronze Age Industrial Waste and Bronze Casting Possible Interpretations At first glance, the features characteristic of these mysterious fragments – burning, multiple ceramic layers, preservation of the interior contours and, periodically, the use of highly porous clay – can be found in other comparanda. cookware One interpretation might be that these pieces are fragments of cookware, which, due to its constant exposure to heat, often appears blackened both on its exterior and within the fabric itself (fig. 10.4). It is rare, however, that the heat caused by cooking becomes sufficiently intense or prolonged actually to vitrify the clay, as is seen in the Pylos fragments. Furthermore, cookware, because it is manufactured to undergo such harsh treatment, is generally not made from fine clays. The Englianos pieces, on the other hand, while partially constructed of medium and coarse clays, also have a layer of finer clay on the interior, a feature unparalleled in cookware. Finally, the complex convex/concave shapes preserved on the interiors of the some of these fragments would not have been suitable for the interior of cooking vessels, which need broad, smooth surfaces for even heating and cooking.

10.1 Triple-layer terracotta fragments, showing the fine, sliplike layer.

kiln debris Another possibility is that the Pylos pieces are the remains of kiln and firing debris. Debris from pottery production has some of the traits seen in the pieces from Englianos, most specifically burning and vitrification. These can be seen not only in wasters, but also frequently in fragments of kiln walls. Beyond this, however, the comparisons

10.2 Double-layer terracotta fragment.

10.3 Double-layer terracotta fragment showing the porous, spongy main layer.

10.4 Pithos fragment exhibiting blackening to surface and interior fabric.

276

Possible Interpretations between the Englianos pieces and ceramic industry debris are few. A number of genuine wasters came to light during the MARWP excavations, and the difference in appearance between them and the fragments under discussion is immediately apparent. Despite their collapsed and warped appearance, the original shapes of the pieces can often be recognised within the waster, but none of the Englianos terracotta fragments can be identified with any known vessel type. It is, of course, possible that they are all body sherds from very large vessels, but this does not explain why their interior and exterior surfaces retain different contours that do not parallel each other; while the surface of wasters might become bubbly and vitrified in places, the exterior surface nevertheless generally mirrors the shape of the interior surface. Moreover, in no instance do wasters ever split so neatly into separate layers as these fragments often do. Indeed, the clay of over-fired pottery often condenses and becomes quite heavy in proportion to its size – quite the opposite of the peculiar spongy material from Englianos (Rye 1981).1

In their general form, these ceramic moulds resemble the Englianos fragments. Nevertheless, because the ceramic moulds are fired only once, they show no evidence of the burning that appears on the Pylos pieces, let alone vitrification. Moreover, the moulds are made entirely of a finer clay than was used for the Pylos fragments, as fine clay was needed to produce the detail required in objects such as terracotta reliefs and Megarian bowls. There is, as well, no layering of clays in the moulds, nor any coating of fine slip. A pertinent point to make here is that ceramic moulds were always made to produce open shapes – shapes that do not fold back on themselves at any point; a closed shape would require that the mould be broken to remove the object within it, thus making it impossible to reuse. Similar in form to ceramic moulds are moulds used in glass production. Excavations in the Workshop of Pheidias at Olympia have yielded moulds for the production of glass and gold drapery for the cult statue in the Temple of Zeus (Schiering 1991). Like ceramic moulds, these resemble the Englianos fragments in their surface contours, but the same objections to glass moulds being the original function of the Pylos fragments can be made.

Fragments of kiln walls are a more tantalising solution. Not only do they usually show traces of burning and vitrification, but some kiln walls appear to have been built up in numerous layers, as in those from Amphipolis (Lazarides 1973, 43–54) and Haghia Triadha (Levi and Laviosa 1979–80, 7–47). Nevertheless, the outer layer of the kiln wall that covers the underlying mudbrick substructure in such kilns is rather sloppily applied and does not at all resemble the smooth, broad curves of the Englianos fragments. That these pieces could be fragments of the underlying mudbrick wall of the kiln (as seen at Haghia Triadha) is equally unsatisfactory. There is no precedent for individual mudbricks being constructed in layers, nor do any of the Englianos pieces preserve a squared face or corner, as one might expect were these mudbrick fragments.

industrial bronze casting debris The closest points of comparison in the suggestions made above for the Englianos fragments were with pottery and kiln debris (in terms of burning and vitirification) and with ceramic moulds (in terms of unusual surface contours). A suggestion for the function of the Pylos fragments that combines these features is that of bronze casting debris. The Lost-Wax Method of Bronze Casting The method used to cast ancient bronzes (see Mattusch 1988) explains many of the peculiarities of the Englianos pieces. The sculptor began with a wooden or clay core, upon which was laid a thin wax mantle that was carefully carved to replicate exactly the form and surface details of the eventual bronze sculpture. Over this were painted several coats of fine, liquid clay. This slip was, in turn, coated with a layer of coarser clay, which was then covered by a layer of even coarser clay (though sometimes only two layers of clay – one fine, one coarse – were used). These coarse outer layers, known as ‘investment’ layers, served as a refractory material, helping to reduce cracking and shrinkage in the subsequent firing and casting of the bronze. Nails were driven through the investment layers into the core to keep it in place during firing. The mould was allowed to dry and then was fired. During this firing the wax melted and was released through holes left in the investment layers (giving the name of ‘lost-wax method’ to the process).

ceramic or glass moulds Another possibility is that the Englianos pieces are fragments of moulds used in ceramic production. In order to expedite the production of terracotta relief sculpture and certain types of vessels, such as Megarian bowls (see Akamates 1993), ceramic moulds were used to enable mass production of these objects. These moulds are usually flat and/or smooth on one side, while on the other they preserve the reverse image of the shape they are intended to reproduce. Thus the contours of their interior and exterior surfaces do not match, the interior surface often showing a complexity of convex and concave forms. Before use, the mould itself was fired, so that it could be reused indefinitely. The desired object was created by pressing wet clay into the mould, removing it when it had air-dried, then firing it.

After cooling, the mould was transferred to a casting pit, where it was buried in earth and sand and heated again

1 My thanks to Kostas Chrysogelos, a potter from Limenas Thasou, who kindly discussed pottery firing techniques and kiln mishaps with me.

277

Downey – Post-Bronze Age Industrial Waste and Bronze Casting to prepare it for receiving the bronze. The molten metal was poured into the mould through holes and channels left in the investment layers. Once the mould cooled, the investment was broken off to reveal the bronze object beneath. The broken mould debris, scorched from the firing and from contact with the molten bronze, was usually thrown back into the casting pit and eventually buried. The end product of this process was a bronze sculpture that was not solid metal, but only a skin of bronze around an inner core. While small figurines could be formed in solid bronze, the process would have been economically and structurally prohibitive for larger pieces, making the lost-wax method the standard technique for large-scale bronze sculpture throughout antiquity. Mould fragments from lost-wax casting have been uncovered at a number of sites in Greece, for example the Athenian Agora, Palaikastro and Corinth (see Mattusch 1977a, 1977b, 1991; Hemingway 1996). Like the Englianos pieces, these fragments preserve a distinct convex and/ or concave surface on one side. The exterior surface roughly mirrors that of the interior surface, but the two are never perfectly parallel to each other. Some pieces are scorched on their surfaces and in the interior, and many have patches of vitrification on the surface or running through the fabric of the pieces. Sometimes the first, fine layer of clay has been lost, but normally at least two layers of the investment are preserved. On the inner surface of these pieces, fine parallel lines can sometime be seen – the impression made by the marks of a rasp used to work the wax mantle underneath the mould. The Englianos Fragments as Bronze Casting Waste The parallels to the Englianos fragments are striking. The only difference between the mould fragments found at the sites just mentioned and those from Pylos is the presence of flecks of bronze clinging to cracks in the mould surface on some of the non-Pylian pieces. No bronze has been found on the Englianos fragments, but there are numerous pieces from both Palaikastro and the Athenian Agora that, while indisputably bronze moulds, also do not preserve any bronze.2 While many of the pieces from Palaikastro, the Athenian Agora and Corinth are too fragmentary to determine the shape of the object that would have come from the mould, enough pieces have been joined to identify moulds for tripod legs at Palaikastro and for a kouros at Athens. The Englianos pieces, too, are incomplete, but it has been possible to join nine fragments to form a piece that appears to have been a mould for a forearm and, indeed, fits snugly in position in this location (fig. 10.5). Other pieces have preserved enough of their contours to suggest other possibilities. One retains the shape of two fingers 2

My thanks to the staff of the Athenian Agora for assistance in locating and examining fragments of bronze casting moulds not on display in the Agora museum.

10.5–10.7 Fragments forming possible forearm mould, showing the shape of two fingers and possible drapery mould.

278

Dating the Fragments

10.8–10.9 Examples of other joining terracotta fragments.

and fits neatly between the thumb and forefinger of a hand (fig. 10.6). Another preserves a convoluted surface that might have been a mould for crinkly drapery or perhaps tree bark (fig. 10.7). Other fragments are more amorphous (figs 10.8–10.9), but their broad, smooth, curving surfaces may indicate moulds for drapery (as is the suggestion for many of the Athenian Agora pieces) or perhaps even for human musculature. It would seem, therefore, that we have at Englianos moulds for large-scale, figural statuary.

If the terracotta pieces from Pylos are fragments of bronze casting moulds, the ramifications are significant: they suggest that someone was casting large-scale bronzes at some time up on the Englainos ridge. Because these fragments were not found in situ, they cannot be dated stratigraphically, and we must look to other factors in an attempt to construct a time-frame for this activity.

Dating the Fragments There is still the problem, however, of the spongy fragments. Although comparisons to this material in the Greek world have not been forthcoming,3 moulds made of clays that are mixed with organic material such as dung, straw, hair and sand do occur beyond the Mediterranean.4 This particular combination of refractory material allows the moulds to be super-heated before the molten bronze is poured in, enabling the creation of very fine, thin bronzes. When the moulds are fired, the organic material burns, leaving the clay with a highly porous appearance. The subsequent super-heating of these moulds in preparation to receive the bronze results in vitrification of their outer surface. Because bronze casting debris has often been overlooked in archaeological excavations, it seems likely that the material we have represents only a small fragment of the types of moulds used in antiquity. These spongy fragments suggest that the bronze-workers at Pylos were experimenting with different investment compositions, and had perhaps discovered that a particular combination of clay and organic material allowed them to produce thinner and therefore more economical bronzes.

It is unlikely that the bronze casting activity evidenced on the hilltop dates to the Bronze Age, despite the fact that most of the finds from the site belong to the Mycenaean occupation of the palace. Although the Minoans used the lost-wax method for hollow casting, there is no evidence that sculptors on the Greek mainland cast anything larger than small, solid bronze objects. Hemingway (1996, 243), in his publication of the Minoan mould material from Palaikastro, noted that the discovery there of moulds for large tripod stands suggests that ‘we may well be missing an entire class of metal objects that simply have not been preserved in the archaeological record’. This is a possible scenario for Minoan Crete, where we know that lost-wax casting was being practiced and where there is secure evidence for large-scale bronze working (for a mould for a three-quarter life-size hand and clenched fist from Phaistos, see Laviosa 1967–68, 499–510). Nevertheless, although it is conceivable that a similar situation holds true for Mycenaean metalwork, we would have to theorise from negative evidence entirely, and there are other, more likely dates for the material here.

3

A type of Medieval ceramic called ‘spongy ware’ was identified at Nichoria and on MARWP’s investigations at Pylos (see pp.261–62). Th is fabric, however, is entirely different from the spongy material under discussion here. Unlike our fragments, the ‘spongy ware’ from Nichoria and Pylos is pricked only on the surface with small pin-holes, probably the result of overheating in the kiln; the fabric itself is not spongy. Furthermore, the pieces from Nichoria are much thinner, are glazed and are clearly vessel sherds. See McDonald, Coulson and Rosser 1983, 379–80. 4 My thanks to Robert Poor of the Art History Department of the University of Minnesota for discussing traditional Chinese bronze casting methods with me and for taking the time to examine photographs of the debris fragments from Pylos.

Because bronze casting debris has often been overlooked in the past, there is no typology for bronze casting moulds. In general, Roman moulds are thinner and finer than Greek moulds, with less bronze adhering to them, but little more than this has been determined (Mattusch 1977a and 1977b). The moulds from Englianos are fairly thick, and so can be cautiously dated to between the Mycenaean period and the Roman era – a vast chronological range. The sculpture suggested by the shape of the mould surfaces may prove useful for determining a chronological range for the objects produced by the Englianos moulds. Figure

279

Downey – Post-Bronze Age Industrial Waste and Bronze Casting 10.4 shows a mould for a hand with at least the forefinger and thumb separated and outstretched, and the forearm piece in Figure 10.5 indicates an arm with the hand curving upwards. We are not, therefore, looking at sculpture that is rigidly stiff, with arms and hands clenched at the sides; the form of the sculpture suggests something postArchaic. If Figure 10.7 does in fact preserve the impression of crinkly drapery, we have another indication of late date. Although drapery of this sort was depicted in garments worn by Archaic korai, the Pylos fragment appears less stylised and more animate than that seen on korai figures. It resembles more closely the sort of crinkly, water-soaked drapery that appears first in the late 5th century bc, such as is seen on the pedimental sculptures of the Parthenon or from the late Hellenistic period.

Mattusch, C. 1988. Greek Bronze Statuary from the Beginnings through the Fifth Century BC. Cornell University Press, Ithaca, NY. Mattusch, C. 1991. ‘Corinthian Metalworking: The Gymnasium Bronze Foundry’, Hesperia 60, 383–95. McDonald, W. A., Coulson, W. D. E., and Rosser, J., eds. 1983. Excavations at Nichoria in Southwest Greece, III: Dark Age and Byzantine Occupation. University of Minnesota Press, Minneapolis. Schiering, W. 1991. Olympische Forschungen, XVIII: Die Werkstatt des Pheidias in Olympia. Walter de Gruyter, Berlin. Rye, O. S. 1981. Pottery Technology. Taraxacum, Washington, DC.

It appears that there is evidence of a large-scale bronze casting operation at Pylos in historic times, sometime after the Archaic period. It is clear from other material found at the Palace of Nestor that the ridge was occupied well after the Mycenaean period, but it is less clear why such a bronze casting operation might have existed on the hilltop. If there was indeed a later temple on the site, as is argued in Chapter 8, it is possible that this bronze casting activity might have been associated with it, perhaps in the form of a workshop for temple sculptures and/or votives. Whether or not this is the case, what is clear is that there was significant activity on the hilltop in the historic period; these pieces are part of the mounting evidence that the Englianos ridge had a life far before and far beyond the Mycenaean period.

Submitted 2002

Bibliography Akamatis, G. M. 1993. Πηλινες Μητρες Αγγείων από την Πέλλα. Athens. Hemingway, S. 1996. ‘Minoan Metalworking in the Postpalatial Period. A Deposit of Metallurgical Debris from Palaikastro’, Annual of the BSA 91, 213–43. Laviosa, C. 1967–68. ‘Una forma minoica per fusiona a cera perduta’, ASAtene 45–46, 499–510. Lazaridis, D. 1973. ‘Ανασκαφή και ερεβναί Αμφιπόλεως’, Praktika, 43–54. Levi, D., and Laviosa, C. 1979–80. ‘Il forno minoico da vasaio di Haghia Triadha’, ASAtene 57–58, 7–47. Mattusch, C. 1977a. ‘Bronze and Ironworking in the Area of the Athenian Agora’, Hesperia 46, 340–79. Mattusch, C. 1977b. ‘Corinthian Metalworking: The Forum Area’. Hesperia 46, 380–89.

280

THE MINNESOTA PYLOS PROJECT, 1990–98

PART II THE ARCHITECTURE OF THE PALACE OF NESTOR MICHAEL C. NELSON

1 THE PALACE AT ENGLIANOS

Today, some 3500 years after its abandonment and 65 years after its initial discovery, the late Bronze Age palace at Englianos remains the best preserved such edifice of its kind. Its buildings and ancillary structures cover an area of more than 13,500 square metres (plans I and II). Extensive plaster floors and well-built masonry walls, some of which still stand to a height of more than a metre, clearly define the complex and delineate the interior arrangement of the rooms and the circulation therein. Where walls are missing, regular gaps in the plaster floors outline their original footprints; cut-stone doorjambs and threshold blocks mark the numerous doorways. In its final phase, at the end of the Late Helladic IIIB period, the palace was Mycenaean in form. Most of the remains visible today at the site date to this period. At the centre of this final complex lay its most impressive suite of rooms: the megaron (rooms 4, 5 and 6), entered via a grand and formal route composed of a columned propylon (rooms 1 and 2) and open court (3). Sister palaces at Mycenae (fig. 1.1) and Tiryns (fig. 1.2) incorporated this same distinctive sequence.

The site was discovered in 1939 by Carl W. Blegen and excavated from 1954 to 1966 under the direction of Blegen and Marion Rawson. The palace was published as, essentially, a single-period LH IIIB building, and very few state plans – detailed, stone-by-stone drawings of the in situ remains – were provided (Blegen and Rawson 1966; hereafter PN I). Nevertheless, multiple building methods and phases are obvious in the surviving remains and attest to the hilltop’s long and varied history. The aim of this study was to re-examine the buildings at Englianos firsthand, and to produce detailed state drawings of all the surviving architecture.

1.1 Mycenae palace plan (after Wace 1921–23, pl. II).

1.2 Tiryns palace plan (after Muller 1930, vol. III, pl. 4).

283

Nelson – The Architecture of the Palace of Nestor Chapters 2 and 3 discuss the building materials employed in the construction of the palace and the methods by which these materials were combined to create structural elements, namely load-bearing walls. Stone, worked in various ways, was the primary building material. Unworked stone was used in rubble masonry, partially worked slabs were stacked to form large retaining walls, and finely cut blocks were used to create impressive facades. (Englianos preserves the only extant orthostate walls in mainland Bronze Age architecture.) Sophisticated jointing techniques fastened wood to stone blocks in coursed ashlar masonry. The use of half-timbering in many of the palace’s walls, which is thought to be typical of Mycenaean architecture, is brought into question because of the significant lack of evidence for wooden members in the walls at the time of the palace’s destruction. Instead, a new building system is proposed, consisting of rubble bound in a matrix of strong mortar. Buildings were founded on bedrock or strong foundations, and water was carefully drained away from the hilltop through built channels. All these practices insured architectural stability and durability, and demonstrate a sophisticated and well-organised building industry.

The Englianos ridge is one of the many finger-like projections of land in southwestern Messenia that reach down from the inland hills to the coastal plain surrounding the Bay of Navarino. At its southern end is a small, flat plateau approximately 4800 square metres in area; the fence shown in Plan I marks its edges. On the northwest and southeast side, the topography descends sharply into deep valleys; to the southwest and northeast, the downward slope is gentler. The view from the plateau is magnificent: the bay lies to the south and southeast, and Mount Aigaleon rises to the northeast. Numerous terraces planted with olive trees fi ll the panorama in all directions. The cluster of buildings forming the LH IIIB palace complex occupies the southwestern edge of the plateau. At the centre is the Main Building (rooms 1–58), its southeast facade pierced by a monumental entranceway (plan ii). In plan, the propylon (rooms 1 and 2; plan viii) is H-shaped and strictly symmetrical: two column bases stand at either end of its long axis, each of which is positioned between antae on the propylon’s short axis. One anta block remains in situ at the northwest end of the northeast wall of Room 2; Blegen and Rawson reconstructed the other three antae based on the existing one and on the remains, or lack thereof, of the plaster floors in rooms 1–3 and Court 58 (PN I, 54–55, 61). The entrance is further emphasised with the arrangement of buildings and open spaces on the southeast side of the palace complex. The large plastered court (58), the portico (92 and 94) of the Northeast Building on the northeast, and the tall Building X on the southwest all focus attention on the entrance to the Main Building . A single column base (labelled 58SW on plan v) may be the remains of a larger colonnade on the southwest side of Court 58, which would have complemented Portico 94 on the opposite side.

Chapter 4 defines the major Bronze Age building phases at Englianos, based on the development of specific building systems in concert with structural changes in the built spaces and alterations to the spatial configuration of the complex. In its early phases, the palace builders adopted construction methods and design concepts from Crete: the use of ashlar and orthostate masonry at Englianos is strikingly similar to that of Minoan architecture. A change occurred at the end of LH IIIA, however, when the site lost much, if not all, of its foreign architectural influences and fell into line with its fellow Peloponnesian palaces. At Englianos, Mycenae and Tiryns, design focused on hilltop or plateau settings, prominent megara and controlled circulation. Nonetheless, the architecture at Englianos stands alone among its contemporaries for its development and use of cut stone masonry and its invention, or at least widespread exploitation, of a new type of rubble and mortar building system.

The propylon leads into a large open-air court (3) surrounded on all sides by rooms of the Main Building. Its plaster floor exhibits rain damage from the roofline around its edges. Beyond, to the northwest, lies the megaron, the largest suite of rooms in the palace; it comprises a porch (4), vestibule (5) and the so-called Throne Room (6; no actual throne was discovered in the excavations), which is itself the largest single room in the palace complex. Like the propylon, the megaron’s architecture is symmetrical: the two columns of the porch and the four columns and hearth of Room 6 align along a central axis that runs northwest–southeast down the middle of the suite.

A Brief Description of the LH IIIB Palace The following discussion uses as points of reference the building names and room designations assigned by Blegen and Rawson in their publication of the palace. I add additional labels where necessary for clarification. Some rooms and areas were never assigned numbers by the original excavators; these are given room designations here. The presumed functions of the various buildings and ancillary structures are not discussed here at any length, and the reader is directed to the bibliography for more thorough examinations of the possible uses of the spaces described below.

The megaron, open court and propylon together create a strikingly formal sequence of related spaces and rooms. They must surely represent the deployment of a clear architectural concept and the imposition of organised, and perhaps controlled, circulation patterns.1 In plan, the megaron and propylon are each symmetrical around their own axes, but they do not share the same axis. This non-axial approach to the megaron seems to be standard among Peloponnesian palaces: at neither Mycenae (fig.

284

The Palace at Englianos 1.1), Tiryns (fig. 1.2) nor Englianos (fig. 1.3) can one advance directly from the propylon towards the megaron along its axis. At Mycenae, the propylon is some distance away from the open court, and the whole arrangement is more complicated than that at Englianos and Tiryns; at Tiryns, the approach consists of two propylaia and two open courts. At Englianos, the slightly offset axis of the propylon in respect to that of the megaron forces the visitor to approach the porch and the rooms beyond from an angle rather than straight on. That visitor would have

been impressed by the megaron’s facade, with the two fluted columns of Room 4 standing in antis and colourful frescoes adorning its walls (for which only the dado courses are certain (PN III, 192)). To either side of the megaron run long corridors that granted access to the flanking suites of rooms to the southwest and northeast and to magazines 23 and 24 at the rear of the building; they also provided an alternative route to the Throne Room via Vestibule 5. The southwest corridor once led to a series of storage rooms (19–21), but in a later remodelling phase it was blocked by two cross walls (plan vii). These created one long room (18) and effectively cut off all access from Court 3 to the storage rooms in the corner of the building. Entrance into the Main Building

1 Blegen and Rawson thought that the plaster platforms found in the palace – two sunk into the plaster floor (rooms 1 and 64) and the other two raised (rooms 4 and 5) – were mounted by sentries or doormen (PN I, 57, 68, 74–75, 253). There is no evidence for this suggestion.

1.3 Circulation in the main building.

285

Nelson – The Architecture of the Palace of Nestor may also have been gained from Room 12. Although a doorway is reconstructed here, based on Blegen and Rawson’s conjecture (PN I, 108–109), the bedding stone foundations of the building’s southwest facade continue uninterrupted along this stretch and suggest a solid wall. On the other side of the building, the northeast corridor leads to the nicely finished suite of rooms on the northeast flank. Despite their walls and floors being covered in smooth plaster, these rooms seem to have functioned as storage facilities (PN I, 152 (Room 30); 158 (Room 32)).

tower (55, 56 and 57; PN I, 221; see also Part I, pp.80–82 above for Cooper’s interpretation of this area.) In the east corner of the complex is a bathing room (43) and a private suite consisting of a large, lavishly decorated hall (46) with a central hearth and two smaller rooms (50 and 53) connected by a T-shaped corridor (48 and 49). In an earlier phase, the bathing room was associated with a major entrance into the Main Building, Room 41. The doorway in Room 41 still retains a column or pier base and a stone threshold (fig. 1.4).

In addition to the formality of the sequence of propylon, court and megaron, the numerous well-built doorways attest to a major concern by the occupants of the Main Building for controlling circulation. With few exceptions, every passage from one room to the next was equipped with a door, the evidence for which are the threshold and jamb blocks, and the pivot holes cut into the latter, in doorways 18–20, 20–88, 21–88 and 29–30 (fig. 1.3). Doorways 1–7 (PN I, 92–93) and 57–58 (PN I, 226) were found badly damaged, with no thresholds or jambs surviving. Obviously, individual doors controlled access into individual rooms, but on a broader scale and in conjunction with the formal arrangement of prominent rooms, a planned circulation pattern becomes clear (fig. 1.3). The vestibule of the megaron (5) and the two long flanking corridors form a principal circulation hub. To the southwest and northeast, the vestibule opens onto a long corridor and a staircase (14 to the southwest and 36 to the northeast) leading to the upper floor(s); to the northwest it opens to the Throne Room (6). Except for the staircases, opening and closing doors could control all access points to rooms on the ground floor.2 Indeed, Vestibule 5 would appear to have functioned solely as a circulation nucleus. (For further discussion of circulation patterns in the palace, see Part I, chapter 3.)

Finally, courts 42 and 47 abut the northeast facade (plans ii and xxii and fig. 1.5). These were late additions and effectively sealed off all ingress into the Main Building from the northeast. Proposals for the functions of these two spaces range from private courts for a king and queen (PN I, 208; Blegen and Rawson 1967, 20), to perfume manufacturing spaces (Shelmerdine 1984, 59– 62) to gardens (see Part I, pp.165–70). The first, rather romantic view of court life in the Mycenaean world has no supporting archaeological evidence, but the other two suggestions may have merit. There are hydraulics in Court 42 (see Part I, pp.139–40 above), and the floor in Court 47 has pebble-lined holes of varying depth. Five holes aligned in a row running southwest–northeast may have been postholes for a light partition wall (PN I, 208); the arrangement of the other holes reveals no apparent order. There is little evidence for gardens in Mycenaean palaces,

At the southern corner of the Main Building and flanking the propylon is the Archives Complex (rooms 7 and 8), where Blegen discovered the bulk of the Linear B tablets from the site. A doorway, reconstructed by Blegen and Rawson (PN I, 92–93), connects rooms 1 and 7 and is shown on the plans included in this volume. However, the wall separating the Archives Complex from the propylon had been destroyed and removed prior to discovery and excavation, and the original circulation pattern is unknown. It is possible that multiple entrances served this suite (Palaima and Wright 1985). Northwest of rooms 7 and 8 and forming the southwest side of Court 3 is another two-room suite (9–10), thought perhaps to be a reception room of sorts (PN I, 105). On the northeast side of the propylon and southeast of Court 3 is a confusing set of rooms that may have been the lower portions of a 2

No remains of doorjambs or a threshold were found in the wall separating rooms 5 and 13, but a built doorway is likely to have been located there, judging from the uniformity of construction and strong axial symmetry of the megaron unit. The doorway between rooms 5 and 35 preserves both doorjambs and a threshold (PN I, 72).

1.4 Doorway of Room 41 showing column or pier base and stone threshold.

286

The Palace at Englianos

1.5 Plan of courts 42 and 47, showing drainage holes.

or even for a Mycenaean affection for nature, such as one can detect in Minoan frescoes (M. Shaw 1993), but the suggestion is certainly tantalising. The juxtaposition of plants and flowers with the finely cut and coursed ashlar masonry of the northeast facade of the Main Building would no doubt have been a pleasing sight.

pots recovered from within and the orderly arrangement of postholes along the inner faces of its four walls, suggesting shelving (plan xvii; PN I, 237–41). In its present condition, Pantry 60 is a single-roomed structure, but in the middle of the exterior face of the eastern wall, another wall, represented by only a few stones, appears to have bonded. In addition, the east and west walls may have continued to the southwest, creating a second adjoining room (62).4

Clustered about the Main Building are four ancillary structures. To the north, situated on the edge of the plateau, is a small, two-room structure that served as a wine magazine (plans ii and xxii–xxv; PN I, 342–49). The door, on the north facade, opens into a small anteroom (104), which in turn leads to a rectangular storage room (105) fi lled with pithoi. Some of these line the walls while others run in two rows down the centre of the room.

Beyond Pantry 60 and further south from the Main Building lies Building X (plans xiv and xv). This may have been another tower-like structure similar to the three-room suite 55–57 in the Main Building (PN III, 40). The building stands just at the bottom of the southern portion of the hilltop, at the point where the ground level begins to rise steeply towards the plateau to the northeast. It is relatively small (c.7 x 3 m), about the size of Pantry 60, but built with extremely thick walls (1.40–1.70 m). In comparison, the walls of Room 6 are about 1.20 m thick, and the room itself measures approximately 13 x 11 m. The disparity between the thickness of its walls and the size of Pantry 60 argues for a solidly built structure such as a tower, which may have risen several storeys. Building X is earlier than the LH IIIB structures so far described, but it was probably, at least in part, still in use during the last phases of the palace’s existence (PN I, 283; PN III, 12–13, 40–41).

The Northeast Building (92–100) lies off the eastern corner of the Main Building (plans ii, xxii and xxvi). The building may have undergone two architectural phases (Westerburg 2001), but in its last phase a corridor (95) with doorways on both of its long walls forms a central spine and runs the entire length of the building. All rooms access the corridor except Room 93, which faces an open, plastered court to the southeast. Architecturally, Room 93 is perhaps the most elaborate in this building. At least one, if not more, cut limestone risers step up from the plaster court to the room’s open southeast facade.3 Large anta blocks terminate the southwest and northeast walls. Positioned in front and outside of the room is a plastered and frescoed limestone block partially sunk into the plastered Court 92. The excavators thought that the building served as a workshop, based on recovered artefacts (PN I, 303–05; see also Lupack 1996, 1997, 1999). However, a more recent study advocates a storage and administrative function for the building rather than its use for production activities (Bendall 2003).

Situated on the southwestern-most edge of the plateau is the Southwestern Building (rooms 63–81; plans ii, xiv and xviii–xx). Deep plowing by the former landowner levelled most of the walls down to one or two courses – in some places down to the small foundation or bedding stones. Fortunately, the massive southwestern facade, built of large limestone slabs, was preserved to a height of several courses by dense vegetation that once concealed

Near the south corner of the Main Building and separated from it by a plaster ramp (59) is Pantry 60 (plans ii, xiv and xvii). It received its name because of the quantity of

3

Blegen and Rawson suggested two or three steps (PN I, 304). Blegen and Rawson (PN I, Key Plan) reconstruct the second room as well as a possible third room.

4

287

Nelson – The Architecture of the Palace of Nestor and protected it.5 The focal point of the building is an impressive two-roomed hall (64 and 65). The monumental double-columned (possibly in antis) facade of Hall 64 serves as the only entrance to the building. Frescoes once covered its northwest and northeast walls, the latter of which once bore a violent battle scene (PN I, 249, fig. 197; PN II, 71–74, pl. M). Two doorways in Hall 64 access the remaining rooms. The wider doorway, in the southwestern wall, leads to Hall 65, the largest room in the building. Four column bases remain inside this room, though six are possible. The other, narrower doorway in the northwest wall of Hall 64 grants access to the western portion of the building (66–81). A stairway (69), with one step still in situ, once led to an upper storey.

of soil above the LH IIIB material, and an Archaic Lakonian roof tile is still embedded in the wall separating rooms 84 and 86. (For a detailed discussion of the stratigraphy and post-Bronze Age structures in this area see Part I, chapter 8.) In addition to the buildings just discussed, fragments of walls and buildings lie scattered about the site. These were never assigned area designations or room numbers by the original excavators; for ease of discussion, they are labelled herein (plan ii). The Northwest Area6 is behind, to the northwest, of the Main Building; here was built a series of grid-like walls and drains that run underneath and to the northwest of rooms 82, 83–86 and Circular Structure 87. The dates of these structures range from the Middle Helladic to the post-Bronze Age periods (see Part I, chapter 8 and pp.47–73).

Plaster courts 63 and 88 separate the Main Building and the Southwestern Building. A small structure consisting of rooms 89 and 90 was built between and incorporated the northeast facade of the Southwestern Building and southwest facade of the Main Building (plans xiv, xvii and xix). This dwelling, built in the Geometric period, reused the fallen debris from the LH IIIB palace (Griebel and Nelson 1998; see also Part I, p.229).

In previous MARWP publications, Area 103 has comprised the remains of several walls and structures between the Wine Magazine and the Northeast Building. Like the walls in the Northwest Area, these date to various periods; few seem to have been built in LH IIIB (Cooper, F. 1994). In this discussion, Area 103 is limited to the rectangular building containing four separate chambers, labelled a through d (PN I, 337–38, fig. 427). Area 106 comprises the fragmentary walls and other features, including three column bases, a plaster floor and a larnax, just to the south and east of the Wine Magazine; like those of Area 103, these constructions belong to various building periods. (For a discussion of phasing here, see Part I, pp.35–44.)

The Southwestern Building extends to the southeast beyond halls 64 and 65, but the walls on this side of the building are difficult to decipher. The original excavators did not assign numeric designations to this area, and as is evident from the published plan, it is fi lled with walls dating to different periods and built using various construction techniques. Nevertheless, sections of the southwestern facade can be traced for some distance (plan iii). Although fragmented, it is clear that this wall was built with several offsets before coming to an end at the southeast, at the point where the Main Drain cut through it. The entire facade was dated to the LH IIIB period (PN I, 277–81), but the variety of construction methods used in this wall, as well as in the entire Southwestern Building, indicate several different building periods.

Finally, returning to the southwestern portion of the hilltop, the area southeast of halls 64 and 65, southwest of Pantry 60 and northwest of Building X is identified herein as the Southwest Quadrant. This area, like the Northwest Area and Area 106, contains a maze of walls from various periods, but with few recognisable structures. One outstanding feature is a large gateway, mentioned in the original publication only in passing (PN III, 40). For the most part, only its foundations survive, but there also remains a plaster ramp, one ashlar tread from the flight of steps and several courses of ashlar wall pierced by the stairway (plans ii and xvi-xvii; Nelson and Cooper 1998).

Room 82 lies off the western corner of the Main Building (plans ii, xix and xxi). This is a single-roomed structure, larger than Pantry 60, with no apparent doorway. Its purpose is unknown, though it may have been an ancillary structure perhaps serving a function similar to that of the Wine Magazine or Pantry 60 (PN I, 289).

Beyond the palace complex proper, other Bronze Age structures survive on the hilltop. At its northeastern edge, another stairway – the Northeast Gateway – leads down from the plateau to the small valley that lies between the hilltop and Tholos IV (plans i, ii and xxiii). A flight of steps, reconstructed with 23 treads, pierces two flanking

Directly behind the Main Building, to the northwest, is a small circular structure (87), 3.10 m in diameter, with no obvious function (plans ii and xxi; PN I, 293). To its southwest lie the remains of a small building that consists of at least three and perhaps four or more rooms (83–86). The building’s slightly skewed alignment in comparison to the normal northwest–southeast orientation of the palace complex is similar to that of the Wine Magazine and probably reflects a respect for the dropping slope of the plateau in this area. Like rooms 89 and 90, rooms 83–87 post-date the Bronze Age. Their walls rest on a thin layer

5

Personal communication with Blegen’s excavation foreman, Dionysios Androutsakis, and gleaned from the Pylos Excavation Archive photographs now housed at the Blegen Library at the University of Cincinnati; see also PN I, pl. 205. 6 In the Blegen publication this area is referred to as ‘Exporation of the Stratigraphy Northwest of the Main Building’ (PN III, 43–47).

288

The Palace at Englianos walls similar to the gateway in the Southwest Quadrant. An ashlar block lies fallen near the northwest wall, perhaps indicating that the upper portions were ashlar masonry; the lower portions were built of flat fieldstones. Two large flat slabs form a broad tread near the midpoint of the flight and may have been a threshold for a door or gate.

beyond (plans xxx–xxxii; PN I, 332–36). For the most part, the aqueduct consists of two parts: a channel, which along its entire length is cut into the marly limestone bedrock, and flanking sidewalls. The construction of the lining walls is far from consistent: in some places, the channel is bordered on one side, in other places it is lined on both sides, and still in other places no lining was built at all. Furthermore, although in most sections the lining walls were generally built of rubble masonry, one short section employs ashlar blocks. The use of cover slabs was also inconsistent: some sections were covered, but others seemingly remained open to the sky.

The Belvedere Area lies at the southeastern corner of the plateau (plans ii and xxxiv; PN III, 20–23). Preserved here are a small kiln (figs 1.6–1.7; see also Part I, p.45, fig. 1.23), fragmentary walls, and vestiges of five drains.7 The walls and drains may be the remains of a small group of LH I domestic structures, all of which were built of rubble masonry except wall C, which was built in part of ashlar blocks. The Belvedere structures, indeed this general area of the hilltop, go out of use at the end of the LH IIIA period.

Lastly, just outside of the Southwestern Building on the slope of the plateau are remains of walls in an area identified by the original excavators as the Lower Town. This area was not extensively explored by MARWP; only the end of one rubble wall was exposed (plan xxvi); the Blegen trenches are plotted on plans i and ii. Many walls and floors came to light in the five trenches that were opened in this area by Blegen’s team. Most of the walls were built with rubble, but some used cut stone blocks interspersed among the rubble. No complete structure was excavated, so no definite plan of individual houses has been made, nor has the layout of the town been ascertained. The size of the town is also unknown, though survey work by the Pylos Regional Archaeological Project (PRAP) has gone a long way towards defining the extent of occupation. According to PRAP’s research, the settled area surrounding the Englianos plateau, including the Lower Town, increased steadily from around 5.5 hectares in the Middle Helladic period to a minimally estimated 12.4 hectares in the LH IIIB period (PRAP 1997a, 427–30). The published plans indicate that habitation concentrated on the western side of the plateau in the LH IIIB period (ibid., 429, fig. 12).

