THE AMERICAN MINERAI,OGIST, VOL.

53, JANUARY_FEBRUARY, 1968

CHEMICAL DIFFERENCES AMONG THE SERPENTINE

,.POLYMORPHS''1

Nonuex J Pacn, LI.5. GeologicolSuruey, Menlo Park,

California

Arsrrlcr
Existing wet chemical analyses and structural studies of the serpentine minerals, chryso-
tile, Iizardite, and antigorite suggest that these minerals have different chemical composi-
tions and could be identified by composition alone. Triangular composition diagrams, a
statistical treatment of the oxide components, MgO, FeO, FezOa,AIzOa, and HsO, and cal-
culated mineral formulae from chemical analyses all suggest that chrysotile, iizardite, and
antigorite are not polymorphs. Antigorite is distinguished by its comparatively low H:O
and high SiO2 contents. Chrysotile is characterized by a relatively high HrO and MgO con-
tent and by a small ratio of FezOato FeO; while lizardite has hieh SiOz and low FeO con-
tents.

INrnonucrroN
Chrysotile, lizardite, and antigorite compose the serpentine-group
minerals. Much discussionhas been centered around the nature of the
relations among these mineral speciesand the ultimate question: Are the
serpentinespolymorphs or are they separate but similar mineral species
which can be distinguished by chemical analyses?Though the answer to
this question is controversial at present, it is important for the interpre-
tation of alteration phenomena of ultramafic rocks. Although past chem-
ical and crystallographic studies indicate that the serpentine-group
minerals have essentially the same structural foundation and similar
formulae, the many existing chemical analyses and structural studies
have not been consolidated into a cohesiveinterpretation. These results
are synthesizedin this paper to show that the serpentine-groupminerals,
based on the available analyses,are either members of a complex solid-
solution seriesor are separate chemical species,but are not simply poly-
morphs. A polymorph is a substance that crystallizes in more than one
form or crystal structure, all forms or structures having identical chemi-
cal compositions.
Unfortunately, in the past, a large number of names have been applied
to individual samples of serpentine-group minerals, generally based on
textures and appearanceof the specimens.Faust and Fahey (1962) pro-
vide a comprehensivediscussionof this aspect so that it need not be con-
sidered here. The best classification of the minerals in the serpentine
group is that of Whittaker and Zussman (1956) based on the X-ray
diffraction results. This classification is used here as given below:

I Publication authorized by the Director, U. S. Geological Survey.

201
202 NORMAN J PAGE

The present study was begun at the University of California where

most of the work was done; researchhas been continued at the U.S.
Geological Survey.

Srnucrunn oF TEE SBnpBNrtNo Gnoup MrNBnars

Although other structures for serpentines have been proposed, the
evidencepresented by Warren and Hering (194I), Aruja (1945), and
Whittaker (1953) demonstratesthat all serpentineminerals, including
the fibrous varieties, have a trioctahedral structure analogousto diocta-
hedral kaolinite. As discussedby Bates (1959) and Deer, Howie, and
Zussman (1962), a major factor repsonsiblefor the different serpentine
minerals is the amount of mis-matching involved in joining a brucite-
Iike layer to a tridymite-like layer to form an unit cell of serpentine. At
Ieast four possibilities for deriving such structures have been postulated:
(1) curving the composite sheets with the tridymite layer forming the
concavepart of the cylinder, (2) distorting one or both layers, (3) sub-
stituting various ions in either layer to adjust the sizes of the layers,
and (4) creating cation vacanciesin the octahedral layer and maintaining
charge balance by removing an equivalent amount of hydroxyl ions. The
structure of chrysotileis controlledby (1) and (3) with (1) the dominant
mean of matching the sheets(Whittaker, 1956a).Method (4) combined
with (1) and (3) is responsiblefor the structure of antigorite (Kunze,
1961). Another factor in the development of the different structures of
serpentine minerals is the possibility of both regular and irregular stack-
ing of the layers.

