Mineral Sciences Investigations 1976-1977

SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES • NUMBER 22

Mineral Sciences Investigations1976-1977

Robert F. FudaliEDITOR

SMITHSONIAN INSTITUTION PRESS

City of Washington

1979

A B S T R A C T

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 therelationships within this class; a comparative chemical study of sixty Australiantektites from widely separated localities; a description of a new, rapid techniqueof sample preparation for whole-rock analyses using the electron microprobe; aninterlaboratory comparison of the precision and accuracy of electron microprobeanalyses; 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 UlawanVolcano, New Britain.

Library of Congress Cataloging in Publication DataMain 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 theSmithsonian's Department of Mineral Sciences."Bibliography: p.1. Mineralogy—Addresses, essays, lectures. I. Fudali, Robert F. II. National Museum of Natural

History. 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

PageMETEORITES

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 14T H E 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 CharlesR. Obermeyer 46

MICROPROBE ANALYSES OF FOUR NATURAL GLASSES AND ONE MINERAL:

AN INTERLABORATORY STUDY OF PRECISION AND ACCURACY, by EugeneJarosewich, 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

i n

Mineral Sciences Investigations1976-1977

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 andlunar 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 hereinbetween Lonar basalts and impact glasses suggestsome probable trends in lunar analogs. Conversely,the isochemical nature of much of the Lonar glassand basalt demonstrates that lunar glasses may oftenbe chemically equated to their parent rocks.

Introduction

As the importance of meteorite impacts as amechanism producing lunar land forms, rocks, and

Kurt Fredriksson, Phyllis Brenner, Joseph Nelen, Departmentof Mineral Sciences, National Museum of Natural History,Smithsonian Institution, Washington, D.C. 20560. AnandaDube, Geological Survey of India, Calcutta 13, India. DanielMilton, 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 ofLonar Crater, India, is coming to be seen as one ofearth's more fortuitous catastrophes. This impactcrater is located in the Deccan Trap basalts ofIndia (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, andchemistry of shocked Lonar and lunar materials.

ACKNOWLEDGMENTS.—We are indebted to WalterBrown, Andrea Eddy, Becky Fredriksson, and JulieNorberg for technical and/or editing assistance.Support from the Geological Survey of India, theSmithsonian Research Foundation and Foreign Cur-rency Program is gratefully acknowledged. One ofus (DM) received partial support under a NASAcontract. We also thank Laurel Wilkening andHans 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 mostof the crater floor. The rim of the crater is raisedapproximately 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 excavatedafter throwout from the main crater landed, butduring settling of fallout. Various data on generalmorphology, conclusive evidence for the impactorigin, and preliminary studies of petrography,stratigraphy, and the chemistry of the impact rockshave been published elsewhere (Fredriksson et al.,1973). Recent fission track dating of Lonar shock-melted glasses indicates an age less than ~ 50,000years (Wilkening, pers. comm.). Carbon-14 datingshows an age of more than 30,000 years for organicmaterial recovered 50 m down in one of the ~90 mthick sediment columns in the lake (Suess, pers.comm.). The crater's age then lies between 30 and50 thousand years.

The Geological Survey of India has drilled aseries of five boreholes approximately on a linetrending NE-SW in the floor of the crater. Eachdrillhole 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 materialwas probably eroded from the crater rim. Afterpenetrating the sediment, each of the drillings re-turned cores of coarser breccia apparently composedof blocks up to meters in size which are unshockedor slightly shocked and contain crude shatter cones.Beneath the coarse breccia, the first four drillingsencountered a layer that yielded essentially no corerecovery, apparently because it was composed ofunconsolidated to extremely friable microbreccia.The assumption was confirmed in the fifth hole forwhich an improved core catcher was used, resultingin approximately 100% core recovery. In addition,one deep (~100 m) and several shallower holes weredrilled in the smaller crater and on its rim. Currentevidence indicates that the structure, although arti-ficially modified by former dams and ditches, forpurposes of irrigation, is of impact origin.

Comparative Petrography, Morphology,and Chemistry

The Deccan Trap basalts at Lonar exhibit twomain textural varieties: (1) "hard" trap which haseuhedral plagioclase phenocrysts with strong oscil-latory zoning, often glomero-porphyritic, in anaphanitic 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 (ilmeniteand/or magnetite) 2%-12%. Olivine may be pres-ent in small amounts. Alteration minerals includechlorite 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 fromcoarse to fine to vitrophyric, and are often vesicular.In coarse lunar basalts pyroxene may occur as poiki-litic grains in plagioclase. Plagioclase constitutes25%-3O%, clinopyroxene 45%-55%, and opaques15%-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 Lonarand lunar impact-produced rocks and glasses dis-cussed below may be explained by the chemicaldifferences between the target basalts (see Table 1).

Shockwaves generated by a hypervelocity impactcause a variety of deformations and changes in thetargets. Thus, slight shocking of Lonar basalt(<250 kb) produces minor cataclasis in plagioclaseand pyroxene, wavy extinction in pyroxene, andshock-induced twinning in plagioclase. Increasingpressures within this range produce closely spacedtwins in pyroxene (this twinning seems to be relatedto possible strong shear loci). Large plagioclase andpyroxene crystals also tend to recrystallize or granu-late into smaller domains, optically independent. Insome cases planar elements in plagioclase are alsoproduced, often accompanied by strong cataclasis,warping of the grain, and reduced birefringence.In "moderately" shocked basalt (~250-450 kb),plagioclase is converted to disordered maskelynitein which the original oscillatory zoning is preserved.Moderately strong shocking (~450-600 kb) producesmicrobreccia clasts in which maskelynite exhibitsincipient flow. With intense shock, (>800 kb), thebasalt may be converted to glass. The mechanismsinvolved and necessary pressures for these varioustransformations with special reference to Lonarbasalts have been studied in great detail by Kiefferet 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 electronmicroprobe 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 basal t 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 17 a

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

Lonarbasalt

50.13

13.69

14.09c

5.80

10.20

2.52

0.55

2.72

Lonar

LNRshockedbasalt

50.35

15.08

12.74C

5.99

11.55

1.70

0.25

2.31

unit ("throw-out") is crudely stratified and showsno evidence of shock. The overlying unit ("fallout")contains clasts from different bedrock units, Fladle-like bombs, glass spherules, and fragments. Thelithic clasts show varying degree of shock frombarely perceptible through medium to (rarely) in-tense. As a practical matter the ejecta can be as-signed to only two bedrock units representing thetwo basalt types described above: (1) hard, freshtrap exposed in the upper 50 m of the crater wall;and (2) soft, weathered trap exposed in the lower50 m which presumably continues downward almostto the lake level and is probably followed by an-other dense basalt. Ejecta patches dominantly ofone or the other unit are juxtaposed throughout theejecta blanket with, at best, a weak tendency to-ward inverted stratigraphy. The patches vary widelyin size, but are characteristically on the order of100 m across near the rim and decrease to severalmeters across near the outer edge.

The hard trap typically forms breccia of smallangular clasts; the soft trap is in large masses thathave much less internal disruption. Concentrationsof ejecta relatively abundant in glass and shocked

fragments were found up to ~2 crater radii fromthe rim expecially to the E and ESE.

The ejecta blanket does not feather out evenlyat the margin, but breaks into clumps increasinglyscattered in ejecta-free terrain. Some of these arecomposites of hard and soft trap or even polymictbreccias which, surprisingly, cohered during a flightof two or more kilometers. Some clumps remain in-tact; others are spread out inside secondary cratersthey have excavated. By analogy, secondary crater-ing on the moon depends neither on ejection oflarge individual blocks nor absence of atmosphere(Milton and Dube, 1977).

In Figures 1-6 the similarities between Lonarand lunar shocked basalts and breccias are illus-trated. Lithic clasts in the Lonar microbrecciasmay exhibit severe cataclasis without reduction ofbirefringence in the plagioclase, or may show littleor no cataclasis and still contain plagioclase com-pletely converted to maskelynite. Deformation ofplagioclase in lunar breccias is directly analogousto that found in Lonar samples including cataclasis,increased density of twin lamellae, granulation ordevelopment of a mozaic texture, development of


FIGURE 1.—a, Lonar trench sample LRT 18VII-1, a medium-grained basalt with plagioclasetotally converted to maskelynite; pyroxenes exhibit moderate cataclasis; section is ~0.9 mm inlength, b, Apollo 17, #79155,68, a coarse-grained basalt with plagioclase partially converted tomaskelynite (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 betweencoarse-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 crystalsand 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 Lonarcounterparts ( Figures 1, 2).

Lonar microbreccias from both cores and trenchesconsist of lithic and crystal fragments embedded ina 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 mayall exhibit similar shock characteristics or may ex-hibit a wide range of shock features (e.g., a slightlyshocked fragment juxtaposed to a moderatelyshocked one). Fragments may also be very similar intexture and grain size ("monomict" breccia) or ex-tremely dissimilar ("polymict" breccia) within onethin section. Occasionally, abraded breccia frag-

NUMBER 22

FIGURE 3.—a, Lonar core sample LNR-2-319 is a microbreccia with angular to subangular crystalfragments 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 ofsection 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 tomaskelynite; section is 3.4 mm in length, b, Apollo 16, #68822,1; note abraded lithic fragmentsin 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 consistsalmost entirely of plagioclase; similar lunar "an-orthosite" breccia clasts may, thus, not always rep-resent individual rock types. Microbreccia rock flourmatrix material often shows evidence of flow, and

may contain turbid shock glass (Figure 6). Crystalfragments in the matrix are usually angular to sub-angular and exhibit shock effects ranging from mildto moderately strong. The unconsolidated varietyof microbreccia appears to contain little or no rockflour matrix, a major component of consolidatedmicrobreccia, and instead, consists almost entirelyof small crystal fragments of a fairly uniform size.Shock effects in these fragments are generally slight,occasionally moderate. Lithic and breccia fragments


FIGURE 5.—a, Lonar core sample, LNR-2-319A, with abraded lithic fragments in fine-grainedrock flour; note high degree of crystal angularity in matrix contrasting with roundness of thelithic fragments; length of section is 3.4 mm. b, Apollo 16, #67943,1-J; note abraded lithicand 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 brecciationin situ.

The material found in core 5 between ^200 and300 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. Occasionallymore indurated (whether by shock or later dia-genetic processes is uncertain) sections of core (up

to tens of centimeters) are found. Thin sectionsshow these breccias to resemble simple lunar soilbreccias and to contain unshocked to moderatelyshocked minerals and rock fragments.

Lonar glass ejecta consist of bombs, spherules, andfragments. Black glassy bombs, evidently still softwhen 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, andsimilar in appearance to microtektites, consist ofdroplets 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; saapleis 0.6 mm long.

banding indicated by contorted schlieren. Featheryrecrystallization in the spherules occurs rarely.Though most Lonar spherules seem to be compo-sitionally homogeneous, several were found withcompositional zoning. The trends most often ob-served include progressive alkali depletion (partic-ularly Na2O) accompanied by FeO enrich nent. fromthe center to the edge of the spherules. Glass shardsand angular fragments mostly consist of turbid,

brown to colorless glass, often vesicular and contain-ing contorted schlieren. The production of shockmelts requires intense shocking, >800 kb (Schaaland 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 beenstudied by a number of authors (see especiallyCarter, 1971). Figures 7-10 show some characteristic


FIGURE 9.—a, A spherule from Lonar trench sample LRT 18IIIA-4 # 1 ; SEM photo shows smallmound of glass possibly extruded as spherule cooled; analysis in Table 3; spherule is 1 mm inlong dimension, b, Thin section of spherule shown in a; note apparent inhomogeneity of internalstructure; the glass, however, has almost constant (basaltic) composition; the skeletal crystallitesdecorating 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 shownin c reveals a partially devitrified texture.

types illustrating the similarities between Lonar andlunar glass particles. A few chemical analyses aregiven in Tables 2-4.

A large number of electron probe analyses weremade to determine trends in composition amongLonar basalts, "bombs," and the other glass par-ticles. The glasses were subdivided into two some-what arbitrary groups based on composition and/ormorphology: (1) "normal" glasses (Table 2), thecompositions of which were comparatively uniformand resemble the basalts; and (2) "selected" glasses

(Table 3) mostly with compositions deviating fromthe "normal" glasses. Observed trends were variousdegrees of alkali, water, and sometimes silica deple-tion in the spherules and some "unusual" glasses.The majority of analysed glasses, however, closelyapproximate a homogeneous basalt melt althoughthey appear somewhat depleted in FeO + MgO andenriched in A12O3 relative to the parent basalt(Figure 11, Tables 1, 2). In addition, ferrous ironis more prevalent in the glasses, indicating areduction 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 trenchsample LRT 18IIA-3 # 3 , a spherule similar to that shown in a; analysis in Table 3. c, Apollo17, #74220,86 (>60 mesh), a spherule; SEM photo; note microcrater; sample is ~0.43 mm indiameter, d, Thin section of spherule shown in c, showing almost complete devitrification; lengthof section is 0.25 mm.

(Fe2O3/(FeO + Fe2O3) is -0.42 in the basalts, -0.37in the glasses). According to Morgan (1978) theheavy volatile elements Se and Cd are also signifi-cantly depleted in the glasses.

Lunar trends, though inherently more difficult todiscern (e.g., initially low alkalies) seem also to indi-cate a depletion in alkalies in glasses relative to theparent basalts (Fredriksson et al., 1971). As at Lonarsome heavy volatile elements, such as Pb (Silver,1975), may also have been redistributed in the lunarrocks and soils by impact volatilization. If lunar

fractionation trends are indeed analogous to Lonartrends (even ignoring the "unusual" Lonar glasseswhich were likely produced by complete or partialmelting of individual phenocrysts) then lunar glassfragment chemistry may or may not represent un-altered lunar rock types.

Interpretation of impact cratering and ejecta stra-tigraphy has been attempted throughout the Apollomissions. 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, thislaboratory

Constituent

SiO2

AI2O3

FeOa

MgO

CaO

Na20

K20

TiO2

LRT19VB-1

52.02

13.84

13.58

5.43

10.27

2.30

0.63

2.69

LRT19VB-1

51.65

14.89

12.76

5.12

9.94

2.56

0.53

2.09

LRT19VB-1

52.61

14.36

12.91

4.95

10.40

1.99

0.47

2.27

D215-4

50.48

13.99

12.96

5.46

10.13

2.59

0.53

2.23

SL-31.5-4

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

FeO d

MgO

CaO

Na2O

K 2 0

TiO2

LRT19VIB-3

37.69

22.22

28.41

4.82

6.96

0.02

0.02

0.10

LRT18HA-3 a

49.1

12.8

16.5

7.24

9.0

1.76

0.28

2.34

LRT18111A-4 b

47.79

12.69

15.88

7.67

9.13

1.49

0.23

2.29

LRT19VB-1

51.81

5.27

18.30

11.64

12.13

0.62

0.29

1.95

LRT181VB-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 reworkingthe lunar regolith (e.g., Gault et al., 1974) maydeserve reconsideration. Secondary impacts playedan important role in mixing and reworking pre-existing Lonar soils with ejecta of both basalt frag-ments and loosely consolidated microbreccias. Thisshould be considered in the interpretation of theevolution of the lunar regolith. Secondary crater-ing may also explain the discrepancy between the

long exposure ("suntan") ages of lunar soil relativeto excavated rock fragments (Gold and Williams,1974). Detailed descriptions and maps of the Lonarejecta distribution are being prepared by the Geo-logical Survey of India in cooperation with some ofus, and will be published elsewhere.

Some findings near the bottom of the fifth andlast drillcore may justify also a reevaluation ofcurrent 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 withlithium tetraborate flux (A, B, C, and D designate individual glass analyses that appear to berepresentative of distinctive groupings; it appears that the B glasses generally correspond to thelocal bulk soils, but the other groupings are not readily identifiable)

Consti tuent

SiO2

A12O3

FeOMgOCaONa20K2OTiO2

Constituent

SiO2

AI2O3FeOMgOCaONa20K20TiO2

Consti tuent

SiO2

AI9O3FeOMgOCaONa20K20TiO2

Bulk Soi l

41.412.816.3

9.910.60.280.157.2

Bulk Soil

41.213.015.7

9.910.70.390.106.4

Bulk Soil

42.015.114.310.011.10.440.204.7

700019,72

A

38.16 . 1

24.414.2

7.20.300.228.7

A

39.46.2

22.714.6

7.60.320.099 .0

A

38.65 . 8

22.414.7

7.40.420.109 . 1

B

40.112.716.810.110.80.340.117.9

79035,29

B

41.913.815.810.011.30.330.117.4

79135,71

B

42.013.614.310.510.80.400.066.0

C

41.422.0

6.414.713.10.110.031.20

C

45.221.5

8 . 111.713.30.040.011.16

74220,86

A

39.96.0

22.314.8

7 . 10.330.108.8

D

44.036.0

0 . 10 . 1

20.10.250.020.03

D

43.635.5

0 . 10 .3

19.60.240.080.02

plexes with layers of different competency (chemi-cally similar or not). At ~450 m below the rim solidbasalt was encountered as in previous drillings. Atabout 470, 490, and 520 m, however, rock flourlayers apparently up to ~1 m thick were encoun-tered between several meters of solid basalt withessentially no shock features. The rock flours inthese 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 weatheredbasalt layers of low competency is unclear. Ongoingstudies should yield a better understanding of thisand also give more realistic comparisons (scalingfactors) between real, large impacts and artificial ex-periments on small targets (e.g., Gault et al., 1974).


t.25

2 4

- 2 3

22

21

2 0

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 X

oo

a °<f o oo o

1.0% Na2O

2.0 3.0

FIGURE 11.—Plot of Na2O versus FeO + MgO, both in weight percent, for Lonar basalts andglasses; note that two "selected" glasses with high FeO + MgO plot outside the diagram.

NUMBER 22 13

Literature Cited

Apollo 17 Preliminary Examination Team (PET)1973. Apollo 17 Lunar Samples: Chemical and Petro-

graphic Description. Science, 182:659-672.Carter, J. L.

