Ancient and Modern Volcanics and Proposed Miscibility in Silicate Systems Exploration Géologique

Home , Variolite

DPV 427 ANCIENT AND MODERN VOLCANICS AND PROPOSED MISCIBILITY IN SILICATE SYSTEMS EXPLORATION GÉOLOGIQUE

MINISTERE DES RICHESSES s\ NATURELLES

DIRECTION GÉNÉRALE DES MINES

J l ç i

ANCIENT AND MODERN VOLCANICS AND PROPOSED MISCIBILITY IN SILICATE SYSTEMS

L . GELINAS C . BROOKS

D PV-427 ANCIENT AND MODERN VOLCANICS AND PROPOSED MISCIBILITY IN SILICATE SYSTEMS

L. GELINAS Département de Génie Minéral, Ecole Polytechnique

C. BROOKS Département de Géologie, Université de Montréal

Document déposé au fichier ouvert, le 4 octobre 1976 Distribution sur demande seulement Document placed on Open File, October 4, 1976 Distribution on request only ABSTRACT

Evaluation of ancient and modern volcanic suites in terms of an experimentally based proposed miscibility gap

(PRIG) reveals that of the main classes of volcanism it is only the tholeiites (ancient and modern) and ancient calc-alkaline volcanics which overlap or closely flank the PMG. This may explain the recently numerous observations of immiscibility-produced-textures in tholeiitic rocks, and leads to a prediction that for modern tholeiites such observations will be most common in the more diversified low-K tholeiites of island-arcs. The restriction of the best and most common evidences of immiscibility to Archaean volcanic terrains

(variolitic lavas) is related via partial melting and mantle- water considerations to a time-dependant, decreasing susceptibility of tholeiitic and low-K calc-alkaline magmas to undergo liquid immiscible-splitting. A time-independant tendency however is the apparent need for some degree of differentiation of a tholeiitic parent magma prior to immiscible splitting. — II —

CONTENTS

Page ABSTRACT I

INTRODUCTION 1

PROPOSED MISCIBILITY GAPS 3

MODERN VOLCANIC SERIES AND THE PMG 6

Tholeiite suites

Calc-alkaline suites

Alkaline suites

ANCIENT VOLCANICS AND THE PMG 13

DISCUSSION 15

ACKNOWLEDGEMENTS 24

REFERENCES 25

FIGURES

Page 1 - Pseudo-ternary diagram 5 2 - Modern tholeiite suites and the PMG 7 3 - Modern cale-alkaline and alkaline suites and the PMG 10 4 - Ancient tholeiitic and low-K talc.--alkaline volcanics and the PMG 12 5 - South African Komatiites and the PMG 14 6 - Ar.hean variolites of Rouyn-Noranda region and the PMG 17 7 - Composite fields for modern clay volcanics in relationship to the PMG 18 8 - Composite modern and ancient tholeiite and low-K calc-alkaline fields and the PMG 19 INTRODUCTION

Because of the renewed interest in immiscibility

(e.g. Roedder and Weiblen, 1970; McBirney, 1975; Gélinas, 1974;

Gélinas et al., 1975) many phenomena which texturally resemble

globules produced by liquid segregation have been re-examined,

and some found demonstrably immiscible with their enclosing fractions (e.g. Philpotts and Hodgson, 1968; Ferguson and

Currie, 1971). Many of these examples come from mafic alkaline

intrusives; however, surprisingly, some have been unearthed in

tholeiitic associations. For instance Roedder and Weiblen

(1971) report the occurrence of two immiscible silicate liquids

(one enclosed as droplets in another and quenched to form two

immiscible-glasses) in volcanics from Hawaii, California,

Greenland and Canada. Similar observations have been made in

the high-alumina olivine-tholeiite of Hat Creek, California

(Anderson and Gottfried (1971) and in the tholeiitic Deccan

Traps (De, 1974). De (1974) also suggested that the trend of

differentiation in the Upper Zone of the Layered Series of the

Skaergaard Intrusion was in part controlled by liquid immiscibility, and this was subsequently experimentally verified by

McBirney (1975). Further evidence linking immiscibility and tholeiitic magmatism has been advanced by Gélinas et al.,

(1975) who explain the origin of variolitic lavas in tholeiitic suites of Canadian Archaean volcanic piles by liquid immiscibility. 2

It would seem therefore that some form of association exists between liquid immiscibility and tholeiitic magmas. In many cases this association links immiscible splitting to only the late stages in magmatic history, however the tholeiitic Archaean variolites of Canada, the komatiitic Archaean variolites of South Africa (Ferguson and Currie, 1972) and the Skaergaard intrusion of Greenland all involve immiscibility on a much grander scale. Are these examples merely artifacts of some peculiar geologic conditions, or is immiscibility a ubiquitous phenomenon that continually shadows tholeiitic magmatism, awaiting only the right physio-chemical conditions before manifesting itself? The purpose of this study is to examine this question by an appraisal of certain ancient and modern volcanic suites representative of the main classes of volcanism (tholeiitic, calc-alkaline, alkaline) with the direct aim of a) evaluating the possible link between tholeiitic magmatism and immiscibility, and b) comparing the main classes of volcanism in the light of

