SMITHSONIAN CONTRIBUTIONS TO THE
EARTH SCIENCES
NUMBER 9
wuiiam Q. Mehon Mineral Sciences
Investigations,
1969-1971
EDITOR
I S S11 E D
AUG .161972
SMITHSONIAN INSTITUTION PRESS
CITY OF WASHINGTON
1972
ABSTRACT
Melson, William G., editor. Mineral Sciences Investigations, 1969-1971. Smith-
sonian Contributions to the Earth Sciences, number 9, 94 pages, 34 figures, 1972.?
Seventeen short contributions from the Smithsonian's Department of Mineral Sci-
ences from 1969 to 1971 are gathered together in this volume. The scientific and
technological subjects treated in these contributions include studies of lunar samples
from Apollo 12, meteorites, petrology and volcanology, and descriptive mineralogy,
as well as of the history and description of one of the Smithsonian Institution's most
important recent acquisitions?the Carl Bosch Collection of Minerals and Meteorites.
Official publication date is handstamped in a limited number of initial copies and is recorded
in the Institution's annual report, Smithsonian Year.
Library of Congress Cataloging in Publication Data
Melson, William G., 1938- comp.
Mineral sciences investigations, 1969-1971.
(Smithsonian contributions to the earth sciences, no. 9)
"Seventeen short contributions from the Smithsonian's Department of Mineral Sciences."
1. Mineralogy?Addresses, essays, lectures. 2. Petrology?Addresses, essays, lectures. I. Na-
tional Museum of Natural History. Dept. of Mineral Sciences. II. Title. III. Series:
Smithsonian Institution. Smithsonian contributions to the earth sciences, no. 9.
QE1.S227 no. 9 [QE369] 55O'.8s [549'. 1] 72-355
For sale by the Superintendent of Documents, U.S. Government Printing OfficeWashington, D.C. 20402 - Price $1.25 (paper cover)
Contents
Page
LUNAR STUDIES
Mineralogy and Petrography of Some Apollo 12 Samples, by Brian Mason,
William G. Melson, E. P. Henderson, Eugene Jarosewich, and Joseph
Nelen 1
Microhardness of a Lunar Iron Particle and High-Purity Iron Samples, by
E. P. Henderson, R. D. Buchheit, and J. L. McCall 5
PETROLOGY AND VOLGANOLOGY
Basaltic Nuees Ardentes of the 1970 Eruption of Ulawun Volcano, New
Britain, by William G. Melson, Eugene Jarosewich, George Switzerland
Geoffrey Thompson 15
Experimental Data Bearing on the Paragenesis of Two Hawaiian Basalts
from Kilauea Volcano, by R. F. Fudali 33
Iron-rich Saponite: Additional Data on Samples Dredged from the Mid-
Atlantic Ridge, 22? N Latitude, by Harold H. Banks, Jr 39
Olivine Crystals from the Floor of the Mid-Atlantic Ridge near 22f'N Lati-
tude, by George Switzer, William G. Melson, and Geoffrey Thompson .... 43
Origin and Composition of Rock Fulgurite Glass, by George Switzer and
William G. Melson 47
METEORITES
The Estherville Meteorite, by Joseph Nelen and Brian Mason 55
Forest Vale Meteorite, by A. F. Noonan, Kurt A. Fredriksson, and Joseph
Nelen 57
Iron Meteorite Compositions, by Roy S. Clarke, Jr., and Eugene Jarosewich 65
The Nejo, Ethiopia, Meteorite by Roy S. Clarke, Jr., Eugene Jarosewich, and
Joseph Nelen 67
The Nakhon Pathom Meteorite, by Joseph A. Nelen and Kurt Fredriksson . . 69
MINERALOGY
Gorceixite in the Buffels River, Namaqualand, South Africa, by J. J. Gurney . 77
Tocornalite, by Brian Mason . - 79
STANDARDS FOR CHEMICAL ANALYSIS
Chemical Analysis of Five Minerals for Microprobe Standards, by Eugene
Jarosewich 83
THE COLLECTIONS
Acquisition of the Carl Bosch Collection of Minerals and Meteorites, by Paul
E. Desautels 87
THIN-SECTION PREPARATION
Preparation of Doubly Polished Thin Sections, by Grover Moreland, Frank
Ingram, and Harold H. Banks, Jr. 93
LUNAR STUDIES
Brian Mason, William G. Melson,
E. P. Henderson, Eugene Jarosewich,
and Joseph Nelen
Mineralogy and
Petrography of Some
Apollo 12 Samples
ABSTRACT
Samples of the Apollo 12 fines contain fragments of
crystalline rocks and breccias, mineral grains, and
glassy particles. The mineral grains are mainly pyrox-
ene (pigeonite and augite) and olivine, with minor
amounts of plagioclase and ilmenite, and rare metal
grains. The compositions of the metal grains indicate
that some are of meteoritic origin and some are de-
rived from lunar rocks. A single pyroxene crystal has
a pigeonite core, an augite mantle, and a rim of
ferrohedenbergite, with a narrow zone of pyroxfer-
roite along a central fissure. One rock fragment from
a sample of coarse fines (12032) is a silica-rich hed-
enbergite granophyre, consisting of sanidine (50%),
quartz (30%), hedenbergite (10%), plagioclase
(7%), fayalite (3%), and accessory apatite, whit-
lockite, and zircon. A complete chemical analysis of
12001 fines has been made, and also a determination
of the water-soluble cations, a noteworthy feature
being the presence of 103 ppm water-soluble calcium.
The Smithsonian received the following samples from
the Apollo 12 collections:
Fines (-1 mm): 12001 (3.0 g); 12003 (0.5 g); 12033
(0.5 g); 12042 (0.5 g); 12057 (2.0 g);
12070 (2.0 g)
Coarse fines: (1 mm-1 cm): 12032 (0.5. g)
Rock sections: 12009, 12022, 12045, 12051
Each sample of the fines was sieved through 60-
mesh (0.250-mm opening) and 200-mesh (0.074-mm
opening) sieves. The grain-size distribution was sim-
ilar in all, the weight proportions being: +60 mesh,
Brian Mason, William G. Melson, E. P. Henderson, Eugene
Jarosewich, and Joseph Nelen, Department of Mineral Sci-
ences, Smithsonian Institution, Washington, D.C. 20560.
10-20%; 60-200 mesh, 20-30%; ?200 mesh, 50-
60%. The fines contained fragments of crystalline
rocks and breccias, mineral grains, and glassy par-
ticles. The coarser fractions were density-separated,
using methylene iodide (D = 3.32). The fraction with
D>3.32 comprised about 30%, and consisted almost
entirely of grains of pyroxene (major), olivine (mi-
nor), and opaque minerals (small amount). The
fraction with D<3.32 (about 70%) consisted of rock
and breccia fragments, glass spherules and fragments,
and feldspar grains.
The ?200-mesh fraction of 12,001 was further
separated with a 325-mesh sieve (0.044-mm open-
ing), and a portion of the ?325-mesh fraction ana-
lysed. The results are given in Table 1, along with
two analyses of Apollo 11 fines. The compositions of
the Apollo 12 and the Apollo 11 fines are rather
similar, except that the Apollo 12 fines contain about
half as much titanium as the Apollo 11 material;
this has already been noted in earlier reports (LSPET
1970). Our analyses also show distinctly higher po-
tassium and chromium in Apollo 12 fines. Since
microscopic examination showed minute particles of
metallic iron in the glassy fraction of the fines, we
attempted to determine this in the Apollo 12 fines
(by dissolution in KCUCI3 solution). The figure
obtained, 0.33% Fe, is certainly a minimum value,
since the procedure would only dissolve those metal
particles exposed on the surface of the glass frag-
ments, and not those totally enclosed. The high sum-
mation of the analysis of the Apollo 12 fines prob-
ably results from the reporting of some metallic iron
as FeO, and possibly some trivalent titanium as TiO2.
The calculated mineral compositions (norms) of
these analyses show about 40% feldspar of anorthitic
composition and about 45% pyroxene. The composi-
1
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 1.?Analyses and norms of Apollo 11 and Apollo 12 fines
(Eugene Jarosewich, analyst)
Constituent
SiOa
A12O3
TiO2
FeO
MgO
CaO
Na2O
K2O
P2O5
H2O-
MnO
Cr2O3
Ni
S
Femet
Sum
S = O
Sum
Analyses
12001
?325 mesh
45.83
13.45
2.99
15.78
10.28
10.63
0.47
0.25
0.22
0.00
0.21
0.45
0.03
0.17
0.33
101.09
0.09
101.00
(weight percent;
10084 10084
?525 mesh glass concentrate
42.02
14.61
6.38
15.59
7.92
12.39
0.45
0.15
0.27
0.00
0.15
0.17
0.03
100.13
43.28
14.83
5.43
14.91
7.63
12.56
0.48
0.19
0.17
0.00
0.18
0.10
0.03
0.13
99.92
0.07
99.85
Constituent
An
Ab
Or .
Di
Hy
Ol
11
Cr
Ap
Metal
FeS
Norms
12001
?325 mesh
33.9
4.0
1.5
14.4
33.0
7.0
5.7
0.7
0.5
0.3
0.5
(weight percent)
10084 10084
?325 mesh glass concentrate
37.4
3.8
0.9
18.4
20.4
6.2
12.1
0.3
0.6
37.8
4.1
1.1
19.4
22.5
4.1
10.3
0.2
0.4
0.4
tions are slightly undersaturated in SiOo, giving small
amounts of olivine. The Apollo 12 fines contain the
equivalent of about 5% ilmenite, compared to about
10% in the Apollo 11 material. After subtracting the
ilmenite, the FeO/FeO + MgO molecular percentage
is remarkably similar in all these analyses (42 for
the Apollo 12; 41 and 43 for the Apollo 11 analyses).
We also determined the water-extractable cations
in both Apollo 11 and Apollo 12 fines. A control ex-
periment without sample yielded less than 1 ppm of
the cations determined. A sample of the Allende
meteorite was extracted for comparison. The results
are as follows (in ppm) :
Na K Mg Ca Fe Ni
12001 14 5 31 103 5 5
10084 21 5 16 325 5 5
Allende 238 5 25 20 5 5
The notable feature is the comparatively high amount
of extractable calcium, considerably less, however, in
12001 than in 10084. The extractable cations are not
present in the fines as halides, since Reed and Jo-
vanovic (1970) report less than 2 ppm water-soluble
halogens in 10084. The high surface area and reactiv-
ity of the fresh lunar glass may be responsible. The
water-extractable calcium perhaps explains the re-
ported stimulating effect of lunar dust on plant
growth.
Thirty grains of olivine in the Apollo 12 fines were
analysed with the microprobe. The range in composi-
tion, in mole percent Fe2SiO4 (Fa), is Faae-eoj with
no marked compositional preference; the mean is
Fa4i. Minor elements detected with the microprobe
were Ca (average 0.28%), Mn (0.24%), Cr (0.14%)
and Ti (0.08%). The contents of calcium are notably
high for olivine; in meteoritic olivines which we have
measured Ca is usually 0.01-0.02%, and Cr 0.02-
0.04%. This probably reflects a high temperature of
crystallization for the lunar olivine, and rapid chilling
preventing readjustment at lower temperatures.
Some sixty pyroxene grains in the Apollo 12 fines
were also analysed. They showed two compositional
clusters, one for pigeonite and the other for augite-
subcalcic augite. The mean composition of the pi-
geonite cluster is \V011En55Fs33, of the augite-sub-
calcic augite cluster \V033En40Fs27. The principal mi-
nor elements are aluminum and titanium. The pi-
geonite grains average 1.4% A12O3 and 0.9% TiO2,
the augite and subcalcic augite 2.3% A12O3 and
1.3% TiOo.
The range of compositions shown by the isolated
NUMBER 9
pyroxene grains are essentially duplicated in a single
pyroxene crystal fragment picked from 12057. This
crystal fragment, 1.1 mm by 0.7 mm, has a central
core of pigeonite of rather uniform composition
(\V010En57Fs33) with a mantle of subcalcic augite
(\V037En37Fs26), and on one edge a narrow rim of
ferrohedenbergite (Wo3oEn7Fse3). A narrow rift in
the central core is lined with a thin zone of pyrox-
ferroite. The individual minerals are clearly distin-
guished in crossed polars by the increase in birefrig-
ence in passing from pyroxferroite to pigeonite to
augite to ferrohedenbergite. The compositional vari-
ation across this grain is given in Figure 1 in terms
of the pyroxene end members. A similar pyroxene
crystal from Apollo 12 material has been described
by Bence et al. (1970).
Plagioclase grains from the Apollo 12 fines show
a range of composition from Ans2 to Ange, with a
mean of An90. The plagioclase contains detectable
iron (average 0.4% FeO), potassium (average 0.1%
K2O), and titanium (average 0.08% TiO2).
Several metal-rich particles were extracted from
the Apollo 12 fines by means of a hand magnet. The
total amount in any sample is small, making up much
less than 1%. Two types of metal were distinguished;
one type is finely granular and shows deformation
structures, the other is coarse grained and unde-
formed. Microprobe analyses of the deformed grains
give figures for nickel and cobalt comparable with
those for meteoritic nickel-iron, suggesting that these
grains have been derived from meteorite impacts on
the lunar surface. The coarse-grained undeformed
metal usually has lower nickel concentrations than
meteoritic nickel-iron, and sometimes more cobalt
than nickel;' these fragments are presumably indige-
nous to the Moon, probably formed by localized
reduction of basaltic material. One such fragment is
essentially pure iron, with no detectable amounts of
other elements; it is extremely soft, comparable with
the purest strain-free iron we have been able to ob-
tain. This metal fragment contains inclusions of oli-
vine and ilmenite similar to these minerals in the
lunar basalts, from which it was presumably derived.
Our 0.5-g sample of coarse fines (12032, from the
rim of Bench Grater) was washed with acetone to
remove dust, and then density-fractionated in methy-
lene iodide-acetone mixtures. The density fractions
were sorted by hand under a binocular microscope
into breccia fragments and crystalline rock fragments.
The results are set out in Table 2. The average
weight of individual fragments is 3.5 mg, and the
percentages of breccia and crystalline rocks are 82
and 18 respectively.
Several crystalline rock fragments were selected
for detailed examination. The most remarkable one
was a lone white fragment in the D<2.70 fraction,
which otherwise consisted of highly vesiculated black
breccia fragments. This crystalline fragment, 3-mm
long, proved to consist largely of sanidine, with minor
amounts of quartz, plagioclase, fayalite, and heden-
EN
Dl
Ferrohedenbergite
FS
FIGURE 1.?Compositional zoning across a single pyroxene crystal from sample 12057, projected
in the system enstatite (En), ferrosilite (Fs), wollastonite (Wo); arrows indicate crystallization
sequence from core (pigeonite) through mantle (augite) to rim (ferrohedenbergite); a thin
zone of pyroxferroite occurs along the central fissure.
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
bergite, and accessory zircon, apatite, and whitlockite.
The composition of the individual minerals, deter-
mined by the microprobe, is given in Table 3, along
with the calculated bulk composition.
The hedenbergite and fayalite are present as phen-
ocrysts, up to 1-mm long, in a fine-grained ground-
mass of quartz and feldspar, much of it in micro-
TABLE 2.?Fractionation of 0.5-g coarse fines from
the rim of Bench Crater (12032, 46)
Density
2.70
2.70-2.90
2.90-3.10
3.10-3.32
3.32
Totals
Weight percent
8
33
44
12
3
100
Number
Breccia
11
43
58
5
?
117
of fragments
Crystalline rocks
1
3
6
12
4
26
TABLE 3.?Analyses (microprobe) of minerals in hed-
enbergite granophyre (12032, 46?2)
Constituent
SiO2
TiO2
AI2O3 .
Gr2Os
FeO
MnO
MgO
CaO
Na2O
K2O
SrO
BaO
Microprobe Analyses
1*
65.4
18.6
0.82
n.d.
n.d.
0.31
2.96
11.6
n.d.
1.21
100.9
2*
57.0
26.8
1.10
n.d.
n.d.
8.74
6.85
0.15
n.d.
n.d.
100.6
3*
99.72$
0.28
n.d.
n.d.
n.d.
n.d.
n.d.
(100.0)
(weight percent)
4*
48.8
0.75
0.45
0.06
28.0
0.39
1.55
19.3
0.10
99.4
5*
29.8
0.21
0.04
0.13
67.8
0.95
1.22
0.67
n.d.
100.8
6*
72.0
01
11.1
5.3
0 1
0.2
?7
2.1
5.8
06
100.0
n.d. = not detected (<0.05%).
* Column 1: Sanidine (OrooAbasCn^ni); 50%.
Column 2: Plagioclase feldspar (AbssAnwOri); 7%.
Column 3: Quartz; 30%.
Column 4: Hedenbergite (Wo47En5Fsso); 10%.
Column 5: Fayalite (FaOT) ; 3%.
Column 6: Bulk composition (calculated).
J Calculated by difference.
graphic intergrowth. The percentages of sanidine and
quartz in this intergrowth are approximately 65:35,
a composition close to the temperature minimum on
the liquidus in the system KAlSiaOs?NaAlSiaOg?
S1O2. The rock is comparable to some terrestrial
granophyres, specifically the hedenbergite grano-
phyres of the Skaergaard intrusion. One of these
(EG 3021) has essentially identical mineralogy, con-
sisting of hedenbergite (Wo44En6Fs5o), fayalite
(Fa95), plagioclase (An40), potash feldspar, and
quartz and tridymite. Lindsley et al. (1969) estimate
temperatures of 950??980? C and an oxygen fugacity
ranging from 10~13? to 10~13-6 atm for the crystalli-
zation of the Skaergaard rock, and these figures
should be equally applicable to the lunar granophyre.
The chemical, mineralogical, and textural features
of this rock strongly suggest that it represents the last
liquid differentiate from the fractional crystallization
of a large body of lunar basaltic magma (or possibly
an initial product of partial melting). Its composition
is similar to that of the high-silica glassy mesostasis of
many of the Apollo 11 igneous rocks, described by
Roedder and Weiblen (1970). It is also comparable
with the light-colored, silica-rich parts of the unique
rock 12013. Whether it was formed at or near the
locale of the Apollo 12 landing, or whether it is an
exotic fragment, possibly from a Copernicus ray, re-
mains an open question.
This work has been supported by NASA Contract
9-8016.
Literature Cited
Bence, A. E., J. J. Papike, and C. T. Prewitt
1970. Apollo 12 Clinopyroxenes: Chemical Trends.
Earth and Planetary Science Letters, 8:393.
Lindsley, D. H., G. M. Brown, and I. D. Muir
1969. Conditions of the Ferrowollastonite-Ferroheden-
bergite Inversion in the Skaergaard Intrusion, East
Greenland. Mineralogical Society of America
Special Paper 2:193.
LSPET (Lunar Sample Preliminary Examination Team)
1970. Preliminary Examination of Lunar Samples from
Apollo 12. Science, 1967:1325.
Reed, G. W., and S. Jovanovich
1970. Halogens, Mercury, Lithium and Osmium in
Apollo 11 Samples. Geochimica et Cosmochimica
Ada, Supplement 1:1487.
Roedder, E., and P. W. Weiblen
1970. Lunar Petrology of Silicate Melt Inclusions,
Apollo 11 Rocks. Geochimica et Cosmochimica
Acta, Supplement 1:801.
E. P. Henderson,
R. D. Buchheit,
and J. L. McCall
Microhardness of
a Lunar Iron
Particle and
High-Purity Iron
Samples
Introduction
Lunar soil recovered by the Apollo 11 and 12 mis-
sions is continually being subjected to critical analyses
and examinations by numerous scientists and engi-
neers. A sample of the lunar soil, Sample 12032, 46-1
from the Apollo 12 mission which had been sieved
to a particle size of about 3 mm, was received for
studies by the Smithsonian Institution from the
Manned Space Flight Center. During the studies,
one particle of the lunar soil sample attracted at-
tention because it was so strongly magnetic. Metal-
lographic examinations showed the magnetic particle
to be metallic iron possessing an unusually low micro-
hardness. A comparison of the microhardness of this
particle of metallic iron with the microhardness of
high-purity iron samples is reported herein.
Examinations of Lunar Iron Particle
The particle of lunar soil which exhibited a strong
magnetic attraction is shown in Figure 1 as a metal-
lographic specimen etched in 1% Nital for about 20
seconds. Polishing was done using mechanical meth-
ods. The maximum diameter of the particle was about
2.3 mm. The microstructure of the particle consisted
of a coarse-grained mass of iron (kamacite) with
numerous nonmetallic inclusions concentrated in a
E. P. Henderson, Department of Mineral Sciences, Smith-
sonian Institution, Washington, D.C. 20560. R. D. Buchheit
and J. L. McCall, Battelle Memorial Institute, Columbus
Laboratories, Columbus, Ohio 43201.
peripheral zone. The dark inclusions are essentially
olivine but some ilmenite is associated with the
olivine.
The dark inclusion, left of center, Figure 1, which
is surrounded with a swathing zone, is shown at
higher magnification in Figure 2. The electron micro-
probe study of the inclusion showed that calcium and
phosphorus are present in about the proportions in
which they exist in apatite; since the surface of the
black material was chipped, a satisfactory probe
study was impossible.
A small metallic inclusion with a slightly different
reflectivity than the metal was found along the inter-
face of the swathing metal around the apatite (?)
and the matrix. This grain was probed by Dr. K. F.
J. Heinrich and Mr. Charles Fiori of the National
Bureau of Standards who reported that although it
has essentially the composition of the surrounding
metal it probably contains a light element?carbon,
boron or nitrogen?in amounts too small to be mea-
sured. On this evidence, we suspect that this inclu-
sion is a carbide.
The average diameter of the metal grains in the
lunar particle, determined by a lineal intercept
method, was found to be about 0.25 mm. We ob-
served no work hardening or other physical damage
in the microstructure of this lunar iron.
MICROHARDNESS DETERMINATIONS
The microhardness values obtained on this Apollo 12
sample were lower than the values normally reported
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 1.?Microstructure of lunar iron particle. Maximum dimension is 2.3 mm.
for commercially pure iron and much below the mi-
crohardness of kamacite in iron meteorites.
Recently E. P. Henderson found what is assumed
to be kamacite grains in stony meteorites which are
unusually soft. Some of these grains, when measured
with the Vickers diamond indenter, using a 25-gram
loan, gave a Vickers Hardness Number (V.H.N.)
between 120 and 130. The study of these soft Ni-Fe
grains in stony meteorite has not been completed.
The Vickers 136? diamond pyramid was chosen
for all these microhardness measurements because it
required a smaller area for a given load than the
Knoop indenter and it generally provides more pre-
cise data than ball-type indenters (Lea 1936). The
average value of the microhardness found with the
weight loads given below are:
NUMBER 9
. ?A1ft9sf!
?./.mm
.-????-*>V4t??SS
FIGURE 2.?The dark rough area at the left is tentatively identified as apatite. Along the inter-
face of the apatite and metal there is some ilmenite. The metallic inclusion between the two
metallic areas may be a carbide. Width of field is about 0.1 mm.
Weight load
(grams)
25
50
100
V.H.N.
80.6
78.6
72.6
Indentations
3
8
8
Mr. T. B. Shives of the National Bureau of Stand-
ards measured the microhardness of the lunar particle
using a 100-gram load and reported an average value
of 69 V.H.N. Various loads were used for all these
and other microhardness measurements reported in
this paper because of the known variation in hardness
number when low-load microhardness testing is per-
formed (Small 1960; Lysaght and DeBellis 1969;
Edmond 1958).
A recent study involving laboratory annealing of
kamacite with 7% nickel for long periods of time, at
temperatures of 300-700 ?C, showed the lowest mi-
crohardness attained was 153 Knoop Hardness Num-
ber, which is equivalent to about 140 V.H.N. (Staub
and McCall, 1969).
It is widely accepted that the hardness of mete-
oritic iron (kamacite) is approximately proportional
to the nickel content suggesting that the low hardness
values obtained from the lunar iron particle could be
a result of high purity, essentially fully annealed, iron.
To check this out, electron microprobe analyses were
made on the lunar iron particle and the microhard-
ness of the lunar iron was compared with the micro-
hardness of three different high-purity iron samples.
ELECTRON MICROPROBE ANALYSES
Electron microprobe analyses of the particle of lunar
iron detected only iron and failed to detect the pres-
ence of Ni, Co, or P, elements generally found in
meteoritic iron. The detectability limits of the elec-
tron microprobe analyser used were 0.1% for Co and
0.05% for Ni. No other analytical techniques were
used on this particle because of its small size and the
desire to preserve as much of it as possible.