1971. Chemistry and Surface Morphology of Fragmentsfrom Apollo 12 Soil. Proceedings of the SecondLunar Science Conference, 2:873-892.

Fredriksson, K. J., Nelen, A. Noonan, C. A. Anderson, andJ. R. Hinthorne

1971. Glasses and Sialic Components in Mare ProcellarumSoil. Proceedings of the Second Lunar Science Con-ference, 1:727-735.

Fredriksson, K., A. Dube, D. J. Milton, and M. S. Balasun-daram

1973. Lonar Lake, India: An Impact Crater in Basalt.Science, 180:862-864.

Gault, D. E., F. Horz, D. E. Brownlee, and J. B. Hartung1974. Mixing of the Lunar Regolith. Proceedings of the

Fifth Lunar Science Conference, 3:2365-2386.Gold, T., and G. J. Williams

1974. On the Exposure History of the Lunar Regolith.Proceedings of the Fifth Lunar Science Conference,3:2387-2395.

Kieffer, S. W., R. B. Schaal, R. Gibbons, F. Horz, D. J.Milton, and A. Dube

1976. Shocked Basalt from Lonar Impact Crater, India,and Experimental Analogues Proceedings of theSeventh Lunar Science Conference, 2:1391-1412.

Milton, D. J., and A. Dube1977. Ejecta at Lonar Crater, India. Meteoritics, 12:311.

Morgan, J.1978. Siderophile and Volatile Trace Elements in "High-

Magnesian" Australites and in Glasses From LonarCrater, India. In Lunar and Planetary Science, IX,pages 754-756.

Schaal, R. B., and F. Horz1977. Shock Metamorphism of Lunar and Terrestrial

Basalts. Proceedings of the Eighth Lunar ScienceConference, 2:1697-1729.

Silver, L. T.1975. Volatile Lead Components in Lunar Regolith at the

Apollo Mare Sites. In Lunar Science VI, pages 738-740.

Chemical Variation among Australian Tektites

Brian Mason

ABSTRACT

Microprobe analyses for major components of 60australites from specific locations covering most ofthe strewnfield show a wide range of composition.Regional variations are seldom clearly defined, andaustralites of closely similar composition may befound at widely separated locations. Three distinc-tive geographic-compositional groups can, however,be distinguished: (1) a group with high Fe and Mgfrom the Lake Wilson-Mt. Davies region in north-west South Australia; (2) a group with high Ca/Alratio centered around Charlotte Waters on theSouth Australia—Northern Territory border and ex-tending into southwest Queensland; (3) a groupwith high Al and low Si from Pindera in New SouthWales (australites from elsewhere in New SouthWales analyzed by other investigators also show thisfeature). Australites show a relatively restricted com-position range, whereas micro tektites from deepseasediments south and west of Australia show a muchmore extensive composition range; thus claims fora common origin of these contrasted materials arenot supported by compositional comparisons.

Introduction

Beginning in 1963 and continuing for some years,Mr. R. O. Chalmers of the Australian Museum,Dr. E. P. Henderson of the Smithsonian Institution,and I travelled widely throughout Australia (exceptTasmania) for the purpose of collecting tektites(australites) and establishing the limits of thestrewnfield. The results of these travels are reportedin Chalmers et al. (1976). An important result wasthe acquisition of large collections from clearly-defined locations; one scientific drawback of many

Brian Mason, Department of Mineral Sciences, NationalMuseum of Natural History, Smithsonian Institution, Wash-ington, D.C. 20560.

collections of australites is the vagueness of geo-graphic attribution, such as "Nullarbor Plain,""Birdsville Track," etc. Our collections from clearlydefined locations (Figure 1) thus provide the op-portunity to investigate compositional variationswithin a specific location, and to make comparisonsbetween locations. By virtue of the fact that indi-vidual tektites are glassy objects usually of uniformcomposition and that australites exhibit a limitedrange of compositions (67%-79% SiO2), australitesare especially suitable for microprobe analysis.

TECHNIQUES USED.—From a single-locality collec-tion, individual specimens were selected for analysisby separation according to specific gravity. The spe-cific gravity of australites (except for the rare andenigmatic HNa/K group) ranges from about 2.36to 2.52, and is inversely related to SiO2 content,79% to 67% (Chalmers et al., 1976, fig. 17). Collec-tions were processed by flotation in bromoform-acetone mixtures of progressively increasing specificgravity, and individual specimens were selected tocover the whole range. This provided a group ofspecimens fairly evenly spaced over the total rangeof SiO2 content for that locality. A total of 60 speci-mens representing 11 localities were selected foranalysis.

A polished thin section was prepared from eachselected specimen. Another portion of the samespecimen was crushed and the refractive index ofthe crushed material determined by the immer-sion method. This determination also provided anopportunity to evaluate the homogeneity of the in-dividual australite. Thin sections of australites fre-quently show schlieren of slightly differing refrac-tive index (see, e.g., Chalmers et al., 1976, fig. 16).Most of the analyzed australites showed only minorvariation (±0.001) in the refractive index of crushedfragments, and this limited variation was confirmed

14

NUMBER 22 15

-14°

-22°

-30°

-38°

(

\

110°

1120°

\

• EARAHEEDY

1130°

MT.DAVIES*

HUGH• L YINDARLGOOOA

120°

• L. WILSON

ES«

130°I

1140°

i

ii

iiii

ii

•CHARIOTTE WATERS SPURRIE

• MACUMBA 1

o. 1 »PINDERAV # P I N E DAM

h V ) y • FROME DOWNS

^ A 1

I. /S «MANNAHILL

|40o \ ^

/

J150°1

1150°

14°-

f 22°-

\

/ 30°-

/

38°-

FIGURE 1.—Map showing locations of analyzed australite collections.

by the limited variation between individual spotsin the microprobe analysis. A few, specifically somewith high SiO2 content (>75%), showed a greatervariation in refractive index, sometimes as much as±0.003. Although Glass (1970) reports large compo-sitional variations (SiO2 60%-85%) within a singleaustralite, it must be noted that he specifically se-lected small, rare inclusions (two high-silica andthree low-silica in two thin sections) to establishthis extreme range, and also that he states (p. 241),

"The bulk of the glass has an average compositionof about 76% SiO2."

The polished thin sections of the individual aus-tralites were analyzed with an ARL-SEMO electronmicroprobe using an operating voltage of 15 kVand a sample current of 0.15 /xA; a broad beam(approximately 0.05 mm diameter) was used tocounter the local inhomogeneities discussed in thepreceding paragraph. A previously analyzed aus-tralite containing 68.35% SiO2 (Chalmers et al.,


1976, tbl. 8) was used throughout as a referencesample. Measured intensities were adjusted by com-puter using Bence-Albee factors. The followingcomponents were determined: SiO2, A12O3, FeO (allFe reported as FeO) MgO, CaO, K2O, Na2O, TiO2,MnO; MnO is not reported in the analyses, sinceit was uniformly at or near the lower limit ofmeasurability, approximately 0.1%.

Results

The analytical results are set out in Table 1 andare plotted in Figures 2, 3. The results for specificlocations are discussed below.

EARAHEEDY.—The collection from this localitycomprised 97 specimens. The specific gravity rangewas from 2.432 to 2.459, most specimens (87) fallingin the range 2.455-2.459. The four specimens ana-lyzed (Table 1, columns 1-4) cover the full specificgravity range, which is relatively small; the analysesshow a correspondingly small composition range.

LAKE YINDARLGOODA.—The collection from thislocality 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 threespecimens analyzed (Table 1, columns 5-7) coveralmost the full specific gravity range.

A considerable number of analyses of australitesfrom the Kalgoorlie region have been published,most of them by Taylor (1962) and Taylor andSachs (1964). These analyses confirm the restrictedrange of australite composition in this region. OfTaylor's 13 analyses, 11 have SiO2 contents rangingfrom 70.34% to 72.13%; the remaining two have75.17% and 77.39% respectively.

HUGHES.—This collection, from the NullarborPlain close to the South Australia-Western Aus-tralia border, comprised 248 specimens, ranging inspecific 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 lessthan 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 similarspecific-gravity distribution patterns, with a singlepeak at 2.45-2.46 on the frequency curve. Thelargest collection, that from Hughes, shows thegreatest range in specific gravity, as might be ex-

pected. Chapman et al. (1964) have shown that thisspecific-gravity distribution pattern is uniform forall collections they examined from the Kalgoorlieand Nullarbor Plain locations. The Earaheedy col-lection shows that this uniformity extends far to thenorth. Within this vast area australites with com-parable SiO2 content generally show a close cor-respondence in other components; compare forexample in Table 1, analyses 1 and 5, 3 and 9,7 and 12.

CHARLOTTE WATERS.—A small collection of 33specimens from Charlotte Waters showed a widerange in specific gravity, 2.390-2.483, with a ratherill-defined frequency peak at 2.440. Chapman et al.(1964), who examined a much larger collection(420 specimens), found a specific-gravity range of2.395-2.495, with a peak at 2.445. The analyzedspecimens (Table 1, columns 14-20) show SiO2

ranging from 69.0% to 77.6%; a significant featureis a relatively high CaO content compared to aus-tralites of similar SiO2 content from the NullarborPlain and Western Australia. Taylor (1962) andTaylor and Sachs (1964) published six analyses ofaustralites from Charlotte Waters; SiO2 rangedfrom 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 easternmargin of the strewnfield. The 99 specimens in thecollection ranged in specific gravity from 2.384 to2.489, with a mean of 2.440. Six specimens coveringthe 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 similarto those of Charlotte Waters, but extend to a higherSiO12 content; it is significant that the low-SiO2

australites from Durrie also show the high CaOcontents of Charlotte Waters australites (cf. analyses14 and 21).

MACUMBA.—This small collection (19 specimens)is of particular interest because it has a wide rangeof specific gravity (2.386-2.461), and the specific-gravity distribution plot shows two peaks, one at2.395 and one at 2.440 (Chalmers et al., 1976, tbl. 7).This two-peaked specific-gravity distribution plot ischaracteristic of australite collections from the LakeTorrens-Lake Eyre region of South Australia and isnot 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 andrefractive 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, USNM2537; 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, USNM2536; 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


TABLE 1.—Continued.

Const i tuent

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


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

A l l Fe a s FeO.


70.9%-77.3%, similar to but somewhat less exten-sive than that for Durrie. Other components arecomparable for those recorded for australites ofsimilar SiO2 content from Charlotte Waters andDurrie; in particular, the low-SiO2 (<75%) aus-tralites from Macumba also show relatively highCaO (except for analysis 28).

PINE DAM.—Australites from Pine Dam, nearLake Torrens, show a specific-gravity range of2.368-2.476, with peaks at 2.395 and 2.462(Chalmers et al., 1976, tbl. 7). Six specimens, cover-ing practically the full specific-gravity range, wereanalyzed (Table 1, columns 33-38). These analy-ses resemble those of similar SiO2 content fromMacumba, except that they are generally lower inCaO and higher in A12O3.

MANNAHILL.—A collection of 89 australites fromMannahill was available for study. The specific-gravity range was 2.377 to 2.468, with a single peakat 2.419 in the distribution curve. This single peakat 2.40 to 2.42 is characteristic of collections fromsoutheast South Australia and Victoria (Chalmerset al., 1976, tbl. 7). Seven specimens, covering thewhole specific-gravity range, were analyzed (Table 1,columns 39-45). The analyses are comparable tothose of Pine Dam australites with similar SiO2

contents, although the Mannahill australites appearto have somewhat lower CaO contents.

FROME DOWNS.—A small collection (14 specimens)from Frome Downs station was acquired after themain part of this manuscript was completed. It isof 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 gravityranges 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, peak2.450). The 14 Frome Downs australites have aspecific gravity range of 2.380-2.429; no peak isevident, probably because of the small number ofspecimens, but the mean is 2.402. The FromeDowns collection clearly differs in specific gravityrange and therefore in chemical composition fromthe Pine Dam and other collections in the LakeTorrens region to the west, and to the nearest loca-tion in the east, the Pindera area in New SouthWales. A search of the region between FromeDowns and Pindera in August 1978 found noaustralite occurrence, although the terrain appeared

favorable in many places. I conclude therefore thata significant chemical hiatus exists between thesetwo locations. The Frome Downs occurrence seemsto be the farthest north location for the low specificgravity Victoria-type australites unmixed with thehigher specific gravity population, as around LakeTorrens and Lake Eyre.

LAKE WILSON.—This location, in the far north-west of South Australia, is known to carry australitesof unusual composition. This was first recognizedby Taylor and Sachs (1964:250), who analyzed onespecimen and commented: "Sample No. 26 is un-usual in possessing, in addition to the high valuefor nickel, the highest amounts of Mg, Fe, Co andCr. The concentrations of the other elements arenormal for the silica content (69.8 per cent)."Chapman et al. (1964) published a specific-gravitydistribution diagram for Lake Wilson australitesand noted that it was different from those of allother australite locations, with a more extensivespecific-gravity range (2.395-2.505) and the highestpeak (2.465).

Six specimens from Lake Wilson were analyzed(Table 1, columns 46-51), ranging from 2.415 to2.489 in specific gravity and 74.9% to 68.7% inSiO2 content. The data plotted in Figure 2 showthat the Lake Wilson australites are uniformlyhigher in FeO and MgO than other australites ofsimilar SiO2 content. It was on this basis thatChapman and Scheiber (1969) established theirHMg group of tektites; they recognized tektitesbelonging to this group not only in Australia butin Indonesia and at one locality in the Philippines.Chapman and Scheiber (1969) and Chapman (1971)published analyses of HMg australites from thefollowing 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 GilesCreek (25°02'S, 128°18'E). These locations, togetherwith the Mt. Davies collection mentioned below,define a triangular area straddling the SouthAustralia-Western Australia border and extendingover 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 LakeWilson, 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, andalthough its limits remain to be precisely defined,they do not appear to overlap into surroundingareas of normal compositions.

MT. DAVIES.—Mt. Davies is a geographical areacentered around the Western Australia-SouthAustralia—Northern Territory conjunction, about40 miles west of Lake Wilson. Many thousands ofaustralites have been collected from this area byaborigines and sold through the Department ofAboriginal Affairs, Adelaide. I obtained a smallcollection of 12 specimens directly from the localaborigines in 1965. Five of these were analyzed(Table 1, columns 52-56), and these analyses arevery similar to those of the Lake Wilson australitesof comparable SiO2 content.

PINDERA.—This is an isolated location in north-western New South Wales, rather far removed fromother areas of tektite concentration. Australites fromthis location have a unique appearance, with pittedsurfaces resulting from the presence of small inter-nal bubbles, a feature rare in australites from otherlocalities. As a result, specific gravities of Pinderaaustralites show a wide range (2.270-2.470 in acollection 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 verysimilar 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 ofcomposition, 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 Pinderaspecimen and two other analyzed australites fromNew South Wales have uniquely high A12O3 con-tents (>14.9%), and this may be related to theirunusually low SiO2 contents (66.8%-68.5%). Figure2 shows that apart from the low SiO2 and highA12O3 contents, the only other distinguishing fea-ture in the analyzed Pindera australites is a rela-tively low CaO content. Nevertheless, it is possiblethat the New South Wales portion of the strewnfielddoes have a distinctive chemical composition; moreanalyses from additional locations are needed toestablish this.

Discussion of Analyses

Most of the available analyses of australites havebeen 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; (sourceof data by column: 1, this paper, 60 analyses; 2, Taylor (1962) and Taylorand Sachs (1964), 32 analyses for SiO,2, A1,,,O3, FeO and 43 analyses for othercomponents; 3, Chapman and Scheiber (1969) and Chapman (1971), 23analyses)

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.


tained by these investigators with those reportedin 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 jsspecific-gravity screening of approximately 47,000tektites), and their identification as tektites has beenquestioned by Chalmers et al. (1967:37-38).

The comparison shows that the ranges of com-position for the major components are very similar qin each of the three groups. It can therefore be ^stated with some confidence that the compositionalrange of australites is now well established, andspecimens falling outside this range are unlikely tobe found.

The extent of the variation in major componentsis clearly demonstrated in Figure 2, in which thepercentages of the individual components in each <Sanalysis are plotted against the SiO2 percentage ofthat analysis. The first impression is the limitedscatter within each component. This is especiallycharacteristic for FeO, MgO, K2O, and Na2O (notethat the vertical scale for K2O and Na2O is twice

0

that for the other components, which tends to em- £phasize the scatter for the alkalies). Another obvious **feature is the general decrease in A12O3, FeO, MgO,and CaO with increasing SiO2 content; however,this trend is not marked for Na2O and may beabsent in K2O.

Taylor and Sachs (1964) computed correlation |coefficients for their 32 analyses for which SiO2 de-terminations were available. The correlation coeffi-cients with SiO2 were Al, -0.907; Mg, -0.955; Ti,-0.842; Fe, -0.817; Na, -0.805; K, -0.764; Ca,-0.221. For all these elements except Ca, the nega-tive correlation is significant at less than 0.01 percent.

Chapman has used his analyses of Australasiantektites to establish a chemical classification of theseobjects. Within the Australian strewnfield he has 3distinguished the following classes: normal, HCa, ^HMg, HA1, HNa/K (not considered here). His defi-nitions of these classes are not unequivocal; in factthey vary somewhat from time to time, as the fol-lowing quotations show.Normal: The predominant type in Western Australia and

the Philippines, with a restricted range of CaO and MgOrespectively about 3.1% and 2.2% (Compston and Chap-man, 1969:1024).

SiO2 70% to 73%, alkalies higher than in, for example, the

•*»**»%•**

©o

•••o•

mo•o®

EarahxdyL. YindarlgoodaHughttCharlotte WatertDurri.MacumbaPin* DamMannahillPind.ral.WiltonMl. Daviw

B*

FIGURE 2.—Plot of SiO, versus Na2O, K2O, CaO, MgO, FeO,and A12O3, all in weight percent, for australite analyses inTable 1.

NUMBER 22 23

HCa group (Chapman and Scheiber, 1969:6761).GaO > MgO and Ni < 41 ppm (Chapman, 1971:6313).

