On the tectonic regimes of ophiolite genesis

On the tectonic regimes of ophiolite genesis

Earth and Planetary Science Letters, 43 ( 1979) 93- 102 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands 93 [6] ON ...

916KB Sizes 0 Downloads 8 Views

Earth and Planetary Science Letters, 43 ( 1979) 93- 102 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

93

[6]

ON THE TECTONIC REGIMES OF OPHIOLITE GENESIS H.D. UPADHYAY Department of Earth Sciences, Northeastern Illinois University, Chicago, IL 60625 (U.S.A.) and E.R.W. NEALE Institute of Sedimentary and Petroleum Geology, Calgary, Alta. T2L 2A 7 {Canada)

Received July 18, 1978 Revised version received January 2, 1979

Ophiolites and their associated rocks are examined in terms of field relationships and petrochemistry. Many pre-obduction plagiogranite and dioritic dykes in the ophiolites of southern Quebec, Vourinos, California, Oregon and Newfoundland are judged not to be the end phase of differentiation of tholeiitic magma. Repetitive sedimentary-volcanic formations that overlie many ophiolite suites resemble Cenozoic volcaniclastic sequences of the New Hebrides and Tonga arcs and possibly an analogy may be drawn to the several seismic layers detected in modern marginal basins. Variation diagrams for Ti, Zr, and Cr show that dykes and pillow lavas associated with ophiolite suites from Scotland, Greece, Cyprus and Newfoundland have island arc or calc-alkaline trends. It is difficult to reconcile some of these data with origin at major oceanic ridges. Most evidence indicates proximity to island arcs and origin within marginal basins of western Pacific type. Evolution of such basins was probably initiated by distension along a continental edge resulting in the formation of a graben so that island arcs were formed on break-away continental fragments and the grabens replaced by marginal basins in which komatiitic, calc-alkaline, tholeiitic and other lavas were erupted. The rocks are more diverse than those of major ocean basins due to the interaction of downgoing oceanic plate and adjacent continental margin, the role of subducted sediments and the insulating effect of a thick sedimentary-volcanic cover.

1. Introduction An interesting debate has been provoked by Miyashiro's [ 1] suggestion that the Troodos ophiolite was formed beneath an island arc, and not at a spreading ridge as commonly stated. Since the Troodos ophiolite is similar to many ophiolites around the world, it is important to re-examine the environments in which ophiolites appear to have originated in the light of field and petrochemical data. Miyashiro [1,2] uses the following characteristics of island arc series to distinguish them from spreading-ridge assemblages: (a) an FeO (t) */MgO ratio of more than 2 ; ( b ) 52.5% SiO2 (which increases

with increasing FeO(t)/MgO ratio during fractionation of calc-alkaline suites); (c) relatively lower TiO2 that progressively decreases with increasing FeO(t)/ MgO ratios; and (d) presence of silicic and intermediate rocks. However, these criteria unfortunately do not apply easily to those rocks that have undergone metamorphism, metasomatism or hydrothermal alteration, as some or all of the chemical elements cited by Miyashiro become mobile during such processes. Hence, Miyashiro's conclusions have been disputed by many (see Smith [4] for summary). It is imperative, therefore, that the problem of the * FeO(t) means total iron as FeO.

94 petrochemical environment of ophiolite genesis be resolved by using a combination of factors, such as field relationships, petrology, lithologies of the overlying sedimentary]pyroclastic formations, major element geochemistry (provided the extent of alteration is taken into account) and the relatively less mobile trace elements. In this article, we discuss the problem of ophiolite origin in the light of all the above-mentioned factors, with emphasis on the evidence from certain Newfoundland occurrences. All existing information leads to the conclusion that most well-studied ophiolites around the world represent the crvst of marginal ocean basins that evolved adjacent to island arc systems.

2. Field relationships and magmatic rock types

2.1. Field relationships The sheeted dyke member represents probably the most spectacular feature of many ophiolite suites and is generally taken as evidence for their genesis through a process of lateral spreading. The simple geometry of this process suggests that the sheeted dykes should strike parallel or nearly parallel to the trend of the spreading ridge. Thus, unless the closing of an ecean was highly asymmetrical or the oceanic crust underwent a rotation, the best approximation of the trend of a former oceanic ridge should be furnished by the trends of the sheeted dykes in preserved remnants of ophiolites. However, in the Newfoundland Appalachians, where the trends of the sheeted dykes are well documented, a summary of the data [6] shows that in most cases the dykes are oblique or at high angles to the trend of the ophiolite belt in which they are located. Since such discordances are rare or unknown in the modern major ocean floors, as manifested by the magnetic anomaly patterns, we infer that the discordant Appalachian ophiolites were generated in marginal ocean basins behind island arcs where the spreading may have taken place in a disorderly manner along discontinuous zones that are perhaps randomly located or randomly oriented [7,8]. It has also been suggested that the spreading in such marginal basins could have taken place from a ridge that was oblique to the zone