8 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
Impurity element
Carbon
Nitrogen, N2
Oxygen
Aluminum
Calcium
Chlorine
Cobalt
Chromium
Copper
Germanium
Potassium
Manganese
Sodium
Nitrogen, N
Phosphorus
Sulfur
Weight, ppm
lla
7?
12b
0.2c
2.1c
12.7?
15.7?
8.4c
1.7c
5.2c
7.0c
1.0c
20.6c
15.8c
4.4c
0.6c
* Combustion analysis.
b Vacuum fusion analysis.
c Mass spectrographic analysis.
Microhardness Determinations of High-Purity Iron TABLE 1.?Impurity content of the Bain high-purity
Samples iron sample
A literature search for microhardness data for high-
purity iron samples revealed a complete absence of
useful data. Therefore, an effort was made to procure
some samples of high-purity iron and to determine
their microhardness for comparison with the micro-
hardness of the lunar iron particle. Three samples
of high-purity iron were obtained, one from the
Edger C. Bain Laboratory for Fundamental Research
of the U. S. Steel Corporation,1 one from the Colum-
bus Laboratories of Battelle Memorial Institute,2 and
the third sample from Lehigh University.3
BAIN HIGH-PURITY IRON SAMPLE
Details concerning the preparation of the Bain sample
of high-purity iron were not available; however, a
chemical analysis was reported and the impurity
elements and their contents in parts per million
(ppm) by weight are given in Table 1. The analyses
for all of the impurity elements except carbon, nitro-
gen, and oxygen were determined by mass spectros-
copy. The analyses of carbon and nitrogen were made
by chemical determinations, and the analysis of oxy-
gen was made by vacuum fusion. These analyses in-
dicate the sample to be 99.985 + percent iron.
The Bain sample was mounted in Bakelite at a
pressure of 5000 psi and a temperature of about
130?C. The mounted sample was ground and pol-
ished metallographically and etched for 30 seconds in
1% Nital. Subsequently, microhardness tests were
made in several places on the polished and etched
surface.
All of the microhardness indentations were made
with a Vickers diamond microhardness tester at-
tached to a Vickers metallograph. The microhardness
determinations were performed by one operator over
an interval of several days. Each day the balance pan
on the microhardness tester was reset. In order to
evaluate the uniformity of the microhardness values
and their reproductibility in the Bain high-purity iron
sample, four series of microhardness determinations
were made. The first series consisted of four identa-
tions equally spaced along a straight line and made
Courtesy of Dr. John Bromback.
Courtesy of American Iron and Steel Institute.
Courtesy of Dr. Joseph Goldstein.
with a load of 50 grams. The second series was made
near the first series and consisted of four rows of
four indentations in each row. The indentations in
the first and fourth rows were made with a 50-gram
load, and the indentations in the second and third
rows were made with a 25-gram load. The third
series consisted of two rows of four indentations in
each row made with a 100-gram load. The fourth
and last series of microhardness determinations con-
sisted of three rows of five indentations in each row
made with a 25-gram load. The microhardness values
obtained from each of the four series of determina-
tions in the Bain high-purity iron sample are given
in Table 2, and the average values are plotted versus
load in Figure 3. Included in Figure 3 is a plot of the
average Vickers microhardness values of the lunar
iron.
BATTELLE HIGH-PURITY IRON SAMPLE
The Battelle high-purity iron sample was floating
zone purified (Rengstorff 1968). The process of puri-
fication begins with commercially available iron, usu-
ally high-purity electrolytic iron. The starting mate-
rial is electron-beam melted in a vacuum furnace and
NUMBER 9
TABLE 2.?Vickers Microhardness Numbers for three high-purity iron samples and the Apollo 12 lunar iron
particle and the total impurity content for the three high-purity iron samples
Total
Sample impurities,
ppm
Load grams
10 25 50 100 200 300 400 ^00
Battelle
Iron
Lunar Iron
Bain Iron ..
Lehigh
Iron
35-50 78.9 74.5 70.0 65.4 60.6 59.5 58.1 57.4
(70.5-83.8) (73.7-75.5) (69.0-71.2) (64.7-66.3) (58.5-62.4) (57.2-60.9) (56.3-61.3) (55.5-60.0)
unknown ? 80.6 78.3 72.6 _____
(78-85) (74.2-82.5) (70.7-74.5)
124.4 ? 81.9 78.3 77.9 ? ? ? ?
(77.2-85.7) (76.0-80.0) (75.7-79.8)
700.5 135.5 118.5 110.7 105.2 ? ? ? ?
(122-159) (109-131) (106-116) (100-112)
160r
140
120
A
W)
111
Z
?100
80
60
40
? LEHIGH HIGH-PURITY IRON
? BAIN
O APOLLO 12 '<
? BATTELLE "
25 50 100 200
LOAD IN GRAMS
300 400 500
FIGURE 3.?Vickers microhardness number versus load for three high-purity iron samples and
the Apollo 12 lunar iron particle.
cast into V/% -inch-diameter ingots. Four zone-melting
passes of the electron-beam-melted ingots complete
the purification process. No chemical analysis of the
elemental impurity content was obtained for the
particular sample made available, but the total im-
purity content was known to be 35?50 ppm by
weight. This means the sample was 99.9950+ per-
cent iron.
10 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
The Battelle sample was mounted in a room-tem-
perature setting epoxy resin, ground and polished
metallographically, and finally electropolished. Sub-
sequently, microhardness determinations were made
on the electropolished surface.
All of the microhardness determinations were made
with a Vickers diamond indenter or an MO Tukon
Microhardness Tester using loads of 10, 25, 50, 100,
200, 300, 400, and 500 grams. The various loads
were used because of the previously shown variations
of hardness reading with load. The microhardness in-
dentations were placed in 11 rows of 6 indentations in
each row. The spacing between rows and between
adjacent indentations was approximately 4 mils for
low load (50 grams and less) indentations and 8
mils for the heavier load indentations. Twelve micro-
hardness determinations were made for each of the
10-, 25-, and 50-gram loads; six determinations were
made for each of the 100-, 200-, 300-, 400-, and 500-
gram loads. The microhardness indentations were
measured with a filar micrometer eyepiece in con-
junction with a 50 X objective. The results of the
microhardness survey on the Battelle high-purity iron
sample are given in Table 2 and the average micro-
hardness values versus load are plotted in Figure 3.
LEHIGH SAMPLE
The Lehigh sample, although regarded as high-purity
iron, differed from the Bain and Battelle samples in
that it consisted of a fine-grained microstructure con-
taining many dark nonmetallic inclusions. The Le-
high iron was cast as a cylinder and then heat treated
at 1000?C for 12 hours. The reported analysis of the
Lehigh iron sample, shown in Table 3, indicates it
to be 99.9 + percent iron. The sample was mounted
in Bakelite and prepared metallographically for the
microhardness measurements.
The microhardness measurements were made using
loads of 10, 25, 50, and 100 grams by the same pro-
cedure described for the Battelle iron sample. The
microhardness values obtained are given in Table 2,
and the average values are plotted versus load in
Figure 3.
Discussion of Results
The comparison, shown in Figure 3, of the micro-
hardness of the lunar iron particle with the micro-
TABLE 3.?Impurity content of the Lehigh iron sample
Impurity element
Carbon
Oxygen
Nitrogen
Hydrogen
Tungsten
Nickel
Al, Ba, B, Ca, Zr
Co, Cu, Ge, Mn, Mo, V
Gr, Pb, Mg, Si, Sn, Ti, Zn
ppm
70?
200b
10b
<0.5b
<300c
<40c
<50c
<20c
<10c
a Combustion analysis.
b V&cuum fusion analysis.
c Emission spectrographic analysis.
hardness of three different high-purity iron samples
indicates that the lunar particle also is high-purity
iron. This indication is supported further by the
electron microprobe analyses made on the lunar par-
ticle which revealed no detectable elements other
than iron.
The hardness of a material usually is affected by
only three factors: the structure and grain orienta-
tion, the amount of strain (cold work), and the
composition. The structures of the three high-purity
iron samples, as well as the structure of the lunar
iron particle, consisted of single-phase alpha iron
(kamacite). The only difference among the structures
which might affect hardness (O'Neill 1967) was the
difference of grain sizes, but, since the grain sizes of
all the samples were large enough to entirely contain
a microhardness impression at all the loads used, this
effect is not believed to be significant. Another factor
known to affect hardness is the crystallographic ori-
entation of the grains with respect to the plane on
which the microhardness measurements are made.
This effect is particularly significant for anisotropic
materials, but has been shown to be relatively minor
for isotropic materials such as alpha iron. Further-
more, the hardness measurements made on the sam-
ples in this study were found to be rather consistent
from reading to reading (grain to grain) within each
sample, indicating little or no orientation effect.
The amount of strain in a material very signifi-
cantly affects hardness with the lowest hardness al-
ways being obtained for the condition of 0 percent
strain, i.e., fully annealed. The three high-purity iron
NUMBER 9 11
samples were intentionally prepared in this condition
and, in all cases, care was taken to metallographically
prepare their surfaces such that no strain was intro-
duced (Vitovec and Binder 1955). The only indica-
tion that the lunar iron particle was strain free, other
than its low hardness, is its microstructure which re-
vealed equiaxed grains with no "strain lines." Further
confirmation of this probably would require a destruc-
tive technique of examination, such as transmission
electron microscopy which was not permitted in this
study.
Finally, the hardness of materials are affected by
their compositions. This is well known for alloys and
probably is the most important reason for alloying
(strength is directly related to hardness (Lenhart
1955)). Less well kno\vn, however, is the effect of
small amounts of impurity elements on the hardness
of otherwise pure elements. Since the other contribut-
ing factors to hardness described above are believed
to provide only minor effects on the materials in
this study, it is believed that the present data (Figure
3) show only the effects of impurities on the hardness
of pure iron. The three high-purity iron samples were
found to have different impurity levels; the Battelle
iron had about 35?50 ppm of impurities, the Bain
sample had about 110 ppm of impurities, and the
Lehigh sample had about 700 ppm of impurities. The
microhardness data show that, for these samples, the
microhardness increases with increasing impurity
content.
Assuming that the microhardness of the three iron
samples and the lunar iron particle is affected pri-
marily by the impurity content, as discussed above,
and that the effect is related only to the total im-
purity content and not to the content of selected im-
purities (perhaps a poor assumption since interstitial
elements generally have a more pronounced influence
on the hardness of iron than do substitutional ele-
ments), the close match between the microhardness
of the lunar iron particle and the Bain sample sug-
gests the total impurity content of the lunar iron
sample is in the order of 100+ ppm. Unfortunately,
the small size of the particle and the requirement not
to destroy it prevented analyses for impurities to ver-
ify the correlation indicated between microhardness
and the total impurity content.
Acknowledgments
We are indebted to Dr. K. F. J. Heinrich and Mr.
Charles Fiori of the Bureau of Standards for the
electron microprobe studies made on the small metal-
lic inclusion, and also to Messrs. Grover C. Moreland
and Harold Moore for mounting and preparing the
polished surface we studied.
Literature Cited
Edmond, L.
1958. Vickers-Knoop Hardness Conversion. Metals
Progress, 74(3):97.
Lea, F. G.
1936. Hardness of Metals. London: C. Griffin and Co.
Lenhart, R. E.
1955. The Relationship of Hardness Measurements to
the Tensile and Compression Flow Curves. Wright
Air Development Center Technical Report, 55:
114.
Lysaght, V. E., and A. DeBellis
1969. Hardness Testing Handbook. New York: Ameri-
can Chain and Cable Company.
O'Neill, H.
1967. Hardness Measurements of Metals and Alloys.
2nd edition. London: Chapman and Hall, 238
pages.
Rengstorff, G. W. P.
1968. Zone Purification and Analysis of Large Bars of
Iron. Memoires Scientifiques de la Revue de
Metallurgies LXV: 85-90.
Small, L.
1960. Hardness, Theory and Practice. Ferndale, Michi-
gan: Service Diamond Tool Company.
Staub, R. E., and J. L. McCall
1969. Metallographic Studies of an Iron-Nickel Meteo-
rite Following Long-Time Heat Treatments. Pro-
ceedings of the International Metallographic So-
ciety Second Annual Meeting, San Francisco,
California.
Vitovec, F. H., and H. F. Binder
1955. The Effect of Surface Preparation and Condition
on Microhardness. Wright Air Development Cen-
ter Technical Report, 55:372.
PETROLOGY AND VOLCANOLOGY
William G. Melson, Basaltic
Eugene jarosewkh, Nu?es Ardentes
George Switzer, 1O7A
and Geoffrey Thompson OI the lb>/U
Eruption of
Ulawun Volcano,
New Britain
ABSTRACT
Ulawan volcano, New Britain, (5?02'40"S.,
151?2O'15"E.) erupted between January 9, 1970,
and February 10, 1970, producing a small nuee ar-
dente about 0405 hours local time, January 22, 1970.
Ash eruptions of various intensities occurred through-
out the episode, and a 5.6-kilometer-long lava flow
was produced during the final phases. The main nuee
ardente was essentially a hot avalanche of basaltic
scoria which could not have emitted large volumes of
gases during flow, but had the high mobility and
marginal hot "hurricane" winds which characterise
nuees ardentes. These features lend support to the
experimental evidence of McTaggart (1956) that
high temperatures alone can greatly increase the
mobility of avalanching materials.
The chemical composition of the eruptives which
produced the nuees ardentes and the lava flows are
identical. The nuees ardentes evidently occurred
when exteremely large volumes of scoria were dis-
gorged at low ejection velocities, such that they fell
back as coherent fragmental flows on the upper slopes,
creating hot avalanches. The preferential direction of
the main avalanche, down the Northwest Valley re-
William G. Melson, Eugene Jarosewich, and George Switzer,
Department of Mineral Sciences, National Museum of Nat-
ural History, Smithsonian Institution, Washington, D.C.
20560. Geoffrey Thompson, Woods Hole Oceanographic
Institution, Woods Hole, Massachusetts 02543. (Mr. Jarose-
wich made the chemical analyses; Dr. Switzer, the electron
probe analyses; and Dr. Thompson, the trace element
analyses.)
fleets the existence of a "notch," or low point in the
crater wall, at the head of this valley.
The ejecta of the 1970 eruption are plagioclase
phyric high-alumina basalts, which are composition-
ally like previous eruptive rocks. Three new analyses,
petrographic features, and electron probe analyses are
given. Of special interest are the compositions of glass
inclusions in phenocrysts, which give evidence for the
existence of high-alumina basaltic liquids, and may
indicate mixing of basaltic and dacitic magmas prior
to eruptions.
This eruption of Ulawun was a comparatively small
one. The combined volume of the nuee ardente de-
posits and lava flows probably are between 0.002 and
0.010 cubic kilometers. The area devastated by the
nuees ardentes is about 2 square kilometers?a small
area compared to most historic nuees ardentes.
Introduction
Ulawun volcano, New Britain, had an eruption early
in 1970. First reports to the Smithsonian Center for
Short-Lived Phenomena indicated that a well-de-
veloped nuee ardente (glowing-cloud eruption) had
been produced on January 22, 1970. This led to a
field study of the eruption as part of an ongoing proj-
ect on the classification, properties, and origin of
nuees ardentes. Two other nuee-ardente-producing
eruptions were recently studied as part of this project
(1968 eruption of Mayon volcano, Phillippines, see
Moore and Melson 1969; 1968 eruption of Arenal
15
16 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
volcano, Costa Rica, see Melson and Saenz 1968).
This is an interim report dealing briefly with the re-
sults of the field observations at Ulawun volcano, and
of the laboratory studies of the samples from the
nuee ardente deposits and from an associated lava
flow.
Mr. R. Citron and the writer were at the volcano
from February 7 to 14, 1970. In addition to the
results reported here, extensive still photography and
16-mm cinematography were obtained by Mr. Citron.
We had hoped to use cinematography to get quanti-
tative information on the velocity and other features
of nuees ardentes, but we arrived too late. Nonethless,
extensive photographic coverage was obtained of the
nuees ardentes deposits and the surrounding devas-
tated zone. The still photographs (35-mm slides) and
motion pictures are now on deposit at the Smithso-
nian Institution.
We were accompanied throughout the field inves-
tigations by R. A. Davies, and part of the time by
G. A. M. Taylor and David Palfreyman, of the
Rabaul Volcanological Observatory, an organization
which is maintained by the Administration of the /
Territory of Papua, New Guinea, and draws its pro/
fessional staff from the Australian Bureau of Mineral
Resources, Geology, and Geophysics. We acknowledge^
the kind assistance and free exchange of information^
given by these gentlemen. R. A. Davies (192^0 and
1971) has discussed this eruption, and R. W. Johnson
(1970a) has described the volcano and its petrology
prior to the 1970 eruption. Further publications on
this eruption are contemplated by Davies, who
reached the scene of the eruption almost at its in-
ception, and remained at hand until its termination.
Nuees Ardentes
Nuees ardentes ("glowing-cloud eruptions") have re-
ceived much attention because of their incredible
destructive potential, as demonstrated by the oblitera-
tion of St. Pierre, Martinique, and its 30,000 inhabi-
tants by a nuee ardente from Mount Pelee, an erup-
tion described in the classic work on nuees ardentes
by A. Lacroix (1904). Since that time, numerous
other nuees ardentes have been described, and much
has been written about their mechanism of flow. One
of the most recent reviews is by McTaggart (1960).
Nuees ardentes are defined somewhat differently
by different authors, but perhaps the following defi-
nition is acceptable to most; it is the one used in
this report:
A nuee ardente is a type of volcanic eruption in-
volving two components: (1) a hot, normally but not
always incandescent, highly mobile avalanche com-
posed of either essential or accidental material or a
combination of the two, and (2) a hot envelope of
gases and ash, which is derived from the avalanche
but includes much heated air, which expands verti-
cally in billowing, ash-laden clouds, and also travels
laterally, predominantly parallel and extending be-
yond the sides of the avalanche, and may extend very
far beyond the final terminus of the avalanche; these
lateral, ground-following ash and gas clouds can at-
tain "hurricane" velocities, on the order of 100 mph,
and are responsible for most of the loss of life and
extensive damage associated with historic nuees
ardentes.
Ulawun Volcano: Background Information
The following summary is drawn largely from John-
son (1970a), who gives an account of the morphol-
ogy, geology, and petrology of Ulawun prior to its
1970 eruption, and gives a schematic topographic
map. Johnson (1970b) also has reviewed the rela-
tionship between lava composition, volcano position,
and the Benioff zone beneath New Britain. Ulawun
(The Father) is a symmetrical stratovolcano on the
north coast of central New Britain which rises from
a base at sea level to 2,300 meters. It has a pro-
nounced summit crater about 300 meters wide and up
to 130 meters deep, with notches which feed into the
major valleys, of which the Northwestern Valley
(Figure 1) was the course of the 1970 nuee ardente.
This Northwestern Valley is a pronounced "scar" on
the mainly smooth cone (Figure 1), and appears to
have been eroded by hot avalanches as well as by
normal erosive processes.
A pronounced older, deeply eroded, volcanic edific
trends roughly east-west along the south side of the
volcano (Figure 1), and is marked by a 160-meter-
high escarpment, which may be the remains of a
caldera or linear graben.
Eruptions of Ulawun appear to be characterized by
a mixture of strombolian and vulcanian eruptions.
Lava flows typically occur, and move as thin ribbons
down one or more valleys. These flows are interlay-
ered with the volcanic fragmental debris, mainly
NUMBER 9 17
FIGURE 1.?Vertical aerial photograph of Ulawun volcano, New Britain, taken June 17, 1948.
scoria, and buttress the volcano with an interleaved
framework emanating from the summit. The rela-
tive proportion of fragmental deposits and coherent
flows is unknown, but they are perhaps on the order
of 1:1.
Observations of the 1970 eruption show that nuees
ardentes at Ulawun are of the avalanche type (St.
Vincent type, or d'explosion vulcanienne type of
Lacroix, 1904), and occur when the volcano disgorges
large quantities of fragmental debris at low ejection
velocities, such that large quantities fall immediately
back on the slopes, coalescing to form hot avalanches.
Table 1 summarizes some of the principal features
of Ulawun volcano.
Summary of the 1970 Eruption
The eruption was a brief one (about 32 days) and,
in terms of total volumes of ejecta, was not a major
eruption. Based on preliminary measurements, the
hot avalanche and associated marginal ash deposits
18 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
total about 0.005 cubic kilometers, and the main lava
flow, that in the Southwest Valley, totals about 0.003
cubic kilometers. The total area devastated by the
nuee ardente of January 22 is about 2 square kilom-
eters. These estimates are calculated from the pre-
liminary measurements of Davies (1970). No esti-
mate can yet be made of the volume of airfall ash.
The area devastated by the nuee ardente is ex-
tremely small compared to the destruction caused
a major nuee ardente, such as that of January 21,
1951, at Mount Lamington, which ravaged over 200
square kilometers (Taylor 1958).
TABLE 1.?Summary of features of Ulawun volcano, mainly from Johnson (1970a)
and Fisher (1957)
SYNONYMS: Ulawon, Uluwun, The Father
TYPE: Stratovolcano
LOCATION: New Britain, Territory of New Guinea and Papua; opposite Lolobau Island (also
an active volcano), and 16 kilometers west of Bamus, an older dissected stratovolcano.
COORDINATES: 5?02'40"S., 151?20'15"E.
HEIGHT ABOVE SEA LEVEL: 2,300 meters, which slopes down to sea level; height above adjacent
sea floor 3,000 meters.
PREVIOUS ERUPTIONS: A number of mildly explosive events were recorded between 1915 and
1967, the most violent of these were in April 1915, and January 1967. No lava flows or
nuees ardentes were reported during these eruptions. The initial explosions seem to have
been most violent, and were followed by less violent eruptions for several months. Zones
of ash deposition are dependent on wind directions, which have ranged widely on the
different times of eruption.
PETROLOGY: Mainly porphyritic basic lavas, with plagioclase as dominant phenocryst, although
a few percent of pyroxene and olivine phenocrysts occur in most. Chemically, these lavas
are basalts, which are tholeiitic, i.e., they have low total alkalies. High alumina content
(17?19%) is a characteristic feature. Johnson (1970a) gives 4 chemical analyses, and 32
modal analyses of Ulawun rocks.
Davies (1970) has broken the eruption of Ulawun
into three overlapping phases:
1. January 9?14, 1970. Quiet vapor and ash
emission, white to grey vapor, occasional ejections of
ash, no audible activity.
2. January 15?27, 1970. Audible explosions; cli-
mactic explosion at 0405 local time, on January 22,
1970. Flows observed at 1650 the same day, moving
down Northwest Valley.
3. January 27-February 10, 1970. Quiet, effusive
phase, characterized by the major aa lava flow which
moved down the Southwest Valley. Activity slowly
declined, until February 10, 1970, when the main
flow ceased moving, and vapor-phase activity con-
tinued to decline.
A number of vents were active during the eruption,
all located circumferentially to the summit crater.
The elevation of Ulawun grew through cone build-
ing during the eruption, starting at about 2,250
meters high, and ending at about 2,340 meters high.
GEOPHYSICAL CHANGES ASSOCIATED
EMISSION OF THE NUEES ARDENTES
WITH THE
Davies (1970) also reports the changes in tilt and
seismicity associated with the major explosive activ-
ity of January 22, 1970. Tiltmeter readings of the
tangential component indicate a high in the inflation
of the volcano prior to this event, and a rapid de-
flation of 28 seconds of arc between February 5 and
8, 1970. Continuous harmonic tremor characterized
the seismicity. Impulsive tremors were not noted.
Ground motion increased steadily from January 16,
1970, when measurements began, and reached a
peak in excess of 25 microns during the night of
January 21-22, 1970, which coincided with the first
eruption of a nuee ardente. Ground motion decreased
after that, although significant peaks were recorded
on the afternoon of January 25 and the night of
January 27, 1970, the latter coinciding with the
production of three nuees ardentes, and the beginning
of the quiet effusive phase.
NUMBER 9 19
Account of the Major Nuee Ardente
Davies (1970), who was near the volcano, gives this
account of the event:
A "major" eruption from the northwest vent occurred at
0405 hours LT. The eruption, of Pelean type, was not pre-
ceded by any abnormal audible activity and explosive ejec-
tions of lava fragments and ash were observed prior to the
eruption. Dense ash-laden vapour clouds accumulated above
the summit, successive emissions gradually enveloping the
northwest vent. An eye witness described how some 40 sec-
onds later, abundant spots of light could be seen on the
lower northwestern slopes. This can be ascribed to the ava-
lanche (ash-flow) component of the nuee emerging from
the Northwest Valley. Incandescent blocks and igniting
vegetation marked the path of the nuee down into the
forested region. The dark ash cloud component, which had
blanketed the northwestern slopes, expanded rapidly to above
3,000 meters within 10 minutes. Between 0412 hours and
0432 hours an electric storm raged within the ash cloud and
at 0430 hours a light shower of accretionary lapilli (mud
pellets) and rain fell on the Mission.
