American Mineralogist, Volume 59, pages 231-243, Ig74

Lunar Mineralogy: A Heavenly Detective StoryPresidential Address, part 11

Josnpn V. SurrsDepartment ol the Geophysical Sciences, Uniuersity ol Chicago,

Chicago,Illircis 60637

Abstract

A general introduction emphasizes the strolg interactions between mineralogy and relatedsciences, and the need for cooperative programs among the various mineralogical groups.

To provide maximum access to the literature, the bulk of lunar rock types are classifiedinto ANT, FETI, and KREEP groups. The Anorthositic-Noritic-Troctolitic compositions mostlyoccur as metabreccias, and are believed to come from the crust. The basalts rich in Fe and Tiresult from late lava flows in mare basins. The KREEP suite of basaltic compositions occursmostly as glasses and metabreccias rich in K, REE, P, and other large-ion elements; it prob-ably results from a late liquid in primary differentiation of the Moon, or partial melting ofANT rocks, or both. Some lunar rocks are hybrids formed by melting of soils, Metamorphismis widespread. Minor rock types include peridotite, spinel allivalite and troctolite, and otherfragments containing Mg-rich minerals. Granitic material occurs rarely. Minor meteoriticfragments occur.

Major mineral groups are pyroxene, feldspar, olivine, spinel, pseudobrookite (armalcolites),silica (quartz, tridymite, cristobalite), Fe-rich types (iron, troilite, ilmenite), phosphates(apatite, whitlockite), and Zr-minerals (zircon, baddeleyite). Minor silicates include pyroxfer-roite, amphibole' garnet' and tranquillityite. Minor sulfides are mackinawite, pentlandite,cubanite, chalcopyrite, and sphalerite. Minor oxides are rutile and corundum. Minor metalsinclude copper, brass, and tin. A Zr-rich mineral is either zirkelite or zirconolite. Schreibersite,cohenite, and niningerite probably derive from iron meteorites. Rusty areas probably consistof goethite perhaps associated with hematite and magnetite; presence of Cl suggests possibleterrestrial oxidation of meteoritic lawrencite.

Important features of feldspars are: mostly calcic plagioclase, but rare K,Ba-feldspar inresidua; correlation of Mg,Fe,K, and Na with rock type giving test of differentiation models;Eu anomaly depends on temperature, oxidation state, etc; solid-solution of ca(Mg,Fe)sLo, idFETI plagioclase; exsolution of pyroxene, silica, and Fe in ANT plagioclase.

Important features of olivines are: correlation of mg [- atomic Mg /(Mg + Fe)] with rocktype giving test of differentiation models; Ca may distinguish volcanic from plutonic environ-ment; cr enters as divalent ion, and may provide guide to oxidation state; Mg,Fe orderingof some olivines indicates long annealing; melt inclusions give magma composition; high Alin some olivines may correlate with spinel exsolution; melt inclusions give magma cornporitioo.

Introduction

This written address differs considerably lrom theoral presentation. Because ol time limitations, thegeneral introduction was omitted. Seueral handredmin:erdogists and peffologists worked o,n the lunarprogram, and most ideas were discouered almo,stsimultaneously by seaeral persons or groups. In theoral presentation, relerences to indiuidual scientistswere deliberately omitted to auoid tedious repetitionand inuidious selection of names. In this written

l Presidential Address, Mineralogical Society of America.Delivered at the 54th Annual Meeting of the Society, 13November 1973.

presentation many relerences ore giuen, but theseare highly selectiue and chosen for conuenience andnot lor priority or prestige. In particular, relerencestend to concentrate on controaersial items and onpapers submitted lor publication, Matters of com-mon krwwledge ffe not fally documented.

The Presidential Address provides an opportunityfor the Retiring President to philosophize or to rvview some aspect of his work or of mineralogy ingeneral. I shall briefly pontificate, and then use areview of lunar mineralogy as an implicit illustrationof my main theme: that the distinctions betweenmineralogr and allied sciences (including geochemis-

231

IOSEPH V. SMITH

try, geophysics, petrology, and crystallography) are

blurring so rapidly that research mineralogists should

become competent in as many areas as possible, and

should collaborate closely with scientists lvith related

and overlapping interests.It is quite impossible to understand the phase

equilibria of (say) plagioclase feldspars without in-formation on the following: (1) controlled labora-

tory syntheses at high temperature with X-ray dif-

fraction and electron-optical characterization of theproducts; (2) chemical, X-ray diffraction, and elec-

tron-optical characterization of plagioclases from

natural rocks ranging from volcanic to metamor-phic; (3) heating studies of natural specimens. In

addition, calorimetry and resonance techniques, plus

many others, provide further knowledge. A crystal-lographer without a thermodynamic backgroundmight misinterpret the phase relations, while a geo-

chemist might overlook the importance of domain

texture in the occurrence of trace elements. The

electrical conductivity and diffusion coefficients of

ferromagnesian minerals depend strongly on struc-

tural defects and the oxidation state, thereby havinga profound effect on geophysical interpretation ofphenomena such as creep in the mantle and the tem-perature profile of the Moon. The routine use of the

electron microprobe (hopefully to be joined by the

ion microprobe for trace elements and isotopic ra-

tios) brings geochemical data within the immediate

control of the mineralogist and petrologist. The or-ganization of research is beginning to utilize the

power of coordinated studies, as in the mineralogy-petrology-geochemistry consortia for lunar speci-

mens. Many mineralogical problems must be studied

by groups of investigators, often geographically dis-persed; such investigators should be diplomats as

well as scientists.

Unquestionably, it is important to develop strongcooperative programs among the various mineralog-ical groups (and also among the various geologicalsocieties); otherwise prolonged isolation leads todistrust and antagonism. It would be tragic if theworld of mineralogy split up into separate camps,each fearful and contemptuous of the others.

Finally, I want to mention the practical uses ofminerals. Agriculture depends on the proper inter-action between the inorganic minerals and the or-ganic content of soils. Many chemical and all metal-lurgical industries depend on extraction of minerals,some of which are becoming scarce. Novel uses of

minerals, such as the employment of zeolites ascatalysts in the petroleum industry and the appliea-tion of mineral analogs as lasers and solid-statedevices, should attain increasing importance; forexample, several rare minerals with strong pleochro-ism may have potential electronic uses' For this rea-son, mineralogical societies should also collaborateclosely with chemical and physical societies.

No mineralogist can rest on the knorvledge gainedin his formal education. One of the great advantagesof the lunar program was the opportunity for thescientists to meet closely with practitioners in otherdisciplines. Some mineralogists gained a workingknowledge of such apparent esoterica as a europiumdeficiency and the time relation of BABI and ADOR.It is extremely important to provide opportunitiesfor mineralogists to expand their knowledge, and theMineralogical Society of America should become ac-

tive in organizing short courses and symposia, and

in publishing critical monographs and treatises'So much for the pontificating; now for lunar min-

eralogy with its illustration of the strong intercon-nections between mineralogy, petrology, geochemis-

try, and geophysics.Fortunately, not all mineralogy requires the big

battalions with the kilo- or mega-buck. A perennial Lunar Mineralogyjoy of mineralogy is the scope for the collector andthe artist, for the field man and the museum curator. Introduction

Some modern artists examine the world for existing The completion of the Apollo missions provides a

natural beauty; collectors of minerals have been suitable time to evaluate the status of lunar miner-

doing this foi centuries. Mining operations contin- alogy. An article "How the Apollo program changed

ually reveal new mineral deposiis of spectacular in- the geology of the Moon: suggestions for future

terest, such as the one at St. Hilaire. Some mineral studies" by Smith and Steele (1973) gives a general

museums expand their collections and put on mag- background. Briefly, the Moon was in orbit around

nificent dispiays; others deteriorate. Probably most the Earth at least 4 giga-years ago; it underwent

mineralogists recognize the value of a pluralistic ap- major chemical differentiation at least 3'9 giga-years

proach, *a uppr.iiute the mutual interests and dif- ago (and perhaps as early as 4.5 giga-years ago)

ierent styles of the various types of mineralogists. producing a crust, a mantle, and perhaps a core; in-

