Differentiation of calc-alkaline magmas: Evidence from the Coquihalla Volcanic Complex, British Columbia

Journal of Volcanology and Geothermal Research, 9 (1981) 151--179 Elsevier Scientific Publishing Company, Amsterdam

--

151

Printed in Belgium

DIFFERENTIATION OF CALC-ALKALINE MAGMAS: EVIDENCE FROM THE COQUIHALLA VOLCANIC COMPLEX, BRITISH COLUMBIA

ROBERT G.

BERMAN

Department of Geological Sciences, The University of British Columbia, Vancouver, B.C., V6T 1 W5 (Canada) (Received December 27, 1979;revised and accepted June 3, 1980)

ABSTRACT Berman, R.G., 1981. Differentiation of calc-alkaline magmas: evidence from the Coquihalla Volcanic Complex, British Columbia. J. Volcanol. Geotherm. Res., 9: 151--179. The Coquihalla Volcanic Complex is part of the Miocene, calc-alkaline Pemberton Volcanic Belt, whose formation in southwestern British Columbia is associated with subduction of the Juan de Fuca plate. The complex consists of voluminous rhyolitic ash-flow tufts cross-cut by hypabyssal andasitic to dacitic intrusive bodies and a diorite stock. Whole-rock compositions of members of the Coquihalla Volcanic Complex show a range in silica from 54 to 76 wt.% (volatile-free). Uniform initial strontium isotopic ratios (0.7037 +- 0.00008) argue against derivation of magmas by crustal melting or mixing of basaltic magmas with crustal material. Trace element concentrations of whole-rocks cannot be modelled by varying degrees of partial melting of mantle material. Compositional variations among whole-rocks and constituent phenocrysts form the basis for major element mass-balance calculations which indicate that: (1) hornblende dacites can be derived from basaltic andesites by 50% crystallization of a mixture of plagioclase, hornblende, clinopyroxene, titanomagnetite, and apatite, and (2) rhyolites can be derived from dacites by roughly 45% crystallization of a mixture of plagioclase, hornblende, biotite, titanomagnetite, and apatite. Mineral/liquid distribution coefficients for eight trace elements and the observed mineral assemblages (plagioclase, clinopyroxene, hornblende, titanomagnetite, biotite, apatite, and zircon) have been compiled in an appendix. Large uncertainties in these values have led to a reformulation of the Rayleigh fractionation model so that an internally consistent set of distribution coefficients has been calculated for the proposed differentiation sequence.

INTRODUCTION

In recent years controversy has surrounded the problem of magma genesis in areas of lithospheric subduction. Less attention has been focussed on the petrogenesis of voluminous rhyolitic ash flows c o m m o n l y associated with calc-alkaline volcanism. The purpose of this paper is to present and interpret petrologic and geochemical evidence which bears on the petrogenesis of the calc-alkaline andesites and rhyolites of the Miocene Coquihalla Volcanic Complex of southwestern British Columbia (Fig. 1).

0377-0273/81/0000--0000/$02.50 © 1981 Elsevier Scientific Publishing Company

152

,~0OUIHALLA VOLCANIC COMPLEX

Fig. 1. Present plate t e c t o n i c setting o f s o u t h w e s t e r n British Columbia, showing e x t e n t o f P e m b e r t o n Volcanic Belt (squares), Garibaldi Volcanic Belt (triangles), A n a h i m Volcanic Belt (circles), Alert Bay Volcanic Belt (stars), and oldest K-Ar ages in million years. Stippled area shows the extent o f Miocene plateau lavas ( S o u t h e r , 1977). PP = Pacific plate, EP = E x p l o r e r plate, J d F P = J u a n de F u c a plate, NAP = N o r t h A m e r i c a n plate, Bp = Brooks Peninsula, P R f z = Paul Revere fracture zone, Sfz = Sovanco fracture zone. D o t t e d line is possible EP = J d F P boundary. Diagram modified f r o m Bevier et al. (1979).

The Coquihalla Volcanic Complex (CVC) forms part of the Miocene calcalkaline Pemberton Volcanic Belt (Souther, 1977; Berman and Armstrong 1980). Formation of this volcanic belt, as well as that of the Pleistocene-toRecent Garibaldi Volcanic Belt is ascribed to subduction of the Juan de Fuca plate beneath the North American plate (Souther, 1977; Green, 1977; Bevier et al., 1979). The stratigraphic and structural aspects of the CVC are described by Berman and Armstrong (1980), who also discuss regional geologic setting and tectonic implications. The complex consists of approximately 1600 m of variably welded crystal-lithic, lithic-crystal, and vitric tuffs intruded by hypabyssal andesite and dacite domes, dykes, and sills. A late-stage diorite stock in the centre of the complex forms Coquihalla Mountain. PETROGRAPHY

AND MINERALOGY

General petrographic features, and the range of phenocryst and microphenocryst compositions of rocks comprising the CVC are presented in Table 1. Representative modal analyses are presented in Table 2. Detailed petrographic descriptions are presented by Berman and Armstrong (1980).

plag cpx timt

plag hbl timt cpx

plag hbl timt cpx ap

plag biotite timt ap zr

Pyroxene andesite

Hornblende andesite

Hornblende dacite

Rhyolite

MP, G MP MP, G MP, G

P, MP P, MP MP, G Acc Acc

P, MP, G P, MP P, MP, G MP P, MP

P, P, P, P,

P, MP, G P, MP, G P, MP, G

P, MP P, MP P. MP MP, G MP, G G G

Occurrence

epidote + calcite chlorite

sericite chlorite + biotite chlorite --

Alteration

Uspxs.ls

An40.~0

An60.~s Wo2TEn49Fs24 Usp2 s Wo4,En4~Fsls

--

albitization

rare albitization rare oxide rims rare sphene + rutile chlorite

ANTI.30 epidote + calcite Wo2~Ens0Fs23 oxide rims, chlorite Usp4~.44 Wo43En#~Fs~4 chlorite

AnT~.4a Wo41En45Fs,4 Uspso_41

An6~.~ Wo4sEn40Fsls Wo0,En6~Fs~4 Usp12.11

Composition

porphyritic; vitroclastic matrices lithic fragments c o m m o n ; abundant ap and zr inclusions in biotite

porphyritic; intergranular to intersertal groundmass; timt and ap inclusions in plag

porphyritic, glomeroporphyritic; intergranular groundmass; timt inclusions in plag, hbl, and cpx

porphyritic, glomeroporphyritic; intergranular groundmass; timt inclusions in plag and cpx

hypidiomorphic granular; graphic textures in quartz and kspar

Textures and comments

plag = plagioclase, cpx = clinopyroxene, opx = orthopyroxene, hbl = hornblende, timt = titanomagnetite, ilm = ilmenite, kspax = orthoclase, ap = apatite, zr = zircon, P = phenocryst, MP = microphenocryst, G = groundmass, Acc = accessory,

plag cpx opx timt ilm kspar quartz

Diorite

Mineral

Petrographic characteristics

TABLE 1

c¢

154 TABLE

2

Representative m o d a l analyses o f No.

Rock type

632

diorite

62.9

251

61.8

192

p y r o x e n e andesite p y r o x e n e andesite h o r n b l e n d e andesite h o r n b l e n d e andesite h o r n b l e n d e dacite h o r n b l e n d e dacite

173

vitric

712

vitric

61 4 283

Volcanic

Complex igneous rocks

Phenocrysts plag

252

Coquihalla

cpx

opx

7.6

hbl

bio

mt

Q

kspar

18.0

--

--

5.8

3.0

29.3

--

--

--

8.8

--

8.4

ap

gndms

3.0

--

--

--

--

43.1

55.5

36.1

--

--

--

--

--

--

60.9

61.9

22.6

--

5.5

--

I0.0

--

--

~

60.0

64.2

22.3

--

4.3

--

13.1

--

--

0.9

56.9

52.9

8.0

--

35.8

--

11.3

--

--

0,4

60.7

65.3

--

--

23.5

--

10.1

--

--

1.2

67.2

tuff

83.6

--

--

--

15.1

1.4

--

--

P

93.0

tuff

76.1

--

--

2.2

4.3

4.4

13.0

--

--

94.2

P h e n o c r y s t m o d e s recalculated to

100%.

