American Mineralogist, Volume 74, pages 1032-1037, 1989
Diopside-tridymite liquidus boundary line in the systemMgrSiOo-CaMgSi'Ou-SiO' at
atmosphericpressure*
Tnn-CnrNc Lru, Dn.ls C. PnnsN.q,LL
Geosciences
Program,The University of Texasat Dallas,P.O. Box 830688,Richardson,Texas75083-0688,U.S.A
AssrRAcr
The presenceor absenceof a temperature maximum and minimum on the diopsidetridymite liquidus boundary line in the system MgrSiO.-CaMgSirOu-SiO,at atmospheric
pressurehas been a matter of dispute. To resolve this controversy, we have determined
the temperatureprofile and compositions of coexistingdiopside and liquid along the diopside-tridymite liquidus boundary line from the invariant point tridymite-diopside-pigeonfrom the invariant
ite (1376'C) to a point on the oppositeside of the join CaMgSirOu-SiO,
point. The temperature profile along the diopside-tridymite liquidus boundary is constrained by (l) close brackets ofthe first appearanceofdiopside for starting compositions
in the tridymite primary phasefield and (2) determination of the compositions of glasses
along this boundary line at known temperatures. Becausethe temperature gradient is
extremely small and near the limit of experimental resolution, our data do not definitively
resolve the issue. However, the data are most consistent with continuously decreasing
temperaturesalong the diopside-tridymite boundary from the tridymite-diopside-pigeonite invariant point. Also, three-phasetrianglesof glass+ diopside * tridymite, determined
for a range of liquid compositions along the boundary line, show the same orientation
over the entire length of the diopside-tridymite boundary line and are consistent with a
continuously decreasingtemperature profi le.
INrnooucrroN
The systemMgrSiOo-CaMgSirOu-SiOr,
a classicternary
system used to discussthe petrogenesisof basaltic magmas, was first investigated at atmospheric pressure by
Bowen (1914).In his original diagram,temperaturesalong
the boundary line between the pyroxene and silica primary phase fields were shown continuously decreasing
from the MgrSiOo-SiO,side to the CaMgSirOu-SiO,
side.
Schairer and Yoder (1962) recognizeda boundary line
separatingthe fields of diopside and protoenstatite and
showed a temperature minimum on the diopside-tridymite boundary line. Schairerand Kushiro (1964) subsequentlyfound that the join CaMgSirOu-SiO,
is not binary
and therefore cannot be a thermal barrier. In their diagram, they omitted the temperatureminimum and showed
the temperature continuously decreasingfrom the tridymite-protoenstatite-diopsideinvariant point to the join
Kushiro (1972) recognizeda third pyCaMgSirOu-SiOr.
roxene, pigeonite, at the liquidus, and redetermined the
location of the diopside-tridymite boundary line using
electron-microprobe
analyses.He againshowedthe temperature profile along the diopside-tridymite boundary
line to be continuously decreasingfrom the tridymitepigeonite-diopsideinvariant point (previously thought to
be the tridymite-protoenstatite-diopsideinvariant point)
* Contribution no. 587, GeosciencesProgram,The University
of Texas at Dallas.
I 0-1032$02.00
0003-o04x/89/09
Longhi and Boudreau(1980)
to thejoin CaMgSirOu-SiOr.
located an orthoenstatite liquidus field but did not study
the diopside-tridymite boundary line. In their diagram,
they took the diopside-tridymite boundary from the work
of Kushiro (1972).