LUNAR MINERALOGY 233

tense bombardment devastated the surface to depthsof several tens of kilometers, perhaps with significantlate accretion of the incoming projectiles; lava flowscovered impact basins with the last activity occurringnear 3 giga-years ago (though some unexploredareas of the Moon may have later extrusions); andcontinuing bombardment by relatively small projec-tiles produced a surface covered by debris (regolith).Geophysical and geological data show that thepresent crust of the Moon consists of a loose regolithsome meters thick lying on some tens of kilometersof material with low seismic velocity. In the mareregions, the upper parts are dominated by basaltsrich in Fe and Ti, whereas in the highland regions,the dominant rocks are probably rich in calbicplagioclase with subordinate pyroxene and olivine.The nature of the hypothetical mantle is controver-sial, but many workers believe that it is dominatedby magnesian olivines and pyroxenes. Figure I, amodel of the Moon proposed after the Apollo 11mission, may provide a crude description for a timeprior to 4 giga-years ago. There are some very dif-

Thin c rus l (5 -2Okm)under lo in by l iqu idImpocls creote

P log ioc lose - r i chrock ond bosolfs

P =2.8- l . t gm/cmJ

lorge lovo lokes( seos

._ ei?,r' - Fors ideick Crus t ,

l i f t l e o r n ounderlying l iquid

Liquid feirobosotl

Impocls creotedeep c ro le rswi lh l i f f le "o rno lovo .p = 3 . 3 g m / c m 3

(drown to neors ideby Eor th 's g rov i tydue to h igh dens i l y

t ' t h i n l o y e r o f s u l f i d e

ficult geochemical and geophysical problems, espe-cially concerning (1) the sources of energy and thetiming of igneous events on the Moon, and (2) theultimate origin of the bulk chemistry of the Moon.

The mineralogy and petrology of the Moon pro-vides a fascinating challenge to the lunar detective.Few samples returned to Earth are simple igneousrocks. Most consist of complex breccias or fine-grained soils. In spite of careful selection of prospect-ing sites and detailed documentation (too detailedfor the temper of many lunar investigators faced withsample description forms!), it is very difficult to pindown the actual starting point of most lunar samples.Fortunately the mineralogy of the Moon is rathersimple compared to that on the Earth, comprisingonly some few dozens of species compared to sev-eral thousand (ultimately, of course, the full rangeof meteoritic minerals should be found on theMoon). There are.few rock types (though onemight not think so from the confusing nomenclaturesin the literature!). In general, it is possible to ar-range most of the information on lunar rocks andminerals into plausible sequences consistent withcrystal-liquid difterentiation, metamorphism, hybridi-zation. The major uncertainties relate to geochemicaland geophysical problems: what were the time se-quences of crystal-liquid fractionation; what was theextent of late accretion; how important were re-melting and metamorphism; what are the implica-tions for the internal structure and the bulk com-position of the Moon? In attempting to answer thesequestions, the mineralogic detective must bear inmind that the regolith is dominated by crustal ma-terial, and that deepseated material should be sparseand found mainly as fragments in breccias frommajor impacts.

General statement on lunar rock typesond i lmen i ie - r i chr o c k . p : 5

compored lo overlying crusf') The nomenclature of the lunar rocks is in a sham-Ftc. 1. A possible model for the Moon during primary bles. Unfortunately many lunar rocks are actually

differentiation prior to 4 giga-years ago. This m9!-e] was metamorphosed breccias or remelted regolith, aniproposed in l97o before detailed characterizatio" "l-*fP!" ,o,n" ,ru.". are given on the basis of bulk chemistryrocks. Major reworking of the outer part of the Moon "".:'- ":"'-

would occur in the neit giga-year, and the FEn-(;;.;; rather than the petrology and mineralogy' Further-basalts would appear in this period. The ANT group might more, the chemistry of the minerals is biased towardsconsist of the '"plagioclase-rich

rock and basalts" plus fur- alkali-poor compositions and reduced states of transi-ther differentiates from the ferrobasalt. The KREEp group tion metals with respect to terrestrial analogs, therebymight arise from both the late "ferrobasalt" and reme.lting causing f,oUt"11, when terrestrial names are appliedof the ANT group. The FETI basalts might result from --l;

'* t "remelting of cumulates of pyroxene and oxide ;";;; to lunar rocks.followed by complex fractional crystallization. Minor ultra- The following nomenclature and description ofbasic rocks might be cumulates from the upper mantle and the major rock types are chosen to maximize initialcrust. From Smith er al (1970). access to the literature:

234 IOSEPH V. SMITH

ANT suite: (name from Anorthositic-Noritic-Troctolitic; Keil el aI, 1972) comprised mainly ofcalcic plagioclase, orthopyroxene, and olivine; rang-ing from anorthositic to noritic and troctolitic inbulk composition. Mostly occurs as metamorphosedbreccias (often monomict), fragments, or glasses.Commonly assigned to crust. Believed to result fromearly differentiation with gravity separation of crys-tals and liquid. Terms like gabbro, anorthositic gab-bro, and highland basalt are also used.

FETI suite: (new name referring to high Fe andTi content) comprised of basalts rich in Fe and Ti;dominated by zoned pyroxenes (mostly augitic),calcic plagioclase, and ilmenite. Olivine may or maynot occur. The final residuum contains several rareminerals and has a granitic component. Mostly oc-curs as rock fragments, but some glasses occur. In-terpreted as the result of late lava flows in marebasins, probably mostly frorn the tops of lava col-umns several kilometers thick.' Origin controversialbut perhaps from remelting of cumulates rich inpyroxenes and opaque minerals, modified by near-surface differentiation in lava suites. Usually calledmare basalt in literature.

KREEP suite: (name is acronym given by Hub-bard et aI, l97la) a description introduced for ba-saltic compositions rich in K, REE, P,Zr,Ba, U, andTh. Mostly occurs as glasses and meta-breccias buta few crystalline rocks with basaltic texture occur.Dominated by plagioclase, low-Ca pyroxene andminerals such as apatite and zircon which containthe above characteristic elements. Origin controver-sial but perhaps results from either liquid producedat the crust-mantle interface during primary difter-entiation, or liquid from partial melting of ANTcrustal rocks, or both. Other terms including high-alumina basalt are also used.

There are some problems with this three-foldnomenclature: (1) There is evidence of both low-and high-K anorthosites (e.9., Hubbard et al.,l97lb); perhaps the former were produced by flota-tion of plagioclase during formation of the crust, andthe latter from crystal-liquid differentiation of aK-rich liquid produced by partial melting of thecrust; if so, the former might be assigned to ANT,and the latter to KREEP. (2) Some rocks such as14310 almost certainly were produced by remeltingof regolith (e.g., Dence and Plant, 1972), and there-fore have a bulk composition derived from manycomponents including meteoritic infall (Morganet aI, 7972); many breccias show metamorphic tex-

tures (e.g., Anderson et al, 1972; Chao, Minkin, andBest, 1972 Warner, 1972), and it is not alwaysobvious what rocks and minerals composed theoriginal breccia. (3) The applicability of the termshigh-alumina basalt and very high-alumina basalt(e.g., Bansal et al, 1973) with respect to other rockdescriptions is not fully clear at the moment (seelater); whether such rocks represent a distinct igne-ous suite, or whether they are hybrids formed bymelting of regolith, or both, is unclear.

