ca.200
rma
ng IA
LAUR
us Box
29 E. U
ry, Sm
pted in
d wino
ion hi
te ass
, and
ies for
pyroxe
yroxen
r the fi
een the
and sa
parent
ough
ded at
y more
sis
magma. It is widely accepted that these trends formed during
crystallization of a metallic core after differentiation, with each
gro
19
IA
wi
con
(13
fol
wt
Ir)
20
che
19
and
tw
mo
pic
the
der
me
oth
(C
tio
sou
geochemical trends in the metal. Takeda et al. (2000) studied
gabbroic-textured regions in the Caddo County IAB iron and
* A
(gb
? P
Mu
Geochimica et Cosmochimica Acta, Vol. 69, No. 21, pp. 5123?5131, 2005
Copyright ? 2005 Elsevier Ltd
Printed in the USA. All rights reservedup forming in a distinct asteroid (Wasson, 1972; Buchwald,
75). There are, however, several iron meteorite groups (e.g.,
B, IIE, IIICD) whose geochemical trends are not consistent
th this scenario and many members of each of these groups
tain silicate inclusions. IAB is the largest of these groups
1 members) and second largest of all iron meteorite groups,
lowing the well-studied IIIAB group.
IAB iron metal is characterized by wide ranges in Ni (5.5
% to 60.8 wt%) and siderophile trace element (e.g., Ga, Ge,
concentrations (Choi et al., 1995; Wasson and Kallemeyn,
02). The common silicate inclusions are similar in bulk
mical composition to chondritic material (Bunch et al.,
70; Wasson, 1972; Benedix et al., 2000). However, olivine
orthopyroxene are intermediate in FeO concentration be-
een H and E chondrites. The stony winonaites share a com-
n silicate mineralogy, mineral chemistry and oxygen isoto-
composition (Benedix et al., 1998, 2000 and references
rein) with silicate material in IAB irons and are thought to
ive from the same parent asteroid (Bild, 1977).
The features of IAB irons indicate formation by one or more
chanisms that may not have operated on the parent bodies of
suggested formation by segregation of silicate and metal partial
melts at depth on a parent body early in the history of the solar
system. Benedix et al. (2000) advocated a hybrid model that
explains the apparently contradictory features by invoking in-
complete melting and separation of the Fe, Ni-FeS cotectic and
basaltic partial melts from the chondritic to ultramafic residues,
followed by catastrophic impact mixing to produce the silicate-
metal mixtures.
While these models provide a broad framework for interpret-
ing the genesis of IAB irons and winonaites, they cannot
elucidate the details of formation, including such intrinsic geo-
chemical parameters as peak temperatures, oxygen fugacity and
pressure. Particularly enigmatic is the nature of the precursor
chondritic material. Do the mineral and bulk compositions of
the IAB silicate inclusions and winonaites reflect those of the
precursor chondrite or were they originally more oxidized and
became reduced as a result of interaction with the carbon-rich
metal in which they are now embedded? In this study, we
attempt to determine these parameters in the IAB silicate in-
clusions and Winona, to better understand the thermal history
and starting composition of the parent body.
2. SAMPLES AND TECHNIQUES
2.1. Analytical Techniquesuthor to whom correspondence should be addressedenedix@levee.wustl.edu).doi:10.1016/j.g
Thermodynamic constraints on the fo
silicate-beari
GRETCHEN K. BENEDIX,1,*,? DANTE S.
1Dept. of Earth and Planetary Sciences, Washington University, Camp
2Lunar and Planetary Laboratory, Univ. of Arizona, 16
3Dept. of Mineral Sciences, National Museum of Natural Histo
(Received November 14, 2003; acce
Abstract?Silicate inclusions in IAB irons and relate
mineral chemistries that indicate a complex format
metamorphism. Using olivine-orthopyroxene-chromi
(Caddo County, Campo del Cielo, Copiapo, Lueders
calculated closure temperatures and oxygen fugacit
Fe-Mg exchange temperatures are compared to two-
peratures range from
590?C to
700?C, while two-p
Oxygen fugacities of these meteorites, determined fo
units below the Fe-FeO buffer and define a line betw
temperatures were experienced by these rocks on the h
consistent with the idea that the winonaite-IAB iron
reassembly after peak temperatures were reached. Alth
the oxygen fugacities and mineral compositions recor
chondritic precursor for this parent body was initiall
? 2005 Elsevier Ltd
1. INTRODUCTION
Most iron meteorite groups exhibit geochemical trends con-
tent with fractional crystallization of a common metallicW
Lu
resent address: Department of Mineralogy, The Natural History
seum, Cromwell Road, London, SW7 5BD, UK.
51235.03.048
tion conditions of winonaites and
B irons
ETTA,2 and TIMOTHY J. MCCOY3
1169, One Brookings Drive, St. Louis, MO 63130-4899, USA
niversity Blvd., Tucson, AZ 85721-0092, USA
ithsonian Institution, Washington DC 20560-0119, USA
revised form March 21, 2005)
naite meteorites have textures, mineralogies and
story of heating, followed by brecciation and
emblages in five IAB iron silicate inclusions
Udei Station) and one winonaite (Winona), we
these meteorites. Calculated olivine-chromite
ne temperatures. Olivine-chromite closure tem-
e temperatures range from
900?C to
1200?C.
rst time in this study, range from 2.3 to 3.2 log
Fe-FeO and Cr-Cr2O3 buffers. Highly variable
mple, and sometimes even the thin section, scale
body experienced collisional fragmentation and
modest reduction likely occurred during cooling,
peak metamorphic temperatures suggest that the
reduced than ordinary chondrites. Copyright
er iron meteorites (Wasson, 1972). Wasson and co-workers
hoi et al., 1995; Wasson and Kallemeyn, 2002) favor forma-
n of IAB irons by selective impact melting of a chondritic
rce, suggesting rapid heating and cooling to produce the
0016-7037/05 $30.00  .00e studied the IAB irons Caddo County, Campo del Cielo, Copiapo,
eders, and Udei Station, as well as the type winonaite, Winona
(Ta
fou
phi
me
res
the
P
refl
sio
JEO
cur
siti
(Ca
ton
con
20
the
2.2
T
env
sili
inte
C
dif
mit
Gh
chr
min
Th
bar
up
for
atu
Kre
hig
O
the
par
oliv
and
Eq
can
Lo
and
Lo
In
fay
aSiO
cal
sili
spe
Th
mo
Re
com
com
sol
3.1
3.1
hib
all
Be
thr
sili
clu
Co
bas
ma
Ud
(Fi
tog
som
par
cat
par
oft
sili
me
gra
con
mi
ext
cal
Sta
ite
occ
me
coe
mo
3.1
eac
ch
co
Ca
Ca
Co
Lu
Ud
Wi
5124 G. K. Benedix, D. S. Lauretta, and T. J. McCoyble 1). These meteorites span the complete range of textural types
nd in winonaites and IAB irons, including recrystallized metamor-
c unmelted rocks (Winona, Campo del Cielo, Copiapo), partial
lts (basaltic-gabbroic silicates in Caddo County), and partial melt
idues (Udei Station) (Benedix et al., 1998, 2000 and references
rein).
