Geochemistry, petrology and ages of the lunar meteorites Kalahari 008 and 009: New constraints on early lunar evolution

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 72 (2008) 4845–4873

Geochemistry, petrology and ages of the lunar meteoritesKalahari 008 and 009: New constraints on early lunar evolution

A.K. Sokol a,b,*, V.A. Fernandes c,d,1, T. Schulz b,e, A. Bischoff a, R. Burgess c,R.N. Clayton f, C. Munker g,b, K. Nishiizumi h, H. Palme e, L. Schultz i,

G. Weckwerth e, K. Mezger b, M. Horstmann a

a Institut fur Planetologie, Wilhelm-Klemm-Str. 10, 48149 Munster, Germanyb Institut fur Mineralogie, Corrensstr. 24, 48149 Munster, Germany

c School of Earth, Atmospheric and Environmental Sciences, The University of Manchester, Oxford Road, Manchester M13 9PL, UKd Centro de Geofısica, Universidade de Coimbra, Coimbra 3000-134, Portugal

e Institut fur Mineralogie und Geochemie, Universitat zu Koln, Zulpicher-Str. 48b, 50674 Koln, Germanyf Enrico Fermi Institute, University of Chicago, Chicago, IL 60637, USA

g Mineralogisch-Petrologisches Institut, Universitat Bonn, Poppelsdorfer Schloss, 53115 Bonn, Germanyh Space Sciences Laboratory, University of California, 7 Gauss Way, Berkeley, CA 94720-7450, USA

i Max-Planck-Institut fur Chemie, Abt. Kosmochemie, Postfach 3060, 55020 Mainz, Germany

Received 28 January 2008; accepted in revised form 8 July 2008; available online 26 July 2008

Abstract

Kalahari 008 and 009 are two lunar meteorites that were found close to each other in Botswana. Kalahari 008 is atypical lunar anorthositic breccia; Kalahari 009 a monomict breccia with basaltic composition and mineralogy. Based onminor and trace elements Kalahari 009 is classified as VLT (very-low-Ti) mare basalt with extremely low contents ofincompatible elements, including the REE. The Lu–Hf data define an age of 4286 ± 95 Ma indicating that Kalahari009 is one of the oldest known basalt samples from the Moon. It provides evidence for lunar basalt volcanism priorto 4.1 Ga (pre-Nectarian) and may represent the first sample from a cryptomare. The very radiogenic initial 176Hf/177Hf(eHf = +12.9 ± 4.6), the low REE, Th and Ti concentrations indicate that Kalahari 009 formed from re-melting of man-tle material that had undergone strong incompatible trace element depletion early in lunar history. This unusuallydepleted composition points toward a hitherto unsampled basalt source region for the lunar interior that may representa new depleted endmember source for low-Ti mare basalt volcanism. Apparently, the Moon became chemically very het-erogeneous at an early stage in its history and different cumulate sources are responsible for the diverse mare basalttypes.

Evidence that Kalahari 008 and 009 may be paired includes the similar fayalite content of their olivine, the identical initialHf isotope composition, the exceptionally low exposure ages of both rocks and the fact that they were found close to eachother. Since cryptomaria are covered by highland ejecta, it is possible that these rocks are from the boundary area, wherebasalt deposits are covered by highland ejecta. The concentrations of cosmogenic radionuclides and trapped noble gasesare unusually low in both rocks, although Kalahari 008 contains slightly higher concentrations. A likely reason for this dif-ference is that Kalahari 008 is a polymict breccia containing a briefly exposed regolith, while Kalahari 009 is a monomict brec-ciated rock that may never have been at the surface of the Moon.

Altogether, the compositions of Kalahari 008 and 009 permit new insight into early lunar evolution, as both meteoritessample lunar reservoirs hitherto unsampled by spacecraft missions. The very low Th and REE content of Kalahari 009 as

0016-7037/$ - see front matter � 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2008.07.012

* Corresponding author. Address: Institut fur Planetologie, Wilhelm-Klemm-Str. 10, 48149 Munster, Germany. Fax: +49 251 8336301.E-mail address: [email protected] (A.K. Sokol).

1 Present address: Berkeley Geochronology Center, Berkeley, CA 94709, USA.

mailto:[email protected]

4846 A.K. Sokol et al. / Geochimica et Cosmochimica Acta 72 (2008) 4845–4873

well as the depletion in Sm and the lack of a KREEP-like signature in Kalahari 008 point to a possible source far from theinfluence of the Procellarum-KREEP Terrane, possibly the lunar farside.� 2008 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Lunar meteorites are an important source of new datacomplementing the information obtained from samples re-turned from the Moon by the Apollo and Luna missions.The special value of lunar meteorites derives from the factthat they may come from previously unsampled regions ofthe Moon such as the farside and they thus may improveour understanding of processes active during formation of lu-nar crust, the lunar mantle composition and volcanic history.

In September 1999, Kalahari 008 and 009 were recov-ered during geological field work in Botswana together withnine other meteorites. So far, these are the only knownmeteorite samples from Botswana (Grady, 2000; Koblitz,2003). Kalahari 008 (598 g) is a typical lunar anorthositicbreccia. Kalahari 009, a single rock weighing about13.5 kg, resembles a brecciated very-low-Ti (VLT) lunarmare basalt (Sokol and Bischoff, 2005a,b). Both meteoriteswere found about 50 m apart in front of a small dune.

This study presents a comprehensive dataset includingthe mineralogy, texture, mineral chemistry, major, minorand trace element whole-rock chemistry, noble gases, oxy-gen isotope composition, 40Ar–39Ar and 176Lu–176Hf-chro-nology of Kalahari 008 and 009. Preliminary results havebeen presented earlier by Sokol and Bischoff (2005a,b),Nishiizumi et al. (2005), Fernandes et al. (2006, 2007) andSchulz et al. (2007). The goals of this contribution are tocompare these two meteorites with the existing dataset onlunar samples and to provide constraints on their petrogen-esis and shock history.

2. ANALYTICAL PROCEDURES

2.1. Petrography and mineral compositions

Three polished thin-sections of each sample were studiedby optical microscopy in transmitted and reflected light. AJEOL 840A scanning electron microscope (SEM) was usedto resolve the fine-grained texture of the rocks. Mineralanalyses were obtained with a JEOL JXA-8600 S and aJEOL 8900 electron microprobe (EPMA) at the UniversitatMunster as well as with a JEOL 8900RL electron micro-probe at the Universitat zu Koln. Accelerating voltageswere set to 15 kV, the beam current to 15 nA, and the beamsize was 1 lm. Data from the JEOL JXA-8600 S and JEOL8900RL electron microprobes were corrected using theZAF correction procedures; data from the JEOL 8900 elec-tron microprobe were corrected using the PRZ-method(Armstrong, 1991).

2.2. Major and trace element analyses

Major and minor elements were determined by XRFanalyses. Splits from Kalahari 008 and 009 were cleaned

in an ultrasonic bath and each split was homogenized bycrushing in an agate mortar. All samples were preparedas fused beads made from 120 mg finely ground samplepowder (an aliquot from �2 g of crushed and homogenizedsample material) and 3600 mg Li2B4O7 flux. All analyseswere performed using a Philips PW 2400 XRF spectrometerat the Universitat zu Koln. Concentrations of Si, Mg, Al,Ca, Ti, Cr, Mn, Fe, P and V were determined with a preci-sion of about 5%.

For instrumental neutron activation analysis (INAA),powders (�150 mg aliquots from 2 g of homogenized sam-ple material) of Kalahari 008 and 009 were encapsulated ina silica tube and irradiated for 6 h in the TRIGA researchreactor at the Universitat Mainz with a thermal neutronflux of 7 � 1011 cm�2 s�1. Gamma ray spectra were col-lected using high-purity germanium detectors at the Univer-sitat zu Koln. The procedures were essentially the same asthose that have been used at the cosmochemistry depart-ment of the Max-Planck-Institut fur Chemie in Mainz foranalyzing lunar samples and meteorites for more than 20years. A detailed description of these procedures can befound in Wanke et al. (1977). In addition, data was reducedwith the program of Kruse and Spettel (1982). Analyticalprecision for most of the 30 analyzed elements ranges from5% to 10%. Some elements were analyzed by both XRF andINAA (Fe, Ca, Cr, etc.). In general, agreement to within5% was found between the results of both methods.

The two meteorite samples (100–200 mg of powder orig-inating from gram-sized chips) were also analyzed for theirNb, Ta, Zr and Hf concentrations using the high precisionisotope dilution method of Munker et al. (2001) and Weyeret al. (2002). A mixed 94Zr–176Lu–180Hf–180Ta tracer wasused and 2r external reproducibilities are ±4% for Nb/Taand ±0.6% for Zr/Hf.

2.3. Isotope measurements

2.3.1. Oxygen isotopes

Oxygen isotopes were measured on O2 sample gas, pre-pared by BrF5 reaction of 4–6 mg powdered samples for16 h at 600–650 �C, following the procedure of Claytonand Mayeda (1983). Gas samples were purified by the13� molecular sieve procedure. The meteorite samples werenot pretreated to remove possible terrestrial weatheringproducts.

2.3.2. Rb–Sr

For mineral separation, about 2 g of Kalahari 008 and009 were crushed to <180 lm using an agate mortar andpestle. After sieving using a nylon sieve stack, the 180–60 lm size fraction was handpicked under a binocularmicroscope.

Plagioclase fractions, handpicked to be optically free ofimpurities, from Kalahari 008 (2.7 mg) and 009 (10.38 mg)

Kalahari 008 and 009: New constraints on early lunar evolution 4847

were leached in order to remove any fine-grained terrestrialalteration products. Feldspars from Kalahari 008 were lea-ched only with hot aqua regia (for 3 h), whereas stepwiseleaching in progressively stronger acids was applied to feld-spars from Kalahari 009. The first step involved treatmentwith cold acetic acid in an ultrasonic bath for 30 min(2 ml; Leachate 1 in Table 12). In the next step, the sameplagioclase fraction was leached for 30 min in an ultrasonicbath with cold 3 N HCl (2 ml; Leachate 2 in Table 12). Sub-sequently, the plagioclases were subjected to aggressiveleaching with hot aqua regia (2 ml, 70 �C) for 30 min(Leachate 3 in Table 12). The acid leachates were separatedfrom the residue and a 10% aliquot of each leachate solu-tion was removed and spiked with Rb–Sr tracers for thedetermination of Rb and Sr concentrations. The residualplagioclase was washed with ultrapure H2O prior to disso-lution in HF–HNO3 (5:1). After dissolution, this final resi-due was dried down three times with small amounts ofconcentrated HNO3 to break down insoluble fluorides. Fi-nally, all solutions (leachates and residue) were dried downon a hot plate under a flow of clean air and taken up with3 M HNO3. Strontium was separated from other elementsusing Sr-spec resin. Rubidium was separated on cation ex-change resin (DOWEX AG-50W-X8, 100–200 mesh, resin)using 2.5 N HCl. The leachates and the residue were ana-lyzed separately for Rb and Sr. Strontium was loaded onW filaments with TaF5 and measured on a multi-collectorthermal ionization mass spectrometer (Finnigan Triton atthe Universitat Munster) in static mode. 87Sr/86Sr ratioswere normalized to 86Sr/88Sr = 0.1194 (Steiger and Jager,1977). The average 87Sr/86Sr for the NBS 987 standard dur-ing the course of the measurement was 0.710278 ± 0.000012(2r, n = 96). Rubidium was measured with a MC-ICPMS(Micromass Isoprobe at the Universitat Munster) followingthe method described in Nebel et al. (2005). The Sr and Rbisotope data for feldspar separates are given in Table 12.

2.3.3. Lu–Hf

For the Lu–Hf analyses, two pyroxene-rich fractions,two fractions of opaque phases, two powdered whole-rockaliquots and one plagioclase-rich fraction were separatedfrom a �4 g fragment of Kalahari 009. Lu–Hf analysis ofKalahari 008 was carried-out on a powdered whole-rocksplit of 100 mg. The pyroxene-rich fractions and the frac-tions of opaque phases were leached in cold 2.5 N HClfor 15 min. Due to their very small grain sizes (<60 lm),the plagioclase and the whole-rock aliquots were notleached.

The samples were spiked with a mixed 176Lu–180Hf tra-cer and digested in HF–HClO4–HNO3 (3:1:1) at 180 �C inclosed Savillex� vials placed on a hot plate for 48 h. Thesolutions were then evaporated to dryness and the residuesrepeatedly treated by adding 1–2 ml of concentrated HNO3

with a trace of HF. After evaporation, the samples were re-dissolved in 6 ml of 6 M HCl–0.06 M HF, the vials wereclosed and heated on a hot plate at 6150 �C to ensure fullsample spike equilibration. Subsequently, the samples werecompletely dried down. Finally the samples were dissolvedin 4–5 ml 3 M HCl–0.1 M ascorbic acid and separation ofLu and Hf from the matrix was accomplished with a sin-

gle-column separation technique using Eichrom Ln-spec re-sin (Munker et al., 2001). The Hf and Lu isotope ratios ofall samples were determined by MC-ICPMS in static modeusing the Micromass Isoprobe at the Universitat Munster.For Lu isotope dilution measurements, mass bias correc-tion was done by normalizing the measured and interfer-ence-corrected 176Lu/175Lu to the measured 173Yb/171Yb.For interference correction of 176Yb, the ratio 176Yb/171Ybof pure Yb was measured in the same run. The externalreproducibilities reported in Table 10 include effects of er-ror magnification due to nonideal spike-sample ratios aswell as uncertainties due to Yb corrections.

Hafnium isotope ratios were corrected for interferencesand mass bias following the procedure in Munker et al.(2001), employing the exponential law and 179Hf/177Hf =0.7325 for mass bias correction. All 176Hf/177Hf data are gi-ven relative to a value of 0.282160 for the Munster AMESstandard that is isotopically indistinguishable from theJMC-475 standard. The typical external reproducibility isca. ±50 ppm for �20 ng of Hf. To estimate the externalreproducibility an empirical relationship was used wherethe 2r external reproducibility equals 4� the 1r internalstandard error.

Measured blanks for the analytical procedure were 30 pgHf and 11 pg Lu. Measured isotope ratios and concentra-tions are listed in Table 10. The age and its uncertainty werecalculated using the Isoplot/Ex 2.49 program of Ludwig(2001) and a 176Lu decay constant of k = 1.867 �10�11 yr�1 (Scherer et al., 2001; Soderlund et al., 2004).

2.3.4. 40Ar–39Ar, noble gases and cosmogenic radionuclides

Two samples of Kalahari 009 were prepared for Ar–Arage determination. A thick-section (0.01646 g; �2.9 ��2.3 mm) was used for petrographic study using SEMand chemical analysis using EPMA. After completion ofthe petrographic and chemical study the section was re-moved from the glass slide and thoroughly cleaned withacetone to remove all adhesive. The section, together witha whole-rock fragment (0.02767 g) of Kalahari 009 weresealed in quartz vials and irradiated in position B2W atthe SAFARI-1 reactor in Pelindaba, South Africa. The fastneutron flux was monitored by Hb3gr monitors, in closeproximity to the samples, at 1.8 � 1018 neutrons cm�2

(J = 0.009248 ± 0.000128).Initially, it was intended to use the thick-section of Kal-

ahari 009 for laser Ar–Ar analysis of individual minerals,however, the yields of Ar-gas produced with the laser wereonly just above background levels and hence this approachwas abandoned. The thick-section was subdivided into twoparts: one containing the brecciated portion of the meteor-ite including a K-rich vein that is probably of terrestrial ori-gin, and the other portion representing a basalt fragment.Argon gas was extracted from the Kalahari 009 fragmentand thick-section using resistance furnace step heating overthe temperature interval of 300–1600 �C using steps of 100or 50 �C. The blank at low temperature (300–1200 �C), inunits of 10�12 cm3 STP was 40Ar = 964.6 ± 7.5; 38Ar =1.7 ± 0.6; 36Ar = 3.3 ± 0.5 (uncertainties: 2r). These valuesincreased at high temperature (P1300 �C) to 40Ar =2606.1 ± 5.2; 38Ar = 3.1 ± 0.6; 36Ar = 9.0 ± 0.3. Most Ar


blanks have isotopic compositions of approximately terres-trial atmospheric composition. Data were corrected forblank contributions, mass discrimination, radioactive decayand neutron interference reactions. Uncertainties on Ar–Arstep ages include the 1% difference in J-value obtained forthe monitors and, for total ages, an additional 1.4% uncer-tainty introduced by the uncertainty on the age determina-tion of the Hb3gr monitor (1072 ± 11 Ma; Turner, 1971).Ar–Ar ages are reported with uncertainties of two standarddeviations (2r). Further details of the experimental meth-ods and data reduction procedures are given by Fernandesand Burgess (2005).