On the whole the above three-fold grouping coversmany of the lunar rocks, correlates fairly well withthe chemistry of the component minerals, and per-haps also correlates with chronological developmentof lunar rocks. It certainly allows a preliminary ex-amination of the petrogenetic significance of min-eralogical properties.

The nature and extent of the minor rock types israther uncertain. Crystal-liquid differentiation ofbasaltic compositions leaves a final immiscible liquidwith granitic affinity (e.g., Roedder and Weiblen,1970). Minor glassy or crystalline representativeswere found in the lunar samples, but whether largerock bodies of granitic composition occur is uncer-tain. Grains of olivine and pyroxene with very Mg-rich compositions occur as a minor constituent in allthe lunar soils (see many papers), and spinel alliva-lite and troctolite samples (e.g., Agrell et al, 1973;Prinz et aI, 1973a), a peridotite clast (Anderson,1.973) and ultrabasic fragments (e.9., Steele andSmith, 7972\ have been described. All these mate-rials are more Mg-rich than equivalents from thebroad ANT suite. A possible source is the uppermantle, but crustal processes or accretion must beconsidered. Rare fragments of iron meteorites havebeen found with characteristic textures, and otheriron specimens are believed to repfesent melted ormetamorphosed meteorites on chemical evidence.Contributions of stony meteorites to the lunar fineshave been recognized from bulk chemistry (e.9.'

Ganapathy et al, l97O), and an enstatite chondritehas been found (HaggefiY, 1972a).

Types ol Lunqr Minerals

The major minerals to be described in detail laterare:

Pyroxene

Feldspar

(Mg, Ca, Fe, Al, Ti, Cr, etc)(Si, elc)O3. Very low in Na and K.Mostly anorthite and bytownite, butrare K, Ba-feldspar in residua.

Olivine

Spinel

Mostly Mg-rich but rare fayalite inresidua. Divalent Cr.N MgAlrOn in Mg-rich rocks.Chromite zoned to ulviispinel inFETI basalts.

Pseudobrookite Armalcolite group with divalentiron. Low Cr and Zr in FETIbasalts and high Cr and Zr inmetamorphosed Mg-rich rocks.

Silica Tridymite, cristobalite, and quartz.Fe-rich Iron, troilite, and ilmenite occur,

especially in FETI basalts.Phosphates Apatite and whitlockite, especially

common in KREEP rocks. High inREE, Cl, and F.

Zr-minerals Zircon and baddeleyite typical ofKREEP rocks; also in residua ofFETI basalts.

Minor silicate minerals are: pyroxferroite (analogof pyroxmangite, common in residua of FETI ba-salts; Chao et al,1970; Burnham, 1971), amphiboles(single occurrences of alumino-tschermakite (Denceet al, l97l), and magnesio-arfvedsonite-richterite(Gay, Bancroft, and Bown, 1970) in FETI basalts;high in halogens), garnet (almandine-rich grains ina FETI basalt; Traill, Plant, and Douglas, l97O),tranquillityite (residua of basalts; Lovering et al,I97l), melilite? (uniaxial grains in fines, no X-rayor chemical data; Masson et al, 1972), phyllosilicates?? (in fines, probably terrestrial contaminants;Drever et al, l97O), sphene?? (perhaps in devitrifiedspherule, tentative X-ray powder data; Gay, Ban-croft, and Bown, l97O\.

Minor sulfide minerals are: mackinawite, pent-landite, cubanite, chalcopyrite (all associated withtroilite at margins or as inclusions consistent withexsolution or inversion; optical and chemical iden-tification but no X-ray data; e.g., Taylor and Wil-liams, 1973), sphalerite (e.g., El Goresy, Taylor,and Ramdohr, 1973).

Minor oxide minerals are: rutile (associated withilmenite as lamellae and inclusions indicative ofexsolution, e.9., Haggerty, 1972b; as distinct grainsin a peridotite, Anderson, 1973; and as blue grainsin soils, Jedwab, 1973), corundum (in fines asaggregates, Kleinmann and Ramdohr, 1971; andas large crystal twinned on (0001), Christophe-Michel-Levy bt al, 1973; vapor deposit?, meteoritic?or terrestrial contaminant?), hematite? (lamellaein ilmenite, and associated with goethite? in "rusty"

235

rocks, e.g. Taylor, Mao, and Bell, 1973), and mag-netite? (in "rusty" rocks).

Hydroxylated mineral: "rusty" areas on somerocks contain goethite? or another iron oxide hy-droxide (Agrell et al, 1972), perhaps with hematiteand magnetite; presence of Cl suggests possibility ofterrestrial oxidation of lawrencite, originally derivedfrom cometary or meteoritic contamination on theMoon (El Goresy et al, 1973; Taylor, Mao, andBell, 1973), but lead isotope data suggest that thematerial is indigenous to the Moon (Nunes andTatsumoto, 1973).

Minor minerals, probably largely from iron meteo-rites, are: schreibersite, cohenite, and niningerite.

The nature of the lunar mineral variously com-pared with terrestrial zirconolite or zirkelite is con-troversial, and X-ray structural data are needed toprovide a control for assignment of atoms: either(Ca, Fe, etc) (Zr, etc)zOz for zirconolite or 1:1:5for zirkelite (Wark et aI, 1973; Busche et al, 1972;Peckett, Phillips, and Brown, 1972; Brown et aI,1972). Other Zr, Ti-rich minerals may occur (e.9.,Zl: Haggerty. 1972b).

Minor metallic minerals are: copper (tiny grainsnear Fe metal, ilmenite, or troilite), brass (variablecomposition extending outside terrestrial range), tin(embedded in Fe metal), and indium (contaminantfrom rock box).

Aragonite (Gay, Bancroft, and Bown, 1970), SiC(Jedwab, t973), and graphite may be contaminants.

Important Aspects ol Feldspars

In this and succeeding sections, some aspects ofthe major mineral groups important for understand-ing the genesis of lunar rocks and minerals are out-lined.

The great majority of lunar feldspars are calcicplagioclases. K,Ba-feldspars crystallize in the residuaof FETI basalts and occur in KREEP basalts; theymay occur with silica minerals as a granitic rocktyp @.g., fragment of hedenbergite granophyre,Mason et al, l97L). Sodic plagioclases are very rareand may result from minor meteoritic contamination.

The plagioclases in the FETI basalts range fromsector-zoned hollow tubes to anhedral grains inter-gro\ryn with pyroxene. Plagioclases from ANT rocksusually show textures ranging from fine-grained"equilibrium" mosaics to irregular shock textures;probably most textures result from one or moreepisodes of metamorphism (both thermal and shocktypes) following an earlier igneous stage.

LUNAR MINERALOGY

|OSEPH V. SMITH

Electron microprobe analyses of ANT plagioclaseare consistent with the MTaOs formula, but inter-pretation of analyses of FETI plagioclase is unclear.The location of the Mg and Fe atoms is uncertain,and the closest fit to an MT4O8 formula is obtainedif these elements are mostly in the tetrahedral sites.Weill el ol (1971,) outlined the problem, and Wenkand Wilde (1973\ coordinated the available data.Mcissbauer and electron-spin resonance data showthat a small part of the iron is ferric while most isferrous, and that the iron enters two structural sites,one of which is tetrahedral (e.9., Schiirmann andHafner, 1972a; Hafner, Niebuhr, and Zeka, 1973);see also optical absorption data (Bell and Mao,1973).

