Method in Microwear Analysis: Prehistoric Sickles and Other Stone Tools from Arjoune, Syria 9780860545613, 9781407347189

Citation preview

Method in Microwear Analysis Prehistoric Sickles and Other Stone Tools fro� Arjoune, Syria

Roniana Unger-Haniilton

BAR International Series 435

1988

B.A.R.

5, Centremead, Osney Mead, Oxford OX2 ODQ, England.

GENERAL EDITORS A.R. Hands, B.Sc ., M.A., D.Phil. D.R. Walker, M.A.

BAR S435,1988 :Xethod in Xicr owear Analysis © Romana Unger-Hamilton

The author’s moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN 9780860545613 paperback ISBN 9781407347189 e-book DOI https://doi.org/10.30861/9780860545613 A catalogue record for this book is available from the British Library This book is available at www.barpublishing.com

C ontents

P age L ist o f T ables

5

L ist o f P lates

6

L ist o f F igures

1 6

P reface

2 6

P art I .

T heory a nd M ethod o f M icrowear A nalysis

I ntroduction C hapter 1 C hapter 2 C hapter 3 C hapter 4

-

-

-

-

E xperimentation

2 9

M icroscopy a nd P hotography

3 1

F eatures O bserved i n aM icrowear A nalysis

3 3

E dge D amage

3 4

( 1. O utline o f R esearch, 2 . E xperimental R esults, 3 . V ariables, 4 . E dge D amage d ue t o O ther F actors t han U se, 5 . C onclusions) C hapter 5

C hapter 6

C hapter 7

-

-

-

P olish, S triations a nd G loss O utline o f R esearch P olish, S triations a nd G loss A ims o f t he I nvestigation

4 3

4 8

-

T he M echanism o f P olish F ormation

5 0

( 1. R ecent T heories, 2 . E xperimental I nvestigat ion u nder t he L ight M icroscope, 3 . I nvestigation u nder t he S EM I 4 . C onclusions) C hapter 8 C hapter 9

-

-

A M odel o f P olish F ormation

5 6

S triations

5 7

( 1. F ormation, 2 . T he M icrowear A nalysis)

V alue

o f

S triations

t o

C hapter 1 0

M acroscopic G loss

-

( 1. F ormation, A nalysis) C hapter 1 1

-

C hapter 1 2

-

C hapter 1 3

-

C hapter 1 4

-

2 .

6 0 T he V alue o f G loss t o N icrowear

R esidues

6 2

C leaning

6 4

T he I nfluence o f S oil C hemistry o n P olishes

6 7

E xperimental P olishes

6 8

( 1. M ethod, 2 . W ood, 3 . B one, 4 . H ide, 5 . A ntler, 6 . H orn, 7 . M eat, 8 . F ish, 9 . S hell, 1 0. S tone, 1 1. P ottery, 1 2. S and, 1 3. F eather, 1 4. W ool, 1 5. " Ivory", 1 6. C opper, 1 7. P lant, 1 8. H afting a nd P rehension T races, 1 9. C onclusions) C hapter 1 5

-

T he V ariables A ffecting t he D evelopment a nd t he A ppearance o f P olish O utline

8 2

T he P reliminary S eries o f E xperiments w ith P lants

8 3

-

C hapter 1 6

-

( 1. M ethod, C hapter 1 7

2 .

R esults,

3 .

C onclusions)

V ariables o f P olish F ormation F urther E xperiments

-

-

8 7

( 1. F lint, 2 . D uration o f W ork, 3 . T ool A ction, 4 . T ool S hape, 5 . P ressure, 6 . T ool A ngle, 7 . M oisture C ontent, 8 . S pecies, 9 . A brasives, 1 0. C onclusions) C hapter 1 8

P olish a nd S triations d ue t o F actors O ther t han U se

-

9 6

( 1. M anufacture, 2 . N atural A gencies, 3 . P olish a nd S triations f rom P ost-excavational P rocedures) C hapter 1 9

-

C hapter 2 0

-

C onclusions

-

S ampling

1 01

T he B lind T ests

1 03

T heory a nd M ethod o f M icrowear A nalysis

1 07

S ickle B lades a nd O ther T ools f rom A rjoune

P art I I. I ntroduction C hapter 1 C hapter 2

-

-

A rjoune a nd i ts E nvironment T he E xcavations

C hapter 3

-

C hapter 4 C hapter 5

-

-

C hapter 6 C hapter 7 C hapter 8 C hapter 9

-

1 12

T he F aunal A nalysis

1 13

N on-Flint A rtefacts

1 14

T he F lint

1 15

T he R aw M aterial

-

-

-

C hapter 1 0

T he B otanical E vidence

-

( Flint)

1 17

P reservation o f t he E xcavated F lint

1 19

T he M ethod

1 20

S crapers

1 22

( 1. P revious R esearch , 2 . E xperiments, 3 . R esults, 4 . S crapers f rom A rjoune, 5 . C onclus ions) C hapter 1 1

P erforators

-

( Borers,

P iercers,

D rills)

( 1.Previous R esearch, 2 . E xperiments, R esults, 4 . B orers, P iercers, D rills A rjoune, 5 . C onclusions) C hapter 1 2

R etouched B lades,

-

C hapter 1 2 A

-

1 32 3 . f rom

B ladelets a nd F lakes

1 39

B lades, B ladelets a nd F lakes w ith L ateral R etouch

1 39

( 1. P revious R esearch, 2 . E xperiments, 3 . R esults, 4 . B lades, B ladelets a nd F lakes w ith L ateral R etouch f rom A rjoune, 5 . C onclusions) C hapter 1 2 B

-

S napped B lades

1 48

( 1. P revious R esearch, 2 . E xperiments a nd R esults, 3 . S napped B lades f rom A rjoune, 4 . C onclusions) C hapter 1 3

N otches a nd D enticulates

-

C hapter 1 3 A

-

-

I ntroduction

N otches

1 51

( 1. E xperiments, 2 . R esults, A rjoune, 4 . C onclusions) C hapter 1 3 B

-

-

3 .

N otches

f rom

D enticulates ( 1. E xperiments, R esults, A rjoune, 4 . C onclusions)

C hapter 1 4

1 50

B urins

1 54 3 .

D enticulates

f rom

1 57

( 1. P revious R esearch, 2 . E xperiments, 3 . R esults, 4 . B urins f rom A rjoune, 5 . C onclus ions) C hapter 1 5

-

A rrowheads a nd B arbs

1 61

( 1. P revious R esearch, 2 . E xperiments, 3 . R esults, 4 . A rrowheads f rom A rjoune, 5 . C onclus ions) C hapter 1 6

-

A xes,

A dzes a nd C hoppers

( 1. P revious R esearch, 2 . E xperiments, R esults, 4 . A xes, A dzes a nd C hoppers A rjoune, 5 . C onclusions) C hapter 1 7

-

1 65 3 . f rom

B lades w ith G loss f rom A rjoune a nd O ther S ites i n t he L evant

1 68

( 1. I ntroduction, 2 . P revious R esearch, 3 . P lants, 4 . E xperiments, 5 . B lades w ith G loss f rom A rjoune, 6 . P olishes o n N atufian B lades w ith G loss, 7 . P olishes o n B lades w ith G loss f rom J ericho P PNA t o E B, 8 . C onclusions) C hapter 1 8

B urnt F lint I mplements f rom A rjoune

2 06

C omparison o f T ypology a nd N icrowear A nalysis

2 07

C onclusions

l .A

T he F lint T ools a t A rjoune

2 08

C onclusions

1 .B

T he S ite o f A rjoune

2 10

C onclusions

2 T he B enefits o f H igh-Power M icrowear A nalysis

C hapter 1 9

-

-

-

-

-

B ibliography A ppendix

-

2 11

2 14 L ist o f E xperiments

2 31

L ist o f T ables T able 1 2 3 4 5 6

7

8 9 1 0

-

-

-

-

-

P age E dge D amage

2 44

P lant P olishes

2 45

B lind T ests 1 -10

2 46

B lind T ests 1 1-20

2 46

B lind T ests 2 1-30

2 47

E xperimental H arvest a nd W eeds

-

-

-

-

-

P roportions o f C ereals,

-

G rass 2 47

L ithic A rtefacts f rom A rjoune, M ounds B a nd C ( Surface)

T renches 1 -1V ,

L ithic A rtefacts f rom A rjoune,

T rench V

2 49

L ithic A rtefacts f rom A rjoune,

T rench V I

2 50

L ithic A rtefacts f rom A rjoune,

T rench V II

2 51

5

2 48

L ist o f P lates ( When n ot i ndicated, 2 00x.)

t he

o riginal

m agnifications

a re

a t

P age P late 1

-

E dge D amage ( cutting a nd s craping, e xperiments, 5 0x )

a ll

2 52

a -Stone 1 7 b -Stone 1 8

P late 2

-

c -Bone 1 ,

d ried

d -Bone 5 ,

d ried

e -Plant 2 6,

r eed

f -Plant 2 3,

r eed

g -Antler 1 8,

d ried

h -Antler 2 0,

d ried

E dge D amage ( cutting a nd s craping, e xperiments, 5 0x ) a -Antler 1 4,

s oaked

b -Antler 1 3,

s oaked

c -Wood 2 ,

d ried

d -Wood 6 ,

d ried

a ll

2 53

e -Hide 1 6 f -Hide 5 g -Meat 1 4

P late 3

-

R esidues

( SEN )

2 54

a -bone r esidues b -"phytolith"

( Anderson-Gerfaud,

( Anderson-Gerfaud,

c -experimental b one r esidues, a ccelerated v oltage 2 5V d -"phytolith"

1 981),

4 000x

5 000x

4 000x,

( Anderson-Gerfaud,

6

1 981),

1 981),

4 400x

e -lower p ower v iew o f 2 -c,

1 000x

f -object o n f lint b lade u sed o nly t o p olish o ther f lint, 4 400x

P late 4

R esidues

-

( SEM )

2 55

a -"cellular v egetable m atter" G erfaud, 1 981), 1 800x

( Anderson-

b -object o n f lint b lade u sed o nly t o p olish o ther f lint, 5 400x c -similar o bject o n u nused f lint b lade,

6 000x

d -similar o bject o n u nused f lint b lade,

5 400x

e -object p rotruding f rom t he s urface o f a n u nused f lint b lade, 6 000x

P late 5

P late 6

V ariables: F lint

-

-

(a ll e xperimental)

a -Wood 1 /1,

B randon,

u nused

b -Wood 1 /1,

B randon,

1 400 s .m.

c -Wood 1 /2,

S yr .C,

u nused

d -Wood 1 /2,

S yr.C,

1 400 s .m.

e -Wood 1 /4,

S yr.A,

u nused

f -Wood 1 /4,

S yr.A,

1 400 s .m.

g -Wood 1 /5,

S yr.B,

u nused

h -Wood 1 /5,

S yr.B,

1 400 s .m.

M anufacturing T races a nd B urning a -exp.

l imestone r etouch,

b -ARJ 7 02.3B, c -exp.

n otch,

l OOx

q uartzite r etouch,

d -ARJ 2 10.1,

f lake,

l OOx

l OOx

e -exp.

s napped b lade,

f -exp.

a ntler r etouch,

7

l OOx

2 00x 2 00x

2 56

2 57

g -exp.

P late 7

-

b urnt f lake,

l OOx

h -ARJ 2 21.10-232.4,

l OOx

N atural T races a -exp.

S and 1 ,

b -ARJ 8 03.3, c -exp.

-

b lade,

f lake,

+

2 00x w ater,

2 00x

f -ARJ 7 01.3A,

2 00x

g -exp.

u nused,

S yr.D,

l OOx

l OOx

S tone 1 /6,

h -JPF 1 12.4,

P late 8

2 00x

S tone 1 /8,

d -ARJ V . e -exp.

2 58

b lade,

l OOx

l OOx

S crapers

2 59

a -exp. W ood 1 /4 b -ARJ 1 008.3 c -exp.

W ood 1 /5

d -ARJ V I, e -exp.

S urface

W ood 1 5,

d ried s ycamore

f -ARJ 8 00.1 g -exp. W ood 1 6/1,

s ycamore

h -ARJ 2 20.1

P late 9

-

S crapers a -exp. W ood 3 7,

2 60 +

b ark

b -ARJ 1 14.2 c -exp.

S tone 2 5

d -ARJ M ound C e -exp.

S hell 5 ,

s awing

8

f -ARJ 7 03.3B

P late 1 0

-

g -exp.

A ntler 1 3,

h -exp.

B one 1 3

s oaked

S crapers a -exp.

2 61

B one 1 6

b -ARJ 1 000.1 c -exp.

B one 2 2,

c ooked

d -ARJ 1 005.3 e -exp. H ide 1 2,

+

o chre

f -ARJ 8 01.3 g -exp.

H ide 5

h -ARJ 5 00.1

P late 1 1

-

P erforators a -exp.

2 62

W ood 7 ,

b oring

b -ARJ 7 00.1 c -exp. A ntler 9 ,

b oring

d -ARJ 8 00.2 e -exp.

B one 4 2,

g rooving

f -ARJ 5 00.1 g -exp.

P ottery 6 /1,

d rilling

h -ARJ 1 000.1

P late 1 2

-

P erforators

2 63

a -exp. W ood 5 0, b -ARJ 7 01.3B, c -exp.

l OOx

l OOx

W ood 5 0,

d -ARJ 7 01.3B,

d rilling,

d rilling,

2 00x

9

2 00x

e -exp.

S tone 2 8,

b oring

f -ARJ 4 03.2 g -ARJ 7 00.1 h -ARJ 1 -1V ,

P late 1 3

-

B lades, a -exp.

S urface

B ladelets,

F lakes

2 64

W ood 2 1

b -ARJ 7 03 .3B c -exp.

W ood 3 5

d -ARJ 1 -1V , e -exp.

S urface

W ood 4 7,

w hittling

f -ARJ 2 01.2 g -ARJ 7 01.3A h -ARJ 7 00.1

P late 1 4

-

B lades,

B ladelets a nd F lakes

2 65

a -exp. M eat 3 b -ARJ,

B aulk 6 00-700.3

c -exp. M eat 4 ,

+

b one

d -ARJ 8 03.3 e -exp. W ool 1 f -ARJ,

B aulk 2 01-210

g -exp.

H ide 1 6

h -ARJ 7 00.1

P late 1 5

-

B lades, a -exp.

B ladelets a nd F lakes S tone 2 2

b -ARJ M ound C c -exp.

B one 2 0

1 0

2 66

d -ARJ 7 01.3B e -exp.

F ish 1 ,

s caling

f -exp.

F eather 1

+

g -exp. A ntler 1 8, h -exp. A ntler

P late 1 6

-

s oaked

3 /1 ,

d ried

B lades a nd S napped B lades a -exp. M eat 1 1/1, b -exp. M eat 9 ,

2 67

t endon

d ried s inew

c -exp. V egetable 1 ,

c arrot

d -exp. H orn 1 e -exp.

b roken b lade

f -ARJ 1 02.4,

b reak

g -exp. W ood 1 /5,

s craping

h -ARJ 9 00.1

P late 1 7

-

N otches

2 68

a -ARJ 1 03.10 b -ARJ 2 02.3 c -exp. W ood 4 0,

s having

d -ARJ ? ?5.2 e -exp.

B one 4 4,

s having

f -ARJ 8 01.3 g -exp.

S tone 2 1,

s craping

h -ARJ 1 011.2

P late 1 8

-

D enticulates a -exp. W ood 4 5,

2 69 s craping

b -ARJ 1 03.3

1 1

c -ARJ 2 13.3 d -ARJ 5 00.1 e -exp.

H ide 1 9,

s craping

f -exp.

S tone 1 6,

g rooving

g -ARJ 1 002.3 h -ARJ 1 01.1

P late 1 9

-

B urins a -exp.

2 70 B one 2 7,

g rooving

b -ARJ 7 00.1 c -exp.

S tone 2 ,

g rooving

d -ARJ 7 03.3B e -exp.

S hell 4 ,

i ncising

f -ARJ 7 01.3A g -exp. W ood 8 ,

g rooving

h -ARJ 2 12.3

P late 2 0

-

P rojectiles

2 71

a -exp. M eat 1 2, b -ARJ 7 01.3B,

f -ARJ 7 01.3B,

P late 2 1

-

A xes, a -exp.

l OOx

2 00x

2 00x

M eat 1 3,

h -ARJ I ,

b arb,

5 0x

e -exp. M eat 1 2,

g -exp.

5 0x

5 0x

c -exp. M eat 1 2, d -ARJ 7 00.2,

a rrowhead,

a rrowhead,

S urface,

2 00x

A dzes a nd C hoppers B one 4 3

1 2

2 00x

2 72

b -ARJ 1 -1V, c -exp.

S urface

W ood 2 6

d -ARJ 2 22.2 e -exp.

W ood 5 1

f -ARJ V I,

P late 2 2

P late 2 3

P late 2 4

S urface

g -ARJ V II,

S ounding 1

h -ARJ V II,

S ounding 4

E xperiments:

-

-

-

P lants

2 73

a -Plant 4 0,

w ild e inkorn,

b -Plant 4 0,

2 00x

c -Plant 3 6,

d om.

d -Plant 3 6,

2 00x

e -Plant 4 1,

d om.wheat,

f -Plant 4 1,

2 00x

g -Plant 4 3,

d om.

h -Plant 4 3,

2 00x

E xperiments:

b arley,

l øøx

l OOx

l OOx

e inkorn,

P lants

2 74

a -Plant 1 5,

b ulrush,

b -Plant 1 5,

2 00x

c -Plant 2 8,

r eed,

d -Plant 2 8,

2 00x

e -Plant 1 4,

d ried r eed,

f -Plant 1 4,

2 00x

g -Plant 2 3,

r eed s craping,

h -Plant 2 3,

2 00x

E xperiments:

l OOx

l OOx

P lants

1 3

l OOx

l OOx

l OOx

2 75

P late 2 5

-

a -Plant 4 5,

C yperus,

b -Plant 4 5,

2 00x

c -Plant 3 1/1,

c ane,

d -Plant 3 1/1,

2 00x

S parganium,

f -Plant 3 0,

2 00x

g -Plant 4 8,

p oppy,

h -Plant 4 8,

2 00x

E xperiments:

P lants

2 76

S tipa l OOx

b -Plant 4 6,

2 00x

c -Plant 2 ,

g rass,

d -Plant 2 ,

2 00x

l øøx

e -Plant 3 5/2,

h orsetail,

f -Plant 3 5/2,

2 00x

A rjoune: a -exp.

l OOx

l OOx

a -Plant 4 6,

h -Plant 4 4,

-

l OOx

e -Plant 3 0,

g -Plant 4 4, w eeds,

P late 2 6

l OOx

l OOx

l OOx

2 00x

P lants

P lant 4 1,

2 77 d om. w heat

b -ARJ 7 00.2 c -exp.

P lant 8 ,

d om.

d -ARJ 5 00.2 e -exp.

P lant 2 8,

r eed

f -ARJ 9 00.1 g -exp. W ood 3 5 h -ARJ 2 16.1

1 4

b arley

P late 2 7

P late 2 8

-

-

K ebara a nd E l W ad:

P lants

a -WB2,

5 0/4739,

N atufian,

b -WB1,

5 0/4740,

L ate N at.,

c -as a bove

( a ),

2 00x

d -as a bove

( b ),

2 00x

l OOx

N atufian,

l OOx

f -KB 5 8/2010,

N atufian,

l OOx

g -KB 5 8/2607,

N atufian,

2 00x

h -KB 5 8/1976,

N atufian,

2 00x

J ericho:

P lants

2 79

b -JPF P PNB X 8 .12A ,

d -JPF P PNB X 1 00.1,

f -JPF P PNB X 8 .12a,

h -JPF E B B 1 .5,

l OOx

l øøx

e -JPF P PNA P Pii 3 00.15,

g -JPF P N J 1 01.5,

l OOx

l øøx

c -JPF P PNA S Sii 3 00.27,

-

l OOx

e -KB 5 8/1995,

a -JPF P PNA S Siv 3 00.25,

P late 2 9

2 78

l OOx

l OOx

l OOx

l øøx

M iscellaneous

2 80

a -exp.

r ubbing r eeds,

1 0 m ins.

b -exp.

c utting r eeds,

1 0 m ins.

c -exp.

H ide 2 ,

s oaked l eather

d -exp.

F ish 1 ,

s caling,

e -exp.

C opper 1 ,

f -exp.

B one 4 5,

g -exp.

S tone 1 0,

h -exp.

I vory 1 ,

5 0x

d rilling c hicken +

a brasives

d rilling

1 5

L ist o f F igures

F igure 1 F igure 2

F igure 3

-

-

-

F igure 6 F igure 7

F igure 8

2 81

P olish I ntensity

-

F igure 4 F igure 5

M odel o f P olish F ormation

-

-

-

-

2 82-7

a -Plant,

5 0x

2 82

b -Plant,

l OOx

2 83

c -Plant,

2 00x

2 84

d -Wood,

5 0x

2 85

e -Wood,

l øøx

2 86

f -Wood,

2 00x

2 87

M ap o f S yria

2 88

T he R egion o f A rjoune

2 89

L ocation o f A rjoune

2 90

P lan o f t he S ite o f A rjoune

2 91

E xperimental S ickles

2 92-3

a -Rectilinear S ickle

2 92

b -Curved S ickle

2 93

T ool M easurements a nd T erminology

2 94

a -Measurement o f S ickle B lades b -Burin T erminology c -Measurement o f S craper E dge A ngle F igure 9 F igure 1 0

D istribution o f B urnt F lint:

-

-

T rench V I

S crapers a -exp.

2 95 2 96

W ood 1 /4,

b -exp. W ood 1 /5,

1 6

s ee s .

( s.1 P l.8:a P l.8:c

c -ARJ 1 008.3, d -exp.

s .

W ood 1 5,

e -ARJ 8 00.1,

P l.8:b s .

s .

P l.8:f

f -exp.Wood 1 6/1, F igure 1 1

-

2 97 S urface,

b -ARJ 2 20.1, c -exp. -

s .Pl.8:g

S crapers a -ARJ V I,

F igure 1 2

s .

c -exp.

n o p hotograph 2 98

W ood 3 7,

s .

s .

S tone 2 5,

e -ARJ 7 03.3B, f -exp.

s .

s .

S hell 5 ,

S .

B one 1 3,

s .

P l.9:d

P l.9:e s .

s .

P l.9:g

P l.9:h

S crapers

2 99

a -exp. B one 1 6, b -ARJ 1 000.1, c -exp.

e -exp.

B one 2 2,

s .

H ide 1 2,

h -ARJ 5 00.1,

s .

P l.10:a

P l.10:b s .

s .

H ide 5 ,

f -ARJ 8 01.3, g -exp.

s .

s .

d -ARJ 1 005.3,

-

P l.9:c

P l.9:f

9- e xp. A ntler 1 3, h -exp.

P l.9:a

P 1.9:b

d -ARJ V II H ound C ,

F igure 1 4

P l.8:d

P 1.8:h

W ood 2 0,

b -ARJ 1 14.2,

-

s .

S crapers a -exp.

F igure 1 3

P l.8:e

P l.10:c

P 1.10:d

s .

P l.10:g

P 1.10:f s .

P l.10:e

P 1.10:h

S crapers

3 00

1 7

a -ARJ 1 02.11 b -ARJ I ,

S urface

c -ARJ 1 04.2 d -ARJ 2 10.1 F igure 1 5

-

S crapers

3 01

a -ARJ 5 00.1 b -ARJ 7 02.3B c -ARJ 9 00.1 d -ARJ 1 000.1 F igure 1 6

-

P erforators a -exp.

3 02

W ood 7 ,

b -ARJ 7 00.1, c -exp.

s .

P l.11:b

A ntler 1 9,

e -ARJ 5 00.1,

s .

P l.11:f

B one 4 2,

s .

s .

P l.11:e

P 1.11:h

P ottery 6 /1,

i -ARJ 1 -1V ,

s . P l.11:g

S urface,

( L .C.),

s . P 1.12:h

P erforators

3 03

a -exp. W ood 5 0, b -ARJ 7 01.3B, c -exp.

-

P l.11:c

P l.11:d

h -exp.

F igure 1 8

s .

s .

g -ARJ 1 000.1,

-

P l.11:a

d -ARJ 8 00.2,

f -exp.

F igure 1 7

s .

s .

P 1.12:a,c

( L.C.),

S tone 2 8,

s .

d -ARJ 7 00.1,

( L.C.),

e -ARJ 4 03.2,

s .

s .

P 1.12:e s .

P 1.12:g

P 1.12:f

P erforators a -ARJ I ,

P l.12:b,d

3 04

S urface,

b -ARJ 3 13.2

1 8

( L.C.)

c -ARJ 7 03.38 d -ARJ 9 00.1 e -ARJ 1 000.1 F igure 1 9

-

B lades,

B ladelets a nd F lakes

a -exp. W ood 2 1, b -ARJ 7 03.3B,

s .

s .

F igure 2 0

-

B lades,

-

s .

P 1.13:c

S urface,

s .

a -exp. W ood 4 7,

s .

b -ARJ 2 01.2,

s .

P 1.13:f

c -ARJ 7 00.1,

s . P 1.13:h

B lades, a -exp.

s .

P l.13:g

B ladelets a nd F lakes M eat 3 ,

s .

c -exp. M eat 4 , d -ARJ 8 03.3, -

B lades, a -exp.

s .

s .

s .

P l.14:b

P 1.14:c

P 1.14:d

s .

P 1.14:e

b -ARJ B aulk 2 01-210, c -ARJ 7 00.1,

s .

s .

P . 1 4:f .

P l.14:h

d -exp.

H ide 1 6,

e -exp.

F ish 1 ,

f -exp.

F eather 1 ,

s .

P 1.15:f

g -exp. A ntler 1 8,

s .

P l.15:g

h -exp.

3 07

P 1.14:a

B ladelets a nd F lakes W ool 1 ,

3 06

P 1.13:e

b -ARJ B aulk 6 00-700.3,

F igure 2 2

P 1.13:d

B ladelets a nd F lakes

d -ARJ 7 01.3A, F igure 2 1

P 1.13:a

P 1.13:b

c -exp. W ood 3 5, d -ARJ 1 -1V ,

3 05

s . s .

A ntler 3 /1,

1 9

P 1.14:g P 1.15:e

s .

P 1.15:h

3 08

F igure 2 3

B lades,

-

a -exp.

B ladelets a nd F lakes S tone 2 2,

b -ARJ M ound C , c -exp.

F igure 2 4

B lades,

-

a -exp.

s .

B one 2 0,

d -ARJ 7 01.3B,

s .

P 1.15:a

P 1.15:b

s .

s .

P 1.15:c

P 1.15:d

B ladelets a nd F lakes M eat 1 1/1,

b -exp. M eat 9 ,

s .

s .

H orn 1 ,

e -ARJ 9 00.1, f -exp.

s .

s .

s .

h -exp. W ood 1 /5, F igure 2 5

B lades,

-

P 1.16:a

s .

P 1.16:c

P 1.16:d

P 1.16:h

b roken b lade,

g -ARJ 1 02.4,

s .

P 1.16:e

P 1.16:f s .

P l.16:g

B ladelets a nd F lakes

a -ARJ 1 -1V,

3 10

P 1.16:b

c -exp. V egetable 1 , d -exp.

3 09

S urface,

3 11

( L.C.)

b -ARJ B aulk 1 01-401 c -ARJ 1 02.2 d -ARJ 1 04.2 e -ARJ 1 04.3 f -ARJ 2 00.0 g -ARJ 2 02.2 h -ARJ 2 02.2 F igure 2 6

-

B lades,

B ladelets a nd F lakes

a -ARJ 2 11.2 b -ARJ B aulk 6 00-700.3 c -ARJ B aulk 6 00-700.3 d -ARJ B aulk 7 00-800.3

2 0

3 12

F igure 2 7

B lades,

-

B ladelets a nd F lakes

3 13

a -ARJ 7 02.3B b -ARJ 8 01.3 c -ARJ 8 01.3 d -ARJ 1 001.3,

( L.C.)

e -ARJ 1 002.3 f -ARJ 1 007.2 g -ARJ 1 007.3 h -ARJ 1 007.3 i -ARJ 1 03.3 F igure 2 8

N otches

-

3 14

a -ARJ 1 03.10, b -ARJ 2 02.3,

s .

P l.17:b

c -ARJ ? ?5.2,

s .

P l.17:d

d -exp. W ood 4 0,

s .

e -ARJ 8 01.3,

P l.17:f

f -exp. F igure 2 9

-

B one 4 4,

s .

S tone 2 1,

b -ARJ 1 011.2, c -ARJ 1 03.3,

-

s .

P 1.17:c

P l.17:e

N otches a nd D enticulates a -exp.

F igure 3 0

s . P 1.17:a

S.

s .

s .

P l.17:g

P l.17:h P 1.18:b

d -exp. W ood 4 5,

s .

e -ARJ 2 13.3,

P 1.18:c

s .

P 1.18:a

D enticulates a -ARJ 5 00.1,

3 15

3 16 s .

P 1.18:d

b -exp.

H ide 1 9,

s .

c -exp.

S tone 1 6,

2 1

S.

P l.18:e P 1.18:f

d -ARJ 1 01.1,

S .

e -ARJ 1 002.3, F igure 3 1

-

P l.18:h

( L.c.),

s .

P 1.18:g

N otches a nd D enticulates

3 17

a -ARJ B aulk 6 00-700.3 b -ARJ 7 01.3B c -ARJ 7 02.38 d -ARJ 1 000.1 e -ARJ 1 01.1 f -ARJ 7 00.1 g -ARJ 8 00.1 F igure 3 2

-

D enticulates

3 18

a -ARJ 8 04.3 b -ARJ 9 00.1 c -ARJ V II,

S ounding 4

d -ARJ 1 001.3 F igure 3 3

-

B urins a -exp.

3 19 B one 2 7,

b -ARJ 7 00.1,

s .

s .

P 1.19:b

P 1.19:d

d -exp.

S tone 2 ,

s .

P l.19:c

e -exp.

S hell 4 ,

s .

P 1.19:e

f -ARJ 7 01.3A, g -exp.

s . P l.19:f

W ood 8 ,

h -ARJ 2 12.3, -

P l.19:a

( L.C.),

c -ARJ 7 03.3B,

F igure 3 4

s .

s .

s .

P 1.19:h

B urins a -ARJ I ,

P 1.19:g

3 20 S urface

b -ARJ 7 03.3B c -ARJ 8 00.1

2 2

d -ARJ M ound C ,

S ounding

e -ARJ 1 001.2 F igure 3 5

-

P rojectiles a -exp.

3 21

M eat 1 2,

b -ARJ 7 01.3B,

g -ARJ I , F igure 3 6

-

A xes, a -exp.

A xes,

-

B one 4 3,

s .

P 1.20:h

s .

S urface, s .

3 22

P l.21:a s .

P l.21:b

P l.21:c

s .

b -ARJ 2 22.2,

P l.21:d

s .

3 23

P l.21:e

S ounding 1 ,

s .

P 1.21:g

A d es a nd C hoppers S urface,

b -ARJ V II, -

P l.20:g

a -exp. W ood 5 1,

A xes,

P 1.20:d

A dzes a nd C hoppers

a -ARJ V I,

F igure 3 9

s .

A dzes a nd C hoppers

c -ARJ V II, F igure 3 8

P l.20:b,f

( L.C.)

c -exp. W ood 2 6, -

s .

P l.20:c

s .

S urface,

b -ARJ 1 -1V ,

F igure 3 7

s .

( L.C.),

e -exp. M eat 1 3, f -ARJ 8 03.3,

P l.20:a,e

( L.C.),

c -exp. M eat 1 2, d -ARJ 7 00.2,

s .

( L .C.),

S ounding 4 ,

E xperiments:

s .

P lants

a -Plant 4 0,

S .

P 1.22:a,b

b -Plant 4 3,

s .

P l.22:g,h

c -Plant 1 5,

s .

P l.23:a,b

d -Plant 1 4,

s .

P l.23:c,d

e -Plant 2 3,

s .Pl.23:g,h

2 3

3 24 s .

P 1.21:f

P 1.21:h 3 25

f -Plant 2 ,

s .

g -Plant 4 5,

P l.25:c,d

s .

h -Plant 3 1/1,

P l.24:a,b s .

iPlant 4 8,

s .

P l.24:g ,h

j -Plant 4 4,

s .

P l.25:g,h

k -Plant 3 0,

s .

P l.24:e,f

1 -Plant 4 6,

s .

P l.25:a,b

r n -Plant 3 5/2, F igure 4 0

-

A rjoune: a -exp.

c -exp.

s .

s .

s .

P l.26:a,

P 1.26:c

s .

P 1.26:e,

2 3:c,d

P 1.26:f

g -exp. W ood 3 5,

s .

h -ARJ 2 16.1,

P 1.26:h

s .

2 2:e,f

P 1.26:d

P lant 2 8, s .

s .

P 1.26:b

P lant 8 ,

f -ARJ 9 00.1,

-

3 26

P lant 4 1,

d -ARJ 5 00.2, e -exp.

s .Pl.25:e,f

P lants

b -ARJ 7 00.2,

F igure 4 1

P 1.24:c,d

P 1.26:g

A rjoune B lades

3 27

a -ARJ 1 04.3 b -ARJ 1 12.2 c -ARJ 1 15.2 d -ARJ 1 15.2 e -ARJ B aulk 2 01-220 f -ARJ 4 02.3 g -ARJ 7 05.3B h -ARJ 8 03.3 i -ARJ 8 01.3 j -ARJ 9 00.1

2 4

F igure 4 2

F igure 4 3

K ebara a nd E l W ad:

-

s .

P l.27:a,c

b -WB1 5 0/4740,

s .

P l.27:b,d

c -KB 5 8/1995 ,

s .

P l.27:e

d -KB 5 8/2010,

s .

P l.27:f

e -KB 5 8/2607,

s . P l.27:g

f -KB 5 8/1976,

s .

P l.27:h

P lants

3 29

a -JPF P PNA S Siv 3 00.25,

s .

P l.28:a

b -JPF P PNA S Sii 3 00.27,

s .

P l.28:c

c -JPF P PNA P P±i 3 00.15,

s .

P l.28:e

d -JPF P PNB X 8 .12A, F igure 4 4

-

J ericho:

s .

3 30

s . P l.28:d

b -JPF P PNB X 8 .12 A ,

s .

c -JPF P N J 1 01.5,

P l.28:g

d -JPF E B D 1 .5, -

P l.28:b

P lants

a -JPF P PNB X 1 00.1,

F igure 4 5

3 28

a -WB2 5 0/4739,

J ericho:

-

P lants

s .

s .

P l.28:f

P l.28:h

M iscellaneous

3 31

a -exp.

r ubbing r eed,

s .

P 1.29:a

b -exp.

c utting r eed,

s .

P l.29:b

c -exp.

H ide 2 ,

d -exp.

I vory 1 ,

e -exp.

F ish 1 ,

f -exp.

C opper 1 ,

s .

P l.29:e

g -exp.

S tone 1 0,

s .

P 1.29:g

h -exp.

B one 4 5,

s .

P 1.29:c

s . s .

p l.29:h P l.29:d

s . P l.29:f

2 5

P reface

T his w ork i s b ased o n a t hesis w ritten a s a r esult o f r esearch c arried o ut a t t he I nstitute o f A rchaeology , L ondon, b etween 1 981 a nd 1 984. T he r esearch w as f unded b y t he D epartment o f E ducation a nd S cie nce o f G reat B ritain. M y r esearch d ealt w ith e xperimental " high-power" m icrowear a nalysis. T his a pproach t o t he f unctional a nalysis o f s tone t ools i s c omparatively n ew a nd, i n m y o pinion, a g reat d eal o f r esearch s till n eeds t o b e d one. A lthough I a m f ully a ware o f t he i nnovative q uality a nd i mportance o f w ork b y o ther r esearchers, I h ave s ometimes f ound m yself i n d isagreement w ith t hem. I e xpect ( and h ope ) t o b e d isagreed w ith i n t urn. S ince t he c ompletion o f t his t hesis i n 1 984 I h ave c onsidered a n umber o f t he p roblems i t a ddresses i n g reater d etail. M uch o ther w ork h as a lso b een d one o n t his s ubject. F or o bvious r easons i t h as n ot b een p ossible t o t ake a ccount o f t hat w ork h ere. Iw ould l ike t o t hank t he f ollowing: D r. M ark H . N ewcomer f or h is e xpert g uidance a nd h is r eading, d iscussion a nd c orrection o f t he m anuscript; P eter P arr f or p roviding a ccess t o t he m aterial a nd m oral s upport; C hristopher B ergman f or h is g enerous h elp w ith t he m anufact ure o f t ools a nd e xperimentation; G ordon C . H illman f or h is p ractical h elp a nd a dvise o n t he b otanical a spects o f t his s ubject; J onathan M offett a nd P eter D uckworth f or t heir h elp w ith t he c omputer s ystem; C hris C layton f or h is h elp w ith s ome o f t he g eological a spects o f t his s ubject; P eter D orrell f or h is n ever-ending g ood s pirits. Iw ould l ike t o t hank t he f ollowing f or t heir h elp i n v arious w ays: P rof. O fer B ar-Yosef, D r. J ohn B ryant, M ichael C harles, S usan C olledge, L orraine C opeland, J oan C rowfoot P ayne, J ohn D umont, N ick B arton a nd o thers a t t he I nstitute f or Q uaternary R esearch, O xford, A fif D akermanji, A .S. G ale, D r. I an G raham , P hil H arding, A nna K ers ten, D r. M . K islev, C hristopher r ay , D r. R obert M iller, A rlene M iller-Rosen, N aomi N icolas, D r. P eter R eynolds a nd s taff a t B utser A ncient f arm, S imon R ichardson, D r. D on R obins, R ichard W hitcombe, C hristine W ilson, a nd t he N icrowear R esearch G roup a t t he I nstitute o f A rchaeology , L ondon. M ost o f a ll I a m g rateful t o m y s ons, F elix a nd F erdy , m y m other a nd J ohn H ope M ason f or t heir p atience, e ncouragement a nd s upport.

2 6

Part

I .

Theory and Method of Microwear Analysis

Introduction At the beginning o f my research i nto microwear analysis i n 1 980 I d id not think that I would have t o write a substantial chapter dealing with theory and method: t he principles and procedures s eemed to have been e stablished over t he previous t en t o f ifteen years. From the middle o f the nineteenth c entury research into functional analysis of s tone tools was carried out through ethnographic analogy or use-wear study with or without experimentation by observation with the naked eye or the magnifying glass.(For a detailed and comprehensive review o f research into f unctional analysis of stone t ools the reader i s advised to t urn to Vaughan ( 1981, pp. 1-76).) This research was revolutionized in 1 964 when the work o f S emenov, begun in the 1 930s, became available in an English t ranslation. S emenov carried out a comprehensive experimental program and used high power microscopy in order t o identify t ool function by comparing wear t races ( polish, striations and edge damage) on experimental t ools with those on prehistoric tools ( Semenov, 1 964) Keeley ( e. g 1 976, 1 980) developed this method by focussing on one a spect, t he polish, to the extent that i t seemed possible t o identify with considerable a ccuracy the material on which the ancient tools had been used. A blind t est carried out by Keeley and Newcomer ( 1977) seemed to demonstrate the viability o f K eeley's method. S imilarly Tringham et a l. ( 1974) carried out a s tudy, f ocissing on another a spect, the formation o f edge damage, and again the viability o f that approach was t ested by Odell and Odell- Vereecken ( 1980) with c laims o f some success. Many microwear aia1yses have been carried out in recent years, notably the Ph i ) theses o f Anderson-Gerfaud ( 1981) Vaughan ( 1981) and Moss ( 1983) all of whom concentrated on Keeley's"High Power Approach"in which polishes were observed. Anderson-Gerfaud ( Anderson, 1 980, Anderson-Gerfaud, 1 981, 1 982, 1 983) s eemed t o have solved the riddle o f the mechanism of polish f ormation and a lso s eemed t o have f ound a method ( by using s canning e lectron microscope ( SEM) techniques) for identifying t he exact worked material by means o f residues f rom worked materials t rapped in t he polish on the surfaces o f u sed stone t ools. Most recently researchers have begun t o i nvestigate t he possibilities o f mechanisation of " High Power"microwear analysis ( by c omputer digitization ( Grace e t a l. 1 986) .

,

,

‚

2 7

and by i nterferometry ( Dumont, confidence w ith which r n icrowear regarded.

that

1 982) This i llustrates analysis has c ome t o .

However, very soon after I b egan my t here were a considerable number o f

research problems

t he be

I f ound with e ach

of t he methods referred to above: t he"Low Power"i nethod, i .e. t he analysis o f t ool function by observation o f edge damage, the"High Power"method, i .e. t he a nalysis o f t ool f unction by observation o f polish and striations, and t he S EM method, i .e. t he analysis of tool function by observati on of residues following

on t ool surfaces under t he a publication which s eemed

SEM. t o

I a lso f elt t hat, contain evidence

opposed t o Anderson-Gerfaud's t heory o f polish f ormation ( Masson et a l. 1 981) I had to i nvestigate the mechanism of polish f ormation, a lthough such an investigation would be l imited by t he fact that I am not a chemist or geologist. ‚

,

As a result this a ccount o f my work begins with t he following t opics: the principles of experimentation, t he t echniques o f microscopy and photography, and t he different k inds of wear traces and their value to f unctional analysis. Particular emphasis i s placed on t o K eeley's method o f observation of polish and striations at h igh magnifications s ince this has s timulated many areas o f d iscussion; attention i s a lso paid to the t heory of polish f ormation and t o the v iability o f Anderson-Gerfaud's method i n i dentifying worked materials from residues, t o c leaning methods, and t he effect o f soil chemistry on polishes on ancient t ools. Also discussed a re the difficulties o f t ransferring c onclusions about experimental work onto archaeological samples. F inally, accounts of two s eries of blind t ests are given, one designed t o t est experimental r esults on the same l ine a s Keeley's and Newcomer's t est ( Keeley and Newcomer, 1 977) and the other to t est the e fficiency o f Keeley's method applied t o archaeological t ools.

i s

Part I provides conducted.

the

basis

2 8

on which my work

i n

Part

I I

Chapter

1

-

Experimentation

I t i s now generally a ccepted by microwear analysts ( e. g. K eeley, 1 980, pp. 5 -7, Anderson-Gerfaud, 1 981, Vol-I, p .6-12, Vaughan, 1 981, pp. 7 8-82, Moss, 1 983, p .54-56) t hat an experimental program must not only be carried out before the wear traces on prehistoric t ools can be identified, but a lso t hat such a program must be designed t o be".. . relevant a ) to t he ecological s ituation and other general c onditions o f the s ite and s ites f rom which the study materials originate, b ) t o the l ikely worked materials ( hide, bone, meat and so on) and c ) t o t he rock types from which t he a rchaeological implements are made"(Keeley, 1 980, p . 5 ). Publications o f earlier work in which inappropriate materials were u sed for experimentation have often been criticised: s ee for example Keeley's criticism o f a paper by Ahler ( 1971) ( Keeley, op. c it., pp.6-7). ,

Therefore, in order t o carry out the appropriate experimentation it i s necessary to i nvestigate t he environment o f t he present s ite o f Arjoune and carry out experiments there, using l ocal f lint and materials in a ccordance with the evidence o f f inds excavated f rom t he s ite. I t i s a lso necessary t o investigate t he past environmental conditions and resources, and the probable methods of resource exploitation and manufacture. The f irst task was easily achieved, i .e. the s ite was v isited, local f lint was collected and used experimentally on materials suggested by the archaeological evidence. The s econd t ask, investigating past conditions, proved more difficult: an analysis of t he environmental evidence indicating past conditions and resources i s s till in progress ( see P art I I, Chapters 4-5) The l ast t ask, t o" imitate" methods o f resource exploitation and methods of manufacture, was t he most d ifficult, i f not an impossible t ask: the only .

comprehensive ethnographic studies which could be used a s evidence are t hose available f or areas in Palestine ( Turkovski, 1 969) and in I ran ( Wulff, 1 966) I t i s questionable to what extent one can rely on ethnographic evidence from completely different environmental a reas a nd t raditions, such a s f rom Australian hunter-gatherers or American Indians, for an analysis concerning t he environments and ways o f l ife o f Neolithic people f rom Arjoune in Syria. Keeley ( op. c it. p. 6 ) in f act criticised a paper by Gould e t a l. ( 1971) who had examined wood-adzes used by Australian aborigines and suggested that European Mousterian s crapers had been used a s wood-adzes, on a ccount of their shape which was s imilar to that o f the Australian t ools .

.

Moss ( op. c it., p .55) pointed out that one of t he most problematic a spects o f experimentation i s the way in which s tone t ools a re used, e specially by inexperienced experimenters. In f act, opinion seemed t o be divided amongst researchers and I f ound i t difficult t o decide a s t o whether experiments should be c arried out mechanically w ith

2 9

t he variables held c onstant i n order t o i nvestigate t he process of wear t race formation ( see Tringham 1 974) or whether i t would be better to"imitate"tasks, a s f ar a s known. Keeley ( 1980, p .8) stated:"If we wish t o replicate wear traces f ound on prehistoric i mplements, t hen our implements must be used in a human, not a mechanical fashion."Vaughan ( 1981, pp. 9 5-96) a ttempted t o combine t he human and mechanical approach and s tated t hat".. . the controlled, l imited version o f the aboriginal t asks was conducted with the s ame materials and motions a s would be t he case i f the researcher were performing i t"for real"and not under"laboratory conditions". ,

As I could not s ee how Vaughan's c ombined approach would work out in practice I c arried out two s eries o f experiments the f irst a s mechanical as possible i n order t o investigate the theory of wear t race formation, the s econd, in what seemed the manner which was most natural and which would be c losest to possible prehistoric usage, such a s by using s ickle blades mounted in a wooden handle. I will describe such procedures fully i n connection with the archaeological tools ( see Part I I, Chapters 1 0-17) -

3 0

Chapter

2

-

Microscopy and

Photography

Vaughan ( 1981, PP. 1 4-15) discussed the early research into functions of s tone tools which was e ssentially conducted with t he naked eye or e lse with a magnifying glass or a hand l ens. I n h is work ( 1964) S emenov employed a binocular microscope with diagonally reflected external l ight which he used to view cracks, edge chipping and striations up to 1 80x magnifications, whereas he used a monocular microscope with perpendicular incident l ighting f or detailed examination o f very small areas at magnifications of 3 00x to 5 00x or more ( Semenov, 1 964, p . 2 2) S ince then microwear researchers have employed a variety o f l ight microscopes: Tringham et a l. ( 1974) and Odell ( e. g. Odell and Odell-Vereecken, 1 980) used a .

s tereomicroscope with generally lower magnifications to v iew primarily the edge damage. Keeley ( e. g. 1 980) who concentrated on polishes viewed at higher magnifications used two microscopes: a stereomicroscope ( Wild M 5 ) and a l ab microscope ( Wild M 2 0) with an incident l ight attachment ( ibid, p .12) which he found vastly superior t o the stereomicroscope ( ibid, p . 1 3) Vaughan ( op. c it. p .81) used only one metallurgical microscope ( Wild M 5 0) Moss ( 1983, pp.79-80) discussed the f act that researchers such a s Keeley, Anderson-Gerfaud, Vaughan and herself used s imilar microscopes the Wild M 2 0 or M 5 0, the Leitz Metallux 2 or the Olympus Vanox microscope, a ll with magnifications ranging from about 3 5x to about 3 60x ( in the case of the Olympus Vanox 5 0x to 4 00x), a ll with incident l ight and bright f ield i llumination and that therefore results were comparable. .

.

-

-

I used the Olympus Vanox microscope because I had a ccess t o one and found i t easy to use. However, because o f the small distance from the stage to the l ens, in cases where prehistoric tools were large, I used the Wild M 2 0. I did f ind that using different microscopes, even two microscopes o f the same model, a ltered my perception o f wear traces s lightly ut not to the extent that observations were not comparable. The reason f or this optical difference i s unclear to me. Anderson-Gerfaud ( Ander o , 1 980, Anderson-Gerfaud, 1 981, 1 982, 1 983), f ollowed by other researchers, e .g. Mansur-Franchomme ( 1983), studied the mechanism of polish formation and residues on used f lint surfaces under a s canning e lectron microscope ( models: S iemens and Cameca) mostly at an accelerated voltage o f kv 2 0 and a t magnifications o f 5 00x to 1 0,000x ( Anderson-Gerfaud, 1 981, Vol.1, p .94). I t oo carried out investigations under the s canning e lectron microscope ( model: JEOL) using an a ccelerated voltage of 2 5 V and s imilar magnifications t o Anderson-Gerfaud ( see Part I , Chapter 7 and Unger-Hamilton, 1 984) i n order to see whether I could detect residues f rom ‚

,

3 1

t he worked materials on t he f lint s urfaces and i n order t o investigate polish f ormation. I did not use SEM t echniques t o investigate surfaces o f archaeological t ools, a s t he experimental research suggested i t would be unprofitable and t ime consuming ( see Part I , Chapter 7 ) and a s some o f t he ancient t ools would have had t o be f ragmented in order t o be inserted into the SEM chamber. I photographed a ll tools with I lford FP4 f ilm. I f ound that photographs of t he s ame wear t races could l ook very different according to t he f ilm and printing processes used, and according to the angle o f the photographed surface t o the microscopic lens. I t herefore tried t o keep these factors constant.

3 2

Chapter

3

-

Features

Observed

in a Microwear Analysis

The f ollowing f eatures have in recent years been s tudied a s i ndicative o f t ool f unction, each type of f eature requiring a d ifferent range o f magnifications: 1 )

Edge

damage

on

t ools

was

s tudied

by

Tringham

et

( 1974), Odell ( e. g 1 978, 1 981, 1 983, Odell and Odell-Vereecken, 1 980) Roy ( 1982) and by other researchers. In this"Low Power"approach magnifications between l Ox t o 5 0x were usually used on a l ight microscope. ,

2 ) Polish and striations were f irst and foremost studied by S emenov ( 1964) then by K eeley ( 1976, 1 978, 1 980, 1 982, 1 983, K eeley and Newcomer, 1 977, Keeley in Cahen et al. 1 979, 1 980), Anderson-Gerfaud ( 1981, 1 982, 1 983), Vaughan ‚

( 1980, 1 981, P erles and Vaughan, 1 983) and by many other researchers. For this"High Power"approach magnifications o f 5 0x to 4 00x were usually used on a l ight microscope. 3 ) Residues seen under t he SEM were s tudied by Anderson-Gerfaud ( Anderson, 1 980, Anderson-Gerfaud, 1 981, 1 982, 1 983) and by Mansur-Franchomme ( 1983) For t his .

approach used.

magnifications

of

between

5 00x

and

1 0,000x were

I will discuss each of t hese f eatures and its value in functional analysis in the f ollowing chapters. The observations o f residues, using SEM t echniques, a re discussed together with the t heory o f polish formation, a s the viability o f the former i s l inked to the understanding of the latter.

3 3

Chapter 1 .

4

Outline

-

Edge Damage of

Research.

There has been a heated debate during the past t en years whether edge damage on a stone tool when viewed under relatively l ow magnifications of up t o 5 0x i s i ndicative of the action with which, and o f the material on which a stone implement had been used ( Tringham e t a l. 1 974; Odell and Odell-Vereecken, 1 980; Roy, 1 982), or whether i t i s an unreliable indicator and should only be considered a ".. . useful check on an interpretation a lready based on t he microwear t races..." ( Keeley, 1 980, p .83), that i s, t he polish and striations observed under higher magnifications of 5 0x to 4 00x ( Keeley, ibid; Vaughan, 1 981; Moss, 1 983) ‚

S everal researchers argued that edge damage a lone was a reliable indicator o f both t ool action and worked material. Tringham et a l. ( 1974) referred t o edge damage a s"microflaking", to be distinguished from " abrasion in the form o f striations and polish" ( ibid, p .171) They decided to concentrate on this f eature of microwear a s i t was faster and easier t o observe than polish and s triations, needing less magnification and l ess sophisticated equipment ( ibid, p .175). They used 1 05 experimental f lint t ools with various actions divided into longitudinal ( cutting, s awing), transverse ( scraping, shaving, planing) and c ircular ( boring) actions on various materials ( including antler, .

-

-

bone, wood, meat and skin) classified according to whether they were considered hard, medium or soft. Tringham et a l found that the location and the direction o f microflaking varied according to the a ction, and that t he scar morphology ( i.e. s ize, depth and shape o f the s car) varied according t o .

the relative hardness of the worked material. They a lso pointed out the many variables other than t ool a ction and worked material such a s f lint type, edge angle, edge shape, angle o f the tool to the worked material, which a ffect microflaking. They investigated microflaking due t o causes other t han use, such a s microflaking due t o intentional retouch, a ccident or natural causes. They c laimed to be able t o distinguish t his type o f f laking from f laking due to use, because of the randomness of l ocation and orientation o f the f ormer and, in the instance o f intentional retouch, because o f the s ize o f the s cars. They concluded that microflaking a lone, when viewed under magnifications of l Ox to 6 0x, would be evidence of the action with which, and the relative hardness o f the material on which, archaeological tools had been used. However, they also concluded that by this method a lone one could not go any further and specify the exact movement o f t he tool or the exact nature o f the worked material ( e.g. meat or skin). ‚

Odell and Odell-Vereecken ( 1980) followed Tringham e t a l. ' s basic method which they referred t o a s the"Low Power Approach". They conducted a blind t est c omparable to the blind t est o f the"High Power Approach"(ibid., p .88)

3 4

conducted by K eeley and Newcomer ( 1977) Odell-Vereecker i used 3 1 basalt implements i n a greater diversity of ( 8) a ctions t han Tringham e t a l. on a wide variety o f materials which had been subdivided into four categories : hard, medium-hard, medium-soft and soft. Odell analysed the implements using magnifications of l Ox t o 2 0x, but a lso up t o l OOx i n order t o view"problem areas". The Odells stated that they referred deliberately t o"edge damage" a s they f elt edge rounding was a s indicative of a ctivities such a s s craping a s was"flaking". Their observations agreed with those o f Tringham et a l. ( perhaps not surprisingly a s Odell was one o f the " alii"). The results of their blind t est s eemed encouraging despite the f act that t he s ample was rather small and such statistical evaluations a s a "69.4%"rate o f success seemed to me rather misleading. Odell a chieved a success rate o f 7 9% in distinguishing the worked a rea of t he tools, of 6 9.4% in distinguishing the a ction and o f 6 1.3% in distinguishing the relative hardness o f the worked material, a lthough only o f 3 8.7% in distinguishing t he precise worked material. The Odells too .

c laimed t hat edge damage due to other factors than use was distinct, although at the end of the s ame paper ( ibid., p .117) t hey admitted that Keeley and Newcomer ( op., c it., p .35) may have been right in saying that"utilisation damage.. . cannot usually be distinguished on retouched edges." Roy ( 1982) went further than Tringham et a l. and Odell and Odell-Vereecken, claiming that by viewing edge damage on unretouched blades and f lakes she could not only t ell the a ction but the precise worked material, by f itting the materials into more precise hardness categories. Roy too c laimed to be able to differentiate between edge damage due to use and edge damage due to other agencies. She only briefly mentioned the large number of variables involved in edge damage formation and stated, " Ii n 'est pas question de l es enumerer i ci" ( ibid, p .168) a s i f they had no particular relevance to her investigation. Keeley, Vaughan and Moss have separately argued that edge damage a lone i s not a particularly reliable indicator o f either tool a ction or worked material, and that polish and striations observed at E Ox to 4 00x were far more diagnostic of both: Keeley ( 1980) referred to"edge damage" ( ibid., pp.24-25) but, a s opposed to Odell and Odell-Vereecken, did not take edge rounding into account. He noted edge rounding a s a "microwear feature"of e . g. hide polish ( 1977, p .42). He classified individual edge damage s cars a ccording to s ize, shape and depth and counted them individually on e ach experimental t ool. Keeley's f indings differed considerably f rom those of Tringham et al. Odell and Odell-Vereecken and Roy, in that he could not see any particular pattern of scars according to action and no variability o f s car shape and s ize a ccording t o the worked material. He c laimed that other variables such a s edge angle a ffected microflaking a s much a s the a ction and the ‚

3 5

worked

material

( 1980,

p .83)

.

He

a lso

pointed

out,

and

i n

some cases demonstrated, that f laking due t o other causes than use could look indistinguishable f rom f laking due t o use ( ibid., pp.25-28). He concluded t hat t he observation o f edge damage proved a useful check on an interpretation based on the " High Power"investigation of polish and s triations ( ibid,

p .83).

Vaughan ( 1981, pp.83-86, pp.106-120) referred t o " microchipping"and, l ike Keeley, did not include edge rounding in his category of"Edge Damage". The results o f his attribute analysis o f microscarring on 2 49 tools, made of three different f lint types and viewed a t magnifications o f up t o 2 80x, showed considerable differences to t he results o f Tringham et a l. those of the Odells and those of Roy: scar types and s izes were f ar more varied in h is analysis and he found that sometimes even hard materials did not s car edges at a ll; ( however, t his happened with only 6 of his experimental tools ( ibid., p .114).) Vaughan a lso demonstrated the other variables a ffecting t he formation o f edge damage, notably the different f lint types, and referred to the many causes ( other than use) o f edge damage. He therefore concluded that microchipping a lone was not a precise indicator o f the action with which and the material on which a prehistoric tool had been used ( ibid., p .115). ‚

Moss ( 1983, pp.76-79) referred t o"edge damage" without t aking edge rounding into a ccount. She discussed t he differences between the f indings o f Odell a nd Odell-Vereecken and of Vaughan in detail and, in my opinion quite r ightly, pointed to the f ollowing differences a s possibly responsible: the different stone they used, t he different l evel o f expertise o f the experimenters which was demonstrated by Moss and Newcomer ( 1981) to cause different f requency of edge damage and the fact t hat Vaughan used much higher magnifications. Vaughan himself ( op. c it.) l isted different sample s ize a s a possible reason. Despite the fact that Moss attempted to relate these different results t o controllable f actors, she agreed with Vaughan and Keeley that"edge damage due t o use i s neither very diagnostic o f the worked material, nor i s i t always present"(Moss, op. c it., p .76). -

-

2 .

Experimental Results

In order to decide whether there i s a distinct variation in edge damage according t o a particular type o f action with which, and a particular material on which, a f lint implement had been used, I investigated under t he microscope at 5 0x the edges of 7 2 o f my own experimental tools. They had been knapped f rom three types of f lint: f ine-grained Brandon f lint, f ine-grained Syrian f lint and medium-grained f rost shattered S yrian f lint. I used t he tools on hard l imestone, dried bone, f resh reeds, dried antler, s oaked antler, f resh and s easoned wood, medium f resh hide

and

fresh meat

with

l ongitudinal

3 6

a ctions

( sawing

I

s licing) and w ith t ransversal a ctions ( scraping) Most implements were unretouched b lades or f lakes of varying shape and with varying edge angles, a lthough most s crapers had been retouched with a ham i nerstone. As opposed t o other researchers ( Tringham et a l. op. c it. ; Brink, 1 978) who observed the edge damage on the dorsal a spect o f t he s crapers, I concentrated on t he ventral ( the contact) a spect .

‚

a s a ) t here were no misleading intentional retouch s cars on this a spect and b ) I wanted to observe t he edge rounding which l ike Odell and Odell-Vereecken ( op. c it., p .90) I considered a s belonging t o " edge damage". -

-

I c lassified individual damage s cars according t o t heir t erminations i nto three basic types ( modifying t he"Ho Ho Classification" ( Hayden ed., 1 979, pp.133-135) which had distinguished between four types o f t erminations) f eather ( F) s car i s"a f lake ( scar) which t erminates edge with a minimal margin"(crabtree, 1 972, p .64); 2 )

: 1 ) A in an A s tep

( ST) scar i s"a f lake or f lake s car that t erminates abruptly in a r ight angle break"(Crabtree, 1 972, p . 9 3). I c ould not differentiate a t 5 0x between step and hinge s cars which have different terminations when viewed in cross-section ( Hayden ed., 1 979, p .134) and result f rom s lightly different processes of formation ( Lawrence, 1 979, p .117) Lawrence t oo ( ibid.) found that these scars were".. . similar in appearance and sometimes difficult t o distinguish". Tringham et a l. ( op. c it., p .188) did not distinguish between t hem. 3 ) Snap ( Sn) t erminations result f rom the edge being broken away .

through"bending initiations". This category incorporates labels such a s"break"and " half-moon breakage"(Hayden ed., op. c it.). Unlike Keeley ( 1980, p .24) I did not count scars individually but observed the predominance o f l arge or small scars, thereby observing a pattern rather t han t he individual unit. Table 1 and P lates 1 -2 show the results: a ll s craping tools ( regardless of the worked material) showed edge rounding, which was never seen on sawing or s licing tools. Cutting l imestone caused a lot o f edge damage with l arge scars, predominantly with step t erminations. ( However, when l imestone was scraped, the scraper edge became abraded rather than f laked.) The number of f eather and snap t erminations seemed t o depend on the angle of contact between the t ool and the worked material. Dried bone caused a lot of edge damage with l arge s cars, predominantly with step t erminations. The tools which showed no edge damage t ended t o be of the same medium-grained f rost shattered f lint f rom Syria. Fresh reed caused a lmost a s much and the same types and s izes of s cars a s dried bone, a lthough i t caused somehwat less damage than did bone t o s craping tools. Dried antler caused somewhat l ess damage than stone, reed or bone. Soaked antler caused smaller scars, f ewer s tep t erminations and a ltogether l ess damage than did dried antler. Wood ( both fresh and seasoned) caused less damage, smaller s cars with f ewer step t erminations t han t he harder materials l isted above. I d id not t est the d ifference

3 7

between fresh and seasoned wood. Medium fresh hide caused much smaller scars than the above materials, mainly with feather terminations, but also some tiny scars with step terminations. Meat caused the least damage and the smallest scars of the materials listed. The scars had mainly feather terminations, although a few very small step terminations could also be seen. I saw distinct variations in edge damage according to tool action: longitudinal actions did not result in any noticeable edge rounding but in considerable microflaking which was usually bifacial, although it could be almost unifacial when the tool was held at an acute angle to the worked material. Transverse actions alwayi resulted in considerable edge rounding, but in very little microflaking on the ventral (contact) aspect, as opposed to the dorsal aspect. My findings agreed with those of Odell and Odell-Vereecken (op. cit. p.90) who stated that edge rounding was diagnostic of scraping. My findings disagreed with those of Tringham et al. (op. cit., pp.188-189) and with those of the Odells (Odell and Odell-Vereecken, op. cit., pp.98-99, p.109) which had led them to assume that sawing actions almost always caused bifacial, while transversal actions almost always caused unifacial microflaking. My findings also accorded well with my earlier observations with the naked eye : two similarly shaped and angled flint edges developed a different shape and angle according to whether they were used to saw or scrape limestone. After one minute's use the sawing edge became finely retouched and remained acutely angled, while the scraping edge became invasively retouched and steeply angled, i.e. the sawing edge looked like that of a denticulated blade while the scraping edge looked like that of an intentionally retouched scraper. This experiment had indicated at a macroscopic level that tool action affected the character of the edge damage on that tool. I also saw distinct variations in edge damage according to the hardness of the worked naterial. Hard materials such as stone, bone or dry antler n arly always produced most microflaking, the largest scars and the most frequent step fractures. The soft materials , such as fresh hide and meat, produced the least microflaking , the smallest scars and only a few and minute step fractures. My findings disagreed with those of Keeley (1980, p.83) who reported that the size of scars and the occurrence of particular types of scars were indicative of other variables, e.g. edge angle, as much as of the worked material. My findings also disagreed with those of Tringham et al. (1974, p.191) who stated that soft materials could not produce stepped removals. I could not see any

38

variations i n edge damage which might indicate t he precise material a s was c laimed by Roy ( 1982) ; materials such a s wood o r antler occur naturally, or can be t reated in such a way t hat they can vary enormously in hardness. Roy ( ibid.) c lassified plants a s medium-soft, but my experimental t ools used on reeds which are surely hard materials, showed edge damage which c ould not be distinguished f rom that f rom bone. ‚

3 .

Variables

Just a s Vaughan ( 1981, p .114) and Moss ( 1983, p .76) I found t hat on some t ools edge damage was notably absent or unexpectedly different t o t hat on most of t he other tools used w ith the same actions on t he same materials. These exceptions could be related to the following variables: The edge s carring t he edge a ll.

angle o f the tool. The frequency and type o f varied according t o the a cuteness and steepness of angle. I found that steeper edges hardly s carred at

The edge shape. hardly damaged This

agreed

I found that absolutely straight edges were at a ll even when used on hard materials.

with Moss'

observations

( 1983,

The angle of the edge to the worked researchers including Tringham et e ffect o f this variable, the latter

p .76).

material. While a ll a l. agreed about the insisted that sawing

a lways produced bifacial retouch ( 1974, p .188) However, my blades when held at an angle o f 450 against l imestone during sawing became unifacially retouched, while blades with the same edge angles used with the same actions on the same material for t he same amount of t ime, but held a t an angle o f 900, developed bifacial retouch. .

The f lint shattered

type. Syrian

I found f lint was

that the invariably

medium-grained f rost less damaged than were

the other two f ine-grained f lints f rom Brandon and Syria. This i s in a ccord with the different fracturing properties of different cryptocrystalline stones referred to by other researchers ( Tringham, op. c it., p.178). Vaughan ( 1981, pp.107-108) in his experiments t ested such properties. However, he found that"in the majority o f cases ( 17 out o f 2 1) the f actor o f l ithic raw material was not found t o exert a s ignificant influence in causing the differences noted among s carring patterns on the three varieties o f f lint..."(ibid.). Recent experimental work showed t hat naturally recrystallised f lint ( chert) ( Bradley and C layton, 1 986) and thermally a ltered f lint ( Bradley and C layton, ibid.; S eitzer-Olaussen, 1 983) chipped more easily than unaltered f lint. The condition o f the worked material. More and larger s cars and more step f ractures occurred on the f lint edges used on dry antler ( Pl.1: g, h), than on those used on antler soaked f or three days ( Pl.2:a, b) Other r esearchers have noted t he .

3 9

s ame phenomenon p .102). Vaughan conditions o f the varying results.

( Odell and Odell-Vereecken, ( op. c it., p .114) referred to worked materials a s a possible

op. c it., t he varying reason f or

The duration of the action. The longer a t ool was used more edge damage formed. However it s eemed that s ize t ype of s carring were not notably a ffected ( Tringham et op. c it., p .191).

t he and a l.