The so-called aqueduct snakes its way across the hilltop from Area 101 to the Northeast Gateway and perhaps 7 The drains are labelled L, G, A–B and W in PN III, fig. 307. Drain W was not observed by us and therefore does not appear in plan xxxiv.

The Published Chronology It is evident that there was much building activity of various phases on the Englianos hilltop over the long period of its occupation. Table 1.1 summarises the dates given to individual buildings by the original excavators

1.6 Kiln (top to bottom): plan (PN III, fig. 308); central tongue, south face; central tongue, north face.

1.7 Mudbricks of central tongue of kiln, north face.

289

Nelson – The Architecture of the Palace of Nestor and others who have studied the chronology of the site; the majority of the extant remains were built or were in use during the LH IIIB period. The debate concerning the date of the final, Bronze Age destruction of the palace complex has recently been revived. The latest proposal agrees with the excavators’ original conclusion – that the palace came to an end during the ceramic transition from LH IIIB to LH IIIC (Mountjoy 1997). Previously suggested dates for the destruction range from early to mid-LH IIIB (Warren 1975, 134; Popham 1991, 322) to the end of LH IIIB (Hope Simpson and Dickinson 1979, 127).

plowing; deep cuts in the upper surfaces of the ashlar blocks of the northwest wall of Room 32 attest to the devastating effects of the modern plow (PN I, 145). As a result, very little soil covered the remains – in some areas, none at all. The plowed surface soil contained a mixture of artifacts from all periods and was very shallow, ranging from 0.20 to 0.40 m in depth. Aside from this stratum and one or two identifiable post-Bronze Age layers (Griebel and Nelson 1998; see also Part I, pp.225–40 above), the rest was all Middle to Late Bronze Age. Because the surface soil was shallow, modern plowing had scraped the tops of the walls and confined the LH IIIB strata within rooms and above plaster floors. As has been noted (J. Wright 1984, 19), this stratigraphy and Blegen and Rawson’s method of publication contribute to the impression that Englianos is a single-period site dated to LH IIIB. In the room-by-room descriptions of the excavation, finds are arranged by types rather than stratigraphically, and the pottery catalogue contains only material considered to be LH IIIB, omitting earlier, later or foreign potsherds and vessels; these, where noted, are recorded in the individual room descriptions and appear unlabelled in the plates.

For the most part, extant plaster floors prevented Blegen and his team from exploring underneath many of the rooms and buildings, but they did sink exploratory trenches in some areas. In the Main Building, earlier floors and walls were discovered beneath the south and west corners, under Corridor 26 and below the conjectured floor levels in rooms 56, 57 and 58. A massive Middle Helladic wall runs underneath and perpendicular to Corridor 26,8 and a likely companion to this wall lies deep under the vestibule of the Wine Magazine (PN III, 32). Other earlier walls lie scattered about the site: under and near the western portion of the Southwestern Building, in the Southwest Quadrant, in areas 103 and 106, and in the Northwest Area (where some walls date to the Middle Helladic period (F. A. Cooper 1996; see also Part I, pp.55–56 above). For the early Late Helladic period, plenty of ceramic evidence survives, but little securely dated architecture. The Northeast Gateway seems to have been constructed in LH I and went out of use prior to LH IIIB (PN III, 6–7). Tholos IV, just to the northeast of the hilltop, was built sometime between the Middle Helladic period and LH I (PN III, 107).

The bulk of the LH IIIB strata is debris from collapsed walls and floors. Where digging was possible below LH IIIB floors, pre-LH IIIB strata were discovered, but they were only correlated in very general terms with earlier walls and features. Most of these earlier constructions were simply labelled ‘early’ or ‘earlier’; one of the few areas where earlier stratigraphy was examined in some detail was the Northwest Area (PN III, 43–47). Here, five building phases and seven individual strata, ranging in date from the Middle Helladic to the Late Helladic IIIB periods, were identified by Blegen (PN III, fig. 311), but the walls and the strata in which they lay were never correlated. (For Cooper’s analysis of the phasing of the Northwest Area see Part I, pp.47–73.)

Evidence for activity after the palace was destroyed is sparse, for like the earlier material it is represented by plenty of ceramic remains but few datable walls or structures. As noted, rooms 89 and 90 were built in the Geometric period. The later occupants of the hilltop used existing walls and fallen building debris to erect a small two-room structure here. Also later is the building formed by rooms 83–86, perhaps built sometime in the Archaic period (see Part I, pp.67–68).

Despite the lack of specific dates in the publications, the Englianos plateau clearly has a long history. The earliest surviving architecture belongs to the Middle Helladic period, though earlier occupation cannot be ruled out. The hilltop flourished in the Late Helladic period, when monumental building activity appeared early and continued in several stages. The palace complex reached its maximum size in LH IIIB, at which time it covered most of the southwestern half of the plateau. Throughout this history, building materials and methods changed and developed. The chronology of these changes, combined with the little stratigraphic evidence available, provide valuable evidence for the overall history of the site.

Dating of the pre- and post-LH IIIB phases is difficult because the excavated stratigraphy did not receive proper analysis or publication. For LH IIIB and later, the stratigraphy was relatively simple. By the time Blegen and Rawson sunk their first shovel into the earth, the hilltop had been subjected to more than 2500 years of wind and rain erosion, as well as 30 or so years of mechanical 8

In the stratigraphy of this pit, the lowest levels are recorded as containing ‘almost exclusively fragments of the Middle Helladic wares’ (PN III, 33). MARWP cleared the pit in 1996; the original excavation did not extend any deeper than the bottom of the massive wall, which would suggest that if Blegen’s interpretation of the stratigraphy is correct, the wall was standing in the Middle Helladic period.

290

The Palace at Englianos Building or Area 9

MH

LH I

LH II

Main Building Wine Magazine

LH IIIA

LH IIIB

post-LH IIIB

◊10

☐

◊11

◊12

☐13

Northeast Building

☐14

Pantry 62

☐

Building X

◊15

☐

Southwest Building

◊16

☐

Room 82

☐

☐

Northwest Area17

◊

◊

◊

◊

◊

◊

Area 10318

◊

◊

◊

◊

◊

Area 10619

◊

◊

◊

◊

◊21

◊

Southwest Quadrant 20 Northeast Gate23

◊

◊

◊

Belvedere Area 24

◊

◊

◊

◊22

☐25

Aqueduct Lower Town 26

◊

◊

Table 1.1 Published dates for the buildings at Englianos. 9

18

Squares indicate main periods of use; diamonds indicate walls and features in an area or building that were built or in use in that particular time period. 10 Ashlar wall beneath Room 7 (PN I, 94). 11 Dark Age activity in rooms 89–90 (Griebel and Nelson 1998; see also Part I, p. 229). 12 Large rubble wall beneath room 104 (PN III, 32). 13 Late LH IIIB addition (Shelmerdine 1987, 563). 14 Late LH IIIB addition (Shelmerdine 1987, 563–64). 15 Originally built in LH IIIA, continued in use through LH IIIB (PN I, 283; PN III, 12–13, 40–41). 16 Blegen and Rawson (PN I, 34–35, 423) suggested that the Southwest Building may have been the first palace on the hilltop. 17 Seven strata were uncovered in the Northwest Area, ranging in date from MH to LH IIIB (PN III, 43). Rooms 83–86 and 87 have now been determined by MARWP to post-date the Bronze Age.

F. A. Cooper (1994) traced seven phases of wall construction in this area. See now his more recent analysis in Part I, pp. 35–44. 19 Ibid. 20 PN I, 281–83, and Key Plan. 21 Walls southwest of and beneath Court 63 (PN I, 282). 22 MARWP identified the following post-LH IIIB features: an isolated construction built of re-used ashlar, which sits above the LH IIIB strata in a way similar to the fallen blocks of the northeast facade of the Main Building); the Geometric walls of rooms 89 and 90; and the even later walls of rooms 83–87. 23 PN III, 6–7. 24 PN III, 20–23. 25 No specific date for the aqueduct was published, though the material recovered ranges in date from MH to LH IIIB (including one Linear B tablet fragment; PN I, 332–35). 26 PN III, 52–62.

291

2 BUILDING MATERIALS

Structural walls at Englianos were built primarily of stone (both worked and unworked), mudbrick and wood. Terracotta was also used, as a protective lining for the kiln and for drainage systems. Terracotta was not part of any of the building systems covered here and is discussed in Appendix A (pp.369–71 below).

Another type of stone, represented by only one example at the site, is red in colour, very hard, and granite-like in appearance. It is a rough slab with no signs of stone working, which covers a short stretch of section A of the aqueduct (plan xxviii). Poros Limestone

Stone

Poros limestone, common in Late Bronze Age Messenian architecture,2 was the principal worked stone used in the construction of the palace. Its relatively soft composition makes it easy to shape and work with bronze tools, a convenience that may account for its popularity among Late Bronze Age builders. Ashlar walls, door thresholds and jambs, anta blocks, column bases, stair treads and drain tiles were cut from this material.

Rubble and Fieldstone Of the palace’s extant walls, more than 94 per cent (based on the total linear length of all features recognised as walls) were constructed with rubble as the primary building component. The rubble is composed of fieldstones of several grades of limestone, including poros. In the early phases of building at the palace, the rubble was probably gathered from nearby locations rather than quarried. Later, rubble masonry walls were constructed of both gathered fieldstones and stones re-used from previous structures. Stone size varies considerably, with the maximum dimension ranging from c.0.05 to 0.65 m. A few larger stones were occasionally incorporated, but never in a consistent fashion that would indicate a specific structural or aesthetic function.

The present colour of the poros blocks and block fragments at the palace depends on their state of preservation. Those beneath the protective shed or covered by backfi ll are much better preserved, retaining a colour that varies from Moderate Yellowish Brown (10YR5/4) to Pale Yellowish Orange (10YR8/6) on the Munsell scale. Readings taken on various blocks throughout the palace include 10YR5/4 (Moderate Yellowish Brown), 10YR6/6 (Dark Yellowish Orange), 10YR7/4 (Greyish Orange) and 10YR8/6 (Pale Yellowish Orange). Blocks or fragments lying beyond the confines of the protective shed that covers the Main Building have endured more than 30 years of wind, rain

Conglomerate was used sparingly and only in rubble masonry walls.1 In the work conducted by MARWP at the palace, only three examples were noted, and they do not distinguish themselves as either part of a particular building method or a specially chosen material. None of the conglomerate pieces were worked, and unlike the limestone fieldstones, they must have been carried to the site from farther afield. One of the nearest outcroppings of conglomerate is on the north side of the modern town of Chora, which is about 3 km north of Englianos on the modern road.

1

Conglomerate was used much more frequently at other Mycenaean sites, particularly at Mycenae and Tiryns. Küpper (1996, 3–6) briefly surveyed building stones in Mycenaean architecture; see also Dörpfeld 1885, 253–54, and Wace 1949, 135–36. 2 Poros limestone was worked for the tombs near the palace and further afield (Tragana, Peristeria, Nichoria and Malthi) and reused in the last phase of the Menelaion (Barber 1992, 11–13); the surviving wall blocks at Kakovatos were also cut of this material. Poros was also commonly used in Minoan architecture.

293

Nelson – The Architecture of the Palace of Nestor and sun exposure since they were uncovered, in addition to any ancient exposure. As a result, they have darkened considerably to 5Y5/2 (Light Olive Grey). Efflorescence has also left behind a white powdery crust on some blocks and block fragments.

which pottery was placed for firing, extends about 1.20 metres from the kiln’s south–southwest wall. It was built with three courses of mudbrick, the topmost of which was damaged by deep plowing; the two lower courses remain intact. Each course measures 0.10 m in height. The other dimensions of the bricks, where they could be determined, vary considerably, from 0.18 x 0.28 to 0.48 x 0.30 m.3 The reasonably consistent widths of the bricks (0.28–0.30 m) is reflected in that of the tongue, which is one brick wide.

Beginning in 1992, investigations were carried out by MARWP into the possible origins of the stone used in the construction of the palace (see Part I, chapter 5). These studies identified ancient quarries near the modern town of Gargaliani and around 5–8 km from Englianos as being a highly likely source of the poros found at the palace.

Mudbrick fragments recovered by MARWP are rustcoloured (5YR5/6, Yellowish Red) and contain inclusions of very small, round stones (granule size 0.002–0.004 m) and very small voids left by rotted or burned out vegetable matter. One fragment preserves a height of 0.096 m, which compares nicely with the bricks in the kiln. From these few examples, it may be postulated that a height of c.0.10 m was a standard dimension for bricks at Englianos. Mudbricks recovered at other mainland Late Helladic sites have similar heights. Those from Eleusis, Zygouries, Tiryns, and the House of Sphinxes and the Panagia area at Mycenae range in height from 0.08 to 0.09 m (Dörpfeld 1885, 256–63; Shear 1968, 432, 484). Larger bricks, 0.12–0.13 m high, came from rooms XVII and XXX at Tiryns, but the smaller dimension, just below 0.10 m, is more frequent, even for bricks from earlier periods (Shear 1968, 484). More is known about Minoan mudbricks, which also generally have heights near 0.10 m, though they can be up to 0.14 m in height (Shaw 1973a, 187–98, 231–34). Brick lengths and widths vary for all of the Aegean samples, though within individual Minoan sites or buildings brick sizes tend to be standardised (ibid., 198); but the similarity in height suggests a common unit that was shared, knowingly or not, by both mainland and Cretan builders.

Other Types of Limestone Two other types of limestone were used at the palace, but neither one as frequently as poros. Both are much harder and finer in grain than poros, and their colours differ. The first is light in colour, Moderate Yellowish Brown (10YR5/4), and was worked for column bases, door thresholds and jamb bases. The upper surfaces of the bases and blocks were dressed semi-smoothly, since they were either exposed to view or were intended to receive wooden members such as columns and doorjambs. The other surfaces, where they can be seen, were hammer-dressed. Door thresholds were smoothly finished, and some were covered with plaster (PN I, 71, 76). Another type of limestone was used in the large slabs of sections 1, 2 and 3 in the southwestern facade of the Southwestern Building, and in the small dado slabs on the walls of Room 4. The stone is similar in grain and hardness to the bases, thresholds and jamb blocks (particle size: fine, 0.00025–0.000125 m), but darker in colour, ranging from Dark Reddish Brown (10R4/6) to Light Olive Grey (5Y5/2). Hardness and durability must have determined its selection, for the southwestern facade functioned as a massive retaining wall, and the small dado slabs served a decorative purpose requiring a hard, smooth face (see Nelson 2001, 56–58, for further discussion of the dado slabs).

Wood The wood used in the palace for columns, roof and ceiling structures, doors and doorjambs is now missing. Only a few fragments of wooden architectural members (including charcoal and ash) survived the Bronze Age, and even fewer were found in situ (PN I, 71, 77, 103, 117, 125, 151, 156, 173, 209, 319, 343). A burned and decayed column fragment was recovered still standing within the plaster column ring in Room 2. Other fragments came mainly from doorways, including an unburned portion of a wooden doorjamb in the doorway between rooms 46 and 48 (PN I, 209).

Mudbrick Numerous fragments of mudbrick recovered in the original excavations attest to its frequent use as a building component (PN I, 36–37; MARWP recovered more pieces during its re-excavations). One complete brick (0.52 x 0.38 x 0.09 m) was discovered in Room 32, where it was part of a small construction that apparently supported a large jar (PN I, 36).

Among the fallen debris were discovered a few architectural members with partially preserved dimensions. From Room 44 came three pieces that were probably structural beams used to support the ceiling and floor above. Their finished lengths cannot be measured (preserved lengths are c.0.75, 1.90 and 0.70 m), but the cross-sections were

Examples were preserved in situ only in the lower parts of the small kiln located in the Belvedere Area (figs 1.6–1.7 above; plans ii and xxxiv; PN I, 19, figs 44 and 308). This area of the hilltop has very little surface soil, and the kiln’s superstructure is entirely absent. Its substructure consists of a somewhat irregular oval pit (the fire chamber) sunk into bedrock and lined with terracotta plaques. A central tongue, which probably supported a perforated rack upon

3

The mudbrick dimensions were originally published as 0.52–0.54 m long by 0.30 m wide by 0.10 m high (PN III, 19).

294

Building Materials Present-day tree coverage in the Pylos region is dominated by the olive, and large trees of other species are lacking, especially those suitable for timber construction; aside from the few stands of cypress in the region, one would be hard-pressed to find a tree from which could be cut a timber member with a cross-section comparable to those recovered from palace debris. Oak species (mainly smaller evergreen oaks) are still present, but pine is not (Zangger et al. 1997, 578–79). Even the larger oak trees currently found in the area would only be able to provide small architectural timbers. From the surviving evidence, the palace builders seem to have structurally over-compensated with wooden spanning members, the size of which can be calculated from wall-to-wall or wall-to-column distances in the palace. Assuming that the builders spanned the shortest distance between two parallel walls, or between wall and column, the longest ceiling span in the LH IIIB palace is 6.23 m, over Room 46. According to modern heavy timber-frame construction guidelines (Allen 1985, 101, fig. 4.7, and 109–10), six metres is usually considered the maximum span for a single beam;

extant and measured c.0.13–0.14 m square, 0.27 x 0.35 m and 0.13 x 0.26 m (PN I, 192). A number of wood fragments recovered in the original excavations were analysed by Peter Kuniholm at the Malcolm and Carolyn Wiener Laboratory for Aegean and Near Eastern Dendrochronology, Cornell University. Fourteen of the samples collected were identified (table 2.1), resulting in some differences between the assessments of Blegen and Rawson and those of the Weiner Laboratory. About half of the samples belong to the oak genus, the rest to the pine genus. The Late Bronze Age tree coverage, according to recent studies, varied over time and was, of course, different from today. During the period 1600–1400 bc, pine is absent in the pollen record, and the presence of oak drops significantly (Zangger et al. 1997, 576–95, esp. 589; see also Zangger et al. 1998, 5–6). During the later years of the Late Bronze Age, pine remains relatively infrequent in the pollen count, but the presence of oak increases. Blegen Ref.

MARwp Ref.

Dendrochronology Laboratory Ref.

Blegen Description

WC 2

147

PYL-2

Burned wood

Charred wood

Room 2, southwest

Quercus

WC 20

165

PYL-20

Fragments of carbonised wood

Burned wood

Room 44

Quercus

WC 25

170

PYL-25

Wood fragment G, carbonised

Burned wood, almost charcoal

Room 44

Coniferous

WC 26

171

PYL-26

Carbonised wood

Burned wood, almost charcoal

Room 4

Coniferous

WC 36

181

PYL-36

Burned wood

Burned wood w/ some earth matrix

Area between court 42 and Wine Magazine

Quercus

WC 45

190

PYL-45

Fragments of carbonised wood

Burned wood, almost charcoal

Doorway between rooms 64 and 66

Quercus

WC 51

196

PYL-51

Burned wood

Wood charcoal

?

Pinus / Coniferous

WC 52

197

PYL-52

Carbonised wood

Burned wood

Room 3, near room 4

Coniferous

WC 58

203

PYL-58

Wood

Charcoal

?

Coniferous

WC 59

204

PYL-59

Charcoal

Burnt wood

Room 61(?)

Coniferous

WC 62

207

PYL-62

Burnt wood

Burnt wood

Room 63 or 64

Quercus

WC 63

208

PYL-63

[illegible]

Decaying wood fragments in earth

Room 6

Quercus

WC 65

210

PYL-65

Charcoal

Charcoal

Room 3, near room 4

Coniferous

WC 66

211

PYL-66

Carbonised wood

Charcoal

?

Coniferous

Table 2.1 Analysed wood samples from the palace at Englianos.

295

Archaeometry Lab. Description

Discovery Location

Genus

Nelson – The Architecture of the Palace of Nestor the palace builders were therefore pushing the limits of structural safety in Room 46. This span would have required a wooden beam with a minimum cross-section of 0.23 x 0.39 m according to modern guidelines, and one of the beam fragments recovered in Room 44 does, in fact, have similar dimensions: 0.27 x 0.35 m. The fragments recovered in Room 44 were probably used in the ceiling span for that room. The method by which this room was spanned is not readily apparent, and two beam arrangements are possible: column-to-column or column-to-wall with a short cantilever. If a beam were laid over the columns and the space was therefore spanned in a northwest–southeast direction, then one end of the beam rested on the lintel of the door opening into Stair 54 and the other on the lintel of the door between rooms 37 and 44. We might expect this beam placement to be avoided because of the large load placed upon the lintels. The second arrangement places a beam over each column with one end fi xed in the northeast wall of the stoa and the other end extending slightly over and beyond the column. A secondary cross-beam, running northwest-southeast, would then rest on the two beams just mentioned and fi x each end into the northeast anta of Room 4 and the southwest anta of Stair 54. This alternative was favoured by Blegen and Rawson, who based their reconstruction on the location of recovered fragments of wooden beams among the fi ll (PN I, 192). In the latter beam configuration, the maximum span length is c.2.81 m, measured from either of the columns to the northeast wall of Room 44. A beam with a minimum cross-section of 0.09 x 0.18 m would have adequately carried floor or roof loads over this distance. The recovered fragments, therefore, indicate beams greatly oversized for the structural requirements of the span, a condition of which the builders would surely have been aware. This is the only surviving evidence for measurable wooden beams, but it may indicate that the palace builders had little trouble acquiring the timber they needed, even if it was not available locally.

296

3 BUILDING METHODS

too deep to be reached.2 Stepped foundations, which incorporated both the bedrock and built foundations, were used on the fringes of the hilltop where the topography slopes downward.

The methods with which the materials discussed in Chapter 2 were worked and with which they were combined to create load-bearing entities (i.e. walls) are set forth here in a typology of building systems. It is possible that some of the walls at the palace were non-load bearing and simply supported their own weight, for example the thin walls enclosing rooms 50 and 53 of the Main Building (plans vi and ix). Little of these walls survive, and they are very thin in comparison to the other walls of the palace. They may have been partitions, but their method of construction is now lost.

Klaus Kilian, in his study of Tiryns, has completed the most extensive examination of Mycenaean foundations to date, and distinguished two types: none, or where walls were built directly on earth with no change in building technique or material within the walls themselves; and built foundations employing various, and sometimes complicated, construction methods (Kilian 1990; see also Küpper 1996, 52–53). The same division applies to the Englianos palace, though its built foundations are not quite as large as those surveyed by Kilian at Tiryns and used at other Mycenaean sites, especially in comparison to the foundations of fortification walls and terraces at those sites.

Foundations Foundations are structural members that transfer load from the walls built upon them to the ground beneath. In the Bronze Age they were often invisible because they were covered by the structures they supported or the earth into which they were sunk, or both. The same applies to column bases, which at Englianos are essentially foundations.

Bedrock Foundations and Bedding Stones The working of bedrock in preparation for construction was not a common practice at Englianos, despite the relatively soft nature of the stone. Trimmed bedrock was found in only a few places: beneath a thick rubble wall on the northwest edge of the hilltop (plan xxi, wall B; PN III, 8) and beneath the wall flanking the northwest side of the Northeast Gate (ibid., 6). Cuttings, but no walls, were found in the bedrock at the bottom of an exploratory trench laid out on the northeast side of the hilltop (ibid., 65).

The excellently preserved plaster floors of the Main Building conceal most of the built features below them. The ancillary buildings and outlying structures provide more information about foundation construction, though still relatively little. It appears that the palace builders took great care to ensure stable and secure foundations for their walls and buildings.1 The type of foundation employed depended largely on the topography of the plateau. In areas more or less level, the builders placed walls directly on the marly limestone bedrock (Zangger et al. 1997, 604– 05 and 635). Under a few walls, the bedrock was trimmed flat; under others, bedding stones were used to even out the bedrock, levelling it in preparation for receiving a wall or adjusting the level of the first course of blocks or stones. Later structures, which were laid over older walls and buildings, required built rubble foundations, occasionally including worked stones, because the bedrock was often

1

Walls that predate LH I and post-date LH IIIB are not considered here; in general, pre-LH I walls appear to rest not on built foundations but on bedrock. 2 Shear (1968, 437–38) came to the same conclusion regarding the foundations of Mycenaean houses and the land upon which they sit, though she did not make architectural distinctions between foundations built on bedrock and those built on fi ll.

297

Nelson – The Architecture of the Palace of Nestor At Mycenaean sites elsewhere, there is more evidence for the working of bedrock, usually to provide either a level footing for walls or a level platform for both walls and floors (Shear 1968, 435–36, note 801). The bedrock was trimmed flat for the north portion of the floor of the megaron at Mycenae (Wace 1949, 77), and at the House of Kadmos in Thebes it was levelled by both trimming the bedrock and fi lling in shallow holes and depressions within the bedrock with clay (Symeonoglou 1985, 43–44).

Both the early wall built of limestone slabs that runs underneath and perpendicular to the southeast wall of Room 65 (plans xiv and xviii) and sections 2 and 3 of the southwest facade of the Southwestern Building (plan iii) rest directly on bedrock. No bedding stones were used or needed beneath these walls. The nature of their construction, which is similar to rubble construction, did not require a level surface upon which to build, so the bedrock was left uneven under both walls.

Bedrock foundations supplemented with bedding stones (usually no larger than cobbles, 0.06–0.25 m) support the northeast ashlar facade of the Main Building (plan III, wall 11L; PN I, 48). Here, a shallow, intermittent course of small flat stones lay directly on the unworked bedrock, presumably to level it where necessary. Some ashlar blocks or portions of blocks rest directly on the bedrock, while beneath others stones protrude visibly beyond the face of the wall (PN I, figs 30 and 31).

Built Foundations New walls erected over earlier structures required built foundations. Substantial foundations, c.1.20 m deep and consisting of large, roughly squared limestone blocks and rubble, support the northwest ashlar wall (wall 10C; see plan iii) of Room 32 (fig. 3.1 and plan x). Coursing is crude: the topmost course is formed largely of squarish blocks with smaller unworked stones, which together create a level surface for the ashlar wall laid upon them. Near

3.1 Plan and elevation, wall 10C, Main Building (detail).

298

Building Methods

3.2 Elevation, southwest facade of Southwest Building (wall 17L), sections 1, 2 and 3.

the wall’s southwest end, the Late Bronze Age foundation straddles a Middle Helladic wall that runs perpendicular to wall 10C; the top of the earlier wall lies approximately 0.25 m below wall 10C, and it rests on bedrock (PN III, 32). The builders of wall 10C may not have known this last fact, but they nonetheless deemed the strength of the earlier wall sufficient for the new construction. Excavations were not carried out below the floor of Corridor 26 near its northeast end, but built foundations for wall 10C probably continued to the northeast of the earlier wall and were sunk to the same level.

Stepped Foundations Stepped foundations employed both bedrock and rubble to facilitate the construction of walls on sloping topography. They were built as a series of steps to keep walls roughly level over short lengths. Stepped foundations were built most frequently at the sloping southwestern edge of the plateau, particularly beneath the Southwestern Building. For a little less than half of its length, section 1 of this facade rests on bedrock (fig. 3.2). Near its northwest end, where it forms the western corner of Room 81 (plan xx), stepped foundations provide support. The ground at this corner rises towards the northwest and northeast (plan i). Large stones, varying in size from 0.25 to 0.60 m and laid on bedrock, create short, level steps that rise with the topography. Since the stones are partially concealed, full measurements could not be taken, but they appear unworked except for their upper surfaces; these may have been hammered relatively flat.

Built foundations consisting of several courses of rubble masonry support the four exterior walls of the Wine Magazine. They are slightly thicker than the walls they support, and vary in depth from c.0.50 to 0.75 m (PN I, 342); they do not sit on bedrock. A substantial Middle Helladic wall, no doubt a section of the earlier wall running under wall 10C, underlies the Wine Magazine (plan xxiii, Room 104); it rests on bedrock and runs parallel to, and approximately between, the two long walls of the magazine. The foundations of the magazine’s southwest wall straddle this earlier wall and utilise its strength in the same way as the foundations beneath wall 10C.

At the return of the facade, where it becomes the northwest wall of Room 81 (SW39 on plan iii; see also plan xx), the ground rises more steeply (c.0.85 m over a distance of 7 m), and the wall continues to step up towards the northeast (fig. 3.3). Three steps can be distinguished at c.1.00, 4.10 and 5.95 m from the southwest end of the facade; these correspond to rises of c.0.48, 0.28 and 0.10 m, respectively. The steps are partially cut into the bedrock and partially built up with a few foundation stones. The stones are not laid at every step, nor are they uniformly distributed beneath the course of the wall but instead are set to fill out a step or maintain its level. Foundation stones also cluster beneath the facade corners, which suggests that the builders were concerned about the stress and load inflicted at these points. The facade returns again to the southeast to form the short, northeast exterior wall of

Another example of built foundations lies beneath the Southwestern Building. A portion of this building sits on the sloping southwest side of the plateau, where many walls, including sections of the massive southwestern facade, required support. Sections 5, 6 and 10 (plan iii) all sit on rubble-built foundations. Three, perhaps even four, courses of rubble laid on bedrock provide support for section 10. The foundations beneath sections 5 and 6 were also well built and include some roughly squared blocks. At the intersection of these two sections, six courses of rubble stand to a height of 0.88 m above the bedrock. The wall that the foundation supported is entirely missing except for one ashlar block that remains in situ at the northeast end of section 6.3

3

Sections 5 and 6 were never fully explored by Blegen and Rawson. They were careful to point out the worked blocks in the upper course of the foundations, but they did not discuss the in situ wall block (PN I, 279, fig. 204).

3.3 Elevation, northwest facade of Southwestern Building (wall SW39).

299

Nelson – The Architecture of the Palace of Nestor

3.4 Elevation, southwest facade of Southwest Building (wall SW52).

Room 81. The single course of foundation stones returns and runs beneath the entire length of this short facade.

are slightly wider (c.0.20 m) than the wall itself (partially shown in plan xvii, on the inside corner of the intersection of wall SW52 and the southwestern facade). Except for its width, the foundation’s construction is similar to that of the wall it supports.

Similar stepping occurs beneath the sturdy southeast wall (SW52 in plan III) of Room 65 (fig. 3.4 and plan xviii). The underlying bedrock in the general area of Room 65, as far as can be determined, is unworked and slopes upwards towards the Main Building. Three broad steps were carved into the bedrock beneath wall SW52, on which foundations were laid. At the wall’s very northeast end there is presently only one course of bedding stones resting on the bedrock. At the southwest end of the wall the bedrock drops about 1.4 m, and more substantial rubble foundations were built. The foundations run four to six courses deep and

Stepped foundations also support section 9 (plan iii) of the Southwestern Building’s southwestern facade, for the slope of the ground in this area necessitates such construction (figs 3.5–3.6).4 Stepping occurs both in the built foundations and the ashlar courses. In the latter, the first course contains two blocks, and the second, four. Rubble foundations lie beneath the first course and beneath the two northeasternmost blocks of the second course. The

3.5 Elevation, southwest facade of Southwestern Building (wall 17L), section 9.

3.7 Elevation, southwest facade of Southwestern Building (wall 17L), section 10.

3.6 Southwest facade of Southwestern Building (wall 17L), section 9, from the southeast.

3.8 Southwest facade of Southwestern Building (wall 17L), section 10, from the southeast.

300

Building Methods rubble foundation turns the corner and supports section 10 (figs 3.7–3.8), under which are four courses of rubble laid on bedrock. It is not possible to determine the depth of the rubble foundations underneath section 9, but the massiveness of this wall suggests that they probably run deep or sit on bedrock, like those beneath section 10.

foundations, or clay beds with small levelling stones (Kilian 1990, 110–11; see also Küpper 1996, 95). Of the 20 remaining column bases at Englianos (plan V and Table 3.1), 14 are concealed by plaster floors, which prevent close examination. The remaining six fall into two types based on construction method: single block bases and platform-built bases – categories that essentially match those established by Kilian (though instead of clay beds, levelling stones were used in conjunction with bedrock). All the bases fall under the definition of foundations because they were not meant to be seen when construction of the palace was completed.

Column Bases and Column Foundations In his study of foundation construction at Mycenae and Tiryns, Klaus Kilian distinguished two types of column foundation: cut stone bases supported on rubble

For the single block bases, only the upper surface is finished, either worked entirely flat or trimmed at its edges. The latter method resulted in a raised circular disk, which was presumably slightly larger in diameter than the column it supported. Column bases c58SW and c65N, are

4 Th is wall, among others, has shifted and heaved since the Bronze Age. The ashlar courses were certainly level in the wall’s original condition. 5 A question mark in Table 3.1 indicates that the plaster floor conceals details of the column base; ‘n/a’ indicates that the platform does not contain a cut stone base.

Column

Foundation type

Raised foot

Raised foot diameter (m)

Platform diameter (m)

Column diameter (m)

Elevation (masl)

PN reference

1

Base (?)

yes

?

0.43–0.47

192.14

I, 56–57

2

Base (?)

yes

?

0.47

192.34

I, 61

4NE

Base (?)

yes

0.68

0.48

192.29

I, 67–68

4SW

Base (?)

yes

0.70

0.49

192.41

I, 67–68

6E

Base (?)

?

?

0.51–0.52

192.43

I, 80–81

6N

Base (?)

?

?

0.51–0.52

192.37

I, 80–81

6S

Base (?)

?

?

0.51–0.52

192.39

I, 80–81

6W

Base (?)

?

?

0.51–0.52

192.43

I, 80–81

41

Pillar base

yes

n/a

192.55

I, 179

44NW

Base (?)

yes

?

0.40–0.41

192.49

I, 192

44SE

Base (?)

yes

?

0.41–0.42

192.51

I, 191–192

58

Base

no

0.64

0.64–0.66

191.38

I, 228; III, 40

64NE

Base (?)

?

< 0.64

0.50

192.35

I, 250–251

64NW

Base (?)

?

< 0.52

0.50

192.32

I, 251–252

64SW

Base (?)

yes

< 0.64

0.50

192.34

I, 250–251

65E

Platform

n/a

n/a

?

192.15

I, 256–257

65N

Base

yes

n/a

0.52

192.31

I, 256

65S

Platform

n/a

n/a

1.360–1.520

?

191.89

I, 256–257

65W

Platform

n/a

n/a

1.220–1.458

?

191.85

I, 256–257

106SW

Composite

no

< 0.47

1.253–1.490

0.45–0.47

192.61

III, 30

106NE

Base

no

< 0.48

0.46–0.48

192.71

III, 30, fig. 62

NW.Cb.A

Base

no

0.58

192.02

III, fig. 311

NW.Cb.B

Base

no

0.62

191.67

III, fig. 311

1.336–1.585

Table 3.1 Column base dimensions.5

301

Nelson – The Architecture of the Palace of Nestor single, roughly disk-shaped blocks of worked limestone. The upper surface of c58SW is finished flat, whereas its sides are only hammer-dressed. Base c65N is cut with a raised circular surface on its upper face to receive the foot of the actual column.6 Its sides appear hammer-dressed. Both c58SW and c65N rest on a bedding of cobbles (see plans xv and xviii respectively). Like the bedding beneath the northeast facade of the Main Building, the cobbles do not form a solid and complete layer, and portions of the bases rest directly on bedrock. Some of the cobbles must have been used for levelling purposes and were wedged underneath the edges of the bases once they were set in place. A similar practice is reported for some of the column bases at Tiryns, though there the bedding layer contains earth and clay (Küpper 1996, 95).

3.9 Plaster column ring, Main Building, Room 1.

Single column bases with raised circular disks were used frequently at Mycenae and more so at Tiryns (ibid., 94–104); raised disks are also common in Minoan column bases (J. W. Shaw 1973a, 111–21). Unlike the bases at Englianos, the raised, disk-like surfaces at these sites sometimes remained visible in the finished state of construction; among others, the four column bases of the large megaron at Tiryns all have visible dressed and raised circular surfaces (Küpper 1996, 96). Küpper argues both a practical and aesthetic function for the raised areas (ibid., 96–98; see also J. W. Shaw 1973a, 115). By lifting the column above the floor, the raised surface keeps the foot of the column dry and thus prevents rot; at the same time it visually enhances the separation between column base and floor. These suppositions do not apply at Englianos because the raised surfaces of the bases were completely covered, either by a plaster floor (columns 6W, 6N, 6S, 6E, 64NW, 64SW and 64NE) or a plaster ring (4SW, 4NE, 44NW, 44SE, 2, and 1) as well as by the column itself. The bases could not have functioned to shed water or create a visual demarcation of any sort. Instead, Küpper suggests that the column rings (e.g. fig. 3.9) around many of the Englianos columns compensated for the lack of a raised surface projecting above the floor by sealing the foot of the column from moisture. They also imitated an actual column base. However, since column base 64NW in Room 64, with its column ring (for the most part destroyed; fig. 3.10), is approximately centred in the middle of the room and thus protected from the elements, it is difficult to argue a weather-proofing function for the stucco ring. For the palace’s columns then, the function of the raised and dressed surface must have been to provide a flat and even surface for the column foot and perhaps simply to save stone-cutting time.

3.10 Column base 64NW, Southwestern Building, Room 64.

columns portrayed in the fresco from the Room of the Frescoes in the Citadel House area (Immerwahr 1990, pl. 60); the column within a window depicted in the Falling Warrior scene from the megaron (ibid., 65); and the engaged columns of the Treasury of Atreus and the Tomb of Clytemnestra (Wace 1949, figs 5, 6, 49, 51). However, at Englianos the column bases of Room 6, the Throne Room, are sunk c.0.07–0.12 m below the plaster floor and have no surrounding plaster ring; they would have originally appeared to rest directly on the floor.7 Its position within the building, its architecture and its fresco decoration indicate that this room was very important, and one would expect some articulation of column and floor if indeed such an aesthetic was significant to the palace occupants. While Küpper’s argument seems logical, the palace columns do not provide convincing proof. The other four column bases (c65E, c65W, c65S and c106SW) are circular rubble platforms. Column base 6

Blegen and Rawson (PN I, 256) reported that a circle of rubble surrounded base c65N. When cleaned by MARWP, there were no stones around the base. 7 Blegen and Rawson (PN I, 80) thought that the column bases were probably level with the floor when the columns were originally erected and the fi rst plaster floor laid. Successive coats of plaster, laid for remodeling or repair, continually raised the floor level, which covered more and more of the column shaft s.