Kushiro and Schairer
On the join CaMgSi.Ou-MgSiOr,
(1963) found a temperaturemaximum on the diopside
liquidus that lies slightly toward MgSiO, from CaMgSirOu.Kushiro (1972) redeterminedthe location of this
maximum and placedit slightly toward the MgrSiOoapex
On the basis of the
from the join CaMgSirOu-MgSiOr.
existenceof this maximum, Morse (1980, p. 203-206)
argued that there must exist a special diopside-silica tieline correspondingto a point of maximum thermal stability (point M' u Fig. l) on the diopside-tridymite
boundary line. Becausetemperaturesmust decrease
slightly along this boundary line as it extends from the
tridymite-pigeonite-diopsideinvariant point, Morse correctly argued that his postulated temperature maximum
would also require a temperature minimum (point m in
Fig. 1). Longhi (1987, p. 304, 313) stated that Tucker,
working in Longhi's laboratory, has confirmed the existence of the maximum M' and the minimum m (Fig. l),
but that the results are inconclusive.
In the absenceof diopside solid solution, the argument
of Morse would hold. However,because(l) diopsidesolid solution occurson both sidesof the liquidus maximum
and (2) temperaturesalong the diopside-tridymite boundary line are lower than the diopsideliquidus maximum,
r032
t 033
LIU AND PRESNALL: DIOPSIDE-TRIDYMITE LIQUIDUS BOUNDARY LINE
c
A S c h a i r e r& K u s h i r o , 1 9 6 4
t r K u s h i r o ,1 9 7 2
This study
O Glass
c o m p o si t i o n
o Starting
c o m p o si t i o n
I
I
I
Mg2SiOa
MgSi03
weight
itr
5io2
percenl
Fig. 1. Liquidussurfaceof the systemMgrSiOo-CaMgSirOupressure
SiO,at atmospheric
(redrafted
afterMorse,1980).Point
M is a temperature
maximumon the forsterite-diopside
boundary line. PointsM' and m area temperature
maximumand a
Mg2SiOa MsSiO3
temperature
minimum,respectively,
on the diopside-tridymite
weight percenl
boundaryline proposed
by Morse(1980).
the diopsideapex of diopside-tridymiteJiquid three-phase
triangles is not constrainedto lie at the diopside liquidus
maximum. Stated another way, the diopside liquidus
maximum would fix the compositionof diopsidein equilibrium with tridymite and liquid only at the temperature
of the diopsideliquidus maximum. However, this maximum occursat a temperatureabovewhich all diopsidetridymite tie-lines have vanished.Thus, no topological
requirementexistsfor a temperaturemaximum alongthe
diopside-tridymiteboundaryline. In this case,a temperature minimum also would not occur, and temperatures
alongthe boundary line would decreasecontinuouslyfrom
the pigeonite-diopside-tridymite invariant point toward
the join CaMgSirOu-SiOr.
The controversy could normally be resolved experimentally with little difficulty. However, temperaturegradients along the diopside-tridymite boundary line are extremely small,and the increasein temperaturecausedby
the temperaturemaximum, if it exists,would be near the
limit of temperaturereproducibility (+t "C). Nevertheless, we have carried out a careful experimental study
designedto resolvethe controversy.Two approacheshave
been used. First, we have determined the temperature
profile along the diopside-tridymiteboundary line. Second, we have analyzedthe compositionsof coexisting
glassand diopside in experimentsthat consist of glass,
tridymite, and diopside.Theseanalysesdefinea seriesof
three-phase
trianglesalongthe diopside-tridymiteboundary line. The liquid apicesofthese three-phasetriangles
must alwayspoint toward lower temperaturesalong the
boundary line and would reveal the existenceof a temperaturemaximum or minimum.
ExpBnrunNTAL METHOD
Starting mixtures (Table I and Fig. 2) were prepared
accordingto proceduresdescribedby Presnall(1966)and
Presnallet al. (1972).Ten grams of each bulk composition (Table l) were taken through two cycles of firing in
a Pt cruciblefor 2h at about 1550'C, quenchingto a
SiO2
portionof the systemMg,SiOoFig.2. The CaMgSi,Ou-rich
pressure.
