The early geological history of the Moon inferred from ancient lunar meteorite Miller Range 13317

Miller Range (MIL) 13317 is a heterogeneous basalt‐bearing lunar regolith breccia that provides insights into the early magmatic history of the Moon. MIL 13317 is formed from a mixture of material with clasts having an affinity to Apollo ferroan anorthosites and basaltic volcanic rocks. Noble gas data indicate that MIL 13317 was consolidated into a breccia between 2610 ± 780 Ma and 1570 ± 470 Ma where it experienced a complex near‐surface irradiation history for ~835 ± 84 Myr, at an average depth of ~30 cm. The fusion crust has an intermediate composition (Al2O3 15.9 wt%; FeO 12.3 wt%) with an added incompatible trace element (Th 5.4 ppm) chemical component. Taking the fusion crust to be indicative of the bulk sample composition, this implies that MIL 13317 originated from a regolith that is associated with a mare‐highland boundary that is KREEP‐rich (i.e., K, rare earth elements, and P). A comparison of bulk chemical data from MIL 13317 with remote sensing data from the Lunar Prospector orbiter suggests that MIL 13317 likely originated from the northwest region of Oceanus Procellarum, east of Mare Nubium, or at the eastern edge of Mare Frigoris. All these potential source areas are on the near side of the Moon, indicating a close association with the Procellarum KREEP Terrane. Basalt clasts in MIL 13317 are from a very low‐Ti to low‐Ti (between 0.14 and 0.32 wt%) source region. The similar mineral fractionation trends of the different basalt clasts in the sample suggest they are comagmatic in origin. Zircon‐bearing phases and Ca‐phosphate grains in basalt clasts and matrix grains yield 207Pb/206Pb ages between 4344 ± 4 and 4333 ± 5 Ma. These ancient 207Pb/206Pb ages indicate that the meteorite has sampled a range of Pre‐Nectarian volcanic rocks that are poorly represented in the Apollo, Luna, and lunar meteorite collections. As such, MIL 13317 adds to the growing evidence that basaltic volcanic activity on the Moon started as early as ~4340 Ma, before the main period of lunar mare basalt volcanism at ~3850 Ma.


INTRODUCTION
Determining the distribution and duration of ancient lunar volcanism is vital to our understanding of the early evolution of the interior and thermal history of the Moon, and in the wider context of volcanism on other planetary bodies. The products of these volcanic eruptions infilled large impact basins, most of which are situated on the lunar near side. The earliest timing of basaltic volcanism is currently unknown, although the majority of the Apollo mare basalts erupted betweeñ 3850 and~3200 Ma (Head 1976;Head and Wilson 1992;Nyquist and Shih 1992;St€ offler and Ryder 2001;Hiesinger et al. 2003). Ancient lunar volcanism (>3900 Ma) has been documented in the Apollo sample collection by KREEP basalt fragments (Ryder and Spudis 1980;Taylor et al. 1983;Terada et al. 2007) and the~4300 to 3900 Ma Apollo 14 high-Al mare basalt group (Taylor et al. 1983;Nyquist and Shih 1992;Snyder and Taylor 2001;Neal and Kramer 2006). Lunar basaltic meteorites extend the range of these magmatic processes in representing both the youngest (e.g.,~2930 Ma, Northwest Africa [NWA] 032: Borg et al. 2009) and oldest (~4350 Ma, Kalahari 009: Terada et al. 2007;Sokol et al. 2008) known samples. Stratigraphic relationships inferred from remote sensing data and crater counting model age calculations have also indicated occurrences of ancient >3850 Ma volcanism over widespread regions including Mare Frigoris, Mare Australe, Mare Tranquillitatis, and Mare Serenitatis (e.g., Spudis 1979, 1983;Hiesinger et al. 2000Hiesinger et al. , 2003Hiesinger et al. , 2008Whitten and Head 2015).
Lunar meteorites represent invaluable additional samples that provide information about different rock types, originating from regions that are remote from the near side equatorial Apollo and Luna landing sites, including samples from the lunar far side (Warren and Kallemeyn 1991;Korotev et al. 2003;Korotev 2005;Joy and Arai 2013;Calzada-Diaz et al. 2015). In this paper, we report a detailed petrographic, geochemical, and noble gas investigation of lunar meteorite Miller Range (MIL) 13317, coupled with in situ U-Pb analyses of zircon-rich (Zr-rich) and calcium-phosphate (Caphosphate) phases. We demonstrate that this meteorite contains various clasts originating from ancient basalts, shedding further light on the earliest phase of the Moon's magmatic history.

Miller Range 13317
Miller Range 13317 is a 32.2 g lunar meteorite that was found in Antarctica in 2013 by the Antarctic Search for Meteorites program (ANSMET). The meteorite was initially classified as an anorthositic breccia (NASA 2015). However, preliminary examinations (Curran et al. 2016;Shaulis et al. 2016;Zeigler and Korotev 2016) indicated that the sample is a mixture of both anorthositic and mafic components, and likely belongs in the category of a "mingled" regolith breccia (e.g., Korotev et al. 2009). MIL 13317 can be further subclassified as moderately mafic (FeÕ 9.0 to 16.0 wt%) and Th-rich (5.4 ppm), similar to only seven other lunar stones/meteorite groupings including Lynch 002, the NWA 4472/4485 paired stones, Calcalong Creek, Sayh al Uhaymir (SaU) 169, Dhofar (Dho) 1442, and NWA 6687 (Korotev 2018). Initial studies of Zr-bearing and Ca-phosphate minerals indicated crystallization ages of basalt clasts of 4332 AE 2 Ma (Snape et al. 2018), 4351.8 AE 8.7 Ma, and 4270 AE 24 Ma (Shaulis et al. 2016), which are similar in age to the youngest lunar ferroan anorthosites and ancient magmatic rock suites.

ANALYTICAL TECHNIQUES
Analyses of MIL 13317 were performed on a polished thin section (MIL 13317,7), a small polished block (MIL 13317,11a), and bulk chips (from MIL 13317,10 and MIL 13317,11b), all provided by the NASA Meteorite Working Group. The MIL 13317,10 subsplit is from a regolith portion of the sample and MIL 13317,11 is from a fragmental portion. The methodology we employed is outlined below and in more detail in Data S1 and Table S1 in the supporting information.

