Oxide Minerals 9781501508561, 9780939950034

Citation preview

REVIEWS in MINERALOGY (Formerly: "Short Course Notes")

Volume 3

OXIDE MINERALS DOUGLAS RUMBLE, III, Editor

The Authors: Ahmed El Goresy Max-Planck-Institut fur Keruphysik Heidelberg, West Germany

Stephen E. Haggerty

Department of Geology University of Massachusetts Amherst, Massachusetts 01003

J. Stephen Huebner 959 National Center United States Geological Survey Reston, Virginia 22092

Donald H. Lindsley Department of Earth and Space Sciences State University of New York Stony Brook, New York 11794

Douglas Rumble, III Geophysical Laboratory 2801 Upton St., N.W. Washington, D.C. 20008

Series Editor: Paul H. Ribbe Department of Geological Sciences Virginia Polytechnic Institute and State University Blacksburg, Virginia 24061

MINERALOGICAL SOCIETY OF AMERICA

COPYRIGHT 1976 Reserved by the authors (Second printing, 1981)

PRINTED BY BookCrafters, Inc. Chelsea, Michigan 48118

REVIEWS IN MINERALOGY (Formerly:

SHORT COURSE NOTES)

ISSN 0275-0279 Volume 3:

OXIDE MINERALS

ISBN 0-939950-03-0

Additional copies of this volume as well as those listed below may be obtained at moderate cost from Mineralogical Society of America 2000 Florida Avenue, NW Washington, D.C. 20009 No. of Pages

Vol. 1

SULFIDE MINERALOGY, P.H. Ribbe, Editor (1974)

2

FELDSPAR MINERALOGY, P.H. Ribbe, Editor (1975; revised 1981)

284 ^350

3

OXIDE MINERALS, Douglas Rumble III, Editor (1976)

4

MINERALOGY and GEOLOGY of NATURAL ZEOLITES,

5

0RTH0SILICATES, P.H. Ribbe, Editor (1980)

381

6

MARINE MINERALS, R.G. B u m s , Editor (1979)

380

F.A. Mumpton, Editor (1977)

502

232

7

PYROXENES, C.T. Prewitt, Editor (1980)

525

8

KINETICS of GEOCHEMICAL PROCESSES, A.C. Lasaga and R.J. Kirkpatrick, Editors (1981)

391

9A

AMPHIBOLES and Other Hydrous Pyriboles - Mineralogy, D.R. Veblen, Editor (1981)

372

9B

AMPHIBOLES: Petrology and Experimental Phase Relations, D.R. Veblen and P.H. Ribbe, Editors (1981) (ii)

V375

FOREWORD Oxide Minerals was first printed in 1976 as Volume 3 of the Mineralogical Society of America's "SHORT COURSE NOTES."

That series was renamed "REVIEWS

IN MINERALOGY" in 1980, and for that reason this, the second printing of Oxide Minerals, has been reissued under the new banner.

Only minor corrections have

been made in this printing. Paul H. Ribbe Series Editor Blacksburg, VA October 1981

EDITOR'S PREFACE and ACKNOWLEDGEMENTS The purpose of this volume is to provide, in a rapidly-printed, inexpensive format, an up-to-date review of the mineralogy and petrology of rock-forming opaque oxide minerals. It was the textbook for the short course on rock-forming oxide minerals sponsored by the Mineralogical Society of America at the Colorado School of Mines, November 5-7, 1976.

The contributors hope that the work will

be valuable not only to participants in the short course, but also to others desiring a modern review of the subject. The editor is grateful to Don Bloss, Paul Ribbe, and Southern Printing Co. for invaluable assistance in preparing the notes for publication.

Especial

thanks are due Mrs. Margie Strickler for her outstanding work in typing the entire text.

Elsevier Scientific Publishing Co. granted permission to reprint

figures from their journal Earth and Planetary Sciences Letters. J.J. Finney, George Fisher, J.F. Hays, Jim Munoz, Paul Ribbe, S.B. Romberger, and E-An Zen contributed vital advice and assistance in organizing the short course.

E. Leitz, Inc., C. Reichert (American Optical Corp.), Vickers Instruments,

Inc., and Carl Zeiss, Inc., provided demonstration models of their microscopes for use in the ore microscopy workshop. Finally, I wish to acknowledge the assistance provided by the resources of the Carnegie Institution of Washington through the good offices of H.S. Yoder, Jr., Director of the Geophysical Laboratory, in organizing the short course and publishing this volume. Douglas Rumble, III Washington, D.C. November 1976

(iii)

TABLE OF CONTENTS FOREWORD and EDITOR'S PREFACE AND ACKNOWLEDGMENTS

(iii)

USEFUL REFERENCES

(ix)

Chapter 1.

The CRYSTAL CHEMISTRY and STRUCTURE of OXIDE MINERALS as EXEMPLIFIED by the Fe-Ti OXIDES

INTRODUCTION

Donald H. Lindsley L- 1

Techniques

L- 1

Magnetic properties

L- 2

THE CUBIC OXIDE MINERALS

L- 4

Monoxides (Space group Fm3m) Spinel group Magnetite (Fes04> UlvSspinel (FegTiOg) Magnetite-ulv'ôspinel solid solutions Maghemite (y-Fe20s) Magnetite-maghemite solid solutions Titanomaghemites THE RHOMBOHEDRAL OXIDES

L- 5 L- 7 L-12 L-15 L-15 L-18 L-22 L-24 L-31

Hematite Magnetic structure of hematite Curiej Nêel, and Morin temperatures of hematite Ilmenite Crystal structure of ilmenite Magnetic structure of ilmenite Crystal and magnetic structure of hematite-ilmenite solid solutions ORTHORHOMBIC OXIDES—THE PSEUDOBROOKITE GROUP

L-34 L-36 L-37 L-38 L-38 L-40 L-41 L-44

The structure of pseudobrookite (Fe2Ti05)

L-45

The Fe2Ti05-FeTi205 series

L-45

STRUCTURES OF RUTILE, ANATASE, AND BROOKITE

L-47

REFERENCES

L-52

Chapter 2.

EXPERIMENTAL STUDIES of OXIDE MINERALS

INTRODUCTION

Donald H. Lindsley L-61

Control of experimental conditions Oxygen fugacity Container problems Experiments at very high pressures or high temperatures Minerals and phases considered

L-61 L-62 L-64 L-64 L-65

FE-TI-0 SYSTEM L_66

Fe-0 join Wustite Fe20s in magnetite

L 67 ~ L-67 L 67 ~ L_68

Fe0-Ti02 join Fe2Û3-Ti02 join

u

(iv)

Fe0-Fe203-Ti02(-Ti203) join 1300°C isotherm Ti-Maghemite Reduction of Fe-Ti oxides

L-69 L-69 L-75 L-75

FE-0-MG0-TI02 SYSTEM

L-79

Fe0-Fe203-Mg0 join Fe0-Mg0-Ti02(-Ti203) join Fe0-Fe203-Mg0-Ti02 join

L-80 L-80 L-81

FE0-FE 2 03-AL 2 03 SYSTEM

L-81

CR203-BEARING SYSTEMS

L-82

Fe0-Fe203-Cr203 system MgAl20^-Mg2Ti0^-MgCr20^ system FeCr204-Fe304-FeAl20^ system

L-82 L-83 L-84

REFERENCES

L-84

Chapter 3. OXIDE MINERALS in METAMORPHIC ROCKS

Douglas Rumble, III

INTRODUCTION

R- 1

MINERALOGY

R- 1

Spinel solid solutions Hematite-ilmenite solid solutions Pseudobrookite solid solutions Rutile and polymorphs

RRRR-

OXIDE MINERALS IN RELATION TO METAMORPHIC MINERAL ZONES

R- 7

Metamorphic zones of low to intermediate pressure ChXorite and biotite zones Garnet and staurolite zones Sillimanite zone Sillimanite-potash feldspar zone Chromian spinel composition in relation to metamorphic zones . High-temperature contact metamorphism High-pressure metamorphism PHASE PETROLOGY

1 3 6 6

R- 9 R- 9 R- 9 R-10 R-10 R-10 R-ll R-ll R-ll

Element partitioning Oxide mineral equilibria Oxide-silicate mineral equilibria Deduction of conditions of metamorphism

R-ll R-12 R-15 R-16

OXIDATION-REDUCTION PROCESSES IN METAMORPHISM

R-19

REFERENCES

R-20

Chapter 4.

OXIDATION of OPAQUE MINERAL OXIDES in BASALTS

Stephen E. Haggerty

INTRODUCTION

Hg- 1

OXIDATION PARAGENESIS

Hg- 3

Primary oxide mineralogy Oxidation of ulvbspinel-magnetite solid solutions Trellis type Composite types (v)

HgHgHgHg-

3 4 4 8

Sanduich type Oxidation of titanomagnetite-ilmenite intergrowths C4 stage C5 stage C6 stage C7 stage Oxidation of discrete primary ilmenite Ilmenite oxidation classification

Hg-16 Hg-16 Hg-17 Hg-20 Hg-21 Hg-24 Hg-28 Hg-28

PHASE CHEMISTRY OF OXIDATION ASSEMBLAGES

Hg-37

Introduction Titanomagnetite-ilmenite assemblages Ferrian ilmenite and ferrian rutile Titanohematite, rutile, and magnetite Pseudobrookite, titanohematite, magnetite and Al-magnesioferrite OXIDE SYSTEMATICS AND PHASE COMPATIBILITY RELATIONSHIPS

Hg-37 Hg-46 Hg-49 Hg-50 Hg-52 Hg-55

Introduction Ilmenite systematics Titanomagnetite systematics Applications

Hg-55 Hg-61 Hg-61 Hg-69

ASSOCIATED MINERAL OXIDATION

Hg-74

Introduction Chromian spinel oxidation Olivine oxidation

Hg-74 Hg-74 Hg-75

OXIDE DISTRIBUTIONS IN BASALT PROFILES

Hg-78

Introduction Mean oxidation numbers Oxide distributions Mechanism of oxidation

Hg-78 Hg-79 Hg-82 Hg-94

IMPLICATIONS FOR ROCK MAGNETISM AND OXIDE PETROGENESIS

Hg-95

Introduction Oxidation Single cooling units

Hg-95 Hg-96 Hg-98

SUMMARY AND CONCLUSIONS

Hg-98

Chapter 5.

OXIDE MINERALS in LUNAR ROCKS

Ahmed El Goresy

INTRODUCTION

EG- 1

MINERALOGY

EG- 1

Chromite-ulv8spinel series Ilmenite-geikielite series Armalcolite-anosovite series Rutile

EGEGEGEG-

OXIDE RELATIONS IN DIFFERENT ROCK TYPES RECOVERED FROM THE MOON Opaque oxides in Ti02~poor basalts Spinels Cationic relationships and substitutions Cr-Al substitutional trends V-Cr and V-Al substitutional trends

(vi)

. .

1 3 4 4

EG- 4 EG- 4 EG- 5 EG-14 EG-14 EG-17

Fe-Mg substitutional trends Ti-(V+Cr+Al) substitutions Opaque oxides in Ti02-rich basalts Armaloolite relationships Origin of ilmenite rims around armaloolite in olivine porphyritio basalts Chemistry of armaloolite Chromian ulv'dspinel Ilmenite Rutile Opaque oxides in anorthositic rocks and highland breccias . . . . Spinels Ilmenite Rutile SUBSOLIDUS REACTIONS

EG-25 EG-30 EG-34 EG-35 EG-3 6 EG-36 EG-3 7 EG-37 EG-37 EG-38

Subsolidus reduction reactions in the lunar rocks Nature of reducing agent in Apollo 17 basalts

EG-39 EG-42

REFERENCES Chapter 6.

EG-19 EG-19 EG-24 EG-24

EG-43 OPAQUE OXIDE MINERALS in METEORITES

Ahmed El Goresy

INTRODUCTION

EG-47

MINERALOGY

EG-49

Spinel group minerals Ilmenite-geikielite-pyrophanite series

EG-49 EG—56

Rutile

EG-5 7

OXIDE ASSEMBLAGES IN VARIOUS METEORITE GROUPS

EG-57

Chondrites Spinels Ilmenite Achondrites Ilmenite Stony irons Iron meteorites REFERENCES

EG-57 EG-5 9 EG-6 2 EG-64 EG-6 5 EG-65 EG-67 EG-71

Chapter 7.

The MANGANESE OXIDES - A BIBLIOGRAPHIC COMMENTARY J. Stephen Huebner

INTRODUCTION

SH- 1

Crystal structures and chemistry Tetravalent oxides Trivalent oxides "Spinel-type" oxides WUstite-type oxide Geologic occurrences Thermochemistry and phase relations Petrology of manganese oxides REFERENCES

SH- 1 SH- 1 SH- 2 SH- 3 SH- 4 SH- 4 SH- 6 SH- 8 SH-11

(vii)

Chapter

8.

OPAQUE MINERAL OXIDES in TERRESTRIAL IGNEOUS ROCKS Stephen

E.

Haggerty

GENERAL INTRODUCTION

Hg-101

SYSTEMATIC MINERALOGY

Hg-101

Nomenclature Spinel series Ilmenite series Pseudobrookite series TiOg polymorphs Exsolution Subsolidus reactions Oxide assemblages and textures Two-dimensional mineral morphology Chromian spinelss Ilmenite-hematitess Reactions involving ilmenite-hematitess UlvOspinel-magnetitess Titanomagnetite-pleonaste^s Magnetite-chronrite-hercymtess Reactions involving magnetite-ulvSspinelss Oxides derived from silicates PRIMARY OXIDE DISTRIBUTIONS

Hg-101 Hg-101 Hg-104 Hg-104 Hg-104 Hg-104 Hg-108 Hg-108 Hg-108 Hg-109 Hg-118 Hg-119 Hg-123 Hg-128 Hg-129 Hg-129 Hg-135 Hg-140

Introduction Chromian spinel distributions Pseudobrookite distributions Ilmenite distributions Magnetite-ulvSspinel distributions Oxidites

Hg-140 Hg-142 Hg-150 Hg-152 Hg-158 Hg-158

T AND fo 2 VARIATIONS IN IGNEOUS ROCKS

Hg-160

Introduction Extrusive suites Intrusive suites Magmatic ore deposits Data summary Experimental determinations of T and fQ 2 Silica activity

Hg-160 Hg-160 Hg-160 Hg-167 Hg-167 Hg-169 Hg-169

TABLES OF MINERALOGICAL, PETROLOGICAL, AND CHEMICAL PROPERTIES OF OPAQUE MINERAL OXIDES IN IGNEOUS ROCKS

Hg-176

REFERENCES

Hg-277

(viii)

USEFUL REFERENCES General

Deer, W. A., R. A. Howie, and J. Zussman (1962) Rock Forming Minerals, Vol. 5, Non-sillcates. John Wiley, New York. Elsdon, R. (1975) Iron-titanium oxide minerals in igneous and metamorphic rocks. Minerals Sci. Eng. 7, 48-70. Freund, H. (Ed.) (1966) Applied Ore Microscopy, Theory and Technique. millan Co., New York. Ramdohr, P. (1969) The Ore Minerals and Their Intergrowths. Oxford.

Mac-

Pergamon Press,

Meteorites Bunch, T. E., K. Keil, and K. G. Snetsinger (1967) Chromite composition in relation to chemistry and texture of ordinary chrondites. Geochim. Cosmochim. Acta 31, 1569. Bunch, T. E. and K. Keil (1971) Chromite and ilmenite in nonchronditic meteorites. Am. Mineral. 56, 146-157. Mason, B. (1962) Meteorites.

John Wiley and Sons, New York.

Ramdohr, P. (1973) The Opaque Minerals in Stony Meteorites. Amsterdam. Wasson, J. T. (1974) Meteorites.

Springer-Verlag, New York.

Elsevier,

The CRYSTAL CHEMISTRY and STRUCTURE of OXIDE MINERALS as EXEMPLIFIED by the Fe-Ti OXIDES Donald H. Lindsle}:

Chapter I INTRODUCTION Crystal structures play a vital role in the interpretation of chemical reactions and magnetic properties of the oxide minerals.

For most purposes it

is useful to treat the oxide minerals as ionic crystals that consist of oxygen frameworks (nearly cubic or hexagonal close-packed) with cations occupying octahedral or tetrahedral interstices.

Both iron and titanium as well as

manganese are members of the first transition metal series; each can therefore exist in more than one valence state. 3d electrons in Ti to these ions.

, Fe

, Fe

, Mn

Furthermore, the existence of unpaired , and Mn

imparts net magnetic moments

Thus a complete characterization of the structure of an oxide

mineral must include the determination of valence states and magnetic orientation as well as the position of each atom in the unit cell.

In some cases it

is necessary to chose a magnetic unit cell that is a multiple of the crystallographic cell, or, alternatively, to view the magnetic cell as having lower symmetry than the crystallographic cell of the same size.

No one technique

can uniquely characterize the structure and chemistry of an iron-bearing oxide mineral, but a combination of methods has yielded detailed information on the most important structures.

Techniques The primary method of determining the structures is of course x-ray diffraction.

Single-crystal x-ray studies yield the (non-magnetic) symmetry of

the structure, the positions of the metal ions, and with lesser precision, the positions of the oxygen ions.

Thus in 1915, only three years after the

discovery of x-ray diffraction, Bragg and Nishikawa were able independently to determine the main details of the magnetite (spinel) structure. However, 2+ 3+ they were not able to assign Fe and Fe to specific sites in the structure (the assignment they assumed is incorrect), nor were they able to determine the oxygen positions with great precision.

Nevertheless, the basic structure

provided by x-ray diffraction makes possible the utilization of other

L-l

techniques, for these merely provide refinements of the structure rather than independent determinations of it. Neutron diffraction is an important technique in determining the structures of the oxide minerals; three characteristics make it especially useful. It can yield information on (1) electron spin orientation and thus on magnetic structures; (2) oxygen parameters, since oxygen has a relatively large scattering cross-section for neutrons; and (3) the distribution of iron and titanium in the structure.

(The electron densities of iron and titanium are sufficiently

similar that x-ray diffraction cannot always distinguish between them, whereas the scattering cross-section of iron for neutrons is three times that of titanium in magnitude and of the opposite sign.)

Most determinations of mag-

netic structure and precise oxygen parameters utilize neutron diffraction. Measurement of saturation magnetization at low temperature can test models of magnetic structures, and is particularly useful in studying solid solution series.

Electrical conductivity measurements have also been used to predict

magnetic symmetry and cation distribution; this technique has been superseded for the most part. Naturally occurring

in the iron-titanium oxides permits application

of the Mossbauer effect, which has been particularly useful in determining cation valencies, and, to a lesser extent, magnetic structure.

Determination

of the isomer shift for "^Fe in the sample relative to that in the source provides a quantitative (or at least semiquantitative) measure of the valence of the iron—and, by difference, that of other cations.

Magnetic properties The ensuing discussion of crystal structures must refer to several concepts of magnetism of the solid state.

These concepts are briefly reviewed

here; a more extensive review is given by Nagata (1961, p. 1-39) and by Stacey and Banerjee (1974). An electron in an atom (or ion) has a magnetic moment that results from its spin, from its orbital motion, or from a combination of both.

In iron and

titanium, the moment results mainly from spin rather than orbital motion.

The

moments of electrons with opposite spins but with otherwise identical quantum states are self-cancelling.

Atoms and ions in which all the electrons are thus

paired have no net magnetic moment and are called diamagnetic.

Atoms or ions

with one or more unpaired electrons, on the other hand, have a net magnetic moment and are termed paramagnetic.

The magnetic moment of an unpaired elec-20

tron is one Bohr magneton (y ) which equals 0.9274 x 10 B L-2

emu.

Most para-

magnetic particles have but one unpaired electron and therefore have moments of 1 yg.

However, in the first transition-metal series each of the five states

in the 3d level tends to be filled first by a single electron, and to become doubly occupied only after all five have been filled singly. as the high-spin condition.

(This is known

When as many 3d electrons as possible are paired,

the atom or ion is said to be in the low-spin condition.)

The spins of elec-

trons occupying the same state must be opposed, in accord with the Pauli ex3+ and Fe , each having five unpaired 3d electrons 2+ with parallel spins, have magnetic saturation moments of 5y_. Fe has six D 3d electrons,'two of which are paired and therefore self-cancelling, so the elusion principle.

Mn

2+

resultant saturation moment due to the four unpaired electrons is 4 yD. Fe° likewise has six 3d electrons (as well as two paired 4s electrons) and also has a saturation moment of

with one 3d electron has 1 y B , but Ti^ + with

no 3d electrons has no net moment. It might be noted here that the saturation moments differ slightly from the susceptibility moments of the same ions.

Susceptibility moments are some-

what larger, and tend not to be integral numbers of Bohr magnetons. If paramagnetic atoms or ions are incorporated into a crystal structure with their spins (and hence their magnetic moments) randomly oriented, the resulting solid has a low magnetic susceptibility and no net moment, and is termed paramagnetic. But in crystals of a-iron, for example, exchange interaction between neighboring iron atoms results in a parallel alignment of moments throughout each of several small volumes (domains) of the crystal. Such crystals that have high magnetic susceptibilities and can acquire remanent (permanent) magnetic moments, are termed ferromagnetic. The theory of ferromagnetism is reviewed by Bozorth (1951). Exchange interactions are strongly effective only between nearest-neighbor atoms.

In oxide structures, however, the nearest neighbors of the metals are

always oxygen ions, and exchange coupling cannot be invoked.

Kramers (1934)

proposed that the spins of next-nearest neighbor metal ions are coupled through the intermediate oxygens, a mechanism that he called superexchange. Anderson (1950) provided a firm theoretical basis for superexchange interactions, which now play a central role in most interpretations of the magnetic properties of oxides.

Superexchange becomes more effective as the metal-oxygen-metal angle

approaches 180°, and in general the interaction is negative:

the moments of

tyo paramagnetic ions coupled by superexchange through an oxygen are antiparallel and hence self-cancelling.

Crystals in which superexchange inter-

actions produce two equal magnetic substructures with opposite spin directions are termed antiferromagnetic (Hulthen, 1936). L-3

A perfect antiferromagnetic

crystal has no magnetic moment, but if impurities, defects, or cation deficiencies are concentrated in one substructure there may be a net moment, as was postulated, for example, by Neel (1949,1953) to explain the ferromagnetism" of hematite.

"parasitic

Similarly, if the spin directions of the magnetic

substructures are not precisely antiparallel, there can be a resultant moment perpendicular to the spin axes; this phenomenon is termed

"spin-canting."

For every antiferromagnetic substance there is a characteristic

temper-

a t u r e — the Neel p o i n t — a t and above which the antiferromagnetic structure is destroyed through thermal vibration and the m a t e r i a l becomes paramagnetic. Some substances, such as ilmenite, are paramagnetic at room temperature but become antiferromagnetic at lower temperatures.

Magnetic susceptibility

and

specific heat of an antiferromagnetic material usually reach a m a x i m u m at or near the Neel point. For some oxide structures superexchange interaction results in the formation of two magnetic substructures with opposite but unequal moments.

Such

structures have a net magnetic moment and were called ferrimaonetir;

by Neel

(1948), the term being derived from the ferrites

ferrimag-

netism is best displayed.

(spinels) in w h i c h

Ferrimagnetic substances, like ferromagnetic

ones,

each have a characteristic t e m p e r a t u r e — t h e Curie p o i n t — a t and above which they become paramagnetic; there is a close analogy between the Curie point and the Neel point of antiferromagnetic materials.

Many m a c r o s c o p i c properties

of ferrimagnetism are closely similar to those of ferromagnetism, and the former is sometimes conveniently if inelegantly considered a subclass of the latter. Thus it is common to speak of the "ferromagnetic properties" of a rock or mineral, even though those properties are almost certainly due to ferrimagnetism.

Both ferrimagnetism and antiferromagnetism w i t h parasitic

ferromag-

netism result from two opposite but unequal magnetic substructures, but in ferrimagnetism the inequality is a result of the crystal structure and thus would exist even in a perfect crystal.

THE CUBIC OXIDE MINERALS

Many oxide minerals have a cubic structure; others are perhaps best viewed as slight distortions or modifications of isometric structures.

In

general the oxygens approach cubic close packing.

frame-

A cubic-close-packed

w o r k of 32 oxygens contains a total of 64 tetrahedral sites and 32 octahedral sites.

If only octahedral sites are occupied by divalent cations, as in peri-

clase (MgO), the crystal has the NaCl structure

(Fig. L-l). In a very

group of o x i d e s — t h e s p i n e l s — b o t h octahedral sites and tetrahedral L-4

important

sites

^ ^

Oxygen

O

Octahedral cations

•X-

Origin

Figure L-l. A view of the monoxide (periclase, manganosite, ideal wiistite) structure, emphasizing the alternating (111) planes of oxygens and cations. Eight unit cells are shown to facilitate comparison with the spinel structure (Fig. L-3).

are occupied.

Not all of both sets of sites can be filled, however, for this

would require face-sharing between octahedra and t e t r a h e d r a — w h i c h is energetically unstable because of electrostatic repulsion of the cations.

The

filling of 8 tetrahedral sites precludes occupancy of more than 16 octahedral sites; conversely, the filling of 16 octahedral sites permits occupancy of only 8 of the 64 possible tetrahedral sites, if the symmetry of the space group is maintained.

Monoxides (Space group

Fm3m)

Like periclase, manganosite (MnO) has the NaCl structure, with the divalent cations occupying the octahedral interstices.

Viewed along the [111]

axis, the structure consists of alternating (111) layers of oxygen and metal (Fig. L-l). L-5

—

X

Figure L-2. Unit cell dimension of wustite as a function of composition.

F e / O (atomic) = I - X

Although rare as a mineral (but see, for example, Walenta, 1960), the phase wustite (nominally FeO) is of considerable interest. troversy raged over whether stoichiometric FeO exists.

For years a con-

There is now general

agreement that stoichiometric FeO cannot exist as a stable phase at low pressures and that it is always cation-deficient. 2+ the replacement of 3x Fe 2+ 3+

Charge balance is maintained by

3+ by 2x Fe

F e ^ _ ^ x F e 2 x 0 , or, more simply, Fe^

, so that the formula may be written 0.

The most iron-rich wustite stable at

one atmosphere is about Fe^ 954^ (Darken and Gurry, 1945). a range from 4.309 A for F e Q 949°

t0

4.292 X for Fe^

depending on the thermal history of the samples

^0,

Lattice

constants

the exact values

(Foster and Welch, 1956) ; see

Fig. L-2. Katsura et at. sures above 36 kb.

(1967) reported synthesis of stoichiometric FeO at presThe value a - 4.323 + 0.001 A determined by them is very

close to that predicted for pure FeO by extrapolation

(4.326 1) .

Roth (1960) has made a detailed study of the defect structure of wustite by neutron diffraction.

The diffraction patterns not only confirm the presence

of vacancies in the octahedral iron positions, but indicate more such vacancies than are required by the chemical analysis.

For example, the neutron diffrac-

tion data indicate 0.08 vacancies per octahedral iron site for the composition Fe n / 0 . Evidently 0.02 iron ions per formula unit must be accommodated in u. y4 tetrahedral sites interstitial to the close-packed oxygen framework.

Cubic

symmetry is maintained if the unit cell is doubled; the resulting space group is Fd3m.

The symmetry and oxygen content of such a cell are identical to that L-6

of magnetite, and, indeed, Roth visualized the local structure in the vicinity of octahedral vacancies as essentially that of magnetite. Roth's model has been generally confirmed by single crystal x-ray diffraction studies (Koch et at. , 1966; Koch and Fine, 1967, Koch and Cohen, 1969) which suggested the presence of ordered complexes of octahedral vacancies and tetraShirane et at. (1962) have found Mossbauer evidence that the 2+ symmetry around the Fe irons is less than cubic; they infer that the

hedral cations. local

lower symmetry is due to the vacancies.

Some studies have shown that quenched

wiistites may have one of three slightly different structures (Vallet and Raccah, 1965; Kleman, 1965; Carel, 1967).

Swaroof and Wagner (1967) could find no evi-

dence of phase changes in the wiistite field at 950-1250°C; however, high-temperature x-ray diffraction studies by Manenc (1968) show that at least one of the ordered structures persists to 1000°C for Fe-poor compositions.

Manenc sug-

gests that at temperatures within the wiistite stability field, Fe-rich wiistites have the true NaCl structure (with the vacancies random) whereas Fe poor wiistites have an ordered structure. Spinel group A large number of oxide minerals, some sulfides, and many artificial substances crystallize with the spinel structure, which is extraordinarily flexible in terms of the cations it will accept.

At least 30 different elements, with

valences ranging from +1 to +6, can serve as cations in oxide spinels.

Some

geologically important spinel end members are listed in Table L-l. Of the irontitanium oxides, the magnetite-ulvospinel solid solution series has the spinel structure, as do the cation-deficient oxides—at least as a first approximation— maghemite and titanomaghemite.

The unit cell is face-centered, cubic, and in

oxide spinels contains 32 oxygens, which form a nearly cubic-close-packed framework as viewed along the cube diagonals ([111]), the space group being Fcfim. The cations occupy interstices within the oxygen framework (Fig. L-3). In space group Fd3m there are two alternative sets of compatible tetrahedral and octahedral sites:

16d (octahedral) and 8a (tetrahedral) or 16e and 8b (International

Tables, 1952, p. 340-341).

These sets, although mutually exclusive, are iden-

tical in that a translation of the origin by j, j, | will bring one set into coincidence with the other. the spinels.

By convention, the set 16d and 8a is chosen for

There is considerable variation in the literature regarding the

notation used to describe these sites.

Most workers have employed the Wyckoff

(1922) notation, others that of the International Tables—which is followed here—and still others a notation based on the concept of magnetic substructures

L-7

\

a0 Oxygen Q

Octahedral cations

Q

Tetrahedral

•X -

Origin

cations

Figure L-3. The spinel unit cell, oriented so as to emphasize the (111) planes. Atoms are not drawn to scale; the circles simply represent the centers of atoms.

The origin in this diagram lies at the center of symmetry, as recom-

mended by the International Tables (1952); it differs by origin used in much of the literature.

from the

The arrow at the top indicates the

cation and oxygen layers shown in Figure L-5.

L-8

Table L - l . Some s p i n e l end members.

Mineral Name

Cell Edge a, in A

Formula

Oxygen Parameter u

Structure

Magnetite

Fe3+[Fe2+Fe3+]04

8.396

0.2548

I

Magnesioferrite

Fe3+[Mg2+Fe3+]04

8.383

0.257

I

3+

2+

3+

Jacobsite

Fe

8.51

I

Chromi te

Fe2+[Cr3+]04

8.378

N

Magnesiochromite

8.334

0.260

N

Spinel

Mg2+[Cr3+]0. 2+ 3+ Mg Alp 0 4

8.103

0.262

7/8 I

Hercynite

Fe2+[A?3+]04

8.135

Ulvöspinel

Fe

2+

[Mn

Fe

]04

[Fe2+Ti4+]04

8.536

N I

0.261

N, "normal" cation distribution, XIY^JO^, where [ J indicate octahedral cations.

I, "inverse" distribution, YfXYJO^. Many spinels

are probably intermediate between these extremes.

Data from Burns

(1970, p. 110).

A and B.

For the convenience of those who wish to pursue the original litera-

ture, TableL-2 correlates these various notations. Table L-2. Nomenclature o f c a t i o n s i t e s in s p i n e l s .

Tetrahedral

Octahedral

Example of Usage

f

c

Wyckoff (1922, 1965)

A

B

N6el

a

d

I n t e r n a t i o n a l Tables

(1948) (1952)

The sites occupied by cations in the spinel unit cell are special sites; that is, they lie at the intersections of symmetry elements.

These sites are

therefore fixed, and their coordinates are given by rational fractions of the cell edge, a.

Thus the coordinates x, y, z of one tetrahedral (a) site are

^ a, ^ a, g a (abbreviated

g).

The coordinates of all the sites are

given in Table L-3, based on an origin at the center of symmetry, rather than the more usual origin at - g ,

-3.

The 32 oxygens, on the other hand,

occupy sites whose exact coordinates must be determined experimentally.

Only

one parameter, conventionally called u, needs to be determined; the oxygen

L-9

Table L-3. Coordinates of anion and cation sites in spinels. Space group Fd3m. Origin at center (3m).

Type of Site

No. of Sites

Anion

Notation

32

Note:

1 1 1 1 u,u,u; u.^-u, ^ - u ; ^-u.u.^-u;

3m

Tetrahedral Cation

Octahedral Cation

16

Because of the face-centered

Coordi nates

Symmetry

- - - 3 u , u , u ; u , ^ + u, j + u;

1

1

|+

u, G , | + u; J +

u, | +

43m

1 I I. 7 7 7 o> O' Q' Q> Q» Q

3m

111. Ill- 111. I l l 2'2'2' 2'4'4' 4'2'4' 4'4'2

lattice, there are four points

u, G

equivalent

to the origin: 0,0,0; 0,—, —; ' 2'2

—,0,~; 2' ' 2

—,—,0 2'2'

The full sets of sites are generated by applying the coordinates table to each of the equivalent Reference:

International

Tables

in the

points. (1952, p. 341).

coordinates are then derived from symmetry considerations.

The oxygens lie in

32 e sites whose coordinates in terms of u are given in Table L-3. If the oxygen parameter u happened to equal jj, the oxygens would occupy special sites, which correspond to rigorous cubic close-packing.*

Inasmuch as u does not deviate

greatly from ¡j, it is a reasonable and useful approximation to view the oxygens as close-packed.

Variations in u correspond to displacements of oxygens along

cube diagonals, and reflect adjustments to the relative effective radii of cations in the tetrahedral and octahedral sites.

An increase in u corresponds

to an enlargement of the tetrahedral coordination polyhedra and a compensating diminution in the octahedra (see Fig. L-A'). In addition, the entire oxygen framework (i.e., the unit cell) can expand or contract to accommodate cations of larger or smaller average effective radius.

It is this flexibility of the

*Many workers choose an origin for the spinel structure at 43m, which is ( - 5 , - ¿ , - j j ) removed from the origin at the center (3m) adopted here. The values for u listed here should be increased by 0.125 to compare them with most values in the literature. L-10

Figure L-4.

Details along a body diagonal of the spinel unit cell in Fig. L - 3 ,

illustrating how changes in the oxygen parameter u change the relative sizes of

oxygen framework that permits such a large number of elements to occur as important cations in oxide spinels. X

The general chemical formula for ideal spinels is X Y 2 ° 4 unit cell), where X and Y are cations of different valence. 2j Fe and Y = Fe ; in ulvospinel, on the other hand, X = Ti

8Y16°32

p6r

In magnetite X = 21 and Y = Fe

The original determinations of the spinel structure (Bragg, 1915a,b; Nishikawa, 2+ 3+ 1915) were made on magnetite and on spinel (X=Mg , Y=A1 ); it was impossible to distinguish between the X and Y cations on the basis of x-ray intensities 2+ 3+ because of the similar scattering powers of Mg and A l and of Fe

and Fe

For simplicity it was assumed that the eight X cations occupy the 8a sites and the sixteen Y cations occupy the 16d sites, giving a structural formula: X [ Y 2 ] 0 4 , where the brackets denote cations in octahedral

(16c?) sites.

Barth

and Posnjak (1931,1932) discovered by careful x-ray intensity measurements that this "normal" structure was correct for several aluminate spinels, but not for many other spinels.

The latter have 8 of the 16 Y cations in the 8a

sites, yielding the structural formula Y[YX]0^.

Verwey and H e i l m a n n

(1947)

termed these spinels "inversed," and at present they are known as "inverse spinels."

It is now known that most spinels are intermediate between these L-ll

two extremes.

Table L-l indicates which end member spinels are normal and which

are inverse. Verwey and Heilmann (1947) also gave empirical rules regarding the distribution of cations in the tetrahedral and octahedral sites of the spinel structure.

Electrostatic and ionic radius considerations alone are insufficient to

explain the observed distributions.

Goodenough and Loeb (1955) showed that 3 3+ bonds—such as Fe —are favored

cations having a tendency to form hybrid sp in tetrahedral sites.

There have been suggestions that ions thus bound in

tetrahedral sites are relatively immobile (Frolich and Stiller, 1963; O'Reilly and Banerjee, 1966), but self-diffusion experiments indicate otherwise (Lindner and Akerstrom, 1956). Magnetite

(Fe^O^).

Both Bragg (1915a,b) and Nishikawa (1915) determined

the magnetite structure using Laue x-ray photographs obtained from single 2+ 3+ crystals. They assumed that the structural formula was Fe [Fe2 and that the oxygen parameter u was 0.25, although they recognized that it would vary with cation composition. 0.001.

Claassen (1926) refined the value of u to 0.254 +

Hamilton (1958) further refined u to 0.2548 + 0.0002 at 23°C from

neutron diffraction data. A wide range of values for the unit-cell parameter (a) of magnetite has been reported.

Several reasons for this variation exist.

Many of the magne-

tites were not adequately characterized chemically, and2+the discrepancies 3+ probably reflect the presence of cations other than Fe and Fe , or of cation vacancies.

Precise x-ray work on natural and synthetic magnetites of known

composition has yielded values of a ranging from to 8.3963 X. 2+ 8.3933+ The similarity in scattering power of Fe and Fe made it impossible for the early workers to determine the structural formula of magnetite by x-ray methods.

Verwey and deBoer (1936) used measurements of electrical 3+ conductivity 2+ 3+ to demonstrate that magnetite has the inverse spinel structure, Fe [Fe Fe Neel (1948) predicted that the irons in the 8a and the 16d sites form two magnetic substructures with antiparallel moments along [111]. If magnetite had 3+ the normal structure, Fe2+ [Fe„ 2 ]0,, h there would be 8x4 u„ B in the 8a sites and 8x(5+5) Pg oppositely directed in the 16d sites, for a net moment of 48 yg per Oi per formula unit]. In the case of inverse structure, Fe

unit cell [=6 y 2+ 3+ [Fe Fe

the net moment would be 8x(4+5)-8x5 = 32

per formula unit].

per unit cell [=4

The empirical value (extrapolated to 0°K) , of 4.07

(Neel, 1948, p. 179) per formula unit thus supports the Neel model regarding both the inverse structure and the ferrimagnetic structure. the excess moment (.07

Neel assumed that

per formula unit) reflects an orbital contribution L-12

to the moment of Fe

; an alternative explanation is that the structure is not

completely inverse.

Final confirmation of the model was provided by the neutron

diffraction experiments of Shull et at. (1951b). For many purposes it is useful to view the magnetite structure along the [111] axis (Fig. L-5).

It is along this axis that the (nearly) cubic close-

packing of the oxygens is observed and that the ferrimagnetic moment tends to be directed.

In the wiistite structure normal to [111], planes of oxygens al-

ternate with planes of iron ions.

In magnetite the sequence is somewhat more

complex because tetrahedral and octahedral cations do not lie in the same VI IV VI IV VI normal to [111] the sequence is O-Fe -O-Fe -Fe -Fe -O-Fe ...,

planes:

where the superscripts refer to coordination numbers.

The distance along [111]

between F e ^ layers and the adjacent Fe'"*" layers is small (0.675 A , or 0.08 a),

Figure L-5.

Diagram, approximately to scale, showing an oxygen layer of a

spinel and the cation layers on either side of it, projected onto a (111) plane. The large open circles are oxygens. are drawn for magnetite.

The ionic radii (Shannon and Prewitt, 1969)

Decimal fractions show height above the lower octa-

hedral layer, expressed as a portion of the body diagonal of the unit cell (a/3). Note that the oxygen plane is puckered slightly; for ideal cubicclose-packing, the height of the oxygens would be 0.144 (=

L-13

.

and for some purposes it is a good approximation to consider the structure as VI successive planes of O-Fe

IV -0-(Fe

VI Fe

IV Fe

VI )-0-Fe

....

The structure of mag-

netite as viewed in this way bears many similarities to that of hematite, and the (111) planes play an important role in the textural relations between these two minerals. The Curie temperature of magnetite corresponds to a transition from longrange ferrimagnetic ordering at lower temperatures to disorder at higher temperatures.

Values of the Curie temperature ranging from 570° to 581°C have

been reported.

The transition is accompanied by m a x i m a in the coefficient of

thermal expansion, the specific heat and the neutron scattering

cross-section.

At low temperatures magnetite undergoes another transition w h i c h involves a decrease in crystallographic symmetry, electrical conductivity, and magnetic properties.

Probably the best value of the transition temperature is 119.4°K,

obtained for stoichiometric, synthetic single crystals

Figure L-6.

Variation of cell dimension w i t h composition in the

xFe2TiO^ series. x = 0.8.

(Calhoun, 1954).

(l-x^e^O^-

There appear to be changes in slope at or near x = 0.2 and

Height of symbols shows uncertainties

L-14

(Lindsley, 1965).

Ulvospinel (Fe^TiO^).

Pure synthetic ulvospinel was shown to have the

inverse spinel structure (Fe[FeTi]04), with u = 0.265 + 0.01, by Barth and Posnjak (1932).

The cell dimension a is 8.536 + 0.001 (Lindsley, 1965;

Akimoto and Syono, 1967).

Forster and Hall (1965) determined u to be 0.261

+ 0.001 by neutron diffraction.

The sequence of planes normal to [111] for an

VI

ideal Fe[FeTi]0^ is 0-(Fe,Ti) -0-Fe IV -(Fe,Ti) VI -Fe IV -0. of the iron and titanium are evidently F e ^ t T i ^ F e ^ J O ^

The valence states (Gorter, 1957 , p. 345;

Verwey and Heilmann, 1947; Rossiter and Clark, 1965). The ideal inverse struc2+ 2+ ture has 8 Fe in 8a with spins antiparallel to 8 Fe in 16d and is antiferromagnetic with a net moment of 0

. D

Magnetite-ulvospinel solid solutions.

Solid solutions (l-x)Fe20^-x(Fe2

TiO^) between magnetite and ulvospinel have the spinel structure.

Variations

of the unit-cell parameter a with composition for synthetic solid solutions are presented in Figure L-6.

Syono (1965, p. 136) reported that neutron dif-

fraction measurements gave u = 0.261 + 0.001 for x = 0.99 and u = 0.255 + 0.001 for x = 0.56. Several different structural formulae for the solid solution have been 4+ 3+ proposed. The Akimoto model assumes that Ti always replaces the Fe in 2+ 3+ 16d in magnetite whereas Fe always replaces the Fe in 8a; the general formula would be Fe? + Fe^ + [Fe^ + Fe? + Ti^+]0,, yielding a predicted saturation 1-x x 1-x x 4 moment of 4(l-x) p B (Akimoto, 1954); see Figs. L-7 and L-8. Neel (1955, p. 196) and Chevallier et al. (1955) assumed that all Ti^ + enters the d sites, 2+ 3+ 3+ and that Fe first replaces Fe in d sites and only then the Fe in a sites.

This model, which is in accord with the empirical rules of Verwey and

Heilmann (1947) as well as the theoretical considerations of Goodenough and Loeb (1955), yields the structural formulae: for 0 < x < 1/2 (li-1) 4

for 1/2 £ x £ 1.

The corresponding saturation moments are (4-6x) y,. for 0 < x < 1/2 and (2-2x) Pg for 1/2 _< x £ 1.

The observed moments for many synthetic solid solutions

(Fig. L-7) fall about half-way between the values predicted for these two models (Akimoto et al. , 1957a, p. 173; Akimoto, 1962).

A model that fits

the magnetic data has been proposed by O'Reilly and Banerjee (1965; Banerjee and O'Reilly, 1966).

Their model is identical with the Neel-Chevallier-

Bolfa-Mathieu model over the composition ranges 0 £ x £ 0.2 and 0.8 2) -

it is always metastable with respect to

Its distinguishing characteristics are a composition close

to i ^ O ^ combined with high magnetic susceptibility and remanence.

Much work

on maghemite has been spurred by the discovery of its usefulness in making magnetic tapes. Robbins (1859) showed that a "magnetic peroxide of iron," prepared by the oxidation of magnetite, had the composition Fe203 and thus was a polymorph of hematite.

Hagg (1935a,b), Verwey (1935a,b), and Kordes (1935) independently

proposed that maghemite has an iron-deficient spinel structure with a unit-cell 3+ formula Fe^-^ 33332"

distribution

or both—has remained controversial.

the vacancies—octahedral, tetrahedral, Hagg interpreted his x-ray intensity data

(1935b) as indicating that the vacancies are randomly distributed in the octahedral 16d sites, yielding a structural formula Fe [Fe^ 6 7 • Q 3 3 ] ° 4 and a unit-cell content Fe„ + [Fe?i .,,[ |„ ,_]0__, where I I indicates a cation o 1 3 . 3 3 — ¿ . 6 / 3z — vacancy.

Verwey's interpretation of his intensity data was somewhat equivocal:

In three papers (Verwey, 1935a,b; Verwey and vanBruggen, 1935) he states that the vacancies are distributed over all 24 cation positions (16 r- 4J '—CQ O CD O O i— 3 O i— •H e -a -a -c: ^ ^ ^ cu O O O to "O ^ C C u r O f O N 3 OO C_3 -C CT) E= 3 in id io i/i "ï «t in S- r— IO CO CTI CTI CT) *U •OTJ 0) * CTÌ a) aj-oi— i—i— c C C •!-> sA! i— in co c '— id ID ID to CO >, ai 3 ID o lt> _q -a •"H CO IO H) O O C C C "O >)S-0 CT) cn-c -c: i— o o 3 -r- a; id i—^ s£ 0)JÏ3wluinJ> > -C c ^O -C 30a) QÍ OO M b 0 13 .c+J .c • +> +J-C +J •!-> +-> +-> -I-) 9 C4->4->CC4->.CC C C C C C co >>iD(0>,>,ia>>>) cu >> =») I/lZZÜll/lZI/lW OO 00 00 00 oo M Di S 0 Ih ai £ S•a cu co oo co •p ^ u 3- >a•— SC\J E CL Û_ CL P 6 +-> m co to m si 1— 1— 1— q CL sa) 0 L• ia _ CL — a_ o_ 0 CL cu oo M u — m

ID E

lo i- >> t_> •(->

>>

+>1 J E •i— IO 4-> laj u c rj OCt HJ

o o u o u o o o JDJDJDJDJ2JD-QJD CJIt-J0C_)C_>l_)00

O O O +1

•h to CTI CTIOl CTI cu u io m IO IO •e: •i- S- S- S— Í. 4J -O +J +-> -u q 3 ai Qj ai ai (U U h h 1— 1— tH IO

C\J o o +1

CTi '—* oo co c\j r^ coC\J cocNJ«3-co^-cooo^i- in inj oo oo cocococooooocooo oo oo oo co oo CO 03 CO CO CD OO CO COco co co 00 00

L-21

•H CO ai M io t>

cw «

u -tu 0

The instability of maghemite at high temperatures makes it impossible to determine the Curie point directly.

Two different values that differ by nearly

100°C are presently in use; each has been determined by extrapolation.

Michel

and Chaudron (1935) determined the value 675 + 10°C by preparing a series of Na-stabilized "maghemites" and extrapolating their Curie temperatures to pure y-Fe-O-.

This value is accepted by most workers.

However, in 1949, Maxwell 2 + 3 +

et al. observed that "certain unreduced iron oxide catalysts" with Fe

/Fe

varying from 0.352 to 1.276 all had Curie temperatures equal, within a few degrees, to that of magnetite, 583°.

From this they concluded that the Curie

temperature of maghemite is likewise approximately 583°.

This conclusion may

be correct, but based on the information published (which is in an abstract only), it should be considered unproved.

Aharoni et al. (1962) claim to have

measured directly a Curie point of 590°C for maghemite; but as their major means of identifying the sample as maghemite is by the Curie temperature, their results must be viewed with caution.

Frolich and Vollstaedt (1967) have

calculated a Curie temperature for maghemite of 675°C, based on a measured value 650°C for LiFe_0 D oo .

The best conclusion at the present time is that none

of the methods for determining the Curie temperature of maghemite is completely satisfactory, and as a result no published value can be considered definitive. However, the 675° value seems tentatively more acceptable. Magnetite-maghemite intermediate between

solid solutions.

Hagg (1935b) prepared three spinels

and y - i ^ O ^ by partial oxidation of magnetite; the

relation of their compositions to cell parameters is shown in Figure L-9.

On

this basis Hagg proposed that there can exist a continuous solid solution series between magnetite and maghemite—which corresponds to a continuous increase from 0 to 2.67 vacancies per unit cell. Several workers (Feitknecht and Lehmann, 1959; Basta, 1959; Colombo et al., 1965; and Davis et al, 1968) have proposed an oxidation mechanism that can produce such a continuous series.

During oxidation, oxygen cannot be added

within

the existing structure which already consists of a (nearly) cubic-close-packed oxygen framework.

However, free oxygen at the surface

of a crystal can be

ionized by the addition of two electrons which are derived from ferrous ions 2+ 3+ = 2Fe + 2e . The charge within the

within the crystal by the reaction 2Fe

original crystal is no longer balanced; to restore the balance it is necessary 3+ to remove 2Fe for every 3 oxygens added at the surface—corresponding to one 3+ of every three newly formed Fe ions. The removal is accomplished by diffusion of the Fe 3+ through the oxygen framework to the surface of the crystal, where one of several possible fates awaits the expelled Fe

L-22

3+

and newly formed

Figure L-9.

Variation of cell dimension with composition for spinels in the

F e ^ - F e ^ series.

The data of H'dgg (1935b) and Feitknecht (1965) are for

samples oxidized from magnetite at relatively low temperatures (1200°C).

Hagg's data and those of David

and Welch are converted from KX; Greig et al. (1935, pp. 296-297) reported that the cell edge of their specimen was approximately 0.01 & smaller than that of pure magnetite, which is 8.394-8.396 A.

0

:

(1) They may be removed in solution if a liquid phase is present.

(2)

They may be deposited epitaxially as hematite on the surface of the crystal. (3) They may be deposited on the surface of the crystal as an extension of the original spinel structure, thus forming new maghemite.

Ideally these three

different mechanisms lead to three different products:

Case (1) produces a

crystal of maghemite containing the same number of oxygens, but smaller than, the parent magnetite, since the unit cell of maghemite is smaller than that of magnetite.

Case (2) produces a similar maghemite crystal that is surrounded by

hematite, and case (3) results in a crystal of maghemite containing the same number of iron atoms but 1/8 more oxygens than the parent magnetite; the crystal will have increased in size by (§)(a . .la ), or by a factor H maghemite magnetite of approximately 1.118.

In practice, of course, it is unlikely that any one

of these mechanisms would operate alone; more likely two or even all three would be involved. Colombo et al. (1965) have shown that case (3) is unlikely to obtain if there is any hematite in intimate contact with the magnetite; the 3+ 2Fe and 0 will crystallize as stable hematite if seed crystals are present. L-23

It is worth noting that the Fe

3+

ions expelled during oxidation are not 2+ 3+ necessarily those that have been converted from Fe to Fe . As Verwey et al. 2+ 3+ (1947) have pointed out, the eight Fe and eight Fe in octahedral sites per unit cell of maghemite are better viewed as a cloud of eight electrons distri3+ . The electrons that ionize the exterior oxygens are drawn 2+

buted around 16 Fe

from this cloud, not from any given Fe

ion.

