SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES ? NUMBER 10
Mineralogy, Mineral-Chemistry, and
Composition of the
Murchison (C2) Meteorite
Louis H. Fuchs, Edward Olsen,
and Kenneth J. Jensen
SMITHSONIAN INSTITUTION PRESS
City of Washington
1973
ABSTRACT
Fuchs, Louis H., Edward Olsen, and Kenneth J. Jensen. Mineralogy, Mineral-
Chemistry, and Composition of the Murchison (C2) Meteorite. Smithsonian
Contributions to the Earth Sciences, number 10, 39 pages, 19 figures, frontispiece,
1973.?The Murchison meteorite shower, September 28, 1969, occurred in and
around Murchison, Victoria, Australia. Chemical and mineralogical analyses
established it as a type II carbonaceous chondrite (C2). Murchison consists
largely of fine-grained black matrix which has been identified as primarily a
mixture of two iron-rich, low-aluminum chamosite polytypes. Contained in the
matrix are four main types of inclusions: (1) single crystals and crystal frag-
ments, (2) loosely aggregated clusters of crystals ("white inclusions"), (3) discrete
true chondrules, (4) xenolithic fragments of two other meteorite types (mostly
a unique kind of C3 chondrite).
The first type of inclusions consists of unzoned and highly zoned olivines,
unzoned (disordered and ordered) orthopyroxenes, clinoenstatite, and rare diop-
side. Prominent minor phases are calcite, chromite, metal (with occasional traces
of schreibersite), troilite, pentlandite, and two phases that could not be fully
characterized.
The second type of inclusions consists primarily of grains of olivine (Fa 0
to Fa 40), lesser low-Ca pyroxenes, and minor spinel, calcite, whewellite, hib-
onite, perovskite, chromite, pentlandite, and rare Ca-pyroxene.
The true chondrules consist of olivine, Ca-poor pyroxene, occasional metal,
and, in rare instances, one of the poorly characterized phases. The chondrules
are not texturally typical of the ordinary chondrites, but resemble more closely
those chondrules seen in C3 and C4 chondrites.
The fourth type of inclusion consists mainly of distinct xenolithic fragments
of a light blue-gray chondrite type that resembles certain C3 chondrites (like
Vigarano), though not in all aspects. These xenolithic fragments consist of dis-
equilbrated olivines and pyroxenes, abundant pentlandite and troilite, and
virtually no metal. In addition, a single xenolithic fragment was found of an
unknown meteorite type.
Ca- and Al-rich glasses (of varying compositions) are found as blebs, with
or without gas bubbles, contained within olivine crystals. The average Ca/Al
ratio of these glasses approximates that for all meteoritic matter. They may
represent early (nonequilibrium) subcooled condensates from the solar nebula.
This nonequilibrium stage was apparently followed by equilibrium condensation
through intermediate to low temperatures at which the layer-lattice phases con-
densed in abundance and incorporated crystals and fragments of the higher
temperature phases.
OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recordedin the Institution's annual report, Smithsonian Year. SI PRESS NUMBER 4784. SERIES COVER
DESIGN: Aerial view of Ulawun Volcano, New Britain.
Library of Congress Cataloging in Publication DataFuchs, Louis H.
Mineralogy, mineral-chemistry, and composition of the Murchison (C2) meteorite.(Smithsonian contributions to the earth sciences, no. 10)
I. Meteorites. I. Olsen, Edward John, 1927?joint author. II. Jensen, Kenneth J., jointauthor. III. Title. IV. Series: Smithsonian institution. Smithsonian contributions to the
earth sciences, no. 10.QE1.S227 no. 10 [QE395] 550'.8s [523.5'1] 72-13427
For sale by the Superintendent of Documents, U.S. Government Prin'ing Office, Washington, D.C. 20402Price 75 cents domestic postpaid or 55 cents GPO Bookstore
Contents
Page
Introduction 1
General Description 2
Mineralogy 3
Layer-Lattice Silicates 3
General Considerations 3
Compositional Evidence 4
X-ray Evidence 5
Other Layer-Lattice Silicate . 8
Olivine 9
Ca-Poor Pyroxene 9
Calcic-Pyroxene 9
Metal 9
Sulfides 10
Phosphorus Minerals 11
Calcite 11
Whewellite 11
Spinel Group Minerals . . 11
Lawrencite (?) 13
Hibonite (CaAl12O19) . 13
Perovskite 13
Sulfates . 13
Glasses 15
Poorly Characterized Phases 17
Textural Features . 19
Type 1 Inclusions . 19
Type 2 Inclusions 22
Type 3 Inclusions (True Chondrules) 23
Type 4 Inclusions (Xenolithic Fragments) 26
C3 Chondrite Inclusions 26
Xenolith of Unknown Meteorite Type 27
Mineralogical-Compositional Relationships 28
Calcium in Olivines . 28
Chromium in Olivines 28
Manganese in Olivines 28
Nickel in Olivines . 29
Nickel in Pyroxenes 30
Chromium and Phosphorus in Metal 30
Minerals Not Present in Murchison 30
Wet Chemical Analysis of Murchison . 31
Summary and Discussion of Observations 31
Appendix: Description of Wet Chemical Analytical Procedure 36
Literature Cited 37
in
FRONTISPIECE: A fragmental grain of olivine removed from the matrix of the Murchison me-
teorite; its resemblance to an "organized element" is purely coincidental. The "eye" of our
friend is a calcium aluminum silicate glass inclusion (21 microns long), the "pupil" is a void
bubble. The presence of the bubble, the ovate shape ot the "eye" together with its composition,
all indicate that the inclusion formed from a liquid state at temperatures between 1250? and
1450?C.
Mineralogy, Mineral-Chemistry, and
Composition of the
Murchison (C2) Meteorite
Louis H. Fuchs, Edward Olsen,
and Kenneth J. Jensen
Introduction
The Murchison meteorite shower occurred on
September 28, 1969, 1045 hours local time, 85
miles north of Melbourne, Australia. Details of
the fall and an initial description are given by
Lovering et al. (1971). In addition, Clarke (per-
sonal communication) has assembled details
pertinent to the strewnfield. A chemical analysis
was reported by Jarosewich (1971) and a short
description was published by Ehmann et al. (1970).
A somewhat more detailed description was given
by Fuchs, Jensen, and Olsen (1970) before the
1970 meeting of the Meteoritical Society.
Murchison is the largest known fall of a type II
carbonaceous chondrite. The major portion, ap-
proximately 65 kg., is located at the Field Museum
of Natural History, Chicago. Other major portions
are located at the National Museum of Natural
History, Smithsonian Institution, Washington, D.C.
(approximately 30 kg), the Australian Museum,
Sydney (approximately 4 kg), and the University of
Melbourne (approximately 7 kg). In addition,
Louis H. Fuchs, Chemistry Division, and Kenneth J. Jensen,
Chemical Engineering Division, Argonne National Laboratory,
Argonne, Illinois 60439; Edward Olsen, Field Museum of
Natural History, Chicago, Illinois 60605 and Department of
Geological Sciences, University of Illinois-Chicago Circle,
Chicago, Illinois 60680.
smaller quantities, up to several kilograms, have
been acquired around the world by a number of
museums, universities, and private persons.
In making this study we were fortunate in having
such a large quantity of this meteorite available.
Several hundred hand specimens were examined
and seven polished microscope mounts were used.
Mineral identifications were made by optical ex-
amination of handpicked grains and confirmed by
X-ray powder diffraction in all instances.
ACKNOWLEDGMENTS.?We would like to thank
Dr. Milton Blander, Argonne National Laboratory,
for many stimulating discussions on some of the
features observed in the Murchison meteorite and
Dr. E. Roedder, U.S. Geological Survey, for his
penetrating comments on portions of the manu-
script. We are grateful to members of the staff of
the National Museum of Natural History: Dr. Brian
Mason, Eugene Jarosewich, Joseph Nelen, Grover
Moreland, and Roy S. Clarke, Jr., for helpful coun-
sel and for making available unpublished data
on Murchison. Dr. Warren Forbes and Mrs. Mar-
garet Meyers, University of Illinois?Chicago Circle
Campus, helped in a portion of the microprobe
work and data reduction, and this is gratefully
acknowledged. We thank Mr. Karl Anderson, Ar-
gonne National Laboratory, for help on the micro-
probe at ANL, and Irene M. Fox (ANL) and
Ralph W. Bane (ANL) who obtained the carbon,
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
hydrogen, and nitrogen data reported. We espe-
cially thank Mr. Francis (Bert) Eliason of Murchi-
son, Australia, for making available the most
significant portion of the Murchison fall and point-
ing out features we might otherwise have missed,
and Mr. Glenn Commons and Mr. R. Groh who
provided the financial assistance to obtain the
major portion of this significant meteorite. Edward
Olsen was partly supported by a grant from the
National Aeronautics and Space Administration
(NGR-014-052-001), and Louis H. Fuchs was
partially supported by a NASA research grant. This
grant is gratefully acknowledged. Part of the work
was supported under the auspices of the United
States Atomic Energy Commission. The cost of
publication was paid by Argonne National Lab-
oratory and the Field Museum of Natural History.
General Description
Murchison consists of numerous individuals rang-
ing from a few grams to seven kilograms. Most
of the pieces are surrounded by fusion crust; how-
ever, because of low mechanical strength many are
broken and cracked, some possessing no crust at
all because it has spalled off at impact and after-
ward due to normal handling. Many specimens
show severe desiccation cracking, presumably due
to loss of volatiles. Several observers have noted
gradual weight changes, both losses and gains,
that are presumably due to changes in atmospheric
conditions in the several laboratories where weight
checks were made. Fine-grain size coupled with
good porosity and permeability permit the me-
teorite to "breathe." Although this has only modest
effects in the mineralogy, major composition, and
petrography, it can have serious implications for
any work that involves gases and organic chemistry.
In comparing hand specimens with other C2
chondrites Murchison shows a similarity only to
Murray. The ratio of inclusions + chondrules to
black matrix is clearly higher in both these meteor-
ites (Figure 1) in contrast to others of this group,
such as Mighei, Cold Bokkeveld, etc. Typical of
IFICURE 1 Hand specimens (left to right) of Murchison, Murray, and Mighei C2 meteorites.
Murchison and Murray show a closely similar content of inclusions, in contrast to other C2s
such as Mighei. The Mighei specimen is 7 cm long.
NUMBER 10
C2s the matrix is deep black and featureless. Dis-
persed throughout the matrix are several distinct
types of inclusions: (1) single crystals and crystal
fragments of individual minerals ranging from a
few microns up to millimeter size; (2) multi-grained
mono- and polyminerallic inclusions up to 2mm,
which we will call white inclusions; (3) true chon-
drules that are spherical or fragmented spherical
segments ranging in size from 0.05 to 0.8 mm; and
(4) angular xenolithic fragments of another c-hon-
drite type that range from a few millimeters up
to 13 mm, which we will call xenolithic fragments.
In making the distinction between the type 2
and type 3 inclusions above, we have relied upon
clear textural characteristics: the white inclusions
(type 2) have a granular texture?the loosely held
mineral grains can be raked out easily with a
needle. They are of various shapes and break with
the fine-grained black matrix that surrounds them.
True chondrules (type 3), on the other hand, are
relatively hard, well-defined spherical objects?the
minerals cannot be scraped out with a needle. They
do not break with the matrix but detach as distinct
beads. Figure 2 illustrates the difference between
the white inclusions and the true chondrules. We
emphasize this distinction because casual observ-
ers have tended to refer to all the inclusions as
chondrules. In certain instances this distinction is
not clear-cut.
Of these four types, the first two comprise over
98% of the inclusions seen in the black matrix.
Each type will be described after the total miner-
alogy has been presented.
FIGURE 2.?Chondrule (0.9 mm in diameter) in matrix with
white inclusion next to it.
Mineralogy
LAYER-LATTICE SILICATES
The major mineral in Murchison is the black
matrix layer-lattice silicate. It comprises 77 vol. %
as determined by linear measurements across two
hand specimens. Whether this phase is truly black
cannot be determined; it may be that the color is
due to disseminated submicroscopic amorphous
carbon and organic compounds. In addition, finely
divided sulfides (such as troilite) are black. Our
microprobe results show a consistent matrix car-
bon content, 1.6-2.2 wt. %, and sulfur, from
1.6-4.1%.
Because the layer-lattice silicate is the major
phase in all C2 chondrites, its definite character-
ization is highly desirable, especially as it has im-
plications for the genesis of this primitive group.
In spite of reported attempts by numerous inves-
tigators over the past twenty years, definite and
consistent results have not been obtained. The
only consensus is that it is a hydrated layer-lattice
silicate, either of the septechlorite group or serp-
entine group. Bostrom and Fredriksson (1966) pre-
sented microprobe and X-ray data suggesting it is
a ferric chamosite (septechlorite). Bass (1971) has
published a thorough review of the literature on
this matter.
GENERAL CONSIDERATIONS.?Because the matrix
phase comprises the dominant mineral making up
Murchison and because the other phases are no-
tably iron-poor, the bulk of the iron oxide in the
whole meteorite analysis must be in the matrix
phase. Serpentines in general are most often iron-
poor and magnesium-rich. In C2 meteorites it has
been believed that magnetite, submicroscopically
disseminated in a magnesian serpentine, could ac-
count for the bulk iron content. In Murchison,
however (and probably in other C2 meteorites),
magnetite is only a trace mineral (cf., "Spinel
Group Minerals"); the weak magnetic properties
of the matrix appear to be due to an Fe-S-O phase
(cf., "Poorly Characterized Phases").
On the other hand, the low total aluminum con-
tent of the meteorite argues against normal
septechlorites. The only known (terrestrial) sep-
techlorites with low Al and high Fe are cronstedite
and greenalite. These, however, have too low mag-
nesium to account for the bulk Mg content of the
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
meteorite analysis. Frondel (1962) has reported on
the serpentine, ferroantigorite (formerly called
jenkinsite), a platy polytype or serpentine. On
compositional grounds it might be comparable to
the matrix material in Murchison.
COMPOSITIONAL EVIDENCE.?The H2O content of
the matrix can be estimated from that fraction of
the water content of the meteorite liberated above
105?C. This fraction is calculated from the amount
of hydrogen liberated above this temperature and is
not a uniquely determined quantity, for some hy-
drogen may be released from a small amount of or-
ganic matter. In our analysis the total water is 12.1
wt. %; 8.8% is H2O+ (105?C) (Table la) which is
in good agreement with the 8.95% reported by Jaro-
sewich (1971) on a drier specimen of Murchison.
The matrix constitutes 77 vol. % of the meteorite
with the inclusion (mainly low iron olivine and
pyroxene) occupying the remaining 23 vol. %.
Assuming an average density of 3.3 for the in-
clusions, the density of the matrix is then 2.71 as
deduced from the measured density of 2.85 for the
TABLE la (wt. %) .?Comparison of chemical analyses of
Murchison meteorite
TABLE lb (atom %) .?Comparison of chemical analyses of
Murchison on a volatile (C, H, N, O, S) free basis
This Work
25.85
1.636
0.0713
31.99
0.146
2.822
0.374
0.218
33.12
2.20
1 .30
0.05]
0.21,
Calculated from the
Data of Jarosewich (1971)
26.46
1.564
0.0715
32.29
0.108
2.815
0.422
0.188
33.04
2.25
0.518
0.057
0.21,
meteorite as a whole. Thus the matrix comprises
73.4 wt. % of the meteorite and contains most of
the H2O + , which amounts to 12.0 wt. % (circum-
stantially almost equal to the total analyzed H2O).
TABLE 2a.?Electron microprobe analysis of Murchison
matrix
Fe
Cu
NiO
CoO
FeS
sio2
TiO2
Al,0,2 3
Cr2?3
FeO
Fe2O3
MnO
MgO
CaO
Na2O
K20
P2?5
2H
20(-)
SO,
3S
S=
c
CO,
2
N
Total Fe
Total S
(1) See discussion in
(2) A calculated value
This Work
0.03
0.015
1.73
0.076
(1)
27.22
0.099
2.05
0.402
(1)
(1)
0.22
18.90
1.75
0.57
0.034
0.23
8.80
3.26
2.38
0.49
1.81(2)
1.91 (Total)
n.d.
0.096
(5)
20.44
3.24H
8section - Wet Chemical Analysis .
. Total S -[S as SO, + elemental
(3) Sulfide sulfur included in FeS value.
(4) No elemental sulfu
(5) No summation since
(6) n.d. - Not determi
r detected.
values for FeS, FeO and Fe?Co are
I 3
ned.
Data of Jarosewich (1971)
0.13
n. d. (6)
1.75
0.08
7.24
29.07
0.13
2.15
0.48
22.39
n.d.
0.20
19.94
1.89
0.24
0.04
0.23
8.95
1.14
0.90
(4)
(3)
1.85
1.00
n.d.
99.80
22.13
3.00
S]
not assigned.
SiO,
C.
A12?3
Cr203
Fe,0,2 3
FeO
MgO
CaO
MnO
NiO
H,02
Sum
Si
Al
Cr
Fe
Fe+2
Mg
Ca
Mn
Ni
(OH)
1. Average values i
2. Oxides in no. 1
H20+ (see text)
3. 20% of Fe in no
1.
23.4
3.1
0.3
-
34.1
15.1
0.9
0.2
1.5
_
78.6
Number of
1.49>
10.23 1
> 1.73
0.01
JJ
1.81 ?>
1.43
I 3.39
0.01
0.08 J
4.54
For 40 spot analyses
recalculated to 88%
+3. 1 assigned as Fe
4. No. 3 recalculated to 88%, H20 = 12
2.
26.2
3.5
0.3
_
38.2
16.9
1.0
0.2
1.7
12.0
100.0
ions*
by J. Nelen,
, H20 content
.0%.
3.
23.4
3.1
0.3
7.6
27.3
15.1
0.9
0.2
1.5
79.4
1.45 >
0.22
0.01
0.36-
1.42 ^
1.39
0.06
0.01
0.08.
4.49
U.S.N.M., NiO
is calculated
4.