olished thin sections were studied optically in both transmitted and
ected light. Elemental X-ray maps of entire thin sections or inclu-
ns were acquired with a Cameca SX-50 electron microprobe and a
L JSM-840 scanning electron microscope at 15kV and 10nA beam
rent. Olivine, orthopyroxene, clinopyroxene and chromite compo-
ons were obtained on the electron microprobes at Virginia Tech
meca SX-50), the Smithsonian (JEOL JXA-8900R), and Washing-
University (JEOL Superprobe 733). All mineral analyses were
ducted with a fully focused beam at 15 kV accelerating voltage and
?30 nA beam current. Well-known mineral standards were used and
data were corrected with manufacturer-supplied ZAF routines.
. Thermodynamic Modeling
he primary goal of this study is to provide constraints on the
ironmental conditions experienced by the IAB iron meteorites with
cate inclusions and related winonaites. Such constraints allow for
rpretations of the origin and evolution of these materials.
losure temperatures are determined using thermometers for two
ferent mineral systems. First, we used co-existing olivine and chro-
e to determine temperature using the MELTS calculator (Sack and
iorso, 1991a, 1991b Ghiorso, pers. comm. 2004). We input our
omite and olivine compositions (Table 2a) and the calculator deter-
es the activities of the endmembers and a closure temperature.
ese calculations are also dependent on pressure, which we set to 1
, which is a reasonable assumption for the interior of a small body
to 100km in radius. Variation in temperature results are negligible
pressures up to 10000 bars. We also determined a closure temper-
re(s) using the two-pyroxene mineral system and the technique of
tz (1982), which is based on the transfer of Ca between the low- and
h-Ca pyroxenes.
nce closure temperatures have been determined, we can calculate
oxygen fugacity recorded by the major silicate minerals. The key
ameter for calculating the oxygen fugacity is the FeO content of the
ines and pyroxenes. The following reactions are relevant:
2Fe
 SiO2
O2(g) Fe2SiO4 (1)
Fe
 SiO2
 0.5 ? O2(g) FeSiO3 (2)
uilibrium constants as a function of temperature for these reactions
be expressed as
g K1 Log


aFe
aFe
2
? aSiO2 ? fO2

29558
T
 7.48
(737 T 1073) (3)
Table 1. Samples used in this study.
Meteorite Section Loaning Institution
ddo County USNM 6936-1 Smithsonian Institution
mpo del Cielo USNM 5615-4 Smithsonian Institution
USNM 5615-6 Smithsonian Institution
piapo USNM 3204-1 Smithsonian Institution
USNM 3204-2 Smithsonian Institution
eders UH 245 University of Hawaii
UH 255 University of Hawaii
ei Station USNM 2577 Smithsonian Institution
nona USNM 854 Smithsonian Institution
UH 133 University of Hawaii
UH 195 University of Hawaiig K2 Log


aFe
aFe ? aSiO2 ? 0.5fO2

29482
T
 7.43
(737 T 1073) (4)
these equations, T is the temperature in Kelvin, aFa is the activity of
alite, aFs is the activity of ferrosilite, aFe is the activity of Fe metal,
2 is the activity of silica, and ?O2 is the oxygen fugacity. Thus, to
culate an oxygen fugacity based on the FeO contents in these
cates, the activities of fayalite, ferrosilite, iron and silica must be
cified. The values for these parameters will be discussed below.
ermodynamic calculations were performed using the equilibrium
dule in the HSC Chemistry package, produced by Outokumpu
search Oy. This module calculates multi-component equilibrium
positions in heterogeneous systems using a database of over 17000
pounds and the GIBBS Gibbs energy minimization equilibrium
vers (White et al., 1958; Eriksson and Hack, 1990).
3. RESULTS
. Petrographic Features
.1. Textures
Silicate inclusions in IAB irons and winonaites generally ex-
it equigranular textures with abundant 120? triple junctions and
of the meteorites studied here fit this description (Fig. 1).
nedix et al. (1998, 2000) suggested that these textures formed
ough metamorphic cooling from near or slightly above the
cate peritectic. In addition to metamorphic textures, some in-
sions exhibit evidence of partial melting. For example, Caddo
unty contains gabbroic inclusions formed by crystallization of a
altic partial melt as well as inclusions of unmelted chondritic
terial (Takeda et al., 2000), while a single silicate inclusion in
ei Station is depleted in plagioclase and enriched in olivine
g. 2) and is thought to be a residue (Benedix et al., 2000). Taken
ether, the variation in textures of these meteorites indicate that
e silicate partial melting occurred on the winonaite-IAB iron
ent body.
The textures of the opaque minerals in winonaites and sili-
e inclusions in IAB irons are also indicative of heating to
tial melting. In winonaites, both Fe,Ni metal and troilite
en occur as veins cutting through silicate materials while
cate inclusions in IAB irons, which are generally depleted in
tal, contain abundant troilite found in veins or as individual
ins. Chromite is of particular importance in our attempt to
strain the conditions of formation. Coarse, euhedral chro-
te is often found in contact with troilite and, to a lesser
ent, metal, while chromite in contact with silicates is typi-
ly subhedral. A common mineral assemblage from Udei
tion is illustrated in Figure 3a, with troilite, chromite, graph-
, orthopyroxene and olivine. Anhedral chromite can also
ur, usually in contact with or in the presence of clearly
lted troilite. This is the case in Winona (Fig. 3b). The
xistence of these minerals allows us to apply the geother-
meters and thermodynamic equations outlined above.
.2. Mineral compositions
Mineral compositions are listed in Tables 2a and 2b. For
h meteorite, we measured the compositions of olivine,
romite, orthopyroxene, and clinopyroxene to produce a
nsistent data set. The only exception was chromite in
Table 2a. Olivine and chr
theses.