Probably some tetrahedral cations

also migrate, since there may be some tetrahedral vacancies. The cation-deficient F e ^ ^ spinels produced by Kullerud and Donnay (1967) The cation-defi by reactions such as 3Fe 3 0 4 + 2S = FeS 2 + 4Fe2C>3 (spinel) must have formed in the absence of excess oxygen.

In this case it is sulfur,

rather than oxygen, that effectively removed electrons and iron from the magnetite.

Kullerud and Donnay have argued that an important part of the vacan-

cies in these spinels occur in tetrahedral sites.

If this interpretation were

correct, it would indicate a subtle difference between oxygen and sulfur in 3+ inducing diffusion of Fe

in magnetite; but it is difficult to envision a

mechanism by which the difference would be effective, since the diffusing cations may be drawn from sites many unit cells away from the new anions. Basta (1959) has tried to draw a distinction between the low-temperature phases produced by oxidation of magnetite and the limited solid solution of Fe20.j in at high temperatures (Greig et al. , 1935), calling the first the y-Fe-O.-Fe.-O, series and the second the a-Fe„0 o -Fe.0, series. It is true ¿ 3 3 4 2 3 3 4 that one of these series forms metastably at low temperatures and the other stably at high temperatures; but there seems to be no reason why they should be different structurally.

In each the decrease in Fe/O must be accomplished

by the formation of vacancies in the spinel structure.

Basta argues (1959, p.

700) that the data of Greig et al. show there is less change in a with composition in the a-Fe o 0„-Fe_0, series than in the y-Fe_0 n -Fe o 0, series. 2 3 3 4

A com-

2 3 3 4

parison of the Greig et al. data for magnetite containing 29 mole percent (22 weight percent) F e ^ ^

with

HSgg's data (Fig. L-9) shows no significant dif-

ference. It is mainly a matter of personal preference whether spinels intermediate in composition between Fe^O^ and

are considered as omission solid solu-

tions between magnetite and y-(or a ^ F e ^ ^ , or simply as cation-deficient magnetites.

I find the latter concept more meaningful, especially insofar

as it is extended to the titanium-bearing "titanomaghemites." Titanomaghenrites.

Spinels intermediate in composition between the Fe^O^-

Fe 2 TiO^ join and the Fe^^-FeTiO^ join are known to form in two ways: L-24

(1) by

stable—but quite limited—solid solution of Fe„0,-FeTi0- in the Fe.O.-Fe-TiO. 2 3 3 3 4 2 4 series at temperatures above 1000°C, and (2) by metastable oxidation of essentially binary Fe20^-Fe2TiO^ solid solutions at lower temperatures—probably well below 600°C.

As is the case with spinels intermediate between magnetite

and Fe2025 there is no a priori reason to assume that the structures formed at high and at low temperatures are basically different.

It is possible that

those Fe-Ti spinels formed stably at high temperatures may have a more nearly random distribution of Fe^ + , Fe^ + , Ti^"*", and vacancies over the octahedral and tetrahedral sites; but apparently no x-ray work on quenched specimens (let alone at high temperatures) has been done to test this possibility.

In the

absence of any data on the structures of the high-temperature phases, this discussion will concentrate on the structures of those formed by oxidation at lower temperatures.

These phases have been given a variety of names:

"Gener-

alized titanomagnetite," "abnormal" or "very abnormal titanomagnetite," "titanomaghemite," "y-titanohematite," and "y-phases."

The term "titanomaghemite" will

be used here to apply to any Fe-Ti spinels lying off the Fe20^-Fe2TiO^ join— that is, spinels with vacancies in some cation positions—and will not be restricted, as has sometimes been done, to spinels lying outside the Fe^O^Fe2Ti0Zt-FeTi03 compositional field. In the past, many workers have assumed that there was extensive, essentially binary, solid solution of FeTiO^ in magnetite.

To account for this

supposed solid solution, Dunn and Dey (1937, p. 160) proposed a hypothetical end member—cubic FeTiO^—by analogy with maghemite, cubic Fe203.

This pro-

posed end member was revived and named yFeTiO^ by Nichols (1955, p. 143) and by Chevallier et at. (1955, p. 322).

As subsequent experimental work and

analyses of volcanic Fe-Ti spinels have shown, however, there is no unique solid solution of FeTiO^ in magnetite, but rather an extension of the field of Fe-Ti spinels from the Fe,,0,-Fe„Ti0, join toward—and in some cases even 3 4 2 4 past—the Fe202~FeTi02 join. Consequently, this field of spinel has been viewed by some workers as reflecting solid solution of yFe202_yFeTi02 in Fe^O^Fe2TiO^.

There is nothing inherently incorrect in this viewpoint:

The concept

of omission solid solution is valid, and the use of a hypothetical end member is likewise acceptable.

However, I question the utility of this approach,

partly because it has led to pointless discussions as to the nature and structure of yFeTiO^, and partly because it tends to obscure the oxidation origin of titanomaghemites.

Chemically, the concept of yFeTiO^ is of little use:

is sufficient, for example, to refer to spinels lying in the Fe20^-Fe2TiO^FeTi0j-Fe202 field, without reference to a gamma end member.

The concept of

cation-deficiency, applied to spinels oxidized from essentially binary L-25

It

magnetite-ulvospinel solid solutions, is a much more useful (although no more "correct") approach to the structure of the titanomaghemites. In keeping with this viewpoint, compositions of titanomaghemites can conveniently be expressed in terms of two parameters, one that expresses in some way the Fe/Ti ratio, and one that indicates the degree of oxidation.

The first

parameter can be expressed either as an atomic ratio or as the mole fraction of Fe2TiO^ (or Fe^O^) in the parent binary solid solution, and is generally considered to remain constant during oxidation. (1) the number of Fe

The oxidation parameter can express

2+

ions oxidized (which equals twice the number of oxygens 2+ consumed), (2) the atomic fraction of oxidized Fe relative to initial total 2+ Fe

, or (3) the number of cations per unit cell; this number plus the number

of vacancies equals 24.

Table L-5 correlates some of the parameters that have

been used by various workers.

The parameter (Fe2TiOj!() / (Fe^O^ + Fe 2 TiO^) (this

is the x of the Japanese workers and of O'Reilly and Banerjee, and the q of Verhoogen) is particularly useful in that it is appropriate for the unoxidized magnetite-ulvospinel series as well as the oxidized 2+ spinels. Perhaps the most useful oxidation parameter is z = (number of Fe ions oxidized)/(initial number

2+

of Fe

).

A structurally significant aspect of the titanomaghemites is their cubicclose-packed oxygen framework, inherited from their titanomagnetite parents. Evidently the oxidizing oxygen is adsorbed to this framework and forms an extension of it, thereby creating new cation sites that are filled by diffusion of cations from the parent crystal (Sasajima, 1961, p. 201).

Sites in the

parent formerly occupied by the diffused cations become 2+ vacancies; 3+ and overall charge balance is maintained by the oxidation of two Fe oxygen added.

to Fe

for every

Note that the analytical data of Akimoto and Katsura (1959)—

which show that natural titanomagnetites retain 3+ their Fe/(Fe+Ti) ratios upon oxidation—demonstrate that the diffusing Fe ions must be retained in the overall expanded structure; if they were not the Fe/(Fe+Ti) ratio would change with oxidation.

In the absence of any detailed x-ray studies one assumes that

titanomaghemites do retain the spinel structure; it is possible, however, that all or some of the structures reported for maghemite (ordered vacancies, primitive cubic, tetragonal) may ultimately be found.

The most important aspect of

the structure that remains to be determined, therefore, is the distribution of cations plus vacancies in the 8a (tetrahedral) and 16d (octahedral) sites of the spinel structure. Several early models for cation distribution in titanomaghemites make use of assumed structural formulae for the hypothetical end member yFeTiO^» an approach that hindsight shows to be of little use. L-26

More appropriate are models

p-s.

«3 "O .c i/i o >oO

so

(O

>

0)

o

O 4-> O E «t.

CT) r LT) « 0 CT> CTl

« i3 1/) 4-> fl ^

-C (/) 3 r ^ « CD "O "O '— C C-—'

-o o c «0 >> c >> S «— E

•o — ' c (O •

c 0) CT) o o .c Sai

iv

O) M

>

CT)

3 -C O tS)

O

vil

v||

N

c

vil

vil

O

Ç 0 •H •u ü M

^

c 1 1— c o +-> (0 •o X o

0) M

+J

£ O

4-> c

•o + a> :\j N a;

y

•o

•r- X 1 1 1

4-

A»

A» O

0

+

e

CT) So

X o

+

oo

Í—

L-27

N

re

a ; +->

Ll. il

X vil

*

X o

Li_

evi m

-o 0) CT» N >> X *o o

Ll_

CSJ ai

ai u.

(/) c ai

c

c

c c a) 3 CT) >> -

O)

N

X 0

+

0

N

fc.

o

co C " O Oí 0) CT) N >,

5. Wechsler et al.

a and b of the unit cell are shown in dashed outline.

(1976).

L-46

Figure L-16.

Unit-cell dimensions for the Fe2Ti0,.-FeTi205 (Psb-Fpb) series

and the FeT^O^-MgT^O^ (Fpb-Kar) series.

Lines for the Fpb-Kar series are

linear least-square fits to the data of Lindsley et

al.

(1974).

The crosses

are data for a natural armalcolite from lunar sample 74243,2,2; it contains 3+ Al, Cr, Mn, and probably Ti (Wechsler et al. , 1976). Lines for the Psb-Fpb series are "eyeball" fits to the data (dots) of Akimoto et are constrained to the Fpb values of Lindsley et

al.

al.

(1957), but

(1974).

2+ and Mg

randomly distributed over the 4e (Ml) sites.

Wechsler et

at.

agree with

this arrangement for natural and annealed specimens, but claim that substantial disorder is present in armalcolites quenched from high temperatures.

Unit-cell

parameters for the FeTijOj-MgTi^O^ series are shown in Figure L-16. STRUCTURES OF RUTILE, ANATASE AND BROOKITE The structures of the three TiC^ polymorphs—rutile, anatase, and brookite— are discussed together to emphasize the similarities and differences between them.

Brookite is orthorhombic, whereas the others are both tetragonal, with

c/a < 1 for rutile and c/a > 1 for anatase. L-47

The first determination of the structures of rutile and anatase was made by Vegard (1916a,b), who observed that each could be considered as made up of O-Ti-O "molecules."

In rutile the O-Ti-O units lie in (001), but in anatase

they are parallel to [001]; Vegard considered this the explanation for the differences in c/a. The now more familiar view of these structures as sequences of coordination polyhedra was set forth by Pauling (1928; Pauling and Sturdivant, 1928).

In both structures Ti is coordinated by six oxygens, and each oxygen by

three titaniums; the difference is in the sequence of the coordination octahedra.

In rutile only two edges of the octahedra—those lying in (001)—are

shared (Fig. L-17), whereas in anatase four edges are shared (Fig. L-18). As would be expected, the shared edges are shorter than the unshared edges.

The

relationships between the "molecular" structures and the coordination structures are also illustrated in the figures.

In rutile the octahedra form chains paral-

lel to [001]; the chains are linked by the sharing of corners.

In anatase, on

the other hand, there are no distinct chains (Fig. L-18). On the basis of Vegard's data, Niggli (1921) assigned rutile to space group Pkylmnm and anatase to Ih^/amd", Parker (1923) provided an exhaustive confirmation of the latter assignment.

In each structure the Ti ions occupy unique

special sites, but the positions of the oxygen ions must be fixed by one experimentally determined parameter (Table L-10). Values of the unit cell

Figure L-17.

Two aspects of the

structure of rutile (TiOj).

O °2

The

circles show only the centers of ions and are not to scale. a.

The tetragonal unit cell as

determined by Vegard (1916a;1926), showing the O-Ti-O "molecules" lying in (001). b.

Four unit cells showing the

coordination octahedra of oxygens about the Ti at the center of each cell.

Only those two edges of the

octahedra that lie in (001) are shared with adjacent octahedra. The shared edges are shorter than the unshared edges.

L-48

M + ,-| particularly for armalcolites, the presence

requires the component Ti^O^ as well.

Experimental studies include those

of Lind and Housley (1972), Hartzman and Lindsley (1973), and Lindsley et at. (1974).

Ideal armalcolite (Fe Q

5MgQ

j T i ^ ) is stable down to 1010 + 20°C; below

that temperature it breaks down to Mg-ilmenite + rutile (Fig. L-28).

Kesson and

Lindsley (1975) showed that armalcolite is stabilized to still lower temperatures 3+ 3+ 3+ by A1 , Cr , and Ti .

Ilm+Rut Figure L-28.

. . . Q/ Mole %

Geik+Rut

Phase relations along the join F e T i ^ - M g T i ^ at low pressures

(Lindsley et at., 1974), showing the stability field for armalcolite.

Only the

upper curve Is binary; the lower curve marks the intersection of the join with the Ilm -Rut edge of Arm -Ilm -Rut three-phase triangles in the system FeOss ss ss MgO-TiO . Abbreviations as in Table L-12. L-80

Fe0-Fe 2 Q 3 -Mg0-Ti0 2 .join The Fe0-Fe 2 0 3 -Mg0-Ti0 2 join was studied in air by Woerman et at. (1969) and at 1160-1300°C and a variety of oxygen fugacities by Speidel (1970). Microprobe analysis of coexisting (Usp-Mt) g s and Ilm in Speidel 1 s experiments indicated that Mg preferentially enters the magnetite, in contrast to most natural pairs in which Mg is enriched in the ilmenite.

Pinckney and

Lindsley (1976) studied the join hydrothermally at 700-1000°C using reversed exchange equilibria and found that Mg preferentially enters the ilmenite, the preference becoming more pronounced with decreasing temperature.

It is not

clear whether the discrepancy with Speidel's results is simply a temperature effect or whether his experiments may have failed to reach equilibrium.

FEO-FE 2 O 3 -AL 2 O 3 SYSTEM Portions of the F e O - F e ^ ^ - A ^ O j system were studied hydrothermally by Turnock and Eugster (1962).

They determined the extent of the miscibility gap

between magnetite and hercynite and also the composition of magnetite in equilibrium with hematite and corundum (Fig. L-29).

These results are particularly

applicable to the genesis of emery deposits.

1000

800

600 -

400

80 Mole %

Figure L-29. Mt

-He

The join F e ^ - F e A l ^

miscibility gap.

ss ss position of Mt from 1-4 kbar.

ss

100 FeAI204

(Turnock and Eugster, 1962) showing the

The curve labeled (Mt

(Hem,Cor) shows the comss in equilibrium with hematite and corundum. Pressures ranged

L-81

CR 2 0 3 "BEARING SYSTEMS

Chromites are ideally solid solutions between F e C ^ O ^ and MgCr^O^, but most terrestrial chromites also contain appreciable amounts of F e A ^ O ^ , MgAl^O^, Fe^O^

an
+ Hem

ss

.

(c) Oxidation is now extended to include the magnetite which is

replaced by hematite, and the exsolved spinel lamellae are scattered throughout as finely disaggregated blebs.

(d-e) These photomicrographs are at a lower and

higher magnification, respectively, illustrating the oxidation of original ilmenite to Pb

» the exsolution of pleonaste^ in magnetite, the replacement of

M t g s by H e m ^ , and the resistance of unresorbed spinels which are dispersed within the H e m ^ .

(f) A subhedral grain of original titanomagnetite with

satellite skeletal extensions consisting of Pb ss + Hem ss and two large rectangular areas of magnesioferrite^ (black, lower left). None of the original magnetite is any longer present.

(g) A higher magnification photomicrograph

of a grain similar to that of grain (f) but with the magnesioferrite^ now being partially replaced along {111} by hematite. Scale bar = 25 ym.

Hg-26

Hg-27

Oxidation of discrete primary ilmenite Seven clearly defined oxidation assemblages emerge for ilmenites in the same suites of basalts and in the same specimens described for titanomagnetite above.

Examples of these assemblages are illustrated in Figures Hg-7-9, and

the classification of assemblages with progressive high temperature oxidation is as follows: 1) homogeneous ilmenite 2) ferrian ilmenite + ferrian rutile 3) ferrian rutile + (ferrian ilmenite) 4) rutile + titanohematite + ferrian rutile + ferrian ilmenite 5) rutile + titanohematite 6) rutile + titanohematite + (pseudobrookite) 7) pseudobrookite + (rutile + titanohematite) Although this is the most commonly observed sequence, an alternative sequence of ilmenite oxidation results in the formation of pseudobrookite directly, thus bypassing other intermediate assemblages.

Factors controlling

the development of either of these progressive sequences are largely temperature dependent and by implication, fg^ dependent; the former develops at lower temperatures and the latter at correspondingly higher temperatures (+ 800°C). Shielded ilmenite, or ilmenite which has remained unreacted at elevated temperatures, will follow the sequence of assemblages outlined above provided that levels of ÌQ2 are still sufficiently high and with the provision that slow cooling ensues.

Ilmenite

oxidation

classification

Details of the seven stages of ilmenite oxidation observed optically are as follows; each stage is prefixed by R (rhombohedral) to distinguish this classification from the cubic (C) ulvospinel-magnetite classification.

Hg-28

R1 stage.

Homogeneous, unoxidized primary ilmenite is the lowest state

of oxidation in this classification (Fig. Hg-7a). R2 stage.

The first signs of ilmenite oxidation observed optically are

an increase in reflectivity and a change in the color of ilmenite from reddishbrown to light-tan. develop along

Fine wisp-like sigmoidal lenses (1-5 um) of ferrian rutile

{0001} and {0111} parting planes of the ilmenite.

These lenses

generally develop in equal concentrations throughout the entire ilmenite crystal although aggregations towards the grain boundaries and along fractures are also present.

There is a tendency for lamellae parallel to {0001} to be

coarser than those developing along the rhombohedral planes, and by analogy with exsolution in ilmenite-hematitess these finer lenses may develop at lower temperatures and under conditions of lower diffusion activity (Fig. Hg-7b). The distribution of larger and smaller lenses, however, is not an ordered or seriate arrangement as is commonly the case in Ilm s s

or

® e m s s which have ex-

solved from Ilm-Hem ss R3 stage. more abundant.

With more advanced diffusion the lenses become thicker and There are slight changes in color from pale gray-white to

white as the lenses thicken, with marked increases in reflectivity, the appearance of white internal reflections, and a pronounced increase in anisotropy.

Lenses within any one crystallographic plane are in optical continuity,

are doubly terminated, and are relatively well defined at each of the progressive growth stages.

Ferrian rutile lenses are mantled by haloes of bleached

ferrian ilmenite that are gradationally lighter in color as the ferrian ilmenite host is approached. This variation in color is due in part to differential +2 +3 concentrations of Fe

and Fe

in the host, and in part to the concomitant

diffusion of Ti from the ferrian ilmenite to the ferrian rutile lamellae. The {0001} and {0111} lenses rarely intersect or continue along precisely the same planes for any distance; tapering occurs on contact and the sigmoidal form of the lenses gives rise to classic syneusis textures (Fig. Hg-7c-d). R4 stage.

This stage of ilmenite oxidation is considerably more complex

than the preceding stage or the stage that follows.

The assemblage is a four-

phase metastable assemblage comprising ferrian ilmenite, titanohematite, ferrian rutile and rutile.

The rhombohedral ferrian ilmenite and titanohematite

are present in approximately equal concentrations and constitute the host; ferrian rutile and rutile are present as sigmoidal lenses or as finely disseminated lamellae and these phases are oriented along presumed {0001} and {0111} planes of the original ilmenite. Hg-29

Optical continuity of phases along

Figure Hg-7.

Oxidation of primary ilmenite.

discrete crystal of primary ilmenite.

Stages R1 to R4.

Stage Rl.

(a) Homogeneous

(b) Subhedral crystal of il-

menite showing one coarse and many fine lamellae of ferrian rutile. (c) Fine sigmoidal lenses of ferrian rutile in ferrian ilmenite.

Stage R2.

Stage R3.

(d) Similar to grain (c) but with one particularly well-developed set of ferrian rutile lamellae along {0001} parting planes of ferrian ilmenite. planes are {0111} or {0112}.

Stage R3.

The subsidiary

(e-g) Three metailmenite crystals il-

lustrating the progressive transformation of ferrian rutile to rutile, and of ferrian ilmenite to titanohematite.

All grains contain the four-phase meta-

stable assemblage typical of stage R4.

The light-gray sigmoidal lamellae are

rutile, the lighter-gray lamellae (of. grains (b), (c), (d)) are ferrian rutile and the host phases are ferrian ilmenite (medium gray) and titanohematite (gray to white).

The oxide complex to the upper right of grain (g) is an oxidized

olivine crystal (see Fig. Hg-17). Scale bar = 25 um.

Hg-30

Hg-31

these planes is maintained, but there are marked reflectivity increases in both the host and the lamellar components when this assemblage is compared 2+ 3+ with the R3 assemblage. Diffusion rates and the redistribution of Fe , Fe and Ti are highly variable even within the limits of a single crystal, and as a consequence the compositions of phases are also highly variable.

The process

reflected in this four-phase metastable assemblage is the transition of ferrian ilmenite to titanohematite and of ferrian rutile to rutile (Fig. Hg-7e-3). B5 stage.

Rutile and titanohematite develop extensively at this stage of

oxidation (Fig. Hg-8a).

Ferrian rutile and ferrian ilmenite may appear as

finely disseminated unreacted phases but these gradually disappear with increased diffusion and oxidation. phase assemblage:

This stage is represented by a simple two-

The titanohematite host is, in general, optically homog-

eneous and each grain has the properties of a single crystal.

Titanohematite

is considerably whiter, has a higher reflectivity and is more strongly anisotropic than its incipient counterpart in R4.

These properties result from an

+3 increase in Fe

and a loss of TiO^.

The ferrian rutile lenses are sharply

defined and crystallographically controlled along {0001} and {0111} parting planes.

These lenses are in optical continuity along their respective orien-

tations, are strongly anisotropic, and show distinct yellow or yellow-red internal reflections.

From a textural standpoint stages R3, R4 and R5 are

identical, insofar as each assemblage contains a member of the ilmenitehematite series and each assemblage is characterized by lenses of a Ti-rich oxide which are crystallographically controlled within the host. RS stage.

This stage is characterized by the development of P b ^ from

the R5 rutile-titanohematite assemblage.

In the incipient stages, pseudo-

brookite is concentrated along cracks and grain boundaries and gradually replaces the central-most regions of the grain by preferential migration along rutile lenses.

Both titanohematite and rutile participate in the reactions,

but the development of pseudobrookite in frond-like protuherances is most clearly controlled by the distribution and orientation of rutile (Fig. Hg-8 b-e).

Pseudobrookite develops as a result of solid-state diffusion and counter

diffusion of cations, with the addition of oxygen, between the rutile and the titanohematite.

Because the development of pseudobrookite is controlled by

pre-existing rutile and because of the compositional differences between the two primary phases, the resulting pseudobrookite is optically and compositionally heterogeneous.

The pseudobrookite is polycrystalline and varies

Hg-32

from pale to dark gray in color, being slightly darker in close proximity to rutile and lighter in regions adjacent to or in contact with titanohematite. These optical differences reflect higher TiC^ contents and correspondingly higher Fel^O^. contents in those pseudobrookites developing from nuclei of rutile.

Similarly, pseudobrookites which develop predominantly by excess

TiC^ diffusion into titanohematite result in an enriched Fe2TiO^ component. Variations in color are the dominant differences in these pseudobrookites; reflectivity and the degree of anisotropy are almost identical, as are the reddish internal reflections. The well-defined lensoidal texture of rutile is totally disrupted by the replacement of pseudobrookite.

Replacement is rarely complete and even in

the most advanced stages unreacted subgraphic inclusions of rutile and titanohematite tend to persist in the pseudobrookite host.

These inclusions are

neither ordered in distribution nor crystallographically controlled, which is a useful distinguishing textural feature between generative pseudobrookite (i.e. , Pt>ss from pre-existing assemblages) and pseudobrookite which is undergoing decomposition (to rutile + hematite^). R7 stage.

This is the most advanced stage of oxidation in the ilmenite

classification and is represented by the predominance of pseudobrookite (Fig. Hg-8f-h).

Intermediate stages between complete pseudomorphs and the R6 as-

semblage (ferrian rutile + titanohematite + pseudobrookite) are frequently observed in single polished sections; in fact, the entire spectrum from totally unoxidized ilmenite to pseudomorphs of pseudobrookite after ilmenite although common in many basalts are also occasionally present in single hand specimens. Somewhat similar optical and textural variations as those noted in R6 apply here.

Pseudobrookites are optically heterogeneous,and faint ghost-like

traces of original ferrian rutile are sometimes apparent; these relic lenses are particularly evident in crossed-nicols, and two distinct sets of extinction occur on rotation of the microscope stage.

Pseudobrookite generally consti-

tutes 80-90% of the pseudomorph, ferrian rutile is next in abundance, and titanohematite is a minor component or totally absent.

Ferrian rutile and

titanohematite occur as drop-like inclusions or as subgraphic intergrowths and show no preferred orientation in the pseudobrookite.

As noted in R6,

these textures do not result from the breakdown of pseudobrookite but simply reflect the incomplete formation of pseudobrookite from a pre-existing assemblage (viz. ferrian rutile + titanohematite). tion experiments on pseudobrookite

Our isothermal decomposi-

(Haggerty and Lindsley, 1970) show that

Hg-33

Figure Hg-8.

Oxidation of primary ilmenite.

Stages R5-R7.

(a) This grain is

a classic example of stage R5 with coarse rutile lenses in a host of titanohematite.

(b-c) P b s s (gray) mantles on the assemblage R + H e m ^ after original

primary ilmenite.

Grain (b) contains abundant rutile in the mantle assemblage

whereas grain (c) is mostly pseudobrookite.

These mantles develop progressively

from the two-phase assemblage illustrated by grain (a), and the assemblage is characteristic of the R6 stage of oxidation. (c) Pb

(d) In contrast to grains (b) and

has developed along cracks in an ilmenite grain which is otherwise at

ss

the R5 stage (R + H e m ^ ) . Note that the protruding crystals of Pb, which are normal to the cracks, are parallel to the directions of the rutile lamellae. Stage R6.

(e) The assemblage in this grain is similar to those in grains (b),

(c) and (d) and illustrates the disruption of rutile lenses by the encroaching Pb which replaces the assemblage R + Hem . Cuniform inclusions in Pb are ss ss ss rutile (light gray) and H e m ^ (white). Stage R6. (f) The core of this oxidized crystal consists of R + Pb with associated Hem , whereas the outermost part J ss ss of the crystal consists of a Hem g s matrix with Pb g g lamellae along a crude {111} relic fabric.

The relative ratios of R:Pb

core region was originally H

m

s s

an

and Hem :Pb suggest that the ss ss ss d that the mantle was originally Usp-Mt ss>

an example comparable to the relationship illustrated in Figure Hg-3a. C7 and R7. menite. Hem

ss

.

Stages

(g-h) Completely pseudomorphed crystals after original primary il-

Dark gray is Pt>ss> medium gray is rutile, and the whitish areas are Stage R7.

Hg-34

Hg-35

rutile + hematite^ or rutile + ilmenite^ develop along {100} or {110} planes as fine lamellae or as lenses in the pseudobrookite host; textures of this form do exist in basaltic ilmenite and must therefore result by decomposition at lower temperatures. An alternative oxidation trend to the seven stages described above are present in a small percentage of basalts and examples and are illustrated in Figure Hg-9.

In this trend Pb g s develop directly from ilmenite, and other in-

termediate assemblages are not observed except in these ilmenite crystals which are partially oxidized, or in ilmenites which have escaped initial oxidation.

The distinction between these trends appears to be temperature

dependent, with the direct formation of Pt>gs developing, at higher temperatures, and the rutile-titanohematite to Pb s g trend developing at somewhat lower temperatures (i.e., above 800°C and below 800°C).

The assemblage

rutile + titanohematite which may be present in partially reacted ilmenite is paragenetically later than Pb s g developing directly from ilmenite, suggesting that oxidation is not isothermal but is continuous with cooling. The fact that second generation Pt>ss (from rutile + titanohematite) are rarely observed in basalts displaying the direct formation of Pt>ss from ilmenite suggests furthermore that high temperature oxidation is kinetically rapid and that lower-temperature oxidation is expectedly sluggish. There are several distinctive features in pseudobrookites which develop at elevated temperatures: a) Crystals are a darker gray, rarely show the typical reddish internal reflections, have slightly higher reflectivities and are distinctly less anisotropic than those pseudobrookites associated with the indirect formation of Pb

from rutile + titanohematite. ss b) The major included phase in pseudobrookite is rutile.

A crude

lineation of bleb-like inclusions is sometimes present but generally no well-defined crystallographic ordering is apparent. c) The pseudobrookite pseudomorphs are either single crystals or are coarsely polycrystalline. d) Pseudobrookite is considerably more homogeneous and shows no preferred orientation in the ilmenite that it replaces. e) The directly formed Pb

are enriched in FeTi20j, whereas the indiO i

rectly formed Pb contain higher concentrations of Fe , lower concentrass 2+ tions of Fe and TiO„ and contain more of the Fe-TiO, component.

Hg-36

The distinctions are in accord with experimental data for the join Fel^O,.Fe 2 Ti0 5 (Lindsley, 1967; Haggerty and Lindsley, 1970) and on heating experiments of natural ilmenites which confirm that the direct formation of Pb

ss

is rate

controlled and temperature dependent (Haggerty, unpublished).

PHASE CHEMISTRY OF OXIDATION ASSEMBLAGES

Introduction There are wide variations in the compositions of the oxidation assemblages which develop from primary titanomagnetite and from ilmenite. are dependent upon and result from:

These variations

(a) the composition of the initially crys-

tallizing magma; (b) the composition of the primary oxides; (c) the degree to which decomposition has proceeded; and (d) the absolute values of T°C and of initial crystallization and of deuteric cooling.

In most instances a

single variation is difficult to attribute to a single chemical or kinetic parameter but in order to outline in general the basic principles of oxide decomposition, a small but judicious choice of examples will be considered. These examples do not cover the entire, or the expected, compositional ranges of the assemblages discussed in the previous section, nor indeed does the choice address many of the outstanding problems in textural interpretation. Discussion is limited to the mineral chemistry of two basalts from Iceland, an olivine basalt from western Iceland (WA14-1), and a tholeiitic basalt from eastern Iceland (JS-10); two coarse-grained diabasic basalts from Oregon (DHL 296/69 and DHL 296/69); and a picritic basalt from Hawaii (H-5).

The

Iceland basalts are chosen because of their initial compositional variations, because both basalt samples contain a wide variation in oxidation assemblages and because JS-10 is one sample from a continuous sampling of a single lava profile displaying systematic oxidation variations from base to top.

The

Snake River plateau basalts illustrate an equally wide variation in oxidation assemblages, but the distinction for these samples is that oxidation can be demonstrated to be related to joints and contrasts with the Iceland examples where oxidation is particularly pervasive towards the central-most regions of lava flows.

The Hawaii sample is highly oxidized and although neither

the primary mineralogy nor other intermediate assemblages are present, the end products of oxidation are particularly well developed, coarse grained, closely approximate equilibrium, and are more magnesian and aluminous than the other examples.

Representative analytical data of a selection of assem-

blages for these samples are given in Tables Hg-1 through Hg-4, and these data are plotted in Figures Hg-10 through Hg-12. (Text resumes on p. Hg-46.) Hg-37

Figure Hg-9.

Rapid oxidation of primary ilmenite.

micrographs illustrates the formation of bypassing oxidation stages R2 to R6.

(a-f) This series of photo-

members directly from H m s s members

Grains (a) and (b) show Pt>ss in associa-

tion with large areas of partially oxidized ferrian ilmenite which contain oriented lamellae of ferrian rutile. the

In these grains, and in grains (d) and (e),

is the product of rapid oxidation at high temperatures, the unoxidized

portions of the ilmenite follow the sequence of progressive oxidation stages illustrated in Figures Hg-7 and Hg-8, so that grains (a) and (b) contain Pb s s typical of stage R7 as well as areas classified as R3; grain (c) contains R7 as well as R4 (uppermost portion:

R + H e m s s + ferrian rutile + ferrian ilmenite);

grain (d) has gone one stage further and the unoxidized portion is now at R6 (R + Hem s g ) with the remainder of the grain at R7.

For grains (e) and (f) high

temperature oxidization proceeded very nearly to completion, with the entire grain advancing uniformly to the R7 stage.

Hg-38

Hg-; -39

Hg-40

ri CMrio o ® UI CR, tn (N

«FF NFL>

s «sassate R L N R I O O F F L R-|O

&

O O ^ U

ì

3

I * O

a »»sgRsste

H

saaasgi

H CM

ssgggg

D _.

.

.

.

.

o © «a- o o

ass|gs!

ri CM ^ o o ^ o o r j

0.25

h ° ° 5

2.63

H

K 3 G 5 E R K

C M

43.83

S S H

ri

3

43.38

S S

Y ^ ^ O O ^ O O H V

51.85

sssgsss|flNH5a§3s

H H M O O « OLA* PI O ~ • ' R R

51.17

a

« S I S I S S I L S © O ©

O O ©

OL RI

S i i Issili O O O O ©

©

O CM

N G G S G G S S I G O ©

S S J O S S : N

N O N O O R T P

O O O O O CM

§ 8 £ I § S S ] I

O O O O O O

ss § s s s s a ais

Ori

" S S S S S S S L I © © O O O © O CM

T

L

I

L

I

S

L

L

L

£ S

ss

ss

ia

ajsassacj^sgsjiis? CS O CM O O e OO HOO O r

s-s

s a s s a s ^ S i S i i f ! es © O © O rH »"*" O O H O O O r

£> Ol 1 N O H) O o o es o olo n

yo o o o
O l-l O O O r

S® OOO' ° ^ ° ° °C O O I-H O O O .

ss3sass|s„s|g|§: d S 0 0 " : OOOOOO-

SÏÏggSSS^SSggSSS en \o o o o -a- c> œ PÎ . ° ° ""í ^ es m t-h ON rv M O O O O es

fl O es u"l O i-H "i Mo>ON o O o o o o i,-hum

o

>3S

^ 0\ S S S O N N ¡O H fri vO * rH ' O O '

OO— II O O O '

s sp g i a ^ i i g i i g i OaOn r-s os O 3 OO OOOC

sr

cMvOooesor-on eses -fl-O r» r» © o © © eo

¥>» u' \0

o o rH o o o h n

3SSS3" S§3SSg3 .NC1Í1HOU1HN co u-i O O es co voló en

t l f l i s l l l 2:3 s ö ¡sä ¿¡a

Hg-42

Ti0 2

WEIGHT Figure Hg-10.

PERCENT

Phase chemistry of coexisting assemblages in sample WA14-1.

Hexagonals are ilmenites and oxidation products after ilmenite. titanomagnetites and products after titanomagnetite. are present:

Triangles are

The following assemblages

(1) titanomagentite + composite ilmenite; (2) titanomagnetite +

sandwich ilmenite; (3) titanomagnetite + trellis ilmenite; (4) ferrian ilmenite + ferrian rutile; (5) magnetite + R + Hem (Hem obtained from average analss ss yses of R + Hem ; R + H on diagram—see text); and (6) Pb + R + Hem (obSS ss ss tained from average value of R + H e m s s ; R + H on diagram). values are for wt% MgO. neous ilmenite).

The percentage

Primary unoxidized ilmenite is marked as HI (homoge-

Analytical data in Table Hg-1.

Hg-43

TiOp

FeTi205

i

FeTiO, 'T1-6%\

\ R3 0%

VR+H

Fe2Ti04

PbVv\R+H . Fe2Ti05

4 6% Samp!»•

FeO Figure Hg-11.

4 and 5

JS-tO

V WEIGHT

PERCENT

Fe3^4 3 0,

F e2, 0U :3

Phase chemistry of coexisting assemblages in sample JS-10.

semblages are as follows:

As-

titanomagnetite + trellis ilmenite; (2) titanomag-

netite + composite ilmenite; (3) ferrian ilmenite + ferrian rutile; (4 & 5) Mt + R + Hem s s (Hem^ values obtained from R + H on diagram—see text) ; (6) Pb g s + Hem s s + R; and (7) R + H e m ^ . values are wt % MgO.

Primary ilmenite is HI.

Analytical data in Table Hg-2.

Hg-44

Percentage

Ti02

WEIGHT PERCENT Figure Hg-12.

Phase chemistry of coexisting assemblages in sample DHL 295/69

and DHL 296/69.

The former sample is relatively unoxidized and consists of:

(1) titanomagnetite + composite ilmenite; (2) titanomagnetite + trellis ilmenite; (3) ferrian ilmenite + ferrian rutile.

The second sample is more

highly oxidized and the assemblages are as follows: + R; (5) Pt>ss +

R

(4) Mt

SS

+ Hem

S3

+ Pb

S S

after primary ilmenite (I), after composite ilmenite (C),

and after trellis ilmenite (T); (6) ferrian ilmenite + titanohematite + Pb (7) R + Hem

ss

after primary ilmenite.

ss Analytical data are listed in Table

Hg-3 for both samples which were collected approximately 80 cms apart across an oxidized joint selvage (see Fig. Hg-22).

Hg-45

;

Titanomagnetite-ilmenite

assemblages

Homogeneous titanomagnetite is absent in all samples, and the compositions of discrete primary ilmenite are remarkably similar for the Iceland basalts and for the Oregon basalt. MgO contents

The latter have slightly lower initial

(1 vs. 2 wt % MgO) and the former h a v e higher H e m s s

(8 vs. A wt % Fe20j).

contents

T h e first textural intergrowths to be considered are

titanomagnetite-ilmenite assemblages for ilmenite in trellis (T), sandwich (S) and composite (C) relationships.

For the Iceland basalts the m e a n MgO

contents of C, T, and S ilmenites in WA14-1 are 2.2, 2.7, and 2.6 wt X MgO respectively.

In terms of I ^ O ^ contents the C ilmenites contain the smallest

proportion of H e m g s >

the S ilmenites the maximum, and the T ilmenites are

intermediate between these two extremes. virtually identical.

Minor element concentrations are

For the second basalt

difference between T and C ilmenites

(JS-10) the only significant

(sandwich laths are absent) is the lower

Fe^Oj contents of trellis lamellae; this difference is the reverse of that shown for WA14-1 and is also m o r e pronounced.

The relatively unoxidized

Oregon basalt (DHL 295/69) has a range of MgO contents in T lamellae between 2.4 and 2.6 w t % whereas the C inclusions are close to 2 wt % MgO.

Ranges

in minor element concentrations are similar for b o t h T and C ilmenites. Trellis lamellae have an average i ^ O ^ content of 3.2 wt % and a range b e tween 3.2 and 5.3 wt %; the latter maximum value is clearly atypical and is restricted to a single grain.

Significantly different H e m ^

contents are

observed for the composite inclusions which fall into two populations: first has a m e a n F ^ O ^

the

2.2 wt X w h i c h is lower than the T lamellae and a

second w h i c h has a m e a n i ^ O ^ v a l u e of 6.2 w t X .

For these three basalts

two exhibit trellis compositions w h i c h are m o r e oxidized than the coexisting composite inclusions and the O r e g o n basalt suggests that some C inclusions m a y h a v e resulted by oxidation exsolution while others are primary precipitates.

This complex distribution is n o t at all atypical,and it is noteworthy

that the differences between discrete and intergrowth ilmenites and the differences among ilmenites in the second category are subtle.

Expectedly w i d e

variations in primary precipitated ilmenites from ilmenites w h i c h develop at subsolidus temperatures are in general not observed, except for MgO contents which tend to be higher in exsolved

ilmenites.

Turning our attention now to the magnetites w h i c h host these ilmenites we are confronted w i t h an equally complex distribution.

Dealing

firstly

w i t h WA14-1 w h i c h contains the three textural types of ilmenites, w e find satisfactorily that the titanomagnetites which coexist w i t h "exsolution"

Hg-46

trellis lamellae (Tm = trellis magnetite) are m o r e oxidized than those titanomagnetites w i t h either S (Sm) or C (Cm) inclusions, w h i c h are almost

identical.

N o t e also that significant differences are present in the Al^O^ contents of T m w h e n compared w i t h either of the Sm or the compositions of the C m (2.5 w t % for T m and 1.5-1.9 wt % for Cm and Sm).

M a g n e s i a is not as clearly defined:

Trellis and sandwich magnetites are similar (0.9 w t %) whereas composite magnetites are enriched in MgO

(1.3 wt %).

For the second basalt, JS-10,

the oxidation trend is repeated and is consistent w i t h WA14-1; namely, that magnetites hosting T ilmenites are m o r e highly oxidized and contain higher concentrations of Al^O^ than those hosting composite inclusions.

MgO con-

tents, a l t h o u g h different by a factor of two, are 1 wt % MnO.

Ferrian ilmenite and ferrian rutile This assemblage is characteristic of oxidation stages R2 and R3 for discrete ilmenite, and for stage C4 in the titanomagnetite classification.

In

general the assemblage—which is restricted to trellis, composite and sandwich intergrowths in titanomagnetite—is extremely fine grained, and individual mineral components are beyond the resolution of the electron microprobe.

A

comparison, therefore, of the assemblage ferrian ilmenite-ferrian rutile in titanomagnetite intergrowths is not possible with similar assemblages in discrete ilmenite.

Notwithstanding this limitation, the close proximities in

mineral compositions of the four ilmenite textural types still permit the basic compositional trends which are established for discrete ilmenite to be related to those which develop in titanomagnetite.

As noted previously, an

optical comparison of ilmenite in titanomagnetite with discrete ilmenite shows that the trellis lamellae are initially more susceptible to oxidation than other textural types, and from the composition of these lamellae the explanation for this difference is most probably related to higher Hem

ss tents since all other elements are in approximately equal concentrations.

con—

Assemblages of ferrian ilmenite + ferrian rutile are present in the two Iceland basalts, and the analyses quoted are respectively for grains with minor ferrian rutile (WA14-1) and for grains with abundent ferrian rutile (JS-10).

The assemblage is also present in the second more highly oxidized

Oregon basalt (DHL 296/69), and the data reported are for grains of approximately equivalent high ferrian rutile densities. teristics to be noted are as follows:

The compositional charac-

(1) For both samples the ferrian il-

menites are more highly oxidized and contain larger concentrations of MgO than their respective unoxidized counterparts; (2) grains containing lower densities of ferrian rutile (WA14-1) are less oxidized than grains with abundant ferrian rutile (JS-10), and the differences in MgO contents with respect to discrete ilmenite are proportionately larger for JS-10 (3.1 vs. 2.2 wt % MgO) and correspondingly smaller for WA14-1 (2.5 vs. 2.3 wt

MgO).

This MgO

enrichment is pronounced in the Oregon sample where the increase is a factor of four larger (1.0 vs. 3.8-4.1 wt % MgO) than that of discrete ilmenite; Hg-49

(3) partitioning characteristics of other minor elements (Cr^O^, A ^ O ^ , MnO) are ill defined for the Iceland samples, but noticeable increases in A ^ O ^ and MnO contents are evident for DHL 296/69.

The compositions of the ferrian ru-

tiles are variable and difficult to determine quantitatively because of x-ray excitation of the host ferrian ilmenite.

Qualitative estimates of the FeO

contents suggest a variation of between 5 and 10 wt % FeO, and minor element concentrations are equivalent to background levels.

The essential features

of ilmenite oxidation at this stage in the sequence are an increase in the hematite component, enrichments of MgO, MnO and A ^ O ^ in ferrian ilmenite, and a separation of the titanium-rich phase ferrian rutile. Titanohematite, rutile, and magnetite Titanohematite + rutile are characteristic of the R5 stage of ilmenite oxidation, and the assemblage titanohematite + rutile + accessory magnetite with exsolved aluminous magnesioferrite is characteristic of the C5 stage of titanomagnetite oxidation.

These assemblages are present in the two Iceland

basalts and in DHL 296/69. Considering firstly the decomposition products of discrete ilmenite (R5) we observe that the Ti02 contents of the hematite constituent are typically between 20 and 25 wt % TiO^, that MgO varies between 3 and 4 wt % and that MnO is in the range 1-1.5 wt %.

Compositions of rutile for this assemblage

in the Iceland samples show that minor concentrations of FeO are present (ss after ilmenite-free titanomagnetite in DHL 296/69. In WA14-1, Pb

ss

are restricted to outermost mantles on R + Hem

ss

cores after

discrete ilmenite, and the broad beam composition of this latter assemblage, which was discussed earlier, is plotted in Figure Hg-10 with the deduced titanohematite composition and with the composition of coexisting P b ^ .

The

discrepancies in composition are the result of incomplete reaction, but note that the R + Hem s s composition is precisely within the range of Pb crete, composite and trellis lamellae in DHL 296/69.

f° r dis-

These similarities are

applicable to most basalts, a result perhaps of narrow limits for T°C, fg2 and the composition of the oxidizing medium, with minor differences in initial chemistry playing a small but influential role in mineral stability. Data for pseudobrookites and titanohematites in the picritic basalt (H-5) are listed for three grains of oxidized titanomagnetites in Table Hg-4.

Grain

1 also contains Mg-Al magnetite + exsolved ferrian pleonaste, whereas in grains 2 and 3 the spinel assemblage has decomposed to form aluminous magnesioferrite. Differences in titanohematite compositions among these grains are as follows: The aluminum contents vary between 1.6 and 2 wt % A ^ O ^ ; magnesium contents in grain 1 (coexisting with magnetite) are similar to MgO contents in grain 3 adjacent to A1 magnesioferrite (2-2.6 wt %), but these are a factor of two

Hg-54

less than the MgO contents associated with P b s s (4 wt % MgO); TiO^ values vary between 4.6 and 10.4 wt %, and if stoichiometry is assumed, grain 1 titanohematites contain on the order of 5 wt % FeO whereas those in grains 2 and 3 approach zero; MnO contents in the less reacted grain 1 are and MgO during the growth of the chromite from the basaltic liquid (Figs. EG-6a, EG-7a).

These

chemical zonings may indicate precipitation of chromite from a silicate liquid with continuous build-up in A1 2 0 3 > V 2 0 3 , and MgO contents.

Crystal-

lization of this Ti-chromite evidently took place before crystallization of EG-8

Figure EG-4.

A spinel with gradational zoning on the left-hand side and

abrupt and corroded boundary to late Cr-ulvospinel on the right-hand side. Length of photograph 300 microns.

Figure EG-5,

Spinel shielded by pyroxene on the left-hand side with grada-

tional zoning. right-hand side.

Corroded core with sharp contacts to ulvbspinel on the Length of photograph 300 microns.

EG-9

plagioclase and pyroxene and probably either before or during olivine precipitation.

Decrease in the FeO/(FeO+MgO) ratios of those spinels may, however,

be also due in part to subsolidus equilibration between Ti-chromite and host olivine, thus allowing considerable Mg diffusion from the olivine to the chromite.

Spinels displaying both gradational and sharp zoning and occurring

in the same basalts with early chromites are characterized by the second chemical trend (Figs. EG-6a, EG-7a).

Comparison between early spinels and this

spinel type indicates a distinct difference in textural features and chemical trends and thus puts important constraints on the behavior of basaltic liquids during cooling and crystallization.

The FFM initial ratio is usually higher

than that of the early chromites, indicating:

(a) Precipitation from a liquid

with higher FFM ratio and thus later in the crystallization sequence; and (b) the cooling basaltic liquid did indeed subsequently precipitate various chromite generations with various FFM ratios due to changes in its composition and at a certain cooling rate (El Goresy et at., 1976).

On the sides of the grains

with gradational zoning there is an increase in the TÍQ2/ ( T i C ^ + C ^ O ^ + A ^ O ^ ) ratio from core to rim from 0.02 up to 0.48 (Figs. EG-6a, EG-7a, EG-8a).

On

the other side of the grains with abrupt zoning, evidently, reaction took place between basaltic liquid and the Ti-rich outer rims thus corroding the spinels on this side down to the chromite core prior to new precipitation of the chromian ulvBspinel.

Compositional variation of chromian ulvüspinel rim on the

other side of the chromite with sharp contacts is usually chemically continuous with the outermost rims of the uncorroded side (Figs. EG-6b, EG-7b, EG-8b). The process of precipitation of Ti-chromite with gradational zoning followed by resorption of the Ti-rich layers could have taken place several times in the course of crystallization (e.g. Fig. EG-7b) and as much as three generations of Ti-chromites with various initial FFM ratio could be encountered in the same basalt.

Increase in the TiO^/(Ti02+Cr202+Al20^) ratio, decrease in

content, and increase in FFM ratio of these spinels may be indicative of co-precipitation with pyroxene but before massive crystallization of plagioclase took place.

Resorption of the Ti-rich zones of the same spinel should

result in an increase in the Cr concentration and to a lesser extent Ti-concentration of the basaltic liquid.

This slight but sudden increase in both

Cr and Ti concentrations may be reflected in the change of chemistry of the co-precipitating zoned pyroxene. The third chemical trend was encountered in olivine basalts and coarseand fine-grained pigeonite basalts. initial FFM ratio (0.80-0.85).

The Ti-chromite cores usually have a low

This ratio increases sharply from core to rim EG-10

0.70

0.60 OSO FeO/FeO*MgO

OSÒ"

1.00

oso"

FeO/FèO •MgO (b)

(a)

Figure EG-6. Projections of the rectangular face of the spinel prism displaying the first and second zoning trends in an olivine basalt, (a) (left) Compositional variation of early spinel (hachured area) and gradationally zoned spinels on their shielded side with the zoning filling the Apollo 12 gap. (b) (right) Compositional variations of the chromite corroded side and the late ulvbspinel demonstrating the gap after removal of the gradationally zoned equivalents.

080 0.85 030 0.95 0.70 0.80 FeO/FeO + MgO FeO/FeO^MgO (a) (b) Figure EG-7. Projections of the rectangular face of the spinel prism with the first and second zoning trends in a pigeonite basalt. (a) (left) Compositional variation in early spinels (hachured area) and gradational zoning on the shielded sides of spinels of the second trend. (b) (right) Compositional variation on the corroded side and the late ulvbspinel demonstrating the gap after removal of the gradationally zoned layers.

EG-11

0.70

0.80 0.90 FeO/FeO+MgO

0.80 0.90 FeO/FeOMgO (b)

(a)

Figure EG-8. Projections of the rectangular face of the spinel prism displaying (a) (left) the second zoning trend on the uncorroded and (b) (right) on the corroded side of spinels in an Apollo 12 ilmenite basalt.

0.70

0.80 0.90 FeO/FeO*MgO

070

1.00

(a)

0.80 0S0 FeO/FeOMgO (b)

Figure EG-9. Projections of the rectangular face of the spinel prism displaying the third zoning trend in an olivine basalt. (a) (left) Sharp increase in the FFM ratio a n d sharp rise in the T i 0 2 / ( T i 0 2 + C r 2 0 3 + A l 2 0 3 ) only in the terminating stage. (b) (right) Corrosion of the Ti02~rich zones and compositional variation of the late stage ulvospinel.

