American Mineralogist, Volume 72, pages 12-28, 1987

Compositions of anhydrous and hydrous melts coexisting withplagioclase, augite, and olivine or low-Ca pyroxene from

I atm to 8 kbar: Application to the Aleutianvolcanic center of Atka

DoN R. BlrnRrr D.c.vro H. Eccr,nnDepartment of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A.

Ansrru,cr

Phase relations and partial melt compositions have been determined for a suite of high-alumina basalts and andesites from the Aleutian island of Atka and from the Marianaislands. Experimental conditions varied from anhydrous, at pressures from I atm to 8kbar, to hydrous (2o/o HrO in the melt) at 2 and 5 kbar, although not all compositionswere studied at all conditions.

Partial melt compositions at 1 atm, projected onto pseudoternary phase diagrams, definephase fields for plagioclase, augite, olivine, and orthopyroxene. Hydrous partial melts, 2wto/o HrO in the melt, aI2 and 5 kbar define equivalent phase fields. At all those conditions,the four phases coexist with melts of andesitic composition, and olivine reacts with thosemelts to produce orthopyroxene. Anhydrous partial melt compositions at 8 kbar, however,define phase fields for plagioclase, olivine, augite, and pigeonite. Phase compositions in-dicate the presence of thermal divides on both the plagioclase-olivine-augite and plagro-clase-augite-pigeonite phase boundaries.

Crystallization fractionation at deep-crustal pressures of 8 kbar can only account for aportion of Atka basalt compositions, but it cannot account for the total array of Atka rockcompositions, because of the presence of a thermal divide involving pigeonite cyrstalli-zation. Even though the olivine reaction relation has been demonstrated to occur, lower-pressure anhydrous fractionation cannot account for andesitic and dacitic melts. In reality,all Atka rock compositions are considerably more aluminous (more plagioclase-rich onpseudoternary diagrams) and less Fe-enriched than low- or high-pressure anhydrous mul-tiply saturated melts. That mismatch can be alleviated, at least for andesites and dacites,on the hypothesis that andesite and dacite melts containing about 2o/o H2O have evolvedfrom basaltic andesite parents in crustal magma chambers (2-5-kbar pressure). The arrayof Atka basalt to andesitic basalt compositions is also consistent with crystallization frac-tionation at 5-8-kbar pressures and low HrO contents, although experimental details arecurrently lacking.

INrnooucrroN

Andesites may represent partial melts of either mantleperidotite (Mysen, 1982) or subducted oceanic slab (Greenand Ringwood, 1968), mixing of basaltic and daciticmagmas (Eichelberger, 197 5), contamination of basalticmagmas by crustal rocks, or crystallization fractionationof basaltic magmas. Both hydrous (Kay et al., 1982) andanhydrous (Osborn, 1959,1979; Gill, l98l) fractionatingcrystal assemblages have been proposed. The former as-semblage contains hornblende (Kay et al., 1982). Am-phibole is, however, a rare phenocryst in andesitic rocksfrom oceanic island arcs (Gill, I 98 I ; Ewart, 1982; Marsh,1982), but is more common in andesitic rocks eruptedthrough continental crust (Coulon and Thorpe, 198 l;

t Present address: Department of Geology, Rensselaer Poly-technic Institute, Troy, New York 12180-3590, U.S.A.

0003404x/87l0102{012$02.00 12

Ewart, 1982; Leeman, 1983). The anhydrous assemblageconsists of plagioclase, olivine, clinopyroxene, and pos-sibly magnetite (Gill, 1981).

Hydrous experiments performed at crustal pressureshave primarily investigated amphibole crystallization inrocks of basaltic composition (Holloway and Burnham,1972;Helz, 1976; Cawthom, 1976:, Cawthorn et aI., 1973;Allen et al., 19751' Allen and Boettcher, 1978) with watercontents greater than approximately 4 wto/o. Experimentalresults provide some support for the hydrous crystalli-zation fractionation hypothesis, although, unlike mostandesites, experimentally produced andesitic melts aretypically peraluminous (Ewart, I 982).

Anhydrous crystallization fractionation has been in-vestigated by Grove et al. (1982) and Grove and Bryan(1983); these investigators located a reaction point at Iatm involving olivine, augite, plagioclase, pigeonite, and

BAKER AND EGGLER: ATKA MELT COMPOSITIONS

Table 1. Compositions of basalts and andesites studied

Ar-1121 Ar-41t AT-29t Ar-2st M-461 AG8-4f

1 3

AT-1t P-229

sio,Tio,Al,o3FeO-MnOMgoCaONaroKrOP.ou

Total

50.41 .29

19.79.080.285.059.593.420.790.31

gE-ST

4 9 01.02

19.09.470216 . 1 8

11.42.56U . O J

0 . 1 1gg50

54.70.79

18.77 9 00.313.297 0 4

4 . 1 00.99025

gFSg

56.81 .01

16 .98.030 .173 0 97 0 5q q a

2050.28

9937

57.20,92

17.16.980.213.327.O73.902.470 .16

9935

52.41.33

14.513.00.434.309.033.150.85

96-5B

57.40.88

15.99.860.222.4'l6.293.881.880.28

9632

60.10.81

16.08.800.221.875.814.252.060.35

166fr

51.62.05

15.011.70.225.30

10.763.020.520.0

TOO:T7. All Fe determined as FeO.t Microprobe analyses of superliquidus or nearliquidus glasses.+ XRF analyses from Stern (1979).$ Microprobe analysis trom Walker et al. (1979).

melt that apparently prevents basaltic compositions fromproducing andesitic residual melts by crystal fractiona-tion.

To better understand andesite petrogenesis beneath theAleutian island of Atka, compositions of partial melts ofhigh-alumina basalts and andesites were determined atpressures from I atm to 8 kbar under anhydrous condi-tions and from 2 to 5 kbar at water-undersaturated con-ditions, extending the study ofBaker and Eggler (1983).

S.lnnpr,r DEScRrprIoN AND pREpARATToN

Sample description

Rocks studied represent a suite of high-alumina basalts andandesites from the Aleutian island of Atka (Marsh, 1982), var-ious islands in the Mariana island arc, and one sample from theOceanographer fracture zone (Table 1).

Atka rocks. Samples AT-1, AT-112, and AT-41, which aresimilar to other high-alumina basalts (Kuno, 1960), all containphenocrysts of plagioclase, olivine, and magnetite; AT-41 alsohas phenocrysts of clinopyroxene. Samples AT-25 and AT-29,which are typical oceanic island arc andesites (Ewart, I 982), con-tain phenocrystic plagioclase, clinopyroxene, and orthopyrox-ene.

Mariana and Oceanographer fracture zone rocks. The basaltfrom the Marianas, M-46-10B (hereafter referred to as M-46),is from Esmerelda Bank (Dixon and Stern, 1983), a seamountthat lies behind the volcanic arc. The andesites, AG7-l and AG8-4, are from the island ofAgrigan (Stern, 1979) and are typicalof the andesites found in the Marianas (Dixon and,Batza, 1919).The Oceanographer fracture zone rock is a tholeiitic basalt, P-22;l-atm melting experiments have previously been performed onthis rock by Walker et d. (1979).

Mineral separates, Mineral separates (Table 2) were used in"forced saturation" experiments. Minerals used were olivine fromthe Duke Island ultramafic complex of southeast Alaska (Irvine,1974), plagioclase from Crystal Bay, Minnesota (U.S. NationalMuseum no. R2912), and augite from the Frosty Peak volcaniccomplex on the Alaskan Peninsula (collected by D.R.B.). Anal-yses ofthese minerals are presented in Table 2.

Sample preparation

Rock samples. Rocks were received as chips (AT-1, AT-25,A^t-29, AT-41, AT-112, P-22) or as powder of approximately-200 mesh (M-46, AG7-1, AG8-4). Rock chips were crushed

and ground in a tungsten carbide-lined ball mill until greaterthan 750/o of the grains were less than 5 pm in their longestdimension (Baker, 1985). Rock powders were stored in either adrying oven or dessicator.

Mineral separates. Mineral separates were crushed in a per-cussion moruar and ground for I to 2 h under ethanol with anagate mortar and pestle. The size fraction less than 400 meshwas used to make starting mixes. Mixes of combined mineralseparates were then ground together for I h under ethanol withaL alate mortar and pestle and stored in either a drying oven ordessicator.

ExprnruBNTAL TECHNTeUES

Anhydrous experiments

One-atrnosphere experiments. One-atmosphere experimentswere performed in either silicon carbide-heated (PennsylvaniaState University) or platinum-wound (Smithsonian Institution,experiments 3,4, ll, and l4) gas-mixing furnaces in which CO,and H, gases were used to control the oxygen fugacity. Rockswere melted on pretreated Pt loops (Baker and Eggler, 1983;Baker, 1985). Other experimental techniques were the same asMahood and Baker (1986) with the exception that all experi-ments were performed within 0.3 log units of the nickel-nickeloxide buffer [NNO).

Durations of experiments varied from a few hours at near-liquidus temperatures to 480 h at 1060"C. Experiments of 24-and 480-h duration at 1060"C on andesite AT-29 did not vary

Table 2. Compositions of mineral separates

Plagio-Olivinet clase* Augite+

37.10.00.2

17.40.1

44.2

. All Fe determined as FeO.t Analysis from lrvine (1974).+ Microprobe analysis.

sio,Tio,Alro3FeO-MnOMgoCaONa.OKrO

Total

49.70.06

31.00.400.120.04

15.32.470.07gg:re

50.20.813.567.86o.24

14.721.10.240.0

967i

0.30.0

95-3

l 4 BAKER AND EGGLER: ATKA MELT COMPOSITIONS

in melt composition and phase assemblage (App. Table 1, ex-periments 1 I and l4), indicating that andesites and basalts canreach steady state, ifnot equilibrium, with respect to both phaseassemblage and melt composition within 24 h.

Solid media apparatus experiments. Experiments at 8 kbar andhigher pressures were performed in a single-stage solid mediaapparatus (Boyd and England, 1960). Talc-Pyrex<rushable alu-mina assemblies, 2.54-cm diameter, were used. Graphite cap-sules inside 3-mm Pt capsules were used to contain rock powdersin all experiments for which partial melt compositions were de-termined (App. Table 1). Loaded capsules were dried at 110'Cfor I h and fired to red heat for 30 s before welding. Furtherdetails of encapsulation techniques and pressure-temperaturecontrol are discussed in Baker (1985) and Mahood and Baker(r 986).

Runs at 8 kbar varied in duration from t h near 1300"C to196 h near I 100"C. Experiments performed by partially meltingrock powders ("melting experiments") or by completely meltingrock powders and then lowering the temperature to the sametemperature as the partial melting experiments ("crystallizationexperiments") produced the same phase assemblage and similarphase compositions for rocks AT-1 (App. Table l, experiments891 and 953) and AT-29 (App. I'able 1, experiments l0l0 and989). These experiments suggest close approach to equilibriumalthough near-solidus phase proportions and compositions pro-duced in different experiments om rock AT-112 were inconsis-tent (App. Table l, experiments 940 and 1098).

For the 8 kbar "forced saturation" experiments, rock powderswere placed between layers of a plagioclase-augite-olivine mix-ture (Takahashi and Kushiro, 1983; Baker, 1985). A series ofvariable duration experiments, 24 to 96 h at 1200'C, producedsomewhat variable melt compositions (App. Table l), appar-ently due to inhomogeneous Fe distribution in starting materials;nevertheless, melts appear to reach a steady-state compositionwithin 24 h (Baker, 1985).

Hydrous experiments

Internally heated pressure vessel techniques. Experiments wereperformed at 2 and 5 kbar in internally heated pressure vessels(Holloway, 1971). Temperature was monitored with two Pt-PtnoRh,o thermocouples calibrated against an NBS-certified ther-mocouple and held within 5 deg of the desired temperature. Thethermal gradient along the samples was below 3 deg. No pressurecorrection was made on the thermocouple emf. Pressures, mon-itored with manganin cells, were precise to 60 bars and believedaccurate to 100 bars. Oxygen fugacity was controlled within 0.7log units of NNO by controlling H, partial pressrre using a Ptmembrane; this technique was checked against NNO and fay-alite-magnetite-quartz (FMQ) buffer capsules containing HrO-CO, mixtures (Baker, 1985). Experiments were quenched at ap-proximately 500'C/min.

Capsules were 3-mm diameter and either AgroPdro or Pt, whichwere treated with wiistite (Helz, 1978; Baker, 1985) in order tomitigate Fe loss from the samples. Despite these precautions, Feloss occurred in the 2- and 5-kbar experiments (App. Table 1).

"Melting experiments" were at 2 and 5 kbar, except for oneforced-saturation experiment. Experiments varied in durationfrom 2 h at superliquidus conditions to 96 h at temperaturesnear 1000"C. In experiments performed on AT-25 and AT-29at 1020'C for both 24- and 96-h duration, the phase assemblagewas the same and compositions of AT-25 partial melts (App.Table l, experiments ll47 and 1248) were similar. Composi-tions of AT-29 partial melts were different (App. Table 1, ex-

periments I 148 and I I 74), presumably as the result of variableamounts of Fe absorption by the capsule.

Control and calculation ofHrO contents ofexperimentally pro-duced melts. HrO contents of experimentally produced meltswere controlled by changing compositions of HrO-COr fluidspresent in experiments (cf. Eggler, 1972; Baker and Eggler, 1983).Capsules were loaded with oxalic acid dihydrate and silver ox-alate in correct proportions to yield a fluid with mole fraction ofHrO near 250/o (2-kbar experiments) or 180/o (5-kbar experi-ments). Fluid weight was kept approximately equal to rock pow-der weight to minimize changes in the fluid composition due todiferential melt solubilities of HrO and CO, (Holloway, 1976).Capsules were welded with no loss of volatiles.

After each experiment capsules were weighed; many capsuleswere very brittle due to Fe absorption and lost fluid duringquench. The fluid HrO:CO' ratio in unfailed capsules was de-termined by a weightJoss technique (Baker and Eggler, 1983).

For capsules that did not fail, HrO contents of melts can becalculated from fluid compositions by assuming HrO activity-composition models for silicate melts (Burnham, 1979, l98l)and for HrO-CO, fluids. An ideal-mixing model was used forfluids, in part because the system quartz-sanidine-HrO-CO, sug-gest near-ideality (Bohlen et a1., 1983) and in part because amore complicated Redlich-Kwong model (Holloway, 1977, pers.comm.) yields nearly identical activities.

For failed capsules, H2O contents can still be calculated byiterating the above procedure with the constraint of mass bal-ance (Merzbacher and Eggler, 1984; Baker, 1985). The iterativeprocedure can also be performed for the subset ofcapsules thatdid not fail, serving as a check of the melt and fluid activity-composition models and measurement techniques. In that check,measured and calculated fluid compositions are entirely com-parable.

As a test of these techniques, the liquidus of AT-l +2.5 wt%HrO was determined at 8 kbar in the solid media apparatus,both when HrO was added directly and when silver oxalate +HrO was added in the correct proportions to produce a meltwith the desired HrO content. Agreement of the liquidus deter-minations, within l0 deg, for both methods of H'O additionindicates an uncertainty of 0.35 wto/o in the calculated HrO con-tent of the melt.

Analytical techniques

Selected run products were analyzed by electron microprobe,at Pennsylvania State University and the National Museum ofNatural History, Smithsonian Institution, at 15-kV acceleratingpotential and 0.012-pA sample current using the data-reductiontechniques of Bence and Albee (1968) with matrix-conectionfactors of Albee and Ray (1970).

Run products were imaged using sEM attachments on micro-probes, and crystals were analyzed with a 2-pm beam. Olivinesand most pigeonites were compositionally homogeneous. Augiteand plagioclase crystals were invariably heterogeneous, even incrystallization experiments. Except where quench overgrowthson the rims of crystals were observed, rim compositions (anal-yses made 5-15 pm from the rim of the crystal) are presented.

Quench overgrowths were only apparent (by secondary-electronimaging and quantitative analyses) in a few experiments per-formed in the internally heated vessel.

To minimize alkali loss during analysis, 1G-20-pm diameterbeams were used to analyze glass in run products wherever pos-sible. Comparison between microprobe analyses of AT-29 andAT-25 superliquidus glasses and wet chemical analyses (8. D.

o

J

o

o| ' '0N >, , t< E

oo

ooo,l

oo

o -

o r !:E

F

ooc)

oo(\

ooo

ooo

ooo)

oor,

oo6l

oo

ooo,ooa)

ooc.l

ooO)ooct

eoo

ooo)

o( \ t O @ ( o ! l Noo

ooo c L

o-

ooo

l r_--o------

^ q

J

j 6

\o o+ ' ;

:-ai

J 1-:--::

o

o

osiF C< G

o6

o

o

oo6l

oo

Ooo c iF FF d

F

ooo

ooO)oo

t -

o J

q a x+ . Y ' q

J o O

+

ah

o

< f , >, , E

AF

j a - t . P - r Et .r ?= t iJ O J : J O

E.9 to 9)J +. - - - . - - -_ : '

-- \-_--r--t--->';--+- a 6 , x

+ f FJ N V

snplr;--

tt

oo)t tN >, ' GF F

< d

rc

r c t

++d 3 , .J d -

o

o

ott' ; > r

F - C<E

oo

leq r 'd

l 6 BAKER AND EGGLER: ATKA MELT COMPOSITIONS

Marsh, pers. comm., 1980) indicate the success of using theselarge-diameter beams in minimizing, or even eliminating, alkaliloss.