Chrysotile. Fankuchen and Schneider (1944) suggested,on the basis of

small angle X-ray scattering, that chrysotile consists of fibers on the
order of 200 A in diameter packed in trigonal array along the fibers.
Since Turkevich and Hiller (1949) first used an electron microscope to
demonstrate that chrysotile apparently had a "hollow tube" morphology,
abundant evidencebased on the electron microscopeobservationshas
accumulated (Bates, Sand, and Mink, 1950; NolI and Kircher, 1950,
1951; and Maser, Rice, and Klug, 1960). These observationsare sup-
ported by the detailed structural studies of Whittaker (1957). In a re-
cent paper Huggins and Shell (1965)reported bulk densitiesof chrysotile
ranging ftom 2.2 to 2.4 g/cc. Using cylinder diametersof 340 A and 80 A
as outer and inner diameters, respectively, the theoretical bulk density,
 SERPI]NTIND "POLYMORPHS" 203

including pore spaceis 2.19 to 2.25g/cc. This demonstratesthat chryso-

tile has hollow or partially filled hollow tubes.
Three varieties of chrvsotile are recognized from single-.frberstudies:
clinochrysotile, orthochrysotile, and parachrysotile (Whittaker, 1953,
1955, 1956a,b,c,and 1957).Ortho- and parachrysotileare orthorhombic
while clinochrysotileis monoclinic (neglectingthe curvature). In clino-
and orthochrysotile, the fibers are curved about the o-axes,and any
regular stacking or translational-ordered arrangements would be in a
la or -a direction. By the nature of the fiber, translational order cannot
occur in the 6 direction, so that the atoms at the base of layers would
have to be restricted to circumferential grooves. Clinochrysotile (Whit-
taker, 1956a) has a two-layered cell in which the successivesheets are
offset in the same direction (either -la or -o), with alternate layers
offset by different amounts. Orthochrysotile (Whittaker, 1956b) has a
two-layered cell and has an orthorhombic cylindrical lattice. Alternate
Iayers are offset in the la and -a direction. The third variety, para-
chrysotile, is defined with 6 as the axis of curvature and has an ortho-
rhombic cylindrical lattice. Whittaker (1956c) shows there can be no
analogouspartner to parachrysotileas there is for clino- and orthochryso-
tile.
From Whittaker's study (1956a,b,c) and Whittaker and Zussman
(1956), it appears that clinochrysotile is the most abundant, ortho-
chrysotile is next most abundant, and parachrysotile is fairly rare or
mixed with the other two in small amounts.

Lizardite. Observationsunder the electron microscopeshow that lizardite

has a platy morphology (Deer, Howie, and Zussman, t962). Early
powder work by Whittaker and Zussman (1956) on samplesdescribed
but not named by Midgely (1951),suggestedthat,lizardite has a single-
Iayered unit cell and is pseudo-orthohexagonal,individual layers having
trigonal symmetry. This was confirmed by the single-crystal analysis of
Rucklidge and Zussman (1965). They found thatlizadite is disordered
in three ways: (1) the crystalsare macroscopicallybent about more than
one crystallographicaxis; (2) some layers are rotated by 1800;and (3)
some layers are displaced by *(b/3). There are two possibleorienta-
tions of alizardite layer which can be combined in three ways to produce
stacking arrangements in an individual crystal.

Antigorite. The structure of antigorite, the other distinct serpentine min-

eral, has been elucidatedby Aruja (1945),Zussman(1954, 1956),Kunze
(1956, 1957, 1958, 1961), Zussman, Brindley and Comer (1957) and
Chapman and Zussman(1959).Under the electronmicroscopeantigorite
204 NORMAN J PAGT,

exhibits both a platy and fibrous morphology, but there are no sugges-
tions of cylindrical tubes. Although antigorite is a single-layeredserpen-
tine whose c and b dimensions are very simiiar to lizardite, it's o param-
eter is large and in many casesapproximates40 A. On the basis of the
single-crystalstudies of Kunze (1956, 1958, 1961) and Zussman (1956),
the structure involves deficiencies of hydroxyl and magnesium ions.
These deficienciesare repeatedalong the d axes.The resulting structure
is then composed of "half-waves joined by means of Mg-bridges"
(Kunze, 1956).
The most frequently occurring a parameters for antigorites studied
by electrondiffraction (Zussman,Brindley, Comer, 1957; Chapman and
Zussman,1959)are 33.7,35.8, 4t.2, and 43.0 A. Zussmanand co-workers
found that different grains of one sample yielded values oI a of 41, 90,
and 110 A, while other fibrous antigoriteshad o parametersin the range
of 16 to 10 A (Chapman and Zussman,1959).Kunze (1961) has treated
the variation of o parameters from a structural and theoretical approach
and considers that antigorites have compositions between talc and
chrysotile which accountsfor the octahedralcation and (OH)- deficien-
cies.From his structural approach,the various observedvalues of the a
parameter can then be derived theoretically.