Modal and mineral compositions are very similar to those observed in other orogenic lavas of continental margins and island arcs (Jake~ and White, 1972a; Ewart, 1976). Amphiboles and pyroxenes exhibit a very limited range in composition among the various members of the volcanic suite. The most striking differences appear between pyroxene phenocrysts and pyroxenes in crystal clots. The latter contain m or e A1203 (as much as 4.6%), more TiO2 (as much as 1.1%), and a larger proportion of Ca-Tschermak's molecule (as much as 6.5%). These differences can be accounted for by crystallization of clot pyroxenes at higher pressures (Aoki and Kushiro, 1968), or at a stage prior to plagioclase precipitation (Barberi et al., 1971). Feldspars and titanomagnetites show the most significant compositional variations among the different members of the suite. Although phenocryst zonation produces considerable overlap (Fig. 2), there is a systematic decrease in anorthite content through the series diorite-andesite (ANT6 to An30), dacite (An60 to An3s), and rhyolite (An4o to An20). Similarly, ulvospinel contents of titanomagnetites from volcanic and hypabyssal rocks decrease continuously with increasing silica content of host rocks, although titanomagnetites of Coquihalla stock are displaced to lower ulvospinel contents relative to other CVC rocks (Fig. 3). These systematic decreases can be accounted for by steadily decreasing temperatures of crystallization throughout the suite. Assuming oxygen fugacities (ilmenite rarely coexists with titanomagnetite) along the Ni-NiO buffer for pyroxene andesites (Gill, 1978), and slightly above it for amphibole- and biotite-bearing dacites and rhyolites (Carmichael, 1967), temperature estimates vary from 925°C for andesites to 675°C for rhyolites. Lower ulvospinel contents of the plutonic diorites suggest crystallization at lower temperatures under conditions of higher PH, O, or subsolidus reequilibration. It is important to note that early crystallization of CVC oxides observed petrographically is inconsistent with the arguments of Eggler and Burnham (1973) that oxygen fugacities will not be sufficiently high to cause the appearance of oxides on the liquidi of calc-alkaline magmas.

155 An

Z___

"

Ab

Or

Fig. 2. Feldspar phenocryst and mierophenocryst compositions in terms of mole percent end-members. Diorite (squares), pyroxene andesite (apex-up triangles), hornblende andesite (apex-down triangles), dacite (solid circles), rhyolite (open circles).

Tip2

T i Fe204

/ geO

"-,.

",% g%04

\ Fe203

Fig. 3. Compositions of titanomagnetite and ilmenite phenoerysts and microphenocrysts. Symbols as in Fig. 2.

156 GEOCHEM~TRY Whole-rock analyses of members of the CVC were performed by X-ray fluorescence spectrometry, using fused discs for major elements (except sodium), and pressed powder pellets for trace elements and sodium (Armstrong and Nixon, 1980). Tables of whole-rock and phenocryst analyses are available upon request from the author. Members of the CVC display a range in silica contents (volatile-free) from 54 to 76 wt.%. Compositional variations within the suite (Figs. 4 and 5) are characterized by enrichment of K20, Na20, Rb, and Nb with increasing silica, and depletion of TiO:, A1203, MgO, FeO, MnO, CaO, P2Os, Cr, Ni, V, and Sr. The elements Ba, Ce, Nd, and Zr all show enrichment throughout most of the suite, and depletion in the most felsic members of the suite. Rocks of the CVC display the typical calc-alkaline trend on an AFM diagram. Major and trace element compositions of rocks within this suite are very similar to compositions within the Cascade volcanic system (McBirney, 1978), and to average compositions of orogenic lavas of continental margins (Jake~ and White, 1972b; Ewart, 1976). The main difference between the CoquihaUa volcanic rocks and those in the Pemberton-to-Reeent Garibaldi Volcanic Belt is the lower K20 contents of the latter (see also Berman and Armstrong, 1980). In order to estimate the effect of alteration processes on whole-rock compositions, four rocks, which showed pronounced development of secondary minerals (calcite, epidote, and chlorite), were analyzed and plotted with other "unaltered" rocks in Figs. 4 and 5. The main effect of alteration is an increase in H20 and CO2. The Harker diagrams were constructed on a volatilefree basis, and they indicate that altered rocks plot slightly outside the main trend of compositional variation for the elements Na, K, Ca, Mn, Ba, and Sr. The Na20-silica diagram is the only plot in which significant scatter appears among non-altered rocks. Depletion of alkalies in some CVC rhyolites could be due to leaching of glasses (Lipman, 1965), but these variations are not accompanied by changes in Rb concentrations, which is generally regarded to be mobile during hydrothermal alteration (Chikhaoui et ah, 1978). Moreover, the depletion of alkalies is accompanied by decreases in Ce, Nd, Ba, and Zr, and the latter element is considered to be highly immobile during metamorphism (Smith and Smith, 1976; Winchester and Floyd, 1977). In view of the small amount of scatter in compositional variation diagrams, and the small effects on highly altered rocks, the compositional variations exhibited by members of the CVC are regarded as reflecting the primary igneous history of these rocks. A fractionation scheme to account for the late-stage depletion of the above-mentioned elements is proposed below. PETROGENESIS

Models proposed for the genesis of silicic members of calc-alkaline suites (dacites and rhyolites) fall into three groups:

157 I

1.0

C

0.6

k-

0

0.2

0

18

~-:~ ~o ~

o

r6

¢,j

<:Z

0

CO

0

14

0 o

0

9 0

8 0

6

LL

4

2

o

o0

o

o

oo

o

%

o

~o

4

0

v~

3

x

~e v

2 I

v•

0.2

0 C:

o

0.1

o

8 © 121

6

L)

4

o

vv

•

•

v o~

v

o

2

o n° o

0

5

•

4

Z

3

n

o

o

o

o

o o

o

0

3

Z

2

o

o

o

~

~,o

o

r o y

o

o

o

o

~0.3

vv o

•

0.2

EL OJ

o

50

I

I

I

I

55

60

65

70

SiO

o

oO I 75

° 80

2

Fig. 4. Harker variation diagrams w i t h values in w e i g h t percent, volatile-free. S y m b o l s as in Fig. 2. Solid triangles = altered rocks.

(1) (2) 1971; (3)

Crustal anatexis (Ewart and Stipp, 1968; Ewart et al., 1971). Low-pressure fractional crystallization of andesitic liquids (Nicholls, Ewart et al., 1973; Lambert et al., 1974). Partial melting of quartz eclogitic upper mantle (Wilkinson, 1971).

158 I

I

20 16 12

o

I

I

,;,o •

%+, ~ o + .

8

•

4

O

I

OO O

60 40 20

a o

m

0 9

+D 200 >

o

o

o

v

SV%

°I

°

IZ0 40

o

12 8

o

o

I0 Z

ooo

o O

o

o

o

6

>.-

36 32 2B 24 20

¢+" .'++

a

7q

VA

500

o

0 o o o

240 2O0 160 120 80 600

.

~

o

-°-°

o

+

a

•

=

A

•

400 U3 3 0 0 20O IO0

o o

o o

o

80 -O

o

D =

a

60 40

.'=

° ~ = v

v v

•

o o

o

0

0

o

20 I.~0o 150o ( ~ HOO

•

700

•

0

v A

900

o .I

~" ,°"

o ,z

:o"

0

45

•1~ Z

35 25 15

°

"

o

A

o

Ioo oo ¢u

60 I 4o 20i 50,

o

,~+ "o=n

o

.

•

o

o

o

l~ •

o

I

I

55

60

I

I

I

65

70

75

80

SiOz

Fig. 5. Trace element Harker variation diagrams, with trace element concentrations in parts l~er million. Symbols as in Fig. 4.

A genetic relationship between members of the CVC is suggested by relatively smooth compositional variations and strontium isotopic uniformity within the suite (Berman and Armstrong, 1980). Systematic variations of modal abundances, and the correlation of whole-rock compositions with con-

159

stituent feldspar and titanomagnetite compositions, suggest that crystal fractionation played a dominant role in generating the petrological diversity. Although partial melting of an isotopically uniform source could also account for such diversity, trace-element data contradict this hypothesis. Trace element concentrations within a suite of igneous rocks can be used to discriminate between partial melting and fractional crystallization as discussed by Ferrara and Treuil (1974). The two processes can be distinguished on plots illustrating the correlation of two trace elements which have slightly different degrees of proclivity for the liquid phase. Rayleigh fractionation and equilibrium partial-melting equations (Shaw, 1970) were used to calculate the distribution of two trace elements with different, but constant bulk distribution coefficients (D), as a function of the a m o u n t of remaining liquid (F). It can be seen from Fig. 6 that the partial melting process causes a much greater departure from linearity in trace element correlation diagrams than the crystal fractionation process, even if neither element is hygromagmatophile (highly concentrated in the melt). It should be noted that calculated trace element distributions for other types of partial-melting processes produce similar patterns to those shown in Fig. 6 for the equilibrium case. 14

FC 30 A

d u

C~

PM

v

6 1 7 '°

!