25.9
3.4
0.3
8.4
30.3
16.7
1.0
0.2
1.7
12.0
99.9
\
I 2.04[
)
1
I 2.96
1
is our datum.
from bulk
Calculated to 9(0, OH).
NUMBER 10
The major element composition of the matrix was
obtained from electron microprobe analyses. The
average of three analyses kindly supplied by J.
Nelen of the National Museum of Natural History
are given in Table 2a. The low summation is due
to several factors: the major one is the water con-
tent; oxides present in amounts less than 0.3% were
omitted because of their extreme variability. Our
microprobe analyses contain amounts for C and S
of 1.7 and 3.3 wt. %, respectively, these elements
are probably not a part of the layer-lattice silicate;
additional factors involve the porosity of the speci-
men and the nonideality of the polished surfaces
of the fine-grained matrix. The averaged analysis
was recalculated to 88 wt. % to allow for the esti-
mated water content of 12 wt. %.
The composition (Table 2a) is adjusted to assign
approximately 20% of the analyzed iron as ferric
so as to satisfy a tetrahedral to octahedral site ratio
of 2/3. The low silica content corresponds more
closely to that of a chlorite than to that of a ser-
pentine. The estimated water content may be too
high since the calculated oxygen to OH ratio of
TABLE 2b.?Electron microprobe analyses of "spinach" phase,
Murchison meteorite
unity is less than the ideal ratio of 5 to 4. Based on
composition alone it does not appear to compare
favorably with any known layer-lattice silicate.
X-RAY EVIDENCE. ? Approximately twelve X-ray
powder patterns (film) were taken of the black
matrix from several areas on different specimens of
the meteorite. Although sampling was done under
the binocular microscope so as to avoid visible
contamination, extra lines appeared on some pat-
terns which were not common to others. All pat-
terns, however, contain approximately nine lines in
common and these provided a starting point for
the identification (Table 3). In some cases the im-
purity lines could be assigned to either olivine or
calcite, but two prominent lines (intensity 3 to 5)
at 6.1 A and 5.4 A were not positively correlated
with each other. Since the 6.1 A line was generally
present on samples selected from unusually black
areas of the matrix, it was thought that this might
be due to an organic compound. Overnight treat-
ment of these samples with various organic solvents
produced no effect on the patterns. But a five min-
ute treatment with 2% nital eliminated the 6.1 A
line with no other effects to the overall pattern.
Except for the conclusion that the unknown phase
was acid soluble, no further attempt was made to
A12?3
Cr2?3
FeO
MgO
CaO
MnO
Sum
Si
Al
Cr
Fe+3
(OH)
28.0
2.6
0.4
28.0
2.6
0.4
2.3
27.3
19.9
23.4
3.7
1. Associated with Fo 99 in chondrule, E. Olsen analyst.
2. Recalculated, assuming 7% Fe as Fe+3
3. Composition of associated olivine not determined. J. Nelen, analyst.
4. Recalculated, assuming 5% Fe as Fe
* (OH) set to 4.00, formula calculated on the basis of 7 oxygen equivalents.
TABLE 3.?X-ray powder patterns* of the matrix material
from the Murchison, Murray, and Mighei meteorites
Mur
d(A)
7.2
-
5.4
4'7?4.3 j
3.9
3.62
3.58
-
-
-
-
2.70
-
2.53
2.43
2.30
2.16
2.03
1.79
-
1.57
1.54
1.49
1.44
chison
'I
10
-
1
Broad
Band
1
1
5
-
-
-
-
2
-
8
3
1
4
1
2
-
3
3
1
1
Murr
d(A)
7.1
6.0
5.4
4.7")
4.3 )
3.9
3.62
3.58
-
-
2.88
-
2.70
-
2.54
2.43
2.29
2.16
2.03
1.785
1.68
1 .575
1 .540
1.49
1 .44
ay
I
10
1
1
Broad
Band
1
1
8
-
-
1
-
3
-
6
4
1
3
1 +
2
2B
3
2+
1
1
Miqhei
d(A)
7.2
6.0
5.4
4.7
4.51
-
-
3.57
3.43
3.33
2.86
2.77
2.69
2.61
2.53
2.44
2.36
2.16
2.03
1.785
1.67
1.57
1.535
1.49
1.438
I
10
5
1
1
2
-
-
8
FT
FT
1
1
2-
1
10
1
1
6
1
4
IB
3
3
2
1
Assignments
C
A.S.
U
c
A.S.
C
?
c
?
?
?
A.S.
CM+CH
A.S.
CH
CM
C
CH
CM
CH
C
C
C
CH
C
*Fe radiation, Mn-filter, 114 mm Philips powder camera. Intensities are visual.
Abbreviations: B = broad; CM = monoclinic chamosite, CH = Hexagonal chamosite,
C = lines common to CH and CM; A.S. = acid soluble. U = unidentified fibrous phase.
6 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
identify it. Samples showing a 5.4 A spacing were
mounted and polished. These were found to con-
tain several percent of two phases, which displayed
massive amorphous and fibrous textures, respect-
ively. They are ubiquitous in all sections of Murchi-
son studied (cf. "Poorly Characterized Phases").
The fibrous phase is characterized by a strong 5.4 A
reflection.
Based on the nine lines common to all patterns
of the matrix material certain similarities exist with
the septechlorite, chamosite, and the serpentine
(ferroantigorite).
Carroll (1970) noted that a trio of medium in-
tensity X-ray peaks at 2.67, 2.40, and 2.15 A are
diagnostic for chamosites. Our matrix material pat-
terns show this trio at 2.70, 2.44, and 2.16 (the small
shifts in d-spacings due to compositional differ-
ences). This, in itself, could be compelling evidence
to establish this material as chamosite. Our X-ray
pattern of ferroantigorite from O'Neil Mine,
Orange Co., New York (Field Museum sample
M3641), however, shows the same three lines pres-
ent, 2.72, 2.43, 2.17 A. These, however, are not
uniformly of medium intensity, rather the 2.72
line is exceedingly weak. Because our 2.70 line is
stronger, and the trio approximates the intensity
pattern of chamosite, we regard a chamosite desig-
nation of the matrix material as a slightly better
possibility than ferroantigorite (or any other ser-
pentine-group mineral). Bostrom and Fredriksson
(1966) have concluded that the matrix material in
the Cl meteorite, Orgueil, is a ferric chamosite
primarily on compositional grounds.
For comparative purposes X-ray powder patterns
were obtained on the matrix material of Murray
and Mighei, and these are presented in Table 3
together with that of Murchison.* Aside from a
few impurity lines the d-spacings of corresponding
lines on the three patterns are in agreement, but
significant differences exist in the relative intens-
ities of some lines, especially those that correspond
to the prominent lines of either monoclinic or
hexagonal chamosite. The intensities of the line
at 2.43 A?monoclinic?vary inversely as those at
2.53 and 2.16?both hexagonal?(for a comparison
* Samples invariably contained impurities, as evidenced by
the spotty character of some lines; it was found that non-
rotation of samples during exposure simplified the distinc-
tion between the fine-grained matrix and coarser grained
impurities.
of both forms of chamosite, see Brindley, 1951 or
ASTM card 7-315). It can be seen that the ratio of
hexagonal to monoclinic chamosite is greatest for
Mighei, intermediate for Murchison, and least for
Murray. Although this relationship may not be
representative of the matrix material in these three
meteorites, the fact that a mixing ratio exists for
the samples X-rayed supports the preference of
chamosite rather than a serpentine mineral as the
predominant mineral in the matrix material. When
present, the line at 6.1 A is removed from all sam-
ples when treated with 2% nital for five minutes,
in the case of Mighei, lines at 4.51, 2.77, and 2.61
are also eliminated by this treatment. It should be
noted that our pattern of Mighei shows a number
of differences when compared to the pattern pub-
lished as "Murray F" by DuFresne and Anders
(1962) which they obtained from Mighei. Their
pattern contains 9 extra lines that can be assigned
to the strongest lines of olivine. Aside from the
olivine impurity, the remainder of the pattern is in
good agreement with ours.
As a further check on the properties of the matrix
material a series of heating experiments were un-
dertaken. A sample of Murchison matrix was
TABLE 4.?X-ray powder patterns* of Murchison matrix ma-
terial in an evacuated silica capillary following each stepwise
heating period.
Not
Heat
d(A)
7.2
5.4
3.58
2.70
2.54
2.44
2.17
2.12
1.79
1.57
1.54
ed
I
10
3
7
2
7
1
2
1
FT
3
2
3. 7 Days
245
d(A)
7.1
5.4
3.58
2.70
2
2
2
1
1
1
.54
44
.17
79
57
54
C
I
10
1
8
1
8
1
2
FT
3
2
7 7
?45
d(A)
7
3
3
3
1
9
58
33
2.97
2
2
2
85
70
54
2.44
2 16
2.04
1
1
1
1
Days
?C
I
10
1
7
FT
1
1
2
8
1
2
1
79 FT
57
54
48
2
1
1
1.7 Days
340
d(A)
7.1
3.58
2.97
2.70
2.54
2.44
2.16
2.05
1.79
1.72
1.61
1.57
1.54
1.48
C
I
10
7
2
1
8
2
2
2
1
1
FT
2
1
1
5.8 Days
340
d(A)
7.1
3.9
3.58
2.97
C
I
10
FT
7
3
2.70 FT
2.62
2.54
2.45
2.17
2.05
1.79
1.72
1.61
1 .57
1.54
1.48
1.31
FT
9
1
2
2
1
1
FT
2
1
FT
FT
24 Days
340
d(A)
7.1
3.58
2.97
2.70
2.64
2.54
2.44
2.16
2.05
1.79
1.72
1 .61
1.57
1.53
1.48
C
I
8
6
3
FT
1
8
1
2
3
1
2
1
2
1
1
1 .32 FT
12 Days
400?
d(A)
7.1
3.58
3.50
2.97
-
2.65
C
I
8
3
2
3
2
2.53 10
2.16
2.06
1.79
1.72
1.61
1.57
-
2
3
1
1
2
2
1.53 FT
1 .48
-
2
-
19 Days
450?
d(A)
7.4
I
2
3.60 FT
3.50
2.97
-
2.65
2.52
-
-
2.09
2.06
-
1.72
1.62
1.48
1.32
2
4
3
10
-
-
2
3
-
1
4
-
-
5
1
33 Da
450?C
d(A)
3.50
2.96
2.65
2.52
-
-
2.09
2.06
-
.72
.61
-
.48
.32
.092
.048
ys
I
-
FT
4
-
2
10
-
-
2
2
1
3
-
4
1
3
2
0.856 1
0.854 1
Phase
C
(1)
C
C
?
M
?
CM+CH
FeS
(2)
CM
CH
M
FeS
CH
M
M
c
c
M
FeS+M
M
M
Ma.
Ma2
*Cu radiation, Ni-filter, 114 mm Phil
Abbreviations: FT = faint trace; CM
C = lines common to CM and CH; A.S. =
fibrous phase; (2) = line common to M
Inte
: monoclinic chamosite; CH = hexagonal chamosite;
unidentified acid-soluble phase; (1) = unidentified
and CH; M = Magnetite.
NUMBER 10
sealed in vacuo in a silica capillary with a mini-
mum of empty space and heated for various times
at increasing temperatures. After each heating
period the capillary was air quenched to room
temperature and X-rayed. The results are shown in
Table 4. The major effect is the decomposition of
the layer-lattice structure, which begins at 340?C
and is practically complete after the final heating
period at 450?C. Magnetite, not present in the un-
heated material, begins to form about 100?C lower
than the apparent decomposition temperature of
the silicate and becomes the major phase at 450?C.
A significant side effect involves the decomposition
of the unidentified fibrous phase (with the charac-
teristic 5.4 A line) at 245?C to form a new phase,
Avhich is most probably troilite (lines at 2.65, 2.06,
and 1.32 A). Of minor importance is the occasional
appearance of weak lines as noted in the footnote to
Table 4. These may be due to local impurities
present along the length of the capillary.
A sample of ferroantigorite from the O'Neil
Mine, Orange County, New York, sealed in an evac-
uated silica capillary and heated at 450?C for 50
days showed that some alteration occurred to form
fayalite, but that ferroantigorite persisted as the
major phase. These differences observed for the
thermal stability of matrix Murchison material and
ferroantigorite support the X-ray results that they
are not identical. Unfortunately a clean specimen of
chamosite was not available on which to perform
similar experiments.
Aside from all of these considerations there are
problems attending the use of terrestrial minerals of
the septechlorite and serpentine groups as refer-
ence materials. Often many polymorphic forms exist
in a single terrestrial occurrence of a mineral that
may be characterized by one published powder
pattern. An example in point is the reported ex-
istence of eight polymorphs of cronstedite by Stead-
man and Nuttall (1963) for the type-locality in
Cornwall. A powder pattern for this mineral as
presented on ASTM card 17-470 could contain con-
tributing reflections from several polymorphs. Since
these polymorphs were discovered on single crystal
photographs there is visually no assurance that each
has the same composition as the bulk material.
The relative amounts of these phases of uncertain
compositions may vary sufficiently in a meteoritic
environment to mask or obliterate a strict com-
parison with terrestrial counterparts.
Ultimately there may be little justification for
discussion of the nature of the matrix material for
carbonaceous chondrites; there may be several
minerals and/or coexisting polymorphs of several
minerals of various compositions. Kerridge (1969),
using selected area electron diffraction techniques,
has found a wide range of unit cell parameters for
meteoritic layer-lattice silicates, which suggested to
him corresponding variability in compositions.
A polished matrix specimen was treated with
warm concentrated phosphoric acid for 30 minutes.
A microscopic gridlike pattern of matrix material
was removed leaving "islands" of matrix standing
in relief. About half of the matrix was dissolved.
a
11
OJ mm
FIGURE 3.?a, Plate of olivine (Fo100), parallel to (100) , con-
taining oriented glass stringers and blebs; b, plate of green
hydrated silicate called the "spinach phase" in text. Alterna-
tions of a and b impart a micaceous appearance to this type
of inclusion.
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
This suggests that the matrix consists of at least two
phases.
OTHER LAYER-LATTICE SILICATE.?Outside of the
matrix, within some inclusions, we found a separate
and unique occurrence of a layer-lattice silicate.
This silicate may or may not be the same, miner-
alogically, as the matrix material, for its appearance
and mode of occurrence set it apart from the
matrix material; the two may not be related genet-
ically. This silicate is characteristically a transpar-
ent spinach-green color and to distinguish it in our
discussion we will call it the "spinach" phase.
It is found in four distinctive settings: (1) in-
terstitially between, and coating, forsterite (Fo 99)
grains; (2) as complete inclusions of "spinach"
within the black matrix; (3) interstitial to forsterite
(Fo 99) grains within some true chondrules (type
3 inclusions); and (4) as rhythmically spaced lam-
ellae interleaved with olivine (Fo 99-50 plates)
(Figure 3).
The last type of setting with its distinctive platy
texture can almost be considered a separate type of
inclusion. Plates of unaltered olivine alternate with
isotropic or weakly anisotropic plates which vary in
color from light to dark green. The color is appar-
ently related to the degree of development of the
spinach phase. The darker green platelets are very
fragile and give an X-ray powder pattern resem-
bling (but not the same as) the patterns of the
matrix; whereas the lighter green plates give broad
lines of both a layer-lattice silicate and olivine. The
colorless olivine plates give good olivine patterns.
This material resembles that described by DuFresne
and Anders (1962) for Mighei and Murray: ". . .
olivine has been divided by the penetrating mate-
rial into a series of thin parallel plates with rather
a micaeous appearance, although diffraction shows
nothing present but olivine and Murray F." The
authors consider Murray F to be an alteration prod-
uct of olivine produced by the intervention of
liquid water, "the alteration having taken place in
the parent body." Our observations on these specific
occurrences in Murchison provide additional infor-
mation but still leave some unanswered questions.
The texture of these "lit par lit" spinach-olivine
inclusions ends abruptly at the surrounding matrix.
In section there are no obvious adits leading to the
inclusion; additionally, nearby olivine crystals or
fragments within the matrix may be unaltered,
suggesting that the alteration is pre-Murchison in
origin.
The plates of olivine are always parallel to the
(100) plane, as evidenced by a centered acute bi-
sectrix interference figure, and they contain nu-
merous glassy inclusions of various shapes from
distorted spheres to elongated blebs or ribbons
(Figure 3). Glassy blebs within some of the plates
of olivine are oriented in directions parallel to or
45? to the extinction directions and thus occur
along crystallographic directions. We have not an-
alyzed these glasses as most are only a few microns
wide in their shortest direction. Lacking compo-
sitions, we can say nothing about their origin or
why they are so numerous. Their abundance, how-
ever, suggests that they are somehow related to the
degree of olivine alteration. It is possible that the
spinach formed by the replacement of olivine plates
which were loaded with nonresistant glass. The
surviving unaltered olivine either contained dis-
crete glass blebs protected by olivine, or perhaps
the unaltered glass was simply more resistant.
One clue that possibly illustrates the initial de-
velopment of the spinach phase comes from a
chondrule in a thin section of Murray. Parallel
bands of olivine are interleaved with alternating
bands of light-green mottled isotropic glass. The
green is confined to the glass; bands of olivine are
unaltered.
The quality of the X-ray patterns for the matrix
material did not permit comparison of subtleties
that would allow more definitive interpretation.
Only 9 X-ray lines were obtainable in that case. On
the other hand, X-ray patterns of the green spinach
phase yield 14 relatively sharp lines that can be
assigned to a mixture of monoclinic and hexagonal
polymorphs of chamosite; a few weak unidentified
impurity lines are usually present.
In the case of the fourth setting, it was noted
that the plates of layer-lattice silicate visibly con-
tained impurity phases. An X-ray pattern consisted
of 21 lines and was more difficult to interpret. It
appeared to consist of a mixture of chamosite,
serpentine, and other unidentified phases.