EG-12

(Figs. EG-9a,b) with slight increase in the Ti02/(Ti02+Cr203+Al2C>3) ratio and decrease in VjO^ content thus demonstrating crystallization from a liquid continuously enriched in FeO content.

In the coarse-grained olivine and pigeonite

basalts the Ti-chromite cores are rounded and corroded indicating that likewise the second trend reaction between basaltic liquid and chromite took place.

In

olivine basalts with spinels indicative of this trend numerous spinels display both gradational and abrupt zoning.

The chemical trend on the continuously

zoned side is indicative of enrichment in the Fe 2 TiO^ end member just before termination of the overgrowth, yet the trend is distinct from the second type discussed above.

The compositional variation of late stage ulvbspinel is almost

identical in all basalts regardless of grain size or type:

sharp increase in

Ti0 2 /(Ti0 2 +Cr 2 0 3 +Al 2 0 3 ) ratio, decrease in Cr20y(Cr2C>3+Al203) ratio, decrease in V 2 0 3 , and increase in FeO/(FeO+MgO) ratio (Figs. EG-6,7,8,9,10).

0.8 0.9 FeO/FeO *MgO

0.8 0.9 FeO/FeO+MgO

(a) Figure EG-10.

(b)

(a) (left) Base of the spinel prism showing the third composi-

tional variation trend of idiomorphic chromite core and late ulvospinel in an Apollo 12 pigeonite basalt.

(b) (right) Rectangular face of the prism dis-

playing the third zoning with the sharp increase in FFM ratio but without any significant increase in TiO,,/(Ti02+Cr,,03+Al203).

EG-13

Cationic relationships and substitutions The coupled oxide relationships of the three chemical trends discussed above substantiate the application of zoning trends in spinels for a better understanding of the crystallization histories of mare basalts.

It is not our

intention to present a detailed treatment of distribution of the various cations among the tetrahedral and octahedral sites of the spinel structure. Rather important is the substitutional relationship in zoned individual spinel grains along with the zoning as a direct function of change in chemistry of the cooling basaltic liquid due to crystallization of silicate phases. Chromite, spinel and hercynite are structurally normal spinels with the trivalent cations (Cr,Al,V) octahedrally coordinated in the B site and the divalent cations (Fe,Mg) tetrahedrally coordinated in the A site.

In contrast,

ulvBspinel is an inverse spinel with half of the B sites occupied by tetravalent Ti and the divalent cations distributed among the A site and the other half of the B site. Cationic substitutions in a zoned spinel in the B site involving the trivalent cations Cr, Al and V are extremely sensitive to changes in the concentration of these elements in the liquid.

Changes in the substitutional trends

may signal the crystallization of a major phase like pyroxene (Cr,V) or plagioclase (Al).

Changes in the substitutional trends for the divalent cations Fe

and Mg are also dependent on changes in the relative concentration of these elements in the cooling liquid due to crystallization of olivine and/or pyroxene.

However, subsolidus equilibration between chromite inclusion and

olivine host may have slightly modified or even obscured the initial trends and such an effect cannot be ruled out.

Substitutional trends for the pairs

Cr-Al, Cr-V, Al-V, and Fe-Mg offer an accurate way to trace the crystallization sequence (El Goresy et al. , 1976).

However, the substitutional trends

of these pairs cannot be treated independently since precipitation of a major phase should be documented in the majority of trends involving the different pairs.

The Cr-Al substitutional ratio is a function of this ratio in the

liquid whereby decrease in Al reflects decrease of the concentration of this element in the liquid due to crystallization of plagioclase.

Decrease in the

Cr and V ratios, on the other hand, indicates crystallization of pyroxene.

Cr-Al substitutional trends In olivine and pigeonite basalts with spinel generations displaying the first and second coupled chemical trends, three distinct Cr-Al substitutional EG-14

/ 3:1

/

15555.42 SPINELS CATIONS 152 ANALYSES

Figure EG-11.

Cr-Al substitutional trends

in various spinels in an Apollo 15 olivine basalt,

(a) (top) Cr-Al substitutions for

all spinels. are obscured.

Note that the relationships (b) (middle) Cr-Al substi-

tutional relationship for early spinels displaying the negative substitutional trend.

(c) (bottom) Cr-Al substitutional

trends of the second Ti-chromite type and late ulvospinel (lower branch). 2.40

3.60

4.80

6.00

AL 15555.42 SPINELS CRT IONS 31 ANALYSES

i £ I £

trends were encountered (Figs. EG-11,12). The substitutional trends become obscured if all the Cr and A1 cations of both spinel generations, along with that of late ulvospinel, are plotted altogether in the same diagram (Figs. EG-lla, EG-12a).

Early chro-

mites in a coarse pigeonite basalt (Fig. EG11b) and an olivine basalt (Fig. EG-12b) show a unique trend that varies from core to rim from Cr/Al ratios of 4:1 to 2:1 and of 3:1 to 2:1, respectively (Figs. EG-llb, 2.40

EG-12b).

3.60 AL

/ 3.,

/

15555.42 SPINELS CATIONS 121 ANALYSES

These trends demonstrate that

early chromites in these two basalts grew from a liquid with continuous build-up of A1 and thus before crystallization of plagioclase.

The Cr/Al substitutional trend

for early chromites in both rocks is entirely different from the trends of other chromite generations and ulvbspinel.

However,

it is quite similar to the substitutional trend reported for MgAl^O^-rich spinels in Apollo 14 samples (Haggerty, 1972a).

Spinels

with the second chemical trend display a curved substitutional trend starting for 2.40

3.60 AL

cores at a Cr/Al ratio of 4:1 with a steep

positive slope to 2:1 and with a sharp turn back to ratios higher than 4:1. This curvature was interpreted as indicative of change in the activities of

EG-15

Figure EG-12.

Cr-Al substitutional

trends in spinels in an Apollo 15 pigeonite basalt.

(a) (top) Cr-Al

substitutions for all spinels.

(b)

(middle) Cr-Al substitutions for early spinels.

Note the slight dif-

ference in slope from Fig. EG-llb. (c) (bottom) Cr-Al substitutional trends for the second chromite type and late ulvospinel.

AL

Cr^Oj and Al^O^ due to subsequent crystallization of pyroxene and plagioclase (El Goresy et at. , 1976). Along the first branch of this trend simultaneous crystallization of pyroxene and plagioclase is likely; however, the steep drop in Cr is indicative that at this stage pyroxene along with Ti-chromite was the main crystallizing phase.

The sharp

turn back to much lower Al concentration signals the entry of plagioclase as a major crystallizing phase RL

(Figs. EG-llc,EG-12c).

The Cr/Al

substitutional trend for chromian ulvospinel is indicative of crystallization from a liquid continuously depleted in both Cr and Al with the trend beginning at a Cr:Al ratio of 3:1, but changing continuously to a ratio lower than 2:1 for the outer rims of the ulvospinels.

Probably

at that stage of crystallization both pyroxene and plagioclase coprecipitated.

The crystallization

sequence deduced from the Cr-Al

EG-16

substitutional trends is also consistent with V/Cr, V/Al, and Fe/Mg substitutional trends for both rocks thus substantiating the application of these trends in zoned spinels as an indicator of the crystallization histories.

The conclu-

sions drawn from the analyses of the substitutional trends are in excellent agreement with experimental results obtained by Kesson (1975).

V-Cr and V-Al substitutional trends Though a trace element in spinels, vanadium is a valuable element because of its partitioning between chromite and ulv5spinel and its preference for 15555,42 EARLY SPINEL CATIONS 31 ANALYSES

15555.42 SPINELS CRIIONS 31 ANALYSES

2.140

4.80

7.20

9.60

1.00

12.00

2.00

3.00

4.00

5.00

AL

CR 15555.42 SPINELS CATIONS 119 ANALYSES

15555.42 SPINELS CATIONS 119 ANALYSES

4.60

7.20

9.60

2.00

12.00

CR

3.00

4.00

5.00

AL

Figure EG-13. V-Al and V-Cr substitutional relationships in an olivine basalt, (a) (upper left) and (b) (upper right) substitutional trends of early spinels indicating increase in V and A1 from core to rim. (c) (lower left) and (d) (lower right) substitutional trends for later chromite and ulvospinel. EG-17

p y r o x e n e (Laul and Schmitt, 1973) and hence it provides an additional check for p y r o x e n e entry as a crystallizing phase.

V-Cr and V - A l substitutional

trends

p r o v i d e a good control for the p o s i t i o n of b o t h pyroxene and plagioclase in the crystallization sequence.

Figures EG-13 and EG-14 document the substitutional

trends in two different rocks. spinels

The unique substitutional trend of the early

(Figs. EG-13a,c, Fig. EG-14a,c) is also w e l l developed thus indicating

crystallization before pyroxene and plagioclase as evidenced by the p o s i t i v e V - A l sympathetic trend and n e g a t i v e antipathetic V-Cr relationship.

The trends

for chromites w i t h the second chemical trend and the late ulvQspinels are antipathetic b o t h for V - C r and V-Al. 15065.93 SPINELS CRTIQNS 106 ANALYSES

2.40

1.60

7.20

9.60

1

15065.93 SPINELS CRT IONS 106 ANALYSES

£ * P

1.00

12.00

2.00

3.00

15065.93 SPINELS CATIONS 156 ANALYSES

4.80

4.00

5.00

BL

CR

7.20

9.60

15065.93 SPINELS CATIONS 158 ANALYSES

12.00

2.00

3.00

1

£ : f

AL CP Figure EG-14. V - A l and V-Cr substitutional relationships in a pigeonite basalt, (a) (upper left) and (b) (upper right) Substitutional trends of early spinels indicating increase in V and A 1 from core to rim. (c) (lower left) a n d (d) (lower right) Substitutional trends for later chromite and ulvospinel.

EG-18

Fe-Mg substitutional trends Several features can be recognized in the Fe-Mg substitutional trends shown in Figures EG-15 and EG-16.

All the substitutional trends observed in the Cr-Al,

V-Cr, V-Al diagrams are also encountered in the Fe-Mg relationships.

All Ti-

chromite and ulvospinel generations with various initial FFM ratios emerge. Early spinels again display a unique antipathetic trend with negative slope with increasing Mg substitutions for Fe from core to rim.

This trend very probably

indicates growth from a liquid with decreasing FFM ratio.

Haggerty argued

(1972a,b,c) that the divalent Fe versus Mg relationship is poor to totally incoherent and that distinct linear slopes do emerge if the Ti content of the spinels is considered.

Each slope proposed by Haggerty would correlate spinel

compositions with similar TiO^/(TiC^+C^O^+A^Oj) ratios.

This attempt does not

reflect the real substitutional trend and even obscures the Fe-Mg relationship in the zoned spinels.

This conclusion is demonstrated in Figures EG-15 and EG-

16, where vertical trends with sharply increasing Fe substitution for Mg from core to rim for the later Ti-chromite and chromian ulvospinel generations are evident.

In fact, these steep trends cross the several slope lines for compo-

sitions with various Ti-contents proposed by Haggerty.

Furthermore, slopes

constructed for compositions with similar Ti-content connect spinels of the various generations crystallized at different times and the relationship observed for early spinels would then completely disappear.

The Fe-Mg substi-

tutional trends of the second zoning trend demonstrate a continuous increase in the FFM ratio of the liquid after precipitation of olivine (El Goresy et al. , 1976).

Ti-(V+Cr+Al) substitutions The Ti-(V+Cr+Al) substitutional ratio displays the occupancy in the B site in the normal-inverse solid solution series of tetravalent and trivalent cations.

Non-stoichiometry of the spinels should cause a departure of the data

points from the

8

(Ti) to the 16 (V+Cr+Al) ratio.

No evidence of departure

from the 8:16 ratio was found and hence cation deficiency is questionable (Nehru et al., 1974; El Goresy et al. , 1976) (Fig. EG-17).

Neglecting to

include V and Si in the spinel analyses is responsible for speculation that either the B site is deficient or there is increase in the octahedral site occupancy for divalent cations. The Luna 16 mare type basalts are characterized by their high A ^ O ^ content (^16%).

Spinel analysis in these rocks indicate similar compositional

variation trends to Apollo 12 and Apollo 15 basalts (Haggerty, 1972). EG-19

However,

15555.42 SPINELS CATIONS 146 ANALYSES

F i g u r e EG-15.

Fe-Mg

substi-

t u t i o n a l trends for v a r i o u s z o n e d s p i n e l s in a n olivine basalt.

Note

antipathetic

n e g a t i v e t r e n d for early = 5 7 i 6 -7122

s p i n e l s towards h i g h e r M g substitutions.

Zoning

trends

of later c h r o m i t e s a r e steep a n d i n d i c a t e sharp Fe substit u t i o n for M g . = 2.108-5.648

Substitutional

trends of zoned g r a i n s

cross

s e v e r a l of the s l o p e lines by Ti=0.666-2.027 H a g g e r t y

1.00

.50

1.50

2.00

(1972a,b,c).

2.50

MG

15065,93 SPINELS CATIONS 263 ANALYSES F i g u r e EG-16.

Fe-Mg

1

£

2

S

substi-

t u t i o n a l trends for v a r i o u s zoned s p i n e l s in a p i g e o n i t e basalt.

M i = 5.746-7122

Antipathetic nega-

tive t r e n d s for e a r l y

spinels

is m o r e p r o n o u n c e d t h a n in Fig. EG-15.

All

chromite

generations with various FFM ratios

emerge.

Ti= 2.106 -5.64a

Ti = 0 . 6 6 6 - 2 . 0 2 7

.50

1.00

1.50 MG

EG-20

2.00

2.50

Figure EG-17.

Plot of Ti+Si against Cr+Al+V cations for spinels from Apollo 15

rake samples.

Open circles, spinels; filled circles, chromite and ulvtispinel

(from Nehru et al. , 1974).

they are characterized by considerable solid solution towards FeAl^O^ and MgAl^O^ (Fig. EG-18).

Evidently, these spinels crystallized from a liquid

with high but continuously decreasing Al concentrations (Fig. EG-19) as documented by the sharp decrease in the Al/Cr ratio.

This may indicate that in

comparison to Apollo 12 and 15 basalts, anorthitic plagioclase and pyroxene co-precipitated from the Luna 16 magmas (Bence et al. , 1972).

Continuous

crystallization of anorthitic plagioclase would explain the continuous and drastic decrease in the Al/Cr ratio.

Spinel compositions from various basalt

types from Apollo 12, Apollo 15 and Luna 16 sites are shown in Table EG-1. Ilmenite textures in basalts of the three landing sites are quite similar (Haggerty, 1971; El Goresy et al. , 1971).

The concentration of geikielite is

usually in direct relationship with the total MgO content of the rock, e.g. , high geikielite concentrations in rocks with high MgO content.

Ilmenite in

the majority of the Ti02~poor basalts is usually late in the crystallization sequence probably after the late Cr-ulvospinel.

EG-21

Figure EG-18.

Compositions of Luna 16 spinels in the modified spinel prism

(from Haggerty, 1972b).

Figure EG-19.

Atomic proportions of Cr as a function of A1 for Luna 16 spinels.

The dashed lines indicate the zonal trends (from Haggerty, 1972b).

EG-2 2

Table EG-1.

1 SiO Ti02 Cr20 A1 2 Û3 V2O3 FeO MgO MnO CaO Total

2

-

6.74 42.20 11.10 0.97 36.60 1 .89 0.40 0.05 99.50

25 .80 16 .30 4 .37 P .62 51 .30 1 .70 0 .36 0 .05 99 .80

Chemical analyses of spinels.

3 0,.52 3..60 47.. 10 10..70 0,.87 32,.80 2..94 0..35 0..34 99..22

4 0.. 1 1 32..60 1,.28 2..58 0 ..01 62..70 0 ..52 0 ..31 0 ..26 100..36

6

5 0 ..17 7..09 30..96 21..51

0.76 0.92 51 .49 14.48

34..42 4..60 0 ..36 0 ..08 99..19

25.85 6.67 0.46 0.38 101.01

7 0..35 29..70 5..54 1 .89 .

-

61 .05 . 0,.17 0 ..46 0 ..17 99..33

1: Titanian chromite, Apollo 12 (Taylor et al.1971, Table 5, p.865) 2: Chromian ulvöspinel, Apollo 12 (Taylor et al. 1971, Table 5, p. 865) 3: Magnesian-aluminian chrotnite, Apollo 15 (Nehru et al.1974, Table 2, p. 1225) 4: Chromian ulvöspinel, Apollo 15 (Nehru et al.1974, Table 2, p. 1225) 5: Magnesian-aluminian chromite, Luna 16 (Haggerty, 1972 b, Table 1, p. 335) 6: Cr-rioh aluminian-magnesian chromite, Luna 16 (Haggerty, 1972 b, Table 1, p. 334) 7: Chromian ulvöspinel, Luna 16 (Haggerty, 1972 b, Table 1, p. 335)

EG-23

Opaque oxides in TiO^-rich basalts TiC^-rich basalts were collected during the Apollo 11 and 17 flights from Mare Tranquillitatis and the Taurus-Littrow Site, respectively.

These high-

titanium basalts were formed in the period from 3.82 to 3.55 G.Y. and are mainly confined to the eastern half of the side of the moon facing the earth. Textural variations and compositional similarity among the Taurus-Littrow hightitanium basalts are suggestive of a similar source for both landing sites (LSPET, 1974).

The TiO^-rich basalts can be classified into two major types:

(1) plagioclase poikilitic ilmenite basalts; and (2) olivine porphyritic ilmenite basalts.

So far, the first basalt type was not encountered in the Apollo

11 site and is restricted to the Apollo 17 samples. tered in the studied basalts are:

The opaque oxides encoun-

ilmenite, armalcolite, chromian u1vospinel,

secondary titanian chromite, and rutile.

Of special interest are textural re-

lationships between armalcolite and the coexisting silicates and opaque oxides in the different rocks as well as variations of armalcolite chemistry and opaque oxides in the different rocks.

Armalcolite in the two different major rock

types show differences in their Cr and Fe/Mg ratios. Armalcolite relationships Two optically different armalcolite types were reported in several Apollo 17 basalts (Haggerty, 1973; El Goresy et at. , 1974):

(a) a grey variety usu-

ally mantled by Mg-rich ilmenite present in olivine porphyritic ilmenite basalts; and (b) a tan variety encountered in medium- and coarse-grained plagioclase poikilitic ilmenite basalts.

Preliminary investigations indicated

that the grey armalcolite shows higher MgO and Cr^O^ contents than the tan variety and this may be responsible for the difference in color (El Goresy et al. , 1973).

Smyth and Brett (1973) demonstrated that the two armalcolite

types are indistinguishable in terms of crystal structure ruling out the possibility that the types are different polymorphs.

Their study, however, showed

similar differences in MgO and C^O^ contents as reported by El Goresy et al. (1974).

Textural relationships of armalcolite-bearing assemblages in Apollo

11 Ti02_rich basalts indicate that only the grey variety mantled by Mg-rich ilmenite is present.

Apollo 17 plagioclase poikilitic ilmenite basalts do not

have equivalents in the Apollo 11 landing site.

Papike et al. (1974) report

that armalcolite morphology in olivine porphyritic and plagioclase poikilitic ilmenite basalts is due to local variation in silicate crystallization rather than to major paragenetic differences.

According to El Goresy et al. (1974)

there are indeed differences in the paragenetic sequence between plagioclase EG-24

poikilitic and olivine porphyritic basalts.

These textural and paragenetic

differences are outlined below. 1.

Plagioclase poikilitic ilmenite basalts: This rock type is charac-

terized by the presence of two pyroxenes:

(a) titanaugite with sectoral

zoning and (b) pigeonite as single crystal overgrowths on augite (Hodges and Kushiro, 1974).

Olivine and Cr-ulvospinel were among the first minerals to

crystallize followed by tan armalcolite followed by titanaugite and pigeonite. Ilmenite, plagioclase and then cristobalite were the last minerals to crystallize.

Armalcolite occurring in these rocks is only of the tan variety and is

exclusively present as inclusions in the titanaugite (Fig, EG-20); occasionally, it is present with ilmenite in sealed grain boundaries with armalcolite morphology (Haggerty, 1973).

The dominant feature, however, is the idiomorphic

blocky appearance of armalcolite crystal clusters occurring only in the cores of titanaugite.

In these rocks massive ilmenite precipitation started after

the majority of the pyroxenes crystallized.

In none of the studied fragments,

regardless of their grain size, were ilmenite reactions rims around armalcolite observed, although Papike et al. (1974) report ilmenite reaction rims in sample

7

Figure EG-20. clinopyroxene.

Cluster of idiomorphic tan armalcolite crystals enclosed in a Apollo 17 plagioclase poikilitic ilmenite basalt.

field 400 microns.

EG-25

Length of

70035.

The above-described crystallization path is not dependent on the

cooling rate.

The few exceptions reported by Papike et al. do not negate the

crystallization sequence described above. 2.

Olivine porphyritic ilmenite basalts: These rocks are characterized

by the presence of olivine phenocrysts (partially as a quench phase) with Crulvospinel inclusions.

These two minerals, as in the plagioclase poikilitic

basalts, were the first phases to crystallize (El Goresy et at. , 1974).

Both

were then followed by grey armalcolite, then ilmenite and at last augite, plagioclase and then tridymite. colite were not encountered.

In coarse-grained rocks olivine and armal-

The textures in both Apollo 11 and Apollo 17

samples are identical. Two main features differentiate the two rock types. a.

Olivine porphyritic ilmenite basalts contain only the grey armalcolite variety regardless if armalcolite is mantled by ilmenite or not.

b.

In olivine porphyritic ilmenite basalts, ilmenite precipitated directly after armalcolite, whereas in plagioclase poikilitic ilmenite basalts, ilmenite crystallized after the major part of titanaugite and pigeonite precipitated.

In olivine porphyritic ilmenite basalts the majority of the armalcolite grains are surrounded by ilmenite rims. tle vary from grain to grain.

Shape and width of the ilmenite man-

Usually, the armalcolite grains are surrounded

by continuous ilmenite mantles (Fig. EG-21) and the composite grain still displays armalcolite morphology.

The origin and shape of the mantling ilmenite

will be discussed separately in detail in a later section. The above-described differences between the two rock types, especially the presence of two pyroxenes in plagioclase poikilitic basalts and the inverted pyroxene-ilmenite crystallization sequence in the olivine porphyritic basalts, regardless of the grain size of the rock, is suggestive of different mehcanisms other than cooling rate to explain these features. Origin of ilmenite rims around amalcolite in olivine porphyritic basalts Studies in reflected light on numerous basalt samples (El Goresy et al., 1974; Papike et al. , 1974) indicate that the ilmenite mantles around grey armalcolite grains are formed according to one or a combination of the following processes: 1.

Reaction between the cooling basaltic liquid and early crystallized armalcolite FeTi.O. + FeO (from melt)

EG-26

2FeTiO,

Figure EG-21.

Several idiomorphic armalcolite crystals displaying bireflection

and ilmenite reaction rims.

Olivine prophyritic ilmenite basalt.

Length of

field 200 microns.

2.

Reaction between chromian ulvospinel and armalcolite according to the idealized reaction: Fe2TiO/, + FeTi^Oj -*• 3FeTi03

3.

Reaction between metallic Fe° and armalcolite as suggested by Harzman and Lindsley (1973) 4FeTi205 + Fe° ->• 5FeTi03 + "TijOj" (solid solution in armalcolite)

4.

Breakdown of armalcolite to ilmenite and rutile

5.

Simple overgrowth of ilmenite around armalcolite.

The majority of the above-described processes may be present in the same basalt sample in Apollo 11 and Apollo 17 material. 1.

The reaction between the cooling basaltic liquid and armalcolite is

evidently the major process responsible for the formation of ilmenite rims around armalcolite in the TiC^-rich basalts of the Taurus-Littrow and Apollo 11 sites (Lindsley et al., 1974; El Goresy et al. , 1974).

Ilmenite formed

due to this reaction was apparently enriched in TiO. in contrast to primary

EG-2 7

ilmenites p r e c i p i t a t e d directly f r o m the basaltic liquid, since the ilmenite mantles usually show numerous rutile inclusions w h i c h probably exsolved on cooling.

M a n y armalcolite grains show reactions only on certain sides, namely

w h e r e the b a s a l t i c liquid had a free p a t h to the armalcolite crystal (El Goresy et at.,

1974; P a p i k e et at. , 1974).

This feature is indeed strong evidence

that reaction (1) is responsible for the formation of the ilmenite rims around armalcolite.

I n the fine-grained vitrophyres a few armalcolite grains display

no or little r e a c t i o n although these armalcolites w e r e n o t protected by other silicates.

This could b e due to the v e r y fast cooling of the basaltic liquid

and the deposition of pyroxene a n d plagioclase quench crystals before the reaction started. 2.

Textures strongly suggestive of reaction (2) w e r e observed in a few

lithic fragments a n d large basalts from the Apollo 17 landing site.

In an

ideal case p u r e ulvbspinel w o u l d react w i t h armalcolite to form ilmenite according to the equation F e 2 T i 0 4 + FeTi^Oj

3FeTiC>3

Since ulvbspinel in the A p o l l o 17 basalts is in a b r o a d sense a member of ulvospinel-chromite solid solution series, secondary titanian chromite w i l l p r e c i p itate

in a d d i t i o n to ilmenite as a result of this reaction.

Thus, the chromian

ulvbspinel w i l l change its composition according to the degree of the reaction. Normally, the boundaries of the original ulvbspinel grain are still visible (Fig. EG-22) w h e r e b y the newly formed chromite deposited b e t w e e n ulvbspinel and ilmenite.

I n advanced stages the reaction is also accompanied by exsolu-

tion of ilmenite f r o m the host ulvbspinel (Fig. EG-23).

The chromite is con-

fined to ulvbspinel boundaries w h e r e the reaction w i t h armalcolite took place. Drastic enrichment in MgAl^O^ in the secondary spinel took p l a c e due to this reaction (20.5 w t . % Al^O^ versus 9.3% in the original ulvbspinel).

Ilmenites

formed due to this reaction w e r e also probably rich in TiOj, since rutile exsolved from the ilmenites. 3.

H a r z m a n and Lindsley (1973) and Lindsley et at.

(1974) report that

armalcolite h e a t e d w i t h i n its stability field w i t h m e t a l l i c iron yields il4+ menite + a different armalcolite in w h i c h part of Ti is reduced to oxidize ss o 3+ Fe

and the Ti

et at.

p r o d u c e d enters the armalcolite as Ti^O^ component.

El Goresy

(1974) o b s e r v e d in m a n y lithic fragments textures strongly suggestive of

this reaction.

S e v e r a l armalcolite grains mantled by ilmenite w e r e observed

w i t h small iron globules at the boundary between the armalcolite core and the ilmenite mantle.

E l e c t r o n m i c r o p r o b e analyses of these armalcolites indicate

a significant enrichment of Ti compared to the coexisting armalcolite in the EG-28

Figure EG-22.

Armalcolite (center, gray) which reacted in its lower part with

ulvospinel to form ilmenite and secondary titanian chromite deposited between ilmenite and ulvospinel.

Original boundaries of ulvospinel' are still visible

as marked by small aligned silicate inclusions above the ulvospinel-ilmenite boundary.

Length of field 200 microns.

Figure EG-23.

A very advanced stage of reaction 2, gray at top is armalcolite

surrounded by secondary ilmenite which exsolved rutile (light gray). left is ulvospinel with ilmenite exsolutions.

Gray at

Big patches of secondary titan-

ian chromite (dark gray) are located between ulvospinel and armalcolite. Length of field 150 microns. EG-29

same lithic fragment.

The total cations (Ti is calculated as Ti^+) are also

slightly lower than those of coexisting armalcolites. 4. Pure ferropseudobrookite (FeTi^O^) decomposes to ilmenite and rutile at and below 1140 + 10°C (Harzman and Lindsley, 1973). Harzman and Lindsley also report that armalcolite of a given Fe/Mg ratio decomposes first to Mgenriched armalcolite + ilmenite + rutile.

This breakdown requires the presence

of ilmenite and rutile in almost 1:1 ratio.

Apollo 11 and 17 basalts were in-

spected for this reaction and only very few grains of armalcolite in the 17 material were found mantled by ilmenite and rutile satisfying this reaction. The breakdown of armalcolite to rutile + ilmenite is, however, a common phenomenon in Apollo 11 TiC^-rich basalts. 5.

In olivine porphyritic ilmenite basalts, ilmenite precipitates after

armalcolite and before titanaugite.

According to this crystallization sequence

simple overgrowths of ilmenite around pre-existing armalcolite should be expected in these rocks.

An important criterion to recognize this texture is that the

composite ilmenite-armalcolite grain does not show any resemblance to the armalcolite morphology.

However, this kind of overgrowth is quite rare compared to

reactions 1, 2, and 3. Study of equilibrium pjiase relations of synthetic TiC^-rich basalts as a function of oxygen fugacity (fc^) indicates that there is a direct relationship between the crystallization sequence and fo2 f o r basaltic liquids with the same composition (Usselman et al. , 1975).

The observed difference in the composi-

tion of armalcolites and the reversal of ilmenite and pyroxene in the crystallization sequence in plagioclase poikilitic and olivine porphyritic basalts was found to be a function of fOj (Usselman et al. , 1975) (Fig. EG-24).

Chemistry

of

armalcolite

Haggerty (1973) reports that tan armalcolite and gray armalcolite are compositionally indistinguishable in terms of major element abundances.

More

than 400 complete analyses (El Goresy et al. , 1974) strongly suggest that both tan and gray armalcolite are cation deficient since the number of cations calculated on the basis of 5 oxygens never totalled 3. The total number of cations for all armalcolites analyzed range from 2.91 to 2.97. According to Lind and Housley (1972) and Smyth (1973), armalcolite crystallizes in the space 4+ group Bbmm with the cations strongly ordered whereby Ti cations occupy the 8f(M„) and Fe2+ and Mg2+ cations are randomly distributed among the 4C(M1) 2+ sites. Following this model, the majority of the analyses revealed that Fe and Mg 2+ do not satisfy the 4C site occupancy since they never totalled 1,

EG-30

Figure EG-24.

Melting relations of

Apollo 17 sample 74275.

Triangular

points are those of O'Hara and Humphries (1975) at their stated oxygen fugacities.

The iron-wiistite (Fe-

FeO) curve is shown as reference.

although there is a full complement of almost two Ti cations per five oxygens 3+ (El Goresy et at. , 1974). Smyth suggested that Ti may be present as Ti and 2+ 3+ perhaps Cr as Cr . Wechsler et at. (1975) calculated 4-10% Ti, TiO- compo3+ nent for many lunar armalcolites, thus supporting the presence of Ti rather than cation deficiency of the armalcolite structure. The gray armalcolite variety is characterized by relatively higher Cr^O^ and MgO contents than the tan variety (El Goresy et at., Papike et at. et at.

1974).

However,

(1975) indicate that many gray armalcolites are zoned.

Papike

report a decrease in C ^ O ^ and increase in FeO content from the core

to the rim of an armalcolite grain.

The compositional variation of a zoned

crystal was found to overlap a major part of the separate fields assigned for tan and gray armalcolite. The Mg versus Fe and Mg versus Cr cationic distributions for tan and gray armalcolites are shown in Figures EG-25 and EG-26, respectively. tant features are recognized:

Two impor-

(a) There is indeed a compositional bimodal

distribution of tan and gray armalcolites with slight overlap of the fields; (b) the Mg-Fe substitutional relationship for the tan armalcolite variety is almost coherent with a negative slope, indicating that Fe is substituting for Mg.

Probably, in analogy to olivine, armalcolites which crystallized earlier

from the silicate melt have a higher Mg/Fe ratio than those crystallized later.

The gray armalcolites, regardless if they are mantled by ilmenite or

not, show generally higher Mg concentrations than tan armalcolite. EG-31

Compared

APOLLO 17

Mg

samples

osou 0.45

70017,125 70215,159 72015,21 74242,19 74243,4 79155,63 70135,60 71135,29

0.40

0.35-

030A Tan armalcolite + Gray a r m a l c o l i t e 0.25 0.35

Figure EG-25.

0.40

—i— 0.45

050

060 Fe

0.55

Mg-Fe cationic substitutional relationship (based on 5 oxygens)

for tan and gray armalcolite.

Mg 0.50

samples

0.45 "

A P O L L O 17

II,f 74242 19 74243 4 79155 63 70135,60 71135,29 .

AA

» M i Vf,

A

0.30 •

0.25 " A Tan armalcolite + G r a y armalcolite

JM

.040

fl45

"So"

£55

060 Cr

Figure EG-26.

Mg-Cr cationic substitutional relationship for tan and gray

armalcolites.

EG-32

to tan armalcolite the Mg-Fe substitutional relationship of the gray armalcolite is not coherent.

This scatter was interpreted (El Goresy et al. , 1974) as due

to enrichment of armalcolite in Mg resulting from reactions 1, 2, 3, or A described above.

Electron microprobe analyses indicate such strong preference

of Mg for armalcolite rather than for mantling ilmenite (El Goresy et al. , 1974). The compositional bimodality is also demonstrated in Figure EG-26 which shows the Mg-Cr substitutional relationship for both armalcolite types.

Despite

the scatter in the data points for tan armalcolite, this figure is suggestive of a positive relationship between Mg and Cr.

The Mg-Cr substitutional relation-

ship for gray armalcolite is coherent, indicating that these two elements have similar partitioning behavior.

Armalcolites in olivine porphyritic basalts have

higher MgO and C^O^ contents than tan armalcolites.

The data presented here are

also strongly suggestive of a partitioning of Mg and Cr between armalcolite and mantling ilmenite with strong preference of these elements for armalcolite.

Il-

menite rims around gray armalcolite cores were also analyzed with the microprobe (El Goresy et al., 1974), and the suggested partitioning is confirmed in the plot of MgO in armalcolite MgO in mantling ilmenite (Fig. EG-27).

C ^ O j in armalcolite C ^ O j in mantling ilmenite

The coherent positive slope of the data points indicates a

preference of both Mg and Cr for armalcolite.

This slope, however, may not

represent the actual partitioning relationship between armalcolite and the mantling ilmenite, since the majority of the ilmenite mantles exsolved rutile after their formation which may have caused an additional redistribution of

LUNAR ARMALCOLITES and coexisting ILMENITES

samples 70215,69 72015.21 74220,68 74242,19 74243,4

Figure EG-27.

MgO Arm./MgO Ilm.

versus Cr^O^ Arm./C^O^ Ilm. re-

A A

lationship for gray armalcolites

¿A

and mantling ilmenites. & Aá

2.0

25

lili

CrgO^ in Armalcolite Cr203 vi Ilmente

3D

EG-3 3

both Cr and Mg between ilmenite and rutile.

The partitioning of Cr and Mg

between ilmenite and rutile does not necessarily need to be similar to the partitioning of these two elements between armalcolite and ilmenite.

Thus,

gray armalcolite shows higher Mg and Cr concentrations than tan armalcolite. Reflection measurements on numerous armalcolite types from several Apollo 17 basalts indicate that the reflection curves of tan armalcolite show a continuous increase above 600 nm compared to a flat curve for gray armalcolite. This is responsible for the difference in color.

Armalcolite compositions

from various rocks are displayed in Table EG-2.

Table EG-2.

1 Si02 Ti0 2 Cr 2 0 3 AI2O3 V203 FeO MgO MnO CaO Total

0.33 70.44 1 .63 1 .72 0.18 18.43 5.47 0.08 0.57 98.85

2 0 .30 71 .61 1 .69 1 .77 . 0,.27 16..30 6,.63 0..09 0..44 99., 10

Chemical analyses of armalcolite.

3 0.08 74.13 2.00 2.10 0.10 14.08 7.86 0.14 0.04 100.53

4 0.08 73.91 2.01 1.99 0.05 14.44 7.75 0.17 0.03 100.43

5

6

7

74 .30 2..17 1 .93 ,

73,.00 1 .72 , 1 .96 ,

72 .50 1 .43 , 1 ,91 ,

13,.4 7..95 0..00

16,,30 6.,27 0..00

17.,60 5..32 0..00

99.,80

99,.20

98..80

1: 2: 3: 4: 5:

Tan armalcolite, Apollo 17 (El Goresy et al. 1974, Table 2, p. 641) Tan armalcolite, Apollo 17 (El Goresy et al. 1974, Table 2, p. 641) Gray armalcolite, Apollo 17 (El Goresy et al. 1974, Table 2, p.641) Gray armalcolite, Apollo 17 (El Goresy et al. 1974, Table 2, p. 641) Core of a gray armalcolite, Apollo 17 (Papike et al. 1974, Table 6, p. 492) 6: Same gray armalcolite few microns away from the core, Apollo 17 (Papike et al. 1974, Table 6, p. 492) 7: Same gray armalcolite just at the ilmenite rim, Apollo 17 (Papike et al. 1974, Table 6, p. 492)

Chromian ulvospinel Chromian ulvSspinel occurs in both basalt types not only as idiomorphic crystals but also as clusters of grains enclosed in olivine or pyroxene. both rock types ulvospinel is probably an early quench phase.

In

In coarse-

grained ilmenite basalts this early crystallized ulvospinel was not encountered; instead, tiny discrete late-stage ulvospinel grains were observed in the mesostasis.

The early crystallized ulvospinels in Apollo 17 basalts

resemble the Apollo 11 ulvospinel since their composition is intermediate EG-3 4

between chromite and ulvospinel.

The Apollo 17 ulvospinels, however, show con-

siderable solid solution towards M g A l ^ (Fig. EG-28).

Figure EG-28 displays

the composition of ulvospinels and chromite exsolution lamellae in ilmenite in the spinel prism.

It is noteworthy to mention that ulvSspinel shows systema-

tically higher MnO and V^O^ contents than the coexisting armalcolite. Fe2T104 APOLIO

17

SPINELS

M9AI2q,'

MgCr 2 0 4

Figure EG-28.

Composition of chromian ulvospinels in Apollo 17 basalts as pro-

jected in the multicomponent spinel prism (left); projection of the spinel prism base (middle); and projection of front rectangular face of spinel prism (right).

Ilmenite Ilmenite is the most abundant opaque mineral both in Taurus-Littrow and Apollo 11 basalts.

The amount of ilmenite in the studied samples range from

15% to 20% by volume.

Ilmenite occurs in both rock types as primary blocky

crystals and in olivine porphyritic basalts as reaction rims around armalcolite formed according to processes 1, 2, 3 or 4 as described above.

This latter

ilmenite type is designated secondary ilmenite to distinguish it from the former primary ilmenite which precipitated directly from the silicate melt. The majority of ilmenite formed by processes 1 and 2 show numerous rutile inclusions exsolved from the TiO^-rich secondary ilmenite on cooling.

Analy-

ses of primary and secondary ilmenites formed by processes 1 and 2 indicate a systematic difference in the composition of both ilmenites (El Goresy et at. , 1974).

Secondary ilmenites were found to contain significantly higher Cr^O^

contents (0.93-1.05%) than coexisting primary ilmenites (0.39-0.65%).

EG-3 5

Rutile Rutile occurs mainly as exsolution lamellae, as a product of reaction 4 described above, or is formed due to subsolidus reduction of magnesian chromian ilmenite to rutile + spinel + metallic Fe (El Goresy et al. , 1975).

Many of

these rutile grains show anomalous optical properties with prominent bireflection and anisotropism and a typical blue reflection color (El Goresy et al., 1975).

These features are characteristic of oxygen-deficient rutile.

The

majority of rutile grains were found to be highly enriched in ZrO^ (Table EG-3) regardless if they were formed by exsolution or subsolidus reduction reaction (El Goresy et al. , 1975). Table EG-3.

Si02 Ti02 Cr 2 0 3 AI 2 O 3 V2O3 FeO MgO MnO CaO Zr02 Total

Chemical analyses of rutile.

1

2

3

0.42 97.80 0.23

0.11 98.60 0.23

0.00 0.00

0.00 0.00

0.93 0.04

0.59 0.06

0.1 1 98.30 1.54 0.26 0.05 0.86

0.00

0.00

0.05 0.82 100.29

0.07 0.32 99.98

0.00

0.14 0.04 1.3! 102.61

1: Stoichiometric rutile, Apollo 17 (El Goresy et al. 1975, Table 4, p. 744) 2: Stoichiometric rutile, Apollo 17 (El Goresy et al. 197S, Table 4, p. 744) 3: Blue oxygen deficient rutile, Apollo 17 (El Goresy et al.1975, Table 4, p. 744)

Opaque oxides in anorthositic rocks and highland breccias On the basis of the occurrence of a few anorthositic lithic fragments in Apollo 11 fines, Wood et al. (1970) proposed that the Highlands of the moon are largely composed of anorthosite.

Prinz et al. (1973) demonstrated that

the anorthositic rocks range in composition and mineralogy from anorthositic to noritic and troctolitic. of the lunar crust.

These rocks very probably constitute a major part

Due to continuous bombardment of the lunar surface in the

past, the original textures of the rocks of this group are obscured and altered. Rocks belonging to this group were collected at the Luna 20 landing site (in the area of Apollonius C crater), Apollo 16 at the Descartes site, and at the EG-36

Apollo 14 (Fra Mauro) site.

Two major rock suites were recognized in the sam-

ples collected from the three landing sites (Prinz et at. , 1973b). 1.

ANT (anorthositic-noritic-troctolitic) suite which could be subdivided

to anorthosites, norites, and (spinel) troctolite). 2.

High-alumina basalt suite (> 45% < 60% plagioclase).

Opaque oxide minerals are rare and generally fine grained.

They also occur as

discrete rounded to angular grains in microbreccias and agglutinates and as rare euhedral to anhedral grains in crystalline fragments (El Goresy et at., 1973; Brett et at., 1973; Haggerty, 1973).

In the ANT suite members of the spinel

series are the most abundant opaque oxides followed by ilmenite and rutile. In the high-alumina basalt suite, ilmenite is quite frequent, usually occurring in the assemblages ilmenite-rutile-baddeleyite (Zr02) and ilmenite-rutile-baddeleyite-chromite (El Goresy et at. , 1973).

Spinets Two distinct spinel types occur in the highland rock suites:

(a) Members

of the spinel (MgA^O^-hercynite (FeAl^O^) series mainly encountered in the ANT suite, especially in spinel troctolites (Brett et at., 1973). ranges from 93 to 56 mole % MgAl^O^ (Brett et at. , 1973). ized by a low Ti02~content (0.02-1.0 wt %).

Their composition

They are character-

A minor amount of the chromite

component is also present but their compositions are clearly separated from the chromite-ulvospinel solid solution series.

(b) Members of the chromite-

ulvospinel series occurring as a minor component in the ANT suite but more abundant in the high-alumina basalt.

Although no optical zoning was observed

in these spinels, quantitative electron microprobe analyses indicate a zonational trend similar to the third trend present in mare-type basalts (El Goresy et at., 1973).

The zoning trend of these spinels shows, however, a sharp in-

crease in the C^Oj/(Cr^Oj+Al^O^) ratio from core to rim (El Goresy et at. , 1973) (Fig. EG-29).

The Cr/Al cationic relationship is similar to that found

in Luna 16 rocks (Haggerty, 1973).

Itmenite Ilmenites encountered in highland rocks contain appreciable amounts of the geikielite component just as do those from mare basalts.

The MgO-content

of ilmenite varies between 0.02 and 8 wt % (Brett et at., 1973).

Rutite Rutile is usually confined to oriented lamellae and irregular areas in

EG-3 7

Fe2Ti04 Mg2TiO.

LUNAR SPINELS APOLLO 16 SAMPLE 67075,45

MgAi2o4< 'fisCr204

MsCrjO, Figure EG-29.

Composition of spinels in an anorthosite sample as projected in

the multicomponent spinel prism (left); projection of the spinel prism base (middle) showing the crystallization trend of the spinels; and front rectangular face of the spinel prism (right) also depicting the trend with increase of Ti0 2 ~content.

ilmenites (Haggerty, 1973).

The mineral contains between 2.3 and 2.9 wt % FeO

and between 0.4 and 0.5 wt % Cr^O^.

SUBSOLIDUS REACTIONS

Lunar rocks differ from the majority of terrestrial counterparts mainly in their extremely low oxygen fugacities.

This feature is manifested by the

ubiquitous occurrence of metallic iron as a primary phase, the presence of textures suggestive of subsolidus reduction, and the extremely low amount of 3+ Fe

in silicate and oxide minerals.

Very probably, many of the lunar magmas

were already in a reduced state at the time of extrusion (Sato et at. , 1973). Considerable variation in the intensity of reduction has been reported among rocks of the same landing site (El Goresy et at. , 1972; Haggerty, 1972a) reflecting the complexity of the formation history of the lunar rocks and supporting the existence of various mechanisms for reduction.

An attempt to understand the

processes responsible for reduction reactions is due to Mao et at. (1974), who demonstrated that trapped solar wind hydrogen is probably an important reducing agent on the surface of the moon, especially during the flow of basaltic liquid on hydrogen-enriched regolith. EG-3 8

Subsolidus reduction reactions in the lunar rocks Evidence for five subsolidus reactions for which experimentally-determined buffer curves exist

was reported from the lunar samples of all landing sites.

The experimentally-determined buffer curves involve the following reactions (Sato et al., 1973; Williams., 1971; Taylor et al., 1972; Lindsley et al., 1974): 1.

Fe 2 Ti0 4 + FeTi0 3 + Fe + j0 2

2.

FeTiOj

3.

Fe„SiO. -»• 2Fe + SiO„ + 0„ 2 4 2 2

4.

2FeTi03

5.

FeTi 2 0 5 -»• Fe + 2Ti02 + ¿0

->• Fe + Ti0 2 + ¿0 2

F e T i ^ + Fe + |o 2

In applying these reactions, only estimations of the degree of reduction and temperature can be made, since the reduced lunar minerals are by no means pure compounds as compared to the experimentally-investigated synthetic phases. Furthermore, the host minerals subjected to reduction change their composition continuously as the reaction proceeds to the right-hand side of the above equations.

These compositional variations were reported by El Goresy et al.

(1972) and Haggerty (1972a) who demonstrated that chromian ulvBspinel changes its composition continuously upon subsolidus reduction to become enriched in the chromite component.

In comparison to equation 2, lunar magnesian ilmenites

break down to rutile + spinel + Fe (Haggerty, 1973a).

The activity of Fe 2 SiO^

in lunar fayalite decreases continuously upon reduction to Fe + Si02-

Host

ilmenite also changes its composition to become enriched in the geikielite component.

Reactions 4 and 5 were found to be induced upon flow of basaltic

liquid on regolith loaded with solar wind hydrogen (Mao et al., 1974) and hence are regarded as exogenic reactions.

The reduction described by Mao et al.

(1974) indicates that extensive reduction of lunar basalts probably first took place on the surface of the moon during the eruption process. Figure EG-30 displays schematically the oxygen fugacity curves of the univariant reactions listed above, after Sato et al. (1973) and Taylor et al. (1972), and extrapolated down to 700°C.

Among the univariant reactions de-

scribed, the I-Q-F curve holds a key position.

The quartz-fayalite-iron

(I-Q-F) univariant curve intersects the ulvbspinel-ilmenite-iron (Fe-il-uv) curve at 950C and the ilmenite-rutile-iron (Fe-ru-il) curve at roughly 830°C (Taylor et al. , 1972).

Any understanding of the temperature and reduction his-

tory of the lunar samples requires the study of all assemblages involved in the first three reactions.

Apollo 14 samples 14053 and 14072 were previously

EG-39

800'C

700 *c

*K

Figure EG-30.

ITTO

96

92

900*C

S8

8A IO^/T'K

1000"C

80

IMOt

7.6

7.2

1200*0

i I! s

68

Oxygen fugacity curves for the reduction reactions 1-5.

taken from Sato et al. (1973) and Taylor et al. (1972). below 830°C is extrapolated and hence is uncertain.

Data

Path of fugacity curves

Symbols are:

Fe or I =

iron, Wii = wlistite, il = ilmenite, fb = ferropseudobrookite, and ru = rutile.

considered to be the most reduced basalts, due to the presence of the breakdown of fayalite to silica + Fe in addition to the extensive reduction of chromian ulvospinel to Ti-chromite + ilmenite + Fe (El Goresy et al., 1972; Haggerty, 1972a).

Taylor et al. (1972) conclude from their experimental investigations

that the presence of fayalite reduction in these two samples is direct evidence that these two rocks have undergone more reducing conditions than any other rocks, terrestrial or lunar

reported from Apollo 11 through Apollo 14.

A

comparison between the breakdown assemblages in these two rocks and in several TiO^-rich basalts from the Taurus-Littrow region as well as the application of the first three reactions negates a severe reduction.

Table EG-4 displays the

assemblages involved in the univariant reactions 1 to 3 in sample 14053 as compared to assemblages in Apollo 17 samples 70017, 70035, and 70135.

Table EG-4.

Comparison

Assemblages in reduced lunar basalts. 70017;

14053 1. Ulvospinel * ilmenite+Fe 2. Fayalite — » Fe+silica 3. Incipient breakdown of ilmenite (only very few grains)

70035;

70135

1. Ulvospinel ? ilmenite+Fe 2. Ilmenite * spinel+rutile+Fe 3. No fayalite breakdown

EG-40

between the assemblages in 14053 and in the three Apollo 17 basalts strongly suggests that the rocks of the two landing sites had different thermal histories.

Very probably, samples 14053 and 14072 were subjected to reduction

during a heating event to a temperature in excess of 830°C and/or close to 950°C.

A mechanism of reduction upon heating to these temperatures at the

given oxygen fugacity explains best the simultaneous breakdown of fayalite to Fe + silica and ulvBspinel to ilmenite + Fe.

Incipient breakdown of ilmenite

to rutile + Fe may suggest that the opaque fayalite assemblages were placed below the Fe-ru-il fugacity curve above 830°C and very probably close to 950°C in order to account for the extensive subsolidus reduction of both fayalite and chromian ulvbspinel followed by rapid cooling, thus preventing further ilmenite breakdown.

Textures and mineralogy of sample 14053 do not support

reduction during an initial cooling process.

Fayalite in this sample always

occurs in a mesostasis assemblage, which is widely accepted to be the last assemblage to crystallize from the silicate melt.

Reduction of fayalite must

have taken place after solidification of the mesostasis assemblage, since the breakdown texture still displays the morphology of fayalite grains (El Goresy et al. , 1972). Apollo 17 samples 70017, 70035, and 70135 display extensive subsolidus reduction of chromian ulvospinel to Al-Ti chromite + ilmenite + Fe and of Mg ilmenite to Al-Ti chromite + rutile + Fe.

Accessory fayalite in these samples

does not show any sign of reduction to Fe + silica.

On the basis of the per-

vasive reduction of ilmenite alone, and neglecting fayalite (El Goresy et al. , 1972; Taylor et al., 1972; Haggerty, 1972a), one may conclude that these samples were severely reduced.

Such a conclusion is unrealistic.

