SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ? NUMBER 22 Mineral Sciences Investigations 1976-1977 Robert F. Fudali EDITOR SMITHSONIAN INSTITUTION PRESS City of Washington 1979 ABSTRACT Fudali, Robert F., editor. Mineral Sciences Investigations 1976-1977. SmithsonianContributions to the Earth Sciences, number 22, 73 pages, 22 figures, 20 tables, 1979.?This volume is comprised of six short contributions reporting the resultsof some of the research carried out by the Department of Mineral Sciences, Smithsonian Institution, during the period 1976-1977. Included are: a comparisonof impact breccias and glasses from Lonar Crater (India) with very similar speci- mens from the moon; petrographic descriptions and chemical analyses of virtuallyall the known pyroxene-plagioclase achondrite meteorites and a discussion of the relationships within this class; a comparative chemical study of sixty Australiantektites from widely separated localities; a description of a new, rapid technique of sample preparation for whole-rock analyses using the electron microprobe; aninterlaboratory comparison of the precision and accuracy of electron microprobe analyses; and a tabulation of the chemical compositions of some electron micro-probe reference samples. OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recordedin the Institution's annual report, Smithsonian Year. SERIES COVER DESIGN: Aerial view of Ulawan Volcano, New Britain. Library of Congress Cataloging in Publication Data Main entry under title:Mineral sciences investigations, 1976-1977. (Smithsonian contributions to the earth sciences ; no. 22)"Six short contributions reporting the results of some the research carried out by the Smithsonian's Department of Mineral Sciences."Bibliography: p. 1. Mineralogy?Addresses, essays, lectures. I. Fudali, Robert F. II. National Museum of NaturalHistory. Dept. of Mineral Sciences. III. Series: Smithsonian Institution. Smithsonian contribu- tions to the earth sciences ; no. 22.QE1.S227 no. 22 [Q364] 560'.943 [549] 78-24474 Contents Page METEORITES PETROLOGY, MINERALOGY, AND DISTRIBUTION OF LONAR (INDIA) AND LUNAR IMPACT BRECCIAS AND GLASSES, by Kurt Fredriksson, Phyllis Brenner, Ananda Dube, Daniel Milton, Carol Mooring, and Joseph A. Nelen ... 1 CHEMICAL VARIATION AMONG AUSTRALIAN TEKTITES, by Brian Mason 14 THE PYROXENE-PLAGIOCLASE ACHONDRITES, by Brian Mason, Eugene Jarosewich, and Joseph A. Nelen 27 ANALYTICAL LABORATORY FUSION OF ROCK AND MINERAL POWDERS FOR ELECTRON MICROPROBE ANALY- SIS, by Peter A. Jezek, John M. Sinton, Eugene Jarosewich, and Charles R. Obermeyer 46 MICROPROBE ANALYSES OF FOUR NATURAL GLASSES AND ONE MINERAL: AN INTERLABORATORY STUDY OF PRECISION AND ACCURACY, by Eugene Jarosewich, Alan S. Parkes, and Lovell B. Wiggins 53 ELECTRON MICROPROBE REFERENCE SAMPLES FOR MINERAL ANALYSES, by Eugene Jarosewich, Joseph A. Nelen, and Julie A. Norberg 68 in Mineral Sciences Investigations 1976-1977 Petrology, Mineralogy, and Distribution of Lonar (India) and Lunar Impact Breccias and Glasses Kurt Fredriksson, Phyllis Brenner, Ananda Dube, Daniel Milton, Carol Mooring, and Joseph A. Nelen ABSTRACT Chemically and mineralogically, the shock meta-morphosed breccias and melt rocks at Lonar Crater (India) are the closest terrestrial analogs to theimpact-generated rocks and soils of the moon. Com- parative studies reveal a wide range of virtuallyidentical forms and textures between Lonar and lunar samples. Thus, the advantage of knowing theunshocked target rocks at Lonar, especially with re- gards to alternating layers of different competency, may yield invaluable insights into lunar impactprocesses. Also, chemical differences reported herein between Lonar basalts and impact glasses suggestsome probable trends in lunar analogs. Conversely, the isochemical nature of much of the Lonar glass and basalt demonstrates that lunar glasses may oftenbe chemically equated to their parent rocks. Introduction As the importance of meteorite impacts as a mechanism producing lunar land forms, rocks, and Kurt Fredriksson, Phyllis Brenner, Joseph Nelen, Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Ananda Dube, Geological Survey of India, Calcutta 13, India. Daniel Milton, United States Geological Survey, National Center, Reston, Virginia 22092. Carol Mooring, Department of Ge- ology, University of Wisconsin, Madison, Wisconsin 53706. soils is more clearly recognized, the existence of Lonar Crater, India, is coming to be seen as one of earth's more fortuitous catastrophes. This impact crater is located in the Deccan Trap basalts of India (19?58'N, 76?31'E), and affords unique op- portunities for earthbound scientists to study ana- logs to lunar processes. We report herein a com- parative study of the petrography, morphology, and chemistry of shocked Lonar and lunar materials. ACKNOWLEDGMENTS.?We are indebted to Walter Brown, Andrea Eddy, Becky Fredriksson, and Julie Norberg for technical and/or editing assistance. Support from the Geological Survey of India, the Smithsonian Research Foundation and Foreign Cur- rency Program is gratefully acknowledged. One of us (DM) received partial support under a NASA contract. We also thank Laurel Wilkening and Hans Suess for access to their unpublished radio- metric data from Lonar. General Geology The Lonar crater is an almost circular depression, 1830 m in diameter, approximately 150 m deep. Lonar Lake, a shallow saline lake, occupies most of the crater floor. The rim of the crater is raised approximately 20 m above the surrounding plain. A smaller circular depression in the traps, 300 m in SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES diameter and about 700 m north of Lonar Crater, apparently is a second crater, probably excavated after throwout from the main crater landed, but during settling of fallout. Various data on general morphology, conclusive evidence for the impact origin, and preliminary studies of petrography, stratigraphy, and the chemistry of the impact rocks have been published elsewhere (Fredriksson et al., 1973). Recent fission track dating of Lonar shock- melted glasses indicates an age less than ~ 50,000 years (Wilkening, pers. comm.). Carbon-14 dating shows an age of more than 30,000 years for organic material recovered 50 m down in one of the ~90 m thick sediment columns in the lake (Suess, pers. comm.). The crater's age then lies between 30 and 50 thousand years. The Geological Survey of India has drilled a series of five boreholes approximately on a line trending NE-SW in the floor of the crater. Each drillhole encountered 90 to 100 m of lake sediment, which was found to contain small amounts of im- pact glass in the form of fragments and spherules, and also rounded lithic fragments exhibiting vary- ing degrees of shock metamorphism. This material was probably eroded from the crater rim. After penetrating the sediment, each of the drillings re- turned cores of coarser breccia apparently composed of blocks up to meters in size which are unshocked or slightly shocked and contain crude shatter cones. Beneath the coarse breccia, the first four drillings encountered a layer that yielded essentially no core recovery, apparently because it was composed of unconsolidated to extremely friable microbreccia. The assumption was confirmed in the fifth hole for which an improved core catcher was used, resulting in approximately 100% core recovery. In addition, one deep (~100 m) and several shallower holes were drilled in the smaller crater and on its rim. Current evidence indicates that the structure, although arti- ficially modified by former dams and ditches, for purposes of irrigation, is of impact origin. Comparative Petrography, Morphology, and Chemistry The Deccan Trap basalts at Lonar exhibit two main textural varieties: (1) "hard" trap which has euhedral plagioclase phenocrysts with strong oscil- latory zoning, often glomero-porphyritic, in an aphanitic groundmass; and (2) medium to coarse- grained, slightly to moderately vesicular or amygda- loidal basalt which has weathered to "soft" trap. Major minerals in both textural types are plagio- clase (An 50-70) 33%-48%, clinopyroxene (augite ? pigeonite) 17%?34%, and opaques (ilmenite and/or magnetite) 2%-12%. Olivine may be pres- ent in small amounts. Alteration minerals include chlorite minerals, iron oxides, serpentine, and epi- dote. In addition, within both groups more coarse- grained sub-ophitic basalt may contain 20%-30% interstitial fresh glass or palagonite. In contrast, lunar basalts range in grain size from coarse to fine to vitrophyric, and are often vesicular. In coarse lunar basalts pyroxene may occur as poiki- litic grains in plagioclase. Plagioclase constitutes 25%-3O%, clinopyroxene 45%-55%, and opaques 15%-2O% (including ilmenite, armalocolite, chrome- spinel, ulvospinel, rutile, metallic iron, nickel-iron, and troilite). Minor olivine may also be present. Many of the observed differences between Lonar and lunar impact-produced rocks and glasses dis- cussed below may be explained by the chemical differences between the target basalts (see Table 1). Shockwaves generated by a hypervelocity impact cause a variety of deformations and changes in the targets. Thus, slight shocking of Lonar basalt (<250 kb) produces minor cataclasis in plagioclase and pyroxene, wavy extinction in pyroxene, and shock-induced twinning in plagioclase. Increasing pressures within this range produce closely spaced twins in pyroxene (this twinning seems to be related to possible strong shear loci). Large plagioclase and pyroxene crystals also tend to recrystallize or granu- late into smaller domains, optically independent. In some cases planar elements in plagioclase are also produced, often accompanied by strong cataclasis, warping of the grain, and reduced birefringence. In "moderately" shocked basalt (~250-450 kb), plagioclase is converted to disordered maskelynite in which the original oscillatory zoning is preserved. Moderately strong shocking (~450-600 kb) produces microbreccia clasts in which maskelynite exhibits incipient flow. With intense shock, (>800 kb), the basalt may be converted to glass. The mechanisms involved and necessary pressures for these various transformations with special reference to Lonar basalts have been studied in great detail by Kieffer et al, 1976, and Schaal and Horz, 1977. Ejecta on the rim of the main Lonar crater con- sist of two contrasting types of debris. The lower NUMBER 22 TABLE 1.?Chemical compositions of some lunar and Lonar basalts; Lonar tabulations are electron microprobe analyses, done by this laboratory, using powders fused to a glass with lithium tetra- borate flux Apollo 17 PET, 1973:659-692. -Average of 9 basalt samples. All Fe as FeO; average Fe2O3/(FeO + Fe2O3) = 0.42, Constituent SK>2 A12O3 FeO MgO CaO Na2O K2O TiO2 75055,6 41.27 9.75 18.24 6.84 12.30 0.44 0.09 10.17 Apollo 17a 70035,1 37.84 8.85 18.46 9.89 10.07 0.35 0.06 12.97 70215,2 37.19 8.67 19.62 8.52 10.43 0.32 0.04 13.14 Average'3 Lonar basalt 50.13 13.69 14.09c 5.80 10.20 2.52 0.55 2.72 Lonar LNR shocked basalt 50.35 15.08 12.74C 5.99 11.55 1.70 0.25 2.31 unit ("throw-out") is crudely stratified and shows no evidence of shock. The overlying unit ("fallout") contains clasts from different bedrock units, Fladle- like bombs, glass spherules, and fragments. The lithic clasts show varying degree of shock from barely perceptible through medium to (rarely) in- tense. As a practical matter the ejecta can be as- signed to only two bedrock units representing the two basalt types described above: (1) hard, fresh trap exposed in the upper 50 m of the crater wall; and (2) soft, weathered trap exposed in the lower 50 m which presumably continues downward almost to the lake level and is probably followed by an- other dense basalt. Ejecta patches dominantly of one or the other unit are juxtaposed throughout the ejecta blanket with, at best, a weak tendency to- ward inverted stratigraphy. The patches vary widely in size, but are characteristically on the order of 100 m across near the rim and decrease to several meters across near the outer edge. The hard trap typically forms breccia of small angular clasts; the soft trap is in large masses that have much less internal disruption. Concentrations of ejecta relatively abundant in glass and shocked fragments were found up to ~2 crater radii from the rim expecially to the E and ESE. The ejecta blanket does not feather out evenly at the margin, but breaks into clumps increasingly scattered in ejecta-free terrain. Some of these are composites of hard and soft trap or even polymict breccias which, surprisingly, cohered during a flight of two or more kilometers. Some clumps remain in- tact; others are spread out inside secondary craters they have excavated. By analogy, secondary crater- ing on the moon depends neither on ejection of large individual blocks nor absence of atmosphere (Milton and Dube, 1977). In Figures 1-6 the similarities between Lonar and lunar shocked basalts and breccias are illus- trated. Lithic clasts in the Lonar microbreccias may exhibit severe cataclasis without reduction of birefringence in the plagioclase, or may show little or no cataclasis and still contain plagioclase com- pletely converted to maskelynite. Deformation of plagioclase in lunar breccias is directly analogous to that found in Lonar samples including cataclasis, increased density of twin lamellae, granulation or development of a mozaic texture, development of SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 1.?a, Lonar trench sample LRT 18VII-1, a medium-grained basalt with plagioclase totally converted to maskelynite; pyroxenes exhibit moderate cataclasis; section is ~0.9 mm in length, b, Apollo 17, #79155,68, a coarse-grained basalt with plagioclase partially converted to maskelynite (lower right); pyroxene exhibits moderately strong cataclasis; length of section is ?3 mm. FIGURE 2.?a, Lonar core sample LNR-2-317, ~460 m below crater rim; note contact between coarse-grained ophitic basalt and area of strong brecciation; section is ?0.9 mm in length. b, Apollo 17, #78155,7; note contact between coarse-grained plagioclase and pyroxene crystals and strongly brecciated area; section is ~0.9 mm in length. planar features and conversion to maskelynite. Lunar pyroxenes exhibit close-spaced twins, granu- lation, and wavy extinction as do their Lonar counterparts ( Figures 1, 2). Lonar microbreccias from both cores and trenches consist of lithic and crystal fragments embedded in a generally fine semi-opaque rock flour (Figures 3-5). The lithic fragments vary in size from several centi- meters to approximately 0.2 mm. Small lithic frag- ments tend to be more abraded than large frag- ments. Lithic fragments in a single thin section may all exhibit similar shock characteristics or may ex- hibit a wide range of shock features (e.g., a slightly shocked fragment juxtaposed to a moderately shocked one). Fragments may also be very similar in texture and grain size ("monomict" breccia) or ex- tremely dissimilar ("polymict" breccia) within one thin section. Occasionally, abraded breccia frag- NUMBER 22 FIGURE 3.?a, Lonar core sample LNR-2-319 is a microbreccia with angular to subangular crystal fragments and rounded lithic fragments; length of section is ~3.2 mm. b, Apollo 17, #73235,63, also with angular to subangular crystal fragments and rounded lithic fragments; length of section is ?3.6 mm. FIGURE 4.?a, Lonar core sample LNR-3-300, ~440 m below crater rim, has abraded lithic frag- ments in fine-grained rock flour; fragment at lower edge has plagioclase partially transformed to maskelynite; section is 3.4 mm in length, b, Apollo 16, #68822,1; note abraded lithic fragments in fine-grained rock flour and partially devitrified glass spherule; section is 0.9 mm in length. merits are observed along with normal lithic frag- ments indicating a process of brecciation, consolida- tion, abrasion, and final disposition. Apparently, one impact can produce a kind of the breccia-within- a-breccia texture so common in lunar samples (Fig- ure 5). One notable lithic fragment type consists almost entirely of plagioclase; similar lunar "an- orthosite" breccia clasts may, thus, not always rep- resent individual rock types. Microbreccia rock flour matrix material often shows evidence of flow, and may contain turbid shock glass (Figure 6). Crystal fragments in the matrix are usually angular to sub- angular and exhibit shock effects ranging from mild to moderately strong. The unconsolidated variety of microbreccia appears to contain little or no rock flour matrix, a major component of consolidated microbreccia, and instead, consists almost entirely of small crystal fragments of a fairly uniform size. Shock effects in these fragments are generally slight, occasionally moderate. Lithic and breccia fragments SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 5.?a, Lonar core sample, LNR-2-319A, with abraded lithic fragments in fine-grained rock flour; note high degree of crystal angularity in matrix contrasting with roundness of the lithic fragments; length of section is 3.4 mm. b, Apollo 16, #67943,1-J; note abraded lithic and breccia fragments in fine-grained rock flour; length of section is 1.7 mm. FIGURE 6.?a, Lonar core sample LNR-2-138, ~280 m below crater rim, is a layered microbreccia; light areas are plagioclase crystal fragments; section is 3.4 mm in length, b, Apollo 17, #72215,7, also a layered microbreccia; note subangular breccia fragment; section is 3 mm in length. are small and uncommon. Crude stratification sug- gests flow or sedimentation rather than brecciation in situ. The material found in core 5 between ^200 and 300 m below the lake level ("no core" zone in pre- vious holes) is indeed best characterized as micro- breccia, although it is mostly unconsolidated. Mineralogically it resembles the microbreccias de- scribed in the previous paragraph. Occasionally more indurated (whether by shock or later dia- genetic processes is uncertain) sections of core (up to tens of centimeters) are found. Thin sections show these breccias to resemble simple lunar soil breccias and to contain unshocked to moderately shocked minerals and rock fragments. Lonar glass ejecta consist of bombs, spherules, and fragments. Black glassy bombs, evidently still soft when deposited, are 10-15 cm in diameter and gen- erally have a dense outer shell and a vesicular core. Spherules, millimeters in diameter and smaller, and similar in appearance to microtektites, consist of droplets of brown to colorless glass, often with flow NUMBER 22 FIGURE 7.?a, Part of a glassy teardrop from Lonar trench sample LRT 18IVB-3 #2. SEM photo; analysis in Table 3; note microcrater and mound of glass; sample is ~2 mm long, b, Apollo 17, #74220,86 (>60 mesh) #1, a teardrop-shaped spherule; note internal structure; SEM photo; sample is 0.5 mm long. FIGURE 8.?a, A cone-shaped fragment from Lonar trench sample LRT 18VA-4 #1; SEM phc.o; note mound of glass, probably extruded upon cooling; sample is 1.3 mm long, b, Apollo 17, #74220,86 (>100 mesh) #1, a cone-shaped fragment; SEM photo; analysis in Table 4; saaple is 0.6 mm long. banding indicated by contorted schlieren. Feathery recrystallization in the spherules occurs rarely. Though most Lonar spherules seem to be compo- sitionally homogeneous, several were found with compositional zoning. The trends most often ob- served include progressive alkali depletion (partic- ularly Na2O) accompanied by FeO enrich nent. from the center to the edge of the spherules. Glass shards and angular fragments mostly consist of turbid, brown to colorless glass, often vesicular and contain- ing contorted schlieren. The production of shock melts requires intense shocking, >800 kb (Schaal and Horz, 1977), and yet glass shards frequently in- clude crystal fragments or rarely entire lithic frag- ments that are only slightly shocked. Lunar glass spherules and fragments have been studied by a number of authors (see especially Carter, 1971). Figures 7-10 show some characteristic SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES FIGURE 9.?a, A spherule from Lonar trench sample LRT 18IIIA-4 #1; SEM photo shows small mound of glass possibly extruded as spherule cooled; analysis in Table 3; spherule is 1 mm in long dimension, b, Thin section of spherule shown in a; note apparent inhomogeneity of internal structure; the glass, however, has almost constant (basaltic) composition; the skeletal crystallites decorating flow lines are titanomagnetite; length of section is 1 mm. c, Apollo 17, #74220,86 (]>100 mesh); SEM photo illustrates similarities in surfaces features of lunar and Lonar spherules (also of many meteoritic chondrules); sample is 0.4 mm long, d, A thin section of sample shown in c reveals a partially devitrified texture. types illustrating the similarities between Lonar and lunar glass particles. A few chemical analyses are given in Tables 2-4. A large number of electron probe analyses were made to determine trends in composition among Lonar basalts, "bombs," and the other glass par- ticles. The glasses were subdivided into two some- what arbitrary groups based on composition and/or morphology: (1) "normal" glasses (Table 2), the compositions of which were comparatively uniform and resemble the basalts; and (2) "selected" glasses (Table 3) mostly with compositions deviating from the "normal" glasses. Observed trends were various degrees of alkali, water, and sometimes silica deple- tion in the spherules and some "unusual" glasses. The majority of analysed glasses, however, closely approximate a homogeneous basalt melt although they appear somewhat depleted in FeO + MgO and enriched in A12O3 relative to the parent basalt (Figure 11, Tables 1, 2). In addition, ferrous iron is more prevalent in the glasses, indicating a reduction upon melting of the parent basalt NUMBER 22 FIGURE 10.?a, Almost perfectly round spherule from Lonar trench sample LRT 18VA-4 #2; SEM photo; note microcrater; spherule is 0.8 mm in diameter, b, Thin section of Lonar trench sample LRT 18IIA-3 #3, a spherule similar to that shown in a; analysis in Table 3. c, Apollo 17, #74220,86 (>60 mesh), a spherule; SEM photo; note microcrater; sample is ~0.43 mm in diameter, d, Thin section of spherule shown in c, showing almost complete devitrification; length of section is 0.25 mm. (Fe2O3/(FeO + Fe2O3) is -0.42 in the basalts, -0.37 in the glasses). According to Morgan (1978) the heavy volatile elements Se and Cd are also signifi- cantly depleted in the glasses. Lunar trends, though inherently more difficult to discern (e.g., initially low alkalies) seem also to indi- cate a depletion in alkalies in glasses relative to the parent basalts (Fredriksson et al., 1971). As at Lonar some heavy volatile elements, such as Pb (Silver, 1975), may also have been redistributed in the lunar rocks and soils by impact volatilization. If lunar fractionation trends are indeed analogous to Lonar trends (even ignoring the "unusual" Lonar glasses which were likely produced by complete or partial melting of individual phenocrysts) then lunar glass fragment chemistry may or may not represent un- altered lunar rock types. Interpretation of impact cratering and ejecta stra- tigraphy has been attempted throughout the Apollo missions. The results of the present study are cer- tainly relevant to such questions as the extent, homogeneity, and composition of lunar ejecta blan- 10 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 2.?Chemical compositions of Lonar "normal" glasses; electron microprobe analyses, this laboratory Constituent SiO2 AI2O3 FeOa MgO CaO Na20 K20 TiO2 LRT 19VB-1 52.02 13.84 13.58 5.43 10.27 2.30 0.63 2.69 LRT 19VB-1 51.65 14.89 12.76 5.12 9.94 2.56 0.53 2.09 LRT 19VB-1 52.61 14.36 12.91 4.95 10.40 1.99 0.47 2.27 D 215-4 50.48 13.99 12.96 5.46 10.13 2.59 0.53 2.23 SL-3 1.5-4 51.59 14.53 13.45 4.85 9.97 2.32 0.46 1.90 All Fe as FeO; average Fe2O3/(FeO + Fe2O3 ) = 0.37. TABLE 3.?Chemical composition of Lonar "selected" glasses (note that LRT 18IVB-3 is a "normal" glass); electron microprobe analyses, this laboratory Constituent SiO2 A12O3 FeOd MgO CaO Na2O K20 TiO2 LRT 19VIB-3 37.69 22.22 28.41 4.82 6.96 0.02 0.02 0.10 LRT 18HA-3a 49.1 12.8 16.5 7.24 9.0 1.76 0.28 2.34 LRT 18111A-4 b 47.79 12.69 15.88 7.67 9.13 1.49 0.23 2.29 LRT 19VB-1 51.81 5.27 18.30 11.64 12.13 0.62 0.29 1.95 LRT 181VB-3C 51.78 13.43 13.55 5.08 9.41 2.31 0.65 2.38 aSee Figure 10b. b See Figure 9a,b. See Figure 7JI. 