Canad.ian Mineralo gist
Vol. 16, pp.239-244 (1978)
THEEFFECT
OF BORONON THEGRANITESOLIDUS
LESLEY B. CHORLTON* EUO ROBERT F. MARTIN
Department of Geological Sciences,McGill University, 3450 University Street,
Montreal, Que, H3A 2A7
perature of granite, COz and HCI led to an increase in solidus temperature, presumably beThe addition of boron to a rrater-saturatedgra- cause of a reduction in the activity of water in
nitic systemlowers the solidus at 1 kbar by 125"C; the melt and coexisting vapor phase. Addition
the liquidus seemssimilarly affected. This drastic
led to a decreasein the solidus
reduction reflects tle incorporation of tetrahedrally of HF and LirO
coordinatedboron (Bj in the place of aluminum in temperature by partitioning strongly into ttre
the liquid and the influence of tiis substitution on melt and modifying it structurally. Boron is one
the thermal stability of the feldspars. The ratio of minor component excluded from their preliBtv to Bir in the melt increaseswith increasingpres- minary suryey, but of local importance in nature,
sure and alkalinity and with decreasingt€mperature. especially in pegmatitic environments. This
Reedmergnerite-bearingperalkaline pegmatites in paper provides some insights into the efficiency
alkaline complexes could have crystallized from o1 this constituent as a flrD( in granitic bulk
melts at temperaturesbelow 600"C.
compositions.
Arsrnecr
SorranrernB
ENERIMENTAL
L'addition de bore i un systdmegranitique satur6
d'eau en abaissele point de fusion de 125"C I 1
kbar et semble 6galementaffecter le liquidus. Cet
effet r6sulte de la substitution du bore i I'aluminium dans les sites t6tra6driques du liquide et de
I'influence qu'exercecette substitution sur le champ
de stabilit6 thermique des feldspaths.Le rapport de
Biv e Bil augmenteavecla pressionet I'alcalinit6du
milieu, et varie inversement avec la temp6rature.
Les pegmatiteshyperalcalinesi reedmergperitedes
complexesalcalins ont pu cristalliser d partir de liquidesr6siduelsau-dessous
de 600'C.
TECHNIQUB
The synthetic starting materials chosen for
this study are glassesprepared by the method of
coprecipitating gels (Luth & Ingamells 1965),
and used in a previous study of the system
(Luth et al.
KAlSigOs-NaAlSiaOrSiOrH,O
1964). These glasseswere ground, dried in a
vacuum oven for four to eight hours, then stored
in a desiccator until required. Boron was introduced as analytical-gradecrystalline IIsBOa,nonhygroscopic and stable at room temperature.
]]ire SrOr content of each run was calculated
frorn the amount of HgBOs added. Total water
content, generally between 10 and 20 wt. Vo,
INTRoDUcTIoN
was calculated to include the water added in
'system"
Many experimental studies of the
the form of H'BOg. Solids and distilled water
granite-water have been undertaken since the were loaded into 2mm Au tubing 10 to 15mm
pioneering efforts of Goranson (1932): Tuttle & long, and welded shut. Experiments were run at
Bowen (1958) and l-uth e/ al. (1964) determined P(H,O) = 1 kbar in externally heated 2.5cm
solidus temperatures and phase relationships in O.D. stellite cold-seal pressure vessels;temperthe quaternary haplogranite system KAlSisOs- atures fluctuated :t5 or 6oC during the 5- to
NaAISLOs-SiOg-HaO as a function of total
ll-day isothermal exPeriments.
pressure; Wyllie & Tuttle (1959, 1960, I96t,
Other experiments were run at atmospheric
1964) extended the applicability of these fundapressure
on a tourmaline-bearing granitic segre'
mental studies to natural situations by adding a
gation from a pegmatite in the Boulder bathonumber of components to the lime-free.granite
Montana (Knopf 1957). This granite
SOe and PsOr produced iith,
system. Whereas \ft,
(L697A in Knopfs collection) consistsof quartz,
negligible effects on the minimum melting temmicroline (OrrJ lom albite and 2O vol. Vo dark
green tourmaline; the rock contains 72.68 wt.
% SiOo 14.63% AlzOq l.74%o BzOs,O'56Vo
*Present address: Department of Geology, MemoCaA, 2.49Vo NaaO and 4.36Vo KzO (8. B.
rial University of Newfoundlando St. John's, Nfld.
Knopf, priv. comm.; full analysis in Chorlton
Alc 5$7.