"susceptibility to immiscibility". We will demonstrate that not only is there a good case for tholeiitic magmas to be more commonly associated with immiscibility than either the calc- alkaline or alkaline volcanics, but also that there is a stronger likelihood of encountering evidences of immiscibility in the more ancient volcanics, as a result of a polarized, time- dependant "proneness" of certain magmas to undergo liquid segregation. 3

PROPOSED MISCIBILITY GAPS

Grieg (1927) showed that of the major rock forming oxides, Na2O, K2O and Al2O3 are miscible with S102 whereas

CaO, FeO, MgO and Fe2O3 are immiscible with SiO2 at high SiO2 concentrations. Grieg presented a pseudo-ternary diagram of

(CaO + MgO + EFeO(+TiO2)) versus (Na2O + K2O + Al 2O3) versus

(SiO2) to show the limited extent of the primitive immiscibility field in the polycomponent system. In this system, Grieg's primitive field is restricted to a narrow zone high on the silica portion of the cafemic oxide-SiO2 edge (figure 1).

Although experimentally verified, Grieg considered immiscibility unlikely as a process in the origin of igneous rocks, on the bases that: (1) the experimentally established minimum temperatures were too high (nearly 1700° C); (2) no igneous rocks approached the compositions found to show immiscibility in the laboratory; and (3) the textural evidence for immiscibility in natural rocks could be readily assigned to other processes. It was not until Roedder (1951) described silicate immiscibility in the system leucite-fayalite-silica, and in the parent quaternary system K2O - FeO - Al2O3 - SiO2 (1953), at much lower temperatures (minimum < 1100° C), and in compositions much closer to igneous rocks, that the geological possibility of immiscibility became reasonable.

In the development of knowledge concerning this second field of immiscibility (denoted PTIG in figure 1), both

Ilolgate (1954) and Roedder (1956) noted that two systems relevant to silicate melts display flat liquidus surfaces which according to Grieg (1927) are indicative of possible metastable fields of

liquid immiscibility. These systems are NaA1SiO4 FeO - SiO2 (Bowen and Schairer, 1938), where the fayalite field displays

a flat: liquidus surface, and the CaNgSi2O6 - SiO2 (Schairer and Bowen, 1938), where the plateau-shaped diopside field shows a peculiar inverse curvature. With the aid of Greig's (1927) pseudo-ternary diagram, Ilolgate (1954) presented Roedder's (1951) stability field of immiscibility within the system leucite-fayalite-silica, together with the locations of the axes of the metastable miscibility gaps for the systems NaA1SiO4

FeO - SiO2 and KA1Si2O6 - CaMgSi2O6 - SiO2. The coincidence of these axes with Roedder's field of immiscibility was striking.

The final validity of this field was obtained in the laboratory. Roedder and Weiblen (1970), Ferguson and

Currie (1972), Massion and Van Groos (1973) and McBirney (1975) have obtained immiscible, liquid-pairs during experiment which plot within the PMG shown in figure 1. Hence the PMG can be confidantly considered as a verified, geologically-applicable, two-liquid field. Presumably, magmas with compositions plotting in or on the flank of the PMG would be prone to split into two-liquids and we use this property throughout our evaluation of the volcanic rock series. Because we use the PMG in our appraisal of volcanic rock suites, it is pertinent to here consider some of Si0 2

PRIMITIVE T.WO LIQUID FIELD

EN f" M, DI /" PMG B+' 50, / 50 FO ~

2 Ca0+Mg0+EFe0+Ti0 Na20 + K20 + A1203

WEIGHT

FIGURE 1 - Pseudo-ternary diagram showing location of the primitive and proposed miscibility gaps (latter denoted as PMG). Tie-lines within the PMG join experimentally established liquid immiscible fractions (data from Roedder and Weiblen (R) 1970; Massion and Van Gross (M) 1973; McBirney (B) 1975). The open fields are for lunar glass inclusions (N 50) of Roedder and Weiblen). 6 the more important properties of this field as revealed on the pseudo-ternary diagram of figure 1. The PMG is an extended ellipsoid that roughly parallels the silica-cafemic oxides boundary. This means that a magma composition plotting on the alkaluminous side of the PMG would be driven into the PMG by any processes leading to a relative decrease in the alkali and alumina contents of that magma. If this process also leads to a lower alumina to alkalis ratio, then immiscible splitting is even further favored, although as pointed out by Watson and Dickey (1975) the ratio, more appriopriately, should be alumina to alkalis plus water. These features have importance in the comparison of volcanic suites to follow.