The aesthetic value of a ring or base – to define clearly and to sharpen the division between column and floor – may have been quite significant in Mycenaean architecture, since columns depicted in Mycenaean art often have this distinguishing feature. The most prominent example is the column and base depicted in the Lion Gate at Mycenae (Wace 1949, pl. 73), but others include the two

302

Building Methods bedrock lay far below the intended column. The platforms vary in diameter from 1.22 to 1.60 m; the diameter of the columns they originally supported can be estimated from the diameter of the raised circular area on base c65N in the same room. Its diameter is c.0.52 m, which should make the diameter of the actual column slightly less. Assuming that all of the columns in this room were approximately the same size, the platforms are more than double that diameter.

Block Masonry Construction Poros limestone blocks were used in a variety of building systems at Englianos: coursed ashlar masonry, ashlar shell walls, orthostate construction and pseudo-ashlar masonry. With the exception of pseudo-ashlar, these systems and the terms used to describe them are not unfamiliar in any discussion of Mycenaean architecture, nor in regard to the architecture of other Bronze Age Mediterranean cultures. Nevertheless, the words used to describe building systems and their material components vary in the literature; I therefore define below the terms as they are used in the present discussion. My definitions depend as much on the cross-sections of walls as on their outer appearance; that is, I take into account not only the outer, exposed face of a wall but also its interior composition. Whether a wall is a retaining wall or is part of a larger structure (called herein ‘freestanding’, since both faces of the wall are exposed) is also a factor in the definition of its system of construction, since each satisfies different structural requirements.

3.11 Column base 106SW.

c106SW is unique among these four because it contains a cut circular base (fig. 3.11). It is not perfectly centred on the platform, but it no doubt supported the actual column. The three platforms in Room 65 may also have had worked bases in their centres, but erosion and modern plowing has destroyed a good portion of their height (plan xviii).8 The elevation of the single stone base c65N is 192.31 m above sea level, which, assuming that the floor was level across the entire room, should have been the elevation of the feet of the three other columns. The present upper surface of c65E is 0.16 m below this elevation, and the two southwestern bases fall short by c.0.28 m. Thus, at least one course of rubble is missing from c65E and several courses from the other two platforms.

I distinguish between building system and building component because the terms for each are often used interchangeably. The word ‘ashlar’, for instance, can be applied to a single block or an entire wall, yet there is a fundamental difference between the two – an ashlar wall consists of more than just an outer face of finely cut blocks. Masonry styles are crucial to the understanding of the architecture of the Palace of Nestor and its development, and clear distinctions must therefore be made between them.

Platform foundations transferred a column’s load to bedrock where the bedrock was some distance below the surface. At the north corner of Room 65, bedrock lies c.0.31 m below the level of the plaster floor, and the single stone base, c65N, sufficed to convey the load down this short distance. But south and southwest of here, the bedrock drops off rapidly. Platform c65E is sunk down to bedrock c.0.44 m below the floor, whereas c65W and c65S are partially built on top of earlier walls, which transfer the load of the columns indirectly to bedrock. (The walls are founded on bedrock and lie at depths of c.0.89 and 1.24 m below the two platforms.)

Building Components An ashlar9 is a worked block of stone used as a component of a masonry building system. Ashlar blocks vary greatly in size and at Englianos are cut from poros limestone.10 In ashlar masonry, the block’s exposed face or faces are cut square or rectangular and are worked smooth and flat 8

The number of columns in this room remains a matter of debate; the northeastward placement of the existing column group suggests the presence of a third pair in the southwestern portion of the room. However, these were not found; see Nelson 2001, 29–31. 9 For the use of this term elsewhere, see J. Wright 1978, 134; Hult 1983, 91; J. W. Shaw 1973a, 92. 10 Elsewhere they are cut from hard limestone, sandstone and conglomerate, as well as poros limestone (Hult 1983, 49–50).

To insure that they provided adequate support, the platforms for the columns in Room 65 were built with their diameters substantially greater than that of the columns that once stood upon them. The builders may have felt it necessary greatly to enlarge the platforms because the

303

Nelson – The Architecture of the Palace of Nestor (fig. 3.12). Bedding joint surfaces (top and bottom) are usually cut parallel to one another; vertical joint surfaces (sides) are cut back obliquely from the front face of the block in a form of anathyrosis. As a result, blocks look wedge-shaped or trapezoidal in plan. Surfaces other than the front face of the block are rarely finished smoothly, but generally left roughly worked.

An orthostate11 is a worked block of stone used as a component of a masonry building system; it has all the characteristics of an ashlar block save its dimensions. An orthostate block is cut as an upright slab, such that its thickness (measured from the front, exposed face to the back of the block) is less than its length or height; thus the block appears thin.

Blocks reused in pseudo-ashlar masonry have their exposed face or faces cut square or rectangular. Contrary to typical ashlar blocks, however, their joint faces lack anathyrosis; the blocks are roughly cubic or brick-like in shape. None of the pseudo-ashlar blocks at the palace could have their back faces examined, as they are all in situ.

In general, orthostate blocks are rare in Mycenaean architecture. At Tiryns, blocks with appropriate orthostate proportions were reused in rubble construction at the extreme north portion of the east wall of the court (IV) before the megaron (fig. 3.13). The blocks are broken, and their original exposed outer faces are turned inward; it is possible that they are fragments of a single block. Anathyrosis was used on the vertical joint surfaces, and the back faces are roughly worked.12 The blocks of the lowest course of masonry in the dromos walls of the Treasury of Atreus (fig. 3.14) and the Tomb of the Genii (fig. 3.15), both at Mycenae, give the impression of being orthostates, but their thicknesses are unknown – they could be cut with dimensions proportional to a typical ashlar wall block (Wace 1949, 43, figs 5 and 15; Wace does not use the term ‘orthostate’). Note also that these are retaining walls, not freestanding walls. Building Systems The ashlar style building system13 for freestanding walls consists of three elements: an exterior face of ashlar blocks laid in horizontal courses, combined with at least one timber course (probably more) and an inner face of rubble or rubble-based masonry. Course heights vary but usually remain the same for individual courses. Section 9 of the Southwestern Building’s southwestern facade (figs 3.7– 3.8) is a good example of the ashlar style building system.

3.12 Measuring and defining ashlar blocks.

11

Th is term is used infrequently in discussions of Mycenaean architecture. Blegen and Rawson (PN I, 206) described the reused ashlar blocks at the base of the northeast wall of courts 42 and 47 (their original exterior faces now turned downwards) as orthostates. J. Wright (1978, 144) applied the term to the ashlar blocks of the northeast facade of the Southwestern Building (wall 18L in plan III). But these blocks are no different than those of the northeast facade of the Main Building or the southeast wall (wall 5E) of Court 3. In both instances, the reference to the blocks as orthostates seems to depend on their location and use (at the base of a wall) rather than to the blocks themselves. Minoan orthostates (J. W. Shaw 1973a, 83) are defi ned similarly, and the term applies to Bronze Age architecture farther afield as well. For other Bronze Age cultures see Hult 1983; the definition of orthostate on his page 46 is perhaps better than the glossary entry on page 91. For Egyptian architecture see Arnold 1991, 296. 12 The side faces were cut with a flat bladed chisel, c.0.04–0.06 m wide, applied in a series of short strokes. 13 Th is defi nition does not apply to the ashlar masonry of stomion facades and dromos walls of tholos tombs, because there is no evidence for the use of timber in those. In Minoan architecture, ashlar masonry usually sits on a worked stone socle and usually does not incorporate long, horizontal timber beams between courses (J. W. Shaw 1973a, 92–107).

3.13 Orthostate block from east side of court to the south of the megaron, Tiryns.

304

Building Methods

3.14 Elevation of the Treasury of Atreus, Mycenae, showing north dromos wall, south face (after Wace 1949, fig. 5).

3.15 Elevation of the Tomb of the Genii, Mycenae, west dromos wall, east face (after Wace 1921–23, pl. LX).

Ashlar shell construction14 of freestanding walls comprises an interior and exterior face of ashlar masonry and an inner core of rubble and mud. An example of this type of construction can be seen in the northwestern end of wall 11L (see fig. 3.3). This system made use of wooden beams; at Englianos, mortises indicate where a horizontal timber

was fastened atop the second course of blocks in a wall. This method of construction seems to be unique on the mainland to this palace, where two ashlar shell walls still stand; it has yet to be found at any other Mycenaean site. On Minoan Crete it was used infrequently (J. W. Shaw 1973a, 92–107, esp. 104), but it was more popular in the Bronze Age architecture of cultures farther east.15

14

Th is term is little used in Bronze Age architecture to describe this type of building system. For Egyptian architecture see Arnold 1991, 148–151; for later Greek architecture see Coulton 1977, 31–32. 15 For similar Cypriot, Ugaritic and Anatolian building systems see Hult 1983. 16 Minoan orthostate masonry uses the same elements, though the socle is more fi nely fi nished and often cut square (J. W. Shaw 1973a, 83–92). At Kommos, the horizontal timber beam was replaced with a course of ashlar masonry in perhaps the last use of this system on Crete (J. W. Shaw, 1983, 213–16). For Bronze Age orthostate masonry of nearby regions, to which this defi nition also applies, see Hult 1983. G. Wright (1992, 413, figs 246–48) refers to this style of construction on Cyprus as ‘orthostate socle construction’. There, the tops of the socle blocks were sometimes covered by built floors. The Egyptians cut and used orthostates, but their orthostate building system did not incorporate wood (Arnold 1991, 142, 164, 175). 17 Since the blocks are defi nitely reused, the mortises may have been carved for their original use or added later. Küpper (1996, 73) partially addresses this dilemma and recognises two of them as probably original, but leaves the question open. Since there are no signs of the use of wood in this wall, it seems most likely that the mortises were cut for the blocks’ original use.

Orthostate masonry in freestanding walls has four components: a single course of orthostate blocks on the outer or exterior face, a horizontal timber beam course fastened atop the orthostate course, a backing or interior face of rubble or rubble-based masonry, and a semiworked stone socle upon which the orthostates and rubble masonry sit; see, for example, the early wall beneath Room 7 in figures 3.26 and 3.27 below.16 The socle course projects slightly beyond the face of the orthostates, adding to the structural stability of the wall. This projection may also have served an aesthetic function, perhaps emphasising the base of the wall. The material(s) and method(s) of construction for the portion of the wall that rose above the timber beam no longer survive at Englianos, and nothing is known of them. Orthostate masonry is nearly absent in Mycenaean architecture, the closest example being the Tiryns blocks mentioned above. These give every impression of being part of an orthostate masonry building system, especially given the mortises sunk into their upper surfaces.17

305

Nelson – The Architecture of the Palace of Nestor poros blocks were worked to three grades of finishing: smooth, semi-smooth and rough. Every ashlar block has a minimum of six surfaces, and the grade of finishing each received depended on its position within the wall. The exposed or exterior sides of wall and anta blocks20 were finished smooth, even and flat. After these faces had been finely chiselled, they were rubbed with an abrasive material of some sort, perhaps simply sand, which erased the tooling marks and created flat, smooth surfaces. At stone-to-stone or stone-to-wood joints the blocks were usually finished semi-smoothly, and where stone abutted either a rubble backing or an inner rubble core the blocks received only rough working. In some instances the rough working may simply be the finish left behind by quarrying and extraction of the block.

3.16 Random-range ashlar masonry.

Otherwise, the retaining dromos walls of the Treasury of Atreus and the Tomb of the Genii appear to be of orthostate construction, though the large blocks in the first course may not be slabs, and there is no timber course in any wall. These walls are retaining walls and not freestanding and are therefore products of a different building system than the orthostate building system at Englianos.

Upper and lower bedding joints received a semi-smooth dressing to permit a tight join between courses. Vertical joints were tight also, but only at the exterior face of the wall, since the blocks were cut so that only their exterior edges were in contact (see fig. 3.12). This type of anathyrosis was a timesaving technique that assured tight joints at the exterior face of the wall and produced the appearance of a finely cut ashlar wall.

Pseudo-ashlar masonry combines both squared-block masonry and rubble masonry for wall construction (the south facade of Building X is an example; see figs 3.24– 3.25 below). The squared blocks are usually laid in groups, not singly, and their location within a wall appears to be arbitrary; they may be coursed or laid in random range fashion, in which blocks of various heights and lengths are laid in irregular courses (fig. 3.16).18 Blocks used in pseudo-ashlar masonry are reused from earlier structures. They lack anathyrosis and are generally smaller than those used in ashlar masonry. The rubble stones vary greatly in size, from cobbles to large slabs, and some appear to be hammer-dressed. As with ashlar shell construction, this building system seems to have been unique to the Englianos palace within Mycenaean architecture.19

From the chisel marks still preserved on both semi-smooth and rough-cut surfaces, it is evident that at least two different instruments were used to the shape the blocks, one of which was applied to the stone in two different ways. The tool mark most frequently left behind is that of a flat-bladed chisel, the width of which varied from 0.06 to 0.07 m. It was applied in a series of short strokes in a direction diagonal to the face of the block (fig. 3.17).21 The number of strokes per series varies from two to nine, and the length of each stroke is relatively consistent at 0.005 m. The depth of each cut varies but is no deeper than 0.006 m. This chisel must have been a percussion tool that was held in the hand and tapped with a hammer. The repeated hammering left behind a series of short, shallow strokes.

Construction Methods and Devices

Other visible tool marks indicate that the same type of chisel – flat-bladed, with a width varying from 0.060 to 0.072 m – was sometimes applied to the stone in a slightly

The ashlar blocks used at Englianos were shaped and finished to various degrees for specific purposes, whether structural or aesthetic, and many walls preserve evidence of the methods and devices used in their construction. Ashlar style, ashlar shell and orthostate walls were fitted with wooden members, and mortices and beddings were cut into the blocks to fasten and secure those components; blocks were finished in specific ways to accommodate them. Before examining block masonry building systems in detail, and because these devices figure prominently in the building process, a brief survey of the evidence for them follows.

18

Th is style appears in the dromos walls of the Tomb of Clytemnestra and the inner sidewalls of the stomion of the Kato Phournos tomb (Wace 1949, pls 6 and 16 respectively). See also Fleming, Honour and Pevsner 1980, 258; Phillips 1994, 20. 19 Th is particular style of masonry is not mentioned as a building system or a construction method in discussions of Mycenaean architecture. Hult (1983, 21–31) recognised the style in Ugaritic architecture, but provided few details. J. W. Shaw (1973a, 83) distinguished it in Minoan architecture as rubble masonry with reused cut blocks collected from older structures. 20 Anta blocks were set at wall corners, intersections and terminations. 21 In figure 25, the stroke direction is diagonal to the upper face of the block. Diagonal stone fi nishing was noted by G. Wright (1992, figs 216 and 219) in Cypriot masonry. For the possible use of chisels and adzes in Minoan masonry, see J. W. Shaw 1973a, 70–75.

Stoneworking Techniques The ashlar blocks at the palace were chiselled and abraded; none appear to have been cut with saws. In general,

306

Building Methods different manner. Again, the chisel left a series of strokes diagonally across the face of the block, but they are not as uniform as those noted above. The lengths of the individual strokes are much longer, varying from 0.008 to 0.12 m (fig. 3.18), and the chisel was driven into the stone to a greater depth, averaging c.0.008 m. These marks, and particularly the variance in stroke length within a single series, seem to indicate that the mason attempted to apply the chisel generally in one direction and in a continuous series, but was unable to maintain a consistent and uniform application. This tool may not have been a percussion chisel but rather a hand-swung or hafted instrument, such as an axe or adze. If so, the marks suggest that the mason stood over the block and made short swings down along its face, shaving off relatively thin slivers with each swing.22

(PN I, fig. 290:4) appears to be only the lower, or cutting, end of a chisel, and its blade shape matches that of the intact chisel from Room 62. The Room 62 chisel (PN I, fig. 300:11) and the one recovered in Room 99 (PN I, fig. 316:6) are both flat-bladed with slightly convex cutting edges, and both taper upward from the cutting edge to the butt end (the end that receives the blow from the hammer). The Room 62 chisel has a blade width of 0.018 m and is 0.135 m long. The one from Room 99 is 0.087 m long, with a blade width of 0.031 m. From the difference in sizes, it seems that the Room 62 chisel was used for rough cutting because its relatively small blade could be driven more deeply into both soft stone and wood than a wider blade. The Room 99 chisel may have been used for finer finishing, such as shaving off slivers of either soft stone or wood. These two chisels are similar to those found elsewhere in the Aegean. That from Room 62 belongs to Type III in Keith Branigan’s typological study of Aegean implements from the Early and Middle Bronze Age. Such chisels have flatedged butt ends and concave sides that flare out towards the cutting edge, which is usually convex (Branigan 1974, 25). The chisel from Room 99 conforms to Branigan’s Type III flat-axe; this has the same characteristics and dimensions of his Type III chisel, though flat-axes tend to be broader throughout their entire length, and their sides flare a bit more. In Branigan’s categories of bronze tools, the difference between a flat-axe and chisel is the length-to-maximum width ratio – for the former, this number is less than 3.5. In other words, chisels tend to be

The third tool mark is one left by that of another flat-bladed chisel, but one with a slightly convex cutting edge. Its width varies from 0.057 to 0.065 m, and it too was applied in a series of short strokes varying in length from 0.005 to 0.006 m. This chisel was neither widely used nor does it appear to have been reserved for any particular block face or area nor required for any special cutting.23 Five bronze artifacts recovered in the original excavations may be chisels or chisel fragments (table 3.2).24 Judging from the published photograph (PN I, fig. 307:2), the fragment found in Room 94 is quite dubious as a stoneworking tool because of its shape: square in section and apparently without any sort of cutting edge. It may be a fragment of a square nail such as the one recovered outside the Southwestern Building (PN I, fig. 302:8). The chisel discovered in the area of the Southwest Quadrant (PN I, fig. 302:6) is quite small and may have been used for detailed work, perhaps for cutting the numerous mortises on the tops of many ashlar blocks or, more likely, for ivoryor wood-carving. There are no visible marks of a pointed chisel on any of the poros blocks. The other three bronze tools, one of which is only a fragment of a blade tip, are certainly chisels, but probably not stoneworking tools, judging from the marks on limestone blocks. The fragment recovered from Room 52

3.17 Ashlar block 33, northeast facade of Main Building (wall 11L). 22

Using G. Wright’s categories of Striking Percussion Tools and Struck Percussion Tools, the tool that left behind the series of short strokes with consistent stroke lengths was a struck percussion instrument; the other tool, applied with varying stroke lengths, was a striking instrument (G. Wright 1992, 365–66). 23 Marks from a claw or toothed chisel are evident on the poros block ashlar facade rebuilt by the Greek Archaeological Service at the eastern end of the northeast facade of the Main Building (the current exterior wall of rooms 48 and 50). Since only these stones reveal this type of chisel mark and since they were reused in 1960, this type of chisel should not be assigned to the Bronze Age, at least at this palace. My thanks to Robin Rhodes for his comments on the chisel marks at the palace during the 1995 MARWP field season. 24 Bendall (2003, 189–91, fig. 5) plots the fi ndspots of chisels and chisel fragments recovered in the excavations.

3.18 Ashlar block 35, northeast facade of Main Building (wall 11L).

307

Nelson – The Architecture of the Palace of Nestor longer and thinner, while flat-axes tend to be shorter and broader. According to J. Deshayes’ typology, which covers a broader chronological range and geographical area than Branigan’s, the intact chisels from Englianos fall roughly into the same categories as in Branigan’s study: flat-axes and chisels.25

Clamps may have been used in block masonry construction at Englianos, but since the blocks with preserved clamp cuttings are no longer in situ, and their original position within ashlar walls is now lost, they are discussed separately in Appendix B (pp.371–72 below). • The Evidence of Mortises

The blade widths of the Room 99 and Room 62 chisels, whether they are designated chisels or flat-axes, do not match the marks on the blocks (0.057–0.070 m). The blade widths are smaller, so these chisels could not have been used to work and shape the ashlar blocks. As flat-axes generally have wider blades than chisels (Branigan 1974, 24–26; Deshayes 1960, 85–100), the finishing of the blocks at Englianos may have been completed with flat-axes or similar instruments that had broad, thin cutting edges. Although the thinness of the flat-axe is not conducive to repeated blows by a hammer on its butt end – it may deform after only a short period of use – poros limestone is relatively soft when it is freshly quarried and could have easily been planed with a flat-axe tapped in a series of short, controlled strokes. Also, the flat-axe might have been mounted in a wooden handle that would reduce and absorb any damage inflicted by repeated strokes. As already noted, the irregular strokes evident on some ashlar blocks were probably inflicted by a hand-swung instrument such as a small flat-axe or adze.

Square or rectangular mortises were cut into the upper surfaces of anta blocks, ashlar wall blocks and orthostates. The mortises received dowels that once fastened either wooden vertical posts or horizontal beams, or a combination of both, to a block or group of blocks. The cutting of mortises and their positioning on the upper surfaces of blocks follow several recognisable rules. First, there are more mortises in ashlar blocks than in orthostate slabs, though this may be due simply to the larger number of extant ashlars. Second, antae contain more mortises than wall blocks. Third, not every wall block in the full length of a course has a mortise, but those that do have only one. The exceptions to this rule are the blocks of the northeast facade (plan iii, wall 18L, and fig. 3.39 below) of the Southwestern Building, where each block contains a pair of mortises. Each of the mortises in a pair belongs to separate phases of construction, however; the paired mortises were not used in conjunction (Nelson 2003). In another exception, when a wall block turns a corner it may have two mortises, as in block 11L-45 (fig. 3.37 below). Fourth, mortises in wall blocks are usually sunk near the centre of the block’s upper surface, though mortises near the ends of a block are not uncommon. Finally, there is a relationship between the placement of mortises and a block’s exterior face. For instance, if a block’s northeast face is exposed, the mortise will be sunk on its upper surface near the northeast edge of the block. The two mortises of block 11L-45 each relate to one of the block’s two exterior faces.26 Only two blocks (11L-26 and 11L-30) form an exception to this last rule; they are discussed below.

The Use of Wood in Ashlar and Orthostate Walls In the orthostate, ashlar and ashlar shell building systems, wood was fastened to worked stone with dowels (of which the only evidence is their accompanying mortices) and bedding cuts. Anta blocks contain many more mortises and bedding cuts than other building blocks, but many of these were reused in later building phases, at which time wooden members were not attached to them. These reused blocks are examined below in the discussions of pier walls and rubble walls. 25

Flat-axe types B, C, D, G, J and I4 and chisel types A2, A3, C3a, E and F (Deshayes 1960, 51–84 and 85–100 respectively).

Findspot

Type

Condition

Length

Blade width

Handle width

Thickness

Room 52 (PN I, 218, fig. 290:4)

Flat-blade, convex

Fragment

0.016 m

0.016–0.02 m

?

?

Room 62 (PN I, 244, fig. 300:11)

Flat-blade, convex (?)

Intact

0.135 m

0.018 m

0.01 m

0.001-0.008 m

SW Quad (PN I, 285, fig. 302:6)

Pointed

Intact

0.053 m

0.003 m

0.005 m

?

Room 94 (PN I, 306, fig. 307:2)

?

Fragment

0.019 m

0.003 m

Room 99 (PN I, 322, fig. 316:6)

Flat-blade, convex

Intact

0.087 m

0.031 m

Table 3.2 Bronze chisels and chisel fragments recovered in the original excavations.

308

? 0.007 m

0.0025 m

Building Methods • The Evidence of Bedding Cuts

The actual cutting of the mortises into the top surfaces of the blocks was done precisely. The masons used setting lines, lightly inscribed into the stone surface, to indicate the placement of the mortises. Although most have been erased, probably during the final trimming or finishing of the blocks, some setting lines are still preserved (fig. 3.38 below).

Bedding cuts are smoothly chiselled surfaces on the upper faces of blocks, onto which horizontal wooden beams were laid (fig. 3.19). Beddings were cut on both anta and wall blocks near the block’s exterior or exposed face(s), similar to the placement of mortises. Beddings do not span the entire depth of the blocks into which they were chiselled, but presumably were cut just slightly deeper than the depth of the beams to be laid in them. The remainder of the block’s upper surface was left only roughly chiselled, because it received the mud and rubble of the wall’s backing or interior core. Bedding cuts are not as prolific as mortises and may have had several functions. They provided a level and even surface for the wooden beam, and they were a time-saving technique for the masons, who had only to finish smoothly and evenly approximately half the upper surface of a block. Beddings may have been created during the process of levelling an entire wall course (Nelson 2003). At Mycenae and Tiryns, beddings were cut only on anta blocks (Dörpfeld 1885, 268–69; see also Küpper 1996, figs 6, 9–11, 16, 18–19 and 24).

Despite the evident precision and care taken in the placement and cutting of each individual mortise, the layout of the mortises as a group appears far less controlled. Aside from the general positioning rules noted above, the distances between mortises do not conform to any recognisable pattern and must not have been measured according to a repeated or standard unit. In modern American balloon-frame house construction, the builder fastens the foot of the wooden-framed wall to the masonry block foundation with bolts placed according to a repeated unit, usually every 16 inches. This sort of consistency is not evident in the placement of the mortises at the palace, and they appear to have been laid out haphazardly at irregular intervals. From measuring the extant dowel holes preserved in the northeastern facade, particularly their setback distance from the exterior face, and the setting lines, J. Wright (1978, 137–39, 141–43, fig. 204) determined that dowel hole spacing relates to the dimensions of specific horizontal beams. In other words, the cross-section and length dimensions of the wooden members were known in advance of the cuttings being made, and not vice versa, as one might expect. Such a supposition suggests that wood may have been a more valuable resource than stone.

• The Function of Wooden Members Mortises and bedding cuts indicate the presence of horizontal beams. Where appropriate, these probably served as lintels and window sills (Blegen 1965, 120–21),

The securing of wood to cut stone with dowels also occurred at Mycenae and Tiryns, but neither site preserves long stretches of cut-stone walls. The two longest walls are the Great Poros Wall around the Tomb of Clytemnestra and a similar wall around the Treasury of Atreus (Taylour 1955, 209–23, and Wace 1956, 116–19, respectively). Neither wall seems to have used wood in its construction.27 Most of the evidence for wood-to-stone connections comes from the many anta blocks at both palaces (Küpper 1996, fig. 181–202); as with the Englianos antae, mortise placement corresponds to the exterior, exposed faces of the blocks. The builders at Tiryns also used circular mortises that were cut with a bore or, in the case of harder stones, particularly conglomerate and the denser limestones, a drill (Dörpfeld 1885, 265–67). Circular dowel holes and worked conglomerate do not exist at Englianos.

26

In the building chronology of this facade, block 45 was in place before block 44, which abuts it to the northwest (fig. 3.35 below). Along with several others (particularly changes in masonry style), this fact suggests that Room 32 was an addition to an already standing structure; thus the northwest walls of rooms 30 and 31 were once exterior and possibly faced with ashlar. 27 Neither wall has been fully excavated, no mortises have been reported, and the published photographs reveal no dowel holes or bedding cuts.

3.19 Ashlar block 49, northeast facade (wall 11L) of Main Building.

309

Nelson – The Architecture of the Palace of Nestor but mortises in the Main Building’s northeast facade indicate that beams ran the full length of the wall here and even returned at antae and wall setbacks (J. Wright 1978, 137–39, fig. 204; J. Wright 1984, 26–27, note 4, fig. 9; Küpper 1996, 80–84, pls 16, 17, 19, 21). A continuous course of horizontal beams goes well beyond the simple function of spanning openings and probably indicates another or supplemental function. The height of the timber course in the northeast facade cannot be reconstructed, but the bedding cuts indicate beams varying generally from 0.25 to 0.35 m in depth.28 Above the beam, the wall would have continued in ashlar masonry, but fallen blocks, also with cut mortises, point to another timber course higher up in the wall. The beams probably had several functions: to tie together blocks within a single wall, to serve as lintels over doors and windows, or simply to create decorative string courses. They were presumably visible in the finished wall, but this must remain speculative, since no wooden members were found in situ.

All of the anta blocks at Englianos have mortises, and some have both mortises and bedding cuts. The antae of the palace can be classified into five groups: one-, two-, three- and four-sided antae, and doorjambs (table 3.3; the locations of the various antae are shown in Plan iv). The four ‘sided’ types are classified according to the number of exposed faces, that is, those faces that were not covered by wall matrix and were meant to be visible. The only example of a one-sided anta block is the northwest anta (18L.b) of the Southwestern Building’s northeast facade. In plan, the block is T-shaped (plan xix), corresponding to the T-intersection of the northeast facade with the wall separating rooms 64 and 67. The northeast facade originally continued northwestward, but was robbed out beyond the T-shaped anta; only the bedding stones of its foundations remain. A horizontal beam running along the northeast face of the facade would have been fastened to the top of the block at its northeast edge with a single mortise. The beam probably extended northwesterly, for if another beam had been intended to run southwest, at 90 degrees to the first, a mortise would have been cut near the northwest side of the top surface.

Evidence for the use of vertical posts incorporated in cut stone masonry is sparse in comparison to that for horizontal members. A noted exception occurs in wall 10C, which separates Corridor 26 from Room 32 (fig. 3.1 above).29 In the second course of ashlar, there is a 0.22 m gap between blocks 10C-11 and 10C-12, which could accommodate a narrow wooden upright springing from the top of the first course. The wall matrix of the inner backing almost entirely fi lls the gap, however, leaving little room for any type of insert.

The two-sided anta blocks have two exposed, exterior faces and were set at wall returns. The three-sided antae have three exposed, exterior faces and were set at wall ends. Doorjamb antae have two opposite exposed faces. These probably did not serve as actual doorjambs but provided support for wooden jambs. Many of these blocks contain cuttings that apparently served as sockets for the upright timbers of the jambs. Three of the largest and no doubt originally very impressive doorways of the Main Building (1–2, 4–5, and 5–6) have doorjamb anta blocks with such sockets, in which Piet de Jong reconstructed large wooden uprights; doorway 4–5 is shown in figure 3.20 (PN I, figs 429–31).

Another possible piece of evidence for wooden uprights are two deeply cut sockets in the first, and only surviving, course of the Main Building’s northeast facade outside of Room 46 (fig. 3.35 below). The socket in block 11L-34 is sunk to a depth of c.0.19 m and has a width of c.0.24 m. Though obscured by the inner backing of the wall, the socket seems to extend across thickness of the block. The socket in block 11L-35 is cut at the block’s southeast end and is sunk to a depth of c.0.17 m. It has the same width as the socket in 11L-34, c.0.24 m, and also seems to span the entire thickness of the block. Each socket has a single mortise, which either fastened an upright post that rose up the face of the wall or a horizontal beam that ran through the thickness of the block. Blegen and Rawson concluded that the sockets spanned the entire thickness of the wall (PN I, 205), but since the rubble backing of the facade clearly limits the length of any beam, this seems unlikely. Upright timbers could have framed a window opening here (ibid.). The slots are approximately centred in the wall relative to the room, as is the hearth, suggesting an orchestrated arrangement (Küpper 1996, 84).

Mortise placement relates directly to anta type, corresponding to the exposed faces of the block. Wooden members attach to two-sided and doorjamb anta blocks only at the two exposed faces, and at the three exposed faces on three-sided antae. Bedding cuts have the same relationship to exposed faces as mortises do. The doorjamb antae of doorway 1–2 (5D.a and 5D.b) are good examples of this association: both have bedding cuts indicating the presence of horizontal beams running southwest– northeast at the northwest and southeast exterior edges (figs 3.21–3.22). Another example is the badly damaged anta block at the southeast end of the northeast facade 28

The southeast bed of anta 18L.a in the Southwestern Building’s northeast facade (plan IV) is c.0.69 m deep. A beam with a cross-section of this dimension would be quite substantial. 29 Wall 10C was rebuilt in a slightly different fashion in a later phase; see below.

• Anta Blocks In block masonry construction, anta blocks were set at wall returns, intersections and terminations. They also appear to have served a decorative function, as shown by two examples in the northeast facade of the Main Building.

30

Mortises indicate a 2-sided anta. Mortises indicate a 3-sided anta, but the cutting of the block’s vertical joint surfaces clearly specify its position as a wall return.

31

310

Building Methods Anta

Wall

Location

Anta Type

Bedding Cuts (corresponding face)

5B.a

5B.1

Southwest anta, door 5–6

Doorjamb

5B.b

5B.2

Northeast anta, door 5–6

Doorjamb

5C.a

5C.1

Southwest anta, door 4–5

Doorjamb

5C.b

5C.2

Northwest anta, door 4–5

Doo jamb

5D.a

5D.1

Northeast anta, door 1–2

Doorjamb

Northwest, southeast

5D.b

5D.2

Southwest anta, door 1–2

Doorjamb

Northwest, southeast

5E-1

5E

Southwest anta, Stair 54

Custom cut

Southwest, northwest

8L.a

8L

Southeast wall end

3-sided

9.a

9L

Northwest anta, stair 36

2-sided

9.b

9L

Northwest anta, stair 36

2-sided

11L-20

11L

Ashlar block 20 (PN I, 49, table)

3-sided

11L-25

11L

Ashlar block 25 (PN I, 49, table)

3-sided30

11L-26

11L

Ashlar block 26 (PN I, 49, table)

2-sided31

14L.a

14L

Room 93, southwest wall, southeast end

badly damaged, particularly upper surface

15L.a

14L

Room 93, northeast wall, southeast end

3-sided

15L.b

15L

North corner, room 93

4-sided

18L.a

18L

Southeast anta, northeast facade of SW Bldg.

3-sided

18L.b

18L

Northwest anta, northeast facade of SW Bldg.

1-sided

SW51.b

SW51

Northwest anta, doorway 64–65

badly damaged, particularly upper surface

Southwest, northwest

Northeast, southeast and southwest

Table 3.3 Anta blocks.

of the Southwestern Building (18L.a), which has three bedding cuts (fig. 3.23). They correspond to the anta’s three exposed faces (northeast, southeast and southwest) and outline the U-shaped formation of the beam that once fastened to the block.

and rubble masonry in a style of wall construction that at first glance seems irregular. But despite the mixing of construction styles and the lack of concern for a uniform wall appearance, the pseudo-ashlar system does seem to follow two general rules. First, squared blocks were placed in the walls on the downhill side of a structure. All of the pseudo-ashlar walls sit on the southwestern edge of the hilltop, where the topography slopes downwards to the southwest. The builders may have preferred stacked squared blocks over rubble because the downhill wall needed to be taller and thus more structurally sound than its uphill counterpart. Second, squared blocks were usually laid in the lower courses of walls.

One anta, apparently cut to fill a unique position in the architecture of the palace, cannot be categorised. The southwest anta block of the doorway between Court 3 and stairway 54 (5E-1) gives the impression that it is a three-sided anta, but the bedding cuts and the mortises clearly indicate that it is a two-sided anta with only two exposed faces: the southwest and northwest (plan ix). The northeast side, which in part served as a doorjamb, has a cut recess that could have been a socket to support a wooden doorjamb.

Three structures built in the pseudo-ashlar masonry style survive in situ. Just to the southeast of Court 63, and running partially underneath it, lay a series of nicely laid and well built walls (SW58a–f and SW59 in plan iii; see also plan xvii). They represent the remains of an earlier structure or structures founded directly on the marly limestone bedrock.32 Wall SW58a runs northwest–

Pseudo-Ashlar Masonry Examples of pseudo-ashlar walls at the palace are few. As defined above, the system combines ashlar blocks

311

Nelson – The Architecture of the Palace of Nestor

3.20 Doorway 4–5, rooms 4 and 5, Main Building (after PN I, fig. 430).

3.21 Anta 5D.a, northeast anta of doorway 1–2, rooms 1 and 2, Main Building, from the southwest.

3.22 Anta 5D.b, southwest anta of doorway 1–2, rooms 1 and 2, Main Building, from the northeast.

312

Building Methods the bedrock and preserves only a few courses: an ashlar course (founded on bedrock) and a course, in some places two, of rubble masonry. The south wall of Building X is also of pseudo-ashlar masonry (plan xv, figs 3.24–3.25). Its west and east ends, where the wall returns, are constructed with squared blocks, perhaps as quoins, while the middle stretch of wall is rubble masonry.34 Unlike the walls just surveyed, the ashlar sections do not rest on bedrock, but on a foundation of several courses of rubble, which is slightly wider than the wall above.35 The large blocks at the east end are laid in random range fashion. Most blocks used in pseudoashlar walls are smaller than those of the ashlar style, but some of the blocks of Building X are comparable in size to the blocks of the northeast ashlar facade of the Main Building. Small stones are inserted between some blocks, in both butt and bedding joints. One block is cut with a crotch joint, frequently found in random range ashlar. This type of joint occurs in blocks the upper surfaces of which are cut to two different heights to accommodate changing course heights.36 Section 8 of the Southwestern Building’s southwestern facade (wall 17L in plan iii; see also plan xiv) is the third pseudo-ashlar building wall on the site (PN I, fig. 215).37 The excavators considered section 8 part of the southwestern facade (PN I, 279–80, fig. 412), but there is no physical connection, either abutting or bonding, of this section with sections 7 and 9. Its entire length has not been traced;38 beginning from the west corner of Building X it runs more than 18 metres almost directly westward, though its course is not straight. Three courses of random

3.23 Southwest facade (wall 18L) of Southwestern Building, from the southeast.

southeast and returns at its northwest end, where it appears to end abruptly. Here it may have once stepped up in a northeasterly direction with the rising topography of the hilltop. At its southeast end, SW58a was destroyed to make room for a new wall, SW59, which runs perpendicular to it. Another, later wall (SW58b) abuts SW58a’s northwest return and continues underneath court 63. There it possibly intersects wall SW58e, either in a bond or an abutment. Wall SW58f follows the same course as SW58a but is slightly thicker, so it may not be a continuation of the latter (PN III, fig. 306). Farther to the southeast lies wall SW60. It is built in the same construction style and seems to have been part of the same series of walls as SW58a–f and SW59, though it is not as nicely laid, nor does it follow such a straight and even course.