239
240
THE CANADIAN MINERALOGIST
1973).The 1 atm isothermalexperimentslasted
from 4 days (experimentsabove 10'00'C) to 22
days (experiments at 743 t soc, the lowest
temperature in this series).These run durations
are not sufficiently long for equilibrium to have
been attained.
P ( H g O )" I t l o r
x+L+v
Rrsur-rs oF HyDRoTHERMAL
ExpERTMENTs
These experiments were designed to evaluate
the effect of boron on the temperature at which
granite begins to melt. This was achieved by
holding a series of up to four synthetic bulk
compositions isothermally and estimating the
proportion of melt produced; one composition
was boron-free, and a maximum of. L0 wt. %
BsOgwas added. Details of bulk composition,
temperature, duration and run products are presented in Table 1 for selected experiments; full
details on all experimentsare provided by Chorlton (1973).
The data points 1n the region of the solidus
allow a boundary to be drawn, separating the
field of crystals * vapor from that of the vaporsaturated, partially melted assemblage(Fig. 1).
The minimum melting temperature at P(HzO) =
1 kba.r for boron-free compositions is close to
72O"C, as determinedby Tuttle & Bowen (1958)
and confirmed in this study. The compositions
used have predicted solidus temperatures be-
Amount ot boron oddod (rt. % BzO!)
Frc. l. Trace of the minimum melting temperature
in the systemKAlSigOs-NaAlSisO8--SiOr-HrO
rBzOaat I kbar, as a function of the amount of
boron addedto the system.The data pertain to
bulk compositionsnear the quartz-feldsparfield
boundaryin the boron-freesystem.Thesecompositions vary in Ab:Or ratio from 0.167 to I on
a weight basis.Empty circles indicate subsolidus
run products;the proportion of black in a data
point is indicative of the proportion of glass
(quenchedmelt) in the run products.
tween 720 and 74O"C, depending on the ratio
Ablor (in the range of 0.167 1o 1 on a weight
basis).The addition of 5 wt. %oBzO"to the granitic
bulk compositionslowers the melting temperTABLE1. RESULTS
OF SELEETED
HYDROTHERMAL
EXPERIMENT5
ON
ature drastically to approximately 595"C, a drop
+ Bz0sAT P(H20)= I KBAR
HAPL0GRANITIC
MIXTURES
of. 125". Additional boron does not lead to furCOMPOSITIONBr0r
TEMPDURATION RUN
ther reduction (a horizontal solidus is drawn in
(DAYS) PRODUCTS'
Ab 0r Q* A D b E D T ( . c )
Figure 1 on the basis of the data point at 8.3%io
q+San+V+b
t0 50 40 A.9l S93.5
7
BgOg)
but does lead to the appearanceof specks
8.35
Q+San+V+b
of unidentified birefringent borates in tiny white
30 30 40 4 . 7 0 5 9 5 1 1 5 l 3
Q+San+V+b
I 0 .1 1
Q+San+35%gl+V+b vitreous spheresthat are interpreted as quenched
30 40 30 1 . 7 5 6 3 7 ! 2
( tr)+V
9
vapor phase. These findings suggestthat saturaQ+San+Ab
5.00
Q+San+Ab(tr)+30%g'l+V
tion in tetrahedrally coordinated boron has been
] 0 5 0 4 0 9.5
gl + V
640i5
7
achieved with SVo BsOs, and that additional
30 30 40 1 . 9 4 6 5 5 1 5
(tr)+V
I
Q+San+Ab
boron partitions more strongly into the vapor
4.96
(tr)+50%g
l +v
Q+san+Ab
phase,where it occurs in part as Brtr,coordinated
20 30 50 0
665i5
I1
Q+san+V
1 07
to three oxygen atoms. The differences in proQ+San+25%gl+V
4,98
Q+San+60%gl+V
portion of melt produced in bulk compositions
Q+70%gl+V+b
containing 5 and lOVo Bzo' (FiB. 1) suggestthat
]0 40 50 0
70915
9
Q+san+V
'I
the liquidus is also drastically depressedupon
q+San+25%gl
.98
+V
9.74
l +V+b
Q(tr)+98%s
addition of boron to the systern KAlSiaOs*
20 40 40 0
71515
7
Q+san+v
NaAISLOg-SiOg-H,O.
