Significance of hornblende in calc-alkaline andesites and ...mineral in some basaltic bulk compositions if P",o exceeds about l0 kbar (Yoder and Tilley, 1962; Green and Ringwood, 1968;

American Mineralogist, Volume 65, pages 837-851, 1980

Significance of hornblende in calc-alkaline andesites and basalts

A. T. ANDERSoN. JR.

Department af the Geophysical Sciences, The University of Chicago5734 South Ellis AvenueChicago, Illinois 60637

Abstract

Hornblende occurs in some andesitic and basaltic rocks of calc-alkaline affini1y, where it iscommonly associated with olivine. The texture, eruption history, and composition are inter-preted to indicate that hornblende commonly forms as a product of reaction between olivineand basaltic and andesitic liquids. The natural hornblendes probably formed within the crustat temperatures between 960o and 1080'C from liquids with less than about 6 weight percentHrO. The formation of hornblende from basaltic liquid within the crust has implications forevolution of continental crust, as well as for the origin of andesite and thermal conditions insubduction zones.

Introduction

Boettcher (1973) and others (Mueller, 1969; Hol-loway and Burnham, 1972; Green, 1972; Helz, 1976;Cawthorn and O'Hara, 1976; Allen and Boettcher,1978) argue that amphibole plays an important rolein governing the compositions of calc-alkaline ande-sites. The mechanisms whereby amphibole suppos-edly exerts its role are not clear. Is it left behind in acrystallins residue-where is it: in the crust or in themantle? Does amphibole crystallize and separatefrom a parental liquid-what are the compositionand temperature of the parental liquid? Does amphi-bole crystallize from a derivative liquid which mixeswith a parental liquid to yield andesite? Amphiboleoccurs in some calc-alkaline volcanic and plutonicrocks with basaltic and andesitic bulk compositions,but the significance of such occurrences is uncertainwithout textural evidence of how it got there: is theamphibole a primary igneous mineral or a product ofalteration? If igneous, over what interval of crystalli-zation did it form, and was the composition of theliquid in that interval andesitic, dacitic, or basaltic? Ifamphibole can be shown to have actually formedfrom a fiquid with a certain composition, what doesthat portend for the concentration of HrO in and thetemperature of the liquid? My aim is to help pointthe way to answers to some of the above questions bycritically sxxmining evidence on the growth of am-phibole from basaltic and andesitic liquids, both ex-

Definitions

In using the word liquid, I mean to refer to a singlestate of matter (a homogeneous liquid phase). It is vi-tal to distinguish between the composition of a liquid(generally a residual liquid) and the composition ofthe bulk material. Petrographically a certain horn-blende may be associated with a particular residualliquid. Experimentally a certain bulk compositionmay crystallize hornblende at or below the liquidus.The residual liquid associated with hornblende willgenerally differ in composition from the bulk. Ther-modynamically the composition and temperature ofa certain liquid can remain constant as the propor-tions of crystals and liquid vary. To interpret a givenassociation of hornblende and liquid, the petrogra-pher must seek experimental conditions which yieldhornblende and liquid with compositions similar tothose observed. If analogy is drawn with a particularbulk composition (and liquidus hornblende) the im-plied pressure (and concentration of HrO in the liq-uid) is larger than if analogy is made with a residualliquid. The implications for temperature are complexand depend on the concentration of HrO in the liquidas well as on compositional features such as Fe/Mgand concentrations of alkalis.

Hornblende in experimental products

Experimental studies yielding hornblende in prod-ucts with either calc-alkaline bulk compositions orwith residual liquids having analogous compositionsperimental and natural.

0003-o04x/80/09 I 0-0837$02.00 83'l

ANDERSON: HORNBLEN DE IN CA LC-A LKA LI N E A N DES ITES

reveal: (l) Hornblende is a liquidus or near-liquidusmineral in some basaltic bulk compositions if P",oexceeds about l0 kbar (Yoder and Tilley, 1962;Green and Ringwood, 1968; Allen et al., 19751' Sternet al., 1975) and if H,O dissolved in liquid exceedsabout 12 weight percent. (2) In some andesitic bulkcompositions hornblende is a liquidus mineral rf Pn,oexceeds about 2 kbar (Piwinskii, 1973). (3) As P",odecreases below P,o,u,, the hornblende-out curve liesprogressively farther below the liquidus (Robertsonand Wyllie, l97l; Holloway and Burnham, 1972;Green, 1972; Eggler and Burnham, 1973; Allen andBoettcher, 1978). (4) The proportion and composi-tion of residual liquid change dramatically over aninterval of temperature less than 50"C after the be-ginning of crystallization of hornblende (Hollowayand Burnham, 1972). (5) Near the liquidus low oxy-gen fugacity favors the crystallization of hornblendecompared to pyroxene (Helz, 1973, 1976). (6) The re-actions between hornblende, liquid and other crystalsare complex (Bowen, 1928, p. 61, 85, lll; Helz,1976). (7) Hornblende is stable to the highest temper-atures and lowest concentrations of HrO in liquid insystems (a) where olivine is present (Helz,1976), (b)where P,,o ( P,",", (Holloway, 1973), and (c) wheresignificant F is present (Holloway and Ford, 1975).

The above relations are summarized on Figure l.The coordinates of Figure I are qualitative. Quan-titative values will vary with bulk composition.Lesser MgO and Mg/(Fe+Mg) will diminish thetemperature of olivine, pyroxene, and hornblendeliquidus surfaces. The field for liquidus hornblendewill then extend farther into the vapor-absent portionof the diagram. Lesser concentrations of AlrO. willshift more of the liquidus surface into the olivine andpyroxene fields and diminish the area of the liquidusfield for plagioclase. Consequently, the combinationof low Al,O, and high MgO may obliterate the liq-uidus surface having hornblende, olivine, and plagio-clase (see Bowen, 1928, p. 1l l). For a natural magmawith CO, as well as HrO, there will be a range in X",oin addition to a boundary marking vapor-absent andvapor-present regions. Reactions between olivine,pyroxene, hornblende, and liquid add complicationsnot shown on the diagram. Spinel and magnetite areprobable additional liquidus minerals but are notshown. For simplicity only one pyroxene is depicted.Calcium-rich clinopyroxene is the liquidus pyroxenebut is joined atd/or replaced by orthopyroxene be-low the liquidus at temperatures below or near thatof the appearance of hornblende. Except for such

Fig. l. Simplified hypothetical P-T-XH2o equilibrium diagram

for a high-alumina basalt. The diagram is drawn in perspective.

The origin is not shown. The concentration of H2O refers to the

bulk material. The stippled and dashed regions are saturated with

vapor rich in H2O. Below the liquidus surface the pyroxene (Px)-

out and olivine (Ol)-in surfaces are assumed to coincide for

simplicity. Fields for spinel and magnetite are omitted for

simplicity. Hornblende (Hb) is a liquidus mineral in both vapor-

absent and vapor-present regions. The liquidus hornblende is

accompanied either by liquidus Ol or by both plagioclase (Pl) and

Ol. Only one symbol is shown for pyroxene (Px). Liquidus Px is

calcium-rich. Low-calcium pyroxen€ probably predominates if Hb

is abundant.

simplifications, the diagram is consistent with exist-ing experimental and natural data.

It would be helpful to know the positions of sur-faces of equal SiO, in the liquid (H,O-free basis). Atpresent the coordinates of such surfaces in Figure Iare uncertain. Qualitatively, it is possible to givesome limits. I consider a surface of SiO, : 60 weightpercent on an anhydrous basis, correspondingroughly to andesite. Because the composition of theresidual liquid changes steeply with temperature af-ter the beginning of crystallization of hornblende(Holloway and Burnhan, 1972), it is reasonable toexpect the 60 percent SiO, surface to lie close to thehornblende-out surface. In the region ofthe diagramwhere hornblende is a liquidus mineral (hornblende-

1T

f-t--ii:)

a\*> 2,,/ ./'//1:1a(ry2

ANDERSON: HORNBLENDE IN CALC-ALKALINE ANDESITES

out corresponds to liquidus), the 60 perc€nt SiO, sur-face must lie below the hornblende-out surface. Inthe region ofthe diagram where the hornblende-outsurface is far below the liquidus (for example, sub-aerial basalts with about 0.1 weight percent HrO), re-sidual liquids with more than 60 percent SiO, maydevelop in the absence ofhornblende. Therefore, the60 percent SiO, surface may intqisect the horn-blende-out surface, unless crystallization intervenes.

It is probable, but uncertain, that residual liquid inanhydrous high-alumina basalt attains 60 percentSiO, before complete crystallization if equilibrium ismaintained. Residual glasses in natural high-aluminabasalts have 62 (Fuego volcano, Guatemala-Rose e/al., 1978) and 67 (Hat Creek flow, California-An-derson and Gottfried, 19'71) percent SiOr. However,equilibrium is not maintained throughout the naturalrocks. Because the minerals [plagioclase (An55), aug-ite, olivine (Fo66), and magnetitel associated withthe residual glass at Fuego are compositionally simi-lar to the'normative minerals [plagioclase (An50),pyroxene, olivine (Fo70), magnetite, and ilmenite,based on the weighted average composition with l/3of the iron trivalent], I conclude that anhydrous,calc-alkaline, high-alumina olivine tholeiites ex-emplified by Fuego probably yield equilibrium resid-ual liquids with 60 percent or more SiO, before com-plete crystallization. Therefore the equilibriumsurface for residual liquids with 60 percent SiO,probably intersects the hornblende-out surface.