Aerial inspection at 0700 hours (LT) revealed an area of
almost total devastation 2.4 kilometers long and 0.8 kilom-
eters wide, extending down through the forested region on
the lower northwestern flanks of the volcano.
The northwest and .northeast vents were continually active
throughout the day, with little response from the southwest
vent. Ash-laden vapour emission was again voluminous, dense
clouds rising to over 5,000 meters. Cumulus cloud, together
with smoke haze rising from burning vegetation marginal to
the devastated area, rendered observation difficult. At 1650
hours (LT) a lava flow was seen descending the Northwest
Valley, formed by the merging of two streams from the
northeast and northwest vents, respectively. This flow was
probably initiated during the early morning and by 1800
hours it had reached an altitude of 1,000 meters above sea
level.
During the night all four vents were active, the lava
temperature on emission being estimated at 1000?1100? C
(by comparison with temperature-colour charts). Occasion-
ally explosive ejections from two or more vents were syn-
chronised, the more rhythmic ejections from the northeast
and northwest vents littering the upper slopes with great
volumes of incandescent blocks. Pulses of lava were fre-
quently expelled from the north and northwest vents, the
incandescent fronts surging down into the Northwest Valley.
Audible explosions, which had persisted at irregular intervals
throughout the day, increased in intensity during the night.
OBSERVATIONS ON THE DEVASTATED ZONE AND HOT
AVALANCHE DEPOSITS
There was insufficient time during our field work to
prepare a detailed account of our observations on the
"%**
FIGURE 2.?Hot avalanche (light gray deposit in middle of photograph) filling lower portion
of Northwest Valley, February 7, 1970. Low altitude, oblique aerial photography.
FIGURE 3.?Hot avalanche deposits in Northwest Valley,
looking southeast toward Ulawun, at an elevation of about
500 meters. February 10, 1970.
devastated zone and hot avalanche deposits, or to
refine our initial observations. Some of these observa-
tions, recorded in a field journal immediately after
our first entry into the devastated zone, follow. Fig-
ures 2-12, all taken by Mr. R. Citron, illustrate
some features of the devastated zone and the hot
avalanche deposits.
We reached the devastated zone on February 7, 1970, some
16 days after the major nuee ardente had occurred. To
reach the zone, we travelled south from a point near the
airstrip, which is north of Ulawun and on the coast, through
heavy rain forest along a gradually rising stream course,
which, at an altitude of 450 meters, intersected the devas-
tated zone. The first evidence of nearing the devastation was
the smell of smoldering wood and evidence of a clearing of
the rain forest ahead. The first damage we found was yel-
lowed, nvithered vegetation. This withering became ubi-
quitous' within ten to a few tens of feet after these first,
irregularly distributed signs, and gave way to a totally dev-
astated zone, where all trees had been either uprooted or
snapped off a short distance up their trunks. Uprooted trees
were aligned pointing down slope, that is, away from the
FIGURE 4.?Water vapor rising from the hot avalanche deposits after a rainstorm,
February 8, 1970.
NUMBER 9 21
volcano. These fallen trees included eucalyptus trees which
were up to 150 feet long. We noted very little ash in this
zone, although a thin veneer, less than a centimeter thick,
was present in places on the ground.
We traversed this zone of felled trees with difficulty, mov-
ing southwest so as to intersect the avalanche itself. After
about 300 to 400 meters, we reached a blackened, ash-
covered zone, where trees were slightly charcoalized on the
side facing the volcano, and where some, those that had
been ignited at the time of the avalanche, were still smol-
dering. This zone was a few hundred feet wide, and was
confined to the margin of the channel down which the hot
avalanche had traveled.
Remarkably, a small number of tree trunks were still
standing near the margin but within the avalanche depos-
its, although their tops had been snapped off and carried
away. These trees were smoldering from the portion buried
in the avalanche deposits and, where their bases had been
burned through, had collapsed.
Smoke rose from buried,' smoldering wood in the ash flow,
and there were small, only weakly active fumeroles. The
avalanche here appeared to be several hundred feet wide,
and was composed largely of dark grey> to black, sometimes
red, scoriaceous basaltic bombs in a matrix of coarse ash,
mainly composed of seriate fragments derived from the
shattering and breaking up of the bombs during the ava-
lanche. The bombs commonly had a knobby, "puffed-up"
appearance, and typically their interiors were more vesicular
than the oujter portions, as if the outer vesicles had been
flattened and compressed by pounding with other bombs
and blocks during the avalanche. The "puffed-out" appear-
ance suggests that expansion and gas release continued dur-
ing the avalanche, and that the bombs were in a thick,
pasty state during their ejection and subsequent avalanching.
The bombs ranged widely in size, from large, up to five feet
wide, on down to a few inches across.
Less abundant, but still conspicuous components, were
angular light-colored, apparently holocrystalline medium-
grained fragments, ranging up to large blocks, a few feet
across. These appear to be accidental inclusions, perhaps
from the plug and walls of the vent.
Subsequent field work gave a better overall view
of the avalanche deposits. From place to place, the
deposits range widely in the amount of coarse (e.g.,
greater than 6-inch maximum diameter) and the
amount of fine clasts. There is no evidence of sorting
in some places, whereas elsewhere there are areas,
FIGURE 5.?Hummucky topography of the avalanche near terminus. Direction of "flow" from
left to right, and about parallel to the direction indicated by the alignment of fallen trees in
the background. At an elevation of about 500 meters, February 8, 1970.
22 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 2.?Analyses of some 1970 eruptives of Ulawun
volcano (Chemical analyses by E. Jarosewich; trace
element analyses by G. Thompson)
Constituent
SiO2
A12O3
Fe2O3
FeO
MgO
CaO
NasO
K2O
H2O+
H2O-
TiO2
P2O5
MnO
Totals
Q
OR
AB
AN
DI
HY
MT
IL
AP
Fe/Mg
B
Ba
Co
Gr
Gu
Ga
Li
Ni
Sr
V
Y
Zn
Zr
Chemical analyses (weight percent)
1* 2* 3*
52.77 52.62 51.51
18.00 18.31 18.96
2.99 3.23 2.72
7.14 6.81 7.50
4.98 4.94 4.92
10.48 10.45 11.14
2.29 2.40 2.24
0.37 0.40 0.32
<0.1 <0.1 <0.1
0.01 0.01 0.03
0.96 0.96 0.79
0.08 0.08 0.07
0.14 0.14 0.16
100.21 100.35 100.36
CIPW Norms
7.10 6.60 4.50
2.19 2.36 1.89
19.38 20.31 18.96
37.74 38.01 40.74
11.25 10.88 11.59
16.20 15.49 17.06
4.34 4.68 3.94
1.82 1.82 1.50
.19 0.19 .16
.36 .34 .40
Trace element analyses (parts per million)
30 28 15
100 110 75
40 40 38
30 35 35
160 110 90
20 18 18
10 8 7
21 23 22
350 300 280
300 300 300
25 '24 22
180 190 190
25 28 20
TABLE 3.?Comparison of the 1970 Ulawun e'jecta,
represented by the analysis of the scoria (Table 2)
in column 1, with an average oceanic deep-sea basalt
in column 2 (Melson et al. 1968), and overall average
basalt and andesite (Taylor and White 1966) in
columns 3, and 4 respectively
Constituent
SiO2
A12O3
Fe2O8
FeO
MgO .
CaO
Na2O
K2O
TiO2
P2OS
MnO . .
B
Ba
Co
Cr
Cu ...
Ga
Li
Ni
Sr
V
Y
Zn
Zr
Chemical analyses (weight percent)
1
52.77
18.00
2.99
7.14
4.98
10.48
2.29
0.37
0.96
0.08
0.14
2
49.21
15.81
2.21
7.19
8.53
11.14
2.71
0.26
1.39
Trace element analyses
30
100
40
30
160
20
10
21
350
300
25
180
25
7
12
32
296
87
18
8
123
123
289
43
122
100
5
48.9
15.7
10.7
8.70
10.8
2.32
1.02
1.82
4
59.5
17.2
6.10
3.42
7.03
3.68
1.60
.70
(parts per million)
5
250
48
200
100
12
10
150
465
250
25
110
10
270
24
56
54
16
10
18
385
175
21
110
* Column 1: Scoriaceous basalt bomb from 1,600 feet
level of main avalanche produced by nuee ardente of Jan-
uary 22, 1970. Sample number USNM 112480.
some almost dune-shaped with their long axis in
places perpendicular and in other places parallel to
the direction of flow, which appear to be well sorted,
consisting mainly of coarse?sand-sized fragments.
Typically, the deposits were estimated to be about
30 feet thick, although there was no way to accu-
Column 2: Massive basalt from lava flow of 1970 eruption.
Collected from terminus of slowly moving flow on February
8, 1970, at elevation of 1,450 feet along the upper dry
course of the Matisibu River, the Southwest Valley of Ula-
wun volcano. Sample number USNM 112481.
Column 3: Light-gray holocrystalline basalt xenolith from
main avalanche deposit at 1,600 foot level of the Northwest
Valley. Sample number USNM 112482.
rately measure this thickness. The deposits do not
thin noticeably toward the terminus, but rather thin
abruptly, in a distance of a few hundred feet, ending
in a chaotic melange of large blocks, bombs, and
partly shattered and charcoalized tree trunks, with
perhaps about 60 to 70 percent of finer material. To
a first approximation, the deposits show no great
range in thickness from their point of first "heavy"
deposition, near the base of the northwest "shoot,"
to their terminus, another kilometer or so away.
The overlaying hot-ash "hurricane" proceeded be-
yond the terminus of the avalanche, withering vege-
tation up to 300 meters further down the "avalanche"
course. Beyond the avalanche terminus, the dry, flat-
tish stream bed curved gently. The totally withered
zone gave way to zones where the withering alter-
nates from one side, at one curve, to the other side,
at the next curve, as if the "blast" became narrow,
and ricocheted at each turn.
Composition of the Eruptives
Here, we search for significant chemical differences
between the melts which produced the nuee ardente
and those that produced the lava flow. We also se-
lected a sample of a common variety of xenolith in
the main nuee ardente avalanche deposit, which, in
the field, appeared to be quite different from the
essential ejecta.
Table 2 shows that all these eruptives are chemi-
cally basalts (between 45 and 55 percent silica), that
they are slightly quartz normative (4.5 to 7.1 per-
cent), and that they are remarkably similar to one
another, even in their trace element compositions.
The slightly higher ferric/ferrous iron ratio in the
flow compared to the bomb is an expected conse-
quence of the longer high-temperature exposure of
the flow to the atmosphere. The very slightly lower
contents of B, Ba, Sr, and Zr in the xenolith com-
FIGURE 6.?Felled trees, uprooted, or snapped off, during nuee ardente of January 22, 1970.
Northeastern side of avalanche channel at about 400 meters elevation. Photograph taken
February 8, 1970. Vapor from steaming avalanche deposits in foreground.
24 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 4.?Comparison of the 1970 eruptives (columns 1 and 2) with those of earlier
eruptions (columns 3-6) compiled by Johnson (1970a)
Constituent
SiOa
A12O3
Fe2O8
FeO
MgO
GaO
Na2O . . . .
KaO
HaO +
H2O-
TiO2
p2o?
MnO
/?
52.77
18.00
2.99
7.14
4.98
10.48
2.29
0.37
0.1
0.01
0.96
0.08
0.14
100.21
Chemical
2
52.62
18.31
3.23
6.81
4.94
10.45
2.40
0.40
0.1
0.01
0.96
0.08
0.14
100.35
analyses (weight percent)
3
51.1
17.5
4.30
6.35
5.85
10.8
2.40
0.30
0.07
0.23
0.92
0.07
0.13
100.15
4
52.47
19.16
3.40
5.82
4.90
10.76
2.45
0.32
0.14
0.03
0.75
0.09
0.17
100.51
5
52.10
16.78
3.07
7.32
6.52
10.84
2.19
0.32
0.08
0.03
0.75
0.09
0.19
100.39
6
51.79
18.06
2.51
7.25
5.85
10.98
2.19
0.32
0.17
0.06
0.74
0.0$
0.17
100.26
?Column /: Basaltic bomb, from hot avalanche deposits, 1970 eruptions. E. Jarosewich,
analyst.
Column 2: Lava flow, sampled at terminus, 1970 eruption. E. Jarosewich, analyst.
Column 3: Ash, 1967 eruption, deposited on Lolobau Island (Jorgensen, analyst, in John-
son, 1970a).
Columns 4-6: A.J.R. White, analyst, in Johnson (1970a).
Columns 4 and 5: Prehistoric flows and a sill.
Column 6: Bomb from southwestern slope.
pared to the essential ejecta are also small but note-
worthy features of the analyses.
The Ulawun rocks have high alumina contents
compared to most basalts. This is reflected modally by
the abundance of plagioclase. Other differences with
"average" basalts are brought out in Table 3, and
include higher boron, and lower nickel, chromium,
and zirconium.
The trace element abundances show both basaltic
and andesitic affinities (Table 3). The very low
nickel and chromium contents are close to the values
for andesites, and low compared to most basalts. On
the other hand, the cobalt and vanadium contents are
close to those of basalts. The zirconium is anomalously
low, and the copper, vanadium, boron, and zinc are
anomalously high compared to both average basalts
and average andesites. Gallium, yttrium, and lithium
show little differences between the Ulawun samples,
and the average basalts and andesite.
The Ulawun samples have low sodium contents
compared to deep-sea basalts and to most island and
continental basalts, and their potassium content, al-
though higher than most deep-sea basalts, is lower
than most island and continental basalts. The iron/
magnesia ratio is higher, and the total iron is close
to and magnesia is less than the average basalts.
Table 3 indicates that the Ulawun basalts cannot
be related to either of the average basalts by a simple
near-surface (low pressure) fractionation scheme. If
we regard the average basalts as parents, we see that
separation of some olivine and chromite would ac-
count for the lower Mg, Ni, and Cr, and higher Si
of the Ulawun samples compared to these hypothetical
parents. However, the low Na and Zr content of the
Ulawun basalts is inconsistent with such near-surface
olivine fractionation of these hypothetical parents.
The recent and prehistoric eruptives of Ulawun are
of relatively uniform composition (Table 4). All are
NUMBER 9 25
TABLE 5.?Hand-specimen and petrographic description of the analyzed basalts from
the 1970 eruption of Ulawun Volcano. Mineral abundances in volume percent
Bomb Flow Xenolith
STRUCTURE (hand-specimen description)
Black, highly vesicular, nubby,
"puffed-out" outer portion;
more vesicular in interior than
on surface. Black to reddish
oxidized glass on surfaces.
Dense; black; scattered irreg-
ular cavities; porphyritic.
Light gray; porphyritic; lacks
significant pore space. Modal
analysis (volume %): Plagio-
clase phenocrysts: 36.2; Pla-
gioclase microlites: 21.4; Py-
roxene (total) : 29.3; Olivine:
2.0; Oxides: 2.7; Silica (cris-
tobalite): 1.5; Interstitial glass
and unresolvable material:
5.9; Pores: 1.0.
GROUNDMASS
Clear glass, plus microlites Brownish glass, with micro- Mainly crystalline, with about
and quench crystals of plagio- lites of plagioclase, pyroxene 40 percent pyroxene, 50 per-
clase and pyroxene. and opaque(s). cent plagioclase, and a few
percent oxide, and possibly
clear interstitial glass.
PHENOCRYSTS
Plagioclase (1-5 mm; about
35% volume percent); olivine
(mainly small, 1-2 mm across;
less than 5 volume percent),
rare rounded phenocrysts of
clinopyroxene, and still rarer
grains of hypersthene.
Plagioclase (1?5 mm across)
about 35 volume percent;
rounded olivine grains (main-
ly small, 1-2 mm, but some
up to 7 mm across; less than
5 %); rare rounded pheno-
crysts of clinopyroxene, and
still rarer grains of hyper-
sthene.
Plagioclase (1-5 mm; about
35 volume percent); olivine
(small, 1-2 mm across; less
than 5 percent), pyroxene
(dark green clinopyroxene, a
few large and rounded, up to
8 mm long).
high-alumina, quartz-normative basalts, with rela-
tively low total alkali contents, and K2O contents be-
tween 0.30 and 0.40 percent. Johnson (1970a) notes
that they are comparable in chemistry and mineral-
ogy to alumina-rich rocks with abundant plagioclase
phenocrysts from various parts of Japan.
Mineralogy and Petrography
The three samples selected for chemical, electron
microprobe, and petrographic analyses are all basalts,
and all are porphyritic, with up to 40-volume-percent
phenocrysts (crystals greater than about 0.5 mm in
maximum dimension). The dominant phenocryst is
plagioclase, which composes about 35-volume per-
cent. These have highly calcic cores (bytownite, An8s,
to anorthite, An90) with weak zoning to An80 at their
rims. In volume, there are less than 5 percent large,
normally rounded pyroxene and olivine grains, which
may be partially resorbed phenocrysts. Olivine occurs
in all the samples, and is around F074. No complete
analyses were made of the opaques, but they appear
to be mainly homogeneous titanmagnetite with about
17 percent TiC>2. Table 5 compares the hand speci-
26 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 6.?Pyroxene analyses via electron microprobe. George Switzer, analyst. First
three pairs (1?2, 3-4, 5?6) are from the rims and cores of a single phenocryst in
each of the three analyzed samples in Table 2, the basaltic bomb, the flow, and
xenolith, respectively. The final analysis (Column 7), is of a pigeonite microlite in
the groundmass of the basaltic xenolith
Constituent
SiOa .
TiO2
AI2O3
Fe
MnO
MgO
GaO
Na2O
K2O
Fe
Mg
Ca
Bomb
1
51.5
0.4
2.3
7.4
0.2
17.2
20.6
0.2
0.05
14.4
46.1
39.5
Electron
2
51.4
0.4
2.5
8.4
0.2
17.3
19.3
0.2
0.05
16.3
46.5
37.2
microprobe
Flow
3
53.2
0.5
3.2
6.6
0.2
16.7
19.5
0.2
0.05
analyses
4
52.8
0.5
3.3
8.0
0.2
15.2
19.0
0.2
0.05
Molar ratios (mole
13.4
47.0
39.6
16.7
43.9
39.4
(weight
5
52.5
0.4
3.5
7.8
0.2
15.6
20.0
0.2
percent)
15.8
43.7
40.4
percent)
Xenolith
6
52.7
0.4
3.4
7.8
0.3
17.0
18.8
0.2
15.6
47.0
37.4
7
52.6
0.3
1.0
19.9
0.4
16.9
6.3
0.2
40.1
47.2
12.6
FIGURE 7.?Uprooted trunk of eucalyptus tree. Northeastern
margin of avalanche channel. Note abrasion of trunk. Feb-
ruary 10, 1970.
men and petrographic features in the three analyzed
samples, and Figures 13 and 14 give comparisons of
petrographic features.
The pyroxene phenocrysts are typically augite in all
three samples (Table 6). Rarely, hypersthene pheno-
crysts occur, and intermediate pigeonite microlites
(analysis 7, Table 6) were found in the matrix gran-
ules of the only holocrystalline basalt examined?the
basaltic xenolith. Analyses of the rim and core of one
augite phenocryst was made in each of the three sam-
ples (Table 6, analyses 1?6). These show that there
is slight inverse zoning in the flow and bomb augites.
The cores are slightly higher in Fe and lower in Ca
than the rims.
The 1970 eruptives are petrographically similar to
those of earlier eruptions described by Johnson
NUMBER 9 27
FIGURE 8.?Felled and snapped-off tree trunks near terminus devastated zone. Avalanche de-
posits to right (note vapor); looking northwest (downslope). Unaffected trees in background.
February 8, 1970.
(1970a), except he reports that pigeonite phenocrysts
are commonly present, although in small amounts.
Johnson (1970a) gives more detailed descriptions of
reactions between phenocrysts and liquids than given
here, including oxidation and resorption of the oli-
vine phenocrysts.
COMPOSITION OF GLASS INCLUSIONS IN PHENO-
CRYSTS: EVIDENCE FOR HIGH-ALUMINA BASALTIC
LIQUIDS AND OF MIXED MAGMAS
Some of the phenocrysts and large rounded grains
contain inclusions of clear to light brown glass (Fig-
ure 14). Electron probe analyses of some of these
glass inclusions are in Table 7. The glass (analysis 3,
Table 7) in a rounded olivine crystal in the analysed
basaltic bomb is quite close to the bulk composition
of the bomb, and suggests that this bulk analysis is in
fact representative of a liquid composition of a high
alumina basalt. In other words, the phenocrysts grew
from a high alumina basaltic liquid, and phenocrysts
have been neither added nor subtracted from this
parent during crystallization. The glass inclusions in
the plagioclase and augite phenocrysts are more diffi-
cult to understand, particularly the high silica
(65.4%) glass in the augite phenocryst from the lava
flow. This does not appear to be a residual glass, as
the alkalies are not noticeably higher than in the bulk
sample. One can explain this relationship by assuming
that the clinopyroxene crystal grew in a different
magma, e.g., a dacite, and that this magma was later
mixed with basaltic magma. Indeed, cases of erup-
tions of mixed magmas have been reported (e.g.,
MacDonald and Katsura, 1965). All analysed Ula-
wun lavas (Table 4) have remarkably similar bulk
composition, however, and thus if mixing of two
28 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
magmas is occurring, they must mix in the same
proportions in each eruption, and involve mixing of
basaltic magma with very small amounts of dacitic
magma.
Significance of Nuees Ardentes from Basaltic
Volcanoes
Normally, nuees ardentes are thought of as products
of explosive eruptions of acid to intermediate rocks.
Taylor (e.g., 1960) has pointed out that nuees ar-
dentes are by no means restricted to such composi-
tions, and reports of basaltic nuees ardentes of
Manam volcano, New Guinea. He also notes the
reports of nuees ardentes from basaltic eruptions
recorded by Lacroix (1904) and by Hantke (1951,
also at Manam volcano).
Gas emission during avalanching may be important
in giving the characteristic high mobility and billow-
ing ash emission in nuees ardentes (see Taylor, 1958,
page 47, for a review of this mechanism). However,
McTaggart (1960) has demonstrated that high tem-
perature alone without gas emission during flow can
account for the high mobility of nuees ardentes. He
experimented by pouring sand at various tempera-
tures down an incline and recording its distance of
flow and geometry after deposition on a plane at the
TABLE 7.?Composition of glass inclusions in plagio-
clase (column 1), augite (column 2), olivine (column
3), compared to the bulk composition (column 4, an-
alysis 1, Table 1). Analyses by electron microprobe by
G. Switzer. The glass inclusions are in the following
samples: columns 1 and 3 are in sample 1, Table 2;
column 2 is in sample 2, Table 2
Constituent
SiO2
AI2O3
TiO2
FeO*
MgO
CaO
NaaO
K2O
MnO
Q
OR
AB
AN
DI
HY
IL
Electron microprobe analyses
(weight percent)
12 3 4
55.1 65.4 51.1 52.8
13.3 18.5 17.9 18.0
1.6 1.0 1.1 1.0
15.8 5.0 10.3 10.8
4.2 2.5 4.7 5.0
7.1 5.9 10.7 10.5
2.0 2.1 n.d. 2.3
0.7 0.6 n.d. 0.4
0.2 0.1 n.d. 0.2
CIPW Norms
11.3 33.1 ? ?
0.4 0.4 ? ?
16.9 17.8 ? ?
27.1 29.3 ? ?
7.0 ? ? ?
33.7 14.0 ? ?
3.0 1.9 ? ?
* All Fe recalculated as FeO.
n.d. = not determined.
FIGURE 9.?Splintered tree trunk in avalanche deposits near
terminus. This splintering occurred during transport of the
trunk in the avalanche. February 10, 1970.
-"' ? " '"* ' ?; *:-*' , Ai
? ??, ??? x $i/
FIGURE 10.?Surface of the avalanche deposits, showing abundance of lapilli-sized fragments.
February 10, 1970.
FIGURE 11.?Broken, scoriaceous bomb. Surface of avalanche deposits. February 10, 1970.