I .93
Q+San+50%gl+V
4,97
Although Maitrallet (1976) failed in his efforts
Q+San+70%gl+V
Q+90%gl+V+b
to synthesizetourmaline in a granite * water *
20 50 30 0
73816
5
Q+San+30%gl+V
HrBOr mixture held at P(HzO) = 2 kbar, he did
107
Q+San+60%gl+V
confirm that at 625"C, the rock was partially
V a l u e sa r e g l v e n i n w e i g h t% .
melted; without boron in the system, the mini'
S y n r b o luss e dt o d e s c r l b et h e r u n p r o d u c t s :Q ( q u a r t z ) ,
mum melting temperature is 680oC (Tuttle &
S a n ( s a n l d l n e ) , V ( v a p o rp h a s e ) ,b ( u n l d e n t i f i e d b o r a t e s
Bowen 1958). Maitrallet found an even gteater
t h a t c r y s t a l l l z e . f r o m t h e v a p o r p h a s e ) ,t r ( t r a c e ) ,
A b ( a l b l t e ) ,g l ( g l a s s ) .
degreeof melting when he added 2.5 wt. % NaF
to the .mixture, suggestingthat the influence of
EFFECT OF BORON ON GRANTTE SOLIDUS
fluorine and boron on melting phenomena is
additive.
As there is ample evidence that tetrahedrally
coordinated boron (Bt") can replace aluminum in
feldspars (e.9., Eugster & Mclver 1959, Barsukov 1961,Sheppard& Gude 1965,Martin 1971),
some indication of B-for-Al substitution was atil
sought in the unit-cell dimensions of the Krich feldspars in the run products. Cell parameters were determined on products of subsolidus
experiments,so that boron would be partitioned
only between crystals and vapor phase. X-ray
le96o 12.980 |3.OOO
. l3.O2O l3.O4O
diffraction patterns of powders were obtained
,(A)
with a Guinier-Hlgg camera, CuKar radiation'
Fro. 2. AnomalousD-c coordinatesof K-rich feldat room temperature; indexed 2d values corspars synthesizedfrom boron-bearinghaplograrected against a quartz internal standard were
nitic compositions;apparentmol Vo Or is shown
used as input for the Applernan & Evans (1973)
next to eachdot. The sanidineallrtaining 78.6Vo
cell-refinement program.
Or was synthesizedat 715'C from a boron-free
mixture; its positionin the b-c plot is as expected
A plot of the cell edgesb vs c is commonly
for its composition,calculatedfrom the a cell
used to obtain inforrnation on composition and
edge. Sanidinescontaining79.2, 8L6, 82.2 alt.d
degree of SiAl order of feldspars.The resulting
at 595, 593, 593 and
84.4VoOr were synthesized
quadrilateral can be contoured for composition
506'C, respectively,from mixes that had 10.1,
with the a cell edge; a should increase with de8.4, 4.9 and lO.2Vo(rfi.) BOr, respectively.Four
creasing proportion of Na in the K-feldspar.
sodium-freesyntheticK-feldspars,representedby
However, a is found to contract slightly, by
a diamondsymbol,containO, 5.5, 9,7 and 18.1
0.013A, between KAlSigOe and (KBSisOa)re.r mol lo KBSisOs.HS and LM refer to high sani
(KAISisOa)ar., (Martin L97L); this is approxidine and low microcline,respectively.
mately the equivalent of.2.7 mol Vo Or. Bi"-forAl substitution leads to greater reduction in D
The Fc.plot cannot be used to determine the
and c cell edges,and should thus be reflected in
extent of Bt'-for-Al substitution. However, a
anomalous trends in the D-c plot.
feldspar synthesizedat 506'C from a mixture
Compositions calculated from the a cell edge containing L0J97o BrOa differs from its exas if the feldspars were boron-free suggestthat pected location in a b-c plot contoured for a
the sanidines become increasingly K-rich with
(Stewart L975) by roughly the same extent as
decreasing temperature of synthesis, as fully
the separation between compositions O and
expectedin view of the widening solvus between 5.5% KBSLO' (Fig. 2). We tentatively conclude
KAlSisOs and NaAlSisOe.The trend of increas- that the K-rich feldspar contains approximately
ing a, leading to calculated compositions from this proportion of the boriferous end-member.