of distensions in the adjacent main Proto-Atlantic Ocean [9]. An excellent example of an ophiolite representing marginal-basin crust comes from the Mesozoic terranes of southern Chile where the ophiolite belt is flanked by a continuous island arc of andesitic bodies on one side, and stable continental crust on the other [10]. Since the internal stratigraphy of Chilean ophiolites is similar to many other around the world, their tectonic setting merits consideration in any evaluation of the problem of ophiolite genesis.

2.2. Magmatic rock types The presence of diorites and plagiogranites within ophiolites has been suggested as a reason for including them as part of island arc-type suites [1 ]. This criterion must be used with caution since small amounts of such rocks can be produced through the extreme fractionation of a tholeiitic magma or through post-magmatic alterations. However, it is a fact that significant amounts of primary dioritic rocks occur in the ophiolites of Papua, Troodos, Vourinos and the Bay of Islands and, in some, these rocks are not necessarily restricted to the top part of the gabbro member as would be expected if they were an end product of the differentiation of a tholeiitic magma (Fig. 1). Such dioritic rocks by no means constitute incontrovertible evidence for an island arc environment but their presence must be carefully weighed in any geochemical interpretation of the ophiolite environment. There are several well-documented cases of silicic intrusions cutting across and showing chilled margins against mafic and ultramafic members of the ophiolite suite. They include those in southern Quebec [11,12], Vourinos [13,14], Canyon Mountain [15] and Preston Peak, California [17]. In the Betts Cove ophiolite, dyke swarms of calc-alkaline affinity cut across the basal ultramafic and gabbroic members and since such dykes also occur within the overlying main sheeted complex they must be coeval with the latter. As argued earlier for the Vourinos ophiolite, these dyke swarms could not have evolved from the same magma chamber as that in which the basal cumulates settled down. Primary silicic minerals, some of which formed through the devitrification of glass, occur in many dyke and pillow lava rocks, indicating the

95 16

l
.15

2.2

2.9

-"2.01 10 ~::~::.:~1

4.0 5.0~L •"0.6~ iDodm ii ~ q P '

11

4.9

6.69 +0.26

7.0

¢

6.7 10

>8.0

7.7

7.8

=£,

15

a

b

c

d

e

f

g

IZ ] I1NIN!17-2 I--;I

h

i

FVI

j

k

1

m

I" v 11';,, I

n

o

p

q

r

s

ln;q

Fig. 1 Comparative sections of ophiolites and related island arc assemblage (columns 1 - 1 3 ) and modern oceanic crust-mantle (column 14-16). 1 = Thetford Mines ophiolite, Quebec (synthesized from Laurent [11] and St. Julien and Hubert [22]); 2 = Betts Cove ophiolite and related rocks of the Snooks Arm Group, Newfoundland (modified from Upadhyay et al. [18]); 3 = Lush's Bight ophiolite Newfoundland [20] ; 4 = pillow lava-pyroclastic sequence of Western Arm Group, Newfoundland with Lush's Bight ophiolitic rocks (LB) at the base [19]; 5 = Bale Verte ophiolite, Newfoundland [6]; 6 = NOrth Arm Mountain complex, Bay of Islands, Newfoundland (J. Malpas, personal communication, 1973). 7 = island arc assemblage of Cutwell Group, Newfoundland [ 30]; 8 = Moretons Harbour island arc succession, Newfoundland [31]; 9 = Troodos ophiolite, Cyprus [33]; 10 = Papuan ophiolite [34]; 11 = Canyon Mountain complex, Oregon [ 15,16]; 12 = Vourinos ophiolite, Greece [13,14]; 13 = Sarmiento complex, southern Andes [ 10] ; 14 = typical section of oceanic crust and mantle (as synthesized in Dewey and Bird [261); 15 = Balearic basin [24]; 16 = Aleutian basin [25]. Explanation of symbols: a = ultramafic rock; b = gabbro; c = diorite/quartz diorite/trondhjemite; d = older sialic material assimilated in gabbro; e = sheeted dykes; f = pillow lava; g = pillow breccia; h = deep-sea sediments; i = volcaniclastic aggiomerates and pyroclastics;j = volcanogenic fly sch and tuff; k = mafic siUs;1 = calc-alkaline volcanics; m = silicic volcanics of island a~c origin; n = limestone; o = dioritic/calc-alkaline dykes, coeval with ophiolite suite (width exaggerated); p = explosive breccia; q = ~61ange; r = unconformity; s = fault.