HCa: CaO varying from about 2.0 to 10.0% and MgO from1.3 to 2.5% (Compston and Chapman, 1969:1024).

FeO 3.5-4.9%, Ni 14-42, Cr 56-110, Co 7-18 ppm, Cr/Ni2.0-4.9 (Chapman and Scheiber, 1969:6741).

LSG-HCa streak . . . stretching from Tasmania to CentralAustralia . . . 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% andcomparatively restricted variation in CaO (Compston andChapman, 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 > 190ppm (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 afew in the HMg class. The criterion CaO varyingfrom about 2.0% to 10.0% and MgO from 1.3% to2.5% for the HCa class applies to practically allaustralites; reference to Table 1 and Figure 2 showsonly three australites with less than 2% CaO, andthe only australites with MgO >2.5 belong to theHMg class. Australites generally contain less than6% CaO; tektites with CaO greater than 6% seemto 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 areconfined to Western Australia. The HMg australitesdefine a streak extending northwest from the LakeWilson-Mt. Davies region. The HCa australites de-fine a streak from Victoria through South Australiato Charlotte Waters and Henbury. The HA1 austra-lites are found only in New South Wales, with theexception of one from Tasmania.

Figure 2 provides the opportunity to try to fit theanalyses 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 (andFeO) than other australites of similar SiO2 content.The utility of an HA1 group is doubtful; few are

known, and their A12O3 content (>14.9%) beingcoupled with low SiO2 content puts them on thesame 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 thanthe remainder. Thus an HCa class can possibly berecognized, but as pointed out above the criteriagiven for this class by Compston and Chapman(1969) apply to practically all analyzed australitesexcept those of the HMg class.

Robert Fudali has demonstrated to me that a plot(Figure 3) of AL>O3 versus CaO values from theanalyses in Table 1 shows a well-defined hiatus.All the values plotting above the hiatus are ofaustralites from Charlotte Waters, Durrie, andMacumba. This pattern defines a coherent and re-stricted geographical region extending from north-central South Australia to adjoining southwestQueensland. Not all analyses from these locations,however, yield values that plot above the hiatus;two of seven from Charlotte Waters, four of sixfrom Durrie, and one of six from Macumba plotbelow it. The values which plot above the hiatusprobably all represent HCa australites according toChapman, but many of those plotting below thehiatus also belong to this class. The values from fiveanalyses of HCa australites published by Chapmanand Scheiber (1969) have been plotted on Figure 3;three (all from the Charlotte Waters area) plotabove the hiatus, one (from Victoria) plots below it,and one (from near Macumba) plots just within itslower limit.

One purpose of this research was to see whetheraustralites from a specific locality showed a distinc-tive range of chemical composition. Table 1 andFigure 2 show that this is not generally the case,although some regional trends can be distinguished.The Lake Wilson-Mt. Davies group in the farnorthwest of South Australia are uniformly some-what higher in MgO and FeO than other australitesof similar SiO2 contents, but for the other majorcomponents they fall within the range of most aus-tralites. The Pindera australites have uniquely highA12O3 contents. Some but by no means all the aus-tralites from localities extending from Mannahill toCharlotte Waters in South Australia and Durriein Queensland have high CaO (>4%); some ofthese also have lower A12O3 contents than most


EaraheedyL. YindarlgoodaHughesCharlotte WatersDurrieMocumbaPin* DamMannahillPinderaL. WilsonMt. Davies

12 13% ALO,

14 15

FIGURE 3.—Plot of Al O3 versus CaO, both in weight percent, for the analysisin Table 1 (c = HCa australites by Chapman and Scheiber (1969); diagonallines define locality-related hiatus).

other australites. Australites from Western Australiaand adjacent parts of the Nullarbor Plain tend tobe slightly higher in K2O and Na2O than australitesfrom other localities.

These results may be compared with the observa-tions of Taylor and Sachs (1964:241). They detectedregional differences between average australite com-positions from localities near Kalgoorlie, CharlotteWaters, and in Victoria. They commented:

Compared to the Charlotte Waters group (II), the Kalgoorliegroup (I) have a distinctly lower average for SiO2 and arehigher in those elements which show significant negativecorrelations with SiO2 . • . The Victorian group (III), althoughcontaining low values for the alkali elements, are generallyintermediate in average composition. It thus appears thattwo distinct groups, at Kalgoorlie and Charlotte Waters,may be distinguished. Perhaps a mixture of these two typesis found in Victoria.

My results are consistent with these comments. I didnot analyze any Victorian australites, but many ofmy analyses of Pine Dam and Mannahill australitescan be matched with published analyses of Victorianaustralites.

Australites and Microtektites

In 1967 Glass announced the discovery of micro-scopic (< 1 mm diameter) glassy objects in deep-seasediments from south and west of Australia: "Onthe basis of their geographical distribution, appear-ance, physical properties, and age of deposition, itis concluded that the glassy objects discussed in thisreport are microtektites, and that they constitute aportion of the Australasian strewn field whichextends from Thailand to Tasmania" (Glass, 1967:374). Not all tektite researchers have accepted theidentity of microtektites with tektites; a summaryof opposing viewpoints is provided by Chalmerset al. (1976:16).

Cassidy et al. (1969) published microprobe anal-yses of 60 microtektites and Glass (1972b) reported44 analyses of a particular variety which he calls"bottle green microtektites." Cassidy et al. andGlass both conclude that their chemical analysessupport the identification of these microtektiteswith Australian tektites. A rigorous comparison,however, of their analyses of microtektites with the

NUMBER 22 25

analyses of australites reveals many inconsistenciesand may negate their conclusion.

The range of composition of microtektites ismuch wider than that for australites. For the prin-cipal component, SiO2, the range for microtektitesis 48.1% to 77.0%; of the 104 microtektite analysesfrom the Australian region, only 46, or less thanhalf, 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 theAustralasian tektites and served to extend theknown composition range of tektites. Glass claimedthat the bottle green microtektites overlap andextend all major-element trends found for the HMgtektites of Chapman and Scheiber (1969). Even ifthis claim were valid, the non-existence of austra-lites with less than 66% SiO2 renders it of littlerelevance.

When the data, however, for microtektites withinthe 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 onlybe described as poor. Discordances are particularlyprominent for Na,O (most microtektites have lessthan 1%, most australites more than this amount);for K2O, where only four microtektites fall withinthe field for australites and two of these are dis-cordant with australites in other components; andfor MgO, where practically all the bottle greenmicrotektites have much higher contents than theHMg australites with which they are claimed tocorrespond. The chemical discordance betweenmicrotektites and australites is clearly displayed inFigure 4, which plots FeO versus MgO percentagefor australites and microtektites. The australitesshow a very limited range in FeO and MgO con-tents and an extremely strong positive correlationbetween these two components. Note the contrastin the microtektites, which show a very wide rangein 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 ofaustralites, argues against rather than for a com-mon source material. If one is to claim that micro-tektites and australites are cogenetic, the evidencemust be from sources other than chemical compo-sition.

Literature Cited

Cassidy, W. A., B. Glass, and B. C. Heezen1969. Physical and Chemical Properties of Australian

Microtektites. Journal of Geophysical Research,74:1008-1025.

Chalmers, R. O., E. P. Henderson, and B. Mason1976. Occurrence, Distribution, and Age of Australian

Tektites. Smithsonian Contributions to the EarthSciences, 17: 46 pages.

Chapman, D. R.1971. Australasian Tektite Geographic Pattern, Crater and

Ray of Origin, and Theory of Tektite Events. Journalof Geophysical Research, 76:6309-6336.

Chapman, D. R., H. K. Larson, and L. C. Scheiber1964. Population Polygons of Tektite Specific Gravity for

Various Localities in Australasia. Geochimica etCosmochimica Ada, 28:821-839.

Chapman, D. R., and L. C. Scheiber1969. 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.


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 Geophysical1970. 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. SachsD. 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 providechemical analyses for 20 of them, and have selectedanalyses 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 thisdifinition, are distinct in their chemical and min-eralogical composition, and the term shergottiteshould be retained for them. A large number ofeucrites (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 notmechanical mixtures of eucrites and diogenites;their individual mineral and rock clasts belong toa differentiation sequence that is compositionallymore extensive than the eucrites and diogenites.

Introduction

The pyroxene-plagioclase achondrites are a groupof meteorites consisting essentially of pyroxene andcalcic plagioclase, the pyroxene usually dominantover plagioclase. They compose a majority of theachondrites, totalling over 50 of approximately 90known. Prior (1920) grouped them together withAngra 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 tocall the calcium-rich achondrites the basaltic-typeachondrites (now generally shortened to basalticachondrites), to emphasize their general similarityto terrestrial basalts (although the latter are clearlydifferent in the amount and composition of theplagioclase).

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 usageis the mineral now known as pigeonite.) In thecurrent 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 beingpigeonite in eucrites and orthopyroxene (hyper-sthene) in howardites. Mason (1962) noted that thisdistinction could be correlated with the bulk chem-istry of these meteorites, most clearly with the CaOcontent and the FeO/(FeO + MgO) molecular per-centage; he suggested that a dividing line could bedrawn at FeO/(FeO + MgO) = 50, eucrites showinga higher ratio, howardites a lower one.

Lacroix (1926) felt there were insufficient groundsfor separating eucrites and howardites, but that ifsuch separation were desired, it would be bettermade on structural grounds, howardites beingclassed 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


achondrites. This distinction was adopted by Dukeand Silver (1967) and several other researchers.Bunch (1975) provided a comprehensive account ofthe large variety of rock and mineral clasts foundin 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 themineralogical criterion is usually sufficient to class-ify unambiguously a pyroxene-plagioclase achon-drite as a howardite or a eucrite, since meteoriteswith dominant orthopyroxene are also polymictbreccias (except Binda), and most monomict brec-cias have pigeonite as the dominant pyroxene(polymict breccias with dominant pigeonite areMacibini, Bialystok and Nobleborough). In usingthe terms howardite and eucrite, however, it isdesirable to state in what sense these terms areused; in this paper we use them in the structuralsense. We also distinguish Shergotty and Zagamifrom the eucrites on account of their distinctivechemical and mineralogical composition.

In discussing the eucrites and howardites, fre-quent reference will be made to the diogenites (alsoknown as hypersthene achondrites). The diogenites,although classed by Prior as calcium-poor achon-drites, occupy one end of a spectrum of chemicaland mineralogical compositions extending throughhowardites to eucrites. Prior did not specify a limit-ing value of CaO between the calcium-poor andcalcium-rich achondrites; currently the most cal-cium-rich diogenite is Garland with 2.60% CaO(Fredriksson et al., 1976), and the most calcium-poorhowardite is Yamato-7307 (also known as Yamato(l)with 3.83% CaO (Wanke et al., 1977). The amountsof 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-greenorthopyroxene clasts with only trace amounts ofother minerals, whereas Yamato-7307, while con-taining a large amount of pyroxene similar in com-position to that in Garland, also has a considerableeucritic component (pigeonite + plagioclase). Onecharacteristic of howardites is that they frequentlycontain clasts of diogenite composition. A closerelationship, therefore, exists between these twoclasses, which will be further discussed later in thispaper.

Interest in the pyroxene-plagioclase achondrites

has increased greatly in recent years, largely becausethey 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 ofthe Apollo missions have negated this suggestion,and therefore established that these meteoritesprove the previous existence of at least one otherbody in the Solar System with analogous chemicaland mineralogical composition (and presumablysimilar evolutionary history) to the moon. The na-ture and composition of this body ("eucrite parentbody") has been the subject of intensive research(see, for example, Dreibus et al., 1977; Consolmagnoand 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 meteoritesand supplemented them with analyses of as manypreviously unanalysed meteorites as possible. Wherepublished analyses are either incomplete or showinconsistencies with the mineralogical compositionwe have attempted to rectify this. Some of the newanalyses (by E. J.) are by classical wet-chemicaltechniques; where only small samples were avail-able the material was fused with lithium tetraborateand the resulting glasses analysed with the electronmicroprobe (by J. N.).

ACKNOWLEDGMENTS.—We are indebted to thecurators of the meteorite collections in the follow-ing institutions for providing material of meteoritesnot represented in the Smithsonian Institution col-lection: Arizona State University; Australian Mu-seum, Sydney; British Museum (Natural History);Field Museum, Chicago; Geological Survey ofIndia; Museum d'Histoire Naturelle, Paris; Natur-historisches Museum, Vienna.

Results

The analyses are presented in Table 1. In thistabulation K2O and P2O5 are omitted (recent analy-ses show that K2O is consistently below 0.05% andP2O5 is usually less than 0.1%); free nickel-iron isalso 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 haveFeO/(FeO + MgO) 30-45, eucrites 40-69, therebeing a slight overlap). The ratio also correlatespositively with CaO and A12O3, and negatively withMgO. Other parameters derived from the analyticaldata are the composition and amount of normativeplagioclase, the composition and amount of norma-tive pyroxene, and the amount of normative olivineor silica. From a consideration of texture the mete-orites are classified as unbrecciated, monomict brec-cias, and polymict breccias; if texture is used as thediscriminant factor between eucrites and howard-ites, then unbrecciated meteorites and monomictbreccias are eucrites, while polymict breccias arehowardites.

Notes on the individual meteorites (bearing theanalysis numbers used in Table 1) follow.

1. YAMATO-7307.—This meteorite, also known asYamato(l), was collected by the Japanese AntarcticExpedition in 1973, and is of special interest sinceit has the lowest FeO/(FeO + MgO) mole percentageof 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 recognizedin a few), with uniform pyroxene composition(Fs 25-27). Sections of Yamato-7307 which we haveexamined show that it is a polymict breccia, witha wide range in pyroxene composition (Fs 19-49);two grains of olivine were seen (Fa 17, 37), and adiogenite enclave consisting essentially of pyroxenewith 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 rangeof pyroxene compositions (Fs 22—46); plagioclasealso shows a considerable range in composition(Ab 4-21). It should be mentioned that specimensof a chondrite mislabelled as Massing have been dis-tributed (the specimens in the Museum d'HistoireNaturelle, Paris, and Eidgenossische TechnischeHochschule, Zurich are such); the analytical dataon Massing A reported by Schmitt et al. (1972) areprobably of this mislabelled material.

3. CHAVES.—Thin sections of this meteorite showthat it is a typical polymict breccia. The ground-mass consists largely of comminuted orthopyroxeneenclosing larger shocked clasts of the same mineral;

eucrite enclaves (unbrecciated, with ophitic texture)are not uncommon, and one fragment of diogenitewas seen. Jeremine (1954) has provided a thoroughpetrographic description of this meteorite; she notedthe presence of accessory tridymite and quartz,which we have confirmed with the microprobe (notethe 0.5% of free SiO2 in the norm). Our microprobeanalyses show that the pyroxene consists largely oforthopyroxene, 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 describedby Mason and Wiik (1966a). It is a polymict breccia,with pyroxene clasts (mainly orthopyroxene, somepigeonite) in a comminuted groundmass; one smalleucrite enclave was observed. Our microprobe analy-ses show pyroxene compositions ranging from Fs 24to Fs 39; similar results were reported by Takeda,Miyamoto, Ishii, and Reid (1976).

5. ZMENJ.—Thin sections show a typical polymictbreccia, consisting largely of clasts of hypersthene(major) and pigeonite (minor) in a comminutedgroundmass. Our microprobe analyses show pyrox-ene compositions ranging from Fs 23 to Fs 44, andplagioclase ranging from Ab 4 to Ab 10 (normativeplagioclase is Ab 13, which suggests that the Na2Oin the analysis may be a little too high). Desnoyersand Jerome (1973) analysed a number of grains ofolivine in Zmenj, and found compositions rangingfrom Fa 15 to Fa 39.

6. BINDA.—This meteorite, originally classified asa howardite because the pyroxene is largely hyper-sthene, was reclassified by Duke and Silver (1967) asa eucrite because it is a monomict breccia. Thepyroxene in Binda has been described by Takedaand Ishii (1975); they comment (page 500):

The bulk composition of low-Ca pigeonitesuggests the crystallization temperature is above 1110°C. Thetemperature where augite blebs Ca44.7Mg42.4Fe12.9 developedin the host orthopyroxene Ca2.4Mg64.4Fe33.2 is about 970°C.The Binda howardite contains the most magnesian pigeoniteamong 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 Bindameteorite may be classified as the most Mg-rich and Ca-pooreucrite.

7. PAVLOVKA.—Thin sections show a polymictbreccia largely of mineral grains (hypersthene withminor pigeonite and plagioclase), but small- tomedium-sized eucrite enclaves are not uncommon;some plagioclase grains are unusually large, larger


than could be derived from common eucrites, sug-gesting a possible anorthositic parent. Our probeanalyses indicate an unusually wide range ofpyroxene compositions; orthopyroxene is dominant,with Fs ranging 14-36, with lesser amounts ofpigeonite, 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 grainsof olivine, Fa 33, 35 were analyzed. Desnoyers andJerome (1973) analyzed 15 grains of olivine inPavlovka, 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 givethe range of composition of the pyroxenes asFs 20-47. They describe several kinds of lithic clasts,among them eucritic fragments with an ophitic orsubophitic texture, and a cm-sized dunite xenolith(olivine composition Fa 12.8).

9. BHOLGATI.—A thin section shows a breccia withnumerous clasts of orthopyroxene up to 1 mmacross, small less numerous clasts of pigeonite andplagioclase, and rare grains of chromite, in a fine-grained groundmass. Our microprobe analyses showa range of Fs 23-62 in orthopyroxene and pigeonite

KEY TO TABLE 1Identification 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); Wankeet 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 normativeplagioclase; d = mole percent Wo, Fs in normative pyroxene; e = weight percent normativepyroxene; 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 onfacing 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



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


Const i tuent

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; onegrain 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 polymictbreccia mostly of individual mineral grains (hyper-sthene, pigeonite, plagioclase) but with occasionaleucrite clasts; one section had an anorthosite clast(~90% plagioclase, 10% pyroxene). Microprobeanalyses showed pyroxene compositions rangingFs 33-52, plagioclase ranging Ab 4-19.