The composition of this spinach phase from a
true chondrule is presented in Table 2b, which
includes an analysis of the same phase obtained by
f. Nelen from an undetermined site in one Murchi-
son section. It is apparent that this phase varies
NUMBER 10
in composition. Although sufficient data are lack-
ing, this variation apparently has no relation to
the composition of the associated olivine; in our
occurrence (29.3 wt. % FeO) the associated olivine
is practically Fe-free. Both analyses can be calcu-
lated to 7 oxygen equivalents as no estimates of
H2O contents can be made. Compared with the
matrix material (Table 2a), smaller amounts of
Fe+3 are required to fill the tetrahedral sites, but
clearly there are too many octahedral ions in both
spinach analyses. This indicates that nonlayer-
lattice silicate impurities are present. They are
either amorphous or, if crystalline, are present at
levels not permitting X-ray identification. These
uncertainties nullify the possibility that the spin-
ach phase is a simple in situ alteration product of
olivine.
OLIVINE
Olivine is the next most abundant phase, oc-
curring in all four types of inclusions. It ranges in
composition?from grain to grain?from Fa 0 to
Fa 85, both as internally homogeneous crystals and
as strongly zoned crystals. Olivine will be discussed
in detail under "Textural Features" and "Miner-
alogical-Compositional Relationships."
CA-POOR PYROXENE
Both the orthorhombic and monoclinic forms of
the enstatite-bronzite series occur, compositions
ranging from Fs 1 to Fs 50. No zoning was observed.
Many of the orthorhombic crystals gave X-ray pat-
terns of a disordered form (Pollack and Ruble,
1964) with considerable line broadening. Some,
however, showed normal ordered patterns. Pyrox-
ene will be treated in more detail later.
C ALCIC-P YROXENE
Fragments of calcic-pyroxene occur within the
matrix, but are exceedingly rare. Partial analyses
performed on some grains gave Fe/ (Fe+Mg) ratios
of 0.06 to 0.09. Diopside has been found surround-
ing hibonite in minute white inclusions (cf.,
"Hibonite" and "Perovskite"). In one instance a
dark green augite grain was found.
FIGURE 4.?Polished section of an exceptionally metal-rich
area. Some troilite is present, but it is less than one percent.
METAL
Metal is not abundant. In the bulk analysis
(Table la) it amounts to only 0.03%. A few sec-
tions, however, contain abnormally high amounts
(Figure 4). It is found mainly within some olivine
grains, both in white inclusions and in true chon-
drules. Some rare isolated grains occur in the black
matrix or within Ca-poor pyroxene. Metal grains
range in size from submicron blebs up to 0.42 mm
in sections we studied. The very few large grains
were found isolated in the matrix and not in
olivine or pyroxene. The grains are smooth ovate
to spherical in shape, which is distinctly different
from the angular shapes commonly found in the
ordinary chondrites.
A group of metal analyses are given in Table 5.
These analyses are of metal grains within olivine
(both from white inclusions and true chondrules)
and isolated in the matrix. There are no systematic
chemical differences between the metals from these
three sites.
The outstanding features are the unusually high
Cr and P contents, from 0.20 to 0.96 wt. % and
0.28 to 0.37 wt. %, respectively. Metal from all
other classes of meteorites, where measurements
are available, carry only trace amounts of Cr (less
than 0.001%). Indeed, in most instances it is not
reported because it is presumably at low trace
levels only. Some of the metal in lunar lavas runs
up to 0.42% Cr (Taylor, Kullerud, and Bryan,
10
TABLK 5.?Electron microprobe analyses of metal grains in
Murchison
Location of grain
Four grains
within
single
chondrule
Two grains
within
single
chondrule
Six grains
in
black matrix
1
2
3
4
1
2
1
2
3
4
5
6
f Center
\Edge
f Edge
\ to
[Edge
Fe
93.8
94.6
94.0
93.7
90.8
92.2
92.1
93.5
94.0
93.6
93.7
93.1
92.0
Ni
5.6
4.3
4.0
4.9
7.4
5.6
5.5
5.0
4.8
4.8
4.8
6.3
7.0
Co
0.40
0.32
0.32
0.35
0.41
0.74
0.40
0.35
0.38
0.35
0.32
0.42
0.43
Cr
0.96
0.85
0.79
0.80
0.86
0.69
0.78
0.83
0.64
0.73
0.49
0.77
0.20
0.55
P
0.28
0.32
0.32
0.37
0.29
0.37
0.35
0.37
0.32
0.36
0.35
0.31
0.32
Sum
101.0
100.9
100.3
99.4
100.2
99.6
99.7
99.2
99.9
100.2
99.6
99.9
100.3
100.3
1971), which is still a factor of 1.7 less than the
average Murchison metal (0.7%). Only two grains
were large enough to test for Cr homogenity. Scan-
ning across a large (0.42 mm) grain (isolated in
the matrix) showed a steady Cr profile of about
0.2% for about half the grain, then rising smoothly
to 0.55% at the opposite edge. In the other case, a
grain of metal within forsterite in a true chon-
drule showed a central core of 0.96% Cr falling to
0.85% at the edges.
The average P content of these metal grains is
0.33%. The large 0.42 mm matrix grain was
scanned for phosphorus homogeneity. The grain
ran generally around 0.32% with a few spots show-
ing up to 0.60%. These spots may be due to small
specks of phosphide, although none were visible.
No such spot was wide enough for microprobe
analysis.
Because of the unusual chemistry of the Murchi-
son metal careful wavelength scans were performed,
especially for Si, Ti, V, Mn, Cu, Mo, W, and Pt.
None of these was present above detection limits
(about 0.05%) although one grain showed a pos-
sible Pt content of about 0.07%.
We assume the metal is kamacite; however, we
do not have any data that totally rule out other
structural states. The grains analyzed were generally
too small to scan for homogeneity. The one large
(0.42 mm) grain was relatively homogeneous in Ni
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
content, showing a slight rise from 6.3% to 7.0%
going across the grain; however, a few isolated
spots within it showed lows of 4.7%. Etching of
larger grains revealed no observable structural fea-
tures, plessite areas, Neumann bands, etc.
The recent study of the Fe-Ni-P system by Gold-
stein and Doan (1972) shows that kamacite with up
to 0.4% P and up to 7% Ni can exist without
schreibersite being formed down to a temperature
of 550?C. These are the upper limits of P and Ni
values in the Murchison metal. Grains of kamacite
with as little as 0.3% P and 4% Ni could repre-
sent even lower temperatures (450-500?) without
the formation of schreibersite. The effect of up
to 0.9% Cr on this is unknown.
The unusual content of Cr in the Murchison
metal will be discussed under "Chromium and Phos-
phorus in Metal." We also note that a complete
chondrule made of metal, with some schreibersite,
was found in one single instance. It is described
under "Phosphorus Minerals."
SULFIDES
Both troilite and pentlandite were found, the
former several times more abundant than the lat-
ter. The predominance of troilite over pentlandite
in the case of Murchison disagrees with the reverse
relation in other C2s (Ramdohr in Mason, 1962-
63). The troilite commonly shows a mottled appear-
ance due to finely disseminated pentlandite
(Figure 5). The mottling ranges from submicron
FIGURE 5 -Pentlandite (light) within troilite matrix,
immersion, scale bar is 10 microns.
NUMBER 10 11
size up to 5-micron patches of pentlandite. The
nickel content of the troilite ran as low as 0.2 wt. %
where the mottling was absent entirely, up to
1.7% where the mottling was very apparent. It
would appear that the nickel content of the troilite
is quite low, comparable to values in other com-
mon meteorite types, and that values above 0.2%
are due to the minute admixture of pentlandite.
Separate analyses of the larger pentlandite grains
give Ni=24.4%.
Texturally most of the troilite occurs as angular
grains in the matrix; it has a distinct fragmental
appearance and is not associated with metal grains.
This is in marked contrast to some C3 chondrites,
where troilite commonly rims metal grains. Point
counting of troilite and pentlandite grains gives
an abundance of 0.13 wt. % which is considerably
less than the amount that is computed from the
bulk chemical analysis, 7.24 wt. % (Jarosewich,
1971). This disparity is because the bulk of the
sulfur is not present as troilite but must be finely
divided and dispersed in the black matrix. As such
it is not visible in an interior section, but its pres-
ence is revealed within a band 0.3-0.6 mm wide
adjacent to the fusion crust. Atmospheric heating
has formed profuse amounts of a light-yellow phase,
softer and less reflecting than troilite, which result-
ed from a reaction between matrix sulfur and
iron in unknown states. This soft-yellow phase may
be related in composition to the poorly character-
ized sulfur-bearing phase (see "Poorly Character-
ized Phases) where it is not associated with the
fusion crust, but the two phases differ markedly in
mode of origin.
PHOSPHORUS MINERALS
No phosphates were seen; however, the phos-
phide schreibersite occurs as blebs in the chond-
rule of metal noted earlier, and as rare specks
within metal grains in the black matrix. It was
never observed in metal grains contained within
olivine crystals or contained in true chondrules.
This is the first time schreibersite has been reported
from any carbonaceous meterorite.
CALCITE
Calcite grains were ubiquitous in all sections
studied. It occurs principally in the black matrix;
however, X-ray patterns of some white inclusions
showed it to be present there in small amounts.
Grains range from a few microns up to 80 microns.
X-ray, optical, and microprobe work show it to be
free of any cations other than calcium. The grains
are frequently polysynthetically twinned.
WHEWELLITE
Whewellite, the calcium oxalate monohydrate,
Ca (C2O4).H2O, was found in one white inclusion
between olivine grains. The X-ray pattern matches
the ASTM reference pattern perfectly with no
extraneous lines present. Over ten lines were ob-
served. The occurrence in Murchison is logical
considering the relatively high organic and calcite
content of the meteorite. Terrestrially whewellite
is found in organic-rich sediments, plant cells, some
kidney stones, and occasionally in hydrothermal
deposits.
At 230?C whewellite decomposes to the anhy-
drous salt (Freeman and Carroll, 1958). This tem-
perature, however, cannot be used as a maximum
temperature indicator here because the anhydrous
salt is hygroscopic (Miller, 1953). At lower tempera-
tures it would revert to the monohydrate in the
water-rich environment within the meteorite. At
best a maximum of 480?C may be indicated where
the anhydrous salt decomposes to form CaCO3+CO
(Freeman and Carroll, 1958).
SPINEL GROUP MINERALS
Three minerals of this group are found in
Murchison: spinel proper is found as inclusions
within olivine and in white inclusions; chromite
is found in white inclusions and as fragments in the
black matrix; spherules of magnetite are found in
the matrix and occasionally within troilite. The
NiO content of the magnetite within the troilite
is 0.8%, whereas no nickel is found in the matrix
magnetite spherules.
Spinel proper varies considerably in composition,
color, and cell size from grain to grain, from pure
MgAl2O4 to compositions with considerable Fe
and Cr (Table 6).
All these members of the spinel group occur in
trace amounts only. Chromite is somewhat more
abundant than the others. Magnetite is quite
variable in quantity from section to section, al-
12 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 6.?Electron microprobe analyses of spinels
A12O3
Cr2?3
Tl02
FeO
MgO
CaO
Sum
n
a(A)
Al
Cr
Ti
Fe+2
Mg
Ca
R+V
R+2
Notes
Nos. 1
Compos
Nos. 4
Compos
No. 6
n.d. =
Mn, Zr
1
71.1
0.7
0.3
0.1
28.8
n.d.
101.0
1.722
8.091
Numbe
15.78
0.10
0.05
0.01
8.08
-
4 15.93
8.09
to Table 6.
,2, and 3 are 50-120 urn
itions of associated ol
and 5 are 6 ym grains '
2
46.5
25.4
0.3
10.9
18.3
n.d.
101.4
1.855
8.182
rs of ions
11.75
4.30
0.05
1.96
5.85
-
16.10
7.81
grains fror
ivines are:
3
59.9
12.8
0.3
2.5
25.7
n.d.
101.2
1.780
8.135
on the basis i
13.95
1.99
0.05
0.42
7.57
_^_
15.99
7.99
n 3 different
FaQ(No. 1),
inside of glass inclusior
ition of host olivine is Fan ,.0.6
is a 12 urn inclusion in
not determined
Fa0.3-
, K, Na are below background.
4
67.4
2.6
0.2
0.6
27.5
1 .6
99.9
n.d.
n.d.
Df 32(0)
15.34
0.39
0.03
0.09
7.91
0.34
15.76
8.34
olivine-spinel
Fa1Q (No. 2),
I la and lb res
5
67.3
2.1
0.2
0.6
28.1
0.1
98.4
n.d.
n.d.
15.46
0.33
0.04
0.09
8.16
0.02
15.83
8.27
white ir
Fa2-7 (*
p. Table
6
69.6
0.7
0.2
0.5
28.3
0.0
99.3
n.d.
n.d.
15.73
0.12
0.03
0.08
8.09
0.0
15.88
8.17
iclusions.
j. 3).
8.
though generally lower than chromite. The general
low abundance of magnetite runs counter to many
statements in the literature that it is a significant
phase in all carbonaceous chondrites. Small chips
of Murchison are noticeably magnetic. The magne-
tism, however, may be due to a relatively abundant
phase which we have been able only to poorly
characterize as a possible Fe-S-O phase (see "Poorly
Characterized Phases").
The chromite varies moderately in composition
from grain to grain (Table 7). The compositions
are notably different from chromites in all other
meterorite types (Bunch, Keil, and Snetsinger,
1967; Snetsinger, Keil, and Bunch, 1967; Bunch,
Keil, and Olsen, 1970; Bunch and Keil, 1971). The
analyses reported in Table 7 are for individual
angular fragments dispersed within the matrix in
one probe section. These chromites are generally
characterized by their A12O3 contents (av. = 14.8
wt. %) and in this respect they are markedly differ-
ent from those in the ordinary chondrites (av.
about 6%), and resemble those from terrestrial
igneous rocks (av. about 16%). Of all the meteoritic
chromites analyzed in the four references cited,
exclusive of those in the brecciated mesosiderites,
only a few have A12O3 contents approaching those
in Murchison, and they occur in a few members
from highly fractionated classes such as the hypers-
thene achondrites and some irons. If the C2s
contain primitive matter, it is surprising that the
compositions of the Murchison chromites are so
different from those in the abundant chondrites.
It is difficult to imagine some process whereby
these compositions might be brought to the com-
positions of ordinary chondritic chromites, especial-
ly at temperatures around 800?C where most chon-
drites are thought to have equilibrated. Only small
differences in the compositions of chromites exist
between the various metamorphic grades of chon-
drites, hence metamorphism has only a minor effect
in altering chromite compositions. The composi-
tional differences between Murchison and ordinary
chondritic chromites were consequently established
in some primary process. We can only speculate
that the temperatures involved were rather high
and close to those in the magmatic range that
effected the fractionation of some meteorites with
TABLE 7.?Electron microprobe analyses of chromites
Cr2?3
Al203
V2?3
Ti02
FeO
MgO
MnO
Cr
Al
V
Ti
Fe+2
Mg
Mn
47.4
17.3
0.6
0.9
26.7
6.2
0.3
99.4
9.93
5. 40
0. 13
0. 18
5.92
2.45
0.06
46.8
17.7
0.5
0.7
28.4
5. 7
0.2
100.0
Nun
9.80
5.52
0.11
0.14
6.28
2.24
0.05
47.1
18.5
0.4
0.6
25.4
8.1
JLJL
100.3
nbers of
9.65
5.65
0. 08
0.12
5.51
3.13
0. 05
60.2
3.2
0.6
0.5
32.9
1 .3
0.3
99.0
ions or
14.07
1.12
0.14
0.11
8.14
0.57
0.07
48.9
15.7
0.8
1 .6
27.4
5.7
0.3
99.4
l the
10.25
4.90
0. 18
0. 32
6.07
2.24
0.06
51 .4
10. 7
0.7
1 .7
29.3
4.5
0. 3
98.6
basis of
11 .33
3.52
0.15
0.35
6.84
1 .89
0.07
41 .9
22.3
0.6
0.9
27.7
6.9
0.2
100.5
"32(0)
8.49
6.73
0.12
0.17
5.95
2.63
0.05
50.3
14.0
0.7
1 .2
27.2
6.7
0.3
100.4
10.60
4.40
0.14
0.24
6.07
2.66
0.06
49.2
14.2
0. 7
1 .4
28.2
5.8
0. 3
99.8
10.47
4.51
0.15
0.29
6.34
2.33
0.06
15.
8.
64
43
15.
8.
57
57
15.
8.
50
69
15.
8.
44
78
15.
8.
65
37
15
8
.35
.80
15
8
.51
.63
15.
8.
38
79
15
8
.42
.73
24.07 24.14 24.19 24.22 24.02 24.15 24.14 24.17 24.15
Redistribu
Fe as Fe
;Fe+++ 8
!+3+R+4 15
!+2 7
,um 23
tion
+
.0
.99
.89
.88
of
8.
15.
8.
23.
lor
0
95
01
96
is to
11 .
15.
8.
23.
al 1
0
96
01
97
ow
9
1 5
7
23
tor
. 0
.98
.95
.93
a n
8
16
7
23
ass
.0
.01
.83
.84
1 gn
10
15
8
23
ed
.0
.87
.02
.89
perc
1 0
1 5
7
23
enta
.0
.97
.93
.90
ge o
1 1 .
15.
8.
23.
f total
0
91
03
94
11 .
15.
7.
23.
0
93
94
87
Note to Table: The number of divalent ions exceeds the ideal number of
8 and the sum of the quadrivalent and trivalent ions is less than the
ideal number of 16. This indicates that either ferrous or ferric ions
are in 6-fold coordination. For illustrative purposes, we have recalcu-
lated each analysis assuming that from 8 to 11 wt. % of the total Fe may
be present as Fe but no case is made for its presence. The departure
in the total number of cations from 24 may be the result of analytical
error.
NUMBER 10 13
distinctive compositions such as the irons and
hypersthene achondrites. These observations clearly
argue against the derivation of ordinary chondrites
from C2 chondrites by any process short of melting.
LAWRENCITE (?)