'All Fe as FeO. kets. Thus the mechanisms involved in reworking the lunar regolith (e.g., Gault et al., 1974) may deserve reconsideration. Secondary impacts played an important role in mixing and reworking pre- existing Lonar soils with ejecta of both basalt frag- ments and loosely consolidated microbreccias. This should be considered in the interpretation of the evolution of the lunar regolith. Secondary crater- ing may also explain the discrepancy between the long exposure ("suntan") ages of lunar soil relative to excavated rock fragments (Gold and Williams, 1974). Detailed descriptions and maps of the Lonar ejecta distribution are being prepared by the Geo- logical Survey of India in cooperation with some of us, and will be published elsewhere. Some findings near the bottom of the fifth and last drillcore may justify also a reevaluation of current concepts of the impact effects in rock com- NUMBER 22 11 TABLE 4.?Chemical compositions of selected lunar soil and glasses (Apollo 17); electron micro- probe analyses, this laboratory; bulk soil analyses derived from 20 mg splits fused to a glass with lithium tetraborate flux (A, B, C, and D designate individual glass analyses that appear to be representative of distinctive groupings; it appears that the B glasses generally correspond to the local bulk soils, but the other groupings are not readily identifiable) Constituent SiO2 A12O3 FeO MgO CaO Na20 K2O TiO2 Constituent SiO2 AI2O3 FeO MgO CaO Na20 K20 TiO2 Constituent SiO2 AI9O3 FeO MgO CaO Na20 K20 TiO2 Bulk Soil 41.4 12.8 16.3 9.9 10.6 0.28 0.15 7.2 Bulk Soil 41.2 13.0 15.7 9.9 10.7 0.39 0.10 6.4 Bulk Soil 42.0 15.1 14.3 10.0 11.1 0.44 0.20 4.7 700019,72 A 38.1 6.1 24.4 14.2 7.2 0.30 0.22 8.7 A 39.4 6.2 22.7 14.6 7.6 0.32 0.09 9.0 A 38.6 5.8 22.4 14.7 7.4 0.42 0.10 9.1 B 40.1 12.7 16.8 10.1 10.8 0.34 0.11 7.9 79035,29 B 41.9 13.8 15.8 10.0 11.3 0.33 0.11 7.4 79135,71 B 42.0 13.6 14.3 10.5 10.8 0.40 0.06 6.0 C 41.4 22.0 6.4 14.7 13.1 0.11 0.03 1.20 C 45.2 21.5 8.1 11.7 13.3 0.04 0.01 1.16 74220,86 A 39.9 6.0 22.3 14.8 7.1 0.33 0.10 8.8 D 44.0 36.0 0.1 0.1 20.1 0.25 0.02 0.03 D 43.6 35.5 0.1 0.3 19.6 0.24 0.08 0.02 plexes with layers of different competency (chemi- cally similar or not). At ~450 m below the rim solid basalt was encountered as in previous drillings. At about 470, 490, and 520 m, however, rock flour layers apparently up to ~1 m thick were encoun- tered between several meters of solid basalt with essentially no shock features. The rock flours in these horizons differ in grain size but are all rela- tively well sorted. Whether these loose layers repre- sent a peculiar kind of shock veins or injected ma- terial or explosively disrupted porous or weathered basalt layers of low competency is unclear. Ongoing studies should yield a better understanding of this and also give more realistic comparisons (scaling factors) between real, large impacts and artificial ex- periments on small targets (e.g., Gault et al., 1974). 12 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES t. 25 24 -23 22 21 20 O 18 G)(VI ? Spherules A Glass bombs o Normal glasses a Selected glasses X Basalts ? Shocked basalt x* x D o X X * a Ao "o A X ?A A a A Xo o a ?75%), showed a greater variation in refractive index, sometimes as much as ?0.003. Although Glass (1970) reports large compo- sitional variations (SiO2 60%-85%) within a single australite, it must be noted that he specifically se- lected small, rare inclusions (two high-silica and three low-silica in two thin sections) to establish this extreme range, and also that he states (p. 241), "The bulk of the glass has an average composition of about 76% SiO2." The polished thin sections of the individual aus- tralites were analyzed with an ARL-SEMO electron microprobe using an operating voltage of 15 kV and a sample current of 0.15 /xA; a broad beam (approximately 0.05 mm diameter) was used to counter the local inhomogeneities discussed in the preceding paragraph. A previously analyzed aus- tralite containing 68.35% SiO2 (Chalmers et al., 16 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 1976, tbl. 8) was used throughout as a reference sample. Measured intensities were adjusted by com- puter using Bence-Albee factors. The following components were determined: SiO2, A12O3, FeO (all Fe reported as FeO) MgO, CaO, K2O, Na2O, TiO2, MnO; MnO is not reported in the analyses, since it was uniformly at or near the lower limit of measurability, approximately 0.1%. Results The analytical results are set out in Table 1 and are plotted in Figures 2, 3. The results for specific locations are discussed below. EARAHEEDY.?The collection from this locality comprised 97 specimens. The specific gravity range was from 2.432 to 2.459, most specimens (87) falling in the range 2.455-2.459. The four specimens ana- lyzed (Table 1, columns 1-4) cover the full specific gravity range, which is relatively small; the analyses show a correspondingly small composition range. LAKE YINDARLGOODA.?The collection from this locality near Kalgoorlie comprised 61 specimens. The specific gravity range was from 2.432 to 2.467, with practically all the specimens (57) having spe- cific gravities between 2.450 and 2.460. The three specimens analyzed (Table 1, columns 5-7) cover almost the full specific gravity range. A considerable number of analyses of australites from the Kalgoorlie region have been published, most of them by Taylor (1962) and Taylor and Sachs (1964). These analyses confirm the restricted range of australite composition in this region. Of Taylor's 13 analyses, 11 have SiO2 contents ranging from 70.34% to 72.13%; the remaining two have 75.17% and 77.39% respectively. HUGHES.?This collection, from the Nullarbor Plain close to the South Australia-Western Aus- tralia border, comprised 248 specimens, ranging in specific gravity from 2.398 to 2.473 (two specimens, described in columns 9 and 11 of Table 1, had in- terior bubbles and their specific gravities were less than 2.398). Of the total collection of 248 specimens, 203 had specific gravities between 2.440 and 2.460. The three collections from Hughes, Lake Yin- darlgooda, and Earaheedy all show very similar specific-gravity distribution patterns, with a single peak at 2.45-2.46 on the frequency curve. The largest collection, that from Hughes, shows the greatest range in specific gravity, as might be ex- pected. Chapman et al. (1964) have shown that this specific-gravity distribution pattern is uniform for all collections they examined from the Kalgoorlie and Nullarbor Plain locations. The Earaheedy col- lection shows that this uniformity extends far to the north. Within this vast area australites with com- parable SiO2 content generally show a close cor- respondence in other components; compare for example in Table 1, analyses 1 and 5, 3 and 9, 7 and 12. CHARLOTTE WATERS.?A small collection of 33 specimens from Charlotte Waters showed a wide range in specific gravity, 2.390-2.483, with a rather ill-defined frequency peak at 2.440. Chapman et al. (1964), who examined a much larger collection (420 specimens), found a specific-gravity range of 2.395-2.495, with a peak at 2.445. The analyzed specimens (Table 1, columns 14-20) show SiO2 ranging from 69.0% to 77.6%; a significant feature is a relatively high CaO content compared to aus- tralites of similar SiO2 content from the Nullarbor Plain and Western Australia. Taylor (1962) and Taylor and Sachs (1964) published six analyses of australites from Charlotte Waters; SiO2 ranged from 71.26% to 75.10%, CaO from 3.11% to 5.49%, figures similar to those obtained in this investi- gation. DURRIE.?This collection is of considerable in- terest, since it is closest to the northern and eastern margin of the strewnfield. The 99 specimens in the collection ranged in specific gravity from 2.384 to 2.489, with a mean of 2.440. Six specimens covering the full range of specific gravity were analyzed (Table 1, columns 21-26), and show a range of SiO2 from 69.2% to 78.9%. The compositions are similar to those of Charlotte Waters, but extend to a higher SiO12 content; it is significant that the low-SiO2 australites from Durrie also show the high CaO contents of Charlotte Waters australites (cf. analyses 14 and 21). MACUMBA.?This small collection (19 specimens) is of particular interest because it has a wide range of specific gravity (2.386-2.461), and the specific- gravity distribution plot shows two peaks, one at 2.395 and one at 2.440 (Chalmers et al., 1976, tbl. 7). This two-peaked specific-gravity distribution plot is characteristic of australite collections from the Lake Torrens-Lake Eyre region of South Australia and is not found elsewhere. The six specimens analyzed (Table 1, columns 27-32) cover an SiO2 range of NUMBER 22 17 TABLE 1.?Analysis of 60 selected australites for major constituents and for specific gravity and refractive index (locality of analyzed specimens, by column: 1-4, Earaheedy, USNM 5796; 5-7, Lake Yindarlgooda, USNM 5733; 8-13, Hughes, USNM 2548; 14-20, Charlotte Waters, USNM 2537; 21-26, Durrie, USNM 5799; 27-32, Macumba, USNM 2535; 33-38, Pine Dam, USNM 5802: 39-45, Mannahill, USNM 4831; 46-51, Lake Wilson, USNM 2534; 52-56, Mount Davies, USNM 2536; 57-60, Pindera, USNM 2242) Constituent SiO2 AI2O3 FeOa MgO CaO K20 Na2O TiO2 Total SG n Constituent SiO2 A12O3 FeO MgO CaO K2O Na20 TiO2 Total SG n 1 69.9 14.5 5.24 2.34 3.24 2.74 1.54 0.80 100. 3 2.459 1.517 11 73.3 12.0 4.64 2.04 3.52 2.74 1.51 0.71 100.5 1.508 2 70.2 14.2 5.01 2.16 2.51 2.77 1.74 0.84 99.4 2.452 1.514 12 74.3 12.0 4.42 1.81 2.23 2.39 1.10 0.73 99.0 2.410 T.504 3 71.8 12.8 4.54 2.15 3.70 2.29 1.14 0.72 99.1 2.445 1.513 13 75.2 11.0 4.27 1,80 2.79 2.71 1.35 0.65 99.8 2.398 1.500 4 73.0 12.9 4.60 2.02 3.14 2.63 1.37 0.78 100.4 2.432 1.511 14 69.0 12.9 4.78 2.29 5.93 2.34 1.12 0.74 99.1 2.483 1.522 5 69.9 14.1 5.06 2.31 3.36 2.49 1.32 0.80 99.3 2.464 1.517 15 70.8 13.3 4.74 2.22 3.80 2.43 1.26 0.77 99.3 2.455 1.514 6 70.6 14.0 4.99 2.31 3.15 2.48 1.35 0.79 99.7 2.454 1.514 16 73.3 10.9 4.11 1.86 4.40 2.40 1.20 0.64 98.8 2.442 1.512 7 74.3 11.4 4.02 1.69 2.89 2.44 1.27 0.67 98.7 2.432 1.509 17 73.6 10.6 4.16 1.93 5.03 2.37 1.23 0.57 99.5 2.439 1.512 8 70.2 13.4 4.93 2.39 4.63 2.62 1.42 0.78 100.4 2.467 1.518 18 74.3 10.1 3.98 1.77 5.00 2.35 1.19 0.59 99.3 2.429 1.508 9 71.7 12.6 4.64 2.14 3.63 2.62 1.43 0.68 99.4 -- 1.515 19 75.6 9.95 3.91 1.70 4.19 2.39 1.16 0.54 99.4 2.417 1.506 10 72.4 12.7 4.55 1.98 3.20 2.71 1.43 0.70 99.7 2.445 1.513 20 77.6 10.6 3.65 1.58 2.17 2.23 1.08 0.60 99.5 2.390 1.499 18 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Continued. Constituent SiO2 A12O3 FeO MgO CaO K20 Na2O TiO2 Total SG n Constituent SiO2 A12O3 FeO MgO CaO K20 Na20 TiO2 Total SG n 21 69.2 12.9 5.26 2.42 5.79 2.36 1.08 0.74 99.8 2.489 1.522 31 76.0 10.1 3.76 1.64 3.31 2.34 1.20 0.59 98.9 2.406 1.504 22 70.3 12.5 4.72 2.28 5.81 2.41 1.28 0.69 100.0 2.473 1.519 32 77.3 9.48 3.55 1.46 3.10 2.33 1.20 0.53 99.0 2.398 1.503 23 71.7 13.1 4.90 2.10 4.05 2.40 1.20 0.72 100.2 2.452 1.514 33 68.9 13.8 5.10 2.35 5.19 2.60 1.43 0.74 100.1 2.476 1.520 24 73.1 12.8 4.73 2.04 3.12 2.45 1.31 0.74 100.3 2.429 1.510 34 70.1 13.4 4.91 2.26 4.18 2.54 1.36 0.76 99.5 2.466 1.518 25 76.1 10.9 4.44 1.65 2.40 2.55 1.15 0.68 99.9 2.404 1.504 35 72.0 12.9 4.74 2.21 4.05 2.45 1.28 0.71 100.3 2.454 1.514 26 78.9 10.0 4.08 1.49 1.89 2.36 1.00 0.59 100.3 2.384 1.499 36 74.2 12.0 4.52 1.94 3.06 2.25 1.09 0.57 99.6 2.416 1.507 27 70.9 10.8 4.42 2.03 5.91 2.44 1.27 0.58 98.4 2.461 1.515 37 76.3 10.9 4.42 1.70 3.00 2.55 1.25 0.66 100.8 2.400 1.502 28 71.4 12.9 4.65 2.12 3.51 2.44 1.37 0.77 99.2 2.454 1.514 38 78.3 10.3 3.83 1.31 1.57 2.44 1.14 0.61 99.5 2.371 1.496 29 72.3 10.8 4.24 1.82 5.10 2.51 1.33 0.64 98.7 2.442 1.513 39 69.5 14.5 5.04 2.41 3.97 2.48 1.22 0.77 99.9 2.468 1.518 30 74.9 10.1 3.90 1.80 3.97 2.41 1.21 0.58 98.9 2.422 1.507 40 70.6 14.1 4.92 2.28 3.59 2.35 1.13 0.72 99.7 2.460 1.515 NUMBER 22 19 TABLE 1.?Continued. Constituent SiO2 AI2O3 FeO MgO CaO K20 Na20 TiO2 Total SG n Constituent SiO2 A12O3 FeO MgO CaO K2O Na2O TiO2 Total SG n 41 72.3 12.5 4.55 2.02 3.71 2.49 1.26 0.71 99.5 2.446 1.513 51 74.9 10.9 4.75 2.41 2.68 2.37 1.11 0.64 99.8 2.415 1.506 42 73.5 12.5 4.45 1.85 3.49 2.41 1.25 0.67 100.1 2.434 1.510 52 69.0 13.2 5.50 3.30 4.14 2.33 1.17 0.75 99.4 2.493 1.523 43 75.2 12.0 4.28 1.70 2.11 2.52 1.13 0.69 99.6 2.399 1.501 53 69.9 13.2 5.76 3.13 3.22 2.52 1.26 0.72 99.7 2.477 1.520 44 76.0 11.7 4.08 1.64 2.07 2.32 1.09 0.67 99.6 2.391 1.500 54 70.6 13.1 4.99 2.47 3.59 2.66 1.41 0.74 99.6 2.460 1.515 45 76.8 11.5 4.32 1.54 1.60 2.55 1.10 0.65 100.1 2.377 1.498 55 71.4 12.5 5.33 2.96 2.92 2.37 1.21 0.66 99.4 2.445 1.512 46 68.7 13.0 6.10 3.53 3.91 2.28 1.29 0.79 99.6 2.489 1.522 56 73.0 12.2 5.25 2.99 2.87 2.16 1.10 0.66 100.2 2.425 1.506 47 70.7 12.8 5.85 3.33 3.28 2.16 0.97 0.72 99.8 2.466 1.519 57 67.3 15.9 5.57 2.68 3.22 2.78 1.38 0.85 99.7 -- 1.519 48 71.1 12.0 5.39 2.92 3.59 2.37 1.18 0.65 99.2 2.446 1.513 58 68.2 15.7 5.37 2.57 3.08 2.56 1.30 0.83 99.6 -- 1.518 49 72.2 11.8 5.33 2.70 3.40 2.48 1.16 0.64 99.7 2.437 1.513 59 68.9 15.3 5.27 2.33 2.99 2.61 1.13 0.83 99.4 -- 1.518 50 74.1 11.6 5.26 2.98 2.51 2.08 0.91 0.65 100.1 2.420 1.507 60 69.3 15.4 5.25 2.45 2.90 2.55 1.38 0.82 100.1 -- 1.518 All Fe as FeO. 20 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES 70.9%-77.3%, similar to but somewhat less exten- sive than that for Durrie. Other components are comparable for those recorded for australites of similar SiO2 content from Charlotte Waters and Durrie; in particular, the low-SiO2 (<75%) aus- tralites from Macumba also show relatively high CaO (except for analysis 28). PINE DAM.?Australites from Pine Dam, near Lake Torrens, show a specific-gravity range of 2.368-2.476, with peaks at 2.395 and 2.462 (Chalmers et al., 1976, tbl. 7). Six specimens, cover- ing practically the full specific-gravity range, were analyzed (Table 1, columns 33-38). These analy- ses resemble those of similar SiO2 content from Macumba, except that they are generally lower in CaO and higher in A12O3. MANNAHILL.?A collection of 89 australites from Mannahill was available for study. The specific- gravity range was 2.377 to 2.468, with a single peak at 2.419 in the distribution curve. This single peak at 2.40 to 2.42 is characteristic of collections from southeast South Australia and Victoria (Chalmers et al., 1976, tbl. 7). Seven specimens, covering the whole specific-gravity range, were analyzed (Table 1, columns 39-45). The analyses are comparable to those of Pine Dam australites with similar SiO2 contents, although the Mannahill australites appear to have somewhat lower CaO contents. FROME DOWNS.?A small collection (14 specimens) from Frome Downs station was acquired after the main part of this manuscript was completed. It is of considerable significance because of its approxi- mately equidistant position between three loca- tions (see Figure 1) with distinctly different austra- lite populations, as indicated by specific gravity ranges and peaks (in parentheses): Mannahill (2.377-2.468, peak 2.419); Pine Dam (2.368-2.476, peaks 2.395, 2.462); Pindera (2.270-2.470, peak 2.450). The 14 Frome Downs australites have a specific gravity range of 2.380-2.429; no peak is evident, probably because of the small number of specimens, but the mean is 2.402. The Frome Downs collection clearly differs in specific gravity range and therefore in chemical composition from the Pine Dam and other collections in the Lake Torrens region to the west, and to the nearest loca- tion in the east, the Pindera area in New South Wales. A search of the region between Frome Downs and Pindera in August 1978 found no australite occurrence, although the terrain appeared favorable in many places. I conclude therefore that a significant chemical hiatus exists between these two locations. The Frome Downs occurrence seems to be the farthest north location for the low specific gravity Victoria-type australites unmixed with the higher specific gravity population, as around Lake Torrens and Lake Eyre. LAKE WILSON.?This location, in the far north- west of South Australia, is known to carry australites of unusual composition. This was first recognized by Taylor and Sachs (1964:250), who analyzed one specimen and commented: "Sample No. 26 is un- usual in possessing, in addition to the high value for nickel, the highest amounts of Mg, Fe, Co and Cr. The concentrations of the other elements are normal for the silica content (69.8 per cent)." Chapman et al. (1964) published a specific-gravity distribution diagram for Lake Wilson australites and noted that it was different from those of all other australite locations, with a more extensive specific-gravity range (2.395-2.505) and the highest peak (2.465). Six specimens from Lake Wilson were analyzed (Table 1, columns 46-51), ranging from 2.415 to 2.489 in specific gravity and 74.9% to 68.7% in SiO2 content. The data plotted in Figure 2 show that the Lake Wilson australites are uniformly higher in FeO and MgO than other australites of similar SiO2 content. It was on this basis that Chapman and Scheiber (1969) established their HMg group of tektites; they recognized tektites belonging to this group not only in Australia but in Indonesia and at one locality in the Philippines. Chapman and Scheiber (1969) and Chapman (1971) published analyses of HMg australites from the following locations: Lake Margaretha (25?26'S, 125? 14'E), West Serpentine Lakes (28?50'S, 128?35'E), near Young Range (25?06'S, 125?15'E), and Giles Creek (25?02'S, 128?18'E). These locations, together with the Mt. Davies collection mentioned below, define a triangular area straddling the South Australia-Western Australia border and extending over about 4? of both latitude and longitude. Hughes, with australites of "normal" composition, is about 140 miles south of West Serpentine Lakes, and Earaheedy, also with australites of normal com- position, is about 150 miles west of Young Range; at Charlotte Waters, 300 miles to the east of Lake Wilson, the australites are also of normal composi- tion except for relatively high CaO. The HMg NUMBER 22 21 tektites thus occupy a defined geographic field, and although its limits remain to be precisely defined, they do not appear to overlap into surrounding areas of normal compositions. MT. DAVIES.?Mt. Davies is a geographical area centered around the Western Australia-South Australia?Northern Territory conjunction, about 40 miles west of Lake Wilson. Many thousands of australites have been collected from this area by aborigines and sold through the Department of Aboriginal Affairs, Adelaide. I obtained a small collection of 12 specimens directly from the local aborigines in 1965. Five of these were analyzed (Table 1, columns 52-56), and these analyses are very similar to those of the Lake Wilson australites of comparable SiO2 content. PINDERA.?This is an isolated location in north- western New South Wales, rather far removed from other areas of tektite concentration. Australites from this location have a unique appearance, with pitted surfaces resulting from the presence of small inter- nal bubbles, a feature rare in australites from other localities. As a result, specific gravities of Pindera australites show a wide range (2.270-2.470 in a collection of 155 specimens) according to Chapman (in Chalmers et al., 1976, tbl. 6). This range is, how- ever, almost entirely a function of bubble content, since the ten specimens I analyzed all had very similar compositions, with SiO2 ranging from 67.3% to 69.3% and A12O3 ranging from 15.9% to 15.3%; four of these analyses, covering the full range of composition, are given in Table 1 (columns 57-60). Chapman and Scheiber (1969) report a similar com- position for the Pindera specimen they analyzed. Chapman and Scheiber note that their Pindera specimen and two other analyzed australites from New South Wales have uniquely high A12O3 con- tents (>14.9%), and this may be related to their unusually low SiO2 contents (66.8%-68.5%). Figure 2 shows that apart from the low SiO2 and high A12O3 contents, the only other distinguishing fea- ture in the analyzed Pindera australites is a rela- tively low CaO content. Nevertheless, it is possible that the New South Wales portion of the strewnfield does have a distinctive chemical composition; more analyses from additional locations are needed to establish this. Discussion of Analyses Most of the available analyses of australites have been published by Taylor (1962), Taylor and Sachs (1964), Chapman and Scheiber (1969), and Chap- man (1971). Table 2 compares the results ob- TABLE 2.?Compositional range, in weight percent, of australites; (source of data by column: 1, this paper, 60 analyses; 2, Taylor (1962) and Taylor and Sachs (1964), 32 analyses for SiO,2, A1,,,O3, FeO and 43 analyses for other components; 3, Chapman and Scheiber (1969) and Chapman (1971), 23 analyses) Constituent SiO2 A12O3 FeO a MgO CaO K20 Na20 TiO2 67.3 9.5 3.55 1.31 1.57 2.08 0.91 0.59 1 - 78.9 - 15.9 - 6.10 - 3.53 - 5.93 - 2.78 - 1.74 - 0.85 2 69.6 - 9.35 - 3.83 - 1.49 - 2.13 - 2.07 - 1.05 - 0.55 - 78.7 14.0 5.38 2.49 5.09 2.57 1.52 0.90 3 66.9 - 9.9 - 3.57 - 1.31 - 1.72 - 2.00 - 1.00 - 0.48 - 79.7 16.1 6.06 4.28 5.62 2.62 1.58 0.93 All Fe as FeO. 22 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES tained by these investigators with those reported in Table 1. Table 2 excludes the HNa/K austra- lites of Chapman and Scheiber; these are very rare 3 (Chapman and Scheiber found nine during the js specific-gravity screening of approximately 47,000 tektites), and their identification as tektites has been questioned by Chalmers et al. (1967:37-38). The comparison shows that the ranges of com- position for the major components are very similar q in each of the three groups. It can therefore be ^ stated with some confidence that the compositional range of australites is now well established, and specimens falling outside this range are unlikely to be found. The extent of the variation in major components is clearly demonstrated in Figure 2, in which the percentages of the individual components in each MgO and Ni < 41 ppm (Chapman, 1971:6313). HCa: CaO varying from about 2.0 to 10.0% and MgO from 1.3 to 2.5% (Compston and Chapman, 1969:1024). FeO 3.5-4.9%, Ni 14-42, Cr 56-110, Co 7-18 ppm, Cr/Ni 2.0-4.9 (Chapman and Scheiber, 1969:6741). LSG-HCa streak . . . stretching from Tasmania to Central Australia ... low modal SG (< 2.41); Na2O < 1.25% (Chapman, 1971:6312). Nearly HCa: CaO near lower boundary and with NasO > 1.25% (Chapman, 1971:6313). HMg: MgO varying continuously from 1.5 to 8.0% and comparatively restricted variation in CaO (Compston and Chapman, 1969:1024). MgO > 3.4%, Ni > 210 ppm, Cr > 210 ppm (Chapman, 1971:6312). FeO 3.9-8.6%, Ni 91-390 ppm, Cr 100-460 ppm, Co 14- 63 ppm (Chapman & Scheiber, 1969:6741). Nearly HMg: MgO > 2.8%, Ni > 200 ppm, Cr > 190 ppm (Chapman, 1971:6312). HA1: A12O3 > 14.9% (Chapman, 1971:6312). Some comments may be made on these criteria. The criterion CaO >MgO given for normal austra- lites applies to all australite analyses except for a few in the HMg class. The criterion CaO varying from about 2.0% to 10.0% and MgO from 1.3% to 2.5% for the HCa class applies to practically all australites; reference to Table 1 and Figure 2 shows only three australites with less than 2% CaO, and the only australites with MgO >2.5 belong to the HMg class. Australites generally contain less than 6% CaO; tektites with CaO greater than 6% seem to be limited to the Philippines. Chapman (1971) presented a map (his fig. 2) show- ing the distribution of his chemical classes of aus- tralites. Practically all his normal australites are confined to Western Australia. The HMg australites define a streak extending northwest from the Lake Wilson-Mt. Davies region. The HCa australites de- fine a streak from Victoria through South Australia to Charlotte Waters and Henbury. The HA1 austra- lites are found only in New South Wales, with the exception of one from Tasmania. Figure 2 provides the opportunity to try to fit the analyses of this investigation into Chapman's classi- fication scheme. The only class that can be un- equivocally recognized is the HMg class; the analy- ses of australites from the Lake Wilson and Mt. Davies regions are consistently higher in MgO (and FeO) than other australites of similar SiO2 content. The utility of an HA1 group is doubtful; few are known, and their A12O3 content (>14.9%) being coupled with low SiO2 content puts them on the same Al2O3-SiO,2 trend as other australites (Fig- ure 2). Figure 2 shows that some australites, specifi- cally most of those from Charlotte Waters, Durrie, and Macumba, are significantly higher in CaO than the remainder. Thus an HCa class can possibly be recognized, but as pointed out above the criteria given for this class by Compston and Chapman (1969) apply to practically all analyzed australites except those of the HMg class. Robert Fudali has demonstrated to me that a plot (Figure 3) of AL>O3 versus CaO values from the analyses in Table 1 shows a well-defined hiatus. All the values plotting above the hiatus are of australites from Charlotte Waters, Durrie, and Macumba. This pattern defines a coherent and re- stricted geographical region extending from north- central South Australia to adjoining southwest Queensland. Not all analyses from these locations, however, yield values that plot above the hiatus; two of seven from Charlotte Waters, four of six from Durrie, and one of six from Macumba plot below it. The values which plot above the hiatus probably all represent HCa australites according to Chapman, but many of those plotting below the hiatus also belong to this class. The values from five analyses of HCa australites published by Chapman and Scheiber (1969) have been plotted on Figure 3; three (all from the Charlotte Waters area) plot above the hiatus, one (from Victoria) plots below it, and one (from near Macumba) plots just within its lower limit. One purpose of this research was to see whether australites from a specific locality showed a distinc- tive range of chemical composition. Table 1 and Figure 2 show that this is not generally the case, although some regional trends can be distinguished. The Lake Wilson-Mt. Davies group in the far northwest of South Australia are uniformly some- what higher in MgO and FeO than other australites of similar SiO2 contents, but for the other major components they fall within the range of most aus- tralites. The Pindera australites have uniquely high A12O3 contents. Some but by no means all the aus- tralites from localities extending from Mannahill to Charlotte Waters in South Australia and Durrie in Queensland have high CaO (>4%); some of these also have lower A12O3 contents than most 24 