11. YURTUK.—Thin sections show a polymict brec-cia with large (up to 1 mm) orthopyroxene andsmaller pigeonite clasts; lithic fragments are alsopresent, both eucritic and diogenitic; rare grains ofolivine were seen. Microprobe analyses show pyro-xene ranging in composition from Fs 31 to Fs 50.

12. BUNUNU.—This meteorite was described byMason (1967a), with an analysis by E. Jarosewich.We have made microprobe analyses of the pyrox-enes and the plagioclase. The range of plagioclasecomposition is Ab 6-17, average 10. Orthopyroxeneand 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 ahowardite because of its high content of Ca-poorpyroxene; however, Hutchison et al. (1977) reclassifyit as a eucrite, commenting (page 149): "Furtherunpublished work . . . indicates that it is a eucrite,mineralogically similar to Moore County and Serrade Mage." The analysis is noteworthy for the verylow TiO2 and CraO3 and high A12O3.

14. LE TEILLEUL.—Thin sections show this mete-orite is a polymict breccia; hypersthene clasts (somewith augite exsolution lamellae) predominate, withminor amounts of pigeonite and rare eucrite en-claves; one enclave of coarse plagioclase, reminiscentof an anorthosite, was seen. Microprobe analysesshow a range of Fs 24-41 in hypersthene andpigeonite, and Ab 10-24 in plagioclase. Desnoyersand Jerome (1973) analyzed 13 grains of olivine inLe Teilleul, and found a composition range ofFa 28-42.

15. MOLTENO.—This meteorite was described byFrost (1971) as a howardite, with pyroxene rang-ing from Wo2Fs22En76 to Wo4Fs58En38 and then toWo41Fs26En33, with plagioclase (An 84), and withsmall quantities of olivine (Fa 12), troilite, nickel-iron, ilmenite and chromite. Desnoyers and Jerome


(1973) analyzed 9 grains of olivine in Molteno, andrecord a composition range of Fa 11-56.

16. MOAMA.—The analysis of this meteorite showsthat it is very similar to Medanitos. Lovering (1975)describes it as an unbrecciated adcumulate of pyrox-ene and plagioclase, with accessory amounts ofchromite, troilite, nickel-iron, and tridymite (thepresence of olivine in the norm and free silica astridymite in the mode is probably due to all Febeing reported as FeO, whereas some is presentas troilite and nickel-iron). Lovering's microprobeanalyses 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 similarto other howardites with two unique exceptions: 1, largecm size lithic clasts of cumulate ferrogabbros, and 2, frag-ments of C2 carbonaceous chondrites {8 vol. %). The cumulateferrogabbros contain large 1 to 5 mm ferrohypersthene orferropigeonite grains with exsolved ferroaugite lamellae, calcicplagioclase (An 84-89), ilmenite, FeS, and rarely free silicaand phosphates. Bulk analyses of these clasts are similar tolithic clast ferrobasalts in other howardites . . . The C2fragments range in size from 1/t to 6 mm and are similarin bulk composition and mineralogy to C2 inclusions in theAbbott and Plainview chondrites and the Kapoeta andWashougal 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 grainedanorthositic 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 insection . . . It is highly equilibrated and the pyroxenes recorda complex subsolidus history.

Serra de Mage* is unique in having the highestA12O3 content and hence the highest plagioclasecontent of any meteorite. In its high plagioclasecontent, and in other compositional and texturalfeatures it is clearly related to Medanitos, Moama,Nagaria, and Moore County.

19. KAPOETA.—The polymict character of thismeteorite was clearly established by Fredriksson andKeil (1963), in one of the first applications of micro-probe analysis to meteorite research; they analyzed132 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 Ferange of 7.5 to 26.3; they also analyzed five olivinegrains, with Fa ranging from 8.5 to 37.7 molepercent. Their work has been greatly extended byDymek et al. (1976), who comment:

Kapoeta is a "regolith" meteorite, and mineral-chemical andpetrographic data were obtained for numerous other rockand mineral fragments in order to characterize the surfaceand near-surface materials on its parent body. Rock clastscan be grouped into two broad lithologic types on the basisof modal mineralogy—basaltic (pyroxene- and plagioclase-bearing) and pyroxenitic (pyroxene-bearing). Variations inthe compositions of pyroxenes in rock and mineral clasts aresimilar to those in terrestrial mafic plutons such as Skaer-gaard, and indicate the existence of a continuous range inrock compositions from Mg-rich orthopyroxenites to veryiron-rich basalts. . . . We interpret these observations toindicate that the Kapoeta meteorite represents the com-minuted remains of differentiated igneous complexes togetherwith "primary" undifferentiated basaltic rocks. The presentlyavailable isotopic data are compatible with the interpretationthat this magmatism is related to primary differentiation ofthe Kapoeta parent body. In addition our observations pre-clude the interpretation that the Kapoeta meteorite is asimple 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 prominentcontrasted light-dark areas. It has been described bySimpson (1975) and Desnoyers and Jerome (1977),who clearly established its nature as a polymictbreccia.

21. POMOZDINO.—Kvasha and Dyakonova (1972)describe this meteorite as a eucrite. A section wehave examined shows that the meteorite is a mono-mict breccia, the breccia fragments consisting of asubophitic intergrowth of plagioclase and pigeonite.

22. BRIENT.—This meteorite is described byZavaritskii and Kvasha (1952) as a howardite, andtheir illustrations (page 230) show that it is apolymict breccia. Thus in spite of the close simi-larity in chemical composition between Brient andPomozdino they are quite different structurally.

23. PETERSBURG.—Thin sections show that thismeteorite is a polymict breccia; the dominant clastsare subequal amounts of orthopyroxene and pigeon-ite (up to 1 mm across); plagioclase clasts are alsopresent, and occasional olivine clasts (one olivineclast 2 mm across); small ophitic eucrite enclavesare not uncommon. Microprobe analyses showorthopyroxene 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 showingthe 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 MooreCounty in chemical and mineralogical compositionand texture. The minerals are highly equilibrated.Microprobe analyses show plagioclase of uniformcomposition, Ab 8; the pyroxene evidently crystal-lized originally as pigeonite, but the individualgrains are now hypersthene (Wo1.3Fs5o.4En,48-.3) withexsolved augite (Wo43.3Fs18.9En37.8).

25. MOORE COUNTY.—A comprehensive descrip-tion of this meteorite was published by Hessand Henderson (1949). Additional information onthe pyroxenes has been provided by Ishii andTakeda (1974); they showed that the primarypigeonite (Wo101Fs44.3En4g 6) has exsolved augite(Wo415Fs217En36 8) and the host pyroxene now hasthe composition Wo5 8Fs47 5En40 7. Plagioclase hasuniform 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 meteoritesin this group. It contains much more Na2O thanthe other meteorites, and as a result the plagio-clase composition is that of labradorite instead ofbytownite-anorthite; the plagioclase has been shock-transformed into the isotropic form maskelynite. Italso contains magnetite, which indicates a muchhigher oxygen fugacity in its place of crystalliza-tion than for the eucrites and howardites. Duke(1968) has provided a comprehensive descriptionof Shergotty, with analyses of the bulk meteorite,the maskelynite (Or:2Ab49An49), and the pyroxene(Wo22Fs41En37); the pyroxene is strongly zoned, asshown by Duke's microprobe analyses.

Binns (1967) has shown that Zagami is closelysimilar to Shergotty both in texture and mineralogi-cal composition; for an analysis of Zagami, seeEaston and Elliott (1977). We have compared thinsections of the two meteorites, and find that Zagamidiffers from Shergotty only in being somewhat finer-grained (feldspar laths average twice as long inShergotty); it is certainly possible, if not probable,

that both had a common parent body. The chemi-cal composition of Shergotty places it well awayfrom the main trend of the eucrites and howardites(Figure 3). It is significant that Bunch (1975) foundno shergottite clasts in the howardites, although hefound numerous clasts of eucrites, and less numer-ous clasts of other meteorite types. All these factsindicate that Shergotty (and Zagami) probably comefrom a different parent body than the otherpyroxene-plagioclase achondrites.

27. MACIBINI.—This meteorite has been describedby Reid (1974), as follows:

The Macibini meteorite is a complex polymict breccia withrock and mineral clasts that are extraordinarily diverse intexture and composition . . . . Pyroxenes encompass a rangefrom magnesian orthopyroxene through pigeonite and ferro-pigeonite to ferroaugite and to hedenbergite. The trend mim-ics, and extends the pyroxene trend shown by the basalticachondrite group of meteorites. The more magnesian py-roxenes are coarser grained and tend to occur as mineralclasts; the more iron-rich pyroxenes are commonly presentwithin fine-grained eucritic lithic fragments that have sub-ophitic textures. . . . From the coherence of the mineraldata, the parent materials are not a random set of unrelatedrocks but appear to be derived predominantly from a suiteof related samples, that may have differentiated from asingle source The sequence of mineral compositions appearsto be continuous and the basaltic achondrites appear to berelated by fractional crystallization, involving dominantlypyroxene. There is evidence for the existence of achondriteswith compositions intermediate between the hyperstheneachondrites and the eucrites; howardites are not simply two-component mixtures.

We have examined sections of Macibini and agreewith Dr. Reid's comments. Our microprobe analysesshow orthopyroxene and pigeonite with a composi-tion range Fs 19-61, together with occasional grainsof ferroaugite; plagioclase composition range isAb 7-21. Desnoyers and Jerome (1973) analyzed11 grains of olivine in Macibini, with a compositionrange Fa 56-83 (unusually iron-rich).

28. BIALYSTOK.—This meteorite resembles Maci-bini closely in composition and texture. It is apolymict 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 compositionof Fs 25-62 for orthopyroxene and pigeonite, andAb 7-12 for plagioclase; three grains of olivine,Fa 71-73, were analyzed, and rare grains of a silicapolymorph, probably tridymite, were seen.

29. YAMATO-74450.—This meteorite, collected by


the Japanese Antarctic Expedition in 1974, has beenanalyzed by Wanke et al. (1977), who describe it asa eucrite. We have been informed (H. Takeda, pers.comm.) that Yamato-74450 is a monomict breccia,with pyroxene compositions similar to those inPasamonte.

30. IBITIRA.—This meteorite has been describedby Wilkening and Anders (1975) and Steele andSmith (1976). The latter authors describe it asfollows:

Ibitira meteorite is interpreted as a strongly metamorphosed,unbrecciated, vesicular eucrite with a primary variolitic andsecondary hornfelsic texture dominated by 60% pyroxene(bulk composition En37Fs48Woi5 exsolved into lamellae sev-eral micrometers wide of augite En32Fs27Wo4i and pigeoniteEn40Fs56Wo4) and 30% plagioclase An94 (mosaic extinctionand variable structural state). Minor phases are 5% tridymiteplates one-quarter occupied by plagioclase (An94) inclusions;several percent intergrowths of ilmenite and Ti-chromitewith 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 indicatevapor deposition. These properties can be explained by aseries of processes including at least the following: (1) igneouscrystallization under pressure low enough to allow vesicula-tion, (2) prolonged metamorphism, perhaps associated withvapor 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 andtridymite but not the pyroxene, (4) sufficient annealing toallow recrystallization of the plagioclase and tridymite, andpartial 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 areasretaining the igneous subophitic texture of inter-grown plagioclase and pyroxene. Our microprobeanalyses (Figure 1) show pyroxene ranging in com-position from Wo2Fs65En34 to Wo47Fs24En29, the Encontent remaining almost constant while the Woand Fs contents vary inversely; this pattern ischaracteristic for most eucrites, as noted by Dukeand Silver (1967) for Sioux County, Juvinas, andNuevo Laredo. Plagioclase composition ranges fromAb 7-17, with an average of Ab 11.

32. Sioux COUNTY.—Duke and Silver (1967) de-scribe this meteorite as a monomict breccia madeup of coarse-grained lithic fragments with equi-granular and ophitic textures. They show the rangein composition of the pyroxenes in the form of adiagram (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 Olsenet 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 bulkcomposition so close to the average for eucrites, weclassify this meteorite as a eucrite" (p. 223). This isclearly a meteorite whose classification as a eucriteor as a howardite depends on a subjective evalua-tion of the amount of foreign admixture.

34. NOBLEBOROUGH.—Thin sections show thismeteorite to be a polymict breccia, with clastsdominantly of pyroxene (mostly pigeonite), minorplagioclase (Ab 7-17, average 9), some tridymitegrains, and occasional lithic fragments of eucriticcomposition and texture; one grain of unusuallyiron-rich olivine (Fa 88) was found. The polymictcharacter is well illustrated by a plot of pyroxenecompositions 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. Theextreme contrast in pyroxene compositions betweenJonzac (a monomict breccia) and Nobleborough(a polymict breccia), in spite of their almost identi-cal bulk compositions, supports the division of thepyroxene-plagioclase achondrites into eucrites andhowardites on the basis of structure rather thancomposition. Of course, the mere existence in thesemeteorites of pyroxenes with a wide range of com-position is not necessarily evidence for a polymictorigin; a similar range of composition for a zonedsingle crystal has been recorded for many pyroxenesin lunar basalts (e.g., Mason et al., 1972). The widerange of pyroxene compositions in Nobleborough,however, has been recorded on discrete grains, notzoned crystals. This indicates that at least some ofthe rocks from which these grains were derived wereprobably not cogenetic.

35. CHERVONY KUT.—This meteorite has beendescribed by Zavaritskii and Kvasha (1952). Theirillustration (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 theJonzac eucrite, two pyroxene-plagioclase achondrites with very similar bulk compositions.

large areas, however, are unbrecciated, and show anequigranular to subophitic association of pigeoniteand plagioclase. Cachari has been described byFredriksson and Kraut (1967) as a rather typicalbrecciated eucrite, consisting of orthopyroxene(Wo2Fs59En39) and clinopyroxene (Wo4l2Fs27En31),plagioclase (An 88), with minor amounts ofchromite, ilmenite, and troilite.

37. KIRBYVILLE.—This meteorite has never beendescribed; a small portion was kindly made avail-able to us by its owner, Mr. Oscar Monnig. It is amonomict breccia; pyroxene compositions, deter-mined by the microprobe, are similar to those inJonzac, ranging from Wo,Fs,66En32 to Wo40Fs29En31.A silica polymorph, probably tridymite, was de-tected with the microprobe.

38. JUVINAS.—This meteorite has been extensivelydescribed in the literature, most recently by Dukeand Silver (1967). They record it as a monomictbreccia, and provide a plot of pyroxene composi-tions, which show a continuous range from pigeon-ite through subcalcic ferroaugite to ferroaugite withrelatively constant MgSiO3 contents, similar to thatof Jonzac and other eucrites.

39. MILLBILLILLIE.—A sawn surface of this me-teorite presents a brecciated appearance, with aprominent light-dark structure. The brecciated ap-pearance is not so marked in a thin section, butcan be seen as areas of coarser and finer grain.Our microprobe analyses show pryoxene composi-

tions similar to those in Jonzac, ranging fromWo2Fs66En32 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 achondriteenclaves, which have been described by McCall(1966). We have examined one of these, an un-

brecciated hypidiomorphic-granular aggregate ofpyroxene and plagioclase. Its bulk composition isessentially identical with many eucrites. Microprobeanalyses show pyroxene compositions ranging fromWo2Fs64En34 to Wo20Fs49En32, i.e., pigeonite to sub-calcic ferroaugite; plagioclase composition is veryuniform, Ab 8.

41. PADVARNINKAI.—This meteorite has beenclassed as a shergottite, because it contains mas-kelynite; however, Binns (1967) showed that it dif-fers significantly in chemistry and mineralogy fromShergotty, and it still retains some untransformedplagioclase, which has a bytownitic compositionquite different from the Shergotty maskelynite. Thechemical analysis shows that its composition is es-sentially identical with other eucrites. Our micro-probe analyses show pyroxene compositions rangingfrom Wo,2Fs,67En31 to Wo44Fs27En29, similar to thosein Millbillillie; the plagioclase composition rangeisAb 10-23, average 11.

42. ADALIA.—This meteorite is unbrecciated, withan ophitic texture. Our microprobe analyses show


pyroxene ranging from WoaFs^Engg to Wo36Fs34

En30, closely similar to the preceding meteorites;plagioclase ranges Ab 9-18, average 11; accessorygrains of a silica mineral were noted.

43. PALO BLANCO CREEK.—This meteorite is de-scribed by Lange and Keil (1976) as a monomictpigeonite-plagioclase achondrite. Their plot ofpyroxene compositions show them closely clusteredin two areas, one pigeonite and one ferroaugite; ourmicroprobe 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 byBerwerth (1903); it is a typical monomict eucriticbreccia. Our microprobe analyses show pyroxenecompositions clustered around Wo2Fs63En35 andWo43Fs27En30, 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). Itis a monomict breccia; some areas retain the originalophitic texture of pigeonite and plagioclase, butmuch of the meteorite is strongly crushed and brec-ciated. Our microprobe analyses show pyroxeneranging in composition from Wo3Fs65En32 toWo26Fs46En28, and plagioclase from Ab 12 to Ab 20,average 17.

46. HARAIYA.—This meteorite is a monomictbreccia, with areas of original ophitic texture sep-arated by areas of crushed material; some of thecrushed material appears to have recrystallized intoa 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 lesseramounts of plagioclase, and lithic fragments withophitic texture. It has been classified as a monomictbreccia by Duke and Silver (1967) and by Takeda,Miyamoto, and Duke (1976). Takeda et al. haveshown that Pasamonte is almost unique among themonomict breccias in having strongly zoned pyro-xenes with no visible exsolution lamellae. Ourmicroprobe analyses show pigeonite compositionsranging from Wo6Fs33En61 to Wo6Fsr,5En29, withsome Ca-rich pyroxene ranging up to Wo33Fs39En28.

