Chronology and complexity of early lunar crust

Tectonophysics.

157

161 (1989) 157-164

Elsevier Science Publishers

B.V., Amsterdam

- Printed

in The Netherlands

Chronology and complexity of early lunar crust * E.J. DASCH ’ NRC/NASA

I, G. RYDER

2 and L.E. NYQUIST

3

JSC, Houston, TX 77058, and Oregon State University, Corvallis, OR 97331 (U.S.A.)

’ Lunar and Planetary Institute, 3303 NASA Road One, Houston, TX 77058 (U.S.A.) ’ SN4/NASA

(Received

JSC, Houston. TX 77058 (U.S.A.)

January

2,1988;

accepted

June 24, 1988)

Abstract Dasch,

E.J., Ryder,

(Editor),

G. and Nyquist,

Growth

The least equivocal was formed

It was formed

Ga-and magmas

in a large magma

cannot

rare-earth

pattern.

mare basalt

crystallized

by about

This residue

of lunar

fashion,

depths

materials

in the mantle

crust.

that: (1) the lunar

hundred

kilometers

in depth.

(2) at least some members

In: L.D. Ashwal

reworked

Ga, whether

from anorthositic trace-element from

the previous

in melting

episodes)

demonstrate

that lunar

crustal

to at least several hundreds

However,

of

their parental

with a remarkably

processes

in immediate

Ga ago). The sources

not

Mg-suites

(3) a trace-element-rich

and impact

impact or volcanic

crust earlier.

with rare earths

of the diverse

crust formation. patterns;

4.33 Ga) than was once assumed,

(3.9-4.0

anorthostic

a very few 100s of Ma after 4.56

origin;

(4) the onset of ferrous

for this mare volcanism formation

of several

happened

material constant

such that most samples

and was in process

and were also in place by 4.4 or 4.3 Ga ago at a depth

These characteristics and affected

incompatible

began much earlier (about

to KREEP

suggest

very soon after lunar accretion-within

was subsequently

of the most intense period of bombardment in characteristics

at least several abundances;

4.3 Ga ago (as residues 3.9-4.0

of early lunar

of the Moon at 4.56 Ga; it may have been completed

chronologically

from chondritic

it have ages around volcanism

accretion

probably

be distinguished

evolved

was formed

which contain

dunites)

presently

(KREEP)

data on lunar rock samples

from chondritic-relative

troctohtes,

had clearly

system

and complexity

Tecronophysics, 161: 157-164.

120 Ma after the primary

fractionated

rocks (norites,

Crust.

age and petrographic

by about

appreciably

L.E., 1989. Chronology

of the Continental

before the end

are complementary hundred

quickly.

kilometers. in a complex

of kilometers.

introduction

produced a plagioclase-bearing crust including genuine anorthosites; it was probably over in about

In its earliest history the Moon had a global or near global, complex magma system. The evidence

200

is in the form of preserved, “pristine” igneous rocks from the massive differentiation of the Moon, as well as isotopic data from subsequent volcanic samples demonstrating early differentiated deep sources. The system may have been a magma ocean or a magmasphere; even in the former case it was superposed by smaller but also very complex magma systems. The main episode

probably some kind of feldspathic norite. The processes of crust building, reviewed by Warren (1985), remain in serious dispute.

* This paper is LPI Contribution 0040-1951/89/$03.50

No. 696.

0 1989 Elsevier Science Publishers

B.V.

Ma,

tinued.

although

The average

diminished composition

magmatism

con-

of the crust

is

Although the time of formation of these earlier lunar rocks is critical to an understanding of their petrogenesis, few well documented samples of pristine plutonic lunar rocks (or PPLRs) have been dated unequivocally (Nyquist, 1981). Isotopic information on several of these rocks and their minerals do not form systematic arrays on isochron or other graphs. Commonly, the more aberrant data are excluded from the final re-

158

gressions criteria,

or other though

calculations,

perhaps

rock, are not standardized common

problem

internal

(mineral)

departure

but the selection

reasonable

for a given

or systematic.

for many

Rb/Sr

The most

and

stitute

Grieve, norites,

a significant

though

1980). Mafic and gabbros)

poorly

quantified

Sm/Nd

whatever

from the best-fit

magma

anorthosites,

system

nor

produced

in most

the

cases

rather

niques

have, in some cases, resulted in markedly different “ages” for the same rock or even the

of these crustal

rocks suggests

same sample.

low, bulk lunar Ca/Al.