The composition of the calcic plagioclases cor-relates with the host rock. Figure 2 is a plot of wtpercent Fe os mole percent Ab inferred from the Nacontent. The plagioclases from the FETI basaltshave the highest Fe contents, and there is a definitetendency for the Fe content to increase from about0.5 wt percent at Ab5 to 1 wt percent near Abrz.The plagioclases from the KREEP basalts tend to bemore sodic, mostly occurring in the range from Abroto Abzs: again the Fe content correlates with Ab,but the concentration is lower than for FETI basalts.The third main group of plagioclases, that from ANTrocks, tends to be highly calcic with mcst compos!tions near Abs. The composition ranges of ANT and

t . 2

- 0.8r

0

KREEP

b o s o l l sx

- - - - - - z -

- - - - -

x x x - - \ __ I l o r i - - t l t" I x i - - J - b - , x _ _ _ )-20 r00

mol . % AbFrc. 2. Fe us Na content of lunar plagioclases. Schematic

drawing showing principal ranges for FETI basalts, KREEPbasalts, and ANT materials. P shows the plagioclase inperidotite clast 15445,10 (Anderson, 1973) and PST thatfor pink-spinel troctolite (Prinz et al, 1973a). The rangesfor plagioclases from KREEP and ANT rocks probablyoverlap. The crosses show data for 67747 feldspathic basaltwhich contains large poikilitic olivines enclosing feldsparlaths and intergranular pyroxenes (Steele and Smith, 1973).Data from many sources. Data for Mg, Mg/Fe, and K to bepublished elsewhere.

KREEP plagioclases appear to overlap. The Mg andK contents appear to be distinctive for the three typesof plagioclase, but the data are probably confused byrelatively low accuracy of electron microprobe analy-ses for these elements. Smith (1971,) noted that thecomposition bands seemed to diverge from a com-position near Ab3; K 0.03, Fe 0.1, Mg 0.01 wt per-cent, and this turned out to be a common composi-tion for plagioclase in anorthosites (e.9., Steele andSmith, l97l). Plagioclases from a peridotite and apink-spinel troctolite fall near the main concentra-tion range for ANT rocks. As an illustration that thesituation is not simple, data points are shown inFigure 2 for a feldspathic basalt with igneous texturebut Fe content even lower than for KREEP basalts(Steele and Smith, 1,973).

The well-known preference of Na for liquid overcoexisting plagioclase (e.9., Kudo and Weill, 1970)allows the Na content of lunar plagioclase to be usedas a test of differentiation models. The low Ab con-tent of ANT plagioclase is consistent with their originas cumulates from liquids which later producedKREEP plagioclase (note the very calcic plagioclasein the pink-spinel troctolite); alternatively, partialmelting of ANT rocks could yield liquids which giverise to KREEP plagioclase. If FETI basalts resultfrom partial melting of a plagioclase-bearing sourcematerial, that plagioclase would be extremely low inAb content, assuming that equilibrium occurred.

Unfortunately there are few laboratory data onthe partition of Fe, Mg, K, etc between plagioclaseand basaltic liquids, though one might expect that thesubstitution of Mg and Fe in plagioclase would de-crease with temperature, and that the Mg/Fe ratiowill depend both on the oxidation state and the Mg/Fe ratio of the coexisting liquid. Green et al (1972)and Akella and Boyd (1972) reported 0.3-0.6MgO and 0.5-1.0 wt percent FeO in plagioclasessynthesized in the presence of liquid from lunarbasaltic compositions, which values are comparablewith those observed in FETI basalts. Systematicdata are badly needed to allow use of these ele-ments for testing difierentiation models. The mgratio of the plagioclase (where mg : atomic Mg/(Mg + Fe) ) may prove valuable in giving anindication of the parent liquid (e.g., Walket et aI,1973); sparse data of low accuracy suggest thatlng tends to be higher for KREEP than for FETIplagioclase, as expected by comparison with thedata for pyroxenes. Crawford (1973) found thatmg of zoned FETI plagioclase was the same as

d-S

=

for pyroxene crystallizing simultaneously. Detailedmeasurements of the alkali and alkaline-earth ele-ments are badly needed as tests for differentiationmodels; unfortunately the concentrations of theseelements, except for Na and K, are too low foraccurate measurement with an electron microprobe,but hopefully detailed ion-microprobe data will beforthcoming (see Andersen, Hinthorne, and Fred-riksson, 1970, for indicative data).

Interpretation of the minor element content ofplagioclases from ANT rocks, metamorphosedbreccias, and fines must take into account the effectof solid-state reactions which change the contentof minor and trace elements. Steele and Smith(1974) pointed out that the pyroxene and silicaminerals occurring as inclusions inside ANT plagio-clase, and as grains at plagioclase boundaries (e.g.,James, 1972), may result from solid-state expul-sion from plagioclase which originally showed highcontents of Fe and Mg, perhaps comparable to thosefor FETT plagioclase. They proposed that withfalling temperature the component Ca( Fe,Mg) SisOebroke down into pyroxene Ca(Fe,Mg)Si2O6 andsilica SiOe. If this is true, the pyroxenes in somelunar anorthosites consist partly or perhaps wholelyof material exsolved from plagioclase. Neverthelessthe mg content of the pyroxene may provide aguide to that of the liquid from which the plagio.clase crystallized. The best evidence for exsolutionis that of Lally et al (1972) who obtained electronmicrographs of pyroxene grains lying in twin bound-aries of plagioclase. Brett et al (1973), and others,reported tiny oriented rods of FeeeNi, metal inplagioclase grains from Luna 20, and preferredsub-solidus reduction as the explanation.

The europium anomaly has figured in many geo-chemical discussions of lunar rocks. Unfortunatelyall accurate data obtained to date are on bulksamples using neutron activation or mass spectro-metric analyses. Under reducing conditions, euro-pium becomes divalent, whereas its neighboringrare earths remain trivalent. In the divalent state.europium tends to enter the feldspar structurepreferentially, thus giving a positive anomaly in afeldspar and a negative anomaly in a coexistingbasaltic liquid. Unfortunately, the europium anomalydepends in a complex way on the temperature,oxygen fugacity, and bulk composition (e.g., Weilland Drake, 1973) and careful study will be neededbefore quantitative evaluation of lunar rocks canbe made. Nevertheless it seems reasonablv certain

237

that the europium anomalies in lunar rocks can beascribed largely to plagioclaseJiquid fractionation.

There are many electron-diffraction and X-raydifiraction studies of lunar plagioclases which showcomplex domain textures indicative of exsolutionand order-disorder. Unfortunately, the phenomenadepend in a complex way on the bulk chemicalcomposition and on the crystallization and post-crystallization history. At the moment the data areof more value in providing evidence on the phaserelations of plagioclase in general than on the spe-cific problems of lunar mineralogy.

It is quite certain that lunar plagioclase canprovide many clues to the discerning lunar de-tective.

Important Aspects ol Oliaines

The mg value of lunar olivines provides a valu-able guide to crystal-liquid differentiation. Figure 3summarizes the principal ranges of mg for lunarolivines together with corresponding data for pyrox-enes (see next section).

The ANT group of rocks carries olivines mostlyin the range of mg = 0.85 - 0.65 whereas theFETI group carries olivines ranging mostly from0.7 to 0.3 with rare fayalites in the final residua.

(mo l .%)

LUNAR MINERALOGY

EnFor00Ftc. 3. Schematic drawing of major-element concentra-

tions of lunar olivine and pyroxene. Note that considerablezoning occurs and that the ranges cover the principal butnot the total ranges of composition. The pyroxenes of FETIbasalts show final zoning to pyroxferroite or Fe-richpyroxenes, while the olivines zone into fayalite. BothKREEP and FETI pyroxenes give compositions whichtend to concentrate at the Mg-rich side of the ranges. Datafrom many sources. The data for ultrabasic fragments(filled circles) are from Steele and Smith (1972). P andPST refer to peridotite and pink-spinel trocolite (see Fig.2). The rectangle shows the composition range of largecrystals of orthopyroxene in Apollo 17 fines which mayresult from a true norite with coarse texture: veins ofdiopside have the same composition as the filled circle(Irving et al,1974).