CnBlrrcar CoMposrtroNs ol rHE SBnpnNrtNB MrNBner-s

Althouth the formula normally quoted in mineralogical textbooks for
serpentineis MgaSizOr(OH)a, or two times this formula, natural serpen-
tines are rarely, if ever, composedonly of magnesia,silica, and water. In
the generalformula X6Y4O10(OH)s, X representsthe possibility for the
substitution of the ions Mg2+, Fe2+,p.a+, Co'+, lrllz+,n{n2+,Mn3+, Cr3+,
Cu2+,Al3+,and Tia+ in octahedral coordinationl Z representsSia+,AI3+,
; n d O H r e p r e s e n t s( O H ) - ,
B t + ( ? ) , a n d F e 3 +i n t e t r a h e d r a lc o o r d i n a t i o n a
Cl-,F-(?), and Br-(?).

Criteria and selectionof analyses.In addition to the ions mentioned above,

serpentineanalysesfrequently show some K1+, Nar+, and Ca2+,that are
probably not in the structure but are due to a mineral impurity in the
analyzed concentrate. Even the exact amounts of Fe2+,ps3+, Q13+,fia+,
and AI3+in the structure may not always be known becauseof the fine-
grain size of the serpentine and its intimate association with fine-
grained spinels or hydrated oxides, such as goethite or lepidocrocite,
which makes physical separation of the minerals almost impossible. It is
also suspectedthat in somecasesthe amount of magnesiareported is in-
correct and is really due to the mineral impurity brucite. In considering
the analytical results, both structural state and purity require careful
 SERPENTINE I'POLYMORPHS" 205

attention and definition. Analyses of serpentines for which there is not

sufficient X.:.ay, optical and textural, DTA and infrared evidence to de-
termine the mineral speciesare not consideredhere.
The group of analysesconsideredhere was selectedfrom 248 analyses
of serpentine mineral done between 1834 and 1962as compiled by Faust
and Fahey (1962).Their discussion(p. 7,79) ol the sourcesof the analy-
ses indicates that the literature was well covered and that a wide as-
sortment of types was selectedby them. For the presentstudy, analyses
(Appendix) made on material apparently free of mineralogical impurities
and identified as to mineral type, were chosen.Five new analyses(Page,
1967) were added to this group. Comments concerning each analysis
used are given in the Appendix; these comments indicate why every
analysiswas not used in each of the following figures.All of the analyses
were recalculated by computerl to mineral formulae, using the hydro-
gen-equivalentmethod. Alkalis and HzO(-) were rejected for the pur-
poses of the calculations.

Composit'ionalvariations of serpent,ine.In Figure 1, the variation in the

composition of serpentine is represented as weight percent of a specific
oxide versus the normalized number of analyses. The five major com-
ponents, MgO, SiOz, FeO, FerOa; and AlzOg, are the only oxides con-
sidered.Water is discussedlater. The number of analysesout of the 86
for which the amount of the oxidesis available,is indicated in the figure.
The frequency diagrams show that serpentines have a fairly limited
range of composition.An inspection of the percentageof sample versus
weight percent of oxides indicates little chemical difference among the
various species.The analyseswere recalculatedto total weight percen-
tages of the sets of componentsMgO, SiOz,HzO; MgO, FeO, SiO2;and
MgO, FeO, H2O. The amount of total iron was calculated as FeO and
only HrO(f ) was used for the HzO component except where only total
water was determined. Figure 2 shows the recalculated analysesfor
chrysotiles, lizardites, and antigorites.
On the MgO, SiO2,HzO diagram (Fig, 2a),lizardite and chrysotile
compositions do not differ significantly. Antigorites tend to be higher in
SiOz and lower in HzO and MgO than lizardite and chrysotile but some
overlap of the rangesexistswhich could be causedby errorsin separation,
identification, or analysis. Lizardites and antigorites tend to have higher
total iron contents than chrysotilesas is demonstratedin the diagram
for MgO, FeO, SiO2(Fig. 2b). Again, antigoritesare grouped separately
from lizardites.In the compositionaldiagram for MgO, FeO, H2O (Fig.