The trace elements

differentiated

or

In addition,

the several

dating

tech-

What is the lunar crust?

processes

to Apollo,

the nature

of the lunar

crust

sources

(Norman

later) have negative

4.3 to about

trace-element

Plagioclase

is rarely

Moon

trace element

10% of its volume,

overlying

composing

an ultramafic

about mantle

and other non-

relative

near

their

characteristics

to be necessarily

assimilation 3.0 Ga or even

Eu anomalies

chondritic

crust

a very require

and Ryder, 1980). Mare basalts

from perhaps

was anyone’s guess, and apparently no one surmized correctly. Geophysical data show that the has a low density

melting,

or even

complex

other. in most

large-scale

sources,

to

ferroan

than clinopyroxene

previously-differentiated

(erupted Prior

por-

to each

Orthopyroxene

isochrons.

rocks con-

tion of the crust. They are not easily relatable

or whole rock ages is significant

of one or several points

more than 50 km; (dunites, troctolites,

abundances.

liquidii,

are believed

inherited

from

the sources,

not

from complex

km thick on the nearside;

These sources are differentiated, having lost plagioclase and in some cases gained ilmenite.

the offset of the lunar

center of mass from its center of the Earth) is generally interpreted thicker (more than 75 km?) crust Apollo samples demonstrate that

figure (towards to result from a on the farside. the low density

phase is plagioclase; for pure anorthite a little over 20 km worth would be in the outer regions of the nearside (Wood, 1986). The lunar crust is not anorthosite,

although

at least

its upper

part

is

anorthositic (26-28% Al,O,); at greater depths it is probably more noritic or troctolitic (Spudis and Davis, 1986).

It is most easily explained fractional

crystallization,

scale or by a commonly reworked and

large proportion The lunar

crust

is very rich in plagioclase.

It

partially melted zone. Much of the plagioclase occurs in the rock type ferroan anorthosite; mafic complements to this rock type have not been found, so the differentiation (and probably highdegree melting) reached far below the depth of penetration of even large basins (i.e. certainly

sources

km (Delano,

1986).

complex; nonetheless were ultimately felt

through at least the outer half of the Moon. The trace element-rich KREEP is a common chemical component of highland polymict rocks. Wherever found, it has a strangely uniform trace element abundance pattern (Warren and Wasson, 1979b).

volcanism

requires complete extraction of plagioclase components from about 30% of the lunar volume, or a depth of 200 km, assuming a Moon similar to the silicate portion of chondrites, which it resembles in geophysical properties. More plagioclase in the crust, or less efficient extraction, requires a thicker

of 200-400

Their origin was certainly the effects of differentiation

later Evidence for massive differentiation

Mare basalt

the

(the Earth’s crust is only 1% of its volume). Apollo seismic data indicate that this crust is about 55-65

were at depths

later processes.

and

by most

by

as the end-product probably repeated

large-degree

by impact

on process. partial

redistribution.

of the bulk

lunar

of

a global It was melting A very

incompatible

element budget is now in the crust, much of it as some form of KREEP. Petrology and chronology of crustal materials The intense cratering of the highlands up to 3.8 Ga ensures that almost all of the samples collected are polymict, commonly melted, breccias. Only rare samples can be recognized as pristine; that is having preserved an igneous chemistry, and lacking meteoritic siderophile contamination, or admixtures of polymict materials. Few of these have retained vestiges of an igneous texture and the

159

identification is

of a particular sample as “pristine”

not always cut-and-dried.

AI-Ar

rl

It is not always clear

whether or not a rock is pristine. Additionally,

N

72255

our limited sampling of pristine rocks cannot be considered

representative

I

of major crustal rock

---

I

types. Our limited sampling of pristine rocks apparently is not representative of the entire popula-

I

;

I

A

collection,

A

mainly as clasts in breccias,

continues. So do the deconvolutional

T

N

t

A

77115

N-A 87075

15415

attempts to

73233

’ 772’5

153e2

more (small) igneous samples hidden away in the Apollo

N

1

; 15455 ---

i

tion of the pristine lunar crust. The search for

1 N

A

22013

T

30025

A?