Weight percent modes of some experiments were calculatedusing bulk-rock compositions and the compositions of run-prod-uct phases as input to a least-squares mixing program (Bryan etal., 1969). Fe loss in hydrous experiments was calculated byusing wtistite as a potential phase. Modes for such experimentswere corrected for Fe loss by normalization of silicate modes toa total free of modal wiistite.

ExpnnnrnNTAL RESULTS

Compositions of multiply saturated melts and coexist-ing crystals at pressures from I atm to 8 kbar are listedin Appendix l. Additional experimental data used for thecon5truction of phase diagrams are presented in Baker( l e85) .

Anhydrous phase relations

High-alumina basalts. High-alumina basalts AT-1, AT-41, and AT-l12 have similar phase relations (Fig. l),except for the lack ofclinopyroxene in the l-atm phaserelations of AT-41. At I atm along the NNO buffer, pla-gioclase is the liquidus phase followed by olivine, thenclinopyroxene. The clinopyroxene that crystallizes fromAT-l and AT-ll2 compositions is an augite. At S-kbar,plagioclase remains the liquidus phase for all composi-tions studied. The plagioclase liquidus for AT-l in "crys-tallization experiments" was 25 deg below the liquidustemperature of plagioclase in "melting experiments,"probably due to inhibited plagioclase nucleation at near-liquidus temperatures (Gibb, I 974).

The l-atm clinopyroxene that crystallized from AT-lis augite (Wo.nEnnrFs,n; Baker and Eggler, 1983), whereasat 8 kbar, pigeonite (Wo,rEnu.Fsro) crystallized. Pigeonite(Wo,oEnroFsoo) also crystallized from AT-41 at 8 kbar.Clinopyroxene that crystallized from AT-ll2 at 8 kbar ison the boundary between augite and subcalcic augite(WorrEnruFs,r). Changes in pyroxene chemistry in AT-lare due to the expansion of the low-Ca pyroxene primaryphase field, relative to augite, as pressure increases (cf.Kushiro, 1969).

Andesites. At I atm, Atka andesites AT-25 and AT-29exhibit similar phase relations (Fig. l): plagioclase fol-lowed by olivine and augite (WorrEno.Fs,n in AT-29). Or-thopyroxene and magnetite join the sequence, and olivinedisappears at temperatures near I 100"C.

At 8 kbar, plagioclase is still the liquidus phase for AT-29 (Fig. l), but crystallizes at approximately I163'C, atemperature below the l-atm plagioclase liquidus. Thisdiscrepancy is attributed to difficulties in plagioclase nu-cleation; the estimated 8-kbar plagioclase melting tem-perature is I 190'C. At ll25C, plagioclase, melt, and twoclinopyroxenes (Wo,oEnrrFs' and WoroEnouFsro) coexist;this phase assemblage is present in both "melting" and"crystallization" experiments at I 125"C. Because micro-probe analyses were only made of AT-29 mn products,it is not known which clinopyroxene(s) is present in AT-25 at I150'C.

At 8 kbar, plagioclase is the liquidus phase (at approx-imately 1190"C) for Agrigan andesites AG7-l and AG8-4 followed by pigeonite (WorEn,rFsrr, -1160'C). Theseminerals crystallize from AG7-l and AG8-4 to the lowesttemperatures investigated at 8 kbar (1050'C).

Hydrous phase relations

Previously published phase relations of high-aluminabasalt AT- I at 2 kbar (Baker and Eggler, I 983) are heresupplemented by experiments on Atka rock compositionsat 2 and 5 kbar with -2 wro/o HrO in the melt. Owing tovariable Fe loss, phase relations determined in this studyare approximate, and melting temperatures of maficphases should be considered as maxima.

The liquidus mineral of AT-l and AT-41 at 2 kbar and2 wto/o HrO is plagioclase (Fig. l), followed by olivine(AT-1) or augite (AT-41, WornEnorFs'.). (Exact liquidustemperatures were not determined.) Orthopyroxene ap-parently crystallizes from both high-alumina basalts be-tween 1030'C (AT-41) and 1020"C (AT-l). Andesites AT-25 and LT-29 also have plagioclase as the liquidus phaseat 2 kbar (Fig. l), followed by augite (WoooEnorFs,'). AT-29 crystallizes orthopyroxene (WorEnunFsr') but not oliv-ine. AT-25 crystallizes olivine at 1080"C and possiblyorthopyroxene at 1030'C. (Orthopyroxene was found ina 1020C run in which greater than 500/o of the originalFe content was lost to the capsule.)

Anhydrous liquid lines of descent

Compositions of a limited number of experimentallyproduced partial melts can be compared to a wide rangeof natural rock compositions (presumed to be melt com-positions) in pseudoternary diagrams (O'Hara, 1968). Thiscomparison has been very useful in elucidating the petro-genesis of mid-ocean ridge basalts (MORBs) (Walker etal., 19791- Grove and Bryan, 1983).

Pseudoternary diagrams used in this study (cf. Bakerand Eggler, 1983) are a modification of the diagrams cre-ated by Walker et al. (1979). Baker and Eggler (1983)combined normative orthoclase with quartz as a "mag-maphile" component (SiOr) enriched in residual melts,whereas Walker et al. (1979) combined orthoclase withplagioclase and Grove et al. (1982) used orthoclase as acomponent. In both cases, projection from Plag + Or orfrom Or incorrectly implies saturation of residual meltswith orthoclase.

One-atrnosphere liquid lines of descent. Compositionsof partial melts (AT-25, AT-29, AT-41, and AT-l12 atI atm along the NNO buffer, App. Table l; Baker andEggler, 1983; Eggler and Osborn, 1982), allow subdivi-sion of a plagioclase-saturated pseudoternary diagram intofields ofplag + ol, plag + aug, and plag + opx (Fig. 2).

At approximately I100'C, olivine undergoes a reactionrelation with the melt, wherein orthopyroxene and mag-netite, without olivine, crystallize from andesitic bulkcompositions AT-25 and AT-29. Melts coexisting withplag * aug + opx + mt are significantly enriched in silicaand depleted in FeO*, compared to melts formed 20 deg

BAKER AND EGGLER: ATKA MELT COMPOSITIONS

o tm

t h i s s t u d y

C i n r

a r . ( 1 9 7 9 )

a l . ( 1 9 8 2 )

Fig. 2. Pseudoternary diagrams at I atm. Pseudoternary diagrams constructed using the techniques of Baker and Eggler (1983)with the exception that the Qz + Or component has been renamed SiOr. (a) (Plag + Mt)-saturated pseudoternary with analyses ofglasses produced in the l-atm experiments along the NNO buffer. (b) (Di + Mtfsaturated pseudoternary diagram; +, glassescoex is t ingwi thp lag+o l *aug; t r ,g lassescoex is t ingwi thp lag+aug+opx+mt .The long-dashed l inesare thep lag+o l+augLLMSs determined by Walker et al. (1979) and Grove et al. (1982). The intersection of the short-dashed and long-dashed lines inFig. 2b is the intersection of the plag + ol + opx LLMS and the plag + ol + pig LLMS determined by Grove et al. (1982). Thearrows on the LLMS point in the direction of decreasing temperature. (c) (Plag + MtFsaturated diagram based on the results ofWalker et al. (1979) and Grove et al. (1982); +, melts coexisting with plag + ol + aug; r, melts coexisting with plag + aug + pig;*, melts coexisting with plag + ol + aug + pig, O, melts coexisting with plag + ol + pig; crosses within diamonds, melts coexistingwith plag * ol * opx.

t 7

higher and coexisting with plag * ol + aug, presumablybecause of magnetite crystallization (Osborn, 19 59, l9'1 9).

The coexistence of augite, orthopyroxene, plagioclase,and andesitic melts (Fig. 2a) differs from the results ofGrove et al. (1982) and Grove and Bryan (1983). Theseauthors found that the plag + ol + aug liquid line ofmultiple saturation (LLMS) is terminated by a reactionpoint where ol + melt react to form pig + aug + plag. Afield ofpig + plag separates the fields ofaug + plag andopx * plag (Fig. 2c). Grove et al. (1982) stressed thatbasaltic melts become trapped at the reaction point anddo not undergo silica enrichment, but instead becomeenriched in Fe (cf. Fenner, 1931).

We attribute differences in phase diagrams (Figs. 2a,2c) to differences in bulk compositions and in particular

to differences that are not displayed in pseudoternary dia-grams. The two chief such differences are KrO contentsand Mg/(Mg * Fe2+), which are plotted versus SiOr forthe various LLMS (Fig. 3).

Values of Mg/(Mg + Fe'z+) in melts of this study areintermediate between values of Grove et al. and Walkeret al. for equivalent phase assemblages, and hence Mg,/(Mg + Fe'z+) differences are probably not significant. Therelatively small (and overlapping) differences in Mg/Mg +Fe'z+) also suggest that the slightly different oxygen fugac-ities in the sets of experiments (FMQ for Walker et al.and Grove et al.; NNO for this study) are not significant,because such diferences should principally change Mg/(Mg + Fe'z+) of melts at any particular SiOr content.

It is apparent, however, that melts of Walker et al.

P l o g

G r o v e

. p l ol aug Walker and Grove^ pl ol opx Groveo pl ol aug lhis study o

, 'ooo pl opx aug mt this study ,

. pl aug pig Walker , 'o

. pl ol pig Grove o /

this sludy

Walker et al.

/ . .l - - > ' * - r

D .-/14... - - 1 ; .

,t Grove et al.

1 8 BAKER AND EGGLER:ATKA MELT COMPOSITIONS

Table 3. Fe-Mg partitioning between crystalsand melts at 1 atm and 2 kbar with fo,

near the NNO buffer

Ko:

[(re,JMg)""","]/[(Fer",/Mg)*,t]

s.d tO r

Y

1

. pl ol aug Walker and Grover Pl ol opx GroveE Pl ol auE this studyo Pl oPx aug mt this study. Pl aug Pig walker. pl ol pig Grove

\ Grove et a. .

t

Walker el al.

SiOr

Fig. 3. KrO and M/(Mg + Fe'*) versus SiOr values of l-atmmelts. (a) The KrO content of melts along the plag + ol + aug,plag + ol + pig, plag + ol + opx, plag + aug + pig, and plag +aug + opx + mt LLMS are plotted against SiOr values, from thePlag + Mt projection. (b) Mg/(Mg + Fe2*) values [zof the Mg/(Mg + Fe,.,)l of melts plotted against SiOr values. The amountofFe2* was calculated using the method ofSack et al. (1980) atthe same conditions oftemperature and oxygen fugacity as usedto construct the pseudoternary diagrams. "Walker"-data fromWalker et al. (1979); "Grove"-data from Grove et al. (1982).

(1979) and Grove et al. (1982) are less potassic than theandesitic melts used in this study; their melts were in factMORBs and an unusual high-alumina basalt anomalous-ly low in KrO (0.09 wto/o) compared to typical high-alu-mina basalts (cf. Grove et al., 1982; Kuno, 1960). Weattribute the differences in phase diagrams to these dif-ferences in KrO.

Although a rationale of the different phase diagrams isnot necessary to our arguments, we propose that the phasediagrams reflect different intersections of solidus surfaceswith pyroxene phase relations. An analogy in a syntheticsystem is the demonstration by Kushiro (1969) in CaO-MgO-SiOr-HrO that the assemblage Fe-free pigeonite +diopside crystallizes at 20 kbar under anhydrous condi-tions but that orthopyroxene + diopside crystallizes un-der HrO-saturated conditions. The difference lies in the

OlivineAugiteOrthopyroxene

OlivineAugiteOrthopyroxene

AugiteOrthopyroxene

AugiteOrthopyroxene

1 atm4 0.299 0.215 0 .192kbal5 0.298 0.262 025

1 0 . 1 71 0 .18

2 kbar: experiment 1 1 74| 0 .14r 0 . 1 5

2 kbar: exoeriment 1148

0.030.050 0 3

0.020.020.04

0

oL

== 0 4

. Number o{ experiments used to calculate the meanand standard deviation.

f One standard deviation of the calculated meanoartition coefficient.

failure of the lower-temperature HrO-saturated solidussurface to intersect the subsolidus pigeonite field becausetemperatures are below the reaction pigeonite : augite *orthopyroxene (Lindsley, 1980). By analogy, we propose

that more potassic liquids, with lower liquidus tempera-tures, crystallize orthopyroxene. Higher potash contents,independent of temperature, may also affect low-Ca py-

roxene compositions (cf. Kushiro, 1975).Eight-kilobar liquid lines of descent. In comparison to

I atm, fields of aug + plag and low-Ca pyroxene + plaghave expanded at the expense ofthe ol + plag field (Fig.4a); plag + aug stability remains virtually unchanged (Fig.4b). Pigeonite is the stable low-Ca pyroxene at 8 kbar,presumably because of high Fe/Mg ratios of residual melts(cf. Fujii and Bougault, 1983).

Phase relations (Figs. 4a,4b) were constructed from allthree types of experiments- "melting," "crystallization,"and "forced-saturation." At 8 kbar, residual melts of AT-l, AT-l12, and AT-41 do not undergo SiO, enrichment,unlike residual melts at I atm (App. Table l; Figs. 4a,4b). Tie lines between coexisting crystals and melts (Fig.

5) indicate why no SiO, enrichment occurs: there is athermal divide on the LLMS for plag + aug + pig and apseudoternary eutectic for plag + ol + aug + pig. Mod-ally, the lack of SiO, enrichment is reflected in lower pro-portions ofolivine and higher proportions ofclinopyrox-enes in crystallizing assemblages.

The topology ofthe phase diagram (Fig. 4a) is consis-tent with the system olivine-diopside-silica under anhy-drous conditions at 20 kbar (Kushiro, 1969). The LLMS(Figs. 4a, 4b) are, moreover, consistent with literaturedata provided that melt compositions are selected forwhich Mg/(Mg + Fe,",) values are less than 0.55, like themelts in this study. Such data include melting of tholeiites(Thompson, 1974; Stolper, 1980) and an alkalic basalt

1 0 0

BAKER AND EGGLER: ATKA MELT COMPOSITIONS

8 kbo r

t h i s s t u d y

Mg '< o . ss

3 kbo r

l i t e r a t u r e

S i O r O l o p x

d

Fig. 4. 8-kbar pseudoternary diagrams. (a) Plag + Mt pseudoternary; (b) Di + Mt pseudoternary; +, (plag + ol * aug)- and(plag + ol + pigfsaturated melts; n, (plag + aug + piglsaturated melts; A, (plag + pigFsaturated melts. Solid line, 8-kbar LLMSs( m e l t + p l a g + o l + p i g , m e l t + p l a g + o l * a u g , m e l t + p l a g + a u g + p i g ) ; d a s h e d l i n e , l - a t m L L M S s ( m e l t + p l a g + o l + a u g ,mel t+p lag+aug+opx+mt) . (c )Compos i t ionsofh igh-Mgmel ts [Mg/ (Mg+Fe.J>0.55 ]coex is t ingwi thp lag+o l+aug(+) ,plag * aug + opx (o), and plag + ol + aug * opx (cross within square) on (Plag + MtFsaturated diagram. Analyses of melts fromBender et at. (1978), Stolper (1980), Fujii and Solgalt (1983), and Elthon and Scarfe (1984). The composition ofglass coexistingwith plag + ol + cpx (either augite or pigeonite) at 5 kbar from Takahashi and Kushiro (1983) is also plotted (r). (d) Compositionsof low-Mg melts [Me/(Mg + Fe,",) < 0.55] coexisting with plag + ol + aug (+, S-kbar melts; o, lo-kbar melts) from Thompson(1974, 1975) and Stolper (1980) on (Plat + Mtfsaturated diagram. Numbers next to certain experiments are the temperatures atwhich the melt was produced; the nonunderlined numbers refer to the l0-kbar experiments, and the underlined numbers refer tothe 8-kbar experiments. 8-kbar LLMS of this study are also plotted.

t 9

o p x

a

o p x

(Thompson, 1975) at 8-10 kbar (Fig. ad). Thompson's(1974, 1975) experiments, when compositions are con-sidered together with temperatures (plotted next to pro-jected melt compositions; Fig. 4d), in fact also confirmthe presence of two thermal divides.

On the other hand, the LLMS (Fig. 4a) are not consis-tent in phase assemblage with literature-derived meltcompositions for which Mg/(Mg * Fe.,) is greater than0.55 and for which the low-Ca pyroxene is orthopyrox-ene, not pigeonite. Such experiments, largely on MORBcompositions, typically did not involve plagioclase (e.g.,Takahashi and Kushiro, 1983), but the few experimentsat 7.5-10 kbar that did involve plagioclase (Fujii and

Bougault, 1983; Elthon and Scarfe, 1984; Bender et al.,I 978) constrain Figure 4c. It may be noted that the pseu-do-eutectics in Figures 4a and 4c, although involving dif-ferent low-Ca pyroxenes, are at essentially the same po-sition.