30 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 12.?Light-colored holocrystalline basalt xenoliths in the avalanche deposits.
base of the incline. The existence of basaltic nuees
ardentes mentioned above and of the 1970 eruption
of Ulawun volcano give strong support to McTag-
gart's idea. Basalts characteristically contain low con-
tents of dissolved volatiles, and hence could not emit
large volumes of gases during avalanching. True, the
fragments in the avalanche deposits at Ulawun are
vesicular, but are scoriaceous rather than pumiceous,
and could not have emitted large quantities of vol-
canic gases. Gas emission from the avalanching ma-
terial will enhance the mobility, but evidently is not
a prerequisite of fragmental materials which will pro-
duce nuees ardentes. Recently, nuees ardentes were
produced totally of nonessential, but hot material,
ejected during the explosive opening of a new vent
on the slopes of Arenal volcano, Costa Rica (Melson
and Saenz 1968). Here, as at Ulawun, gas emission
from the fragments in the avalanche had little to do
with the production of nuees ardentes. The mobility
at these volcanoes appears to have something to do
with the constant inclusion and heating of air, with
concomitant expansion. To my knowledge, an analy-
sis of precisely how this mechanism works is not yet
available, and is an objective of our further studies.
NUMBER 9 31
FIGURE 13.?Comparison of textures in the three analyzed ejecta of the 1970 eruption of Ula-
wun volcano. Width of field equals 2 mm in each photomicrograph, A. Central, vesicular por-
tion of the analyzed basaltic breadcrust bomb (analysis 1, Table 2). Colorless phenocrysts are
plagioclase; gray phenocrysts are augite. Matrix, crowded with microlites and minute oxide
grains, is opaque, B. Basalt, from lava flow, showing abundance of devitrified glass inclusions
in plagioclase phenocrysts and with clear overgrowths on margin. Plagioclase also abundant as
microlites in. matrix. Note scarcity of vesicles, c. Basalt, from lava flow, with large rounded
phenocryst of augite. Analyzed sample (analysis 2, Table 2). D. Basaltic xenolith. Holocrystal-
line, with comparatively large grains of opaques (magnetite). Inclusions in plagioclase are
holocrystalline, and mineralogically similar to the bulk sample (dominantly pagioclase and
augite).
FIGURE 14.?Basalt, with plagioclase phenocrysts rich in glass inclusions. Phenocryst is about
1 millimeter wide. Analysis 1 (Table 7) is of one of these glass inclusions.
Literature Cited
Davies, R. A.
1970. The 1970 Eruption of Mt. Ulawun, New Britain.
Smithsonian Institution Center for Short-lived
Phenomena Report, May 15, 1970: 14.
1971. The 1970 Eruption of Mt. Ulawun, New Britain.
Note on Investigation, 70501. 30 pages. Territory
of Papua of New Guinea, Geological and Vol-
canological Branch.
Fisher, N. H.
1957. Catalogue of the Active Volcanoes of the World
Including Solfatara Fields, Part 5, Melanesia. In-
ternational Volcanological Association, Naples.
Hantke, G.
1951. Ubersicht uber die Vulkanische Tatigkeit 1941?
1947. Bulletin Volcanologique, Series II, 11.
Johnson, R. W.
1970a. Ulawun volcano, New Britain: Geology, Petrology,
and Eruptive History Between 1915 and 1967.
Australian Bureau of Mineral Resources, Geology
and Geophysics, Report 1970/21: 43.
1970b. Volcanoes of New Britain: a Summary of Pre-
liminary Petrological Data. Australian Bureau of
Mineral Resources, Geology and Geophysics, Rec-
ord 1970/72: 2-5.
Lacroix, A.
1904. La Montagne Pelee et Ses Eruptions. Paris: Mas-
son et Cie. 626 pages.
MacDonald, G. A., and T. Katsura
1965. Eruption of Lassen Peak, Cascade Range, Cali-
fornia, in 1915: Example of Mixed Magmas.
Geological Society of America Bulletin, 76:475-
482.
McTaggart, K. C.
1960. The Mobility of Nuees Ardentes. American
Journal of Science, 258:369-382.
Melson, W. G., and R. Saenz
1968. The 1968 Eruption of Volcano Arenal, Costa
Rica: Preliminary Summary of Field and Lab-
oratory Studies. Smithsonian Center for Short-
Lived Phenomena, Report, November 7, 1968.
Melson, W. G., G. Thompson, and T. H. van Andel
1968. Volcanism and Metamorphism in the Mid-Atlantic
Ridge: 22?N. Journal of Geophysical Research,
73:5925-5941.
Moore, J. G., and W. G. Melson
1969. Nuees Ardentes of the 1968 Eruption of Mayon
Volcano, Philippines. Bulletin Volcanologique,
33(2):600-620.
Taylor, G. A.
1958. The 1951 Eruption of Mount Lamington, Papua.
Australian Bureau of Mineral Resources, Geology
and Geophysics, Bulletin, 38:117.
1960. An Experiment in Volcanic Prediction. Austra-
lian Bureau of Mineral Resources, Geology, and
Geophysics, Record, 74.
Taylor, S. R., and A. J. R. White
1966. Trace Element Abundance and Andesites. Bulle-
tin Volcanologique, 24:177-194.
R. F. Fudali Experimental Data
Bearing on the
Paragenesis of
Two Hawaiian Basalts
from Kilaeua Volcano
Introduction
Mineral paragenesis (the order of formation of min-
erals in time succession) in igneous rocks is a matter
of much interest to petrologists concerned with com-
positional trends in the residual melts, and with the
genetic relations between different rock types which
are associated in time and place.
Paragenesis is, of course, largely a function of the
major element composition of the silicate melt but
can also be influenced by such variables as: (1) Total
confining pressure; (2) partial pressures (or fugaci-
ties) of H2O and O2; and (3) whether or not equi-
librium between crystals and melt is maintained as
the crystallization proceeds.
In 1968, Dr. T. L. Wright of the United States
Geological Survey's Hawaiian Volcano Observatory
sent me specimens of two basalts which had issued
from the Summit of Kilauea Volcano in 19-19 and
1954, respectively. Although separated in time by 35
years the two lavas are virtually identical in their
major element content (Table 1).
However, these "isochemical" lavas apparently had
different paragenetic sequences, presumably due to
differences in one or more of the other variables
mentioned above. This conclusion was based on the
microscopic examination of samples from both rock
units which exhibited varying crystal/glass ratios.
R. F. Fudali, Department of Mineral Sciences, National
Museum of Natural History, Smithsonian Institution, Wash-
ington, D.C. 20560.
TABLE 1.?Chemical analyses of two Kilauean basalts
(weight percent)*
Constituent
SiO2
A12O3
TiO2
Fe2Oa
FeO
MgO
MnO
CaO
Na2O
K2O
H2O+ ....
H2O-
P2OS
CO2
Cl ...
F
Subtotal ..
Less 0
Total
1
50.20
14.04
2.74
1.83
9.50
7.03
0.17
11.49
2.25
0.57
0.1
0.02
0.27
0.02
?
100.14
1919
2
50.02
13.87
2.76
1.35
9.76
7.45
0.17
11.45
2.33
0.53
0.08
0.02
0.26
0.02
0.02
0.04
100.07
0.02
100.05
3
50.20
13.73
2.72
2.65
8.80
7.20
0.17
11.56
2.25
0.57
0.00
0.00
0.28
0.01
0.02
0.05
100.21
0.02
100.19
)54
4
50.09
13.79
2.68
1.80
9.59
7.31
0.17
11.51
2.28
0.53
0.08
0.02
0.26
0.01
0.01
0.05
100.18
0.02
100.16
* All analyses were performed at the U.S. Geoldgical Sur-
vey Rock Analysis Laboratory, Denver. Column 1: Analyst,
Lucille Tarrant. Column 2: Analyst, Ellen Danials. Column
3: Analyst, Lois Trumbull. Column 4: Analyst, Elaine
Munson.
33
34 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
The change in crystal/melt ratio in the different
specimens is normally a function of the temperature
that portion of the rock unit attained before it was
effectively quenched. Thus, by judicious selection of
samples one can examine, at least in a crude way, a
temperature and crystallization sequence for the rocks
in question. Dr. Wright (personal communication)
wrote:
The Halemaumau lava lake overflows [1919] show sub-
equal amounts of pyroxene and plagioclase at crystallinities
varying from 5-20%. Evidently the two phases came out
simultaneously . . . and the two minerals crystallized at
nearly the same rate with falling temperature. By contrast,
in the current series of flows [1954] which come up quickly
from a depth of about 3 km* (,-^1000 bars total pressure)
clinopyroxene comes out distinctly earlier than plagioclase
(Cpx=8% by volume when plagioclase first appears) and
the rate of plagioclase crystallization does not equal that of
pyroxene until several tens of degrees lower temperature. I
would like to explain the contrast by saying that the over-
flows from Halemaumau lava lake were degassed and equili-
brated at approximately 1 atm. total pressure whereas the
current flows retained a paragenesis on eruption character-
istic of some PH o lower than 1000 bars but much greater
than 1 atm.
This theory was supported by the first chemical
analysis of the 1954 basalt (Table 1, column 3) which
reported a total iron content identical to that of the
1919 basalt but a substantially lower ferrous/ferric
iron ratio?i.e., the 1954 basalt was apparently oxi-
dized relative to the 1919 basalt, a natural conse-
quence of higher associated water (and oxygen)
fugacities.
However, subsequent chemical analyses and labora-
tory high-temperature, controlled atmosphere experi-
ments performed here demonstrated this hypothesis to
be incorrect. Further, total pressure differences and
possible disequilibrium effects were also eliminated as
possible causes of the observed paragenetic differ-
ences. The inescapable conclusion is that an appar-
ently trivial difference in A12O3 content (virtually
within the analytical error of the analyses) was re-
sponsible for stabilizing plagioclase at significantly
different temperatures in the two melts.
FeO and Total Iron Analyses
Divalent iron can be partially oxidized in sample
preparation and digestion so that correct FeO deter-
minations are notoriously difficult. For this reason one
should always view reported FeO values with some
scepticism, especially if the reported FeO value is
lower than expected. One of the first things I did
upon receiving the rock samples was to reanalyze
them for FeO and total Fe as FeO.
The samples were very carefully ground to ?80
mesh in an agate mortar, under alcohol. The analyti-
cal procedures used were standard wet chemical
methods previously described by Muan and Osborn
(1956) and myself (1965). Results of these analyses
are shown in Table 2. My results indicate an identical
ferrous/ferric iron ratio for the two rocks; a subse-
quent U.S. Geological Survey analysis (Table 1, col-
umn 4) supports this conclusion. The low ferrous iron
reported in analysis number 3 must be due either to
an analytical error or a nonrepresentative sample.
The identical ferrous/ferric ratios of the two rocks
eliminates the possibility of different associated fo/s,
since such a difference would be reflected in a large
difference between the two rocks in this ratio. The
possibility of differences in f|H2o is also eliminated
since, in the absence of significant amounts of other
gas species, H2O and O2 are inextricably related
through the dissociation of water equation (^O^*
H0 + 14O0). Different total confining pressures and/
or disequilibrium effects remained as possibilities.
TABLE 2.?Ferrous iron and total iron analyses of
1919 and 1954 basalts (weight percent)
Year and Iron Content
1919:
FeO
Total Fe as FeO
1954:
FeO
Total Fe as FeO
Fudali analysis
9.89
11.20
9.87
11.18
USGS analyses
9.50- 9.76
10.98-11.15
8.80- 9.59
11.18-11.21
* Determined from analyses of ground tilt.
Experimental Procedure
Representative samples of the two rocks were care-
fully crushed, under alcohol, to ?80 mesh in an
agate mortar, as for the chemical analyses. One to
two hundred milligram charges of the homogenized
powders were dryed at 110?C and placed in 60%
Ag-40% Pd crucibles which, in turn, were placed in
NUMBER 9 35
vycor (silica glass) tubes, the ends of which had been
previously sealed. The open ends of the vycor tubes
were connected to a vacuum pump and, after the
system was evacuated, short lengths of tubing (con-
taining the Ag-Pd crucibles) were necked off in the
flame of an oxy-hydrogen torch. The atmospheres in-
side the sealed vycor tubes were reduced to approxi-
mately l/2 mm Hg by this procedure. This prevented
the glass tubes from bursting at run temperatures.
More importantly, the amount of gas remaining in
the tube was so small, with respect to the weight of
the powdered charge, that this rock charge could
generate an equilibrium atmosphere at run tem-
perature without a measurable change in its own
composition.
The charges, so prepared, were placed in a stand-
ard, platinum-wound, vertical quench furnace. Tem-
perature control was maintained within d=l?C dur-
ing the course of the run. After sufficient time for
the charge to equilibrate (in all but a few cases) at
run temperature?usually 24 to 48 hours?the charges
were quenched by plunging them in a beaker of cold
water and the quenched phases examined under the
petrographic microscope.
Temperatures were measured immediately before
and after each run with a Pt-Pt/lORh thermocouple,
previously calibrated against the melting points of
pure gold and diopside. Corrections were empirically
determined for the difference in thermal profile of
the reading thermocouple and the more massive vycor
tube et al. Absolute accuracy of the indicated tem-
peratures, as measured by this method, may be no
better than ?2?G; however any error will be sys-
tematic rather than random. The indicated possible
temperature spreads shown in Table 3 do not reflect
such a systematic error. The indicated spread results
largely from the spread between the temperatures of
the two runs which most closely bracket the appear-
ance or disappearance of given phase. One should
therefore recognize that, although the absolute tem-
perature scales could be roughly 2? higher or lower
than those shown in Table 3, the relative temperature
differences shown will not change.
Both rocks were run in the same furnace, using
the same thermocouple, controller, and temperature-
sensing instrument. Length of runs, quenching, and
examination techniques were identical.
The paragenetic sequences of the two rocks, deter-
mined by these methods, are shown in Table 3.
Discussion of Experimental Results
The experimental results represent crystallization se-
quences under conditions of constant total composi-
tion. The ratio of ferrous to ferric iron in the charge
does not change with falling temperature and the
TABLE 3.?Crystallization sequences of 1919 and 1954
Kilauean basalts?constant composition conditions
1919:
1182?2?
1174?2?
1162?3?
1105 + 5?
1078?5?
1050?5?
1954:
1179?3?
1163 + 2?
1161?2?
1100 + 3?
1078?5?
C olivine-liquidus
plagioclase
pyroxene
Fe oxide
Solidus 1
Solidus 2
C olivine-liquidus
pyroxene
plagioclase
Fe oxide
Solidus
associated gas phase has a composition which is a
function of the charge composition, rather than vice-
versa. Down to about the appearance of an oxide
phase the crystallization sequence is also an equilib-
rium sequence. With glassy starting charges it is an
equilibrium sequence to the solidus. With crystalline
starting materials crystallization at lower tempera-
tures departs somewhat from equilibrium due to zon-
ing effects in the crystals and the consequent inability
of the largely crystalline charge to equilibrate in the
experimental runs because of the low rate of solid
state diffusion at these temperatures.
The 1954 lava sample used in this work was largely
glass (^80%) and the determined sequence is
therefore an equilibrium sequence. Both completely
crystalline and glassy charges of the 1919 lava were
used. The only difference between the two is that,
with the crystalline charge, nonequilibrium effects as
discussed above result in the persistence of a small
amount of residual liquid to '?30? lower (solidus 2)
than is the case with glassy charges (solidus 1). The
sequence determined with crystalline starting charges
represents the actual sequence that the rock ex-
perienced in nature.
The modal analyses of samples of the two rocks
equilibrated at 1160?C are entirely consistent with the
36 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
modal observations reported by Dr. Wright (Table
4), and the experimental crystallization sequences do
indeed show a significant difference with respect to
the first appearance of plagioclase. However, the lab-
oratory data does not support the interpretation based
upon the modal analyses of the natural rocks.
TABLE 4.?Modal analyses (volume percent)
Constituent
1919:
glass
pyroxene + olivine
plagioclase
1954:
glass
pyroxene + olivine
plagioclase
Experimental runs
equilibrated at
1160?C
*Original
rock modes
91.7
4.4
3.8
88.7
8.8
2.5
(Analysis #\,
Table 1)
89.2
5.1
5.7
(Analysis #4,
Table 1)
91.8
7.1
1.1
* Original rock modes supplied by Dr. T. L. Wright.
First, although plagioclase and pyroxene are pres-
ent in subequal amounts in 1919 (at 8% total crys-
tals) they did not begin to crystallize simultaneously.
Plagioclase has a crystallization interval of ^12?
before pyroxene becomes stable. Pyroxene then crys-
tallizes more rapidly over a short temperature interval
to approximately equal plagioclase in amount at
1160?C.
More importantly, since the observed differences in
crystallization histories have been duplicated experi-
mentally under low-pressure, equilibrium conditions,
it is clear that neither confining pressure differences
nor nonequilibrium effects need be invoked to explain
the modal differences observed in the natural rock
specimens. Clearly the true explanation is simply that
the very slightly higher AI2O3 content of the 1919
lava stabilized plagioclase at a temperature approxi-
mately 13? higher than in the 1954 lava. Although
the determined sequences are not precise enough to
warrant further conclusions, the modal analyses at
1160? suggest that the temperature interval between
the appearance of pyroxene and plagioclase in the
1954 lava is larger than the nominal 2? indicated in
Table 4. Here again, very minor differences in major
element content (most likely slightly lower MgO and
slightly higher GaO as compared with the 1919 lava)
probably favored pyroxene over olivine in the 1954
lava in comparison with the 1919 lava.
Conclusions and Implications
An obvious conclusion of this study is that paragenetic
sequences are best determined in the laboratory,
rather than by examination of natural, quenched
rock samples supposed to represent different (but
unknown) temperatures.
Equally obvious is that, while gross differences in
paragenetic sequences may reflect significant differ-
ences in the governing parameters, modest differences
in paragenesis (e.g., a 13?C change in the appearance
of plagioclase out of a total crystallization range of
<~'100oC) may be caused by very small, indeed
perhaps virtually undetectable, major element differ-
ences. Clearly, not a great deal of significance should
normally be attached to such modest paragenetic
differences.
The experimental data gives a reasonably precise
estimate of the temperature of the 1954 magma prior
to its expulsion from the 3 km deep magma chamber.
The 1954 lava specimens contain abundant pheno-
crysts of olivine and pyroxene plus rare phenocrysts of
plagioclase embedded in a matrix consisting of glass
and thin plates and needles of plagioclase and pyrox-
ene (obviously formed in the rapidly cooling lava
after extrusion). In the absence of water (less than
0.1% by wt.) the low confining pressure in the shal-
low magma chamber ('-'1000 bars) would not cause
a measurable change in the temperature of first ap-
pearance of the various mineral phases?with respect
to the temperatures determined in the laboratory.
Therefore, the temperature in the magma chamber
must have been lower than 1160? (for plagioclase to
be present) but no lower than 1155? (at which tem-
perature plagioclase is approaching the abundance
of pyroxene).
The situation is not as clear-cut with the 1919 lava
since the crystallization which occurred at the surface
in a slowly cooling lava lake is not readily distin-
guishable from that which occurred at depth. Only
olivine phenocrysts can be identified with certainty
as being a product of crystallization at depth. So the
possible temperature spread for the 1919 lava prior
to extrusion is 1155-1175?C. Interestingly enough,
NUMBER 9 37
the flank eruption that began on May 24, 1969, and
is still continuing in the east rift zone of Kilauea, has
produced lavas throughout the eruption "whose close-
range, optical pyrometer temperatures are normally
1160-1165 ?C,"* in excellent agreement with the
experimentally deduced temperatures of the 1919 and
1954 eruptions.
The similarity in the major element composition of
the 1919 and 1954 lavas, and their deduced tempera-
tures prior to eruption provide considerable informa-
tion about the nature and location of the magma
source, given the reasonable assumption that both
magmas come from the same source.
The possibility that a constant composition silicate
melt existed from 1919 thru 1954 in a shallow magma
chamber and was thus able to erupt isochemical lavas
35 years apart is not tenable. At 1160? the magma
contains significant amounts of olivine and pyroxene.
Given these circumstances it would be remarkable
that separation of crystals and liquid did not occur,
yielding a more recent magma or magmas of sub-
stantially different composition(s) from the 1919
lava. Melt viscosity and density differences between
crystals and melt result in effective laboratory sepa-
rations in a matter of days, with temperatures and
compositions similar to those indicated herein. Ac-
tually, volcanic activity during this period included
two additional summit eruptions (1934 and 1952)
and also periods of inflation and inferred injections
into the east rift zone (T. L. Wright, personal com-
munication) .Neither the 1934 nor the 1952 lavas are
chemically identical with the 1919 and 1954 lavas.
No east rift lavas were actually erupted during this
period but two separate flows erupted in the east rift
zone in 1955 appear to be clearly derivative rocks,
presumably formed by removal of pyroxene and pla-
gioclase from a summit basalt composition. Thus not
only should a shallow-seated magma differentiate,
but given the opportunity, some Kilauean magmas
have differentiated, to a recognizable extent, within
a period of time much shorter than 35 years.
Since all the Hawaiian lavas clearly originate by
partial melting in the upper mantle, a reasonable
explanation of the isochemical 1919 and 1954 lavas
is that they represent the primary melts of such a
* The Smithsonian Institution Center for Short-Lived
Phenomena, Event no. 14-17, card 1075, 12/22/70. Source:
Swanson and Peterson, U.S. Geological Survey, Hawaiian
Volcano Observatory.
process and were erupted before they had time to
differentiate.
The partial melting of mantle material most likely
occurs at, constant temperature, triggered by some
sort of pressure release mechanism. Since the lavas
carry phenocrysts rather than residual crystals they
must have been entirely liquid at one time, pre-
sumably when they originated. This necessitates an
absolute minimum source temperature of /~'1180?C
(liquidus temperature at 1 atm. pressure). A more
realistic source temperature*estimate is 1200? 1220?C
To keep basaltic material solid (prior to the melting
event) at such temperatures requires a depth of 60?
80 km for the source rock (Yoder and Tilley, 1962).
Such a depth is well within the earth's mantle and is
consistent with the frequent occurrence of ultramafic
xenoliths (considered by most geologists to be locally
derived mantle material) in some Hawaiian basalts.
As the magmas migrated upward from their mantle
source into the shallow chamber at 3 km depth they
cooled to ^1160?C, with some attendant crystalliza-
tion. In most cases some differentiation took place
but the 1919 and 1954 lavas were erupted before the
melt-crystal mixture could separate appreciably.
A remarkable aspect of this volcanic complex is
demonstrated by the fact that two lavas, separated in
time by 35 years, could move 60-80 km upward
through the mantle and crust, from the same source,
to issue at the same surface vent. Either the conduit
system from the mantle source to the surface has
remained continuously open because it is filled with
magma, as suggested by T. L. Wright (personal com-
munication) or the same conduit system has been
repeatedly opened, along a continual zone of weak-
ness, by a periodic pressure release mechanism. In
any case it seems indisputable that the conduits have
been well-established, stable features from the sur-
face to a depth of 60?80 km over at least the last
fifty years.
Acknowledgments
I thank Dr. T. L. Wright of the United States Geo-
logical Survey for supplying samples of the Hawaiian
basalts and making available to me their major ele-
ment analyses prior to publication. His enlightening
discussions with me and his penetrating criticisms of
this manuscript are also greatly appreciated.
38 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
Literature Cited 1956. Phase Equilibra at Liquidus Temperatures in the
System MgO-FeO-F2O3-SiO?. Journal of the
Fudali, R. F. American Ceramic Society, 39:121-140.
1965. Oxygen Fugacities of Basaltic and Andesitic Mag- Yoder, H. S., arid C. E. Tilley
mas. Geochimica et Cosmochimica Acta, 29: 1962. Origin of Basalt Magmas: an Experimental Study
1063-1075. of Natural and Synthetic Rock Systems. Journal
Muan, A., and E. F. Osborn of Petrology, 3(3):342-532.
Harold H. Banks, Jr. Iron-rich Saponite:
Additional Data on
Samples Dredged from
the Mid-Atlantic Ridge,
22? N Latitude
Banks and Melson (1966) reported the occurrence of
saponite in hydrothermally altered basalts and ba-
saltic tuffs from the crest of the Mid-Atlantic Ridge
and in the median valley between 22? and 23? N
latitude. The present paper gives the results of a
chemical and X-ray study of that material as well as
of identical material tentatively identified as non-
tronite by Melson and van Andel (1966).
The 22? N latitude area was first examined in
detail during cruise 44 of R. V. Chain (September-
October 1964) of the Woods Hole Oceanographic
Institution and revisited in 1965 on cruise 1965-1 of
R. V. Thomas Washington (December 1965) of
Harold H. Banks, Jr., Department of Mineral Sciences, Na-
tional Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560.
Scripps Institution of Oceanography. The geo-
logical and geophysical results were presented by van
Andel et al. (1965) and van Andel and Bowin
(1968). The petrology of the area has been discussed
by Melson and van Andel (1966) and Melson et al.
(1968). Siever and Kastner (1967) have described
the mineralogy of the sediments.