The intersection of the 60 percent SiO, and horn-blende-out surfaces may have major petrological sig-nificance. In a given bulk composi$on residual liq-uids with 60 percent SiO, will vary compositionallyin accordance with the amounts of associated miner-als, particularly hornblende. Many workers havepointed out that crystallization of hornblende canyield calc-alkaline residual liquids (Green and Ring-wood, 1968; Boettcher, 1973; Holloway and Burn-ham, 197 2; Heh, 197 6; Cawthorn and O'Hara, 1 976).Andesitic residual liquids not associated with horn-blende may be alkalic or alkali-calcic. The residualliquid with 60 percent SiO, may have both a smallerratio of Fel(Mg+Fe) and a lower temperature ifhornblende is present than if it is absent. For a givenbulk composition the amount of hornblende associ-ated with a residual liquid having 60 percent SiO,will be maximized: (l) if the initial concentration ofHrO is large; (2) if crystallization occurs at constantor increasing pressure; and (3) if equilibrium (reac-tion) is maintained. Consequently, many differencesbetween various andesitic liquids may be explained

by either contrasting initial concentrations of HrO inparental high-alumina basaltic magma, variabledepth of crystallization, or extent of reaction duringcrystallization.

Some natural calc-alkahne liquids associated withhornblende are less siliceous and hotter than mostexperimental analogs. The natural occurrenees sug-gest that the field of liquidus hornblende extends tolower concentrations and partial pressures of HrOthan is indicated by most existing experimental re-sults.

Hornblende associated with natural andesitic liquid

Many andesites have phenocrysts of hornblende(see for example, Moorhouse, 1959, p. 189-190;Turner and Verhoogen, 1960, p.272-282) but mostdo not. Many, perhaps most, andesites contain clotsof anhydrous crystals which may be reaction prod-ucts of hornblende (Stewart, 1975); however, the oc-currence of oscillatory zoned plagioclases withinsuch clots suggests a different origin (Garcia and Ja-cobson, 1979).

Most hornblende andesites have more than about60 percent SiO, (Kuno, 1950; JakEs and White, 1972;Sakuyama, 1979); the groundmass probably is daciticto rhyolitic in composition. The hornblendes com-monly are rimmed or cored by crystals of pyroxene,plagioclase, and other minerals. The compositions ofthe hornblendes may be modffied after crystalliza-tion. Thus the compositional relations between horn-blende and andesitic liquid are difficult to infer frommost occurrences.

Below I describe an association ofhornblende andandesitic glass included in olivine from a lapillus ofporphyritic andesite.

The 1783 eruption of Asama Volcano, Japan. Back-ground: the eruption and its products

Asama volcano is a composite andesitic strato-volcano in central Honshu, Japan. Its geology isdocumented by Aramaki (1963) and briefly surnma-rued by Anderson (1979).

The history of the 1783 eruption helps constrainthe history of cooling and decompression which af-fected the extruded products. Although complex, thiseruption probably is typical of the principal wayAsama has grown. It involved a succession of Plinianexplosions which produced about 0.17 km3 of pyro-clastic material (Minakami, 1942) and an equal vol-ume of lava (Aramaki, 1956). The early productswere air-fall pumice lapilli and ash; ash flows formed


on and after August 3 when the culminating blast oc-curred; an andesitic lava flow formed during the finalphase of the eruption. The details are reported byAramaki (1956, 1957), who interpreted the eruptionsequence in terms of an initial vent-clearing phasebetween June 25 and August 2, 1783 followed b! ex-trusion from a subvolcanic body of magma on andafter August 3, 1783.

Petrography. The 1783 andesitic pumice lumpsvary from light to medium brown; some have streaksof different shades of brown. Lapillus AM2-2 (An-derson, 1979) is medium brown. It was collectedfrom a layer of air-fall lapilli which probably formedon or before August 2.

The pumice contains mm-sized phenocrysts ofplagioclase, augite, hypersthene, magnetite, and il-menite. Plagioclase is the principal phenocryst andmakes up about 20 volume percent of the pumice ona void-free basis. Most of the plagioclases are sub-hedral to anhedral crystals with irregular cores sievedwith vermicular inclusions of glass and gas. Thecores are surrounded by a normally-zoned rim whichis free of inclusions and l0 to 50 pm thick. Otherplagioclases are poor in inclusions and oscillatory-zoned with skeletal protrusions. The augites containinclusions of glass and lamellae of pyroxene in irreg-ular, pale green cores. Some augites occur in clotswhere they are subhedral and irregular against eachother. Commonly a zone of inclusions of plagioclaseand pyroxene marks the boundary between a palegreen core and deeper green rim. The rims are 50 toabout 150 pm thick. Glasses included in pyroxene,plagioclase, and oxide phenocrysts range from about70 to 79 weight percent SiOr; the host glass outside ofthe phenocrysts has abont 72 percent SiO, (Ander-son, 1979).

Crystals of olivine (Fo79-Fo85) are rare. The oli-vines have 5O-pm-thick rims of hypersthene or aug-ite. There is more olivine in medium-brown pumicethan in light brown pumice. Inclusions of glass in oli-vine crystals have 54 to 6l weight percent SiO, on ananhydrous basis (Anderson, 1979).

Hornblende associated with andesitic glass. Horn-blende occurs together with andesitic glass includedin an olivine crystal from the lapillus AM2-2 of me-dium-brown andesitic pumice (Fig. 2, Table 1). Thehornblende is pale brown and faceted against glass.A second mineral (probably orthopyroxene) occurswith hornblende in the inclusion. It is coloiless, longand prismatic, and has a refractive index close to thatof the hornblende. The included glass has a roundcontact with the otvine host crystal. The microprobe

Fig. 2. Inclusion of glass (G), hornblende (H), and probable

pyroxene (P) in o l iv ine (O) The long dimension of the

photograph is about 200 pm. Dip and strike symbols refer to the

attitude of the contact between olivine and glass referred to the

polished surface as horizontal The olivine crystal is rimmed by

orthopyroxene and is from a lump of pumice extruded from

Asama volcano, probably on or before August 2, 1783.

analysis yields an estimated 4 percent of HrO in theglass (Anderson, 1979).

A second, smaller inclusion in the same olivinecrystal contains a pale brown crystal which is prob-ably hornblende (Fig. 3). It comprises about 2 vol-ume percent of the inclusion. Anisotropic andopaque crystals line the inside wall of the vaporbubble in the small inclusion.

The otvine host crystal of the hornblende-bearinginclusion is Fo83 (Table l) compared to Fo79 toFo85 for olivines from other lumps of pumice of the1783 eruption (Anderson, 1979). The composition ofthe preserved glass (Table l, column 2) is close to butmore silicic than that of other inclusions in olivinewhich lack hornblende (Table l, column 4).

Hornblende comprises about 30 volume percent ofthe large inclusion (Fig. 2), some of which was re-moved by sectioning. Probable orthopyroxene com-prises about l0 volume percent. Most inclusions inAsama olivines are subspherical to ellipsoidal. Thehornblende-bearing inclusion has walls which dipaway from the inclusions at high angles (Fig. 2).

Attached phenocrysts of magnetite and ilmeniteoccur in the same lump of pumice. Their composi-


Table l. Chemical compositions of hornblende and andesitic liquid with comparisons

841

C o l . I

S i 0 z

A 1 2 0 3

Fe02

Mgo

Ca0

N a 2 0

Kz0

T i 0 2

C I

Sum3

H r 0 al 5

p 6

As soc .

4 ? . 6

8 . 2l q 6

12.7

L . d

n ?

2 . 4

0 . 0 5

98.2

3

1080-960

0 . 1 - 1 ?

F o 8 3 , P x ?

5 6 . 0

t9.7

6 . 9

3 . 0

1 0 . 0

2 . 9

0 . 6

0 . 8

0 . 0 8

9 5 . 5

3

1080-960n 1 - 1 ?

F o79

6 0 . 5

2 0 . 0

4 . 2

1 . 1

8 . 4

3 . 2

0 . 8

1 . 3

n . o .

e 7 . 3

i0

1 0 1 5

5

H b , C p x ,0 l

5 5 . 5

1 7 . 3

8 . 5

3 . 4

9 . 2

L . 6

0 . 7

2 . 2

n . o .

8 5 . 7

1 0

1045

5

H b , C p x ,01 ,0x

63.2

20.3

3 . 0

0 . 3

7 . 4

3 . 0

0 . 9

1 . 2

n . o .

8 4 . 0

10

1050

8 . 0

H b , C p x ,0 l ,Mt

6 1 . 3 5 7 . 0

2 t . 0 2 0 . 9

3 . 4 3 . 6

2 . 6 3 . 4

5 . 5 8 . 9

4 . 3 4 . 0

1 . 5 1 . 2

0 . 3 1 . 2

0 . 1 s 0 . I 2

9 7 . 4

4

1080-960

0 . 1 - l ?