p r e s e n c e o f m u c h m o r e silica t h a n is k n o w n f r o m t h e m a j o r m i d - o c e a n i c ridges. Dacitic lenses a n d h u g e g r a n o d i o r i t i c screens w i t h i n t h e s h e e t e d d y k e s a n d t h e p r e s e n c e o f m a g m a t i c q u a r t z in t h e " l e u c o -

g a b b r o " o f the B e t t s Cove o p h i o l i t e is f u r t h e r e v i d e n c e o f t h e p r e v a l e n c e o f a siliceous e n v i r o n m e n t . A l t h o u g h in this article we have b e e n able t o cite o n l y a few o p h i o l i t e s c o n t a i n i n g p r i m a r y dioritic/plagio-

96 granitic material that was not necessarily produced as the end phase of differentiation of the tholeiitic magma chamber in which the cumulate members were laid down, it is probable that more detailed work in other places may yield further examples of this.

3. Volcano-sedimentary sequences above ophiolites The composition of the sedimentary/pyroclastic formations that invariably overlie the ophiolite suites and, in some places, alternate with pillow lava formations at higher levels, is an extremely important parameter in evaluating the tectonic setting of ophiolite genesis. The occurrence of such repetitive pillow lava and/or pyroclastic sequences is possibly best exemplified by Newfoundland's Snooks Arm Group (Betts Cove ophiolite [18]), Western Arm Group (Lush's Bight ophiolite sequence [19,20]) and the Mings Bight ophiolite [6]. In the ophiolites of southern Quebec, the pillow lava member is overlain by a thin cherty argillite unit which, in turn, is followed by the Lower Ordovician "wildflysch-type sediments" of St. Daniel Formation [21 ];this is capped unconformably by the Middle Ordovician flysch of the Magog Group that interdigitates with the calc-alkaline lavas of the Ascot/Weedon Groups [22]. The uppermost lavas and volcaniclastic rocks have been recognized as remnants of an island arc sequence [231. Such successions are not known to exist on any of the present-day floors of the major oceans. On the other hand, some of the modern marginal ocean basins [7,8] exhibit several seismic layers ([24,25]; Fig. 1) that may be interpreted as analogous to the repetitive volcanic sedimentary formations overlying the ophiolites under discussion. The renewed volcanic phases associated with these marginal basins may have been triggered by intermittent subduction of the major ocean floor into the nearby trench [26]. The composition of many of these sedimentary pyroclastic formations consistently indicates the presence of a nearby island arc. For example, the pyroxene-bearing andesitic agglomerates overlying the Betts Cove, Mings Bight [27], Bale Verte [6] and Lush's Bight [19] ophiolites show remarkable

similarity to some of the slumped turbidites, structureless rudites, agglomerates, and quartz-poor volcaniclastic rocks described from the Miocene volcanic arc of New Hebrides [28] and the tufts, agglomerates and conglomerates of the Eocene island arc of Eua, Tonga [29]. This is further attested by the occurrence of Ordovician andesitic volcaniclastics in the established island arc successions of the Long Island and Moretons Harbour areas of Newfoundland [30,31 ]. Volcaniclastic sediments are also known to occur above the ophiolites of the Bay of Islands, Newfoundland [32], and those of Chile [10], where in the latter they show an andesitic affinity. Turbidites derived from a volcanic source, common in many Appalachian ophiolites, have been suggested as indicative of an island arc environment even by those [3] who favoured a mid-oceanic ridge origin for the Troodos ophiolite.

4. Geochemistry It is recognized that secondary processes could mobilize certain major elements in basaltic rocks and, therefore, such elements might not be reliable discriminants of tectonic environments of various rock suites. We suggest that the reliability of major elements be evaluated by carefully comparing them with the less mobile (or immobile) trace elements. In the case of Betts Cove, plots of trace elements that are known to have minimal or no mobility during secondary processes [35] show a fairly high degree of scatter in some variation diagrams (Fig. 4) and therefore may indicate that the geochemical diversity is a primary feature. This is consistent with the observation that, at this locality, composition of dykes in the sheeted complex varies from plagiogranite to peridotite although diabase is predominant. It is not unlikely, therefore, that the scatter in the major element plots is, to some extent, a reflection of such contrasting lithologies rather than modifications by post-magmatic processes. This may perhaps indicate that the Betts Cove ophiolite was generated from "composite" magmas possibly tapped from different chambers sited beneath the zone of distension, a model similar to that envisaged by Moores and Vine [33] for the Troodos massif.