32

The somewhat schematic published plans of these walls (PN I, 282, fig. 412; PN III, 39, fig. 306) do not accurately reflect their various abutments or bonds. 33 The original excavators did not provide a relative chronology of these walls, possibly because they were all constructed in the same manner (PN III, 39). The destruction of one wall and the abutment of another may represent either the passage of time or perhaps a method of constructing the building in small portions. 34 Blegen and Rawson identified some of the rubble masonry in this wall as repair work or reinforcement (PN I, 283; PN III, 40). Around 1.60 m of the rubble section may be later repair work, since there appears to be a break in the wall. Otherwise there is nothing, such as a change of wall thickness or building technique, that suggests reinforcement. 35 The original excavators referred to this entire wall as a foundation resting on a foundation (PN III, 40, fig. 75). Since the rubble wall beneath the pseudo-ashlar wall is slightly thicker, it probably served as the actual foundation. 36 See, for example, the dromos walls of the Tomb of Clytemnestra at Mycenae (Wace 1949, pl. 6). 37 Blegen and Rawson (PN I, 279) described this wall as if it were built in the ashlar style, but it clearly contains rubble masonry along with squared blocks on its outer face and does not show any sign of having held wooden members. It could not possibly have been constructed in the same manner as, for instance, the Main Building’s northeast facade.

Although these bonding and abutting walls point to several phases of construction,33 all of them were built in the pseudo-ashlar style, with some consistency in the placement of squared blocks. For the most part, ashlar blocks rest directly on bedrock and form the first course. As the topography rises, the first course steps up accordingly. For example, two courses of ashlar and three to four courses of rubble masonry survive in wall SW58f. The squared blocks were laid in the first two courses, and they step up towards the northwest, following the steep rise of the bedrock. The masonry work is random range ashlar (fig. 3.16 above), probably in response to the rising topography. Wall SW58a runs parallel to the contours of

313

Nelson – The Architecture of the Palace of Nestor

3.24 Elevation, southwest facade of Building X, Southwest Quadrant.

the palace. Here, they are laid neither at the corners nor in the lower courses, in both of which situations they would have added greatly to the structural integrity of the wall. Orthostate Masonry There are few examples of orthostate walls at the palace. This type of construction went out of use several building phases before the palace’s destruction, and as a result, walls constructed in this system survive only in fragments and can only be seen at points below the later floor levels. All of the orthostate blocks are cut slabs set upright with their exposed face finished smoothly and their rear and vertical joint surfaces left roughly worked, the latter cut back with anathyrosis. All of the socles were meant to be partially visible, and their front faces were worked appropriately. In one instance, the portion of the face that was visible was worked smoothly, while the remaining face, below ground, was left only roughly worked. The socle’s upper surface was usually worked more or less smoothly in order to receive the orthostate. In places where socle blocks abut one another, their joint surfaces were cut back with anathyrosis. Elsewhere, there seems to have been no apparent concern for smooth and even joints, and gaps often exist between blocks. In these cases, vertical joint faces were left roughly finished or even uncut.

3.25 Southwest facade of Building X, Southwest Quadrant, from the southwest.

range ashlar are preserved in the westernmost portion of the wall, which is about four metres in length and was not fully exposed by MARWP. Several courses of rubble masonry support the squared blocks; above the blocks, rubble masonry probably continued. The reason for the positioning of the squared blocks in this wall is not as readily apparent as in the two other pseudo-ashlar walls at 38

There is some confusion in the presentation of this wall and this area of the site by the excavators. In PN I, the wall is thought to be part of the Southwestern Building’s southwest facade. In PN III (11–12), the wall is described as part of the circuit wall around the hilltop; it runs from the corner of sections 5 and 6 to the western corner of Building X and is composed of several sections of rubble and pseudo-ashlar masonry. The published drawing (PN III, fig. 306) contradicts the existence of a rubble wall (partially revealed by MARWP) presented in PN I (fig. 412) and PN III (fig. 303), which takes approximately the same course as the pseudo-ashlar wall. The courses of these two walls are represented with dashed lines in plan xvi. 39 The wall lies in the area dubbed by Blegen (1953, 63) the ‘chasm’, the outline of which can best be seen in the plan shown in PN I, pl. 406. Th is drawing depicts the broken edges of the plaster floor in the area of the Archives Complex, edges that must have traced the footprint of the original walls about rooms 7 and 8. When the soil was cleared from the area between the broken edges, there appeared a gaping hole in the smooth plaster floors, hence the term ‘chasm’. Blegen did not undercut the plaster floors when he excavated this feature, but the socle here appears to return in PN I, fig. 17, lower left corner (a portion of the northeast–southwest stretch is obscured by loose stones). MARWP recleared the northeast portion (between rooms 1 and 2 and 7 and 8) of the chasm in 1995; for a detailed discussion of its stratigraphy, see Part I, pp.76–80 above.

Beneath Room 7, just to the northwest of the reconstructed southeast facade of the Main Building, is a fragment of an orthostate wall (plan xii, figs 3.26–3.27; see also PN I, 94). Only two orthostates survive, both of which still stand in situ. The socle for this early wall is much better preserved, running 1.25 m beyond the northeastern orthostate and more than a metre to the southwest of the southwest orthostate. At the southwest end it turns to the southeast and continues in that direction.39 The slabs and blocks of the socle course were not finely cut, but were only roughly squared. Their exposed faces were for the most part worked to a smooth finish. Slab sizes vary considerably, from 0.19 x 0.12 m to 0.68 x 0.24 m; the original excavations did not reveal the depth of individual slabs. The slabs were loosely fitted together, with gaps of 0.010–0.05 m between them, and the interstices were fi lled with earth. Both orthostates are fully preserved and sit back 0.034 m from the front edge of the socle course. The block to the northeast is 1.14 m long by 0.52 m high; its thickness varies

314

Building Methods from 0.26 m at the southwest to 0.38 m at the northwest.40 Its northeast joint face was cut with anathyrosis, and a single mortise was sunk into the top of the block, near its northeast end.41 The southwestern block, with its inscribed double-axe mason’s mark on its front face, measures 1.20 m long, 0.52 m high and 0.42 m thick (PN I, 94). The excavators reported shallow cuttings on the tops of both blocks, which they interpreted as pry holes. The cuttings are quite small and would have required a prying instrument with a sharp, strong edge and considerable tensile strength, such as a modern steel crowbar. The poros limestone used at the palace is much too friable for prying by any type of sharp instrument unless it was a wooden log or tree branch. The cuttings are irregular and do not appear to have been cut with a chisel of any sort; they are probably simply the result of wear or accident.

be seen extending southeastward from beneath the later rubble wall (fig. 3.28). The socle was finished in the same fashion as that beneath Room 7: its face and upper bedding surfaces were worked quite smoothly with hammerdressing. Another orthostate and its socle can be seen extending northwestward from beneath the later rubble wall (fig. 3.29). Some of the socle blocks are not slab-like; that is, they do not have two roughly parallel bottom and top surfaces. Because of this, small stones were needed to level their upper surfaces. The orthostates here in section 7 are also smaller than those beneath Room 7. The thickness of the southeastern-most slab measures c.0.20 m, with a height of c.0.54 m; accurate measurements cannot be made because of the presence of the overlying wall. The visible vertical joints of both blocks exhibit anathyrosis. There are no visible mortises, but the later wall that straddles the orthostates may conceal one or more. Part of the rubble backing remains and is visible behind the orthostates, but the entire wall thickness cannot be measured; in Blegen’s published plan (PN III, fig. 306) the wall was drawn with a thickness of c.1.15 m. Likewise, the presence of a plaster coating on the interior face of the wall is unknown.

A third block, in situ, abuts the southwestern block on its southwest face (PN I, 94).42 It is damaged, but it is definitely not an orthostate block, since its finished height is little more than half that of its neighbor. The block was finished on its exposed face; its top appears to be flat. The change in block style perhaps indicates a feature in the wall that is now missing, such as a window or recess. The inner, rubble backing of the wall is still extant behind all three blocks and gives a total wall thickness of c.1.20 m. Behind the northeasternmost block, this rubble backing has partially tumbled, but there are no signs of plaster or clay coating the interior face. Abutting the wall’s socle is a cobblestone floor or pavement that spreads out towards the southeast. It is bordered at the southwest by the return of the socle course, but its extent in the other direction is unknown because it continues into the scarp beneath the plaster floor of Court 58 (now concealed by the reconstructed southeast facade; see PN I, fig. 17, for its original excavated state). On the other side of the wall, the floor is missing. All that is left is the bottom of the chasm (plan xii), which penetrated the floor associated with the orthostate wall.43

3.26 Plan and elevation (from northeast), orthostate wall beneath Room 7, Main Building (detail).

A portion of section 7 of the Southwestern Building’s southwest facade is also an orthostate wall (figs 3.28–3.29). This section of wall, preserved to slightly more than two metres, is partially concealed by a later rubble wall that runs roughly perpendicular to and over its top (shown in outline in plan iii; see also plan xvi). One orthostate can 40

Blegen and Rawson (PN I, 94) measured the block as 1.13 m long and 0.26–0.375 m thick. 41 The mortise measures 0.058 x 0.042 m (length is measured in relation to the front face or length of the block); its depth is 0.050 m. It is offset 0.106 m from the front face of the block and 0.340 m from the block’s northeast edge, measured from the block’s east corner. 42 MARWP’s permit did not allow the full uncovering of the two southwestern blocks. 43 The pottery recovered from the chasm dates to MH, LH IIIB and LH IIIC periods (PN I, 100). See further Part I, pp.76–80 above.

3.27 Early orthostate wall beneath Room 7, Main Building, from northeast.

315

Nelson – The Architecture of the Palace of Nestor Sometime later, the northwest end of section 7 was destroyed, and an ashlar-style wall built against the northwesternmost remaining orthostate block (fig. 3.29). The new addition was obviously meant to continue the course of the earlier orthostate wall, since its outer face is in line with that of the orthostate’s face, and the upper surfaces of both blocks reach the same elevation (the height of the ashlar block is c.0.27 m). Finally, at some later point still, the wall seems to have gone completely out of use when the later rubble wall was built over the top of it.

A surviving fragment of an orthostate wall stands just northeast of the northeast facade (wall 17L) of the Main Building outside of rooms 34 and 40 (fig. 3.30 and plan x). The wall seems to have been completely destroyed save for a short section running parallel to the northeast facade, into which these few blocks were partially incorporated. One orthostate and one anta (11L-20) remain, both of which sit on a socle of worked blocks (fig. 3.31).44 Unlike those of the orthostate wall beneath Room 7, the three visible socle blocks were nicely cut and well laid.45 They were worked smoothly on their upper surfaces, and their exterior (northeast) faces were cut square and even. They were set in a straight course, and the exaggerated anathyrosis cut into their side faces – emphasised particularly in the northwesternmost block, much more so than in any other block at the palace – created tight joints. The lone orthostate was apparently pushed out of its original position so that it could form the last stone, and perhaps serve as an anta of sorts, for the later, northwest rubble wall of Court 42. The gap between the anta block and orthostate is now fi lled with rubble and mud. The orthostate, 0.72 m long by 0.55 m high, varies in thickness from 0.29 to 0.34 m and sits back 0.12 m from the exterior face of the socle. It exhibits anathyrosis on both vertical joint surfaces, but not as exaggerated as that of the socle blocks. It was smoothly finished on its upper and front faces but was left only roughly cut on the back and vertical joint surfaces. Sunk into the top of the stone is a mortise (0.038 m long x 0.038 m wide x 0.075 m deep), cut approximately in the centre of the block near its exterior, northeast face. The anta block is rectangular and juts out from the northeast facade; its dimensions are 1.05 m long x 0.92 m.wide x 0.6 m high (PN I, 182). It too was set back 0.12 m from the front face of the socle so that it falls in line with the front face of the orthostate block. The anta sits on two socle blocks, the actual shapes of which are concealed by the anta itself. The southeasternmost socle block returns 90 degrees at its east corner and continues southwesterly in a direction parallel to the southeast face of the anta. The socle block does not run the entire length of the anta, but stops just shy of the northeast face of block 11L-21, where it seems to have been purposely cut or broken away. Thus, the socle block and the anta once formed an exterior corner in the orthostate wall. The anta later became part of the northeast facade, and certain adjustments were made to it. The socle block was cut to make room for block

3.28 Early orthostate block, section 7, southwest facade of Southwestern Building, from northeast.

44

Blegen and Rawson only reported the anta (11L-20) and its socle, referring to the latter as ‘a foundation of smaller poros blocks’ (PN I, 181–82). 45 The number of blocks is unknown; Blegen and Rawson did not report them, and the MARWP field permit did not allow for further cleaning of this area of the site. The plaster floor of Court 47 and the unexcavated earth between this wall and the northeast facade obscure the bottoms of the socle blocks and the material – earth or otherwise – that they rest on.

3.29 Another early orthostate block from the same section of wall.

316

Building Methods 11L-21 and was covered by the later plaster floor of Court 42, some of which creeps over the top of the socle at the north corner of anta 11L-20.

above are certainly orthostate slabs, but they were reused and probably no longer sit in their original positions. Otherwise, the orthostate masonry of Minoan Crete is the closest parallel to the Englianos masonry (for a more thorough discussion see J. W. Shaw 1973a, 83–92; J. W. Shaw 1983). The Minoan system combined the same components in the same manner: orthostate slabs with mortises for horizontal beams and a rubble backing, sitting on a socle of cut stones. Some differences are noteworthy, however. Minoan orthostate blocks tend to be larger, and socle blocks are more finely cut than those of the Englianos palace. At Phaistos, plaster sealed joints between blocks, and anathyrosis was not rigorously applied. At Kommos, in the last known use of Minoan orthostate construction, the usual timber course was replaced by a course of ashlar

The remainder of the surviving orthostate masonry walls at the palace lack standing orthostates. Their socles were all cut relatively smoothly on their front and upper faces and were set in straight courses. Some blocks fit tightly, whereas small gaps separate others. They resemble the socle courses just described and they clearly distinguish themselves as such among the remains of the palace. Beneath Room 57 at its southeast side, just to the northwest of the reconstructed facade, lies a course of socle blocks (plan xiii; see also PN III, fig. 310, wall y). As far as can be determined from the published photographs (PN I, 226–27, figs 13, 19, 143), they were all cut fairly smoothly on their upper and exterior faces.46 The joint faces exhibit anathyrosis, but the technique was not applied as consistently as it is in the later ashlar style building system. Nonetheless, the blocks were well laid, more so than in the socle beneath Room 7, and their course was set straight. A slight setback of c.0.08 m occurs in the line of blocks and certainly indicates a similar setback in the original orthostates. From the published photograph of the excavations in this area (PN I, fig. 13) it is clear that the socle course turns at its northeast end and runs towards the southeast. After continuing for c.1.42 m, it turns to the northeast and runs in that direction for approximately another five metres. At its southwest end, socle y continued southwesterly but was destroyed and replaced by the bedding stones laid for a later facade. Blegen and Rawson used the course of this later wall to reconstruct the northeast wall of Room 1 (wall 8LD in plan iii). The elevation of socle y is 191.55 m above sea level, which is close to the elevation of the socle of the orthostate wall beneath Room 7, at 191.60 masl. The two walls may represent two different buildings laid at the same time or one long continuous wall.47 Another socle course lies beneath and near the northwest end of Court 88 (plan xix). Fragments of the court’s plaster floor still cover some of the blocks, but one large block and four or five smaller ones are visible. Several other orthostate socle courses, particularly in the Main Building, were unfortunately covered by walls reconstructed in 1956 by the Greek Archaeological Service. The reconstructed southwest facade of the Main Building conceals a socle course with at least two setbacks beneath and near Room 10 (PN I, figs 20–21).

3.30 Plan and elevation of orthostate wall northeast of Room 34, Main Building (detail).

Orthostate masonry of the type used at the palace is rare in Mycenaean architecture. The Tiryns blocks mentioned 46

These blocks were not fully described in the original publication, and the modern (1956) reconstructed wall now prevents adequate observation. 47 The original excavators preferred the latter explanation (PN III, 35– 36).

3.31 Anta 11L-20 in Court 42, northeast facade of Main Building (wall 11L), from the northeast.

317

Nelson – The Architecture of the Palace of Nestor blocks (J. W. Shaw 1983). Otherwise, the conception and execution of orthostate masonry at the Englianos palace is identical to that in Minoan architecture (see further pp. 353–55 below).

In section 11Lb the blocks of the second course are uniform in height, 0.44 m (PN I, 49), but those of the first are not. These vary in height from 0.435 to 0.515 m and rest on a bedding of small stones laid directly on bedrock.

Ashlar-style Masonry

Many ashlar blocks were found among the debris just to the northeast of the northeast facade.49 The original excavators removed the blocks from courts 42 and 47 but left in place those lying outside of rooms 26, 27, 31 and 32. It was clear to them that the entire wall, at least up to the first storey, was built in the ashlar style. For the northwestern portion of the facade, from blocks 11L-20 to 11L-40, Blegen and Rawson calculated at least eight ashlar courses with two courses of timber (one between the second and third courses and another higher up) to a total height of c.3.70 m (PN I, 50).

Of the building systems that employ cut stone masonry at the palace, the ashlar style is best represented. This system was in use when the palace was destroyed, so walls constructed in this manner are much better preserved than either of the other cut stone masonry systems so far surveyed. In the extant examples, all components of the system are evident: an exterior face of coursed ashlar, mortises for dowelling horizontal timbers inserted between ashlar courses, and an interior backing of rubble masonry. Anathyrosis insured tight vertical joints between every block, and plaster covered the interior rubble face.

Measurements of the fallen blocks indicate that above the first timber beam, block heights remained consistent throughout an entire course, though course heights varied from 0.37 to 0.44 m.50 Assuming that the wall tipped straight over, at least three courses can be distinguished among the fallout.51 The course closest to the northeast facade is assumed to be the course that mounted the horizontal timber beam. The next two courses also seem to be fairly well represented, but beyond these the blocks are very disordered and courses are difficult to discern. Course 4 (course 3 is the first timber course) had a height of c.0.41 m; course 5, c.0.39 m; course 6, c.0.40 m.52 If the fallen courses have been interpreted correctly, course heights diminish towards the top of the wall but not, apparently, in an orderly fashion (PN I, 50).

All of the ashlar style walls are exterior walls with their cut, outer faces set to confront the natural elements. The southeast wall of Court 3 is inside the Main Building, but since Court 3 was open to the sky, this wall too functioned as an exterior wall. Discussed below are four examples of the ashlar style at the palace. The Northeast Facade of the Main Building The Main Building preserves two ashlar style walls: the northeast facade and the southeast wall of Court 3 (walls 11L and 5E on plan iii). The longest is the northeast facade, which, with the exception of the opening at Room 41, stretches the entire length of the building – approximately 40.19 m (figs 3.32–3.35).48 At its northwest end, the wall is constructed in the ashlar shell wall system (discussed below), but beginning with blocks 11L-4 and 11L-41 (fig. 3.35), the wall continues to the southeast in the ashlar style. From the exterior, the shell wall and ashlar style systems appear identical, so the transition from one to the other must have gone unnoticed by the palace’s occupants and visitors. Plaster covered the interior faces of both systems, at least in the final remodelling phase of the palace, so here too there was little to distinguish one system from the other.

At least three of the blocks in the course lying closest to the northeast facade possess mortises; more might be found if the blocks were lifted. This indicates a second timber beam above the fourth course. Thus there were two timber courses in the facade, separated by only one course of ashlar. The arrangement seems awkward, particularly if the beam courses were used as windowsills or lintels. A similar configuration seems to have been constructed 48

The last c.5.15 m of ashlar facade at the southeast end, corresponding to rooms 48 and 50, was rebuilt by the Greek Archaeological Service (PN I, 213). 49 The stretch of wall outside of rooms 31, 32 and 33 did not come down, by whatever means, until some time after the Bronze Age. The blocks rest on a layer of soil, present also in other areas of the site, which contains post-Bronze Age artefacts (Griebel and Nelson 1993 and 1998. 50 Measurements and observations of these fallen blocks should be considered preliminary. The blocks lie close together and require moving to document them properly, activity that was outside the remit of the MARWP permit. 51 Blegen and Rawson suggested that it was ‘perhaps an explosion of the olive oil stored in Magazines 27 and 32 that, with tremendous force, blew the wall out as a whole’ (PN I, 50). Regardless of whether or not this proposed scenario was possible, if the wall tipped over, we would expect that the course lying closest to the standing portion of the wall to be the next course in the wall, and so forth, following the rows of

From blocks 11L-41 to 11L-57 (figs 3.34–3.35), the wall stands two courses high. The upper bedding surfaces of the blocks of the second course preserve mortises and bedding cuts for a horizontal timber beam that ran from block 11L-40 to anta 11L-20 – a length of 19.38 m. From block 11L-21 to 11L39 (figs 3.32–3.33), only one course still stands. Here too, the upper surfaces of the blocks were cut with mortises and beddings. Because the different placements of the beam represent two phases of construction, I distinguish between the two sections of the facade: 11La (in which the timber beam lay atop the first course of ashlar) and 11Lb (in which the timber beam ran atop the second course).

318

Building Methods for the southeast wall of Court 3 (wall 5E): two horizontal beams left behind wood grain impressions in the mud and rubble backing of the ashlar style wall (PN I, 63). The top beam was only 0.15 m above the lower one, and between them there may have been a course of rubble or cut stone masonry. Similarly, the eastern section of the rear wall of the porch of the megaron at Mycenae preserves two beam slots, one c.0.39 m high, between the first and second courses of ashlar; one above the second course of ashlar (Wace 1949, 76; J. Wright 1978, 140–41, figs 207–08).

The joining of blocks 11L-38A, 11L-38B and 11L-39 was handled quite skillfully, particularly the cutting of block 11L-39 (fig. 3.32). Doorway 46–47 was punched between blocks 11L-37 and 11L-38 of the northeast facade as a part of a later remodelling phase, during which time adjustments were made to the wall (PN I, 203–04). The southeast face of block 11L-37 and the northwest face of 11L-38A both exhibit anathyrosis, indicating that other wall blocks once abutted them. Block 11L-38A’s northeast face was originally exposed. The abutting addition of 11L-39 increased the thickness of the wall and created a setback, but then block 11L-38B was added to the front of 11L-38A, which shifted the setback northwest to the doorway. This is unusual and further supports a hypothesis of remodelling, since it is the only example in the northeast facade of a setback occurring at a doorway. Block 11L-38B is approximately the same length as 11L-38A but was slightly shifted towards the northwest. As a result, the northwest and southeast ends of both blocks are not flush. To adjust for the discrepancy, the mason cut the northwest face of block 11L-39 with a notch to accommodate the east corner of block 11L-38A. The cutting is one of the few examples where two blocks join (11L-38A and 11L-39) without anathyrosis, the two joining vertical surfaces instead being cut square and flat so as to abut flush with one another.

Among the blocks of the tumbled wall are several that lack anathyrosis on one of their vertical joint faces. Instead, they were finished square in relation to the exposed front face. These blocks probably served either to frame an opening, such as a window, or to return the wall a short distance for a recess or setback. A similar short return was cut into block 11L-55 (fig. 3.34). Two blocks, one on top of the other (in relation to their fallen position), have matching square vertical faces, which would put an opening or recess approximately centred in the facade wall of Room 32. Little evidence survives for the upper courses of the other half of the northeast facade, section 11La. The fallen blocks were removed from courts 42 and 47 and stacked, unmarked, in piles around the site. The in situ blocks cannot be fully measured because the plaster floor of courts 42 and 47 abuts the facade and conceals their lower portions.53 This section of the facade is not continuous but is pierced by the formal and monumental entrance 41 and the doorway leading into Room 46. A setback of 0.17 m. occurs between blocks 11L-29 and 11L-30 and corresponds to the intersection of the facade and wall 10J, which separates rooms 43 and 46 (see plan ix). We would expect the setback to have been handled in a similar way as the recess let into block 11L-55, by simply cutting the northwest face of block 11L-30 square in relation to the block’s front face. Instead, it was cut back from the exterior face with anathyrosis, which suggests that the block was reused.

Wood was a part of the ashlar style, and the blocks of the northeast facade of the Main Building preserve some 50 mortises and nine bedding cuts (table 3.4). As indicated by these devices, a timber beam course was fastened to the tops of the blocks near the face of the wall and ran the entire length of the facade.54 According to the placement of mortises and bedding cuts, the beam course followed the line of the facade and returned at several setbacks, antae and at the recess in Room 32. Wall setbacks occur at the junction of blocks 11L-44 and 11L-45 and at blocks 11L29 and 11L-30. Two antae jut out from the wall at Room 40, and the mortises clearly trace the path around the outer, exposed faces of each anta. The transition from the third course level of the beam in section 11Lb to the second of section 11La was handled rather straightforwardly – anta 11L-20 has a notch sunk 0.32 m deep into its south corner. The beam that ran along the first course fit snugly into and came to an end within the notch.

fallen blocks as they lie farther away from the wall base. But if the wall collapsed by some other means, for example by the strong shear force of an earthquake, then the fallen blocks may be entirely jumbled or the fallen courses mixed. If the wall buckled, some courses may even have fallen in reverse order. Because the blocks line up, it appears that fallen courses are present, but whether they represent the original course order remains unknown. 52 J. Wright (1978, 143) reported course heights of 0.39, 0.41 and 0.43 m, measured from the fallen blocks. 53 The facade wall beyond block 11L-39 was reconstructed by the Greek Archaeological Service. 54 The use of the term ‘beam’ is a misnomer here. ‘Beam’ is traditionally used in reference to a structural member spanning and carrying a load over any distance. The beams in the upper courses of ashlar style walls may have spanned window openings and thus served as lintels, but a beam sitting on the first or second course of ashlar would not have spanned anything; it simply rested on the blocks.

Not every block has a mortise, especially the smaller ones (11L-23 and 11L-42). A pattern of mortise placement emerges near the northwest end of the facade outside of Room 31. Here, every other block, beginning with 11L-45 and ending with 11L-51, was cut with a mortise. Beyond this block, all of the rest have mortises. Bedding cuts and setting lines are confined to the portion of the facade between and including blocks 11L-25 and 11L-47. Setting lines are shallow incised lines cut by the masons to mark the placement of mortises and beams. All of the setting lines are straight, so they must have been laid with the aid of a straight edge of some sort. Setting lines were not employed rigorously; not every block that

319

Nelson – The Architecture of the Palace of Nestor

320

Building Methods supported a beam, with or without a mortise, preserves one, and erosion or later stone finishing probably erased some incisions. The setting line on anta 11L-20 (fig. 3.33) is set back 0.05 m from, and runs parallel to, its northeast face. The line originates at the block’s northwest face and stretches approximately 0.64 m, about two-thirds the length of the anta. Two setting lines 0.15 m apart, both running parallel to the face of block 11L-56, define the course and depth of the beam. The northeasternmost line is set back 0.02 m from the front face of the block, which suggests that the beam was not quite flush with the outer face of the block. Block 11L-55 has only a single setting line, which is set back from the front of the block 0.15 m. This incision defined the interior edge of the beam.

cuts on blocks 11L-21, 11L-22 and 11L-24 are difficult to distinguish because of erosion, but the two types of stone finishing can still be seen: rough and semi-smooth. Unlike setting lines for beams, bedding cuts simply provided adequate room for the width of the beam rather than precisely marking its position. From these cuts and lines, the beam in section 11Lb of the facade was between 0.15 and 0.22 m. thick. The beam in section 11La was approximately 0.27 m. thick, measured from the extant wall matrix in the section of the northeast facade outside of room 46. The height of the beams remains unknown. The mortises and bedding cuts greatly help to reconstruct the beam in the northeast facade – it simply followed the path that these devices dictated. Two scholars, J. Wright and M. Küpper, have proposed reconstructions. Küpper’s attempt (1996, 80–84, pls 15–21) is more successful, since his version faithfully follows the beam course as dictated by the mortises and bedding cuts. He also reconstructed the possible joints between the various beam lengths and beam returns throughout its course, though there is no physical evidence of such in Mycenaean architecture. Wright (1984, 26–27, note 4, fig. 10) reconstructed the facade without the change in elevation of the beam from the first to second course, running the beam along the second course of ashlars throughout the entire length of the facade. Such a reconstruction cannot be reconciled with the preserved state of the wall in situ.

Only four blocks preserve setting lines for the positioning and cutting of mortises; these were not inscribed in a recognisably efficient method or pattern. On the top of block 11L-53 the mason etched two sets of parallel lines, one set perpendicular to the other. The intersection of the two sets creates the outline of the mortise. Here, the mason measured four times to mark the position of the mortise. For the other three mortises that show setting lines, the mason seems to have been concerned only with the placement of the mortise in relation to the length of the block, not with the distance that it was set back from the block’s face. Setting lines mark only the southeast and northwest edges of the mortises in blocks 11L-57, 11L-55 and 11L-51; there are no setting lines to measure the width of the mortises. On blocks 11L-55 and 11L-57, the setting lines extend only southwestward beyond the mortise, but on block 11L-51, they extend both southwest and northeast.

Despite the seemingly straightforward reconstruction that is possible, several anomalies in the wall’s construction need reconsideration. Some of these oddities have already been mentioned. First, the beam changes level in its course, from the third course in section 11Lb to the second course in 11La. Second, with one exception, bedding cuts cluster

Beddings for beams were not cut or marked with straight instruments. In all instances, the line between the finished and unfinished surface is irregular and jagged. Bedding

3.32–3.34 (left, top to bottom) and 3.35 (above). Plan and elevation of the northeast facade of the main building (wall 11L), from Room 46 to Room 32.

321

Nelson – The Architecture of the Palace of Nestor Mortise 20A 20B 20C 20D 20E 20F 20G 21 22 24 25A 25B 25C 25D 26A 26B 26C 26D 27 28 29 30A 30B 30C 31 32 33 34 35A 35B 36 44 45A 45B 47 49 51 52 53 54 55 56 57

Block 11-20 11-20 11-20 11-20 11-20 11-20 11-20 11-21 11-22 11-24 11-25 11-25 11-25 11-25 11-26 11-26 11-26 11-26 11-27 11-28 11-29 11-30 11-30 11-30 11-31 11-32 11-33 11-34 11-35 11-35 11-36 11-44 11-45 11-45 11-47 11-49 11-51 11-52 11-53 11-54 11-55 11-56 11-57

Block type anta anta anta anta anta anta anta wall wall wall anta anta anta anta anta anta anta anta wall wall wall anta anta anta wall wall wall wall wall wall wall wall corner corner wall wall wall wall wall wall wall wall wall

Length (m) 0.044 0.044 0.038 0.025 0.058 0.029 0.037 0.037 0.040 0.042 0.039 0.042 0.051 0.041 0.072 0.050 0.044 0.060 0.059 0.044 0.048 0.037 0.038 0.049 0.029 0.048 0.047 0.042 0.054 0.029 0.039 0.035 0.052 0.070 0.069 0.068 0.051 0.067 0.061 0.062 0.050 0.054 0.054

Width (m) 0.031 0.034 0.069 0.026 0.049 0.062 0.049 0.039 0.044 0.043 0.052 0.033 0.042 0.043 0.037 0.057 0.035 0.054 0.031 0.034 0.052 0.046 0.053 0.051 0.041 0.045 0.041 0.055 0.054 0.034 ? damaged 0.030 0.028 0.030 0.030 0.028 0.028 0.026 0.042 0.062 ? 0.034 0.034

Depth (m) 0.065 0.094 0.069 0.063 0.044 0.086 0.077 0.086 0.063 0.052 0.074 0.050 0.060 0.058 0.065 0.061 0.071 0.054 0.066 0.069 0.077 0.067 0.066 0.078 0.068 0.068 0.080 0.092 0.072 0.055 0.020 0.060 0.060 0.060 0.072 0.060 0.074 0.055 0.063 0.066 ? 0.055 0.061

Setback (m) 0.101 0.105 0.163 0.175 0.077 0.023 0.020 0.144 0.142 0.141 0.109 0.112 0.104 0.089 0.144 0.129 0.144 0.306 0.143 0.129 0.148 0.148 0.149 n/a55 0.152 0.145 0.145 0.124 0.148 0.129 0.149 0.130 0.158 0.063 0.152 0.150 0.150 0.152 0.078 0.078 ? 0.068 0.058

Table 3.4 Details of mortises, northeast facade (wall 11L) of Main Building.

in one section of the wall, at Room 31. Third, setting lines also cluster in one portion of the wall, from anta 11L-20 to block 11L-51. Fourth, block 11L-30 has anathyrosis on its northwest face. A setback occurs here, and the face should be cut back square in relation to the northeast face.

exposed northwest face, and their position indicates a face beam running along the wall separating Room 41 from Room 43 (wall 10I). Though the excavators noted a horizontal groove positioned c.0.32 m above the floor in its northwest face (PN I, 179), there is no evidence for a slot large enough to accommodate a beam of any size. Furthermore, if a beam occupied this wall prior to the final destruction of the palace, it could not have fastened into the anta block without some sort of adjustment in its

More peculiarities can be added to this list. Fift h, in its present position as a wall return, block 11L-26 (fig. 3.33) contains more mortises than are necessary. Mortise A corresponds to the exposed northeast face of the wall and was used to fasten a beam running along the northeast face of the facade. Mortises B and C fall near the block’s

55

In the block’s present position, this mortise was not cut in relation to the front exposed face of the block.

322

Building Methods course and height. The height of the block is 0.40 m, 0.08 m higher than the groove.

block 11L-57. Further northwest, block 11L-45 is without a doubt a corner block, as indicated by the anathyrosis on its southwest and southeast faces and the position of its mortises. Thus, the first section of the facade to be built was the stretch from anta 11L-20 to block 11L-45 (i.e., at rooms 34, 33 and 31), where the wall turned southwest and continued. It is also in this section that the setting lines concentrate. These devices occur nowhere else along the facade and indicate a single building phase here.

Mortise D also presents a problem: what beam, if any, did it secure? The mortise is cut near the rear and centre of the block, where it falls along the centreline of wall 10I. Yet there is no evidence for a beam running down the centre of this wall, and no other wall on the Englianos hilltop gives any indication of having had a central beam. Sixth, block 11L-45 has more mortises than expected (fig. 3.35). Its mortises and stone finishing identify it as a corner block, and its northeast and northwest faces are smoothly finished as exposed faces. A dowel set in mortise 11L-45A once secured the face beam running on the northeast face of the northeast facade. Mortise B could have fastened a beam running southwest–northeast, but the presence of this beam and mortise combination is doubtful in its present position. A beam may have been in this position, but it could only have spanned the thickness of the wall, since doorway 31–32 lies immediately to the southwest of block 11L-45; a beam could not have extended further through and into this doorway. Moreover, a beam here would have served as a chase beam to connect the northeast doorjamb of doorway 31–32 to the face beam running along the outer edge of the northeast facade. In that case, however, the mortise would not have been needed because the wall matrix would have prevented the beam from moving in a southeast or northwest direction, and the doorjamb and northeast facade face beam would have stopped it from slipping in the other two directions. In the present state of this wall, this mortise, similar to mortise D in block 11L26, does not have a function and instead indicates that in an earlier phase of this wall, the facade returned at block 11L-45 and continued southwesterly.

At the end of this section of wall, the facade was extended farther to the northwest with the addition of Room 32. The new wall abuts the corner block, 11L-45, and was built using two different building systems: ashlar style and ashlar shell wall construction (discussed below). At the other end, another new addition extended the facade to the southeast with blocks 11L-21 through 11L-24 and anta 11L-25. One of the socle blocks beneath anta 11L-20 was chipped away to accommodate the abutting block 11L-21, and a notch was cut in anta 11L-20 for the second course beam of the new extension. The section of the facade alongside rooms 43 and 46 may also have been added at this time, since the beam here also lies atop the first course of ashlar. This section contains reused blocks: block 11L26, which has more mortises than are necessary, and block 11L-30, which has anathyrosis on its northwestern edge, where it should have been finished square at least in part. It is unknown when these changes occurred, but the masonry clearly shows at least three phases. More importantly, this facade documents a change in building system from orthostate construction to the ashlar and ashlar shell wall styles, with the orthostate building system never appearing again in the architecture of Englianos. Before leaving this facade, two features of section 11Lb deserve comment. Four gaps were left between several blocks in the first course, spaces that Blegen and Rawson identified as weepers (PN I, 51–52; shown in figs 3.34– 3.35). The gaps span the entire height of the first course, from the bedding stones (evident in all four gaps) to the underside of the second course. They do not penetrate the thickness of the wall but stop at the rubble backing of the interior face. Since the ashlar blocks flanking each gap are not cut with anathyrosis, the openings must have been intentionally created. Their function is far from clear, but they could not have acted as weepers, which serve to evacuate accumulated water from the cavity between two faces of a wall. Though the wall is composed of two faces, it is nonetheless solid throughout its entire width and length; there is no space within the wall for water to accumulate. If the gaps had originally pierced the entire thickness of the wall, they may have had some function in connection with the rooms behind the wall. Three of the four gaps align with three rooms (32 (fig. 3.36), 31 and 33) and could have provided a conduit for some material coming in or going out, or both. But the rubble backing of the facade clearly cuts off any connection to the rooms behind, and the fourth gap, at the southeast end of section

Finally, and in contrast to 11L-45, there is one block that should contain at least one, if not two, mortises, but has none. At its northwest end, the northeast facade returns and runs southwesterly (wall 10C; fig. 3.35). Block 11L-40 is a corner block here, its short side facing northeast and its long side northwest; but no mortise was cut near either exposed face. If a timber beam was present at this corner and returned as well (which the rest of the wall seems to indicate), then one would expect mortises here to secure it; this is particularly important in a corner position, where the stresses on a beam would come from several more directions than in a normal face beam situation. Blegen and Rawson concluded that the entire facade was built in a single building phase (PN I, 47), but all of these anomalies make sense only if the wall was constructed in several phases. It was explained above that anta 11L20 was part of an orthostate wall that predated the entire northeast facade. When the orthostate wall was destroyed, by whatever means and for whatever purpose, all that remained in situ was this anta and the socle blocks underneath and to the northwest of it. The northeast facade was joined to this, starting with the abutting

323

Nelson – The Architecture of the Palace of Nestor 11Lb, corresponds not to a room but to the doorway between rooms 33 and 34. Nevertheless, the facade blocks were reused in one of the later remodellings of the palace, and the gaps may have been intentionally closed up then. For now, their purpose is difficult to determine.

supported only at its ends would have been sufficient to span the distance. Furthermore, the recess reduced the thickness of the wall. If a window had been placed within it, the thinner reveal would have allowed more light into the room. Recesses may have had an aesthetic value by creating strong shadow lines on the facade, as has been suggested in respect to Minoan architecture (Graham 1960; 1987, 162–64), but they could also have served a much more practical function.

