lunar mineralogy: a heavenly detective story. part ii1 · lunar mineralogy: a heavenly detective...

American Mineralogist, Volume 61, pages l0S9-l I16, 1976

Lunar mineralogy: a heavenly detective story.Part II1

JosrpH V. Snt t ru eNo Int t M. SrnslE

Department of the Geophysical Sciences, Uniuersity of ChicagoChicago, I I I inois 606 3 7

Contents

AbstractIntroductionGeneral discussion of rock tvoes: relation to model of Moon

Lunar specimens with high MglFe(a) selected specimens of special interest(b) others

Lunar specimens with lower MglFe(a) high-KREEP material, mostly norit ic(b) low-KREEP norit ic material(c) anorthositic(d) granitic and rhyolit ic(e) mare basalts(f) ANr

Origin of rock typesDiscuss ion o f minera ls

(b) spinels from non-mare rocks(c) spinels from mare rocks .(d) miscel laneous

IlmeniteArmalcol i te, zirkel i te, and possible related phases .Baddeleyite, corundum, and rut i leApat i te and wh i t lock i te . . .Sulf ides, carbide, phosphide, and metals

(a) sulf ides(b) cohenite and schreibersite(c) metal l ic Fe,Ni,Co minerals

Conclus ionA p p e n d i x : a d d i t i o n a l l u n a r m i n e r a l s . . . . .AcknowledementsReferences

FeldsparsOlivines 107'7Pyroxenes 1079

(a) relation of major elements to rock type and crystall ization conditions 1079(b) minor and trace elements 1080(c) exsolution, inversion, and site population l08l(d) deformat ion 1082

O t h e r s i l i c a t e s . . . . 1 0 8 2Spinels 1084

(a) general chemistry 1084

| 060t06 rt06 rl 066r 066I 070r07 |t07 I107 Ito121074t074I 075I 075lo'76r 076

I 084I 0861086I 087I 089l09 l1092I 093I 093I 095109'7I 106r r07I r08I 108

r059

1060 J. V. SMITH AND I. M. STEELE

Abstract

This second part is a revised and expanded version ofthe 1973 Presidential Address, andincorporates new data publ ished up to November 1975.

The earl ier l ist of lunar minerals is expanded to include cordieri te, molybdenite, farr ington-ite, akagan6ite, unidenti f ied Cl,Fe,Zn-bearing phases, bornite(?), aluminum oxycarbide(?),monazite, thori te(?), pyrochlore(?), t i tanite, and graphite.

A general discussion of rock types is set in the context of the phase relat ions in thesi l ica-ol ivine-anorthite and Fe-FeS systems, and a model for early primary dif ferentiat ionfol lowed by part ial melt ing of cumulates. Most of the returned rocks are metamorphosedbreccias composed of several materials. Lunar specimens with Mg-rich minerals are reviewedin detail because they may represent early crystalline differentiates. Emphasis is placed on thechemical composit ion of ol ivine and pyroxene as a means of sort ing out the host rocks, whiletextures and modes are given less signif icance because of errat ic mechanical and metamorphicprocesses. Rocks with Mg-rich minerals are classif ied as dunit ic [with ol ivine For, ,n,Cr,Al,Fe,Mg-spinel, and sometimes a t i tanate (armalcol i te?)f, toctol i t ic (Mg-subgroup withol ivine Foru-", and Fe-subgroup with variable ol ivine mostly Foro-ru), spinel-troclol i t ic (di-verse ol ivine Fo", ,0, rare Mg-rich pyroxene, and Mg,Al-r ich spinel). Rocks with minerals oflower Mg-content are classified as high-KREEP material, mostly norilic (pyroxenes mostlyEnuo-roFsru-ruWor-u, rare augite, relatively Na-rich plagioclase Anro rr), low-KREEP noritic(pyroxenes mostly Enrr-rrwo-3 but some Enrr-o.Wonu-.r, plagioclase Anrr-rr), anorthosit ic,granitic and rhyolitic, mare basalts, and ,4NI group.

Important aspects of the fol lowing minerals are discussed.Feldspars: trace element analysis by ion microprobe; cation vacancy in anorthite; experi-

mental reproduction of textures in basalts; twin distr ibution; domain textures from electronmicroscopy : unusua l te rnary compos i t ions .

Oliuines: correlat ion of Al,Ca,Ti,Cr,Mn, and Ni with type of host rock; sl ight Mg,Feordering; magnetic hyperf ine peaks; reaction of ol ivine clasts in metamorphosed breccias.

Pyroxenes: correlat ion of Ca,Mg,Fe content with host rock and interpretat ion in terms ofcrystal- l iquid dif ferentiat ion; low content of Na and K; content of minor elements includingTi,Al, and Cr; est imation of crystal l izat ion and cool ing condit ions from exsolut ion phenom-

ena and site populat ions; interpretat ion of disorientat ion and cracking.Other si l icates; possible l imit on pressure from occurrence of si l ica polymorphs; l imit on

availabi l i ty of water from rari ty and composit ions of amphibole; possibi l i ty of garnet occur-r ing in deep-seated rocks; uncertain identi f icat ion of meli l i te.

Spinels: apparent absence of tr ivalent iron; new scheme for plott ing composit ions on atetragonal pyramid with ulvdspinel at the apex; composit ion trend from Mg,Al-spinel towardchromite-r ich spinel and then to ulvdspinel as the host rock trends from troctol i t ic to marebasalt; correlat ion of zoning from chromite to ulvcispinel in mare basalt with relat ive stage ofcrystal l izat ion of si l icates; reduction of Ti-r ich spinel to iron plus rut i le or i lmenite plus Ti-poor spinel; part i t ion of Zr between i lmenite and ulvdspinel as a thermometer.

I lmenile: Mg/Fe rat io which probably depends on temperature and Mg content ofcoexist ing (Mg,Fe)-si l icates; deformation occurring at low shock pressure; interpretat ion ofintergrowths with rut i le, spinel and pyroxene in terms of exsolut ion, reduction or shock.

Armalcol i te: contains about 5 percent Ti3+Tia+Ou; breaks down to i lmenite and rut i le belowabout 800-l000oC; has an Mg/Fe rat io which correlates with that of coexist ing si l icates; hasan upper l imit for pressure stabi l i ty at several kbar. Zr-r ich and Cr,Zr-Ca-rich grains inbreccias were commonly described as armacoli te, but may not be isostructural; their Mg/Feratios correlate with coexist ing si l icates; because they are found with Mg-rich si l icates theymay occur deep in the Moon.

Baddeleyite and rutile: Wide compositional ranges.Phosphate and phosphide: relative occurrence depends on the redox conditions. Whit-

lockite carr ies substantial REE. Apati te is r ich in F and Cl, and OH is probably absent or very

low. Schreibersite apparently derived from meteorit ic debris and direct ly from lunar rocks,and the part i t ion of Ni and P between schreibersite and iron mineral gives a useful thermome-ter for metamorphosed breccias.

Troi l i te: ubiquitous in lunar rocks; anticorrelat ion of Fe and S in mare basalts indicates

LUNAR MINERALOGY

loss of S; posit ive correlat ion of Fe and S in non-mare rocks indicates meteorit ic con-tamination; Ti part i t ioning with i lmenite is temperature dependent. Cohenite: results frommeteorit ic debris, and probably from solar wind.

Iron metals'. ubiquitous; indicate a low redox state; result from both meteoritic con-tamination and indigenous processes for which Co and Ni contents may provide a fair lyrel iable clue; occur in veins and vugs indicating presence of vapor.

The sum total of data provides important controls on geochemical and petrologic models.Extensive crystal- l iquid fract ionation, impact dif ferentiat ion, shock and thermal metamor-phism, vapor transfer and change of redox condit ions, and late meteorit ic addit ion are neededto explain the mineralogy. Although considerable uncertaint ies remain, primary dif ferentia-tion of an olivine-rich dry Moon is reaffirmed as a valuable model upon which many subsidiaryprocesses including part ial melt ing of cumulates, formation of an Fe,S-rich core, and impactmetamorphism can be added as subsidiary parameters. The clues offered to the lunar detectiveare heavi ly biased by sampling and confused by mult iple processes, and much further researchis needed.

Introduction

Part I (Smith, 1974) set lunar mineralogy in thecontext of a model Moon, crudely classified the rocktypes, enumerated the observed minerals, and re-viewed the properties of feldspar and olivine. Taylor(1975) provided a well-balanced view of the generalproperties of the Moon, but his detailed geochemicalmodel has some controversial features especially con-cerning the origin of mare basalts and KREEP-richrocks. The development of the early Earth and Moonwas reviewed by Smith (1975), and a model wasdeveloped for origin of the Moon by catastrophiccapture followed by accretion of the Earth-orbitingdebris. Simultaneous accretion of solar-orbiting ma-terial by a proto-Earth and an Earth-orbiting proto-Moon was also considered l ikely, while origin of theMoon by fission from or volati l ization of the Earthwas thought to be unlikely. Uncertainties in modelsfor the bulk composition of the Moon were empha-sized: in particular, existing estimates for Ca differfour-fold.

'Note by J. V. Smith Revised and expanded vers ion of Presi-dent ia l Address, Mineralogical Society of America, del ivered at the54th Annual Meet ing of the Society, l3 November 1973. The f i rs tpart was publ ished in Vol . 59, p. 231-243, 1974, but var iouscircumstances made i t impossib le to prepare the second part imme-diately thereafter. This was actually fortunate because this secondpart , completed in November 1975, could incorporate the exten-s ive data and ideas which marked the end of the Apol lo program.

[Final revis ion of the manuscr ipt in May 1976 al lowed addi t ion ofreferences to several papers appearing after November 1975.] Thefirst part is brought up to date, the rock types are evaluated, andthe remaining minerals are reviewed. Al though a President ia l Ad-dress is personal , few recent Presidents could maintain that theirwork was carr ied out tota l ly by themselves. Ian Steele has been aremarkably product ive col league s ince the second year of my lunarprogram, and i t is appropr iate that we prepared th is ar t ic le jo int ly

even thouqh i t modi f ies a t radi t ion

The appearance of the review of lunar minerals byFrondel (1975) makes it unnecessary to elaboratehere on the strictly mineralogic features, and the pres-ent review concentrates on crystal-chemical, petro-logic, and geochemical aspects of lunar mineralogy.Minerals not discussed in detail, and not l isted in PartI, are given in the appendix.

General discussion of rock types:relation to model of Moon

The confusion in lunar rock nomenclature hasbeen partly reduced as petrologists and geochemistsrealized that most lunar rocks are hybrids which haveundergone sequential changes including erratic shockmetamorphism. Most lunar petrologists believe fhatthe outer part of the Moon (perhaps l0 km thick) isdominated by debris produced by repeated impactfrom -4.5 to -4.0 Gy into a complex assemblage ofigneous and metamorphic rocks. Impacts by bodiessome tens of kilometers across yielded superimposedMoon-wide blankets ranging from a few kilometersthick at the edge of an impact basin to some metersthick at great distances. The debris consisted of com-plex mixtures of material excavated down to depthsof some tens of kilometers mixed in with dispersedmaterial from the projecti le. Secondary impacts anddifferentiation during transport caused further com-plication, and could account for the bulk of presentregolith. Impact basins were partially f i l led with hotdebris which became modified by metamorphism andvolcanism. Mantle material uplifted the basin fi l l , andlate mare basalts provided a cap. The ejecta blanketsare dominated by metamorphosed breccias rangingfrom poorly consolidated material to partly-meltedsamples. These complex assemblages were further

1062 J V. SMITH AND I, M. STEELE

modified by smaller impacts to produce a veneer ofregol i th .'

Most non-mare rocks are inhomogeneous, andsome controversy among lunar petrologists arosemerely because they had examined different sub-samples. Other controversies arose from different in-terpretations of the same complex texture. Thus somethin sections from rock 14310 show sporadic clots ofplagioclase whose irregular grain boundaries are con-sistent with solid-state annealing, whereas the greatbulk of the area is dominated by an undoubted ig-neous texture. After considerable dispute on frag-mentary evidence, most investigators now believethat 143 l0 was produced by near ly complete mel t ingof regolith in which only the clots and trace elementsprovide evidence of the earlier hybrid state (see pa-pers by Ridley et al., 1972, James, 1973, Gancarz eta l . ,1972, and many papers in the Proc.3rd Lunar Sci .Conf. for confl icting ideas: note that Crawford andHolfister, 1974, from the presence of Na-rich cli-nopyroxene and low-Al hypersthene, interpreted14310 as a product of mel t ing 200 km deep in theMoon). We accept the general opin ion in 14310, andsuspect that many lunar rocks with igneous texturesare the products of impact melting of complexbreccias.

Most Apol lo and Luna specimens are metamor-phosed breccias in which detailed mineralogical studyhas revealed extensive chemical and physical changesbetween mineral grains accidentally thrown together.A good idea of the complexity can be obtained fromthe.following papers: Albee et al., 1973; Anderson e/al., 1972: Bence et al., 1973, 1974; Cameron andFisher , 1975; Chao, 1973; Gr ieve et a l . , 1975: Kuratet al., 1974; Simonds et al., 1973, 1974:' Steele et al.,1972; Warner, 1972, Warner et ql., l9'74. All grada-tions occur from (a) monomict breccias with angularmineral clasts to (b) low-grade polymict breccias withopen pores and res idual g lass to (c) medium-gradepolymict breccias with partial recrystall ization of themineral clasts along with complete elimination of theglass to (d) high-grade breccias ranging from oneswith hornfelsic textures to ones with igneous textures.As the grade increases, the mineral compositions be-come homogenized, and information on the primaryconstituents becomes lost. Partial melting and vaportransport may affect the textures and bulk composi-t l ons .

Because so few even-textured rocks occur on theMoon, petrologists and geochemists were forced touse rock terms in a chemical rather than a textural ormineralogical sense; furthermore many acronyms

were invented and terrestrial terms were used in aloose or extended way. A glossary is given in pagesi-xi of the Proceedings of the Fifth Lunar ScienceConference. Many names are confusing, and we shalltry to use clear terms such as "polymict breccia withanorthosite bulk composition."

Apart from the mare basalts, which can containlarge amounts of i lmenite and significant amounts ofd iopside component , most lunar rock composi t ionscan be represented closely on the ternary plot ofsil ica-olivine-plagioclase. The liquidus relations forthe pure SiOr-MgrSiO.-CaAlzSizOs system at I at-mosphere dry conditions (Fig. I ) show crystall iza-tion fields for forstQrite, Mg,Al-spinel, anorthite,Mg-pyroxene, and tridymite or cristobalite (Ander-sen, l9l5). These crystall ization fields change litt lewith substitution of Fe for Mg and NaSi for CaAl(e.g. Walker et al., 1975), and a single generalizeddiagram is sufficient for consideration of melting oflunar highlands material at low pressure. Key featuresof the diagram are (a) incongruent melting of bulkcompositions in the region pqt to Mg,Al-spinel andMg,Al-depleted l iqu id; equi l ibr ium cool ing resul ts inreaction of spinel with the l iquid to produce olivineand plagioclase below the solidus (Fig. 2) for pressureless than Skbar; (b) incongruent mel t ing of pyroxeneto olivine and Si-rich l iquid; with increasing pressurethe incongruent melting is reduced, and is f inallyeliminated at several kilobars, (c) decreasing temper-atures along the cotectic curves qr and rs to the

- mg 0.71.0

- - T i - r rch moreboso l t

Mg- r icho l i v i n e

Ftc l . Crystal l izat ion f ie lds of SiOr-Mg-r ich ol iv ine-Ca-r ichplagioclase system for dry conditions at I atmosphere (Fe-free,

Andersen, l9l5; mg 0.7, Walker et al , 1973b) Temperatures forFe-free system See Lip in (1975) for abstract on data for mg 06Labelsp-w indicate reference points d iscussed in thetext H and Mrepresent possrble bulk composi t ions of the lunar crust and thewhole Moon

Co-r ichfe ld spor

PL46IOCL ASE

LUNAR MINERALOGY 1063

k m depth

grov i to l iono l p ressureFIc. 2 Some relevant phase equi l ibr ia for a Moon wi th o l iv ine,

pyroxene, and plagioclase as the dominant minerals See text foror ig in of curves.

eutectic s at 1222"C for the Fe-free system; (d)decreasing temperatures over the whole diagram asFe substitutes for Mg and NaSi for CaAl.

Figure 2 shows some relevant controls on the min-eralogy of the Moon. The present-day selenothermresults from interpretation of the electrical con-ductivity profi le deduced from perturbation of thesolar wind (e.9. Sonet t , 1975; Shankland, 1975). As-sumption that the mantle consists largely of Mg-richolivine essentially free of Fe3+ leads to a model-de-pendent estimate of the present temperature profi le(Dyal et al., 1974) using olivine data of Duba et al.

Q97a); assumption of Mg-rich pyroxene gives a sim-i lar prof i le (Heard et a l . ,1975). Al though the deta i lsare controversial, the general trend of the seleno-therm is consistent with seismic evidence for meltingbelow - l000km depth: if the molten l iquid is basal-tic, the required temperature for 1000 km depth is-1500' C which is consistent wi th the proposedcurve. The selenotherm is estimated for the entireearly Moon reaching the temperature for extractionof basaltic melt from an ultrabasic bulk compositionwhen account is taken of near-surface cooling overthe past 4 Gy. Complete melting of Mg-rich olivinewould require temperatures up to 1800-2000' C,which would exacerbate problems of heat sources.Figure 3 is a suggested mineralogical model of theMoon expressed as a density-depth profi le. Detailswil l be given elsewhere in a paper on the chemicalcomposition of the Moon. Complete differentiationof the Moon would result in (a) extraction of

Fe,Ni,S-rich l iquid to form a core, (b) gravitationals ink ing of Mg-r ich o l iv ine and minor amounts ofheavy complex oxides to form a mantle, and (c) de-velopment of a late l iquid rich in plagioclase andpyroxene components which would ult imately formthe complex crustal rocks. The Fe-rich core mightbecome a magnetic dynamo. The Fe-FeS eutectic isclose to 1000' C for all lunar depths, whereas themel t ing curve for i ron r ises f rom -1520 to -1650"

(Bret t , 1973; Usselman,1975). A S-r ich l iqu id wouldbe strongly superheated if the supposed selenothermis correct, thereby reducing the viscosity and easingfluid-dynamical problems of an early lunar dynamo.During progressive crystal-l iquid differentiation theFe content of the l iqu id would increase, as would theamount of the " incompat ib le" e lements (e.9. K,REE, Ba, P, Zr, and U) which enter the octahedraland tetrahedral sites of common sil icates only in tri-vial amounts. Late partial melting and volcanismwould confuse the situation greatly, as would theeffects of major and minor impacts. Figure 3 showshow the mean densi ty of the Moon could be ex-plained by this kind of petrological model. Even ifonly the outer part of the Moon melted, as proposedby many workers (e.g. Taylor, 1975), many of these

Solid

k bG P o

dens il yg/cn3 4.5

m0re

gronit-=--='ii*E 1500 t250 t000 500 0rod ius km ( r3 p lo t )

Flc 3. Mineralogical model of Moon expressed as densi ty pro-f i le The smal l d iagram shows the var iat ion of densi ty of o l iv ineand low-Ca pyroxene f rom 20oC and 0kbar to I 500"C and 38kbar.the lat ter of which are possib le condi t ions at - l00km depth. Apossib le lunar core might consist of e i ther Fe,Ni or Fe,S or amixture of both A mant le might consist dominant ly of o l iv ine andlow-Ca pyroxene. The crust is probably dominated by rocks r ich inpyroxene and plagioclase (point l / , F ig. I ) . Mare basal ts mayresul t f rom melt ing of cumulates including ones r ich in i lmeni te.Note that many smal l var iants would sat is fy the mean densi ty of3.343 gcm 3, inc luding a mant le whose ol iv ine was uni form at-Fo, , From Smith (1976).

deplh (km)

' \very rore- t i lmbnite - rich cumulote

,1t

Forr* Co,Al-Enuuz / rore mullrple 0xtdest /

Ty^^;i;1,., ---;;/'

/-oia,nlnsiie Foru

Fe,Ni7.5

FeS mellneor 5.0Densily ofolivine ondpyroxene

[-l Fo*________

705:-:::eo 51F-:gobi|

=-too'"

zo"c tsoo"cokb 38 kb

- 1000 kmdepth

4.0

features of crystal-l iquid fractionation would occur(e.g. Murthy et al., l97l).

Ignoring the possible Fe-rich core and the minorconstituents, the bulk composition of a completelydifferentiated Moon might approximate to point Mon Figure I, which l ies in the field outl ined by Walkeret al. (1973a). Olivine would be the l iquidus phasewhatever the depth in the Moon. Sinking of olivinewould result in the bulk composition of the remain-der moving away from the olivine corner. Dependingon the exact position of M and the extent ofcrystal-l iquid reequil ibration, the subsequent differentiationcould take different paths. Figure 2 shows phase rela-tions for a bulk composition with equal amounts offorsterite and anorthite (based largely on Kushiroand Yoder, 1966). At low pressure, anorthite andforsterite are the subsolidus phases. Anorthite dis-appears at the sol idus ( -1320' ) g iv ing spinel * for-s ter i te t l iqu id, then o l iv ine d isappears at -1380' Cleaving spinel * l iquid, with complete melting at1500-1600' C. In the subsol idus region, anor th i teand forsterite are replaced by anorthite, spinel, andtwo pyroxenes above about Skbar, and then by gar-net and pyroxene above about 15kbar. The details arecontroversial, but the general relations are sufficientfor the present purpose. Depending on the bulk com-position M and the temperature profi le of the Moon,garnet might form in the lunar mantle. Assuming thatthe temperature profi le of the early Moon is con-trolled by the "basalt" solidus, garnet should notform at depths less than -300km. Only half of thelunar volume is inside this depth, and olivine may bethe sole crystall izing sil icate there, thus eliminatinggarnet except perhaps for the products of trappedliquid. On the other hand, the l iquid may havereached a composition sufficiently rich in An mole-cule to yield some primary garnet. Further complica-tions can arise in the depth range 100-300 km be-cause of the possible assemblage anorthite * twopyroxenes * spinel.

For depths less than 100 km, crystall ization wouldbe controlled principally by the relations in Figure 1.Point f l on the olivine-plagioclase cotectic wouldmark the final disappearance of olivine when accountis taken of elimination of the incongruent melting ofpyroxene at several kilobars pressure (or of sub-solidus reaction). The lunar crust would be domi-nated by Mg-rich pyroxene and Ca-rich plagioclasewith a bulk composition near -F1. Subsequent differ-entiation would terminate at the eutectic s. Of course,this ideal theoretical situation could not occur, andmanv idiosvncratic variations must occur. It is our

J V. SMITH AND T. M STEELE

task to find out whether lunar rocks and mineralsgive clues on the applicabil ity or otherwise of thismodel. [Elsewhere we shall consider the validity ofmodels proposed by other workers based on bulkcompositions rich in pyroxene (Ringwood) or Ca,Al-rich refractory condensate (D. L. Anderson).]

A simple approach is to examine all the bulk analy-ses of lunar samples without taking account of themineralogical features. Wood (1975) separated marefrom highlands material by a plot of Ca/Al us. TiO,.Those specimens with CalAl (atomic) ) [0.786-O.229log TiO, (wt.7o)l were classified as mare ba-salts. The remaining samples were split into high-KREEP and low-KREEP var iet ies by the l ine KzO *P2Ou (wt.Vo : 0.945 -0.00893 (7o normative plagio-

clase). Figure 4 shows the resulting modal plots ofolivine-plagioclase-quartz for the two groups of pre-

sumed highlands samples. Many of the bulk analysesby broad-beam electron microprobe analysis are forsmall l i thic fragments, and both the sampling andanalytical errors are significant. Nevertheless the datashow interesting trends. The KREEP-rich specimenstend to l ie near the cotectic qr of Figure l, consistentwith proposals for origin from cotectic primary orsecondary melts, but the spread is so wide that otherorigins are indicated as well (e.9. hybridism). TheKREEP-poor specimens tend to l ie in the lower right,and this might be explained by derivation mainlyf rom p r imary co tec t i c l i qu ids i n t he ano r -thite-plagioclase-sil ica system, with a bias caused bythe tendency for plagioclase to rise to the lunar sur-face because of its lower density. A few samples arevery rich in olivine, and we shall focus attention onthem because they may derive from the deep crust orupper mantle.

The remainder of this section wil l concentrate onevaluation of the rock types in the lunar samples,using the chemistry of olivine, pyroxene and spinel(Figs. 5-12) as the principal mineralogic clues. Be-cause all lunar petrologists assume that the lunarhighlands are dominated by Ca-rich plagioclase andlow-Ca pyroxene as expressed by rock terms such ashighland basalt and norit ic anorthosite, we shall de-liberately focus attention on the mineralogy of con-troversial materials which may represent deep-seatedrocks. Study of rocks from large igneous complexesand granulite terrains on Earth, together with in-vestigation of xenoliths transported by kimberlitesand alkali basalts from the upper mantle, provides anexcellent basis for establishing mineralogical andchemical criteria. Under plutonic conditions, thegrain size is often coarse, and cumulate texture occurs

LUNAR MINERALOGY

HIGH KREEPHIGfILRNDS SRHPLES

1065

LOh KREEPHIGHLRNDS SBI1PLES

O L I V I N E

in some igneous complexes; however, granulationand recrystall ization can lead to a variety of texturesfrom porphyroclastic to equant (e.g. Mercier andNicolas, 1975; many papers in Phys. Chem. Earth,Vol. 9). Chemical features should provide easier cri-teria especially for lunar specimens in which impactmetamorphism is so common. The chemical featuresindicative of deep-seated origin include lack of zon-ing, presence of high-pressure minerals, chemical par-tit ioning indicative of high-pressure, coarse ex-solution and strong cation ordering indicative ofprolonged annealing, low content of certain minorefements (e.g. Ca in olivine), and chemical featuresinherited from a high-pressure ancestor. In addition,certain chemical features indicate early crystallizationfrom primitive l iquids of which the simplest isthe mgratio (: atomic MglMg+Fe) of the ferromagnesianminerals. Rocks composed of early crystals should below in the "incompatible elements" which tend toend up in granitic residua, and should have accessoryminerals rich in most transition metals of the firstseries (Ti-Ni).

PLOT OO?5

I,.IHOLE ROCKRNRLYS I S

L I T H I CFRHC,IIENT

FNORTH I TE

Because there are so few rocks with simple textureand chemistry, the lunar detective is forced to workwith complex rocks. As an example Figure 5 showstextural features of 67075, a polymict breccia withanorthosite bulk composition (Brown et al., 1973;Smith and Steele, 1974; McCal lum et a l . ,197Sb).Theoverall view (a) shows the angular clasts of plagio-

clase, pyroxene, and olivine, while the magnified por-tions (b-f) show special features described in thefigure legend. Particularly interesting is the coarseexsolution of pyroxene clasts (d-f), which indicatesprolonged subsolidus annealing (this is the coarsestrecorded in lunar pyroxene). Although individualpyroxene grains have uniform compositions (except,

of course, for the exsolution) the pyroxenes of differ-ent grains in the same thin-section range in composi-tion from EnnoFsuuWon to En.uFs.rWos and fromEnroFsrrWon, to EnorFsrnWonn while the olivinesrange from Fono to Fouu. The simplest explanation isthat 6'707 5 is composed of the compacted debris fromdifferent parts of a layered igneous complex whichunderwent prolonged annealing before disruption

,o,rt . FNORTH I TE

FIc. 4. Principal components of highlands samples selected from computer store. Wood (1975).

OURRTZ

1066 J. V. SMITH AND I, M STEELE

FIc. 5. Photomicrographs of 670'75, a polymict breccia wi thanorthosi te bulk composi t ion(a) Dark port ion at lef t consists of o l iv ine (-Fouo, f ractured) andplagioclase (unfractured) clasts set in plagioclase-rich matrix.L ight port ion shows var ious c lasts of p lagioclase grains plus alarger c last of polycrystal l ine micro-anorthosi te wi th smal l equantpyroxenes; minor o l iv ine c lasts ( -Fotu) a lso occur(b) Polycrystal l ine micro-anorthosi te c last , most ly wi th t r ip le junc-tions and polygonal texture Small pyroxene grains are hard torecognize Same texture as c last in (a) .(c) Polycrystal l ine c last dominated by shocked 2 mm grain ofanorth i te Matr ix shows at upper lef t , whi le other anorth i te grainswith interstitial pyroxene grains occur at the right of the clast.(d) Pyroxene clast showing coarse and fine exsolution lamellaeThe matr ix contains angular p lagioclase c lasts.(e) Clast composed of 2 mm pyroxene grain wi th coarse and f ineintergrowth (central region) and portion of large plagioclase (darkregion at lower lef t ) . The composi te c last is surrounded by matr ix .Matte areas are cracks f i l led wi th epoxy The pyroxene grain istraversed by a band of plagioclase grains with curved boundaries(gray and whi te grains running NNW through the center) . Thecentral part is enlarged in ( f ) .( f ) Central part of (e) . Spinel occurs as black inclusions (E-shapedat lower left and rod-shaped at upper and lower right). Pyroxeneexsolution lamellae occupy the left and right parts of the field ofv iew. The centra l one-th i rd consists of p lagioclase grains, the outerones of which tend to be scal loped whi le the inner ones tend to beel l ipsoidal . Simi lar p lagioclase grains occur as dark-gray ent i t ies atthe margin and just ins ide the lower edge of the pyroxene grain in(e) . The or ig in of th is p lagioclase is unknown but in ject ion ofplagioclase into fractures followed by prolonged annealing is onepossib i l i ty . Faint hor izontal l ines are blemishes on f i lm.

(a) (b) (c) and (e) 2 mm across, (d) I mm across, and ( f ) 0.5 mmacross. (a) and (c) plane-polarized light; others crossed or partly-crossed polars. (a) and (c) f rom 67075,2; others 67075,51

See McCaflum et al (1975b) for other photographs.

(Brown et a l . ,1973). l f 67075 had been st rongly meta-morphosed or melted, the above evidence would havebeen lost.

Lunar specimens with high Mg/Fe

All the lunar missions brought back soils andbreccias with rare grains of olivine whose highly mag-nesian composition (-Fono) is l ike that of olivinesfrom the Earth's upper mantle (e.g. Heiken, 1975).Steele and Smith (1972) located four Apollo 14 frag-ments with ultrabasic affinit ies, and many more havesince been discovered including some larger speci-mens with indicative textures. All these specimens aremore magnesian (mg -0.9) than the common material(bulk mg mostly < -0.75). We first describe somespecimens with particularly interesting textures andmineralogy (Table I and Figures 6, 7, and l2), andthen summarize pertinent mineralogical data forother Mg-rich materials.

(a) Selected specimens of special interest. Theclearest textural evidence for deep-seated processes isfrom 76535 (Gooley et al., 1974), a 155 g rake samplecomposed main ly of -5mm grains of p lagioc lase(587o, Anr6), olivine (37Va, Fo'6) and orthopyroxene(4Vo, Eng.). Bence et al. (1974) found similar 2-4mmfragments in Apollo l7 soils. The grain boundariesare curved and tend to meet at 1200, and the textureis that of a thoroughly annealed rock which has notbeen brecciated or shock-metamorphosed. Thepyroxene hae the space group P2,ca (Smyth, 1974a)which results from unusually effective segregation ofMg and Fe. Although the depth of crystall izationinferred from thermodynamic estimates is con-troversial (Albee et al., 1975), all authors agree that76535 was annealed in the solid state for a long timeat considerable depth: indeed the textural featurescan be matched in terrestrial granulite rocks. Inter-esting details are: (a) the plagioclase contains orien-ted incfusions of iron needles (5-15 wt.Vo Ni and up to5Vo Co), chromite, and an unidentif ied phase; (b) theolivine contains discrete chromite inclusions andsymplectic intergrowths of chromite and pyroxene;(c) symplectic intergrowths of orthopyroxene, diop-side and Al-Mg-chromite occur between olivine andplagioclase grains; (d) mosaic clusters of orthopyrox-ene, diopside, chromite, Fe-Co-Ni metal, troil i te,whitlockite, chlorapatite, baddeleyite, plagioclase,and K-feldspar occur sporadically, and are most eas-ily interpreted as the products of trapped primaryliquid augmented by exsolved material; (e) somemetaf particles consist of a-7 assemblages whose

LUNAR MINERALOGY

Trslr I Analyses of minerals in selected lunar peridoti tes

r067

o t n l 1 T2

s i o 2 3 9 ' 9 5 5 . 9 5 5 . 5 5 2 . 5 4 4 . 9 - 4 0 . 1 5 ! . 3 5 2 . 7 4 4 . 8 0 . 1 9 + 2 . 4 4 0 . 8

T i o 2 0 . 0 3 0 . 4 0 . 3 o . T o . o 2 a . g - r . 2 < 0 . 0 1 0 . 3 1 0 . 5 4 < o . 0 1 a . 4 9 0 , 0 < 0 . o 1

A L z o 3 - L . 5 1 ' o r . , 3 5 . 3 1 4 . 7 - 1 9 . 4 < o . 0 r l ' 0 7 2 . 7 3 3 5 . a 1 9 , 9 0 ' 0 o ' 2 2

c r 2 o 3 < 0 . 0 2 a . 5 0 , 5 a . 9 - 4 5 . 3 + 2 . 8 < o . o 1 0 . 8 4 o . B 5 - 4 8 . 3 0 . o l

F e o 1 2 . 0 7 . 4 7 . 8 2 . 8 o . 0 5 2 r . 3 - 2 4 . 4 r L . 9 8 . r 3 . o o 0 . 1 4 t B . 6 8 . 3 8 . 5

M n O 0 , t 0 . 1 0 . r 0 . 0 8 - 0 , 3 - 0 . 3 o . 1 r 0 . 1 1 o . 1 3 - A . 7 5

r , r s o 4 T . t 3 2 . T 3 3 . 2 r B . r - 8 . a - r r . p 4 8 . 1 3 3 . 7 1 8 . 5 a . 2 3 l l . 1 5 a . T 5 0 . 2

c a o < o . o 2 L . 5 o . 8 2 3 . 2 1 8 . 8 - o . 1 3 5 . 5 ' z r . i r 9 . 3 - 0 . 0 5 < 0 , 0 1

N a 2 o - 0 . 4 - - < 0 . o 1 o . o 5 a . 5 2

r o r a l 9 9 . 2 1 0 0 . 1 1 0 0 . 2 9 9 . 9 9 9 . 5 - 1 0 0 . 4 1 0 1 . 0 9 9 . 9 r o a . z 9 9 . 8 L a L . 5 9 9 . 8

t ?

s i0^

T i0^

4 .0^o-

Fe0

Mn0

Mgo

ca0

Na-0

14 L5 16 17

5 5 . a J 6 . 5 + 5 . 2 o . o0 . 7 8 0 . 1 5 - 0 . 0 35 . r 2 . 4 + 3 4 . 3 t T . 3o . 4 1 1 4 . 05 . 8 5 . 7 a o . 0 2 9 . 3

3 3 . 5 3 5 . 7 o . 0 2 2 a . \0 . 0 < 0 . 0 1 1 8 . 4 0 . 0

^ a ^

18 rg 2a 2L 22

l a ^- - t . t 5 3 . 2o . o r 9 8 . o 7 6 . 9 - o . 9

5 B . o o . o o 5 . 91 2 . 7 0 . 3 O . 5 O . 0 4 o . 4

9 . 4 0 . 1 9 . 6 9 . 8 5 . L

0 . 0 J . 0 9

2 0 . 4 < 0 . 1 1 3 , 1 4 9 . 3 3 3 . 1< 0 . 1 < o . l o . 0 4 0 . 4

2 ?

] l L q

5 " . +

0 . 0 5

0 . o l

1 9 . 8

O . 3 8

2 ) J 2 \

0 . 0 0 0 . 0 0

o . 0 4 t 7 . oA T T D Q E

8 . 6 0 . 4 38 . 9 2 8 . 9- 0 . 4 3

2L.4 11 .5

T o t a l 1 0 0 . 7 1 0 o , 5 9 9 . 7 1 O r . O l O O . 5 I O O . 1 1 O 0 . 1 1 O O , 5 1 0 0 . 1 l o c . O 1 0 0 . 0 1 0 1 . 2

L - 6 7 6 5 3 5 t r o c t o l i t i c g r a n u l i t e ( G o o l e y e t a l . r 1 9 7 4 ) , f o l i v i n e F o g 7 . 5 , 2 , 3 b r o n z i t e a s

s i n g l e g r a i n s E r g 6 . 2 F " l O . 9 t d o 2 . 9 r r d i n m o s a i c i n l e r g r o w t h s a n d s y m p l e c t i t e s E n g T , 2 F s 1 f . 5 W o f . 3 ,

4 d j o p s j d e i n s J - m p l e c t i l e s a n d n o s a i c s L n . ^ f s . W o , o . a n o r r h i L e A b " / A n - t r , 0 r ^ - .K2o o .09 r AZ- rvs-chromiLe ranse ro . a l r

' l ; I . , " ' j o r " i ' . \ -o . " , u r "o t r ^ . " i i l : J t ' .e -E ' .8 . t

t , 2 - 6 , L , t a e n i t e c o 1 . 8 - 4 . 2 I ' t i 1 8 , 0 - 3 3 , 0 ; 7 - l ] 7 2 4 1 5 b r e c c i a t e d d u n i t e ( A l b e e e L a 1 . , 1 9 7 4 b ) ,

7 o l i v ine FoBBr l l i o 0 .04 , 8 low-Ca pyroxene EngL ls l l l , i ogr t h ieh-Ca pyroxene En54Fs5 Wo4I ,

l o a n o r t h i t e A b r A n r 2 o r l o t h e r s 2 , K 2 O o . 0 9 B a O 0 , 0 4 , 1 1 A l - l ' { g - c h r o n i t e , V r O , A . 3 5 , Z r O 2 o . A B ,

a l s o F e - m e t a l C o f . 3 - 2 . 2 N i 2 4 . 5 - 3 I . 8 t ' I g 1 , 3 C r o . 3 5 S i 0 . 0 4 ; 1 2 - 2 0 b r e c c i a t e d p e r i d o t i t e

c l a s t s i n 1 5 4 4 5 b r e c c i a ( A n d e r s o n , 1 9 7 3 ; R i d r e y e L a L . , L 9 7 3 ) L 2 o l i v i n e ( A ) a f s o N i O 0 , 0 3 ,

1 3 o l i v i n e ( R ) , f 4 o r t h o p y r o x e n e ( A ) , 1 5 o r t h o p y r o x e n e ( R ) 1 6 a n o r t h i t e ( A ) a l s o K 2 0 0 . 1 2 ,

1 7 C r - s p i n e l ( A ) , f 8 c r - s p i n e l ( i l ) m i d d l e o f f o u r a n a l y s e s j 1 9 r u t i l e ( e ) a f s o N b Z a 5 I . 6 ,

2 0 a r n a l c o L i l e ( A ) ; 2 - - 2 5 b r e c c i a r e d p e " i d o ' i ' e ' r a g m e n ' s i n 3 2 4 3 s o i l ( B e n c e e L a l . ,

1 9 . 1 \ , 2 1 o l i v i n e ( m e a n o l - o , 2 2 p y - o * " n e ( m e a n o f l j r s r I a n a l l s e s ) . 2 3 a n o r L l i i e ( m e a n o f

2 , K A A O . A T ) , 2 4 s p i n e l ( m e a n o f 2 ) , 2 5 i l n e n i t e .

composition indicates a cooling rate consistent withsome tens of kilometers below the lunar surface.