I
Jackson (in press) designed the computer program.
206 NORMAN T PAGT,:

30 40
86 serpentines 70 serp- serpelF
e n ti n e s t ines
o o o

4 e

G 5zo G

lso is \s
42 44 46
o / o ' Si 0 2
weight % Fe203
024
neight % Al20a

cnrys0- 43 chryso-
30 t i les t i les
57 chrysotiles

o o
:20t
9l
of G @
ol
I
is | -f \e is
0Eu 0
0 < 3I 10 12 44 46 24 0 )246

r e i g ht % s i 0 2 weight % A1203 weight % Fer0t

22 ant igor ites l9 ant igor- 16 antigorites

ites
o?
:20
e

\s \s
0 ^o
0{ 3 8 40 42 44 46 024
||eight % si02 weight % Al 203 weight % F e 20 3

I izat d-
lizard- ites
rtes
6 lizardites

o o
e

G a

I
:s rs

24
w ei g h t % S i 0 2 w ei g h t 0l F e ro ,
weight % Al20g

(a) (b) (c/

 SERP EN TI N E " POLY MORPH S" 207

s e r p e n ti n e s 8 6 s e r p e n ti n e s
62
o o
e
E
G G

\s so
2468 36 40 44

l{eight % Fe0 w ei g h t % t i l g 0

chrysoti les 57 chrysotiles

@ o2
e
E

\R \s

weighl % Fe0 weight % llgg

2l antigorites 22 antigor ites

o
B e
E
e G

\s \R
0l , 0
0 248 3ie oP/, oo
weight % Fe0 ignt nro
toI
n
ll
6 lizardites
6 lizardites

o o

e
E
o

\R iS

0 2468

wci3ht % Fe0 w ei g h t % i l g 0

(d) (e)

Frc. 1. Histograms for each serpentine mineral group showing the variation in (a) SiOz,
(b) AhOr, (c) Fe:O:, (d) FeO, (e) MeO weight percent.
208 NORMAN J PAGE

I
o
- T?
a

o
-

=l a-

bo

Y+U
HM
o
.I ^cd
anE

t3:
o
-{ :o
t I
E
I
6j
o
d^
-;
'-:
h .ii'

- 5
botr
OO dtr
oN
r-
cd
^@

i;i
OF

o x.o
6-m
ro : xii
e
d
?a
- o F^
60
G
a -

"T
cd
I
I
I
I q
<l
!

o
a
I

o a
- N
 SERPENTINE "POLYMORPHS" 2W

2c),lizardite and antigorite analysesplot differently than the chrysotile

analysessince they have higher total iron content than chrysotiles. The
lizardite and antigorite analyses tend to divide in two groups in which
antigorite is distinguishedby having a lower water content.
Other methods of plotting mineral analyses make use of calculated
mineral formulae. In this study, the hydrogen equivalent method was
used to calculate formulae. It is based on the hydrogen-equivalents,
that is, on the sum of the anions, O2-, (OH)-, and F- equals 18. This
method for finding the number of ions per formula involves calculating
ratios of hydrogen equivalents of oxide weight percents and normalizing
these values to twice the sum of anions per unit formula. It is preferred
becauseit usesthe HzO(*) of an analysis.Mineral formulae for the ser-
pentine minerals can also be calculated on the basis of 28 negative
chargesand assuming that the OH content of the cell is 8 as was done by
Faust and Fahey (1962). Originally Page (1966) used this method and
came to the same general conclusions regarding chemical compositions
and polymorphism, although the number of cations per formulae were
different. Since this method neglects the water reported in the analysis,
the decisionwas made not to waste this information in this report.
Figure 3 demonstrates that antigorite and lizardite contain less water
than chrysotile and that lizardite and antigorite differ chemically from
chrysotile by havinglower ratios of (Fe2+f Mg'+) to (Fe3+f Als+).