40

I

60 Ct(Dt-- O. I)

I

80

I00

Fig. 6. T h e calculated c o r r e l a t i o n curves of two trace e l e m e n t s C 1 and C 2 w i t h d i f f e r e n t bulk distribution coefficients during partial melting (PM) and fractional crystallization ( F C ) processes. Numbers indicate the fraction o f m e l t f o r m e d or remaining.

160

Of the trace elements analyzed in CVC rocks, Zr, Nd, Ce, and Nb are commonly considered as hygromagmatophile (Ferrara and Treufl, 19"74; All~gre et al., 1977). The only elements, however, which increase in concentration throughout the CVC sequence andesite to rhyolite are Rb and Nb, and their enrichment rates indicate that bulk~iistribution coefficients are significantly greater than zero. Although the observed distribution of Rb and Nb (Fig. 7) is not quantitatively matched to that calculated in Fig. 6 because distribution coefficients are not constant over the range of compositions of CVC rocks (discussed below in the section "Trace element model"), the observed trend is very similar to that calculated for fractional crystallization, and supports the model that the volcanic suite has been generated by crystal ffactionation processes. In order for a partial-melting process to produce this nearly-linear pattern, distribution coefficients for Rb and Nb would have to be almost equal; such a near-equality is inconsistent with steadily increasing Rb/Nb ratios observed in CVC rocks. I

I

I

I

I

I

I

I O

12

o o I0•

0

0

•

i @

Z v vV

6 ¸

I 20

!

I 4£,

I

I 60

I

l 80

I I0(

Rb

Fig. 7. The observed correlation of rubidium and niobium in Coquihalla Volcanic Complex rocks. Symbols as in Fig. 4.

In addition to the above argument, it would be difficult to account for the depletion of Ce, Nd, Ba, and Zr in some rhyolites by partial melting of mantle or crustal material, in which bulk distribution coefficients for these elements are close to zero.

161 MAJOR ELEMENT MODEL

In order to test quantitatively the fractional crystallization hypothesis, solutions were sought to least-squares mass-balance equations which test whether the major element composition of a rock can be mathematically derived from that of another rock by subtraction of the compositions of phenocrysts observed in these rocks. A program written by Albar~le and Provost (1977} was used in this study and all mineral and whole-rock analyses were assigned standard deviations equal to 2% of the concentrations of each element present. Numerical modelling of the spectrum of CVC compositions was divided into two steps: basaltic andesite
162 TABLE 3 Major e l e m e n t m o d e l : step 1 - - basaltic andcsite to dacite

Mineral (%) Standard d e v i a t i o n

ap

hbl

plag

mt

cpx

0.82 0.21

26.35 5.77

42.20 3.09

7.80 0.62

22.84 5.16

Percentage o f dacite remaining = 5 1 . 1 Standard d e v i a t i o n = 3.9 Initial c o m p o s i t i o n s

SiO 2 TiO 2 AI~O 3 FeO MgO CaO Na20 K~O P2Os MnO

ap

hbl

plag

mt

cpx

dac

an

0.0 0.0 0.0 0.0 0.0 54.00 0.0 0.0 40.00 0.0

45.20 2.63 9.40 12.46 14.54 11.52 2.19 0.49 0.0 0.36

49.17 0.07 32.29 0.40 0.06 15.95 2.55 0.16 0.0 0.04

0.07 8.19 3.08 88.59 0.16 0.01 0.02 0.0 0.0 0.80

50.38 1.09 4.51 9.12 14.93 20.16 0.32 0.0 0.0 0.24

65.63 0.58 16.69 4.40 1.77 3.51 4.39 2.38 0.23 0.12

54.54 1.03 17.44 8.36 4.51 8.80 2.99 1.33 0.28 0.17

R e d u c e d residuals*

SiO 2 TiO 2 AI203 FeO MgO CaO Na20 K20 P~O~ MnO

ap

hbl

plag

mt

cpx

dac

an

0.000 0.002 --0.000 --0.000 ---0.000 0.015 0.001 --0.000 --0.015 0.000

0.024 0.204 --0.071 --0.067 -0.203 0.110 0.074 --0.022 -0.012 0.010

0.042 0.097 ---0.365 ---0.011 -0.022 0.238 0.132 -0.027 --0.019 0.013

0.000 0.153 --0.008 -0.132 -0.004 0.003 0.007 -0.004 -0.003 0.004

0.023 0.102 -0.033 -0.044 --0.181 0.161 0.027 -0.013 -0.010 0.008

0.140 0.356 --0.485 --0.108 -0.145 0.158 0.501 -0.199 ---0.058 0.034

--0.228 --0.896 0.989 0.366 0.565 ---0.672 ---0.727 0.268 0.118 ---0.070

* A b s o l u t e residuals (difference b e t w e e n initial and final c o m p o s i t i o n s ) divided b y error (2%) o f analyses. V a l u e s less t h a n 1.0 indicate residuals less than errors.

perimental work by Kushiro (1978) indicates that all fractionating phases are more dense than hydrous andesitic liquids. The observation that plagioclase occurs in slightly greater amounts than predicted by these calculations may be related to the lower relative density, and hence slower settling rates o f plagioclase. Experimental studies by Green (1972) and Eggler (1974) on natural and synthetic andesitic compositions demonstrate the feasibility of the proposed fractionation model. It should be noted that the age relationships of rhyolites (early) and intermediate igneous rocks (later) are incompatible with derivation of all igneous rocks in stages from a single, stagnant, fractionating magma chamber. CVC rhyolites and dacites appear to be derived by crystal fractionation of andesitic magmas, but the actual process may involve more than one magma chamber, or periodic replenishment of a single magma chamber by mafic melts (O'Hara, 1977).

163

TABLE 4 Major e l e m e n t m o d e l : step 2 - - d a c i t e t o r h y o l i t e

Mineral (%) Standard d e v i a t i o n

ap

hbl

plag

mt

bi

1.18 0.24

17.46 4.03

66.67 5.69

6.71 0.24

7.98 3.55

Percentage of rhyolite remaining = 56.70 Standard d e v i a t i o n = 5.46 Initial c o m p o s i t i o n s

SiO~ TiO 2 A1203 FeO MgO CaO Na20 KzO P2Os MnO

ap

hbl

plag

mt

bio

rhy

dac

0.0 0.0 0.0 0.0 0.0 54.00 0.0 0.0 40.00 0.0

43.62 2.93 11.11 11.76 14.80 11.05 2.50 0.35 0.0 0.31

58.23 0.10 26.79 0.12 0.03 7.56 7.23 0.21 0.0 0.0

0.31 5.35 2.10 89.90 0.67 0.14 0.08 0.08 0.0 0.90

38.62 4.07 14.55 13.75 16.15 0.10 0.75 7.01 0.0 0.55

77.02 0.14 13.23 0.84 0.14 0.68 3.79 3.80 0.04 0.05

65.63 0.58 16.69 4.40 1.77 3.51 4.39 2.38 0.23 0.12

R e d u c e d residuals I

SiO 2 TiO 2 AI~O 3 FeO MgO CaO Na20 K20 P~O 5 MnO

ap

hbl

plag

mt

b io

rhy

d ac

--0.000 0.003 ---0.000 --0.000 --0.001 0.045 0.000 0.001 ---0.045 --0.004

---0.026 0.146 --0.029 --0.020 --O.171 0.145 0.001 0.025 --0.016 --0.070

--0.133 0.156 --0.255 ~0.007 --0.043 0.394 0.010 0.087 ---0.062 --0.205

--0.000 0.091 --0.003 ---0.055 ---0.007 0.005 0.000 0.008 --0.006 --0.039

--0.011 0.086 --0.017 --0.011 --0.085 0.006 0.000 0.069 --0.007 --0.038

--0.354 0.326 --0.263 ---0.022 --0.095 0.156 0.012 0.695 ---0.130 ---0.434

0.534 --0.797 0.578 0.115 0.407 --0.737 ---0.023 --0.863 0.271 0.816

1 R e s i d u a l s and errors as in Table 3.

TRACE

ELEMENT

MODEL

Serious difficulties exist in the use of trace elements to test the consistency of the results of major element modelling. The problem of whether analysed rocks represent true liquids can be circumvented only with aphyric rocks. Errors related to analysis of a rock composed of cumulate phases and liquid are small for elements with bulk distribution coefficients (D) near zero, but quite considerable for elements with large D values (All~gre et al., 1977). The greatest problem in quantitative treatment of trace element distributions is the uncertainty in, and wide range of values for, mineral/liquid distribution coefficients (k). Many k values have been determined from analyses of phenocryst-matrix pairs in natural rocks. Distribution coefficients determined in this manner implicitly assume bulk equilibrium between mineral and liquid, and can give rise to spurious results when applied to Rayleigh