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Earaheedy L. Yindarlgooda Hughes Charlotte Waters Durrie Mocumba Pin* Dam Mannahill Pindera L. Wilson Mt. Davies 12 13 % ALO, 14 15 FIGURE 3.?Plot of Al O3 versus CaO, both in weight percent, for the analysis in Table 1 (c = HCa australites by Chapman and Scheiber (1969); diagonal lines define locality-related hiatus). other australites. Australites from Western Australia and adjacent parts of the Nullarbor Plain tend to be slightly higher in K2O and Na2O than australites from other localities. These results may be compared with the observa- tions of Taylor and Sachs (1964:241). They detected regional differences between average australite com- positions from localities near Kalgoorlie, Charlotte Waters, and in Victoria. They commented: Compared to the Charlotte Waters group (II), the Kalgoorlie group (I) have a distinctly lower average for SiO2 and are higher in those elements which show significant negative correlations with SiO2 . ? . The Victorian group (III), although containing low values for the alkali elements, are generally intermediate in average composition. It thus appears that two distinct groups, at Kalgoorlie and Charlotte Waters, may be distinguished. Perhaps a mixture of these two types is found in Victoria. My results are consistent with these comments. I did not analyze any Victorian australites, but many of my analyses of Pine Dam and Mannahill australites can be matched with published analyses of Victorian australites. Australites and Microtektites In 1967 Glass announced the discovery of micro- scopic (< 1 mm diameter) glassy objects in deep-sea sediments from south and west of Australia: "On the basis of their geographical distribution, appear- ance, physical properties, and age of deposition, it is concluded that the glassy objects discussed in this report are microtektites, and that they constitute a portion of the Australasian strewn field which extends from Thailand to Tasmania" (Glass, 1967: 374). Not all tektite researchers have accepted the identity of microtektites with tektites; a summary of opposing viewpoints is provided by Chalmers et al. (1976:16). Cassidy et al. (1969) published microprobe anal- yses of 60 microtektites and Glass (1972b) reported 44 analyses of a particular variety which he calls "bottle green microtektites." Cassidy et al. and Glass both conclude that their chemical analyses support the identification of these microtektites with Australian tektites. A rigorous comparison, however, of their analyses of microtektites with the NUMBER 22 25 analyses of australites reveals many inconsistencies and may negate their conclusion. The range of composition of microtektites is much wider than that for australites. For the prin- cipal component, SiO2, the range for microtektites is 48.1% to 77.0%; of the 104 microtektite analyses from the Australian region, only 46, or less than half, fall within the SiO2 range of australites (66%- 79%). Cassidy et al. claimed that the "normal" microtektites (excluding the bottle green ones) showed concordant compositional trends with the Australasian tektites and served to extend the known composition range of tektites. Glass claimed that the bottle green microtektites overlap and extend all major-element trends found for the HMg tektites of Chapman and Scheiber (1969). Even if this claim were valid, the non-existence of austra- lites with less than 66% SiO2 renders it of little relevance. When the data, however, for microtektites within the australite composition range of 66%-79% SiO2 are studied closely, the correspondence between ? Aultralitn ? Normal microtoktitu * Bottl* gra*n mkrataktlm FIGURE 4.?Plot of FeO versus MgO, both in weight percent, for australites (Table 1) and microtektites (Cassidy et al., 1969; Glass, 1972b) (trapezoid defines scatter of australites). australite and microtektite compositions can only be described as poor. Discordances are particularly prominent for Na,O (most microtektites have less than 1%, most australites more than this amount); for K2O, where only four microtektites fall within the field for australites and two of these are dis- cordant with australites in other components; and for MgO, where practically all the bottle green microtektites have much higher contents than the HMg australites with which they are claimed to correspond. The chemical discordance between microtektites and australites is clearly displayed in Figure 4, which plots FeO versus MgO percentage for australites and microtektites. The australites show a very limited range in FeO and MgO con- tents and an extremely strong positive correlation between these two components. Note the contrast in the microtektites, which show a very wide range in FeO and MgO contents, and essentially no cor- relation between these components. The extreme scatter of microtektite analyses, com- pared to the relatively restricted composition of australites, argues against rather than for a com- mon source material. If one is to claim that micro- tektites and australites are cogenetic, the evidence must be from sources other than chemical compo- sition. Literature Cited Cassidy, W. A., B. Glass, and B. C. Heezen 1969. Physical and Chemical Properties of Australian Microtektites. Journal of Geophysical Research, 74:1008-1025. Chalmers, R. O., E. P. Henderson, and B. Mason 1976. Occurrence, Distribution, and Age of Australian Tektites. Smithsonian Contributions to the Earth Sciences, 17: 46 pages. Chapman, D. R. 1971. Australasian Tektite Geographic Pattern, Crater and Ray of Origin, and Theory of Tektite Events. Journal of Geophysical Research, 76:6309-6336. Chapman, D. R., H. K. Larson, and L. C. Scheiber 1964. Population Polygons of Tektite Specific Gravity for Various Localities in Australasia. Geochimica et Cosmochimica Ada, 28:821-839. Chapman, D. R., and L. C. Scheiber 1969. Chemical Investigation of Australasian Tektites. Journal of Geophysical Research, 74:6737-6773. Compston, W., and D. R. Chapman 1969. Sr Isotope Patterns within the Southeast Austral- asian Strewnfield. Geochimica et Cosmochimica Acta, 33:1023-1036. 26 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Glass, B. P. Australian-New Zealand Sector, pages 335-248. Wash- 1967. Microtektites in Deep-Sea Sediments. Nature, 214: ington: American Geophysical Union. 372-374. 1972b. Bottle Green Microtektites. Journal of Geophysical 1970. Comparison of the Chemical Variation in a Flanged Research, 77:7057-7064. Australite with the Chemical Variation among Taylor, S. R. "Normal" Australasian Microtektites. Earth and 1962. The Chemical Composition of Australites. Geo- Planetary Science Letters, 9:240-246. chimica et Cosmochimica Ada, 26:685-722. 1972a. Australasian Microtektites in Deep-Sea Sediments. In Taylor, S. R., and M. Sachs D. E. Hayes, editor, Antarctic Oceanology II: The 1964. Geochemical Evidence for the Origin of Australites. Geochimica et Cosmochimica Ada, 28:235-264. The Pyroxene-Plagioclase Achondrites Brian Mason, Eugene Jarosewich, and Joseph A. Nelen ABSTRACT The pyroxene-plagioclase achondrites (also knownas eucrites and howardites, as calcium-rich achon- drites, and as basaltic achondrites) form the largestclass of achondrite meteorites, and are of spe- cial interest because they are the closest meteoriticanalogs of lunar rocks. In this paper we provide chemical analyses for 20 of them, and have selected analyses of 31 others from the literature. Brief ac-counts are provided for them, together with micro- probe analyses of their principal minerals, pyroxeneand plagioclase. We have adopted the structural dis- tinction between eucrites (unbrecciated and mono- mict breccias) and howardites (polymict breccias).Shergotty and Zagami, although eucrites by this difinition, are distinct in their chemical and min-eralogical composition, and the term shergottite should be retained for them. A large number of eucrites (which we call main-group) have almostidentical chemical and mineralogical composition; they are silica-oversaturated, having excess SiO2 inboth the mode and the norm. Howardites are not mechanical mixtures of eucrites and diogenites; their individual mineral and rock clasts belong toa differentiation sequence that is compositionally more extensive than the eucrites and diogenites. Introduction The pyroxene-plagioclase achondrites are a group of meteorites consisting essentially of pyroxene and calcic plagioclase, the pyroxene usually dominant over plagioclase. They compose a majority of the achondrites, totalling over 50 of approximately 90 known. Prior (1920) grouped them together with Angra dos Reis and the nakhlites as calcium-rich Brian Mason, Eugene Jarosewich, Joseph A. Nelen, Depart- ment of Mineral Sciences, National Museum of Natural His- tory, Smithsonian Institution, Washington, D.C. 20560. achondrites. Urey and Craig (1953) preferred to call the calcium-rich achondrites the basaltic-type achondrites (now generally shortened to basaltic achondrites), to emphasize their general similarity to terrestrial basalts (although the latter are clearly different in the amount and composition of the plagioclase). Prior recognized three classes of pyroxene-plagio- clase achondrites: (1) eucrites or clinohypersthene- anorthite achondrites; (2) shergottites or clinohy- persthene-maskelynite achondrites; and (3) howardites or hypersthene-clinohypersthene-anorth- ite achondrites. (Clinohypersthene in Prior's usage is the mineral now known as pigeonite.) In the current British Museum catalog (Hutchison et al., 1977) the eucrites are denned as plagioclase-pigeon- ite achondrites and the howardites as plagioclase- hypersthene achondrites (the two shergottites, Sher- gotty and Zagami are included with the eucrites). Thus the distinction between eucrites and how- ardites was originally made on the basis of the na- ture of the pyroxene, the dominant pyroxene being pigeonite in eucrites and orthopyroxene (hyper- sthene) in howardites. Mason (1962) noted that this distinction could be correlated with the bulk chem- istry of these meteorites, most clearly with the CaO content and the FeO/(FeO + MgO) molecular per- centage; he suggested that a dividing line could be drawn at FeO/(FeO + MgO) = 50, eucrites showing a higher ratio, howardites a lower one. Lacroix (1926) felt there were insufficient grounds for separating eucrites and howardites, but that if such separation were desired, it would be better made on structural grounds, howardites being classed as brecciated eucrites. Wahl (1952) intro- duced the terms monomict and polymict breccias, and classed the eucrites as unbrecciated or mono- mict, howardites as polymict pyroxene-plagioclase 27 28 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES achondrites. This distinction was adopted by Duke and Silver (1967) and several other researchers. Bunch (1975) provided a comprehensive account of the large variety of rock and mineral clasts found in the polymict pyroxene-plagioclase achondrites, and clearly demonstrated the desirability of distin- guishing these meteorites as howardites. It is noteworthy that either the structural or the mineralogical criterion is usually sufficient to class- ify unambiguously a pyroxene-plagioclase achon- drite as a howardite or a eucrite, since meteorites with dominant orthopyroxene are also polymict breccias (except Binda), and most monomict brec- cias have pigeonite as the dominant pyroxene (polymict breccias with dominant pigeonite are Macibini, Bialystok and Nobleborough). In using the terms howardite and eucrite, however, it is desirable to state in what sense these terms are used; in this paper we use them in the structural sense. We also distinguish Shergotty and Zagami from the eucrites on account of their distinctive chemical and mineralogical composition. In discussing the eucrites and howardites, fre- quent reference will be made to the diogenites (also known as hypersthene achondrites). The diogenites, although classed by Prior as calcium-poor achon- drites, occupy one end of a spectrum of chemical and mineralogical compositions extending through howardites to eucrites. Prior did not specify a limit- ing value of CaO between the calcium-poor and calcium-rich achondrites; currently the most cal- cium-rich diogenite is Garland with 2.60% CaO (Fredriksson et al., 1976), and the most calcium-poor howardite is Yamato-7307 (also known as Yamato(l) with 3.83% CaO (Wanke et al., 1977). The amounts of the other major components are equally close. Nevertheless, these two meteorites are easily dif- ferentiated both macroscopically and microscopic- ally, Garland being an aggregate of pale-green orthopyroxene clasts with only trace amounts of other minerals, whereas Yamato-7307, while con- taining a large amount of pyroxene similar in com- position to that in Garland, also has a considerable eucritic component (pigeonite + plagioclase). One characteristic of howardites is that they frequently contain clasts of diogenite composition. A close relationship, therefore, exists between these two classes, which will be further discussed later in this paper. Interest in the pyroxene-plagioclase achondrites has increased greatly in recent years, largely because they are the closest analogs to lunar rocks. Indeed, Duke and Silver (1967) presented arguments sug- gesting that the eucrites, howardites, and meso- siderites were actually lunar rocks. The results of the Apollo missions have negated this suggestion, and therefore established that these meteorites prove the previous existence of at least one other body in the Solar System with analogous chemical and mineralogical composition (and presumably similar evolutionary history) to the moon. The na- ture and composition of this body ("eucrite parent body") has been the subject of intensive research (see, for example, Dreibus et al., 1977; Consolmagno and Drake, 1977; Allegre et al., 1977; Stolper, 1977; Anders, 1977). In view of this great interest in the pyroxene- plagioclase achondrites, we have collected the pub- lished major-element analyses of these meteorites and supplemented them with analyses of as many previously unanalysed meteorites as possible. Where published analyses are either incomplete or show inconsistencies with the mineralogical composition we have attempted to rectify this. Some of the new analyses (by E. J.) are by classical wet-chemical techniques; where only small samples were avail- able the material was fused with lithium tetraborate and the resulting glasses analysed with the electron microprobe (by J. N.). ACKNOWLEDGMENTS.?We are indebted to the curators of the meteorite collections in the follow- ing institutions for providing material of meteorites not represented in the Smithsonian Institution col- lection: Arizona State University; Australian Mu- seum, Sydney; British Museum (Natural History); Field Museum, Chicago; Geological Survey of India; Museum d'Histoire Naturelle, Paris; Natur- historisches Museum, Vienna. Results The analyses are presented in Table 1. In this tabulation K2O and P2O5 are omitted (recent analy- ses show that K2O is consistently below 0.05% and P2O5 is usually less than 0.1%); free nickel-iron is also omitted, since it is a rare and erratically dis- tributed accessory and is seldom determined; FeS, if determined, is given in the table key. The analyses are arranged in order of increas- ing FeO/(FeO + MgO) mole percentage. This is a NUMBER 22 29 useful factor in distinguishing eucrites and howar- dites in the chemical sense (howardites have FeO/(FeO + MgO) 30-45, eucrites 40-69, there being a slight overlap). The ratio also correlates positively with CaO and A12O3, and negatively with MgO. Other parameters derived from the analytical data are the composition and amount of normative plagioclase, the composition and amount of norma- tive pyroxene, and the amount of normative olivine or silica. From a consideration of texture the mete- orites are classified as unbrecciated, monomict brec- cias, and polymict breccias; if texture is used as the discriminant factor between eucrites and howard- ites, then unbrecciated meteorites and monomict breccias are eucrites, while polymict breccias are howardites. Notes on the individual meteorites (bearing the analysis numbers used in Table 1) follow. 1. YAMATO-7307.?This meteorite, also known as Yamato(l), was collected by the Japanese Antarctic Expedition in 1973, and is of special interest since it has the lowest FeO/(FeO + MgO) mole percentage of any of these meteorites. As such, it is composi- tionally very similar to the diogenites. The dio- genites, however, are essentially monomict breccias (a small eucritic component has been recognized in a few), with uniform pyroxene composition (Fs 25-27). Sections of Yamato-7307 which we have examined show that it is a polymict breccia, with a wide range in pyroxene composition (Fs 19-49); two grains of olivine were seen (Fa 17, 37), and a diogenite enclave consisting essentially of pyroxene with uniform composition (Fs 27). 2. MASSING.?The analysis shows that this mete- orite is very similar in composition to Yamato-7307; an incomplete analysis by Fukuoka et al. (1977) con- firms this. It is a polymict breccia with a wide range of pyroxene compositions (Fs 22?46); plagioclase also shows a considerable range in composition (Ab 4-21). It should be mentioned that specimens of a chondrite mislabelled as Massing have been dis- tributed (the specimens in the Museum d'Histoire Naturelle, Paris, and Eidgenossische Technische Hochschule, Zurich are such); the analytical data on Massing A reported by Schmitt et al. (1972) are probably of this mislabelled material. 3. CHAVES.?Thin sections of this meteorite show that it is a typical polymict breccia. The ground- mass consists largely of comminuted orthopyroxene enclosing larger shocked clasts of the same mineral; eucrite enclaves (unbrecciated, with ophitic texture) are not uncommon, and one fragment of diogenite was seen. Jeremine (1954) has provided a thorough petrographic description of this meteorite; she noted the presence of accessory tridymite and quartz, which we have confirmed with the microprobe (note the 0.5% of free SiO2 in the norm). Our microprobe analyses show that the pyroxene consists largely of orthopyroxene, with lesser amounts of pigeonite; the Fs mole percent ranges from 20 to 51. Plagio- clase also shows a range of composition, Ab 9-24. 4. FRANKFORT.?This meteorite has been described by Mason and Wiik (1966a). It is a polymict breccia, with pyroxene clasts (mainly orthopyroxene, some pigeonite) in a comminuted groundmass; one small eucrite enclave was observed. Our microprobe analy- ses show pyroxene compositions ranging from Fs 24 to Fs 39; similar results were reported by Takeda, Miyamoto, Ishii, and Reid (1976). 5. ZMENJ.?Thin sections show a typical polymict breccia, consisting largely of clasts of hypersthene (major) and pigeonite (minor) in a comminuted groundmass. Our microprobe analyses show pyrox- ene compositions ranging from Fs 23 to Fs 44, and plagioclase ranging from Ab 4 to Ab 10 (normative plagioclase is Ab 13, which suggests that the Na2O in the analysis may be a little too high). Desnoyers and Jerome (1973) analysed a number of grains of olivine in Zmenj, and found compositions ranging from Fa 15 to Fa 39. 6. BINDA.?This meteorite, originally classified as a howardite because the pyroxene is largely hyper- sthene, was reclassified by Duke and Silver (1967) as a eucrite because it is a monomict breccia. The pyroxene in Binda has been described by Takeda and Ishii (1975); they comment (page 500): The bulk composition of low-Ca pigeonite suggests the crystallization temperature is above 1110?C. The temperature where augite blebs Ca44.7Mg42.4Fe12.9 developed in the host orthopyroxene Ca2.4Mg64.4Fe33.2 is about 970?C. The Binda howardite contains the most magnesian pigeonite among the basaltic achondrites, and these pyroxenes are in- terpreted to be crystallized from magma and not the mechni- cal mixture of diogenite and eucrite pyroxenes. The Binda meteorite may be classified as the most Mg-rich and Ca-poor eucrite. 7. PAVLOVKA.?Thin sections show a polymict breccia largely of mineral grains (hypersthene with minor pigeonite and plagioclase), but small- to medium-sized eucrite enclaves are not uncommon; some plagioclase grains are unusually large, larger 30 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES than could be derived from common eucrites, sug- gesting a possible anorthositic parent. Our probe analyses indicate an unusually wide range of pyroxene compositions; orthopyroxene is dominant, with Fs ranging 14-36, with lesser amounts of pigeonite, Fs 36-55, and a few grains of ferroheden- bergite, averaging Wo38Fs47En15; a single grain, Wo19Fs72En9, possibly pyroxferroite, was analyzed. Plagioclase ranged Ab 7-13, average 9. Two grains of olivine, Fa 33, 35 were analyzed. Desnoyers and Jerome (1973) analyzed 15 grains of olivine in Pavlovka, with Fa ranging 36-49. An incomplete analysis by Fukuoka et al. (1977) is very similar to that given in Table 1. 8. WASHOUGAL.?Jerome and Michel-LeVy (1972) describe this meteorite as a very heterogeneous poly- mict breccia. From the refractive indices, they give the range of composition of the pyroxenes as Fs 20-47. They describe several kinds of lithic clasts, among them eucritic fragments with an ophitic or subophitic texture, and a cm-sized dunite xenolith (olivine composition Fa 12.8). 9. BHOLGATI.?A thin section shows a breccia with numerous clasts of orthopyroxene up to 1 mm across, small less numerous clasts of pigeonite and plagioclase, and rare grains of chromite, in a fine- grained groundmass. Our microprobe analyses show a range of Fs 23-62 in orthopyroxene and pigeonite KEY TO TABLE 1 Identification of Analyses by Columns 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Yamoto-7307, also known as Yamoto(l); Wanke et al., 1977 (FeS = 0.30%). Massing; J. Nelen, analyst. Chaves; McCarthy et al., 1972. Frankfort; Mason and Wiik, 1966b (FeS = 0.69%). Zmenj; Wiik, 1969 (FeS = 0.48%). Binda; McCarthy et al., 1973. Pavlovka; J. Nelen, analyst. Washougal; Jerome and Michel-Levy, 1972 (FeS = 0.99%). Bholgati; E. Jarosewich, analyst. Luotolax; Wiik, 1969 (FeS = 1.08%). Yurtuk; J. Nelen, analyst. Bununu; Mason, 1967a. (TiO2) revised, 0.51%). Medanitos; Symes and Hutchison, 1970. Le Teilleul; J. Nelen, analyst. Molteno, non-magnetic fraction; Frost, 1971. Moama; Lovering, 1975. Jodzie; J. Nelen, analyst. Serra de Mage; E. Jarosewich, analyst. Kapoeta; Mason and Wiik, 1966a (FeS = 0.98%). Malvern; McCarthy et al., 1972. Pomozdino; Kvasha and Dyakonova, 1972 (FeS = 2.79%). Brient; Dyakonova and Kharitonova, 1961 (FeS = 0.55%). Petersburg; E. Jarosewich, analyst (FeS = 0.74%). Nagaria; J. Nelen, analyst. Moore County; Hess and Henderson, 1949 (FeS = 0.82%). Shergotty; Duke, 1968. 27. Macibini; McCarthy et al., 1973. 28. Bialystok; E. Jarosewich, analyst (FeS = 0.55%). 29. Yamoto-74450; Wanke et al., 1977 (FeS = 0.65%). 30. Ibitira; Wanke et al., 1974. 31. Jonzac; J. Nelen, analyst. 32. Sioux County; Duke and Silver, 1967 (FeS = 0.45%). 33. Allan Hills No. 5; E. Jarosewich, analyst (FeS = 0.27%). 34. Nobleborough; J. Nelen, analyst. 35. Chervony Kut; Zavaritskii and Kvasha, 1952 (FeS = 0.25%). 36. Cachari; McCarthy et al., 1973. 37. Kirbyville; J. Nelen, analyst. 38. Juvinas; Duke and Silver, 1967 (FeS = 0.53%). 39. Millbillillie; J. Nelen, analyst. 40. Mount Padbury enclave; J. Nelen, analyst. 42. Adalia; J. Nelen, analyst. 43. Palo Blanco Creek; J. Nelen, analyst. 44. Peramiho; Berwerth, 1903 (FeS = 0.80%). 45. Stannern; Von Engelhardt, 1963 (FeS = 0.72%). 46. Haraiya; E. Jarosewich, analyst (FeS = 0.27%). 47. Pasamonte; Duke and Silver, 1967 (FeS = 0.06%). 48. Bereba; McCarthy et al., 1973. 49. Emmaville; Mason, 1974. 50. Nuevo Laredo; Duke and Silver, 1967 {FeS = 0.21%). 51. Lakangaon; McCarthy et al., 1974, with Cr2O3 and Na2O by J. Nelen. TABLE 1.?Analytical data on pyroxene-plagioclase achondrites (a = mole percent FeO/ (FeO + MgO); b = mole percent Ab in normative plagioclase; c = weight percent normative plagioclase; d = mole percent Wo, Fs in normative pyroxene; e = weight percent normative pyroxene; f = weight percent normative olivine (ol) or silica (q); g = texture (u = unbrecciated, m = monomict breccia, p = polymict breccia); key to identification of analyses by columns on facing page) Constituent SiO2 TiO2 A1203 Cr203 FeO MnO MgO CaO Na2O a b c d e f g Constituent SiO2 TiO2 A12O3 Cr203 FeO MnO MgO CaO Na20 a b c d e f g 1 50.82 0.23 4.27 1.02 16.46 0.52 21.37 3.83 0.13 30.2 10 12 4,29 80 5.1 ol P 11 49.0 0.54 9.1 0.78 16.5 0.45 14.8 7.26 0.22 38.4 8 26 7,36 69 2.0 ol P 2 51.1 0.31 4.1 0.91 17.6 0.49 21.3 4.1 0.08 31.7 6 12 4,30 80 6.4 ol P 12 48.67 0.11 8.87 0.56 16.04 0.53 14.20 6.77 0.34 38.8 12 26 6,36 69 0.5 ol P 3 51.79 0.37 6.84 0.61 15.33 0.48 18.39 5.64 0.22 31.8 10 20 5,30 78 0.5 q P 13 47.74 0.03 15.09 0.14 13.72 0.51 12.16 10.38 0.20 38.8 4 42 8,36 49 8.8 ol m 4 49.48 0.46 5.10 1.34 17.39 0.53 20.50 4.02 0.17 32.2 10 15 3,31 73 8.9 ol P 14 50.1 0.42 7.4 0.72 17.6 0.46 15.0 6.44 0.17 39.7 7 21 7,37 74 1.9 q P 5 50.47 0.46 6.97 1.04 16.22 0.55 17.85 5.35 0.29 33.8 13 20 5,32 75 1.8 ol P 15 47.54 0.41 8.83 0.53 15.94 0.41 13.15 6.65 0.29 40.5 10 25 6,38 66 1.1 q P 6 50.41 0.23 6.95 0.75 16.83 0.54 17.75 5.82 0.20 34.8 9 20 5,33 76 3.8 ol m 16 48.58 0.22 13.74 0.63 14.85 0.50 11.89 9.47 0.22 41.2 5 38 7,38 58 2.4 ol u 7 49.9 0.29 6.0 0.86 18.1 0.52 17.8 5.3 0.17 36.3 9 17 5,35 76 4.6 ol P 17 48.3 0.48 7.9 0.63 18.3 0.50 14.6 6.64 0.20 41.3 8 22 7,38 70 3.7 ol P 8 50.22 0.32 6.70 0.83 17.34 0.47 16.95 5'. 