The second odd feature of section 11Lb is a shallow recess let into the second course of blocks outside of Room 33 (fig. 3.34). The 2.49 m-wide recess is set back 0.18 m from the exterior face of the wall. It was carefully planned and executed, with the southeast face of block 11L-52 cut back square rather than with anathyrosis, and block 11L55 notched at its northwest end to create the southeast corner of the recess. Blocks 11L-53 and 11L-54 are typical ashlar blocks, no different in size than the others, and are simply pushed back 0.18 m. from the outer face of the wall. Mortises cut into the tops of these blocks clearly trace the wooden beam running around the recess. More questions arise than can be answered regarding the function of the recess, but some comments can be made. The excavators associated the feature with a possible window opening in the wall (PN I, 52). If the recess did form the lower parts of a window,56 then it appears that only one window pierced this section of the facade, though other windows, probably small, may have been located higher up in the wall.

The Southeast Wall of Court 3 (Wall 5E) A single course of four ashlar blocks forms the base of the southeast wall (wall 5E) of Court 3 (plans iii and ix, and fig. 3.37). At its southwest end, anta block 8LD.a turns the corner into wall 8LD, and at its opposite end, anta 5E-1 terminates the wall’s course at Stairway 54. The blocks vary in length, but their finished upper surfaces rise 0.38 m above the plaster floor (elevation 192.72 masl). The full height of the individual blocks cannot be ascertained because the plaster floor of Court 3 runs flush up against their exposed faces. The backing of the wall varies. Behind blocks 5E-1, 5E-2 and 5E-3 is the pier-wall building system (described below), while rubble masonry lies behind the remaining two blocks. The wall’s depth also varies: the northeast part is c.0.80 m thick, the southwest c.1.28 m.57

The two notches sunk into the first course of ashlar blocks outside Room 46 may also have been parts of a window opening. These devices and the recess at Room 33 are architecturally distinct, however, and suggest two different forms of window construction. The two notches imply vertical posts, but no such cuttings are present in the recess. In the section of the facade alongside Room 46, there may have been a structural necessity for vertical supporting members that were secured by the dowel holes in the notches, such as in a long band window broken into smaller openings. The length of a band window usually requires intermediary supporting members for its lintel, and the two notches may have corresponded to two vertical mullions in a band window consisting of three openings. If a window did occupy the Room 33 recess, and assuming that the size of the opening directly matched the size of the recess itself, then a lintel of moderate size

The joints between the blocks are intricate and demonstrate the mason’s skill not so much in cutting as in planning. All of the wall blocks exhibit anathyrosis. Block 5E-3 is slightly unusual and probably reused. Its northeast face was cut with reverse anathyrosis: rather than the angle between the front face and the side being less than 90 degrees (acute), it is obtuse. Its southwest face was cut back square for c.0.20 m and then cut with anathyrosis. Block 5E-5 (fig. 3.38) is a corner block; thus its northeast face was cut with anathyrosis and its southwest face was cut square. At the rear of the latter face, a notch accommodates the north corner of anta 8LD.a. At the other end of the wall, another notch was cut into anta 5E-1 to receive the northeast face of block 5E-2. Every block contains mortises, and as to be expected, the two antae have more than the other blocks: anta 5E-1 has three mortises and anta 8LD.a has four (dimensions are given in table 3.5). Two blocks, antae 5E-1 and 5E-5, have beddings on their upper surfaces (perhaps the two most finely cut beddings at the palace). The placement of mortises and bedding cuts indicate that a horizontal wooden beam originally ran the entire length of the wall 56

Graham (1960; 1987, 162–64) came to the conclusion that recesses for the most part indicate the presence of a window in the Minoan palaces of Knossos, Phaistos, Mallia and Gournia – if not in the fi rst storey, then surely in the second. For Minoan windows and their construction details, see J. W. Shaw 1973a, 174–83. 57 It is possible that the rubble backing of the southwest portion of the wall continues beneath the raised plaster floor in the north corner of Room 55 (see plan ix).

3.36 So-called ‘weeper’ outside Room 32, northeast facade (wall 11L) of Main Building, from the northeast.

324

Building Methods

3.37 Plan and elevation of southwest wall (wall 5E) of Court 3, Main Building.

and over both antae. The beam course returns twice as indicated by the bedding cuts: at the west corner of anta 5E-1 and at the west corner of 5E-5. Based on the mortises and bedding cuts, as well as the preserved impressions of wood grain in the wall matrix above the ashlar blocks, the excavators assumed the presence of a horizontal wooden beam spanning the entire length of this wall and continuing over both antae (PN I, 63). Küpper (1996, 85, fig. 23) concurred and offered a reconstruction of the beam course, but neither he nor Blegen and Rawson properly account for the state of the wall. The wall matrix makes it impossible for a continuous beam course to have existed between anta 5E-1 and block 5E-2. If there were beams in this wall, there must have been two independent systems: one beam placed atop anta 5E-1 and another over the ashlar blocks. One would expect the beam on anta 8LD.a to return southeastward and run along wall 8LD’s southwest face, but, again, wall matrix blocks its course. Furthermore, none of these scholars considered mortise H in block 5E-5 (figs 3.37–3.38). This mortise defies explanation. It obviously was meant to receive the dowel of some sort of wooden armature placed above, but it was cut into the upper surface of the block in an area that, according to its rough finish, should have received the rubble and mud backing of a typical ashlar-faced wall. Küpper (1996, 85) thought that the mortise was incomplete, because its interior surface was only roughly finished. This is unlikely, since it would be the only example of a block at the palace left unfinished by the builders. The block may have been reused, but even if this were true, the

3.38 Ashlar block 5E-5, southwest wall of Court 3 (wall 5E), Main Building.

325

Nelson – The Architecture of the Palace of Nestor Mortise A B C D E F G H I J K L

Block 5E-1 5E-1 5E-1 5E-2 5E-3 5E-4 5E-5 5E-5 8LD.a 8LD.a 8LD.a 8LD.a

Block Type Anta Wall Wall Wall Wall Wall Wall Wall Anta Anta Anta Anta

Lenth (m.) 0.040 0.041 0.065 0.046 0.059 0.048 0.043 0.036 0.038 0.025 0.033 0.042

Width (m.) 0.046 0.044 0.017 0.043 0.037 0.037 0.044 0.031 0.048 0.050 0.043 0.035

Depth (m.) 0.048 0.049 0.038 0.038 0.053 0.068 0.050 0.000 0.065 0.052 0.065 0.059

Offset (m.) 0.109 0.107 0.702 0.206 0.166 0.160 0.169 0.399 0.137 0.127 0.121 0.079

Table 3.5 Details of mortises, southwest wall (wall 5E) of Court 3, Main Building.

placement of mortise H is very awkward, since it was cut nearly in the centre of the block and not near any exposed edge, as the general rules for mortises would dictate. A third possibility is that the mortise was used to fasten a vertical post. This would be unusual, but the placement of mortise H at the corner of a bedding cut indicates some relationship with the wall return, and a vertical post might help turn the corner here. As noted above, the wall matrix above the in situ ashlar course preserves wood grain impressions left by rotted or burned away horizontal timbers. The two sets of impressions mark the locations of two beams: one resting directly on top of the ashlars and another set slightly higher up. Apparently, the two beams were separated by a course of ashlar or rubble masonry.58 A vertical post fastened at the corner with mortise H may have been used to secure the upper beam, and perhaps this system extended farther up the wall with more horizontal members.

except at the very northwest end of the wall, where the pattern is broken with mortises cut into adjacent blocks, including anta 18L.b. Unusually, all of the wall blocks cut with mortises have two each, rather than the usual one. Furthermore, each mortise of a pair falls into a separate group based on size and location. The A-group mortises are larger and are cut near one end of the block. The B-group mortises are smaller than their counterparts and lie near the centre of the block. Each and every mortise is cut near the exterior exposed face of the wall and so fastened a beam running along this face. In comparison to the other ashlar style walls, the number of mortises here is far greater than normal and seems quite excessive for their function. But if the wall was reused and the configuration of the beam changed, then a new set of mortises would have been needed.59 It was mentioned above that wood may have been more valuable than stone, and that the mortises were adjusted to fit the available beams. If this wall was reused or remodelled, the dimensions of the wooden timbers that were to be incorporated must have been known, and the stone, particularly the mortises, cut to fit them.

The Northeast Facade of the Southwestern Building (Wall 18L) The northeast facade of the Southwestern Building is also partially constructed in the ashlar style (plans xiv, xvii and xix, and fig. 3.39; see also Nelson 2003). Nine ashlar wall blocks and two antae remain. At the southeast, the wall is terminated with a wall-end anta block, 18L.a, which has three exposed faces. At the opposite end is the T-shaped anta block 18L.b. The facade once continued northwesterly but was robbed out at a later date (its foundation of bedding stones can be seen in plan xix; see also PN I, 249). All of the wall blocks exhibit anathyrosis except 18L-7, the southeast face of which is cut back square from its front face. The backing of the wall is the usual rubble masonry faced with plaster. The thickness of the wall, c.1.02 m, remains uniform throughout its entire length, although anta 8L.a is slightly thicker at 1.06 m.

Based on the breadth of the bedding cuts, the beam here was no thicker than 0.29 m. The beddings cluster at the southwest end of the wall, where they are cut into blocks 18L-8, 18L-9, 18L-10 and anta 18L.a. The anta bedding is U-shaped; thus the beam that ran along the northeast face of the facade returned twice to wrap around the wall’s southeast end. The beam does not appear to have continued along the southwest, interior face of the wall. Sections 9 and 10 of the Southwestern Facade of the Southwestern Building Sections 9 and 10 intersect to form a return angle in the southwestern facade (plans iii, xvi and xvii, and figs 3.5–3.8 and 3.40). Both sections are massive: section 9 measures c.3.10 m thick (by far the most substantial wall

A horizontal beam ran along the top of the first course at the wall’s northeast face. If setting lines were used, they are no longer evident, but 16 mortises and several beddings are cut into the blocks (dimensions are given in Table 3.6). With one exception, the placement of the mortises follows a simple pattern: they are sunk into every other block

58

Blegen and Rawson did not report any fallen ashlar in this area of the court. 59 Küpper (1996, 86–87, pls 26–27) agreed, but suggested that there were two phases of half-timbering rather than a rebuilding or remodelling.

326

Building Methods Mortise A B A B A B A B A B A B A B C D

Block 18L.b 18L.b 8-2 8-2 8-3 8-3 8-5 8-5 8-7 8-7 8-9 8-9 18L.a 18L.a 18L.a 18L.a

Block Type Anta Anta Wall Wall Wall Wall Wall Wall Wall Wall Wall Wall Anta Anta Anta Anta

Length (m) 0.087 0.054 0.062 0.056 0.068 0.063 0.076 0.059 0.071 0.058 0.076 0.077 0.096 ? 0.038 0.026

Width (m) 0.048 0.027 0.041 0.038 0.045 0.029 0.046 0.029 0.047 0.031 0.057 0.038 0.033 0.039 0.028 0.037

Setback (m) 0.083 0.150 0.092 0.134 0.084 0.156 0.090 0.144 0.095 0.145 0.083 0.141 0.110 0.321 0.632 0.335

Table 3.6 Details of mortises, northeast facade (wall 18L) of Southwestern Building.

built at Englianos), and section 10, though not very well preserved, seems to have had roughly the same thickness. The exterior of both sections is faced with ashlar blocks. Two courses remain in each section, though they have been heaved and pushed out of position, perhaps the result of earthquake, erosion and/or modern plowing. (At the foot of the walls are deep ruts left behind by flowing rainwater that probably disturbed the integrity of the walls.) Like the ashlar style walls so far surveyed, all of the blocks of these two sections exhibit anathyrosis, except for blocks 17.10-7 and 17.10-8. Both of these are damaged, and their heights cannot be restored, but their abutting joint faces are cut back square.

in length and thickness, but those in the same course maintain the same height: 0.45 m for the first course and 0.47 m for the second (table 3.7). The corner between the two wall sections was handled straightforwardly. The outer, east vertical edge of block 17.9-6 simply abuts the outer, west edge of block 17.10-1. In the now missing upper courses, the walls may have bonded together by overlapping blocks. Of the two surviving courses in section 10, the uppermost is damaged and only two blocks remain (figs 3.7–3.8). The upper surface of block 17.10-8 is broken and ragged, probably as a result of deep plowing. Its neighbor, block 17.10-7, may not be in it original position, because a thin layer of soil separates it from the one (17.10-2) below. The second course is better preserved, with five wall blocks and a stair tread at its southeast end. The section continued towards the southeast, but as a stairway or gateway. Block length and thickness varies, and unlike section 9, heights differ for individual blocks. This course seems to have been levelled in a manner similar to the levelling of section 11Lb of the northeast facade of the Main Building, with a bedding of small stones laid directly on bedrock.

Section 9 runs perpendicular to the topography of the hilltop – its foundations and ashlar courses step down towards the southwest (figs 3.5–3.6). Although the wall was clearly destroyed at its southwest end, it probably once continued in that direction.60 The first course contains two blocks in situ and the second, four. The blocks vary 60

Blegen and Rawson (PN I, 279–80, figs 416 and 417) reconstructed a connection between sections 8 and 9 of the facade, but there is no physical evidence of such.

3.39 Plan of the northeast facade (wall 18L) of the Southwestern Building.

327

Nelson – The Architecture of the Palace of Nestor Mortises are cut into blocks 17.9-3, 17.9-4 and 17.9-5 of the second course in section 9, and in section 10 on blocks 17.10-1 and 17.10-4 of the first course, indicating that a horizontal beam rested above them. The beams of both sections no doubt joined at the corner. In section 10, block 17.10-8 obstructs the beam’s course, making it impossible to place a continuous beam here that runs the entire length of the wall to the stair tread.61 Since block 17.10-8 is not cut with anathyrosis, a beam placed to either side of it may have stopped and abutted this block. Ashlar Shell Construction The palace preserves only two ashlar shell walls. Some 3.20 m southeast of the north corner of Room 32, the northeast facade (wall 11L) of the Main Building exhibits a change in building method from the ashlar style to ashlar shell wall construction (plan x and figs 3.1 (above) and 3.41). The wall bonds and forms a return with wall 10C, the northwest wall of Room 32. Before the addition of Room 27 at the northern corner of the Main Building, these two walls originally formed an exterior corner. Two courses of wall 10C still stand at its northwest face (fig. 3.1 above). The entire first course of its southeast face remains in situ, but only a single block from the second course is preserved in this face.62 The northwest end of wall 11L preserves one

Block

Length (m)

Height (m)

Course

17.9-1

0.92

0.45

1

17.9-2

1.03

0.45

1

17.9-3

0.97

0.47

2

17.9-4

0.76

0.47

2

17.9-5

0.93

0.47

2

17.9-6

0.37

0.47

2

Block

Length (m)

Height (m)

Course

17.10-1

0.75

0.35

1

17.10-2

0.67

0.33

1

17.10-3

0.86

0.33

1

17.10-4

0.57

?

1

17.10-5

0.44

0.33

1

17.10-6

?

0.37

1

17.10-7

0.49

?

2

17.10-8

0.88

?

2

Table 3.7 Dimensions of ashlar blocks in wall 17L, sections 9 (above) and 10 (below).

61 Küpper (1996, 86, pls 24 and 25) regarded these blocks as later building activity and reconstructed a beam on the first course of ashlars of section 10. He also identified two bedding cuts, one on block 17.9-3 and the other on 17.10-1. They look to be plow marks.

3.40 Plan of southwest facade (wall 17L) of Southwestern Building, sections 9 and 10.

328

Building Methods exterior corner block of the second course and both the interior and exterior face of the first course. The interior of both walls was covered with plaster in the palace’s last phase of occupation.

Minoan palatial architecture. In the palace at Knossos, the Middle Minoan II east wall of the lightwell in the Queen’s Megaron is a shell wall (Evans 1964, vol. III, fig. 249). The blocks are thick, leaving little space for the in-fi ll matrix. Since the lightwell was open to the sky, the interior face of the wall was as exposed to the elements as the exterior, which accounts for the shell wall construction in this room. Shell walls were also used, sparingly, at Mallia (rooms iii, 5 and 6) and Phaistos (north and west walls of room 49; for both sites see J. W. Shaw 1973a, 104 and note 2). In a slight variation of the shell wall system, the MM III west facade of the palace at Knossos is an orthostate shell wall with its outer faces held together by dovetail clamps (ibid., 88–90). The two faces of the ashlar shell walls at the Englianos palace do not appear to have been held together in a similar way.

The first course of wall 10C returns at its southwest end and indicates that an ashlar shell wall originally formed the southwest wall of Room 32 (fig. 3.1 above). This wall was replaced in a later remodelling or rebuilding phase with wall 9L, which was built with pier-wall construction. Wall 10C was also altered; at least one block of its second course was pushed out of position to accommodate the new pier-wall construction built on top of the wall. Two ashlar blocks in the second course of wall 10C (10C9 and 10C-13) have mortises that once fastened a timber beam running along its northwest face (fig. 3.3 above). They follow the same general rules for mortises and have the same characteristics as those on the northeast facade (wall 11L). The oddness of the corner block (11L-40) of the second course with its missing mortises has already been noted above. Since only one block of the interior second course remains, it is not known if a timber course ran along the interior face of the wall.

Pier-wall Construction Blegen and Rawson believed that the interior walls of the Main Building and of some of the ancillary buildings were built with a heavy timber framework of evenly spaced vertical and horizontal members, the latter of which both spanned the thickness of the wall (chase beams) and ran along its face (face beams). Altogether, the system of beams and posts were thought to form a three-dimensional wooden grid into which rubble and mud were packed (fig. 3.42). This system would then have been concealed with plaster and sometimes frescoed; the decoration may have reflected the underlying wooden members (fig. 3.43). They called the building system by a modern Greek term, xylodesia (Blegen 1965, 118; in PN I, xylodesia is referred to as the ‘Mycenaean Style’ of construction). Good examples of it are the megaron walls (plan xi): vertical moulds and impressions in the mud and rubble, and in some instances the surviving plaster, seem to indicate vertical timbers spaced c.0.80–1.00 m apart along the course of the walls, and horizontal moulds and impressions appear to show chase and face beams connecting the posts (fig. 3.50 below).

The palace at Englianos stands alone among other mainland structures for its use of the ashlar shell wall building system,63 but close comparisons can be found in

The excavators concluded that an intense fire destroyed the palace (PN I, 32). The fire was believed to have burned all of the wood out of the xylodesia system and left behind only moulds in the stone, mud and plaster, along with occasional impressions of wood grain, as in the mud matrix of the wall above the ashlar blocks of the southeast 62

In the state plan of this wall, there appears to be a second ashlar block just to the southwest, but this is part of the later remodelling in which ashlar fragments were often used to line chases in the pier-wall building system, discussed below. 63 Hult (1983, 50–51) reported that the north wall of the court leading to the megaron at Mycenae is constructed in the ashlar shell wall building system. Th is wall is a retaining wall, however, and does not have an interior face (in this case, the north face) of ashlar blocks, but one of rubble (Wace 1949, 72, fig. 4). The wall partially collapsed in the winter of 1997, bringing to light its rubble fi ll. See also Küpper 1996, 55.

3.41 Northeast facade (wall 11L) at Room 32 of Main Building, from the southeast.

329

Nelson – The Architecture of the Palace of Nestor wall of Court 3. This fire was said to have been so intense and of such duration that it ‘melted much of the stone into lime; the latter (lime) flowed over the slots in which the beams were laid, covering the ashlar blocks below with a fused mass of migma’ (PN I, 49). Furthermore, in reference to the use of wood in the doorway separating rooms 5 and 6,‘[stone and mudbrick] must have been fused into a molten mass by the tremendous heat of the fire and flowed down from above the [wooden] casing. When it cooled, it hardened into something very much like concrete and extremely difficult to excavate’ (PN I, 77). 3.44 Chase 36, northeast wall (wall 8L) of Room 6, Main Building (southwest face at norwest end), from the southwest.

The evidence of timber in the supposed xylodesiaconstructed walls at the time of the palace’s destruction is less than convincing, however. Limestone (calcium carbonate, CaCO3) melts or becomes molten at a temperature of 1339º Celsius at 103 atmospheres (103 x the earth’s atmosphere).64 No fire, fuelled only by wood and possibly olive oil (PN I, 50, 92, 158), could have reached this temperature, nor did the earth’s atmosphere become so heavy. If the palace was destroyed by fire, the intensity of the flames could not have melted the limestone; at most, the flames might have turned the stone into lime. The process is more commonly known as calcination; the result is calcined stone, or quicklime (CaO), which has a white crumbly surface (CRC Handbook of Chemistry and Physics, 1993, 4–7; Adam 1994, 65–66). The excavators mentioned calcined stone regularly throughout their descriptions of the architecture, and it can still be seen today.65 During the process of calcination the stones do not lose their original shape, and the resulting lime does not flow; rather, the colour of the stones lightens, and they lose weight with the release of carbon dioxide. Thus,

the so-called migma identified by the excavators, which is still visible in almost all the moulds today (e.g. fig. 3.44), cannot be the result of a fire but must be considered original building material produced as such and put in place during the construction process. Few of the moulds are empty or have sufficient space to have once housed portions of a timber framework. The moulds in the walls of Room 6 provide good examples. Of the 45 chases (the interstices that span the thickness of the wall between piers in pier-wall construction) spaced approximately every 0.80 m. in its four walls (drawn schematically in fig. 3.45; see also plan xi), three are completely empty (32, 34 and 39, the last shown in fig. 3.46). These could have been occupied by vertical posts and horizontal beams, but the rest are almost wholly filled with mud mortar and small stones. In 16 of these slots (1–3, 6, 8, 11, 13, 29, 32, 34–39, 43) there is adequate space for one vertical post at the interior face of the wall. But if vertical posts actually stood in them at the time of the destruction, the posts would have had dimensions more like a plank set on end than a post. For instance, in chase 38 the vertical post would have measured no more than 0.14 x 0.26 m. Any load placed on this member, which would have been very long in comparison to its crosssection, would have caused it to buckle outward, blow the plaster off the wall and eventually collapse. In chase 36, the post could have had dimensions no greater than 0.11 x 0.23 m. Thus, of these 45 slots, there are only three chases with sufficient available space for load-bearing vertical posts with adequate, square cross-sections. The walls could not have been built using the xylodesia method if the vertical wooden members in this system were load-bearing, as implied, for there is simply not enough room in the chases for a sufficient number of posts. Instead, it is obvious that the mud and rubble functioned as the major load-bearing elements in these walls.

3.42 Reconstructed xylodesia building system (after PN I, 78–79).

64

CRC Handbook of Chemistry and Physics, 1993, 4–7. My thanks to Craig Glenn of the Department of Geology and Geophysics at the University of Hawaii for his insightful comments. 65 One might also expect vitrification of the mud and plaster to have occurred during the intense fi re, but this was not reported by Blegen and Rawson and is not seen anywhere in the palace remains today.

3.43 Reconstruction of fresco decoration, northeast wall of Room 64 (after PN II, ID64, 39–38C64, 22H64 and 1F2).

330

Building Methods

3.45 Schematic pier and chase plan, Main Building.

The evidence for horizontal timber members is also less than convincing. As noted above, there are two types of horizontal beams in a xylodesia system: those that span the thickness of the wall (chase beams) and those that run along and are embedded in the wall faces (face beams). The evidence for chase beams has been misinterpreted. A glance over the tops of the megaron walls quickly distinguishes the chases from their flanking piers because more often than not the matrix of mud and small stones within the chases is lower than that of the flanking piers (fig. 3.47). In their current state of preservation, the chases would seem to indicate a missing beam – the difference in elevation would be corrected with the addition of a (now supposedly missing) chase beam. But this is not the case if we take into account the preserved elevations of the walls over their entire lengths.

depth was only 0.70 m.’ (PN I, 89). The elevation at the intersection of the walls at the south corner of Room 6 is 193.44 masl, and at its west corner, 192.96 masl. Hence, a direct relationship exists between the height of the extant walls and the original pre-excavation topography of the Englianos hilltop. Therefore, the present elevations of the chases depend not on any missing material (i.e., chase beams) but on two other factors. First, the material in the chases – small stones, mud and sand – is not as durable as that of the piers and was more susceptible to erosion and the excavator’s pick. Second, like the piers, the elevations of the chases correspond to the natural and manmade topography of the hilltop prior to excavation. For example, the northeast wall of Room 6 descends from the southeast to the northwest. The chases in this same wall descend as well, the highest measuring 193.51 masl at the southeast end of the wall and the lowest 193.00 masl at the northwest (table 3.8). This wall once ran perpendicular to the original, pre-excavation contours; the graph in Table 3.8 shows the descending relationship between pier and chase height. The southwest wall is similar, with the wall and chases descending in the same direction. Conversely, the chases are all approximately the same height in the rear, northwest, wall of Room 6 (wall 5A; fig. 3.47 and table 3.9), because this wall once ran parallel to the preexcavation contours.

The highest elevation among the extant walls is 193.66 metres above sea level, measured at the intersection of the walls at the north corner of Room 4. From this point the elevations of the walls and other built features drop in every direction. Thus, the highest point of the hill before excavations commenced was over this intersection; from this point the contours gradually dropped in all directions (fig. 3.48). Blegen and Rawson described the strata in Room 6 as follows: ‘The deposit covering the Throne Room had a maximum depth in its southeastern section of 1.25 m., but it grew much shallower, following the descending slope of the hill toward its northwestern end, where the

331

Nelson – The Architecture of the Palace of Nestor

3.46 Chases 38 and 39, and pier 37, in northwest wall (wall 5A) of Room 6, Main Building, from the southeast.

Evidence for face beams also exists, but they do not seem to have been placed and used as consistently and systematically as originally thought. For the most part, the excavators relied on uniform breaks in the plaster wall surface as evidence for underlying but now missing face

beams, since it was assumed that the original slots were fi lled with flowing migma at the time of the fire. Few real slots can be observed. If we take Room 6 as an example again, Blegen and Rawson noted horizontal face beams in both faces of the southwest wall, wall 4L (PN I, 78). On the northeast face, they reconstructed a beam at floor level and another 0.75 m above. On the southwest face, they placed a beam at floor level and two more higher up: at 0.75 m in the southeast section of the wall and at 0.60 m in the northwest section. For the two proposed beams at floor level, there are no empty slots or any other convincing gaps or suitable openings, and only a few piers preserve actual slots for wooden beams placed higher up on the wall. On the southwest face of the southwest wall, in piers 22, 23 and 24 (fig. 3.49), there remains a rectangular slot of sufficient size for a beam with a cross-section of c.0.11 m high by 0.15 m deep. The slot originates at chase 23 and continues on to pier 25 moving northwest, bringing the total length of the beam to approximately 3.40 m. At the northwest end of the same wall, and on the same face, the slot noted by Blegen and Rawson at 0.60 m above the floor could have accommodated a face beam c. 0.15 m deep (fig. 3.50). This beam could only have spanned piers 31, 32 and 33, which gives it a total length of c.3.28 m; pier 30 rises to a higher elevation and does not preserve any signs of a face beam at the same level. On the opposite side of Room 6, a face beam slot is preserved 0.63 m above the floor in piers 3 and 4 of the northeast wall (8L).66 The height of the slot cannot be determined because the wall is missing above the slot,

3.47 Northwest wall (wall 5A) of Room 6, Main Building, from the southwest.

66

Blegen and Rawson measured the slot at c.0.70 m above the floor (PN I, 78).

332

Building Methods

3.48 Pre-excavation topography over the Main Building (contour elevations shown as metres above sea level).

333

Nelson – The Architecture of the Palace of Nestor Room 6, Northeast Wall

but its depth varies from 0.23 to 0.27 m; it could have held a beam with one dimension equal to or less than the slot’s minimum depth (fig. 3.51). This beam may have originated on pier 1 and continued over pier 2, but there is no evidence for a slot. The beam could not have spanned much farther southeastward over pier 5, since this pier’s elevation rises without a slot on its face. On the other face of the same wall, another face beam appears to have spanned piers 5 and 6 at height of c.0.83 m. The depth of this slot is 0.22 m, and the beam it housed could have extended farther northwest, but because of the decrease in elevation, its length is unknown. The beam could not have extended southeasterly over pier 7, since the pier is higher in elevation. Thus, the beam would have been relatively short in comparison to the length of the wall.

193.80 193.60 193.40 193.20 193.00 192.80 192.60 192.40

Piers Chases

1

3

5

7

9

11 13

Chase/Pier No.

Going beyond Room 6, there is a face beam slot in the section of the rear wall of Room 4, northeast of the door in the wall’s southeast face (fig. 3.52). This slot clearly demonstrates that the beam was relatively short. The beam appears to have spanned the wall from the northeast doorjamb over three piers, stopping at the northeasternmost chase. The beam would have been no more than 3.12 m long; the slot dictates a beam with a depth of c.0.32 m. The slots surveyed in rooms 4 and 6 indicate the presence of only a few beams of relatively short lengths, placed in no consistent manner within the walls at the time of the palace’s destruction. An exception is the northwest face of the rear wall (5A) of Room 6, which preserves a face beam slot c.7.82 m in length and 0.25–0.33 m deep (fig. 3.53). The beam originally abutted the southwest face of the northeast wall of Room 24 and continued as far as chase 39. A single continuous beam of the length prescribed by the slot was probably not used, but rather one composed of multiple short timbers.

Number

Chase Elevation (masl)

1 2 3 4 5

192.88 192.83 193.00 193.01 193.04

Pier Elevation (masl) 193.00 193.02 193.03 193.07 193.25

6 7 8 9 10 11 12

193.17 193.23 193.22 193.30 193.35 193.34 193.36

193.41 193.47 193.53 193.57 193.48 193.48 193.51

13

193.37

Table 3.8 Pier and chase elevations, wall 8L, Room 6, Main Building.

Room 6, Northwest Wall 193.20 193.00 192.80 192.60 192.40

From the evidence recounted above, it is clear that a heavy timber framework was not present in the palace walls at the time of the destruction. Based on the appearance of these walls, I suggest the term ‘pier-wall’ construction rather than xylodesia, which gives a false impression of the building method.

Piers Chases

34

36

38

40

42

44

Chase/Pier No.

Number

The majority of the interior walls in the Main Building, a few in the Southwestern Building, and three exterior walls of the small building containing rooms 60 and 62 are pier-walls. All of these walls, as they appeared when they were originally uncovered and as they appear today, are somewhat deceptive. In their original, finished state a thick, concealing plaster coat sheathed both faces of the walls. When the palace was excavated, much of the plaster had either already fallen off or, if painted, was purposely removed by the excavators to the museum in Chora, though some plaster still remains on the lower portions of the walls and testifies to their original condition. For the most part, what remains today is the part not meant to be seen – the wall matrix, consisting of a series of close-set

34 35 36 37 38 39 40 41 42 43 44

Chase Elevations (masl)

192.87 192.86 192.87 192.61 192.91 192.89 192.91 192.88 192.85

Pier Elevation (masl) 192.96 192.92 192.95 192.99 193.01 192.90 192.92 192.96 192.95 192.91 192.93

Table 3.9 Pier and chase elevations, wall 5B, Room 6, Main Building.

334

Building Methods

3.49 Southwest face of southwest wall (wall 4L) of Room 6, Main Building, from the southwest.

3.50 Southwest face of southwest wall (wall 4L) of Room 6 at Room 18, Main Building, from the southwest.

335

Nelson – The Architecture of the Palace of Nestor rubble and mud piers separated by filled, c.0.20 m-wide interstices. The original plaster coat that covered the faces of the pier-walls rendered the alternating piers and gaps indistinguishable.

consolidated the walls both by fi lling gaps between the stones and by preventing individual stones from shifting and moving. In some rubble walls on the site, it is difficult to see the mud between courses. But in the piers, the mud amounts to 20–40 percent of the matrix, substantially more than for typical rubble walls. In the interstices and the piers that form doorjambs, the mud content is as much as 60–70 percent of the matrix, and often the large stones of the piers are entirely concealed by the mud. In the piers, stone does not rest on stone; instead, a thick layer of mud (sometimes up to 0.20 m) separates them. The mud is clearly a strong material with mortar-like qualities, capable of bearing structural loads, because the settling of the matrix during the building of the pier-walls would otherwise have pushed most of the mud out from between the stones.67

The piers consist of rubble and ashlar block fragments set in a mud mortar. The matrix of the interstices is mud and small stones, which vary in size from pebbles to fist-sized cobbles (much smaller than the stones in the piers). The large amount of mud between stones in both piers and interstices is remarkable. A typical rubble wall at Englianos consists of stones stacked one on top of the other. Mud, laid in thin layers between courses,

Samples of wall matrix from pier-walls and rubble walls were analysed by inductively coupled plasma atomic emission spectrometry (ICP-AES) for elemental composition (Appendix C). As shown in table C.4 in Appendix C below, the mud-like material of the piers and the interstices generally contains a higher proportion of calcium carbonate (CaCO3) than the mud used in rubble construction, ranging from 30 to 70 percent calcium carbonate in the piers, to a generally higher amount in the material of the interstices and the piers forming

3.51 Southwest face of the northeast wall (wall 8L) of Room 6, Main Building, from the southwest.

3.52 Northeast section of northwest wall (wall 5C.1) of Room 4, Main Building, from the northeast.

3.53 Northwest wall (wall 5A) of Room 6, Main Building, from the southwest.

336

Building Methods hydroxide, Ca(OH)2). In the next step the lime was mixed with sand, clay and/or pebbles. When the mixture was applied to walls and allowed to dry, it absorbed carbon dioxide (CO2) and reverted back to calcium carbonate. During wall building, the addition of lime to mortar would seem very appropriate because of its ability to harden and bind – it literally turns back into the stone from which it came.74

doorjambs; in the mud of the rubble walls, the calcium carbonate content ranges from 25 to 42 percent. The contrast is particularly evident when comparing the sample taken from the later rubble wall separating Room 18 from Room 22 (sample 51) and those from the pier-built walls in the same building (samples 41, 46, 49 and 50). It seems that calcium carbonate was intentionally added to the mud of the pier-walls. The Mycenaeans were familiar with this material because it was one of the principle components of the plaster with which they covered their walls and floors. Mabel Lang’s analysis shows that the plaster in the palace is composed mostly of calcium carbonate.68 Similar results were found in plasters from Tiryns and Mycenae.69 Many of the floors in and around Mycenae, including at the palace,70 in two of the houses within the walls,71 and in tholoi, were surfaced with a lime (calcium)-based material often referred to as ‘cement’.72 Tiryns, too, has floors laid with this material.

67

I am unaware of a defi nition of mortar used by scholars of Mycenaean architecture. I use the term to refer to a material that has the ability to bond with strong adhesive qualities and is capable of bearing loads. In this sense, the mortar must have load-bearing capabilities similar to that of the material it binds. I know of no structural studies of masonry consisting of mud and stone, or of mud as a building material, but it is a given that the strength of masonry depends on the mortar used to consolidate it (Salvadori and Levy 1981, 29). The mud, then, should have physical characteristics equal to or stronger than the stone. See also G. Wright 1992, 409. 68 Lang did not give exact percentages but referred to ‘calcium carbonate (CaCO3) with a very low percentage of calcium sulfate (anhydrite) (CaSO4)’ (PN II, 229).

Farther afield on Minoan Crete, the excavators at Phaistos identified a type of cement (calcestruzzo) composed of stones, clay, lime and pottery fragments (J. W. Shaw 1973a, 222–24; Graham 1987, 147–48). The builders must have recognised the strength of this mixture, since it was used as a fi lling and levelling material for new building and continued to be produced for later building phases. Calcium carbonate was also used in Minoan plaster production (J. W. Shaw 1973a, 206–18, 226). In the earlier periods, calcium carbonate content was low (40–65 percent), but later, with the development of a finer-textured plaster, the amount of calcium carbonate rose to as much as 94 percent of the plaster composition.