Sample Petrography and Chemistry
Two sections (MIL 13317,7 and MIL 13317,11a) were carbon-coated and analyzed at the University of Manchester.
Petrographic characterization was performed using a FEI XL30 field emission gun environmental scanning electron microscope with electron dispersive spectrometer and EDAX Genesis software. Montaged backscattered electron (BSE) images (Fig. 1a) and element maps (Fig. 1b) were processed using the ImageJ software package (Rueden et al. 2017). Optical microscope-cathode luminescenceimaging (CL-imaging) was acquired using a CITL 8200 mk3 "cold" CL system coupled to a transmitted-light microscope; typical conditions were~10 to 15 kV and a current of~300 mA.
Mineral phases were analyzed using a Cameca SX 100 electron microprobe (EMPA) equipped with a BSE detector and five wavelength-dispersive spectrometers. Major (Si, Al, Fe, Ca, Mg, Ti) and minor (Na, K, P, Mn, Cr, S, Ni) elements were analyzed with a 15 kV accelerating voltage, a beam current of 20 nA, and a focused beam diameter of 1 lm. Counting times between 10 and 50 s were used for each element.
Typical errors were~0.1 wt% for major elements and 0.05 wt% for minor elements for both calibration and sample analyses. The instrument was calibrated against well-characterized mineral standards. Acceptable analytical totals were taken to be between 98 and 101 wt% and mineral compositions were checked for stoichiometry. A defocused beam of 10 lm was used to assess the bulk composition of fine-grained clasts and homogenous glass areas within the fusion crust.
To further assess the bulk rock composition of MIL 13317, we used glass beads that were produced by fusing MIL 13317,10 (Bulk 1) and MIL 13317,11b (Bulk 2) by laser heating for noble gas analysis (see the Noble Gas Analysis section). Glass beads were mounted on glass slides using superglue adhesive, polished, and carbon-coated. A 10 lm defocused beam was used to determine major elements by EMPA (Table 1). It is noted that as a result of the high temperatures experienced during melting, there is a potential for the loss of some volatile elements using this approach, although we assume these losses to be negligible as we detected Na in the glass beads made after laser heating (Table 1).  232 Th, and 238 U. Samples were ablated using a spot size of 60 lm, at a repetition rate of 8 Hz with the laser energy at the target (fluence) regulated at~5.5 J cm À2 . Laser ablation times of 60 s were used with 10 s dwell times between analyses. SRM NIST 612 was used as the calibration standard. Calcium and Si were used as the internal normalizing standards for plagioclase and other phases, respectively. Repeat NIST 610 and NIST 614 analyses were measured over the analytical campaign in order to assess the data quality (supporting information Table S2). Typical %RSD for most elements in NIST 610 is better than 5% (except for Ni-19%, Ti-8%, Cr-6%, Sr-10%). Typical %RSD for elements in NIST 614 is better than 10% (except for Sc-38%, Cr-27%, Mn-31%, Co-23%, Zr-24%, Mo-14%). Detection limits varied by <10 ppb. The abundances obtained for NIST 610 and 614 are within range of certified values reported within the GeoReM database (georem.mpch-mainz.gwdg.de). Data reduction was conducted using Iolite software (v3) using the "Trace Elements IS" data reduction scheme, where the time-resolved display and segment-picking for data integrations allow signal spikes from inclusions to be identified and avoided.

Noble Gas Analysis
Two rock chips, from subsplits MIL 13317,10 and MIL 13317,11b, were investigated for their bulk rock noble gas content. Neon, Ar, and Xe were analyzed on a Fig. 1. a) Montaged backscattered electron images of thin section (MIL 13317,7) and polished block (MIL 13317,11a) of subsplits of MIL 13317. b) False-color element maps of two sections of MIL 13317 where colors correspond to: Ca = yellow (phosphates), Mg = green (pyroxene cores), Si = blue (silica), Fe = red (olivine, pyroxene rims), Al = white (plagioclase), Ti = pink (ilmenite and spinel), and K = cyan (k-feldspar). The false-color maps were used to assess the diversity of clasts within MIL 13317 (see approach of Joy et al. 2011c). Figure S6 in supporting information shows the difference between the two subsplits of MIL 13317. c) Backscattered electron image of MIL 13317,7 with clasts studied in this work outlined in red. The clasts are described in Table S1. (Color figure can be viewed at wileyonlinelibrary.com.) Thermo Scientific TM HELIX-MC noble gas mass spectrometer at the University of Manchester. A Photonmachines TM fusion diode laser (3 mm beam diameter) was used to step-heat samples from 0.6 to 21 W (in six to nine steps depending on mass of the sample) corresponding to the complete melting of the sample. Prior to analysis, chips were loaded into an aluminum holder in the laser port and baked for 24 h at 200°C to remove any absorbed atmospheric gases. During analysis, any active gases released from the sample were removed by two SAES getters (NP10 and GP50) during the gas preparation stage (see Section S1.2 in Data S1 for full experimental technique). Blanks and air calibrations were performed before and after every sample analysis. Typical blanks were 6.1 9 10 À12 cm 3 .STP 20 Ne, 8.6 9 10 À11 cm 3 .STP 40 Ar, and 1.2 9 10 À14 cm 3 .STP 132 Xe. All isotopes were background and blank corrected. Neon-22 was corrected for interferences from CO 2 ++ . The sensitivity and mass discrimination of the mass spectrometer were determined using 0.2 cc aliquots of pure air at a known pressure.

Secondary Ion Mass Spectrometry
The Ca-phosphate (apatite and merrillite), zircon, and baddeleyite U-Pb systems were analyzed using a CAMECA IMS 1280 ion microprobe at the NordSIMS facility at the Swedish Museum of Natural History, Stockholm, using a methodology similar to that outlined in previous studies (Whitehouse et al. 1997(Whitehouse et al. , 2005Nemchin et al. 2009b;Snape et al. 2016). Full details of the secondary ion mass spectrometry (SIMS) methodology are included in Section S1.3 in Data S1.
These data were processed using in-house SIMS data reduction spreadsheets and the Excel add-in Isoplot (version 4.15;Ludwig 2012). Three different sets of ages were calculated (1) Pb-Pb isochron ages, using the Pb isotope compositions uncorrected for contamination with terrestrial common Pb, (2) terrestrial common Pb corrected 207 Pb/ 206 Pb ages (using the model of Stacey and Kramers [1975] and their values for present-day terrestrial Pb isotopic ratios), and (3) U-Pb concordia ages. A U-Pb concordia age was not calculated for the baddeleyite and merrillite grains due to the lack of analyses on appropriate matrixmatched external standards and the well-known matrixorientation problems associated with U-Pb calibration of baddeleyite (Wingate and Compston 2000). All age uncertainties are reported at the 95% confidence limit in the following discussion. Following SIMS analysis, SEM images were acquired of the targeted phases to assess the exact locations of the SIMS pits, to ascertain if any overlapped with mineral grain boundaries (see supporting information Fig. S1). Four such analyses were identified (supporting information Table S3) and excluded from the combined age calculations (e.g., weighted averages of 207 Pb/ 206 Pb ages or Pb-Pb isochrons), and are not considered further.