60 0.18 36.5 8 19 5,35 77 0.5 ol P 18 46.69 0.11 20.89 0.33 9.97 0.36 7.52 13.09 0.30 42.7 5 58 9,39 38 2.1 ol u 9 49.35 0.41 8.30 0.71 16.27 0.53 15.56 6.14 0.27 37.1 10 24 5,35 72 0.2 ol P 19 48.47 0.37 9.46 0.63 17.16 0.53 12.00 8.08 0.46 44.5 15 28 10,40 67 0.4 ol P 10 50.42 0.64 9.03 0.87 16.09 0.52 14.97 6.83 0.37 37.6 13 26 6,35 71 0.6 q P 20 49.15 0.49 9.96 0.56 18.03 0.53 12.40 8.06 0.42 44.9 13 29 9,41 67 2.3 ol P SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Continued. Constituent SiO2 TiO2 A12O3 Cr2?3 FeO MnO MgO CaO Na20 a b c d e f g Constituent SiO2 TiO2 A12O3 Cr203 FeO MnO MgO CaO Na20 a b c d e f g 21 47.97 0.73 11.31 0.10 14.61 0.53 9.96 10.76 0.41 45.1 11 32 16,38 62 0.9 q m 31 50.7 0.98 11.7 0.33 17.6 0.47 6.9 10.4 0.38 58.9 10 33 15,50 57 6.3 q m 22 49.00 0.79 11.02 0.59 17.02 0.58 11.27 8.63 0.24 45.9 7 31 9,42 64 1.4 q P 31 48.29 0.60 12.84 0.31 18.25 0.53 7.08 10.39 0.42 59.2 10 37 13,51 59 2.9 q m 23 49.18 0.57 10.82 0.56 17.00 0.52 11.24 8.49 0.35 45.9 10 31 9,41 65 1.3 q P 33 48.92 0.73 12.22 0.39 17.56 0.56 6.74 9.63 0.43 59.4 11 35 13,52 55 4.8 q m? 24 47.4 0.38 15.7 0.51 15.0 0.40 8.97 11.0 0.40 48.4 8 44 10,44 50 3.8 ol u 34 50.0 0.78 12.5 0.40 18.5 0.46 7.1 10.0 0.46 59.4 11 36 13,52 58 4.0 q P 25 48.16 0.32 15.57 0.44 15.02 0.31 8.41 11.03 0.45 50.0 9 44 11,45 54 0.04 ol u 35 48.80 0.71 13.44 0.21 18.47 0.63 6.98 11.48 0.52 59.8 12 39 16,50 61 0.3 q u 26 50.10 0.92 6.68 0.18 18.66 0.50 9.40 10.03 1.28 52.7 48 23 21,42 72 0.2 ol u 36 48.26 0.63 12.85 0.32 19.10 0.59 7.14 10.25 0.51 60.1 12 37 13,52 60 0.7 q m 27 49.32 0.72 12.12 0.42 18.31 0.54 8.37 9.97 0.48 55.1 12 35 13,48 62 1.6 q P 37 49.8 0.67 11.9 0.40 18.7 0.54 6.96 10.0 0.45 60.2 12 34 14,52 59 4.1 q m 28 48.97 0.56 12.63 0.72 17.83 0.58 7.88 9.72 0.32 56.0 8 36 11,50 58 3.1 q P 38 49.32 0.68 12.64 0.30 18.49 0.53 6.83 10.32 0.42 60.2 10 36 14,52 58 3.3 q m 29 48.28 0.90 11.43 0.40 18.03 0.53 7.58 9.95 0.51 57.2 14 33 14,49 60 2.4 q m ? 39 52.0 0.68 12.8 0.33 18.3 0.61 6.76 10.2 0.36 60.3 9 36 13,52 57 6.5 q m 30 49.0 0.78 12.6 0.36 18.1 0.50 7.1 10.9 0.19 58.8 5 35 15,50 59 3.6 q u 40 48.0 0.55 13.1 0.39 19.0 0.44 7.0 10.3 0.32 60.4 8 37 12,53 59 1.6 q u NUMBER 22 33 TABLE 1.?Continued. Constituent SiO2 TiO2 A12O3 Cr2O3 FeO MnO MgO CaO Na2O a b c d e f g 41 49.5 0.67 12.8 0.32 18.3 0.57 6.70 10.4 0.36 60.6 9 36 14,52 58 3.9 q m 42 50.1 0.78 11.8 0.39 19.0 0.51 6.91 10.3 0.41 60.7 11 34 15,52 60 4.2 q u 43 48.7 0.36 12.4 0.28 20.4 0.63 7.41 9.44 0.31 60.8 8 35 10,55 62 1.5 q m 44 49.32 0.42 11.24 19.99 7.15 10.84 0.40 61.0 11 32 16,51 64 2.2 q m 45 49.33 0.96 12.34 0.28 17.92 0.50 6.36 10.58 0.60 61.3 15 39 16,51 54 3.8 q m 46 48.30 0.57 11.92 0.33 19.95 0.43 6.96 9.67 0.37 61.7 9 34 12,54 61 2.3 q m 47 48.59 0.65 12.70 0.33 19.58 0.56 6.77 10.25 0.45 61.8 11 36 13,54 60 1.6 q m 48 49.13 0.70 12.75 0.32 19.83 0.55 6.80 10.48 0.47 62.C 11 37 14,54 61 1-.6 q m 49 52.6 0.66 12.2 0.36 18.5 6.12 10.2 0.51 62.8 13 35 15,54 57 7.6 q ? m 50 49.46 0.95 11.78 0.29 20.10 0.56 5.46 10.40 0.57 67.4 15 34 16,57 59 4.0 q - m 51 49.03 0.85 11.54 0.35 22.81 0.60 5.87 10.27 0.49 68.6 13 33 14,59 65 1-.6 q m compositions; plagioclase Ab 5-17, average 10; one grain of augite, Wo44Fs26En30; one grain of olivine, Fa 31; and one grain of a silica polymorph, prob- ably tridymite. 10. LUOTOLAX.?Thin sections show a polymict breccia mostly of individual mineral grains (hyper- sthene, pigeonite, plagioclase) but with occasional eucrite clasts; one section had an anorthosite clast (~90% plagioclase, 10% pyroxene). Microprobe analyses showed pyroxene compositions ranging Fs 33-52, plagioclase ranging Ab 4-19. 11. YURTUK.?Thin sections show a polymict brec- cia with large (up to 1 mm) orthopyroxene and smaller pigeonite clasts; lithic fragments are also present, both eucritic and diogenitic; rare grains of olivine were seen. Microprobe analyses show pyro- xene ranging in composition from Fs 31 to Fs 50. 12. BUNUNU.?This meteorite was described by Mason (1967a), with an analysis by E. Jarosewich. We have made microprobe analyses of the pyrox- enes and the plagioclase. The range of plagioclase composition is Ab 6-17, average 10. Orthopyroxene and pigeonite range is Fs 24-63; one grain of augite, Wo43Fs25En32, was analyzed. 13. MEDANITOS.?This meteorite was described by Symes and Hutchison (1970), who classed it as a howardite because of its high content of Ca-poor pyroxene; however, Hutchison et al. (1977) reclassify it as a eucrite, commenting (page 149): "Further unpublished work . . . indicates that it is a eucrite, mineralogically similar to Moore County and Serra de Mage." The analysis is noteworthy for the very low TiO2 and CraO3 and high A12O3. 14. LE TEILLEUL.?Thin sections show this mete- orite is a polymict breccia; hypersthene clasts (some with augite exsolution lamellae) predominate, with minor amounts of pigeonite and rare eucrite en- claves; one enclave of coarse plagioclase, reminiscent of an anorthosite, was seen. Microprobe analyses show a range of Fs 24-41 in hypersthene and pigeonite, and Ab 10-24 in plagioclase. Desnoyers and Jerome (1973) analyzed 13 grains of olivine in Le Teilleul, and found a composition range of Fa 28-42. 15. MOLTENO.?This meteorite was described by Frost (1971) as a howardite, with pyroxene rang- ing from Wo2Fs22En76 to Wo4Fs58En38 and then to Wo41Fs26En33, with plagioclase (An 84), and with small quantities of olivine (Fa 12), troilite, nickel- iron, ilmenite and chromite. Desnoyers and Jerome 34 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES (1973) analyzed 9 grains of olivine in Molteno, and record a composition range of Fa 11-56. 16. MOAMA.?The analysis of this meteorite shows that it is very similar to Medanitos. Lovering (1975) describes it as an unbrecciated adcumulate of pyrox- ene and plagioclase, with accessory amounts of chromite, troilite, nickel-iron, and tridymite (the presence of olivine in the norm and free silica as tridymite in the mode is probably due to all Fe being reported as FeO, whereas some is present as troilite and nickel-iron). Lovering's microprobe analyses show plagioclase of uniform composi- tion (Ab 6); the major pyroxene is hypersthene (Wo3 4Fs416En55 0) with exsolution lamellae of augite (Wo44.9Fs18.4En36.7). 17. JODZIE.?Bunch et al. (1976) describe this me- teorite as follows: The Jodzie howardite is a polymict-regolith breccia similar to other howardites with two unique exceptions: 1, large cm size lithic clasts of cumulate ferrogabbros, and 2, frag- ments of C2 carbonaceous chondrites {8 vol. %). The cumulate ferrogabbros contain large 1 to 5 mm ferrohypersthene or ferropigeonite grains with exsolved ferroaugite lamellae, calcic plagioclase (An 84-89), ilmenite, FeS, and rarely free silica and phosphates. Bulk analyses of these clasts are similar to lithic clast ferrobasalts in other howardites . . . The C2 fragments range in size from 1/t to 6 mm and are similar in bulk composition and mineralogy to C2 inclusions in the Abbott and Plainview chondrites and the Kapoeta and Washougal howardites. 18. SERRA DE MAGE.?This meteorite has been de- scribed by Prinz et al. (1977) as follows: The Serra de Mage meteorite is a medium to coarse grained anorthositic norite cumulate (eucrite) with a mosaic texture. . . . Phases present are plag (An95), opx (Enss 5), augite (Wo45En39Fsi6), chromite (Cr2O3, 50-54%), SiO2, Ni-Fe (Ni, 0.3-1.5%; Co, 0.8-1.2%), troilite, and a trace of ilmenite; zircon and phosphate are present but were not found in section ... It is highly equilibrated and the pyroxenes record a complex subsolidus history. Serra de Mage* is unique in having the highest A12O3 content and hence the highest plagioclase content of any meteorite. In its high plagioclase content, and in other compositional and textural features it is clearly related to Medanitos, Moama, Nagaria, and Moore County. 19. KAPOETA.?The polymict character of this meteorite was clearly established by Fredriksson and Keil (1963), in one of the first applications of micro- probe analysis to meteorite research; they analyzed 132 pyroxene grains from the light portion and found Fe (weight percent) ranging from 8.3 to 28.4, and 132 grains from the dark portion with an Fe range of 7.5 to 26.3; they also analyzed five olivine grains, with Fa ranging from 8.5 to 37.7 mole percent. Their work has been greatly extended by Dymek et al. (1976), who comment: Kapoeta is a "regolith" meteorite, and mineral-chemical and petrographic data were obtained for numerous other rock and mineral fragments in order to characterize the surface and near-surface materials on its parent body. Rock clasts can be grouped into two broad lithologic types on the basis of modal mineralogy?basaltic (pyroxene- and plagioclase- bearing) and pyroxenitic (pyroxene-bearing). Variations in the compositions of pyroxenes in rock and mineral clasts are similar to those in terrestrial mafic plutons such as Skaer- gaard, and indicate the existence of a continuous range in rock compositions from Mg-rich orthopyroxenites to very iron-rich basalts. . . . We interpret these observations to indicate that the Kapoeta meteorite represents the com- minuted remains of differentiated igneous complexes together with "primary" undifferentiated basaltic rocks. The presently available isotopic data are compatible with the interpretation that this magmatism is related to primary differentiation of the Kapoeta parent body. In addition our observations pre- clude the interpretation that the Kapoeta meteorite is a simple mixture of eucrites and diogenites. 20. MALVERN.?This meteorite is extremely simi- lar to Kapoeta in chemical and mineralogical com- position, and in texture, having the same prominent contrasted light-dark areas. It has been described by Simpson (1975) and Desnoyers and Jerome (1977), who clearly established its nature as a polymict breccia. 21. POMOZDINO.?Kvasha and Dyakonova (1972) describe this meteorite as a eucrite. A section we have examined shows that the meteorite is a mono- mict breccia, the breccia fragments consisting of a subophitic intergrowth of plagioclase and pigeonite. 22. BRIENT.?This meteorite is described by Zavaritskii and Kvasha (1952) as a howardite, and their illustrations (page 230) show that it is a polymict breccia. Thus in spite of the close simi- larity in chemical composition between Brient and Pomozdino they are quite different structurally. 23. PETERSBURG.?Thin sections show that this meteorite is a polymict breccia; the dominant clasts are subequal amounts of orthopyroxene and pigeon- ite (up to 1 mm across); plagioclase clasts are also present, and occasional olivine clasts (one olivine clast 2 mm across); small ophitic eucrite enclaves are not uncommon. Microprobe analyses show orthopyroxene and pigeonite ranging in composi- NUMBER 22 35 tion from Fs 22 to Fs 55; plagioclase ranges in com- position from Ab 8 to Ab 14; two grains of olivine, Fa 38 and Fa 42, were analyzed. Duke and Silver (1967, fig. 15) have published a diagram showing the compositional range of pyroxenes in Petersburg. 24. NAGARIA.?A thin section shows that this me- teorite is unbrecciated with a cumulus texture, con- sisting of approximately equal amounts of plagio- clase and pyroxene. It is very similar to Moore County in chemical and mineralogical composition and texture. The minerals are highly equilibrated. Microprobe analyses show plagioclase of uniform composition, Ab 8; the pyroxene evidently crystal- lized originally as pigeonite, but the individual grains are now hypersthene (Wo1.3Fs5o.4En,48-.3) with exsolved augite (Wo43.3Fs18.9En37.8). 25. MOORE COUNTY.?A comprehensive descrip- tion of this meteorite was published by Hess and Henderson (1949). Additional information on the pyroxenes has been provided by Ishii and Takeda (1974); they showed that the primary pigeonite (Wo101Fs44.3En4g 6) has exsolved augite (Wo415Fs217En36 8) and the host pyroxene now has the composition Wo5 8Fs47 5En40 7. Plagioclase has uniform composition, Ab 9. 26. SHERGOTTY.?Although it belongs mineralogi- cally with the pyroxene-plagioclase achondrites, Shergotty (and the closely-related Zagami) has sev- eral features distinguishing it from other meteorites in this group. It contains much more Na2O than the other meteorites, and as a result the plagio- clase composition is that of labradorite instead of bytownite-anorthite; the plagioclase has been shock- transformed into the isotropic form maskelynite. It also contains magnetite, which indicates a much higher oxygen fugacity in its place of crystalliza- tion than for the eucrites and howardites. Duke (1968) has provided a comprehensive description of Shergotty, with analyses of the bulk meteorite, the maskelynite (Or:2Ab49An49), and the pyroxene (Wo22Fs41En37); the pyroxene is strongly zoned, as shown by Duke's microprobe analyses. Binns (1967) has shown that Zagami is closely similar to Shergotty both in texture and mineralogi- cal composition; for an analysis of Zagami, see Easton and Elliott (1977). We have compared thin sections of the two meteorites, and find that Zagami differs from Shergotty only in being somewhat finer- grained (feldspar laths average twice as long in Shergotty); it is certainly possible, if not probable, that both had a common parent body. The chemi- cal composition of Shergotty places it well away from the main trend of the eucrites and howardites (Figure 3). It is significant that Bunch (1975) found no shergottite clasts in the howardites, although he found numerous clasts of eucrites, and less numer- ous clasts of other meteorite types. All these facts indicate that Shergotty (and Zagami) probably come from a different parent body than the other pyroxene-plagioclase achondrites. 27. MACIBINI.?This meteorite has been described by Reid (1974), as follows: The Macibini meteorite is a complex polymict breccia with rock and mineral clasts that are extraordinarily diverse in texture and composition .... Pyroxenes encompass a range from magnesian orthopyroxene through pigeonite and ferro- pigeonite to ferroaugite and to hedenbergite. The trend mim- ics, and extends the pyroxene trend shown by the basaltic achondrite group of meteorites. The more magnesian py- roxenes are coarser grained and tend to occur as mineral clasts; the more iron-rich pyroxenes are commonly present within fine-grained eucritic lithic fragments that have sub- ophitic textures. . . . From the coherence of the mineral data, the parent materials are not a random set of unrelated rocks but appear to be derived predominantly from a suite of related samples, that may have differentiated from a single source The sequence of mineral compositions appears to be continuous and the basaltic achondrites appear to be related by fractional crystallization, involving dominantly pyroxene. There is evidence for the existence of achondrites with compositions intermediate between the hypersthene achondrites and the eucrites; howardites are not simply two- component mixtures. We have examined sections of Macibini and agree with Dr. Reid's comments. Our microprobe analyses show orthopyroxene and pigeonite with a composi- tion range Fs 19-61, together with occasional grains of ferroaugite; plagioclase composition range is Ab 7-21. Desnoyers and Jerome (1973) analyzed 11 grains of olivine in Macibini, with a composition range Fa 56-83 (unusually iron-rich). 28. BIALYSTOK.?This meteorite resembles Maci- bini closely in composition and texture. It is a polymict breccia, with mostly pyroxene clasts (pi- geonite > orthopyroxene) in a fine-grained ground- mass; small ophitic eucrite enclaves are present. Microprobe analyses showed a range in composition of Fs 25-62 for orthopyroxene and pigeonite, and Ab 7-12 for plagioclase; three grains of olivine, Fa 71-73, were analyzed, and rare grains of a silica polymorph, probably tridymite, were seen. 29. YAMATO-74450.?This meteorite, collected by 36 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES the Japanese Antarctic Expedition in 1974, has been analyzed by Wanke et al. (1977), who describe it as a eucrite. We have been informed (H. Takeda, pers. comm.) that Yamato-74450 is a monomict breccia, with pyroxene compositions similar to those in Pasamonte. 30. IBITIRA.?This meteorite has been described by Wilkening and Anders (1975) and Steele and Smith (1976). The latter authors describe it as follows: Ibitira meteorite is interpreted as a strongly metamorphosed, unbrecciated, vesicular eucrite with a primary variolitic and secondary hornfelsic texture dominated by 60% pyroxene (bulk composition En37Fs48Woi5 exsolved into lamellae sev- eral micrometers wide of augite En32Fs27Wo4i and pigeonite En40Fs56Wo4) and 30% plagioclase An94 (mosaic extinction and variable structural state). Minor phases are 5% tridymite plates one-quarter occupied by plagioclase (An94) inclusions; several percent intergrowths of ilmenite and Ti-chromite with trace kamacite Fe99Co05Nio.2 and narrow olivine (Fa8;i) rims; one grain of low-Ti-chromite enclosed in tridymite; trace troilite with kamacite FeggCox.oNio.g. Euhedral ilmenite, Ti-chromite, plagioclase and merrillite in vesicles indicate vapor deposition. These properties can be explained by a series of processes including at least the following: (1) igneous crystallization under pressure low enough to allow vesicula- tion, (2) prolonged metamorphism, perhaps associated with vapor deposition and reduction, to produce the coarse exsolu- tion of the pyroxene and the coarse ilmenite-chromite inter- growths, (3) strong shock which affected the plagioclase and tridymite but not the pyroxene, (4) sufficient annealing to allow recrystallization of the plagioclase and tridymite, and partial conversion to the low structural state of the former. 31. JONZAC.?Thin sections show that this mete- orite is strongly brecciated, only a few small areas retaining the igneous subophitic texture of inter- grown plagioclase and pyroxene. Our microprobe analyses (Figure 1) show pyroxene ranging in com- position from Wo2Fs65En34 to Wo47Fs24En29, the En content remaining almost constant while the Wo and Fs contents vary inversely; this pattern is characteristic for most eucrites, as noted by Duke and Silver (1967) for Sioux County, Juvinas, and Nuevo Laredo. Plagioclase composition ranges from Ab 7-17, with an average of Ab 11. 32. Sioux COUNTY.?Duke and Silver (1967) de- scribe this meteorite as a monomict breccia made up of coarse-grained lithic fragments with equi- granular and ophitic textures. They show the range in composition of the pyroxenes in the form of a diagram (their fig. 15); the pyroxenes range continu- ously from pigeonite through subcalcic augite to ferroaugite with relatively constant molecular per- centage (~30) of MgSiO3. 33. ALLAN HILLS NO. 5.?This meteorite was col- lected in Antarctica in 1977 and described by Olsen et al. (1978). They state: "Thus, in a strict sense, Allan Hills #5 is polymict, however, the excep- tional anorthosite clasts are so few, and the bulk composition so close to the average for eucrites, we classify this meteorite as a eucrite" (p. 223). This is clearly a meteorite whose classification as a eucrite or as a howardite depends on a subjective evalua- tion of the amount of foreign admixture. 34. NOBLEBOROUGH.?Thin sections show this meteorite to be a polymict breccia, with clasts dominantly of pyroxene (mostly pigeonite), minor plagioclase (Ab 7-17, average 9), some tridymite grains, and occasional lithic fragments of eucritic composition and texture; one grain of unusually iron-rich olivine (Fa 88) was found. The polymict character is well illustrated by a plot of pyroxene compositions determined by the microprobe (Fig- ure 1); most of the pyroxene grains are pigeonite, but compositions falling in the fields of hyper- sthene, augite, subcalcic ferroaugite, ferroaugite, and ferrohedenbergite were recorded, and one grain (Wo14Fs82En4) is almost certainly pyroxferroite. The extreme contrast in pyroxene compositions between Jonzac (a monomict breccia) and Nobleborough (a polymict breccia), in spite of their almost identi- cal bulk compositions, supports the division of the pyroxene-plagioclase achondrites into eucrites and howardites on the basis of structure rather than composition. Of course, the mere existence in these meteorites of pyroxenes with a wide range of com- position is not necessarily evidence for a polymict origin; a similar range of composition for a zoned single crystal has been recorded for many pyroxenes in lunar basalts (e.g., Mason et al., 1972). The wide range of pyroxene compositions in Nobleborough, however, has been recorded on discrete grains, not zoned crystals. This indicates that at least some of the rocks from which these grains were derived were probably not cogenetic. 35. CHERVONY KUT.?This meteorite has been described by Zavaritskii and Kvasha (1952). Their illustration (fig. 252) shows an unbrecciated struc- ture, with pigeonite and plagioclase in an equi- granular to subophitic texture. 36. CACHARI.?Cachari is a monomict breccia, extensively crushed, and veined with brown glass; NUMBER 22 37 * *o? FIGURE 1.?Pyroxene compositions (microprobe analyses) in the Nobleborough howardite and the Jonzac eucrite, two pyroxene-plagioclase achondrites with very similar bulk compositions. large areas, however, are unbrecciated, and show an equigranular to subophitic association of pigeonite and plagioclase. Cachari has been described by Fredriksson and Kraut (1967) as a rather typical brecciated eucrite, consisting of orthopyroxene (Wo2Fs59En39) and clinopyroxene (Wo4l2Fs27En31), plagioclase (An 88), with minor amounts of chromite, ilmenite, and troilite. 37. KIRBYVILLE.?This meteorite has never been described; a small portion was kindly made avail- able to us by its owner, Mr. Oscar Monnig. It is a monomict breccia; pyroxene compositions, deter- mined by the microprobe, are similar to those in Jonzac, ranging from Wo,Fs,66En32 to Wo40Fs29En31. A silica polymorph, probably tridymite, was de- tected with the microprobe. 38. JUVINAS.?This meteorite has been extensively described in the literature, most recently by Duke and Silver (1967). They record it as a monomict breccia, and provide a plot of pyroxene composi- tions, which show a continuous range from pigeon- ite through subcalcic ferroaugite to ferroaugite with relatively constant MgSiO3 contents, similar to that of Jonzac and other eucrites. 39. MILLBILLILLIE.?A sawn surface of this me- teorite presents a brecciated appearance, with a prominent light-dark structure. The brecciated ap- pearance is not so marked in a thin section, but can be seen as areas of coarser and finer grain. Our microprobe analyses show pryoxene composi- tions similar to those in Jonzac, ranging from Wo2Fs66En32 to Wo45Fs:26En,29; plagioclase composi- tion range is Ab 9-18, average 12; a silica poly- morph was observed. 40. MOUNT PADBURY.?This meteorite is a meso- siderite containing a varied range of achondrite enclaves, which have been described by McCall (1966). We have examined one of these, an un- brecciated hypidiomorphic-granular aggregate of pyroxene and plagioclase. Its bulk composition is essentially identical with many eucrites. Microprobe analyses show pyroxene compositions ranging from Wo2Fs64En34 to Wo20Fs49En32, i.e., pigeonite to sub- calcic ferroaugite; plagioclase composition is very uniform, Ab 8. 41. PADVARNINKAI.?This meteorite has been classed as a shergottite, because it contains mas- kelynite; however, Binns (1967) showed that it dif- fers significantly in chemistry and mineralogy from Shergotty, and it still retains some untransformed plagioclase, which has a bytownitic composition quite different from the Shergotty maskelynite. The chemical analysis shows that its composition is es- sentially identical with other eucrites. Our micro- probe analyses show pyroxene compositions ranging from Wo,2Fs,67En31 to Wo44Fs27En29, similar to those in Millbillillie; the plagioclase composition range isAb 10-23, average 11. 42. ADALIA.?This meteorite is unbrecciated, with an ophitic texture. Our microprobe analyses show SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES pyroxene ranging from WoaFs^Engg to Wo36Fs34 En30, closely similar to the preceding meteorites; plagioclase ranges Ab 9-18, average 11; accessory grains of a silica mineral were noted. 43. PALO BLANCO CREEK.?This meteorite is de- scribed by Lange and Keil (1976) as a monomict pigeonite-plagioclase achondrite. Their plot of pyroxene compositions show them closely clustered in two areas, one pigeonite and one ferroaugite; our microprobe analyses confirm this, with mean com- positions Wo2Fs65En33 and Wo44Fs28En28; plagio- clase is very uniform in composition, Ab 10. 44. PERAMIHO.?This meteorite was described by Berwerth (1903); it is a typical monomict eucritic breccia. Our microprobe analyses show pyroxene compositions clustered around Wo2Fs63En35 and Wo43Fs27En30, similar to those in Palo Blanco Creek; feldspar composition range is Ab 9-14, average 11. 45. STANNERN.?This meteorite has been compre- hensively described by von Engelhardt (1963). It is a monomict breccia; some areas retain the original ophitic texture of pigeonite and plagioclase, but much of the meteorite is strongly crushed and brec- ciated. Our microprobe analyses show pyroxene ranging in composition from Wo3Fs65En32 to Wo26Fs46En28, and plagioclase from Ab 12 to Ab 20, average 17. 46. HARAIYA.?This meteorite is a monomict breccia, with areas of original ophitic texture sep- arated by areas of crushed material; some of the crushed material appears to have recrystallized into a granular aggregate of pyroxene and plagioclase. Our microprobe analyses show pyroxene composi- tions ranging from Wo2Fs67En31 to Wo43Fs29En2S; plagioclase compositions range Ab 9-18, average 11. 