Compositions
of liqCaMgSirOu-SiO,
at atmospheric
boundaryline determinedin
uids alongthe diopside-tridymite
this studyareshownassoliddots.Startingmixturesusedin this
studyareshownasopencircles.Thedashedline is thediopsidelinedetermined
by Kushiro(1972).Thepotridymiteboundary
sitionof theuppermostopensquareis basedon thecomposition
from
listedin TableI of Kushiro(1972)andis slightlydisplaced
the plottedpositionofthis point in FigureI ofthe samepaper.
boundaryline determinedin the present
Thediopside-tridymite
studyis shownasa solidline.
glass,and crushing to a powder. The glasspowders were
loadedinto Pt envelopes,suspendedin a l-atm verticalquenching furnace, and subsequentlyquenchedby dropping them directly from the hot zone ofthe furnaceinto
water. All temperatureswere measuredwith a Pt-PteoRhro
thermocouple calibrated by comparison against a standard Pt-Pte'Rh,othermocouple,which was in turn standardized by the National Bureau of Standards(NBS). All
temperatures,including those of previous investigators,
were correctedto conform to the International Practical
TemperatureScaleof 1968(Anonymous,1969).The liquidus temperature of diopside (CaMgSi'Ou)on this scale
is 1394.3"C (1391.5'C on the GeophysicalLaboratory
scale).Precisionof the temperaturemeasuremenlsis estimated to be + I 'C, and the calibration of the standard
thermocoupleis statedby NBS to be accurateto +2 "C.
The close agreementof the temperaturesof our quenching experimentswith those of previous investigations
Treu 1. Compositions
of startingmixtures(wt%)
Mixture
sio,
Mgo
Mg.SiOn
CaMgSi,O6
sio,
63.00
19.00
18.00
10.6
63.00
17.00
20.00
4.5
ov.o
19.9
It.5
18.1
63 00
20.50
16.50
15.0
63.8
2'l 2
63.00
18.00
19.00
7.5
73.4
19.0
63.00
16.50
20.50
3.0
79.2
1 7. 7
LIU AND PRESNALL: DIOPSIDE-TRIDYMITE LIQUIDUS BOUNDARY LINE
1034
3.98 and 4.02 cations and between 1.98 and 2.02 Si atoms per six oxygenatoms (Turnock and Lindsley, l98l).
wr.% coM95i2o6
1()(}
*
o
o
r
r
13
8()
d i - t r b o u n d o r y ( S c h o i r e r& K u s h i r o , 1 9 5 4 )
gl+ fr (Thisstudy)
gl+lr+dr+pl
il+ tr+di (Ihis siudy)
( K u s h i r o ,1 9 7 2 )
gl+ tr+di+pi (Thissludy)
Temperolure precision
I
*
o
o.
E
9()
13
gB e c
{
(,
<-
60
CoSiO3
50
Wr.% C o S i O 3
40
MsSiO3 +
portion
Fig. 3. Projectionfrom SiO, of the CaMgSi,Ou-rich
of the pyroxene-tridymite
boundaryline onto thejoin CaSiOrMgSiO..The plus symbolis the positionof the diopside-tridymite boundaryat the join caMgSirou-Sio,determinedby
Schairerand Kushiro(1964).The locationof the tridymite-pigeonite-diopside
invariant point is from Kushiro (1972).The
curveis identicalto that in Fig.4 andis basedon the combined
datain Figs.3 and4 (seetext).
suggeststhat the NBS accuracy limits are very conservative. Durations of experiments ranged from 72 to
93 h.
Phaseswere identified in polished sectionsby reflectedlight microscopy.Compositionsof glassesand diopsides
coexisting with tridymite were analyzed with a rEorelectron microprobe at Southern Methodist UniJXA-733
versity. All data were collectedby wavelength-dispersive
methods and reducedby the Bence-Albeeprocedurewith
the alpha lbctors of Albee and Ray (1970). Operating conditions were an acceleratingpotential of 15 kV, a beam
current of 20 nA, and a beam diameter of I pm. A glass
chip of mixture B (Table l) was used as a standard for
analysesofboth glassesand diopsides.