MODERN VOLCANIC SERIES AND THE PMG

Phenocryst-free, rapidly-cooled volcanics can be used to tell much about the composition of their parent magma. In gathering data from the literature we have undoubt- edly included numerous Phenocryst-bearing volcanic rocks, however this is of secondary importance since we are primarily interested in liquid compositions and these may be either original melt compositions or compositions of late, residual- liquids. Thol.eiite suites. These suites from different tectonic environments are plotted in figure 2. The environments include oceanic island (fig. 2-A), continental (fig. 2-B), sea-floor (fig. 2-C), and island arc (fig. 2-P,G,H,I). A composite tholeiite field for the low-K tholeiites of island Si02

,' o •': , ,~~ • 50 50

~ J ~ CaO+MgO Na10+K20 +E Fe O+Ti O 2 +AI z0

WEIGHT %

FIGURE 2 -Modern tholeiite suites and the PMG. A -- Thingmuli Iceland (Carmichael, 1964) - 30 analyses. B - Columbia Plateau (Waters, 1961) - 40 analyses. C - Abyssal tholeiites (retrieved from data-file of B. Gunn, U. de M.) - 80 analyses. D - Spilite (Valiance, 1960) - 90 analyses. E - Suggested immiscible globules in glassy-pillow margins (Yeats et aZ., 1973) - bulk compositions as open circles, globules as closed circles. F - Nasu zone, Japan (Kawano et al., 1961) - 27 analyses. G - Hachijo-jima, Izu Islands (Isshiki, 1963) - 43 analyses. H - Tonga-Kermadec Group (Ewart et al., 1973 - 42 analyses. I Raoul Island - Kermadec Group (Brothers and Searle, 1970) - 22 analyses. J - Composite tholeiite field with superposed Columbia River (horizontal shading) and abyssal (vertical shading) fields - 284 analyses. 8

the Columbia River and abyssal basalts. Because these fields overlap to a great extent we subsequently treat them as a single

composite tholeiite field, the outline of which is contoured to

embrace the greatest density of data points. Occasional out- liers have not been included within the field which incorporates 284 individual analyses. Examination of figure 2-J reveals that while the

more mafic tholeiite suites such as abyssal or continental do not plot near the PMG, the more diversified tholeiites of island k arcs both flank and overlap the PMG at highish Si02 concentrations. This however is not simply a phenomenon of diversifica-

tion in the suite, since while the tholeiites of oceanic islands (such as Thingmuli) show much the same silica range, they have

a consistently higher relative alumina plus alkalis content, and

hence plot well outside the PMG over all of its major axis

length (figure 2-A). Liquid immiscibility in abyssal tholeiites has

been proposed by Yeats et al., (1973) to explain the occurrence

of reddish brown and opaque globules found in glassy, pillow- margins. The bulk composition of these basalts and the co- existing globules are shown in figure 2-E. They all plot out-

side the PMG and the tie-lines between co-existing globules cross-cut the long-axis trends of both the composite tholeiite field, and the PMG. These tie-lines trends are dis-similar to most experimentally derived tie-lines that join immiscible

fractions plotting within the PMG, and hence another explanation - 9 - for their origin may be necessary. A similar proposal has been advanced to account for the formation of very small variolites in the pillow margins of abyssal basalts (Furnes, 1973).

However, these features involve spherulitic growth usually nucleated about a quench plagioclase-crystal, and most probably originate by simple, rapid immature crystallization without the intervention of liquid splitting (Géli_nas and Brooks, 1974).

It would seem that good textural evidences of immiscibility in most modern tholeiites is restricted to the inclusion type liquid-splitting such as observed by Roedder and Weiblen (1970).

Pseudo-ternary diagram considerations of the positioning of the tholeiite suites however, indicate that the best modern textural evidences of immiscibility are most likely to be encountered in the low-K tholeiites of island arcs.

Calc-alkaline suites. These suites have been taken from examples occurring in both continental and island arc environments. A distinction has been made (using Kuno's (1959) index) between normal and low-K talc-alkaline suites (figures 3-A to C and

E to G respectively). None of these suites plot very close to the PMG, and examination of the resulting composite fields

(figures 3-D and 3-H) indicates that the low-K and normal calc- alkaline suites are similarly placed with respect to the PMG

(with the low-K talc-alkaline field plotting slightly closer to the PMG). This is quite different from the low-K tholeiite series of island arcs which actually overlap and closely flank the PMG. - 10 -

510?

f L CaO•Mg0 Ne,O•K,O •LFeO•T1O, .A1103

WEIGHT %

FIGURE 3 - Modern Calc-alkaline and alkaline suites and the PMG. A - Aleutian Islands (Byers, 1959; Coats, 1952, 1959) - 111 analyses. B - Cascade (open field) (McBirney, 1968; Smith and Carmichael, 1968; Wise, 1969) - 126 analyses. Patricutin (shaded field) (Wilcox, 1954; Williams, 1950) 35 analyses. C - Chokai calc-alkaline suite (shaded field) (Kawano et aZ., 1961) - 22 analyses. D - Composite calc-alkaline field - 168 analyses. E - Talasea, New Britain (Lowder and Carmichael, 1970) - 25 analyses. F - Mariana Islands (Schmidt, 1957) - 12 analyses. G - Nasu Zone, Japan (Kawano et al., 1961) - 36 analyses. H - Composite low-K talc-alkaline field - 294 analyses. I - Tristan da Cunha (Baker et aZ., 1964) - 54 analyses. J - Gough and Oceanic Islands (Le Maitre, 1962) - 45 analyses. K - St. Helena Island (Baker, 1969) - 20 analyses. L - Composite alkaline field showing pronounced chemical gaps. - 119 analyses. Also shown are ocelli (closed circles) matrix (crosses) and bulk compositions (open circles) of alkaline intrusives (data from Philpotts (P) 1971; Furguson and Currie (F), 1971). Alkaline suites. These suites are taken from oceanic-island

environments (fig. 3-I to L). For the most part they plot far

removed from the PMG but they do trend toward it in the more

mafic compositions. A composite field based on all three suites

is given in figure 3-L, which also contains matrix-ocelli

compositions reported by Philpotts (1971) and Ferguson and

Currie (1971) for mafic alkaline intrusives. While the tie-

lines joining these fractions do not plot in the PMG and do not

parallel tie-lines for the experimentally produced immiscible

glasses plotting within the PMG (figure 1), they do conform with

the axial trend of the composite alkaline field. The abundance of chemical gaps in the composite volcanic field is noteworthy,