On many (not all) freshly broken and fusion-
crust-covered surfaces we observe scattered, spotty
"bleeding" taking place slowly. The bleeding
consists of patches, up to several millimeters, of
hydrated red iron oxide. In the past such bleeding,
especially in irons, has been attributed to the pres-
ence of unstable lawrencite. Our attempts failed
to definitely identify lawrencite. Whatever unstable
phase is causing this, its distribution is not uniform.
Individual specimens vary from no spots, to a few
scattered spots, to numerous ones. In general, the
spots tend to cluster into areas up to a centimeter
across.
sion had higher indices of 1.80 and 1.77. Hibonite
in Murchison marks the first occurrence of this
mineral in a C2 meteorite. The blue color and
absence of associated gehlenite serves to distinguish
this occurrence from those reported in the C3s.
PEROVSRITE
Scattered euhedral crystals of perovskite, CaTiO3
(2 to 4 microns), occur with diopside and spinel
which surround patches of hibonite. This is similar
to the occurrences in Allende and Leoville in which
hibonite is rimmed by perovskite. In those C3
chondrites, however, gehlenite is an accessory min-
eral. No gehlenite was found in Murchison. Perov-
skite was identified by its characteristic X-ray
powder pattern. A partial probe analysis (uncor-
rected amounts vs. rutile and grossular standards)
gave, in wt. %, 55 TiO2 and 43 CaO as compared
to an ideal composition of 59 TiO2 and 41 CaO.
HIBONITE (CaAl12O19)
Some very rare white inclusions, 100-150 microns
in diameter, have intensely colored blue centers of
microcrystalline hibonite needles, the largest 3 x 10
microns. X-ray powder patterns of the blue con-
centrates were predominantly those of hibonite
with minor perovskite. Line spacings and relative
intensities of the hibonite pattern were identical
to those reported for hibonite from the Leoville
meteorite (Keil-and Fuchs, 1971). The surrounding
white material is spinel and perovskite which may
be associated with diopside. One diopside-bearing
inclusion was mounted and probed. Diopside forms
the outer surface with an intergrowth of diopside,
spinel, and perovskite making up the mantle which
encloses in turn a core of hibonite. Because of the
small width of the hibonite needles, only a partial
analysis could be obtained: in wt. %, A12O3 74.5,
CaO 8.2, MgO 2.9. It is suspected that several per-
cent of titania may be present analogous to one of
the hibonite compositions found in both Leoville
and Allende; however, during the course of the
analysis epoxy-bubbling obscured the surfaces and
forced us to terminate the analysis.
Optically, the needles are strongly pleochroic: e
= colorless, to ? intense blue; refractive indices
are 1.78 (e) and 1.75 (co); grains from one inclu-
SULFATES
Glistening plates (up to 40 microns) of gypsum
were found in localized areas on the surfaces of
some hand specimens (Figure 6). X-ray patterns
were of excellent quality and contained no extra
lines. Qualitative microprobe scanning showed no
cations other than calcium present above the 2%
level. By a series of observations on fresh surfaces
FIGURE 6 Scanning electron microscope photograph of gyp-
sum crystals that occur in localized areas on some freshly
broken fragments. Scale bar is 20 microns.
14 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
it appears that gypsum does occur in Murchison as
a pieterrestrial phase filling localized short, discon-
tinuous, thin veins. Evaporation on surfaces
apparently causes redeposition of intergranular
gypsum as large plates. We examined specimens of
the C2, Cold Bokkeveld, which have fracture sur-
faces created by sampling breakage in museum
collections over many years. All of these post-
terrestrially made surfaces are covered by numerous,
outward-projecting plates of gypsum.
No other sulfate phases were found; however, it
is possible others may be present. This is inferred
from the appreciable amounts of Na and Mg in
the water-soluble fraction determined in the bulk
chemical analytical results (see "Chemical Analysis
Discussion and Comparison").
FIGURE 7.?Assorted calcium aluminum silicate glass inclusions within olivine: a, largest in-
clusion found is 36 microns in diameter and contains a void bubble. Smaller inclusions are
present but are not in focus; b, 18-micron inclusion with void bubble and spinel crystal at-
tached to outside surface of inclusion. Smaller inclusions (not in sharp focus) do not have
spinel crystals; c, elongated inclusion (11 by 29 microns) containing void bubble and euhedral
spinel crystal, d, olivine fragment (220 microns in diameter) containing numerous inclusions
elongated parallel to extinction directions.
NUMBER 10 15
GLASSES
Rare shards of glass that yielded no X-ray pattern
were observed in some polished sections. In addi-
tion, most true chondrules (type 3 inclusions)
contain small amounts of glass. The compositions
vary moderately and average around 77% SiO2,
14% AlaO3, 9% CaO, and less than 1% of Na2O
+ K2O. Of special interest is the occurrence of
ovate glass inclusions (up to 36 microns in diameter)
within single olivine crystals that are isolated in the
matrix (type 1 inclusions, Figure 7). Olivine grains
within the loosely consolidated multigrained aggre-
gates (type 2 inclusions) also contain these inclu-
sions. The latter occurrence has not been studied in
detail, so our description and discussion of the
inclusions apply only to those within the isolated
subhedral to euhedral olivines.
These inclusions are similar in appearance to
those noted in the Chainpur chondrite by Dodd
(1969) and to the melt inclusions in lunar and
terrestrial olivines described by Roedder and Weib-
len (1971). An inclusion may or may not contain
void bubbles, spherules of metal, an iron sulfide, or
euhedral spinel crystals. Interior walls of some void
bubbles are lined with an iron sulfide (presumably
troilite), suggesting that a sulfur-containing vapor
occupied the bubble volume at elevated tempera-
tures. The size of the bubbles varies considerably,
and there appears to be no correlation with the size
of the inclusion. For twelve inclusions measured, the
volume percentage of bubble to inclusion varies
from 1 to 14; some inclusions have no bubble. This
variation in bubble size cannot easily be accounted
for by a simple differential contraction mechanism
between inclusion and host during cooling. Alter-
natively, bubble occupancy by vapors under vary-
ing pressures may account for the observed varia-
tion. The smooth ovate shapes of the glass inclu-
sions (most are spherical but some are elongated)
indicate that they formed from a liquid phase.
Aside from the occasional presence of spinel (Table
6), inclusions contain no other nonopaque crystal-
line phases; the glasses have not devitrified. The
persistence of a glass during cooling is in agreement
with the observations by Roedder and Weiblen on
lunar olivines where, with one exception, inclusions
less than 22 microns remained all glass. Our largest
glass inclusion is somewhat larger, 36 microns.
Microprobe analyses of these glasses (Table 8)
TABLE 8.?Electron microprobe analyses of glass inclusions
in olivine
Incl. No.
la
lb
2
3
4a
4b
4c
5
6
7
8
9
10
11
Av.
SiO,
48.2
48.1
45.4
63.1
45.4
43.6
54.7
44.9
53.1
49.4
74.2
57.3
47.7
53.7
52.1
CaO
17.0
21.3
21.6
10.7
22.9
23.8
17.3
20.4
17.4
21.1
9.0
17.4
20.9
12.8
18.1
A12O3
27.4
21.6
25.7
12.9
25.2
21.0
25.0
28.0
23.7
24.1
8.7
17.3
24.6
16.5
21.6
FeO
0.2
0.3
0.2
7.1
0.1
0.2
0.1
0.2
0.2
0.1
2.0
0.1
0.2
6.8
1.3
MgO
6.7
3.6
4.6
2.6
3.3
9.4
1.6
3.8
4.0
3.7
2.2
7.5
3.7
1 .7
4.2
MnO
0
0
0
0.1
0
0.03
0
0
0
0
0.1
0.03
0.03
0.1
0.03
K20
0
0
0
0
0
0
0
0
0
0
0.2
0.1
0
0.4
0.05
Na2O
0.01
0.01
0.01
0.3
0
0
0.2
0
0
0.03
1.4
0.5
0.03
1.8
0.3
TiO2
0.9
1.1
0.9
0.6
1.0
2.1
0.3
0.9
0.6
0.6
0.4
0.5
0.8
0.8
0.8
Cr2O3
0.2
0.1
0.2
0.2
0.2
0.1
0
0.2
0.3
0.1
0.2
0.3
0.3
0.03
0.2
Sura
100.6
96.1
98.6
97.6
98.1
100.2
99.2
98.4
99.3
99.1
98.4
101.0
98.3
94.6
98.7
Notes: (1) la and lb are 2 separated inclusions in one olivine grain; 4a is
a homogeneous inclusion which is isolated from the intergrowths
of 4b and 4c in the same olivine grain; remainder are single inclusions
in different olivine grains.
(2) The fayalite content (in mol. I) of host olivines is 0-1 except for
the following: No. 3 (25-45), No. 8 (3-8), No. 11 (23-35).
show wide variability in composition, though indi-
vidually they are of uniform composition. One
bleb, however, contained two glass phases exhibit-
ing distinctive reflectivites in the section. Their
compositions are given as 4b and 4c in Table 8.
Both phases are intergrown. Compositional features
common to most of the individual analyses are the
high Ca and Al and low alkali contents and thus
are very different from the alkali-rich glasses ob-
served in unequilibrated ordinary chondrites
(Kurat, 1967; Fuchs, 1968; van Schmus, 1967). Ca
and Al rich inclusions are found in C3 chondrites,
but consist of several crystalline phases. The en-
richment of Ca and Al within small areas in both
C2 and C3 meteorites suggests a common source
for this material in both types of meteorites, but
not necessarily a genetic relation between them.
Since the melt inclusions in Murchison olivines
differ in composition, it might appear that they did
not originate from a common source; this conclu-
sion, however, may not be valid. The average
weight ratio of Ca/Al of all glasses is 1.13, which
16 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
is not too different from that of 1.09 for all chon-
drites and basaltic achondrites (Ahrens, 1970).
Thus these glasses may be primordial condensates
from a gas cloud that supplied all Ca and Al to
meteoritic matter. Further implications will be
discussed later under "Summary and Discussion of
Observations."
It should be pointed out, incidentally, that there
is a relationship between the FeO content of a
given glass bleb and its host olivine. Glasses con-
taining less than 0.3 wt. % FeO occur in unzoned,
iron-poor olivines. The three relatively FeO-rich
glasses occur in zoned olivines, the average FeO
content of which increases with that in the glass.
This correlation probably reflects the FeO environ-
ment of the source material and does not result
from an in situ diffusion of iron between inclusion
and host. Microprobe traverses (1 micron beam
diameter) show that no compositional gradients
exist between host and inclusion. Abrupt changes
in composition are confined to bands 2 microns
wide. Zoning in host olivines is not influenced by
the presence of inclusions.
Melt inclusions in minerals from igneous rocks
may be formed by trapping fluid during fractional
crystallization, the composition of a melt inclusion
is then in equilibrium with its host at the tempera-
ture of the liquidus. Moreover, all inclusions within
a single host should have the same composition.
With regard to the latter condition, we have data
for two such olivines, both of which contain glasses
that differ in composition (glasses 1 and 4, Table
8). While it is possible that an olivine crystal grow-
ing in a magma could trap melt inclusions of
different compositions, we can find no other evidence
for a magmatic origin for these olivines.
A few preliminary heating experiments on
olivines with unanalyzed inclusions were performed
at 100?C intervals from 1200 to 1400?C for periods
of several hours followed by air-quenching to room
temperature. About six inclusions within each of
two crystals were observed microscopically and
photographed before and after each heating period.
In a few cases, crystals of unknown compositions
appeared in glass (which initially was all glass)
and then disappeared after heating to higher tem-
peratures. Either solution of some olivine had
occurred to change the composition of the inclusion
(some inclusions did become larger), or some an-
nealing occurred and was erased at the higher tem-
perature. The lowest liquidus observed was be-
tween 1200 and 1300?C as indicated by a shift in
bubble position, but bubbles in other glasses in
the same olivine did not move even after 1400?C.
These observations support the compositional evi-
dence above that different inclusions within an
olivine crystal have different compositions. Further-
more, their compositions in some cases would have
been altered if they coexisted with their host under
these heating conditions.
Phase diagrams can be examined to test for in-
clusion?host equilibria. The normative composi-
tion (wt. %) for the average for all glasses in
Table 8 is 60% anorthite, 27% diopside, and 13%
silica. But because the compositional range is
appreciable, each can be treated separately. For
simplicity FeO is added to MgO. The major com-
ponent oxides then lie within the quaternary
system CaO-Al2O3-MgO-SiO2. The diagrams at one
atmosphere of Prince (1954) and Morey (1964)
are applicable for a few compositions, but for most
compositions the diagrams of Osborn et al. (1954)
are more appropriate as they present planes in the
tetrahedron for constant alumina contents in in-
crements of 5% in the range from 5 to 35%. The
major conclusion deduced from these glass com-
positions is that they are not in equilibrium with
olivine at liquidus temperatures and therefore can-
not represent compositions of melts from which
olivine crystallized. There is no olivine primary
phase field in planes with constant alumina equal
to or greater than 25%. Eight of the glasses lie
close to or within these planes. Three glasses (lb,
4b, 9) lie within the anorthite field; cristobalite or
tridymite are primary crystallates for glasses 3 and
8, and number 11 has too low a summation for
consideration.
Two melt inclusions (la and lb) contain euhe-
dral spinel crystals. The question arises whether
spinel crystallized in situ from a melt higher in
MgO and A12O3 than the present melt composi-
tions. If this be true, then the initial composition
of la (for example), adjusted to include the
amount of observed spinel (33 wt. percent), be-
comes in wt. percent: SiO2 32.1, CaO 11.8, A12O3
40.7, FeO 0.4, MgO 13.7, TiO2 0.7, and Cr2O3 1.1.
This composition has not been observed for any
spinel-free inclusion. Furthermore, euhedral spinels
(Fe- and Cr-free) are found within olivine not asso-
ciated with a melt inclusion. This suggests that
NUMBER 10 17
euhedral spinel predated formation of melt inclu-
sions and olivine. Early-formed spinel could have
nucleated later condensates.
As was mentioned earlier, in appearance these
glassy inclusions resemble the melt inclusions
found in olivines that crystallized from magmas.
Roedder (personal communication) in a prelimi-
nary review of our observations believes that both
host and inclusions can be explained as being de-
rived from a magma. We suggest an alternate
hypothesis?that the glassy inclusions were frozen
liquid droplets incorporated into olivine, which it-
self did not crystallize from a magma. Following
are several reasons for this suggestion:
1. Single and relatively large olivine crystals in the
matrix are not associated with any other adhering
crystalline material. A meteorite type consisting of
only olivines (most of which are practically Fe-
free) and contain melt inclusions is unknown.
2. Aggregates of olivine grains that may contain
glassy inclusions are loosely consolidated. This state
of aggregation appears contrary to a magmatic
origin.
3. We have noted olivines with inclusions within
1-2 mm chondrules in the Allende C3 meteorite.
The inclusions are similar in appearance to those
in olivines from both igneous rocks and to those
in Murchison. Similarity in appearance, however,
does not necessarily signify a common mode of
origin. Clearly, chondrules solidified as discrete
systems of millimeter dimensions and did not
crystallize in a magma chamber.
4. The C2s, next to the Cls, contain elemental
abundances that approximate solar abundances.
Most, if not all, of the components of C2s must
have condensed from primordial matter. Thus,
their silicates and metal must have been closely
associated with a condensation process and not
with a fractionating magma. Further implications
of this condensation process will be discussed later
in the summary.
POORLY CHARACTERIZED PHASES
Two ubiquitous minor phases are present in all
polished sections studied. Both are soft and ex-
hibit a dull beige color in reflected light, but they
differ considerably in form and mode of occurrence.
The first of these is massive in appearance, mot-
tling is usually present, masses are optically iso-
tropic though sometimes showing faint undulatory
extinction. Characteristic shapes observed vary:
(1) regular spheres resembling chondrules, (2)
angular to irregular pieces with serrated edges (Fig-
ure 8a), and (3) irregular rounded blebs (Figure
8b). The first two are most often found within
the matrix, while the blebs occur within or between
olivine grains in white inclusions. Based on shape
alone, the blebs might have replaced preexisting
metal where it was exposed to attack along frac-
tures or grain boundaries. It is typical that metal
securely protected within olivine shows no altera-
tion. On the other hand, isolated blebs of this
FIGURE 8.?Various shapes of massive Fe-S-O phase: a, ir-
regular form (0.3 mm long) in Murchison matrix. Note the
variegated texture and what appear to be drying cracks; b,
oblate and oval forms in granular olivine inclusion. Smaller
bright spots are metal. Width of field is 0.5 mm.
18 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
massive phase show no channel fillings leading to
them. If metal within these inclusions were the
precursor, then the replacement must have occurred
prior to incorporation of the inclusions in Murchi-
son, since unaltered metal coexists within the
matrix. As can be seen from the photographs, what
appear to be drying cracks are usually present.
Blebs of this phase can sometimes be removed easily
from white inclusions; they are highly magnetic
and give no X-ray powder pattern. The overall
weak magnetism of the matrix material of Murchi-
son may be due to inclusions of this phase.
In occurrences (1) and (2) above, the edges of
these masses are slightly corroded and embayed.
This could represent either of two things: direct
reaction with the matrix material or reaction with
some liquid or gaseous phase that, at some time in
the preterrestrial history of the meteorite, passed
through the permeable, fine-grained matrix and
attacked this phase or some preexisting phase. Most
of the masses show diffuse edges which may repre-
sent more advanced stages of reaction. This massive
phase occasionally contains minute blebs of metal
or troilite and pentlandite.
In a few instances patches of a highly anisotropic
phase consisting of a criss-cross matting of fibers or
plates (5 to 10 microns long) are contained within
an isotropic groundmass resembling the massive
phase. This fibrous or sometimes platy phase is
the second poorly characterized phase seen in
Murchison. Some typical occurrences are shown in
Figure 9.
The anisotropy and reflectivity of these fibers in
polished sections approaches that of graphite and
may easily be mistaken for it. X-ray patterns show
one sharp line at 5.4 A and a weaker line at 2.72
A. The pattern could not be identified.
It is not clear if these two phases are related. It
is possible the massive material represents an alter-
ation of the fibrous material, or the converse, that
the fibers represent a crystallization from the amor-
phous phase. On the other hand a cluster of fibers
do occur alone within the matrix in one instance,
and in several other instances clusters of fibers were
found rimming and projecting into large calcite
crystals (Figure 9c). The fibers are clearly much
lower in abundance than the massive, amorphous
phase.