Caddo County
(Takeda et al., 2000)
SiO2 41.25
FeO 5.40
MnO 0.37
MgO 52.95
Total 99.97
XFaa 0.054
N
Caddo County
(Takeda et al., 2000)
TiO2 0.38
Al2O3 1.72
Cr2O3 68.75
V2O3 0.31
FeO 12.96
MnO 5.46
MgO 8.14
ZnO 1.10
Total 98.80
XChrb 0.393
XPChrc 0.440
N
N (grains)
a XFa  Fe/(FeMg) in mole frac
b XChr  Fe/(FeMgMn) in m
c XPChr  Mg/(FeMgMn) inomite compositions for the meteorites examined in this study. Also shown are data from previous studies (in bold) for comparison. 1-sigma errors are italicized in paren-
Olivine
Campo del Cielo
USNM 5615-4
Campo del Cielo
USNM 5615-6
Campo del Cielo
(Bunch et al., 1970) Copiapo
Copiapo
(Bunch et al., 1970) Lueders Udei Station
Udei Station
(Bunch et al., 1970)
Winona
USNM 854
Winona
UH 133
Winona
UH 195
41.59 (0.15) 41.45 (0.61) 42.10 41.73 (0.11) 42.40 42.26 (0.13) 41.50 (0.22) 41.30 42.02 (0.29) 42.04 (0.29) 41.94 (0.11)
3.92 (0.25) 4.03 (1.03) 4.10 4.50 (0.25) 5.10 4.45 (0.11) 5.61 (0.34) 7.50 4.88 (0.52) 5.19 (0.72) 4.63 (0.61)
0.43 (0.03) 0.42 (0.03) 0.40 0.42 (0.03) 0.37 0.41 (0.02) 0.42 (0.04) 0.41 0.38 (0.03) 0.37 (0.02) 0.38 (0.02)
52.77 (0.58) 51.81 (1.37) 54.10 52.76 (0.18) 52.40 53.43 (0.13) 52.50 (0.42) 51.00 52.45 (0.59) 52.64 (0.66) 52.57 (0.41)
98.71 97.70 100.70 99.40 100.27 100.55 100.03 100.21 99.98 100.50 99.74
0.040 (0.003) 0.042 (0.011) 0.044 0.046 (0.002) 0.052 0.045 (0.001) 0.057 (0.003) 0.076 0.0495 (0.005) 0.052 (0.008) 0.047 (0.006)
18 17 8 14 6 5 57 5 21 11 6
Chromite
Campo del Cielo
USNM 5615-4
Campo del Cielo
USNM 5614-6
Campo del Cielo
(Bunch et al., 1970) Copiapo
Copiapo
(Bunch et al., 1970) Leuders Udei Station
Udei Station
(Bunch et al., 1970)
Winona
USNM 854
Winona
UH 133
Winona
UH 195
0.23 (0.04) 0.63 (0.1) 0.42 0.52 (0.28) 0.48 0.64 (0.24) 0.54 (0.17) 0.70 0.25 (0.04) 0.22 (0.03) 0.28 (0.02)
0.79 (0.39) 2.88 (0.65) 0.79 0.45 (0.43) 0.42 0.32 (0.26) 0.13 (0.16) 1.73 0.51 (0.36) 0.16 (0.05) 0.19 (0.12)
71.15 (0.86) 69.08 (2.15) 74.00 70.17 (0.8) 71.70 68.56 (1.13) 70.26 (1.33) 68.20 71.53 (0.63) 71.04 (0.80) 70.89 (1.13)
0.67 (0.10) 0.38 (0.05) 0.24 0.52 (0.04) 0.57 0.70 (0.04) 0.79 (0.12) 0.68 0.23 (0.04) 0.25 (0.03) 0.25 (0.05)
11.90 (0.73) 13.31 (1.75) 12.30 14.43 (0.51) 15.10 16.20 (1.64) 16.83 (0.75) 16.40 12.94 (0.44) 14.20 (0.23) 13.44 (0.95)
3.59 (0.41) 1.83 (0.12) 2.32 2.91 (0.86) 3.40 3.1 (0.49) 2.90 (0.35) 2.66 2.21 (0.41) 2.93 (0.56) 2.79 (0.40)
11.00 (0.26) 11.14 (0.45) 9.20 8.51 (0.19) 7.10 8.00 (0.31) 7.48 (0.41) 7.50 10.32 (0.42) 9.25 (0.22) 9.78 (0.24)
1.58 (0.26) 2.20 (0.17) 1.44 1.99 (0.17) 1.70 2.34 (0.31) 2.18 (0.27) 1.88 2.26 (0.24) 2.32 (0.13) 2.12 (0.20)
100.99 101.53 100.71 99.67 100.47 99.89 101.17 99.75 100.3 100.4 99.8
0.338 (0.015) 0.379 (0.036) 0.396 0.443 (0.019) 0.484 0.481 (0.032) 0.509 (0.017) 0.505 0.386 (0.016) 0.422 (0.004) 0.422 (0.004)
0.558 (0.017) 0.568 (0.035) 0.528 0.466 (0.011) 0.406 0.425 (0.027) 0.403 (0.019) 0.412 0.549 (0.018) 0.490 (0.016) 0.490 (0.015)
41 10 31 9 84 53 11 45
8 2 5 5 5 9 2 5
tion.
ole fraction.
mole fraction.
5125
Form
ation
co
nditions
of
w
inonaites
and
IA
B
irons
Ca
ine
al.
al.
me
ne
the
oli
Ca
of
val
Wi
con
one
wh
val
thi
ges
rat
rim
Ca
gra
exh
obs
zon
mi
tho
ibr
acr
tha
oli
cat
F
IAB
gra
gra
Wi
F
min
Th
sho
vin
pho
in t
loc
figu
5126 G. K. Benedix, D. S. Lauretta, and T. J. McCoyddo County, which was absent in the section we exam-
d. In addition to our data, we report analyses of Bunch et
(1970) and Takeda et al. (2000). The data from Takeda et
(2000) are taken from a chondritic lithology found near a
lted lithology. In most cases, the agreement between our
w compositions and those of previous authors was within
range of analytical uncertainty, although our analysis of
vine in Udei Station and chromite and low-Ca pyroxene in
mpo del Cielo and Copiapo differ significantly from that
Bunch et al. (1970).