Saponite is a major constituent of massive basalts,
metabasalts, basaltic tuffs, and metatuffs dredged at
four stations in this area of the Mid-Atlantic Ridge
(Table 1).
In massive basalts, olivine phenocrysts are entirely
replaced, in some, by dark, gray-green saponite with
iron oxide, possibly goethite, localized at the crystal-
matrix interface and in fractures in the pseudo-
morphs. The pseudomorphs are commonly connected
TABLE 1.?Location of dredge hauls on the Mid-Atlantic Ridge in the area between 22? and
23? N latitude containing abundant saponite-bearing rocks
Station
Thomas Washington, cruise 1965-1:
THV-4 (dredge 2)
THV-14 (dredge 17)
THV-18 (dredge 10)
Chain, cruise 44:
DR-2
Latitude and longitude
22?31.5'N, 45?W to 44?59.8'W
22?49.2'N, 45?8.8'W to
22?48.6'N, 45?10.1'W
22?10'N, 45?14.2'W to
22?11.4'N, 45?16.6'W
22?38.2'N, 44?58.6'W to 44?58'W
Depth (m)
1985-1735
3500-3000
3100-2600
2400-2100
39
40 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
by veinlets of optically identical material. Saponite
also occurs as thin films along cleavages and fractures
in subcalcic augite grains and along grain boundaries.
Plagioclase (ca. An65) often exhibits peripheral al-
teration to saponite, and any glass inclusions in the
plagioclase are entirely altered to saponite. Saponite
is also present as vesicle, interstitial, and vein filling.
Such relations may be seen in a single thin section.
Saponite occurs associated with albite (Ani?An4),
relict calcic plagioclase (as high as An75), ripidolitic
chlorite (MgO = 20%, FeO=19%, A12O3=18%),
actinolite (47-64% tremolite molecule), epidote (av-
erage composition about 75% clinozoisite molecule),
quartz, sphene, pyrite, and relict spinel in the meta-
basalts (Melson and van Andel 1966). It is com-
monly intergrown in vesicles and veins with material
which optically resembles chlorite. Veinlets and
shear planes in mylonitized and brecciated basalts
which have been only slightly altered in mineral
composition contain trace amounts of saponite asso-
ciated with albite, actinolite, chlorite, epidote,
pumpellyite, and zeolites (probably mainly natrolite).
Marine tuffs dredged from the 22? N area show
considerable variation in texture and mineralogy, but
most show an abundance of saponite in the matrix,
absence of highly vesiculated fragments, and abun-
dance of fragments of glassy, variolitic rinds as op-
posed to coarser-grained fragments.
The metatuffs are undeformed and consist of
lapilli of originally glass-rich basalt in a fine-grained
ashy matrix. The glass is completely replaced by ripid-
olitic cholorite which is commonly altered to saponite.
The matrix is composed of abundant saponite and
traces of actinolite.
Optically, the saponite is micaceous to fibrous in
habit, gray-green to yellow-brown in color, exhibits
moderate to high birefringence (ca. 0.02 ;NZ= 1.56)
and is weakly pleochroic. In hand-specimen, it is light
to dark gray-green with a greasy luster. Massive pieces
are extremely soft when wet but become harder and
brittle upon drying.
The saponite was identified using the identification
scheme of Warshaw and Roy (1961) for layer sili-
cates. Samples were ground in a porcelain mortar,
and care was taken to prevent or at least minimize
the effects of fine grinding which can destroy the
ordered crystalline structure. The powders were then
mixed with a small amount of finely ground silicon
as an internal standard for X-ray study. Oriented
slides, glycolated and unglycolated (with ethylene
glycol) were analyzed on a diffractometer at 1??
20/minute, with Cu radiation and a nickel filter.
Resolution of the (060) smectite* reflection (1.49?
1.52 A for dioctahedral smectite and 1.53-1.54 A for
trioctahedral smectite) was obtained by powder pho-
tographs using nickel filtered copper radiation in a
114.6 mm diameter camera. Powder data for two
massive samples of saponite (samples 2?17 and 10?
116) are given in Table 2.
TABLE 2.?X-ray powder data for saponite
dA
Indices1 Cahuenga Mid-Atlantic Ridge Mid-Atlantic Ridge
Pass, Station THV-4 Station THV-18
California1 (Sample 2-17) (Sample 10-116)
001
002
003
11 ,02
004
005
13,20
006
?
?
04,22
15,24
31
06,33
26,40
19,46
53
60,39
15.4
7.9
5.28
4.60
3.93
3.13
2.648
?
2.560
2.452
2.298
1.740
?
1.541
1.330
1.001
?
0.888
12.5382
4.605
2.640
2.520
2.311
1.739
1.536
(10)8
(4)
(3)
(1)
(1)
(4)
12.5032
?
?
4.563
?
?
2.618
?
?
?
?
?
?
1.527
?
?
?
?
(10)8
(4)
(3B)
(5)
1Data for griffithite, an iron-rich saponite (Faust 1955).
2 X-ray pattern was made using the following instrumental
factors: GuKa radiation, \= 1.54178A; 38KV; 18MA;
silicon served as an internal standard. Peak shifted to 16.61 A
after treatment with ethylene glycol.
3 Numbers in parentheses are estimated relative intensities.
B-broad.
* I use the term "smectite," proposed by British clay
mineralogists (Brown 1955), rather than "montmorillonite-
type mineral," an approved, identical, but cumbersome
name.
NUMBER 9 41
Semiquantitative chemical analyses of saponite
(Table 3) from veins in a mylonitized olivine basalt
(sample 2?17) and from hydro thermally altered ba-
saltic glass (sample 10?116) were performed by elec-
tron microprobe and X-ray fluorescence analysis,
respectively, of lithium tetraborate fusions, with ana-
lyzed oceanic basalt lithium tetraborate fusions serv-
ing as standards: the fusions were prepared following
the procedure outlined by Rose et al. (1963). For
sample 10-116, H2O was determined by the Penfield
method, Na2O and K2O by flame photometry: the
FeO and Fe2C?3 contents of both samples were de-
termined by classical methods.
The chemical analyses of the samples were plotted
TABLE 3.?Semiquantitative chemical analyses of
iron-rich saponite, Thomas Washington 1965?1, Sta-
tion THV-4, Sample 2-17*, and Station THV-18,
Sample 10-116f
Constituent
SiOa
AUO,
Fe2Os
FeO
MgO
CaO
Na2O
K2O
TiO2
H2O (total)
Chemical analyses (weight percent)
2-1T 10-1162
43.0
7.2
7.79
2.53
21.0
0.4
3.3
0.4
<0.05
18.9
41.6
5.0
11.32
0.78
18.3
0.6
2.2
?0.5
0.3
20.6
* Saponite from veins in mylonitized olivine basalt.
f Saponite derived by the hydrothermal alteration of
basaltic glass.
1 Performed by electron microprobe analysis of lithium
tetraborate fusion using an enlarged electron microprobe
spot and classically analyzed oceanic basalt lithium tetra-
borate fusions as standards. FeO and Fe2O3 contents were
determined by classical methods.
2 Performed by X-ray fluorescence, using the method of
Rose et al. (1963). H2O determined by Penfield method;
NassO and K2O by flame photometry; FeO and Fe2O3 by
classical methods.
on a FeO-Fe2O3-MgO ternary diagram (Figure 1)
which graphically distinguishes between the two smec-
tites in which iron is an essential and major con-
stituent: nontronite (dioctahedral; FeO = O-l%)
and iron-rich saponite (trioctahedral; FeO=l?8%).
Chemical analyses to delineate the smectite fields
shown in Figure 1 were selected from various sources:
saponites are from Ross and Hendricks (1945, anal-
yses 77-79) and Deer et al. (1962, analysis 13);
nontronites are from Ross and Hendricks (1945,
analyses 55?67) ; iron-rich saponites are from Cal-
liere and Henin (1951, analyses 4-5), Faust (1955),
and Sudo (1959).
The combined X-ray and chemical results indicate
that the clay is iron-rich saponite. The (001) reflec-
tion for the clay shifts from nearly 12.5A to 16.6A
upon treatment with ethylene glycol, and the critical
(060) reflection is greater than 1.52A which is in-
dicative of trioctahedral smectite (see Table 3). In
addition, the semiquantitative chemical analyses of
the clay plot very nicely in the iron-rich saponite
region of the FeO-Fe2O3-MgO diagram (Figure 1).
I suspect that the high Fe2O3 content of sample 10?
116 (Table 3) is the direct result of the participation
of heated sea water with a high Eh in the formation
of the clay. The high Na2O content and the low GaO
content of both samples may very well be a primary
feature, but Na saturation may also result from ca-
tion exchange during weathering of a normal, Ca-
rich saponite in contact with sea water.
Textural relations of saponite in rocks from the
22?N area suggest several modes of origin for the
saponite: (1) hydrothermal alteration, by heated sea
water, of basaltic glass and silicates in the matrix of
basaltic tuffs during cooling and movement of the
tuffs; (2) low-grade metamorphism of basalts and
basaltic tuffs at a maximum temperature of 300 ?G
(Melson and van Andel, 1966) and a pH2O of
about 500 bars; and (3) deuteric alteration of ba-
saltic glass, olivine, plagioclase and augite in the
massive basalts.
I am indebted to Eugene Jarosewich, who provided
the semiquantitative chemical data for sample 10?116
and the total and ferrous iron analyses for sample 2?
17, and to Dr. Joel Arena who kindly assisted in the
preparation of samples for X-ray analysis. Special
thanks are due to Kenneth Towe and William G.
Melson for their helpful discussions of many of the
problems encountered during the course of this study.
42 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FeO FeO
IRON-RICH SAPONITE
NONTRONITE
o ^a-
MgO Fe2O3
FIGURE 1.?Semiquantitative chemical analyses of saponite from the Mid-Atlantic Ridge (sam-
ples 2-17 and 10-116, open circles with solid centers) plotted on the ternary diagram FeO-
Fe2O?-MgO. Open circles are nontronites selected from Ross and Hendricks (1945); open
squares are saponites selected from Ross and Hendricks (1945) and Deer et al. (1962); solid
triangles are iron-rich saponites selected from Calliere and Henin (1951), Faust (1955), and
Sudo (1959).
Literature Cited
Banks, H. H., Jr., and W. G. Melson
1966. Saponite from the Mid-Atlantic Ridge, 22? N
Latitude. 1966 Annual Meeting Geological Soci-
ety of America, Abstracts: 9?10.
Brown, G.
1955. Report of the Clay Minerals Group Subcommittee
on Nomenclature of Clay Minerals. Clay Min-
erals Bulletin, 2:294-302.
Caillere, S., and S. Henin
1951. The Properties and Identification of Saponite
(Bowlingite). Clay Minerals Bulletin, 1:138-
144.
Deer, W. A., R. A. Howie, and J. Zussman
1962. Rock-Forming Minerals, 3:226-245. New York:
John Wiley & Sons.
Faust, G. T.
1955. Thermal Analysis and X-Ray Studies of Griffith-
ite. Washington Academy of Science Journal,
45:66-70.
Melson, W. G., and Tj. H. van Andel
1966. Metamorphism in the Mid-Atlantic Ridge, 22? N
Latitude. Marine Geology, 4:165-186.
Melson W. G., G. Thompson, and Tj. H. van Andel
1968. Volcanism and Metamorphism in the Mid-Atlan-
tic Ridge, 22? N Latitude. Journal of Geophysi-
cal Research, 73:5925-5941.
Rose, H. J., Jr., I. Adler, and F. Flanagan
1963. X-Ray Fluorescence Analysis of the Light Ele-
ments in Rocks and Minerals. Applied Spec-
troscopy, 17:81-85.
Ross, C. S., and S. B. Hendricks
1945. Minerals of the Montmorillonite Group. U.S.
Geological Survey Professional Paper, 205B:23-
77.
Siever, R., and M. Kastner
1967. Mineralogy and Petrology of some Mid-Atlantic
Ridge Sediments. Journal of Marine Research,
25:263-278.
Sudo, Toshio.
1959. Mineralogical Studies on Clays of Japan. To-
kyo: Maruzen. Pages 227-268.
van Andel, Tj. H., and C. O. Bowin
1968. The Mid-Atlantic Ridge Between 22? and 23?
North Latitude and the Tectonics of Mid-Ocean
Rises. Journal of Geophysical Research, 73:
1279-1298.
van Andel, Tj. H., V. T. Bowen, P. L. Sachs, and R. Siever
1965. Morphology and Sediments of a Portion of the
Mid-Atlantic Ridge. Science, 148:1214-1216.
Warshaw, C. M., and R. Roy
1961. Classification and a Scheme for the Identification
of Layer Silicates. Geological Society of Amer-
ica Bulletin, 72:1455-1492.
George Switzer,
William G. Melson,
and Geoffrey Thompson
Olivine Crystals
from the Floor of
the Mid-Atlantic
Ridge near
22? N Latitude
Remarkably fresh, small, gemmy, single crystals of
olivine were found in the heavy mineral fraction of
foraminiferal sand dredged from the eastern flank of
the mid-Atlantic Ridge near 22?N. This occurrence
of euhedral, free olivine crystals in a marine sediment
is highly unusual in our experience. These olivine
crystals clearly grew in basaltic lava, and were then
freed by growth of vesicles, with gas exsolution com-
monly initiating at crystal-liquid boundaries. Ultimate
freeing of the crystals from the lava may entail more
than this. Conditions in a vesiculating, turbulent,
submarine lava fountain may be particularly favor-
able to the final freeing of these crystals.
Olivine phenocrysts, along with plagioclase pheno-
crysts, are the most common phenocrysts in deep-sea
basalts from the mid-ocean ridges. Thus, the occur-
rence of olivine in the central, volcanically active zone
of the mid-Atlantic Ridge is not surprising, but its
occurrence free of a basaltic matrix is.
The dredge which recovered the olivine crystals is
from the eastern slope of the median valley of the
mid-Atlantic, a zone of active sea-floor spreading,
characterized by frequent earthquakes and volcanic
eruptions. The rocks of the 22?N area are dominantly
fresh basalts and greenschist facies metabasalts (Mel-
George Switzer and William G. Melson, Department of
Mineral Sciences, National Museum of Natural History,
Smithsonian Institution, Washington, D.C. 20560. Geoffrey
Thompson, Woods Hole Oceanographic Institution, Woods
Hole, Massachusetts 02543.
son and van Andel 1966, Melson et al. 1968). The
median valley here is essentially a rifted submarine
basalt field, where volcanism is manifested mainly
by fissure eruptions, some on the flanks, and others
along the floor of the median valley. The ridge from
which the olivine was obtained is a major bathymetric
feature, rising to within 1400 meters of the sea sur-
face. This same ridge elsewhere produced pillow
lavas, massive basalts, and massive to schistose green-
schist facies metabasalts. The bathymetry and geology
of this portion of the mid-Atlantic Ridge is also
described by van Andel and Bowin (1968). Details
of the recovery are as follows:
Cruise: R. V. Chain, cruise 44, of the Woods Hole
Oceanographic Institution.
Location: Dredge tract from 22?50.0'N, 45?05.6'W
at 3200 meters, to 22?50.6'N, to 45?04.2'W at 3400
meters.
Description: 450 grams of coarse foraminiferal
sand from a 4-inch pipe dredge, taken in dredge 8.
At Woods Hole, a portion of the sample was
leached with dilute hydrochloric acid and subjected
to grain size analysis and to density separations. The
grain size distribution was as follows: 27.35 gms
greater than 590 microns; 0.54 gms greater than 297
microns, but less than 590 microns; and 43.32 gms
less than 297 microns. Remarkably pure concentrates
of euhedral olivine crystals were produced in the size
fraction between 297 and 590 microns in the density
range of 2.9 to 3.3 gm/cc. These crystals compose
less than 1 percent of the total untreated sample.
43
44 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
The crystals occur with fresh fragments of basaltic
glass (n= 1.596). Olivine commonly occurs in this
glass as phenocrysts, which have in some places been
partially separated from their matrix by growth of
gas pockets (vesicles) along the crystal-lava contact.
It is thus quite clear that the free olivine crystals
originally crystallized from basaltic lava, but were
freed by growth of gas cavities around their surfaces,
evidently a favored site for gas exsolution during
vesiculation.
The olivine crystals are transparent, pale yellowish
green and range from about one-half to one milli-
FIGURE 1.?A-B, Complete individuals showing typical olivine morphology. XlOO. C-D, TWO
views of a crystal showing pitting and skeletal growth, (c X95, D X140). Scanning electron
microscope photographs by Walter Brown, Smithsonian Institution.
NUMBER 9 45
FIGURE 2.?A, Crystal showing pitting and skeletal growth, B, Crystal with attached glass
fragments X95. c, Crystal in cavity in vesiculated glass, X40. D, Spinel crystal in glass
matrix, X140. Scanning electron microscope photographs by Walter Brown, Smithsonian
Institution.
meter in maximum length. Their edges are remark-
ably sharp, and they have highly reflective faces.
There is no evidence whatsoever of abrasion. These
features, and the interesting skeletal forms shown by
some crystals are illustrated in the accompanying
scanning electron microscope photographs. The com-
position from microprobe analysis is Fo82. We did
not determine nickel and other minor and trace ele-
ment contents. Small octahedra of a chromian spinel
are associated with the olivine.
The olivine crystals are similar in composition to
olivine phenocrysts (Fo85) described from pillow
lavas in this region (Melson et al. 1968, page 5929).
Acknowledgments
We wish to acknowledge the assistance of Dr. V. T.
Bowen, Chief Scientist, and the officers and crew of
46 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
the R. V. Chain, for their assistance in obtaining
these samples. Dr. Tj. H. van Andel has contributed
substantially, both in published accounts and in dis-
cussions, to our information on the 22?N area. Sup-
port for various aspects of this study aboard ship and
at Woods Hole came from the United States Atomic
Energy Commission, the Office of Naval Research,
and the National Science Foundation. The work at
the Smithsonian was supported partially by the
Smithsonian Research Foundation. This paper will
be contribution No. 2665 from the Woods Hole
Oceanographic Institution.
Literature Cited
Melson, W. G., and Tj. H. van Andel
1966. Metamorphism in the Mid-Atlantic Ridge, 22? N
Latitude. Marine Geology, 4:165-186.
Melson, W. G., G. Thompson, and Tj. H. van Andel
1968. Volcanism and Metamorphism in the Mid-Atlan-
tic Ridge, 22? N. Latitude. Journal of Geophys-
ical Research, 73:5925-5941.
van Andel, Tj. H., and C. O. Bowin
1968. The Mid-Atlantic Ridge Between 22? and 23?
North Latitude and the Tectonics of Mid-Ocean
Rises. Journal of Geophysical Research, 73:
1279-1298.
George Switzer Origin and
and William a Mdson Composition of
Rock Fulgurite
Glass
A great deal of attention has been given recently to
natural glasses, especially tektites, impact glasses, and
the glass phases of lunar rocks. However, fulgurites,
natural glasses formed as a result of fusion of rock
by lightning, have not been the subject of modern
study.
Fulgurites may be conveniently divided into two
types, those formed by fusion of loose sand, or sand
fulgurites, and those formed by fusion of solid rock,
usually referred to as rock fulgurites.
Sand fulgurites are normally of simple composi-
tion, from about 90 to 99 percent SiC>2 depending
on the purity of the sand. They are quite common
and have been described in detail by numerous in-
vestigators beginning^ the year 1711.
Rock fulgurites on the other hand appear to be
relatively uncommon, or at least have been observed
less frequently than sand fulgurites. Also, they are
petrologically more interesting since they may result
from fusion of any rock type. They have been de-
scribed as formed from rocks ranging as widely as
andesite, granite, hornfels, glaucophane schist, and
serpentinite.
Investigation of rock fulgurite glass composition
by microprobe has revealed that these glasses are
extremely hetergeneous. As examples, we have looked
at the composition of glasses formed by fusion of
George Switzer and William G. Melson, Department of
Mineral Sciences, National Museum of Natural History,
Smithsonian Institution, Washington, D.C. 20560.
1 Presented orally at the International Mineralogical As-
sociation Seventh General Meeting, Tokyo and Kyoto,
Japan, August 1970.
andesite from Little Ararat, Turkey; quartz diorite
porphyry from Crested Butte, Colorado; and horn-
fels from Castle Peak, Colorado.
Little Ararat
Little Ararat, in eastern Turkey, elevation 12,877
feet, has the most extensive development of fulgurites
reported anywhere in the world. The fulgurites are
in hypersthene diopside andesite.
Detailed chemical analyses were made of areas in
two thin sections of specimen USNM 52094. Figure
1A shows a partially melted plagioclase phenocryst.
A microprobe step scan, at five micron intervals, was
made along the line shown, and the results of the
analysis as given in Figure 1B. The plagioclase-glass
interface shows clearly in the photograph but is not
so clearly discernible from the chemistry. For some
distance beyond the interface the glass has plagioclase
composition. It then changes abruptly to a composi-
tion representing melted groundmass.
Figure 2A shows a partially melted diopside pheno-
cryst. Two interfaces are clearly discernible, one be-
tween diopside and diopside glass, and the other be-
tween diopside glass and glass derived by melting of
groundmass. The same boundaries are reflected in
the chemistry, as shown in Figure 2B.
Crested Butte, Colorado
Crested Butte, elevation 12,172 feet, is a Tertiary
quartz diorite porphyry laccolith lying about 3 miles
northeast of the town of Crested Butte in west
central Colorado. The specimens used in this study
47
48 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
LITTLE ARARAT,TURKEY
- *? -Diopside- -?~-Diopside
Glass
LITTLE ARARAT, TURKEY
150 200 250 300
MICRONS
%NajO
0 50 100 150 200 250 300 350 400
Q Microns
FIGURE 1.?a, (Photo) Partially melted plagioclase pheno-
cryst. Andesite, Little Ararat, Turkey, b, (Graph) Com-
position of glass along AB, Figure la.
FIGURE 2.?a, (Photo) Partially melted diopside phenocryst.
Andesite, Little Ararat, Turkey, b, (Graph) Composition
of glass along AB, Figure 2c.
(USNM 52981) were collected by Whitman Cross in
1885.
The extreme hetergeneity of rock fulgurite glass,
resulting from essentially no mixing while in the
NUMBER 9 49
molten state, is especially well exemplified by this
material. In thin section, shown in Figure 3A, the
plagioclase-glass interface is clearly seen, and hetero-
geneity of the glass is indicated by swirl marks, not
visible in the photograph.
A step scan was made along the line shown and
the results are shown in Figure 3B. There has been
so little mixing that areas of glass representing the
original phase can be recognized. Moving from left
to right the composition is first that of plagioclase,
then plagioclase glass. It then changes to quartz
glass (as shown by SiC>2 nearly 100 percent, all other
elements near zero), back to plagioclase glass, then
quartz glass, plagioclase glass once again, and finally
a glass high in Fe and Mg, probably representing
fused biotite. The areas cannot be detected optically,
but their dimensions correspond closely to the ground-
mass grain size of the rock.
FIGURE 3.?a, (Photo) Fusion crust on quartz
diorite, Crested Butte, Colorado, b, (Graph)
Composition of glass along AB, Figure 3a.
i ?? ',y
CRESTED BUTTE.COLORADO
%SiO.
300 350 400 450
M I C R 0 N S
600 650 700
50 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
Castle Peak, Colorado
Castle Peak, elevation 14,265 feet, is in west central
Colorado and is the highest point in the Elk Moun-
tains. The specimens studied (USNM 52980) were
collected by Whitman Cross in 1887. The rock is
an albite-quartz hornfels, with minor epidote, andra-
dite, hornblende, calcite, and pyrite.
A photomicrograph of the glass analyzed, a bleb
attached to the wall of a tube, is shown in Figure 4A.
Because of the fine grain size, the glass appears opti-
cally homogeneous, and isochemical melting of in-
dividual grains cannot be seen. The results of a step
scan analysis along the line shown are given in Fig-
ure 4B. This analysis reflects the fine grain size of
the rock, but shows that even on this scale there
has been relatively little homogenization.
Discussion
A remarkable feature of some of the rock fulgurites
studied is the physical effect of the lightning strike.
For example, the hornfels from Castle Peak has a
nearly straight, glass-lined tube, 4 mm in diameter,
and extending entirely through 9 cm of rock. The
"entrance" of the tube is spalled and glass spattered,
and the "exit" has a raised glass lip.
m mm
FIGURE 4.?a, (Photo) Fusion crust on horn-
fels, Castle Peak, Colorado, b, (Graph) Com-
position of glass along AB, Figure 4a.
CASTLE PEAK, COLORADO
50 100 150 200 250 300 350 400 450 500 550
Microns
NUMBER 9 51
Specimens from other localities, such as Little
Ararat, show both broad surfaces covered by a thin
fusion crust, as well as irregular, sometimes branching,
glass-lined tubes.