The concentrations and isotope compositions of He, Neand Ar, as well as the concentrations of the main isotopesof Kr and Xe were determined in two bulk samples of Kal-ahari 008 and 009 (54.1 and 52.8 mg). The experimentaltechniques and methods were described by Scherer andSchultz (2000). Results are given in Table 8.

Fragments of Kalahari 008 (135.8 mg) and 009(165.2 mg) were used for cosmogenic radionuclide measure-ments. In order to eliminate terrestrial weathering products,each sample was etched with a 0.5 N HNO3 solution in anultrasonic bath for 10 min. After powdering the samples,each sample was then dissolved in an HF–HNO3 mixturealong with 0.13 mg of Be and 1 mg of Cl carriers. The con-centrations of Mg, Al, Ca, Mn, and Fe in the sample weredetermined by atomic absorption spectroscopy. Beryllium,Al, Cl, and Ca were chemically separated and purified foraccelerator mass spectrometry (AMS) measurements. TheAMS measurements of 10Be (half-life = 1.36 � 106 yr),26Al (7.05 � 105 yr), 36Cl (3.01 � 105 yr), and 41Ca(1.04 � 105 yr) in each sample were conducted at LawrenceLivermore National Laboratory (Davis et al., 1990).

3. RESULTS

3.1. Petrography and mineral chemistry

3.1.1. Kalahari 008

Kalahari 008 is a lunar anorthositic breccia composed ofvarious types of lithic clasts and isolated mineral fragmentsembedded in a fine-grained clastic matrix. The matrix con-sists of fine-grained, rounded and angular mineral fragmentsthat are well consolidated (Sokol and Bischoff, 2005a,b).

The majority of clasts in Kalahari 008 are individualmineral fragments. Plagioclase fragments are the mostabundant; some are as large as 1 � 1.5 mm (Fig. 1a). Cata-clastic pyroxene and olivine are present as well. Otherabundant fragments are fine-grained impact melt brecciashaving a crystalline or glassy matrix. On the basis of theirmineral compositions, they can be subdivided into feld-spar-rich and mafic breccias. There is a great variety of tex-tures for the impact melt breccias ranging from very fine-grained to microporphyritic. A large number also showsrecrystallized subophitic to variolitic textures. Fig. 1bshows a crystalline microporphyritic melt breccia. It con-sists of equant plagioclase fragments set in a fine-grainedmatrix. The matrix itself is composed of microcrystallineplagioclase and an interstitial mafic phase (Fig. 1e). A typ-ical feldspar-rich breccia with a subophitic texture is shown

in Fig. 1d. Here, the dendritic plagioclase is partly enclosedby a mafic phase. In contrast to the feldspathic melt brec-cias, mafic melt breccias display higher amounts of maficminerals and contain plagioclase as well as pyroxene asfragments (Fig. 1e). The groundmass in such breccias is ex-tremely rich in mafic phases and ranges in texture from fine-grained microporphyritic to partly glassy. Granulitic lithol-ogies and fragments of primary igneous rocks have alsobeen observed within Kalahari 008. Most granulitic frag-ments consist predominantly of plagioclase, with variableamounts of olivine, pyroxene, and trace amounts ofFe,Ti-oxides. Only one impact melt spherule was observedin the studied thin-sections (Fig. 1f). It is devitrified andconsists of plagioclase, olivine and pyroxene.

The major minerals in the clasts and in the matrix areplagioclase, followed by pyroxene and minor olivine.Accessory phases are ilmenite, troilite, FeNi metal, chro-mite and ulvospinel. The opaque phases appear as irregulargrains and are homogeneously distributed within the brec-cia. Most plagioclase is anorthite (An92–99; Fig. 2). Pyrox-ene crystals display a wide range of compositions (Fs14–77

Wo0.5–39 En8–76; Fig. 3). Several pyroxene fragments pos-sess exsolution lamellae that are up to 8 lm wide. Olivineoccurs mostly in the matrix and is much less abundant thanpyroxene. It displays a distinct bimodal chemical composi-tion (�Fa42–74 and Fa78–98; Fig. 2). A few grains have Fa-contents as low as 34 mol %. The Fa-rich olivines occuras individual mineral fragments within the clastic matrix.Lithic clasts containing these olivines have not been ob-served. Representative chemical compositions of mineralphases are listed in Table 1.

3.1.2. Kalahari 009

Kalahari 009 is a monomict breccia consisting of frag-ments of basaltic lithologies embedded in a fine-grained,heterogeneous matrix (Fig. 4). The basaltic clasts have acoarse-grained subophitic texture (Fig. 4a and c) and reachsizes of up to 3 � 4 mm. The matrix consists of compacted,fine- and coarse-grained mineral debris (Fig. 4b and c), thatis occasionally cemented by impact melt. The main constit-uent is pyroxene followed by plagioclase; olivine is lesscommon. Accessory minerals are ilmenite, chromite, troi-lite, Cr-rich ulvospinel, baddeleyite, phosphates, and Fe,Nimetal (Sokol and Bischoff,2005a,b). The modal mineralogyincludes 50–55% pyroxene, �35% plagioclase, �10% oliv-ine, and <5% accessory phases.

Symplectitic intergrowths of hedenbergite + faya-lite + SiO2 are common in this sample (Fig. 4d) and theyoccur typically between pyroxene grains. Some representrounded objects up to 300 lm in size. Symplectites havebeen reported in several lunar and martian basalts, suchas Asuka-881757 (Oba and Kobayashi, 2001), NWA 773(Fagan et al., 2003; Joliff et al., 2003), Los Angeles andQUE 94201 (Aramovich et al., 2002). They are usuallyinterpreted to result from pyroxferroite breakdown duringcooling (Papike et al., 1998).

Assemblages of K-feldspar–silica–plagioclase–pyroxene,which are common in many lunar basalts (e.g., NWA 773;Fagan et al., 2003; Joliff et al., 2003), were not observed inthe studied thin-sections, but phases that are usually associ-

Fig. 1. Backscattered electron (BSE) images of some typical fragments in Kalahari 008. (a) Large individual plagioclase fragment. (b)Crystalline microporphyritic melt breccia consisting of plagioclase fragments set in a fine-grained matrix. (c) Enlarged image showing thematrix of the melt breccia shown in (b); it consists of microcrystalline plagioclase (dark gray) and an interstitial mafic phase. (d) A typicalfeldspar-rich breccia with a subophitic texture; the dendritic plagioclase is partly enclosed by a mafic phase. (e) Mafic melt breccia withplagioclase and pyroxene as fragments and fine-grained microporphyritic to partly glassy matrix. (f) Devitrified impact melt spheruleconsisting of plagioclase, olivine and pyroxene. Plag, plagioclase; px, pyroxene; ol, olivine.


ated with these late-stage assemblages such as baddeleyiteand phosphates were found. Pyroxene is the most abundantmineral in the Kalahari 009 basalt breccia and occurs inbasaltic fragments as well as in the matrix. Its compositionshows strong variation in Ca–Mg–Fe contents, following atypical mare basalt fractionation trend. The grains are com-posites of pigeonite and augite ranging in composition fromFs22–67 En10–64 Wo6–41 (Fig. 3). (Ferro-)Hedenbergite oc-curs within symplectitic intergrowths. The majority ofpyroxenes display exsolution lamellae (Fig. 4e). Most ofthem are <5 lm wide (Fig. 4e); in some cases submicronexsolution lamellae are just barely visible under opticaland electron microscopes. Some grains without visible exso-lution lamellae using an electron microscope clearly showzonation (Fig. 4a and 6). Fe/Mn in randomly selectedpyroxene grains show some scatter and are plotted in Fig. 5.

Most feldspars are anorthites (An>90), but someare more sodic (Table 2 and Fig. 2) and An-contents

as low as 72 mol % have been measured. Some individ-ual plagioclase crystals show twinning; in some of themthe transformation of plagioclase to maskelynite isvisible.

The fine-grained olivine is Fe-rich, ranging from Fa46 toFa99.9 (Fig. 2). Most olivines occur as very fine-grainedangular or rounded crystals in the matrix. Some occur asphenocrysts in the basaltic fragments and are surroundedby pyroxene and plagioclase (Fig. 4c). Fayalitic olivine alsooccurs as intergrowths in symplectites. Representative com-positions of the mineral phases in Kalahari 009 are listed inTable 2. The Fe/Mn ratios of randomly selected olivinegrains are shown in Fig. 5. Ilmenite is the most commonopaque phase and occurs as rounded grains. Oxide compo-sitions do not vary significantly. Average compositions ofspinel and ilmenite are listed in Table 3. Metal grainsoccur in very small masses and are not Ni-rich(60.8 wt%; Table 4).

0

8

16

24

32

71 75 79 83 87 91 95 99An in plagioclase (mol %)

An in plagioclase (mol %)

No.

of M

easu

rem

ents

0

2

4

6

8

24 32 40 48 56 64 72 80 88 96Fa in olivine (mol %)

Fa in olivine (mol %)

No.

of m

easu

rem

ents

a bKalahari 008 Kalahari 008

0

6

12

18

71 75 79 83 87 91 95 99

No.

of M

easu

rem

ents

0

7

14

21

28

35

24 32 40 48 56 64 72 80 88 96

No.

of m

easu

rem

ents

c dKalahari 009 Kalahari 009

Fig. 2. Histograms showing the frequency of olivine composition (Fa-contents) and An-contents of plagioclases in Kalahari 008 (a and b) and009 (c and d).

Fig. 3. Pyroxene quadrilateral for Kalahari 008 (a) and 009 (b).


3.2. Shock effects

Kalahari 008 and 009 contain abundant petrographicevidence for impact-induced shock effects. The observedshock effects are typical of shock pressures in excess of30–35 GPa according to the calibration scheme of Stoffleret al. (1991) for ordinary chondrites (S4). Characteristic fea-tures are mosaicism and planar fracturing in feldspar andolivine. Rarely the transformation of plagioclase to mask-elynite is visible. In Kalahari 008 all fragments are shocked

to almost the same degree. In some areas of Kalahari 009the clastic matrix was shock-melted which has led to stronglithification of the entire rock.

3.3. Terrestrial weathering

Only weak evidence of terrestrial oxidation or hydrationin metal and troilite has been detected. The weathering scaledeveloped by Wlotzka (1993) for ordinary chondrites isbased on petrographic criteria and when applied to Kalaha-ri 008 and 009 indicates a weathering grade of W1. Frac-tures due to terrestrial weathering are abundant. Most arefilled with almost pure Ca-carbonate. Some K-rich cauli-flower-like structures have been observed as fracture fillingsin both rocks (Fig. 7a). Fig. 7b also presents a K-map of avein filled with terrestrial alteration products obtainedusing EPMA. The newly formed phase is dominated bySi, Al, K, Na Ca and Mg and gives low analytical totals(about 46 wt% SiO2, 13 wt% Al2O3, 8 wt% K2O, 4 wt%Na2O, 2 wt% CaO and 1.2 wt% MgO). The low analyticaltotal of �75 wt% may indicate that the weathering productsare hydrous, but inter- or intragranular porosity may alsocontribute to the low totals. Silica was also found withina meshwork of alteration products that may have formedby dissolution of olivine and pyroxene.

3.4. Whole-rock compositions and oxygen isotopes

3.4.1. Kalahari 008

The bulk-rock composition of Kalahari 008 is given inTable 5. In the same table its chemical composition is

Table 1Representative electron microprobe analyses of olivine, pyroxene, and plagioclase in Kalahari 008 (wt%); n.d., not detected

Olivine

SiO2 34.5 33.5 35.3 35.8 34.7 32.7 34.5 29.9 31.6 31.2 33.9TiO2 <0.03 0.06 <0.01 0.07 0.07 0.14 0.08 0.12 0.10 0.05 0.09Al2O3 0.20 0.06 <0.02 0.05 <0.02 <0.04 0.31 0.26 1.21 0.07 n.d.FeO 45.3 48.4 39.1 38.0 38.8 47.1 42.3 63.6 60.1 56.7 40.5MnO 0.46 0.62 0.50 0.48 0.47 0.63 0.48 0.72 0.74 0.40 0.58MgO 20.2 17.5 25.9 26.4 26.2 19.0 23.1 3.7 4.3 11.4 23.4CaO 0.41 0.44 0.35 0.35 0.50 0.38 0.29 1.00 0.42 0.58 0.30Na2O <0.02 n.d. <0.02 n.d. <0.02 <0.02 n.d. <0.03 0.04 <0.03 n.d.K2O <0.02 n.d. n.d. <0.02 <0.01 n.d. n.d. n.d. 0.06 n.d. n.d.Cr2O3 n.d. n.d. 0.13 <0.07 <0.07 <0.02 <0.06 n.d. <0.03 <0.03 <0.04

Total 101.07 100.58 101.28 101.15 100.74 99.95 101.06 99.30 98.57 100.40 98.77

Fa 55.75 60.86 45.91 44.62 45.36 58.24 50.69 90.55 88.61 73.65 49.29

Pyroxene

SiO2 47.2 50.4 50.4 49.7 52.0 35.0 49.1 49.6 50.9 47.7 50.2TiO2 0.55 0.22 0.58 0.86 0.60 0.11 0.94 0.90 0.55 1.03 1.18Al2O3 1.74 3.26 2.18 0.89 1.17 0.02 1.32 2.56 0.82 0.99 1.81FeO 36.3 27.8 21.2 30.3 22.6 40.7 28.9 19.0 24.5 26.7 15.4MnO 0.54 0.53 0.31 0.57 0.32 0.23 0.76 0.28 0.16 0.45 0.29MgO 3.4 14.9 18.7 10.7 18.7 22.9 12.5 15.7 16.3 4.9 14.8CaO 9.2 2.56 3.92 6.9 4.1 0.31 5.8 10.1 6.6 17.5 14.5Na2O 0.06 <0.03 0.64 <0.04 <0.04 <0.04 <0.04 <0.02 0.13 0.07 0.08K2O n.d. <0.02 0.06 <0.01 n.d. n.d. n.d. <0.01 <0.02 n.d. n.d.Cr2O3 0.10 0.63 0.57 0.18 0.49 0.06 0.19 0.73 0.87 0.23 0.62

Total 99.09 100.30 98.56 100.10 99.98 99.33 99.51 98.87 100.83 99.57 98.88

Fs 66.86 48.23 35.66 52.03 36.98 39.69 49.36 31.73 39.47 46.16 25.48En 11.34 46.07 55.89 32.66 54.50 59.72 38.03 46.69 46.82 15.13 43.63Wo 21.80 5.70 8.45 15.31 8.52 0.59 12.61 21.58 13.70 38.70 30.88

Plagioclase

SiO2 46.1 46.4 44.7 44.0 45.3 44.3 44.2 43.8 44.8 44.8 44.8TiO2 n.d. n.d. <0.01 n.d. 0.08 0.05 <0.02 <0.03 n.d. 0.04 <0.02Al2O3 34.3 34.5 35.9 36.5 35.2 36.1 35.9 36.1 35.4 35.4 35.4FeO 0.11 0.12 0.21 0.08 0.15 0.09 0.24 0.10 0.27 0.27 0.27MnO n.d. <0.03 <0.02 <0.02 n.d. n.d. <0.03 n.d. <0.03 n.d. n.d.MgO 0.02 0.04 0.12 <0.01 0.14 0.09 0.15 <0.01 0.06 0.09 0.21CaO 17.7 17.9 19.3 19.4 18.9 19.5 19.6 19.8 18.9 18.9 18.9Na2O 1.12 1.12 0.35 0.36 0.76 0.38 0.28 0.26 0.64 1.08 1.00K2O <0.02 <0.02 0.12 n.d. <0.07 <0.02 <0.01 <0.01 <0.02 <0.03 0.11Cr2O3 n.d. n.d. <0.06 0.12 <0.07 n.d. <0.02 <0.02 n.d. <0.06 n.d.