75535

I

A?

87435

37435

I

Rb-Sr

understand the chemistry of the polymict rocks. Highland

(crustal)

igneous rocks are almost

N

entirely plutonic, perhaps because volcanic rocks,

7823e

being surficial units, are the most susceptible to impact comminution. They form three distinct(?)

J-g&

,

suites: Ferroan anorthosites, Mg-suite, and KREEP. At least the Mg-suite itself cannot be from a single magma system, but is a polyglot. Unfortunately,

Sm-Nd

radiogenic isotopic data for many

of the Mg-suite samples are lacking and some of that is confusing. Nyquist

(1981)

compiled

a histogram

of the 1

available age data on PPLRs (updated version in Fig. 1). (The system for recognizing PPLRs, vised by Warren and Wasson (1979a),

I

de-

I

and Nor-

:t

I

3.9

man and Ryder (1980) among other workers, is based on low siderophile abundances, cumulate mineralogy and textures, and, in some cases, on “ primitive” chemical characteristics such as un-

AGE (AE)

Fig. 1. Ages of Pristine

Plutonic

Lunar

Rocks.

evolved trace element patterns, that is, no KREEP

to 3.9 Ga. There is no apparent explanation

component. This select group of rocks constitutes

why Rb/Sr

as to

ages should be greater than Sm/Nd

a very reduced population of lunar samples. These

ages, in some cases on the same sample (Nyquist,

ages, all > 3.9 Ga (26 numbers) were determined

1981)

by 39Ar/40Ar, Sm/Nd, and Rb/Sr techniques. Of

tively young Sm/Nd

special interest are the oldest apparent

ostensibly

crystallization.

ages of

Of the three techniques, the oldest

although Carlson et al. (1988) report relaages on eucrite cumulates,

a result of low blocking temperatures

for this system.

ages as well as the oldest average ages were Rb/Sr ages. Three of the six Rb/Sr ages plot near 4.5 Ga. The seven Sm/Nd ages range from 4.2 to 4.5

Ferroan anorthosites

Ga, averaging 4.3 Ga. 39Ar/40Ar ages are distinctly younger, with a peak at 3.9 Ga, but with three ages of 4.4 Ga, and some evidence of older events in the higher temperature releases of some of the more complex spectra. The 39Ar/40Ar ages may be explained by excavation-cooling during basin-creating events such as the Imbrium or by reheating during the high impact flux. continuing

Ferroan anorthosites probably formed from magmas with roughly chondritic trace element patterns (Warren and Wasson, 1979a) very early in lunar history. They are vestigiously coarsegrained; mafic minerals (rarely more than a few percent) are homogeneous within a sample, except for rare, mixed samples. Ferroan anorthosites are among the most slowly-cooled crustal rocks known

160

in the solar system.

at least according

Ca in their olivines

(Ryder,

fractionated

sequence

1984). They

(Mg’

70-40)

form a

but

trapped

abundances al.. 1981;

of incompatible elements (Haskins Ryder, 1982). The lack of mafic

ultramafic

rocks (or even mineral to the

that they formed the depths basins They and

containing

have

negligible

plimentary

liquid,

to the low

ferroan

Davis,

system

et to

com-

suggests larger than

km; Grieve,

to be of global

1986).

low

of even the largest lunar

(on the order of 50-80 also appear

fragments)

anorthosites

in a magma

of excavation

very

even should

extent

1980). (Spudis

making

allowance

for

not

attributed

to

plagioclase

that

anorthosite,

i.e. the plagioclase

be

in norites,

trocto-

the ferroan anorthosites. yet are more magnesian for a given plagioclase composition. A wide range of textures cooling

and mineral

chemistry

environments;

e.g.,

suggests

dunite

slightly zoned olivine crystals

which contain

calcium

absolutely

contents

neous

olivines

1975; Ryder, indicating and

than

in troctolite

magmas,

of ferroan

Ryder,

homoge-

76535 (Dymek

an origin from evolved,

et al.,

1980).

Most

of ilmenite

Ti/REE

and Ca/Al.

unlike

the

Some and augite.