The three as-

semblages shown in Table EG-4 suggest that the reduction of these rocks took place below 830°C, probably during cooling.

These assemblages may have been

reduced to a fQ^ below the Fe-ru-il curve, but above the I-Q-F curve.

This

is only possible below 830°C, where the fayalite buffer curve intersects the ilmenite buffer curve. It is worthwhile to mention that the above-given temperatures are subject to some correction since the activities of Fe„TiO. and FeTiO, in the lunar 2 4 3 ulvospinel and ilmenite, respectively, are by no means equal to 1. Haselton and Nash (1975) discussed a model (formerly proposed by Lindsley et al. , 1974) of isochemical reaction of opaque oxides with part of Ti present 3+ as Ti

in the initial phases to explain the subsolidus reactions observed.

An important aspect of this model is the fact that the components of the assemblages are indeed not in their standard states. EG-41

Such a system would display

deviation from ideality.

This deviation from stoichiometry indicates that the

f02 curve will not be a straight line.

Experimental studies in the system Fe-

Ti-0 with assemblages coexisting with metallic Fe

(Simons, 1973) indeed indi-

cate a compositional variation as a function of temperature.

The mechanism

presented by Haselton and Nash (1975) could explain the textures in the assemblage armalcolite-ilmenite-rutile-Fe°.

However, it could hardly explain the 3+ breakdown features of chromian ulv8spinel since Ti was not detected in an ulvospinel-magnetitess coexisting with metallic Fe° between 1000 and 1300°C (Simons, 1974). Nature of reducing agent in Apollo 17 basalts The causes for the reduction of lunar basalts have been a subject of debate since recovery of the Apollo 11 samples.

Several mechanisms were proposed:

(1)

internal redistribution of valence of states of Fe, Ti, and Cr by assuming in3+ 2+ itial presence of Ti and Cr in the melt (Brett et at., 1971,1972); (2) vacuum pumping of oxygen (O'Hara et at.,

1970; Biggar et at.,

1971; Ford et

at.,

2-

1972; Haggerty, 1972a); (3) sulfur loss from lunar magmas as S

(Brett, 1975);

(4) alkali volatilization from lunar magmas during extrusion (O'Hara et 1970; Biggar et at.,

1971,1972, Ford et at.,

at.,

1972; (5) reduction by carbon or

carbon monoxide at a depth shallower than a few kilometers (Sato et at.,

1973).

Vacuum pumping of oxygen could never be a plausible cause of reduction because of the preferential escape of Fe vapor at low oxygen fugacities (Sato et at. , 1973).

Nor could alkali loss explain the reduced state of the lunar

basalts since escape of alkalis as elements would cause oxidation, whereas volatilization of alkalis as oxides (Na20) would not change the oxidation state of lunar basalts (Sato et at. , 1973).

The mechanisms of sulfur loss from lunar

magmas or of redistribution of the valence of states of Fe, Ti, and Cr may well explain the precipitation of metallic iron from the lunar basaltic liquids, but they are unable to account for subsolidus reduction reactions of iron-titanium oxide assemblages. El Goresy et at.

(1975) propose that gaseous activity took place after

solidification of the lunar basalts.

They found evidence for genetic relation-

ship between this gaseous activity and subsolidus reactions of the opaque oxides. Many of the open cracks penetrating silicate and oxide minerals as well as cleavage in pyroxenes in several Apollo 12, Apollo 15, and Apollo 17 samples were found to be filled with metallic iron.

The metallic iron forms a network

system of veins across several mineral grains along several hundred microns. El Goresy et at.

(1975) believe that these features exclude the possibility

EG-4 2

that iron liquid has been injected in the crack system.

These observations can

be explained as due to deposition of iron from a gaseous phase which permeated these rocks after crystallization.

Ilmenite grains penetrated by such iron-

filled cracks display textures typical of subsolidus reduction to spinel + rutile + Fe°. These reduction products always originate from the cracks radiating into the ilmenite grains.

El Goresy et al. (1975) consider these features as strongly

suggestive that a single event is responsible for deposition of iron in the cracks and reduction of opaque oxides.

A gaseous mixture of CO and carbonyle

iron compounds (Fe(CO),.) was proposed to have initiated these reactions.

Upon

cooling, many of those complex gaseous compounds become unstable and break down at different temperatures to Fe°+ CO.

Upon breakdown, metallic iron would fill

many of the open cracks, whereas carbon monoxide released would account for the reduction processes.

The breakdown and the reduction processes may have con-

tinued during cooling down to 200°C.

It is also plausible that these gases may

have been released from the same magma chamber from which the basalts originated and permeated those rocks after solidification and formation of tension cracks. The reduction process is probably not the result of a release of solar wind hydrogen during an impact process since no evidence of reheating or shock features in the plagioclase or ilmenite were found.

All these observations do indeed

support the reduction mechanism proposed by Sato et at. (1973) and emphasize the possibility of endogenic gaseous activity during eruption of lunar lavas.

REFERENCES Anderson, A. T. , T. E. Bunch, E. N. Cameron, S. E. Haggerty, F. R. Boyd, L. W. Finger, 0. B. James, K. Keil, M. Prinz, P. Ramdohr, and A. El Goresy (1970) Armalcolite: A new mineral from the Apollo 11 samples. Proa. Apollo 11 Lunar Soi. ConfGeochim. Cosmochim. Acta, Suppl. 1, 1, 55. Bence, A. E., W. Holzwarth, and J. J. Papike (1972) Petrology of basaltic and monomineralic soil fragments from the Sea of Fertility. Earth Vianet. Soi. Lett. 13, 299. Biggar, G. M., M. J. O'Hara, A. Peckett, and D. J. Humphries (1971) Lunar lavas and the achondrites: Petrogenesis of protohypersthene basalt in the mare lava lakes. Proo. Lunar Soi. Conf. 2nd, 617. Brett, R. (1975) Reduction of mare basalts by sulfur loss (abstr.) Science VI, 89. The Lunar Science Institute, Houston.

In, Lunar

, P. Butler, Jr., C. Meyer, Jr., A. M. Reid, H. Takeda, and R. J. Williams (1971) Apollo 12 igneous rocks 12004, 12008, 120022: A mineralogical and petrological study. Proo. Lunar Soi. Conf. 2nd, 301. El Goresy, A., P. Ramdohr, and L. A. Taylor (1971) The opaque minerals in the lunar rocks from Oceanus Procellarum. Proo. Lunar Soi. Conf. 2nd, 219.

EG-4 3

El Goresy, A., P. Ramdohr, and L. A. Taylor (1972) Fra Mauro crystalline rocks: Mineralogy, geochemistry and subsolidus reaction of opaque minerals. Proo. Lunar Sci. Conf. 3rd, 333. site:

,

, and 0. Medenbach (1973) Lunar samples from the Descartes Opaque mineralogy and geochemistry. Proo. Lunar Sci. Conf. 4th, 733.

, rich basalts: 5th, 627.

, , and H.-J. Bernhardt (1974) Taurus-Littrow Ti02Opaque mineralogy and geochemistry. Proo. Lunar Sot. Conf.

and P. Ramdohr (1975) Subsolidus reduction of lunar opaque oxides: Textures, assemblages, geochemistry, and evidence for a late-stage endogenic gaseous mixture. Proo. Lunar Sci. Conf. 6th, 729. , M. Prinz, and P. Ramdohr (1976) Zoning in spinels as an indicator of the crystallization histories of mare basalts. Proc. Lunar Sci. Conf. 7th, in press. Ford, C. E., G. M. Biggar, D. J. Humphries, G. Wilson, D. Dixon, and M. J. O'Hara (1972) Role of water in the evolution of lunar crust; an experimental study of sample 14310; an indication of lunar calc-alkaline volcanism. Proo. Lunar Sci. Conf. 3rd, 207. Haggerty, S. E., F. R. Boyd, P. M. Bell, L. W. Finger, and W. B. Bryan (1970) Opaque minerals and olivine in lavas and breccias from Mare Tranquillitatis. Proc. Apollo 11 Lunar Sci. Conf. , 513. and H. 0. A. Meyer (1970) Apollo 12: Sci. Lett. 9, 379.

Opaque oxides.

Earth Planet.

(1972a) Apollo 14: Subsolidus reduction and compositional variations of spinels. Proc. Lunar Sci. Conf. 32*d, 305. (1972b) Luna 16: An opaque mineral study and systematic examination of compositional variations of spinels from Mare Fecunditatis. Earth Planet. Sci. Lett. 13, 328. (1972c) Solid solutions, subsolidus reduction and compositional characteristics of spinels in some Apollo 15 basalts. Meteoritics 7, 353. (1973a) Apollo 17: Armalcolite paragenesis and subsolidus reduction of chromian-ulvBspinel and chromian-picroilmenite (abstr.). ESS (Trans. Am. Geophys. U.) 54, 593. (1973b) Armalcolite and genetically associated minerals in the lunar samples. Proo. Lunar Sci. Conf. 4th, 111. (1973c) Luna 20: Mineral chemistry of spinel, pleonast, chromite, ulvospinel, ilmenite, and rutile. Geoohim. Cosmochim. Acta 37, 857. Harzman, M. J. and D. H. Lindsley (1973) The armalcolite join (FeTi205-MgTi205) with and without excess Fe°: Indirect evidence of TiJ on the moon (abstr). Ann. Meeting Geol. Soc. Am. 5, 593. Haselton, J. D. and W. P. Nash (1975) Observations on titanium in luna oxides and silicates (abstr.). In, Lunar Science VI, 343. The Lunar Science Institute, Houston. Hodges, F. N. and I. Kushiro (1974) Apollo 17 petrology and experimental determination of differentiation sequences in model moon compositions. Proc. Lunar Sci. Conf. 5th, 1, 505. Kesson, S. E. (1975) Mare basalts: Melting experiments and petrogenetic interpretations. Proc. Lunar Sci. Conf. 6th, 921. EG-44

Knorring, 0. V. and K. G. Cox (1961) Kennedyite, a new mineral of the pseudobrookite series. Mineral. Mag. 32, 672. Kushiro, I., Y. Nakamura, and S. Akimoto (1970) Crystallization of Cr-Ti spinel solid solutions in an Apollo 12 rock, and source rock of magmas of Apollo 12 rocks (abstn%. Am. Geophys. U. Ann. Meeting, 64. Laul, J. C. and R. A. Schmitt (1973) Chemical composition of Apollo 15, 16, and 17 samples. Proe. Lunar Soi. Conf. 4th, 1349. Lindsley, D. H., S. E. Kesson, M. J. Hartzman, and M. K. Cushman (1974) The stability of armalcolite: Experimental studies in the system MgO-Fe-Ti-O. Proa. Lunar Soi. Conf. 5th, 1, 521. LSPET (Lunar Sample Preliminary Examination Team) (1974) Preliminary Examination of Lunar Samples, pp. 7.1-7.46. L. B. Johnson Space Center. Mao, H. K., A. El Goresy, and P. M. Bell (1974) Evidence of extensive chemical reduction in lunar regolith samples from the Apollo 11 site. Proo. Lunar Soi. Conf. 5th, 673. Marvin, U. (1975) The perplexing behavior of Niobium in meteorites and lunar samples. Meteoritios 10, 452. Muan, A., J. Hauck, and T. Lofall (1972) Equilibrium studies with a bearing on lunar rocks. Proo. Lunar Soi. Conf. 3rd, 1, 185. Nehru, C. E., M. Prinz, E. Dowty, and K. Keil (1974) Spinel-group minerals and ilmenite in Apollo 15 rake samples. Am. Mineral. 59, 1220. O'Hara, J. M. , G. M. Biggar, S. W. Richardson, and C. E. Ford (1970) The nature of seas, mascons, and the lunar interior in the light of experimental studies. Proo. Apollo 11 Lunar Soi. Conf. , 695. and D. J. Humphries (1975) Armalcolite crystallization, phenocryst assemblages, eruption conditions and origin of eleven high titanium basalts from Taurus-Littrow (abstr.). In, Lunar Science VI, p. 616-618. The Lunar Science Institute, Houston. Papike, J. J., A. E. Bence, and D. H. Lindsley (1974) Mare basalts from the Taurus-Littrow region of the moon. Proc. Lunar Soi. Conf. 5th, 1, 471. Prinz, M., E. Dowty, K. Keil, and T. E. Bunch (1973a) Mineralogy, petrology and chemistry of lithic fragments from Luna 20 fines: Origin of the cumulative ANT suite and its relationship to high-alumina and mare basalts. Geochim. Cosmochim. Acta 37, 979. , , , and (1973b) Spinel troctolite and anorthosite in Apollo 16 samples. Science 179, 74. Ramdohr, P. and A. El Goresy (1970) Opaque minerals in the lunar rocks and dust from mare Tranquillitatis. Science 167, 615. Sato, M., N. L. Hickling, and J. E. McLane (1973) Oxygen fugacity values of Apollo 12, 14 and 15 lunar samples and reduced states of lunar magmas. Proo. Lunar Sci. Conf. 4th, 1061. Simons, B. (1974) Zusammensetzung und Phasenbreiten der Fe-Ti-Oxyde in Gleichgewicht mit metallischem Eisen. Diplomarbeit, Technisohe Eoohschule, Aachen, 104. Smyth, J. R. and P. R. Brett (1973) The crystal structure of armalcolites from Apollo 17 (abstr.) Ann. Meeting Geol. Soc. Am. 5 (7), 814. (1974) The crystal chemistry of armalcolites from Apollo 17. Earth Planet. Soi. Lett. 24, 262.

EG-45

Taylor, L. A., G. Kullerud, and W. B. Bryan (1971) Opaque mineralogy and textural features of Apollo 12 samples and a comparison with Apollo 11 rocks. Proc. Lunar Sci. Conf. 2nd, 855. , R. J. Williams, and R. H. McCallister (1972) Stability relations of ilmenite and ulvospinel in the Fe-Ti-0 system and applications of these data to lunar mineral assemblages. Earth Planet. Sci. Lett. 16, 282. Dsselman, T. M. (1975) Ilmenite chemistry in mare basalts, an experimental study. Origin of mare basalts and their implications for lunar evaluation (abstr.). In, Lunar Science, 164. The Lunar Science Institute, Houston. __

and G. E. Lofgren (1976) Phase relations of high-titanium rare basalts as a function of oxygen fugacity (abstr.). Lunar Science VII, 888. The Lunar Science Institute, Houston.

Wechsler, B. A., C. T. Prewitt, and J. J. Papike (1975) Structure and chemistry of lunar and synthetic armalcolite (abstr.) In, Lunar Science VI, 860. The Lunar Science Institute, Houston. Williams, R. J. (1971) Reaction constants in the system Fe-Mg0-Si02-02 at 1 atmosphere between 900°C and 1300°C: Experimental results. Am. J. Sci. 270, 334.

EG-4 6

OPAQUE OXIDE MINERALS in METEORITES Ahmed El

Goresy

Chapter 6 INTRODUCTION Records of stones falling from the sky can be traced back to classical Chinese and ancient Greek or Latin literature.

The stone of Ensisheim in

Alsace, France, is the oldest preserved meteorite fall.

A stony meteorite

weighing 127 kg fell on November 16, 1492 in Ensisheim and the event was recorded by chroniclers of the town in a detailed illustrated description.

How-

ever, it was only after the meteorite shower of L'Aigle, France which fell on April 26, 1803 that the majority of the scientific community finally accepted the extraterrestrial origin of meteorites, due to the convincing demonstration by J. B. Biot to the members of the Academie Française that the L'Aigle stones fell from the sky.

Ever since that time the scientific community all over the

world has continuously and carefully recorded meteorite falls.

The increasing

interest in the study of meteorites, especially in the last 100 years, is mainly due to the fact that these objects document the wide variety of solar system processes, e.g.,

(1) processes in the solar nebula

prior to formation of

planets; (2) processes in planet-like bodies, analogous to the earth; and (3) those resulting from collisional events between interplanetary objects creating shock and fragmentation.

Little is known about the exact source of meteorites,

although it is widely accepted that they probably come from the asteroid belt. The wide variation in the compositions of the primary objects from which meteorites were derived is well demonstrated by the variations in meteorite compositions.

The most commonly used classification of meteorites among

petrologists is the scheme of Mason (1962) which is based on that of Prior (1920).

This classification is shown in Table EG-5.

Chondrites and achondrites

could be grouped together as stony meteorites since silicate and oxide minerals comprise the major part of these meteorites.

The distinction between the two

categories of stony meteorites is straightforward:

"chondriti.c" meteorites

contain spherical silicate or oxide objects called chondrules whereas these spherical objects are absent in achondrites.

However, a few stones considered

as chondrites actually contain no chondrules but are so classed because they are chemically and mineralogically similar to the chondrule-bearing stones of the same type (Mason, 1962).

The chondrites are the most abundant of all

EG-4 7

Table EG-5.

Classification of meteorites.* Subclass

Class

Enstatite chondrites Olivine-bronzite chondrites Olivine-hypersthene chondrites Carbonaceous chondrites

I. Chondrites

A. B. C. D.

II.¿chondrites

A. Calcium-poor achondrites 1. Enstatite achondrites (aubrites) 2. Hypersthene achondrites (diogenites) 3. Olivine achondrites (chassignites) 4. Olivine-pigeonite achondrites (ureilites) B. Calcium rich achondrites 1. Augite achondrites (angrites) 2. Diopside-olivine achondrites (nakhlites) 3. Pyroxene-plagioclase achondrites a) Eucrites b) Howardites

III. Stony irons

A. B. C. D.

IV. Irons

A. Hexahedrites B. Octahedrites 1. Coarsest octahedrites 2. Coarse octahedrite 3. Medium octahedrites 4. Fine octahedrites 5. Finest octahedrite C. Nickel-rich ataxites

*From Mason,

meteorites.

Olivine stony irons (pallasites) Bronzite-trydimite stony irons (siderophyres) Bronzite-olivine stony irons (lodranites) Pyroxene-plagioclase stony irons (mesosiderites)

1S62, 196?

A metal phase occurs in the majority of meteorites; it consists of

two nickel-iron alloys:

(a) Alpha iron or kamacite is an iron-nickel alloy with

constant composition of about 5.5% Ni; it crystallizes in a body-centered cubic lattice.

(b) Gamma iron, or taenite is a nickel-iron alloy of variable compo-

sition ranging from about 27 to about 65% Ni; it crystallizes in a face-centered cubic lattice.

Most iron meteorites are mixtures of both kamacite and taenite.

The classification of iron meteorites into hexahedrites, octahedrites, and Nirich ataxites is mainly based on the configuration of kamacite and taenite in a polished etched surface.

Hexahedrites consist of large crystals of kamacite

with cleavage parallel to the faces of a cube.

In octahedrites the orientation

of kamacite and taenite bands are parallel to octahedral planes.

EG-48

This structure

is also known as "Widmanstatten pattern" after its discoveror, Baron Alois von Widmanstatten.

The texture of the pattern is in direct but inverse re-

lationship with the Ni content of the meteorite. the finer the octahedral structure.

The higher the Ni content

As the nickel content of octahedrites in-

creases, the bands of kamacite become narrower.

At 12-14% Ni, they become ex-

tremely narrow and discontinuous, and the Widmanstatten pattern disappears. Meteorites of this type are classified as nickel-rich ataxites. In stony meteorites, chondrites and achondrites, opaque oxides occur as accessory minerals usually evenly dispersed in the silicate groundmass.

In

chondrites the abundance, assemblages, and textures of opaque oxide minerals in chondrules {e.g. , in many ordinary chondrites) or in Ca, Al-rich inclusions (e.g. , carbonaceous chondrites) can vary drastically from those in the groudmass of the same meteorite.

In iron meteorites, opaque oxides are extremely

rare compared to the mass of meteorites in which they are encountered.

The

opaque oxides occur intergrown with silicates in silicate-rich inclusions or with troilite (stoichiometric FeS) and graphite which also occurs as macroscopic inclusions in the NiFe alloy groundmass. Members of the spinel group (chromite, spinel, and to a lesser extent magnetite) are the most abundant opaque oxides in meteorites. rutile constitute only a small fraction of the oxides.

Ilmenite and

Members of the pseudo-

brookite series have never been encountered in meteorites.

Enstatite chon-

drites and enstatite achondrites appear to be barren of opaque oxides. MINERALOGY Spinel group minerals Chromite is by far the most abundant end member of the spinel series in meteorites.

Chromite occurs in various quantities in more than 90% of all

stony meteorites (chondrites and achondrites) and is the most abundant oxide accessory in mesosiderites, palasites, and iron meteorites (El Goresy, 1965). However, it is rare or almost absent in carbonaceous chondrites.

Instead, the

groundmass of carbonaceous chondrites is especially enriched in magnetite (Jedwab, 1971; Ramdohr, 1973).

In a few calcium-rich achondrites (Nakhlites)

titanomagnetite is the dominant spinel (Bunch and Reid, 1975; Boctor et at., 1976).

Spinel (MgA^O^) occurs as a minor component especially in a few

chondrules in ordinary chondrites (Ramdohr, 1973); however, the mineral is a major constituent of Ca and Al-rich inclusions in some carbonaceous chondrites (Sztrokay, 1960; Christophe Michel-L6vy, 1969; Marvin et at., 1970).

EG-49

Many of

these Ca- and Al-rich inclusions enriched in melilite, spinel, pyroxene, and perovskite are widely accepted as high-temperature condensates from the primordial solar nebula (Grossman, 1975). Based on textures and assemblages, Ramdohr (1967,1973) recognized the following various types of chromites in ordinary chondrites:

(1) coarse chromite;

(2) clusters of chromite aggregates; (3) pseudomorphous chromite; (4) exsolution chromite; (5) chromite chondrules, and (6) myrmekitic (or symplectitic) chromite. The number of these groups would increase markedly, if subtle morphological details were taken into account. 1.

Coarse chromite:

This chromite type occurs in the overwhelming majority

of ordinary chondrites, achondrites, stony irons, and iron meteorites.

It is

present as coarse euhedral to subhedral grains interlocked in the silicate matrix.

Coarse chromite exhibits idiomorphlc features only if in contact with

troilite or metallic iron (Ramdohr, 1973). usually anhedral.

Its boundaries to silicates are

According to Ramdohr (1973), coarse chromite is later than

olivine and pyroxene in the crystallization sequence.

Planimetric integration

of 73 chondrites (Keil, 1962) indicates that the chromite content of chondrites varies between 0.01 and 0.61 wt %.

No apparent relationship was found between

the chromite content and the chondrite type (L = low iron, or H = high iron groups).

Exsolution lamellae of ilmenite along (111) planes of coarse chromite

is a rare feature in chondrites, but not rare in chromites in mesosiderites or pallasites (Fig. EG-31).

Figure EG-31.

Mount Padbury (stony iron), Australia.

A grain of coarse chromite

in pyroxene with rutile exsolution lamellae (white); length of photograph 150 microns (from Ramdohr, unpublished). EG-50

2.

Clusters of chromite aggregates:

ordinary chondrites.

This type seems to be restricted to

The cluster consists of medium- to fine-grained idio-

m o r p h i c to subhedral grains of chromite usually embedded in plagioclase EG-32).

(Fig.

This chromite type is also later in the crystallization sequence than

olivine and pyroxene.

Figure EG-32.

Nardoo

(olivine-bronzite chondrite), N e w South Wales.

Cluster

of aggregate chromite, w h i t e is kamacite; length of photograph 350 microns (from Ramdohr, 1973).

3.

Pseudomorphous chromite:

This chromite type appears to b e restricted

to chondrules in ordinary chondrites. and iron meteorites.

It is absent in achondrites, stony irons,

The assemblage in w h i c h this type occurs consists chiefly

of albitic plagioclase and chromite, a n d sometimes m i n o r clinopyroxene 1967,1973). volume.

(Ramdohr,

The chromite content of such chondrules may b e as h i g h as 40% by

This chromite occurs as fine grains (less than 10 microns in diameter)

frequently oriented in fan-shaped laths of a former unknown m i n e r a l w h i c h broke down to albite + chromite (Fig. EG-33)

(Ramdohr, 1973).

characteristic for breakdown.

(1976) suggested the precursor is very

Ramdohr

The texture is indeed

probably ureyite (kosmochlor), N a C r S i 2 0 g , to account for the 1:3 ratio of chrom i t e to albite.

Y o d e r and K u l l e r u d (1971) report that at 700°C and 2 kb albite

+ chromite, as w e l l as albite + eskolaite ( C ^ O ^ ) , w e r e found to b e stable.

How-

ever, a m i x t u r e of kosmochlor + anorthite + enstatite reacted at the same conditions to form an albitic plagioclase + eskolaite + chromite +

EG-51

clinopyroxene.

Figure EG-33.

Bachmut (olivine-hypersthene chondrite), Ukraine.

A chondrule

with pseudomorphous chromite ill lath-shaped radiating pseudomorphs in albite (dark gray) matrix; length of photograph 1.2 mm (from Ramdohr, 1967).

Yoder and Kullerud (1971) thus propose that an explanation for the assemblage chromite + albite (+ minor clinopyroxene) could be that kosmochlor is a stable product or metastable quench product that reacts with anorthite as well as bronzite and possibly olivine to form albitic plagioclase, chromite, and a clinopyroxene. 4.

Exsolution chromite:

This type is encountered in clinopyroxene-rich

chondrules as fine-grained clusters of chromite at the boundaries between clinopyroxene and plagioclase.

Ramdohr (1973) proposes exsolution from a chromium-

rich silicate to account for this assemblage. 5.

Chromite chondrules:

Though uncommon in chondrites, the chromite type

is spectacular with regard to its possible origin and the chemical variation of chondrules.

This type of chondrule was first discovered by Ramdohr (1967).

The chondrules consist usually of a two-phase assemblage: clase.

chromite and plagio-

The chromite:plagioclase ratio can vary drastically, and in some cases

almost pure chromite chondrules are encountered, e.g. , Loot and Burdette meteorites (Ramdohr, 1967).

The chromite occurs in concentric alternating layers

of different grain size, compactness, and plagioclase content.

The outermost

shell of the chondrule is usually surrounded by a thin feldspar layer of fairly

EG-5 2

uniform thickness, overgrown bv FeNi alloy and troilite.

Some chondrules exhibit

an atoll-like feature with a plagioclase core and a uniform chromite shell, e.g. , Harrisonville chondrite.

A very common type of chromite chondrule is charac-

terized by the presence of large idiomorphic chromite crystals in a plagioclase groundmass which, in turn, is loaded with numerous small anhedral chromite grains dispersed throughout the plagioclase (Fig. EG-34).

The chondrule is

sealed by an almost continuous shell of chromite, which in turn is separated from the meteorite groundmass by a continuous layer of plagioclase (Fig. EG-34).

Figure EG-34. Texas.

Plainview (polymict olivine-bronzite chondrite), Hale County,

Chromite chondrule with idiomorphic chromite crystals,

fine-grained

chromite-plagioclase intergrowth, and chromite rim; length of photograph 450 microns (from Ramdohr, 1967).

The amount of chromite usually exceeds 50% by volume of the chondrule. common type is shown in Figure EG-35.

A less

Although chromite and plagioclase are

the major constituents, clinopyroxene occurs in minor amounts.

A major part

of the chromite occurs as skeletal crystals, indicative of quenching due to rapid cooling.

These chondrules also contain pseudomorphous chormlte, indica-

ting a genetic link between the two chromite types (Fig. EG-35).

The pre-

dominance of chromite and plagioclase in these chondrules document the unusual composition of the liquid from which the chondrules were derived.

EG-5 3

Figure EG-35.

Mangwendi (polymict olivine-hypersthene chondrite), Rhodesia.

A

chromite chondrule with skeletal crystals of chromite and pseudomorphous chromite; length of photograph 450 microns (from Ramdohr, 1973).

6.

Myrmekitic (symplectitic) chromite:

to mesosiderites and iron meteorites. and chromite (Fig. EG-36). observed (Ramdohr, 1973).

This type seems to be restricted

The assemblage consists of major olivine

So far, neither plagioclase nor pyroxene has been Texture and assemblage are indicative of eutectic

intergrowth of olivine and chromite. Magnetite is by far the major opaque oxide in the groundmass of numerous carbonaceous chondrites.

Usually, it occurs as reaction rims around metallic

FeNi alloy or troilite (Ramdohr, 1973).

However, the main mass is present as

spherules of various sizes dispersed in the groundmass (Fig. EG-37).

A detailed

study with the scanning electron microscope (Jedwab, 1971) indicates that these magnetite spherules are in fact composed of stacking platelets or spirals suggestive of condensation from the vapor phase.

This interpretation would indi-

cate that these magnetites are one of the latest components to condense directly from the cooling solar nebula. Titanomagnetites occasionally occur in Ca-rich achondrites (Bunch and Reid, 1975; Boctor et al. , 1976); they usually exhibit ilmenite lamellae parallel to (111) of the magnetite and very fine lamellae of ulvospinel parallel to (10Q) of the magnetite host.

These features document the close similarity between

nakhlites and many terrestrial igneous rocks (Boctor et al. , 1976). EG-54

Figure EG-36.

Vaca Muerta (mesosiderite), Chile.

Myrmekitic (symplectic) in-

tergrowth between chromite and olivine; length of photograph 800 microns (from Ramdohr, unpublished).

Figure EG-37.

Esebi (carbonaceous chondrites), Zaire.

Magnetite spherules in

carbonaceous and silicate groundmass; length of photograph 300 microns (from Ramdohr, unpublished).

EG-55

Hmenite-geikielite-pyrophanite series Members of this series were reported from ordinary chondrites, stony irons, and iron meteorites (Ramdohr, 1963,1973; Snetsinger and Keil, 1969; El Goresy, 1965; Bunch and Keil, 1971).

They seem to be absent in carbonaceous chondrites

and are very rare in achondrites.

In nakhlites, ilmenite g s usually occurs as

lamellae along with ulvospinel in titanomagnetite (Boctor et at., 1976).

Only

in very few iron meteorites is it also present as exsolution lamellae in coarse chromite.

In contrast to terrestrial igneous and metamorphic rocks, meteoritic

ilmenite never displays titanohematite lamellae.

Rutile lamellae are very rare

but reported to be frequent in some mesosiderites, e.g., Vacu Muerta (Ramdohr, 1973).

Many of the ilmenites reported from meteorites were found to exhibit

twin lamellae presumably formed by shock due to meteorite collisions in space (Fig. EG-38) (Ramdohr, 1973).

Figure EG-38.

Adlie Land (olivine-hypersthene chondrite), Antarctica.

Xeno-

morphic ilmenite with shock-induced twin lamellae, white is troilite; length of photograph 600 microns (from Ramdohr, 1973).

The amount of the components geikielite (MgTiO^) and pyrophanite (MnTiO^) in ilmenite is usually higher in chondritic ilmenites than in nonchondritic (achondritic, stony irons) ilmenites (Snetsinger and Keil, 1969; Bunch and Keil, 1971).

The FeO content of ilmenite in ordinary chondrites tends to

EG-56

increase with increasing FeO/(FeO+-MgO) in coexisting olivine, but MgO tends to decrease (Snetsinger and Keil, 1969).

Rutile Rutile is quite rare in ordinary chondrites and achondrites and seems to be absent in carbonaceous chondrites.

However, it is a frequent accessory

mineral in the majority of mesosiderites (Ramdohr, 1965; Marvin and Klein, 1964) and many pallasites.

The mineral is also not rare in silicate inclusions in

iron meteorites (El Goresy, 1965,1971).

Busek and Keil (1966) report that among

chondrites, rutile is most abundant in the Farmington chondrite.

In chondrites

and stony irons the mineral has a colorless to pale blueish appearance in reflected light with frequent internal reflections.

Both in the Farmington

chondrite and Vaca Muerta mesosiderite it occurs as lamellae in ilmenite (Ramdohr, 1965; Busek and Keil, 1966).

However, in several mesosiderites rutile

lamellae in chromite were also reported (Ramdohr, 1965; Busek and Keil, 1966). Ramdohr (1964) suggested that rutile occurring in the assemblage rutile-ilmenitea F # w a s probably formed due to the subsolidus reaction 2FeTi03 •+• 2Ti02 + 2Fe° + 0 2 El Goresy (1965,1971) reports that rutile present in iron meteorites has anomalous optical properties with a greenish color in reflected light.

These rutiles

are characterized by a relatively high Cr^O^ (1.23%) and NbO^ (2.93%) contents. OXIDE ASSEMBLAGES IN VARIOUS METEORITE GROUPS Chondrites The classification of chondritic meteorites in the four subclasses given in Table EG-5 is based mainly on the mineralogical composition of the meteorites. The abundance of Fe, Mg, Si, and 0 altogether in chondrites is very high and may total 90% in the majority of chondrites (Wasson, 1974).

As mentioned before,

in contrast to terrestrial rocks the majority of meteorites contain metallic FeNi alloy.

Prior (1916) noted that "the less the amount of Ni-Fe in chondritic

stones, the richer it is in Ni, and the richer in Fe are the magnesium silicates." These two relationships are known as Prior's rules. of view, Prior's rules are expressions of the

From the petrological point

conditions under which mete-

orites were formed, since the prevailing oxygen fugacity at a given temperature will control the partitioning of iron between the coexisting silicates and the metal phase.

In reduced chondrites, e.g., enstatite chondrites, most of the

EG-5 7

iron is present in the metal phase, and the Ni content of the metal is correspondingly low.

With increasing degree of oxidation more of the iron is oxi-

dized; the iron content of the silicates increases, and since the amount of metal decreases while the amount of N i remains constant, the Ni content of the metal increases.

Prior's rules should be considered as qualitative and hold only for

ordinary chondrites (but not for enstatite or carbonaceous chondrites) because the fractionation in the Fe/Si ratio discovered by Urey and Craig (1953) demonstrated that the ordinary chondrites could not be understood in terms of such a simple model.

Urey and Craig (1953) demonstrated the existence of two groups

of chondrites, one having a low total iron content (22.33 wt 7.), the other a high content (28.58wt %) which they designated the L group respectively.

and the H

group,

Determination of the fayalite content of olivine in some 800

chondrites (Mason, 1963) and the iron and magnesium distribution in coexisting olivines and rhombic pyroxenes (Keil and Fredriksson, 1964) established the existence of three major groups of ordinary chondrites:

(1) high iron (H)-

group (olivine-bronzite chondrites); (2) low iron (L)-group

(olivine-hypersthene

chondrites); (3) low iron-low metal (LL)-group (olivine-hypersthene

chondrites).

Figures EG-39 and EG-40 demonstrate the existence of these three groups as

0

2

4

£

8 i0 13 14 16 16 20 MOL' PER

Figure EG-39. (Fe+Mg) in

CENT

24 26 28 30 32

IN OLIVINE

Ratios of Fe/(FefMg) in olivine plotted against ratios of Fe/

rhombic pyroxene for 86 chondrites (from Keil and Fredriksson, 1964). EG-5 8

i

1

1

1

1

'

40 30 •

L L GROUP 4 • jmm ao

L GROUP H GROUP

c ^ c f f f e o 8 0 "a.

k ! 20 10

0

0-2

i 0'6

0-4

0-8

FeVFe Figure EG-40.

Plot of mole percent of fayalite in olivine versus Fe°/Fe ratios

for ordinary chondrites (from Van Schmus and Wood, 1967).

obtained from the distribution of Fe and Mg among coexisting olivines and pyroxenes (Fig. EG-39) and the mole percent fayalite in olivine versus Fe°/Fe ratios in the meteorite.

This classification, however, does not account for the

degree of equilibration among silicates.

Van Schmus and Wood (1967) introduced

a simplified classification using two parameters: (degree of equilibration). eter classification.

chemical and petrologic

Figures EG-41 and EG-42 summarize this two param-

Increasing number indicates increase in the degree of

equilibration, e.g. , LL3 means low to medium equilibrated LL chondrite, whereas LL6 is a highly equilibrated LL chondrite. Spinels.

Bunch et al. (1967) carried out a detailed electron microprobe

survey of chromite chemistry in ordinary chondrites as a function of chemistry and texture (equilibration) of the chondrites.

Their study, however, is re-

stricted to the coarse chromite, and hence the variability in chemistry of the other chromite types described by Ramdohr (1967,1973) is not known.

The com-

positional variability of chromites in equilibrated chondrites (subgroups 5 and 6) was found to be within the precision of the method except for Al and Ti. The largest compositional variabilities were observed in chromite in unequilibrated chondrites (subgroups 3).

Zoning in chromites in unequilibrated chon-

drites is also a frequent feature (Bunch et al. , 1967).

These results indeed

support the classification introduced by Van Schmus and Wood (1967).

A direct

correlation exists between chromite composition and the classification of

EG-59

Petrologio t y p e 1

2

3

E

El

E2

li 3

C

CI

C2

C3

H

HI

4

4

16

C5

LI

L2

L3

H4

H5

LL1

LL2

LL3

H6

35

74

L4

L5

9 LL

6 C6

2

H3 7

L

E6

2

C4

8

H2

6

E5

4

1*

Chcmical group

5

E4

18

43

LL4

LL5

4

44 L6

152 LL6

3

7

21

* N u m b e r of examples of each meteorite t y p e now known is given in its box.

Classification of chondrites.

Figure E6-41.

Number of each meteorite type now

known is given in its box (from Van Schmus and Wood, 1967). Petrologie type 1 E

|

•

2

3

»

1 1 Carbonaceous clion dritcs

C H

*

L

*

"

1

H3J

* *

LL

4

5

j

L3}

Hyporsthene chondrites

LL3*

Amphoteric chondrites

6

t

• Unpopulated •f Ordinary chondritos. J Unoquilibratod ordinary chondrites.

Figure EG-42.

|

Enstatito chondritos i 1 • 1 04 ! • > : [. Bronzite chondrites

Location of meteorite type in the petrological chemical classi-

fication (from Van Schmus and Wood, 1967).

equilibrated chondrites into H, L, and LL groups.

From Table EG-6 it is ap-

parent that FeO and TiC^ increase and C^O^, MgO, and MnO decrease from H to L to LL groups.

Bunch et al. (1967) did not include chromites from subgroups 3

and 4 since they were found to vary in composition and are apparently not in equilibrium with the silicates.

Correlation between chromite composition and

classification of equilibrated chondrites becomes more evident if major oxide contents are compared to the iron oxide content of coexisting olivine expressed as mole percent FeO/(FeO+MgO) (Fig. EG-43). are well separated.

In most cases, H, L, and LL groups

Both FeO and Ti09 in chromite increase from subgroups 3 to

EG-60

Table EG-6 Composition of chromite from equilibrated chondrites.*

CR 2°3 AI2O3 V

2°3 TI02

H group (10)**

L group (7)

LL group (6)

56.9

56.1

54,,4

5.9

5.3

5..7

0.68

0.72

0,.73

2.33

2.81

3,.23

31.2

FeO

33.0

34,,5

MgO

2.66

1.99

1.,62

MnO

0.94

0.74

0,.63

100.61

100.66

100,.81

Total

Analyses for subgroups Numbers in parentheses

Figure EG-43.

5 and 6; all from Bunch et al. 1967, Table indicate number of meteorites analysed

1, p. 1571

Correlation between chromite composition and classification of

equilibrated H-, L - , and LL-group chondrites (subgroups 5 and 6).

With increase

of FeO/(Fe04-Mg0) in coexisting olivine, FeO and TiO^ of chromite increase (a and b) and Cr.O, and MgO decrease (c and d) (from Bunch et al. , 1967).

EG-61'

5 in H and L groups.

However, FeO in chromite decreases slightly from subgroups

5 to 6, whereas TiC^ continues increasing (Bunch et al. , 1967).

The classifica-

tion of Van Schmus and Wood (1967) implies that their petrographic subgroups within a major chondritic group {e.g., H3 through H6) represent metamorphic series, i.e., mineral compositions in H6 chondrites were established by metamorphism of H3-type assemblages after chondrite formation.

Keil and Fredriksson (1964), Ramdohr

(1967), and many others indicate, however, that chemical and mineralogical equilibration may well be attained during a primary process {e.g., by slow cooling). However, Bunch et at. (1967) indicate that relations between chromite compositions and subgroup classification are consistent with either a metamorphic or a primary origin of chondrites. Spinels found in Ca-, Al-rich inclusions in carbonaceous chondrites coexist with a Ti-rich fassaite, gehlenite, anorthite and perovskite (sometimes with hibonite, Ca. ((Al,Ti)„,0,o) 2 24 Jo

(Marvin et at., 1970; Keil and Fuchs, 1971).

Two

types of coarse-grained Ca-, Al-rich inclusions with spinel as a major constituent were reported from the Allende meteorite (Grossman, 1975) .

Type A con-

tains 80-85 percent melilite, 15-20 percent spinel and 1-2 percent perovskite. Clinopyroxene, if present, is usually restricted to thin rims around inclusions or surrounding cavities in the interior (Grossman, 1975).

Type B contains 35-60

percent clinopyroxene, 15-30 percent spinel, 5-20 percent plagioclase, and 5-20 percent melilite.

The main differences between the two types of inclusions are

abundance and composition of clinopyroxene. -4 phases condensing at 10 sequence:

Thermodynamic calculations for

atmospheres in the solar nebula indicate the following

corundum, hibonite, perovskite, gehlenite, spinel, Fe-metal, diopside,

forsterite, Ti^O^., anorthite, enstatite, rutile, albite, and nepheline (Grossman, 1972). The spinel-bearing Ca-, Al-rich inclusions are considered to be early condensates from the solar nebula.

Spinels occurring in these inclusions in Allende

are almost pure MgAl^O^ but contain traces of FeO, C ^ O ^ , TiO^, and CaO (Marvin et at. , 1970; Grossman, 1975; El Goresy, unpublished). CaO, and TiO^ contents are well below 1 percent.

The Cr 2 0 3 and most FeO,

No systematic difference in

composition was found between spinels enclosed in melilite and those enclosed in pyroxene.

Figure EG-44 displays the Cr^O^ content versus TiO^ content of

spinels in type A and type B inclusions (Grossman, 1975). Itmenite.

Snetsinger and Keil (1969) indicate that the average composition

of ilmenite from equilibrated ordinary chondrites is, like chromite, related to chondrite classification:

FeO increases and MgO and MnO decrease from H to L

to LL groups (Table EG-7).

These relationships are displayed in Figure EG-45

EG-62

1.0h o o

CM 0.5

i i i i I i i i i | i i i i | i i i i ALLENOE SPINEL COMPOSITIONS d 5 Type A Inclusions ° 2 Type B Inclusions o o

H

m

h 0

a o • o SpD PI . ! I I I I ' I I I I I I I I I 0.5 1.0 1.5 WT. %

Ti02

2.0

contents Figure EG-44. Cr 2°3 an •t 3

4 5 40 41 42 WEIGHT PERCENT IN ILMENITE

43

44

45

Figure EG-45. Mole percent FeO/(Fe+MgO) in olivine versus weight percent of MgO, MnO and FeO in ilmenite of equilibrated chondrites (from Snetsinger and Keil, 1969).

where FeO/(FeCH-MgO) in olivine is plotted against MgO, MnO, and FeO in coexisting ilmenite.

Although the statistics are poor due to the rarity of ilmenite, com-

positional trends similar to those observed in chromites can be recognized.

The

correlation indicates that the four phases olivine, orthopyroxene, chromite, and ilmenite did indeed equilibrate in these chondrites. Achondrites Chromite is the most abundant opaque oxide in achondritic meteorites, followed by ilmenite and then rutile (Ramdohr, 1973).

Chromite compositions in

diogenites, chassignites, and several eucrites have been published by Ramdohr and El Goresy (1969), Bunch and Keil (1971), and Lovering (1975).

Spinel com-

positions in Angra dos Reis (angrite) were recently reported by Keil et al. (1976).

Chromian titanomagnetites from nakhlites were also reported by Bunch

and Reid (1975), and Boctor et al. (1976).

A detailed study of mineral assem-

blages in different clasts in the Kapoeta Howardite was recently published by Dymek et al. (1976).

Bunch and Keil (1971) report compositional variability

(particularly for Ti and Al) of chromites from achondrites (without specifying the meteorite group), although Lovering (1975) indicates a complete chemical homogeneity of all chromites he analyzed in the Moama eucrite.

EG-64

Bunch and Keil

(1971) indicate some distinctions between chromites in diogenites and chromites in eucrites:

Chromites in diogenites are higher in Al^O^ and MgO, whereas

chromites in eucrites are higher in TiO^, FeO, and V^O^.

Spinel in Angra dos

Reis is a magnesian, chromian hercynite (Keil et at., 1976).

The howardite

Kapoeta was found to contain various basaltic clasts which could be grouped into two broad lithologic types on the basis of modal mineralogy—(a) basaltic (pyroxene- and plagioclase-bearing) and (b) pyroxenitic (pyroxene-bearing) (Dymek et at. , 1976).

Type b clasts are unusually enriched in opaque oxides

(chromite and ilmenite).

Spinel compositions in various types of achondrites

(other than nakhlites) are given in Table EG-8. recognized in Table EG-8:

Two important features can-be

(1) Chromites in achondrites do not show any corre-

lation due to the strong variability of mineralogy and history of the various subclasses; and et at. (1976)

(2) comprehensive studies similar to those presented by Dymek of additional achondrites

are needed for better understanding

of phase petrology. The nakhlites are characterized, compared to other meteorites, by the presence of titanomagnetite with ilmenite and ulvBspinel exsolution lamellae (Bunch and Reid, 1975; Boctor et at., 1976).

Ilmenite in the Lafayette mete-

orite was found to have broken down to rutile + hematite.

Boctor et al. (1976)

report that the maximum calculated hematite content corresponds to a temperature of about 740°C and fg the Skaergaard gabbro.

of 10 ^ .

Such values are similar to those obtained for

Boctor et al. conclude that the phases they reported

from the Lafayette meteorite represent the late stages of differentiation of the original toagmas and that the parent body from which Lafayette was derived may have undergone major primary differentiation of a nature similar to that believed operative in the early crystallization history of the lunar crust. Ilmenite.

Zoning and minor grain-to-grain compositional variability of

ilmenite in several achondrites was reported by Bunch and Keil (1971).

Achon-

dritic ilmenite is usually depleted in MgO and MnO compared to ilmenite in ordinary chondrites.

Stony irons Chromites in pallasites tend to occur as coarse to very coarse grains (e.g. , up to a few centimeters in the Brenham pallasite) either coexisting with olivine or completely embedded in the iron mass.

They are usually characterized

by their low TiO,, content (0.18% average) but variable Cr 2 0 3 (60.5 to 69.0%) and A1 2 0 3 (1.5 to 9.1%) contents (Bunch and Keil, 1971).

Compositional variability

of chromites in mesosiderites is more pronounced than in pallasites. EG-65

The TiO.

Table EG-8.

Cr 2°3 A1203 V

2°3 Ti02 FeO

Fe 2°3 MgO

MnO Total

Composition of spinels from various

1

2

3

55.50

46 .10

10.10

9,.80

0.43

0,.28

1 .09 28.80

4

5

6

7

8

48..90

3,.30

38.87

40 .25

47.35

46.18

9,.30

54,.50

4.92

5 .33

7.49

9.19

0,.75

0..07

0.58

0,.38

0.34

0.24

0,.65

10.86

11,.01

3.12

5.57

36,.50

35..70

28,.40

40.57

41,.45

36.04

36.61

2..86

0,.59

8,.00

0,.71

2.78

3,.70

4,. 10

3..40

-

3.90

achondrites.

0.68

0..54

0,.59

0,.18

100.50

99..97

99.,93

98.,50

-

0.43 1 .04 97.26

-

0..95 100..08

1 .06 98.26

-

1.02 0.98 99.79

1: Average chromite composition in diogenites (Bunch and Keil, 1971, Table 4, p. 149) 2: Chromite composition in Chassigrry (chassignites) (Bunch and Keil, 1971, Table 5, p. ISO) 3: Average chromite composition in eucrites (Bunch and Keil, 1971, Table 4, p. 149) 4: Spinel composition in Angra dos Reis (Angrites) (Keil et al., 197S, Table 1, p. 444) 5: Chromite composition in basaltic clast A in Kapoeta Howardite (Dymek et al., 1976, Table 1, p. 1117) 6: Chromite composition in pyroxenitic clast B in Kapoeta Howardite (Dymek et al., 1976, Table 2, p. 1118) 7: Chromite composition in fine grained pyroxenite clast C in Kapoeta Howardite (Dymek et al., 1976, Table 3, p. 1120) 8: Chromite composition in fine grained povphyritic basalt clast P in Kapoeta Howardite (Dymek et al ., 1976, Table 4, p. 1121)

EG-66

content of chromites in mesosiderites is relatively higher than that of chromites in pallasites (Bunch and Keil, 1971).

Table EG-9 shows a comparison between chro-

mite compositions in pallasites, mesosiderites and lodranites.

Compared to chro-

mites in other stony irons, chromites present in lodranites are unique due to their high ZnO content (0.78-1.68%).

In this respect, they are analogous to

chromites in iron meteorites. Compositions of ilmenites in mesosiderites are generally similar to ilmenite compositions in achondrites.

However, ilmenites in mesosiderites tend to be

higher in MgO, MnO, and Cr^O^ than ilmenites in achondrites (Bunch and Keil, 1971). Iron meteorites Chromites in iron meteorites occur both in the metal groundmass and together with silicates in troilite and graphite inclusions (El Goresy, 1965).

In many

troilite inclusions it coexists with the thiospinel daubreelite, FeCr^S^ (El Goresy, 1965).

This documents the chalcophile and lithophile behavior of chro-

mium in the same meteorite.

In the metal groundmass chromites always form per-

fect euhedral crystals with sharp crystal faces to the iron.

In troilite inclu-

sions with silicates, chromite exhibits idiomorphic boundaries only against

Table EG-9.

Cr.O. 2. 3 A1

2°3

V

2°3 Ti02

Composition of chromites in pallasites, mesosiderites, and lodranites.

1

2

3

64.00

52.00

61.83

5.60

11 .50

4.43

0.54

0.54

0.46

0.18

I .84

0.91

FeO

23.20

31.00

22.28

MgO

5.80

2.29

6.47

MnO

0.65

0.77

1.10

ZnO Total

-

-

100.13

99.94

1.27 98.75

1: Average chromite composition in pallasites (Bunch and Keil, 1971, Table 7, p. 152) 2: Average chromite composition in mesosiderites (Bunch and Keil, 1971, Table 7, p. 152) 3: Average chromite composition in Lodran (lodranits ) (Bild, unpublished analyses).

EG-67

troilite but not against silicates (Fig. EG-46).

This feature may indicate

equilibration between chromite and coexisting rhombic pyroxene and olivine in the early history of the meteorite.

Bunch et at. (1970) classified silicate

inclusions in iron meteorites into several groups:

(1) Odessa type; (2) Copiapo

type; (3) Weekero Station type; and three other "miscellaneous" types: Kendall County, and Netschaevo.

Enon,

The classification introduced by Bunch et at.

(1970) is based on mineralogical composition, mineral abundance, texture, and shape of inclusion.

Due to the problem of small sample populations, such

classifications are difficult and somewhat ambiguous.