H b , F o B 3 ,P x ?

6 3 . 8 6 0 . 0

1 8 . 6 2 0 . 4

4 . 3 4 . 5

0 . 7 4 . 3

6 . 9 7 . 5

2 . 8 0 . 9

1 . 0 0 . 9

t . 2 7 . 2

n . d . n . d .

8 3 . 7 n . r .

6 ? 1 0

999 1010

5 . 2 1 3

C p x , H b , H b , C p xO l , M t 0 x

rsee key to co lumns be low.

' A l I r r o n a s F e u .

35um o f the or ig ina l ana lys is . Repor ted ana lyses are reca lcu la ted to 100 percent except fo r round ing er ro rs .Or ig ina l sum fo r co lumn 9 is no t pub l i shed. Sums are Hz0- f ree .

4 E s t i m a t e d c o n c e n t r a t i o n o f H z 0 . E s t i m a t e d b y d i f f e r e n c e ( A n d e r s o n , 1 9 7 3 ) f o r c o l u m n s I , 2 , 4 ; e s t i m a t e d f r o mpressure , 9as compos i t ion and so lub i l i t y o therw ise .

s E s t i m a t e d t e m p e r a t u r e i n d e g r e e s C e l s i u s . S e e t e x t f o r c o l u m n s 1 , 2 , 4 . 0 t h e r w i s e p u b l i s h e d r u n t e m p e r a t u r e .

6Est imated pressure in Kbar . See tex t fo r 1 and 2 . o therw ise run pressure ( to ta l p ressure no t necessar i l y samea s P H " 0 / .

'Co. * i r t inq minera ls : Fo = o l i v ine w i th a tomic ra t io o f . l00

t lS / (Mg+Fe) ; t tU = hornb lende, Px = pyroxene, Cpx =

c a l c i u m - r i c h p y r o x e n e , 0 l = o l i v i n e , M t = m a g n e t i t e , 0 x = o x j d e .

Key to Co1 umns:

L = H o r n b l e n d e i n c l u d e d i n o l i v i n e t o g e t h e r w i t h g l a s s o f c o l u m n 2 ( a n a l y s i s 9 o f T a b l e 3 o f A n d e r s o n , 1 9 7 9 ) .

2 = A n d e s i t i c g l a s s i n c l u d e d i n o l i v i n e , A s a m a 1 7 8 3 ( a n a l y s i s 8 o f T a b . l e 3 o f A n d e r s o n , 1 9 7 9 ) '

3 = Resu l t o f add ing 0 .4 par ts hornb lende (co lumn 1) , 0 ,7 par ts g lass (co lumn 2) , and 0 .1 par t o f o r thopyroxene( s t o i c h i o m e t r i c E n 8 3 ) a n d s u b t r a c t i n g 0 . 2 p a r t s o l i v i n e ( F o B 3 ) .

4 = A n d e s i t i c a l a s s i n c l u d e d i n o l i v i n e , A s a m a 1 7 8 3 ( a n a l y s i s 1 0 o f T a b l e 3 o f A n d e r s o n , 1 9 7 9 ) .

5 = S y n t h e t i c a n d e s i t i c g l a s s ( a n a l y s i s 1 5 a o f T a b l e 7 b o f H e l z , 1 9 7 6 ) .

6 = S y n t h e t i c a n d e s i t i c g l a s s ( a n a l y s i s 1 6 a o f T a b l e 7 b o f H e l z , 1 9 7 6 ) .

7 = S y n t h e t i c a n d e s i t i c g l a s s ( f r o m T a b l e 8 o f H o l l o w a y a n d B u r n h a m , 1 9 7 2 ) .

8 = S y n t h e t i c a n d e s i t i c g l a s s ( f r o m T a b l e B o f H o l l o w a y a n d B u r n h a m , 1 9 7 2 ) .

9 = Synthe t ic andes i t i c A lass (average o f ana lyses 2 l and ?2 o f Tab le 6a o f A l len and Boet tcher , 1978) .

tions (Hm24, Usp34) suggest a temperature of equili-bration of 1080'C, according to the unpublished geo-thermometer of Lindsley and Rumble (1977). Thesame compositions suggest a temperature of 10600 to

1070"C, according to my visual interpolation of databy Buddington and Lindsley (1964) and Taylor(1964). Similar compositions but lower temperatures(990"-1020"C) were reported by Oshima (1976).

842

Fig. 3. Photograph of a second, smaller inclusion in the samecrystal as that shown in Fig. 2. The inclusion is about 30 pm longand contains a pale brown crystal of probable hornblende (H) anda dark bubble of vapor.

Interpretation

The andesitic glass (Table 2, column 1) probably isin a state of quenched equilibrium with olivine(about Fo83) and hornblende, because the pumicelump in which it occurs was explosively erupted andcooled in air before falling to the ground. The in-clusion probably cooled from about 1080'C to a fewhundred degrees C in a few minutes, judging fromthe Plinian explosions which probably erupted it inlate July or early August of 1783 (Aramaki, 1956;written communication, 1979). Possibly the inclusionwas contained in a froth of magma at a few hundredatmospheres or less for a few days, because it was ex-truded during the early part of the eruption when thevent probably was intermittently filled with froth.

I interpret the textural, compositional, and histori-cal facts to suggest that andesitic or basaltic liquidcarried crystals of olivine into a crystal-rich body ofandesitic magma during or before the early stages ofthe eruption (May 9 to August 2). The irregular an-hedral cores of many plagioclase crystals suggestfragmentation. The abundance of gas in the sievedcores suggests that fragmentation was accompaniedby and/or caused by effervescence. The rims on thecrystals of plagioclase, pyroxene, and olivine indicate


that reaction and crystallization took place after frag-mentation. The faceted (not skeletal) shape and sizeof hornblende included in olivine suggest it to be adaughter crystal or reaction product rather than aquench crystal or product of devitrification. Theround contact between the olivine and glass associ-ated with hornblende is consistent with a reaction re-lation between olivine, liquid, and hornblende. Crys-tals intergrown with a residual rhyolitic liquid werebroken apart and overgrown probably because in-vading, olivine-porphyritic magma caused efferves-cence of the interstitial liquid (due to pressure releaseand/or temperature increase) and contaminated theliquid with relatively refractory constituents.

Liquids trapped within the olivine crystals becamemodified in composition as a result of growth ofdaughter crystals of hornblende and/or pyroxeneand reaction with olivine, because the temperaturedecreased from that in the initial environment tonear that of the invaded andesitic magma with resid-ual rhyolitic liquid.

The initial composition of the trapped liquid be-fore crystallization of hornblende and pyroxene canbe estimated by adding 0.4 parts hornblende, 0.1parts pyroxene and 0.7 parts andesitic glass, and sub-tracting 0.2 parts olivine. The result (Table l, column3) is somewhat similar in composition to other in-clusions of andesitic glass in olivines which lackhornblende (Table l, column 4). If other minerals areincluded in the calculation, the similarity can be in-creased. Some inclusions of glass in the Asama 1783olivines have as little as 54 percent SiO, (Anderson,1979). lt is not possible to relate all the glasses to aconmon precursor without including either mineralsnot now in contact with the glasses, or residual rhyo-litic liquids as mixing components, or both. The ori-gins of the glasses included in the olivines are com-plex if considered in detail. Such detail is beyond thescope of this paper. Subtraction of olivine in theabove calculation corresponds to resorption of oli-vine concomitant with crystallization of hornblendeand pyroxene from a less silicic liquid precursor.Probably the andesitic liquid was in equilibrium withhornblende, pyroxene, and olivine at the temperatureof eruptive quenching; however, the relation with oli-vine probably was one of reaction.

Two kinds of experimental results can be com-pared with the natural association of hornblende, py-roxene, olivine, and andesitic liquid. Consequently, itis possible to obtain two estimates of the temperatureand the concentration of HrO in the silicate liquid. Inthe first case comparison is made with the liquidus

ANDERSON: HORNBLENDE IN CALC.ALKALINE ANDESITES

Table 2. Chemical compositions of hornblendes and associated basaltic materials

843

C o l . 1

s i02A l z 0 g

Fe02

Mgo

Ca0

Na20

K z 0

T i 0 2

c l

5Um"

Hr0q

T 5p 6

Assoc . 7

4 2 . 8

1 3 . 6

1 0 . 5

1 . 5 . J

12.22 ?

0 . 4

2 . 7

0 . 0 5

9 8 . 9

1

10301 '

An92

1 8 . 19 . 4

3 . 5

o . J

3 . 7

1 . 0

1 . 2

0 . 1 2

9 3 . 8

4

10 30r - a

An86

5 i . 0

1 9 . 1

1 0 . 0q 4

v . J

3 . 0

0 . 6

1 . 1

0 . 0 9

9 4 . 3

4

1030

I - 2F o79

4 2 . 1

1 4 . 3

i 0 . 8

1 5 . 6

1 1 . 82 . 5

0 . 32 ' l

n . d .

9 9 . 8

1 . 8

n . 0 .

8

0 l , C p x ,P I , M t

5 r . /

1 8 . 4q ' 1

1 0 . 32 A

0 . 5

1 . 0

n . o .