Statistical analys'isof serpent'inem'ineral'compositions.Given a superior

analysis of a serpentine mineral the structural type of the mineral can
usually be ascertained.Analysesmarked with an asterisk in the Appendix
were selectedfor a statistical treatment. The 52 analyses were assigned
to one of three groups, lizardite, antigorite, or chrysotile, by X-ray, DTA,
or other information, but not on the basis of the chemicalanalysis.The
weight percent oxides for MgO, FeO, Fe2Os,AlzOa,SiOz and HrO(*)
were treated as variables.These data were analyzedwith a computer by
a linear discriminant analysis(Program BMD05M, Discriminant Analy-
sis for Several Groups, available from the Stanford Computation Cen-
ter). The program treats the data (Anderson, 1958) by first comparing
the way in which each variable (oxide weight percent) varies within a
group and the way in which each variable varies with respect to the en-
tire set of observations (52 analysesof serpentineminerals). The dif-
ferences in the way in which the group and set vary are used to deter-
mine linear equations for the groups. Once these equations are deter-
mined they are evaluated for each analysisor observation.Using these
functional values for each analysis, a probability is calculated and is used
to indicate the likelihood that each observation belongs to any given
 NORMAN ] PAGE

X
X
X
-- X
X
+
+o X
r

\
10 o
X
+4
- X
+ XX
x X
r
q
x
aa

o
a a
a oo
a

l0
I
8
7
6
3
x C h r y s o ti l e
o Lizardite
o Antigor ite

15.0 16.0
l fe i g h t % H20

l-rc. 3. (Fe2+f Mg,+)/(Fe3+{A13+) versus weight percent H:O for serpentines.

group. Also statistics are calculated whi-ch allow one te test whether
arithmetic means of the three groups are the same.
The evaluation of the statistical analysis for each individual is that out
of 31 chrysotiles (originally grouped together) only two were included
 ''POLYMORPHS'' 211
SERPENTINE

Taelr 1. AvnnLcn SonprNttNr CouposrrroNs exn Cnr.cur,emo Mnwn.lr- Fonuur-,tn

Besoo on rnu HronocnN Eeury.q.lrNr MntIton. Ats-vsns Usno Ap.r Irtorceren
nv .tn AsrrnrsK rN Tr{E APPnuorx

Average weight percent

Chrysotile Lizardite Antigorite

SiOs 41.530 41.024 42 136

AlzOs 0.716 1.395 1.639
Fezoa 0.718 4. 100 I .165
FeO o.624 o 419 3.729
Mso 40.928 39.437 38.369
HzO* t3 542 13.286 12.098

t 98.058" 99.66r 99.136

FormuJae
Telrahedral ions

si4+ 3.897 3 .834 4.010

Al3+ 0.079 0 .1 5 4

t 3.976 3 .988 4.010

Octahed.ralions
A[3+ 0 .1 8 4
F-eBr 0.051 0.288 0.083
re'' 0.049 0.033 0.297
Mgz+ 5.724 5.494 5.43

I 5.824 5. 8 1 5 6.007
Ani.ons
o2- 9.524 9.7t7 10.320
(oH)- 8 476 8.283 7.680

! chargeson anions 27.524 2 7. 7t 4 28.320

! chargeson cations 27.524 27.716 28.32r

u Average HrO (-) for chrysotile is 1.25 weight percent.

in other groups than chrysotile, for six lizardites only one was put in
another group and for 15 antigoritesonly one was put in another group.
If the set of analysesused is representative,then the structural type of
serpentine mineral can be found from the mineral analysis. In Table 1
the means of the analysesfor each of the three groups is presented.The
statistical analysisindicate that the means are not the same at greater
than 99.95percent confidencelevel.
2I2 NORMANT PAGE

Besides this type of treatment, all of the analyses were presented

randomly to the same statistical program and three groups defined by a
random process.These three groups were nearly identical to the groups
chosenon a non-chemicalbasis.Therefore,very little bias was probably
introduced by the method of selectionof the groups.