Many of the properties we have noted for these
poorly characterized phases resemble those de-
tv.
FIGURE 9.?Photographs of unidentified highly anisotropic
fibrous and platy phase. Oil immersion, reflected light: a,
fibers and plates embedded in matrix. Scale bar is 20 mi-
crons; b, interlocking plates within matrix. Scale bar is 20
microns; c, fibers within and bordering polysynthetically
twinned calcite crystal.
NUMBER 10 19
scribed for several types of an unknown phase (or
phases?) noted by Prof. Ramdohr (1963) in the
Mighei C2 meteorite. He suggested that it might
be a new Fe-C-S mineral with low Ni content and
that its low hardness indicated a layered structure.
Our X-ray results for the fibrous or platy phase
would appear to confirm a layered structure, but
we have not attempted to gather any compositional
data because of the narrow width (1-2/*) of the
fibers and platelets. Our probe results for the
massive phase indicates a more complicated compo-
sition.
Averaged microprobe results for the massive
phase are, in weight percent (with ranges, where
available, in parentheses): S 17.6 (16-19), Fe 42.0
(37-45), Ni 6.1 (5.9-8.2), Cr 1.6 (1.4-2.5), C 0.2,
P. 0.3, O 31.0, sum = 99.7. Approximately fifteen
areas of this phase on two separate probe sections
were analyzed. Wavelength scanning showed no
other elements present in detectable amounts (spe-
cifically no Cl, F, or Mg and only a faint trace of
Si). The fact that the sum is close to 100% is
fortuitous, as our accuracy for oxygen was low. The
result for oxygen may at best be semiquantitative,
but even this is significant. The stoichiometry does
not fit that of any known oxysalts of sulfur: the
atomic Fe+Ni + Cr to sulfur ratio is 1.6, whereas
most known compounds are unity or less. Some
hydrated oxysulfates exist with similar stoichiome-
tries but none close enough to draw any con-
clusions. This massive phase was determined to be
water insoluble, hence it does not contribute to
the soluble sulfate content of the bulk chemical
analysis of Murchison.
Regardless of what kind of compound this repre-
sents, the association of Fe with both Ni and Cr
make it impossible to postulate a simple altera-
tion of troilite or pentlandite.
In the heating experiments on the matrix ma-
terial described earlier ("X-ray Evidence" and Ta-
ble 4), we note that this fibrous phase appeared
to break down to troilite. The steps in the reaction
could not be followed; they could involve reaction
with the surrounding matrix rather than straight
thermal decomposition.
Textural Features
In the "General Description" we distinguished
four distinct inclusion types. Each of these types
exhibits features that bear upon the preterrestrial
history of the whole meteorite.
TYPE 1 INCLUSIONS
This type of inclusion consists of (1) fragmental
pieces and single crystals of individual minerals
contained directly within the black matrix, and (2)
individual mineral phases that appear to have
grown within the black matrix. In the first category
the minerals are clearly debris that was incorpo-
rated into the matrix. In all, these debris frag-
ments comprise approximately 8 vol. % of the
meteorite.
The majority of type 1 inclusions are in the first
category above, and are represented by euhedral
and angular olivines, angular pyroxenes, angular
chromites, and angular pieces of glass. Minerals
in the second category are calcite, gypsum, magne-
tite, and the two poorly characterized phases noted
earlier.
The olivines and pyroxenes vary in composition
from grain to grain, and a small percentage of the
olivines are moderately to strongly zoned. A micro-
probe grain count of the homogeneous (i.e., un-
zoned) olivine and pyroxene fragments indicates a
ratio of olivine to pyroxene of approximately 2/1.
The olivine fragments average larger in size than
the pyroxenes so the actual volume ratio will be
larger than this. Histograms of these olivine and
pyroxene compositions are given in Figure 10. The
diagrams are closely similar to the ones determined
by Fredriksson and Keil (1964) for the Murray
meteorite. Wood (1967a) made an extensive sur-
vey of this kind for all inclusions in ten C2
chondrites. He discussed some subtleties in the dis-
tribution from one meteorite to another; however,
the overall patterns are the same as for Murchison.
Figure 11 illustrates a special grain to grain sur-
vey that was made on an unusually metal-rich
section. Grains of olivine not associated with grains
of metal show the same kind of histogram as just
discussed above. Olivines associated with metal,
however, show an absence of fayalitic compositions.
Fredriksson and Keil (1964) suggested for the Mur-
ray C2 meteorite that metal was formed by a Prior's
Rule?type of process by reduction of the fayalitic
component of olivine to form metal and the more
forsteritic olivine. As compelling as this argument
20 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
is, it meets with some serious mass balance prob-
lems. In Murchison, for example, a fayalitic olivine
of 30-40% Fa contains about 0.1% NiO (Figure
18, and "Nickel in Olivines"). Reduction of the
FeO and NiO from this olivine would yield asso-
ciated metal containing only 0.4% Ni, which is
over an order of magnitude smaller than that ob-
served in any C2 meteorite (Wood, 1967). Metal
in Murchison, for example, ranges from 4% to 7%
nickel. Thus, the nickel balance does not work
out. In addition, the ratio of masses of metal to
forsterite in associated grains shows wide variation
with, in extreme cases, grains consisting of up to
50% (by vol.) of metal. It appears that reduction
in situ does not account for the metal-olivine asso-
OLIVINES
oo
E
3
5
0
10
0
20
10
0
30
20
10
0
0
ciations. It would appear more likely that they are
the result of the initial condensation process from
the solar nebula. As will be discussed in the sum-
mary, depending on the total (hydrogen) pressure,
metal condensation accompanies or slightly pre-
cedes the condensation of the first forsteritic oli-
vine. Thus, metal grains would be associated with,
and possibly acting as nuclei for, the precipitation
of forsterite. As condensation continues metal
ceases to form, while small amounts of more faya-
litic olivine, not associated with metal, condenses.
Fragments of zoned olivines were found that
showed zonal variations from as little as 5 mole
% Fa to as much as 40 mole % Fa. It should be
noted, however, these results are largely from frag-
PYROXENES
ISOLATED GRAINS in C3 INCLUSION
CLUSTERED GRAINS in C3 INCLUSION
GRAINS in MURCHISON CHONDRULES
ISOLATED GRAINS in MURCHISON MATRIX
10 20 30 40 50 580 10 20
Mole % Fe2Si04 | Mole % FeSiO3
FIGURE 10.?Histograms of olivine and pyroxene compositions. A few more Fe-rich compositions
(up to Fags) are omitted for brevity.
NUMBER 10 21
ments so that it is impossible to determine what
is rim and what is core, or indeed if the full varia-
tion of the original crystal is represented in the
fragment that was encountered. In many cases
fayalite values did not show a concentric arrange-
ment but a monotonic change from one Mg-rich
edge across the fragment to the opposite Fe-rich
edge. Such asymmetric changes show that Fe-rich
edges cannot be the result of interaction with the
high iron matrix material. In some instances, how-
ever, concentric symmetry was observed with a
Mg-rich center and Fe-rich rims. In a few rare cases
reverse zoning was observed, with cores more Fe-
rich than the rims!
One of the striking features observed in this type
OLIVINES
of inclusion is the presence of rare whole single
crystals of olivine in the matrix (Figure 12). In
many instances these crystals are perfectly euhedral,
showing a full set of terminating faces. They range
in size from 0.3 mm to f mm and are either equant
in shape or somewhat flattened prisms. This is, as
far as we know, the first reported occurrence of this
phenomenon in any C2 chondrite. Specimens of
Murray and Cold Bokkeveld we examined, how-
ever, exhibited the same kinds of olivine crystals.
It may be this is a characteristic feature of other
C2 chondrites. In some instances, otherwise perfect
crystals are broken or chipped. The crystals are
never attached to other grains but are isolated in
the matrix such that their removal leaves behind a
PYROXENES
-O
30
20
10
0
40
30
20
10 -
0
GRAINS ASSOCIATED WITH METAL
GRAINS NOT ASSOCIATED WITH METAL
10 20 30 40
Mole % Fe2Si04
50 580 10 20
Mole % FeSiO3
FIGURE 11.?Histograms of olivine and pyroxene compositions in Murchison. These data were
obtained on a metal-rich section different from that used for Figure 10.
22 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 12.?Euhedral olivine crystal (0.5 mm) embedded in
matrix.
smooth-walled negative crystal cavity. An interest-
ing feature of these crystals is that the faces always
appear slightly frosted.
Some of these crystals are zoned. Measurements
of four of them showed core to rim values from Fa
25 to Fa 45, 3 to 8, 22 to 31, and 23 to 35. Other
crystals, however, were unzoned and showed con-
sistent compositions of less than Fa 1. Of the
crystals that were selected for microprobe work and
study a few contained internal blebs of glass (up
to 36 microns), spherules of metal, and euhedral
crystals of spinel. These glass inclusions were de-
scribed in detail under "Glasses."
It is clear these euhedral crystals did not form
within the matrix. Olivine is not a mineral that
develops crystal faces against impeding matrix as
do minerals like staurolite and garnet. Euhedral
olivines are found in terrestrial rocks only in in-
stances where they crystallize within a melt (e.g.,
magnesian olivines poikilitically contained in py-
roxene or plagioclase in some basalts, troctolite, or
peridotites) or subhedral crystals projecting into
open cavities (as fayalitic olivines in some volcanic
obsidians and trachytes, or coarse gabbro pegma-
tites). Furthermore, some of the fragments of oli-
vine within the Murchison matrix show several
crystal faces still present on them. These were
originally fully euhedral crystals that were broken
and then incorporated into the black matrix. We
believe it highly likely that all of the angular
olivine fragments (most of which show no crystal
faces) are pieces of larger, once euhedral, crystals.
Only a small percentage of the olivine crystals man-
aged to be incorporated into the Murchison matrix
before being fragmented.
Single crystal X-ray patterns made of these
euhedral olivines are extremely sharp, indicative of
a high degree of order and, by inference, relatively
slow cooling. They are clearly not quench prod-
ucts.
TYPE 2 INCLUSIONS
The most prominent inclusions in Murchison are
these so-called "white inclusions." Because they
tend to be rounded in outline, they are frequently
called "chondrules" by casual observers. Closer
study, however, clearly distinguished them from
true chondrules, and the presence of true chon-
drules in Murchison (type 3 inclusions) further
accentuates the distinction.
"White" is somewhat of a misnomer. The great
majority of the inclusions are snow-white, how-
ever, a few are blue or spinach green. Type 2 in-
clusions range in size from approximately 0.5 mm
to 4.5 mm and are generally rounded to elliptical
in cross section, though a high proportion of them
have intricate, crenulated or lobate outlines (Fig-
ure 13). In a few hand specimens we observed a
large number of elongated lenslike shapes that
showed a slight lineation pattern suggestive of solid
flow.
The boundaries of type 2 inclusions are gener-
ally sharp, although on a fine scale grains at the
edges jut out into the black matrix. Often the
matrix rimming an inclusion is unusually black
and fine grained. In no case is it possible to work
loose one of these inclusions from the matrix and
remove it as a whole unit, as can be done with true
chondrules.
Most of the type 2 inclusions are quite granular
in texture, the grains being only loosely packed
together. A needle gently raked over the surface
of one easily produces a small pile of grains. Most
of these inclusions consist of clusters of anhedral to
subhedral grains of usually more than one phase.
The dominant phase is highly magnesian olivine
(Figure 10), and it is present in almost all of these
inclusions. In rare instances, however, inclusions
are found that contain no olivine at all.
The most common inclusion consists of sugary
NUMBER 10 23
FIGURE 13.?Typical white inclusions in matrix: a, inclusion
(1.7 mm wide) surrounded by a narrow band (0.2 mm wide)
of fine-grained, inclusion-free matrix; b, two white inclusions,
each is surrounded by a narrow band of fine-grained matrix.
Long dimension of oval-shaped inclusion is 1.7 mm; c, pol-
ished section showing granular white inclusion surrounded
by fine-grained matrix. Width of field is 1.5 mm.
textured, granular, white forsterite which may con-
tain blebs of metal. Within a given inclusion the
forsterite shows some variations in composition
from grain to grain. The average variation we
have observed is from Fa 1 to Fa 5, although in a
few instances fayalite contents up to Fa 60 were
observed. Accessory spinel, calcite, diopside, troi-
lite, pentlandite, whewellite, and chromite may be
present as intergranular phases in type 2 inclusions.
Some of these inclusions contain appreciable
amounts of low calcium pyroxene with the olivine,
and in some instances the pyroxene exceeds the
olivine. Other, less abundant, inclusions consist of
white forsterite laced through with flakes of the
spinach phase (layer-lattice silicate; see "Other
Layer-Lattice Silicate"). In a few cases such in-
clusions exhibited the peculiar, regular interleaved
olivine-spinach texture noted earlier, in which it is
possible to break the plates of olivine and spinach
from each other like separating pages of a book. A
small number of inclusions consisted entirely of
clusters of the spinach phase. We did not, how-
ever, observe a full range of proportions from all
olivine to all spinach. The spinach was, in all
cases, iron-rich (approximately 30% FeO), while
the coexisting olivine was generally iron-poor. It
is, therefore, not possible to derive the spinach* (a
mixture of polymorphs of chamosite ? serpentine)
from in situ hydration alteration of the olivine
without invoking an improbable sequence of ca-
tion additions and subtractions.
A very small number of white inclusions con-
tained no olivine but consisted of spinel with cal-
cite, or spinel with clinopyroxene. The former rep-
resented the largest sizes of white inclusions ever
seen in the meteorite, up to 4.5 mm across.
Finally, blue patches in white inclusions (ap-
proximately 0.15 mm), consisting mainly of hibo-
nite, perovskite, and spinel with or without
diopside, were found in trace amounts. These high
temperature Ca-rich assemblages were described
and discussed in detail under "Hibonite" and
"Perovskite."
TYPE 3 INCLUSIONS (TRUE CHONDRULES)
True chondrules are in low abundance in Mur-
chison. Because they do not break with the matrix,
it is relatively easy to distinguish them. They
stand out as spherical or ellipsoidal projections on
24 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 14.?Assorted true chondrules which were easily sepa-
rated from matrix. Largest shown is 0.75 mm in diameter.
fresh fracture surfaces and can, with care, be
worked loose and removed (Figure 14). The ma-
jority of the chondrules are dark gray to black on
the outside, although in a few instances very small
(0.1 mm and smaller) pure white spheres occur in
the matrix.
The white spheres are mechanically weak and
break open at attempts to remove them from the
matrix. Internally they consist of very fine-grained,
silky white shreds of either clinoenstatite or or-
thoenstatite (almost iron-free) with minor olivine
and calcite sometimes present. X-ray patterns of
the orthoenstatites showed disordered or ordered
forms in individual instances (Pollack and Ruble,
1964).
The more common black chondrules ranged
from 0.05 to 0.8 mm in diameter. Separated in-
dividual ones were mounted and sectioned. These
consisted of a variety of internal textures and min-
eralogies. Mostly they contained olivine as granu-
lar aggregates of subhedral to anhedral grains
surrounded by streaks of glass, or stringers of the
spinach phase layer-lattice silicate. Small patches
of the poorly characterized (Fe-S-O) phase are pep-
pered through some individual chondrules. Ca-
poor pyroxene was found but is rare. A microprobe
survey of olivine and pyroxene compositions shows
most of the olivine to be close to pure forsterite
with a few more iron-rich compositions, and a small
peak at Fa 58?59 (Figure 10). This peak is due to
one single chondrule that contained this composi-
tion of olivine as scattered subhedral crystals within
a matrix of a pigeonitic pyroxene (approximately
Fs 50) that contains stringers of glass. Compositions
within individual chondrules were uniform.
Most chondrules contained some minor specks of
sulfide and metal, and few were quite rich in metal
with blebs of 50 to 150 microns. Individual metal
grains are uniform in composition, but the nickel
content did vary somewhat from grain to grain
within individual chondrules, from 4.5 to 5.7% Ni.
One chondrule contained a single metal grain that
ran 9.6% Ni.
Many of the chondrules contained from minor
to major amounts of irregular patches, blebs, and
streaks of glass that were quite variable in compo-
sition from point to point. Microprobe analyses
show it to be a high Ca-Al-silicate glass that is
alkali-poor. The compositions are all within the
range of values for the glass found in type 1 in-
clusions (Table 8).
Some unique chondrules were found. One con-
sisted almost entirely of the poorly characterized
(Fe-S-O) phase discussed earlier. Another con-
sisted of small subhedral to anhedral grains of
olivine (Fo 99) surrounded by massive spinach
layer-lattice phase (Table 2b). Once again it is
clear that the layer-lattice phase could not be de-
rived by direct hydration alteration of virtually
Fe-free olivine because of the high iron content of
the layer-lattice mineral. Another chondrule was
seen that contained a fairly large, lobate void space
cavity. Small cavities were evident in several oth-
ers; however, we could never be certain they had
not been created during the sectioning process by
the plucking out of small metal grains. The large
cavity, however, is an original feature by virtue of
its unique shape and size.
It is clear that the individual true chondrules
represent much of the same range of chemical con-
ditions as observed in the types 1 and 2 inclusions
described earlier. One significant difference, how-
ever, is the smaller number of different mineral spe-
cies found in the true chondrules in contrast to the
white inclusions, especially the very high tempera-
ture phases such as perovskite, hibonite, and spinel,
which were not observed in any true chondrule.
NUMBER 10 25
a
FIGURE 15.?C3 xenolithic inclusions in Murchison: a, largest fragment found is 1.3 cm in
diameter and is contained in an 18-cm piece of Murchison; b, close-up of same fragment (note
the thin separation of fragment from surrounding matrix); c, smaller (7 mm) xenolithic frag-
ment of same C3 meteorite type found in another sample of Murchison.
