CLAY MINERALS BULLETIN

JULY, 1960

Vol. 4, No. 23

C H A N G E S EFFECTED I N L A Y E R S I L I C A T E S B Y
H E A T I N G B E L O W 550~ *
By C. M. WARSHAW,P. E. ROSENBERG and R. RoY.
The Pennsylvania State University, University Park, Pa., U.S.A.
[Received 5th January, 1960]
ABSTRACT
The changes in X-ray diffraction patterns of layer silicates which result
from dry heat-treatments below 550~ have been studied and the results
are tabulated in order of increasing temperature at which these changes
are observed; the structural implications are briefly considered. In
order to evaluate the effects of composition, particle size and disorder,
synthetic clay minerals have been used in addition to natural clays.
This summary of X-ray data may be of value in the identification of clay
minerals in mixtures.
INTRODUCTION
It has been well known for many years that the different types of
clay minerals exhibit characteristic behaviour on heating. Earlier
work consisted for the most part of correlating the amounts of water
lost on ignition with the temperature, of determining the thermal
energy associated with the dehydration, and of identification of the
products formed at high temperatures. Much of this work is briefly
summarized by Grim (1953) and in the symposium edited by
Mackenzie 0957). All the reactions described in this paper refer to
" d r y " heating in air; even although a phase persists under these
conditions up to a certain temperature this cannot be used to indicate
its thermodynamic stability under the conditions. Thus, there is no
correspondence whatsoever between the temperature cited herein and
those obtained for the upper stability limits of clays under hydrothermal conditions. In genera[, the clays can be heated a few
hundred degrees above the true stability temperature of the phase in
air.
In the last decade much of the emphasis in the thermal investigation of layer silicates has been on determining the structural
changes during the loss of hydroxyl water--see, for example, Bradley
*Contribution No. 59-48, College of Mineral Industries, The Pennsylvania
State University.
113

114

c.M.

WARSHAW, P. E. ROSENBERG AND R. ROY

and Grim (1951) on montmorillonite, Brindley and Ali (1950) on

chlorites, Roy (1949) and Sundius and Bystrom (1953) on micas,
Walker (1956) on vermiculite. Recently, Brindley and Nakahira
(1959) have examined the structural changes in kaolinite and have
summarized all the significant previous research on this subject.
The changes in structure on heating are, of course, accompanied
by changes in the X-ray diffraction patterns which are characteristic
for the different mineral families, and which are used as a means of
identification of some of the clay minerals. However, Nelson and
Roy (1954) showed that certain variables other than temperature
must be considered if one is to use these heating experiments for
identification. Thus, in heating normal chlorites and 7 A minerals,
chemical composition, particle size, degree of crystallinity, and length
of heat treatment all affect the final product. At present, however,
the data which have been accumulated on the thermal behaviour of
different clay minerals are sufficient for difficulties not to be encountered when the chemical composition and degree of crystallinity
are taken into account. Moreover, if samples are heated for longer
times at lower temperatures than those suggested in the past, particle
size need not be considered. The principal difficulty in attempting
to utilize the information in the literature for identification by heat
treatment is that very few authors have reported the duration of their
he~ting experiments.
Although positive identification of a clay mineral can rarely be
made solely on the basis of its thermal behaviour, the changes
observed in X-ray diffraction patterns as a result of heat treatment
are frequently of considerable use as an auxiliary identification
method or in confirming an identification. For this reason, heat
treatments of clay minerals have been investigated as part of the
identification scheme presented in a recent review (Warshaw and Roy,
1960). As a result of these experiments, it has been possible to summarize the changes which occur on heating all the various layer
silicates with respect to temperature of heating. The previous compilations of heating data, on the other hand, deal separately with
each family of clay minerals. The summary in this paper should be
of considerable use to those investigators who are dealing with
assemblages of clay minerals.
The nomenclature employed in this paper is the same as used in
the above mentioned review article by Warshaw and Roy (1960) and
agrees, except in minor detail, with that used by Mackenzie (1959)
in his classification of the clay minerals.