'cp rip,?iv since, of all magmatic compositions, some of the best textural and experimental evidence comes from continental mafic alkaline intrusives (e.g. Philpotts and Hodgson, 1968;

Philpotts, 1971; Ferguson and Currie, 1971; Strong and Harris,

1974; Carman et al., 1975). Chemical gaps have long been noteworthy in the literature (e.g. Chayes, 1963, Thompson, 1972), and the question arises as to whether these gaps are merely an artifact of sampling (119 samples in the field) or are real. If the latter, then the gaps could represent a surface manifestation of immiscibility which otherwise produces in continental intru- sive equivalents, convincing textural evidence of liquid- splitting. The lack of chemical gaps in the calc-alkaline composite field based on a comparable number of analyses (168) does not support the sampling artifact argument. - 12

Cao+MgO Na,0+K,0 +EFe0+TuO, +AI,O3 WEIGHT % EFeO

Na,0+K,0 Mg0

WEIGHT %

FIGURE 4 - Ancient tholeiitic and low-K calc-alkaline volcanics and the PMG. Data from Gélinas et ca., 1975. A to D - pseudo-ternary diagrams. E to F - respective AFM diagrams (the dividing lines on these diagrams distinguish the calc-alkaline from the tholeiitic series according to Kuno (1968) or to Irvine and Baragar (1971). A and E - The Dufault calc-alkaline series; 80 analyses. B and F - The Dufresnoy tholeiitic series, the open circles in B are variolites; 79 analyses. C and G - The Rouyn-Noranda tholeiitic series; 57 analyses. D and H - The Ruisseau Deguisier tholeiitic series (open fields and circles) and the Duparquet - Destor - Manneville tholeiitic series (closed circles and solid fields) - 33 analyses. - 13 -

ANCIENT VOLCANICS AND THE PMG

Archaean terrains are notably deficient in alkaline

rocks (Hart et al., 1970) and neither do they contain many normal

calc-alkaline suites. For the most part, the Archaean metavolcanic

piles consist of tholeiitic and low-K calc-alkaline series. Our

knowledge of the Abitibi metavolcanic belt has allowed us to

confidently define a series of alternating tholeiitic and low-K

calc-alkaline series (Gélinas et al., 1975), and hence we use this

data to furnish a comparison with the modern volcanics.

The data are plotted in figure 4, which also shows

the corresponding AFM diagrams. Examination of the pseudo-ternary

diagrams (fig. 4-B to D) reveals that the tholeiitc series form fields that both flank and overlap the PMG in much the same way as the modern low-K tholeiites from island arcs do. The low-K calc-

alkaline series (figure 4-A) however does not flank the PMG,

although, like the tholeiites, it does touch the PMG at very high

SiO2 concentrations. The only ancient volcanics that plot consistently in the PMG are certain komatiites, certain lunar rocks and

variolites from Archaean tholeiitic series. Plotting the Viljoen's

(1969) komatiite data on the pseudo-ternary diagram, reveals that it is the basaltic komatiites which are confined to the PMG, while the peridotitic komatiites plot outside it on the lower left flank

(figure 5-A). The Barberton and Badplaas komatiites cluster on opposite flanks of the PMG and define zones that trend directly along clinopyroxene fractionation lines. The Geluk type however Si02

Ca0+Mg0 Na20+120 +EFe0+Ti0 2 +A1203 WEIGHT %

FIGURE 5 - South African Komatiites and the PMG. A - Total rock compositions distinguishing Barberton (Ba), Badplaas (Bp), and Geluk (G) basaltic komatiites from peridotitic komatiites (P). Lunar rocks are shown as shaded field (for data sources see text). B - Variolitic basaltic komatiites (Ferguson and Currie, 1972) (variole (closed circle), total-rock (open circle), matrix (cross)). - 15 - shows an olivine effect, which is even more clearly portrayed in the peridotitic komatiites. Variolitic komatiites have been shown to be linked with immiscibility (Ferguson and Currie, 1972) and variole-matrix pairs from these rocks plot within or just on the right flank of the PMG (fig. 5-B).

The lunar rocks plot on the lower right flank of the immiscibility field (fig. 5-A; analyses from Brett et al., 1971; Biggar et al., 1971; Engel et al., 1971; Bence and Papike,

1972; and Mason et al., 1972). The liquid inclusions found in lunar rocks (Roedder and Weiblen 1971, 1972) occupy restricted domaines at each extremity of the PMG (figure 1). The variolite data from Archaean tholeiite series

(Gélinas et al., 1975) plot within the PMG defining a general trend of silica enrichment (figure 6-B). The same data, when plotted on the AFM diagram show a marked trend of iron-enrichment which coincides almost exactly with the Skaergaard trend (fig. bc).