48. BEREBA.—The mineralogy and petrology ofthis meteorite were comprehensively described byLacroix (1926). It is a monomict breccia; our mi-

croprobe analyses show pyroxene of compositionWo3Fs61En36 with narrow exsolution lamellae offerroaugite (Wo36Fs35En29), and plagioclase withcomposition 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 brecciawith veinlets of brown glass. It is noteworthy forbeing the finest-grained of all the eucrites, with ahornfelsic hypidiomorphic-granular texture, thepyroxene grains averaging 0.02-0.03 mm across andset in a plagioclase matrix. Our microprobe analysesshow hypersthene (Wo2Fs66En32) coexisting withaugite (Wo47Fs25En28). Plagioclase compositionsrange from Ab 9 to Ab 20, and average Ab 13. Thepyroxene compositions indicate that Emmavilleprobably 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 monomictbreccia in which the lithic clasts are rather finegrained. Our microprobe analyses show pyroxenecompositions 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 theanalysis by McCarthy et al. (1974). It is a monomictbreccia, the individual lithic clasts being coarse-grained granular to subophitic intergrowths ofpyroxene and plagioclase. Pyroxene compositionsrange from Wo5Fs60En27 to Wo43Fs33En24; plagio-clase from Ab 10 to Ab 19, average Ab 15.

Neither analytical data nor a description areavailable for Muckera, which is catalogued byHutchison et al. (1977) as a howardite. Melrose(b), which we have not examined, was briefly de-scribed and identified as a howardite by Sibray etal. (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 highcontent of plagioclase). The FeO content is alsofairly 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 thereverse effect. The Na2O content is low through-out, except in Shergotty (1.28%). Of the minorcomponents, MnO is rather constant (Dymek et al.(1976) demonstrated that the FeO/MnO ratio inKapoeta pyroxenes is fairly uniform (~35) andsignificantly lower than in lunar pyroxenes (~60),and Marvin (1976) has confirmed this for otherpyroxene-plagioclase achondrites); Cr2O3 decreaseswith increasing FeO/FeO+MgO) molecular per-centage, whereas TiO2 increases.

A notable feature of Table 1 is the marked clusterof 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 becalled the main-group eucrites.

Additional correlations are revealed by the cal-culation of normative mineralogy (because the

actual minerals present are essentially identical withthose calculated in the norm, normative and modalmineralogy are closely similar). Since practicallyall the A12O3 is combined in plagioclase, A12O3 andplagioclase contents are directly related; plagioclasecontent increases with increase in FeO/(FeO + MgO)mole percentage, from 12% in Yamato-7307 toalmost 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 ratheruniform Na2O content; in individual meteoritesactual plagioclase composition may vary from grainto grain and within grains but will average close tothe 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 widelyin the howardites, but generally show a limited andsystematic range in the eucrites (Figure 1).

70 -

60 -

Q 50 -

30 -

- •o

©

-

Howarditat

EucritosStwrgotty

1a 9 *

7» •06

• 5• 3

I

17

* •12 • • «

• to

23

I

O51SOo

28 _• • 2 7

© 3 6

O24

0 16

o13

1 1

8CaO

10 12

FIGURE 2.—CaO (weight percent) plotted against FeO/(FeO + MgO) (mole per-cent) from analyses of the pyroxene-plagioclase achondrites (the ellipse enclosesvalues for 20 eucrites and one howardite; numbers are those of analyses inTable 1).


A significant feature that becomes evident fromthe norm calculation is the close approach to silica-saturation in all the pyroxene-plagioclase achon-drites, so close in many instances that a change ofless than 1% in SiO2 percentage would change ananalysis from olivine-normative to quartz-normative.The eucrites, however, are consistently quartz-normative (except Binda, Medanitos, Moama,Nagaria, Moore County), whereas most howarditesare olivine-normative. Accessory olivine is presentin most howardites, but we have not found it inany of the eucrites we have examined. Most eucrites(even the olivine-normative Binda, Moama, andMoore County) contain accessory free silica, usuallyas 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. Theclose groupings of analyses 29 through 49 is ap-parent; there are too many analyses to be shownindividually, but the one howardite within thiscluster is shown. Nuevo Laredo (50) and Lakangaon(51), although identical to main-group eucrites inCaO content, are clearly differentiated on FeO/(FeO + MgO) mole percentage. Outside the main-group eucrites are Binda (6) and six probablecumulates: Medanitos (13), Moama (16), Serra deMage (18), Pomozdino (21), Nagaria (24), and MooreCounty (25). Another intriguing feature is thathowardite compositions are fairly continuous in the

• 4

#3

.7 nl

10.17 ' «

• 12

• «

• 1?

• Howarditeo Eucrites0 Shvrgotty

.20

• 2 2

.27 O*8

• 26

to 12 20

AI2O,

FIGURE 3.—MgO (weight percent) plotted against A12O3 (weight percent) from analyses of thepyroxene-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 selectedeucrites (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 andSmith, 1976); 5 = Peramiho (this paper); 6 = Haraiya (this paper); 7 = Lakangaon (this paper)).

FeO/(FeO + MgO) range 30-46, but there is then ahiatus to FeO/(FeO + MgO) = 55, with only threebeyond 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 linearrelationship between the howardites and the main-group eucrites, i.e., although the howardites arepolymict breccias, their bulk composition is dearlyconstrained. The MgO-Al2O3 relationship can beprojected to include the diogenites, whose compo-sitions cluster around 26% MgO, 1% A12O3; therelationship between diogenites, howardites, andeucrites will be discussed in a later section. Figure3 shows how the Shergotty composition falls wellaway from the eucrites and howardites, and justifiesplacing it in a special category, as already indicatedby its mineralogy. Relative to the eucrites andhowardites, Shergotty is enriched in pyroxene anddepleted in plagioclase; its texture suggests apyroxene cumulate.

As demonstrated in Figure 1, howardites are dis-tinguished from eucrites by their great range inpyroxene composition; in eucrites pyroxene com-positions may range in Wo content from pigeoniteto augite or ferroaugite, but the En content remainspractically constant. The extremes in pyroxene com-

position are plotted in Figure 4 for seven eucrites,ranging from the most Mg-rich (Binda) to the mostiron-rich (Lakangaon). In most of these meteoritesall intermediate compositions between the extremesmay be found by microprobe analyses, because ofthe 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 almostcontinuous sequence has long been known andcommented on. Mason (1962) suggested that thesethree classes might be related as successive mem-bers of a fractional crystallization sequence from amelt of approximately chondritic composition (thepallasites being the initial member of the sequence).Jerome and Goles (1971) pointed out that howard-ites are polymict breccias, and therefore are notdirect products of magmatic crystallization; fromthat observation they proceeded to the postulatethat howardites are mechanical mixtures of diog-enites and eucrites. This postulate has been supported by McCarthy et al. (1972) and later re-searchers, most recently by Dreibus et al. (1977:211),


who state: "The hypothesis of Jerome and Goles(1971) and McCarthy et al. (1972), according towhich howardites are thought to be mechanicalmixtures 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 byMason (1963) and Fredriksson et al. (1976). Cal-cium-poor pyroxene compositions in eucrites (Fig-ure 4) range from Fs 33 to Fs 68. Therefore, ifhowardites are mixtures of diogenites and eucrites,their pyroxenes should comprise a large componentwith composition Fs 26±1 (the diogenite fraction)and additional calcium-poor "pyroxenes possiblyranging from Fs 33 to Fs 68 (the eucrite fraction).That this is not the case was already evident in 1971from the work of Fredriksson and Keil (1963) onKapoeta; figures 7 and 8 of their paper showpyroxene compositions extending as low as Fs 14and no large component at Fs 26. These results onKapoeta have been confirmed and extended byDymek et al. (1976). Similar evidence is availablefrom Macibini (Reid, 1974) and Malvern (Simp-son, 1975); both Reid and Simpson conclude thatthese meteorites could not have been produced bythe physical mixing of diogenite and eucrite (Des-noyers and Jerome (1977), however, disagree withSimpson in the case of Malvern). Our data on thehowardites confirm and extend the conclusions ofReid and Simpson. Not only do the howardites con-tain calcium-poor pyroxenes much more magnesianthan those in diogenites, but some of them (nota-bly Pavlovka and Nobleborough) have pyroxenesmore 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)—alsoargues for a component not represented either inthe diogenites or the eucrites. We agree, neverthe-less, with Reid's comment (1974:398) on Macibini:"The parent materials [of the howardites] are nota random set of unrelated rocks but appear to bederived predominantly from a suite of related sam-ples, that may have differentiated from a singlesource."

The Eucrite Parent Body

Discussions of the eucrite parent body usuallyimply the parent body of what we have called the

main-group eucrites, i.e., the compositional clusteraround FeO/(FeO + MgO) molecular percentage of60. Within this group a wide variety of primary(i.e., pre-brecciation) textures are seen, indicating aconsiderable variety of crystallization conditions inthe parent body. Nevertheless, most, if not all,crystallized under low pressure (the common oc-currence of tridymite indicates a confining pressureof 3 kbar or less).

Although the textures of the clasts in most of theeucrites resemble those of the lunar basalts, the ex-solution and inversion features show significantdifferences. The pyroxenes in lunar basalts arestrongly zoned, and can range from orthopyroxenethrough pigeonite to augite and hedenbergite in asingle crystal. In most eucrites the pyroxenes areessentially unzoned, and consist of pigeonite of uni-form composition with exsolution lamellae ofaugite. Possibly the eucrites crystallized originallylike the lunar basalts, as thin surface flows, but laterwere annealed (perhaps by burial under successiveflows), whereby originally zoned pyroxenes werehomogenized, and then exsolved augite on slowcooling. If true, it is noteworthy that this equilibra-tion occurred without textural change, except forIbitira and Emmaville, which have a hornfelsictexture. Presumably the absence of water and othervolatiles 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 asteroidconsisting of a pallasitic core and a diogeniticmantle. The radius of the hypothetical asteroid wasset at 300 km, to be consistent with the measuredcooling rates of pallasite meteorites (Buseck andGoldstein, 1967). It had a pallasitic core with radius207 km, a diogenitic mantle 80 km thick, and aeucritic crust 13 km thick. The overall mineralogicalcomposition of this asteroid was calculated to be(weight percent): nickel-iron, 14; olivine, 25; hyper-sthene, 46; pigeonite, 9; plagioclase, 6. The bulkcomposition of this asteroid was then calculatedfrom the average mineral compositions in thesemeteorites, with the results as given in Table 2. Itis gratifying to see that recent deductions as to thecomposition of the eucrite parent body, based onsophisticated geochemical arguments (e.g., Morganet al., 1978) are in good agreement with the simplepetrological model. The composition derived byMorgan 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 pyroxenethan that of Mason, i.e., one with a larger pallasitecore and smaller diogenite mantle.

Literature Cited

Allegre, C. J., N. Shimuzu, and M. Treuil1977. Comparative Chemical History of the Earth, the

Moon and Parent Body of Achondite. PhilosophicalTransactions of the Royal Society of London, seriesA, 285:55-67.

Anders, E.1977. Chemical Compositions of the Moon, Earth, and

Eucrite Parent Body. Philosophical Transactions ofthe 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 theSixth Lunar Science Conference, 1:469-492.

Bunch, T. E., S. Chang, J. M. Neil, and A. Burlingame1976. Unique Characteristics of the Jodzie Howardite.

Meteoritics, 11:260-261.Buseck, P. R., and J. I. Goldstein

1967. Olivine Compositions and Cooling Rate of PallasiticMeteorites. 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. Jerome1973. 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. Wanke1977. The Bulk Composition of the Moon and the Eucrite

Parent Body. Proceedings of the Eighth LunarScience 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 NaturalMaterials, pages 613-621. Baltimore, Md.: MonoBook Corporation.

Duke, M. B., and L. T. Silver1967. Petrology of Eucrites, Howardites, and Mesoside-

rites. Geochimica et Cosmochimica Acta, 31:1637-1665.

Dyakonova, M. I., and V. Y. Kharitonova1961. 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 theKapoeta Howardite and Petrologic Constraints onthe Evolution of Its Parent Body. Geochimica etCosmochimica Acta, 40:1115-1130.

Easton, A. J., and C. J. Elliott1977. Analyses of Some Meteorites from the British Mu-

seum (Natural History) Collection. Meteoritics, 12:372-374.

Fredriksson, K., and K. Keil1963. The Light-Dark Structure in the Pantar and

Kapoeta Stone Meteorites. Geochimica et Cosmo-chimica Acta, 27:717-740.

Fredriksson, K., and F. Kraut1967. 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 EightHypersthene Achondrites. Meteoritics, 11:278-280.

Frost, M. J.1971. The Molteno Meteorite. Mineralogical Magazine,

38:89-93.


Fukuoka, T., M. S. Ma, and R. A. Schmitt1977. 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 Studywith Comment on Its Primordial Environment.American Mineralogist, 34:494-507.

Hutchison, R., A.W.R. Bevan, and J. M. Hall1977. 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. Takeda1974. 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 andE. Steinnes, editors, Activation Analysis in Geo-chemistry and Cosmochemistry, pages 261-266. Oslo,Norway: Universitetsforlaget.

Jerome, D. Y., and M. Christophe Michel-Levy1971. Quelques aspects remarquables de la meteorite

achondritique de Malvern (Afrique du Sud). Bulletinde 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 meteoritesfeldspathiques en general. Archives du Museumd'Histoire naturelle, series 6, 11:15-58.

Lange, D. E., and K. Keil1976. Meteorites of Northeastern New Mexico. New

Mexico Geological Society Guidebook, 27th FieldConference, 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 Mesosideriteand Its Achondritic Enclaves. Mineralogical Maga-zine, 35:1029-1060.

McCarthy, T. S., L. H. Ahrens, and A. J. Erlank1972. Further Evidence in Support of the Mixing Model

for Howardite Origin. Earth and Planetary ScienceLetters, 15:86-93.

McCarthy, T. S., A. J. Erlank, and J. P. Willis1973. New Chemical Analyses of Six Achondrites and One

Chondrite. Meteoritics, 9:215-221.Marvin, U. B.

1976. A Unique Eucritic Gabbro from the DescartesHighlands. 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 etCosmochimica 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 Compositionof Lunar Samples 15085, 15256, 15271, 15471, 15475,15535, 15555, and 15556. Proceedings of the ThirdLunar Science Conference, 1:785-796.

Mason, B., and H. B. Wiik1966a. 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. Hertogen1978. A "Chondritic" Eucrite Parent Body: Inference

from Trace Elements. Geochimica et CosmochimicaActa, 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 theOrigin of Basaltic Achondrites. Meteoritics, 9:398-399.

Schmitt, R. A., G. G. Goles, R. H. Smith, and T. W. Osborn1972. Elemental Abundances in Stone Meteorites. Meteo-

ritics, 7:131-124.Sibray, S., K. Keil, P. M. Starzyk, R. A. Schmitt, and G. I.

Huss1976. The Melrose (b) Howardite: A Complex Meteorite

Parent Body Regolith Breccia. Transactions of theAmerican 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 Comparisonwith Other Eucrites and Lunar Samples. Earth andPlanetary 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. Ishii1975. 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 LunarMare Pyroxenes. Meteoritics, 11:372-374.

Takeda, H., M. Miyamoto, T. Ishii, and A. M. Reid1976a. Characterization of Crust Formation on a Parent

Body of Achondrites and the Moon by PyroxeneCrystallography and Chemistry. Proceedings of theSeventh 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. Craig1953. 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. Thacker1974. Chemistry of Apollo 16 and 17 Samples: Bulk Com-

position, Late Stage Accumulation and Early Dif-ferentiation of the Moon. Proceedings of the FifthLunar 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, aRe-evaluation. Proceedings of the Eighth LunarScience 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. Anders1975. 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 ElectronMicroprobe 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 anitrogen atmosphere. The fusion times ranged from5 to 30 sec at temperatures between 1650° C and1750° C. The fusion bridge was powered by a1.5 kVA welding transformer.

This fusion technique in combination with theelectron microprobe uses very small amounts ofsample and is suitable for very rapid analysis ofrock samples ranging in composition from 45 toabout 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 betterthan 5 percent (relative) for all major componentsexcept TiO2, which is better than 13 percent (rela-tive). A major part of the error in the TiO2 analysesis probably due to the presence of minute residualor quench Fe-Ti oxides in the glass.

The samples investigated suggest that if fusiontimes of less than 20 sec are used, Na2O loss is pre-vented or greatly restricted and homogeneous glassesstill result.

Introduction

Brown (1977) summarized techniques used inelectron microprobe major elements analysis ofwhole rock samples. He also described a new tech-nique for fusing rock powders in molybdenumboats in a pressurized (60 psi) argon atmosphere.

Peter A. Jezek, John M. Sinton, Eugene Jarosewich, andCharles 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 andmineral powders in molybdenum and tungstenboats in a nitrogen atmosphere (at atmosphericpressure), using a strip furnace which is easily builtand inexpensive to operate. These experiments wereaimed at developing a rapid, simple, and inex-pensive method of fusion of rock powders intorepresentative, homogeneous beads or shards suit-able for analysis by electron microprobe, withoutthe need of an argon pressure chamber.

INSTRUMENTATION.—The fusion bridge (Figures1, 2) is powered by a 1.5 kVA welding transformer.The primary of the transformer is connectedthrough a Variac autotransformer to 115 VAC. Thesecondary of the transformer, rated at 11.5 V, isconnected to a 100 A silicon diode bridge rectifier.The DC output of the rectifier then powers thefusion bridge (Figure 3). To reach a temperatureof 1670° C the strip receives 40 A at 4 VDC. Tem-peratures were measured by an optical pyrometer.The particular welding transformer employed inthis 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 allmolten glasses, and are especially useful in restrict-ing the how of low viscosity basaltic glasses. Thesize 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 iscontinuously flushed by nitrogen (Figure 1). Afterplacing the powder, ground finer than 100 mesh,

46

NUMBER 22 47

FIGURE 1.—Front view of the fusion furnace: the fusion boatis mounted between the two posts and the whole assembly iscovered by a glass bell jar that is continuously flushed bynitrogen (for symbol explanation see Figure 3).

FIGURE 2.—Detail of the fusion bridge (P = positive andnegative terminals of the bridge; B = fusion boat; Q = out-let of the nitrogen quench line; O = outlets through whichthe gas under the glass bell jar is forced out; S = verticalshield behind which is located the outlet for the purge linethrough which nitrogen enters the glass bell jar during theflushing cycle (see Figure 3).

in the fusion boat and covering the fusion assemblyby the bell jar it takes about 1 min to establish anatmosphere of almost pure nitrogen. This is docu-mented by the lack of oxidation of a Mo stripheated in this atmosphere at 1750° C for 25 min.