The origin of the chromite-pyroxene symplectitesis controversial. Gooley et al. (1974) suggested thatthey formed by a reaction Ol + An * chromite -Opx f Cpx * Al-Mg-chromite, whereas Albeeet a l . (1975) argued for co-precip i ta t ion dur ingmagmat ic crysta l l izat ion. Bel l and Mao (1975) pro-posed tha t s im i l a r symp lec t i t es i n12415 b recc ia teddun i t e and i n seve ra l l una r o l i v i nes resu l t ed f romsol id-state breakdown of pr imary garnet t rappedin o l iv ine. The garnet would be very Cr-r ich(-Ca' rMgo rFeo rCr, rAlo. rSi .O, , for 72415 duni te) .Chromite forms vermicular intersrowths with several

sil icates in terrestrial peridotite xenoliths (Basu andMacGregor, 1975; Dawson and Smith, 1975), andexsolution, especially from pyroxene, combined withsolid-state recrystall ization provides a plausible butnot definit ive explanation of the terrestrial occur-rences. Perhaps chromite tends to intergrow with sil i-cate under deep-seated metamorphic conditionswhatever the source of the sil icate; if so, the origin ofthe lunar symplectites wil l be hard to discover.

Alf large grains in 7 6535 have an inclusion-free rimwhich results from diffusion of exsolved material tothe grain boundaries. Whatever the details of theannealing mechanisms, we endorse the mineralogicimplications that 76535 is the product of prolonged

1068 J. V. SMITH AND I, M. STEELE

ST TzO

7o q. t -

- l

Sp ine lTrocto I i t ic

t62295

3Sp inel-A north itt

Opx -0 l i v i neCotoc lo s i te

(I

T'oc to l i t ic- + 7 6 5 3 5 ' I

, , t at l

t09e9 - ?

o -. + _' 1

_ *72415l l

D un itic

t00 90 80 70 60nol. % Fo

FIc. 6. Composi t ions of o l iv ine in selected samples, many ofmg-r ich composi t ion.

Oliuine-rich (" dunitic" )

l . Soi l f ragmenl73242,4,Al4: o l iv ine breccia wi th minor p lagio-c lase and spinel . One c last of p ink spinel and plagioclase (Steeleand Smith, 1975b). 2. Dunite clast in 14068,10 breccia: polyg-ranular mosaic of olivine For" with single chromite grain Helz(1972). 3. Dunite clasts 72415-18: see text (Albee et al., 1914b). 4.Soil fragment 78422,2,A18: olivine breccia, I mm across, withminor plagioclase, chromite and armalcolite (Steele and Smith,1975b). 5. Soi l f ragment 14002,7,E- l -8: shocked s ingle o l iv inegrain wi th minor chromite (Steele and Smith, 1972).6. Soi l f rag-ment 78502, 17,A34: o l iv ine breccia wi th minor chromite, armalco-l i te, and plagioclase (Steele and Smith, 1975b). 7. Soi l f ragment78442,2,A15: o l iv ine breccia, I mm across, wi th minor chromite,armalcol i te, d iopside, and plagioclase (Steele and Smith, 1975b). 8Soil fragment 78442,2,A16: olivine breccia, I mm across, withminor p lagioclase and armalcol i te (Steele and Smith, 1975b) 9.Breccia 14063,14. Fine-grained clast of 7070 olivine, 30Vo plagio-c lase An." , and minor bronzi te and chromite (Steele and Smith,1972). 10. Breccia clast 78502,17,815: olivine breccia with plagio-

c lase c last and minor chromite wi th TiOu ( l7o. Unpubl ished. I l .

Polycrystalline inclusion, lmm across, in mare basalt 74275 (Meyer

and Wilshire, 1974). 12. Dunitic fragment in polymict breccia

14318 Ol iv ine Fouz " ,

wi th average For" . Minor p lagioclase

An". ".

with cluster at Anr, (Kurat et al., 1974).

Oliuine-plagioclase-rich (" troctolilic" )

I Clast in rake sample 15335. " lgneous" texture (Steele et a l ,1972\. 2. Troctol i t ic granul i te 76535 and 2-4 mm f ines 76503. SeeTable I and text . 3 Clast in breccia 14321. Granular assemblage of

olivine, anorthite Anrr, and accessory K-feldspar (Compston et al ,1972). 4. Several 2-3 mm fragments in breccia 14320,4. Oli-

vine-plagioclase An"u intergrowth with texture similar to cres-

cumulate of layered ultrabasic rocks of Rhum (Brown et al., 1972).5 Fragment, I mm across, in l -2 mm f ines 14166,6 70Vo plagio-

c l a s e A n , u , 3 0 % o l i v i n e , 2 % d i o p s i d e , 8 t i n y g r a i n s

Cr-Zr-armalcolite (Steele and Smith, 1972). 6. Fine-grained an-nealed clast in recrystallized noritic breccia fragment 76503,6,21.

Plagioclase Anr. , o l iv ine, minor or thopyroxene, and augi te (Bence

et a l . , 1974).7. Troctol i t ic breccia 77017. Ol iv ine Fo" ' and plagio-

clase An"u, patches in feldspathic breccia Interpreted as xenolith inmare basal t (Brown el a l . , 1974).8. Clasts in st rat i f ied Boulder l ,

Stat ion 2, Apol lo l7 Most have f ractured ol iv ines in granulated

groundmass of o l iv ine and plagioclase, but some have basal t ictexture Plagioclase is An.. ,uOro , r n with An usually over 8870(Stoeser et a l . ,1974) 9. Clasts in st rat i f ied Boulder l , Stat ion 2,Apol lo 17. Granulated ANT (Stoeser et a l . , 1974).10. Luna 20 soi lf ragments whose bulk composi t ion is nor i t ic or t roctol i t ic anortho-s i te (Pr inz et a l , 1973a). l l . d i t to for anorthosi t ic nor i te andtroctol i te composi t ions. For specimens 8- l l , the analyses are dis-played as histograms. 12. "Basal t " f ragment 14162,41-1407- l l30% skeletal o l iv ine For. microphenocrysts, l07o Na,Al ,Cr-r ichpyroxenes En.oFsuWo.. and EnurFsrrWo2T5, 5070 plagioclase and

l0% glassy mesostasis (Powel l and Weiblen, 1972).

S p ine l-ano r t hi t e -o rt ho p y ro xe ne-ol iu i ne c at ac lasit e

L 15445,10 c last in breccia (Anderson, 1973).2. 15445 Type B

clasts in breccia (Ridley et a l , 1973) 3.73263,1,11 2-4 mm soi lfragment (Bence el al , 1914). See Table I and text for details.

Spinel troctolitic

1.62295 rock. lVo xenocrysts of o l iv ine Forr-ro, anorth i te

Ao"" "u,

p ink spinel (mg O.75, 3 5 mol .7o CrzO.) and Fe-Ni-Co

meLal.99Vo fine-scale intergrowth of olivine Fo"r ,", plagioclase

Arr"u " , ,

spinel (zg 0.9-0.76, 2 mol9o CrzO"), and glassy residuum(Walker et al., 1973b).2. 62295 rock Olivine phenocrysts Forr-".with overgrowths For. ,, set in groundmass of anorthiteAn"oAb"rOro. skeleta l o l iv ine For"-r r , spinel , i lmeni te, e lc (Agrel l

et al., 19'73). 3. 62295 rock. Olivine (xenocrysts Fo".-.", inter-growth Fo., .r), plagioclase (xenocrysts An".-"r, intergrowths andquench grains Anro

"r ) , spinel (xenocrysts 9-16 mol 7o chromite,

intergrowth 2-4 mol % chromite) (Hodges and Kushiro, 1973). See

also Brown et a l (1973)and Weiblen and Roedder(1973) 4.4 mmclast 67435-14 in breccia. Cumulus of o l iv ine Fo" ' , , , o and spinelpoikilitically enclosed in plagioclase AD6 u ,, o Prinz et al., 197lb).

5 Luna 20 plagioclase Anrr,rr-olivine Fo"-spinel fragment (Tara-

sov et al., 19'73). 6 I mm fragment 14321,76 of olivine -Fo"u,

anorthite An"n, and spinel in breccia (Steele, 1972).7. Rake sample

65785 with coarse-grained cenler (65Vo An"r,30Vo Fo"o-"", spinelzoned outwards f rom 2.6 to l2.6Vo CrzOs and FeO l0 to l2%). Rare

pyroxenes, i lmeni te, Ni- i ron, t ro i l i te, Zr-rut i le , whi t locki te,

Cr-Zr-REE armalcolite, K-feldspar, chromite, and farringtonite(Dowty et al., 1974b).8. Thirty-four fragments of spinel troctolite


metamorphism at considerable depth (tens of kilome-ters?), probably of an igneous cumulate, in whichsubtle effects resulted from exsolution, diffusion andreduction. Haskin et al. (1974) concluded from analy-ses of REE, K, Rb, Sr, and Ba that 76535 was origi-nally an olivine-plagioclase cumulate containing8-16 percent parent l iquid with no appreciable Euanomaly and about 20 times chondrit ic abundance ofREE. The age data are inconclusive (reviewed byHaskin et al., 1974, p. l2l4), but original crystall iza-tion at -4.3 Gy is indicated.

Five olivine-rich fragments 72415-72418 were col-lected from a l0 X 20cm clast of metaclastic breccia72435. Olivine Foru-r, accounts for 9l percent of72415 as crystals up to one cm set in a granularmatrix of olivine with the same composition (Albee eral., 1974b). Some large olivines contain strain bandsmarked by small recrystall ized olivines 50pm across.Plagioclase Anrr-rr, Cr-spinel, high- and low-Capyroxene and metal occur as clasts and as inclusionsin the olivine, sometimes as symplectic intergrowths.The compositions of the two pyroxenes are closerthan for those in 76535 (Table l, Fig. 7) and the low-Ca pyroxene may be a clinopyroxene (EnrrFsrrWor).

in Luna 20 sample wi th p lagioclase Anr,- r , mean An"" , o l iv ineForr , ro mean Fo"n, low-Ca pyroxene Enrr-r rFsrr- r ,Wor,r , andspinel rzg 0. I -0 5 cr 0-0.2 (Prinz et al., 1973a).9. Fragment 100 I 9-22. Plagioclase An"u. , o l iv ine For, , spinel and glass (Kei l et a l . ,1970). 10. I mm clast l43l9, l la-3, Plagioclase An"u, o l iv ine Fo, .and spinel mg 0.54 cr 0.027 TiO,03-1.0 wt.7o (Roedder andWeiblen, 1972a). l l . 0.5 mm breccia c last 14303,508. Ol iv ineFo"" u, p lagioclase An,o and spinel zg 0 55 cr 0.068 TiO, l . l wt .%(Roedder and Weiblen, 1972a). 12. Type A spinel-brear ing l i th icf ragments in Apol lo 16 aod 17 soi ls . Ol iv ine euhedral to subhedraland plagioclase as later laths. Trace metal, troilite (?), mesostasis,and rutile(?). No pyroxene (Weiblen et al , 1974). 13. Type Bspinel-bear ing l i th ic f ragments in Apol lo l6 and l7 soi ls . Ol iv ineinterstitial to plagioclase. Minor spinel, trace metal, troilite (?),ilmenite and l-8Vo mesostasis (Weiblen el a/., 1974). 14."Spinel-o l iv ine anorthosi te" 60618, l -1. Rake sample. Photo-micrograph shows 4 mm shocked plagioclase grain Anr" with at-tached i r regular grains of p lagioclase, minor o l iv ine, Mg-Alspinel ,and pyroxene. Pyroxene also as veins in the large feldspar grain.Assigned to the Mg-troctolite group because of mineral composi-t ions (Dowty et a l . , 1974d). 15. One Apol lo l5 and two Luna 20fragments. Data for spinels and one ol iv ine (Reid, 1972) 16Histograms of olivine compositions for spinel troctolite fragmentsin Luna 20 and Apollo 16 (Taylor et al , 1973). l'7. 2-4 mmfragments in Al6 soils, denoted Type A Feldspathic IntersertalIgneous Rocks. Quench, diabasic, and poikilitic textures, with rareevidence of thermal metamorphism. An*-*, Fo 2 74, pyroxene mg- 0.2, Mg-Al spinel . Only other mineral data is h istogram ofol iv ine composi t ions (Delano et a l . , 1973). 18. Luna 20 f ragmentwith very fine-grained igneous texture. 6570 ADr"-,00, l5Vo Fo"o,2VoMg-Al spinel and -20Vo interstitial phase. Interpreted as impactmeft which lost volatiles (Dowty et al., 1973a).

m0lo/lo o dunitic

o t rocto l i i ictr sp-0n-opx-01

cotoc lositex spinel - t rocto l i t ic+ 60618, l - |

Frc 7. Compositions of pyroxene in selected samples, many of

Mg-rich compositions. Numbered specimens can be identified

from the appropr iate sub-group of specimens l is ted in Fig. 6.

Specimen 60618,t - l was descr ibed as a spinel-o l iv ine-anorthosi te,

and is l is ted as No. l4 in the spinel- t roctol i t ic sub-group. Other

data for the spinel-troctolitic group are lumped together without

numbered identification N and TA and AN and T are items l0

and I I of the t roctol i t ic group

This evidence plus the relatively high CaO content of

the olivine (0. l3 wt.7o) suggests a higher temperatureof equil ibration for 72415 than 76535. The 72415

brecciated dunite is particularly important because of

the tentatiue conclusion from the Rb-Sr isotopic data

that crystall ization occurred very early (- - 4.6 Gy).Polymict breccia 15445 contains white clasts un-

usually rich in olivine, spinel, and pyroxene (Ander-

son, 1973; Ridley et al., 1973) which were calledperidotite. Bence el al (1974) found similar minerals

in 2-4mm fragments in 73263 soil and incorrectly

stated that they represented a new rock type, spinel

cataclasite, not found in other missions. Although the

textures were badly altered by shock, some coarse-grained relics (- I mm) of unbrecciated olivine andCr-spinel indicate an original coarse texture. The

chemical analyses of the minerals differ considerablyin detail (Table l), but the general features are quite

similar including the very low CaO and CrzOa in the

olivines, low CaO and high but variable AlrO, in the

low-Ca pyroxenes, and the low TiO, and moderateCrrO, in the spinels. Whereas Anderson reported mi-

croprobe analyses of ruti le (?) and low-Zr "armalco-lite" (?), Bence et al. reported high-Mg ilmenite. TwoFe-rich pyroxenes in73263 were ignored because theycannot be in equil ibrium with the other minerals, andpresumably represent erratic grains incorporatedduring brecciation. The mg ratios of all the otherolivines and pyroxenes are mostly 0.90-0.92. Ander-son (1973) concluded that the 15445 clasts were ig-neous cumulates formed at moderate pressure (e.9.

2kbar) followed by partial recrystall ization duringgranulation at 950 + 50'C and P > 1.3 + 0.5 kbar.Bence e t a l . a rgued tha t t he assemb lage o f

40OU80En

-.a-N ond TA

1070 J V. SMITH AND I. M. STEELE

Al-enstatite, forsterit ic olivine, pleonaste, and anor-thite is stable from 3 to 5kbar. Whatever the details,t he ove ra l l ev idence sugges ts t ha t t hesespinel-plagioclase pe ridotit ic clasts originated atsome tens of kilometers depth. Anderson made inter-esting speculations about the opaque minerals andthe minor element content of the sil icates, and theseideas should be tested by thorough study of all speci-mens in this group: in particular the relations be-tween armalcolite and ilmenite need further ex-planation in view of the stabil ity data obtained byLindsley et al. (1974a). The pattern of rare earths in a15445 clast was found to be inconsistent with coexis-tence of olivine, orthopyroxene, and Cr-pleonastewith any known basaltic lunar l iquid, and Ridley elal. (1973) suggested that garnet had been present byanalogy with the pattern found for pyropic garnet interrestrial peridotites. Further work is needed, butone should ponder the possible presence of garnetperidotite and eclogite on the Moon. For Mg-richbulk compositions, pressures over l5kbar and depthsover 300km would be needed, assuming a temper-ature over 1000'C (Fig. 2) .

(b) Others. From the viewpoint of bulk composi-tion, other Mg-rich specimens could be classifiedmostly as dunitic, troctolit ic, and spinel-troctolit ic.Figures 6, 7, and l2 show major elements of olivine,pyroxene, and spinel (s./.) from these types.

Dunitic: Polycrystall ine olivine-rich fragments oc-cur in breccias and soils, and the legend to Figure 6lists all the specimens for which chemical composi-tions and some textural data were given. All are fromApollo 14 and 17. Specimens 1-8 have olivrne com-positions in the range Fo.u to Forr, suggesting originfrom a near-uniform source. Chromite occurs innearly all these dunitic fragments, and the composi-t ions fa l l in a s ingle region on a p lot of Mg,Al ,Cr , andFe (Fig. l2); furthermore the TiO, contents are below4 weight percent, which distinguishes them from mostchromites of mare basalts. Armalcolite occurs in sev-eral fragments, and its mg value ranges from 0.6 to0.7 and correlates with mg of the coexisting olivinealong the t rend of Steele (1974). Specimens 10, l l ,and 12, and perhaps 9, have olivine compositionswhich range to lower Fo values.

Troctolit ic: Polycrystall ine fragments rich in pla-gioclase and olivine occur in breccias and soils, andthe legend to Figure 6 l ists all specimens with pub-lished chemical compositions and textural data.There are two principal types, one with non-brec-ciated textures and olivine compositions restricted toForu-., (specimens l-6, 12) and the remainder with a

variety of textures (described as fine-grained basaltic,annealed cataclastic, metamorphosed, brecciated,cataclastic, and altered) and a wide range of olivineand pyroxene compositions. One aberrant specimenwith Fo., (No. 7) could be interpreted as a cognatexenolith in a mare basalt. The large troctolit ic gran-ulite 76535 has the same olivine composition as thefirst type, and indeed might serve as a guide forinterpretation of all the first group. The second groupof fragments can be interpreted as the products ofimpact mixing of diverse materials near the lunarsurface. Whereas the CaO content of olivines of thefirst group is mostly below 0.I weight percent sugges-tive of plutonic crystall ization, that of the secondgroup is mostly above 0. I weight percent (e.g. Prinzet ql., 1973a, Fig. l2) suggestive of near-surface con-ditions. The first group was found only in Apollo 14and Apollo l7 samples, whereas the second groupwas strongly represented in Luna 20 samples as well.Tentatively we define a group of Mg-troctolites withunzoned o l iv ines in the range 85-91 Fo and a groupof Fe-troctolites with diverse textures and olivinesless magnesian than Forr. The Apollo l7 fragmentsdescribed by Stoeser et al. (1974) show a bimodald is t r ibut ion on the h is togram of l ine 8 in F igure 6,and may represent a mixture of the proposed twotypes of troctolites. Further study, especially of mi-nor e lements, is desi rable. L ines 9, 10, and 1 l ofFigure 6 are histograms for specimens described as"anorthosite," and "ANT." Without further study oftextures, mineralogy, and minor elements, it is notpossible to interpret these specimens in detail, but wesuspect that most of them are polymict brecciaswhich have undergone various degrees of meta-morphism. Preferring to rely on chemistry ratherthan on texture and mode, we shall assume that speci-mens 8, 9, 10, and ll are mixtures of several types ofmater ia l .

Spinel-troctolit ic: The olivine compositions inFigure 6 are widely scattered from Fo' to Foro, withthe majority from Foro to Foru. Pyroxenes are vol-umetrically unimportant and tend to scatter widely incomposition (crosses in Fig. 7). Spinels are rich in Mgand Al and low in Fe and Cr; thei r p ink color isdiagnostic. Rock textures are often igneous, with adistinction between early megacrysts and lategroundmass. Most, if not all, of this group can beexplained by melting of mixtures of olivine andplagioclase whose bulk composition l ies in the areapqt of Fig. 1, followed by cooling sufficiently rapid toallow preservation of the spinel. Impact melting ofregolith and partial melting of olivine-plagioclase

cumulates near the lunar surface are the two favoredmechanisms (e.g. Agrell et al., 1973; Hodges andKushiro, 1973; Walker et a l . , 1973b). The extensiveliterature on specimen 62295 provides an excellentguide to the evidence and ideas. In this specimen aretwo generations of olivine: an early generation vari-ously described as xenocrysts or phenocrysts withcomposition Fo.n_ru and a later generation of micro-crystals often skeletal and with composition Foro_rr.At least some of the spread of olivine compositions inFigure 6 results from a contrast between early Mg-rich and late Fe-rich crystals, but the total spread isso large that diverse source materials must be in-volved. Until a thorough study has been made oftextures and mineral compositions of all members ofthis group, we feel that interpretation is diff icult. Onthe whole, we suspect that some are impact melts,while some are the products of rapidly-crystall izedmagmas from the lunar interior. Perhaps the Fe-richones tend to fall into the first group while the Mg-richones mostly or entirely fall into the second group.

Lunar specimens with lower MglFe

In part I, a crude distinction was made betweenANT, KREEP, and mare-type specimens. After1973, lunar petrologists became increasingly dis-satisfied with these simple distinctions, though it wasdiff icult to establish precise criteria for a better classi-f ication. The following attempt to sort out the miner-alogical evidence is rather subjective.

(a) High-KREEP material, mostly noritic. Essen-tially all lunar rocks have detectable KREEP com-ponents, and the key feature is whether the KREEPcontent is high or low. Unfortunately the KREEPcomponents (and related ones including Zr and Ba)occur mainly in the interstit ial and late phases whichare very fine-grained and diff icult to characterize. lnsmall and coarse-grained specimens, non-randomsampling is a problem. Some bulk analyses were madeby broad-beam electron microprobe techniques forwhich the accuracy of K and P was marginal. Bearingall these factors in mind, it is not certain whether theconcentrations of KREEP components show a bi-modal or a continuous distribution. The plot ofweight percent (KrO + P2Ou) us. weight percentp lagioc lase by Wood (1975, F ig. 3) shows a cont inu-ous distribution, though there is a strong tendencyfor the population density to drop abruptly as theKREEP content increases past the l ine (KrO + PrOd): 0.945-0.00893 (norm. 7o p lagioc lase) . Orbi ta ldatashow a marked concentration of radioactivity nearthe Imbrium region, but unti l greater areal coverage

l 07 t

Frc. 8 Composi t ion ranges of o l iv ines and pyroxenes tn

KREEP-r ich rocks. The shading and st ippl ing are qual i tat ive

indicat ions of the f requency of analyses of pyroxenes; by far the

major i ty occur in the dark region. Data sources: Apol lo l2 gray-

mott led breccias (Anderson and Smith, l97l) ; Apol lo l2 fe ldspa-

th ic nor i tes (Brown el a l . , l97l ) ;12032,45 l i th ic f ragments (Quaide

et a l . , 1971); nor i t ic c lasts in 14313 breccia (Floran et a l . , 19 '12);

14053 and 14073 basal ts (Gancarz et a l . , l97l ) ; 14276 basal t (Gan-

carz et a l , 1972\:67'749 basal t c last (Steele and Srni th, 1973);4-10mm Apol lo l5 f ines wi th basal t ic texture (Powel l et a l . , 1973);

basal t c lasts in 15465 breccia (Cameron and Delano, 1973); Apol lo

l2 coarse f ines, Type A crystal l ine nor i te-anorthosi tes (Wood el

a l , l 9 ' 7 l \ :Apo l l o l 2 KREEP f r agmen ts (Meye r e l a / . , l 97 l ) . No te

that many other data are avai lable.

is obtained it is not certain what is the Surfacedist r ibut ion of KREEP-r ich mater ia ls . Textura l ly ,KREEP-rich materials range from glasses to brecciasto rare basalts. Most crystall ine specimens have amineralogy dominated by plagioclase and low-Capyroxene (commonly orthopyroxene, but pigeoniteoccurs), and can be called norit ic. Figure 8 summa-rizes the composition range of pyroxenes obtainedfrom the specimens listed in the legend. Compilationof erratic data from diverse sources cannot give astatistically meaningful frequency distribution. How-ever, when account is taken of the tendency to l istu n u s u a l c o m p o s i t i o n s , m o s t p y r o x e n e s f r o mKREEP-rich norit ic material l ie in the rangeEn.o-roFsr.-ruWor-u. A fairly strong concentrationband extends to augites - EnnrFs'uWono, while a fewscattered analyses extend to Fe-rich compositions.Olivines are uncommon and mostly l ie in the rangeFo* to Forr. Plagioclases tend to be sodic comparedto other lunar specimens, with a range mostly fromAnnoAbro to AnruAbru (Part I , F ig. 2) .

(b) Low-KREEP noritic material. The legend toFigure 9 summarizes selected specimens. ThreeApol lo l7 specimens (Civet Cat c last in72255, rock78235, and orthopyroxene megacrysts in soils) haveminerals with similar compositions: plagioclaseAtnr-ro, orthopyroxene Enrr-rrWo -r, and augiteEno, nrWo..-n.. All are coarse-grained, and norite

LUNAR MINERALOGY

- - . . * - - i \

' . i. !:. i

!_ "______ . j

Fo2 0 0

1072 J. V. SMITH AND I M. STEELE

to Fs

to Fo

Ftc 9. Composi t ions of pyroxene, o l iv ine, and plagioclase inlow-KREEP nor i te and 67075 anorthosi t ic breccia Detai ls of spec-imens : l . C i ve t Ca t c l as t , 2 . 5 cm 1ong ,72255 , Bou lde r l , S ta t i on 2 ,South Massif, Apollo 17. 35Vo light streaks of pure plagioclaseAnr, ,nOro r- , o part ly shocked to maskelyni te. 6590 dark streaks ofor thopyroxene Enr, ,oFsrr_rrWor_o wi th abundant i lmeni te p latesin c leavages, rare augi te Ena2_nuFs",roWoq6_a6 ?.S smal l grains andlamel lae in opx, accessory cr is tobal i te, baddeleyi te, i lmeni te, chro-mite, metal l ic i ron, t ro i l i te, and Nb-rut i le (Stoeser et a l . , 1974) 2.Nor i te rock 78235 wi th l -5 mm plagioclase and or thopyroxene,chipped f rom boulder, Stat ion 8, Apol lo l7 Cumulate texture.Heav i l y shocked 60Vo ano r t h i t e Ah* u_ "n u , 30VoEnr" ,oFs,r- roWo, , , l }Vo glassy veins, interst i t ia l augi teEnnrFs"Wonu, s i l ica, whi t locki te (MgO 3.6 CezOg 2.0 wt %), f luo-rapat i te, rut i le (Cr,O3 3.0 and Nb?), i ron (Co 3 Ni 2) , chromite,baddeleyi te (TiO,4), and t ro i l i te (McCal lum et a l . , l975a; l rv ingeta l . , 1 9 7 4 ) . 3 . S i x l - 2 m m o r t h o p y r o x e n e c r y s t a l sEnr"- . ,Fs, . -2oWor(AlrO. l . l CrrO, 0.6 wt.Vo) wi th minor at tachedplagioclase Anrr rn, (Fe 0.1 K 0.1 wt.7o), mesostasis, and veins ofs i l ica mineral , K-Ba fe ldspar, phosphate, chromite, d iopsideEnn"Fsn-uWo."-c, t ro i l i te, and i ron (Ni 3.9 Co 2.8) . Probablyrefated to 78235 ( l rv ing et a l . , 1974).4. Micronor i te f ragments in| 432 I polymict breccia. Granoblast ic , hornfels ic and vest ig ia l sub-ophi t ic textures Unzoned pyroxenes wi th tota l composi t ion rangeEn.rFs, ,Wo, to Enu.FsrrWon (TiOr 0 5 Alros 0.6 CrrO.0 3) , p lagio-c lase inclusions Anru-ru, rare high-Ca pyroxene, i lmeni te, i ron,z i rcon and whi t locki te (Gr ieve et a l . , 1975) 5. Nine recrystal l izednor i t ic f ragments wi th granular pyroxene occurr ing as c lasts inmicrobreccias of Apol lo l5 soi ls . Interpreted as monomictbreccias Plagioclase Ahru

"u, pyroxene Enuo-"oFsru ,"Wor-rr, oli-

vine Fouu-ss (Cameron et al., 1973').

78235 has a definite cumulus texture. An obvioussource would be a disrupted plutonic complex. Therange of mg of the pyroxenes is close to that forpyroxenes in spinel troctolites, but the compositionsof the two pyroxenes are more highly separated asbefits a lower temperature of equil ibration. The or-thopyroxenes fall to the extreme Mg-rich end of thecompositional maximum for KREEP pyroxenes (Fig.8). The augite is rare, and occurs as discrete grains

and as lamellae in the Civet Cat clast, as interstit ialgrains in 78235, and as a component of rare KREEP-rich veins in the orthopyroxene megacrysts fromsoils. Although the bulk KREEP content is low, thereare KREEP-rich spots or veins in the specimens.Perhaps these result from permeation of the noritesby minor amounts of KREEP-rich fluids, or erraticconcentration of residual l iquids.

Specimens 5 and 6 in the legend to Figure 9 areclasts in breccias, and the wide range of plagioclasecompositions suggests a diverse origin.

(c) Anorthositic. Petrographically, anorthositesshould have more than 90 percent plagioclase, but theterms anorthosite and anorthositic have been usedloosely. Few photomicrographs have been publishedof anorthositic specimens, but we suspect that mostare polymict breccias. The 19 specimens listed in thelegend to Figure l0 probably show the entire range oftextures and compositions. Many specimens crudelylumped into the ANT category are rich in feldspar,

Frc. 10. Composi t ions of pyroxene, o l iv ine and plagioclase inanorthosi t ic specimens l . 2 mm fragments of microanorthosi te10085-4-10a. Microphenocrysts of anorth i te An", u in granul i t ic

mosaic of anorth i te, o l iv ine For, , r r , c l inopyroxene, i lmeni te (9 6wt.7o MgO), rut i le , and t race of t ro i l i te and i ron (Agrel l et a l . ,1970) 2 Anorthosi t ic basal t f ragment 10059-27. Anorth i te Ane6 4,m i n o r o l i v i n e F o r r , , a u g i t e E n o u

" F s r u r W o r " r , p i g e o n i t e

Enu, nFsrorWo.r , and i lmeni te (6 17" MgO) (Kei l e l a l . , 1970).3.Anorthosi te f ragment 10085-07,6. Anorth i te Anrr * and or-thopyroxene En"" oFsru ,Won . wi th lamel lae of c l inopyroxeneEnoroFs, .uWoruo (Reid et a l . , 1970). 4. Anorthosi te f ragment

14258,28,1419-7 Bytownite Anr" oOr, ,, I7o diopside Enn, uFsu uWonu o as inclusions and interst i t ia l grains, and l lVoi lmeni te (8.2 wL.Vo MgO) (Powel l and Weiblen, 1972) 5 FragmentD of brecciated anorthosite in 14257,2 coarse fines AnorthiteAD"or-rur , minor o l iv ine Fo"o, p igeoni teEnuo uuFsrr , roWo, , , andtrace ilmenite (Klein and Drake, 1972) 6. 23 anorthositic frag-ments of 67712 soi ls . Anorth i te An**r , o l iv ine FoTa*r , and pyrox-

enes wi th narrow composi t ion ranges at Eno.FsroWon, andEnrrFs,rWo, (Simkin et a l . ,1973).7. Cataclast ic anorthosi te 61016.

At

BO5 J

60 40

I

HdD

En

t o AAn

mol.

4 o t An65


Plagioclase An"u " ,

contain ing blebs of separate pyroxenesEnn,Fs,uWon, and Enu"FsroWo, Trace s i l ica mineral (Steele andSmith, 1973) 8. Cataclast ic anorthosi te 60215 97Vo anorthosi teAhru-"r , or thopyroxene En.r-r rFsro_rrWor_r, interst i t ia l to granu-lated fe ldspar, very rare i ron. Clasts of Anru, En. .FsrrWo, andCr-spinel (mg O.27 cd.65) (Meyer and McCal l is ter , 1973). 9. Cata-c last ic anorthosi te 60025 3 mm tablets of p lagioclase An"o ,u infine-grained recrystallized matrix. 2Vo Enro-.rFsno-nrWo, ,o andEn.rWonnFs,, occurr ing part ly as crushed grains in matr ix , andpart ly as inc lusions and recrystal l ized grains. Trace s i l ica andopaques (Walker et a l , 1973b).10. Cataclast ic anorthosi te 60025.95Vo An""-"n, Enuo , .Fsn, . rWor- . and Enr, , rFs,u-r ,Wonr_n. most lyas discrete grains, t race i lmeni te and Cr-spinel (Hodges and Kush-i ro, 1973). I l . Granoblast ic anorthosi te 606t9 Rake sample In-homogeneous dist r ibut ion of minerals but composi t ions uni formAng u-" , , , Foro , r , Enuu-rrFsrr-r .Wo3 11, E[aa o"Fsro_,uWo., nu, i l -meni te (MgO 5-8 wt.%), chromire (TiOz 3-5Vo MgO 5.67o Al ,O312-15%) and rut i le (Dowty et a l . , 1974d). l2 Granoblast icanorthosi te 60677 2 mm clast in rake sample breccia. Anr, , , ,

" ,FoTa 7. (Dowty et a l . , 1974d). 13. Nine Apol lo l6 ferroan anortho-s i te rake samples wi th cataclast ic textures and veins of mel t , usu-al ly recrystal l ized. 62275 is most ly maskelyni te. For brevi ty, onlycomposi t ion ranges given here. An"" 2 s? o, except for An. , u , r , ln62275 Rare ol iv ine most ly Fo* u", but Fono_uu for 65789. High-and low-Ca pyroxenes as separate grains or intergrowths, wi thlat ter most common: most are Enro-roWo^,, and En."-nrWo-4s.Rare chromite in 3 specimens (TiO" l -3%o MgO l -4% Al ,O,l l -2 l%o) and i lmeni te in 3 specimens (MgO 2-5Vo) (Dowty et a l . ,1974d\. 14. Genesis rock t5415 Polymetamorphic texrure wi thevidence of cataclasis and recrystallization. 97-99Vo anorthiteAfl ,u ,", 2% EnnoFsr"Wor4, trace EnrrFsooWo3, trace silica mineral,t race i lmeni te (MgO | .0 MnO 0 6 wt Vo) (James, 1972; Hargravesand Hol l is ter , 1972; Stewart et a l ,1972;- Steele and Smith, l97l ;Smith and Steele, 1974) 15. Cataclast ic anorthosi te rake sample15362 . 97Vo An r " , , r rO ro , -o

" , 2Vo Ens "Fs r rWonr , 0 57o

Enu.Fso,Wo,, 0.370 i lmeni te (Al ,Oa 1.6 MnO 0.7 MgO 1.9 wt %),two grains Al-chromite (AlrO, 9.1 Cr,O,48.8 FeO 35.1 MgO I 2)(Dowty et a l ,1972a' , Nehru et a l . , 1974) [6 Cataclast ic anortho-s i te 15264,19 f ragment )907o Ango ee! average 96, or thopyroxenemean EnrrFsrrWo3, t race i ron and chromite (Mason, 1972). 17.Cataclast ic anorthosi tes f rom Apol lo 16. Abstract reports var ioustextures, pyroxene and plagioclase composi t ions, and absence ofs i l i ca m ine ra l . Spec imens 64819 , 60215 , 61016 , 65315 and 60015have Enu, , -uu oFs' o s.5Woq 5_r o and En' n nn rFsr" , , ,

"Woo* n, ,

(Dixon and Papike, 1975). l8 Four anorthosi t ic f ragments inApol lo l5 soi ls , one f rom Luna 20 and three f rom Apol lo 16.Plagioclase At t r r -" r , o l iv ine scat tered values between Fonu and Forr ,h igh-Ca pyroxene Enoo-n"Fs,o-roWo."-r , and low-Ca pyroxenemost lyEnro suFs12-2rWo, u. Notplot ted in Fig. l0(Cameronet a l . ,1973). 19. Anorthosi t ic polymict breccia 67075. Contains c lasts of

microanorthosi te wi th d i f ferent amounts of o l iv ine and pyroxene(Fig. 5 ) Plagioclase is An"u

"" irrespective of textural state. Olivine

as isolated unzoned c lasts wi th b imodal composi t ion range Fono-. ,and Fouo-uu Pyroxene as augi te, p igeoni te, or thopyroxene, andpart ia l ly inverted pigeoni tes ei ther as indiv idual grains or lamel larintergrowths up to 40pm across. Composition ranges for clastsand matr ix may show clusters. Blue-gray chromites occur as rn-c lusions in o l iv ine or as crushed grains in groundmass, whi le

chromite- i lmeni te- i ron intergrowths occur as smal l inc lusions inpyroxenes. Data combined from Brown et al. (1973), El Goresy

et al (1973a), McCallum et al. (1975b), Ghose e/ al (1975) and

Smi th and S tee te ( 1974 ) .

and show a similar range of mineral compositions.Probably most ANT specimens have bulk composi-tions corresponding to norit ic anorthosite andanorthositic norite. Historically, the Genesis Rock,15415, is the most interesting, and indeed its texturehas been interpreted in different ways. Even on Earth,the origin of anorthosites is very controversial,though most petrologists probably favor origin bycumulation from a l iquid followed by solid-state an-nealing which produces many textural variants. Fig-ure 5 shows several textures in the polymict anortho-sit ic breccia 67075, which was interpreted by Brownet al. (1973) as the product of a disrupted layeredigneous complex dominated by plagioclase inter-spersed with minor augite (now exsolved to two pyrox-

enes) . In spi te of the problems, we assume that a l l

lunar anorthosites had an igneous origin which has

been obscured to different degrees by brecciation,shock metamorphism, mix ing and anneal ing.

Pyroxene, olivine, i lmenite, Cr-rich spinel, si l icaminerals. ruti le. iron, and troil i te occur erratically,and sampling problems preclude estimates of modes

and mineral associations. Although incorporation of

extraneous grains cannot be ruled out, we believe that

all these minerals occur in the anorthositic assem-

blage, though not necessarily in equil ibrium. Thepyroxenes occur either as separate grains or inter-grown lamellae, and the wide composition gap in-

dicates substantial solid-state annealing. Augite pre-

dominates several-fold over low-calcium pyroxene,

and there is a wide spread of mg from 0.8 to 0.4.

Olivines occur sporadically, and the range of mg is

similar to that for the low-Ca pyroxenes. Mostplagioclase compositions are between Anrn and Anee,

but some analyses go into the bytownite range. The

spinels (F ig. l2) are r ich in Fe and Cr.The origin of the pyroxenes and the significance of

their mg ratio is controversial. Smith and Steele(1974: see also Part l, p. 237) suggested that theplagioclase incorporated substantial Ca(Fe,Mg)SLOawhich exsolved at lower temperature to give pyroxene

Ca(Fe,Mg)Si2O6 and a sil ica mineral. Compilation ofavailable data (e.g. legend to Fig. l0) shows thatsil ica minerals occur in only some anorthositic speci-mens and are always much less abundant than thepyroxenes. Of course, one can suggest that primary

olivine combined with exsolved sil ica to give pyrox-

ene. The mg rutio of plagioclase is lower than that forpyroxene crystall izing from the same liquid, and

Smith and Steele deduced that the source l iquid oflunar anorthosites would be more Fe-rich if thepyroxenes were primary rather than secondary. The

t074

modal and textural data on anorthositic speclmensare too poor to give a test, but we now believe thatthe major part of the pyroxene is primary: the reasonsare (a) the olivines in Figure l0 have a similar rangeof mg to the low-Ca pyroxenes in spite of erraticmodal variations and (b) the spinels are very Fe-rich(Fig. l2) . I f th is conclus ion is va l id , the parent l iqu idsof the anorthosites could have been very Fe-rich.Predicted mg values are 0.6-0.2 for extreme crys-tal-l iquid fractionation (Part I, Fig. 4), and highervalues 0.8-0.4 for complete crystal-l iquid equil ibra-t lon.