RESULTS
The petrology and geochemistry of MIL 13317 are consistent with a lunar origin. The Fe/Mn ratios of pyroxene and olivine in MIL 13317 plot on previously determined lunar meteorite trend lines ( Fig. 2) (Papike 1998;Karner et al. 2003Karner et al. , 2006Joy et al. 2014). MIL 13317 also contains trapped "solar" noble gases with similar compositions to other lunar meteorites (see the Regolith History section).

Petrography and Mineral Chemistry of MIL 13317
Miller Range 13317 is a well-consolidated regolith breccia, comprising a diverse array of polymict lithic clasts (black, dark-gray, clear, and cream) fused together in a heterogeneous matrix of the same material, with shocked melt veins and mineral fragments (supporting information Figs. S2 and S3). Lithic clasts are up tõ 7 mm in size and include regolith breccias, basalts, feldspathic fragments, symplectites, norites, and granulites (Figs. 1 and 3). Mineral fragments are up to 1.5 mm in diameter and typically consist of pyroxene, plagioclase, and olivine. Impact melt breccias are abundant in the sample, including clast-rich and crystalline types. Melt veins are variable in size and crosscut both the matrix and clasts in random orientations (Fig. 3a). Rare glass spherules (<100 lm) were found in the sample matrix (see Fig. 3i inset), but no agglutinates were identified. Various clast types are described below (see Fig. 1c for clast location and number). Mineral and bulk clast compositions are given in supporting information Tables S4-S8.

Basaltic Clasts
There are numerous medium-and fine-grained basaltic clasts in MIL 13317 (Fig. 1). The largest clasts (Clasts 1 and 4, Figs. 3a and 3b) are~2.0 9 1.5 mm in size, and are composed of an intergrowth of pyroxene, plagioclase, silica, and areas of mesostasis (e.g., K-glass, phosphates, zircons) associated with late-stage crys-tallized residual melt material of basaltic magmas. Other basaltic clasts range from fine-grained (minerals <0.2 mm) clasts with "pods" of late-stage mesostasis, to coarse-grained (minerals <1 mm) clasts and fragments with both intergranular and subophitic textures. These textures suggest a range of cooling histories for basaltic clasts in MIL 13317.

Mineral Fragments
The matrix of MIL 13317 consists of single mineral fragments up to 1.5 mm in size. The majority of mineral fragments are plagioclase and pyroxene with a wide range of compositions (An 78-99 and Wo 1-44 En 1-70 Fs 12-88 , respectively, Figs. 4 and 6). Trace element compositions of the matrix pyroxene are similar to pyroxene within the basaltic clasts (Fig. 5b). The REE patterns display negative Eu-anomalies (Eu/Eu* = 0.01-0.05) and variable LREE depletions of [La/Sm] CI = 0.09-0.17 (Fig. 5b). Silica and ilmenite are also present with minor amounts of Mg-rich olivine up to 1.5 mm in size.
Accessory minerals in the matrix are mainly Caphosphate (merrillite and apatite) and Zr-bearing phases <100 lm in size. Three Zr-bearing phases were identified: zircon, baddeleyite, and the rare lunar mineral tranquillityite (see Fig. S1). These accessory phases are typically associated with symplectic areas with pyroxene, minor olivine, K-rich glass, and K-feldspar.

Evolved Lithologies
Highly evolved lithologies of granitic (granophyric) K-feldspar and silica intergrowths, up to 0.7 mm are present in MIL 13317 ( Fig. 1-cyan colored phases). The K-feldspar occurs as blocky phases with lath-like silica. These assemblages are commonly associated with regions of mesostasis and areas of K-rich glass, and are most similar to the Apollo high-alkali suite rock types. It is possible that some of these are fragments of mesostasis from the basaltic clasts.

Fusion Crust and Bulk Rock Analysis
A dark-brown vesicular fusion crust is present on the MIL 13317,7 section, which has an average composition shown in Table 1 (Fig. S2). The fusion crust displays LREE enrichment ([La/Sm] CI = 1.40), fractionated HREE profile ([Dy/Yb] CI = 1.15), and a negative Eu-anomaly (Fig. 5e). The Th-composition of the fusion crust is 5.41 ppm (Table S2).

Noble Gas Inventory
Five separate chips of MIL 13317 (three from MIL 13317,10 and two from MIL 13317,11b) were analyzed for Ne and Ar isotopes and one chip (from MIL 13317,10) for Xe isotopes. A summary of the noble gas data is given in Table 2 and Tables S9-S11 in the supporting information.
Neon data for MIL 13317,10 are dominated by trapped 20 Ne/ 22 Ne and 21 Ne/ 22 Ne ratios (Fig. 8a), whereas MIL 13317,11b shows a trend of increasing 21 Ne/ 22 Ne and decreasing 20 Ne/ 22 Ne ratios with increasing laser heating steps (see Table S9). Trapped 20 Ne/ 22 Ne ratios ( 20 Ne/ 22 Ne) tr were determined using a least squares fitting to the data on a three-isotope mixing diagram (Fig. 8a). Overall, Ne in MIL 13317 has a ( 20 Ne/ 22 Ne) tr ratio between 10.8 and 12.6 (for both subsplits, Table 2). The concentrations of 20 Ne tr in MIL 13317,10 are higher than in MIL 13317,11b ( The concentration of cosmogenic 38 Ar c is an order of magnitude higher in MIL 13317,10 compared to MIL 13317,11b (Table 2). There is also a factor of approximately 2 difference in 38 Ar c within the three subsplits of MIL 13317,10. In contrast, 21 Ne c concentrations for all analyzed chips are similar and mostly within error of each other (Table 2).
Three-isotope mixing diagrams of 124 Xe/ 130 Xe versus 126 Xe/ 130 Xe (see Fig. 8b) illustrate the components present in MIL 13317,10. The sample is dominated by a solar wind component (98%), but contains minor amounts of spallation products from Ba: the bulk spallation Xe signature corresponds to (30 AE 11%) (1r) of 126 Xe derived from Ba. Heavy Xe isotopes formed by fission of U or Pu are not evident in any of the samples (supporting information Fig. S4). Petrographic observations revealed that MIL 13317 contains REEbearing phases like zircon, apatite, merrillite, and K-rich glass, which account for the presence of some spallogenic xenon components. The depth-sensitive shielding indicator 131 Xe/ 126 Xe c is 5.6 AE 0.4 (1r), assuming an average lunar regolith density of 1.5 g cm À3 (Carrier et al. 1991), this corresponds to an average burial depth on the Moon of about 30 cm. This depth was used to calculate production rates of P 21 = 0.1789 9 10 À8 cm 3 STP/Ma, P 38 = 0.1369 9 10 À8 cm 3 STP/Ma, and P 126 = 0.2250 9 10 À8 cm 3 STP/Ma (based on the method of Hohenberg et al. 1978). These production rates yield exposure ages for neon (t 21 ) of between 71 AE 28 and 320 AE 117 Myr, argon (t 38 ) of between 457 AE 1.4 and 5652 AE 159 Myr, and xenon (t 126 ) of 835 AE 84 Myr (Table 2).