The rock fulgurite glasses examined contain no
crystallites. This feature, combined with the observed
isochemical melting of individual mineral grains and
the extreme heterogeneity of the glasses, indicates
rapid melting and cooling. These features are pre-
dictable since the duration of a lightning strike is
of the order of one millisecond.
Measurements of lightning made by many investi-
gators have shown that:
1. The potential difference between cloud base and
ground is of the order of 100,000,000 volts.
2. Peak current may be as high as 10,000 amperes,
and during most of the duration of the strike is of the
order of 1,000 amperes.
3. Air particles within a radius of 2 cm of the
channel are ionized at a temperature of about
30,000?K.
The tubes, such as the one described in hornfels,
if formed by lightning, must also have been formed
in a time interval of the order of millisecond. This
is a phenomenon difficult for us to comprehend even
knowing the vast amount of energy released by a
lightning stroke. Other explanations for the forma-
tion of these remarkable tubes must be considered,
particularly for the unusually straight tube in the
sample from Castle Peak, although as yet none more
plausible has occurred to us.
The extremely high temperatures conceivably
reached by fulgurites might lead to volatilization
of metallic oxides, particularly silica and those of
the alkalies. All the sand and rock fulgurites we
examined contain gas cavities, but these appear to
have formed mainly by boiling off of surface and
combined water; we have detected no loss of metallic
oxides.
METEORITES
Joseph Nelen
and Brian Mason
The Estherville
Meteorite
The Estherville meteorite fell near the town of the
same name in eastern Iowa on May 10, 1879. It be-
longs to the small class of mesosiderites, and is one
of only six that were observed falls. The material is
particularly suitable for research since it was collected
immediately after the fall and has not suffered from
the weathering that has seriously deteriorated most
other mesosiderites. No complete chemical analysis
has been published of the silicate material of this
meteorite, hence the present investigation.
Material was taken from specimen number 1722
in the meteorite collection of the United States Na-
tional Museum of Natural History. The material was
crushed to pass through a 50-mesh sieve, and nickel-
iron particles were removed with a hand magnet.
This procedure evidently failed to remove all the
included metal, since 0.14% Ni was determined in
the analysed sample; using the figure of 8.57% Ni in
the metal phase of Estherville (Powell 1970), this
corresponds to 1.6% of nickel-iron. The analysed
sample also showed 1.05% S, equivalent to 2.88%
troilite. From the total iron determined by analysis
the appropriate amounts were allotted to nickel-iron
and troilite, and the remainder is reported as FeO.
The analysis is set out in Table 1, together with par-
tial analyses of Estherville silicates by Powell (1970)
and Jerome (1970). Powell's data were obtained by
X-ray fluorescence analysis; Jerome's data by semi-
micro wet chemical analysis except for Cr2O3, ob-
tained from instrumental neutron activation anlysis.
The normative mineralogical composition calcu-
lated from the analysis is (in weight percent) : pyrox-
ene (En71Fs2-)Wo4), 72.6; plagioclase (An89), 18.4;
troilite, 2.9; nickel-iron, 1.6; tridymite, 1.6; chromite,
1.1; whitlockite, 0.7; ilmenite, 0.5. This is consistent
Joseph Nelen and Brian Mason, Department of Mineral
Sciences, National Museum of Natural History, Smithsonian
Institution, Washington, D.C. 20560.
with the observed mineralogical composition, except
that a little olivine (estimated 2%) is present; the
amount of modal pyroxene is therefore less than
normative pyroxene by about 2%, and the amount
of modal tridymite will be somewhat greater than the
1.6% in the norm. Modal chromite and ilmenite will
be somewhat less than the normative amount, be-
cause some chromium and titanium are combined in
the pyroxenes. Fuchs (1967) has shown that the
whitlockite in Estherville is accompanied by stan-
fieldite, Ca4Mg3Fe2(P04)o.
Estherville is a polymict breccia, and different sam-
ples of the silicate material are likely to give somewhat
different analyses. This certainly explains, at least
TABLE 1.?Analyses of silicate material of the
Estherville meteorite
Constituent
SiO2
TiO2
Al2Os
Cr2O3
FeO
MnO
MgO
CaO
Na2O
K2O
p2o5
FeS
Fe
Ni
Total
FeO/FeO + MgO
(mole percent)
Chemical analyses
1*
50.03
0.29
6.24
0.76
12.27
0.51
18.88
5.18
0.22
0.02
0.31
2.88
1.46
0.14
99.19
26.8
(weight percent)
2*
55.42
0.43
8.44
11.80
0.47
14.16
5.62
0.74
31.9
3*
0.29
7.10
0.95
18.68
0.43
17.40
5.50
0.26
0.02
37.6
* Column 1: Analyst, Joseph Nelen.
Column 2: Powell (1970).
Column 3: Jerome (1970).
55
56 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
in part, the discrepancies between the different anal-
yses in Table 1. The FeO/FeO + MgO molecular
percent given by Jerome's analysis is too high to be
consistent with the average composition of the pyrox-
enes; probably his sample contained some troilite
and nickel-iron whose iron was determined and re-
ported as FeO. The SiO2 figure in Powell's analysis
appears to be somewhat too high; he gives a figure
of 10.1% for normative tridymite, which is con-
siderably higher than the amounts we have observed.
Literature Cited
Fuchs, L. H.
1967. Stanfieldite, a New Phosphate Mineral from
Stony-Iron Meteorites. Science, 158:910-911.
Jerome, D. Y.
1970. Composition and Origin of Some Achondritic
Meteorites. Thesis, University of Oregon.
Powell, B. N.
1970. Petrology and Chemistry of Mesosiderites. I. Tex-
tures and Composition of Nickel-Iron. Geochim-
ica et Cosmochimica Ada, 33:789-810.
A. F. Noonan, Forest Vale
Kurt A. Fredriksson,
and Joseph Nelen
ABSTRACT
The Forest Vale meteorite is a typical olivine-bronzite
chondrite. Chemical, mineralogical, and textural evi-
dence indicates that all phases are genetically related,
crystallized at high temperature, and the later formed
agglomerate was lithified with no subsequent thermal
metamorphism.
Introduction
At about 3 p.m. on August 7, 1942, a shower of
meteorites fell near Forest Vale, New South Wales,
Australia. Forest Vale is about 12 miles (19.2 km)
northeast of Tullibigeal (Latitude 33?30'S; Longi-
tude 146?48'E), which is on the Wyalong-Cargellico
railway line. The countryside has low relief with hills
of Devonian sedimentary rocks, granite, and Tertiary
lava flows.
Details are reported by Hodge-Smith (1942, 1943).
Five fragments were recovered which fitted together
to form an almost complete stone. Of these, three
have been preserved. The two smaller fragments have
been destroyed or lost since their discovery. It is es-
timated that the two lost specimens weighed nearly
4.5 kg, and the total weight of the stone was about
26 kg. In addition, another stone was discovered in
the immediate vicinity and its probable weight was
about 2 kg. Hodge-Smith (1943) wrote:
Both stones are chondritic. On the fractured surface the color
is bluish-grey and the texture is fine granular somewhat re-
sembling a grey sandstone in appearance. The chondrules
A. F. Noonan, Kurt A. Fredriksson and Joseph Nelen, De-
partment of Mineral Sciences, National Museum of Natural
History, Smithsonian Institution, Washington, D.C. 20560.
vary in size from 1 mm. to 0.1 mm. in diameter. They are
very numerous, and many consist of olivine or enstatite, or
both. The groundmass consists mainly of the same two min-
erals. Nickel-iron is present in a very finely divided state,
fairly evenly distributed throughout the mass. It constitutes
roughly 20% by weight of the stone. Chemical analysis and
further optical studies have yet to be undertaken.
Mason (1963) determined the composition of the
olivine in the Forest Vale meteorite. The purpose of
this study is to expand the mineralogic and petrologic
data and, if possible, obtain some indications about
the conditions of formation of this meteorite and of
chondrites in general.
Chemical Composition
In addition to the wet chemical analysis by Jarose-
wich (1967), detailed phase analyses were made with
an ARL (Applied Research Laboratories) electron
probe microanalyzer. The instrument was operated
at 20 KV with a specimen current of about 0.03
microamps. All measurements were corrected for
deadtime, drift, background, and mass absorption.
In addition, secondary fluorescence corrections were
made for the metal phases. For details on the tech-
nique see Keil (1967). For the quantitative analyses,
pure metallic iron and nickel, Marjalahti olivine
(Fe = 8.7, Mg = 29.0 weight percent), and Johnstown
hypersthene (Ca = 0.99, Mg= 16.40 weight percent)
were used as standards.
WET CHEMICAL ANALYSIS
The chemical analysis in Table 1 was published by
Jarosewich (1967). However, the reported FeO
value (5.84%) is too low to be consistent with the
composition of the olivine and pyroxene as deter-
mined from microprobe analyses. This is probably due
57
58 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
to analytical error, since FeO was determined as the
difference between total iron and iron in metal and
sulfide. The analysis has been corrected by calculating
the amount of FeO (8.75%) required to make Fai7
from the acid-soluble MgO (23.95 weight percent).
Combined with the acid insoluble FeO, the total
FeO would then be 8.61%. This will in turn necessi-
tate lowering the metallic iron to 17.24%, which is
much closer to the average of two recent analsyes for
metallic iron?17.57%?obtained by Jarosewich on
two other small pieces from this meteorite.
In Table 2 the data are also presented in the con-
ventional form expressed as metal, troilite, and oxides;
in terms of the elements in weight percentages; and
in atom percentages after the elimination of O, C, H,
and S. The presentation in conventional form (col-
umn A) assumes that all S exists as FeS, that Fe in
excess of free metal and FeS is present in the ferrous
TABLE 1.?Analysis of the Forest Vale meteorite (Ja-
rosewich 1967, with revised figures for Fe and FeO)
Constituent
Fe
Ni ..
Go
FeS
SiO2
TiO2
AUO,
Cr2O8 .
FeO
MnO
MgO
CaO
Na2O
KaO .
P2O8
H2O ( + )
H,O (-)
C
Total
Total Fe
Chemical analyses (weight percent)
1
Acid Soluble
56.59
24.00
0.03
0.14
8.75
23.95
0.58
43.48
2
Acid Insoluble
43.41
55.01
0.27
4.38
8.44
22.31
3.15
6.56
3
17.24
1.76
0.09
5.54
37.46
0.13
1.98
0.46
8.61
0.31
23.23
1.69
0.85
0.11
0.31
0.72
0.06
0.18
100.73.
27.45
state, and that all Ni and Co are present in the metal
phase. The second method of presentation (column
B) more closely reflects the results actually obtained
in analysis. The amounts of the different elements
are determined with the exception of O, which is
later added to make 100 percent. The expression of
the analysis in atom percentages after the elimination
of O, C, H, and S, was introduced by Wiik (1956).
This provides a method for comparing all stony me-
teorites regardless of the abundance or lack of vola-
tiles. The normative mineral composition, calculated
from the chemical composition as recommended by
Wahl (1950), and expressed in weight percentages, is
given in Table 3 (A).
CARBON DISTRIBUTION
The carbon content as reported by Moore and Lewis
(1967) (0.18%?Table 1) appears somewhat higher
than normal. In an attempt to determine the distri-
bution of carbon within this meteorite, a sample was
selectively drilled under a low-power microscope with
a dental drill (Fredriksson and Nelen 1969). The
collected material consisted of some finely powdered
matrix material and a few small chunks with adher-
ing chondrules. After separating powder and chunks,
both fractions were analyzed for carbon content with
a Leco low-carbon analyzer. The powdered matrix
fraction contained 1.04% carbon, the chunks 0.19%
carbon.
In a second experiment, two small samples were
gently crushed in a mortar. The coarser pieces were
then manually picked out and the remaining material
separated in several size fractions by sieving. With
this technique the softer, fine-grained matrix material
tends to concentrate in the finer fractions, while the
harder chondrules and metal tend to remain with the
coarser material. Carbon content of the separates is
given in Table 4. Similar results have been obtained
for other normal (i.e., with constant olivine-pyroxene
composition) H and L group chondrites. Apparently
these chondrites have a carbonaceous matrix similar
to that in unequilibrated or carbonaceous chondrites,
although in smaller amounts. The presence of a
carbonaceous low-temperature fraction (Fredriksson
and Keil 1964; Anders 1964) would argue against a
high-temperature metamorphic event such as sug-
gested by Dodd (1969) after the agglomeration of
the chondrites.
NUMBER 9 59
TABLE 2.?Chemical composition of the Forest Vale meteorite
A*
Constituent Weight percent Constituent Weight percent Constituent Atom percent
Fe 17.24
Ni 1.76
Co 0.09
FeS 5.54
SiO* 37.46
TiO2 0.13
A12O, 1.98
Cr2O, 0.46
FeO 8.61
MnO 0.31
MgO 23.23
GaO 1.69
Na2O 0.85
K2O 0.11
P2O? 0.31
H2O ( + ) 0.72
H2O (-) 0.06
C 0.18
Total 100.73
Fe 27.45
Si 17.44
Mg 14.01
S 2.02
Ni 1.76
Ga 1.21
Al 1.05
Na 0.63
Cr 0.31
Mn 0.24'
G 0.18
P 0.14
Co 0.09
K 0.09
H 0.09
Ti 0.08
O 33.21
Total 100.00
Si 33.84
Mg 31.39
Fe 26.78
Al 2.11
Ca 1.65
Ni 1.63
Na 1.49
Cr 0.32
P 0.25
Mn 0.24
K 0.13
Ti 0.09
Co 0.08
Total 100.00
a Chemical analysis expressed as nickel-iron, troilite, and oxides (Jarosewich 1967, with re-
vised figures for Fe and FeO).
b Chemical analysis expressed as elements, with oxygen added to make 100 percent.
c Chemical analysis expressed as atom percentages, with the elimination of O, C, H, and S.
TABLE 3.?Normative composition of the Forest Vale
meteorite
TABLE 4.?Carbon content of separates of the Forest
Vale meteorite
Constituent
Nickel-iron
Troilite
Bronzite
Olivine
Diopside
Albite
Anorthite
Orthoclase
Chromite
Apatite
Ilmenite
A
Bulk
19.1
5.5
31.6
28.5
4.0
7.2
1.3
0.7
0.7
0.7
0.2
B
Matrix
11.1
2.9
36.2
17.2
9.0
16.2
2.9
1.5
0.4
1.5
0.4
A. Normative composition calculated from chemical anal-
ysis according to Wahl (1950).
B. Matrix (44.6%)?Bulk normative minerals (column
A) minus modal amounts in weight percent of recognizable
mineral grains.
Size fraction
Sample 1
(% Carbon)
Coarse pieces
>80 Mesh
80-100 Mesh
< 100 Mesh
Sample 2
(% Carbon)
0.13
0.51
0.86
1.27
0.190.40
0.78
1.23
Mineralogy and Petrology
OLIVINE
In a crushed sample, olivine grains are abundant,
colorless, and frequently inclusion-filled. The fractures
are characteristically conchoidal.
The refractive indices are a =1.665 ?0.002, "Y =
1.702 ?0.002, indicating a content of 16 mole percent
60 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
of the Fe2SiO4 (Fa) component, according to the
determinative curve of Poldervaart (1950). The
(130) peak at 2.771A appears sharp and symmetrical
on the diffractograms indicating uniform composition.
According to the data of Yoder and Sahama (1957),
this peak shows the composition of olivine to be 16.5
mole percent Fe2SiO4.
Five or six points on each of about ten randomly
selected olivine grains, or a total of 55 points, were
analyzed with the microprobe. The olivine proved
Ni-poor (less than 0.02%) and with an essentially
homogeneous composition of Fal7.6.
In thin section, the crystal habit of olivine is vari-
able. Subhedral grains are most common, although
euhedral crystals do exist. The well-developed crystals
are commonly elongated parallel to the C-axis, six-
sided, and show (010) cleavage. Poikilitic inclusions
of small olivine grains in pyroxene are common. Grain
orientation is random, and crystals attain a maximum
length of 0.7 mm. Euhedral grains are length-slow
with parallel extinction.
PYROXENE
Pyroxene is also abundant and occurs as large pris-
matic crystals or fibrous aggregates. The grains are
colorless to opaque depending on the abundance of
inclusions, often part along cleavage planes, and fre-
quently display both simple and polysynthetic twins.
The refractive indices are a =1.670 ?0.002, and
y= 1.684 ? 0.002, indicating a composition of 15 mole
percent of the FeSiO3 component, according to
Kuno's data (1954). This falls in the compositional
range of bronzite. The diffractograms indicate both
abundant bronzite and one or more clinopyroxenes.
These are probably clinobronzite with accessory
diopside.
Both orthorhombic and monoclinic pyroxene re-
vealed a near constant composition of Fsis.g in both
twinned and untwinned crystals. A few anomalously
high Ga areas prompted the use of electron beam
scanning pictures taken of pyroxene grains and meas-
ured for Ca and Fe distribution, only one revealed
distinct Ca-rich rims (Figure 1). The enrichment oc-
Fe Ca Al
FIGURE 1.?Photomicrograph of twinned and untwinned pyroxene viewed by polarized trans-
mitted light at 450 X, upper left, the nearly correlative secondary electron emission photograph,
upper right, and electron probe elemental scans, lower three photographs, of iron, calcium, and
aluminum, moving from left to right.
NUMBER 9 61
curred on both twinned and untwinned crystals, and
the rims had a uniform composition of Wo34.9Fs10.6
En54.5. This Ca-armoring noted on only some rhom-
bic and monoclinic pyroxene grains seems to eliminate
any significant equilibrating influence which might be
attributed to thermal metamorphism.
The pyroxene crystals range in external shape from
finely fibrous to the characteristic prismatic habit.
The (110) cleavage and (010) partings are present
on the numerous large distinct crystals. Orthopyrox-
ene is most abundant, but half of the well-developed
grains, although of similar composition, are clino-
pyroxene and these display what appears to be poly-
synthetic twinning. The grains are randomly oriented
with the exception of the eccentrically radiating
chondrules made of fibrous crystals. The large crys-
tals, which attain a maximum length of 0.8 mm, are
riddled with inclusions, and are length-slow with
parallel extinction.
OLIGOCLASE
Although oligoclase was not identified optically, the
X-ray diffractograms suggest it to be a minor con-
stituent. The recalculated normative composition also
suggests that the dark matrix and devitrified glass in
barred chondrules is in part Na-rich plagioclase.
NICKEL-IRON
In reflected light, nickel-iron appears white, struc-
tureless, and in some grains bi-reflectant (possibly as
a result of deformation). The alloy appears grey and
nearly always isotropic under crossed nicols.
The maximum dimension of these irregularly
shaped masses, which occur most frequently between
the chondrules, is approximately 0.5 mm. The silicate
chondrules appear to control the shape and distribu-
tion of the large metal grains, but smaller, rounded,
nickel-iron particles are found in the matrix and are
included in chondrules. It is generally inclusion-free
and a limonitic alteration product usually exists
around the perimeters of the grains.
A very limited microprobe study of the Fe-Ni
phases was undertaken. Three tracks across two dis-
tinct grains were run at a speed of eight micron/
minute (Figure 2). The intensities for both Fe Ka
and Ni Ka were recorded while the pulses were simul-
taneously counted on the sealers at 10-second inter-
vals. The taenite appeared Ni-deficient when com-
pared to normal chondritic taenite, whereas the
kamacite was rather high in Ni. Also, there were no
low-Ni zones in the kamacite at the taenite-kamacite
interfaces. This suggests that equilibration of the Ni-
Fe was retained only through relatively high tempera-
tures and that the metal was probably subject to
rapid cooling across the taenite-kamacite phase
boundary, thus restricting Ni diffusion.
TROILITE
Plane polarized light reveals this iron sulfide to be
distinctly yellow in color and slightly bi-reflectant.
Its anisotropism is distinct, with colors ranging from
silver-grey to brown revealing polycrystalline troilite
as the typical texture. This accessory mineral, which
usually occurs in close association with nickel-iron,
attains maximum dimensions of approximately 0.5
mm. Like nickel-iron, the borders are highly irregu-
lar, and its distribution seems to be controlled by
the surrounding silicate chondrules. Besides occurring
as large, distinct masses within the matrix, minute
grains are included in the chondrules and the larger
of the nickel-iron fragments.
CHROMITE
The accessory chromite grains, which are grey, iso-
tropic, and structureless, occur most commonly as
extremely fine-grained aggregates within the matrix.
They do, however, attain a maximum dimension of
0.2 mm when included in chondrules.
MATRIX
The fine-grained, nearly isotropic matrix has very low
bi-refringence and occurs as interstitial aggregates
between chondrules. It does, however, merge with the
boundaries of some fine-grained chondrules. Although
the material could not be identified optically, the
recalculated norm suggests it is principally pyroxene,
plagioclase, and metallic iron.
GHONDRITIC TEXTURE
The Forest Vale meteorite is composed of olivine and
pyroxene chondrules uniformly distributed throughout
the stone and quite variable in their degree of crys-
tallinity. They often range up to 1 mm in diameter,
62 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
20- ,~ 10:..,"
a.
U
L.
ou
1
10 50
2
. 20 f ?? :
10:..., V?.....
1 ? J '10 40
20- ?
10--_-
10 30
Distance (u)
FIGURE 2.?Photomicrograph of nickel-iron observed at 190X and corresponding electron probe
compositional tracks. The low nickel content of the taenite and the lack of Ni depletion in the
kamacite at the taenite-kamacite interface is noteworthy.
retain their sphericity, and are set in a matrix of
fine-grained crystallites and opaque minerals (Figure
3).
MODAL ANALYSIS
A total of 5000 points were counted on two polished
thin sections. The mineralogical composition as cal-
culated from the mode (in weight-percent) is: nickel-
iron 14.2; chromite 0.5; troilite 4.2; olivine 20.9;
pyroxene 15.6; and 44.6 percent fine-grained matrix
material which could not be identified. The abun-
dance of pyroxene is higher in the norm than in this
mode, which suggests a large amount of the pyroxene
exists as fine-grained groundmass material, or as fine-
grained or glassy material within the chondrules, or
NUMBER 9 63
FIGURE 3.?Chondritic texture of the Forest Vale meteorite at 30 X using
polarized transmitted light.
both. Assuming that the insignificant amount of ma-
terial was lost during polishing, and recalculating
the norm with the elimination of nickel-iron, troilite,
and silicates noted in the modal analysis, the fine-
grained matrix material is probably rich in pyroxene,
plagioclase, and metallic iron and poor in olivine
(Table 3, column B) . The gross similarity of the
norms, however, seems to suggest that matrix and
chrondules are cogenetic.
Classification
If the amount of nickel-iron and the ratio of MgO
to FeO in the silicates are considered, as suggested by
Prior (1920), the Forest Vale meteorite is an olivine-
bronzite chondrite. With regard to both the total iron
content and the oxidation state of the iron, it would
be placed in the high-iron (H) group of Urey and
Craig (1953), while the amount of reduced iron
would place it in Wiik's group (1956) of chondrites
with 15 percent to 19 percent metallic iron. The ratio
Fe/Fe + Mg) in both olivine and pyroxene, as sug-
gested by Keil and Fredriksson (1964), would again
classify it as an H group chondrite. The Forest Vale
meteorite corresponds to that of the petrologic type 4
in the chemical-petrologic classification of Van
Schmus and Wood (1967). This group is character-
ized by well-developed chondrules, olivine and py-
roxene of uniform or nearly uniform composition,
and an abundance of clinopyroxene.
Conclusions
1. If not attributable to shock, the high-tempera-
ture nickel-iron and devitrified glass in barred chon-
drules indicate a brief cooling history for the
chondrite.
2. Ca-rich coronas occurring on only some pyrox-
ene grains indicate that the crystallization conditions
were not identical for all fragments.
3. The presence of metal particles in, and adjacent
to, chondrules, and the gross similarity in normative
mineralogy of both matrix and included fragments
suggest that although crystallization conditions were
not identical they may have been nearly the same.
4. The degree of chondrule integration, the hetero-
64 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
geneous pyroxene, the high C content in the matrix
and the high-temperature nickel-iron suggest accumu-
lation and lithification of the genetically related frag-
ments but with no subsequent metamorphism.
Acknowledgments
This manuscript is an extension of preliminary data
gathered and submitted in thesis form by A. F.
Noonan in partial fulfillment of the M.S. degree at
the University of Tennessee. This initial research was
carried out under a Smithsonian Research Founda-
tion Grant. Additional costs for extended studies were
supported by NASA and United States Air Force
grants. We are indebted to the Department of Geol-
ogy faculty at the University of Tennessee for an early
review of the paper, and to Dr. Brian Mason of the
Smithsonian Institution for many helpful suggestions
throughout the study.