26 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 16.?Polished section of C3 inclusion (type 4) shown
in Figure 156. All bright spots are troilite or pentlandite.
This unusually high-sulfide-containing inclusion is practically
metal free. Width of field is 1.5 mm.
It is significant that among the chondrules col-
lected and studied from Murchison, not one is like
the common chondrules of the ordinary chondrites:
barred olivine, radiating pyroxene-plagioclase,
porphyritic olivine. Furthermore, the glass within
Murchison chondrules is very different in its range
of compositions from the sodium-rich glasses of the
chondrules in ordinary chondrites.
TYPE 4 INCLUSIONS (XENOLITHIC FRAGMENTS)
Within Murchison we encountered xenolithic
fragments of two other different meteorite types:
(1) six fragments of one meteorite type, ranging
from a few millimeters up to 13 mm (Figure 15);
and (2) a single fragment, 1.2 mm, of yet another
meteorite type (Figure 17).
C3 CHONDRITE INCLUSIONS.?Fragments of this
meteorite type are a distinctive light blue-gray
color. The largest fragment is slightly separated
from the surrounding matrix such that it could be
completely removed as a unit. Within these frag-
ments we observed both distinct true chondrules
and inclusions, the latter with the same morphol-
ogies as the type 1 and type 2 inclusions found in
the body of Murchison proper.
The main matrix of these xenolithic pieces con-
sists of finely comminuted, submicron size olivine
and pyroxene with numerous coarser irregular
patches of abundant troilite and pentlandite (Fig-
ure 16), and a trace of ilmenite and magnetite.
Nothing can be said about the matrix olivine and
pyroxene compositions; they are too fine-grained
for microprobe analyses.
Contained within this matrix were inclusions
that consisted of single phase angular fragments,
rounded pieces, anhedral, subhedral, and (rarely)
euhedral grains of mainly olivine, with lesser Ca-
poor pyroxene. The only metal seen in any xeno-
FIGURE 17.?Second type of xenolithic inclusion in Murchi-
son: a, inclusion shown in upper right is 1.2 mm in diameter
and contains large euhedral crystals of olivine (Fo100) in a
partially devitrified groundmass (note contrast in texture
with Murchison matrix) ; b, enlarged area of inclusion showing
devitrified glass groundmass. Large crystals are olivine, small
euhedral crystal is chromite (lighter gray), bright spots are
metal and sulfide. Scale bar is 100 microns.
NUMBER 10 27
lithic fragment occurred as two micron-size grains
in an olivine inclusion.
In addition, inclusions of multigrained aggre-
gates of olivine, pyroxene, minor sulfide, and a
minor unidentified phase were found. All of these
inclusions appear analogous to the type 1 and type
2 inclusions seen in Murchison proper.
The olivines and pyroxenes in these xenolithic
fragments show wide compositional variations from
grain to grain with, however, a different distribu-
tion of compositions than in Murchison proper
(Figure 10). Many of the grains are zoned. Olivine
zones ranged from Fa 2-20 to Fa 51-57, with all
combinations of values between. Pyroxenes ranged
from Fs 2-8 to Fs 15-22, with other values in be-
tween.
Most of the inclusions in the xenolithic frag-
ments were rimmed by nonopaque, ultra fine-
grained matrix (of submicron olivine and
pyroxene) that is virtually free of sulfides and
coarser grained silicate fragments.
True chondrules are present in the xenolithic
fragments; however, they are rare. Only a few
were observed on the largest fragment. No detailed
study was made of them.
It is worthwhile to highlight the complete ab-
sence of any layer-lattice silicate in the xenolithic
meteorite fragments. At the contacts of the frag-
ments with the surrounding Murchison matrix
there is a complete break, with no interfingering of
layer-lattice material into xenolithic fragments.
Thus it is impossible that the hydrated layer-lattice
phases in Murchison proper are derived from the
alteration of olivine in situ. If that were the case
then the olivines in the xenolithic fragments
would have been pervasively altered also, or at least
attacked at their rims in contact with Murchison
proper. The very smallest xenolithic fragments
(approximately 1 mm) show just as sharp boun-
daries with the black matrix of Murchison as do
the larger ones.
In addition, the xenoliths contain no other low
temperature phases like those observed in the sur-
rounding Murchison proper: calcite, whewellite,
gypsum, poorly characterized phases.
Macroscopically these appear to be xenoliths of
a C3 chondrite. Comparisons were made with speci-
mens of a number of C3s, of both the Ornans and
Vigarano subtypes (Van Schmus, 1969). In general
appearance the fragments looked more like the
Ornans subtype; however, the very high ratio of
matrix to inclusions plus chondrules and the non-
opacity of the matrix material contrast markedly
with the Ornans-subtype as examined by us directly
and as described and characterized by Van Schmus.
Furthermore the almost complete absence of
metal, and the abundance of sulfides adds to the
problem of comparison. Sulfides make up about 7
vol. %, half of which was troilite and the other
half pentlandite.
A partial wet chemical analysis was performed:
Mg = 14.5%, Fe = 25.2%, Ni = 1.55%. Mason
(1971) noted some chemical groupings of Vigarano
and Ornans subtypes on a plot of Mg against Fe.
A plot of the values given above lies exactly be-
tween the two groupings as illustrated in Mason's
paper. On this basis it is not possible to put it
distinctly into either group.
In the histograms of olivine compositional vari-
ations for C3 subtypes presented by Van Schmus
(1969), these xenolithic fragments (Figure 10)
show their main peak at 5-10 mole % Fa, which is
similar to the Kainsaz meteorite in the Ornans
subtype group.
We conclude that the xenolithic fragments in
Murchison are of a C3 chondrite that is similar to
the Ornans subtype in terms of olivine-pyroxene
compositional distribution, is intermediate between
the Ornans and Vigarano subtypes in terms of the
bulk composition, and is unlike either subtype in
terms of texture.
G. Mueller (1966) noted the presence of an angu-
lar xenolithic inclusion in Mighei C2 chondrite.
Megascopically it appeared similar to a C3 chon-
drite. It was, however, a darker colored fragment
and contrasted less markedly with the surrounding
Mighei matrix than the fragments in Murchison
(G. Mueller, pers. comm,). It is possible that C3
inclusions in C2 meteorites is more widespread
than heretofore imagined.
XENOLITH OF UNKNOWN METEORITE TYPE.?One
section of Murchison contains a single small (1.2
mm) xenolithic fragment of a second meteorite type
(Figure 17). It consists of blocky, jointed pieces
(some subhedral crystals) of pure forsterite set in
a matrix of glass, with traces of troilite, metal, and
euhedral chromite. The glass is partially devitrified
into a dendritic pattern of submicron needles.
In texture and mineralogy this fragment is iden-
tical to a unique chondrule observed and photo-
28 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
graphed by Ramdohr (1966) in the Adelie Land
ordinary chondrite (L5). He referred to it as an
"unknown (or nearly so) type" of chondrule. This
occurrence in Murchison may be a fragment of this
rare type of chondrule.
all zones regardless of fayalite content. As an ex-
ample, CaO = 0.26% as fayalite goes from 36 to
48. Fredriksson and Keil (1964) observed no CaO-
fayalite correlation in olivines of the Murray C2
chondrite.
Mineralogical-Compositional Relationships
In studying the Murchison meteorite a number
of relationships were observed between mineral
phases and between compositional components of
individual mineral phases, which bear on the his-
tory of the meteorite.
CALCIUM IN OLIVINES
Weight percent CaO was obtained for a group
of 21 olivines, of which 17 were homogeneous and
4 were zoned. These were chosen at random and
included olivines in both type 1 and type 2 in-
clusions. Fayalite contents ranged from 0 to 85,
with a distribution like that shown in Figure 10.
CaO contents ranged from 0.08 to 0.77%. No cor-
relation exists between CaO and fayalite contents
for the homogeneous olivine grains. For the zoned
olivines one crystal showed a positive correlation:
CaO increased from 0.15 to 0.48% as fayalite con-
tent increased from 54 to 85. The other zoned
crystals, however, show constant CaO contents in
CHROMIUM IN OLIVINES
Twelve random olivines (9 homogeneous, 3
zoned) were analyzed for Cr2O3 values. These
ranged from 0.06 to 0.3 wt. %. Homogeneous oli-
vines showed no correlation between chromium
and fayalite contents. Similarly, chromium re-
mained constant across all zones in zoned crystals
(e.g., Cr2O3 = 0.3% for fayalite from 25 to 35).
MANGANESE IN OLIVINES
Thirteen random olivines (9 homogeneous, 4
zoned) were analyzed for MnO. These ranged from
zero to 0.3 wt. %. Although there was a slight
tendency for MnO to increase with fayalite con-
tent, the data were too few to draw a definite con-
elusion. Such a correlation would be likely because
of the closely similar behavior of iron and manga-
nese in most sets of conditions. Terrestrial (crustal)
and most meteoritic conditions would not be ex-
pected to show any substantial difference in this
trend (Simkin and Smith, 1970).
0.20-
0.1 5 -
0.10 -
0.05 -
20 30 40
Mo I. % Fayalite
50 60 70
FIGURE 18.?NiO vs fayalite content in olivine grains in Murchison.
NUMBER 10 29
NICKEL IN OLIVINES
Nickel analyses were performed on 31 olivines
randomly selected from one polished section o?
Murchison (Figure 18). A correlation exists: NiO
increases with the fayalite content. As fayalite goes
from essentially zero to 80 mole percent, NiO goes
from less than 50 ppm up to 2300 ppm. There is
some scatter to the data, part of which is analytical
error, and a smaller portion of which is due to er-
rors in always returning to the same grain to ana-
lyze for both Fe and Ni, since both were done on
the same microprobe spectrometer channel, with all
the Fe counts obtained first. The correlation is,
nevertheless, quite evident and is exactly opposite
to the correlation observed in terrestrial rocks by
Simkin and Smith (1970).
The explanation of the opposite correlation is
clear from the calculations of R. Mueller (1965) in
the system Fe-MgO-SiO2-O2. In terrestrial rocks, oli-
vines (and coexisting pyroxenes) coexist at equilib-
rium with ferrite phases such as magnetites or
chromites but not metal. With increasing oxida-
tion state the fayalite content of the olivine de-
creases as iron is oxidized from the + 2 to the + 3
state into ferrite phases. Since the oxidation po-
tential of nickel beyond the +2 state is extremely
high it becomes progressively enriched in the oli-
vine in which it forms an ideal solid solution
(Campbell and Roeder, 1968). Thus, Mg-rich ter-
restrial olivines contain more NiO as observed by
Simkin and Smith (1970).
In a more reducing environment than the ter-
restrial crust, at oxygen partial pressures less than
those of the ferrite-metal boundary, just the reverse
happens. With decreasing oxidation state the faya-
lite content of the olivine (and pyroxene) decreases
as iron is reduced into the metallic state (or a re-
duced sulfide phase such as troilite). Because the
oxidation potential of nickel from the +2 state to
the neutral state is higher than that of iron (from
+2 to neutral), nickel is more readily reduced than
iron. Thus, for oxidation states that do reduce
iron, nickel is reduced along with it. As the oxida-
tion state drops the olivine becomes more Mg-rich
and poorer simultaneously in nickel.
It is clear from studies on equilibrated chon-
drites that the fayalite content decreases with lower
oxidation state and that Mg-rich (Ni-poor) silicates,
such as olivines and pyroxenes in iron meteorites,
TABLE 9. Equilibrium condensation sequence of phases ob-
served in Murchison (total pressure 10-* atm) *
Phase
Hibonite
Perovskite
Spinel
Metal
Diopsi de
Fors ten' te
Ertstati te
Intermediate compositions
of olivines and pyroxenes
Troilite (+ pentlandite)
Hagneti te
Layer-lattice phases
Poorly-characterized phases
Whewel1i te
Calci te
T?C (rounded values)
1500 (approx. )
1375
1240
1200
1180
1170
1075
700-250
425
130
<150
?1 00
<<100
<<1 00
Most data after Grossman, 1972. Temperatures of phases
condensing below 150?C (except magnetite) are estimated
by the authors.
olivines in pallasites, and pyroxenes (and rare
olivine) in enstatite chondrites and enstatite achon-
drites, represent reduced environments (and
generally higher temperature environments than
more oxidized meteorites). The olivine and py-
roxene fragments found in the matrix of Murchi-
son clearly did not originate there in equilibrium
with each other (because of the large grain to
grain variations in composition, and the clastic tex-
ture of these inclusions). They represent a variety
of oxidation conditions. By virtue of the correla-
tion we observe (Figure 18) these conditions were
established in equilibrium with metal and not in
equilibrium with the ferrite phases which occur in
small amounts also in the matrix and inclusions of
Murchison. On the basis of the kind of condensa-
tion sequence which formed the minerals of Mur-
chison, metal and olivine formed closely together
in temperature, while ferrites formed about 1000?C
lower. This condensation sequence will be dis-
cussed in detail in the summary (also cf. Table 9).
Figure 19 is a plot of the NiO content in olivines
from one of the xenolithic fragments of the C3
chondrite found in Murchison. It shows the same
trend as for Murchison proper. From our earlier
description (C3 Chondrite Inclusions) this meteor-
ite is also disequilibrated and clastic-textured. The
almost complete absence of metal and the high
amounts of troilite and pentlandite indicates it is
30 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
0.20
O 0.15
0.10
0.05
? ?????-
10 20 50 60
Mo I. % Foyolite
FIGURE 19.?NiO vs fayalite content in olivine grains in C3 xenolithic inclusion.
70
an agglomeration of grains from a higher sulfur
environment than for Murchison proper. Thus,
most of the iron and nickel reduced out of the sili-
cate phases (for the Mg-rich reduced compositions,
which dominate the histogram, Figure 10) went
into iron and nickel-rich sulfide phases.
NICKEL IN PYROXENES
In doing the survey of NiO in olivines of Murchi-
son, pyroxene grains were also examined. No de-
tectable NiO was found in any of them. This is
logical in view of the histogram (Figure 10): vir-
tually all the pyroxenes are very Mg-rich, hence
very reduced and thereby Ni-poor.
CHROMIUM AND PHOSPHORUS IN METAL
The high Cr and P contents of the metal grains
are different from all known meteoritic metal ana-
lyzed thus far. It is similar to some of the metal
grains from the lunar samples, though they never
reach the values observed in Murchison metal. Su-
perficially it would appear that the metal represents
somewhat more reducing conditions than for other
meteorites, and that Cr and P were alloyed into
early-forming metal rather than coming out in
lower temperature Fe- and Cr-rich spinels and sili-
cates, and phosphates, respectively. Certainly the fact
that almost all the metal occurs as included beads
within very iron-poor olivine would tend to bear
this out. On the other hand, the metal shows no
detectable Mn or Si. The latter would require ex-
treme reduction such as observed in the Horse
Creek iron or in the enstatite chondrites. Mn, how-
ever, would be expected to be reduced into the
metal along with Cr and P. Its solar abundance
(Goles, 1969) relative to Fe is about the same as P
and is only a factor of two less than Cr. In this
regard, metal in enstatite chondrites (Keil, 1968),
though notably high in Si, shows no Cr or P of
unusual significance. We are currently investigat-
ing the metal in a large sampling of carbonaceous
chondrites and hope to be able to draw conclusions
in a later paper. For the time being we have no
solution to the unusual compositions of the metal
grains in Murchison.
MINERALS NOT PRESENT IN MURCHISON
Perhaps as significant as the minerals that have
been found is the absence, or only trace presence,
of minerals common to the ordinary chondrites:
plagioclase feldspar, chromite, ilmenite, or phos-
phates such as whitlockite. All sections of Murchi-
son were searched carefully for these and only
chromite was found in traces. Ilmenite was found
NUMBER 10 31
as a trace within one of the C3 xenolithic in-
clusions, but not in Murchison proper.
Wet Chemical Analysis of Murchison
As in overall appearance, so too in bulk chemical
composition, Murchison is closely similar to Mur-
ray. The results obtained are shown in Table la
and compared with the published data of Jarose-
wich (1971). These data are also compared on a
volatile-free (C, H, N, O, S) basis in Table lb. On
this basis the agreement between the major con-
stituents is better since the effect of our larger
H2O (-) value is eliminated. Jarosewich has kindly
supplied a sample of his analyzed material to allow
for a comparison of some of the analytical results
(especially Na and Fe) on the sample. These data
follow:
SiO2
Fe (total)
MgO
Na2O
H2O (total)
C (total)
Jarosewicli
29.07
22.15
22.29
19.94
0.23
10.05
2.12
this work
28.95
21.73
21.73
19.94
20.01
0.21 (flame emission)
0.21 (atomic-absorption)
9.66
1.96
Analysis by Jarosewich of two separate portions
of our sample confirmed our high Na value. We
have further investigated this point by determining
Na in five separate interior fragments from five dif-
ferent pieces of Murchison and found that the
Na2O content ranged from 0.19% to 0.71%. The
weighted average was 0.48%. Sodium in our ana-
lyzed sample is relatively high and variable. Such
a result could be due to contamination after fall
or to the fact that sodium is indeed heterogeneously
distributed in this meteorite.
We have not assigned values for FeS, FeO, and
Fe2O:{ in our analysis for the following reasons:
1. Sulfide sulfur is estimated by subtracting the
sum of elemental and sulfate sulfur from the total
sulfur content. There is a tendency for both the
elemental and sulfate sulfur values to be low due
to incomplete extraction of the elemental sulfur
by carbon tetrachloride or of sulfate sulfur by hot
water. This results in a high estimate of sulfide
sulfur.
2. As described earlier an appreciable amount of
sulfur is present as the so-called poorly character-
ized Fe-S-O phase, rather than as FeS. Also some
sulfur is present as pentlandite.
3. Mineralogical examination by point-counting
on polished sections shows that the total sulfide
content is much less than that calculated by sub-
tracting elemental and sulfate sulfur from the total
sulfur. Point-counting of four sections gives troilite
+ pentlandite ranging from 0.07% to 0.23% by
weight (average 0.13%).