U-Pb and Pb-Pb Chronology
Analyses of a zircon and five baddeleyite grains in MIL 13317,7 yield terrestrial common Pb corrected 207 Pb/ 206 Pb ages between 4326 AE 12 Ma and 4349 AE 16 Ma (Table 3). The zircon analyses were carried out on a single~70 lm grain. Calculating either a weighted average 207 Pb/ 206 Pb age (4337 AE 17 Ma; MSWD = 2.9; P = 0.034; Fig. 9), or a Pb-Pb isochron age (4331 AE 49 Ma; MSWD = 4.1; P = 0.016; Fig. 10c) from these data results in high MSWD values and low probabilities of fit, potentially due to partial resetting of the U-Pb system within different parts of the grain. Nonetheless, a U-Pb concordia age of 4334 AE 17 Ma is determined for the three zircon analyses with less than 10% discordance (concordance and equivalence MSWD = 1.4; P = 0.24; Fig. 11a). Despite being located in separate areas within the breccia matrix, the five baddeleyite grains analyzed were all associated with similar symplectite assemblages (Fig. S1) Fig. 9) and a Pb-Pb isochron age of 4344 AE 5 Ma (MSWD = 0.12; P = 0.95; Fig. 10d).
Two of the 12 matrix apatite analyses were obtained from a single~40 lm apatite grain, within which it was possible to place two SIMS spots. These two SIMS spots yielded the lowest 207 Pb/ 206 Pb ratios in the sample, with a  (Table 3). It was also noted that one of these analyses was an apatite grain within a large symplectite assemblage (Clast 16, Fig. 3h), similar to the smaller symplectite assemblages containing the baddeleyite grains. If these 10 matrix apatite analyses are considered to form a single population, a weighted average 207 Pb/ 206 Pb age of 4342 AE 4 Ma (MSWD = 0.59; P = 0.81; Fig. 9) is obtained. A U-Pb concordia age of 4342 AE 13 Ma (concordance and equivalence MSWD = 1.2; P = 0.28; Fig. 11c) is obtained for the six analyses that are less than 10% discordant; however, the 204 Pb/ 206 Pb and 207 Pb/ 206 Pb ratios of these matrix apatite analyses do not provide sufficient spread to constitute a 10-point isochron. The similar ages of the matrix zircon, baddeleyite, and apatite mean that it is possible to combine all of these analyses into a single weighted average 207 Pb/ 206 Pb age of 4342 AE 3 Ma (MSWD = 0.99; P = 0.47; Fig. 9), or a Pb-Pb isochron corresponding to an age of 4342 AE 4 Ma (MSWD = 1.14; P = 0.31; Fig. 10f). Apatite analyses were also obtained from two of the basaltic clasts (six analyses from Clast 4 and one from Clast 10), which give similar 207 Pb/ 206 Pb ages (between 4329 AE 4 Ma and 4351 AE 16 Ma) to the majority of the matrix apatite grains. A Pb-Pb isochron age of 4334 AE 4 Ma (MSWD = 1.02; P = 0.39; Fig. 10e) was calculated for Clast 4, although calculating a weighted average of the 207 Pb/ 206 Pb ages (4333 AE 5 Ma) results in a relatively high MSWD and low probability of fit (MSWD = 2.2; P = 0.055; Fig. 9). The relatively good statistical fit of the Pb-Pb isochron for apatite in Clast 4 compared with the scatter in 207 Pb/ 206 Pb ages may reflect an inappropriate common Pb correction, assumed to have a modern terrestrial Pb composition, and the grains actually contain a small proportion of more radiogenic (when compared with terrestrial Pb) lunar initial Pb (Fig. 10e).
The merrillite grains show a significantly wider range of Pb isotopic compositions than the apatite, zircon, or baddeleyite grains, resulting in 207 Pb/ 206 Pb ages from 4277 AE 12 Ma to 4401 AE 38 Ma. This is shown on a plot of 207 Pb/ 206 Pb versus 204 Pb/ 206 Pb (Figs. 10a and 10b), with the merrillite grains forming a separate trend below and to the right of the other phases.

DISCUSSION
Miller Range 13317 represents a mixture of lunar material which is of basaltic, anorthositic, and impact origin. The ages of the mineral phases (>3.9 Ga) within the sample provide unique insights into the nature of ancient lunar volcanism. In the following discussion, we synthesize the results of our study to understand MIL 13317's geological history within the context of other lunar meteorites and Apollo samples.

Constraining the Source Region of MIL 13317
Due to the regolithic nature of MIL 13317, remote sensing chemical data can be used to help locate potential source regions on the Moon. Previous work has used this approach to constrain the source regions of specific chemical types of lunar rocks (Kramer et al. 2015) and lunar meteorites (e.g., Gnos et al. 2004 13317 (a, b). These data were also used to determine Pb-Pb isochron ages for the zircon (c), baddeleyite (d), and Clast 4 apatite (e) grains. It is possible to construct a single combined isochron (f) through all of the matrix zircon, baddeleyite, and apatite data, assuming that all of these grains originated from the same igneous precursor. PBC = common terrestrial Pb, calculated following the model assumptions of Stacey and Kramers (1975). Uncertainties on the Pb-Pb isochron ages are stated at the 95% confidence level. (Color figure can be viewed at wileyonlinelibrary.com.) which features FeO and Th abundances in bins of 2°e qual area per pixel have been used to identify areas of the lunar regolith showing compositional similarities to MIL 13317 (Fig. 12). A range of FeO (9.6-12.3 wt%, Table 1) and Th (4.6-5.4 ppm) was used to account for compositional heterogeneities between sample subsplits. Regions with similar FeO contents to MIL 13317 (red pixels, Fig. 12) are located mainly in the Procellarum KREEP Terrane (PKT) and a few outcrops in the South Pole-Aitken basin region. However, the high-Th content (blue pixels, Fig. 12) constrains MIL 13317 to the near side of the Moon with matches for the Th content only outcropping in the PKT region. The overlap of both the FeO and Th contents of MIL 13317 (green pixels, Fig. 12) occurs only in a few locations in the near side PKT region. These locations, where the FeO and Th content overlaps, are the most likely source regions of MIL 13317.
Regions on the Moon that are most similar in composition to MIL 13317 (i.e., the green pixels in Fig. 12), include the northwest region of Oceanus Procellarum, the eastern edge of Mare Frigoris, and to the east and northeast of Mare Nubium. Crater counting data sets from basalt surface units of these two regions (Hiesinger et al. 2003(Hiesinger et al. , 2010 indicate that the lava flows are no more than~3700 Ma. This suggests that if MIL 13317 originates from one of these regions, then the basalt component in MIL 13317 represents ancient cryptomare basalts that were buried by subsequent lava flows and/or impact ejecta within the PKT (e.g., Head and Wilson 1992).