Dowty et al. (1974d) used the termferroan anortho-site for a "widespread and distinctive lunar rocktype," but we are not convinced that a clear dis-tinction has been made from other anorthosites. Wedeliberately placed their spinel-olivine anorthosite60618,1- l wi th our spinel t rocto l i te group because ofchemical similarit ies and inconclusive texture (theirFig I b). They distinguished two granoblasticanorthosites whose ferromagnesian minerals aremore magnesian than those in the ferroan anortho-sites. Hubbard et al. (1974) used bulk analyses ofminor and trace elements to distinguish betweenanorthositic rocks, but the relation to the detailedmineralogy and mineral chemistry is not clear. At thistime we prefer not to attempt a subdivision of theanorthositic specimens, but agree that the presentevidence suggests more than one origin. Because ofthe mineralogical evidence for a hybrid origin ofmany anorthositic specimens, we argue that chemicaldistinctions made from bulk compositions should beless powerful than those made from the individualminerals. Whatever the future of research onanorthositic specimens, the tendency for mg to be lowin the pyroxenes, olivines and spinels must be ex-p la ined.

(d) Granitic and rhyolitic. Although no large rocksor fragments are known, numerous tiny clasts (up to2mm) of granitic material occur on the Moon. Mostspecimens are quenched melts or f ine-scale inter-growths of K-rich feldspar and a sil ica mineral, butoccasional clasts (e.g. Anderson et al., 1972, Fig. 3)show coarse grains in a equigranular texture. Becauseof the small sample size, the number of accessoryminerals is hard to determine, but minor pyroxene,olivine, phosphates, zircon, troil i te, and iron havebeen recorded. Ryder et al. (1975) reported numerousgranitic clasts in Boulder l, Station 2, Apollo 17, anddescribed various textural variants and unusual ter-nary feldspars. K-feldspars in granitic specimens usu-ally carry about 1-2 weight percent. BaO and vari-

J V. SMITH AND I M. STEELE

EnFo

able but low amounts of NarO. The sparse data on

olivine and pyroxene compositions (Fig. I 1) are scat-

tered and tend to be Fe-rich. The presence of K-rich

feldspar, phosphates, and zircon obviously gives a

KREEP-like affinity to the granitic and rhyolit ic

specimens. All workers agree on an origin fromhighly-differentiated l iquids but the relative roles ofprogressive fractional crystall ization (Ryder et al.,1975), partial remelting, and liquid immiscibil i ty arecontroversial.

(e) Mare bqsalts. We shall not attempt to review

in detail the mineral data for mare basalts. Becausemare basalts are undoubted igneous rocks crystal-l ized from magmas, the bulk compositions are mean-ingful. Readers could consult Taylor (1975, Chapter

D i Hd

FsFo

r00FIc I l . Composi t ions of pyroxene and ol iv ine in grani t ic and

r[yol i t ic specimens. I Grani t ic and rhyol i t ic f ragments in

14306,14321 and 14270 breccias. Several contain cores of

K-fe ldspar and quartz, and one (14306,53) has 0.4 mm grains wi th

planar faces Most are glassy or have granophyr ic textures. Analy-

s is of Fe-augi te microphenocryst and K-fe ldspar (NarO 0.5 CaO

0 8 BaO 2.6) in 14306,53 grani te c last (Anderson et a l . , 1972).2

Rhyol i t ic c lasts (0 5-2 mm) in breccias f rom Boulder I , Stat ion 2,

Apol lo l7 Maf ic, non-maf ic, holocrystal l ine, and glassy types Fig.' 3 shows tota l range of pyroxene composi t ions in a l l c lasts, but indi-

v idual ranges are smal ler . K,Ba fe ldspars (BaO l . l -3.6) , p lagio-

c lase, and unique ternary fe ldspars near AnroAbloOrno. Cr istobal-

i te, o l iv ine For,-n. , i lmeni te, t ro i l i te, i ron, REE-whi t locki te, and

zircon (Ryder et a l . , 1975) 3 Rhyol i t ic f ragment in 12070 wi th

granophyr ic texture, o l iv ine Forr , ferroaugi te and i lmeni te (Marvin

et a l . , 1971) 4. Polymict KREEP-r ich breccia 12013 permeated by

"grani t ic" component. Dist inct ion between pr imary and second-

ary minerals was di f f icul t . Drake et a l . (1970, Fig. 5g) showed two

ranges o f py roxenes assoc ia ted w i t h g ran i t i c a reas -

Enoo-uoFsnr r rWoa-16 and Enrr-r .Fs, noWoro-.0, K-fe ldspars

Oruu-roAn, 15, quartz and t r idymite, i lmeni te, Cr-ulvc ispinel , t ro i l -

i te, i ron, apat i te, whi t locki te, and z i rcon. See also Lunat ic Asylum

(1970). 5. Clasts of grani t ic composi t ion in t4303, 14319, 14320,

and 14321 breccias. Glassy to crystalline K-feldspars (Na,O

0.6-1.4 BaO l -3 CaO 0 2-1.0) and plagioclase (Ab 8-27Vo Or

l -6%) (Roedder and Weiblen, 1972b).6. Granoblast ic c last of

"micrograni te" in 14321 breccia I X 1.5 mm. Mosaic of equant

grains of K-fe ldspar (NarO 0.3, CaO 0.3 BaO 1.9) , s i l ica mineral ,

minor z i rcon, apat i te, and whi t locki te. Also rhyol i t ic g lasses, some

devi t r i f ied (Gr ieve et a l ,1975).

80 60 20

LUNAR MINERALOGY IO75

4) for a useful introduction. Idiosyncrasies in the core, an olivine-rich mantle, and a plagioclase-cooling history of fragments torn out of different pyroxene-rich crust;levels of the same lava flow produced extreme com- (b) intense brecciation of surface rocks, and com-plexity in the zoning of the pyroxenes. Many of the plex volcanic, plutonic, and metamorphic activity indetails are irrelevant from the viewpoint of overall the upper 200km during the next 0.5Gy; minorlunar petrogenesis. Mare basalts definitely have di- amounts of incoming projecti les were incorporated;verse bulk compositions and mineralogy requiring (c) KREEP material tended to be dispersed; on thedifferent source regions, and partial melting of hybrid largest scale, Fe-rich ANT rocks would tend to over-rocks is a popular but not fully accepted model. l ie Mg-rich peridotit ic rocks with late oxide-richAlthough there are subtle differences in the chemical cumulates in the middle, however, on a smaller scale,zoning which wil l be mentioned in later sections, layered plutons would develop sporadically in thepyroxene compositions tend to l ie at the left of the near-surface region;region labeled FETI in Figure 3 of Part I: indeed all I (d) after -4.OGy, consolidation of a lunar crust;mare basalts have pyroxenes with mg < -0.7. Oli- igneous activity mostly confined to partial meltingvines occur only in some mare basalts and also obey and migration of interstit ial l iquid; extrusion of marethe boundary at -0.7. Strong zoning occurs to very basalts produced by remelting mixtures of earlyFe-rich compositions, sometimes with formation of cumulates and trapped liquids; the early melts wouldpyroxferroite. Plagioclases tend to l ie in the range tend to become mixed up in the surface debris, andAnru-ru. Spinels are very Fe-rich and zoned from only the later melts would be easily recognizable aschromite to ulvdspinel. KREEP minerals occur at the mare basalts.low concentrations in the late residua. Some of the above ideas have become widely ac-

(J) ANf. To conclude this section, a few words cepted, but the nature of the lunar interior and theare needed on the catch-all acronym for anorthositic, origin of mare basalts are highly controversial. Manynorit ic, and troctolit ic specimens. Materials with of the mineralogical and petrological data can bethese bulk compositions dominate the non-mare lu- explained satisfactori ly by early melting of only thenar samples, but most are metamorphosed polymict outer part of the Moon, followed by migration of abreccias. We have found no simple way of sorting out molten zone into the interior. All simple models forthe mineral properties and in despair refer readers to the origin of mare basalts have apparently been aban-the summary diagrams in Part I. Most ANT speci- doned, and multi-stage models are being explored. Atmens have ferromagnesian minerals withmg 0.5-0.8. one extreme are models involving remelting of latePyroxenes dominate over olivines, and low-Ca pyrox- cumulates produced during primordial differenti-enes dominate over high-Ca pyroxenes. Most plagio- ation (e.g. the use of phase-equil ibrium data byc laseshaveAn)90pe rcen t .A fewspec imenscanbe Wa lke r e ta t . , 1975 , t oexp la inh igh -T i l una rbasa l t sclassified into the groups given earlier. The remainder from remelting of i lmenite-rich cumulates), and atcan be interpreted as a heterogeneous rag-bag of the other extreme are models involving late melting inmaterial dominated by plagioclase and low-Ca pyro- the interior of the Moon (e.g. Kesson and Ringwood,xene probably produced by mixing up the products of 1976). We refer readers to the Proceedings of the 1975multiple igneous and metamorphic processes. De- Mare Basalt Conference (obtainable from the Lunartailed study of minor and trace elements of mineral Science Institute) and of the Seuenth Lunar Scienceand rock clasts should enable progress to be made, Conference. Whatever the problems concerning thebut the task is daunt ing because of the loss of ev i - ins ide of the Moon, we bel ieve that manydence by partial melting and solid-state diffusion.

Origin of rock types

Although very complicated models can be devel-oped, and indeed must be necessary for individualbreccias, we suggest that all the data on the rocktypes and mineral compositions can be accom-modated in the following simple framework:

(a) crystal-l iquid differentiation of the entireMoon about -4.5Gy with formation of Fe,S-rich

petrographic and mineralogic data can be fitted intoany scheme involving early differentiation of all ormost of the Moon (e.9. 76535 and 15445 from aplagioclase peridotite zone, 67075 from a disruptedlayered pf uton, 62295 from melting of an Mg-richpfagioclase-peridotite zone, 14310 from incompletemelting of plagioclase-rich debris).

We shall consider the petrologic model in detail ina discussion of the controls on the bulk chemicalcomposition of the Moon (Smith and Steele, in prep-aration). The remainder of this review is devoted to

t076 J. V SMITH AND I, M. STEELE

o nor i t i c+ onor ihos i t i c. dun i t i c

spinel -ol-plog cotoclositespinel troctol i t ictroctol i t ic I+

2a

t n

. a l

TPXK4;.! ! .lYc-

. ochondr i teso Murchisono A l lende

I 9tr

X

o

B 1 "Wl.

o/o

Ti02Ia

o' -hd;3 2

TICBC( - 't

lllgCr20a

discussion of the mineral groups, especially on the

clues they provide to petrologic and geochemical dif-

ferentiation.

Discussion of minerals

Feldspars

The sections on feldspar and olivine bring up to

date the corresponding sections in Part I.Meyer et al. (1974) investigated the potential and

limitations of a low-resolution ion microprobe for

analysis of trace elements in lunar plagioclase. Only

the elements Li, Mg, K, Ti, Sr, and Ba occur insufficient concentrations to give statistically valid sig-nals that can be corrected accurately for overlapping

masses, especially those from troublesome molecularions. From plots of Li us. Ba and Mg us. Ba, Meyer etql. recognized distinct compositional ranges forplagioclases from anorthositic and gabbroic rocks (Li

l -2ppm, Ba 5-25ppm, Mg 150-350ppm), KREEP

basalts (Li l0-30, Ba 60-220, Mg 750-1500), marebasaf ts (L i2-14, Ba l5-350, Mg > 1700) and a h igh-Ba group of plagioclase clasts from Apollo 14breccias (Li 14-50, Ba 200-750, Mg 200-1100). Theyconcluded that these Apollo 14 breccias could nothave derived from a pre-Imbrium regolith becausethey should then have contained more plagioclase ofthe first type believed to represent the original lunarhighlands. [Data for 14063, a fine-grained clastic rockwith several components, do not fall with data for the

FeAl20a

o2

0.4mg

0.6

0.8

Mqlzoootomic cr

Frc. 12. Major-e lement composi t ion of spinels f rom non-mare rocks mg and cr are atomic Mgl(Mg*Fe) and Cr l (Cr*Al) respec-

tively. The numbers refer to the specimen sequence in the appropriate figure legend for the rock type (see Fig. 6 for dunitic, troctolitic,

spinel-anorth i te-or thopyroxene-ol iv ine cataclasi te and spinel t roctol i t ic ; F ig.9 for nor i t ic ; F ig. l0 for anorthosi t ic) Composi t ion ranges

are given for the fo l lowing: M al l meteor i tes (Bunch and Olsen, 1975), EC equi l ibrated chondr i tes (Snets inger et s l . , 1967) ' UC

unequi l ibrated chondr i tes (Bunch e/ at . ,1967), achondr i tes (Bunch and Kei l , l97l ; Lover ing, 1975), Murchison C2 meteor i te (Fuchs el

at , 1973), Type A and B inclusions, Al lende meteor i te (Grossman, 1975), TPXK terrestr ia l per idot i te xenol i ths in k imber l i te (Smith and

Dawson, 1975), TXAB terrestr ia l xenol i ths in a lkal i basal ts (var ious sources especia l ly Basu and MacGregor, 1975), TPAC terrestr ia l

podi form deposi ts in a lp ine complexes (var ious sources including those t is ted by Smith and Dawson, 1975), TICBC terrestr ia l igneous

complexes wi th basal t ic composi t ions (many sources including those for thole i i t ic composi t ions, e.g. Bushveld, and those for metamor-

phosedArcheancomp lexes ,eg F i skenaesse t : seeSmi thandDawson , l gT5 , f o r re fe rences ) ,Rhumu l t r abas i ccomp lex (Hende rson , 1975 )

In the auxi l iary p lots for TiOr, t rends for most terrestr ia l specimens are omit ted to avoid over lap. Most such specimens have less than I

wtVo TiO,.

other Apollo 14 breccias.l Apollo 16 samples withmetaclastic texture contain many plagioclase clastswith low content of trace elements even though thebulk content of trace elements is high. Microprobeanalysis of minor and trace elements is a powerfultool for deciphering the origin of plagioclase espe-cially in anorthositic rocks and breccias.

Detailed crystal structure refinement by Smyth( I 975) of anorthite Nao e36Ks.ee5Cae.errFeo.oorAlr.nrrSir.o,from troctolitic granulite 76535 demonstrated l0 per_cent vacancy in the M site of type (000). This result issurprising for a strongly-annealed specimen, thoughperhaps it is inherited from primary crystallization ofa defect anorthite from an earlier igneous melt. Smithsuggested that the vacancy might result from ex-pulsion of Fe upon reduction to metal.

Lofgren et al. (1974) reproduced experimentallythe textures and mineral chemistry of Apollo 15quartz-normative basalts. Megacryst-groundmasstexture could be produced by steady cooling and doesnot require two discontinuous periods of crystalliza-tion. All the synthetic plagioclases were rich in minorelements and were found from electron microprobeanalyses to contain I l-23 percent Cao.uAlSirO, for-mula-unit. Crawford's (1973) description of the zon-ing and chemical variation of plagioclase in marebasalts is closely matched by that of Bryan ( I 974) forquenched oceanic basalts. Delayed growth of plagio-clase from basaltic melts is a serious problem fordetermination of stable phase-equilibria (Gibb,1974\.

Dowty et al. (1974a) studied the occurrence oftwins in Ca-rich plagioclase of feldspar-rich rocksattributed to the lunar highlands. Carlsbad andCarlsbad-Albite twins as well as Albite and periclinetypes were found in rocks with igneous texture which

.- had not been severely brecciated. In cataclastic rocks,only Albite and Pericline twins were found, andDowty et al. suggested that original Carlsbad andCarlsbad-Albite twins had disappeared by breakagealong twin boundaries. Albite and Pericline twinscommonly showed features attributed to origin bydeformation. Anorthites can undergo such twinningvery easily and reversibly because the Si,Al orderdoes not block the twin process. Carlsbad and Carls-bad-Albite twins are not infrequent in granoblasticrocks, and probably arise from high-temperaturesolid-state recrystallization. In terrestrial plagioclase,such twins tend to occur in igneous specimens and areabsent in low and medium grades of metamorphism:however, some do occur in the granulite facies. Manyother reports of plagioclase twinning occur in the

LUNAR MINERALOGY t07'l

l i terature: e.g. Wenk et al. (1972) found that forp lagioc lase in rocks 12051, 14053, and 14310 the twinfrequencies were generally Albite ) Carlsbad-Albite- Carlsbad - Pericline (or Acline) - Baveno )Albite-Ala ) Manebach, but there were minor dif-ferences from rock to rock.

Many papers described the domain textures of lu-nar plagioclases. Wenk et al. (1973) compiled dataobtained by electron microscopy on out-of-step do-mains, and proposed a phase diagram to explain theoccurrence of b and c domains and of the Huttenlo-cher intergrowth. Subtle features of diffuse X-raydiffractions of both lunar and terrestrial plagioclasewere interpreted by Jagodzinski and Korekawa(1973,1975) in terms of stepped domain boundaries.

Shock and recrystall ization produced many com-plex textures in lunar plagioclase (e.g. Heuer et al.,1974), and high-speed particles left tracks. All theseeffects reported in numerous papers testify to thecomplex history of the lunar regolith with regard toimpact of both meteorites and nuclear particles.

Ryder et al. (1975) reported feldspars with unusualternary compositions from Apollo 17 granitic clasts(see earlier). Those in holocrystall ine clasts are com-positionally uniform in any clast and lie nearAnuoOrnoAbro whereas those in a rim between plagio-clase crystals and granitic groundmass range fromplagioclase composition to AnurOrroAbru. Some clastscontain intergrown lamellae of K-feldspar andplagioclase, probably the result of unmixing. Theternary compositions l ie well within the stabil ity f ieldof two coexisting feldspars, and some kind of non-equil ibrium crystall ization or metasomatism seems tobe necessary.

Niebuhr et al. (1973) interpreted the ESR spectrumof I 5415 plagioclase in terms of l ines from Mn,+ andfrom Fe3+ in two structural sites. Only I percent ofthe iron was trivalent.

Oliuines

Following up the ideas on minor elements in Part I,Steele and Smith (1975b) measured Al , Ca, T i , Cr ,Mn, and Ni down to the 50ppm level in Mg-richolivines which may derive from the early lunar crustor deeper environments. Low Ca contents (0 to 0.1wt.Va CaO) consistently occur only in olivines fromdunitic and troctolit ic rocks, especially breccias,whereas olivines from spinel troctolites, mare basalts,and ANT materials from Luna 20 have CaO over 0.1weight percent (Fig. l3). Although the effect of bulkcomposition is not known, simple analogy with ter-restrial specimens indicates that high-Ca olivines de-

l07E

rive from volcanic environments and low-Ca onesfrom annealed ones. The distinctly lowest Ca content(0.01 wt.7o) is for olivine in 15445 peridotitic catacla-site (see earlier).

The highest contents of AlzOr found in new analy-ses of lunar olivines are 0.08 weight percent, and mostvalues are below 0.04 weight percent. The high valuesin the literature, collected in Part I (Fig. 7), should bediscarded unless they are confirmed by new analyses.Currently it appears that serious revision of tech-niques (probably with respect to determination of thebackground) is needed in some electron microprobelaboratories.

High Cr analyses (up to 0.15 wt.7o) and low Nianalyses (up to 0.05 wt.7o) in Mg-rich olivines con-firm earlier analyses. A weak anti-correlation be-tween Cr and Ni might be explainable in terms of theoxidation state being so low that Ni2+ is removedupon reduction to the metallic state while Cr'+ isreduced to the divalent state which should favor en-trance into olivine. However, the data scatter somuch, and there are so many indications of com-plexities in lunar rocks, that such an idea is presentedwith great hesitation. Furthermore a weak positivecorrelation indicates coupled substitution between Crand Al. Titanium ranges up to 0.1 weight percent andappears to be higher in olivines from plagioclase-richvolcanic rocks; this may result from plagioclase tend-ing to crystallize from later differentiates of the bulkMoon, whose Ti content has been enriched by exten-

J. V. SMITH AND I. M. STEELE

sive earlier crystall ization of Ti-poor olivine. System-

atic study of other transition metals (V,Co) and other

minor and trace elements in olivine should place fur-

ther controls on models for the crystall ization of

olivine.Burns et al. (1973) concluded that there is no

unequivocal optical-absorption evidence for sub-

stitution of Cr2+ in lunar olivines and pyroxenes be-

cause of possible overlap of absorption bands with

those for Fe'+. Furthermore they concluded that

there is no reliable evidence for assigning bands to

Cr'+. If the chromium in lunar olivines is indeed all

tr ivalent, it wil l be necessary to ascribe the high Cr

content of some lunar olivines to high crystall ization

temperature rather than to low oxidation state. The

occurrence of high Cr content in olivine from terres-

trial komatiite lavas could be explained in the same

way (e.g. Green, 1975). Detailed study of olivines

synthesized from Cr-bearing systems at controlled

temperature, pressure, and oxygen fugacity is badly

needed.Hewins and Goldstein (1974) studied the Ni and

Co contents of Ni-Fe metal grains and their olivine

host from Apollo 12 mare basalts. Cobalt concentra-

tions in olivine ranged from 300ppm at the core to

70ppm at the rim. The ColNi distribution between

olivine and metal indicated non-equil ibrium, and

showed that the metal did not crystall ize first. Per-

haps late reduction by sulfur loss (see later), affects

the distribution of transition elements. Hewins and

o.r o.2 0.3CoO in Olivinc (wt.%)

Frc. 13. Calc ium in lunar o l iv ines. Steele and Smith (1975b).

Ave.=O.O75

Sp.fio.t. Auno iOt.-Prinz etol. (19730)

sp.Troci. (a)_ _---zweibren etol. (1974)

tic",?octolific-

o.4

Goldstein (1975a) concluded that the olivine andmetal of 76535 troctolit ic granulite equil ibrated at900 + l00oc, a range not inconsistent with thatdeduced from the coexisting pyroxenes.

Ghose et al. (1975) observed slight Mg,Fe segrega-tion in olivine from cataclastic anorthosite 67075, arock which contains orthopyroxene whose Mg,Fe or-der ind icates equi l ibr ium at 640.C. In the o l iv ine,0.525 + 0.002 Fe atoms enter the Ml site and 0.4g0the M2 site.

Huffman et al. (1975) interpreted magnetic hyper-fine peaks of Mdssbauer spectra of olivine-rich lunarsamples taken at 5oK in terms of superparamagneticclusters of Fe2+ spins in mirror sites. Annealing at1000"C for I day reduced the hyperfine peak intensityby 30 percent in an olivine from a lunar basalt. Ter-restrial olivines of unspecified origin showed no hy-perfine peaks. When detailed time and temperaturestudies have been made, the hyperfine spectra mayprovide another thermometer for high temperatures.

Olivine clasts in high-grade metamorphosedbreccias react with the matrix. Cameron and Fisher(1975) descr ibed coronas of pyroxene, i lmeni te, andplagioc lase around o l iv ines in Apol lo l4 breccias,and attributed them to diffusion of Ti, Mg, and Fe.

The distribution of mg between olivine and basalticl iquid was used by many workers in models for crys-tal-l iquid fractionation in the Moon. Roeder andEmslie (1970) found that rng (olivine)/rng(tiquid) wasindependent of temperature from I 150 to 1300.C andof oxygen fugacity from l0-0.? to l0-r2 atm. Theirratio of 0.30 was found to apply approximately tosynthetic data for lunar basalts except that high i lme-nite content caused some displacement. For hydroussystems at l5kbar, the ratio ranges from 0.52 at theiron-wiistite buffer to 0.04 at the magnetite-hematitebuffer (Mysen, 1975). Because the oxidation state andwater content of the Moon may have changed greatlysince its inception, the fractionation between olivineand early l iquid may have changed with time.

Pyroxene

Pyroxene is the major mafic phase in the returned

(a) Relation of major elements to rock rype andcrystallization conditions. For classification or char-acterization of rocks, soils, and breccias, plots of


Mg,Fe,Ca on the we l l -known quadr i la te ra lEn-Fs-Di-Hd are useful, as illustrated by Reid el a/.(1970), Prinz et al. (1973a), and Steele and Smith(1973). The use of pyroxene (and olivine) composi-tions to distinguish rock types was illustrated earlier.In general, the mg value decreases with progressivecrystal-liquid differentiation, thereby allowing test-ing of petrologic models. The pyroxene grains fromcoarse-grained rocks are individually homogeneousexcept for exsolution, and the bulk compositionprobably gives a good clue to the original parentliquid. Highly metamorphosed breccias have rela-tively uniform pyroxenes, whereas poorly-metamor-phosed ones have diverse pyroxenes (e.g. Warner,1972, Fig. l0). Rapidly-crystallized pyroxenes showcomplex zoning trends which are not even uniformover the same thin section or even in different tra-verses of the same crystal. The details are very com-plex, and only a brief account can be given here.However, the data are consistent with the generalconclusion that pyroxene nucleates easily and thatthe mg value of the first pyroxene depends fairlyclosely on the bulk mg. Subsequent zoning dependson many factors including competition with otherminerals for the components in the declining reser-voir of liquid, melt temperature, viscosity, and cool-ing rate.

The main emphasis has been on pyroxenes frommare basalts which show many textures ranging fromskeletal crystals in vitrophyres to megacryst-ground-mass assemblages. Initially it was believed that thelatter texture resulted from early slow crystallization(perhaps in a deep-seated magma chamber) followedby rapid crystallization (perhaps upon extrusion atthe lunar surface as a lava flow). However, Lofgren etal. (1974) were able to reproduce all the textures ofApollo l5 quartz-normative basalts merely by vary-ing the rate of monotonic cooling. Indeed many spec-imens of lunar basalts probably came from the sameindividual lava flow and differ merely in their close-ness to the surface; of course, initial quenching com-bined with multiple extrusion could give complexcooling histories. The observations by Dowty et al.(1973b, 1974c) on Apollo l5 basalts il lustrate thecomplexities.

Sector zoning, often associated with skeletalgrowth, produces quite different chemical trends inthe same crystals (e.g. Hollister et al., l97l; Bence etql., 197 l; Boyd and Smith, I 97 I ). Delayed crystalliza-tion of plagioclase (see earlier) results in a highercontent of CaAlrSizO, in the residual magma, whileonset of plagioclase crystallization results in a sharp

J, V. SMITH AND I. M STEELE

drop of Ca and Al in the zoning trend of the pyroxene(e .g . Ho l l i s t e r e t a l . , l 97 l ;Bence e t a l . , l 97 l ; Dowty

et al., 1974c). After about two-thirds of the basaltic

l iquid has crystall ized, the remainder tends to be

trapped in pockets, and the composition of the crys-

tall izing pyroxenes is affected by the idiosyncrasies of

the local environment. Most pyroxene crystall izationoccursjust above the vault-shaped solvus, but discon-tinuities in some zoning trends from pigeonite to

augite compositions suggest straddling of the solvus.Finally, very fine oscil latory zoning probably resultsfrom diffusion effects at the crystal-l iquid surface.After the init ial surge of studies of pyroxene zoning,most investigators have settled for random analyseswhich provide a statistical averaging of the above

effects. Bence and Papike (1972) gave details ofzon-

ing t rends for Apol lo l l , 12, 14, 15, and Luna 16

speclmens.Random Ca, Mg, Fe plots of pyroxenes from mare

basalts mostly fall in the region labeled FETI in Part

I, Figure 3. There are many dozens of such plots, and

those of Dowty et al. (1973b) are good examples.

Although there are slight differences from one text-

ural type to another, all plots show mg < - 0'7 with

the major population between 0.7 and 0.6. Extreme

zoning trends ultimately produced the pyroxenoidpyroxferroite. Relative concentrations of Ca-rich and

Ca-poor pyroxenes definitely vary from one type of

mare basalt to another: thus Ti-rich basalts 70215

and 75035 are strongly dominated by Ca-rich pyro-

xenes (Longhi et al., 1974) whereas most basalts have

both Ca-poor and Ca-rich pyroxenes.

To summarize'. the peridotite suite is characterized

by high-Mg, low-Ca pyroxenes accompanied by only

sporadic rare diopside, the KREEP suite is also

dominated by low-Ca pyroxenes, but these are some-

what less rich in Mg and are accompanied by rather

more Ca-rich pyroxenes, the mare basalt suite is even

poorer in Mg and has a greater proportion of Ca-rich

pyroxene, and the ANT conglomeration occupies a

middle region. This general trend from Mg-rich, Ca-

poor to Fe-rich, Ca-medium compositions can be

explained by cotectic crystall ization moving off the

simple olivine-plagioclase-sil ica ternary towards a

Ca-rich pyroxene end-member. Of course, partial

melting is involved as well as progressive crystal-

l iquid fractionation, metamorphism, and hybridism.(b) Minor and trace elements. In conformity with

the low content of volati le elements in the Moon, Na

and K are much lower than for terrestrial pyroxenes'

Potassium is below detection in routine electron mi-

croprobe analyses (- < 0.05 wt.7o) and probably is

below 0.01 weight percent in most samples. Sodium is

tailed analyses of Na and K are needed.

Broad overall correlations between the major, mi-

nor and trace elements are rather similar to those

found for terrestrial pyroxenes; e.g. Cr tends to

correlate with Mg, while Al, Ti, and Cr tend to

near-absence of Na and K, the substitutions of

Ti, Al, and Cr can be interpreted in terms of the

following components (Bence and Papike , 1972):

W+Tto* Alrou, _rl l+Cr3+SiAlO., and .&f+AlSiAlO6.

The components i l+Ti3+AlSiO6 and Cr2+SiOt

have also been invoked. Zoning trends of pyrox-

enes from mare basalts were plotted by many in-

vestigators and interpreted in terms of. factors such

as: beginning of plagioclase crystall ization involving

decrease of ,41*AlSiAlO. component and approach

to the l:2 ratio of Ti to Al for the rl l+Tia+AlzO.

component; presence of Ti3+ by a TilAl ratio greater

than 0.5; early removal of Cr'+ leaving Cr-depleted

liquid; correlation with crystall ization of i lmenite or

spinel; and changes in crystall ization rate. Ea4Y

pyroxenes from 14310 (see earlier for references) are

inriched in M'?+AlrSiO. probably because of the high

AlrO, of the bulk rock, but crystall ization at high

pressure has also been invoked. Most ANT rocks

are metamorphosed breccias with diverse origins and

textures, and the significance of minor elements has

not been evaluated. Careful evaluation of the minor

and trace elements in the pyroxenes from unusual

Mg-rich rocks is needed: perhaps the most noticeable

feature in Table I is the high ALO' content of py-

roxenes in 15445 and'73263 cataclasites.

Qualitative areal scans with an ion microprobe

showed t rends for Mg, Ca, T i , A l , Mn, L i , Na, K,

and Ba in some pyroxenes from mare basalts (Bence

and Autier, 1972). Higher concentrations of some

elements were found at crystal edges or at pigeon-

ite-augite interfaces, probably from liquid trapped in

defects.(c) Exsolution, inuersion and site population. Prior

LUNAR MINERALOGY l08 l

to the Apollo program, the phase relations and poly-morphism of the pyroxenes were only partly under-stood, and major new discoveries were made on lunarsamples. The state of knowledge in 1969 is ex-emplified by Special Paper Number Two of the Miner-alogical Society of America, J. J. Papike, Ed. papikeet al. (1973) interpreted pyroxene structures from theviewpoint of topology and structural l inkages.

Ross e/ al.(1973) delineated the augite-pigeonitesolvus at I atm, using pyroxenes from mare basaltl2l2l. Lindsley et al. (1974b) determined the compo-sitions of augite and hypersthene coexisting at 810"Cand l5kbar under hydrothermal conditions, andfound that the composition gap was similar to thoseobserved for coexisting augite and either pigeonite orhypersthene in various lunar specimens (aug-ite-pigeonite-orthopyroxene intergrowths in crystalsfrom 67075 anorthositic breccia, Brown et al., 1973,McCallum et al., 1975b; inverted pigeonite clastsfrom 14082 and 14083 breccias, Papike and Bence,1972; inverted pigeonite in fractured clast of"anorthositic gabbro" in breccia 15459, Takeda,1973). All these specimens with coarse intergrowthsof augite and low-Ca pyroxene(s) are the result ofhigh-temperature crystall ization followed by pro-longed annealing in the solid state. Inversion frompigeonite to orthopyroxene is sporadic. Althoughprecise conditions cannot be determined, the coolinghistory and other evidence require burial at somekilometers depth probably of gravitationally-layeredplutonic bodies. Analogies were drawn with the well-known Skaergaard and Bushveld complexes onEarth.

Low-Ca pyroxenes with Mg-rich compositionshave orthorhombic symmetry (e.g. those in the Mg-rich specimens described earlier, and in fragmentsascribed to KREEP-rich rocks, Fuchs, l9?l). Thesepyroxenes show no evidence of inversion from pi-geonite, and presumably crystall ized directly with or-thorhombic symmetry. This is consistent with theoccurrence of low-Ca, Mg-rich pyroxenes in terres-trial rocks. However, bronzites from troctolit ic gran-ulite 76535 and norite 78235 (Smyth, 1974a; Steele,1975) have the space group P2rca rather than Pbcainterrestrial bronzites. Crystal-structure analysis bySmyth (oral communication, 1975) showed segrega-tion of Mg and Fe into four sites instead of the usualM(l) and M(2) sites. The new space group was alsofound in pyroxene from the Steinbach meteorite. pro-longed subsolidus annealing must be needed to pro-mote the strong ordering of Mg and Fe, as was infer-red from the textural features of the lunar andmeteorit ic specimens. Perhaps the new space group

developed only in pyroxenes annealed at some tens ofkilometers depth in rocks ascribable to a peridotit icfacies.

Coarse exsolution of lunar pyroxenes is uncom-mon, and most specimens have exsolution which is sofine that X-ray diffraction and electron microscopictechniques are needed. In general for mare basalts,the pyroxenes crystall ized mostly as augite or pigeon-ite (often as zoned crystals) which exsolved smallamounts of pigeonite and augite, respectively, duringcooling. Pigeonites crystall ize with C2/c symmetryand invert to P2r/c symmetry upon cooling below atemperature ranging from 1000'C for Mg-rich speci-mens to 500' for Fe-rich ones (Prewitt et al., l97l).Passsage through the inversion results in an antiphasedomain texture which is revealed by dark-field elec-tron micrography (e.g. Champness and Lorimer,l97l ; Champness e, a l . , l97 l ; Chr is t ie et a l . , l97 lLally et al., 1972: Ghose et al.,1972). Very complexlamellar structures were observed which require sev-eraf stages of exsolution (e.g. Ghose et al., 1973).lnslowly cooled rocks, inversion from pigeonite to or-thopyroxene may occur. Unfortunately it is diff icultto determine from textures whether exsolution pro-ceeded by spinodal decomposition or heterogeneousnucleation. Lally et al. (1975) suggested that the fol-lowing experimental data yield the best informationon the cooling rate: difference ofB angle for exsolvedpyroxenes in quickly-cooled rocks, textural evolutionand domain size in moderately-cooled rocks, and thewidth of diffusion zones and the domain morphologyin slowly-cooled rocks.

Finally, the distribution of cations between theM(l) and M(2) sites depends on temperature, andcan be examined by either Mrissbauer resonancestudy of the Fe distribution or by X-ray crystal struc-ture analysis ofthe electron density. Interpretation ofboth types of data is diff icult for various technicalreasons (e.9. Dowty et al., 1972b; Schiirmann andHafner, 1972), but the evidence for equil ibration ofcations between the M(l) and M(2) sites down totemperatures of around 600'C even for pyroxenesfrom mare basalts seems convincing (e.g. Hafner etal., l97l; Ghose et al., 1972; Brown and Wechsler,1973; Yajima and Hafner, 1974). Such low temper-atures have appeared surprising in view of the per-sistence of pigeonite and the expected rapid coolingof thin basaltic f lows. However, the Mg, Fe distribu-tion cdn be modified by atomic diffusion on a unit-cell scale, whereas polymorphic inversion requiresmajor structural changes over much larger distances.Suggestions of a late episode of metamorphic heatingof basaltic specimens may require further thought.

1082

Some non-pyroxene phases were found as ln-

clusions in pyroxenes and probably result f iom ex-

solution: oriented i lmenite and green spinel in or-

thopyroxene clasts from 14321 breccia (Gay et al.,

1972); oriented i lmenite lamellae in orthopyroxenesfrom Apol lo KREEP-r ich soi ls (Fuchs, l97 l ) ;p late-

lets of chromite in pyroxene En..FsrrWo, grains of

14258 soils (Haggerty, 1972a); chrome-spinel la-

mel lae paral le l to (001) and perpendicular to (001) in

pigeonite from Luna 20 soil (Ghose et al., 1973);

oriented lamellae of i lmenite in a hedenbergite crystal

from Luna 20 soil (Haggerty, 1973a). A cluster of

iron needles in pyroxene from 12021 basalt was inter-

preted by Walter et al. (1971) as the result of late-

stage introduction of Fe rather than from exsolution.(d) Deformation. Disorientation and cracking of

lunar pyroxenes were explained by several factors:

differential thermal contraction of pigeonite and aug-

ite (Carter et al., l97l), inversion from C to P lattice

symmetry of pigeonite (Brown et al., l97l), rapid

growth (e.g. Carter et a\.,1970), and shock deforma-

tion (e.g. Dence et al., 1970; Engelhardt et al., 1970).

Careful study is needed to separate all these factors as

emphasized by Carter et al. (1971). Although some

lamellae in pyroxenes are reasonably ascribed to

shock deformation, most arise from exsolution' Even

the most casual study of breccias reveals that pyrox-

enes are much less affected by shock than plagio-

clases: indeed some soil fragments contain plagio-

clase which was converted completely to maskelyniteglass whereas the pyroxene was merely fractured.

Although the possibil i ty should be considered that a

plagioclase-rich magma might be produced by fi l ter-

pressing the l iquid produced by shock deformation of

plagioclase-pyroxene rocks, no evidence has been

adduced.