The ages of the Mg-suite

mixed

chondritic

(Norman

have low Ca/Al

thopyroxene-dominant). amounts

ratios

probably

anorthosites

has

higher

1984). Most have low Ti/REE

assimilated

Ti/REE

the

varied

72415

and

ratios

have

(or-

significant

and hence higher

samples

appear

to be

lites, and basalts. The very low measured a7Sr/a6Sr ratios of selected anorthosites (e.g. Papanastassiou and Wasserburg, 1969) lead most workers to believe that they are very early lunar differentiates, per-

varied, but also confusing. These rocks are recognized as important constituents of the early crust, but their ages and chemical characteristics have been clouded by one or more kinds of natural or

haps from a magma

ocean (even though 39Ar/40Ar

analytical

dates are apparently (1987) has reported age of 4.44 + 0.02

younger). Recently, Lugmair a Sm/Nd mineral isochron Ga for ferroan anorthosite

60025. The date is important not only for its precision but also for its significantly “young” age, relative to the age of the Moon. The age implies that, if the global magma ocean hypothesis is valid, the early anorthositic crust did not fully

data

open system behavior.

selected

by

crystallization although

ages

most

uncertainties,

Nyquist range

of these

Mineral

(1981) from dates

several appear

isochron

indicate

that

4.1 to 4.5 Ga; rather

large

to be reasonably

have

pre-

cise, and, of these, a few are the same as the accretion age of the Moon. These problems, along with new data for Apollo 15 norites, have been recently discussed by Dasch et al. (1987a).

crystallize, or reach blocking temperatures for Sm/Nd, until 120 Ma after the accretion of the

KREEP

Moon. Adding to the confusion are more recently determined Pb-Pb ages for the same rock, 60025 (Hannon and Tilton, 1987). These workers obtain a result distinctly older, relative to the stated uncertainties, than the Sm/Nd analytical

KREEP is rare as igneous rock; its presence, however, is moonwide, varying is concentration, and dominates the trace elements and radiogenic isotopes of polymict breccias. The most KREEP-

result-4.52 + 0.007 Ga (model Pb-Pb 4.50 f 0.007 Ga (concordia intercept Sm/Nd result is in possible conflict parently younger

older ages for possibly derivative or rocks (eg. the Sm/Nd age is 4.52 f 0.10

Ga for anorthositic al., 1979); note barely overlap). Mg-suite

age) and age). The with ap-

norite that

the

15455,228 uncertainty

(Nyquist

et

envelopes

rocks

These rocks (norites, troctolites, and dunites) all have trace element patterns more evolved than

rich samples (La > 150 X chondrites) basaltic in major element composition, around

the

pyroxene-plagioclase

tend to be clustering

cotectic

even

when they are breccias. KREEP is not uniformly distributed around the Moon; samples from the Imbrium

Basin

region,

especially

Apollo

12 and

Apollo 14, are the most KREEP-rich. The Apollo 16 samples have less KREEP; the lunar meteorites, of unknown location, have virtually no KREEP (e.g., Warren et al., 1983). The petrogenesis of these enigmatic rocks is imperfectly known, but is multistage. KREEP samples have a paradoxical

161

evolved

incompatible

combined

with

suggestive

of

processes.

trace

a fairly

element

primitive

complex

mixing

Most of the isochron

abundance Mg

of

number,

assimilative

ages cluster around

(4) A plot of initial Sr isotopic composition vs. eNd results in a value. within experimental error. of 0.05 (Nyquist (5) A T-I 14 mare basalts

3.9 Ga, with model ages near 4.3 Ga.

et al., 1981).

plot for a selected

group

of Apollo

also suggests a ratio very near 0.05

(0.06: Dasch et al., 1987b). (6) Isotopic

Mare basalts

is consistent

Not commonly

included

rocks, mare basalts about

tense bombardment Apollo

extruded

between

the end of widespread,

basalt

fragments.

however,

has

uncovered

a series of seven clasts with crystalliza-

tion

from

ages

1987b). during

3.96 to 4.33 Ga

The existence this period

our understanding

of mare

(Dasch

basalt

thus has somewhat of early crustal

et al..

volcanism complicated

evolution

of the

Moon.