9 9 . 8

n , o .

n . o .

8

0 1 , C p xH b , P I ,Mt

41.4

1 6 . 5

12.4

1 4 . 1

1 1 . 0L . O

0 . 1

1 . 8

n . d .

98 .7

n . o .

1000

12-30

Fo83

5 4 . 0

1 7 . 7

9 . 2

4 . 8

8 . 3

3 . 8

0 . 8

n . o .

9 9 . 8

n . d .

1000

L2-30

O I , P I

40 .7 55 . 5

1 3 . 3 1 7 . 3

L 4 . 2 8 . 5

1 3 . 0 3 . 4

1 0 . 7 9 : 3

2 . 0 2 . 8

0 . 5 0 . 7

4 . 4 2 . 2

n . d . n . d .

9 8 . 7 8 5 . 7

n . d . 8

1045 1045

5 5

Fo82-84 Fo82-84C r , C p x H b , C r , C p x

Key to Co lumns (see Tab le 1 fo r foo tno tes) :

1 = Hornb lende inc luded in core o f p lag ioc lase phenocrys t f rom Fuego (ana lys is repor ted in Iegend to F igure 16 ,p . 2 0 o f R o s e e t a l . , 1 9 7 8 ) .

2 = G l a s s i n c l u d e d i n r i n o f p l a g i o c l a s e p h e n o c r y s t ( A n 8 3 ) o f c o l u m n 7 ( a n a l y s i s N o . 6 o f T a b l e 6 o f R o s e e t a l , ,1 9 7 8 ) .

3 = G l a s s j n c i u d e d i n o l i v i n e ( F o 7 9 ) f r o m F u e g o ( a n a l y s i s 4 o f T a b l e 6 o f R o s e e t a l . , . l 9 7 8 , a n d c l o s e i n c o m p o s i t i o nto the we igh ted average o f the 1974 erup ted ash) .

4 = H o r n b l e n d e f r o m g a b b r o i c b l o c k , L a S o u f r i E r e , S t . V i n c e n t ( a n a l y s i s 1 o f T a b l e 6 o f L e w i s , 1 9 7 3 a ) .

S = f ! lg g ra ined, in te rs t i t ia l mater ia l f rom same gabbro ic b lock as co lumn 9 (ana lys is 1 o f Tab le 4 o f Lewis ,I 9 7 3 b ) .

6 = H o r n b l e n d e i n c l u s i o n s i n o l i v i n e p h e n o c r y s t s , L i t t l e l Y t . H o f f m a n b a s a l t , C a l i f o r n i a ( c o l u m n 2 , T a b l e 2 o fMertzman, i978) .

7 = Groundmass o f L i t t le Mt . Hof fman basa l t (ca lcu la ted by subt rac t ing 10 percent o l i v ine IFo80] and 25 percentp lag ioc lase [An70] f rom the bu lk rock- -ana lys is 1 o f Tab le I o f l4er tzman, 1978) .

8 = S y n t h e t i c h o r n b l e n d e ( t i e l z , t g Z O , T a b l e 1 3 a , N o . 2 , p . 1 9 2 ) .

9 = S y n t h e t i c l i q u i d a s s o c i a t e d w i t h h o r n b l e n d e ( c o l u m n B ) c a l c u l a t e d b y H e l z ( 1 9 7 6 , T a b l e 7 b , a n a l y s i s N o . 1 6 a ,p , t 5 z ) .

phases of an andesitic bulk composition. In the sec-ond case comparison is made with the residual liquidin a basaltic bulk composition.

For the first comparison two assumptions are nec-essary: (l) the hornblende is a hydroxy-hornblende;(2) the minerals in contact with the natural andesiticliquid are thermodynamically equivalent to the liq-uidus phases of a system composed of the liquidalone.

The first assumption is justifiable because: (l) thenatural hornblende contains negligible chlorine; (2)

its pale color suggests a negligible concentration ofoxycomponent; and (3) its fluorine content is un-known, but probably small, because fluorine is a mi-nor constitutent of most magmas. The second as-sumption may be questioned because olivineprobably is in a reaction relation. For the presentpurposes I assume that the presence of olivine maybe ignored. Consequently, I seek experimental condi-tions which yield hornblende and pyroxene on theliquidus of andesite. The compilation of Stern et al.(1975) suggests that compositionally similar andesitic


liquids have hornblende and clinopyroxene on theliquidus if the temperature is below about 980oC andthe liquid contains more than about l0 weight per-cent HrO. The results of Allen and Boettcher (1978)on andesitic bulk compositions are consistent withthe above estimate: more than l0 weight percent H2Oin the liquid.

For the second comparison only one assumption isnec€ssary: the hornblende is a hydroxy-hornblende(ustffied above). I now seek experimental conditionswhich yield hornblende, pyroxene, and andesitic liq-uid in a reaction relation with olivine. Holloway andBurnham (1972) and Helz (1973, 1976) encounteredandesitic residual liquidus associated with horn-blende, pyroxene, and olivine (and in some casesmagnetite) at temperatures between about 1000 and1050'C and with about 6 to l0 weight percent HrOin the silicate liquid (Table l, columns 5-8). Allenand Boettcher (1978) found a comparable residualliquid (Table l, column 9, but no olivine). However,the compositions of residual liquids are difficult to es-tablish for experimental products. The detailed as-sessment by Helz (1976) suggests that the residualliquids at temperatures greater than about l0l5'C(at P",o : 5 kbar) are less siliceous than the naturalliquid. The results of Allen and Boettcher (1978) re-veal that the maximum temperature of hornblendestability increases from about 990o to l050oC as themole fraction of HrO in the gas decreases from I toabout 0.25, consistent with the findings of Holloway(19'73). At the mole fraction of 0.25 and a total pres-sure of l0 kbar, the partial pressure of HrO would beabout 2.5 kbar and there would be about 6 weightpercent HrO dissolved in the liquid. The Hawaiianbasalt investigated by Holloway and Burnham(1972), H.elz (1973, 1976) and Allen and Boettcher(1978) probably is neither appropriate nor optimalfor the generation of natural andesitic residual liq-uids because of its low concentrations of AlrO, andalkalis. Consequently, it is likely that high-aluminabasalt would yield andesitic residual liquids associ-ated with hornblende, pyroxene, and olivine at pres-sures less than 5 kbar and with P",o less than P,o.^, attemperatures greater than that found for Hawaiiantholeiite at 5 kbar : Pr,o: Ptor.r (namely, greaterthan about l0l5"C).

The above comparisons suggest that andesitic liq-uids and hornblende are stable to higher temper-atures and lesser concentrations of HrO in basalticbulk compositions than in andesitic bulk composi-tions. Probably the availability of olivine (or augite)

as a reactant is a principal cause ofthe greater stabil-ity of hornblende in basaltic systems. Other composi-tional differences [greater AlrO, (Bowen, 1928, p.lll), and alkalis (Cawthorn and O'Hara, 1976)l mayhelp stabilize hornblende, but they cannot be eval-uated quantitatively with published information. Insum, comparison of the natural andesitic liquid andassociated minerals with experimental products sug-gests that the natural liquid contained less than 6weight percent HtO and quenched from a temper-ature greater than l0l5'C.

Sekins et al. ll979l studied experimentally theAsama 1783 andesite. They found no hornblende onor near the liquidus, but did estimate that the resid-ual dacitic liquid (69-72 weight percent SiO,) of thisandesite contained 3.3 weight percent of HrO at960'C before extrusion.

Could the concentration and fugacity of H'O inthe included andesitic liquid exceed that of the sur-rounding residual dacitic liquid? Roedder's (1965)work reveals that olivines are strong enough at1200'C to contain excess pressures ofa few thousandatmospheres in an inclusion. However, a significantgradient of HrO fugacity from inclusion to rim ofcrystal would probably disappear by diffusion in afew days or weeks at temperatures near l000oc,judging from data on Fe, Mg, and O diffusivity inolivine (Buening and Buseck, 1973; Muehlenbachsand Kushiro , 197 5). The diffusion of HrO probably iscomparable to that of oxygen because of the similar-ity in size. The concentration of HrO (or OH-) in oli-vine probably is low in comparison to oxygen; con-sequently, the concentration gradient and diffusionof HrO in olivine may be small. Laboratory heatingexperiments on natural olivines with inclusions ofglass suggest that diffusive transfer of HrO is effectiveover tens of microns in a few hours at temperaturesabove l000oC (Anderson,1974). Probably both tem-perature and the fugacity of HrO in the trapped an-desitic liquid associated with hornblende are similarto those in the surrounding melt. Consequently, themost probable conditions for growth of the horn-blende from the trapped andesitic liquid are 1080"(Fe-Ti oxides) to 960oC and 100 atm (plausible pres-sure of froth in vent) to about 1000 atm of HrO pres-sure (Sekine et al., 1979).

In sum, the eruptive history, texture, and composi-tional relations of the hornblende-bearing inclusionsuggest that its glass represents an andesitic liquidquenched from a temperature of 1020+60"C and apartial pressure of HrO between about 100 and 1000

ANDERSON: HORNBLENDE IN CALC.ALKALINE ANDESITES

atm. The liquid was in equilibrium with hornblende,probably orthopyroxene, and olivine. The andesiticliquid probably contained a few percent HrO.