Other silicates

The sil ica minerals cristobalite and tridymite occur

as late crystals in mare basalts. Several papers in the

Proceedings of the Apollo I I and l2 conferences de-

scribe stacking faults and complex inversion twinning

of cristobalite. Quartz is very rare in basalts, but is a

characteristic mineral in granitic fragments where it is

usually intergrown with a barian sanidine. Electron

microprobe analyses report 0.0n-0.n weight percent

of AlrOs, TiO2, FeO, CaO, NarO, and KrO in cristo-

balite and tridymite, but the accuracy is unknown

because of problems from secondary fluorescence

and small grain size. On Earth, cristobalite and tridy-

mite have not been found in ancient rocks because of

inversion to quartz. Persistence of cristobalite and

J, V. SMITH AND I. M. STEELE

tr idymite on the Moon (and in meteorites) might be

explained by absence of the catalytic effect of water,

but other effects cannot be ruled out. Rare quartz

grains in lunar soils may be relics from coarse-

grained granites. Charles et al. (1971) suggested that

quartz in granitic specimens implies crystall ization at

over 400 bars for a dry system, so long as the quartz

crystall ized from a melt and did not invert from

tridymite or cristobalite. This pressure corresponds

to a depth of over 10km.Pyroxferroite, a pyroxenoid with a seven-repeat

sil icate chain, commonly crystall ized instead of pyrox-

ene from the late residua of mare basalts. Its compo-

sition range was not delineated precisely because of

complex textural relations with Fe-rich pyroxene, but

is probably Mgo-o.ruFeo.r -o. rCao.r -o. r rSiOe' Table 2 '

No. 1, gives a representative analysis' Crystal-struc-

ture refinement showed cation disorder (Burnham,

1971), and the number of te t rahedra in the chain

repeat was not controlled by the cation stoichiome-

try. Pyroxferroite crystall ized metastably (Lindsley

and Burnham, 1970).Tranquil l i tyite, Fer(Zr,Y)rTiasisor4, also occurs in

the late residua of mare basalts and is important for

its high content of REE and U (Table 2' No. 2)' The

crystal structure is unknown, but hexagonal geome-

try was indicated in preliminary X-ray work (Lover-

ing et a l . , l97 l ) . Tranqui l l i ty i te broke down to bad-

deleyite, i lmenite, and pyroxene in 75055 basalt (El

Goresy et al., 1974).Zircon promises to provide useful clues to the ori-

gin of soils and breccias. As pointed out by G' M'

Brown, Zr is always high in KREEP-rich material,

and indeed he tried to popularize the acronym

KREPZ. The most abundant minor e lement is Hf

which ranges from 0.n to about 3 weight percent in

some dozen analyses (e.g. Table 2, Analysis 4)' Ter-

restrial zircons carry similar amounts of Hf, which

tends to increase from gabbroic to granitic rocks'

Various other minor elements including Al' P, Ti, Fe,

Mg, and Ca were reported, and the P-rich zircon

from 14321polymict breccia is noteworthy (Table 2'

Analysis 3). Pink, colorless, and brown grains were

found in lunar soils, and several types in breccias' A

detailed study of zircon from all rock types is ur-

gently needed to test whether it can be used to dis-

lntangle components of polymict breccias' Unlike

terrestrial zircon, the lunar varieties are unzoned, and

there is no evidence for more than one stage of

growth.Amphiboles are extremely rare in lunar samples in

conformity with the absence or near-absence of hy-

LUNAR MINERALOGY

Some representative analyses of lunar minerals

1083

TlsI-s 2

I I I5 17t ?

P^O-s io2 46 .0 14 .0T io2 o .4 19 ,5Z I O Z - L 7 . AHf02 - a .2A t z a 3 0 , 3 1 . 1

Feo 46 .0 42 .0

Mr ,o 0 .8 0 .3M e o a . 7 o . 0c a o 6 . 0 1 . 3Y ^ 0 - - 2 . 8N b ^ o - - 0 . 3Na20

Kzo

Total IOo.2 99.2

o , 6 0

3 2 . 7

6 6 . 9

a . 2 a

o . 0 9 0 . 4

0 , 0 2

0 . 1 5- n f

54 .5 42 .70 . 1 6 0 . 1

a . 7 5 1 6 . 7

12,2 14.2

o . 1 6 o . z

1 6 . 3 9 . 92 , 1 3 1 2 . 0

o . 2 52 7 . 5 7 r . 2

32.80 . 8 90 . 4 3 ) - . 9

a . 2 + r . 3

7 . 4 + 1 6 . 2

o , 0 8 o . a 2

o , o 3 8 . 8

3 . 5 51 0 . 5

0 , 1 4 . 27 r . 3 6 2 . r

r . 7 1 0 . 31 3 . 4 8 . 10 . 0 4 0 . 0 99 . 2 2 . 80 . 4 3 . 2

0 . 1 8 o . 6 r1 . 8 2 8 5 . 3

9 4 . 7 0 . 7 0

a . o t

0 , 5 4 0 , 8 2

o , 1 3 2 . 6 5

0 . 4 5 0 . 6 r .

0 . 1 7 A . a 2

o . 1 4 0 . 0 3o . 1 6 a . j 2

0 . 4 9 7 . L

4L.2 39.8

3 . B B o . 7 0

0 . 0 7 < 0 . 0 1

r . 9 5 0 . 1 04 2 . 2 5 1 . 3r . 9 5 a . 5 7

0 . 0 9 < 0 . 0 1

o J . t

J . a

4 . 2 4

3 a , 2

0 , 1 5o , 2 92 2 . A

0 . 3 04 . 3 3

4 . 3 2

1 0 . 4

0 , 1

n l

r . 2

8 . 7n f

nf

n f

99 .+

39.9

1 2 . 3

0 . 4

8 . 6 9 r . 3 2 . 7r . 4 9 a . 5 1 . 8

raa "6 99 "9 97 . r 97 .7 100 .5 9 9 . 7 9 9 . I t 9 8 . 8 9 D . a 1 0 o . 3 1 0 0 . 5 9 8 . E 9 8 . 3 9 9 . I

1 P y r o x f e r r o i t e , T a b l e 4 , C h a o e t a 1 . ( t 9 7 A ) . p T r a n q u i t l i t y i t e , l a b t e 2 , l o v e r i n g e t a l , ( 1 9 7 1 ) , U 7 o p p n . 3 Z i T c a ^ ,

m i c r o n o r i t e c l a s t i n 1 4 3 2 1 p o l y n i c t b r e c c i a , G r i e v e e t a I . ( 1 9 1 5 ) . 4 Z i r c o n , 1 2 0 1 3 p o l t m i c r K l F I p y b r e c c i a , H a i n e s e t a l . ( 1 9 7 1 ) ,

L r - 1 0 0 p p n . 5 A m p h i b o l e , f O O 5 8 b a s a t t , G a y e t a l . ( 1 9 7 0 ) , t L 2 + 0 , 3 . a A n p h i b o t e , g r a i n s f r o n 1 2 0 2 1 b a s a t t , D e n c e e t a I . ( 1 9 7 1 ) ,

4 0 . 4 c 1 0 . 2 . 7 A m p h i b o l e , b r e c c i a f r a g m e n t l n 1 4 3 1 9 f i n e s , l , l a s o n e l a l . ( 1 9 7 2 ) . I G a r n e i , f o o s e g r a i n s f r o n f 2 o 2 l b a s a l t ,

T r a l l ] e t a 1 . ( 1 9 7 0 ) . 9 z l r k e f i t e , b a s a l t 1 5 5 3 8 , t t r a r k e - ! a l . ( t 9 7 3 ) , L a 2 a 3 0 . 2 9 c e z o 3 l - . 6 3 ! r 2 o 3 0 . 4 3 N d 2 c a 2 . 1 3 s g o 3 t . 0 8

: 1 9 : ' 0 9 G d 2 0 3 1 . 4 5 T b a o 3 a . 4 2 r y 2 o 3 2 . a 9 H o 2 o 3 o . 3 8 F r 2 o 3 1 . L b T n 2 o 3 o . 1 r y b 2 o 3 r . 2 4 L u 2 c a c . 4 4 u o 2 0 . 1 4 . 1 0 A r n a l c o l i t e , b a s a l t

1 0 0 2 2 , T a b l e 2 , A n d e r s o n e t a r . ( 1 9 7 0 1 , t l " z i - a r m a 1 c o 1 i i e " , u r e c c i a 7 8 { + : , g r e e t e ( 1 9 / 4 ) , t 2 " c t - z t - a i l M r c o f i t e l , b r e c c i a

7 8 4 4 2 , s t e e l e ( 1 9 7 4 ) . 1 3 " c r - z r - a w r c o l i t e r ' , b r e c c l a ' 1 8 4 4 2 , s t e e l e ( 1 9 7 4 ) . 1 3 z , r , 1 5 1 0 2 s o 1 1 r r a g m e n t , l i a g g e r r y ( r 9 - i 2 c \ .

1 4 B a d d e l e y i t e ' L u n a 2 0 g r a i n , B r e t t e t a ] . ( 1 9 7 3 ) , . V 2 A 3 a . 0 6 . f 5 R u t i l e , l n c l u s i o n s i n i l m e n i t e o f K R F I l p y I l + 1 6 2 n i c r o b r e c c i a ,

H l a v a e t 4 L ( 1 9 7 2 ) , x 2 a 3 O . 2 2 . 1 6 u h i t l o c k i t e , 1 4 3 1 0 b a s a l t , G r i f f i n e t a L . ( \ 9 1 2 ) , F < O . f C l C . O t r € 2 0 3 0 . 7 4

c e 2 0 3 2 . r 5 P t 2 0 3 a . 3 5 l l d 2 0 3 f . l - 1 s n r 2 0 3 0 . 3 0 G d 2 o 3 0 . 5 1 D y 2 0 3 0 . 5 0 . 1 7 A p a t i t e ' 1 4 3 1 c b a s a l t ' G r l f f i n e r a r . ( \ ) 1 2 ) ' F 3 . 3 . tc L a , 2 9 L d 2 a 3 0 . 3 1 c e 2 0 3 0 . 5 I p r 2 0 3 o . o g N d 2 o 3 o ' 5 2 s n 2 o 3 o ' 1 4 0 d 2 0 3 o , o g t _ r y 2 o 3 0 . 1 1 .

drous fluids which have such strong effects on Earth.Gay et al. (1970) chipped a small grain from a vug rnmare basalt 10058 and reported the electron mlcro-probe analys is No. 5 of Table 2. This magnesioar f -vedsonite composition is surprisingly high in Na andlow in Ca wi th respect to other lunar minerals . Anion microprobe analysis for H and an electron mlcro-probe analysis for Cl would be interesting. Charles etal. (1971) synthesized the hydrous equivalent, anddiscussed the stabil ity relations for lunar amphibole.Dence et al. (197 1) reported aluminotschermakitegrains from a bag containing 12021 mare basalt,and suggested that the amphibole together withgarnet (see later) crystall ized metastably. The elec-t ron microprobe analys is (Table 2, No. 6) showslow but significant alkalies, low F and Cl, and near-equal CaO, MgO, and FeO. Mason et a l . (1972)observed loose grains in 14163 fines with chemicaland optical properties indicative of Fe-rich richterite(Table 2, No. 7) and argued against it being a terres-trial contaminant. It is surprising that further amphi-boles have not turned up in the extensive scanning-electron micrographs of vugs and in mineralogicalstudies of lunar fines. Further work is needed. Am-phiboles from enstatite chondrites are richterites witha general similarity to the lunar magnesioarfvedso-

nite, except for their very low FeO and very highMgO (Olsen et al., 1973b). A fascinating speculationis whether amphibole occurred in an early hydrousMoon which ultimately lost its water, or whetheramphibole was a lways very sparse in a Moon whichwas a lways dry.

Existing reports of mica and layer sil icates in lunarsamples are too sketchy to have any genetic signifi-cance. The possib i l i ty of meteor i t ic input in to lunarf ines must be borne in mind.

Loose gra ins of a lmandine-r ich garnet (Table 2,No. 8) from mare basalt 12021 were interpreted asrare constituents of late differentiates (Trail l et al.,1970). Of more general in terest is the speculat ion thata garnet precursor was involved in the generation ofol-opz-plag cataclasite 15445 (Ridley et al., 1973).For a l l models of the Moon which involve composi-tions dominated by olivine, pyroxene, and plagio-clase, garnet must be considered as a possible con-stituent at depth (see earlier). As the outer part of ahot Moon cooled to the present estimated seleno-therm, the boundary for garnet stabil ity could havemoved up by 50 to 100 km. Many interesting specula-tions can be made on possible changes of lunar den-sity and radius with implications for surface mor-phology (e.g. l ineaments) and heat content. If the

J. V SMITH AND I, M, STEELE

lunar interior is dominated by olivine, these effectscould be rather small, but if plagioclase coexistedwi th o l iv ine at substant ia l depth, the react ion of o l tan * gt -| px could be important. The possible role ofthe basalt-eclogite transition was discussed exten-s ive ly , e.g. by Ringwood and Essene (1970) and An-de rson (1975 ) .

The/speculation that the Moon was derived from ahigh-temperature condensate of the solar nebula wasforcefully advocated by D. L. Anderson (e.5. 1975),but runs into severe problems from phase-equil ib-r ium constra ints . Mel i l i te is one of the minerals typ i -cal of high-temperature condensates, as exemplif iedby dkermanite-rich ones from the Allende meteorite(e.g. Grossman, 1975). Eleven uniax ia l gra ins, f iveand six respectively with optical properties similar tothose of ikermanite and gehlenite, were reported byMasson et a l . (1972) in Apol lo 14 f ines, but no mel i -l i tes have been reported in lunar rocks. Further studyis needed to verify whether these grains are indeedmel i l i te , and whether they have composi t ions l ikethose of the Al lende mel i l i tes. Whether mel i l i te oc-curs in the lunar in ter ior is pure ly speculat ive. Rele-vant phase re lat ions were g iven by Yoder (1973). Ofcourse, the chance of hypothetical meli l i te from thedeep interior of the Moon being found in surfacerocks is very small. However, a detailed study oflunar fines is desirable to identify the crystals re-ported by Masson el a/.

Spinels

(a) General chemistry. There are no direct chem-

ical analyses for Fd+ and Fe'+ in lunar spinels, but

electron microprobe analyses can be calculated to

u I vijsp ine IFe2Tr 0o {Fe ,Mg)

F e l

concen l r0 l l0nSfor more bosol ls

onorthosi i ic

cnr0m lTeFeCr20a

sp rne l oiom percentMq Cr20aMg A l20a

Frc. l4 Graphical representat ions of lunar spinels and composi-

t ion ranges for spinels f rom several lunar rock types. ST spinel

t roctol i t ic , SOPC spinel-o l iv ine-plagioclase cataclasi te, DNT du-

ni t ic , nor i t ic , and t roctol i t ic . See Figs. 12, 16, and l7 for detai ls

AB,O4 compositions without invoking substantial

Fd+. This is consistent with the generally reduced

state of lunar rocks, and indeed Haggerty (1912a)

suggested that some Cr is divalent in lunar spinels

though firm evidence is lacking. The major elements

are Mg, Al, Ti, Cr3+, and Fd+. Manganese increases

from 0.0r to 0.r weight percent as Fe increases' Van-

adium (expressed as VrOr) has a similar range but

there are insufficient data upon which to base correla-

tions. Calcium and sil icon were reported at the 0.0tt

to 0.r weight percent level (e.g. Nehru et al. ' 1974)'

and presumably enter the crystal structure, though

analytical error must'be considered for small spinel

crystals. Most published analyses of lunar spinels

omit one or mote minor elements, and a systematic

study is badly needed. There is no valid evidence for

deviation of any lunar spinel from the ABrO4 compo-

s i t ion.The following scheme for calculating end members

is based on that of Busche et al. (1972) and is unam-

biguous: (a) calculate the number of metal atoms for

4 oxygens; (b) add Si to Ti, V to Cr' and Ca * Mn to

Fe g ivrng TI , CR, and FE; (c) use up the TI as

FerTIOn (all lunar spinels contain enough FE); (d)

calculate FECRTO. and assign either excess CR to

MgCRrOo or excess FE to FEAlzOo; (e) calculate

remaining Mg as MgAlrOo; (f) note whether residual

Mg or Al occurs; (g) recalculate end members to 100

percent.Both Busche et a l . (1972) and Haggerty (1971a)

plotted analyses on a Johnston type oftrigonal prism

with rectangular base between spinel (sensu stricto)

MgAlrOa (Sp), picrochromite MgCrrOo (Pc), chro-

mite FeCrrOn (Cm), and hercynite FeAlrO. (Hc),

and a roof-l ine between FerTiO, (Up) and MgrTiOo'

Unfortunately this geometry is unnecessarily com-

plex because the MgrTiOn end-member is not needed

in lunar spinels (except for skeletal crystals in a

shocked basaltic glass, Sclar et al., 1973), but is im-

plied by the method of plotting' The simplest geome-

try is a tetragonal pyramid with FerTiOn at the apex

(Fig. 15). To plot an analysis (a) calculate the metals

to a total of 4 oxygens; (b) calculate Up from

Ti(+Si) ; (c) add Mn and Ca to Fe and subtract 2 (T i

+ Si) to give Fe', then calculate mg' = Mg/(Mg +

Fe'); (d) add V to Cr to give Cr', then calculate cr' =

Cr'/(Al * Cr'); (e) plot point p in either of the two

diagrams at the upper left of Figure 14 using mg' and

u', (f) plot point 4 between p and the end member

Up using the Up component. Because data points

on a perspective drawing of the tetragonal-pyramid(middle left diagram) are not located unambiguously

hercyn i te IFeAl20a\,

unless the l ine pqUp is drawn, the triple projection(upper left diagram) is usually best for detailed study.When Ti is low, Figure 12 shows an easy approxima_tion. Calculate mg : atomic Mg/(Mg * Fe) and cr :atomic Cr/(Cr + Al), and plot on a rectangle ratherthan a square in order to fit lantern slides better.Then plot TiO, on adjacent rectangles as weight per-cent. For specimens rich in FeCrrOn, I weight percentTiO, corresponds to about 3 mole percent Up. Forspinels from mare basalts, a simple approximation isthe l ine between (Fe, Mg, Mn) (Cr, Al , V)rOn and(Fe, Mg, Mn)r(Ti, Si)On in the upper right diagram ofFigure 14. The posi t ions of lunar i lmeni te (Fe, Mg)TiO, and armalcolite (Fe, Mg) TirOu are also shown.

The large diagram in Figure l4 summarizes thepr inc ipal composi t ion t rends of lunar spinels . Thosefrom non-mare rocks l ie in two clusters, one nearspinel (s .s . ) and one t rending towards chromite. AsFe and Cr increase, so does the ulvdspinel com-ponent. Spinels from mare basalts tend to l ie in a tubebetween magnesian-aluminian chromite and ulv6spi-nel with relatively low population in between.

Crystal-structure refinements of chromite, ulv6spi-nel, and a pink Mg, Al-rich spinel showed that Crs+and Tia+ enter the octahedral B site, that Mg2+ entersthe tetrahedral I site, and that Fe2+ enters both siteswith a strong preference for the I site (Takeda et al.,1974). There are no X-ray data indicating lower thanisometric symmetry for spinels, but several authors(e.g. Haggerty, 1972a) reported anisotropic opticalreflectivity for Ti-rich spinels.

(b) Spinels from non-mare rocks. Figure 12coordinates the data which have already been dis-cussed in relation to petrologic differentiation of non-mare rocks. The analyses tend to l ie in a band fromMg, Al-rich spinels to Fe, Cr-rich ones with higherconcentrations at the two ends. Titanium tends toincrease with increasing Fe and Cr content but thereis a wide scatter which should be much greater thanexperimental error. Detailed data for an anorthositeand an anorthositic norite from Apollo l5 are shownin Figure 16.

Ranges for various types of meteorit ic and terres-trial spinels are shown for comparison.. Chromitesfrom equil ibrated chondrites (EC) and unequil i-brated ones (UC) are richer in FeCrrOa than anylunar spinel, thereby providing a possible criterionfor characterizing such types of meteorit ic spinels inlunar regolith; however metamorphism might changethe composition in breccias or grains broken out ofbreccias. Spinels from achondrites have lowerFeCrrOa, and their range partly overlaps that of spin-

LUNAR MINERALOCY

els from lunar anorthosites and mare basalts: otheraspects of the mineralogy of achondrites, of course,resemble those of lunar minerals . Some composi-tional overlap occurs with lunar spinels and Cr-richspinels from other meteorites whose compositionswere reviewed by Bunch and Olsen (1975). The Al-lende and Murchison meteorites carry spinels whichare mostly very rich in Mg and Al, but a few havesubstantial Fe and Cr (e.g. Fuchs el al., 1973:- Gross-man, 1975): further study may extend the composi-t ion range.

Terrestrial spinels carry substantial Fe8+ and tendto zone to magnetite rather than ulvdspinel. Speci-mens with low Fe and Ti are summarized in Figure12, and most have only a few weight percent of FerO.and TiOr. Spinels from deep-seated rocks, terrestrialperidotite xenoliths in kimberlite (TPXK) and terres-t r ia l xenol i ths in a lka l i basal t (TXAB), l ie in bands atlower mg than for lunar spinels except for Al-richcompositions where overlap occurs. Podiform depos-its in alpine complexes give spinels (TPAC) whichrange to lower mg values at higher Cr-content. Spi-nels from terrestrial igneous complexes of basalticcomposition cover a very wide range, and the factorswhich govern their composition are not fully under-stood. Probably the oxygen fugacity, water content,and extent of later metamorphism have importanteffects. Spinels from layered complexes produced un-der reducing conditions from tholeiit ic basalt or ul-tramafic composition tend to have cr -0.6-0.8 andmg -0.2-0.7 (e.g. Irvine, 1974), while those frommetamorphosed Archean complexes tend to havespinels with lower cr.The closest f it to low-Ti spinels

0.7

0.8mg

0'o

o.'r:'- ,,Qfi1 zt:l)

,r. '6.i$----. 3

o - c l ?

Mg A l20a 0.1c r

0.2 0 0.4 0.8 t.2wl o/" I i02

FIc. 15. mg vs. cr and wtVo TiO, for spinels from spinel trocto-l i t ic samples (dots) and f rom spinel-anorth i te-or thopyroxene-ol i -v ine cataclasi tes (squares). The numbers are keyed to the legend ofFig. 6.

1 2 c Bx t aI x J o l 2

l-l.'ls 's

.12,18r:_-..-15

1086 J. V SMITH AND I. M STEELE

from lunar dunitic, norit ic, and troctolit ic rocks is

with spinels from the Rhum layered complex whose

dominant rock is all ivalite (i.e. mostly bytownite and

olivine): however, facile comparison is to be es-

chewed especially in view of the textural and compo-

s i t ional complexi t ies repor ted by Henderson (1975)

and Cameron (1975): several other papers in Volume

39, Number 6/7, Geochim. Cosmochim' Acta areper-

t inent .With this background one can hazard a model for

the origin of the lunar spinels. Those from anortho-

sit ic rocks are remarkably high in FeCrzOo and lie at

the opposite end from the spinels from spinel trocto-

l ites and the spinel-olivine-plagioclase cataclasites,

while those from dunitic, norit ic, and troctolit ic rocks

lie in between. A possible explanation is that during

primary differentiation of the Moon, the early l iquids

were low in Fe. Cr, and Ti. Prolonged precipitation

of low-Cr, Mg-rich olivine joined later by plagioclase

resulted in ult imate build-up of l iquids relatively rich

in Fe, Cr, and Ti. The oxidation state would be high

enough to prevent formation of substantial Cr2+ ' The

sp-ol-pl cataclasites represent the early mineralogy,

while the anorthosites represent cumulates crystall iz-

ing from the later l iquids. Spinel troctolites with

Mg,nt-rich spinels might result from remelting of

early cumulates, while the norit ic' dunitic, and troc-

tolit ic spinels would represent intermediate stages'

Mare basalts would result from remelting of late

cumulates rich in i lmenite and Fe,Ti-rich spinel, and

would precipitate the chromite-ulvcispinel series

see earlier for comments on garnet).

The composition of spinels may provide a clue to

the origin of spinel troctolites. Although the available

data are scattered and probably of variable accuracy'

from the cataclasites, and differ only in the absence of

pyroxene. On the mg,TiO2 plot of Figure 15, the

ipinels from all these materials have low TiOr' The

oiher spinels tend to l ie in broad bands in which cr

and TiOz increase asmg decreases' Specimens l0 and

I 1 from breccia clasts have unusually low mg for low-

cr spinels. Detailed investigation should be made of

minor and major elements of spinels from all speci-

mens listed under spinel troctolit ic in the legend to

Figure 6.Mg,Al-rich spinel crystall izes from the l iquid in the

system SiOr-anorthite-olivine (Fig. 1), but should

rlact completely with the l iquid at equil ibrium to

produce anorthite * forsterite at lower temperature

ipig. Z) unless cumulation of spinel moves the bulk

composition out of the system' Walker et al ' (1973b)

and Hodges and Kushiro (1973) interpreted rock

62295 as the rapidly quenched partial melt from a

parental olivine-spinel-anorthite rock, in which melt

iome phenocrysts or xenocrysts of spinel' olivine, and

plagioclase were carried' Metamorphism of this and

rirnitut spinel troctolites should result in partial or

complete elimination of the Mg,Al-spinel' Because

spinel is much denser than sil icates, it should sink

rapidly in lunar l iquids thereby causing depletion of

Al rO' and MgO.Mg,Al-rich spinels from Apollo l4 and other poly-

mict breccias commonly show reaction rims against

the matrix; these rims are richer in Ti, Cr, and Mn

than the parent spinel (Roedder and Weiblen, 1972a)'

(c ) Spinels from mare rocks- Numerous composi-

tion plots exist for spinels from mare rocks (e'g'

Haggirty, 19726, 1973a, Nehru et al ', 1974), but the

situation is confusing because of incomplete correla-

tion between the zoning trends of the spinel and co-

crystall izing sil icates. In general, early. spinels are

Mg,Al-chromites which are euhedral and enclosed in

olivine when that mineral is present' Late spinel has

ulvcispinel composition and occurs either in the

groundtutt as separate crystals or more commonly

as rims on earlier spinels. The intermediate history

varies from one rock to another and from one place

to another in the same rock. The morphology of the

zoning depends on whether crystall ization is progres-

sive or whether resorption or reaction occurs' The

composition depends on competit ion with other min-

era ls .In broad outl ine, the composition range is shown

pyroxene.Many mare basalts contain spinels which zone dis-

continuously (e.g' those from Apollo l5 basalts)'

Nehru et at. (1974) discussed the existing ideas for the

A pol loboso I ts

discontinuity and favored early crystall ization ofchromite followed by cessation of spinel growth,while pyroxene continued to crystall ize followed byresumption of spinel growth producing an ulvdspinelrim. Laboratory syntheses show that the solvus in theFerTiOn-FeCrrOn-FeAlrO4 system (Muan et al.,1972) crests below l000oC for compositions with cr> 0.6, thereby ru l ing out exsolut ion as a possib leexplanation for spinels growing above I 100"C fromlunar basaltic melts. Nehru et ql. (1974\ correlatedthe presence or absence of a composition gap withtexture and composition of Apollo l5 rake samples,and for pyroxene-phyric basalts suggested that thegap is smaller for rocks which cooled more slowly.Dal ton and Hol l is ter (1974) found that in 15555 ba-salt the M'?+CrSiAlO, content of pyroxene decreasedand the M'+Al2SiO. content increased when Cr-richspinel crystall ized, and that M'?+TiAlrOu content in-creased sharply when ulvdspinel crystall ized. ElGoresy et al. (1975) used spinel zoning as an in-dicator of the crystall ization relations with olivine,pyroxene, and plagioclase in Apollo l2 and l5 marebasalts.

LUNAR MINERALOGY

o t56t0ol microgobbro

o 15105 ,5ol-phyr ic bosol i

,.- Apollo i5px- phyr icbosol tso l5 l l8 ,1 -1

l5 l l6 , t -4

Figure l6 shows details for several unusual rocks.Basalt 12036,9 (upper left) contains three types ofspinel, one of which is unusually rich in hercynite(Busche et al., 1972): the unusual spinels occur as 3grains in early Al-rich pyroxene melt inclusions. Twoof the high-Al, KREEPy basalts from Apollo 15 havespinels fall ing in the general range from chromite-richto ulv6spinel-rich compositions, but the third one hasMg,Al-rich spinels l ike those from spinel troctolites.

(d) Miscellaneous. Ulvcispinel in mare basalts isoften reduced to i lmenite + iron (e.g. Haggerty,l97 lb) and very rare ly to rut i le * i ron (ElGoresy e/al.,1972). Ti-rich spinels frequently were reduced toTi-poor spinel * i lmenite * iron, and Haggerty(1972a) suggested a reduction indicator TAC :

TiOr/(TiO, + AlrO3 f CrrO3). Experimental studiesby Taylor et al. (1972) showed that the former reac-tion can proceed under slightly less reducing condi-tions than the latter. The stabil ity curves range fromlog oxygen fugaci ty - -16 at 1000'C to - -18 at850'C. McCallister and Taylor (1973) concludedfrom laboratory studies that the reduction of ulvcispi-nel to iron and ilmenite usually involves nucleation

l0E7

high-Al KREEPybo so lts/.15382,7

a -r5366,r-r:15359,2-2

| 5308onorthos iticn0r i tet5362.1

nblinositeUp

Flc 16. Composi t ion ranges of spinels in selected lunar rocks. Apol lo l2 and Apol lo l5 data were calculated f rom atomic f ract ionsgiven in Busche et a l . (197 l \ and Nehru et a l . (19'73) using the procedure descr ibed in the text .

Sp

t205 t , 59t2039,3( )\ l

fe ldspoth icm icrogo bbro

. t5388, t0

f e ldspoth icpendottte

' 15385 ,1 -1

L\ \ *".'r*

1088 J. V SMITH AND I M. STEELE

and growth, and that i lmenite lamellae develop inulvcispinel only under conditions slightly more reduc-ing than for equil ibrium of all three phases.

The partit ion of Zr between ilmenite and ulv6spi-

nel in the presence of iron and ZrOz depends on

temperature, and favors i lmenite 3-fold near I 100'Cand 1.5-fold near 850oC (Taylor and McCallister,1972). The highest reported concentrations for lunarbasalts are 0.4 weight percent Zr in i lmenite and 0.2weight percent in ulvdspinel. Observed ratios forApollo l5 rocks indicate that some contain i lmeniteand ulvcispinel which equil ibrated down to - 850"Cwhereas others were quenched from near-l iquidustemperatures.

El Goresy et al. (1972) suggested that the partit ionof Mg between ilmenite and ulvdspinel approachesequil ibrium more closely during subsolidus reductionthan during init ial crystall ization, and interpretedchemical analyses in terms of phase equil ibrium dataofJohnson et a l . (1971).

Spinel lamellae occur in i lmenite (see later). Gay et

al. (1972) reported trails of beads about lpm acrossin olivine from Apollo l4 breccia, and Roedder andWeiblen (1971) reported rosettes up to lOpm acrossin o l iv ine in 12018 basal t . Both types of inc lus ionmay be spinel, but definit ive identif ication awaitselectron-optical study.

The occurrence of magnetite in lunar specimens ishighly controversial. Trace element data for bulk lu-nar soils indicate incorporation of - I percent carbo-naceous chondrite which should bring in some mag-netite: indeed Wood et al. (1971) found magnetitespherules in a fragment of carbonaceous chondrite in12037 f ines. However, Jedwab and Herbosch (1970)

did not f ind any crysta ls in Apol lo 1 l so i ls wi th thepeculiar morphology of magnetite from Type I car-bonaceous meteorites. Numerous studies by mag-netic, M6ssbauer, and electron spin resonance tech-niques reported in Volume 3 of the Proceedings of

both the Fourth and Fifth conferences are subject tovarious interpretations. Magnetite may have been in-

corporated into lunar soils but destroyed by reduc-tion involving solar protons or lunar-derived reduc-ing mater ia ls .

Ilmenite

Ilmenite can be found in essentially all types oflunar specimens ranging from scattered grains inanorthosite and ultrabasic materials to about 20 per-

cent of i lmenite-rich mare basalts. Particularly im-portant are its relations to armalcolite, ruti le, and Ti-

rich spinel, while the partit ion of elements between

ilmenite and sil icate may provide useful indicators of

physical conditions when laboratory calibrations

have been made.The crystal structures of lunar and terresterial i l-

menites are essentially the same in being based on

highly ordered Fd+ and Tio+ (Raymol{ and Wenk,

197 l). Whereas terrestrial i lmenites commonly con-

tain FerOa in solid solution, lunar i lmenites show no

definit ive evidence for such substitution, and indeed

there is evidence for minor substitution of TirOr.

Pavicevic et al. (1972) interpreted Ihe L X-tay spectra

of i lmenites from 10047 and 12063 basalts in terms of

Fe2+ and Tia+, but small intensity differences with

spectral l ines for i lmenite from 14053 basalt were

interpreted in terms of Tis+ substitution. Bayet et al.(1972) interpreted electron microprobe, X-ray and

Mossbauer data of terrestrial and lunar i lmenites'

and concluded that Fe8+ was present in the former

and Ti3+ in the latter. These and other data on the

occurrence of Ti'+ in lunar i lmenites should be aug-

mented by systematic study to check for possible

correlation with paragenesis and oxidation state.

Electron microprobe analyses of i lmenites from non-

chondrit ic meteorites showed no excess or deficiency

of T i (Bunch and Kei l , 1971). Most lunar i lmeni tes

appear to be stable towards reduction, but Mao et al.

(1974) reported breakdown to ferropseudobrookite

and iron, and to ruti le and iron in Apollo 17 soils.

These reactions indicate very reducing conditions(e.g. log oxygen fugacity below -16 if the tempera-

ture were 1000"C), perhaps resulting from incoming

solar protons.The main compositional substitution is Mg for

Fe2+. In i lmenites from mare basalt, MgO ranges

from about 2 to 0.2 weight percent, the amount ap-parently decreasing from early arborescent crystals to

tiny ones in the late residuum (e.g. Cameron, 1970;

Smith el al., 1970).l lmenite rims on armalcolite con-

tain up to 5 weight percent MgO, and the presence of

oriented ruti le lamellae indicates decomposition of

armalcolite. Steele (1974) found that the i lmenites

from metamorphosed Apollo l7 breccias were con-

sistently higher in MgO (4.6-7.6, average 6.5 wt. Vo)

than those from ilmenite basalts (0.8-3.5, average

1.8); this confirms the interpretation of Brett et al '(1973) for Luna 20 specimens. The former i lmenites

were chemically homogeneous and had apparently

equil ibrated with the sil icates during metamorphic

recrystall ization. I lmenites from unmetamorphosed

breccias showed considerable compositional scatter,

and presumably have diverse origins. Nehru et a/.

(1974) plotted histograms of MgO for i lmenites fromApollo l5 rake samples. The wide scatter of data formost rocks indicates disequil ibrium but neverthelessMgO of the i lmenite tends to increase with bulk rzg ofthe host rock. I lmenite from brecciated peridotitefragments in 73263 soil (Table 1, analysis 25) con-ta ins 11.6 weight percent MgO. Under equi l ibr ium,the principal factor which determines the partit ion ofMg between ilmenite and coexisting sil icates (e.g.pyroxene and olivine) is temperature. Anderson et c/.(1972) proposed a thermometer for the distributionof Mg/Fe between ilmenite and pyroxene, and gave atentative calibration based on thermochemical calcu-lation and some observations on natural minerals.F. C. Bishop (personal communication) confirmedthe general relations from laboratory syntheses, butfound considerable shifts in the proposed relations.The MglFe ratio of the i lmenite increases with bulkMg/Fe ratio of the rock and with increasing temper-ature as Mg is transferred from the sil icate to theilmenite. The lunar data can be explained satisfac-tori ly by the new calibrations, and values for equil i-bration temperatures in metamorphic lunar rockswill ensue. Interpretation of i lmenite from mare ba-salts wil l require careful examination of textures tocoordinate the crystall ization times of the i lmeniteand sil icates.

The reported ranges of minor elements in i lmeniteare MnO - 0.2-0.6 weight percant, Cr2O, - 0-1.0,CaO - 0-0.4, Al,Oa - 0-2.8, ZrO, - 0-0.4, but somevalues may be too high because of f luorescence oroverlap onto other minerals during electron micro-probe analysis. The high values of AlrO3 especiallyshould be investigated in view of the errors found inanalyses of olivine. No reliable patterns emerge fromcorrelation of the minor element concentration withrock type, but systematic study may well prove fruit-ful when erratic errors are eliminated.

The partit ioning of Zr between ilmenite and ul-vcispinel was mentioned earlier, and interpretation ofobservations for several well-known basalts (e.g.12052 and 14310) indicated cooling rates near thesolidus in the region of 100'C per day. Arrhenius elal. (197 1) noted that lunar i lmenites contain 6 to 8times more Zr than terrestrial ones, but the severalproposed explanations remain untested by laboratorysyntheses. El Goresy et al. (1974) found that CrrO,was higher (0.93-1.05 wt. 7o) in secondary i lmeniterims around armalcolite than in primary i lmenite(0.39-0.65) of i lmenite basalt 74243. The CaO con-tent of i lmenite may correlate positively with that ofthe bulk rock, but careful measurements are needed

LUNAR MINERALOGY l0E9

to test this. Because of overlap between the TiK6 andV K" l ines, V is diff icult to detect in Ti-rich minerals,but reported values range up to 0.4 weight percent.

Shock deformation of i lmenite produces multipletwinning, which can be seen in i lmenite from manylunar breccias. X-ray study of natural and experimen-tally deformed ilmenite by Minkin and Chao (1971)showed that distortion is greatest in reciprocal latticedirections perpendicular to a. Sclar (1971) producedsigmoidal bending and microfaulting in experimen-tally-shocked ilmenites; these features are common inilmenite from lunar breccias. Sclar er al. (1973) ob-tained further evidence on the twinning produced byshock deformation of terrestrial i lmenite, and foundthat optically visible twins were produced at 35-75kbar peak pressure, and that init ial twinning on{0001} was fo l lowed by twinning on l l0 i l } . This peakpressure is much lower than for production of twin-ning in pyroxenes and feldspars. Sclar et al. alsoobtained evidence of shock-induced reduction andcation disordering.

Ilmenite frequently occurs with oriented ruti le, andtextural studies suggest that most intergrowths resultfrom breakdown of armalcolite upon fall ing temper-ature (see next section) rather than reduction. How-ever, Scfar et al. (1973) attributed some occurrencesto shock reduction. Exsolution of spinel from ilme-nite was reported by several investigators e.g. Denceet al. (1971). Haggerty (1973a) noted i lmenite ex-solution in pleonaste spinel. Fuchs (1971) and Gay etal. (1972) interpreted oriented i lmenite plates in or-thopyroxene as exsolution products. I lmenite in-clusions in plagioclase from anorthosites were as-cribed to exsolution of Ti from plagioclase by Smithand Steele (1974).

Ilmenite of euhedral shape occurs in vugs (e.g. inApollo 14 breccias: McKay et al. 1972). Scanningelectron microscopy of specimens from 10072 basaltshowed growth steps and globular over-growths ofsil icate and sulfide (Jedwab, l97l). These observa-tions, and those on other mineral types testify to thepresence of a widespread vapor phase in the outerpart of the Moon.

Armalcolite, zirkelite and possible related phases

The identif ication of some fine-grained Ti- and Zr-rich minerals in the lunar rocks is uncertain becauseof lack of diffraction data. Peckett et al. (1972) madeconsiderable progress by plotting compositions on aZrO"-FeO-TiO, ternary diagram, and, Brown et ul.(1975) extended the data to eight contrasted com-positions. There is no problem with i lmenite

1090

(Fe,Mg)TiO, described in the preceding section. Theminerals zircon (ZrSiO.; HfO, 0-3 wt. 7r), baddeley-ite (ZrOz; FeO 0-7, TiO" 2-3, HfO, 1-3, YrOs 0.2),

ruti le (TiOr, NbrO, 0.n-7, Fe"O"0-8, CrzO' 0-3, ZrO,0-2) and tranquil l i tyite Fe2"+(Zr I Y)zTirSi'Ozo arewell established (see next section) by electron micro-probe and optical methods though systematic studywith diffraction techniques is needed. Zirkelite andzirconolite are two names assigned to a mineralwhich occurs in the late residuum of mare basalts andin the groundmass of KREEP-rich breccias. The ter-restrial equivalents have not been fully established,and there may be only one actual mineral for which

the name zirkelite may have priority. Chemical analy-ses of the lunar mineral show ZrOz 30-45, TiO,25 -35 , CaO 3 - l l , FeO 4 - l l , and Y rO3 3 - l l we igh tpercent with substantial amounts of REE, Th, and U(e.g. Table 2, Analysis 9). Wark et al. (1973) gave thegeneral formula A2+B24+Oz based on data for naturaland synthetic specimens. X-ray study showed mon-

oclinic symmetry. The above Zr-rich minerals occurfate in the differentiation when Zr and other "in-compatible" elements have been concentrated byprior crystall ization of sil icates. For details, see Fron-de l ( 1975 ) .