Differences to Other Miller Range Lunar Meteorites
MIL 13317 is the seventh lunar meteorite found in the Miller Range region of Antarctica to date. Other Miller Range brecciated meteorites include MIL 090036 and the grouped MIL 090034/090070/090075 (MIL 09 group) stones, which are feldspathic regolith breccias grouped on the basis of their chemical composition and petrography (Korotev et al. 2011;Zeigler et al. 2012), and exposure ages (Nishiizumi and Caffee 2013). The lack of basaltic material in the group of MIL 09 samples (Liu et al. 2011;Korotev and Zeigler 2014;Calzada-Diaz et al. 2015;Martin et al. 2017), and low bulk rock Th and KREEP-rich elements suggest that they are unlikely to be source paired with MIL 13317 (supporting information Fig. S5a). MIL 090036 is chemically distinct from MIL 13317, as it is more feldspathic (27.1 wt% Al 2 O 3 , 5.0 wt% FeO -Korotev and Zeigler 2014: compared to MIL 13317 from this work with 15.9 wt% Al 2 O 3 , 12.3 wt% FeO) and has no basaltic material (Fig. S5a) (Liu et al. Fig. 11. U-Pb concordia diagrams for data that are <10% discordant in the zircon (a), younger matrix apatite (b), and older matrix apatite (c) grains in MIL 13317. Uncertainties on the concordia ages are stated at the 95% confidence level and include the decay constant errors. (Color figure can be viewed at wileyonlinelibrary.com.) 2011; Korotev and Zeigler 2014;Calzada-Diaz et al. 2015;Martin et al. 2017). Similar to MIL 13317, the MIL 07006 stone is a regolith breccia that contains basaltic clasts (Korotev et al. 2009;Liu et al. 2009;Joy et al. 2010a;Robinson et al. 2012). However, MIL 07006 basalts are much more Mg-rich, Ti-depleted, and KREEP-poor (e.g., Th = 0.36 ppm, Fig. S5b) compared to MIL 13317, ruling out a pairing relationship.
Miller Range 13317 is petrographically similar to the mingled anorthositic and basaltic regolith breccia Meteorite Hills (MET) 01210, which is grouped with MIL 05035 (as well as Yamato-793169 and Asuka-881757, collectively known as the YAMM meteorites: Day et al. 2006;Joy et al. 2008;Arai et al. 2010aArai et al. , 2010b. MIL 13317 bulk rock Ti compositions are low (1.0 wt% TiO 2 , Table 1) like MET (1.53 wt% TiO 2 ; Korotev and Zeigler 2014), which is indicative of a very low-Ti (VLT) to low-Ti source region. However, MIL 13317 has lower FeO (12.3 wt%) and higher K 2 O (0.27 wt%) content than MET (16.4 wt% FeO, 0.05 wt% K 2 O; Korotev and Zeigler 2014). Furthermore, MET 01210 contains agglutinates (Arai et al. 2010a(Arai et al. , 2010b, which were not found in MIL 13317. In summary, we exclude a pairing relationship with any of the other Miller Range lunar meteorites found to date.
Concentrations of Sc and Sm in the MIL 13317 fusion crust are similar to those in Th-rich meteorites like Calcalong Creek and Lynch 002 (Fig. 5f). Calcalong Creek (Hill et al. 1991;Hill and Boynton 2003) and Lynch 002 (Smith et al. 2012;Korotev 2013;Robinson et al. 2016) are the only lunar meteorites found in Australia (located over 800 km apart from each other). Although not paired with each other, they have features in common with MIL 13317. All three meteorites have a mixed clast assemblage of anorthosite, mare basalts, and KREEP-rich material, and have comparable Th (~4-5 ppm) and FeO (~9-10 wt%) contents (Korotev 2018). Calcalong Creek was potentially launched from around the edge of the Imbrium region (Calzada-Diaz et al. 2015), we suggested a similar region of the Moon for MIL 13317 (Fig. 12). Like MIL 13317, pyroxene in basaltic clasts of Lynch 002 show basaltic chemical fractionation trends typical of VLT to low-Ti mare basalts of the Apollo suite, and the basaltic clasts are rich in silica. These petrographic and compositional similarities may indicate a source-pairing relationship between Lynch 002 and MIL 13317, although this needs to be tested further by radiogenic and cosmogenic nuclide analysis (e.g., crystallization age, ejection age, terrestrial age, and trapped and cosmogenic components) of Lynch 002.

Regolith History
Miller Range 13317 is a heterogeneous mix of phases that have undergone different burial and reexposure histories. The few spherules present affirm a regolith origin for the sample, but their scarcity suggests that the parent regolith from which the sample consolidated was relatively immature (McKay et al. 1991).
The two subsplits analyzed for noble gases have sampled different components within MIL 13317: one being dominated by a regolith breccia component (MIL 13317,10) and the other (MIL 13317,11b) dominated by a more fragmental breccia, with smaller amounts of regolith material (see Fig. S6). Analysis of the two subsplits of MIL 13317 yields exposure ages that range from 71 AE 28 to 2860 AE 74 Myr (supporting information Table S11). The anomalously high argon cosmic ray exposure durations (2860 AE 74 and 5652 AE 159 Myr: Table S11) from two of the subsplits, are likely to be a result of the sample being saturated with implanted parentless 40 Ar tr and solar wind implanted 36 Ar tr . As a result, we cannot adequately correct for the presence of these components in these subsplits, which has an effect on the calculated 38 Ar cos . Although the Ne exposure ages are generally consistent between subsplits, they are lower than those obtained from cosmogenic Ar and Xe isotopes. This may be explained by partial loss of Ne, potentially during an impact event. Thus, a cosmic ray exposure age of 835 AE 84 Myr is adopted for MIL 13317 based on the t 126 exposure age, as heavy noble gases are less likely to have been affected by shock-related diffusional gas loss or subject to large corrections for solar wind-derived trapped components.
The lithification of MIL 13317 from soil into breccia components is estimated to be between 2610 AE 780 and 1570 AE 470 Ma (determined from the 40 Ar/ 36 Ar tr ratio antiquity indicator). MIL 13317 was consolidated into its current form and exposed to the cosmic environment for an overall period of 835 AE 84 Myr, where it has experienced a complex history at an average depth of~30 cm.