Literature Cited
Anders, E.
1964. Origin, Age and Composition of Meteorites.
Space Science Reviews, 3:583-714.
Dodd, R. T.
1969. Metamorphism of the Ordinary Chondrites: A
Review. Geochimica et Cosmochimica Acta, 33:
161-203.
Fredriksson, K., and K. Keil
1964. The Iron, Magnesium, Calcium, and Nickel Dis-
tribution in the Murray Carbonaceous Chondrite.
Meteoritics, 2:201?217.
Fredriksson, K., and J. Nelen
1969. Carbon Distribution in Chondrites. Meteoritics,
4:271-272.
Hodge-Smith, T.
1942. A Fall of Meteorites at Forest Vale, New South
Wales. Australian Museum Magazine, 8:45?48.
1943. A Shower of Meteoritic Stones at Forest Vale,
New South Wales. Australian Journal of Science,
5:154-156.
Jarosewich, E.
1967. Analysis of Seven Stony Meteorites and One Iron
with Silicate Inclusions. Geochimica et Cosmo-
chimica Ada, 31:1103-1105.
Keil, K.
1967. The Electron Microprobe X-Ray Analyzer and
its Application in Mineralogy. Fortschritte der
Mineralogie, 44:4-66.
Keil, K., and K. Fredriksson
1964. The Fe, Mg, and Ca Distribution in Coexisting
Olivines and Rhombic Pyroxenes in Chondrites.
Journal of Geophysical Research, 69:3487-3515.
Kuno, H.
1954. Study of Orthopyroxene from Volcanic Rocks.
American Mineralogist, 39:30-46.
Mason, B.
1963. Olivine Composition in Ghondrites. Geochimica
et Cosmochimica Ada, 27:1011-1023.
Moore, C. B., and C. F. Lewis
1967. Total Carbon Content of Ordinary Chondrites.
Journal of Geophysical Research, 72:6289-6292.
Poldervaart, A.
1950. Correlation of Physical Properties and Chemical
Composition in the Plagioclase, Olivine, and Or-
thopyroxene Series. American Mineralogist, 35:
1067-1079.
Prior, G. T.
1920. The Classification of Meteorites. Mineralogical
Magazine, 19:51-63.
Urey, H. C, and H. Craig
1953. The Classification of the Stone Meteorites and
the Origin of the Meteorites. Geochimica et
Cosmochimica Acta, 4:36?82.
Van Schmus, W. R., and J. A. Wood
1967. A Chemical and Petrologic Classification for the
Chondritic Meteorites. Geochimica et Cosmo-
chimica Acta, 31:747-765.
Wahl, W.
1950. The Statement of Chemical Analysis of Stony Me-
teorites and the Interpretation of the Analysis in
Terms of Minerals. Mineralogical Magazine, 29:
416-426.
Wiik, H. B.
1956. The Chemical Composition of Some Stony Meteor-
ites. Geochimica et Cosmochimica Acta, 9:279?
289.
Yoder, H. S., and T. G. Sahama
1957. Olivine X-Ray Determinative Curve. American
Mineralogist, 42:475-491.
Roy s. Clarke, Jr. Iron Meteorite
and Eugene Jarosewich Compositions
During the past several years chemical analyses for the
elements nickel, cobalt, and phosphorus have been
performed on twelve meteorite specimens from the
Smithsonian collection. Classification or curatorial
considerations have normally prompted individual
investigations. The data are summarized in the ac-
companying table in order to make them available
to others.
Recent analyses of four of these same meteorites
Roy S. Clarke, Jr. and Eugene Jarosewich, Department of
Mineral Sciences, Division of Meteorites, National Museum
of Natural History, Smithsonian Institution, Washington,
D.C. 20560.
have been given by Moore et al. (1969), and their
paper includes a general discussion of the problems of
iron meteorite analysis. Our results on these four
meteorites are in excellent agreement with their
results. Nickel analyses have also been performed on
all twelve of these meteorites by Wasson as part of
his comprehensive study of chemical groupings among
iron meteorites (John T. Wasson, personal communi-
cation, 1970). Agreement is again good, but it was
noted that our nickel values are normally 0.1 to 0.2
percent higher than Wasson's. Older analyses of these
meteorites and references to descriptive papers are
cited by Hey (1966).
TABLE 1.?Chemical analyses for nickel, cobalt, and phosphorous on twelve meteorite
specimens from the Smithsonian collections
Name
NMNH
cat. no.
Structure
class* Ni
Weight percent
Co Notes
Angelica, Wisconsin
Arlington, Minnesota
Bodaibo, Siberia
Cacaria, Mexico
Cruz del Aire, Mexico
Cumpas, Mexico
Elbogen, Bohemia
El Capitan, New Mexico
Monahans, Texas
New Westville, Ohio
Obernkirchen, Germany ...
St. Genevieve County,
Missouri
2177
627
2149
1480
1441
1591
309
2754
1303
1412
1611
454
Om
Om
Of
Om
Of
Om
Om
Om
D
Of
Of
Of
7.50
8.55
8.05
7.74
9.19
8.14
10.25
8.80
10.79
9.38
7.67
7.96
0.50
0.41
0.36
0.46
0.50
0.47
0.64
0.48
0.52
0.34
0.36
0.36
0.11
0.02
0.04
0.11
0.30
0.18
0.22
0.43
0.09
0.18
0.02
0.24
(1)
(0(4)
(1)
(2)(5)
?Buchwald and Munck (1965).
(1) Recent analysis reported by Moore et al. (1969).
(2) Recent analyses' reported by Wasson (1967).
(3) A doubtful analysis for Ni in Monahans was reported by Cobb (1967).
(4) Analysis reported by Dyakonova and Kharytonova (1963).
(5) Recent analyses reported by Lewis and Moore (1971).
65
66 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
All of the meteorites in this group are finds and
have, therefore, been subjected to varying degrees of
terrestrial weathering. In each case material was
carefully selected to be free of oxidation and obvious
inclusions. Samples were in the size range of 1.2 to
3.0 g, large enough to be representative of the metal-
lic phases. They were dissolved in an HCI-HNO3
mixture, and then evaporated to dryness several times
with HC1. Residues were taken up with 5 ml of con-
centrated HG1, diluted with H2O and filtered into
a 100 ml volumetric flask. Appropriate aliquots were
taken for the determination of nickel, cobalt, and
phosphorus. Nickel was precipitated with dimethyl-
glyoxene and determined gravimetrically, cobalt spec-
trophotometrically with nirosto-R salt, and phos-
phorus spectrophotometrically using the molybedenum
blue method (Moss et al., 1961).
Literature Cited
Buchwald, V. F., and Sole Munck
1965. Catalogue of Meteorites. Analecta Geologica, 1:
1-81.
Cobb, James C.
1967. A Trace-Element Study of Iron Meteorites. Jour-
nal of Geophysical Research, 72:1329-1341.
Dyakonova, M. Y., and V. Y. Kharytonova
1963. Composition of Nickel-Iron in Some Iron and
Stony-Iron Meteorites of Various Types. Mete-
oritika, 23:42-44.
Hey, Max H.
1966. Catalogue of Meteorites, lxviii+637 pages. The
British Museum (Natural History).
Lewis, Charles F.3 and Carleton B. Moore
1971. Chemical Analyses of Twenty-eight Iron Meteor-
ites. Meteoritics, 6:195-205.
Moore, Carleton B., Charles F. Lewis, and David Nava
1969. Superior Analyses of Iron Meteorites, in Meteor-
ite Research. P. M. Millman, editor. Pages 738-
748. Dordrecht, Netherlands: D. Reidel Publish-
ing Company.
Moss,, A. A., M. H. Hey, and D. I. Bothwell
1961. Methods for the Chemical Analysis of Meteorites:
I. Siderites. Mineralogical Magazine, 32:802-
816.
Wasson, John T.
1967. The Chemical Classification of Iron Meteorites: I.
A Study of Iron Meteorites with Low Concentra-
tions of Gallium and Germanium. Geochbnica
et Cosmochimica Ada, 31:161-180.
Roy S. Clarke, Jr.,
Eugene Jarosewich,
and Joseph Nelen
The Nejo,
Ethiopia,
Meteorite
A stony meteorite fell 16 km due west of Nejo, near
the village of Jarso, Wollega Province, Ethiopia, at
11:30 a.m. local time on May 11, 1970. The co-
ordinates of the fall site are 9?30'N, 35?20/E. The
fall was observed by local residents and reported to
Professor Pierre Gouin, Director, Geophysical Observ-
atory, Haile Selassie I University, Addis Ababa, Ethi-
opia. Professor Gouin visited the site and reported
obtaining three pieces weighing a total of 2.4 kg that
had been recovered1, apparently from an individual
stone. The stone had made a small hole in hard
ground, and he estimated that its impact weight
was 4-4.5 kg.
Professor Gouin's reports on the meteorite fall and
visit to the site were circulated by the Smithsonian
Institution Center for Short-Lived Phenomena on
July 23, 1970, and August 4, 1970. On July 21, 1970,
Professor Gouin sent the Smithsonian Institution a
287 g specimen (NMNH 5420) of the Nejo meteorite
to be made available for time-dependent research.
Material was distributed for this purpose to Edward
Fireman, Smithsonian Astrophysical Observatory,
Cambridge, Massachusetts 02138, Louis A. Rancitelli,
Battelle Memorial Institute, Richland, Washington
99352, and G. D. O'Kelley, Oak Ridge' National
Laboratory, Oak Ridge, Tennessee 37830. Reports on
these investigations will appear elsewhere. We have
undertaken a brief mineralogical and chemical char-
acterization that will be summarized here. A detailed
Roy S. Clarke, Jr., Eugene Jarosewich, and Joseph Nelen,
Department of Mineral Sciences, National Museum of Nat-
ural History, Smithsonian Institution, Washington, D.C.
20560.
petrographic investigation is being undertaken by
Dr. Bill Moreton, Geology Department, Haile Selassie
I University, Addis Ababa.
Mineralogy and Chemistry
The Nejo meteorite is a hypersthene chondrite, an
L6 meteorite in the classification of Van Schmus
and Wood (1967). It is coarsely crystalline with a re-
crystallized matrix, and no distinct chondrules were
observed. The olivine and pyroxene were examined
by the electron microprobe and found to be equili-
brated. The olivine compositions clustered closely
around 25 mole percent fayalite (Fa2s) and the py-
roxene around 20 mole percent ferrosilite (Fs2o)-
Plagioclase, chlorapatite, and chromite were iden-
tified by the microprobe and the troilite was observed
to be essentially free of nickel. The presence of both
kamacite and taenite was observed microscopically
on etched polished sections and confirmed by the
microprobe. Kamacite and taenite occur as indi-
vidual grains, in association with one another, and
in association with troilite. Troilite grains are com-
paratively large and monocrystalline or coarsely crys-
talline. The larger kamacite grains frequently contain
slightly distorted Neumann bands. Evidence of strong
shock appears to be lacking.
Our specimen contained a small area of well-
developed fusion crust. The vesicular glass crust
ranged up to 0.3 mm thick and was seen in polished
section to contain precipitated magnetite. It is under-
lain by a zone injected with sulfide veins penetrating
as far as 1 mm below the surface (Ramdohr 1967).
Within this zone, particularly near the surface, are
67
68 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 1.?Chemical composition of the Nejo, Ethi-
opia, meteorite, NMNH 5420 (Eugene Jarosewich,
analyst)
Constituent
Fe
Ni
Co
FeS
SiO3
TiOa
AlaO,
Gr2O3
FeO
MnO
MgO
GaO
NaaO
K2O
PaO,
H2O( + )
ftO(-)
G
Total
Total Fe
D(g/cm8)
Percent
7.10
1.27
0.06
6.47
39.98
0.12
2.27
0.56
14.22
0.32
24.83
1.82
0.92
0.10
0.24
<0.1
0.03
0.01
100.32
22.26
3.59
eutectic structures resulting from the melting of
metallic phases and troilite. Kamacite grains through-
out this zone have been transformed to alpha-2
structure, indicative of heating to temperatures
around 500?C.
The bulk chemical analysis of the Nejo meteorite,
typical of hypersthene chondrites, is given in the ac-
companying Table 1. The analytical procedures em-
ployed are outlined in an earlier paper (Jarosewich
1966).
Literature Cited
Jarosewich, Eugene
1966. Chemical Analyses of Ten Stony Meteorites. Geo-
chimica et Cosmochimica Ada, 30:1261-1265.
Ramdohr, Paul
1967. Die Schmelzkruste der Meteoriten. Earth and
Planetary Science Letters, 2:197-209.
Van Schmus, W. R., and J. A. Wood
1967. A Chemical-Petrologic Classification for the Chon-
dritic Meteorites. Geochimica et Cosmochimica
Acta, 31:747-765.
Joseph A. Nekn The Nakhon
and Kurt Fredriksson
Meteorite
ABSTRACT
The Nakhon Pathom meteorite from Thailand has
been chemically analyzed and found to be an olivine-
hypersthene chondrite, belonging to the L-group, with
24 mole percent Fe2SiO4 in olivine and 21 mole per-
cent FeSiOg in orthopyroxene. It shows dark-light
structure of light gray areas surrounded by dark gray
bands, probably caused by hypervelocity shock. The
chemical composition of both parts is similar.
Introduction
The meteorite fell on December 21, 1923, at about
2100 hours. It was found in the province of Nakhon
Pathom, Thailand, probably shortly after the fall,
since it is fresh and unaltered. The place (13?44'N,
100?05'E) is a swampy area in the Don Yai Nom
subdistrict about 10 kilometers south-south-east of
the town of Nakhon Pathom. Two stones were placed
in a case of curios in the National Museum in Bang-
kok, because of their unusual properities. When Dr.
and Mrs. Edward P. Henderson were visiting Thai-
land in 1960, Mrs. Henderson saw the stones and
suspected them to be of meteoritic origin. The picture
showing both stones (Figure 1) was taken at that
time by Dr. Henderson. Eventually, the smaller stone
was loaned to the United States National Museum
for study. Mason (1967) has described the meteorite
briefly, giving full data and olivine composition. It is
not otherwise reported in the literature. The main
Joseph A. Nelen and Kurt Fredriksson, Department of Min-
eral Sciences, National Museum of Natural History, Smith-
sonian Institution, Washington, D.C. 20560.
mass is in the National Museum in Bangkok, Thai-
land. The United States National Museum has 413
grams of the smaller piece in its collection (NMNH
3192).
The total weight of the two pieces was 32.2 kilo-
grams. The meteorite shows light-dark structure, the
density of the lighter part being 3.56 gm/cc of the
darker part, 3.51 gm/cc. The average density of a
piece containing 41% dark material and 59% light
is 3.53. Stones with a light-dark structure were first
described by Reichenback (1865). Similar meteorites
include Walters (Roy et al. 1962), Stalldalen (Fred-
riksson et al. 1963), and Lanzenkirchen (Kurat and
Kurzweil 1965).
Description
The meteorite Nakhon Pathom broke in two pieces.
The smaller piece, weighing 9.6 kilograms, is covered
on two sides with a dark brown fusion crust 2 milli-
meters thick. The side oriented toward the front
during the fall shows well-formed "thumbprints,"
covered by a smooth dull crust (Figure 2). The back
side is flat and is covered with a rougher crust, tra-
versed by cracks.
A cross-section (Figure 3) shows a brecciated struc-
ture of light gray material surrounded by areas of
darker gray. In the darker parts many smaller round
to angular light gray bodies can be observed. The
lighter parts are broken pieces of a well-crystallized
chondrite. They contain only a few recognizable
chondrules or parts of chondrules, strongly intergrown
with the coarsely crystalline matrix (Figure 4). The
chondrules and crystals show strong mechanical de-
formation. The cracks are frequently filled with
droplets or angular particles of troilite and nickel-
69
70 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 1.?The two pieces of the Nakhon Pathom meteorite in the National Museum, Bangkok,
Thailand (smaller stone 210 X 160 mm).
iron. The black veins penetrating the lighter mass
consist mostly of comminuted silicates and opaque
minerals.
Mineralogy
The minerals identified in a polished thin section of
the meteorite are olivine, orthopyroxene, clinopyrox-
ene, plagioclase, whitlockite, chromite, kamacite, and
taenite.
Olivine is mostly xenomorphic, has a wavy extinc-
tion, and is often, especially in the dark part, mixed
with nickel-iron and troilite. The composition in both
dark and light parts as determined by electron
microprobe analysis is given in Table 1.
The average fayalite content, Fa24, is the same as
that given by Mason (1967) who obtained his figure
by X-ray diffraction. Orthopyroxene is also mostly
xenomorphic, shows strong wavy extinction, and is
mixed with nickel-iron and troilite in the same man-
ner as olivine. The composition is given in Table 2;
it averages 21 mole percent FeSiOa.
Most of the clinopyroxene exhibits similar proper-
ties. A little calcium-rich clinopyroxene was also ob-
served. Plagioclase with wavy extinction occurs oc-
casionally as interstitial grains, both in chondrules
NUMBER 9 71
FIGURE 2.?The smaller piece of the Nakhon Pathom meteorite, showing comparatively thick
fusion crust and prominent regmaglypts (210 X 160 mm).
and in the groundmass. Whitlockite is common as
interstitial crystals or aggregates of fairly large size,
especially in the light gray areas. Chromite occurs
together with nickel-iron and troilite, as well as
intergrown with silicates; it is fairly large and almost
xenomorphic. Troilite is mixed with nickel-iron in
the larger aggregates. Separately it occurs as small
droplets and irregular-shaped aggregates in the inter-
stices of the silicates. Kamacite and taenite occur
separately from each other and could not be distin-
FIGURE 3.?Polished surface on the main mass of the
Nakhom Pathom meteorite, showing prominent light-dark
structure, together with veining and brecciation (130 X
90 mm).
72 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 4.?Photomicrograph of thin section of the Nakhon
Pathom meteorite in .ordinary light. The dark part, upper
half, has barely recognizable chondrules and a brecciated
texture. In the light, lower part chondrules and matrix are
almost entirely intergrown. The black areas consist of nickel-
iron and troilite. The length of the section is 9 mm.
TABLE 1.?Olivine composition of the Nakhon
Pathom meteorite in light and dark portions
Area
Light
Dark
FeO weight percent
Range
21.9-22.5
21.9-23.4
Average*
22.2
22.5
FeO
FeO+MgO
Mole percent Fa
24.0
24.3
* The averages represent ten to fifteen grains with about
three to four measurements in each grain.
TABLE 2.?Orthopyroxene composition of the Nakhon
Pathom meteorite in light and dark portions
Area
Light
Dark
FeO weight percent
Range
13.8-14.7
13.9-14.8
Average*
14.2
14.2
FeO
FeO+MgO+CaO
Mole percent Fs
21.1
21.2
guished microscopically, but were identified with the
electron microprobe.
Chemistry
The sample selected for chemical analyses had a total
weight of 18.8 grams. The light and dark material
was carefully separated and each portion reweighed.
Total recovery was about 18.6 grams of which 41.1%
was dark and 58.9% light.
Each portion was analyzed separately by techniques
previously described by Jarosewich (1966, 1967). The
results are expressed as oxides, troilite, and metal in
Table 3. The elemental composition with oxygen by
difference and the elements calculated on a volatile-
free basis are given in Table 4 under A and B. The
normative mineral content, expressed as weight per-
centages in Table 5, was calculated as suggested by
Wahl (1950). It agrees well with the observed min-
eral abundances. The phosphorus was calculated as
apatite. Microscopic observation and semiquantitative
microprobe measurements indicate that the phos-
phorus mineral is whitlockite. The titanium is prob-
ably associated with the silicate phases, since ilmenite
was not found.
TABLE 3.?Chemical composition of the Nakhon
Pathom meteorite (weight percent) (Joseph Nelen,
analyst)
Constituent
?The averages represent ten to fifteen grains with one
to three analyses in each grain.
Fe
Ni
Go
FeS
SiO2
TiO2
A12O,
Cr2O8
FeO
MnO
MgO
GaO
Na2O
KaO
P3O?
H2O+
HjO-
G
Total
Fe (Total)
Dark portion 41.06% Light portion 58.94%
7.08
1.17
.05
6.46
39.98
0.17
1.81
0.60
13.52
0.31
24.71
1.84
0.98
0.10
0.26
0.58
0.07
0.08
99.77
21.69
8.15
1.22
.06
5.84
40.00
0.17
1.79
0.51
13.26
0.31
24.45
1.83
0.97
0.10
0.24
0.66
0.04
0.14
99.74
22.17
NUMBER 9 73
TABLE 4.?Elemental composition of the Nakhon
Pathom meteorite
Constituent
H
C
O
Na
Mg
Al
Si
P
S
K
Ca
Ti
Gr
Mn
Fe
Go
Ni
Total
Dark portion
A B
0.07
0.08
37.04
0.73 1.21
14.90 24.70
0.96 1.59
18.69 30.98
0.11
2.36
0.08 0.13
1.32 2.19
0.10 0.16
0.41 0.67
0.24 0.40
21.69 35.95
0.05 0.08
1.17 1.94
100.00 100.00
Light portion
A B
0.08
0.14
36.90
0.72 1.19
14.75 24.32
0.95 1.57
18.70 30.83
0.10
2.13
0.08 0.13
1.31 2.16
0.10 0.16
0.35 0.58
0.24 0.40
22.17 36.55
0.06 0.10
1.22 2.01
100.00 100.00
A. Chemical analysis in weight percent of the elements,
with oxygen to bring the sum to 100.
B. Atomic percent of the elements on a volatile (C,H,O,
S)-free basis.
TABLE 5.?Normative minerals of the Nakhon
Pathom meteorite (weight percent)
Constituent
Nickel-iron
Troilite
Olivine (Fa 24) . .
Hypersthene (Fs 21)
Diopside
Orthoclase
Albite
Anorthite
Apatite ...
Ghromite
Ilmenite
Dark portion
8.30
6.46
41.14
26.07
5.81
0.61
8.20
0.22
0.60
0.95
0.32
Light portion
9.43
5.84
39.40
27.38
5.83
0.61
8.17
0.22
0.57
0.83
0.32
Summary
The chemical and mineralogical composition of the
Nakhon Pathom meteorite shows that it is an olivine-
hypersthene chondrite in Prior's classification (1920).
The total iron content, about 22 percent, places it
in the low-iron (L) group of Urey and Craig (1953).
The olivine composition, Fa24, determined by micro-
probe analysis, also confirms that this meteorite is
an L-group chondrite (Keil and Fredriksson 1964).
The minerals present are the same in the light and
dark parts both according to microscopic observations
and chemical composition. The microprobe analyses
appear to show a slightly higher iron content and
more variability for olivines in the dark portion. The
difference, however, is practically within the limits
of analytical error and is not considered significant.
The bulk chemical composition, Table 3, is also sim-
ilar in both parts and the meteorite may therefore be
classified as a monomict breccia (Wahl 1952).
The meteorite was severely shocked. The dark and
light structure and the black veins, containing small
troilite bodies, are similar to the black veins and
the darkening produced in the gray chondrite Stall-
dalen by artificial shock as described by Fredriksson
etal. (1963).
It should also be noted that the carbon content
is lower in the dark part than in the light (Table 3),
and consequently the dark color is not caused by
carbon as suggested by Wlotzka (1963) for the
Pantar meteorite. The carbon values are averages of
five analyses of each part, of which two of each were
carried out by Dr. Gharleton Moore, Arizona State
University. All analyses agreed with ?0.03%C.
Acknowledgments
We wish to thank Dr. Louis Walter and Dr. Brian
Mason for their critical review of the paper.
Part of the expenses of this paper have been met
by grants from the National Aeronautics and Space
Administration (Grant NsG?688) and the United
States Air Force (Contract AF19(628)-5552).
Literature Cited
Fredriksson, K., P. S. De Carli, and A. Aaramae
1963. Shock-Induced Veins in Chondrites. Space Re-
search 111:974-983. North Holland Publishing
Company, Amsterdam.
Jarosewich, E.
1966. Chemical Analyses of Teh Stony Meteorites. Ge-
ochimica et Cosmochimica Ada, 30:1261?1265.
1967. Chemical Analyses of Seven Stony Meteorites and
One Iron with Silicate Inclusions. Geochimica et
Cosmochimica Ada, 31:1103-1106.
74 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
Keil, K., and K. Fredriksson
1964. The Iron, Magnesium and Calcium Distribution
in Coexisting Olivines and Rhombic Pyroxenes of
Chondrites. Journal of Geophysical Research,
69:3487-3515.
Kurat, G., and H. Kurzweil
1965. Der Meteorit von Lanzenkirchen. Annalen des
Naturhistorischen Museums in Wien, 68:9?24.