Compositions of HrO-undersaturated melts

Limitations imposed by Fe loss. Because HrO-CO, fluidcompositions cannot be widely varied in graphite cap-sules, hydrous experiments were performed using noble-metal capsules and, despite all precautions (see above),samples lost Fe to capsules. Loss of Fe affects melt com-oositions in two wavs. The first is the concentration effect

r r 0 g

20 BAKER AND EGGLER: ATKA MELT COMPOSITIONS

o p x S i O r

Fig. 5. Compositions of coexisting melts and mafic crystalsat 8 kbar projected from Plag + Mt; +, (plag + ol + augfsaturated melts; a, (plag + aug + pigFsaturated melts; A, co-existing augites and pigeonites or olivines and augites. Tielinesconnecting the phases indicate the presence of thermal divideson both the plag + ol + aug and plag + aug + pig liquid linesof multiple saturation.

that occurs in any multicomponent solution when onecomponent is removed; the second is that increasing Mg/(Mg + Fe.",) values lead to higher liquidus temperaturesfor ferromagnesian minerals, relative to plagioclase andother nonferromagnesian minerals, and therefore lead todifferent liquid lines ofdescent. Because ofthese effects,a liquid line of descent for a specific bulk compositionwith 2o/o HrO in the melt cannot be determined.

Continuous Fe loss during experiments might be ex-pected to result in disequilibrium compositions of meltsand crystals; to investigate this possibility, partitioning ofFe and Mg between melts and crystals was investigatedat NNO oxygen fugacity at both I atm and 2 kbar. Thisshould be a reliable test of equilibrium as partitioningbetween olivine and melt is known to be relatively insen-sitive to pressure and temperature (Takahashi and Ku-shiro, 1983). In fact the l-atm equilibrium, or at leaststeady-state, Ko values [(Fe,",/Mg)"r",", divided by (Fe,.,/Mg)-rJ and the 2-kbar Ko values compare favorably (Ta-ble 3) (except for experiments 1148 and ll14 which havemuch lower Ko values than observed in other experi-ments). All olivine Ko values are near the "ideal" 0.30when melts are corrected for Fe3* content by the methodof Sack et al. (1980). On this basis, which is admittedlynot perfect, most 2-kbar experiments appear to demon-strate equilibrium rK" values for olivine, augite, and or-thopyroxene. The two experiments with low Ko valuesalso have the highest Fe/Mg ratios of the melts; it is pos-sible that high Fe/Mg ratios affect the partitioning of Feand Mg between crystals and melt (cf. Mysen, 1982). Theplag * aug + opx LLMS determined from I 148 and 1174

is consistent with plag + aug + opx LLMS of other 2-kbarexperiments and was therefore used in constructing thepseudoternary diagrams.

Even though experiments may closely approach equi-librium, if Fe loss has created melts falling outside of therange of natural rock compositions, then application ofthe experimental results to igneous petrogenesis is highlyquestionable. To test the experimentally produced melts,Mg/(Mg + Fe,.,) and SiOr values of a suite of rocks fromthe Aleutian island of Atka were determined. Ratios ofthose values from the rocks were compared with the ratiosdetermined for experimentally produced melts. Except forthe only 5-kbar experiment containing olivine, all meltsused to construct the hydrous pseudoternary diagramshave ratios of Mg/(Mg * Fe.",) to SiOr similar to ratiosof Atka rocks. Thus, despite Fe loss, the hydrous meltsproduced in this study are similar in composition to an-desites found in intraoceanic island arcs.

Hydrous melt compositions projected onto pseudoter-nary diagrams. Baker and Eggler (1983) suggested thatfirst-order effects of total pressure and HrO pressure on1-atm pseudoternaries are the following: (l) in Plag + Mtprojection (Figs. 4a, 6a), pressure shifts the plag * ol +aug LLMS toward olivine composition, but HrO contenthas little effect; pressure also shifts the plag + ol + low-Ca pyroxene LLMS toward olivine, but HrO content shiftsit away from olivine; (2) in the Ol + Mt projection, po-sitions of the LLMS are functions of both pressure andHrO content; however, in Di + Mt projections (Figs. 4b,6b), HrO content shifts the plag + aug + low-Ca pyrox-ene LLMS toward the plagioclase apex, owing to dra-matic lowering of plagioclase liquidi relative to ferromag-nesian mineral liquidi, whereas total pressure has a minorefect. Thus, to a first approximation, the Plag t Mt pro-jection can be used as a geobarometer and the Di + Mtprojection as a geohygrometer. Baker and Eggler (1983)based their predictions on synthetic systems (e'9., Kushi-ro, 1969), on existing natural hydrous systems (e.g., Eg-gler and Burnham, 1973), and on new experiments. Rel-atively meager amounts of data available in 1983necessitated the use of olivine-saturated phase equilibriaand therefore the pressure- and H'O-sensitive Ol + Mtprojection. Data of this study allow use of the superior,relatively pressure-insensitive, Di + Mt projection.

The new hydrous data at 2 kbar (Figs. 6a, 6b) com-pletely follow the predictions ofBaker and Eggler (1983).Compared to the 2-kbar results, a surprising element ofthe 5-kbar data is the expansion of the augite field; theexpansion presumably can be attributed to pressure rath-er than to HrO content. The experiments also demon-strate that orthopyroxene melts incongruently from I atmto at least 5 kbar and that no thermal divide exists toprevent production of andesitic melts from basaltic meltsby crystallization fractionation (cf. Kushiro, 1969; War-ne r ,1973 ) .

Other investigations. Investigations at crustal pressuresof hydrous melting of basalts or of compositions of par-tial melts, or both, have been performed by Holloway

o l

augi tc

pigeonite

BAKER AND EGGLER:ATKA MELT COMPOSITIONS

bo r

kbo r

2 l

Fig.6. H,O-undersaturatedmelts (2o/oH,Oinmelt) coexistingwithplag + ol + aug(+)andplag * aug + opx(n). (a)2-kbarPlag + Mt pseudoternary; (b) 2-kbar Di + Mt pseudoternary. a, composition of a melt coexisting with plag + aug + opx (opticallyidentified only; no orthopyroxene was found with the probe. Solid line, 2-kbar LLMS. Dashed lines, l-atm LLMS for reference. (c)5-kbar Plag + Mt pseudoternary; (d) 5-kbar Diop + Mt pseudoternary. Solid line, 5-kbar LLMS. Dashed line, l-atm LLMS.

o g r

c

and Burnham (197 2), Helz (197 6), Cawthorn (197 6), andSpulber and Rutherford (1983). Typically these experi-ments contained large amounts of water (-4 to 8o/o) re-sulting in amphibole stability and consequent dramaticchanges in liquid lines of descent. Only Spulber andRutherford (1983) obtained analyses ofhydrous basalticto rhyolitic melts coexisting with plagiqshr., olivine, andpyroxenes (augite and pigeonite, without orthopyroxene),but these phases were synthesized at much lower oxygenfugacities than in this study. Nevertheless, calculation oftheir l-2-kbar melt compositions locates an LLMS forplag + ol + aug (with oxide phases) that agrees with theLLMS of this study. Andesite experiments in this studyare comparable to studies by Eggler (1972), Eggler andBurnham (19'13), Sekine et al. (1979), and Merzbacherand Eggler (1984), who located a family of LLMS for plag+ opx + cpx and for HrO contents between 0 and 4 wto/0.Merzbacher and Eggler (1984) did not determine the typeof clinopyroxene coexisting with the melts, but similari-ties between the melt compositions of Merzbacher and

o P r

d

Eggler (1984) and those ofthis study strongly support theassumption by Merzbacher and Eggler that the clinopy-roxene was augite.

Appr,rc.q.rroN oF EXpERIMENTAL RESULTS ToPETROGENESIS OF ATKA HIGH-ALUMINA

BASALTS AND ANDESITES

The possible petrogenetic link, through a process ofcrystallization fractionation, between high-alumina ba-salt and associated andesites has long been recogrrized(Bowen, 1928; Kuno, 1950; Osborn,1959), although notaccepted by all petrologists (e.g., Fenner, l93l; Mysen,I 982). The results ofthis investigation are consistent withthe hypothesis that andesites, in particular Atka andes-ites, can be produced by crystallization fractionation fromparental high-alumina basalts at crustal pressures.

Crystallization fractionation models

One atmosphere. On a Plag + Mt projection (Fig. 7a),basalt analyses do not fall along the LLMS for plag +

P l o g

P l o 9

22 BAKER AND EGGLER: ATKA MELT COMPOSITIONS

P l o g

Atko

+ o l p h e n o s .

o p x p h e n o s .o

t z5 k b a r

o p t

b

2% H2O

N o Z O + K 2 0 M g 0 N o a 0 r K a 0 M c o

Fig. 7. Comparison of experimental LLMS and compositions of rocks from the Aleutian island of Atka. (a) Plag + Mt pseu-doternary; O) Di + Mt pseudoternary. Large arrows on Plag + Mt pseudoternary indicate pressure effects on locations of the LLMSand the olivine-Iow-Ca pyroxene boundary point. Large arrow on Di + Mt pseudoternary indicates the effect of increasing H'Opressure on the LLMS. Solid line, 8-kbar LLMS; dashed line, l-atm LLMS; dotted line, 2-kbar LLMS; dash-dot line, 5-kbar LLMS;*, olivine-orthopyroxene boundary point on the 2-kbar LLMS; O, olivine-orthopyroxene boundary point on the 5-kbar LLMS; +,compositions of rocks with phenocrystic olivine; a, compositions of rocks with phenocrystic orthoplT oxene. (c) AFM diagram ofbulk compositions of starting materials (filled squares) together with compositions of melts produced at 1 atm along the NNO buffer(the asterisks) and at 8 kbar (filled circles). (d) AFM plot of Atka rock compositions. Rock compositions courtesy of B. D. Marsh(pers. comm., 1980).

ol + aug, but some of the andesite analyses do plot along cannot be responsible for generation of the complete Atkathe LLMS for plag + aug + opx. On a Di + Mt projec- rock suite, because of the presence of a thermal dividetion, however, all natural rocks plot at much higher Plag on the plag + aug + pig LLMS and because, with crys-values than the l-atm LLMS. Furthermore, Fe enrich- tallization, the 8-kbar melts become too impoverished inment of experimental melts (Fig. 7c) is greater than the AlrO, and too enriched in FeO* and TiO' relative to Atka"calc-alkaline" AFM trend of the Atka rock suite (Fig. rocks.7d). Andesites of Atka could not be produced from dry Two and five kilobars at HrO-undersaturated condi-high-aluminabasaltsbyfractionationof plag + ol + aug tions. The 2- and S-kbar LLMS either coincide with orat low pressure, nor could silica-rich andesites or dacites bracket basaltic andesite to dacitic rock compositions ofbe produced from silica-poor andesites by fractionation the Atka suite on both Plag + Mt projections and Di +

of a plag + aug + opx + mt crystalline assemblage from Mt projections (Figs. 7a, 7b). Therefore andesite meltsanydrous melts at low pressure. can fractionate from evolved high-alumina basalts or ba-

Eight kilobars. The 8-kbar LLMS (Figs. 7a,7b) overlap saltic andesites (similar in composition to AT-41) by theAtka rock compositions in the Plag + Mt projection, but crystallization of plag + ol + aug between 2 and 5 kbarnot in the Di + Mt projection. High-pressure fractiona- for HrO contents of about 2 wto/o (it the resulting andes-tionof plag + ol + augorplag + aug + pig, moreover, it icmelts). Thosepressureestimatesarealsocompatible

I kba r / ' . ' i - "

F o o r

l o tm .I

8 kbo r

F c O r

BAKER AND EGGLER: ATKA MELT COMPOSITIONS 23

with the relative positions of Atka olivine-bearing versusorthopyroxene-bearing andesites (Fig. 7a). Presumably,HrO content of parental basaltic or basaltic andesite mag-mas was near I wt0/0. Andesitic melts can produce daciticresidual melts by the fractionation of plag + aug + opxat the same conditions of pressure and similar, but slight-ly higher, HrO contents.

A petrologically favorable consequence of slightly hy-drous fractionation is that proportions of plagioclase incrystallizing assemblages are reduced (App. Table l) withconcomitant increases in proportions ofaugite. Increasein modal augite is important in production of residualmelts that are not strongly enriched in TiOr, because au-gite is the only crystal in these experiments that is a suit-able host for Ti.

Although basaltic andesites of Atka can be linked toandesites and dacites of Atka through crystallization frac-tionation at crustal pressures with approximately 2o/oHrOin the melt, a mechanism linking more primitive high-alumina basalts (e.g., AT-l and AT-l 12) to basaltic an-desites (AT-41) is equivocal because of the lack of directexperimental data on the hydrous plag + aug * ol LLMSfor melts with composition similar to the more primitivehigh-alumina basalts. It should be noted, however, thatextrapolation of the hydrous 2- and 5-kbar plag + ol +aug LLMS to more basaltic compositions is on trend withAtka basalts in the Di + Mt projection (Fig. 7b), sug-gesting the hydrous nature of Atka basalts. In Plag + Mtprojection (Fig. 7a), the array of Atka basalts suggestsfractionation at pressures near 5 to 8 kbar.

The HrO content of Atka andesitic and dacitic meltsis unlikely to have been much higher than 2 wto/o. Merz-bacher and Eggler (1984) clearly demonstrated the sen-sitivity of the plag + aug + opx LLMS to water contentwhen displayed on a diopside-saturated projection anddeveloped a geohygrometer for use on andesitic to daciticmelts. On the basis of their research, the plag + aug +opx LLMS with 4 wto/o HrO would plot at higher Plagvalues than observed for the dacitic portion of the Atkarock suite.

Hydrous experimental results have not defined the roleof magnetite in the generation of Atka andesite with mag-netite. Magnetite was absent in hydrous experiments be-cause of relatively low oxygen fugacities (Baker and Eg-gler, 1983; Baker, 1985). Magnetite is required for manycrystallization fractionation calculations (e.g., Gill, I 98 I )and is present in Atka rocks, but in both cases magnetiteonly occurs in small amounts; such small amounts ofmagnetite are not inconsistent with the conclusions ofthis study which stress the importance of crystallizalionfractionation ofplag + ol + aug or plag + aug + opx inthe petrogenesis ofAtka rocks.

Depth of magma chambers beneath the Aleutianisland of Atka

As argued above, comparison of Atka rocks and ex-perimental LLMS on pseudoternary diagrams indicatesthat Atka andesites evolved in magma chambers at therelatively shallow depths of 7 to 17 km (2-5 kbar). Evo-

lution of basaltic andesites also probably occurred inchambers at similar depths. Basalts, however, may haveevolved in deeper magma chambers at -26 km (8 kbar;see above). In addition to that relatively deep history,basalts also record residence in more shallow magmachambers for a time period sufficient to dissolve any cli-nopyroxene phenocrysts originally present in the moreprimitive members of the suite (Baker and Eggler, 1983).Therefore, at least two levels of magma chambers existbeneath Atka. Primitive high-alumina basaltic magmas,produced by partial melting, appear to undergo limitedfractionation at -26 km; they then ascend to crustal mag-ma chambers (7-17 km) where the melts evolve to moresilica-rich high-alumina basalts, basaltic andesites, an-desites, and dacites. Magmas then probably rise into athird level of subvolcanic magma chambers that aretapped during eruption.

CoNcr-usroNs

Anhydrous experiments indicate the lack of a l-atmphase field of pigeonite for andesitic and basaltic com-positions studied, whereas such a pigeonite field exists at8 kbar. The lack ofthe pigeonite stability field precludesthe existence of a low-pressure thermal divide for thecompositions studied along a LLMS where olivine co-exists with pigeonite. At I atm, andesites exhibit a rcaa-tion point where olivine reacts with the melt, and ortho-pyroxene and magnetite begin to crystallize. At 8 kbar, athermal divide exists along both the plag * aug * ol andthe plag + aug + pig LLMS; these thermal divides pre-vent the formation of alkali basalts from tholeiites andthe production ofandesites from basalts under anhydrousconditions and deep-crustal pressures.

At 2 and 5 kbar, with 2o/o HrO in the melt, orthopy-roxene is the observed low-Ca pyroxene, and no thermaldivides exist on the LLMS; under these conditions ofpressure and HrO content, basaltic and andesitic magmascan evolve toward silica enrichment through the crystal-lization fractionation ofplagioclase, augite, and either ol-ivine or orthopyroxene.

Comparisons of LLMS and Atka rock compositions onsuitable pseudoternary projections suggest that andesiticmembers of the suite evolved in magma chambers atpressures between 2 and 5 kbar from magmas of basalticandesite composition. HrO contents of 2 wto/o are esti-mated for Atka andesites. Evolution of Atka high-alu-mina basalts occurred in deeper magma chambers, at 5-S-kbar pressure.

AcxNowr,nncMENTs

This research would not have been possible without the sam-ples donated by T. N. Irvine, B. D. Marsh, R. J. Stern, D. Walk-er, and J. S. White of the National Museum of Natural History.R. F. Fudali is thanked for the use ofthe experimental petrologylaboratory at the Smithsonian. Technical support was providedby M. Wilson, L. Eminhizer, E. Jarosewich, J. Nelen, and J.Collins. Early versions ofthis paper were thoughtfully reviewedby J. Allen, F. Dudas, S. Sorensen, and W. G. Melson. A. D.Johnston and F. J. Spera are also thanked for constructive re-views. This research was supported by Sigma Xi and P. D. Kry-

24 BAKER AND EGGLER: ATKA MELT COMPOSITIONS

nine grants and a Smithsonian postdoctoral fellowship to Bakerand NSF Grant EAR83-08292 to Eggler.

RnnnnrNcnsAlbee, A.L., and Ray, R. (1970) Correction factors for electron

probe microanalysis of silicates, oxides, carbonates, phos-phates and sulphates. Analytical Chemistry, 42, 1408-1414.