Olivine compositions have a narrow span (Table 2a) of Fa
ig. 1. Crossed-polarized photomicrographs of silicate inclusions in
irons and Winona illustrating metamorphic textures and range in
in sizes. Black areas are either opaque minerals (metal, sulfide, or
phite) or extinct minerals. Field of view for each  2.5 mm. (a)
nona, (b) Campo del Cielo, (c) Udei Station.ues ranging from 4.0 to 5.7 mol% across all meteorites studied.
thin a single thin section, the olivine composition is tightly
strained (1 for Fa is less than 0.6 mol%). An exception is in
thin section of Campo del Cielo (USNM 5615-6, Table 2a)
ich has a standard deviation of the mean of 1.1 mol% on Fa
ue. Wlotzka and Jarosewich (1977) reported zoned olivine in
s meteorite, prompting Kallemeyn and Wasson (1985) to sug-
t that silicate compositions in IAB irons recorded reduction,
her than metamorphic equilibration. We analyzed cores and
s of 20 randomly selected olivine grains in two thin sections of
mpo del Cielo (both from the El Taco mass). We found three
ins that exhibit significant reverse zoning, while the remainder
ibited no significant zoning. In addition, in Winona, we have
erved olivine grains with normal and reverse as well as no
ing co-existing. The extent of the zoning, when present, is
nimal (less than 1 mol%, typically less than 0.5 mol%). Or-
pyroxene and clinopyroxene compositions are also well equil-
ated amongst themselves and exhibit small ranges in Fs values
oss entire thin sections (Table 2b). For both minerals, 1 is less
n 0.7 mol%. It is interesting to point out that Fa values of
vine are lower than Fs values of low-Ca pyroxene. This indi-
es they are not in equilibrium with each other, as has been
ig. 2. Combined X-ray element map showing the distribution of all
erals in the silicate inclusion in Udei Station (PTS USNM 2577).
is image combines X-ray maps of Mg, Na, Ca, S, P, and Cr to
wcase the distribution of the major minerals (orthopyroxene, oli-
e, plagioclase, troilite, clinopyroxene, chromite, phosphate, and
sphide). The metallic matrix which encloses the inclusion is black
his image. The white box in the upper right of the image outlines the
ation of Figure 3a. Scale is depicted in the upper right portion of the
re.
dis
Ho
and
wit
tha
the
gen
alth
Fe
Mg
and
gra
ant
to
(Ch
Al
lar
Cr/
me
2 w
3.2
(19
chr
ties
tem
clo
the
are
chr
are
les
ma
clo
of
to
hav
tio
and
19
cal
Fe
tha
the
and
fay
is
Th
act
low
sys
(19
oli
ox
iro
13
rim
tha
sys
det
the
fro
fro
To
bo
bu
cho
and
the
Cr
Th
and
up
cur
pac
F
mit
bla
inc
ite
are
che
for
5127Formation conditions of winonaites and IAB ironscussed previously (Benedix et al., 1998 and references therein).
wever, this result is somewhat puzzling since both the olivine
pyroxene grains have relatively homogenous compositions
hin individual inclusions without significant zoning, suggesting
t they have equilibrated among themselves. We return to this in
discussion.
Individual IAB irons and winonaites exhibit relatively homo-
eous chromite compositions (1 for XChr of 2?4; Table 2a),
ough they exhibit significant variation between meteorites. The
endmember (XChr) values range from 33.8 to 50.9 mol%. The
endmember (XPchr) values range from 40.3 to 56.8 mol%. Fe
Mg exhibit minor to no variation across individual chromite
ins. MnO and ZnO abundances are variable and exhibit weak,
icorrelated variations in some grains. Chromites can contain up
5 wt.% MnO in the chondritic lithology of Caddo County
ikami et al., 1999; Takeda et al., 2000). The trivalent cations
and Cr, which have the slowest diffusion rates, exhibit the
gest variations (i.e., biggest standard deviations), although the
CrAl ratio stays relatively homogeneous at 
0.98 across all
teorites. Absolute abundances of Al2O3 in chromite can reach
t.% but are generally 1 wt%.
ig. 3. Back-scattered electron (BSE) images of olivine (Ol)-chro-
e (Chr)-troilite (Tr)-orthopyroxene (OPX)-graphite (Gr) assem-
ges. (a) Close-up image of assemblage in Udei Station silicate
lusion (USNM 2577). (b) Image of silicate-chromite-troilite-graph-
assemblages in Winona (USNM 854). Olivine and orthopyroxene
indistinguishable in this BSE. The labeled olivine point was
cked with an X-ray energy dispersive spectrometer (EDS). Scales
100 m are on each image.. Thermodynamic Modeling
We use the olivine-chromite thermometer of Sack and Ghiorso
91a, 1991b) to determine a closure temperature. The olivine-
omite closure temperatures and corresponding oxygen fugaci-
are shown in Table 3 and illustrated in Figure 4. These
peratures range from 
590?C to 
700?C with the highest
sure temperature recorded in Winona (PTS USNM 854) and
lowest in Lueders. Uncertainties of
50?on the temperatures
determined from the standard deviations of the olivine and
omite compositions (Table 2a). Olivine-chromite temperatures
highly sensitive to small changes in olivine composition and
s dependent on changes in chromite composition of comparable
gnitude.
Also shown in Table 3 are the calculated two-pyroxene
sure temperatures which were determined using the method
Kretz (1982). These range from 
900?C (Campo del Cielo)

1100?C (Winona). Uncertainties on these temperatures
e been estimated to be around  100?, based on composi-
nal variability within a given sample, as well as precision
accuracy errors in the calculation of the temperature (Kretz,
82).
Plugging the closure temperatures into reactions 3 and 4, we
culate oxygen fugacities at these temperatures based on the
O content of the olivines and low-Ca pyroxenes. We assume
t the activity of Fe (aFe) is 0.91 due to the presence of Ni in
metal of the IABs. We used the MELTS calculator of Sack
Ghiorso (1991a, 1991b) to determine the activities of
alite (aFa) and ferrosilite (aFs). Thus, the only parameter that
unconstrained is the activity of silica in this system (aSiO2).
ere have never been quantitative constraints placed on the
ivity of silica in systems containing Fe metal, olivine, and
-Ca pyroxene. To determine the activity of silica in our
tem we examined the experimental results of Larimer
68). In that study, the distribution of Fe between metal,
vine, and orthopyroxene was determined as a function of
ygen fugacity (ranging from 1 to 3 log units below the
n-w?stite buffer) and temperature (ranging from 1100 to
00?C). Every experimental condition investigated by La-
er (1968) yielded a silica activity value of 0.9. We assume
t this silica activity is characteristic of the three component
tem metal-olivine-low-Ca pyroxene and use this value to
ermine the oxygen fugacities recorded by the FeO content of
olivine. Our calculated olivine oxygen fugacities range
m 
2.9 to 
3.2 log units and orthopyroxene ?O2 range
m 2.3 to 2.7 log units below the iron-w?ustite (IW) buffer.
gether these two datasets define a linear array intermediate in
th absolute ?O2 and slope between the IW and Cr-Cr2O2
ffers. These results are not unreasonable given that ordinary
ndrites have fugacities that fall 
1 log unit below IW (Brett
Sato, 1983).