Two melting procedures were used. (1) Sampleswere heated for about 10 sec at low temperature(dark red glow of the strip) to drive off adsorbedwater 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 secbut total heating at the final temperature for about10-30 sec promotes diffusion and thus formation ofmore homogeneous glasses. (2) Rapid heating of thecold powder to about 1700° C was also investigated.When subjected to this technique the powders fusemore 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 homogeneityof the glasses produced by these two techniques.

Quenching of the melts is accomplished by simul-taneously shutting off the power to the fusionbridge and directing a stream of nitrogen onto thebottom of the fusion boat. Cooling is sufficientlyfast to prevent crystallite formation in all but themost mafic samples. Quench olivine is common inthe more mafic samples.

In order to avoid either strip-contaminated glass,glass possibly depleted by elemental diffusion intothe strip (Brown, 1977), or elemental volatilizationinto the nitrogen atmosphere, only glass in the in-terior of the bead or shard was analyzed.

The glasses were analyzed by a 9-spectrometercomputer-automated ARL-SEMQ electron micro-probe. The spectrometers were equipped with thefollowing analyzing crystals. Si-EDDT, Al-EDDT,

48

F1 T1

QUENCH PURGE


R1T2

NITROGEN

FIGURE 3.—Wiring diagram of the fusion furnace (SI = DPDT115-VAC 10 A switch; Fl = 10 A fuse; Tl - VARIAC auto-transformer; T2 = welding transformer; BR1 = 100 A 50 PIVsilicon diode bridge rectifier; Rl = meter shunt, 150 A 50 mV;Ml = 50 mV full scale meter calibrated to read 150 A at fullscale). Lower left: a diagram of the nitrogen system (the purgeline is located behind the vertical shield (S in Figure 2); thequench 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). Duringthe analysis 15 kV accelerating potential and 0.02fiA sample current were used. A defocused beam,20-50 pjm in diameter, practically eliminates Navolatilization during analysis and also helps to aver-age local, small inhomogeneities which may existin the glass. It is recommended that 10 sec counttimes be used and the analysis repeated at least tentimes to accumulate a statistically significant bodyof data. The repeated measurements also help toaverage possible larger scale inhomogeneities.

Results

Eleven rock powders and three mineral sampleswere fused by the described technique. The rockswere selected to cover as wide a compositional range

as possible. Analytical data for the eleven rocksamples are presented in Table 1. The electronmicroprobe data show good agreement with the wetchemical values, which were recalculated volatilefree in order to facilitate direct comparison withthe 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 instrumentalinstability. The standards used in the glass analyseswere analyzed by wet chemical techniques at theDepartment of Mineral Sciences, Smithsonian In-stitution. Their compositions are given in Jarose-wich, Nelen, and Norberg (this volume).

The average differences between the wet chemicalanalyses and the electron microprobe analyses areshown in Table 2. The average accuracy in weightpercent is better than 0.2 for all major componentsexcept SiO2 which is accurate to better than 0.3. In

TABLE 1.—Analyses of 11 fused rock powder samples by wet chemical and electron microprobemethods (A = wet chemical analysis (wt %); B = electron microprobe analysis (wt %), numbersin parentheses are the numbers of point analyses averaged; CV (approximate measure of glasshomogeneity) = (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 wetchemical analyses by E. Jarosewich and J. Norberg, Department of Mineral Sciences, SmithsonianInstitution; all wet chemical analyses recalculated as volatile free to facilitate direct comparisonwith 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



Consti tuent

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) betweenthe wet chemical and electron microprobe analyses presented in Table 1

Constituent

sio2

TiO2

A12O3

FeO

MgO

CaO

Na2O

K20

Range ofSamples

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

AverageAbsolute

Difference(wt %)

0.21

0.10

0.14

0.16

0.12

0.09

0.13

0.05

RangeAbsolute

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

AverageRelative

Difference

a)0 .4

12.7

0 .9

2.2

2 .8

1.9

4 . 5

4 . 5

RangeRelative

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 than5 for all major components except TiO2 which isbetter than 13. The somewhat poorer accuracy inmeasuring TiO2 is due in part to some very lowmeasured 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, themajor part of the error is probably due to the pres-ence of residual or quench Fe-Ti oxides observed inthe form of opaque spots in some glasses. Due totheir very small size a positive identification isdifficult.

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 than100 mesh powders were used. Subsequent testingshowed 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 fusionof high-silica samples impractical. The MgO-rich,low-silica samples could not be prevented fromgrowing skeletal olivine crystals during quenching.

Three mineral samples (Table 3) and a sampleof natural rhyolite glass (VG-568, Table 1) werefused to evaluate: (1) Na^O loss, (2) the behaviorof hydrous minerals during fusion, and (3) the im-pact of rapid dehydration on the composition of theresulting glass.

It is clear that prolonged fusion of the samplesresults in Na2O loss. But fusion times ranging up to20 sec produce good glasses with minimal Na^O loss.The three fusion times used in the fusion of theLake County plagioclase sample (Table 3) did notproduce significantly large differences in the Na2O

TABLE 3.—Comparison of analyses of crystalline minerals with analyses of the same mineralsfused for various times (A = electron microprobe analyses of the unfused minerals; B = electronmicroprobe analyses of fused minerals; numbers in parentheses are the numbers of point analysesaveraged; C = wet chemical analyses of crystalline mineral recalculated water free, based ondata 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

Plagioclase115900

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


concentration. However, the 25 sec fusion probablyrecords a small Na2O loss, and the 30 sec fusion ofomphacite has clearly resulted in a significant Na2Oloss.

Arenal hornblende frothed extensively duringfusion due to the escape of structural water (1.21 wt% H2O). The composition of the resulting glass isvery close to the crystalline mineral compositionrecalculated water free. Although it was anticipatedthat Na2O may be volatilized more rapidly duringH2O release from hydrous minerals, no Na2O losswas observed in three separate glasses, each pro-duced by 5 sec fusion. The hornblende samplewas not fused longer than 5 sec because the waterpresent in the mineral facilitates rapid fusion.

No Na2O was lost during the 10 sec fusion ofVG-568. This sample is a natural glass and thereforeprovides information only on Na2O stability duringfusion.

Summary and Conclusions

The described fusion technique uses very smallamounts of sample and is rapid. The furnace canbe built and operated inexpensively. When used inconjunction with an electron microprobe, majorelement concentrations can be measured rapidlyand quite accurately for rock samples ranging incomposition from basalts to high-silica andesites.More mafic samples have a tendency to crystallizeolivine on cooling. Granites and other high-silica,high-alkali samples produce glasses of high viscositypreventing 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 suitablyaccurate to allow routine use of the technique. Theaverage accuracy of better than 5% relative for allmajor components except TiO2 (better than 13%)is quite acceptable.

The rapidity and convenience with which thesamples can be prepared and analyzed allowed ma-jor element reconnaissance analyses of rock suitespreviously not analyzed due to time and cost factorsinvolved. The rock powders and the three mineralsamples investigated suggest that if fusion times ofless than 20 sec are used Na2O loss is prevented orgreatly restricted and glasses result that are suffi-ciently homogeneous to be suitable for microprobeanalysis.

Literature Cited

Bence, A. E., and A. L. Albee1968. 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. Geochimicaet 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 naturalglasses and one mineral is reported in this compila-tion. The results obtained by the three participatinglaboratories are in good agreement with chemicalanalysis values (where available) for SiO2, A12O3,FeO, CaO, and K^O (> 2 wt %). One labora-tory shows a considerable bias for Na^O and aslight bias for MgO. Some results for MnO, K2O(/—0.2 wt %), and P2O5, 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 theavailable 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 ofnatural glasses is of great importance in the study ofvolcanic and deep sea rocks. Melson et al. (1977)briefly summarized this need, referring specificallyto the analyses of glasses from the Deep Sea DrillingProject. Since a standardized method for analysisof such glasses is not available, each laboratory hasdeveloped its own techniques and has been using its

Eugene Jarosewich, Department of Mineral Sciences, NationalMuseum of Natural History, Smithsonian Institution, Wash-ington, D.C. 20650. Alan S. Parkes, Department of Earthand 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 resultsobtained 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 analyticalerrors. Since more than one institution may partici-pate in the study of volcanic or deep sea rocks andthe analytical data may be used interchangeably forpetrologic studies, it is of utmost importance thatthe 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 threelaboratories, each using its own standard operatingparameters; these results are compared for precisionand accuracy. The participating laboratories were:Massachusetts Institute of Technology (MIT),Smithsonian Institution (SI), and U.S. GeologicalSurvey (USGS).

The data presented in this paper are meant toserve primarily as a general guide for the expectedaccuracy and precision of microprobe analyses ofglasses; however, they can serve as a guide for suchanalyses 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 workand Dr. J. S. Huebner for revising the data. Wewould 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 grainof Kakanui hornblende and only one grain of each

53


of the following natural glasses:

VG-2 (USNM 111240-52), Basalt with a few rare micro-phenocrysts of olivine and plagioclase in section, lessthan 1%. Chemical analysis, Table 4.

VG-A99 (USNM 113498-1). Basalt from Makaopuhi LavaLake, 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 fromDeSteiguer 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 dredgeD5, 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 threelaboratories for microprobe analysis.

These four glasses were selected because theyrepresent the approximate elemental ranges en-countered in the study of natural basaltic glasses.Classical chemical analyses are available for Kakanuihornblende, VG-2, and VG-A99 for comparison withthe microprobe results (see Table 1, Jarosewich,Nelen, and Norberg, this volume). Although nochemical analyses are available for VG-999 andVG-D08, these two samples were included in thisstudy to obtain additional data on the precisionamong the microprobe analyses themselves. Thesame 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 toobtain analytical values for the samples as they are

obtained under the laboratories' normal operatingconditions, the operating parameters (kv, /iA, beamsize, etc.) were not specified. The samples wereanalyzed 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 referencesample. In this manner discrepancies due to the useof different reference samples were eliminated andany deviations could be ascribed to the instrumen-tal parameters. Since Kakanui hornblende is not anideal reference sample for obtaining accurate anal-yses of these glasses, the accuracy was determined byanalyzing these samples using each laboratory's pre-ferred reference samples and comparing results withthe chemical analyses of these glasses. The analyticalresults were compiled in the following manner. (1)Each sample was analyzed several times usingKakanui hornblende as the reference sample; thenumber of individual analyses and the averages aregiven in Tables 5-7. (2) The samples were analyzedseveral times using each laboratory's preferred ref-erence samples; the number of individual analyses,the averages, and the type of reference samples usedin these analyses are also given in Table 5-7. (3)The uncorrected averages obtained using Kakanuihornblende as the reference sample are summarizedin Table 3 for comparisons of precision; these re-sults indicate the variation of the data due only tothe instrumental variation. (4) Table 4 summarizesthe 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-SEMQ9 spectrometers

ARL-EXM-SM

AcceleratingPotential

(kV)

15

15

15

SampleCurrent(yA)

0o03

0o05

0.05

Beam Size(ym)

4

2, 30

20

Counting Time

30 sec or 60,000 counts

Seven 10 sec counts foreach analysis

20 sec or 20,000 countsfor each element

NUMBER 22 55

TABLE 2.—Analyses of reference samples used by participating laboratories

Reference sample

DJ35P-140AN-60MAC APOrthoclaseDi2TiMnllmCoss

Kak hornblendeVG-2ApatiteFayalite

RhodoniteGarnet-PyropeOrthoclaseDi2TiDiopside-Jadeite

SiO2

56.8840.8653.050.0564.3954.36

40.93

40.3750.810.3429.22

46.7641.4563.4254.3956.10

A12°3

8.82

30.01

18.58

14.9014.060.07

0.9623.5019.24

3.78

FeO

7.23

0.03

45.4140.99

10.9211.830.05

67.54

12.4910.760.10

MgO

12.1051.63

0.01

18.260.25

12.806.71

"0.4218.80

18.2615.84

CaO

MET

16.83

12.3838.82

25.38

SI

10.3011.1254.02

USGS

5.625.090.0825.3822.02

Na2O

5.36

4.560.181.14

6.93

2.602.620.23

0.36

2.31

K

14

20

15

2°

.92

.05

.19

.34

TiO2

2.0051.618.71

4.361.85

0.04

0.51

2.00

P

17

040

0

2°5

.86

.20

.78

.49

MnO

1.50

0.090.22

2.14

33.340.33

these results with the chemical values indicate theaccuracy of the microprobe results.

Discussion

Comparisons for precision of the results obtainedusing Kakanui hornblende as the reference sample(Table 3) indicate that careful microprobe analysesfor all major elements, titanium, and high potas-sium (>2 wt %) produce relatively precise results.The results also indicate that minor elements andsodium vary considerably and more careful workmust be done if the results for these elements are tobe used for exacting petrological work. There is nopronounced bias in the results by any of the threelaboratories for the major elements except for aslight bias in the magnesium results. There is abias in the results for sodium, low potassium(~0.2 wt %), and phosphorus. The sodium andphosphorus results analyzed by laboratory 1 runconsistently 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 usingthe 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 preferredreference samples were used (Table 4) also checkfavorably for the major elements, titanium, andpotassium (> 2 wt %); more important, these re-sults are in excellent agreement with the values forthe three samples analyzed by classical chemicalmethods, thus showing a high level of accuracy. Theresults for sodium show a much larger discrepancythan those given in Table 3; only the lowest valuesof the microprobe analyses check well with thechemically analyzed sodium values. Sodium fre-quently presents a problem in microprobe analysisand 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 somesamples (Table 6, results in brackets), the lattergives acceptable values. Kakanui hornblende so-dium results are not affected by the beam size. Oneof the possible explanations for the higher resultsof laboratory 1 and lower results of laboratory 2with the small beam is the unique behavior ofsodium in the electron beam for both the referencesamples and the unknown samples. If for examplea reference sample in which sodium is easily "vola-tilized" (accepted description of decreasing inten-


TABLE 3.—Summary of uncorrected data from microprobe analyses of the 5 samples, usingKakanui hornblende as the reference sample, for interlaboratory comparison of precision (figurein parentheses indicates number of analyses averaged; each analysis represents up to 10 individualpoint analyses; laboratory 1 = MIT, 2 = SI, 3 = USGS; data based on Tables 5-7)

Const i tuent

SiO2

A 12°3

FeO

MgO

CaO

Na2O

K20

TiO 2

P2°5

MnO

Lab

123

123

123

123

123

123

123

123

123

123

KH

40.4540.5840.31

14.8814.9714.98

10.8510.9110.81

12.8312.6712.92

10.2010.3910.24

2.742.592.67

2.022.051.96

4.364.384.27

0.090.060.08

0.070.090.19

(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

51.0751.6251.05

14.6915.0614.85

11.9611.8411.78

7.026.906.87

11.1211.1611.01

2.902.652.78

0.180.200.06

1.661.671.64

0.240.200.20

0.160.180.25

(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

52.4952.4552.54

13.2313.5413.13

13.5713.6013.46

5.294.984.99

9.389.298.80

2.912.712.71

0.770.830.72

3.653.703.77

0.440.390.42

0.160.180.22

(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

52.2752.2151.88

13.6114.1214.41

13.7913.7113.66

5.985.815.57

10.3910.4010.26

2.892.602.66

0.140.170.03

1.721.741.66

0.220.190.17

0.170.200.13

(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

50.1750.5650.36

16.6017.2716.71

9.008.978.88

9.368.718.70

12.4812.3912.13

2.772.382.40

0.060.080.00

0.981.010.92

0.160.120.12

0.160.140.16

(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 anunknown sample which does not show this vola-tilization effect, the sodium results for the unknownsample will be high. If on the other hand the so-dium in the reference sample is "stable" and sodiumin the unknown sample is easily volatilized, low re-sults will be obtained. The use of a larger beam sizemay 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 by15%-20% in most cases, giving much higher valuesthan those of the chemical analyses. These higherphosphorus corrections are evident with both theup-dated Bence-Albee (Albee and Ray, 1970) andMagic IV correction procedures. Both correctionprocedures 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 eachlaboratory's preferred reference samples, along with chemical analyses of 3 of the samples, forinterlaboratory comparison of precision and accuracy (figures in parentheses indicates number ofanalyses averaged; each analysis represents up to 10 individual point analyses; laboratory1 = MIT, 2 = SI, 3 = USGS; Chem. = chemical analysis from Table 1, Jarosewich, Nelen, andNorberg, this volume; microprobe data based on Tables 5-7)

Consti tuent Lab

SiO9Z

Al90oZ j

FeO

MgO

CaO

Nao0z

K00z

TiOoz

P90

MnO

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

123

(Chem.)

KH

40.6540.7240.4240.37

14.6114.9514.7614.90

10.4510.9310.5410.92

12.9512.7013.0312.80

9.9910.4710.1210.30

2.802.602.722.60

2.102.061.942.05

4.874.404.654.36

0.140.070.040.00

0.100.110.060.09

(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

50.8550.7250.7550.81

13.8114.1513.9814.06

11.2611.7911.7911.83

7.016.787.026.71

10.8511.1410.7211.12

3.172.662.752.62

0.200.210.180.19

1.861.911.861.85

0.320.230.190.20

0.220.220.230.22

(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

51.0551.2250.8050.90

12.5912.6612.8012.97

13.2413.4713.4113.18

5.244.955.165.18

9.089.288.979.38

2.812.702.732.73

0.820.900.760.80

4.044.053.774.06

0.540.460.310.41

0.190.220.190.19

(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

51.1651.4150.70

13.0613.3613.54

12.9413.6813.30

5.995.776.13

10.1210.3610.18

3.062.612.57

0.140.190.04

1.931.961.80

0.290.220.15

0.290.250,20

(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

50.2950.1849.77

15.8016.1715.44

8.678.968.78

8.888.468.78

12.2312.4112.05

2.792.252.39

0.070.090.09

1.091.141.03

0.230.150.08

0.180.170.13

(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)


TABLE 5.—MIT laboratory individual analyses of Kakanui hornblende and 4 natural glassesusing Kakanui hornblende as reference sample and also the laboratory's preferred referencesamples

Run No.