All crystalsand pools ofglass used for analysishave a
diameter larger than 30 pm except for some glassesand
diopsides in experiment A-4, which have a diameter of
about l0 to 25 pm. Analyses of diopsides were accepted
if (l) the sums of element oxides were between99 and
l0 I wto/oand (2) the structural formulae showedbetween
Rrvnns^q.r-ExPERIMENTS
The diopside-tridymite boundary line has been reversed for mixtures B, X, and Z (Table 2). All of these
mixtures lie in the tridymite primary phase field toward
the silica apex from the diopside-tridymite boundary (Fig'
2). On cooling,they initially crystallizetridymite and then
diopside. Reversalswere accomplishedby the method of
Presnallet al. (1978).By this procedure,the first appearance of diopside was bracketed by two experim€nts of a
specificduration,for example,72h.To obtain a reversal,
one experiment was made in which the sample was first
held above the diopside-appearancetemperature for 72
h. Then, without taking the sample out of the furnace,
the temperaturewas changedto a temperaturelower than
the diopside-appearancetemperatureand held for another
72h. Asecondexperimentwasmadein which the sample
was first held at a temperature below the diopside-appearancetemperature for 72h and then raisedto a higher
temperature for another 72 h. Successfulreversal experiments showedthe phaseassemblagepresumedto be stable at the temperatureof the secondhalf of the experiment.
This type ofreversal has the advantageofbeing independent of the nature of the starting material. The shortest
reversal time attempted was 58 h (Table 2); all experiments were held for at least 72 h.
TnNrpnn-c,ruREPRoFILE ALoNG THE
BOUNDARY
DIOPSIDE-TRIDYMITE
LINE
Two methods were used to determine temperatures
along the diopside-tridymite liquidus boundary. In one
method, temperatureswere determined by closely bracketing the first appearanceof diopside for five starting
compositions located near the boundary line but within
experiments
Tael.e3. Quenching
Expt.no.
Temoerature Duration
(h)
(qc)
Phases.
Trele 2. Reversalexperiments
A-26
A4
A-15
1376
1374
1370
74
72
74
gl +tr
gl+tr+di
tr+di
Initial condition
B-27
B-16
B-6
1377
1374
1370
13 6 1
72
73
73
93
gl+tr
gl+tr+di
gl+tr+di
tr+di
x-25
x-17
1377
1374
72
72
gl+tr
gl+tr+di+pi
Y-37
Y-31
Y-23
Y-21
1377
1375
1373
1369
72
72
72
72
gl+tr
gl+tr+di
gl+tr+di
gl+tr+di
248
z-24
1375
1372
1370
1369
72
72
72
72
gl+tr
gl+tr+di
gl+tr+di
gl+tr+di
Expt
no.
Temp.
fc)
Mirture B
8-46
1367
B-47 1380
Mixture X
x-35
1373
x-33
1379
Mixture Z
1361
z-38
1378
z-39
Final condition
6-t
Duratron
(h)
Temp.
(rc)
Duration
(h)
73
58
1380
1365
58
73
gl+tr
gl+tr+di
72
1379
1373
72
72
g t + tr
gt + t r + d i + p i
1 a70
73
72
g t + IT
gl + t r + d i
73
72
1363
Phases"
. gl : glass, tr : tridymite, di : diopside, pi : pigeonite
z-22
z-30
'Abbreviationssame as in Table 2.
LIU AND PRESNALL: DIOPSIDE-TRIDYMITE LIQUIDUS BOUNDARY LINE
the tridymite primary phasefield (Fig. 2). The results of
the quenchingexperimentsare in Table 3. Figure3 shows
the data of Table 3 projected from SiO, onto the join
CaSiOr-MgSiO..As shown in Figure 3, our resultsagree
with the resultsof Schairerand Kushiro (1964) for the
temperatureof the diopside-tridymite boundary where it
crossesthe join CaMgSirOu-SiOr.
Bowen's(1914)result,
however,is l0'C lower than that of Schairerand Kushiro
(1964).Other liquidus temperaturesdeterminedby Bowen (1914) in this region of the phasediagram are also
slightly lower than modern determinations.