Orzeat 71,5"C to Orsaat 5O6oC,should therefore The coexisting sodic feldspar, not studied by
lead to displacementstoward the high sanidine- X-ray diffraction, presumablyalso containsB'",
low microcline sideline (Fig. 2), reflecting larger although extent of solid solution toward reedb and c values. Instead, both D and c contract, mergnerite, NaBSisOs, may be limited by a
defining a trend away from this sideline; shrink- miscibility gap
@ugster & Mclver 1959). The
agein b and c is expectedas a result of Blrforspecksof borates depositedfrom the fluid phase
Al substitution due to the small size of the boron presumably contain boron primarily as Bttt.
atom. The trend in Figure 2 is analogous to that
defined by four synthetic monoclinic KrooNao
feldsparscontaining0, 5.5,9.7 and 18.1 rnol Vo
Rssurrs on DnY MBruNc EXPERTMENTS
KBSiaOa (Martin 1971). As the value of No.
Upon heating, the microcline (Orss)in tourquoted in Figure 2 for the feldspars synthesized
from the granitic mixtures are slightly low be- maline granite 1697A reequilibrates compolicause the feldspars are boron-bearing, the ano- tionally,-presumably by homogenization with
maly in location of points in the D-c plot even nearby albite grains: Orzsat 841'C and Oror at
exceedsthat suggestedby the apparent No" cotr- 1005"C, as revealed by a cell edges.Microcline
tents. These results suggestthat the K-rich feld- also attempts to re-equilibrate structurally, by
spars do incorporate Bt", and increasingly so at disordering the interrnediate microcline, a metastable state that persists until K-feldspar is :qi'
progressivelylower temperatures.
242
THB CANADIAN MINERALOGIST
pecially in bulk compositions seqlaining more
tban 5Vo BaOa.Approximately 5 mol Vo of the
boriferous feldspar end-member could be present in the K-feldspars forming upon crystallization of such low-temperature granitic fiquids.
Although the importance of boron is wellappreciated in low-temperature sedimentary environments, little pertinent information is available on the distribution of boron among the
minerals of granites and pegmatites. Boron is
unfortunately not among tle commonly sought
constituents of rocks and minerals, nor are Bot/
Bt" ratios readily available. Tourmaline-bearing
granites are well-known, especially in Cornwall,
England. Brammall & Harwood (1925) and
Exley & Stone (1964) have noted the general increase in tourmaline content, commonly accompanied by topaz and, occasionally, fluorite, from
the earliest to tle latest intrusive units. Textural
relationships suggest that two generations of
tourmaline, one possibly magmatic and the other
post-magmatic, may characterize many rocks in
this association.Whether or not the accompanying feldspars contain Bt" substituting for Al
would provide an interesting subject of research
and a possibletest of the likelihood that some of
thesegranites formed from liquids whose solidus
was depressedbecauseof appreciableboron conDrscusstoN
centrations. If a tourmaline granite contains only
Work done on industrial borosilicate and bo- Btit, enrichment in boron may have occurred
roaluminate glasses(e.9., Warren 1941, Riebling entirely via a fluid phase, after the primary
1964, Bray 1967) suggeststhat boron occurs in crystallization of the feldspars and thus clearly
the melt both ffiply coordinated, in BOa groups, at the postmagmatic stage.Late development of
and tetrahedrally coordinated, in BOa groups. tourmaline may also result from the release of
The ratio of Bil to Bt" in a melt, as determined Bt" from structures of primary silicates during
by proton NMR spectrq depends on temper- episodes of rock-water interaction.
atureo pressure and bulk composition. IncreasAt least one pegmatitic complex is known in
ing temperature favors Bttt over Bt", whereas in- which boron enrichment seemsto have occurred
creasing .pressure favors Bi" due to the denser at the end stages of magmatic crystallization.
packing of atoms in this configuration @ockris Pockets in the Dara-Pioz alkaline pegmatite,
& Kojonen 1960). Addition of alkali atoms to Alai range, Tadzhik SSR, contain quartz, albite,
melts in the system BqOg-SiO,can be correlated perthitic microcline, aegirine and reedmergnewith increasing proportion of tetrahedrally co- rite (Dusmatov et aJ. L967).In such peralkaline
ordinated boron @ray 1967). Thus, low temper- environments, Bt" may well predominate over
ature, high pressureand peralkalinity can be ex- Btn; for example,tourrnalineis not expectedin
pected to favor the occurrence of Br" in natural peralkaline pegmatites, presumably because it
environments.