69

Cameron, Jones and Philippakis 1977; their Table 1 (p. 128) lists lime (CaO) rather than calcium carbonate (CaCO3) content. 70 Almost every floor from the southern doorway of the propylon to the porch of the megaron, including the rooms about the Grand Staircase, is paved with cement, some of which now is very fragmentary (Wace 1949, 70–76). 71 Tsountas’ House and the House of the Columns (Wace 1949, 67 and 91–93, respectively). No scientific analysis is reported, and Wace did not distinguish between cement floors of Hellenistic date and those of the Bronze Age (ibid., 67). For lime in clay and plaster applied to floors and walls of houses outside of the fortification walls at Mycenae and elsewhere, see Shear 1968, 443–45. 72 Mylonas and Graham cautioned against the use of this term, implying that it may cause confusion with our ‘modern cement’ or ‘true cement’. I believe the confusion lies not in the application of the term but the misunderstanding of it. At its simplest, ‘cement’ refers to any material or combination of materials that, when allowed to dry and set, hardens with adhesive or binding qualities (Phillips 1994, 46). To a certain degree mud can act as a cement, albeit a poor binder. Lime can also function as a cement, and indeed, it was part of the plaster that covered Mycenaean and Minoan floors. It is not, therefore, inappropriate to apply the term to a floor that is built of a material containing a binding agent. Today, ‘cement’ is often used as an abbreviation for Portland cement, to which Mylonas and Graham must have been referring. The process of making Portland cement is slightly different than that for a lime-based mortar or plaster. In the former, lime (usually from limestone) and clay (containing iron, silica and aluminium) are burned together and then crushed to form a powder that is then mixed with water and aggregate (ibid., 127; Allen 1985, 394–96). For the latter, the lime is burned separately, crumbled or powdered, slaked (hydrated with water) and then mixed with other materials, which may include clay or mud (for the process see J. W. Shaw 1973a, 213; Adam 1994, 65–66). 73 Th is amount includes calcium oxide plus carbon dioxide, which were bound together according to Müller (1930, 179); as such, they formed calcium carbonate (CaCO3). Dörpfeld (1885, 254–55), in discussing the architecture of Tiryns, did not think that the Mycenaeans used lime mortar. He did not submit any samples for laboratory analysis.

Wall mortar has not been as extensively analysed as wall and floor plaster. K. Müller (1930, 178–80) submitted one mortar sample from Tiryns for laboratory testing (table 3.10). The mortar he collected came from an unspecified rubble wall and was found to contain a high concentration of calcium carbonate (more than 55 percent).73 This mortar was used in the lower parts of the wall only; the rest of the wall was carried up in rubble with a clay mortar. Elsewhere, a sand and lime mortar was reported to have been used in the tombs at Deiras in the Argolid, but it was not submitted for laboratory testing (Aupert 1996, 406; Deshayes 1966). Despite there being only these two tested examples and a possible untested third, other structures on the mainland may also have used a calcium carbonate-based mortar, the presence of which has gone unnoticed, untested or both. The pier-wall building system dates to the last major remodelling phase of the palace. By this time the production of lime must have been mastered and standardised for the making of plaster – a process already achieved on Minoan Crete (Adam 1994, 65–76; Allen 1985, 245; J. W. Shaw 1973a, 213). Limestone (calcium carbonate, CaCO3) was burned to produce lime (calcium oxide, CaO). This was then hydrated, producing slaked lime (calcium

337

Nelson – The Architecture of the Palace of Nestor Element

Percentage of Sample

Silica (Si02) Titanium Dioxide (Ti O2) Aluminum Oxide (Al2O3) Iron Oxide (Fe2O3) Calcium Oxide (CaO) Magnesium Oxide (MgO) Potassium Oxide (K 2O) Sodium Oxide (Na2O)

30.24 0.30 2.53 3.32 31.00 1.45 1.61 3.26

Carbon Dioxide (CO2) Organic Substances Total

24.52 1.32 100.13

Amount insoluble in hydrochloric acid 30.03 0.30 2.50 1.16 0.35 N/a Fixed

Amount soluble in hydrochloric acid .021 0.03 2.16 31.00 1.10 N/a Fixed

Table 3.10 Analysis of soil samples from Tiryns (Müller 1930, 178–80).

The mortar in the matrix of the pier-walls is very hard: while collecting samples with an electric drill, I broke several bits and bores in attempts to penetrate the core of the pier-walls. Blegen and Rawson frequently encountered the hardness of the lime mortar during their excavations, but they referred to the material as fallen chunks of migma. In volume I of The Palace of Nestor, migma is reported in 23 locations; of these, 20 are confined by or near the remains of pier-walls. These chunks of migma were not fused and molten masses of hardened limestone, mud and mudbrick, as the excavators thought, but original wall matrix consisting of rubble and mortar. They were, in fact, fallen parts of pier-walls, which came down in large, solid masses. Blegen himself recognised this, writing in an early field report (1954, 30) that ‘a compact concrete-like mass of fused and calcined stones and crude brick, which here [area of rooms 9–15; the exact area is unclear] projected above ground, was seen to constitute a large section of a fallen wall’; but he continued with the assumption that ‘under the effects of terrific heat, [it] had coalesced into a solid block.’75 Later he referred to migma as ‘cement-like’ (PN I, 217), which ‘had to be chiselled and chopped out by sheer force’ (ibid., 211). The wall matrix of the pier-walls is without a doubt very hard, and the fallen sections revealed its material and construction, though they were mistakenly identified. If simple mud had been used to consolidate the rubble of the pier-walls, then fallen sections should neither have appeared cement-like nor have required much force to break apart. The high concentration of calcium carbonate turned the mud in the piers and interstices into a true mortar, or close to it. Once hardened, it strongly bound all of the material encased within it, functioning as a load-bearing material.

to set. The walls could have been built up in short sections, perhaps a metre at a time, with the mortar allowed to set before another section was built.76 If formwork (similar to that in modern poured-concrete construction) was used in the construction process, then walls could be built in relatively high sections or even to their full height with little settling and little concern for collapse (fig. 3.54). Some of the inside faces of some piers (e.g. chases 32 and 34, seen in fig. 3.50) are extremely flat and vertical, suggesting that boards or some other flat and stiff material may have been used to form shells within the frame. This is the wall building method that Blegen and Rawson proposed – a heavy timber framework into which rubble and mud were packed. Such a system explains the moulds and chases present in the wall matrix and the use of the hardening agent, calcium carbonate. But the gaps between the piers show that walls were built of single, independent piers, not as long, continuous masses of rubble and mud, as the excavators thought. The interstices are not simple moulds left behind by chase beams used to tie together two faces of a timber framework, for they span the entire thickness of the wall and rise from floor to ceiling. As has been noted, there was an insignificant amount of wood within the pier-walls when the palace was destroyed – certainly not enough to represent an extensive and massive timber framework. Blegen and Rawson’s xylodesia proposal for wall construction is nevertheless correct, except for two details – individual piers were formed, 74

Additives to the lime, such as sand or pebbles, affect its strength and binding capabilities to a greater or lesser degree; they also make the drying process uneven and may cause cracking (Allen 1985, 245). 75 Th is sort of confusion occurred sporadically during the original excavations when the excavators could not determine if the walls they were digging were in situ or fallen. For example, Blegen reported on 14 June 1955 that the area of Room 44 was ‘very difficult to dig, hard fused material – almost impossible to tell what is wall and what is fallen mass of mbrick [mudbrick]’ (excavation notebook entitled ‘C.W. Blegen, Pylos 1955’, 63). 76 Th is building technique has not been reported in Mycenaean architecture, but J. W. Shaw (1973a, 78–79) identified the technique for some rubble walls in the Minoan palace of Mallia.

With the use of calcium-rich mortar, the need to ensure structural integrity by having stone rest upon stone (a requirement established millennia before with rubble wall building) was considerably reduced. Instead, the large amount of mortar in the pier-wall building system required a different method of construction. In piers where there is 0.20 m. or more of mortar between stones, settling would have occurred, with the eventual collapse of the pier, unless sufficient time was allowed for the mortar

338

Building Methods The timber formwork was erected in the second step of the building process and dismantled in the fourth step. The formwork must have been very simple. For each pier, four posts were erected and tied together with transverse (running between piers) and face beams (along the exposed faces of the pier). Neither the posts nor the beams need have been squared timbers – some were logs left uncut or split longitudinally (PN I, 38). Tacked to this formwork of posts and beams were boards that formed temporary shells to hold the rubble and mortar while the latter set. The boards were probably relatively narrow and nailed close together, much like those used in poured-concrete construction in Greece today. As indicated by the surviving horizontal slots in the faces of the walls, wooden face beams may have been used not only to tie together the posts of a single pier, but also to separate the posts of several piers. There is some indication that the spacing of the timber posts was pre-calculated in such a way that the formwork could be reused as movable units; the face beams may have been used to measure these units.

rather than entire walls, and the timber framework was removed after the piers had set. A timber framework was necessary to build pier-walls because the mortar required external support before it set. Afterwards, however, there was no need for such support: the hardened piers were just as capable of supporting load as a cut stone masonry or rubble wall. The strength of the pier-wall building system is demonstrated by its use in walls (such as those of Room 6) that bear some of the heavier, if not the heaviest, loads in the palace complex. The pier-wall building system comprised the following steps (fig. 3.54): setting down a layer of flat stones to serve as a foundation; erecting a timber formwork; building the piers with rubble and mortar; removal of the formwork; fi lling the interstices with mortar, sand and small stones; and plastering both sides of the walls. Unfortunately, plaster floors cover the foundations of many of the pier-walls, but as can be seen in the few empty interstices of Room 6 and in the lower portions of wall 8LD (separating rooms 1 and 55), a layer of flat stones does lie immediately beneath both piers and interstices. Whether these were laid on bedrock or built foundations is unknown.

The walls of Room 6 vary greatly in preservation, and it is possible to see the finished as well as the deteriorated state. In some sections of the walls the plaster coat remains intact, while in other areas it is completely missing. Between

3.54 Pier-wall construction method.

339

Nelson – The Architecture of the Palace of Nestor the cross walls 5A and 5B, the long walls 8L and 4L each consist of 12 piers (fig. 3.45 above). Wall 5A is built with nine piers and wall 5B, which is pierced by doorway 5–6, with seven piers (three in the stretch of wall to the northeast of the doorway and four in the southwestern portion). Three of the four piers at the wall intersections are L-shaped, the others are rectangular or square in plane. The piers are uniform in thickness, but vary in length (where possible, all piers were measured at the base of the wall). The long walls 8L and 4L (1.30 m. thick)77 are thicker than the cross walls 5A and 5B (1.05 m. and 1.02 m. respectively)78. Tables 3.11–3.14 record the length of each pier in these four walls. Lengths vary considerably overall, particularly for walls 5A and 5B, but there are two distinguishable patterns in walls 4L and 8L. When charted, it becomes obvious that the piers of wall 8L were built in groups of three, with each group containing a series of one long and two short piers (table 3.14). The same pattern is continued around the corner into wall 5A, but only for the three northeasternmost piers (table 3.13). For wall 4L, there are two distinctive groups, each consisting of one long and four short piers; each group is separated by a pier shorter than those in the flanking groups (table 3.11).

wall was partially built in the ashlar shell wall style, which was then replaced by a pier-wall. Three of the pier-walls in the Southwestern Building, SW42 (table 3.18), SW43 (table 3.19) and SW46 (table 3.20), also show no consistently repeated pier length (plans xviii and xix). These are relatively short sections of walls and might not have been long enough to warrant the reuse of a formwork unit. The three walls are also part of a later remodelling, all added to pre-existing rubble walls. The existing walls, and other features possibly missing, probably required a uniquely dimensioned formwork. The pier lengths in the pier-walls of rooms 60 and 62 also lack signs of a pattern (plan xvii and table 3.21). The small building containing rooms 60 and 62 was built in a single phase and entirely with pier-walls (figs 3.55–3.57). Those of Room 62 are for the most part missing, but those of Room 60 are much better preserved and still retain their thick coat of mud plaster. The depth of the plaster and its good state of preservation render the pier-and-gap rhythm nearly invisible here. Only the tops of the walls reveal the pier-wall system; the tops of the interstices are particularly distinctive because they are currently conserved with a coat of white plaster (now obscured by accumulated dust and in many instances displaced rubble).

Elsewhere in the Main Building, other patterns of pier length are evident. Wall 10J (the northwest wall of Room 46; plan ix) is 0.80 m thick and contains seven piers (table 3.15), of which every other pier is approximately 0.70 m in length. For wall 9L, on the northeast side of area 44, the pattern of groups of three piers with one long and two short piers appears again, similar to that in wall 8L (table 3.16). The two piers at the end of the wall vary greatly in length and probably represent formwork that was measured and erected specifically for this location.

Like those of Room 6, the corner piers of Room 60 are L-shaped, but they are turned so that they bond with their intersecting walls (fig. 3.55). Wall P1 is the longest uninterrupted wall of the small building, consisting of five piers. The length of each pier is different (table 3.21). The piers of walls P2 and P4 are also inconsistent in length. Only those of wall P3, which measure 0.73–0.74 m, show some consistency.

The spacing of the timber posts in the megaron was apparently pre-calculated in the second step of the construction process, that of erecting the timber formwork (the actual laying out of the walls and piers), in order that the timber units might be reused. The length of walls 4L and 8L at Room 6 was known before they were built,79 because both walls begin and end with interstices and L-shaped piers. Walls 4L and 8L abut the elbows of the L-shaped piers, not their arms, and thus corner piers were set in place first. With these established, the builders calculated the lengths of the piers between. Instead of laying out 12 piers of equal length, they opted for a short series of repeated pier lengths. A three-unit formwork was erected for wall 8L, and piers 1, 2 and 3 built. When these set, the formwork was dismantled and re-assembled to build piers 4, 5 and 6. This process continued until the wall was complete. Similar formwork units were used in walls 4L, 9L and 10J.

The reason for not employing a re-usable formwork in this small building cannot be explained. As a single, free-standing structure, it would seem to have been an ideal candidate for such a method of construction. In the planning stages of the building, the piers could have been staked out according to the dimensions of a re-usable formwork – but they were not. It could be that the building was built entirely at once rather than in sections like the walls of the megaron. This would have required all of the formwork to be erected first. 77

Blegen and Rawson measured 1.25–1.30 m (PN I, 78). For the northwest wall, Blegen and Rawson measured 0.70–0.90 m, and for the southeast, 0.97–1.02 m (PN I, 78–79). Our measurements are from fi nished plaster face to fi nished plaster face where preserved; in some areas where the plaster face is missing, the walls vary in thickness. 79 As Graham (1960, 52) noted, the dimensions of the three mainland megara in the Peloponnese (Englianos, Tiryns and Mycenae) are approximately the same. Th is must be more than coincidence and suggests that the builders of Peloponnesian megara had preconceived plans, which included room sizes, for this portion of a Mycenaean palace. 78

However, not all pier-walls show a pre-calculation of pier lengths. For the northwest section of wall 9L, at rooms 30 and 32, there is no recognisable pattern (plan x and table 3.17). This section is a later remodelling: the original

340

Building Methods

Wall 5A

Wall 4L 1.00

1.00

0.80

0.80

0.60

0.60

0.40

0.40

0.20

0.20

0.00

0.00 35

22 23 24 25 26 27 28 29 30 31 32 33

36

37

38

Wall 4L 4L 4L 4L 4L 4L 4L 4L 4L 4L 4L 4L

Length (m.) n/a 0.77 0.76 0.76 0.74 0.91 0.66 0.82 0.83 0.80 0.78 0.89 0.75 n/a

40

41

42

43

Pier Number

Pier Number

Pier 21 22 23 24 25 26 27 28 29 30 31 32 33 34

39

Thickness (m.) n/a 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 1.25–1.30 n/a

Pier 34 35 36 37 38 39 40 41 42 43 44

Wall 5A 5A 5A 5A 5A 5A 5A 5A 5A 5A

Length (m.) n/a 0.72 0.81 0.85 0.92 0.93 0.76 0.89 0.77 0.75 n/a

Thickness (m.) n/a 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 n/a

Table 3.13 Pier dimensions of wall 5A.

Wall 8L

Table 3.11 Pier dimensions of wall 4L. 1.00

Pier 13 14 15 16 17 18 19 20 21

Wall 5B.1 5B.1 5B.1 5B.2 5B.2 5B.2 5B.2

Length (m.) n/a 0.76 0.78 1.02 0.93 0.92 0.51 0.43 n/a

Thickness (m.) n/a 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 n/a

0.80 0.60 0.40 0.20 0.00 1

2

3

4

5

6

7

8

9

10 11 12

Pier Number

Table 3.12 Pier dimensions of wall 5B.

Pier 1 2 3 4 5 6 7 8 9 10 11 12 13

Wall 8L 8L 8L 8L 8L 8L 8L 8L 8L 8L 8L 8L

Length (m.) 0.76 0.78 0.89 0.77 0.79 0.88 0.71 0.72 0.87 0.75 0.72 0.87 n/a

Table 3.14 Pier dimensions of wall 8L.

341

Thickness (m.) 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 0.97-1.05 n/a

Nelson – The Architecture of the Palace of Nestor Pier (from northeast) 1 2 3 4 5 6 7

Wall

Length (m)

Thickness (m)

10J 10J 10J 10J 10J 10J 10J

0.72 0.82 0.69 0.61 0.70 0.51 0.70

0.80 0.80 0.80 0.80 0.80 0.80 0.80

Pier (from northeast) 1 2 3 4

Wall

Length (m.)

SW42 SW42 SW42 SW42

0.74 0.54 0.70 0.73

Thickness (m.) 0.83 0.83 0.83 0.83

Table 3.18 Pier lengths of wall SW42 between rooms 76 and 78.

Table 3.15 Pier dimensions of wall 10J. Pier (from northwest)

Wall

Length (m.)

Thickness (m.)

1 2 3 4 5

SW43 SW43 SW43 SW43 SW43

0.73 0.77 0.76 0.78 0.80

0.84 0.84 0.84 0.84 0.84

Wall 9L at Stoa 44 0.8 0.6

Table 3.19 Pier lengths of wall SW43 between rooms 71 and 76.

0.4 0.2

Pier (from northwest) 1 2 3 4 5

0 1

2

3

4

5

6

7

8

9

10

11

Pier Number

Pier (from northwest) 1 2 3 4 5 6 7 8 9 10 11

Wall 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44) 9L (Stoa 44)

Length (m.) 0.45 0.44 0.63 0.52 0.42 0.63 0.53 0.48 0.63 0.71 0.45

Thickness (m.) 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.90

Wall 9L (room 32) 9L (room 32) 9L (room 32) 9L (room 32) 9L (room 32) 9L (room 32) 9L (room 30) 9L (room 30) 9L (room 30) 9L (room 30) 9L (room 30)

Length (m) 0.72 0.80 0.93 0.75 0.54 0.46 0.77 0.42 0.56 0.50 0.94

Length (m.)

SW46 SW46 SW46 SW46 SW46

0.66 0.78 0.74 0.84 0.81

Thickness (m.) 0.87 0.87 0.87 0.87 0.87

Table 3.20 Pier lengths of wall SW46 between rooms 71 and 73.

Pier 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Table 3.16 Pier lengths of wall 9L at stoa 44.

Pier (from northwest) 1 2 3 4 5 6 7 8 9 10 11

Wall

Thickness (m) 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20

Wall n/a P1 P1 P1 P1 P1 n/a P3 P3 P3 n/a P2 P2 P2 P2 n/a P4 P4 P4

Length (m.) n/a 0.88 0.7 1.13 0.74 0.76 n/a 0.74 0.73 0.73 n/a 0.26 0.93 1.1 1.13 n/a 0.53 0.64 0.75

Thickness (m.) n/a 0.90 0.90 0.90 0.90 0.90 n/a 0.90 0.90 0.90 n/a 1.02 1.02 1.02 1.02 n/a 1.00 1.00 1.00

Table 3.21 Pier lengths of walls P1–P4, Room 60.

Table 3.17 Pier lengths of wall 9L at rooms 30 and 32.

342

Building Methods Returning to the process of construction, the fift h step was to fi ll the interstices (after the timber was removed) with lime mortar, sand and small stones – a simple grout mixture. Probably a simple formwork consisting of a braced board on either side of the interstice was used to contain the grout while it hardened. Many of the slots and chases are still filled with this matrix. In some walls the excavators reported carbonised remains found in the chases.80 It no doubt happened, as it does today in concrete construction, that when the piers set and hardened it became impossible to remove some of the timbers of the formwork. Those left behind, which subsequently rotted or burned out, explain both the inconsistent placement and the irregular lengths of slots that held face beams, as well as the few empty chases that may have contained timber posts.

these walls is uniform throughout – there is no change of building material or matrix, such as there is between the piers and interstices in the pier-wall system at Englianos. The west wall of the bathroom (room XI) at Tiryns, however, does seem to have been built in a pier-wall system similar to that at Englianos. In its present state, the wall rises c.0.30 m. and consists of three recognisable piers, approximately 0.40–0.45 m in length, separated by interstices of c.0.16–0.20 m. The interstices, currently empty, span the thickness of the wall and may have risen from floor to ceiling without interruption. Unfortunately,

The earth material of the interstices is a different matrix than that of the piers. Since the interstices were not load-bearing, they could be filled in with grout, which is a homogenous mixture rather than a mortar. The narrowness of the interstices also made it difficult to include large stones or pieces of ashlar fragments within the matrix. The same grout may have been used to fill in the face beam slots left behind from horizontal wooden members of the formwork. Finally, in the last step of construction, both sides of the pier-walls were encased in plaster, completely concealing the piers and interstices. When the excavators uncovered the pier-walls, they were reminded of similar Mycenaean building systems, particularly at Mycenae and Tiryns (PN I, 37; see also Küpper 1996, 67–69). However, it is only in appearance that these walls are comparable to the pier-walls of Englianos. The clearest and most well illustrated examples from the Argolid sites are walls of the South House (fig. 3.59; Wace 1949, 66, fig. 24a), Tsountas’ House (ibid., 66– 67), the House of Columns (ibid., 95, fig. 108b) and the large ramp just inside the Lion Gate (fig. 3.58; ibid., 55, fig. 24b), all at Mycenae; and the west wall of the bathroom (room XI) and the east wall of Magazine XLVI at Tiryns (Müller 1930, 180–81, fig. 83). Except for the bathroom wall at Tiryns, all of these walls show a systematic application of wood with more or less regularly spaced interstices that once contained vertical and horizontal timbers. In crosssection, however, it is notable that the interstices do not span the entire thickness of the wall, either face-to-face or floor-to-ceiling, as they do at Englianos. Thus there are no piers at Mycenae or in Magazine XLVI at Tiryns, but simply shallow slots and chases in the wall matrix into which timbers were fitted. Moreover, the wall matrix in

3.55 Wall and pier designations, Room 60.

3.56 Room 60, from the northwest.

80

Blegen 1964, 117–118, and PN I, 38. In the detailed descriptions of PN I, only one chase – in the southwest wall of Room 4 – is reported to have contained carbonised matter when it was uncovered (PN I, 67). Most of the carbon matter from the palace was the remains of doors and doorway constructions (PN I, 71, 77, 103, 116, 117, 125, 151, 156, 173, 209, 319, 343). Other carbon matter came from actual wood fragments (PN I, 130, 192) and black-colored strata (PN I, 107, 120, 144, 239).

3.57 Northeast face of southwest wall of Room 60, at southeast end, from the northeast.

343

Nelson – The Architecture of the Palace of Nestor

3.58 West wall of ramp at Mycenae (after Wace 1949, fig. 24b).

Staircase 36 taken in 1954 (Blegen 1955, pl. 26, fig. 9) depicts nicely squared wall corners on top of the stair’s flanking antae. A similar squared corner made of original building material (the outside, south corner of Room 39; plan ix) still remains extant today and clearly demonstrates that some (if not all) wall corners were finished in such a fashion. In the summer of 1961, the excavators completed a round of ‘final cleaning’ throughout the palace (Blegen 1962, 147; PN I, 22), at which time extant portions of rubble and mud mortar were removed in search of dowel holes or underlying blocks.81 A photograph taken in 1961 (PN I, fig. 127) reveals the results of the cleaning and the exposed dowel holes (fig. 3.60 shows the same view in 1991). Other photographs reveal similar wall conditions prior to cleaning.82

3.59 Wall of South House at Mycenae (after Wace 1949, fig. 24a).

the matrix of the piers was not tested for lime content or other inclusions. In Minoan architecture, timber framework was also used in rubble-based masonry, but not as methodically as at Englianos (J. W. Shaw 1973a, 139–50). Examples date as early as the Early Minoan period, and the evidence points mainly to the use of horizontal members. Later, vertical posts placed at intervals along wall faces were used more frequently. In the Hall of the Double Axes at Knossos, a systematic timber framework was incorporated in cut stone masonry walls. However, these walls are not composed of individual piers with chases spanning wall face-to-wall face and floor-to-ceiling, but are long continuous masses of wall matrix with shallow slots and chases for timbers.

At the time of the fire that destroyed the palace, the seven anta blocks under discussion here did not support any kind of wooden armature or framework, and their dowel holes indicate the re-use of these large blocks from older constructions. Their purpose in one of the last phases of construction must have been to re-enforce the wall ends and perhaps partially to support the wooden doorjambs that may have socketed into them.83 For example, anta 9L.a doubles as a wall return and a doorjamb; its southwest face has a recess cut into it, and the northeast jamb block of doorway 28–35 fits snugly into the recess (PN I, 446). Two other anta blocks in the pier-walls clearly did support timbers above them. Antae 5D.a (fig. 3.21 above) and 5D.b (fig. 3.22 above) were found with no wall matrix on them. Both have mortises cut near their exposed faces (northwest and southeast) for a total of four on 5D.a and five on 5D.b. Both also have beddings cuts, which relate

Wood and Antae in the Pier-wall Building System Anta blocks terminated walls and served as doorjambs in pier-walls. As has been discussed above in reference to the chases and face beams in these walls, evidence is doubtful in the palace’s last phase for wooden timbers as either vertical posts or horizontal beams fastened to the anta blocks. When antae 5B.a, 5B.b, 5C.a, 5C.b, 8.a, 9.a and 9.b (plan iv) were initially excavated, the tops of these blocks were covered with wall matrix – rubble and mud mortar. Blegen and Rawson interpreted the concealment of the cuttings as the result of flowing lime or limestone and later removed the original wall material to reveal the underlying mortises and bedding cuts. For instance, a photograph of

81

Pick marks can be seen in the wall matrices above and around these antae blocks.

82

PN I figure references for antae: anta 5B.a, figs 9, 22, 23, 45, 59, 62, 88, 93; anta 5B.b, figs 9, 22, 23, 45, 59, 61, 88, 93; anta 5C.a, figs 9, 22, 23, 45, 59, 88, 93; anta 5C.b, figs 9, 22, 23, 45, 57, 59, 88, 93; anta 8.a, figs 9, 11, 45; anta 9.a, figs 58, 126, 127; anta 9.b, figs 22, 23, 58, 126, 127. 83 They may also have been used in the laying out of the wall courses. The antae would be laid fi rst and the wall between them fi lled in, with the antae serving as benchmarks for the wall course.

344

Building Methods no drystone-laid walls. They were constructed simply by stacking stones one atop the other with a thin layer of mud spread between. The settling of the stones eventually pushed out the mud so that stone rested upon stone, and the mud fi lled the gaps and cavities. The mud stabilised and consolidated the wall to a certain extent by preventing the stones from shifting. Coursing within a rubble wall depended upon the unworked shape of the building stone. Flatter, slab-like stones were easier to lay in courses than more rounded rubble, but crude coursing is evident in just about every rubble-built wall. Plaster covered some rubble walls and concealed their coarse appearance.84 In general, three types of rubble walls were built: rubble shell walls, uniform rubble walls and rubble slab walls.85 Rubble shell walls are similar to ashlar shell walls and consist of an outer shell of large stones and an inner core of smaller stones. In uniform rubble walls, stone sizes vary, but they are evenly spread throughout the wall. Rubble slab walls employ small, uncut limestone slabs. The difference between rubble slab walls and the limestone slab construction discussed below is largely the size of individual slabs, but the latter also uses small stones both between courses and to fill interstices. It is not unusual for rubble-built walls to incorporate a very large stone or two. Wood was not a component of any of the rubble masonry systems at the palace.

3.60 Stairway 36 in 1991, from the southwest.

directly to the mortises and in turn correspond to the anta blocks’ exposed faces. Anta 5D.a has two bedding cuts that span the entire length of the block: on the southeast a c.0.22 m bed, and on the northwest a wider bed measuring 0.34 m. The northwest bed was sunk c.0.04 m deeper than the southeast one, and the area between them is slightly higher, c.0.01 m. Mortises were cut into each bed: two on the southeast and five on the northwest; two of these may be the result of remodelling (PN I, 60). In the wall beyond the block (wall 5D.1) towards the northeast, there are preserved two face beam slots, one on each side of the wall, which match the height of their corresponding bedding cuts in the block and run the entire length of the wall (plan viii). The beam slots are narrower than the bedding cuts: c.0.17 m on the northwest and c.0.15 m on the southeast. From this evidence it is possible to reconstruct two face beams that fastened to the anta block and spanned the distance from the southwest edge of the anta block to wall 8LD.

The rubble-built walls on the hilltop are far too numerous to examine in detail, but several deserve comment. Many of the rubble walls of the Southwestern Building are rubble shell walls. For the most part, larger stones line the outer faces of the walls, but in some wall sections there seems to have been little regard to distinguishing the exterior and the inner core. The northwest and southeast walls of Room 65 (SW50 and SW52; PN I, fig. 205) are substantial, averaging c.1.50 m and c.1.30 m thick respectively (plan xviii and fig. 3.6 above). Rough coursing is evident. Both walls run perpendicular to the topography, stepping down towards the southwest. As a result, the number of courses varies from one or two at their northeast ends to as much as twelve near their southwest ends.86 Larger stones, reaching rough dimensions of 0.37 x 0.17 m, appear frequently in the bottom courses. Although the walls are

A similar situation exists on the opposite side of the doorway. Anta 5D.b also has two bedding cuts: the one on the northwest side is c.0.33 m deep and that on the southeast c.0.40 m deep. The southeast bed is higher than the northwest by 0.05 m. The excavators reconstructed face beam slots on both sides of the wall beyond anta 5D.b to the southwest. The evidence is not quite as clear as it is in the wall on the opposite of the doorway because some in situ stones would both block a beam lying in the southeast bed and prevent it from extending southwesterly into the wall. Nonetheless, it seems reasonable to reconstruct wood fastened to this anta block.

84

Wall 2E, southeast face, in Room 15; wall 3L, southwest face, wall 2D, southeast face, and wall 2E, northwest face, in Room 17; and wall 1L, northeast face, in Room 19. 85 Th is typology is by no means complete, since rubble masonry by its very nature has many subtle variations. It does, however, provide some general categories. The first two categories are based on the cross-section of the wall. The third category relies on the appearance of the wall face, since a rubble slab wall may have an inner core of smaller slabs or may consist of slabs uniformly throughout. Discussions of rubble typology can be found in Walsh and McDonald 1986; Tzavella-Evjen and Rohner 1990; Küpper 1996, 26–31 (mainland Greece); J. W. Shaw 1973a, 77–83 (Minoan Crete); Netzer 1992, 22–23 (ancient Israel). 86 The present height of both walls is the result of erosion and modern plowing.

Rubble Construction Rubble masonry was used throughout the palace complex: in the Main Building, portions of the Southwestern Building, Area 106, the Northwest Area and the Belvedere Area. The walls of the Wine Magazine and the Northeast Building may be rubble socles for mudbrick superstructures, but in their present condition they are no different than other rubble walls. The rubble walls consist of two components, rubble and mud; there are

345

Nelson – The Architecture of the Palace of Nestor substantial, they are poorly laid out, deviating from a straight course by as much as 0.10 m in some sections.

majority of the walls of the Main Building are constructed in the pier-wall building system, so those of Room 27 and the southwest suite of rooms are easily distinguished.

The northeast wall of courts 42 and 47 is a rubble shell wall that includes reused ashlar blocks (plan xxii; PN I, 182–83, 206). For approximately 7.20 m, ashlar blocks were placed face down in the first course of the wall (figs 3.61– 3.62). The blocks do not span the entire thickness of the wall, but were laid only at the interior face. Above them the wall continues in rubble masonry with very rough and uneven coursing. A good portion of this rubble contains small fragments of ashlar blocks.

Wood and Antae in Rubble Masonry The only examples at the palace of antae used in rubble masonry are found in the Northeast Building; all of these blocks are incorporated in the walls of Room 93 (plan xxvi). Two three-sided anta blocks (14L.a and 15L.a) terminate Room 93’s southwest and northeast walls at their southeast end. Anta 14L.a has no recognisable mortises,89 thus wood does not seem to have been fastened to it. Its partner on the opposite side of the room, anta 15L.a, has three mortises that correspond to the block’s southwest, northwest and northeast sides. This arrangement is slightly odd according to the rules stated above. The northwest side of the block is not an exposed face because the northeast wall of Room 93 abuts it, yet there is a mortise near this edge of the block. Moreover, the exposed southeast face has no corresponding mortise, as would be expected, suggesting that anta 15L.a is reused.

The northeast wall of Court 42 is later in date than the southwest wall of Room 102, and when the former was built, it engulfed the latter (plan xxiv). The rubble of the earlier wall comprises small slabs, and the later wall is rubble shell wall construction (PN I, 183). In its original state, two reused ashlar blocks perhaps served as quoins or antae at its northwest end; the later rubble shell-constructed wall simply abuts. In its finished state, the entire face of the court’s northeast wall, both interior and exterior, was covered in plaster (PN I, 183, 299), so the differences in the masonry style were not visible.

The third anta, 15L.b, sits at the T-intersection of the southwest wall of Corridor 95 and the wall separating Room 93 from Room 96. Four mortises are cut into the block’s upper surface, and their arrangement seems to indicate four exposed sides. Each mortise is approximately centred

The walls of the southwest suite of rooms (14, 15, 17, 19–21; plan vii)87 and Room 27 (plan vi)88 of the Main Building are built of uniform rubble masonry. Room 27 is a late addition, and the southwest suite of rooms is a late remodelling (PN I, 45). The walls of the southwest suite are thinner (0.86–0.97 m) than those of the rest of the building, and they are poorly laid out; they neither follow straight courses nor do they intersect at right angles. In the longest stretch of rubble wall in this suite, wall 3L, three to five courses of rubble remain, with stone sizes measuring up to 0.53 m. Coursing is rough at best. The

87 The extreme southwestern portions of walls 2B, 2C, 2D, 2E and 2F were reconstructed by the Greek Archaeological Service. 88 The southeast half of Room 27’s southwest wall and the southwest section of the southeast wall were reconstructed by the Greek Archaeological Service. 89 Blegen and Rawson (PN I, 304) suggested that there may have been a mortise on the southwest side.

3.61 Southwest face of northeast wall of Court 42.

3.62 Southwest face of northeast wall of Court 47.

346

Building Methods in relation to its corresponding face. This arrangement would place a horizontal timber along each face, but since the rubble masonry of the Northeast Building did not include timber, it is difficult to find a relationship between the wood of anta 15L.b and its abutting walls. If anta 15L.b was reused, which seems to be the case, it may have once been the base of a free-standing pier.

underneath and perpendicular to the southeast wall of Room 65. One course is preserved in the section of the wall to the southeast of wall SW52, and in the northwest section, under the southernmost column base, two courses still stand. The wall is c.1.08 m wide (measured at the southeast section),94 and the slabs here are quite large: the biggest measures c.0.93 x 0.47 m. The slabs are laid in header fashion, with their short dimension at the exterior and their long dimension extending into the thickness of the wall. Small bedding stones underlay some of the large slabs, but for the most part the wall is founded on bedrock. The last metre or so of the very northwest end of the northwest section is built on an even earlier rubble wall. This wall runs parallel to the contours of the southwestern side of the hilltop and probably once served as a retaining wall.

Rubble Socle Construction Rubble socle construction consists of a rubble masonry socle upon which sits a mudbrick, or mudbrick and halftimber superstructure; this system has a long history in mainland Greece, reaching as far back as the Neolithic period.90 The socle provides both a durable, water-resistant wall base that could withstand rain and ground-born dampness, and a level surface upon which to erect the superstructure. There is little tangible evidence for the use of rubble socle construction at the palace, and no mudbricks were discovered in situ on a rubble socle. The excavators nevertheless relied on strata of disintegrated mudbrick to posit rubble socle construction for the walls of the Wine Magazine (plans xxii–xxv)91 and the Northeastern Building (plan xxiv; PN I, 36–37).

The second limestone slab-built wall is later in date and consists of sections 1, 2 and 3 of the Southwestern Building’s southwestern facade (plan xx; PN I, 278). Section 1 is U-shaped and comprises the northeast, northwest (SW39) and southwest walls of Room 81. Altogether, these sections supported the west corner of the Southwestern Building where the terrain drops towards the south and southwest. Unlike the earlier wall, the slabs are laid occasionally as headers and occasionally as stretchers. The stones appear to have been laid with no apparent order or arrangement, though this will remain uncertain unless the wall is dismantled. Moreover, the wall is also not laid uniformly

The walls of the Wine Magazine were carelessly built, neither following straight lines nor maintaining a uniform thickness (PN I, 342).92 They are rubble shell walls containing stones of various sizes, including small (0.16–0.41 m) slab-like stones. The walls of the Northeastern Building are similarly built, with little concern for coursing or uniformity in either stone size or placement. No single wall of these buildings preserves a level upper course, which would seem necessary and efficient if indeed the wall was a socle. These walls, like those of the rest of the palace, have suffered from erosion and modern plowing, and their current height reflects the pre-excavation topography of the hilltop. Although none of these characteristics excludes the possibility that the wall superstructures were indeed built of mudbrick, there are no structural or constructional differences that distinguish them from ordinary rubble walls. Without more convincing evidence, complete rubble construction should be assumed for these two buildings.