47. PASAMONTE.?This meteorite is a breccia con- taining clasts dominantly of pigeonite with lesser amounts of plagioclase, and lithic fragments with ophitic texture. It has been classified as a monomict breccia by Duke and Silver (1967) and by Takeda, Miyamoto, and Duke (1976). Takeda et al. have shown that Pasamonte is almost unique among the monomict breccias in having strongly zoned pyro- xenes with no visible exsolution lamellae. Our microprobe analyses show pigeonite compositions ranging from Wo6Fs33En61 to Wo6Fsr,5En29, with some Ca-rich pyroxene ranging up to Wo33Fs39En28. 48. BEREBA.?The mineralogy and petrology of this meteorite were comprehensively described by Lacroix (1926). It is a monomict breccia; our mi- croprobe analyses show pyroxene of composition Wo3Fs61En36 with narrow exsolution lamellae of ferroaugite (Wo36Fs35En29), and plagioclase with composition Ab 8-10; a silica polymorph (quartz, according to Lacroix) was detected. 49. EMMAVILLE.?This meteorite was briefly de- scribed by Mason (1974). It is a monomict breccia with veinlets of brown glass. It is noteworthy for being the finest-grained of all the eucrites, with a hornfelsic hypidiomorphic-granular texture, the pyroxene grains averaging 0.02-0.03 mm across and set in a plagioclase matrix. Our microprobe analyses show hypersthene (Wo2Fs66En32) coexisting with augite (Wo47Fs25En28). Plagioclase compositions range from Ab 9 to Ab 20, and average Ab 13. The pyroxene compositions indicate that Emmaville probably recrystallized and equilibrated at tempera- tures well below the liquidus. 50. NUEVO LAREDO.?This meteorite has been de- scribed by Duke and Silver (1967). It is a monomict breccia in which the lithic clasts are rather fine grained. Our microprobe analyses show pyroxene compositions ranging from Wo5Fs68En27 to Wo41Fs34 En24, similar to those reported by Duke and Silver. 51. LAKANGAON.?There is very little published information about this meteorite, except for the analysis by McCarthy et al. (1974). It is a monomict breccia, the individual lithic clasts being coarse- grained granular to subophitic intergrowths of pyroxene and plagioclase. Pyroxene compositions range from Wo5Fs60En27 to Wo43Fs33En24; plagio- clase from Ab 10 to Ab 19, average Ab 15. Neither analytical data nor a description are available for Muckera, which is catalogued by Hutchison et al. (1977) as a howardite. Melrose (b), which we have not examined, was briefly de- scribed and identified as a howardite by Sibray et al. (1976). We accept the identification by Tscher- mak (1872) of Constantinople as a mislabelled frag- ment of Stannern. Chemical and Mineralogical Relationships Examination of Table 1 reveals a number of reg- ularities in the composition of the pyroxene- plagioclase achondrites. The SiO2 content is re- markably uniform at 50 ?3 percent (Serra de Mage", with 46.69% SiO2, is exceptional because of its high content of plagioclase). The FeO content is also fairly uniform (Serra de Mag? is again exceptional), NUMBER 22 39 and the MgO content shows a fairly regular de- crease with increasing FeO/(FeO + MgO) molecu- lar percentage, whereas CaO and A12O3 show the reverse effect. The Na2O content is low through- out, except in Shergotty (1.28%). Of the minor components, MnO is rather constant (Dymek et al. (1976) demonstrated that the FeO/MnO ratio in Kapoeta pyroxenes is fairly uniform (~35) and significantly lower than in lunar pyroxenes (~60), and Marvin (1976) has confirmed this for other pyroxene-plagioclase achondrites); Cr2O3 decreases with increasing FeO/FeO+MgO) molecular per- centage, whereas TiO2 increases. A notable feature of Table 1 is the marked cluster of analyses (analyses 29-49) with FeO/(FeO + MgO) mole percentage of 60?3, i.e., almost half the anal- yses group together, with very limited internal vari- ation. They are unbrecciated or monomict breccias, except Nobleborough (34), and form what may be called the main-group eucrites. Additional correlations are revealed by the cal- culation of normative mineralogy (because the actual minerals present are essentially identical with those calculated in the norm, normative and modal mineralogy are closely similar). Since practically all the A12O3 is combined in plagioclase, A12O3 and plagioclase contents are directly related; plagioclase content increases with increase in FeO/(FeO + MgO) mole percentage, from 12% in Yamato-7307 to almost 40% in main-group eucrites. A few eucrites (Medanitos, Serra de Mage, Nagaria, Moore County) contains more than 40% plagioclase, and are be- lieved to have a cumulate origin. Normative plagio- clase composition is rather uniform at Ab 10?5, except for Shergotty (Ab 48), reflecting the rather uniform Na2O content; in individual meteorites actual plagioclase composition may vary from grain to grain and within grains but will average close to the normative composition. Normative pyroxene in- creases in amount and becomes more calcic and iron- rich with increase in FeO/(FeO + MgO) mole per- centage. Actual pyroxene compositions vary widely in the howardites, but generally show a limited and systematic range in the eucrites (Figure 1). 70 - 60 - Q 50 - 30 - - ? o ? - Howarditat Eucritos Stwrgotty 1 a 9* 7? ? 06?5 ? 3 I 17 *?12 ? ?? ? to 23 I O51SOo 28 _??27 ?36 O24 0 16 o13 1 1 8 CaO 10 12 FIGURE 2.?CaO (weight percent) plotted against FeO/(FeO + MgO) (mole per- cent) from analyses of the pyroxene-plagioclase achondrites (the ellipse encloses values for 20 eucrites and one howardite; numbers are those of analyses in Table 1). 40 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES A significant feature that becomes evident from the norm calculation is the close approach to silica- saturation in all the pyroxene-plagioclase achon- drites, so close in many instances that a change of less than 1% in SiO2 percentage would change an analysis from olivine-normative to quartz-normative. The eucrites, however, are consistently quartz- normative (except Binda, Medanitos, Moama, Nagaria, Moore County), whereas most howardites are olivine-normative. Accessory olivine is present in most howardites, but we have not found it in any of the eucrites we have examined. Most eucrites (even the olivine-normative Binda, Moama, and Moore County) contain accessory free silica, usually as tridymite, sometimes as quartz or cristobalite. Additional relationships are clarified by compo- sition diagrams. Figure 2 plots CaO weight percent- age against FeO/(FeO + MgO) mole percentage. The close groupings of analyses 29 through 49 is ap- parent; there are too many analyses to be shown individually, but the one howardite within this cluster is shown. Nuevo Laredo (50) and Lakangaon (51), although identical to main-group eucrites in CaO content, are clearly differentiated on FeO/ (FeO + MgO) mole percentage. Outside the main- group eucrites are Binda (6) and six probable cumulates: Medanitos (13), Moama (16), Serra de Mage (18), Pomozdino (21), Nagaria (24), and Moore County (25). Another intriguing feature is that howardite compositions are fairly continuous in the ? 4 #3 .7 nl 10 .17 '? ? 12 ? ? ? 1? ? Howardite o Eucrites 0 Shvrgotty .20 ? 22 .27 O*8 ? 26 to 12 20 AI2O, FIGURE 3.?MgO (weight percent) plotted against A12O3 (weight percent) from analyses of the pyroxene-plagioclase achondrites (the ellipse encloses the values for 20 eucrites and one howardite; numbers are those of analyses in Table 1). NUMBER 22 41 FIGURE 4.?Compositions of coexisting calcium-poor and calcium-rich pyroxenes in selected eucrites (1 = Binda (Takeda, Miyamoto, Ishii and Reid, 1976); 2 = Moama (Takeda, Miyamoto, Ishii, and Reid, 1976); 3 = Moore County (Ishii and Takeda, 1974); 4 = Ibitira (Steele and Smith, 1976); 5 = Peramiho (this paper); 6 = Haraiya (this paper); 7 = Lakangaon (this paper)). FeO/(FeO + MgO) range 30-46, but there is then a hiatus to FeO/(FeO + MgO) = 55, with only three beyond this?Macibini, Bialy stole, and Noble- borough. The MgO-Al2O3 diagram (Figure 3) demon- strates many of the same relationships as Figure 2. It shows even more clearly than Figure 2 the linear relationship between the howardites and the main- group eucrites, i.e., although the howardites are polymict breccias, their bulk composition is dearly constrained. The MgO-Al2O3 relationship can be projected to include the diogenites, whose compo- sitions cluster around 26% MgO, 1% A12O3; the relationship between diogenites, howardites, and eucrites will be discussed in a later section. Figure 3 shows how the Shergotty composition falls well away from the eucrites and howardites, and justifies placing it in a special category, as already indicated by its mineralogy. Relative to the eucrites and howardites, Shergotty is enriched in pyroxene and depleted in plagioclase; its texture suggests a pyroxene cumulate. As demonstrated in Figure 1, howardites are dis- tinguished from eucrites by their great range in pyroxene composition; in eucrites pyroxene com- positions may range in Wo content from pigeonite to augite or ferroaugite, but the En content remains practically constant. The extremes in pyroxene com- position are plotted in Figure 4 for seven eucrites, ranging from the most Mg-rich (Binda) to the most iron-rich (Lakangaon). In most of these meteorites all intermediate compositions between the extremes may be found by microprobe analyses, because of the presence of microscopic to submicroscopic ex- solution lamellae of calcium-rich pyroxene in a cal- cium-poor host. The Relationship between Diogenites, Howardites, and Eucrites The fact that the bulk compositions of diog- enites, howardites, and eucrites form an almost continuous sequence has long been known and commented on. Mason (1962) suggested that these three classes might be related as successive mem- bers of a fractional crystallization sequence from a melt of approximately chondritic composition (the pallasites being the initial member of the sequence). Jerome and Goles (1971) pointed out that howard- ites are polymict breccias, and therefore are not direct products of magmatic crystallization; from that observation they proceeded to the postulate that howardites are mechanical mixtures of diog- enites and eucrites. This postulate has been sup ported by McCarthy et al. (1972) and later re- searchers, most recently by Dreibus et al. (1977:211), 42 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES who state: "The hypothesis of Jerome and Goles (1971) and McCarthy et al. (1972), according to which howardites are thought to be mechanical mixtures of eucrites and diogenites, is strongly con- firmed by mixing computations." The fatal flaw in this hypothesis is the remark- able uniformity of pyroxene compositions in diog- enites (Fs 25-27), as has been demonstrated by Mason (1963) and Fredriksson et al. (1976). Cal- cium-poor pyroxene compositions in eucrites (Fig- ure 4) range from Fs 33 to Fs 68. Therefore, if howardites are mixtures of diogenites and eucrites, their pyroxenes should comprise a large component with composition Fs 26?1 (the diogenite fraction) and additional calcium-poor "pyroxenes possibly ranging from Fs 33 to Fs 68 (the eucrite fraction). That this is not the case was already evident in 1971 from the work of Fredriksson and Keil (1963) on Kapoeta; figures 7 and 8 of their paper show pyroxene compositions extending as low as Fs 14 and no large component at Fs 26. These results on Kapoeta have been confirmed and extended by Dymek et al. (1976). Similar evidence is available from Macibini (Reid, 1974) and Malvern (Simp- son, 1975); both Reid and Simpson conclude that these meteorites could not have been produced by the physical mixing of diogenite and eucrite (Des- noyers and Jerome (1977), however, disagree with Simpson in the case of Malvern). Our data on the howardites confirm and extend the conclusions of Reid and Simpson. Not only do the howardites con- tain calcium-poor pyroxenes much more magnesian than those in diogenites, but some of them (nota- bly Pavlovka and Nobleborough) have pyroxenes more iron-rich than have been found in any eucrite. The presence of accessory olivine in many howard- ites, with a wide range of composition?Fa 11-83, according to Desnoyers and Jerome (1973)?also argues for a component not represented either in the diogenites or the eucrites. We agree, neverthe- less, with Reid's comment (1974:398) on Macibini: "The parent materials [of the howardites] are not a random set of unrelated rocks but appear to be derived predominantly from a suite of related sam- ples, that may have differentiated from a single source." The Eucrite Parent Body Discussions of the eucrite parent body usually imply the parent body of what we have called the main-group eucrites, i.e., the compositional cluster around FeO/(FeO + MgO) molecular percentage of 60. Within this group a wide variety of primary (i.e., pre-brecciation) textures are seen, indicating a considerable variety of crystallization conditions in the parent body. Nevertheless, most, if not all, crystallized under low pressure (the common oc- currence of tridymite indicates a confining pressure of 3 kbar or less). Although the textures of the clasts in most of the eucrites resemble those of the lunar basalts, the ex- solution and inversion features show significant differences. The pyroxenes in lunar basalts are strongly zoned, and can range from orthopyroxene through pigeonite to augite and hedenbergite in a single crystal. In most eucrites the pyroxenes are essentially unzoned, and consist of pigeonite of uni- form composition with exsolution lamellae of augite. Possibly the eucrites crystallized originally like the lunar basalts, as thin surface flows, but later were annealed (perhaps by burial under successive flows), whereby originally zoned pyroxenes were homogenized, and then exsolved augite on slow cooling. If true, it is noteworthy that this equilibra- tion occurred without textural change, except for Ibitira and Emmaville, which have a hornfelsic texture. Presumably the absence of water and other volatiles inhibited textural alteration. Some years ago one of us (Mason, 1967b) pro- posed that the eucrites were derived from a rela- tively thin surface crust of a differentiated asteroid consisting of a pallasitic core and a diogenitic mantle. The radius of the hypothetical asteroid was set at 300 km, to be consistent with the measured cooling rates of pallasite meteorites (Buseck and Goldstein, 1967). It had a pallasitic core with radius 207 km, a diogenitic mantle 80 km thick, and a eucritic crust 13 km thick. The overall mineralogical composition of this asteroid was calculated to be (weight percent): nickel-iron, 14; olivine, 25; hyper- sthene, 46; pigeonite, 9; plagioclase, 6. The bulk composition of this asteroid was then calculated from the average mineral compositions in these meteorites, with the results as given in Table 2. It is gratifying to see that recent deductions as to the composition of the eucrite parent body, based on sophisticated geochemical arguments (e.g., Morgan et al., 1978) are in good agreement with the simple petrological model. The composition derived by Morgan et al. (Table 2, analysis 2) would give a NUMBER 22 43 TABLE 2.?Composition of (1) an asteroid consisting of palla- site core, diogenite mantle, and eucrite crust (Mason, 1967b); and (2) the eucrite parent body (Morgan et al., 1978) Constituent SiO2 A12O3 FeO MgO CaO Na20 Fe Ni 1 42.7 2.3 12.7 26.2 2.0 0.1 12.5 1.5 2 35.0 2.2 23.2 24.8 1.8 0.04 11.2 1.7 parent body with more olivine and less pyroxene than that of Mason, i.e., one with a larger pallasite core and smaller diogenite mantle. Literature Cited Allegre, C. J., N. Shimuzu, and M. Treuil 1977. Comparative Chemical History of the Earth, the Moon and Parent Body of Achondite. Philosophical Transactions of the Royal Society of London, series A, 285:55-67. Anders, E. 1977. Chemical Compositions of the Moon, Earth, and Eucrite Parent Body. Philosophical Transactions of the Royal Society of London, series A, 285:23^0. Andersen, O. 1915. The System Anorthite-Forsterite-Silica. American Journal of Science, 39:407-454. Berwerth, F. 1903. Der meteorische Eukrit von Peramiho. Sitzungs- berichten der kaiserlichen Akademie der Wissen- chaften in Wien, mathematische-naturwisseenschaft- liche Klasse, 112(l):739-777. Binns, R. A. 1967. Stony Meteorites Bearing Maskelynite. Nature, 213: 1111-1112. Bunch, T. E. 1975. Petrography and Petrology of Basaltic Achondrite Polymict Breccias (Howardites). Proceedings of the Sixth Lunar Science Conference, 1:469-492. Bunch, T. E., S. Chang, J. M. Neil, and A. Burlingame 1976. Unique Characteristics of the Jodzie Howardite. Meteoritics, 11:260-261. Buseck, P. R., and J. I. Goldstein 1967. Olivine Compositions and Cooling Rate of Pallasitic Meteorites. Transactions of the American Geophysi- cal Union, 48:165-167. Consolmagno, G. J., and M. J. Drake , 1977. Composition and Evolution of the Eucrite Parent Body: Evidence from Rare Earth Elements. Geo- chimica et Cosmochimica Acta, 41:1271-1282. Desnoyers, C, and D. Y. Jerome 1973. Olivine Compositions in Howardites and Other Achondritic Meteorites. Meteoritics, 8:344-345. 1977. The Malvern Howardite: A Petrological and Chem- ical Discussion. Geochimica et Cosmochimica Acta, 41:81-86. Dreibus, G., H. Kruse, B. Spettel, and H. Wanke 1977. The Bulk Composition of the Moon and the Eucrite Parent Body. Proceedings of the Eighth Lunar Science Conference, 1:211-227. Duke, M. B. 1968. The Shergotty Meteorite: Magma tic and Shock Meta- morphic Features. In B. M. French and N. M. Short, editors, Shock Metamorphism of Natural Materials, pages 613-621. Baltimore, Md.: Mono Book Corporation. Duke, M. B., and L. T. Silver 1967. Petrology of Eucrites, Howardites, and Mesoside- rites. Geochimica et Cosmochimica Acta, 31:1637- 1665. Dyakonova, M. I., and V. Y. Kharitonova 1961. The Chemical Composition of Eighteen Stony Meteorites in the Collection of the U.S.S.R. Acad- emy of Sciences. Meteoritika, 21:52-59. Dymek, R. F., A. L. Albee, A. A. Chodos, and G. J. Wasser- burg 1976. Petrography of Isotopically-dated Clasts in the Kapoeta Howardite and Petrologic Constraints on the Evolution of Its Parent Body. Geochimica et Cosmochimica Acta, 40:1115-1130. Easton, A. J., and C. J. Elliott 1977. Analyses of Some Meteorites from the British Mu- seum (Natural History) Collection. Meteoritics, 12: 372-374. Fredriksson, K., and K. Keil 1963. The Light-Dark Structure in the Pantar and Kapoeta Stone Meteorites. Geochimica et Cosmo- chimica Acta, 27:717-740. Fredriksson, K., and F. Kraut 1967. Impact Glass in the Cachari Eucrite. Geochimica et Cosmochimica Acta, 31:1701-1704. Fredriksson, K., A. Noonan, P. Brenner, and C. Sudre 1976. Bulk and Major Phase Composition of Eight Hypersthene Achondrites. Meteoritics, 11:278-280. Frost, M. J. 1971. The Molteno Meteorite. Mineralogical Magazine, 38:89-93. 44 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Fukuoka, T., M. S. Ma, and R. A. Schmitt 1977. Genesis of Howardites as Mixtures of Eucrites and Diogenites. Lunar Science VIII, pages 340-342. Hess, H. H., and E. P. Henderson 1949. The Moore County Meteorite: A Further Study with Comment on Its Primordial Environment. American Mineralogist, 34:494-507. Hutchison, R., A.W.R. Bevan, and J. M. Hall 1977. Appendix to the Catalogue of Meteorites with Special Reference to Those Represented in the Col- lection of the British^ Museum (Natural History). 297 pages. London: British Museum (Natural His- tory). Ishii, T., and H. Takeda 1974. Inversion, Decomposition and Exsolution Phe- nomena of Terrestrial and Extraterrestrial Pigeon- ites. Memoirs of the Geological Society of Japan, 11:19-36. Jeremine, E. 1954. Eucrite de Chaves (Portugal). Boletim da Sociedada Geologica de Portugal, 11:127-138. Jerome, D. Y., and G. G. Goles 1971. A Re-examination of Relationships among Pyro- xene-Plagioclase Anchondrites. In A. O. Brunfelt and E. Steinnes, editors, Activation Analysis in Geo- chemistry and Cosmochemistry, pages 261-266. Oslo, Norway: Universitetsforlaget. Jerome, D. Y., and M. Christophe Michel-Levy 1971. Quelques aspects remarquables de la meteorite achondritique de Malvern (Afrique du Sud). Bulletin de la Socie'te francaise de Mineralogie et de Cris- tallographie, 94:156-161. 1972. The Washougal Meteorite. Meteoritics, 7:449-461. Kvasha, L. G., and M. I. Dyakonova 1972. The Pomozdino Eucrite. Meteoritics, 31:109-115. Lacroix, A. 1926. L'eucrite de Bereba (Haute-Volta) et les meteorites feldspathiques en general. Archives du Museum d'Histoire naturelle, series 6, 11:15-58. Lange, D. E., and K. Keil 1976. Meteorites of Northeastern New Mexico. New Mexico Geological Society Guidebook, 27th Field Conference, pages 293-299. Lovering, J. F. 1975. The Moama Eucrite?a Pyroxene-Plagioclase Ad- cumulate. Meteoritics, 10:101-114. McCall, G. J. H. 1966. The Petrology of the Mount Padbury Mesosiderite and Its Achondritic Enclaves. Mineralogical Maga- zine, 35:1029-1060. McCarthy, T. S., L. H. Ahrens, and A. J. Erlank 1972. Further Evidence in Support of the Mixing Model for Howardite Origin. Earth and Planetary Science Letters, 15:86-93. McCarthy, T. S., A. J. Erlank, and J. P. Willis 1973. New Chemical Analyses of Six Achondrites and One Chondrite. Meteoritics, 9:215-221. Marvin, U. B. 1976. A Unique Eucritic Gabbro from the Descartes Highlands. Meteoritics, 11:329-330. Mason, B. 1962. Meteorites. 274 pages. New York: John Wiley and Sons. 1963. The Hypersthene Achondrites. American Museum Novitates, 2155: 13 pages. 1967a. The Bununu Meteorite, and a Discussion of the Pyroxene-Plagioclase Anchondrites. Geochimica et Cosmochimica Acta, 31:107-115. 1967b. Meteorites. American Scientist, 51:429-455. 