EFFECT OF HEAT ON LAYER SILICATES

l 15

EXPERIMENTAL

Specimens Investigated. The behaviour of most of the natural

layer silicates upon heating is well established. Since many of these
vary considerably in chemical composition, particle size, and crystallinity, it was decided to study synthetic clay minerals (in which
these variables can be controlled) and their natural analogues to
determine the influence of these factors. The synthetic specimens
are all of very fine particle size regardless of their degree of crystaUinity; thus, a synthetic well-ordered kaolinite is much finer
than a natural one and may be comparable in particle size with
natural disordered kaolinite (kaolinited).
Over the past decade a large number of papers from this laboratory
have described the pressure-temperature conditions for synthesis,
the range of compositions and the properties of synthetic clay
minerals (see Warshaw and Roy, 1960). The methods of synthesis
are described in some detail by Koizumi and Roy (1959) and by
Warshaw (1960). For the synthetic clays used in the present investigation, the starting materials were gels of the desired composition. These were reacted in sealed gold tubes at the appropriate
temperature and at pressures of one to three thousand atmospheres.
Before use the products were checked for phases present and degree
of crystaUinity.
The synthetic minerals in Table 1 and the natural minerals in
Table 2 were used in this study. Various combinations of these
natural and synthetic clay minerals were also studied by X-ray
diffraction following various heat-treatments. Binary, ternary, and
a few quaternary mixtures were prepared by mixing together in a
mortar for a short period of time (about one minute) small weighed
portions of the separate phases. Of the many mixtures possible,
only a limited number were made, the combinations being~ for the
most part, those which are commonly found in nature. CombinaTABLE 1--Synthetic Minerals employed.
Mineral
Kaolinite
Septeclinochlore
Chrysotile
Beidellite
Saponite
Muscovite
Clinochlore

Composition
AI 4Si401 o(O H) s
MgsA1Si3A1010(OH) 8
Mg6SilOlo(OH) 8. . . . . . .
Nao. aaA| uSi a. 67AI o.33~ 10(UIT1)2
Nao-33MgzSia. 67A10.330 10(OH)2
KA12Si3A1Olo(OH)2
MgsAl Si3A 101 o(OH) s

116

C. M. WARSHAW, P. E. ROSENBERGAND R. ROY

TABLE2--Natural Minerals employed.
Mineral

Kaolinite
Kaolinited
Chlorite (septechlorite?, Fe-rich)
Montmorillonite (or beidellite?)
Montmorillonite (or beidellite?,
natural organic complex)
Montmorillonite (or beidellite?,
natural organic complex)
Illite
Morttmorillonite
Dioctahedral vermiculite
Vermiculite
Clinochlore
Chlorite (very small 14/~ peak)
Corrensite
Mixture of montmorillonite (or
beidellite?), mica (illite), kaolinite
(disordered?)
Dioctahedral chlorite
Corrensite (dioct.?)

Occurrence
Flint clay
Unknown
Green pellets from Tertiary sediment
Recent marine sediment
Recent fresh water sediment
Recent sediment
Uaderclay, Fithian, Ill.
Bentonile
Soil from Shenandoah Valley,
Virginia
Westtown, Pennsylvania
Westchester, Pennsylvania
amClay fraction of sandstone
b---Green pellets, Cretaceous sediment
Residue from acid-treatment of
limestone
Tertiary sediment
Clay fraction of sandstone
Clay fraction of sandstone

tions were selected with the following questions in mind: (a) Is

there good resolution of the basal reflections of more than one
mineral with the strongest diffraction peak near 7/k, and is there
more, or less, resolution after heat treatment below their decomposition temperatures? (b) Are spacing changes observed with
unresolved diffraction maxima when one component contributing to
these maxima is decomposed before another? (c) Do the intensity
changes which accompany changes in spacing upon heating render it
impossible to detect certain phases in mixtures? (d) Do possible
reactions between clays of different composition (e.g. between
high-A1 and high-Mg clay minerals) tend to promote changes at
lower temperatures?
Heating and Examination Procedure. The clay samples were
allowed to sediment from water on to glass slides, dried at room
temperature, and examined by X-rays using a Norelco Diffractometer. The slides in a special slide holder were heated at the desired
temperature for a period of 11-16 hours and then transferred to an
oven at 110~
Each slide was removed from the oven in turn and
again examined by X-rays. Rehydration while the slide was being
examined was prevented by covering the opening in the scatter shield