90

A good example of the continuous use of this construction method is at Korakou, in the Corinthia, where the system can be traced in residential architecture from the Early to Late Helladic periods (Blegen 1921, 75–99). Some modern buildings erected with this system still stand today in the town of Chora, 3 km north of Englianos. The system is rarely discussed in detail, but for general comments see Dickinson 1994, 144; Lawrence and Tomlinson 1983, 28–34; Shear 1968, 438–41; J. Wright 1978, 132–33. 91 Blegen and Rawson (PN I, 342) reported chases in the southwestern and northeastern walls that were not observed during MARWP’S surveying and drawing of this building. Erosion may have obliterated the evidence. If indeed there are chases in these walls, this would indicate a building system of half-timbering on a stone socle (surveyed in Küpper 1996, 67–69). 92 A state plan of the Wine Magazine was prepared by Blegen and Rawson (PN I, fig. 428); it does not vary considerably from that prepared by MARWP. The MARWP plan shows the partially preserved plaster floor of Room 104 and the Middle Helladic wall beneath the same room. The locations of the pithoi and the remains of the benches in which they sit should be read from Blegen and Rawson’s plan, as the fragility of the remains precluded their uncovering by MARWP. 93 J. Wright (1978, 162–66) identified this style of construction on early sites (Malthi and Peristeria) in Messenia, including the flanking bastions of the Northeast Gateway at the Englianos palace, as the origin of Cyclopean masonry. The common element among the walls Wright surveyed is the use of large, naturally flat limestone slabs used on the exterior faces. 94 The northwest section is partially destroyed and partially concealed by the southernmost column base.

limestone slab construction Limestone slab construction is a heavy form of wall building, consisting of two parts: an outer facing of large, minimally-worked, limestone slabs and chinking stones, and an inner backing of rubble masonry.93 The two walls at Englianos constructed in this manner served as retaining or terrace walls, with the inner backing of rubble completely concealed by earth backfi ll. The earliest of these two walls was partially destroyed and built over when Room 65 was constructed (plan xviii). The wall is nearly 8 m long and runs northwest–southeast,

347

Nelson – The Architecture of the Palace of Nestor but the bonding is doubtful (ibid., 98).97 The proposed compartment-like construction method requires a strong and secure bond between the retaining wall and the anchor wall, but only wall SW39, the northwest facade of the building, forms a clear, bonded corner with section 1 of the southwest facade. The wall separating Room 80 from Room 81 is badly preserved, with only one or two courses remaining; its southwest end, near the facade wall, is almost entirely missing. Furthermore, if one were to extend the course of this wall and continue it to the facade, one half of the wall would intersect with section 1 and the other half with section 2. In contrast to Blegen and Rawson (PN I, 276), the implied intersection does not correspond to the offset in the wall. Wall SW43 (separating Room 74 from Room 80) is also badly preserved at its southwest end, and no clear evidence exists that the wall continued to the facade (PN I, 271). If the line of the wall were extended, it would indeed correspond to the offset between sections 2 and 3. Nevertheless, if the two walls did at one time bond at the corner, two different building systems would have bonded: rubble masonry and limestone slab construction. This connection could not have been very secure, because of the vastly different sizes of stones employed in each system.

with large slabs. In some portions there is simple rubble masonry built with fairly large unshaped stones. The slabs naturally lend themselves to coursing, but the rubble often interfered and made coursing difficult. The builders seem to have compensated for the irregularity of the rubble by inserting levelling courses and using chinking stones to level individual slabs. Instead of attempting to level every course as it was laid, they built the wall up in sections, with the top of each section capped with a levelling course (figs 3.4–3.5 above). In section 1, the top of the fourth course is level, and below it the coursing is irregular. On top of this, the wall continues, but again the coursing becomes irregular. In section 2, the same method seems to have been applied, though most of the upper portions of the wall are missing. Coursing is irregular until the fourth course, which is composed of four of the largest slabs in the wall, all with their upper surfaces at roughly the same elevation. At the northwest end of this course, the smaller, chinking stones were used to maintain the level created by the four large slabs. Section 3 is badly damaged and does not preserve as many courses as its neighbors to the northwest, but it is possible that here levelling occurred at the third course, since the coursing below is irregular.

J. Wright’s L-shaped units are thus doubtful as structural entities; perhaps the heaviness of the limestone slab construction of the facade effectively retained the earth behind them. No doubt the weight of the walls above the slab construction also contributed to their retaining capabilities. However, the proposed compartment system is definitely valid for the U-shaped wall configuration around Room 81. All three facades of this compartment were built using the same system, and the bonds are strong and tight.

Stones that do not have slab-like qualities, with parallel or closely parallel bedding surfaces, are disruptive to coursing. With chinking, the builders levelled the upper surfaces of irregularly shaped slabs. For instance, a slab with a triangular exposed face was inverted so that its apex pointed downwards and its flat surface became the upper bedding joint; the smaller stones propped it up and levelled the upper surface. This technique occurs several times in section 1, but only a few times in sections 2 and 3 (perhaps the result of poor preservation in these sections). The need to maintain the level and sturdiness of the southwestern facade stems from its function, which was to retain earth and support the west and southwest portions of the Southwestern Building. It has been proposed that sections 1, 2 and 3 form compartment-like wall arrangements built to counteract the loads placed on these walls (Wright, J. 1978, 61–62, 65– 68; 1980).95 In this scenario, section 1, wall SW39 and the northeast wall of Room 81 form a U-shaped compartment (plan xx). Sections 2 and 3 are L-shaped compartments, each composed of a section of the facade and a wall return; the wall separating Room 80 from Room 81 is associated with section 2, and that separating Room 74 from Room 80 with section 3.96 These L- and U-shaped wall sections would have served to compartmentalize the terrace fi ll retained by the walls and to transfer the load of the upper storeys directly to the bedrock upon which the southwestern facade sits (Wright, J. 1978, 61–62). The compartments would have resisted the pressure of the retained soil because the return walls served as anchors. The relationship between the anchoring walls and the facade is not very clear, however. Published plans depict these walls as aligning with the offsets in the facade,

95

Blegen and Rawson (PN I, 278) saw the facade as simple repair work that replaced the original ashlar facade. 96 These are best seen in Blegen and Rawson’s plans of the Southwestern Building, e.g. PN I, fig. 417. 97 Blegen and Rawson seem hesitant to fi ll in these walls in their published plan (PN I, figs 410 and 411).

348

4 BUILDING HISTORY

Masonry Development at Englianos

unknown building system, the original locations of which have not been discovered, but which employed cut ashlar blocks; second, pseudo-ashlar masonry, which reused blocks from the first phase; third, orthostate construction; fourth, the ashlar style, which includes ashlar shell wall construction; and fift h, pier-wall construction.

The history of masonry at the palace set forth here derives from the few published dates available, the original excavated stratigraphy and the ‘stratigraphy of walls’. The excavators concentrated their efforts on the Late Helladic IIIB palace, but throughout their investigations they encountered earlier and later material. Where sufficient and clear stratigraphy permitted, they established dates for these remains. Often dates were given for whole structures (see table 1.8 above) rather than individual features. The stratigraphy of walls, or the physical relationship of one wall or feature to another, is thus important because it adds further details to the sequence of construction for both the walls of individual structures and those of the whole site. The different periods of wall construction are marked by changes in material and building method, by the building of one wall or structure on top of another, and by the abutting or bonding of one wall or structure with another.

Phase I: Original Cut Ashlar Masonry Although there are no remains in situ, there must have been a phase of ashlar masonry construction on the site prior to the reuse of blocks in pseudo-ashlar masonry. Its seems unlikely that the pseudo-ashlar style was the earliest form of wall construction using ashlar blocks at Englianos, since combining rubble with cut stone in wall building would defeat the purpose of quarrying and squaring stone. Little is known about the initial use of these cut blocks, particularly how they were set and if they served specialised functions, such as antae or quoins. They lack anathyrosis on their vertical joint faces, where they would abut other blocks, and they are cut to various sizes. Most of them are smaller than and not as finely worked as those of the later orthostate and ashlar styles (PN III, 13). Because the blocks vary in height, they do not lend themselves easily to coursing and were probably used in a random range masonry style (fig. 3.16 above).

Much of the original stratigraphy at the site was never published and is now lost; for the most part, it cannot be correlated with the stratigraphy of walls presented here, though there is one exception. Around the site, some original strata, both Bronze Age and later, can still be seen in the few scarps and baulks left behind by the excavators (at present mostly covered by backfi ll). One stratum stands out and is most important for dating because it represents the levelling and grading of the southwestern half of the hilltop in preparation for the LH IIIB remodeling of the palace (PN I, 32, 423).

Phase 2: Pseudo-ashlar Masonry The excavators thought that this second phase of construction belonged in some cases to the Late Helladic I period (PN III, 32–33). Firmly dated to this period is the pseudo-ashlar section 8 of the Southwestern Building’s southwest facade (plan iii; see also plan xvi), also referred to by the original excavators as the ‘circuit wall’ at the southwestern edge of the hilltop (PN III, 13); whether or not the wall served this function is discussed below. Dates for the other pseudo-ashlar walls are not secure. The excavators assigned the pseudo-ashlar walls of Building

The history of cut stone masonry at the palace is straightforward and can be summarised as follows:1 first, an 1

Rubble masonry is excluded from this history because it was used during all periods at Englianos with few changes in building materials or methods.

349

Nelson – The Architecture of the Palace of Nestor X and the walls beneath Court 63 (SW58a–f and SW59 in plan iii) to the LH IIIA period.2 The supposed circuit wall abuts the southwest wall of Building X at its west corner (plan xv; see also PN III, fig. 22). It is clear from the abutment that the southwest wall of Building X was built before the circuit wall, so it too must date to LH I or earlier. The pseudo-ashlar walls SW58 and SW59 were all built in the same fashion with reused ashlar blocks, which suggests that they also belong to this early period.

the ashlar style appeared in the buildings on the hilltop. All of the earlier orthostate walls were replaced with or altered by ashlar style walls. The transition is best seen in the northeast facade of the Main Building. The orthostate building represented by corner anta 11L-20 (figs 3.30–3.31 above) was destroyed, and both facades were dismantled. The corner anta then became a projecting anta for the ashlar style wall, 11L, built just to the southwest of it and incorporating its southwest face.

Phase 3: Orthostate Construction

The orthostate wall to the southeast of Room 65 was also altered, when an ashlar wall was tacked onto the northwesternmost orthostate and continued the wall’s course farther towards the northwest (fig. 3.29 above). The orthostate wall beneath Room 7 was partially dismantled and never again used; a later floor was placed over it and a new facade erected on the cobblestone pavement in front of the orthostates. Other orthostate walls, such as those beneath rooms 50 and 53 and Court 88, were completely dismantled save for their socles.

At some point after the building of the pseudo-ashlar walls, orthostate construction appeared in the architecture on the hilltop. The orthostate wall to the southeast of Room 65 attests to the introduction and later date of this style (figs 3.28–3.29 above). The wall runs approximately northwest–southeast and was built directly over the top of the existing LH I pseudo-ashlar wall (section 8, or the supposed circuit wall).3 This is the only instance in which a wall built in the new orthostate style displaced an older pseudo-ashlar wall. Elsewhere on the site, many of the other pseudo-ashlar walls continued in use.

Only a few comparable Late Helladic examples of orthostate construction exist, but dates are uncertain for all of them. The two blocks preserved in the east wall of the court in front of the large megaron at Tiryns were reused in one of the last phases of construction, not as orthostates but apparently simply as rubble matrix. The dromos walls of the Tombs of the Genii and Atreus at Mycenae (figs 3.14– 3.15 above) resemble most closely the orthostate masonry system of Englianos, but their LH IIIB date is based on architectural typology and not stratigraphy.4 If, however, these walls do belong to LH IIIB, then they date to a time when cut stone masonry was no longer employed in new wall construction at Englianos.

Orthostate construction brought more sophisticated wall construction techniques. Blocks increased in size, with some becoming quite large. One block near the Main Drain in the Southwest Quadrant area measures c.1.30 m long across its front face and has a maximum depth of 0.96 m (in plan xvi, this block has an elevation tic of 187.58 masl; it was never fully excavated). It is wedge-shaped and resembles the socle blocks beneath the orthostate wall northeast of the northeast facade. Engineering also improved in this style of building. Anathyrosis assured smooth and flush joints at the exterior faces of the walls, and mortises and dowels secured wooden members to the blocks.

One of the orthostate slabs beneath Room 7 is engraved with a double-axe mason’s mark (fig. 3.26 above). Like orthostate masonry itself, such marks are rare in mainland architecture. A branch-like sign adorns the Treasury of Atreus, and an inscribed trident marks a block from the South House at Mycenae, but these marks are later than the double-axe on the Englianos block. The date of the Atreus tomb is probably at least LH III, and the South House belongs to the LH IIIB period (Wace 1949, 16–19, 66, 119–31; Mylonas 1968, 197–98). Closer to the palace, and firmer in date, are a similar double-ax and a branchlike sign inscribed on two blocks of the stomion facade of Tholos I at Peristeria. These marks belong to the LH II period (Marinatos 1964; 1965; 1966; 1967, 10–14; ; Korres 1976, 470–550; Vermeule 1961, 121–22). Admittedly, the

The date of the introduction of orthostate masonry can only be approximated. The excavators found no conclusive evidence for the date of the orthostate wall beneath Room 7 (figs 3.26–3.27 above). It must belong either to a building earlier than the LH IIIB palace or to early LH IIIB (PN I, 94; PN III, 36), because the later LH IIIB floor was laid 0.18 to 0.30 m above the tops of the orthostates. For wall y beneath Room 57, of which only the socle remains (plan xiii), the stratigraphy and the general area had been disturbed before excavation, obliterating all means of dating. Nevertheless, the excavators assigned this wall to LH IIIA or early LH IIIB (PN III, 35–36). No mention is made of the dates of the other orthostate walls and socles, though the socle beneath the LH IIIB plaster floor of Court 88 must predate the floor (plan xix). As already noted, the orthostate wall to the southeast of Room 65 belongs to a phase of construction that immediately post-dated the pseudo-ashlar wall it displaced.

2

‘Building X appears to be older than Mycenaean IIIA’ (PN I, 12, 283); for walls SW58 and SW59, see PN I, 282. 3 In PN I, fig. 215, the pseudo-ashlar wall cuts diagonally across the foreground in the photograph; behind it and somewhat concealed by shadow stands the southeasternmost orthostate slab. 4 These are retaining walls and therefore they served a different structural purpose than the free-standing orthostate walls at Englianos (Wace 1949, 16–19, 119–31).

More certain than its introduction is the date of the cessation of orthostate construction, which occurred when

350

Building History evidence is sparse, but it does seem to indicate that the transition from pseudo-ashlar masonry to orthostate construction occurred in the Late Helladic I–II period. The orthostate phase must have ended well before the end of the LH IIIA period, when the ashlar style appeared.

Phase 5: Pier-wall construction In the Main Building, most of the interior walls were constructed in the new pier-wall style of construction, but some earlier ashlar walls, notably the southeast wall of Court 3, were skillfully incorporated in the new building. In the Southwestern Building, mainly in its northern corner, the new pier-walls seem to have been used to redefine the interior arrangement of the rooms. They were added to the rubble walls SW42, SW43 and SW49, and three new pier-walls were constructed: SW46, SW47 and SW48 (plans xviii and xix). Between the Main Building and the Southwestern Building, a new structure, containing rooms 60 and 62, was erected entirely with pier-walls.

Phase 4: Ashlar Masonry The transition to the ashlar style occurred before the end of the LH IIIA period. At the end of LH IIIA and the very beginning of LH IIIB, a fire swept over the hilltop and destroyed the earlier structures (PN I, 32–33, 339). The hilltop then underwent extensive levelling and grading, ending with a layer of earth being spread over the debris. This levelling stratum is still partially preserved throughout the site and is easily distinguished by its greenish colour, clay-like consistency and inclusions (potsherds, bits of plaster, bone, and small stones and pebbles).5

The ashlar style, which goes out of use at the beginning of LH IIIB, represents the last use of cut stone masonry on the hilltop. The builders of the last building phase, the pier-wall building system, did not quarry and shape any new blocks; instead, they remodelled and reused some standing walls, reused individual shaped blocks as door jambs and anta blocks, and reused broken blocks and fragments in rubble construction. All of the standing ashlar walls were built in LH IIIA and reused in LH IIIB.

The northeast ashlar facade of the Main Building belongs to LH IIIB, along with the rest of the building (PN I, 32–33). The levelling stratum abuts this wall, and so the northeast facade must date earlier than that stratum.6 The northeast facade of the Southwestern Building was built in the same ashlar style and probably belongs to the LH IIIA period, making the Southwestern Building, at least in part, earlier than the Main Building (PN I, 423). The Southwestern Building underwent several phases of remodelling, and its earliest phase is represented by this ashlar style northeast facade. Remodelling was undertaken later, with new walls built and old walls repaired using the pier-wall building system (similar to the remodelling that took place in the Main Building) and limestone slab construction.

To summarise, there are five styles of cut stone masonry at Englianos, which were used in four successive chronological phases. Construction with cut stone masonry appeared first sometime prior to LH I and ended at the end of LH IIIA, although existing ashlar walls continued in use in LH IIIB. At some point between LH I and LH IIIA, orthostate construction began and ended. Exactly when the changes took place cannot be determined with certainty because of the lack of stratigraphy.

There were several building phases during the period when the ashlar style was prevalent. At least three phases can be seen in the northeast facade of the Main Building, the last of which includes the introduction and use of ashlar shell walls. Room 32 was an addition to an existing building, and its northwest end was built with ashlar shell walls (plan x and fig. 3.35 above). At least two phases of ashlar style are evident in the northeast facade of the Southwestern Building as well. The clearest indication of this is the two sets of mortises, which represent two different wooden beam configurations (fig. 3.39).

Comparisons with Minoan Masonry Masons’ marks occur more frequently in Minoan than in Mycenaean architecture (Sakellarakis 1967; Graham 1987, 154–55; J. W. Shaw 1973a, 109–11). Their purpose is unknown both on Crete and the mainland, but they do represent a shared building practice. In addition to the presence of these marks, the masonry styles of Englianos are similar to those of Minoan Crete. Orthostate construction, the ashlar style and ashlar shell wall construction were all used in the Minoan palaces and other buildings, with only a few details in these masonry systems varying slightly between Minoan Crete and Mycenaean Englianos. For orthostate construction, the Minoan builders preferred larger blocks overall and finely cut and squared stone socles. Otherwise, all of the components of the Englianos orthostate wall system are present in Minoan orthostate construction and are assembled in the same manner. All contain a socle of cut blocks that supported a wall faced with orthostate slabs and backed by rubble masonry (J. W. Shaw 1973a, 83–92). Mortises in the tops of the slabs once secured dowels used to fasten horizontal and perhaps vertical timbers.

5

Th is stratum is still present in areas 103 and 106, beneath some of the floors of the Northeast Building, in the Northwest Area, and in some areas of the Southwestern Building. 6 The stratum was observed during the 1993 MARWP season while cleaning and clearing the area between the northeast facade, the Wine Magazine, the Northeast Workshop and beneath Court 88. Its presence, particularly its relationship to the northeast facade, was noted and recorded by Marion Rawson in her 1954 excavation notebook (p. 172).

351

Nelson – The Architecture of the Palace of Nestor Orthostate construction appeared in Minoan architecture in the Early Minoan period and continued to be built into the Late Minoan I period (J. W. Shaw 1983). The last known wall in the orthostate style is a portion of the north wall of Building T at Kommos in southern Crete. The wall dates to LM I and is slightly unusual in that the timber course above the orthostates was probably replaced with a course of ashlar blocks. The LM I date of the Kommos wall corresponds approximately to the suggested date of orthostate construction at the Englianos palace.

borrowed heavily from Minoan, and certainly the Englianos walls favour a Minoan legacy. The masonry styles were approximately concurrent with the tail end of the same styles in Minoan architecture; moreover, the sequence of the palace’s masonry styles matches similar transitions in Minoan architecture. At the Englianos palace, however, the changeover from orthostate to ashlar style masonry seems succinct and abrupt, whereas on Crete, the former was only gradually discontinued and replaced by the latter (ibid., 83). Cut stone masonry has a much longer history on Minoan Crete, but both the builders at Englianos and those on Crete were using the same styles and bringing about the same changes in stone masonry at roughly the same time.

The ashlar style in Minoan architecture appeared in the Middle Minoan I and II periods and was somewhat different than the same masonry system at Englianos (J. W. Shaw 1973a, 92–109). Minoan ashlar walls, like those at Englianos, rose from floor to ceiling with an exterior face of coursed ashlar and an interior face of rubble covered with plaster. However, the Cretan walls usually sit on socles of cut stone, include mud mortar between courses and usually lack timber inserts. A few examples of timber members do exist, though, including the timber beam serving as the fourth course in the western light well of the Hall of the Double Axes in the palace at Knossos (Evans 1964, vol. III, fig. 225), and the beam inserted between the second and third courses of masonry in the west wall of the Little Palace, also at Knossos.

The similarities in architecture noted here, particularly in terms of structural and chronological development, support strong communication, in whatever form, between Englianos and Minoan Crete. The contact between the two (and the means by which it was established and endured), and between Crete and Messenia in general, has been a topic hotly debated; for extensive discussion, often with opposing views, see Hägg 1982 and Korres 1984. The pithos style of burial and the Minoanising or imported Minoan pottery, jewelry and swords discovered in the Pylos Grave Circle, which lies about 145 m south–southwest of the palace, suggested to its excavator, Lord William Taylour, that a few of the tomb’s occupants may have been ‘naturalised’ Minoans (PN III, 134–76, esp. 153). The date of the interments of the Grave Circle spans the late Middle Helladic/early Late Helladic period to the transition from LH II to LH III (ibid., 148–53), which roughly corresponds to the masonry changes that took place farther up the hill. The extent of Minoan occupancy, Minoan influence or any form of interaction in the Middle and Late Bronze Ages, however, is beyond the scope of this study. Nevertheless, purely from the point of view of the masonry and the layout of the various buildings and rooms, it seems that there was some direct connection that permitted specific construction practices and techniques to be used, exchanged or transferred between Englianos and Crete. (A separate case for further Minoan architectural elements at Englianos is presented by Cooper in Part I, chapter 2.)

The attachment of wooden members to ashlar blocks was handled in a similar fashion by both the Englianos and Minoan masons: both used bedding cuts and mortises. In Minoan ashlar walls, these devices secured windowsills to ashlar blocks (J. W. Shaw 1973a, 174–85). Minoan masons seem to have cut mortises into the timber beams first, and then the matching mortises in the blocks, just as the builders at Englianos did. This is shown by the lack of fi xed intervals between mortises in Minoan ashlar walls (ibid., 175, 179–80); at Englianos, the spacing between mortises along the same stretch of wall also follows no discernible pattern. It was suggested above that the palace masons knew the lengths of the timber beams and adjusted the ashlar blocks to accommodate them, implying that wood was the more valuable material to both the Englianos and Minoan masons. If true, it also identifies a common, and very specific, shared construction practice. Minoan ashlar shell walls are unusual. Examples include the monumental west facade of the palace at Knossos, the walls of rooms iii, 5 and 6 of the palace at Mallia, and the north and west walls of room 49 at the palace at Phaistos (ibid., 88, 104). In all cases, both faces of the wall are built with cut stone masonry; between the blocks is a matrix of rubble and mud. In the west facade of the Knossos palace, the shell is composed of orthostates held together with dovetail clamps. Above the orthostate shell, the wall continued with a course of timber and then rubble construction.

The Development of the Englianos Hilltop The history of the architecture of the hilltop presented here is derived from the masonry styles discussed above and the evidence of such styles still preserved in the remains on the site. I distinguish between the hilltop proper and the ridge. The hilltop is the highest, plateau-like, area of the ridge (c.192–194 m above sea level) and covers approximately 15,000 square metres (plan i). It is relatively flat, and in the last phase of Bronze Age inhabitation, late LH IIIB, it was occupied exclusively by the palace complex. The edges of the plateau drop off rapidly on all sides save for the southwestern, where it slopes gently down into the Lower Town area. For my purposes, the modern fence around

Similarities in wall construction techniques and materials often draw the comment that Mycenaean architecture

352

Building History the site, erected by the Greek Archaeological Service, separates the ridge from the hilltop proper. Throughout all of the periods discussed below, both the hilltop and ridge were occupied.7

As was the case for the earlier MH wall on the north edge of the hilltop near the gateway (Blegen’s trench W6), the bedrock was cut to receive the wall. The L-shaped section on the southeast side of the steps is poorly preserved. Its thickness cannot be determined, but it can be traced for c.3 m to the southeast, where it stops abruptly. Between the flanking walls and among the steps there is no evidence of pivots or the like for an operable door or gate.

Middle Helladic Period The first certain building activity on the hilltop occurred in the Middle Helladic period and is represented by fragmentary walls and drains surviving largely on the plateau’s northwest edge. These remains are surveyed by Cooper in Part I of this volume, in which he proposes a large, quite sophisticated building on the hilltop in the Middle Helladic period (see Part I, pp.140–45). Though possible, the reconstruction hinges on hypothetical drain inlets (c, d and e) for which there is no evidence surving in situ. All of the Middle Helladic walls were built of rubble masonry, and some are quite substantial, for example wall b and the wall beneath Room 104 of the Wine Magazine. The width of these walls, upwards of 1.40 m thick, and their location at the edge of the hilltop, where they run parallel with the site contours, imply a fortification of retaining function and bring to mind the Middle Helladic hilltop settlement at Malthi, with its irregular rubble masonry circuit wall surrounding an agglomeration of small buildings (Valmin 1938). Among the Middle Helladic walls, one of them, wall B, is perplexing because it includes a squared limestone block within its rubble matrix. During the next major building phase, in LH I, these squared blocks were reused in pseudo-ashlar masonry. Prior to their re-use, therefore, they must have been part of a MH masonry building system that, unfortunately, does not survive anywhere on the hilltop.

The presence of the steps and the massive walls to either side prompted the excavators to search for a circuit wall. Of the more than 50 sondages sunk into the edges of the hilltop proper, only eight revealed traces of walls (PN III, 8–18). That in trench W6 has been shown to be MH, and another is LH I (the pseudo-ashlar supposed circuit wall, section 8 in the Southwest Quadrant). The others cannot be dated securely. The MH wall in trench W6 and the one beneath Corridor 26 probably continued in use, but wall B in the Northwest Area was immediately built over by new walls and structures. (See Part I, pp.55–73, for Cooper’s analysis of the Norwest Area stratigraphy.) The LH I pseudo-ashlar wall, section 8 (plan xvi), has a thickness of no more than 0.95 m. That thickness and the wall’s construction method do not match that of the walls of the Northeast Gateway; nor was it sunk or cut into the bedrock, but instead rests on a layer of sherds. Neither this wall, which may or may not belong to the same building program as the gateway, nor the gateway itself are convincing components of a protection system. Nevertheless, Blegen and his team regarded the hilltop as having been fortified in the Early Mycenaean period (PN III, 3). Until more tangible evidence is forthcoming, this conclusion should be regarded with caution.

Late Helladic I–II Perhaps the key to the interpretation of the gateway is its relationship to Tholos IV (plan i). The gateway aligns with the dromos of this tomb, which was built at approximately the same time: MH–LH I (PN III, 95–134). The steps may simply be a staircase that facilitated communication between the tomb, the area before it and the hilltop proper. Well-built retaining walls were placed to either side of the gateway to insure stability and durability, and as a result the gateway appears monumental and perhaps fortified. Tholos IV and the Northeast Gateway were clearly planned to align with one another, and since the staircase was by its very nature transitional (i.e., built to move people from one elevation to another), there were no doubt some structure(s) aligned farther up the hill towards the southwest. The gateway then may not have been a barrier, but part of a formal arrangement of landscape and built architecture.

Building activity on the hilltop during the Early Mycenaean period was greater and more extensive than during the Middle Helladic period. In LH I, the Northeast Gateway was set on the northeast edge of the hilltop (fig. 4.1; PN III, 7). The stairway is flanked by two L-shaped wall sections (plan xxxiii). The section to the northwest of the steps is much better preserved than its counterpart to the southeast, extending c.10 m northwestward and following the contours of the edge of the hilltop. No evidence was found for the continuation of this wall beyond this stretch. Its minimum thickness is reported as 1.40 m, but the published drawing (PN II, fig. 304) depicts it slightly larger. 7

For mainland absolute and relative chronology, see Rutter 1993, 755–57; Dickinson 1994, 9–22; and Shelmerdine 1997, 539–41. It must be kept in mind that the conclusions set forth here are based more on architecture and the sequence of building than on stratigraphy and pottery. Where possible, I correlate the former with the latter according to the published, dated material in PN I and III. J. Bennet (1999, 11–13) has also determined a history of Englianos based in part on the work conducted by PRAP. Y. Lolos (1994, 45–52) offers a brief history of Englianos and the immediate area (‘the heart of the kingdom of Nestor’) from the Neolithic to Byzantine periods.

Close to the Northeast Gateway and belonging to the same building phase are fragmentary walls in the Belvedere Area (plan xxxiv). Those that remain are humbly built rubble walls founded directly on bedrock. They may be remnants of domestic structures, though no complete building was found (PN III, 18–23).

353

Nelson – The Architecture of the Palace of Nestor

4.1 Early Late Helladic I phase plan.

354

Building History

4.2 Late Helladic I–II phase plan.

355

Nelson – The Architecture of the Palace of Nestor On the other side of the hilltop, the walls beneath Court 63 also belong to the early LH I period (plan xvii). These were nicely built, and they all reused cut blocks in the pseudo-ashlar style. The building or buildings that these walls represent must have been important structures, for the walls follow straight courses and intersect at right angles, showing a degree of care and skill in their erection. They were founded solidly on bedrock to insure their stability.

remains of magazine-like spaces (plan xiii and fig. 4.3). Fragments of two walls (γ and l) survive, which enclose a long narrow space. Sunk into the dirt floor and lining the walls are five pithoi bases. More bases remain on the other side of wall γ. The magazines had a facade of orthostate masonry, and the entire configuration looks suspiciously like the west magazines of the Minoan palace at Knossos. Cooper reconstructs a large building on the hilltop in the LH I period, a building that incorporates walls of both the pseudo-ashlar and orthostate styles. Within this building he also reconstructs a staircase wrapped around a lightwell within rooms 56 and 57 (see Part I, pp.80–82 above), the area in which I posit magazines. In accordance with Blegen (PN I, 221), I would put a staircase here in a later phase, because the cuttings on the ashlar blocks of wall 5E indicate that the stairs, Room 54, were inserted after the construction of this wall; moreover, one of the supporting walls of the staircase incorporated reused ashlar block fragments within a rubble matrix similar to the type of masonry used in the last major building phase of the hilltop. In sum, prior to the LH IIIA period, there existed at Englianos a very large building or group of buildings, built with Minoan-like masonry walls that were inscribed in a Minoan fashion with mason’s marks, and containing Minoan-like magazines.

On the northwest edge of the hilltop, more walls and drains were built (see Part I, chapter 2 above), and on the southwest edge a large, limestone slab-built wall, centred approximately beneath Room 65 (plan xviii), was laid parallel to the contours of the hilltop. Because it was built like the later retaining wall 17L, it may indicate a desire to expand the hilltop in this early period. The wall preserves two phases of construction, both entirely replaced with a later rubble wall farther down the slope, also running along the edge of the hilltop. The later rooms 65, 69, 74 and 80 covered this wall in the LH III period. One plaster floor belongs to this phase; it sits beneath the northwest and north portions of Court 88 (beneath the orthostate socle) and runs beneath the west corner of the Main Building (PN I, 293–94). Two small remnants of this floor can be seen in Plan xix, each labelled with an elevation of 192.37 m above sea level. The floor’s extent is unknown, but it appears to have been quite large. Nevertheless, it did not stay in use for long. Later in the Early Mycenaean period a new wall, using the new orthostate masonry style, was built on the plaster floor and formed part of a large building program that covered a good portion of the southwestern half of the hilltop. As mentioned above, an orthostate wall was built over the LH I pseudo-ashlar wall at the south and southwestern edges of the hilltop, and another was erected just northeast of the Main Building’s northeast facade at Room 34.

Beyond the hilltop proper, the first extant evidence of building activity in the Lower Town appears in the Early Mycenaean period (PN III, 47–68, esp. 52–61). Five trenches were sunk in the terrace immediately below the southwestern facade of the Southwestern Building (plan i). Though occupation may have begun in Middle Helladic

The remains of the orthostate walls that were being built at this time are not sufficiently preserved to determine their extent or the outline of any structure, but they may have comprised more than one building. The walls have their orthostate slabs facing outward, that is, towards the southwest, southeast, northeast and northwest, as if they once were facade walls for a large rectangular building (fig. 4.2). The area bordered by the orthostate walls measures more than 2200 square metres. If this represents a single building, it would have been approximately 500 square metres larger than the later Main Building. Though exact dimensions and layout are unknown, the architecture on the hilltop at this time was monumental both in masonry and in size. In addition, the multiple drains operating at this time (see Part I, pp.145–51) suggest a preconceived notion of a large building prior to its construction. The original layout of this building or buildings is now gone, since later building activity either destroyed or concealed it. However, beneath rooms 56 and 57 are the

4.3 Early orthostate walls and rubble walls beneath rooms 55–57, Main Building.

356

Building History times, the earliest walls belong to the LH I–II periods. Most of the walls are rubble masonry, and the few floors uncovered are of clay and pebbles. Walls average 0.57 m in thickness8 and were probably part of small structures. The largest wall, the northeasternmost wall in Lower Town trench II, is a bit more substantial, measuring 0.91 m in width; it rests on an earlier, thicker predecessor. Both walls run parallel to the contours of the terrace and may have been built as retaining walls. In 1996, MARWP opened up the very north end of Lower Town trench I, which is now partially within the area enclosed by the site’s fence (plan xvi). Wall 1 was built with rubble;9 if its present state is any indication of its original construction, the wall was quite thin and built with minimal concern for its appearance and stability. In contrast, wall 3, of which MARWP exposed only the northeasternmost end, was more sturdily built, with rubble masonry that included some cut stone. Occasional, partially worked blocks in rubble masonry are not unusual; in the Belvedere Area, at least one worked stone was used, laid in the lowest course as a quoin at the corner of a small structure formed by walls M and N (plan xxxiv).

suggesting that they were set in place at the same time. In addition, many walls in this suite of rooms bond with one another. The northwest wall (SW50 in plan iii) bonds with the southwestern and northeast facades, and the southeast wall (SW52) bonds with southwestern facade. Parts of walls SW50 and SW51 were later remodelled, so the original configuration of the doorways leading in and out of the rooms is unknown. Smaller rooms were laid out to either side of rooms 64 and 65. The northeast facade extended northwestward, as indicated by the T-shaped anta block 18L.b and the bedding stones upon which the ashlar blocks were set. This section was later robbed of its cut blocks (plans xiv and xix). Behind the facade, the walls were constructed in several different masonry styles and represent both original and later remodelled walls. The southwestern facade probably also extended northwestward, in the same direction as the northeast facade (wall 18L), but its original course and construction style are uncertain. Sections 1, 2 and 3 of wall 17L (plan xx) were built of heavy limestone slabs; since they are not ashlar style walls, the excavators thought they represented later repairs to the facade (PN I, 278). The limestone slab walls are definitely later than section 4 of the southwestern facade, since they abut it, but how late cannot be said for sure. Nonetheless, Building A must have contained rooms laid out to the northwest of Room 65. If the northwest wall (SW50) of this room had been built as a facade wall, it should have been built in the ashlar style, like the other facades; its rubble construction makes it an interior wall. Sections 1, 2, and 3 of the facade, although they represent later building activity, no doubt replaced an earlier wall following approximately the same alignment.

Building on the ridge expanded throughout the Early Mycenaean period, and by the end of the period the estimated settlement size had reached some seven hectares (Zangger et al. 1997, 430). The lower terraces immediately below the hilltop proper were still the main areas of settlement, but sherd densities increased in the areas to southeast and east along the more gently sloping sides of the ridge. Occupation of the Lower Town area continued through the LH IIIB period, coinciding with the last major phase of settlement on the hilltop proper. Late Helladic IIIA The Northeast Gateway remained in use in LH IIIA, and building activity continued in the Belvedere Area and Area 106 (fig. 4.4; see Part I, pp.35–43, for Cooper’s analysis of the phases of Area 106). At least two, and perhaps three buildings (labelled A, B, and C in fig. 4.4), were constructed in this period, and all used the new ashlar style masonry that was introduced at this time. On the southwestern edge of the hilltop, Building A was erected with a massive southwestern facade. The wall represents the last phase of expansion of the hilltop in this direction, serving both to retain earth and to support rooms built above it. Six sections of the wall can be reconstructed, all built in the ashlar style. The northeast facade (wall 18L in plan III) was built in the same style.

Even greater difficulty is encountered in attempting to interpret the profusion of walls to the southeast of Room 65 (plans xiv–xviii). The excavators attempted a reconstruction (PN III, fig. 306), but some of the walls used in their plan run over the top of the southeast wall of Room 65 (SW52) and therefore must belong to a later phase. Others abut wall SW52, so they too are later, and still others were destroyed to make room for SW52. Despite these problems, since the southwestern facade extends southeastward beyond Room 65, there must have been rooms and built spaces in this area, and some of the abutting walls, though later, may have been built shortly after the erection of wall SW52.

The interior arrangement of most of the rooms of Building A is difficult to reconstruct. The exceptions are halls 64 and 65, which were probably laid out as a unit. The surviving column base in Hall 65 and the three bases in Hall 64 all lie within centimetres of the same elevation, c.192.33 masl, 8 9

Building A’s southwest facade ends with a new set of steps at its southeast end. In its original configuration, the staircase was flanked on the northwest by section 10 of wall 17L (see plan iii) and on the southeast apparently by the northwest wall of Building X. Building X either had remained standing from an earlier period or had been partially rebuilt. In its present condition, section 10 is better preserved, with two courses of ashlar in situ;

Measured from the published drawings in PN III, figs 313–18. Wall designations are Blegen’s (PN III, fig. 314)

357

Nelson – The Architecture of the Palace of Nestor

4.4 Late Helladic IIIA phase plan.

358

Building History the northwest wall of Building X is stripped of most of its upper courses. The steps, too, are poorly preserved. Only one step, a poros block cut to form a tread and riser, remains in situ and bonds with section 10. However, some of the steps’ foundations are still in place, and these permit a plausible reconstruction (fig. 4.5).