Implications for the Age Resetting History of MIL 13317
The ages determined from the zircon, baddeleyite, and the majority of the phosphate grains in MIL 13317 provide evidence of ancient lunar volcanism occurring at or before~4340 Ma (see also Snape et al. 2018). Although it is possible that these ages could represent the timing of impact induced resetting of the U-Pb system, this would still require formation of the igneous protolith(s) from which the grains originated, prior to this time. Equally, while it is possible that different grains within the meteorite breccia matrix originate from different protoliths, the association of the matrix apatite and baddeleyite grains with similar symplectite assemblages, combined with the similarity in their 207 Pb/ 206 Pb ages, suggests that the simplest explanation for the similar ages identified in the majority of the grains is that they originated from the same igneous protolith. Based on this interpretation, the weighted average 207 Pb/ 206 Pb age of 4342 AE 3 Ma (MSWD = 0.99; P = 0.47; Fig. 9), or the combined Pb-Pb isochron age of 4342 AE 4 Ma (MSWD = 1.14; P = 0.31; Fig. 10f) for the matrix zircon, baddeleyite, and apatite analyses provides the best estimate for the crystallization age of the igneous protolith. These ages are similar to those reported by Shaulis et al. (2016) for matrix phosphate and zircon-bearing phases (4351 AE 9Ma) in a separate thin section (MIL 13317,5) of the same meteorite. The apatite grains within basalt Clast 4 appear to be slightly younger (an average 207 Pb/ 206 Pb age of 4333 AE 5 Ma and a Pb-Pb isochron age of 4334 AE 4 Ma) than the matrix grains, but are consistent with Pb-Pb isochron ages (4332 AE 2 Ma) determined for several other basaltic clasts in the MIL 13317 meteorite in a separate study ). These younger ages may indicate that the clasts represent fragments of a separate mafic protolith; however, this is difficult to confirm given the tight clustering of ages, and the uncertainties associated with correcting for terrestrial Pb contamination.
The range of Pb isotope compositions in the merrillite grains may indicate that the U-Pb system in this phase is more easily reset than in apatite. However, this is contrary to the behavior observed in Caphosphate minerals in Apollo breccias (e.g., Merle et al. 2014;Snape et al. 2016;Thiessen et al. 2017Thiessen et al. , 2018, where breccia formation ages based on consistent 207 Pb/ 206 Pb ages in both apatite and merrillite phases, indicate that the U-Pb system is reset in both phases to a similar extent. Furthermore, the youngest 207 Pb/ 206 Pb ages of the merrillite analyses (~4280 Ma) in MIL 13317 are consistent with younger ages reported by Shaulis et al. (2016) of 4276 AE 21 Ma for two phosphate grains and 4270 AE 24 Ma for 22 baddeleyite grains, which they interpret as representing the crystallization age of a separate basalt lithology.
Although the mechanism by which the younger apatite grain was introduced to the breccia is unclear, it is notable that the U-Pb age of 3890 AE 33 Ma is close to that determined for the formation of the Imbrium impact basin in several studies (Gnos et al. 2004;Liu et al. 2011) including analyses of phosphates in breccias from Apollo 12, 14, and 17 (Snape et al. 2016;Thiessen et al. 2017Thiessen et al. , 2018. Without precisely knowing the provenance of the meteorite or this particular apatite grain, it is unclear if this age is (1) indicative of a link with the formation of the Imbrium basin, (2) a reflection of the general last stage of basin formation, or (3) a coincidental intermediate partial resetting age between the original crystallization of the apatite and a later impact event. As discussed earlier, considering the susceptibility of the apatite U-Pb system to be reset by impact events, and the observed tendency for phosphate phases analyzed in Apollo breccias to be completely reset (e.g., Nemchin et al. 2009b;Snape et al. 2016;Thiessen et al. 2017Thiessen et al. , 2018, the third interpretation seems slightly less likely. Without a clearer understanding of the Moon's impact history, and a resolution to the long-standing debate surrounding the potential~3900 Ma lunar cataclysm (Bottke and Norman [2017] and references therein; Morbidelli et al. [2018] and references therein), it is difficult to say which of the first two interpretations are more likely.

Geological Context and Significance of MIL 13317
Petrology and composition of MIL 13317 suggests that the sample is formed from a regolith in which there was a mixture of mare-like basaltic material and highland anorthositic material, with a minor KREEP component (Figs. 5 and 7). Based on comparisons with remote sensing data (Fig. 12), our best estimate is that the MIL 13317 meteorite is derived from the edge of the PKT on the lunar near side. Assuming this is the case, the mineral and clast chemistry, and age of components can provide new information about the geological evolution of the lunar crust in this region of the Moon.

What Types of Highland Rocks and Impact Melts Are Distributed in Different Crustal Regions?
MIL 13317 contains highland igneous rocks and impact melt breccias that are mainly associated with the FAN suite found at many of the Apollo landing sites. For example, plots of [Eu/Sm] CI versus Sm in plagioclase can be used to distinguish parent lithologies (Fig. 5d) (Russell et al. 2014) and indicates that plagioclase in Clast 5 falls within the field of lunar anorthosites as sampled by the Apollo missions and lunar meteorites. Trace element systematics for plagioclase from granulitic Clast 12 overlap with that of plagioclase from Clast 5 (Eu/ Eu* = 15.83; [La/Sm] CI = 1.77) and are also similar to the Apollo FAN suite (Figs. 5c and 5d). Granulitic Clast 14 is more similar to magnesian granulite found in feldspathic lunar meteorites (Treiman et al. 2010;Gross et al. 2014;Kent et al. 2017) that fall into the compositional gap between Apollo FAN and HMS (Fig. 7a). Furthermore, Clast 9 in particular has sampled material which is a mix of KREEP, feldspathic, and basaltic material with trace element characteristics that are distinct from the Apollo impact melts. These compositions and petrology are generally represented by the mineral grains in the matrix which have an affinity to ferroan anorthosite parent rocks.
In general, there is an apparent lack of HMS rock types in MIL 13317 considering the likely source region. The distribution of HMS rocks has been strongly linked to an origin in and surrounding the PKT region due to their possible association with KREEP-rich material (Snyder et al. 1995;Korotev 2000;Wieczorek et al. 2006). Similarly, the FANs have been associated with rocks typical of the feldspathic lunar highlands. However, feldspathic lunar meteorites, which represent a more global sampling of the feldspathic lunar highlands, generally lack Apollo-like FAN, KREEP, and HMS material, and have a feldspathic component that is more magnesian (e.g., the magnesian anorthosites) than the Apollo FAN suite (Korotev et al. 2003Korotev 2005;Zeigler et al. 2012;Gross et al. 2014). This suggests that the typical highland rocks and impact melts found during the Apollo missions (e.g., FANs, HMS) may not be globally distributed across the Moon (Gross et al. 2014). The highland components of MIL 13317 are compositionally more similar to the Apollo highland rock suite than typical feldspathic lunar meteorites, supporting this suggestion.