FsFo

(v406080

,

KREEP

238

Ultrabasic rocks and fragments carry olivines withmg = 0.93 - 0.83; seven compositions found bySteele and Smith (1972) for clasts and fragmentsare shown in Figure 3 along with that of 0.91 forthe peridotite clast of Anderson (1973) and of0.92 for the pink-spinel troctolite of Pinz et al(1973a). Other data in the literature for frag-ments, clasts, and single grains fall in the samerange. Olivine is rare or absent in KREEP rocks.Figure 4 is a simplified version of Figure 6 of Steeleand Smith (1973\ which shows data on mg ofolivines, Ca-poor pyroxenes, and basaltic liquids.Roeder and Emslie (1970) determined the frac-tionation of Mg between olivine and terrestrialbasaltic liquids, and Medaris (1969) evaluated thefractionation of Mg between olivine and low-Capyroxene. Electron microprobe analyses of syn-thetic lunar samples yield a single band displacedto the left of the one for terrestrial samples, andthere is no clear distinction between olivines andlow-Ca pyroxenes. At liquidus temperatures, mg issimilar for low-Ca and high-Ca pyroxene, but atlow temperature, mg is higher for the low-Ca pyrox-ene. In general, for crystalJiquid fractionation onthe Moon this diagram provides a good test ofdifferentiation models, irrespective of the precisedetails of temperature and pressure.

Consider the olivines of the ultrabasic group.If these are phenocrysts which accumulated from abasaltic liquid, that liquid must have mg rangingfrom about 0.6 to 0.8. This is actually the range

- ^ f o l i v i n e" ' e l l o w - C o

i n t e r e d :o l i v ines !0n0 pxs :of source Irocks of :

tL | | DOSdI$

olivines ofTultrobosic lfrogments-

px

0.6

0.8

ANT ---KR EEPFETI

mg l i qu i d

Ftc. 4. Relation between mg for coexisting liquid andeither olivine or low-Ca pyroxene. The shaded area showsthe range of synthetic data for lunar basaltic compositionsobtained by various workers including Biggar et aI (1971),Green et aI (1971, 1972) and Kushiro (1972). The rangesbelow the abscissa show the bulk composition for the threemajor rock types. For KREEP-type material, the mg valuedrops to zero for extremely-differentiated residua. See textfor further explanation. [Modified from Steele and Smith( 1 9 7 3 , F i e . 6 ) 1 .

of mg for the bulk composition of most ANT rocks.If the FETI basalts result from partial melting ofa rock dominated by olivines and pyroxenes, themg range of the latter should be about 0.6 to 0.8.Partial melting of the ANT suite should give liquidswith mg mostly from 0.4 to 0.6, which overlapsthe ranges for the bulk values of FETI and KREEPbasalts.

The minor elements in olivines vary considerably,and have petrogenetic value. Smith (1977) cau-tioned about problems arising from inaccurate elec-tron-microprobe analyses, especially for secondaryfluorescence of transition metals. Review of theextensive data in the literature shows the following:

( 1) Mn correlates positively with Fe and hasno petrogenetic significance (indeed major devia-tions from the correlation probably indicate an in-correct analysis).

(2) Ca tends to increase with Fa content rang-ing from 0. 1-0.4 wt percent CaO near forsteriteto 0.4-0.9 wt percent for the rare fayalites in residuaof FETI basalts (Fig. 5). The ranges for the FETIand ANT groups overlap, but the latter tend to bericher in Mg and poorer in Ca. Note that manyspecimens in the ANT group are breccias, oftenrecrystallized. Simkin and Smith (1970) found thatthe Ca-content provides a distinction between ter-restrial olivines from plutonic rocks and those fromvolcanic or hypabyssal rocks. In Figure 5 all theolivines from FETI basalts lie in the region forvolcanic and hypabyssal terrestrial olivines; further-more, the dominant range of FETI olivines is closeto the data points found by several investigatorsfor olivines synthesized in the presence of liquidfrom lunar basaltic compositions. The ANT groupoverlaps both the plutonic and the volcanic-hypa-byssal ranges suggesting that this group derives bothfrom volcanic and plutonic rocks. Smith (1971)noted a distinction between a high-Ca and a low-Cagroup on the basis of just Apollo 11 and 12 data,but examination of all available data shows thatthere is a continuous gradation-e.9., the data byPrinz et al (1973b) for ANT fragments in theLuna 20 sample. The effect of metamorphism onthe Ca content of olivine has not been quantifiedso far, but available data on diffusion rates (e.9.,Buening and Buseck, 1973) suggest that meta-morphism in a thick ejecta blanket may cause sig-nificant migration of Ca from the olivine. Hencea low Ca content may indicate either a plutonicorigin, or metamorphism, or both. The possibility

TOSEPH V. SMITH

FETI , ,

ANT\ ? |l t t; /l r /

l - /

_ - - ' i volconic g---;;L-rffiyt__

t ^I P S T A

AA

\---__X P | |

239

relates nicely with data obtained by several workersfor olivines synthesized from melts of lunar basalticcomposition. The band for ANT olivines is displacedto higher Mg values and lower Cr concentrations.Particularly noticeable are the low Cr concentra-tions for the peridotite and pink-spinel troctolite.Currently there is no definitive explanation for thewide range of Cr-contents (when normalized forthe Mg correlation). A simple explanation mightbe that outgassing of the Moon resulted in reductionof Cr3* to Cr2*. Alternatively one might invokeefiects from the bulk composition of the rock, fromthe temperature of equilibration, and from the effectsof simultaneous crystallization of minerals such asspinel and pyroxene (e.9., Butler, 1973). Dodd(1973) reported Cr contents in olivines from theSharps (H-3) chondrite. These occupy the rangefrom 0.02-0.2 wt percent Cr2O3 and 100-70 molepercent Fo, without any correlation between Mgand Cr. Dodd found a weak correlation betweenCr and Al. and favored the occurrence of some ofthe Cr as trivalent with a coupled substitution ofcr3* * Al3* - (Mg,Fe)r- + Si4..

(4) Al is a puzzle. Accurate analysis of Al byelectron microprobe methods is difficult because

n PFI

Fomol. % Fo

Frc. 6. Relation between wt percent CrrO" (or CrO) andmole percent Fo for lunar olivines. The two ringed areasshow the ranges of electron microprobe analyses for ANTand FETI olivines. The majority of the specimens fall inthe regions of Fo marked by the continuous lines. P andPST stand for a peridotite and a pink-spinel troctolite (seeFig. 3). The triangles show synthetic olivines crystallizedfrom liquids or lunar rock compositions (Biggar et a/,1971, Green et al, 1971, 1972; Kushiro, 1972). PB and Cstand for olivine from two picritic basalts and a clast (Gayet al. 1972) .

LUNAR MINERALOGY

o nAo

O

he0.4

E

r .0

0.8

0.2

nFo 80 60 40 20 Fo

mol . % FoFIc. 5. Relation between wt p€rcent CaO and mole per-

cent Fo for lunar olivines. The two ringed areas show theranges of electron microprobe analyses for ANT and FETIolivines. Most specimens fall in the regions of Fo markedby the continuous lines. P and PST stand for a peridotiteand a pink-spinel troctolite (see Fig. 3). The dashed linemarks the boundary between terrestrial olivines fromplutonic us volcanic and hypabyssal rocks. The trianglesshow olivines crystallized from liquids of lunar rock com-positions (Biggar et al, l97l; Green er al, 1971, 1972:Kushiro, 1972).

of metastable crystallization with interaction betweenolivine and other Ca-bearing minerals must be evalu-ated; furthermore the bulk Ca content of the hostrock must be considered. Butler (1972) gave de-tailed data on the zoning trends of four FETIbasalts, and found a correlation of Ca content withtexture indicating the effect of temperature ofcrystallization or of cooling rate.