The tie-lines for the separated fractions (variole and matrix) from some of these variolites are identical to most of the experimentally derived tie-lines which relate fractions that show strong partitioning of total Fe, Mg and Ca into the mafic fraction and SiO2 and alkalis into the felsic fraction (figure 6-p).

DISCUSSION

The composite fields for modern volcanics are plotted in figure 7. These volcanics define a vast domaine on the right flank of the PMG, with their respective positions being - 16 -

controlled by their relative alkali plus alumina contents. Successively from the alkali-alumina corner towards the PMG, the

suites encountered are alkaline, cale-alkaline, low-K calc- alkaline and tholeiitic. The only modern suite to overlap the PMG is the tholeiite suite and it does so only for the more silic

low-K tholeiites of island arcs. No other modern volcanic suite approaches the PMG, although the low-K and normal cale-alkaline fields closely parallel the solves-liquids intersection portrayed as the boundary of the PMG. The composite Archaean tholeiite and low-K calc- alkaline suites are shown in figure 8. The ancient tholeiite suite

is essentially identical with the modern tholeiite suite, and most esperinlly with the low-.K tholeiite suite of island arcs. Both of these suites overlap the PMG at higher SiO concentrations, 2 with the ancient suite displaying a slight trend to even higher contents than those encountered in the modern day examples. SiO2 The composite ancient low-K calc-alkaline field differs from the modern-day field. While the latter never approaches the PMG, the former is positioned significantly closer

to the PMG and actually touches it at high Si02 concentrations. For the most part the ancient volcanics possess lower alumina plus alkalis than their modern, chemical-counterparts. Relating these observations to immiscibility, it

would appear that ancient and tholeiitic magmas have behaved

similarly. On examining the abundance of textural phenomena indicative of immiscibility, it is clear that the Archaean terrains 502

C TOTAL ROCK

• VARIOLE

4- MATRIX

,' , DI O ~0 +/~,C ,or/4111!,,/ ~ - 50 FOAir_,-

Ca0+Mg0 Na20 + K 20 +EFe0+TI0 2 +A1203 WEIGHT %

FIGURE 6 - Archean variolites of Rouyn-Noranda region and the PMG. A -- variole - total rock-matrix fractions B - variolite total-rocks. Si02

Ca0+Mg0 Na20+ K20

+EFe0+Ti0z +A1203 WEIGHT To

FIGURE 7 — Composite fields for modern day volcanics in relationship to the PMG. Si02

0 ~~~k 0 ~\~~O ~ ~ / ÷~ ~\~`~ ~P~ ~~P~,~ «•«;-

~~p\,~ P~~Q^~ <<, GP~c; CJP ,~C, & O~~- ~ ~~ o \, \.. ~

Ca0+Mg0 Na 20+K20 +EFe0+Ti0 2 +A120 3 WEIGHT %

k'IGURE 8 - Composite modern (A) and ancient (B) tholeiite and low-K calc- alkaline fields and the PMG. - 20 - contain numerous examples whereas evidence in modern terrains is lacking. The best examples of immiscibility found in Archaean terrains, are the variolitic lavas of Canada and South Africa, and they are widespread. Is this a real phenomenon indicating a polarized, time-dependant susceptibility of tholeiitic and ancient calc-alkaline magmas to undergo liquid immiscibility? There is evidence to suggest that it may be so. Lower alumina and alkali contents are primarily responsible for driving rock fields into the PMG. It has been suggested previously (e.g. Hart et al., 1970; Brooks and Hart, 1972) that because of a higher geothermal gradient, high rates of convective upwelling and reduced lithospheric thickness were probable in the Precambrien. Extensive and relatively shallow partial melting is a direct prediction of such thermal considerations and this has already been used in discussing the origin of komatiite (e.g. Viljoen and Viljoen, 1969; Brooks and Hart, 1974;

Cawthorn and Strong, 1974). Because extensive partial melting produces magmas with lower alkali and alumina contents, this model implies that the production of immiscibility-prone magmas was more common in the Archaean than today. Further support for a time dependant polarized likelihood of encountering immiscibility-prone magmas is provided by consideration of models of mantle water-content with time. Strong and Stevens (1974) in discussing the change in Earth's behavior at the Archaean-Proterozoic boundary postulate that the differences between Archaean and younger sections result from a - 21 - water pressure effect on the peridotite solidus. With degassing of the mantle, smaller and smaller degrees of partial melting of mantle material occur. This is in direct accord with the previously discussed thermal model. However, water not only has a role in partial melting considerations, it has been shown to have a direct effect on the field of immiscibility. Holgate (1954) proposed that the field of immiscibility was enlarged at high water pressure, and as pointed out by