1974. Notes on Australian Meteorites. Records of the Australian Museum, 29:169-186. Mason, B., E. Jarosewich, W. G. Melson, and G. Thompson 1972. Mineralogy, Petrology, and Chemical Composition of Lunar Samples 15085, 15256, 15271, 15471, 15475, 15535, 15555, and 15556. Proceedings of the Third Lunar Science Conference, 1:785-796. Mason, B., and H. B. Wiik 1966a. The Composition of the Bath, Frankfort, Kakangari, Rose City, and Tadjera Meteorites. American Mu- seum Novitates, 2272: 24 pages. Morgan, J. W., H. Higuchi, T. Takahashi, and J. Hertogen 1978. A "Chondritic" Eucrite Parent Body: Inference from Trace Elements. Geochimica et Cosmochimica Acta, 42:27-38. Olsen, E., A. Noonan, K. Fredriksson, E. Jarosewich, and G. Moreland 1978. Eleven New Meteorites from Antarctica, 1976-1977. Meteoritics, 13:209-225. Prinz, M., C. E. Nehru, J. L. Berkley, K. Keil, E. Jarosewich, and C. B. Gomes 1977. Petrogenesis of the Serra de Mage Cumulate Eucrite. Meleoritics, 12:341. Prior, G. T. 1920. The Classification of Meteorites. Mineralogical Mag- azine, 19:51-63. Reid, A. M. 1974. The Macibini Meteorite and Some Thoughts on the Origin of Basaltic Achondrites. Meteoritics, 9:398- 399. Schmitt, R. A., G. G. Goles, R. H. Smith, and T. W. Osborn 1972. Elemental Abundances in Stone Meteorites. Meteo- ritics, 7:131-124. Sibray, S., K. Keil, P. M. Starzyk, R. A. Schmitt, and G. I. Huss 1976. The Melrose (b) Howardite: A Complex Meteorite Parent Body Regolith Breccia. Transactions of the American Geophysical Union, 56:1017. Simpson, A. B. 1975. Electron Microprobe Investigation of the Howardite Malvern. Meteoritics, 10:489-491. Steele, I. M., and J. V. Smith 1976. Mineralogy of the Ibitira Eucrite and Comparison with Other Eucrites and Lunar Samples. Earth and Planetary Science Letters, 33:67-78. Stolper, E. 1977. Experimental Petrology of Eucritic Meteorites. Geochimica et Cosmochimica Acta, 41:587-611. Symes, R. F., and R. Hutchison 1970. Medanitos and Putinga, Two South American Me- teorites. Mineralogical Magazine, 37:721-723. NUMBER 22 45 Takeda, H., and T. Ishii 1975. Pyroxene Geothermometry Applied to Lunar Crus- tal Rocks and Meteorites. Meteoritics, 10:499-500. Takeda, H., M. Miyamoto, and M. B. Duke 1976. Pasamonte Pyroxenes, a Eucritic Analogue of Lunar Mare Pyroxenes. Meteoritics, 11:372-374. Takeda, H., M. Miyamoto, T. Ishii, and A. M. Reid 1976a. Characterization of Crust Formation on a Parent Body of Achondrites and the Moon by Pyroxene Crystallography and Chemistry. Proceedings of the Seventh Lunar Science Conference, 3:3535-3548. Tschermak, G. 1872. Die Meteoriten von Stannern, Constantinopel, Sher- gotty und Gopalpur. Mineralogische Mittheilungen, 1872:83-100. Urey, H. C, and H. Craig 1953. The Composition of the Stone Meteorites and the Origin of the Meteorites. Geochimica et Cosmoclii- mica Ada, 4:36-82. Von Engelhardt, W. 1963. Die Eukrit von Stannern. Beitrage zur Mineralogie und Petrographie, 9:65-94. Wanke, H., H. Palme, H. Baddenhausen, G. Dreibus, E. Jagoutz, H. Kruse, B. Spettel, F. Teschke, and R. Thacker 1974. Chemistry of Apollo 16 and 17 Samples: Bulk Com- position, Late Stage Accumulation and Early Dif- ferentiation of the Moon. Proceedings of the Fifth Lunar Science Conference, 2:1307-1355. Wanke, H., H. Baddenhausen, K. Blum, M. Cendales, G. G. Dreibus, H. Hofmeister, H. Kruse, E. Jagoutz, C. Palme, B. Spettel, R. Thacker, and E. Vilcsek 1977. On the Chemistry of Lunar Samples and Achond- rites: Primary Matter in the Lunar Highlands, a Re-evaluation. Proceedings of the Eighth Lunar Science Conference, 2:2191-2213. Wahl, W. 1952. The Brecciated Stony Meteorites and Meteorites Containing Foreign Fragments. Geochimica et Cos- mochimica Ada, 2:91-117. Wiik, H. B. 1969. On Regular Discontinuities in the Composition of Meteorites. Commentationes Physico-Mathematicae, Societas Scientarium Fennica, 34:135-145. Wilkening, L. L., and E. Anders 1975. Some Studies of an Unusual Eucrite: Ibitira. Geo- chimica et Cosmochimica Ada, 39:1205-1210. Zavaritskii, A. N., and L. G. Kvasha 1952. Meteorites of the U.S.S.R. 245 pages. Moscow: Acad- emy of Sciences of the U.S.S.R. Fusion of Rock and Mineral Powders for Electron Microprobe Analysis Peter A. Jezek, John M. Sinton, Eugene Jarosewich, and Charles R. Obermeyer ABSTRACT Rock and mineral powders, finer than 100 mesh,were fused in a molybdenum or tungsten boat in a nitrogen atmosphere. The fusion times ranged from5 to 30 sec at temperatures between 1650? C and 1750? C. The fusion bridge was powered by a1.5 kVA welding transformer. This fusion technique in combination with theelectron microprobe uses very small amounts of sample and is suitable for very rapid analysis ofrock samples ranging in composition from 45 to about 65 weight percent SiO2. In general, the glassesare sufficiently homogeneous to be suitable for elec- tron microprobe analysis. The electron microproberesults, when compared to analyses obtained by clas- sical chemical methods, have average accuracy better than 5 percent (relative) for all major componentsexcept TiO 2, which is better than 13 percent (rela-tive). A major part of the error in the TiO 2 analysesis probably due to the presence of minute residual or quench Fe-Ti oxides in the glass. The samples investigated suggest that if fusiontimes of less than 20 sec are used, Na 2O loss is pre-vented or greatly restricted and homogeneous glasses still result. Introduction Brown (1977) summarized techniques used in electron microprobe major elements analysis of whole rock samples. He also described a new tech- nique for fusing rock powders in molybdenum boats in a pressurized (60 psi) argon atmosphere. Peter A. Jezek, John M. Sinton, Eugene Jarosewich, and Charles R. Obermeyer, Department of Mineral Sciences, Na- tional Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. We have investigated direct fusion of rock and mineral powders in molybdenum and tungsten boats in a nitrogen atmosphere (at atmospheric pressure), using a strip furnace which is easily built and inexpensive to operate. These experiments were aimed at developing a rapid, simple, and inex- pensive method of fusion of rock powders into representative, homogeneous beads or shards suit- able for analysis by electron microprobe, without the need of an argon pressure chamber. INSTRUMENTATION.?The fusion bridge (Figures 1, 2) is powered by a 1.5 kVA welding transformer. The primary of the transformer is connected through a Variac autotransformer to 115 VAC. The secondary of the transformer, rated at 11.5 V, is connected to a 100 A silicon diode bridge rectifier. The DC output of the rectifier then powers the fusion bridge (Figure 3). To reach a temperature of 1670? C the strip receives 40 A at 4 VDC. Tem- peratures were measured by an optical pyrometer. The particular welding transformer employed in this instrument was used because of its availability; two parallel-wired filament transformers, 25 A each, could be used instead. The Mo and W boats used in the fusion are inex- pensive and readily available from most SEM sup- pliers. They are very practical for holding all molten glasses, and are especially useful in restrict- ing the how of low viscosity basaltic glasses. The size best suited for routine sample fusion is a 1% in. strip containing a boat 7/16 in. long, 3/16 in. wide and 3/32 in. deep. TECHNIQUE.?The melting boat, connected be- tween two posts, is covered by a glass bell jar that is continuously flushed by nitrogen (Figure 1). After placing the powder, ground finer than 100 mesh, 46 NUMBER 22 47 FIGURE 1.?Front view of the fusion furnace: the fusion boat is mounted between the two posts and the whole assembly is covered by a glass bell jar that is continuously flushed by nitrogen (for symbol explanation see Figure 3). FIGURE 2.?Detail of the fusion bridge (P = positive and negative terminals of the bridge; B = fusion boat; Q = out- let of the nitrogen quench line; O = outlets through which the gas under the glass bell jar is forced out; S = vertical shield behind which is located the outlet for the purge line through which nitrogen enters the glass bell jar during the flushing cycle (see Figure 3). in the fusion boat and covering the fusion assembly by the bell jar it takes about 1 min to establish an atmosphere of almost pure nitrogen. This is docu- mented by the lack of oxidation of a Mo strip heated in this atmosphere at 1750? C for 25 min. Two melting procedures were used. (1) Samples were heated for about 10 sec at low temperature (dark red glow of the strip) to drive off adsorbed water and water of hydration if present. The tem- perature was then rapidly increased to approxi- mately 1700? C. The powders usually melt in 3-5 sec but total heating at the final temperature for about 10-30 sec promotes diffusion and thus formation of more homogeneous glasses. (2) Rapid heating of the cold powder to about 1700? C was also investigated. When subjected to this technique the powders fuse more rapidly probably due to the presence of water, which is driven off when the first technique is used. There is no apparent difference in the homogeneity of the glasses produced by these two techniques. Quenching of the melts is accomplished by simul- taneously shutting off the power to the fusion bridge and directing a stream of nitrogen onto the bottom of the fusion boat. Cooling is sufficiently fast to prevent crystallite formation in all but the most mafic samples. Quench olivine is common in the more mafic samples. In order to avoid either strip-contaminated glass, glass possibly depleted by elemental diffusion into the strip (Brown, 1977), or elemental volatilization into the nitrogen atmosphere, only glass in the in- terior of the bead or shard was analyzed. The glasses were analyzed by a 9-spectrometer computer-automated ARL-SEMQ electron micro- probe. The spectrometers were equipped with the following analyzing crystals. Si-EDDT, Al-EDDT, 48 F1 T1 QUENCH PURGE SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES R1T2 NITROGEN FIGURE 3.?Wiring diagram of the fusion furnace (SI = DPDT 115-VAC 10 A switch; Fl = 10 A fuse; Tl - VARIAC auto- transformer; T2 = welding transformer; BR1 = 100 A 50 PIV silicon diode bridge rectifier; Rl = meter shunt, 150 A 50 mV; Ml = 50 mV full scale meter calibrated to read 150 A at full scale). Lower left: a diagram of the nitrogen system (the purge line is located behind the vertical shield (S in Figure 2); the quench line is located below the fusion boat (Q in Figure 2); VI = 3 way valve, toggle actuated). Fe-LiF, Mg-ADP, Ca-LiF, Na-RAP, K-LiF, P-ADP. The data were corrected by an on-line computer us- ing the method of Bence and Albee (1968). During the analysis 15 kV accelerating potential and 0.02 fiA sample current were used. A defocused beam, 20-50 pjm in diameter, practically eliminates Na volatilization during analysis and also helps to aver- age local, small inhomogeneities which may exist in the glass. It is recommended that 10 sec count times be used and the analysis repeated at least ten times to accumulate a statistically significant body of data. The repeated measurements also help to average possible larger scale inhomogeneities. Results Eleven rock powders and three mineral samples were fused by the described technique. The rocks were selected to cover as wide a compositional range as possible. Analytical data for the eleven rock samples are presented in Table 1. The electron microprobe data show good agreement with the wet chemical values, which were recalculated volatile free in order to facilitate direct comparison with the electron microprobe analyses. The CV values (CV = standard deviation x 100/mean) reflect sig- nificantly the homogeneity of the glasses but also in- clude a component of variation due to instrumental instability. The standards used in the glass analyses were analyzed by wet chemical techniques at the Department of Mineral Sciences, Smithsonian In- stitution. Their compositions are given in Jarose- wich, Nelen, and Norberg (this volume). The average differences between the wet chemical analyses and the electron microprobe analyses are shown in Table 2. The average accuracy in weight percent is better than 0.2 for all major components except SiO2 which is accurate to better than 0.3. In TABLE 1.?Analyses of 11 fused rock powder samples by wet chemical and electron microprobe methods (A = wet chemical analysis (wt %); B = electron microprobe analysis (wt %), numbers in parentheses are the numbers of point analyses averaged; CV (approximate measure of glass homogeneity) = (SD x 100)/x; ND = no data; Std (standard or reference sample) 1 = VG-568 (glass), 2 = VG-A99 (glass), 3 = tektite glass (synthetic), 4 = VG-2 (glass), 5 = Kakanui horn- blende; wet chemical analyses of W-1, BCR-1, and AGV-1 from Flanagan, 1973; remaining wet chemical analyses by E. Jarosewich and J. Norberg, Department of Mineral Sciences, Smithsonian Institution; all wet chemical analyses recalculated as volatile free to facilitate direct comparison with microprobe analyses; standard analyses given in Table 1, Jarosewich, Nelen, and Norberg, this volume) Constituent SiO2 TiO2 A12O3 FeO MnO MgO CaO Na2O K2O P2?5 Total Temp. (?C) Time (sec) I Constituent sio2 TiO2 A12O3 FeO MnO MgO CaO Na2O K2O p2o5 Total Temp. (aC) Time (sec) U3180/B159(b) A 47.32 2.87 16.99 11.12 0.19 7.60 10.26 2.79 0.69 0.28 100.11 A 54.50 0.64 19.49 7.28 0.16 4.87 9.46 2.90 0.62 0.16 100.08 B(14) 47.41 2.67 16.85 10.97 ND 7.65 10.13 2.96 0.63 ND 99.27 1700 30 CV 1.07 3.19 1.99 1.43 2.40 1.55 2.42 10.17 113419 B(16) 54.51 0.60 19.75 7.04 0.17 4.84 9.30 3.17 0.60 ND 99.88 1670 30 CV 1.63 6.32 2.20 3.05 4.07 1.35 3.03 8.27 Std 4 5 4 4 5 4 3 3 Std 3 5 5 2 2 2 1 3 A 51.09 0.92 18.92 11.26 0.15 3.16 10.17 2.49 1.42 0.14 99.72 A 55.37 2.24 13.83 12.30 0.18 3.52 7.03 3.32 1.73 0.36 99.88 D-14-53 B(26) 51.19 0.98 19.14 11.11 0.14 3.34 10.11 2.30 1.30 ND 99.61 1750 30 CV 0.42 3.95 1.77 1.93 4.99 1.14 3.51 4.73 Std 3 5 5 2 2 2 1 3 FUSION BCR-1 B(29) 55.21 2.09 13.79 12.26 ND 3.48 7.06 3.26 1.66 ND 98.81 1670 CV 0.78 4.06 2.42 2.35 4.33 1.79 2.25 2.75 Std 4 5 4 4 5 4 4 3 FUSION 10, 20 A 50.99 1.87 19.27 9.60 0.16 4.05 9.35 3.50 0.57 0.23 99.59 MI-8 B(46) 51.23 2.06 19.02 9.81 ND 4.01 9.33 3.58 0.57 ND 99.61 CONDITIONS A 56.64 0.95 19.80 6.33 0.11 3.48 6.92 4.58 1.18 0.15 100.14 1650 20 CV 1.30 4.54 2.96 2.21 3.99 2.73 2.40 9.13 SH-384-2 B(25) 57.86 0.80 18.77 5.88 0.11 3.73 6.78 4.46 1.20 ND 99.59 CONDITIONS 1670 21 CV 1.99 9.29 3.65 3.81 3.84 1.34 2.92 6.75 Std 4 5 4 4 4 4 4 3 Std 3 5 5 2 2 2 1 3 A 53.00 1.08 15.10 10.05 0.17 6.67 11.04 2.16 0.65 0.14 100.06 A 60.14 1.06 17.58 6.23 0.10 1.56 4.99 4.34 2.95 0.50 99.45 W-1 B(29) 53.14 1.05 15.44 9.85 ND 6.51 11.15 2.14 0.62 ND 99.90 1670 15, 20 CV 1.29 5.33 1.90 2.56 2.51 1.46 2.92 7.77 AGV-1 B(22) 60.46 0.92 17.59 6.20 ND 1.55 5.09 4.05 2.84 ND 98.70 1670 15 CV 0.27 6.48 1.56 3.24 4.35 1.72 2.49 2.21 Std 4 5 4 4 5 4 4 3 Std 4 5 4 4 5 4 4 3 50 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 1.?Continued. Constituent SiO2 TiO2 A12O3 FeO MnO MgO CaO Na20 K20 P2?5 Total Temp. (?C) Time (sec) A 63.76 0.66 17.69 4.48 0.08 1.91 5.12 4.85 1.40 0.13 100.08 SH-281 B(24) 63.67 0.60 17.78 4.33 0.08 1.87 5.05 4.95 1.33 ND 99.66 1670 10 CV 1.01 5.30 1.23 2.16 4.71 1.09 2.10 2.82 Std 3 2 5 3 3 3 1 1 A 64.01 0.32 12.95 6.79 0.13 5.47 7.51 1.83 0.95 0.06 99.96 111108 B(16) 63.37 0.36 13.06 6.65 0.13 5.84 7.34 2.01 0.86 ND 99.62 CV 1.65 7.99 2.72 2.35 2.73 1.92 3.16 7.04 Std 3 2 5 3 3 3 1 1 FUSION CONDITIONS 1700 15 A 76.70 0.12 12.42 1.23 0.03 <0.1 0.50 3.74 4.88 <0.01 99.73 VG-568 B(21) 76.48 0.18 12.33 1.24 ND ND 0.54 3.71 4.89 ND 99.37 1750 10 CV 1.19 15.58 1.65 2.49 12.15 3.11 4.36 Std 1 1 1 1 1 1 1 TABLE 2.?Absolute differences and relative differences (percents of the amounts present) between the wet chemical and electron microprobe analyses presented in Table 1 Constituent sio2 TiO2 A12O3 FeO MgO CaO Na2O K20 Range of Samples 47.32 - 0.12 - 12.42 - 1.23 - <0.01 - 0.50 - 1.83 - 0.57 - Analyzed (wt %) 76.70 2.87 18.92 12.30 7.60 10.26 4.85 4.89 Average Absolute Difference (wt %) 0.21 0.10 0.14 0.16 0.12 0.09 0.13 0.05 Range Absolute Difference (wt %) 0.09 - 0.64 0.04 - 0.20 0.01 - 0.34 0.01 - 0.45 0.01 - 0.37 0.02 - 0.17 0.02 - 0.18 0.00 - 0.12 Average Relative Difference a) 0.4 12.7 0.9 2.2 2.8 1.9 4.5 4.5 Range Relative Difference (%) 0.1 - 1.0 2.8 - 50.0 0.1 - 2.3 0.3 - 7.1 0.6 - 7.2 0.2 - 8.0 0.8 - 9.8 0.0 - 9.5 NUMBER 22 51 relative percent the average accuracy is better than 5 for all major components except TiO2 which is better than 13. The somewhat poorer accuracy in measuring TiO2 is due in part to some very low measured concentrations. In such cases a small error (e.g., 0.06 wt % of TiO2 in sample VG-568) repre- sents as much as 50% relative error. However, the major part of the error is probably due to the pres- ence of residual or quench Fe-Ti oxides observed in the form of opaque spots in some glasses. Due to their very small size a positive identification is difficult. Attempts to fuse granitic samples and MgO-rich, low-silica samples to homogeneous glasses were un- successful. USGS samples G-2, GSP-1, and G-l pro- duced very inhomogeneous glasses when less than 100 mesh powders were used. Subsequent testing showed that if very fine powders are used more homogeneous glasses result. The required prepara- tion of very fine powders and the increased possi- bility of Na2O loss on fusion make routine fusion of high-silica samples impractical. The MgO-rich, low-silica samples could not be prevented from growing skeletal olivine crystals during quenching. Three mineral samples (Table 3) and a sample of natural rhyolite glass (VG-568, Table 1) were fused to evaluate: (1) Na^O loss, (2) the behavior of hydrous minerals during fusion, and (3) the im- pact of rapid dehydration on the composition of the resulting glass. It is clear that prolonged fusion of the samples results in Na2O loss. But fusion times ranging up to 20 sec produce good glasses with minimal Na^O loss. The three fusion times used in the fusion of the Lake County plagioclase sample (Table 3) did not produce significantly large differences in the Na2O TABLE 3.?Comparison of analyses of crystalline minerals with analyses of the same minerals fused for various times (A = electron microprobe analyses of the unfused minerals; B = electron microprobe analyses of fused minerals; numbers in parentheses are the numbers of point analyses averaged; C = wet chemical analyses of crystalline mineral recalculated water free, based on data in Table 1, Jarosewich, Nelen, and Norberg, this volume) Constituent SiO2 TiO2 Al2O.j FeO MgO CaO Na2O K20 Total Temp. (?C) Time (sec) Lake Co. USNM A (13) 51.29 ___ 31.07 0.50 ___ 13.79 3.33 ... 99.98 (17) 51.27 30.78 0.50 ? 13.70 3.38 .? 99.63 1670 5 Plagioclase 115900 B (13) 51.04 -__ 30.78 0.50 --_ 13.64 3.45 ? 99.41 1670 (21) 51.30 31.01 0.49 ___ 13.77 3.30 ? 99.87 1730 25 1 A (14) 55.36 0.37 8.81 4.58 11.69 13.86 4.95 0.15 99.77 Omphacite JSNM 110607 B (18) 55.01 0.37 8.69 4.56 12.01 13.64 5.10 0.17 99.55 FUSION CONDITIONS 1700 10 (12) 55.48 0.37 8.60 4.63 12.03 13.67 4.20 0.15 99.13 1730 30 Arenal Hornblende USNM 111356 C 42.36 1.44 15.81 11.72 14.55 11.80 1.95 0.21 99.84 B (16) 42.18 1.22 15.56 11.79 14.97 11.63 2.12 0.21 99.68 1670 5 52 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES concentration. However, the 25 sec fusion probably records a small Na2O loss, and the 30 sec fusion of omphacite has clearly resulted in a significant Na2O loss. Arenal hornblende frothed extensively during fusion due to the escape of structural water (1.21 wt % H2O). The composition of the resulting glass is very close to the crystalline mineral composition recalculated water free. Although it was anticipated that Na2O may be volatilized more rapidly during H2O release from hydrous minerals, no Na2O loss was observed in three separate glasses, each pro- duced by 5 sec fusion. The hornblende sample was not fused longer than 5 sec because the water present in the mineral facilitates rapid fusion. No Na2O was lost during the 10 sec fusion of VG-568. This sample is a natural glass and therefore provides information only on Na2O stability during fusion. Summary and Conclusions The described fusion technique uses very small amounts of sample and is rapid. The furnace can be built and operated inexpensively. When used in conjunction with an electron microprobe, major element concentrations can be measured rapidly and quite accurately for rock samples ranging in composition from basalts to high-silica andesites. More mafic samples have a tendency to crystallize olivine on cooling. Granites and other high-silica, high-alkali samples produce glasses of high viscosity preventing homogenization during the short fusion times necessary to prevent Na2O loss. Fusion of ex- tremely fine powders may decrease the inhomo- geneity of the resulting glass and improve the re- sulting analysis. In general the fused glasses are sufficiently homo- geneous and the electron microprobe is suitably accurate to allow routine use of the technique. The average accuracy of better than 5% relative for all major components except TiO2 (better than 13%) is quite acceptable. The rapidity and convenience with which the samples can be prepared and analyzed allowed ma- jor element reconnaissance analyses of rock suites previously not analyzed due to time and cost factors involved. The rock powders and the three mineral samples investigated suggest that if fusion times of less than 20 sec are used Na2O loss is prevented or greatly restricted and glasses result that are suffi- ciently homogeneous to be suitable for microprobe analysis. Literature Cited Bence, A. E., and A. L. Albee 1968. Empirical Correction Factors for Electron Micro- analysis of Silicates and Oxides. Journal of Geology, 76:382-403. Brown, R. W. 1977. A Sample Fusion Technique for Whole Rock Analysis with the Electron Microprobe. Geochimica et Cosmochimica Ada, 41:435-438. Flanagan, F. J. 1973. 1972 Values for International Geochemical Refer- erence Samples. Geochimica et Cosmochimica Ada, 37:1189-1200. Microprobe Analyses of Four Natural Glasses and One Mineral: An Interlaboratory Study of Precision and Accuracy Eugene Jarosewich, Alan S. Parkes, and Lovell B. Wiggins ABSTRACT An interlaboratory study for precision and accuracyof electron microprobe analyses of four natural glasses and one mineral is reported in this compila- tion. The results obtained by the three participatinglaboratories are in good agreement with chemical analysis values (where available) for SiO2, A12O3,FeO, CaO, and K^O (> 2 wt %). One labora- tory shows a considerable bias for Na^O and a slight bias for MgO. Some results for MnO, K2O(/?0.2 wt %), and P 2O5, considering their low con-centrations, are in reasonably good agreement; others show scatter. In general, considering thatdifferent reference samples and operating parame- ters were used to obtain these analyses, the correla-tion of results is encouraging. On the basis of the available data, however, it is evident that the preci-sion and accuracy of these results cannot be im- proved much without more elaborate data acquisi- tion techniques. Introduction The need for accurate microprobe analyses of natural glasses is of great importance in the study of volcanic and deep sea rocks. Melson et al. (1977) briefly summarized this need, referring specifically to the analyses of glasses from the Deep Sea Drilling Project. Since a standardized method for analysis of such glasses is not available, each laboratory has developed its own techniques and has been using its Eugene Jarosewich, Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Wash- ington, D.C. 20650. Alan S. Parkes, Department of Earth and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139. Lovell B. Wiggins, U.S. Geo- logical Survey, Reston, Virginia 22092. own preferred reference samples for these analyses. It is rather remarkable that many of the results obtained on the same or similar samples, using dif- ferent techniques and reference samples, are gen- erally in good agreement. Occasionally, however, some of these results exceed accepted analytical errors. Since more than one institution may partici- pate in the study of volcanic or deep sea rocks and the analytical data may be used interchangeably for petrologic studies, it is of utmost importance that the precision and accuracy of these analyses be de- termined. As a start in obtaining this type of data, a set of the same samples has been analyzed by three laboratories, each using its own standard operating parameters; these results are compared for precision and accuracy. The participating laboratories were: Massachusetts Institute of Technology (MIT), Smithsonian Institution (SI), and U.S. Geological Survey (USGS). The data presented in this paper are meant to serve primarily as a general guide for the expected accuracy and precision of microprobe analyses of glasses; however, they can serve as a guide for such analyses of minerals as well. ACKNOWLEDGMENTS.?We would like to acknowl- edge Drs. W. G. Melson, G. R. Byerly and T. L. Wright for their suggestions pertaining to this work and Dr. J. S. Huebner for revising the data. We would also like to express our appreciation to Ms. J. Norberg of the Smithsonian Institution for com- pilation and arrangement of the data. Experimental Procedure One disc was prepared containing only one grain of Kakanui hornblende and only one grain of each 53 54 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES of the following natural glasses: VG-2 (USNM 111240-52), Basalt with a few rare micro- phenocrysts of olivine and plagioclase in section, less than 1%. Chemical analysis, Table 4. VG-A99 (USNM 113498-1). Basalt from Makaopuhi Lava Lake, Hawaii (Wright and Okamura, 1977). Micropheno- crysts of plagioclase in section, about 1% or less. Chem- ical analysis, Table 4. VG-999 (USNM 113155-614). High-titanium ferro-basalt from DeSteiguer dredge D6, Galapagos Spreading Center (Byerly et al., 1976). Phenocrysts of plagioclase, augite, and olivine. Microphenocrysts of plagioclase and augite, about 10%?15% in section. VG-D08 (USNM 113154-557). Basalt from DeSteiguer dredge D5, Galapagos Spreading Center (Byerly et al., 1976). Phenocrysts of olivine, plagioclase, and minor spinel. No microphenocrysts in section. This disc was carbon coated and sent to the three laboratories for microprobe analysis. These four glasses were selected because they represent the approximate elemental ranges en- countered in the study of natural basaltic glasses. Classical chemical analyses are available for Kakanui hornblende, VG-2, and VG-A99 for comparison with the microprobe results (see Table 1, Jarosewich, Nelen, and Norberg, this volume). Although no chemical analyses are available for VG-999 and VG-D08, these two samples were included in this study to obtain additional data on the precision among the microprobe analyses themselves. The same grain was analyzed by each laboratory to elimi- nate the variable of possible inhomogeneities be- tween different grains of the same sample. Since the primary purpose of this work was to obtain analytical values for the samples as they are obtained under the laboratories' normal operating conditions, the operating parameters (kv, /iA, beam size, etc.) were not specified. The samples were analyzed employing the standard operating param- eters and reference samples used in each laboratory (Tables 1, 2). The precision was obtained by analyzing all sam- ples using Kakanui hornblende as the reference sample. In this manner discrepancies due to the use of different reference samples were eliminated and any deviations could be ascribed to the instrumen- tal parameters. Since Kakanui hornblende is not an ideal reference sample for obtaining accurate anal- yses of these glasses, the accuracy was determined by analyzing these samples using each laboratory's pre- ferred reference samples and comparing results with the chemical analyses of these glasses. The analytical results were compiled in the following manner. (1) Each sample was analyzed several times using Kakanui hornblende as the reference sample; the number of individual analyses and the averages are given in Tables 5-7. (2) The samples were analyzed several times using each laboratory's preferred ref- erence samples; the number of individual analyses, the averages, and the type of reference samples used in these analyses are also given in Table 5-7. (3) The uncorrected averages obtained using Kakanui hornblende as the reference sample are summarized in Table 3 for comparisons of precision; these re- sults indicate the variation of the data due only to the instrumental variation. (4) Table 4 summarizes the corrected results obtained using each labora- tory's preferred reference samples; comparisons of TABLE 1.?Instrumental parameters used by participating laboratories Laboratory MIT SI USGS Instrument ETEC MAC 5 ARL-SEMQ 9 spectrometers ARL-EXM-SM Accelerating Potential (kV) 15 15 15 Sample Current (yA) 0o03 0o05 0.05 Beam Size (ym) 4 2, 30 20 Counting Time 30 sec or 60,000 counts Seven 10 sec counts for each analysis 20 sec or 20,000 counts for each element NUMBER 22 55 TABLE 2.?Analyses of reference samples used by participating laboratories Reference sample DJ35 P-140 AN-60 MAC AP Orthoclase Di2Ti Mnllm Coss Kak hornblende VG-2 