Total 99.35 100.08 100.7 100.46 100.53 100.51 100.37 100.06 100.07 100.58 100.69

An 89.57 89.76 96.12 96.76 92.82 96.45 97.48 97.58 94.09 90.04 90.07Ab 9.34 9.22 3.09 3.14 6.37 3.31 2.43 2.3 5.47 8.93 8.55Or 0.14 0.10 0.71 0.00 0.40 0.13 0.03 0.07 0.13 0.17 0.65


compared with the compositions of Dar al Gani 262 (DaG262), Dhofar 081 (Dho 081), the average of lunar highlandmeteorites and the average lunar highland composition(Taylor, 1982; Palme et al., 1991; Bischoff et al., 1998).These data show that the chemical composition of Kalahari008 is very similar to that of other lunar highland meteor-ites. In common with the majority of lunar highland rocksKalahari 008 is rich in aluminum (�28% Al2O3) and haslow concentrations of incompatible elements. The elementratios Fe/Mn, Fe/Sc, Mg/Cr, Al/Ga, Na/Eu also fall withinthe range of other lunar highland rocks (Table 6). Fig. 8gives the bulk Fe versus Mn contents of different rocks inthe solar system, including Kalahari 008. Zr/Hf = 35.7

and Nb/Ta = 18.5 are shown in Fig. 9. The measured oxy-gen isotopes are as follows: d18O = +6.52&;d17O = +3.32&; D17O = �0.07&. These data points fallwithin analytical uncertainties of ±0.1& on the terres-trial–lunar-fractionation line.

Fig. 10 shows the chondrite-normalized rare earth ele-ment (REE) patterns for all the lunar meteorites Kalahari008, DaG 262 and Dho 081 and the average for lunar high-land meteorites. There is a close similarity between the REEpatterns of all samples. The concentrations levels are about5–10 times chondritic. The LREE show a slight enrichmentin all samples while the variations among the heavy REE(HREE) are more limited. Kalahari 008 shows lower

Fig. 4. Backscattered electron (BSE) images of typical fragments and textures in Kalahari 009. (a) Large, coarse-grained basaltic clast. (b) Anarea illustrating the matrix consisting of compacted, fine- and coarse-grained mineral debris. (c) A basaltic clast having olivine as phenocrystssurrounded by pyroxene and plagioclase. (d) Rounded symplectitic intergrowth of hedenbergite (gray) + fayalite (white) + SiO2 (black). (e) Apyroxene grain showing exsolution lamellae. (f) Glassy matrix containing schlieren and opaque grains. Plag, plagioclase; px, pyroxene; ol,olivine; calc, calcite.


REE abundances than Dar al Gani 262, but higher concen-trations than Dhofar 081. All samples display a positive Eu-anomaly of the same magnitude indicating a high abun-dance of anorthite.

3.4.2. Kalahari 009

The whole-rock major and trace element data for Kala-hari 009 are presented in Table 7. This table shows thechemical composition of Kalahari 009 compared to Asu-ka-881757 (low-Ti) and Y-793169 (very low-Ti, VLT) aswell as Apollo 14 (low-Ti), Apollo 17 (VLT) and Luna 24(VLT) basalts. Based on major and minor element compo-sition and following the mare basalt classification scheme ofNeal and Taylor (1992), who classify mare basalts accord-ing to their titanium content, Kalahari 009 can be classifiedas a VLT basalt, similar to VLT Apollo 17 and Luna 24samples (Fig. 11). In common with some Luna 24 VLT bas-alts (Neal and Taylor, 1992), Kalahari 009 contains morethan 11 wt% Al2O3 and thus falls within the range of

high-Al basalts. Most of the variations in major elementsof lunar rocks result from the variable contents of plagio-clase and pyroxene, the most common minerals, plus someolivine (Korotev, 2005). Therefore, in a plot of Al2O3 versusFeO + MgO all lunar rocks should lie on a line connectingthe composition of a pure feldspathic highland material andthe olivine plus pyroxene association in mare basalts. Asshown in Fig. 12, Kalahari 009 plots close to this line withinthe region for mare basalts. In total, the composition ofKalahari 009 is quite ferroan, with molar Mg/(Mg + Fe) = 0.49, consistent with the relatively evolvedbasaltic assemblages as shown in Fig. 4.

Fig. 8 gives the bulk-rock Fe versus Mn contents of Kal-ahari 009 and other rocks in the solar system. The elementratios Zr/Hf = 30.2 and Nb/Ta = 17.4 are given in Fig. 9relative to ratios for lunar and terrestrial rocks.

Incompatible elements including REE, Ti and Th arelow in Kalahari 009 with Th being exceptionally stronglydepleted (Fig. 13). These low abundances are in contrast

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.0 0.5 1.0 1.5 2.0Fe (afu)

Mn

(afu

)

Kalahari 008 olivine Earth

Moon

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.0 0.5 1.0 1.5 2.0

Fe (afu)

Mn

(afu

)

Kalahari 008pyroxene

Earth

Moon

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.0 0.5 1.0 1.5 2.0Fe (afu)

Mn

(afu

)

Kalahari 009 olivine Earth

Moon0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.0 0.5 1.0 1.5 2.0

Fe (afu)

Mn

(afu

)

Kalahari 009pyroxene

Earth

Moon

a b

c d

Fig. 5. Concentrations of Fe and Mn in olivine and pyroxene from Kalahari 008 (a and b) and 009 (c and d). The atoms per formula unit (afu)are based on 4 oxygens for olivine and 6 oxygens for pyroxene. The ratios of Fe/Mn for Earth and Moon associated with specific parentbodies are from Papike et al. (2003).


with the high concentration of K2O, because K-enrichedlunar rocks usually also have elevated REE concentra-tions. Thus, post-fall weathering is most likely responsiblefor this high abundance of K (Fig. 7). In Fig. 10b, REEpatterns of Kalahari 009 are compared with those of A-881757 (low-Ti), Y-793169 (VLT), Apollo 14 (low-Ti),Apollo 17 (VLT), and Luna 24 (VLT) basalts. Kalahari009 displays a small positive Eu-anomaly and the LREEare more depleted than the HREE. A positive Eu-anom-aly is quite uncommon among lunar mare basalts, but hasbeen observed in some small Luna 24 basalts samples(Fig. 10; e.g., Ma et al., 1978), that may not berepresentative.

The whole-rock has a d18O of +6.87& and a d17O of+3.45&. This composition plots along the terrestrial–lu-nar-fractionation line. The d18O values are slightly higherthan the usual lunar range. Since the rocks are feldspar-rich, this may be due to mineral-specific isotopefractionation or desert weathering as has been suggestedpreviously in the case of meteorites from Algeria (Stelzneret al., 1999).

3.5. Noble gases and 4p cosmic ray exposure ages

The concentrations of noble gases in Kalahari 008 and009 are given in Table 8. Concentrations of He in bothmeteorites are extremely low compared to other lunar mete-orites. Helium may have been lost by shock or thermalevents. Kalahari 008 shows trapped solar gases (Fig. 14)as demonstrated by its high 20Ne/22Ne ratio (Table 8).The relative noble gas abundances of Kalahari 008 are sim-ilar to Dar al Gani 400 and Yamato 82192 (Fig. 15). ForKalahari 009, however, the measured trapped componentis consistent with a terrestrial atmospheric contamination,which is masked in Kalahari 008 by solar gas.

The cosmogenic nuclide concentrations (10Be, 26Al, 36Cl,41Ca) in Kalahari 008 and 009 are given in Table 9. Theseare the lowest activities of cosmogenically produced radio-nuclides ever measured in a stony meteorite. The 41Ca con-centrations in both samples are below detection limit.Cosmogenic radionuclide concentrations of Kalahari 008are 40–70% higher than those of Kalahari 009 but nuclideratios are very similar for both samples indicating different

Table 2Representative electron microprobe analyses of olivine, pyroxene, and plagioclases in Kalahari 009 (wt%); n.d., not detected

Olivine

SiO2 35.4 32.8 29.9 31.6 29.6 30.0 30.4 29.7 30.5 30.3 29.5TiO2 0.02 0.08 <0.03 <0,03 0.15 0.15 0.09 0.09 0.09 0.12 0.18Al2O3 0.12 0.07 n.d. <0,01 0.05 <0.01 n.d. 0.04 <0.03 0.05 <0.03FeO 41.7 50.2 65.8 57.1 68.0 65.5 63.3 66.3 61.7 65.3 67.9MnO 0.46 0.61 <0.6 0.77 0.74 0.82 0.80 0.95 0.51 0.86 0.84MgO 21.6 16.0 3.8 9.9 1.50 3.6 5.2 2.6 6.3 4.0 0.9CaO 0.34 0.33 0.47 0.64 0.78 0.79 0.65 0.84 0.68 0.42 1.10Na2O n.d. 0.02 <0.03 0.02 <0.02 0.04 0.02 n.d. <0.01 <0.02 n.d.K2O <0.02 <0.01 n.d. <0.01 n.d. n.d. 0.02 0.01 n.d. n.d. n.d.Cr2O3 n.d. <0.02 n.d. n.d. 0.11 <0.07 <0.04 n.d. <0.06 <0.02 <0.03

Total 99.64 100.11 99.97 100.03 100.93 100.90 100.48 100.53 99.78 101.1 100.42

Fa 52.05 63.81 90.70 76.38 96.18 91.07 87.29 93.36 84.64 90.10 97.60

Pyroxene

SiO2 53.6 53.4 50.6 53.8 52.3 49.7 53.2 52.2 52.0 51.6 50.1TiO2 0.15 0.14 0.43 0.13 0.24 0.55 0.16 0.20 0.26 0.21 0.82Al2O3 1.82 2.38 1.87 2.45 1.74 2.01 1.74 1.95 1.80 2.37 1.26FeO 17.7 17.6 28.3 17.4 14.8 31.2 17.5 20.4 15.3 15.6 27.9MnO 0.44 0.39 0.50 0.34 0.36 0.63 0.31 0.35 0.28 0.41 0.32MgO 21.2 20.9 11.7 21.8 15.0 8.4 22.6 16.3 14.1 14.1 12.1CaO 5.3 5.0 6.8 4.1 14.9 7.8 4.0 8.7 15.9 15.3 7.3Na2O 0.06 0.15 <0.03 0.09 0.10 0.06 0.08 0.10 0.07 0.12 0.04K2O 0.06 0.05 <0.03 0.04 <0.01 <0.01 <0.02 0.04 <0.01 <0.03 <0.01Cr2O3 0.79 0.70 0.51 0.56 1.07 0.27 0.76 0.68 0.68 0.87 0.37

Total 101.12 100.71 100.71 100.71 100.51 100.62 100.35 100.92 100.39 100.58 100.21

Fs 28.41 28.75 48.91 28.24 24.4 55.55 27.77 33.65 25.12 25.79 47.43En 60.68 60.74 35.98 63.24 43.99 26.71 64.03 47.98 41.39 41.72 36.75Wo 10.91 10.51 15.11 8.51 31.52 17.74 8.20 18.37 33.48 32.49 15.82

Plagioclase

SiO2 45.9 47.9 46.6 44.5 44.8 45.2 47.0 45.3 45.2 45.5 45.2TiO2 n.d. <0.04 n.d. n.d. n.d. n.d. 0.09 <0.02 n.d. 0.05 n.d.Al2O3 34.0 32.7 30.1 33.4 34.2 34.0 33.7 34.3 34.3 34.2 34.0FeO 0.51 0.99 2.19 0.83 0.46 0.67 0.41 0.40 0.47 0.63 0.44MnO n.d. <0.01 0.13 n.d. <0.05 <0.03 n.d. <0.05 <0.03 <0.05 <0.05MgO 0.16 0.70 2.48 0.10 <0.05 0.19 0.15 0.16 0.17 0.14 0.13CaO 18.7 16.5 16.7 19.0 19.4 19.0 18.6 18.9 18.9 19.0 18.8Na2O 0.82 1.02 0.69 0.70 0.53 0.76 0.67 0.66 0.73 0.73 0.72K2O <0.03 0.28 0.11 <0.01 <0.01 <0.03 <0.01 <0.01 0.05 <0.02 <0.02Cr2O3 <0.02 n.d. <0.02 n.d. n.d. <0.07 n.d. n.d. n.d. n.d. n.d.

Total 100.09 100.09 99.00 98.53 99.39 99.82 100.62 99.72 99.82 100.25 99.29

An 92.46 88.30 92.37 93.68 95.23 93.05 93.00 94.03 93.25 93.37 93.38Ab 7.34 9.89 6.93 6.25 4.73 6.78 5.77 5.94 6.45 6.50 6.48Or 0.20 1.81 0.70 0.07 0.03 0.16 1.24 0.03 0.30 0.13 0.13


shielding conditions in the same object in space, either dur-ing Moon–Earth transit or on the Moon before ejection.The 26Al/10Be activity ratios of Kalahari 008 and 009 areboth 1.0 ± 0.3 thus significantly lower than the productionrate ratio of �4. A large portion of the 10Be may be mete-oric 10Be contamination, even though both samples wereacid leached. The combined 26Al and 36Cl 4p exposure agesare 350 ± 120 yr for Kalahari 008 and 220 ± 40 yr for Kal-ahari 009 which are the shortest exposure ages of any mete-orite. These ages imply that the transition time from theMoon to the Earth was 230 ± 90 yr and ejection depthwas more than >1100 g/cm2 (367 cm assuming density of3 g/cm3) on the Moon.

However, small amounts of cosmogenic nuclides arealso produced in-situ on the Earth’s surface. The 26Aland 36Cl concentrations in Kalahari 009 can be explainedby �0.3 Ma exposure time in the Kalahari Desert(1000 m elevation and 21�S). Long terrestrial ages, 0.3–0.5 Ma, were, for example, found for Dhofar lunar andmartian meteorites (Nishiizumi et al., 2002). In the caseof Kalahari 009, all cosmogenic nuclides could have beenproduced on the Earth’s surface without previous expo-sure in space. However, the 36Cl concentration in Kala-hari 008 is �15% higher than saturation of 36Clproduction on the Earth’s surface, therefore Kalahari008 must have been exposed in space. A more detailed

Table 3Representative electron microprobe analyses of oxides in Kalahari009

Ilm 1 Ilm 2 Ilm 3 Sp 1 Sp 2 Sp 3

SiO2 <0.01 <0.03 <0.03 n.d. <0.02 0.07TiO2 52.8 52.3 52.3 21.7 20.6 31.9Al2O3 0.06 0.08 0.07 3.24 3.35 2.18FeO 45.5 45.5 45.3 49.3 49.7 60.8MnO 0.29 0.32 0.28 0.88 0.79 0.13MgO 0.36 0.40 0.40 0.46 0.39 0.39CaO n.d. n.d. n.d. n.d. n.d. 0.28Na2O n.d. n.d. n.d. n.d. n.d. 0.04K2O <0.01 n.d. n.d. n.d. <0.01 n.d.Cr2O3 0.13 0.08 0.09 22.1 23.3 2.60

Total 99.14 98.68 98.44 97.68 98.13 98.39

n.d., not detected; Ilm, ilmenite; Sp, spinel.

Table 4Representative electron microprobe analyses of metal (wt%) inKalahari 009

P n.d. n.d. n.d. n.d. n.d.Co 0.69 2.10 0.73 0.68 1.53Cr <0.01 0.02 0.04 n.d. 0.02S 0.02 n.d. 0.02 <0.01 0.04Ni 0.17 0.31 0.05 0.25 0.80Mn n.d. n.d. n.d. <0.01 n.d.Fe 98.0 97.0 98.4 98.4 97.8

Total 98.88 99.43 99.24 99.33 100.19

n.d., not detected.


discussion on cosmogenic nuclides will be providedelsewhere.

Due to these low concentrations and the associateduncertainties in blank contributions only upper limits canbe given for cosmogenic 21Ne in both meteorites: <10�10

ccSTP21Ne/g for Kalahari 008 and <3 � 10�11

ccSTP21Ne/g for Kalahari 009. Assuming meteoritic pro-duction rates for 4p-irradiation (e.g., Wieler, 2002) expo-sure ages of <300 and <100 ka, respectively, are obtained.These upper limits for exposure ages are orders of magni-tude above the limits from 26Al. The cosmogenic 21Ne thatwas produced during the 350 yr of transit time from Moonto Earth is negligible compared to the 21Ne measured.

3.6. Chronology

3.6.1. Lu–Hf age

The Lu–Hf results are listed in Table 10 and shown inFig. 16. Two pyroxene-rich and two opaque-mineral-richfractions from Kalahari 009 plot on a well-defined isochron(MSWD = 1.4) with an initial 176Hf/177Hf ratio of0.28037 ± 0.00013 and a slope corresponding to a Lu–Hfage of 4286 ± 95 Ma (2r). The plagioclase and the whole-rock fractions do not plot on this isochron. These fractionswere not leached and may have been contaminated by theobserved terrestrial weathering products as inferred fromSr-isotope and trace element systematics. Terrestrial crustalrocks are characterized by low 176Hf/177Hf ratios (negativeeHf; Patchett, 1983). Therefore, it is likely that the inherited

terrestrial component lowered the measured 176Hf/177Hf ofthese fractions and caused them to plot below the regres-sion line. Inclusion of these contaminated fractions wouldgive an age of 4293 ± 370 Ma, with a large MSWD(=31), and an initial 176Hf/177Hf ratio of0.28024 ± 0.00047. In the following discussion, only theisochron obtained by excluding plagioclase and whole-rockfractions is considered.