Hornblende associated with natural calc-alkalinebasaltic liquid

Observational data

Hornblendes associated with calc-alkaline basalticmaterial are known from La Soufridre, St. Vincent(Lewis, l973a,b), Fuego (Rose et al.,1978), and Med-icine Lake Volcano, California (Mertzman, 1978).Analyses of these hornblendes and related materialsare compared in Table 2. The compositions of thehornblendes are similar and have 13.6 to 16.5 weightpercent Al2O3, 2.3 to 2.6 weight percent NarO, and1.8 to 2.7 weight percent TiO, on an anhydrous basis.

Interpretation

It is uncertain whether any of these hornblendesare in a state of quenched equilibrium with basalticliquids, because the textural relationships are not de-finitive and because quenched basaltic matrix istermed andesite by some workers. A case-by-caseanalysis and interpretation is necessary.

Fuego volcano, Guatemala

The analyzed hornblende (Table 2, column l) is inthe anorthitic core (An92) of a plagioclase pheno-cryst (An92 to An80). The most closely associatedanalyzed glass (column 2, Table 2) is an inclusion inthe rim of the same crystal. The glass and the horn-blende are not in contact. The crystal is depicted inFigure 16 of Rose et al. (1978). The glass in theplagioclase is transitional in composition between ba-salt and andesite.

Rose et a/. suggested that the hornblende includedin the anorthitic core actually grew from a liquidwith a more basaltic composition than the inclusionof glass found within the rim of the plagioclase. Theirreasoning was as follows: many cores of anorthitefrom the same eruption contain inclusions of olivine.Subhedral crystals of anorthite (An92-90) are com-mon in olivine phenocrysts. The anorthitic cores ofthe plagioclase phenocrysts have round corners, com-prise a low proprotion of the total plagioclase, aresubequant and little zoned. Most of the Fuegoplagioclase is oscillatory zoned. The textures of theanorthite cores suggest that it is an early mineral toform in the Fuego magma. Some olivine phenocrystscontain inclusions of basaltic glass (51 percent SiOr)which are close to the weighted average composition

of the bulk rock extruded in the October eruption.The basaltic inclusions and bulk rock probably aresfunilar in composition to the parental liquid fromwhich the earliest crystals formed. The anorthiticcores and the included hornblende probably formodfrom a lquid with a composition near that of the ba-saltic glasses (Table 2, column 3).

The temperature of the parental basaltic liquidwas about 1030'C; it contained about 4 weight per-cent HrO (Rose et al., 1978). Subsequently, Harris(1979) has directly measured 1.6 to 3.2 weight per-cent of HrO in inclusions of glass in olivine pheno-crysts from Fuego. The lower concentration of 1.6weight percent HrO applies to parental liquid inFo77 olivine. AccordinEly, a revised estimate of1030150"C and 1.6 weight percent HrO is preferablefor the Fuego liquid which probably precipitated thehornblende.

La Soufriire, St. Vincent

Lewis's (1973a,b) descriptions show that vesicularand microcrystalline dark interstitial materialtouches hornblende as well as olivine, plagioclase,pyroxene, and magnetite in xenolirh 770. The inter-stitial material (termed scoria by Lewis) analyzed byhim is basaltic in composition (column 5, Table 2).Lewis (written communication, 1979) considers hisseparate to be a good one, little affected by productsof weathering and contaminating pieces of the crys-tals from the gabbro. His analyses show that inter-stitial scoria from different blocks differ in composi-tion, consistent with an indigenous origin of theinterstitial material. The most siliceous interstitialscoria comes from a block of gabbro lacking horn-blende but containing olivine, plagioclase, pyroxene,and magnetite. I would expect progressive solidi-fication to increase the number of minerals. s6lilizehornblende, and enrich the residual liquid in KrOand SiOr. Possibly the gabbro with the most siliceousinterstitial material originates from a lesser depth.Lewis's data suggest the crystallization of hornblendefrom basaltic liquid.

Liltle Mt. Hoffman, California

Olivine phenocrysts in a basaltic lava from a ventof Modoc Basalt (Powers, 1932) contain subround toirregular inclusions of hornblende (Mertzman, 197 8).Olivine phenocrysts in the scoria of the cinder con€(Little Mt. Hoffman) at the source of the flow haveround inclusions of devitrified glass, but no horn-blende (A. T. Anderson, unpublished data). Thecompositions of the hornblende and the groundmass

ANDERSON: HORNBLENDE IN CALC-ALEA,LINE ANDESITES

of the basalt are given in Table 2, columns 6 znd 7.Because the hornblende occurs included in olivine

phenocrysts in basaltic lava, Mertzman (1978) rea-sonably argued that the hornblende formed from ba-saltic liquid. The occurrence of hornblende includedin olivine together with andesitic glass (see Asama,above) suggests possible alternative interpretations:(l) the olivines with hornblende are xenocrysts de-rived from an andesitic environment, (2) the horn-blendes are daughter crystals which crystallized afterthe trapped melt attained an andesitic compositionby crystaltzing olivine. Mertzman's figures show thatthe hornblende is associated with microcrystallinematerial (probably formerly tiquid). Because of itsround shape and association with microcrystallinematerial, it is likely that the hornblende grew fromtrapped liquid after it had crystallized olivine.Growth of hornblende from supercooled liquid at apressure less than 1000 atmospheres is a third possi-bility. The composition of the liquid from which thehornblende in the Little Mt. Hoffman basalt fcrmedis uncertain. Mertzman (written communication,1979) notes that the 0.1 weight percent of K,O in theLittle Mt. Hoffman hornblendes suggests that the liq-uid from which they formed had less than about 0.5percent KrO. Possibly the groundmass of the LittleMt. Hoffman basalt is enriched in KrO by con-tamination (compare studies of East Sand Butte andcinder cone M39 of Anderson, 1976). Additionalwork on the Little Mt. Hoffman materials is rteededto evaluate the alternatives. I suggest that the liquidfrom which the hornblende grew was richeriin SiO,than is the bulk rock because of crystallization of oli-vine. Probably the liquid was similar in compositionto that of the calculated groundmass given as column7 of Table 2. The composition is that of a basaltic an-desite.

Mertzman (1978) extrapolated the results of Yoderand Tilley (1962) on the Warner high-alumina basaltto estimate that the hornblende initially formed atpressures greater than about 12 kbar. In my judge-ment Mertzman's apptcation and extrapolation ofYoder and Tilley's work is not justified, because thehornblende in question probably grew from a moresiliceous liquid with a greater Fel(Fe * Mg) ratiothan the Warner basalt.

Summary

Evidence for the crystallization of hornblendefrom natural calc-alkaline liquids at least as basic asbasaltic andesite (about 54 weight percent SiOr) iscompelling. The data of Rose et al. (1978) and Lewis

(1973a,b) on ejecta of Fuego and Soufriire volca-noes, respectively, suggest that hornblende growsstably from some high-alumina basalt liquids withabout 5l weight percent SiO, at temperatures near1030'C. The basaltic melts are comparatively rich inFe and poor in Mg and crystallize anorthitic plagio-clase (commonly greater than An85), olivine (Fo84to 66), and magnetite, as well as hornblende.

Experimental results (Helz, 1973, 1976) demon-strate the crystallization of hornblende at crustalpressures from basaltic andesite but not from basalticliquids. As argued above for andesitic liquids, thecrystallization of hornblende is fostered by greaterconcentrations of alumina and alkats and P,"o, >P",o probable for natural high-alumina basaltic liq-uids compared to the Hawaiian basalt studied byHelz. The Warner high-alumina basalt studied byYoder and Tilley (1962) has atypically greater MgO,MgO/(MgO + FeO), and low alkalis for basalticmembers of calc-alkaline magma series. Rose et a/.(1978) argued that the smaller MgO/MgO + FeO)ratio for Fuego high-alumina basalt would have theeffect of making its olivine liquidus temperatureabout 80oC less than that of the Warner basalt. Thedifference might permit liquidus hornblende for ba-salt with a concentration of about 5 percent HrO inthe liquid if PH,o : P,o,or. Experimental results are notconsistent with crystallization of hornblende fromsome natural high-alumina basaltic liquids with lessthan a few weight percent HrO.