Apatite Fayalite Rhodonite Garnet-Pyrope Orthoclase Di2Ti Diopside-Jadeite SiO2 56.88 40.86 53.05 0.05 64.39 54.36 40.93 40.37 50.81 0.34 29.22 46.76 41.45 63.42 54.39 56.10 A12?3 8.82 30.01 18.58 14.90 14.06 0.07 0.96 23.50 19.24 3.78 FeO 7.23 0.03 45.41 40.99 10.92 11.83 0.05 67.54 12.49 10.76 0.10 MgO 12.10 51.63 0.01 18.26 0.25 12.80 6.71 "0.42 18.80 18.26 15.84 CaO MET 16.83 12.38 38.82 25.38 SI 10.30 11.12 54.02 USGS 5.62 5.09 0.08 25.38 22.02 Na2O 5.36 4.56 0.18 1.14 6.93 2.60 2.62 0.23 0.36 2.31 K 14 2 0 15 2? .92 .05 .19 .34 TiO2 2.00 51.61 8.71 4.36 1.85 0.04 0.51 2.00 P 17 0 40 0 2?5 .86 .20 .78 .49 MnO 1.50 0.09 0.22 2.14 33.34 0.33 these results with the chemical values indicate the accuracy of the microprobe results. Discussion Comparisons for precision of the results obtained using Kakanui hornblende as the reference sample (Table 3) indicate that careful microprobe analyses for all major elements, titanium, and high potas- sium (>2 wt %) produce relatively precise results. The results also indicate that minor elements and sodium vary considerably and more careful work must be done if the results for these elements are to be used for exacting petrological work. There is no pronounced bias in the results by any of the three laboratories for the major elements except for a slight bias in the magnesium results. There is a bias in the results for sodium, low potassium (~0.2 wt %), and phosphorus. The sodium and phosphorus results analyzed by laboratory 1 run consistently high. The potassium results by labora- tory 2 run slightly high and by laboratory 3, low. The discrepancy for these three elements is signifi- cant because all elements were determined using the same reference sample, indicating that the tech- niques for acquiring the data should be examined. Comparisons for precision of the results of micro- probe analyses in which each laboratory's preferred reference samples were used (Table 4) also check favorably for the major elements, titanium, and potassium (> 2 wt %); more important, these re- sults are in excellent agreement with the values for the three samples analyzed by classical chemical methods, thus showing a high level of accuracy. The results for sodium show a much larger discrepancy than those given in Table 3; only the lowest values of the microprobe analyses check well with the chemically analyzed sodium values. Sodium fre- quently presents a problem in microprobe analysis and to demonstrate this problem, laboratory 2 ana- lyzed sodium using a 2 /im beam and a 30 /on beam; the former gives somewhat lower values for some samples (Table 6, results in brackets), the latter gives acceptable values. Kakanui hornblende so- dium results are not affected by the beam size. One of the possible explanations for the higher results of laboratory 1 and lower results of laboratory 2 with the small beam is the unique behavior of sodium in the electron beam for both the reference samples and the unknown samples. If for example a reference sample in which sodium is easily "vola- tilized" (accepted description of decreasing inten- 56 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 3.?Summary of uncorrected data from microprobe analyses of the 5 samples, using Kakanui hornblende as the reference sample, for interlaboratory comparison of precision (figure in parentheses indicates number of analyses averaged; each analysis represents up to 10 individual point analyses; laboratory 1 = MIT, 2 = SI, 3 = USGS; data based on Tables 5-7) Constituent SiO2 A12?3 FeO MgO CaO Na2O K20 TiO2 P2?5 MnO Lab 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 KH 40.45 40.58 40.31 14.8814.97 14.98 10.85 10.91 10.81 12.83 12.67 12.92 10.20 10.39 10.24 2.74 2.59 2.67 2.02 2.05 1.96 4.36 4.38 4.27 0.09 0.06 0.08 0.07 0.09 0.19 (2) (4) (3) (2) (4) (3) (2) (4) (3) (2) (4) (3) (2) (4) (3) (2) (2) (3) (2) (4) (3) (2) (4) (3) (2) (2) (3) (2) (2) (3) VG-2 51.07 51.62 51.05 14.69 15.06 14.85 11.96 11.84 11.78 7.02 6.90 6.87 11.12 11.16 11.01 2.90 2.65 2.78 0.18 0.20 0.06 1.66 1.67 1.64 0.24 0.20 0.20 0.16 0.18 0.25 (2) (4) (4) (2) (4) (4) (2) (4) (4) (2) (4) (4) (2)(4) (4) (2) (2) (4) (2) (4) (4) (2) (4) (4) (2) (2) (4) (2) (2) (4) VG-A99 52.49 52.45 52.54 13.23 13.54 13.13 13.57 13.60 13.46 5.29 4.98 4.99 9.38 9.29 8.80 2.91 2.71 2.71 0.77 0.83 0.72 3.65 3.70 3.77 0.44 0.39 0.42 0.16 0.18 0.22 (1) (4) (2) (1) (4) (2) (1)(4) (2) (1) (4) (2) (1) (4) (2) (1) (2) (2) (1) (4) (2) (1) (4) (2) (1) (2) (2) (1) (2) (2) VG-999 52.27 52.21 51.88 13.61 14.12 14.41 13.79 13.71 13.66 5.98 5.81 5.57 10.39 10.40 10.26 2.89 2.60 2.66 0.14 0.17 0.03 1.72 1.74 1.66 0.22 0.19 0.17 0.17 0.20 0.13 (2) (4) (1) (2) (4) (1) (2) (4) (1) (2) (4) (1) (2) (4) (1) (2) (2) (1) (2) (4) (1) (2) (4) (1) (2) (2) (1) (2) (2) (1) VG-D08 50.17 50.56 50.36 16.60 17.27 16.71 9.00 8.97 8.88 9.36 8.71 8.70 12.48 12.39 12.13 2.77 2.38 2.40 0.060.08 0.00 0.98 1.01 0.92 0.16 0.12 0.12 0.16 0.14 0.16 (1)(4) (2) (1)(4) (2) (1)(4) (2) (1)(4) (2) (1)(4) (2) (1)(2) (2) (1)(4) (2) (1)(4) (2) (1)(2) (2) (1)(2) (2) slty) is used for standardization to analyze an unknown sample which does not show this vola- tilization effect, the sodium results for the unknown sample will be high. If on the other hand the so- dium in the reference sample is "stable" and sodium in the unknown sample is easily volatilized, low re- sults will be obtained. The use of a larger beam size may diminish this discrepancy. Low potassium, manganese, and phosphorus re- sults vary just as much as those in Table 3. It should be noted that phosphorus corrects up by 15%-20% in most cases, giving much higher values than those of the chemical analyses. These higher phosphorus corrections are evident with both the up-dated Bence-Albee (Albee and Ray, 1970) and Magic IV correction procedures. Both correction procedures performed by one of us (A. P.) give ex- cellent agreement with each other for all elements (Table 5). Since for this study some grains were analyzed up NUMBER 22 57 TABLE 4.?Summary of corrected data from microprobe analyses of the 5 samples using each laboratory's preferred reference samples, along with chemical analyses of 3 of the samples, for interlaboratory comparison of precision and accuracy (figures in parentheses indicates number of analyses averaged; each analysis represents up to 10 individual point analyses; laboratory 1 = MIT, 2 = SI, 3 = USGS; Chem. = chemical analysis from Table 1, Jarosewich, Nelen, and Norberg, this volume; microprobe data based on Tables 5-7) Constituent Lab SiO9 Z Al90o Z j FeO MgO CaO Nao0 z K00 z TiOo z P90 MnO 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) 1 2 3 (Chem.) KH 40.65 40.72 40.42 40.37 14.61 14.95 14.76 14.90 10.45 10.93 10.54 10.92 12.95 12.70 13.03 12.80 9.99 10.47 10.12 10.30 2.80 2.60 2.72 2.60 2.10 2.06 1.94 2.05 4.87 4.40 4.65 4.36 0.14 0.07 0.04 0.00 0.10 0.11 0.06 0.09 (5) (4) (2) (5) (4) (2) (5) (4) (2) (5) (4) (2) (5) (4) (2) (5) (2) (2) (5) (4) (2) (5) (4) (2) (5) (2) (2) (5) (2) (2) VG-2 50.85 50.72 50.75 50.81 13.81 14.15 13.98 14.06 11.26 11.79 11.79 11.83 7.01 6.78 7.02 6.71 10.85 11.14 10.72 11.12 3.17 2.66 2.75 2.62 0.20 0.21 0.18 0.19 1.86 1.91 1.86 1.85 0.32 0.23 0.19 0.20 0.22 0.22 0.23 0.22 (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (2) (1) VG-A99 51.05 51.22 50.80 50.90 12.59 12.66 12.80 12.97 13.24 13.47 13.41 13.18 5.24 4.95 5.16 5.18 9.08 9.28 8.97 9.38 2.81 2.70 2.73 2.73 0.82 0.90 0.76 0.80 4.04 4.05 3.77 4.06 0.54 0.46 0.31 0.41 0.19 0.22 0.19 0.19 (4) (4) (1) (4) (4) (1) (4) (4) (1) (4) (4) (1) (4) (4) (1) (4) (2) (1) (4) (4) (1) (4) (4) (1) (4) (2) (1) (4) (2) (1) VG-999 51.16 51.41 50.70 13.06 13.36 13.54 12.94 13.68 13.30 5.99 5.77 6.13 10.12 10.36 10.18 3.06 2.61 2.57 0.14 0.19 0.04 1.93 1.96 1.80 0.29 0.22 0.15 0.29 0.25 0,20 (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (2) (1) VG-D08 50.29 50.18 49.77 15.80 16.17 15.44 8.67 8.96 8.78 8.88 8.46 8.78 12.23 12.41 12.05 2.79 2.25 2.39 0.07 0.09 0.09 1.09 1.14 1.03 0.23 0.15 0.08 0.18 0.17 0.13 (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (4) (1) (1) (4) (1) (1) (2) (1) (1) (2) (1) 58 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 5.?MIT laboratory individual analyses of Kakanui hornblende and 4 natural glasses using Kakanui hornblende as reference sample and also the laboratory's preferred reference samples Run No. 1 2 4 5a 5b 6 8 1 Conditions Standards Uncorrected B-A Corr. Magic IV Corr. a (5) Uncorrected B-A Corr. Magic IV Corr. a (6) Standards Uncorrected B-A Corr. Magic IV Corr. o (10) Uncorrected B-A Corr. Magic IV Corr. o (5) Uncorrected B-A Corr. Magic IV Corr. a (5) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) SiO2 KH 40.39 40.35 40.35 0.17 40,51 40.55 40.56 0.53 DJ35 38.21 40.29 40.36 0.34 38.92 40.97 41.07 0.21 38.71 40.75 40.85 0.32 DJ35 38.32 40.43 40.52 0.30 DJ35 38.71 40.82 40.92 0.19 KH 51.36 49.98 50.08 0.51 A12O3 KH 14.69 14.68 14.68 0.11 15.07 15.07 15.08 0.08 AN-60 12.61 14.45 14.54 0.08 12.66 14.48 14.57 0.08 12.68 14.50 14.59 0.08 AN-60 12.83 14.71 14.80 0.09 AN-60 13.00 14.89 14.97 0.09 KH 14.51 14.03 14.03 0.13 FeO KH 10.99 10.99 10.99 0.15 10.70 10.71 10.70 0.22 P-140 10.48 10.37 10.43 0.27 10.34 10.23 10.29 0.12 10.29 10.18 10.24 0.13 Coss. 10.26 10.68 10.60 0.12 Cost.. 10.35 10.. 77 10.70 0.18 KH 12.15 12.15 12.15 0.14 MgO Kakanui KH 12.57 12.59 12.59 0.13 13.09 13.08 13.07 0.22 P-140 11.68 12.98 13.04 0.26 11.58 12.86 12.88 0.13 11.51 12.78 12.82 0.17 DJ35 12.38 13.10 13.18 0.10 DJ35 12.34 13.05 13.14 0.15 KH 6.83 6.89 6.88 0.08 CaO Na2O hornblende KH 10.13 10.13 10.13 0.15 10.27 10.28 10.28 0.07 DJ35 10.11 9.92 9.96 0.14 10.20 10.02 10.05 0.18 10.18 9.99 10.03 0.13 DJ35 10.18 9.98 10.02 0.07 DJ35 10.23 10.04 10.09 0.10 VG-2 KH 11.03 11.03 11.04 0.08 KH 2.71 2.71 2.71 0.12 2.77 2.76 2.75 0.12 DJ35 2.55 2.87 2.90 0.09 2.55 2.86 2.89 0.11 2.59 2.91 2.93 0.08 DJ35 2.49 2.80 2.83 0.11 DJ35 2.29 2.57 2.60 0.07 KH 2.82 2.84 2.84 0.11 K20 KH 2.03 2.03 2.04 0.02 2.00 2.00 2.00 0.04 Ortho 2.10 2.05 2.05 0.02 2.16 2.11 2.10 0.04 2.15 2.10 2.09 0.04 Ortho 2.16 2.11 2.10 0.05 Ortho 2.16 2.11 2.10 0.05 KH 0.19 0.20 0.19 0.00 TiO2 KH 4.34 4.34 4.35 0.03 4.37 4.38 4.37 0.03 Di2Ti 5.13 4.94 4.94 0.08 5.05 4.87 4.87 0.05 5.11 4.92 4.93 0.05 Di2Ti 5.03 4.84 4.85 0.05 Mnllm 4.20 4.79 4.73 0.05 KH 1.67 1.67 1.67 0.02 p2o5 AP 0.10 0.12 0.12 0.00 0.08 0.10 0.12 0.00 MAC AP 0.10 0.12 0.12 0.00 0.11 0.13 0.15 0.00 0.11 0.15 0.15 0.00 MAC AP 0.11 0.15 0.15 0.02 MAC AP 0.13 0.16 0,15 0.00 AP 0.23 0.30 0.30 0.02 MnO KH 0.04 0.04 0.05 0.01 0.10 0.10 0.10 0.03 Mnllm 0.10 0.11 0.11 0.03 0.12 0.13 0.12 0.01 0.11 0.12 0.12 0.03 Mnllm 0.04 0.04 0.04 0.04 Mnllm 0.11 0.11 0.11 0.03 KH 0.14 0.14 0.14 0.04 NUMBER 22 59 Table 5.?Continued Run No. 2 5 2 4 5 6 8 Conditions Uncorrected B-A Corr. Magic IV Corr. a (10) Standards Uncorrected B-A Corr. Magic IV Corr. o (6) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) Uncorrected B-A Corr. Magic IV Corr. o (10) Standards Uncorrected B-A Corr. Magic IV Corr. o (5) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) SiO2 50.77 49.52 49.61 0.41 DJ35 49.61 50.85 51.05 0.83 KH 52.49 50.68 50.84 0.43 DJ35 50.22 51.03 51.30 0.28 50.25 51.00 51.42 0.49 DJ35 50.32 51.17 51.43 0.32 DJ35 50.09 50.99 51.26 0.53 A12O3 14.87 14.39 14.39 0.19 AN-60 12.49 13.81 13.88 0.13 KH 13.23 12.78 12.78 0.15 AN-60 11.18 12.36 12.44 0.09 11.20 12.34 12.44 0.08 AN-60 11.50 12.71 12.78 0.06 AN-60 11.73 12.95 13.03 0.11 FeO 11.76 11.77 11.76 0.18 P-140 11.39 11.26 11.33 0.14 KH 13.57 13.53 13.55 0.09 P-140 13.25 13.06 13.16 0.19 12.85 12.68 12.77 0.18 Coss. 13.11 13.61 13.53 0.19 Coss. 13.09 13.59 13.50 0.14 MgO 7.21 7.26 7.25 0.10 P-140 6.27 7.01 7.02 0.06 KH 5.29 5.42 5.41 0.08 P-140 4.61 5.24 5.26 0.10 4.54 5.16 5.17 0.07 DJ35 4.88 5.27 5.31 0.03 DJ35 4.88 5.27 5.30 0.10 CaO 11.20 11.20 11.21 0.07 DJ35 11.05 10.85 10.91 0.08 VG-A99 KH 9.38 9.36 9.37 0.10 DJ35 9.11 8.91 8.96 0.06 9.24 9.04 9.10 0.08 DJ35 9.29 9.08 9.14 0.01 DJ35 9.51 9.30 9.36 0.11 Na2O 2.97 2.99 2.99 0.12 DJ35 2.80 3.17 3.19 0.11 KH 2.91 2.99 2.99 0.08 DJ35 2.63 3.05 3.08 0.08 2.55 2.95 2.98 0.15 DJ35 2.24 2.61 2.63 0.11 DJ35 2.24 2.61 2.63 0.12 K20 0.17 0.17 0.17 0.01 Ortho 0.20 0.20 0.20 0.00 KH 0.77 0.77 0.77 0.02 Ortho 0.84 0.82 0.82 0.01 0.84 0.83 0.82 0.02 Ortho 0.83 0.81 0.81 0.02 Ortho 0.85 0.83 0.83 0.02 TiO2 1.65 1.66 1.65 0.03 Di2Ti 1.93 1.86 1.86 0.07 KH 3.65 3.64 3.64 0.03 Di2Ti 4.30 4.12 4.12 0.05 4.28 4.10 4.11 0.05 Di2Ti 4.27 4.08 4.09 0.05 Mnllm 3.41 3.87 3.83 0.07 P2?5 0.24 0.31 0.33 0.00 MAC AP 0.25 0.32 0.33 0.02 AP 0.44 0.56 0.56 0.02 MAC AP 0.42 0.54 0.53 0.02 0.41 0.53 0.53 0.02 MAC AP 0.43 0.55 0.56 0.00 MAC AP 0.42 0.54 0.53 0.02 MnO 0.18 0.18 0.18 0.04 Mnllm 0.20 0.22 0.22 0.03 KH 0.16 0.16 0.15 0.03 Mnllm 0.19 0.20 0.20 0.03 0.17 0.18 0.18 0.03 Mnllm 0.18 0.20 0.19 0.01 Mnllm 0.16 0.18 0.18 0.03 60 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Table 5.?Continued Run No. 1 2 5 2 5 Conditions Standards Uncorrected B-A Corr. Magic IV Corr. a (5) Uncorrected B-A Corr. Magic IV Corr. o (10) Standards Uncorrected B-A Corr. Magic IV Corr. o (10) Standards Uncorrected B-A Corr. Magic IV Corr. a (10) Standards Uncorrected B-A Corr. Magic IV Corr. o (10) SiO2 KH 52.20 50.64 50.79 0.34 52.34 50.77 50.91 0.68 DJ35 50.07 51.16 51.42 0.45 KH 50.17 49.25 49.30 0.43 TXT35 48.68 50.29 50.43 0.60 A12O3 KH 13.66 13.24 13.24 0.17 13.56 13.15 13.16 0.42 AN-60 11.80 13.06 13.16 0.08 KH 16.60 16.02 16.06 0.55 AN-60 14.35 15.80 15.86 0.06 FeO KH 14.03 13.99 14.01 0.20 13.55 13.52 13.53 0.17 P-140 13.11 12.94 13.03 0.17 KH 9.00 9.04 9.04 0.09 P-140 8.73 8.67 8.72 0.10 MgO KH 5.85 5.97 5.98 0.05 6.11 6.24 6.24 0.07 P-140 5.30 5.99 6.02 0.07 KH 9.36 9.20 9.20 0.18 P-140 8.13 8.88 8.88 0.12 CaO VG-999 KH 10.23 10.21 10.24 0.07 10.54 10.52 10.54 0.08 DJ35 10.33 10.12 10.17 0.06 VG-D08 KH 12.48 12.53 12.54 0.08 DJ35 12.41 12.23 12.29 0.06 Na2O KH 2.75 2.82 2.82 0.09 3.02 3.09 3.08 0.09 DJ35 2.66 3.06 3.09 0.20 KH 2.77 2.70 2.70 0.09 DJ35 2.55 2.79 2.81 0.12 K2? KH 0.15 0.15 0.16 0.00 0.13 0.13 0.13 0.00 Ortho 0.14 0.14 0.14 0.01 KH 0.06 0.07 0.06 0.01 Ortho 0.07 0.07 0.07 0.00 TiO2 KH 1.70 1.69 1.70 0.03 1.73 1.73 1.73 0.03 Di2Ti 2.01 1.93 1.94 0.03 KH 0.98 0.99 0.99 0.03 Di2Ti 1.12 1.09 1.08 0.03 *2?5 AP 0.21 0.27 0.27 0.02 0.23 0.30 0.30 0.00 MAC AP 0.22 0.29 0.30 0.02 AP 0.16 0.21 0.21 0.00 MAC AP 0.18 0.23 0.24 0.02 MnO KH 0.13 0.13 0.13 0.01 0.20 0.20 0.21 0.03 Mnllm 0.27 0.29 0.29 0.03 KH 0.16 0.16 0.17 0.03 Mnllm 0.17 0.18 0.18 0.03 to 40 times to obtain good precision, it is reasonable to assume that, on the basis of the data in Table 3, the precision of microprobe analyses of basaltic glasses cannot be much improved without more elaborate data acquisition techniques. The standard deviations for various elements given in Tables 5-7 are of similar magnitudes for the three laboratories and they realistically indicate the expected scatter of analytical results. It is obvious that longer counts would not improve the precision very much and, indeed, extended counting times are impractical in day-to-day operation. The accuracy, on the other hand, can be in some cases improved by the selection of suitable reference samples for a given unknown sample. From the data in Tables 5-7 it is evident that the microprobe re- sults for some elements are in better agreement with the chemical results when reference samples other than Kakanui hornblende are used. The standard deviation here again is similar to that of the precision and thus could also be used as an indication of the expected scatter of the results. Analysis of elements below 0.1 wt % should be done with extra care and accepted with caution. NUMBER 22 61 TABLE 6.?SI laboratory individual analyses of Kakanui hornblende and 4 natural glasses using Kakanui hornblende as reference sample and also the laboratory's preferred reference samples Run No. 1 2 3 4 1 2 3 4 1 2 3 Conditions Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. a (7) Standards Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. a (7) Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) SiO2 KH 41.09 41.09 0.49 40.45 40.46 0.28 40.03 40.13 0.38 40.76 40.76 0.23 KH 41.09 41.09 0.49 40.45 40.46 0.28 40.44 40.51 0.17 40.83 40.83 0.37 KH 52.04 51.18 0.42 51.32 50.49 0.73 50.89 50.19 0.48 A12O3 KH 15.05 15.02 0.23 14.83 14.82 0.24 14.95 14.98 0.21 15.06 15.02 0.28 KH 15.05 15.02 0.23 14.83 14.82 0.24 15.00 15.02 0.20 14.95 14.92 0.22 KH 15.10 14.55 0.18 14.90 14.38 0.16 15.12 14.64 0.19 FeO KH 10.87 10.88 0.26 11.11 11.11 0.27 10.88 10.88 0.12 10.76 10.77 0.08 KH 10.87 10.88 0.26 11.11 11.11 0.27 10.95 10.96 0.14 10.76 10.76 0.07 KH 12.01 12.03 0.10 12.01 12.03 0.07 11.82 11.83 0.11 MgO Kakanui KH 12.78 12.75 0.43 12.60 12.60 0.38 12.79 12.83 0.26 12.51 12.49 0.21 KH 12.78 12.75 0.43 12.60 12.60 0.38 12.77 12.82 0.17 12.63 12.61 0.25 KH 6.85 6.87 0.19 7.00 7.03 0.09 6.91 6.98 0.21 CaO Na2O hornblende KH 10.70 10.72 0.26 10.22 10.24 0.04 10.22 10.23 0.07 10.40 10.42 0.14 KH 10.70 10.72 0.26 10.22 10.24 0.09 10.28 10.30 0.11 10.57 10.60 0.14 VG-2 KH 11.22 11.32 0.13 11.19 11.29 0.14 10.96 11.06 0.17 KH T.51" 2.58 0.08 2.60 2.59 _0.0_7. 2.60 2.60 0.02 2.58 2.58 0.01 KH T.58" 2.58 0.08 2.60 2.59 0_.0_7_ 2.62 2.62 0.04 2.58 2.57 0.03 KH F.53 2.53 0.11 2,53 2.53 .0.06 2.69 2.69 0.03 K20 KH *1.98 1.98 0.08 2.13 2.13 0.07 2.04 2.04 0.05 2.05 2.05 0.06 KH 1.98 1.98 0.08 2.13 2.13 0.07 2.03 2.03 0.04 2.08 2.08 0.06 KH 0.19 0.19 0.02 0.20 0.20 0.02 0.21 0.21 0.02 TiO2 KH 4.33 4.34 0.12 4.45 4.45 0.11 4.43 4.43 0.04 4.29 4.30 0.07 KH 4.33 4.34 0.12 4.45 4.45 0.11 4.51 4.51 0.09 4.30 4.30 0.06 KH L.67 1.68 0.07 1.71 1.72 0.01 1.69 1.70 0.03 P2?5 AP 0.05 0.07 0.02 0.06 0.07 0.02 AP 0.05 0.07 0.02 0.06 0.07 0.02 AP i 0.20 0.25 0.03 0.19 0.24 0.03 MnO KH 0.09 0.09 0.01 0.09 0.09 0.01 Fay 0.11 0.12 0.03 0.10 0.10 0.02 KH 0.17 0.17 0.02 62 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Table 6.?Continued Run No. 4 1 2 3 4 1 2 3 4 1 2 Conditions Uncorrected B-A Corr. a (7) Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. o (7) Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. o (7) Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. ? (7) SiO2 52.21 51.32 0.43 VG-2 50.56 50.57 0.48 50.69 50.71 0.51 50.93 50.99 0.59 50.52 50.60 0.61 KH 52.56 51.39 0.45 52.76 51.53 0.29 51.65 50.63 0.66 52.82 51.63 0.58 VG-2 51.08 50.78 0.75 51.71 51.38 0.28 A12O3 15.11 14.55 0.19 VG-2 14.14 14.13 0.25 14.08 14.08 0.26 14.15 14.16 0.19 14.21 14.23 0.24 KH 13.72 13.19 0.12 13.51 12.99 0.22 13.47 13.03 0.27 13.46 12.96 0.24 VG-2 12.68 12.65 0.23 12.65 12.62 0.20 FeO 11.53 11.55 0.09 VG-2 11.72 11.73 0.15 11.54 11.56 0.09 11.87 11.88 0.19 11.99 12.00 0.11 KH 13.90 13.88 0.08 13.59 13.57 0.06 13.62 13.59 0.12 13.30 13.29 0.17 VG-2 13.48 13.45 0.08 13.42 13.39 0.26 MgO 6.83 6.85 0.12 VG-2 6.57 6.57 0.11 7.01 6.99 0.15 6.69 6.70 0.11 6.83 6.84 0.08 KH 4.89 4.97 0.20 4.98 5.06 0.11 5.07 5.19 0.13 4.98 5.06 0.10 VG-2 4.78 4.83 0.06 4.95 5.00 0.44 CaO 11.28 11.39 0.14 VG-2 11.10 11.11 0.09 11.18 11.19 0.13 11.13 11.13 0.16 11.13 11.13 0.18 VG-A99 KH 9.30 9.36 0.08 9.34 9.38 0.18 9.07 9.10 0.17 9.46 9.50 0.08 VG-2 9.18 9.13 0.04 9.36 9.32 0.45 Na2O 2.60 2.59 0.03 VG-2 T.62 2.62 0.11 2.47 2.46 JD.21 2.64 2.65 0.06 2.66 2.66 0.05 KH T.3T 2.37 0.10 2.22 2.26 2.72 2.75 0.03 2.70 2.75 0.03 VG-2 2~.4T 2.47 0.17 2.27 2.32 0.10 K20 0.20 0.20 0.03 VG-2 0.22 0.21 0.02 0.22 0.21 0.02 0.23 0.22 0.07 0.22 0.21 0.03 KH 0.84 0.84 0.12 0.82 0.83 0.04 0.80 0.81 0.02 0.84 0.85 0.02 VG-2 0.93 0.90 0.06 0.91 0.88 0.04 TiO2 1.62 1.63 0.05 VG-2 1.94 1.94 0.06 1.92 1.93 0.05 1.92 1.92 0.03 1.85 1.85 0.07 KH 3.66 3.66 0.06 3.68 3.68 0.03 3.81 3.81 0.10 3.65 3.66 0.06 VG-2 4.05 4.02 0.09 4.12 4.09 0.11 p2o5 AP 0.19 0.24 0.02 0.18 0.22 0.03 AP 0.37 0.45 0.04 0.41 0.51 0.04 AP 0.38 0.47 0.04 0.37 0.45 0.01 MnO 0.18 0.18 0.02 Fay 0.20 0.22 0.02 0.19 0.22 0.02 KH 0.18 0.18 0.02 0.18 0.18 0.02 Fay NUMBER 22 Table 6.?Continued Run No. 3 4 1 2 3 4 1 2 3 4 Conditions Uncorrected B-A Corr. CT (7) Uncorrected B-A Corr. o (7) Standards Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. ? (7) Standards Uncorrected B-A Corr. ? (7) Uncorrected B-A Corr. ? (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) SiO2 51.77 51.51 0.90 51.46 51.21 0.62 KH 52.46 51.47 0.70 52.17 51.20 0.77 51.89 51.02 0.61 52.31 51.35 0.44 VG-2 51.64 51.52 0.58 51.77 51.67 0.64 51.46 51.42 0.38 51.06 51.04 0.41 A12O3 12.74 12.72 0.22 12.64 12.65 0.13 KH 14.22 13.69 0.29 14.04 13.54 0.18 14.01 13.55 0.07 14.20 13.69 0.17 VG-2 13.32 13.30 0.13 13.35 13.34 0.22 13.37 13.37 0.10 13.39 13.42 0.03 FeO 13.39 13.36 0.11 13.73 13.69 0.14 MgO 4.83 4.89 0.11 5.00 5.07 0.14 VG-999 KH 13.91 13.89 0.11 13.76 13.75 0.13 13.59 13.58 0.03 13.56 13.55 0.13 VG-2 13.53 13.51 0.17 13.59 13.56 0.20 13.75 13.73 0.07 13.95 13.92 0.09 KH 5.71 5.77 0.11 5.96 6.03 0.06 5.83 5.93 0.15 5.72 5.80 0.11 VG-2 5.57 5.62 0.10 5.85 5.89 0.16 5.66 5.72 0.17 5.80 5.86 0.20 CaO 9.38 9.31 0.12 9.41 9.35 0.14 KH 10.59 10.66 0.19 10.43 10.50 0.03 10.10 10.17 0.13 10.47 10.54 0.07 VG-2 10.30 10.28 0.11 10.43 10.42 0.10 10.33 10.29 0.15 10.46 10.43 0.22 Na2O 2.64 2.70 0.03 2.63 2.69 0.06 KH T.31 2.35 0.11 2.32 2.36 ?.?92. 2.60 2.63 0.01 2.60 2.63 0.05 VG-2 rr.4ir 2.52 0.11 2.38 2.41 2.54 2.57 0.09 2.60 2.65 0.06 K20 0.94 0.91 0.04 0.92 0.89 0.04 KH 0.16 0.16 0.03 0.17 0.17 0.02 0.18 0.18 0.03 0.16 0.16 0.00 VG-2 0.21 0.20 0.02 0.16 0.16 0.02 0.20 0.20 0.03 0.18 0.18 0.03 TiO2 4.05 4.03 0.17 4.08 4.05 0.14 KH 1.71 1.71 0.06 1.75 1.75 0.08 1.75 1.76 0.07 1.73 1.73 0.03 VG-2 1.93 1.93 0.06 2.00 1.99 0.04 1.97 1.96 0.09 1.95 1.94 0.03 P2?5 AP 0.19 0.23 0.00 0.18 0.22 0.02 AP 0.18 0.22 0.03 0.18 0.22 0.03 MnO 0.19 0.21 0.02 0.20 0.22 0.01 KH 0.19 0.19 0.02 0.20 0.20 0.02 Fay 0.21 0.24 0.02 0.22 0.25 0.02 The data in Tables 5-7 contain a wealth of in- formation for statistical evaluation that is beyond the scope of this paper. One point, however, should be emphasized: there is a general apprehension on the part of the staff responsible for the microprobe analyses regarding use of the observed standard de- viation of a single day's analyses, based on one standardization, as a measure of precision. Since most microprobe users complete the analysis of a particular mineral within one day, this is often the 64 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Table 6.?Continued Run No. 1 2 3 4 1 2 3 4 Conditions Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Standards Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. a (7) Uncorrected B-A Corr. o (7) Uncorrected B-A Corr. o (7) SiO2 KH 50.61 50.02 0.58 50.39 49.84 0.73 50.42 49.90 0.20 50.83 50.26 0.31 VG-2 49.37 49.63 0.46 50.22 50.49 0.37 50.08 50.37 0.56 49.93 50.22 0.24 A12O3 KH 17.37 16.69 0.19 17.25 16.62 0.19 17.10 16.51 0.20 17.36 16.71 0.13 VG-2 16.09 16.06 0.13 16.05 16.04 0.20 16.30 16.25 0.21 16.33 16.31 0.12 FeO KH 9.06 9.11 0.11 8.97 9.02 0.06 8.89 8.94 0.23 8.97 9.02 0.06 VG-2 8.85 8.89 0.12 8.89 8.93 0.26 8.91 8.96 0.20 9.01 9.05 0.05 MgO KH 8.34 8.22 0.21 8.95 8.80 0.11 8.79 8.68 0.30 8.77 8.63 0.24 VG-2 8.36 8.20 0.12 8.99 8.81 0.44 8.45 8.28 0.11 8.73 8.56 0.25 CaO VG-D08 KH 12.46 12.64 0.15 12.45 12.63 0.13 12.21 12.39 0.18 12.45 12.64 0.04 VG-2 12.25 12.33 0.05 12.24 12.33 0.45 12.39 12.46 0.17 12.44 12.51 0.16 Na2O KH T.3"4 2.28 0.06 2.31 2.24 0.03 2.32 2,. 25 0.09 2.44 2.37 0.06 VG-2 T.38 2.31 0.03 2.37 2.30 _0.16j 2.30 2.23 0.04 2.33 2.26 0.01 KH 0.07 0.07 0.02 0.09 0.09 0.02 0.08 0.08 0.07 0.07 0.07 0.02 VG-2 0.09 0.09 0.02 0.09 0.09 0.01 0.09 0.09 0.02 0.10 0.09 0.02 TiO2 KH 0.96 0.97 0.05 1.01 1.02 0.05 1.05 1.06 0.07 1.00 1.01 0.03 VG-2 1.13 1.14 0.06 1.11 1.12 0.02 1.14 1.15 0.05 1.12 1.13 0.06 P2?5 AP 0.11 0.14 0.02 0.12 0.15 0.02 AP 0.11 0.14 0.03 0.12 0.15 0.01 MnO KH 0.13 0.14 0.02 0.15 0.15 0.03 Fay 0.16 0.18 0.01 0.14 0.16 0.01 2 results in brackets, obtained using a 2ym beam, are not included in compilations of Tables 3 and 4. See explanation in text. only measure of precision that can be observed. It is clear, however, that the average values of analyses made on two separate days, with separate standard- izations, may occasionally be different. For example, in the analysis of Kakanui hornblende, one of us (A. P.) performed a Student's t-test for runs 1 and 2 (Table 5). Run 1 gives a mean for MgO of 12.57 with a standard deviation of 0.13 and run 2 gives a mean of 13.09 with a standard deviation of 0.22. The Student's Mest indicates that the difference be- tween the two means (i.e., the error) is significant at the 99% confidence level. Of course, this is to be expected occasionally, given the limited stability of the instruments and the vagaries of data acquisition in general. For microprobe users, however, it is diffi- cult to realize, and frequently even more difficult to accept, the fact that their data might not be as good as it appears. Concern about the precision and accuracy of microprobe analyses is ever-present with the discriminating worker. One way to monitor the NUMBER 22 65 TABLE 7.?USGS laboratory individual analyses of Kakanui hornblende and 4 natural glasses using Kakanui hornblende as reference sample and also the laboratory's preferred reference samples Run No. 1 2 3 4a 4b 1 2a 2b 3 Conditions Standards Uncorrected B-A Corr. a (10) Uncorrected B-A Corr. a (10) Uncorrected B-A Corr. a (10) Standards Uncorrected B-A Corr. a (5) Uncorrected B-A Corr. a (10) Standards Uncorrected B-A Corr. a (10) Uncorrected B-A Corr. o (4) Uncorrected B-A Corr. a (10) Uncorrected B-A Corr. a (5) Standards Uncorrected B-A Corr. a (10) SiO2 KH 40.43 40.42 0.28 40.35 40.39 0.56 40.14 40.15 0.64 Di85 38.61 40.73 0.24 37.94 40.11 0.24 KH 50.63 49.58 0.68 51.38 50.25 0.77 51.00 49.91 0.48 51.18 50.05 0.28 Di85 49.19 50.75 0.42 Al2 KH 14. 