is invariably a peraluminous phase, i.e., (Na *
The data presentedin this paper confirm that K) < Al. AII silicate minerals in the pegrnatitic
in granitic bulk cornpositionsin which (Al * B) pockets, including aegirine and the alu'ninous
exceeds (Na * K) only slightly, more tetra- feldspars, contain appreciable quantities of tehedrally coordinated boron could be expected trahedrally coordinated boron. The mineral asat temperaturesclose to 500o than above 700oC, semblagein these pockets could have crystallized
as suggestedby the trend of feldspars in Figure from a water-saturated,Br"-enrichedgranitic melt
2. The various run product further confirn the at temperatures below 600oC or frorn the copresenceof the two coordination statesof boron existing high-pH fluids. Cell dimensions of the
in granitic bulk compositions: B!" occurs in aluminous feldspars suggest that Bi'-for-Al subfeldspars and possibly in mullite, Bt't is found stitution is more extensive in albite than in the
in borates deposited from the vapor phase, es- coexistingmicrocline (R. F. Martin & V. D. Duscompletely melted. In experimentsat the lowest
temperatures,hematite appears as a coating on
the surface of an unidentified high-relief greenish mineral, apparently an isotropic breakdown
product of tourmaline. A diffraction peak identified as the (1.22) reflection of tourmaline is
present in run products up to 877oC. Melting
begins in the interval 858 to 877'C; quartz, microcline, albite, hematite and bundles of mullite
fibres coexist with melt. The solidus temperature of lourmaline granite L697A is evidently
about 80 to 90'C lower than 96O"C" the minimum melting temferature at 1 atm in the system KAlSisOe-NaAlSisOrSiO, (Iuttle & Bowen
1958). Interestingly, none of the unsealed
charges showed significant weight loss on heating, even to temperatures above the breakdown
of tourmaline. Run products were essentially the
same whether the capsuleswere sealed or left
open. This suggeststhat boron derived by breakdown of tourmaline partitions strongly among
coexisting liquid * crystals; very little is lost
as volatile borates. Anomalies in the cell dimensions of mullite (Chorlton 1973) are tentatively
attributed to Bt"-for-Al substitution.
EFFECT OF BORON ON GRANITE SOLIDUS
243
matov, unpubl. data). Other examplessimilar to Bnev, J. P. (L967): Magnetic resonance studies of
bonding, structure, and diffusion in crystalline
the Dara-Pioz pegmatite may have been overand vitreous solids' fto Interaction of Radiation
lookedn as the reedmergnerite there looks very
with Sofds. (A. Bishay, ed.), Proc. 1966 Solid
much like an ordinary feldsPar.
State Conf., Cairo, Pletum Press, New York.
This investigation was planned to explore
L. B, (t973): The Effect of Boron on
empirically the effect of adding boron to granitic CHoRLroN,
in the Granite-VFater Systern.
Phase
Relatiow
bulk compositions. Besausethis componeat has
M.Sc. thesis, McGill Univ., Montreal, Que.
a tendency to occur in two types of coordination
A. & KesANove, L.
polyhedra, general statements concerning the DusMATov, V. D., Popove, N.
(1967) I The first occurreoce of reedmergnerK.
and
liquid
disiribution of boron among solids,
ite in the USSR. Do*/. Akad, Nauk Tadzh. SSR
vapor are difficult to make at this stage. Ho-w10, 51-53.
ever, our observationson depressionof melting
II. P. & Mclvr& N. L. (1959): Boron
Eucsrnn,
temperatures, coupled with the relatively low
analogues of alkali feldspars and related silicates'
reedof
temperature of the incongruent melting
Geol, Soc. Amer. Bull. ?0, 1598-1599(Abstr')'
mergnerite (NaBSiaOs+ tridymite * L at 862"C,
E<r-By, C. S. & SroNg M. (1964): The granitic
1 atm: Milton et al. L96O) and with preliminary
rocks of southwest England. /z Presont Views on
documentation of a miscibility gap between
Some Aspects of the Geology of Cornwall- and
(Eugster
Mclver
&
NaAISLO8 and NaBSLOg
Devon. Roy, Geol. Soc. Cornwall, Oscar BlackL959), are directly applicable to pegmatites, esford Ltd., Truro, England.
pecially those in which (Na + K) > Al. The inR. W. (1932): Some note$ on the meltGoRANsoN,
?luenci of boron on the melting temperature of
ing of granite. Amer. t. Sci, 223' 2n'n6.
granitic compositionswill likely be in addition-to
KNoPF, A. (1957): The Boulder bathvlith of Mondepressions in melting points caused by IIrO
tana. Amer, l. Sci. 255' 81-103.