EFFECT OF HEAT ON LAYER SILICATES

117

with cellophane tape and keeping a dish of magnesium perchlorate

in the scatter shield.
With the stainless steel holder it was possible to heat up to sixteen
slides in the furnace at once while keeping all of them supported on a
flat surface to prevent warping.
Various temperatures between 270~ and 500~ were employed.
One set of slide preparations was heated at a series of temperatures
(e.g., periods of about 12 hours each at 400~ 450~ and 500~
while other slides of the same samples were heated directly at 500~
without prior heat-treatment. The lowest temperature of 270~
was selected because it is welt below temperatures at which most clay
minerals begin to dehydroxylate and it is high enough for interlayer
water to be essentially removed from smectites and for vermiculite to
undergo profound changes. Only reactions up to 500~ have been
considered, since these are usually sufficient to distinguish one type of
clay mineral from another. Moreover, ordinary glass slides can be
used up to this temperature. Some high-temperature reactions are
also useful in the identification of "pure" specimens, but it is not yet
known to what extent high-temperature (1000~ heat-treatments
can be used in the identification of clay minerals in mixtures, since
interaction may occur.
A heating period of 15 hours is in practice more convenient than
periods of one to two hours, since this is approximately the length
of time that samples can be heated unattended overnight.
RESULTS

The changes which occur in the diffraction patterns as a result of

heating at various temperatures for periods of 11-16 hours are
summarized in Table 3 where the minerals are grouped in order of
increasing temperature of the first observed change. Only the
changes which occur in the low-angle portions of the patterns (below
20~ 20 with CuK~t) are necessary in identification procedures; thus,
for simplicity, only these changes are noted, although others occur.
It was found that prior heating at one of the lower temperatures
makes little difference in the results obtained for any higher temperature.
DISCUSSION

Single Phases. The expanded 2:1 layer silicates or minerals containing some expandable layers (expandable clay minerals) exhibit
changes, involving only interlayer water, at the lowest temperatures
cited in Table 3. Smectites with essentially only monovalent and

"0

(j

~-~

..~2

~,

~o

.~

0~0

~==
0 0..

.o
o
g

~J

"~'
~

.=

~~
~.~.~

= ~ E
~ ~

-~'~=

o~=~o
~

o~=

~n

~ '~

.~:

b
-0
"0

.o

o'x

.'0

~j
&

"~o

-~.~

~=~J ~

~o~

<o ~

-o.s~

~o

=~

~-~

~,*-.~

0-~
0
~

.~

"0

"0

~o

~. ,~

,-~0

~..~

(j

g g
"0

o'Fn 0s

~J
g

~Z

~o
"~
0

~o
~0~
o o~..

_~

.~
~o~
0~.- o
o o~-~

~%~

r~,

"0

~J
0

>

0 ~,.~:~ 0

r~

~ ~'n.

,~'~.~
Q

~.~

~o~

._= ~,m.~.<
~'~ ~ Z
~"~

"~ 0
cj ~ ~ , m

~..~ o,=

.<

.~

~.~

o~

=_

.~'~
t2

0 ~ ' ~ c.,,,

"<o

I=
am

o . . ~.+.
m

o
Z

o'~
Z

c~

~J
._9
~
0
t2
0
r.a9

t~

120

C . M . WARSHAW, P. E. ROSENBERG AND R. ROY

divalent interlayer cations and containing no complexed organic

material collapse readily with a decrease in peak intensity (cf. Milne
and Warshaw, 1956, Fig. 2). In illites and related minerals and in
non-complexed montmorillonite in young sediments the collapse is
accompanied by a sharpening of the 001 reflection (cf. Milne and
Warshaw, 1956, Figs. 5 and 7, respectively). The natural montmorillonite-organic complex exhibits only a partial collapse at
270~ a temperature greater than 300~ (Milne and Shott, 1958,
Fig. 4) being necessary to effect the type of change shown by the
montmorillonite itself. When trivalent interlayer ions (+organic
material?) are present only a partial collapse can ever be obtained
before dehydroxylation begins. An example of this behaviour is
shown by dioctahedral vermiculite (Hathaway, 1955, Fig. 4), where
the gradual collapse of the basal spacing is accompanied by a
broadening of the 001 reflection.
When vermiculite is heated at low temperatures (e.g., 120~ a
collapse similar to that observed with montmorillonite occurs. The
very strong 14.3/~ reflection is replaced by a less intense reflection at
11-6 A. When heating is carried out at the temperatures included
in Table 3, however, the changes no longer involve simply the loss of
interlayer water. Walker (1956) has proposed that a regular mixedlayer phase is formed with a basal spacing of 20-6/~ (11.6 + 9.0 A),
the strongest reflection being at 10-2 A (002); the changes which
occur at increasingly higher temperatures reflect the gradual breakdown of this phase. Since the term vermiculite is used to cover a
wide variety of phases both with regard to composition and degree
of mixed-layering, or heteropolytypism (Weaver, 1958), the behaviour of various specimens on heating may vary considerably,
necessitating some differences in interpretation of the changes.
Thus, the sample of vermiculite examined here developed a peak at
8-7 A in addition to the one at 10-2 A, and a peak at 9.6 A replaced
the 10-2A peak after heating at 500~
Walker (1956) has mentioned only a spacing of 9.0 A in addition to the 20.6 A phase.
One feature of the dehydration of vermiculite which has not received
sufficient emphasis is the actual appearance of the diffractometer
traces of the heat-treated material; those obtained here are reproduced in Fig. 1.
Corrensite is a regular mixed-layer mineral (heteropolytype)
containing approximately equal numbers of chlorite and expanded
2:1 layers. Because of the presence of the latter, corrensite exhibits changes at low heating temperatures, which have been dis-