(3) Cr tends to correlate with Mg, but mayalso correlate with the Cr content of the host liquidand more importantly with the oxidation state. Interrestrial olivines, the Cr content is undetectableby normal electron microprobe methods presumablybecause the Cr is trivalent and is not acceptedeasily by the olivine structure. In lunar olivines, theCr content may rise to 0.n wt percent, presumablybecause divalent Cr is similar to Mgz* and is ac-cepted by the olivine structure. Haggerty et ol(1970) reported optical absorption bands character-istic of Cr2*, but not of CrS-, in olivine from a FETIbasalt, but no comprehensive study of lunar olivineshas been reported. Figure 6 shows the ranges forlunar olivines. The band for FETI olivines cor-

0.4 bs

B

0.2

0.6

0.4

U

nO

Gj

O

d-e<j

=

. 6 o

(J

Fo204060BO

240 IOSEPH

of the non-linear background. Gay, Bown, andMuir (1972) reported curved threads of tiny beadsof spinel from a clast (C) and a picrite basaltxenolith (PB) in breccia 14321 (see Fig. 6 forCr content). The Al contents (Fig. 7) of theolivines (1.12 wt percent in C, 0.38 and 0.19 inPB) are unusually high, and should be muchgreater than any possible experimental error. Sim-kin and Smith (1970) found no Al at the 0.01 wtpercent level in terrestrial olivines, but more recentanalyses by several investigators show O.0n wtpercent. Dodd (1973) reports up to 0.07 wt percentAl2O3 in olivines from the Sharps chondrite. Figure7 summarizes the data for olivines from FETIbasalts and ANT materials. Most of them fall below0.2 wt percent but a few are higher. The syntheticolivine synthesized by Kushiro et aI (1971) at 10kbar and l29O"C has 0.07 wt percent AlzOg whichis near the average for all the olivines from FETIbasalts. Akella and Boyd (1972) found 0.04 and0.43 wt percent AlzOs in olivines synthesized atI 100'C and I bar. and at 1350'C and 30 kbar.Perhaps the substitution of Al increases rapidly withtemperatures over 1300'C. Attention is now focused

FETI boso l lA N T

V. SMITH

on olivines with high analyses of AlzO:r. The brec-ciated anorthosite and pink-spinel troctolite frag-ment from Apollo 16 samples (Prinz et aI, 1973a)are coarse-grained rocks for which there shouldbe no difficulty in obtaining an accurate analysis ofAl. The 14301 microbreccia is fine-grained but theanalyses are very consistent (Prinz et al, 1973c).Only one datum is given for an olivine from thel43l0 basalt. A few scattered analyses of ANTolivines lie above 0.2 wt percent AlzO;r. At thistime, there is no explanation for these very highanalyses of AlzO:;. The analysts have excellent repu-tations but the data should be rechecked in otherlaboratories. Perhaps some problems arise if samplesare polished with alumina instead of preferreddiamond.

(5) Ti analyses are mostly in the range below0.1 wt percent, but it is possible that careful studieswill show interesting features. Unquestionably, care-ful re-analysis of minor elements in olivines is de-sirable by electron microprobe methods togetherwith data on trace elements (especially transitionmetals) by other techniques.

Other interesting features of olivines are: (1) theM1 octahedral site of some lunar olivines is favoredby Fe over Mg atoms, suggesting prolonged an-nealing (see references in Brown and Prewitt,1973); (2) late-stage reduction of fayalite yieldsFe metal * silica mineral * glass (e.s.,El Goresy,Taylor, and Ramdohr, 1972); (3) occurrence ofinclusions yields an estimate of the composition ofthe parent magma when homogenized (e.9., Roedderand Weiblen, 1972\.

AcknowledgmentsSincere thanks go to A. T. Anderson and I. M. Steele for

their creative collaboration on lunar studies, to many otherfaculty members at Chicago for their advice, to I. Baltuska,R. Banovich, O. Draughn, T. Solberg, and others for meet-ing so many technical guidelines. An unpublished glossaryof lunar minerals by J. W. Frondel was very helpful. Tech-nical and financial help from the National Aeronautics andSpace Administration is gratefully acknowledged; indeed, Ihave been amaznd at the general competence of the Apolloprogram, and not surprised at the occasional incompetencewben compared to problems in my own laboratory. Manythanks go to the National Science Foundation for providingbasic support which allowed establishment of necessaryequipment and know-how. I. M. Steele kindly reviewed themanuscript.

ReferencesAcnrrt, S. O., J. E. Acnelr, A. R. Anvoro, eNo J. V. P.

LoNc (1973) Some observations on rock 62295. LunarSci. 4. 15-17 . Lunar Science Inst.

d o.aN

I o.o=

o A

- t 4 3 1 0rU a 1350 'C oa

-Fo 80 60 40 20 Fo

mol . % Fo

Fra. 7. Relation between wt percent Al,O' and mole per-cent Fo for some lunar olivines. A representative selectionof data for FETI basalts and ANT materials is shown: notethat the former data actually should concentrate near Foo-Fo"o, but that there is a tendency for tables of analysis toquote unusual Fe-rich specimens. For the ANT specimens,some data points from the main cluster have been omitted.The unusual data are: C and PB, clast and picritic basaltin 14321 breccia (Gay, Bown, and Muir, 1972); PST and A,pink-spinel troctolite and anorthosite (Prinz et al, 1973a);14301, microbreccia (Prinz et al, l9 '73c);14310, basalt(Gancarz, Albee, and Chodos, 1971). The triangles are forolivines synthesized by Kushiro et al (1971) at l29O"C and10 kbar (Fo"a Al,O" 0.07 wt percent) and by Akella andBoyd (1972) at 1100'C and I bar (Fo*, 0.04) and at1350"C and 30 kbar (Fo* ,0 .43) .

PBoo ̂te.Sc .-.

LUNAR MINERALOGY 24r

J. H. ScooN, J. V. P. LoNc, eNo I. N. Corns (1972)The occurrence of goethite in a microbreccia from theFra Mauro formation. Lunar Sci. 3r 7-9. Lunar ScienceInst.

AxELu, J., rNo F. R. Bovo (1972) Partitioning of Ti andAl between pyroxenes, garnets, oxides and liquid. TheApoIIo 15 Lunar Samples, pp. 14-18. Lunar Science Inst.

ANoenseN, C. A., J. R. HrNrnonnr, ,rxo K. Fnnonrrsson(1970) Ion microprobe analysis of lunar material fromApollo 11. Proc. Apollo II Lunar Sci. Conf. l, 159-167.Pergamon Press.

ANornsoN, A. T. (1973) The texture and mineralogy oflunar peridotite, 15445,10. I. Geol. El,219-226.

T. F. BnsruNAs, J. Jecony, eNo J. V. Sr"rrrn(1972) Thermal and mechanical history of breccias 14306,14063,14270, and 14321. Proc. Third Lunar Sci. Conl. l,819-835. The M.I.T. Press.

BANSAL, B. M., P. W. Gesr, N. J. Husslnp, L. E. Nyqursr,J. M. Rnooss, C. Y. Snrn, lNo H. WTESMANN (1973)Lunar rock types. Lunar Sci. 4, 48-50. Lunar Science Inst.