14. 0. 15. 15. 0. 14. 14. 0. ?3 93 93 40 09 10 26 91 93 45 Ortho 12. 14. 0. 13. 15. 0. 71 48 43 21 04 42 KH 14 14 0 14 14 0 14 14 0 15 14 0 .87 .46 .21 .71 .30 .43 .75 .34 .40 .07 .61 .30 Grtho 12 13 0 .65 .98 .19 FeO KH 10. 10. 0. 10. 10. 0. 11. 11. 0. 50 5114 86 86 03 06 05 19 Garnet 10. 10. 0. 10. 10. 0. 53 50 44 59 57 19 KH 11 11 0 11 11 0 11 11 0 11 11 0 .94 .95 .26 .82 .83 .30 .77 .77 .19 .60 .61 .13 Garnet 11 11 0 .82 .79.14 MgO Kakanui KH 13. 12. 0. 12. 12. 0. 12. 12. 0. 02 99 28 89 89 35 84 86 17 Di2Ti 12. 13. 0. 12. 12. 0. 34 13 58 15 93 17 KH 6 6 0 6 6 0 6 6 0 6 6 0 .93 .98 .10 .94 .98 .10 .88 .91 .18 .71 .73 .03 Di2Ti 6 7 0 .55 .02 .05 CaO Na2O hornblende KH KH 10.66 10.66 0.49 9.71 9.72 0.63 10.34 10.34 0.08 Di2Ti 10.57 10.20 0.32 10.40 10.04 0.11 VG-2 KH 10.90 10.91 0.20 10.97 10.98 0.28 10.94 10.95 0.21 11.22 11.24 0.08 Di2Ti 11.11 10.72 0.06 2.68 2.68 0.13 2.76 2.76 0.12 2.58 2.58 0.07 Di85 2.47 2.68 0.12 2.54 2.76 0.05 KH 2.77 2.79 0.11 2.75 ..77 0.09 2.85 2.86 0.15 2.75 2.75 0.05 Di85 2.55 2.75 0.12 K20 KH 1.89 1.89 0.12 1.98 1.98 0.02 2.02 2.02 0.05 Ortho 1.95 1.91 0.04 2.01 1.97 0.02 KH 0.02 0.02 0.06 0.05 0.05 0.07 0.05 0.05 0.05 0.10 0.10 0?06 Ortho 0.19 0.18 0.06 TiO2 KH 4.31 4.32 0.18 4.30 4.30 0.08 4.20 4.20 0.13 Di2Ti 4.90 4.74 0.15 4.70 4.55 0.10 KH 1.63 1.63 0.03 1.65 1.66 0.08 1.66 1.66 0.07 1.63 1.63 0.03 Di2Ti 1.92 1.86 0.08 P2?5 AP 0.09 0.11 0.02 0.09 0.11 0.02 0.06 0.07 0.05 Ortho 0.02 0.02 0.07 0.05 0.05 0.05 AP 0.22 0.27 0.05 0.17 0.22 0.05 0.22 0.28 0.05 0.18 0.21 0.07 Ortho 0.20 0.19 0.05 MnO KH 0.08 0.09 0.12 0.11 0.11 0.10 0.37 0.37 0.41 Rhod 0.07 0.07 0.04 0.04 0.04 0.01 KH 0.23 0.23 0.06 0.36 0.36 0.03 0.32 0.32 0.05 0.08 0.08 0.10 Rhod 0.21 0.23 0.05 66 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES Table 7.?Continued Run No. 1 2 3 ?-I 3 Conditions Standards Uncorrected B-A Corr. ? (10) Uncorrected B-A Corr. o (6) Standards Uncorrected B-A Corr. ? (10) Standards Uncorrected B-A Corr. a (10) Standards Uncorrected B-A Corr. ? (10) Standards Uncorrected B-A Corr. ? (10) Uncorrected B-A Corr. 0 (5) Standards Uncorrected B-A Corr. ? (10) SiO2 KH 52.39 50.80 0.51 52.69 50.69 0.28 Di85 49.58 50.80 0.60 KH 51.88 50.62 0.32 Di85 49.37 50.70 0.09 KH 50.30 49.47 0.73 50.41 49.60 0.66 Di85 47.97 49.77 0.24 A12O3 KH 13.25 12.84 0.30 13.00 12.62 0.08 Ortho 11.61 12.80 0.26 KH 14.41 13.98 0.40 Ortho 14.24 13.54 0.04 KH 16.38 15.86 0.47 17.04 16.46 0.34 Ortho 14.02 15.44 0.17 FeO KH 13.39 13.36 0.15 13.53 13.50 0.27 Garnet 13.48 13.41 0.23 KH 13.66 13.63 0.36 Garnet 12.75 13.30 0.21 KH 8.75 8.79 0.22 9.01 9.05 0.19 Garnet 8.76 8.78 0.18 MgO KH 4.89 .4.99 0.10 5.08 5.18 0.08 Di2Ti 4.76 5.16 0.10 KH 5.57 5.65 0.08 Di2Ti 5.67 6.13 0.10 KH 8.86 8.71 0.18 8.54 8.40 0.12 Di2Ti 8.36 8.78 0.07 CaO VG-A99 KH 8.44 8.42 0.48 9.16 9.13 0.18 Di2Ti 9.33 8.97 0.17 VG-999 KH 10.26 10.25 0.14 Di2Ti 10.57 10.18 0.03 VG-D08 KH 12.03 12.10 0.66 12.22 12.29 0.15 Di2Ti 12.42 12.05 0.03 Na2O KH 2.80 2.88 0.20 2.62 2.69 0.07 Di85 2.44 2.73 0.08 KH 2.66 2.72 0.11 Di85 2.30 2.57 0.03 KH 2.44 2.39 0.04 2.36 2.30 0.08 Di85 2.23 2.39 0.07 K20 KH 0.77 0.78 0.11 0.67 0.67 0.08 Ortho 0.77 0.76 0.07 KH 0.03 0.03 0.04 Ortho 0.04 0.04 0.00 KH 0.00 0.00 0.00 0.00 0.00 0.11 Ortho 0.09 0.09 0.06 TiO2 KH 3.81 3.79 0.10 3.73 3.71 0.10 Di2Ti 3.91 3.77 0.12 KH 1.66 1.66 0.08 Di2Ti 1.87 1.80 0.17 KH 0.88 0.89 0.03 0.96 0.97 0.07 Di2Ti 1.05 1.03 0.05 p2o5 AP 0.40 0.49 0.04 0.44 0.54 0.02 Ortho 0.32 0.31 0.04 AP 0.17 0.21 0.01 Ortho 0.15 0.15 0.11 AP Q.13 0.16 0.01 0.10 0.12 0.02 Ortho 0.08 0.08 0.02 MnO KH 0.27 0.27 0.05 0.16 0.16 0.04 Rhod 0.18 0.19 0.06 KH 0.13 0.13 0.30 Rhod 0.20 0.20 0.04 KH 0.18 0.19 0.01 0.14 0.15 0.21 Rhod 0.12 0.13 0.01 NUMBER 22 67 precision and accuracy is to present data of samples with known composition together with the micro- probe analyses of the unknown samples. This sug- gestion has been made frequently and is followed by many, but it would serve the scientific commu- nity much better if practiced even more extensively. Literature Cited Albee, A. L., and Lily Ray 1970. Correction Factors for Electron Probe Microanalysis of Silicates, Oxides, Carbonates, Phosphates, and Sulfates. Analytical Chemistry, 42(12): 1408-1414. Byerly, G. R., W. G. Melson, and P. R. Vogt 1976. Rhyodacites, Andesites, Ferro-Basalts, and Ocean Tholeiites from the Galapagos Spreading Center. Earth and Planetary Science Letters, 30:215-221. Melson, W. G., G. R. Byerly, J. A. Nelen, T O'Hearn, T. L. Wright, and T. Vallier 1977. A Catalog of the Major Element Chemistry of Abyssal Volcanic Glasses. Smithsonian Contributions to the Earth Sciences, 19:31-60. Wright, T. L., and R. T. Okamura 1977. Cooling and Crystallization of Tholeiitic Basalt, 1965 Makaopuhi Lava Lake, Hawaii. United States Geological Survey Professional Paper, 1004:1-78. Electron Microprobe Reference Samples for Mineral Analyses Eugene Jarosewich, Joseph A. Nelen, and Julie A. Norberg ABSTRACT A table is presented containing compositional datafor 25 minerals, four natural glasses, and one syn- thetic glass prepared and analyzed for use as micro- probe reference samples at the Smithsonian Institu-tion. The table includes new chemical analyses of minerals and some updated analyses of mineralspublished previously. Detailed descriptions of sample preparation andevaluation of homogeneity are given. Introduction Microprobe analyses are an essential part of present-day mineralogical and petrological studies. It can be said that the application of the micro- probe to mineral studies and material sciences in general is one of the most significant advances since the first use of the petrographic microscope in the middle of the last century. The technique is now well established, widely used, and capable of high-quality analyses. As with all comparative in- strumental techniques, however, it requires well- characterized reference samples. Prime prerequisites for microprobe reference samples are homogeneity at the micrometer level and availability in rea- sonable quantities for standard chemical analyses. Either prerequisite is usually easily satisfied by itself but together are difficult to achieve. One of the problems with some minerals used as microprobe reference samples is a lack of proper documentation. Even if well-described minerals are from the same locality and/or are obtained from a reliable source, they may vary in chemical com- Eugene Jarosewich, Joseph A. Nelen, and Julie A. Norberg, Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. position. Therefore, a mineral sample intended as a reference sample should be carefully selected and used only when analytical data on this particular specimen are available. Since natural materials ful- filling all the above requirements are not always available, synthetic minerals and glasses have occa- sionally been prepared as substitutes. Again, homo- geneity of these materials should be checked and chemical analyses performed. The assumption that the precalculated composition is correct is certainly not always valid. In general, the most reliable microprobe analyses are obtained when a reference sample of composi- tion and structure close to that of the unknown is used because the matrix and possible wavelength shift effects are minimized, and only small correc- tions are needed. It is generally accepted that, regardless of the type of correction used, results corrected by more than 10 percent should be viewed with caution. Difficulties with correction procedures in the Si-Al-Mg system have been pointed out by Bence and Holzwarth (1977). Similar discrepancies have been observed by other probe users. All minerals and glasses described here, except one, are of natural origin. Most have been obtained from the Smithsonian collections and were selected either in conjunction with specific projects or for use in silicate analyses in general. ACKNOWLEDGMENTS.?We wish to acknowledge those curators and others listed in Table 1 who provided us with samples for use as microprobe reference samples. Brian Mason's careful examina- tion and assistance in separation of minerals is also greatly appreciated. Preparation of Reference Samples When a sufficient quantity (at least 2 g) of a min- eral or glass is available for use as a microprobe 68 NUMBER 22 69 reference sample, a thin section is prepared for microscopic examination. Next, a microprobe analy- sis for composition and homogeneity is performed. If preliminary results are favorable, the material is gently crushed, sized usually between 20 and 80 mesh, and further purified using either a heavy liquid separation or a Franz magnetic separator or both. In some instances cleaning with a suitable acid is also useful. As a final step, the material is examined under a low-powered microscope and most remaining foreign grains are removed by hand. The purified grains are again checked by microprobe for homogeneity (sigma ratios) within and among grains (Table 2). Finally, a chemi- cal analysis using classical methods (Peck, 1964; Hillebrand et al., 1953) is performed on the same separate that is to be used as the reference sample. Discussion In Table 1 are presented the data for newly ana- lyzed minerals, earlier published analyses, and up- dated analyses for several minerals that have been in use for some time. Johnstown meteorite hypers- thene and Springwater meteorite olivine have been re-analyzed using what we believe to be much cleaner separates. Kakanui hornblende has been re-analyzed for TiO2. Even after the most careful preparation of the reference sample, a grain of accessory mineral or matrix may remain in the sample, which, in the course of preparation of the standard discs, could be included with the reference sample. Occasional grains of the standard itself will be "off composi- tion," due to inhomogeneity. These problems can never be totally eliminated. The user should be aware of the possible presence of such "impurities" and make a thorough check for them. For example, occasional grains are found that are lower in sodium and higher in potassium than usual in the reference sample microcline, lower in manganese than usual in Rockport fayalite, and lower in sodium than usual in Lake County plagioclase. Infrequent inclu- sions in the glasses 72854, 111240/52, 113498/1, and 113716 are also found. The overall homogeneity of each sample was de- termined using the criteria given by Boyd et al. (1967) whereby the sample is considered to be homo- geneous if the sigma ratio (homogeneity index) of observed standard deviation (sigma) to the standard deviation predicted from counting statistics alone does not exceed 3. The sigma ratios were calculated with reference to ten ten-second counts on each of ten randomly selected grains. Table 2 gives sigma ratios for the ten grains of each reference sample for major and some minor elements. The values in parentheses indicate the worst sigma ratio observed for an element in a single grain. This does not, however, imply a single worst grain, as different grains may exhibit differing degrees of homogeneity for each element present. When the criterion of sigma ratios is used as a measure of homogeneity, all the reference samples prove to be very homo- geneous provided a reasonably large number of counts are taken on a reasonably large number of grains. In practice, however, fewer counts and grains are normally used for standardization, and under these circumstances a grain having a slightly dif- ferent composition may influence the microprobe results adversely. For this reason, grains showing some discrepancy in composition should be avoided. The percentages of these "impurities" in the whole samples are minimal and the effects on the bulk analyses of the samples are negligible. These samples were prepared in only small quan- tities, but they can be judiciously made available to microprobe users interested in the analysis of geo- logic materials. Potential users should remember that the purified samples differ in bulk chemistry from the specimens from which they were separated and should be very specific in their requests? i.e., the requests should be made for microprobe standard USNM no. n rather than simply material from specimen USNM no. n. Literature Cited Bence, A. E., and W. Holzwarth 1977. Non-Linearities of Electron Microprobe Matrix Cor- rections in the System MgO-Al2O3-SiO2. Proceedings of the Eighth International Conference on X-Ray Optics and Microanalysis and Twelfth Annual Con- ference of the Microbeam Analysis Society, page 38. Boyd, F. R., L. W. Finger, and F. Chayes 1967. Computer Reduction of Electron-Probe Data. Car- negie Institution Year Books, 67:210-215. Hillebrand, W. G., G.E.F. Lundell, H. A. Bright, and J. I. Hoffman 1953. Applied Inorganic Analysis. 2nd edition, 1034 pages. New York: John Wiley and Sons. Peck, L. C. 1964. Systematic Analysis of Silicates. U.S. Geological Sur- vey Bulletin, 1170:66. 70 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES KEY TO TABLE 1 Analysts, Sources, Analysts: 1. 2. 3. 4. 5. 6. 7. 8. 9. E. Jarosewich, Dept. of Mineral Sciences, Smith- sonian Institution J. Nelen, Dept. of Mineral Sciences, Smithsonian Institution J. Norberg, Dept. of Mineral Sciences, Smith- sonian Institution E. L. Munson, N. M. Conklin, J. N. Rosholt, and I. C. Frost, U.S. Geological Survey B. Wiik, Geological Survey, Finland U.S. Geological Survey, Geochemistry and Pet- rology Branch D. Mills, X-Ray Assay Laboratories, Ontario, Canada; J. Nelen; J. Norberg E. Kiss, Dept. of Geophysics and Geochemistry, Australian National University J. J. Fahey and L. C. Peck, U.S. Geological Survey Sources: 1. 2. 3. 4. 5. 6. 7. P. Desautels, J. S. White, Jr., and P. J. Dunn, Dept. of Mineral Sciences, Smithsonian Institu- tion B. Mason, Dept. of Mineral Sciences, Smith- sonian Institution G. Switzer, Dept. of Mineral Sciences, Smith- sonian Institution W. G. Melson, Dept. of Mineral Sciences, Smith- sonian Institution T. L. Wright, U.S. Geological Survey H. Staudigel, Massachusetts Institute of Tech- nology R. S. Clarke, Jr., Dept. of Mineral Sciences, 8. References Smithsonian Institution J. H. Berg, Northern Illinois University References for previously published analyses: 1. 2. 3. 4. 5. Stewart, D. B., G. W. Walker, T. L. Wright, and J. J. Fahey 1966. Physical Properties of Calcic Labradorite from Lake County, Oregon. American Mineralogist, 51:177-197. Young, E. J., A. T. Myers, E. L. Munson, and N. M. Conklin 1969. Mineralogy and Geochemistry of Fluora- patite from Cerro de Mercado, Durango, Mexico. U.S. Geological Survey Profes- sional Paper, 650D:84-93. Mason, B., and R. O. Allen 1973. Minor and Trace Elements in Augite, Hornblende, and Pyrope Megacrysts from Kakanui, New Zealand. New Zea- land Journal of Geology and Geophysics, 16(4):935-947. Jarosewich, E. 1972. Chemical Analysis of Five Minerals for Microprobe Standards. In William G. Melson, editor, Mineral Sciences Investi- gations, 1969-1971. Smithsonian Contri- butions to the Earth Sciences, 9:83-84. Jarosewich, E. 1975. Chemical Analysis of Two Microprobe Standards. In George S. Switzer, editor, Mineral Sciences Investigations, 1972- 1973. Smithsonian Contributions to the Earth Sciences, 14:85-86. TABLE 1.?Chemical analyses of electron microprobe reference samples; analysts, sources, and references identified in "Key to Table 1" on facing page; these purified samples all differ, to greater or lesser degree, from the bulk chemistry of the USNM specimens from which they were separated (see text). Mineral Ahorthite, Great Sitkin Island, AL USNM 137041 Anorthoclase, Kakanui, New Zealand USNM 133868 Apatite (Fluorapatite), Durango, Mexico1 USNM 104021 Augite, Kakanui, New Zealand USNM 122142 Benitoite, San Benito County, CA2 USNM 86539 Chromite, Tiebaghi Mine, New Caledonia3 USNM 117075 Corundum, synthetic1* USNM 657S Diopside, Natural Bridge, NY USNM 117733 Fayalite, Rockport, MA USNM 85276 Garnet, Roberts Victor Mine, South Africa USNM 87375 Garnet, Roberts Victor Mine, South Africa USNM 110752 Glass, Basaltic, Juan de Fuca Ridge USNM 111240/52 VG-2 Glass, Basaltic, Makaopuhi Lava Lake, HI USNM 113498/1 VG-A99 Glass, Basaltic, Indian Ocean5 USNM 113716 Glass, Rhyolitic, Yellowstone Nat. Pk., WY6 USNM 72854 VG-568 Glass, Tektite, synthetic7 USNM 2213 Hornblende, Arenal Volcano, Costa Rica USNM 111356 Hornblende, Kakanui, New Zealand8 USNM 143965 Hypersthene, Johnstown meteorite USNM 746 Ilmenite, Ilmen Mtns., Miask, USSR9 USNM 96189 Magnetite, Minas Gerais, Brazil10 USNM 114887 Microcline, location unknown USNM 143966 Olivine (F090). San Carlos, Gila Co., AZ11 USNM 111312/444 Olivine (F083), Springwater meteorite USNM 2566 Omphacite, Roberts Victor Mine, So. Africa USNM 110607 Osumilite, Nain, Labrador USNM 143967 Plagioclase (Labradorite), Lake County, OR USNM 115900 Pyrope, Kakanui, New Zealand USNM 143968 Quartz, Hot Springs, AR12 USNM R177O1 Scapolite (Meionite), Brazil13 USNM R6600-1 SiO2 44.00 66.44 0.34 50.73 43.75 54.87 29.22 39.47 40.16 50.81 50.94 51.52 76.71 75.75 41.46 40.37 54.09 64.24 40.81 38.95 55.42 60.20 51.25 41.46 99.99 49.78 AI2O3 36.03 20.12 0.07 7.86 9.92 99.99 0.11 22.27 22.70 14.06 12.49 15.39 12.06 11.34 15.47 14.90 1.23 18.30 8.89 22.60 30.91 23.73 25.05 Fe2O3 0.06 3.69 1.32 2.77 2.17 2.23 1.87 1.12 0.48 0.64 5.60 3.30 11.6 67.5 0.00 1.35 0.34 FeO 0.62 0.20 0.00 3.45 13.04 0.24 66.36 13.76 9.36 9.83 11.62 8.12 0.80 4.32 6.43 7.95 15.22 36.1 30.2 0.04 9.55 16.62 3.41 6.38 0.15 10.68 0.17 MgO <0.02 0.01 16.65 15.20 18.30 6.55 7.17 6.71 5.08 8.21 <0.1 1.51 14.24 12.80 26.79 0.31 0.03 49.42 43.58 11.57 5.83 0.14 18.51 CaO 19.09 0.87 54.02 15.82 0.12 25.63 14.39 18.12 11.12 9.30 11.31 0.50 2.66 11.55 10.30 1.52 0.02 <0.05 13.75 <0.03 13.64 5.17 13.58 Na2O 0.53 9.31 0.23 1.27 0.34 2.62 2.66 2.48 3.75 1.06 1.91 2.60 <0.05 1.30 5.00 0.39 3.45 5.20 K20 0.03 2.35 0.01 0.00 0.19 0.82 0.09 4.89 1.88 0.21 2.05 <0.05 15.14 0.15 4.00 0.18 0.94 TiO2 0.03 0.74 19.35 0.04 0.39 0.35 1.85 4.06 1.30 0.12 0.50 1.41 4.72 0.16 45.7 0.01 0.37 0.18 0.05 0.47 P2O5 40.78 0.20 0.38 0.12 <0.01 0.00 <0.01 0.00 0.00 MnO Cr2O3 O.Ox 0.13 0.11 60.5 0.04 2.14 0.59 0.19 0.22 0.15 0.17 0.03 0.11 0.15 0.09 0.49 0.75 4.77 0.04 0.14 0.30 0.02 0.10 0.01 0.28 H20 <0.05 0.01 0.04 0.1 <0.01 <0.01 0.02 0.02 0.18 0.12 0.10 1.21 0.94 0.00 <0.05 0.02 0.02 0.05 <0.01 0.21 Total 100.33 99.29 99.94 100.38 100.15 98.89 99.99 99.53 99.18 100.19 100.22 99.86 99.39 100.07 99.56 99.88 99.64 100.02 100.25 99.40 98.16 99.12 100.29 99.47 100.03 99.60 100.17 100.30 99.99 99.86 Analyst 1 3 4 5 2 2 1 2 2 1 1 1 3 3 3 6 1 1 3 7 3 8 1 3 1 1 9 1 1 2 Source 1 2 1 2 1 1 1 1 1 3 3 4 5 6 4 7 4 2 7 1 1 3 2 7 3 8 1 2 1 1 Reference j 2 3 4 4 5 5 4 4 1 4 1 SrO 0.07; RE2O3 1.43; ThO2 0.02; As2O3 0.09; V2Os 0.01; C02 0.05; S03 0.37; F 3.53; Cl 0.41; sub-total: 101.52; 0 equivalent to Cl, F = 1.58; final total: 99.94. 2 BaO 37.05. 3 Total Fe reported as FeO. "? Emission spectrometric analysis: Si 0.03; Fe 0.003; Mg 0.007; Ca 0.003; Na 0.005; K 0.005. 5 S 0.12; sub-total: 100.13; 0 equivalent to S = 0.06; final total: 100.07. 6 Cl 0.13; sub-total: 99.59; 0 equivalent to Cl ? 0.03; final total: 99.56. 7 C02 not determined (insufficient sample); Cl 0.00; F 0.01. Synthetic glass prepared by Corning Glass Company. 8 New TiO2 value: 4.72. 9 Nb2O5 0.92. 10 Preliminary values: MgO 0.05; TiO2 0.16; MnO <0.01; Cr2O3 0.25. 11 NiO 0.37. 12 Emission spectrometric analysis: Al 0.0005; Fe 0.01; Mg 0.005; Ca 0.001; Na 0.001; K 0.0003. 13 C02 2.5; S03 1.32; Cl 1.43; sub-total: 100.18; 0 equivalent to Cl = 0.32; final total: 99.86. 72 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES TABLE 2.?Sigma ratios (homogeneity indices) for all analyzed grains of each reference sample observed sigma for n grains (sigma ratio for n grains = ?; least homogeneous grain in sigma predicted from counting statistics parentheses; dashes = not evaluated) Mineral Anorthite Anorthoclase Apatite (Fluorapatite) Augite Dnni f-r\1 t-CkDcIilLUJ-Lc Chromite Diopside Fayalite Garnet, 87375 Garnet, 110752 Glass, 111240/52 VG-2 Glass, 113498/1 VG-A99 Glass, 113716 Glass, 72854 VG-568 Glass, 2213 Hornblende, Arenal Ho rnblende, Kakanui Hypersthene Ilmenite Magnetite Microcline Olivine (Fogo), San Carlos Olivine (F083), Springwater Omphacite Osumilite Plagioclase (Labradorite) Pyrope Quartz Scapolite (Meionite) SiO2 0.96 (1.51) 1.09 (1.60) 0.99 (1.37) 1.07 (1.37) 0.95 (1.46) 0.89 (1.26) 0.88 (1.32) 0.94 (1.10) 0.94 (1.32) 1.12 (1.42) 0.97 (1.61) 1.05 (1.72) 1.07 (1.67) 1.01 (1.38) 1.07 (1.55) 0.94 (1.13) 0.81 (1.13) 0.96 (1.42) 0.89 (1.23) 0.96 (1.90) 1.09 (1.49) 1.08 (1.46) 0.99 (1.29) A12O3 0.81 (1.26) 0.79 (1.38) 0.97 (1.66) 1.00 (1.47) 1.01 (1.42) 0.94 (1.28) 0.89 (1.11) 1.10 (1.46) 1.00 (1.30) 1.00 (1.47) 0.87 (1.24) 0.97 (1.66) 1.00 (1.24) 1.04 (1.52) 0.95 (1.64) 1.27 (1.89) 0.95 (1.40) 0.95 (1.20) 0.95 (1.41) FeO 0.84 (1.26) 1.01 (1.66) 1.14 (2.32) 1.06 (1.41) 0.90 (1.20) 0.86 (1.13) 1.07 (1.38) 0.94 (1.34) 1.01 (1.34) 1.12 (1.36) 1.30 (1.67) 1.10 (1.37) 1.72 (3.60) 0.84 (1.16) 0.90 (1.29) 1.06 (1.51) 0.96 (1.87) 1.20 (2.19) 1.09 (1.59) MgO 0.94 (1.23) 1.11 (1.50) 0.97 (1.50) 1.01 (1.49) 1.00 (1.34) 0.96 (1.61) 0.92 (1.38) 1.01 (1.36) 1.11 (1.67) 1.16 (2.38) 0.93 (1.27) 1.00 (1.64) 0.99 (1.12) 0.91 (1.30) 1.00 (1.70) 0.98 (1.21) CaO 0.92 (1.23) 1.02 (1.51) 1.00 (1.25) 0.95 (1.50) 0.86 (1.11) 0.87 (1.47) 1.00 (1.27) 0.93 (1.34) 0.83 (1.19) 1.05 (1.61) 1.01 (1.27) 1.10 (1.73) 1.02 (1.51) 1.04 (1.65) 0.97 (1.18) 0.91 (1.16) Na2O 1.11 (1.57) 1.05 (1.31) 1.15 (2.10) 1.25 (2.59) 2.31 (3.45) 0.98 (1.32) 1.15 (2.15) 0.99 (1.31) 0.91 (1.33) 0.96 (1.41) K20 0 (1 0 (1 1 (1 1 (1 .98 .36) .90 .29) .09 .59) .13 .64) TiO2 0. (1. 1. (1. 1. (1. 97 44) 01 49) 34 98) P2O5 MnO Cr2O3 0.97 (1.51) 1 (1 1 (1 1.12 (1.49) .03 .58) .21 .53) observed sigma for all grains Sigma ratio for 10 grains = sigma predicted from counting statistics observed sigma for this particular grain Sigma ratio for least homogeneous gram = sigma predicted from counting statistics (in parentheses) Smithsonian Contributions to the Earth Sciences 1. George Switzer and William G. Melson. "Partially Melted Kyanite Eclogite from the Roberts Victor Mine, South Africa." 9 pages, 5 figures, 6 tables. 15 April 1969. 2. Paul A. Mohr. "Catalog of Chemical Analyses of Rocks from the Interaction of the African, Gulf of Aden, and Red Sea Rift Systems." 1322 entries. 16 December 1970. 3. Brian Mason and A. L. Graham. "Minor and Trace Elements in Meteoritic Minerals." 17 pages, 1 figure, 17 tables. 17 September 1970. 4. William G. Melson, Eugene Jarosewich, and Charles A. Lundquist. "Volcanic Eruption at Metis Shoal, Tonga, 1967-1968: Description and Petrology." 18 pages, 13 figures, 3 tables. 16 October 1970. 5. Roy S. Clarke, Jr., Eugene Jarosewich, Brian Mason, Joseph Nelen, Manuel Gomez, and Jack R. Hyde. "The Allende, Mexico, Meteorite Shower." 53 pages, 36 figures, 6 tables. 17 February 1971. 6. Daniel J. Stanley and Noel P. James. "Distribution of Echinarachnius parma (Lamarck) and Associated Fauna on Sable Island Bank, Southeast Canada." 24 pages, 8 figures, 6 plates, 1 table. 27 April 1971. 7. William G. Melson. "Geology of the Lincoln Area, Lewis and Clark County, Montana.'' 29 pages, 13 figures, 8 tables. 15 October 1971. 8. Daniel J. Stanley, Donald J. P. Swift, Norman Silverberg, Noel P. James, and Robert G. Sutton. "Late Quaternary Progradation and Sand Spillover on the Outer Continental Margin off Nova Scotia, Southeast Canada." 88 pages, 83 figures, 6 tables. 11 April 1972. 9. William G. Melson, editor. "Mineral Sciences Investigations, 1969-1971." 94 pages, 34 figures. 16 August 1972. 10. Louis H. Fuchs, Edward Olsen, and Kenneth J. Jensen. "Mineralogy, Mineral-Chemistry, and Composition of the Murchison (C2) Meteorite." 39 pages, 19 figures, 9 tables. 14 August 1973. 11. Daniel J. Stanley and Peter Fenner. "Underwater Television Survey of the Atlantic Outer Continental Margin near Wilmington Canyon." 54 pages, 18 figures, 2 August 1973. 12. Grant Heiken. "An Atlas of Volcanic Ash." 101 pages, 15 figures, 33 plates, 3 tables. 12 April 1974. 13. Nicolas A. Rupke and Daniel J. Stanley. "Distinctive Properties of Turbiditic and Hemi- pelagic Mud Layers in the Algero-Balearic Basin, Western Mediterranean Sea." 40 pages, 21 figures, 8 tables. 10 September 1974. 14. George S. Switzer, editor. "Mineral Sciences Investigations: 1972-1973." 88 pages, 29 figures. 2 July 1975. 15. Daniel J. Stanley, Gilbert Kelling, Juan-Antonio Vera, and Harrison Sheng. "Sands in the Alboran Sea: A Model of Input in a Deep Marine Basin." 51 pages, 23 figures, 8 tables. 16 June 1975. 16. Andres Maldonado and Daniel Jean Stanley. "Late Quaternary Sedimentation and Stra- igraphy in the Strait of Sicily." 73 pages, 39 figures, 5 tables. 3 August 1976. 17. R. O. Chalmers, E. P. Henderson, and Brian Mason. "Occurrence, Distribution, and Age of Australian Tektites." 46 pages, 17 figures, 10 tables. 9 September 1976. 18. Arthur Roe and John S. White, Jr. "A Catalog of the Type Specimens in the Mineral Col- lection, National Museum of Natural History." 43 pages. 22 November 1976. 19. Brian Mason, editor. "Mineral Sciences Investigations: 1974-1975." 125 pages, 48 figures, 37 tables. 9 March 1977. 20. Daniel J. Stanley. Henri Got, Neil H. Kenyon, Andre Monaco, and Yehezkiel Weiler. "Cata- lonian, Eastern Betic, and Balearic Margins: Structural Types and Geologically Recent Foundering of the Western Mediterranean Basin." 67 pages, 33 figures. 20 September 1976. 21. Roy S. Clarke, Jr., and Joseph I. Goldstein. "Schreibersite Growth and Its Influence on the Metallography of Coarse Structure Iron Meteorites." 80 pages, 28 figures, 20 tables. 14 April 1978.