4. Earlier, in the discussion of the layer-lattice
matrix material, crystal-chemical considerations in-
dicated an appreciable amount of ferric iron in the
matrix other than that due to the minute amount
of magnetite. These facts, along with the inherent
errors in the estimate of FeS make the assignment
of iron as FeO difficult.
Summary and Discussion of Observations
Many of the foregoing observations bear out
what has been described in C2 meteorites by others
(DuFresne and Anders, 1962; Wood, 1967; Mason,
1962-63 and 1971). Some of the observations, how-
ever, are new and raise questions regarding inter-
pretation and how they fit into existing schemes of
carbonaceous chondrite genesis. Four of these will
be summarized and discussed in this section.
First, it has been recognized for some time that
C2 meteorites possess a high temperature fraction
contained in a matrix of low temperature material.
The high temperature fraction is generally consid-
ered to be forsterite with some enstatite and spinel.
In addition to this assemblage we have observed
two other high temperature fractions: the perov-
skite-hibonite-spinel-(?)diopside assemblage; and
blebs of Ca-Al-silicate glasses, of variable composi-
tions, contained within olivine (and occasional
spinel) crystals.
The perovskite-hibonite-spinel-(?)diopside as-
semblage is closely similar to the assemblage ob-
served in some C3 chondrites (Fuchs, 1971; Keil
and Fuchs, 1971); in contrast to them, however, no
gehlenite or anorthite was found. In Murchison
the assemblage is very rare and is not found with
the frequency it is in such C3s as Allende. The
fact that it is present at all in a C2 is significant,
and together with the glasses it provides a partial
answer to the question raised by Mason (1971) re-
32 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
garding the mineralogical form of Ca and Al in
C2 meteorites.
The sequence of direct condensation of crystal-
line solids from a solar nebula was originally com-
puted by Lord (1965), later recomputed by Larimer
(1967), and most recently recalculated with a re-
vised scheme by Grossman (1972). Because of the
lack of thermodynamic data on hibonite no calcu-
lation includes it in the crystallization sequence,
although an educated guess would place its sta-
bility in the 1350?-1450?C range (Grossman, 1972).
In Grossman's calculation, at a total gas (primarily
hydrogen) pressure of 10'3 atm, perovskite con-
denses out at 1374?C, followed by melilite and
spinel at 1352?C and 1240?C, respectively. Melilite
does not occur in Murchison; however, the associa-
tion of diopside with spinel could represent the
breakdown products of melilite by later reaction
with gas at 1177?C, or the bypassing (metastably)
of the melilite field. Forsterite and enstatite do not
appear until 1171?C and 1076?C, respectively.
Thus, the perovskite-hibonite-spinel-(?) diopside
assemblage, though minor in quantity, clearly rep-
resents a higher temperature stage than heretofore
observed in any C2 meteorite. It may actually oc-
cur in other C2s, but not yet recognized.
Whether metal precedes, accompanies, or follows
the condensation of forsterite depends on variations
in the total (hydrogen) pressure, with less than an
order of magnitude controlling the sequence
(Grossman, 1972). In Murchison, most of the metal
occurs within olivine crystals, indicating pressures
above 104 atm where the metal condenses before
forsterite. The euhedral morphology (cf. Textural
Features) of many of the forsterite crystals sug-
gests condensation from a vapor phase. Metal
grains must have acted as preferred nucleation sites
for some of the olivine because the high surface
tension of metal surfaces makes them ideal nuclei.
The few zoned olivines represent disequilibrium
with the gas with falling temperature.
The form of the metal grains within olivine
crystals is always spherical or subspherical blebs,
or lobate clusters of two or more blebs. The shapes
strongly suggest the metal was originally present
in liquid drops. This has also been noted by Wood
(1967b), who drew the same conclusion. The shapes
contrast sharply with the angular metal grains
and occasionally stringy metal veins of the ordinary
chondrites. In experimental systems iron condensed
from the gas phase usually forms elongate crystals
and needles. The textural suggestion of a liquid
phase for the metal in Murchison is further raised
by the Ca-Al-silicate glass inclusions that are rela-
tively abundant within olivines (and some spinels).
As noted in "Glasses" under "Mineralogy," these
glass bleb inclusions vary in composition and oc-
casionally contain void space bubbles (Table 8,
Figure 7). In one instance a single bleb was found
that consisted of two glasses of different composi-
tions. Compositionally, the average glass computes
to normative anorthite and diopside with excess
silica. In a solar nebula condensation (at 103 atm
total pressure) crystalline anorthite should condense
out at 1089?C, after forsterite, but before enstatite.
The mineral anorthite readily crystalizes in experi-
mental systems and does occur in the high temper-
ature fraction of some C3 chondrites, as well as in
euhedral to subhedral crystals in some achondrites.
The presence of glasses of these compositions
within olivines and some spinels in Murchison
points to an early liquid stage. As discussed in the
section on glasses, one possibility is that the glass
inclusions represent trapped blebs of a coexisting
rest-magma from some magmatic stage in a plan-
etary object. We noted there, however, four major
difficulties with postulating a magmatic state that
requires complicated, highly contrived conditions
that are contrary to the primitiveness of a meteor-
itic object such as Murchison. It appears more
reasonable to suggest that these glasses are part of
the normal formational processes that condensed
primitive solid matter in the solar system, as many
others have suggested (e.g., Anders, 1971), rather
than a magmatic stage on some early planetary
object.
The variability in the compositions of the glass
blebs clearly indicates that they were not formed
in a single event, or stage, sampling a homogeneous
solar nebular composition.
The distinct indication of a liquid phase raises
the question whether the spherical habit of the
metal grains in olivines might also represent origi-
nal liquid drops. Since liquid metals cannot be
quenched to glasses, as can most silicate liquids,
there is no way to decide other than to point to
the morphology.
Wood (1963) raised the question of condensation
NUMBER 10 33
of solar nebular matter initially into a liquid stage.
The idea was generally discarded when it became
clear that unreasonably high total pressures would
be required for the solar nebular gas, of the order
of 10 to 103 atmospheres (Anders, 1971). Recently,
however, Turekian and Clark (1969) have invoked
total pressures in the range of 0.1 to 1 atm as a
means of accounting for density differences and
differences in accretion histories among the terres-
trial planets, Mercury, Earth, and Mars. Toward
the center of the primitive solar nebula both pres-
sure and temperature would increase (Cameron,
1962).
Although two of the glasses in Murchison reach
7.5% and 9.4% MgO, they are all much too de-
pleted in it. Thus, they are fractionated relative to
nebular abundances. Formation of liquids of these
compositions at pressures of 1 atm or more would
require temperatures in the range of 1350-1450?C,
as noted under "Glasses." In a condensation of
these liquids from the solar nebular gas they may
be the compositions that liquefy first. With activity
data for the elements in liquids in these compo-
sitional ranges, a calculation of the first liquid
composition to condense is possible. Such calcu-
lations are currently in progress by M. Blander
(personal communication).
The most logical sequence to account for these
glasses, as well as the apparent initial liquid drop-
lets of metal, would be for direct condensation of
liquids near the center of the solar nebula, at
pressures and temperatures commensurate with
the formation of silicate and metal liquids. Turbu-
lent conditions and solar-flare activity could cause
variation in the compositions of small regions of
space as well as the necessary fluctuations in pres-
sures to cause liquids to form. The majority of
these drops did not survive but, with falling pres-
sure, were resorbed into the gas phase. Some very
small droplets did survive, however, and were
moved outward in the ecliptic plane. There, with
falling pressure and temperature, additional re-
sorption into the gas phase would occur until the
droplets became cool enough to act essentially as
solids, or at least develop a surface skin that would
inhibit reevaporation. Hoyle (1960) described the
process of separation of an ecliptic plane disk of
matter from the proto-sun, and the transfer of
angular momentum to it. He calculated that early
condensed matter could easily migrate outward
away from the inner part of the disk, as long as the
condensed matter does not accrete into objects
above 100 cm in diameter.
Thus, these tiny quenched drops could move out-
ward into less turbulent regions where, Hoyle esti-
mates, the average pressure of matter in the gaseous
disk is 10"3 atm. Here the droplets that survive be-
come incorporated into the dominant early con-
densing solids, forsterite and spinel.
Attractive as this scheme is, it still requires at
least transient pressures of above 100 atm, held
long enough to allow some droplets to condense. In
addition, as noted under "Glasses," some of the
liquid blebs are deformed from spherical shapes
and consist of elongated blebs that lie parallel to
crystallographic directions in the olivine crystals
in which they are contained. This would mean that
the glass drops would have to remain plastic and
deformable while the temperature fell 200?C (from
approximately 1350?C, where the glasses condense,
to 1170?C at which forsterite sublimes), and the
pressure fell five to six orders of magnitude (102 or
103 down to 10-3 atm).
An alternative scheme would be for the earliest
crystalline phases, like perovskite, melilite, spinel,
hibonite, diopside, to condense from the solar
nebula at equilibrium gas pressures of 103 atm,
and then suffer a high transient heat, such as sug-
gested by Whipple (1966) and Cameron (1966).
This would melt the particles already condensed,
and they would quench to a metastable state as
glass droplets, later to be incorporated into olivine
once the normal sequence of equilibrium condensa-
tion resumed. The major problem with this scheme
is that no combination of solids that condense from
the solar gas at temperatures above that at which
forsterite condenses can form glasses of the compo-
sitions that were measured (Table 8). The silica
content is plainly too high.
This leads us to the interesting approach of
Blander and Katz (1967), in which they established
an argument that the kinetics of nucleation pro-
cesses favor the metastable nucleation of liquid
drops initially, at pressures of 103 to 104 atm.
These would form by supersaturation of the neb-
ular gas, component by component, as subcooling
took place below the temperatures where equilib-
rium solids should have crystallized out. Under such
34 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
a regime, variability in compositions among the
droplets of liquid would be a direct consequence.
Furthermore, droplets formed by such a non-
equilibrium process could maintain their plasticity,
stiffening only when they survive to lower temper-
atures (i.e., below the glass transition temperature).
As such, they could become nuclei for forsterite and
spinel that later crystallized via equilibrium con-
densation from the gas, and in individual cases still
be plastic enough to become crystallographically
elongated within the forming olivine. Under con-
ditions of supersaturation, metal would be ex-
pected to form spherical to subspherical grains
also; this would happen even if the metal con-
densed directly to a solid state. Blander (personal
communication) states that iron vapor can be sub-
cooled as much as 300?C before condensation.
This treatment of Blander and Katz offers many
attractive advantages in explaining these glasses
in Murchison. The only disadvantage is the lack
of quantitative thermodynamic data for the liquids
at the present time. For the supersaturation conden-
sation process to work for Murchison, about 100
to 200?C (at 103 atm) of subcooling would be re-
quired. Blander (personal communication) believes
this amount is probable in this process.
Second, details of the accretion process are, of
course, matters of interest and speculation. We
observed that the white (type 2) inclusions are
almost always surrounded by thin rims of ex-
ceptionally black matrix material (Figure 13). The
deep blackness is due to the complete ab-
sence of light-colored grains of olivine, pyroxene,
etc. These inclusions typically do not have a great
deal of mechanical integrity, as noted earlier. They
appear to have clumped together into commonly
ellipsoidal, but occasionally irregular, masses be-
fore incorporation into the accreting Murchison
parent body. The low temperature, submicron size
(smoke) layer-lattice material appears to have
coated the white masses forming the thin, black
rim. In a hard vacuum, surface-attractive forces of
newly formed layer-lattice phases, with high surface
area per unit mass, would be quite "sticky." The
formation of the adhering film of layer-lattice ma-
terial would act to seal each clump of coarser
grained silicates and allow each of them to act
as a unit until incorporation into the accreting
body. Even at that the low mechanical integrity of
the white (type 2) inclusions shows that the accre-
tion process for the parent body of Murchison must
have been an exceedingly gentle process; clumps
of silicates, smoke-size layer-lattice particles, true
chondrules, single crystal grains and grain frag-
ments, and fragments of at least one other body
(type 4 xenolithic inclusions) must have^been mov-
ing with near zero relative velocities, otherwise the
white inclusions would never have survived. In-
deed, the single grains of silicates in the matrix
(type 1 inclusions) may be, in part, the remains of
white clusters that were disaggregated in the ac-
cretion process.
Third, it is worthwhile to reemphasize our ob-
servation of the peculiar form of green (spinach)
layer-lattice silicate that occurs as regular inter-
leaved plates within forsterite grains (described
under "Other Layer-Lattice Silicate" and "Type 3
inclusions") (Figure 3). It should be noted that these
easily separable green plates are not connected to
the exterior black matrix material. DuFresne and
Anders (1962) observed this spinach phase within
olivine in Murray and from it concluded that oliv-
ine was pervasively altered by hydration, and these
represented the process caught short of completion.
As we have shown in great detail earlier the
layer-lattice phase (or phases) is not a simple
chemically hydrated equivalent of olivine. In addi-
tion, we wish to emphasize here the importance of
the type 4 inclusions (xenolithic fragments). These
fragments were clearly derived from other accreted
objects that were fragmented, dispersed, and later
incorporated into the Murchison body. If pervasive
hydration-alteration then attacked almost all of the
Murchison olivine to form the layer-lattice material,
these olivine-rich xenolithic inclusions should have
been altered also; they are not. In addition, Onuma
et al. (1972) on the basis of his oxygen isotope
work found "no direct genetic relationship between
any pair of the three fractions: matrix [i.e., layer-
lattice material], high-iron olivine, and low-iron
olivine." Clearly, the layer-lattice material formed
in space as a low temperature condensate at under
150?C (Anders, 1971; Onuma et al., 197,2) and
accreted into a body, accompanied by high temper-
ature mineral fragments and xenolithic fragments.
The development of the green plates of spinach
within olivine may be related to strings of elongat-
ed glass blebs arranged in parallel rows within
NUMBER 10 35
condensed olivine crystals. It these glasses were
considerably more Mg- and Fe-rich, and Ca- and Al-
poor, than the glasses discussed earlier they could
have been attacked at low temperatures (below
150?C) when the fine, smoke-size particles of layer-
lattice matrix material were condensing. Water
was the major oxygen-bearing, gas-phase com-
ponent, and it preferentially attacked these glass
layers within olivine. Such Mg- and Fe-rich, Ca-
and Al-poor glasses would have formed at somewhat
lower temperatures than the other glasses that did
survive within olivine crystals. It is, in our opinion,
highly unsatisfactory to postulate the existence of
a glass that is now completely destroyed in order
to account for the peculiar olivine-spinach texture
shown in Figure 3. Nevertheless, it is clear that the
textural relation between high temperature olivine
and the low temperature, layer-lattice spinach
phase, together within single grains, must precede
the accretion of the host olivine into the parent
body. We have shown earlier that the alteration
could not have taken place after accretion, other-
wise other olivines, both in the matrix and in the
xenolithic inclusions, would have been altered
also.
Fourth, by whatever process the mineral
components of Murchison condensed (i.e., by
equilibrium condensation or by nonequilibrium
subcooling-supersaturation-condensation), there was
a more or less continuous fractionation of mineral
phases as the nebular gas temperature fell. If we
use the equilibrium condensation model (as com-
puted by Grossman, 1972) plus estimates from
other sources (e.g., Onuma et al., 1972) as bases
for comparison only, then the estimated range of
temperatures for the entire mineral assemblage
observed in Murchison would be as shown in Table
9. Clearly, a range is represented from very high to
very low temperatures, and statements made in the
literature in the past that C2 meteorites consist of
a high temperature fraction and a low temperature
fraction are somewhat misleading. What is meant,
obviously, is that the most abundant minerals are
the forsterites and the layer-lattice material; all
others are relatively minor in quantity. This fre-
quency distribution might be interpreted as a rapid
drop in temperature from the 1000?C range to the
150?C range, but is more likely a function of bulk
composition of the nebular gas and its state of oxi-
dation with falling temperature. Clearly the cooling
was rapid enough that early formed higher tempera-
ture phases did not adjust their solid solution com-
positions continuously with falling temperature; a
small percentage developed zoned overgrowths in-
stead. The puzzling feature of the sequence of
minerals in Murchison is the complete absence of
feldspar phases, which should appear (under the
perfect equilibrium situation) in the 1100?C-700?C
interval. Their absence may be due to the same
kinetic factor that kept olivine and pyroxene from
adjusting compositions at lower temperatures. Most
of the original Ca and Al are tied up in
the early glasses and spinel. The metastable per-
sistence of these with falling temperatures would
not allow the gas to resorb enough of these elements
to establish the correct composition for condensa-
tion of feldspars.
In summary, we consider that early-formed glass
and metal condensed directly by a process like that
described by Blander and Katz (1967), at pressures
of approximately 10"3 atm and temperatures around
1200?C. Hibonite, perovskite, spinel, additional
metal, forsterite, and enstatite formed by approx-
imatedly equilibrium condensation shortly after,
from 1200?C down, so as to poikilitically include
the glass blebs and most of the metal within mainly
forsterite crystals. All lower temperature phases
formed by condensation that may or may not have
been in strict equilibrium with the cooling gas, but
might have involved subcooling, there is no way
to tell. The most puzzling mineralogical aspects
are the very low temperature phases. The spinach
layer-lattice silicate and the poorly characterized
Fe-S-O phase, together exhibit unusual textural
relations with the high temperature phases, includ-
ing low temperature partial replacement of them.
Restrictions on the thermal history of Murchison
after final consolidation can be deduced from the
stepwise heating experiments of the matrix ma-
terial as reported in "X-ray Evidence." The max-
imum temperature at which no changes were
detectable is 245?C. This compares favorably with
thermometers proposed by others for different C2s.
Mueller (1953) found this temperature to be 100?-
350?C for Cold Bokkeveld based on the thermal
decomposition of organic extracts. DuFresne and
Anders (1961) give temperatures between 200? and
300?C for Mighei based on still other considera-
36 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
tions. Thus, reasonable agreement has been
obtained for three C2s from different lines of
evidence.