The name armalcolite was used for several kinds oflunar minerals with about 70 percent TiO, (Haggerty,

19'736). The kind occurring as a primary phase in Ti-rich mare basalts from Apollo 11 and l7 definitelyhas the pseudobrookite type of structure and con-tains l itt le Zr. Because of the reduced state, it con-tains no significant Fe'*, and its composition(Fe,Mg)TirOu is obtained from the terrestrial equiva-lent FeB+TiOu by substituting Ti + (Fe,Mg) for2Fe+ . This armalcolite exists as two textural variants(a) in both Apol lo 1 l and l7 i lmeni te basal ts as b lue-grey discrete grains with or without a mantle of Mg-r ich i lmeni te; (b) only in Apol lo 17 medium- tocoarse-grained ilmenite basalts as tan-colored dis-crete grains, either in mutual and sealed but cuspategrain-boundary contact with i lmenite or in coresmantled by i lmenite. Type (b) is frequently twinned.Haggerty (1973d) invented the terms "ortho" and"para" which were not accepted by other workersbecause (a) single-crystal X-ray data were in-distinguishable (Smyth, l9'74b), and (b) the ranges ofchemical composition overlap (Papike et al., 1974:.Taylor and Will iams, 1974). However, the grey vari-

ety apparently tends to have higher MgO and CrrO,and lower FeO than the tan variety (Taylor and Wil-l iams, 1974; El Goresy et al., 1974). The terms grey

and tan low-Zr armalcolite appear suitable.El Goresy et al. (1974) noted a correlation between

J . V S M I T H A N D I M S T E E L E

the type of armalcolite and the bulk composition and

crystall ization path of the host basalt. Perhaps the

morphology depends on the nucleation conditions.

Armalcolite shows Mg/Fe zoning which may be cor-

relatable with the composition of the host l iquid and

the crystall ization temperature. Taylor and Will iams(1974\ correlated the ColNi ratio of coexisting iron

with the presence of armalcolite, and suggested that

armalcolite occurs only in rocks depleted in Ni.

The low-Zr armalcolites (e.g. Table 2, Analy'

sis 10) range from about Mgo..rFeo r.Tirou to

Mgo ,uFeo.rTirou, and carry minor CrrOr (l-2 wt. Vo)

and AlrOe( l -3 wt . %). Traces occur of MnO, VrO.,

CaO, and ZrOr, probably always (0'2 weight per-

cent. The trivalent elements Cr and Al should sub-

stitute as Ml+TiOu, thereby reducing the atomic frac-

tion of Ti. High contents of Ti in electron microprobe

a n a l y s e s s u g g e s t s u b s t i t u t i o n o f T i ' + a s t h e

Ti;*Ti'+Ou component (Lindsley et al., 1974a) as a

consequence of the highly reduced state of mare ba-

salts, perhaps to the extent of 5 to l0 mole percent.

The alternative of structural vacancies is inconsistent

with crystal structure analyses of lunar and synthetic

armalcolites (Wechsler et al., (1975). Preliminary

analyses suggest considerable disorder of Fe, Ti, and

Mg between the two cation sites, but similarit ies in

ionic radii and X-ray scattering factors caused severe

problems in estimation of site population. Terrestrial

armalcolites from kimberlites may contain some fer-

ric iron (Cameron and Cameron, 1973; Haggerty,

1973c, 1975) in response to a more oxidizing environ-

ment .The stabil ity relations of low-Zr armalcolite were

included in the study by Lindsley et al. (1974a) of thejo in f e r ropseudob rook i t e FeT i rOu-ka r roo i t e

MgTirOu. Armalcolite is stable at higher temperature,whereas i lmenite solid solution and ruti le are stable at

lower temperature. Armalcolite begins to break down

at l l30 'C for mg 1.0 and 1000'C for mg 0 '5, and

breakdown is complete at 800'C for mg 0.5. Lipin

and Muan (1974), Lindsley et al. (1974a), and Kesson

and Lindsley (1975) showed that trF+ ions (Al, Cr

and Ti) stabil ize armalcolite to lower temperatures at

low pressure (approx. by 100'C for lunar armalco-

lites). Pressure raises the stabil ity l imit for armalco-

lite with an increase of 300"C for pure MgTirOu at

7.5kbar and about 100-150"C for armalcol i tes con-

taining substantial Al, Cr, and Ti3+ (Kesson and

Lindsley, 1975). Phase relations for Ti-rich mare ba-

salts were highly controversial, but the original prob-

lems appear to have been largely resolved by Walker

et a l . (1975) and O'Hara and Humphr ies (1975). Ar-malcolite is a primary phase on the l iquidus only at

LUNAR MINERALOGY l09l

low pressure for certain bulk compositions rich inTiO, component. Papike et al (1974) observed zoningin armalcolite from Apollo 17 basalts in which mgdecreased outwards in response to decreasing mg ofthe crystall izing basalt. Because the low-temperaturelimit of armalcolite solid-solution decreases fasterthan the temperature of the basaltic l iquid, armalco-lite ceased to crystall ize at the "suicidal" value of ng: 0.66.

Anderson (1973) reported microprobe analysesconsistent with low-Zr armalcolite and ruti le occur-ring in ol-sp-opx-plag cataclasite 15445, while Benceet al. (1974) reported i lmenite in similar cataclasitesfrom 73263. These occurrences need further studybecause the oxides may provide l imits on crystall iza-tion conditions. For example, the Mg-rich composi-tion of the 15445 armalcolite(?) would be unstablebelow I 100'C at Okbar and about 1400.C for 7kbar.

Low -Zr arm alcolite underwent several post-crys-tall ization reactions in i lmenite basalts (many papers:e.g. El Goresy et al., 1974; Haggerty, 1973b). Somearmalcolite reacted with l iquid to form ilmenite.Some underwent solid-state breakdown to Mg-richi f meni te * rut i le (e.g. Haggerty , t973b, p.781) whi lesome reacted with ulvdspinel to form ilmenite. Reac-tion of ulvcispinel with iron to form ilmenite andTirOu has also been observed. Haggerty (1973b) de-scribed the reaction of niobian ruti le and ilmenite toform a vein of armalcolite in grain 12070,35.

Usselman et al. (1975), on the basis of cooling rateexperiments, suggested that armalcolite rimmed byan oriented intergrowth of equimolar i lmenite f ru-ti le formed at depth, but with a maximum of 100 km.Ringwood and Essene (1970) found that low-Zr ar-malcolite was not stable in compositions similar tothose of i lmenite basalts above pressures of -4kbar.It seems fairly safe to conclude thatlow-Zr, Fe-richarmalcolite is not present in the lunar mantle thoughfurther experiments are desirable. Walker et al.(1975) concluded that i lmenite rather than armalco-lite was in the source region of mare basalts. See laterfor comments on Mg-rich armalcolites.

Sclar e, al. (1973) reported armalcolite rich inFe(mg 0.26) and Alros (7 wt. Vo) as minute skeletalcrystals in a fragment of shocked basalt.

The last group of Ti-rich minerals occurs as tinygrains in breccias from Apollo 14, 16, and 17. Themaximum size is only about 50pm, and no X-raydiffraction data have been obtained. Several reportsemphasize complex textural relations (e.g. Haggerty,1973b; Albee er al., 1973) and the correlation betweenmg of the sil icates and the complex oxides indicatesthat a metamorphic equil ibration has occurred

(Steele, 1974). Haggerty (1973b) classified these Ti-rich minerals as Type 2 andType 3 armalcolites anddistinguished'them chemically by relatively high con-tents of Cr, Zr, and Ca for Type 2 and Zr for Type 3.The principal ranges of analyses are as follows(Cr,Zr,Ca-rich first; Zr-rich second): SiO, 01.-0.6,0.1-0.3; TiO, 65-71, 68-72; ZrO" 4-8,0.5-4; ALO,l -2 , 0 .5 -1 .3 ; Cr ,O ' 4 - l l , l -2 ; FeO 6-13, 8 -18 ; MnO0.1-1 .3 , 0 -0 .1 ; MgO 2-4 , 7 -12 , CaO 3 .1-3 .7 , 0 .3 -0 .7 ;REE 0.5-1.5, not detected; Nbrou 0-0.4,0-0.6. Rep-resentative analyses are given in Table 2, Nos. I I and12. These ranges can be calculated approximately toM+ Mt+Oi formulae when Zra+ is added to Tin+ andaccount is taken of substitution of the trivalent ele-ments as Mt+ M1+Oi. However, the cation totalstend to lie about 0-0.15 below the ideal value of 3,perhaps because some Ti is trivalent. Steele (1974)obtained low oxide totals especially for theCr,Zr,Ca-rich specimens, and Peckett et al.(1972) had a similarproblem which led them to suggest presence of lightelements particularly Li; however, other analysts ob-tained good totals. Perhaps choice of standards andcorrection formulae is responsible. The simplest pro-posal is that these two composition clusters representvariants of the pseudobrookite structure, perhapswith a strong tendency to formation of an orderedsuperstructure. However, armalcolite and otherpseudobrookites are stable only at high temperature,whereas the Zr- and Cr,Zr,Ca-rich specimens appearto be stable under metamorphic conditions, perhapsoccurring in the range 500-900"C. Levy et al. (1972)reported isometric symmetry for reflection optics ofthe Ca-Cr-Zr variety, instead of the biaxial optics fora pseudobrookite structure, but this may result frommetamictization (unpublished X-ray data of I. M.Steele). Analogies have been drawn between theCr,Zr,Ca-rich specimens and zirkelite, but the lattercontains much more Ca and less Ti . Steele (1974)used inverted commas for the terms "Zr-" and"Cr-Zr-armalcolite", but the rules of the IMA Com-mission on New Mineral Names indicate the need forneutral terms. Peckett et al. (1972) used the termPhase X for the Ca-Cr-Zr-rich mineral, while Hag-gerty (1973b) used the term Zr-armalcolite for theZr-rich one. Only structural data can resolve theproblem, but in the meantime, we shall use "Zr-"and "Ca-Cr-Zr-titanate" as neutral terms. Haggerty(1972c) reported phase Zl in 15102, and discussed itspossible relation to zirkelite (1973b); see Table 2,Analysis 13.

Both "Zr-" and "Ca-Cr-Zr-titanate" occur incomplex intergrowths with other oxides. Albee et al.(1973) showed the minerals occurring within 50pm of

J. V, SMITH AND I M. STEELE

each other, but not in contact, in metaclastic frag-ment 61 156; Mg-rich i lmenite, Nb-rich ruti le and sil i-cates formed part of the cluster. Haggerty (1973b)described complex assemblages including (a)"Ca-Cr-Zr-titanate", Mg-rich i lmenite, chromite,ruti le, and iron and (b) "Zr-titanate", i lmenite, chro-mite, and Zl . Other assemblages were reported whichinvolved zirkelite, but not the unknown titanates. Allthese assemblages, and ones involving baddeleyite areexplainable by a sequence of solid-state reactions in-volving Ti-Cr-Zr-Fe-Mg oxides affected by changesof temperature and oxidation state. Steele (1974) de-liberately selected some Apollo l7 metabreccias withnarrow composition ranges of sil icates. These showeda regular partit ion of zg between titanate and olivinewith Ko (oxide/olivine) -0.3 for "Zr-titanate" and-0.1 for "Ca-Cr-Zr-titanate." I lmenites had similarKo to the latter. These partit ion data may be relevantto the possible coexistence of complex Ti-oxides andMg-rich sil icates in the lower crust or upper mantle,but the effect of pressure must be evaluated.

The titanates carry a substantial amount of rareelements including up to 0.6 weight percent Nbrou.The "Ca-Cr-Zr-titanates" are characterized byabout 1 weight percent total REE. Other mineralssuch as ruti le and apatite, however, contain greaterquantit ies of one or more of these rare elements.

ZrO" 0.7 but no REE in ruti le from an Apollo 14

KREEP f ragment (Table 2, Analys is l5) . Stoeser e l

al. (1974) reported a mineral from the Civet Cat

norite clast with TiOz 73 percent, Nb2Ou 20, CrrO" 4,

ZrO" 2, and FeO I percent. The highest reported

content of ZrO" is for a ruti le spinel troctolite 65785(ZrO, 3.8, NbzO, not found, CrrOs 0.6, FeO 0.7,

Al ros 1.1; Dowty et a l . ,1974b). Probably a l l Nb-r ich

specimens occur in KREEPy rocks or patches of

rocks, but the factors which govern the other ele-ments are unknown. The ruti le(?) in the 15445sp-ol-opx-plag cataclasite contained 1.6 weight per-

cent NbzOu, but only trivial amounts of other ele-

ments. Jedwab (1973) reported several types of ruti le

in lunar soils including a blue ruti le with no detect-

able N b and Zr. The blue ruti le occurred as an 0.5mm

aggregate of micrometer grains, and may have re-

sul ted f , rom condensat ion.Corundum aggregates from 10084 fines were in-

trepreted as condensates of impact-derived vapor(Kle inmann and Ramdohr, 1971). A 200pm crysta lwas found in 14163 fines by Christophe-Michel-Levyet al. (19'12), but no corundum has been reported in

lunar rocks.

Apalite and whitlockite

Whitlockite -Car(POn), is apparently more abun-

Baddeleyite, corundum, and ruti le dant than apatite -Cau(POn)'(F'Cl) in lunar rocks'Both phosphates occur in the high-KREEP rocks or

These three oxides are much less abundant than the in KREEP-rich areas of low-KREEP rocks, some-

multiple oxides. t imes as intergrowths readily characterized by cath-

Baddeleyite (ZrOr) is found in the KREEPy assem- odoluminescence. Particularly beautiful are the

blages. Although there are no systematic studies, scanning electron micrographs of vugs containing

scattered electron microprobe analyses reported hexagonal prisms of apatite or hexagonal tablets of

HfO, l -3 weight percent , FeO 0-7, and TiO, l -8 whi t lock i te(e.g. McKay eta l . ,1972). Manybreccias

weight percent (e.g. Table 2, Analysis l4). Brown et and lunar rocks contain veins and vugs l ined with

al. (1975) reported from an Apollo l7 basalt a phosphates, sil icates, i lmenite, and iron;these testify

Ti-Fe-baddeleyite with YrO, 1.8 weight percent and to ubiquitous vapors penetrating the surface rocks.

HfO, 0.8. Prewitt and Rothbard (1975) compared terrestrial,

Rutile (TiOr) generally occurs in association with lunar, and meteorit ic whitlockites by X-ray and

ilmenite. Lamellae parallel to {0112} of i lmenite in- chemical methods. Whitlockite from the Estacado

dicate exsolution, whereas irregular intergrowths of meteorite yielded a crystal structure with one phos-

Mg,Fe-rich ruti le and Mg-ilmenite indicate break- phate tetrahedron inverted with respect to the crystal

down of armalcolite (Haggerty, 1973b). Primary ru- structure of a terrestrial whitlockite: the ideal for-

ti le tends to occur as twinned, randomly-oriented, mula is CarrMgzNazP'rOu.. A terrestrial whitlockite

euhedral crysta ls associated wi th i lmeni te. Some f rom the Palermo quarry had the formula

analyses of primary ruti le are rich in Nb, and Marvin Carr r(Mg,Fe)sH'.6P14O86, while synthetic ones could

(1971) repor ted NbrO,6.4 weight percent , Cr"Or3.2, be made e i ther wi th or wi thout hydrogen

Ta"OuO.2, VrOs 0.4, LarOs0.4, and CerOs 0.8 in ruti le (CarrMgrHrP,rOu., CarsMgrPrrOuu). Unfortunately

grains from microbreccia fragment 12070,35, while Prewitt and Rothbard could not locate a lunar

Hlava et al. (1972) reported Nbrou 7.1, CrrOe 2.6, whitlockite large enough for single-crystal X-ray


study. About 80 electron microprobe analyses of lu-nar whitlockites record up to 25 elements with weightconcentrations up to two atomic percent (e.9. Table2, Analysis l6). Lunar whitlockite may have invertedfrom a high-temperature polymorph (Griffin et al.,1972), which would tolerate excess Ca and extensivesubstitution of REE. Prewitt and Rothbard (1975)believed that the old name merrillite should be usedfor the meteoritic mineral, and that a new namewould be needed for the lunar mineral when detailedX-ray work had given a proper characterization. Inthe meantime, however, the name whitlockite is re-tained for the lunar mineral . Columns l6 and l7 inTable 2 contain analyses of whitlockite and apatitefrom 143 l0 basalt, for which special care was takenwith standardization. Fluorine and chlorine are al-ways low in lunar whitlockites; SiO, is consistentlyreported at the 0.n percent level or greater; FeOranges from 0.r to 6 weight percent, while MgO tendsto vary inversely from 4 to 0.n weight percent; YrO,ranges from I to 3 percent; CezOs and NdrO, occurup to 3 percent, while other REE occur at the O.n to Ipercent level; NarO ranges from 0.0n to 0.6 weightpercent; SrO usually was not reported, but 1 percentin 15475 whitlockite (Brown et al., 1972). Probablythe composition of lunar whitlockite depends on lo-cal idiosyncrasies during the crystallization of lateresidua, and the mg ratio flor example may be a guideto the degree of fractionation during such crystalliza-tion. Vapor-deposited whitlockite may have a differ-ent composition than whitlockite crystallized fromliquid: in particular, the REE may be absent in theformer.

Some 40 electron microprobe analyses of lunarapatites record 0.0n-2 weight percent Cl and -2.5-4percent F, and when account is taken of the higherror for F there is no need to invoke the presence ofOH. However, ion microprobe analyses for H wouldbe very interesting to check for the relative availabil-ity of Cl,F, and OH. Apatites from recrystallized soilsexposed to solar protons might incorporate signifi-cant OH, and vapor-deposited apatite might alsohave a distinctive composition. The ranges of otherelements are: SiO, 0.n-2 weight percent indicatingsubstantial replacement of P by Si; FeO 0.0n-2; MgO0.0n-0.n: Y zOe 0.n-2i CerO, 0.0n-3; NdrO3 0.0r-1, U20-1000 ppm. The high U contents might permit agedetermination by ion microprobe techniques.

The relative occurrence of phosphate and phos-phide (e.g. schreibersite) depends on the redox condi-tions (next section). Fuchs (1968) recorded severalother types of phosphates in meteorites, of which

only one (farringtonite) has been observed in lunarrocks so far.

Suffides, carbide, phosphide and metals

The detailed review by Frondel (1975) emphasizedthe crystallographic and mineralogic aspects of theseminerals and allows us to concentrate on petrologicand geochemical features. Many papers discussedwhether the minerals have a lunar, meteorit ic, orhybrid origin, but simple answers cannot be expectedbecause of complex effects produced by shock andthermal metamorphism and by melting of polymictbreccias. Furthermore the bulk Moon was itselfprobably accreted from early bodies, and many typesof such bodies were available, not necessarily thesame as those now represented by meteorites in mu-seums. The extensive evidence for precipitation ofiron and troil i te in vugs, for reduction of Fe'+ to Feoby sulfur volati l ization and carbon oxidation, and forredox conditions allowing movement of phosphorusbetween oxidized minerals (phosphate and sil icate)and metall ic phases (schreibersite and iron minerals)adds to the problems. However, considerable prog-ress has been made in establishing possible criteriafor distinguishing lunar from meteorit ic phases, andat least some specimens can be assigned with consid-erable confidence. In addition, various textures andelement partit ionings are useful as indicators of meta-morphic equil ibration. An understanding of the me-tall ic minerals and alloys is important for inter-pretation of magnetic phenomena.

(a) Sulfides. The principal sulfide is troilite, essen-tially FeS. In mare basalts, troil i te can occur withblebs of iron mineral, and Skinner (1970) interpretedtroil i te-iron intergrowths in 100'72 basalt as productsof an immiscible l iquid which separated at I 140" Cfrom the sil icate magma after substantial crystall iza-tion of i lmenite and pyroxene. In Apollo l2 basalts,troil i te commonly occurs with i lmenite and spinelrather than with metall ic iron, but does occur inter-grown with iron as a late product of crystall ization(e.g. Taylor et al., 1971). Probably its occurrencedepends on a combination of factors including thebulk S content and the state of oxidation. Troil i te isone of the minerals occurring in cracks in norite78235 (Irving et al., 1974) and in the mosaic assem-blages interpreted as the final products of crystall iza-tion of troctolit ic granulite 76535 (Gooley et al.,1974), to mention just two occurrences in coarse-grained rocks. Indeed troil i te seems to be ubiquitousin lunar rocks (except perhaps for some anortho-

sites). This is not surprising when consideration isgiven to the factors governing the migration of S onthe Moon.

During primary differentiation of the Moon, a l iq-uid rich in Fe, FeS, and the metal- and sulfur-seekingelements would form wherever the temperature ex-ceeded about 1000' C (Fig. 2), and its high densitywould cause it to sink. Coexisting sil icate l iquidsshould contain sulfur at the level of at least 0.0nweight percent by analogy with data for terrestrialrocks (summarized by Bishop et a l . ,1975); indeed theS content of mare basalts indicates 0.n percent (Brett,personal communication). Crystall ization could leadto rocks relatively richer or poorer in S depending onthe efficacy of differentiation. Particularly importantfor interpretation of the sulfide and iron minerals isthe extent of post-crystall ization migration of S andFe, especially the escape of S vapor and associatedreduct ion of Fe2+ into Feo. Gibson and Moore(1973) noted a negative correlation between S andFe" (not to ta l i ron) in Apol lo l5 and l7 basal ts , andBret t (1975) found s imi lar corre lat ions for Apol lo l2basalts. The incremental ratio AS/AFe' was close tobut substantially less than the theoretical ratio ofunity for atom-for-atom reduction of Fe2+ in basalticmelt as follows:

S'?-(melt) * Fe2+(melt) : S' (vapor) f Feo (crys-tal). Disproportion of troil i te to iron mineral * Swould also require a unit ratio. Brett supported thesuggestion by Gibson and Moore that the observedcorrelations resulted principally from outgassing ofS, but the ultimate fate of the S is unknown. Sato e/al. (1973) made detailed measurements of the redoxconditions of lunar basalts and argued from ther-mochemical calculations and the expected presenceof C in the Moon that reduction of lunar rocks byconversion of carbon or carbide to carbon monoxidewould be a second plausible mechanism at the lunarsurface. Probably more than one process occurs inthe reduction of lunar rocks, but the extensive evi-dence of migration of S in lunar surface rocks and thegreater abundance of S than C suggest that S vol-atization plays the stronger role in determining thepresent redox state. Gibson and Moore (1973) foundthat 12 to 30 percent of the S and C in lunar brecciascould be volati l ized under vacuum in I day at 750" Cand a lmost a l l a t I100' C. The extensive observat ionsof troil i te in vugs (e.g. in Apollo 14 breccias; McKayet al., 1972) testify to the migration of S-bearingvapor. On an even larger scale, the migration of S inthe lunar mantle and especially from the hypotheticallunar core must be considered seriously. For the

J, V SM]TH AND I M. STEELE

Moon model in Figure 3 and the possible seleno-therm in Figure 2, S could exist as an Fe, FeS eutecticliquid at depths greater than 300km in the presentMoon and at even shallower depths before the Mooncooled down.

Troil ite is a significant constituent of many types ofmeteorites, but the contribution of impact debris tothe S content of lunar surface rocks, especially impactbreccias, is hard to evaluate because of the ease ofvolati l ization of S. About 5 percent (range I to 1Vo) ofcoarse metal-rich particles from Apollo 14 and 16soi ls conta in sul f ide (Goldste in and Axon, 1973). Al -though the Co and Ni contents of metall ic iron maynot distinguish rigorously between meteorit ic and lu-nar sources, the wide range of Co and Ni contents ofmetall ic iron mineral associated with sulfide in thesecoarse particles indicates both lunar and meteorit icsources. Spherules, 50pm to 5mm across, showed Scontents from 0.05 to 26 weight percent, and hadcomplex textures which were reproduced experimen-tally by rapid cooling of drops of synthetic l iquid(Blau and Goldste in, 1975). These spherules occur inmost lunar soils, and were interpreted as the productsof shock melting of metall ic and non-metall ic mate-r ia l , probably of both lunar and meteor i t ic or ig in.Synthetic compositions rich in both S and P appar-ently yielded two immiscible l iquids, one of whichcrystall ized to troil i te with small inclusions ofnickel-iron while the other crystall ized to nickel-irongrains surrounded by schreibersite (Fe,Ni)sP and afew blebs of troil i te: rare lunar fragments showedevidence of similar textures. Brett (1975) observed astrong positive correlation between S and Feo (zottotal iron) in non-mare rocks, and suggested thatmeteorit ic material with approximately l0 times Feoto S had been incorporated.

Troil i te from a vug was found to have a super-structure with a and c doubled over those for thesimple NiAs st ructure type (Evans, 1970).

Electron microprobe analyses of minor elements introil i te are questionable because of problems fromsecondary fluorescence and overlap into inclusions.Published data show Mg 0.02-0.13 weight percent, Si0-0.8, P not detected, Ca 0-0.6, T i 0-0.7, Cr 0-0.1,Mn 0-0.4 but mostly <0.05, Co 0-0.3, Ni 0-6.3, Cu0-0.4. High values of Co (0.3) and Ni(6.3) werefound only for troil i te occurring with iron in a glob-ule in 12001 fines (Champness et al., l97l), and allother values are 0-0.1. High values of Cu were re-ported only for troil i te coexisting with copper metalin 10045 basalt (Simpson and Bowie, 1970). Highvalues of Mg, Si, Ca, and Mn should be rechecked,

but consistently high values of Si and Ca were ob-tained for troil i te from 12051 basalt by Keil et al.( 1 9 7 1 ) .

The Ti content of troil i te is greater (up to 0.5 wt.%)when it coexists with i lmenite or ulvcispinel thanwhen it coexists with metall ic iron (0.0n) in lunarbasalts, and although fluorescence error cannot beruled out, exchange reactions may be responsible (ElGoresy et al., l97l; El Goresy et al., 1972;Taylor etal., l97l). Troil i te annealed with iron, ulvdspinel,and vapor at 700 to 1000o C showed systematic re-duction of Ti content from 0.8 weight percent at1000" C to 0.3 at 700o C, while the Ti content ofmetall ic iron dropped from 0.8 to 0.4 (Taylor et al.,1973a). Observed Ti contents of lunar troil i te coexist-ing with i lmenite could be interpreted as the result ofmetamorphism to temperatures from about 900 tosubstantially below 700" C.

Mackinawite was tentatively identif ied from itsstrong bireflectance as tiny spots associated withtroil i te from basalts 12018 and 12063 (El Goresy etal., l97l; Taylor et al., l97l). Subsolidus reactionbelow 150' C is ind icated. Taylor and Wi l l iams(1973) reported opticaland chemical identif ication ofchalcopyrite and cubanite along cracks and grainboundaries of troil i te in 12021 basalt. BothCu-Fe-sulfides may have exsolved at low temper-ature from Cu-bearing troil i te, and both contain 0.9weight percent Co. Clanton et al. (1975) reported thatin vugs of 76015 breccia the order of crystall izationfrom vapor was sil icate, i lmenite, Ni-Fe mineral,troil i te, chalcopyrite (Cr 0.5 wt.Vo), pentlandite (Fe35.5 Ni 29.1 Cu 0.3 Cr 0.2) , and chromite (?) . Sphal-erite in "rusty" rock 66095 contained 28 mole percentFeS and occurred as reaction rims in and aroundtroil i te grains (El Goresy et al., l973a,b) or as indi-vidual grains attached to or enclosed by troil i te grains(Taylor et al., 1973b). The latter authors argued thatthe uniform composition and the latter texture wereinconsistent with the idea of alteration by aZn-bear-ing solution proposed by the former authors. Theidentif ication of all other species of lunar sulfides istentative.

(b) Cohenite and schreibersite. The carbon chem-istry of lunar samples is extremely complex becauseof the input of carbon from the solar wind and be-cause reaction of carbon with iron in sil icates in-volves oxidation-reduction processes. Grains ofcohenite large enough to be visible by optical micros-copy were reported only in a few fragments in finesand 66095 "rusty" breccia. The coexisting mineralsand textures indicate that the host frasments were

1095

derived from iron meteorites. Electron microprobeanalyses showed Ni 0.8-1.7 weight percent, Co<0.02-0.25, P 0.03-0.08 (Goldstein et al., 1972; ElGoresy et al., l973a,b; Goldstein and Axon, 19731'Taylor et al., 1973b). Carbide from a melted frag-ment of iron meteorite had 6 percent Ni and up to 8percent P (Goldstein et al., l97O). Carbon from thesolar wind should tend to reach a steady level onmineral surfaces as proton stripping balances the in-coming carbon. Meteorite impact should result intransfer of some carbon from surface to interior ofgrains, while some could be converted to gaseousspecies especially carbon monoxide. Sputtering anderosion by micrometeorites should change the grainsize. Many papers (e.g. Moore et al., 1974) havereported the C and S contents of lunar specimens.Pilf inger et al. (1974) found that the content of ironcarbide in lunar fines (as measured by evolution ofCDo upon dissolution by deuterated acid) is propor-t ional to the amount of f ine-gra ined ( ( lpm) i ronmetal measured by ferromagnetic resonance spectros-copy. Both the carbide and iron metal when normal-ized to the solar wind exposure are proportional tothe bulk iron content of the fines, which indicates thatcarbide synthesis occurred only with Fd+ reduced toFeo by the solar wind. In general, lunar carbide tendsto be higher in the fines, and only a small part is ofmeteont lc ongln.

Schreibersite, (Fe,Ni)rP, is important because itforms only under fairly reduced conditions, and itsrelationship to phosphate should place a l imit on theoxygen fugacity-temperature relation. Furthermoreit is a common constituent of meteorites, especiallyirons, and its occurrence in lunar rocks might providea measure of meteorit ic contamination. Unfortu-nately solid-state reactions cause problems. The dis-tribution of Ni between schreibersite and metall iciron provides a thermometer.

The Type I enstatite meteorite fragment in anApollo l5 soil (Haggerty, 1972d) contains I percentschreibersite together with kamacite, troil i te, and ni-ningerite (Mg,Fe,Mn,Cr)S. A fragment from ApolloI I f ines containing iron mineral with Neumann lines,schreibersite, cohenite, and troil i te is probably mete-orit ic, and Widmanstiitten l ines were probably re-moved by shock (Frondel et al., 1970). A fragmentfrom breccia 10046 contained symplectic inter-growths indicative of crystall ization from a P-richeutectic (Goldstein et al., 1970). A particle fromApollo l2 fines contained coarse kamacite grains plusan intergrowth of small kamacite grains set in schrei-bersite matrix, while another particle showed den-

LUNAR MINERALOGY

r0%

drites of kamacite and taenite surrounded by schrei-bersite and troil i te (Goldstein and Yakowitz, l97l).For the latter particle, the Ni distribution betweenthe taenite, kamacite, and schreibersite (Fig. l7) in-dicated equil ibration to 500-600" C based on experi-mental data of Doan and Goldstein (1970) for theFe-Ni-P system. Metamorphism in an ejecta blanketfollowed by mechanical removal would be a possibleexplanation of the equil ibration to relatively low tem-perature. These and other data (e.g. Goldstein et al.,1972) indicate that iron meteorites were impactingthe lunar surface and that their debris was under-going complex reactions including melting and meta-morphism.

The state of oxidation of some iron meteorites wasestimated by Olsen and Fuchs (1967) using calcu-lations for the assemblage chlorapatite-metall iciron-pyroxene. These showed that schreibersite isstable for conditions more reducing than the ironwi.istite buffer, while phosphate is stable for moreoxidizing conditions. McKay et. al. (1973) heatedsynthetic sil icate glass doped with Car(PO4)2, NiO,and CorO, in a controlled stream of H, and CO2.Charges quenched from 1250' at oxygen fugacitieslower than l0-13-10 ra atm contained spherical inter-growths of schreibersite and metall ic iron interpretedas the product of immiscible Fe-P liquid. The oxida-tion conditions in lunar rocks probably range overseveral orders of magnitude around the iron-wiistitebuffer, and straddle the phosphide-phosphate equi-l ibrium. Because of the possibil i ty of growth ofschreibersite from calcium phosphates, P.idley et al.(1972) argued that presence of schreibersite was not asufficient criterion for meteorit ic contamination. Lu-nar schreibersite could result from one or more ofthree processes: (a) simple incorporation of mete-orit ic schreibersite, (b) shock heating of meteorit icphosphate (Goldstein et al., 1972), and (c) conversionof lunar phosphate. Conversely meteorit ic schreiber-site could be converted to lunar phosphate.

Gooley et al. (1973) studied the occurrence ofschreibersite and iron in Apollo l6 rake samples, andtheir data were supplemented and endorsed by thoseof Brown et al. (1973), El Goresy et al. (1973a),Grieve and Plant (1973), McKay et al. (1973), Misraand Taylor (1975), and Reed and Taylor (1974) forother Apollo l6 rocks. Figure l7a. shows the data onNi and P distribution obtained by Gooley et al.(1973), and Figure l7b shows tabulated data formany types of specimens given by various authors.Coexisting schreibersite and metall ic iron irr samplesquenched from high temperature (open circles in

J. V, SMITH AND I. M. STEELE

Figs. l7a and b) show low P in the schreibersite andhigh P in metallic iron (except for ltem 8, Fig. l7b);these compositions are close to those for the 1000" Cisotherm in the synthetic Fe-Ni-P system (Doan andGoldstein, 1970). The insert to Figure l7b shows the

wl.o//o

P

uto20304050

wt. % NiFtc. 17. Part i t ion of P and Ni between coexist ing metal l ic iron

and schreibersite. (a) Apollo l6 rake samples, Gooley et al. (1973).(b) Other data: l-4 Brown e, al. (1973), I 62295 spinel troctolite, 260335 feldspar-phyric troctol i te, 3 15082 soi l fragment, 4 68815troctolite clast in remobilized breccia (phosphide interpreted asproduct of l iquid quenched from 1080"C) 5 14310 basalt, ElGoresy e/ al (1972). 6 64455 highland basalt, Grieve and Plant(1973).7 66095 metamorphosed breccia, mean of several analyses,Ef Goresy et al. (1973a).8 66065 breccia, typical analyses forspherical intergrowths, McKay e! al. (1973). 9 66095 metamor-phosed breccia, metal-schreibersite-cohenite association, Tayloret al. (1973b). l0 14310 basalt, Axon and Goldstein (1972). l l -14phosphide-metal part icles in 14003 and 14163 f ines, Axon andGoldstein (1972). 15 Type I enstatite meteorite in Apollo l5 fines,Haggerty (1972d). l6 Apollo l2 soi l part icle F2, kama-cite-schreibersite eutectic, Goldstein and Yakowitz (1971'). 17Apollo l2 soi l part icle F3-2, kamacite-taenite dendrites in schrei-bersite-troilite matrix, Goldstein and Yakowitz(1971). See Misraand Taylor (1975, Fig. 5) and Reed and Taylor (1974, Fig.9) forgraphical data on Apollo l6 samples. Data for 7 particles from soil78501 in which iron-phosphide assemblages coexist with anortho-site, basalt, or agglutinate were given by Goldstein et al. (1974):equif ibration temperatures of 375-475'C were indicated. Inset dia-gram shows 3-phase triangles between taenite, kamacite, andschreibersite at 550, 650, and 750oC, Doan and Goldstein (1970).

o devi t r i f ied g loss. mesosiosis - rich

LUNAR MINERALOGY t09'1

experimental data for the three-phase triangle fortaenite, kamacite, and schreibersite at 750, 650, and550" C. For bulk Ni contents to the right of thekamacite-schreibersite side of the triangle, the tielines between kamacite and schreibersite stay almostparallel, whereas for bulk compositions to the left ofthe taenite-schreibersite join, the tie l ines betweentaenite and schreibersite are also almost parallel. Thedata of Gooley et al. (1973) indicate that schreibersiteand iron exchanged Ni to -600" C in mesostasis-richrocks, to -550' C in diabases, and to various temper-atures from 600 to below 550' C for rocks withpoikil i t ic texture. Reed and Taylor (1974) reported asimilar temperature range for 100 metall ic particlesfrom Apollo 16 soils and rocks. Both Gooley et al.(1973) and Misra and Taylor (1975) emphasized thatschreibersite was rare or absent in the poikil i t ic rocks,and Misra and Taylor noted that whitlockite wasunusually abundant in them. Gooley et c/. suggestedthat this resulted from conversion of phosphide tophosphate as a result of slower cooling, while Misraand Taylor pointed out that the distinctly high Pcontents of the metall ic iron in poikil i t ic rocks in-dicated a different thermal history which includedequil ibration at a higher temperature. No phos-phide-phosphate intergrowths hav€ been reported.Reed and Taylor (1974) stated that the P contents ofmetall ic iron minerals in the Apollo 16 rocks werehigher than those for metall ic iron minerals frommeteorites, and suggested that P was incorporatedinto metall ic iron minerals from meteorit ic debrisupon reduction of lunar phosphate.

Spinel troctolite 62295 has a quench texture, butthe Ni distribution between schreibersite and taenite(Brown et a l . , 1973; Misra and Taylor , 1975) in-d icates equi l ibrat ion below 550' C (F ig. l7b, No. I ) .

The coarseness of schreibersite grains (called rhab-dite when occurring as needles and elongate crystals)provides an estimate of t ime-temperature relations.Axon and Goldstein (1972) concluded that severalmetal-phosphide fragments from Apollo l4 soils hadbeen annealed to 500-600' C for periods up to Imonth to 50 years. Gooley et al. (1973) concludedthat rhabdite lamellae, 4pm wide, in poikil i t ic rock64657 would have required growth for about 2months at 600' C.

Basalt 143 l0 was controversial because of the highcontent of trace elements characteristic of meteorit iccontamination. Axon and Goldstein (1972) and ElGoresy et al. (1972) presented evidence for more thanone generation of schreibersite, one type occurring ascoarse grains in contact with iron, troil i te, and Ca

phosphate while a second type occurs as tiny grains.The assemblage plotted as Number l0 in Figure l7bapparently annealed at 700o C while assemblageNumber 5 may not be in equi l ibr ium. When accountis taken of metamorphism and oxidation-reductionreactions, the complex data are not inconsistent withderivation of 14310 by melting of a regolith (but seeCrawford and Holl ister, 1974, p.aft).

In summary, the coexistence of schreibersite andiron provides a useful parameter on thermal anneal-ing, while the relative amount of schreibersite andphosphate provides information on the redox condi-tions. Meteorit ic schreibersite can apparently be ei-ther augmented by lunar-derived schreibersite or dis-sipated by conversion to phosphate; unless it isaccompanied by other iron-bearing minerals in rec-ognizable textural or chemical assemblages, it is not areliable indicator of meteorit ic contamination.