. General discussion

The facts and interpretations outlined above sug-gest that hornblende crystallizes from andesitic andbasaltic liquids at crustal pressures. Most such horn-blende probably is a product of a reaction betweenliquid and olivine, as first suggested by Bowen (1928,p. 6l). The field and textural evidence of plutonicrocks is consistent with crustal formation of horn-blende by reaction between liquid and olivine(Miller, 1938; Best and Mercy, 1967; Lewis, 1973b,Ikeda, 1976;Walawender, 1976; Mullan and Bussell,1977). The l0o- to 106-year lifetimes of subduction-zone volcanoes (McBirney et al., 1974; Rose et al.,197 7 ; Mertzman, 197 7 ; McBirney, 1 978), the shorter-term variations in compositions of extruded products(Crandell and Mullineaux, 1973; Crandell et al.,1962; Newhall, 1979), and the I km' and smaller vol-umes of products of most individual eruptions, aswell as the larger volumes of rare caldera-formingeruptions (Smith, 1979), suggest that the roots of sub-duction zone volcanoes have thicknesses less than


about I km and solidify (and differentiate) on timescales appropriate for cold (crustal) environments.Geophysical models of Cenozoic island arcs suggestthey consist mostly of rock denser than andesite (forexample, see Grow, 1973). Gabbroic rocks, includinghornblende gabbro, have appropriate densities con-sistent with Kuno's (1968) model of a gabbroic crustin island arcs. Dominant hornblende gabbro in is-land arcs is particularly appealing because it canyield derivative magmas with the calcium-rich coresof plagioclase crystals, the isotopic ratios of stron-tium and lead, and the bulk compositions typical ofmany rocks in granodioritic batholiths (Piwinskii andWyllie, 1968; Faure and Powell, 1972; Doe, 1967;Wyllie, 1977).The basaltic and andesitic liquids fromwhich hornblende forms appear to contain 1.6i0.3(Harris, 1979) to 4t-2 wei5ht percent of HrO (Rose eral.,1978; Anderson, 1979). The temperature at whichhornblende forms in such liquids is probably be-tween 1080o and 960oC (see above).

Various factors probably contribute to the crystal-lization of hornblende from natural basaltic and an-desitic liquids with smaller concentrations of HrOthan the minimum of about 6 weight percent sug-gested by experinents (Yoder and Tilley, 1962; Hol-loway and Burnham, 1972; Green, 1972; Helz, 1973,1976; Stern et al.,1975; Allen and Boettcher, 1978):(l) olivine commonly is present as a reactant in natu-ral associations; (2) natural liquids have some CO, inaddition to HrO (Harris, 1979), consequently P^,o IP,"*,i (3) natural basaltic liquids associated with calc-alkaline magma series have greater concentrations ofalumina and alkalis than Hawaiian basalts used inexperiments; (4) natural calc-alkaline basaltic liquidshave smaller MgO/(MgO + FeO) and greater con-centrations of alkalis than the Warner basalt used inexperiments. All these differences probably enlargethe field of stability of hornblende beyond that foundin experiments. Reaction with olivine and other fac-tors probably foster the crystallization of hornblendein some natural basaltic and andesitic liquids withless than 6 weight percent of HrO, as suggested bythe natural occurrences described above.

Reaction between olivine and liquid yielding horn-blende acts to minimize the increase in Fel(Mg +Fe) ratio in derivative liquids. Two factors are signif-icant in minimizing the increase in Fel(Mg + Fe):(l) reaction between liquid and olivine; (2) crystalli-zation of a hornblende-bearing assemblage of miner-als in place of an equivalent anhydrous assemblage.

The effect of reaction was emphasized by Bowen(1928, p. 6l). With reaction magnesium-rich olivine

tends to buffer the Fel(Mg + Fe) ratio of crystalliz-ing minerals and residual liquid. Helz (1976) notedthat the Fel(MB * Fe) ratio of pyroxenes and resid-ual liquids decreases with crystallization in part ofthe range of solidification of hornblende, for ex-ample. Reaction with magnesium-rich minerals is aneffective mechanism which minimizes increases inthe Fel(Mg + Fe) ratio of residual liquids.

The influence on Fel(Mg + Fe) of the crystalliza-tion of a hornblende-bearing assemblage in place ofan equivalent anhydrous assemblage of minerals iscomplex and disputed (see Ringwood, 1974 and Al-len and Boettcher, 1978 for contrasting views). Whatis meant by "equivalent"? Consider the hornblende-out surface of Figure l. The mineralogy above thesurface and the composition of the liquid on the sur-face vary with P, T and X,,o. Many definitions of"equivalent" are possible. I adopt the following defi-nition because it seems consistent with most petrolo-gical discussions regarding the role of hornblende inthe formation of calc-alkaline rock series: on any sur-face marking identical bulk compositions of liquids,regions with different solids are equivalent.

For a given increment of crystallization how do theFel(Mg * Fe) ratios of residual liquids in two equiv-alent fields compare? For a given bulk composition(Hro-free), the liquidus surface is one (and possiblythe only) surface with identical compositions of liq-uids. Consider the difference between the Ol + Pland the Ol + Hb fields. I assume that the equilibriumcomposition of olivine on the liquidus is everywherethe same (Roeder and Emslie, 1970). In general theFel(Mg + Fe) of hornblende is greater than for asso-ciated olivine (Helz, 1973; Lewis, l973a,b). However,the effect on Fel(MB + Fe) of the residual liquid de-pends on the proportions of the crystallizing mineralsas well as on their compositions. The net iron enrich-ment caused by crystallization of Ol + Hb possibly isgreater than that for Ol + Pl. Probably the variationof Fel(Mg * Fe) of residual liquids with equilibriumcrystallization of equivalent hornblende-free andhornblende-bearing assemblages is complex, but notas large as the variations caused by fractional (as op-posed to equilibrium) crystallization. I agree, there-fore, with Cawthorn and O'Hara (1976) that theomission of iron from their experiments probably isnot crucial to an evaluation of the role of horn-blende. My emphasis is on reaction, however, ratherthan fractionation.

Minor increase in Fel(Mg + Fe) is a characteristicfeature of calc-alkaline magma series (Fenneg 1926;Kuno, 1950; Miyashiro, 1974). Consequently, the


spectrum of rock series found in subduction zone en-vironments (Peacock, l93l; Kuno, 1950, 1959) can berelated in part to variable reaction between olivineand liquid yielding hornblende: in general the morehornblende. the more calcic and the less iron-en-riched (less tholeiitic) the series. Many tholeiitic se-ries are calcic, however (for example Higashiyama,Isshiki, 1963) and have basaltic members with smallconcentrations of alkalis. The concentration of al-kalis in the parental liquid is important. Series lack-ing hornblende probably are tholeiitic or alkalic andtend to have groundmass pigeonite. Variable produc-tion of hornblende by reaction between liquid andolivine can help explain variable iron-enrichmentand lime-alkali index of manv subduction zonemagma series.

Implications

The probable crystallization of hornblende withinthe crust from subduction zone basaltic liquids withless than 6 weight percent HrO has implications for:(l) the genesis of andesite, (2) the temperature abovethe Benioff zone, (3) dehydration of subducted oce-anic crust, and (4) the origin of continental crust.

Genesis of andesite within the crust is implied bythe crustal formation of hornblende gabbro. As dis-cussed above, the Fel(Mg + Fe) ratio and calc-alkalic index for subduction zone magma series areexplicable according to crystallization of hornblendeby reaction between olivine and liquid. Although thestorage and release of HrO in and from hornblendein subducted oceanic crust may be important in ini-tiating subduction-zone magmatism, reactions be-tween hornblende and liquids in the Benioff zoneprobably have no direct influence on the character-istics of magma series in subduction zones (compareGreen and Ringwood, 1968; Green, 1972;'Boettcher,1973, with Stern et al., 1975; Cawthorn and O'Hara,1976; Allen and Boettcher, 1978).

Temperatures greater than about 1000"C are im-plied by basaltic liquids with less than about 6 weightpercent HrO. If subduction zone basaltic liquids orig-inate between the Benioff zone and the surface. thenthe temperature must be at least as high as l000oCsomewhere in the interval. Geophysical models ofsubduction zone processes generally yield lesser tem-peratures, particularly close to the relatively coldBenioff zone (compare Andrews and Sleep, 1974;Anderson et al., 1976, with Hasebe et al., l97O; Ito,1978). In order to be consistent with volcanologicaland geological data, geophysical models should yieldtemperatures in excess of l000oC in the magma path

between the Benioff zone and the volcanic front, ashas been emphasized by Hasebe et al. (197O).

Some of the HrO subducted as oceanic crust isprobably not returned to th€ surface by the out-gassing of parental basaltic liquids in subductionzones. There is an average ofabout I to 2 weight per-cent of mineralogically-bound HrO in the combinedoceanic crustal layers 2 (Melson et al., 1968; Hum-phris and Thompson, 1978) and 3 (Melson andThompson, l97l-Pnnz et al.,1976;lto, 1979; Ander-son et al., 1919). Parental basaltic liquids of sub-duction zones probably contain an average of about2 to 4 weight percent H'O (Eggler, 1972; Anderson,1973, 1979; but see also Marsh, 1976; Garcia et al.,1979; and Sekine et al., 1979, who estimate lowerconcentrations of HrO). The ratio of mass productionof crust in subduction zones (continental growth) tothat at oceanic ridges is about 0.2 (Anderson, 1974),based on the ratios of cross-sectional areas of Ceno-zoic oceanic island arcs (Murauchi et ul., 1968; Grow,1973) and the estimated cross-sectional areas of sub-ducted oceanic crust implied by late Cenozoic localrates of plate convergence (Chase, 1972). If all theHrO bound in subducted oceanic crust returns to thesurface by first entering into basaltic parental liquids,the expected concentration of HrO in the liquidsranges from 5 (l/0.2) to l0 (2/0.2) weight percent.The estimates of HrO actually present (2 to 4 per-cent) are uncertain, but are consistent with the notion(Anderson et al.,1979) that H'O is released from de-hydrating chlorite (the principal mineralogical resi-dence of HrO in oceanic crust) at too shallow a depth(Delany and Helgeson, 1978, but see Jsnkins, l!l!)to contribute to melting in subduction zones. TheHrO associated with the parental liquids of sub-duction zone magmas is similar in amount to theHrO supplied by amphiboles in the oceanic crust(Ito, 1979).