Analysis of earliest lunar ages and petrogenesis

Many of the pristine ages with analytical

plutonic

be the intense phism

comminution

of meteoritic

A critical parameter in understanding the evolutionary history of lunar rocks is the “Rb/ *%r ratio of the undifferentiated or whole Moon.

exotic components.

Ratios for derivative

Kellog.

and mixing

melting,

fractional

or contamination

cryscan, in

principle, be evaluated. The NASA/JSC laboratory has used a value of 0.05 for this ratio for the past 15 years. There are, at present, at least six lines of evidence supporting this value; two are the result of recent

studies:

(1) In a plot of Sm/Eu ratios vs. Rb/Sr ratios of carefully selected rocks, Nyquist et al. (1973) derived the value of about 0.05 for the 87Rb/R6 Sr ratios from the intercept of the Rb/Sr curve with the curve for chondrites, assuming no Eu anomaly. (2) Time of crystallization

vs. initial

“Sr/a6Sr

ratios (T-Z plots) of Apollo 12 basalts indicate 0.05 as a reasonable number for the bulk Moon (Nyquist et al., 1979, 1981). (3) Modeling of a variety of lunar processes by the JSC group during the period 1978-1987 have used this value without introducing any fundamental problems.

and shock metamor-

bombardment

continued

and the effects of partial

rocks yield

too large for an

that these rocks have undergone

but difficult to quantify Additional contributing

lunar rocks can be compared,

lunar

uncertainties

unequivocal assignation of dates for the formation of the lunar crust. The main problem appears to

mRb/86Sr ratios for lunar rocks and the undifferentiated Moon

tallization,

60025

1987).

in-

about 3.9 Ga. Recent work on

14 mare

anorthosite

with the earliest crustal

were mainly

3 Ga ago and

work on pristine

with this value (Lugmair.

and

as a result

the consequent

redistribution of nuclides. problems are partial or

equilibration

of some

components

in

the subsolidus state owing to elevated temperatures deep in the lunar crust. and inmixing of cannot

completely

paration.

such as other be excluded

The diversity for many

uncertainties

preclude

for the fundamental

sample pre-

of cNd (e.g., Turcotte

1986) also contribute

assignment

rock types. that

during

lunar

and

to an equivocal rocks.

Resulting

an unequivocal problems:

age age

chronology

the validity

of the

magmasphere/ magma ocean hypotheses. the age of the most ancient anorthosites relative to the time or times of intrusion and the delimiting formation.

of the Mg-suite

of the period

of lunar

of rocks. crustal

Our present understanding is as follows. The ferroan anorthosites date near to the origin of the Moon: according to model ages such as BABI they cannot be much younger than 4.56 Ga. Ferroan anorthosite 60025 has an age of 4.44-4.52 Ga; some of the more mafic rocks (“Mg-suite”) apparently have the same or even older ages. Mare basalt sources closed at about 4.35 Ga for Pb, Sm. and Sr systems. and the complementary isotopic and trace element characteristics of KREEP, the most evolved component. were formed by this time. A few mare basalt samples show that at 4.3

162

Ga there was a stable

crust (not just

floating

and such an origin best explains

the chemical

rockbergs) onto which lavas could flow. Thus it is

bimodal character of the ferroan anorthosite

evident that the main part of the crustal formation

suite rocks. The anorthosite parental ocean had a

was complete

roughly chondritic

abundance

cooled below the closure temperatures of the main

tory incompatible

elements

non-gaseous

have floated,

by 4.35 Ga, and the system had radiogenic

isotopic

systems, i.e. to

unlike

most

Mg-

pattern of refrac(they

could indeed

Mg-suite

samples-

depths of the order of a few hundred kilometers.

Warren, 1979; Warren and Wasson, 1979a). The

After 4.35 Ga, partial melting of mantle and crustal

parent

sources continued, but not on a global scale. By

chondritic

3.9 Ga at least portions of the crust had cooled

partial

enough for the Apennine Front (Imbrium ring) to

olivine and some pyroxene.

had

evolved

from

(bulk Moon)

a

volatile-depleted

composition

melting and fractional

mainly by

crystallization

of

A very large scale

remain isostatically uncompensated (Ferrari et al.,

system is required to account for the absence of

1978); nonetheless the crust appears to have been

mafic cocumulates,

simultaneously

the crust, and the non-chondritic

hot enough in the same region to

have partially melted and produced KREEP volcanism (perhaps by a combination of heat input and pressure release from the Imbrium impact). The essential point of the chronology however is that crustal and mantle Sm-Nd