In addition to temperature and ?O2 for these IAB irons and
winonaite, Figure 4 includes buffer curves for Fe-FeO (IW),
-Cr2O3, and for CO-C (at pressures of 1, 10 and 100 bars).
e IW buffer is the upper limit for ordinary chondrites (Brett
Sato, 1983), while Cr-Cr2O3 is a reasonable proxy for the
per limit of oxygen fugacity of enstatite meteorites. Buffer
ves were calculated using data from the HSC Chemistry
kage.
Table 2b. Orthopyrox
in parentheses.
Caddo County
(Takeda et al., 2000
SiO2 57.65
TiO2 0.25
Al2O3 0.27
Cr2O3 0.38
FeO 4.25
MnO 0.46
MgO 35.58
CaO 1.02
Na2O 0.04
Total 99.89
XFsa 0.062
XWo 0.019
N
Caddo County
(Takeda et al., 2000
SiO2 54.10
TiO2 0.68
Al2O3 0.81
Cr2O3 0.90
FeO 1.82
MnO 0.27
MgO 18.44
CaO 22.00
Na2O 0.64
Total 99.66
XFs 0.029
XWo 0.448
N
a XFs  Fe/(FeMgCa) i
b No high-Ca pyroxene was5128ene and clinopyroxene compositions for the meteorites examined in this study. Also shown are data from previous studies (in bold) for comparison. 1-sigma errors are italicized
Orthopyroxene
)
Campo del Cielo
USNM 5615-4
Campo del Cielo
USNM 5614-6
Campo del Cielo
(Bunch et al., 1970) Copiapo
Copiapo
(Bunch et al., 1970) Lueders Udei Station
Udei Station
(Bunch et al., 1970)
Winona
USNM 854
Winona
UH 133
Winona
UH 195
57.62 (0.24) 57.26 (0.33) 58.40 58.00 (0.29) 57.00 58.89 (0.24) 57.37 (0.36) 56.50 58.65 (0.32) 58.80 (0.23) 58.39 (0.27)
0.23 (0.02) 0.22 (0.02) 0.18 0.25 (0.03) 0.18 0.25 (0.06) 0.26 (0.02) 0.22 0.25 (0.03) 0.26 (0.02) 0.26 (0.05)
0.31 (0.03) 0.43 (0.31) 0.22 (0.04) 0.55 0.24 (0.08) 0.27 (0.03) 0.85 0.29 (0.08) 0.32 (0.03) 0.28 (0.07)
0.37 (0.03) 0.35 (0.04) 0.34 0.22 (0.04) 0.29 0.29 (0.10) 0.35 (0.04) 0.37 0.26 (0.04) 0.28 (0.06) 0.24 (0.04)
4.36 (0.42) 4.06 (0.49) 4.20 4.11 (0.30) 4.90 4.30 (0.40) 5.34 (0.32) 5.60 4.46 (0.18) 4.53 (0.16) 4.48 (0.15)
0.52 (0.02) 0.51 (0.02) 0.49 0.47 (0.03) 0.42 0.48 (0.03) 0.48 (0.03) 0.45 0.41 (0.02) 0.40 (0.02) 0.39 (0.03)
35.11 (0.46) 34.88 (0.44) 36.30 35.68 (0.41) 35.30 35.82 (0.6) 34.75 (0.36) 34.70 35.42 (0.49) 35.52 (0.28) 35.31 (0.44)
0.80 (0.11) 0.74 (0.13) 0.70 0.78 (0.19) 0.58 0.79 (0.15) 0.87 (0.11) 0.92 0.92 (0.19) 0.96 (0.08) 0.93 (0.15)
bdl bdl 0.20 bdl 0.22 bdl bdl 0.11 bdl bdl bdl
99.35 98.51 100.81 99.67 99.44 101.10 99.74 99.72 100.78 101.18 100.43
0.064 (0.006) 0.060 (0.007) 0.060 0.060 (0.004) 0.071 0.062 (0.006) 0.078 (0.005) 0.082 0.065 (0.003) 0.066 (0.002) 0.065 (0.002)
0.015 (0.002) 0.014 (0.002) 0.013 0.015 (0.004) 0.011 0.015 (0.003) 0.016 (0.002) 0.017 0.017 (0.004) 0.018 (0.002) 0.017 (0.003)
27 30 18 29 8 16 46 8 11 13 9
Clinopyroxene
)
Campo del Cielo
USNM 5615-4
Campo del Cielo
USNM 5614-6
Campo del Cielo
(Bunch et al., 1970) Copiapo
Copiapo
(Bunch et al., 1970) Lueders Udei Station
Udei Station
(Bunch et al., 1970)
Winona
USNM 854
Winona
UH 133
Winonab
UH 195
54.05 (0.19) 54.05 (0.14) 55.3 54.54 (0.17) 54.3 55.31 (0.09) 54.39 (0.19) 54.3 55.13 (0.16) 55.63
0.71 (0.03) 0.66 (0.06) 0.58 0.76 (0.08) 0.53 0.67 (0.03) 0.68 (0.03) 0.57 0.87 (0.06) 0.93
0.84 (0.02) 0.86 (0.07) 0.17 0.70 (0.03) 1.17 0.96 (0.04) 0.83 (0.03) 1.72 0.85 (0.03) 0.86
1.40 (0.04) 1.44 (0.03) 1.27 0.74 (0.13) 1.15 1.39 (0.09) 1.28 (0.04) 1.16 0.90 (0.08) 0.77
1.67 (0.27) 1.69 (0.26) 1.52 1.63 (0.19) 1.98 1.86 (0.18) 2.05 (0.06) 1.84 1.63 (0.03) 1.49
0.30 (0.04) 0.28 (0.03) 0.27 0.27 (0.03) 0.29 0.30 (0.02) 0.30 (0.03) 0.28 0.24 (0.02) 0.24
17.67 (0.26) 17.45 (0.07) 18.3 17.91 (0.09) 18.1 18.04 (0.14) 17.97 (0.31) 18.6 18.32 (0.10) 18.82
21.49 (0.42) 21.68 (0.56) 21.3 22.04 (0.11) 21.1 21.09 (0.20) 21.30 (0.39) 20.2 22.14 (0.12) 21.69
0.83 (0.04) 0.83 (0.04) 1.4 0.64 (0.003) 1.01 0.82 (0.03) 0.80 (0.03) 1.06 0.63 (0.03) 0.66
98.96 98.94 100.11 99.22 99.63 100.44 99.69 99.73 100.8 101.09
0.028 (0.004) 0.028 (0.004) 0.025 0.026 (0.003) 0.032 0.030 (0.003) 0.033 (0.001) 0.030 0.026 (0.001) 0.024
0.454 (0.010) 0.459 (0.009) 0.444 0.457 (0.004) 0.441 0.443 (0.005) 0.445 (0.008) 0.425 0.453 (0.002) 0.442
6 5 17 2 8 7 10 9 5 1
n mole fraction; XWo  Ca/(FeMgCa) in mole fraction.
found in this PTS.