164 fractionation models. Analytical results o f Philpotts and Schnetzler (1970) clearly show phenocryst zonation with respect to trace elements, and k values calculated from zoned minerals in their study show differences as great as an order of magnitude. As pointed out by Albarkie and Bottinga (1972), apparent distribution coefficients calculated from zoned minerals will be higher than equilibrium values if h > 1, and lower than equilibrium values if k < 1. In recognition of this problem, Zielinski (1975) has modelled trace element abundances using a bulk equilibrium fractionation equation. Although this approach is internally consistent, results are probably inappropriate for application to natural systems where Rayleigh-type processes appear to be active. In addition to the problem of selecting appropriate k-values, recent experimental studies (reviewed by Irving, 1978) demonstrate that such values vary considerably with intensive parameters such as composition of mineral and melt, pressure, temperature, and oxygen fugacity. Increases in k-values with decreasing temperature and increasing silica content o f melts appear to reflect increasing polymerization and Si/O bond ratios in the melt structure (Irving, 1978). Discrepancies between apparent and experimental values can in part be resolved by more rigorous definition of distribution coefficients in terms of proposed equilibrium exchange reactions incorporating activity estimates of melt components (i.e., k-values for Ni in olivine, Leeman and Lindstrom, 1978). At the present time, however, m u c h more experimental work is necessary to extend this technique to other trace elements and the c o m m o n minerals in volcanic rocks. In addition, the effect of water content of melts on distribution coefficient values is in need of experimental investigation. In view of the above qualifications, it makes little sense to calculate trace element distributions using best estimates for k-values. Graphical techniques can be used to calculate bulk
(1)

where C~ is the concentration of i in the remaining liquid, COis the concentration of element i in the original liquid, F is the fraction of liquid remaining, a n d D i is the bulk solid/liquid distribution coefficient, defined as: D i = xikii

(2)

where x. is the fraction of mineral j in the cumulate, and h.- is the solid/ • . l . . . . . . . ~J hqmd d~stnbutmn coeffmmnt for mineral j and element ~. The bulk distribution coefficient for any trace element can be calculated from the concentration ratio of whole-rock pairs, combined with values of F

165

derived from the major element model (equation (1)). Values for mineral distribution coefficients can then be computed using the mineral proportions given by the results of the major-element model (equation (2)). Consistency of the trace element model can be judged by comparing these calculated kvalues with the ranges of values quoted in the literature (see Appendix 1). While providing a test for the proposed fractionation model, this approach has the added advantage of making explicit the assumptions concerning the most sensitive parameters in trace element modelling -- the values of the mineral/liquid distribution coefficients. In addition, calculated distribution coefficients can provide useful suggestions for further experimental investigations. Trace element modelling of this nature can be formulated such that solutions can be obtained by least-squares fitting akin to the major element procedure. This approach has recently met with some success in quantitative modelling of batch partial-melting processes (Minster and All~gre, 1978). In the present case, however, attempts to solve the problem in this manner led to non-converging solutions owing to the large errors associated with k-values for many minerals. TABLE 5 Assigned and calculated distribution coefficients S t e p 1: b a s a l t i c a n d e s i t e t o d a c i t e

Cr V Ni Sr

Ba Rb Nd Ce

D*

cpx

hbl

plag

mt

3.80 2.53 3.01 1.16 0.19 0.07 0.62 0.30

3.0 2.8 2.5 0.12 --0.2 0.1

5.0 4.0 2.9 0.46 0.42 0.19"* 0.5 0.3

---2.4** 0.19"* 0.05 0.08 0.09

23** 10.7"*

---

21.4"*

--

hbl

plag

mt

5 4 2.9 0.46 0.42 0.19 0.5 0.3

---5.6** 0.35 0.05 0.12 0.2

24.2** 46.5** 21.7"* ------

------

ap

1.0 --49.9* 19.6"

S t e p 2: d a c i t e t o r h y o l i t e D* Cr V Ni

Sr Ba Rb Nd Ce

3.05 4.55 2.25 3.86 1.13 0.26 1.93 1.98

bio 7 9.3** 3.7 0.4 10.2"* 2.4** 0.34 0.32

ap

m

1

147"* 153"*

* B u l k d i s t r i b u t i o n c o e f f i c i e n t ( c a l c u l a t i o n e x p l a i n e d in t e x t ) . * * C a l c u l a t e d v a l u e ( c a l c u l a t i o n e x p l a i n e d in t e x t ) . - - = d i s t r i b u t i o n c o e f f i c i e n t value c l o s e t o z e r o .

166 The results of the trace element modelling are presented in Table 5. The calculation was again divided into t w o steps using the trace element concentrations of the same rocks used in the major element models. Bulk distribution coefficients were calculated in the manner described above. For each element, one mineral/liquid distribution coefficient was calculated (indicated by an asterisk in Table 5) from the bulk distribution coefficient after other minerals, with better known or less sensitive k-values, were assigned values taken from the literature (Appendix 1). An effort was made to utilize experimentally determined values, but calculated k-values can vary dramatically depending on the chosen distribution coefficients. Decisions made in assigning k-values, and results of the calculations are discussed in detail in Appendix 2. DISCUSSION The general agreement o f calculated distribution coefficients (Table 5) with reported ranges of values found in the literature (see Appendix 1) lends support to the proposed fractional crystallization model. Many distribution coefficients appear to show large increases with increasing silica content of whole-rocks. It should be noted, therefore, that additional uncertainties are likely to have been produced by the assumption that distribution coefficients are constant within each modelling step adopted in this paper. Discrepancies in trace element k-values for the rare earth elements may reflect sampling problems, where apatite phenocrysts in dacites could be considered as cumulate phases. The problem is compounded in rhyolites where rocks with higher modal percentages of apatite show much greater concentrations of Ce and Nd. If the Rb content is used as an indicator for degree of differentiation, it is evident that Ce and Nd increase steadily in abundance until a sharp decrease is encountered in the most differentiated rhyolites (Fig. 8). Similar trends are recorded for Zr and Ba, and lead to the conclusion that the depletion of these elements marks a point of increased fractionation of biotite, with concomitant removal of the abundant zircon, apatite, and rutile inclusions observed in these rhyolitic biotites. Late decreases in Zr content of rhyolites in New Zealand have been interpreted in an analogous fashion b y Ewart et al. (1968). It is interesting to note that the two rhyolite samples which show the greatest range of Ce, Nd, Zr, and Ba concentrations are only separated stratigraphically b y approximately 75 m. The enriched sample (173) has roughly 15% biotite phenocrysts while the depleted sample (712) has less than 5%. The lower stratigraphic position of the depleted sample suggests that eruptions tapped a zoned magma chamber in which biotite had accumulated in lower levels, leaving the overlying residual liquids markedly depleted in biotite-associated trace elements.

167 I

I

f

I

I

I

I

220 @

180

l'q

g

140

~ 0 vvz~ ~O~

0

0

V z~

IO0

0 0

1500 0 CO

1300 I100 900

0 [3 v

700

v v z~

o

V

0 @'7 •

0

0

Do 0

80 (D

L)

60

z~

o

OooV

0

v

40

0

o

o

v

30

0

t,

"0

z

0

25 o

20

~,

15 I0

•

v '~v

0

0

@o

0 A

O

I

I

20

3O

I

1

I

I

I

I

40

50

60

70

80

90

I00

Rb

Fig. 8. The correlations o f rubidium with zirconium, barium, cerium, and neodynium in CVC rocks. Symbols as in Fig. 4. CONCLUSIONS

The association of the Coquihalla Volcanic Complex with subduction of the Juan de Fuca plate, and andesites in general with zones of lithospheric subduction, implies a genetic relationship such as outlined by Green (1977) or Mysen (1978b). The low Mg/(Mg+Fe) ratios, and low Cr and Ni contents of CVC andesites are consistent with their derivation through olivine and clinopyroxene or spinel fractionation from tholeiitic liquids produced by hydrous melting o f mantle peridotite above the subducted plate (cf. Nicholls and Ringwood, 1973). Generation of dacites and rhyolites by low-pressure fractionation o f andesitic liquids is supported by quantitative major- and trace element modelling, as well as the results of experimental studies. This low-pressure fractionation implies long-term residence of magmas in the crust, and may be a consequence o f the increased viscosity of hydrous andesitic liquids as they rise through the crust (Kushiro, 1978). The observed stratigraphic sequence o f the CVC suggests a model in which tapping o f the upper, differentiated portion of a high-level magma chamber occurs, producing voluminous early ash-flow eruptions. Production of latestage dacite, andesite, and diorite reflects decreasing residence time in the crust, possibly as a response to the increased development of conduit systems and fractures resulting from earlier eruptions.