Chromite is common in

Odessa and Toluca (El Goresy, 1965; Bunch et at. , 1970) but was not found in several members of the Copiapo-type inclusions.

Chemistry of chromites in

the majority of iron meteorites is unique compared to chondritic, achondritic, pallasitic, and mesosideritic (except for Lodran) chromites.

Usually, they

contain appreciable amounts of MnO (up to 4.2%) and ZnO (up to 2.31%).

This

again puts some constraints on the geochemical behavior of both Mn and Zn in iron meteorites.

In the same inclusions these chromites coexist with alabandite,

MnS, and sphalerite, ZnS (El Goresy, 1965).

This establishes both the chalco-

phile and lithophile behavior of these two elements in iron meteorites.

Com-

positions of chromites from various iron meteorites are given in Table EG-10.

Figure EG-46.

Mundrabilla (coarse octahedrite), Australia.

silicate-bearing troilite nodules.

Chromite in

Note sharp boundaries to troilite and

subhedral features towards silicates;' length of photograph 200 microns (from Ramdohr, unpublished).

EG-68

Table EG-iO.

Composition of chromite in various iron meteorites.

1

Cr

2°3 ai2o3 V

2°3 Si02

2

3

4

69.40

71 .90

71 .70

68 .40

2.51

1 .24

0.42

10,.90

0.31

0,.26

0.57

0,.68

0.21

-

0..02

Ti02

1 .01

0..40

0.48

FeO

7.00

12,.60

15.10

2.,50

MgO

16.00

10..20

7.10

14.,20

MnO

2.22

2.,28

3.40

4.,20

ZnO

1.37

1 ,39 .

1.70

0.,02

100.03

100. 27

100.47

100. 88

Total

1: Average of 25 grains in Mundrabilla (Ramdohr et at., unpublished data). 2: Odessa chromite (Bunch et al., 1970, Table 7, p. 314) 3: Copiap.o chromite (Bunch et al., 1070, Table 7, p. 314) 4: Kendall County chromite (Bunch et al., 1970, Table 7, p. 314)

The very high Cr^O^ content (except in Kendall County) is striking.

In Weekero

Station, Colomera, Kodiakanal, Enon, Kendall County, and Netschaevo, chromites are characterized by their relatively high Al 0 content (between 2.68 and +2 -4-9 .9 17.6%). Figure EG-47 displays the Fe /(Fe ^+Mg in olivine versus Fe /

0.5

c I

0.4

-

0.3

o u> s

V u. •r

.Chassigny

4L

0.2

4H

0.1

Mundrabilla • 0.2

PaltasitesD Silicate i Netschaevo Inclusions^

8

0.4

0.6

1.0

0.8

Fe'2/(Fe*2+Mg) in Chromite Figure EG-47.

+2

Fe

+2

/(Fe

+2

+Mg) in olivine versus Fe

+2

/(Fe

+Mg) in chromite in

Chassigny, LL-, L-, H-chondrites, pallasites, silicate inclusions in iron meteorites (modified from Bunch et al., 1970).

EG-69

+2 +2 (Fe +Mg ) in coexisting chromite with chondrites, pallasites, Chassigny, and silicate inclusions in iron meteorites.

Both chromite and olivine in iron

meteorites show the lowest Fe/(Fe+Mg) ratios compared to other meteorite groups. Also evident is the decreasing Fe/(Fe+Mg) trend from Chassigny (achondrite) to LL, L, H chondrites, pallasites, and iron meteorites. Rutile in iron meteorites occurs as separate grains in silicate-rich troilite inclusions or with chromite in silicate-free nodules (El Goresy, 1965). Rutile in iron meteorites is unique compared to rutiles in chondrites (Busek and Keil, 1966), since El Goresy (1971) found that rutile in iron meteorites is usually enriched in Cr.O, and NbO. (Table EG-11).

Table EG-11.

Composition of rutile in iron meteorites and Vaca Muerta mesosiderite.x

1

2

3

4

92.48

95.58

Ti0 2

95.06

95.10

Nbt>2

2.93

2.89

1 .63

FeO

0.73

0. 10

I .00

2.27

MgO

Fe^ + , or to the rectangle on the right if Fe^1" > Al; this second plot is specific to the divalent ion The scale values represent the number of cations in tetrahedral and octahedral coordination for the spinel formula based on 32 oxygens and 24 cations.

Hg-105

Figure Hg-25 Mineral Morphology (a)

Dark gray chromian-spinel core mantled by titanomagnetite which is partially oxid along the peripheral margins to titanomaghemite (white).

(b)

0.09 mm.

Euhedral primary pseudobrookite crystal containing a cylindrical silicate glass i: elusion with an attached crystal of chromian-titanomagnetite.

0.09 mm.

(c)

Skeletal ilmenite with glass and finely crystalline inclusions.

(d)

An ilmenite crystal with glass inclusions and with renewed second generation grow: of T-shaped crystallites along the basal plane.

(e) - (h)

0.09 mm.

0.09 mm.

Skeletal growth morphologies of titanomagnetite illustrative of a progressi'

trend towards euhedral morphology.

This trend is classified as the cruciform

typ
2, and TiC^; these distributions are apparent in the electron microprobe element distribution x-ray scanning images shown in Figures Hg-26c-f for a chromian spinel core mantled by titanomagnetite of the type comparable to the assemblage shown in Figures Hg-25a and Hg-26a.

These mantles result by the reaction of early formed chromite with an Fe-Ti rich

liquid and contrast with those chromites which are included in either olivine or pyroxene where the silicate-enclosed crystals commonly exhibit very thin reaction mantles or no mantles at all, as illustrated in Figure Hg-26b.

For kimberlites, the basaltic trend of

early Cr, Mg, Al and later Fe + Ti is also observed (Fig. Hg-26g-h); but the paragenetically later mantling phase can also become the cores for nucleation and growth of picroilmenite as shown in Figure Hg-26i.

In rare instances for groundmass spinels, but commonly 3+ in garnet kelyphitic rims, the cores show a preferred enrichment of Mg, Al, and Fe , with 2+ intermediate zones which are enriched in Cr and Fe , and with outermost zones which are Ti enriched.

These complex distributions are illustrated in Figure Hg-26j for a ground-

mass crystal, and in Figures Hg-26k and 1 for two crystals in kelyphite. Figure Hg-27 illustrates a variety of textures between chromites and silicates, and chromites and titanomagnetites.

Silicates, glass and sulfide inclusions are shown in

Figure Hg-27a; the effects of magmatic corrosion and resorption are in Figures Hg-27b-c; mantling by titanomagnetite in Figures Hg-27a-c; coprecipitating growth of chromite + olivine in Figure Hg-27; the oxidation at high temperatures (>600°C) of chromite + titanomagnetite in basalts (Figs. Hg-27e-f); and the alteration associated with serpentinization in partially decomposed harzburgite in Figures Hg-27g-h.

Additional discus-

sions of the relationships of these overgrowths and of the oxidation trends are to be found in Table Hg-19 and in Chapter 4.

Neither the resorption characteristics, nor the

high-temperature oxidation assemblages of chromian spinels have received attention in the literature which is in contrast to the reaction of chromite to titanomagnetite

{e.g.,

Evans and Moore, 1968; Beeson, 1976) and of the alteration of chromite to ferritchromite and magnetite {e.g.,

Onyeagocha, 1974; Engin and Aucott, 1971; Springer, 1974; Ulmer,

1974; Hamlyn, 1975; Bliss and Maclean, 1975).

Hg-109

Figure Hg-26 Chromian Spinel^ (a) Asymmetrically mantled, chromian-spinel cluster partially enclosed in olivine. Thi mantles are titanomagnetite and the intermediate zones are chromian-titanomagnetit< The adjacent discrete crystal of titanomagnetite, although incomplete, would be classified as the cruciform type.

0.16 mm.

(b) Euhedral spinel core mantled by titanomagnetite.

M t ^ are also present along the

crack and in patches associated with inferred cracks below the polished surface. 0.16 mm. (c-f)

X-ray scanning images obtained by electron microprobe illustrating the distribut:

of major elements for a chrome spinel-titanomagnetite core-mantle relationship. T1 intensity of the spots are approximately proportional to concentration.

The elemei

displayed are Fe (in c), Ti (in d), Cr (in e) and A1 (in f); the core is thus virti ally Ti-free (d) and the mantle virtually Cr (e) and Al-free; both the mantle and : core contain Fe (c).

0.11 mm.

(g-h) Multiple zoning in chromian-spinels from kimberlites with chromite cores and magnetite mantles,

g = 0.15 mm; h = 0.08 mm.

(i) Euhedral magnetite core, epitaxially overgrown by picroilmenite.

0.09 mm.

(j) Oscillatory zoning in spinel where the white areas are Mg, A1 and Fe 2+ darker areas are Cr+Fe enriched, and the mantle is Ti enriched.

3+

enriched, tl

(k-1) Compositionally similar to (j), but these spinels are present in garnet kelyphiti rims associated with Ti-phlogopite.

k = 0.09 mm; 1 = 0.07 mm.

Hg-110

Figure Hg-26

Hg-111

Chromian Spinel

(a)

Reactions

Glass inclusions in a phenocrystic chromian spinel mantled by titanomagnetite; an island of chromite in the glass suggests magmatic corrosion. titanomagnetite.

(b)

The lighter mantle

0.14 mm.

Coarse web-shaped chromian spinel with partially crystalline, and dominantly glas inclusions.

Note that the extent of the titanomagnetite mantle is a function of

size of chromite and that the internal cores have a similar morphology to the out' edges.

0.16 mm.

A symplectic internal core of chromian spinel mantled by successive barriers of

(c)

later chromite and an outer margin of titanomagnetite; most of the cuniform segme: in the core are also mantled and the smallest areas contain the largest mantles o Mt (d)

ss

.

0.15 mm.

Euhedral chromian spinel core mantled by subgraphic chromite + glass + olivine, overall outer-morhpology of the symplectite is broadly similar to that of the cor 0.16 mm.

(e)

Irregular glassy inclusions in chromite mantled by titanomagnetite which has unde gone oxidation "exsolution" and subsequent partial decomposition to R + Hem gs . 0.16 mm.

(f)

The dark central core is chromian spinel, the outer assemblage is oxidized titano magnetite, and the white attached areas were I l m ^ but are now R + H e m ^ .

The li,

oriented {111} trellis lamellae were also originally Ilm residual host Mt g s rods. (g-h)

but are now R + Hem ; ss ss are still apparent, and these contain dark, oriented pleonaste

0.15 mm.

These two examples illustrate the features typical of "ferritchromit" which res

during the serpentinization of chromite in ultramafic rocks. 3+ Fe

The outer mantles a

-rich chromian magnetites, and the exchange chemistry is complex and varied

(see Table Hg-21).

0.15 mm.

Hg-112

Figure Hg-27

Hg-113

Figure Hg-28 Ilmemi te-Hemat i te ss (a-b) The light-gray hosts are titanohematite and the oriented darker-gray lenses are ferrian-ilmenite.

In (a) individual grains show

centrations of ilmenite at the grain boundaries.

sharp terminations and large con In (b) the grains are rounded,

and these preferred concentrations are not as evident.

Each of the larger ilmenit

lenses is surrounded by a depletion zone, and the regions between the large lenses are occupied by finer lenses which are assumed to have formed at T°C lower than those of the larger bodies. (c-d)

This distribution is the synneusis texture.

0.15 mm.

The cores of these Hemgg grains contain similar distributions in H m s s lenses to

those shown in (a-b), but here the development of thick mantles of ilmenite has re suited from very extensive migration during exsolution.

These mantles contain a

second generation of exsolved Hem ss> and for all sets which are in optical continuity, the plane of exsolution is {0001}. (e-f)

0.15 mm.

Low and high magnifications, respectively, of H e m ^ cores with exsolved lenses o

Ilm

and with mantles of Ilm . In addition to these lenses are coarse lamellae ss ss of pleonaste which share the same plane of exsolution as those of the ilmenite. The outer margins result in a symplectic intergrowth with Mt gg .

At high magnifi-

cation it is evident that each of the pleonaste lamella is surrounded by Ilm^^, suggesting that the spinel predated Ilm^ exsolution.

Whether these spinels are

the result of exsolution or an exsolution-like process is not clearly understood, e = 0.15 mm; f = 400 ym. (g-h) The reverse relationships are illustrated here for ilmenite hosts and H e m ^ exsolution.

The plane of exsolution is still {0001} but the notable difference here

is that the lenses are extremely fine grained and that depletion zones, or zones o non-exsolution, are concentrated along the grain boundaries.

Hg-114

0.15 mm.

Figure Hg-28

Hg-115

Figure Hg-29 "Exsolution" in Ilmenite

ss

, Hematite , and Rutile ss

(a-b) Discontinuous oriented rods of titanian chromites in kimberlitic picroilmenites Both grains show a varied distribution of lamellae with respect to Ilm^ grain boundaries but the rods are uniform in size. to be parallel to {0001}.

The plane of exsolution is assumed

0.15 mm.

(c-d) Although the oriented lamellae in these picroilmenites have distinctly differen optical contrasts with respect to each other and with respect to their hosts, the lamellae in both grains are Mg-Al-titanomagnetites; the lighter lamellae are high in Fe

, and the darker lamellae are enriched in Mg and Al.

The bleached zones i

(c) are the result of partial oxidation and the surrounding groundmass crystals a perovskite.

Peripheral rutile is present in grain (d) and a small amount is also

present in grain (c).

0.15 mm.

(e-h) The host in these grains is titanohematite; the darker gray lenses are Ilmss; t z-shaped lighter gray lamellae are rutile in the typical blitz

texture.

One ruti

lamella is present in (e), along which abundant Ilm have exsolved; no Ilm is ss present in (g). The ilmenite exsolved plane is {0001} and the rutile exsolved planes are assumed to be {0111} and {0112}; the former results by exsolution sens stvicto,

whereas the latter is more likely the result of an "exsolution"-like

process related to oxidation.

0.08 mm.

(i) A rutile phenocryst containing sigmoidal lenses of Ilm^ and mantled by Ilmsg> Mt and perovskite.

0.15 mm.

(j) A picroilmenite crystal with a core of rutile; the rutile contains irregular incl sions of Ilmss rather than the oriented arrays typical of the association. ginal and groundmass grains are perovskite.

Hg-116

0.15 mm.

The m

Figure Hg-29

Hg-117

A review of the textural assemblages associated with members of the Ilm-Hemss series are illustrated in Figures Hg-28-31.

Exsolution characteristics are illustrated in Figure

Hg-28,and the experimental phase relationships are discussed in Chapter 2 for the series Fe„0o-FeTi0-,. 2 3 3

Exsolution of Hem

ss

from Ilm

ss

, and of Ilm

ss

from Hem

ss

is restricted to

deep-seated intrusions and are particularly characteristic of anorthosite associations and other basic suites, but are also present in granitic suites.

The plane of exsolution

is parallel to the {0001} rhombohedral direction, and exsolution takes place in accord with decomposition as a consequence of slow cooling and solvi-intersection.

The growth and

distribution of large and finer lamellae results in a synneusis texture, i.e., with finer lenses of the solute (exsolving phase) distributed between thicker lenses in the solvent (host-dissolving medium).

Under conditions of extreme and prolonged cooling, diffusion

of the solute migrates towards crystal boundaries and atoll textures result.

Equally

common, although not within the resolution of these photomicrographs, are second and in some cases tertiary exsolution bodies within the primary lamellae; these relationships suggest a process of exsolution which yields successive sets of compositional pairs of Ilm ss and of H e m ^ which conform respectively to conjugate sets of lamellae whose compositions are controlled by the slopes and the limiting boundaries defined by the immiscible solvus region.

Textural interpretations of the series are considered by Edwards (1965)

and by Ramdohr (1969), and the magnetic self-reversal properties of the series are discuss' by Carmichael (1961,1962).

Solvus relationships were established by Carmichael (1961) on

natural material, and the revised experimental solvus by Lindsley (1973) is discussed in Chapter 2.

A detailed treatment of exsolution mechanisms in the series has been pub-

lished by Kretchsmar and McNutt (1971) in which they suggest that exsolution of compositioi with >Ilm Hem

ss

OJ

is continuous, whereas with initial compositions which are more enriched in

, exsolution proceeds discontinuously.

This interpretation is consistent with the

textures illustrated in Figure Hg-28 inasmuch as there is a greater preponderance of discontinuous or second generation lenses in grains where hematite solid solution members fori the host and Ilm members the exsolved solute: for the reverse relationship of Ilm hosti r ss ss the exsolved Hem g s solute is a relatively evenly-distributed single generation constituent (compare for example, Figs. Hg-28g and Hg-28h with Figs. Hg-28b and Hg-28f). Other examples of previously assumed exsolution are shown in Figure Hg-29, and in these crystals the exsolving solutes are spinels from picroilmenite and of rutile from titanohematite. acid suites.

These examples are typical of

i n kimberlites, and of Hem sg in more

The exsolved phases have an exsolution appearance with respect to crystal-

lographic control, and with respect to the morphology of crystal lamellae. both the H m s s

an

However, for

d the Hem g s the solvent and the solute have differing crystal symmetries

and neither the host nor the exsolved phase form members of either a continuous or a discontinuous solid solution series.

The exsolved constituents of the kimberlitic

picroilmenites are titanian-chromites (Fig. Hg-29a-b), and Mg-Al-titanomagnetites (Fig. Hg-29b-d); these phases are oriented along {0001} rhombohedral planes which is consistent with the parallel planes of exsolution exhibited in Ilm-Hem ss (i.e., exsolution sensu striato). In the case of Hem (Fig. Hg-29e-h), two of the crystals shown contain Ilm ss ss lenses in addition to rutile. The grain in Figure Hg-29e has a single thick lamella and many finer lamellae which are clearly earlier than the exsolution of Ilm

. In Figure ss Hg-29f the paragenetic relationships are not definitive but here rutile is in far Hg-118

in Hem

with well-defined crystallographic control along {0111} rhombohedral planes ress suiting in Ramdohr's (1969) "blitz" texture. The abundance of rutile in these two latter instances is far in excess of the limits for the solid solubility of TiC^ in members along the join Fe,0,-FeTi0_, in the absence of Pb members. Therefore, these assemblages are Z 3 3 ss most likely the result of an exsolution-like process which is related to oxidation in much the same sense that Ilm

"exsolve" by oxidation exsolution from Dsp-Mt members. ss ss The reverse relationship of H m s s lamellae in rutile hosts is illustrated in Figures

Hg-29i and j and in two groundmass grains from kimberlites.

For both grains the perplexin

problem of whether true exsolution or an exsolution-like process is responsible for this fabric is compounded by the fact that rutile and H m s s

are

at

times present in approximate

equal concentrations; this bears a much closer relationship to the expected modal concentrations which are likely to develop if decomposition of a Pt>sg member, in the system Fe2TiOj-FeTi20^-MgTi20g, is invoked as the precursor to such bimodal assemblages, and sine such compositions have been recognized {i.e., compositions approaching those of lunar armalcolites (FeMg)Ti2C>5 (Haggerty, 1975), the likelihood of decomposition rather than exsolution is favored.

For the cases of minor Il» ss in rutile the probability of unmixing

from a system with limited solid-solubility may still be possible from an initially high temperature and homogeneous one-phase mineral.

The mechanisms remain unresolved.

Reactions involvinq ilmenite-hematite "

ss

In contrast to the uncertainties aired above, the effects of magmatic and metasomatic reactions and the resulting assemblages which form from early and late stage deuteric oxidation and reduction are reasonably well defined.

Examples of the assemblages and the

textures derived from each of these modifying processes are illustrated in Figures Hg-30 and Hg-31. The first process to be considered is magmatic metasomatism which falls into at least the following three distinct categories characterized by the major newly developing Tienriched constituent:

(1) the formation of sphene (CaTiSiO,.) ; (2) the formation of perov-

skite (CaTiOj); and (3) the formation of aenigmatite (Na2FejTiSigC>2Q).

Each of these

reactions

Sphenitization

is illustrated in sets of photomicrographs in Figure Hg-30.

is relatively common in regional low-grade metaraorphic terrains, but in many underformed or tectonically undisturbed deep-seated intrusives and in recent hypabyssal suites, the development of sphene as a reaction product of H m s g °r of titanomagnetite (discussed later) is pervasive, suggesting that it constitutes a product of autometasomatism by reaction of residual liquids with early formed Fe-Ti oxides.

The formation of primary

and secondary perovskite is restricted to undersaturated suites, and the most spectacular examples are in melilite basalts, kimberlites, and in carbonati^tes.

For the latter two

suites, Ilross are enriched in MgTiO^ and decomposition results also in the formation of Mg- and Ti-rich spinels some of which are magnesian-titanomagnetites in composition while others show variable enrichment in Cr^O^, Al^O^ and FeO.

In the cases of sphene and

perovskite reactions, the additive metasomatic ions are Ca + Si, and Ca, respectively. For perovskite reactions Fe from the ilmenite is accounted for in the formation of a spinel, and in low abundance levels of Fe in CaTiO^.

However, for sphene reactions the

excess Fe is restricted to the very limited formation of H e m ^ ; Fe is most commonly absent in analyses of CaTiSiO^, and the formation of associated late-stage amphiboles and phyllosilicates are usually regarded as the potential mineral sinks that develop Hg-119

Ilmenite

Reactions

Sphene (CaTiSiOj) replacement of H - m s s

b)

are

illustrated in both grains; rutile is

an associated phase in (a), and H e m ^ lamellae are exsolved from ilmenite in (b). The variation

in color which is particularly evident in (a) results from variation

in crystal orientation and reflection pleochroism. f)

0.15 mm.

The progressive decomposition of picroilmenite illustrated in this series results

initially in the marginal formation of Mg-rich M t ^ + perovskite; the spinels are dark gray, and perovskite (CaTiO^) is white. Hem h)

ss

+ R are associated phases.

With more intense decomposition (f)

(c, d & f) = 0.15 mm; (e) = 750 ym.

The grains illustrate the replacement of Ilm

which appears as the dark gray constituent. crystal, whereas grain (h) has {111} oriented ternal composite ilmenite.

by aenigmatite (Na 2 Fe 5 TiSi 6 0 2Q )

Grain (g) is a discrete ilmenite lamellae in M t ^ and an ex-

Note that it is only the H m s s which is selectively

replaced by aenigmatite (also known as cossyrite) and that the M t ^ remains unaffected.

0.15 mm.

Hg-120

Figure Hg-30

Hg-121

In many cases, however, the Fe-exchange process is not evident in the metasomatized halos adjacent to individual crystals, so that the Fe is either redistributed among other Febearing minerals, which may be possible at elevated temperatures, or alternatively removed from the system in aqueous solutions if the reactions take place at correspondingly lower temperatures. Aenigmatite reactions differ from those of sphene and perovskite insofar as the problem of Fe reconstitution is accounted for, and in the sense that the metasomitizing ions are Na and Si rather than Ca + Si or Ca.

Aenigmatite replacement of ilmenite and

of the ilmenite constituents of oxidized titanomagnetites are observed in pegmatoid zones in thick differentiated basalts, and these zones in common with the primary crystallizatio of aenigmatite are typical of peralkaline host rock compositions.

Examples of aenigmatite

replacement are illustrated in Figure Hg-30g-h and in the case of the titanomagnetite grain (Fig. Hg-30h), it is clear that aenigmatite replacement post-dates the oxidation exsolution of H m s s from the titanomagnetite.

This is consistent with experimental data

for aenigmatite which show that synthetic compositions have a maximum stability below 900°C (Lindsley, 1971), and with the textural evidence that aenigmatite is a late-stage interstitial liquid reaction product. Reduction "exsolution" and the oxidation of ll®

are

illustrated in Figure Hg-31.

Buddington st at. (1963) and Buddington and Lindsley (1964) have discussed the formation of Mt

from Ilm-Hem by subsolidus reduction. These authors have noted the occurrences ss ss of subsolidus reduction of Ilmgs in hornblende granite gneisses from the Adirondacks and in paragneisses of the Franklin, New Jersey area.

Magnetite^ develop as lenses or

lamellae along planes parallel to the basal planes of the ilmenite; the texture has been simulated experimentally (Buddington and Lindsley, 1964), and the only known widespread occurrences in igneous rocks are in trachybasalts from Teneriffe (Haggerty et at. , 1966) and in andesites from Peru (Haggerty and Wilson, unpublished).

Examples of the

latter are illustrated in Figure 31a-b, and in the three ilmenite grains shown, the reduction exsolution lamellae of Mt gg are partially oxidized to titanomaghemite.

An explanatioi

for the trachybasalts cannot be given, but the widespread occurrence of these intergrowths in the Peru andesites suggests that this assemblage may have resulted during deuteric cooling; an alternative mechanism could also be invoked which would relate reduction to sulfide hydrothermal activity. Apart from these examples of subsolidus reduction, spinel^ are also present in kimberlitic picroilmenites as discussed above with reference to Figure Hg-29a-d, and a second example from the deeply weathered pipes of west Africa are shown in Figure Hg-31c-d.

These

examples are for ilmenite xenocrysts and contrast with the groundmass ilmenites discussed previously.

Surface weathering is highly selective in replacing the M t ^ , with a tendency

to leave the more resistant Mg-rich ilmenites virtually untouched.

The overall similarity

between these kimberlitic ilmenites and those in andesites would tend to favor a subsolidu; reduction mechanism, and the reducing agent is perhaps more clearly defined in the former and is related to COiCO^ equilibria.

However, the possibility still exists that the

formation of Mt g s is not related to subsolidus reduction but to a process of low-temperatui chemical weathering.

It is relevant to note:

(1) that oxidation exsolution lamellae of

Ilm in Mt , that coexist with Ilm having reduction lamellae of Mt , are moderately ss ss ss ss ' common in the Peru andesite suite but extremely rare in kimberlites; and (2) that there are no visibly apparent reduction modifications to coexisting spinel^. Hg-122

For Ilm

ss

in other igneous rocks the progressive sequences of oxidation are less well es-

tablished, and in rock types other than those of the extrusive basic suite, the prevalent decomposition assemblage is Hem s g + TiO^.

Examples in which the TiO^ polymorph is rutile

(based on x-ray data) are illustrated in Figure Hg-31e-g for three grains from a gabbroic anorthosite; additional examples in which the TiC^ polymorph is anatase (based on x-ray data) are illustrated in Figure Hg-31h-j for a progressive sequence of oxidation in a kimberlitic ilmenite megacryst.

Adjacent to the ilmenite the assemblage is finely tex-

tured, and is anatase + Hem g g (h); in an intermediate zone the H e m ^ is progressively removed and coarse crystals of anatase result (i); at the outermost grain boundary the Hem is largely removed and a coarse cavernous network of anatase is the end product, ss This example is most typical of kimberlitic I l m ^ and is most likely the result of lowtemperature magmatic fluidal interaction and later supergene dissolution. Ulvdspinel-magnetite The consolute point for the solvus of the series Fe^TiO^-FejO^ is at approximately 600°C and compositions along this join are sensitive to T°C and fj^ conditions in magmas which co-crystallize members of the Fe202_FeTi0^ solid solution series as discussed in Chapter 2.

The relatively low temperature of the critical point yields compositions which

may be quenched from above 600°C in most extrusives and in some high-level hypabyssal suites; hence, compositions which span the entire series are theoretically possible although the extremely low f()2 values required for the formation of stoichiometric Fe^TiO^ dictates that it is relatively rare as a discrete mineral.

Under conditions of slow

cooling, however, for members within the low temperature immiscible region, unmixing or phase exsolution into Usp gg -rich and Mt gg -rich solid solution members form when the solvus is intersected.

This results in exsolution and the plane of exsolution nucleation

is parallel to {100} spinel planes. lit-par-lit.

Textures are described as woven, cloth, parquet or

Exan^iles of Mt g g exsolving U s p ^ and of the reverse relationship are illus-

trated in Figure Hg-32.

These photomicrographs are most typical of those examples de-

scribed in the literature by Ramdohr (1953,1969), Vincent and Phillips (1954), Nickel (1958), Vincent (1960), Vaasjoki and Heikkinen (1962), and Tsvetkov (1965).

These

examples, and from the reports by many other investigators, e.g. , Anderson (1966), Morse and Stoiber (preprint), there is a clearly defined relationship between exsolution of members along the join and deep-seated basic intrusives of the gabbroic and anorthositic suites.

Hypabyssal occurrences have been described by Vaasjoki and Heikkinen (1962) and

by Jensen (1966) but in general even dykes appear to quench in members with intermediate compositions and exsolution is inhibited. The examples illustrated in Figure Hg-32 are from the Kiglapait intrusion, Labrador, and from the Cumberland stock, Rhode Island.

In samples from both localities H m s s

ar

e

an associated oxidation "exsolution" constituent and the relative parageneses between exsolution of Usp

ss

, or of Mt

ss

, and of Usp

ss

oxidation to Ilm

ss

are variable.

For

example, in some samples Hi" s s predate Usp gg exsolution, and oxidation must therefore have taken place above 600°C; in other instances the formation of Usp g g predates that of the Ilm g g , and oxidation was hence below 600°C. Exsolution produces two constituents which are both more highly enriched in Usp and therefore FeO, and a second phase which ss 3+ is more highly enriched in Fe^O^ and therefore in Fe . The former is more susceptible to oxidation and the latter consequently less susceptible to oxidation than the original Hg-123

Figure Hg-31 Ilmenite

(a-b)

Reactions

ss

The hosts in these grains are I l m ^ , and the oriented lamellae along {0001} plane

are Mt

ss

which show partial to complete decomposition to titanomaghemite.

the cores abound with lamellae whereas the margins are lamellar-free. result by subsolidus reduction and subsequent low T°C oxidation. (c-d)

The lamellae

0.15 mm.

lamellae in picroilmenite which show partial oxidation to H e m ^ ( c )

Deformed Mt

and dissolution of these lamellae as shown in (d). free of lamellae are Mg-titanomagnetites. (e)

Note the

The darker mantles which are

0.15 mm.

The dark gray host is ilmenite, the white ellipsoidal bodies are Hem s g , and the lighter gray sigmoidal phase is rutile.

Note that R alone is absent in the ilmenit

so that although the texture is suggestive of exsolution, the reaction 2FeTi0^ + l/202 = Fe 2 0 3 + 2Ti02 is more likely. (f)

A core of Hem bodies of Hem

ss ss

0.09 mm.

with oriented Ilm ; the mantle is ilmenite with fine exsolution ss Ilmenite within the core as well as ilmenite in the same outer

.

rim, show decomposition to an intimate intergrowth of H e m ^ + R.

The resulting

assemblage is recognized optically by yellow to red internal reflections which are characteristic of rutile, and by the distinction that the Hem ss are whiter than Because the decomposed hematite

either the exsolving Hem g s or the exsolved H e m ^ .

is associated with rutile, it is Fe^+-rich and Ti-poor, in contrast to the exsolution-associated Hem s g which are titanohematites. the former also have a higher reflectivity. (g)

A more advanced stage of H m s s

(h-j)

R +

Hem

ss

Apart from color differences

0.13 mm.

than that illustrated in (h) .

750 ym.

This series illustrates the decomposition of Hn> s s to anatase (Ti02) + H e m s s

with the preferred dissolution of F e 2 0 j and the development of a porous network of euhedral Ti0 2 crystals.

A portion of the unaltered ilmenite is shown in the

upper left hand corner of (h).

0.1 mm.

Hg-124

Figure Hg-31

Hg-125

Ulvo spinel-Magnetite^ (a) A parquet-textured intergrowth of exsolved Mt gg (white) and Usp sg (gray in the photc micrograph but brown in reflected light oil immersion) from an initial composition which must have been close to Usp^gMt^Q.

0.15 mm.

(b) Blitz-textured Usp gg in Mt gg with an attached grain of pyrrhotite. (c) A transitional series from lit-par-lit Usp ss to lamellar Uspsg. solution are parallel to {100} spinel planes.

0.15 mm.

The planes of ex-

The assemblage is weakly anisotropic

and Ilm

are inferred, although these cannot be identified as discrete oxidation ss constituents. 0.15 mm. (d) The black, rods are pleonaste and the host is exsolved Usp gs from M t ^ ; the pleonasl results from a higher temperature solvi intersection than the Usp-Mt^ assemblage. Note that Dsp gs are absent in zones adjacent to pleonaste.

0.15 mm.

(e) The area to the right of this grain contains abundant lamellae of Ilm gs whereas the area to the left contains almost equal proportions of Usp gg + texture.

in a cloth-like

This is indicative of extremely steep oxidation gradients and although

ilmenite cannot be observed within the Uspgg-rich portion, weak optical anisotropy is apparent and incipient H m g g are hence inferred.

0.15 mm.

(f-g) Blocky Mt gg (white) with finely dispersed original Uspgg. gray and display very distinct optical anisotropy; H ® s s

These areas are dark once again inferred.

The term pvotoilmenite (Ramdohr, 1963) has been employed to describe these optically unresolvable areas.

These grains show a distinct marginal texture which is

apparent also in (a) and in (e).

In addition to protoilmenite, grain (g) also has

a sandwich lath of Ilm . 0.15 mm. ss (h) A thick sandwich lath of Ilm^ divides this grain into:

a relatively unoxidized

and protoilmenite-rich zone (left); and a highly oxidized zone with Ilm lamellae and minor Usp

ss

exsolution.

0.15 mm.

Hg-126

trellis

Figure Hg-32

Hg-127

oxidation of the Usp

ss

takes place and that the Mt

ss

remains unaffected.

Each of the

>600° and pleonaste

exsolution is either earlier than the oxidation of titanomagnetite (c), or contemporaneous with oxidation (d) .

H m s s lamellae are either lighter or darker

gray in the photomicrographs depending on the orientation. are offset by H m s s

Some pleonaste rods

(d)> whereas in other regions (d also) it appears that oxida-

tion and nucleation of I l m ^ has occurred along the discontinuities created by high-temperature exsolution.

In these cases the exsolved spinels are fragmented

and are totally enclosed in ilmenite (thick sandwich lath in grain d). (e-f)

0.15 mm.

These are two examples of coexisting Cr-Mg-hercynite (FeA^O^, black) and Al-Cr-

magnetite.

The cloth-like fabric in the lighter

Mt

ss

areas, although similar in

appearance to Usp-Mt gs exsolution, has not been identified and is not Usp because of analytically determined low Ti contents.

An Ilni^^ lath is present in (e) and

the bright fleck in the center of (f) is pyrrhotite. (g-h)

0.15 mm.

The relationships of chromite and of magnetite differ only in the degree of phase

separation.

Although the series is known to be complete at high temperatures and

that coexistence is dependent on ff^ (see Fig. Hg-39), these assemblages are present in highly reduced rocks (serpentinites) which probably formed at T < 500°C. Exsolution is inferred but the assemblage may have originated by an entirely different process, e.g., recrystallization at high pressure or by oxidation.

Mt

appear white in (g) and light gray in (h); the white areas in (h) are Ni^Fe. 0.15 mm.

Hg-130

ss


600°C) and the constituent phases are gray Pb

lamellae along relic {111} planes, which

were originally Ilm^^, with the associated white phase which is Hem g .

Grain (c)

contains some unreacted Mt gg with abundant rods of pleonaste. Note that where the magnetite is oxidized (SSW in c) that the resulting assemblage is hematite and that beads of undigested

spinel persist because of the limited solid solubility of Mg

(but also Al) into the Ti-poor Fe2°3 structure.

0.15 mm.

(See Chapter 4 for ad-

ditional and more complete coverage.) h)

Sphene (either light gray or dark gray) replacement of titanomagnetite-ilmenite

assemblages which have resulted by oxidation-"exsolution."

Grain (e) is an unusual

example because of the preferred replacement of the M t ^ constituent.

Grain (f)

is most typical because replacement by CaTiSiO^ is restricted to the Ti-rich lamellae which were originally Ilm^.

Grains (g) and (h) show the partial to com-

plete replacement of Mt-Ilm; in (g) the Mt is totally replaced and the thick Ilm lamellae are partially replaced; in (f) a sphene pseudomorph has resulted with ghost-like lamellae of original H m s s still apparent along {111} planes of chemical destruction but maintained structural integrity).

(a result

The contrast in

optical properties of these sphenes is related to the degree of crystallinity; sphene after Ilm is coarsely crystalline (c/. Fig. Hg-30a-b), whereas sphene after Mt is extremely fine grained.

0.15 mm.

Hg-132

Figure Hg-34

Hg-133

illustrated in Figures Hg-34g-h arid in these there are distinctive optical properties for each of sphene replacing Mt

and of sphene replacing Ilm . For Ilm replacement the ss ss ss sphene is coarsely polycrystalline and compact, in contrast to sphene replacing host titanomagnetite which is microcrystalline and porous; the former yields a well-defined

x-ray powder pattern whereas the latter yields a diffuse pattern.

The compositions, as

determined by electron microprobe for both types, show that FeO is rarely present in concentrations which exceed 0.5 wt. % and that the two types are similar in every other respect.

The example shown in Figure Hg-34e is from an altered basalt from Jamaica, and

the remaining examples are from the mafic body of the Kalgoolie Gold Mine, Australia. Similar examples are typically present in both deformed and undeformed metasomatized mafic complexes, and it is considered by some authors {e.g. , Desborough, 1963; Davidson 2+ 3+ and Wyllie, 1968) that the extensive removal and remobilization of Fe + Fe may result in the formation of large iron-ore deposits of the Cornwall type.

The temporal relation-

ships of sphenitization are difficult to determine but in some cases the process may result from autometasomatism with late stage fluids, or from autometasomatism associated with synkinematic intrusions.

Other examples of sphenitization can be more clearly re-

lated to post-supergene oxidation or to post low-temperature deuteric oxidation.

For

these associations, sphene paramorphs the curvilinear planes of pre-existing maghemite as shown in Figure Hg-35a-b.

This form of replacement, which appears as the murky gray

to dark gray phase in Figure Hg-35b has not been fully characterized, and diverse opinions range from R + H e m ^ , to sphene, and to amorphous Fe-Ti oxide.

More recent studies by

electron microprobe and coupled x-ray analyses show that sphene is present in many instances, but in an equal number of cases neither the detection of Ca nor the characteristic peaks for rutile or for sphene have been found in Debye-Scherrer powder patterns. In all cases, however, both Fe and Ti are typically present, but data yield low analytical totals and a complex submicroscopic assemblage is hence inferred. Maghemitization (^620^) and martitization

(aFe^O^) have been discussed rather

extensively in the literature (e.g., Basta, 1959; Akimoto and Kushiro, 1960; Katsura and Kushiro, 1961; Ramdohr, 1969).

Examples of each of these oxidation products of Mt ss Maghemite is a cationic

are illustrated respectively in Figures Hg-35c-d and Hg-35e-f.

deficient spinel, is hence cubic and optically isotropic, and is commonly either white or bluish-white in reflected light oil immersion.

Oxidation planes are typically ill

defined,and the characteristic fabric as illustrated in Figure Hg-35c is along curved, conchoidal directions which are usually accompanied by fine hairline cracks; these cracks are considered to develop as a result of the reduction in volume from M t ^ to maghemite^ and are perhaps also the consequence of octahedral cationic deficiencies in yFe20j.

With

increasing distance from these cracks there are gradational color changes from white immediately adjacent to the cracks, to tan at contacts with unoxidized titanomagnetite. These variations represent a continuous sequence of oxidation solid solution from titanomaghemite to titanomagnetite, and members of the series are therefore somewhat akin to the omission solid solution series of kenotetrahedral and keno-octahedral magnetites described by Kullerud et at. (1969).

This topic is treated by Lindsley in Chapter 1.

A second

example of yFe20j is shown in Figure Hg-35d in association with the high-temperature oxidation "exsolution" of I l m ^ lamellae along {111} spinel planes.

In this example

wide variations in the intensity of maghemitization, which are similar to those shown

Hg-134

magnetite crystal into subunits of smaller areas results in a patch-quilt appearance with each subarea showing a higher or lower degree of oxidation intensity; the white areas are at the most advanced stage, and the tan areas (gray in the photomicrograph), or areas of intermediate color, are at low intensity levels, or are at intermediate levels in the transition from titanomagnetite to titanomaghemite.

Oxidation to maghemite is typically

associated with more titaniferous Mt g g components, whereas aFe20j (hematite) is more typical of magnetitegs (i.e., Ti-deficient) compositions.

The inversion of yFe^Oj to

aFe^Oj takes place between 250° and 550°C and is dependent on grain size and a number of other factors as discussed in Chapter 1. An additional distinction, apart from the characteristic association of -yFe^O^ and titanomagnetite, is that hematite oxidation of Mt g g takes place preferentially along {111} spinel planes (Fig. Hg-35e), and this fabric bears a close resemblance to the orientation and form of H m s s which develop by oxidation exsolution (see, e.g., Fig. Hg-34a) at elevated temperatures from Uspgg-rich starting compositions.

As oxidation becomes more

intense, the {111} primary fabric is disrupted (Fig. Hg-35f) and pseudomorphs of hematite after magnetite result.

The end product is colloquially referred to as martite and the

oxidation process as martitization.

Note that in contradistinction to the high-temperatur

oxidation of titanomagnetite, no additional phases such as Pb gg or R form (see Chapter 4). The optical distinctions between hematite and maghemite are a higher reflectivity constant for hematite, marked pleochroism and optical anisotropy; maghemite is optically isotropic or very, very weakly anisotropic and has a reflectivity value which is close to that of titanomagnetite but substantially less than that of hematite. Oxyhydration is the final process to be considered, and in ideal cases the hydrated products of magnetite or of hematite is goethite, aFeO.OH (Fig. Hg-35g); oxyhydration of maghemite (yFe^Oj) results in lepidocrocite (yFeO.OH); and akaganeite (BFeO.OH) which is relatively rare appears to be more closely associated with FeCl decomposition in meteorite rather than with oxide hydration.

Hydration oxidation of titanomagnetite or of titano-

hematite is uncommon in basic igneous rocks and appears to be most commonly associated with the less titaniferous varieties in more silica-rich suites, and as a byproduct of sulfide decomposition. The processes of maghemitization and oxyhydration are low-temperature events.

These

may result from the final stages of alteration associated with deuteric cooling, but the more likely associations are in supergene weathering and in burial metamorphism.

Hematite

derived from low titanium magnetites, however, can be either a high-temperature product or alternatively it may form under conditions which are similar to those of maghemite formation. Oxides derived from silicates The replacement of ilmenite by aegnigmatite and by sphene is reviewed in Figure Hg-30 and in Figures Hg-30 and Hg-34, respectively.

The reverse relationships of silicates

yielding oxides by exsolution, partial decomposition through processes of magmatic interaction, or of oxidation are illustrated in Figure Hg-36. Exsolution of Mt g g and in some instances H m s s Is relatively common in plagioclase and in pyroxene from plutonic suites.

Particularly striking examples are in larvikites

and in labradorite-bearing anorthosites where oxide exsolution is in part responsible for the blue irridescent Schiller effect.

The planes of exsolution are typically parallel to Hg-135

Magnetite and Titanomagnetite Reactions

(a-b)

The gray and white areas are sphene but have also been described as amorphous

Fe-Ti oxide, and this is so particularly for grain (b).

A comparison of the

curvilinear control on the distribution of these secondary products with textures typical of maghemite replacement (c) suggests that sphene postdates the formation of yFe202- Ilm is absent in grain (a) but is present in (b). of Mt

ss hemite. (c)

Small residual area;

are still present in (b) but the grain is largely oxidized to titanomag0.15 mm.

This is the most typical development of titanomaghemite (white) after titanomagnetite, i.e., along curved conchoidal fractures which are inferred to result from changes in structural volume induced by the inversion of an inverse spinel to a cationic deficient spinel.

(d)

0.15 mm.

Variable compositions and a spectra of intermediate phase inversions are apparent in this partially oxidized titanomagnetite grain.

The lamellae are I l m ^ and the

variations in color among segments which are bounded by these lamellae are representative of stages in the transition from titanomagnetite to titanomaghemite. Note that the volume-change cracks are not necessarily associated with the most intensely maghemitized areas, and that the crystal interior is more highly oxidize than the grain boundaries. (e-f)

0.15 mm.

Hematite is the white phase in both grains, and oxidation of these Ti-free crys-

tals along {111} spinel planes is in stark contrast to the lack of crystallographi control so typical of maghemitization (b & c).

These distinctive textural habits

and the anisotropy of Hem (maghemite is cubic and hence isotropic) are the optical distinguishing properties for identification. (g-h)

0.15 mm.

The dark gray regions in these photomicrographs are goethite; in (g) it develops

as a botryoidal (or reniform = kidney shaped) overgrowth on Mt but some areas of the Mt also exhibit partial oxyhydration; in (h) hematite is an associated phase, and this assemblage is typical of vesicle infillings in volcanic suites.

Hg-136

0.15 mm.

Figure Hg-35

Hg-137

Figure Hg-36 Oxides Derived from Silicates

(a)

Usp-Mt

exsolved as rods and discontinuous blebs in plagioclase.

cent oxides are Ilm

ss

.

The large adja

0.15 mm.

(b)

Symplectic arrays of Al-rich spinels in clinopyroxene.

(c)

Magnetite derived from biotite and concentrated along cleavage planes.

(d)

Magnetite associated with the peripheral decomposition of amphibole.

0.15 mm.

Coarse symplectic magnetite in a highly oxidized crystal of olivine.

Oxidation

(e)

0.15 mm.

occurred at T > 600°C based on the presence of associated (f)

0.15 mm.

^50 pm.

The outer margin of this highly oxidized olivine crystal (T > 600°C) has a diffusJ mantle of Hem (white) + magnesioferrite (gray).

The highly reflective zone adjace

to this mantle is predominantly Hem; the core of the olivine is dark and diffuse £ at high magnifications, symplectic magnetite can be resolved which is similar to that shown in grain (e). (g-h)

0.15 mm.

These olivine crystals are typical of those which are generally described as

having undergone iddingsitizatvon.

Because iddingsite is not mono-mineralic, as

is clearly apparent from the photomicrographs, the name has only descriptive connotations.

The identifiable white phase is goethite and the associated darker

phases (note cleavages in grain g) are smectite-related constituents.

This assem-

blage parallels maghemitization of titanomagnetite; ilmenite remains unaffected as shown in Figure Hg-35c.

0.15 mm.

Hg-138

Figure Hg-36

Hg-139

plate-like in form as illustrated in Figure Hg-36a.

These finely textured oxides are an

important constituent in magnetic studies, and several investigators (e.g., Hargraves and Young, 1969; Evans and McElhinny, 1969) have demonstrated that the major stable component of remanence resides in these microscopic arrays rather than in the larger discrete crystals.

More aluminous oxides in the form of green or brown spinels are also relatively

common in pyroxenes (Fig. Hg-36b) but appear to be rare in plagioclase (Whitney and McLelland, 1973) in cases other than those which involve the reaction of olivine + plagioclase = aluminous pyroxene + spinel.

These latter reactions are typically the result of

metamorphism although convincing arguments have been made, for example by Griffin (1971) and by Griffin and Heier (1973), which suggest that a pressure increase during cooling is a probable mechanism to trigger the reaction at subsolidus temperatures.

The products of

the reaction result in a symplectic intergrowth and the assemblage, which may also be accompanied by garnet, develops as a corona around olivine or at the grain junctions between plagioclase and olivine. Figures Hg-36c and Hg-36d illustrate the partial decomposition of biotite and of amphibole to magnetite; these magnetites are typically very close to stoichiometric Fe^O^, but minor element concentrations of 1-2 wt. % (MgO + A ^ O ^ + TiO^) may also be present. Hydration of these oxides is also very common and this leads to the formation of goethite. The most widespread form of silicate decomposition is the oxidation of olivine, which at high temperatures results in magnetite + hematite or in hematite + magnesioferrite (Fig. Hg-36e-f).

At lower temperatures the typical breakdown product is iddingsite which

has a major constituent of goethite intimately associated with smectite (Fig. Hg-36g-h). Iddingsite formation may develop during the volatile saturated stages of deuteric cooling but is more clearly demonstrated to result as a consequence of supergene weathering processes.

The distinction between high and low temperatures for these olivine byproducts

is also reflected in the oxidation assemblages of the discrete oxides which show that the magnetite-hematite-magnesioferrite suite is accompanied by rutile + hematite, or by pseudobrookite + hematite, whereas iddingsite develops more typically in parallel with the oxidation inversion of titanomagnetite to titanomaghemite (Baker and Haggerty, 1967; Haggerty and Baker, 1967).

For all of the above cases, the earlier comments related to

magnetic stability and remanence apply here also. PRIMARY OXIDE DISTRIBUTIONS Introduction Paragenetic sequences, compositions and model abundances of primary opaque oxides in igneous rocks depend for the most part on the initial bulk chemistry of the host rock, on the depth of emplacement, and on the prevailing oxygen fugacity of the crystallizing magma. As a general principle because FeO, TiO^ and Cr^O^ contents increase with decreasing SiO^, basic rocks tend to contain larger concentrations of oxides than either intermediate suites or acid end members.

To a first approximation this distribution is controlled

largely by the increases in FeO contents which result with increasingly more mafic compositions.

Titanium variations, on the other hand, in the range from acid to basic suites,

determine both the distribution and composition of mineral solid solutions within the system Fe0-Fe90,-Ti0„.

Members of each of the solid solution series Fe0Ti0.-Fe,0. Hg-140

9+

in the ratio of FeiT:FeJ . Therefore, oxide distributions and compositions depend on initial titanium abundances and on ÍQ^• These coupled parameters lead to Mt-richss and hematite-rich Ilm

in acid and intermediate rocks, and to Usp-rich and Ilm-rich in ss ss ss basic and ultrabasic suites. Because both Fe :Fe and Fe:Ti ratios are dependent on temperature and f()2> f°r which coequilibrated oxide pairs yield unique solutions, the initial temperatures of crystallization and the cooling paths followed within the subsolidus are the limiting factors chiefly responsible for compositional variations between titanomagnetites and ilmenites.

This distribution implies, therefore, that suites of

initially lower temperature origins, e.g., acid extrusives, result in Mt-rich solid solutions + Ilm^, but by the same token suites which reequilibrate over long periods of time result in products of closely comparable compositions.

Members of the Ti-enriched FPb-Pb ss series are restricted to more basic rock types and are more abundant as secondary oxidation products, of Usp-Mtss and Ilm-Hem^, than as primary precipitates.

The TiC^ polymorphs

(rutile, anatase, brookite) are also commonly of secondary origin but minor primary concentrations are occasionally present in granites, syenites and diorites.

Chromium abun-

dances vary from low to very low in acid and intermediate rocks, increase with more mafic character, and peak in the ultrabasic.

Chrome-bearing members of Mt-Usp^ are restricted

almost exclusively to the basalt suite, whereas more complex solid solutions of A ^ O ^ and MgO in chromite (FeCr^O^) are common and abundant in picrites, peridotites, dunites and in kimberlites. The major controls on the distribution of oxides as a function of depth dependence are the effects of crystal settling processes and the likely but controversial differences that may exist in

levels among rocks of equivalent chemistry which have evolved as

batholiths, minor intrusions, extrusives, or explosive tephra.