(and other

trace element) evolution departed from chondritic composition early, when major crustal development took place. That departure has been preserved, and its effects influence and are demonstrated by subsequent melting events. The ferroan anorthosites appear to have formed prior to the isotopic departures from chondritic composition, i.e. they appear to have been the earliest rocks.

basalt

the amount of anorthosite

sources. (A magma ocean

in

nature of mare might not be

required if the Eu anomaly of mare basalts can be explained by near-surface

complex processes and

if the upper crust does not have a positive Eu anomaly overall (Walker, 1983). It should be noted that the chemistry of the lunar meteorites suggest that the average, non-mare surface has a positive Eu anomaly.) Production (i.e. initial crystallization) of plagioclase crust from a magma ocean is not itself very simply explained (Morse, 1982). The Mg-suite rocks require a separate set of origins, presumably later than the magma ocean, because their origin requires the presence of evolved materials (KREEP or KREEP-like) at an early stage. They formed from multiple magmatic

Discussion

episodes, but the primitive Mg’s (90) of some of the Mg-suite rocks suggests large-scale melting. It is possible that a massive overturn of a density-un-

Even now the processes which produced the lunar crust, reviewed by Warren (1985), are disputable. Their elucidation depends on gaining a bet-

magma ocean caused extensive melting of uprising

ter understanding of the igneous rocks in the highlands, the unravelling of the chemistry of

rich, Eu-depleted sources for the later mare basalts. The residual liquid from the earliest ocean (Ur-

polymict rocks, and understanding the lateral and vertical variations of rock types in the crust (e.g., Spudis and Davis, 1986). Walker (1983) abandoned the concept of a lunar magma ocean (or magmosphere), concluding that serial magmatism would have been adequate to explain the crustal rock types. Longhi and Ashwal (1985) suggested the mechanical (diapiric) separation of anortho-

KREEP) may have played a considerable role in influencing the chemistry of varied Mg-suite rocks. Whatever the nature of these events, they were pretty much history by 4.3 Ga, by which time KREEP and the sources of mare basalts had cooled below their closure temperatures. Later magmatic activity included possible Mg-suite plutonic magmas, and certainly remelting of KREEP to form

sites from their mafic complements. Nonetheless, the characteristics of ferroan anorthosites are most compatible with production in a magma ocean,

volcanic rocks. Even polymict KREEP rocks have essentially basaltic (cotectic) compositions indicating magmatic rather than solely impact control.

stable mantle following rapid crystallization

of a

Mg-mantle as well as causing sinking of the Fe-

163

Comparison

of earliest lunar and terrestrial events

Dasch, E.J., Nyquist, L.E., Ryder, G., Steele. A.M., Wiesmann, H., Bansal, norites.

With respect

to bulk composition

and earliest crustal formation, about that

the Moon there

ocean

much more is known

than about

could

have

(e.g., Warren,

of the crust

Earth.

been 1985).

It seems likely

an early

of earth’s the data are

very few, about 4.28 Ga-Compston 1986),

still

accretion

is about rather

the first 280-760 larger

obliterated tance

than

younger

than

and

more

existing

suggests, evolution

primitive

crust.

Ma of its existence, active

its

repreDuring

the Earth’s

mantle

evidently

traces of crust, with the assis-

of a high impact

record crustal

and Pidgeon,

age, and these rocks and minerals

sent evolved much

280 Ma

flux. The Moon’s

earliest

however, that the Earth’s was probably petrolo~cally

early com-

plex and rapid. The Moon ferentiation,

preserved

its evidence

in apparent early

with the Earth,

shut-down

of its magma

systems,

or at least major

important

point

complex

things

of early dif-

contrast

of the

mantle

is that

convection.

on the Moon

happened

very

An

a lot of

rapidly,

even

though it is a rather small body (the eucrite parent body may not have such complexity, although diverse eucrite lithic units have recently been reported-Mittlefehldt and Lindstrom, 1988). Does that the early terrestrial

Delano,

Philips Lunar

evolution

Are there history?

terrestrial Clearly

was

relics of this

the Earth’s

could not have been very similar

the Moon,

because

have contributed

to that of

not so much plagioclase to a terrestrial

early could

crust.

and

G.J.

tive petrology origin:

anorthosite

Planet.

of the eucrites Sci. Conf.,

N.Z.,

Nuevo

1988. Radiometric

72415,

Origin

Houston,

Grieve,

surroundings.