168

Production of dacites and rhyolites by crystal fractionation requires the accumulation of a large amount of gabbroic material within the crust, but there are no geophysical data available to assess this possibility. ACKNOWLEDGEMENTS

Field and analytical expenses were met by NCERC operating grant A-8841 awarded to R.L. Armstrong. J. Knight and L.C. Pigage assisted with microprobe analyses, and the manuscript was greatly improved by the suggestions of D.J. Whitford. APPENDIX 1

Compilation of mineral/liquid distribution coefficients (K) of clinopyroxene hornblende, plagioclase, biotite, apatite, and zircon for the elements chromium, vanadium, nickel, strontium, barium, rubidium, neodymium, and cerium. Chromium K

Reference

Remarks*

2.6** 2.5--3.7**

Lindstrom and Weill (1978) Schreiber and Haskin ( 1976)

0.3--32

Ewart et al. (1973)

Di-Ab-An; 1290°C Fo-An-Di; 1350°C; 1 atm; k dec with dec fo basalti~ andesite to dacite; k dec with inc Fe/(Fe+Mg) of cpx estimated from range of experimental and natural data

Clinopyroxene

Leeman (1976)

10

Hornblende 23--36 12

Gill (1978) Leeman (1976)

andesite to dacite calculated from hb/cpx and cpx/ liquid ratios

Lindstrom (1976)

alkalic basalt; 1111--1168°C, 1 atm fo = 10 -4 to 10 -12 atm an~esite to dacite calculated from Rayleigh model of basalts andesite to dacite

Magne tire 100---620"* 27--58 9--26 1--58

Ewart et al. (1973) Leeman et al. (1978) Gill (1978)

Biotite 12.6 ± 4.8 17 7

Higuchi and Nagasawa (1968) Andriambololona et al. (1975) Leeman (1976)

dacite dacite calculated from mica/cpx and cpx/ liquid ratios

*Abbreviation " i n c " = increases, "dec" = decreases. **Indicates experimentally determined value.

169 Vanadium K

Reference

Remarks*

0.03--10"*

Lindstrom (1976)

0.8--2

Ewart et al. (1973)

basalt; 1125--1315°C; 1 atm; k dec with inc fo andesite to hacite

Gill (1978)

andesite to dacite

0--67**

Lindstrom (1976)

24--63 4.9--17

Ewart et al. (1973) Leeman et al. (1978)

basalt; II12--I135°C; 1 atm; k dec with inc fo andesite to hacite calculated from Rayleigh model of basalts

Clinopyroxene

Hornblende 18--45

Magnetite

Biotite 50

Andriambololona et al. (1975)

dacite

Reference

Remarks*

2.2--4.4**

Hakli and Wright (1967)

1.5--11.7"*

Lindstrom and Weill (1978)

2.55** 2.0

Mysen (1978a) Leeman (1976)

3.5--8

Gill (1978)

Makaopuhi basalt; 1050--1160 ° C; 1 atm; k dec with inc T ° C Ab-An-Di; 1150--1350°C; 1 a t m ; k dec with inc TOC Ab4sAn,sFo~Q3; 1025°C; 20 kbar estimate from experimental and natural data andesite

Nickel

K

Clinopyroxene

Hornblende 2.9** 3.7

Mysen (1978a) Leeman (1976)

7--8

Gill (1978)

An41Ab41Fot~Q2; 1000°C; 15 kbar calculated from hb/cpx and cpx/liquid ratios andesite to dacite

Magnetite 12.2--19.4"* 20--77**

Leeman (1974) Lindstrom (1976)

4--19

Gill (1978)

picritic tholeiite; 1300--1252°C; 1 atm alkalic basalt; 1111--1168°C; fo~ = 10 -4 to 10 -13 atm andesite to dacite

Biotite 13 3.7

Andriambololona et al. (1975) Leeman (1976)

andesite calculated from mica/cpx and cpx/ liquid ratios

170 Strontium K

Reference

Remarks*

0.06--0.08**

Shimizu (1974)

0.18---0.3"* 0.07--0.11 0.12--0.43 0.01--0.06 0.52 0.11 0.12

Sun et al. (1974) Hart and Brooks (1974) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Onuma et al. (1968) Sun and Hanson (1976)

DisoAb25An~s; 1100--1200 ° C; 15--30 k b a r ; f o = 10 -8 to 10 -4 atm basalt; 1~ 10---1140 ° C; 1 arm ankaramite, basaltic andesite basalt-zoned phenocrysts andesite rhyodacite

Clinopyroxene

megacryst in basalt

Hornblende Griffin and Murthy (1969) Ewart and Taylor (1969) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970)

basalt andesite to rhyolite basalt andesite

1.5--2.2"* 1.2--3.3"*

Sun et al. (1974) Drake and WeiU (1975)

1.4--2.8 1.3--1.8 2.8 2.4--4.5 1.5--7 1.4--1.8 1.9--2.6 2.6 4.6 2.0--3.9 2.3

Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Dudas et al. (1971) Korringa and Noble (1971) Ewart et al. (1973) Ewart et al. (1973) Ewart et al. (1973) Sun and Hanson (1976) Duchesne (1978) Duchesne and Demaiffe (1978)

basalt; 1110--1140°C; 1 atm tholeiite, andesite; 1150--1400°C; 1 atm; k inc with dec T ~C basalt andesite dacite dacite k inc from Ang0 to An~0 An64_ss in basaltic andesite Ans0.s5 in dacite Anss in rhyolite basalt k inc from Ans0 to A n ~ An4v in jotunite

Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970)

basalt dacite rhyolite

Duchesne (1978) Sun and Hanson (1976)

calculated from anorthosite basalt

Reference

Remarks*

Shimizu (1974)

Dis0Ab2sAn2s; 1100--1200°C; 15--20 kbar

0.31 0.2--0.5 0.55---0.64 0.19

Plagioclase

Biotite 0.08 0.12 0.67

Apatite 0.41 1.5

Barium K

Clinopyroxene <0.01"*

171 <0.01 <0.01 0.03--0.05 0.01--0.39 0.13 0.01--0.04 0.03

Hart and Brooks (1974) Onuma et al. (1968) Schnetzler and Philpotts Philpotts and Schnetzler Philpotts and Schnetzler Philpotts and Schnetzler Sun and Hanson (1976)

(1968) (1970) (1970) (1970)

ankaramite and basaltic andesite alkali-olivine basalt basalt basalt-zoned phenocryst rhyodacite andesite basalt megacryst

Hornblende

0.45 0.42--0.73 0.10 0.4

Griffin and Murthy (1969) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Leeman (1976)

0.32

Sun and Hanson (1976)

basalt basalt andesite calculated from hb/cpx and cpxtliquid ratios kaersutite megacryst in basalt

Plagioclase

0.2 --0.7**

Drake and Weill ( 1975 )

0.15--0.59 0.05--0.24 0.36 0.16--0.42 0.34--1.43 0.12--0.17 0.11--0.17 0.16 1.47 0.39

Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Korringa and Noble ( 1971 ) Dudas et al. (1971) Ewart et al. (1973} Ewart et al. (1973) Ewart et al. (1973) Sun and Hanson (1976) Duchesne and Demaiffe (1978)

natural tholeiite, andesite 1150-1400°C; 1 a t m ; k inc with dec T~C basalt andesite dacite k inc from Ang0 to An30 dacite basaltic andesite to andesite dacite rhyolite basalt An47 in jotunite

Higuchi and Nagasawa (1969) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970)

dacite phlogopite in basalt biotite in dacite biotite in rhyodacite

Reference

Remarks*

<0.01"*

Shimizu (1974)