Crystal fractionation

of oxides is widespread in basic intrusives, less common in rocks of intermediate composition, and rare in acid intrusives.

This generalized distribution also applies to suites

intruded at hypabyssal levels and for volcanic equivalents, with rare exceptions, is absent in all but exceptionally thick basic lava flows.

In the absence of crystal set-

tling, the oxides remain a minor component regardless of the mode of emplacement but assume radically different textural and paragenetic relationships with the silicates. These textures may vary from equigranular in intrusives, to inequigranular in dikes and sills, and to interstitial in extrusives.

With respect to the effects of oxidation

control on oxide distribution with depth, the problem is one of open and closed systems and on the retention capacities of the host rock to volatile loss during magmatic emplacement.

Volatile retention at depth in the closed system model will lead to higher 3+ oxidation and greater proportions of Fe , and with the higher probability of an approach to equilibrium for the partitioning of iron species and of titanium between oxides and silicates.

In a generalized sense the extrusive situation is one of rapid volatile loss

and of partial disequilibrium among oxides and silicates.

In hypabyssal regions, con-

ditions may vary between these two extremes although it is probably closer to that of the plutonic environment. The final and perhaps the most fundamental parameter in controlling the distribution, abundance and compositions of opaque oxides in igneous rocks is the level of initial oxygen fugacity and the fg^ path which is followed with cooling.

At magmatic temperatures

ÍQ2 is a major controlling factor in elemental partitioning among the oxides, and in elemental partitioning between the oxides and silicates. Hg-141

For high values of fo£> oxides

f()2 the available iron Is competitively partitioned between these crystallizing phases. The former results in the Bouen trend and the latter in the Fenner trend.

In the case

of titanium and for coexisting oxides and silicates, titanium will preferentially enter oxide solid solutions at low values of fj^, and hence the precipitation of Fe2TiO^-, FeTiO,-, and FeTi.O_-rich solid solutions are favored.

With progressively increasing 3+

f02> which is the generalized progression from basic to acid suites, the ratio of Fe

:Fe

also increases, and this results in the stabilization of Mt

The

and of Hem-rich Ilm

.

ss ss is also a strong interdependence between silica activity and fQ2 which dominates in highl; undersaturated rocks.

In these suites TiC>2 is depleted in Usp g s and in H m s g

extent that members of the Ilm ss series are commonly absent.

to the

Either sphene (CaTiSiO-) J

or perovskite (CaTiO^) may precipitate, with the former developing at relatively high silica activities and high oxygen fugacities, and the latter at correspondingly lower values of

an
500°C is considered necessary for crystallization, probably from a vapor phase.

(8)

T-f„ data for the rhyolites form part of a detailed study of °2

pumice (see appropriate Table), and Lipman's demonstration that T°C correlates with SiOj (higher silica with lower T°C) in the pumices is borne out also by two suites of rhyolites; in addition,these rhyolites have similarly high MnO contents in the ilmenites (3-4.5

wt%

MnO) to those in the pumices but

T = 700-780°C is relatively low; the correlation patterns are similar and H^O is not regarded as being a significant factor in the differing modes of eruption.

A third suite of rhyolites (810-860°C) are considered to have had

a low magmatic water content and these plot off the T-SÌO2 trend.

Post-

caldera lavas have Fe-Ti phenocrysts with very large compositional variations and these are considered to have formed by extensive magma mixing as a result of caldera collapse.

Hg-179

i

X X

X X X

X

X

X

X

X

Hg-IBO X X X X X X X X

X

X X X X

X

X

X

X

Nevada

Nevada

Rhyolites

0)

X

Norway

X

Rhombporphyry Rhyolites

|

u-|

o J m

e

X

CM

M 4) 01

o C (0s: FH

«

-14.5 -13.0 -12.0 -10.5

-

2. 00

X

ai

0>

X

J1

X

B 570670 700780 810860

X

-11.7

-tf

JRhyolite Rhyolite

1 1

j

1

S ON

X X X X X X X

CM CO

X X

to

vO

to

References

Carmichael (1967b)

Carmichael (1967b) Carmichael (L967b)

(1967b)

(1967b)

Carmichael

(1967b)

Carmichael (1967a)

Carmichael

Carmichael

Lipman (1971)

Storetvedt and Petersen (1970) Lipman (1971)

Bauer et al. (1973) 1 Nicholls (1971) 1 Carmichael (1967b) 1 Lowder (1970) Lufkin (1976) Klootwijk (1975)

t 3 w

Hawaii Ii Greece : Pantelleria New Britain New Mexico India

o ON X

fC

Ryodacltes [Rvodaclte

X

New Zealand

X

-12.6 -15.1 to -14.7 -15.2 to -13.8

-11.4

M f>

Rhyolites

o CO 00

H"

790810 735780 855865 . 890 1 1025 1

1 1

Notes

ÎN

California

X

Iceland

o «sì m S ON ON 00

X

815985 900925

f: -12.2 -13.9 to -10.6 -12.9 to -12.2

Emphasis 1

rH

Rhyolites

X

Scotland

X X

Arran

O. X X X

California

|

0) c 0) h -

O

6

M X

n a l=>

M X

X

X X X X X

S

«a e

4-1

e a CJ

0» Q. >•. H O

tu

T3 C ^ « a)

U W i-i « 0) c o i-i -C a E B e H Cu b cu

Pumice

O) o

d o 4J d ~ x: CO S T3ftC •d fl•a ° > a M > o > > Z H w Z î* Z H

cO » X

•d ai >

c

Z oS

m

0) 4-( > m e c M 3 M jî u a u ë à S 5

Hg-182

ON

d) 1 4J « 4J K-H ai -H (Xi s N M Ol S o*

Notes to Table Hg-7 (concluded)

730° (rhyolltes) and 130°C (qtz latites) correspond closely to those obtained for coequilibrated oxide pairs; a significant observation is that 1o Mt g s become consistently richer in 0 ° as the unit becomes more mafic, a trend which is not exhibited by coexisting feldspars; the latter is considered to have reequilibrated in the subsolidus as a result of the introduction of meteoric water to the crystallizing magma in the interval between rhyolite eruption and qtz latite eruption.

(10)

are

Values for

^rom

Table 5, p. 9 in Heming and Carmichael (1973) but the absolute range in T is 1035-835°C

= l1

O

OO

O

O

ON

«1

o m o oO OO O ON CT\ Os O ON £> © . - I CN O OO CO I N 1 1

O

-15. -19,

O

to

O CM

00 o

© oo o CM NO O 3 o Z

*

X

X

«J 3 " o O Q Q

w £ o O Q|

(0 4-" (0 w r-l "O M

to 3 c 4J T-t 0) 01 o as J2 •H co h M T3 co Z H
4J M S 0 and

^2^3

at

t le

'

expense of increasing FeO, Fe^O^, TiO^

a trend which is comparable to that shown for chromites in the ancient

Makaopuhi lava lake (see note 7).

These latter chromites are poorer in Cr^O^

(38 vs. 43 wt%) and this is reflected in bulk rock Cr^O^ contents which range from 0.14 - 0.20 for Kilauea Iki, and from .056 - .059 for Makaopuhi. ranges in Al^O^, MgO and TiO^ are approximately equal.

The

Eruption T = 1200-1225°C,

but attempts to utilize coexisting 01 + Chr yield 2000-2300°C.

(19)

These

papers consider the magnetic and oxide variations in drill cores from the Alai and Makaopuhi lava lakes, Hawaii. (20)

The oxide data are discussed in Chapter 5.

This, in common with Creer (1971)—see note 17, is an excellent review of

correlative parameters between magnetic property data and oxide mineralogy; cooling histories and (21)

their relationship to T and f„ are also considered. °2

Compositional data for Pb s g are reported from 29 localities; this early

report is still the most comprehensive study in the literature and was the first to draw attention to the high MgO contents typical of the series, and the fact that preferred distributions may be present in lavas exhibiting differential high temperature oxidation profiles.

See chapter 5.

(22)

For

three generations of primary titanomagnetite systematic variations are observed which show progressive increases in TiO^ and MnO contents, and progressive decreases in A l ^

= 9.5 wt%; MgO = 5.1 and C r ^

= 1.8 wt%.

The systematic

increase in TiO^ from U s p ^ ^ to Usp^g -j is considered to result from a decrease in f A with crystallization. (23) Columnar basalt study. (24) U2

On

the suitability of basalts of variable oxidation states for magnetic field intensity studies; these data show that only the most highly oxidized samples from a single lava study (lava HS—Chapter 5) obey the necessary criteria for reliable paleointensity determinations.

(25)

Although it had long heen recog-

nized that the particle sizes of M t g s are related to magnetic stability, this paper was the first to point out that particle size reduction (and hence increased stability) may be accomplished by oxidation exsolution, i.e. large crystals being reduced to a large number of finer sub-crystals separated by {111} I l m s s lamellae.

(26)

A detailed electron microprobe study of Ti-Al-

chromian-ulvospinels show that crystals are zoned from cores (high Cr) to

4J -1

m rJ r -

CM ai CM ai

r - r-»

C

a o

S

e

^ ON

u

(NI

«4-1

ON

«5

»

n

d m e in ON

> « IJ u h

en

•H >

'

0

B

tfl «3 c

J

rn r-

'm

? CJ

ON >

—

— _a

CO

O •u 3 M B Û a h i o m e (O CO e S

CM

CN»

g « •o •a ai XI o b) h h O o 00 ' ! N O CM

X

M h 0> ai o o

Hg-198

Hg-199

X X

X!

4J

Ul (0 x>

X

X X

U rt (0 nJ

X

Hg-200

Fe Pss 2°35 self-reversal processe? deep-sea and sub-aerial basalt comparisc mineralogical dependencies in the context of oceanic magnetic quiet zones.

e •H w 5 co e >> 0) MX (0 6 s •H M o< 1180-660 I

4> CM

OXIDATION MECHANISMS

X

X

basalt, 1 harzburgiti ; gabbro, \ troctolite1

co (0 o. (0 X

Macquarie ridge

00 o EE O

K X X

X X X X

rH o

«s

»H

d

c^

c

ë

Davidson and Wvllie (1968)

1 Ade-Hall et al. (1972)

1 Jensen (1966) 1 Unan (1970)

Vaasloki and Puustinen CL966)

Wilson and Haggertv (unpublished)

Beck (1966) Hargraves and Young (1969) Helsley and Spall (1972)

1' •

Puffer and Peters (1974) 1 Rice et al. (1971) I Watkins and Cambray (1971)

00 ON o»

X X

X

X X

X

Smith (1970)

c o 3 Wilkinson (1965)

\D r»

I

m X X X

t> o b t) c -

X X

«

1

o IO

en -a-

1

580-590

X X X

s>

s 0

I 1 I I

00 o O M-J

« a

1 1 1 1

o co

o>

Pennsylvania New Jersey Arizona S. Rhodesia Australia

CO

,

X X X

o u *) 01 c 01 a) X S "d c 3 (0 X» co G 0 c < Vi 0) I 0 4J 1 ce 01 3 s o> \D ON o o

X

o

H X

O X

B

««

X X X X X X X

X

X X X X

X X

|

1

Acid-Intermediate Intrusive Suites Notes:

(1) Wet chemical analyses; (2) fayalite - ferrohedenbergite - mesoperthite

granite; (3) hornblende (+ biotite) mesoperthite granites; (4) b i o t i t e m e s o p e r thite granites; (5) EMPA Manganoan-ilmenite p

in qtz-biotite-diorite; Mn-Ilm ( I l m 7 2 (6) M t g s

(Hem^

h

=

y 28^

12

'

^^g) 7

in plag, in hornblende and in biotite.

wtX

M n 0

in

=

12.7 wt% MnO

a d a m e l l i t e

;

(7) EMPA. Mn-Ilm: granite, MnO =

29.8-30.1 wt%; granodiorite, MnO = 10.8-19.5 wt%; m o n z o n i t e , MnO = 2.9-4.1%. V a l u e s for MgO are uniformly low, max. = 0.17 wt%.

Coexisting magnetites for

the three rock types contain max. MnO contents of 0.75, 0.54, and 0.18 respectis

ively; T i 0 2 + MgO + A 1 2 0 3 + C r 2 0 3 +

t

y P i c a l 1 y O i-I

o

CTN

CTS

OS

c

C

C

e

tí

a) (4-1

efl

•H

£

eu to ai

Z sí

M

e

P

i—(

c

S 3

§§

Z

z

z

í¡

-

vO

id

y

z

c

§

i

«

-

M

4J 4)

4-t

CO CO

CO

I

e

c •H

T3 •H

sO

sf sO

d

CO ON •O Cd

4)

e •H

- o • d c o T3 SO c «d «d cri

tí C

C 60

0

« 1

s

3 'cj

•Ih •H

3

H

3 •H

• o T3 • a T3

z

z

z

H

co

sO

£

CO iH

os o

Pu

•rt

«


CO CO O r s CO CO IO vO IO I O - Í i / ì

I

1—1

7 7

7

o o m o o CO 0 0 so C ) r-» s o r - . CO so

LO

m o i—i SO

o o •» 1—1

etí co

4) S

E D CO CO

o •H

> H S

(0

M

O.

n M

1 4J E

o

i ex

•H

cd

cd

(9 O O i-J

O O o

Z

»

co

o z

>>

3J i

3

3

O

O

z

o

z z z

O

3

o

z

•H

(il •H

à

4J

0)

4J O

H

M

•H 0)

cd Q

O

jj

Ö c cd cd o

-a

o

e e flj ^ o

N N 4J oíá

«

> !>>

o z

m

« o

Ü

o .o

O

ed

x> A (Il (Il O" o

Ö

00 Z

33

co

0) •H

O -H 0) tí (O CO S O a) 0)

OI O)

4-1

H -H

c

O cd 4) co O 4-) o> o 4) e o 4) d JJ 1—1 4-1 4) ¿J •2 , -H M •H > CO e s E c W 11 0) oc •H 4> ffl N «M > C/Î co eu O - H Z Cu o o - ä Q

4

tí

anorthosite complex in which he also reports T = 620 C and fu an oxide-rich gabbro.

2

-19 5

for

In the sequence from anorthosite (early), to gabbroic

anorthosite, to Fe-Ti oxide rich gabbro, to syenite (latest): poverished

= 10

M t g s become im-

in V^Oj/Fe^O^ by a factor of 0.5; and H m s s become enriched in

MnO/FeO by a factor of 4.3. These changes are associated with a progressive decrease in Hem contents which are: Hem., (anorthosite) to Hem, (syenites ss 34 6 and gabbros). f„ . at 880°C = 2000 bars for the gabbro. (23) EMPA. The 2 ranges in T and f are for initial and final conditions of crystallization. For the interval of T = 700-985°C the estimate of f„ = 310 bars. (24) The H2U value of T and f„ represents one of several values obtained from a study of °2 the association of Mt g g + apatite (see Tables on Magmatic Ore Deposits). (25)

Textural discussion of the distribution of exsolved Usp g s and of the

oxidation to Ilm exsolved Usp

. (26) Magnetite is present in two forms: (a) with ss + Ilm ; and (b) Ti-poor Mt free of exsolution. Sphene is ss ss ss

a selective replacement product and the authors consider that Fe-mobilization may lead to Cornwall type magnetite deposits.

(27)

Hornblende-biotite-qtz

diorite contains a reversely magnetized H e m ^ with exsolved I l m ^ ; and a normally magnetized Mt.

The Hem-Ilm ss association is in the range I l m ^ ^ ; self-

reversal although non-reproducible in the laboratory is considered to involve a negative charge interaction between a Ti-ordered constituent and a Ti-disordered constituent (i.e. within the Ilm-Hem exsolution array).

(28)

Syenites are

within an alkaline complex (2111 my) and because these rocks have high blocking temperatures (500-650°C) the directions of magnetization are those at the time of emplacement.

(29)

The oxide assemblage is similar to that reported by

Merrill and Gromme (1969) - note 27.

The two rhombohedral constituents are

Ilm^Q and I l m ^ respectively in exsolution association; intrusion T = 800°C.

Hg-207

rOCICICOCÌC-JsOv N vO (M in N VD n f H ri ri H H H t I I I I I I

m o io o m m o < 00 O CD N CO N CM V CTNCOCT\coo"xcocr>r

Hg-208

.. O "O r-N. f— OÍ AI i— > • aj E

co c ra •P - o raCT) a) o -c to o •p LO S- rato i— en o -p ra c s-C •Pa> CJ+ I ra en "O a>

en

E •*> O-—»

$- o a> r-E er»f— ra CJ —' •« E

---. o -—co i- en IX) ai LO Oí E en i—rai— >—C_J -—' E •* E

O S_ a>

E

E ra ) ra OÍ s- — . o

.Q

o sC3

s:

CM r-. Oí

ra

X 3

•P

a»

CU E >i o

C C >> LL.

S:

(/)

IO •p o Q. ai -o ai •p E o Í. sz o 4o E o -p ra N sO) • +J © o E ra o

S- NI ra .E i—

o ra o

"O -IOJ -P

u

ra o •p O) E

3 3 E "O 3 E O ra -P CO a> o •p Q. E O SE^ Oo -P ra O í. 0J E "O O to •P E ra O iO to >> +-> ai 3 -p er ra ai o .O to

co E a; sai s -o E

ra

OJ E O N

ra0) u

to

E EN -P 2 FC OFL)3 (O O CJ

ra.p -P -o CL x ra E C CLLÜ ID ai E CU nj 3

E en

E + J DL+J •R- OI E LO •*•

4— d> — i o o

3 $-rao ti i Eo C M ai C\jco s a> to 3 E co > a> O

0 s- c •i-rao •P O) 4— ra OE Q- •« •i-rao E o í. -P c +O ->í to CM

R D .E "O > O C O I— -P C LO I— C\L 00 E -P

ai a

+

.Q E ai to to •st

•P

•P

E o i^ C_)

E O s^ o

ai

ai

•P E

O i. ^ c_>

E ra ra +j E

— i

Hg-209

E

+-» C O -Q .««I— O -R-A I Í H S- E •r-

to ai -p E o iJZ u

•

li w a o O •PO U CMS«fc«.

I

ai E en ra s: ai

ra

E

-P

ra

Oí co en

E

E

a) -p

LO

o es -a

LO R-

-p a> E en

ra E

0)

ra a. ra i

-P

ai E >

O

ai •p •p CU E en ra E to 3 O a> ai -P

E

E

ra ai

•P E

jü

ai •p sz

io

ra E

+ ai -p ai -R•P NI •I- E Eo O t. í . JÛ sz o +

> r-» oi uo .c cr>

•p -M C C +-> OJ ai .e

fO id iai 0) Oi Ol (Ü fO 3 3

J

O

>,

Ol

3 S- f— > O ì I -O -M JÛ 3 ¡.(COTOWIÌÌÌ o i- en t- +-> o - i - fO lo r c\J «d- «.o r*«. r>» CJ - i- Oí O *t- u Q_ OiCTi r ? .O M- SOJ CO i/i O CO co co _Q t- • • Q. 1 W S- +J 1— «J j DÌ o 4— '— aj S"o 3 i- I— S- . I "O O •r- OJ 2 O) -Q O) T3 1— Q--i— S_ c = 1— J 3 + J U • O V) E a) X c "O fO -i— "r— r— > O O > OJ .O .Q fd-O - .o fd (O ,-0 CT) S- 01 fO "O .Q } Oil— i- C CM • I u c en-o ra : O O ID ID • O r— 0JS- 1 'i— Oi O X) XJ fC > O) E J- > C— O "O i- i— E 0 sE o .Q fO .Q I o> S- XI "•r- +J JD E • O _Q O-"OO) O O S-J3 a) fd a> > S- fd >— l£> +-> Í0 >> E 4- cn E : -i— o oí«—i co OÍ_E fO -Q O ì-r— JZ O * O «3 Ol i> JO S-TJ ove—* 0>4J i' " i - c o s- i- ai "o 0 a; Ji a) —-fl)15 C .Û QJ o o V r - M- J3 O ) c fljf- (OJ2 M- 3 -r- •i- 10 o S- 0 s • flJCL) J j n m -P 4> £ i- ai z oiQjai'r- Eco-p E J +j 4J í. £U rI +JJ3T3 T3 ai ai 10 O TD a» (1) t- ai r O I— 4J r E E -I-i I w i— fd I— QJ C O S- a «3COMa) >) 10 w ' CO $O l/)'r >, fO +J a» co .0 ai -P OJ f— J= 4- ifl .. Q. C C JD S-O I C l f l r 0) O «3 o s: o Líai c U-—' o 00 . >i _ c -r-

EOÌ a;E >o E X X» •1- Q. C 0 0J •r- 1 E E +•> s- O 0 y) cu • 3 E « ti i- E > > a) fd E S X ^ •i-

L£> O KO R-> CTI O Í

aj f0 •o c a) o E s•P ai «3 -p C •P •r-P

C

- P R— -pre»i

en -o 4-> r— C C fO -r-

•I- c a) •rT3 S. 1 L04 . c a» . c c f o " " O U : ai ai -c (tjs-soooj c p> cvjfO i s : i 5 +J W O) 4-1 o co m cti-i- o ojli_ (—o m (0 •»+) — _ . O -i- (T3 -C m 3 _i Ci— -p -p *o0 J | — O i - E VI |Q a i — — -• (fl c > -J «o -M a» a» • a -i- x o c o i c i û - u a s . a i +-> CL) m - p o o • f—LO i •!— o . (O a o •t- fC C LO

+-> s c

I •f- >, ul > w ir> >> ai*«- o c o u S10 C i— r-» . c SW "O O C0 X 4-> Lf> i-< •I- C C0rtJ•!- Q..C • 03 ; ai ai o ai > C 3 «— ï f0"0 s : 3 on o -Q 4 — O ai c -p -i- i( I— I i- rtJCTt J. - N S- , 3 en •P >» (C< a> • i - i— • ai a i i x i • = a> s- +-» ai «.vi •p "o « t/>> i—ai «— WQJ a + J f O r - r3 s: a) ta .c 4- r— l/l • 5 I— a c P - r a a a i ( o a i — vi ia • i h cl - o c «o w " O i i o u un o t r s "O «4S» • «53- c«— o c o - i - a i v j c • c -paio£ ) o -c ai •, LO ti i- -m c: en c •1-^3 > •i- -r- ai a > a » x : s - ro co t ra r- C r - a» jz x • 01 C to O l/> OJ «J +J to "O -o > "O rtJ QJ •«-

r s i f i « t

n e n o c o

3DCTï«d- LO I— un Lf) o -po 4J o o+J of) o+J o+J om op o

+J

w o M f l i n o n

u n lo co lo lo o o o o LO f— CO r— UD ^ ^ 0100 C_)[LDLOOOCÛCTiCOr^OLOLO O O i i h-!o o l o o o o l o ^ l o o c o cr> i— c û .— c o f— r-^r». LO LO co œ LO LO LO LO

ai

s- 4-> s-

p ai o ai nj -p L i/i Q- C >> 0) n3 as o . E

Hg-212

+J

«^-r-»

>>r* >,-a to fO E O -E >— O +-> f0 O E a a i • « "O4— C1J 0) E s_ N O 4- tO -r-O • o E +J a; CJ -p 0) E 4- 3 -I- E E E CD iE O O a> IO a n s o -P ai a> io s- O S3 Q- S- -P•r- O.E o N e ai cr 3 cn o s- 01 O.•r- X E « 3 E N -Q E O O) O 0) • : e o o -o o _ at u w _ JD •«>> e • < ia i - O 4-> ii- -P (T3 > > i- _o "O 4- to en -r-s-T a o i- C Ü (O i- o N I -Q ) 4- «_o -P «3 (O DI E 01 4) £ ) ts) O) lO "O O CI S. rO D) i s-r c E m T- a> e ) O) CL-i(O t. ^ OI î c T - to o > 1« «« O i- O O ,u c ai i •, O -Pr i-rQ--E >— ro CJ 4-> to E X O o ai S- s_ ai a» t. o -E >> o o o. O CL E 0) o E 0J -E 3 4J JD ai co CO1 CT» CT 3 X» •p ai ••>E ra E — CD en to E O ' S- CM•r- Dì o •P ^ 3 XI •> J3 E ra N (a S- ra ai -I-E í. cu S- rra to ra _Q >>*-> O ra E l/lr > E E E cn4- O 3 E ç a>

ro

-P "O CJ O -P . O) "D i- (Osw r W C ^ E (j ~ o o. •I- OJ Q. «3 Q--P E i•I- o cr * crio o uss•i- 3 +J -P -O • OI •I- N O) CL O iE S- -P 0) (Tí 13 ai ra -ien > -p xr 3 w o to ai S- - -P >>r- 0 - r i- +-> c 3 CL J -P O ->.—-. ra -I- s_ flj c r ' CO "O .Q O RD E 3

•

• "O ÜL «L. T/I «O CL 3

AJ T«J

OJOOOOLDOLOO

A> E ( 0 0 1 3 VI1* «— O A> R O — C O AJ O 3

O O O O O O O O OOJIDR-COCOOM LO Iß ( £ H ß IFL Iß I S N

3 O

-O >

(D NJ O

is> c.

3

FO 3

VI M +J 1)

I—

J O O 00 3 : +-» O 3 E > C -I— 3 : •I- Q) O U ' R O II

: AJ O. • TJ 3 3 A3 1/1 C D 1- CT O O D ) 13 •I- 4- +-> A» S_ O 3 I - S- > O E: -R- " LH TR-. D-R— *O CN E A» 1 C R WR " O- T T 3 AI Z) C C O CN CN -MO •«- S_ O R- •«- AJ -— +-> => =N -O

.,_

O O O O O O O O LDLOOLOOLDLNUR» A» N. RS Iß R- R- OI IJD^OKOIOIRIFOOJM

HG-214

r^. in LO P-. c i CTÌ

I— -C +J rO ro

JZ +J

«3- LO

l o r-«* r-«. r-. 0 0 0 0 OfOCO^OWr-CONfOOJ COr-»r-»OOCOLOCOOOLOOCO

OJ c ' r o a) ••-• o > -c c C O O-r W o Ol CT i - 00O r— N >> D C £1 to O O >— CJ> c r - i - -Q E " o 0 1 0 O) o a)i-«JJ c — JD 00 C\J Ni E T - + J Cd .c ro a> «— cj c cn c o . e i—i i— •r- -i- o u 10 ai aj aj S_N*>nJ-0 < D i - a > < L ) i - í — «i- o > S - O C J > S - C J U > C J ' r i . 3 O ' - - i - =3 O -r- 3 .— o j û j s - « — < u a > s - r - a j o u il—ou i j— o —i

00 ZD CU 1— Z

o 00 et CD

3

d) cu

4- O "O s- Q. a> c •r- .C LO +->> > 3 O r— fd fd .Q O" 0) a •r- 01 4— "O c en T3 i— 3 CD fd fd c N 0) 3 LO 13 E 4S_ fd i— 3 O i— 1! 0) 3 O -P u E efl

u o • o ai o s- en s_ c •«- >i fÖ fd d) o +-> ra c sz -P -R- O I— o Q. C +-> c 3 4- a; t/) •O O •P -p a» O «3 4— c • o. 0) CL a» o •P -P E O c u 3 -P ai U O r-» o Q- C co .c l/> T- +-> a> -P C D) co ) S- CL 3 «3 1 u CL.Q 00 (J .a

>,

CM fd 3 S00 J_ E O) c u w 3 +-> o 2 U c • 3 3 0) c 3 S- o E E3 Ll_ cr id N CU U•a CL CU a» 3 D c E O O 4-» CO + u .c 0) fd 4XI Ra> O "O -p 0) IT) r - O sz C cu >i E 4-> O C Oí o S- •i- fd 3 S3 +-» CL.— ai ^ LL. •p- 00 CLC o -P

CO 4O 0) curo CO OD•p c a> E coro LL. O E CL o CO4- -P M o O 0) fd fd •p ^ ai • • QJ E C O O S- vi c c 3 aj e E ai O a i a) E o e s+-) " O c o > sen LO E - r - f d >) >> e c c o . o c j . o •>O f0 to E to o E + J CM " O 4 - " O to E i t o O) a i +-> 0) r e 0 ) i íC ra > > m ZD u O o CJ 4 - _E ai E • C\J

tO - r CT) 0 ) «3 W (O W E +J "O O co O E O -Q " c u O c O ai i o at o OJ c fO a i CT) _Q t o i - +•> E -M + J CT) C U ' r U C CJ E e n o f0 - Û O CJ 0J O E O t o CJ • re o Ji (/> o a> s - a i to C\J _Q - E C L E E E O .O Ol E r— a j a j re T - o • n : •o c n ^ o

"SL

-P e

re

re re re

re

i to X +-> CU

i/) oo E to e E o •rE to Ol O E •P to Q to 0) O- O

•p 0) - P to e n e T D a> o 00 E •!- E Ì . o - p CL i o i - a) o T D a> to 00 E ^ re •p •p d) - P -a cu E TD 4 •p X •r- CJ O " O io ai to a j E +-> so eu to > , o u " O > •p re P z to s s:

re re re

re rere re

a; -p

re

o -a

-O

re

E O

-P 3 O to X a¡ O E to O II m .E O a> - P 3 S- o o o E O > re o i

Scn o •P

tos-P toai s_ E . E o ai E i

re

re

re

LO • en l o E • r - CO •H Q to a» ZD ire il

ai E

re re

>, u

Lü

re E O

H g - 2 1 8

ai o o JÛ LO r-«. oo e n to o o Ci- o a; oto 3 re to 4 u 0) o

. o «L. cu 3 OI QJ U E

re

tnZD

CU I 5 TD - P 4 -

+-> i- •

CO O E E I—

- O d ) t / V P SZ i- E fl 4) •) » ( O l - c +•> -j a» i < -i— m— re 3 +-» E O E : - r - •'•4- S-rtO 00 O 0) to X j O to - P C O > E £ ai o s• + j ijj o " ö T a 1 i- X TO") Q . 5 O I - P ( M U ) (O

X to

-p io E

re

^ 4 to T D > r - -a • E 1 a j •—i cu o o -p X oo O l -p "o o re (/) E r— ^ re o N -P o re u to o o •p to t o - r - - p r—LO t o E en E a) re re o o E •p i— E d l r 00 I CL >>-o E 4— r— E S E to o •— O Q. a t/1 0) a> Ol re CM s0) E > cu < i

re

r— JD E O) to to re

re re

re

E

•P •o E 5-5 E O U a> E T D co O o -P E to CO a; o co CTI E o r— S- ^ O M A 2 co

Uì 0) en

a i O •P •p S - O 00 J D ( J -D 3 00-1J D T D •p to s_ a i 9 J Z v i •p sz

to to 3 E O O >> 3 -rto to E - P 'ÖJ

re re

re

OO 3 o 3 E -I-P E o (J OI

E O •P E 3 O Í. O 4to X E Lü O

E O •r4_> 3 i— •P o o 3 00 O l -Q +-> o 0) o ai ai m r w w i. ai -o cu l/l-C -C E +-> i— t/)|— +-> QJ (TJ ~ ~ C 7>4fd O -*-> E O O O o O LO I— ^ • ••• E c E O QJ E O 0+-> • w ai w o r i/i o s r-. x ta ro 31 QJrtJOJ -P OI -O en i t—i a> o. i/i o s- 3 i- r- WT3 C W Ur- IC 0) E LO LO d ai o c o -e i— E QJ «1 W +)HI KO LO •r w u p oí x o +J a>s »-pe • w s m o LA e •o N í- • i Vcoco-—-a> I ai i- CVJ t Co r co s_ o O) t/i OI34J-OWDOIC o -C ^ 4-» +-> a) >v- c i O' ra >, s- i— cvj i too 2 c +-> i- Ll_ O) E -P -r- i— E "O S_ "O ai 'I- CinM-h- Eí— 3 .C E s_ ai ai qj v p -r a o ají— o c; 4- S e a>:r o -r4- © i. -r- irtOi/i-C — i/I w 3 i/o c • lo o a> >) E E o C O I— Q- 3 — « l O í O E C to QJ >—' 1 *r* E r— i- QJ O) OI 00 > "O 0) s4- a; i c < cni-i a i a s- ai o • ai c o Í . .. _ . .. E E => ai c oo +-> s- -p oom -p 3 i— o vi o O) x o o (O (J m -r- i/l E H X T 3 O U >> o E=J3CE"a-)>>íO Si— i— +J X (D W O H J3 00 -E 00 "O uo -i- -P -—- QJ -P E C 4- — "O "O C\J-P -Icn O — o CLI OJ O to 3 ^ O •!- -P Q- 0J -P N N -r- i— w r i/i c e m s- •> 1+Jfll-rOD+JQJS-ls- 4- E fO -P -P C\J o >+J S - 0 Í - < U 3 0 i - X 3 0 U v a *rQ. QJ "O I— O

i- c*>ro cn-r- o e 4- E i ai >— c > w o Oíora o. o lo • cu >> (J J3 - E ai i— E M t— -i— si— a» co ì- cm e ai ^ ai ai ' r flj 3 JZ II > , P P P X ii- ai o - - a r o « >> rO -O O E O E SQ(O CM-r- p >) S- S- co 4 - S- . CO C_> Q . o ra l o ai E > r — p • - - o ai "O O •>E sz 0) CO U_ E P fO u IO E 3 3 E o o co - o Q - ' i - -rCM U E P u O O -r- aj i - CO c j co p p r o ai o » 4 - CTI w a c c j o E •r E ( U t . rO O X !— C JZ > , O O co o U "O S_ 0) Il co o ai > > p •«E CL-r- CO i - ro •r~ O E O ro E co > e o M a E E JZ t— 1— «t o 1— O U U — ü E

Hg-221

a; OÌ (O sa> > fO

O o un LO ro

" O E S aj ro \ W D Di

«3r». ) i— CM LT) O "o iO

fO J=

O l ' I CJ

+J O l f l J c J_ S- fO 4 - 3 rO ' t O JD 0) •

E O

E co Í0 3 4"O E fO ro

na aj + o. >ÍM ro a>

a» «— Q. E

So

O CTI CTì

- r-» «a- r

CM CO

J CM CO »

CT» CTi

i r--» l o c

CM C£>

. CO I"-« r

cn cr.

co co M û t

CO CM

ro

Ol en

m CO N C

-

un r

> o

m c

o ro en

rO .

' s e . u l o «3- < (O 01 00 P +J C W — o lo c co io o o : CM I— r E T - -r- E S- I— -P O O I t 11 4 - 4 - S- . C P P >)T3 c ai r o r o m P CU 0 J - C T 3 O O O •i- N - CM CM CMO C i— ' f - c - r - X 1.1— ai a> ro c o 3 o ü < u - ü . ¡ o •«- u O P P I— C 1— CO ai a) - r on c l c X •«- S- -r- 03 . e a> Q- o I— co co o

o CPi cr>

ro

P &(O CL

O) i E LO Sen 3 "O u > o co i - -rs- O P P -rfO u sQ . ro C_> 4 - -7¡L. 0) CL -E sO 1 fO O CO CJ ro T p 43 T3 ro -Q 0) E P (O to O ro o S- ai 4 ai > o p «a ai . E E JZ o O

> , ra P E ro M rO ai • -p o o ^ cnvxruffli-T •o O O I— O -M : a) oj csj •• ai ai E ra 0+Ji— O oo O -C SE (O < -r C Ol+J ai C •r- I— O S 00 S- o U «r- "O -i— ra 4— o cs-e +-> -a c tO -P -r- C fO Cl— O) w ï -p w m o o .c rei 3 o o •>— -p a. roo -p "o -e •«— E 0 0 < a c - 0 •m oo • o o j o -r- , c -a c-i--«- o c c .c w a> o o -p •>- o -p i ra o oo s•«- o ra en -i- -r-r- -r- ai oo -P 'r- Jtoo-i-Pt/itocr— E c c * - — « o a» WIE ^ fl 3 J O ' r O T T Ï 5 T CI4oj o >> o o r— o •>- s- a ai-p V -p c S- O -P -r- U -C -P .Cr— > 3 -P -i- *iQ. •»->).— C Iti U O < W 3 ^ Ol tO X O.T- c-i-4M -r- j : >QJ .sai i— a» > -P ai o-i- 10 C + J ' i - e i U O i — rtJ • -i— OJ OJ 4- -I- U i•«— •«— f— E ' - r M fO W +J OJ OQ- ra r- 5 w l - r - ' i - o to > • , ra f— oo o en i ¡z oj ra +J o o c l sic i - c v j z r ai E -— SE O to - i- o -— 03 Q. -—- ai + o 1— +-> 4o o o .o "a *cji-aj « c w r -—' Cl— to -I- Cr— o •--+-> a» 3 o oj ai oj -r- ra u + J c-m^»— >01.». -u > ^ Q. ra a» 3 4-> O 0 3 i- E13wO3 0) "5> -oa>o>i— i - o s - a j •+jo)s-e+- + cccaio4--Mcx-«-cr>s-ro3ro m •— i O -r- C 4- -r— C rO Q-00C n3i— OO Q. > -O -f- -M O C fC C 3COU- E u E ra o. to o ra>ii— • •— i O-C to 0 0 . 0 3 " 0 oo ra lo a» c +JO +-> •>- E i— ra xi r - 1 : ra • a) -r- X s- ra o M— -r- a i o c 1 1- r - o C P a O E - M K J V l - i - O +J d) ITJ CTI +-> QJ Oi— o 4— -r— E >>l— •>- > • Sa» 3 -p p m w J3 m . s_ o OES-+ tocs-E-cO'i-^:ro—'ï«o> ioco ai ai p s. u jd r co +->o>ra +JO o E>— o X jz co3t-MO ^ I— Ll_ OÜ-M 3 4 - ^ O a i O C O ra ex o \ o ta s-»-5 o OJO r~ 00 I— M- ai O * 00 JZ -1- i— . -rra -C < S-+J|— O i - C 3 C < O O J O+J • +-> «3 'I- m o Z5 I— OJ to «r— "o o "o i- -M ai -i— 00 c i 4- o ra en 3 c c o o cl 00 o 1— o c ra iaracOQ-3+^-D • oo — T-+J -oo q>+J(o c w N i-roi ai t- 00 + c •>- »> ovip e mu o c 3 T-ro rtj4->co S--OCO1— oi c p 3 i-oxai o o c o c - i - j - t o c o o o Q.+J a)Ll>•!— CL ) S- C O to « ITJ ÛJ So -M >s Uf000-C30 -(-> o E ^ D D m I— ra M — S3i— _û>N EO» E -I--M S--M >) +J +j JD !— -I- O Û I 2 «— -c: o o - o s - ia-i- e -i- i— o s S-'i-*^.CO S- OS- S--0 OS-I— -M4-> o - c i - o - o c - c + j o - p o i-roiara x oa.3oufd-i-OfaotoQ.4->4-Eo

Hg-222

0 0 TD to II 4— +J LO 00 0 0 O XI O a. co 0 •p 0 0 •p i— -P ta OJ > 0) +J •r- a> •0 s» < +-> to 4- •— TD 10 rO c < i- 0 a. ra + ^-i c •p -p í. 0 c •1- 40 0 to 0 1— O 4->\ ra 0 0 0 N O J- O í . +J 0 O cl c i- CT ia 0 O C s_ .c ra +j to ra sz 1— c 0 +J to •a r— ra i0 +J 3 a) 0 to c O c s3 0 4- O 4-•1- 0 CL > c O -a c ra 0 to 0 Q. -M i- t/1 O 0 E 3 O -CO LL. t- 0 •p 0 ra ra O N 0 +J 3 C OJ +j +J 0 1 ra J= ai s_ +J c 0 > 3 0 0 IL. 10 0 to -P r— 0) 0 LO +-> > id re» n ai to ra r-*. hi- JZ O '1- Ol O O 1— c 4- to LO 1— ra ai 0 Í . •0 0 * •1— 3 4-> la i- O ÍÜ o o s- o • 00 E ^ 3 -I- +•> o o p - o w o 10 u S- 3 C to C N 0.4m o •>- •!OJ S- !— » m ^ a. o i— +-> i- (C ) Il 2 1.rtJ-P O t/) > coin 3 .c >> • O CO O O i>00 •>—•>- u : o +-> C 1— o •r- CO o ai 4. 3 o T- c OI O N U) 4-

) co >, e i— LO s- «I> . . o 0)4-> : o ^ o -r- in

o

ai •p

ro


J ai i— so c o 3 io O 0J ai ai o o s- LI- > J- CO Q . Il co o m 3 i— • o O O •!- c\j a C X I OJ o ai £= O - p ai MiO u . x : •rN •P - P SQ. CO ai o • i O w IO Uì U • p O co LO 1— CO 01 E •r- «d- > , ro ro MS- 3 • sO -i0 f O U ' C0 II t - o c > 01 ) -P o cn > > < c l o j • IC j "O • *> x » • -r- O c 1 i - CO O o Q . Q . O di > e

•» o n • cn

W 0) 3 s- . a 3 cltO io co - p

il OO QJ Ol 0 O C C l/l n Q . o CVJ'i- -r- «r: i— ai 01 Ì - Q . C L - E • < LL. O W V) I—

Hg-223

01 - Q + •«- r N 0 O i— . J_ 0J 01 OJ io l i . C QJ -C ^ + -I- +-> • -P I— C L - I «o: co e • XJ • + I 3 * C 01 S- 01 "O tO - P E ' O 11 S- ifl - c n \ . s_ j j - s- . e - p » r u u - x> sz n N II CD • 3 - S3 ro i - ir> o 3 r r r v r j o • -P

ro • — o c o '—- x > : o Q. r + Mcoro ai : i• ai x :

ai r o + S - O iS- e o o 0J X I -P c rO ro E co X o i_ o . co CL rO co ^ ai U ^ f

3

tO

XI co rO

•P ai C - p ai

-p o x>

d co

s_ •P o E II O ro Sai t U - JZ o o -P LO U c s: o +J • CSJ co r-. O Ì O XJ ro Ll. O l ai iIl OMO O X LU ai CL •p O II o

E l/> O ai u ^

x> sro ai sIO E CSJ co ai >> CL >> I— fO o co a i x> O CL o ai s- • p $3 rO £3 N io rO ai CL

OJ O O CL a i c •• u j ai -p ii •«c X 3 ai oo 1— CL O ^ "O o •p O Q . co E 0J 0J ai r—UO O X3 CL c •p OMO C 4 - io O CL c l ì . O 1 co C/5 ai 3 c OJ 0J < X ) ii ro o ^ ic in ^ «t o O X I 4-> t . CO ai Q. ^ 0J u CL >>-C •p 3

TD

a) •p ci s3 J D N S0 3 m

o

0) •p ra "O II •r- >> E -P r0) XJ fO "O he C 3 •i- J D O) fo ai O) e «o .e e O -a •1- -OC J 0 3 C — i .E • * ••>>i •p r — C U 0 3 1 — to < D C O O 4to co c isO) O •r- foXI -E + c C OtO h» l£> r» C T I «dto e +•> co o 0 3 to -—»m Q. oí a> c > rs (\J IÛ roo o Il 2 o a» r — 4o : «a- — . — . — +-> O 0 3>, O 1 O) II Q > p E S-f— E O Kl .a o ai "Ou er Q. + S- 0)(O o + 0 3 to U -r- " < Û C &C U JZ C 4 — E -P s + O 4O ai c m ro ro ^t r 4- 0 3en ni ai a TE ai co E N L O 4-> C E U — ctí ai o o "O to-i- • CT1 — 1 — i- 1 (3 t/) Cl cr> > M 1 O O» 0) E to E O •r- «r•P t— r- -P t. O fO< — ZD E 3 a> i-P O C U X c u s.fO to «5Í--C tO rtJ a> 3 -a +-> en r--co •o ai to • O -P O O O C M C L JZ >, t + s- ai o3 » iC M P C U J D co o E i 0) E < T > J3 3o +•> «-3 + O— O to fO COtcM- O uai O X C OO ül U (D — < • — L O Li_ — ' • 1--. O -p • r 0 3 — tû rit^S -a > ^û CTI CTI (V, r o>L O C M N d) r •p —c S-"O • z: • 0— i r- 0 S C L Ö 4 » en o Í. 4J- a> 0 0 U »• i- C o s: iU CJ C U O O"O 0) •r- < _^=4- E -r4 —•r- S X 3 o tO 3U fO •(- + O CM -r- 0) C L O o s- ai 4- i. — O . C M C \ J CTI L c u O Xio Cl) o J i-H U_ 4 — _E C M 3 u C U1 O 0) < u - > — i C J > i- E S-, 0 3 -Ijz o a; co o s4u_ O) L O -P -M C o C — ( L O O -c en o X to ra u to o +J s- C S JX EO o E co o co r-- o o oo O P i — i •1- E •>0) -1-3 SI 0) + L O — o _ 3 LJ- + -o •p a>c a c i/) jD < h- X C M 0) Cd O e •r- 1_o to ra ovi— X S— i s i- 1 — O) fO to ai•i- ai a» s. C Û. O 4-•P s. e +-> O 51 o u- 0 3 Q. 4 —0 C S I 3 O II l/l X L X oC m • IO IO M 00 r L OL O e E T-N -C i- •«+ S- C M 4-> O "O•r- 4-> ^ «SÍ--0 II 1 — (J O .c a» O 4 — en O Eu cr> o U- < T 3o s Csi E i>¿ fö "O -—- 3 c X E+ < U r O C O ^ O Sa> .o 3fljLi_ — i a> C L~ V 0 0 3 < sJ o— i E C Û — i Í. to 1 0O Ol C M in 1 dì lo ^ o> — ' > Ncooioroi-io3 sz x¡ CJ) -P cOÌ > o fO fO £- 0) •r- • tO -P +J N V• )o vi tu io a» ai .e -IE s- -p o S- 3 o O -P E (O 14- X • fO Oí < u •' 00 O J < L > ID U O L -E -P E fd U - < U -P O •r- C : c c •«- o E t- j= - •(- o .a-p o a) o J Q- O i. 4- -iXJ .E 4- -E C O 3OIU-I- 3 i a» o o : _e est i- u p to i i - > ja xi • 3 a; •1- -O o +-> ) .e + «o .c «I• en O o to IO ,—, o L J ai iE< U -E oC en D. to .E a> t o aj u >>z •o +J^ 0) fa -p 0) S- o ra -p O) to o E 3 Oí-o o -p O ^ «a E to fO E E -P fO L E a) C •p E m aiO ra X a> d) c L•p o o C U o E S- X -p to 4 —O)4O >> a) fC C JC L S- 4- 0 0O (O c u aj O 4•p -p "O •p t / >C MfO to c u 3 Sto 3 io •o 3 to u E fa -o aj ra ra E 4-E "Oa> > fO4c u4-> O) o •p s: to fO E a> 3 x> ^ 4o o XJ ca a s_ ai s: S- u E o JC X» / >o -p ra t Í . to -p en ra •P ^ -p ra to 3 E X 3 3 ai 3 >> o o -p > > E fO< J > Sa; 0) »-p -p to E -Û o O o o 3L cn oC M(O E S- o fa O X3 E at O J3 0) JZ X1 —en o N t- o to o m -P S- E 0 fO J «o o ra SI Lia>-C ai C u + •p •p Í . 0 J ai •i-CM E 3 E j= JZ c u O 4- •P h- 3 L Uu O «a— E

>,

1— r— CTì c ro N (O i— - r-- en : OÍ i— -P c co cu c o c o w li u X 3 -P

co un o «s-

ctì ou c\j! ro CO •— 1— O O O O] O

U"D (D —ICO) «O Jc oo o • •— I V> •!+-> i O ro O O • > — : •- _£ZS- CO CL Q) •— P -P -MISO "O P J= 0 CJ CM C -P 0) < X) Ífd V) -P Cr- 01 0 O 3 •r- O 0) •(- c c -a -o H •1-4- •r- Il ro c E c V)4- CL ro fO CM V) O -P O-r- OO CM Q.-0 O CU w •P CO -P I— •!E O to • h—s CM -a a» u -r- X id •• • o c QJ «E V) d) OJ 1 '—3 1CM 2 CO E COí O r— r- 4— : co o *>— E CO • o o r-* CL • co r— fO O -P U 3 vi (J'I- O • c •r- P í- a-e co->nj -P O i- VI IV -P >, >, CL 2 CM en V) î_ U E a. 3 fO iV) Dì 0 sO -P O Sc a) 4- --- .