Using the initial 176Hf/177Hf ratio of 0.28037 ± 0.00013,the initial eHf at 4286 Ma is +12.9 ± 4.6 (calculated using(176Hf = 177Hf)CHUR(0) = 0.282772, (176Lu/177Hf)CHUR(0) =0.0332 and the 176Lu decay constant of Scherer et al.(2001) and Soderlund et al. (2004)). For Kalahari 009, ahighly depleted source region is required to account for theradiogenic initial Hf composition. Assuming a formationage of the depleted source reservoir at �4.50 Ga (maximumformation age of the Moon; Touboul et al., 2007), a176Lu/177Hf value between 0.09 and 0.16 would be required.If the reservoir formed later, the 176Lu/177Hf value requiredfor the source region of Kalahari 009 would be even higher.Other mare basalts, ca. 3.2–3.8 Ga old, show initial eHf from0 to +45 at their time of formation (Patchett and Tatsumoto,1981; Unruh et al., 1984; Beard et al., 1998). At 3.8 Ga, thesource region of Kalahari 009 would have an initial eHf be-tween +28 and +60, at 3.2 Ga between +54 and +113 whichis higher than previously observed for mare basalts.

In Fig. 16 the whole-rock Lu–Hf isotope composition ofKalahari 008 falls close to the isochron defined by thepyroxene-rich and opaque-rich fractions from Kalahari009. Although both samples represent completely differentrock types, this observation may be evidence for an identi-cal initial Hf isotope composition. This would imply thatboth rocks formed at the same time from the same reservoirin the lunar interior.

3.6.2. 40Ar–39Ar systematics for Kalahari 009

Results of the 40Ar–39Ar step heating analyses of a bulksample, brecciated material and a fragment of a basalt clastseparated from the thick-section are given in Table 11. Ageand Ca/K spectra for the bulk sample are shown in Fig. 17and compared with the breccia and basalt in Fig. 18. Theapparent 40Ar–39Ar age spectrum for the bulk sample (Fig.17) shows a complex pattern. Only a minor amount oftrapped lunar Ar is present but a relatively large amount ofterrestrial atmospheric Ar was released, particularly at lowto intermediate temperatures (300–650 �C), comprising42% of the 39Ar released. The terrestrial origin of this Arwas established from the 40Ar/36Ar ratios of 285–340 whichare close to the present day terrestrial atmospheric value of295.5. At 650 �C a significant (18%) 39Ar release occurred(Ca/K = �0 in Fig. 17), again associated with atmospheric40Ar/36Ar ratios (Fig. 17). This Ar is most likely released fromsecondary K-rich minerals present in terrestrial alterationveins as seen in the K-map of the section (Fig. 7b). At inter-mediate and high temperatures apparent Ar–Ar ages rangefrom 0.4 to 2.7 Ga and correspond to 58% of the 39Ar re-leased. During the intermediate and high temperature re-leases, the Ca/K increased from 16 to 510 suggesting thatAr was being progressively released from plagioclase andpyroxene.

Fig. 7. (a) Cauliflower-like structures in fractures in Kalahari 008 and Kalahari 009; (b) K-map obtained by EPMA showing a K-rich veinwithin a basaltic clast (up to 4.3 wt% K2O from possibly terrestrial contamination). The colors from blue to red indicate increasingconcentration of K.

Fig. 6. (a) BSE image of Kalahari 009 and (b) Mg-map obtained by EPMA showing the zoning in the large pyroxenes. The colors from blueto red indicate increasing concentration of Mg.


The 40Ar–39Ar spectra for all three samples of Kalahari009 show considerable variation in apparent age from 0.89to >4.5 Ga for temperature steps above 50% 39Ar release(Fig. 18). At least some of this variation can be attributedto subtle changes in the relative amounts of terrestrial, radio-genic and minor trapped lunar Ar components that cannotbe adequately resolved using the present dataset. In particu-lar, the breccia aliquot shows large releases of Ca-derived37Ar at high temperature likely related to terrestrial Ca-richveins with anomalously high apparent ages >4.5 Ga. The ini-tial 26% of 39Ar released from the basalt fragment shows highapparent ages associated with a terrestrial 40Ar/36Ar. Theseages decrease to a minimum of 890 ± 10 Ma. For theremaining 74% Ar, released over the temperature interval850–1600 �C there is shows a progressive increase in appar-ent ages up to a maximum of 3.09 ± 0.18 Ga. This increasein apparent ages is correlated with the steady increase inCa/K from 228 to 1520 similar to values obtained by EMPanalyses of plagioclase and pyroxene.

The main release of 38Ar and 36Ar at the 1250 and1300 �C steps is coincident with major release of Ca-derived

37Ar. However the 36Ar/38Ar ratio for these, and in factnearly all other temperature steps, is lower than the cosmo-genic ratio of 0.65 (it is 0.34 and 0.32 for 1250 and 1300 �Csteps, respectively). This implies the presence of Cl-derived38Ar. It is therefore not possible to calculate an exposureage from these data. This does not mean cosmogenic Aris not present, as there is a relatively large release of 36Arat high temperature coincident with 37Ar release. This36Ar could be cosmogenic or represent an unknown trappedAr component. It could also be related to a Cl-rich phaseformed by terrestrial contamination. The absence of Neand He suggests that Kalahari 009 may have never been ex-posed directly at the lunar surface, consistent with the lackof solar gases. Alternatively, the lack of cosmogenic He andNe may be due to loss during the (impact) event that dis-turbed the K–Ar system.

3.7. Sr-isotopes in feldspars

Plagioclase crystallizing from a melt generally hasvery low Rb/Sr ratios. As a consequence the Sr-isotope

Table 5Chemical composition of Kalahari 008 (INAA and XRF) compared to Dar al Gani 262, Dhofar 081, average lunar highland and averagelunar highland meteorites

Element Kalahari 008 s.d. in % Dar al Gani 262 Dhofar 81 Average highlandsamples

Average lunar highlandmeteorites

Ref. (1) (2) (3) (4) (5)

wt%SiO2

b 44.4 5 44.9TiO2

b 0.28 5 0.22 0.15 0.57 0.27Al2O3

b 27.74 5 27.26 30.43 24.57 26.29FeO 4.44 5 4.58 2.93 6.60 5.24FeOb 5.06 5MgOb 4.44 5 5.21 2.82 6.80 6.12MnO 0.06 10 0.07 0.05 0.07MnOb 0.07 5CaO 14.9 5 16.89 16.80 15.80 15.40CaOb 15.5 5Na2O 0.54 3 0.36 0.31 0.45 0.33K2O 0.16 3 0.06 0.02 0.07 0.02

ppmPb 130 5 260 87Sc 10.0 3 7.85 5.4 9.06Cr 690 3 639 410 680 699Crb 701 5Co 10.8 3 22 9.8 15 17Ni 64 10 270 85 100 177Ga 2.8 15 4.3 2.4 3.4As 0.28 20Se 0.34 25Br 0.70 20Rb 3.6 15Sr 200 8 245 240 120 251Zr 34 30 34 63 35Zrc 16.78Nbc 0.944Cs 0.063 25 0.12 0.07 0.14Ba 65 10 240 19 66 66La 1.3 8 2.44 1.43 5.3 2.58Ce 4.5 10 7.25 3.4 12 6.6Nd 3.5 20 3.85 1.9 7.4 3.8Sm 0.90 5 1.15 0.63 2 1.17Eu 0.82 8 0.73 0.7 1 0.81Tb 0.17 8 0.24 0.15 0.41 0.24Dy 0.95 30 1.75 2.6 1.46a

Ho 0.24 20 0.3 0.18 0.53 0.28Yb 0.70 5 0.91 0.51 1.4 0.97Lu 0.11 5 0.13 0.073 0.21 0.14Luc 0.09636Hf 0.58 5 0.85 0.44 1.4 0.87Hfc 0.4699Ta 0.07 15 0.11 <0.1 0.13Tac 0.0511Ir 0.0035 15 0.012 0.005 0.039Au 0.003 15 0.004 0.07 0.003Th 0.18 10 0.43 0.2 0.9 0.41U 0.07 30 0.21 0.07 0.24 0.13

(1) This study; (2) Bischoff et al. (1998); (3) Warren et al. (2001); (4) Taylor (1982); (5) Cahill et al. (2004).a Taken from Palme et al. (1991).b XRF analysis.c Isotope dilution method.


composition does not evolve much over time andthus preserves a 87Sr/86Sr that is close to the initialvalue for the sample. Therefore plagioclase was sepa-

rated and analyzed for its isotope composition toobtain information on source of Kalahari 008 and009.

Table 6Element ratios in lunar meteorites and lunar highland rocks

Kalahari 008 Dar al Gani 262 Dhofar 81 Average highland samples Average lunar highland meteoritesRef. (1) (2) (3) (4) (5)

Fe/Mg 1.32 1.13 1.34 1.25 1.10Na/Ca 0.0267 0.0223 0.0192 0.0295 0.0222Sr/Ca 0.0018 0.0020 0.0020 0.0011 0.0023Eu/Ca 7.38 � 10�6 6.04 � 10�6 5.83 � 10�6 8.85 � 10�6 7.36 � 10�6

Fe/Mn 64 66 63 75Fe/Sc 3539 4535 4217 4492Mg/Cr 39 49 41 60 53Na/Eu 3618 3685 3286 3339 3023Al/Ga 52,433 33,551 67,104 40,923

(1) This study; (2) Bischoff et al. (1998); (3) Warren et al. (2001); (4) Taylor (1982); (5) Cahill et al. (2004).

Fig. 8. Fe/Mn versus Mn for different rocks in the solar system (after: Ostertag et al., 1986; from: Palme et al., 1991).

chond ritic

chondritic

Terrestrial array

MoonLunar Meteorites

Kalahari 008Kalahari 009

10 15 20 25 30 35 40 45 50 10

12

14

16

18

20

22

24

Zr/Hf

Nb/

Ta

Fig. 9. Nb/Ta versus Zr/Hf of the Kalahari meteorites comparedto values of lunar and terrestrial rocks (Munker et al., 2003).


The Rb and Sr concentrations and 87Sr/86Sr data forleachates and residues are listed in Table 12. The low Srconcentrations in the residues from both Kalahari 008and 009 suggest that a large amount of Sr was removed dur-ing the leaching procedure. The residuum (which is thoughtto represent clean plagioclase) from Kalahari 008 displaysan 87Sr/86Sr ratio of 0.699826 ± 31. The leachate has aslightly more radiogenic 87Sr/86Sr, because it probably con-

tains Sr from the (terrestrial) carbonates located infractures.

The results of the leaching experiments for Kalahari 009are slightly more complex. For this sample, the plagioclasewas leached progressively (see ‘‘Analytical Procedure”).The 87Sr/86Sr values decrease from Leachate 1 (acetic acid)to Leachate 3 (aqua regia) indicating that most of the ter-restrial calcite contamination was dissolved with aceticacid. Leachate 3 yielded the lowest 87Sr/86Sr of0.699242 ± 43 and thus is closest to the initial isotope valueof the sample. The residuum from Kalahari 009 shows avery radiogenic 87Sr/86Sr ratio of 0.713974 ± 13. Clearly,the residuum contained some material with high Rb/Sr,which however could not be identified.

4. DISCUSSION

4.1. Evidence for a lunar origin

Igneous textured clasts in Kalahari 008 and 009 indicatethat these rocks originate from a differentiated planetaryobject. Based on the oxygen isotope composition and con-sidering the known parent bodies only the Earth, the Moonand the enstatite parent asteroid can be considered as theparent body of Kalahari 008 and 009. The high Fe-concen-trations in olivine and clinopyroxene rule out an enstatite

1

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

chon

drite

nor

mal

ized

con

cent

ratio

n

average lunar highland meteoritesDaG 262Dho 81Kal 008

a

1

10

10 0

1000

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

chon

drite

nor

mal

ized

con

cent

ratio

n

Luna 24 Y-793169

Apollo 17 A-881757

Apollo 14 Kalahari 009

b

Fig. 10. Rare earth element concentrations (normalized to CI chondrites) of (a) Kalahari 008 compared to other highland meteorites and (b)Kalahari 009 compared to selected mare basalts. For references see Tables 5 and 7.


parent body as a potential source of these meteorites. Con-sequently, only the Earth and the Moon remain as possibleparent bodies for Kalahari 008 and 009.

Both rocks lack a fusion crust. Mineralogical and tex-tural similarities (e.g., impact melt breccias, melt spherules)of Kalahari 008 with other lunar anorthositic brecciasstrongly point towards a lunar origin. Kalahari 009 is abasaltic breccia with fragments having textures that resem-ble both lunar and terrestrial rocks.

A fundamental difference between lunar and terrestrialrocks is that lunar plagioclase is richer in Ca (mostly anor-thite; An > 90) compared to plagioclase of terrestrial rocks(An 6 90, Table 13). The anorthite-rich composition of pla-gioclases in Kalahari 008 and 009 makes a lunar originmore likely. Furthermore, Kalahari 009 contains FeNi me-tal, not present in terrestrial basalts, a common late-stagecrystallization product of lunar basalts and a direct effectof the low fO2

in the lunar mantle. Contamination withmeteoritic metal is a possibility; however, this can be ex-

cluded as meteoritic metal has a much higher Ni contentat 4 wt% (Table 4). The ancient age of �4.3 Ga obtainedby Lu–Hf (this study), U–Pb (Terada et al., 2007) andSm–Nd (Shih et al., 2008) also makes a terrestrial originof Kalahari 009 very unlikely (Table 13).

Further evidence for a lunar origin comes from the87Sr/86Sr ratios in feldspars. The initial 87Sr/86Sr ratios forlunar rocks are extremely low (Table 13). The typical valuefor anorthosites is 0.698949 (Warren, 1985). For lunar marebasalts the initial 87Sr/86Sr ratios range from0.69906 ± 0.00004 (Luna 16 basalts) to 0.69957 ± 0.0005(Apollo 12) (Papike et al., 1998). These ratios are close tothe estimated LUNI value of 0.69903 (Papanastassiou etal., 1970), derived from highland rock data. In general,the measured 87Sr/86Sr values in plagioclase from variouslunar rocks do not exceed 0.70000 (Compston et al.,1970; Papanastassiou et al., 1970; Snyder and Taylor,1994; Nyquist et al., 1996; Norman et al., 1998) and thelowest 87Sr/86Sr ratio ever measured in any terrestrial mate-

Table 7Chemical composition of Kalahari 009 (INAA and XRF) compared to other lunar VLT mare basalts and some Apollo and Luna 24 marebasalts

Element Lunar mare basalts Apollo and Luna basalts

Kalahari 009 (1) s.d. in % A-881757 (2) Y-793169 (3) Apollo 14 (4) Apollo 17 (5) Luna 24 (6)

wt%SiO2

a 46.2 5 47.1 43.7 46.8TiO2

a 0.45 5 2.45 1.52 2.45 0.70 1.1Al2O3

a 12.8 5 9.96 11.1 12.8 11.6 12.8FeO 15.8 5 22.0 21.23 16.4 15.5 21.1FeOa 16.0 5MgO 8.53 5 6.20 5.8 10.1 11.0 8.0MnO 0.27 10 0.34 0.32 0.26 0.25 0.25MnOa 0.27 5CaO 10.2 5 11.5 12.0 10.3 9.3 13.5CaOa 10.7 5Na2O 0.53 3 0.29 0.27 0.74 0.15 0.24K2O 0.19 5 0.038 0.062 0.10 0.014

ppmPa 110Sc 55 5 99.4 87 57.8 47 43Va 116 5 88 53 114 256 178Cr 2700 3 1887 1754 2780 6361 1642Cra 2674 5Co 26.5 3 27.9 21.4 27.2 30 38Ni <12 5 52 53 40 30 ± 20Ga 2.20 15 2.50 3.0As 0.63 10 0.065Br 0.94 15 0.11 0.16Rb 7.0 30 2.6 2.1 13Sr 110 8 115 220 50Zrb 14.46Nbb 0.675Cs 0.07 35 0.037 <0,05 0.36Ba 66 15 27 <81 175La 0.80 8 3.69 5.3 19.7 0.72 1.1Ce 2.2 10 10.9 15.6 53.8Nd 1.9 20 8.31 10.2 36Sm 0.56 15 2.88 4.4 10.2 0.55 1.2Eu 0.40 8 1.1 1.37 1.2 0.17 0.63Gd 0.70 30 3.86Tb 0.17 10 0.76 1.1 2.36 0.14 0.25Dy 1.0 40 5.3 7.5 1.5Ho 0.28 20 1.63Yb 1.2 5 3.26 4.8 6.81 0.82 1.0Lu 0.19 8 0.52 0.72 1.03 0.13 0.15Lub 0.2042Hf 0.34 5 2.2 3.08 8.13 0.49 0.9Hfb 0.4783Tab 0.0388Au 0.0017 15 0.002 0.05Th 0.09 20 0.42 0.75 1.52U 0.14 30 0.11 0.30

(1) This study; (2) Koeberl et al., 1993; (3) Warren and Kallemeyn, 1993; (4) Apollo 14 basalt from breccia 14321 (clast # DV-7; sample #1210), Shervais et al., 1985; (5) Apollo 17 VLT basalt (70006,371), Wentworth et al., 1979; (6) Luna 24 VLT basalt (24109,78), Ma et al., 1978.

a XRF analysis.b Isotope dilution method.


rial is as high as 0.700510 (early Archean barite from Pil-bara, Australia, McCulloch, 1994). Thus, the unradiogenic87Sr/86Sr ratios of feldspars from Kalahari 008(0.699826 ± 31) and Kalahari 009 (0.699242 ± 43) are an

unambiguous indication that the samples cannot be of ter-restrial origin and further support a lunar origin.