<0.01 0.02--0.28 0.03 0.01--0.04

Hart and Brooks (1974) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970)

Dis0Ab25An~5 ; 1100--1200 ° C; 15--30 kbar ankaramite and basaltic andesite basalt rhyodacite basaltic andesite

Griffin and Murthy (1969) Nagasawa and Sehnetzler (1971)

basalt dacite

Biotite

9.7 -+ 1.3 1.09 6.36 15.3

Rubidium K Clinopyroxe ne

Hornblende

0.27 0.01

172 0.41---0.43 0.05 0.4

Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Leeman (1976)

basalt andesite calculated from amph/cpx and cpx/ liquid ratios

McKay and Weill (1976) McKay and Weill (1977) Duchesne (1978) Philpotts and Schnetzler (1970) Dudes et al. (1971) Sun and Hanson (1976) Duchesne and Demaiffe (1978)

synthetic lunar basalt; 1200°C; 1 atm synthetic low-K basalt; 1240 ° C; 1 atm k inc from Ans0 to An3, andesite dacite basalt k inc from Ans0 to An~0

Higuchi and Nagasawa (1969) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Philpotts and Schnetzler (1970) Leeman (1976)

dacite phonolite dacite rhyodacite calculated from mica/cpx and cpx/ liquid ratios

Reference

Remarks*

Grutzeck et al. (1974) Tanaka and Nishizawa (1975) Schnetzler and Philpotts (1968) Sehnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts ( 1970 ) Nagasawa and Schnetzler ( 1971 ) Sun and Hanson (1976)

Ab-An-Di; 1265°C; 1 atm basalt; 1200°C; 20 kbar basalt basalt andesite rhyodacite dacite basalt

Frey (in Irving, 1978) Schnetzler and Philpotts (1968) Sehnetzler and Philpotts (1970) Nagasawa and Schnetzler ( 1971 ) Sun and Hanson (1976)

tholeiite; 1000°C; 5 kbar basalt andesite dacite basalt

0.04--0.06**

Weill and McKay (1975)

0.08--0.13"*

Drake and Weill (1975)

0.04 0.11 0.02--0.07 0.02---0.2

Higuchi and Nagasawa (1969) Schnetzler and Philpotts (1968) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970)

synthetic lunar basalt; 1200--1340°C; k dec with dec T ~C tholeiite, andesite, and Ab-An-Di; 1150--1400°C; 1 atm, k inc with dec T~C basalt basalt basalt andesite

Plagioclase 0.02** 0.08** 0.12---0.25 0.03--0.19 0.06--0.49 0.05 0.12--0.25

Biotite 2.24 ± 0.47 3.06 3.26 0.94 3.0

Neodymium K

Clinopyroxene 0.21--0.24"* 0.35** 0.26--0.32 0.17---0.18 0.O7--O.65 1.28 0.94 0.38

Hornblende 0.5** 0.16 0.19 1--4.25 0.85--3.21

Plagioclase

173 0.17 0.14-0.29 0.17

Schnetzler and Philpotts (1970) Dudas et al. (1971) Sun and Hanson (1976)

dacite dacite basalt

Biotite 0.03 0.04 0.34

Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970)

basalt dacite rhyodacite

Apatite 27.4--81.1 21 1.4--16

Nagasawa (1970) Nagasawa and Schnetzler (1971) Sun and Hanson (1976)

dacite dacite calculated for basalt

Zircon 2--6.5

Nagasawa (1970)

dacite

K

Reference

Remarks*

Clinopyroxene 0.1--O.12"* 0.2** 0.3** 0.17 0.12-0.18 0.36 0.08--0.1 0.04--0.51 0.65 0.1

Grutzeck et al. (1974) Tanaka and Nishizawa ( 1975 ) Mysen (1978c) Onuma et al. (1968) Schnetzler and Philpotts (1968) Nagasawa and Schnetzler ( 1971 ) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Leeman (1976)

0.18

Sun and Hanson (1976)

same conditions as for Nd same conditions as for Nd basalt; 950°C; 20 kbar alkali-olivine basalt basalt dacite basalt andesite rhyodacite estimated from experimental and natural data basalt

Hornblende 0.3** 0.04** 0.12 0.34 0.09 0.49--1.98 0.43--1.77 0.2

F r e y (in Irving, 1978) Mysen (1978c) Schnetzler and Philpotts (1968) Higuchi and Nagasawa ( 1969 ) Schnetzler and Philpotts (1970) Sun and Hanson (1976) Nagasawa and Sehnetzler ( 1971 ) Leeman (1976)

tholeiite; 1000 ° C; 5 kbar An41Ab41FoI6Q2; 1000°C; 15 kbar basalt basalt andesite basalt dacite estimated from experimental and natural data

Plagioclase 0.07--0.14"* 0.05--0.07** 0.09 0.02--0.11 0.08--0.3 0.24

Drake and Weill (1975) Weiil and McKay (1975) Higuchi and Nagasawa (1969) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970)

same conditions as for Nd same conditions as for Nd basalt basalt andesite dacite

Cerium

174 0.16--0.4 0.22 0.1

Dudas et al. (1971) Sun and Hanson (1976) Leeman (1976)

dacite basalt estimated from experimental and natural data

Higuchi and Nagasawa (1969) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970) Schnetzler and Philpotts (1970)

dacite basalt dacite rhyodacite

Nagasawa (1970) Nagasawa and Schnetzler (1971) Sun and Hanson (1976)

dacite dacite calculated for basalt

Nagasawa (1970)

dacite

Biotite 0.32 0.03 0.04 0.23

Apatite 18--52.5 16.6 1.1--12

Zircon 2.3--7.4

APPENDIX 2

Method, assumptions, and results of calculations of mineral/liquid distribution coefficients Chromium Step 1: Experimental work demonstrates that k(cpx) decreases with decreasing fO~ as Cr changes from +3 to +2 valence state (Schreiber and Haskin, 1976). The Table 5 value is adjusted to a higher value to account for the higher silica content and lower temperatures of crystallization in CVC rocks as compared with the study cited. No experimental data are available on k(hbl); the Table 5 value is calculated from the average hbl/cpx concentration ratio for Cr in CVC phenocrysts, analyzed by microprobe. The calculated k(mt) value (23) falls well within the range (1--58) found from studies of natural phenocrystmatrix pairs (Appendix 1).

Step 2: k(hbl) is taken from the results of step 1. No experimental work has been carried out on k(bio), and the Table 5 value is taken from calculations by Leeman (1976). The calculated k(mt) value (24.2) is similar to that of step 1.

Vanadium Step 1: h(cpx) decreases with increasing f o (Lindstrom, 1976) and the Table 5 value is taken from this study, k(hb) is taken from t~ie average hbl/cpx concentration ratio for V in CVC phenocrysts. The calculated k(mt) value (10.7) agrees well with experimentally determined values at oxygen fugacities just above the QFM buffer (Lindstrom, 1976).

Step 2: Microprobe analyses of coexisting biotite and magnetite in CVC rocks indicate that k(mt) is roughly five times k(biotite). The calculated k(mt) value (46.5) is within the range of values (24--63) calculated in natural systems; the higher valuie than calculated in step 1 may reflect a real variation of k(mt) with increasing silica content and lower temperature. The k(bio) is much lower than that calculated by Andriambololona et al. (1975) for natural rocks.

175

Nickel Step 1: k(cpx) and k(hbl) are taken from the experimental results of Mysen (1978a) which indicate that variations with temperature and pressure are slight. The calculated k(mt) value (21) is similar to values determined experimentally (Leeman, 1974; Lindstrom, 1976) and in natural systems.

Step 2: No experimental data are available for k(bio) and the Table 5 value is taken from calculations of Leeman (1976). The calculated k(mt) value (22) is similar to results of step 1.

Strontium

Step 1: k(cpx) is taken from the average value of experimental work by Sun et al. (1974) and Shimizu (1974). k(hbl) is taken as an average value from studies on natural systems by Schnetzler and Philpotts (1970) and Ewart and Taylor (1969). k(ap) is adopted from calculations by Sun and Hanson (1976). The calculated k(plag) value (2.6) agrees with estimates calculated for An-rich plagioclase (Korringa and Noble, 1971), and with experimental results of Drake and Weill (1975) for temperatures between 1100 and 1200°C.

Step 2: k(bio) is taken from the average value for phenocrysts in dacites and rhyolites (Philpotts and Schnetzler, 1970). The calculated k(plag) value (5.4) is consistent with the results of Korringa and Noble (1971) for intermediate-composition plagioclase, and with Drake and Weill's (1975) experimental results when extrapolated to temperatures estimated for crystallization of CVC rhyolites.