*— S- 3 10 1 3 s-Í. Il 0 u0 C C c •1- en CO • < — CM fO fO a> • -P S •1- t- S- VI 4- JZ o ro > 10 M f— O CM o V) V) ro c 1 -r- -P• U CM C l- .C (/> 3 fO l/l i -o o a> tu t— I W 3 S- >> CL S- 1 (O i — vi -i- i— en c Q. U lû 0 f ) • < — 3 "O oj o • < — fO 0 . •.- 0 C LD O- CNJ +-> u a> •« sC 0) >> a) a ai c iC X Il -P Lt_

•r-0 E ra c si 0 +-> 0) es u s: LO "O •E o "O -P 00 •rT- i— • "O fO ra QJ 3 00 ra CJ> c i o o" o c o a > • =£ •—• aj coa» •«- c -p : C r 0 0 \ CL o o fc— -p •I+-> "O E 3 E • w o +J m 1> "o a» -P i- -4-• > "O r mt-ui—ai • • > Q J a > I D U i + > ai e t: > -p s- i co o c oro E (O fO OI l/l ai • 1— airoracneoooa) E W O W O" U_ CSJ -r— C > Z r- Q.CO E o »i— en c o c o cu -i- i «=3-un /> -CE -c +-> XJo o t í- O r a) o -p o ii a • 1 en __ - . «S^ QJ J- V- C (O rt]c ní - C /> +•> O -r1—OJ E o r w ai c o-p -p « qj c r> • -r+ > r a r a 4hi ai o a) en o 3 •»- cu o. 3 U £ O I OrC -o en -o nj-i-x c-Q _Q o. +-) ai en a) CO l/) C - -P -P O LT> OJ i + ai c c fC i— x •- ai -0 X cu o lo coco.— en a> a » jd a » o > — Q-i— C - QJ r— E +•> OJOQ.I r- -P r— •o -i- s- u +-» oo i— 0 ai ai •>- -o -p o. — -fJ O -Q «a-a O E .E O) E -P 00 -r. £=i-Q)3 >i0) 0I-S"0 a -o Z3 4J-0 3«—.JZ-i-r— ••-> r- E I O) 00 CJ • IB OSC f H"—"i- QJ E E x -r r E a+J ai o. «C t/> 0 aiauowifli-"Dfl en at o i— o .c c -i- -rW lO 5. D -r- a ri— V " '+ fO s—«ro E — 0) i• —> nCO3 C/)"aE 4-a»rCa E I CL — O -r- >, ta O E a> 0) i- r • E-o co-p -i . uic-ia •!- ^ u u/ >- C +-> >,ra+a» e+j 10 t O U E O O +-> •—O >, CL 3 •(-»•r- CL •»- C " > E "ta tu i- 00 -P O -i— wnj w «>, O»N X O) -r- ro o • E o • < • • N • • > 0 0 T 3 S-,CClflr-J3W'r • ai 0 vi É JD -P Q.04- C O N r— -P E a» r I -C CL • 0 c ai i/i x q vi i/ii— ^ •" cl DI C P r Q.IO +J O CL i/i ai 1/1 •p(u>aiuaj3iouw O CO M- -I- >— C •i- +-> O E CO-M +-> E E r- O) E "P oo E E QJ -"O O CL O co C -i- S- CL + O r— l/l >, C O o o • ra a> ai a ai e n cl >>.— a» -r-E 3 > • w _ w 0) >> o < 0 3 s 4 -QJE "O o VI W Q. +-> t/> EOQ-OdJ S-S~3+-> + J 0 4l O 1 3 i. + J i- • • u ra i—1E a; ÍO >> 4-> -r-S- CL CLO" I— D_ +J (U P CLl/l -r- x ai •,0 0) C ^ S- • I _Q ra 00-P 3 ojo O ta e 01000X1T3 CL l/) -P -C -M S- •«- O U U 0J OI CT)*r— E 4- oí en ra O" o •«- c n o> E -p o • • r a 3-ri o o c s .c 5 O Q. U •>•!- OO E O •— 4- "O p s. x •r- -M i— I O O U r— d) 0J -r- -ra> Q-o ta ra •>- i o ai 3 > i a> -i"O O N • c o •r— -P tO O XJ • -p -p "O 1/1>> ra en O E •• oo - c+j i-ts>a> -P W X£ f0Es_T-C2 O I «— 'r- E 3 0) C c o 3 f tu ••— cu en ai XJ •>"D > -P E El— -r- QJ a i e n o a j o o < _ > • • — ~o o n * i — < i > i — r a c ^ LO E O •" i- «a ra os-c-c-m-i- a» s- .— en >> c e a) 5 -p i. w x u -a • 1 -ro «i-«!- i— e +•> t- -m "O «a o i - a i E > o a > o " o 1 cl o u e o en xi (O r- nO- i 0 0) E Lf>sd- ra - O T3 -I- r- I— E U C-P cu E 3 coo-oEcaj-i-Eo • • -p ai oo i- o -r > r i. c ai4-+J •!- O'— 05 O) 00 C 0 c CU4->E>0raCS-0- •1— en -P QJ C X en-i rO 3 Sa> 1— OJ co XJ C 4co E O QJ QJ QJ -r- Sen -p c E co —s "O — a» — is- => -r- o a) to jz 1 -M C\J ai E s- +j a) 2 LO-r- 1 to • to CJ1 u O to o a. -oe e oi co oX o. 4-» -P r0 > > • » - o «O ro _E .e •1- a E -i- •» E ai -r -p n u i- JZ to 2 o •"O) 2 (O P CO r to 4- — o. ai to ~~ " X) ro •r- -a •o o-1 ai a. ai a) .. . E -Pto .E X> U C -I- IO -r- •«-:o -p O) O tO -iO — i r•Ti— i 3 o"— i >O» c +J Oí o.— a u iai — u cn4- : O) 13 3 Q. -E U rtJ to + E "O •P -> top "O -P u P to Oí ai E J_ S- tta. to E -— SO 4— O ai aj ai ai • • px:•p to Î . E XJ-3 •p o 0) O rO •O u io fO O C -P U X -—- fO E >> E ai ra •p -Q io o O iO -o -p -a E (O E to u a» ro cj > E r O E Q J •i- ra Si•P rtJ «o ai 4— to ai E 4- J_ CO •*•to E LT) 3 < 4- rO 0) to CT en c E to ai 3 C to T3 3D•«-o O E E O) O to to to +-> •P O +-> •P < TJ 3 •r-OTvJ + QJ Sa?

E LO a> + — •i E O O E -P .Q co O Sc -E P o -P S- ro O D- t- ai ro co ro .Û (J ro a> eCU ro Pro to -p u ai E E CL E «O E (O E oO) CL JZ Q. E a> •rO 4- Ol • E 4-> E rO s: oo 2

o o ro re u e < s- «p- a) - ai • to i wa io a) e s_ a) cj c e ai a c ai m E (O -i- .E IO o-i- E -E CL-r- 4-> i. U ai-r (O Dl+j +•>•«- -I- £ -D £ Í0 (0 ai c 4-> oo jr +-> i to Sx: O) co s 1- 0) 4: O to «C CD to E P P 4-) I 10 O «T3 rO -P S- 3 > O) tO > O OO OE i--p to — i -P E o COCO) i a> — i 4-ai c «i- to I p a>c r i a) — '•r • E 'r- O ' O Q--rI i- efl 3-E -P > a> E CT) x: U -p 3 fO -p E O cu CU ro O E " QJ to O "O a_ í. JD O sO E 2 e O -P Íai rü (E O OCTI -P j*: I O ( V to E 4-> CLE Cl CL O S- •r— 1 O O < 0 E •P S- CL O o •P SI ro fO Oí•1-3 E ro 1 C_>0) S- to E 0) SE O a) 4- to 0) E E (N O O io < ai S- rO CL i Í. lO -a ai E to E CT) 4— a)ai a> s_ -o E E o to ro to E CT) O ro to O IO -Eo to +J E í0). O E O) CL -P 2 o 0) CT> +J to X JD a> o E IO O 0) Í . to +J >QJ E O s- u to 0) tlro O •L. -D -p cu J=> 4--p x: ai 0) io to f O E -E rö to xi X) E O p Ì 3 S- a> E E O O E ai 4.E E "CT o ro JD aj ^ S- rO E o ^ Ol E +-> «o h- C ai 3o a. IO >"O to a> O i- ai o CT) E < o -p ai a> 0) E CL CT) iO (J SE io -p -M > E 3to -a "O 3 S- rO > u ••• fO o < to a; io S S- a> QJ 2CD ai ea> a> -p "CT •P -p O LI-s- E a; E E ro to h- CU NI • CL -CT ra E Xi o ai -E t—

Hg-227

O E CL O 00 to IIO C to oC\J aE> íÜ < Q. o en O CMO O C\J o Cn to — j E -p O lO rO to ai ro E XI CL5-S •P SCO 2 ro E :0 > LO 3 _J 3J LO rO SJ o S- 1 ocn O 4- to en s: < tos: ai 4- a; O Cl to E O -P CU • f— co £ ro O •«-.£>> : +J O 3 to E +-> r£ 4-> • > O)

vi

£

>

• JD s- e E c

c o

M- CJ S£ -r- O)

m 3 • 4- o ai -Ir— .O 3 •o CO o a> o E •M £ c s- E ra O £ O 4U O a> 0) o s_ Svi 4-> o 4- ai i- ra "O to J_ E O 3 •13 a. 10 s- > +J a> JD ra o ai 3 4O O E S_ cu s_ 0) J3 ra «a £ E o +-> entto to s- c ra £ 0) 4- E ra o o 01 E -r- 4CL £ a> -o O a» +-> ai>> ra XJ E 0J O to co cu >> 4- XJ o +->-M >> u a> +J O iIO o > to fO •—- ta to aj •i- en c E s_ sra to O) O 0) QJ -C 4 » i-4 > cu ai o o sz o cu í. E o C o -a í- i. to _c: e= CU E "O O) O) +J r to +•> c E 4-> CL e a to ra _J O a* ra •(i- 4—+j CU o 4- O E U c _c ÎT5 o ra c 4-J < c: aj r >> toto a; o a CvJ 3 i- +-> a> c E 4- en cu O -Q QJ +j o c > a> O e +-> Sa CH4- ra c +J +-> a; O ai >> £ Q_ "O itO O E to E V) 3 ra 0J cu >, < D a> o Q. > x) to 4-> 4-» JD 3 ra c CL e E to e E í- í . en UÌ +J s- «a 0) c E ra s- ^ 4-J O o O CT) 4->to C ai o C 4J 0) u Ít3 ra +-> -M SI -C «a 3 4-> 4-> a> o -a _c -iz aj "O 3 i/i U S- c •I- o ra U £ ro+ ai "OCO r en o 0) 4-> eu u. s- sto 3 o •a ai to tieu r to ai IO h- ai 4-> < — to 3 ai £ -p 0) ai to •o en to en ai X w c i. a> a» E to ra W ai "O c O JZ ^ en ai 44->4-5 to ra -C o to E i— O to ra ai O E to ai c _i 4->-o co . 3 c ai o > o > O) QJ •O 4-> 4-1 aj o X) sra ai s 3 o ra 4-> "O C •r- O aj to U CM E to ai 4- .c O ai ai Sc 4->4-> -o > oo O X rd ai o C r-. —• ea E 3 ai o O O E c •a en £ to c s ra Q_ +-> s: c ra ai to 4-> ra c 4-> 4-> ai O r o X ai 3 -c CU 3 +j a» u ai o ir» ro 4- E ai c 3 > ai ai O • S- s- to in < c co CU 3

>,

Hg-228

C -P (/> -C ai • o to ai QJ to S- £ 4-51— QJO 3 QJ XI rsi JD C T•1- to ai i- CL C s4-5 to O ra tO O x: + XI ai ! ai to ai 3 to "O to 4-> ra «3 ai a» 4- E £ QJ C -o 4-4-> ai £ •>- £ en 3 ra o E o ai E u Q •a £ Jen ITS a> s- £ -o a> to ra c to ai E CL -d 4 > .e ai IO 4-> ra ID U 4-1 s_ to ai £ i- a> 0) 4c 4- o J- O ^ 4-J ai 4- ai 4-> XJ ra s: ai t/1 4-) £ rn x: 4-> rarO XJ 4-> £ CL to ra ai c 4- E ai r ra O ai 4-> S- OCJ — i £ o >> ai ai o s_ T3 So QJ 3 U. O) to o a 4-> O ^ to CL Q. to £ O S•>- U x; i< E —' CJ o cu x: h-

LO Lf> OÌ IC _Q fO i—

00 ÜJ hZ3 00 C_J Li_ «sC S C o: I— _i a z: c o U_ s:

co .o c 3 «— r-> CD o ai 4-> ro en LO -a Ti 3 t— +-> O) c ^ -p to to f0 E to o C (O 3 IO a» OJ 4- 0) - — o a> E •i- c a) CT) O ai sIO e 3 e CT) h- aj 4_) C (Ü CD f0 S- 3 a> +-> E © to o CL) (ts ai -a O CO CU S- •P Qcu S- ro 4— 2 "P to 2= "O cu E O fD JD cu cu s_ -p n •P cu fd fD "O o O a> to fd en s- -o > E to c +J S «J c O fd CU £ -P -r"O 3 2 ro iÔ-S Ul a> TJ O (T3 0) >•»4-p r - e ca- E •P -P 2 cu o o c: to (O to O 2 ro >> S- O "TJ i- to — to o 0) •r- ^ E > c -P -P E O a> a> o s- a> aj a> s_ O _c: o •p _e a» cu p E c: c o o > ^ to a» -a 0) S_ p a> cn4- ai E • to d -p c fO 1— +J -p cu o E cn o ^ "O cu rtj u c O OVi_Q E a) i- E 00 c Q. •r- E 4a> «O to CU E O •r- fc« E e XJ 0) c a) O. -P -iCM S- O ia> CT u o e CU O 2 O o. O to E >1 4-» to o . 0) 4^ c (J 1-CVJ C p ai "O E CL "O O 4- E S_ _Q 4- •P O CU O O •P u aim ^ a» -p E u CU >>LD 2 e to -i- ai • r-. JZ o •>4- i - .*(/}+-> O) o ai en L-—- > O O O Er c ai cn •— " ) a» i- o — • O) -P i- -a a) 1X3 •P 03 2 E io . c o to 0) rO O Q. a» ai in lt> a> 3 +j E CD 03 E c O a) CU U to >> e cos E E E

•P l/l fd E E -P •r- Id E U n a) • •PO a> i TD o CJ o + 1 4-I—

ta CJ s4-


- O 00 4- 00 S- CXi t/1 CM Q..C --—« > ) 00 i— + CNJ l (0 O O 0> «3 II •> c + co a) o i- m H 0) ra i — í co en +-> o -P O I—• O) LO o 4-> •r M OI • II 4- 4- IO 10 fd c (ij s «t co — i o < at -P -—E 3 cu i - E rmicoOTO+Jioicifl h—-O Eu—>>CÛ rd JZ .O "O fO

C rd x: CO o ai ai iai en -p >> oE û. to iC i— 2 + -c: rd to o id ai 0) - P U- o x : -p -»->••— 00 ai co eu C L ai ai -> to aic e O -p LL. 2 to Sai cn^í en o ai ai S- 03 o s- 00 ^ co .c o E a» o id m id •p O -C -C Lu ai o +-> CL > ) ^ i. T3 t— ai o C to ro J- to CL -P c «d coo o to C S- T- •P 4-> C M U . rd o rd 3 21 i— E _C to -p c c ai "O •p to C fd ai o SO O 2 CM •1- -p Q . S- O) to rd E ai E o id d c o Sc +-> < ai o O id o i- • 4co i4— X O rd >» O S- to ai id - P s- C rd C0 Q. "O E to c o •i- > O oo 3 TSICT ^ CL N 3 O to -P 1— cu S-1 ro r-» O E S -C •P ai c rd u -p O en s- •> eu JZ 2 r+--> - +J C -p X 2 Q. ai co O c +J ai 2 i/lU'i" ^ U1C > ai Q. co E fd CL'»tO -M -—.E --cp .c en z w o Soo Oh- O "O 1 *rC C — i S- c 4-> "O S o 3 o -c 4- u id s: r - CL O -p O •-P Q_ c ai

O O + « i/i í- w aj -p id c -M 2 "O U -Ien >» a> to i- ro -Q .e -i- ra o. hE ra fe-S X o a> ai c un • i— .c «o 00 CL 4-> o E ^ +-> jH o * q_ u s« a> o- c n +-> +j 3 ~o ai 2 •»' c JZ S_ 4 - O» rd +J i— O O ^ VU to 4— to to 0) Vi O CM o Cl— E o .e o -O i— a» -i- O.T- ra •I-

- «s o c 3 +•> ai "O - « o ai > • "O C 4-> : +j ai o oo > 2 1- < -> T) Í1— >> ai c J J2 4- -r- ai : ai J^ S. -P lO o c o co ai ai -a i• co i co : l o i — ai -m )r IO s- m

C S - + J ( D a ) i T 3 V ) S .

+•> • E 0) c -c en-—» •ia>f d£-0 0 • P rXO oS - T 3 ito- O3 - O OJE^-aiOJiEi— c—' N > en a) — i o> to -i- ai _c o -— i f a C O) I— i I— + j 0 0 E its 00 .— c o m «j ( U t ai c a) ai c ai 4-> jz en . s- -r- wd!i.w-Ps.+Jo4J «3CU+J ai ai u >) -CCCJ-tOEOJ c — i (O-i-uaj to e o = +•> o -Q O 4o) a» «i-p ra c • "o :~: > i— 4- a> —. c ~ • /i> CD •f— to fO i ^ra i-a) aii- -ito .o o a> a; . o E o • to -c c en o at +J < •— cnro+J " en li_ E o cu i . . ^ 4 Û . C +j •, 2 3 T3 •• ai -• ~ ro t/) • • en > co W A3 WfflL i c c fd en + «j : o o «) (0 flJrai X • ' U J3 1. Ll_ • ! Q- (O O "O \ E to E : irt 5C H T j Q :=)-—•• • r - OID " Q) O id LI- CO

Hg-231

a» »ioí 0 O) t- >, id 3 C O fd rd ^ E to 3 i- to to J o 111 -P •> ai 4- +j s 3 JZ -r; - p >>+j + jn ai 3 C CO) • T - O cn-P J ai ItJ-r3 -O N E A3

» Id 2 : oo X E oo ai — i + u ) s : a . CM E 0) O : • - o c T3 ai u -i- — i iu

û

ro +-> o (d

• ìrs 00 (O LO-p CJ 1—E CO *o IO 3 E -a O e o cr + n3 4- cn s: cu in E cn.c ai enr O E jd ai C/1 +j * •r- a» JZ -E 3 a» +-> CL-P s_ E a» rd E ro -Q — 03— in (O + ,+•> o • ai «j a» o iE +-> rO OOS -OC -—• sun U3 » £ o ' Cvjjp < cu "O I— E -Q

+-> +-> 3 C W -—« +-> cj o «O W 11 iQ. 3 CJ 3 .—

> UD 00 1 • -i— - C\l 00 ro 4 2 « fd W (Ö•p£ e o + W D O s: QCD E -O >5 t/> tn ro c/i >> io E r— .O - +-> f0 S SI 5 oo o aj .C • CJ 00 00 CO 3 >>

ro -P en o I •i- -rcn U J (U >>m

a> >) E 1— Cn ro

CO 3 cno IO CVJr—

Hg-232

> CL) +J I— E .E 'i •r- -P 5 "O S O 4-> +-> ül O ai E ro 001— Q. O -E -P (D

s_ o Li-

CQ

+-> aj c to U o S-

ai fd c OI o. to to to

to

to o 3T

c o QJ ro C en O CL E E O O

c t o LO 3 ai •p O 3 c cd to +-> to (J ra O to • P a O) •p O) o 3 fd to o +J 0) c (J a> "O c •p c H c r— ' ai U X O QJ O

OJ4S- O

i- C o a) C E r d «— T-

CO

r - • « - - P s- a c T a) i — t o a i s (O a . i E ^ m Q j i — QJ D C ' r -P QJ I «3 COX r -P •o O) e n •rc o N -P >,•>O e s a a 0 o to >» J_ C M C I 1— _Q - i - O) - Q 1 CXk— - P r a i— O J O - i - - Q < ai

O) •IOÍ

3 fO i i—
to 3

o (/) JD 3 to +J «O

1

-a 0J +•>

1

TO

.

i. jq • •!p— r • DC

X) a> c sS- (C 0 .C ». Q.' 0 C •1 -P ai - û -r£ QJ a> 4-> to S- JZ p •0 O QJ O) -P M P CL «o 0 N 3 i0) N •r0) •1- p i_ s- «a- X O S- P -rO) ai • 1— N (U s- 2 JZ -P CT> £ro U P OJ to a . - o «J • 0J X i- 10 c OJ OJ O) (TJ > Sc 4-» x : aj C 40 co 0 O) a . 0 en i - to •r- QJ 4P P a» r-» ro • x Xi ITJ — to p 0 to P N 0) QJ O O t. to to P C i_ CL-a (TJ .— a) to SE SI 0) P (O fO i - 0 sc •r- c e n o u QJ 40 a> CL O 0. 0 OJ 0 •• TD to N P C í_ Ûto •1- p ai X 0 ai c a i ai c p x: O T- C p OJ •1— t- i . CL p P O "TJ to CT) 0 en c p •rc s(0 CO c a; 3 O • E X 10 a..—- 0 N QJ E f c S S- tu í - ai X O P 4- - p O 3

OJŒ: •r- Ll_ «-—• C > — CM a> + o • XC\J • S o - — - . s- o s- r ^ o i— >> «a4Q. • CO o en ai ai a i T• D 2 r-ai c: i— c CU -R: o ai m • QJ CM > LL. . C

43

C 00I-i— ti) 3 Q L L S_

O

c (TJ

H g - : 234

•I- en qj c l s : i-Q to - O.

I— I Ol LL.

O

(TJ 'f— P i— CO (TJ

O S- i— r i o o m +-> r>. O •i-CTtP 3 i/> P •• a> c CTW i/i Oí V en

-Q "O G io C IQ V V) •I"O QJ co c O •rtO 3 rU

•rE > lO P QJ X en OJ (TJ p Í. O J_ 10 O >> QJ TD CL Q. "O c iaj O O C X O 0 XI C c U en Sc 0 O Cn X P C P to 4- St. O 0 «3 X 0) to u XI 0 ai CL 0 a> 10 P TD P QJ QJ X u p P O) 3 (0 p to to Sto í. X) QJ a> Sai P (T3 ro 0 .

X •!c o 3 S- Q O •«-> (C E

tO > , •r- OJ i i - P > :

P

OJ ,0 -r-

3 -r- O

3 3 Sai CT E aj S- LU 3 P C_> (TJ (TJ

P

•, .— «3 C (TJ

Lf) ai

a» "O -IX O

u i— P CMP o OJ 4— LO

ai P T3 t/l

m m o o CM CM

. IO CO Oí C a> •>- «— I i - r— CO +•> O.»— •«•CO E O S• « c o id • o -i- o c a» 00 CO fd QJ r— c o o -c > a» o en cm-p i— t - c U s: i-r- 00 O C T3 O Q. «4O id i- co » o ~o •«— o co oí . Ln 4-> a« 3 e a) c a. o I fd -p .— a» co i/> c co i- 3 a> • • -o o -1— -— c c o r o • +J c (O >>

,

> CO -P id -P O • -í- •r3 r— -P "O u c a» co 3 >> O o o a. Ll_ E o c fd J3TJ< ftj ai +-> S- u fd C id -c i— +J ai " •«- TD +J 3 00 _Q.c IO .C -r- 3 >, LO O cu •1- 1— .C « 4- c C E +-> CO -p a» s- -t- cm c u CL co O) X o 3 1-1- fd +J «d O 4co cni+CO S_ -i- -M • -r- V- "O C 3 Oì U 3 c 3 fd CD a 3 £1'— o CO 4- o > O -rC LO -P o oo «d • i—

O «ïfr r

i— c a i w O N N I— 3 O-

ai • • O) -m > - 3T3 — 3 3O O L •i— UÌ -o • s c a) co co L 3 .. o a> COCTICO «ä- 1—3 o LO CO co STD 4c o CM 0) O o co •r- o c OJ CL O 1 1 1 1 1 ai o. fd to +-> co siCL-M a> a) • N S- 3 co S- o -p "O to •M 4_ i - ai eu co ai E -P aj o o a) co co ro-c s- • a) 4id .c IT C cc •r- A o 1— -C i—

>,

rt

•r CT) O •• sfl Q. l/l f •sa>ij )OfO'r -—ESO)«. a o • lo o o o ro

ai -o c 3

1 i i i 1 a> a. Ol 4c 0) ro o e: to C «d-r-. io CNJ >,C> o a» o «Sj-o o o E Û.O < ü U- 2 2!r-o _Q ro E a o o Q.LD co a> o < Oí < s- o

• "•r- sv+J 3 -—- o to to ro • a) v> ro o CO CT) O) CM 4- en i- IO r- en 3 a a; «I S V) . CM E CO "-O i- o c < r— 1 • 0J So S-5 O. 00 O I 4CO p oo «d3 3 ro • .c , ai c o s: • «oro c ro er» O E to IO i. «O a> --E CO_Q X >, o CL. «a-o a) SÍ- 4-> S3 a. rO'i> o. +J (/) .+J RO o a> co ai c c i— c i- • +J O) lO ro «3" Q) S ro E o en i +J lû-r COO a) • •>(— +-> . C • co'—» • OJ ro ro -M E -p CO CM C CL 4->O IO O '— CT) 00 -R> CM (0 D -P jz. o S-^ co E —- •M KO 4U C QJ O ro C 0) O 2 O CT»Ô« • ro a> +J C P w w c. s: Ï 41 -R «J +JO 4E +-» C C CL o Q. OI c ai u E T- +•> ro c O S«/) X r- —» r0 roo o ro o «1 •, en oí i- i—• O o. i- t— a> +J i -i— '—» rro- Q. S_ -n^ S- 0) < O 1— co -— 4->O TJ

r-- s- . E f— O O • •> o CNJ O OJ C "C O S- CNJ RA c i—4— < CNJ C O O «a- • ro o • Oí O • (/» CMS S- I- C i— i- o o S- ro 0J ••CD ID 4- CM (/) O 3 cn-r- + Ol— •Z • O c>—- E i— C >-M CNJ A. O OLD ' C Q. Orro •A +-> + CCL t/> s: >,11 -P C O Í— (O u (Di O ra (O CM • C en E en o ai r- S -rE 4- O +J O O CM ro OJ •O N a IO S• o 0) O O i— -M í . -Q O -, 3 LO «o ai Q_ +•> o P-V+J E o roto en . •>> a) c «— ai S- TD OJ •i- < u_ u • C 4•r- CT) CTi+J r— £ f0 CTl LO a» 3 (O -P • «C to to CFLJT S. >, r0 O Lu Z ra X ca o —CD O

X un a. a. cooo a; o • • aj LOtO TD il 0 io M O Ll_ -O 1 CJ1 CM tO OHS.4J CT» (O O C o o a> u_ —" • *> E —' «3- CT) X CO ro ai a. c O LO Ll_ > II to av-> o •o O CM-P t-*~ 0) s to a) Lu 0)Ll_ LO .«o 3 tOLO • -a •«d- O o co • c COLO II en«^o; s: o c II c co s: • ro > oo ro cm en o Ü r -i- o•p li_ cl en oo s:p O o o c • c >»~o ai Oí rO o^ X CO-r- tO sen E • ai S o.— CL S- i— •p uir ~a • u il0J , +-> JD r— ro O O Q

Hg-236

C CM ío en • * i co -o rö ur> to « L O C r- "ï UD

co CD •p u m - c _ (U cn-1I— O c/7 M3 C -M

Q- er-»- r— )ll I— l/l f— o o> O i «a- CO «O M « c 0)+ 1 Il •> Q . •o 1— 3 X 3 3 WÌ o c SCL o -p CM tO CM »cm O II •M i- M r0 >> CD o s- a> ai o -Q c •i- CM • C 0) C o hO Il co ai -m -o C o E 4- a» O "O o X o to o. CL +•> ai -— itJ a> z «o a» •!Q. a> S- ro -O a> .. to .¿z e ro CL c c C UE IO r— CL 3 O 3 E •r- c O > cm S« o ro to to rO (J (O _Q a) o cm +J CM U O ro •M (J U iQ. eneo • 3 i •i— • o CM E M aj C o c >> -C CD • 0) -r- SZ SM ro O 1 C • -P j c o o i aj -i- c CM - E 3 J t< o "O co ai 4- C O M • •* fO CM-M *P- COJD 10) i- 3 to OJ .C S_ tO C "O LO 0) O • - -C T- E « 2 CO •> 4-> en «o fe^ «d•4 -I— S«sf i- +•> I - 3 «TJ CO I O 3 CO U E O «3- 4cm co o I co - o ai r— • y> f-» r— cvjx: < co ai • r— en to II 3 i— -M : < c cm 3 -M • C (D I O. -a «o c a; s- r-. i c o a» «sí- "o u « i- -r- 3 en • c j aj +j -i—> c to ro - co 3 3 a) • , 0 C C Û N J O "O -M i— a r t J O l M c c to ai to CM co o r0 •!- > 0) CO u o r— io lo 3 a ro to aj o «t i— o u a) -M S- o . i— > > ai «-mii e - a> s- a) î- co 3 o i s- c_> so î•> cm co 4— 3 -C 4_ 0) i- J- r— -a - -M O -C (O O • r— h— ) • > cm a> 2 to s-s «3- -i— c i a) en oí 4-> i >>•«J -o i- c .— 3 co cm a) ai j ai o -r o - c o 3 c o • +-> en j o o •— r-» -i- c: > N a) -C o. - M 3 m r w w o to Sj - ai +-> lo a» . î. s- ai i s- ai to ) ro c a i - M t o tt+J-P i. T- +J .1- Oï T- to to +Jí— • to I— > , - i— l. (O O O l r - o Sj ai a).— m r— 3i N u : c c 5- m - -r- en ai •r—• D

a

"O oo o o •i— ai ci s_ -p x +->

ud r +-> id •i- -p S- m • U >> +-> s" (/> o 3 CO O O LT)

0) -c -M c • P i - w ai

CXL

< ÜJ CO LU Q X O

I— S- -o ai c p a m i/i CM i— CM CM E a; fd o fd i— fd co u o fd íCL aj cu -p -P -o 0J en oo E oo E -o fd >> c O 0) o a; CO CO CO O) oo a> O) X c i . C -P c •o O O O c +J T- 00 -P CM CM CM O O O O3 O fd fd _Q E o. c > o to •

vi un 3 CSJ •r- H 1— O 0-51— c o ai i. • •«>>LO 4-> Q. LO u 4-> 0 -— co O -c a> -0 •M COû£ s- 0 SO CM a> o. 00 < 3 1— O • • •O) C-—-..— c •t— Ol CM E cu 1 0. 3 í • «3 1— O 0 nj CM+J + c—• 0) 4-> O -P

a> H +-> JCO -P 0) 0 ai SCM c 0 s- en0 to 0 ro Q. ai E TD O 0) C £ -M 00 ai e to aj en*« +•> E 3 i— fO O 1— Oí10 • X 03 a> «st

ra 5-9 • - "O 2 co í-

o 3 5-5 en +•> s lo 5

•1LO sco -o 0 •c 0 " - « o io Q_ r— -—. 1 1 O) O CTi -Û fO +J lf) r- ^ O 3 -r- - • O E LO O O • O 1 LO 0 S- Oí •1- • - -C r— 0 II z a «e r--. 1 0 s cu PU-< ">-a 3 co 1 CM—.í— oí en 0 «3- 0) ^J- C S -r- • •> 0 1—"O co O G •• 1 >•—1 '— i -i- 01— i— fd •1- +-> J_ re fO+J Z fO >, C -P c — o. ra o OJ O) to +j E

•I-

to c

o -M nícn

• LO -, • to S- Li_ to Oí >,0 CMC • CO SLO o o c «3 " C -ÇJ- o 3 E r- LU 00 CO lfi •r- < «Ñ»--«eoa) ito CO JCL * o o O íd nj 3 a. csja •• o-— E s- E 0 0 0 S QJ U s-s i0-O O) '—» +J O)x: o •>- — tO + -> LO • S j= CL-r- - •—• O E co i— •"OC CMT- O CM • •i- 03 OVO S- I LO _ . a« Q-x: «J o V) O r lÛr- • > OJ +J i +J LU d> O i— I O Oí CM OJ J- +-> —'E •) < CO O C LO I--» O co i o ai S«LO — . r, !/) • -i- CM C -C CTi O LO I—» i— O) -—» on O) r O (U CO-—1— C CO 3 -r> dìo 1- * C CM O II — C rü I— O) o s o : < S h

o. o sc E a> MO OH •1-0 c 4-> O 10 (UU3-P 0) C . TS. W CLi— C

< Oí CT CO

Hg-239

• o m o ai

co CO CO W e CMO 0 0 T- :2 O CMO CM CMO O J• ^ +-> OJ ai x i- +-> u E c X a> CJ > fO OJ CO >> o a» c CL— +-> CO O E Li- +J (O IO 3 c a; -r- ei o «3 n3 CO a. ta -C +-> O S- s- co -Q IO +-> IO •«- it> ai ai 0) -C S- 0) -C ai •r- i. co "O s» ai c cu o -t->T3 CL cn-c s- o ai ro ^ fO • h > CJ E ai ai i- .A — u. ai > w dj o o — fO -P C O 4• o «I- T- ai + • fO CL r S- O") IO IO IO IO S r io o 4O ' r J J to ov— a + > .o c ld o e «3 e E O •!- CO •— ! E cn+-> i— Qi io io "o i— a> • i - s : o o G c o a > a > •i- •> 3 t- •(UI—e 01 o> + C>J ITJ i— 4- + «— c i—• w a x ' r i - o r o Q-'«O rt>ai a> E >- E : I— OH- £ fO -r+ O i) -r C IO IO 4— CO *r- «r- 5 • 3 +-> S -C l- X 3 +•> S: +-> o : +J U 4J V) flJ CO O l •1- ' .— C CO LU 4- L X (DJ j ai r— e i— \ CO 0J •— ( 01 : |— a. s o s. io c a w ol i E r c o ai t w io • o r a> x o +-> co X X «U Q.LD *0 IO SI 10 c fO O O O £ £ u CU s JZ £ «3 .E «— fO CO +J • +J T~£ 1 -M Oì +J o JD n) a> JD ai X IO £ CL o X «3+ + CO CTI ai c IO t- o IO 0J •I-CO co O uo £ ai ai fO fd £ -o ai ai CNJT3 .— CT> e .a ai lu li. i— £ 1 X > co CL E o «Ì «3 O ÌL- £ a> co s- + + o s- a» 4- +-> ai JT IO +j r— r~- p r o O 4- ai «d _a o IO CO «a £ < < s- iE E -Q 5 «j «a ai a> -a " " + + «— u fO S£ £ .a >> o •>>> -M QrS_ 1— £ "O E c E a. CL 3 '— C_J (_>i— ai e fd co IO < 3 (0 — «0 X -i£

+-» CO S•I- oi a> "o i. li E c •r 3-D a> fO O O Q_ •a "O ^ > "O "O e CO O c cn c cn «3 O fO +J LO CO >, O u o IO CL i— "O o a) -o -OIO e IO o u ^ £ a> o fd il «3 +J O o s+J Dì a. +-> > a> i—IO «0 c 2 io cn-i- IO S- -r- i—13 cn co h l O C O D ra ^r > cn rO -r-

-P r- >» Sc ro a> 4a> o O >—

aj TT E • ro «3 "a QJ CO (_> "O OIO •«+J O > to o co C

3 CSJ -p ai s O) LL. co •o .O 1 CO ai "O >»o «a- s_ S- -ro t0 1— u > a» TD a> Lie j_ to rd -t-> E e o CO 4-> •I- +-> o w> LO Q- >, r-. to e co O) a) aj 5- d) 3 .£ •M 4- O +J aj o s_ E • o >1 3 "O 4E Q) .£ to ra > X co E ra•3r- Q>t. to +J E ai .E e ra c +-> ai •r- O +j . £ +-> t/) e +-> t/> o o JT3 ai o S -M JO) O O 3 s: i— C O E o 4E

oo LU 1— _J o: LÜ co s: z co I 1— _l o z UJ X CD z


a; a:

to II CO o o •• en i o to 4-> s: CM 3 . +J r- O C c r-. a> •i- • Il 4-> LT) C IO i— O O 5 1 V- U o r^ z i— CM o to er* a> 4- CO CD II to CM o ra co i O o >> CM CM -O S- CL) • C rO Lu O fO > • »>11 O 1— CM to co co E OJ • O Oí CO M I •I- CM SE 1 O «+0) O o E O •« .— . co E •r in rs d . e e o •O • tO i—a> •r-CT)CM CD -P «¡3- • to •i- CT» •!-c to 11 i— O) O 1 (U(U Q- CM CM en3 E O r-* c •M O -r- • raa> o f— oo oc

t/) 3 (V O 10 +-> £ > O a> CM Û£ o +J £ te (V to co X o t/) Q. CD u E +-> +J o u C 0) to E OJ u 0) o sE Ü T. +J on +-> JD o E £ .£ i-a fO E 3 t/) t/î (U 4-> co 4-> •a £ -M to c O) C E c en 4-> to 43 "O O o £ 4- (0 a> o ai E +•> sO r— C_> 4- >«atO i. LO CT. fO • 1— > O to CO E £ O O O • to -i- o 3 +-> O -r- II •o to •o O co •r- CLO 3 E CM O i— TJ U < £ (O • en oí +-> £ U 3 • O Tí O _l £ 1 «O i— 4— -C CM O O II cr c •i- -I- co -M to o •1- CM S- E iC_J -r— (_)

A3 h— a> i—O JD JZ 0)o CO sZ o s^ •O o o -M fO O E U co i. «d- fO 1 O -o Oi •» O SCM t/) to z 4-î a il II E • en 0J "O S- C_) 4- +-> ai0) I £ -M O t/) s- u a) co en 3 O CT) en 3 4-^ E O s: to u t/> +-> a;.E CO +•> -r-E rO QE 3 + Q. a» 0J LO sE to 0) Z LO o -M u rtJ + £ «d- ai >

to u u 0) O CO II E u , co CT) o •f o s- ( _ ) CL E o + E •o E o o . 1— o to OJ ai en TD en t/)co E . 4-E f0 £ o O f— o < Í. 1 3 LO 4->to £ O 10 • O to c O -r- i+J £ 3 > _a -r-+J O +-> •M Z 4-> 0) E \ CA 3 +-> ^ fO a> L- J3 CT) E M 3 >> í- £ a> to ai

tv

£ E Sra ra o CT) • CT) tO 4-> £ ai 3 a> a; 3 CM 3 "a co O • 0) E OJ JÛ O •r- I) £ 40J ra co +-> E O E ra cm O S- i— a +-> «a: o 3 CT) CD £ -o co a» • E +J co ra rE to O II ai _Q i- O raora sen

>

E

• +J T3 E O i— í. co r>. ^ aj • u o i 3 "a CD C o LT) 1 E . CT» o s I I £ a> Ì. co ra o o +-> CM £ i- o o o o

O; O OVI— - P

E

SO) "O • C I 3

E

E

+

I—

O O O O « 3 - U D O A I O ~ O • SO 4 -

CU E

>

X O CL O "O C FO E X CL O T3 A> +-> + J RO CU E

T3 O +J >> . E •P

E

I—

5 FO M U O M + > + J I— - O N O CM — R— -F— O •I— I — H < O

FO O C\J SU -

FO O CM O C AI

E

CM - M UO 2 E M

>>

S0)

R««. R-. " O >> E RO ' Û J

CO N CO CM O « S O

+-> E

E AI E -O E « RA

CO I N • P CO CO 1

O -P

>

+-> U AI CL CO AI S-

O IA O CN-P O) E -P E CU AI • P I— • R - «A E > A> - R E 3 I— CR « AJ

O TD CD IT) C Í- -ÍO E «— O O >, 0 3

H G - 2 4 2

-O 0) •P l/) QJ CT> cn 3 -a VI c

re

0 c •r- O .E •«3 -P 1 s"D -Q +J .— 3

CD F— >> C •1- JH s3 ÛJ - C A> s_ a. s- 3 to 1— .c l/> «3 - P O -P 3 s_ c 0 O0) sE cn-c •1- s_ cn S- QJ 0) -P -c CL C X •>- «3 ÛJ X 4"O 0. O c 0 (0 1 c E 0 C r— O ' +•> •p a) O..C 0 •r- -P Q. E O +•> O (/) re u ai -c aj a -p "O

to "D 3 S_ •P 0 to 0 1 to aj 1/1 - P ra c a> fd E s3 +J X c QJ fd 4-> 4•a O c U -P •1- (O S- 0 •p S- i— Q- E -C J2l -p , to (O .e 1/1 ( j en oj a> 0 -a 0 c 3 O X i- 3 CÛ sz JZ Q. 4— a>re•!Í. -a -P cu >, X •!O O -P r— to >—» 1— to i- Q; E c a» 1— re 0 QJ Q- +•> C 3 a> O cnr-. c j en cri

•p 3 0 (1) O -p CO 3 X QJ a>

rere

re

(/) >>

JD XI "O 0) QJ .Q s- O 0 4— I/) aj TO to sz 0 -M .c -P 3 TJ

„

E 0 s4-

re

•

O a) «D i_ -p c: cn t/) cn 0 - P cn Jc =3 >> 1/) 0.

re rere re

Hg-243

JZ

3 0 QJ O S•P O m c c E en "Z. E 0 •P 0 E J= 4— en 3 -p O) 0 0 s_ 4- to cn E 0 S- +-> Sa» -p 0 4- •P aj 4-p c to to 0 O QJ E cn c E i- C7> cn 0 3 to 0) - P E QJ 0) CJ -P to - P to c E cn 0 a. ia. -P 0 QJ +J E to cn -P 0 X 0 -P 3 0 CL ai Q) CJ X : CL E -a 3 0. C E O) I0 t- . C jr 0 -O -p 0 -p Q. 3 c 0 3 JD J2 CO 3 - P 0 S- cr-a -P O CSJ+J CL) QJ E "O QJ < C JÍ. T3 O -M 4LU LU (/) 3 LU c cc QJ . Q UJ QJ QJ •P csL 0) > •0 Xi c c U -C ai Q) E S- aj > 0 en - P >Ì CM .c -p 0 >> -P IO c T3 QJ aj C 1— E QJ 3 ^ E a> tO O SI. Dì a> m (J u c c cn > c SS< U 3 to E s~ to c X QJ - P QJ E 0 to C s- -a QJ E S- >> C S^ •1- Q . Q. Q. s:

re re

re

re

re rerere rere re

re

re rere

re

re

rere

re re re re re re re

re

re

o

+->

•a O) iaj -o co e o

a> a) s- -a a» •r- E en ss- o 3 4JD N C U S_ > ta ro rn -c

+ X e o en cl. a> ra o - u p fO a> aj .Q + — - p "O E e CD a> E Q. o r - "O XI X e COo e o o e CO CO fT3 E 0J 4O < - -Q (O CO O U1 u « f i- c i Qj vi a>ro•— •>o 3 o O) co îO en o CM X) CÛ NI CSJ "Oro 4S- «—» c O u o ro ro JÛ .e E E •P Q. ai C •p -p 0J i/) -c co 0J u ro io E 0) N 4fO + — + s_ (_> 1—E i- X • ro o dJ Q. X o -O O Q-4- Q. •p E + O eu V) t/l -p 00 s_ 3 >, ra 1— • c o s- co o > o « r + o. io u 3 X 3 aj > O 4CD c E i— + o E 10 ••- •p O a i r— -P O CL •p 3 • r - 4 - C/Ì aj l/l -a •p -P c r- i- E X — 3 0) C s- - r ro 0J E •P c co e ro ro - r N E 0J - O C o -p c ro S Cro1— o * 4— ro * co -P U 0J

y» E O •P IO O CL E O O

ro co 4- ro < -E 3-0 CO ai e • ro z : ro co _ j

0J E cito

• o i— • 0) 0 E •p cu ja> 4- 0 cu 0 Cd _i

ro

1— .c a» o C -r•1- -C o. 3 co E •D S- E O f0 00 O -Û -M 4-> CL i -aCO •i- CUO i0 h- +->Û- to o OJ fO 4- •!-X» JD c 4) J: row o +-> aj E tO «r- 01-P O E CO to+j •I- S- O CM r Ufli r - i. ai cu 00 s-E CO < O O E on s- 4i— X» s_ O ro "O O 1 OTO -i— ra e o a> i- +•> co ra -a o i- o '—» i. dl (í "O Sai i. o c«5Î- a) -a -o i— ai 3 •i- ra «i- i- f «O E XJ (04- t1.4- < i«O + O . Ero 4) ra cu X) -i- Q. O•r— +J +J "D E jr x: to >— S s- O. (O X E ai co CL o s~ ra • r o 3 re 004- O O) í + f0 c o i. a» ro r o '43

a> ••• +-> "D CO»i— i— CO co • p o c a ) c 4» CSJ O) •!a> X» e -p ai c cu o >>r- E Hc .c ro > -I- < o o i to C 0« +J 1 íD u a ai O ai -o •1- -M -r> xi +J o c D) 3 • rV» ro « O S_ co 4-> o •rO) CO E

£ i-T3 X) c +-> u N 0) a_O i— S-•o 4-> ai S- +J N •>- ai e t J_rtj. 4- CO rO X S- +J t >> I >i— p— i- o i r— - di j— " _e m « S.+J' rO E CO+-> " "D a» -i- o •>.E «O OJIO f o o X c: 3 £ C Er+-> O) O 4J afl c D) O «— £ c LL E a4 t O) C_J i- eo 3O X> O US.O oT c o A3 u : CU +-> L S.-P 3TJ O S- E E D)i— (0 sO 4- E D)ro co a» m o 'f— •r- .E E co 4_> ro-—.+->O ai ro o in E 0-+J csji— aito 3 E (Or- «tO E -o O -C < «a- Sì o O +-> —'i— rOí . E

>,

,,

fO ro .o .E

O C .E co ai • E

O I 1S- -i- 3 4-> W Z (/) QJ SI E a j í ? (D 01 •P-PT3 (O E 3 -r- E

E E c +•> +J +J s» C C fl l 0) O N >} Q. (J -r- Q. +

4-> • o

4-> ift E QJ >> O E S- *.ÜIO+J

E -M Q' ° E 4- O OJ O o -M -M O •i- O 0) cu D E U"0 W E — 3 O f0 -Q +J S~ -rO c 4- 4-» i_ aj CU a. 00 fO 4-> CL O) > = >» (/) _Q 0) +-> 00 0)

S- r— tü -rU 1IT3 S. ) i- O O i- 3 • E -O ^ E -rO - >>•«3 r"CT E QJ -i- a» O CL) Q_ Ll_ a)fl B +J 3 i . O i- r - • M+J 4J s. u (5 m o

4-> -i- I II QJ S_ CL re CO +-> C QJ • 4-> • . P N 3CM ) CL' ~a -a : o H r-» c Qj > CT) • (C (/) O O CsJ O ) r - i/i : >— • S- 3 ) 0) UD C_> O O .E J C «i«sí- o +-> 4-> : -r • CL II co • o O +J ~ a>rea> s- e ai IO UD OJ LO 4- ra U • ) co ..— o c ra -a E -i- r— -reaj ) E O r CM lA s_ o. O "D O "CTre•i• • a a» > i- +•> a. co e s- n £•«-+-> S OJ fOr—fO -i5 -CT O "•a i - 3 (o .c i 1 a> o m h e oj er aj ut +J n o u .E aj 4-» : u CM l l c ai in c t j i — o J -r- CO * O 4-> D i— -f- C O - i- • CJ rtJ - C CO OJ is_ i .e re e co o a> -e • 3 O c/i i—J i— QJ tO re o s- o CL QJ II ai N T J V Q J «rC +-> - o re - 3o -Q s- s_ a> T- •I+-> O -O Q. CT) i- "CT oí s : .c CD c a) M- 3 CO 3 to O. +-> -C O 3 to O O QJ

O) 1— 1— C\J OJ"CT 0) H to +-> .c a» í . r e o E i3 4-J a) i— fO V) 4-> «3 u m ai -O 1 3 re 3 co X 4ra "CT • +-> o E +J o QJ en O T- 3 00 E O) o O o X co o E E to 4-> "CT E • O OJ o E -E QJ r— O S_ CT> +J< 0> ai (0 to n o to i— 0) +J +-> E s- ai o O o -—• E S- O") +-) .—K II O i— +•> +-> "O hrO _ l to 4J 3 3 «3 E co o 3 E re T E QJ E JD > •«N CO O O) E S- i— E KO N S- O Oí E CT) 0) í . O) 4- O r-. 4-> í. M . fO E to •CT 0) C S_ 4-» -E E O-CM E -E r— •i- n} -Q •M S- _E E QJ — m i - < U 3 " 0 i - E 0) Lf) OJ 4-> CT>«5ÍE aj E • o re o 4- í . 1 co E O CT) CO O QJ to E • +-> •M•1- O •¡3 E 3 3 en OJ O II O O 4-> E 1— O to o O O OJ X UJ O 4- Ll_ QJ

re re re

re

re

0 ( 0 • • j i— o o jd : • o o + o ) aj 4-> j il i— •i- Û.-0 • oo+j o ai >o 3 4$ CMO¿ + S. ia>

re re

E +->

OJ ai "o ro to o s- -o "O XI E G fO «3 fO a) M M 4-> O •i- -r- CM "O X) i— •r- < c .cr

~o

LLJ f -

a> E -û 3 O C O S- -O -r- fO i— E 4•ru o O c on s- C O o C ai X -C => "O a) o r - C0 O • s- ai t/> en -—fO . (I) c a+j X o c UJ • — ' r - 4- • — i c:

•i- +•» s"O + J= 2 ai a; f0 c s- 3 -c (0 sfO o -M -C XI • r— C_J rtJ Û_ JD E i— y> •— i o ^ CL +•> C -M (/I a» -rC 4- X 1— s- O o CD c (T3 C7> Ifl c ai E o X «O E eu • 3 0J 0) X c S- en 3 o c +-> CL c ro a> OO *r-S- XI

co co

ro O CM
c •r- •«> E en CL to eu ai II ai -r- ni Q. (/) "O ^ E JZ > > >»0 en +-> O ai 'ai a> aj s :

s_ to •r- O X •p o Q. -r- CL 4fO 3 fO CJ CO-C oo -o o 4-> O o

S- O O i - r fO tO > "O c • • a) o M ' ) +-> -r- I « i u -PCO • L-1- CL CD : S- I l • ) +•> O CM LD +j =3 ) X cr > «O O >> (0 E E • CL E •«- C\J O •p E . E E Il + u 10 r •r- -a +-> • -P SI E -i- X E 2 3 2 ro OJ xi e a. • 00 ra ai — ai r— t. + -P (0 E 3 CO •!-P -r- Q - O O 3 (/) «3 CVJÛ. to >,-P >>ri- c i- « t to a» u 0 a» a) -P 0 > - -0 -Ia> »>1- cn •P to >,^5 4- O •r— "O f— +-> (— r— -P C r— 2 3 O a» O IO tO ai C E LOI CT Q.

II C

X 0 JÛ 1 +J e r- C r - ro

J*:

c

0

• +->

3

0 "O ai CT) go r s0 0 0 • •M +•> rO E E •M O O ••> S- 0J CO 4- Q - O (/> -r>,0)1— S- eu u_ "O ai JZ ^ •P u Sra 2 i+J E >> 0) S- ^ E 0) s- 0 4- ro - +-> c S- IO 0 ITJ Q. SZ +-> U " f0 to to 0 0 aj sa . +J o . E •r- QO 3 r0 • O to to -—« O) . e a> s« +-> +-> +j -p •r- -f- 2 C 1— C a> 0 ai co E c E r-» 0) r— . < X t-i 0

Hg-:•248

(O SI -p 2 •P 2 «51O II

•P E ,

f0 X E rO < E

CU

f

C 1—

•I-

-P rO Il 3 E

-P

4-» csir-» *iai 2 3 Ll_ 00 o «r- E s : "a o w c ci E «o O E IS) (SJ o Q J i í OI U

i- co ai * c/) • r ai ai 4— i— 2 i— -P O •«Il (A +J 3 CO E 3 tO E CO O Oí O O -rOí • » i- "O -M

i/> O oí 2 rO -P "O + M r - 4) CT) n3 Ol £C O) S- «O) oo a . o

O

il to

CLl— < •

+-> C 0) •I- o S- E !» (O O C í— IOT3S D S- C I -O 3

fO O r— 1- -r- O CT) +J i/| 3 X r— dl J -r- X +-> (Ut ) 3 5 > o a j «=1 c o ) X «i- csj > CU CLr— 1/1 e t > 00 —O) l/) O Dl

•

o ai (/) c X

o A i— "

H g - :• 2 4 9

io o r +J S-

«/) C : o n jz. ai )

} +->

>

o i 0) O "O • • U -I+-> .Q X • r £1 O U) (O o en-a

OI i -M E c

• a> O s+j eu S> CT) 0) 3 4n to O 3 O S- to to a; 4-> 3 X fO E O LU 3 C a* aj E +J ÍT3 a. C a> to o +j u c C (/> eu O 3 +J £ -o •u ra x> •M -O O c S- «a c X .Q a) o .a to E ro 3 Ü5 O o 3 c -C C a) +-> E •M •P C TD 3 O ai C O u c Í0

•r"O tO C O

•P I/Í o -E 4-> S_ O e rC E

o -E •p jo e «3 C

(Ti ro

O co

a»

-p

-p •p «3 Q.