A complex origin of continental crust is implied byparental basaltic liquid and hornblende gabbro insubduction zones. A principal idea regarding the for-mation of continental crust involves the postulatedproduction and accretion of andesitic island arcs(Wilson, 1954, p.206), because average continentalcrust is andesitic in composition (Goldschmidt, 1933;Taylor, 1967). The chemical composition of the bulkand lower continental crust is uncertain and debated(see Taylor and Mclennan, 1979, and Tarney andWindley, 19'79 for contrasting views). If the bulk ofCenozoic island arcs is composed of hornblende gab-bro, then how can the andesitic composition of aver-age continental crust be explained? The implication

of dominant hornblende gabbro in island arcs is thatcontinental crust either was made diferently beforeCenozoic times or is further differentiated bv addi-tional processes.

Note added in proof

The following important article appeared after thispaper was accepted for publication: Ritchey, J. L.(1980) Divergent magmas at Crater Lake, Oregon:products of fractional crystallization and verticalzoning in a shallow, water-undersaturated chamber.J. Volcanol. Geothermal Res., 7,373-386.

AcknowledgmentsI thank P. J. Wyllie for helpful discussion and review of an early

version of this paper. C. Bacon (Menlo Park) provided a helpfulreview. I have benefited from rock samples and advice contributedby Shigeo Aramaki arid Naoki Onuma (Asama) and by W. I. Rose(Fuego). The work reported here was supported by NSF grantEAR76-15016 A0l .

References

Allen, J C. and A. L. Boettcher (1978) Amphiboles in andesiteand basalt: II. Stability as a function of P-T-1..,6-fo,. Am.Mineral., 63, lW4-1087.

--, and G. Marland (1975) Amphiboles in andesiteand basalt: I. Stability as a function of P-T-f o,. Am. Mineral.,60. 1069-1085.

Anderson, A. T. (1973) The before-eruption water content of somehigh-alumina magmas. Bull. Volcanol., 37, 53V552.

- (1974) Chlorine, sulfur, and water in magmas and oceans.Geol. Soc. Am. Bull., 85, 1485-1492.

- (1976) Magma mixing: petrological process and volcano-logical tool. J. Volcanol. Geotherm. Res., 1,3-33.

- (1979) Water in some hypersthenic magmas. J. Geot., g7,

509-53 l.- and D. Goufried (1971) Contrasting behavior of p, Ti and

Nb in a differentiated high-alumina olivine tholeiite and a calc-alkaline andesitic suite Geol. Soc. Am. Bull., 82,1929-1942.

--, D. M. Harris and E. Ito (1979) Toward geologic cycles forH2O, CO2, Cl and S. (abstr.). Int. Union Geodesy and GeophysicsAssembly, Canberra, Interdisciplinary Abstr., 26.

Anderson, R. N., S. Uyeda and A. Miyashiro (1976) Geophysicaland geochemical constraints at converging plate boundaries.Part I. Dehydration in the downgoing slab. Geophys. J. R. As-tron. Soc., 44, 333-357.

Andrews, D. J. and N. M. Sleep (1974) Numericat modelling oftectonic flow behind Island Arcs. Geophys. J. R. Astron. Soc., 38,237-25r.

Aramaki, S. (1956) The 1783 activity of Asama volcano, part l.Jap. J. Geol. Geogr.,27,189-229.

- (1957) The 1783 activity ofAsama volcano, part2 Jap. J.Geol. Geogr., 28, ll-33.

- (t963) Geology of Asama volcano. J. Fac. Sci. (Jniv To-kyo,14,229-443.

Best, M. G. and E. L. P. Mercy (1967) Composirion and crystalli-zation of mafic minerals in the Guadalupe igneous complex,California. Am. Minbral.. 52. 41647 4.

849

Boettcher, A. L. (1973) Volcanism and orogenic belts-the originof andesites. Tectonophysics, I 7, 223-240.

Bowen, N. L. (1928) The Evolution of the Igneous Roc&s. Dover,New York.

Buddington, A. F. and D. H. Lindsley (1964) Iron-titanium oxideminerals and synthetic equivalents. J. Petrol., 5,310-35'1.

Buening, D. K. and P. R. Buseck (1973) Fe-Mg lattice diffusion inolivine. "/. Geophys Res., 78, 6852-6862.

Cawthorn, R. G- and M. J. O'Hara (1976) Amphibole fractiona-tion in calc-alkaline magma genesis. Am. J. Sci., 276,309-329.

Chase, C. G. (1972) The N plate problem of plate tectonics.Geophys. J. Roy. Astron. Soc., 29, ll'l-122.

Crandell, D. W. and D. R. Mullineaux (1973) Pine Creek volcanicassemblage at Mount St. Helens, Washington. U.S. Geol. Sum.Bull., 1383A, l-23.

Recent age at Mount Rainier, Washhgton. U.S. Geol. Sum.Prof. Pap., 450-D, D64-D68.

Delany, J. M. and H. C. Helgeson (1978) Calculation of the ther-modynamic consequences of dehydration in subducting oceaniccrust to 100 kb and >800oC Am. J. Sci., 278,638-686.

Doe, B. R. (1967) The bearing of lead isotopes on the source ofgranitic magma. J. Petrol., & 5l-83.

Eggler, D. H. (1972\ Water-saturat€d and undersaturated meltingrelations in a Paricutin andesite and an estimate of water con-tent in the natural magma. Contrib. Mineral. Petrol., 34,261-27 t .

- and C. W. Burnham (1973) Crystauization and fractiona-tion trends in the system andesite-H2O-Co2-O2 at pressures tol0 Kb. Geol. Soc. Am. Bull., 84,2517-2532.

Faure, G. and J. L. Powell (1972) Strontium Isotope Geology.Springer-Verlag, Berlin.

Fenner, C. N. (1926) The Katmai magmatic province. J. Geol., 34,673-7',12.

Garcia, M. O and S. S. Jacobson (1979) Crystal clots, amphibolefractionation and the evolution of calc-alkaline magmas. Con-trib. Mineral. Petrol., 69, 319-327.

N W. K. Liu and D. W. Muenow (1979) Volatiles fromsubmarine volcanic rocks from the Mariana Island arc andtrough. Geochim. Cosmochim. Acta, 4 3, 305-3 12.

Goldschmidt, V. M. (1933) Grundlagen der quantitativen Geo-chemie. Fortschr. Mineral. Kristollogr. Pelrogr., 17, 112-156.

Green, T. H. (1972) Crystallization of calc-alkaline andesite undercontrolled high-pressure conditions. Contrib. Mineral. Petrol.,34. t5Ft66.

- and A. E. Ringwood (1968) Gehesis of the calc-alkaline ig-neous rock suite. Contilb. Mineral. Petrol., 18,105-162.

Grow, J. A. (1973) Crustal and upper mantle structure of the cen-tral Aleutian arc Geol. Soc. Am. Bull., 84,2169-2192.

Harris, D. M. (1979) Pre-eruption variations of H2O, S and Cl, ina subduction zone basalt (abstr.). EOS, 60,968.

Hasebe, R., N. Fujii and S. Uyeda (1970) Thermal processes un-der island arcs. Tectonophysics, 10,335-355.

Helz, R. T (1973) Phase relations of basalts in their melting rangeat PH2o:5 kb as a function ofoxygen fugacity. J. Petrol., 14,249-302.

- (1976) Phase relations of basalts in their melting ranges atPr,o : 5 kb. Part IL Melt compositions. J. Petrol., 17,139-193.

Holloway, J. R. (1973) The system pargasite-H2O-CO2: a modelfor melting of a hydrous mineral with a mixed volatile fluid. I.Experimental results to 8 kbar. Geochim. Cosmochim. Acta, 37,65 l'666.


850 ANDERSON: HORNBLENDE IN CALC-ALKALINE ANDESITES

- and C. W Burnham (1972) Melting relations of basalt with

equilibrium water pressure less than total pressure. J. Petrol.,

13. l -29.- and c E. Ford (1975) Fluid-absent melting ofthe fluoro-

hydroxy amphibole pargasite to 35 kilobars. Earth Planet. Sci.

Lett.. 25.4448Humphris, S. E. and G. Thompson (1978) Hydrothermal altera-

tion of oceanic basalts by seawater. Geochim. Cosmochim. Acta,

42, t07-126.Ikeda, Y. (1976) Petrology and mineralogy ofthe lritono basic in-

trusive rocks, central Abukuma plateau. J. Fac. Sci. Univ. To-

kyo, 19,229-251Isshiki, N. (1963) Petrology of Hachijo-jima volcano group, Seven

Izu Islands, Japan. J. Fac. Sci. Univ. Tokyo,15,91-134.

Ito, E. (1979) High-temperature alteration of oceanic gabbros

frorn the mid-Cayman rise (abstr.). Geol. Soc. Am. Abstracts

with Progrums, 11,449.Ito, K (1978) Ascending flow between the descending lithosphere

and the overlying asthenospherc. J. Geophys. Res.,83,262-268.