Contact through pressure or percussion. These different types o f contact were discussed by Moss ( op. c it., pp.78-79) using f lint knapping by pressure f laking and percussion f laking a s a model. Moss stated that contact through pressure, such a s during drilling, c reated f ar smaller f lake s cars t han did contact through percussion, such a s when a projectile i s f ired. I found this t o be t he case with 1 5 experimental drilibits which showed very small scars i f any, but lots of f ine abrasion, while experimental projectiles showed the opposite: very l arge scars ( see Bergman and Newcomer, 1 983) and hardly any abrasion. I t seemed to be the general consensus of a ll researchers mentioned that the variables a ffecting the f ormation of edge damage are the f ollowing: t he action o f the t ool, the duration of this action, the hardness of t he worked material, the moisture content of t he worked material, t he s tone type of the tool, the edge angle, the edge shape, t he angle o f t he t ool to the worked material and the k ind o f contact, i .e. whether i t i s through pressure or percussion. These variables seem t o be the same a s those affecting t he formation and distribution o f polish ( see Part I , Chapter 1 5) a lthough differing in degree or extent; contact through pressure creates only small damage scars but it c reates a lot o f polish, whereas contact through percussion l arge damage s cars but very l ittle polish.

creates

‚

4 .

Edge Damage due

to Other

Factors

than Use

Again there seemed to be a consensus amongst researchers that there are many causes for edge damage other than u se. They only seemed to disagree about t he d istinctiveness of such damage. Researchers who used edge damage a lone in t heir microwear analyses c laimed t o be able t o distinguish in most c ases between edge damage due t o use and that due t o other f actors ( Tringham et a l. op. c it., ‚

p .181; Odell and Odell-Vereecken, op. c it., pp.96-97 and 1 19; Roy, op. c it., p .168) while those who relied mainly on polish and s triations c laimed they c ould o ften not distinguish between use and non-use edge damage ( Keeley, op. c it., p.83; Vaughan, op. c it., pp.116-120; Moss, op. cit., p .76) Amongst these causes of edge damage are"spontaneous retouch"during manufacture ( Newcomer, 1 976; Keeley, 1 980, pp.25-28) and a ccidental retouch due t o a f lint implement

4 0

f alling f rom a height. I n r elation t o t he l atter c ause I f ound t hat"single blow"tools such as notches and blade s egments could be c reated in this way. I c ould not d ifferentiate between t hese and intentionally shaped t ools, a lthough more i ntensive research suggested that i t might be possible t o distinguish between intentional and a ccidental b reaks on blades, even without a microscope ( Bergman 1 983) Accidental dropping o f a tool could a lso create edge d amage indistinguishable f rom that due to use. .

Handling and t ransport or f lint t ools could create c onsiderable edge damage a ccording t o Vaughan ( op. c it. p .86, referring t o Hayden and Kamminga, 1 973, p .4) Edge damage due to t rampling ( Miller, 1 982) a gricultural activities ( Betts, 1 977, pp.14-15) and natural a gencies such a s water-rolling could also be quite regular, a t l east when viewed with the naked eye, and therefore be c onfused with i ntentional retouch and use. My own experiment, leaving 2 0 f lint pieces with varying edge angles i nside a perforated plastic basket for f ive weeks in a f ast-running s tream, demonstrated the occurrence o f often r egular retouch, e specially where edge angles were a cute. All these potential problems are o f course very f amiliar to researchers who concentrate on the observations o f polish and s triations a lone, a s the problems are precisely the same. 5 .

Conclusions

Because the variables a ffecting edge damage and polish a re t he same, and s ince the agencies imitating edge damage o r polish due t o use are the same and s ince i t will be a rgued t hat polish i s largely due to abrasion of the f lint s urface ( see Part I , Chapter 7 ) i t seems t o me t hat edge damage in the f orm o f microflaking, edge rounding, and polish are all l argely due t o attrition o f the t ool and are t herefore interrelated and should be viewed t ogether. ( The p rocess o f microflaking i s o f course governed by the l aws o f f racture mechanics ( e.g. Odell, 1 981)). In fact I cannot understand why edge damage and polish were not usually s een t ogether by most researchers, a s the location of polish i s v isible with most types of microscope at magnifications of 5 0x and edge damage i s visible at magnifications o f 2 00x. Not only should edge damage and polish be l ooked at t ogether but a lso their quantitative r elation to each other s hould be examined: on t his particular point Moss' discussion of d ifferent e ffects of contact due t o pressure and percussion was very suggestive. Both she and I found that a t ool used w ith pressure would t end to have l ess edge damage but more a brasion ( i.e. polish) while a tool used with percussion would t end to have larger f lake scars, but very l ittle s igns o f abrasion ( i.e. polish) S imilarly, where edge damage has o ccurred the polished edge i s removed. Where this i s not r ecognised, the perception o f polish distribution and .

subsequent identification o f polish c an be c ompletely mistaken. Therefore I conclude that the relation o f edge damage t o polish i s important in i n icrowear analysis.

4 2

Chapter 5 Research

-

Polish,

Striations

and

Gloss

-

Outline

of

Microscopic wear-trace analysis o f f lint t ools was f irst used i n t he l ate 1 920s when Curwen ( 1930) i nvestigated " sickle gloss"on b lades under low magnifications of approximately 2x ( ibid. p 1.1-4) At about the same t ime S emenov began t o develop his experimental approach to wear-trace analysis using h igher magnifications o f 3 00x or more ( 1964, p .22). S emenov defined three different k inds of microwear:"(l) Polishing ( small specific pressures with dispersion o f minute particles and micro-plastic a lterations o f the surface), ( 2) grinding ( higher specific pressures with dispersion o f more substantial particles) and ( 3) rasping ( large s pecific pressures with macroscopic destruction o f t he surface)"(ibid., p .14). S emenov a lso pointed to striations a s a f orm o f microwear ( ibid., p .15) u sually a ssociated with contact with hard materials or the presence o f abrasives during work. He concentrated on the observations o f s triations a s evidence f or the"kinematics" o f t ool use ( ibid., p .17). .

,

Keeley s tated"To restrict oneself to the k inematic approach l eads only to a refinement in typology and not necessarily to any better information about ancient e conomics" ( 1974, p .328) Keeley developed his experimental method, a lso using high power magnifications of 5 0x t o 4 00x, i n which he concentrated not just on the observation of s triations a s indicators of t ool action ( and in s ome cases o f the worked material) but mainly on the polish which a ccording t o him ( 1976, p . 50 ; 1 977, p .37 ; 1 980, p .83) i s distinct according t o the exact material ( such a s bone, meat or antler) on which a f lint t ool had been used. Although he admitted a f ew convergences which may present s ome .

,

difficulties:"Use on meat may not a lways be distinguishable f rom use on fresh, wet hide; use on wood may not a lways be d istinguishable from use on soaked antler, unless t he microwear polish i s well developed; bone and antler s aws will sometimes be indistinguishable from one another."He c ontinued,"But t nese f ew problems ( my emphasis) should hardly e ffect the overall ability of the microwear analyst correctly and fully to i nterpret f lint implements in f unctional t erms, and there may be other k inds o f evidence t o help him choosE between a lternative i dentifications"(1980, p .83). Keeley's"High Power Approach" ( a t erm used by Odell and OdellVereecken ( 1980)) was put to the ( blind) t est by Keeley and Newcomer ( 1977) Keeley c laimed a success rate o f " approximately 1 0 out o f 1 6 correct"(ibid., p .60) in d istinguishing the precise worked materials. Holley and Del Bene ( 1981) a sked whether"Keeley's i dentifications o f worked materials were based on recognition o f his described " characteristic polishes"or had their basis in a suite o f .

other a ttributes of h is reply ( 1981) ,

wear"(ibid., p .346). Keeley insisted as he had earlier, that, " it was

4 3

in t he

character of t he polishes found on t he t est implements t hat provided the main basis for t he inferences concerning t he worked materials"(1977, p .61). In addition, Keeley thought depth o f striations c an vary material i tself ( 1980, p .23)

that the a ccording

width and t he t o the worked

Keeley s eemed to a ccept W itthoft's theory ( Witthoft, 1 967) t hat,"corn"gloss on an implement i s a sure s ign that the implement was used on grass..."(Keeley, 1 980, p .61). On t he other hand he reported similar wear-traces from bracken and bamboo ( ibid.). Microwear analysts who largely followed Keeley's method seemed to agree with his statement that most polishes a re distinct according to t he worked material ( Anderson-Gerfaud, 1 981, Vol. I , p .35; Vaughan, 1980, p .88 ( although he pointed t o a systematic overlap of polishes in Vaughan, 1 981, P .382); Moss, 1 983, p .80) and mostly with the characteristics of the polishes which Keeley described ( Keeley, 1 980, pp.35-61; Anderson-Gerfaud, 1 981, Vol. I , pp.44-63; Vaughan, 1 981, pp.127-179; Moss, 1 983, pp.83-105) a lthough in some instances variations were observed by different observers. Moss ( ibid.) d iscussed these and a lso noted that,"interpretation of polish type includes much more than an analysis of one small patch of polish." ( ibid, p .80). Researchers seem to vary extent striations indicate 1 981, p .121; Fedje, 1 979; Del Most x nicrowear analysts seem implements which i s v isible grasses or at l east p lants worked.

i n their opinion a s t o what t he worked material ( Vaughan, B ene, 1 979; Mansur, 1 981) to accept that gloss on f lint to the naked eye indicates that ( Vaughan, 1 981, p .156) were

In 1 980 Anderson-Gerfaud published her research using SEM t echniques. They revealed on the used edges o f f lint t ools objects she believed to b e residues from the worked materials which had become embedded in the f lint surface during use ( Anderson, 1 980, p .183; Anderson-Gerfaud, 1 981, Vol.1, P .100) S ince then microwear analysts have apparently accepted that polish forms a s a result not of attrition a lone ( as S emenov had believed ( 1964, p .15)), but a s a result of the formation o f a deposit of amorphous s ilica in which residues from t he worked materials become embedded ( Keeley, 1 980, p .43 and 1 981, p .349; Vaughan, 1 981, p .179, Moss, 1 983, pp.16-18). Moreover, this f inding was important evidence for of the d istinctiveness of polish a ccording t o worked material when viewed under t he l ight microscope. .

In addition Anderson-Gerfaud Anderson-Gerfaud, 1 981, 1 982, 1 983) ( Mansur-Franchomme, 1 983) attempted

4 4

( Anderson, 1 980; and another researcher to identify such

residues on out t he exact

f lint worked

s urfaces material

under the SEM on which the

i n o rder t o t ools had

f ind been

used. Despite their a cceptance of Anderson-Gerfaud's hypothesis, an implication o f which was that residues f rom worked materials a s well a s amorphous s ilica c ould be removed by chemical c leaning, most microwear researchers c leaned their experimental and archaeological t ools with chemicals such a s solutions o f sodium hydroxide ( NaOH) and hydrochloric acid ( HCl) ( Keeley, 1 980, p -11) However, Anderson-GerfaUd ( 1981, Vol-I, pp.28-29) Vaughan ( 1981, p .97) and Moss ( 1983, p .105 ) were cautious and to a certain extent restricted t heir use o f chemicals. .

‚

In s tated

a critical review of ‚ "lt i s a matter o f

t races that controlling

are not the for t he f actors

microwear in 1 974 Keeley had controlling against microscopic result of that a ffect

utilization and utilization traces,

such a s the raw materials of the artefacts . , the manner o f use and the material on which the a rtefacts were u sed. I f these controls a re not introduced, then s erious and confusing errors c an result."(1974, p .327). In the publication of his thesis Keeley discussed the e ffects o f abrasion f rom wind and water, patination, natural gloss, soil movements and frost contortions ( 1980, pp.28-35) He investigated such e ffects in some cases experimentally and found t hat in s ome instances they rendered the archaeological t ools unusable for microwear analysis by .

obliterating all t races c aused by use. He c laimed t o be able t o distinguish between t races due to such e ffects and t races due t o use:"Most natural processes, then, l eave traces which are unlikely t o cause much confusion f or the microwear analyst, once he is f amiliar with them and with t rue microwear f eatures..." ( ibid., pp.34-35). Other microwear analysts Vaughan,

s eemed not quite a s 1 981, pp.173-179; Moss,

confident a s Keeley 1 983, p .81-83).

( see

Polishes due to t echnological processes had a lso been experimentally demonstrated and discussed by Keeley ( 1980, pp.25-28) He concluded that"technological e ffects, examined .

with due care, demonstrate many appreciable true microwear..." ( ibid., p .28). This general acceptance ( Anderson-Gerfaud, 1981,

with Vol.

I ,

the p .40)

differences with s tatement f ound

other

researchers

Variables possibly i nvolved in polish formation, other than the worked material, include: different f lint types f rom which the t ools a re made; the tool action and duration of such action; freshness, dryness or hydration o f t he worked material; t he species o f the worked material; t he shape o f the t ool and the presence of external abrasives. S ome o f these variables have been investigated by various researchers: Keeley used s everal types of f lint including f ine-grained f lint from Brandon and coarse-grained f lint f rom Denmark, and s tated that"in the experiments, the t ype ‚

4 5

o f f lint i nvolved had no e ffect on the formation o r appearance of any of the microwear f eatures."(ibid., p .16). In an earlier publication, however, he did r efer t o t he distribution o f wood polish d iffering a ccording to t he texture o f t he f lint surface ( 1977, p .39). Vaughan, who used three f lint types in h is experiments, s aw a "quantitative", not a "qualitative"difference i n polish f ormation a ccording to the grain s ize of the f lint ( 1980, p .90 and 1 981, pp.129-130 and p .184). Anderson-Gerfaud d id s ee a r elation of the amount of"dissolution o f t he t ool edge"to the f ineness of the grain s tructure o f the f lint ( Anderson, 1 980, p .190) but d id not mention particular sources of raw materials f or her experimental t ools in her paper, which dealt not only with a rchaeological f lint but a lso with archaeological quartz implements. However, s he mentioned the sources for her experimental implements in her thesis ( 1981, Vol. I , p .6) Moss s tated that t he raw material of the ancient tools must be duplicated ( 1983, .

p .55) Different actions seemed t o produce a t most some differences in the location and extent o f the a reas o f polishes ( Keeley, 1 977, p .37) Distinct polishes f rom different actions were rarely mentioned; an exception was the difference between sawing and s craping or planing o f materials o f anisotropic structure, such a s the s awing a nd scraping or planing o f antler which l ed to"rough"and"smooth"antler polishes respectively ( Keeley, 1 977, p .44; 1 980, p .56). Other researchers agreed with t his ( Anderson-Gerfaud, 1 981, Vol. I , pp.61-62; Vaughan, 1 981, .

p .143) The e ffects of different durations of work were hardly discussed by Keeley in any o f his publications s ince he thought that the intensity of polish was an indication o f a particular worked material. He did not measure the e ffect o f duration of work when he measured the l ight reflectivity o f his polishes ( 1980, p .62) Vaughan ( 1981, pp.133-137) dealt with the t ime factor and pointed t o the s imilarity o f polishes at the beginning o f their f ormation. .

All researchers noted some variations in the polishes according t o the state o f the material, the polishes varying with the ( Keeley,

f reshness or dryness 1 980, p .44), hide

of such materials a s bone ( ibid., p .49) or wood ( ibid.,

p .56) However, Keeley for instance did not describe polish f rom working dry antler a s he had f ound that antler could not be worked efficiently unless fresh or soaked i n water .

( 1977,

p .44).

S imilarly, Keeley only speculated t hat t here i s probably no difference in polish according t o the species o f antler ( 1980, p .56) a lthough he did refer t o quantitative differences in polishes between those f rom"dense" woods and ,

those from"less dense"woods ( ibid., p .36). pp.154-155), Anderson-Gerfaud ( 1982 and

4 6

Vaughan ( 1981, 1 983, pp.90-91)

Moss ( 1983, p .94) a nd I ( l inger-Hamilton, 1 983, pp.245-246) noted d ifferences in plant polishes a ccording t o species of p lant. Variations in polish a ccording to t ool shape s eemed t o have been overlooked s ince Keeley ( in connection with experiments on antler) s tated,"Variations of the edge angle do not, a s was t he case with use on other materials, result i n any recognizable variation in the appearance of the microwear polish or s triations." ( 1980, p .59). Moss, however, ( Moss and Newcomer, 1 981; Moss, 1 983, p .55) did say t hat t he"i norphology o f the piece, particularly the working edges"should be duplicated experimentally"(1983, p .55). The e ffect on polish abrasives during use was

formation o f the rarely mentioned by

addition of researchers.

Those who did study such e ffects found that they c onsiderable ( Vaughan, 1 981, pp.163-164; Brink, pp.88-93 and pp.106-110; Mansur-Franchomme, 1 983)

were 1 978,

In summary, t here s eemed to be a general agreement between microwear analysts ( although Vaughan and Moss were more cautious than the others) that polishes are distinct a ccording to worked materials, and that neither the e ffects o f natural agencies nor o f t echnological processes, nor the e ffects of variables other than the worked materials, are s ufficiently strong to impede the functional analysis of ancient

f lint

tools.

It s eemed to me that this confidence was responsible f or the recent publications by microwear analysts who, a lthough arriving at very detailed conclusions ( such a s that implements had been used to cut meat c lose to the bone, or t hat implements had been hafted with bone hafts ( Bueller, 1 983)), omitted details o f experimentation, evidence in the f orm of comparative photographs, drawings of experimental t ools ( Bueller, ibid.) or s eemingly any specific experimentation at all ( Coqueugniot, 1 983; Keeley, 1 983) Whether such confidence i s justified will be subject to investigation in the following chapters.

4 7

Chapter

6

-

Polish,

Striations

and

Gloss

-

Aims

of

the

Investigation As indicated in the outline o f research ( Part I , Chapter 5 ) there were in my opinion s everal important questions which needed t o be answered before t he"I iigh P ower Approach"the diagnosis o f the u sed area of the t ool, the action with which and the material on which a tool had been used, by observation o f polish magnifications of 5 0x t o 4 00x could be The most important questions which -

and striations at a ccepted a s v iable. emerged were the

following: 1 ) Does polish really vary consistently a ccording t o an exact worked material ( such a s bone, antler or meat) a s Keeley ( 1980, p .83) would have us believe, i .e. does polish vary according to a man-made c lassification? In

order

to

investigate

this

problem

I

used

a

f ourfold

approach a ) I wanted to f ind out whether such a variation was l ikely by looking at the mechanism o f polish formation. I f polish was due to a l asting deposition of worked material on the tool surface, then a variation i n polish a ccording t o a precise worked material seemed logical. I a lso a ttempted t o construct a model of polish f ormation. In addition I t ried to a ssess the extent to which striations and macroscopic gloss were indicative o f an exact worked material. With the results of these investigations in mind I went on t o consider three topics: f irstly, whether r esidues f ound on f lint t ools by using SEM t ecniques, could be u sed t o identify materials ancient tools had been used on, secondly, to what extent c leaning with chemicals a ltered polishes and whether such c leaning was necessary, and f inally how the chemical composition of the soil could a ffect polishes on ‚

ancient

tools.

b ) I wanted t o f ind out whether c learcut differences i n polishes a ccording t o the exact worked materials were discernible in practice by observing my own f lint t ools. These had been used i n controlled experiments on a variety of

materials.

c ) I wanted to f ind out which variables other t han the worked material a ffected polish formation. I f they were f ew and of minor importance this would argue in f avour of consistent variations of polishes a ccording t o a n exact worked material. d ) In order t o t est my experimental r esults I took part in a s eries of blind t ests. I wanted t o investigate whether Keeley's success in determining exact worked materials could be replicated and what precisely the c ircumstances were under which microwear analyses t urned out to be most a ccurate.

4 8

2 ) The s econd most i mportant question was whether t he"High Power Approach"can be applied with equal c ertainty t o both a rchaeological and experimental t ools, or whether s urface a lterations on the f lint tools due t o natural ( or other) agencies not only render a good proportion o f the excavated f lint material unusable f or a microwear analysis, but a lso positively mislead t he microwear analyst. This question method o f

bears on the problem of sampling. t ackling this question was a " blind

The principal t est"in which

I analysed microwear polishes on tools excavated from Abu S alabikh before r eading the excavation report, i n order to f ind out whether the results o f a microwear analysis of ancient t ools matched the evidence f rom the excavation.

4 9

Chapter 1 .

7

The

-

Mechanism of

Polish

Formation

Recent Theories

Despite the long history of research into use-wear on f lint t ools the nature of polish formation i s s till a matter o f debate. Various hypotheses have been put f orward, most recently centring on the question whether use-wear polish on f lint i s primarily due t o the formation o f a deposit ( Witthoft, 1 967; Anderson, 1 980; Keeley, 1 980; Anderson-Gerfaud, 1 981, 1 982; P lisson, 1 983; Bradley and C layton, in press) or whether it i s due t o a ttrition ( Curwen, 1 930; Kamminga, 1 979; Masson et a l. 1 981; Meeks ‚

et a l. 1 982) i rreconcilable. ‚

.

These

hypotheses

appeared

at

f irst

s ight

There were a number o f arguments for the hypothesis of an additive use-wear polish on f lint. Anderson-Gerfaud Anderson, 1 980; Anderson-Gerfaud, 1981, 1 982) observed used edges on experimental f lint tools under the l ight microscope and under the SEM and suggested t hat under the l ight microscope the polish l ooked inflated ( 1982). Also, under the SEM what she took to be residues f rom contact materials such a s bone, or phytoliths f rom s iliceous plants appeared to be s inking into or melting onto the"dissolved"flint surface; these residues were used to identify the contact materials on which the archaeological f lint t ools had been used. Anderson-Gerfaud a lso compared objects seen on experimental f lint tools, which had been used t o cut plants, with photomicrographs o f phytoliths ( Wynn Parry and Smithson, 1 964) and then attempted with the help o f different"phytoliths" to identify d ifferent p lant species cut by archaeological tools. Anderson-Gerfaud ( e.g. -

-

,

1 982, pp.152-153) suggested that during the use o f a f lint t ool on plants the s ilica o f the f lint surface dissolves and amorphous s ilica i s precipitated, f orming a deposit o f s ilica gel around the phytoliths. The proposed cause was a chemical reaction involving a combination o f factors: the concentration of s ilica in water; high abrasion; a pH over 9 ; certain plant acids and other

f lint

s ilica

temperature; contact with

gels.

Keeley ( 1980, p .43) f ound that polish f rom bone on looked a ltered and"pitted"after treatment with hot

hydrochloric acid ( HC1) and deduced that the material r emoved was inorganic as i t had not been removed with sodium hydroxide ( NaOH). He suggested that i t probably consisted of c rystals of bone apatite which may have become incorporated ‚

in

the

polish.

P lisson removed by

( 1983) found that use-wear polishes could prolonged chemical attack with solutions

be o f

sodium hydroxide ( NaOH), calcium oxide ( CaO) and sodium carbonate ( Na CO) The speed o f removal varied according t o the contact material and the geological age o f the f lint. .

5 0

Bradley and C layton ( 1986) in their investigation into t he influence o f t he microstructure o f f lint on the f ormation of microwear polish referred t o the depositional e lement o f the polish ( as envisaged by Anderson-Gerfaud), a nd the possibility o f t he redeposition o f amorphous s ilica i n the interstices a round the l epispheres and quartz grains o f the f lint. In c ontrast, t here were a lso many arguments f or the hypothesis that use-wear polish on f lint i s due attrition. Masson e t a l. ( 1981) pointed out that while polish f ormation on f lint due t o attrition was well documented, additive polish was not, and had yet to be proven. They examined burnt s ickle blades from Mureybet using X-Ray D iffraction but could not detect any evidence for a surface l ayer o f amorphous s ilica and phytoliths in the form o f amorphous organic opal ( opal-A) and opal-CT, a s would be expected. The authors admitted the possibility of a s uperficial dissolution o f the f lint surface from attrition, but on too small a s cale to be detected by X-Ray Diffraction or to contain phytoliths. They stated that the history of a p iece of f lint begins with i ts formation, and not with its use, and t hat many structures interpreted as use-wear traces may be geological in origin. Meeks et a l. r esearch on plant s ectioned edges of polished with build-up on the Del Bene some polishes polishes from 2 .

( 1982) published, a s part of their polish, SEM micrographs showing the f lint tools used to cut plants or

diamond polished

paste. surface.

They

found

no

trace

of

a

( 1979; Holley and Del Bene, 1 981) thought that ( e. g. from plants) are due to deposition while other materials are due to attrition.

Experimental

Investigation under

the Light Microscope

The apparently i rreconcilable views outlined above led me to investigate some of my own experimental and a rchaeological f lint implements under the l ight microscope a t l OOx 200x f or a preliminary study. I l ater used SEM t echniques for a more detailed investigation. I decided t o t est the following suppositions: 1 ) I f polish i s due t o deposition then a major chemical reaction must take place during i ts formation. In order to f ind out whether a major chemical reaction i s involved, I tested the e ffects o f heat and pH both of which are known to affect -

chemical reactions on polish formation, by rubbing 1 2 B randon f lint pieces 2 00 t imes against each other with the addition of e ither cold or hot ( boiling) water, or with the addition o f acid ( 10% HC1, pH 4 -5) or a lkaline ( 10% NaOH, pH 1 0-11) s olutions. The resulting polishes were t oo s imilar t o conclude that a major chemical reaction i s involved i n polish f ormation. -

5 1

2 ) I f polish i s due t o deposition, then the physical s tructure of the worked material should not a ffect polish f ormation noticeably, i .e. polish f rom contact with antler should be the same regardless o f whether f lint i s rubbed against the surface of t he antler, or whether antler, a f ibrous material, i s cut across or l engthwise. Experiments ( Pl.29:a, b) and Blind Test 1 1-20 ( see P art I , Chapter 2 0) showed t hat polishes l ooked very different a ccording t o whether f lint had been used f or rubbing a material or cutting i t. This indicated that attrition i s the major cause o f polish. The variation took place i rrespective of whether the worked material was plant or a nimal, d iscrediting the hypothesis that polishes f rom p lants are due to deposition while polishes from some other materials are due t o attrition ( Del Bene, 1 979; Holley and Del Bene, 1 981) Two observations led me t o believe perhaps not only due t o attrition but a lso of amorphous s ilica on the f lint surface:

t hat polish i s due to a c oating

1 ) Experimental blades used t o work f resh or hydrated materials, such a s wood, plants, bone or antler, s howed a lmost"melted" looking surfaces ( e.g. P 1.23:d). 2 ) Blades with"sickle f rom the PPNA l evels

gloss", e specially at Jericho, revealed

those i n some

excavated i nstances

a "cracked"looking microscopic polish ( Pl.7:h) which r emained unaltered despite cleaning in soapy water and White S pirit, which would have removed most organic substances while l eaving the f lint surface intact. In summary, I deduced from t he preliminary study under the l ight microscope t hat polish i s f oremost the r esult o f attrition a lthough there were a lso indications o f the presence o f a coating o f amorphous s ilica on the surfaces of both the experimental and the archaeological tools. 3 .

Investigation under

the

SEM

In order t o f ind out whether there was any evidence for residues from worked materials embedded in a thick l ayer of amorphous s ilica such a s would constitute a depositional polish, I decided to investigate, using SEM t echniques, s ome of my own experimental blades which had been knapped f rom Brandon f lint ( see a lso Unger-Hamilton, 1 984) Two blades had been used to cut fresh English grass and bone respectively, each f or thirty minutes, while a third blade had been rubbed against another f lint blade f or ten m inutes. The blades had been washed in hot water and detergent. In addition s everal unused f reshly knapped blades were chosen for observation under the SEM. The results were a s f ollows: .

Evidence

concerning

the

The residues f rom t he Anderson-Gerfaud ( 1981,

residues

f rom worked materials:

working o f bone ( as shown by p .150, P 1.22:1, reproduced here a s

5 2

P 1.3:a) appeared only on t he f lint blade I had used t o saw bone ( Pl.3:c,e) They had not sunk into t he f lint s urface, but r ested on top o f i t, and t races of the"partings"betweef l the residues and the f lint surface were c learly v isible .

( Pl.3:e) The"phytoliths"and some other"cellular vegetable matter" shown by Anderson-Gerfaud ( 1981, p .122, P 1.8:2 ( the arrow indicates the area o f apparent embedding of the object in s ilica, t he t riangle points to e ither embedded constituents or an a rea o f recrystallised amorphous s ilica) and P 1.8:1 ( the a rrow picks out an a rea of supposed dissolution and r ecrystallisation of s ilica), p .140, p1.17:2, reproduced here a s P l.3: b, d and 4 :a) t o have"melted onto"or"sunk into"the f lint surface, appeared not only on f lint blades used t o cut plants; objects resembling her"phytoliths"were observed on the surface o f the f lint blade which had had no contact with any plant, but had been used exclusively t o polish other f lint ( Pl.3:f and 4 : b) The s ize o f the objects ( c. 2 0-30 microns) the characteristic ridges, and the attachment to the f lint surface a ll looked very s imilar ( Pl.3;b, d,f). In addition the shape and the"twist" o f the objects shown in P l.4: a, b looked identical, a lthough the s izes differed s lightly ( 20 and 1 5 microns respectively) S imilar objects, o f a variety of shapes and s izes, were a lso observed on the surfaces o f freshly knapped f lint which had had no contact with any other material ( Pl.4:c,d) The photograph shown in P1 .4 : e shows c learly one such object projecting f rom an unused f lint surface. ,

.

.

The s imilarity between t he objects found on f lint which had been in contact only with f lint, those f ound projecting f rom f reshly knapped unused f lint surfaces, and those s hown a s"phytoliths", l eads me t o suggest that they a re not phytoliths f rom plants melting onto or s inking i nto the f lint surface, but c omponents o f the f lint itself which project f rom the f lint surface. These objects might be relic grains o f s keletal calcite o f e chinoderms or Inoceramids ( A. S . Gale, K ing's College, London, pers. comm.) or relic organic t issues such as o f dinoflagellates -

which have not been fully replaced by s ilica ( Holdaway and Clayton, 1 982) The r ing o f dissolved s ilica around t he objects ( arrowed, P l.3: d) might well be due t o the deposition o f amorphous s ilica in the interstices between the f lint and the objects. The attachments around the objects ( arrowed, P l.3:b) might be due to a build- up o f amorphous s ilica on the f lint surface around the objects. Another possibility, t hat these objects are due t o recent contamination, such a s dust particles or pollen, c an be excluded on the grounds that dust or pollen would c learly

-

.

s ettle

on

top

of

the

f lint

surface,

rather

t han

s ink

into

it. Evidence concerning polish f ormation:

the

additive

5 3

or

abrasive

nature

o f

The s urfaces o f t he u sed f lint edges ( P1.3: b, d,f) appear t o be smoothed a nd p laned, and the"rings"immedjately surrounding the objects ( arrowed, P l.3:d) appear t o be"melted". This evidence suggests t he presence o f amorphous s ilica f illing t he interstices around t he objects and, possibly, a very t hin coating of amorphous s ilica on t he general abraded surface o f the f lint. I have suggested ove that the projecting objects on t he f lint surfaces might be r elic grains o f unreplaced skeletal calcite or relic organic t issues, and t he experiment o f one f lint being rubbed against another had shown that these objects remain projecting from t he polished surface. This l eads me t o suppose t hat polish formation i nvolves a strong chemical interaction rather t han mechanical abrasion a lone. This i s because t he softer relics of organisms would t end t o be worn away before the harder f lint ductile, i n which case t hey might i nto the hollows of the f lint)

t he

The results following:

The

only

o f

residues

the

SEM

c learly

surface ( unless they were g ive way and be pushed

i nvestigation

due

t o

l ed me

use-wear

t o

which

conclude

I

have

detected on the three blades investigated are deposited on t op of the f lint surface, and have not sunk into i t. I t t herefore seems unlikely that such residues would be preserved on a f lint surface f or a l ong t ime, or that due t o the preservation o f the residues, polishes on archaeological f lint tools derive their characteristic appearance from a specific worked material such a s bone or hide. This f act a lso l eads me to suggest that c leaning with chemicals, such a s was carried out by Keeley ( 1980, p .43) and by P lisson ( 1983, P lisson and Mauger, 1 986) would remove any residues which did remain. In addition, I concluded f rom t he s imilarity between the objects on unused f lint surfaces and t he so-called residues f rom worked materials such a s " phytoliths"that a s Masson et a l. ( 1981) had pointed out the genesis and diagenesis o f the f lint i tself need to be known in order t ' at the researcher can differentiate between residues f rom the use o f f lint and organisms i ncorporated i n t he f lint during i ts formation ( see Part I , Chapter 1 1). -

-

The f act that the objects remai ed protruding f rom t he f lint a fter one f lint was rubbed against another leads me t o conclude that there does s eem t o be a chemical dissolution of the f lint surface a s Anderson-Gerfaud ( Anderson, 1 980; Anderson-Gerfaud, 1 981, 1 982) had suspected. However, there was no evidence for a l ayer of amorphous s ilica thick enough ( i.e. on t he scale of several microns) t o contain residues such a s phytoliths f rom worked materials. A well-formed Polish could be produced by a very thin and generally invisible coating of amorphous s ilica, particularly i f infilling i nterstices on an a lready abraded surface. Most o f this coating would s eem t o be deposited in t he interstices between the f lint grains around the l epispheres of t he ‚

f lint

( Bradley

and

Clayton,

5 4

1 986)

and

a lso

in

the

interstices around t he r elics o f organisms ( see above) The thinness o f such a l ayer ( detectable on an Angstrom s cale only) would explain why neither the analysis by X-Ray Diffraction ( Masson et a l. 1 981) nor t he observation under the S EM ( Meeks et a l. 1 982) registered it. .

‚

‚

4 .

Conclusions

I t therefore seemed investigation under the

f rom l ight

both the microscope

preliminary and the

investigation under the SEM that use-wear polish on f lint i s the result of both abrasion and deposition. Other experiments ( Unger-Hamilton, 1 983, s ee P art I , Chapter 1 6) l ed me t o suggest that among other f actors water plays an e ssential role in t he process of polish f ormation, a s others have a lso suggested ( Anderson, 1 980, p .181; Gysels and Cahen, 1 982) Two recent publications seem to support my conclusions: Mansur-Franchomme ( 1983) has investigated the e ffects o f abrasion and humidity on polish formation under the SEM and has found that polish was influenced by both. Anderson and Whitlow ( 1983), two physicists, have used hydrogen profiling ( by means o f I on-Beam Analysis) to study water uptake o f f lint surfaces. They found that the modification of hydrogen profiles through the working o f different materials appeared to be strong evidence in f avour of the gel-formation hypothesis. The maximum depth o f such a layer ( caused by working wet and hard materials) which t hey mentioned was 0 .7 microns ( i.e. not on a scale large enough to contain phytoliths, as I stated above). Anderson and Whitlow ( ibid., p .471) a lso thought that the thinness of such a l ayer of amorphous s ilica would not have been registered by X-Ray Diffraction analysis a s carried out by Masson et a l. ( 1981). Although Mansur-Franchomme ( op. c it.) and Anderson and Whitlow ( op. cit.) s eemed to accept Anderson-Gerfaud's identification o f residues a s phytoliths on f lint surfaces, there was in my investigation no evidence for a deposit in which residues frDm the worked materials are embedded. This means that there s eems t o be no magical ingredient in the use-wear polish which would make i t characteristic of a worked material.

5 5

Chapter My

8

-

A Model

of

investigation

Polish o f

Formation

experimental

f lint

t ools

u sing

S EM

t echniques l ed me to conclude that polish was the result o f abrasion and a coating o f amorphous s ilica ( Unger-Hamilton, 1 984) Other observations ( Unger- Hamilton, 1 983) had shown t hat water plays an e ssential role in this process and t hat i t seems instrumental in the f ormation o f amorphous s ilica ( see Part I , Chapter 7 ) .

Observations under the l ight microscope had l ikewise shown that the presence o f amorphous s ilica s eemed l argely dependent on the water content o f t he worked material. Surfaces of experimental f lint t ools s eemed f lattened when very hard materials were worked, while they were only gently polished when softer materials were worked ( see Part I , Chapter

1 4)

I would formation:

therefore

propose

the

f ollowing model

of

polish

1 ) ( Fig. la) When the worked material i s very hard then the t ops of the microtopography o f the f lint are completely f lattened. 2 ) ( Fig. ib) When the worked material i s medium hard then only the tops o f the r n icrotopography are polished, l eading to the p itted or reticular look often mentioned. 3 ) ( Fig. i c) surface of polished.

When the

the material i s f lint, including

soft the

t hen t he whole depressions, i s

The amorphous s ilica probably l iberated by the humidity in the worked material c ollects a round the f lint grains and around the l epispheres of the f lint ( Bradley and Clayton, in press) and around t he micro-organisms in t he f lint ( see Part I , Chapter 7 ) I t i s unknown how such a m orphous s ilica i s bonded to the f lint surface. Other f eatures, e . g. the horizontal distribution of polishes ( related t o the extent o f contact with the worked material) or characteristics l ike"bumpyness"in the case of hide polish ( probably related t o the structure of t he material) will l ead to a more precise identification o f the worked material. -

-

.

This model o f polish formation was t ested by applying these criteria in Blind Test 1 1-20 ( see Part I , Chapter 2 0) with a success rate of approximately 7 out of 1 0 correct identifications o f the worked materials.

5 6

Chapter 1 .

9

-

Striations

Formation

Mansur - Franchomme ( Mansur, 1 981) i nvestigated t he f ormation o f s triations u sing S EM t echniques. S he s tated t hat s triations were caused by extraneous sand and dust particles c aught in between t he surfaces o f t he f lint and t he worked material, and a lso by microflakes r emoved f rom the t ool during u se. She defined t hree categories f or t he s tate o f t he surface o f t he f lint t ool, based on the degree o f f luidity a t t he moment o f utilisation. To each c ategory corresponded a special t ype of s triation. These were:"fluid-gel s tate" s triations f rom working wet or f resh materials such a s f resh plants,"intermediate-gel state"striations f rom working wood and fresh hide, and"solid-gel s tate"strjations f rom working dry materials, such a s dry hide ( Mansur, 1 981; Mansur- Franchomme, 1 983, p .229) I did not investigate t he f ormation o f striations using SEM t echniques, a s her and my understanding o f polish f ormation s eemed t o be l argely in agreement ( see Part I , Chapter 7 ) and because the f ollowing observations by others and myself seemed t o support her model o f the formation of s triations: When I rubbed f lint against quartzite, many striations were visible on the f lint, however when I rubbed t he same material f or the s ame number of s trokes with the addition of water, hardly any s triations were v isible. P lisson ( 1983; P lisson and Mauger, in 1 986) demonstrated that previously i nvisible s triations could be"uncovered" by t reating t he f lint surface with chemicals which a ttack amorphous s ilica. 2 .

The Value of

Striations

to Microwear Analysis

Moss ( 1983, p .74) discussed striations in S emenov's s ense i .e. a s"any k ind o f l inear depressions in the f lint surface, which i s not a f eature o f f lint i tself"(jbjd j. She pointed to the fact that for S emenov". . striations were the most important functionally d iagnostic f eatures" with which he not only located u se-wear on a t ool, but a lso identified the motion during t ool u se. All microwear analysts s eemed t o agree t hat t he l ocation of use-wear on a tool and t he motion during u se c an be identified with the help o f s triations, a lthough Keeley ( Keeley and Newcomer, 1 977, p .37) had s tated c autiously that striations should only be c onsidered t o be t he result of i ntentional work when they are a ccompanied by microwear polishes. In addition, Moss ( op.cit., p .74) observed ( according to my experiments quite r ightly) t hat there might be s ometimes"jncjdental s triations"during work which do"not a lways conform t o the overall expected pattern" ( ibid.) ‚

.

5 7

However t he extent t o which t he occurrence and nature of striations indicate the worked material s eemed t o be open to question. The f ollowing f actors might be the cause o f striations: Microchipping of the l argely responsible

f lint i tself during use was held for t he formation o f s triations

researchers ( see Mansur-Franchomme, 1 981, 1 979; and Fedje, 1 979, who observed striations on obsidian.)

t o by

be a ll

1 983; Del Bene, t he f ormation of

The accidental presence or deliberate addition of grit during use of a stone t ool was a lso held r esponsible f or the formation of s triations by a ll researchers ( see S e i nenov, 1 964, p .15; Fedje, op. c it.; Mansur, 1 981; Moss, 1 983). The worked material i tself ( presumably its s tructure) caused in some instances characteristic striations a ccording t o Keeley ( 1980, p .23) He described such striations t ogether with the microwear polishes: e . g. broad ( approximately 1 5 micron) and shallow striations on woodworking tools ( ibid., p .35) deep and narrow s triations on boneworking tools ( ibid., p .43), narrow, deep and relatively broad and shallow striations sometimes on hideworking tools ( ibid., p .50), very f ew and very minute striations on tools used on meat ( ibid., p .54), rare, narrow, not particularly deep and short s triations on antlerworking tools ( ibid., p .56) ‚ "filled in"striations ( following Witthoft, 1 967) on plantworking tools. Vaughan ( 1981, p .121) and Moss ( 1983, p .76) a lso t hought that in some instances striations were specific t o a particular worked material, while Mansur had l isted the worked material only a s an indirect cause o f striations ( 1981) .

,

My

own

experiments

showed

the

f ollowing:

Microchipping a s a cause o f striations was demonstrated by the fact that in general the quantity of s triations on a f lint tool related directly t o the hardness of the worked material. However there were exceptions ( see below). Abrasion f rom external agents such a s grit and dust ( see Korobkova, 1 981, p .331), rather t han from the presence o f s ilica in the plants ( Steensberg, 1 943) s eemed to be l argely the cause of striations in plant polish. R iverine plants, regardless of whether harvested in England or Syria, ,

did not cause any striations on f lint surfaces, even when bristling with s ilica, such a s was t he case with Equisetum f luviatile ( Pl.25:f) aptly named"The S couring Rush"or when very hard, such a s was the case with cane ( Pl.24:d). Land plants, such a s grass ( pl.25:d) when harvested f rom compact soil held together by a grassy cover, caused f ew s triations, regardless of whether t hey were f resh or dry. P lants harvested from upturned, but humid and humus-rich ( and therefore compacted) soil in England caused more striae ‚

-

‚

5 8

( occasionally i n g reat quantity) t han did l and plants cut f rom a grass c over ( Pl.22:f). ( Plenty o f s triations a re l ikely t o o ccur when p lants are cut f rom upturned l oose s oils i n t he Near East.) The worked material

( i.e.

i ts

structure)

did

s eem

t o

c ause

s triations; this was most evident when f ibrous materials, s uch a s sheep's wool ( Pl.14:e) and s inew ( Pl.16:b) were cut. I n neither case d id any edge damage o ccur and a t l east t he s inew was completely grit-free. However, the polish c onsisted a lmost s olely o f s triations, which a re perhaps t he negative i mpressions of t he f ibers. S ome materials such a s s hell and f allow-deer antler caused s triations which were c onsistent and r egularly spaced.

o f o f

I t t herefore s eems t o me that s triations a re not only value i n the l ocation of use-wear and t he i dentification t he motion during use, but that t hey can a lso be

i ndicative o f t he hardness and the humidity o f t he worked material, the presence o f abrasives ( which i n my opinion i s a particularly i mportant indicator for t he investigation i nto plant cultivation), and the structure o f the worked material i tself. Striations due to causes other t han u se, such a s manufacture, natural agencies and handling during and a fter excavation are discussed in Part I , Chapter 1 8 together with post-depositiona] . polishes.

5 9

Chapter 1 .

10

Macroscopic Gloss

-

Formation

Gloss which i s visible t o the naked eye has often been referred t o a s"sickle gloss, s ickle polish, s ickle sheen, corn gloss..." ( Diamond, 1 979, p .159) indicating t hat the occurrence of such a gloss was usually a ssociated w ith t he harvesting of plants and especially o f grasses ( Vaughan, 1 981, p .156) The reason for this i s t hat this g loss had been seen primarily on blades whose shape indicated their use a s s ickle blades, and that experiments ( Curwen, 1 930, 1 935) had shown that while prolonged use o f f lint on wood and straw caused gloss on the f lint surface, prolonged use on bone did not. Researchers therefore concluded that gloss i s due to the presence o f s ilica in t he plants or due t o the presence o f s ilica in other materials such a s sand, f linty soil and wood ( Curwen, 1 937, p .93). Witthoft ( 1967, p .385) in fact thought that gloss i s due to an accumulated l ayer of fused s ilica glass and opaline molecules. .

My own experiments l ed me to agree for some t ime with the conclusion that gloss i s entirely due t o the presence o f s ilica in the worked material. I had found that cutting plants with a high s ilica content, such a s Equisetum f luviatile ( see 1 1er, 1979, p .742) l ed to gloss f ormation much f aster than cutting modern c ereals ( see 1 1er, 1 979, ,

p .744; H illman, pers.comm.) which are bred f or a low s ilica content: i t t ook 2 00 s . m. ( sawing movements) on Equisetum and 5 000 s . m. on modern barley to get gloss on the cutting edge of the b lades. I also found t hat cutting or s craping s iliceous materials such as stone, s and or wood l ed t o gloss formation on the f lint tool while cutting or s craping materials such a s bone, antler or hide, which contain very l ittle s ilica ( generally 2 -10% in t he a sh, s ee 1 1er, 1 979, p .754) d id not. However, I later noticed that rubbing the same materials with f lint led to gloss formation a ccording to relative hardness of the worked material, i .e. rubbing bone caused considerable gloss, while rubbing meat d id not. The difference between the effects o f s awing, scraping and rubbing such anisohydrous materials a s bone and antler may have been due to the f act that when such a material i s cut or scraped, l ess of the worked material comes into c ontact with the f lint, and l ess intensely, t han when the harder outer surface i s rubbed. This l ed me to conclude t hat the hardness of formation. very hard.

the worked material i s Phytoliths and other s ilica

important bodies are

i n o f

gloss course

In order t o f ind out whether i t i s the hardness o f such bodies or t he f act that they consist o f s ilica which l eads to gloss formation, I cut another h ighly s iliceous, but soft, material, a f eather ( feathers contain up to 7 7% s ilica in the a sh ( 11er, 1 979, p .753)). No g loss f ormed even a fter prolonged contact. Conversely, intense gloss developed on the f lint t ool when copper ( which c ontains at most t race

6 0

amounts

o f

s ilica)

was

drilled.

In addition I had observed that greatly i nfluenced by t he moisture

gloss formation was content o f t he worked

material ( Unger-Hamilton, 1 983, s ee a lso Chapter 1 7). I t herefore conclude that gloss, l ike microscopic polish, i s t he result of abrasion and moisture and t hat i t i s s imply the macroscopic manifestation o f a strongly polished f lint surface. 2 .

The Value

of

Gloss

to Microwear Analysis

The above conclusion, that gloss i s t he result o f abrasion and moisture implies that it i s not indicative of a particular worked material, such as grass or plant. However, the distribution of gloss did s eem to indicate the worked material: gloss f rom hard materials such as s tone or bone, looked streaky, while gloss from soft materials, such as plants, l ooked evenly distributed. In practice, I have never seen any s tronq and evenly distributed macroscopic gloss on experimental f lint tools other than those used on plants. ‚

6 1

Chapter

11

-

Residues

Vaughan ( 1981, p .91) s tated,"Under c ertain ideal conditions o f preservation, organic r esidues such a s vegetal f ibers or amino acids may be l eft adhering t o prehistoric stone tools and can be s tudied by relatively straight-forward methods of chemical or physical examination."He referred t o the publications o f Briuer ( 1976) Broderick ( 1979) and Shafer and Holloway ( 1979) ,

I myself observed under the l ight microscope inorganic residues, such as from l apis l azuli, on experimental f lint drills and possibly on drills excavated from Abu S alabikh. S imilar residues had been reported by Tosi and P iperno ( 1973) on drills excavated from Shahr-i-Sokhta. Such inorganic residues presumably do not need"certain ideal conditions"(Vaughan, op. c it.) such humidity for their preservation.

a s

extreme

aridity

or

As a lready outlined in P art I , Chapter 7 , Anderson-Gerfaud had observed under the SEM residues f rom worked materials such a s residues f rom bone, or phytoliths from s iliceous plants which appeared to be s inking into or melting onto the"dissolved"flint surface ( Anderson, 1 980; -

-

Anderson-Gerfaud, 1 981, Vol.1, pp.100-101, 1 982). S imilar residues ( from hide) were observed, a lso under the SEM, by Mansur-Franchomme ( 1983) Most microwear analysts seemed t o accept that the residues observed were in all cases residues .

from the worked materials, and that such residues could be used t o identify the worked materials on which archaeological tools had been used ( Keeley, 1 981; Vaughan, 1 981, p .91), a lthough Moss ( 1983, pp.17-18) expressed some reservations. However as a lso outlined in Chapter 7 , there were some arguments against the survival o f such residues f rom the worked materials on f lint, and I came to the following conclusions: ‚

As Masson et a l. ( 1981) have pointed out, the genesis and diagenesis o f the f lint i tself need to be known i n order that the researcher can differentiate between residues from the use of f lint and organisms incorporated in the f lint during i ts formation. Provenance and geological age must a lso be taken into account. As"flint"(or more specifically"chert") has formed in different areas and different s ediment types during different geological periods, it naturally shows a great variety of f orms and particularly o f different quantities and types of organisms. Species and quantities of macro- and microorganisms may vary even within one f lint source. Therefore i t i s desirable to study unused f lint f rom the relevant sources under the SEM before residues on the experimental or archaeological t ools can be identified with certainty.

6 2

Masson e t a l. ( op. c it.) concluded f rom their i nvestigations o f burnt blades with"siljca gloss" that there were no phytoliths on the surfaces of t hese tools. Their r esults agree w ith my f indings which suggest that objects r esembling supposed phytoliths appear on f lint surfaces used only t o polish o ther f lint. The possibility t hat phytoliths m ight be preserved on a f lint surface c annot at present be ruled out. Convincing evidence in the form o f SEM micrographs of phytoliths f rom plants, however, has yet to be presented. ( The only SEM micrograph o f a phytolith Anderson-Gerfaud s eems to have published until now ( 1983, p .103, pl.1:7b) did not convince me of i ts s imilarity to a c orresponding " phytolith"on a f lint tool ( ibid., pl.1:7a)).

6 3

Chapter

12

-

Cleaning

At the beginning o f this microwear analysis i t was widely a ccepted t hat c leaning with chemicals o f experimental and archaeological f lint tools must b e carried out before examining the tools under the microscope. Keeley ( 1980, pp.10-11) had c leaned most of h is experimental and archaeological tools with detergent, White Spirit, a 1 0 96 solution of warm HC1, and a 2 0-30% solution o f NaOH. Anderson-Gerfaud ( 1981, Vol.1, p .28-29) had f ollowed Keeley's NaOH, a s

procedure initially, she was aware o f the

but had substituted H 2 Ox for fact that NaOH a ttacks s ilica.

She also restricted chemical c leaning to the experimental tools. Vaughan had cleaned his experimental ( 1981, p .97) and archaeological ( ibid., p .235) tools i n soapy water, a cetone and in some cases i n a 1 5% solution of HCl. Moss ( 1983, p .105) had initially followed Keeley's method o f c leaning with dilute HC1 and NaOH, but had l ater used acetone, and occasionally dilute NaOH as she discovered that she"found absolutely no difference in the appearance of t he wear t races"after comparing chemically c leaned and uncleaned, experimental and archaeological implements. ‚

Chemical c leaning had been carried out in order to remove remnants from the worked materials on experimental tools and a lso extraneous material which"is o ften deposited on an implement surface while the implement i s buried in the archaeological deposit or during the handling of i t a fter recovery by the archaeologist and the microwear analyst. Such material obscures existing wear t races and may even be l iable t o misinterpretation as an e ffect o f use" ( Keeley, ibid.,

p .10).

However, a s mentioned above in Chapter 7 , Keeley ( 1980. p .43) had found that cleaning with HC1 a ltered bone polish. Vaughan stated that,"Systematic use o f the a cid ( HCl) was avoided because it removed diagnostic residues, such a s those f rom bone/antler, dried beef and stone...; and i t destroyed the surface of f lints i n which there were even minute l imestone impurities within the s ilicate matrix"(1981, p .97). P lisson ( 1983, P lisson and Mauger, 1986) had found that use-wear polishes could be removed by prolonged chemical attack with solutions of NaOH, CaO and Na. CO 3 , and that the speed of removal varied according to contact material and the geological a ge of the f lint.

the

My own observations under the S EM had shown that the only residues c learly due to use-wear which I had detected on the three blades investigated were deposited on t op o f the f lint surface and had not sunk into i t ( see Part I , Chapter 7 ). This would make their r emoval with chemicals, such a s reported by Keeley ( 1980, p .43), Vaughan ( 1981, p .97), and by P lisson ( 1983; Mauger a nd P lisson, 1 986) very easy. Keeley ( ibid., p . 11) had used solutions of HC1 which i s known to a ttack the calcium of which the bone residues may consist. He a lso noted ( ibid.) t hat prolonged immersion

6 4

o f t he f lint in immersed t ool, c leaning would a rtefacts preserved,

s olutions o f NaOH l ed t o patination of the a nd pointed out ( ibid., p .10) that chemical r emove organic residues surviving on

recovered f rom such a s dry cave

environments s ites ( Wylie,

in which 1 975)

they

are

P lisson ( op. c it.) u sed solutions o f NaOH and Na CO which ( as he was aware) to some extent may a ttack amorphous s ilica such as would form a deposit ( such a process was demonstrated by Anderson and Whitlow, 1 983, p .471) and to a lesser extent may attack the f lint grains themselves. Moss ( 1983, p .92) mentioned the fact that she did not f ind a polish"bevel"from bone ( such a s observed by Vaughan) on her experimental f lint implements. This might well be due to the fact t hat Moss cleaned her experimental tools in dilute NaOH, a s in my opinion a "bevel"is a build-up of amorphous

( see Chapter silica.

1 4)

such

While the above evidence suggested to me that neither the experimental nor the archaeological tools should be c leaned in chemical solutions the following problems remained: Obvious extraneous deposits ( identified a s such by virtue of their appearance, which differed from that o f polish) were s een on some of my experimental tools. However they tended to fall o ff after some t ime in storage, and, a s I discovered when I viewed the f lint t ools of Blind Test 1 -10 before and a fter immersion in dilute NaOH, their presence or absence did not a lter my interpretation of the polishes ( see Part I , Chapter 2 0) Residues from hafting, such a s wax, resin or c ellulose adhesive ( Uhu), had to be removed in order to look for hafting t races. I initially attempted to remove these with a cetone, but soon f ound that the acetone combined, especially with Uhu, to l eave acetone"bloom" on the f lint surface. I therefore c leaned hafted tools in White Spirit which removed wax and resin satisfactorily and which does not attack f lint ( Dr. Don Robins, pers. comm.). Limestone concretions-obscured the edges of a few of the archaeological implements completely. I did c lean such tools in a 1 0% solution of HC1 but a lways noted the exact c leaning procedure in order t o t reat experimental tools in the same way for comparison. Grease water.

f rom handling

o f

the

tools

was

c leaned

with

soapy

I therefore decided to c lean the experimental and the archaeological t ools only in water and ammonia free detergent, and then, i f necessary, in an ultrasonic c leaning t ank. In the f ew instances where chemicals had to be used, i .e. when experimental tools had been hafted, or when

6 5

archaeological t ools were covered with concretions, I a lways noted t he exact c leaning procedure. F rom the observed ( see B lind Tests 1 -10, Part I , Chapter 2 0) and the reported a lteration of polishes due t o c leaning with chemicals ( Keeley, 1 980, p .43; Vaughan, 1 981, p .97; P lisson, 1 983; P lisson and Mauger, 1 986) it s eems t o me t hat the exact c leaning procedure should a lways be s tated in publications.

6 6

Chapter

13

-

The

Influence

of

Soil

Chemistry on

Polishes

P lisson ( 1983) remarked that his research into the e ffect of chemicals on polished f lint surfaces suggested t hat the chemical composition o f the soil might s ignificantly a ffect the polish on the f lint tools excavated f rom it. This process was demonstrated by Anderson and Whitlow ( 1983, pp.471-472) who found that archaeological f lints excavated f rom a lkaline environments showed s imilarly changed amorphous s ilica layers a s did experimental f lints which had been r insed in NaOH. In ( 1981, Vol.1, p .7) stated that experimental tools chemically in natural environments from which However she did not state what the these various layers were.

fact, Anderson-Gerfaud she c leaned some of her order t o imitate the the t ools chemical

were excavated. c ompositons of

It s eems to me that while it would be impossible to analyse t he composition of a particular soil l ayer over the ages ( as such compositions change according to such f actors a s the c limate), and while it would be very t ime-consuming t o analyse the chemical composition of the soil in its present state down to every chemical, it i s nevertheless advisable t o collect soil samples f rom t he site in order to get some idea of i ts pH which, i f very a lkaline, would tend to a ffect amorphous s ilica. The analysis o f the soil f rom Arjoune, Trench VI, carried out by means o f a B .D. H. Barium Sulphate Test, showed the following pH values: the soil from the pre-occupation l ayer had a pH o f 7 .5-8.0; the soil f rom the occupation l ayer had a pH of 7 .0; the soil of the a lluvium above the o ccupation layer had a pH of 7 .0-7.5. As the occupation layer showed a neutral pH value and a s the investigation of the archaeological tools showed that in most instances the amorphous s ilica coating seemed to be intact most obviously on the s ickle blades I did not c lean my experimental tools in a lkaline solutions such as of -

-

NaOH.

6 7

Chapter 1 .

14

-

Experimental

Polishes

Method

Four Hundred and t hirty-six implements made of f lint f rom Brandon in Suffolk, Surrey and P otter's Bar, London, i n England, o f f lint and chert f rom Arjoune in Syria, a nd f rom t he Dordogne and P incevent in France, were used to s aw, cut, s crape, whittle, plane, bore, drill, p ierce, chop, grave and rub the f ollowing broad categories o f materials: wood, bone, hide, antler, horn, meat, f ish, shell, stone, pottery, sand, f eather, wool, ivory, copper and p lants. S ome experimental t ools were used hafted. The experimental procedures are described in Part I I in relation to t he tool c lasses. The experimental t ools were cleaned in warm water and ammonia-free detergent and o ften for up to 1 0 minutes in an u ltrasonic c leaning tank in distilled water. No chemicals were used. The tools were observed under the l ight microscope at magnifications of: 5 0x look

-

in order to locate the at the edge damage ( see

l OOx in order to the polish and the -

polish, Part I ,

look at the striations.

define i ts Chapter 4 ).

directional

extent

d istribution

and

o f

2 00x in order t o look at the character of the polish ( i.e. the degree of abrasion of the f lint surface, the amount of amorphous s ilica and the distribution o f both) and t he depth and the width of striations. -

Unless otherwise indicated, the descriptions and the photographs of the polishes are based on magnifications o f 2 00x, with the investigated surface held parallel to the microscope l ens. The understanding o f polish formation, and therefore its description, i s based on observations under the SEM which indicated that polish i s formed as the result of abrasion and a very thin coating o f amorphous s ilica ( see Part I , Chapter 7 , Unger-Hamilton, 1 984). Although I described the polishes f rom a s lightly different viewpoint from that of other researchers, i .e. I described what I thought had happened to the f lint surface rather than just the appearance o f the a ltered f lint surface, I found that my observations o f polishes generally agreed with those o f other microwear analysts ( Keeley, 1 980; Anderson-Gerfaud, 1 981; Vaughan, 1 981; Moss, 1 983) 2 .

Wood

Fifty-six experiments had been carried out sawing, s craping, whittling, boring, drilling, graving, chopping and rubbing the f ollowing kinds of seasoned and f resh wood ( in some instances soaked in water f or one day) : oak, ash, sycamore, birch, cherry and pine.

6 8

The resulting polishes were quite varied. In general t he r nicrotopography o f the f lint was gently polished, l eading t o the"domed"look ( Keeley, 1 980, p .35) o ften described a s characteristic of wood polishes. The undulating, brilliant look of the f lint surface i s probably caused by t he amount o f amorphous s ilica which I believe i s due t o the amorphisation of the f lint surface. The brilliance and amount o f"liquid" looking polish varied with the moisture content of the wood ( P1.8:e, g). Some woods, l ike sycamore, a lso caused a greater degree o f a x norphisation o f the f lint surface than did others. Cutting pine produced only l ittle polish on t he tool surface. The l ikely reason f or this was the f act that the f lint became covered in pine resin. Contact with bark caused rougher looking polish than did contact with heartwood. The distribution o f wood polish was generally wide, presumably because wood i s a yielding material. I t depended on the action with which the tool was used, i .e. when wood was s craped, a continuous l ine of polish, a k ind of " bevel", probably of amorphous s ilica, was f ound at the s craping edge ( Pl.9:a). Such a "bevel"was not f ound on wood saws. The distribution of wood polish i s generally described a s " reticular"(Moss, 1 983, p .91). However, I found the " reticularity"or the degree of horizontal l inkage of the polish varied with the hardness of the wood, the grain s ize of the f lint, and with the pressure applied and drilling led to completely l inked smooth polish with few s triations at the drill t ip ( Pl.12:a), while boring l ed to an incompletely l inked polish with striations at the borer t ip ( Pl.11:a) Striations were not common on .

woodworking tools. This agreed with Moss' experimental results ( 1983, p .91) and disagreed with those o f Keeley ( 1980, p .35) who found a distinctive broad shallow type o f s triation on his woodworking tools. However, I f ound that s triations apparently f illed with amorphous s ilica ( Mansur, 1 981) were the most common types of striations from woodworking. S imilarity

to

other

polishes:

In Blind Test 1-10 I consistently mistook wood polish for ( reindeer) antler polish. The same had happened to other microwear analysts ( e.g. Keeley, in Keeley and Newcomer, 1 977, p .55) I f eel that the following could have been the reasons for this mistake: 1 ) I was not familiar with .

reindeer antler polish at the t ime. 2 ) Newcomer, who carried out the experiments for these blind tests, had used the s craper at an angle of 9 0 degrees while I had used my own at angles of c . 4 5 degrees and therefore the distribution of polish on the t est s craper had been s lightly unusual for ‚

‚

me. I a lso found a s imilarity between fresh wood and reed polishes ( Pl.13:c and 2 3: d). However, reed polishes were a lways distinct because of the complete absence of s triations on them. Leather, when soaked in water for one day, caused a polish virtually identical to wood polish

69

( Pl.29:c and 9 :a) The"bevel", o r apparent build-up o f amorphous s ilica found on some woodworking t ools c ould be confused with such bevels on t ools u sed on soaked antler and f resh or cooked bone ( Pl.9:a, g, h). .

3 .

Bone

Fifty-five experiments were carried out s awing, scraping, whittling, boring, drilling, graving, chopping, rubbing bone ( long bones, shoulder b lades and ribs) o f t he following animals: cattle, pig, bones were worked fresh, dried,

sheep, deer and chicken. roasted and cooked.

The

L ike Newcomer ( 1974) I f ound that soaking bone in water did not make bone working easier, and therefore I did not soak bone a fter my initial experiments. Bone polish in general unyielding material.

was

i solated,

a s

bone

i s

an

When a f lint had been used on dried bone, the polish looked " pitted", probably because bone, which i s hard, had f lattened the very tops of the microtopography of the f lint only, l eaving the depressions in the f lint surface unpolished. Very l ittle amorphous s ilica was s een, which i s probably why

the

polish

looked matt

( Pl.15:c).

However, when fresh or cooked bones were worked, the f lint surface looked more brilliant, probably due t o t he presence of amorphous s ilica which often c ollected a s a " bevel"(Vaughan, 1 981, p .141) on the tool edges ( Pl.9:h and 1 1:e). ( But note the"bevel"from dry bone on P 1.19:a). The fact that Moss ( 1983, p .92) had not detected such a bevel on her experimental tools ( although she had seen it on prehistoric tools) could well be due to t he fact that she had c leaned her experimental t ools in dilute NaOH which can attack amorphous s ilica ( see Part I , Chapter 1 2) Keeley ( 1980, p .43) discovered that a fter c leaning his boneworking tools in solutions of HC1, t he polishes looked pitted, and he suggested that the removed e lements had been bone apatite which had become embedded in the polish. ( This i s discussed in Part I , Chapters 7 and 1 2 : I had found bone residues adhering to, but not embedded i n the f lint). I f such residues merely adhere to the f lint ( albeit for a l ong t ime 6 months as was shown in my experiments) i t follows t hat such residues might eventually fall off and that bone polishes will look"pitted", a s Keeley had described, even when HCl i s not used to c lean boneworking tools. .

-

-

The distribution o f bone polish varied a ccording t o the action; a s Vaughan ( op. c it., p .168) had observed, polish bevels only resulted from t ransverse a ctions, while a "smooth-pitted" bone.

polish

I only observed

one

( ibid,

p .135)

instance

7 0

resulted

where

a

from

sawing

difference

in

polish might have been due t o a difference in species o f animal: sawing t he l ong bones o f ( roast) chicken a lways c aused wide bands o f bright polish ( Pl.29:f) not seen on f lint used t o saw bones from other animals. S triations t ended to be deep and o ften s een on polishes f rom sawing working bone with t ransverse actions, a lthough polish bevels on such s craping l ined ( Pl.9:h, s ee a lso P 1.11:e) S imilarity

to

other

narrow. They were bone, but not f rom such a s s craping, edges t ended to look

polishes:

As Vaughan ( 1981, p .135) pointed out, polishes from sawing o f bone and antler could be virtually indistinguishable. I a lso found that bevels o f bone polish could l ook l ike bevels f rom wood or antler s craping or graving. 4 .

Hide

Twenty experiments had cutting, piercing, drilling

been carried out and rubbing dry and

s craping, f resh deer

hide. Also, two experimental tools used by Bergman on hide with the addition o f ochre were examined. Some experiments were carried out s craping pig skin, dry and soaked in water f or one hour. In general hide i s fairly yielding, and working it l ed t o the gentle polishing of t he f lint surface which g ives i t the"bumpy","matt" l ook ( Keeley, 1 980, p .49) often described ( Pl.10:g and 1 4:g). There were great variations according t o the state of the hide: dry hide l ess y ielding than f resh caused t he"pitted"(Keeley, ibid.) look by only polishing the t ops of the microtopography of the f lint. Fresh hide more y ielding, and therefore polishing the f lint depressions a s well a s the tops caused the"bumpy" ( Keeley, ibid.) polish with some amorphous s ilica, I believe, making the polish l ook somewhat brighter than dry hide polish. -

-

-

-

Dry l eather c aused very l ittle dry and a lso s ofter than dry hide i n water caused a very brilliant,

.

polish, because i t was However, l eather soaked f luid l ooking polish,

t otally different to the other polishes polish resembles wood polish ( Pl.29:c).

f rom hide.

Hide scraped with the addition of ochre led greater abrasion o f the scraping edge ( Pl.10:e).

t o

This

f ar

There were no differences in polishes according to different a ctions, except when dry hide was rubbed. The polish was not typically"pitted"lookjng, a s one would have expected f rom dry h ide. Not enough the e ffect o f Kamminga ( 1978,

experiments were carried out t o investigate h ides f rom different species of animals. p .137) mentioned that marsupial skins caused

7 1

wear-traces which were skins of other animals.

d ifferent

Striations s eemed rare when diffuse shallow l inear f eatures be seen on h ide s crapers. S imilarity

t o

other

t o

hide was ( Keeley,

those

reported

f rom

cut, but l ots of ibid., p .50) could

polishes:

Anderson-Gerfaud ( 1981, Vol.1, p .53) combined h ide and meat polishes, while Vaughan ( 1981, p .160) combined f resh hide and meat polishes. However, I agree with Keeley ( op. c it.) and with Moss ( 1983, p .86) that the two are f airly distinct. L ike Moss ( ibid.) I found a s imilarity between the reticular distribution of dry hide polish and t hat of wood ( or in my case dry reed) I a lso would have mistaken the polish f rom soaked l eather for f resh sycamore polish. I did confuse hide and antler polishes at t imes. .