Allen, J.C., and Boettcher, A.L. (1978) Amphiboles in andesiteand basalt: II. Stability as a function of P-T-JH.O-JO,. Amer-ican Mineralogist, 63, 1074-1087.

Allen, J.C., Boettcher, A.L., and Marland, G. (1975) Amphi-boles in andesite and basalt: I. Stability as a function of P-T-.p,. American Mineralogist, 60, 1069-1085.

Baker, D.R. (1985) The compositions of melts coexisting withplagioclase, olivine, augite, orthopyroxene, and pigeonite atpressures from one atmosphere to 20 kbar and application topetrogenesis in intraoceanic island arcs. Ph.D. thesis, Penn-sylvania State University, University Park.

Baker, D.R., and Eggler, D.H. (1983) Fractionation paths of Atka(Aleutians) high-alumina basalts: Constraints from phase re-lations. Journal of Volcanology and Geothermal Research, 18,387-404.

Bence, A.E., and Albee, A.L. (1968) Empirical correction factorsfor the electron microanalysis of silicates and oxides. Journalof Geology, 7 6, 382-403.

Bender, J.F., Hodges, F.N., and Bence, A.E. (1978) Petrogenesisof basalts from the Project FAMOUS area: Experimental studyfrom 0 to 15 kbars. Earth and Planetary Science Letters, 41,277-302.

Bohlen, S.R., Boettcher, A.L., Wall, V.J., and Clemens, J.D.( 1 983) Stability of phlogopite-quartz and sanidine-quartz: Amodel for melting in the lower crust. Contributions to Min-eralogy and Petrology, 83, 270-2'17 .

Bowen, N.L. (1928) The evolution of the igneous rocks. Prince-ton University Press, Princeton, N.J.

Boyd, F.R., and England, J.L. (1960) Apparatus for phase-equi-librium measurements at pressures up to 50 kilobars and tem-peratures up to 1750"C. Journal ofGeophysical Research, 65,74r-748.

Bryan, W.B., Finger, L.W., and Chayes, F. (1969) A least-squaresapproximation for estimating the composition of a mixture.Carnegie Institution of Washington Yearbook, 67 , 243-244.

Burnham, C.W. (1979) The importance of volatile constituents.In H.S. Yoder, Ed. The evolution of igneous rocks, fiftiethanniversary perspectives, p. 439-482. Princeton UniversityPress, Princeton.

-(1981) The nature of multicomponent aluminosilicatemelts. In D.T. Rickard and F.E. Wickman, Eds. Chemistryand geochemistry at high temperatures and pressures, p.197-229. Pergammon Press, New York.

Cawthorn, R.G. (1976) Melting relations in part of the systemCaO-MgO-AlrOr-SiOr-NarO-H.O under 5 kb pressure. Jour-nal of Petrology, l'1 , 44-72.

Cawthorn, R.G., Curran, E.B., and Arculus, R.J. (1973) A petro-genetic model for the origin of the calc-alkaline suite of Gre-nada, l€sser Antilles. Journal of Petrology, 14, 327-337 .

Coulon, C., and Thorpe, R.S. (1981) Role of continental crustin petrogensis of orogenic volcanic associations. Tectono-physics, 77,79-94.

Dixon, T.H., andBatiza, R. (1979) Petrology and chemistry ofrecent lavas in the Northern Marianas: Implications for theorigin of island arc basalts. Contributions to Mineralogy andPetrology, 70, 167-181.

Dixon, T.H., and Stern, R.J. (1983) Petrology, chemistry andisotopic composition of submarine volcanoes in southernMariana arc. Geological Society of America Bulletin, 94, ll59-tt^t2.

Eggler, D.H. (1972) Water-saturated and undersaturated meltingrelations in a Paracutin andesite and an estimate ofwarer con-tent in the natural magma. Contributions to Mineralogy andPetrology, 34, 26 l-27 l.

Eggler, D.H., and Burnham, C.W. (1973) Crystallization andfractionation trends in the system andesite-Co2-HrO at pres-sures to l0 kb. Geological Society of America Bulletin, 84,25r7-2532.

Eggler, D.H., and Osborn, E.F. (1982) Experimental studies ofthe system MgO-FeO-Fe.Or-NaAlSirO'-CaAlrSirOs-SiOr-Amodel for subalkaline magmas. American Journal of Science,282, t]r2-r04t.

Eichelberger, J.C. (1975) Origin ofandesite and dacite: Evidenceof mixing at Glass Mountain in California and at other cir-cum-Pacific volcanoes. Geological Society of America Bulle-t i n . 8 6 . 1 3 8 1 - 1 3 9 1 .

Elthon, D., and Scarfe, C.M. (1984) High-pressure phase equi-libria of a high-magnesia basalt and the genesis of primaryoceanic basalts. American Mineralogist, 69, 1-15.

Ewart, A. (1982) The mineralogy and petrology of Tertiary-Re-cent orogenic volcanic rocks, with special reference to the an-desitic-basaltic compositional range. In R.S. Thorpe, Ed. An-desites: Orogenic andesites and related rocks, p. 25-95. Wiley,New York.

Fenner, C.N. (1931) The residual liquids of crystallizing mag-mas. Mineralogical Magazine, 22, 539-560.

Fujii, T., and Bougault, H. (1983) Melting relations of a magne-sian abyssal tholeiite and origin of MORB's. Earth and Plane-tary Science Letters, 62,285-295.

Gibb, F.G.F. (1974) Supercooling and the crystallization of pla-gioclase from a basaltic magma. Mineralogical Magazine,39,64 t -653.

Gill, J. (1981) Orogenic andesites and plate tectonics. Springer-Verlag, New York.

Green, T.H., and Ringwood, A.E. (1968) Genesis of the calc-alkaline igneous rock suite. Contributions to Mineralogy andPetrology, 18, 105-162.

Grove, T.L., and Bryan, W.B. (1983) Fractionation of pyroxene-phyric MORB at low pressure: An experimental study. Con-tributions to Mineralogy and Petrology, 84, 293-309.

Grove, T.L., Gerlach, D.C., and Sando, T.W. (1982) Origin ofcalc-alkaline series lavas at Medicine Lake Volcano by frac-tionation, assimilation and mixing. Contributions to Miner-alogy and Petrology, 80, 160-182.

Helz, R.T. (1976) Phase relations of basalts in their melting rangesat P H.O : 5 kb. Part II. Melt compositions. Journal of Pe-trology, 17, 139-193.

-(1978) The petrogenesis of the Ice Harbor Member, Co-lumbia River Plateau, Washington: A chemical and experi-mental study. Ph.D. thesis, Pennsylvania State University,University Park.

Holloway, J.R. (1971) Internally heated pressure vessels. In G.C.Ulmer, Ed. Research techniques for high pressures and hightemperatures, p. 217-258. Springer-Verlag, New York.

- (I97 6) Fluids in the evolution of granitic magmas: Con-sequences of finite CO, solubility. Geological Society of Amer-ica Bullet in. 87. 1513-1518.

-(1977) Fugacity and activity of molecular species in su-percritical fluids. In D.G. Fraser, Ed. Thermodynamics in ge-ology, p. 161-180. D. Reidel, Dortrecht.

Holloway, J.R., and Burnham, C.W. (1972) Melting relations ofbasalt with equilibrium water pressure less than total pressure.Journal ofPetrology, 13, l-29.

Irvine, T.N. (1974) Petrology ofthe Duke Island ultramafic com-plex, southeastern Alaska. Geological Society of AmericaMemoir 138, 240 p.

Kay, S.M., Kay, R.W., and Citron, G.P. (1982) Tectonic controlson tholeiitic and calcalkaline magmatism in the Aleutian arc.Journal of Geophysical Research, 87 , 4051-4072.

Kuno, H. (1950) Petrology ofHakone Volcano and the adjacentareas, Japan. Geological Society of America Bulletin, 61,957-1020.

- ( 1960) High alumina basalt. Journal of Petrology, 1, 12l-r45.

Kushiro, I. (1969) The system forsterite-diopside-silica, with and

BAKER AND EGGLER: ATKA MELT COMPOSITIONS 25

without water at high pressures. American Journal of Science,267 A, 269-294.

- (1975) On the nature of silicate melt and its significancein magma genesis: Regularities in the shift of the liquidusboundaries involving olivine, pyroxene, and silica minerals.American Journal of Science. 27 5, 4ll-432.

Leeman, W.P. (1983) The influence of crustal structure on com-positions of subduction-related magmas. Journal of Volcano-logical and Geothermal Research, 18, 561-588.

Lindsley, D.H. (1980) Phase equilibria at pressures > I atmo-sphere. Mineralogical Society of America Reviews in Miner-alogy, 7 , 289-308.

Mahood, G.A., and Baker, D.R. (1986) Experimental constraintson depths offractionation ofmildly alkalic basalts and asso-ciated felsic rocks: Pantelleria, Strait of Sicily. Contributionsto Mineralogy and Petrology, 93,251-264.

Marsh, B.D. (1982) The Aleutians, In R.S. Thorpe, Ed. Andes-ites: Orogenic andesites and related rocks, p. 99-113. Wiley,New York.

Merzbacher, C.I., and Eggler, D.H. (1984) A magmatic geohy-grometer: Application to Mount St. Helens and other daciticmagmas. Ceology, 12, 587-590.

Mysen, B.O. (1982) The role of mantle anatexis. In R.S. Thorpe,Ed. Andesites: Orogenic andesites and related rocks, p. 489-522. Wiley, New York.

O'Hara, M.J. (1968) The bearing of phase equilibria studies insynthetic and natural systems on the origin and evolution ofbasic and ultrabasic rocks. Earth Science Reviews, 4,69-133.

Osborn, E.F. (1959) The role of oxygen pressure in the crystal-lization and differentiation of basaltic magma. American Jour-nal of Science, 257, 609-647.

-(1979) The reaction principle. In H.S. Yoder, Ed. TheEvolution of the igneous rocks, fiftieth anniversary perspec-tives, p. 133-170. Princeton University Press, Princeton, N.J.

Sack, R.O., Carmichael, I.S.E., Rivers, M., and Ghiorso, M.S.

(1980) Ferric-ferrous equilibria in natural silicate liquids at Ibar. Contributions to Mineralogy and Petrology, 7 5, 369-37 6.

Sekine, T., Katsura, T., and Aramaki, S. (1979) Water saturatedphase relations of some andesites with application to the es-timation of the initial temperature and water pressure at thetime of eruption. Geochimica et Cosmochimica Acta, 43, 1367 -r376.

Spulber, S.D., and Rutherford, M.J. (1983) The origin of rhyoliteand plagiogranite in oceanic crust: An experimental study.Journal ofPetrology, 24, l-25.

Stern, R.J. (1979) On the origin of andesite in the NorthernMariana island arc: Implications from Agrigan. Contributionsto Mineralogy and Petrology, 68,20'1-219.

Stolper, E. (1980) A phase diagram for mid-ocean ridge basalts:Preliminary results and implications for petrogenesis. Contri-butions to Mineralogy and Petrology, 74, 13-27.

Takahashi, E., and Kushiro, I. (1983) Melting of a dry peridotiteat high pressures and basalt magma genesis. American Min-eralogist, 68, 859-879.

Thompson, R.N. (1974) Primary basalts and magna genesis, I.Skye, north-west Scotland. Contributions to Mineralogy andPetrology, 45, 317-341.

- (1975) Primary basalts and magma genesis. II. Snake Riverplain, Idaho, U.S.A. Contributions to Mineralogy and Petrol-ogy, 52,213-232.

Walker, D., Shibata, T., and Delong, S.E. (1979) Abyssal tholei-ites from the Oceanographer fracture zone. II. Phase equilibriaand mixing. Contributions to Mineralogy and Petrology, 70,1 I l -125.

Warner, R.D. (1973) Liquidus relations in the system CaO-MgO-SiO,-HrO at l0 kb P H,O and their petrologic significance.American Journal of Science, 27 3, 925-946.

Mann;scnrpr REcETvED J,wuanv 28, 1986MaNuscnrpr AccEprED Sestsl,rnsn 2, 1986

Appendix Table 1. Experimental conditions, results, and analyses

Run Temp. Phase n SiO2 TiO2( h r s . ) o C ( m o d e )

AI2o3 FeO' Mno MgO CaO Na2O K2O P2O5 Total

o l i v ( 1 . 2 ) 4 3 8 . 0 ( 4 ) 0 . 0 0 . 2 2 ( 2 ) 2 6 . 4 ( 3 ) 0 . 5 6 ( 9 ) 3 3 . 5 ( 5 ) 0 . 6 3 ( 5 ) 0 . 0 0 . 0 0 . 0 9 9 . 3 1

0 . 0 1 0 0 . 1 8

f l ^ d f n ? l

startang composrtj-on: 4r:l lz 1 atm Experiments

1 3 0 4 1 1 2 2 c l a s s ( 3 4 ) 7 5 1 . . 4 ( 4 ) 2 . 1 2 ( 1 8 ) 1 3 , 0 ( 4 ) 1 3 . s ( 4 ) 0 . 3 9 ( 5 ) 4 . 5 s ( 2 0 ) 9 . 5 0 ( 3 ) 2 . 6 5 \ 1 2 ) ! . 2 3 1 8 ) 0 . 3 5 ( 4 ) 9 8 . 6 9( 9 6 ) ( - 8 . 5 ) P l a g ( 4 9 ) 6 5 2 . 3 ( 9 ) 0 . 0 2 9 . 2 \ 4 ) 1 . 3 1 ( 1 2 ) o . o 0 . 1 5 ( 7 ) 1 3 . 5 ( 3 ) 3 . 4 s l 1 2 l 0 , 2 3 ( 2 ) 0 . 0 1 0 0 . 1 4

A u s ( 4 ) 6 s 1 . 3 ( 5 ) 0 . 5 1 . ( 1 2 ) 2 . 1 7 ( 2 5 ) 1 0 . 3 ( 5 3 ) 0 . 3 6 ( 9 ) 1 4 . 3 ( 8 ) 2 0 . 1 ( 3 ) 0 . 1 3 ( 4 ) 0 . 0 0 . 0 9 9 . 3 71 2 8 0 1 1 0 0 c l a s s ( 1 3 ) 5 s 2 . 3 ( 7 ) 3 . 2 7 ( 1 7 ) 1 2 . 3 ( 3 ) 1 3 . 7 ( 5 ) 0 . 3 4 ( 4 ) 3 . 8 3 ( 1 3 ) 8 . 6 1 ( 2 1 - ) 2 . 1 5 ( 1 8 ) 1 . 4 6 ( 8 ) 0 . 4 6 ( 8 ) 9 a . 4 2( 1 9 2 ) ( - 8 . 8 ) P f a s ( 6 2 ) s s s . 2 ( 9 ) 0 . 0 2 8 . 0 ( 7 ) 1 . 0 4 ( 6 ) 0 . 0 0 . 0 9 ( 4 ) 1 2 . 0 ( 5 ) 4 . 3 2 ( s 7 ) 0 . 4 1 ( 4 ) 0 . 0 1 0 1 . 0 6

o l i v ( 1 6 ) 4 3 ' 7 . 2 ( 3 ) 0 . 0 5 ( 5 ) 0 . 2 9 ( 1 0 ) 3 1 . 6 ( 2 ) 0 . 7 1 ( 7 ) 2 9 . ' 7 l . 3 ) 0 . 6 3 ( 2 ) 0 . 0 0 . 0A u s ( 1 0 ) 4 5 L . 2 1 6 ) 0 . 9 2 ( 1 6 ) 1 . 9 5 ( 2 s ) 1 1 . 9 ( s ) 0 . s 4 ( s ) 1 . 4 . 3 ( 3 ) 1 9 . 1 ( 3 ) 0 . 2 2 ( 1 3 ) 0 . 0 0 . 0

1086 t -106( 1 9 2 ) ( - 8 . 7 )

3 1 0 9 5( 1 9 5 ) ( - 8 . 9 )

a '7 4 1083( 1 9 2 ) ( - 9 . 0 )

1 1 1 0 6 0( 4 8 0 ) ( - 9 . 4 )

! 4 1 0 6 01 2 4 ) ( - e . 4 )

S tar t ing1 2 2 2 1 1 0 0( 1 9 2 ) ( - 8 . 8 )

4 1 0 9 5( 1 e 5 ) ( - 8 . 9 )

Composit ionG l a s s ( 5 9 )P I a q ( 2 5 )o l i v ( 5 )A u g ( 1 )G l a s s ( 5 1 )P l a q ( 3 4 )Aug( I )Opx( 4 )opaq( 3 )c l a s s ( 4 7 )P l a s ( 3 6 )A u s ( 8 )o p x ( 3 )o p a q ( 5 )c l a s s ( 3 7 )P I a g ( 4 2 )A u g ( 1 2 )o p x ( 5 )opas( 5 )

Conposit ion:c l a s s ( 6 8 )P l a q ( 2 6 )o l i v ( 4 )Aug( 5 )c l a s s ( 4 4 )P l a g ( 3 8 )A u s ( 1 0 )o p x ( 5 )O p a q ( 3 )

AT-29s 9 . 2 ( 1 )5 5 . ? ( 3 )3 7 , 7 ( 2 ts 1 . 8 ( s )5 3 . 7 ( 6 )5 6 . 7 1 2 )5 2 . 2 ( 9 15 2 . 4 ( 9 1

0 . 3 9 ( 6 )