121

EFFECT OF H E A T O N LAYER SILICATES

cussed by Earley et al. (1956)and by Bradley and Weaver (1956).

The change in the X-ray pattern for this mineral heated for periods
of 11-16 hours at temperatures below 500~ is similar to that obtained by Bradley and Weaver, whose sample contained smaller
amounts of impurities than others which have been described.
The pattern they obtained after heating at 550~ for only two hours
corresponds to the patterns obtained here after heating for longer
times at lower temperatures. The change observed after heating at

3o~

co"

~o"

FIG. 1--X-ray diffractometer traces of vermiculite from Westtown,

Pennsylvania, before and after heat treatment: a--before heat treatment, b---after treatment at 120~ c--after heat treatment at 445~
d--after heat treatment at 500~ The same slide preparation was
used for all heat treatments (12-15 hours). Traces were obtained
with a Norelco Diffractometer using CuKa radiation, scanning speed
2~ 20 per minute, scale factor 64.

500~ for 15 hours reflects the chlorite nature of the mineral. The
interpretation of the basal spacings at low angles may be summarized
as in Table 4. The spacings in Table 4 do not agree precisely with
those obtained by Earley et al. (1956) and by Bradley and Weaver
(1956), but provide an illustration of the degree of precision which
can reasonably be expected in different laboratories working with
natural materials which are mixtures of similar phases.
In Table 3 dioctahedral chlorite has been grouped with the expanded 2:1 layer silicates or minerals containing expandable layers.

122

c. M. WARSHAW,P. E. ROSENBERGAND R. ROY

The exact nature of this mineral has not yet been established, but the
fact that its 001 reflection shows some collapse suggests that it may
contain some vermiculite-type layers.
Since most of the 7 A minerals exhibit changes in their X-ray
patterns as a result of long periods of heating in the temperature
range 400-500~
they can be grouped together. The changes
observed are due to the loss of at least some of the hydroxyl water.
The minerals examined are listed in order of increasing temperature
of the beginning of this water loss.
TABLE4--Basal reflections of corrensite before and after heat treatment.
001

d (A)

Components

Room temperature
001
29.4
002
14-5
003
9.8
004
7'25
006
4"85
Average basal spacing 29"2/~

w
s
w
m
m

14.3A
+
chlorite

14-9A
saponite

Dehydrated below 500~

002
12'3
003
8.1
005
4-85
Average basal spacing 24'3/~

m b
m b
m

14.3/~
chlorite

-k

10-0A
saponite

13-8/~ +
chlorite

10-0/~
saponite

Heated at 500~
002

11.9

s b

The iron-rich sedimentary chlorite which occurs as green pellets

(somewhat similar to glauconite in appearance) exhibits some change
at considerably lower temperatures than do the other minerals of
this group; this is probably related to the oxidation of the iron as
well as to dehydration.
Sedimentary chlorites, even those which are or which contain
normal chlorite, begin to decompose at a lower temperature than
natural well-crystallized kaolinite. In general, it can be said that
sedimentary chlorite and kaolinite show decreases in the intensity