5 .

Antler

Twenty-five experiments had been carried out s awing, scraping, boring, drilling, graving and rubbing red deer, reindeer and f allow deer antler, dry and soaked for three days in water. Unfortunately t he faunal evidence, f or red deer at Arjoune, was only completed a week before this thesis. I carried out two experiments with red deer antler and the resulting wear-traces were s imilar to those from fallow deer antler.

f rom

In general, the both species

distribution of soaked antler was quite wide and reticular,

polishes very much

l ike that o f wood polish, probably because both wood soaked antler were in my experiments s imilarly yielding.

and

Contrary to Keeley's a ssumption ( 1980, p .56) the polish from reindeer antler and from f allow deer antler l ooked somewhat different: contact with f allow deer antler caused the f lint surface to be quite f lattened, and in a ll c ases widely spaced f ine striations were visible ( Pl.9: g) while reindeer antler which i s perhaps l ess dense, did not f latten the f lint surface considerably, but caused a "domed"look ( Vaughan, 1 981, p .144), esperially when antler had been scraped or planed, and striations were not a lways s een ( however, s ee P l.11:c). The species

polishes from dry antler l ooked matt and reticular

( Pl.15:h) of l ike dry bone

e ither polish,

while the polishes from soaked antler o f e ither species invariably became more l inked, bright and more undulating, probably because the antler had become more yielding and more s ilica had become amorphous due to the increase of moisture in the antler.

tool

There were some action: a s a ll

variations researchers

7 2

in had

polishes according to f ound ( Keeley, op. c it.

‚ p.56; Vaughan, 1 981, p . 1 45; Moss, 1 983, p .87) sawing reindeer antler c aused a "rough and pitted"(Keeley, op. c it., p .56) polish. ( I did not f ind such pitting, possibly because I had not c leaned my tools in chemicals such a s HC1.) I did notice a l ot o f striations on antler sawing t ools, e specially on those used t o saw fallow deer ( Pl.15: g). Transverse a ctions, such a s scraping or planing antler, t ended t o cause polish"bevels"on the used f lint edges. Rubbing caused an a ltogether different polish which could be identified on Blind Test streaky distribution and a somewhat with some amorphous s ilica. S imilarity

t o

other

tool 1 2 by virtue of its f lattened f lint surface

polishes:

I f ound ( as had Keeley, 1 977, p .55; Vaughan, 1 981, p .144) that " smooth antler"polish from transverse actions on antler could be confused with wood polish ( Pl.9: g and 8 :e). I found ( as had Vaughan, 1 981, p .135) that polish from antler sawing could look indistinguishable from that of bone sawing, and a lso that the bevels of polish from scraping antler and bone could look s imilar ( Pl.9: g and h ) ( Vaughan, op. cit., p .145). Although Keeley appeared to doubt this ( op. cit., p .56), I thought t hat undeveloped antler polish could look l ike dry hide polish. Pressure f laking with antler caused a polish which could be antler use-wear polish 6 .

confused ( Pl.6:f).

with

undeveloped

hide

or

Horn

Four experiments were graving and boring cow horn

carried out sawing, s craping, soaked in water f or f ive hours.

Although the f lint edges were blunt, very l ittle polish developed. Only a f ew areas of relatively f lat polish could be seen, together with a few short tracks ( Pl.16:d) polish would be impossible to detect on archaeological with post-depositional surface modifications ( a term used by Holmes, 1 986). .

S imilarity

to

other

polishes:

The polish from s craping horn hide or antler polish. 7 .

This tools f irst

could be

confused

with

weak

Meat

Eleven experiments were carried out cutting meat ( cooked and uncooked) beef, pork, chicken and lamb, and scraping meat off deer hide. In addition, projectiles ( points and transverse arrowheads) and barbs which Bergman, and Newcomer, had shot into meat were examined. Meat cutting tools used by Miller were a lso examined. ,

Polish from animal species

the contact with a ll or whether cooked or

7 3

meat, f resh,

regardless of looked s imilar:

t he f lint surfaces were only s triations and very l ittle ( Pl.14:a) by other p .93;

.

gently polished, some f ine amorphous s ilica were present

This observation s eemed t o agree with those researchers ( Keeley, 1 980, pp.53-54; Moss,

Vaughan,

1 981,

made 1 983,

pp.160-162).

Variations i n polishes appeared t o depend on whether bones or t endons were cut a s well a s meat ( Pl.14:c), or whether a wooden cutting board had been touched by the tool: bones and t endons ( Pl.16:a) s eemed t o cause some bright striations, while touching the wooden board l eft i solated patches o f wood polish a t the tool edge. S craping meat off the f lint blade

of

while

hide

polish

hide caused meat polish on the a spect which had had contact with t he meat,

could be

seen

on

the

other

s ide.

Most of the arrowheads shot into a deer carcass showed what Moss ( 1983, p .95) had called"microscopic l inear impact t races ( MLIT)"(P1.20:e, g). She found ( as did I ) that i n most cases the t arget could not be determined f rom such t races. Apart from these traces found a long t he impact axis, I noted that l inear polish and striations were a lso located at other angles on projectiles than those of the impact. Such traces presumably originated when the projectile was deflected inside the carcass, e .g. by bone, or e lse when it was pulled out of the t arget. S imilarity

t o

other

polishes:

As Vaughan had pointed out ( 1981, p .162), meat polish i s weak and could be disguised by general post-depositional t races ( which he referred to as"soil sheen"). On coarsegrained f lint I sometimes could not detect meat polish at a ll. Vaughan ( ibid., p .161) and Anderson-Gerfaud ( 1981, Vol.1, p .53) could not f ind any difference between fresh hide and meat polish. However, I agree with Moss ( op. cit., p .93) that on experimental tools the two polishes looked quite different. 8 .

Fish

Five experiments-were carried s caling bream, gurnet and trout.

out

gutting,

cutting

and

When f ish was gutted, hardly any polish was visible; when f ish was cut, the resulting polish consisted o f bright l ine at the f lint edge, the f lint microtopography l ooked quite f lattened, and some f ine striations were visible. When f ish was s caled, in some cases a very distinct pattern of polish developed. It consisted o f wide polish streaks running obliquely from the f lint edge, presumably where the f ish scales had come into contact with the f lint ( Pl.15:e and 2 9:d). A s imilar pattern of"cross-hatching"had been reported by Moss ( 1983, p .105) However, in some instances, .

when

f lint

was

coarse-grained

7 4

or

in

B lind

Test

1-10

( although could

be

the

S imilarity I could

f ish

s caled

had

hard

s cales)

no

such

pattern

e ither

the

polish

,

s een. t o

not

other

f ind

polishes:

a ny

s imilar

from cutting f ish, However, the frequent

polishes

nor to absence

t o

t he polish from scaling f ish. o f the s treaky polish pattern

from f ish s caling indicates conclude f rom i ts absence that

caution, and one should not f ish was not the worked

material. 9 .

Shell

Twelve experiments were carried out cutting, s craping, drilling, boring incising and rubbing c ardium, s callop shell and ormar shell. ‚

shell

The resulting polishes looked s imilar, regardless o f variety and tool action: shell polish affected only

i solated areas o f the f lint surface. The high points of the f lint microtopography were quite f lattened, more so than from contact with bone, less so than f rom contact with stone. Characteristic of

f ine,

narrowly

This ( 1983,

p .

S imilarity

polish

and

a lways

spaced looked

1 04,

p1.

t o

other

6 :8a)

present were

striations exactly

parallel

bundles

( Pl.9:e)

l ike

that

and by Yerkes

shown

by

Moss

( 1983).

polishes:

I f ound shell polish very characteristic on a ccount o f i ts narrowly spaced s triations, a lthough i t could possibly be confused with stone, bone or antler polishes, i f these caused 10.

s imilarly

spaced

striations.

Stone

Fourty-one experiments were carried out cutting, scraping, polishing, graving, drilling the following varieties of s tone: l imestone, quartzite, schist, metamorphic rock, lapis lazuli, malachite and f lint. In addition, t races from the manufacture o f t ools with quartzite or l imestone hammers were observed. The general polishes consisted of a complete f lattening o f the t ool surface resulting in very bright, wide streaks of polish and bundles o f striations ( Pl.9:c, 1 2:e, 7 :e) s ee a lso Moss ( 1983, p . 1 03, p1. 6 :8c) Some amorphous s ilica was visible, probably causing the hammerstone"smears" shown ,

.

by

Keeley

( 1980,

P 1.8)

The polishes according to tool

( Pl.6:a, b).

did not show any particular action, only a ccording t o s tone

7 5

variance type: t he

polish l ooked more d iffuse ( i.e. the t ool s urface was f inely scratched rather than completely f lattened) f rom contact with softer stones, such a s l imestone or malachite.

from

Drilling both l apis lazuli and malachite left residues these materials on the tools even a fter c leaning.

S imilarity

t o

other

polishes:

Polishes f rom soft stones, such a s malachite, could be confused w ith polish f rom sand and pottery, probably due to the sand-like consistency o f the soft s tones. In one instance ( Blind Test 1 -10) the traces f rom scraping stone l eft on a s craper made of medium to coarse-grained French f lint were confused by me with antler polish. This might have been due t o the f act t hat at the t ime I was not f amiliar with s tone s crapers, nor with wear traces on that particular f lint. Hammerstone t races were usually distinct. However, t hey might well be the dominant type of wear t race on such t ools especially t o

11.

a s the

scrapers, thus potentially misleading, inexperienced microwear analyst.

Pottery

Eight experiments were carried out cutting, boring, graving and drilling pot sherds from Arjoune, Trench VI and from Abu Salabikh. Only the tops o f the f lint microtopography were f lattened and the polish looked bright, but diffuse and f inely s cratched, with lots of striations ( Pl.11:g). Residues f rom th e pottery could still be seen adhering to the f lint surface after c leaning. S imilarity

t o

other

polishes:

I found the polish indistinguishable f rom sand polish ( Pl.7:a), and therefore potentially indistinguishable f rom natural sand polishes, except for the residues f rom the pottery ( Pl.11:g) S and did not seem t o l eave such residues. The polish a lso looked s imilar to polishes f rom soft stone. .

12.

Sand

S ix experiments were carried out to show wear traces from sand: f lint was rubbed against sand, drill bits were used on various materials: copper, shell and lapis lazuli, with the addition of sand, and sand t ogether with water. In addition, examined.

a projectile

shot

into

earth

by

Newcomer

was

Sand polish from a ll these activities ( except drilling) looked s imilar: the s urfaces of t he f lint t ools were diffusely polished with scratches and l ots of f ine striations running in t he direction o f use ( Pl.7:a) Drilling with

the

addition

7 6

of

s and

however

c aused

an

even ( but not smooth) polish without any visible s triations which t otally obliterated the wear t races f rom the worked material i tself. A s imilar phenomenon was described by Vaughan ( 1981, p .163). Drilling with t he addition of sand a nd water a lso obliterated t he t races f rom the worked materials, but caused a d ifferent polish f rom that o f drilling with the addition o f sand a lone: when water was added, the wear t races consisted o f a completely smooth tool t ip with a concentric ring pattern ( Pl.29: g) A s imilar phenomenon had been reported by Gwinnet and Gorelick ( 1979) who had examined drill casts made by wooden drills with the addition o f sand under the S EM. .

-

-

S imilarity

t o

other

polishes:

In Blind Test 1 -10 I confused shot into earth with bone

polish on a polish, on

striations. As confused with

above, working

pointed out polish from

sand both

pottery. S triations in sand polish due to would presumably be randomly orientated i s an area for potential mistakes. 13.

polish can s oft stone

be and

natural however,

agencies t his too

of a pigeon polish a long

feather the edge

Feather

One experiment cutting l eft a thin, indistinct, ( Pl. 15: f 14.

projectile point a ccount o f the

the quill band of

Wool

One experiment was carried out cutting wool f rom a Shetland sheep. The sheep's wool might be different to that of the ancient Near Eastern sheep, as the ancient breeds had long, coarse hair with a downy undercoat, unlike the thick-fleeced breeds of today ( Broudy, 1 979) It i s unlikely that the ancient breeds who moulted ( Ryder, 1 983) were shorn with f lints, but t he spun f ibers may have been cut. .

just f ine

The wear traces consisted o f a brilliant l ine of polish on the edge of the f lint blade, t ogether with lots of striations i n the cutting direction ( Pl.14:e). This

15.

polish

l ooked

s imilar

t o

s inew polish

( Pl.16: b).

Hippopotamus"Ivory"

One tooth.

experiment

was

carried

out

drilling

hippopotamus

The resulting polish was very distinct: i t had f lattened the f lint t ip gently while very small pits remained v isible ( Pl.29: h) I did not s ee any s imilarity between t his polish and any other polishes. .

7 7

1 6.

Copper In

two

experiments

copper

was

drilled.

The resulting polish ( Pl.29:e) was very distinct: t he drill t ip was completely and evenly f lattened. S ome copper residues were seen adhering t o t he t ool surface. I d id polishes. 1 7.

not

s ee

any

s imilarity

between

this

and

other

Plant

One Hundred and e ighty-seven experimental f lint edges were used to cut cereal and non-cereal plant species. ( The complete l ist o f these plants i s given in P art I I, Chapter 1 7.4 . ) The plants were cut in some instances both fresh and dried. Reeds were a lso scraped and c ane was scraped and bored. I n two experiments root vegetables ( carrots) were cut. Most plants caused macroscopic g loss • as soon a s polish was visible under the microscope, with the exception o f occasions when very coarse grained f lint blades had been used. On t hese gloss was visible somewhat later than microscopic polish. ( For a detailed account o f the variables involved in Chapter 1 6 . )

the

f ormation

of

plant

polish

see

Part

I ,

In general, contact with plants caused a gentle polishing o f the f lint surface which l eft the tops of the f lint microtopography rounded. The polishes l ooked buoyant, probably due to the amorphous s ilica l iberated by the moisture f rom the plants ( see P 1.22-25). Contact with f resh plants caused a greater amount brilliant polish, developing at a much faster rate, than did contact with dried plants ( Pl.23:c,e). There were considerable variations in polish f ormation a ccording t o plant species ( see Table 2 ) . This agreed with observations of Anderson-Gerfaud ( 1983) . Hard-stemmed plants such as reeds and e inkorn c aused a reticular polish due to t he f act that the unyielding s tems could only polish the very tops o f the f lint microtopography ( Pl.23: c and P 1.22:a) . However, cane, perhaps because it was very hard, polished the entire f lint edge ( Pl.24:c), while very soft plants,such a s bulrushes, horsetails and grass, a lso polished the entire f lint surface ( Pl.23:a, 2 5:e,c). Thick-stemmed plants, such a s reeds and bulrushes caused f lint surfaces t o be widely and relatively evenly polished ( P1.23:c,a), while thin-stemmed p lants, such a s grass, caused only a narrow band o f polish on the very edge of the f lint ( Pl.25:c) . Stems o f medium width, such a s the stems o f c ereals and of cane ( as cane s tems, a lthough thick, are hollow) c aused both a concentrated band of polish on the ,

7 8

v ery edge

edge o f the f lint and some ( Pl.22:a,c,e,g, P l.24:c).

weaker

polish

Polishes f rom plants i n t heir wild and f orm had a s imilar distribution ( Pl.22: b, h).

away

f rom

the

domesticated

The occurrence of s triations, comet-shaped p its ( Witthoft, 1 967; D iamond, 1 979) and o f micro-pitting l ooking l ike the e ffect of sandblasting, s eemed related not t o the s tructure o f the p lant, but to the presence of loose soil during harvest ( see Part I , Chapter 9 ) Korobkova was the f irst to suggest that s triations i n plant polishes a re due t o abrasion from t he soil rather than from s ilica i n the p lants ( 1981, p .331). .

In my experiments, striations were absent in polishes f rom cutting plants growing in water ( in England a s well as i n Syria) such a s reeds, bulrushes or horsetails ( Pl.23:c, a, 2 5:e). A f ew s triations were usually present on blades used to cut plants growing on a grassy soilcover ,

( Pl.22:a), while many more striations were present on polishes f rom plants cut at the base from upturned soil in England ( Pl.22:e,f) P lenty of striations are l ikely to o ccur when plants are cut from upturned loose soil in the Near East. .

S craping reeds c aused e ssentially p lants.

( Pl.23:g, h) and scraping and boring cane the same polishes a s did cutting these

One water plant species, Sparqanium ramosum cut both i n England and in Syria, caused very unusual wear t races in t he form of a cracked l ooking deposit ( or polish?) ( Pl.24:e,f) which not even twelve hours immersion i n White S pirit known to be a powerful organic solvent could r emove. Very l ittle i s known about this plant, and according t o Dr. John Bryant, Reader in P lant Physiology at Univ. College, Cardiff, ( pers. comm.) there may be some p lant ‚

‚

waxes

which

‚

do

not

dissolve

i n White

Spirit.

The polish f rom cutting carrots for approximately 1 5 m inutes ( Pl.16:c) was indistinct. No macroscopic gloss was v isible on the f lint blades. S imilarity

to

other

polishes:

The divisions between plant polishes were by no means c lear cut. A large sample of archaeological f lint tools would need t o be examined before attempts can be made to i dentify d ifferent plant polishes. Reed polish could be s imilar to fresh wood polish, but i t i s a lmost never striated, and the gloss from reeds t ends t o be much stronger. However, dry plant polishes could well be confused with wood polishes.

7 9

The presence o f s trong macroscopic g loss makes most plant polishes easily distinguishable. Other materials which c aused g loss caused polishes which l ooked different t o p lant polishes. 1 8.

have

Rafting and

Prehension Traces

Hafting traces i n t he form of polish often been reported, an example

and being

s triations, the t races

Anderson-Gerfaud s aw a s hafting t races on Mousterian implements ( 1981, Vol.1, p .41) Keeley identified t he raw material from which a lleged hafting t races originated, i .e. of the haft itself ( 1982, f igs.1 and 2 ) Cahen and Gysels ( 1983, p .44) reported hafting evidence in the form o f hide and bone or antler polish on unused a reas of Neolithic t ools f rom Blicquy. Bueller ( 1983, pp.114-124) defined on a lmost .

.

a ll the Natufian tools which he had studied microscopically t he exact material of the hafts. In none of these examples did the authors report any hafting experiments nor did t hey refer to any possible post-depositional surface modifications ( see Holmes, in press) of t he ancient f lint tools they had studied. P lisson performed 2 0 hafting experiments ( 1981) and f ound t hat traces were so i solated that on ancient tools t hey could be confused with post-depositional traces. Only when ochre had been mixed with the hafting agent and the t ool loose had been loose in the haft were hafting t races present. Moss ( Moss and Newcomer, 1 981; Moss, 1 983, pp.101-102) a lso found that t races f rom f irmly f ixed hafts were rare and minimal. Odell ( 1978, p .42) observed t hat f riction in a loose haft a lso caused not only polish, but a lso diagnostic edge damage. However, a s Keeley h imself s tated ( 1979, p .681) a properly hafted tool should l oose, and in my opinion would be immediately because of its reduced e fficiency.

not he rehafted

Some authors ( Semenov, 1 964, p .115 ; M .-C. Cauvin, 1 973) referred to the distribution of polish on the f lint, and to a "sharp demarcation between the polished and mat surfaces.. . "(Semenov, op. c it.) a s evidence f or hafting. Prehension traces ( i.e. traces from holding an implement) were reported i n the form of edge damage ( Odell and Odell-Vereecken, 1 981) or in the form o f polish and striations ( Semenov, op. c it., pp.14 and 1 07; Keeley, 1 982, p .807; Vaughan, 1 981, pp.164-165). My own experiments showed that I could not detect any hafting t races on end s crapers s ecurely hafted with r esin, wax and s inew, nor on those s imply wedged into a wooden haft. Nor were any such t races visible with the exception o f a f ew striations across t he bulb of one t ool on 1 5 drills used for up t o 1 0 minutes to drill very hard materials, such as stone or copper. The drills had been hafted not only with resin, wax and s inew, but a lso w ith the -

-

8 0

a ddition o f ochre in some cases. Neither could I detect any hafting t races on ancient drills which on account o f their s mall s ize and concentrically polished t ips must have been u sed in a haft. Had there been any minute t races, such a s described by Moss i n connection with f irmly s ecured hafts, t hey would have been obliterated by the postdepositjonal t races found a ll over these tools. On t he tools of Blind Test 1 -10 a ll manner of traces were observed on the areas which were obviously not used. However, t here was no consistency in the nature of these t races a ccording t o whether they had been used hafted or handheld. This indicated to me ( as Moss had observed in c onnection with t races from refitting ( op. c it., p .102) that i t might be fruitless to a ttempt to i solate prehension t races on ancient tools. I did f ind could indicate B lind Test 1 -10 s traight l ine

that the distribution of the use-wear polish hafting; for example, on a blade used in t o cut meat, polish formed a completely a long the entire edge o f the tool. This

i ndicated

( correctly)

to me

that

the

tool

had been hafted.

I

f ound the same phenomenon, and in some cases a sharp demarcation between polished and unpolished areas on experimental s ickle blades which I had used hafted. However, I do agree with Moss ( ibid., p .102) that"... hafting wear t races are inconsistent.. " and I found i t t herefore hard to believe ( in the absence of any evidence) t hat the researchers quoted at the beginning of this section could not only t ell whether tools had been hafted, but a lso t he precise material from which the haft had been made. I a lso agree with Moss ( ibid.) that other"more consistent f actors"should be l ooked for a s evidence for hafting, such a s the t ool shape and/or the frequency of resharpening of a .

t ool, and f inally the recovery a lthough a s Moss herself realised rarely be preserved. -

19.

of -

the hafts themselves, wooden hafts would

Conclusions

I f ound that in practice experimental polishes differed a ccording to hardness, width, structure and water content of t he worked material, and that therefore it i s l ikely that such differences can be recognised. However, I a lso found that quite often such attributes on one worked material could be s imilar to those on a t otally different worked material, and that therefore polishes from two very d ifferent materials could look a like. On the whole, my observations agreed with those of the other r esearchers. Where they did not, t he differences could be explained at l east in theory. -

8 1

Chapter

15

Appearance

-

The Variables Affecting the Development and

of

Polish

-

the

Outline

Despite the f act that microwear analysts had noticed the e ffects o f variables other than t he worked material on the rate o f development and appearance o f polish ( Anderson, 1 980, p .181; Vaughan, 1 981, p .100 ; Moss, 1983, p.55) there had been no systematic series o f experiments a imed at discovering which variables are involved and t o what extent they are involved in polish formation. The only exceptions had been Vaughan, who had experimented with three different f lint types ( op. c it., pp.127-132) and Gysels and Cahen ( 1982) who had investigated the e ffect o f hydration on a variety of materials. However, both these i nvestigations came to my attention a fter I had conducted most o f the experiments described in this chapter. I ran a preliminary controlled s eries o f experiments concentrating on plants a s a worked material, in order t o investigate how the rate and the appearance o f polish were a ffected by the following variables: plant species; the water content of the plant; the point at which the s tem was cut; the t ime o f year o f harvest; the number of stems cut with one movement; the s ource of f lint; the f orce applied by different workers; the resistance or hardness of the stems; the e ffect o f different environments on t he same plant species. Other variables which were not t ested at kept a s constant a s possible: the direction movements; the dimensions, shape and angle o f the the

blades;

the

experimental

t emperature,

humidity

the t ime were o f the cutting the edge o f

and windforce during

work.

After this preliminary investigation concentrating on plants, I carried out further series o f experiments i n order to: -

see

whether

the

same

variables

a ffected

polish

formation

due t o contact with materials other than plants. This I did in view o f the fact that Del Bene ( 1979; Holley and Del Bene, 1 981) had argued that materials such a s plants c aused additive polishes, while other materials caused abrasional polishes. -

investigate

their e ffect sensitivity.

more

fully

with

some

the

test some variables which preliminary experiments. -

help

variables, o f

had not

8 2

and

demonstrate

measurements

yet

been

t ested

o f

l ight

in

the

Chapter

1 6

The

-

Preliminary

Series

of

Experiments

with

Plants 1 .

Method

The b lades were s truck with an antler hammer, caught by hand, and examined under the microscope before u se. Most blades were made of vitreous-looking black f lint f rom Brandon, o thers o f opaque-looking grey f lint from Potter's Bar in London, others of coarseand very coarse-grained f lint f rom Arjoune in Syria. S ixty-six experimental blades were used between May 1 982 and May 1 983 in South England, Wales and Syria. The following plant species were cut f resh and dried, in some instances at intervals o f four t o s ix weeks: reed ( Phraqmites communis Trin.), bulrush S choenoplectus lacustris L . ( Palla) grass ( several species i ncluding the Sweet Vernal Grass ( Anthoxanthum odoratum L .)), e inkorn ( Triticum monococcu i n L . emend. Lam.) and domesticated barley ( Hordeum disticu r n L . emend. Lam.). The non-cereal species were c lassified according to Clapha r n et a l. ( 1962) and the cereal species according t o S chiemann ( 1948) ‚

The blades were used a cross the plant stems

unretouched and unhafted which were bundled in one

t o cut hand. At

a ll t imes a sawing motion was used at right angles a cross t he plant ste i n. One s . m. equals one movement f orward and one backward. The sawing movements were counted and the b lades examined every 1 00 s . m. f or appearance o f gloss by gently wiping the surface of the f lint with distilled water and l etting it dry naturally. As the investigation o f the ancient"sickle"blades had shown that gloss and plant polish nearly a lways corresponded ( with the exception of very coarse-grained f lint, s ee below) it was a ssumed that whatever caused gloss a lso caused polish. Therefore the f irst appearance o f gloss was taken as a f ixed point at which to compare the polish on the blades. The following may s erve a s an example: it took 1 00 s . m. through fresh reeds in July to cause the beginning o f gloss and polish on Brandon ,

f lint, and 2 000 s . m. ( also in July) through f resh e inkorn to cause the same, therefore 1 00 s . m. in the case o f reeds, and 2 000 s . m. in the case-of e inkorn were taken a s one unit of development, and t he polishes were compared a s representing an equal s tage of development. The

method

the

tools

I I,

Chapter

2 .

was

o f

storage,

the

same

a s

c leaning that

o f

and all

investigation the

tools

( see

of P art

9 )

Results

The plant species s eemed to a ffect the rate o f development of polish: i t took approximately 5 0-100 s . m. through reeds, 2 000 s . m. through bulrushes, 6 00 s . m. t hrough grass, 2 000 s . m. through e inkorn, and 4 000 s . m. through domesticated

barley,

before

gloss

8 3

and

polish

could be

s een.

There were a lso considerable o f polish, e specially when the

differences i n cutting edges

the o f

appearance t he f lint

tools were compared a t 2 00x; in particular t he degree o f smoothness varied. Grass caused t he smoothest polish, f ollowed by bulrushes, barley, reeds and e inkorn which caused the most reticular polish ( Pl.25:c, 2 3:a, 2 2:c, 2 3:c, 2 2:h). The distribution varied. The polishes f rom reeds and e inkorn were more evenly distributed over t he f lint surface than those from the other plants, which were concentrated on the edge. After continued experimental work the differences seemed to increase, t he grass polish becoming increasingly smooth, while reed polish remained quite reticular though showing a continuous edge, while e inkorn polish, despite developing s ome reflective r idges, retained a discontinuous edge. The water content of the plant seemed t o a ffect the rate o f development of the polish: i t took 2 000 s . m. on reeds stored for one month, 3 000 s . m. on reeds s tored f or two months, 6 000 s . m. on bulrushes stored for one month, 1 200 s . m. on grass stored for one month, and 8 000 s . m. on e inkorn stored for f our days, before gloss and polish could be seen. In the case of reeds this increased the number of s . m.s necessary for gloss and polish twenty t imes after one month and thirty t imes a fter two months, while in the case of bulrushes i t increased three t imes after one month, in t he case o f grass twice a fter one month, and in the case of e inkorn f our t imes after only four days. The polishes also l ooked d ifferent f rom those caused by fresh plants: whereas polishes from f resh plants looked buoyant and smooth, those f rom the same plant species dried looked relatively f lat and l ess smooth. Hardly any increase in polish seemed to t ake place a fter subsequent use. It was impossible t o distinguish between polishes from different species of dried plants. S triations were seen only once, on dried reed polish. I t did not seem to make any difference to e ither development or appearance of polish whether reeds at the base or at higher points of the stem. This was not tested on other plants.

rate of were cut variable

The t ime of year of harvest seemed t o a ffect the rate of polish formation but not its appearance. Experiments were carried out harvesting reeds at intervals of s ix weeks and in May the polish was very strong, in July i t was weaker, in S eptember only faint at 1 00 s . m., and in November 2 00 s . m. were needed to form gloss and polish. This variation seemed related to the water content of the plant. The

number

o f

stems

cut

with

one

movement

( i.e.

a

s ingle

stem or a bundle of f ive stems) did not a ffect the rate of polish formation or appearance ( 18 experiments with reeds only). The and

source of a lthough

f lint a ffected the polish appeared

8 4

rate of polish on Brandon f lint

f ormation a fter 2 000

S .M. through e inkorn, i t did not on the s lightly coarsergrained Potter's Bar f lint. However, a fter prolonged use ( 9000 S . M.) on grass the Potter's Bar f lint showed the same amount of polish as t he Brandon f lint. Brandon blades used t o cut reeds in November developed polish and gloss a fter 2 00 S . M.. However, coarse-grained f lint and very c oarse-grained S yrian f lints used on the same reeds developed gloss only a fter 2 500 S . M. and 5 000 S . M. r espectively. Microscopic polish had developed earlier than g loss, but had been difficult t o spot as it looked unconnected, a ffecting only a f ew high points of the f lint microtopography only. I t therefore s eemed that the appearance of polish was influenced by the topography o f the f lint surface; even on the f ine-grained f lints polish i s most c learly evident a long the ridges of f eather f ractures. The force applied by any difference t o reeds only)

different workers polish formation

did not seem to make ( 18 experiments with

The resistance or hardness of the stem was only superficially judged. It did not bear any relation to the rate of polish formation. The s tems are l isted in order of hardness and are compared to the approximate amount of S . M. needed for the appearance of gloss and polish: e inkorn ( hardest) 2 000 S . M., reeds 5 0-100 S . M., barley 4 000 S . M., bulrushes 2 000 S . M., grass 6 00 s .x n.. However, there seemed t o be a relation to the appearance o f polish and the s equence o f plants graded a ccording to t heir hardness corresponds to the sequence graded according to roughness ( reticularity) o f polish. Different environments did not s eem to a ffect polish f ormation and plants of the same species ( reeds, and another plant, Sparganium ramosum which was not used in the other experiments) caused s imilar rates of development and identical wear t races, regardless of whether they were harvested in Eng' and, Wales or Syria. ‚

3 .

Conclusions

The rate of polish formation s eems to be governed most importantly by t he plant species, by the water content and by the type of f lint from which the tool was made. The appearance o f plant polish seems governed by the plant species, perhaps because of the varying hardness and width o f the stems, by t he water content of the plant and by the t opography of the f lint surface. The rate o f polish development and i ts appearance do not seem noticeably influenced by the point at which the stem i s cut, by the number of s tems cut at a t ime, by the strength of the worker or by different environmental conditions.

8 5

The absence o f gloss visible t o the naked eye on coarse grained f lint does not necessarily mean t hat plant polish will not be v isible under the microscope on the same f lint. I t seems t hat water f rom the plant ( and/or substances contained in i t) acts a s a medium for polish f ormation. This conclusion i s based on the f act that drying the p lants causes a drastic reduction in the rate of polish f ormation. This reduction cannot s imply be explained by l oss of suppleness of plant t issue and therefore l ess contact with t he f lint surface, a s was suggested by Masson e t a l. ( 1981) This i s evident f rom the f act that the reduction in the rate o f polish formation was four t imes in einkorn a fter only four days of storage but during this t ime the s tems had lost hardly any suppleness. .

8 6

Chapter 1 7 Experiments 1 .

Variables

-

of

Polish

Formation

Further

-

Flint

As the preliminary investigation had shown t hat both t he rate of development and the appearance o f plant polish were a ffected by t he f lint types I wanted to s ee whether the s ame variables applied when materials other than p lants were worked, in order t o f ind out how important i t i s t o use experimental replicas made o f the same raw material a s the a rchaeological t ools. Other researchers had reported widely d iffering results: Keeley ( 1980, p .16) f ound that different f lint types did not cause different polishes, while Vaughan ( 1981, pp.131-132) f rom his experiments with three f lint t ypes of varying grain s izes concluded that polish f ormation was s imilar in rate o f development and appearance, a lthough s omewhat s lower and l ess connectedlooking on f lint of l arger grain s ize, and that therefore different f lint types would not pose any problem for the microwear analyst. Holmes ( 1986) demonstrated the differing rates o f development of English chalk f lint and of Egyptian l imestone chert in paired experiments, but a lso concluded that polishes on the two stone types l ooked s imilar, when well developed. Moss ( 1983, p .55) s imply stated that the raw material o f t he t ool had to be replicated. Experiments I used f ive f lint types: Brandon, S yrian"Type A ", S yrian"Type B ", Syrian"Type C ", and Syrian"Type D ". O f the S yrian f lints two types had been f ound on t he surface o f t he s ite of Arjoune, while two types had been knapped f rom cores excavated f rom Trench V . They ranged f rom f ine t o very c oarse-grained. Ten experiments were conducted on bone and wood, and comparisons were made after u sed for the same number of strokes.

each

f lint

had

been

Differentiation between f lint types was made a ccording t o the f ollowing criteria: grain s ize, surface t exture, opaqueness and colour. Of these attributes grain s ize and surface t exture turned out to be interrelated and except f or inclusions t hey were found to be consistent throughout each f lint piece. Opaqueness was found to be unrelated to grain s ize/surface t exture and not consistent t hroughout e ach piece. Colour was found to be unrelated t o grain s ize/surface t exture. The last two attributes were t herefore d isregarded. -

-

I thought i t s implest t o define grain s ize/surface t exture i n relation t o the focus under magnifications of 2 00x to 5 0x, a s t he depth of f ield increases with decreasing magnifications. Therefore f lint easily f ocussed a t 2 00x ( Brandon and Syrian"Type D ") was defined a s f ine-grained; f lint easily f ocussed at l OOx ( Syrian " Type C ") a s medium-grained; f lint easily f ocussed at 5 0x ( Syrian"Type

8 7

A ") a s coarse-grained; and focussed at 5 0x ( Syrian"Type ( Pl.5:a,c,e, g).

f lint which B ") a s very

can hardly b e coarse-grained

Results The results from the experiments sawing bone up t o 8 00 S .M. and s craping wood up to 3 200 S .M. ( Pl.5:b,d,f, h) showed the following: The rate o f polish development seemed at f irst s lower on t he coarse grained f lints. However, c loser inspection r evealed that the polish in i ts initial stages was very difficult t o detect on coarse-grained f lint, a s the uneven t exture o f such f lint l ed to only i solated a reas of polish. L ight s ensitivity measurements of polishes on coarse-grained f lints ( Fig.2) showed that the rate of polish development was much s lower on these. This was f ound with all materials worked, regardless of whether they were plants, wood or bone. The appearance of polish was found t o differ according t o the f lint type: the relative f ineness or coarseness o f the f lint was nearly a lways found to correspond t o the d egree o f l inking o f the polish. The exception was the medium-grained Syrian"Type C "which was patinated: polish on this f lint was a t least a s developed and connected than the polish f rom equivalent use on the f iner-grained Brandon f lint. The o ther exception was t he f ine-grained Syrian"Type D " which developed very unconnected polish. This was probably due t o i ts very i rregular edge, again a consequence o f patination. Conclusions The evidence from my experiments with different f lint t ypes supported Vaughan's conclusions ( 1981, pp.130-132) t hat both the rate of development and the appearance o f polish vary with the grain s ize o f the f lint. However I did f ind exceptions to this rule, when f lint was patinated. S eitzer Olaussen ( 1983) and Bradley and Clayton ( 1986) a lso reported different rates of polish development on t hermally a ltered and recrystallised l int. I would therefore agree with

Moss

( 1983,

p .55)

that

the

raw

material

o f

the

archaeological tools should be replicated, in my opinion a t l east until such t ime that the microwear analyst i s f amiliar with the polishes on the f lint f rom his s ite. The f lint both the

grain size proved to be an important development o f polishes f rom contact

variable i n with plants,

and of that from contact with other materials, such a s bone. I t herefore saw no evidence in f avour of the hypothesis that polish from plants and wood i s depositional while polish f rom other materials i s attritional ( Del Bene, 1 979; Holley and Del Bene, 1 981)

8 8

As had been the c ase with plants, the f aint g loss visible on woodworking implements made of f ine-grained f lint was not s een on coarse-grained implements used on the same wood for t he same number o f strokes, a lthough wood polish was visible a t a microscopic l evel. 2 .

Duration of Work

While most researchers reported that at the very beginning o f their development polishes from different worked materials c ould look indistinguishable, only Vaughan investigated the e ffect of t ime on polish f ormation more fully. He found three developmental stages:"generjc weak polish"(1981, p .133 and pp.137-138), then"smooth-pitted polish"(ibid., pp.135 and 1 39), followed by fully developed polish. Although he referred to the initial"generic weak polish"as"dull"(ibid., p .138), he did not discuss i n polish reflectivity related to the duration of

a change work, but

only a s the f irst amongst the various attributes by which"i nicropolishes induced by certain c lasses of worked materials may be d istinguished in most instances.. . "(ibid., p .129). Anderson-Gerfaud had made a s imilar statement ( 1981, Vol.1, p .35). In fact, reflectivity, brightness, or ( the t erm I prefer) intensity of polish, was generally s een a s indicating primarily the worked material. Keeley ( 1980) used t he t erm"brightness"in his description o f polishes of precise worked materials, e . g. he described wood polish a s"very bright"(ibid., p .35) and dry hide polish a s"dull"(ibid., p.49). He did not seem to see brightness of polish a s related to t he duration of work: he left out the t ime factor a s a variable in his"Light Reflectivity Measurements"of polishes ( ibid., p .62). Moss ( 1983, p .80), however, f ound that such attribution of a distinct polish i ntensity to a precise worked material only hindered her identification of polishes. Intensity o t polish was a lso seen a s an indicator of the water content of the worked material ( Anderson, 1 980, p .181; Gysels and Cahen, 1 982; l inger-Hamilton, 1 983) Gysels and Cahen ( op. c it., p .224), in their paper ( which appeared a fter I had begun my experimental work) did refer to a variation o f polish intensity related to the duration o f work. .

Observations

the

I wanted to f ind out to what extent polish was influenced by the duration

the intensity of work:

of

Fifteen experimental tools of different grain s ize were used to work plants, wood and bone, and were examined before use and at regular intervals during use. The intensity of the polishes was measured using the exposure meter attached to the microscope.

8 9

Measurements were t aken at 5 0x ( this measurement would reflect primarily the distribution o f the polish) ; a t l OOx ( this measurement would reflect both the d istribution and the intensity of t he polish) and a t 2 00x ( this measurement would reflect primarily the i ntensity o f the polish). A mean f rom measurements on three areas of the polish was calculated f or each intensity value. The aperture o f the microscope was held open and observations. Care was taken:

constant

during

t o measure only even surfaces parallel f lint surface held a t an angle to reflective.I therefore measured only the the f lint implements; -

to ensure that the the f lint surface;

-

background was

to ensure that the f lint colour and opaqueness;

-

-

to

avoid

polish

due

t o

surfaces

agencies

to use s imilarly shaped tool uneven distribution of polish. -

this

measured

edges

of

t o the lens, a s a the l ens i s less ventral a spects of

not measured

other

s eries

a s

were

well

of

a s

even

than use; in

order

t o

avoid

The measurements o f the exposure needed ( TIME) were plotted against the number o f s trokes ( WORK) Absolute measurements o f polish intensity cannot be made, but intensity could be measured by c omparison with grades of white to black on a f ilm t est strip ( P. Dorrell, pers. .

comm.) Results ( See

Fig.2:a-f)

The duration of work a ffected the i ntensity o f the polishes f rom a ll worked materials ( see plant and wood polishes) The relative f ineness and coarseness o f the f lint a ffected the intensity o f the polish. In a ll cases the measurements of polishes on f ine-grained Brandon f lint produced s teeper curves than did those o f polishes on coarse-grained Syrian f lints"Type A"and"Type B ". The curves of t he latter show hardly any deflection ( compared to those o f the polishes on f ine-grained f lint) although they become a l ittle steeper at 2 00x. This probably reflects t he fact that polish on coarse- grained f lint was less connected compared to t hat on f inegrained f lint. The medium-grained patinated " Type C "flint had a polish the intensity o f which was s imilar to that on the f ine-grained Brandon f lint. The f ine-grained patinated f lint"Type D " showed a n increase in polish intensity l ater than did the f ine-grained Brandon f lint, and then

very

suddenly.

9 0

The c ontact materials t o s ome extent on t he

a ffected the intensity f ew measurements made.

The amount o f moisture i n the variation in the i ntensity o f

o f

the

polishes

worked material l ed t o a great a ll the polishes measured.

Conclusions Variations in polish intensities were caused by f actors: the duration o f work, the f lint grain s ize, worked material and the moisture of the worked material.

all the

The question o f whether one could work out the duration o f work f rom measurements of polish intensity on ancient f lint tools depends on t he certainty with which one could know the other variables a ffecting the intensity of the polish. These variables include the precise worked material, the moisture content o f this material, the f lint grain s ize,as well as a lso other f actors, such a s the brightness o f the unused f lint surface ( the presence of post-depositional polish would obstruct this measurement) and the edge damage and possible resharpening, e ither of which would cause the removal o f polished surfaces. I ntensity of polish a lone i s therefore not a good indicator o f the duration of use. Gysels and Cahen ( op. c it., p .224) arrived at a s imilar conclusion. 3 .

Tool Action

Although he never d iscussed this variable in polish f ormation systematically, Keeley mentioned that polishes f rom sawing antler were"rough" while polishes from s craping/planing/graving antler were"smooth"(1980, p .56). Anderson-Gerfaud ( 1981, Vol.1, pp.61-62) agreed and s tated t hat this difference was probably due to t he anisotropic s tructure o f antler. Vaughan ( 1981, p .135) pointed out that polishes f rom s awing s olid materials such a s bone, antler, wood and reeds never developed past the " smooth-pitted"stage ( obviously he used this t erm differently f rom Keeley who had called the polish f rom s awing antler " rough and pitted") and t hat they looked often indistinguishable f rom each other. Vaughan thought that the cause o f this l ay e ither in the a nisotropic structure o f these materials, or e lse in t he"dispersal of contact forces over a wider area in a s awing a ction t han i s the case with t ransverse motions and grooving, where a small part of the edge i s submitted to more intense, sustained contact than the rest of t he used a rea"(ibjd j. Vaughan ( ibid., pp.167-170) discussed the f ormation o f polishes a ccording to"Types o f Actions in I ntentional Use"(ibid., p .167). He described how the d istribution of polishes varied a ccording to tool a ction. A lthough he never explicitly stated it, he s eemed t o refer mostly to variations in the distribution o f polishes when v iewed under low magnifications, i .e. the entire polish l ocation was observed and the distribution o f polish on edge damage s cars. In one instance only ( apart f rom

9 1

the"smooth-pitted polish"referred t o above) did Vaughan s eem t o refer to a variation in polish a ccording to t ool a ction, when t he polish was viewed under high magnification: t his was a polish"bevel"which he could only f ind on t ools which had been used with t ransverse a ctions. Moss ( 1983, p .3) w isely headed a s ection on experimental polishes " The wear-traces o f certain uses and worked materials". She a ppeared to be largely in agreement with the other r esearchers ( as discussed above) and in addition referred t o specific wear-traces, which she called microscopic l inear impact t races ( MLIT) ( ibid., p .95), on projectile points and barbs. My own experiments showed s imilar variations i n polishes from longitudinal ( cutting/sawing) a ctions and t ransversal ( scraping/planing/graving) a ctions on anisotropic materials such a s bone and antler ( although not a s much on wood and reeds a s Vaughan ( op. c it., p .135) had r eported.) In addition, Blind Test 1 1-20 ( see Part I , Chapter 2 0) i n which tools had been rubbed against antler, wood, bone, hide and f ern, i llustrated the importance o f tool a ction a s a variable most c learly: the resulting polishes were completely different from those of tool a ctions s uch a s s awing or s craping, and a whole new series o f experiments rubbing various materials had to be carried out, before polish identification could be attempted. The cause o f t his d ifference was probably the f act that the s tructure o f the material influences the polish considerably a s Anderson-Gerfaud ( 1981, Vol.1, pp.61-62) and Vaughan, ( 1981, p .135) had pointed out. ‚

I a lso found that drilling ( mechanically) and boring ( by hand) the same plank o f wood for the same amount o f t ime, with tools made o f t he same raw material, caused different polishes. Drilling caused the formation o f a completely smooth polish with very f ew striations on the t ip o f the t ool, while boring caused only i solated polish-areas with lots of striations on the t ip of the borer. The difference was probably caused by s everal factors ( discussed below) such a s different pres ure during work, d ifferent real duration of work ( i.e. t he number o f rotations) and a s lightly different angle of the tool to the worked material. In summary, I agree with Moss ( 1983, p .55) t hat"the motion or a set of motions, of t he tool in use"should be repeated a s c losely a s possible, and therefore one should not compare polish on an experimental f lint edge used t o cut antler with polish on an archaeological tool, which because o f its shape and retouch i s c lassified a s a scraper. I t follows that the exact a ction with which an experimental t ool had been used should be i ndicated in microwear r eports.

9 2

4 .

Tool

Shape

Keeley did not think that variations of the edge angle resulted i n any r ecognisable variation in the appearance of polish ( 1980, p .59) Most microwear analysts did not pay t oo much attention t o this variable, and in recent publications any indication o f t he shape of the experimental t ool has been omitted ( Keeley, 1 983; Bueller, 1 983) The notable exception was Moss who stated that the"morphology of the piece, particularly the working edge"(1983, p .55) should be duplicated. .

.

In my own experiments I found that the edge angle influenced the distribution of the polish, and so a thin-tipped borer penetrated deeper into the worked material than a broad-tipped borer and therefore the polish had a wider distribution. I a lso found that the edge shape ( when v iewed in section) a ffected polish distribution. Polish t ended to form primarily on the higher part of an uneven edge. I therefore found myself agreeing with Moss that the shape of t ool edges should be duplicated a s f ar a s possible. 5 .

Pressure

Although this variable when t ested in connection with reed cutting ( see Part I , Chapter 1 6 Unger-Hamilton, 1 983) did not s eem to be important, the opposite was t rue in other experiments: for instance, it was f ound that rubbing materials, using d ifferent amounts of pressure ( while a ll the other variables were the same) produced different amounts of polish on the f lint implements. ‚

6 .

Tool

Angle

The e ffect o f this variable was i llustrated when I used a s craper on wood at an angle of 9 0 degrees t o the worked material and the r esulting polish was confined to the very edge o f the tool. On the other hand I had used s crapers on wood at an angi ' . of approximately 4 5 degrees and the resulting polishes extended considerably over the ventral surface of the s crapers. While we cannot know at which angle ancient t ools were held, we must nevertheless be aware that different angles of the tool to the worked material can l ead to s lightly d ifferent polish distributions. 7 .

Moisture Content

Most microwear analysts mentioned that polishes varied a ccording to the moisture content of the worked materials, such as bone ( Keeley, 1 980, p .44) hide ( ibid., p .49) wood ( ibid., p .36; Anderson, 1 980, p .181) and antler ( Vaughan, 1 981, pp.145-146). However, only a f ew systematic experimental investigations into the e ffect of the moisture content o f the worked material on polish formation have been carried out t o date: Gysels and Cahen ( 1982) studied polishes f rom a variety o f materials, such a s c lay, both dry ,

9 3

,

and moist; Mansur-Franchomme ( 1983) studied the e ffect moisture on dry hide under the SEM; while I carried out study on plant polishes ( Unger-Hamilton, 1 983).

o f a

My experiments ( see Part I , Chapter 1 6 for details) with fresh and dried plants, cut in some instances a t different t imes o f the year, showed t hat the rate of polish development and the appearance o f polish varied considerably according t o the moisture content o f t he plants. I observed the same phenomenon with practically a ll materials ( such a s fresh and dry bone, f resh and dry hide, soaked and dry leather, fresh and dry wood) ( e. g. P l.23:d,f, P l.8:e, g) In some i nstances ( e. g. from u se polishes l ooked totally different t o

on soaked l eather) those f rom use on the

same dry material and looked much more l ike polishes from use on other soaked or f resh materials ( in this instance wood). I t f ollows that we should pay more attention t o this variable. For instance, the t ime during which antler was soaked for experimental use s hould be stated i n publications. 8 .

Species

Keeley s tated that there i s probably no difference i n polish according to the exact species o f antler ( 1980, p .56) a s different species of other worked materials such a s hide, wood, bone had not led to variations in polishes. However, Keeley did refer to quantitative differences in polishes according t o whether " dense wood"or"less dense wood" ( ibid., p .36) had been cut. Vaughan ( 1981, pp.154-155), Anderson-Gerfaud ( 1983) and I ( Unger-Hamilton, see Part I , Chapter 1 6) had found that plant polishes varied a ccording to the species o f the plants c ut. In addition I had observed s light differences in t he appearance of polishes f rom reindeer and f allow deer antler and wood ( see Part I , Chapter 1 4) S uch variations appeared to be related t o s light variations in the s tructure and the density o f t he worked materials. ,

‚

.

It therefore seems to me that t he exact species of the worked material should be used i n experiments i f possible, and that i t should be reported in publications, even i f only a f ew categories, such a s plants o r antler, cause specific variations i n polishes. 9 .

Abrasives

Brink, in his microwear study o f s crapers, found t hat when external abrasives had been a dded during the s craping of bone ( 1978, pp.81-82 and 8 8-93) a nd o f dry hide ( ibid., pp.97-98 and 1 06-110) that the rounding process of t he f lint

9 4

surface occurred much more rapidly than when abrasives had not been added. The same phenomenon was demonstrated by Mansur-Francho r nme ( 1983) in her study using S EM t echniques. Vaughan ( 1981, pp.163-4) devoted a s ection i n his thesis to the"Influence o f Grit on Polish Characteristics"which he f ound to be considerable: " A substantial amount of grit will a ct to replace existing polishes with grit polish, or it will completely dominate the polish formation t o l eave only i ts own characteristics..."(ibid., p .163). My own experiments drilling various materials, such a s stone, with or without the addition of abrasives, such a s s and, or sand and water together ( Unger-Hamilton et a l. 1 987) ( Pl.29:g), led me to agree wholeheartedly with Vaughan' s s tatement. 10.

Conclusions

The results f rom the investigation into the variables involved in polish formation proved that a great many variables other than the worked material are involved. This implied that a great range of controlled experiments should ideally be carried out before a microwear analysis i s attempted. However, i t would be impossible t o t est every combination of variables, for reasons of t ime and the l imitations o f human visual memory. This in t urn means that the researcher should investigate and construct experimental programmes for one t ool group at a t ime, or ideally that s everal researchers cooperate and establish l arge reference collections o f experimental tools. It a lso implies that variables should be reported in detail i n publications dealing with microwear analyses.

9 5

Chapter Use

1 8

-

Polish

and

Striations

due

to Factors

Other

than

All s cholars agreed that polish and striations, a s well a s edge damage ( see Part I , Chapter 4 ) could form on f lint surfaces which had not been used at a ll. Wear traces could f orm during the manufacture of a f lint tool, c ould be produced by natural agencies and by t rampling, handling and processing by archaeologists. ,

1 .

Manufacture

Such t races have been discussed and demonstrated by Keeley ( 1980, pp.25-28) in his s ection on"technological e ffects". He stated that"these t races can happen during the removal of a f lake from a core, when a f lake strikes the ground, or, e specially, when an implement i s being retouched. They are important. . because they can resemble " true"microwear and need to be distinguished from i t"(ibid, p .25). Keeley observed the effects o f various experimental manufacturing processes, e . g. hard hammer knapping ( ibid., p1.7-8), " spontaneous retouch"(ibid, p1.9) during knapping ( Newcomer, 1 976) retouch with hard hammer ( stone) and soft hammer ( antler), and found that t he resulting edge damage could be confused with edge damage due to use, but that this was not the case with polish and striations. He stated that the l atter ( e. g."smears"(ibid., p1.8) from hard hammer retouch and"abrasion"from spontaneous retouch) were unlikely to be confused with polish and striations due to use ( ibid., p .28) Vaughan discussed polish and striations from retouch with stone, antler, bone and wood ( 1981, pp.170-172) He thought that t races from soft hammer retouch would be too weak t o be recognised on ancient tools"given the problem of soil sheen or grit polish on prehistoric f lints" ( ibid., p .172) He had only detected a f ew instances of t races attributable to hard hammer retouch on the Lower Magdalenian tools f rom Cassegros he had studied ( ibid., p .172), a s had Moss ( 1983, p .104) on her t ools f rom the French F inal Paleolithic from P incevent and Pont d ' Ambon. .

,

.

.

.

and

My own experiments indicated that a ll striations resulted fror knapping

manner of and hard

polish hammer

( quartzite, see P l.6:c, and l imestone, s ee P l.6:a) retouch and soft hammer ( antler) retouch and pressure f laking with an antler t ine ( Pl.6:f). Such traces conformed to the use-wear t races from the corresponding materials outlined in Part I , Chapter 1 4. As was the case manufacture t ended f lint than on coarse

that f rom with

with use-wear to be more grained f lint

t races, t he traces from developed on f ine grained ( see Part I , Chapter 1 7)

I agree with the other researchers inasmuch a s I found t races from manufacture were u sually distinguishable those due to use, on a ccount of t he former's a lignment observable retouch scars, and on account of t he fact

9 6

t hat they o ccurred in i solation, compared t o the u se-wear polishes, which t ended t o cover the entire surface o f the working edge of t he f lint tool. However, I did f ind t races f rom manufacture potentially indistinguishable with those f rom use when the retouch was considerable, when the retouch was carried out with a material which may have been worked, or when the retouch was on the same a lignment a s the tool t o t he worked material during use, i .e. on s crapers or drills. I a lso found that snapping tools, such a s blades, could l eave a thin l ine o f polish with striations running in the direction o f the snapping, i .e. a long the t ools long axis ( Pl.6:e). Such o f s craping.

t races

could

be

misinterpreted

It follows that the microwear f amiliarize himself with a ll manner of in order to avoid misinterpretations. 2 .

a s

the

result

analyst has to manufacturing traces

Natural Agencies

Keeley discussed the e ffects of various natural agencies on surfaces o f used f lint implements, such a s the e ffects of patination, abrasion by water-borne sediments, abrasion by wind-borne sediments and soil movements ( 1980, pp.28-35) He based his observations mainly on the e ffects o f patination on his own experimental f lints which he had patinated by immersion into a 3 0% solution of NaOH. ( Despite his knowledge t hat NaOH patinates f lint, he used s imilar solutions o f NaOH to c lean his f lint tools ( ibid., p .11); this i s discussed in Part I , Chapter 1 2) He a lso referred t o the f indings of other researchers, for example Semenov ( 1964, p .11), and to the detailed research by Rottlander ( 1975) Keeley found that patination e ffects such a s pitting and a granular appearance, were not l ikely to be confused with use-wear t races. He also found that severe patination rendered implements"unusable for x nicrowear analysis" ( op. .

.

c it.,

p .29).

Keeley wind-borne

a lso discussed abrasion and of

the effects o f waterand soil-movements on the basis o f

other researchers' results ( Shackley, 1 974; Stapert, 1976) and on the basis of his own observations of excavated f lint implements. He investigated the effects of t rampling experimentally ( this has recently a lso been investigated by Holmes, in press). Keeley concluded that such t races in the f orm o f abrasion, percussion craters, striations, and in t he case of soil movements,"white s cratches", would be easily distinguished because of their random and widespread distribution over the tool surfaces ( Keeley, op. c it., pp.34-35, p1.11-17). He d id point out, however, that S tapert had been"somewhat pessimistic on some points concerning potential confusions between natural t races and microwear..."(ibid., p .35).

9 7

Vaughan ( 1981, pp.173-179) described the e ffects f rom contact with s oil and grit on h is experimental f lints l eading t o"smooth-type grit polish", ( in his opinion potentially s imilar to certain weak use-wear polishes, i f s triations are absent) and" rough-type grit or soil polish"(the"glossy patina"of other researchers) f rom such agencies a s water-rolling. This type of polish was according t o him unmistakable, but rendered archaeological ,

implements useless f or microwear analysis. He s tated that"the same general soil sheen i s present i n l esser degrees on every archaeological f lint t hat has resided in a sedimentary. or aqueous matrix" ( ibid., p .174). He a lso thought that such polishes could be distinguished f rom t rue microwear due t o use on account of the widespread d istribution of the f ormer. However, he l isted some cases ( meat and hide polishes) which could look l ike weak soil sheen. He a lso pointed t o the fact t hat natural agencies can remove polish ( ibid., p .176). He quite rightly deplored the f act that the implications of such damage had been a lmost completely neglected in microwear publications ( ibid.). Vaughan therefore adopted Anderson-Gerfaud's system ( 1981, Vol.11, p .69) which reflected"the degree of certitude in the determination of the worked material f rom prehistoric microwear polishes" ( Vaughan, op. c it., p .177) Vaughan's identifications ranged from"definitely or very highly probable", " most l ikely"to"unspecified . .materials" ( ibid., P .177) or"too abraded"(ibid. ,, p .178; see a lso Vaughan's Table 3 5) .

.

Moss ( 1983, pp.81-83) agreed with Keeley inasmuch a s she stated that i f".. . natural phenomena do not conform t o microwear patterns they can confidently be a scribed t o some other causal f actor t han human use"(ibid., p .81). She agreed with Vaughan that"all archaeological collections have some sort of soil sheen"(ibid.). She mostly discussed one particular phenomenon which she called"bright spots" which may be due to unknown agencies 1 976, p .38).

f riction ( see Shepherd, ( see Vaughan, op. c it.,

1 972) p .365;

or other, S tapert,

I t oo observed a ll t he natural t races discussed ( apart f rom the"white s cratches") on my archaeological from Arjoune ( Pl.7: b, d,f).

above t ools

Although microwear analysts had agreed that most post-depositional t races, on a ccount o f their random distribution over the f lint surfaces, are d istinguishable f rom use-wear polishes, and a lthough t his s eemed t o be t he case on most of my archaeological pieces, I conducted my own s eries o f experiments, designed t o t est t he e ffects o f burning, and o f water and/or sand a ction on f lint.

9 8

Experiments An experiment was carried out in which Brandon and Syrian f lint had been thrown into

pieces of a wood f ire

and l eft overnight in the cooling f ire. The resulting " wear"traces included complete greying o f the f lint, and cracks visible under the microscope ( Pl.7: g). The s ame phenomena were observed on several f lint implements f rom Arjoune. Also experiments were carried out rubbing f lint against s and and s tone, with and without water. The polishes varied according t o the contact material ( Pl.7:a,c,e). An experiment was carried out in which a perforated plastic basket c ontaining 2 0 f lint pieces ( fine-grained f lint f rom Brandon, f ineto coarse-grained f lint pieces from Syria, and burnt f lint) was l eft i n a f ast-running stream in Wales f or f ive weeks in the spring 1 984. A considerable amount of sediment had collected a t the bottom o f the basket by t he t ime it was retrieved. The resulting wear traces were surprisingly i solated, random striations

very and

f ew a

and very weak. f ew polish

A f ew spots

resembling the"bright spots"discussed by Moss ( op. c it., pp.81-83) were confined to the f ine grained f lint pieces. This experiment was perhaps too short in duration t o judge whether the f lints f rom Arjoune had been water-rolled or not. Removal

of Use-Wear

Polish

Far more disturbing than the positive t races left by natural agencies on f lint surfaces, was the phenomenon f irst mentioned by Vaughan ( see above) and by Gysels and Cahen ( 1982) that polishes on archaeological f lint implements never l ooked a s well developed a s those on t he experimental t ools. Also worrying was the publication by P lisson of his experimental work ( 1983, P lisson and Mauger, 1 986) which showed that polishes can be removed by various a lkaline chemicals. S imilar effects on experimental f lint implements, from rinsing these with solutions of NaOH, and e ffects o f a lkaline environments on ancient f lint implements, were reported by Anderson and Whitlow ( 1983) ( See a lso Part I , Chapter 1 3.) In P lisson's experiments the rate o f removal o f polishes varied a ccording to the worked material. Hide polishes survived l onger than other polishes, and i t i s therefore somewhat disturbing that on both the s ites which Vaughan ( 1981) and Moss ( 1983) analysed, hide working tools were reported t o be in the majority The reason f or .

P lisson's r esult i s probably not the f act ( as he believed ( op. c it.)) t hat the chemical compositions o f the polishes vary with the worked material, but that a lkaline chemicals such a s NaOH attack amorphous s ilica and t o a l esser extent the f lint grains; and therefore that t he more amorphous s ilica i s present, the f aster the polish s eems to disappear.

9 9

L ike Gysels and Cahen ( 1982) I had f ound t hat polishes on archaeological f lints rarely l ooked exactly l ike t he experimental polishes, and i t seemed t o me that s ometimes the amorphous s ilica had disappeared. However, t he distribution of the polish ( in the f orm of f lattening o f t he f lint surfaces) remained in most cases visible. 3 .

Polish and

Striations

from Post-excavational

In addition t o traces from natural take into account wear t races procedures:

Procedures

agencies one has t o f rom archaeological

S tone polish can occur f rom bagging f lints together in one bag; metal polish can occur from t rowelling and s ieving; striations can occur from c leaning w ith brushes and polish can be produced by t ouching the t ools with a pencil during drawing. Such traces are rarely referred t o in microwear reports. I did f ind such traces, but they t ended to be i solated and random ( as in the case of polish f rom bagging tools together) or they could be detected with the naked eye ( as in the cases of contact with pencil) It i s however conceivable that brushing a f lint edge a long that edge might cause s triations which could be mistaken f or striations due to use. .

I f a microwear analysis i s intended for a s ite which i s in the process of being excavated, then the microwear analyst can ensure that most of such t races can be avoided, by insisting that the tools he wants t o analyse are not s ieved, not c leaned with brushes and t hat they are wrapped separately, and that care i s exercised during drawing. 4 .

Conclusions

On a ccount of the problems discussed above, I believe that archaeological implements would only rarely show intact use-wear polishes. This means that a lot o f archaeological implements cannot be assigned t o a particular worked material. This in turn means that a statistical evaluation of particular functions in relation t o particular areas of a s ite ( i.e. a distribution analysis) i s impossible.

1 00

Chapter

1 9

-

Sa m pling

Moss ( 1983, pp.106-107) discussed the problems related t o s ampling f or the microwear analyst. As we are dealing with a very s low method involving experimentation and s canning implements a t high magnifications we are dealing nearly a lways with sub-samples of the samples excavated by a rchaeologists. Moss stated,"We can e ither sample i n a ccordance with t he patterning of a particular s ite or we can choose a s ite in a ccordance with particular microwear questions"(ibid., p .106). She l isted f ive categories o f investigation: t echnological e fficiency; spatial variability; a s traightforward description of a total a ssemblage ( such a s carried out by Vaughan on the Magdalenian 0 a ssemblage at Cassegros, possible because of t he relatively small number o f f lints excavated) ; t emporal changes; and the"affirmation of some experimental phenomena" ( ibid.) I had chosen the s ite of Arjoune because it had been excavated recently by excavators known to me ( so their methods could be checked) because a ll the excavated l ithic material had been kept, because the entire a ssemblage was theoretically available f or study, because i t was possible f or me t o get to the s ite in order to f ind the f lint sources and use the l ocal f lint experimentally, and because environmental evidence was available a s an independent check on the microwear analysis. As the a ssemblage f rom Arjoune comprised 1 6,351 f lints ( see Part I I, Chapter 6 ), possibly f rom two s lightly different periods ( Trenches I -V and VII, and Trench VI) i t was impossible t o look at a ll of t he f lints, and I decided to sample with the following t opics in mind:

-

-

One

particular

The

use

o f

t ool

other

c lass,

tool

the

types

at

s ickle

blades,

Arjoune;

The comparison of c lassifications based on microwear analysis with those based on morphological typology;

-

-

-

The

use

of

the

l ithic

The

s ignificance

of

raw material.

burnt

f lint

at

the

site.

I examined t he following t ools under the microscope: a ll the available lustered blades from Trenches V ( 40) and VI ( 38) a s well a s 8 8 lustered blades from other s ites and periods ; a ll t he t ools f rom Trench VI ( 241) a ll the tools drawn by Copeland ( in prep.) and a s election of tools f rom Trenches I -TV ( 27) and VII ( 47) and a selection o f tools f rom Trench V ( 77) ( These counts do not include t he lustered b lades.) ,

,

,

.

In order t o investigate the s ignificance Arjoune I examined ( macroscopically) a ll

1 01

of burnt f lint a t the l ithic material

f rom Trench VI ( 1,477 pieces). The t otal number o f t ools f rom Arjoune examined under t he microscope was 4 70. I could only come to an opinion about 1 80 tools ( including a ll t he s ickle blades ( 78) examined) Originally I had intended a f ar more ambitious program involving the microwear analysis o f a ll the f lint f rom several loci o f Trenches V and VI, i n order t o be able t o chart the horizontal distribution o f polishes according to worked material. It was hoped that t his would i llustrate possible activity areas on the s ite. In t he event i t was not possible . to carry out this task; f or a variety of reasons ( see Conclusions: Theory and Method of Microwear) the identifications o f worked materials were only rarely precise and therefore a statistical analysis of polishes on a ll f lint proved impossible; in addition the drawn f lints from Trench V had disappeared and I was unable to examine the most distinctive t ools from the most productive t rench.

1 02

Chapter

t ools

2 0

-

The B lind Tests

A f ew blind t ests ( involving by one r esearcher, and

the the

use o f experimental identification o f the

wear-traces on such tools by another researcher) had been carried out. Newcomer ( Keeley and Newcomer, 1 977) had used f ifteen unretouched f lakes and s imple tools made of Brandon f lint i n a variety o f tasks, on a variety o f materials. Keeley identified 1 4 out 1 6 used areas, 1 2 out o f 1 6 tool movements and 1 0 out o f 1 6 materials correctly. ( For a s imilar b lind t est s ee Gendel and P irnay ( 1982, pp.251-265).) Holley and Del Bene ( 1981) criticised the methodology of K eeley and Newcomer's blind test. They applied more stringent c riteria in defining the success rate of Keeley's identification of worked materials and thought that instead of"approximately 1 0 out o f 1 6", Keeley had only identified 8 out o f 1 6 worked materials correctly ( ibid., p .341). Keeley ( in the same volume, 1 981, p .351) replied that"Holley and Del Bene show an obvious bias t oward counting e rrors". Holley and Del Bene ( op.cit.) also c laimed that Keeley had based his interpretation not only on polish identification, but on the perceived resistance o f the worked material ( ibid., p .342), a c laim which Keeley vigorously denied ( 1981, p .350) As I discussed in Part I , Chapters 7 , it s eems that the resistance o f the worked material ( i.e. t he abrasion) influences the appearance o f .

polishes, and I t herefore believe that Holley and Del Bene were wrong in s tating that identification o f this attribute invalidated Keeley's method, and that Keeley was wrong in c laiming that he did not take resistance o f the worked materials

into

a ccount.