Mason, B.
1965. The Chemical Composition of Olivine-Bronzite
and Olivine-Hypersthene Chondrites. American
Museum Novitates, 2223:38.
1967. Olivine Composition in Chondrites?A Supple-
ment. Geochimica et Cosmochimica Acta, 31:
1100-1103.
Prior, G. T.
1920. The Classification of Meteorites. Mineralogical
Magazine, 19:51-63.
Reichenbach, C. V.
1865. Die schwarzen Linien und Ablosungen in den
Meteoriten. Poggendorff's Annales der Physik
und Chemie, 125:308-325, 420-441, 600-618.
Roy, S. K., J. J. Glass, and E. P. Henderson
1962. The Walters Meteorite. Chicago National History
Museum Fieldiana. Geology, 10:539-550.
Urey, H. C, and H. Craig
1953. The Composition of the Stone Meteorites and the
Origin of the Meteorites. Geochimica et Cosmo-
chimica Acta, 4:36-82.
Wahl, W.
1950. The Statement of Chemical Analyses of Stony
Meteorites and the Interpretation of the Analyses
in Terms of Minerals. Mineralogical Magazine,
29:416-426.
1952. The Brecciated Stony Meteorites and Meteorites
Containing Foreign Fragments. Geochimica et
Cosmochimica Acta, 2:91-117.
Wlotzka, F.
1963. Uber die Hell-Dunkel-Structur der urgashaltigen
Chondrite Breitscheid and Pantar. Geochimica
et Cosmochimica Acta, 27:419-429.
MINERALOGY
j. J. Gumey Gorceixite in the
Buffels River,
Namaqualand,
South Africa
Gorceixite is a hydrated basic phosphate of barium
and aluminium of uncertain formula but approxi-
mating to [BaAl3(PO4)2(OH)5H2O] (Hussak 1906).
The mineral has been reported to be present in
carbonatites such as Mrima Hill, Kenya, (Binge and
Joubert 1966), but is better documented as a "Fava"
or pebble in the diamantiferous gravels of Brazil,
British Guiana, West Africa, Rhodesia, and now
South Africa. Gorceixite is of no economic value in
itself but is generally thought to be a good indicator
of the presence of diamonds (Beard 1941).
In Brazil the mineral has been reported to occur
as pebbles in the diamantiferous gravels of the rivers
Abiete Bagagem and Douradinhos in the Minas
Geraes, the Paranhijba and Verissimo in Goyaz, and
the Patrocinio de Sapucahy and the Canoas in Sao
Paulo. Gorceixite has also been reported to be asso-
ciated with diamonds at Issineru, British Guiana;
Bonsa River, Gold Coast; Oiyi District, Sierra Leone;
and at Somabula, Rhodesia (Palache, Berman, and
Frondel 1951). More recently it has been discovered
in the Bongou River in French Equatorial Africa
(Miller 1959).
Pebbles of gorceixite contain some impurities, usu-
ally of silica, iron, and titanium. Calcium and the
rare earth elements may substitute for barium and a
strontium analogue of the mineral (goyazite) has
been reported from the same gravel deposits.
The pebbles found in the Buffels River were ob-
tained from the heavy mineral concentrate of the
/. /. Gumey, Department of Mineral Sciences, National
Museum of Natural History, Smithsonian Institution, Wash-
ington, D.C. 20560.
diamond mining operations of the Buffelsbank Dia-
mante Edms. Bpk. This company is recovering dia-
monds from a quaternary drainage channel of the
Buffels River in the Spektakel Valley, west of Spring-
bok, Namaqualand.
The pebbles range from greyish white to light
brown to dark brown and are up to 2 cm across. An
analysis of a light brown pebble is given in Table 1.
This pebble had a porcelaneous fracture, hardness of
about 6, specific gravity of 3.14, and vitreous lustre.
An X-ray diffractogram confirmed that the mineral
was gorceixite.
In the analysis reported, some calcium, strontium,
and rare earth substitution of barium is apparent.
Neither the ideal formula, nor that suggested as an
alternative by Palache, Berman, and Frondel (1951)
are identical to the analytical result seen after the
obvious impurities are ignored (see Table 1).
The source of the gorceixite in the Buffels River
is unknown and there is no reported association be-
tween gorceixite and diamond except in river gravels.
It seems probable that similarity in specific gravity
was the chief agent in bringing the two minerals into
close association in this environment. Possibly gor-
ceixite is more common than might be supposed from
its sparsely reported occurrences. It is unimpressive in
appearance and can easily be overlooked.
Acknowledgments
The author would like to thank Mssrs F. Hoffman
and F. Mostert, Managing Director and Mine Man-
ager of the Buffelsbank Diamante Edms. Bpk, for
77
78 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
TABLE 1.?Analysis of gorceixite from the Buff els
River, Cape Province, South Africa
Constit-
uent
BaO
CaO
Total
R. E. ..
SrO
AUO,
PaO8
HSO+ .
SiOa
TiO,
FesO, ....
Na2O
KaO
HsO-
Total ...
Chemical anlaysis
1*
Ba Ah(POi)i
(OH)sHsO
29.98
29.96
27.75
12.31
100.00
2*
Gorceixite
Buffels
River
23.38
0.92
1.71
1.50
27.64
19.76
16.11
2.63 '
0.95
2.49
0.29
0.15
0.32
97.85
(weight percent)
3* 4*
Recalcu- Ba AL(POi)i
lated less (OH)sHzO
impurities
25.12 26.02
1.00
1.86
1.62
29.70 34.59
21.23 24.10
17.11 15.29
> Impurities
97.64 100.00
* Column /: Theoritical composition.
Column 2: New analysis, J. J. Gurney.
Column 3: Recalculation after subtraction of probable
impurities (SiO2, T1O2, FesOj, NasO, K2O, H2O-).
Column 4: Ideal composition.
submitting the samples investigated for examination.
Permission to publish the results was kindly granted
by the Board of Directors of the operating company
and the X-ray diffraction work was carried out by
P. Hofmeyr.
Literature Cited
Beard, E. H.
1941. Recent Discoveries of Gorceixite. Bulletin of
Imperial Institute, 39(2): 160-164.
Binge, J, and P. Joubert
1966. The Mrima Hill Niobium Deposit, Coast Province
Kenya. Information Circular No. 2 Republic of
Kenya Geological Survey.
Hussak, E.
1906. Uber die Sogenannten 'Phosphat?Favas' der Dia-
mantfuhrenden Sande Brasiliens. Mineralogische
und Petrographische Mittheilungen, 25:336.
Miller, M.
1959. Sample Submitted to Smithsonian Institution 3/
2/59. Locality Bongou River, French Equatorial
Africa.
Palache, C, H. Berman, and C. Frondel
1951. Dana's System of Mineralogy, 7th edition, 2:833.
New York: John Wiley and Sons.
Brian Mason ToCOmallte
In 1968 I received from Sir Maurice Mawby (Mel-
bourne, Australia) a mineral specimen from Broken
Hill, New South Wales, for identification. The min-
eral for identification filled a small cavity in a coral-
loid mass of embolite; it was soft, mealy, bright yellow
in color but extremely photosensitive, rapidly turning
gray-green and then black on exposure to bright
light. Sir Maurice had obtained it in 1922 from the
underground manager of the Block 14 mine at Broken
Hill (the Block 14 mine is now part of the North
Broken Hill property).
The association with embolite and the photosensi-
tivity immediately suggested a silver halide. However,
an X-ray powder photograph gave a distinctive pat-
tern (Table 1) which could not be matched with any
of the silver halides in the ASTM index. On check-
ing against the silver halide minerals in the literature,
my attention focused on the obscure mineral tocorna-
lite, as described in Palache et al. (1951) :
A name given to a supposed iodide of silver and mercury
found with other ill-defined chloride-iodides of silver and
mercury at Chanarcillo, Chile. Granular massive. Color pale
yellow, becoming darker on exposure. Streak yellow. An
analysis of impure material gave Ag 33.80, Hg 3.90, I 41.77,
siliceous residue 16.65, total 96.12; the loss is due to some
water belonging in the residue, and probably some iodine.
A mineral referred to tocornalite was found at the Proprie-
tary Mine, Broken Hill, New South Wales, as a deep yellow
powder, blackening in sunlight, associated with iodyrite and
cinnabar.
In the mineral collection of this department was one
small specimen (No. C906) labeled "Tocornalite,
Chanarcillo, Chile"; material from this specimen
gave an X-ray powder photograph identical with that
from the Broken Hill specimen.
The published identification of tocornalite at
Broken Hill was the work of Mr., George Smith
(1926), who was associated with the mines for many
Brian Mason, Department of Mineral Sciences, National
Museum of Natural History, Smithsonian Institution, Wash-
ington, D.C. 20560.
TABLE 1.?X-ray powder photograph of tocornalite,
Broken Hill, Australia (Cu Ka radiation, intensities
visually estimated)
Line Intensity d-spacings
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
7
2
2
9
9
1
4
4
3
4
10
1
2
7
3
5
4
2
2
3
4
3
6.37
4.67
4.51
3.76
3.61
3.39
3.29
3.19
2.998
2.894
2.644
2.468
2.309
2.254
2.127
2.036
1.877
1.834
1.743
1.695
1.634
1.583
years from 1888 onward, first as an ore buyer and
assayer, then as a mine manager, and subsequently
as Inspector of Mines for twenty-one years. It is
eloquent testimony to his discerning eye and his min-
eralogical knowledge that he was able to recognize
tocornalite in the Broken Hill material from the de-
scription of the Chilean mineral published many years
earlier. The Proprietary Mine, from which his mate-
rial came, was contiguous with the Block 14 Mine.
Tocornalite was described by Ignacio Domeyko in
1867. I have attempted to find type material of this
mineral, both in Chilean and in European collections,
without success. However, the Chanarcillo specimen
in our collection does correspond well with the origi-
79
80 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
nal description. Unfortunately, neither in this speci-
men nor in the Broken Hill material is there sufficient
tocornalite for a quantitative chemical analysis to
check the composition arrived at by Domeyko. Ex-
periments at synthesizing this mineral have so far been
unsuccessful; a variety of products has been obtained
by precipitating mixed silver and mercury solutions
with potassium iodide, but none of the precipitates
has given an X-ray powder photograph agreeing with
that of tocornalite. Attempts at indexing the powder
photograph of tocornalite and thereby determining its
crystal system, cell dimensions, and ultimately its struc-
ture have also been unsuccessful.
Tocornalite is evidently a valid species, occurring at
both Chanarcillo and Broken Hill, but additional
work on better material is needed to characterize it
completely.
I am indebted to John S. White, Jr., for the prepa-
ration of numerous X-ray powder photographs.
Literature Cited
Domeyko, Ignacio
1867. Mineralojia de Chile, Appendix 11:41.
Palache, C, H. Berman, and C. Frondel
1951. Dana's System of Mineralogy, 7 th edition, 2:1124.
New York: John Wiley and Sons.
Smith, George
1926. A Contribution to the Mineralogy of New South
Wales. New South Wales Mines Department,
Mineral Resources, 34: 145.
STANDARDS FOR CHEMICAL ANALYSIS
Eugene Jarosewich l Analysis
of Five Minerals
for Microprobe
Standards
There is a growing need for the availability of reliable
microprobe standards for chemical analysis of various
materials, in our particular case for the analysis of
minerals.
Unfortunately, there are a very limited number of
acceptable standards. Perhaps one of the main rea-
sons for this is lack of homogeneous materials, an
essential property of microprobe standards. Homo-
geneity should be the micron level, and the material
should be stable under the electron beam. Also, there
Eugene Jarosewich, Department of Mineral Sciences, Na-
tional Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560.
should be sufficient material for a chemical analysis.
We have analyzed five minerals which have been
successfully used as microprobe standards. These
standards served very well in plotting the calibration
curves with standards from other sources, and excel-
lent results were obtained. Nevertheless, we still are
contemplating a more thorough study of homogeneity.
These standards were obtained by crushing hand
specimens of a rock or mineral and then selecting a
fraction between 80 and 60 mesh for heavy liquid and
Franz magnetic separations. Each sample was visually
inspected for impurities. In this manner between 0.8 g
to 1.5 g of sample was obtained. Approximately 20
grains of each sample were mounted on a one-inch
TABLE 1.?Chemical analyses of minerals for use as microprobe standards (weight
percent)
Constituent
Kakanui Kakanui Omphacite Garnet Garnet
Hornblende Pyrope (USNM110607) (USNM 87375) (USNM110752)
SiO,
A1*O,
Fe?Os
FeO
MgO
CaO
Na3O
K,O
H*O +
H^O-
TiO,
P?O.
MnO
Total
40.37
14.90
3.30
7.95
12.80
10.30
2.60
2.05
0.90
0.04
4.38
0.00
0.09
99.66
41.46
23.73
N.D.
10.68
18.51
5.17
N.D.
N.D.
N.D.
<0.01
0.47
N.D.
0.28
100.30
55.42
8.89
1.35
3.41
11.57
13.75
5.00
0.15
N.D.
0.02
0.37
ND.
0.10
100.03
39.47
22.27
2.77
13.76
6.55
14.39
N.D.
N.D.
N.D.
<0.01
0.39
0.59
100.19
40.16
22.70
2.17
9.36
7.17
18.12
N.D.
N.D.
N.D.
<0.01
0.35
0.19
100.22
83
84
glass and checked for major elements homogeneity on
the microprobe. The chemical analyses were per-
formed employing classical methods (Hillebrand 1953
and Peck 1964). Sodium and potassium were deter-
mined by flame photometry.
Omphacite USNM 110607 was separated from ec-
logite, garnet USNM 87375 and garnet USNM 110752
were separated from kyanite eclogite, all from Robert
Victor Mine, South Africa. Kakanui hornblende and
pyrope were prepared from nodules described by
Brian Mason (1966).
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
Literature Cited
Hillebrand, W. F., G. E. F. Lundell, H. A. Bright, and J. I.
Hoffman
1953. Applied Inorganic Analysis. 2nd edition, 1034
pages. New York: Wylie and Sons.
Mason, B.
1966. Pyrope, Augite and Hornblende from Kakanui,
New Zealand. New Zealand Journal of Geologv
and Geophysics, 9.
Peck, L.
1964. Systematic Analysis of Silicates. Geological Sur-
vey Bulletin, 1170.
THE COLLECTIONS
Paul K DesautdsAcquisition of the
Carl Bosch Collection
of Minerals
and Meteorites
A five-year effort by the Division of Mineralogy to
acquire ownership of the Carl Bosch Collection of
minerals and meteorites was brought to a successful
conclusion in 1970. This collection, once belonging
to the late Nobel Prize winner, Professor Dr. Carl
Bosch (Figure 1), is a very large and important col-
lection of about 25,000 specimens including almost
600 meteorites. It is especially strong in specimens
from old, exhausted, Central European localities. The
representative suites of older species in a wide range
of variation from many now-extinct mines are un-
matched in any other private collection. These and
many other specimens in it are of materials that are
most difficult or impossible to acquire today. The
meteorite segment of the collection is a major collec-
tion in its own right, consisting largely of falls which
are unavailable now. Considering the difficulty and
expense of acquiring any new meteoritic material
this is an incredible accumulation. As is true with
most large collections, the Bosch Collection contains
several pre-existing collections. Among the best of
these is the well-known Seligmann Collection which
contains a large number of specimens that have been
described in the mineralogical literature of the past.
Assembled and enjoyed by Dr. Carl Bosch, the
collection was transported, for safekeeping, to the
United States from Germany in 1951. Arrangements
were made for its deposition at Yale University on a
fifteen-year loan. By 1966, with the loan period ex-
piring and Yale University having made no move to
acquire the collection, the heirs of Dr. Bosch decided
Paul E. Desautels, Department of Mineral Sciences, National
Museum of Natural History, Smithsonian Institution, Wash-
ington, D.C. 20560.
to dispose of the collection through sale. Although
little used by Yale in the fifteen-year period, the col-
lection had been kept intact and in relatively good
condition. Its original catalog was in rather good
agreement with the contents of the collection.
For several reasons it was considered imperative to
acquire and place the collection in the Smithsonian
Institution for conversion to maximum use as part of
the National Collection of minerals and meteorites.
Firstly, the collection was of major size with its 25,000
specimens including the nearly 600 meteorites. Collec-
tions of such size rarely become available for whole-
sale reinforcement of constantly changing specimen
needs of the National Collection. In numbers, though
not in quality or species representation, it would be
the largest single collection of minerals ever acquired
by the National Museum of Natural History. Also,
the collection contained extensive suites of rare refer-
ence specimens unobtainable through normal chan-
nels. Even those specimens duplicating others already
in the Museum were considered usable for supplying
additional material to meet ever-increasing research
and exchange demands. Several hundred pieces were
thought to be of public exhibit quality. The strong
representation of old Central European mineral oc-
currences was specially significant because of the new
support which would be available for the weakest part
of our Museum collection. Being an older collection,
much of the specimen material was of a type totally
unavailable today through normal sources. Of course,
the presence of the included Seligmann Collection
with its many described specimens increased the num-
ber of scientifically significant pieces generally un-
available at any price. The meteorite materials had
87
88 SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
FIGURE 1.?Dr. Carl Bosch?late Nobel Prize winner?who assembled the Bosch Collection of
minerals and other natural history collections in the early 1900s. Photograph supplied by Dr.
Henning Borchers.
unusually high, current, scientific importance and
needed to be carefully conserved. Our meteorite col-
lections were already under such careful control that
conservation-research procedures were already set up.
Finally, the excellent state of preservation and organiza-
tion of the collection would make it possible to con-
vert it to maximum use in minimum time (Figure 2).
The Bosch Collection had already been openly
offered for sale by the time the decision had been
made to attempt acquisition. When it appeared that
there was a distinct possibility the collection would be
sold outside the United States, negotiations were ac-
celerated. There was no established price but our own
appraisal and independent appraisals were arranged
and a reasonable price of $250,000 was agreed upon.
In retrospect, faced with a continuing rapid rise in
NUMBER 9 89
FIGURE 2.?Cafarsite?a newly (1966) described mineral species. This very fine specimen was
found in the Bosch Collection where it had been labeled as an "unknown." Thus, Dr. Bosch had
discovered the species long before its "official" discovery.
the market price of mineral specimens, this was a very
fair agreement. By this time, Dr. Carl Bosch, Jr., had
concluded that the Smithsonian was the logical re-
pository for his father's collection. When it became
obvious that there was little hope of our raising the
total payment, he was willing to help as much as
possible. Accordingly, Dr. Bosch agreed to donate
$70,000 of the value with the remaining $180,000 to
be paid over a period of five years. With the aid of
a $47,000 grant from the National Science Founda-
tion and $13,000 in available private funds, an initial
payment was made in the spring of 1966. Necessary
packing and moving arrangements were made and
the collection was transferred promptly to Washington
in two large vans. The problems of packing and mov-
ing such a large collection containing much fragile
material were anticipated in advance and the mov-
ing company was so cooperative that there were no
mishaps of any account during the transfer. Most of
the actual handling of specimens was, of necessity,
done by our own staff.
Without waiting for complete ownership, the sort-
ing, relabeling, cataloging and integration of the col-
lection into our existing operation were begun. To
date, about 14,500 specimens have been processed.
Since the original collection's cataloging and labeling
are not up to our present standards, the task is diffi-
cult. German script of names and localities?often
abbreviated?must be translated and related to mod-
ern speciation and geography. These difficulties, to-
gether with the necessary physical handling and the
shortage of trained personnel, have made cataloging
the collection slow work but completion of the proj-
ect is expected in 1971. The size and nature of this
collection as a unit suggested the possibility of using
it for a pilot program in computerization of the
mineral collection. Accordingly, cataloging has been
arranged in such a way as to make the necessary
information available eventually for the computer.
Meanwhile, this important collection has already
been pressed into service for our normal research,
educational, and exhibits programs.
THIN-SECTION PREPARATION
Grover Moreland,
Frank Ingram,
and Harold H. Banks, Jr.
Preparation of
Doubly Polished
Thin Sections
A technique for preparing polished thin sections
which show greater textural and mineralogical detail
has been developed for stony meteorites, rocks, and
minerals.
The preparation of polished thin sections?sections
which can be examined in both reflected and trans-
mitted light and used for electron microprobe
studies?has been discussed at length by Moreland
(1968). This procedure, in summary, is to cement
an optically flat surface of a specimen, which has been
sized for a petrographic section or for a circular
microprobe section, to a glass slide and then to hand-
grind the specimen to a thickness of 30 microns. Next,
the opaques are given a rough polish, then the non-
opaques are polished, and lastly the section is sub-
jected to a final polish.
In the course of preparing polished thin sections
of deep-sea basalts, volcanic ejecta, and Apollo 11
and 12 samples, it was observed that the polishing of
both sides of a sample enhanced the optics by
yielding sharper images, particularly of groundmass
minerals in fine-grained rocks and of inclusions in
minerals.
The first step in the preparation of a doubly pol-
ished thin section is to cut a wafer of the specimen,
approximately 7/8 inch square and 3/16 inch thick
with smooth parallel surfaces, using a 5-inch by 0.015-
inch diamond saw blade. The wafer is then ground
on one face to an optically flat surface, using a water
slurry of 600 aluminum oxide on an 850 rpm, 12-inch
cast-iron lap.
Grover Moreland and Harold H. Banks, Jr., Department of
Mineral Sciences, National Museum of Natural History,
Smithsonian Institution, Washington, D.C. 20560. Frank
Ingram, Ingram Laboratories, Griffin, Georgia 30223.
After washing the wafer with water and drying it
for approximately 30 minutes at 200 ?F, it is im-
mersed, under vacuum, in a mixture composed of
equal portions of Araldite AY105 and Hardner 935F
(Moreland 1968:2071) dissolved in three parts of
toluene. The vacuum occasionally is broken and re-
stored to help force the impregnating solution into the
pores and cavities of the sample. The wafer is re-
moved from the vacuum and the impregnating solu-
tion, after two to three hours, and placed on a hot
plate or in a drying oven to allow the impregnating
medium to cure, which takes about one hour at
200?F and 15 minutes at 300?F. The wafer is re-
ground to expose the sample and the impregnation
procedure is repeated.
The next step is polishing the ground face of the
wafer. For this, a double thickness of kraft, heavy-
duty, wrapping paper is used instead of a lap cloth.
The paper is cut to the size of the lap and evenly
smeared with a one-half-inch length of 3-micron
diamond abrasive and a small amount of polishing oil,
until it appears moist. Unlike a lap cloth, the kraft
paper minimizes plucking and subsequent scratching
of the sample and ultimately gives a much better
polish; in addition, it is inexpensive. The polishing
technique is described by Moreland (1968:2073).
However, it must be noted that the lap should turn
at 200 rpm. After the final polish, the wafer is
mounted with the polished face down on a glass slide,
using Araldite AY105 and Hardner 935 (Moreland
1968:2071).
An Ingram Model 103 thin section cut-off saw is
used to reduce the wafer to a thickness of 0.020 inches*
and an Ingram Model 303 thin section grinder with
a diamond wheel is utilized to still further reduce the
thickness. Additional grinding is done by hand, using
93
94
a slurry of 600 aluminum oxide and water on a 850
rpm horizontal lap, until the desired thickness is
reached. A standard thickness of 30 microns is ob-
tained using a glass plate with a slurry of 1500 alumi-
num oxide and water.
Final step in the preparation of the doubly polished
thin section is polishing of the wafer, which is de-
scribed by Moreland (1968).
Doubly polished thin sections of grain mounts are
prepared in essentially the same manner. Grains be-
tween 177 and 105 microns (80 to 150 mesh) are
sprinkled one layer thick onto a glass slide, which
has been smeared with a mounting medium of equal
portions of Araldite AY 105 and Hardner 935F and
heated to 200 ?F on a hot plate.
After curing, an Ingram thin section grinder is
used to expose the individual grains which are then
SMITHSONIAN CONTRIBUTIONS TO THE EARTH SCIENCES
ground to an optically flat surface and polished in the
same manner as a doubly polished thin section. A
0.015-inch diamond saw blade is used to remove all
outer regions of the glass slide devoid of grains; the
slide is then mounted, polished face down, on another
glass slide. Once the mounting medium has cured,
the sample is placed in an Ingram thin section cut-off
saw and the thickness of the slide containing the
unpolished surface of the grains is reduced so that the
grains are just barely exposed. The section is then
ground and polished to a standard thickness of 30
microns.
Literature Cited
Moreland, G.
1968. Preparation of Polished Thin Sections.
Mineralogist, 53:2070-2074.
American
U.S. GOVERNMENT PRINTING OFFICE: 1972 -484- 316/1B