The geochemical twins Nb–Ta and Zr–Hf (Shannon,1976; Jochum et al., 1986) can fractionate in the silicate

Fig. 11. TiO2 versus FeO plot, showing the classification schemefor mare basalts from Neal and Taylor (1992). Kalahari 009 can beclassified as very-low-titanium basalt, similar to the Apollo 17 andLuna 24 basalts.

Fig. 12. A plot of FeO + MgO (total Fe as FeO in wt%) versusAl2O3 in lunar rocks. As a consequence of their simple mineralogy(plagioclase, pyroxene, olivine are the main mineral phases) alllunar rocks plot on a straight line connecting the anorthositichighland with the mafic mare basalt compositions. All data arefrom Basaltic Volcanism Study Project, 1981.

mare basaltsbasaltic brecciasmafic brecciasfeldspathic brecciasKREEP melt breccia

0 5 10 15 20 25 30 35Al2O3 (wt. %)

0.050.10.2

1

10

50

Th (p

pm)

Kalahari 009Kalahari 008

Fig. 13. Concentrations of Al2O3 and Th in lunar meteorites(modified after Korotev, 2005).


portions of terrestrial planets and form distinct, positivelycorrelated arrays for terrestrial and lunar rocks (Munkeret al., 2003). At a given Zr/Hf ratio, lunar rocks are dis-placed towards systematically higher Nb/Ta ratios. Fig. 9

shows that bulk Kalahari 008 plots exactly in the field of lu-nar rocks. Kalahari 009 plots slightly above the lunar fieldbut definitely outside the field of terrestrial samples.

One of the important geochemical indicators commonlyused to argue for lunar origin is the bulk Fe/Mn ratio inhighland rocks and mare basalts as well as Fe/Mn in olivineand pyroxene (Papike et al., 2003; Table 13). The bulk Fe/Mn ratio in Kalahari 008 of 74 is close to the value for typ-ical lunar highland rocks (Fig. 8 and Table 6). Although inFig. 8 Kalahari 009 plots between lunar and terrestrial bas-alts, the absolute Fe and Mn concentrations match thecomposition for lunar mare basalts (Table 7). The Fe/Mnratios in olivine from both Kalahari 008 and 009 plot alongthe trends found for lunar rocks, although the Fe/Mn ratiosin pyroxene show some scatter (Fig. 5).

Other element ratios that are characteristic of lunarrocks are pairs of highly incompatible elements with differ-ent volatilities such as K/U, Rb/U, Cs/U and Tl/U. Be-cause of the volatile-depleted nature of the Moon, theseratios are much lower in lunar than in terrestrial rocks(Fig. 19a and Table 13). However, these ratios are only sig-nificant in uncontaminated lunar rocks. Due to terrestrialweathering, the concentrations of some elements, e.g., Kand Rb can be strongly increased because of their relativelyhigh concentrations in the terrestrial surface environmentsand their highly mobile character in solutions. As shownin Fig. 19a the K/U and Rb/U ratios in Kalahari 008 and009 are much higher than the average ratios for lunar rocks,which is consistent with post-fall weathering, as shown inFig. 7. However, the Cs/U ratio is within the range of lunarrocks. Another characteristic feature of lunar rocks is theirlow P content. Since the elements P and Nd exhibit a similardegree of incompatibility on the Moon and Earth, the Pdepletion can be most easily recognized by comparing P/Nd ratios. The P/Nd ratio of Kalahari 008 of about 37 issignificantly below the terrestrial ratio of 50 (Weckwerthet al., 1983; Table 13). The P content of Kalahari 009 is ele-vated and the P/Nd ratio is 58, which is higher than the ter-restrial ratio but, again, terrestrial contamination cannot beexcluded. The source regions for mare basalts are depletedin the siderophile elements Ni and Ga. For a given MgOcontent, terrestrial volcanic rocks have a Ni/MgO that ishigher than that of mare basalts by a factor of �5 to 10(Ruzicka et al., 2001; Table 13 and Fig. 20). The Ga abun-dances in lunar basalts, for a given La content, are dis-placed to lower values by a factor of �5 (Fig. 19b;Ruzicka et al., 2001). Kalahari 009 shows a Ni and Gadepletion though only an upper limit for Ni is given in Ta-ble 7 and Fig. 20.

Each rock suite defines distinct compositional arrayscharacterized by positive correlations between the Cr andMg abundances (Table 13 and Fig. 19c). For a given Mgconcentration, terrestrial basalts have between one andtwo orders of magnitude lower Cr concentrations thanmare basalts. The lowest Mg/Cr ratios (or the highest Crcontents) in terrestrial volcanic rocks are observed in Ar-chean basalts, but they plot in the field defined by the lunarhighlands, well resolved from the ‘‘mare array”. In Fig. 19cKalahari 009 plots, together with lunar mare basalts on aline corresponding to a Mg/Cr ratio of 20, whereas Kalaha-

Table 8Gas concentrations (in 10�8 cm3STP/g) and isotope ratios of Kalahari 008 and 009

3He 4He 20Ne 36Ar 40Ar 20Ne/22Ne 21Ne/22Ne 36Ar/38Ar 84Kr 132Xe 129Xe/132Xe

Kalahari 008 0.03 ± .02 3 ± 1 27.3 ± 1.3 17.5 ± 1.5 920 ± 30 11.90 ± .20 0.037 ± .002 5.2 ± .2 0.06 ± .02 0.010 ± .005 0.97 ± .03Kalahari 009 60.02 61 0.14 ± .02 0.40 ± .06 340 ± 40 8.57 ± .20 0.188 ± .011 4.2 ± .8 0.7 ± .2 0.008 ± .002 1.01 ± .05

21Ne/22Ne0.0 0.2 0.4 0.6 0.8 1.0

20N

e/22

Ne

0

2

4

6

8

10

12

14 SWC

SEP

cosmogenic

Air

Kalahari 009

Kalahari 008

Fig. 14. Neon three-isotope diagram. Indicated are the solarcompositions SWC (solar wind) and SEP (solar energetic particles),the range of cosmogenic Ne, and the composition of terrestrialatmospheric Ne (AIR). Kalahari 008 shows solar gases but Ne inKalahari 009 is a mixture of atmospheric Ne and a small amount ofcosmogenic Ne.

10-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2DaG 262QUE 93069QUE 94281MAC 88104Y 791197ALHA 81005Y 82192DaG 400Dhofar 081Kalahari 008Kalahari 009

4He 20Ne 36Ar 84Kr 132Xe

Conc

entr

atio

n [c

m3 ST

P/g]

Fig. 15. Comparison of trapped solar noble gas concentrations inKalahari 008 and 009 with some other lunar meteorites (data fromBogard and Johnson, 1983; Ostertag et al., 1986; Bischoff et al.,1987, 1998; Eugster et al., 1991; Scherer et al., 1998; Greshake etal., 2001).

Table 9Cosmogenic radionuclide concentrations (dpm/kg) in Kalahari 008and 009

10Be 26Al 36Cl 41Ca

Kalahari008

0.021 ± 0.001 0.020 ± 0.006 0.022 ± 0.001 0.40 ± 0.37

Kalahari009

0.014 ± 0.001 0.014 ± 0.004 0.010± 0.002 0.03 ± 0.21

The quoted uncertainties are 1r AMS measurement errors.


ri 008, other highland rocks and terrestrial basalts have dis-tinctly higher ratios.

It has long been known that the abundances of V and Crcorrelate in lunar rocks (Laul et al., 1972; Wanke et al.,

1978). Terrestrial basalts show no correlation and an over-all enrichment in V compared to lunar mare basalts. Duringthe formation of lunar basalt V does not fractionatestrongly from Cr which is in contrast with terrestrial basaltswhich have similar V but much lower Cr contents (Ruzickaet al., 2001). Chromium versus V is shown in Fig. 19d whereis can be seen that Kalahari 009 plots in the field for marebasalts and again is clearly resolved from the terrestrialsamples.

It is a general feature of all lunar basalts that they have asignificantly higher FeO-content and lower Al2O3 than ter-restrial basalts (Table 13; e.g., Neal and Taylor, 1992; Ruz-icka et al., 2001). Concentrations of these elements inKalahari 009 (Table 7) are in the range of lunar basaltsand thus are further evidence for a lunar origin.

Cosmogenic radionuclide concentrations in Kalahari008 and 009 are extremely low. However, the results revealthat Kalahari 008 must have been exposed in space. For thecase of Kalahari 009, the concentrations are even lower andthe cosmogenic nuclides could have also been produced onthe Earth’s surface, without previous exposure in space.The arguments in favor of a lunar origin for Kalahari 008and 009 are summarized in Table 13. While no single lineof evidence is considered to be conclusive, considerationof all the data provide strong evidence for the lunar origin.

4.2. Chemical composition, formation history and possible

source regions

4.2.1. Kalahari 008

As a meteorite find, Kalahari 008 has been affected by ter-restrial weathering. Weathering of meteorites in hot desertenvironments may cause enrichment in the highly water-sol-uble elements such as K but also Sr, Ba, Ca, As, Br, Sb and Cs(Bischoff et al., 1998; Korotev et al., 2003). Thus, these ele-ments are not very useful to reconstruct the petrogenesis ofKalahari 008. However, in contrast to the very high K con-tent, the Sr, Ba and Cs concentrations in Kalahari 008 indi-cate a relatively small terrestrial contribution of theseelements. Additionally, the observed alteration products(mainly calcite deposits) are not expected to contain signifi-cant amounts of the ferromagnesian elements or REEs.

Table 10Lu–Hf isotope data

Sample ppm Lu ppm Hf 176Lu/177Hfa ±2r 176Hf/177Hf ±2r eHf(0)

Kalahari 009

Px-rich 0.2358 0.5005 0.06678 0.00012 0.285937 0.000021 110.7Opaques 0.3604 1.077 0.04741 0.00009 0.284320 0.000035 47.5Opaques 0.4336 3.488 0.01764 0.00004 0.281699 0.000100 �37.9Px-rich 0.1291 0.2136 0.08586 0.00043 0.287490 0.000020 166.9

WR 1 0.2042 0.4783 0.06058 0.00061 0.285241 0.000012 87.3WR 2 0.2038 0.4199 0.06894 0.00034 0.285894 0.000012 110.4Plag-rich 0.1219 0.2650 0.06511 0.00017 0.285475 0.000022 95.6

Kalahari 008

WR 0.09636 0.4699 0.02911 0.00016 0.282590 0.000010 �6.5

a Errors of 176Lu/177Hf are external errors including spike error magnification and uncertainties of Yb corrections.

0.279

0.281

0.283

0.285

0.287

0.289

0.00 0.02 0.04 0.06 0.08 0.10176Lu/177Hf

176 H

f/177 H

f

opaqueswhole-rock and plagioclase fractionspyroxene-richKalahari 008 (whole-rock)

Age = 4286 ± 95 Ma (2j )Initial 176Hf/177Hf =0.28037 ± 0.00013

MSWD = 1.4

Fig. 16. Lu–Hf isochron for Kalahari 009. Whole-rock and plagioclase fractions were excluded from regression calculations; error bars (2r)are smaller than the plotting symbols.

Table 11Summary of Ar–Ar age data for bulk, breccia and basalt fragmentsof Kalahari 009 obtained by IR-stepped heating

Sample Weight (mg) K (ppm)a Ca (%)a Age (Ga)

Bulk 27.67 585 6.3 1.787 ± 0.032Basalt 9.85 445 24.4 3.090 ± 0.180Breccia 3.11 244 51.0 —

Calculated ages show 2r errors and include errors for J-value.a K and Ca content reported were calculated based on the 39Ar

and 37Ar released during furnace step heating.


In common with DaG 262 and Dho 081, Kalahari 008seems to be depleted in Mg and Fe compared to the averagelunar highland meteorites and to average highland rocks.This relative depletion may reflect a lower fraction of maficphases and a correspondingly higher plagioclase compo-nent in these meteorites. Additionally, the Al concentrationis higher in these meteorites also indicating a higher plagio-clase component with Dhofar 081 displaying the highest Aland the lowest Fe and Mg contents. The low abundance of

mafic melt breccias found in Kalahari 008 explains the lowconcentrations of Mg and Fe and the high Al contents.

Despite similar Al, Fe and Mg contents in Kalahari 008and DaG 262, Kalahari 008 exhibits a higher Fe/Mg ratio(1.32, Table 6), similar to the value in Dho 081 (which isalso richer in Al2O3). The different ratios may reflect a dif-ferent composition and/or source of the mafic phases. Inparticular, the higher ratios may relate to higher Fe/Mg ra-tios in the mafic phases (e.g., Fe-rich olivine). Indeed, theolivine compositions in Kalahari 008 are notably richer inFe (Fa28–87) than those from DaG 262 (Fa20–71; Bischoffet al., 1998) and Dhofar 081 (Fa29–48; Bischoff, 2001). Oliv-ine from the Apollo 16 impact melt breccias ranges fromFa10 to Fa40 (Stoffler et al., 1985).

The generally low content of incompatible elements indi-cates that Kalahari 008 does not contain a KREEP-compo-nent. Fig. 21 is a plot of molar Mg/(Mg + Fe) versus the Ti/Sm ratio of Kalahari 008 and the major highland rocksuites (after Taylor, 1982). From this diagram it is clear thatKalahari 008 and other highland meteorites are distinctfrom the KREEP fields, with far higher Ti/Sm ratios thanany KREEP-bearing rocks.

Fig. 17. Apparent age and Ca/K versus K released for bulk Kalahari 009 during step heating.

Fig. 18. Apparent age and Ca/K versus K released during step heating for bulk, and breccia and basalt fractions obtained from the thick-section of Kalahari 009.

Table 12Rb–Sr analytical data for Kalahari 008 and 009

Sample 87Rb/86Sr 87Sr/86Sr

Kalahari 008

Residue 0.018 0.699826 ± 31Leachate 0.014 0.700781 ± 16

Kalahari 009

Residue 0.349 0.713974 ± 13Leachate 1 n.d. 0.709628 ± 134Leachate 2 0.037 0.702860 ± 34Leachate 3 n.d. 0.699242 ± 43

n.d., not determined.


According to Korotev (2001, 2005) the lunar regolith isa mixture of three major components: KREEP-rich materi-als, feldspathic materials and mare volcanics (Apollo andLuna mare basalts). As a result, all samples plot insidethe triangle defined by these components (Fig. 22). In Fig.22, Kalahari 008 together with Dar al Gani 262 and Dhofar081 plot within in the field of lunar highland rocks.

Although a few basaltic fragments were observed withinKalahari 008, these do not seem to contribute significantlyto the bulk composition.