(c) Metallic Fe,Ni,Co minerals. Introduction.Most metall ic minerals in lunar specimens areFe,Ni,Co alloys with up to 100 percent Fe, 60 percentNi, 8 percent Co, and small amounts of other sider-ophile elements. Trivial amounts of metall ic copperoccurring in some specimens (l isted by Dwornik elal., 1974) appear to be indigenous, while indium andaluminum metal probably are contaminants. Most ofthe Fe,Ni,Co alloys were characterized only by op-tical and chemical methods, and djstinction betweenkamacite (a type with low Ni) and taenite (r typewith high Ni) varieties was not always made. Al-though details of interpretation were controversial,magnetic techniques yielded much additional chem-ical information, especially for submicroscopic ironproduced by reduction. The literature is confusingbecause early data were interpreted in terms of simpleideas now seen to require further evaluation, andbecause it is diff icult to cross-relate the data obtainedby the different techniques. By 1975, it was certainthat iron minerals derived from at least four partlyinterrelated sources: (a) rocks and magmas from thelunar interior in which the composition of the ironminerals correlates with the sil icates. sulfide. and va-por phase, (b) simple meteorit ic contamination inwhich the iron minerals survived the impact, (c) com-plex meteorit ic contamination in which the iron min-erals were transmuted during the impact and mixedwith material already on the Moon, (d) reduction ofFe2+ in sil icates, sulfide, and magma by various proc-esses involving shock, escape of S vapor, addition ofH and C from the solar wind, and conversion of Cand H to escaping CO and HrO. Furthermore ironwas transported in the vapor phase, was mixed up

1098 J. V. SMITH AND I M. STEELE

and recrystall ized in breccias, and was generally re-mobil ized during all the complex processes in thelunar crust. There is no need to belabor the problemsresulting from oxidation-reduction reactions in viewof earlier comments. However, it is necessary to em-phasize the problems of determining the amount ofmetall ic Fe,Ni,Co resulting from meteorit ic con-taminat ion.

Recent evidence from oxygen isotopes (Clayton

and Mayeda, 1975) indicates that the precursors ofthe Earth, Moon, and "differentiated meteorites" (i.e.

eucrite, howardite, hypersthene achondrite, enstatitechondrite and achondrite. and nakhlite) could havebeen genetically related in the solar nebula, whereasthe ordinary chondrites (types H, L and LL), carbo-naceous chondrites (C2 and C3), and ureil i tes musthave formed separately. The simplest explanation isthat the Earth, Moon, and differentiated meteoritesformed largely from material which was originallyinteracting in the same general region of the solarnebula, whereas the other meteorites formed else-where, but other possibil i t ies must be considered in-cluding a time variation. There is considerable argu-ment and some observations in favor of present-daymeteorites arising from either the asteroid belt orcometary debris or both. Because the capture cross-section depends strongly on the relative orbital ele-ments, the relative population of meteorites impact-ing the Earth and the Moon should have changedconsiderably with time. A search for meteorit ic debrison the lunar surface must bear in mind how thesurface developed with time. In terms of our favoredmodel, most of the Moon would have accreted about-4.5Gy, and shortly afterwards severe differentiationwould have resulted in sinking of most of the Feo,Coo, Nio, and other siderophile elements into thelunar interior. The nature of the remaining metalphase in the upper mantle may be revealed by theultrabasic fragments emphasized earlier, or by themare basalts if they really derive from late remeltingof early deep-seated cumulates. Down to -3.9Gy thelunar surface would be changed drastically by a fluxof incoming projecti les of which the large basinswould provide the most dramatic evidence. Eachejecta blanket should contain some of the projecti legreatly diluted by excavated crustal materials, theratio probably varying considerably from place toplace. Scouring of the lunar surface by the debriscloud, metamorphism, reduction, and vapor transfershould cause further complications. Smaller impactswould rework the ejecta blankets, and melting wouldproduce plagioclase-rich and other rocks. Flooding

of the maria by internally-derived basalts would pro-

vide fresh surfaces which should receive only minor

amounts of earlier debris from late impacts in the

highlands. The regolith on the mare surfaces should

be especially suitable for recording meteorit ic con-

tamination since about -3Gy. In the context of this

complex situation, (a) Ganapathy et al. (1970) and

others noted the very strong depletion of certain ele-

ments including Ir and Au in Apollo I I mare basalts,and calculated that the higher abundance of these

elements in Apollo I I f ines could be produced by

about 2 percent undifferentiated material from the

solar nebula such as carbonaceous chondrite, (b)

Morgan et al. (1974) interpreted the Au, Ir, Ge, Re,

and Sb contents of rock fragments from highland

soils in terms of debris believed to have arisen from 6

basin-forming projecti les of distinctive chemical com-position, and (c) Goldstein and co-workers (latest

papers: Goldstein et a|.,1974; Hewins and Goldstein,1975b) interpreted the Ni and Co contents of lunar

metal in terms of f ive components (meteorit ic, high-Fe and high-Co basaltic, supermeteorit ic, and sub-meteorit ic-to be defined later). The (c) data were

obtained by electron microprobe analysis of individ-

ual metal grains, whereas the former data were ob-

tained by neutron activation analysis of bulk frag-

ments whose very minor component of metal

undoubtedly carries the trace siderophile metals(W2inke et al., l97l; Wlotzka et al., 1972). Partic-ularly important is the conclusion by Morgan et al.(1974) from plots of Au,Ge,Ir and Au,Sb,Re that the

basin-forming projecti les do not resemble iron mete-orites and show litt le resemblance to chondriteswhich have recently impacted the Earth. Wlotzka et

al. (1972, 1973) made neutron activation analyses of

Ir,Au,As,W and other trace elements of bulk metal

separated from Apollo 14, 15, and 16 fines and

breccias, and also made electron microprobe analysesof Co and Ni for individual particles. After reviewingseveral possibil i t ies, they concluded that, althoughthe meteorit ic component added during the past 3 to4 Gy had a siderophile content l ike that of undiffer-entiated solar material, earlier meteorit ic componenthad a different composition. Wasson et al. (1975)

concluded that earlier meteorit ic contamination was

richer in siderophile elements than material arriving

after -3.7Gy, and went on to question the identi-

fication of debris from basin-forming projectiles by

Morgan el al. Although several fragments of stonymeteorites were discovered in lunar fines (e.g. a car-bonaceous chondrite, Wood et al., l97l1' an enstatitechondrite, Haggerty, 1972d), and many iron-rich


fragments appear to be heated relics of iron mete-orites (e.g. Goldstein and Axon, 1973; Goldstein etal., 1975), most meteorit ic debris may have beentransformed during incorporation into lunar regolith.It is obvious that great caution is needed in attempt-ing to interpret the chemistry of lunar iron, and in-deed in using the term "meteorit ic."

The magnetic properties of lunar specimens arelargely determined by the metall ic iron minerals.Much iron in unmetamorphosed breccias, glasses,and fines occurs as tiny particles peppering a glassymatrix, as displayed in transmission electron micro-graphs (e.g. Agrell et al., l97O; Lally et al., 1972Housley et al., 1973). Widely-separated iron particlesabout 0.0lpm across in a diamagnetic matrix aresuperparamagnetic and show viscous remanent mag-netization (e.g. Gose et al., 1972). For increasinggrain size, the magnetic phenomena are governedfirst by a single domain (-0.03rrm) and then bymulti-domain assemblages ()lpm) typical of ferro-magnets. Gose e/ al. (1972), Cisowski et al. (1974),and others correlated the increasing grain size of ironin metamorphosed breccias with the metamorphicgrade deduced from the sil icate and oxide minerals,and Pearce et al. (1972), Tsay and Live (1974), andothers mimicked the observed phenomena with meas-urements on glassy fines or synthetic glasses annealedat 600 to 1045" C. Wasilewski (1974) pointed out thepossible magnetic effects from antiferromagnetic7-Fe,Ni, and referred to the demonstrated occur-rence of 7-iron in lunar breccias (Lally et al., 1972).Early interpretations of Mcissbauer and electron fer-romagnetic resonance spectra in terms of magnetiteor other Fe3+-bearing phases appear to have beenreplaced satisfactori ly by new interpretations (e.g.Tsay et al., 1973b) based on tiny grains of super-paramagnetic iron; however, the controversy hingeson subtle arguments involving computer simulationof resonance profi les, chemical composition, andshape, and on observed temperature variation. Somerelevant recent papers are: Forester (1973), Griscomet al. (1973), Tsay et al. (1973b), Weeks (1973),Weeks and Prestel (1974), Friebele et al. (1974), andTsay and Live (1974). Many mineralogical propertiesof lunar fines and breccias require a high state ofreduction, and any ferric-oxide phases should tend tobe reduced to either ferrous phases or iron metal.

Dunlop et al. (1973) concluded that single-domainparticles of iron, l0-20nm across, are the major car-riers of natural magnetic remanence in soil breccia14313. From detailed analysis of viscous remanence,thermoremanence and magnetic granulometry, they

also concluded that the population frequency variesapproximately as the inverse square of the grain vol-ume, and that there are two separate populations,one nearly equidimensional with coercive forces>2000 Oe and one very elongated. The former can beidentif ied with spheres enclosed in glass, and the lat-ter perhaps with needles or whiskers observed in sev-eral optical and electron microscopic studies.

Electron microprobe analyses of submicroscopicparticles are diff icult, but some analyses indicate lowNi and Co for iron particles in glasses. The Curiepoints for pure Fe, Co, and Ni are 770, 1131, and358' C, respectively. The latest thermomagnetic data(Nagata et a l . ,1974,1975) show Cur ie points of lunarsamples straddling the 770" C transition for pureiron. Because of the opposite temperature effects ofthe subsitution of Co and Ni, all interpretations arecompletely ambiguous: however, the histograms areconsistent with all lunar rocks containing some pureiron, and with the breccias, f ines, and some but notall igneous rocks containing Ni-bearing iron as well(see later). Mdssbauer resonance measurements ofthe hyperfine field allowed Housley et at. (1972) toconclude that the iron metal in 10084 fines containedless than 1.5 percent Ni on auerage and perhaps noneat all; because electron microprobe analyses of coarseparticles from 10084 fines showed particles with up tol6 percent Ni (see later), the fine iron particles shouldbe very low in Ni.

Although there are some technical subtleties, theamount of Feo was estimated from the saturationmagnetization, the ferromagnetic resonance, and theMcissbauer spectrum. The principal features of thedata given by Housley et al. (1974), Huffman et al.(1974), Pearce et al. (1974), Cisowski et al. (1974),and Pearce et al. (1975) are: (a) low values (<0.05) ofFeo,/(Feo * Fe'2+) in most mare basalts, (b) diversevalues for highland rocks (0 to 0.15), (c) relativelysmall range for lunar soils (0.02 to 0.08) no matterwhether they derive from highland or mare locations,and (d) increasing values as the grain size of the soildecreases and the grain size of the iron increases, witha specific trend for each of the Apollo 14, 16, and l7suites. The latter trends were explained by increasingreduction of the iron as a soil became more matureduring meteorite bombardment.

The formation of iron minerals in breccias andsoils must involve many complex problems (e.5. Ag-rell et al., 1970). Reduction of ferrous iron from thesil icates, oxides, and sulfides augments the iron al-ready available from pre-existing lunar rocks andmeteoritic debris. Such reduction can occur as the re-

^ Juvinos eucri te

_ o

bu lkA I lende

cosm ic r o l i o

l 100 J V. SMITH AND I. M STEELE

da

+'=

Frc. 18. Co and Ni content of n ickel- i ron: summary of mete-

or i t ic mater ia l . Note change of scale for Ni at 2Owl.Vo and for Co

a I 2 w t V alrons'. dala for bulk compositions taken from Scott (1972, Fig l6)

wi th pr incipal regions for types I , I Ia, b, t ld, I I Ia, I l lc , IVa, IVb,

and others out l ined. The unusual But ler i ron wi th Co-r ich kama-

cire (Ni 6-6 5 co I 7) , taeni te (Ni l6-45 Co 0.6) , and plessi te is

shown separately (Goldstein, 1966). Euuites and howardiles'. dala

for homogeneous kamacite grains from the Binda howardite and

Macibin i eucr i te f rom Lover ing (1964) Five grains f rom the Ju-

v inas eucr i te showed diverse bulk composi t ions ( f i l led t r iangles;

WZinke e! a l , l97l ) . l ron in micrometer intergrowth wi th t ro i l i te in

the Moama eucr i te has Ni 0.04 Co 0.16 wt.7o (Lover ing, 1975)

Duke (1965) reported Ni for Ni-Fe metal in l0 basal t ic achon-

dr i tes. but Co was not measured: 6 have Ni { l wL%o,3 have2-49o

Ni, and one contains taeni te (3890) wi th kamaci te (47o) We ob-

tained new data for iron in the Ibitira eucrite described by Wilken-

ing and Anders (1975): i ron associated wi th t ro i l i te has Ni 0.9 Co

08 -1 .1 , and t ha t w i t h i lmen i t e and sp ine l has N i 0 .1 -03 Co

0.4-0.6.Carbonaceous chondrites'. large filled circles are for metal grains

(h i gh C r 016 -1 .0 and h i gh P 0 -3 .2 w t%) usua l l y enc losed i n

forster i te crystals in b lack matr ix of Type I I (C2) carbonaceous

chondrites (Olsen el al., 1973a). Inclined crosses show composi-

t ions of kamaci te and taeni te (some as aggregates) in 7 Type I I I

carbonaceous chondr i tes (Fuchs and Olsen, t973). Kamaci te

dominates in 5 specimens, but in Al lende and Ornans the metal is

most ly taeni te g iv ing bulk composi t ions near 67 and 6l wt .7o Ni

Note that one cross for kamacite lies outside the ringed area at Ni

5 2 Co 6.9.Other meleoriter: the hexagons show compositions of bulk Ni-Fe

metal in enstatite chondrites (ec), pallasites, bronzite-enstatite

chondrites (bec) and hypersthene-olivine chondrites (hoc) as re-

ported by Lover ing (1964) f rom a Ph.D. thesis by Creenland. The

dotted l ine labeled enc shows the composi t ion range for Ni-Fe in

enstat i te chondr i tes (Kei l , 1968) which l ies at h igher Co than the

mean reported by Lover ing (1964); the hor izontal cross at Ni 3 3

Co 0.61 is for the Kota-Kota enstat i te chondr i te. The only Ni-Co

pair for an enst^tite achondri te is f or the Bustee specimen (Wai and

Knowles, 1972), but Ni analyses and general chemical propert ies

suggest that enstatite achondrites have a similar range of Co-Ni to

sult of loss of CO, S, HrO, and O, either uponmeteorit ic impact or from thermal metamorphism. In-creasing grade of metamorphism and recycling of lunarbreccias by fresh impacts should result in a tendencytowards hybridization of all the different types ofiron. Reduction of sil icate. oxide. and sulfide couldyield iron low in Ni and Co, and this could dilute theNi and Co concentrations of iron from most mete-orit ic debris and most lunar rocks. The solar windimplants the outer 0. lpm with protons and carbonatoms, and the present-day flux of protons couldreduce about lffglcm2 Fe during the past 3 Gy if noprotons were lost during the impact. However, theescape velocity for the Moon is only 2.4 km/sec, andsome impacting bodies, especially present-day micro-meteorites, should hit at such a high velocity (10-25

km/sec) that a substantial fraction of material is lost.The situation before -4Gy is especially uncertain,but processes associated with impacts must havecaused substantial reduction of iron in rocks,breccias, and fines perhaps for depths in the range ofkilometers.

The next sections describe some of the miner-alogical properties of iron metal in lunar rocks in thecontext of the above introduction. Emphasis wil l beplaced on the Co and Ni content. Data for meteorit icspecimens recently fallen on Earth are shown in Fig-ure 18 to provide a possible reference point. Forconvenience, the scales for Ni and Co change at 20and 2 weight percent respectively, causing two kinksin the l ine for the cosmic ratio Ni/Co - 20. A de-tailed explanation is given in the figure legend. Thebulk compositions of metal from irons and non-car-bonaceous chondrites have more than 5 percent Ni,and the Ni/Co trend cuts across the l ine for the

enstat i te chondr i tes (Wasson and Wai, 1970) An abstract by Goo-

ley and Moore (1973) reported general ly low Ni in d iogeni tes(13Vo;O. lVo in Gar land and Shalka; up to 527o in lbbenbi i ren) and

high Co () l%o; up to 870 in lbbenbt i ren and 27Vo in Roda). The

metal phases of the ordinary chondr i tes show complex textures and

zoning prof i les of Ni and Fe, but detai led data are not avai lable for

Co (e g . Wood , 1967 ; Tay lo r and Heymann , l 97 l ; Dodd , l 97 l ;

Goldstein and Doan. 1972\.The Ni contents of bulk metal f rom

mesosider i tes range f rom 7.0 to 9.0 wt.7o (Powel l , 1969; Wasson el

al., 1974); although no specific Co analyses were made, general

chemical re lat ions indicate s imi lar levels to those in the i ron mete-

ofltes.The star at Ni 5 Co 0.3 is the predicted value for conversion of

al l Fe to the metal l ic state in normal or carbonaceous chondr i tes(Wlotzka et a l . , 1972). Not p lot ted are the fo l lowing data for bulk

metal analyzed by neutron activation (W?inke et al ,1910): Juvinas

eucr i te Ni 29 Co 0.62, Norton County aubr i te 10.3, 027 and9. l ,

0.29, Pul tusk H-chondr i te: 9.1, 0.46, Ramsdorf L-chondr i te 13.9

0 63.

LUNAR MINERALOGY i l01

cosmic ratio. Reduction of all the Fe to the metallicstate would bring all the chondritic metal close to thestar at Ni 5 Co 0.3 weight percent (Wlotzka et al.,1972), but only partial reduction is to be expected forchondrites incorporated into lunar surface rocks, be-cause some Fe should remain in the silicates; indeedthe enstatite chondrites should become oxidized withloss of the Si from the metal to form silicate. Kama-cite and taenite in Type III carbonaceous chondritesare respectively poorer and richer in Ni/Co than thecosmic ratio, and the bulk composition of the metalfrom the Allende meteorite is very rich in Ni. Themetal from most eucrites and howardites has rela-tively low Ni and high Co, but metal from theMoama eucrite has no detectable Ni and only a littleCo. In general the metals from the eucrites and how-ardites except for the Macibini eucrite can bematched with those from the mare basalts.

Mare basalts. Two types of metal occur in marebasalts (Fig. l9). That in the Ti-rich mare basaltsfound at Apollo ll and l7 sites occurs only as blebsin troilite (see earlier). The final liquid should be verylow in Ni and fairly low in Co as a result of priorcrystallization of olivine and/or pyroxene. Althoughfew analyses were made, Co apparently ranges from1.4 to 0 weight percent while Ni decreases slightlyfrom 0.4 to 0.00 weight percent. The metal in therelatively low-Ti basalts from Apollo 12 and l5 sitescrystallized throughout the entire crystallizationrange of the silicates, and its Ni content drops fromas high as 30 weight percent for inclusions in earlyolivine to as low as 0 weight percent for crystals in thefinal residuum where a little troilite may occur. Dif-ferent rocks show different composition ranges, anddifferent investigators reported inconsistent rangesfor the same rock (e.9. for 12004). However, thegeneral trend is clear, and the data can be explainedby the expected tendency for the early ferromagne-sian minerals to extract Mg and Ni out of the liquid,allowing Fe to increase, while Co tended to be inter-mediate in behavior (Hewins and Goldstein, 1974).Idiosyncrasies may ultimately be correlated with thecomplex behavior of the other minerals, especiallypyroxene and oxide minerals (see earlier). Whetherthe negative correlations between S and Feo in marebasalts result from loss of S during crystallization orwhether they result from inheritance of properties ofthe source region of mare basalts (perhaps earlycumulates) is undecided (Gibson et al., 1975). Tayloret al. (1973a) correlated the Cr content of Fe,Nimetal in 15536 basalt with that of coexisting oxidesand other minerals; the highest values of 0.2 to 0.7

r0wr. % Ni

Frc. 19. Co and Ni content of n ickel- i ron: mare basal ts. Note

change of scale for Ni at 20 wt.7o and for Co at 2 wL.Vo. Apollo I l:

numerous general statements that Ni is very low ((0. l ) and Co

var iable between 0 and 0.7; I datum from Simpson and Bowie

(1970). Apollo l2: ranges given by different authors for same rock

do not agree completely; ranges for various rocks are outlined

from data of Brett et al. (1971), Busche e/ al. (1971), Cameron

(1971), El Goresy et a l (1971), Reid et a l . (1970),Taylor et a l .(1971 ) , Wlotzk a et a l . (1972), and Wood et a l . (1971); inc lusions in

olivines from 12004, 12009, and 12022 are shown by special sym-

bol; ranges for metal enclosed by olivine from 12004 and 12022

(Hewins and Goldstein, 1974) are st ippled. Luna 20: one datum,

Tarasov et al. (1973) Apollo 15: rake samples, Dowty et al.

(1973b); Fig 8 shows plots for indiv idual rocks; o l iv ine-phyr ic

samples tend to have Ni-rich metal compared to pyroxene-phyric

ones. Four gabbros show simi lar composi t ion ranges but extending

to lower Co contents when Ni is zero (Taylor et a l . , 1975;Taylor

and Misra, 1915). Apol lo l7: Taylor and Wi l l iams (1974), Meyer

and Boctor (1974)

weight percent were for metal coexisting with chro-mite, but f luorescence error must be considered.

Selected non-mare rocks. Figure 20 shows theCo/Ni contents of metal from some of the key rocksmentioned earlier. The metal in mosaic assemblagesof the troctolit ic granulite 76535 consists of bothhigh-Ni and low-Ni varieties. Most particles consistof one or the other, but some form a 2-phase assem-blage similar to the "clear taenite" found in ordinarychondr i tes (Taylor and Heymann, 1971) and in lunarsoils (Axon and Goldstein, 1973). A strong diffusionprofi le was found at the 76535 kamacite-taenite

b.Q

+-=

20 40

2 l

II

4fr-.-

)lz

t202

1 i . . . .i 2 /ty';/

L20 n t 20 t3

14310 P - f ree14310 P - beo r i nq

l 102 J. V. SMITH AND I. M, STEELE

, n

b-e 1.5+'=

wr. % NiFtc 20. Co and Ni content of n ickel- i ron: selected non-mare

rocks. Note change of scale for Ni at 20 wt.Va and for Co al2wt.Vo.?6535 t roctol i t ic granul i te, s ingle grains and kamaci te- taeni te as-

semblages, Cooley el al. (1974), oriented particles in the interior ofp lagioclase grains contain Ni 5- I 5 Co up to 5Vo:72415 duni te c last ,

Albee et a l . ( l97ab); 78235 nor i te, interst i t ia l grains associated wi th

whi t locki te, apat i te, rut i le , baddeleyi te, chromite, and t ro i l i te,

McCal lum et a l . (1975a); "nor i t ic" f ragments in Apol lo l7 soi lappear to be s imi lar to 78235, metal in vein and as inclusion in

orthopyroxene, I rv ing el a l . (1974);67435 spinel t roctol i te f rag-

ment with coarse texture, Prtnz et al. (1973):62295 spinel troctolitewith quench texture, metal associated with schreibersite, Brown e,

al . (1973), Misra and Taylor (1975), hor izontal crosses show indi-

v idual data; 14310 plagioclase-r ich basal t , many data, heavy l ines

outline ranges for P-bearing and P-free metal measured by James(1973), st ippled area shows range obtained by El Goresy et a i .(1971) as reported by Hewins and Goldstein (1975a, Fig. l ) , cross

shows mean value of range Ni 5.5+5.4 Co 0.8+0 4 obtained byRidfey er a l . (1972),see also Gancarz et a l . (1971); 68415 plagio-

c lase-r ich basal t , large f i l led c i rc les show data obtained by Taylor

et a l . (1973a) and Gancarz er a l . (1972): 14276 plagioclase-r ich

basal t , f i l led squares, Gancarz et a l . (1972); grani t ic part of 12013KREEP-r ich polymict breccia contained metal wi th 21.5-2370 Ni

but no data for Co, Drake et al. (1970); metal in fine-grained

aggregates of plagioclase in 60015 anorthosite believed to result

f rom shock mel t ing, Ni 0. l l -0.13 Co 0.20-0.32, Sclar and Baur

fig',l4\.

boundaries. Metal also occurs as oriented needles inthe interior but not at the grain margins of the plagio-clase. If analogy with the pure Fe-Ni system is valid,the kamacite-taenite assemblages reacted down to400oC, and the inferred cooling rate is a few tens ofdegrees per mill ion years for the range 800 to 400'C.The724l5 dunite clast contains metal with Co and Niranges which overlap those for the troctolite. Themetal from the 78235 norite and Apollo 17 "norit ic"fragments has only 2 percent Ni but relatively highCo (2.5-3.0). All these composition ranges fall in the

high-Co part of the range for mare basalts, and onemight conclude generally that the metal in the uppermantle and lower crust has equil ibrated with theFe,Mg sil icates and tends to be relatively rich in Ni

and Co. The 67435 spinel troctolite has very Ni-richmetal, while the 62295 spinel troctolite has metal with7-17 weight percent Ni and 0.7-l. l percent Co whoserange overlaps that for the 14310 plagioclase-rich

basalt. Unfortunately the data for this basalt aresomewhat inconsistent. James (1973) showed tworather wide ranges, one for P-free metal fairly close tothe cosmic l ine and centered on Ni l4 Co 0.9, and aparticularly wide range for P-bearing metal which

tends to straddle the cosmic l ine and which extends topure Fe. Hewins and Goldstein (1975b) explained anarrow trend obtained by El Goresy et al. (1971) in

terms of fractional crystall ization involving simulta-neous crystall ization of metal and sil icate. The meanvalue of 14310 metal given by Ridley et al. (1972)

appears relatively high in Co. Data for metal inplagioclase-rich basalts 14276 and 68415 mostly l ie in

the range for 14310 metal. Perhaps all three are de-rived from melted regolith, and are somewhat in-homogeneous.

Plagioclase crystals in Luna 20 fines also containrods of Fe about 0.5pm wide and several to l00rrmlong (Brett et al., 1973 Bell and Mao, 1973). Up tofour orientations of the rods occur. Rounded blebsformed when the rods reached the surface. Electronmicroprobe analyses revealed about I percent Ni(Brett er al.) and 2.1 percent Ni (Bell and Mao).

Exsolution of Fe from plagioclase is the favoredmechanism for growth of theSe rods because theyoccur in the interior but not at margins of plagioclasegrains in the 76535 granulite. Presumably Fe exsolvedfrom grain margins migrated into the mosaic assem-blages. Exsolution of Fe requires a reduction mecha-nism. Many lunar rocks, fragments, and clasts con-tain veins or intergranular clots of complexassemblages containing several of the followingphases: iron, troil i te, baddeleyite, apatite, whit-lockite, several sil icates including diopside. Perhapsthe whole crust is pervaded by vapors which react insubtle ways with the main minerals. Metal grains inthe granitic part of 12013 KREEP-rich polymict

breccia contain 22 percent Ni (Drake et al., 1970),and the Co content should be measured.

Anorthosites (sensu stricto) contain l itt le or nometal. Most workers did not report iron metal in the15415 anorthosi te (Genesis Rock) , but Hewins andGoldstein (1975b) located very tiny grains devoid ofNi and Co, possibly attributable to shock reduction.

Indeed Sclar and Baur (1974) reported only 0.1 Niand 0.3 weight percent Co in metal enclosed in fine-grained aggregates of plagioclase believed to haveresulted from shock melting of 60015 anorthosite.Figure 2l shows ranges of Co and Ni for metal in"anorthositic" rocks studied by Hewins and Gold-stein (1975b). Those for coarse-grained specimensmostly fall in the range for metal in mare basalts (Fig.19), whereas those for recrystallized specimens tendto be richer in Ni and lower in Co. Considerableoverlap occurs for the latter and the metal in theplagioclase-rich basalts and 62295 spinel troctolite(Fig. 20).

Fines. The iron in the fines ranges from coarsefragments up to - lmm across down to sub-microscopic particles enclosed in glass. As describedin the introduction, the latter appear to be essentiallypure iron formed by reduction of silicate. Many mor-phologies were reporte d (e.9. Frondel et al., 1970).This section is concerned principally with the coarseparticles for which extensive metallographic and elec-tron microprobe studies were made mostly by J. LGoldstein, F. Wlotzka, and their respective co-work-ers. Figures 22,24,25, and 26 summarize the analysesof the bulk Co and Ni for several hundred particlesfrom soils at the Apollo sites, while Figure 23 showsthe assignment scheme proposed by Goldstein et a/.(1974). Because of space restrictions, we have com-bined together data for soils of similar nature, butthere are subtle differences discussed in detail byGoldstein et al. Although we agree with the generalconclusions, we adopt a rather more sceptical atti-tude to assignment of metal in lunar fines. Based onthe data plotted in Figures 18, 19, 20, and 21, weconclude that (a) the high-Fe and high-Co basalticranges are continuous, and the lower limit of thehigh-Co basaltic field should be placed horizontallyat Co - I weight percent, (b) the meteorite fieldapplies only to a restricted group of relatively "un-differentiated" meteorites with ColNi near the cos-mic ratio, and does not encompass the very differenti-ated eucrites and howardites which tend to overlapthe lunar basalts, (c) non-basaltic lunar rocks such asanorthosite, norite, and granulite contain iron in thebasaltic high-Co range, (d) recrystallized rocks suchas 14310 and the recrystallized "anorthositic" rocksof Figure 2l occupy wide ranges of uncertain mean-ing. Furthermore we note that shock-reduced lunariron has low Ni from magnetic data, and probablyfalls in the high-Fe basaltic range of Gol dstein et al.

Metallic spherules, 50pm-5mm across, fromApollo ll, 12, 14, 15, and 16 fines show complex

I 103

wt. % NiFrc. 21. Co and Ni content of nickel-iron: feld-

spar-rich (anorthositic) specimens. Note change ofscale for Co at 2 wt.Vo. Described in abstract (Hewinsand Goldstein, 1975b). Full lines show ranges forcoarse anorthosites and dashed ones for recrvs-tallized anorthosites.

microtextures which were studied systematically byBlau and Goldstein (1975) using optical and scanningelectron microscopy plus electron microprobe analy-sis. Cooling rates were deduced by comparing tex-tures with those of quenched droplets of syntheticmolten alloys. About 30 percent of the lunar spher-ules had dendritic or cellular textures indicative ofrapid quenching, and Blau and Goldstein concludedthat 6 out of 65 spherules partially or completelysolidified by radiation loss during flight in the lunarvacuum, that 3 solidified upon conduction loss to asilicate substrate, and that about 9 cooled by bothradiation and conduction. The remaining 70 percenthad globular metal areas, and some had solid-stateprecipitates; a variety of processes including rela-tively slow cooling or even reheating in an ejecta

LUNAR MINERALOGY

a(-)

b.S

+=

t5455

6443.5 ;78t55

i' l, ..rr,r,l i i T r t r ' t ' tt t ! I| / ! L - - - - ' .' 1 , " '

, t /

.//

(oo9-" ,,,Irt'z'5

x2.35

lunor spheru les - - { 'x

x

).i,,

"

X x . t t x x

. . . x . t I x x x

x txxx x x xx)40( x F( xx x

n$( xx* ) f f i x- x x

X z 'x

t.0o

c)

bs# 0.5=

n +u5t0 1520

wt olo NiFrG. 22. Co and Ni content of n ickel- i ron: lunar spherules. Bulk

composi t ions. Many textural and chemical detai ls in Blau and

Goldstein (1975). One datum at Ni 0.2 Co 2.35 omit ted

blanket could be possible. The Co and Ni contentsmostly lie in a band (Fig. 22) from Ni 6 Co 0.35 to Ni20 Co 0.8, which closely matches the band for ironsand non-Qarbonaceous chondrites (Fig. l8), includ-ing the tendency for low Ni specimens to have moreCo than the cosmic ratio while the high Ni ones haveless Co than the cosmic ratio. The simplest ex-planation is that the spherules mostly result frombulk iron in a range of meteorites, and that con-tamination with lunar iron is insignificant. Threespherules with Ni ( 5, one with high Co(2.3), andtwo with low Co(0.2) may result from lunar or mete-oritic sources or a mixture of both; for example theone at Ni 3 Co 0.25 falls near the range for theMacibini eucrite.

DOS;high-Fe

wt%NiFrc. 23. Metal classification of Goldstein et al. (1974) based on

Ni and Co content. Abbreviated as M (meteoritic), SM (super-

meteor i t ic) , Fe (basal t ic h igh-Fe), Co (basal t ic h igh-Co), and l -4(submeteoritic with Ni l-49o and Co below basaltic high-Co). See

their F igures 6 and 7 for h istograms of metal in lunar soi ls .

J. V, SMITH AND I M. STEELE

The S and P contents average at2.7 and 3.5 weightpercent with most values below 8 and l0 weight per-

cent. All six radiation-cooled spherules are S-rich.Phosphorus could have derived from either reductionof phosphate or from original phosphide. Experimen-tal data indicated that Fe,Ni,Co,S,P alloy forms twoimmiscible l iquids upon melting. This explained boththe sulfide husks found around many lunar spherulesand the enrichment of Co and Ni in the phosphide-rich areas. That the spherules occur in all Apollosoils, irrespective of whether they are of mare orhighland type, confirms assignment to a meteorit icorigin. Because most spherules have Ni > 5, hybrid-ism with reduced metal from lunar sil icate sourcesshcfuld be small. Furthermore the extreme paucity ofhigh-Co metal argues against formation of spherulesfrom lunar rocks except in rare instances. This studyof large lunar spherules gives confidence in attempt-ing to interpret the more complex data for other typesof i ron in lunar soi ls .

Blau and Goldstein (1975) l isted the many refer-ences to small metal spherules in and on glasses ofwhich the mounds revealed by scanning electron mi-croscopy are particularly spectacular. Although re-quiring further study, it is l ikely that the smallerspherules tend to be poorer in Ni and Co, and to haveformed by reduction of Fe2+ in sil icates. Carter andMcKay (1972) simulated the mounds experimentally.

As a crude approximation, we brought together thenumerous data for large irregular particles in finesfrom highland sites (Apollo 14, Apollo 16, and Luna20; Fig 24), sites with high-Ti mare basalt (Apollo I Iand l7; Fig. 25), and sites with low-Ti mare basalt(Apollo 12 and l5; Fig. 26).lt must be emphasizedthat in the later Apollo missions, the fines from thedifferent subregions should and do vary considerably.The crystallization ages of the basalts at the sites areApollo 14 (Al-rich) -4.OGy, Apollo I I (Ti-rich)-3.8 to -3.6, Apol lo l7 (T i - r ich) -3.8 to -3.7,

Apollo 12 (quartz and olivine) -3.4 to -3.2, andApol lo l5 (quar tz and o l iv ine) -3.4 to -3.3 (Taylor ,1975). Naively one might test whether meteorit ic con-tamination had changed with time by comparing themetal from the various fines, but the uncertainties areimmense. All the plotted data are for large metalparticles mostly near 100pm across, but the size pop-

ulations are different for each soil (e.g. Apollo 14 andl5 soils, Wlotzka et al.,1972, between 20 and l00pmafter grinding the soils; Apollo 15, Goldstein andAxon, 1972, 5}-250pm; Apollo 16, Goldstein andAxon, 1973, over 200pm).

For these relatively large, irregular particles, the

o(-)

be+'=

"":sGHA clossi f icoi ion

l f r a '

- - -aglgoritic

LUNAR MINERALOGY

t 6o r . w

()

d-e

E r.0

I r05

simple distribution for spherules is replaced by com-plex distributions. Immediately obvious is a strongtendency for all soils to have a concentration of metalparticles with bulk compositions near Ni 6 Co 0.45,and indeed the tendency is so high that individualpoints could not be plotted therein without overlap.Actually the cluster for each Apollo site occupies aslightly different region, and the clusters for the oldersites tend to center about Ni 6, whereas those for theyounger Apollo 12 and 15 sites center about Ni 5.The reason for the clustering is unclear. Certainlythere is no simple match with the spherules, and onemight speculate that an early meteorit ic componenttends to be dominated by relatively Ni-poor material,whereas a later meteorit ic component shows a widerange of Ni l ike that for meteorites now impactingthe Earth. Wlotzka et al. (1972) pointed out thatcomplete reduction of all Fd+ to Feo in a "normal orcarbonaceous chondrite" would give iron with 5 per-cent Ni and 0.3 percent Co (star in several of Figs.l8-26.) Neutron activation analyses of Co, Cu, Ga,Ge, As, Pd, Ir, and Au for bulk metal from 14163fines matched quite well with calculated values foreither a reduced carbonaceous chondrite or for ahexahedrite. The details are complex, but in generalthe agreement results because all these types of mete-orite are relatively undifferentiated.

The highland soils contain very l itt le iron with Co) l, indicating l itt le contribution from high-Co ba-saltic, anorthositic, norit ic, and granulit ic rocks. Thefairly strong concentration of pure iron may resultmore from reduction of Fe'+ than from contributionof high-Fe basaltic component. Most of the data fallin a parallelogram from Ni 0 Co 0 weight percent to10,0.3; 10,0.9; and 0,0.6. Much of th is range over lapsthose for recrystall ized "anorthosites" in Figure 2land feldspar-rich basalts in Figure 20. The simplestexplanation is that the iron of these highland soilsconsists mainly of a complex mixture of metals froman ancient meteorit ic component and a variety offeldspar-rich rocks. Perhaps thorough study of thechemical composition of sil icates and glasses adher-ing to some of the metal particles would allow furtherprogress to be made.

The soils from Apollo 12 and 15 mare regionscontain much Co-rich metal, and there is a highconcentration for 15601 soil centered on Co 1.3 Ni0.25. A high concentration of pure Fe occurs in allthese soils. Soil 15601 obtained from the edge ofHadley Ril le is shown separately from the otherApollo l5 soils because it has no concentration ofmetal grains near Ni 5 Co 0.5 and is presumably

h igh lond so i l s

/ \ . 14003,14163

U x 63501,65701,68501o L20

xx

2 t x/ x

. . x tF . . ^

x { ( } . a

u 5 l0 15 20 30

u/t % NiFrc. 24. Co and Ni content of n ickel- i ron: soi ls f rom highland

si tes. Note change of scale for Ni at 20 wtVo and for Co at2wt.Vo.Apol lo l4 (Goldstein et a l . , 1972; Wlotzka et a l . , (1972). Apol lo l6(Goldstein and Axon, 1973). Luna 20 (Goldstein and Blau, 1973).The r ing out l ines a very heavy concentrat ion in a l l three Apol lo l6so i l s .

2.0

o 1.5

5r015wt . % Ni

FIc. 25. Co and Ni content of n ickel- i ron: soi ls f rom Apol lo I I

and 17. Note change of scale for Ni at 20 wt.Vo and for Co aI 2

wt.7o. 10084 (Wlotzka et a l . , 1972).78501 and 78510 (Goldstein e l

al., 1974). Strong concentrations are ringed.

2.0

bQ

#=

I t06 J V. SMITH AND I, M STEELE

J

Uu 5 t0 15 2030wt . % Ni

FIc 26 Co and Ni content of n ickel- i ron: soi ls f rom Apol lo l2and 15. Note change of scale for Ni at 20 wI.Vo and for Co at 2wI.Vo. 12O01 (Wlotzka et a l . , 1972). Apol lo l5 (Goldstein andAxon, 1972) Strong concentrat ions are r inged. The one labeled15071 occurs only for that soi l . The concentrat ion at the or ig inoccurs for a l l soi ls . See Axon and Goldstein (1973) and text fordata on 2-phase Co-r ich part ic les

dominated by mare basalt. The soils from Apollo I Iand l7 mare regions contain a wide range of metalcompositions even though the local rocks are high Tibasalts whose metal has Ni -0 and Co ( - l. Sub-stantial contribution from either lunar non-marerocks or "differentiated" meteorites is indicated; ofcourse, fragments of feldspar-rich rocks includinganorthosites and norites have been reported by manyworkers.