1 . 2 3 ( 1 3 )0 . 00 . 1 s ( 1 2 )0 . 4 3 ( 7 )r . . 1 4 ( 9 )0 . 0 9 ( 6 )0 . 5 3 ( 9 )0 . 3 0 ( 5 )1 1 . 5 ( 6 )1 . 1 6 ( 1 5 )0 . 00 . 5 5 ( 1 4 )0 . 3 2 ( 1 1 )4 . 9 2 ( 1 9 l0 . 7 0 ( 2 )0 . 00 . s 2 ( 5 )0 . 2 9 ( 2 J

1 5 . 3 ( 8 )0 . 7 6 ( 8 )

s . 8 ( 2 )s . 3 ( 5 )1 . 8 ( 5 )4 . 1 ( 2 )0 . ' 7 6 t 7 )6 . 2 1 7 4 1' 7 . 4 ( e l1 . 5 ( 1 0 )2 . 9 1 6 )0 . 2 0 ( 1 1 )6 . 8 ( 8 )

L 4 . 6 \ 2 ) 8 . 6 1 ( s ) O . 2 2 1 1 )2 ' , 7 . 2 1 6 \ 0 . 9 1 ( 2 ) o . o

0 . 1 9 ( 3 ) 3 0 . 6 ( 1 ) 0 . 7 6 ( 4 )1 . 7 6 ( 3 ) 1 1 . 3 ( 5 ) 0 . s 8 ( 7 )

1 4 . 3 ( 3 ) 7 . 3 6 ( 1 e ) o . o2 8 . 3 1 4 ) 0 . 9 9 ( 1 1 ) 0 . 0

r . 6 ' 7 1 2 6 ] . L L . 2 ( L I 0 . 00 . 8 5 ( 9 ) 2 0 . 5 ( 4 ' t 0 . 03 . 5 0 1 2 2 ) 7 ' 7 , 4 ( 3 1 0 . 0

1 4 . 9 ( 6 ) 4 . 3 1 ( 1 ) 0 . 1 3 ( 3 )2 7 . 2 1 1 ' ) 0 . 9 s ( 9 ) 0 . 0

2 . 0 3 ( 3 ) 1 0 . 8 ( 5 ) 0 . 5 8 ( 4 )1 . 0 9 ( 1 4 ) 1 4 . 9 ( 5 ) 0 . 9 0 ( 4 )3 . 6 9 ( 1 6 ) 7 5 . 4 ( 8 ) 0 . 8 4 ( 9 )

1 4 . 7 ( 3 ) 4 . 9 4 ( 2 8 ) 0 . 02 6 . 4 ( 6 1 0 . 9 3 ( 1 5 ) 0 . 0

1 . 9 8 ( 2 9 ) 1 1 . 2 ( 3 ) 0 . 00 . 7 7 1 2 4 ) 2 0 . ' t ( 3 J 0 . 02 . s 6 ( 3 ? ) 7 3 . 5 ( 3 ) 0 . 0

1 4 . 4 ( 9 ) 4 . 9 5 ( 6 5 ) 0 . 0

2 . 7 4 1 9 ) 5 . 8 2 ( 1 5 ) 3 . 7 0 ( 2 3 )0 . 1 4 ( 8 ) 1 , 7 . 2 1 2 ) 4 . ' 7 6 ( 2 7 J

3 0 . 7 ( 1 ) o . 5 0 ( 1 ) o . o1 4 . 6 ( 3 ) 1 8 . 4 ( s ) 0 . 2 7 ( 1 0 )

1 . 6 1 ( 1 s ) 3 . 7 8 ( 3 8 ) 4 . 2 r ( 2 7 10 . 1 6 ( 4 ) 1 0 . 6 ( s ) 5 . 3 8 ( 1 s )

1 s . s ( s ) 1 9 . 0 ( 3 ) 0 . 4 2 ( 5 )2 3 . ' 7 1 7 \ 2 . 0 2 ( 1 s ) 0 . 0 6 ( 2 )

3 . 4 0 ( ] . 8 ) o . 2 0 ( 4 ) 0 . 0L . 7 2 1 9 ' ) 3 . 6 4 / . 2 6 ' t 3 . ' 7 6 1 2 4 )0 . 0 3 ( 2 ) 1 1 . 0 ( 6 ) 4 . 7 6 ( 1 4 )

L 4 . 4 l 2 l 1 9 . 2 1 4 \ 0 . 2 6 ( s )2 6 . 2 1 3 t 2 . L 6 ( 2 J 0 . 0

s . 7 4 ( 3 0 ) 0 . 2 1 ( 5 ) 0 . 00 . 7 8 ( 9 ) 1 . 9 8 ( 4 ) 3 . ' 7 L l 2 6 lo . 0 8 ( 3 ) 9 . 4 3 ( 1 6 ) 5 . 8 1 ( 2 2 )

1 4 . 5 ( 4 ) 1 9 . 3 ( 5 ) 0 . 3 3 ( 5 )2 2 . 0 ' 6 1 1 . 8 7 ( 8 ) 0 . 0 5 ( 1 )

2 . 3 2 ( 1 0 ) 0 . 2 0 ( 3 ) 0 . 00 . 8 0 ( 9 ) 2 . 1 2 ( 2 4 1 3 . 1 1 ( 4 6 )

2 . 4 3 1 5 ) 5 . 5 3 ( 1 4 ) 3 . 5 6 ( 2 0 )o . 0 8 ( 7 ) 1 0 . 9 ( 3 ) 4 , 6 9 1 2 3 )

3 2 . 5 1 1 J 0 . 5 1 ( 1 ) o . o1 4 . 3 ( 5 ) 1 9 . 9 ( 3 ) o . 3 s ( 1 2 )

1 . 4 2 ( 1 r . ) 3 . 0 4 ( 1 3 ) 4 . 0 0 ( 1 1 )o . o ? ( 2 ) 9 . 5 1 1 2 0 1 5 . 5 1 ( 2 9 )

1 4 . 6 ( 3 ) 2 0 . o ( 3 ) 0 . 4 3 ( 7 )2 3 . 6 1 - 7 t 1 . 8 3 ( 9 ) 0 . 0 7 ( 3 )

3 . 5 3 ( 2 6 ) 0 . 1 7 ( 6 ) o . o

2 . 7 5 1 5 1 0 . 3 2 ( 5 ) 9 9 , r 90 . 5 3 ( ? ) 0 . 0 1 0 0 . 4 40 . 0 0 . 0 1 0 0 . 0 00 . 0 0 . 0 9 9 . 7 44 . 2 2 ( 2 5 ' ) 0 . 3 3 ( 1 4 ) 1 0 0 . 5 50 . 7 0 ( 1 2 ) 0 . 0 ! 0 2 . 9 20 . 0 0 . 0 1 0 0 . s 20 . 0 0 . 0 9 9 . 8 70 . 0 0 . 0 9 6 . 3 93 . 9 9 ( 1 ^ 5 ) 0 . 5 3 ( 6 ) 9 9 . 9 40 . 5 3 ( 7 ) 0 . 0 9 9 . 7 70 . 0 0 . 0 9 9 . 6 20 . 0 0 . 0 9 9 . 6 ' 70 . 0 0 . 0 9 2 . 5 65 , L 9 1 2 2 ) 0 . 2 6 ( 3 1 9 8 . 4 60 . 6 ? ( 1 2 ) 0 . 0 1 0 0 . ' 7 20 . 0 0 . 0 9 9 . 3 30 . 0 0 . 0 9 8 . 5 80 . 0 0 . 0 9 4 . 1 85 . 1 r . ( 4 0 ) 0 . 3 1 ( 2 9 ) 9 8 . 3 6

3 . 3 5 ( s ) 0 . 3 3 ( 3 ) 9 8 . 6 70 . 5 5 ( 4 ) 0 . 0 9 9 . 5 20 . 0 0 . 0 9 9 . 5 80 . 0 0 . 0 9 8 . 9 7s . 1 1 ( 2 1 ) 0 . 3 4 ( 8 ) 9 9 . 0 10 . 9 1 ( 1 3 ) 0 . 0 9 9 . 6 00 . 1 0 ( 4 ) 0 . 0 9 8 . 3 90 . 1 1 ( 3 ) 0 . 0 9 7 . 8 50 . 0 0 . 0 9 3 . 0 6

AT-255 9 . 3 ( 7 ) r . 2 2 1 6 ) 1 , 4 . 7 1 2 ) 8 . 1 3 { 4 3 ) 0 . 1 1 ( s )5 4 . 9 ( 7 ) 0 . 0 2 7 . 5 ( 3 ) 0 . 8 0 ( 1 6 ) 0 . 03 7 . 6 ( 1 ) 0 . 1 4 ( 4 ) 0 . 2 0 ( 2 ) 2 8 . r - ( r . ) 0 . 5 3 ( 9 )5 1 . 3 ( s ) 0 . 4 9 ( 1 2 ) 1 . 6 7 ( 1 4 ) 1 0 . 5 ( 3 ) 0 . 4 6 ( 7 )6 3 . 6 ( 9 ) 1 . 0 9 ( 8 ) ] . 4 . 6 ( 4 ) 5 . 8 1 ( 4 3 ) 0 . 05 6 . 0 ( 9 ) 0 . 0 2 6 . 9 1 ' 7 ) 0 . 6 4 ( 6 ) 0 . 0s 1 . 1 ( 3 ) 0 . 4 9 ( 3 ) 1 . 7 0 ( 1 3 ) 9 . 9 7 ( 6 s ) 0 . 0s 1 . 7 ( 1 ) 0 . 2 5 ( 1 ) 0 . 8 9 ( 4 2 ) 1 e . 4 ( 9 ) 0 . 0

0 . 2 6 ( 1 0 ) 1 1 . 3 ( 5 ) 3 . 5 0 ( 2 4 ) ? 4 . 3 ( 7 ) 0 . 0

26 BAKER AND EGGLER:ATKA MELT COMPOSITIONS

Appendix Table 1. Continued

Run Temp. Phase n SiO2( h r s . ) o c ( m o d e )

( I o q f o 2 )1 4 ' 9 ( 8 ) 4 . 2 . 7 ( 2 . 7 J o . 2 | | 6 ) 1 ' 4 o ( 8 ) 3 . 6 1 ( 3 1 ) 3 . 6 4 ( 2 2 | 4 , 4 8 | 7 ) o ' 4 7 ( 5 ) 9 8 . 4 1

( 1 9 2 ) ( - 9 . 0 ) P l a q ( 3 5 ) 4 5 6 . 0 ( 9 ) 0 . 0 2 7 . 5 ( 3 J 0 . 6 4 ( 8 ) 0 . 0 0 . 0 2 ( 2 ) 1 0 . ? ( 3 ) 4 . 8 3 ( l - 7 ) 0 . 7 2 ( 3 ) 0 . 0 1 0 0 . 4 1 .A u s ( ? ) 6 5 1 . 8 ( 4 ) 0 . s 2 ( 2 0 ) 1 . 9 0 ( 4 8 ) 1 0 . 1 ( 6 ) 0 . 4 9 ( 1 3 ) 1 4 . 2 ( 8 ) 2 0 . 1 ( 6 ) 0 . 4 0 ( 1 4 ) 0 . 0 0 . 0 9 9 . 5 1o p x ( s ) 4 5 4 . 2 1 9 ) 0 . 3 6 ( 1 5 ) 1 , 3 0 ( 1 4 ) 1 4 . 5 ( 9 ) 0 . ' l 4 l I ' 7 J 2 5 . 9 ( ' 7 ) 2 . 1 6 l 2 r l 0 . 2 1 1 ' 7 ) 0 . 0 0 . 0o p a q ( 4 ) 4 0 . 6 9 ( s ) s . s 4 ( 5 1 ) 3 . 9 2 ( 1 1 7 5 . 4 ( 5 ) 0 . ? 1 ( 8 ) s . s l ( 2 8 ) 0 . 1 7 ( 4 ) 0 . 0 0 . 0 0 . 0

TiO2 A12O3 FeO' MnO MSO c a o N a 2 O K 2 o P 2 O 5 T o t a l

9 9 . 3 ' 79 l - . 9 4

9 3 4 1 3 0 0 , /t1 /24) 1 -1-75

8 9 1 1 1 . 5 01 2 4 )

9 s 3 1 3 0 0 /l t / 2 4 ) , 1 t 5 o

1008 1L25( 9 6 )

9 4 0 1 1 7 51 2 4 )

1 0 9 8 1 3 o o /( 7 / 2 4 ) L 1 - 7 s

1 0 1 0 I I 2 5( 9 6 )

9 8 9 1 3 0 0 , /l 7 / 2 4 ) 7 1 2 5

910 1725( e 6 )

9 r ' 7 1 1 0 0t 1 9 2 )

9 4 4 L L 1 5 /1 2 4 / 1 1 0 09 6 )

1 0 7 5 1 0 7 5( 7 9 2 )

911 71-25( 9 6 )

' l -0 '74 10?5( l e 2 t

AT-1+Mix8 0 2 1 2 5 0| 2 4 )

8 0 5 1 2 0 0| 2 4 J

8 0 8 1 2 0 0( 9 6 )

8 1 5 1 1 2 o O( ' 7 2 )8I'7 12 0 0( ' 7 2 )8 1 8 1 2 A al ' 7 2 )

M- 4 6+Mix8 1 3 1 2 0 0( 4 8 )8 1 9 1 2 0 0l ' 7 2 )

8 6 9 1 1 5 0( 9 6 )

Composrt lonG l a s s ( 6 0 )P l a q ( 3 5 )o l i v ( 4 )P i q ( 4 )G l a s s ( 3 8 )P l a q ( 4 8 )o l i v ( 5 )P i q ( 9 )c l a s s ( 4 4 )P 1 a 9 ( 4 6 )o l i v ( 3 )P i q ( 7 )

C o m p o s i t i o n :c l a s s ( 4 1 ) 6P l a g ( 5 2 ) 6P i g ( 7 ) s

3 . 0 9 ( 2 9 \ 6 - 9 ' 7 ( 7 2 )0 . 1 2 ( 6 ) 8 . 3 5 ( 2 6 )

1 6 . 9 ( 5 ) 4 . 9 4 1 ' 7 5 )

' 7 . 4 5 t 3 1 \ 9 . 2 3 ( 1 3 )0 . 1 9 ( 6 ) 1 3 . 4 ( 1 s )

3 8 . 4 ( 3 ) 0 . 3 6 ( 4 )1 8 . 9 ( 6 ) 1 1 . 8 ( 3 )

3 . 5 3 ( 2 9 ) 8 . 2 0 ( 3 6 )0 . 1 1 ( 3 ) 1 3 . ? ( 1 3 )

2 4 . r ( 4 ) 0 . s s ( 4 )1 2 . 9 ( 6 ) 1 4 . ? ( 1 1 )

2 . 2 3 ( 1 ) 5 . 6 1 1 2 4 \0 . 0 7 ( 4 ) 8 . 5 4 ( 1 1 )

1 9 . 2 ( 1 5 ) 6 . 6 1 \ 6 2 )2 . 0 1 1 2 2 \ 5 , ' 7 6 1 1 , 4 )0 . 0 4 ( 2 ) 8 . 5 5 ( 1 6 )

1 5 . 3 ( 1 ) 1 1 . 1 ( 4 )1 6 . 7 ( 3 ) 7 . 9 3 ( 4 8 )

3 . 1 1 ( 2 0 ) 1 . 8 5 ( ] . 8 )6 . 0 4 ( 1 8 ) 0 . 9 s ( 4 )0 . 1 6 ( 6 ) 0 . 0

2 . 3 3 ( 3 r ) 1 . 0 2 ( 5 )3 . 4 4 ( 6 1 0 . 2 8 ( 1 )0 . 0 0 . 00 . 0 6 ( ? ) 0 . 02 . 2 ' 7 ( l ' 7 ) 2 . A 6 ( t 7 )3 . 6 9 ( 1 6 ) o . s 6 ( 1 1 )0 . 0 0 . 00 . 5 7 ( 1 9 ) 0 . 0

3 . 7 8 ( 1 7 ) 3 . 4 2 ( L 2 )s . 8 9 ( 1 ? ) 1 . 1 8 ( 6 )0 . 2 4 0 . 00 . 1 5 ( 4 ) 0 . 03 . 2 s ( 1 5 ) 3 . 2 9 ( 1 3 )s . 8 1 ( 6 ) 1 . 4 0 ( 1 7 )0 . 3 3 ( 6 ) 0 . 00 . 2 2 \ 6 ) 0 . 0

Composit ionc l a s s ( 5 9 )P l a s ( 3 4 )o l i v ( 1 )A u q ( 6 )c l a s s ( 6 )P 1 a 9 ( 6 0 )o 1 j - v ( 1 1 )A u g ( 2 2 )

C o m p o s a t a o nc l a s s ( 4 3 )P l a q ( 4 1 )A u s ( 4 )P i s ( 1 2 )c L a s s ( 4 3 )P I a g ( 4 1 )A U g t r JP i s ( 1 4 )

Composit ionc l a s s ( 7 4 )P l a q ( 2 3 )P i s ( 3 )G 1 a s s ( 5 1 )P l a g ( 3 8 )P i s ( 1 1 )G l a s s ( 6 9 )P L a q ( 2 6 )P i s ( 4 )c l a s s ( 7 3 )P l a g ( 2 3 )P i s ( 3 )