G
.K
.B
enedix,D
.S.Lauretta,
and
T.J.M
cCoy
est
tur
com
et
abo
mo
tem
atu
the
19
to
tw
for this
IWe
Ca 3.0
Ca 3.1
C 3.1
C 3.1
Co 3.1
Lu 3.2
Ud 2.9
Wi 2.9
W 2.9
W 2.9
W 3.0
a
b action (
c
d eaction
e .
f temper
g
h
i
F
T(K
Ox
sho
Ind
nam
pia
ent
sec
Cr-
ref
IW
5129Formation conditions of winonaites and IAB irons4. DISCUSSION
Previous studies of silicates in IAB irons and winonaites
imated peak temperatures and calculated closure tempera-
es from a combination of mineral abundances, textures and
positions (Bunch et al., 1970; Takeda et al. 2000; Benedix
al., 2001, 2002). These temperatures range from slightly
ve 900?C to over 1450?C. Abundant textural evidence for
bility of Fe,Ni-FeS and basaltic partial melts suggests peak
peratures in excess of both the Fe,Ni-FeS eutectic temper-
re of 
950?C (Kullerud, 1963; Kubaschewski, 1982) and
basaltic partial melting temperature of 
1050?C (Morse,
80).
Closure temperatures, such as those calculated here, allow us
interpret the cooling history of the parent body. In this study,
o-pyroxene temperatures based on Ca-transfer span 
900?

1100?C and olivine-chromite temperature ranges from
90?C to 
700?C. It is interesting to compare these systems.
-transfer temperatures range less than 200?C (Table 3) and
not correlate with the olivine-chromite temperatures. How-
r, the Ca-transfer temperatures are all substantially (
280?

480?) higher than the olivine-chromite temperatures. Un-
cooling conditions, the two-pyroxene Ca-transfer system
uld close before the Fe-Mg exchange of the olivine-chro-
te.
It is odd that there is no correlation between the 2-pyroxene
thermometers and the olivine-chromite geothermometer
oss this suite of rocks. One explanation may be that heating
/or cooling was localized and any correlations would only
found within different inclusions or sections of a given
teorite. This appears to be the case for Winona, where one
n section (USNM 854) has a smaller difference (T 
80?; Table 3) between the two-pyroxene and the olivine-
omite temperature than another section from a different part
the rock (UH 133; T 
420?; Table 3). These differences
ld be caused by variations in cooling rates, where a smaller
results from a faster cooling rate. The other meteorite for
study using data from Tables 2a and 2b.
Two-pyroxenec
Temperature (?C)
QMFsd
log fO2 IWe Tf
1016 17.2 2.5 426
939 18.6 2.6
965 18.1 2.6 277
913 19.2 2.7 279
933 18.8 2.7 322
1067 16.3 2.5 481
1043 16.6 2.3 401
994 17.0 2.5
976 17.9 2.5 276
1084 16.0 2.4 416
n/ai n/ai n/ai
eqn. 2) using the olivine-chromite T.
(eqn. 1) using the two-pyroxene T.
atures in degrees.to

5
Ca
do
eve
to
der
sho
mi
geo
acr
and
be
me
thi

2
chr
of
cou
T
ig. 4. Plot of oxygen fugacity (log ?O2) vs. temperature (10,000/
)) for the five IAB irons and one winonaite analyzed in this study.
ygen fugacity determined from orthopyroxene (Eqn. 2) in text) is
wn in blue, and that determined from olivine (Eqn. 1) in text) in red.
ividual points are labeled with abbreviations for the meteorite
es: Cad?Caddo County; CdC?Campo del Cielo; Cop?Co-
po; Lue?Lueders; Ude?Udei Station; Win?Winona. The differ-
points labeled ?Win? and ?CdC? are data from the different thin
tions examined. Also shown in this plot are the Fe-FeO (IW),
Cr2O3, and three CO-C (1, 10 and 100 bars) fugacity buffers for
erence. The data lie on a line (R2  0.9997) that nearly parallels the
buffer.Table 3. Thermodynamic properties determined
Meteorite
Olivine-Chromitea
Temperature (?C)
QMFab
log fO2
ddo County 590 28.1
mpo del Cielog 661 25.8
dC (USNM 5615-4) 688 24.9
dC (USNM 5615-6) 634 26.7
piapo 611 27.4
eders 586 28.4
ei Station 642 26.2
nonah 674 24.9
inona (USNM 854) 700 24.4
inona (UH133) 668 25.3
inona (UH195) 653 25.9
Calculated using thermometer of Sack and Ghiorso (1991a,b).
QMFa  Oxygen fugacity derived from the Quartz-Iron- Fayalite re
Temperature based on Ca-exchange thermometer of Kretz (1982).
QMFs  Oxygen fugacity derived from the Quartz-Iron-Ferrosilite r
IW is the deviation of fO2 from the iron-wustite buffer in log units
T is the difference between the two-pyroxene and olivine-chromite
Average of the 2 thin sections studied.
Average of the 3 thin sections studied.
This thin section did not contain any high-Ca pyroxene.
wh
Ca
mo
Th
ma
ho
be
inc
tur
his
sam
ab
me
sis
wi
be
hig
atu
?C
to
atu
fea
co
po
tha
trix
co
tro
all
ch
loc
loc
(19
an
an
an
en
an
tw
3).
co
tai
an
he
sio
ex
an
gra
ex
pre
the
Lu
the
on
era
gra
3)
co
of
res
Th
alo
tha
int
plo
eit
me
be
a c
sm
ox
an
ev
up
ou
in
tro
ox
sil
an
of
for
na
be
Kr
IA
ord
ph
IA
pe
bu
sam
tem
rea
Gi
an
sur
co
res
on
me
IA
Th
rec
py
ph
rev
tha
wi
of
ev
na
thi
of
sug
ch
5130 G. K. Benedix, D. S. Lauretta, and T. J. McCoyich we have multiple sections from different inclusions is
mpo del Cielo. In this case the T between the two ther-
meters is remarkably similar for each section (
280?).
ese sections come from inclusions separated by the metallic
trix. The metallic matrix separating inclusions may create
mogenous thermal systems in the IAB silicates. This could
tested with other IAB irons which contain many silicate
lusions across a large area. The variations in the tempera-
es of each of these meteorites indicate that the thermal
tory of the parent body is localized, down to even hand
ple or thin section scale.