In the second method, the temperature profile along
the diopside-tridymiteboundaryline was determinedfrom
electron-microprobe analysesof glassescoexisting with
diopside and tridymite (Table 4). In Figure 4, the glass
compositionsin Table 4 are projectedfrom the silica apex
onto the join CaSiOr-MgSiOr,as in Figure 3.
There is a suggestionof a temperature maximum in
Figure 3 between 45 and 500/oCaSiOr, but the temperature uncertainty is large enough that a continuously decreasingtemperaturealongthe boundaryline is also consistentwith the data. Conversely,at the same location
along the boundary line in Figure 4 (45-5Ao/oCaSiOr),
there is a suggestionof a temperature minimum. Again,
the data are also consistentwith a continuously decreasing temperature profile toward CaSiO.. The temperature
profile drawn in Figure 3 is the same as that shown in
Figure 4. It is a compromise betweenthe two setsof data
and is drawn to be consistentwith both.
LocArroN oF THE DropsIDE-TRTDyMI't.E
BOUNDARY
LINE
1035
Wt.% CoMgSi2O5
100
90
80
gl+ tr+ di+pi
( K u s h i r o ,1 9 7 2 )
El Kushiro, 1972
a This study
lJ 1390
o
:1380
o
I
Y
F
6 1370
't360
60
a:
CoSiO3
50
Wt.% CoSiO3
40
M9S|O3 ->
Flg.4. Compositions
of glasses
coexistingwith diopsideand
tridymite projectedfrom the silicaapexonto thejoin CaSiOrMgSiO,.Seealsocaptionto Fig. 3.
the one data point ofSchairer and Kushiro (1964)on the
join CaMgSirO6-SiOr.
The degreeof homogeneity of the glassesis indicated
by the standard error of the mean shown for each value
in Table 4. Most of the valuesshown are the mean of 4
to 6 individual spot analyses,but the homogeneityof
experimentY-23 wasexaminedin greaterdetail (22 spots
analyzed).The standarddeviationsfor the oxidesin this
glass(not given in Table 4) are similar to those of the
other glassesanalyzed,but the listed standarderror ofthe
mean, as expected,is considerablysmaller becauseof the
larger number of spots analyzed.Except for experiment
A-4, the listed values for the standard error of the mean
are within the rangeexpectedfrom counting statistics,so
with this one exception, compositional heterogeneityof
the glasseswas not detected.The larger uncertainties in
the analysesof glassesin A-4 may be causedby the small
areas(about 20 p.min diameter) available for microprobe
analysis.
Figure 2 shows the diopside-rich region of the system
MgrSiOo-CaMgSirOu-SiO,
with our redeterminedlocation for the diopside-tridymite boundary line. Our data
consistof electron-microprobe
analysesof the glassin six
experiments(Table 4) that contain the coexisting phases
Trrnrn-pn,lsE TRTANGLES
ALoNG THE
diopside, tridymite, and liquid. The new determination
DIOPSIDE-TRIDYMITE
BOUNDARY LINE
of the position for the boundary line removes the serpentine shapeofthe boundarydrawn by Kushiro (1972)but
Electron-microprobeanalysesof diopside coexisting
is still essentiallyin agreementwith his data points and with glassand tridymite are in Table 5. On the basis of
Taaue 4.
compositions of glasses (wt%) in equilibrium with diopside and tridymite
Expt. no :
No.ofspots:
sio,
Mgo
A-4
4
61.73(0.20)
18 . 4 0 ( 02 1 )
18.07(0.43)
B-7
6
s8.20
62 92(0.09)
16.70(0.09)
19.92(0.08)
99.s4
sio,
71.1
19.3
41
774
18.5
MgSiO3
CaMgSi,Ou
16.2*
83.8'.-
72
92 8
CaO
Total
Mg,SiO4
CaMgSi,Ou
y_29
22
8_16
7
62.62\0 12)
14.03(005)
22.75(0.061
99.41
_t
-f
62.23(0.071
18.2q0.03)
19.32(0.02)
99.74
y_3.t
S
Z_24
5
7.5
74.9
17.6
8.7
72.3
19.0
62.46(0.1
3)
16.87(0.05)
20.25(0.05)
99.s8
4.0
78.6
17.4
126
87.4
14.8
8s.2
6.9
93.1
Note; Standard error of the mean in oarentheses.