C2 meteorites represent very early, and very prim-
itive, accreted matter. They may be less primitive
than Cl meteorites. No sequence in time, C1-C2-
C3, however, can be implied since Murchison con-
tains well constituted xenolithic fragments of C3
meteorites. Clearly the formation of at least one
C3 object preceded the formation of at least one
C2 object in the time sense. This suggests that, al-
though a time sequence between the carbonaceous
chondrites does not apply, a spatial sequence must
be the important factor. This would, in turn, indi-
cate spatial fractionation and inhomogeneity with
respect at least to volatiles (C, H, O, N) in the cool-
ing solar nebula.
Note: As this manuscript was being typed, a
paper, "Calcium Variations in Olivines of the
Murchison and Vigarano Meteorites," by R. Hutch-
ison and R. F. Symes, appeared in Meteoritics, 7:
23-29 (1972). Their results for the Ca contents in
Murchison olivines are in accord with those re-
ported here.
Appendix: Description of Wet Chemical Analytical Procedure
A 24-gram piece of the meteorite, free of crust,
was ground in an agate mortar to pass 50 mesh.
Portions of this material were taken for the several
determinations.
H2O- and SiO2
These oxides were determined on one gram sam-
ples. The weight loss at 105?C was measured, after
which the material was ignited, fused with sodium
carbonate and silica determined by double de-
hydration with HC1, ignition to SiO2 and volatiza-
tion with HF. Silica in the combined filtrates was
determined colorimetrically.
Total Fe, Ti, Mg, Al, Ni, Ca, Na, K, Cr,
Mn, Co, and Cu Determinations
These elements were determined on aliquots of
a solution of 6 grams of the ground sample. The
sample was dissolved in a HF-HC1-HNO3-HC1O4
acid mixture and evaporated to dryness. Repeated
additions of perchloric acid and evaporation to
near dryness with the final addition of a small
amount of hydrochloric acid yielded a water-solu-
ble residue that was diluted to volume. Aliquots of
this "master solution" were used for the various
determinations.
Total Fe, including Ti, was determined gravi-
metrically by preliminary precipitation with excess
ammonia followed by precipitation of a HC1
solution of the ammonia precipitate with cupferron.
This precipitate was ignited to constant weight at
1100?C and weighed as Fe2O3+TiO2. It was then
dissolved in HC1, iron was separated from titanium
by electrodeposition into a mercury cathode, and
Ti was determined colometrically with peroxide.
Al, Mg, and Ca were determined on a second
aliquot of the master solution. Aluminum was sep-
arated from iron by electrodeposition of the iron
into a mercury cathode. It was separated from cal-
cium and magnesium by a double ammonium hy-
droxide precipitation at pH 7. The hydroxide
was dissolved in acid, evaporated to dryness,
and heated to volatilize ammonium salts. Alumi-
num was then determined by the method of
Watts (1958). Manganese was removed by the pro-
cedure of Peck (1964) prior to the calcium de-
termination. Calcium was separated from mag-
nesium by precipitation as calcium oxalate. The
oxalate was ignited to the oxide, and this residue
was dissolved and then titrated at pH 12.5 with
0.01 molar EDTA, using Acid Alizarin Black SN
indicator (Belcher et al., 1958). Magnesium was
determined by titration with 0.1 molar EDTA at
pH 10 using Calmagite indicator in the filtrate
from the calcium separation.
Cr, Mn, Cu, Co, and Ni were determined color-
metrically on separate aliquots of the master so-
lution. Chromium was oxidized to chromate and
reacted with diphenylcarbazide. Manganese was
oxidized to permanganate using periodate. Copper
was reacted with Neo-cuproine (2,9 dimethyl-1,10
phenanthroline). Cobalt was reacted with Nitroso
R salt and nickel with dimethylglyoxime after oxi-
dation with bromine.
Na and K were determined in the master solu-
tion by atomic absorption spectroscopy (AAS).
NUMBER 10 37
Total Sulfur, Soluble Sulfur (SO3), and
Elemental Sulfur
Each of these was determined in a separate por-
tion of the sample. For total sulfur approximately
1 gram of sample was sealed in a modified Carius
tube with approximately 10 ml HC1 and 0.5 ml
HC1O4 and heated at 300?C overnight to disinte-
grate the sample and oxidize all sulfur to sulfate.
The disintegrated sample was transferred to a
Teflon beaker, HNO3, HF, and HC1 added, and
the mixture evaporated to dryness to volatilize
silica. After removal of iron by electrodeposition
into a mercury cathode and a separation of alumi-
num by double precipitation at pH 7 with NH4OH,
sulfur was determined by precipitation as barium
sulfate.
To determine soluble sulfur a 1 gm sample was
leached for about 20 minutes with 150 ml of boiling
water and then filtered and washed with about 100
ml of water. The filtrate was diluted to volume and
sulfur determined gravimetrically as barium sul-
fate in a large fraction of the diluted filtrate. The
remaining solution was shown by spectrographic
analysis to contain Ca, Mg, Na, K, and Al. These
elements (except K) were estimated more accurately
by AAS and shown to be present in the molar ratios
(with Na normalized to unity), 7.0 Mg: 2.2 Ca: 1.0
Na: 0.2 Al: 0.1 K*. These cations were equivalent
to 87 percent of the sulfate found.
Elemental sulfur was extracted from the mete-
orite in a small Soxhlet-like extractor. After evapo-
ration of the solvent the sulfur was oxidized to
sulfate and determined gravimetrically as BaSO4.
About one gram of the material was extracted
for about 1.5 hours with 100 ml carbon tetrachlo-
ride. The solvent was then evaporated to dryness,
oxidized with bromine, and sulfur determined. A
blank was run on the carbon tetrachloride. The
sample was then dried at 100?C and reextracted
* K determined by spectrograph only.
with 100 ml of carbon disulfide and sulfur deter-
mined in the residue after evaporation of the
carbon disulfide. Most of the extracted sulfur
found, 0.46%, was in the carbon tetrachloride ex-
tract while 0.025% was found in the carbon di-
sulfide extract.
Phosphorous Determined Colorimetrically as
Molybdi-vanadophosphoric Acid
About 0.2 grams of the powdered material in a
Teflon beaker was dissolved in a HNO3-HF-HC1O4
mixture and taken to dryness. The residue, free of
silica, was taken up in 6 M hydrochloric acid in
0.001 M sodium bromate, diluted to volume, and an
appropriately sized aliquot passed through a small
anion exchange column to adsorb iron. After rins-
ing sample from the column the eluate was evapor-
ated to dryness, taken up in water, and phos-
phorous determined colorimetrically.
C, H, Total H2O, and N
These were determined using a Perkin-Elmer
carbon, hydrogen, nitrogen analyzer, Model 240.
Relatively small samples, 10-15 mg., were taken
for analyses. Ignition at 900?C was employed. An
independent carbon determination was made on a
125 mg sample by ignition at 1100?C in oxygen and
manometric measurement of the evolved carbon
dioxide. This latter determination yielded a carbon
content (1.91%) slightly higher than the value
(1.85%) found using the analyzer at 900?C.
Estimation of Metallic Iron
About 0.75 gram of the sample was extracted
with 3 grams of mercuric chloride in 150 ml of
boiling water in a flask stoppered with a Bunsen
valve for about 2 hours while magnetically stirred.
The solution was then filtered through glass and
iron determined colorimetrically in the filtrate.
Literature Cited
Ahrens, L. H.
1970. The Composition of Stony Meteorites (IX) Abun-
dance Trends of the Refractory Elements in Chon-
drites, Basaltic Achondrites and Apollo 11 Fines.
Earth and Planetary Science Letters, 10:1-6.
Anders, E.
1971. Meteorites and the Early Solar System. Annual Re-
view of Astronomy and Astrophysics, 9:1-34.
Bass, M. N.
1971. Montmorillonite and Serpentine in Orgueil Me-
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
teoritc. Geochimica el Cosmochimica Ada, 35:
139-148.
Relcher, R., R. A. Close, and T. S. West
1958. The Complexometric Titration of Calcium in the
Presence of Magnesium: A Critical Study. Talanta,
1:238.
Blander, M., and J. L. Katz
1967. Condensation of Primordial Dust. Geochimica et
Cosmochimica Ada, 31:1025-1034.
Bostrom, K., and K. Fredriksson
1966. Surface Conditions of the Orgueil Meteorite Parent
Body as Indicated by Mineral Associations. Smith-
sonian Miscellaneous Collections, 151 (3) : 1-39.
Brindley, G. W.
1951. Crystal Structures of Chamosites. Mineralogical
Magazine 29:502-525.
Bunch, T. E., and K. Keil
1971. Chromite and Ilmenite in Non-chondritic Meteor-
ites. American Mineralogist, 56:146-157.
Bunch, T. E., K. Keil, and E. Olsen
1970. Mineralogy and Petrology of Silicate Inclusions in
Iron Meteorites. Contributions to Mineralogy and
Petrology, 25:297-340.
Bunch, T. E., K. Keil, and K. G. Snetsinger
1967. Chromite Composition in Relation to Chemistry
and Texture of Ordinary Chondrites. Geochimica
et Cosmochimica Ada, 31:1569-1582.
Cameron, A. G. W.
1962. The Formation of the Sun and the Planets. Icarus,
1:13-69.
1966. The Accumulation of Chondritic Material. Earth
and Planetary Science Letters, 1:93-96.
Campbell, F.E., and P. Roeder
1968. The Stability of Olivine and Pyroxene in the Ni-
Mg-Si-O System. American Mineralogist, 53:257-
268.
Carroll, D.
1970. Clay Minerals: A Guide to Their X-ray Identifica-
tion. Geological Society of America Special Paper,
126: 80 pages.
Dodd, R. T.
1969. Metamorphism of the Ordinary Chondrites: A Re-
view. Geochimica et Cosmochimica Ada, 33:161-
203.
DuFresne, E. R., and E. Anders
1961. The Record in the Meteorites, V: A Thermometer
Mineral in the Mighei Carbonaceous Chondrite.
Geochimica et Cosmochimica Ada, 23:200-208.
1962. On the Chemical Evolution of the Carbonaceous
Chondrites. Geochimica et Cosmochimica Ada,
26:1085-1114.
Ehmann, W. D., D. E. Gillum, J. W. Morgan, R. A. Nad-
karni, T. Rabagay, P. M. Santoliquido, and D. L.
Showalter
1970. Chemical Analyses of the Murchison and Lost City
Meteorites. Meteoritics, 5:131-136.
Fredriksson, K., and K. Keil
1964. The Iron, Magnesium, Calcium, and Nickel Dis-
tribution in the Murray Carbonaceous Chondrite.
Meteoritics, 2:201-217.
Freeman, E. S., and B. Carroll
1958. The Application of Thermoanalytical Techniques
to Reaction Kinetics: The Thermogravimetric
Evaluation of the Kinetics of the Decomposition of
Calcium Oxalate Monohydrate. Journal of Physical
Chemistry, 62:394-397.
Frondel, C.
1962. Ferroan Antigorite (Jenkinsite). American Min-
eralogist, 47:783-785.
Fuchs, L. H.
1968. X-ray Crystallographic Evidence for the Meteoritic
Occurrence of Nepheline. Earth and Planetary
Science Letters, 5:187-190.
1971. Occurrence of Wollastonite, Rhonite, and Andra-
dite in the Allende Meteorite. American Mineralo-
gist, 56:2053-2068.
Fuchs, L. H., K. J. Jensen, and E. Olsen
1970. Mineralogy and Composition of the Murchison
Meteorite. Meteoritics, 5:198 [Abstract].
Goldstein, J. I., and A. S. Doan
1972. The Effect of Phosphorus on the Formation of the
Widmanstatten Pattern in Iron Meteorites. Geo-
chimica et Cosmochimica Ada, 36:51-70.
Goles, G. G.
1969. Cosmic Abundances. In Wedepohl, editor, Hand-
book of Geochemistry, 1:116-133, Berlin: Springer-
Verlag.
Grossman, L.
1972. Condensation in the Primitive Solar Nebula. Geo-
chimica et Cosmochimica Ada, 36:597-620.
Hoyle, F.
1960. On the Origin of the Solar Nebula. Quarterly
Journal of the Royal Astronomical Society, 1:28-55.
Jarosewich, E.
1971. Chemical Analysis of the Murchison Meteorite.
Meteoritics, 6:49-52.
Keil, K.
1968. Mineralogical and Chemical Relationships Among
Enstatite Chondrites. Journal of Geophysical Re-
search, 73:6945-6976.
Keil, K., and L. H. Fuchs
1971. Hibonite [Ca2 (Al, Ti) 24O:W] from the Leoville and
Allende Chondritic Meteorites. Earth and Plane-
tary Science Letters, 12:184-190.
Kerridge, J. F.
1969. The Use of Selected-Area Electron Diffraction in
Meteorite Mineralogy. Pages 500-504 in Millman,
editor, Meteorite Research. Dordrecht, Holland:
D. Reidel Publishing Company.
Kurat, G.
1967. Einige Chondren aus dem Meteoriten von Mezo-
Madaras. Geochimica et Cosmochimica Ada, 31:
1843-1857.
Larimer, J. W.
1967. Chemical Fractionations in Meteorites. I: Con-
NUMBER 10 39
densation of the Elements. Geochimica et Cosmo-
chimica A eta, 31:1215-1238.
Lord, H. C.
1965. Molecular Equilibria and Condensation in a Solar
Nebula and Cool Stellar Atmospheres. Icarus,
4:279-288.
Lovering, J. F., R. W. LeMaitre, and B. W. Chappell
1971. Murchison C2 Carbonaceous Chondrite and Its
Inorganic Composition. Nature, 230:18-20.
Mason, B.
1962-1963. The Carbonaceous Chondrites. Space Science
Reviews, 1:621-646.
1971. The Carbonaceous Chondrites-A Selective Review.
Meteoritics, 6:59-70.
Miller, C. C.
1953. Anhydrous Calcium Oxalate as a Weighing Form
for Calcium. The Analyst, 78:186.
Morey, G. W.
1964. Phase Equilibrium Relations of the Common Rock-
Forming Oxides Except Water. U. S. Geological
Survey Professional Paper, 440-L:L70.
Mueller, G.
1953. The Properties and Theory of Genesis of the Car-
bonaceous Complex Within the Cold Bokevelt Me-
teorite. Geochimica et Cosmochimica Ada, 4:1-10.
1966. Significance of Inclusions in Carbonaceous Me-
teorites. Nature, 210:151-155.
Mueller, R. F.
1965. System Fe-MgO-SiO2-O2 with Applications to Ter-
restrial Rocks and Meteorites. Geochimica et Cos-
mochimica Ada, 29:967-976.
Onuma, N., R. N. Clayton, and T. K. Mayeda
1972. Oxygen Isotope Cosmothermometer. Geochimica
et Cosmochimica Acta, 36:169-188.
Osborn, E. F., R. C. DeVries, K. H. Gee, and H. M. Kraner
1954. Optimum Composition of Blast Furnace Slag as
Deduced from Liquidus Data for the Quaternary
System CaO-MgO-ALO3-SiO,. Transactions of the
American Institute of Mining and Metallurgical
Engineers, 200:33-45.
Peck, L. C.
1964. Systematic Analysis of Silicates. U. S. Geological
Survey Bulletin, 1170:66.
Pollack, S. S., and W. D Ruble
1964 X-ray Identification of Ordered and Disordered
Ortho-Enstatite. American Mineralogist, 49:983-
992.
Prince, A. T.
1954. Liquidus Relationships on 10 Percent MgO Plane
of the System Lime-Magnesia-Alumina-Silica.
American Ceramic Society Journal, 37:402-408.
Ramdohr, P.
1963. The Opaque Minerals in Stony Meteorites. Journal
of Geophysical Research, 68:2011-2036.
1966. Observations on the Opaque Ore Content of some
Meteorites, Especially from New South Wales.
Journal and Proceedings of the Royal Society of
New South Wales, 99:45-55.
Roedder, E., and P. W. Weiblen
1971. Petrology of Silicate Melt Inclusions, Apollo 11
and Apollo 12 and Terrestrial Equivalents. Geo-
chimica et Cosmochimica Acta, Supplement, 2:
507-528.
Simkin, T., and J. V. Smith
1970. Minor Element Distribution in Olivine. Journal
of Geology, 78:304-325.
Snetsinger, K. G., K. Keil, and T. E. Bunch
1967. Chromite from "Equilibrated" Chondrites. Ameri-
can Mineralogist, 52:1322-1331.
Steadman, R., and P. M. Nuttall
1963. Polymorphism in Cronstedite. Acta Crystallo-
graphica, 16:1-8.
Taylor, L. A., G. Kullerud, and W. B. Bryan
1971. Opaque Mineralogy and Textural Features of
Apollo 12 Samples and Comparison with Apollo 11
Rocks. Geochimica et Cosmochimica Acta, Supple-
ment, 2:855-871.
Turekian, K. K., and S. P. Clark
1969. Inhomogeneous Accumulation of the Earth from
the Primitive Solar Nebula. Earth and Planetary
Science Letters, 6:346-348.
Van Schmus, W. R.
1967. Polymict Structure of the Mezo-Madaras Chondrite.
Geochimica et Cosmochimica Acta, 31:2027-2042.
1969. Mineralogy, Petrology, and Classification of Types
3 and 4 Carbonaceous Chondrites. Pages 480 to 491
in Millman, editor, Meteorite Research. Dordrecht,
Holland: D. Reidel Publishing Company.
Watts, H. L.
1958. Volumetric Determination of Aluminum in Pres-
ence of Iron, Titanium, Calcium, Silicon and Other
Impurities. Analytical Chemistry, 30:967.
Whipple, F. L.
1966. Chondrules: Suggestion Concerning the Origin.
Science, 153:54-56.
Wood, J. A.
1963. On the Origin of Chondrules and Chondrites.
Icarus, 2:152-180.
1967a. Olivine and Pyroxene Compositions in Type II
Carbonaceous Chondrites. Geochimica et Cosmo-
chimica Acta, 31:2095-2108.
1967b. Chondrites: Their Metallic Minerals, Thermal His-
tories, and Parent Planets. Icarus, 6:1-49.
U. S. GOVERNMENT PRINTING OFFICE: IB73? 516-22l/lM