C o n p o s i t i o n :G l a s s { 5 7 ) 5P l a g ( 3 2 ) ' 7

P i s ( 1 1 ) sG l a s s ( 6 0 ) 4P l a q ( 2 5 ) 5P i s ( 1 5 ) 5

P l a g

Aug

Augc l a s sP l a go 1 i vAugc l a s s

1 1 . 6 ( 4 ) 0 . 3 4 ( 4 )0 . 4 7 ( 5 ) 0 . 0

2 4 . 0 . 4 ) 0 . 7 6 ( 8 )8 . 7 1 ( s s ) 0 . 3 1 ( 7 )0 . 6 ? ( 1 6 ) 0 . 0

2 s . 0 ( 4 ) 0 . 7 5 ( 3 )

1 . 8 3 ( 1 4 )0 . 0 1 ( 2 )

1 6 . 3 ( 6 )

0 . 0 3 ( 3 )1 6 , 0 ( 7 )

AT-14 9 . 1 ( 6 ) 1 - . 6 2 \ 7 6 ) 7 6 . 2 { 3 )5 4 . 0 ( 5 ) 0 . 0 2 8 . 5 ( 2 )3 ' / . 2 0 . 0 0 . 05 1 . 2 ( 3 ) 0 . 6 4 ( 7 ) 5 . s 3 ( 3 3 )4 ' 7 . 8 1 4 ) 2 . 6 2 1 2 4 ) L 4 . 0 ( 2 )s s . 8 ( 6 ) 0 . 0 2 7 . 3 ( 3 13 6 . 5 ( 4 ) 0 . 1 0 { 1 ) o . 0 3 ( 3 )s 1 . 3 ( 1 1 ) 0 . 5 3 { 9 ) r . 7 5 ( 6 9 )4 ' 7 . 6 1 4 1 2 , 0 7 1 2 8 ) L 4 . 9 ( 2 )s 4 . 0 ( 3 ) 0 . 1 s ( 1 1 ) 2 ' 7 , 9 ( 2 2 )3 ? . 3 ( 3 ) 0 . 1 6 ( 4 ) 0 . 05 1 . 9 ( 4 ) 0 . 5 0 ( 7 ) 4 . 5 7 ( s 6 )AT-414 9 . 8 ( 6 ) 3 . 0 6 1 L 7 ) 1 2 . 5 \ 2 15 9 . 6 ( 8 ) 0 . 0 2 4 . 8 ( 3 )5 2 , 4 1 ' 7 ) 0 . 6 1 ( 2 4 ) 2 . 2 4 ( 5 6 )AT-1124 8 . 3 ( 4 ) 1 . 2 0 { 1 6 ) 1 4 . 9 ( 8 )s 1 . 9 ( 3 ) 0 . 0 7 ( 2 ) 2 9 . 5 ( 3 )3 9 . 0 ( 4 ) 0 . 1 s ( 5 ) 0 . 0 7 ( 3 )5 2 . 0 ( 6 ) 0 . 6 7 ( 1 3 ) 5 . 1 8 ( 8 )4 2 . 8 ( 8 ) s . 6 3 l 3 r ) 1 2 . 2 ( 7 )5 2 . 4 ( 3 ) 0 . 0 2 9 . 7 1 4 )3 5 . 4 ( 2 ) 0 . 1 0 ( 6 ) 0 . 0 9 ( s )4 8 . 8 ( 6 ) 1 . 9 6 ( 1 8 ) 6 . s 0 ( ? )AT-295 7 . 2 1 4 ) 1 . 6 5 ( 1 6 ) 1 3 . 8 ( 5 )5 8 . 0 ( 5 ) 0 . 0 2 s . 1 ( s )5 0 . 9 0 . 9 8 2 . 3 5s 2 . 7 1 ' 7 ) 0 . 3 0 ( 9 ) 2 . 0 6 \ 4 4 )5 6 . 4 ( 7 ) r . 9 6 1 1 ) L 3 . 4 ( 2 )5 8 . 7 ( 4 ) 0 . 0 2 5 . r ( 2 )4 9 . 6 ( 5 ) 0 . 7 1 ( 8 ) 3 . 6 3 ( 3 6 )s 0 . 8 ( ? ) 0 . 5 5 ( 8 ) 2 . 8 1 ( 4 0 )

6 1 . 0 ( s ) 1 . 1 9 ( 7 ) 1 3 . 4 ( 3 )5 9 . 4 ( 3 ) 0 . 1 3 ( 7 ) 2 4 . 1 1 4 )5 2 . 7 0 . 4 2 1 . 7 \6 1 . 7 ( 5 ) 1 . 4 s ( 1 3 ) 1 2 . s ( 3 )6 0 . 8 { 5 ) 0 . 1 6 ( 6 ) 2 3 . 9 ( 4 15 0 . 9 ( 7 ) 0 . 4 6 ( 2 ) 1 . 3 1 ( 3 0 )6 0 . 8 ( 8 ) r . 2 0 ( 2 r J 1 3 . 2 ( 3 J5 9 . 9 1 7 ) 0 . 0 2 4 . 5 ( s )5 1 . 3 ( 4 ) 0 . 4 3 ( 1 0 ) 1 . s 6 ( 8 )6 0 . 8 ( 7 ) 0 . 8 9 ( 2 1 ) 1 3 . 4 ( 3 )s 9 . 5 ( 5 ) 0 . 0 2 4 . ' 7 ( 6 )5 1 . 4 ( 5 ) 0 . 5 ' 7 ( L 2 ) 1 . s 4 ( 1 8 )AG8-4s 9 . 4 ( s ) 1 . 4 4 ( 1 4 ) 1 2 . 8 ( 1 )s 9 . 5 ( 7 ) 0 . 1 s ( 7 ) 2 4 . 9 ( 4 1s 0 . 7 ( 3 ) 0 - 5 2 ( 4 ) 1 . 9 s ( 1 9 )6 0 . 7 ( 8 ) 1 . 1 0 ( 1 6 ) l s . 3 ( 1 )s 9 . 6 ( 5 ) 0 . 0 2 3 . 9 ( 3 )s l . 4 ( 4 ) 0 . 4 7 ( 1 3 ) r - . 4 8 ( 1 4 )

C L h ^ Y F Y h a r i h a n f <

1 3 . s ( 7 ) 0 . 2 s ( 6 ) 6 . 1 3 ( 1 9 )0 . 6 2 ( 6 ) o . o 0 . 1 0 ( 4 )

2 6 . 0 0 . 5 3 3 5 . 71 4 . 8 ( 2 ) 0 . 4 6 ( 6 ) 2 1 - 5 ( 5 )7 6 . 2 1 3 ) 0 . 3 2 ( 5 ) 4 . 6 7 ( 8 )

0 . 7 5 ( 4 2 ) 0 . 0 0 . 1 7 ( 7 )3 3 . 4 ( 8 ) 0 . 5 7 ( 8 ) 2 9 . 6 ( 5 )1 7 . 1 ( 5 ) 0 . 5 3 ( 3 ) 2 0 . 2 ( 6 )1 5 . 7 ( 3 ) 0 . 3 1 ( 4 ) s . 7 0 ( 9 )

0 . 3 8 ( 8 ) 0 . 0 3 ( 2 ) 0 . 0 7 ( 3 )3 o . o ( 2 ) 0 . 5 4 ( 1 ) 3 2 . 0 1 2 ' )1 6 . 8 ( 4 ) 0 . 4 6 ( s ) 2 0 . 9 ( 4 ' )

1 7 . 1 ( 5 ) 0 . 3 8 ( s )0 . 6 9 { 1 0 ) 0 . 0

2 4 . 0 ( ' 7 ) 0 . ? 8 ( 1 1 )

1 3 . 9 { 9 ) 0 . 3 6 ( 8 )0 . 4 7 ( 5 ) 0 . 0 2 ( 1 )

2 2 . 1 \ 7 ) 0 . 3 9 ( s )1 0 . 7 ( 6 ) 0 . 3 3 ( 4 )2 0 . 0 { 9 ) 0 . 3 5 ( 8 )

0 . 6 3 ( 1 4 ) 0 . 04 0 . 5 ( 1 0 ) 0 . 6 6 ( 1 0 )1 5 . 1 ( 7 ) 0 . 3 ? ( s )

1 0 . 9 ( 2 ) 0 . 2 3 ( 8 )0 . 5 4 ( 1 6 ) 0 . 0

r 7 . 9 0 . 4 51 9 . 0 ( 6 ) 0 . s s ( 3 )1 1 . 8 i 6 ) 0 . 3 1 ( s )

0 . 3 6 ( s ) 0 . 01 8 . 3 { s ) 0 . 6 5 ( 4 )2 0 . 1 ( 7 ) 0 . ? 0 ( 1 2 )

8 . 3 8 ( 2 0 ) 3 . 2 2 0 . 8 ) 0 . 9 6 ( 6 ) 0 . 2 3 ( 4 ) 9 9 . 5 91 2 . 0 ( 3 ) 4 . 2 2 ( 1 9 ) 0 . 3 4 ( 5 ) 0 . 0 9 9 . 7 8

0 . 3 4 0 . 0 0 . 0 0 . 0 9 9 . ' 1 9s . 6 6 ( 3 5 ) 0 . 1 8 ( 3 ) 0 . 0 0 . 0 9 9 . 4 1? . 8 5 ( 1 r - ) 3 . 3 2 ( 1 3 ) 1 . s 0 ( 7 ) 0 . 5 3 ( ? ) 9 8 . 8 1

1 1 . 0 ( 6 ) 4 . 8 7 ( 2 9 ) 0 . s 6 ( 6 ) 0 . 0 1 0 0 . 4 60 . 4 ( 6 ) 0 . 0 0 . 0 0 . 0 1 0 0 . 6 05 . 9 8 ( 4 ? ) 0 . 1 3 ( 3 ) 0 . 0 0 . 0 9 9 , 5 28 . 2 3 ( ' , 7 ] , 3 . 0 2 ( 1 4 ) 1 . 1 6 ( 3 ) 0 . 4 8 ( 5 ) 9 9 . r ' 7

L L , 9 ( 2 ) 4 . s 0 ( 1 8 ) 0 . 4 0 ( 4 ) 0 . 0 9 9 . 3 30 . 4 3 ( 8 ) 0 . 0 0 . 0 0 . 0 1 0 0 . 4 3s . 8 0 ( 3 5 ) 0 . 1 8 ( 5 ) 0 . 0 0 . 0 1 0 1 . 1 1

1 0 . 1 ( 4 ) 0 . 2 s ( 6 ) 2 . 2 4 ( 1 2 )0 . 3 8 ( 4 ) o . o o . o

2 3 . 2 0 . 8 2 1 8 . 51 0 . 2 ( 3 ) 0 . 3 0 ( 6 ) 1 . 0 9 ( 1 1 )

0 . 6 2 ( 8 ) 0 . 0 0 . 0 3 ( 3 )2 5 . 1 ( 6 ) 1 . 0 3 ( 1 ? ) 1 6 . 5 ( 4 )1 0 . 1 ( 5 ) 0 . 2 8 ( s ) 2 . 0 2 ( 2 3 )

0 . 4 3 ( s ) 0 . 0 0 . 02 3 . 3 ( s ) 0 . ? 2 ( 1 0 ) 1 7 . 5 ( 4 )1 0 . 2 ( 5 ) 0 . 4 1 ( ? ) 2 . r 4 ( ' 7 )

0 . 4 7 ( 1 3 ) o . o o . 0 5 ( 3 )2 3 . 8 ( 5 ) 0 . 8 5 ( s ) 1 6 . 9 ( 3 )

4 . 9 9 ( 1 1 ) 3 . 6 1 ( 9 )' 7 8 4 1 2 5 ) 6 . 0 9 ( 2 1 )

4 . 1 ' 7 0 . 04 . 2 3 1 1 ) 3 . 7 2 ( 8 )6 . ' 7 5 ( 3 7 ' ) 6 . 6 4 ( 1 5 )4 . 4 s ( 3 2 ) 0 . 0 2 ( 3 )4 . 8 1 ( 2 0 ) 3 . 5 0 ( 1 5 )? . s 8 ( 3 3 ) 6 . 3 1 ( 9 )4 . s 3 ( 2 8 ) 0 . 0 5 ( 5 )4 . 9 6 ( 1 6 ) 3 . 9 8 ( 2 0 )8 . 2 6 ( 4 3 ) 6 . 1 8 ( 2 8 )4 . 8 0 ( 2 6 ) o . 0 8 ( s )

0 . 7 0 ( 9 ) 9 8 . s 60 . 0 1 0 0 . 5 50 . 0 1 0 1 . 6 3

0 . 2 5 ( 3 ) 9 8 . 9 40 . 0 9 9 . 2 ' 70 . 0 1 . 0 1 . 0 70 . 0 9 9 . 6 41 . 5 5 ( 2 3 ) 9 8 . 6 90 . 0 1 0 0 . ? 90 . 0 1 0 1 . 4 00 . 0 1 0 0 . 9 0

0 . 4 0 ( 3 ) 9 8 . 9 80 . 0 9 9 . 3 20 . 0 9 9 . 9 20 . 0 9 9 , 9 ' 70 . 5 4 ( 6 ) 9 8 . 7 80 . 0 9 9 . 9 60 . 0 9 9 . 6 20 . 0 1 0 0 . 5 1

2 . 3 4 1 8 ) 0 . 4 4 ( 3 ) 9 9 . 5 80 . 7 4 ( 6 ) 0 . 0 9 9 . 2 80 0 . 0 1 0 0 . 8 92 . 9 3 ( 5 ) 0 . 6 4 ( 1 0 ) 9 8 . ' 7 61 . 1 0 ( 8 ) 0 . 0 1 0 0 . 0 00 . 0 0 . 0 9 9 , ' l ' l2 . 4 2 ( 1 3 ) o . s 3 ( 6 ) 9 8 . 8 60 . 8 3 ( 7 ) o . o 9 9 . 5 50 . 0 0 . 0 9 9 . 6 02 . 3 4 ( s ) 0 . 4 6 ( 4 ) 9 9 . 5 80 . ? 3 ( 8 ) o . o 9 9 . 8 90 . 0 0 . 0 9 9 . 9 4

0 . 1 2 ( 4 ) 9 9 . 6 ' 70 . 0 9 9 , 6 90 . 0 707.21-0 . 0 1 0 0 . 4 70 . 2 3 ( 4 ) 9 9 . 1 00 . 0 9 8 . 8 50 . 0 1 0 0 . 8 50 . 0 9 9 . 2 20 . 1 9 ( 4 ) 9 8 . 6 40 . 0 9 9 . 2 r0 . 0 9 9 . 4 60 . 0 9 9 . 2 00 . 1 5 ( 7 ) 9 9 , 4 5

0 . 2 0 ( s ) 9 9 . 3 5

0 . 1 9 ( ? ) 9 9 . 3 8

0 . 1 s ( 4 ) 9 9 . 3 0

0 . 1 6 ( 4 ) 9 9 . 5 60 . 0 9 8 . 8 70 . 0 1 0 0 . 9 60 . 0 9 9 . 5 80 . 3 3 ( 5 ) 9 8 . 5 70 - 0 9 a . ' 1 30 . 0 9 9 . 9 60 . 0 9 9 . ' 7 7

5 . 4 4 ( 5 ) 3 . 5 6 ( 8 ) 2 . 2 6 1 3 1 0 . 3 6 ( 4 ) 9 9 . 0 37 . 8 4 ( 3 5 ) 6 . 7 2 ( L 4 , A . 7 6 1 ' 7 ) 0 . 0 9 9 , 7 4s . 3 4 ( 2 3 ) 0 . 1 3 ( 6 ) 0 . 0 0 . 0 9 9 . 7 05 . 9 ' 7 1 2 4 1 4 . 0 9 ( 2 1 ) 2 . 1 0 ( 6 ) 0 . 4 1 ( 1 1 ) 9 9 . 8 47 . 8 0 ( 3 9 ) s . 9 4 ( 1 5 ) 0 . ? 9 ( 7 ) 0 . 0 9 8 . ' 7 9s . 0 5 ( 1 0 ) 0 . 0 6 ( 3 ) 0 . 0 0 . 0 1 0 0 . 2 1

at 8 kbar

P I a g

Auq

P I a gAugP i s

4 8 . 7 ( 3 ) 0 . 6 9 ( 1 9 )s 1 , 1 ( 1 0 ) 0 . 04 0 . 1 ( 4 ) 0 . 05 1 . 2 ( 9 ) 0 . 3 3 ( 1 9 )4 8 . 2 1 2 ) 0 . 9 8 ( I )5 0 . s ( 4 ) 0 . 1 0 ( 6 )3 9 . 8 ( 4 ) 0 . 05 0 . s ( s ) 0 . s s ( ? )4 7 . 8 ( 5 ) 1 . 2 1 ( 1 9 )5 1 . 4 ( 8 ) 0 . 03 8 . 4 ( 5 ) 0 . 05 0 . 2 ( 6 ) 0 . 6 3 ( 1 6 )4 8 . 3 ( 6 ) r . 2 2 l I 2 l

4 9 . 1 ( 5 ) 1 . 3 3 ( 1 ? )

4 ' 7 . 4 1 6 ] , 1 . 5 3 ( 2 3 )

4 9 . 2 1 4 ) 1 . 0 s ( 1 9 )