Rrsur,ts AND CoNCLUSToNS

The chemicaldifferencesbetween the different serpentineminerals as
representedby the present available analysesare shown in Table 1 by
averageanalysesand averageformulaeof chrysotile,Iizardite,and antig-
orite. The featuresobservedhere which differentiatethe mineral species
are the same discussedin the previous section.On the averagethey are:
(1) high SiOzweight percent for lizardite and antigorite, (2) low AlzOe
contentsfor lizardite and chrysotile, (3) a large ratio of FezOsto FeO for
lizardite, (4) low MgO weight percent for antigorite, (5) low HzO weight
percent for antigorite, (6) large numbers of trivalent ions in the tetra-
hedral coordinationfor antigorites,and (7) low Fe2+in lizardites.
Structural studiesof the serpentineminerals,chrysotile,lizardite, and
antigorite indicate that the antigorite structure involves deficienciesof
hydroxyl and magnesium ions whereas the other two mineral structures
do not involve these. If mineral formulae are calculated on the assump-
tion of 28 negative charges (Faust and Fahey, 1962), the average sum
of octahedral cations for the selectedchrysotilesis 5.97;Lizardite,S.92;
and antigorite, 5.65. This is the reverse of Table 1, but the (OH)- for
antigorite in Table 1 is quite low, 7.680as comparedwith the other two
species.The actual distribution of cations and anions is probably some-
where in betweenthesetwo results.Sincethe structure changesthe com-
position of antigorite from the theoretical end member MgrSizOr(OH)a,
antigorite, as is now known, is not a polymorph of chrysotile or lizardite.
Recent microprobe studies by Page (1967) suggestthat lizardite and
chrysotile are separate mineral speciesand that lizardite is consistently
more iron-rich than the associatedchrysotile.Presentanalytical informa-
tion on serpentinesand statistical treatments of the same data suggest
that lizardite and chrysotile are not polvmorphs. These arguments apply
only to the presently available analytical information and as yet no one
has managed to obtain an unquestionedserpentinepolymorphic transi-
tion experimentally(Page, 1966).
This author believesthat all serpentineminerals are composedof three
sets of solid solutions, one representing lizardite, another chrysotile,
and another antigorite, with the possibility of polymorphism or poly-
typism within each solid solution series.
 SERPDNT I N E " POLYMORPH S'' 2t3

AcxrowtnlcltnNts

Special thanks are due to ProfessorsW. S. Iyfe, A. Pabst, and H. Hawkes, of the Uni-
versity of California, Berkeley. The University also provided part of the financial support,
which is gratefuily acknowledged. R. G. Coleman, R. W. Bowen, C. L. Christ, and P B.
Hostetler, of the U. S. Geological Survey gave generously of their time and offered many
helpful suggestions during the research and preparation of the manuscript.

RnlBnnNcns

ANoensoN, T. W. (1958) Introduction to multiztariate statistical anolysis. John Wiley and

Sons. Inc.. New York.
Anuya, E. (1945) An X-ray study of the crystal structure of antigorite Mineral. Mag.,27,
o.)-,/4.
Blrns, T. F. (1959) Morphology and crystal chemistry of 1:1 layer lattice silicates.,4raer.
M'inerol'.44,78-114.
---, SaNr, L. B. exo J. F. MrNx (1950) Tubular crystals of chrysotile asbestos. Science
rrr, 512-513.
Cnaeu.tx, J. A ,lNo J. ZussnaN (1959) Further electron optical observations on crystals of
antigorite. Acta Crystallogr.12, 550 552.
Dnrn, W. A., R A. Howrn, AND J. ZussMAN (1962) Roek-lormi.ng mineral; V 2 Chain
Si,l,icates:V. 3 Skeet Silicates. John Wiley and Sons, Inc , New York.
l'awxucnrx, I. eNn M. ScnNnmen (1944) Low angle X-ray scattering from chrysotiles.
J . Amer . Chem.Soc.66, 500.
Fausr, G. T. eNn J. J. Farrnl (1962) The serpentine group minerals. U. S. GeotrSura. Prof .
Pap.38SA,92p.
Huccrxs, C W. eNo H. R. Snnrr. (1965) Density of bulk crysotile and massive serpentine.
Am,er.Mineral 50, 1058-1067.
J,r.cr<son,E. D , Bowrw, R. W exn R. E. SrnvaNs (1967) A computer-based procedure
for reduction for mineral analyses to mineral formulas. U. S. GeoI Sunt. Prof . Pap.
575-C, C23 C31.
KuNzn, G (1956) Die gewellte Struktur des Antigorits, T. Z. Krirtal,Iogr. 108, 82-107.
-- (1957) Die gewellte Struktur des Antigorits. Fortsch.Mi,nual 35,4G48
- (1958) Die gewellte Strukture des Antigorits, lL Z. Krislallogr. 71O,282.
-- (1961) Antigorit, Sturkturtheoretische Grundlagen und ihre praktische Bedeutung
fur die weitere Serpentin-Iorschung Forstchr. Mineral. 39, no. 2,2OG324.
Lurnnorcn, Haxs (1913) Serpentin in C. Doelter Handbuch d.erMina'alchemle. Dresden
und Leipzig, 2, no. 1,385-424.
Masrn, M., Rrcn, R. V. .q.NnH. P. Kr-uc (1960) Chryostile morphology Amu Minerol.45,
680 688.
Mrocr-nv, H. G. (1951) A serpentine mineral from Kennack Cove, Lizard, Cornwall
MineroL. M og. 29, 52G530.
Nor.r, W. AND H. KrRcHEn (1950) Zur Morphologie des Chrystoilasbestos. Naturwissen-
schaften 37, 51t0-541
- - (1951) Uber die Morphologie von Asbesten und ihren Zusammen hang mit cler
Kristallstruktu r. N anes.J ahrb. M in eraI., M onatsh, 10, 219-24n.
Pacn, N. J (1966) Minerdogy ond. chanistry of tke serpenti.negroup mi.nerals ond tlte ser-
pmti.ni.zation process.Ph.D. Thesis, University of California, Berkeley, California.
-- (1967) Serpentinization at Burro Mountain, California. Contrib. Mineyal. Petrology
14.321-342.
214 NORMAN J PAGL,