(Wyllie
anC ty elements like fluorine and lithium
C. O. (1965): Gel
& Tuttle L961, L964),which modify the structure Lutr, W. C. & INcervrsr.r,s,
materials for hydrolhermal
prepa.ration
starting
of
ways
different
of the silicate melt in completely
ixfirimentation. Amer. Mineral. 50' 255'258'
than does a network-forming cation like boron.
----,
JAHtis, R. H. & Turrr-n, O. F. (1964):-The
granite system at pressures of 4 to 10 kilob'ars'
ACKNOWLEDGEMENTS
l. Geophys, Res- 69, 759-773,
(1976): Les Nodules d Tourmaline
This paper summarizesthe findings of a M.Sc. Merrrer.r.nl P.
dans Quelques Gisements de
des
Granitoidis
thesis wriiten by the first author. We would like
France et tltalie. Thdse 3e cycle, Univ' Pierre
Luth,
C.
Drs.
W.
of
to acknowledge the help
et Marie Curie, Paris, France.
Stanford University, A. S. Marfunin of IGEM,
authigenic feldUSSR Academy of Sciences, Moscow, V. D. MenrrN, R. F. (1971): Dsordered- KB-SLOs.frorn
KAlSiaO8
the.series
of
spars
Dusmatov of the Tadzhik SSR Academy of
siutlern California. Amer. Mineral' 56' 281-29L'
Sciences,Dushanbe and E. E. Senderov of the
J' M'. &
Vernadskii Institute of Geochemistry and Ana- Mu-ror.r, C., CHAo, E. C. T., AXELRoD,
(1960):
Reedme-rgnerite'
S.
F.
GRlvALot,
in
participated
RFM
lytical Chemistry, Moscow.
NaBsloa, the boron analogrre of albite, flgn the
a National ResearchCouncil*USSR Academy of
Green River Formatlon, Utah' Amer' Mineral'
acknowledges
Sciences exchange in 1976, and
45, 188-199.
the continued support of NRC through its grant
E. F. (1964): Structure of borosilicate
furnr-rNc,
N72L.
and borogermanate melts at 1300"C; a viscosity
study. J. Arner. Ceram' Soc' 47'
u"A a""itv
REFERENCES
478-483.
[I (1965): Potash
AprlrrraN, D. E. & EveNs, H. T., Jn. (1973): Job Snnrraxo, R. A. & Guoe, A. J.,
in the Barsvalue
possible
economic
of
feldspar
9214: indexing and least squares refinement of
San Bernardino County, Caliior" To..uiion,
powder diffraction data. U.S. Geol. Surv. Comp'
fornia. U.S. Geol. Surv. Circ' 500'
Contr. 20.
parameters'.comBARsuKo% V. L. (1961): Some problems of the Suwanr, D. B. (1975): Lattice
feldspars'
position,
geochemistry of boron. Geochem. 596-608.
'Mineral'. and AllSi order in alkali
Amer. Short Course Nore't 2' St'
Soc,
(1960):
comThe
Bocrrts, J. O'M. & KoloNnN, E.
1-22.
pressibilities of certain molten alkali silicates and
TurrLB, O. F. & BowrN, N. L. (1958): Ori$n 9t
6orates.l. Amer. Chem. Soc.82, 44934497'
granite in the liebt of experimental -studies m
Bnuraual, A. & Henwooo, H. F. (1925): TourmaNaAtslos-KAlsi,os-sio,-H zo' Geot'
A;G;
granite.
Mag'
Mineral,
linization in the Dartmoor
Mem,74.
Amer.
Soc.
20, 319-330.
244
THE
CANADIAN
WARnrN, B. E. (1941): Summary of work on
atomic arrangement in glass. J. Amer. Cerem.
Soc. 24, 256-261.
Wvnrc, P. J. & TurrLE, O. F. (1959): Effect of
carbon dioxide on the melting of granite and
feldspars. Amer. J. Sci. %7,648-655.
(1960): Experimental investiga& _
tion of silicate systems containing two volatile
components. Part I. Geometrical considerations.
Amer. J, Sci. 258, 498-517.
(1961): Experimental investiga& tion of silicate systems containing two volatile
MINERALOGIST
components. Part II. The effects of NHg and HF,
in addition to HsO, on the melting tempe.ratures
of albite and granite. Amer. l. Sci. 259, L28-I43.
(1964): Experimental investiga& tion of silicate systems containing two volatile
components. Part IfI. The effects of SOg, PrO",
HCl, and LirO, in addition to HO, on the melting temperatures of albite and granite. Amer. J,
Sci. 262, 930-939.
Received November 1977; revised manuscript accepted February 1978.