•

1 o c fO -p . •1- o •P v.o CJ -Q •t- (0 -p en c e a> -rX o a» S_ -P >>•«o. P t0 4- D. O (O LO "O ai a> to P e 10 e to

¡ (O (0 "O s) O- •«- Q. • io w X - I O) o o a» O- to ai ín tO >, i— i e a> - • . (O » o -P cu U«í P 4) O -I- i•r (l) C « üí to ai i f0 O) E -P r— .C i— 3 O. I— .Q

>> cu >•i -c H Co s- s_ > c x: Q-CT Q. 0) E (O t- N to -p •P CTto S- (U 1 a» o. •p -P to o «O •r- 3 sCL E O o cu to O) S- > 1 >í o 0) ai tO -Eto -a •p Q- OJ E tO o cu •p 0) O E > cu a) -ca> (Ö c S CL 3 ^ en -p CT Í0 CU OJCU to E H + J (/) ai su •a -p u o CU co i E 4-i QJ

i- i < N c E i- 4-) fO a» _Q _c to +J iQ_ o o Sa. +J QJ •a •o +-> a> c •a E +•> O toi 3 to sSil .c -o o CJ C l

CM to — 3 < O

•a 0J 10 .O c O. o

c oo O) o +-> Ll_ CSJ to E 3 0) s- o 00 UÜ.UZ to V a» •M l/> 4-> i X a* 0) .c 1— CJ)

o 4-> (T3

OC

to o CL

Í.

n3 o Oí o

o -t-> fd

cu

o

+-> to t.

Ol CM to en

-o c ra e

to ra o

i— c o o o

o

i— cr>|r-*.

r v e n c\i o

>—to

: f— o o un t o cm o k : O

CO m PO o m r-

X LO "-Û

oo » QJ r— CO to 1—I CT» >>

0~> LT) [-«-.¡CTi

ai i>

CO > O . c to o CL "CT c s_ o cn p o i n CJ r-. (V CTì - p JH 0) JD •p

ro

Í0 . o CT) C _ | 3 S_ o O 3 _a CT) • qj a> 4— N > i - -r- >> to o > o A3 o .Q E to c rC >J 3 E to c O 4- O O c 4O o (0 .«. 3 •p c to +-> c ra u i— u

>. ro

CL) o

Q.

c i0) "O o o r-*. E CT) c o o - P "O aj s_ to o i/l ai ai o t - a j ro >> > T 3 CL (O - O O I— «3 ro «r- a» O QJ I— -c -a O I — "O •«c e s z >> ro s_ «— " - I - CL •P - Q • p t— > » to S- A3 - P O . S . Q J f C0 O) - I - -rO 0) o o -û to - o O . T 3 - a i . i— - QJ I— (O 3 o OJ O . - P CL V) 0) 1 S- 3 C - C O S O O 'i

ro

CM cn I—CT)CO CO CM

>n i f i N O l O C O ! O C v j n < î O r - [

C CM CO

t O .O 3 ( D J 2 J 2 S• o «3 N QJ CT) J«3 - P o ss ro q j 3 "O - p «4C ' i - t o o fO C 4J c i - fO • - r - o r0 m a» to •!a» i - io . c to a. re S O) +J >> >> E ai QJ c o OJ +J "CT o ^ u. ai to e n - r - O CM •• C -r- 3 E - O X s• • - E O O O o o O . QJ s_ i - S1— ^ . c QJ < u -t-> u Q .

co co CM O o O) O CM CvJO O O I— "O .— s_ üj c cn ra •

to 3

ro

N tO c c • o o s- o -Q

ro o

ce.

s:

ra +J «3 "CT C ta

1 >> I O r— Í• i— O « O ro

+-> to 3 S

rö -P s: O ien ro s: o.

-P -P O s "CT •r— CM S• QJ C0 CLr— 1 "CT CO C • ro t o >— O

c

QJ

>

a> ^ en .E > •P " O (0 o J2> -a QJ Q> So o 4 - a> QJ • p

>

ro 4 o o C •p o -o p OJ ta S- N Q> "O to C o u

ro p to >ì u QJ u £f0 a) ¿i to - p QJ i- e o

E So 4-

ro

QJ QJ - P -—> - Q

a> •P

QJ CT) ro •P co

+J «3 OJ î- -p - P «3 CO

Q> -P •r- 1 E r4 - 1 o en • O SOJ x : c o to - P o •1- c n to • •p C CD •i- CT) QJ CO •P c Q. • ai r-» • E - P "CT «dO t/) E «fli- •(- •i- CO O CM • ^ X u a> CO • CO O •r— r—^ m QJ o i_ "O to n un ro x » c c io ro co«^t o o >— o X t o CM O Q . • S- Il s- o c o o rco •P c i •• o to QJ t O tO CM O QJ c o 3 Qj . o Li-C S • p c o i— • * Q; oo «fl^ •«- Ll- ro CM o ^ o to en > , o C C £ i- M io V— ro • ro •!- Q > t o 1— .C -P to o ro C II >> a> >> o r— i- u •«- o c - P a» 0) - i - II.

• «5ÍO c u n ' — ' to 3 ro «id1 COSc o -p ai • cm c QJ CM Q) 3 «d- U +-> -—' >> 0) X> CO-—s O "CT CM-P ai t - s - p io O 00 •r- s . Q. S o m " C - C CM C CO 1

íai O - P - P 5-S 4 •1- C - P E Q) J - P ta O E ai a ) . C i— . -p O OJ c o IO

ro ro

ro

U CM •p r-» ai cn

E TD O E ••> to •i-ro-—-p •P a; fa CTI ai JZ i- t o «5Sin > id tn. i— •i-p XJ

co o i— -P O LO -P «d* CD Olf" •

un E i— o sE 4O ' E ai s3 N 4- co -o TJ •1to CM 4O I— O S- CT>«£ fd S to a> -o O- E XJ E o •r- E ra •p r— ra

aj u

>»

o

0) > 0) ra -—co -p •P — - r - cm ra i— t o • - • «5Î- 3 o •p O 1— E • • ID 3 3 O O ai O O -p E SZ -P O tO i— 4cm r^- -p E to • «3 O s- • o E U a) CO 1— CT» >ì t o ra O) m E E S- "O to o -i— 4E O 0) o e a» •p •i- •«- t/t 0) •P ra > E u ai ai O ai to t- •P S- l/) (0 u x: 0) E ai u r— J- -I- CT) fd u CT) "O O O) E 3 0J •i— "O ra 00 XJ -P X» ai a) cm ro co fÜ

CD a> (O a> rd • ra +J •i— a> ro-i— w - p u o • E £ O " 0 c (O H f l > O) OW+JCM(llO£S.^C3 •p i-i; i— E c l m •r+J - < S. C0"0 D) ÛJ Ü ' f W O) >>+•> C CD CD S- -P E - P 3 + J r - Ifl OJ 1.+J D - r a i o c o c c J D f l J d i o -p > s. a j i — r - + j 01.a na ai N o -i- r +J fo en cl S_ "O to , E 3 ( D W X fO • to O E i— •«— • -> 0) U ra Oí — «J i . a» X> QI D E Q-r- "O o 3 Q - j z - P O E . . Q. S_ CL-M c i- ra aj i— o í a . «3 >> a> ai u x i ^ eg id io m +J tû s. o . .o i— r-*. a> S- fd XJ X O) dJ en c eo rd S- o . fd o i en to ra - O E (ö s-cn-i— «» ai cm-—' -p _cr r— _cz CO 01 LO Q-i— QJ -O o • r -P S- > , < f O a > - o c -P •+-> o >)i— - e •< a) ra ai c ai -m cm-p ra n n í ) £ o . 3 c -a oí o 3 3 > > x j -i- a» ai a "a í. oj cm -m cu •p S-+J.C ai o r ^ a) E >>-p ra a i r - p i i - - p o J3 o .Q l/> C •!— Dì fO to O ' r £ ' f c w i - í í >, o -a—- E -i- 4-> o +-> í- to s- ai a> +-> s : i- "o r i— 4— "O • +-> ü 41 O 113 O) ra • E .E W (O « í (O C . OJ *r-f-P 0> tO -— S- íC W C (T3 "D DI ra a) +-> +J ai -ri- ta a> ra «— s~ enjz E e -p s$U f o o o o n a t H - o r o o ai to i— cm -p s- o u r m a i D S - d i c + J +-> cu 3 a i i - u o i - a i x w ^ c r . ( o + j tu a j í - a i - P t o X (O r— rd 0) to en a> a) 3 E c 4- 3 • D. > -i- S_ 3 -r+J H- _C E • a> a» st/l 4J "D O . 4— o a> ai ai -p x j E .C 'i- O en en e û .

-P e • CM -P

0> O >1— e to T— (0 E E

Hg-254

. ra rd • rtJ E -P CT> E -Ol 3 rd • * Q_ id OÍ E "o Q--Q 3 ai i— a i - p o s •!- -E E • 4-r- 0 -Q -P tO +-> 01 fd T) i— E CO Z 5- 3 rtJ -r- •!— fd : c ra X ai a c • o s- OJ O CLfl3SJ E E 4~ A3 3 •!> o • (J E "O -P 3 : 0) 0> •a fd o" ) +J tO E e s- eu . U CO ra eu ai CL

i-

o E •1- 0) CO ai to a; -p -P -P 3

«3 -P

• ra C

' -r- O E ai -P 3 J CM S_ \ -r- tO :.— û£ ai en to o o 4rO CO -P CM Q. O ) 4- O O S- 0) : cm (/i c_> -a • to t. -1o • J E O •P CU 3 E ) O 3 : 4-

(T3 a. XI s : ai -p o E O S- ro S-

to "O -P -p to fd -p ai CL 13 o . ra

ai a> o E •i- o •>> «r- -P > ••- -P fU -rrd -C i— O S- -P O

-o r - . N O) C O) S- CU ra •IE O ^ .E O

cnt— í- 3 co-i3 U •»- o +-> -Q -r- a>Lt_ fd N -Q -— T— J_ 3 u fd o x i a> o • to ' io

c qj o. sQJ to

ai c > o aj x : ni -."O 3 -r. CO'Í U roo w w u •— i O CM C tu (/") • 3 CM i - O 4-> oo : s_ o o T - fd - c o +•> no a> o a> - • o. . - P o m fO «— >> ) fO LO I CL +-> •I- I CM c E j o en «a- o >5 ai • o CO c -

"O co

O) (d CMS« O — r - 4-> S- +-> (0 c S c .c 4_> T - O Sfd co f0 o T 3 S Î W + J OJ 4+-> . C -C . 3 0 ^ o - P "O • ro u OI - LO i +J i/i : r - |N M C flJ O ! J_ • CM O >> Q- < fd LT) • -rO tO O 'r•r- a» . « J 3 Z s- +-> u E r lt 3 ' r c > fd < . „ l o td o. e o o E i - O • fd E • aj cm en í . o . c irt t- i— 4-> o . o 1 o i < — •>-o cr> 3 o 4-> • c +J cr> r-. 3 td (d • «— r-^ .o

:

.

"O a.

'—» t - 4-> Sro j z X w X > ) > o ü o > j i fd ) u u u • C o fd -r- Oí • I— -r- O S- Q. 3 : ai fd o c -m o Q.-C x : ) -r- C +•> : o . o i— a) -r- i— ) co o fd 1— 3 «sC

-O a» S O J_ 3 OSI +J 4-J O" r> • •> co fd to cu ai « o i - ai m -o _c + j . + j s- _a c 4-> 3 o c 3 td a) (d to s.^ c o ist o » QJ C 0) -C i - m— r*- a» o > 0) cm 3 o fd -I.c 3 >> c/i i 4-> (O 1) Cl (U +J r - V) • X ) C IO L0 >»co fd aj a> 0) I— CM O 'r— 't— +-> fd c o «r- "O -o •— i c a> 4- o o E fd o t— •>- -Q .o O 3 c 3 +-> T3 en a> a) r aj ai c - i - s. i o c X ) fd co o o

QJ to QJ QJ "O 4-> x: x: co a> ro +J C O 4-» 3 .C E S- M— >> O ( JD o S4-> 1 O i- ^ +-> t- J- C r— Û J < c id o a» fd 4-> to (ZI O QJ 4-> m -i— -rc o_ u «r- -a X I 4— X CO fO eu O 4-> O eu t u o sCM to "o x : c Q. i - 0) L0 •r- t - 4-> ai o TD E O S- a) S- SO fd +-> QJ QJ 4->O •M 4-J CL 3 E fd (J c c U XI ro' 3 fd aj oo 3 ai +-> fd o -o i . x w E 3 (d -M "O E CM o X U O x: O X c +-> o c a 0J O •o (d fd E « "O a> x : QN O E iQJ S•p (O l / l ' r o D--P E V io +J O wU •O 3 0) C •I- « s- a> -Q -c= O +J ro to N 4-> C 3 O O) i- S~ r•— i _Q CM i - — i - fd Q. SI (d s io — sz : n - S- O id 4-> -o 4-> 4 tO C Ol V) < C r + J Id +-> í- C C O - S- CM 3 L0 CL Q. tO i - +J C 4C 1 to > c o ai O -— o (J si "O O j -I4-> C 0) •(- o QJ +-> +J ^ QJ +-> 0) O fd "4- +-> 4-> fd s- C +J O) i/l o fd ai E u QJ O fd S- CL 00 E O O to to E oí E "o O i- LO "O "O Ía» o o r - X V) O O O LO U Û. fd (J (d Q. C L f -

O CO CM 1— J - co l o r»» co w

n n û O ì i ù oo - CM LO LO CM ro

o

LO Oí LO C 0) X o sQ. fd C

,

LO en i— CO «Sí- • LO CM «d- CM LO «a- CM ro

LO > to ta (/> E f0 (O so +J E 10 IT) «3 to äJ ÇZ .E >> e

a;

oc

a» >j

+->

ra o o _J

ra X O QJ O) to +J aj -P a fO -o S- E a; o )

E s_ u tu to o s_ S a) s o o u •o i— O)XJ o > C +-> M— fO ra to 0J co co CTi co to " s- «a co co o 1— -C rö

c s a> co f— co (/) íO O C 1— cu i- o a • c\j «a-1— co o oo o o 3 -o to E : Uf) i— i— i— o ai •«•i- O ra +-> co s- o > tä- \£> r «3- X i- E a> > I— ( CM e tu x: io co LO l>0 .c +-> CO LO E +-> to c -o c a> c O e c •r- +J «3 n3 O O C O W N i o o 0) ra c 0)+-> o ~a -T3 3 C O r— O)+-> O 0) to QJ O CM > co ra X) o -c O N O CO Ln M CMco(co _Q -M+-> LD «d 4- c o ai id 4-> Sc a: •— CD fdo • — ' fd

CO C\J li") O (*) CM CM r o o o o co oo LO o co OJ LO !— r-,o o •OCNJOONtíOJOO o o

CTI CO CO i— CO «3" CM r

i

c\j co o cj>f0 c CMC O 0WhNI-

Hg-258

OOO-TOJO

lOfOOfOHr o «a- co • LO r- CO • O O O O O

co

• m oí o « t O CTI r

OMO o o o

•=í ro o o o

fOONOOOlOr

CM LT> Lf) CTì I— UO fO f l Cl • I • • • I co cm co io i— co o co o co o • I

r-»

f— o o

o

I J

i—

i— Oï o o r-«.r-«. o i— D JOi-fOOMNWOi-

uo un o

o o

o

o o

o CM o oí

•íroo CTi C0 • I

ü L s r

r-» o o r O «il- co "J3 ID O • • 1 focoo i— en o o CM o co o o o

C\J o

co

o

o co o o «3-

i—

^

co o o co o oo CM • O O I— -Q O

n o o n

C\J O I— co C\J

O

CM

LO

O O O

CO O CM O O

CM

LO

O CM

O

LO

O O O

co co

O

I—

o O

CM

co

o

o

• LO

O O O

CO co o o

«a- o o

o

% m ^ o c i o o

— CO CM ^j- O CO CM

O N CO •ÍOOlO

o in i— Ñ

CO co O CO LO LO

co o

LO

o o r co o o

O » CO CO CO "O CM CMO OO •I-OO CM CM CMO

O N O

CO 00 O Oí CO rOlOOMÛ^

o

RC O F D o o o o o o +•>

X-r-'i-r— i- O) N | -

CM COOOÍOOLOCMOOO

aj co co co -a CM CMO 0 0 •I- O O CM CM CMO X -R- R— i- 0) a>

Hg-260

O

O O

co

O O O O O O OÍ FO C T - CMC

aj ce Sa> • M- O

tu ' (O S- -I- oo O T- +J co ai • o —

LO

LO O LO 00 LO CM «dl ' I l i OJ Oí i—«3-LO LO o co co co

r*«. oj oí co co co O O Oí 1 O Oí LO Oí LO c -o • O d «=r ' rüj -» +• Oí o a> « • * ».1— > c -—•o -\ai a> lo X LO C lo A S 3Û. O Oí I— Oí O í. i— 0) »— •>— •—• >»— +-> u .c i— CT) H +>|— id -O -c to co c « OY—•r- m £ fOs- id r— J- OJ O OJ 3 3 I >,4- N E O {/} 'rcn "O C\l C 0J 1TJ (I i- +J "fli— W Ol'rs- >> to «3 LJ -i- Or c a> +-> 4-rao "O

«3" LO LO i— oí co 00 CO LO «d* CO OJ

- r-«. í. ro —• cu O) l/l u I— W W U -r fl a) IA 0J C 3 i_ Ífl OI W fO D1-P ~ - _| >> C o .—» i/> o • >. oí fd i to . CO I— o -M CO >) a» co no co •— 4» CO CO CO ) «OD I o C C r— 73 OJ OJO o o co 0 1—

r-» r-» CTì

1 1

0 CTì CTì

S

et

«3rv CM CM LO CTì CM i— O 1— LO CO LO LO CM 0 r-» CM 0 O CM 1— LO

co

co co i— «¡a- ctì LO f— LO LO CO CO 1— LO •— 1— CM

cn co 0 0

LO LO cn cn

0

N ^ r s « t OÌ rv n a i in o ì r— CM 1 • 1 O LO CO LO CM CO O CM CO CM

cn cn CTI

td 00

1 1

LO r-*. 0 0

CM CO O O

0

0 0

1 1

S LO

to

r-«.

r ^ CTì l o co CM CTì .— tO LO OCTì 1 • 1 LO CO 00 ^ LO O O

rv 0 0 0

0

>

C x : 4- C 0 E L) 0 D > — +->

fd r— cu c (d CTÌ«C < to 0 S id • .a p-» od co < a> s c LO cn p». a> * > ctì _1 LO 1— r>. 0 CD LO 1— c r-» fd CT) cn oí +J G -M ai 0) r - E IA fd 0 0 s e n o O T3 "O to co c c ai

c CM t/> LO 0. 1 a. •1- 1— CO s*s 1 • " LO a> a . a> r— +-»

—

1 CU c > 0 II CD O to to ra O)

il CD 0) (/) fd 0 0 cn id a. c *o 0) -0 O c It cn fd Q_ a> e a) X 0 >>

Q. C •i— TD 0 1

Q£ 0 LO c

rv ai (/> > c < LO CTì c a> •«> i- S r r>- E 0. 0 c Oí ra V) 0 fd •r- C Q. —-TD 0 C +•> •!— fd c fd •1- c • E r - 3. fd 0) + j ai 1— +-> ra TD 0 a; s- +-» •» 3 a> IA - stA fd to 1— r - -C a E c - 0 O) r - (J C 0 c •r-3 «d CiL •1- CTVrQ. O) tO n «d stO CT) 00 —i a> 0 LÛ c td 1 . r. CD 00 0 +-> û£ '—- rd to LO < LO E LO CM HS pv >> CTì vì

Oi P-. O r— a> •a JZ e fd J- CTì trt ^ a; s- a> 1— 3 Q. 3 O Il 0 < 00 > U- M« t to • o_ S i - CT) co CT)*|- 1— _J 00 6 oi r--.CO oo o tí ONCO CO Vi ÙO rH tí D3 CO ^ o 0) S M rH v£> 3 ON rH o- en »Ss CO •rH ON 3 0) 4J H rH ta H 01 •M un •H T3rH vO 4-1 -H a; r-. 01 0) 4-1 ON tí (¡tirH •H H bO U CO /-s (U a oo tí

ÌJ N W S 01 efl 4J H > O M rH -H ' *H C 0» U « H cO • o y (d •Ö C C0 -H 4J C3 o) ai T3 u n) co -H -U -H O M -H O Pn H M > H 01 0) O 0) 4H tí T3 fl ¿ O M Cfl)U Ü -H ttí O -H -H 4-1 W g 4-1 M ¿ 0) O il C n S C CH M oomoinm ii '—• «d cO O tí -H X 41 - 2 ¡ 3 T3 tí "-tí • H T3 • co o 4J tí tí 41 - J-¡ (U u> Cd 03 ^ r l y H (í tí tí O O II O) PQ to 3 w PM II 01tíw « J a) CU CO HO M 0) > Otí 0) •a o ^ n co (U í-l ctí u PQ oeo a) (tí M-l o o rH 01 >> Ü a) P3 M 354c1 u 0) OS co M o 4H 0) O 0 ) c O M 0) CO 41 - 0 & ctí 43» cHd H>>T3•tí •rl H-—» cd rH4> pi rtí CT\ 01 3 CH N HMHo •H rH co M ai co 4J co < a) a otí H w C 0 1 T> 4J w l/l ^ tí ai H >-l OStícT3 o ex o cr a)«—- cd H 3 M a) o co > 4J tí a» tí cO o CU co cd co CUrH- • H á 0) rH O ¿2 w ss co 41 - co -n rH UW COÜ m 4J 0) a) Oo • H 2 a. c tí m c O -H t3 M Cd •H M cO co XI PU 4J S E-4 CJ i-H 00 "-O

CO N O CO rl CM lO >í) ifl N •J r-» -— CO o ai o o O LO o o o CO

co a> co a>

o

r-x. LO co 00 r». co Q_ 1 LlJ hz LlJ s: 1

oo o D_ S o o LlJ 1— Z LU s: «

co

a; X o

o

ai ai o

co

r-.

«3- CM l o r-» oo CJ co CJ co o CM LO o CO 1 o 00 o o co co

CJ en

o o o LO

r»«. Oí

«3o co CO LO co CM co o r-*. LO o LO LO CM co o o o oo o o en co o «aCJ

LO Oí Oí

m co co CM o j o o O o o CM CJ CM oo o o 1— s- 0) X •>- • a> OÍ 4— X Q• i . -Iz o o-+-> "O • i— c cu < »< (T34-> • LO "O £= i- 00 1 (O —»rv. s- o N O I O Q. i— M- O) > Oí +J i - X i - co • "O s- o ai 3 s- c LO r - 3 A3 i t < -p a) — CD o • 3 a. CM Q_ •!Q. • • X 1— LO CL io 1— o >s aj -Q a> i . c . +-> -a .e fT3 -r- ^ ro Oí E E H— "O ^ «3 •«i- >> . . o r - «3Û. S- CM O en -P •— i LO s_ +-> - ra «3 O •• a; •i- _a -i— LO E -C C0 > s co o r— O •p 1— o r-«. o =tfe s : o • a> LO LO ••> LO i— to to • •O) • T3 -— IO •r- LO _ i < c LO 1 «3 - — C E > > " — • LO a; • I— O r -o P>- X to •• ro => C LO r : O a i -rO I— z s- to a> +J « w O O >) Oí -a 4- c +•> ë ^ c -o c a •• 13 >> fO LO c¿ O -r- +•> >LO -O o 0J • •> • o =«a LO CO o LO ^ V +-> a> a> > —s CO l/i to LO O -P o co i— >,~o f— LO T i— c .o OJ r ai ro O "O - J si— c o c -o —co .a aj r^ to o E •P • Oí CJ ro -iLO 10 I— » X LU 0) ai 4- ai • E to c C a> - c co r ai — (0 r - t—t toa i E

ro

ro

Table Hg-20 (11) " F e r n tchromit" Cases

No Pattern

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Enriched in Chromite

Enriched in "Ferritchromit"

AI, Mg AI, Mg, Cr A I , Mg, Cr AI, AI, AI Cr AI , AI, AI, AI, AI, AI, AI, Cr, AI,

Cr

Mg, Cr Mg

Mg, Mg Mg, Mg, Mg Mg, Mg, Ti, Mg,

Cr, Zn Cr Fe2+ Cr Cr Fe2+, F e 3 + Cr

Fe, Cr Fe Fe Fe Fe Fe Fe Fe, Ti, Ni Fe, Cr, Ni, Mn, Co, T i , Zn, V Fe, C o , Zn, Ni Cr, F e 3 + Fe, Cr Fe Fe A I , Mg F e 2 + , F e 3 + , Ni

Cases 1-10 are as summarized by Onyeagocha (1974); 11. Golding and Bayliss (1968); 12. Engin and Aucott (1971); 13. Springer (1974); 14. Ulmer (1974); 15. Hamlyn (1975); 16. Bliss and Maclean (1975).

Oxide Ti0 2 AI?O3 Cr203 Fe203 FeO MnO MgO CoO ZnO NiO V2O3

Total

Chr 1 0.08 17.5 51.1 4.1 12.4 0.44 14.3 0.07 0.05 0.08 0.16 100.28

"Ferrit" 2 0.13 9.9 55.5 7.9 14.1 0.55 12.4 0.10 0.10 0.13 0.19 101.00

Chr 3 0.05 18.1 50.4 3.9 12.6 0.4 14.2 0.05 0.05 0.09 0.12 99.99

Mt 4 0.02 0.56 4.2 64.9 27.2 0.45 1.7 0.21 0.24 0.60 0.13 100.21

Chr 5 0.29 15.27 48.32 7.47 17.57

6 0.21 9.25 47.24 15.14 18.21

-

_

11.07

9.89

_

-

0.01 -

100.0

_ _

0.06 -

100.0

" Ferrit" 7 0.78 0.08 28.65 41.58 23.57

_

4.88

_ _

8 0.14 0.15 2.93 66.02 29.46

Chr 9 2.14 15.1 39.7 9.2 30.7

0.96

2.14

_

_

0.45 -

100.0

-

100.0

-

_

"Ferrit 10 0.00 63.6 0.88 1.2 18.9 -

15.1

_

-

-

-

-

-

99.0

-

99.7

Analyses are for chromite + "ferritchromit", and chromite + magnetite pairs, or for chromite with successive (intermediate) zones of alteration. The examples quoted are representative of the three contrasting enrichment trends reported for "ferritchromit": (a) Fe(Fe2+, F e 3 + ) + Cr + Ni enrichment, Mg + A1 depletion (# 1-2); (b) Fe ( F e 2 + , F e 3 + ) enrichment, Mg + A1 + Cr depletion (# 3-4 and 5-8); (c) Mg + A1 enrichment, Fe + Cr + Ti depletion (# 5-6). References: Analyses 1-4. Twin Sisters dunite, Washington. Onyeagocha (1974). Analyses 5-8. Serpentinized ultramafic, S.W. of the Manitoba Nickel Belt. Bliss and MacLean (1975). Analyses 9-10. Panton ultramafic sill, Western Australia. Hamlyn (1975).

Hg-268

O

cu S fö CT -o to Id IO0O 0 s_ lü o aj (0 0) Q) Tra l/l CU J-> - - P O) 3 3 E • P id fa o - O (O fO "O 3 c:J3 (A C - P a) rO o fO ÍT3 ra o •p (/) CD -a 0) +-> >i - P •p . •r- +->• P c c to 00 jSO) en a> > A> A; a. o. o O -o 3 V) IO o .C c ÍO ra (0 >> >> C £ h- < X co m o O IC 31

U

O) s-^-«. E JD OJ

. O O O O O O O O O O L O O O O O O O O O O JOJ CMCSJCVJCMCJCMCNJCM

3 «3 fC

S- 4-> +-> 03

to i O cü O) to _Q .c ro •P o

O) en (/> in to OI CU O)

ti "O 4-1 4- to •p O ai to Q3 CL) fO >» crS- JZ 3 o "O C •P -p .E ta ai 1/1 O i. aj a» ai io 3 3 ai o O)- P Q. rO «3 -P i> E ' Etn o - P (V to CJ a) cri LÜ -p c - P U rtJ ai s--a c/) ta fO c 3 o. CL a)-a E a. 3 (U o u -o cr Q. CL-i— cr X >>-«A: JC J= o ~o 0) -p to a) O) fO ai "O 0) > E fO S- o rO o en cr» 3 ta o -p a> c s: TJ E < O) «o -C -o O)•o +-> C •p LL_ 00 Q. ai T 3 0) >>-P C to

Hg-269

a; E > ro 4- CT) O CD » 0 0)11 o s -P C CMIr— J3 C iro O N HJ-r Ü to to cc 4- en u c a> "CT i— m 3 TU •1- -a - P ~o 3 C AJ cr 00 FO O E S- (O 1 ai e u r » . o • S- CT CT» -f- (1)JH -p - P S - • • ro 0) 3 C ^ •P «o o o- P S- > « N E -P 00 SFO X O -rai - P QJ E - P >— to a)a» 4- ta CNJ a_ (Q i— «i— i- o E O s_ C/Ì ra n3• P E to o co 0) (O -P C - P CD to O C O a; c ai >>••- o to E o io 3 S- -P ta CTai -p 3 cr o ftj - PQJ FO E "O to •r- N fO I - E -p o E c .«—••.•«— -ao 3 •p jrun i— QJ rO 0) ai cr sz rd II (J -Iai o -C +-> hCTU-MU c ai C O CO •i— > •o

re

• X r • a) o i o u> iura

i— o i— r - -o so a> c Ho «o

a a) a. to ft e -i a)-p fd

4J to II +J to 0) XJ 1— O to o to c 3 3 O CÜH- O E O r - í) E or-. > id a> E "O O > CT) IO 1 o aj JD fd o So •r— i- fd -M 0) "O fd fd TD TD O) H- 0J N -i- ai -E .c í. •r- tO 4->+-> QJ fOi— E tO 4> i— o •5--- +J O 3 Id CM fd S_ to o > o .e "o +J aj -i— C «3 o o to _E •r- "4— >> S -p- X to r— o to E O u • > 4-> O CU +-> LUfd E CM^ fd JZJ CL TJ fd CO E .a "O 4-> • id to o • c rao E saj (d -M Q+J 1 to rd > o • o o .o LO a> E 3 a> fd i- 3 SO O > f— 0) 0J 3 •i- O 4-> 3 "O O í- +-» a. fd rd f— ai o id 5 -o id 4- O) S- r— ai +J 4- CL+J i— sC 3 10 X «13 O QJ 0) O) -P soo CM (_) E CT> O o í . 3i— O . ai i— o 4O) o_ a> 1 LO . a> > X O CM Ox: rtJ LU +Ji— r -

TD ai r - hT- E c 4-> U IO 4-> o o Sai •r- O " 0J tO E • ai to i-+j fd -—• -r— to id a» ai sz NI fd Q.TD o X a> • • E i— (U E to o o . E ai to 0) to O) i- o +-> 3 CD TD fd 0) en fd i oí a i ai id r— a) fd c .c U-Pi- Oh o E .Q E • E S- +J .. id -I- • O to E S- CX N CO O "OO a» ai c N O r - o o o. to o E LO a> X í . »• rd •!- ai TD to fd o O s- ^ 3 CL +-> c o c - p 1— -P 3 to ai O -Is_>— Ll_ ^ 4E 4-> O • to 3i— to«— «d TD 1

E 4-> \ E O i— E rd co O o S- TD > i— O o CL E •r- id a i o o id i— 4"O o aio E c fd o 0) -E 1 O «d +•> o S- -P 1—o sc o O -r- r - 4- "O ai o 4- 4-» +1 a> fd • — TD c E o to E to . QJ o r T aj o cur-. > N -reu o > _Q Q.O s- s- s_ 1 a> X O 3 ld 3 o ai SLJJ to U O O 1 — "O id

fO u z.

I— fd : i— -p ) •>- c

X Ha) ; ai o o 4-> u o c ai +-> «— i a a o ra 4- cu O C il o j t/> s lo i— í- ai ai ai O -r- CMO 'J O) LL (TI r- OI > i- 5 -r- Q. -i— CU -P u o o > Sa. c 4- lo ai ai 5 E ai 03 co r I: ÛJ rtJ -P JD » V J +j T3 +JT3 co iß rO LO O h- . -i- J o c" •P "O O i25+ Otoxcu>C r— O -P : O E -r- O -P 3 O) Ol O C "Oí— "O (O X •!- -P E •«- i— ) aiaiaiTpcr-iu • N .a i- T5 fO S- ST" >>." •r- is 0) CU • — - • - a) -Q Qra r > E ai

ai sCJÌ CL ro ai • sz «a -O +J u C •!ro E iO 4>4 ai i— 4 o ai > JD 03 fO LO

ai 4- o ai ai c -oP roí -— r-'r ai 5 ai o lo a» +•> o cm ai lo ai o. o O S- Or- «3(-

(O ro S 5

I— "O IS)

Hg-271

c o to

4- "O Oí LO LO C C N OI "a 03 o r "O C M O "r— ra o lo X o lo lo o

3:

co 51 LÜ ht/i >00

< c 1 (O • r - c o to to O 3 O CU 03 i— 3 c o c r CM UO Z3 • 3 S- t o e n o a> e n c s_ 1— o ^ c E -f- ai • r - s : O - =! •1r fd tu - P - — z s- t o 3 f d 3 t o . £ A3 U_ 3= o c r o j 1— - r - CO o w e s : -—. t o CMS: CD E «51- E u . fd O E •«• t o cu E: o D_ 3 - a J D QJ IS) ^ - P *r-• r - e n SC • p f— • P E E C (O + er f d f 0 CU dJ •— QJ - P f d C O^W C E to R s: >,U_ CU i n t o Q_-I— _Q QJ d i •Ie n s : fd t o to E CU E + N Q . X Ì S- r— • o to^s x : 03 a> oo £ fd s- 0J i d W 4 J + J W +-> - o f d 13 s_ E 3 t o CU E TD -N- t o en c r aj ai 3 e X r-1 -P E a; E + • c c +-> CSI +-> x : E t o i — o cu 3 o " a fd tj t o II - C - Q M— C to fd fd CÜ a i CU X i o -P o. u r o X) E 4-> 3 Q_ d i ra t o tO E c r Il l_l_ - ro o • ' to o C -—-1— x : fd QJ rd o X ) CM O O E (J to Q - - C -p + U O «Io O 4- 4- o to to t o •P > t o !— " O S4 - UO E - P c o i— fd 3 •!— o TO O i r - 1 S(fl X V E 1— M - ! - CO i d E a j o O c Il OJ r«- r— *r- s"O o 0J CTf— UO • ! +-> cns: O E — - p aj E a E (O • P E u . ì : a - r O CJ «3 • P S I ^ O fd fd U_ r o u s~ a) OJ E í- 3 IO CU •< CNJ •1- X ) 4 - CL) 3 E .00 4 - -r- X) OJ JC o - r - u n c r i o u n E T- t o 1— - r - E CM D ì 3 t o id . E to E O r - i - i t N • i - O" O t|_ o ' co 00 dt Oí

J 3 O i— J S- CNJ 3 Ol :

o

- ÛO o

CM O 4 - CM CO 0) CO CM - o i— X O

CM o O CM S- 4 . SCO cu Q . r — X 1 LÜ . S«— CU O O l o 1 to i• p aj aj '4x: cri 73 . (u • P LO O c o s_ " O fdo O u aj E > ) S- • p fd f d 4 tO í. CM i d to QJ E 4CL OJ cu - I - o 3 fd aj E 4c o .E 3 to to*x E aj . E I— ( J HO 4) • ! - > r to o f d QJ + f d o S_ 4-> tO • i 0)0 x: > Í - 0J o TJ E i •P "O CLÜCL - P to o E o E tO i d - p f d CM 3 E f d > 1 0) D - p IO QJ QJ CL CT X ) QJ 1 x : í. C_> 1 O CM to LO 3 - P Q - - 0 +J fd o 0 ) - P E +-> E • i - a j 4- O o «aO 4 - E I— O a> O -P to o U •r- fd X) 4 - ai a . xa CM 1— E - a s- to - P 3 O QJ 1— II -P CL sE II h O E fd f d u o fd (d QJ X I cu E 1— P CL XI -o QJ t o • o eu a j o E cu id o je - p o t o +-> Í— i d -P -F—'+-> • P E u n E Xi XI N O un O CL S- f d 03 «d CM CU " O E -P t o 0) •P -R- E CO s- a j O - I - QJ t o - p 00 U id •i> CU •1— CL > - r - o id E fd - o s•p •P •!CM o > Sto o dJ 3 O O E o CJ >iO "O ts> s_ •o •p O SE X I f d O J- E T - i E -P 3 4id o QJ t o CL • r - - P o i d 03 X i o >i

Í. -QJ p

lity

to CU CL O 1— to

atms.

1—

>> QJ •P id E X O QQ. fd tO "C TD E fd c QJ •p

D E C O S- + • IX. X o X QJ a E c u id -P to >» Su

a X •p 4O

-P o 03 - p .E E -P E TD f d OJ .E E -P O SQJ O -P Si- f d fd CL TD >> O Dì 4 E O O J_ O +-> O to

T a b l e Hg-21 T-fQ2

(5)

Curves

Buffer

Reference E u g s t e r and Wones

(1962)

9 - 25,738 = -A T

Wones and G i l b e r t

(1969)

=

-A + .092

E u g s t e r and Wones

(1962)

f02

=

9 . 3 6 - 24930 = -A T

Huebner and S a t o

f02

=

-A + .046

E u g s t e r and Wones

f02

=

1 3 . 3 8 - 25680 T

Huebner and S a t o

f02

=

1 4 . 4 1 - 24912 = - A T

E u g s t e r and Wones

(1962)

f02

=

- A + .019

E u g s t e r and Wones

(1962)

Mn304-Mn203

f02

=

7 . 3 4 - 9265 T

Huebner and S a t o

Makaopuhi DH # 9

f02

=

6 . 4 4 - 21900 T

S a t o and W r i g h t

MW

Pressure

correction

FMQ

Pressure

correction

NNO

Pressure

Mn

correction

i-x°"Mn3°4

MH

Pressure

correction

f02

=

1 3 . 1 2 - 3 2 , 7 3 0 = -A T

f02

=

- A + .083

f02

=

f02

T = °K;

(P-l) T

(P-l) T

(P-l) T

(P-l) T

P = bars.

Hg-273

(1970)

(1962)

(1970)

(1970)

(1966)

ANALYTICAL TECHNIQUES METHOD

COMMENTS

REFERENCES

Magnetic Property Measurements Techniques and Applications: Sampling techniques; instrumentation for measurement of magnetic properties; procedures for magnetic s t a b i l i t y t e s t s ; i s o t r o p i c and anisotropic suscepti b i l i t y measurements; èpplied f i e l d techniques; magnetic component measurements (optical and X-ray; chemical; magnetic; low temperature i d e n t i f i c a t i o n ; magnetic domain s t r u c t u r e ) . Review and update of techniques.

Collinson et a l . (1967)

Col 1inson (1975)

Standard Texts: Magnetism and the chemical bond. Rock magnetism. Physical p r i n c i p l e s of rock magnetism. Physics of the earth. Paleomagnetism - application to geological and geophysical problems. Applications of paleomagnetism. Reviews : Magnetic properties of rocks and minerals. Magnetic properties of minerals. Rock magnetism. Paleomagnetism and rock magnetism. Magnetic properties of rocks. Theory and nature of rock magnetism. Magnetic effects associated with chemical changes in igneous rocks. Recent developments in magnetic materials. Theory of magnetic v i s c o s i t y of lunar and t e r r e s t r i a l rocks. Lunar magnetism. Meteoritic magnetism. Noteworthy Publications: Superparamagnetic and s i n g l e domain threshold s i z e s in Fe,0^; indices of multidomain behavior of Fe,0 4 in igneous rocks; TRM of submicroscopic Fe.O.. Cationic d i s t r i b u t i o n s in titanomagnetites.

Special Techniques: Magnetic c o l l o i d ; ionic bombardment and acid etching.

Goodenough (1963) Nagata (1961) Stacey and Banerjee (1974) Stacey (1969) I r v i n g (1964) McElhinny (1973) Lindsley et a l . (1966) Akimoto (1966) Runcorn (1966) Wilson (1966) Verhoogen (1969) Hargraves and Banerjee (1973) M e r r i l l (1975) Wernick (1972) Dunlop (1973) Fuller (1974) Brecher and Arrhenius (1974); Larson et a l . (1974); Stacey (1976). Dunlop (1973); Dunlop et a l . (1973a); Dunlop et a l . (1973b) O ' R e i l l y & Banerjee (1967); Readman and O ' R e i l l y (1971); Stephenson (1969); Stephenson (1972); B l e i l (1971) Petersen (1962); Soffel (1968); Soffel (1970); Soffel and Petersen (1971); M e r r i l l and Kawai (1969)

Mossbauer Spectroscopy Standard Texts: P r i n c i p l e s and applications.

Frauenfelder (1962); Wertheim (1964); Gol danskii and Herber (1968)

Applications to Fe-Ti-Cr Oxides: Fe and Ti s p i n e l s with local disorder. Behavior of Fe z + in Fe-Ti s p i n e l s . Cationic d i s t r i b u t i o n in Fe Ti0 . Cationic d i s t r i b u t i o n s in pseudobrookites. Fe 3 + : Fe^ + ratios in titanomagnetites. Hg-274

Banerjee et a l . (1967a) Banerjee et a l . (1967b); Jensen and Shrive (1973) Rossiter and Clark (1965) Virgo and Huggins (1975) Watt et a l . (1973)

Table Hg-22 (2)

ANALYTICAL TECHNIQUES

X-Ray Diffraction Standard Texts: Elements of X-ray crystallography and the powder method. Tables for identification. International X-ray diffraction card index. Applications to Fe-Ti Oxides:

Azaroff (1962); Azaroff and Buerger (1963) Berry and Thompson (1962)

See Chapter I (this volume)

Electron, Ion and Laser Microprobe Analyses Standard Texts: Instrumentation, procedures and applications to microchemistry. Noteworthy Publications: Ion microprobe mass analyzer. Fe2+ : Fe' + determinations and L.a X-ray emmission spectra. Correction factors.

Applications to Oxides:

Heinrich (1966); Andersen (1973); Reed (1975); Goldstein and Yakowitz (1975); Smith (1976). Andersen and Hinthorne (1972) Snetsinger (1969); Albee and Chodos (1970); O'Nions and White (1971) Ziebold and Ogilvie (1964); Bence and Albee (1968); Sweetman and Long (1969); Albee and Ray (1970) See Chapters 3-8 (this volume)

Electron, Scanning and Transmission Microscopy Standard Texts: Principles, techniques and applications.

Applications to Oxides: Magnetite-ulvospinel exsolution. Titanomagnetite-ilmenite intergrowths.

Ilmenite-hematite intergrowths.

Hall (1953); Holt et al. (1974); Andersen (1973); Goldstein and Yakowitz (1975); Wenk (1976) Nickel (1968) Lewis (1963); Evans and Wayman (1970); Evans and Wayman (1972); McClay (1974); Hoblitt and Larson (1975); Davis and Evans (1976) Lai ly et al. (1976)

Reflection Microscopy Standard Texts: Physical and optical determinative methods: microhardness indentation techniques, reflectivity measurements, optical rotational properties. Noteworthy Publications: Microhardness. Microhardness vs. composition in magnesian ilmenites Reflectivity vs. microhardness. Textures and Genesis: Microscopic identification, textural intergrowths, comparative tables, and interpretation.

Photomicrographic card index.

Hg-275

Zussman (1967); Freund (1966); Cameron (1966); Hallimond (1970); Galopin and Henry (1972); Nakhla et al. (1973); Bowie et al. (1975) Millman and Young (1963) Morton and Mitchell (1972) Millman and Kingston (1964); Chamberlain et al. (1967) Ramdohr (1969); Oelsner (1966); Uytenbogaardt and Burke (1972); Galopin and Henry (1972); Cameron (1966); Edwards (1965); Bastin (1950); Shouten (1962); Deer et al. (1972) Maucher and Rehwald (1961)

Table Hg-22 (3) N oteworthy P ub1i cat i on s: Textural interpretation of oxides.

ANALYTICAL TECHNIQUES

Ramdohr (1969); Edwards (1965); Buddington and Lindsley (1964) Ramdohr (1953); Vincent and Phillips (1954); Vincent (1960) Vaasjoki and Heikkinen (1962); Tsvetkov et al. (1965); Nickel (1968) Ramdohr (1969); Carmichael (1961); Edwards (1965); Kretchsmar and McNutt (1971) Ramdohr (1969); Vincent and Phillips (1954); Buddington and Lindsley (1964); Duchesne (1970) See Chapter 4 (this volume)

U1vospinel-magnetite exsolution.

Ilmenite-hematite exsolution. U1vos pi nel-magneti te s s oxi dati on.

Ilm s s and Usp s s oxidation. Pétrographie Objectives Oxide Thermometry and Geobarometry: Experimental T°C and fOj determinative curves. Computational methods and procedures for calculating mole % Fe2Ti04 and mole % FeTiOj from microchemical and chemical data. Data Plotting Procedures: Ternary FeO-FepOq-TiOo Ternary F e p O ß - F e f ^ - M g T ^ Spinel prisms

Buddington and Lindsley (1964) Buddington and Lindsley (1964); Carmichael (1967); Anderson (1968); Lindsley and Haggerty (1970); Rumble (1973). See Fig. Hg-23 (this chapter) See Fig. Hg-42 (this chapter) Stevens (1944); Jackson (1963); Irvine (1965); Haggerty (1971); Haggerty (1972); Busche et al. (1972) . See Fig. Hg-24 (this chapter)

Hg-276

Abbott, D. (1962) The gabbro cumulates of the Kap Edvard Holm complex, east Greenland. Unpublished Ph.D. thesis, University of Manchester. Ade-Hall, J. M. (1963) Electron probe microanalyser analyses of basaltic titanomagnetites and their significance to rock magnetism. Geophys. J. Royal astr. Soc. 8, 301. (1964) A correlation between remanent magnetism and petrological and chemical properties of tertiary basalt lavas from Mull, Scotland. Geophys. J. Royal astr. Soo. 8, 403. (1964) The magnetic properties of some submarine oceanic lavas.

Geophys. J.

9, 85. (1969) Opaque petrology and the stability of natural remanent magnetism in basaltic rocks. Geophys. J. Royal astr. Soo. 18, 93. (1974) Strong field magnetic properties of basalts from DSDP Leg 26. In, T. A. Davies and B. P. Luyendyk, Eds., Initial Reports of the Deep Sea Drilling Project, 26, 529, U. S. Government Printing Office, Washington, D. C. _ _ _ _ _ (1974) The opaque mineralogy of basalts from DSDP Leg 26. In, T. A. Davies and B. P. Luyendyk, Eds., Initial Reports of the Deep Sea Drilling Project, 26, 533, U. S. Government Printing Office, Washington, D. C. , and N. D. Watkins (1970) Absence of correlations between opaque petrology and natural remanence polarity in Canary Island lavas. Geophys. 2-02: Intensive parameters in the Skaergaard intrusion, east Greenland. Am. J. Sci. 271, 132. Wilson, R. L. (1966) Palaeomagnetism and rock magnetism.

Earth Sci. Rev. 1, 175.

, and N. D. Watkins (1967) Correlation of petrology and natural magnetic polarity in Columbia Plateau basalts. Geophys. J. R. astr. Soc. 12, 405. , and S. E. Haggerty (1966) Reversals of the earth's magnetic field. 25, 104.

Endeavor

, P. Dagley, and J. M. Ade-Hall (1972) Palaeomagnetism of the British tertiary igneous province: the Skye lavas. Geophys. J. R. astr. Soc. 28, 285. , S. E. Haggerty, and N. D. Watkins (1968) Variation of palaeomagnetic stability and other parameters in a verticle traverse of a single Icelandic lava. Geophys. J. R. astr. Soc. IS, 79. , N. D. Watkins, Tr. Einarsson, Th. Sigurgeirsson, S. E. Haggerty, P. J. Smith, P. Dagley, and A. G. McCormack. Palaeomagnetism of ten lava sequences from Southwestern Iceland. Geophys. J. R. astr. Soc. 29, 459. Wimmenauer, W. (1966) The eruptive rocks and carbonatites of the Kaisersthuhl, Germany, 183. In, 0. F. Tuttle and J. Gittins, Carbonatites. John Wiley & Sons, 591p. Wolejszo, J., R. Schlich, and J. Segoufin (1974) Paleomagnetic studies of basalt samples, deep-sea drilling project, Leg 25. In, E. S. W. Simpson and R. Schlich, Initial Reports of the Deep Sea Drilling Project 25, U. S. Government Printing Office, Washington, D. C. Wones, D. R. and M. C. Gilbert (1969) The fayalite-magnetite-quartz assemblage between 600° and 800°C. Am. J. Sci. 267-A, 480. Worst, B. G. (1960) The Great Dyke of Southern Rhodesia. Bull. 47.

Southern Rhodesia Geol. Surv.

(1964) Chromite in the Great Dyke of Southern Rhodesia. The Geology of Some Ore Deposits in Southern Africa, II, 209.

Geol. Soc. S. Afr.

Wright, J. B. (1961) Solid-solution relationships in some titaniferous iron oxide ores of basic igneous rocks. Mineral. Mag. 32, 778. (1967) The iron-titanium oxides of some Dunedin (New Zealand) lavas, in relation to their palaeomagnetic and thermomagnetic character (with an appendix on associated chromiferous spinel). Mineral. Mag. 36, 425. , and J. F. Lovering (1965) Electron-probe micro-analysis of the iron-titanium oxides in some New Zealand ironsands. Mineral. Mag. 35, 604. Wright, T. L. and P. W. Weiblen (1967) Mineral composition and paragenesis in tholeiitic basalt from Makaoupuhi Lava Lake, Hawaii. Geol. Soc. Am. Abstr. 115, 242. Wu, Y. T., M. Fuller, and V. A. Schmidt (1974) Microanalysis of N.R.M. in a granodiorite intrusion. Earth Planet. Sci. Lett. 23, 275. Wyllie, P. J., Ed. (1967) Ultramafic and Related Rocks.

John Wiley & Sons, 464p.

Yagi, K. (1964) Pillow lavas of Kerflavik, Iceland and their genetic significance. Fac. Sci. Hokkaido Univ. Ser. IV: Geol. and Mineral. 12, 171.

J.

Young, B. B. and A. P. Millman (1964) Microhardness and deformation of ore minerals. Bull. Trans. Inst. Mineral. Met. 73, 437. Zeuch, D. H. and H. W. Green, II (1976, in press). olivine from peridotite xenoliths. Geology.

Naturally decorated dislocations in

Ziebold, T. 0. and R. E. Ogilvie (1964) An empirical method for electron microanalysis. Anal. Chem. 36, 322. Zussman, J., Ed. (1967) Physical Methods in Determinative Mineralogy. New York.

Hg-300

Academic Press,