Odell and Odell-Vereecken ( 1980) accepted Keeley's method a s viable a fter Keeley and Newcomer's blind t est; however, they decided to design a blind test to test their own Low Power Approach. Odell-Vereecken used thirty one basalt implements on a wide range of a ctivities and materials, and Odell's results were a s follows: 79.4% correct l ocation, 6 9% correct t ool a ction, 6 1.3% correct relative hardness of the worked material and only 3 8.7% correct worked material ( ibid., p .116). Despite the last ( expected) low percentage, their results seemed to be s imilar to those o f Keeley and Newcomer's blind test of the more l aborious and costly High Power Approach. Keeley criticised this test which in his opinion was not quite"blind" a s t he researchers were married t o one another, and also pointed out that the Odells had counted"as correct, guesses on implements on which the used portion had been misidentified"(1981, p .351) Vaughan's"Reliability Test"(1981, pp.102-5) was in my opinion even l ess"blind"than that conducted by the Odells. Vaughan had used t hirty two implements himself, subjected them to a ll manner o f a ccidental and natural processes, and subsequently believed that four months later ( when he examined the wear-traces) he had forgotten how he had used

1 03

the tools. He identified the correctly ( ibid., p .103).

f unction

o f

2 3

out

o f

3 2

tools

In my own research I took part i n a s eries o f experimental blind t est which were not completed at t he t ime this thesis was examined but which has in t he meantime been published ( Newcomer et a l. 1 986), and a lso in a microwear-analysis o f ancient tools which I carried out before looking a t other evidence f rom the s ite ( Unger-Hamilton et a l. 1 987) ‚

‚

The

following were

the

results:

B lind t ests 1 -10 ( see Table 3 ): a variety o f t ools, made o f S eine f lint, were used by Newcomer i n a variety of a ctions. The tools were examined, c leaned in a solution of NaOH and reexamined. The chemical c leaning appeared to remove some polish. Nevertheless, my polish identifications remained the same. My results were a s follows: 7 out of 1 0 correct location, 6 out of 1 0 correct tool a ction and 2 out of 1 0 correct worked material. I attribute the poor results identification of the worked material to polishes are not always distinct according

concerning the s everal f actors: to the worked

material ( see Part I , Chapter 1 4); a l arge number o f variables had been involved; I had not experimented with the f lint type nor with a ll the tool t ypes Newcomer had used; instead of using my own interpretation of polishes f rom the point o f view of what happens t o the f lint surface and deciding s imply whether the worked material was hard or soft and dry or moist, I followed t he usual interpretative t echnique of looking at the appearance of polish. I was thus forced into identifying exact materials categories of materials to begin with.

rather

than

B lind t est 1 1-20 ( see Table 4 ): Newcomer used ten Brandon f lint f lakes to rub against various materials, which turned out to be paired, for t en minutes each. The edges of the f lakes had been removed with a copper pressure f iaker, in order t o e liminate edge damage a s a possible indicator. I identified approximately 7 out o f lo worked materials correctly. This positive result was i n my opinion due to the fact that the number of variables was l imited, that I had carried out a ll the relevant experimentation ( with the exception of f ern; however I did identify this polish correctly as plant polish) and that I had used my own t echnique of observation l eading me t o f irst identify categories o f materials ( hard-soft, dry-moist) and only then t he worked materials. The visible gloss was a lso helpful for correct identification. ,

Blind t est 2 1-30 ( see Table 5 ): Newcomer used ten Brandon f lint blades to cut various materials for t en minutes. We knew that the materials had been the s ame a s those in blind t est 1 1-20. There was very l ittle edge damage, and the

1 04

microscopic polishes l ooked mostly i ndistinguishable. My results were a s f ollows: f rom observation of the edge damage and gloss with the naked eye I identified 3 out of 1 0 materials c orrectly, by microscopic examination only 2 out o f 1 0 materials. A blind t est related to prehistoric conditions was c arried out by us ( Unger-Hamilton et a l. 1 987) through an experimental microwear analysis o f f ifteen drills f rom Abu S alabikh, which we carried out before we looked up the published details about the s ite ( Crowfoot Payne, 1 980) Bergman had made copies o f the drills and we carried out a wide range of experiments ( 23) drilling and in s ome instances boring wood, pottery, lapis lazuli, malachite, hippopotamus " ivory", bone, copper, two kinds of shell, hide, with and without the addition of abrasives and/or water. I concentrated on the identification o f polish and s triations only, as the edge damage was negligible in a ll cases, except where the t ips had broken off. Only the wear-traces on 3 out of 15 drills from Abu S alabikh could be identified with certainty; the others were e ither severely damaged, or else covered with a weak polish, probably due t o natural agencies. I thought that two drills had been used ‚

t o drill shell and one drill had been used to drill something with the addition o f sand and water. In the latter case the drilled material was probably stone with a hardness grade of l ess than 7 on the MOH scale, a s abrasives proved ineffective with harder Hodges, 1 976, p .107) We were encouraged to l ooked a t the s ite report, that drills .

drilling with stone ( see a lso see, when we had been f ound

t ogether with shell beads, and lumps and beads of carnelian and lapis ( both s tones with a hardness grade of less than 7 on

the

MOH

scale).

Conclusions The results of the blind tests suggested that the High Power Approach t o microwear analysis can be relied on to identify both the used area of a tool and the tool a ction with

some

certainty.

The identification of the worked material i s more problematic. I t seems that the greater the number of variables ( especially when their effect i s not t ested) the ,

l ess

chance

there

i s

of

identifying

the

worked material

correctly. My worked

own approach, f irst defining materials ( from the f lattening

the categories o f o f the f lint surface

and the amount o f amorphous s ilica on the tool) subsequently guessing the worked material distribution of the polish), may be more helpful ,

a ttempt t o appearance

identify the worked material of the polish.

Wear-traces

f rom

uses

involving

1 05

and only ( from t he than the

immediately

intense

from

pressure,

the

s uch

a s f rom drilling or rubbing, were more developed t han t hose f rom uses involving l ess intense pressure, such a s cutting. This suggests that worked materials are l ess l ikely t o be be identified through a microwear analysis o f c utting implements than through a r n icrowear analysis of drilling or s craping implements. It seems that worked materials can be identified on ancient tools, provided they were used intensely and the number of variables involved i s small. However, even under those conditions, experimentation has to be wide-ranging.

1 06

Conclusions

-

Theory and Method

of Microwear Analysis

As a result o f t he foregoing examination of the t heory and method of microwear analysis I came t o the conclusion t hat both t he Low P ower Approach and t he H igh Power Approach are a ffected by t he same variables and t he same problems. These approaches need to be used t ogether a s the principal e lements o f each edge damage and polish should be v iewed in relation t o one another. -

-

Polish appears to consist of an abraded f lint surface, t ogether with a thin coating o f amorphous s ilica. The f actors involved i n the formation of polish, s triations and gloss seem t o be t he same, that i s to say abrasion and the moisture content o f the worked material. Examination under the SEM seems to suggest that residues f rom worked materials adhered t o the f lint but did not become embedded in the f lint, and that microand macro-organisms incorporated in the f lint during i ts formation could be mistaken for residues. The examination under c leaning with chemicals, should be avoided.

the SEM l ed me t o conclude that e specially a lkaline solutions,

The H igh Power Approach was found usefulness. I t c an be used to identify a t ool and the t ool a ction. However the

t o be of varying t he working edge o f identification of

the worked material was f raught with problems: materials, when worked f or a short t ime, left hardly any wear-traces and several materials l eft only i ll-defined traces, even when worked f or a l ong t ime. Some t ool a ctions l ed t o more s trongly developed polishes than did others. Polishes d id not mysteriously conform to precise worked materials, by virtue o f a magical ingredient, such a s residues f rom the worked materials. Polishes seemed t o be a ffected by the hardness, the moisture content and t he structure o f the worked material. This means that t otally different materials can cause virtually identical polishes, and a lso that the same materials can cause polishes distinct from each other. The

number

o f

variables

was found t o be experimentation has to and that researchers

involved

in

polish

formation

considerable. This means that be well controlled and wide-ranging, should concentrate on a few tool

c lasses only. This was also borne out by the results o f blind t ests. The effect of some variables, such a s f lint type, was f ound t o be considerable.

the the

Many t races due to agencies other than use can a ffect excavated f lint implements. This f act, and other factors l ike coarse-grained f lint, could render many implements unusable f or r n icrowear analysis. This has direct implications on sampling; i t might out a s ite distribution analysis.

1 07

be

impossible

t o

carry

An i solated blind t est produce results, a s equence conducted; only in this way developed and a ssessed.

1 08

i s unsatisfactory, and of b lind t ests needs to can different methods

t o be be

Part

II.

Sickle

Blades

and Other Tools

from Arjoune

Introduction P art I , the investigation into the theory and method o f microwear analysis, suggested that the used areas of t ools and the a ction with which t ools had been used could be i dentified with some degree o f certainty. Among the many variables however, the worked material could only be i dentified under certain c ircumstances and with far l ess c ertainty than had previously been a ssumed. This meant that in Part I I I relatively small chosen, namely Syria.

could only come to an opinion about a s election of t ools from the s ite I had t he 5 th millennium BC s ite o f Arjoune in

In Part I I, I describe the s ite of Arjoune, i ts s etting, environmental evidence and i ts artefacts. I investigate the condition of t he prehistoric f lint tools and t he source of the f lint. ( Although the chipped stone tools at Arjoune were made of both l imestone chert and chalk f lint, I refer t o both as"flint"for the sake of s implicity.) I briefly outline the method of investigation which i s based on Part I . I discuss previous research, my own experiments and selected ancient tools according to t ool type. I devote a large chapter to lustered"sickle blades'as the tool type which interested me most, and compare the wear-traces on lustered blades from Arjoune with those f rom other periods and s ites, namely from Natufian Mugharet El Kebara ( hereinafter referred to a s Kebara) and Mugharet El Wad ( hereinafter referred to a s El Wad), and from the P re-Pottery Neolithic ( hereinafter referred to as PPN) A to the Early Bronze Age ( hereinafter referred to a s EB) at Tell e s-Sultan ( hereinafter referred to a s Jericho) The wear-traces on lustered blades are a lso compared with .

botanical evidence where available. I investigate the occurrence and s ignificance of burnt f lint implements at Arjoune. F inally I report my conclusions a s regards the s ite of Arjoune and reassess the usefulness of microwear analysis in archaeology.

1 09

Chapter

1

-

Arjoune

and

its

Environment

The physical s etting of Arjoune has been described by Dorrell ( in prep.). The s ite i s s ituated north of t he Beqaa Valley in the middle o f the funnel-shaped d ivergence o f t he Lebanon and the Anti-Lebanon ranges ( Fig.3, 4). Just north of Arjoune, the Lebanon range i s interrupted by t he Hor ns-Tripoli Gap which runs east t o west. The s ite l ies on a low terrace, formed of a lluvium and eroded bedrock with outcrops of underlying conglomerate, on t he eastern edge of the f loodplain of the Orontes river. The Orontes aquifer i s fed by rainfall in the largely independent o f rainfall. Near Arjoune, meanders down towards the

Lebanon mountains and i s therefore seasonal variations in l ocal the river forms a s eries of active Lake of Hor ns.

The present-day annual rainfall of 4 00-600 mm i s confined to the period between November and mid-April ( about 5 0 rain days per annum) Summer drought i s a lmost t otal, but westerlies give r ise t o morning mist and dew. Temperatures are more variable and, a s far as cultivation i s concerned, less important than rainfall. Mean monthly maximum temperatures range from 2 0 C degrees in January t o 3 8 C degrees in August, both extremes t empered by westerly winds ( Dorrell, ibid.). .

At present t he economy in this f ertile region i s based on agriculture, f arming mainly New World crops such a s maize and l ivestock such a s sheep, goats, cattle and horses. Dorrell ( ibid.) reported that water-meadows on the f lood plains still support a c losed cover of grasses and herbs, despite as reeds birds in

shift

intensive and canes, and around

grazing. I found many marshy plants and many species of f ish, f rogs the Orontes.

There seemed to be no evidence s ince 6 000 BC ( the evidence

such and

for a major c limatic from pollen cores ( e.g.

Niklevski and Van Zeist, 1 970) i s discussed by Dorrell, ibid.). The country around Arjoune was probably forest-steppe during the Chalcolithic and easily c leared f or cultivation, the f lood plain may well have had dense r iverine vegetation and gallery forest and was probably an area r ich in game, f ish and wild t ree crops. Dorrell pointed out that so far no s ettlement of this period had been found west of the Orontes and that i t was unclear whether the f lood plain of the river had been season. He thought that in the past s ite probably included i ts position northwards.

1 10

crossable in t he wet the attractions o f this on the Beqaa routes

Chapter

but t he

2

-

The Excavations

The f inal excavation report has not yet the present author has seen t he f inal f ollowing summary i s based on this.

been published, manuscript, and

The s ite o f Arjoune, encompassing s everal small natural mounds, f ollows t he edge o f a meander cusp and covers an area o f approximately 4 00 x 2 00 meters ( Fig.5,6). Surface s catters on all t he mounds consisted predominantly o f sherds and l ithics o f t he Late Neolithic! Chalcolithic period, together with some Hellenistic sherds. The h ighest Mound A ( approximately 5 m in height) was i nvestigated i n one sounding and three small probes ( Trenches 1 -1V) i n 1 978 ( Marfoe et a l. 1 981) On t he l arger Mound B Trench V and Trench VII were excavated in 1 979 and 1 982 r espectively. In addition Trench VI was excavated in 1 981 ( Fig.6) The trenches were excavated in squares of 1 x 1 , 1 x 2 or 2 x 2 meters. Some o f the excavated soil was dry-sieved. All the l ithic material was curated. ‚

.

.

According t o the excavators ( Parr et a ] .. in prep.) s everal a reas o f the excavation revealed what were probably the lower portions of s emi-subterranean dwellings or"activity areas", roughly c ircular in shape and sunk up to a meter or so into the natural rock. The sunken a reas were f illed with occupation deposits. Faint t races of horizontal surfaces i n the deposits were noted, but t here i s no c lear evidence that t hese were real l iving f loors. The excavators found the absence o f structures puzzling and thought it possible that permanent structures may have been destroyed by erosion. ‚

The a ccelerated carbon dates were obtained from bone and grain samples. " Arjoune V and VII a re believed t o be shortterm o ccupations, which are of effectively the s ame age, c . 6 600 B?. Arjoune VI c learly represents l ater occupation." ( Gowiett et al., 1 987). In f act the dates obtained f rom Arjoune VI appear to be about one thousand years later than t hose from the other trenches.

1 11

Chapter

3

-

The

Botanical

Evidence

Moffett ( in prep.) carried out a small-scale program o f sampling f or charred botanical r emains. The staple crops represented at Arjoune were e inkorn, emmer, two-row hulled barley and l entils. The wheats were fully domesticated. Free-threshing wheat was f ound in Trench V . but not in Trench VI. This may have been an artefact of s ampling a ccording t o Moffett. The same may have been the case with the scarcity of barley in Trench VI a s opposed to Trench V . Lentils were probably the cultivated f orm, a lthough the possibility that wild l entils ( as well a s wild cereals) had been collected could not be ruled out. Other food plants included the horsebean and grapes, pips of which were f ound to be concentrated in l ocus 1 15.3 i n Trench V . These could have been an early cultivated type, or e lse may have been collected f rom vines growing wild in t he Orontes Valley. In addition s eeds from various fruit trees were recovered: hawthorn, pistacio, plum or cherry, sweet a lmond and f ig. Fruit from these trees was probably collected f rom trees growing on the mountain s lopes o f the Lebanon and Anti-Lebanon.

1 12

Chapter

4

-

The

Faunal

Analysis

Grigson ( in prep.) has shown that domestic ungulates ( goats, sheep, pigs and c attle) made up 9 6.4% o f the f aunal a ssemblage and dogs 0 .5%. The wild animals represented were hare, cat, gazelle, red deer ( shed antler of red deer was f ound in Trench VI) bird, t ortoise and c rab. It was not possible t o say whether the equid bones were of onager or donkey, nor whether they were f rom wild or domesticated animals. ,

About 6 0% of the domestic ungulates were sheep and goats, with goats outnumbering sheep by about 6 :4. In both species f emales outnumbered males, more o f the males having been s laughtered when young. The survivorship curves t ogether with the s exing analyses suggested the sheep and goats f rom Trenches V and VII were kept primarily for meat, whereas those from Trench VI indicated a herding strategy with an emphasis on milk. This suggested to Grigson that Trench VI was perhaps later in t ime than Trenches V and VII. P ig bones constituted 2 0% of the domestic ungulate a ssemblage. Ageing data showed that about a third of the pigs were killed in their f irst year and about a half in their second. It was suggested that most males were k illed i n their f irst year, while more f emales were kept until adulthood, presumably for breeding. Ageing data, and the hypothesis that these pigs, l ike wild pigs, gave birth once a year in March, suggested that most of the pigs at Arjoune were killed in the summer. This indicated either a deliberate culling policy or s easonal occupation or both; t he marshy f lood plain o f Arjoune in the summer might be an ideal area in which to herd pigs, while in the winter the pigs were perhaps driven to a drier area. Cattle bones constituted the remaining 2 0% of t he domestic ungulate a ssemblage. The f ew ageing and sexing data f or cattle suggested a generalised herding policy allowing f or meat and breeding.

1 13

Chapter

5

-

Non-Flint Artefacts

F inds f rom Arjoune consisted of pot s herds, dated t o the 5 th millennium BC including Amuq Dark Faced Burnished and related wares and some painted wares including probably imported Halaf ware ( Marfoe et a l. 1 981; Parr in prep.). Many stone objects were f ound. These i ncluded heavy stone equipment, such a s basalt querns, and a lso grindstones, bowls, rubbers and hammerstones, l imestone bowls and plaques, s tone maceheads, engraved and p ierced pendants and engraved pebble f igurines. Clay objects included spindle whorls or weights, baked c lay f igurines with incised f eatures, and a large quantity o f unperforated clay disks. Objects carved of bone i ncluded awls and a t iered carving. A shell pendant was a lso f ound. No metal objects were recovered. ‚

1 14

Chapter

6

-

The

F lint

Copeland ( 1981; in prep.), who i s publishing the t ypology o f the f lint, stated that a t otal o f 1 6,351 f lint p ieces was recovered f rom the s ite of Arjoune. In 1 978 5 41 pieces were excavated mainly f rom disturbed contexts, Trenches I and I II on mound A ( Marfoe et a l. 1 981). Here, only Phase I in Trench I was an undisturbed Neolithic/Chalcolithic deposit, with just 8 8 f lint pieces, comprising 3 t ools and 8 5 pieces o f debitage. The rest o f the f lint, f rom d isturbed or later contexts was dated t o the s ame period and consisted of 9 8 tools and 3 55 pieces o f debitage ( unretouched f lakes or cores). In 1 979 1 2,248 f lint pieces were excavated from Trench V , and 2 were collected f rom mounds B and C . Of these, 1 ,004 were c lassified as t ools, 8 ,769 a s debitage and 2 ,465 pieces a s non-artefactual. The excavation in 1981 o f Trench VI yielded 1 ,477 f lint pieces, comprising 2 41 t ools, 1 ,236 pieces of debitage and 3 83 non-artefactual pieces. In 1 982 excavation o f Trench VII y ielded 1 02 tools and 1 ,710 pieces o f debitage. I n addition 3 7 f lint artefacts were recovered from S ounding 1 , 1 6 f rom Sounding 3 , 1 17 from S ounding 4 , 1 3 f rom S ounding 5 and 8 8 f rom Sounding Mound C . These were included by Copeland in the l ists of the l ithic artefacts from Trench VII and soundings. ‚

The t ool t otal included f ragments o f non-flint artefacts, presumed to have been used f or milling or f lint knapping, and obsidian. Very large or intact s tone t ools, such a s mortars, were not included ( Copeland, i n prep.). The Tables 7 -10 ( taken from Copeland's r eport, ibid.) lint a ssemblages into tool show the exact breakdown of the f nd non-artefacts. The tool t ypes included types, debitage a ickle e lements axes, chisels, picks, arrowheads, choppers, s ( lustered and unlustered) retouched bladelets, microliths, end-scrapers, s traight-ended s crapers, f lake-scrapers ( including fan-scrapers) core-scrapers, steep-scrapers, racloirs ( side-scrapers) raclettes ( small scrapers on thin blanks) burins, borers ( or piercers or drills) denticulates, notched pieces, backed blades, knives, composites, abruptly retouched pieces, pieces with nibbled retouch, bifacially retouched pieces, pieces with inverse/alternate retouch, truncations and various non-flint artefacts, such a s hammerstones, anvils and obsidian implements. ( Non-flint artefacts were not included in my study.) ‚

,

‚

,

According to Copeland ( ibid.) the l ithic assemblages f rom the various Arjoune t renches resemble each other f airly c losely i n raw materials, debitage t echniques and tool types. Nevertheless, there were differences in detail between a ssemblages from Trenches I -IV, V and VII as against those f rom Trench VI. There were a lso differences f rom t rench t o t rench i n the proportions and presence and absence o f c ertain types. Byblos and Amuq points were apparently

1 15

represented by only a few f ragments i n Trench V and one surface f ind on Mound B . In contrast, t ransverse arrowheads f ormed 1 .67% of t he a ssemblage of Trench VI. Byblos points began i n a s imple form in t he PPM a nd virtually disappeared with t he advent o f the Chalcolithic while t ransverse a rrowheads occurred on Late Neolithic/Early Chalcolithic s ites in the Near East, dated to 5 000 4 000 BC ( Copeland, ibid., a fter Bar-Yosef, 1 981, f ig.4). They related t o t he 4 th millennium BC in S inai s ites ( Copeland, ibid., a fter -

Bar-Yosef, 1 981, p .561). From t he arrowhead typology Copeland suggested a date of between 5 000 and 4000 BC f or t he Arjoune trenches. Other f eatures pointed out by Copeland ( ibid.) included the scarcity of axes at Arjoune, duplicated at other inland Neolithic s ites, in comparison t o the great number of axes on coastal Neolithic sites, such a s Ras Shamra or Byblos, near forested mountains. Axes from Trenches 1-1V and VII resembled those f rom Ard Tlaili and s ites northwards t o the Amuq, while t he adze from Trench VI l ooked s imilar to a type characteristic of t he Chalcolithic at southern s ites. S ickle e lements were more abundant in Trench VI ( 25.1%) and Trench VII ( 22.5%) than in Trench V where they only constituted 1 4.9% of t he tools. This suggested to Copeland that special a ctivities were carried out in the area of Trench V . Equally the presence of many and varied non-flint stone tools in t he area o f Trench V l ed Copeland t o wonder about the s ignificance of such a concentration: were these implements involved in processing vegetable, mineral or animal matter, or did they represent"tools to make tools"? There were traces of t he same heavy t ool k it at Trenches 1 -1V and VII, but i t was virtually absent from Trench VI. Obsidian was absent in Trenches VI and VII and rare in Trench V ( 0.9%) compared t o the small Trench 1 -1V samples where i t constituted 9 % of t he tools.

1 16

Chapter

7

-

The Raw Material

( Flint)

The raw material u sed to make t he ancient t ools a t Arjoune was very heterogeneous and comprised a variety o f t ypes o f f lint ( and chert) o f varying colour, a s well a s l imestone and s ome obsidian. Most o f the tools were r elatively small a nd o f i rregular shape, which would make t heir use o ften d ifficult. They were o ften made from coarse a nd/or variegated f lint which must have been difficult t o knap ( see below) However, a f ew t ools were made o f good quality f lint, t he nodules o f which had been quite large. .

We ( Parr, Dorrell and I ) d iscovered two sources o f f lint i n t he v icinity o f Arjoune, both of them secondary. The f irst, a gravel deposit containing f lint, chert, l imestone, s chistose and quartzite, was found in a dried out r iver-bed, the Wadi Rabiy' ah, approximately one hour's walk f rom Arjoune ( Fig.4) The colour and variegation o f t he f lint and chert matched t he raw material from the excavation exactly. The primary sources o f these gravels a re in t he L ebanese mountains, unfortunately inaccessible at the moment. I t i s probably f rom there that the f iner l arger nodules o f f lint derived. S econdly, some outcrops of f lint were found at Arjoune i tself. However, most f lint pebbles were found to be e ither f rost-shattered or compressed into a c onglomerate and therefore impossible to knap. S imilar nodules ( probably rejects) had been found in the excavated t renches. .

Most o f the variegated ( probably f rost-shattered) f lint and most o f the conglomerate proved impossible to knap a s i t splintered. In some instances the f lint nodules had t o be shattered with a metal hammer, a s the fracture l ines were i nterrupted by variegation. S imilarly some apparently r ecrystallised f lint ( Syrian A and B , see Part I , Chapter 1 7, Pl.5:e-h) was difficult to knap as i t was very dry and broke easily. S ome apparently f rost-shattered dark-purple f lint was quite easy to knap. S o was dark-brown f ine-grained f lint ( Syrian D , see Part I , Chapter 1 7.1) which however only occurred in very small pebbles. Fine-grained brown and orange s triped f lint ( Syrian C , ibid., P l.5:c, d) was easiest t o work. Usually a l imestone hammer was used which worked well at f irst but soon became very abraded. A heavier, harder s tone, such a s quartzite, would have been better. An antler hammer was rarely used. The tools produced were even smaller than those of the excavation. This may partly be due to the inferior quality o f the present day f lint in comparison to the ancient f lint, or e lse to my i nferior knapping s kills. I came material

to

u sed

t he at

following

conclusions

about

t he

r aw

Arjoune:

Most of t he raw material was obtained from a s ource approximately one hour's walk f rom t he s ite, with possible exceptions being t he f ine-grained l arge nodules from which

1 17

t he high-quality s crapers) and the ,

t ools obsidian

were made ( e.g. ( which has not been

the t abular analysed)

The great typological variation and l imitation in s ize o f t he archaeological l ithic material f rom Arjoune c an be explained by the poor quality o f t he raw material. Tool manufacture must have been a difficult business, a nd f rom t hat point of view t he s ite was badly l ocated, compared with s ites in the Jebel Abiad in t he S yrian Desert for i nstance, where seams of superb f lint abound.

1 18

Chapter

8

-

Preservation of

the

Excavated Flint

When v iewed w ith the naked eye, the f lint f rom Arjoune l ooked"often mint-fresh"(Copeland, 1 981, p .9). Only a f ew pieces were s lightly patinated ( i.e. white or shiny) a nd a good number of p ieces l ooked heated or burnt ( Copeland., ibid) However when viewed under t he microscope i t became apparent t hat more post-depositional surface modifications ( Holmes, 1 986; P art I , Chapter 1 8, P l.7: b, d,f) had t aken place than expected. There were many pieces only s lightly shiny t o t he naked eye which turned out t o have a polish consistent with patination. When I knapped a f lake from one such shiny core, i t turned out that t he f resh break was not glossy, and t herefore the surface modification had t aken place a fter deposition on the s ite. Some pieces made on what seemed t o be b lack or grey f lint turned out t o be burnt, judging by the c racks seen under the microscope ( Pl.6:h) Experimental f lint pieces which I had l eft in a f ire overnight developed the same colour, consistency and c racks .

-

-

as those pieces excavated f rom Arjoune ( Pl.6:g) Random striations and polish covering a ll areas of the t ool, including those areas which had most l ikely not been u sed, were found on many tools to a differing extent. S uch t races could well have originated from possible annual f looding of the s ite by the Orontes, or perhaps a lake may have f ormed .

here ( Dorrell, pers.comm.). Although experiments l eaving modern Arjoune f lint in a fast-running s tream for a month, had shown that hardly any t races resulted f rom this ( see Part I , Chapter 1 8) the duration o f t he experiment was perhaps f ar too s hort to be comparable. The soil i tself may have left t races: Grigson's bone e lement analysis ( in prep.) showed that f ragile bones had probably been destroyed. According t o her t his may have been due t o a number of ,

reasons, among t hem physical e ffects such as t emperature changes. As far a s the e ffects o f soil chemistry are concerned i t i s i nteresting that she reported evidence o f a neutralisation o f t he calcareous environment by humic a cid. The barium sulphate analysis I had c arried out gave a neutral value ( pH 7 ) for the occupation l evels f rom Arjoune ( Part I , Chapter 1 3) Polish ( probably partially consisting of amorphous s ilica, s ee Part I , Chapter 7 ) may have been removed in s ome cases, a lthough examination o f the s ickle b lades suggested that the polishes on most of these were relatively intact. A third and f ourth possibility f or t he occurrence of post-depositional surface modifications may have been t he fact that t ools had been s ieved, and that they had not been bagged individually, but had been t ransported and s tored a f ew years, and s o had rubbed against each other.

f or

I therefore examination.

the

had

t o

l eave

1 19

many

t ools

out

o f

Chapter

9

-

The Method

The method u sed in the microwear analysis of the from Arjoune and o f the s ickle blades f rom Kebara, and Jericho i s based on the conclusions of P art I .

tools El Wad

As far a s possible copies were made o f the ancient tools or at l east of t he edges which had most l ikely been used. The raw material f or the copies was mostly f lint f rom Arjoune, but I had a lso used other f lint o f various grain s izes. These included f ine-grained black f lint from Brandon, medium-grained grey f lint from Potter's B ar, f ine-grained black-grey f lint f rom Surrey and various t ypes o f f lint ( fineto coarse-grained) from France As t he f lint collected f rom Arjoune was o f mediocre quality and small, l arger tools had t o be made of other f lint. .

Both hard hammers ( of quartzite and l imestone) and soft hammers ( made of antler) were used to retouch the blanks which were kept i ndividually in plastic bags. The tools were, whenever possible, examined under the microscope before use. A f ew t ools were a lso examined and photographed before and a fter retouch. A wide range of experiments ( 436) was carried out in two separate series. The f irst, well c ontrolled, was designed t o identify the variables involved in polish formation ( Part I , Chapters 1 6 and 1 7); t he second, l ess controlled, was designed to s imulate, a s far a s it i s known, prehistoric a ctivity. As f ar a s possible t he materials used were chosen according to whether they were f ound on the s ite and whether they may have been worked by t he prehistoric peoples at Arjoune. S everal species o f wood ( seasoned and f resh and in one instance soaked in water) bone from various parts and species of animals ( fresh, cooked and dried) hide from various species o f animals ( fresh, dried, tanned and soaked) antler from various species ( dry and soaked) s oaked cow horn, cooked and raw meat, f resh and dried s inew, several species of fish, several species of shell, several varieties o f stone, pottery, sand, f eather, sheep's wool, hippopotamus " ivory", copper and several species of plants ( wild and cultivated, f resh and dried), were worked. The wear-traces from t hese a re described in Part I , Chapter 1 4. As the microwear investigation had shown that the hydration and in some instances the exact species o f the worked material could create differences in polishes I a lways noted the condition and t he species of the worked material. Actions included: sawing and cutting, s craping, whittling, planing, boring, drilling ,

,

,

,

( mechanically), piercing, chopping, graving and rubbing. As different a ctions could cause differences i n polishes the tool action was noted. The experiments a re described in detail together with the tool types f rom Arjoune ( see a lso Appendix). I f ound the i nformation derived f rom the experimentation itself extremely valuable.

1 20

The experimental ( as well a s the archaeological) t ools were washed in water and detergent ( usually ammonia-free) I n addition, the a rchaeological tools were immersed for up t o 1 0 minutes in a n ultrasonic c leaning tank. Chemicals were not used ( Part I , Chapter 1 2) a s a rule. However in a f ew instances a 1 0 96 s olution o f hydrochloric acid ( HC1) was used t o r emove c oncretions f rom f lint edges. As chemicals could ( and did) a lter polishes I noted any t reatment with HCl. Experimental blades when hafted in s ickles with a resin and wax mixture, were immersed in white spirit for up t o twenty-four hours i n order to remove the hafting agent c ompletely. After c leaning, the tools were bagged i ndividually in plastic bags. All t he pieces from Trench VI which were c lassified by Copeland a s tools were examined under the microscope, a s well a s many tools f rom Trench V , all the drawn tools from Trenches 1 -1V and VII and a s election of other tools f rom Trench VII. Unfortunately the drawn tools from Trench V were not available for s tudy. Altogether 4 70 tools from Arjoune were examined under the microscope, as well a s 25 lustered blades f rom Kebara and E l Wad and 6 3 f rom Jericho. However, in only some cases ( 180) could an opinion be formed about the tools f rom Arjoune. This was presumably because o f the s lowness with which coarse-grained f lint ( often used for Arjoune t ools) develops polishes and a lso because of post-depositional t races found to a greater or lesser extent on a ll the tools f rom the s ite. As a result a s ite distribution analysis could not be carried out. Observation o f experimental polishes and the results of t he blind t ests had demonstrated that the worked materials could only sometimes be identified with some c ertainty ( see Conclusions: Theory and Method of Microwear Analysis) I therefore did not a lways attempt to identify the worked .

material. I did not attempt to identify hafting traces ( Part I , Chapter 1 4.18) but t ried t o deduce hafting f rom the distribution of polish on what s eemed to be the used edge, and a lso f rom the retouch or shape of an implement. ,

Edge damage i s described a t magnifications polish and striations at 2 00x. Photographs

of 5 0x, and are shown at

2 00x, unless otherwise stated. The drawings are at a 1 :1 s cale, unless otherwise indicated. A f ew tools were outlined f rom drawings by Copeland. This i s indicated by"(L.C.)"next to the f igures. The edge angles protractor ( Fig.8:c)

of

scrapers

were

measured

with

a

I have not a ttempted to give a comprehensive survey o f previous research into functional analysis o f certain t ool types, a s f ull surveys o f the previous l iterature have been given by Vaughan ( 1981) and Moss ( 1983).

1 21

Chapter

10

-

Scrapers

( P1.8-10,

F ig.10-15)

Flakes or blades with a retouched working edge, usually a t one end, which i s o ften rounded, but s ometimes s traight or denticulated, are usually referred to a s"scrapers". Most s crapers are c learly t ools t hough"scraper retouch" c an a lso occur a ccidentally ( Newcomer, 1 976) 1 .

Previous

s ince

Research

Scrapers have received attention by f unctional analysts t he 1 9th century ( Moss, pp.38-43) The abundance of .

s crapers in Magdalenian and Neolithic t imes l ed some researchers t o believe that they were multi-purpose t ools used as s crapers, saws, chisels or burins ( Pfeiffer, 1 912, p .132). Recently Gould et a l. ( 1971) f rom ethnographic evidence inferred that Mousterian"La Quina"scrapers may have been used a s wood adzes. Microscopic wear-traces s uggested that scrapers f rom Mureybet had been used a s wood choppers ( Coqueugniot, 1 983) and that a few s crapers f rom R ingkloster had been used a s burins ( Juel Jensen, 1 981). Semenov ( 1964, p .87) on the other hand believed t hat a ll t he end-scrapers he had examined were used to s crape or soften hide. An important f eature in the f unctional s crapers proved to be the angle of t he working ( 1968) i ntroduced a hypothetical correlation

analysis o f edge. Wilmsen between edge

angle and function: for instance a r elatively acute angle o f 4 6-55 degrees was in his opinion optimal f or h ide s craping, a steeper edge of 6 6-75 degrees optimal f or hide s oftening and wood and bone scraping. Ethnographic s tudies ( Gould et a l. 1 971; Hayden, 1 979, p .124) and experimental ( low power) r n icrowear studies ( Broadbent and Knutssen, 1 975) confirmed Wilmsen's edge angle hypothesis. The edge-angle s eemed t o be a good indicator o f s craper u se, but t here were some problems: the l ack of a uniform measuring t echnique ( Hayden, 1 979a, p .211) l ed to different measurements when two researchers measured the s ame tools. According t o Wilmsen ( 1968) the edge angle might have been considerably increased by resharpening retouch. However, Broadbent and Knutssen ( 1975) reported f rom their experiments t hat not ‚

only did the edge angle increase w ith each resharpening by a s much a s 1 0 degrees, but a lso that with t he increased edge angle, resharpening by pressure-flaking became very difficult. Another s crapers was This feature

f eature observed in t he functional analysis o f t he sharpness or bluntness of t he working edge. provided t he distinction between hide scraping

and hide softening implements f or Se r nenov ( 1964, p .87). Rounding ( in plan) of the working e dge was an important f eature o f hide s crapers for Nissen and Dittemore ( 1974) a s otherwise t he h ide might have been lacerated. Ka r nminga ( 1978,

p .137),

however,

f ound

1 22

that

t his

was

not

so when he

s craped marsupial

skins.

Hafting too was debated. S emenov ( ibid.) thought t hat wear-traces concentrated on one part o f the working edge c onstituted proof o f handapplied pressure and t hus believed t hat a lmost a ll except the smallest s crapers had been used unhafted. Most other researchers reported t hat hide s crapers were only e fficient when used hafted ( Broadbent and Knutssen, 1 975, p .116) Microscopic s tudies concentrating on material included low-power s tudies, notably

the worked t he excellent

s tudy by Brink ( 1978) carried out on various materials in various conditions at up t o 5 0x. Brink investigated f laking, rounding, polish and s triations on used experimental s crapers. He f ound only polish and rounding t o be ,

consistently influenced by the worked material ( ibid., p .117) However a s regards the other two f eatures, he had observed f laking on the dorsal, i .e. retouched, a spect only ( see Part I , Chapter 4 ) and he had counted l inear abrasion marks not a s striations, but a s rounding ( op. c it., p .117) and t herefore did not f ind any s triations. He noted the considerable differences in wear-traces according t o the .

,

condition o f the worked material ( ibid., pp.118-119). The recognition of the fact that wear-traces f rom hide working can differ considerably from each other was in my opinion important. Such divergences ( apart f rom f resh hide dry h ide) were rarely mentioned with the notable exception o f -

Kamminga

( 1978,

p .137)

working marsupial animals.

who

s kins

reported

that

l ooked different

wear-traces t o

those

f rom

f rom

other

High-power s tudies of s crapers included those by Keeley ( 1979) Vaughan ( 1981) P lisson ( 1981) Gendel ( 1982) Juel Jensen ( 1 981) Moss ( 1983) Coqueugniot ( 1983) Hide, wood, bone and a ntler were inferred a s the worked materials. Problems such as absence of wear-traces due to resharpening or short duration of use, mentioned in connection with l ow power studies ( Hayden, 1 979a, p .207) presumably a ffect high power s tudies. In this context P lisson may have been a l ittle hasty when he c lassified the majority of s crapers ( 69 out o f 9 5) f rom La Tourasse a s unused ( 1981) ,

,

,

2 .

,

,

,

.

Experiments

F ifty-eight experimental scrapers were made of s everal types o f f lint ( see Part I I, Chapter 9 ) using both hard hammer and in some i nstances an antler pressure f laker in order t o produce a cute edge angles. The s izes o f t he s crapers ranged f rom very small ( 2.3 x 3 .0 cm) t o large ( 7.2 x 7 .0 cm). In order t o copy t he variety of scrapers f rom Arjoune ( see Tables 7 -10), f lake-scrapers, end-scrapers, s ide-scrapers and small s crapers were made. Three o f t he s crapers were used i nserted longitudinally in a wooden haft, held e ither with wooden wedges and dental f loss, or e lse with resin, wax and s inew. The following materials were ,

1 23

s craped: f resh and s easoned wood ( oak, a sh, s ycamore, cherry) bone ( long bones and r ibs, c ooked, raw and dried, f rom cow, pig and l amb), hide ( tanned pig's hide, dry and ,

soaked, fresh red and roe deer hide, and the hairy s ides o f the latter hides, soaked f or three days) antler ( reindeer and f allow deer, dry and soaked f or three days) soaked horn, ormar shell, l imestone, f resh c ane and reeds. ,

,

I t was a lways much easier to s crape wood which was f resh or water soaked rather t han seasoned. Relatively steep-angled ( 70-90 degrees) s crapers could be used. Edges used t o s crape f resh or soaked wood could still be used a fter 1 200 strokes, while edges used on s easoned wood became b lunted a fter 6 00 s trokes. ( Broadbent and Knutssen ( 1975) reported exhaustion o f their quartz s crapers a fter f ewer strokes) . As with bone and hide, a hafted s craper was far more e fficient and easier used than an unhafted s craper. Hafting with resin, wax and s inew proved more e fficient than wedging; the wedges became loose and the dental f loss broked quickly. S craping cooked and dried bone was easier than s craping f resh uncooked bone. Relatively steep-angled s crapers could be used, but had to be held at an angle of approximately 8 0 degrees to the bone to be e ffective. Edges u sed on uncooked f resh bone became dull a fter 1 000 strokes, while edges on dried or cooked bone could be used much longer. Brink had experienced s imilar difficulties when working c lean bone ( 1978,

p .79).

Bergman and I scraped a red deer hide t aken f rom an adult animal killed one week before the experiment. The hide had been left in a refrigerator and was s lightly dry with some meat and fat left on it. Water was used t o soften the skin. ( For ethnographic a ccounts o f hide preparation s ee Nissen and Dittemore, 1 974, pp.67-8; Wulff, 1 966, pp. 230-232

. ) Scrapers used in the fleshing process had t o have an a cute angle ( c.45 degrees) and most importantly an overhang ( Fig.8:c) with which t o grip the hide. S teep s crapers without an overhang s imply did not work a t a ll even when held at a steep angle to the hide. Hafting made a visible difference. A hafted s craper was not only easier t o use but scraped a much wider area much c leaner, even when used by a an inexperienced worker l ike myself, compared to the unhafted s craper used for the same t ime by an experienced worker l ike Bergman. It took us two hours t o f lesh the hide. The a cutely angled scrapers were still usable a fter 8 4 minutes o f s craping. According to Bergman ( pers. c omm.) who had scraped a hide with the addition of ochre, ochre did not make hide s craping more e fficient. Moss and I used s crapers to dehair a f resh roe deer hide a fter l etting the h ide soak in water f or t hree days. The hair came o ff quite easily a lthough longer soaking would have been better. I had t ried t o dehair a red deer hide with only a f ew minutes soaking and

no

hairs

came

o ff

at

a ll.

1 24

Antler was much easier s craped when soaked f or several days i n water t han when dry. S craper edges u sed on dry antler became dull a fter 1 000 strokes, while those used on s oaked antler could be used f or two hours ( c.3000 s trokes). Acutely angled s crapers cut better than s teeply angled s crapers.

in my

Only s oaked horn was experiment t he horn

When concretions edges became blunt s crapers could not be

a fter

s craped. It f laked o ff.

was

an

easy

t ask,

but

were scraped off shell, the scraper a fter 2 minutes ( c.50 s trokes) Steep used for this task.

Hard l imestone was 1 0 minutes ( c.200

.

easily s craped. Edges s trokes) scraping.

were

not

dull

Acutely angled scraper edges used t o cut shell and l imestone worked well and remained largely undamaged ( as opposed to unretouched blades used for this task) : a large ormar shell ( 1.5 cm thick) was cut through in 5 minutes. Fresh cane and reeds were s crapers were not blunt a fter 2 000 The

different

k inds

o f

f lint

s craped strokes.

easily

I had used

s eemed

and

t o

the

work

equally well. 3 .

Results

Unused retouched scrapers showed traces in the form of s tone polish, or f aint antler polish in the case o f pressure-flaking, with striations running in the direction o f the retouch on the ventral aspect o f the s crapers. Although such traces were i solated rather than continuous on the edge, they might well be ( and were, s ee Part I , chapter 2 0) confused with use-wear. This suggested that use o f ancient s crapers could only be a scertained i f the use-wear polish

on

them was

well

developed.

Wood s craping l eft a continuous smooth l ine of polish, brilliant probably due to the amount of amorphous s ilica present ( Part I , Chapter 1 4, P l.8:g) S easoned wood caused l ess polish than did f resh wood scraped f or the s ame t ime ( Pl.8:e and g ). There was hardly any micro-chipping on the ventral a spects of the scrapers. However, while well developed wood polish was f airly easy to identify, there were areas of confusion: on coarse-grained large f lake-scrapers the use was distributed over a long edge and t herefore hardly any identifiable polish could be s een even a fter one hour's u se. The same happened with a small s craper o f the s ame raw material which had been used to plane wood f or 1 0 minutes. In one instance wood polish on a s craper had .

a distribution s imilar to hide polish, s imilar t o bone polish. Wood polishes a lso different a ccording to species: s craping

1 25

in another i t was looked s lightly sycamore l ed t o a

very the l ike

f luid l ooking polish ( Pl.8: g). resulting polish ( flat with l imestone polish ( Pl.9:a)

When short

bark was s craped s triations) l ooked

S craping dry bone caused t he f lattened dry bone polish with narrow deep striations ( see P art I , Chapter 1 4) The scraper edge became not rounded ( in profile) l ike t hat o f hide s crapers, but l ooked very j agged due to t he i ntense microflaking o f the edge ( Pl.1:d) Fresh uncooked a nd c ooked bone a lso caused some edge damage ( although not a s much a s dry bone) and a distinctive striated polish"bevel"(Pl.9:h). Sometimes t his bevel was only s een on a small projecting area of the s craper edge and the only other wear-traces were a generally bright polished area. The s ame was the c ase with a s craper which Miller had used t o s crape meat from bone. .

.

S craping pig hide ( dried and t anned) resulted i n a polish distribution s imilar t o t he"pitted"look u sually a ssociated with dry hide polish ( Part I , Chapter 1 4) S craping f resh hide and dehairing h ide l ed t o a rounded edge with"bumpy"polish and short striations , . Very l ittle microflaking could be discerned ( Pl.2:f) Hide polishes could differ very much a ccording t o f lint t ype: sometimes on a f ine-grained f lint-scraper ( used by Bergman f or 3 0 minutes) a lot of polish was s een, while on a medium-grained s craper used by me very l ittle polish was d iscernible even a fter 8 4 minutes' use ( Pl.10:g). I f ound t his worrying i n view of the fact that even l ess polish might be visible on ancient tools, and that such sparse polish could be confused .

with other weak polishes, i ncluding those f rom manufacture. One reason for this variation might be that polishes c ould vary according to the surface on which t he h ide i s s craped. The scraper used with ochre showed a polish s imilar polish with a f ew more s triations ( Pl.10:e, g)

t o

hide

Scraping dry antler l eft a f ew i ll-defined t races ( see Part I , Chapter 1 4) The polish on my s crapers l ooked l ike the beginning of wood polish but somewhat duller, t hat i s, with probably l ess amorphous s ilica. S craping soaked antler l eft a polish which could be confused with wood polish ( cf. P l.9: g and 8 :e) The edge was a lso rounded and short striations were visible. There was l ittle microflaking ( Pl.2: b) .

.

I ,

Scraping soaked Chapter 1 4)

horn

l eft

only

i ll-defined

S craping concretions o ff shell l eft surface with f ine striations ( ibid.).

t races

a generally

( Part

a braded

S craping l imestone c aused considerable abrasion o f the edge, microflaking and f lattened s tone polish with short f lat striations ( ibid.) ( Pl.1: b, 9 :c).

that

Sawing shell i s bright

with a s craper c aused shell polish ( ibid) patches o f polish with parallel groups of

1 26

s triations

on

( Pl.9:e).

S awing l imestone with small patches o f polish.

a s craper

c aused

t hin

S craping c ane caused a f luid l ooking polish s triations and w ith hardly any edge damage. S craping Chapter damage.

1 4,

reeds

caused

P l.23:g, h)

One s craper was resulting wear-traces polish.

the with

domed no

reed

s triations

without

polish

s triations,

but

( Part

any

I ,

s ome

edge

used on bone, t hen on wood. looked l ike a mixture o f bone and

The wood

In a ll cases speed of polish formation varied with t he f lint type ( see Part I , Chapter 1 7, P 1.5) Macroscopic gloss was s trongest on the plant scrapers and was visible on some wood s crapers. However, I suspect that any hard material when worked long enough would cause gloss ( see Part I , Chapter 1 0) Macroscopic rounding was a lways visible when a s craper had been used for a long t ime, say one hour. However, with some s crapers such long use was not possible. On very hard materials, such a s s tone, s crapers became rounded f aster t han when used on soft materials. .

.

Part 4 .

For hafting t races I , Chapter 1 4.18)

Scrapers

( or

rather

the

l ack

of

them)

s ee

from Arjoune

The s crapers discussed below measured from small ( 1.8 x 1 .9 cm) t o large ( 8.3 x 6 .8 cm) One reason f or t his variation in s ize may have been the scarcity of suitable raw material ( see Part I I, Chapter 9 ). Despite the small s ample s ize I t ried to investigate the relation between s ize ( as well as type and edge angle) and worked material. For types, numbers and percentages s ee Tables 7-10 .

Trenches Of

the

1-1V

two

drawn

s crapers

( one"end-scraper"(Copeland,

1 981,

Fig.11:6, here F ig.14:a) and one"truncated blade-section ( or ? scraper)" ( ibid., F ig.11:16, here F ig.14:b) the former was patinated, the l atter a surface f ind. I therefore l eft both these scrapers out of the discussion a lthough they both showed wear-traces compatible with their use a s scrapers. Trench V 1 04.2 A retouched backed f lake ( 4.5 x 3 .4 cm) with a straight regularly retouched edge forming an angle o f c . 4 5 degrees ( Fig.14:c) without an overhang, s eemed t o have been used a s a s ide-scraper. This was indicated by t he directions o f striations and polish. I t was impossible t o -

‚

1 27

identify and

the

a worked polish

material

a s

the

f lint

was

c oarse-grained,

too weak.

1 14.2 A small blade ( 3.6 x 1 .6 cm) with end-scraper retouch f orming a steep angle of c .80 degrees, but with a s light overhang due t o the natural curve o f the blade ( Fig.12: b) had been used a s a scraper, possibly on stone, a lthough use on t ree bark could not be ruled out ( P1.9: b). Thinning f lakes on t he dorsal proximal end suggested -

‚

hafting. s teep s craper 2 10.1 A hinged f lake ( 4.5 x 3 .2 cm) with ( w ith a n angle retouch on one l ateral and on the distal edge o n t he o ther of c .75 degrees) had a l ess steep r etouch a cute ly a ng led l ateral edge ( Fig.14:d) Apparently the more edge had been used to scrape, possibly on wood -

.

2 20.1 A t abular scraper f ragment ( 5.5 x 4 .0 cm) with s craper retouch forming an angle of c . 75 degrees opposite the break ( Fig.11: b), had patches of gloss on a projection of the s craper edge. The l ack of edge rounding and the polish suggested wood a s the worked material. Unfortunately the patch of polish was very small and the rest of the tool was covered with a general polish, which suggested that the patch might be due to some post-depositional surface -

modification

( Pl.8:h)

Trench VI Surface A "circular f lake-scraper or fan-scraper" ( Copeland, in prep., F ig.19:3) ( 6.3 x 5 .2 cm) had a s light overhang and one edge area more acutely angled ( c. 5 0 degrees) than the remaining retouched edge ( Fig.11:a). Unfortunately the -

scraper was covered with a general polish, probably due to natural agencies, and there was extensive microflaking on the ventral a spect. Nevertheless the l ack o f rounding of the edge and the distribution o f polish on the most acute part of the retouched edge suggested that wood had been scraped ( Pl.8: d) 5 00.1 A l arge"circular f lake-scraper, with butt thinned, or fan-scraper"(Copeland, in prep., F ig.19:1) ( 8.3 x 6 .8 cm) had retouch a ll around ( Fig.15:a) One s ide was more a cutely angled ( c.45 degrees) than the other. Heavy f laking on the ventral a spect which was concentrated on the acutely angled edge suggested e ither resharpening retouch o f the ventral surface or e lse that this edge had been reused a s a chopper ( cf. Coqueugniot, 1 983) or adze ( cf. Gould et a l. 1 971). However, a s the entire ventral a spect was covered with general gloss and polish i t was impossible t o identify the worked material. Thinning f lakes had been removed f rom the proximal end on both a spects suggesting that the t ool had -

.

‚

been

used

in

a haft.

5 00.1 A f lake-scraper f ragment rounded retouched edge which -

1 28

( 4.9 x 4 .1 cm) indicated that

had a i t had

o riginally been a n end-scraper ( Fig.13:h) edge had an angle of c . 5 0 degrees and the .

The r etouched f lake's natural

curve gave i t an overhang ideal f or gripping pliable materials. The polish and relative l ack o f edge damage v isible near t he break l ooked exactly l ike t he wear-traces f rom experimental hide working ( Pl.10:h) S everal thinning f lakes had been detached f rom t he proximal dorsal end, .

s uggesting

t hat

t he

s craper

7 02.3B A s ide-scraper ( 5.0 f ractured b lank ( fig.15: b) and formed an angle of c . 5 5

had been

-

.

hafted.

x 3 .1 cm) had been made The s craper-edge looked degrees. No microflaking

on a f resh could

be s een, but a weak polish blended into the general polish f ound a ll over t he tool. S triations indicated that the tool had been used a s a scraper, but I could not identify the worked material. The distal ventral end had been thinned, suggesting that the tool had been hafted. 7 03.3B

A f ragment

-

of

a f lake-scraper

been retouched t o form a degrees) ( Fig.12:e). Heavy

( 5.0

very acute f laking was

x

4 .6

cm)

had

edge angle ( c.30 visible on the

ventral a spect. Domed polish on the ventral a spect, running perpendicular to the edge, suggested that wood had been s craped. However, there were a lso groups o f thin deep striations running parallel to the edge on both a spects, particularly on the dorsal aspect. This suggested that the s craper had been reused to cut something hard, perhaps stone or shell ( Pl.9:f) As f ormed a tang, the tool .

the breaks opposite t he may have been hafted.

s craper

edge

8 00.1 A small f lake-scraper ( 3.8 t ranslucent f lint, had been retouched

x 3 .0 on both

ends. very

the working edge, was other lateral and the

-

One lateral edge, obviously steeply retouched, while

not the

cm) made o f s ides and both

d istal edges had angles of c . 6 0 degrees and no overhang ( Fig.10:e). No rounding could be s een and the polish looked l ike wood polish ( Pl.8:f). The retouch on the"inactive"edge, t oo sharp on the dorsal a spect to be comfortably held by hand, suggested t hat the s craper had been hafted. 8 01.3 A large"composite denticulate with racloir"(Copeland, in prep., F ig-19:4) ( 6.6 x 4 .1 cm) with pronounced overhang and an edge angle, at its most a cute c . 4 5 degrees ( Fig.13:f) had an edge rounded both in profile and in plan and l ittle edge damage. The polish ( Pl.10:f) l ooked identical to experimental hide polish. Thinning f lakes on the dorsal surface at the proximal end and on one l ateral edge indicated that the s craper had been hafted. -

‚

9 00.1 A small"lustered end-scraper"(Copeland, in -

had been made out o f a s ickle blade. The

f ragment, possibly a reused prep., Fig-18:16) ( 2.6 x 1 .9 cm)

a l ustered t runcated blade, apparently edge ( angle c . 70 degrees) had a s light

overhang ( Fig-15:c) and very l ittle polish which looked l ike experimental hide polish. This suggested that i t had been u sed on h ide, a lthough the possibility that i t had been

1 29

r etouched w ith antler and l eft unused c annot be ruled out, a s the beginning o f h ide and antler polishes l ooked s imilar. The f act t hat a notch had been made i nto t he l ateral edge near t he distal end suggested t hat t he converted s craper had been meant t o be used i nside a haft. Trench VII 1 000.1 A "thumbnail s craper"(Copeland, in prep ( 1.9 x 1 .8 cm) s eemed on c loser i nspection retouched not only on the dorsal but a lso on a spect ( Fig.15:d) This would make i ts u se unlikely. No use-wear polish could be s een. -

.

1 000.1

-

A "composite

scraper

and

' epine'

F ig.23:12) to have been the v entral a s a s craper ‚

borer"(Copeland,

in

prep., F ig.23:11) ( 4.3 x 3 .2 cm) had very s teep retouch ( c. 9 0 degrees) f orming a very jagged edge ( Fig.13: b) Polish indicated that it had been used a s Copeland suggested: the polish on t he s craping edge was too general t o i dentify t he worked material. However, t he steepness of the scraper edge together with the striated rotatory polish o f the projection c lassified a s a borer ( Pl.10: b) suggested that a hard material was s craped and bored, possibly bone. There was no indication that the tool had been used hafted. .

1 005.3 A small"straight-ended end-scraper"(Copeland, in prep., Fig.23:8) ( 2.6 x 2 .4 cm) had s imilarly steep retouch ( c. 8 0 degrees) a s the s craper discussed immediately above ( Fig.13:d) S triations indicated that perhaps bone ( Pl.10:d) had been s craped, a lthough t hey could a lso have been due t o retouch. There was no indication o f hafting. -

.

1 008.3 A f ragment of a l arge f lake-scraper ( 7.0 x 3 .8 cm) ( Fig.10:c), with an edge retouched to f orm an angle o f c . 4 5-50 degrees, had no overhang nor noticeable edge rounding or edge damage. The polish ( Pl.8: b) l ooked exactly l ike experimental wood polish. The break opposite the used edge had probably followed a natural l ine in the stone and could therefore not be regarded a s evidence of hafting. -

Mound C ( Sounding) An"end-scraper"(Copeland, in prep., Fig. .24:8) ( 3.6 x 2 .1 cm) had a very abraded steep ( c. 8 5 degrees) scraper-edge ( Fig.12:d) and showed a polish ( Pl.9: d) l ike experimental stone polish. As only the end was abraded i t seemed l ikely that i t had been used on s tone. The f act that the end was visibly abraded s uggested t hat t he stone polish was not due to retouch. 5 .

Conclusions

The s crapers f rom Arjoune s eemed t o have been used o n h ide, wood and possibly bone and s tone. In many i nstances identification of t he material was uncertain, however.

1 30

A lthough t he s ample was t oo small t o come t o any definite c onclusions, there were indications that s ize o f s craper was l inked t o a particular worked material: l arge s crapers had apparently been used on wood and hide, materials with very extensive s urfaces, while small s crapers had been u sed on bone and s tone ( as well a s on wood and hide, perhaps because l arge f lint blanks were at a premium) G iven

the

small

s ize

o f

t he

s ample

there

were

s ome

i ndications that s craper type was l inked to a worked material: round t abular s crapers s eemed t o have been u sed on wood, end-scrapers seemed t o have been used on hide, while s traight-sided s crapers with s lightly denticulated edges had apparently been used on bone. Edge

angle

and

overhang

seemed

to

be

important:

the

s crapers

w ith the s teepest edges had been used t o scrape hard materials. The s crapers with overhang seemed t o have been used on s ofter materials such a s hide and i ndeed my experiments had shown that the overhang of the working edge i s e ssential for hide s craping. However . , I t ried t o avoid being influenced by this in polish identification. Many t ools had been reused: one s craper had been made out o f a s ickle blade, other scrapers had apparently been reused a s s aws, choppers or adzes. One t ool was used t o scrape and bore, probably the same material. Hafting could only be deduced from c ircumstantial evidence, s uch a s thinning o f the butt, and in one instance by l ack o f comfortable accommodating retouch. The smaller t ools, i ncluding one composite, d id not show any i ndications t hat t hey had been hafted. The s ample individual

was t oo small t renches.

t o

i ndicate

1 31

differences

between

Chapter

1 1

Perforators

-

( P1.11-12,

( Borers,

Piercers,

Drills)

F ig.16-18)

Any pointed f lint piece could be used a s a perforator, t hat i s t o pierce, bore or drill a hole. ( By"boring"I mean hand-use, by"drilling"mechanical use) Usually a f lint p iece i s c lassified a s a perforator i f it has a retouched point on i ts central axis and a more or l ess c ircular cross- s ection. .

1 .

Previous

Research

S emenov ( 1964, pp.74-83) examined perforated shell and s tone objects f rom the Upper Paleolithic and t he Neolithic. He i llustrated ( ibid., f ig.25) various methods of boring by hand, drilling with a stone drill inserted i n a dowel and rotated between the palms o f the hand, and a bow-drill which was accepted a s present at l east by Neolithic t imes ( ibid.); Cauvin ( 1968, p .163) stated that bow-drills at Neolithic Byblos were"pr tiquement prouv e") Unfortunately the wear-traces drawn in S emenov's publication ( fig.25) looked identical, regardless of whether they originated f rom hand-boring or bow-drilling. I did not f ind this in practice .

( see

below)

and

Low power studies o f perforators included that by Tosi P iperno ( 1973) who found residues f rom l apis lazuli

embedded in perforators made from r n icrolithic burin-spalls at Tepe Hissar and Shahr-i-Sokhta. High-power ( Yerkes, 1 983)

microwear revealed

analysis of drilling of

blades

and

Paleo-Indian drills shell, bone, wood and

s tone, probably with a bow-drill, judging f rom t he regularity o f the striations ( ibid., p .511). A f ew drills had been used a s gravers rather than borers ( ibid.). Our own microwear analysis o f 1 5 drills f rom Abu S alabikh ( Unger-Hamilton et a l. 1 987) i s reported in some detail in Part I , Chapter 2 0. The drilled materials were apparently shell, and a material, probably stone, which had been drilled with the addition o f abrasives. The wear-traces were consistent with mechanical drilling. Gwinnet and Gorelick ( 1979) in their SEM study o f ancient drilling methods reported concentric patterns s imilar to those observed by us on t he Abu S alabikh drills. Such concentric patterns were ‚

a ccording to characteristic had been the investigated

Gwinnet of the

and Gorelick addition o f sand

case with our a few m ches

a s

( ibid., p .20) an abrasive, a s

experiments. Keeley ( 1983) de f or ts f rom Neolithic Abu

Hureyra and f ound that some the broken specimens had been used to drill wood, possibly with a pump- or bow-drill, while others the complete specimen had been used t o -

-

-

-

enlarge the holes. However, no comparable experiments were reported nor were any photomicrographs shown a s evidence f or this early Neolithic example of"joinery" ( Keeley, i bid.).

1 32

2 .

Experiments

Twenty piercers and borers and 3 2 hafted drills ranging f rom very l arge t o very small had been made of various types o f f lint ( see P art I I, Chapter 9 ) They were used t o perforate wood ( fresh and s easoned) bone ( cooked, uncooked a nd dried) antler ( soaked and dried) hide, soaked horn, pottery f rom Arjoune, various types of stone, various types o f shell, and cane. Some piercers were a lso used t o groove wood, bone, pottery and s tone. Bergman and I had used a bow-drill t o drill the above materials, and a lso l apis l azuli, copper, malachite and hippopotamus"ivory" which had not been found at Arjoune. Lapis lazuli and copper were drilled, and stone and wood were bored with the addition of abrasives, including sand on i ts own, and sand and water. In a ddition a "leather awl"used by Miller, and a piercer used to groove soaked antler by Newcomer were examined. .

,

,

,

Grooving

with

piercers

proved

an

easy

t ask.

For

i nstance the t ip of a piercer was still intact a fter 3 0 minutes when a fresh deer metatarsal was grooved in order to make a bone awl after the method i llustrated by Poplin ( 1974, Figs.7-11). F ive holes were pierced into a relatively f resh h ide which had been moistened with water. This was difficult: the h ide stretched over the piercer t ip which the retouch had somewhat blunted. I t would have been better e ither to pierce dry hide or e lse use a much sharper t ip. Boring holes in f resh wood was, not surprisingly, easier than in s easoned wood. For wood a relatively broad t ip could be used. Tips were still intact a fter boring for 1 5 minutes. I t took 1 0 minutes to make a biconical hole through a wooden plank o f 3 cm thickness, and half that t ime t o make a s imilar hole through fresh wood.

was

Bone could a lso be bored, easily crushed or broken.

Boring through the Horn

even when dried,

soaked antler was easy: eight f lat t ip of a reindeer antler

could

be

perforated,

but

the

holes in 1 0

inner

but

the

t ip

were drilled minutes.

casing was

very

hard.

which

pot

Limestone was much easier perforated the t ip broke immediately.

It t ook sherd.

5 minutes

to

bore

a biconical

Two holes were bored through t hickness in 5 minutes with a large was so blunted t hat a third hole could

1 33

than quartzite,

hole

through

a shell of bec. However, not be made.

0 .2 the

on

a

cm t ip

t he

F ive t hird

holes hole.

were

bored

i nto

cane.

The

t ip

broke

Boring wood and l imestone with t he addition o f not make the t asks easier: i t was l aborious to f eed into the hole.

a fter

s and d id t he s and

Bergman and I ( Unger-Hamilton e t a l. 1 987) u sed a bow-drill made by Bergman on a ll t he materials l isted above, which were f astened to a bead-board usually with b itumen. The drills were hafted i n wooden dowels with r esin and s inew, and in some instances with wooden splints and s inew. The bow-drill was easy t o handle and very hard materials could be drilled quickly. Most of the f lint t ips remained intact. Mechanical drilling was a lways more e ffective and easier than boring by hand. I t took f or i nstance 1 2 minutes to make one hole into l apis and 1 0 minutes to make t hree holes in hippopotamus"ivory" ( see Unger-Hamilton et a l. ibid.). Drilling with s and and water was somewhat easier than with sand a lone as the abrasive s tayed in p lace. The addition of abrasives did not s eem greatly t o i ncrease ‚

‚

efficiency of the drilling process, except perhaps when soft stone was drilled ( see Hodges 1 976, p .107) Bergman found that l arge drills were difficult t o haft and use, and that a great amount of heat was generated during t he drilling o f wood with a mache de forst. ‚

f lint 3 .

.

I t s eemed that, when used to perforate, coarse-grained broke more easily than f ine-grained f lint.

Results

Polishes and striations f rom grooving and rotary movements could be differentiated by t heir d ifferent distribution. In the former case polish and striations were confined t o the contact a spect and ran a long t he longitudinal axis of the tool t ip; i n the l atter c ase t hey were f ound on the s ides o f the t ool t ip a s well, a nd were perpendicularly oriented to the tool's long axis. The edge damage distribution was a lso s imilar but a s with t he -

polish and striations retouch t races.

-

could

be

difficult

Hide piercing l eft l ittle in the way except a s light polish at the tool t ip.

to

o f

i solate

f rom

wear-traces,

Boring and mechanical drilling could be differentiated. Boring o ften involved considerable edge damage but l ittle polish and s lightly random s triations ( Pl.11:a), while drilling caused hardly any edge damage ( apart f rom complete breaking o f the t ips), but s trong polish, o ften rotational s triations ( Pl.12:c, 2 9: g) G loss was s een on borer t ips ( the exception being wood borers) .