In Fig. 23, the Sm and Al2O3 contents of Kalahari 008are compared to those of Dar al Gani 262, Dhofar 081and other lunar highland meteorites as well as to samplesfrom the Apollo 14 and 16 sites (Warren and Wassson,1980; Korotev, 1997; Cahill et al., 2004). The diagramshows that the lunar highland meteorites are depleted inSm compared to Apollo 14 and 16 regolith breccias. Thisobservation indicates that these meteorites are composi-tionally different from the Apollo and Luna highland sam-ples. Cahill et al. (2004) suggested that Dar al Gani 262 andDhofar 081 may represent a type of crustal terrane that dif-fers from the KREEPy impact melt breccias in the Apollocollection and that highland regions of the lunar farsideare likely candidate source regions.

4.2.2. Kalahari 009

General chemical and textural similarities of Kalahari009 with other mare rocks indicate a lunar origin, but in de-tail, the chemical composition and mineralogy of Kalahari

Table 13Comparative geochemistry for terrestrial and lunar rocks

Earth Moon Ref.

O isotopes D17O = 0 D17O = 0 (1)An in plagioclase 690 mostly >90 (2)Initial 87Sr/86Sr P0.700510 60.70000 (3 and 4)Age <4.0 Ga from �4.5 to �1.2 Ga (5 and 6)K/U 10,000 3000 (7)P/Nd 50 <50 (8)Bulk FeO/MnO 60 80 (9)Fe/Mn in pyroxene 40 62 (10)Fe/Mn in olivine 75 103 (10)Mg/Cr P60 20 (11)Nb/Ta At a given Zr/Hf ratio, lunar rocks are displaced towards systematically higher Nb/Ta (Fig. 9) (12)Volatile elementsK, Rb, Cs, Tl

The Moon is depleted in these elements compared to the Earth (by factor of �4–17; Fig. 19a). (11)

Ni/MgO in basalts For a given MgO content, terrestrial volcanic rocks have a Ni/MgO value that is higher thanthat of mare basalts (by a factor of �5 to 10; Fig. 20)

(11)

Cr/V in basalts <10 >10 (11)FeO in basalts 9–13 wt% 16–22 wt% (11)Al2O3 in basalts P15 wt% <15 wt% (13)Ga/La in basalts For a given La content, mare basalts are displaced to lower Ga abundances (by a factor of �5;

Fig. 19b)(11)

FeNi metal grains Absent in terrestrial basalts Common late-stage crystallization product in lunar basalts (14)

(1) Clayton (1993); (2) Ashwal (1993); (3) Norman et al. (1998); (4) McCulloch (1994); (5) Amelin et al. (1999); (6) Hiesinger et al. (2003), agebased on crater statistics; (7) Taylor (1984); (8) Weckwerth et al. (1983); (9) Basaltic Volcanism Study Project (1981); (10) Papike et al. (2003);(11) Ruzicka et al. (2001); (12) Munker et al. (2003); (13) Neal and Taylor (1992); (14) Jones and Palme (2000).

10-5

10-4

10-3

10-2

100

K/U Tl/U

terr. basalt

high-Ti mare basalt

Kalahari 009

low-Ti mare basalt

Kalahari 008

0.1 1 10 100La (ppm)

100

10

1

Ga

(ppm

)

terr. basalt

high-Ti mare basaltKalahari 009

low-Ti mare basalt

a b

1 10 100 1000

V (ppm)

10000

1000

100

10

1

Cr/V= 100

Cr/V= 40

Cr/V= 10

terr. basaltmare basaltKalahari 009

c

Cr(

ppm

)

Mg/Cr = 60

Mg/Cr = 20

CI

0 0.25 0.5 0.75 1 Cr (wt.%)

Mg

(wt.%

)

0

5

10

15

20

25

terr. basalthighlandmare basal tKalahari 009Kalahari 008

d

Cs/URb/U

Fig. 19. Comparative geochemistry of basalts from the Moon and the Earth: (a) average U- and CI-chondrite-normalized bulk K, Rb, Cs, Tland (b) Ga and La abundances in terrestrial and lunar mare basalts (Ruzicka et al., 2001). (c) Bulk-rock Mg versus Cr, showing the nearlyconstant Mg/Cr ratio for highlands and mare basalts. (d) Bivariant element diagram showing only minor fractionation of V/Cr in the marebasalts and a strong fractionation in terrestrial basalts; all data in (c) and (d) are from Basaltic Volcanism Study Project, 1981.


terr. basalt

high-Ti mare basaltKalahari 009

low-Ti mare basalt

MgO(wt. %)10 100

1

10

100

1000

Ni(

ppm

)

Fig. 20. Ni and MgO abundances in terrestrial and lunar basalts.Terrestrial and lunar basalts define arrays that are offset from oneanother. Only upper limit for the Ni concentration in Kalahari 009can be given (modified after Ruzicka et al., 2001).

40 60 80Mg#

10 2

10 3

10 4

anorthosites

dunite

spineltroctolite

troctolites + norites

Apollo 14, 15KREEP

noritic brecciasMAC 88105

Kalahari 008

Dag 262

ALHA 81005

Dho 81Ti/S

m

Fig. 21. Plot of mg number versus the Ti/Sm ratio for the bulk ofKalahari 008, Dhofar 081 (Dho 081), Dar al Gani 262 (Dag 262),MAC 88105 (Koeberl et al., 1991), other lunar highland meteorites(Koeberl et al., 1989), and the main lunar highland lithologies(from Koeberl et al., 1991; after Taylor, 1982).

Kalahari 008

Dho81

DaG 262

0 5 10 15 20 25FeO (wt. %)

Th (p

pm)

18

16

14

12

10

8

6

4

2

0

PKT

FHT

range of Apollo maficimpact-melt breccias

range ofApollo &Lunamare basalts

Fig. 22. Composition of lunar regolith represented by a mixture ofthree classes of terrane components: feldspathic highlands (FHT,feldspathic highlands terrane), mare basalts and volcanic glass andKREEP-bearing rocks (PKT, Procellarum-KREEP Terrane) (afterKorotev, 2005).

Apollo 14RegolithBreccias

Apollo 16RegolithBreccias

Far side Highlands Regolith?

Dh-081

Sm (p

pm)

15 20 25 30 35 0.1

1

10

100

DaG 262

Dh- 025

Kalahari 008

Al2O3 (wt. %)

Fig. 23. Whole-rock Sm versus Al2O3 in lunar meteorites (modifiedfrom Cahill et al., 2004).


009 are somewhat different from most lunar mare basalts.The question arises as to whether Kalahari 009 can beviewed as ‘‘real” mare basalt, implying a similar source re-gion and formation history as other mare basalts. SinceKalahari 008 and Kalahari 009 are very likely paired (seenext section) and since the origin of Kalahari 008 is unam-biguously the lunar highland one could also suspect Kala-hari 009 to be a highland rock. Based on bulk chemicalcomposition, Kalahari 009 indeed shows some similaritieswith the gabbronoritic subset of the high-Mg suite of pris-tine nonmare rocks. The Apollo 16 gabbro rocklet 61224,6first studied by Marvin and Warren (1980) is a quite goodanalog, although it has somewhat lower Fe content. Rocksof this suite have no KREEP-component and are free of ameteoritic component. The low Ni content of Kalahari 009and the absence of Ir (Table 7, indicating less than the

detection limit of 2 ppb Ir) suggest that Kalahari 009 doesnot have a meteoritic component and since there is no evi-dence for KREEP this rock must be considered pristine.The relatively high Au content of 1.7 ppb (Table 7) is mostlikely terrestrial contamination and may have nosignificance.

Mare basalts tend to have superchondritic Ca/Al ratioswhile highland rocks are more Al-rich with, in most cases,subchondritic Ca/Al ratios. Kalahari 009 has a chondriticCa/Al ratio and plots in a Ca/Al versus TiO2 diagram inthe field of highland rocks defined by Warren (2005) andmodified after Wood (1975) (Fig. 24). Furthermore, marebasalts and highland rocks have very different Mg/Cr ra-tios, having subchondritic and superchondritic ratios,respectively (Fig. 19c). The Mg/Cr ratio of 19.2 in Kalahari009 is significantly below the chondritic ratio, whereas mostgabbroic highland rocks have ratios exceeding the chon-dritic ratio. The Apollo 16 61224,6 gabbro with a similarbulk composition to Kalahari 009, has a Mg/Cr ratio of39 (Marvin and Warren, 1980). Ryder (1979) has definedthree components (anorthosites, norites, KREEP) inexplaining the chemistry of various highland rocks. His

Bulk-rock TiO2 (wt. %)

Kalahari 009

Apollo 11Apollo 12Apollo 14Apollo 15Apollo 16Apollo 17Luna 16Luna 24Nonmare

0.1 1 10 100

HighlandWood (1975)

Simpler, better? criterion

Bul

k-ro

ck C

a/A

l, m

olar

ratio

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Fig. 24. Bulk-rock TiO2 versus Ca/Al diagram for the distinction between mare and highland rocks (from Warren (2005); modified afterWood (1975)).


norite component representing pristine Mg-rich rocks has aMg/Cr ratio of 45. All highland rocks listed in Table 6 havesuperchondritic Mg/Cr ratios; whereas mare samples listedin Table 7 have subchondritic ratios.

The Ti/Sc ratio of Kalahari 009 is approximately chon-dritic (Fig. 25a), which is highly unusual for a lunar rock.Pristine highland rocks fall significantly below this ratio.The same is true for the Sc/Sm ratio (Fig. 25a). Superchon-dritic Sc/Sm ratios are extremely rare in lunar highlandrocks. Thus, both ratios demonstrate that Kalahari 009 isnot part of the pristine highland suite of rocks. The highand approximately chondritic Ti/Sm ratio indicates thatKalahari 009 could be a product of re-melting of a very

FAN

chondritic

Apollo14 mafic rocks

Fe- noritesMg- noritesMg- gabbronoritesgabbronoritestroctolites

Kalahari 009

0.1 1 10 100Sc/Sm

10000

1000

100

10

Ti/S

m

a

Fig. 25. (a) Ti/Sm versus Sc/Sm for Apollo 15 and 17 lunar highland rocket al. (1989), modified after Warren (1985)). (b) Anorthite content (Anhighland plutonic rocks and Kalahari 009 (modified after Warren et al.

early cumulate at a stage when ilmenite precipitation hadnot yet occurred. The rather high Sc abundance of Kalahari009 and its low REE contents support this hypothesis.Thus, in terms of trace elements Kalahari 009 is primitivelunar rock.

The comparatively high concentrations of K, Ba, Br,and possibly Sr, which are higher than observed in Apollosamples of similar compositions, may be due to terrestrialweathering.

The mineral chemistry of major phases in Kalahari 009is different from mineral compositions of major pristinerock groups as defined by Warren et al. (1983) and as dis-played in Fig. 25b. The ferromagnesian minerals are signif-

80 90 100An in plagioclase

30

40

50

60

70

80

90

100

AlkaliAnorthosites

Kalahari 009

Mg-suite

FAN

Mg#

in m

afic

min

eral

s

b

s, various Apollo 14 mafic rocks and Kalahari 009 (from Lindstrom) in plagioclase versus Mg# of coexisting mafic silicates for lunar(1983) and Lindstrom et al. (1989)). FAN, ferroan anorthosite.


icantly more FeO-rich than Mg-suite rocks at approxi-mately the same An-content of the plagioclase.

In summary, Kalahari 009 is different from all types ofknown lunar highland rocks in having no KREEP-compo-nent, no meteoritic component and is chemically and min-eralogically different from Mg-rich highland rocks.Despite strong affinities of Kalahari 009 to VLT mare bas-alts, this sample is quite distinct from this group of lunarrocks. It has lower chondrite-normalized REE abundancesthan the majority of mare basalts; and its Th concentration(0.09 ppm) is the lowest of all hitherto known mare basalts(Fig. 13). The Ga content of Kalahari 009 is lower than inany mare basalt shown in Fig. 19b. The positive Eu-anom-aly is unusual among lunar mare basalts and was previouslyobserved only in some small basaltic fragments from theLunar 24 collection (which may be not very representative)(Fig. 10).

The crystallization sequence of Kalahari 009 is similar tothat of mare basalts. Late-stage differentiates are repre-sented by fayalitic olivine, including the hedenberg-ite + fayalite + SiO2 symplectites. Symplectites have beenreported in other lunar samples and are commonly,although not universally, interpreted as the breakdownproducts of metastable pyroxferroite to fayalite and silica.During crystallization of a mafic silicate liquid that hasevolved to compositions with high Fe-content, the crystal-lizing pyroxenes may become zoned toward a Fe-rich end-member. During subsequent cooling this pyroxene canbreak down to the stable assemblage hedenbergite + faya-lite + SiO2 (Aramovich et al., 2002), as observed in thissample. Pyroxene compositions for Kalahari 009 are shownin Fig. 3. They show strong variation and follow a typicalmare basalt fractionation trend. Pyroxenes with composi-tions in the miscibility gap of the pyroxene quadrilateral(Fig. 3) and zoning of the pyroxene indicate fast cooling.The complete breakdown of the pyroxferroite, the exsolu-tion lamellae in pyroxene and the coarse-grained natureof the basalt indicate slow cooling. Thus, an early fast cool-ing episode (where pyroxenes zonation formed) may havebeen followed by a stage of slow cooling, where the pyrox-ferroite breakdown occurred and the exsolution lamellaewere produced. One possible scenario is crystallizationand cooling in the center of a thick lava flow.

The Lu–Hf data yield an age of 4.286 ± 0.095 Ga. With-in the analytical uncertainties this age is identical to the in-situ U–Pb phosphate age of 4.35 ± 0.15 Ga reported byTerada et al. (2007) and the Sm–Nd age of4.30 ± 0.05 Ga (Shih et al., 2008). While ages for Mg-suiterocks range from 4.53 ± 0.29 to 4.11 ± 0.02 (Stoffler et al.,2006), the majority of known mare-basalt samples havecrystallization ages younger than 3.9 Ga, leading to the sug-gestion that mare volcanism occurred mainly after the lateheavy bombardment that ended at 3.9 Ga. However, Tay-lor et al. (1983) found ancient mare basalt clasts in theFra Mauro breccias and proposed that mare volcanism be-gan in the Apollo 14 region at least as early as 4.2 Ga ago.Additionally, Dasch et al. (1987) reported Rb–Sr isochronages for Apollo 14 mare basalts ranging from 3.96 up to4.33 ± 0.13 Ga, also indicating pre-4.1 Ga (pre-Nectarian)mare basalt volcanism. Remote sensing and photo-geologic

evidence also indicates the onset of mare volcanism prior tothe termination of late heavy bombardment (Hartmann etal., 1981). According to Hartmann and Wood (1971) thereare a number of occurrences of intermediate-albedo high-land plains, that are thought to represent early mare basaltdeposits mixed with and covered by crater ejecta of high-land crust. Head and Wilson (1992) designated these hiddendeposits as ‘‘cryptomare”. Taking the Lu–Hf, U–Pb andSm–Nd ages as the crystallization ages of Kalahari 009,then VLT-type magma was being emplaced at the lunar sur-face as early as �4.3 Ga ago. If Kalahari 009 is interpretedas a VLT mare basalt, it is one of the oldest known marebasaltic samples and as such provides definitive evidencefor lunar mare volcanism prior to 4.1 Ga. Additionally,Hawke et al. (2005) suggested that basaltic deposits fromcryptomaria tend to be VLT, which is consistent with Kal-ahari 009. Subsequently, Terada et al. (2007) have also sug-gested that Kalahari 009 derives from a cryptomare on theMoon. The formation of exsolution lamellae in pyroxenesdescribed above support the hypothesis. These lamellaeare coarser than usual for Apollo and Luna mare basaltsand were discussed in the YQ suite of launch-paired marebasaltic meteorites (Queen Alexandra Range 94281, Yama-to 793274, EET 87521/96008) by Arai and Warren (1999).These authors interpret such exsolution features as formingduring annealing beneath a blanket of warm impact-debris,i.e., in a cryptomare.

The much younger age of 1.79 Ga obtained from the40Ar–39Ar data suggests that the K–Ar system was totallyreset by an impact event at �1.79 Ga. Total resetting ofthe K–Ar clock usually requires temperatures of <900 �Cfor a prolonged period of time (higher temperatures wouldstart melting the rock), possibly during burial under ejectablankets (Stephan and Jessberger, 1992).