Barium

Step I: k(cpx) is tal~en from the experimental work of Shimizu (1974). k(hbl) is taken as the average value of Griffin and Murthy (1969) and Leeman (1976). The calculated k(plag) value (0.19) is similar to experimental results of Drake and Weill (1975), and is consistent with results for natural systems.

Step 2: k(plag) is taken from the lower temperature values of Drake and Weill (1975) and from the calculated values of Korringa and Noble (1971 ) for intermediate composition plagioclase. The calculated k(bio) value (10.2) is similar to natural phenocryst values determined by Schnetzler and Philpotts (1970).

Rubidium Step 1: k(plag) is taken from the experimental work of McKay and Weill (1976; 1977), and k(cpx) from the results of Shimizu (1974) and Hart and Brooks (1974). The calculated k(hbl) value (0.19) is within the range of values determined from natural phenocrysts.

Step 2: Using k(hbl) of step 1, the calculated k(bio) value (2.4) agrees well with natural phenocryst results of Higuchi and Nagasawa (1969).

Neodymium

Step 1: k(plag) is taken from the experimental results of Drake and Weill (1975) and

176 Weill and McKay (1975). k(cpx) and k(hbl) are taken from the experimental work of Frey (in Irving, 1978). The calculated k(ap) value (50) is within the range of values determined by Nagasawa (1970) on natural phenocryst-matrix pairs.

Step 2:

k(plag) is taken from the lower temperature experimental results of Drake and Weill (1975), and Weill and McKay (1975). No experimental data are available on k(bio) and the Table 5 value is that of Schnetzler and Philpotts (1970). The calculated k(ap) value (147) is higher than any estimates in the literature, and may, in part, reflect increasing silica content and lower temperatures of the melt. A correction for the presence of zircon inclusions in rhyolitic biotites would lower the calculated value for k(ap).

Cerium Step 1: Sources of data are the same as used for neodymium calculations. The calculated k(ap) value (20) is similar to results on natural systems, but is inconsistently low when compared to the ratio of apatite distribution coefficients for Ce and Nd (Nagasawa, 1970; Nagasawa and Schnetzler, 1971).

Step 2: The calculated k(ap) value (153) is similar to that for Nd, but higher than values found for natural phenocrysts.

REFERENCES

Albarede, F. and Bottinga, Y., 1972. Kinetic disequilibrium in trace-element partitioning between phenocrysts and host lava. Geochim. Cosmochim. Acta, 36: 141--156. Albarede, F. and Provost, A., 1977. Petrological and geochemical mass-balance equations: an algorithm for least-squares fittingand general error analysis. Comput. Geosci., 3: 309--326.

All~gre, C.J.,Treuil, M., Minster, J.-F.,Minster, B. and Albarede, F., 1977. Systematic use of trace elements in igneous processes, 1. Fractional crystallization processes in volcanic suites. Contrib. Mineral. Petrol., 60: 57--75. Andriambololona, R., Lef~vre, C. and Dupuy, C., 1975. Coefficient de partage des dl~ments de transition dans les min~raux ferro-magn~sien extraits des dacites. C.R. Acad. Sci., Ser. D, 281. Aok'i, K.~and Kushiro, I., 1968. Some clinopyroxenes from ultramafic inclusions in Dreiser Weiher, Eifel. Contrib. Mineral. Petrol., 18: 326--337. Armstrong, R.L. and Nixon, G.T., 1980. Chemical and Sr-isotopic composition of igneous rocks from D S D P Legs-59 and 60. In: InitialReports of the Deep Sea Drilling Project. U.S. Government Printing Office, Washington, D.C. (in press). Barberi, F., Bizouard, H. and Varet, J., 1971. Nature of the clinopyroxene and iron-enrichment in alkalic and transitional basaltic magmas. Contrib. Mineral. Petrol., 33: 93--107. Berman, R.G. and Armstrong, R.L., 1980. The geology of the Coquihalla Volcanic Complex, southwestern British Columbia. Can. J. Earth Sci. (in press). Bevier, M.L., Armstrong, R.L. and Souther, J.G., 1979. Miocene peralkaline volcanism in west-central British Columbia--its temporal and plate-tectonic setting. Geology, 7 : 389--392. Carmichael, I.S.E., 1967. The iron-titanium oxides of salic volcanicrocks and their associated ferromagnesiansilicates. Contrib. Mineral.Petrol., 14: 36--64. Chikhaoui, M., Dupuy, C. and Dostal, J., 1978. Geochemistryof Late Proterozoic volcanic rocks from Tassendjanetarea. Contrib. Mineral.Petrol., 66: 157--164.

177

Drake, M.J. and Weill, D.F., 1975. Partition of St, Ba, Ca, Y, Eu 2+, Eu 3+, and other R E E between plagioclase feldspar and magmatic liquid: an experimental study. Geochim. Cosmochim. Acta, 39: 689--712. Duchesne, J.C., 1978. Quantitative modelling of Sr, Ca, Rb, and K in the Bjerkres-Sogndal layered lopolith (S.W. Norway). Contrib. Mineral. Petrol., 66: 175--184. Duchesne, J.C. and Demaiffe, D., 1978. Trace elements and anorthosite genesis. Earth Planet. Sci. Lett., 38: 249--272. Dudas, M.J., Schmitt, R.A, and Harward, M.E., 1971. Trace-element partitioning between volcanic plagioclase and dacitic pyroclastic matrix. Earth Planet. Sci. Lett., 11: 440-446. Eggler, D.H., 1974. Application of a portion of the system CaA12Si~Os-NaAISi~Cs-SiO:MgO-Fe-O2-H~O-CO: to genesis of the calc-alkaline suite. Am. J. Sci., 274: 297--315. Eggler, D.H. and Burnham, C.W., 1973. Crystallization and fractionation in the system andesite-H:O-CO~-O 2 at pressures to 10 kb. Geol. Soc. Am. Bull., 84: 2517--2532. Ewart, A., 1976. Mineralogy and chemistry of modern orogenic lavas -- some statistics and implications. Earth Planet. Sci. Lett., 31: 417--432. Ewart, A. and Stipp, J.J., 1968. Petrogenesis of the volcanic rocks of the Central North Island, New Zealand, as indicated by a study of 87Sr/8'Sr ratios, and Sr, Rb, K, U, Th abundances. Geochim. Cosmochim. Acta, 32: 699--736. Ewart, A. and Taylor, S.R., 1969. Trace element geochemistry of the rhyolitic volcanic rocks of Central North Island, New Zealand. Phenocryst data. Contrib. Mineral. Petrol., 22: 127--146. Ewart, A., Bryan, W. and Gill, J., 1973. Mineralogy and geochemistry of the younger volcanic islands of Tonga, S.W. Pacific. J. Petrol., 14: 429--465. E w a r t , A., Green, D.C., Carmichael, I.S.E. and Brown, F.H., 1971. Voluminous lowtemperature rhyolitic magmas in New Zealand. Contrib. Mineral. Petrol., 33: 128--144. Ferrara, G. and Treuil, M., 1974. Petrological implications of trace element and Sr isotope distributions in basalt-pantellerite series. Bull. Volcanol., 38: 548--574. Gill, J.B., 1978. Role of trace element partition coefficients in models of andesite genesis. Geochim. Cosmochim. Acta, 42: 709--724. Green, N.L., 1977. Multistage andesite genesis in the Garibaldi Lake area, southwestern British Columbia. Ph.D. Thesis, University of British Columbia, 246 pp. Green, T.H., 1972. Crystallization of calc-alkaline andesite under controlled high-pressure hydrous conditions. Contrib. Mineral. Petrol., 34: 150--166. Griffin, W.L. and Murthy, V.R., 1969. Distribution of K, Rb, Sr, and Ba in some minerals relevant to basalt genesis. Geochim. Cosmochim. Acta, 33: 1389--1414. Grutzeck, M.W., Kridelbaugh, S.J. and Weill, D.F., 1974. The distribution of Sr and the REE between diopside and silicate liquid. Geophys. Res. Lett., 1: 273--275. Hakli, T.A. and Wright, T.L., 1967. The fractionation of nickel between olivine and augite as a geothermometer. Geochim. Cosmochim. Acta, 31: 877--884. Hart, S.R. and Brooks, C., 1974. Clinopyroxene-matrix partitioning of K, Rb, Cs, St, and Ba. Geochim. Cosmochim. Acta, 38: 1799--1806. Higuchi, H. and Nagasawa, H., 1969. Partition of trace elements between rock-forming minerals and the host volcanic rocks. Earth Planet. Sci. Lett., 7: 281--287. Irving, A.J., 1978. A review of experimental studies of crystal/liquid trace element partitioning. Geochim. Cosmochim. Acta, 42: 743--770. Jakes, P. and White, A.J.R., 1972a. Hornblendes from calc-alkaline volcanic rocks of island arcs and continental margins. Am. Mineral., 57: 887--902. Jakes, P. and White, A.J.R., 1972b. Major- and trace-element abundances in volcanic rocks of orogenic areas. Geol. Soc. Am. Bull., 83: 29--40. Korringa, M.K. and Noble, D.C., 1971. Distribution of Sr and Ba between natural feldspar and igneous melt. Earth Planet. Sci. Lett., 11: 147--151. Kushiro, I., 1978. Density and viscosity of hydrous calc-alkaline magma at high pressures. Carnegie Inst. Washington Yearb., 77: 675--677. Lambert, R.S.J., Holland, J.G. and Owen, .P.F., 1974. Chemical petrology of a suite of calc-alkaline lavas from Mount Ararat, Turkey. J. Geol., 82: 419--438.