,

i t was s een on drill s tone, malachite or i nstances

impossible

t ips used copper, to

identify

1 34

t o or

drill hard wood. I t

the

worked

with rarely while

materials was i n

materials

on

l ike s ome the

p iercers and borers, a s t he a reas used were s o small. A c omplicating f act was t he retouch l eaving i ts own wear-traces perpendicular t o t he t ip axis, a lthough t hey were minimal i n the case o f t he antler hammer ( Pl.6:f) O ften t he t ips o f hand-held perforators used on hard materials had broken. There were some distinct polishes on hand-held perforators, e .g. f rom wood ( Part I , Chapter 1 4, P 1.11:a), s tone ( ibid.) and pottery ( ibid.). However, a l arge number o f striations, probably f rom the f lint particles, were s een on a ll borer .s, regardless o f the worked material. This made i t impossible t o differentiate shell f rom bone polish, a lthough bone s ometimes but by no means a lways caused t he characteristic bevel ( ibid.). Antler could look l ike wood polish, while hide and horn l eft only weak traces which on ancient tools could be mistaken f or post-depositional t races. Macroscopically, tool t ips were found t o be r elatively intact when hide or wood had been perforated; t hey were splintered when bone was worked, and abraded by l imestone and shell. -

-

Mechanical drilling l eft s tronger t races which ( with t he exception o f pottery and malachite polishes) looked d istinct a ccording to the worked materials ( Pl.12;a,c, 2 9:e, h) They matched the general descriptions of polishes ( Part I , Chapter 1 4). Of the polishes only briefly mentioned t here, malachite polish l ooked l ike that f rom pottery, presumably a s the materials are o f s imilar consistency, copper polish looked unmistakably smooth and f lat ( Pl.29:e) and hippopotamus " ivory"caused a corrugated polish with l ots o f t iny pits ( Pl.29: h) .

Drilling with abrasives l eft distinct wear-traces which were a lways dominant regardless o f t he worked material; drilling with s and a lone l eft a roughly abraded surface, while drilling w ith s and and water l eft a concentric r ing pattern

( Pl.29:g)

In a ll cases,the more yielding the material and the s teeper angled t he tool t ip, t he more penetration o f t he material was a chieved, and the more invasive were the wear-traces down t he t ool t ip. Obviously, thickness o f the material was a lso important: a drill would penetrate much deeper into a wooden p lank o f 5 cm t hickness than into one o f 1 cm thickness.

1 35

4 .

Borers,

P iercers,

S ee piercers

Tables 7-10 and drills

Trenches

1-1V

Drills f or

from Arjoune

numbers

and percentages

o f

borers,

Two of the drawn perforators, a "drill-bit"(Copeland, 1 981, F ig.11:1) and a "borer"(ibid., F ig.11:2) a lthough surface f inds, had what s eemed t o be well preserved wear-traces. The drill-bit ( Fig.16:i), with what l ooked l ike a broken and subsequently abraded t ip, had c lear rotary traces, l ike those f rom s tone drilling, down t o a few mm from the t ip. This suggested that t he tool had been used to drill stone ( P1.12:h). The " borer"(Fig.18:a) had a s lightly abraded intact t ip with l ittle edge damage, a nd polish and rotary t races indicating that i t may have been used a s a wood drill. ‚

Trench V I could only burnt or wear-traces.

f ind had

two perforators which broken t ips, or e lse

were no

not e ither identifiable

3 13.2 A tool made on a truncated b lade had an intact ( Fig.18: b) Judging from the striations and polish, i t probably been used a s a wood drill. -

.

4 03.2 A t ool with used to drill wood, out ( Pl.12:f). -

t ip had

a f lattened t ip ( Fig.17:e) may have been a lthough use on stone could not be ruled

Trench VI 5 00.1 A"borer ' bec' type on retouched f lake"(Copeland, in prep., Fig.20:2, here F ig.16:e) had c lear wear-traces on the"bec" indicating t hat it had been used a s a borer. The microscopic bevel ( Pl.11:f) at the t ip was more typical of wood than of bone boring, as was the t ip which was quite intact. The shape o f the tool suggested that i t had been used by hand. -

7 00.1 A large"borer, drill-bit type"(Copeland, ibid., Fig.20:1, here F ig.17: d) had a very abraded t ip and wear-traces consistent with use in a fast rotary drill, probably to drill stone, although wood could be a possibility ( Pl.12: g). Bergman reported ( pers.comm.) that a copy of this tool could only be hafted with great difficulty. -

7 00.1 A borer with an intact t ip ( Fig.16: b) had very l ittle polish and f ew s triations consistent with wood boring ( Pl.11:b). The irregularity of the proximal end s uggested that the tool had not been hafted. -

1 36

7 01.3B A l ong F ig.21:3, here -

t hin"borer, F ig.17: b)

‚

drill-bit type"(Copeland, ibid., o ften referred t o a s a m ' eche de

f or t had an i ntact t ip and wear-traces c onsistent with t he drilling of wood in a f ast rotary drill ( Pl.12: b, d). The f act t hat the wear-traces affected the t ip down t o 4 .2 cm was surprising a s Bergman ( pers. .com r n.) found that he could not drill wood very deeply with a copy o f this tool. ‚

7 03.3B A s r nall"borer, piercer type"(Copeland, ibid., F ig.20:3, here F ig.18:c) had a somewhat abraded t ip, l ittle e dge damage and very weak polish which could be hide polish. S triations indicated t hat the tool had been used t o pierce a s well a s bore. -

8 00.2 A short projection on appeared t o have been used to f rom the polish ( P1.11: d) and -

a t runcated blade ( Fig.16:d) incise wood or antler, judging relative l ack o f edge damage.

9 00.1 A small perforator with an intact t ip ( Fig.18: d) probably been used on wood, judging by the domed polish. s trong development of the polish and t he small s ize of t ool suggested that it had been used in a haft. -

had The t he

Trench VII 1 000.1 F ig.23:13, -

A"borer, piercer here Fig.16: g)

type"(Copeland, in prep., had apparently been u sed a s a

borer. Brown earthy-looking inclusions and d iffuse polish at t he t ip suggested that pottery may have been bored, a lthough post-depositional traces could not be ruled out ( Pl.11:h). 1 000.1 borers used to 5 .

-

A small borer ( Fig.18:e) had d iscussed immediately above perforate pottery.

t races s imilar to and may a lso have

t he been

Conclusions

I t seems t hat of the few perforators with relatively clear wear-traces most had been used on wood. This may however be due to the f act t hat hard materials, such a s s tone, shell or bone t end t o l ead t o t ip breakage, while relatively soft materials, such a s hide, l eave very weak traces. It i s a lso possible that antler polish could have been mistaken f or weak wood polish, but a s no worked antler had been f ound on t he s ite a t Arjoune, though unworked antler was found, this was unlikely. Other materials apparently perforated were s tone, pottery and perhaps hide. I t seems that there i s c lear evidence f or the use o f drills, i n a mechanical drill, possibly a bow-drill. Copeland ( in prep.) had suspected a s much. The drilled materials had apparently been wood and stone. It s eems that at least a t Arjoune the m 'eche de forst had been a wood drill a s i ts name suggested.

1 37

There

was

no

evidence

f or

t he

use

Hafting had t o be inferred from t ips and t he shape o f t he haft. F inds f rom macehead,

the s ite suggested

at t he

o f

the

Arjoune, use o f

abrasives. wear-traces

such t hick

on

t he

t ool

a s a perforated drills. However

Bergman reported difficulties with s uch thick drills ( see above, S ection 2 ) Further experimentation i s needed. I t i s possible that hollow drills ( Semenov, 1 964) perhaps of c ane, may have been used. Other perforated objects i ncluded .

,

a shell

pendant.

1 38

Chapter A

-

1 2

Blades,

-

Retouched Blades, Bladelets

( P1.13-15,

Bladelets

and

and Flakes with Lateral

1 6:a-d,

F ig.19-24:a-d,

Flakes Retouch

2 5-27:a-h)

A f lake i s a ny e lement which i s s truck off a core. A b lade i s a f lake with parallel edges and ridges. A b ladelet i s s imilarly proportioned and i s l ess t han 1 .2 cm wide ( Tixier, 1 974) Only implements with retouch on one or both l ateral edges are discussed, a s these are most l ikely t o have been used. Lateral r etouch, which may be shallow, s emi-abrupt or abrupt, may e ither provide backing o f t he t ool, that i s, the hafted or hand-held edge, or e lse i t could be the working edge. B lades or f lakes with snapped ends are discussed s eparately i n Chapter 1 2 B . .

1 .

Previous Research

There has been a debate whether steeply retouched edges, usually t hought to be backs of blades, may have been used a s rasps or whether one would cut oneself on t he s harp edge ( see Moss, 1 983 ‚ pp.43-46) In my opinion i t i s possible t hat the sharp edge was inserted in a haft. Retouch could be created a ccidentally, e . g. during knapping ( Newcomer, 1 976) or e lse during use ( Part I , Chapter 4 ) Various uses could be envisaged f or blades and f lakes: Moss ( 1983, pp.43-46) discussed research in which such t ools had .

been c lassified a s s crapers, knives, saws, and in the case o f bladelets a s barbs and projectile points. Wilmsen's hypothetical relation o f angle o f the working edge t o f unction suggested certain optimal edge angles f or s crapers ( see P art I I, Chapter 1 0) a cute angles ( 26-35 degrees) f or ,

cutting tools of soft materials, and wider angles ( 46-55 degrees) for the cutting o f harder materials, a nd he apparently a ccepted S emenov's optimal angle ( 35-40 degrees) f or whittling knives. Such edge angles were l argely confirmed by the ethnographic evidence ( Gould et a l. 1 971) S emenov stressed t he importance of meat knives ( 1964, pp.101-113) f or hunting a ctivities, but Frison's experiments ( 1979) and Gould et a l. ' s ethnographic observations ( 1971) suggested that sharp f lint f lakes need only be used during butchering t o cut t he skin and l igaments of the animal. S tone choppers, bone t ools, wooden wedges and logs could be used t o butcher the carcass. According to Gould et a l. ( ibid.) aborigines small f lake knives in c ircumcision r ituals and l arger knives f or a variety of t asks including domestic ones.

used f lake

Retouch o f t he cutting edge varied: Gould et a l. ( ibid.) reported that f lake knives used for day-to-day a ctivities were mostly d iscarded a fter a f ew uses, and not resharpened. Escalon de Fonton ( 1979) resharpened blades by

1 39

denticulation during wood working experiments, a nd a fter f urther work the denticulated edge became s traightened by t he breaking of t he t eeth. I f ound that denticulation did not make a blade sharper but a lways prolonged the l ife o f the b lade considerably ( see P art I I, Chapter 1 7) A number o f microwear s tudies have been c arried out on b lades and f lakes. The r esults o f the low-power s tudies, concentrating on edge damage a re discussed in P art I , Chapter 4 . High-power s tudies included t hose by Vaughan ( 1981, e . g. t able 5 0), Bueller ( 1983), and Moss ( 1983) who reported t hat at P incevent b lades had been used t o butcher and cut meat, cut and bore hide, and that obliquely t runcated blades had been used to groove bone ( p.132) Moss ( ibid., pp.43-46) referred to t he hafting o f b ladelets, and I have examined wear-traces on blades f rom Kebara and E l Wad, s haped exactly l ike the s ickle blades f rom t he same s ites, but with wear-traces f rom materials other than plant. .

Backed bladelets had probably been used a s barbs and projectile points ( Moss, 1 983, p .115; Barton and Bergman, 1 982) Real examples o f such use, i .e. complete Egyptian pre-dynastic and dynastic arrows, were shown by Clark et a l. ( 1974) .

The difficulty in a ssigning wear-traces on cutting t ools to a specific worked material has been discussed in Part I , Chapter 2 0. I believe that this i s mostly due to the f act that the hardness and surface t exture o f the material cut i s largely unknown to the microwear-analyst, and a lso that a long t ime i s required f or polishes to develop on such implements, except with some worked materials such a s wood or some plant species. 2 .

Experiments

Ninety-one unretouched and denticulated blades and f lakes had been used on wood, bone, h ide, antler, horn, meat, f ish, stone, pottery, f eather, wool, shell and root vegetables. ( For experiments on s iliceous p lants see Part I I, Chapter 1 7) The tools had been used t o cut, s aw, whittle ‚ scale ( fish) and to groove ( hard materials). I used a ll the t ools unhafted, but a lso studied wear-traces on blades and bladelets which had been used by Newcomer, i nserted parallel in groups of three in hafts, to cut meat and wood. .

I n to cut used by

addition

I examined

meat and Bergman a s

f lakes

and

blades

s cale f ish, and s everal arrowheads and barbs.

used

by Miller

backed

bladelets

Experiments sawing wood showed t hat i t was a lmost impossible t o make any deep incisions i nto s easoned wood. Fresh wood could be i ncised down t o a f ew mm deep, but blades had to be denticulated a fter 1 0 minute's u se. When denticulated, they became blunt a fter 4 5 minute's use on f resh wood. The denticulated edge broke e asily and great

1 40

c are

had

t o

be

t aken

during

s awing.

Only

t hin

twigs

could

be cut t hrough. The ease with which bone was sawn depended apparently not so much on whether bone was cooked, dried or f resh, but on the type o f bone and animal: thin chicken l eg bones could be s awn with ease, while cow l eg bones or t he e piphysis of a roe deer metatarsal could hardly be cut a t a ll. On most bones cuts could only be made to a few mm deep. B lades were b lunt a fter 1 000 sm, and the t eeth o f denticulated blades snapped o ff. With antler t he ease o f sawing depended on whether the antler was soaked or not. Soaked antler could easily be sawn ( although the deepest cut I could achieve without the blade breaking was only 2 cm) while dry antler was a lmost i mpossible to cut; edges were worn out a fter 5 00 sm. A r etouched cutting edge worked better than an unretouched one. ,

Hide cutting was easy, e specially when the hide was dry. However an unretouched blade was blunted a fter 5 m inutes. Horn soaked for four hours was a lso quite easily cut a lthough i t would have been better t o soak the horn l onger a s the inner casing was hard. To cut meat one needed a very acutely angled blade and i deally a long blade, or e lse s everal blades hafted t ogether. Other t asks ( all easy) carried out were cutting f at off hide, t endon off bone, and cutting through dried s inew. S everal k inds of f ish were gutted and t rimmed. Limestone was incised with a blade: a s craper edge had been more e fficient for this t ask ( Part I I, Chapter 1 0) The s ame was the case with shell: two f ine incisions each 3 cm l ong, were made with a f lake in 5 minutes. However, a fter t hat the edge was blunt, and a thin scraper or burin edge proved better at cutting through the shell. .

‚

Sheep's wool was cut with an unretouched f lake. However, this was a very l aborious task and the f lake was not sharp enough. Ryder ( 1983) s tated that woolly sheep were only bred when wool shears could be used. From my experiments I would agree with Ryder. Altogether f or cutting i t was best to use very a cutely angled b lades with edges s traight both in section and in profile. S oft materials were easiest cut with unretouched edges, medium-hard materials with denticulated edges very hard materials with edges with s craper retouch. Whittling wood and s caling f ish a lso had to be done with acutely angled blades. Grooving f ine incisions i nto hard materials was easy, but edges broke quickly. ‚

1 41

3 .

Results

Wood s awing ( e. g. a fter 2 000 sm) l eft macroscopic g loss, r elatively l ittle a nd rounded edge damage, a nd t he continuous, bright wood polish, e ither gently domed or very f luid l ooking ( Part I , Chapter 1 4, P 1.13:c) Very t hin edges broke away completely, t hus l eaving hardly any polish even a fter 3 0 minute's use. In t his case striations were not i n s awing direction, but rather i n t he d irection o f t he microflaking. On coarse-grained f lint very l ittle and i solated polish with striations could be s een, a lthough because of i ts continuity, the polish l ooked l ike wood polish. .

Bone polish could vary considerably: uncooked a nd f resh bone l eft a general bright area o f polish w ith a f ew s triations, while dry bone could l eave l ittle polish ( Pl.15:c), but in some instances striations. S tepped microflaking was the most characteristic f eature o f bone s awing ( Pl.1:c). Chicken bones seemed t o l ead invariably t o a special polish with i nflated l ooking polish s treaks ( Pl.29:f) Polishes f rom the cutting of h ide ( Pl.14:g) meat ( Pl.14:a,c) and carrots ( Pl.16: 0 l ooked very s imilar t o each other: in each case a general bright area o f polish with some s treaks was visible. S imilarly vague t races were generated by cutting dry s inew ( Pl.16:b) cutting f at f rom ‚

‚

hide, t endon f rom bone ( Pl.16:a), and f eather ( Pl.15:f). None o f these wear-traces could be detected on coarse-grained f lint and none would in my opinion be distinguishable on excavated t ools with soil sheen. Sawing dry antler generated some i solated a reas o f f lat polish ( Pl.15:h) with s triations, which could look l ike wood polish. S imilarly soaked reindeer, red deer and f allow deer antler c aused a continuous inflated-looking polish ( Pl.15:g) akin to t hat o f wood. However, gloss was never a s obvious a s that caused by contact with wood. Fallow deer antler s eemed t o cause regular shallow s triations not s een ( by me) on blades used t o s aw reindeer antler. Cutting f ish l eft a somewhat f latter and o f polish than did meat. Nevertheless, I doubt polishes could be detected on ancient tools. Wool l eft s triations and

f lat

a narrow band l ots of small

L imestone l eft striations.

t he

Shell l eft bright groups o f s triations. Whittling wood

( and

f lat

brighter band whether these

o f polish ( Pl.14:e) edge damage s cars. polish

patches

soaked

1 42

o f

( P1.15:a)

polish

antler)

l eft

with

with

diffuse

( Pl.9:e)

very

f ine

with

rounded

edge damage and i nvasive polish on t he contact a spect, wood polish i n t he f ormer case, domed and s triated polish in t he l atter c ase. S easoned wood l eft very l ittle f lat polish a t t he edge. S awing and whittling wood with t he s ame edge l eft vague t races o f buoyant polish apparently spread in both d irections o f use ( Pl.13:e) Fish bands o f edge, 1 964, f ish,

s caling polish

l eft in some instances a pattern o f ( Pl.15:e,29:d) running diagonally

presumably where the s cales had touched p .107) Unfortunately this happened l ike gurnet, and not with bream. .

( cf. only

bright t o t he

S emenov, with some

B ladelets u sed a s arrowheads and barbs were often s everely shattered and damaged ( see"Previous Research"in P art I I, Chapter 1 5) Wear t races t ended t o consist o f striations running in various d irections ( Pl.20:g), presumably depending on how the bladelets had hit t he carcass, been deflected in it and been removed from i t. B ladelets shot i nto of sand polish and 1 5)

was held

sand by Newcomer had very bright jagged edge damage ( see Part I I,

specks Chapter

Grooving caused specific polishes, but the used area so small t hat i t could be missed on ancient blades, or to be a ffected by post-depositional t races. Altogether,

the

fact

that

wear-traces

on

blades

used

t o

cut meat, f at, s inew, hide, horn, carrots, and to a certain extent wool and f ish, were very weak or non-existent made me very pessimistic about the detection of such t races on ancient b lades. B lades used t o cut hard materials, such a s dry antler, bone or shell, were subject to a lot o f breakage, and again this meant t hat worked materials could not be i dentified on ancient blades. Antler and wood polishes on blades looked sometimes s imilar while bone polishes showed t remendous internal variability. In fact t he blind t ests ( Part I , Chapter 2 0) had demonstrated t o me that when blades are used for only 1 0 minutes, the chances o f identifying the worked material s eemed higher by examining t he blades with the naked eye than by microscopic examination o f weak or virtually indistinguishable wear-traces. No polishes

b lades were

hafted blades might suggest

were very

examined f or hafting traces, but evenly distributed a long the edge on

( used by Newcomer) and such t hat a blade had been hafted.

Blades, Bladelets Arj oune

and

Flakes

with

Lateral

a

distribution

Retouch

from

S ee Tables 7 -10 f or numbers and percentages o f retouched b lades, bladelets and f lakes at Arjoune. The tools discussed below had been c lassified ( by me) a s unlustered a fter the initial macroscopic examination. In some instances

1 43

c loser i nspection revealed a weak l ustre. blades' are discussed i n P art I I, Chapter 1 7 Trenches

( Lustered"sickle

. )

1-1V

The wear-traces on only two unlustered b lades could be identified a lthough the fact that they were surface f inds lessened the degree o f certainty. On a "?sickle blade ( no sheen)"(Copeland, 1 981, Fig.10:10, here F ig.19:d) from Arjoune I S urface showed a f ew spots of wood or possibly reed polish ( Pl.13:d) . The l ack of g loss may have been due t o the considerable resharpening retouch. A "crescent-shaped backed ? sickle-blade ( no sheen)"(ibid., F ig.10:12, here F ig.25:a) a lso f rom Arjoune I Surface showed s tone polish and striations in the cutting direction. Trench V Baulk 1 01-401 A retouched bladelet ( Fig.25:b) was burnt, an important observation in view of t he f act that only small tools which had probably been hafted had been heavily burnt ( see Partli, Chapter 1 8) -

1 02.2 A small t runcated blade ( Fig.25:c) l ooked s ickle-element, but wear-traces consisted o f a weak polish suggesting that hide or meat had been cut. -

1 04.2 A small regular rounded polish consistent of plant polish. -

1 04.3 A l arge ( Fig.25:e) had -

t runcated backed edge damage at with wood sawing,

l ike band

a o f

blade ( Fig.25:d) with the cutting edge showed or perhaps the beginning

i rregular blade with shallow backing retouch a cutting edge which on c loser inspection

revealed a weak gloss. The wear-traces on t his edge were consistent with wood sawing. Antler a s a s awn material was considered unlikely ( see Part II, Chapter 1 1) 2 00.0

-

A crescentic-backed

blade

( Fig.25:f)

with

a

series

of small notches confined to the middle sharp edge had been apparently used with a t ransverse movement, probably t o whittle wood. P erhaps the notches had been made to s harpen a very f ine point. Baulk 2 01-210 A crescentic backed blade ( Fig.22: b) shaped l ike a s ickle blade revealed rounded edge damage and a quantity of f ine s triations parallel t o the cutting edge, on a weak polish. Perhaps pottery or e lse a f ibrous material -

had

been

cut

( Pl.14:f)

2 01.2 A l arge blade ( Fig.20: b) snapped a t one end, with the distal part shaped l ike a tang, and t he proximal part backed l ike a crescent, had a f inely denticulated cutting edge. The rounded edge damage and t he polish were consistent with wood cutting/sawing ( Pl.13:f) . The presence of the tang suggested that the blade had been used hafted, a lthough -

1 44

polish a nd retouch t he very end.

o f

t he

cutting

edge

was

f ound

down

t o

2 02.2 A very s mall unretouched bladelet snapped a t the d istal end ( Fig.25: h) had s triated weak wear-traces ( rather l ike MLIT, see P art I , Chapter 1 4) a long i ts long axis, and a lso general polish on i ts dorsal r idges. I t may have been u sed as a projectile or possibly t he t races might be post-depositional. -

2 02.2 A very small backed bladelet c ompletely covered with polish. The shape may have been used a s a barb. -

2 11.2 A large f lake ( Fig.26:a) i nversely retouched cutting edge had a wood saw. -

( Fig.25:g) was suggested t hat i t

with a t ang apparently been

and used

an a s

Trench VI Baulk 6 00-700.3 A naturally curved ( in plan) t runcated b lade ( Fig.26:b) with a serrated edge on the distal part had been used a s a s aw on medium-hard material, judging by the o rientation of t he striations and f rom the polish. However, t he f lint was too coarse to be more precise. -

Baulk 6 00-700.3 A denticulated blade ( of very coarse f lint) with a t hinned back ( Fig.26:c) t runcated or snapped a t both ends, had no polish on the denticulated edge, but t he s tepped edge damage and the polish on the a cutely angled edge could be consistent with bone sawing. -

‚

Baulk 6 00-700.3 A long, naturally backed blade showed rounded edge damage, and a thin band o f striations parallel to the cutting edge. I t may -

u sed short

on use

a on

soft wood

Baulk

7 00-800.3

material, could not -

A "backed

( Fig.21: b) polish with have been

such a s meat ( Pl.14: b), be ruled out. t runcated

a lthough

blade-segment

with

' hook'" ( Copeland, in prep., Fig.18:15, here F ig.26:d) had small patches o f polish on the ' hook', s imilar t o that f rom s easoned wood, with striations parallel to the tool's l ong axis. The l ack o f edge damage on the hook together with the polish suggested t hat wood had been incised. 7 00.1 A backed t runcated e lement ( Fig.20:c) with s emi-abrupt retouch on the cutting edge d id not have much polish, presumably because of resharpening. I t had probably been used a s a wood saw ( Pl.13:h) -

7 00.1 A small t runcated e lement ( Fig.22:c) shaped l ike a s ickle-element had a notch on the cortex s ide which had wear-traces, perhaps f rom use. I t had a wide band o f weak polish on the opposite edge, part o f which had broken away. This polish may have been due t o t he cutting o f a soft material, perhaps hide or meat ( Pl.14:h) -

1 45

7 01.3A broken

The t ip of a backed knife ( Fig.20:d) had p robably a ccidentally, a s i ts shape was very s imilar to a

-

complete knife f ound in 8 01.3 ( see below) F rom the rounded edge damage, the domed polish on both a spects, t he s light macroscopic g loss and the direction o f the striations i t appeared t hat the knife had been used t o whittle wood .

( Pl. 1 3 : g) 7 01.3B cutting bone

A backed ( Fig.23: d) edge had been used

-

f lake w ith a heavily r etouched t o cut a hard material, perhaps

( Pl.15:d).

7 02.3B A c rescentic backed bladelet ( Fig.27:a) regularly notched-looking retouch had, j udging polish and deep striations, been used t o saw -

with a f rom i ts a hard

material. 703.3B A "blade f ragment with nibbled retouch"(Copeland, i n prep., Fig.21:7, here F ig.19: b) had a weak g loss a long i ts straight edge, and had apparently been used to s aw wood -

( Pl. 1 3 : b) 8 01.3 A long backed knife ( Fig . .27:b) with a snapped d istal end had an a cutely angled lateral edge with rounded edge damage, s light gloss and polish on both a spects, though more -

i nvasive suggested

on the ventral a spect. S triations and polish that this had been a wood-whittling knife l ike the

s imilarly

shaped

tool

8 01.3 A denticulated

small edge

suggest

s awing.

-

wood

f rom

701.3A

( see

above)

truncated blade ( Fig.27:c) with a had a weak domed polish which could

8 03.3 A "backed knife"(Copeland, in prep., F ig.20:8, here Fig.21:d) had very l ittle edge damage a nd only general polish a long the cutting edge, and had perhaps been used t o -

s lice

meat

or

hide

( Pl.14: d).

Trench VII 1 001.3 A crescentic"backed knife"(Copeland, i n prep., F ig.23:6, here F ig.27:d) had l ittle edge damage, some domed polish on the ventral aspect, some f lat polish with f ine striations on the dorsal a spect, and may have been used t o s aw a medium-hard material. -

1 002.3 A backed t runcated striations on a very thin -

used

t o

1 007.2

cut -

a f ibrous A

b lade

bladelet ( Fig.27:e) had band o f polish and may have

material,

such

( Fig. .27:f)

a s

hair

w ith

or an

f ine been

wool. unretouched

f resh-looking edge and a f inely denticulated edge showed a f ew edge damage s cars and a domed polish on the denticulated edge. I t was probably a wood s aw.

1 46

1 007.3 A crescentic backed blade ( Fig.27:g) with a r etouched cutting edge had deep s triations and i solated f lat polish, and had probably been used on a hard material, perhaps s tone. -

1 007.3

-

The

t ip

o f

a b ladelet

( Fig.27:h)

had

been

burnt.

Mound C ( Sounding) A naturally backed ( Fig.23:b) f lake had considerable edge damage and a quantity o f deep striations, i ndicating that shell or stone had been s awn ( Pl.15: b) -

5 .

Conclusions

The identification of worked materials f rom blades and f lakes had, a s expected ( Part I , above) which

wear-traces Chapter 2 0,

on and

proved t entative or impossible. Most o f the b lades on it did s eem possible to identify wear-traces showed

evidence o f wood s awing or whittling. The f act wood working ( and plant working, see Part I I, could be i dentified, may be due to the f act t hat softer materials are too weak to be recognised, ( with the wear-traces) tend to break away materials are cut.

t hat only Chapter 1 7) t races from while edges when hard

There was evidence that soft materials ( perhaps hide, meat, horn or f eather) and f ibrous material ( perhaps wool) had been cut with truncated e lements shaped l ike s ickle-elements. There was a lso evidence for cutting harder materials ( perhaps pottery, s tone, bone or shell) but it was impossible t o i dentify the exact worked material. ,

The hafting of tools l ike wood-saws and blades used on soft materials could only be deduced indirectly f rom the shapes o f the blades. I t appeared that most hafted e lements had been inserted e ither into a straight or a curved handle, j udging f rom the shape of the backing. A f ew " tanged"tools may have been hafted at the end. A lot of probably

the materials sawn or whittled at been perishable materials, such a s

Arjoune had meat or wood.

However there was evidence, in the form of pebble f igurines, incised l imestone plaques, shell amulets, bone objects and pottery disks, that these materials could have been s awn or i ncised with blades or f lakes. The evidence for the use o f bladelets a s barbs or arrowheads was mainly based on t he shape o f the bladelets and cannot be s een as c ertain.

1 47

B

Snapped

-

Blades

( Pl.16:e-h, 1 .

Previous

F ig.24:e-g,

F ig.27:i)

Research

Bergman et a l. ( 1983) indicating t hat blades had presumably in order to increase

reported c ertain f eatures been deliberately s napped, t he number of r ight-angled

working edges and a lso to create particularly strong edges. The microwear examination ( ibid.) suggested that the breaks on blades f rom Hengistbury Head had been used t o work bone. B eyries and Inizan ( 1982) reported that t runcated blades f rom the Capsian had been used to s crape wood and bone. 2 .

Experiments

b lade

and Results

Experiments showed t hat t he mere could l eave a microscopic polish

act of a long the

breaking a break which

could be confused with a scraping polish on account very f ine s triations running perpendicularly t o the The breaks were best used exactly l ike t hose on s crapers 3 .

Snapped Blades Only

to s crape ( Part I I,

of t he edge.

and polishes Chapter 1 0)

looked

from Arjoune

3 snapped

blades

were

examined.

Trench V 1 02.4 A snapped blade ( Fig.24: g) had a l ateral unretouched edge which showed c lear s triations perpendicular to t he edge, suggesting that a hard material had been c ut. The snapped end had a bright l ine of general polish a long t he -

edge, with random striations suggest t hat the snapped end

( Pl.16:f). Such had been used.

traces

d id

not

1 03.3 The break on an unretouched blade ( Fig.27:i) showed polish and striations, the direction of which suggested t hat the snapped end may have been used to scrape. The denticulated lateral edge had rome general polish suggesting that it had been used. -

Trench VI 9 00.1 with

-

an

The ventral edge of a break ( on a blade ( Fig.24:e) i rregularly retouched l ateral edge) showed t he only

t races possibly due to use on this implement. Rounded polish and striations running perpendicular to the edge s uggested that the break had been used t o scrape wood ( Pl.16:h). The t ip at t he break may have been used.

1 48

4 .

Conclusion

The end o f two may have been u sed wood.

snapped b lades ( out a s s crapers, in one

1 49

o f three examined) i nstance possibly o f

Chapter

1 3

-

Notches

( P1.17-18,

and Denticulates

F ig.28-32)

Introduction A notch i s a f lake or blade with a "notch"made by one or s everal blows f rom the dorsal or ventral a spect. A denticulate consists of s everal such notches on t he same edge. Relatively f ine denticulation can occur s imply a s retouch t o resharpen or strengthen a cutting edge and i s discussed in Part I I, Chapters 1 2 and 1 7. The d ivision between denticulates and denticulated retouch i s a rbitrary and purely based on s ize o f the notching.

Notches, unless present t o f acilitate hafting ( Cauvin, 1 968, p .170) or handling ( Semenov, 1 964, p .108) were generally thought to have been"concave s crapers"(Cauvin, op.cit.) used t o shave or whittle relatively hard materials of small diameter, particularly wood. S emenov ( 1964, p .113) suggested that"concave blades" might have been used to make ,

shafts,

axe-

and

adze-handles,

poles,

stakes

etc.

High-power microwear studies showed the f ollowing uses: Cahen and Gysels ( 1983) reported t hat notches had been used t o cut wood. Bueller ( 1983), without presenting any evidence or t aking post-depositional wear-traces into account, reported that notches f rom E l Wad had been used t o cut

tendons

from bone.

Very l ittle has been written about the functional analysis o f denticulates. Denticulates may of course be s imply notches which had been used consecutively. On t he other hand a denticulate may be a tool on which a ll t he notches were used at the same t ime: Moss ( 1983, p .72) reported that a denticulate worked better t han an end-scraper in her experiment dehairing h ide. Cahen a nd Gysels ( 1983, F ig.4:1-2) i llustrated denticulates which had apparently been used to s crape or plane wood and Bueller ( 1983) reported s imilar uses. The f act that notches and denticulates can be produced a ccidentally i s discussed in Part I , Chapter 4 . Notches ( 13 A ) and Denticulates ( 13 B ) are discussed s eparately a s t hey may have been used in different ways.

-

-

1 50

A

-

Notches ( P1.17,

1 .

F ig.28,

2 9:a-b,

3 1:a-d)

Experiments

Nine experimental notches ( singleand multiple-blow) were retouched with a small ha r nmerstone. The notches were used to shave s easoned and f resh wood, dry and f resh wood, soaked antler and l imestone, and t o cut dried s inew. In addition s everal a ccidental notches were produced by dropping f lakes into a hard f loor, and a notch used by Miller to plane wood was examined. Notches were most e fficient ( more so t han s crapers) when materials such a s wood, bone or s tone had a diameter somewhat smaller than that o f the notch. Nevertheless, the edge of a notch used to shave a fresh birch twig was exhausted a fter 1 000 S m, and those used on bone and stone earlier than that. I found it much more d ifficult to cut s inew with 2 .

a notch

than with

a f lake.

Results

Wear-traces f rom the quarzite-hammer retouch were absent on one notch and present on another where they consisted o f t he typical"stone smears"on the edge, i n direction o f the blow. S imilar t races were observed in some instances

on

a ccidental

notches.

Shaving wood caused rounded and l ittle edge damage, a faint, continuous l ine of polish on the very edge of the notch, and a quantity o f f ine, deep striations ( Pl.17: 0. Shaving bone caused some edge damage and general"bone"polish ( Pl.17:e) ( see Part I , Chapter 1 4). Shaving Chapter

1 4)

antler with

caused weak

f ine

antler

polish

( see

Part

I ,

striations.

Shaving l imestone caused striations in t he direction

general o f use

polish with and an abraded

f ine edge

( Pl. 1 7 : g) Cutting s inew i solated polish and

l eft very superficial

l ittle edge striations.

damage,

some

I f ound the wear-traces on notches difficult t o identify. However, the nature o f these tools would suggest that their use was l imited to wood, bone, antler or stone. 3 .

Notches

from Arjoune

S ee Tables 7 -10 for numbers and percentages o f at Arjoune. All notches were made from the ventral unless otherwise s tated.

1 51

notches a spect,

Trenches

1-1V

Trench I , 1 03.10 A "notched b lade"(Copeland, 1 981, F ig.11:4, here F ig.28:a) with a small s ingle-blow notch may have been used a s a notch. The only wear-traces on t he implement, in the f orm of thin s triations perpendicular t o the notch, l ooked l ike those from stone ( Pl.17:a). They may have been due t o use or e lse due t o a ccidental or deliberate retouch. -

Trench V 2 02.3 A f lake with a multiple-blow notch ( Fig.28: b) which gave r ise t o a projection had apparently not been u sed a s a notch, but t he projection seems t o have been used as a perforator, perhaps on antler and hide ( Pl.17:b) -

? ?5.2 A f lake used on wood, edge ( Pl.17:d). -

with a multiple notch judging by the f ine

( Fig.28:c) had striations on the

been very

Trench VI Baulk 6 00-700.3 A f lake with a s ingle-blow notch on the lateral edge, which had been made f rom t he dorsal a spect, had a concavely truncated distal end, and thus two projections on either s ide o f the distal end ( Fig.31:a) One of these projections, opposite t he notched s ide, was s lightly hooked. The small lateral notch and the t ip next to -

.

i t did not have any wear-traces, while t he truncated end ( especially near the hooked point) had c ontinuous polish, looking l ike wood polish. It therefore s eemed that the t runcated notch was

end had been used, perhaps t o s crape too steeply angled for shaving.

wood,

as

the

7 01.3B A backed blade, truncated at one end and snapped at the other, had been burnt. The small s ingle-blow notch ( Fig.31: b) l ooked unused. The t ool had probably been used a s a cutting blade, and the notch was accidental. -

702.3B

-

A broken blade

with

a s ingle-blow notch

made

from

the dorsal a spect ( Fig.31:c, had a "stone smear"on the ventral a spect of the notch which may mean that the notch had been produced accidentally. 8 01.3 A l ong blade, t runcated a t the distal end with semi-abrupt retouch and with a multiple i rregular notch at the distal end ( Fig.28:e) showed l ittle edge damage at the -

notch, and a polish ( Pl.17:f) I could not t ell whether the been used.

which l ooked l ike bone polish. l ateral retouched edges had

1 52

Trench VII 1 000.1 A blade w ith a multiple-blow notch ( Fig.31:d) had been used a s a notch, judging by the rounding o f t he edge o f t he notch. However t he f lint was s lightly patinated and i t was impossible t o identify t he worked material. -

1 011.2 A f lake with a s ingle notch ( Fig.29: b) had a c oncave t runcated distal end thus forming a point. I t s eemed t hat the notch and t he t ip ( with l ittle edge damage and continuous polish) had been used on wood, a s had perhaps the concave truncation, a lthough the polish ( Pl.17:h) t here may have been f rom retouch. Altogether i t i s l ikely t hat this -

t ool had been notch and the 4 .

used a s a borer, t runcation.

with

the

wood

t ouching

the

Conclusions

Only a few notches had wear-traces which could be identified with some degree o f certainty. Notches appeared t o have been u sed to s crape ( or shave) wood, bone and possibly perforate wood. I t seemed that s ingle-blow notches, and hit from the dorsal a spect, were S ome notches appeared perforate materials.

to

have

especially accidental.

been made

1 53

t o

when

provide

small

a t ip

to

B

Denticulates

-

( P1.18, 1 .

F ig.29:c-e,

3 0,

3 1:e-g,

3 2)

Experiments

used

Four t o

sheep's

denticulates were retouched with a ham r nerstone and s crape f resh and dry wood, dehair hide a nd comb

wool.

I f ound t hat denticulates were not e fficient when used t o s crape wood. Moss ( 1983, p .72) reported that denticulates worked best when she dehaired a hide. However, i n an experiment which she and I carried out together we found that on a f ine roe deer hide ( soaked f or several days in water) denticulates only s craped away t hin l ines o f hair and were in constant danger of tearing the hide. Wool was combed 2 .

e asily.

Results

The retouch

wear-traces were the same

from a ccidental denticulates and a s those on notches and are d iscussed

above. The wear-traces f rom s craping a large plank a ffected mostly the projections and the edges of with a weak continuous l ine of polish on the ( Pl.18:a)

with

‚

broad

striations

running

in

the

of oak wood the notches very edge direction

o f

use. Few wear-traces could be seen on the denticulate used dehair hide for 7 minutes on a stone-surface. Only one

t o

f lake had broken off each polish ( Pl.18:e) looked confined The f ine 3 .

to

the

wear

t ips

traces

of

projection s imilar t o

the

f rom

of t he tool. The s tone polish and was

projections.

combing

wool

consisted

o f

very

striations.

Denticulates S ee

Tables

from Arjoune 7-10

f or

numbers

and

percentages

o f

denticulates. Trench V 1 01.1 A retouched f lake with two notches on one s ide and a pointed end ( Fig.30: d) had c lear stone polish on the projections only. The points may have been used t o incise -

s tone

( Pl.18:h).

1 01.1 A f lake with two notches ( largely cortex) ( Fig.31:e) had l ittle edge damage, continuous polish on the upper -

notch, and may have been used a s a notch rather denticulate. It s eemed t o have been used to s crape

1 54

t han a or plane

wood. t he

The

s ame

l ower

notch

a lso

had

weak

t races,

possibly

due

t o

u se.

1 03.3 A f lake with two notches on one s ide and one notch on the other ( Fig.29:c) had l ittle edge damage, but a c ontinuous l ine o f polish ( like experimental wood polish) a round the two notches on one s ide. The t ip had s imilarly l ittle edge damage and s imilar polish, and l ooked a s i f i t had been u sed t o i ncise wood ( Pl.18: b). The implement could easily be held i n such a way a s to make incising easy. -

2 13.3 A burnt b lade-fragment with three small notches on one lateral edge ( Fig.29:e) had very f ine s triations perpendicular to t he notches and the projections. There was -

hardly any edge damage. The wear-traces indicated t hat this implement may have been used to comb a f ibrous material ( Pl.18:c, c f. P 1.14:e). On the other hand t he notches a s well a s the wear-traces may have been due t o natural agencies,

a s

the

i mplement

was

burnt.

Trench VI 5 00.1 A denticulated f lake ( Fig.30:a) had been used t o s crape wood or h ide, judging from the continuous polish and t he relative lack o f edge damage. The polish ( Pl.18:d) was f ound on the notches and on the projections. The f lint was t oo coarse to permit a more precise reconstruction of this -

t ool's

use.

7 00.1 A denticulated f lake ( Fig.31:f) had been used s crape wood or h ide. I t showed the s ame wear-traces a s t ool discussed immediately above.

t o t he

-

8 00.1 t o have

A denticulated f lake ( Fig.31: g) with been used to scrape and bore wood.

a point

seemed

8 04.3

A denticulated

a point

seemed

-

-

f lake

( Fig.32:a)

with

t o have been used to s crape hide. The point was and I could not t ell whether i t had been used.

crystalline,

9 00.1 A cortex b lade ( Fig.32:b) with two notches on s ide was patinated. The larger notch seemed to have used but i t was impossible to say on what material. -

one been

Trench VII S ounding 4 A naturally backed f lake ( Fig.32:c) with one multiple and s everal small s ingle-blow notches had been used, judging from the rounded polish on the projections. However, i t was not c lear whether the denticulated edge had been used with a t ransversal or longitudinal movement, a s t he s triations ran in both directions. -

1 001.3 polish,

-

A denticulated cortex-flake but a lso rounded continuous

had probably

been used

a s

a

scraper,

1 55

( Fig-32:d) had s tone polish on i ts edge, and perhaps

on

wood

or

h ide. 1 002.3 A "denticulate on a cortex-flake"(Copeland, i n prep., F ig.23:7, here F ig.30:e) had s tone polish a ll over the edge. I t may have been a s tone scraper, but t he possibility that i t had been a core could not be ruled out ( Pl. 1 8 : g) -

4 .

Conclusions

I t s eems t hat tools i n a

denticulates had variety o f ways:

been u sed a s notches

a s multi-purpose t o shave or s crape

wood; a s scrapers perhaps on wood, h ide and s tone; a s"combs"of f ibrous materials; a s perforators, perhaps o f wood; a s gravers, perhaps o f wood and stone. I n

one

instance

a "denticulate"may

1 56

have

been

a core.

Chapter

1 4

-

( P1.19,

Burins F ig.33-34)

A burin ( see Newcomer, 1 972, for a typology of burins) s ometimes a lso referred to a s"graver"(Crowfoot Payne, 1 983), i s a f lake or blade f rom which a burin-spall has been removed by burin-blow t echnique ( see Newcomer, 1 972, p .26). I n this way s ix working edges c an be created ( Moss, 1 983, p .47) As only one blow i s required f or the creation o f a burin, i t can be easily created accidentally ( Moss, ibid. Barton and Bergman, 1 982) In two cases implements with edges l ooking l ike burin-edges were in fact blades with two s napped intersecting edges. I called these .

.

implements"pseudo-burins". 1 .

Previous Research

Newcomer ( 1972, Appendix 3 ) discussed the following u ses ( some more, some l ess l ikely, but a ll possible) o f burins: to groove hard organic materials, a s scrapers, a s boring or drilling tools ( cf. S emenov, 1 964, pp. 66 and 9 8-99) a s stone-engraving t ools, truncations on burins a s s craping edges, burins a s cores and the burin-blow t echnique a s a way of t ipping arrowheads. Newcomer ( 1981) a lso made a s tone bowl, using ( amongst other tools) a burin. In some c ases i t appeared that the burin-spall, the"by-product" o f t he burin-blow t echnique, was in fact the tool: this was the case at Tepe Hissar ( Tosi and P iperno, 1 973) where burin ,

spalls

had been

Low power ( 1978); f rom s tatistically i ndicated by

used

as

studies

drills. of

burins

include

that

her attribute analysis s ignificant relation between edge wear) and various burin

by

S eitzer

she f ound no function ( as types. S eitzer

( ibid.) quoted Bordes who had stated that a fter cutting reindeer antler with a burin, he

15 minutes noted l ittle

microwear, and, what was worse ( for microwear analysts) that he would have resharpened the burin normally beforehand. However, no magnifications were given in this context. High-power

microwear

studies

of

burins

suggested

that

t hey had been used to work reeds or wood ( Moss, 1 983a), or to groove or bore hard organic materials ( Keeley, in: Audouze et a l. 1981, p .139) Moss ( 1983, pp.116-117) reported that burins from P incevent had been used f or bone or antler work, involving the burin edges and facets, and a lso for hide s craping, piercing and cutting, a lthough not necessarily involving the burin edges. In addition, she ‚

.

mentioned wear-traces, f rom butchering, woodworking and other activities, on burins from the same s ite. Other high-power analyses of burins included those by Cahen and Gysels

( 1983)

and Bueller

( 1983)

1 57

2 .

Experiments

Twenty-one burins had been retouched w ith hard hammers and antler hammers, and t he burin-edges ( see F ig.8: b) were used to groove fresh and s easoned wood, f resh and dry bone, soaked f allow deer and r eindeer antler, s oaked horn, pottery f rom Arjoune, l imestone and shell. Two burin-facets ( see F ig.8: b) were used t o cut rough-outs f rom a shell 1 .2. cm thick, and a burin-edge was u sed t o perforate bone. I n addition I s craped bone with a burin-edge, and examined a burin used by Newcomer to s crape wood and a burin u sed by Grace t o s crape reindeer antler. The burins were usually used t o groove w ith a pulling movement of the upper edge. ( The l ower edge could not be used as the upper edge invariably got in the way . ) Pushing was only possible when the worked material was f airly soft. Most materials could be grooved easily, a lthough antler had t o be soaked for several days. Burin-edges u sed on f resh or soaked wood were blunt a fter one hour, while those u sed on dry bone were blunt a fter 1 0 minutes. Very hard materials such a s l imestone and horn ( because of i ts i nner casing) caused the edge t o s lip rather t han to cut deeply. F ine and/or deep grooves in bone, shell or stone had t o be made with piercers ( Part I I, Chapter 1 1) f lakes ( Part I I, Chapter 1 2) or narrow-angled scrapers ( Part I I, Chapter 1 0). ,

were

Burin-facets used t o cut blunt after 1 0 minutes.

shell

worked

very

In contrast to others ( Newcomer, pers.comm.) boring with burin edges a lmost impossible.

3 .

well

I

and

f ound

Results

Grooving wood l eft l ittle edge damage and the t ypical wood polish ( Pl.19: g), in some cases with f aint macroscopic gloss. Grooving fresh and dried bone l eft diffuse bone polish, often with the characteristic striated"bevel"(see P art I , Chapter 1 4, P 1.19:a). Fresh bone c aused l ittle edge damage, dried bone severe edge damage. Grooving soaked reindeer antler left only i solated domed polish areas, a s the f lint used had been coarse-grained. Grooving f allow-deer antler caused a diffuse continuous polish with some edge damage and regularly spaced striations, which I f ound t o be typical of fallow-deer antler considerable t ime, reindeer

( cf. P l.11:c). When used f or a antler polish did l ook l ike

the"snowbank"Keeley ( 1980) described. In this case antler polish looked s imilar t o wood polish. Less developed antler polish however, could look l ike bone polish, a lthough the edge l ooked l ess jagged. Grooving horn l eft weak f lattened polish, while grooving pottery l eft diffuse polish with i nclusions ( from t he pottery) and shallow s triations. Grooving l imestone l eft an abraded t ip and typical l imestone polish

( see

Part

I ,

Chapter

1 4,

1 58

P l.19:c)

‚

while

i ncising

and

c utting

s hell

l eft

t ypical

shell

polish

( ibid.,

P 1.19:e)

S craping bone, wood and soaked antler with t he burin-facet l eft wear-traces l ike t hose on s crapers ( see P art I I, Chapter 1 0) with l ess edge damage, presumably because the e dge a ngles were s teeper. ,

Boring bone

l eft

a general

s triated polish.

On t he whole I found that t he area used could be i dentified, but that the worked materials presented problems: antler polish could l ook s imilar t o bone or wood, depending probably on the duration o f work s hell and stone l eft t races which l ooked s imilar depending on the angle o f t he f lint t o the microscope l ens. The polish f rom grooving horn and f rom perforating bone l ooked t oo weak to be distinguished on ancient t ools. 4 .

Burins S ee

Trench

from Arjoune Tables

7 -10

for

numbers

and

percentages

of

burins.

I

Surface A "single-blow burin"(Copeland, 1 981, F ig.11:11, here F ig.34:a) had s evere edge damage, together with a buoyant polish and striations a t i ts burin-edge. Weak polish was also f ound on the burin-facet, and on both a spects -

bordering the burin-facet. It used t o i ncise s ideways. determined.

s eemed that the The material

tool had been could not be

Trench V 2 12.3 A burin ( Fig.33:h) had apparently been used t o cut or split p lants with the edge on which the burin-facet had been made. This edge had macroscopic gloss, and plant polish ( Pl.19:h) l ooking l ike that from dried reeds ( Pl.23:f) The burin-facet had a weak l ine of polish a long the ventral a spect, indicating that i t had been used for the s ame purpose. -

.

Trench VI 7 00.1 A "co r nposite burin with s craper"(Copeland, in prep., Fig.21:4, here F ig.33: b) made on a patinated blank, had a polish"bevel" with striations on i ts burin-edge. This suggested t hat bone may have been grooved ( Pl.19: b) The s craping edge looked used, but a s the tool was patinated i t was impossible t o identify the worked material. Heavy edge damage suggested bone s craping. -

.

7 01.3A A pseudo-burin on a broken blank ( Fig.33:f) was generally c overed with stone and s and polish, but had at i ts dorsal and ventral burin-edges bright patches o f polish with deep s triations, rather l ike experimental shell polish -

1 59

( Pl.19:f) t o incise

.

This suggested or cut shell.

t hat

this

t ool

may

have

been

used

7 03.3B A p seudo-burin ( Fig.34: b) on a broken blade had a broad burin-edge, with lots of edge damage, buoyant polish and s triations which suggested t hat i t had been used ( sideways) t o incise rather t han groove a hard material. The polished area was too small t o be more precise. S ome weak polish was s een on the retouched l ateral edge opposite t he burin f acet. However, i t remained unclear whether this edge had been used or not. -

7 03.3B A narrow burin-edge ( Fig.33:c) had very l ittle buoyant polish with t hin deep s triations. The edge was abraded rather then chipped which suggested t hat l imestone may have been incised ( Pl.19:d) The edge opposite the burin-facet had some irregular retouch, but was probably -

.

unused. burin-edge with 8 00.1 A burin ( fig.34:c) had a c rushed very l ittle polish. S triations i ndicated that it had been t he polish and used to incise a hard material, judging by edge damage shell or bone. -

Trench VII Mound C ( Sounding) A "dihedral burin"(Copeland, in prep., Fig.24:10, here F ig.34: d) had s ome rounded polish and inclusions which suggested that the incised material may -

have

been

pottery.

1 001.2 A s ingle-blow burin ( Fig.34:e) on a broken, retouched bladelet had apparently not been used, judging by the absence of any definite wear-traces. I t i s possible that the burin-spall came off a ccidentally, or e lse that i t had -

been 5 .

detached

for

easier

hafting.

Conclusion s

Only very f ew burins with reasonably definite wear-traces could be found. Of these most had apparently been u sed to incise ( and in one instance perhaps t o cut) relatively hard materials, perhaps bone, shell were found t o be incised on the have been used to split reeds.

or l imestone. Such materials s ite. One burin seemed to

The burin-break on accidental, unless i t

bladelet a t ang.

Two have

been

implements used

a s

a small served a s

were

broken

burins.

1 60

blades

may

which

have

appeared

been

to

Chapter

15

-

( P1.20,

Arrowheads

and Barbs

F ig.35)

A f ew t ransverse a rrowheads were excavated a t Previous r esearch ( see below) suggested t hat shaped points a nd b ladelets could have been arrowheads 1 .

and

Previous

Arjoune. variously used a s

barbs.

Research

C lark et a l. ( 1974) i nvestigated Late P redynastic and Dynastic Egyptian archery equipment. Arrows were hafted in f letched reed-shafts with hardwood f oreshafts, and in dynastic t imes barbs were added. The stone a rrowheads were t ransversely mounted lunates or naturally backed pieces, but o ther materials l ike bone, ivory or wood, were made into points, and some were probably coated with vegetable poison; even bones from a catfish spine were used a s arrow t ips. Arrowheads were probably hafted with beeswax and resin, ( whereas a rrowheads at Tell Hadidi in Syria had been hafted with gypsum ( Miller, 1 983). Clark et a l. ( op.cit.) reported that some transversely mounted stone a rrowheads and barbs were hafted loosely and obviously designed t o disengage from the haft, in order to cause s evere wounding. Normally such arrowheads were f irmly mounted ( ibid.). The earliest Egyptian bow depicted on Predynastic rock drawings was the self-bow. It was s imilar to those still found in parts o f Africa today ( ibid) -

-

Experimental archery was carried out by Bergman and others with copies o f points from Hengistbury Head ( Barton and Bergman, 1 982) and f rom Upper Paleolithic l evels at Ksar Akil ( Bergman and Newcomer, 1 983) The f racture patterns on t he points, f lute-like, burin-like and transversal bending .

f ractures ( op. c it.) were a lso observed on many of the archaeological points the authors had studied, but sometimes no breakage occurred ( Barton and Bergman, 1 982) The e ffectiveness of small transverse arrowheads, thought of by Woolley a s bird-hunting arrows, was proved on large game ( Miller et a l., 1 982) These researchers suggested the use o f vegetable poison, perhaps on wooden .

arrow 1 974)

,

t ips ( similar s ince only

the s ite o f Shams

to the t ips described by C lark a f ew arrowheads had been excavated

Low-power microwear investigations included those Ahlers ( 1971) who reported that Paleo-Indian projectiles been used not only a s such, but a lso a s s crapers whittling

f rom

their paper referred t o, namely the Halafian s ite ed Din, a s ite r ich in wild f auna ( Uerpmann, 1 982)

t ools,

High-power included two reported that

a nd

a s hort

paper

by

Odell

by had and

( 1978).

microwear analyses dealing with projectiles r ecent s tudies: Anderson-Gerfaud ( 1983) wear-traces and residual hafting agents

1 61

suggested t hat some l unates a nd t riangles f rom t he Mesolithic and Natufian l evels a t Mureybet and a t Abu Hureyra had been shot a s a rrowheads and barbs i nto animal t issue. Moss ( 1983) reported t hat she could only f ind MLIT on 3 0% o f her experimentally shot projectiles ( ibid., p .95). S he s tated ( wisely in my opinion) t hat the precise t arget could usually not be points f rom Pont d ' Ambon Abu Hureyra ( 1983a) 2 .

determined ( ibid.), e . g. on Azilian ( p.147) and on t anged points f rom

Experiments

S everal projectiles with a rrowheads ( 3 o f t hem t ransverse) and barbs had been f ired by Bergman and Barton at a red deer carcass which had been defrosted. The s elf-bow was a copy, but in yew, of t he Mesolithic Holmegaard bow and drew 4 0 lbs at 2 6 inches ( see a lso Barton and Bergman, 1 982). The unfletched a rrows ( 79.5 cm l ong) had been made o f Port Orford Cedar wood, and t he arrowheads ( made of Brandon f lint with a hammer stone and an antler pressure-flaker) had been f astened with wax and s inew, covered by resin, in a notch cut at t he t ip o f t he arrow. S everal pairs o f barbs were inserted at the s ides of t he arrow. The shooting distance was 4 meters. The t ransverse arrowheads were very e fficient: one arrowhead went r ight t hrough the chest c avity of the deer, completely intact. Only t he r esin had c racked. Another penetrated the carcass, got l odged in t he spine and could not be retrieved. I t i s hoped t hat t he arrowhead will be retrieved during future o steological analysis. The third t ransverse arrowhead shattered on impact on the r ight l eg o f the carcass.

once

In addition, several i nto meat and bone,

one shot s even examined. 3 .

t imes

into

pointed a rrowheads and barbs shot and others shot i nto dry sand, and wet

sand

( all

by Newcomer)

will

be

Results

The most characteristic wear-traces on t ransverse a rrowheads which had been f ired were the edge damage; t his consisted of heavy ( macroscopic) chipping of the cutting edge. Bergman ( pers. comm.) who had additionally shot 7 t ransverse f lint and obsidian arrowheads with a s elf-bow drawing 4 0 lbs in weight, reported s imilar breakage, and a lso that spalls were removed a long t he retouched edge. Microscopic t races were very random: MLIT could sometimes be s een in t he direction of the shot, mostly on t he arrowhead r idges ( Pl.20:e, g) However, s triations could a lso be s een i n different d irections, probably originating when t he a rrowhead was deflected i nside t he carcass or pulled out f rom the carcass. Edge damage on t ransverse arrowheads was rounded ( Pl.20:a) A general polish w ith f ine s triations could be seen on t ips o f pointed arrowheads even when t hese had broken, but on ancient tools such polish .

.

1 62

might

be

confused with

post-depositional

polish.

Arrowheads s hot i nto wet s and were covered with s lightly buoyant l ooking polish f lecks. The strength o f t his polish was probably related t o t he f act t hat the projectile had been shot s even t imes. S uch polish, covering a lmost t he entire sheen. damage f lecks

a ctive part of t he tool, could be mistaken for s oil An Azilian point, shot i nto dry sand, had jagged edge on part o f the l ateral edge, and very bright polish characteristic o f s and or metal.

Barbs ( Part I I, Chapter 1 2) were e ither shattered or survived i ntact. No particular t races could be i dentified on barbs, except some general polish covering most of t he i mplement ( Pl.20:c) Hafting 4 .

Arrowheads For

Trench

t races

were

not

s tudied.

from Arjoune

numbers

and

percentages

s ee

Tables

7 -10.

I

Surface Copeland

-

A "fragment inside t he

of a pressure-flaked arrowhead"drawn by outline o f an Amuq Point from Janoudiye

( 1981, Fig.10:1, here F ig.35: g) had some s triations in t he direction o f shooting on i ts dorsal r idge. However, a s i t was a surface f ind there was lots of random s tone polish a ll over both a spects ( Pl.20: h) Also, according to Copeland's reconstruction, I was examining t he hafted end. .

Trench VI 700.2

A burnt"transverse

-

arrowhead"(Copeland,

i n

prep.,

F ig.18:1, here F ig.35:d) with both a spects o f the distal edge s lightly chipped ( Pl.20:d) had polish and deep striations on i ts dorsal r idge. Unfortunately the entire t ool was somewhat polished, probably due to t he burning, but i t may have been s hot. 701.3B

-

A burnt"transverse

arrowhead"(ibid.,

Fig.18:2,

here

Fig.35: b) had s imilar wear-traces ( Pl.20: b,f) to t he arrowhead d iscussed immediately above. There was rounded edge damage on t he distal end and lots o f general polish a nd striations, a s well a s some rounding and striations parallel to the axis o f the i mplement at t he distal end. I t had probably been shot. 8 03.3 an"atypical t ransverse F ig.18:4, here F ig.35:f) showed on the above arrowheads. -

those

a rrowhead"(copeland, ibid., wear-traces s imilar t o t hose

No arrowheads had been recovered f rom from Trench V were not available f or

1 63

Trench s tudy.

VII

and

5 .

Conclusions

L ittle

can

t hat they game. Two been used.

be

said

about

the

f ew

arrowheads

d iscussed

except

could have been u sed e fficiently to hunt l arge t ransverse a rrowheads f rom Trench VI s eem t o have

All the arrowheads appeared t o have c losely t ogether and two were burnt.

1 64

been

f ound

relatively

Chapter

1 6

-

Axes,

( P1.21,

Adzes

and Choppers

F ig.36-38)

Axes are defined by t heir shape. The difference between axes and adzes i s in the method of their use: axes are hafted with their working edges parallel to the haft, while adzes a re hafted t ransversely to the haft. Choppers are usually 1 .

pebbles

Previous

or

cobbles

with

bifacially

retouched

edges.

Research

S emenov ( 1964) s tated that he could differentiate between axes and adzes f rom the distribution o f s triations, which were stronger on the f ront f ace ( away f rom t he handle) i n the case o f an adze, while they were f ound in s imilar quantity on both f aces in the case of an axe ( p.125) He t hought t hat axes were a lways connected with wood working, a lthough use of axes on other materials l ike ivory or bone was not ruled out. Adzes were according to him a lso wood-working tools, and he s tated that adzes used a s hoes, i .e. to dig the ground, could not be reused a s wood-working t ools, a lthough the reverse was possible ( ibid.). Gould et a l. ( 1971) c ited instances of stone hand-axes used a s digging t ools by aborigines, but reported t hat such tools were a lmost a lways used to chop wood ( p.156-157) Flakes used as adzes had edge angles between 4 0 and 8 9 degrees. A hafted adze was used in a s itting position and drawn towards the worker, to make a spear-thrower for instance. The adze was resharpened some 2 0 t imes and the t ool sometimes reversed in the haft. I t t ook the aborigines about e ight hours and 3 0 minutes t o make a spear-thrower with a stone-adze, compared t o four to f ive hours with metal tools. Adzes were a lso used for other purposes ( ibid.) e . g. t o engrave sacred boards. .

.

‚

Microwear s tudies on axes and adzes included that by Keeley ( 1983a) who found that wear-traces on s tone tools f rom PPNA Jericho indicated that they had been used a s wood-adzes ( p.759 and p1.33). Coqueugniot ( 1983) examined wear-traces on f lake-adzes from Mureybet under l ow-power magnifications. He found t hat the adzes had been used t o chop and f rom the

a lso same

Ground

to plane wood. According context had been used in

s tone

hatchets

f rom

PPNB

to him, some s crapers the same way.

Bouqras

had,

a ccording

t o Roodenberg ( 1983), been used i n various ways: a s f elling axes, wedges, chisels, and adzes. Hafting arrangements a s axes and adzes were s hown ( ibid., Figs.3 and 4 ) a fter reconstructions by Mellaart and Cauvin. 2 .

Experiments

Four axes and choppers were used experimentally t o chop wood and bone. In addition f ive hafted t ranchet axes ( one used by Bergman t o chop wood, t he other f our used by Harding

1 65

t o chop hazel wood w ith up t o 1 500 b lows) were examined. Unfortunately, no hoeing experiments were carried out. 3 .

Results

The wear-traces consisted of f airly i solated polish areas and striations running parallel t o t he direction o f u se, on both a spects of t he working edges o f t he choppers and axes. Wood l eft l ittle edge damage and domed wood polish with striations ( Pl.21:c,e), while bone l eft a l ot of edge damage and i solated f lat polish ( Pl.21:a) 4 .

Axes,

Adzes

For

and Choppers

numbers

Trenches

and

from Arjoune

percentages

see

Tables

7 -10.

1-1V

Surface An"axe ( with polish on upper part. . made on a chert pebble"(Copeland, 1 981, F ig.10:3, here F ig.36: b) had regular retouch on one a spect of its sharp edge. Microscopic examination revealed a domed polish, perhaps f rom wood, and invasive stone polish ( Pl.21:b), probably f rom polishing and retouch, a lthough of course post-depositional damage was possible. Wear-traces were f ound over both a spects. This t ool's shape suggested that i t had been hafted. .

-

Trench V 2 22.2 A l arge chopper ( Fig.37: b) made on a f lake s eemed t o have deliberate ( regular) retouch on t he dorsal a spect, and edge damage on the ventral a spect. Domed polish and striations on both a spects of t he working edge suggested that it had been used a s a wood chopper ( Pl.21:d) The shape suggested that it had not been hafted. -

.

Trench VI S urface A "chipped axe or adze with damaged bit"(Copeland, in prep., Fig.21:5, here F ig.38:a) showed domed polish and striations on both ends, suggesting that perhaps both ends had been used. The polish was in each case quite invasive. This fact, together with the nature of the polish ( P1.21:f) suggested that wood had been worked. Striations ran both perpendicular and parallel t o the l ong axis o f t he t ool, t hus indicating two directions o f impact. Wear-traces were f ound on both aspects, suggesting i t was an axe rather than an adze. -

Trench VII S ounding I A "small chopper or axe-roughout"(Copeland, in prep., F ig.24:9, here Fig.37:c) had s ome polish ( P1.21:g) on t he projections of t he cutting edge, which may have been wood or stone polish, but was probably wood considering t he l ack o f crushing of the edge. Judging by t he shape i t may -

1 66

have

been

hafted.

S ounding 4 A chopper ( Fig.38: b) on a f lake s imilar t o the c hopper f rom Trench V had rounded edge damage on both a spects and domed polish ( Pl.21: h) with s triations perpendicular t o the cutting edge. This suggested that t he t ool had been used a s a wood chopper. -

5 .

Conclusion

Unfortunately only a few axes and choppers could be s tudied a s most o f such tools had been made o f l imestone, rather t han of f lint or chert. The t ools examined appeared t o have been used to chop wood. But the f act that no hoeing experiments were carried out and that most o f the a rchaeological t ools were surface f inds made t he conclusions uncertain.

1 67

Chapter 1 7 Blades in the Levant. -

with Gloss

from Arjoune

and Other

Sites

Introduction A s ignificant number of b lades or bladelets with g loss or l ustre a long one or both edges were f ound on s ites i n P alestine and Syria from the Epi-Paleolithic ‚ dating f rom approximately 1 8,000 t o 1 5,000 BC onwards ( Bar-Yosef, 1 970, p .9) These pieces became more numerous and l arger i n s ize .

i n the course of t he Neolithic, that i s, approximately f rom about the beginning of the 8 th millennium BC ( see t he Carbon 1 4 dates i n Mellaart, 1 975, pp.283-288) These blades are commonly referred to as"sickle blades" and have s ometimes .

been regarded a s evidence f or c ereal-gathering, i f cultivation among otherwise pre-agricultural ( Neuville, 1 934, p .18).

o f

S imilarly Neolithic

the and

not plant peoples

mention of"sickle blades"in the context later s ites has become a lmost synonymous

with"agriculture", and where"sickle b lades"have been absent, t his has been seen a s an enigma: were there no c ereals t o harvest? I f there were cereals, how were they harvested? These f allacies

l ast often

two questions encountered

are when

indicative o f two dealing with"sickle

b lades"and harvesting. The f irst f allacy i s that gloss ( often called"sickle gloss", " silica gloss"or"corn gloss", s ee Part I , Chapter 1 0) equals gloss f rom grasses ( Wilke et a l. p .205), especially from wild or domesticated c ereals. The f act i s t hat a considerable, varying amount of s ilica the hardness of which seems to cause the gloss ( see P art I Chapter 1 0) i s present in a ll green plants. Not only grasses, such a s cereals, but a lso other p lants, such a s reeds and rushes, have s iliceous stems; these were common in the Levant and had their uses i n daily l ife. They a ll produce gloss when rubbed against a f lint surface for a l ength of t ime. ‚

In f act, as demonstrated in Part I , Chapter 1 0, gloss on f lint i s caused by contact with plants and with other hard materials. Conversely, there i s the question whether s imilarly shaped blades without g loss are definitely not"sickle blades". Some archaeologists, e . g. Otte ( 1976) in his study o f Neolithic tools from Apamea in Syria, counted blades without gloss a s"sickle blades". However, in Part I , Chapter 1 6 i t was shown that only on coarse-grained f lint did gloss appear later than did microscopic polish. Such blades without gloss could o f course have been unused. The second f allacy rests on t he a ssumption t hat c ereals wild or domesticated are a lways harvested by cutting. In f act there are other methods still u sed today in some parts o f t he world, i ncluding harvesting by beating or by -

-

1 68

uprooting.