In general, the closure temperatures range from slightly
ove that of basaltic partial melting to well below the Fe,Ni
tal-FeS eutectic. Both temperature sets are internally con-
tent with textural and mineralogical features observed
thin these rocks, but record variations in cooling rates
tween the samples. Silicates in Winona (Fig. 1a) exhibit a
hly recrystallized, metamorphic texture, but the temper-
re recorded by olivine-chromite range from 650 to 700
, while the two-pyroxene temperatures range from 
980?

1080?C. One explanation for this difference in temper-
re appears to be recorded in the mineralogy and textural
tures. There are regions within PTS USNM 854 that are
arse-grained and dominated by olivine which were hy-
thesized (Benedix et al., 1998) to be partial melt residues
t were mixed with the metamorphosed, fine-grained ma-
material. In all PTS of Winona, chromite occurs in
ntact with troilite as well as olivine. In these assemblages,
ilite always exhibits melt textures and chromite occasion-
y does as well (Fig. 3b). Troilite not in contact with
romite does not have melt textures. These features imply
alized heating episodes in Winona. While some may link
alized heating to the impact hypothesized by Choi et al.
95), breakup and reassembly after peak heating provides
equally probable scenario to produce localized heating
d cooling histories, as hotter and cooler clasts are mixed
d very localized thermal regions equilibrate. At the other
d of the temperature spectrum, the IAB iron, Lueders has
olivine-chromite closure temperature of 
590?C and a
o-pyoxene temperature of 
1070?C (T  
480?; Table
Lueders exhibits a recrystallized texture that is more
arse-grained than Winona. Mineralogically Lueders con-
ns chondritic abundances of high-Ca pyroxene (
6 vol%)
d plagioclase (
12 vol%), however these minerals are
terogeneously distributed throughout this particular inclu-
n. This sample does not appear to be a residue, but does
hibit evidence of mineral mobilization in veins of troilite
d plagioclase crosscutting the thin section. In addition,
phite is distributed unevenly throughout the entire PTS
amined, indicating that temperatures were low and/or
ssures were high enough to retain it. We hypothesize that
lower temperature and the large T is consistent with
eders cooling slowly, possibly at some depth, thus leaving
mineral systems open longer. Udei Station (Fig. 2) is the
ly sample from this study that exhibits textural and min-
logic evidence suggesting that it is a residue. Its larger
in size, presence of graphite, and large T (
400?; Table
suggests that it, like Lueders, likely experienced slow
oling.
While the closure temperatures expand our understandingthe thermal history of this parent body, a more important
ult of this study is the determination of oxygen fugacity.
e meteorites fall on a line (R2  0.9997), but do not fall
ng any known buffer. One possible explanation might be
t graphite is buffering the system. Chromite often occurs
ergrown with troilite and graphite. On a log ?O2 vs. T
t, graphite buffers fall on a distinctly different slope than
her Fe-FeO or Cr-Cr2O3. At temperatures of interest to the
lting of chondrites, the CO-C buffer occupies the gap
tween these other buffers. In meteoritic terms, melting of
hondrite in the presence of graphite, with subsequent
elting of FeO to form reduced Fe and CO, produces an
ygen fugacity intermediate between ordinary chondrites
d enstatite meteorites (Walker and Grove, 1993). How-
er, if this were the case all graphite should have been used
in the reactions. Graphite is clearly distributed through-
t the silicate inclusions and is also found (although rarely)
Winona. Thus, it seems likely oxygen fugacity is con-
lled by a complex, multi-buffer system.
This is the first study to provide direct constraints on the
ygen fugacities of these enigmatic meteorites. The mafic
icate compositions (e.g., Fa values lower than Fs values)
d sulfide mineralogy (e.g., daubre?lite, but no oldhamite)
IAB irons and winonaites has long been taken to indicate
mation at oxygen fugacities intermediate between ordi-
ry chondrites and enstatite chondrites. It has, however,
en unclear how this reduced nature was established.
acher (1985) postulated that winonaites and silicates in
B irons were reduced from a starting composition of
inary chondrite (
Fa20) composition during metamor-
ism at a pressure of 10 bars. Our work suggests that the
Bs and winonaites cooled to two-pyroxene closure tem-
ratures between 2.5 and 3 log units below the iron-w?stite
ffer. Although, the presence of graphite in many of the
ples indicates that the pressure was at least 10 bars, the
perature was not high enough for the graphite reduction
ction, as hypothesized by Kracher (1985), to take place.
ven that peak temperatures (based on mineral abundances
d textures) were not much higher than two-pyroxene clo-
e temperatures, we can calculate the average Fa value
rresponding to the two-pyroxene closure T and ?O2. The
ulting Fa value (5.2 mol%; Sack and Ghiorso, 1989) is
ly slightly above the average measured Fa value for these
teorites (4.7 mol%). Thus, winonaites and silicates in
B irons experienced only slight reduction during cooling.
is is further supported by the fairly small differences
orded in ?O2 between the olivine-chromite and two-
roxene systems. While modest reduction during metamor-
ism may be necessary to explain some features (e.g.,
erse zoning in some olivine grains), it now appears clear
t the oxygen fugacity and degree of reduction recorded by
nonaites and IAB iron silicates was an intrinsic property
the precursor chondritic material and that they did not
olve through partial melting and reduction from an ordi-
ry chondrite-like (
Fa20) protolith. The reduced nature of
s precursor chondrite and the oxygen isotopic signatures
IAB irons and winonaites are unlike known chondrites,
gesting that we have either yet to sample it as a distinct
ondrite or that all of the parent bodies of this type of
chondrite experienced at least partial melting and differen-
tiation.
5. CONCLUSIONS
wi
iro
Fe
atu
sca
at
be
py
sca
bre
be
reh
mi
mo
wi
inh
wa
pre
ap
Wh
trin
wi
ab
To
bo
ind
tem
Ack
Un
the
Gh
ano
por
(DS
A
Be
Be
Be
Be
Bild R. W. (1977) Silicate inclusions in group IAB irons and a relation
to the anomalous stones Winona and Mt. Morris (Wis). Geochim.
Cosmochim. Acta 41, 1439?1456.
1983) Intrinsic oxygen fugacity measurements of 7 chondrites and a
pallasite and redox state of meteorite parent bodies. Proc. Lunar
Planet. Sci. XIV, 69?70.