4 8 . 4 ( s ) 1 . 1 6 ( 1 5 )5 0 . 9 ( s ) 0 . 03 8 . 6 ( s ) 0 . 0s o . 7 ( 7 ) 0 . 5 4 ( 1 1 )4 9 . 0 ( 1 ) 2 . 1 3 1 2 0 )5 5 . 2 ( 9 ) 0 . 05 2 . 1 1 ' 7 ' J 0 . 3 5 ( 9 )s 2 , 2 1 5 ) 0 . 4 4 ( 2 0 )

7 7 . 2 ( 5 ) 9 . 6 9 1 2 9 12 9 . s ( 1 0 ) 0 . 5 5 ( 4 9 )

0 . 0 1 5 . 5 ( 1 0 )7 . 2 5 ' 9 3 ) 5 . 8 9 ( 6 2 )

1 6 . ' 7 ( 4 ) 9 . 8 3 ( 1 9 )2 9 . 5 ( 3 ) 0 . 3 9 ( ? )

0 . 0 1 6 . 2 ( 3 )? . 0 6 ( 7 8 ) 5 . 2 5 ' 2 8 t

1 5 , 9 1 2 ) 1 1 . 3 ( 3 )2 9 , 9 1 3 ) 0 . 2 ' 7 l ' / t

0 . 0 1 9 . 8 ( 4 )6 . ' 7 5 1 2 6 ) 6 . 3 6 ( 3 3 )

L 6 . 2 ' 7 ) 1 o . 4 ( 3 )

1 6 . 0 ( 1 ) 1 1 . 1 ( 1 )

1 s . 8 ( 2 ) 1 2 . 8 ( 3 )

1 6 . 0 ( 1 ) 1 0 . 6 ( 3 )

1 5 . 6 { 1 ) 7 2 . 7 1 3 )2 9 . 8 ( 3 ) 0 . 4 4 ( 1 0 )

0 . 0 2 1 . 1 ( 9 )6 . 1 0 { 1 1 ) 7 . 5 0 ( 4 6 )

1 4 . 3 ( 1 ) L 4 . 6 ( 2 )2 6 . 9 t 4 ) 0 . s 9 1 2 2 )

2 . 7 s { 3 8 ) 1 3 . 8 ( 8 )2 . 5 2 ( 6 5 ) 1 5 . 5 ( 6 )

0 . 0 9 ( 3 ) 8 . s 1 ( 6 2 )0 . 1 1 ( 9 ) 0 . 4 9 ( 3 7 )4 , 2 4 1 6 ) 4 s . o ( 8 )0 . 1 6 ( 8 ) 1 7 . 7 ( 6 )0 . 1 7 ( 3 ) 8 . 1 3 ( 1 9 )0 . 0 0 . 0 9 ( 2 )0 . 1 9 ( 5 ) 4 4 . 2 1 3 1o . 1 3 ( 6 ) 1 6 . s ( s )0 . 2 5 ( 5 ) ' 7 . r 7 | [ t )

0 . 0 0 . 0 s ( 3 )0 . 3 5 ( 5 ) 4 A . 4 1 4 to . 2 0 ( 5 ) 1 5 . 6 ( 3 )0 . 3 1 ( 1 0 ) ' 7 . ' 7 9 l : . 5 l

0 . 2 3 ( 5 ) 6 . ' 7 2 1 ] . L )

0 . 3 0 { 7 ) 6 . s 9 ( 1 4 )

0 . 1 4 ( 3 ) 7 . 6 8 ( 1 5 )

0 . 3 0 ( 1 0 ) 7 . 2 5 ( 1 3 )0 . 0 0 . 1 7 ( 4 )0 . 3 6 ( 8 ) 4 0 . 5 ( 8 )0 . 1 8 ( 4 ) 1 6 . 5 ( 9 )0 . 1 1 ( 5 ) 5 . 1 2 1 2 1 ' )0 . 0 0 . 1 5 ( 7 )0 . 3 7 ( 3 ) 1 8 . 1 ( 1 0 )0 . 4 2 ( 1 3 ) 1 9 . 7 ( 1 3 )

1 , 1 , . 6 ( 2 ) 2 - 6 3 l L s l A . 4 4 1 2 J1 4 . 6 ( s ) 3 . 0 7 ( 1 7 ) 0 . 1 6 ( 3 )

a , 3 ' 7 1 2 ) , 0 . 0 0 . 01 7 . 5 ( 8 ) 0 . 4 4 ( 1 0 ) 0 . 01 1 . 8 ( 3 ) 2 . 6 4 ( 1 6 ) o . 4 2 t 4 J1 5 . 0 ( 4 ) 3 . 1 2 ( 1 s ) 0 . 1 5 ( 2 )

0 . 4 6 ( 6 ) 0 . 0 0 . 01 8 . 9 ( 8 ) 0 . 3 3 ( 6 ) 0 . 01 1 . 5 ( 1 ) 2 . 7 7 ( 1 8 ) 0 . 5 5 ( 3 )r . 4 . 1 ( 3 ) 3 . 3 1 ( 1 0 ) 0 . 1 8 ( 2 )

0 . 5 1 ( ? ) o . o o . o1 9 . 2 ( 3 ) 0 . 2 6 ( 8 ) 0 . 01 1 . 9 ( 1 ) 2 . 6 9 1 2 t ) 4 . 4 9 \ 2 J

1 0 . 5 ( 1 ) 3 . 3 ? ( 2 0 ) o . 8 0 ( 3 )

1 1 . 0 ( 1 ) 3 - 0 6 ( 1 0 ) 0 . 7 1 { 4 )

1 1 . 2 ( 1 ) 2 . ' 7 7 . 5 ) 0 . 5 7 ( 3 )

1 0 . 8 { 1 ) 2 . 6 5 1 6 ) 0 . 5 4 ( 4 )1 4 . 1 ( 3 ) 3 . 3 0 ( 1 3 ) 0 . 1 6 ( 2 )

0 . 4 0 ( 4 ) o . o o . o1 ? . 6 ( 8 ) 0 . 3 6 ( 5 ) 0 . 0

8 . 7 3 ( 1 2 ) 3 . 1 1 ( 6 ) 1 . 1 4 ( 3 )1 0 . 7 ( 6 ) 4 . 6 9 ( 1 5 ) o . 5 0 ( 5 )1 2 . 1 ( l s ) 0 . 3 8 ( 9 ) 0 . 0

8 . 5 9 ( 1 6 ) 0 . 4 0 ( 1 4 ) 0 . 0AT- 2 9+Mix

4 7 4 1 2 0 0 c l a s s 7 4 9 , 7 ( 4 ) 1 . 2 0 ( 1 1 ) 1 5 . 6 ( 2 )( 4 8 )

1 2 . 0 ( 3 ) 0 . 2 6 t 6 ) 5 . 1 5 ( 1 2 ) 9 . 5 1 ( 1 8 ) 3 . 0 4 ( 1 9 ) 1 . 3 2 { 4 ) 0 . 1 9 ( 7 ) 9 9 . 9 ' 1

BAKER AND EGGLER: ATKA MELT COMPOSITIONS

Appendix Table 1. Continued

27

A ' , h T 6 h ^ D h : c 6

( h r s . ) o c ( m o d e )n s io2 T io2 Al2o3 Feo* Mno Mgo CaO Na2O KZO P2o5 Total

3 . 5 9 ( 1 9 ) 0 . 5 7 { 7 ) o . 2 s ( 9 ) 9 4 . 5 84 . 1 - 6 1 7 ) o . 0 9 ( 3 ) o . o 9 9 . 7 4o . o 0 . 0 0 . 0 9 1 . 9 6o . o 0 . 0 0 . 0 9 8 . 6 8

l 1 ^ d f n t l

) L h : r F v b a r i m a h f <

s tar t ing compos i t ion : AT-11 2 3 4 1 1 2 0 c l a s s ( 3 2 ) 1 0 5 6 . 3 ( 1 2 ) 2 . 4 3 1 2 s J 1 4 . O l 4 ) 6 . 8 ? ( 9 9 ) 0 . 3 1 ( 1 2 ) 4 . 1 9 ( 3 3 ) 7 . 4 9 ( 3 8 ) 3 . 7 5 ( 2 6 ) 1 . 4 9 ( 8 ) a . 4 4 1 7 ) 9 7 . 2 7( 3 s ) ( - 8 . 8 ) P l a s ( 5 9 ) 5 s 4 . o ( 1 s ) o . o 2 S . O ( 6 ) 0 . 5 0 ( 3 ) O . o 0 . 1 8 ( s ) 1 1 . 8 ( 8 ) 4 . 6 1 ( 4 5 ) 0 . 2 6 ( 3 ) 0 . 0 9 9 . 3 s

o l i v ( 9 ) 4 3 9 . 8 ( 3 ) 0 , 2 3 1 7 ) 0 . 1 9 ( 6 ) 1 ? . 5 ( 8 ) 0 . 6 9 ( 6 ) 4 2 . 5 ( 5 1 0 . 3 9 ( 3 ) 0 ' 0 0 . 0 0 . 0 1 0 1 . 3 0A u s ( 1 ) 4 5 1 . 0 ( 2 ) 1 . 8 5 1 1 2 ) 3 . 0 ? ( 3 3 ) ? . 5 3 ( s 3 ) 0 . 5 5 ( 1 0 ) 1 6 . s ( 4 ) 1 8 . 1 ( 6 ) 0 . 3 6 ( 6 ) 0 . 0 0 . 0 9 8 . 9 6H2o conten t o f me l t : 1 .8 t1 l ron Loss 66?J

1 1 8 0 1 1 o o c l a s s ( 4 4 ) s s s . 7 ( 8 ) 1 . 9 9 ( 1 7 ) 1 s . 8 ( 9 ) s . 8 7 ( 3 0 ) 0 . 1 9 ( 7 ) 3 . 6 6 ( 3 7 ) ' 1 . 4 4 ( 2 5 J 3 . 9 ' 7 ( 9 ) 1 . 4 2 ( 1 5 ) 0 . 3 8 ( 5 ) 9 6 . 4 21 2 4 ) ( - 9 . 5 ) P l a s ( 4 7 ) 4 s 2 . 9 1 4 ) o . o 2 9 , 4 ( 5 ) O . s 7 ( 1 1 ) o . o 0 . 1 s ( 3 ) 1 - 2 . 6 \ 4 ) 4 . 0 8 ( 2 9 ) 0 ' 1 6 ( 3 ) 0 . 0 9 9 . 8 6

o l i v ( 7 ) 4 3 9 . 2 1 3 ) O . t 2 l I 2 ) 0 . 3 8 ( 2 3 ) 1 9 . 0 ( 8 ) 0 . 7 s ( 6 ) 4 1 . 1 ( 5 ) 0 . 4 3 ( 1 3 ) 0 . 0 0 . 0 0 . 0 1 0 0 . 9 8A u g ( 2 ) 4 5 0 . 4 ( s ) 1 ' . 5 5 1 1 ' 2 ) 4 . 5 0 ( 3 7 ) 6 . 6 6 ( 1 9 ) 0 . 4 0 ( 9 ) 1 5 . 6 ( 6 ) 1 9 . 0 ( 1 0 ) 0 . 5 2 ( 1 3 ) 0 . 0 0 . 0 9 8 . 6 3H r O c o n t e n l o f m e l t : 1 . 8 9 I r o n L o s s : 6 6 8

Star t rnq Composr t ron : AT- I12i , r 6 2 1 o G 5 c t ; s s ( G ) 5 5 3 : S l 5 ) r . 6 t ( 6 ) 1 5 . 2 ( 5 J 6 . 3 8 1 4 6 \ a . 2 4 1 ' t ) 2 . 2 6 ( L o l 5 . 9 5 ( 4 ) 3 . ' 7 r ( 2 1 ) 1 . s 8 ( 1 1 ) 0 . 3 0 ( 6 ) 9 6 . ' 7 3\ 2 4 ) ( - 9 . s ) P l a q ( 6 6 ) 5 5 3 . 5 ( ? ) O . O 2 s . 2 ( ' 7 ) 0 . 5 7 ( 1 1 ) O . O o . 0 5 ( 4 ) 1 2 . 8 ( 6 ) 4 ' 1 1 ( 2 0 ) 0 . 2 0 ( 4 ) 0 . 0 l ' 0 0 . 4 3

o l i v ( 1 3 ) 4 3 7 . 8 ( 6 ) 0 . 1 6 ( s ) o . o 2 8 . 7 ( ' 7 ) 0 . 6 ? ( 1 3 ) 3 2 . 9 ( 1 ) 0 . 3 4 ( 4 ) 0 . 0 0 ' 0 0 . 0 1 0 0 . 5 7A u s ( 1 6 ) 4 5 1 . 6 { 3 ) 0 . 9 5 ( 1 3 ) 2 . 8 4 ( 4 I ) 9 , 6 9 ( I 2 ) o . 4 L ( 2 ) 7 4 . 4 1 4 ) t - 9 . 2 1 3 ) 0 . 3 9 ( 1 9 ) 0 . 0 0 . 0 9 9 ' 4 8HrO conten t o f me l t : 1 .83 l ron Loss : 432

Slar l lnq ConDos l taon: A I -291 1 6 5 1 0 6 6 G l ; s s ( 6 3 ) 7 s 9 . 9 ( 9 ) 1 . 2 8 ( t s ) 1 5 . 2 ( 2 ) 6 . 9 0 ( 4 ) 0 . 2 5 ( 6 ) 2 . 2 4 1 1 , I ) s . 4 6 ( l . o ) 3 . s 8 ( 1 6 ) 2 . 4 4 ( 7 ) 0 . 3 8 ( 1 2 ) 9 ' 1 . 6 31 2 4 ) ( - 9 . 5 ) P I a s ( 2 8 ) s s ? . 1 ( s ) 0 . 0 2 7 , 0 1 2 ) O . ? 3 ( s ) o . o o . o s ( 4 ) 1 0 . 3 ( 3 ) 4 . 9 2 ( 1 6 ) 0 . 5 0 ( 3 ) 0 . 0 1 0 0 . 6 0

A u s ( 4 ) s s 1 . 9 ( 6 ) 0 . 5 5 ( 1 0 ) 1 . S 6 ( 1 6 ) 1 1 . 1 ( 4 ) 0 . 5 5 ( 1 1 ) 1 4 . 6 ( 4 ) 1 9 . 3 ( 3 ) 0 . 3 9 ( ? ) 0 . 0 0 . 0 1 0 0 . 2 5o p x ( 4 ) s s 3 . 8 ( 6 ) 0 . 3 4 ( 1 2 ) 1 . 3 8 ( 5 9 ) 1 ? . 0 ( 2 ) 0 . 6 1 ( 4 ) 2 4 . 6 ( 3 ) 2 . 2 s ( 1 3 ) 0 . 1 4 ( 8 ) 0 . 0 0 . 0 7 0 0 . 1 - 2H?O conten t o f me l t : 1 .88 I ron Loss : 308

1 1 4 8 1 o 2 O c l a s s ( 5 4 ) 5 6 0 . o ( 3 ) 1 . 1 2 ( 1 0 ) 1 s . 3 ( 3 ) 6 . 7 2 ( 3 0 ) o . L 3 ( 4 ) 1 . 3 2 ( 1 1 ) 4 . 3 4 ( I ' 7 ) 3 . ' 7 9 1 2 2 ) 3 . 2 6 ( s \ 0 . 2 8 ( 7 ) 9 6 . 2 6( 2 4 1 ( - 9 . 9 ) P l a s ( 3 3 ) s 5 6 . 3 ( 3 ) o . O 2 5 - 5 ( 5 ) 0 . 9 5 ( 1 s ) o . o O . 0 1 ( 2 ) 1 0 . s ( 4 ) 5 . 1 3 ( 2 8 ) 0 . 5 2 ( 4 ) 0 . 0 9 9 . 9 7

A u s i ? ) 5 s 1 . 9 ( 7 ) 0 . 5 9 ( ? ) r - . 6 8 c - 6 ) 1 1 . 0 { 4 ) 0 . 3 1 ( 9 ) 1 4 . 3 \ 2 ) 1 9 . 3 ( 5 ) 0 . 3 7 ( 9 ) 0 . 0 0 ' 0 9 9 ' 5 5o p x ( 6 ) 4 5 3 . ' 7 ( ' 7 ) 0 . 5 0 ( 4 ) 0 . 8 6 ( 1 8 ) 2 0 . 3 { 1 ) o . ? 3 ( 1 8 ) 2 2 . 6 ( s ) 2 . 0 2 \ 7 ) 0 . 1 0 ( 7 ) 0 . 0 0 . 0 1 0 0 . 8 1H2O conten t o f me l t : 1 .8% I ron Loss : L '72

r L ' t 4 1 O 2 O G l a s s ( 4 3 ) 4 6 6 . 1 ( 1 3 ) O . ' 7 ' 7 ( 2 6 ) 1 , 3 . 4 ( 4 ) 3 . 6 6 ( 1 9 ) 0 . 0 7 ( 7 ) 0 . 6 5 ( 9 ) 2 . 4 5 ( 1 5 ) 3 . 6 9 ( 3 ) 3 . 6 2 1 3 ) O . 2 3 1 7 ) 9 4 . 6 4( 9 6 ) ( - 1 0 . 3 ) P l a s ( 4 2 ) 5 5 5 . 5 ( 1 0 ) o . o 2 8 . 2 1 6 ) 0 . 7 4 ( 2 8 ) 0 . 0 o . o 3 ( 4 ) 1 1 . 4 1 4 1 4 ' 6 I ( 2 3 ) 0 . 4 0 ( 9 ) 0 . 0 1 0 0 . 8 8

A u s i 8 ) 5 s 1 . 6 ( 1 0 ) 0 . 4 8 ( 1 9 ) 1 . 9 8 ( 2 7 ) 1 1 . 2 ( s ) 0 . 5 2 { 1 3 ) 1 4 . 5 ( 2 ) 1 9 . 3 ( 5 ) 0 ' 2 s ( 4 ) 0 . 0 0 . 0 9 9 . 8 3o p i ( e ) 3 s 3 . 2 1 2 ) o . 4 0 ( 1 5 ) O . S 2 ( 3 6 ) 1 9 . ? ( 6 ) 0 . 8 3 ( 1 3 ) 2 3 . 0 ( 3 ) 2 . 0 3 ( 9 ) 0 . 0 6 ( 1 1 ) 0 . 0 0 . 0 1 0 0 . 0 4H2O conten t o f me l t : 1 .9% I ron Loss : 56?