Goldstein and co-workers reported many texturesand chemical variations too complex to discuss herein detail: among those reported are structurelessmetal (usually kamacite), kamacite * isothermaltaenite, ragged d2, acicular martensite, zoned taenite,d + I t interstit ial phosphide, metal * massivephosphide, metal * phosphide eutectic, and metal *sulfide. In general, some grains have equil ibrated tolow-temperature assemblages such as coexistingkamacite and taenite, whereas others show differentlevels of adjustment to lower temperature. Partic-ularly important was the conclusion (Axon andGoldste in, 1973) that some high-Co par t ic les inApollo l5 soils had annealed to 350'C at conditionsprobably found only at a depth of l0-20km. Sil icon

contents are below 0.2 weight percent, thereby rulingout simple addition of Si-rich metal from enstatitechondrites. Sulfur and P are also very low except formetal particles quenched from high temperature.Tungsten contents are much higher than for metal in"undifferentiated" meteorites (Wlotzka et al., 1972),and addition consequent upon reduction of W inlunar rocks is indicated. Chromium is usually low,but one particle from Apollo l4 soils contained l7weight percent Cr (Goldstein et al., 1972).

Because most data were obtained on coarse par-ticles or on bulk samples, the role of the smallerparticles is rather undefined. Electron microprobeanalyses could be extended to particles only a fewmicrometers or less in diameter. Systematic study ofthe composition and texture of adhering or contain-ing sil icate or glass is mostly lacking, though Gold-stein et al. have provided some qualitative indications(e.9. Goldstein and Axon, 1972). Because silicateshould be more abundant than metal in the averageproduct of the solar nebula, several percent of mete-orit ic sil icate should occur in the lunar regolith. Per-haps criteria can be developed for recognizing suchmaterial when attached to iron of meteorit ic origin.

Breccia. Although there are many data on thecomposition of iron in breccias, we have not foundany obvious patterns different from those in fines. Asmentioned repeatedly, metamorphism of mixed de-bris has yielded very complex materials ranging fromweakly metamorphosed breccias up to partially-melted rocks. Many of the fines are merely disruptedbreccias. Recrystall ization and vapor transport havemodified the iron metal in breccias, and the grain sizeand chemical properties give information on themetamorphic grade. Detailed studies of iron and as-sociated sil icate and other minerals in relatively un-metamorphosed breccias may allow decipherment ofearly processes at the lunar surface.

Conclusion

This review of the mineralogy of the Moon has

been set deliberately in a petrological-geochemicalcontext with auxil iary use of geophysical parameters.

Although the mineralogy can be interpreted plausibly

in the context of an early differentiated Moon whose

surface was extensively reworked by various igneousand metamorphic processes, the provenance of mostsamples is too uncertain for definit ive interpretations.The clues provided by the ultrabasic and basic rocks

and fragments are very tantalizing, and combinedwith the equally fragmentary and tantalizing geophy-

sical data, allow speculation about the lunar interior.

43

2n

. O t2ootx l 5 6 O l

o l 5c) , . "

b.e

# , ^= t .u

n 5

In recent years increasing emphasis on igneous andmetamorphic processes in meteorites has beencoupled with theoretical arguments for catastrophicdynamics to favor high-temperature processes inearly bodies in the solar system. Such high-temper-ature processes unfortunately tend to reduce thenumber of clues to the mineralogy of the early solarsystem and to the origin of the planets. In a gooddetective story, the clues lead inexorably to a uniqueconclusion, but the case of the Origin of the Moon issti l l open. The lunar detectives have hauled in ex-cellent clues, but many remain undiscovered. Furtherexploration of the lunar surface with detailed samplecollection is needed when reconnaissance has beencompleted of other targets in the Solar System; per-haps mankind wil l learn the futi l i ty of expenditureson dreadful war machines and wil l direct its aggres-sions to colonization of parts of the Solar System.Scientif ic study of the Moon might then enter a newphase in which field and laboratory studies wouldconsolidate and extend present knowledge. In themeantime substantial advances should come from thefollowing studies of lunar mineralogy: (a) careful sur-face study of all lunar samples now in storage forfurther coarse-grained rocks or clasts, (b) detailedstudy of all the Mg-rich specimens, especially the"periodotit ic" ones, with particular emphasis onthose minor and trace elements which undergo sub-stantial fractionation, (c) coordinated study of un-metamorphosed polymict breccias to search for min-eralogic and geochemical clues leading to a hierarchyof processes, (d) study of redox conditions as a func-tion of crystall ization age to test for an over-all ten-dency to progressive reduction, (e) detailed study ofthe chemistry of coexisting iron and sil icate mineralsin fines and breccias to obtain further clues to theamount and type of meteorit ic contamination as afunction of t ime, (f) laboratory studies of potentialgeochemical indicators such as the partit ion of transi-tion metals between sil icate, metal, and liquid as afunction of temperature and redox state. These andmany other studies to be invented by ingenious in-vestigators should reduce the uncertainties so oftenmentioned in this review.

Few scientific endeavors can have been so excitingand so full of good fellowship as the study of themineralogy of the Moon. We are greatly indebted toour scientif ic colleagues for their bri l l iant studieswhich made this review possible. Scientif ic historianswill probably evaluate the scientif ic exploration of theMoon as one of themost productive advances in thegeological sciences and indeed in the whole of sci-

I 107

ence . Certainly in the specific f ield of mineralogy, theimplications of lunar mineralogy for the origin anddevelopment of meteorites and planets wil l be minedfor a very long time indeed.

Perhaps some readers wil l feel that we have at-tempted to make too much of the slender clues inlunar minerals to which we reply with a phrase fromRobert Browning's famous soli loquy by the Floren-tine painter Andrea del Sarto: "Ah but a man'sreach should exceed his grasp/Or what's a heavenfor . "

Appendix: Additional lunar minerals

The following minerals extend the l ist in Part I:cordierite (Albee et al., 1974a), inclusion in spinelgrain of a plagioclase-rich l ithic clast of sample72435, no chemical or optical data listed; molybdenite(Jedwab, 1973), micrometer-size particles in fines,probably contaminant from lubricant in sample cabi-nets; farringtonite (Dowty et al., 1974b), (4pm grainassociated with nickel-iron. troil i te. and a Ca-sil icate(probably plagioclase) all enclosed by spinel in spinel

troctolite clast 65785, chemical analysis showed 5lpercent P2OE, 4l percent MgO, 4 percent FeO;' aka-ganbite (Taylor et al., l974a,b), identif ied by X-rayand optical spectra in the "rusty" material previouslyascribed to goethite(?) in references l isted in Part I,attributed to terrestrial oxyhydration of probablyprecursors Fe metal and lawrencite in Apollo l6 spec-imens; unidentified Cl,Fe,Zn-bearing phases, probablya sulfate and a phosphate associated with akagan6ite;still uncharacterized mineralogically (El Goresy et al.,1973b see also Taylor et al., 1973b); bornite(2) (Carter and Padovani, 1973), chemical identif ication dur-ing SEM study of metall ic spherule in 68841 fines,occurs with FeS between interlocking crystals ofschreibersite and metallic iron; aluminum oxycarbide(?) (Tarasov et al., 1973),20pm grain in troctolit icfragment from Luna 20, electron microprobe analysisgave 74 percent Al, 20 percent C, and 4 percent 0;monazite (Lovering et al., 1974), l0pm grain identi-fied chemically by electron microprobe from mesos-tasis of 10047 mare basalt; thorite(?) (Haines et al.,1972), abstract reporting chemical identification ofone U-rich grain in 14259,97 soil; pyrochlore (2)(Hinthorne et al., 1975), abstract describing chemicalidentif ication of 5pm grain in 76535 granulite, 20 per-cent UO, + ThOr. Titanite, alias sphene, reported inPart I as a doubtful mineral, was chemically identifiedby Grieve et al. (1975), and occurs as anhedral, (5pm

grains associated with i lmenite and troil i te in basalticmesostasis of polymict breccia 14321. Zr-rich phases

LUNAR MINERALOGY

I 108

show even further complexities, and Brown el al.(1975) now record 8 contrasted compositions forzirkelite-type phases. Graphite was found with FeSas a nodule in an a-iron fragment of soil 14003 (Gold-stein e/ al., 1972).

Acknowlegments

We thank all our colleagues at the University of Chicago for

help and advice, part icular ly E Anders A. T. Anderson, R. N.

Clayton, R Ganapathy and L. Grossman. Technical help f rom l .Bal tuska, R. Banovich, O. Draughn and T. Solberg was much

appreciated We thank B. Mason and R Bret t for helpfu l cr i t ic ismand the Publ icat ions and Execut ive Commit tees of MSA for theird i f f icul t decis ions on th is unusual ly long manuscr ipt . Generousf inancia l a id came from NASA grant NGL l4-001- l7 l Res, but theresearch could not have been carried out without electron micro-probe and X-ray facilities built up from NSF grants including thegrant to the Mater ia ls Research Laboratory (NSF DMR72-03028A04). The cost of publishing this paper has been met by a specialgrant f rom the Nat ional Aeronaut ics and Space Administrat ion.

ReferencesThe followrng reference abbreviations are used

throughout:Apollo l5 The Apollo I5 Lunar Samples (Lunar Science lnstitute,

Houston, Texas)

LS III, IV, V, VI Lunar Science III, IV, V,Inst i tute, Houston, Texas)

J V. SMITH AND I M STEELE

- . T. F. Bneztures, J JACoBy eNo J. V. SurrH (1972)Thermal and mechanical h istory of breccias 14306, 14063,14270, and 14321. PLC J, 819-835.

- AND J. V Sunu (1971) Nature, occurrence, and exot icor ig in of "gray mott led" (Luny Rock) basal ts in Apol lo l2 soi lsand t t reccias. PLC 2,431-438.

- AND l0 Orsens (1970) Armalcol i te: a new mineral f romthe Apol lo l l samples. PLC I , 55-63.

ANDERSoN, D. L. (1975) On the composi t ion of the lunar inter ior .J Geophys. Res 80, 1555-1557.

AnnHrNtus, G., J . E EvrnsoN, R W. FrrzcrnALD AND H. FuJrrA(1971) Zirconium fract ionat ion in Apol lo l l and l2 rocks PLC2 , 169-t7 6

Axou, H. J. nNo J. l . Got-osruN (1972)Temperature- t ime rela-t ionships f rom lunar two phase metal l ic part ic les (14310, 14163,140O3). Qarth Planet Sci Lett. 16, 439-447.

- AND - (1973) Metal l ic part ic les of h igh cobal tcontent in Apol lo l5 soi l samples. Earth Planet . Sci . Let t . 18,I 73- I 80.

Blsu, A R. nNl I . D. MACGREcon (1975) Chromite spinels f romultramafic xenoliths Geochim. Cosmochim Acta, 39, 937-945.

Blvsn, C. , J Fsr-scHr, H. ScHur-z nNo P. RUrcsrccen (1972) X-ray study and M6ssbauer spectroscopy on lunar i lmeni tes(Apol lo l l ) EarthJlanet . Sci Leq 16,273-274.

Berr , P M rNo H. K Mno (1973) Opt ical and chemical analy-sis of iron in Luna 20 plagioclase. Geochim Cosmochim Acta,31 ,755 -759 .

- AND - (1975) Cataclast ic p lutoni tes: possib le keys

to the evolut ionary h istory of the ear ly Moon. L,S VI ,34-35.

See a l so PLC 6 .231 -248 .Btr lcs, A. E. er . rp B. Aurrrn (1972) Secondary ion analysis of

Z1 (Lunar Science pyroxenes from two porphyritic lunar basalts. Apotlo t5,t 9 r -194 .

PLC I Proceedings Apollo Il Lunar Science Conference, Geochim. _, J W DELAN., J. J. peprrs AND K L. CerunnoN (1974)Cosmochim' Acta Suppl

- petrology of the highlands massif at Tauris-Littrow: an analysisPLC 2,3,4,5,6 Proceedings 2nd etc. Lunar Science Conkrence, of th" i_a.. soil fraction. pLC 5,.:g5-g21..

Geochim. Cosmochim. Acta Suppl. 2, etc. _ AND J. J. pAplKE (1972) pyroxenes as recorders of lunar

basalt petroqenesis: chemical trends due to crystal-liquid inter-AcRELL, S. O., J . E. Acnr l l , A. R Anuolo eNo J. V. P. LoNc act ion. pLC J,431-469.

(1973) Some observat ions on rock 62295. LS IV, 15-17. eNn D H. LrNosLey (1971) Crystal l izat ion-, J . H. ScooN, I . D. Muln, J. V. P LoNc, J. D C. h istor ies of c l inopyroxenes in two porphyr i t ic rocks f rom

McCoNNr r - r nxnA .P rc re r r ( 1970 )Obse rva t i onson thechem- OceanusProce l l a rum. PLC2 ,559 -574 .ist ry, mineralogy and petro logy of some Apol lo l l lunar sam- S. SusNo eNo J. W. DeuNo (1973) Pyroxenep les PLC 1 ,93 -128 . po i k i l ob l as t i c rocks f romthe luna rh igh lands . PLC4 ,597 -611 .

Alnss, A. L. , A. A. CHooos, R. F. Dvrrapr, A. J. G,cNc,o 'nz eNo Brssop, F C., J . V. , Srrr r ru nNo J. B. DnwsoN (1975) Pent lan-D S GoloulN (1974a) Preliminary investigation of boulders 2 dite-magnetite intergrowth in De Beers spinel lherzolite: reviewand 3, Apollo 17, Station 2: Petrology and Rb-Sr Model Ages. of sulphides in nodules phys Chem. Earth,9, 323-337.LS V' 6-8. Bleu, P. J lNo J. I . Gol t>srrrN (1975) Invest igat ion and s imula-

- ' D. A. PepeN,cstrsstou t ionofmetal l icspherulesfromlunarsoi ls .Geochim.cosmochim,nNo G. J. Wlssensunc (1974b) Duni te f rom the lunar h igh- Acta,39,3O9-3i4l ands : Pe t rog raphy , de fo rma t i ona l h i s t o r y , Rb -S r age . lS Z BoyD ,F .R .ANDD.Sna r rH (1971 )Compos i t i ona l zon ing inpy rox -3-5 ' enes f rom lunar rock 12021. Oceanus Procel larum. J Petro l .

- , R. F. Dvvsr AND D. J. Dnpror. r (1975) Spinel sym- lZ.439_464.plectites: high-pressure solid-state reaction or late-stage mag-

matic crystallization? LS vl, l-3. Bnerr, R (1973) A lunar core of Fe-Ni-S Geochim cosmochim.

- , A. J. GANcARZ nNo A. A. cuooos (1973) Metamorphism Acta ' 37 ' 165-170'

ofApol lo 16 and 17 and Luna 20 metaclast ic rocks at about 3.95 - (1975) Reduct ion of mare basal ts by sul fur loss LS

AE: samples 61156, 64423,14-2, 65015, 6'1483,15-2, 76055, VI'89-91'

22N6.22007. PLC 4.569-595. - , P. Burrrn, Jn, C. Mryen, Jn. , A. M. Rsto, H. Tersoe

ANDERSEN, O, (1915) The system anorth i te- forster i te-sr l ica. Am nNo R. WtLt- l . lvs (1971) Apol lo l2 igneous rocks 12004, 12008,

J. Sci., 4th ser ,39, 407-454. t2009, and 12022: a mineralogical and petrological study. PIC

ANoensoN, A. T. (1973) The texture and mineralogy of lunar 2,301-317.

per iodot i te, 15445,10. J. Geol .8 l ,219-226. - , R. C. Goolev, E. Dowrv, M. PntNz AND K. KEIL (1973)

LUNAR MINERALOGY t t09

origin of metall ic mounds on lunarOxide minerals in lithic fragments from Luna 20 fines. Geochim.Cosmochim Acta, 31 ,

'16l-773.

BnowN, G. E. nNo B. A. WpcHsr-en (1973) Crystal lography ofpigeonites from basaltic vitrophyre 15597. PLC 4, 887-900.

BnowN G. M., C. H. ErvrBLrus, J. G. Holr-rNo, A. Prcxrrr lNoR. PHrr-lrps (1971) Picrite basalts, ferrobasalts, feldspathic no-rites, and rhyolites in a strongly fractionated lunar crust. PIC 2,583-600.

AND _ (1972) Mineral_chemicalvariat ions in Apollo l4 and Apollo l5 basalts and granit icfract ions. PLC 3.141-157

-, A. Prcrerr, R. Psrl lrps rNo C. H. EruEr-Eus (1973) Min-eral-chemical variations in The Apollo l6 magnesio-feldspathichighland rocks. PLC 4, 505-518.

C. H. Er*.ru-eus AND R. PHrr-r-rps (1974) Min-eral-chemical propert ies of Apollo l7 mare basalts and terrafragments. LS V,89-91.

R. Psrr- lrps AND C. H. Eusleus (1975) Mineralogyand petrology of Apollo l7 basalts. LS VI,95-97 and PLC 6,r - 1 3 .

Bnynr.r, W. B. (1974) Fe-Mg relat ionships in sector-zoned sub-marine basalt plagioclase. Earth Planet. Sci. Lett.24, l5'l-165.

BuNcs, T. E. nNo K. Kerl (1971) Chromite and i lmenite in non-chondritic meteorites. Am. Mineral. 56, 146-157.

lNo K. G. SNrrrsrNcen (1967) Chromite composi-t ion in relat ion to chemistry and texture of ordinary chondrites.Geochim. Cosmochim. Acta, 31, 1569-1582.

- AND E. Or-srN (1975) Distr ibution and signif icance of chro-mium in meteorites. Geochim. Cosmochim. Acta, 39, 9ll-927.

BunNHlu, C. W. (1971) The crystal structure of pyroxferroitefrom Mare Tranquil l i tat is. PLC 2,47-5' l .

BunNs, R G., D. J. VeucnlN, R. M. Anu-Ero lNo M. WrrNrn(1973) Spectral evidence for Cr3+, Ti8+ and Fd+ rather thanCr2+ and Fe8+ in lunar ferromagnesian silicates. PLC 4,983-994.

Buscse, F. D., G. H. CoNRAD, K. Krr l , M. PnrNz, T. E. BuNcH,J. EnlrcuueN nNo W. L. Queror (1971) Electron microprobeanalyses of minerals from Apollo l2 lunar samples. Units. NewMexico Insl Meteorit.. Soec. Pub.3.

CeupnoN, E. N. (1970) Opaque minerals in certain lunar rocksfrom Apollo l l . PLC 1,221-245.

- (1971) Opaque minerals in certain lunar rocks from Apollo12. PLC 2,193-206.

- (1975) Posrcumulus and subsolidus equil ibrat ion of chro-mite and coexisting silicates in the Eastern Bushveld Complex.Geochim. Cosmochim Acta. 39. l02l-1033.

CnnenoN, K. L. ,qNo M. CeurnoN (1973) Mineralogy of ultra-mafic nodules from Knippa Quarry, near Uvalde, Texas. CeolSoc. Am. Abslr. Programs,5,566.

- AND J. W. DnLeNo (1973) Petrology of Apollo 15 con-sort ium breccia 15465 PLC 4,461466.

A. E. BsNcr rNo J. J. Prexr (1973) Petrology ofthe 2-4mm soil fraction from the Hadley-Apennine region of theMoon. Earth Planet. Sci. Lett. 19,9-21.

- AND G. W. FISHER (1975) Olivine-matrix reactions inthermaf ly-metamorphosed Apollo l4 breccias. Earth Planet. Sci.Lett. 25, 197-20'1.

Clnren, J. L. lrqo D. S. McKlv (1972) Metal l ic mounds pro-duced by reduction of material of simulated lunar composition

and impl icat ions on theglasses. PLC 3, 953-970.

- AND E. PeoovnNr (1973) Genetic implications of someunusual part icles in Apollo l6less than lmm fines 668841,1I and6994t.t3. PLC 4. 323-332.

Cnnrrn, N. L., L. A. FBnNllorz, H. G. Ave' l , t l -LEMANT lr-ro I . S.LruNc (1971) Pyroxenes and ol ivines in crystal l ine rocks fromthe Ocean of Storms. PLC 2,775-795.

-, I . S. Lsur.rc, H. G. AvE'hLLEMANT eNo L A. FERNANDEZ(1970) Growth and deformational structures in silicates fromMare Tranquillitatis. P LC I, 26'l -285.

CHnurNrss, P. 8., A. C. DuNHnv, F. G. F. Grss, H. N. Gtles,W. S. MncKrrzre, E. F. Srunpnl euo J. ZussulN (1971)Mineralogy and petrology of some Apollo l2 lunar samples.PLC 2 .359-376.

- AND G. W. LonrnrEn ( 197 I ) An electron microscopic studyofa lunar pyroxene. Contr ib. Mineral. Petrol.33, l7l-183.

CHeo, E. C. T. (1973) The petrology of 76055,10, a thermallymetamorphosed fragment-laden olivine micronorite hornfels.PLC 4. U5-664.

- AND l4 orsBns (1970) Pyroxferroite, a new calcium-bearingiron silicate from Tranquillity Base. PLC I, 65-79.

CHrnr-ss, R W., D. A. HEwtrr AND D. R. WoNrs (1971) HrO inlunar processes: The stabi l i ty of hydrous phases in lunar samples10058 and 12013 PLC 2,645-664.

CHnrsr re , J . M. , J . S . Ln l l v , A . H. Hpurn , R. M. F Isusn, D. TGntccs lNo S. V. RADcLIFFE (1971) Comparative electron pet-rography of Apollo I l , Apollo 12, and terrestr ial rocks. PLC 2,69-89.

CsnrsropHe-MrcHrr--Levy, M., C. Lnvv, R. Clvp AND R. PIER-nor (1972) The magnesian spinel-bearing rocks from the FraMauro formation. PLC 3, 887-894. Given incorrectly as (1973)in Part I

Crsowsxr, C. S., J. R. DuNr.r, M. Ful lrr. , M. F. Rosr AND P. J.Wnstlrwsxt (1974) Impact processes and lunar magnetism.PLC 5,2841-2858.

Cr-eNroN, U. S., J. L Cnnrrn nNn D. S. McKev (1975) Vapor-phase crystallization of sulfides? LS VI, 152-154. See also PIC6,7 t9 -728

ConrsroN, W. , M. J . VERNoN, H. Benny, R. Ruoowsxr , C . M.Guv nNo N Wnnr (1972) Age and petrogenesis of Apollo l4basalts. LS II I , l5l-153.

Cnnwrono, M L. (1973) Crystal l izat ion of plagioclase in marebasalrs. PIC 4, 705-7 17.

- AND L. S. Horr_rsren (1974) KREEP basalt: A possiblepart ial meft from the lunar interior. PLC 5,399-419

Der-toN, J. lNo L. S. Hot- l tsren (1974) Spinel-si l icate co-crystal-l izat ion relat ions in sample 15555. PLC 5,421429.

Dnwsott, J. B. nNo J. V. SMIrH (1975) Chromite-si l icate inter-growths in upper-mantle peridotites. Phys Chem. Earth, 9,3 39-3 50.

DELANo, J . W. , A E. BErcr , J . J . Pnptxs rNo K. L . CeusnoN(1973) Petrology of the 2-4mm soil fraction from the Descartesregion of the Moon and strat igraphic implications. PLC 4,537-55 r .

DeNce, M R., J. A. V. Doucl ls, A. G. PLANr nno R. J. Tnl l l l(1970) Petrology, mineralogy and deformation of Apollo l lsamDles. PLC I, 315-340.

- , M. PnlNz, K. KelL eNo T. E. BUNcH (1972) Spinels and CuvroN, R. N. luo T. K. Mnvso,c (1975) Genet ic re lat ionsthepe t rogenes i so f someApo l l o12 igneous rocks .Am.M ine ra l . be tween theMoonandme teo r i t es . LSV I , l 55 - l 57andPLC6 ,51 , t 729 -174 '7 . t 76 t - t ' t 69 .

l l t 0

_ AND _ (1971) Mineralogy and petrol-ogy of some Apollo l2 samples. PLC 2,285-299.

DrxoN, J. R. euo J. J. P,rprxr (1975) Petrology of cataclast icanorthosites from the Descartes region of the Moon: Apollo 16.LS VI. t90-192.

DonN, A. S. nNo J. I . Gor-osrstN (1970) The ternary phase dia-gram, Fe-Ni-P. Metal. Trans. l , l '159-1767.

Dooo, R. T. (1971) The petrology of chondrules in the Sharpsmeteorite. C ont rib. M ine ral. Pe t rol. 31, 201 -227 .

Dowrv, E., K. Kerl AND M. PnrNz (1972a) Anorthosite in theApollo l5 rake sample from Spur Crater. Apollo 15,62-66.

AND - (1973a) Major-element vapor fract ion-ation on the lunar surface: an unusual lithic fragment from theLuna 20 fines. Earth Planet. Sci Lett 21,91-96.

AND- (1974a) Plagioclase twin laws in lunarhighfand rocks; possible petrogenetic significance. Meteoritics,9 , 1 8 3 - 1 9 7 .

AND - (1974b) Igneous rocks from Apollo l6rake samples. PLC 5,431-445.

AND - ( | 974c) Lunar pyroxene-phyric basalts:crystalfization under supercooled conditions. J. Petrol. 15,4t9453.

-, M. PnrNz nNo K. Kerl (1973b) Composit ion, mineralogy,and petrology of 28 mare basalts from Apollo l5 rake samples.PLC 4.4234M.

AND - (1974d\ Ferroan anorthosite: A wide-spread and distinctive lunar rock type. Earth Planet Sci. Letr.24, t5-25

-, M. Ross nNo F. Currrrrl (1972b\ Fe'z*-Mg site distribu-t ion in Apollo 12021 cl inopyroxenes: evidence for bias in Mdss-bauer measurements, and relat ion of ordering to exsolut ion.PLC 3 ,48 t -492.

Dure. M. J., L S. McCellurrr. G. A. McK,rv AND D. F. WEILL(1970) Mineralogy and petrology of Apollo 12 sample No.120f3: A progress report. Earth Planet. Sci Lett.9,103-123.

Dusn, A., H. C. Hreno eNn R. N. ScHocK (1974) Electr icalconductivi ty of ol ivine at high pressure and under control ledoxygen fugacity. J. Geophys. Res. 79, 1667-1673.

DurB, M. B. (1965) Metal l ic iron in basalt ic achondrites. , / .Geophys. Res 10, 1523-152'l.

DUNI-oR, D. J., W. A. Gose, G. W. PEARCE nNn D. W. StneNc-wev (1973) Magnetic propert ies and granulometry of metal l iciron in lunar breccia 14313. PLC 4,2977-2990.

Dwonrrr, E. J., C. S. AnNrr-1, R. P. CHnrsrrnN, F. CurrrrrA,R. B. FrNxrlueu, D. T. LrcoN, Jn. nNo H. J. Rose, Jn. (1974)Chemical and mineralogical composit ion of Surveyor 3 scoopsample 12029,9. PLC 5, 1009-1014.

Dvnl, P., C. W PnnrrN AND W. D. Duly (1974) Temperatureand electrical conductivity of the lunar interior from magnetictransient measurements in the geomagnetic tai l . PLC 5,3059-307 t.

El Gonrsv, A., M. Pnrxz eNo P. Rnploosn (1975) Zoning inspinels as an indicator of the crystal l izat ion histories of Apollof 2and l5marebasalts. Geol. Soc. Am.,Abslractswith Programs,7, to67.

-, P. Revoosn rNo O. MroeNBAcH (1973a) Lunar samplesfrom Descartes sites: opaque mineralogy and geochemistry.PLC 4 ,733-750.

- AND H. J. BenNuenor (1974) Taurus-Lit-trow TiOr-rich basalts: opaque mineralogy and geochemistryPLC 5.627-652.

M. Plvrdrvlt . O. MeoeNsrcu. O. Mil l lsn nNo

J. V. SMITH AND I M, STEELE

W GeNrNrn (1973b) Zinc, lead, chlor ine and FeOOH-bear ing

assemblages in the Apol lo l6 sample 66095: or ig in by impact of

a comet or a carbonaceous chondrite. Earth Planet. Sci Lett. lE,

4 l t 4 t 9 .- , P. R nvoonn eNo L. A. TAY LoR ( 197 I ) The opaque miner-

als in the lunar rocks f rom Oceanus Procel larum. PLC 2'

219-235.- , L. A. Tnvlon nuo P RnvooHY. (1972) Fra Mauro crystal -

l ine rocks: mineralogy, geochemistry and subsol idus reduct ion

of the opaque minerals. PLC 3,333-349.

ENcrLHnnor, W. VoN, J. AnNor, W. F. Mi l l l rn AND D. SrdFF-

l rn (1970) Shock metamorphism of lunar rocks and or ig in of

the regolith at the Apollo I I landing site PLC I, 363-384.

EvrNs, H. T. Jn. (1970) The crystal lography of lunar t ro i l i le . PLC

/. 399-408.Fr-oneN. R. J. , K. L. C, tuenou, A E. BrNcn AND J. J. PAPIKE

(1972) Apol lo l4 breccia 14313: A mineralogic and petro logic

report . PLC 3,661-6'71FoRESTER, D. W. (1973) Mt issbauer search for ferr ic oxide phases

in lunar mater ia ls and s imulated lunar mater ia ls. PLC 4,

269't-270r.FnresrLe. E. J., D. L. Gnlscou, C. L. MnnQu,qnor, R. A Wrrxs

nNo D Pnrsret (1974) Temperature dependence of the ferro-

magnet ic resonance l inewidth of lunar soi ls , i ron and magnet i te

precip i tates in s imulated lunar g lasses, and nonspher ical metal l ic

i ron part ic les. PLC 5,2729-2736.

FnoNorr- . C , C. Klerrv Jn. , J . I ro AND J. C. Dnnrr (1970) Miner-

alogical and chemical studies of Apol lo I I lunar f ines and se-

lected rocks. PLC I ,445-474.FnoNosr-, J. W. (1975) Lunar Mineralogy. Wiley-lnterscience.

New YorkFucHs, L. H. (1968) The phosphate mineralogy of meteor i tes. In,

P. M Millman, Ed., Meteorite Research, Reidel, Dordrecht.- (1971) Orthopyroxene and or thopyroxene-bear ing rock

fragments r ich in K, REE, and P in Apol lo 14 soi l samples

14163. Earth Planet Sci . Let t .12,170-174.-AND E OlsrN (1973) Composi t ion of metal in Type I I I

carbonaceous chondr i tes and i ts re levance to the source-assign-

ment of lunar metal . Earth Planet Sci . Let t 18, 379-384

lNn K. J. JeussN (1973) Mineralogy, mineral-

chemistry, and composi t ion of the Murchison (C2) meteor i te.

Smithsonian Contrib Ealth Scl., No. 10, l-39

GeNlpnrHv, R . R. R. Krnvs, J. C. L lur- nNo E. ANorns (1970)

Trace elements in Apol lo I I lunar rocks: impl icat ions for mete-

or i te inf lux and or ig in of Moon. PLC I , l l lT-1142.- , J . W . MoRcAN, H H rcucHr , E ANDERS lNn A . T . ANorn -

soN (1974) Meteor i t ic and volat i le e lements in Apol lo l6 rocks

and in separated phases f rom 14306. PLC 5,1659-1683'

GnNcnnz , A . J . . A . L ALs rE rNo A . A . CHoDos (1971 ) Pe t ro l og i c

and mineralogic invest igat ion of some crystal l ine rocks returned

by the Apol lo l4 mission. Earth Planet Sci . Lel l . 12 ' l -18.

AND - ( | 972 ) Comparative petrology of Apollo

l6 sample 68415 and Apollo l4 samples 14276 and 14310. Earth

Planet . Sci . Let t 16,307-330.

Gey. P. . G. M. Blr r rcnoFr AND M. G. BowN (1970) Di f f ract ion

and Mdssbauer studies of minerals f rom lunar soi ls and rocks.

PLC t , 481497.- , M. G. BowN AND I . D. Murn (1972) Mineralogical and

petrographic features oftwo Apollo l4 rocks. PLC 3,351-362.-, C M. BlNcnopr AND P G. L. Wlrltrtr,ts

(1971) Mineralogical and petrographic invest igat ion of some

Apol fo l2 samples. PLC 2,3 '17-392.

GHosr, S, l . S. McCnl lur* l eNo E. Trny (1973) Luna 20 pyrox-enes: exsolution and phase transformation as indicators ofpetro logic h istory. Ceochim Cosmochim Acta, 37,831-839.

- , G. Nc eNo L. S. Wnlrsn (1972) Cl inopyroxenes f romApol lo l2 and l4: exsolut ion, domain structure and cat ion or-der PLC - i , 507-531.

- , C. WeN ero l . S. McCnrruu (1975) Late thermal h istoryof lunar anorthosi te 67075: Evidence f rom cat ion order in o l i -vine and orthopyroxene. LS V(,282-283.

Ctss, F. G. F. (1974) Supercool ing and the crystal l izat ion ofplagioclase from a basaltic magma. Mineral. Mag.39,641-653.

Grssol . r , E. K. Jn. , S. CHnnc, K. LsrroN, C. W. Moonr eNoG. W. PEARCE (1975) Carbon, sul fur , hydrogen and metal l ici ron abundances in Apol lo l5 and l7 basal ts. LS Vl ,290-292.See also PLC 6, 1287-1301.

- AND G. W. Moone (1973) Carbon and sul fur d ist r ibut ionsand abundances in lunar f ines. PLC 4, 1577-1586.

- AND - (1974\ Sul fur abundances and dist r ibut ions inthe val ley of Taurus-Li t t row. PLC J, 1823-1837.

Golosretn, J. L (1966) But ler , Missour i : an unusual i ron mete-orite. Science 153, 975.

- AND H. J. AxoN (1972) Metal l ic part ic les f rom 3 Apol lo l5soils. Apollo 15, 78-8 l.

- AND - (1973) Composi t ion, st ructure, and thermalhistory of metal l ic part ic les f rom 3 Apol lo l6 soi ls , 65701, 68501 ,a n d 6 3 5 0 1 . P L C 4 , 7 5 1 - 7 7 5 .

AND S. O. AcRELL (1975) The structure and thermalhistory of five large metal particles from the lunar regolith. LSvt. 303-305.

nNo C. F. YeN (1972) Metal l ic part ic les in theApol lo l4 lunar soi l PLC 3, 1037-1064.

- AND P. J. Blnu (1973) Chemistry and thermal h istory ofmetal part ic les in Luna 20 soi ls . Geochim. Cosmochim. Acta,31,847-855.

-AND A S Don r , Jn ( 1972 )Thee f f ec to fphospho ruson theformation of the Widm?instatten pattern in iron meteorites. Geo-chim Cosmochim Acta, 36, 51-69.

-- , E P. HrNornsoN AND H YAKowrrz(1970) Invest igat ionof lunar metal part ic les. PLC 1,499-512.

- , R. H. HEwrNs nNo H. J. AxoN (1974) Metal s i l icaterelat ionships in Apol lo l7 soi ls . PLC 5, 653-6 ' l l .

- AND H Ylrowrrz (1971) Metal l ic inc lusions and metalpart ic les in the Apol lo l2 lunar soi l . PLC 2, 177-191.

GooLEy, R. C. , R Bnrrr nNo J. L. WARNER (1973) Crystal l izat ionhistory of metal part ic les in Apol lo l6 rake samples. PLC 4,'t99-810

- AND J. R Srr lvru (1974) A lunar rock ofdeep crustal origin: sample 76535. Geochim. Cosmochim Acta,3E , 1329 -1339 .

- AND C. B. Moone (1973) The metal in diogenites. Mete-o r i t i c s ,8 ,4 l ( abs t r ) .

Cose, W A , G. W. Penncs, D. W. SrnnNcwAy AND E. E.

Lnnsol (1972) Magnet ic propert ies of Apol lo l4 breccias andtheir corre lat ion wi th metamorphism. PLC 3,2387-2395.

GnteN, D. H. (1975) Genesis of Archean per idot i t ic magmas andconstra ints on Archean geothermal gradients and tectonics Ce-ology,2, 15-18.

Gn revs , R . A . F . , G A . McKny , H . D . Sv r r rH AND D . F WEILL(1975) Lunar polymict breccia 14321: a petrographic study. Geo-chim Cosmochim Acta, 39, 229-245

- AND A. G. PlnNr (1973) Part ia l mel t ing on the lunar

l l l l

surface, as observed in glass coated Apollo l6 samples. PLC 4,667 -679.

GnrrnrN, W. L., R. Aulr nxo K. S. Htrrn (1972) Whit lockite andapatite from lunar rock 14310 and from Odegirden, Norway.Earth Planet. Sci. Lett.15, 53-58.

Gnrscou, D. L., E. J. Fnreser-s nNo C. L. Mlnounnor (1973)Evidence for a ubiquitous, sub-microscopic "magnetite-like"consti tuent in the lunar soi ls. PLC 4,2'109-2727.

GnossMe:r, L. (1975) Petrography and mineral chemistry of Ca-rich inclusions in the Allende meteorite. Geochim. Cosmochim.Acta,39, 433-454.

HerNen, S. S., D. Vrnco nNo D. WARBURror.r (1971) Cationdistr ibutions and cool ing history of cl inopyroxenes fromOceanus Procel larum. PLC 2. 9l-108.

Hnccsnrv, S. E. ( l97la) Composit ional variat ions in lunar spin-els. Nature Phys. Sci.233, 156-160.

- ( l97lb) Subsolidus reduction of lunar spinels. Nature Phys.Sc i .234, I l3 - l 17 .

- 11972a) Apollo l4: subsolidus reduction and compositionalvariat ions of spinels. PLC 3,305-332.

- (1972b) Luna I 6: an opaque mineral study and a systematicexamihhtion of composit ional variat ions of spinels from MareFecunditatis. Earth Planet. Sci. Lett. 13,32E-352.

- (1972c\ The mineral chemistry of some decomposition andreaction assemblages associated with Cr-Zr, Ca-Zr andFe-Mg-Zr titanates. Apollo l5, 88-91.

- (1972d) An enstatite chondrite from Hadley Rille. Apollo/J, 85-87.

- (1973a) Luna 20: mineral chemistry of spinel, pleonaste,chromite, ulvdspinel, ilmenite and rutile. Geochim. Cosmochim.Acta, 37, 857-867.

- ( 1973b) Armalcol i te and genetical ly associated opaque min-erals in the funar samples. PLC 4,777-797.

- (1973c) The chemistry and genesis of opaque minerals inkimberlites. Intl. Conf. Kimberlites, Extended Abstr. 147-149.

- (1973d) Ortho- and para-armalccl l i te samples in Apollo 17.Nature Phys. Sci 242,123-125.

- (1975) The chemistry and genesis of opaque minerals inkimberlites. Phys. Chem. Earth, 9, 245-307.

HlrNss. E. L.. A L. ALBEE. A. A. CHoDos AND C. J. Wnsse nsunc(1971) Uranium-bearing minerals of lunar rock 12013. EarthPlanet Sci. Lett. 12, 145-154.

-. A J. G,qNcenz. A. L. ALBEE AND G. J. Wrssensunc(1972\The uranium distr ibution in lunar soi ls and rocks 12013and 14310. ZS 1 / / , 350-351.

Hencnlves, R. B. nNo L. S. Hollrsrrn (1972) Mineralogic andpetrologic study of lunar anorthosite sl ide I5415,18. Science,t75,430432.

Hesxr r , L . A . , C . -Y . Ssrs , B . M. BeNs l r - , J . M. Rsoors , H.Wrrsrr,rnuN exo L. E. Nyeutsr (1974) Chemical evidence for theorigin of 76535 as a cumulate. PLC 5, 1213-1225.

Hrnno, H C., A. Dusl, A. J. Prwrrsrrr AND R. N. ScHocr (1975)Electrical conductivity studies: refinement of the selenotherm.LS vr, 355-357.

Httrsr, G. (1975) Petrology of lunar soils. Reu. Geophys. SpacePhys. 13, 567-587.

He rz, R T. (1972) Rock 14068: an unusual lunar breccia. PLCJ,865-886.