Examples

exist

f or

a ll

three

methods:

Harvest o f cereals by cutting with non-metallic s ickle b lades: i n Africa domesticated wheat and barley, amongst others, are still harvested with f lint and bone s ickle b lades ( Whitcombe, pers.comm.). Even wild c ereals with their brittle rachis have been c ut quite successfully ( Harlan, 1 967) W ild c ereals were cut not long ago in Mongolia ( Maurizio, 1 927, p .136). -

.

Harvest t hough i t ( ibid.).

by i s

-

beating i s often referred to ( Maurizio, ibid.) an i nefficient method because of t he grain l ost

Harvest by uprooting was practiced in Palestine ( Dalman, 1 928, Vol.1, Part 2 , p .551) and more recently in Syria ( Van Z eist and Bakker -Heeres, 1 979, p .166). It i s seen by Hillman -

( pers. c ereals

comm.) a s the easiest method, are harvested; reeds and rushes

have t o be cut because o f their a re a lmost impossible to pull out.

that

Nevertheless cereals were

metal

There are plenty of present s tone s ickles on c ereals.

-

horizontal

rhizomes

which

s everal arguments favour the possibility cut, even with stone s ickle blades:

The present day use of non-metallic precursors. -

especially when wild on t he other hand

s ickles

day

and

instances

scythes

of

the

points

use

to

of

There are disadvantages t o uprooting cereals: soil f rom the roots would get mixed into the grain or straw. Perhaps worse, erosion of the soil would be severe i f the roots were pulled from the a lready upturned soil, especially in areas such a s Arjoune where winds from the Mediterranean prevail f orcefully a ll day and a ll year around. -

It seems that a ll three methods of harvesting have their advantages and disadvantages, and any could have been practiced at a s ite in the Levant. The absence of s ickle b lades i s therefore not proof o f a non-agricultural economy, and, a s outlined above, the presence of s ickle blades i s not necessarily proof o f an agricultural economy. I therefore wanted to study the blades with gloss and s ome blades without gloss c lassified a s"sickle blades" ( Copeland, 1 981, p .11, and i n prep.) in detail, in order to f ind out whether they had been used a s parts o f s ickles to harvest cultivated cereals. The results of my experimental work ( see Part I , Chapters 1 4 and 1 5) had shown that only contact with plants or wood l ed t o an evenly distributed g loss on f lint cutting-blades, and that most different p lant species l eft different wear-traces on the blades ( Table 2 , P 1.22-25) Nevertheless, t he interpretation of different wear-traces from d ifferent plant species and/or varieties .

1 69

proved quite problematic ( see P art I I, Chapter 1 7.4). I t herefore decided t o compare t he wear-traces on t he Arjoune blades with those on b lades with g loss f rom other periods and environments, and t o compare t he results o f t he wear-trace analyses w ith t he botanical evidence a vailable f or t he relevant s ites o r a t l east t he relevant areas. For this comparison I chose 2 5 Epi-Paleolithic l ustered bladelets and blades f rom Kebara published by Turville-Petre ( 1932) and f rom El Wad published by Garrod and Bate ( 1937) Both t hese s ites were discovered in the coastal l owland o f P alestine a region which a ccording to Vita-Finzi a nd Higgs ( 1970, p .16) . was largely unsuited t o the growing o f c ereals. I a lso chose 6 3 lustered b lades f rom the P PNA, PPNB, P ottery Neolithic A ( hereinafter referred t o a s PNA) and t he EB levels in area Payne ( 1983) 2 .

Previous

F at

Jericho

which were

published

by Crowfoot

Research

Experimental

investigation of blades with gloss.

" Sickle blades"with gloss have f ascinated archaeologists from the beginning o f research i nto functional analysis. Spurrell ( 1892) tried t o r eproduce gloss which he had noted on"early s ickles"experimentally and f ound that i t was only produced by cutting"ripe s traw" ( ibid., pp.57-58) Vayson ( 1919) in his s tudy o f s ickles f rom Europe and Africa demonstrated that gloss was a lso produced by other materials, such a s wood. Curwen ( 1930) demonstrated experimentally t hat wood and corn caused somewhat different patterns of g loss. He a lso r aised t he question whether one type o f f lint polishes a s e asily a s .

another, and whether gloss a lways occurs when plants are cut ( ibid., pp.184-186). I n a paper published in response t o Neuville's a ssertion that gloss was caused by hard usage, Curwen stated that i t was safe t o a ssume t hat diffuse g loss was caused by"corn or other s iliceous grasses" ( Curwen, 1 935, p .65) and that"one may a ssume that i t i s generally agreed t hat the discovery of t he r emains o f a s ickle in an ancient deposit i s evidence that i ts owner knew how t o grow corn"(ibid., p .62). This i nterpretation o f gloss s eemed t o have induced archaeologists to regard lustered"sickle blades"as evidence f or cereal-gathering i f not cereal-cultivation. In recent years however, i t has been demonstrated by several researchers that gloss i s not only caused by corn or grass ( as well a s by other s iliceous materials ( Curwen, 1 937, p .93)) and by rubbing hard material ( see P art I , Chapter 1 0) but by a ll green plants which have so f ar been harvested experimentally, l ike reeds ( Vaughan, 1 981, pp.149-154) bamboo ( Keeley, 1 980, p .61) s edges, rushes, reedmaces and other p lants ( Anderson-Gerfaud, 1 983, p .89) bulrushes ( Unger- Hamilton, 1 983) c ane, horsetails, s tipa, weeds ( see Part I , Chapter 1 4) The exceptions were poppies which may not have been c ut l ong enough f or g loss t o ‚

,

,

.

1 70

d evelop. S tudies have been carried out on t he microscopic polishes which were produced by different p lant species ( Anderson- Gerfaud, 1 983; Unger-Hamilton, 1 983; P erles and Vaughan, 1 983) S tudies have a lso polishes on lustered b lades", e . g. on burins Hureyra ( Moss, 1 983a) Experimental

been carried out on microscopic f lint implements other t han"sickle and tanged points f rom Tell Abu

efficiency studies

In 1 943 S teensberg published his r esults o f various experiments harvesting barley and oats with European S tone, B ronze and Iron Age s ickles: most e fficient was a Viking s cythe, l east e fficient were a crescentic f lint s ickle with a s errated straight edge and a copy of t he S tenild s ickle, a f lint f lake mounted perpendicular to t he handle. The latter s ickle was e fficient in cutting weeds and thistles rather t han corn ( Steensberg, 1 943, pp.23-25). S teensberg noted that pronounced serration of f lint cutting edges was unsuitable for cutting corn a s the s traws s lipped between t he coarse t eeth ( ibid., pp.25-26). Korobkova ( 1981) studied and compared the evolution of harvesting tools f rom the earliest agricultural complexes in Central Asia, Kazakhstan, the Caucasus, the Ukraine and Moldavia. Copies of the tools were made and used t o harvest various c ereals, beans, reeds and grasses. Korobkova pointed t o the differences ( mainly in the number o f striations) between wear-traces from cultivated cereals, wild cereals, grasses and reeds ( ibid., pp.331-334). She a lso examined t he relative e fficiency of the different implements ( ibid., p .343, F ig.8) a nd the approximate plot s ize cut with e ach type of s ickle. Helmer ( 1983) reconstituted various types of prehistoric s ickles f rom Europe, Africa and t he Near East on the basis o f observations made by J . and M .-C. Cauvin; he c lassified these implements i nto straight and curved s ickles, with parallel and oblique hafting a rrangement and with s ingle and multiple retouched and unretouched e lements. Helmer used these s ickles experimentally on c ereals, though by his own admission for a short t ime only. He came to various conclusions: for example, on s ickles with parallel hafted b lades, i t was the retouch and not t he number of blades which made t he s ickle e fficient f or cutting r ipe c ereals, while on obliquely hafted s ickles i t was t he number o f blades which made the s ickle e fficient. He a lso stated t hat steeply angled s ickles were better f or reaping while s traight s ickles were better for cutting. He f ound the overall e fficiency of variously shaped s ickle mounts t o be s imilar.

1 71

Typology Amongst the innumerable typological s tudies o f"sickle b lades" I would l ike t o mention t he s tudy of the l ustered b lades f rom t he Byblos Neolithique Ancien, Neolithique Moyen and Neolithique Recent by J . Cauvin ( 1968, pp.70-73, pp.100-105, pp.128-133) Cauvin i n h is"Analyse Fonctionelle"(ibid.) attempted t o c alculate the number and l ength o f s ickles f rom the number o f one-ended t runcated e lements ( ibid. p .72) He t ried t o reconstruct t he s ickle mounts f rom t he shapes of t he s ickle e lements and f rom t he angle o f their t runcations. He mentioned t he f act t hat i n h is experiments coarsely denticulated blades had proved inefficient in cutting cereals, but rather effective f or .

,

.

cutting reeds. The early habitations a t Byblos were thought t o have been covered with plant material ( ibid., p .73) and i n f act coarsely denticulated blades were f ound i n great numbers i n the Neolithique Ancien, but not i n t he l ater l evels. Recently M .-C. Cauvin ( 1983) based her s tudy o f s ickle blades from the Near East on typology a nd a f ew microwear analyses. She came to the conclusions t hat the earliest s ickle had been used t o cut plants other than cereals. Hafting S tudies of hafts, hafting materials and hafting arrangements of blades include M .-C. Cauvin's study o f s ickle b lades from Tell Aswad ( Cauvin, 1 973) F rom the location of the bitumen used a s a hafting agent, a s well a s gloss and retouch, she deduced the hafting arrangement of the blades. .

Korobkova mentioned that harvesting implements, t he hafts and hafting arrangements of which she had studied, had in some instances been hafted with bitumen ( Korobkova, 1 981, p .328) or resinous vegetable matter ( ibid., p .330). Camps-Fabrer and Courtin i n their paper"Essaie d ' approche t echnologique des f aucilles prehistoriques dans l e bassin Mediterraneen" ( 1982) gave a comprehensive survey o f excavated early s ickles and s ickle f ragments and i n addition made suggestions about hafting methods. 3 .

Plants

In this section I l ist the plants which were common i n the Syro-Palestinian region and which i n my opinion were most l ikely to have been useful t o prehistoric peoples t here. The l ist i s based on botanical, a rchaeological and ethnographic evidence. For reasons o f t ime and place I could not harvest a ll o f these plants myself and have t herefore concentrated only on those plants which a ccording t o available ethnographic evidence and common s ense a re harvested by cutting. This i s problematic: f lax, f or i nstance, i s universally uprooted but i n the Gezer C alendar,

1 72

d ated t o t he 1 0th c entury BC, t he Hebrew word ' zd which i s t ranslated a s"to cut" ( Donner and Roellig, 1 964, p .181) i s u sed i n c onnection w ith f lax. ( It i s possible t hat not f lax, but i ts variety, l inseed i s referred t o ( Hillman, p ers.comm.) . ) An even greater problem is presented by the f act t hat there may have been many plant species which may have died out or may not be deemed important anymore. S ome p lants l ike papyrus are nowadays nearly extinct i n t he Near E ast, while others, for instance the opium poppy, could only be harvested by r isking imprisonment or worse. I could only harvest p lant varieties which a re c losely related to these p lants. The plant species and varieties I have l isted and h arvested are t herefore quite l ikely to be i ncomplete. This f act has been t aken i nto a ccount in the interpretation o f t he results ( see below) Wild Cereals Wild barley ( Hordeum spontaneum C . Koch) had a wide d istribution ( Helbaek, 1 960, p .112) . Nowadays i t i s still f ound in abundance in t he Near East, both in primary and in d isturbed habitats. I t i s relatively tolerant of drought and heat, but not of cold, a ltitudes o f over 1 500 Zohary, 1 966, p .1076) Wild S chiem.) a n area L evant. grow at i n both

and i s not commonly found meters above sea-level ( Harlan

e inkorn wheat ( Triticum i s found nowadays mainly

at and

boeoticum Boiss. emend. in the Zagros-Taurus a rc,

f rom which i t i s thought to have spread t o the I t i s f ar more resistant t o cold than barley and can heights o f 2 000 meters above sea-level. I t i s found primary and disturbed habitats.

Wild emmer wheat ( Triticum dicoccoides ) i s a lso found i n various regions of t he Near East. Two races exist, one in I ran, Iraq and the USSR, t he other in the Jordan Valley. Wild emmer wheat i s very exacting and only grows in a primary, i .e. undisturbed habitat, on s lopes such a s t he l imestone and basaltic and Zohary, 1 966)

s lopes

of

the

Eastern Galilee

( Harlan

All three species have been exploited a s f ood plants, a s i s i ndicated by c arbonised grain samples f rom Epi-Paleolithic village s ites. Wild e inkorn was found at Tell Abu Hureyra ( Hillman, 1 975) and at Mureybet ( Van Z eist, 1 970) where present day s tands of wild e inkorn are s eparated f rom the s ite by a distance o f 1 00 to 1 50 km. This distance was one o f the f acts which suggested to Moore ( 1982) t hat c ereals, t hough morphologically still wild, had been cultivated at Epi-Paleolithic ( or Mesolithic) Tell Abu Hureyra. However, Van Z eist and Woidring ( 1980, p .124) pointed out that t he distribution of wild c ereals might well have been different i n t he past. The Natufian s ites i n Palestine l ie within the presentday distribution zone o f wild barley, and most w ithin the present distribution z one o f wild emmer wheat. The equipment found on these s ites,

1 73

such a s s ickle blades a nd hafts, grinding s tones suggested t o Harlan ( unpublished manuscript) oriented t owards harvest o f w ild c ereals.

a nd mortars a n e conomy

Harvesting of wild c ereals i s probably most e fficiently c arried out by beating because t hese wild varieties have a brittle rachis which enables t he p lants t o disperse t heir s eeds. However, researchers have convincingly argued t hat harvest by beating would not l ead t o the tough rachis o f t he domesticated variety o f t he species, and in f act s elect against i t ( Wilke et a l. 1 972) I t therefore f ollows t hat at some stage before t he Neolithic cereals were cut with blades or e lse uprooted. Nowadays w ild grain o ften harvested by beating, both by non-agricultural and by agricultural peoples who use s ickles on t heir domesticated crops. However, there are exceptions, such a s the Mongolian t ribes who cut cereals ( Maurizio, 1 927, p .136) I t i s a lso l ikely that wild grain i s cut when the s tems are t o be u sed f or such purposes a s making mats, bedding or perhaps a s animal ‚

.

.

fodder, and under humid c onditions which s eed dispersal ( Hillman, pers.comm.).

would

be

a dverse

to

D omesticated C ereals Domesticated barley ( Hordeum vulgare L .) i ncludes s everal varieties. The earliest variety was t he two-row hulled barley with a t ough rachis which was very s imilar t o the wild variety ( Harlan and Zohary, 1 966, p .1076 ). Further genetic change led t o s ix-row hulled barley, f our-row lax-eared barley and t o t he naked barley, the f irst two f orms carrying more seeds, t he l atter easier t o process. Domesticated e inkorn wheat ( Triticum monococcum L .) a lso resembles its wild progenitor very c losely. The t ough rachis i s the most conspicuous difference between t he wild and the domesticated varieties ( Harlan and Zohary, 1 966, p .1076) Domesticated emmer wheat ( Triticum dicoccum ) t oo i s very s imilar to i ts wild form. It i s a glume wheat and arduous processing i s required t o f ree the edible part o f the grain f rom the g lume. Further genetic change i nto f ree-threshing wheat ( T .aestivum T .durum ) ensured t hat t he glume would f all f ree upon threshing ( Harlan, unpublished manuscript) Domesticated spelt ( Triticum spelta L .) was not f ound i n the Neolithic i n t he Near East. The earliest f inds o f spelt are f rom 2nd millennium BC s ites in Europe. Helbaek ( 1960, p .105) referred t o t he possibility that spelt m ight have existed earlier s ince carbonised spelt s eeds l ook very s imilar to those o f emmer wheat. Three of the f our species ( spelt being the exception) have been used a s f ood plants i n the Near East f rom t he beginning of the 8 th millennium onwards. E inkorn i s l ikely

1 74

to have been domesticated in the Northern Levant and Turkey and emmer further south in the Upper Jordan Valley ( Harlan and Zohary, 1966) while barley could have been domesticated anywhere in the Near East. Helbaek argued that barley, which grows as a weed and which i s far more ubiquitous than wheat, ,

may have been domesticated accidentally together with wheat ( Helbaek, 1960, p .11 -2) Harlan and Zohary ( 1966) pointed out that the domestication of plants is not l ikely to have .

occurred

where

The

wild

cereals

archaeological

following:

at

Tell

grew,

record

Aswad

( Van

but in

Zeist

in

adjacent

areas.

Levant

shows

the and

Bakker -Heeres,

domesticated emmer wheat and einkorn wheat were the earliest phase IA dated to around 7800 BC. Hureyra

( Hillman,

1975)

domesticated

emmer

the 1979)

found from At Tell Abu and

einkorn

wheat, six-row hulled barley and naked barley were found in the Neolithic levels. At Jericho domesticated emmer and e inkorn wheat and barley were found in the PPNA levels ( Hopf, 1983, p . einkorn wheat levels

( Van

609) and

Zeist

.

At Tell Ramad domesticated emmer and barley were also found in the Neolithic

and

Bottema,

1966)

from

Van Zeist and Bakker -Heeres ( 1979, p .168) their investigation of plant remains from

that the

agriculture was well established by 8th millennium BC in the Aswad area

to

have

been

particularly

suited

to

concluded Tell Aswad

the beginning of which did not seem

agriculture.

They

also

concluded ( ibid.) that possible exploitation of surface water for agricultural purposes could suggest that plant culivation may have begun as early as the 9th millennium. According to Zohary ( 1969, p .60) the early farmers, for instance at Beidha ( Helbaek, 1966) practiced both collection and cultivation. The fact that two varieties of barley, one with a brittle and the other with a tough rachis, had been found at Beidha did not necessarily mean in Zohary' s opinion that domestication began at precisely that point in t ime. Harlan

( unpublished

domesticated by

early

cereals

manuscript) might

have

been

stated

that

subsequently

some

abandoned

farmers.

Several

ethnographic

studies

have

been

carried

out

in

the Near East, by Dalman in Palestine at the beginning of this century, by Turkovski in the Judean Hills in 1943-1947, by Sweet at Tell Toqaan in Northern Syria in the 1950s and by Wulff in Iran in the 1960s. A variety of harvesting practices were reported and most of these were probably not very different to those in prehistoric t imes. In general cereals were harvested when ripe, but in some instances ( when growing too high or when needed as animal fodder Palestinian barley ( and rarely wheat) was cut when still green ( Dalman, Vol.11, pp.349-350). At Tell Toqaan green wheat was used to prepare a special dish ( Sweet, 1 960, p .78) The usual annual harvest took place in May-June while the summer crops in Iran were harvested in September ( Wulff, 1 966, p .271). In Iran, barley was uprooted, while .

1 75

wheat was mostly cut with s ickles ( Wulff, 1 966, p .272) The s ame was t rue at Tell Toqaan i n S yria ( Sweet, op. c it., p .70), t hough barley was a lso s ometimes cut t here, when t he"grain was thick"(ibid.). In the Judean H ills most c rops were uprooted, a lthough in the mountainous a reas s ickles were used t o cut grain ( Turkovski, 1 969) Dalman r eported .

.

t hat while p .37)

short barley was uprooted ( 1933, Vol.111, pp.34-35) h igh barley and wheat were a lmost a lways cut ( ibid.,

I n both Syria and I ran, c ereals were cut a s c lose t o t he ground a s possible ( Sweet, op. c it., p .71, Wulff, op. c it.). I n P alestine however, l ong cuims were were cut h igher up ( Dalman, op. c it., p .37), l eaving 2 0 to 3 0 cm h igh stubble for the animals to graze. Dalman a lso pointed to the biblical evidence f or the cutting o f the cereal heads f rom t he culms ( op. c it, p .42) The s ickles varied according t o both region and plant. Wulff reported two kinds of s ickles, a l arge hooked grain-cutting s ickle and a smaller, a lmost s traight grass-cutting s ickle, t ooth-edged in some r egions, straight-edged and hardened with horn i neal i n other regions ( op. c it., p .272). Turkovski ( op. c it.) reported two types of cutting implements, a serrated curved knife, mainly used to cut branches such a s vines, and a large curved s ickle, mainly used t o reap corn, but a lso used on o ther plants such a s herbs and grasses. Dalman ( 1933, Vol. I II, pl.1B) showed and described several harvesting i mplements, including a b lunt curved s ickle used to uproot c ereals and weeds ( ibid., p .19), a straighter sharp s ickle ( ibid., pp.20-21) used to cut c ereals, a lthough smaller versions were used t o cut grasses in some areas. These s ickles were sometimes tooth-edged, sometimes not, a s in I ran this depended on the region; in some cases then the t eeth are s tylistic and not functional. In addition, Dalman reported two t ypes of s ickles, both curved, toothed and untoothed, which were commonly used to prune f ruit t rees and vines ( ibid., p .23). The practices f ollowing the harvest were f airly s imilar in a ll countries. The crops were t ied i nto sheaves, dried and taken t o the threshing f loor, where they were threshed by beating, animal hooves or threshing s ledges. Threshing s ledges are o f special interest t o a study o f f lint t ools a s the undersides of the wooden boards were s tudded with f lints ( Crawford, 1 935) However, t he earliest references f or such implements s tem from biblical t imes ( Wulff, 1 966, p .275) and microscopic s tudy of threshing f lints f rom Cyprus revealed distinctive wear-traces so f ar not seen on the prehistoric implements I have s tudied. The grain was winnowed with forks. The chaff was used a s t emper i n bricks and pise ( Wulff, 1 966, pp.108-109) and i n o il-seed presses ( ibid., p .297); t he straw was used to make baskets, mats and a s animal f odder ( e.g. Turkovski, 1 969) The grain was s ieved t o s eparate i t f rom the husks ( e.g. Sweet, 1 960, p .72) The grain was t hen t aken to the mill o r quern. The earliest .

.

.

1 76

r otary millstones so f ar discovered come f rom t he s econd m illennium BC i n Palestine ( Wulff, op. c it., p .277) However, querns were f ound on much earlier s ites, e .g. a t J ericho f rom the Proto-Neolithic onwards ( Dorrell, 1 983, p .488) Different cereals were used i n the Judean Hills ( Turkovski, make bread while barley was only i n

case

Other

o f

food

f or different purposes, and op. c it.) wheat was u sed t o used f or human consumption

s hortage.

Plants

Wild grasses proliferate in the Syro-Palestinian region and were in recent t imes cut with s ickles ( Dalman, 1 932, Vol.11, p .349-350). They were probably used i n prehistoric t imes a s food for humans and animals and in t he manufacture o f household articles such a s mats and baskets, such a s depicted i n Iraq and Egypt i n the 5 th millennium BC ( Wulff, 1 966, pp.219-222) S everal species have been f ound on early s ites in the Levant, including Bromus and S etaria ( Van .

Z eist, 1 970) at Mureybet and Bromus, Lolium and Phalaris at Tell Aswad ( Van Zeist and Bakker -Heeres, 1979) The l ast-named researchers a lso pointed out that wild grasses ( and s everal l eguminous seeds) may have been collected i ntentionally at Tell Aswad ( Van Zeist and Bakker -Heeres, .

1 979,

p .166).

Weeds are a ll plants which compete with crops and i nclude such plants a s grasses or poppies. In order t o a llow growth of the crops, weeds have to be e liminated f rom the f ields. This was often done by uprooting ( Dalman, 1 932, Vol.11, p .324) but a lso with the help o f knives, hoes or spades ( Dalman, ibid.; Wulff, 1 966, p .271). Dalman ( ibid., pp.328-329) l isted ancient references to weed management i ncluding the cutting o f thistles with s ickles. S everal genera, including Polyqonum fp Rumex and Chenopodiu r n ( which I cut experimentally) have been found a t Tell A swad ( Van Zeist and Bakker -Heeres, 1 979, p .162) Tell Abu Hureyra ( Hillman, 1 975) and Mureybet ( Van Zeist, 1 970) Boraginaceous seeds, such a s Arnebia decumbens, Buglossoides t ennuiflora and B .arvensis have been recovered f rom many early Near Eastern s ites, such a s Tell Abu Hureyra ( Hillman, pers.comm.). It i s not c lear whether the seeds were carried t here a ccidentally a s weeds, or whether they had been deliberately gathered for the sake ' of the red dye which t heir roots yield. However, in the latter case they would have been uprooted ( Hillman, pers.comm.). .

,

.

,

,

S everal species of S tipa including Stipa holosericea have a wide distribution including Mongolia ( S . qiqantia a nd P alestine. S eeds o f t his t all tussocked s teppe grass were abundant in t he Mesolithic levels at Tell Abu but rare in the Neolithic l evels. According to ( 1975) S tipa i s very s ensitive to overgrazing. ,

1 77

Hureyra, Hillman L ike t he

r eed, S tipa could be u sed f or a number o f purposes, s uch a s c onstructing shelters ( Hillman, pers.comm.) The s tems c an be pulled from the basal node, cut or e lse uprooted. The presence of S tipa in t he Judean uplands s uggested t o S auer ( 1958, pp.187-189) t hat the s errated blades f rom Jeri cho may have cut plants used i n crafts and building rather t han a s f ood plants. .

Scirpus maritimus i s found in coastal areas a nd on r iver banks. S eeds o f this plant have been found a t many s ites, such a s a t Tell Abu Hureyra ( Hillman, 1 975, p .71). I t i s unlikely , that t he s tems were used t o make basketry or mats a s stems have sharp s iliceous edges, but the s eeds which a re edible can be rubbed out by hand ( Hillman, pers comm.) .

Equisetum f luviatile a species o f horsetail, i s f ound in shallow rivers. The high s ilica content o f horsetails makes them ideal for polishing and s couring. I f ound i t easier t o cut E .fluviatile t han to uproot i t. ‚

Reeds ( Phraqmites communis emend. Lam.) proliferate in a ll the marshy regions and r iverbanks of the Near E ast and t heir distribution has probably not changed much s ince prehistoric t imes. Reeds are found near a ll the s ites I am dealing with: Kebara and El Wad in t he coastal l owland o f P alestine, Jericho in the Jordan Valley and Arjoune i n t he Orontes Valley. Hillman identified stem fragments o f reeds a t Tell Abu Hureyra ( Hillman, 1 975, p .71) Vita-Finzi a nd Higgs ( 1970, pp.21-22) t hought that i t had probably been reeds rather than cereals which had been cut a t Kebara and E l Wad ( see below, Section 8 ) Mat impressions have been reported from various s ites, such a s the Aceramic Neolithic a t Jericho ( Crowfoot, 1 982) Reeds were s till i mportant recently in everyday l ife; t hey were made i nto mats, baskets ( Wulff, 1 966 ‚ pp.219-222) and trays for o ilseeds i n o il-presses ( ibid., p .297). In recent years cane and bamboo were used for relatively heavy-duty thatching ( Wulff, ibid., p .107) a s well a s roof mats, blinds and curtains ( Wulff, ibid., pp.219-222) and a lso a s beater combs i n weaving ( ibid., p .195). However, a s bamboo i s not indigenous i n t he Near East, it i s l ikely that l arge r eeds or stipa were used instead. Reeds could a lso have been used a s arrowshafts ( Clark e t al. 1 974). Reeds ( see Wulff, ibid., p .107) must be cut on account o f t heir horizontal rhizomes. Curved knives were used in Iran to split a nd trim reeds a nd c ane u sed f or matweaving, basketry and comb making ( Wulff, 1 966, p .107 and 2 21). .

.

.

‚

Rushes v . bulrushes ( Juncus spp.) v . ( Cyperaceae) including S choenoplectus l acustris L . ( Palla) and S cirpus c f. t abernaemontanii are normally found on r iverbanks but, because of their over-exploitation by man, are now rare in the Near East, rather l ike Cyperus papyrus t he papyrus p lant. S cirpus c f. t abernaemontanii s eeds were f ound a t Tell Abu Hureyra ( Hillman, 1 975, p .71) Rushes were probably ‚

‚

.

1 78

used

in

the

( Wulff,

past

in

the

p .107 )

1 966,

sane

way

mats,

,

as

recently,

brooms,

for

thatching

baskets

( ibid.,

pp.219-222) and o ilseed trays ( ibid., p .297) rather harder reed. Like reeds, bulrushes have horizontal and must be cut. Sparganium banks of been used or

rushes.

ramosum

the Orontes for basketry It

too

has

grows

near

presentday

l ike the rhizomes

Arjoune

on

the

River. Its soft broad leaves may have and other articles rather l ike reeds to

be

cut.

Certain , species of Cyperus sp. are indigenous in Palestine and a high proportion of seeds found at Tell Aswad ( Van-Zeist and Bakker -Heeres, 1 979) were of this genus. Although mainly mentioned in Egyptian references, the variety of Cyperus papyrus was found by Dalman in Northern Palestine ( Dalman, 1937, Vol. V, p .22) . According to evidence from Egypt, the Bible ibid., p .31), papyrus

and accounts ( which grows

by Pliny ( see Dalman, to 3-4 meter height), was

used in Egypt to make virtually everything from ships, sails, blankets, and ropes to garments. The fact that parchment was developed when Egypt cut off papyrus supplies to the Greeks indicates that papyrus was not a common plant in neighbouring countries ( Ferguson, 1973, p .100 ) A few varieties of cane ( Saccarum . ) are indigenous in the Near East. Two varieties are found growing at present day Arjoune and are still used to make roofing mats. In the past

S .spontaneu xn

( Taeckholm and used to make ( ibid,

had

a wide

distribution

Drar, 1941, Vol.1, mats. Cuims were

in

the

Near

East

p .489). In Egypt cane was found in 5th Dynasty tombs

p .490).

Legumes, such as vetch, wild and cultivated, ( Vicia ervilia the Bitter Vetch, Vicia sativa the Common Vetch, Vicia faba the Horse Bean) have been found in many early ‚

‚

‚

levels

of

,

Near

Estern

s ites,

for

example,

Vicia

ervilia

at

Tell Mureybet ( Van Zeist, 1970) and at Mesolithic Tell Abu Hureyra ( Hillman, 1 975) and Vicia faba at Neolithic Tell Abu Hureyra ( ibid.). Vetch i s still harvested by uprooting, for example at Tell Toqaan in Syria where it is used as animal fodder ( Sweet, 1 960, p .70). Wild lentils ( Lens orientalis) were found at Mesolithic sites, for example at Tell Abu Hureyra ( Hillman, 1975, p .71) and cultivated lentils ( Lens culinaris ,

and/or example and

chickpeas ( Cicer arietinum) at Tell Abu Hureyra ( ibid.),

Bakker -Heeres,

1979),

Tell

at Neolithic s ites ( for Tell Aswad ( Van Zeist

Mureybet

( Van

Zeist,

1970)

and at Jericho ( Hopf, 1983, p .609)). Pulses apparently became a staple in Syro-palestine. As far as I am aware pulses are a lways harvested by uprooting ( Dalman, 1937, Vol.111, p .34; Turkovski, 1969) Varieties Linum

bienne

of

both

wild

flax

( several

) and domesticated flax

1 79

species,

including

( Linum usitatissimum

grew i n P alestine when Dalman c arried out h is s urvey a t t he beginning of t his c entury. The domesticated f orm was f ound on many Neolithic s ites i n the Near E ast. F rom i ts l ikely ocurrence at Ramad, Van Z eist and Bakker- Heeres ( 1975, p .218) concluded t hat cultivation o f f lax must have s tarted i n the s econd half of the 7th millennium BC i f not e arlier. Domesticated f lax grows t o a height o f 1 meter. I t needs humidity or i rrigated land ( Dalman, 1 937, Vol. V, p .19). The s tems contain f ibers u sed for weaving ( ibid.) and u sed a s wicks f or o il-lamps ( ibid., p .28), while t he seeds yield l inseed o il ( ibid., p .19). Dalman r eferred t o G . Crowfoot's observation of f lax harvest in Egypt: f lax was uprooted, combed and l eft in water for f ifteen days. I t was then dried and l eft f or two years and subsequently beaten with wooden mallets and combed again to extract the f ibers ( ibid., p .20 ). According to Egyptian pictorial references ( ibid., p .29) and biblical references ( ibid., p .26 ), f lax was uprooted a s i t i s nowadays ( E. Crowfoot, pers.comm.). However, the related t o f lax ( Donner

Gezer Calendar used t he word the Hebrew word f or knife, in and Roellig, 1 964, p .181).

' zd, which i s connection with

Sesame ( S esamum indicu r n ) i s cultivated for i ts s eeds and the o il contained in the seeds."The o il, the s eeds and even the l eaves have been a scribed s everal medicinal and other desirable properties." ( Sim r nonds, 1 976, p .231). Archaeological evidence indicates that sesame was cultivated in Palestine and Syria as early a s 3 000 BC ( ibid., p .232). However, no wild species of sesame have been found in t he Near East according t o S immonds; he considered Africa a s a possible region where s esame originated ( ibid., p .232). Other o il plants i nclude hemp ( Cannabis sativa ) and the opium poppy ( Papaver somniferam var. globrum ) a ll of which are found in the Near East today. The seeds o f t he s esame a re a lso used in baking, and the hemp and t he opium poppy have much publicised narcotic qualities. However none o f these plants have been f ound on Syro-palestinian s ites until a date much later than the periods I am dealing w ith. Dalman ( 1937, Vol. V, p .30) thought that hemp was imported f rom the Hellenistic world, a s he could not f ind any biblical references. Nevertheless, opium had been important in Egypt at an early date and was probably t raded i n the poppy-shaped"base r ing"jugs in the Late Bronze Age ( Merillees, 1 962) My own experiments harvesting two poppy species ( see below, S ection 4 ) suggested that t hey were difficult to cut, but easily uprooted. The s eedheads were easily snapped o ff by hand. My own f eeling therefore i s t hat poppies were not harvested with f lint blades. ‚

.

V ines a nd F ruit t rees The vine ( Vitis ) grew wild a long Levantine r iver valleys. Remains o f the wild and the cultivated v ine were found f rom Neolithic t imes onward on Near Eastern s ites,

1 80

e . g. at recovered

Tell from

Abu Hureyra ( Hillman, 1 975) the 5 th millennium BC site at

.

Grape Arjoune

pips may

have come from the wild or the cultivated vine ( Moffat, in prep.) The fruit of the vine may have been eaten as raisins ( Helbaek, 1958) or drunk as juice, or may perhaps even have been fermented as wine. Special knives and sickles, both serrated and unserrated, were used in Palestine to trim ,

vines ( Dalman, 1937, Vol.111, p .23) and fruit trees, such as hawthorn ( Crataegus sp. ) pistacio ( Pistacio sp. ) plum or cherry ( Prunus ap) fig ( Ficus sp ) and sweet a lmond ( Amyqdalus dul cis ) listed by Moffat ( in prep.) as occurring in the 5th millennium BC s ite at Arjoune. ‚

.

4 .

‚

.

E xperiments One

or and

Hundred

and

eighty-four

sickle

blades,

unretouched

with f ine or coarse denticulations, 11 straight sickles 7 curved sickles with both unretouched or f inely or

coarsely denticulated plant species, which in dried: Cereal

blades were used to some instances were

cut the following both fresh and

species:

Wild barley ( Hordeum spontaneu r n C . Koch) ( not shown, see below), domesticated einkorn wheat ( Triticum monococcum L .) ( P1.22: g, h) domesticated barley ( Hordeum disticum L . emend. Lam.) ( Pl.22:c, d), domesticated emmer wheat Triticum dicoccum L . ,) ( Pl.22: e, f), domesticated spelt ‚

Triticum

spelta

Non-cereal

L .)

( not

shown).

species:

Grass ( several Anthoxanthum

species including the Sweet odoratum L .)) ( Pl.25:c, d),

Vernal weeds

Grass ( several

species, including Polyqonum convulvulus Chenopodium album Rumex crispus Seh ne alba Potentella arvensis Mentha arvensis Sonchus arvensis, all of which grow in temperate climates but some of which grow in the Near East ‚

‚

‚

‚

‚

( Pl.25: g, h)), a tussocked steppe grass ( Stipa qiqantia ( Pl.25:a, b), horsetail ( Equisetum f luviatile ) ( Pl.25:e,f), reed

( Phraqmites comrnunis

Trin.)

( P1.23:2-h),

bulrush

Schoenoplectus lacustris L . ( Palla)) ( Pl.23:a, b), Sparganium ra r nosu i n ( Pl.24:e, f) Cyperus lonqus ( Pl.24:a, b) cane ‚

‚

Saccarum ) ( Pl.24:c, d), f ield poppy ( Papaver rhoeas ) ‚ oriental poppy ( Papaver orientale ) ( Pl.24: g, h) and various species of wood including wood of the cherry tree ( Prunus sp. ) ( Pl.26: g). .

The experiments were carried out between May 1981 and June 1984 in England, Wales and Syria. In Syria non-cereal species were harvested on the banks of the Orontes River near Arjoune, while cereal species were harvested at the International Centre for Agricultural Research into Dry Areas ( I. C. A. R. D. A.) in Aleppo. Wild barley was harvested by Miller-Rosen in Jersualem. In addition Hillman allowed me to

1 81

examine a s ickle he had used t o cut e inkorn, mostly wild xamine a ( Pl.22:a, b) i n Wales, while Harding a llowed me t o e w heat and s ickle which he had used t o harvest domesticated barley i n England. Bar-Yosef and K islev k indly s ent me w ild and s ickle b lades which they had used t o cut t hreshing domesticated barley. I n addition I examined 4 f lints which had been used i n Cyprus. ‚

The

S ickles

A variety of s ickle hafts, mostly f ragmented, have been recovered f rom Mesolithic and Neolithic s ites i n t he Near East ( see t he comprehensive survey by Camps-Fabrer and Courtin, 1 982) S ickle hafts were made o f various materials, such a s bone, wood, antler, c lay and horn, and were e ither rectilinear or curved in shape. While s ickle hafts have been recovered f rom Palestine, e . g. f rom Kebara and E l Wad ( ibid.), no remnants o f such hafts have so f ar been f ound in Syria a s far a s I know. This indicates that s ickle hafts in Syria may have been made o f wood. Circumstantial evidence for s ickle hafts comes f rom Byblos, where Cauvin ( 1968, pp.70-73, pp.100-105, pp.128-133) calculated the s ize, shape and l ength of composite s ickles f rom t he number and shapes o f s ickle e lements, and postulated a change f rom rectilinear s ickles in the Neolithique Ancien t o both rectilinear and curved s ickles in t he Neolithique Recent. A s imilar case for the coexistence o f two types of s ickles i s made f or t he s ickles at Arjoune ( see below, Section 7 ) I therefore used two types o f s ickles i n my own experiments: .

.

A rectilinear s ickle ( Fig.7:a) 5 0 cm long ( made o f cherry wood a fter a s ickle f ound in t he Fayum i n Egypt, s ee Camps-Fabrer and Courtin, 1 982, f ig.5) in which 8 -10 blades were hafted parallel t o the s ickle edge. -

A curved s ickle ( Fig.7:b), 2 2 cm l ong ( made o f red deer antler by Newcomer a fter a s ickle excavated at Hacilar, s ee M .-C. Cauvin, 1 983, f ig.5:8), in which 4 -5 blades were hafted parallel to the s ickle edge. -

The hafting agent I had used i n the f irst 4 experiments was Uhu, a commercial c ellulose adhesive, but i t proved t o be not very strong and t he b lades f ell out quickly. Another problem was the impossibility of removing t he adhesive with acetone without leaving a n"acetone b loom"which obscured t he wear-traces on the f lint surfaces. The hafting agent used i n a ll t he subsequent experiments was resin mixed with wax in a proportion o f approximately 4 :1. This mixture set quickly and was very s trong: the b lades were f irmly hafted. The resin and wax could be removed easily by immersion i n white spirit without a ltering t he wear-traces on the f lint.

1 82

The Work All harvesting i mplements were approximately 5 m inutes ( 200 s trokes) and s trokes) mostly with both unidirectional cutting movements. ,

u sed between f our hours ( 15,000 and b idirectional

The non-cereal species were cut a ll t hrough the year and in s ome i nstances a fter s torage. They were usually cut a t the base of the s tem. Fresh and s easoned wood was s awn. The cereal species were usually cut when r ipe, a lthough domesticated barley was i n one instance cut green, three weeks before harvest was due. Sometimes cereals were cut away from the f ield and a fter storage. Cereals were harvested in June in t he Near East and i n August- S eptember i n England and Wales. They were u sually cut at t he base o f t he cuim. The bundles. upturned habitat

non-cereal species were No other plants were soils i n which cereals o f weeds. The proportion

harvested s ingly or in cut at the same t ime. The are sown are a lso the of weeds t o c ereal s talks

ranged between l ess than 5 t o 2 0% ( see Table 6 ) I rarely cut any weeds t ogether with t he c ereals, a lthough I cut weeds separately i n order to examine t he wear-traces f rom t hese ( see Part I , Chapter 1 4, P l.25:g, h ‚ Table 2 ) When wild barley was harvested i n Jerusalem, wild oats were cut a s well ( Miller-Rosen, pers.comm.); however, no wear-traces could be d iscerned on these blades ( see S ection 4 ) No weed counts were g iven f or the experiments by staff a t I .C. A.R.D. A. at Aleppo. .

.

.

Efficiency When non-cereal species were harvested, s ingle unhafted blades could be used with ease, provided t hey were quite l arge and had backing retouch. The most e fficient composite s ickle f or plants growing in dense stands, such a s reeds and bulrushes, was the short curved s ickle which was best used t o reap rather than to cut or s aw. Unretouched f lint edges a lways cut best. However, on wooden branches or hard stemmed plants such a s r eeds, unretouched edges became damaged within s econds of use whi1e denticulated f lint edges did not. No plot s izes were calculated. Only thin branches o f wood could be efficiently s awn with f lint blades. Poppies were much easier uprooted than cut. When cereal species were cut, s ingle unhafted blades could be used, but composite s ickles were far more e fficient. A s stands were not a s dense a s those o f the marsh plants, t he long r ectilinear s ickle could be easily used and proved more e fficient t han the short curved s ickle. The former was best used f or cutting, not f or reaping. Again, unretouched f lint edges were the sharpest, but wore out too quickly. Coarsely denticulated b lades could not be used, a s the culms got caught in t he denticulations and were pulled

1 83

out. For s imilar reasons blades had t he haft, at l east in my experiments.

t o

t ouch end t o end i n F inely denticulated

blades were t he most e fficient and l asted l ongest. This principle i s s till recognised t oday: modern c ombine harvesters have f inely denticulated s teel blades. At

I .C. A.R.D. A.

two

rectilinear

wooden

s ickles,

hafted

with f lint e lements with f inely denticulated cutting edges, were u sed by a s ixteen year o ld girl, an experienced f arm worker. She harvested two rows of domesticated barley and t hree rows o f domesticated wheat, by holding bundles of culms f rom two rows o f the same crop t ogether. Each row was 7 0 cm l ong and t he rows stood 2 5 cm apart. Harvesting barley with t he f lint s ickle took her 2 3 minutes, while harvesting the same s ized plot with a modern metal s ickle t ook her 1 2 minutes. Harvesting wheat with the f lint s ickle took her 1 4 minutes, while harvesting t he same s ized plot with a modern s ickle took her 1 1 minutes. She declared that the f lint s ickle was " acceptable". At Butser Hill it took me three hours to cut 1 05 square meters of domesticated e inkorn, one hour and 3 0 minutes t o cut 3 5 square meters of domesticated e r nmer and one hour t o cut 4 5 square meters o f domesticated spelt with rectilinear f lint s ickles ( see Table 6 for ste in counts) The work took longer t han expected a s it was raining heavily at the t ime. .

All of t he than they were, reset.

f lint s ickles could have been a lthough a f ew blades f ell out

used longer and had t o be

Korobkova ( 1981, p .340) l isted the efficiency of the models o f prehistoric s ickles her experienced t eam had used: the most e fficient, a gently curved s ickle of the Late Tripolye type with sawtooth f laking on the edge, had an average productivity o f 1 .1 square meter per minute. I f we calculate that the worker a t I .C. A. R.D. A. would have t o harvest 8 rows of 70 meters i nstead of 2 rows of 7 0 meters ( which meters,

stood then

2 5 it

cm apart) would have

in order to taken her

harvest 7 0 square approximately 9 2

minutes to harvest 7 0 square met ers o f barley and 5 6 minutes to harvest 7 0 square meters o f wheat, a mean of just under 1 square meter per minute. I therefore conclude that the rectilinear wooden s ickle in which s everal f inely denticulated f lint e lements had been inserted parallel t o t he

f ell

haft

was

quite

e fficient.

Hafting was very much a matter out i f they were not hafted

of in

experience: blades a s traight l ine with

their ends touching. The proportions o f resin to wax a lso proved important: when too much wax was added, the b lades e specially when used in a warm c limate became covered with wax and no wear-traces could be seen. This happened with t he experiment cutting wild barley in Jerusalem. I t a lso l ed t o a l oosening o f the blades and i t i s therefore unlikely that -

-

1 84

i t i s t he cause o f the a bsence o f g loss on a ncient b lades otherwise shaped l ike s ickle blades. When too l ittle wax was a dded, e specially in a c old c limate, t he resin became too brittle

and

broke

away.

Results The species

wear-traces a re shown

resulting on Table

f rom working different plant 2 , P lates 2 2-25, and F ig.39,

4 0: a , c , e , f . The edge damage ( see Part I , Chapter 4 ) on the blades depended on the hardness o f the plant stems and t o some extent on other variables such a s the moisture content of the plant and on t he type o f f lint from which the blade was made. Soft-stemmed plants l ike bulrushes l eft hardly any damage, while hard-stemmed plants l ike reeds and wood left considerable

damage.

Polishes ( see Part I Chapters 1 4 and 1 7) differed in t heir horizontal and vertical distribution a ccording to plant species, and in their brilliance mostly a ccording to the moisture content o f the p lant. They a lso differed to some extent according t o c ertain variables such a s t he f lint ( see Part I Chapters 1 5 and 1 6) ‚

‚

S triations

( see

Part

I

‚

Chapter

9 )

were

only

s een

on

polishes f rom contact with plants growing on land, and most were seen on polishes f rom contact with cultivated plants harvested from upturned s oil ( Pl.22:c-h). I t i s most l ikely that polishes from plants cultivated in dry loe ss soils would be the most striated ( see Korobkova's experiments in Moldavia and Odessa Oblast ( 1981, p .331). S ome striations were a lso s een on wood polishes ( see Part I , Chapter 1 4) A "sandblasting"effect polishes from plants

was a lso harvested

only f rom

observed upturned

on soil

( Pl. 2 2: e , f ) Macroscopic G loss ( see Part I Chapter 1 0) nearly a lways began at the same t ime a s microscopic polish but when coarse-grained f lint was used gloss appeared much later than microscopic polish. As with the polishes, the rate o f gloss development varied a ccording to plant species, and moisture content of the plant, a s well a s other variables, such a s the f lint t ype. Thus the beginning o f gloss on Brandon f lint was visible when f resh reeds had been cut f or 2 00 s . m. and when f resh barley had been cut f or 4 000 s . m.. The only p lants which so f ar have not caused any gloss were the two poppy species which up t o 1 500 s . m. produced only ‚

very

f inely

streaked

polish.

No different rates o f development could be detected between g loss and polishes f rom p lants harvested i n England and Wales and those harvested in the Near East.

1 85

No h afting t races c ould be d iscerned. Hafting was i ndicated by t he regularity o f polish d istribution a long t he cutting edge and by the f act t hat polish stopped abruptly a t t he ends o f t he b lade a nd did not a ffect t he e nds themselves. A sharp demarcation between polished and unpolished f lint ( where t he haft had covered the b lade) was not visible on my experimental b lades. Problems I d iscovered analysis dealing in three s ections: P roblems

related

t o

many with

t he

problems related s ickle blades which

t o may

a m icrowear be discussed

plants:

P lants may have died out or may have been overlooked. S ome plants which are commonly uprooted may have been cut in t he past. P rehistoric plants may have differed s lightly i n t heir morphology and physiological balance, with broader or narrower s tems or with s lightly different s ilica content f rom t oday. This may have been the c ase with t he domesticated cereals which a re nowadays bred t o be s ilica-free. S ilica and moisture content o f p lants may vary f rom area to area and wear-traces may d iffer s lightly a ccordingly, but this did not s eem to be t he case i n my experiments with reeds, barley and Sparganium ramosum. S ilica and moisture content do vary a ccording to the t ime o f year and the weather, and we a re not c ertain a t what t ime o f t he year cereals were harvested and what t he weather was l ike. Cereals may well have been harvested when unripe, and s tored. However, I did harvest p lants at varying t imes o f t he year, in order to f amiliarize myself with varying wear-traces. -

P roblems

related

to

the

work:

I could not do t he experimental harvesting o f c ereals myself in the Near East. The blades used at I .C. A.R.D. A. were not available for me to examine. The b lades used a t Jerusalem had no wear-traces on t hem because the wax had run. -

Problems

related

to

the

results:

In my experiments t he polishes from different p lant species ( see Part I , Chapters 1 4, 1 5, Table 2 , P 1.22-25) varied s lightly a ccording t o species. However, a s pointed out in P art I , Chapters 15 a nd 1 6, i f worked materials have s imilar properties, polishes can l ook s imilar. This was f ound t o be t he case with t he polishes from reeds and Stipa q iqantia ( see Table 2 , P l.23:c, d and P 1.25:a, b) As a lso pointed out i n Part I , Chapters 1 5 and 1 6, there are many variables other than t he species which can a ffect polish f ormation. I t i s therefore possible that polishes f rom t he s ame p lant species can differ s lightly. Polishes f rom dry -

.

1 86

p lants ( Pl.23:e,f) were i ndistinguishable f rom e ach other, a nd could a lso b e confused with t he b eginning of wood polishes. They were very weak and chances are that they would not be d iscernible on ancient t ools. Sometimes wear-traces showed a c ombination o f f eatures which was not c onsistent with t hose of the wear-traces observed on t he experimental b lades: e .g. t here were wear-traces most o f whose f eatures s uggested use on rushes ( see Table 2 ) ‚ but t here were a lso many striations which would indicate plants growing i n upturned soil. Here several possibilities o ccurred t o me: a ) several plant species may have been cut w ith t he s ame s ickle; b ) an unknown p lant species may have b een cut; c ) c onditions such a s soil t ype or presence of weeds may have been different; d ) s light changes in the p lant morphology may have occurred; e ) post-depositional damage may have o ccurred. In such cases I c lassified the polish a s " plant"polish. The presence o f striations could i ndicate that plants had been cultivated. Equally, however, s ome plants, such a s barley, occur wild in disturbed soils, s uch as s crees. S triations are therefore not an i nfallible i ndicator o f cultivation. Conversely the l ack o f striations on polishes, otherwise t hought to originate f rom contact with c ereals, c ould indicate that plants had been cut some d istance f rom t he soil, perhaps halfway up t he stem, or e lse t hat the soils had not been loose, perhaps because o f a damp s pell or because d igging sticks were used which do not produce complete upturning o f the soil. Consideration of these problems l ed me t o conclude that many ancient b lades have t o . be examined i n order to avoid m isinterpretation of wear-traces, and that r esults of t he m icrowear analyses should be compared with t he extant p lant remains f rom the s ites, i n order to see whether t he results c orrespond. 5 .

Blades with Gloss

from Arjoune

Aims ( Pl.26: b, d,f,h,

F ig.40: b, d, f , h,

4 1)

All t he excavated lustered blades ( 38) a nd a ll t he available ( 40 out of 6 0) excavated f rom Trench V were examined with under the l ight microscope. A f ew of the f rom both t renches, c lassified by Copeland i n prep.) a s s ickle blades on account of examined f or comparison.

f rom Trench VI lustered blades t he naked eye a nd unlustered blades ( 1 981, p .11, and t heir shape, were

In t his microwear analysis I concentrated on t he a ssemblage f rom Trench VI a s a ll the f lint implements were available f or study and f or this reason I d iscuss Trench VI b efore Trench V . I had

s everal

main

a ims:

1 87

I wanted t o investigate whether t he b lades had been u sed hafted. Apart from a ffecting t he e stimation o f t ool e fficiency ( Korobkova, 1 981) t his f actor has a lso implications for the i nterpretation o f t he nature o f the s ite: a concentration o f s ickle e lements might indicate that the s ite i s an undisturbed l iving s ite ( see Part I I, Chapter -

2 .) As no s ickle hafts have been excavated f rom t his s ite or, as f ar a s I am aware, f rom any Neolithic t o Chalcolithic Syrian s ite, I wanted to t ry and reconstruct t he shapes o f s ickles, t he hafting a rrangement and t he movements used, using f eatures such as t he shape o f the b lades and t he polish distribution.

-

I wanted

-

edge

o f

t o

f ind

a blade

out

was

at

what

s tage

I wanted t o investigate whether the exclusively on cultivated cereals

-

been

cut

f resh

I wanted microscopically -

on

plants

or

or

and

why

t he

cutting

retouched. b lades had been used and whether plants had

dried.

to in

investigate order to see

t he unlustered blades whether they had been used

not.

I wanted t o f ind out whether the b lades had been worn out and to calculate the minimum duration o f use f rom the edge damage and the amount o f polish. However, a s ancient c ereals were probably more s iliceous than t hey are now ( see S ection 4 , Problems) the t ime could not be e stimated very -

,

accurately. Results from this investigation which relate t o general questions about the nature o f t he s ite and t he f lint assemblage, such as the spatial distribution o f t he burnt blades and a comparison o f t he microwear with t he typological analysis will be discussed in the relevant general

chapters.

Results

Hafting

The blades were measured ( see F ig.8:c). Relatively short, straight blades o f uniform width and thickness would, I believe, have been used hafted. ( The thickness o f a blades i s e ffectively increased by i ts c urvature, when v iewed in profile, and this was t aken into a ccount in the measurement o f the thickness.) " Truncation", accompanied by retouch and/or l ikely intentional breaks at the distal and/or proximal ends of the blades indicated to me that t he blades had been intentionally shortened t o be used a s s ickle e lements. Hinge f ractures were counted a s i ntentional breaks, a s the thus shortened blade had been chosen t o be used. Another f eature which indicated t o me that t he blades had been used hafted was a uniform orientation o f microscopic

striations.

1 88

Trench VI Out of o f b ladelets

3 8 l ustered t ools, 3 5 were ( less t han 1 .2 cm wide).

blades

a nd

3

were

The l ength o f the b lades varied from 1 .5 cm to 6 .7 cm. Twenty out o f 3 8 blades were between 2 .0 cm and 4 .0 cm and 1 3 b lades were between 4 .0 and 5 .9 cm long. Only 2 out o f 3 8 b lades were shorter than 2 .0 cm and only 3 were longer t han 6 .0 cm. A ll the b lades, with t he exception of 1 , were e ither t runcated or snapped a t one or both ends. The width o f the b lades ranged f rom 0 .9 t o 2 .6 cm. Twenty nine b lades were b etween 1 .5 and 2 .0 cm, and 6 b lades were between 2 .1 and 2 .6 cm wide. The gap between t he widest bladelet ( 1.0 cm) a nd t he narrowest b lade ( 1.5 cm) suggested t o me a different u se f or t he bladelets. The e ffective thickness ranged f rom 0 .2 t o 1 .2 cm. The majority o f the blades ( 34) were between 0 .3 and 0 .9 cm t hick. The thinnest ( 0.2 cm) was one of t he b ladelets, and t he 3 thickest blades ( 1.0 t o 1 .2 cm) were a ll over 4 cm long. The curvature of the profile was minimal ( less t han 0 .1 cm) in 3 2 of the blades. Four b lades had a curvature o f 0 .1 cm to a t hird of their thickness, only 2 b lades had a s trong curvature o f a t hird o f t heir t hickness or more.

s ome

The measurements standardisation

of t he b lades o f l ength,

from Trench VI suggested width and thickness. All

except 4 blades had been intentionally shortened and of t hese 4 , 3 had broken a fter use ( and therefore may have been i ntentionally shortened before accidental breakage) Most b lades were straight i n profile and usually rectilinear in plan with s traight t runcations. Two blades had s lightly curved backs, 3 blades were s lightly narrower at the proximal end and 3 blades had pointed ends indicating t heir u se a s end pieces. Only 2 blades were exceptionally shaped in that they had t angs, one of these being the unbroken blade referred to above. Three bladelets may have been used s eparately f rom the rest. .

The microscopic examination showed uniformly oriented striations parallel ( Pl. 2 6 : b, d )

that 3 4 blades had t o the cutting edges

Trench V It i s l ikely t hat o f the 4 0 tools examined a ll had been blades before being shortened by truncation or snapping. One blade had neither been truncated nor snapped. S ix b lades were probably broken a fter use. There was only 1 bladelet. One s ickle blade had been made into a s craper ( see Part I I, Chapter 1 0) The l ength o f the blades ranged f rom 1 .6 t o 6 .2 cm., but 2 5 out o f t he 4 0 blades were between 2 .0 and 4 .0 cm long. Eight blades were between 1 .6 and 1 .9 cm, 6 blades were between 4 .0 a nd 6 .0 cm and only 1 blade was over 6 .0 cm

1 89

l ong. t ools

Four o f the 6 pieces which were probably f ragments o f were between 1 .6 and 1 .9 cm l ong. This could decrease

the shortest group to only 4 t ools, t he other 2 f ragments being 2 .1 and 3 .1. cm l ong. The width o f t he blades ranged f rom 1 .1 t o 2 .5 cm. Twenty-four blades were between 1 .5 and 2 .0 cm wide, while 1 1 b lades were between 1 .1 and 1 .4 cm and 5 blades were between 2 .1 and 2 .5 cm wide. The t hickness o f the blades ranged f rom 0 .3 t o 1 .4 cm. Thirty-seven b lades were 0 .3 to 0 .9 cm thick, only 3 o f t he blades being exceptionally t hick. The curvature ( in profile) o f most blades was negligible. Only 3 b lades had a curvature o f between 0 .1 . cm and a third o f their thickness and 4 blades had a curvature o f over a third o f t heir thickness. The measurements o f blades f rom Trench V suggested a standardisation o f l ength, width and thickness. Excluding the f ragments, most blades were intentionally shortened and o f rectangular shape. There were, however, a lso 6 s lightly pointed pieces, probably end p ieces, and a lso 3 curved, 1 triangular and 3 t rapezoidal e lements. Most b lades had uniformly oriented s triations running parallel t o t he cutting edge and therefore most blades appeared t o be s ickle elements, probably hafted in both straight a nd curved s ickles, a lthough a f ew, l ike a bladelet w ith perpendicular striations and a tanged blade, may have been individually used. Comparison Apart f rom the more variable shapes o f t he s egments i n Trench V . t he differences between the t ools o f the t wo trenches were slight: the blades from Trench VI t ended t o be s lightly l onger and wider, but not thicker than t hose f rom Trench V . There was a gap between bladelets and blades f rom Trench VI which was not observed in Trench V . Sickle

Shape,

Hafting Arrangement

and Movements

during Use.

The shapes o f the s ickles and the hafting a rrangement inferred f rom the shape o f the blades ( usually b y b acking r etouch ) t h e a ng le o f t he t runcat ions determined at the ends, f rom the d istribution of polish and or breaks

were

,

spects o f the blades, and gloss over the dorsal and ventral a direction o f t he microscopic s triations. Movement from t he rom t he orientation of t he during use was reconstructed f striations. Trench VI The

shapes

of

most

o f

t he

3 8

b lades

were

rectangular,

a lthough some were pointed, and two b lades had s lightly curved backs. The angle o f t he t runcation t o the back was usually 9 0 degrees, t he exceptions being t he 2 b lades with t angs. The

d istribution

o f

polish

1 90

a nd

gloss

varied;

on

3

b lades r etouch h ad obliterated t he polish a lmost completely and t he d istribution had t o be c lassified a s unknown. S ixteen b lades h ad a s traight and narrow band of polish a long the c utting edge, which t erminated abruptly at the ends. This indicated t hat when used t he blades had been i nserted i n a s traight or gently curving haft. E ight b lades had a wide distribution o f polish; 1 b lade with unusual s triations and w ith polish extending over the ends had probably been h and held. Two blades ( one with a t ang, P l.26:f) w ith unusually oriented s triations and with polish which s topped before t he proximal end were probably hafted at one end. The o ther 5 blades, with a wide distribution o f polish but uniformly oriented striations, might have protruded f rom t he haft, or a lternatively t he wide polish distribution could indicate t hat thick-stemmed p lants were been cut. This l atter possibility i s l ess l ikely s ince the polish on these b lades looked l ike the experimental c ereal polish f ound on most o f the blades. E ight blades with s lightly curved polish and 3 blades with diagonally distributed polish were probably used while inserted in a curved haft. A ll of t hese blades were straight in profile and therefore the uneven distribution o f t he polish was not due to a curved profile. All the blades had microscopic striations parallel to t he cutting edge which i ndicated that the hafting had been parallel and that the s ickles had been used with cutting or sawing rather than with reaping movements. Trench V Most

o f

the

4 0

blades

approximately 9 0 degrees had a tang, 6 blades were blades were c urved, t rapezoidal.

were rectangular with t runcations to the cutting edges, one o f t hese s traight but s lightly pointed, 3 1 was t riangular, and 3 were

The d istribution of polish and gloss varied; on 8 blades t he distribution could not be c lassified because of considerable edge damage. Eighteen blades had a straight l ine o f polish j ust a long the very cutting edge, indicating that most had been hafted in a s traight s ickle. The exception was a bladelet with perpendicular s triations. F ive blades had a wide polish distribution. Two of t hese blades had a polish which t erminated abruptly halfway down the blade, i ndicating-that the blades had been hafted a t one end. One o f t hese was t riangular and very t hick. Four blades had a curved polish distribution, 2 o f t hese were curved i n profile and t he polish d istribution could be due to t his f act rather than the shape of t he s ickle. Three blades had a d iagonal d istribution o f polish. A ltogether 5 blades, 2 with t he curved polish and 3 with the diagonal polish, had probably been inserted in a curved haft.

1 91

Comparison The results of the i nvestigation were s imilar f or t he b lades f rom Trench VI and Trench V i n that most e lements had probably been hafted i n s traight or only s lightly curved handles, possibly made o f wood. B oth antler and bone were preserved on t he s ite and i t i s t herefore unlikely t hat s ickles made o f t hese materials h ad not been discovered. Some e lements had probably been hafted in c urved s ickles, especially t hose f rom Trench V which were backed t o be curved or t rapezoidal. The s ize of t he blades ( Fig.40-41) suggested d ifferent s ized s traight, or e lse pointed s ickle handles. However, given the poor quality o f t he f lint ( see Part I I, Chapter 7 ) found near the s ite, the different s izes of e lements may reflect necessity rather t han choice. The s ickles s eemed to have been used with c utting or sawing rather t han reaping movements. A f ew other plant cutting blades had probably been u sed hand held and some had probably been hafted at one end. The results f rom the different t renches only differed i n a s much a s t he blades from Trench V were s lightly l ess standardised in shape. Retouch I was i nterested to discover whether retouch had been intentional, and i f intentional, whether i t was done before use, perhaps in order to cut specific plants, or whether i t was done during or a fter use, i n order t o resharpen the blade. The i nvestigation a ssumed that intentional retouch i s regular and occurs before and during use, and that unintentional retouch i s irregular and occurs mostly during and a fter use. With intentional retouch, the amount o f microscopic polish would equal the polish on the other a spects of t he blade, while with unintentional retouch, the amount of microscopic polish on the s cars i f not absent, would be considerably l ess than t he polish on t he other a spects of t he blade. ‚

Retouch proved a very difficult a spect t o investigate, s ince there was no sharp division between regular and irregular retouch on the blades f rom Arjoune, and s ince i t was difficult to see microsccpic polish i nside t he retouch scars because of the oblique angle o f the scars to the microscope lens. I t was a lso difficult to decide whether a smaller amount of polish found on t he s cars, compared to the polish on the ventral and dorsal a spects, was due to t he scars f orming during rather than before work, or whether this was due to the oblique angle o f the s cars to t he worked material. The microscopic examination was therefore carried out before the independent examination o f the retouch with t he naked eye in order to avoid bias. The results were then compared.

1 92

Trench VI The microscopic examination indicated that 1 8 b lades had been retouched before u se, 9 blades had been retouched during and a fter u se and 1 1 b lades had been retouched e ither before, during or a fter u se. The macroscopic examination revealed that 2 3 b lades had regular r etouch. O f t hese, 5 blades were definitely denticulated and 2 seemed t o have been heavily retouched, t hat i s, probably resharpened. S ix blades had i rregular, probably unintentional retouch and retouch on 9 b lades could not be c lassified. The f act that a ll 5 t ools c lassified a s denticulated were a lso c lassified under t he microscope a s r etouched before use, and that the 2 resharpened tools were t hought t o have been retouched during and a fter use, and that most t ools c lassified a s retouched before use were a lso t hought t o have been intentionally retouched ( 16 out o f 1 8) s eemed t o indicate that this approach was quite productive. Trench V The microscopic investigation suggested that 1 9 blades had been retouched before use, that 1 0 blades had been r etouched during and after use, and that 1 1 blades had been r etouched before, during and a fter use. The macroscopic investigation indicated that 2 4 blades had been i ntentionally retouched. Of these, 5 blades had been denticulated and 2 had probably been resharpened. E ight b lades had irregular and therefore probably unintentional retouch, while on 8 blades retouch could not be c lassified. Most o f the t ools c lassified under the microscope a s retouched before use were a lso identified as intentionally retouched ( 17 out o f 1 9) All t he blades c lassified a s denticulated were thought t o have been retouched before use .

a s

well.