- AND R. G. ColnueN (1967) Serpentinc mineral anall'ss5 and physical properties.

U. S. Geol'.Suru. ProJ. I'ap. 575-8, B103 8107.
Rucrlrncr, J. C , exo J. ZussuaN (1965) The crystal structure of the serpentine mineral,
iizardite MesSirOz(OH)r. Acta Crystallogr. 19, 381 389.
Tunrnrtcu, J. axn J. Hrr.r.nn (1949) Electron microscopy of colloidal syslems Anal'.
Chen. 21,475 485.
Wamnw, B. E. eNo K. W. Hrnrnc (1941) The ranclom structure ol chrysotile asbestos.
[abstr.] PZ}s. Ret:. 59, 925
Wrrtrtarnn, E. J. W. (1953) The structure of chrysotile. A cta Crystallogr. 6, 747 .
--- (1955) Aclassificationof cylindricallaltices ActaCrystallogr.a,STl 574.
--- (1956a)The structure of chrysotile. IL Clinochrysotile. Acta Crystdlogr. 9, 855 861.
- (1956b) The structure of chrysotile. III. Orthochrysotile. Acta Crystdlogr.9,862-
864.
-- 11956c)The structure of chrysotile. IV. Parachrysolile. Acta Crystallogr.9, 865-
866
-- (1957) The structure of chrysotile. V. Difiuse reflections and fibre Iextrre. Acta
Crystailogr. 10, 149
J. Zussuex (1956) The characterization of serpentine minerals by X-ray dif-
fraction. M,inualog. M ag. 31, 107-126.
ZussuaN, J. (1954) Investigation of the crystal structure of anttgorite. Minerdog. Mag 3O,
498-5t2.
--- (1956) Antigorite: superlattice and structural formula. Amer. Mi,neroL. 41r 148-152.
--, BnrNu.nrr, G W. em J. J. Counn (1957) Electron diffraction studies of serpentine
minerals. Amer. M inual. 42, 133-153.

Monuscr'i,pt receired,May 11, 1967; occepteilJor publi,cotion, September10, 1967.

AppeNorx: Drscussror.r oF SAMpLES UsBo rN THE SruDy

The samplenumberslisted below are from those analysesof Faust and

Fahey (1962) used in this study. Pertinent comments are given below
and include the date, whether FerOaand FeO, Hz(-) and H2O(f ), and
AlzOswere determined,impurities present,and probableserpentinemin-
eral. An asterisk indicates that the analysis was used in the statistical
treatment of the presentpaper.

*6;1936,0.50
CaO,probablylizardite. *35;1930,n.d.Fe2O3,
chrysotile
*10;1936,
HzOlumped,0.73CaO,prob- *37;7936,
chrysotile.
ably lizardite. 42;1936,0 40 CaO, chrysotile.
*72;7936,0.63 CaO, antigorite. 43; 1936, probably antigorite.
13;1936,0.98 CaO, antigorite. x45; 1957, chrysotile.
*15; 1957, antigorite. *49;7954, antigorite.
*20;1906, HzO lumped,0.?2 CaO, antig- 50; 1931,0 75 total impurity, probably
orite. antigorite.
*28;1936, probably lizardite. *51;1947, n.d. FeO, 0.14 NazO, probably
29;1848, n.d. FezOa,H2O lumped, antig- lizardite.
orite. 53; 1885, n d. Fe2O3,n.d. Alroa probably
*39; 1857, n.d. FerOr, H2O lumped, antig- chrysotile.
orite. 5 4 ; 1 9 3 1 ,n . d . F e 2 O 3n, . d . A l r O e , 0 . 4 0t o t a l
33; 1954, sixJayer orthoserpentine. impurity, probably chrysotile.
 SERPEN TI N E " POLV MORP II S'' 21s