1

2

4

5a

5b

6

8

1

Conditions

Standards

UncorrectedB-A Corr.Magic IV Corr.a (5)


Standards

UncorrectedB-A Corr.Magic IV Corr.o (10)



Standards


Standards


Standards


SiO2

KH

40.3940.3540.350.17

40,5140.5540.560.53

DJ35

38.2140.2940.360.34

38.9240.9741.070.21

38.7140.7540.850.32

DJ35

38.3240.4340.520.30

DJ35

38.7140.8240.920.19

KH

51.3649.9850.080.51

A12O3

KH

14.6914.6814.680.11

15.0715.0715.080.08

AN-60

12.6114.4514.540.08

12.6614.4814.570.08

12.6814.5014.590.08

AN-60

12.8314.7114.800.09

AN-60

13.0014.8914.970.09

KH

14.5114.0314.030.13

FeO

KH

10.9910.9910.990.15

10.7010.7110.700.22

P-140

10.4810.3710.430.27

10.3410.2310.290.12

10.2910.1810.240.13

Coss.

10.2610.6810.600.12

Cost..

10.3510.. 7710.700.18

KH

12.1512.1512.150.14

MgO

Kakanui

KH

12.5712.5912.590.13

13.0913.0813.070.22

P-140

11.6812.9813.040.26

11.5812.8612.880.13

11.5112.7812.820.17

DJ35

12.3813.1013.180.10

DJ35

12.3413.0513.140.15

KH

6.836.896.880.08

CaO Na2O

hornblende

KH

10.1310.1310.130.15

10.2710.2810.280.07

DJ35

10.119.929.960.14

10.2010.0210.050.18

10.189.9910.030.13

DJ35

10.189.98

10.020.07

DJ35

10.2310.0410.090.10

VG-2

KH

11.0311.0311.040.08

KH

2.712.712.710.12

2.772.762.750.12

DJ35

2.552.872.900.09

2.552.862.890.11

2.592.912.930.08

DJ35

2.492.802.830.11

DJ35

2.292.572.600.07

KH

2.822.842.840.11

K20

KH

2.032.032.040.02

2.002.002.000.04

Ortho

2.102.052.050.02

2.162.112.100.04

2.152.102.090.04

Ortho

2.162.112.100.05

Ortho

2.162.112.100.05

KH

0.190.200.190.00

TiO2

KH

4.344.344.350.03

4.374.384.370.03

Di2Ti

5.134.944.940.08

5.054.874.870.05

5.114.924.930.05

Di2Ti

5.034.844.850.05

Mnllm

4.204.794.730.05

KH

1.671.671.670.02

p2o5

AP

0.100.120.120.00

0.080.100.120.00

MAC AP

0.100.120.120.00

0.110.130.150.00

0.110.150.150.00

MAC AP

0.110.150.150.02

MAC AP

0.130.160,150.00

AP

0.230.300.300.02

MnO

KH

0.040.040.050.01

0.100.100.100.03

Mnllm

0.100.110.110.03

0.120.130.120.01

0.110.120.120.03

Mnllm

0.040.040.040.04

Mnllm

0.110.110.110.03

KH

0.140.140.140.04

NUMBER 22 59

Table 5.—Continued

Run No.

2

5

2

4

5

6

8

Conditions


Standards


Standards


Standards



Standards


Standards


SiO2

50.7749.5249.610.41

DJ35

49.6150.8551.050.83

KH

52.4950.6850.840.43

DJ35

50.2251.0351.300.28

50.2551.0051.420.49

DJ35

50.3251.1751.430.32

DJ35

50.0950.9951.260.53

A12O3

14.8714.3914.390.19

AN-60

12.4913.8113.880.13

KH

13.2312.7812.780.15

AN-60

11.1812.3612.440.09

11.2012.3412.440.08

AN-60

11.5012.7112.780.06

AN-60

11.7312.9513.030.11

FeO

11.7611.7711.760.18

P-140

11.3911.2611.330.14

KH

13.5713.5313.550.09

P-140

13.2513.0613.160.19

12.8512.6812.770.18

Coss.

13.1113.6113.530.19

Coss.

13.0913.5913.500.14

MgO

7.217.267.250.10

P-140

6.277.017.020.06

KH

5.295.425.410.08

P-140

4.615.245.260.10

4.545.165.170.07

DJ35

4.885.275.310.03

DJ35

4.885.275.300.10

CaO

11.2011.2011.210.07

DJ35

11.0510.8510.910.08

VG-A99

KH

9.389.369.370.10

DJ35

9.118.918.960.06

9.249.049.100.08

DJ35

9.299.089.140.01

DJ35

9.519.309.360.11

Na2O

2.972.992.990.12

DJ35

2.803.173.190.11

KH

2.912.992.990.08

DJ35

2.633.053.080.08

2.552.952.980.15

DJ35

2.242.612.630.11

DJ35

2.242.612.630.12

K20

0.170.170.170.01

Ortho

0.200.200.200.00

KH

0.770.770.770.02

Ortho

0.840.820.820.01

0.840.830.820.02

Ortho

0.830.810.810.02

Ortho

0.850.830.830.02

TiO2

1.651.661.650.03

Di2Ti

1.931.861.860.07

KH

3.653.643.640.03

Di2Ti

4.304.124.120.05

4.284.104.110.05

Di2Ti

4.274.084.090.05

Mnllm

3.413.873.830.07

P2°5

0.240.310.330.00

MAC AP

0.250.320.330.02

AP

0.440.560.560.02

MAC AP

0.420.540.530.02

0.410.530.530.02

MAC AP

0.430.550.560.00

MAC AP

0.420.540.530.02

MnO

0.180.180.180.04

Mnllm

0.200.220.220.03

KH

0.160.160.150.03

Mnllm

0.190.200.200.03

0.170.180.180.03

Mnllm

0.180.200.190.01

Mnllm

0.160.180.180.03



Run No.

1

2

5

2

5

Conditions

Standards



Standards


Standards


Standards


SiO2

KH

52.2050.6450.790.34

52.3450.7750.910.68

DJ35

50.0751.1651.420.45

KH

50.1749.2549.300.43

TXT35

48.6850.2950.430.60

A12O3

KH

13.6613.2413.240.17

13.5613.1513.160.42

AN-60

11.8013.0613.160.08

KH

16.6016.0216.060.55

AN-60

14.3515.8015.860.06

FeO

KH

14.0313.9914.010.20

13.5513.5213.530.17

P-140

13.1112.9413.030.17

KH

9.009.049.040.09

P-140

8.738.678.720.10

MgO

KH

5.855.975.980.05

6.116.246.240.07

P-140

5.305.996.020.07

KH

9.369.209.200.18

P-140

8.138.888.880.12

CaO

VG-999

KH

10.2310.2110.240.07

10.5410.5210.540.08

DJ35

10.3310.1210.170.06

VG-D08

KH

12.4812.5312.540.08

DJ35

12.4112.2312.290.06

Na2O

KH

2.752.822.820.09

3.023.093.080.09

DJ35

2.663.063.090.20

KH

2.772.702.700.09

DJ35

2.552.792.810.12

K2°

KH

0.150.150.160.00

0.130.130.130.00

Ortho

0.140.140.140.01

KH

0.060.070.060.01

Ortho

0.070.070.070.00

TiO2

KH

1.701.691.700.03

1.731.731.730.03

Di2Ti

2.011.931.940.03

KH

0.980.990.990.03

Di2Ti

1.121.091.080.03

*2°5

AP

0.210.270.270.02

0.230.300.300.00

MAC AP

0.220.290.300.02

AP

0.160.210.210.00

MAC AP

0.180.230.240.02

MnO

KH

0.130.130.130.01

0.200.200.210.03

Mnllm

0.270.290.290.03

KH

0.160.160.170.03

Mnllm

0.170.180.180.03

to 40 times to obtain good precision, it is reasonableto assume that, on the basis of the data in Table 3,the precision of microprobe analyses of basalticglasses cannot be much improved without moreelaborate data acquisition techniques. The standarddeviations for various elements given in Tables 5-7are of similar magnitudes for the three laboratoriesand they realistically indicate the expected scatterof analytical results. It is obvious that longer countswould not improve the precision very much and,indeed, extended counting times are impractical inday-to-day operation.

The accuracy, on the other hand, can be in somecases improved by the selection of suitable referencesamples for a given unknown sample. From the datain Tables 5-7 it is evident that the microprobe re-sults for some elements are in better agreementwith the chemical results when reference samplesother than Kakanui hornblende are used. Thestandard deviation here again is similar to that ofthe precision and thus could also be used as anindication of the expected scatter of the results.Analysis of elements below 0.1 wt % should be donewith extra care and accepted with caution.

NUMBER 22 61

TABLE 6.—SI laboratory individual analyses of Kakanui hornblende and 4 natural glasses usingKakanui 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

UncorrectedB-A Corr.a (7)


UncorrectedB-A Corr.o (7)


Standards





Standards




SiO2

KH

41.0941.090.49

40.4540.460.28

40.0340.130.38

40.7640.760.23

KH

41.0941.090.49

40.4540.460.28

40.4440.510.17

40.8340.830.37

KH

52.0451.180.42

51.3250.490.73

50.8950.190.48

A12O3

KH

15.0515.020.23

14.8314.820.24

14.9514.980.21

15.0615.020.28

KH

15.0515.020.23

14.8314.820.24

15.0015.020.20

14.9514.920.22

KH

15.1014.550.18

14.9014.380.16

15.1214.640.19

FeO

KH

10.8710.880.26

11.1111.110.27

10.8810.880.12

10.7610.770.08

KH

10.8710.880.26

11.1111.110.27

10.9510.960.14

10.7610.760.07

KH

12.0112.030.10

12.0112.030.07

11.8211.830.11

MgO

Kakanui

KH

12.7812.750.43

12.6012.600.38

12.7912.830.26

12.5112.490.21

KH

12.7812.750.43

12.6012.600.38

12.7712.820.17

12.6312.610.25

KH

6.856.870.19

7.007.030.09

6.916.980.21

CaO Na2O

hornblende

KH

10.7010.720.26

10.2210.240.04

10.2210.230.07

10.4010.420.14

KH

10.7010.720.26

10.2210.240.09

10.2810.300.11

10.5710.600.14

VG-2

KH

11.2211.320.13

11.1911.290.14

10.9611.060.17

KH

T.51"2.580.08

2.602.59_0.0_7.

2.602.600.02

2.582.580.01

KH

T.58"2.580.08

2.602.590_.0_7_

2.622.620.04

2.582.570.03

KH

F.532.530.11

2,532.53.0.06

2.692.690.03

K20

KH

*1.981.980.08

2.132.130.07

2.042.040.05

2.052.050.06

KH

1.981.980.08

2.132.130.07

2.032.030.04

2.082.080.06

KH

0.190.190.02

0.200.200.02

0.210.210.02

TiO2

KH

4.334.340.12

4.454.450.11

4.434.430.04

4.294.300.07

KH

4.334.340.12

4.454.450.11

4.514.510.09

4.304.300.06

KH

L.671.680.07

1.711.720.01

1.691.700.03

P2°5

AP

0.050.070.02

0.060.070.02

AP

0.050.070.02

0.060.070.02

APi

0.200.250.03

0.190.240.03

MnO

KH

0.090.090.01

0.090.090.01

Fay

0.110.120.03

0.100.100.02

KH

0.170.170.02



Run No.

4

1

2

3

4

1

2

3

4

1

2

Conditions


Standards





Standards





Standards


UncorrectedB-A Corr.° (7)

SiO2

52.2151.320.43

VG-2

50.5650.570.48

50.6950.710.51

50.9350.990.59

50.5250.600.61

KH

52.5651.390.45

52.7651.530.29

51.6550.630.66

52.8251.630.58

VG-2

51.0850.780.75

51.7151.380.28

A12O3

15.1114.550.19

VG-2

14.1414.130.25

14.0814.080.26

14.1514.160.19

14.2114.230.24

KH

13.7213.190.12

13.5112.990.22

13.4713.030.27

13.4612.960.24

VG-2

12.6812.650.23

12.6512.620.20

FeO

11.5311.550.09

VG-2

11.7211.730.15

11.5411.560.09

11.8711.880.19

11.9912.000.11

KH

13.9013.880.08

13.5913.570.06

13.6213.590.12

13.3013.290.17

VG-2

13.4813.450.08

13.4213.390.26

MgO

6.836.850.12

VG-2

6.576.570.11

7.016.990.15

6.696.700.11

6.836.840.08

KH

4.894.970.20

4.985.060.11

5.075.190.13

4.985.060.10

VG-2

4.784.830.06

4.955.000.44

CaO

11.2811.390.14

VG-2

11.1011.110.09

11.1811.190.13

11.1311.130.16

11.1311.130.18

VG-A99

KH

9.309.360.08

9.349.380.18

9.079.100.17

9.469.500.08

VG-2

9.189.130.04

9.369.320.45

Na2O

2.602.590.03

VG-2

T.622.620.11

2.472.46JD.21

2.642.650.06

2.662.660.05

KH

T.3T2.370.10

2.222.26

2.722.750.03

2.702.750.03

VG-2

2~.4T2.470.17

2.272.320.10

K20

0.200.200.03

VG-2

0.220.210.02

0.220.210.02

0.230.220.07

0.220.210.03

KH

0.840.840.12

0.820.830.04

0.800.810.02

0.840.850.02

VG-2

0.930.900.06

0.910.880.04

TiO2

1.621.630.05

VG-2

1.941.940.06

1.921.930.05

1.921.920.03

1.851.850.07

KH

3.663.660.06

3.683.680.03

3.813.810.10

3.653.660.06

VG-2

4.054.020.09

4.124.090.11

p 2o 5

AP

0.190.240.02

0.180.220.03

AP

0.370.450.04

0.410.510.04

AP

0.380.470.04

0.370.450.01

MnO

0.180.180.02

Fay

0.200.220.02

0.190.220.02

KH

0.180.180.02

0.180.180.02

Fay

NUMBER 22


Run No.

3

4

1

2

3

4

1

2

3

4

Conditions

UncorrectedB-A Corr.CT (7)


Standards





Standards





SiO2

51.7751.510.90

51.4651.210.62

KH

52.4651.470.70

52.1751.200.77

51.8951.020.61

52.3151.350.44

VG-2

51.6451.520.58

51.7751.670.64

51.4651.420.38

51.0651.040.41

A12O3

12.7412.720.22

12.6412.650.13

KH

14.2213.690.29

14.0413.540.18

14.0113.550.07

14.2013.690.17

VG-2

13.3213.300.13

13.3513.340.22

13.3713.370.10

13.3913.420.03

FeO

13.3913.360.11

13.7313.690.14

MgO

4.834.890.11

5.005.070.14

VG-999

KH

13.9113.890.11

13.7613.750.13

13.5913.580.03

13.5613.550.13

VG-2

13.5313.510.17

13.5913.560.20

13.7513.730.07

13.9513.920.09

KH

5.715.770.11

5.966.030.06

5.835.930.15

5.725.800.11

VG-2

5.575.620.10

5.855.890.16

5.665.720.17

5.805.860.20

CaO

9.389.310.12

9.419.350.14

KH

10.5910.660.19

10.4310.500.03

10.1010.170.13

10.4710.540.07

VG-2

10.3010.280.11

10.4310.420.10

10.3310.290.15

10.4610.430.22

Na2O

2.642.700.03

2.632.690.06

KH

T.312.350.11

2.322.36

£.•92.2.602.630.01

2.602.630.05

VG-2

rr.4ir2.520.11

2.382.41

2.542.570.09

2.602.650.06

K20

0.940.910.04

0.920.890.04

KH

0.160.160.03

0.170.170.02

0.180.180.03

0.160.160.00

VG-2

0.210.200.02

0.160.160.02

0.200.200.03

0.180.180.03

TiO2

4.054.030.17

4.084.050.14

KH

1.711.710.06

1.751.750.08

1.751.760.07

1.731.730.03

VG-2

1.931.930.06

2.001.990.04

1.971.960.09

1.951.940.03

P2°5

AP

0.190.230.00

0.180.220.02

AP

0.180.220.03

0.180.220.03

MnO

0.190.210.02

0.200.220.01

KH

0.190.190.02

0.200.200.02

Fay

0.210.240.02

0.220.250.02

The data in Tables 5-7 contain a wealth of in-formation for statistical evaluation that is beyondthe scope of this paper. One point, however, shouldbe emphasized: there is a general apprehension onthe 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 onestandardization, as a measure of precision. Sincemost microprobe users complete the analysis of aparticular mineral within one day, this is often the



Run No.