Was There a Single Magmatic Source for Basalt Clasts in MIL 13317?
Basalt clasts in MIL 13317 provide further insight into the magmatic processes occurring in regions remote from the Apollo and Luna landing sites, both in terms of composition and basaltic volcanism through time. The abundances of Sc and Sm in some bulk clasts (e.g., Clast 18, Fig. 5f) show that these clasts have also incorporated KREEP-rich material and are intermediate to the Apollo VLT-mare basalts, KREEP-rich basalts, and Apollo impact melts (see also Zeigler and Korotev 2016). The Fe# to Ti# correlations of pyroxene suggest that all the basalt clasts appear to be derived from lavas intermediate to the Apollo VLT and the low-Ti mare basalt source regions ( Fig. 4b; after Arai et al. 1996;Robinson et al. 2012). Pyroxene in basalt clasts show a typical fractionation trend for lunar basalts of increasing Fe# (atomic Fe/Fe + Mg, Fe# of 0.40-0.99) with increasing Ti# (atomic Ti/Ti + Cr, Ti# of 0.25 to 1) (Fig. 4b). The Al/Ti ratio can be compared with the Fe# to indicate element partitioning from the parental melt to further understand this crystallization trend (Nielsen and Drake 1978;Arai et al. 1996). The trends in MIL 13317 basalt clasts show that pyroxene (Fe# 40 to Fe# 46) was first to crystallize out of the parent melt (Al/Ti~>3) (Fig. 4c). As the melt evolved, plagioclase began to co-crystallize with pyroxene (Fe# 46 to Fe# 70) and the Al/Ti ratio decreased as Al was removed from the melt. An Al/Ti ratio of~1.5 indicates that Ti was removed from the melt and ilmenite co-crystallized with pyroxene and plagioclase, as the Al/Ti versus Fe# trend flattens out from Fe# 70 and Al, Ti, and Fe are all extracted from the melt. At this point, the parent magma would have had an Fe-enrichment allowing for a late-stage formation of symplectites. These trends support a common igneous origin for the basalt clast in MIL 13317 and suggest that the clasts are comagmatic, and that differences in texture must be caused by slightly different crystallization histories.

Implications for Ancient Lunar Volcanism
Based on ages of the Ca-phosphate phases in basalt Clast 4, and the ages determined for the basalt clasts analyzed by Snape et al. (2018), the basaltic material sampled by MIL 13317 is interpreted as having crystallized at~4330 Ma (Figs. 9-11; Table 3). Thus, VLT and low-Ti basaltic volcanism on the Moon recorded by lunar meteorites extended across 1.4 billion years, from~4330 to~2930 Ma (recorded by young lunar meteorite NWA 032: Borg et al. 2009). The ancient ages of MIL 13317 places it among the oldest suite of lunar basalt samples, including the Apollo 14 high-Al basalts (Taylor et al. 1983;Snyder and Taylor 2001), Apollo 17 KREEP basalts (Nyquist et al. 1975;Shih et al. 1992), and the high-Al, VLT lunar meteorite Kalahari 009 (Terada et al. 2007;Shih et al. 2008;Sokol et al. 2008;Snape et al. 2018) (Fig. 13c). Observations from the Apollo suite (e.g., Nyquist and Shih 1992) indicate a correlation between the age of the basalt and Ti-content, with older mare basalts (e.g., Apollo 11 basalts;~3900 to 3500 Ma) having higher Ti-content, and the younger basalts (e.g., Apollo 12,~3400 to 3200 Ma; Luna 24 3200 Ma) having more intermediate to lower Ti-content (Fig. 13a). Using the method of Robinson et al. (2012), we calculate the parent bulk-rock Ti-contents of basalt clasts in MIL 13317 to be between 0.14 and 0.32 wt% making it a VLT basalt (Fig. 13a), as well as having a high-Al content ( Fig. 13b; Table 1). MIL 13317 and another ancient high-Al, VLT lunar meteorite Kalahari 009 (Figs. 13a and 13b) indicate that the most ancient (>3800 Ma) mare and mare-like basalts in lunar meteorites tend to have some of the lowest Ti-contents and highest Al-contents of all lunar sampled basalts (Figs. 13a and 13b; see also Joy and Arai 2013). Further age dating of basaltic clasts in other lunar meteorites is needed to see if this trend is supported across a wider geographic distribution of lunar material. However, the observation of the most ancient lunar basalts being low in Ti (<3 wt% TiO 2 ) is also seen in coupled crater countingchemical remote sensing studies of mare basalt regions (e.g., Kodama and Yamaguchi 2005;Morota et al. 2011;Kramer et al. 2015;Sato et al. 2017). The discovery of new types of high-Al basalts suggests that this variety may be a more common basalt type than suggested by their relatively low abundances in the Apollo and Luna sample collection (Kramer et al. 2015).
Outstanding questions relate to how such early melting is associated with the timing of LMO ilmenitedriven density instabilities and overturn (e.g., Hess and Parmentier 1995;Elkins-Tanton et al. 2002), and the role of urKREEP (Apollo 14: e.g., Neal and Kramer 2006) or lack thereof (MIL 13317 and Kalahari 009: Terada et al. 2007;Snape et al. 2018) driving this earliest phase of magmatic activity. Terada et al. (2007) promoted impact-driven decompression melting as a possible mechanism of such ancient KREEP-poor volcanism. However, partial melting of the earliest olivine and pyroxene LMO cumulates during LMO overturn and adiabatic ascent is also feasible (Joy et al. 2008;Elkins-Tanton and Bercovici 2014).
The relationship between high-Al abundances in the earliest >4300 Ma volcanism is also intriguing. High-Al abundances in Apollo 14 basalts have been modeled via assimilation fractional crystallization, where the assimilant was a high-Al granitic component either added during ascent (Neal and Kramer 2006;Hallis et al. 2014) or via thermal erosion of the underlying regolith during the flow of lava across the lunar surface (Hui et al. 2011). However, both MIL 13317 and Kalahari basalts are KREEP-poor (Terada et al. 2007;Snape et al. 2018; this study), and the high-Al component instead could have been contributed from assimilation of ITE-poor ferroan anorthositic crust. There are challenges with assimilating such refractory materials into early lunar VLT/low-Ti mafic magmas (Finnila et al. 1994) during magma chamber storage (Neal et al. 1988;Head and Wilson 1992) or dyke conduit magma transport (Andrews-Hanna et al. 2013). For example, magmas must be at a higher temperature than the country rock it melts and assimilates: the melting temperature (solidus) of lunar plagioclase in highland rocks (~1250°C: Finnila et al. 1994) is higher than or similar to the liquidus of typical lunar basaltic VLT (~1180°C: Grove and Vaniman 1978) or low-Ti systems (~1273 to 1222°C: Grove and Bence [1977] and references therein). Thus, anorthitic FAN-like plagioclase dissolution by a mare basaltic melt might be thermodynamically challenging. Moreover, on the Moon, FAN host rock would melt to form a more viscous and denser crystal mush than the basaltic ascending fluid, inhibiting ascent and eruption (Finnila et al. 1994). However, the age overlap between ferroan anorthosite samples, the HMS, and ancient basaltic meteorite ages (Figs. 13b and 13c), suggests that the lunar highlands crust may still have been accumulating and was likely warmer and more ductile at ~4300 Ma (Elkins-Tanton and Bercovici 2014). Such elevated crustal thermal conditions may have facilitated dissolution (Reiners et al. 1995) of anorthitic plagioclase from the warm country feldspathic rock. This would allow an Al-rich assimilant to be available in ancient high-Al basaltic lava flows. Alternatively, the Hui et al. (2011) model of post-eruption assimilation by a lava flow thermally eroding the underlying KREEP-poor feldspathic glass-rich regolith might be an attractive alternative hybrid magma-surface interaction model. The nature of transport of basalt magmas through, and onto, hot and cold lunar crust should be explored to better understand ancient basalt petrogenesis (Finnila et al. 1994) and investigate effects of how crustal assimilation could mask isotopic signatures of mantle sources (Reiners et al. 1995) in such ancient basalts.