Jakds, P and A. J. R. White (1972) Hornblendes from calc-alka-.

line volcanic rocks of island arcs and continental margitts. Am.

Mineral.. 57.887-902.Jenkins, D. M. (1979) Experimental phase relations of chlorite-

and amphibole-bearing peridotites (abstr.). GeoI. Soc. Am. Ab-

stracts with Programs, 11,451.

Kuno, H. (1950) Petrology of Hakone volcano and the adjacent

areas, Japan. Geol. Soc. Am. Bull., 61,95'l-1020.- (1959) Origin ofCenozoic petrographic provinces ofJapan

and surrounding areas Bull. Volcanol., 20, j7-76.

- (1968) Origin of andesite and its bearing on the island arc

structure. Bull. Volcanol, 32, 14l-176.

Lewis, J F. (1973a) Mineralogy of ejected plutonic blocks of the

Soufridre volcano, St. Vincent: olivine, pyroxene, amphibole

and magnetite paragenesis. Contrib. Mineral. Petrol., 38, 197-

220.- (1973b) Petrology of ejected plutonic blocks of the Souf-

ridre volcano, St. Vincent, West Indies. J. Petrol., 14,8l-112.

Lindsley, D. H. and D. Rumble, III (1977) Therrnodynamic solu-

tion model for coexisting Ti-magnetite plus ilmenite (abstr.).

E O S , 5 8 , 5 1 9 .Marsh, B. D. (1976) Some Aleutian andesites: their nature and

source. ,/. Geol., 84,2746McBirney, A. R. (1978) Volcanic evolution of the Cascade range.

Am. Rev. Earth Planet. Sci.,6,437456.J. F. Sutter, H. R. Naslund, K. G. Sutton and C. M. White

(1974) Episodic volcanism in the central Oregon Cascade

Range. Geology, 2, 585-589.Melson, W. G. and G. Thompson (1971) Petrology of a transform

zone and adjacent ridge segments. R. Soc. Lond. Phil. Trans.,

268.423441.- and T. H. van Andel (1968) Volcanism and metamorphism

in the mid-Atlantic ridge, 22oN latitude. J. Geophys. Res., 73,

5925-5941.Mertzman, S. A., Jr. (1977) The petrology and geochemistry of the

Medicine Lake volcano, California. Contrib. Mineral. Petrol.,

62 .221 -247 .- (1978) A tschermakite-bearing high-alumina olivine tho-

leiite from the southern Cascades, California. Contrib. Mineral.

Petrol , 67,261-265.Miller, F S. (1938) Hornblendes and primary structures of the San

Marcos gabbro. Geol. Soc. Am. Bull.,49, l2l3-L232.

Minakami, T. (1942) On the distribution of volcanic ejecta (Part

2). The distribution of Mt. Asama pumice in 1783. Bull. Earthq.

Res. Inst.. 20.93-106.

Miyashiro, A. (1974) Volcanic rock series in island arcs and active

continental rnargins. Am. J. Sci., 274,321-355.

Moorhouse, W W. (1959) The Study of Rocks in Thin Section.

Harper and Row, New York.

Muehlenbachs, K. and I. Kushiro (1975) Measurements of oxygen

diffusion in silicates (abstr.). EoS, 56,459.

Mueller, R. F. (1969) Hydration, oxidation, and the origin of the

calc-alkali series. NlSl TN, D-5400, l-27.

Mullan, H. S. and M. A. Bussell (1977) The basic rock series in

batholithic associations. Geol. M ag., 1 1 4, 265-280.

Murauchi, S., N. Den, S Asano, H. Hotta, T. Yoshii, T' Asanuma,

K. Hagiwara, K. Ichikawa, T. Sato, W. J. Ludwig' J. I. Ewing'

N. T. Edgar and R. E. Houtz (1968) Crustal structure of the

Philippine Sea. J. Geophys. Res., 73,3143-fl'71.

Newhall, C. G. (1979) Temporal variation in the lavas of Mayon

volcano, Philippines. J. Volcanol. Geothermal. Res., 6, 6l-83.

Oshima, O. (1976) Fe-Ti oxide minerals of the 1973 eruption of

Asama volcano. Univ. Tokyo, Coll. of Gen' Educ., Sci. Pap.' 26'

39-50.Peacock, M. A. (1931) Classification of igneous rock series' "/.

Geol., 39,54-67.Piwinskii, A. (1973) Experimental studies of granitoids from the

central and southern Coast Ranges, California. Tschermaks

Mineral. Petrogr. Mitt., 20, 107-130.- and P. J. Wyllie (1968) Experimental studies of igneous

rock series: a zoned pluton in the Wallowa batholith, Oregon. .I.

Geol., 76,205-234.Powers, H. A. (1932) The lavas of the Modoc lava bed quadrangle,

California. Am. Mineral., 17, 253-293.

Prinz, M., K. Keil, J. A. Green, A. M. Reid, E' Bonatti and J.

Honnorez (1976) Ultramafic and mafic dredge samples from the

equatorial mid-Atlantic Ridge and fracture zone. J. Geophys.

Res., 81, 408'l-4103.Ringwood, A. E. (19'14) The petrological evolution of island arc

systems. Geol. Soc. Lond. J., 130, 183-2M.

Robertson, J. K. and P. J Wyllie (1971) Rock-water systems with

special reference to the water-deficient tegion. Am. J. Sci., 271,

252-277 -

Roedder, E. (1965) Liquid CO2 inclusions in olivine-bearing nod-

ules and phenocrysts frombasalts. Am. Mineral., 50,1746-1782.

Roeder, P. L. and R. F. Emslie (1970) Olivine-liquid equilibrium.

Contrib. Mineral. Petrol., 29, 2'l 5-289.

Rose, W I., Jr., A. T. Anderson, Jr., L. G. Woodruff and S. B'

Bonis (1978) The October 1974 basaltic tephra from Fuego vol-

cano: description and history of the magma body. "/. Volcanol.

Geotherm. Res., 4, 3-53.N. K. Grant, G. A. Hahn, I. M. Lange, J. L. Powell, J.

Easter and J. M. Degraf (1977) The evolution of Santa Maria

volcano, Guatemala. J. Geol., 85, 63-87.

Sakuyama, M. (1979) Lateral variations of H2O contents in Qua-ternary magmas of northeastern Japan. Earth Planet. Sci. Lett.,

43, t03-lrl.Sekine. T.. T. Katsura and S. Aramaki (1979) Water saturated

phase relations of some andesites with application to the estima-

tion of the initial temperature and water pressure at the time of

eruption. Geochim. Cosmochim. Acta, 43, 1367-1f76.

Smith, R L. (1979) Ash-flow magmatism. Geol. Soc. Am. Spec-

Pap., 180,5-27.Stern, C. R., W. L. Huang and P. J. Wyllie (1975) Basalt-ande-

site-rhyolite-H2O: crystallization intervals with excess H2O and

ANDERSON: HORNBLENDE IN CALC-ALKALINE ANDESITES 851

HrO-undersaturated liquidus surfaces to 35 kilobars, with im-plications for magma genesis. Earth Planet. Sci. Lett., 2g, lgg-196.

Stewart, D. C. (1975) Crystal clots in calc-alkaline andesites asbreakdown products of high-Al amphiboles. Contrib. Mineral.Petrol., 53,195-2A4.

Tarney, J. and B. F. Windley (1979) Continental growth, islatrdarc accretion and the nature of the lower crust-a reply to S. R.Taylor and S. M. Mcl-eonar.. Geol. Soc. Lond. J., 136,5Ol-504.

Taylor, R. W. (1964) Phase equilibria in the system FeO-FerO3-TiO2 at 1300"C. Am. Mineral.,49, 1016-1030.

Taylor, S. R. (1967) The origin and growth ofconrinents. Tectono-physics,4, 17-34.

- and S. M. Mclennan (1979) Discussion on ,.Chemistry,thermal gradients and evolution of the lower continental crust.In J. Tarney and B. F. Windley, Eds., Geol. Soc. Lond. J., 136,49't-500.

Turner, F. J. and J. Verhoogen (1960) Igneous and MelamorphicPetrology. McGraw-Hill, New York.

Walawender, M. J. (1976) Petrology and emplacement of the LosPinos pluton, southern California. Can. f. Earth Sci, 1J, 1288-1300.

Wilson, J. T. (1954) The development and structure of the crust.In G. P. Kuiper, Ed., The Earth as a Planet, p. 138-214. Univer-sity of Chicago Press, Chicago, Illinois.

Wyllie, P. J. (19'17) Crusral anatexis: an experimental review. Tec-tonophysics, 43, 4l-7 l.

Yoder, H. S. and C. E. Tilley (1962) Origin of basalt magmas: anexperimental study of natural and synthetic rock systems. ./.Petrol., 3, 342-532.

Manuscript received, October 27, 1979;acceptedfor publication, March 25, 1980.

Documents

Significance of hornblende in calc-alkaline andesites and ...mineral in some basaltic bulk compositions if P",o exceeds about l0 kbar (Yoder and Tilley, 1962; Green and Ringwood, 1968;