Bu
Bu
Ch
Ch
Eri
Ka
Kra
Kre
Ku
Ku
Lar
Mo
Sac
Sac
Sac
Tak
Wa
Wa
Wa
Wh
Wl
5131Formation conditions of winonaites and IAB ironsThe closure temperatures calculated in this study, along
th textural features, indicate that silicate inclusions in IAB
ns and winonaites likely experienced temperatures above
S-FeNi and, possibly, the silicate partial melting temper-
res. The degree of heating varied at the local (m to dm)
le and, in some cases, heterogeneous heating is preserved
the cm scale, as evidenced by the lack of correlation
tween the olivine-chromite geothermometer and the two-
roxene geothermometers. Heterogeneous heating at these
les is consistent with reassembly after a catastrophic
akup, where the pieces were reassembled at temperatures
low the peak temperatures and experienced a range of
eating and cooling histories. Oxygen fugacities deter-
ned for the first time in this work suggest only very
dest reduction during cooling. The reduced nature of
nonaites and IAB silicate inclusions appears to be an
erent property of their precursor chondritic material and
s not established during reduction of a more oxidized
cursor during metamorphism. The precursor chondrite
pears unsampled among our collection of chondrites.
ile oxygen fugacity and temperature were important in-
sic parameters in determining the evolution of the IAB-
nonaite parent body, our knowledge of the timing (both
solute and relative) of these events remains incomplete.
completely understand the history of this complex parent
dy, a coordinated study that incorporates chronology of
ividual silicate inclusions with information about their
perature and oxygen fugacity history is required.
nowledgments?We are grateful to the Smithsonian Institution and
iversity of Hawaii for allocation of samples. The Excel version of
MELTS Olivine-Chromite calculator was kindly provided by Mark
iorso. Helpful reviews by A. Kracher, D. Mittlefehldt, and an
nymous reviewer improved the manuscript. This work was sup-
ted by NASA grants NAG5?13464 (TJM) and NNG04GF65G
L).
ssociate editor: D. Mittlefehldt
REFERENCES
nedix G. K., McCoy T. J., Keil K., Bogard D. D., and Garrison D. H.
(1998) A petrologic and isotopic study of winonaites: Evidence for
early partial melting, brecciation and metamorphism. Geochim.
Cosmochim. Acta 62, 2535?2553.
nedix G. K., McCoy T. J., Keil K., and Love, S. G. (2000) A
petrologic study of the IAB iron meteorites: Constraints on the
formation of the IAB-Winonaite parent body. Meteor. Planet. Sci.
35, 1127?1141.
nedix G. K., Lauretta D. L. and McCoy T. J. (2001) Constraining the
formation conditions of silicate-bearing iron meteorites. 11th. Annu.
Goldschmidt, Abstr. # 3843 (CDROM)
nedix G. K., Lauretta D. L. and McCoy T. J. (2002) Thermodynamic
constraints on formation conditions of silicate-bearing IAB Iron
meteorites. Lunar Planet. Sci. XXXIII Abstr.#1317 (CDROM).chwald V. F. (1975) Handbook of Iron Meteorites. Univ. of Calif.
Press, Berkeley, Ca. USA 1418 pp.
nch T. E., Keil K., and Olsen E. (1970) Mineralogy and petrology of
silicate inclusions in iron meteorites. Contrib. Mineral. Petrol. 25,
297?340.
ikami J., Miyamoto M., and Takeda H. (1999) The variation of Zn
content in spinel group minerals and daubreelites of primitive
achondrites. Antarct. Meteorite Res. 12, 139?150.
oi B.-G., Ouyang X., and Wasson J. T. (1995) Classification and
origin of the IAB and IIICD iron meteorites. Geochim. Cosmochim.
Acta 59, 593?612.
ksson G., and Hack K. (1990) ChemSage?A computer program for
the calculation of complex chemical equilibria. Metall. Trans. 21B,
1013?1023.
llemeyn G. W., and Wasson J. T. (1985) The compositional classi-
fication of chondrites: IV. Ungrouped chondritic meteorites and
clasts Geochim. Cosmochim. Acta 49, 261?270.
cher A. (1985) The evolution of the partially differentiated plan-
etesimals:Evidence from the iron meteorite groups IAB and IIICD.
J. Geophys. Res. 90 (Suppl.), C689?C698.
tz R. (1982) Transfer and exchange equilibria in a portion of the
pyroxene quadrilateral as deduced from natural and experimental
data. Geochim. Cosmochim. Acta 46, 411?421.
baschewski O. (1982) Iron-Binary Phase Diagrams. Springer-Ver-
lag, New York, New York, USA. 185 pp.
llerud G. (1963) The Fe-Ni-S system. Ann. Rep. Geophys. Lab. 67,
4055?4061.
imer J. W. (1968) Experimental studies on system Fe-MgO-
SiO2-O2 and their bearing on petrology of chondritic meteorites.
Geochim. Cosmochim Acta 32, 1187?1207.
rse S. A. (1980) Basalts and Phase Diagrams. Springer-Verlag. 493
pp.
k R. O., and Ghiorso M. S. (1989) Importance of considerations of
mixing properties in establishing an internally consistent database:
Thermochemistry of minerals in the system Mg2SiO4-Fe2SiO4-
SiO2. Contrib. Mineral. Petrol. 102, 41?68.
k R. O., and Ghiorso M. S. (1991a) Chromite as a Petrogenetic
Indicator. In Oxide Minerals: Petrologic and Magnetic Significance
(ed. D. H. Lindsley), Rev. Mineral. 25, 323?353. Mineralogical
Society of America
k R. O., and Ghiorso M. S. (1991b) Chromian spinels as petroge-
netic indicators: thermodynamics and petrological applications.
Am. Mineral. 76, 827?847.
eda H., Bogard D. D., Mittlefehldt D. W., and Garrison D. H.
(2000) Mineralogy, petrology, chemistry and 39Ar-40Ar and expo-
sure ages of the Caddo County IAB iron: Evidence for early partial
melt segregation of a gabbro area rich in plagioclase-diopside.
Geochim. Cosmochim. Acta 64, 1311?1327.
lker D., and Grove T. (1993) Ureilite smelting. Meteoritics 28,
629?636.
sson J. T. (1972) Parent body models for the formation of iron
meteorites. Proc. Intl. Geol. Cong. 24, 161?168.
sson J. T., and Kallemeyn G. W. (2002) The IAB iron meteorite
complex: A?group, five subgroups, numerous grouplets, closely
related, mainly formed by crystal segregation in rapidly cooling
melts. Geochim. Cosmochim. Acta 66, 2445?2473.
ite W. B., Johnson S. M., and Dantzig G. B. (1958) Chem. equi-
librium in complex mixtures. J. Chem. Phys. 28, 751?755.
otzka F., and Jarosewich E. (1977) Mineralogical and chemical
compositions of silicate inclusions in the El Taco, Campo del Cielo,
iron meteorite. Smith. Contrib. Earth Sci. 19, 104?125.