SUMMARY
Miller Range 13317 is a mingled regolith breccia of lunar origin comprised of fragments of basalts, impact melts, and feldspathic material (including clasts with an affinity to the Apollo FANs). It is one of a handful of lunar meteorites available that have sampled ancient (>3800 Ma) basaltic volcanism. It is distinct from any of the other lunar meteorites collected to date and there is no substantial evidence that MIL 13317 is paired with any other Th-rich lunar meteorite, although further investigation is needed on Lynch 002 to confirm this. As a result, MIL 13317 provides new perspectives on the diversity of rocks on the lunar surface, and in particular lunar volcanism through time. MIL 13317 shows a complex regolith history with a range in trapped and cosmogenic noble gas isotope concentrations. However, all parts of the meteorite have resided in the top millimeter of the lunar surface at some point in history, where trapped gases from the solar wind and lunar atmosphere were acquired. The ( 40 Ar/ 36 Ar) tr antiquity indicator suggests that compaction from the parent soil into breccia components occurred between 2610 AE 780 and 1570 AE 470 Ma, which closed the fragments off from further space weather processes. MIL 13317 then experienced exposure to the cosmic environment for a period of approximately 835 AE 84 Myr, where it resided in the regolith at an average depth of 30 cm.
Comparing Lunar Prospector data with the bulk rock FeO and Th content of MIL 13317, the meteorite was possibly launched from a regolith in a region close to the PKT. It is most compositionally similar (in terms of Th and FeO) to regoliths in the lunar near side around the northwest region of Oceanus Procellarum, the eastern edge of Mare Frigoris, and to the east and northeast of Mare Nubium.
The basaltic clasts in MIL 13317 provide evidence of ancient (~4340 Ma) VLT, high-Al basaltic volcanism that was not sampled by the Apollo missions, possibly derived from cryptomare deposits that were buried by subsequent lava flows and/or impact ejecta. These ancient ages suggest that basaltic volcanism, in particular high-Al, VLT basaltic volcanism, on the Moon started as early as 4350 Ma and before the bulk onset of lunar mare volcanism at~3850 Ma, providing further constraints on the earliest volcanic activity of the Moon.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article: Data S1. Analytical techniques. Fig. S1. Backscattered electron images of grains used in SIMS U-Pb dating work. Image directory and corresponding data for each grain analysis can be found in Table S3. Zirc = zircon; badd = baddeleyite; phos = phosphate.   Korotev and Ziegler 2014). MIL 13317 is also compared to another Th-rich meteorite Calcalong Creek (Hill and Boynton 2003) and Kalahari 009. b) FeO (wt%) versus Th (ppm). Plot shows a comparison of lunar meteorite (blue dots and markers) and Apollo and Luna soils (represented by the colored fields, Korotev 2018) with the fusion crust data for MIL 13317 (Table 1). Note the log scale for the Th concentrations. Fig. S6. Chips of Miller Range 13317,10 and Miller Range 13317,11. Subsplits from both chips were used for noble gas analyses. A polished block was made of MIL 13317,11b. MIL 13317,10 represents a regolith dominated chip and MIL 13317,11 represents a large brecciated basaltic clast. Table S1. Summary and types of analysis conducted on clasts in MIL 13317,7 and MIL 13371,11a. Table S2. Trace elements abundances (ppm) for clasts and matrix minerals and bulk compositions in MIL 13317,7. Errors are 1r SD from the averages of several measurements (number of analysis displayed in brackets with the mineral, where px = pyroxene and plag = plagioclase). Table S3. U-Pb isotope data for phases (zircon, baddeylites, apatites, merrillites) in MIL 13317,7 determined by SIMS analysis. Table S4. Plagioclase compositional summary data for MIL 13317,7 and MIL 13317,11a. Table S5. Olivine compositional summary data for MIL 13317,7. Table S6. Pyroxene compositional summary data for MIL 13317,7 and MIL 13317,11a. Table S7. Fusion crust and vein compositional summary data for MIL 13317,7. Table S8. Bulk values and averages of compositional data for fine-grained clast in MIL 13317,7. Table S9. Summary and raw data of Neon and Argon noble gas data for bulk rock chips of MIL 13317,10 and MIL 13317,11b. Raw data include blank corrected and mass spectrometer sensitivity corrected data. Errors are to 1r. Table S10. Summary and raw data of Xenon noble gas data for bulk rock chips of MIL 13317,10. Raw data include blank corrected and mass spectrometer sensitivity corrected data. Errors are to 1r. Table S11. Summary of endmember noble gas components, production rates, cosmic ray exposure ages, and antiquity ages in MIL 13317,10 and MIL 13317,11b.