EFFECT OF HEAT ON LAYER SILICATES

123

of the 7/~ and 3.5/~ basal reflections in the same temperature

range. Thus, loss of these reflections as a result of heat-treatment
cannot be used to distinguish these minerals in sediments. However,
heat-treatment is useful if the 13-14A region is examined. The
14 A peak of normal chlorites increases in intensity as a result of
heat-treatment, for about 15 hours at temperatures above 500~ *
It has been reported by Nelson and Roy (1954) that all the septechlorites they examined developed a peak in this region as a result
of high-temperature heat treatment; such behaviour is also exhibited by dickite (Hill, 1955) but not by kaolinite. The septechlorites examined in this study (synthetic, iron-free) did not develop the
14 A peak on heat-treatment (which was probably not high enough
in temperature) but previous experience with sediments suggests that
most sedimentary chlorites would show this peak after heat-treatment
at 500~
The spacing of the 14 A peak of the sedimentary chlorite
investigated decreased from 14.2/~ to 13.8/~ with the increase in
intensity. Characteristically, the 14A peaks developed by septechlorites are actually broad peaks centered at about 13.5 A.
Chrysotile, which has the highest thermal stability within the
7 A group, shows only slight changes at 500~ and actually has
higher thermal stability than normal Mg-chlorites, which dehydroxylate in two stages.
Only two minerals were investigated in the group with high
thermal stability. All the 2:1 layer silicates and normal chlorites
(which contain 2:1 layers) belong to this group (the smectites
can be included once the interlayer water is removed.) Some
minor changes in the spacings and intensities of the basal reflections
may be observed when some of these minerals, especially the
dioctahedral ones, are heated for a prolonged period at 500~
but these can be correlated with the loss of a considerable portion of
the hydroxyl water to yield dehydroxylated forms (Bradley and Grim,
1951) which retain the layer structure. Thus the basal reflections
instead of being destroyed, as with kaolinite, undergo very little
change. The talc layers of the normal Mg-rich chlorites are not
affected by heat-treatment at 500~ but the brucite layers are dehydroxylated at this or a slightly higher temperature, producing the
change in X-ray pattern which is characteristic of chlorite, namely
the increase in intensity of the 14A reflection.
*This temperatureis high enoughfor sedimentarychlorites, althougha slightly
nigher temperature may be needed for Fe-freechlorites.

124

C. M. WARSHAW, P. E. ROSENBERG AND R. ROY

Mixtures. Results for the mixtures studied showed that the

characteristic thermal behaviour of the clay minerals discussed is
observed whether they occur singly or in mixture. That is, there is
little interaction between clays in 11-16 hours at the temperatures
employed. Micas, illites, and most smectites present no difficulties
in identification as components of a mixture, but some difficulties are
encountered with mixtures of 7 A minerals, of chlorites and vermiculite, and of mixtures of all of these minerals.
The 3.5 A reflection of a mixture of kaolinite and sedimentary
chlorite (with a negligible 14A peak) may be resolved into two
peaks, but frequently only one rather broad peak is observed at
25.0 ~ 20 (CuKa radiation). Increase in size or development of a
13-14A peak upon heat treatment at 475-500~ indicates that
sedimentary chlorite is present, but the question remains as to
whether kaolinite is also present. It is occasionally possible to
detect a shift in the 25.0 ~ peak to 25.2 ~ (chlorite) or 24.9 ~ (kaolinite)
after heat treatment at 400~ revealing that one component has lost
intensity relative to the other (see Bradley, 1954). If this does
not occur it may be necessary to use other methods, e.g., acid treatment or high-temperature heat-treatment, (see Warshaw and Roy,
1960) to detect kaolinite in mixture with sedimentary chlorite.
Mixtures of chlorite (usually the normal variety) with vermiculite
and/or corrensite may also present some difficulties. Unless large
amounts of the latter two minerals are present the broad peaks or
bands which develop upon heat-treatment may not be detectable.
However, a decrease in intensity of the 14 A reflection relative to
that of the 7 A reflection as a result of heating at 375~ indicates the
presence of vermiculite and/or corrensite in addition to chlorite.
Large amounts of vermiculite or corrensite in a mixture can be
distinguished by their characteristic broad peaks or bands (see
Table 3). It should be noted, however, that the presence of corrensite may be confirmed by the low-angle peak (29 A) observed
at room temperature. If sufficient corrensite is present to be detectable by heat-treatment, then there is also enough to yield the
low-angle peak when the mixtme is examined by X-rays before heattreatment.
The greatest difficulty is encountered with mixtures containing
normal chlorite and septechlorite and/or kaolinite. The kaolinite
may be detected by its greater resistance to acid, or by the development of mullite when the mixture is heated for a short time (1-2
hours) at 1000-1100~
Septechlorites may be separated from nor