. Composition lies
outside the system MgrSiO4-CaMgSiro6-SiO,
.' Composition projected
as
from silica apex onto the join MgSiO3_CaMgSirO".
t Composition as projected from silica apex lies outside theloin MgSiO-,_CLMgSi,Ou.
63.01(0.10)
18.53(0.08)
18.75(0.04)
100.28
l 036
LIU AND PRESNALL: DIOPSIDE.TRIDYMITE
LIQUIDUS BOUNDARY LINE
CoMgSi2O6
CoM9Si2O5
o Kushiro,1972
A Carlson,1988
o This study
Me2sioa
""L]3;r, percenr
sio2
trianglesat I oCintervalsfrom
three-phase
Fig.7. Smoothed
boundaryline.
13'76to 1370"C alongthe diopside-tridymite
of diopsidesarefrom Fig. 6. Smoothed
compositions
Smoothed
Mg2SiOa MgSiO3
SiO2 compositionsof glasses
of the diopside-tridyare intersections
mite boundaryline with linesof projectionbetweenthe silica
weight percenl
in Figs.3 and 4. The
apexandthe projectedglasscompositions
invartriangleat the tridymite-pigeonite-diopside
trianglesshowingthe co- three-phase
Fig. 5. Unsmoothedthree-phase
with the
existence
of diopside,tridymite,and glassalongthe diopside- iant point at 1376'C (Kushiro,1972)is constructed
line.ThedatapointofCarlson(1988)is for compositionof diopsideat this invariantpoint determinedby
tridymiteboundary
(1988).
It is Carlson
forsterite-pigeonite-diopside.
the equilibriumassemblage
of forsteritethat thecompositions
usedhereon the assumption
diopsidesin equilibriumwith
saturated
andtridymite-saturated
pigeonitearethe same.Exclusionofthis point wouldnot change
anyof theconclusions.
The tridymite-diopside-glass tie-lines define a series of
three-phasetriangles along the boundary line whose apicesall point down-temperatureaway from the tridymitebackscattered-electronimages and the uniform compopigeonite-diopsideinvariant point toward increasinglyCasitions of the various spots ar'alyzed,no compositional
rich compositions.
zoning was observedin the diopsideslisted in Table 5.
Figure 6 shows the compositions of diopsides in equiThe raw data for compositions of diopsides in equilibrilibrium with glassand tridymite projected from the silium with liquids along the diopside-tridymite boundary
ca-apexonto the join CaSiOr-MgSiO,.Figure 7 shows
line are plotted in Figure 5. The large variation in the
smoothed three-phasetriangles along the diopside-tridyorientationsofthe tie-linesis probably causedby the exmite boundaryline at I oCintervalsfrom 1376'C at the
tremely small temperaturegradient (seeFig. 3) along the
tridymite-diopside-pigeonite invariant point to 1370 "C.
boundary line. Small temperaturefluctuations of only + 1
'C during an experiment would cause significant move- The triangles are constructed by reading the smoothed
compositionsof glassand diopsidefrom Figures4 and6,
ment of the liquid composition along the boundary line.
respectively,at each temperature.
The unsmootheddata on the three-phasetriangles(Fig.