178 Leeman, W.P., 1974. Experimental determination of partitioning of divalent cations between olivine and basaltic liquid. Ph.D. Thesis, Part H. University o f Oregon, pp. 231-303. Leeman, W.P., 1976. Petrogenesis of McKinney (Snake River) olivinetholeiite in light of rare-earth element and Cr/Ni distributions. Geol. Soc. An~¢ Bull., 87: 1582--1586. Leeman, W.P. and Lindstrom, D.J., 1978. Partitioning of Ni 2÷ between basaltic and synthetic melts and olivines - - an experimental study. Geochim. Cosmochim. Acta, 42: 801--816. Leeman, W.P., Ma, M.-S., Murali, A.V. and Schmitt, R.A., 1978. Empirical estimation of magnetite/liquid distribution coefficients for some transition elements. A correction. Contrib. Mineral. Petrol., 66: 429. Lindstrom, D.J., 1976. Experimental study of the partitioning of the transition metals between clinopyroxene and coexisting silicate liquids. Ph.D. Thesis, University of Oregon, 188 pp. Lindstrom, D.J. and Weill, D.F., 1978. Partitioning of transition metals between diopside and coexisting silicate liquids. 1. Nickel, cobalt, and manganese. Geochim. Cosmochim. Acta, 42: 817--832. Lipman, P.W., 1965. Chemical composition of glassy and crystalline volcanic rocks. U.S. Geol. Surv. Bull., 1201-D: D1--D24. McBirney, A.R., 1978. Volcanic evolution of the Cascade Range. Annu. Rev. Earth Planet. Sci., 6: 437--456. McKay, G.A. and Weill, D.F., 1976. Petrogenesis of KREEP. Proc. 7th Lunar Sci. Conf., pp. 2427--2447. McKay, G.A. and Weill, D.F., 1977. KREEP revisited. Proc. 8th Lunar Sci. Conf., pp. 2339--2355. Minster, J.F. and All~gre, C.J., 1978. Systematic use of trace elements in igneous processes. Inverse problem of batch partial melting in volcanic suites. Contrib. Mineral. Petrol., 68: 37--52. Mysen, B.O., 1978a. Experimental determination of nickel partition coefficients between liquid, pargasite, and garnet peridotite minerals and concentration limits of behavior according to Henry's Law at high pressure and temperature. Am. J. Sci., 278: 217-243. Mysen, B.C., 1978b. The role of descending plates in the formation of andesitic melts beneath island arcs. Carnegie Inst. Washington Yearb., 77: 797--801. Mysen, B.O., 1978b. The role of descending plates in the formation of andesitic melts tween hydrous silicate melt, amphibole, and garnet peridotite minerals at upper mantle pressures and temperatures. Geochim. Cosmochim. Acta, 42: 1253--1263. Nagasawa, H., 1970. Rare earth concentrations in zircons and apatites and their host dacites and granites. Earth Planet. Sci. Lett., 9: 359--364. Nagasawa, H. and Schnetzler, C.C., 1971. Partitioning of rare earth, alkali and alkaline earth elements between phenocrysts and acidic igneous magma. Geochim. Cosmochim. Acta, 35: 953--968. Nicholls, I.A., 1971. Petrology of Santorini volcano, Cyclades, Greece. J. Petrol., 12: 67--119. Nicholls, I.A. and Ringwood, A.E., 1973. Effect of water on olivine stability in tholeiites and the production of silica-saturated magmas in the island-arc environment. J. Geol., 81: 285--300. O'Hara, M.J., 1977. Geochemical evolution during fractional crystallization of a periodically refilled magma chamber. Nature, 266: 503--507. Onuma, N., Higuchi, H., Wakita, H. and Nagasawa, H., 1968. Trace element partition between two pyroxenes and the host volcanic rocks. Earth Planet. Sci. Lett., 5 : 4 7 - - 5 1 . Philpotts, J.A. and Schnetzler, C.C., 1970. Phenocryst-matrix partition coefficients for K, Rb, Sr and Ba, with applications to anorthosite and basalt genesis. Geochim. Cosmochim. Acta, 34: 307--322.

179 Schnetzler, C.C. and Philpotts, J.A., 1968. Partition coefficients of rare-earth elements and barium between igneous matrix material and rock-forming mineral phenocrysts, 1. In: L.H. Ahrens (Editor), Origin and Distribution of the Elements. Pergamon Press, pp. 929--938. Schnetzler, C.C. and Philpotts, J.A., 1970. Partition coefficients of rare-earth elements between igneous matrix material and rock-forming mineral phenocrysts, 2. Geochim. Cosmochim. Acta, 34: 331--340. Schreiber, H.D. and Haskin, L.A., 1976. Chromium in basalts: experimental determination of redox states and partitioning among synthetic silicate phases. Proc. 7th Lunar Sci. Conf., pp. 1 2 2 i - - 1 2 5 9 . Shaw, D.M., 1970. Trace-element fractionation during anatexis. Geochim. Cosmochim. Acta, 34: 237--243. Shimizu, N., 1974. An experimental study of the partitioning o f K, Rb, Cs, St, and Ba between clinoDyroxene and liquid at high pressures. Geochim. Cosmochim. Acta, 38 : 1789--1798. Smith, R.E. and Smith, S.E., 1976. Use of Ti, Zr, Y, Sr, K, P, Nb in classification of basaltic magmas~ Earth Planet. Sci. Lett., 32: 114--120. Souther, J.G., 1977. Volcanism and tectonic environments in the Canadian cordillera -- a second look. In: W.R.A. Baragar et al. (Editors), Volcanic Regimes in Canada. Geol. Assoc. Can. Spec. Pap., 16: 3--24. Sun, C.C., Williams, R.J. and Sun, S.S., 1974. Distribution coefficients of Eu and Sr for .plagioclase-liquid and clinopyroxene-liquid equilibria in oceanic ridge basalt: an experimental study. Geochim. Cosmochim. Acta, 38: 1415--1433. Sun, S.S. and Hanson, G.N., 1976. Rare-earth element evidence for the differentiation of the McMurdo volcanics, Ross Island, Antarctica. Contrib. Mineral. Petrol., 54: 139-155. Tanaka, T. and Nishizawa, O., 1975. Partitioning of REE, Ba, and Sr between crystal and ]iquid phases for a natural silicate system at 20 kb pressure. Geochim. J., 9: 161--166. Weill, D.F. and McKay, G.A., 1975. The partitioning of Mg, Fe, Sr, Ce, Sin, Eu, and Yb in lunar igneous systems and a possible origin of KREEP by equilibrium partial melting. Proc. 6th Lunar Sci. Conf., pp. 1143--1158. Wilkinson, J.F.G., 1971. The petrology of some vitrophyric calc-alkaline volcanics from the Carboniferous of New South Wales. J. Petrol., 12: 587--619. Winchester, J.A. and Floyd, P.A., 1977. Geochemical discrimination of different magma series and their differentiation products using immobile elements. Chem. Geol., 20: 325--343. Zielinski, R.A., 1975. Trace-element evaluation of a suite of rocks from Reunion Island, Indian Ocean. Geochim. Cosmochim. Acta, 39: 713--734.