R.A.F.,

W.L.

J. Geophys.

Lunar

and

record:

Planetary

primary

78235,

and

and Phillips, Apennines

R.J.,

and

re-

Res., 83: 2863-2871. processes

In: 0. James and F. Hoertz (Editors), Crust.

norite

Sci. Conf. 6: 301-341.

of the lunar

1980. Cratering

Tex., pp. 231-247.

A.A., 1975. Compara-

76535,

Sjo8ren,

state

R.J.

of the Moon.

rocks of possible

troctotite

D.L.,

1978. The isostatic gional

nickel, and volatiles

and effects.

The Lunar

Institute,

Highlands

Houston,

Tex., pp.

Hannon,

B.B. and Tilton, G.R., 1987. 60025: relict of primitive

lunar crust? Haskins,

Earth

Planet.

L.A., Lindstrom,

D.S.,

1985.

On

anorthosites. Longhi,

compositiona

Proc. Lunar

J. and Ashwal,

anorthosites: sis. Lunar Lugmair,

Planet.

G.W.,

P.A. and Lindstrom,

variations

Pianet.

to the magma

Sci. Conf., Planet.

Sci. Conf.,

D.W. and Lindstrom,

of diverse

lithologies

Sci. Conf.,

19: 790-791. growth

of the lunar crust. J. Geophys. .%A., 1986. Origin

compositional

model

for lunar

ocean

hypothe-

eucrites.

Lunar

of anorthosite

Planet.

at the base

Res., 87: AlO.

of earliest

convection.

60026-Methuse-

18: 584-585.

M.M., 1988. Geochemistry

in Antarctic

S.. 1982. Adcumulus

lunar

15: 491-492.

1987. Age of the lunar crust: Lunar

among

Sci., 12B: 41-66.

L., 1985. A two-stage

an alternative

lah’s legacy. Mittlefehldt.

Sci. Lett.. 84: 15-21.

M.M., Salpas.

planetary

Earth

Planet.

crust: Sci.

role of

Lett.,

81:

constraints

on

118-126. Norman,

M. and Ryder,

G., 1980. Geochemical

the igneous evolution

of the lunar crust. Proc. Lunar

Planet.

11: 317-331.

L.E., 1981. Radiometric piutonic

lunar

ages and isotope rocks.

systematics

In: D. Walker

and

I.S.

Laredo

and Bareba.

Australia.

of

Nature,

321: 766769. of zircon

IS. and Meyer, from

Sci. Conf.,

late-state

15: 182-183.

C., 1984. Age and lunar

differentiates.

Planetary

Institute

Workshop,

Houston,

L.E.,

defined

Hubbard,

N.J.,

H.,

Rb-Sr

Apollo

1973.

Gast.

P., Bansal,

systematics

15 and 16 materials.

B.M.

and

for chemically

Proc. Lunar

Sci. Conf.,

J., Bansal,

B.M., and

4: 1823-1846. Nyquist,

W., Williams, Planet.

Nyquist,

Wiesmann,

zircons in Western

Lunar

pp. 71-77.

19: 166-167.

W. and Widgeon, R.T., 1986. Jack Hills, evidence

more very old detrital

Lunar

Cosmo-

McCallum (Editors), Magmatic Processes of Early Plane-

geochronology

chemistry

Institute,

A.J., Nelson,

Ferrari,

tary Crusts.

Compston,

(Editors),

62237. Proc. Lunar

of pristine

R.W., Tera, F. and Boctor,

Compston,

Taylor

of lunar cumulate

dunite

Sci. Conf.,

References

Lunar

frag-

and petrogen-

Geochim.

of cobalt,

R.F., Albee, A.L. and Chodos,

Nyquist,

Carlson,

14321; chronology volcanism.