Wr.% CoMgSi2O5
5) show clearly that temperaturesalong the diopside-trid90807
60
ymite boundary line near the CaMgSirOu-SiO,join and
also near the diopside-tridymite-pigeoniteinvariant point
O Kushlro,1972
decreaseaway from the diopside-tridymite-pigeonite inA Carlson, 1988
l, 1390
o
variant point. These data arealso consistentwith the data
a This study
r38o
of Kushiro (19'12)arld Carlson (1988). Therefore,only
i
tr
two possibilitiesexist. Either (l) temperaturesdecrease
-o
l37O
continuously from the diopside-tridymite-pigeonite invariant point or (2) a temperatureminimum and maxiI360
by Morse (1980).A single
mum both occur,as suggested
40
temperatureminimum or maximum is not possible.
50
.+cosio3
wr.% cosio3 MgSio3*
The diopside-liquid tielines at intermediate temperatures along the tridymite-diopside boundary line (Fig. 5)
Fig. 6. Compositions of diopsides coexistingwith glassand
show an apparently consistentclockwiserotation relative
tridymite projected from the silica apex onto the join CaSiOrboth higher and lower temperatures.This
MgSiOr. The open triangle and the right-hand open squareshow to tie-lines at
direction expectedif a temperatureminis
in
the
two different determinations of the composition of diopside at rotation
existed along the boundary line.
maximum
and
imum
the tridymite-pigeonite-diopsideinvariant point. Seecaption to
However, the magnitude of the rotation is not sufficient
Fig. 5 for statementregardingthe datum ofCarlson (1988).
LIU AND PRESNALL: DIOPSIDE-TRIDYMITE LIQUIDUS BOUNDARY LINE
t037
TABLE5. Compositionsof diopsides(wt%) in equilibriumwith glass and tridymite
Expt. no.:
No. of spots:
sio,
Mgo
CaO
Total
Si
Mg
Total
MgSiO3.
CaMgSirO6.
A-4
7
57.23(0.12)
23.10(0.08)
20.24(0.11)
100.57
2013
1.210
0.762
3.985
21.4
78.6
B-7
3
E-to
5
Y-23
12
56.34(0
04)
56.34(0
09)
s6.20(0.09)
2130(0.1
0)
2096(012)
22 08(0.1
1)
2273(0.20)
23 43(012].
22 32(011)
10037
100.70
100.60
Cationsper six oxygens
2.002
1.999
1.992
1.128
1. 1 0 9
1.166
u,60c
0.890
0.847
3 994
4 005
3.998
12.3
10.2
148
877
89.8
852
Y-31
4
s6.74(0.1
s)
22.06(0.1
8)
21.90(0.12)
100.69
2.004
1161
0.828
a ooe
15.7
84.3
z-24
5
s6.2s(0.05)
21.33(0.09)
22.73(O.11)
100.31
2.000
1. 3 1 0
0.865
a ooA
12.4
87.6
Note.'Standarderror oJ the mean in oarentheses.
- Projection pyroxene
of
compositionsfrom silicaapex onto the join MgSiOs-CaMgSi,Ou.
to causea reversal in the orientation of any of the unsmoothedthree-phasetriangles(seeFig. 5), as would be
required if a temperatureminimum or maximum occurred. Some of the three-phasetriangles are very narrow, and the composition and temperature uncertainties
are such that the existence of a temperature minimum
and maximum cannot be positively eliminated. However, becausenone of the unsmoothedthree-phasetrianglesshow a reversal,we arguethat the dataare fit best
by a continuously decreasing temperature along the
boundaryline.
RnrnnBNcnscrrBo
Albee, A.L., and Ray, L. (1970) Correction factors for electron-probe
microanalysisof silicates,oxides,carbonates,phosphatesand silicates
Analytical Chemistry, 42, 1408-14 14.
Anonymous. (1969)The intemational practicaltemperaturescaleof 1968.
Metrologia,5,35-44.
Bowen, N.L. (1914) The ternary systemdiopside-forsterite-silica.American Journalof Science,38,207-264
Carlson, W.D (1988) Subsolidusphaseequilibria on the forsterite saturated join MgrSi,O.-CaMgSirOuat atmospheric pressure. American
Mineralogist,73, 232-24l.