Building B was part of the same building programme that included Building A. The building was smaller than Building A, corresponding roughly in size to the northeast suite of rooms (29–34, 36, 39 and 40) of the later Main Building. Building B’s northeast ashlar facade replaced the earlier orthostate facade located outside Room 34. Part of the facade was built as an ashlar shell wall, which returns twice at Room 32 (plan x). Thus, at least three of Building B’s facades were done in ashlar. The internal layout of the building is unknown because it was obscured by the later LH IIIB remodelling. There is only one break in the floor in this area of the Main Building; it occurs in the doorway between rooms 29 and 33. Here, a rubble wall, probably LH IIIA, seems to have served as the doorway’s threshold in the LH IIIB period (PN I, 161). Stairway 36 also belongs to the later remodelling phase. Its southeast anta block is flush with the riser of the first step, while the northwest block is not. The offset suggests that one block was in place from an earlier building period and that the other anta block and steps were added later.

The staircase comprised a single flight of two groups of steps of different tread dimensions. The first group began about 1.50 m. in front of the ashlar facade and was composed of four steps with low risers, c.0.10 m., and broad treads, 0.50 m. A broad plaster ramp led up from the southwest and abutted the lowest step. The next group began at the exterior face of the facade and continued to the northeast with six steps that had risers of the same height, c.0.20 m., but deeper treads, c.1 m (see also Part I, pp.75 above). The steps and the ashlar facades certainly give Building A an impressive veneer. It was the largest of the three buildings on the hilltop in LH IIIA and later became the LH IIIB Southwestern Building. Blegen and Rawson suggested that the Southwestern Building was the first palace on the hilltop (PN I, 423). With its size and construction material, the building was no doubt an important structure at Englianos.

The ashlar shell wall of Room 32 most likely indicates its original function: such a wall was built to face the elements, particularly the rain. Lightwells in Minoan palaces were built in this way – they brought light into one or more rooms of the building while at the same time keeping out the wind (Graham 1987, 167–70). A lightwell

187.33

18 9. 57 188.80

187.67

187.33

4.5 Reconstruction of stairway in Southwest Quadrant.

359

Nelson – The Architecture of the Palace of Nestor in Room 32 seems appropriate, as a strong wind comes up from the deep valley to the north of the palace every day, which the builders no doubt wished to block.10

his reconstruction belong to various periods, and some certainly were not used in conjunction (see also Dickenson 1994, 153). Nevertheless, Kilian’s idea of an early building that was more akin to Minoan than mainland architecture is evident in these three ashlar buildings – if not in plan, then certainly in wall construction.

The third of the three large buildings of the LH IIIA period was Building C. Portions of three of its facades were incorporated in the later Main Building and can still be seen: the northwest (exterior walls of rooms 43, 46, 48 and perhaps 50; plan ix), northeast (southeast wall of Room 3; plan ix) and southeast (wall m in plan xiii). The building may have been freestanding or perhaps a wing of Building B. The correct interpretation hinges on whether the space between B and C (later Room 41) served as an entrance to or simply an alley between the two buildings. Further excavation is required to determine the relationship.

The alignment of the possible central open space is also interesting in comparison to that of two Minoan palaces. The central court at Knossos aligns towards Mount Juktas, and the central court at Phaistos points towards twin-peaked Mount Ida. If indeed Pylos had a central, open space, it aligned towards the twin peaks of Mount Lykodemos to the southeast. By the LH IIIA period, sites such as the Menelaion and Tiryns already had buildings that were megaron-like in plan;12 Tiryns and Mycenae became LH IIIB palaces with central and prominent megarons fronted by open courts. In comparison to these Peloponnesian palatial centres, the building development at Englianos is slightly unusual. Most notable is the arrangement proposed here and by Kilian for buildings A, B and C, consisting of blocks of buildings flanking a central open space – an organisation which has so far not appeared at any other mainland site. The closest parallel, as Kilian recognised, is the Minoan palace, and although the period of Minoan court-centred palaces had all but passed,13 most having gone out of use by LH IIIA, Englianos did have one architectural comrade – the palace at Knossos, which still endured as a courtcentred building (Dickinson 1994, 21–22; Rehak and Younger 1998, 92–93). As will be seen below, it is not until the following period that the architecture of Englianos comes to resemble that of the other Peloponnesian palaces. In the LH IIIA period the site is notable for its Minoanstyle architecture on the hilltop and for the Minoan and Minoanising material from the Grave Circle on the ridge.

What lay between the three buildings is now covered by the remains of the LH IIIB palace, principally the megaron. The excavators dug a pit underneath Corridor 25 for the foundations of one of the metal posts of the protective shed roof (PN III, 33), but nothing was found except disturbed stratigraphy, which they attributed to the grading and levelling of the hilltop at the end of LH IIIA. They also discovered a plaster floor and some rubble walls beneath Room 21, which they dated to LH IIIA (ibid., 34– 35). However, the walls and floor relate architecturally to, and lie at the same elevation as, the LH I–II plaster floor beneath Court 88 and the rubble wall that runs across the northwest end of the court, meaning that they must belong to a period earlier than LH IIIA. All of the facades of Buildings A, B and C that face the central area were built in ashlar. In Minoan masonry, the cut stone faces of ashlar-built walls generally faced the exterior of the building and bore the brunt of the weather.11 The same may have been true for mainland architecture, though insufficient evidence survives for such a broad conclusion. The north wall of the court before the megaron at Mycenae and the southeast wall of Court 3 at Englianos were ashlar-faced, and both courts were open to the sky. If this general practice applies to LH IIIA Englianos as well, then the area between the buildings may have been one large open space – a configuration that recalls the layout of Minoan palaces. (Cooper also posits large central courts in his two early buildings; see Part 1, pp.139–54 for detailed discussion).

Late Helladic IIIB Earlier Phase More walls and structures survive at Englianos from the LH IIIB period than any other phase (fig. 4.6). Nevertheless, at the beginning of this period the Northeast Gateway went out of use and occupation of the Belvedere

In an attempt to interpret the early buildings at Englianos, K. Kilian (1985, 213–17) reconstructed an early arrangement of a central open space flanked by several buildings, all of which he dated to the end of the Middle Helladic and the beginning of the Late Helladic period (fig. 4.6). His proposal for the plan of the building(s) essentially echoed that of the overall architectural arrangement of a Minoan palace. However, some problems do exist in Kilian’s reconstruction. He thought the patches of plaster floor beneath the west corner of the Main Building were the remnants of a large plastered court, but the true extent of these fragments is unknown. Also, many of the walls in

10

The wind is strong and gusty. Today, the north corner of the protective metal shed is tied to a nearby oak tree with a steel cable to prevent it from blowing away. 11 As noted in the case of Room 32, shell walls and walls of interior courts and lightwells are exceptions (J. W. Shaw 1973a, 92–107). Orthostates were also usually used on exterior wall faces (ibid., 83–92). 12 For thorough discussions, including more sites with early Mycenaean palace features, see Kilian 1985, esp. 209, 212–13 on the Menelaion and Tiryns; Barber 1992, esp. 11–13 on Tiryns; and Dickinson 1994. 13 For discussions of the Minoan palace periods, their dates and some of their prominent buildings, see Dickinson 1994, 145–52; Watrous 1994, 717–53; Rehak and Younger 1998.

360

Building History area ceased (PN III, 7, 23). As a result, the northeast half of the hilltop proper apparently stood vacant during the last phases of the Bronze Age.

No new cut-stone masonry walls were erected in this period. Instead, and where suitable, the new buildings incorporated older ashlar walls.

The opening of the LH IIIB period was marked by a large fire that destroyed the LH IIIA structures (PN I, 34, 423); the Lower Town suffered from the same disaster (ibid., 19). This prompted a major remodelling that included the razing of some structures and the levelling of the hilltop in preparation for new building. The preferred building method for the LH IIIB architecture was the new pier-wall building technique, particularly in the Main Building.

Floor plans also changed. Out of the older Buildings A, B and C, two new structures emerged that together surpassed their predecessors in size. Buildings B and C became multi-room suites in the new Main Building, and Building A became the Southwestern Building. In the Main Building, all remodelling was done with pier-walls. In the Southwestern Building, pier-walls were added either wholly new or tacked onto existing rubble walls.

4.6 Buildings and central court in proposed reconstruction of Palace of Nestor by K. Kilian (1985).

361

Nelson – The Architecture of the Palace of Nestor All associations with Minoan architecture disappeared. Ashlar walls were no longer built, and a new megaron (rooms 4, 5 and 6), which has clear roots in mainland Helladic architecture, was built in the large, open, central space of the earlier period. The plan of the Main Building closely resembles its Peloponnesian contemporaries, most notably in its structured entrance route of propylon, open court and megaron.

former. The engineering survey conducted by MARWP (Part I, pp.31–34) determined that columns 6W, 6S, 4SW, 2 and 1 align along the same line (plan v). Columns 4NE, 6E and 6N also follow a straight course. The precise alignment could have been achieved only if the bases were set or their locations staked before the construction of walls 5B and 5C around Room 5 (plan iii), which would have blocked the line of sight required to lay out each column course.

Other structures were built, aside from the new Main and Southwestern Buildings. The Southwest Quadrant area was reconfigured. The steps at the southeast end of the Southwestern Building went out of use when the new Main Drain sliced through both them and the sloping plaster ramp. A new set of steps was erected at the southeast of end of Building X; their rubble-built foundations can be seen in Plan xv at the south corner of Building X. These replaced the older steps and served the same purpose – to connect the Lower Town area with the hilltop proper. They led directly to the large plaster court (58) before the propylon (1). None of the treads survive, but some of the foundations do and show the reuse of ashlar blocks. Since the northwest edge of the steps was built over the southeast wall of Building X, this structure must have been altered in some way or portions of it re-used at the end of LH IIIA Its prominent location at the edge of the hilltop suggests that the building was not simply left to decay, and it may have supported a now missing structure. The pre-existing southeast wall of the building served as the supporting foundations for the new steps. The northeast wall, abutting the southeast wall, must have been built at this time.

The preplanning and skill used in the setting of the column courses and the doorjambs is also evident in the walls of the Main Building. Many of its long walls (designated with an L in plan iii) deviate no more than a few centimetres from straight lines, and intersections with cross-walls often fall within half a degree of true right angles. The method of laying out the building is unknown, but the anta blocks of the pier-wall system, especially in corner, wall-end and doorjamb locations, may have been positioned first and then used as benchmarks to lay out the walls between them. Whatever the practice may have been, the precision of the surveying techniques, along with the development and use of a calcium carbonate-enriched mortar in the pier-wall building system, designate the early LH IIIB phase of remodeling as a period of highly sophisticated building technology at Englianos. Later Phase The Main Building and the Southwestern Building remained in use in the later phases of LH IIIB, though modifications were made to the former (fig. 4.8). The walls of the southwest suite of rooms are of rubble construction, and since they were not built with the pier-wall building system, Blegen and Rawson thought they may have been late repair work (PN I, 46, 130). During the remodelling, the corridor running along the southwest side of the megaron was blocked with two rubble walls, and two new doors were let into the southwest facade, affecting the circulation patterns through the building. (plan vi). While the Main Building remained a single entity, the new circulation routes served effectively to divide it into two conjoined buildings: rooms 18, 19 and 20, and the rest.

The later wall may have been built at this time to connect the steps with a new freestanding structure containing rooms 60 and 62, built just to the southwest of the Main Building (plan xvii). Its southeast wall runs over the top of the old set of steps and perhaps prompted the construction of the new steps to the southeast. Its northwest wall runs over the top of sections 9 and 10 of the Southwestern Building’s southwest facade, which must have put them out of use. The postholes near the long walls of Room 60, as well as the numerous pots recovered, suggest that the small building served as a storage facility (PN I, 237–41). Another new building, Room 82, was constructed to the northwest of the Southwestern Building (plan xix). The few finds from the building do not provide sufficient evidence for its function (ibid., 289–91), but its similarity to rooms 60 and 62 suggest that it too may have been an ancillary storage facility.

Room 27, built with rubble walls, was added onto the north corner of the Main Building (plans vi and x). Pithoi line all four walls, indicating that the new room provided extra storage space for the Main Building.14 Courts 42 and 47 were also appended to the Main Building (plan xxii). Their walls are rubble, though ashlar blocks were reused in the lowest course of a short section of the courts’ northeast wall. This addition effectively cut off all admittance to the Main Building from the northeast, as well as obscuring the finely-built and impressive ashlar facade.

The hilltop architecture was clearly preplanned in this major phase of remodelling at the beginning of the LH IIIB period. Cooper (Part I, pp.151–52) analyses the drainage system of this period in detail. The doorjamb blocks and many of the column bases were put in place before any superstructure was built. Most of the jamb blocks had to be set in place before the walls were erected, because as is well documented in the doorway state plans drawn by Piet de Jong (PN I, figs 431–69), the latter partially overlay the

Two entirely new buildings were also erected in the late LH IIIB period: the Wine Magazine and the Northeast 14

Because of the fragility of the in situ pithoi, MARWP did not uncover them, and they do not appear on the state plan.

362

Building History

4.7 Early Late Helladic IIIB phase plan.

363

Nelson – The Architecture of the Palace of Nestor

4.8 Late Late Helladic IIIB phase plan.

364

Building History Building. They, like the modifications to the Main Building, increased the hilltop’s storage and manufacturing spaces and/or its administrative capacity. Both buildings were humbly built, with rubble construction. They must have served purely utilitarian functions and apparently did not require the refined building materials and methods of the Main Building and the Southwestern Building.

determining factor. The pier-wall building system is a more time-consuming and difficult construction method than rubble construction, which used only three building materials (rubble, mud and plaster) in a much shorter process of erection. For load-bearing walls, both cost and structural requirements must have determined the building material and construction technique. Limestone slab construction was less time- and energy-consuming than coursed ashlar work because it involved only a single material, as opposed to three in ashlar masonry (ashlar blocks, wood and rubble), and did not require masons skilled in cutting blocks square and laying them in neat and level courses. This building system is stronger and more durable than ashlar masonry for several reasons: the material is heavier and harder than poros limestone, and the construction is uniform throughout the entire thickness and length of the wall. The three materials in ashlar masonry generate a risk that one will separate from another. If the northwestern end of the southwestern facade of the Southwestern Building is indeed repair work, the builders chose a more cost effective and durable building system than the original construction technique.

The two new buildings also affected circulation on the hilltop. The entrance to the Wine Magazine is on its northwest side, so that the structure turns its back completely away from areas 103 and 106 (plan ii). Ramp 91, which preceded the erection of the Northeast Building but antedated section C of the aqueduct (Shelmerdine 1987, 560–65; F. A. Cooper 1994), directed all movement up the ramp northwestward and then around the west corner of the Northeast Building. On the other side of the Main Building, the rubble wall separating area 61 and Court 63 forced circulation through a single doorway. These additions, modifications and circulation paths indicate that by the end of LH IIIB, the primary outdoor space of the hilltop was probably the large plaster court, area 58, on the southeast side of the Main Building. From here, access was granted to the Main Building, funnelled through either ramp 59 or 91, or down the steps into the Lower Town. Further emphasis on this area of the hilltop was created with Room 93, Portico 94 and Court 92 of the newly added Northeastern Building (these spaces together created the north corner of area 58). In Court 92 sits a nearly cubic block of limestone plastered on its upper surface and frescoed on its four vertical faces (plan xxvi). The block may have been an altar (PN I, 301–02); if so, it is the only identifiable religious space on the hilltop. The opposite side of area 58, towards the Bay of Navarino, may have been bounded by a colonnade, of which one column base still survives (58SW on plan v). Whatever its original function, the architectural arrangement about area 58 indicates that it served as both a major transitional space into the palace domain and an organised outdoor space.

Indeed, all later LH IIIB repair work, remodelling and new building were carried out using construction methods that were less expensive and required less skill to erect than the existing walls and buildings. The utilitarian functions of storage sheds such as the Wine Magazine or an administrative building such as the Northeast Building apparently did not require expensive masonry techniques. Equally significant and somewhat unfortunate is the change in aesthetics that some of these repairs seem to indicate. At Englianos, there was a predilection for cut stone masonry in monumental buildings, as seen throughout several building periods. In the final repair and remodelling, however, aesthetically impressive walls were sacrificed. Ashlar was replaced with limestone slab construction, and a good portion of the northeast facade of the Main Building was concealed by the addition of courts 42 and 47. This last phase of architecture at the palace is characterised by more cost-effective building systems and simultaneously by the sacrifice of a longstanding aesthetic tradition.

The new architectural changes and additions that took place during LH IIIB were executed using different building materials and different construction techniques than those already in place. The reason for this is somewhat enigmatic. In the Main Building, simple partition walls covered in plaster were no doubt indiscernible from the surrounding pier-walls, which were finished in the same manner. On the other hand, the supposed repair of the southwestern facade using large limestone slabs roughly fitted together (as opposed to the original ashlar masonry that was still extant in section 4) projected an entirely different aesthetic, replacing an ordered, smooth surface with one that was coarse and irregular.

At the end of the LH IIIB period, the Englianos hilltop suffered a great destruction from which it never fully recovered. There is evidence for some building activity in the post-Bronze Age Geometric to Classical periods (see Part I, chapters 8 and 9), but it compares neither in size nor in masonry skill to the achievements made many centuries before.

The Mycenaean builders no doubt took into account cost and structural requirements. In the case of partitions, which are not load-bearing walls, cost may have been the

365

appendix a TERRACOTTA BUILDING ELEMENTS

Terracotta was not used in the masonry building systems discussed above. However, it was employed for nonloadbearing architectural purposes at Englianos, in drains (both pipes and covering tiles), plaques and chimney pipes.1

was needed to adjust or raise the mudbrick courses it supported. Alternatively, the broken plaques may simply have been used in an early attempt at recycling.

Drain Pipes Terracotta was used in two types of drain systems on the plateau: pipes and channels (the latter formed with inverted Π-shaped tiles). Branch B of the aqueduct was apparently composed entirely of terracotta pipes. The northeastern end of this branch did not survive, but it no doubt connected somewhere at the point where the aqueduct (labeled branch A) forks into branches B and C (plans xxii and xxiv; PN I, 183). Only two sections of pipe remain in situ: a collapsed, though somewhat nicely preserved fragment embedded in a square opening in the northeast rubble wall of Court 42 (ibid., 137; see also Part I, fig. 4.9) and a fragment of another section about 4.50 m directly northeast of the first (Area 101; ibid., 242). The wall thickness of the pipes varies from 0.01 to 0.03 m. The inside diameter of the wall pipe is approximately 0.15 m.3

Plaques Terracotta plaques line the lower portions of the walls of the small kiln near the Belvedere Area of the hilltop (fig. 1.6 above). The plaques have similar material consistencies and colour as the mudbricks that make up the kiln, though with fewer inclusions.2 Though much weathered, they appear to be hand-molded; they vary in length from 0.17 to 0.43 m (PN III, 20). Plaque thickness is consistent at about 0.02 m, measured on the top, exposed edges. As with the top course of mudbrick on the kiln’s central tongue, ploughing and erosion have shaved off the upper surfaces of the plaques, so finished heights cannot be determined; the maximum preserved height is c.0.34 m (ibid.). The kiln is partially subterranean, and the terracotta plaque lining may have served to protect and preserve the marly limestone bedrock into which it was sunk. Repeated exposure to high temperatures during the firing process would have weakened the limestone; eventually, any pressure from above, such as that created by the (now missing) walls of the kiln’s superstructure, would have caused the limestone sides of the kiln to crumble and possibly collapse. The plaques may be regarded as a fireproofing material.

The total length of the pipe fragment in Area 101 is no longer extant, but the largest preserved diameter measures c.0.21 m, and the smallest, c.0.18 m. Tapering sections would have simplified the connection between individual pipes 1

The terracotta chimney pipes recovered in the palace debris are not included here because unlike all of the other architectural elements at the site, I did not study them fi rsthand. See PN I, 81–82, and Nelson 2001, 66–70. 2 The colours vary between 5YR5/6 yellowish-red and 5YR6/6 reddishyellow on the Munsell scale. In comparison to the mudbricks, very little, if any, vegetable matter was mixed with the clay. Rounded and granule size (0.002–0.004 m.) stones are included. 3 Blegen and Rawson (PN I , 183, n. 57) measured the diameter at 0.14 m. More sections and fragments may have been found by the original excavators, but they were not reported.

Inserted between the first and second brick courses of the central tongue of the kiln is a thin course of terracotta tile or plaque fragments, c.0.01–0.02 m high (figs 1.61.7 above). The fragments all appear broken, with no recognisable straight faces or corners. This is a curious construction feature, but perhaps the terracotta course 367

Nelson – The Architecture of the Palace of Nestor in the course, with the narrow end fitting into the wider end of the next pipe. No sealant, such as clay or plaster, appears to have been used in this water system, and the builders may have depended only on the tapering sleeve connections to prevent leaks.4 (See further discussion in Part I, pp.138–39.) Terracotta pipes do not seem to have been employed at other mainland sites, though the term appears in site descriptions.5 The use of the term ‘pipe’ is often misleading in the literature, for in every case it seems to refer to terracotta drain tiles, which are discussed below.6 The Minoans used true pipes, and a number of them are preserved (J. W. Shaw 1973a, 198–204, 235). Not all taper, but those that do were inserted into the corresponding end of another tapered section. Some were equipped with handles and ridges to prevent over-insertion.

Drain Tiles A portion of branch C of the aqueduct was laid with handmade terracotta drain tiles; their cross-sections resemble inverted Π-shapes (labeled C1 in plan xxiv; see PN I, 327, fig. 236). The terracotta tiles were capped with unworked limestone cover slabs.7 The total length of branch C1 is c.5.90 m, and only the southernmost portion, c.4.20 m long, consists of terracotta tiles (figs A.1–A.2). The north end was built with small, slab-like stones. These were set upright to line the channel and covered by slabs laid horizontally on top of them.

A.1 Plan of branch C1 of the acqueduct at the Northeast Building, showing position of terracotta tiles.

Tile dimensions vary (table A.5): the shortest measures 0.47 m (T5), and the longest, 0.89 m (T3; fig. A.3).8 The vertical walls of individual tiles are generally consistent in thickness throughout a tile’s length, but among all five tiles, wall thickness varies from 0.014 to 0.020 m (fig. A.4). Similarly, floor thickness is consistent for individual tiles, but varies among all tiles. In general, the floor of a tile is thicker than its walls. 4

Some water must have seeped through the walls of the pipes and the connections between them. The lack of a sealant may also indicate a relatively low volume of water passing through the pipe. It should also be kept in mind that the two preserved pipes are very fragmentary. 5 Küpper (1996, 64) did not think that they were in use prior to LH III, but provided no examples belonging to this period. 6 Examples include Dörpfeld 1885, 233–34, 246; and Blegen 1928, 35. More published illustrations would help to clear up any misunderstandings. 7 During MARWP’s clearing of this drain, it became apparent that Blegen and Rawson removed the cover slabs and excavated the soil within the drain channels. When they fi nished their investigations, they back-fi lled the drain and replaced the cover slabs seemingly in the same positions in which they found them. MARWP’s permit did not allow for the removal of all of the cover slabs, and so the drain was only partially observed. 8 Blegen and Rawson (PN I, 327, n. 2) reported five tiles with lengths of 0.87, 0.43, 0.87, 0.65 and 0.46 m; their total length (3.28 m) does not equal 4.20 m.

A.2 Branch CI of the acqueduct, showing terracotta tiles, from the north.

368

Terracotta Building Elements Like the pipe fragments in Area 101, the width of some tiles tapers in one direction. Only three channels could be fully observed: two (T3 and T5) taper, and one (T4) maintains a constant width throughout its entire length; tiles T1 and T2 could not be fully measured. Tile T3’s north end is the widest, 0.31 m; its south end is 0.27 m in width (figs A.3 and A.4). In contrast, T5’s south end is wider than its north, 0.27 m versus 0.24 m. Apparently at one time T4 fit

into the north end of T3 and the south end of T5 to form a kind of bridge between the two. Like the pipes, one would expect the taper to correspond to the direction of water flow, with the narrow end of one tile fitting snugly into the wider end of the next tile down-slope.11 The overlapping of tiles would facilitate water flow while preventing water loss. However, the drain is not constructed in such a way because T3 and T5 taper in opposite directions. This whole section of drain appears to have been either installed or repaired in a shoddy manner. Along the same drain and beyond the terracotta tile section, branch C doglegs southeasterly around the outside corner of Room 97 (plan xxiv). At the dogleg, branch C1 ends in a basin of sorts, constructed of ashlar block fragments and rubble (PN I, 327–28, fig. 224). From the south end of the basin, branch C2 continues (built with limestone blocks cut with channels, which also have inverted Π shapes) and runs down Ramp 91 flanking the southwest facade of the Northeast Building. The elevation drops from the corner of Room 97 down Ramp 91 towards the southeast. As with the terracotta drainpipe, no sealant was used on the joints between individual channel segments.

A.3 Branch C1 of the acqueduct, showing terracotta tile T3, from the northeast.

Drain tiles were more widely used in mainland architecture than pipes (Shear 1968, 434–35; Küpper 1996, 64), and in palatial contexts there are examples from Mycenae (Wace 1921–23, 91, 187) and Tiryns (Dörpfeld 1885, 233–34, 246). For the latter, the drain channel of the large slab of the bathroom (room X) issues into a section of drain composed of tiles. All of the tiles are tapered to fit nicely into one another. Dörpfeld illustrated one: it is 0.68 m. long and tapers from 0.22 to 0.28 m at its widest end (ibid., 234, fig. 118). The drain tiles from Room 30 of the Potter’s Shop at Zygouries taper much more severely than those of Englianos and Tiryns, from 0.24 to 0.38 m (Blegen 1928, 35). These tiles, four of which remain, are also longer, averaging 0.92 m, and they do not seem to have been capped by any cover. The Minoans too used terracotta drain tiles, but unlike the mainland examples, most of them maintain a constant width throughout their length (J. W. Shaw 1973a, 201–04, 235–36); only at the palace at Mallia do the drain tiles taper. Instead of fitting snugly into one another, the abutting joints were sealed with plaster.

A.4 Profiles of terracotta tiles of branch C1 of the acqueduct at the Northeast Building.

Tile T1 T2 T3 T4 T5

Length (m) 0.879 0.4310 0.87 0.65 0.46

Minimum width (m) 0.23 0.14 0.27 0.20 0.24

Maximum width (m) 0.23 0.14 0.31 0.20 0.27

Depth (m) 0.15 0.14 0.15 0.12 0.14

9

Blegen and Rawson measurement (PN I, 327, n. 2) Ibid. 11 Blegen and Rawson (PN I, 327) reported that this is the case for some of the tiles, but they did not specify which. MARWP found no tiles overlapping. 10

Table A.1 Dimensions of terracotta drain tiles, branch C1 of the acqueduct.

369

appendix b CLAMP CUTTINGS IN ASHLAR BLOCKS

upper surface of an orthostate block.2 It may have been a corner orthostate or part of a door anta, since its left vertical face has no anathyrosis. Though badly damaged, the cutting’s depth can be measured at 0.08 m; the thickest part of the tail measures 0.11 m. The cutting, like the one

Three ashlar blocks on the hilltop preserve clamp cuttings. Each block contains one dovetail-shaped cutting in its upper surface at its rear edge (fig. B.1). Unfortunately, the original placement within the walls into which these three blocks were built cannot be determined now, because the blocks were moved from their find spots. It is also unknown whether or not the clamps that once fit into the cuttings were carved from wood or made from some other material. Minoan clamp cuttings similar to those at Englianos are assumed to have held wooden clamps, based on the lack of recovered examples and the presence of carbonised remains associated with the cuttings (J. W. Shaw 1973a, 157–61). The largest cutting was carved into an ashlar block that now lies against the site’s enclosing fence, just south of Building X (figs B.1, A and B.2).1 The block’s front face was partially damaged, but otherwise it is fairly well preserved. The block is unusual because its lacks the anathyrosis common on other ashlar blocks and because its rear face was finished as smoothly as its other three. The dovetail mortise was cut near but not quite on the centerline of the rear edge of the block. The block is 0.96 m in length, and the centreline of the mortise falls at 0.40 m, measured from the right face (facing the front of the block). The mortise was also not aligned perpendicular to the rear face of the block, but cut slightly askew. At its thickest dimension, the mortise is 0.19 m broad, and at its narrowest (where it meets the rear face), 0.08 m. Its depth is 0.09 m. The other two clamp cuttings are also dovetail mortises but are smaller, nearly half the size of the one just described. The clamp cutting shown in figure B.1, C was cut into the 1

Th is dovetail mortise is one of two noted and drawn by Küpper (1996, figs 168f, 169l). 2 In 1998 this block fragment was among a large pile of similar fragments near the southeast edge of the site.

B.1 Dovetail clamp cuttings at Englianos.

371

Nelson – The Architecture of the Palace of Nestor the block.6 Many Minoan dovetail clamps are cut in the same way (Shaw 1973a, figs 94, 187, 190). The purpose of the skewed alignment is unclear; it may simply be the result of letting into the block a device that was not meant to be seen, so that fine workmanship could be overlooked. In the north retaining wall of the court before the megaron at Mycenae, it has long been known that one block of the sixth course of ashlar preserved a dovetail clamp cutting. In late 1996 or early 1997 this wall collapsed,7 and during a visit the following summer, I saw two other blocks with dovetail clamp cuttings among those fallen. One block with a clamp belonged to the ashlar course that rested on the horizontal wooden beam above the first course of blocks. For this wall and the encircling and blocking walls of the tombs, tie-back clamps were used at different levels and not restricted to a single course. The dispersal of clamps among different wall courses insured that the entire face was effectively retained.

B.2 Ashlar block with dovetail clamp cutting in the Southwest Quadrant.

described above, is neither in the centre of the block’s rear edge nor perpendicular to that edge, but set askew. The last dovetail clamp cutting (fig. B.1, B) was sunk into the most fragmentary of the three blocks. This stone does not appear to be a typical wall block, for its left face was cut perpendicular to the front, suggesting a return configuration in its original position. Perhaps the block was once placed at a wall return or at a window or door opening. The thick portion of the tail measures 0.05 m, and the cutting is sunk 0.04 m into the top surface of the block. As appears to have been standard practice, its direction is not perpendicular to the rear face of the block, but slightly askew. The block also has a rectangular mortise in its upper surface, located between the front face of the block and the dovetail cutting. A similar combination of dovetail cutting and rectangular mortise was cut into an ashlar block from the Granary at Mycenae (Küpper 1996, fig. 168a). The block there is not in situ, but it appears to have been a corner block, since its front and left face were finished smooth and cut perpendicular to one another. The purpose of the mortise and dovetail combination is unclear, since the block is not in its original position and is now reused in the matrix of a rubble wall. The dovetail clamp may have tied together the two faces of the wall to which the block belonged, and the mortise probably served to secure a horizontal beam running along the block’s front face.3

3

More measurements need to be taken on this block; MARWP was not allowed to move the block, so neither its height nor thickness could be examined accurately. The sunken lip or edge cut into the block’s back face (shown dotted in fig. B1, B) is confusing. The combination of this sunken lip and the dovetail clamp cutting matches a similar pair on a geison block of the Archaic temple at Mycenae (Klein 1997, 282–84 and fig. 15). Since post-Mycenaean activity is known to have occurred at Englianos and is documented with ceramic and roof tile fragments (see Part I, pp.245–53 above), it is possible that this block belongs to the later occupation of the hilltop. 4 None of the clamps survive to demonstrate how they actually functioned. 5 At present, some blocks from the blocking wall of the Treasury of Atreus lie in rows in front of the dromos. Some from the Tomb of Clytemenestra are still in situ, while others lie nearby. 6 Küpper (1996, fig. 169) does not show the front edge of the blocks in his compilation drawing of clamp cuttings. 7 It has since been reconstructed in a way not consistent with its original appearance.

Clamps imply a retaining or securing function, and since the cuttings are at the back of the block, their clamps probably secured the ashlars to an inner rubble core. Dovetail clamps were used in some ashlar walls at Mycenae, and in every instance there the clamps seem to have secured the outer ashlar face with the rubble or earth immediately behind.4 In this way, the ashlar face acted as a retaining wall, and the clamps served as tiebacks (Küpper 1996, 55–56), presumably fairly long ones, so as to penetrate some distance and effectively retain the wall. Many examples come from both the ashlar-faced encircling walls and the blocking walls of the Treasury of the Atreus and the Tomb of Clytemnestra.5 All of the clamp cuttings are similar to those at Englianos, and many are cut slightly askew rather than perpendicular to the face of

372

appendix c ANALYSIS OF EARTH MATERIAL

In 1991 and 1992, under the direct supervision of the Greek Archaeological Service, MARWP collected earth material samples from the archaeological remains at Englianos. These were tested for 27 elements by inductively coupled plasma–atomic emission spectrometry (ICP–AES) at the Soil Testing and Research Analytical Laboratories in the Department of Soil Sciences at the University of Minnesota.

6. Total the volumes in step 5, and then subtract this volume from 30. This gives the volume of 0.76 M extracting solution to be added to the above soil solution. The figure 30 is the total mls of 0.76 M extracting solution that is sought to be added to the 3 g of sample in order to keep the soil/solution ratio close to 10X. The final dilution factor will vary some from 10X. 7. Cover the sample bottles and place on shaker for 30 minutes at a shaking rate of about 180 epm.

The method for extraction of multi-elements from the earth material was as follows.1 1. Weigh 3.00 g sample (air dry) into 180 ml diSPo brand plastic bottles.

8. Filter through Whatman 30 filter paper, discarding the first 2–3 ml of extract that may contain contaminates from the fi lter paper.

2. Add 3 ml of 0.76 molar (M) acid extracting solution (5% HCl and 1% HNO3)2 to determine which samples contain carbonates.

9. Analyse by inductively coupled plasma atomic emission spectrometry (ICP-AES) for 27 elements. Applied Research Laboratories model 3560 simultaneous reading ICP.

3. To the samples that contain carbonates as evidenced by the reaction in step two, add 1 ml of 6 M hydrochloric acid (HCl). Swirl samples container and allow the carbonates to react for about 5 minutes or until effervescence ceases. 4. Test the soil solution in step 3 with pH indicator paper, and if the pH is less than 1.5, it is sufficiently neutralised. If the pH is above 1, add an additional 1 ml of 6 M HCl and allow to react as in step 3 and retest the pH. Continue this sequence until all samples have been neutralised to below approximately pH 1.5 using the pH paper. It is recommended that for samples that test around pH 2 while going through the sequence have the 6 M HCl added in a 0.5 ml increment. It is essential that the volume of acid added in the neutralisation process be recorded.

1

Special Job No. 124. Data Report: 9 March 1992. Compiled and written by Bob Munter, Soil Testing and Research Analytical Laboratories, University of Minnesota. 2 The acid mixture and strength was selected primarily because the laboratory has used this mix on soils analysed for John Foss at the University of Tennessee on archeological samples. The mix was originally selected for convenience, in that this mix is routinely used in this laboratory for making up calibration standards for water analysis according to the EPA method for ICP analysis of water samples. The acid mix and strength is not of importance between studies unless that data is to be compared, but must be kept the same within studies for accurate data comparison.

5. Total the volume of acid required to neutralise the sample. Then add/record the amount of 6 M HCl required to make this volume of neutral soil/salt solution approximately 0.76 M in acidity. 373

Nelson – The Architecture of the Palace of Nestor Sample No.

Batch No.

Collected

Location

12

74

1991

Matrix from northwest end of wall separating rooms 97 and 98

41

74

1991

Matrix from interstice in wall 8L at Room 4

41 (duplicate)

74

1991

Matrix from interstice in wall 8L at Room 4

49 4

74 124

1991 1992

Matrix from pier in wall 4L near its northwest end at Room 6 Black earth from floor near northeast wall of Room 99

4 (duplicate)

124

1992

Black earth from floor near northeast wall of Room 99

20

124

1992

Soil beneath fallen ashlar block, northeast of northeast facade of Main Building

46

124

1992

Matrix from interstice in wall 5A

50

124

1992

Matrix from door jamb in wall 5B.2

51

124

1992

Matrix from rubble wall at northwest end of Room 18

53

124

1992

Matrix from interstice in wall 2H (between rooms 9 and 10)

55

124

1992

Matrix from interstice in wall 2H (between rooms 9 and 10)

59

124

1992

Earth fi lling of pithos in Room 24

60

124

1992

Matrix from wall 2B (between rooms 20 and 21)

61

124

1992

Earth in wall 17L, section 1

64

124

1992

Earth from Great Drain southeast of Room 60

67

124

1992

Matrix from wall of Northeast Gateway

67 (duplicate)

124

1992

Matrix from wall of Northeast Gateway

C.1 Earth material samples.

Sample3

Approximate % CaCO3 Equivalent4

Building System

4

30–35

n/a

12

37–42

rubble

20

35–40

n/a

41 46

70–75 40–45

pier-wall pier-wall

49

40–45

pier-wall

50

55–60

pier-wall

51

25–30

rubble

53

35–40

pier-wall

55

30–35

pier-wall

59

30–35

n/a

60

35–40

rubble

61

25–30

limestone-slab

64

25–30

n/a

67

30–35

rubble (NE Gate)

C.2 Approximate calcium carbonate content of earth material samples.

374

Analysis of Earth Material Sample

AL2 Aluminum

AS Arsenic

B Boron

BA Barium

BE Beryllium

CA Calcium

CD Cadmium

12

3481.320

11.902

6.535

55.559

0.344

130932.00

1.028

41

1314. 167

8.202

5.789

24.020

0.179

113366.00

0.706

41 (dup)

1329.829

8.315

5.828

23.982

0.181

113931.40

0.702

49

4337.190

11.993

6.373

71.213

0.337

126421.90

1.011

Wau

1594.303

3.076

1.844

130.422

0.342

3609.07

0.503

Sample

CO Cobalt

12

4.681

41

CR Chromium

CU Copper

FE2 Iron

K Potassium

LI Lithium

MG Magnesium

8.839

32.312

1551.000

590.304

1.193

1125.228

3.067

5.222

25.441

669.115

654.843

0.662

893.607

41 (dup)

3.094

5.236

27.222

682.631

654.724

0.670

898.753

49

3.264

11.717

13.913

1426.260

573.265

0.608

858.759

Wau

1.743

1.650

- 6.005

1257.977

326.274