Retouch

and Plants

There s eemed t o be a continuum from f inely retouched t o denticulated blades, and i t was therefore difficult to examine the relation between denticulation and plants cut. The denticulated blades f rom Trench VI showed t he c ommon"sickle polish"(see below) which was probably caused by cultivated grain. In only one case did the polish look l ike experimental reed or S tipa polish. Three of the 5 denticulated blades f rom Trench V showed unusual polishes which l ooked l ike experimental reed or Stipa and wood polishes. Conversely, however, uncommon polishes were a lso s een on the common f inely retouched blades. B lades

c lassified

a s

unretouched

1 93

f rom

both

Trenches

showed polishes probably due c ereals, a s well a s t o the B lades with i nverse retouch VI

a lso

( see

showed

both"sickle

t o the harvesting o f c ultivated harvesting o f other p lants. which were only f ound i n Trench

polish"

and

other

p lant

p olishes

below)

I n summary, i t s eemed that b lades had not been r etouched with certain plants in mind, but a f ew b lades which had been denticulated were probably u sed t o cut r eeds or

S tipa

and

wood.

Plants Polishes from different plant species were f ound t o be s lightly different ( see Part 1 , Chapters 1 4, 1 5, Table 2 and P lates 2 2-25) but differences were not a lways c learcut ( see above, S ection 4 ) I therefore c oncentrated on distinguishing polishes which resembled experimental polishes made by cutting cultivated cereals f rom polishes ,

.

made by processing other plants. The f act t hat t he Arjoune material was made o f different f lint types has been t aken into account, but no consistent correlation was found between a particular f lint type and the i dentification o f a particular

plant

species.

Trench VI Of the 3 8 blades excavated from Trench VI, 3 1 showed the"sickle polish" ( Table 2 and Pl.26:a-d) with many s triations and"comet-shaped pits", and many showed numerous t iny holes which are perhaps due t o sand grains hitting t he f lint. This polish was probably due t o contact with cultivated c ereals which had been cut at the base f rom upturned l oose soils. The remaining 7 b lades had widely distributed unstriated polishes which could have been f rom contact with reeds or S tipa ( Pl.26:e,f) However, i n a ll t hese instances t he blades were e ither made o f coarse-grained f lint on which polish distribution c an be quite different ( see Part I , Chapter 1 7) or e lse t hey were .

burnt ( Pl.6:h) or patinated. Both burning and patination c an a ffect the appearance of the polish ( see Part I , Chapter 1 8) However, a ll the blades with polishes other than t he common"sickle polish ' had f eatures which made them s lightly different f rom the rest o f the blades: 2 blades were exceptionally small, 1 blade was e specially thick and had a regularly notched edge, 1 was unusually strongly denticulated, while 1 was exceptionally long, broad and thin, 1 blade had a t ang, and one b ladelet was double-edged. I t i s therefore possible that these blades, a lthough c learly parts of s ickles, may a lso have been used on plants other t han c ereals. No relation could be s een between curved blades or polishes and plant species cut. An unshortened blade which, on account o f i ts thickness, widely distributed polish and random s triations, I had c lassified a s probably hand-held, showed t he com r non"sickle polish"which indicated that i t had been used on c ereals but hand-held. .

1 94

Trench V Twenty-two out o f 4 0 blades showed t he common"sickle polish"with s triations ( Table 2 , P l.26:a-d) which i ndicated t hat cereals had b een cut at t he base f rom upturned s oil. In g eneral, however, polishes were somewhat l ess striated t han t he polishes on t he blades f rom Trench VI. Of t he r emaining 1 8 blades, 8 were burnt ( Pl.6: h) or patinated, 6 had polishes which c ould have been f rom cereals or other plants, 3 had polishes which l ooked l ike experimental reed or S tipa polishes ( Pl.26:e-f) while 1 blade showed wood polish ( Pl. 2 6 : h) ‚

have

All t he lustered b lades f rom both t renches been used on f resh p lants ( cf.Pl.23:e,f).

s eemed

to

" Sickle Blades"Without Gloss Only 2 of the unlustered blades f rom Trench VI c lassified a s"sickle blades"proved to have m icroscopic plant polish on t hem. They were both very coarse-grained. It was i mpossible to a ssign t his polish t o a particular plant a s t he distribution o f the polish governed to a certain extent by the f lint was very i solated and difficult t o a ssess. -

-

Duration of Use Most blades f rom both t renches l ooked well-used and h eavily polished. I cut green barley f or approximately f our hours and r ipe e inkorn f or approximately t hree hours and i n both cases the b lades were quite undamaged and not very polished. I t seems to me that most of t he blades f rom Arjoune must have been used at l east 3 t imes a s long, i .e. f or about 1 2 hours, i f the ancient plants and t he p lants I cut had s imilar properties. Conclusions S everal observations may be made f ollowing i nvestigation o f t he blades with gloss f rom Arjoune:

this

Most of the lustered blades appeared t o have been u sed a s s ickle e lements, while some blades appeared to have been used individually, e ither hafted at one end or e lse hand-held. The l atter had probably been used to cut other plants such a s reeds or S tipa and wood. S tipa i s, however, c ommonly harvested by uprooting rather t han by cutting ( Hillman, pers. c omm.). I could not see any c oncentration o f s ickle b lades o r indeed o f any particular f eature, such a s s ize or polish a ccording to the layout of t he s ite. The

distribution

o f

gloss

and

the

different

o f blades f rom both t renches suggested s traight and curved s ickles may have been t hese may well have been made in l arge

1 95

s izes

and

shapes

t o me t hat both u sed and that and small s izes.

There seemed t o be no r elation between t he s ickle s hape and t he plant species cut. A s no s ickle hafts have been f ound on the s ite i t i s l ikely t hat they were made of wood. Retouch was carried out before u se, possibly, a s experiments suggest, t o prolong t he l ife of t he blades rather t han t o increase s harpness, and the b lades were a lso r etouched during u se, probably t o resharpen t he edges. Retouch r anged from f ine t o coarse denticulation and s eems t o have been random rather than matching any particular p lant. However, t he one instance of wood polish was found on a c oarse denticulate, a s were s ome of t he polishes which looked l ike experimental polishes f rom hardstemmed p lants, such a s r eeds or S tipa. The plants the blades had been used on s eemed t o have been f resh a nd were in my opinion cultivated c ereals, possibly reeds or S tipa and in one c ase wood. However,it must be admitted that I f ound the a llocation o f polishes d ifficult a s p lant polishes seem to form a continuum, and one can easily be confused with another. There may well have been plants which I did not harvest experimentally. Comparison with the botanical evidence therefore s eemed very important. Analysis o f the botanical remains ( see Part I I, Chapter 3 ) showed t hat the staple food crops r epresented a t Arjoune were einkorn, emmer, two-row hulled barley and l entils. The wheats were fully domesticated. Free-threshing wheat s eemed to be present at Arjoune V . but whether i ts absence at Arjoune VI was real or apparent was a ccording to Moffett ( in prep.) uncertain. A s imilar a rtefact o f sampling may have been the relative s carcity o f barley i n Trench VI. The l entils were probably the cultivated species. Evidence for other f ood plants included grape pips in large numbers f rom Trench V , s eeds f rom the horse bean and remains f rom various f ruit t ress, such a s hawthorn, pistacio, plum or cherry, sweet a lmond and f ig. According to Moffett such f ruit was most l ikely collected from t rees growing on the s lopes o f ‚

the

Lebanon

and

Anti-Lebanon.

This l ist matches my evidence fairly well: the microwear analysis pointed t o the harvest of cultivated cereals. In one instance wood polish could indicate that v ines or f ruit t rees had been pruned, a lthough i t might have been any wood which had been cut. Neither l entils nor horse beans are harvested by cutting, as f ar a s I am aware. Reeds and Stipa a re absent f rom the botanical analysis. This could indicate t hat my i dentification was wrong or e lse t hat these plants were not preserved in t he botanical r ecord. C ertainly reeds proliferate nowadays near Arjoune. I could only f ind two unlustered"sickle blades"from Trench VI which showed microscopic polishes which I would a ttribute t o plants. Both had been made o f coarse-grained f lint. This suggested t o me that t he other unlustered b lades, a lthough o f s imilar shape and with edge damage and microscopic polish, had been used on other materials.

1 96

A ll o f t he blades f rom Trench VI and many o f t he blades f rom Trench V appeared t o be worn out. On t he evidence o f experimentation w ith modern cultivated c ereals I would e stimate t hat t hey had been used for about s ix t o twelve hours. As the s ite was not completely excavated and t here was no apparent c oncentration o f s ickle blades, i t would be f utile t o a ttempt t o calculate the l ength o f t he s ickles and t he plot s ize harvested by each. There were a few d ifferences between the b lades o f Trench VI and Trench V : the blades f rom Trench VI were s lightly longer and more uniformely shaped. The"sickle polishes", i .e. probable c ereal polishes, from Trench VI looked to me generally more heavily s triated than those on blades f rom Trench V . This may have been coincidence. Alternatively, it may be t hat the b lades f rom Trench V had been used to cut c ereals higher up t he culms. Could this have been related to t he h igher percentage o f barley recovered f rom Trench V? I t i s unlikely that the l ess striated blades had been used to c ut wild cereals, a lthough on account o f the poor preservation of t he botanical remains the possibility of wild emmer at Arjoune was not ruled out ( Moffett, in prep.). -

-

6 .

Polishes ( P1.27,

on Natufian Blades with Gloss Fig.42)

S ickles have been f ound f rom the Natufian l evels from such s ites as Kebara and E l Wad ( Camps-Fabrer and Courtin, 1 982) in the Mt. Carmel region. At Kebara f ifteen bone s ickle hafts had were excavated, two of them complete, 3 8 cm and 2 8 cm long. At El Wad thirteen s ickle f ragments had were f ound, some of them o f antler. One s ickle haft was more convex than the others. The overall shape, t runcation type, distribution o f gloss and polish and orientation of striations on the bladelets and blades f rom Kebara and E l Wad which I examined a ll indicated to me that the blades had been hafted in s ickles. It has been t he subject of long debate what plants these Natufian s ickles were used on: some s cholars ( e.g. Neuville, 1 934) t hought that they had been used on cultivated cereals, others ( Harlan, unpublished manuscript) thought that they indicated a t rend towards the exploitation o f cereals, while others ( Vita-Finzi and Higgs, 1 970) thought that most o f the t erritory of these s ites in coastal Palestine was not suitable for grain growing ( ibid., p .16) and that therefore the s ickles from these s ites had perhaps been used on reeds and grasses ( ibid., p .22). However, a ccording t o the map published by Vita-Finzi and Higgs ( 1970, p .9, Fig.I), Kebara and El Wad l ie very c lose to the Mt.Carmel s lopes ( 10 and 2 km respectively) and Bar-Yosef ( 1970, p .29) l isted t hese two s ites a s belonging t o a mountainous area. This suggested to me t hat t he s ites were well within reach o f wild cereals. ,

1 97

I examined twenty-five l ustered bladelets i ncluding some b lades f rom the Natufian l evel B and t he mixed l evel A f rom K ebara, which was a ccording t o t he excavators Early Bronze Age t o Recent Arab ( Turville P etre, 1 932), and f rom t he Natufian l evel B2 and t he Late Natufian ( Khiamian) l evel 3 1 at E l Wad ( Garrod and Bate, 1 937, Vol.1, p l.IX a nd p .34, pl. VIII and p .31.) The b ladelets a re part o f the collection o f the I nstitute of Archaeology, London. The l ength o f t he bladelets ranged between 2 .5 and 5 .2 cm, their width between 0 .8 and 2 .2 cm and t heir t hickness between 0 .3 and 0 .6 cm. All t he b ladelets were s traight in profile. Most b ladelets had been t runcated and/or broken a t one or both ends, and had been backed with bifacial ( "Heiwan") retouch. The 3 bladelets f rom the mixed l evel A a t Kebara were i ncluded a s they l ooked identical i n shape and retouch t o t he other Natufian bladelets and blades. Most bladelets were rectangular, with 2 exceptions which were curved. Nine b ladelets had denticulated cutting edges, 2 bladelets had cutting edges with inverse retouch, while t he others were probably not retouched before use. The state of polish preservation varied f rom apparently complete t o poor. In the case of poor preservation t his may be attributed to post-depositional changes, probably natural, and a lso to the f act t hat the blades had been stored unprotected for a l ong t ime. I n some instances the polish distribution was obscured by varnish applied by the excavators. The f lint ranged f rom f ine-grained t o medium-grained.

the

a ll l ike

Examination experimental

of the polishes ( P1.27) and comparison with polishes ( P1.22-25) showed the f ollowing:

The 5 bladelets f rom t he Natufian levels at E l Wad B 2 had unstriated widely distributed polishes which l ooked my experimental reed or S tipa polishes.

Of the 5 Late Natufian bladelets f rom El Wad Bi, 4 showed polishes which looked l ike experimental cereal polishes ( Pl.27: b, d) S triations on 3 of these suggested that l oose soil had got caught between the bladelets and t he p lants during the harvest. This could indicate e ither t hat cereals had been cultivated or that wild cereals had been harvested f rom s cree. The only Late Natufian bladelets which were denticulated showed polish which l ooked l ike experimental reed or S tipa polish. .

The 1 2 Natufian bladelets and blades f rom Kebara B showed a mixture o f polishes: 5 denticulated bladelets ( of which one may have been s imply heavily damaged) showed polishes l ike experimental reed or S tipa polish ( Pl.27:e, g), 6 showed polishes l ike c ereal or perhaps grass polishes ( Pl.27:f, h) 3 o f t hese without s triations, 2 with a f ew s triations and one with many striations, again i ndicating t he presence o f l oose s oil. One of t hese was e ither ‚

1 98

i ntentionally damage.

denticulated

o r

s howed

very

s trong

edge

Of t he 3 b ladelets f rom t he mixed l evel A a t K ebara, 1 b lade s howed c ereal polish with a f ew s triations only and t he 2 b ladelets s howed reed or S tipa polishes, a lthough 1 of t hese could have been unstriated c ereal polish. Conclusions f rom t his investigation had t o be t entative, given t he s mall number o f blades. I could not and d id not s et out t o investigate the problem o f t he beginning o f agriculture. However, r esults o f the microwear analysis did suggest that some of t he Natufian s ickle blades at K ebara a nd some o f the Late Natufian s ickle b lades at E l Wad had been u sed to harvest cereals, which were e ither wild and growing in loose soil, or e lse cultivated. These blades were mostly undenticulated. The results a lso suggested that a ll t he Natufian blades f rom E l Wad B2 had been used to harvest r eed or stipa, a lthough t he l atter i s more l ikely t o have been harvested by pulling ( Hiliman, pers. comm.). There was no published botanical evidence available f rom these s ites and very l ittle botanical evidence f rom the Natufian i n Palestine. At t he neighbouring s ite o f Nahal Oren 3 s eeds o f T . dicoccum ( domesticated wheat) and 1 s eed o f H .spontaneum ( wild barley) had been recovered f rom an undisturbed Kebaran t hat i s, pre-Natufian deposit ( Noy et a l. 1 973, pp.92-93) The f inds f rom t he Natufian a t Nahal Oren were 5 s eeds of Vicia sp 6 s eeds o f Viciae 1 o f Vitis sp 4 of Graminae and 9 other seeds ( ibid.) Because they t hought t hat t his region was not particularly s uited f or agriculture Vita-Finzi and Higgs ( 1970, p .21) t hought that grain was probably gathered not there, but a t upland s ites, l ike Raqefet and Hayonim. But t he s ame c ould be said o f the area a round Tell Aswad where domesticated c ereals had been f ound f rom t he beginning o f the s ite i n the e arly 8 th millennium ( Van Zeist and Bakker -Heeres, 1 979) ‚

‚

.

.

.

‚

‚

Thus the results o f the microwear analysis agree with t he botanical evidence, inasmuch a s domesticated c ereal grains had been f ound in t he v icinity, a lbeit f rom an earlier deposit. However, t he results o f the microwear analysis d id not suppört the opinion o f Vita-Finzi and Higgs based on t heir s ite c atchment analysis. The results were compatible with t he opinion o f Harlan ( unpublished manuscript) who suggested t hat Natufian s ickle blades c ould be s een a s an indicator of an e conomic s hift t owards cereal exploitation and f inally cultivation."Sickle blades" ( more than 1 000) were t he most numerous of the f lint implements f rom Kebara B ( Turville-Petre, 1 932, pp.271-272) and a t E l Wad 6 30"sickle blades"had been f ound i n B 2 and 3 94 had been f ound i n B i ( Garrod and Bate, 1 937, Vol.1, p.34 and 3 1) Therefore a potentially l arge number o f Natufian lustered b lades could be examined, but my comparatively small s ample can not be expected t o i ndicate whether the s triated c ereal polish i s made by cultivated grain or whether i t i s t he ‚

1 99

r esult o f contact with p lants growing in loose soils and t herefore undiagnostic. The examination of a l arge number o f Natufian blades, however, m ight bring u s c loser t o an a nswer t o t he question when cereals were f irst cultivated. 7 .

Polishes ( P1.28,

on

Blades

with Gloss

from Jericho

PPNA

to EB.

F ig.43-44)

I t i s sometimes debated whether a change i n tool type constitutes an adaptation t o d ifferent requirements, s uch a s d ifferent materials to be worked or i s rather a change in f ashion, due to the a rrival o f new people or to t he desire f or novelty among an established people. Another reason f or such a change might be a t echnological advance. Changes in the typology o f Palestinian"sickle b lades"have raised this question. At Jericho ( area F ) ( Crowfoot-Payne, 1 983) blades f rom t he Aceramic Neolithic A ( PPNA): Sultanian ( ibid., pp.649-651) were o f various s izes and often i rregularly shaped. They had mostly unretouched cutting edges, but quite a lot were backed. Lustered blades f rom t he Acera r nic Neolithic B ( PPNB) Tahunian ( ibid., pp.683-686) t ended t o be uniformly l ong and regular. They were mostly unbacked and had i nverse, a lternating or bifacial retouch which gave the b lades a f inely denticulated cutting edge. However, t here were also other forms l ike broken or t runcated blades which r esembled s ickle e lements, plus a variety of retouched and unretouched blades. The

small

lustered

blades

f rom

the

Pottery Neolithic

A :

Yarmukian ( ibid., p p .708-710 ) had coarsely denticulated cutting edges, some of which had been produced by b ifacial retouch, others by pressure f laking on most or a ll of the surfaces o f the blades. S ome e lements with straight ends and f inely denticulated cutting edges had been completely pressure f laked. The Pottery Neolithic B : Ghassulian ( ibid., p .716) b lades had been made with direct retouch and very f ine denticulations, but a s t hey had not been found in a rea F , they were not examined. One of the hallmarks o f the Proto-Urban and Early Bronze Age ( ibid., p .718) i s the Canaanean blade which i s trapezoidal in section and has a f aceted butt. Canaanean blades were e ither unretouched or backed with f ine direct retouch. I examined

6 3

lustered

blades

f rom

area

F ,

i ncluding

1 9

blades f rom t he P PNA, 3 3 blades f rom t he PPNB, 6 blades from the PNA and 5 blades f rom an EB d isturbed deposit which contained Canaanean b lades.

2 00

g iven

The t otal numbers o f by Crowfoot-Payne

P PNB,

1 1

f rom

the

PNA

a nd

excavated blades f rom a s 4 6 f rom t he P PNA, 3 f rom

t he

area F were 1 46 f rom t he

EB.

The b lades had been made f rom a variety o f f lint ranging f rom t he very f ine-grained f lint t ypical o f the P PNB a t Jericho t o t he coarse-grained f lint used f or the Canaanean b lades. I i nvestigated the relation o f f lint t ype to polish i dentification in order t o s ee whether identifications o f worked materials were perhaps i nfluenced by the grain-size o f

the

f lint.

Results The investigation of the PPNA blades showed that polishes were varied: on 4 blades polishes l ooked s imilar to experimental cereal polishes and were mostly unstriated. Only 1 o f these polishes showed the amount of striations a ssociated with the harvest o f cereals f rom loose soil ( Pl.28:e) Polishes on 3 blades looked identical to reed or S tipa polishes ( Pl.28:a), while polishes on a further 7 blades l ooked s imilar to experimental reed or S tipa polishes. Polishes on 5 blades looked s imilar t o experimental bulrush polishes ( Pl.28: 0. Polish on 1 blade l ooked l ike experimental grass polish. The number o f polishes l isted here i s higher than the a ctual number o f blades investigated because a lternative interpretations were made, e .g. one blade might have a polish which looked e ither l ike reed or l ike bulrush polish. Polishes on 2 blades were t oo damaged to be c lassified ( Pl.7: h) Polish on 1 .

.

unlustered blade was not plant polish. present, were a lways oriented parallel o f t he blades. This could indicate that hafted

( see

Crowfoot-Payne,

ibid.,

S triations, where t o t he cutting edges the blades had been

p .651).

Polishes on the lustered blades from the PPNB ( ibid., pp.683-686) were less varied: 2 5 o f t he 3 3 blades had polishes which looked exactly l ike the experimental c ereal polishes ( Pl.28: b, d,f). All were somewhat striated and 1 6 of these were heavily striated, 4 blades having polishes which looked s imilar t o experimental c ereal polishes. Polishes on 3 b lades looked l ike experimental reed or S tipa polishes, while polishes on 3 blades looked s imilar t o experimental bulrush polishes. Polish on 1 blade looked l ike experimental grass polish. On a f ew blades the identification was uncertain because of post-depositional damage and 1 unlustered blade d id not have microscopic p lant polish. All the l ong blades with f inely denticulated cutting edges ( Crowfoot Payne's types 2A and B , ibid.) produced by inverse, a lternating and bifacial retouch had"sickle"polish which was s lightly t o heavily striated ( Pl.28:b, d,f). I t i s possible that the a lternating retouch had been carried out in order t o straighten the cutting edge, while perhaps the bifacial

retouch

represented

a

2 01

s econd

sharpening

o f

the

b lades. A ll were heavily l ustered and had probably been u sed f or a l ong t ime. S triations a lways ran uniformly parallel t o the cutting edge, supporting t he argument t hat t he b lades had been hafted ( Crowfoot P ayne, ibid., p .686) The i nvestigation o f a f ew blades from t he PNA ( ibid., pp.708-710) showed t hat t he pressure f laked s traight-ended b lades which were probably s ickle e lements had polishes which s eemed t o be deteriorating. Judging f rom t he f ew complete patches of polish, t he d istribution looked l ike t hat o f experimental cereal polish. The coarsely denticulated b lades had polishes which l ooked l ike experimental reed or S tipa polishes ( Pl.28:g) a lthough a ‚

f ew striations were present. I t would s eem t hat t hese b lades had probably been used t o cut plants w ith t hick and hard s tems, a s the relatively thin s tems o f c ereals would have been caught i n t he coarse denticulations a nd s imply pulled out. The gloss on t hese blades was t oo s trong f or t he polish t o be c lassified a s wood polish. I n general, polish identification was made difficult by the f act t hat the b lades had been pressure f laked which meant t hat I c ould not a ssess the polish distribution. S triations ran uniformly parallel t o the cutting edges o f the b lades. I c ould not t ell whether the l arge denticulate had been hafted, but t he pressure f laked s traight-ended blades were probably used a s s ickle e lements a s they were t oo small to be held by hand. The Canaanean blades f rom t he EB ( ibid., p .718) had s triated c ereal polishes ( Pl.28:h) S triations which ran uniformly parallel to the cutting edges and t he presence o f l ittle or no backing r etouch suggested t o me t hat t hey had been used hafted. .

The comparison of t he f lint i dentification showed t hat there was t he two.

no

types with polish relationship between

Conclusions There was a variety of polishes on the b lades f rom t he P PNA which were s imilar to experimental polishes f rom cereals cut with or without t he presence of l oose soil, grass, reeds and/or S tipa and bulrushes. ( S tipa i s most commonly harvested by pulling ( Hillman, pers. comm.). I n some cases the identification was doubtful because o f polish disintegration, other post-depositional damage or because polishes d id not f it into distinct categories. This could mean that plants harvested with t hese blades had not been harvested experimentally or t hat p lants or conditions were s lightly different t o t hose t oday. A lternatively t his could mean that s everal plant species had been harvested with t he s ame blades. I t i s o f interest, however, t hat t he s triated c ereal polish suggesting t he presence o f l oose s oil during harvest, and t herefore harvest o f cultivated cereals, was rarely f ound i n t he P PNA, while i t was t he most common polish

i n

t he

PPNB

and

i n

t he

2 02

l ater

periods.

This

change

f rom mainly unstriated polishes i n t he P PNA t o t he s triated polishes i n t he P PNB, could mean s everal t hings, f or example, t hat i n t he PPNA most of t he cereals cut were wild, or e lse t hat cultivated c ereals were collected by hand or cut high up t he cuim. I t might a lso mean t hat i n the P PNA c ereals may have been planted with t he help o f digging s ticks, t hus l eaving the ground f airly i ntact, while in t he P PNB t he s oil was completely t urned over ( digging stick weights were apparently not found before t he PPNB M orrell, 1 983, p .489) A t hird and very l ikely possibility i s that t his reflects t he rarity o f domesticated cereal grains ( einkorn, emmer a nd barley) f ound in the P PNA l evels at Jericho ( Hopf, 1 983, p .609) particularly a s independent evidence, such a s the proportional decrease o f pestles and increase o f stone vessels f rom t he P PNA t o the PPNB l ed Dorrell ( op. c it., p .527) to come to s imilar c onclusions. He t hought t hat t he stone equipment of the P PNA suggested t he processing o f husked or brittle rachis cereals, while that o f the PPNB suggested the processing of naked-grain cereals .

,

and not

tough rachis varieties. contradict t he idea that

The t he

microwear investigation blades had been hafted.

d id

In the PPNB the majority o f polishes looked l ike heavily striated experimental cereal polishes. Such polishes were found on each o f the f inely denticulated s ickle blades. I n some instances unretouched PPNB blades a lso showed t he s triated c ereal polish. As a ll the f inely denticulated b lades were very lustered i t was a ssumed t hat t he retouch was carried out i n order to prolong t he l ife The microwear investigation supported the t he P PNB blades had been used hafted. Some s uggested to me that a variety of p lants, S tipa or bulrushes, had been cut in the PPNB.

o f the blades. a ssumption t hat other polishes possibly reeds, Evidence f or

t he use o f such plants was found through the basketry and matting recovered f rom both the PPNA and the P PNB ( Crowfoot, 1 982). The botanical evidence ( Hopf, op.cit., p .609) showed a great increase o f f inds o f domesticated cereal grains f rom t he PPNA t o the P PNB. In the Pottery Neolithic A the smaller pressure s traight-ended b lades had probably been u sed a s

f laked s ickle

e lements t o cut c ereals, while the coarsely denticulated blades, on a ccount o f ' the distribution o f their polishes and f rom my own experimental evidence, had probably been used t o cut reeds or S tipa. S imilar blades had been excavated f rom P ottery Neolithic s ites a long the bank o f the River Yarmuk. The buildings there were huts and houses, the roofing o f the f ormer probably made of"branches, l eaves and s traw"(Stekeljs, 1 973, p .39); i t i s therefore quite possible t hat such coarsely denticulated blades had been used t o cut material l ike r eeds. The Yarmukian s ites were, however, a lso agricultural a nd one would have t o examine the s ickle b lades in order t o t o s ee whether they were u sed on cultivated c ereals or not. Interestingly, 1 T opf ( op.cit., p .578) pointed t o t he small number o f s eeds and the complete absence o f charcoal f rom the PN at Jericho which she t hought

2 03

supported r K enyon's conclusions that o f s econdary importance i n the PN.

plant

growing

had

been

The Canaanean blades which derive f rom a period when, according t o Kenyon ( Hopf, op.cit.) people were f armers again, were probably used to cut cultivated c ereals, j udging f rom

their

s triated

polishes.

The results suggested that typological d ifferences c an occur a ccording t o the materials cut: examination o f t he i rregular blades of the PPNA suggested a s tage o f experimentation during which many plants were cut; these plants i ncluded cereals, though these may not a lways have been cultivated. Examination o f t he PPNB blades s uggested an increase in s tandardisation and specialisation with f ine denticulation. I t seems that a ll t he f inely denticulated blades had been used to harvest cultivated cereals. They constituted over 75 o f the t otal o f the lustered blades f rom the PPNB ( Crowfoot Payne, 1 983, p .683) In t he PNA l arge denticulates were probably used to cut reeds or S tipa while pressure f laked standardised b lades were most l ikely used t o cut domesticated cereals. Canaanean blades f rom the Proto-Urban t o the EB were probably used to harvest .

domesticated contradict the 8 .

cereals. evidence

The botanical evidence did not f rom the microwear investigation.

Conclusions

The f rom the Natufian t o the

comparison of wear-traces on the lustered blades 5 th millennium s ite at Arjoune, with those f rom t he l evels at Kebara and E l Wad and those from the PPNA EB levels at Jericho l ed me t o the f ollowing

conclusions: The identification of plant polishes using high-power optical microscopy was by no means a lways certain and o ften the polishes formed a continuum. This indicated t o me t hat the s ample o f blades should a lways be large when investigating a spects of plant husbandry. The results of the microwear analysis did not contradict the botanical and other independent evidence, which i n order to avoid bias was s tudied a fter the microwear analysis had been carried out, and s ites f rom which domesticated cereal grains had been recovered a lso yielded blades with s triated cereal polishes. The increase in domesticated cereal grains f rom the P PNA to the PPNB a t Jericho and the change in proportions o f stone vessels and pestles was matched by a considerable i ncrease in blades with striated cereal polishes. From this evidence I conclude that the microscopic examination of polishes on lustered blades i s a viable approach to t he investigation o f a spects o f plant husbandry. Nevertheless, g iven t he different explanations that are possible f or some f eatures of wear-traces, i t i s desirable to s tudy t he wear-traces a t the s ame t ime a s the botanical evidence.

2 04

I t was very i nteresting t o note t hat a f ew o f t he b ladelets f rom t he Natufian l evels at Kebara and E l Wad had s triated c ereal polishes which t ally with t he harvest o f c ereals f rom upturned s oil. I t may be that both the presence and absence o f s triations i n p lant polishes was due to f actors other t han t he c ultivation t o me that many more period before a ny

or the gathering of plants and i t s eems blades need to be examined f rom t his statements concerning t he beginning o f

agriculture can be made. But i f a l arge number o f Natufian b lades with striated polishes i s discovered, then t his would perhaps i ndicate e arly s oil-tillage and t hus cultivation o f c ereals. I n other words, striations in plant polishes could provide crucial evidence f or early agriculture. The i nvestigation o f t ypological f eatures d ifferent plants, and

plant polishes i ndicated that some were correlated with t he harvesting o f the f inely denticulated blades f rom

t he PPNB a lways showed the striated polish which suggests t he harvest o f cultivated cereals. Coarsely denticulated b lades at the same s ite, both f rom the P PNA and f rom the PNA s eemed to have been used t o cut hard-stemmed plants such a s reeds or Stipa. The evidence in both these cases corresponded to t he knowledge gained f rom my experimental work, namely that coarsely denticulated blades were most e fficient in cutting reeds, while f inely denticulated blades were most e fficient in cutting cereals. At Arjoune, however, denticulated blades showed cereal as well a s some other plant

polishes,

which

l ooked

l ike

experimental

reed,

Stipa,

grass and wood polishes. ( Although S tipa i s most easily harvested by pulling ( Hillman, pers. comm.). This suggested t hat typological f eatures need not necessarily correspond t o d ifferent worked p lant species. Most of picked

the unlustered blades out a s"sickle blades"

f rom a ll s ites which I on account of their shape,

had had

apparently not been used on plants, but on other unidentified materials. However, judging by the microscopic polish on them, some unlustered blades made of coarse-grained f lint had been used to cut plants. I t follows that microscopic examination would help to d ifferentiate unlustered s ickle blades f rom blades used on other materials.

2 05

Chapter

1 8

-

Burnt

Flint

Implements

from Arjoune

Copeland ( 1981, i n prep.) had noticed a number o f burnt i mplements a t Arjoune. My own examination of t he t ools r evealed that s ome implements s howed t races o f heavy burning i n t he f orm o f microscopic c racks and/or b lackening o f t he f lint, such a s s een on experimental p ieces l eft in a f ire overnight ( see Part I , Chapter 1 9, P l.6:g,h) These t ools ( from a ll t renches) were invariably very small implements which had probably been hafted, s uch as a rrowheads, s ickle .

blades, unlustered bladelets, and drills. A subsequent examination o f a ll t he f lint f rom Trench VI, t ools, f ragments, debitage and"rubbish", r evealed t hat, apart f rom such tools a s described above, only small f ragments and p ieces o f"rubbish"were burnt. The f act t hat only t he smallest, probably hafted t ools were burnt, suggested t o me that t he tools may have been burnt when they were t aken out o f their hafts and dropped i nto a f ire, which i s necessary t o melt the hafting agent. Another possibility might be t hat these

t ools

were

A horizontal

used

near

hearths

distribution

( Newcomer,

plot

o f

t he

pers.com l n.). burnt

f lint

p ieces from Trench VI ( Fig.9) revealed no s ignificant c oncentration. Unfortunately the f lint had not been plotted precisely during excavation. Also, experiments by Bergman had shown that f lint can be heated a t 3 50 degrees c entigrade without any visible alterations, except greasy looking f lake s cars from f laking a fter burning. I t i s therefore possible that concentrations of burnt f lint did exist but were overlooked.

2 06

Chapter

1 9

-

Comparison of

Typology

and Microwear Analysis

Comparison o f t he results o f t he t ypological a nd the microwear analysis of t he f lint t ools was not a lways possible. O ften t he tools which had been drawn by Copeland were not suitable for microwear analysis ( see P art I I, Chapter 9 ) and t hose which were suitable f or microwear analysis were o ften not drawn. This meant, f or example, that only 1 notch, 1 denticulate and 3 burins could be compared. Only t he c learest types of t ools were selected by me f or microwear analysis,and probably had been s elected by Copeland f or her drawings. I t i s a lso t rue t hat I based some o f my conclusions such a s those concerning hafting on macroscopic observations, that i s, on t he s ame k ind a s had been used by Copeland. ,

I t i s therefore not surprising that no major discrepancies between t ypology and microwear analysis were f ound. Only 2 divergences occurred in t he s craper analysis ( of 7 s crapers compared) : a " thumbnail s craper" ( Copeland) had probably not been used a s a s craper, while a "denticulate racloir"was probably an end-scraper. O f the perforators ( 7 compared) a "borer" had probably been used a s a drill. One"denticulate on a cortex-flake"may have been a core. Out of 3 compared axes/adzes one had apparently been used a s an axe, but i t might have been used a s an adze t o begin with. The major divergence was that some but not a ll of the unlustered blades, c lassified by Copeland a s"sickle-blades" on account of t heir shape, had been used on materials other than plants ( see Part I I, Chapter 1 7) -

2 07

-

Conclusions l .A

Flint Tools

-

at Arjoune

My research l ed me t o reach the following conclusions about t he f lint t ools a t the s ite of Arjoune ( Part I T, Chapters 1 -6) The raw materials ( Part I I, Chapter 7 ) o f which most o f t he t ools had been made a t prehistoric Arjoune were found one hour's walk from the s ite in a gravel t errace on t he banks o f the Wadi ' Rabiy' ah. The material consisted mainly o f small pebbles and was often o f poor quality. S imilar, probably f rost shattered cores have been excavated f rom s everal t renches. The material t o make the f ew l arge chipped stone tools f rom Arjoune, s uch a s t abular s crapers, had probably been collected directly from the chalk and l imestone o f the s lopes o f the L ebanon whence the gravels originated. The preservation of t he ancient f lint implements ( Part I I, Chapter 8 ) was found to be variable. Most edges looked in"r nint fresh" condition. However when examined under the microscope a ll t he implements showed post-depositional surface modifications to a varying extent, and many were of coarse-grained f lint or chert. These f actors made a microwear analysis difficult. Another l imiting f actor was the great quantity of experiments which had to be carried out for each type of implement ( see Part I , Chapters 1 0-17) Therefore only relatively f ew ( 470) undamaged tools, made o f the f iner-grained stone, were selected according to t ool type ( Part I I, Chapter 9 ) I could only come to an opinion about t he use of some 1 80 of the tools from Arjoune. .

The showed

microscopic

evidence

Acutely

of

angled

study

the

at

5 0x

t o

following

uses:

end-scrapers

( Part

2 00x

I I,

o f

Chapter

t he

1 0)

tools

with

an"overhang"of the working edge, had apparently been used t o s crape hides, while steeper angled, o ften tabular scrapers, had been u sed on wood. Large s crapers s eemed to have been used on both these materials, but not on s tone or bone. The l atter a s well a s the former materials had apparently been s craped with smaller scrapers. Many of the scrapers had thinned butts suggesting t hat they had been inserted in a haft; experiments had shown t hat this increased the e fficiency o f the tools, a s well a s the ease of working. Perforators

( Part

I I,

Chapter

1 1)

seemed

t o

have

been

used a s borers ( of wood and pottery) and a s piercers ( of hide). There was c lear evidence o f fast, i .e. mechanical drilling, possibly with a bow-drill. A long thin meche dc f oret had been used to drill wood, while a drill had been used t o perforate s tone, perhaps such an object a s the l arge macehead recovered f rom t he s ite.

2 08

Retouched b lades a nd f lakes ( Part I I, Chapter 1 2) o f various s hapes a nd s izes appeared t o have been u sed a s wood-whittling knives, a s wood-saws, and a s cutters o f various soft ( perhaps hide or meat) or hard ( perhaps bone, shell and s tone) materials. Unfortunately the precise material ( with the exception of wood) could not be i dentified on t his t ype o f t ool. Judging f rom the backing and the t runcations o f many b lades, such a s the wood-saws, a nd the blades which had been u sed on soft materials, they had been used i n hafts, rather l ike the s ickle-blades ( see below) A f ew very small bladelets could possibly have been u sed as barbs. I n one instance t he t runcation on a blade had apparently been u sed to s crape wood. .

Each notch ( Part I I, Chapter 1 3A) s eemed t o have been used in a different way. Some notches appeared t o have been used to shave wood, another to shave bone. On s everal the projections appeared t o have been used a s perforators. Small s ingle-blow notches made f rom t he dorsal a spect appeared to be accidental. Denticulates ( Part I I, Chapter 1 3B) showed a great variety o f wear-traces. They appeared to have been used as s crapers ( of wood and hide) and a s individual notches ( to shave wood) In some instances the projections only appeared t o have been used t o perforate ( wood) or i ncise ( stone) .

Burins ( Part I I, Chapter 1 4) had apparently been used t o incise hard materials ( shell, l imestone or bone). One burin had been used to split p lant material, perhaps reeds. The Chapter

been

f ew 15)

excavated

seemed

Axes

and

used

to

to

have

choppers

t ransverse been

( Part

I I,

arrowheads

( Part

I I,

shot. Chapter

1 6)

had

apparently

chop wood.

Lustered " sickle blades" ( Part I I, Chapter 1 7) had indeed been used to harvest plants, apparently mostly cultivated cereals. Experimentation and comparison with wear-traces on 8 8 s ickle-blades f rom Epi-Paleolithic Kebara and El Wad and f rom PPNA t o EB Jericho, a s well a s with botanical evidence, suggested that certain wear-traces were i ndicative of soil-tillage. A s ignificant increase of such wear-traces ( in the form o f a great quantity o f s triations) was found from t he PPNA to the PPNB at Jericho. A f ew blades with a considerable number of striations were f ound f rom the Epi-Paleolithic s ites. At Arjoune most s ickle-blades had been t runcated and inserted end-to-end in s traight and/or curvilinear hafts. Comparison with experimental results showed that the e fficiency o f my"models"of the Arjoune s ickles was about three-quarters o f the e fficiency o f a modern metal-sickle. A f ew b lades, some of which had probably been hafted l ike knives, appeared t o have been used to cut non-cereal plants, such a s reeds.

2 09

Unlustered blades, c lassified a s"sickle b lades"on a ccount of t heir shape, f ell i nto t wo c ategories: t hose made of coarse f lint or chert o ften r evealed microscopic plant polish and had been used a s s ickle-blades. The macroscopic lustre s imply did not show because o f t he coarseness o f t he f lint. Unlustered blades made o f f ine-grained f lint o ften revealed d ifferent polishes ( Part I I, Chapter 1 2) a nd s eemed to have been used t o cut wood or o ther materials ( see above) The s tudy of the ancient t ools confirmed t hat good quality raw material had been at a premium a t Arjoune: tools had been reused, e . g. s crapers a s saws and choppers or adzes, a s ickle-blade a s a s craper. There were s everal composite t ools, and many t ools had been used or r esharpened to"exhaustion". This l ack o f suitable raw material a lso seemed t o be responsible f or the often unorthodox s hapes o f the tools. Quite a f ew of the tools were o f a greyish colour and revealed microscopic cracks consistent with burning ( Part I I, Chapter 1 8) Such t races were seen on experimental f lint pieces which had been burnt in a woodfire f or s ome hours. I t was interesting t o note t hat the only obviously burnt t ools were the smallest tools which had probably been used hafted: a number o f s ickle-blades, b ladelets, drills and arrowheads were burnt, a s well a s a f ew fragments and pieces which I found when I examined a ll the f lint pieces excavated f rom Trench VI for s igns of burning. This suggested t hat the tools may have been burnt when t hey had been t aken out of their hafts. .

Comparison of the typology and the microwear analysis ( Part I I, Chapter 1 9) revealed no major discrepancies. This was perhaps not surprising, a s only t he c learest types o f tools had been selected by me f or the microwear analysis, and presumably by Copeland, t he typologist, f or her drawings. Also I had based some o f my conclusions on macroscopic observations, i .e. on t he same criteria a s t he typologist. The only"surprise"was t hat some but not a ll of the unlustered backed t runcated blades classified by Copeland a s"sickle-blades"had been used on materials other than plant. -

1 .B

been

-

The

S ite

-

of Arjoune

I t s eemed c lear to me t hat the s ite o f Arjoune chosen f or a ccess to good quality f lint.

had

not

Many o f the examined t ools had been used t o exhaustion, however many had not. This f act, t ogether w ith the s carcity of good f lint, suggested t hat the s ite, a s eries of p its sunk i nto bedrock, was not a s eries o f rubbish pits. Burning o f the f lint implements s eemed t o be due rehafting rather t han t o a general conflagration o f

2 10

t o t he

s ite. The horizontal d istribution p lot o f burnt f lint f rom Trench VI d id n ot reveal any s ignificant concentrations. However, i f t he material was i n s itu, t hen i t i s l ikely t hat a r etooling a rea had been i n c lose proximity. The l arge suggested t hat

amount of f lint t ools

From t he evidence some instances t he

in

debitage had been

and made

t he reject a t the s ite.

o f t he wear-traces on the objects probably made with

c ores

tools and the t ools,

a s well a s the botanical and t he f aunal remains, t he following t asks could be reconstituted: harvesting domesticated crops and non-crop p lants, drilling wood and stone, engraving s tone, bone and shell, s craping wood, hide, stone and bone, chopping wood, perhaps f elling t rees, and shooting, presumably game. 2 .

-

The Benefits

of

High-Power Microwear Analysis

I t will be c lear that one purpose o f adopting t his method, namely t o define possible activity a reas through a distribution analysis o f implements a ccording t o worked materials, was not achieved. This was f or the following reasons:

-

most

tools

did not

have

any definite

wear-traces.

I t has been demonstrated ( Part I , Chapter 2 0) that on some tools, e . g. cutting implements used on materials other t han plant or wood, wear-traces could not be identified with certainty. This could a lso not be done on tools which had not been used for s ome t ime, nor on t ools o f coarse f lint which had been u sed for some t ime ( Part I , Chapter 1 7) Polishes f rom the s ame materials could l ook different s ince many variables a ffected polishes. On the other hand, polishes f rom different materials could look indistinguishable ( Part I , Chapter 1 4) -

P ost-depositional surface modifications were a ll tools to a varying degree ( Part I I, Chapter

present 8 )

on

Mo meaningful s ampling process could be devised. Even sampling had been carried out according t o edge damage, would have excluded tools u sed on soft materials.

i f one

-

-

A high-power microwear analysis o f a ll t he f lint tools and debitage f rom s everal squares would have t aken an enormous amount of t ime, even i f wear-traces could be identified on a ll t he implements. -

It could be a rgued that the results o f t his analysis were o ften negative because I had had years i n which to gain experience and t o complete and/or

perhaps

t hat

the

f lint

2 11

f rom

t he

s ite

was

microwear only t hree my work, unsuitable,

e ither because o f i ts quality or because o f post-depositional surface modifications. However, I would argue t hat t he blind t ests were c onducted a t the end o f t he three years a nd t hat t he results were not markedly d ifferent f rom t hose o f other r esearchers w ith longer experience. The f lint f rom Arjoune, a lthough o ften coarse-grained, d id develop wear-traces a s rapidly a s English or French f lint or chert o f c orresponding grain-size, and l ooked i n better condition than f lint f rom many other s ites which I examined. The s ite o f Arjoune, in prehistoric t erms very recent, d id not undergo t he enormous c limatic and environmental changes found in t hose s ites which existed during or before the l ast glaciation. I t herefore f eel t hat my results are relevant t o high-power microwear analysis i n general. However valuable supplementary information about t he s ite was derived f rom t he h igh-power microwear analysis and t his method could therefore be useful for this purpose. I t a lso s eemed a very good method of investigating individual tool types, e specially when the used edge and a ction o f a tool could be identified with some certainty, and when use-wear polishes were strongly developed. Comprehensive experimentation in which many variables were t aken into account did s eem to produce results which matched other l ines of evidence. This was e specially t he case with t he analysis o f drills ( Part I , Chapter 2 0) and of s ickle blades ( Part I I, Chapter 1 7) Furthermore, i t appeared that through examination o f certain wear-traces ( striations) on a ssemblages of Epi-Paleolithic s ickle-blades a major archaeological problem, namely t he beginning o f plant cultivation, could be solved ( Part I I, Chapter 1 7.10) .

In general,

therefore,

high-power

microwear

analysis

can be o f considerable value to t he archaeologist, provided i ts present l imits are c learly recognised and i t i s not taken t o be the panacea for a ll problems concerning prehistoric s ites. I f eel that the major contribution of this thesis was to call a ttention to such l imits ( see Conclusions: Theory and Method o f Microwear) : polishes do not mysteriously conform to precise worked materials, and a great number o f variables i s involved in polish f ormation. Other l imiting f actors are that polishes are a ffected not only by natural agencies but equally by c leaning with chemicals, and that residues supposedly f rom worked materials appeared t o be constituents of the f lint i tself. I therefore f eel that t his thesis has t aken microwear analysis a step t owards a realistic a ssessment o f what it can and cannot do. The results of my work demonstrate that microwear analysis requires a great deal more painstaken and scientifically controlled research and t hat i t needs to be subjected t o f ar more stringent controls, be i t blind t ests, be i t ' objective" comparison ( by means o f computer s cans ( Grace et a l., 1 986)) of polishes. The results o f a lso i ndicate t hat f ar more documentation t han has rule i s required i n publications dealing with this and

t hat

a

s ite

distribution

2 12

a nalysis

i nvolving

my work been t he subject, a l arge

n umber o f t ools i s s imply not practicable a t t his s tage. Equally i t appears t hat, a s with other specialists such a s a rchaeobotanists, a great deal o f t raining i s required before a microwear analyst can make a valuable contribution t o archaeology. I hope i t i s a lso c lear t hat such a c ontribution i s o f value, and with f uture developments, possibly, o f great value.

2 13

Bibliography Ahier, S . A. ( 1971) Projectile Point Rodgers Shelter, Missouri Missouri ‚

Research Series, 8 , University of Archaeological Society, Columbia.

Anderson,

P .C.

diagnostic Archaeology

( 1980)

A

Missouri

testimony

residues on stone 1 2 ( 2) : pp.181-194.

Form and Function at Archaeological S ociety,

tool

of

and

the

Missouri

prehistoric working

tasks:

edges.

World

Anderson-Gerfaud, P .C. ( 1981) Contribution M4thodoloqigue a 1 ' Analyse des Microtraces d ' Utilisation sur les Outils Pr historiques. Th se present ee a l ' Universit4 de Bordeaux I , 3e cycle, No. 1607.

Anderson-Gerfaud, P . C. ( 1982) Comment pr4ciser l ' utilisation agricole des outils pr historiques? Cahiers de l ' Euphrate 3 pp.149-164. Sophia Antipolis: C . N.R.S.

Anderson-Gerfaud, P .C. ( 1983) A consideration of the uses of certain backed and ' lustered' stone tools from late Mesolithic and Natufian levels of Abu Hureyra and Mureybet ( Syria.) In ( Cauvin, M . -C., ed.), Traces d ' Utilisation sur les Outils N olithiques du Proche Orient : pp.77-106. Lyon: Travaux de la Maison de l ' Orient 5 .

Anderson, H . H. patination on and Methods in Holland,

and Whitlow, H . J. ( 1983) Wear traces and Danish f lint artefacts. Nuclear Instruments Physics Research 2 18 : pp.468-474. North

Amsterdam.

Audouze, F ., Cahen, D ., Keeley, L . H., Schmider, B . ( 1981) Le site Magdale 'nien du Buisson Car npin Verberie ( Oise) Gallia Prehistoire 2 4(1) : pp.99-143.

Bar-Yosef, Palestine.

0 . The

Bar-Yosef, 0 . the Southern eds.)

( 1970) Hebrew

The Epi-Paleolithic University: Jerusalem.

( 1981) The " Prepottery Levant. In ( Cauvin,

Pr histoire

du

Levant

Neolithic" period J ., and Sanlaville

: pp.555-569.

Paris:

Barton, R . N. E. and Bergman, C . A. ( 1982) Hengistbury: some evidence from experimental World

Archaeology

1 4(2)

: pp.237-248.

2 14

Cultures

of

in P .,

C . N.R.S.

Hunters at archaeology.

Bergman,

C . A.

breakage: Archaeology

Bergman,

and

C . A.,

( 1983) La Pal olithique

A .

M . H.

Barton,

R . N. E.,

( 1983) Akil.

8 7(3)

( 1977)

An

Flint

arrowhead

Journal

Collcutt,

S . N.

fracture volontaire dans Sup &ieur tardif du Sud

L ' Anthropologie

Betts,

Newcomer,

examples from Ksar 10 : pp.238-243.

of

and

une de

Field

Morris,

G .

industrie du l ' Angleterre.

: pp.323-337.

Experiment

to

Examine

Some

Effects

of

Agricultural Implements on Flints: Its Results and Implications. Unpublished B .Sc. dissertation, London University, Institute of Archaeology.

Beyries, S . fonction.

and Paper

Inizan, M .-L. ( 1982) Typologie, ochre, presented at ' Recent Progress in Microwear

Studies', Tervuren. pp.313-322.

Bradley, R . and microstructure Paper presented

Studia

Clayton, on the in 1983.

M .B., eds.) The Human The University Press.

Brink, J . Formation 8 3,

2

C .J. ( 1986) The influence of f lint formation of r nicrowear polishes. In ( Sieveking, G . de G . and Hart, Cambridge:

( 1978) An Experimental Study of on Endscrapers. Archaeological Survey

Microwear of Canada,

F . L.

animal

Belgica

Chert,

National

Briuer, and

Praehistorica

Museum

of

( 1976)

Man:

New

residues.

Uses

of

Flint

Mercury

clues

to

American

and

Series,

stone

tool

Antiguity

Broadbent, N . D. and Knutsson, analysis of quartz scrapers. Fornv nnen 70 : pp.113-128.

Ottawa.

function:

41(4)

K . ( 1975) An Results and

plant

: pp.478-484.

experimental applications.

Broderick, M . ( 1979) Ascending paper chromatographic technique in archaeology. In ( Hayden, B . ed.) Lithic Use -Wear Analysis : pp.375-383. Academic Press: New York.

Broudy,

E .

( 1979)

The

Handloom Nostrand

from Ancient Reinhold.

Bueller,

H .

( 1983)

analysis

of

tools

Book

of

Times

Looms:

to

the

Methodological s elected

from

2 15

A

Present.

problems

the

History

in

Natufian

of

New York:

the sites

the Van

microwear of

El

Wad

a nd Ai r i Mall a l la . I n ( Cauvin, M .-C. ed.) Traces d 'T lf f lisation sur ics Out i ls N € o1ithiques du Proche Orient pp.107-126. Lyon: Travaux dc l a Maison dc 1 ' Orient 5 . ‚

Cahen, D . and Gyseis J . ( 1983) Techniques et fonctions dans i 'industrie l ithique du groupe de Blicquy ( Belgique). In ( Cauvin,

M .-C.,

N oiithiques l a

Maison

Cahen, tools,

de

ed.)

du

Traces

Proche

l ' Orient

Cahen, D . more than

2 0(4)

sur

: pp.37-52.

Lyon:

les

Outils

Travaux

Van Noten, behaviour in

F . L. ( 1979) Stone prehistory. Current

: pp.661-683.

and Keeley, L . H. ( 1980) Not less than two, three. World Archaeoloqy 1 2(2) : pp.166-180.

Camps-Fabrer,

H .

de

5 .

D . ‚ Keeley, L . H. and tool kits and human

Anthropoloqy

d 'I Jtilisation

Orient

and

Courtin,

J .

( 1982)

Essaie

not

d 'approche

technologique des faucilles pr historiques dans le bassin M diterran en. Travaux du Laboratoire d ' Anthropoloqie, de Pr t histoire et d ' Ethnolo ie des Pays de la Méditerran e Occidentale, Etude 8 : pp.1-26. Universit' de Provence.

Cauvin, J . ( 1968) Les outillages n olithiques du l ittoral libanais. Fouilles de Byblos

de IV.

Byblos Paris:

et A .

Maisonneuve.

Cauvin, M .-C. ( 1973) Probl mes d ' em r nanchement des du Proche-Orient: les documents de Tell Assouad Syrie) Pal orient 1 : pp.103-106.

faucilles ( Djezireh,

.

Cauvin,

M .-C.

Proche-Orient, Pal orient

9 /1

Clapham, A . R., Flora of the

( 1983)

Les

donn es

faucilles

morphologiques

pr historiques et

du

fonctionelles.

: pp.63-79.

Tutin, T . G. and British Isles.

Warburg, E . T. ( 1962) The Cambridge: The University

Press.

Clark, D ., Phillips, J .L. and Staley, P .S. ( 1974) Interpretation of prehistoric technology from ancient Egypt and other sources. Part I : Ancient Egyptian bows and arrows and

t heir

relevance

for

African

pp.323-388.

2 1

prehistory.

Pal'orient

2 (2)

Copeland, P . J.,

L .

( 1981)

Arjoune

prehistoric pp.1-27.

Copeland,

site

L .

in:

Marfoe,

1978: in

( in

the

prep.)

L .,

Hor ns

in:

grattoirs

E .

et

( 1983)

Parr,

Analyse

herminettes

L .

Parr,

P . J.

et

Site

in

Mureybet,

of

Levant

al.

British

d ' une

Syrie

a

13

Arjoune:

‚

Syria.

trac ologique

de

and

investigation

Basin/Syria.

Excavations at a Prehistoric Archaeological Reports: Oxford.

Coqueugniot,

Copeland,

preliminary

serie

de

( 9 r ne-7 me

r nil14naires). In ( Cauvin, M .-C., ed.) Traces les Outils N olithiques du Proche Orient Lyons: Travaux de la Maison de l ' Orient 5 .

d ' Utilisation : pp.163-172.

Crabtree,

Flintworking.

D .

( 1972)

An

Introduction

Occasional papers at the No.23, Pocatello, Idaho.

Crawford, O . G.S. threshing-machine.

Idaho

to

State

University

( 1935) Notes and News Antiquity 9 : pp.335-339.

-

A

Museum,

primitive

Crowfoot, E . ( 1982) Textiles, Matting, and Basketry. In: Kenyon, K . M. and Holland, T . A., Excavations at Jericho Vol.4 pp. 546-550. Oxford: The University Press. ,

Crowfoot

Payne,

industry

from

Crowfoot

Payne,

J .

Abu

( 1980)

An

Salabikh.

J .

( 1983)

Early

Iraq 42

The

flint

Dynastic

III

industries

of

In: Kenyon, K . M. and Holland, Vol-5 : pp.622-759. Oxford: The

T . A., Excavations University Press.

Curwen, E .C . ( 1930) pp.179-186.

flint

Prehistoric

Curwen, E .C. ( 1935) Agriculture and Palestine. Antiquity 9 : pp.62-66.

Curwen, Sussex.

E .C.

( 1937)

Antiquity

Notes 1 1

and

f lint

: pp.105-120.

News:

sickles.

the

f lint

Tribulum

Jericho. at

Jericho

Antiquity

sickle

flint

4

in

from

: pp.93-94.

Dalman, G . ( 1928-1942) Arbeit und Sitte in Bertelsmann: Guetersloh. Reprinted 1 964, Hildesheim. 7 vols.

2 17

Palaestina. Georg Ohms:

Del

Bene,

of

striation

Lithic

T . A.

( 1979) and

Use -Wear

Once

upon

polish

Analysis

a striations:

formation.

I n

: pp.167-177.

current

( Hayden,

Academic

models

B .,

e d.)

Press:

New

York.

Diamond, on

G .

( 1979)

stone

pp.1 59-166.

Donner,

The

artifacts. Academic

H .

and

Aramaeische

nature In

Press:

Roellig,

of

so-called

( Hayden,

W .

New

B .,

L ithic

surfaces

Use-Wear

York.

( 1962,

Inschriften.

polished

cd.)

3

1 964) Vols.

Kanaanaeische Otto

und

HarrassowitZ

Wiesbaden.

Dorrell, P . G. ( 1983) Stone vessels, tools, and Kenyon, K . M. and Holland, T . A., Excavations at pp.485-575.

Dorrell,

Oxford:

P . G.

( in

The

prep.)

University

The

Press.

physical

et al., Arloune: Excavations at British Archaeological Reports.

objects. In: Jericho Vol.5

background.

a Prehistoric Oxford.

In:

S ite

in

Parr Syria.

Dumont, J . ( 1982) The Qantification of n iicrowear traces: new use for interferometry. World Archaeology 1 4(2)

a

pp.206-217.

Escalon dc Fonton, M . exp erimentale. Bulletin Fran5 aise

Fedje,

D .

76(7)

pp.179-187.

Ferguson,

In

Scanning ( Hayden,

Academic

J .

La l a

retouche Montbani Soci t Pr historiQue

: pp.217-221.

( 1979)

use-striae.

( 1979) dc

( 1973)

e lectron B .,

Press:

The

ed.)

microscopy Lithic

analysis

Use -Wear

of

Analysis

New York.

Heritage

of

Hellenism.

London.

Frison, G . C. ( 1979) Observations on the use of stone tools: dulling of working edges of some chipped s tone tools in bison butchering. In ( Hayden, B ., ed.) Lithic Use-Wear Analysis : pp.259-268. Academic Press: New York.

Garrod, Carmel

D . A. E. Vol I.

and

Bate,

D . M.

( 1937)

The

Stone

Age

of

Mount

Oxford.

Gendel, P . ( 1982) Lauwers, R . and

Functional Vermeersch,

2 18

analysis of P . M., Un s ite

scrapers. In: du M solithique

ancien

6 Neerharen-De

Kip.

Studia

Preahistorica

Beiqica

1

pp. 49-51.

Gendel, P . A. and P irnay, L . ( 1982) Microwear analysis of experimental stone tools: further test results. Studia Praehistorica Belqica 2 : pp.251-265.

Gould, R . A., Koster, D . A. and Sontz, A . H. L. ( 1971) l ithic assemblage of the Western Desert aborigines Australia. American Antiquity 3 6(2) : pp.149-169.

Gowiett, ( 1987)

J . A. J.,

Hedges,

Radiocarbon

archaeometry

R . E. M.,

dates

datelist

5 .

Law,

from

I . A.

the

and

Oxford

Archaeometry

2 9

and

chert.

Paper

presented

in

Perry, AMS

C .

system:

: pp.125-155.

Grace, R ., Graham, I . D. and Newcomer, Preliminary investigations into the characterisation o f wear on f lint tools: the f lint

The of

1983.

M . H. ( 1986) mathematical human uses of

In

de G . and Hart, M .B., eds.) The Human Uses Chert. Cambridge: The University Press.

( Sieveking, of

Flint

G . and

Grigson, C . ( in prep.) The faunal analysis. In: Parr, P . J. et al., Arioune: Excavations at a Prehistoric S ite in Syria. British Archaeological Reports: Oxford.

Gwinnet, A . J. and Gorelick, Expedition 2 2(1) : pp.17-32.

L .

( 1979)

Ancient

lapidary.

Gysels, J . and Cahen, D . ( 1982) Le lustr4 des faucilles les autres traces d ' usage des outils en silex. Bulletin la

Soci4t4

Harlan,

Pr4historiaue

J .R.

Archaeology

Harlan,

( 1967)

20

Fran9aise

A

wild

79(7)

wheat

J . R.

( unpublished

manuscript)

in the Old Archaeology,

World. London)

Harlan,

and

D .

J .R. and

Hayden, Institute

Zohary,

barley.

B . of

: pp.221-224.

harvest

in

Turkey.

: pp.197-201.

Agriculture Institute of

wheats

et de

( 1979)

Science

( 1966) 153

Paleolithic

Aboriginal

Studies:

2 19

The

( In

Origins the

of

Cereal

library

Distribution

of

of

the

wild

: pp.1074-1080.

Reflections. Canberra.

Australian

Humanities

Press

Inc.:

New

Jersey

;

USA.

Hayden, use-wear Use-Wear

B . ( 1979a) Snap, shatter and superfractures: on stone skin scrapers. In ( Hayden, B ., ed.) Lithic Analysis : pp.207-229. Academic Press: New York.

Hayden,

B .,

ed.

nomenclature pp.133-136.

( 1979)

committee

The

Ho-Ho

report.

In:

classification

Lithic

Use -Wear

Hayden, B . and Kamminga, J . ( 1973) Gould, Roster and on " microwear": a critical review. Newsletter of Technology

2 (1 - 2 )

Helbaek,

H .

Tufnell,

0 .,

Helbaek, Europe.

H . In

Plant

Lachish

IV

economy

in

ancient

: pp.309-317.

Iraqi .

Kurdistan

Lachish.

the Near East and eds.) Prehistoric

: pp.99-118.

Chicago.

Helbaek, H . ( 1966) Pre-Pottery Neolithic farming at Palestine Exploration Quarterly 98(1) : pp.61-66.

Helmer,

D .

( 1983)

In ( Cauvin, Neolithiques la Maison de

G .C.

Hureyra:

a

Hodges,

( 1975)

and

The

preliminary

Society

H .

Materials in 1964.

faucilles

et

les

gestes

de

M .-C., ed.) Traces d ' Utilisation sur du Proche Orient : pp.189-198. Lyon: l ' Orient 5 .

Hillman,

Prehistoric

Les

( 1976)

41

plant

In:

London.

( 1960) The paleoethnobotany of ( Braidwood, R . J. and Howe, B ., in

Sontz Lithic

: pp.3-8.

( 1958)

Investigations

and

Analysis

remains

report.

Beidha.

la

moisson.

les Outils Travaux de

from

Tell

Abu

of

the

Proceedings

: pp.70-73.

Artifacts,

Technology.

John

an

Introduction

Baker:

London.

to

First

Early printed

Holdaway, H . K. and Clayton, C . J. ( 1982) Preservation of shell microstructure in silicified brachiopods from the Upper Cretaceous Wilmington Sands in Devon. Geological Magazine 119 : pp.371-382.

Holley, G . A. and Del Keeley' s microwear Science

Bene, T . A. approach.

8 : pp.337-352.

2 20

( 1981) Journal

An of

evaluation of Archaeological

Holmes,

D .

( 1986)

Problems

encountered

m i crowear a rtifacts.

study of some Paper presented in

and Hart, Cambridge:

M .B., eds.) The University

The Human Press.

Hopf, M . ( 1983) Jericho plant Holland, T . A., Excavations Oxford: The University Press.

in

a

high-power

Egyptian Predynastic 1 983. In ( Sieveking, G . Uses

of

F lint

lithic de G .

and

Chert,

remains. In: Kenyon. K . M. and at Jericho Vol.5 : pp.576-621.

1 1er, R . K. ( 1979) The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties, and

Biochemistry. John Brisbane, Toronto.

Juel Jensen, H . s crapers from

Wiley

&

Sons:

New

York,

Chichester,

( 1981) A preliminary analysis of Ringkloster, a Danish Late Mesolithic

blade site.

Paper presented at ' Recent Progress in Microwear Studies', Tervuren. Published 1 982, Studia Praehistorica Belqica 2 pp.323-327.

Kamminga,

J .

Functional Australian 2 Vols.

Analysis of Certain Classes of Prehistoric Stone Tools. Ph. D. thesis, University of Sydney.

Kamminga, smoothing Use -Wear

Keeley, studies.

J .

( 1978)

( 1979)

on

L . H.

The

stone

Analysis

into

nature

tools.

of In

: pp.143-157.

( 1974)

World

Journey

Technique

Archaeology

Keeley, L . H. ( 1976) results. Starinqia

Keeley, ( 1977)

L . H.

( 1977)

Keeley,

L . H.

( 1978)

see

use-polish

Academic

and

L . H.

( 1979)

Keeley,

Preliminary

,

see

and

B .,

Press:

methodology

a

abrasive

ed.)

Lithic

New York.

in

microwear

5 : pp.323-336.

f lint:

L . H.

Cahen

2 21

some

and

r nicrowear

Meer assemblage. In: Van Noten, F . Les pp.78-86. Dissertationes Archaeologicae Tempel, Brugge.

Keeley,

Microcosms:

( Hayden,

Microwear on 3 : pp.49-51.

,

the

et

al.

experimental

Newcomer,

analysis

of

M . H.

the

Chasseurs de Meer Gandenses, 1 8, De

Keeley, L . H. ( 1980) Uses: a Microwear Chicago.

Experimental Determination of S tone Tool Analysis. University of Chicago Press:

Keeley, L . H. ( 1981) Reply of Archaeological S cience

to Holley and Del 8 (3) : pp.348-352.

Keeley, L . H. ( 1982) Hafting archaeological record. pp.798-809.

Keeley, L . H. ethnography Traces Orient 5 .

( 1983) Neolithic novelties: the and microwear analysis. In ( Cauvin,

L . H.

( 1983a)

( Appendix 1 to T . A., Excavations University Press.

Keeley, L . H. experimental

.

Science

4 (1)

view M . -C.,

N4 'olithiques de la Maison

Microscopic

Appendix C ) at Jericho

the

from ed.)

du Proche de l ' Orient

examination

In: Kenyon, K . M. Vol.5 : p .759.

and Newcomer, M . H. f lint tools:

Archaeological

Journal

and retooling: e ffects on American Antiquity 4 7(4)

d ' Utilisatjon sur les Outils : pp. 251-256. Lyon: Travaux

Keeley,

Bene.

of

adzes.

and Holland, Oxford: The

( 1977) Microwear a t est case.

analysis Journal

of of

: pp.29-62.

Korobkova, G . F. ( 1981) Ancient reaping tools and their productivity in the l ight of experimental tracewear analysis. In ( Kohl, P . L., ed.) The Bronze Age Civilisation of Central Asia : pp.325-349. M . E.Sharpe: New York.

Lawrence,

R . A.

s ignificance

( 1979)

of

Experimental

attributes

Hayden, B ., ed.) Lithic Academic Press: New York.