Kushiro, I (l 972) Determination of liquidus relationson syntheticsilicate
systemswith electron probe analysis:The system forsterite-diopsidesilica at I atmosphere.American Mineralogist, 57, 1260-1271.
Kushiro, I., and Schairer,J.F. (1963) New data on the systemdiopsideforsterite-silica.CarnegieInstitution ofWashington Year Book 62, 95103
Longhi, J (1987) Liquidus equilibna and solid solution in the system
CaAlSiOr-MgrSiOo-CaSiOySiO, at low pressure.American Journal of
Science.287. 265-331
Longhi, J., and Boudreau, A E. (1980) The orthoenstatite liquidus field
in the system forsterite-diopside-silicaat one atmosphere.American
Mineralogist, 65, 563-573.
Morse, S.A. (1980) Basaltsand phasediagrams, 493 p. Springer-Verlag,
New York.
Presnall,D C (l 966) The join forsterite-diopside-iron oxide and its bearing on the crystallization of basalticand ultramafic magmas.American
Journal of Science,264, 7 53-809.
Presnall,D.C., Simmons,C.L., and Porath,H. (1972)Changesin electrical conductivity of a synthetic basalt during melting. Joumal of GeophysicafResearch.77, 5665-5672.
Presnall,D.C., Dixon, S.A.,Dixon, J.R., O'Donnell,T.H., Brenner,N.L.,
Schrock,R L., and Dycus, D.W (1978) Liquidus phaserelationson
the join diopside-forsterite-anorthitefrom I atm to 20 kbar: Their
bearing on the generationand crystallization ofbasaltic magma. Contributions to Mineralogy and Petrology,66,203-220.
Presnall,D.C., Dixon, J.R., O'Donnell, T.H., and Dixon, S.A (1979)
Generation of mid-ocean ridge tholeiites Joumal of Petrology,20, 3-
Pnrnor,ocrcAl AppLrcATIoNS
Morse (1980, p. 212) suggested
that the absenceof a
temperatureminimum along the diopside-tridymite
boundaryline would lead to CaSiO, enrichmentin fractionatingliquids.He pointedout that igneousassemblages
containingwollastoniteare rare in nature,which is consistent with the presenceof a temperatureminimum.
However, the lack of wollastonite-bearingassemblages,
even in the absenceof a temperatureminimum, is easily
explainedwhen other componentsare added that make
the ternary system MgrSiOo-CaMgSirOu-SiO,
more directly applicableto natural compositions.For example,
Presnalletal. (1979)showedthat fractionalcrystallization
in the CaMgSirOu-MgrSiOo-CaAlrSirOr-SiO,
tetrahedron
yields a final residual liquid at the enstatite-quartz-anorthite-diopsidequaternaryeutecticpoint (seealso Longhi, 1987).This resultis independentofthe presence
or
absenceof a minimum on the diopside-tridymiteboundary line. Therefore,the systemMgrSiOo-CaMgSirOu-SiO,
is not adequateto constrainthis aspectof the fractional S.irr.r, J.F., and Kushiro, I. (1964) The join diopside-silica.Carnegie
crystallizationof natural liquids, and the absenceof a
Institution of Washington Year Book 63, 13U132.
Schairer,J.F., and Yoder, H.S., Jr. (1962) The systemdiopside-enstatitetemperatureminimum is of no petrologicalconcern.
AcxNowlBocMENTs
This researchwas supported by National ScienceFoundation Grant
EAR-8418685 to D C Presnall.We thank D. Deuring for his assistance
in the microprobe analyses.
silica CamegieInstitution of Washington Year Book 61, 75-82.
Turnock, A.C., and Lindsley, D.H. (1981) Experimentaldetermination of
pyroxenesolvi for = I kbar at 900'and 1000'C.CanadianMineralogist,
1 9 .2 5 5 - 2 6 7.
Memrscnrrr REcETvED
JuNe l, 1988
MrNuscmrr AccEprEDMev 17, 1989