EFFECT OF HEAT ON LAYER SILICATES

125

real chlorites by the method described by Brindley, Oughton and

Youell (1951), i.e., the mixture is heated at a temperature high
enough to dehydroxylate the brucite layer of the normal chlorite,
but not to decompose the septechlorite, and the decomposed normal
chlorite dissolved out with acid. This method has not, however,
been adequately tested to confirm its universal applicability to
chlorite mixtures.
Acknowledgments.--The authors are grateful to the following for donations of
natural clays: Professor G. W. Brindley of the Pennsylvania Slate University,
Dr I. H. Milne and Dr J. W. Farley of the Gulf Research and Development
Company, Dr C. E. Weaver and Dr J. F. Burst of the Shell Development Company, and Dr J. C. Hathaway and Dr L. G. Schultz of the U.S. Geological Survey.
This work was supported by Grant C-294 from the Petroleum Research Fund
administered by the American Chemical Society, and grateful acknowledgment
is hereby made by the authors.
REFERENCES
BRAOt_~Y,W. F., 1954. Clays and Clay Minerals (A. Swineford and N. Plummer,
editors). Nat. Acad. Sci.--Nat. Res. Counc., Washington. Publ. 327, p. 324.
BRAOLEY,W. F., and GRIM, R. E., 1951. Amer. Min., 36, 182.
BRADLEY,W. F., and WEAVER,C. E., 1956. Amer. Min., 41,497.
BRINDLEY, G. W., and Au, S. Z., 1950. Acta cryst., 3, 25.
BmND1.EY, G. W., and NAKAHIRA,M., 1959. J. Amer. ceram. Soc., 42, 311, 314,
319.
BRINDLEY,G. W., OUGHTON,B. M., and YOOELI,R. F., 1951. Acta cryst., 4, 552.
EARLEY,J. W., BRINOt.EY,G. W., MCVEAGH,W. J., and VAND~NHEUVEL,R. C.,
1956. Amer. Min., 41,258.
GRIM, R. E., 1953. Clay Mineralogy. McGraw-Hill, New York.
HATaAWAY, J. C., 1955. Clays and Clay Minerals (W. O. Milligan, editor).
Nat. Aead. Sci.--Nat. Res. Counc., Washington. Publ. 395, p. 74.
HILL, R. D., 1955. Acta cryst., 8, 120.
KOIZtIML M., and RoY, R., 1959. Amer. Min., 44, 788.
MACKENZIE, R. C. (editor), 1957. The Differential Thermal Investigation of
Clays. Mineralogical Society, London.
MACKENZIE,R. C., 1959. Slav Min. Bull., 4, 62.
MILNE, I. H., and SrIOTT,W. L., 1958. Clays and Cla~ Minerals (A. Swineford,
editor). Nat. Acad. Sci.--Nat. Res. Counc., Washington. Publ. 566, p. 253.
MILNE, I. H., and WARSHAW,C. M., 1956. Clays and Clay Minerals (A. Swineford, editor). Nat. Acad. Sci.--Nat. Res. Counc., Washington.
Publ. 456,
p. 22.
NELSON,B. W., and RoY, R., 1954. Clays and Clay Minerals (A. Swineford and
N. Plummer, editors). Nat. Acad. Sci.--Nat. Res. Counc., Washington.
Publ. 327, p. 335.

126

c . M . WARSHAW, P. E. ROSENBERG AND R. ROY

RoY, R., 1949. J. Amer. ceram. Soc., 32, 202.

SUNDIUS,N., and BYSTROM,A. i . , 1953. Trans. Brit. Ceram. Soc., 52, 632.
WALKER,(~. F., 1956. Clays and Clay Minerals (A. Swineford, editor). Nat.
Acad. Sci.~Nat. Res. Counc., Washington. Publ. 456, p. 101.
WAR.SHAW,C. i . , 1960. Clays and Clay Minerals, Proceedings of the Seventh
National Clay Conference (Ada Swineford, editor). Pergamon Press, London,
p. 303.
WARSHAW,C. M., and RoY R., 1960. In press.
WEAVER,C. E., 1958. Amer. Min., 43, 839.