Watson and Dickey (1975), water as a liquidus depressant has been found necessary in many cases of experimental verification of the process (e.g. Philpotts and Hodgson, 1968; Philpotts, 1971; Ferguson and Currie, 1972). Arguments however can be advanced for both wet and dry Archaean volcanism; because a) the Canadian Archaean of 2.7 b.y. contains abundant andesites, and b) the Archaean tholeiites most closely resemble low-K island-arc tholeiites, we are inclined to believe that the volcanism during the formation of the Archaean metavolcanic belts may have been "wet". Whatever the water content however, this model, implicating decreasing water contents in the accessible mantle with time, again favors the likelihood of encountering more evidences of immiscibility in Archaean, compared with younger volcanic terrains. It would seem on the basis of this evaluation of some ancient and modern volcanic suites that a direct link between tholeiitic magmatism and immiscibility is possible for certain tholeiitic and ancient talc-alkaline magma compositions. These compositions are mostly in the intermediate to felsic range which is reminiscent of Grieg's (1927) experimental observation that - 22 -

CaO,. FeO, Mg0 and Fe203 show incomplete miscibility with Si02 at high Si02 concentrations. An important question relates to whether the potential immiscibility in ancient and modern tholeiitic magmas

(and in ancient low-K calc-alkaline magmas) is restricted to the late-stage residual liquids only, or whether it involves much of the magma at some pre-liquidus stage in its history. The previous discussion concerning lowering of relative alumina plus alkali contents, and the role of water in magmas, is primarily applicable to the direct production of an immiscible-prone melt. However differentiation characteristically enhances a residual mamga in liquidus depressants such as iron and alkalis. Very high iron copteptc are especially necessary for successful immiscible splitting experiments in dry systems (e.g. Roedder and Weiblen, 1970, McBirney, 1975). Hence where a tectonic environment does not allow direct enhancement of liquidus depressants in a primary magma, and that magma composition is not favorable to liquid immiscible splitting, it may still reach a two-liquid stage in its residual phases. As yet we cannot distinguish with certainty which alternative was responsible for the Archaean variolites in the Abitibi belt, although their high iron contents removes the necessity of their being the product of "wet" tholeiitic magmas. Because however the

Archaean variolitic lavas indicate immiscibility is restricted to

the higher S102 levels of tholeiitic magma evolution, and because the tholeiite composite field overlaps the PMG only at higher Si02 levels, it is probable that some magmatic differentiation is always - 23 -

required before a tholeiitic (or ancient, low-K talc-alkaline) magma will split into two liquid fractions. This in turn favors

Fe as a more important potential liquidus depressant than water in tholeiite magmas.

The evaluation presented in this paper is not complete. As more and more reliable Archaean data become available

(reliable implying filtered for non-primary rock compositions) we hope to extent our conclusions. At this stage however we can state that the recent abundance of liquid-immiscibility observations associated with tholeiitic magmatism is not in conflict with what is experimentally known about immiscibility. In fact the reverse is true, and immiscibility considerations lead to a direct

»,-pri; rt; nn that textural evidences of the phenomenon should be common in at least the more diversified members of the tholeiitic suites. Furthermore there is evidence to indicate that such textural evidences may be more abundant in ancient rather than in modern terrains. - 24 -

ACKNOWLEDGEMENTS

The study has been supported by the Quebec Department of Natural Resources, the Minister of Education (No -

CRP-293-72,73 to L. G.), the National Research Council of Canada

(No - A5581 to C. B.) and the Carnegie Institution of Washington. We wish to thank G. Daoust for computer related assistance, M. Demidoff for drafting the final diagrams and B. Gunn for access to his data file. Numerous people have participated in discussions leading to the development of this paper, and we are especially grateful to S. R. Hart, A. E. Hoffman and B. Watson. - 25 -

REFERENCES

Anderson, A. T., and Gottfried, D., 1971. Contrasting behavior of P, Ti, and Nb in a Differentiated High-Alumina Olivine- Tholeiite and a Caic-Alkaline Andesite Suite. Geol Soc. America Bull, 82, 1929-1942.

Baker, 1., 1969. Petrology of the volcanic rocks of Saint Helena Island, South Atlantic. Geol. Soc. Amer., Bull. 80, pp. 1283-1310.

Baker, P. E., Gass, I. G., Harris, P. G. and Le Maitre, R. H., 1964. The Volcanological report of the Royal Society expedi- tion to Tristan Da Cunha, 1962. Phil. Trans. Roy. Soc. Lond., Ser. A, Vol. 256, pp. 439-578.

Bence, A. E. and Papike, J. J., 1972. Pyroxene as recorders of lunar petrogenesis chemical trends due to crystal-liquid interaction. Proc. Third Lunar Sci. Conf., 1, pp. 431-469.

Biggar, G. M., O'Hara, M. J., Peckett, A. and Humphries, D. J., 1971. Lunar lavas and the achondrites: petrogenesis of protohypersthcne basalts in the NEARIA Lava Lakes. Proc. ~~ L U ~ 1 ll Lunar U l..L • V V :~.i . 1, ~. 1• " -6'Z

Bowen, N. L. and Schairer, J. F., 1938. Crystallization equilibrium in nepheline-albite-silica mixtures with fayalite. J. Geol., 46, pp. 397-411.

Brett, R., Butler.Jr., P., Meyer Jr., C., Reid, A. M., Takeda, H. and Williams, R., 1971. Apollo 12, a mineralogical and petrological study. Proc. Second Lunar Sci. Conf., 1, pp. 301-317.

Brooks, C. and Hart, S. R., 1972. An extrusive basaltic komatiite from a Canadian metavolcanic belt. Can.J. Earth Sci , 9, pp. 1250-1253.