Berl, P. M., eNo H. K. Meo (1973) An analytical studyof iron in plagioclase from Apollo 15 soils 64501,64502,54802, rock 66095, and Apollo 15 rock 15475. LunarSci. 41 57-59, Lunar Sci. Inst.

Brccln, G. M., M. J. O'H,lnA, A. PEcKErr, eNo D. J.Hurvrpunrps (1971) Lunar lavas and the achondrites:Petrogenesis of protohypersthene basalts in the maria lavalakes. Proc. Second Lunar Sci. Conf. l, 617-643. TheM.I.T. Press.

Bnerr, R., R. C. Goornv, E. Dowrv, M. Pmr.rz, eNo K.KEn (1973) Oxide minerals in lithic fragments fromLuna 20 fines. Geochim. Cosmochim. Acta, 37r 761-773.

BnowN, G. E., lNo C. T. PnBwrrr (1973) High-tempera-ture crystal chemistry of hortonolite. Am. Mineral. 58,577-587.

BnowN, G. M., C. H. EuprBus, J. G. HoueNo, A.PEcKErr, eNo R. Prnrrrrs (1972) Mineral-chemicalvariations in Apollo 14 and Apollo 15 basalts and graniticfractions. Proc. Third Lunar Sci. Conf. l, l4l-157. TheM.I.T. Press.

BurNrNc, D. K., eNo P. R. Buspcx (1973) Fe-Mg latticediffusion in olivine. L Geophys. Res.7t, 6852-6862.

BunNnervr, C. W, (1971) The crystal structure of pyroxfer-roite from Mare Tranquillitatis. Proc. Second Lunar Sci.Conf .1,47-57.The M.LT. Press.

Buscxe, F. D., M. Pnrnz, K. Kerr, lNo G. Kuner (1972)Lunar zirkelite: a uranium-bearing phase, Eerth Planet.Sci. Leu.14.313-321.

BurLEn, P. (1972) Compositional characteristics of olivinesfrom Apollo 12 samples, Geochim. Cosmochim. Acta,36,773:785.

Cueo, E. C. T., eNo 14 Ornpns (1970) Pyroxferroite, a newcalcium-bearing iron silicate from Tranquillity Base. Proc.Apollo 1I Lunar Sci. Conl. 1,65-79. Pergamon Press.

J. A. MrNrnq, lNo J. B. Bpsr (1972) Apollo 14breccias: general characteristics and classification. Proc.Third Lunar Sci. Conf. l, 645-659. The M.I.T. Press.

Cnnrsropse-Mrcnnr-Lnvv, M., C. LBw, R. Cevn, .lNo R.PrBnnot (1973) The magnesian spinel-bearing rocks fromthe Fra Mauro formation. Proc. Third Lunar Sci. Conf. l,887-894. The M.I.T. Press.

Cuwrono, M. L. (1973) Crystallization history of plagio-clase from lunar basalt l2v2l. Lunar Sci. 4, 762-163,Lunar Science Inst.

DENcr, M. R,, J. A. V. Doucres, A. G. PreNr, eNo R. J.Tne'rlr (1971) Mineralogy and petrology of some Apollo12 samples. Proc. Second Lunar Sci. Conf, 1r 285-299,The M.I.T. Press.

rNo A. G. PuNr (1972) Analysis of Fra Maurosamples and the origin of the Imbrium Basin. Proc.Third Lunar Sci. Conf. l, 379-399. The M.LT. Press.

Dooo, R. T. (1973) Minor element abundances in olivinesof the Sharps (H-3) chondrite. Contrib. Mineral. Petrol.42. 159-167.

DREVER, I. L, R. W. Frrzcenero, S. S. LreNc, enn G.Annnerrus (1970) Phyllosilicates in Apollo l1 samples.Proc. Apollo lI Lunar Sci. Conf. lr 34l-345. PergamonPress.

Er- Gonrsv, A., L. A. Tnvron, lxo P. Rlunoun (1972)Fra Mauro crystalline rocks: mineralogy, geochemistryand subsolidus reduction of the opaque minerals. Proc.Third Lunor Sci. Con!. l, 333-349. The M.I.T. Press.

P. R.lr"roonn, M. Plvr6evr6, O. MnoeNaecx, O.Miirren, eNo W. GeNrnEn (1973) Zinc, lead, chlorineand FeOOH-bearing assemblages in the Apollo 16 sam-ple 66095: origin by impact of a comet or a carbonaceouschondrite. Earth Planet. Sci. Lett. [E,411419.

GlNlplrnv, R., R. R. Kuvs, J. C. Llur, lNl E. ANopns(1970) Trace elements in Apollo ll lunar rocks: impli-cations for meteorite influx and origin of Moon. Proc.Apollo 11 Lunar Sci. Conf. 2, lllT-1142. PergamonPress.

GlNcrnz, A. J., A. L. Arnm, rNo A. A. Cttooos (1971)Petrologic and mineralogic investigation of some crystal-line rocks returned by the Apollo 14 mission. EarthPlanet. Sci. Lett. 12. l-18.

Gev, P., G. M. BeNcnoFT, AND M. G. Bown (1970) Dif-fraction and Mtissbauer studies of minerals from lunarsoils and rocks. Proc. Apollo 11 Lunar Sci. Conf. l,3 5 l-362. Pergamon Press.

M. G. BowN, rNo I. D. Mun (1972) Mineralogicaland petrographic features of two Apollo 14 rocks. Proc.Third Lunar Sci. Conf. l' 351-362. The M.LT. Press.

GnsnN, D. H., A. E. RINcwoon, N. G. Wenp, W. O.HrssBRsol.r, A. Ml.ron, rNo E. Krss (1971) Experimentalpetrology and petrogenesis of Apollo 12 basalts. Proc.Second Lunar Sci. Conf . l, 197-206. The M.I.T. Press.

A. E. RtNcwooo, N. G. Wlnr, eno W. O. HBsEn-soN (1972) Experimental petrology and petrogenesis ofApollo 14 basalts. Proc. Third Lunar Sci. Conl. l,197-206. The M.I.T. Press.

HlrNsn, S. S., H. H. Nrerunn, eNo S. Zrrne. (1973) Ferriciron in plagioclase crystals from anorthosite 15415. LunarSci. 4. 325-128, Lunar Science Inst.

Hec,cpnrv, S. E. (1972a) An enstatite chondrite fromHadley Rille. Tfte Apollo 15 Lunar Samples, pp. 85-87.Lunar Science lnst.

(1972b) The mineral chemistry of some decom-position and reaction assemblages associated with Cr-Zr,Ca-Zr and Fe-Mg-Zr titanates. The Apollo 15 LunarSamples, pp.88-91. Lunar Science Inst.

F. R. Bovo, P. M. BEn, L. W. FtNcen, eNn W. B.

242 ]OSEPH V. SMIT'H

BnveN (1970) Opaque minerals and olivine in lavas andbreccias from Mare Tranquillitatis. Proc. Apollo 11Lunar Sci. Conf.lr 5l3-538. Pergamon Press.

Hurreno, N. J., P. W. Gmr, C. Mnven, L. E. Nyeursr,exo C. Sttm-(1971a) Chemical composition of lunaranorthosites and their parent liquids. Earth Planet. Sci.Lett. 13, 7 7--7 5.

C. Mrven, P. W. Gesr, eNo H. WsrsvreNN (1971b)Tho composition and derivation of Apollo 12 soils. EarthPlanet. Sci. Lett. lO,34l-350.

InvrNc, A. I., [. M. Sreerr, eNo J. V. Surrq (1974)Evidence for a noritic igneous compo,nent in Apollo 17fines. Preprint.

hr'rus, O. (1972) Lunar anorthosite 15415: texture, min-eralogy, and metamorphic history. Science, 175, 432436.

feower, l. (1973) Some rare minerals in lunar soils. LunarSci. 41 412414. Lunar Science Inst.