1

2

3

4

1

2

3

4

Conditions

Standards





Standards





SiO2

KH

50.6150.020.58

50.3949.840.73

50.4249.900.20

50.8350.260.31

VG-2

49.3749.630.46

50.2250.490.37

50.0850.370.56

49.9350.220.24

A12O3

KH

17.3716.690.19

17.2516.620.19

17.1016.510.20

17.3616.710.13

VG-2

16.0916.060.13

16.0516.040.20

16.3016.250.21

16.3316.310.12

FeO

KH

9.069.110.11

8.979.020.06

8.898.940.23

8.979.020.06

VG-2

8.858.890.12

8.898.930.26

8.918.960.20

9.019.050.05

MgO

KH

8.348.220.21

8.958.800.11

8.798.680.30

8.778.630.24

VG-2

8.368.200.12

8.998.810.44

8.458.280.11

8.738.560.25

CaO

VG-D08

KH

12.4612.640.15

12.4512.630.13

12.2112.390.18

12.4512.640.04

VG-2

12.2512.330.05

12.2412.330.45

12.3912.460.17

12.4412.510.16

Na2O

KH

T.3"42.280.06

2.312.240.03

2.322,. 250.09

2.442.370.06

VG-2

T.382.310.03

2.372.30_0.16j

2.302.230.04

2.332.260.01

KH

0.070.070.02

0.090.090.02

0.080.080.07

0.070.070.02

VG-2

0.090.090.02

0.090.090.01

0.090.090.02

0.100.090.02

TiO2

KH

0.960.970.05

1.011.020.05

1.051.060.07

1.001.010.03

VG-2

1.131.140.06

1.111.120.02

1.141.150.05

1.121.130.06

P2°5

AP

0.110.140.02

0.120.150.02

AP

0.110.140.03

0.120.150.01

MnO

KH

0.130.140.02

0.150.150.03

Fay

0.160.180.01

0.140.160.01

2 results in brackets, obtained using a 2ym beam, are not included in compilationsof Tables 3 and 4. See explanation in text.

only measure of precision that can be observed. It isclear, however, that the average values of analysesmade 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 12.57with a standard deviation of 0.13 and run 2 givesa 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 atthe 99% confidence level. Of course, this is to beexpected occasionally, given the limited stability ofthe instruments and the vagaries of data acquisitionin general. For microprobe users, however, it is diffi-cult to realize, and frequently even more difficultto accept, the fact that their data might not be asgood as it appears. Concern about the precision andaccuracy of microprobe analyses is ever-present withthe discriminating worker. One way to monitor the

NUMBER 22 65

TABLE 7.—USGS laboratory individual analyses of Kakanui hornblende and 4 natural glassesusing Kakanui hornblende as reference sample and also the laboratory's preferred referencesamples

Run No.

1

2

3

4a

4b

1

2a

2b

3

Conditions

Standards




Standards



Standards





Standards


SiO2

KH

40.4340.420.28

40.3540.390.56

40.1440.150.64

Di85

38.6140.730.24

37.9440.110.24

KH

50.6349.580.68

51.3850.250.77

51.0049.910.48

51.1850.050.28

Di85

49.1950.750.42

Al2

KH

14.14.0.

15.15.0.

14.14.0.

°3

939340

091026

919345

Ortho

12.14.0.

13.15.0.

714843

210442

KH

14140

14140

14140

15140

.87

.46

.21

.71

.30

.43

.75

.34

.40

.07

.61

.30

Grtho

12130

.65

.98

.19

FeO

KH

10.10.0.

10.10.0.

11.11.0.

505114

868603

060519

Garnet

10.10.0.

10.10.0.

535044

595719

KH

11110

11110

11110

11110

.94

.95

.26

.82

.83

.30

.77

.77

.19

.60

.61

.13

Garnet

11110

.82

.79

.14

MgO

Kakanui

KH

13.12.0.

12.12.0.

12.12.0.

029928

898935

848617

Di2Ti

12.13.0.

12.12.0.

341358

159317

KH

660

660

660

660

.93

.98

.10

.94

.98

.10

.88

.91

.18

.71

.73

.03

Di2Ti

670

.55

.02

.05

CaO Na2O

hornblende

KH KH

10.6610.660.49

9.719.720.63

10.3410.340.08

Di2Ti

10.5710.200.32

10.4010.040.11

VG-2

KH

10.9010.910.20

10.9710.980.28

10.9410.950.21

11.2211.240.08

Di2Ti

11.1110.720.06

2.682.680.13

2.762.760.12

2.582.580.07

Di85

2.472.680.12

2.542.760.05

KH

2.772.790.11

2.75..770.09

2.852.860.15

2.752.750.05

Di85

2.552.750.12

K20

KH

1.891.890.12

1.981.980.02

2.022.020.05

Ortho

1.951.910.04

2.011.970.02

KH

0.020.020.06

0.050.050.07

0.050.050.05

0.100.100»06

Ortho

0.190.180.06

TiO2

KH

4.314.320.18

4.304.300.08

4.204.200.13

Di2Ti

4.904.740.15

4.704.550.10

KH

1.631.630.03

1.651.660.08

1.661.660.07

1.631.630.03

Di2Ti

1.921.860.08

P2°5

AP

0.090.110.02

0.090.110.02

0.060.070.05

Ortho

0.020.020.07

0.050.050.05

AP

0.220.270.05

0.170.220.05

0.220.280.05

0.180.210.07

Ortho

0.200.190.05

MnO

KH

0.080.090.12

0.110.110.10

0.370.370.41

Rhod

0.070.070.04

0.040.040.01

KH

0.230.230.06

0.360.360.03

0.320.320.05

0.080.080.10

Rhod

0.210.230.05



Run No.

1

2

3

•-I

3

Conditions

Standards



Standards


Standards


Standards


Standards


UncorrectedB-A Corr.0 (5)

Standards


SiO2

KH

52.3950.800.51

52.6950.690.28

Di85

49.5850.800.60

KH

51.8850.620.32

Di85

49.3750.700.09

KH

50.3049.470.73

50.4149.600.66

Di85

47.9749.770.24

A12O3

KH

13.2512.840.30

13.0012.620.08

Ortho

11.6112.800.26

KH

14.4113.980.40

Ortho

14.2413.540.04

KH

16.3815.860.47

17.0416.460.34

Ortho

14.0215.440.17

FeO

KH

13.3913.360.15

13.5313.500.27

Garnet

13.4813.410.23

KH

13.6613.630.36

Garnet

12.7513.300.21

KH

8.758.790.22

9.019.050.19

Garnet

8.768.780.18

MgO

KH

4.89.4.990.10

5.085.180.08

Di2Ti

4.765.160.10

KH

5.575.650.08

Di2Ti

5.676.130.10

KH

8.868.710.18

8.548.400.12

Di2Ti

8.368.780.07

CaO

VG-A99

KH

8.448.420.48

9.169.130.18

Di2Ti

9.338.970.17

VG-999

KH

10.2610.250.14

Di2Ti

10.5710.180.03

VG-D08

KH

12.0312.100.66

12.2212.290.15

Di2Ti

12.4212.050.03

Na2O

KH

2.802.880.20

2.622.690.07

Di85

2.442.730.08

KH

2.662.720.11

Di85

2.302.570.03

KH

2.442.390.04

2.362.300.08

Di85

2.232.390.07

K20

KH

0.770.780.11

0.670.670.08

Ortho

0.770.760.07

KH

0.030.030.04

Ortho

0.040.040.00

KH

0.000.000.00

0.000.000.11

Ortho

0.090.090.06

TiO2

KH

3.813.790.10

3.733.710.10

Di2Ti

3.913.770.12

KH

1.661.660.08

Di2Ti

1.871.800.17

KH

0.880.890.03

0.960.970.07

Di2Ti

1.051.030.05

p 2o 5

AP

0.400.490.04

0.440.540.02

Ortho

0.320.310.04

AP

0.170.210.01

Ortho

0.150.150.11

AP

Q.130.160.01

0.100.120.02

Ortho

0.080.080.02

MnO

KH

0.270.270.05

0.160.160.04

Rhod

0.180.190.06

KH

0.130.130.30

Rhod

0.200.200.04

KH

0.180.190.01

0.140.150.21

Rhod

0.120.130.01

NUMBER 22 67

precision and accuracy is to present data of sampleswith known composition together with the micro-probe analyses of the unknown samples. This sug-gestion has been made frequently and is followedby many, but it would serve the scientific commu-nity much better if practiced even more extensively.

Literature Cited

Albee, A. L., and Lily Ray1970. 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 OceanTholeiites 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 ofAbyssal Volcanic Glasses. Smithsonian Contributionsto the Earth Sciences, 19:31-60.

Wright, T. L., and R. T. Okamura1977. Cooling and Crystallization of Tholeiitic Basalt,

1965 Makaopuhi Lava Lake, Hawaii. United StatesGeological 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 ofminerals and some updated analyses of mineralspublished previously.

Detailed descriptions of sample preparation andevaluation of homogeneity are given.

Introduction

Microprobe analyses are an essential part ofpresent-day mineralogical and petrological studies.It can be said that the application of the micro-probe to mineral studies and material sciences ingeneral is one of the most significant advances sincethe first use of the petrographic microscope inthe middle of the last century. The technique isnow well established, widely used, and capable ofhigh-quality analyses. As with all comparative in-strumental techniques, however, it requires well-characterized reference samples. Prime prerequisitesfor microprobe reference samples are homogeneityat the micrometer level and availability in rea-sonable quantities for standard chemical analyses.Either prerequisite is usually easily satisfied by itselfbut together are difficult to achieve.

One of the problems with some minerals used asmicroprobe reference samples is a lack of properdocumentation. Even if well-described minerals arefrom the same locality and/or are obtained froma reliable source, they may vary in chemical com-

Eugene Jarosewich, Joseph A. Nelen, and Julie A. Norberg,Department of Mineral Sciences, National Museum of NaturalHistory, Smithsonian Institution, Washington, D.C. 20560.

position. Therefore, a mineral sample intended asa reference sample should be carefully selected andused only when analytical data on this particularspecimen are available. Since natural materials ful-filling all the above requirements are not alwaysavailable, synthetic minerals and glasses have occa-sionally been prepared as substitutes. Again, homo-geneity of these materials should be checked andchemical analyses performed. The assumption thatthe precalculated composition is correct is certainlynot always valid.

In general, the most reliable microprobe analysesare obtained when a reference sample of composi-tion and structure close to that of the unknown isused because the matrix and possible wavelengthshift effects are minimized, and only small correc-tions are needed. It is generally accepted that,regardless of the type of correction used, resultscorrected by more than 10 percent should be viewedwith caution. Difficulties with correction proceduresin the Si-Al-Mg system have been pointed out byBence and Holzwarth (1977). Similar discrepancieshave been observed by other probe users.

All minerals and glasses described here, exceptone, are of natural origin. Most have been obtainedfrom the Smithsonian collections and were selectedeither in conjunction with specific projects or foruse in silicate analyses in general.

ACKNOWLEDGMENTS.—We wish to acknowledgethose curators and others listed in Table 1 whoprovided us with samples for use as microprobereference samples. Brian Mason's careful examina-tion and assistance in separation of minerals is alsogreatly 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 formicroscopic examination. Next, a microprobe analy-sis for composition and homogeneity is performed.If preliminary results are favorable, the material isgently crushed, sized usually between 20 and 80mesh, and further purified using either a heavyliquid separation or a Franz magnetic separator orboth. In some instances cleaning with a suitableacid is also useful. As a final step, the material isexamined under a low-powered microscope andmost remaining foreign grains are removed byhand. The purified grains are again checked bymicroprobe for homogeneity (sigma ratios) withinand among grains (Table 2). Finally, a chemi-cal analysis using classical methods (Peck, 1964;Hillebrand et al., 1953) is performed on the sameseparate 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 beenin use for some time. Johnstown meteorite hypers-thene and Springwater meteorite olivine have beenre-analyzed using what we believe to be muchcleaner separates. Kakanui hornblende has beenre-analyzed for TiO2.

Even after the most careful preparation of thereference sample, a grain of accessory mineral ormatrix may remain in the sample, which, in thecourse of preparation of the standard discs, couldbe included with the reference sample. Occasionalgrains of the standard itself will be "off composi-tion," due to inhomogeneity. These problems cannever be totally eliminated. The user should beaware of the possible presence of such "impurities"and make a thorough check for them. For example,occasional grains are found that are lower in sodiumand higher in potassium than usual in the referencesample microcline, lower in manganese than usualin Rockport fayalite, and lower in sodium thanusual in Lake County plagioclase. Infrequent inclu-sions in the glasses 72854, 111240/52, 113498/1, and113716 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) ofobserved standard deviation (sigma) to the standard

deviation predicted from counting statistics alonedoes not exceed 3. The sigma ratios were calculatedwith reference to ten ten-second counts on each often randomly selected grains. Table 2 gives sigmaratios for the ten grains of each reference samplefor major and some minor elements. The values inparentheses indicate the worst sigma ratio observedfor an element in a single grain. This does not,however, imply a single worst grain, as differentgrains may exhibit differing degrees of homogeneityfor each element present. When the criterion ofsigma ratios is used as a measure of homogeneity,all the reference samples prove to be very homo-geneous provided a reasonably large number ofcounts are taken on a reasonably large number ofgrains. In practice, however, fewer counts and grainsare normally used for standardization, and underthese circumstances a grain having a slightly dif-ferent composition may influence the microproberesults adversely. For this reason, grains showingsome discrepancy in composition should be avoided.The percentages of these "impurities" in the wholesamples are minimal and the effects on the bulkanalyses of the samples are negligible.

These samples were prepared in only small quan-tities, but they can be judiciously made available tomicroprobe users interested in the analysis of geo-logic materials. Potential users should rememberthat the purified samples differ in bulk chemistryfrom the specimens from which they were separatedand should be very specific in their requests—i.e., the requests should be made for microprobestandard USNM no. n rather than simply materialfrom specimen USNM no. n.

Literature Cited

Bence, A. E., and W. Holzwarth1977. Non-Linearities of Electron Microprobe Matrix Cor-

rections in the System MgO-Al2O3-SiO2. Proceedingsof the Eighth International Conference on X-RayOptics and Microanalysis and Twelfth Annual Con-ference of the Microbeam Analysis Society, page 38.

Boyd, F. R., L. W. Finger, and F. Chayes1967. 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.

Hoffman1953. 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.


KEY TO TABLE 1Analysts, 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, SmithsonianInstitution

J. Norberg, Dept. of Mineral Sciences, Smith-sonian Institution

E. L. Munson, N. M. Conklin, J. N. Rosholt, andI. C. Frost, U.S. Geological Survey

B. Wiik, Geological Survey, FinlandU.S. Geological Survey, Geochemistry and Pet-

rology BranchD. Mills, X-Ray Assay Laboratories, Ontario,

Canada; J. Nelen; J. NorbergE. Kiss, Dept. of Geophysics and Geochemistry,

Australian National UniversityJ. 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 SurveyH. Staudigel, Massachusetts Institute of Tech-

nologyR. S. Clarke, Jr., Dept. of Mineral Sciences,

8.

References

Smithsonian InstitutionJ. H. Berg, Northern Illinois University

References for previously published analyses:

1.

2.

3.

4.

5.

Stewart, D. B., G. W. Walker, T. L. Wright, andJ. J. Fahey

1966. Physical Properties of Calcic Labradoritefrom Lake County, Oregon. AmericanMineralogist, 51:177-197.

Young, E. J., A. T. Myers, E. L. Munson, andN. 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. Allen1973. Minor and Trace Elements in Augite,

Hornblende, and Pyrope Megacrystsfrom 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 theEarth Sciences, 14:85-86.

TABLE 1.—Chemical analyses of electron microprobe reference samples; analysts, sources, andreferences identified in "Key to Table 1" on facing page; these purified samples all differ, togreater or lesser degree, from the bulk chemistry of the USNM specimens from which they wereseparated (see text).

Mineral

Ahorthite, Great Sitkin Island, ALUSNM 137041

Anorthoclase, Kakanui, New ZealandUSNM 133868

Apatite (Fluorapatite), Durango, Mexico1

USNM 104021Augite, Kakanui, New Zealand

USNM 122142Benitoite, San Benito County, CA2

USNM 86539Chromite, Tiebaghi Mine, New Caledonia3

USNM 117075Corundum, synthetic1*

USNM 657SDiopside, Natural Bridge, NY

USNM 117733Fayalite, Rockport, MA

USNM 85276Garnet, Roberts Victor Mine, South Africa

USNM 87375Garnet, Roberts Victor Mine, South Africa

USNM 110752Glass, Basaltic, Juan de Fuca Ridge

USNM 111240/52 VG-2Glass, Basaltic, Makaopuhi Lava Lake, HI

USNM 113498/1 VG-A99Glass, Basaltic, Indian Ocean5

USNM 113716Glass, Rhyolitic, Yellowstone Nat. Pk., WY6

USNM 72854 VG-568Glass, Tektite, synthetic7

USNM 2213Hornblende, Arenal Volcano, Costa Rica

USNM 111356Hornblende, Kakanui, New Zealand8

USNM 143965Hypersthene, Johnstown meteorite

USNM 746Ilmenite, Ilmen Mtns., Miask, USSR9

USNM 96189Magnetite, Minas Gerais, Brazil10

USNM 114887Microcline, location unknown

USNM 143966Olivine (F090). San Carlos, Gila Co., AZ11

USNM 111312/444Olivine (F083), Springwater meteorite

USNM 2566Omphacite, Roberts Victor Mine, So. Africa

USNM 110607Osumilite, Nain, Labrador

USNM 143967Plagioclase (Labradorite), Lake County, OR

USNM 115900Pyrope, Kakanui, New Zealand

USNM 143968Quartz, Hot Springs, AR12

USNM R177O1Scapolite (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.


TABLE 2.—Sigma ratios (homogeneity indices) for all analyzed grains of each reference sampleobserved sigma for n grains

(sigma ratio for n grains = —; least homogeneous grain insigma predicted from counting statistics

parentheses; dashes = not evaluated)

Mineral

Anorthite

Anorthoclase

Apatite (Fluorapatite)

Augite

D n n i 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.

9744)

0149)

3498)

P2O5 MnO Cr2O3

0.97(1.51)

1(1

1(1

1.12(1.49)

.03

.58)

.21

.53)

observed sigma for all grainsSigma ratio for 10 grains = s i g m a predicted from counting statistics

observed sigma for this particular grainSigma ratio for least homogeneous gram = s i g m a 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 RobertsVictor 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." 17pages, 1 figure, 17 tables. 17 September 1970.

4. William G. Melson, Eugene Jarosewich, and Charles A. Lundquist. "Volcanic Eruption atMetis 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, andJack 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, 6plates, 1 table. 27 April 1971.

7. William G. Melson. "Geology of the Lincoln Area, Lewis and Clark County, Montana.'' 29pages, 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 Marginoff 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, andComposition 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 OuterContinental 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. 12April 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 theAlboran Sea: A Model of Input in a Deep Marine Basin." 51 pages, 23 figures, 8 tables. 16June 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 ofAustralian 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, 37tables. 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 RecentFoundering 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 theMetallography of Coarse Structure Iron Meteorites." 80 pages, 28 figures, 20 tables. 14 April1978.

Mineral Sciences Investigations 1976-1977

Documents

Transcript of Mineral Sciences Investigations 1976-1977