7 2 4 9 1 0 2 0 c l a s s ( ? s ) 5 5 8 . 8 ( s ) 7 . 7 s l r 2 l \ 6 . 2 \ 1 1 4 . 3 2 ( 1 1 ) o . 2 0 ( 3 ) 2 . 5 0 ( ? ) 4 . 8 5 ( 8 ) 3 . 9 7 1 2 2 ) 2 . 6 4 1 4 ) 0 . 2 5 ( 5 ) 9 4 . 8 6( 9 6 ) ( - 1 0 . 3 ) P t a s ( 8 ) 4 5 2 . 5 ( 4 ) o . o 2 9 . 3 1 2 ! l o . s 2 ( 6 ) o . o o . 0 8 ( 4 ) 1 2 . 8 1 2 ) 3 . 7 8 ( 1 8 ) 0 . 2 8 ( 2 ) 0 . 0 9 9 . 2 6

A u q [ s 1 4 4 s l . 8 ( 2 ) o . ' t 2 ( 2 2 ) z . a 6 l 2 3 ) B . r 3 ( 2 2 ) o . A o l i ) 1 6 . 4 ( 1 ) 1 9 . 3 ( 3 ) o . 0 8 ( 5 ) 0 . 0 0 . 0 9 8 . 8 9H2O conten t o f me l t : 1 ,89 l ron Loss : 40*' 7 0 4 l o o o G i a s s ( 4 3 ) 5 6 5 . 2 ( 5 ) 1 . 1 1 ( 1 0 ) 1 4 . 6 ( 3 ) 2 . 0 ? ( 1 3 ) 0 . 0 8 ( 7 ) 0 . 9 2 ( 1 0 ) 2 . 4 6 ( r - 8 ) 4 . 0 6 ( 1 ) 4 . 1 0 ( 9 ) 0 . 2 6 ( s ) 9 4 . 8 s

( 9 6 ) ( - 1 1 . 8 ) P l a s ( 4 1 ) 4 5 6 . 5 ( 1 1 ) 0 . 0 2 ' , 7 . 0 ( 8 ) 0 . 6 8 ( 1 0 ) O . O O . 0 2 ( 2 ) L O . 2 ( 4 ) 5 - 2 1 1 2 7 ) 0 . 4 9 1 4 \ 0 . 0 l - 0 0 . 1 0A u s i 1 1 ) 5 5 1 . s ( 8 ) 0 . 4 7 ( 1 4 ) 1 . 8 8 ( 4 1 ) L o . 5 ( 8 ) 0 . 4 5 ( 1 4 ) l - 4 . ? ( 2 ) L 9 . 4 \ 4 ) 0 . 3 9 ( s ) 0 . 0 0 ' 0 9 9 . 5 9o p i ( s ) 1 5 3 . 2 a . 3 2 r . 4 4 1 s . 4 r . a 2 2 5 . 0 2 . 4 6 0 . 2 6 0 . 0 0 ' 0 9 9 . 1 0H2O conten t o f me l t : 2 .0% I ron Loss : '732

Star t ing Compos i t ion : AT-251 1 5 4 1 0 6 0 G r a s s ( 6 9 ) 9 6 1 - s T 1 l ) 1 . 0 9 c . s ) l s . 7 ( 2 ) s . 4 7 ( 3 0 ) 0 . 1 1 ( 6 ) 2 . 0 6 ( 1 4 ) 4 . 9 0 ( 3 6 ) 3 . 6 0 ( 1 s ) 2 . 8 7 ( 5 ) 0 . 2 7 1 6 1 9 7 . 5 t -( 2 4 ' ) ( - 9 . 5 ) P r a g ( 2 2 ) 6 5 4 . 0 ( 5 ) O - O 2 s , 5 ( 4 J 0 . 5 5 ( 1 0 ) o . o o . 0 9 ( 3 ) 1 2 . 3 ( s ) 4 . 0 7 ( 1 ? ) 0 . 3 3 ( 4 ) 0 ' 0 1 0 0 ' 9 4

o 1 i ; ( 3 ) 4 3 6 . 6 ( 1 0 ) O . O O . O 2 6 . 2 1 7 7 ) 0 . 4 8 ( 2 ) 3 4 . 5 ( 7 ) 0 . 3 6 ( ? ) 0 . 0 0 . 0 0 . 0 9 8 . 1 4A u s ( 5 ) 7 5 1 . 0 ( 5 ) 0 . 6 1 ( 8 ) 1 . 8 6 ( 5 4 ) 1 0 . 2 ( 1 3 ) o ' 3 5 ( 1 2 ) 1 4 . 8 ( 5 ) 1 9 . 9 ( 8 ) 0 . 3 1 ( 6 ) 0 . 0 0 ' 0 9 9 ' 0 4H2O conten t o f ne l t : 1 .9% I ron Loss : 29%

r I ' 7 3 1 o 2 O G i a s s ( 4 3 ) 5 6 8 . 4 { 8 ) 1 . 2 1 ( 3 s ) 1 4 . 2 ( 4 ) 2 . 1 3 ( 1 9 ) o . 2 0 ( ? ) 0 . 8 5 ( 1 4 ) 2 . 3 7 1 2 8 1 3 . 7 3 ( 2 5 ) 3 . 9 8 ( 1 1 ) 0 . 1 8 ( 1 9 ) 9 7 . I 9( 9 6 ) ( - 1 0 . 3 ) p l a s ( 4 3 ) 4 s s . ? ( 4 ) O . O 2 S . 1 ( 8 ) 0 . 4 1 i 1 4 i 0 . 0 O . O ? ( 4 ) 1 1 . 4 ( 3 ) 4 . 6 r ( 2 2 ) 0 . 4 5 ( 4 ) 0 . 0 1 0 0 . 7 4

A u s ( 8 ) 5 5 1 . 8 ( 6 ) O . 4 0 ( 1 1 ) 1 . 8 4 ( 3 7 ) 9 . ? 1 ( s 4 ) 0 . 4 9 ( 1 0 ) 1 s . 1 ( 4 ) 1 9 ' 8 ( ? ) 0 . 2 6 \ ' t ) 0 ' 0 0 ' 0 9 9 ' 4 0o p i ( z ) 4 5 3 . 8 ( 8 ) O . 2 6 1 2 r ) 1 . 0 9 ( 2 8 ) r ' 7 . 2 \ i 6 ) 0 . 5 8 ( 1 ? ) 2 5 . 1 ( 8 ) 1 . 9 3 ( 3 3 ) 0 . 0 1 ( 2 ) 0 . 0 0 . 0 9 9 . 9 ' 7H 2 O c o n t e n t o f m e l t : 1 . 9 % I r o n L o s s : 6 0 ?' t I 4 ' 7 1 0 2 0 c i a s s ( 6 0 ) 4 6 a . 6 1 2 ) 1 . o o ( 1 9 ) 1 6 . 0 ( 4 ) 5 . 5 5 ( 2 3 ) 0 . 1 4 ( 1 ) 1 . 8 0 ( 1 8 ) 4 . 4 1 \ 3 2 ) 3 . 6 1 ( 2 0 ) 3 . 5 8 ( 9 ) 0 . 1 8 ( 5 ) 9 6 ' 9 4

1 2 4 ) ( - 9 . 9 ) p r a s ( 2 8 ) 6 s ? . 0 ( 6 ) o . o 2 ' t . 6 l 4 1 0 . 6 3 i 4 ) o . o o . o 2 l 2 ) 9 . 9 2 ( 3 4 ' ) s . 0 4 ( 2 2 ) 0 . 6 1 ( 8 ) 0 . 0 1 0 0 . 8 2o l i v ( 3 ) s 3 8 . 0 ( 5 ) 0 . 1 1 ( 3 ) O . O 2 9 . 5 ( j ) 0 . 5 8 ( 3 ) 3 1 . 6 ( S ) 0 . 3 2 ( 9 ) O . O 0 . 0 0 ' 0 1 0 0 ' 1 1A u s ( 9 ) s s 2 . 3 ( 4 ) 0 . 5 7 ( 8 ) 1 . 4 6 ( 1 s ) 1 0 . 8 ( 2 ) 0 . 5 1 ( 1 1 ) 1 4 . 1 ( 3 ) 1 9 . s ( 3 ) 0 . 1 9 ( 3 ) 0 . 0 0 ' 0 9 9 ' 4 3H2o conten t o f me l t : 1 .8? I ron Loss : 24%

r 2 4 B L o 2 o G i a s s ( ? o ) 6 5 8 . 7 ( 8 ) 1 . 1 3 ( 1 9 ) 1 6 . s ( 6 ) s . 6 0 ( 6 9 ) 0 . 1 4 ( 9 ) 1 . s 6 ( 1 4 ) 4 . 5 5 ( 2 9 ) 3 . 8 8 ( 1 s ) 3 . 5 1 ( 1 7 ) 0 . 2 5 ( ] . 2 ) 9 5 . ? ?( 9 6 ) ( - 1 0 . 3 ) p r a s ( 1 9 ) 6 5 3 . 7 ( 6 ) 0 . 0 2 8 . 1 ( s ) o . ? 4 i 1 0 ) o . o o . 0 4 ( 2 ) 1 1 . 5 ( 5 ) 4 . 4 0 ( 4 2 ) 0 . 4 9 ( 6 ) 0 . 0 9 8 , ? ' 7

o r i ; ( 3 ) 5 3 5 . 3 ( 3 ) o . 1 0 ( 8 ) 0 . 1 8 ( 8 ) 3 3 . 2 1 4 ) 0 . 6 0 ( 1 5 ) 2 9 . 1 ( 5 ) o ' 3 5 ( 4 ) o ' o 0 ' 0 0 ' 0 9 9 ' 8 3A u s ( 8 ) 5 5 1 . 4 ( 3 ) 0 . 5 3 ( 5 ) r . 7 ' 7 ( 2 6 ) r A . 2 i 2 ) 0 . 3 ? ( 9 ) 1 4 . 6 ( 3 ) 2 0 . O ( 2 ) 0 . 2 9 ( 1 0 ) 0 . 0 0 ' 0 9 9 ' 1 6H 2 o c o n t e n t o f m e l t : 1 . 8 % I r o n L o s s : 1 4 t

5 kbar EM!149q!g

3 . 8 5 ( 1 4 ) 2 . 5 1 ( 9 ) a . 2 1 ( 4 ) 9 4 . 9 44 . 6 6 ( 2 2 ) 0 . 4 7 ( ' 7 ) o . o 9 9 . 6 ' 70 . 3 3 ( 9 ) 0 . 0 0 . 0 9 9 , 3 6o . o 7 ( 2 ) 0 . 0 0 . 0 9 9 . 2 7

4 . 2 1 - 1 2 4 1 4 . 6 5 ( 1 4 ) 0 . 5 2 ( 4 ) 9 7 . 4 95 . 1 - 8 ( 1 2 ) o . ? 8 ( 4 ) 0 . 0 9 9 . 9 L0 . 2 ' 7 0 . 0 0 . 0 9 ' 7 . 5 30 . 0 1 ( 1 ) 0 . 0 0 . 0 1 0 0 . 3 3

s ta l t ing Compos i t i -on : AT-291 2 8 ' 7 1 0 6 0 c l ; s s ( ? 4 ) s 5 ' 7 . 2 ( 9 , \ . 2 0 \ 7 4 ) 1 5 . ' 7 1 2 ) 6 . 5 8 ( 3 6 ) o ' 1 9 ( 4 ) 2 . r 3 \ 7 2 ) 5 . 3 1 ( l - 6 )( 1 0 3 ) ( - 9 . 6 ) P l a q ( 1 ? ) 4 5 4 . 3 ( 1 2 ) 0 . 0 2 8 . 2 ( 4 ) 0 . 6 3 ( 5 ) 0 . 0 a . 0 7 l 2 l L l - . 4 ( 2 )

A u s ( 6 ) 4 5 0 . 4 ( 5 ) o . ? 5 ( 9 ) 3 . 3 8 ( 8 7 ) 1 . 1 . 6 ( 2 ) 0 . 4 0 ( 8 ) ] ^ 4 . r ( 2 1 l - 8 . 4 ( 6 )o p x ( 3 ) 3 5 3 . 0 ( s ) 0 . 2 4 ( 1 0 ) 1 . 1 2 ( 8 ) 1 8 . 8 ( 3 ) 0 . 5 6 ( 9 ) 2 3 . 3 ( s ) 2 . 7 4 1 1 )H r O c o n t e n t o f m e l t : 1 . 8 9 I r o n t o s s : 2 2 %

Star t ing Compos i t ion : AT-251 2 8 6 1 0 6 0 c l a s s ( 4 8 ) 5 6 2 . U G ) 1 . 8 8 ( 2 8 ) 1 s . 2 ( 4 ) 4 ' 2 9 ( 2 5 ) 0 . 1 6 ( 5 ) 1 . 1 0 ( 8 ) 3 . 0 8 ( 1 7 )( 1 0 3 ) ( - 9 . 6 ) P l a s ( 3 5 ) s s 6 . 4 ( 7 ) 0 . 0 2 ' 7 . 2 1 3 \ 0 . 3 1 ( 4 ) 0 . 0 0 . 0 s ( 3 ) 9 . 9 9 ( 1 4 )

A u q ( 4 ) 7 4 9 . 9 0 . ? 1 4 . L 4 1 0 . 6 0 . 2 1 1 5 . 8 1 5 . 9o p x ( 3 ) 2 5 2 . 3 \ 9 ) o . s 7 ( 1 1 ) 2 . 8 6 1 2 4 ) 1 9 . 0 ( 1 2 ) 0 . 3 2 ( 1 6 ) 2 3 . I ( ' l ) 2 . I ' 7 1 4 1H2o conten t o f ne l t : 2 .02 l lon Loss : 41%

Star t ing Compos i t j .on : P-22+Mi .x7 3 2 4 1 1 o o G l a s s ( 4 8 ) ' 7 s L 9 ( 4 ) 1 . 9 0 ( 1 9 ) 1 8 . 9 2 ( 3 ' 7 J 2 . 3 8 ( 1 3 ) 0 . 0 5 . 8 s ( 2 9 ) 8 . 2 2 ( 2 4 )1 2 4 ) ( - 8 . 6 ) P I a q ( 1 8 ) 2 5 0 . 9 1 2 ) 0 . 0 3 7 . 2 ( r ) 0 . 1 7 ( 1 2 ) 0 . 0 0 ' 2 2 ( I l 1 3 ' 0 ( 3 )

o l i v ( s ) 3 3 9 . 5 ( 4 ) o . o o . o 8 . 2 ( 1 3 ) 0 . 0 5 0 ' 1 ( 1 8 ) 0 ' 1 8 ( 2 )A u s ( 3 0 ) 4 5 2 . 2 1 1 - 0 ) 0 . 7 6 1 2 7 ) 4 . 3 ( 1 5 ) 3 . 3 1 ( 6 2 ) 1 7 . 4 ( 1 3 ) 2 0 ' 3 ( 5 ) 0 . 4 1 1 2 )H 2 o c o n t e n t o f m e l t : L . 7 Z I r o n L o s s : 8 3 %

Note. Explanation of column headings:

Run number, below which in parentheses is listed the run duration in hours. Two duration values separated by a slash indicate acrystallization experiment; the rock was melted at high temperatures for the first duration then crystallized at lower temperatures forthe second.

Temperature of the experiment; for crystallization experiments the two temperatures, corresponding to the duration at each tempera-ture, are separated by a slash. Below the temperature(s), the oxygen fugacity of lhe experiment is listed in parentheses for the 1-atm,2-kbar. and s-kbar exoeriments.

Run(hrs )

Temp., 'G(log f",)

28 BAKER AND EGGLER:ATKA MELT COMPOSITIONS

Appendix Table 1. Continued

Phase The phases observed in the run products are listed, tollowed by the weight percent mode of each phase in the experiment (in paren-(mode) theses; corrected for Fe loss in the 2- and s-kbar experiments). See text for method of calculation.

n Number of analyses of each phase used to calculate mean and standard deviations of analyses.

The remaining columns are the oxide analyses of the run products. The mean value of the analyses is given followed by the standard deviation ofanalyses in terms of the smallest units recorded; thus 51 .4(4) is read as 51.4 + 0.4 wt%. Where a value of 0.0 is given, that oxide was not determined.FeO-: total Fe given as FeO.

1 Melts produced in experiments 813, 814, 816, 817, and 818 coexist with plagioclase, olivine, and augite'?HrO content of melt calculated by method discussed in text.3 Loss of Fe relative to initial bulk-rock Fe content.o Orthopyroxene was optically identified, but not analyzed with the electron microprobe.