Brooks, C. and Hart, S. R., 1974. On the significance of komatiite. Geology, February 1974, pp. 107-110.

Brothers, R. N. and Searle, E. J., 1970. The geology of Raoul island, Kermadec group, Southwest Pacific. Bull. Volc., 34, pp. 7-37.

Byers Jr., F. M., 1959. Geology of Ummak and Bogoslov Islands, Aleutian Islands, Alaska. U. S. Geol. Surv., Bull. 1028-L, pp. 267-369. 26

Carman Jr., M. F., Cameron, M., Gunn, B., Cameron, K. L. and Butler, J. C., 1975. Petrology of Rattleshake Mountain sill, Big Bend National Park, Texas. Geol. Soc. Amer., Bull. 86, pp. 177-193. Carmichael, I. S. E., 1964. The petrology of Thingmuli, a Tertiary volcano in Eastern Iceland. J. Petrol., 5, pp. 435- 460. Cawthorn, R. G. and Strong, D. F., 1974. The petrogenesis of komatiites and related rocks as evidence for a layered upper mantle. E.P.S.L. 74, pp. 369-375. Chayes, F., 1963. Relative abundance of intermediate members of the oceanic basalt-trachyte association. J. Geophys. Res., 68, pp. 1519-1534. Coats, R. R., 1952. Magmatic differentiation in. Tertiary and Quaternary volcanic rocks from Adak and Kanaga Islands, Aleutian Islands, Alaska. Geol. Soc. Amer., Bull. 63, pp. 485-514. Coats, R. R., 1959. Geological reconnaissance of Semiso-Pochnoi Island, Western Aleutian Islands, Alaska. U. S. Geol. Surv., p,01 . 1 n7R-C . nn. .477-519. De, Aniruccha, 1974. Silicate liquid immiscibility in the Deccan traps and its petrogenic significance. Geol. Soc. Am., Bull. 85, pp. 471-474.

Engel, A. E. J., Engel, C. G., Sutton, A. L. and Myers, A. T., •1971. Composition of five Apollo 11 and Apollo 12 rocks and one Apollo 11 soil and some petrogenic considerations. Proc. Second Lunar Sci. Conf. 1, pp. 439-448. Ewart, A., Bryan, W. B. and Gill, J. B., 1973. Mineralogy and geochemistry of the younger volcanic islands of Tonga, S. W. Pacific. J. Petrol, 14, Pt. 3, pp. 429-465. Ferguson, J. and Currie, K. L., 1971. Evidence of liquid immiscibility in alkaline ul.trabasi.c dikes at Callender Bay, Ontario. J. Petrol, 12, Pt. 3, pp. 561-585. Ferguson, J. and Currie, K. L., 1972. Silicate immiscibility in the ancient basalts of the Barberton Mountain land, Transvaal. Nat. Phys. Sci. , 235, pp. 86-89.

Furnes, II., 1973. Variolitic structure in Ordovician pillow lava and its possible significance as an environmental indicator. Geology, 1, pp. 27-30. - 27 -

Gélinas, L., 1974. Textural and chemical evidence of liquid immiscibility in variolitic lavas. (Abstract) Eos. Trans. Amer. Geophys. Union, 55, p. 486.

Gélinas, L. and Brooks, C., 1974. Archean quench-texture tholeiites. Can J. Earth Sci., 11, No. 2, pp. 324-340.

Gélinas, L., Brooks, C. and Trzcienski Jr., W. E., 1975. Archean variolites-quenched immiscible liquids.

Gélinas, L., Brooks, C., Perrault, G., Carignan, J., Trudel, P., and Grasso, F., 1975. Geochemical Trends in Archean Volcanic Rocks, Rouyn-Noranda, Quebec.

Greig, J. W., 1927. Immiscibility in silicate melts. Am. Jour. Sci., Fifth series, Vol. XIII, No. 73, pp. 1-44, and Vol. XIII, No. 74, pp. 133-154.

Hart, S. R., Brooks, C., Krogh, T. E. and Davis, G. L, 1970. Ancient and modern volcanic rocks: a trace element model. E.P.S.L., 10, pp. 17-28.

Holgate, N., 1954. The Role of liquid immiscibility in igneous petrogenesis. Jour. Geol., 62, pp. 439-480.

Isshiki, N., 1963. Petrology of Hachijo-Jima volcano group, Seven Izu Islands, Japan. Journal F. Sci., Univ. of Tokyo, Section II, Vol. XV, pp. 91-135.

Irvine, T. N. and Baragar, W. R. A., 1971. A Guide to the classi- fication of the common volcanic rocks. Can J. Earth Sci., 8, pp. 523-549.

Kawano, Y., Yagi, K. and Aoki, K., 1961. Petrography and petrochemistry of the volcanic rocks of Quaternary volcanoes of Northeastern Japan. The Sci. Reports of the Tohoko Univ., Series III, Vol. VII, No. 1.

Kuno, H., 1959. Origin of Cenozoic petrographic provinces of Japan and surrounding areas. Bull. Volc., 20, pp. 37-76.

Kuno, II., 1968. Origin of andesite and its bearing on the island. arc structure. Bull. Volc. 32, pp. 141-176.