Mansur, M . E. striations:

used

in

Use -Wear

( 1981) Microwear new clues to

evidence

edge-damage Analysis

of

dry

hide

M . E.

working

:

( 1983) tools:

Scanning the

of

humidity in microwear polish formation. Archaeological Science 1 0(3) : pp.223-230.

2 22

In

and use striation

in Microwear Praehistorica

electron

role

the

pp.113-121.

analysis of natural the mechanisms of

formation. Paper presented at ' Recent Progress Studies', Tervuren; published in 1 982, Studia Belgica 2 : pp.213-233.

Mansur-Franchomme,

for analysis.

microscopy

abrasives Journal

and of

Marfoe, L ., Copeland, L . and Parr, P . J. ( 1981) Arjoune 1978: preliminary investigation of a prehistoric site in the Roms Basin/Syria. Levant 1 3 : pp.1-27.

Masson, A ., Coqueugniot, E . and Roy, S . ( 1981) S ilice et t races d ' usage: le lustr des f aucilles. Nouvelles Archives Museum d ' Histoire Naturelle de Lyon 19 ( Suppl.) : pp.43-51.

Maurizio, A . ( 1927) Berlin: Parey.

Meeks,

N . D.,

( 1982)

Die

Geschichte

S ieveking,

Gloss

and

G .

de

use-wear

s imilar phenomena. pp.317-340.

Journal

Mellaart, J . ( 1975) Scribner's: New York.

The

G .,

unserer

Tite,

traces of

Pflanzennahrung.

M .S.

on

Archaeological

Neolithic

Merrillees, R . ( 1962) Opium trade Antiquity XXXVI : pp.287-292.

in

of

the

Miller,

R .

( 1983)

Chisel-ended

Miller, note on ed-Din

of

the

arroweads

Institute

of

R ., Bergman, C . A. and Azouri, I . reconstructing aspects of archery Tannira.

Berytus

30

from

J . and

Near

Bronze

R . ( 1982) Pseudo-tools created by Syria. Journal of Field Archaeology

Cook, sickles

Science

the

Miller, Halawa,

Syria. Bulletin pp.187-190.

and

f lint

9

East.

Age

Levant.

l ivestock from 9 : pp.281-283.

Tell

Hadidi,

Archaeology

20

( 1982) Additional equipment at Shams

: pp.53-54.

Moffett, L . In: Parr,

( in prep.) The botanical remains from Arjoune. P . J., et al., Arjoune: Excavations at a

Prehistoric Oxford.

Site

Moore,

A . M. T.

a model pp.224-235.

in

Syria.

( 1982)

for

the

British

Agricultural 1980s.

223

Archaeological

origins

World

in

the

Reports:

Near

Archaeology

East:

14(2)

Moss,

E . H.

( 1983)

The

Functional

Implements: Pincevent and Pont from the French Final Paleolithic. Reports, International Series 1 77:

Moss,

E . H.

( 1983a)

from Tell Abu d ' Utilisation pp.143-161.

The

functions

Analysis

of

burins

Hureyra, Syria. In ( Cauvin, sur les Outils N'olithiques

Lyon:

Travaux

de

la

o f

Flint

d ' Ambon: two case studies British Archaeological Oxford.

Maison

de

and

tanged

points

M .-C., ed.) Traces du Proche Orient l ' Orient

5 .

Moss, E . H. and Newcomer, M . H. ( 1981) reconstructions of tool use at Pincevent: microwear and experiments. Paper presented at ' Recent Progress in Microwear Studies', Tervuren; published in 1982. Studia Praehistorica Belgica 2 pp. 289-312.

Neuville,

R .

faucille

pr historique

( 1934)

Les

h braigue pp.3-21.

d ' Exploratjon

d buts

en

de

Palestine.

et

l ' agriculture Recueil

d ' Arch ologie

de

la

et

la

Soci ete

Palestinienne

Newcomer, M . H. ( 1972) An Analysis of a Series of Burins from Ksar Akil ( Lebanon) Ph. D. thesis, University of London, Institute of Archaeology. .

Newcomer,

M . H.

from Ksar pp.138-153.

Newcomer, pp. 62-64.

( 1974) Akil

M . H.

Study

and

( Lebanon)

( 1976)

replication World

.

Spontaneous

of

bone

Archaeology

retouch.

tools 6 (2)

Staringia

3

Newcomer, M . H. ( 1981) Stone-carving with flint: experiments with a Magdalenian lamp. Staringia 6 : pp.77-79.

Niklevski, J . and Van pollen diagram from Neerlandica

Nissen,

K .

19

and

Zeist, W . ( 1970) A northwestern Syria.

: pp.737-754.

Dittemore,

M .

( 1974)

wear pattern analysis: a study Tebiwa 17(1) : pp.67-88.

Noy, T ., Legge, A . J. excavations at Nahal Prehistoric

Late Quaternary Acta Botanica

Society

39

and Oren,

of

Ethnographic socketed

eskimo

data

Higgs, E .S. ( 1973) Israel. Proceedings

: pp.75-99.

224

and

scrapers.

Recent of the

Odell, G . ( 1978) Préliminaires des pointes microlithiques Bulletin de pp. 3 7-49.

Odell, tools:

G .

la

Soci 4t4

d ' une analyse de Bergumermeer

Pr historique

fonctionelle ( Pays-Bas)

Fran aise

75(2)

( 1981) The mechanics of use-breakage some t estable hypotheses. Journal

Archaeoloqy

8 : pp.197-209.

Odell,

( 1983)

G .

d ' utilisation.

Probl mes

In

( Cauvin,

dans

M .-C.,

l ' ed.)

sur les Outils N olithiques du Proche Lyon: Travaux de la Maison de l ' Orient

Odell, G . reliability

Syriennes

approach.

Journal

( 1976) Donn es ( sondage A4)

Parr, P . J., Prehistoric Oxford.

by

stone Field

traces

d ' Utilisation

Orient 5 .

Field

nouvelles Annales

.

XXVI

of

des

Traces

and Odell-Vereecken ( 1980) of lithic use-wear assessments

the low-power pp.87-120.

Otte, ‚M . d ' Apamee

tude

of of

:

pp.17-24.

Verifying the "blind tests":

Archaeology

( 7(1)

sur le N olithique Arch4 'ologiques Arabes

: pp.101-118.

et al. ( in prep.) Arjoune: Excavations at a Site in Syria. British Archaeological Reports:

Perl s, C . and Vaughan, P . ( 1983) Pieces lustr es, travail des plantes et moissons Franchthi ( Gr ce) ( X me-IV me mill. B . C.). In ( Cauvin, M .-C., ed.) Traces d ' Utilisation sur les Outils N'olithiques du Proche Orient : pp.209-224. Lyon:

Travaux

P feiffer,

L .

Beziehungen

de

la

( 1912) zur

Maison

Die

de

l ' Orient

5 .

steinzeitliche

Gegenwart.

Technik

und

ihre

Jena.

P lisson, H . ( 1981) Analyse fonctionelle de 95 micro-grattoirs " Tourassiens". Paper presented at ' Recent progress in Microwear Studies', Tervuren; published 1982. S tudia Praehistorica Belgica 2 : pp.279-287.

P lisson,

H .

( 1983)

d ' utilisation. Fran5aise

8 0

De

Bulletin

la

conservation de

: pp.74-77.

225

la

des

Soci 6t4

micropolis Pr historique

P lisson, H . and Mauger, M . ( 1986) Chemical and mechanical a lteration of r nicrowear polishes: an experimental approach. Paper presented in 1983. In ( Sieveking, G . de G . and hart, M . B., eds.) The Human Uses of Flint and Chert. Cambridge: The University Press.

Poplin, usure.

F .

( 1974) Deux cas particuliers In ( Camps-Fabrer, H ., ed.)

de d bitage par Premi er Collogue

International sur l 'Industrie de l ' Os dans l a Pr histoire pp.85-92. Aix-en-Provence: Universit 7 6 de Provence.

Roodenberg,

J . J.

( 1983)

Traces

d ' utilisation

sur

les

haches

polies de Bouqras ( Syrie). In ( Cauvin, d ' Utilisation sur les Outils N olithigues pp.177-186. Lyon: Travaux de la Maison de

M .-C., ed.) Traces du Proche Orient l ' Orient 5 .

Rottlaender,

R .

( 1975)

patina

Archaeometry

1 7

: pp.106-110.

The

formation

Roy, S . ( 1982) M thodologie d ' utilisation sur les lames

of

( 1983)

Sauer, C . O. ( 1958) 32 : pp.187-189.

Schiemann,

E .

Geschichte

und

Sheep

and

Jericho

( 1948)

Man.

and

Weizen,

Verwendung.

f lint.

pour l ' tude des t races et eclats bruts: recherche

exp rimentale d ' apres un niveau pinatoufien ( Syrie). Cahiers de l ' Euphrate 3 : pp.165-176.

Ryder, M . L. London.

on

Gerald

composite

Roggen, Jena:

Mureybet

Duckworth

sickles.

Gerste.

Verlag

de

von

&

Co.,

Antiguity

Systematik,

Gustav Fischer.

Seitzer, D . J. ( 1978) Forms vs. function: microwear analysis and its application to Upper Paleolithic burins. Meddelanden fran Luns universitets historiska museum ( New Series) 2 : pp.5-20.

Seitzer-Olaussen effects of Proceedings

D . J.

heat of the

( 1983)

Experiments

to

investigate

treatment on use-wear on f lint Prehistoric Society 49 : pp.1-13.

Semenov, S . A. ( 1964) Prehistoric Technology, M . W. Thompson. London: Cory, Adams & Mackay.

2 26

the

tools.

translated

by

Shackley, M . ( 1974) Stream Nature 2 48 : pp.501-502.

Shafer,

H . J.

analysis

and

in

Shepherd,

Use -Wear

W .

of

R . G.

( 1979)

Holloway,

determining

ed.) Lithic New York.

abrasion

( 1972)

stone

tool

Analysis

Flint.

S immonds, N . W. ( ed.) Longman Inc .' : New York.

Faber

and

In

Notes on : pp.53-68.

residue

( Hayden,

Academic

Faber:

Evolution

implements.

Organic

function.

: pp.385-399.

( 1976)

Spurrell, F .C. ( 1892) Archaeological Journal 49

f lint

London.

of

Crop

early

Plants.

sickles.

Stapert, D . ( 1976) Some natural surface modifications f lint in the Netherlands. Paleohistoria 1 8 : pp.8-41.

Steensberg, Kopenhagen:

A . ( 1943) Ancient Nordisk Forlag.

Stekelis, M . Period. The

Sweet,

( 1973) Magnes

L . E.

Taeckholm,

( 1960)

V .

and

Tell

G .

Egypt. Bulletin of I University Press,

Harvesting

Togaan,

and

of

Drar,

M .

the Faculty of Cairo. 3 vols.

a

on

Implements.

The Yarmukian Culture of the Press, The Hebrew University:

Anthropological Papers, Museum of Michigan, No.14. Ann Arbor.

B .,

Press:

Neolithic Jerusalem.

Syrian

Village.

Anthropology,

University

( 1941-1954)

Science

1 7,

2 8,

Tixier, J . ( 1974) Glossary for the Description Tools ; translation by M . H. Newcomer, Newsletter Technology, Special Publication No.1.

Flora 3 0.

of of

of

Fouad

Stone Lithic

Tosi, M . and Piperno, M . ( 1973) Lithic technology behind the ancient lapis lazuli trade. Expedition 1 6(1) : pp.15-23.

Tringham,

R .,

Cooper,

G .,

A . ( 1974) Experimentation new approach to lithic Archaeology

1

Odell,

G .,

Voytek,

B .

and

Whitman,

in the formation of edge damage: a analysis. Journal of Field

: pp.171-196.

227

Turkovski, L . ( 1969) P easant agriculture i n Hills. P alestine Exploration Quarterly 1 969(1)

t he Judean : pp.21 33,

pp.101-112.

Turville-Petre, F . ( 1932) e l-Kebarah. Journal o f the 6 2

Excavations i n t he Royal Anthropological

Mugharet I nstitute

: pp.271-276.

Uerpmann, Tannira,

H .-P. ( 1982) a Halafian

Faunal remains s ite in Northern

f rom Syria.

Shams ed-Din Berytus 3 0

pp.3-52.

Unger-Hamilton, R . ( 1983) An i nvestigation into the variables a ffecting the development and the appearance o f p lant polish on f lint blades. In ( Cauvin, M .-C., ed.) Traces d ' Utilisation sur l es Outils Neolithiques du Proche Orient

: pp.

2 43-250.

Lyon:

Travaux de

la Maison de

l ' Orient

5 .

Unger-Hamilton, R . ( 1984) The formation of use-wear polish on f lint: beyond the " deposit versus abrasion" controversy. Journal o f Archaeological S cience 1 1 : pp.91-98.

Unger-Hamilton, R ., Grace, R ., Miller, R . and Bergman, C . A. ( 1987) Experimental replication, use and microwear analysis of spindle-tipped borers from Abu Salabikh, I raq. In ( Stordeur, D ., ed.) Main et Outil Table Ronde, C . N. R.S., ‚

Maison de

l ' Orient Mediterraneen.

Van Zeist, W . ( 1970) The Oriental Institute excavations at Mureybit, Syria: preliminary report on the 1965 campaign. Part I II: the paleobotany. Journal of Near Eastern Studies 2 9

: pp.167-176.

Van Zeist, W . and Bakker -Heeres, J . A. H. ( 1975) Evidence for l inseed cultivation before 6 000 bc. Journal of S cience 2 pp.215-219.

Van Zeist, W . and Bakker -Heeres, J .A. H. ( 1979) Some economic and ecological a spects of the plant husbandry o f Tell Aswad. Paleorient

Van

5 : pp.161-169.

Zeist,

W .

investigations SyrienneS

1 6

and at

Bottema, Ramad.

S .

Annales

: pp.179-180.

2 28

( 1966)

Paleobotanical

ArchdoloqigUeS

Arabes

Van and

Zeist, W . and barley from

Botanica

Van and

Casparie, W . A. ( 1968) Wild einkorn Tell Mureybit in northern Syria.

Neerlandica

Zeist, W . climate

1 7

wheat Acta

: pp.44-53.

and Woldring, H . of northwestern

( 1980) Syria.

Holocene vegetation Paleohistoria 2 2

pp. 1 12-125.

Vaughan, P . ( 1980) Microwear analysis of experimental and obsidian tools. S tarinqia 6 : pp.90-91.

flint

Vaughan, P . ( 1981) Lithic Microwear Experimentation and the Functional Analysis of a Lower Maqdalenian Stone Tool Assemblage. Ph. D. thesis. University of Pennsylvania, Philadelphia.

Vayson de Pradennes, Solferino ( etude pp. 393-422.

A . ( 1919) Faucille pr historique comparative) L ' Anthropoloqie 29

de

.

Vita-Finzi, C . and Higgs, E .S. ( 1970) Prehistoric economy in the Mount Carmel area of Palestine: site catchment analysis. Proceedings of the Prehistoric Society 3 6(1) : pp.1-37.

Wilke,

P . J.,

Bettinger,

R .,

King,

( 1972) Harvest selection and Antiquity XLVI : pp.203-209.

Wilmsen, E . N. ( 1968) artifacts. American

Witthoft, Antiquity

Wulff, M .I. T.

Wylie, Hogup

T . F.

and

on

flint

H . E. ( 1966) The Traditional Crafts Press: Cambridge, Massachussetts.

H . G.

in

seed

( 1975)

Tool

Utah.

Tebiwa

microwear 1 7(2)

Wynn Parry, D . and Smithson, F . silica depositions in the leaves of Botany 2 8 : pp.169-185.

229

and

tools.

of

stone

American

Persia.

functional

J .F.

plants.

Functional analysis of f laked Antiquity 33(2) : pp.156-161.

J . ( 1967) Glazed polish 3 2(3) : pp.383-388.

Cave,

O ' Connell,

domestication

types

The

from

: pp-1 - 3 1.

( 1964) Types of of British grasses.

opaline Annals

Yerkes,

R . W.

Mississippian 4 8(3)

( 1983) craft

Microwear,

microdrills,

specialisation.

American

: pp.499-518.

Zohary, D . ( 1969) The progenitors of wheat relation to domestication and agricultural Old

and

Antiquity

World.

In

Domestication pp.47-66.

( Ucko, and

P . J.

and

Exploitation

Chicago.

2 30

Dimbleby, of

and barley dispersal in G . W.,

Plants

and

eds.)

in the The

Animals

Appendix

-

List of

Experiments

The i tems are l isted in the f ollowing order: t he t ool, t he s tone type o f the t ool, the movement o f the t ool, t he c ondition o f the material, the material, the number of s trokes or the t ime of use, t he experimenter. The

f ollowing

abbreviations

are

u sed

for

the

s tone

t ype: ( B) -Brandon f lint, ( PB)-Potter's Bar f lint, ( 5) -Surrey f lint, ( Syr) -Syrian f lint f rom Arjoune, ( Syr A ,B,C, D) -refers t o different t ypes of f lint f rom Arjoune ( see Part I , Chapter 1 7), ( F) -French chert, ( E)-Egyptian chert, ( L) -Lebanese chert. All blades were used indicated. S crapers, borers, movement i ndicated by indicated. Drills Christopher Bergman. All

materials

the were

were

to cut, drills etc.

unless otherwise were used with the

t ool's name, unless used in a bow-drill

fresh,

unless

otherwise

otherwise made by

indicated.

The duration o f work i s measured in s awing motions ( s.i n.) or s ingle strokes ( str.), or e lse in hours and minutes ( mins.). " To . .. s . m." means that t he t ools were examined at various stages during work. The experiments were carried out by me, unless otherwise indicated. The experimenters made their own t ools. The following abbreviations are u sed for the experimenters: CB-Christopher Bergman, MHN-Dr. Mark Newcomer, GCH-Gordon Hillman, PH-Philip Harding, PD-Peter Dorrell, PP-Peter Parr, RM-Robert Miller, EHM-Dr. Emily Moss, JHM-John Hope Mason, CM-Christopher May, RUH-Romana Unger-Hamilton. Experiments during which I was not present, are only l isted i f the wear-traces resulting from t hem are discussed in the t ext or i llustrated. Experiments carried out for the blind t ests are not l isted. Experiments w ith plants involved the t esting of a great number o f variables, and the information a s regards these experiments i s s lightly different. S ingle b lades were used, unless otherwise s tated. P lants were cut a t varying heights of the stem, unless otherwise stated. Two a dditional i tems are the month and year of the experiment ( expressed i n numbers ) and the place; this was in Great Britain, unless otherwise i ndicated. ‚

Wood 1 /1-5 5 b lades, 3 200 str. each. -

2

-

blade

( B)

,

( B,

dried,

Syr

A-D),

oak,

8 00

2 31

used

s . m..

to

s crape,

beech,

t o

3 4 5 6 7 8 9

-

-

-

-

-

-

1 1 1 2 1 3 1 4 15

( B)

blade

( Syr

A ),

used

to whittle,

beech,

6 00

str..

b lade

( Syr

B ),

used

to whittle,

beech,

6 00

str.-

s craper

-

10

blade

used

,

( PB)

borer

( Syr),

burin

( PB),

blade

( Syr),

-

-

-

-

-

-

to

whittle,

dried,

,

dried, u sed

1 0

groove,

( with

oak,

1 200

wood,

to

oak

oak,

beech,

dried,

bark),

blade

( Syr),

used

to

whittle,

burin

( Syr)

used

t o

groove,

s craper

( Syr)

scraper

( PB)

scraper

( Syr),

dried,

,

oak, oak,

dried,

mins.,

PD.

s . m..

oak, oak,

1 000

2 0 2 0

mins.. mins..

str..

1 000

str..

sycamore

( with

bark),

1 6/1-2 s craper ( Syr) sycamore, 1 distal 6 00 str., 2 lateral edge, 1 200 str.. ,

-

2 0

mins..

dried,

,

PD. wood,

2 000

( L),

s tr.

s . m.

mins.,

blade

,

1 5

6 00

-

6 00

edge

s . m..

of

tool,

-

1 7/1-2 other

-

-

1 8

1 9

burin

-

for

blade ( B), sycamore, l ateral edge, used to

1 2

( S),

hours,

used

a sh,

denticulate

-

to

1 - lateral edge, 1 000 whittle, 1 000 str..

groove

‚

dried

and

soaked

s . m.,

in

2

water

1 hour.

( S)

used

,

to

scrape,

ash

( with

bark)

,

3 0

mins.. 2 0 2 1 2 2 2 3

-

-

ash,

1 hour. ( S),

blade

shave

S . M. -

( PB),

denticulated blade ( S)

scraper

-

5 00 2 4

scraper

-

,

used

( Syr)

,

to

f irst

ash,

3 0

and

cooked

mins..

s aw,

birch,

cow bone,

3 0

mins..

then birch

notch

( Syr)

burin

( Syr),

,

used

to

scrape,

b irch

( with

used

to

groove,

b irch,

1 000

bark)

S . M.. 2 5 2 6 2 7

-

-

-

pick

wood,

each.

( Syr),

scraper

used

( Syr)

,

t o

oak,

chop, 1 5

oak,

mins..

2 32

3 0

mins..

s . m..

,

1 000

2 8

-

2 9

-

3 0

-

3 1

-

3 2

-

3 3

-

3 4

-

3 5 o ak, -

borer

s craper

-

3 8

-

3 9

-

4 0

-

4 1

-

4 2

-

4 3

-

4 4

-

( Syr),

4 9

-

5 0

-

5 1

-

mins..

mins..

burin

( Syr)

used

t o

groove,

( with

bark)

pick

( Syr)

drill

( B)

drill

( B),

,

oak

,

,

mins.

dried,

a sh,

1 7

mins.,

CB.

denticulated

a fter

1 0

( Syr),

( B),

dried,

cherry

,

ash,

( with

used

t o

plane,

dried,

f lake

( F),

used

t o

plane,

wood,

notch

( PB),

blade

( B),

used

to whittle,

notch

( B)

used

to

scraper borer

,

( B),

( B),

blade

with

dried, ( B)

hook ( B)

( B)

1 0

blade

( B)

dried,

drill

( B),

( B),

-

blade

( B),

dried,

mins.,

cherry,

RM.

RM. RM.

cherry,

1 0

mins..

5 mins..

m ins..

scrape,

to

to

dried,

incise,

saw

and

wood,

used

3 0

oak,

cherry

1 0

( with

mins..

whittle,

oak,

mins.

2 2

mins.,

t o

chop,

CB. hazel,

Bone 1

mins.,

mins..

dried wood,

axe

CB.

3 mins..

used

used

,

oak,

,

to

1 0

1 0

mins.,

1 0

dried

mins..

pine,

dried,

dried,

cherry,

3 0

1 0

wood,

cherry,

used

,

( B),

tranchet

cut,

dried,

borer

,

shave,

mins.' use,

mins.,

bark),

( B),

to

CB.

1 0

f lake

used

mins

mins..

2 0

2 piercers

scraper

1 5

1 5

a sh,

blade ( Syr), 1 hour.

-

,

oak,

dried,

-

-

1 5

1 5

4 7/1-2 2 blades mins. each. 4 8

oak,

mins..

oak,

-

-

1 5

( Syr),

4 5 denticulate bark), 3 mins.. 4 6

oak,

borer

3 6/1-2 3 7

( Syr),

cow,

3 0

mins.,

2 33

CB.

1 500

str.,

PH.

1 0

2

blade

-

( B)

,

cooked

3 /1-5 5 blades S . M. each.

( B)

-

4 /1-5 chicken

6

hafted

-

7 5 00

f lake s . m..

-

( Syr

dried,

( Syr)

dried,

cow

( Syr)

,

,

used

to

to bore,

( Syr),

used

9

( Syr),

cooked,

-

1 0

blade -

scraper

1 1/1-2 cooked, 12 1 3 14

-

-

-

-

( Syr),

2 f lakes, 1 pork rib, 8 00

-

,

cooked,

scraper

( Syr)

,

cow

blade, ‚

1 7 1 8 1 9

-

-

-

scraper

( Syr),

scraper

( B),

( S),

used

to

shoulder

,

used

-

1 000

1 000

shin,

1 000

str..

cow

shin,

cooked,

cooked,

,

cooked,

cow

shin,

1 000

str.

s craper

( Syr),

cooked,

cow

shin,

1 000

str.

scraper

( Syr)

cooked,

cow

shin,

5 00

2 3 2 4

-

-

pick

25 notch str.. -

( Syr) ( Syr

,

,

dried,

)

‚

used

cow

shin,

to

s crape,

2 34

s crape,

1 0

and

2 t ool-edges str.. str..

1 000

shin,

( Syr)

-

cooked

cow

scraper

2 2

t o

1 000

2 1

,

5

s . m..

denticulated s . m.

-

( S)

b lade,

s tr.

( with meat),

groove,

blade,

8 00 • str..

2 0 1 000 -

blade

shoulder

-

( S)

1 hour.

5 mins..

-

cow

8 00

c ooked,

blade,

cooked, pork r ib, 1 -2 1 000 str. each, 3 2 00

cooked,

t o

s tr..

shin,

shin

l eg,

s . m..

2

r ib,

2 000

cow

cow

s m..

s crape,

blade,

8 00

rib,

pork

cow

lamb

denticulated blade borer

dried,

r ib,

2 00

shoulder

dried,

pork

shin,

t o

shoulder

used to cut, S . M. each.

( Syr)

-

-

pork

scraper

denticulated

cow

l eg,

chicken

u sed

incise,

cooked,

15/1-3 scraper ( PB) with different angles, 1 6

chicken cooked,

,

( Syr),

8 f lake mins.. -

A-D)

s craper

s craper

-

dried,

5 b lades ( B) ( Syr A-D), l eg, t o 4 00 s tr. each.

-

5

and

dried,

s . m..

1 5

mins..

c ow

shin,

s tr..

mins.. dried,

cow

s hin,

5 00

2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4

-

-

-

-

-

-

-

-

-

burin

( Syr)

,

u sed

t o

groove,

dried,

cow

shin,

1 0

mins..

burin

( Syr)

,

u sed

t o

groove,

dried,

c ow

shin,

1 0

mins..

blade

( Syr)

chicken

l eg,

1 000

drill

( Syr),

1 2

mins.,

drill

( F),

dried,

cow

shoulder

drill

( B),

dried,

cow

r ib,

drill

( B),

bone,

blade

( PB),

f lake

( PB)

unknown, 3 5

unknown, 3 6

-

3 7

-

3 8

-

3 9

-

4 0

CB.

CB.

3 mins.,

s crape,

RM.

meat

f rom

bone,

t ime

( PB)

used

,

to

cut

into

bone

for

marrow,

t ime

RM.

f lake

( Syr),

used

to groove,

dried,

cow

shin,

5 mins..

borer

( Syr)

used

t o

dried,

cow

shin,

5 T h ins..

f lake

( Syr),

used

t o bore,

burin

( B)

piercer axe

-

4 5

t o

mins.,

mins.,

CB.

bone,

u sed

1 4

( B)

-

4 4

mins.,

boiled,

,

1 0

blade,

CB.

notch

-

4 3

1 0

shin,

( PB),

-

4 2

cow

blade

-

4 1

dried,

s .m.

RM.

f lake

-

cooked,

,

-

,

,

boiled used

,

( Syr)

detergent,

to whittle,

used

( B),

in

used

to

chop,

,

used

to

shave,

( B)

blade

( B),

deer

cooked,

foot,

deer

deer

foot,

r ib,

3 0

foot,

cow

1 0

1 000

str..

RM.

5 mins..

5 mins..

T h ins.. 3 0

shin,

leg,

mins.,

shin,

beef

dried,

chicken

1 0

cow

cooked,

to groove,

,

notch

dried,

groove,

to groove,

used

bone,

T hins.. 1 0

T hins..

15 T h ins..

T h ins..

Hide 1 2

-

-

s craper

( Syr)

,

tanned,

scraper

( Syr)

,

tanned

pig, and

soaked

in

str.. 3 4 5

-

-

-

hafted s craper hafted

s craper

( F)

,

( Syr),

deer,

s craper

( Syr)

deer, 3 2 ,

2 hours,

mins.,

deer,

2 35

8 4

CB.

CB. mins..

water,

pig,

1 000

6

-

7

-

8

s craper

( B),

scraper

( Syr),

f lake

-

t issue 9

drill

-

1 0

-

1 1

-

1 2

1 4

-

1 5

-

1 6

1 8

cut, mins.

drill

( B),

tanned,

( F),

seconds, deer, ( with

l eather,

( B)

scraper

( Syr)

( B),

used

,

,

meat, ( See

f allow-deer,

3 2

deer

s craper

blade

-

t o 1 0

t anned,

( F),

mins..

used

cow,

piercer

-

,

5 2

hide,

( B),

blade

-

1 7

( B)

5 mins..

deer,

drill

borer

-

,

deer

s craper

-

1 3

( Syr)

f rom

deer,

2 00

( B)

,

defrosted,

( Syr)

,

used

to

3 0

mins.,

,

soaked

for

deer,

5 mins..

deer,

2 0

holes.

f or

three

5 mins.. 5 holes. defrosted,

soaked

-

-

CB.

5 mins..

1 9 denticulate ( Syr), used to scrape, days, roe-deer, 7 mins., with EHM. 2 0 s craper ( Syr) r ains., with EHM.

CB.

CB.

defrosted,

deer,

pierce,

c onnective 1 4.)

RM.

deer,

deer,

holes,

holes,

dehair,

defrosted,

3 5

ochre),

defrosted,

and Meat

CB.

5 mins.,

to

f at

below,

t hree

days,

roe-deer,

7

Antler 1 /1-3 days, 2

-

burins ( PB) red deer, 5 00

-

2 blades

( PB)

3 /1-2 2 blades 1 000 s . m..

A ,B), each.

dried,

‚

( Syr),

-

4 /1-2 s craper edge 1 000 s . m.,

( Syr S . M.

used

reindeer,

dried,

to groove,

5 00

S . M.

reindeer,

1

-

soaked

5 00

s . m.,

1

( Syr) used to t o 1 200 s . m..

6

borer

( Syr)

,

soaked

for

3 days,

reindeer,

8 holes.

blade

( Syr)

,

soaked

for

3 days,

reindeer,

1 0

7 8

-

-

-

,

burin

reindeer,

( Syr) 1 0

2

-

one

-

5 burin reindeer, -

7

each.

( Syr), soaked for 3 days, reindeer, 2 another edge 2 000 s . m..

-

for

,

used

groove,

to

s oaked

groove,

r a ins..

2 36

f or

soaked

three

days,

r a ins..

for

3

days,

9 borer ( Syr) reindeer, 1 0 mins..

used

,

-

1 0

-

1 1

-

1 2

-

1 3 1 4

deer, 1 5

deer, 1 6

f allow deer,

1 4

mins.,

CD.

drill

( B),

s oaked,

f allow deer,

1 0

mins.,

CB.

borer

( Syr),

1 5

soaked

( Syr),

f or

soaked

deer,

3 0

( Syr),

used

( B)

u sed

,

( Syr)

t o

to

3 days,

fallow deer,

soaked

for

groove,

soaked

whittle,

used

t o

1 8 denticulated blade deer, 3 mins..

soaked

notch

( Syr),

used

f allow-deer,

15

2 0

( Syr),

scraper

groove,

( Syr)

-

-

,

3 0

3

mins. 2 hours

days,

f allow

f or

3 days,

f allow

for

1

day,

fallow

1

day,

f allow

1

day,

f allow

3 mins. .

-

-

for

fallow deer,

days,

mins.

1 7 burin ( Syr), deer, 1 00 str..

1 9

3 days,

3

mins..

blade

-

f or

s oaked,

burin

-

soaked

( B),

denticulated blade

-

groove,

drill

scraper

-

t o

to

soaked

soaked

,

f or

for

whittle,

soaked

for

7

days,

mins.. dried,

red

deer,

1000

s . m..

4 hours,

cow,

1 200

Horn 1 . 2

blade

( Syr),

s craper

-

s oaked

( Syr),

for

soaked

for

4 hours,

cow,

3 burin str..

( Syr),

used

to

groove,

soaked

4

( Syr)

used

to

incise,

soaked,

-

f lake

-

,

for

s . m..

1 200

str..

4 hours,

cow,

2 0

5 mins..

Meat 1

-

2 3 4 5

-

-

-

-

denticulated blade

( B),

blade

( S)

,

beef,

blade

( B)

,

cooked,

l amb

f lake

( F),

used

cut

f lake

( PB),

u sed

pork,

4 00

s . m..

5 mins..

to to

and meat

cut,

pork

from bone,

chicken

RM

2 37

( with bone) 1 0

,

5 mins.

mins.,

( with bone),

15

RM. mins,

6 7 8 9

-

-

-

-

1 0

blade

( PB),

cooked,

beef,

5 mins.,

blade

( PB)

cooked,

beef,

t ime

blade

( PB),

blade

( Syr)

notch

-

1 1/1-4

,

pork,

dried,

,

( Syr)

,

4 blades

-

5 mins.,

to

( B)

t endon

,

unknown,

RM.

RM.

s inew,

used

RM.

5 mins..

cut,

dried,

s inew,

f rom deer

5 mins..

bone,

5 mins.

1 2 s everal arrowheads, including 3 t ransverse ( B), hafted with resin, wax and s inew in f letched arrows, each used to shoot once at d istance o f defrosted, deer carcass, CB.

a rrowheads and barbed 4 meters,

-

1 3 Point once, meat

( B) and

-

,

hafted in bone, MHN.

1 4 f lake ( Syr) t issue f rom deer -

,

used hide,

a

barbed

t o cut, 1 0 mins.

arrow,

each.

used

t o

shoot

meat, f at and connective ( see above H ide 8 )

Fish 1

blade

-

2 /1-2

-

( B),

used

2 blades

to

( Syr

gut

and

A ,B),

scale,

used

t o

gurnet,

s cale,

2 0

mins..

bream,

5

mins.

each. 3 4

-

-

blade

( Syr)

f lake

( PB),

used

,

used

to to

gut

and

s cale,

s cale,

f ish,

1 0

trout,

2 0

mins..

mins.,

RM.

Shell 1 2 3 4 5 6 7 8 9 1 0

-

-

-

-

-

-

-

-

drill

( B),

scallop,

12

drill

( B),

scallop,

7 mins.,

CB.

drill

( B),

cardium,

7 mins.,

CB.

burin

( Syr)

-

t o

incise,

scraper

( Syr)

,

used

s craper

( Syr)

,

concretions

borer

( B)

,

ormar,

burin

( B)

,

used

s craper

-

used

,

mins.,

s craper

( B)

used

,

( B)

to

,

to

saw,

CB.

ormar, ormar,

5 mins.. 5 mins..

f rom ormar

shell,

5 mins.

5 mins.. cut, t o

ormar,

saw,

concretions

1 0

ormar, f rom

2 38

mins.. 1 0

o rmar

mins. shell,

5 m ins..

Stone 1 /1-8 8 f lakes ( B) used t o rub against, 1 f lint ( B) without water, 2 f lint ( B) with cold water, 3 f lint ( B) with hot ( 100 C degrees) water, 4 f lint ( B) with detergent, 5 f lint ( B) with v inegar, 6 quartzite, 7 f lint ( B) without water, 8 f lint ( B) with c old water, 2 00 s .m. each. ,

-

-

-

-

-

-

-

-

-

2

-

3 4

-

-

5

-

burin

( Syr)

u sed

to

groove,

l imestone,

2 0

f lake

( Syr)

u sed

t o

groove,

l imestone,

7 mins..

f lake

( Syr)

used

to

polish,

quartzite , 1 0

T h ins.

borer

( Syr)

u sed

to

incise,

quartzite,

mins.

mins.,

1 0

6 /1-6 6 f lakes ( B) used to rub against f lint with 1 0% NaOH, 2 with 3 0% H H 0 3 with hot with cold water, 5 with 1 0% HCl, 6 with o il, each. ,

-

‚

-

-

7

-

8 9

-

-

-

drill

( B),

metamorphic

drill

( B),

lapis

drill

( B),

laps

lazuli,

( B)

,

lapis

1 1

drill

( B)

,

malachite,

1 2 drill mins..

( B)

,

lapis

1 3

drill

( B),

drill

( PB),

f lake

( Syr)

burin

( B),

used

blade

( B)

l imestone,

-

-

1 4 15 1 6 1 7 1 8 1 9 2 0 2 1 2 2

-

-

-

-

-

-

-

-

-

-

scraper blade

,

notch

( B)

blade

( Syr)

,

lazuli

used to

,

mins., ( with

to

( with

( with

incise,

groove,

to

u sed

to

1 0

-

-

CB. CB. s and

and

water)

,

3 6

sand

and

water)

,

1 5

1 0

1 0

mins.,

CB.

CB.

l imestone,

l imestone,

1 0

5 mins. T h ins..

mins.. 1 0

mins..

mins.

saw,

s crape, to

,

CB.

sand),

7 mins.,

l imestone,

u sed

used

mins.,

l azuli

l imestone,

( B),

1 9

1 hole,

( B) 1 water, 4 1 00 s . m.

5 mins..

malachite, ,

1 2

l azuli

lapis

( Syr),

( B)

scraper

,

rock,

lazuli,

1 0 drill mins., CB. -

-

PD.

l imestone, l imestone,

scrape,

2 39

1 0

mins..

5 mins..

l imestone,

1 0

T hins.

2 3

-

2 4

-

2 5

borer

( B)

,

l imestone,

borer

( B)

,

l imestone

blade

( Syr)

,

used

b lade

( Syr)

,

l imestone,

-

2 6

-

2 7/1-2 2 8

2 blades

-

borer

-

( B),

mins..

( with

t o

( Syr)

1 0

sand)

s crape,

l imestone,

1 0

mins..

l imestone,

1 min.

1 min..

l imestone,

,

,

15

1 5

mins.,

s . m.

e ach.

CB.

Pottery 1

burin

-

( Syr),

used

t o

groove,

sherd

from ARJ V

1 11.3,

5

( Syr),

used

to

groove,

sherd

from ARJ V

1 11.3,

5

borer

( Syr),

sherd

f lake

( Syr),

used

t o

incise,

sherd

from ARJ V

1 11.3,

5

( Syr),

used

to

groove,

sherd

from ARJ V 1 11.3,

3 drills

( B),

mins. 2

f lake

-

mins. 3 4

-

-

f rom ARJ V 1 11.3,

5 mins..

mins. 5

burin

-

mins., 6 /1-3

1 0

PD. -

s everal

holes,

various

sherds

from

Abu

Salabikh,

CB.

Sand 1

2 f lakes

-

( B)

,

used

to

rub

together

with English

soil,

5

mins. 2 the -

4 t hreshing f lints from Cyprus collection of the Institute of

( ethnographic items Archaeology, London ) .

3 2 0 f lint pieces ( B) ( Syr A-D) some burnt, l eft in perforated basket in a fast-running stream f or 5 weeks. ,

-

Feather 1

-

blade

( B)

blade

( Syr)

,

quill

o f

pigeon

feather,

4 00

s . m..

Wool 1 2

-

-

denticulate

,

Shetland ( Syr)

,

sheep,

Shetland

2 40

1 5

mins..

sheep,

5 mins..

i n

a

Ivory 1

drill

( B),

hippopotamus

drill

( B),

copper,

drill

( F),

copper

t ooth,

1 0

mins.,

CB.

Copper 1 2

-

-

7 mins.. ( with

s and

and

water),

7 mins..

P lant 1 t he

curved stem),

-

s ickle, 4 blades ( B), 1 00 s .m., 5 -82, GCH.

reed

2 straight s ickle, 8 blades 5 hours, 1 min., MHN.

( B)

3 t he

s traight s ickle, 7 blades stem), to 5 00 s . m., 7-82.

( B)

4 t he

curved stem),

-

-

-

s ickle, 3 blades ( B) to 5 00 s . m., 7-82.

( cut

( F)

grass

,

( cut

( cut

5 straight sickle, 1 0 blades ( PB) ( E) base of the stem), to 1 0,000 s . m., 7-82. curved s ickle, 4 blades ( PB) base of the stem), 5 00 s . m.,

-

7 8

-

-

( S),

dried

( 1

month),

reed,

( B),

domestic

barley,

7 620

at

3 000

s . m.,

s . m.,

7-82.

1 0

domestic

e inkorn,

to

domestic

e inkorn,

1 1 3 blades 9 -82. -

( B), ( B)

( L)

( PB),

1 2 ( B), 9 -82.

dried

( 4

1 3

( B),

dried

( 12

days),

( B),

dried

( 10

weeks),

( S)

bulrush,

-

1 4 1 5

-

-

-

1 6/1-7 dried blades s tems,

,

days),

to

6 000

( B)

domestic

t o

6 000

reed, s . m.

,

grass,

to

the

the

base

base

( cut

( PB)

reeds,

4 000

einkorn,

s . m., to

6 000

8 -82,

of

at

the

( at

s . m., to

to

to

6 000

9 -82.

4 000

s . m.,

8 000

s . m.,

9 -82. s . m.,

9 -82.

8 -83.

7 blades ( B), 1 -3 grass, to 1 200 s . m., 4 -7 ( 1 month) grass, t o 2 400 s . m. 9 -82. 1 7/1-11 ( B) reed ( cut a t the base o f the stem) 1 t o t o 5 00 s . m., 9 -82, JHM, CM and RUH.

-

-

,

,

,

-

,

2 41

of

8 -82.

( L i)

2 blades

of

,

4 blades

-

base

( B) Sparqaniu x n ramosum 7-82.

9 curved s ickle, s .m., 7-82. -

at

grass

,

-

6 t he

t he

( obsidian),

grass

,

at

-

1 1 5

1 8/1-7 CM 1 9

-

2 0

-

2 1

-

2 2

( B)

reed,

,

1 t o

5 s tems,

( B),

reed,

( Syr

A ),

reed,

2 500

s . m.,

1 1-82.

( Syr

B ),

reed,

5 000

s . m.,

1 1-82.

2 blades

-

2 3

( B),

-

2 4/1-2

t o

( B)

used

2

to

-

2 000

s .m.,

reed,

,

2 blades

-

s .m.,

1 00

-

7 blades

-

t o

1 600

s .m. ‚ JHM,

a nd RUH.

3 00

s crape, ( B)

3 00

s . m.

reed,

reed

,

1 1-82.

s . m.,

e ach,

5 00

( cut

2 -83,

1 1-82.

s tr.,

a t

the

1 1-82.

base

o f

t he

2 5 straight s ickle, 5 blades ( B) ( coarse-grained f lint) wheat with barley and some weeds ( cut a t -

,

of 2 6 the 2 7 at

the

stem)

,

9 hours,

s tem)

PH. ,

straight the base

-

PD,

PP,

2 8

‚

RUH.

Arjoune/Syria, 2 9 straight barley ( cut -

staff

o f

8 blades ( Syr) ( PB) reed ( mainly cut stem), 1 200 s . m., 4 -83, Arjoune/Syria,

s ickle, of the

denticulated

-

1

English t he base

curved s ickle, 6 blades ( Syr) reed ( cut at t he base s tem), 1 800 s . m., 4 -83, Arjoune/Syria, PD, PP, RUH.

-

,

JHM.

at

PD,

blade PP,

( B),

r eed,

2 000

s . m.,

4 -83,

RUH.

s ickle, 8 blades ( Syr) domesticated wheat and at the base o f the ste r n), 6 -83, Aleppo, Syria, ,

I .C. A.R.D. A.

3 0 curved mainly a t

s ickle, 7 blades the base o f

-

( Syr) Sparqaniu r n the stem), 1 200 ,

ramosum s . m.,

( cut 4 -83,

Arjoune/Syria. 3 1/1-3 used

3 blades ( Syr), cane, 1 t o s crape, 2 000 str.,

-

-

4-83, 3 2 o f 3 3

-

-

Arjoune/Syria. curved

the -

used to saw, 2 000 s . m., 2 3 used to bore, 5 holes,

-

s ickle,

stem)

( S),

,

8 blades

( E)

( B)

,

grass

( cut

a t

the

base

6 -83.

bulrush

( cut

at

the

base

o f

the

s tem),

1 000

s .m.,

8 -82. 3 4 straight s ickle, 9 blades ( Syr) wild oat, cut at the base of t he -

Jerusalem, 3 5/1-2 s . m.,

-

A .

,

wild stem),

barley ( with 2 950 s tr., 6 -83,

Miller-Rosen.

2 blades

( B),

horsetail,

7-83.

2 42

1

-

2 00

s . m.,

2

-

6 00

3 6

-

3 7

-

3 8

-

3 9

-

( Syr),

domesticated

( Syr),

domesticated barley,

( Syr),

reed,

t runcation

4 0 straight domesticated s tr., GCH. -

7 000

( Syr),

s ickle, e inkorn

barley,

s . m., used

1 00

s .m.,

s . m.,

8 -83.

8 -83.

8 -83. t o

cut,

reed,

( B) the

wild base o f

7 blades ( cut at

4 1 straight s ickle, ( cut at the base of 8 -84, PD, RUH.

1 0,500

,

3 000

s . m.,

e inkorn and 1 0% the s tem) c .2000 ,

9 blades ( Syr) ( B) domesticated the stem) t o 2 hours, 3 0 mins. ,

-

,

‚

4 2 ( cut

straight s ickle, 9 blades at t he base o f the stem)

( Syr) 9 -83.

4 3 ( cut

straight s ickle, 9 blades at the base o f the stem)

( Syr) domesticated 3 hours, 9 -83.

4 4

( S)

-

-

-

4 5

-

4 6

-

4 7

-

4 8

-

,

,

( B)

crop weeds, ,

Cyperus

2 000

2

-

-

( B)

,

domesticated

,

s . m.,

9-83.

lon us

‚

2 000

s . m.,

8 -84.

qiaantia

‚

3 000

s . m.,

8 -84.

( B),

Stipa

( B),

poppy

( Papaver rhoea

( B),

poppy

( Papaver orientale

Root Vegetable 1

8 -83.

blade

( B),

carrot,

2 00

blade

( B),

carrot,

1 000

s . m.. s . m.

2 43

) ‚

2 000

) ‚

s .m.,

spelt

e inkorn

8 -84.

2 000 s . m.,

wheat 8 -83,

8 -84.

T able 1

-

E dge D amage

T he n umber o f t ools

( total o f 7 2)

i s s hown i n b rackets.

F =feather s car, S t-S tep s car, S n=Snap s car ( see P art I , C hapter 4 ). M aterial S awing/Slicing S craping L i r restone ( 4) ( Pl.l:a), ( Pl.l:b ), m ainly S t,some e dge r ounding, F ( when h eld s om e S t,F ( 2 ) a t a ngle o f 4 5 d egrees) ‚ lots S n ( when h eld a t a ngle o f 9 0 d eg rees ( 2 ) L imestone c aused a l ot o f e dge d amage w ith l arge s cars, m ostly n umber o f F a nd S n s ee m ed t o d epend o n t he c ontact a ngle. c aused a brasion r ather t han f laking. D ried b one ( 18) ( Pl.l:c ), ( P1.l:d ),

S t. T he S craping

m ainly S t,some e dge r ounding ( 12), F ,Sn ( 5), o f t hese F a nd S t ( 10), a lmost n one ( 1) n o s cars ( 2 ) D ried b one c aused a l ot o f e dge d amage w ith l arge s cars, m ostly S t. T he i mpl ements w hich s howed n o d amage w ere o f t he s ame c oarse-grained f rost-shattered f lint. F resh r eed ( 11) ( P1.1:e ), ( P1.1:f ) m ainly S t, s ome e dge r ounding, F , Sn ( 10) S t a nd F ( 1) F resh r eed c aused a lmost a s m uch a nd t he s ame t ypes a nd s izes a s d id b one. D ried a ntler ( 4 ) ( P1.1:g) ( Pl.1:h ),

o f

s cars

m ainly S t, e dge r ounding, s ome F , S n ( 3 ) s ome S t , F ( 1) D ried a ntler c aused s omewhat l ess d amage t han t he a bove m aterials. S oaked A ntler ( 3 d ays ) ( 3)

( P1.2:a ) m ainly S n,St

( Pl.2:b ) e dge r ounding, a f ew F a nd S t

( 2)

S oaked a ntler c aused smaller s cars, l ess S t a nd l ess d ried a ntler. Wood ( fresh ( P1.2:c), ( Pl.2:d ) a nd d ried) ( 18) F ,St,Sn i n s imilar e dge r ounding,

( 1) d amage

t han

d id

p roportions ( 11) m ainly F , b ut f ew ( 7) Wood c aused l ess a nd smaller s cars t han d id h arder m aterials. S t w ere r are. I d id n ot t est t he d ifference b etween f resh a nd d ried w ood. H ide ( 10) ( Pl.2:e), a f ew ( Pl.2:f), s mall S t, m ainly e dge r ounding, a f ew F ( 2) F ( 8 ) H ide c aused m ainly F a nd s maller s cars t han d id t he a bove m aterials. d id n ot t est t he d ifference b etween f resh a nd d ried h ide. M eat ( 4 ) ( P1.2:g), m ainly F , b ut a f ew a nd v ery s mall S t ( 4) M eat c aused t he l east d amage a nd t he s mallest s cars o f l isted. H owever, a f ew v ery s mall S t c ould b e s een.

244

t he

m aterials

T able 2 P lant P olishes ( see P 1.22-25 ) f p olish ; H D=horizontal d istribution o f p olish, D D=depth d istribution o olish l ike S T=striations, C SP=comet-shaped p its, H =small h oles i n t he p t he e ffect o f s and-blasting, O =other a ttributes. c onc.=polish c oncentrated a t e dge b ut q uite i nvasive m ed. =medium H 0 D D S T C SP R D P lant f ew m ed . f ew f ew onc. W ild e inkorn c m any ed. m any m any c onc .wide m D om . e inkorn m any m ed. m any m any c onc. D om .w heat m any m ed. m any m any c onc. D om .b arley d eep f ew f ew n arrow G rass w ear-traces t oo w eak n arrow W eeds h igh w ide S tipa eep H orsetail v ery c onc d h igh w ide R eed d eep f ew w ide B ulrush c racked p olish w ide S parganium m ed. n arrow C yperus v ery c onc. d eep c orrugated e ffect C ane t hin s teaks o f p olish o nly, n o g loss P oppy -

-

-

-

-

-

-

245

T able 3

-

B lind T ests 1 -10

T he i tems a re l isted i n t he f ollowing o rder: t ool t ype, a ction, w orked m aterial. M HN=experiment b y D r. M ark H . N ewcomer R I JH=identjficatjon b y R omana U nger-Hamilton T ool M E IN R UH 1 S craper, v entral d istal e dge, s craper,vent .dist,edge, s craping, w ood. s craping, a ntler. 2 D rill, p roximal t ip , d rill, d istal t ip , d rilling, s hell, d rilling, b one. 3 P iercer, d istal t ip , r eamer , d istal t ip g rooving, a ntler, b oring, a ntler. 4 S craper, v entral d istal s craper, v entral d istal e dge, s craping, s tone. e dge, s craping, a ntler. 5 B urin , b oth b urin b urin , b oth b urin e dges, e dge, s craping, b one. g raving, a ntler. 6 7 8 9 1 0

R etouched p oint , f ired, s and, S craper, u nused. B acked b ladelet, u nret ouched e dge, c utting, n eat. B lade, r ight e dge, c utting a nd s caling f ish , B urin, v entral f acet, s craping, w ood,

u sed

a rea ,

b acked b lade, d istal e dge, s awing, b one. s craper, s craping, h ide. b acked b ladelet, u nret . e dge,cutting, m eat. b lade, r ight e dge, c utting, a ntler. b urin , v entral f acet , s craping, a ntler.

T able 4 B lind T ests 1 1-30 T he v entral a spect o f e ach f lake w as r ubbed a gainst a m inutes. T ool M E IN I JH 1 1 D ry h ide h ard+dry , d ry h ide 1 2 S oaked a ntler n ed.hard+not d ry,soaked a ntler 1 3 W ood s oft ,meat o r f resh h ide -

1 4

S oaked a ntler

1 5 1 6 1 7

F ern D ry h ide W ood

l b 1 9

F resh b one F ern

2 0

F resh b one

( touched s omething h ard ) h ard+not d ry, s oaked a ntler o r f resh b one n ed. s of t +med . dry, w ood s oft , m eat o r u nused i r ed.hard, n ot d ry, w ood o r d ry r eed h ard+not d ry, f resh b one n ed.hard, n ot d ry, r eed o r w ood n ed.hard-hard, n ot d ry, f resh b one

246

m aterial

f or

1 0

T able 5

-

B lind T ests 2 1-30

E ach i mplement w as u sed b y N ewcomer t o s aw a m aterial f or 1 0 m inutes. I t w as k nown i n a dvance t hat t he m aterials w ere t he s ame a s t hose i n t he B lind T ests 1 1-20. T ool M HN R UH ( microscopic ) R UH ( macroscopic ) 2 1 A ntler h ard+fresh, b one m ed.-hard,antler ‚ wood 2 2 W ood s oft , h ide s oft, f ern 2 3

F ern

2 4

B one

2 5 2 6 2 7 2 8

H ide F ern A ntler H ide

2 9

B one

3 0

W ood

m ed.-hard, n ot d ry, a ntler m ed.-hard+fresh, f ern s oft, h ide e dge d a inage:bone h ard+fresh, wood m ed.-hard+fresh , f ern m ad.-hard+fresh, a ntler m ed.-hard+fresh, w ood

m ed.-hard, h ide h ard, b one o r a ntler m ad . -hard ‚ wood ‚ antler h ard, b one m ed.-hard, w ood m ed.-hard, h ide s oft,

f ern

w ood

T able 6 E xperimental H arvest P roportions o f C ereals, G rass a nd W eeds T he c ounts w ere c arried o ut i n 1 983 b y s taff a t B utser Ancient F arm , F ield I V, a t t he t rial p lots w hich I h arvested w ith f lint s ickle b lades. E ach s ample c onsisted o f a 3m ater r un a long o ne r ow o f c rops . -

-

E inkorn

E inkorn

s talks 4 09 a nd 3 72 g rass n il a nd n il w eeds 2 2 a nd 3 7

s talks 4 07 a nd 3 72 g rass n il a nd n il w eeds 1 3 a nd 4 0

E i t u ier

S pelt

s talks 3 70 a nd 3 82 g rass 2a nd n il w eeds 2 8 a nd 5 0

s talks 1 02 a nd 1 22 g rass 2a nd 1 w eeds 2 0 a nd 1 7

S pelt s talks 2 00 a nd 1 97 g rass 5a nd n il w eeds 3 3 a nd 3 0

247

T able 7

-

T he L ithic A rtefacts f rom A rjoune, T renches 1 -1V, M ounds B a nd

C ( Surface ) ( after C opeland, Type

i n p rep .) Number

A rrowhead A xe, p ick, c hisel S ickle e lements E nd-scrapers F lake-scrapers S ide-scrapers B urins B orers, d rills D enticulates, n otches B acked k nives D ivers r etouched p ieces N on-flint t ools Obsidian p ieces T runcations T ool t otal

1 1 1 6 2 1 3 3 3 4 1 2 9 9 9 1 9 8

D ebitage C ores: P rismatic D iscoid D ouble Other K napping p roducts t J nretouched f lakes a nd b lades 1 c ortex-flakes I t f ragments a nd d ebris D ebitage t otal A rtefact t otal

3 7 1 5 1 6 1 62 3 0 1 31 3 55 4 53

T he a bove l ist d oes n ot i nclude t he 8 8 l ithic a rtefacts r ecovered f rom T rench I , P hase I w hich C opeland l isted i n as eparate t able. P hase I w as a pparently t he o nly u ndisturbed p rehistoric p hase f ound i n T renches I I V. O f t he 8 8 a rtefacts o nly 3 w ere c lassified a s t ools .

248

T able 8

-

L ithic A rtefacts f rom A rjoune, T rench V

( after C opeland, i n p rep .) T ype Number A xes, c hisels, p icks 7 A rrowhead f ragmen ts 3 C hoppers 5 5 S ickle-ele r r nts:lustred u nretouched 1 8 1 i t b acked a nd t runcated 7 4 i f t runcated 2 9 I t c rescentic b acked 4

0 .69 0 .29 5 .47 1 .79 7 .36 2 .88 0 .39

R etouched b ladelets, m icroliths E nd-scrapers S traight-ended s crapers F lake-scrapers

1 6 2 4 1 1 3 8

1 .59 2 .38 1 .09 3 .78

C ore-scrapers, s teep-scrapers R acloirs R aclettes B urins B orets, p iercers, d rills

3 1 3 3 1 9 4 2 4 5

3 .08 3 .28 1 .89 4 .17 4 .47

D enticulates 1 08 N otched p ieces 1 03 B acked b lades, k nives 3 C omposites 3 2 N aturally-backed k nives 2 7 A bruptly r etouched p ieces 2 9 P ieces w ith n ibbled r etouch 9 0 B ifacially r etouched p ieces 5 P ieces w ith i nverse/alternate r etouch 9 H am r t rstones 3 6 A nvils 2 4 O bsidian p ieces 7 O ther n on-flint a rtefact f ragments 7 8 T runcations T ool t otal 1 ,004 -

1 0.74 1 0.24 0 .29 3 .16 2 .68 2 .88 8 .95 0 .49 0 .89 3 .58 2 .38 0 .69 7 .76 -

9 9.45%

D ebitage C ores: P rismatic 3 00 D iscoid 4 6 P olyhedric/anorphous 6 0 Double-ended 7 1 M inute 1 05 F ragmentary 5 1 R ough-outs 3 8 K napping p roducts: T ablets a nd s pa L .s 5 6 C rested b lades/flakes 1 02 C ore r efreshment p ieces 7 5 U nretouched f lakes b lades b ladelets c ortex-flakes p art-cortex f lakes p reparation f lakes f ragments, d ebris D ebitage t otal

6 07 1 34 1 84 5 52 1 ,269 2 ,122 2 ,997 8 ,769

249

A rtefact t otal N on-artefacts T otal e xamined

9 ,773 2 ,465 1 2,248

T able 9 L ithic A rtefacts f rom A rjoune, T rench V I ( after C opeland, i n p rep .) T ype N umber % A xes 1 0 .41 A rrowheads 4 1 .65 C hoppers 1 0 .41 S ickle e lements: l ustred u nretouched 6 2 .48 b acked, 2 t runcations 1 4 5 .80 b acked, 1 o r n o t runc. 3 0 1 2.44 t runcated 1 0 4 .14 c rescentic b acked R etouched b ladelets, r n icroliths 4 1 .65 E nd-scrapers 2 0 .82 S traight-ended s crapers 1 0 .41 F lake-scrapers, f an-scrapers 9 3 .73 C ore-scrapers, s teep-scrapers 2 0 .82 R acloirs 4 1 .65 R aclettes 5 2 .07 B urins 7 2 .90 B orers, p iercers, d rills 1 6 6 .63 D enticulates 7 2 .90 N otched p ieces 1 5 6 .22 B acked b lades, k nives 3 1 .24 C omposites o f a bove ( scrapers t o k nives) 1 0 4 .14 N aturally-backed k nives 6 2 .48 M iscellaneous r etouched p ieces: a bruptly r etouched 8 3 .31 n ibbled r etouch; ' utilised' 5 7 2 3.65 b ifacial r etouch 1 0 .41 i nverse/alternative r et. 1 3 5 .39 H ammers, a nvils 1 0 .41 Obsidian -

O ther n on-flint a rtefacts T runcations ( not s ickles )

2 2

T ool t otal

2 41

D ebitage C ores: P rismatic 2 7 D iscoid/flat 9 P olyhedr i c/a i i v rphous 8 D ouble-ended 5 M inute 1 1 F ragmentary 2 3 R ough-outs 5 K napping p roducts: t ablets a nd s palls 3 c rested b lades a nd f lakes 6 c ore-refreshment f lakes 1 5 l i nretouched f lakes 1 17 b lades 1 74 b ladelets 2 1 c ortex-flakes 7 0 p art-cortex f lakes 1 09 p reparation f lakes 2 80 f rag m ents , d ebris 3 53

250

0 .82 0 .82 9 9.80%

D ebitage t otal A rtefact t otal N on-artefacts

1 ,236 1 ,477 3 83

T able 1 0 L ithic A rtefacts f rom A rjoune, T rench V II ( after C opeland, i n p rep .) Type N umber A xes, c hisels, p icks 1 0 .98 A rrowheads -

C hoppers S ickle e lemen ts :

l ustred, u nretouched b acked a nd t runcated t runcated c rescentic b acked R etouched b ladelets , m icroliths E nd-scrapers S traight-ended s crapers F lake-scrapers C ore-scrapers, s teep-scrapers R acloirs R aclettes B urins B orers, p iercers, d rills D enticulates N otched p ieces B acked k nives C omposites N aturally-backed k nives A bruptly r etouched p ieces P ieces w ith n ibbled r etouch P ieces w ith b ifacial r etouch P ieces w ith i nverse/alternate r etouch H a i n r r rstones A nvils O bsidian p ieces O ther n on-flint a rtefacts T runcations T ool t otal

4 2 1 5 2 4 4

3 .92 1 .96 1 4.70 1 .96 3 .92 3 .92

1 6 1 5

0 .98 5 .88 0 .98 4 .90

4 5 3 4 1 2

3 .92 4 .90 2 .94 3 .92 0 .98 1 .96

8 1 4 1 6 2

7 .84 1 3.72 0 .98 5 .88 1 .96

1 6

0 .98 5 .88

1 02

D ebitage C ores: P rismatic

9 9.96%

1 4

D iscoid P olyhedric/a i torphous D ouble-ended M inute F ragmentary R ough-outs K napping P roducts: t ablets, s palls c rested b lades/flakes c ore-refreshment p ieces U nretouched f lakes b lades b ladelets c ortex f lakes p art-cortex f lakes p reparation f lakes F ragments, d ebris D ebitage t otal

6 1 1 1 1 3 5 1 9 2 9 1 0 1 2 1 05 4 4 5 6 7 1 2 09 5 94 5 02

1 ,710

251

A rtifact t otal

1 ,812

P late 1

E dge D amage

-

s cra jirK

cut t f l13

a -Stone 1 7

b -Stone l b

c -Bone 1 , d ried

d -Bone 5, d ried

c -Plant 2 6,

f -Plant 2 3,

j -Ant

t L

i n,

r eed

UL L O

h -Ant1 t a ll e xperi i nts ,

5 0x

r eed

P late 2

-

E dge D amage

a -Antler 1 4,

s oaked

b -Antler

I i,

s oaked

c -Wood 2 , d ried

d -Wood 6 , d ried

c -Hide l b

t -Hiae 5

9 -Meat 1 4 a ll e xperiments,

5 0x

P late 3

-

R esidues

( SEM )

a -rone r esidues ( Anderson- erfaud,

l 9 1),

4 U00x

b -"phytolith" ( Anderson-Gerfaud,

c - xper i€ nt t C one r es 4 000x,

1 9 1),

5 000x

ues,

a ccelerated v oltage 2 5V

d -"phytolitn" ( Ander s on-Ge r faud,

4 400x

o nly t o p olish o ther f lint,

4 400x

P late 4

-

R esidues

( SEM)

a -"cellul ar v eyetable m atter"

b -object o n f lint b lade u sed

( Anderson- erfaud, 1 981), 1 800x

o nly t o p olish o ther f lint 5 4C 1Ox

c -similar o bject o n u nused

d -similar o bject o n u nused

f lint b lade , 6 000x

f lint b l ad e, 5 400x

OLL uJ L 3

o f a n

u nused

f lint

tto a

t he s urface

b lade,

t düüx

P late 5

-

Variables:

F lint

Or

S L

a -good 1 /1, B randon,

u nused

b - ooa i /i, B randon,

1 400 s .in .

LI

P •

4 ;

c - ood 1 /2, S yr.C,

f

u nused

i /4

'

/ 1

A ,

L fUu

P late 6

-

M anufacturing T races a nd B urning

t uU

e -exp .

( iA

L u

exp .

L u UX

2 10.1,

u u

I UUL

I r r ULouch,

L LL.LU

I

1 iuk ,

i. ,

2 0ux

L U)x

P late 7

-

N atural T races

a -exp . S and 1 ,

c -exp . S tone 1 /6,

e -exp. S tone 1 /6,

I

h -kHJ 8 03.3,

2 00x

+

w ater,

f lake,

t -AR J 7 01.3A ,

2 00x

u uau,

d -AR J V ,

iuUX

b lade,

l üüx

2 00x

2 00x

a -exp . w ood 1 /4

OOO

b / J I OOb.3

1 /

t 5,

o oa L a/i,

d -AR J v i, S urface

s yca i fore

S yCa iW )iC

f -kRJ 8 00.1

h -AR J

2 U.1

P late 9

a -ex j.

-

S crapers

40 0d 3 7,

+

b -APO

ork

1 14.2

A HJ M ound C

e -ex

7 03.3 1 3

s aw

A ntLeL

1),

3Oc K

J

20

P t J .0

a -exp.

B one 1 6

c -exp.

B on? 2 2,

exp .

y- eX p.

ii de

12 ,

We 5

b -AR ]

c ooked

+

o cr tr

I 000.1

d -AR J I OUD.i

f -AR J

h ii 'i

u t .3

3 UU.L

P late 1 1

c l -exp .

-

P erforators

wood 7 ,

b oring

b -ARJ 7 00.1

c -exp . Antler 9 ,

b oring

e -exp.

r ooviny

d one 4 2,

A kJ

3 1 )0.1

i u ü o

i

2Y

1

P late 1 2

-

P erforators

a -exp. WO W

U ,

ar iiliny,

t uUx

b -A I . . J

7 01.3B,

l OUx

c -exp . W ood 5 0,

u rilling,

2 00x

c l -ARJ

U i.3b,

10 0x

e 2 d,

e -exp.

Lri n g

t -AR J

u U.

-;

6 . 3

i

1 -1v,

S urf

J2

P late 1 3

a -ex) .

-

B lades, B ladelets,

o od 2 1

F lakes

b -ARJ 7 03.3B

C -tXp .

d -ARJ 1 -1V, S urface

hitt iin y

f -ARJ 2 01.2

P late 1 4

-

B lades, B ladelets a nd F lakes

B aulk 6 00-'OU.3

m eat - 4 ,

+

X )f l

i ')

2 5

P late 1 5

-

B lades, B ladelets a nd F lakes

a -exp . S tone 2 2

c -exp .

B one 2 0

e- ex .

Fi sn

d - exp.

1 ,

b -AR J M ound C

d -ARJ 7 01.3B

+

saL ny

£ -exp. F ea t her I

n tJe

h - xn . I

A ntler 3 /1,

d ried

P late l b

-

B lades a nd S napped B lades

A l

f t

; '

a -exp . M eat 1 1/1, t endon

b -exp . M eat 9 , d ried s inew

•

c -exp . V egetable 1 , c arrot

d -exp . H orn 1

‚ .

e ( -XP .

b ro t c n o lade

f -AR J 1 02.4,

h -AR J 90 0.1

b reak

P late 1 7

-

N otches

a 1 \R J 1 03.10

b -A IU 2 02.3

s having

d -ARJ

? ?5.2

e-e xp . B one 4 4, s having

1—AEJ

s U1.

c -exp . W

9-e xp . S tone 2 1, s ccaping

R J

L UIi.2

P late 1 8

-

D enticulates

UA k