The highly radiogenic initial 176Hf/177Hf ratio (initialeHf of +12.9 ± 4.6) of Kalahari 009 indicates that the ba-salt formed from re-melting of a mantle source that was de-pleted in incompatible trace elements early in the lunarhistory. This depletion is consistent with the low REE con-centrations as well as the low content of other incompatibleelements such as Th and Ti. To evolve an initial eHf of+12.9 ± 4.6 requires a minimal source 176Lu/177Hf of 0.09to 0.16 (assuming that the source had a chondritic initial176Hf/177Hf ratio starting at 4.50 Ga). Similar to most otherlunar basalts, the Lu/Hf ratio measured in Kalahari 009(176Lu/177Hf �0.06) is much lower than the Lu/Hf ratio cal-culated for its source. To account for this low Lu/Hf ratioof the basalt relative to the high Lu/Hf ratio required for itssource, a phase in the residuum is required that selectivelyretains Lu relative to Hf. Beard et al. (1998) argued thatmelting of an initially garnet-bearing source, with garnetbeing consumed as the melting regime moves to shallowerlevels can produce the required Lu/Hf fractionation ob-served in most low-Ti basalts. Though the presence of gar-net in the lunar mantle is controversial, the results in thisstudy may provide further evidence for suggestion.

Although lower than the time-integrated source ratio,the Lu/Hf ratio in Kalahari 009 is still extremely high whencompared to other lunar mare basalts (176Lu/177Hf < 0.03,e.g., Patchett and Tatsumoto, 1981; Unruh et al., 1984;


Beard et al., 1998). Hence, the unusually depleted composi-tion points toward an yet unsampled basalt source region inthe lunar interior. Possibly, Kalahari 009 constitutes a newdepleted endmember of low-Ti mare basalt volcanism. Thisindicates that the Moon became chemically very heteroge-neous at an early stage in its history and that differentcumulate sources that crystallized at various stages duringthe primary lunar differentiation are responsible for the di-verse mare basalt types.

The paucity of noble gases that form by irradiation onthe lunar surface suggests that Kalahari 009 was excavatedfrom deeper regions in the regolith and was thereforeshielded from cosmic rays. This is consistent with the factthat no regolith component or highland material was ob-served in the studied thin-sections and also supports theidea that Kalahari 009 may derive from below the surfacelayer (e.g., cryptomare). In combination with the very lowcosmogenic exposure age the observation distinguishes Kal-ahari 009 from other lunar meteorites that have much olderejection ages from the Moon as well as longer exposures tocosmic rays on the lunar surface.

Remote sensing data (Clementine and Lunar Prospec-tor) indicate that the nearside and the farside of the Moonhave substantially different chemical compositions andlithologies (e.g., Lawerence et al., 1998; Joliff et al., 2000).Several incompatible trace elements (e.g., Th, U) are con-centrated within and around the Procellarum-KREEP Ter-rane (PKT), the origin of most returned Apollo samples.Large regions of the lunar surface, especially the northernfarside highlands are low in Th. The very low Th andREE content of Kalahari 009 (lower than Luna 24 basalts)points to a possible source far from the influence of PKT. Itis possible that this lunar basalt may come from a source onthe farside, far from the PKT.

4.3. Pairing of Kalahari 008 and 009

Kalahari 008 and 009 were found about 50 meters apart.Although they represent different rock types (anorthositicbreccia versus basalt) it is conceivable that they belong toone meteorite fall. It is very unlikely to find two unrelatedlunar meteorites this close together. The compositions ofolivine in both samples also indicate certain similarities.Most olivine in the basaltic rock Kalahari 009 has Fa-con-tents above 80 mol % (Fig. 2d). Such high Fe-concentra-tions in olivine are uncommon in anorthositic highlandbreccias (e.g., Dar al Gani 262 (up to Fa71; Bischoff et al.,1998), Dhofar 081 (up to Fa46; Bischoff, 2001)). As shownin Fig. 2b, the olivine compositions of the anorthositic brec-cia Kalahari 008 show a bimodal distribution. Most olivinehas Fa-contents up to about 70 mol % (with a peak atabout 46 mol % Fa), but several analyzed olivine grainshave significantly higher Fa-contents of about 90 mol %(Fig. 2). It is suggested that the latter olivine derives fromfragments within the anorthositic breccia that were origi-nally derived from fragmented basalts (like Kalahari 009).These fragments must have been incorporated into the Kal-ahari 008 breccia during impact-induced fragmentation,mixing, and re-lithification on the lunar surface. This mayindicate that although Kalahari 008 and 009 represent com-

pletely different lithologies, the Kalahari 008 breccia wasformed close to the Kalahari 009 basalt and thus containsfragments derived from the basalt. In addition, the observa-tion that both rocks may have had identical initial Hf iso-tope composition points to a formation in closeproximity, i.e., in the same reservoir. Hence, both rocksmay have been ejected from the Moon as one polymictmeteoroid.

The concentrations of cosmogenic radionuclides and no-ble gases are unusually low in both rocks, although Kalaha-ri 008 contains slightly higher concentrations. A likelyreason for this difference is that the basaltic portion ofthe meteoroid has never been exposed to the solar wind,whereas at least parts of the anorthositic breccia have beenbriefly exposed at the lunar surface prior to lithification andejection of these very different lithologies. However, bothrocks have the shortest exposure ages of any meteorite everanalyzed. This similarity provides additional support thatthese two meteorites are fragments of the same fall andbroke apart during transit through the terrestrial atmo-sphere or at the find site.

5. CONCLUSIONS

Kalahari 008 is an anorthositic breccia that shares manysimilarities with other highland breccias. Despite an overallsimilarity in chemical composition this meteorite is uniquebecause of its deficiency of solar wind-derived rare gasesthat are prominent in most other lunar highland breccias.Regolith components such as glass spherules and otherglasses are also rare. Thus, Kalahari 008 represents a rela-tively immature highland breccia.

Although, in terms of bulk-rock major element compo-sition, Kalahari 009 shares similarities with some Apolloaluminous mare basalts as well as the gabbronoritic subsetof the high-Mg suite of pristine nonmare rocks, it has dis-tinctive mineralogical and geochemical characteristics thathave not been previously documented in any known lunarbasalt. The very old age, the low trapped noble gas contentand short exposure ages as well as the exceptionally low Thconcentration, and the highly radiogenic initial 176Hf/177Hfratio makes Kalahari 009 a unique and important, if com-plicated, lunar sample. The Lu–Hf isotope systematics ar-gue for the presence of highly depleted reservoirs early inlunar history and indicates that the Moon became chemi-cally very heterogeneous at an early stage in its historyand that different cumulate sources are responsible for thediverse basalt types.

The data obtained in this study indicate that the basaltsolidified �4.3 Ga ago and thus represents an ancient marebasalt deposits (e.g., cryptomare). It experienced at leastone major impact event at 1.7–2.3 Ga resulting in breccia-tion and resetting of the K–Ar system. This impact didnot expose the meteorite at the lunar surface and it re-mained shielded from galactic cosmic rays and solar windparticles until it was ejected and transited to the Earth�230 yr ago. While Kalahari 008 and 009 may be paired,Kalahari 009 probably represents basalt that comes froma locality near a mare–highland boundary region. Sincecryptomare are covered by highland ejecta, it is possible


that these rocks are from the boundary area, where basaltdeposits are covered by highland ejecta.

Using Clementine and Lunar Prospector data, Giguereet al. (2003) and Hawke et al. (2005) report that cryptoma-ria in the Lomonosov-Fleming basin on the lunar farsideare very low to intermediate titanium basalts, consistentwith Kalahari 009. The very low Th and REE content ofKalahari 009 as well as the depletion in Sm and the lackof KREEPy signatures in Kalahari 008 also point to a pos-sible source far from the influence of the Procellarum-KREEP Terrane, possibly the lunar farside.

The discovery of Kalahari 008 and especially Kalahari009, with its unique character, suggests that there is a greatpossibility for further unsampled lithologies on the Moon.Sampling of these rocks and their study will extend theunderstanding of geologic evolution of the Moon. Ancientcryptomaria are therefore considered to be worthy targetsof future sample return missions.

ACKNOWLEDGMENTS

This project was funded by the German Research Foundation(DFG) as part of the priority program ‘‘Mars and the TerrestrialPlanets” (SPP 1115). V.A.F. was funded by the Fundac�ao para aCiencia e a Tecnologia, Portugal. V.A.F. and R.B. were addition-ally funded by Particle Physics and Astronomy Research Counciland the Royal Society, UK. R.N.C. was funded by NASA Grant# NNG05GG86G. The writing of this paper has benefited fromNASA ADS Proceedings Query Service at http://adsabs.har-vard.edu/proceedings_service.html. We greatly appreciate the con-structive reviews by P. Warren and M. Anand, as well as commentsby the Associate editor C. Koeberl. We thank Erik Scherer for per-mission to use his unpublished protocol for high precision Lu con-centration measurements.

REFERENCES

Amelin Y., Lee D.-C., Halliday A. N. and Pidgeon R. T. (1999)Nature of the Earth’s earliest crust from hafnium isotopes insingle detrital zircons. Nature 399, 252–255.

Arai T. and Warren P. H. (1999) Lunar meteorite Queen AlexandraRange 94281: glass compositions and other evidence for launchpairing with Yamato 793274. Meteorit. Planet. Sci. 34, 209–

234.

Aramovich C. J., Herd C. D. and Papike J. J. (2002) Symplectitesderived from metastable phases in martian basaltic meteorites.Am. Mineral. 87, 1351–1359.

Armstrong J. T. (1991) Quantitative elemental analysis of individ-ual microparticles with electron beam instruments. In Electron

Probe Quantitation (eds. K. F. J. Heinrich and D. E. Newbury).Plenum Press, pp. 261–315.

Ashwal L. D. (1993) Anorthosites. Springer, Berlin.Basaltic Volcanism Study Project (1981) Basaltic Volcanism on the

Terrestrial Planets. Pergamon Press, Inc., New York.Beard B. L., Taylor L. A., Scherer E. E., Johnson C. M. and

Snyder G. A. (1998) The source region and melting mineralogyof high-titanium and low-titanium lunar basalts deduced fromLu–Hf isotope data. Geochim. Cosmochim. Acta 62, 525–544.

Bischoff A. (2001) Fantastic new chondrites, achondrites, and lunarmeteorites as the result of recent meteorite search expeditions inhot and cold deserts. Earth Moon Planets 85–86, 87–97.

Bischoff A., Palme H., Weber H. W., Stoffler D., Braun O., SpettelB., Begemann F., Wanke H. and Ostertag R. (1987) Petrogra-

phy, shock history, chemical composition and noble gas contentof the lunar meteorites Y-82192 and Y-82193. Mem. Nat. Inst.

Polar Res. Spec. Issue 46, 21–42.

Bischoff A., Weber D., Clayton R. N., Faestermann T., Franchi I.A., Herpers U., Knie K., Korschinek G., Kubik P. W., MayedaT. K., Merchel S., Michel R., Neumann S., Palme H., PillingerC. T., Schultz L., Sexton A. S., Spettel B., Verchovsky A. B.,Weber H. W., Weckwerth G. and Wolf D. (1998) Petrology,chemistry, and isotopic compositions of the lunar highlandregolith breccia Dar al Gani 262. Meteorit. Planet. Sci. 33,

1243–1257.

Bogard D. D. and Johnson P. (1983) Trapped noble gases indicatelunar origin for antarctic meteorite. Geophys. Res. Lett. 10,

801–803.

Cahill J. T., Floss C., Anand M., Taylor L. A., Nazarov M. A. andCohen B. A. (2004) Petrogenesis of lunar highlands meteorites:Dhofar 025, Dhofar 081, Dar al Gani 262, and Dar al Gani 400.Meteorit. Planet. Sci. 39, 503–529.

Clayton R. N. (1993) Oxygen isotopes in meteorites. Annu. Rev.

Earth Planet. Sci. 21, 115–149.

Clayton R. N. and Mayeda T. K. (1983) Oxygen isotopes ineucrites, shergottites, nakhlites, and chassignites. Earth Planet.

Sci. Lett. 62, 1–6.

Compston W., Arriens P. A. and Vernon M. J. (1970) Rubidium–strontium chronology and chemistry of lunar material. Science

167, 474–476.

Dasch E. J., Shih C.-Y., Bansal B. M., Wiesmann H. and NyquistL. E. (1987) Isotopic analysis of basaltic fragments from lunarbreccia 14321: chronology and petrogenesis of pre-imbriummare volcanism. Geochim. Cosmochim. Acta 51, 3241–3254.

Davis J. C., Proctor I. D., Southon J. R., Caffee M. W., HeikkinenD. W., Roberts M. L., Moore T. L., Turteltaub K. W., NelsonD. E., Loyd D. H. and Vogel J. S. (1990) LLNL/UC AMSfacility and research program. Nucl. Instrum. Methods Phys.

Res. B52, 269–272.

Eugster O., Beer J., Burger M., Finkel R. C., Hofmann H. J.,Krahenbuhl U., Michel Th., Synal H. A. and Wolfli W. (1991)History of the paired lunar meteorites MAC88104 andMAC88105 derived from noble gas isotopes, radionuclides,and some chemical abundances. Geochim. Cosmochim. Acta 55,

3139–3148.

Fagan T. J., Talor G. J., Keil K., Killgore T. L., Hicks M., BunchT. E., Mittlefehld J. H., Wittke D. W., Clayton R. N., MayedaT. K., Eugster O., Lorenzetti S. and Norman M. D. (2003)Northwest Africa 773: lunar origin and iron-enrichment trend.Meteorit. Planet. Sci. 38, 529–554.

Fernandes V. A. and Burgess R. (2005) Volcanism in marefecunditatis and mare crisium: Ar–Ar studies. Geochim. Cos-

mochim. Acta 69, 4919–4934.

Fernandes V. A., Burgess R., Bischoff A. and Sokol A. K. (2006)Lunar volcanism during the Erastothenian I: Kalahari 009.Annual Meeting of the Meteoritical Society 69th, #5297(abstract).

Fernandes V. A., Burgess R., Bischoff A., Sokol A. K. and HalodaJ. (2007) Kalahari 009 and North East Africa 003: Young(<2.5 Ga) lunar mare basalts. Lunar Planet. Sci. XXXVIII.Lunar Planet. Inst. Houston. #1611 (abstr.).

Giguere T. A., Hawke B. R., Blewett D. T., Bussey D. B. J., LuceyP. G., Smith G. A., Spudis P. D. and Taylor G. J. (2003)Remote sensing studies of the Lomonosov-Fleming region ofthe Moon. J. Geophys. Res. 108. doi:10.1029/2003JE002069.

Grady M. M. (2000) Catalogue of Meteorites, fifth edition.Cambridge University Press, Cambridge.

Greshake A., Schmitt R. T., Stoffler D., Patsch M. and Schultz L.(2001) Dhofar 081: a new lunar highland meteorite. Meteorit.

Planet. Sci. 36, 459–470.

http://adsabs.harvard.edu/proceedings_service.html

http://adsabs.harvard.edu/proceedings_service.html

http://dx.doi.org/10.1029/2003JE002069


Hartmann W. K. and Wood C. A. (1971) Moon: origin andevolution of multiring basins. The Moon 3, 3–78.

Hartmann W. K., Strom R., Weidenschilling S., Blasius K.,Woronow A., Dence M., Grieve R., Diaz J., Chapman C.,Shoemaker E. and Jones K. (1981) Chronology of planetaryvolcanism by comparative studies of planetary cratering. InBasaltic Volcanism on the Terrestrial Planets. Pergamon Press,

pp. 1049–1127.

Hawke B. R., Gillis J. J., Giguere T. A., Blewett D. T., LawrenceD. J., Lucey P. G., Smith G. A., Spudis P. D. and Taylor G. J.(2005) Remote sensing and geologic studies of the Balmer-Kapteyn region of the Moon. J. Geophys. Res. 110. doi:10.1029/

2004JE002383.

Head, Jr., J. W. and Wilson L. (1992) Lunar mare volcanism-Stratigraphy, eruption conditions, and the evolution of second-ary crust. Geochim. Cosmochim. Acta 56, 2155–2175.

Hiesinger H., Head, Jr., J. W., Wolf U., Jaumann R. and NeukumG. (2003) Ages and stratigraphy of mare basalts in oceanusprocellarum, mare nubium, mare cognitum, and mare insula-rum. J. Geophys. Res. 108, 5065. doi:10.1029/2002JE001985.

Joliff B. L., Gillis J. J., Haskin L. A., Korotev R. L. and WieczorekM. A. (2000) Major lunar crustal terranes: surface expressionsand crust–mantle origins. J. Geophys. Res. 105, 4197–4216.

Joliff B. L., Korotev R. L., Zeigler R. A. and Floss C. (2003)Northwest Africa 773: lunar mare breccia with a shallow-formed olivine cumulate component, inferred very-low-Ti(VLT) heritage, and a KREEP connection. Geochim. Cosmo-

chim. Acta 67, 4857–4879.