HeNosnsoN, P. (1975) Reaction trends shown by chrome-spinelsof the Rhum layered intrusion. Geochim. Cosmochim, Acta,39,I 03 5- I 044.

Hruen, A . H. , J . M. Cuusr re , J . S . Ln l l y AND C. L . Nonp, Jn .

LUNAR MINERALOGY

I l 1 2

(1974) Electron petrographic study ofsome Apol lo l7 breccias.PLC 5,275-286.

HrwtNs, R. H lNo J I . ColosrnrN (19774)Metal-o l iv ine associa-tions and Ni-Co contents in two Apollo l2 marebasalts. EarthPlanet. Sci Lett. 24, 59-70.

- AND - ( t975a) Compar ison of s i l icate and metal geo-thermometers for lunar rocks. lS VI, 356-357.

- AND - ( I 975b) The provenance of metal in anorthosi t icrocks. LS Vl ,358-360 and PLC 6,343-362.

HrNrHonNr, J. 8. , R. CoNuo euo C. A. ANoenseN (1975)

Lead-lead age and trace element abundances in lunar troctolite,76535. LS VI,373-375.

Huvl , P. F. , M. PnrNz nNo K. Kul (1972) Niobian rut i le in anApollo l4 KREEP fragment. Meteoritics, 7, 479-485.

Hooces, F. N nuu I Kusurno (1973) Petro logy of Apol lo 16lunar h ighlands rocks. PLC4, 1033-1048.

Ho lL r s r sn , L S . , W . E . Tnzc r rNs r r , Jn . , R . B . H lnonnvss nuo

C. G Kulrcx (1971) Petrogenet ic s igni f icance ofpyroxenes tntwo Apol lo l2 samples. PLC 2, 529-557.

Houslrv, R. M., E. H. Crnlr r , N. E. PeroN lNo I . B. Golosenc(1974) Solar wind and micrometeor i te a l terat ion of the lunarregolith PLC 5, 2623-2642.

- ! R. W. Gnnrr AND M. Asor l -Gewno (1972) Study ofexcess Fe metal in the lunar fines by magnetic separation, Mdss-bauer spectroscopy, and microscopic examination. PLC 3,I 065- | 076.

nNo N. E. PlroN (1973) Or ig in and character is t icsof excess Fe metal in lunar glass welded aggregates. PLC 4,2737-2749.

Hunnnno , N . J , J . M . Ruooes , H . Wr r sueNN, C . -Y . SHrH , o rNoB. M. BeNsnr- (1974) The chemical definition and interpretationof rock types returned from the non-mare regions of the Moon.

PLC 5,1227-t246.Hun ru r l r . l , G . P . , G . R . Du l r vyn r AND R . M . F r suen (1975 )

Superparamagnet ic c lusters of Fd+ spins in lunar o l iv ines. l ,SVI, 414-416. See also PLC 6, 757-772.

- , F. C ScHwenrn, R M. FrsHrn nNo T. Nnc,cr l (1974)

Iron dist r ibut ionS and metal l ic- ferrous rat ios for Aool lo lunarsa m ples: M dssbauer aiidhagnetic analyses. P LC 5, 27 7 9 -27 94.

InvrNe, T. N. (1974) Petro logy of the Duke Is land ul t ramaf iccomplex, Southeastern Alaska Geol. Soc. Am Mem. 138.

lRvlNc, A. J. , I . M. Srsr I - r nNo J. V. SlurH (1974) Lunar nor i t icfragments and associated diopside veins. Am Mineral. 59,l 062- l 068.

J lconzrNsxr, H. rNo M. Konxnwn (1973) Di f fuse X-ray scat ter-

ing by lunar minerals PLC 4,933-951.- (1975) Di f fuse scat ter ing by domains in lunar

and terrestr ia l p lagioclase. LS VI,429-431.Jeues, O. (1972)Lunar anorthosi te 15415: texture, mineralogy and

metamorphic history. Science, l7S, 432436.- (1973) Crystal l izat ion history of lunar fe ldspathic basal t

14310. U. S. Geol. Suru. ProJ. Pap.841.Jrowln, J. (1971) Surface morphology of f ree-growing i lmeni tes

and chromites from vuggy rocks 10072,31 and 12036,2. PLC 2,

923-935.- (1973) Rare-micron-size minerals in lunar f ines. PLC 4,

86 I -874.- AND A HERBoscH (1970) Tentative estimation of the con-

tribution of Type I carbonaceous meteorites to the lunar soil.PLC t. 55t-559.

JouNsoN, R. E . , E . WornueuN rNo A. MurN (1971) Equ i l ib r iumstudies in the system MgO-"FeO"-TiOz. Am. J. Sci. 271,278-292.

J. V. SMITH AND I M. STEELE

Kerl, K. (1968) Mineralogical and chemical relat ionships amongenstatite chondrites. J. G eophy s. R es. 73, 6945-697 6.

-, T. E. BuncH rNo M. PnrNz (1970) Mineralogy and compo-s i t ion o f Apo l lo I I lunar samples . PLC 1 ,561-598.

-, M. PRrNz AND T. E. BuNcs (1971) Mineralogy, petrology,and chemistry of some Apollo l2 samples. PLC 2,319-341.

KEssoN, S. E. nNo D. H. LtNosr-sv (1975) The effects of Al+e,Cr+3, and Ti+s on the stability of armalcolite in vacuum and at7.5 kbar. LS VI, 472-474 See also PLC 6,9l l-920.

- AND A. E. Rrxcwooo (1976) Mare basalt petrogenesis in adynamic Moon. Earth Planet. Sci Lelt .30, 155-163.

Krrrr, C. Jn. nNo J. C. Dnnxr (1972) Mineralogy, petrology, andsurface features of some fragmental material from the FraMauro site. PLC 3, 1095-l l l3.

KLrrNueuu, B. nNo P. RAMDoHR (1971) Alpha'corundum fromthe lunar dust Earth Planet. Sci. Letl. 13' 19-22'

KuRAr. G,. K. Kerr nxo M. Pntxz (1974) Rock 14318: a polymictlunar breccia with chondritic textsre. Geochim. Cosmochtm.Acta ,3E, | 133- l 146.

Kusurno, I . nNo H. S. Yoorn, Jn. (1966) Anorthite-forsteri te andanorthite-enstatite reactions and their bearing on the basalteclogite transformation . J. Pe t rol. 7' 33'l -362.

Ln l l v , J . S . , A . H. Hrurn , G- L . Nono, Jn . lNo J . M. Csn ls r te(1975) Subsolidus reactions in lunar pyroxenes: an electron pet-rographic study. Contrib. Mineral. Petrol. 51,263-281.

-. R. M FrsHER, J. M. Cnnrsrtr, D. T. GRIccs, A. H.He urn, G. L. Nono, Jn. , tNo S V. Rnocrrrrt (1972) Electronpetrography of Apollo l4 and l5 rocks. PLC 3,401-422.

Levv. C.. M. CsnrsropHp-Mtcutt--Levv, P. PIcor AND R. CAYE1972\ A new titanium and zirconium oxide from the Apollo l4samples. PLC 3, I I l5- l 120.

LrNosrEv, D. H.lNo C. W. BunrHet' . t (1970) Pyroxferroite: stabi l-ity and X-ray crystallography of synthetic Cao luFeo ruSiOgpyroxenoid. Science, 168, 364-367.

-. S. E. KsssoN. M. J. HlnrzuAN AND M. K. CusHtvt+tt(1974a\ The stability of armalcolite: experimental studies in thesystem MgO-Fe-Ti-O. PLC 5, 521-534.

-, H. E. KrNc AND A. C. TuRNocK (19'14b) Composit ions ofsynthetic augite and hypersthene coexisting at 810'C: appli-cations to pyroxenes from lunar highlands rocks. Geophys. Res.L e u l , 1 3 4 - 1 3 6 .

LrptN, B. R. (1975) The system anorthite-ol ivine-si l ica and i tsbearing on the petrogenesis of lunar crustal rocks. Ceol. Soc.Am. Abstr. with Programs,T, l l '12.

- AND A. MuAN (1974) Equil ibr ia bearing on the behavior oftitanate phases during crystallization ofiron silicate melts understrongly reducing condit ions. PLC 5,535-548.

LorcnsN. G.. C. H. DoNALDsoN, R. J. Wllulus, O. Mult-tNsnNo T. M. UssrlneN (1974) Experimentally reproduced tex-tures and mineral chemistry of Apollo 15 quartz normativebasalts. PLC 5, 549-567.

LoNcHr. J . D. WlI-rsn, T. L. Gnove, E. M. Srolprn AND J. F.H,rvs (1974) The petrology of the Apollo l7 mare basalts' PLC5.447469.

LovERrNc, J. F. (1964) Electron microprobe analyses ofthe metal-l ic phase in basalt ic achondrites. Nature,203'70.

- (1975) The Moama eucrite-a pyroxene-plagioclase-plagioclase-adcumulate. Meteoritics, 10, l0l-l 14.

- AND l4 orHsns (1971) Tranquil l i tyi te: A new si l icate min-eral from Apollo I I and Apollo l2 basalt ic rocks. PLC 2,3945.

-. D. A. Wnnr. A. J. W. Crr.rnow AND R. BRrrrEN (1974)Lunar monazite: a late-stage (mesostasis) phase in mare basalt.Earth Planet. Sci. Lett.2l. 164-168.

LuNntrc AsyLulr (1970) Mineralogic and isotopic invest igat ionson funar rock 12013. Earth Planet Sci . Let t 9,137-163.

Mno, H K. , A. El Gonnsy nNo P. M. BBrt (1974\ Evidence ofextensive chemical reduction in lunar regolith samples from theApol lo l7 s i re PLC J, 673-683.

M,rnvrN, U B. (1971) Lunar niobian rulrle Earth Planet Sci. LettI t . 7 -9

- , J A . Wooo ,G . J Tnv lon , J B . RE ID ,Jn . , B .N powp l l ,

J . S . D rc r r v , Jn . lNo J F Bow tn (1971 ) Re la t i ve p ropo r r i onsand probable sources of rock fragments in the Apollo 12 soilsampfes. PLC 2, 679-699.

MrsoN, B (1972) Mineralogy and petro logy of lunar samples15264 ,19 ,15274 ,12 , and 15314 ,59 . Apo l l o 15 , t 35 -136 .

- , W. G. Mr lsoN eNn J. Nrr-pN (1972) Spinel and horn-blende in Apol lo l4 f ines L,S I I I , 512-514.

MessoN, C. R , L B. Surru, W. D J l l r rssol , J L. Mcl- lucsleNeNo A. VoLsonrx (1972) Chromatographic and mineralogicalstudy of Apol lo l4 f ines. PLC 3, . 1029-1036.

McCnl l rsren, R. H. nNo L. A. Tly lon (1973) The k inet ics ofu lvc ispinel reduct ion: synthet ic study and appl icat ions to lunarrocks Earth Planet Sci. Lett 11,357-364.

McCn l l uu , l . S . , E . A . MATHEZ, F . P Oren rune eNo S . GHos r( | 974) Petro logy and crystal chemistry of poik i l i t ic anorthosi t icgabb ro 77017 . PLC 5 ,287 -3O2 .

_ AND _ (1975a) petro logy of nor i t iccumulates: samples 78235 and 78238. l .S V(,534-536. See alsoPLC 6 ,395 -414 .

- , F. P. Ornuuna AND S. GHosE (1975b) Mineralogy andpetrology of sample 67075 and the origin of lunar anorthositesEarth Planet Sci. Lett 26,36-53

McK,rv, D. S. , U. S. Cr-nNroN, D A. MonnrsoN AND G. H.Ltol t (1972) Vapor phase crystal l izar ion in Apol lo l4 breccia.PLC 3 ,739 -752 .

McKey , G A . , S . J Kn rosLs ,qucH AND D F . Ws r r - l ( 1973 ) Theoccurrence and origin of schreibersite-kamacite intergrowths inm ic rob recc ia 66055 . PLC 4 ,811 -818

Mrncten, J-C C. eNo A. Nrcolns (1975) Textures and fabr ics ofupper-mant le per idot i tes as i l lustrated by xenol i ths f rom basal ts.J Petrol 16,454-487

Mnv rn , C . , JR , D . H . ANosnsoN nNo J . G . Bne t l s y f i 974 ) I onmicroprobe mass analysis of p lagioclase f rom "non-mare" lunarsamples. PLC 5. 685-706.

- , R. Bnrrr , N. J. Huselno, D A MonnrsoN. D SMcK ly , F K . A r r r r x , H . T l r c roe AND E . ScHoNDFELD (1971 )Mineralogy, chemistry, and or ig in of the KREEP componenr insoi l samples f rom the Ocean of Storms. PLC 2, 393-411.

Mrvrn, C. E. nNn H. G. Wrlssrne (1974) "Duni te" inc lusion inlunar basalt 742'75 LS I/, 503-505

MeyEn, H O. A nNo N. Z. BocroR (1974) Opaque mineralogy:Apo l l o 17 , r ock 75035 PLC 5 .707 -7 t6 .

- AND R. H. McCnl l rsrnn (1973) Mineralogy and petro logyof Apol lo l6: Rock 60215,13. PLC 4, 66t-665

MrNx rN , J A . eNo E C . T . Cn ro (1971 ) S ing le c r ys ta l X - rayinvest igat ion of deformat ion in terrestr ia l and lunar i lmeni tePLC 2 ,237 -246 .

Mrsu, K C. nNo L. A. Teylon (1975) Correlat ion between nat ivemetal composi t ions and the petro logy of Apol lo l6 rocks. lSVI, 566-568. See also PLC 6,615-639

Moone, C B , C. F. Lewrs nNo J. D. Cnrpr (1974) Total carbonand sul fur contents of Apol lo l7 lunar samples. PLC 5,l 897- I 906

MonceN, J. W , R. Gnunp,rrHy, H. HrcucHr, U. KnAsrNni jHr_nNo E. ANoens (1974) Lunar basins: tenat ive character izat ion

I I 1 3

of projectiles, from meteoritic elements in Apollo l7 boulders.

PLC 5, t703-1736.MurN , A , J . H ruc r AND T . LdFALL (1972 ) Equ i l i b r i um s tud ies

with a bear ing on lunar rocks. PlC.3, 185-196MunrHv, V. R. , N. M. EveNsoN AND H. T. Hlr- l (1971) Model of

early lunar differentiation. Nature,234,267 and 290MysEN, B. (1975) Part i t ioning of i ron and magnesium between

crystals and partial melts in peridotite upper mantle. Contrib.Mineral. Petrol 52. 69-76.

Nncnr l , T. , R. M. FrsHER, F. C. ScuwenrR, M. D. FulLsn eNn

J. R DUNN (1975) Effects of meteorite impacts on magneticproperties of Apollo lunar materials. LS VI, 587-589 and PLC

6 , 3 l l L - 3 t 2 2 .- , N. Sucrur.e, R. M. Frssrn, F. C. ScswBRER, M D.

Ful l rn AND J. R. DuNr (1974) Magnet ic propert ies of Apol loll-17 lunar materials with special reference to effects of mete-

orite impact PLC 5,2827-2839.NrHnu, C. E, M. PnrNz, E. Dowrv AND K. KEIL (1973) Electron

microprobe analyses of spinel group minerals and i lmeni te inApollo l5 rake samples of igneous origin. Irst. Meteorit., UniD.New Mexico, Spec. Publ. 10.

_ AND _ (19j4) Spinel_group minerals

and i lmeni te in Apol lo l5 rake samples. Am. Mineral . 59,

t 220 - t 235 .NreruHn, H. H. , S. Zrrn l ' eno S. S. Hnrrrn (1973) Ferr ic i ron in

plagioclase crystals f rom anorthosi te 15415. PLC 4,971-982.O'Hnne, M J. eNo D. J. HuvpHntrs (1975) Armalcol i te crystal l i -

zation, phenocryst assemblages, eruption conditions and originof eleven high titanium basalts from Taurus Littrow. LS VI,

6 t 9 - 6 2 1 .Or-srN, E. AND L. H. FucHs ( 1967) The state of oxidat ion of some

iron meteorites. Icarus, 6, 242-253.AND W C. Fonnrs (1973a) Chromium and phos-

phorus enrichment in the metal of Type II (C2) carbonaceouschondrites. G eo chim C os mo c h im. A c ta, 31, 2037 -2042.

- , J . S. HurnNrn, J A. V. Doucles AND A. G. P1aNr(1973b) Meteor i t ic amphiboles. Am Mineral .58, 869-872.

PAprKE, J. J. AND A. E. BsNcE (1972) Apol lo l4 inverted pigeon-

ites: possible samples of lunar plutonic rocks. Earth Planet. SciLet t 14,176-182.

AND D. H. LrNosLev (1974) Mare basal ts f rom the

Taurus-Li t t row region of the Moon. PLC 5,4 ' l l -504.

C. T. Pnrwrrr , S. SunNo AND M. C,cvrnoN (1973) Pyrox-enes: comparison of real and ideal structural topologies. Z.K risrallogr. 138, 254-27 3.

Pnvr6rvr6, M., P. Rl tvtoouR AND A Et- Gonesy (1972) Electronmicroprobe invest igat ions of the oxidat ion states of Fe and Ti ini lmeni te in Apol lo l l , Apol lo 12, and Apol lo 14 crystal l inerocks PLC 3,295-303

Prnncs, G. W., W. A. Gosr AND J. LINDSAv (1975) Reduct ion ofiron and regolith soil maturity. LS V(,634-636

-, D. W. SrnlNcway AND W. A. Gosr (1974) Magnet icproperties of Apollo samples and implications for regolith for-mation. PLC 5, 2815-2826.

-, R. J. Wrllrnns AND D. S. McKAy (1972) The magneticproperties and morphology of metallic iron produced by sub-solidus reduction of synthetic Apollo I I composition glasses.Earth Planet. Sci. Lett. 11,95-104.

Prcrrr r , A. , R. PHtr- l tes AND G. M. BnowN ( i972) New zi rco-nium-rich minerals from Apollo l4 and l5 lunar rocks. Nature,236,2t5-217

Prl r - rNcrn, C. T. , P. R. Dlvrs, G. Ecr-rNror, t , A. P. Gowln, A. J.T. Jur-r - , J . R. Mexwrl t - , R. M. Housr-By AND E. H. CTRLIN

LUNAR MINERALOGY

l l 1 4 J, V. SMITH AND I M. STEELE

(1974) The associat ion between carbide and f inely d iv ided metal- Snro, M., N. L Hlcr l - lNc AND J. E. McLnNo (1973) Oxygen

l ic i ron in lunar f ines. PLC 5,1949-1961. fugaci ty values of Apol lo 12, 14, and 15 lunar samples and

PowsLL, B. N. (1969) Petro logy and chemistry of the mesoside- reduced state of lunar magmas. PLC 4, 106l-1079.

r i tes- l Textures and composi t ion of n ickel- i ron Geochim. Scst jn l tnNN, K. AND S. S. HnnNsn (1972) Dist inct subsol idus

Cosmochim Acta,33,789-810. cool ing histor ies of Apol lo 14 basal ts. PLC 3, 493-506.- , F. K ATTKEN eNo P. W. Wrrs l rNl (1973) Classi f icat ion, Sclen, C B. (1971) Shock- induced features of Apol lo l2 micro-

dist r ibut ion, and or ig in of l i th ic f ragments f rom the Hadley- breccias. PLC 2,817-832.

Apennine region. PLC 4,445460 - AND J. F. Bnurn (1974) Shock- induced mel t ing in anortho--, AND P W Wrrsr-EN (1972) Petrology and origin of lithic sitic rock 60015 and a fragment of anorthositic breccia from the

fragments in the Apol lo 14 regol i th. PLC 3,837-852. "p ick ing pot" (70052). PLC 5,319-336.

Pn rw r t r , C T . , G . E BnowN eNo J . J . P r r r r r ( 1971 ) Apo l l o 12clinopyroxenes: high temperature X-ray diffraction studies. PIC2, 59-68.

- AND D. R. RorHseno (1975) Crystal s t ructures of mete-or i t ic and lunar whi t locki tes. LS Vl ,646-648.

Pn rNz , M . , E . Dowry , K . K r r r AND T . E . Buncu (1973a ) M ine r -alogy, petrology and chemistry of lithic fragments from Luna 20f ines: or ig in of the cumulate ANT sui te and i ts re lat ionships tohigh-alumina and mare basal ts Geochim. Cosmochim. Acta,37,979- r006

_ AND _ (1973b) Spinel t roctol i te andanorthosi te in Apol lo l6 samples. Science, 119,74- '76.

Que ro r , W . , V OBERBEcx , T BuNcn euo G . Po l r ows r r ( 1971 )

Investigations of the natural history of the regolith at the Apollo12s i t e . PLC 2 ,701 -7 t8 .

RAyMoND, K. N. eNo H. R. Wsrr (1971) Lunar i lmeni te ( ref ine-

ment of the crystal structure) Contrib. Mineral. Petrol 3O,I 35- I 40.

Rrro, S J B lNo S R. Tevlon (1974) Meteor i t ic metal in Apol lo[6 samples Meteoritics, 9, 23-34.

Rsro, A. M , J. Z. Fnlzrn, H. FuJrrA rr . ro J. E. Evnnsor (1970)

Apol fo I I samples: major mineral chemistry. PLC 1,749-761.Rrr l , J . B. Jn. (1972) Ol iv ine-r ich, t rue spinel-bear ing anorthosi tes

from Apollo l5 and Luna 20 soils-possible fragments of theear l iest formed lunar crust . Apol lo i ,5, 154-157.

R ro l r v , W . I . , R . Bn r r r , R . J . Wr l l u . vs , H . TAKEDA nNo R . W.

BnowN (1972) Petro logy of Fra Mauro basal t 14310 PLC 3,I 59- I 70.

- , N J Hunreno, J. M. RHoDEs, H. WrrsvrNN eNo B.

BeNsnr- (1973) The petro logy of lunar breccia 15445 and pet-

rogenetic implications. J Geol 81, 621-631.RlNcwooD, A E. eNo E EssENE (1970) Petrogenesis ofApol lo I I

basal ts, internal const i tut ion and or ig in of the Moon. PIC 1,769-799.

Rorpnsn, E. lNo P W. Wprsr-rN (1971) Petro logy of s i l icate mel t

inc lusions, Apol lo I I and Apol lo l2 and terrestr ia l equivalents.PLC 2,507-528

- AND - (1972a') Occurrence of chromian, hercyniticspinel ( "p leonaste") in Apol lo l4 samples and i ts petro logic

implications. Earth Planet Sci. Lett. 15,376-402.- AND - (1972b) Petrographic features and petrologic

signi f icance of mel t inc lusions in Apol lo l4 and l5 rocks. PLC 3,

251-279- AND - (1914) Petrology of clasts in lunar breccia

6T9ts PLC 5, 303-3r8.Rosoen, P. L. nNo R F Etrsup (1970) Olivine-l iquid equi l ib-

r\um. Contrib. M ineral Petrol. 29, 2'7 5-289.Ross, M., J. S. HursNrrn ero E. Dowrv (1973) Delineation of the

one atmosphere augite-pigeonite miscibility gap for pyroxenesfrom lunar basalt 12021 Am. Mineral 5E, 619-635.

Rvorn, G., D B Srossrn, U. B. M,cnvlN lNn J F Bowen (1975)Lunar granites with unique ternary feldspars. PLC 6,435-449.

S. J. Plcrrnr AND H. A. ALPERIN (1973) Shock

effects in experimentally shocked terrestrial ilmenite, lunar ilme-

ni te of rock f ragments in l - l0mm f ines (10085,19), and lunar

rock 60015 .127 . PLC 4 ,841 -859 .

Scorr . E. R. D. (1972) Chemical f ract ionat ion in i ron meteor i tes

and its interpretation. Geochim Cosmochim. Acta, 36,

t205-1236.SH,+Nrl lNo, T. J. (1975) Electr ical conduct ion in rocks and miner-

als: parameters for interpretation. Phys Earth Planel- Interiors,

t0,209-2t9.SruxrN. T. , A. F. NooNAN, G. S Swtrzen, B. M,orsoN, J. A. NEt-sN

nNo W. G. MELSoN (1973) Composi t ion of Apol lo 16 f ines

60051, 60052, 64811, 648t2, 67711, 67712, 68821, and 68822.

PLC 4,279-289.SrruoNns, C. H. , W. C PHtNrrv nNo J. L. WanNrn (1974) Petro-

graphy and c lassi f icat ion of Apol lo 17 non-mare rocks wi th

emphasis on samples f rom the Stat ion 6 boulder. PLC 5,

3 37-353.- , J . L. WnnNen AND W. C. PHrNNrv (1973) Petro logy of

Apol lo l6 poik i l i t ic rocks PLC 4,613-632.

SrvpsoN, P. R. AND S H. U. Bowlr (1970) Quant i tat ive opt ical

and electron-probe studies of opaque phases in Apol lo I I sam-

ples PLC 1, 873-890SrrNNsn, B. J. (1970) High crystal l izat ion temperatures indicated

for igneous rocks f rom Tranqui l l i ty Base. PLC 1, 891-895.

Surru, J. V (1974) Lunar mineralogy: A heavenly detective story.

Presidential Address, Part L Am. Mineral. 59,231-243.- (1976) Development of the Earth-Moon system with impli-

cat ions for the geology of the ear ly Earth. In, B. F. Windley, Ed. ,

Proc NATO Adu. Study 1nsl. Leicester, 1975. John Wiley and

Sons. London.- , A. T. ANDERSoN, R C NswroN, E. J. OlsrN, P. J.

WyLLrE, A. V. CREWE, M. S. Is l lcsoN nlu D. JoHNsoN (1970)

Petrologic history of the Moon inferred from petrography, min-

erafogy, and petrogenesis of Apol lo I I rocks PLC I ,897-925.- AND J. B. DnwsoN (1975) Chemistry of Ti -poor spinels,

i lmeni tes and rut i les f rom per idot i te and eclogi te xenol i ths.

Phys. Chem Earrh, 9, 309-322.- AND I . M. Srnr ls (1974) Intergrowths in lunar and terres-

trial anorthosites with implications for lunar differentiation. Am.

Mineral 59, 673-680.Svvru, J. R. (1974a) Low orthopyroxene from a lunar deep crustal

rock: a new pyroxene polymorph of space group P2rca. Geophy.

Res Let t . 1,27-29. See also PLC 6,821-832.- (1974b) The crystal chemistry of armalcolites from Apollo

17. Earth Planet Sci Letr .24,262-270.- (1975) A crystal s t ructure ref inement ofan anomalous lunar

anorth i te. LS v l ,756-758. See also PLC 6,821-832.

SNnrsrNcrn, K. G., K. Krr l nro T. E. BuNcH (1967) Chromite

from "equilibrated" chondrites. Am. Mineral. 52' 1322-1331.

SoNrrr , C. P. (1975) Solar-wind induct ion and lunar conduct iv i ty .

Phys Earth Planet Interiors, 10, 313-322

SrsEI-p, I . M. (1972) Chromian spinels f rom Apol lo 14 rocks.Earth Planet. Sci Leu 14,190-194.

- (1974) Ilmenite and armalcolite in Apollo 17 breccias. Am.Mineral 59, 681-689.

- (1975) Mineralogy of lunar nor i te 78235: second lunaroccurrence o( P2,ca pyroxene from Apollo 17 soils. Am. Min-e ra l 60 ,1086 -1091 .

- AND J . V . Su r r s ( 1971 ) M ine ra l ogy o f Apo l l o 15415 "Gen -esis rock": source of anorthosite on Moon. Nature, 234,I 38- 140.

- AND - (1972) Ultrabasic lunar samples. Nature Phys.Sci 240,5-6.

- AND - (1973) Mineralogy and petro logy of someApol lo l6 rocks and f ines: general petro logic model of Moon.PLC 4, 5t9-536

- AND - (1975a) Lunar touchstones: I Minerals, I I RockTypes. LS V| ,765-770.

- AND - (1975b) Minor e lements in lunar o l iv ine as apetro logic indicator . PLC 6, 451-467.

- , - AND L. GRoSsMAN (1972) Mineralogy and petro logyof Apol lo l5 rake samples Apol lo 15, 16l-164.

Srswnnt, D B. , M Ross, B. A. MoncnN, D E. Appr-Evnl , J . S.Hue sNsn AND R. F. Covlr r lu (1972) Mineralogy and perrologyof lunar anorrhosi te 15415. LS I I I .726-728.

S ro rssn , D . 8 . , U B . MARvrN , J A Wooo , R . W. Wo lp r nNo J .F Bowrn (1974) Petro logy ofa strat i f ied boulder f rom SouthMassi f , Taurus-Li t t row PLC 5, 355-377.

Surc, C. , R. M. Anu-Ero AND R. G. BunNs (19'14) TI3+/Ti '+rat ios in lunar pyroxenes: impl icat ions to depth of or ig in of marebasa l t magma . PLC 5 ,717 -726 .

TAKEDA, H (1973) Inverted pigeoni tes f rom a c last of rock 15459and basal t ic achondr i tes. PLC 4,875-885.

- , M. Mry,cN,roro AND A. M. Rrro (1974) Crystal chemicalcon t ro l o f e l emen t pa r t i t i on i ng f o r coex i s t i ng ch ro -mite-ulvdsprnel and pigeoni te-augi te in lunar rocks. PLC 5,'127-741

.Tnnnsov, L S. , M A. Nesnnov, I . D SutvelsEvsKy, E. S.

Mnr,cnov AND V. I . IvANov ( t973) Mineralogy of anorthosi t icrocks f rom the region of the crater Apol lonius C (Luna-20).PLC 4,333-349

Tnv lon , G . J . , M . J Du r r , M . E . He l l eu , U . B . Mnnv rN l uo J .A. WooD (1973) Apol lo 16 strat igraphy: the ANT hi l ls , theCayley Plains and a pre- lmbr ian regol i th. PLC 4,553-568.

- AND D. HrynlNr l (1971) The format ion of c lear taeni te inordinary chondr i tes. Geochim Cosmochim. Acta, 35, 175-188.

TAyLoR, L. A. nuo R. H. McCel l rsrrn (1972) An exper imentalinvestigation of the significance of zirconium partitioning inlunar i lmeni te and ulv6spinel . Earth Planet . Sci . Let t . 17,105 -109

AND O. SARDr (1973a) Cool ing histor ies of lunarrocks based on opaque mineral geothermometers. PLC 4,8 t 9-828

-, H. K Mro rNn P. M. Belr (1973b) , .Rust" in rhe Apol lo16 rocks. PLC 4,829-839.

AND _ (1974a) Ident i f icat ion of the hydratediron oxide mineral akagan6i te in Apol lo l6 lunar rocks. Geology,l ,429-432.

AND _ (1974b) B-FeOOH, akagan6i te, in lunarrocks. PLC 5,743-748.

u t 5

- AND K C. Mrsne (1975) Mineralogy and petro logy ofApo l l o l 5 rock 15075 LS 21 ,801 -803 .Seea l so PLC6 , l 65 -179

- , D . R . Un ln ,+NN, R . W. Hopp rn e ro K . C . Mrsn l ( 1975 )

Absolute cool ing rates of lunar rocks based on the k inet ics ofZrdi f fus ion in opaque oxides: appl icat ions to Apol lo l5 rocks f romElbow Crater . LS V| ,798-800 See also PLC 6, l8 l -191.

- AND K. L. Wrl l r l l . rs (1973) Cu-Fe-S phases in lunar rocks.Am Mineral. 58, 952-954.

- AND - (1974) Format ional h istory of lunar rocks:applications of experimental geochemistry of the opaque miner-als. PLC J, 585-596.

AND R. H. McCnlrrsrrn (1972) Stabi l i ty re lat ionsof i lmeni te and ulvdspinel in the Fe-Ti-O system and appl i -cation of these data to Iunar mineral assemblages. Earth Planet.Sci Lett 16,282-288.

TAYLoR, S. R. (1975) Lunar Science A Post-Apollo View Per-gamon, New York.

Tnet l t - , R J. , A G. Pr, rNr AND J A. V. Doucrns (1970) Garnet :f i rs t occurrence in the lunar rocks. Science,169, 981-982.

Tsly, F AND D. H Ltvr (1974) Ferromagnet ic resonance studiesof thermal effects on lunar metallic Fe phases. PLC 5,2737 -2746.

- , S L. MANATT rrrp S I . CHnx (1973a) Magnet ic phases inlunar fines: metallic Fe or ferric oxidesl Geochim. Cosmochim.A c t a , 3 7 , l 2 0 l - l 2 l t .

D . H . L r vs AND S . I . CHrN (1973b ) Me ta l l i c Fephases in Apol lo l6 f ines: Their or ig in and character is t ics asrevealed by electron spin resonance studies. PLC 4, 2751-2761

Ussr lueN, T N (1975) Exper imental approach to the state of thecore: Part l . The l iquidus re lat ions of the Fe-r ich port ion of theFe-Ni-S sysrem from 30 to l00kb. Am. J. Sci .275,278-290

-, G. E. LorcnrN, C. DoN,qlosoN AND R. J Wrr_lrnvs (1975)

Exper imental ly reproduced textures and mineral chemistr ies ofh igh t i tanium mare basal ts LS VI, 835-837. See also PLC 6,997-1020

Wn, C. M rNo C. R. Kxowr-rs (1972) The metal phase of theBustee enstatite achondrite. Mineral Mag 38, 627-629

Wn l r rn , D . , T L . Gnove , J . Lo r cHr , E . M S ro lpsn nNo J F .Hr,vs (1973a) Or ig in of lunar fe ldspathic rocks. Earth PlanetSc i . Le t r 20 , l l 25 -1316 .

- , J . Lo r l cH I , T . L . Gnov r , E . S ro l pen AND J . F HAys(1973b) Exper imental petro logy and or ig in of rocks f rom theDescartes Highlands PLC 4, 1013-1032.

E M. Srolern, T. L. Gnovr AND J F. Hlys (1975)

Origin of titaniferous lunar basalts. Geochim. Cosmochim Acta,39, t2t9-t235

Wl l r r n , L S . , B M . FRENcH, K F . J . HErNRrcH , P . D . LowMAN,Jn. , A. S DolN aN I . Aolsn (1971) Mineralogical studies ofApol lo l2 samples. PLC 2,343-358.

WANre , H , F W lo rz rn , E . Jecou rz AND F BEGEMANN (1970 )Composi t ion and structure of metal l ic i ron part ic les in lunar" f i nes . " PLC l , 931 -935 .

- AND 9 orHsns (1971) Apol lo l2 samples: Chemical compo-sition and its relation to sample locations and exposure ages, thetwo component or ig in of the var ious soi l samples and studies onlunar metal l ic part ic les PLC 2, | 187-1208.

W ln r . D . A . A . F . R r rn . J . F . Lov rn r r c AND A . EL Gonssy(1973) Zirconol i te (versus z i rkel i te) in lunar rocks. Z.S 12,

LUNAR MINERALOGY

-, G KuI- lsnuo l t to W. B. Bnvrr (1971) Opaque miner- '164-766'

alogy and textural features of Apollo l2 samples and a com- WARNER, J.L.(1972) Metamorphism of Apollo 14breccias. PLCpar ison w i th Apo l lo l l rocks . P IC 2 , 855-871. 3 ,623-643.

l l t 6

- , C. H- SrvoNos l rqu W. C. PHrNNey (1974) Impact- inducedfract ionat ion in the lunar h ighlands. PLC 5,379-39'7

Wrsn-Ewsrt, P. (1974) Possible magnetic effects due to fine par-ticle metal and intergrown phases in lunar samples. The Moon,l l , 3 0 t - 3 t t .

WessoN, J. T. , C. Cuou nNo W. V. BoyNroN (1975) Temporaldecrease of s iderophi le/ i r id ium rat ios in mature lunar soi ls . LSvr,857-859.

- , R. Scunuov, R. W. Br lo AND C. CHou (1974) Mesoside-r i tes I . Composi t ions of their metal l ic port ions and possib le

relationship to other metal-rich meteorite groups. Geochim. Cos-mochim Acta. 3E. 135-149.

- AND C. M . Wnr ( 1970) Composi t ion of the metal , schreiber-s i te and perry i te of enstat i te achondr i tes and the or ig in of ensta-t i te chondr i tes and achondr i tes Geochim. Cosmochim. Acta,34.I 69- t 84.

WscHsr-En, B. A. , C. T. Pnpwrrr lNro J. J. Plerrs (1975) Structureand chemistry of lunar and synthetic armalcolite LS VI,860-862

Wrrrs, R A (1973) Ferromagnet ic phases of lunar f ines andbreccias: e lectron magnet ic resonance spectra of Apol lo l6 sam-ples PIC 4,2763-2781.

-AND D. PnrsrEl (1974) Ferromagnet ic resonance propert ies

of lunar f ines and compar ison wi th the propert ies of lunar ana-logues PLC 5,2709-2728

Wsrs r - sN , P W, B . N . PowELL AND F K . A r r r nN (1974 ) Sp ine l -bear ing fe ldspathic- l i th ic f ragments in Apol lo l6 and l7 soi lsamples: c lues to processes of ear ly lunar crustal evolut ion. PLC5 ,749 -767 .

- AND E. Roroosn (1973) Petro logy of mel t inc lusions inApol lo samples 15598 and 62295, and of c lasts in 67915 andseveral lunar soi ls . PLC 4,681-703.

Wrur. E . A. Glnuser. . H ScHwnl len AND V. TRoMMSDoRFF

J V SMITH AND I M. STEELE

(1972) Twin laws, opt ic or ientat ion, and composi t ion of p lagio-

c lases f rom rocks 12051, 14053 and 14310' PLC - t , 581-589

Wrur. H.-R., W F. MUrl ln AND G THoMAs (1973) Ant iphase

domains in lunar p lagioclase. PLC 4,909-923.

Wrlr rNrNc, L L. nro E. ANDERS (1975) Some studies of an

unusual eucrite: lbitira. Geochim Cosmochim. Acta, 39,

1205-1210Wrl l r lns, K. L. nxo L. A. Tnvlon (1974) Opt ical propert ies and

chemical composi t ions of Apol lo 17 armalcol i tes. Geology, I '

5-8.Wlorzre, F. , E. J,ccourz, B. SrrrrEI- , H. BeloeulusrN, A.

BALACEscu eNo H WAurn (1972) On lunar metal l ic part ic les

and their contr ibut ion to the t race element content ofApol lo l4

and l 5 so i l s . PLC 3 ,1077 -1084 .- , B. Spetrs l nro H. WANrr (1973) On the composi t ion of

metal f rom Apol lo l6 f ines and the meteor i t ic component. PIC

4, 1483-1491.Wooo, J. A (1967) Chondr i tes: their metal l ic minerals, thermal

histor ies, and parent p lanets. Icarus,6, l -49.- (1975) Lunar petrogenesis in a wel l -st i r red magma ocean.

PLC 6 ,1087 -1102- . U B . Mnnv lu , J . B . Rs lo , Jn . , G . J . TAYLoR, J . F . Bowrn ,

B. N PowElt- r .No J. S. DIcxev, Jn (1971) Mineralogy and

petrology of the Apollo l2 lunar samples. Smithsonian Astrophy-

sical Obseruatory Spec. ReP 333.

Yurun, T nun S. S. Hnrrsn (1974) Cat ion dist r ibut ion and

equilibrium temperature of pigeonite from basalt 15065. PLC 5''169-784.

Yoosn, H. S. Jn (1973) Mel i l i te stabi l i ty and paragenesis.

Fortschr Mineral. 50' 140-173.

Manuscript receiued, January 12, 1976; accepted

for publ icat ion, May 11,1976.

lunar mineralogy: a heavenly detective story. part ii1 · lunar mineralogy: a heavenly detective...

Documents