H. and

of basaltic

of the Moon. In: W.K. Hartman,

and Planetary

Dymek,

Morse,

history

Wiesmann,

B.M..

provenance

J.W., 1986. Abundances

in the silicate portion

system

again?

Bansal,

mare

IS

18: 219-220.

chim. Acta, 51: 3255-3272.

Morse,

complex

C.-Y.,

L.E., 1987b. Isotopic

ments from lunar breccia

also very complex, but that continued activity and remixing has ironed things out into a simpler early

Shih,

1987a. Age of ApolIo

Sci. Conf.,

27-29.

because

this mean

E.J.,

Nyquist,

Planet.

esis of pre-Imbrium

terrestrial

Evidence

oldest crust, above 3.8 Ga (or, though

Dasch,

B.M. and Shi, C-Y.,

Lunar

L.E., Wiesmann,

H.. Wooden,

Shih, C.-Y., 1977. Age and REE abundances norite

from 15455. In: The Lunar

and Planetary

Institute,

Houston,

Highlands

of anorthositic Crust.

Tex., pp. 122-124.

Lunar

164 Nyquist.

L.E.. Shih. C.-Y.,

Wiesmann. Apollo

12 basalts:

evolution. Nyquist,

B.M.,

basalt

Papanastassiou,

D.A.

time

differences

Earth

Planet.

Ryder,

Isotopic

and

Wasserburg.

Lunar Walker,

H. and of lunar

of aluminus

in the

G.J..

formation

of planetary

Initial of small objects.

Sci. Lett.. 5: 361-376. anorthosite

60025, the petrogenesis

Acta, 46: 1591-1601.

1984.

Olivine cumulate.

in lunar Lunar

of

of the moon. Geo72415.

a rather

Planet.

Sci. Conf..

15:

709-710. Spudis.

P. and

Planet.

Institute,

Sci.

In: W.K. Hartmann.

(Editors),

Origin

Houston,

and terrestrial Conf..

R.J.

of the Moon.

Tex., 311-329.

crust formation.

14.-J.

Geophys.

Proc.

Res.,

88:

B17-B26. Warren.

P.H.,

Warren,

1985. Annu.

The

P.H. and Wasson,

the crystallization implications Warren,

magma

Rev. Earth

ocean

Planet.

concept

J.T.. 1979a. Effects

of a “chondritic”

Sci. Conf.,

P.H. and Wasson,

Rev. Geophys.

lunar

of pressure

magma

for the bulk composition

Planet.

and

Sci.. 13: 201-240. ocean

of the Moon.

on and

Prcc.

10th: 2051-2083. J.T.. 1979b. The origin of KREEP.

Space Phys., 17: 73-88.

P.H.. Taylor,

G.J. and Keil, K.. 1983. Regolith

Allan Hills A81005: phy of pristine

evidence

of lunar origin,

and non-pristine

breccia

and petrogra-

clasts. Geophys.

Res. Lett.,

10: 779-783. Davis,

P., 1986. A chemical

model of the lunar crust. Proc. Lunar

and petrological

Planet. Sci. Conf.. 17:

818-819. Turcotte.

and Planetary

Lunar

Warren.

dunite

G.J.

D., 1983. Lunar

Lunar

chim. Cosmochim.

and

evolution.

1969.

and the resolution

and the composition

shallow-origin

Taylor

of

geochemical

studies

for petrogenesis

abundances

G., 1982. Lunar

G.,

of the moon.

Phillips

Wiesmann.

REE

lunar anorthosites, Ryder.

data for the origin

record

Planet Sci. Lett., 55: 335-355. and

isotopic

lunar

B.M. and

10: 77-114.

J.L., Shih, C.-Y.,

implications Earth

Bansal,

Nd isotopic for

Sci. Conf..

1981.

12038:

mare basalts. strontium

J.L.,

implications

Proc. Lunar

L.E., Wooden,

Bansal,

Wooden,

H. 1979. The Sr and

D.L. and Kellog, L.H.. 1986. Implications

Wood, J., 1986. Moon over Mauna ses of formation Phillips

of isotopic

Lunar

and

Loa: a review of hypothe-

of Earth’s

moon.

Taylor

(Editors),

G.J.

and Planetary

institute,

In: W.K. Hartman, Origin

Houston,

R.J.

of the Moon.

Tex.. pp. 17-55.