Le Maitre, R. W., 1962. Petrology of volcanic rocks, Gough Island South Atlantic. Geol. Soc. Amer., Bull, 73, pp. 1309-1340. - 28 -

Lowder, G. G. and Carmichael, I. S. E., 1970. The volcanoes and caldera of Talasea, New-Britain: geology and petrology. Geol. Soc. Amer., Bull. 81, pp. 17-38. Mason, B., Jarosewich, E. and Melson, W. G., 1972. Mineralogy, petrology, and chemical composition of lunar samples 15095, 15256, 15271, 15471, 15475, 15476, 15535, 15555, 15556. Proc. Third Lunar Sci. Conf. Geochim. Cosmochim. Acta, 1, pp. 785-796. Massion, P. J. and Koster Van Groos, A. F., 1973. Liquid immiscibility in silicates. Nat. Phys. Sci., 245, pp. 60-63. McBirney, A. R., 1968. Petrochemistry of the cascade andesite volcanoes. Oreg. Dept. Geol. Miner. Ind. Bull., 62, pp. 101-107. McBirney, A. R., 1975. Differentiation of the Ska.ergaard Intrusion, Nature, 253, pp. 691-694. Philpotts, A. R., 1971. Immiscibility between feldspathic and gabbroic magmas. Nat. Phys. Sci., 229, pp. 107-109.

Philpotts, A. R. and Hodgson, C. J., 1968. Role of liquid immiscibility in alkaline rock genesis. 23th Int. Geol. Cong., Vol. 2, pp. 1'/5-186. Roedder, E., 1951. Low temperature liquid immiscibility in the system K20-Pe0-A1203-Sî02. Am. Min., 36, pp. 282-286. Roedder, E., 1953. Liquid immiscibility in the system K 0-Pe0- A1203-Sî02 (abst.). Geol. Soc. Amer. Bull., 64, p. 66.14 Roedder, E., 1956. The role of liquid immiscibility in igneous petrogenesis: a discussion. J. Geol., 64, pp. 84-88. Roedder, E. and Weiblen, P. W., 1970. Silicate liquid immiscibility in lunar magmas, evidenced by melt inclusions in lunar rocks. Science, 167, Vol. 10, pp. 641-644. Roedder, E. and R'eiblon, P. W., 1971. Petrology of the silicate melt inclusions, Apollo 11 and Apollo 12 and terrestrial equivalents. Proc. Second Lunar Sci. Conf., Geochim. Cosmochim. Acta, 1, pp. 507-528. Schairer, J. P. and Bowen, N. L., 1938. The system leucite- diopside-silica. Am. Jour. Sci.., Ser. 5, Vol. 35-A, pp. 289- 309. Schmidt, R. G., 1957. Geology of Saipan, Mariana Islands: chap. B, petrology of the volcanic rocks. U. S. Geol. Surv., Pap. 280-B, pp. 127-175. - 29 -

Smith, A. L. and Carmichael, I. S. E., 1968. Quaternary lavas from the Southern Cascades, Western U.S.A. Contr. Mineral. Petrol., 19, pp. 212-238.

Strong, D. F. and Harris, A., 1974. The petrology of Mesozoic alkaline intrusives of Central Newfoundland. Can J. Earth Sci., 11, No. 9, pp. 1208-1219.

Strong, D. F. and Stevens, R. K., 1974. Contrasting Archean and Proterozoic geological regimes: a possible thermal explanation. Nature 249 pp. 545-546.

Thompson, R. N., 1972. Evidence for a chemical discontinuity near the basalt-andesite transition in many anorogenic volcanic suites. Nature, 236, No. 5342, pp. 106-110.

Valiance, T. G., 1960. Concerning spilites. Proc. Linn. Soc. N.S.W., 85, pp. 8-52.

Viljoen, J. J. and Viljoen, R. P., 1969. Evidence for the existence of a mobile extrusive peridotitic magma from Komati formation of the Onverwacht group. Geol. Soc. South Africa, Spec. Publ. 2, Upper Mantle Project, pp. 87-112.

Waters, A. C., 1961. Stratigraphic and lithologie variations in the Columbia River basalt. Am. J. Sci., 259, pp. 583-611.

Watson, E. G. and Dickey Jr., J. S., 1975. Silicate liquid immiscibility in alkaline mafic magmas: evidences in a White Mountain magma series dike, with a discussion of its origin. Geochem. Cos. Acta (in press).

Wilcox, R. E., 1954. Petrology of the Paricutin Volcano, Mexico: U.S. Geol. Survey Bull. 965-C, pp. 281-353.

Williams, H., 1950. Volcanoes of the Paricutin region: U.S. Geol. Survey Bull. 965-B, pp. 165-179.

Wise, W. S., 1969. Geology and petrology of the Mt. Wood area: a study of high Cascade volcanism. Geol. Soc. Amer., Bull. 880 pp. 969-1006.

Yeats, R. S., Forbes, W. C., Scheidegger, K. F., Ross Heart, G. and Van Andel, T. H., 1973. Core from Cretaceous basalts, Central Equatorial Pacific, Leg 16. Deep sea drilling project. Geol. Soc. Amer., Bull. 84, pp. 871-882. Î ç ç ç

J J J J