63;1931,0.43 NarO and 0.19 K2O, prob- *136; 1818, n.d. !'e:Oa, n.d. AlzOa, antig-
ably antigorite. orite.
*64; 7957, chrysotile. *137;1857, n.d. Fe2O3,n.d. AlzOa, chryso-
*65; 1957, chrysotile. tile.
+ 6 7 ; 1 9 3 6 , 1 . 4 1 C a O a n d 4 . 5 4 C O 2 ,c h r y s o - *146;1951,n.d. FeO, HzO Iumped, lizard-
tile. ite.
70;1906, n.d. FeO, HzO lumped, prob- +147; 1926, n.d. l-eo, 1.25 impurities, chrys-
ably chrysotile. otile.
7 l ; 1 9 3 1 , n . d . F e O , n . d . A l 2 O 3 ,p r o b a b l y 152; L834, n.d. IrerOa, H2O lumped, chrys-
chrysotile. otile.
75;1930, H2O lumped,055 total impu- +194; 1936, n.d. FeeO3, n.d. Al:Oa, 0.58
rity, chrysotile. CaO, chrysotile.
76;1906, n.d. FeO, chrysotile. *197;1931,n.d. FezOs,n.d. AlzOs, chryso-
*78; 1856, n.d. FezOr,I{2O lumped, prob- tile.
ably antigorite. 203; 1885, n.d. FezOa, H2O lumped, n.d.
79; t956, chrysctile. AlzOs, chrysctile.
*86; 1930,0.35 CaO, chrysotile. 204;1897, n.d. F-e2O3,H2O lumped, chrys-
89; 1910, lumped FeO-FezOa, and H:O, otile.
chrysotile. 205; 1891, H2O lumped, n.d. Al2Os,chryso-
*97; 7957 chrysotile. tile.
,
*94, tgl0,lumped FeO-FerOa and HzO, 206; 1883, Fe:OrFeO and H:O lumped,
chrysotile. chrysotile.
+96 1936, chinochrysotile. 212;1891, n.d. FeO, H2O lumped, chryso-
;
97;l930,lumped FeO-FerOr and H:O, tile.
average chrysotile. 217; 1955, n.d. HzO-, chrysotile.
100; 1954, lumped HrO, 1.31 CaO, chryso- x222; 1957 chrysotile.
,
tile. 243; 1947, n.d. AlO3, chrysotile.
101; 1891, n.d. Fe2O3Iumped H2O, prob- 248;1918, n.d. FezOs,chrysotile.
ably chrysotile. +250; 1918, n.d. FezOr,chrysotile.
*102;1936, n.d. AlrQr, labelled antigorite, 251;1918, n.d. FerOa,chrysotile.
+260;1890, n.d. FeO, chrysotile.
probably chrysotile.
*F-1; 1956; antigorite.
1lI; 1930,n.d. Fe2O3,290 CaO, chryso-
+\- -15 1962, antigorite.
tile. ;
*F-20; 1956, clinochrysotile.
112; 1954, n.d. FeO, six-layer orthoserpen-
* F -22 1962, clinochrysotile.
tine. ;
*F -23 1962, lizardite.
115; 1910, lumped Fe2O3-FeO,n.d. H2O-, ;
* F -24; 1962, clinochrysotile.
n.d. AlzOa,chrysotile. * F -47 1962, clinoch
* 120; 1952, antigorite. ; rysotile.
*F -43 1962, Iizardite.
* l 2 l ; 1 9 3 7 , n . d . F e 2 O 3 ,n . d . H r O a , c h r y s o - ;
aF - 46 1937, Iizar dite.
tile. ;
*F -47 1937, Iizardite.
*129;1930, n.d. FezOr, 0.16 CaO, chryso- ;
tile. From Page (1967) the following samples
*132;1876, n.d. FeO, H2O lumped, antig- were used: 19-NI-63A, 38-NZ-62-5, 94-NZ-
orite. 62, 19-NI-638, and 14-NI-63A.