Krn, K., G. Kuur, M. PnrNz, ̂ lNo J. A. GREEN (1972)Lithic fragments, glasses and chondrules from Luna 16frnes. Earth Planet. Sci. Lett.13r243-256.

Kr,rrNrvreNN, B., rNo P. Reuoosn (1971) Alpha-corundumfrom the lunar dust. Earth Planet. Sci. Lett.13,19-22.

Kuoo, A. M., eNo D. F. Wsrlr (1970) An igneous plagio-clase thermometer, Contrib. Mineral. Peftol.25, 52-65.

KusHrRo, l. (1972) Petrology of some Apollo 15 marebasalts. The Apollo 15 Lunar Samples, pp. 128-130.Lunar Science fnst.

Y. Nerauuu, K. Krreyer"rA, AND S. Anuoro(1971) Petrology of some Apollo 12 crystalline rocks.Proc. Second Lunar Sci. Conf.1,481-495.

Lenv, J. S., R. M. FrsHrn, J. M. CnnrsnB, D. T. Gnrccs,A. H. Hsunn, G. L. Nono, euo S. V. Re,oclrrrp (1972)Electron petrography of Apollo 14 and 15 rocks. Proc.Third Lunar Sci. Conf . l, 401422. The M.I.T. Press.

LovrnrNc, J. F., eNo 14 Ornpns (1971) Tranquillityite: Anew silicate mineral from Apollo 11 and Apollo 12basaltic rocks. Proc. Second Lunar Sci. Conf. lr 39-45.

MAsott, B., W. G. MErsoN, E. P. Hr,NprnsoN, E. Jlno-sEwrcH, eNo J. Nslrlr (1971) Mineralogy and petro-graphy of some Apollo 12 samples. Smithson. Contrib.Earth Sci.5, 14.

MAssoN, C. R., I. B. Surrn, W. D. JlrrarcsoN, J. L. Mc-Lecnr,eN, AND A. Vorsonrrr (1972) Chromatographicand mineralogical study of Apollo 14 fines. Proc. Third.Lunar Sci. Conl. l , 1029-1036. The M.I.T. Press.

Mroents, L. G. (1969) Partitioning of Fe** and Mg** be-tween coexisting synthetic olivine and orthopyroxene.Am. L Sci. 267.945-968.

MonclN, J. W., J. C. Leur, U. KniHsNniinr, R. GeNe-pATHy, AND E. ANoBns (1972) Major impacts on theMoon: Characterization from trace elements in Apollo 12and 14 samples. Proc. Third Lunar Sci. Conf.2,1377-1395. The M.I.T. Press.

NuNrs, P. D., lNo M. Tlrsurraoro (1973't Excess lead in"Rusty Rock" 66095 and implications for an early lunardifferentiation. Science, lE2. 916-92A.

PECKETT, A., R. Psnllps, AND G. M. BnowN (1972)New zirconium-rich minerals from Apollo 14 and 15lunar rocks. Nat ure, 236, 215-217.

PnrNz, M., E. Dowrv, K. Kerr-, eNo T. E. Buxcn (1973a)

Spinel troctolite and anorthosite in Apollo 16 samples.Science, 179,74-7 6.

AND _ (1973b) Mineralogy,p,etrology and chemistry of lithic fragments from Luna 20fines: origin of the cumulate ANT suite and its relation-ship to high-alumina and mare basalts. Geochim. Cosmo-chim. Acta, 37, 979-1006.

C. E. Nnnnu, G. Kuur, K. Kerr, rNo G. M.CoNneo (1973c) Electron microprobe analyses of lithicfragments, glasses, chondrules, and minerals in Apollo 14lunar samples. Uniu. New Mexico Inst. Meteoritics, Spec.Publ.6.

Roaooen, E., lNo P. W. WsrsrEu (1970) Lunar petrologyof silicate melt inclusions, Apollo 11 rocks. Proc. Apollol1 Lunar Sci.Conf. l ,801-837.

AND - (1972) petrographic features andpetrologic significance of melt inclusions in Apollo 14and 15 rocks. Proc. Third Lunar Sci. Conf. l, 251-279.T\e M.I.T. Press.

Ronorn, P. L., lxn R. F. Euslrs (1970) Olivine-liquidequilibrium. C M P, 29, 27 5-289.

ScniinrvreNN, K., lNo S. S. HlnNBn (1972) On the amountof ferric iron in plagioclases from lunar igneous rocks.Proc. Third Lunar Sci. Conf. l, 615-621. The M.I.T.Press.

SrurrN, T., eNo J. V. Srvrlrrr (1970) Minor-element dis-tribution in olivine. I. GeoI. 7E, 304-325.

Srulrrr, J. V. (1971) Minor elements in Apollo 11 andApollo 12 olivine and plagioclases. Proc. Second LunarSci. Conl. l, 143-150. The M.I.T. Press.

A. T. AlroEnsor.r, R. C. NnwloN, E. J. OrsrN,eNo P. J. Wvrrrn (1970) A petrologic model for theMoon based on petrogenesis, experimental petrology, andphysical properties. l. Geol.7E, 381-405.

lNo I. M. Srrrrr (1973) How the Apollo programchanged the geology of the Moon. Sci. Public Afr. ll-15.

SrsErs, I. M., euo J. V. Srvrlnr (1971) Mineralogy ofApollo 15415 "Genesis rock": source of anorthosite onMoon. N at ure, 234, 738-140.

AND - (1972) Ultrabasic lunar samples. Nar.Phys. !ci.240,5-4.

AND - (1973) Mineralogy and petrology ofsome Apollo 16 rocks and fines: general petrologic modelof Moon. Preprint lor Proc. Fourth Lunar Sci. Conf.

AND - (1974) lntergrowths in lunar anortho-sites with implications for lunar differentiation. Preprint.

Tevron, L. A., H. K. Mlo, eNo P. M. Brur (1973)Apollo 16 "rusty" rock 66095. Lunar Sci. 4' 715-716.Lunar Science Inst.

lNp K. L. Wrlrrlr'rs (1973) Cu-Fe-S phases inlunar rocks. Am. Mineral. 5E,952-954.

Tnllrr, R. J., A. G. PreNr, lNo I. A. V. Doucus (1970)Garnet: first occurrence in the lunar rocks. Science,169'98 1-982.

WALKER, D., J. Lorqcnt, T. L. Gnovr, E. Srorrnn, eNoJ. F. Hlvs (1973) Experimental petrology and origin ofrocks from the Descartes Highlands. Preprint lor Proc,Fourth Lunar Sci. Conf.

Wenr. D. A.. A. F. Rrro. J. F. Lovrnrxc. eNo A. Er

LUNAR MINERALOGY 243

GonEsv (1973) Zirconoli:te (uersus zirkelite) in lunarrocks. Lunar Sci. 4, I 64-7 66.

WARNER, l. L. (1972) Metamorphism of Apollo 14 breccias.Proc. Third Lurnr Sci. Q:onf. t, 623-643. The M.LT.Press.

W.rnr, D. F., eNo M. J. Duxr, (1973) Europium anomaly:a model for prediction. Lunar Sci. 4, 778-780. LunarScience Inst.

I. S. McCellur'r, Y. BorrNcl, M. J. DRAKE, ANDG. A. McKev (1970) Mineralogy and petrology of someApollo 1l igneous rocks. Proc. Apollo 11 Lunar Sci,Conf ,1 ,937-955.

Wnnr, H.-R., eNo W. R. Wrlnr (1973) Chemical anomaliesof lunar plagiociase, described by substitutiol vectorsand their relation to optical and structuial properties.cMP,41,89-104.