Jochum K. P., Seufert H. M., Spettel B. and Palme H. (1986) Thesolar-system abundances of Nb, Ta, and Y, and the relativeabundances of refractory lithophile elements in differentiatedplanetary bodies. Geochim. Cosmochim. Acta 50, 1173–1183.

Jones J. and Palme H. (2000) Geochemical constraints on theorigin of the Earth and Moon. In Origin of the Earth and Moon

(eds. R. M. Canup and K. Righter). Lunar Planet Inst.,

Houston, pp. 197–216.

Koblitz J. (2003) MetBase, Meteorite Data Retrieval Program,version 6.0.

Koeberl C., Warren P. H., Lindstrom M. M., Spettel B. andFukuoka T. (1989) Preliminary examination of the Yamato-86032 lunar meteorite: II major and trace element chemistry.Proc. NIPR Symp. Ant. Met. 2, 15–24.

Koeberl C., Kurat G. and Brandstatter F. (1991) MAC88105-Aregolith breccia from the lunar highlands: mineralogical,petrological, and geochemical studies. Geochim. Cosmochim.

Acta 55, 3073–3087.

Koeberl C., Kurat G. and Brandstatter F. (1993) Gabbroic lunarmare meteorites Asuka-881757 (Asuka-31) and Yamato-793169: geochemical and mineralogical study. Proc. NIPR

Symp. Ant. Met. 17, 14.

Korotev R. L. (1997) Some things we can infer about the Moonfrom the composition of the Apollo 16 regolith. Meteorit.

Planet. Sci. 32, 447–478.

Korotev R. L. (2001) On the systematics of lunar regolithcompositions. Lunar Planet. Sci. XXXII. Lunar Planet. Inst.Houston. #1134 (abstr.).

Korotev R. L. (2005) Lunar geochemistry as told by lunarmeteorites. Chem. Erde 65, 297–346.

Korotev R. L., Jolliff B. L., Zeigler R. A., Gillis J. J. and Haskin L.A. (2003) Feldspathic lunar meteorites and their implicationsfor compositional remote sensing of the lunar surface and thecomposition of the lunar crust. Geochim. Cosmochim. Acta 67,

4895–4923.

Kruse H. and Spettel B. (1982) A combined set of automatic andinteractive programs for instrumental neutron activation anal-ysis. J. Radioanal. Chem. 70, 427–434.

Laul J. C., Wakita H., Showalter D. L., Boynton W. V. andSchmitt R. A. (1972) Bulk, rare earth, and other trace elementsin Apollo 14 and 15 and Luna 16 samples. Proc. Lunar Sci.

Conf. 3rd, 1181–1200.

Lawerence D. J., Feldman W. C., Barraclough B. L., Binder A. B.,Elphic R. C., Maurice S. and Thomsen D. R. (1998) Globalelemental maps of the Moon: the Lunar Prospector gamma-rayspectrometer. Science 281, 1484–1489.

Lindstrom M. M., Marvin U. B. and Mittlefehld D. W. (1989)Apollo 15 Mg- and Fe-norites: a redefinition of the Mg-Suitedifferentiation trend. Proc. Lunar Sci. Conf. 19th, 245–254.

Ludwig K. R. (2001) ISOPLOT/Ex rev. 2.49. United StatesGeological Survey.

Ma M.-S., Schmitt R. A., Taylor G. J., Warner R. D., Lange D. E.and Keil K. (1978) Chemistry and petrology of Luna 24 lithicfragments and <250 lm soils: constraints on the origin of VLTmare basalts. In Mare Crisium: The View from Luna 24 (eds. R.B. Merill and J. J. Papike). Pergamon Press, New York, pp.

569–592.

Marvin U. and Warren P. (1980) A pristine eucrite-like gabbrofrom Decartes and its exotic kindred. Proc. Lunar Planet. Sci.

Conf. 11th, 507–521.

McCulloch M. T. (1994) Primitive 87Sr/86Sr from an Archeanbarite and conjecture on the Earth’s age and origin. Earth

Planet. Sci. Lett. 126, 1–13.

Munker C., Weyer S., Scherer E. and Mezger K. (2001) Separationof high field strength elements (Nb, Ta, Zr, Hf) and Lu fromrock samples for MC-ICPMS measurements. Geochem. Geo-

phys. Geosyst. 2, 10.1029. doi:10.1029/2001GC000183.

Munker C., Pfander J. A., Weyer S., Buchl A., Kleine T. andMezger K. (2003) Evolution of planetary cores and the Earth–Moon system from Nb/Ta systematics. Science 301, 84–87.

Neal C. R. and Taylor L. A. (1992) Petrogenesis of mare basalts: arecord of lunar volcanism. Geochim. Cosmochim. Acta 56,

2177–2211.

Nebel O., Mezger K., Scherer E. E. and Munker C. (2005) Highprecision determinations of 87Rb/85Rb in geologic materials byMC-ICP-MS. Int. J. Mass Spectrom. 246, 10–18.

Nishiizumi K., Okazaki R., Park J., Nagao K., Masarik J. andFinkel R. C. (2002) Exposure and terrestrial histories of Dhofar019 Martian meteorite. Lunar Planet. Sci. XXVIII. LunarPlanet. Inst. Houston. #1366.

Nishiizumi K., Welten K. C. and Bischoff A. (2005) Kalahari 008/009—the shortest exposure age of all meteorites. Meteorit.

Planet. Sci. 40, A113.

Norman M., Borg L., Nyquist L., Bogard D., Snyder G., Taylor L.and Lindstrom M. (1998) Lunar Planet. Sci. XXVIIII. LunarPlanet. Inst. Houston. #1551 (abstr.).

Nyquist L. E., Wiesmann H., Shih C.-Y. and Dasch J. (1996)Lunar meteorites and the lunar crustal Sr and Nd isotopiccompositions. Lunar Planet. Sci. XXVII, 971–972.

Oba T. and Kobayashi Y. (2001) The mineral assemblage ofsymplectites in lunar meteorite Asuka-881757. Antarct. Mete-

orit. Res. 14, 21–27.

Ostertag R., Stoffler D., Bischoff A., Palme H., Schultz L.,Spettel B., Weber H., Weckwerth G. and Wanke H. (1986)Lunar meteorite Yamato-791197; petrography, shock historyand chemical composition. Mem. Nat. Inst. Polar Res. 41, 17–

44.

Palme H., Spettel B., Jochum K. P., Dreibus G., Weber H.,Weckwerth G., Wanke H., Bischoff A. and Stoffler D. (1991)Lunar highland meteorites and the composition of the lunarcrust. Geochim. Cosmochim. Acta 55, 3105–3122.

Papanastassiou D. A., Wasserburg G. J. and Burnett D. S. (1970)Rb–Sr ages of lunar rocks from the Sea of Tranquillity. Earth


http://dx.doi.org/10.1029/2004JE002383

http://dx.doi.org/10.1029/2004JE002383

http://dx.doi.org/10.1029/2002JE001985

http://dx.doi.org/10.1029/2001GC000183


Papike J. J., Ryder G. and Shearer C. K. (1998) Lunar samples. InPlanetary Materials, vol. 36 (ed. J. J. Papike). Mineralogical

Society of America, Washington, DC.

Papike J. J., Karner J. M. and Shearer C. K. (2003) Determinationof planetary basalt parentage: a simple technique using theelectron microprobe. Am. Mineral. 88, 469–472.

Patchett P. J. (1983) Importance of the Lu–Hf isotopic system instudies of planetary chronology and chemical evolution. Geo-

chim. Cosmochim. Acta 47, 81–91.

Patchett P. J. and Tatsumoto M. (1981) The hafniumisotopic evolution of lunar basalts. Lunar Planet. Sci. XII,819–821.

Ruzicka A., Snyder G. A. and Taylor L. A. (2001) Comparativegeochemistry of basalts from the moon, earth, HED asteroid,and Mars: implications for the origin of the moon. Geochim.

Cosmochim. Acta 65, 979–997.

Ryder G. (1979) The chemical components of highlands breccias.Proc. Lunar Planet. Sci. Conf. 10th, 561–581.

Scherer E., Munker C. and Mezger K. (2001) Calibration of thelutetium–hafnium clock. Science 293, 683–686.

Scherer P., Patsch M. and Schultz L. (1998) Noble gas study of thenew lunar highland meteorite Dar al Gani 400. Meteorit.

Planet. Sci. 33, A135–A136 (abstract).

Scherer P. and Schultz L. (2000) Noble gas record, collision historyand pairing of CV, CO, CK and other carbonaceous chondrites.Meteorit. Planet. Sci. 35, 135–154.

Schulz T., Sokol A., Palme H., Weckwerth G., Munker C. andBischoff A. (2007) Chemical composition and Lu/Hf-age of theLunar mare basalt meteorite Kalahari 009. Annual Meeting of

the Meteoritical Society 70th, #5151 (abstract).Shannon R. D. (1976) Revised effective ionic-radii and systematic

studies of interatomic distances in halides and chalcogenides.Acta Crystallogr. A32, 751–767.

Shervais J. W., Taylor L. A. and Lindstrom M. M. (1985) Apollo14 mare basalts—petrology and geochemistry of clasts fromconsortium breccia 14321. Proc. Lunar Planet. Sci. Conf. 15th J.

Geophys. Res. 90, C375–C395.

Shih C.-Y., Nyquist L. E., Reese Y. and Bischoff A. (2008) Sm–Ndand Rb–Sr isotopic studies of meteorite Kalahari 009: an oldVLT mare basalt. Lunar Planet. Sci. XXXIX. Lunar Planet.Inst. Houston. #2165 (abstr.).

Snyder G. A. and Taylor L. A. (1994) Rb–Sr Isotopic systematicsof lunar ferroan anorthosite 62237. Lunar Planet. Sci. Conf.

25th, 1309–1310.

Soderlund U., Patchett P. J., Vervoort J. D. and Clark E. I. (2004)The 176Lu decay constant determined by Lu–Hf and U–Pbisotope systematics of Precambrian mafic intrusions. Earth


Sokol A. K. and Bischoff A. (2005a) Meteorites from Botswana.Meteorit. Planet. Sci. 40, A177–A184.

Sokol A. K. and Bischoff A. (2005b) Mineralogy of the lunarmeteorites Kalahari 008 and Kalahari 009. Annual Meeting of

the Meteoritical Society 68th, #5059 (abstract).Steiger R. H. and Jager E. (1977) Subcommission on geochronol-

ogy: convention on the use of decay constants in geo- andcosmochemistry. Earth Planet. Sci. Lett. 36, 359–362.

Stelzner T. H., Heide K., Bischoff A., Weber D., Schultz L., HappelM., schron W., Neuper U., Michel R., Clayton R. N., MayedaT. K., Bonani G., Haidas I., Ivy-Ochs S. and Suter M. (1999)An interdisciplinary study of weathering effects in ordinarychondrites from the Acfer region, Algeria. Meteorit. Planet. Sci.

34, 787–794.

Stephan T. and Jessberger E. K. (1992) Isotope systematics andshock-wave metamorphism: III K–Ar in experimentally andnaturally shocked rocks; the Haughton impact structure,Canada. Geochim. Cosmochim. Acta 56, 1591–1605.

Stoffler D., Bischoff A., Borchardt R., Burghele A., Deutsch A.,Jessberger E. K., Ostertag R., Palme H., Spettel B., Reimold W.U., Wacker K. and Wanke H. (1985) Composition andevolution of the lunar crust in the Decartes Highlands, Apollo16. J. Geophys. Res. 90, C449–C506.

Stoffler D., Keil K. and Scott E. R. D. (1991) Shock metamorphismof ordinary chondrites. Geochim. Cosmochim. Acta 55, 3845–

3867.

Stoffler D., Ryder G., Ivanov B. A., Artemieva N. A., Cintala M. J.and Grieve R. A. F. (2006) Cratering history and lunarchronology. In New Views of the Moon Reviews in Mineralogy

& Geochemistry, vol. 60 (eds. B. L. Jolliff, M. A. Wieczorek, C.K. Shearer and C. R. Neal), pp. 519–596.

Taylor S. R. (1982) Planetary science: a lunar perspective. Lunar

Planet. Inst., 502.Taylor S. R. (1984) The origin of the Moon: geochemical

considerations. In Origin of the Moon (eds. W. K. Hartmann,R. J. Phillips and G. J. Taylor). Lunar Planet Inst., Houston,

pp. 125–143.

Taylor L. A., Shervais J. W., Hunter R. H. and Laul J. C. (1983)Ancient (4.2 AE) highlands volcanism: the gabbronorite con-nection? Lunar Planet. Sci. 14, 777–778.

Terada K., Anand M., Sokol A. K., Bischoff A. and Sano Y. (2007)4.35 Ga basaltic magmatism on the Moon recorded in Kalahari009. Nature 450, 849–852.

Touboul M., Kleine T., Bourdon B., Palme H. and Wieler R.(2007) Late formation and prolonged differentiation of theMoon inferred from W isotopes in lunar metals. Nature 450,

1206–1209.

Turner G. (1971) 40Ar–39Ar dating: the optimization of irradia-tion parameters. Earth Planet. Sci. Lett. 10, 227–234.

Unruh D. M., Stille P., Patchett P. J. and Tatsumoto M. (1984)Lu–Hf and Sm–Nd evolution in lunar mare basalts. J. Geophys.

Res. 89(Suppl.), B459–B477.

Wanke H., Kruse H., Palme H. and Spettel B. (1977) Instrumentalneutron activation analysis of lunar samples and the identifi-cation of primary matter in the lunar highlands. J. Radioanal.

Chem. 38, 363–378.

Wanke H., Dreibusm G. and Palme H. (1978) Are the siderophileelements found in the lunar highlands of truly meteoritic origin?Lunar Planet. Sci. 9th, 1202–1204.

Warren P. H. (1985) The magma ocean concept and lunarevolution. Annu. Rev. Earth Planet. Sci. 13, 201–240.

Warren P. H. (2005) The Moon. In Treatise on Geochemistry, vol. 1(eds. H. D. Holland and K. K. Turekian.), pp. 559–599.

Warren P. H. and Kallemeyn G. W. (1993) Geochemical investi-gation of two lunar mare meteorites: Yamato-793169 andAsuka-881757. Proc. NIPR Symp. Antarct. Meteorit. 6, 35–57.

Warren P. H. and Wassson J. T. (1980) Further foraging forpristine nonmare rocks: correlations between geochemistry andlongitude. Proc. Lunar Planet. Sci. Conf. 11th, 431–470.

Warren P. H., Taylor G. J., Keil K., Kallemeyn G. W., Shirley D.N. and Wasson J. T. (1983) Seventh foray: Whitlockite-richlithologies, a diopside-bearing troctolitic anorthosite, ferroananorthosites, and KREEP. Proc. Lunar Planet. Sci. Conf. 14th

J. Geophys. Res. 88, B151–B164.

Warren P. H., Taylor L. A., Kallemeyn G., Cohen B. A. andNazarov M. A. (2001) Bulk-compositional study of threeDhofar Lunar meteorites: enigmatic siderophile element resultsfor Dhofar 026. Lunar Planet. Sci. XXXII. Lunar Planet. Inst.Houston. #2197 (abstr.).

Weckwerth G., Spettel B. and Wanke H. (1983) Phosphorus in themantle of planetary bodies. Terra Cognita 3, 79–80 (abstract).

Wentworth S., Taylor G. J., Warner R. D., Keil K., Ma M.-S. andSchmitt R. A. (1979) The unique nature of Apollo 17 VLT marebasalts. Proc. Lunar Planet. Sci. Conf. 10th, 207–223.


Weyer S., Munker C., Rehkamper M. and Mezger K. (2002) Determi-nation of ultra-low Nb, Ta, Zr and Hf concentrations and thechondritic Zr/Hf and Nb/Ta ratios by isotope dilution analyses withmultiple collector ICP-MS. Chem. Geol. 187, 295–313.

Wieler R. (2002) Cosmic-ray-produced noble gases in meteorites.In Noble Gases in Geochemistry and Cosmochemistry, (eds. D.Porcelli, C. J. Ballentine and R. Wieler). Rev. Mineral.

Geochem. 47, 125–170.

Wlotzka F. (1993) A weathering scale for the ordinarychondrites. Annual Meeting of the Meteoritical Society

28th, 460 (abstract).Wood J. A. (1975) Lunar petrogenesis in a well-stirred magma

ocean. Proc. Lunar Planet. Sci. Conf. 6th, 1087–1102.

Associate editor: Christian Koeberl

Geochemistry, petrology and ages of the lunar meteorites Kalahari 008 and 009: New constraints on early lunar evolution

Documents