97

4.1. AFM diagrams and Si02 contents The AFM plots of the dykes and pillow lavas of the Bay of Islands, Mings Bight, Betts Cove and Lush's Bight ophiolites are presented in Fig. 2. Most of the samples can be classified as calc-alkaline, especially if the field suggested by Best [38] is used. Since almost the same number of these also plot within the calc-alkaline or island arc basalt field in the (more stable) trace element diagrams, it appears that the AFM diagrams may indicate reasonably close approximations of their magmatic affinity; metasomatic redistribution of elements, especially alkalies (J.A. Pearce, personal communication, 1978), however, should be regarded as a strong possibility. Miyashiro ([ 1] ; and personal communication, 1978) has suggested that an SiO2 percentage of 52.5 or over can be taken as evidence against the midoceanic ridge origin of a volcanic suite. As shown in Fig. 2, the SiO2 percentages in the Bay of Islands dykes and lavas cluster around 52.5 and in Mings Bight, Betts Cove and Lush's Bight they average approximately 52.5% or over. Judging from the fairly large population of high-SiO2 samples and, more BAY OF ISLANDS

~IHGSBIGHT

BETTS COVE

LUSHSBIGHT

F

A 12F

(s)

3 ~ 40

M ~)

so

[ (e)

[ (H)

ne 6'o SiO2

7'o

Fig. 2. AFM (A = Na20 + K20 , F = total iron as FeO, and M = MgO) and alkali-silica diagrams for the sheeted dyke and pillow lava rocks of the Bay of Island (A, B), Mings Bight (C, D), Betts Cove (E, F) and Lush's Bight (G, H) ophiolites in Newfoundland (after Strong [36]). The solid line divides tholeiitic (TH) and calc-alkaline (CA) fields [37]. The broken lines enclose the ca]c-alkaline field as suggested by Best [38]. Note the relatively high SiO2 content of these rocks in the alkali-silica diagram.

importantly, from their plots in the calc-alkaline and/ or island arc fields in the less mobile trace element diagrams (Fig. 4), it appears that in some of these the S i Q was not necessarily enhanced by metasomatic processes.

4.2. Variation of FeO(t)/MgO with Si02, FeO (t) and Ti02 A second approach is applied to the data of the Betts Cove ophiolite along the lines suggested by Miyashiro [1 ], namely the variations of SiO2, FeO (t), and TiO2 with FeO(t)/MgO ratios. Out of 45 plots (Fig. 3), all but seven fall in the calc-alkaline field. In the latter, the plots maintain a linearity that points to an increase in SiO2 contents as well as in FeO(t)/MgO ratios with advancing stages of differentiation. The rate of increase in the FeO(t)/MgO ratio in relation to the variation in FeO (t) is not as rapid as suggested by Miyashiro [1 ] for a calc-alkaline trend but it does indicate that the FeO (t) contents of these samples drop with increasing FeO(t)/MgO ratios (Fig. 3C). The TiO 2 plots (Fig. 3B) also suggest a trend parallel to, but at a slightly lower TiO2 level than, a "typical calc-alkaline trend". The SiO2 percentages of samples falling in the calc-alkaline field (Fig. 3A) varies approximately from 44% to 57% and textural evidence shows that these silica values are of primary origin in most. Miyashiro's criterion of an FeO(t)/ MgO ratio of more than two is not met by the Betts Cove samples even though they plot in the calcalkaline field. A similar ambiguity occurs in the Baer Bassit [5] and Troodos ophiolites. Many of the Betts Cove samples belong to a komatiitic stock in which the MgO values have been very little modified by metasomatism [39] and hence the FeO(t)/MgO ratios of <2 must reflect their original chemistry. Archean komatiites and associated rocks also yield FeO(t)/ MgO ratios of <2 and yet a variety of evidence points to their origin in island arc settings [40]. The FeO(t)/MgO ratio should therefore be used with caution, especially when dealing with high-MgO intrusive-extrusive sequences. Of the seven samples that fall in the tholeiitic field (Fig. 3), most are also identified as tholeiitic/abyssal tholeiites in the AFM (Fig. 2) and in the trace element (Fig. 4) diagrams. These seven are uniformly distributed within the upper half of the pillow lava of

98 60

15

/

(A)

® I

/

/

/

/

";'

• •

//

/

//

/I 55

(C)

///

I

/

(~)

%

i II

° I

B:

i

(~

12

/

o

iI

/

s01

~'SSAL TBOLEIITES

10

", " //,,//f,'/,.," •

I

x

j

.~

.//

t

~

SKAERGMRD (LIQUID)

x x

t&Si

!

x \ x

/

',

i

I

I

2

3

2

]

)

o

O. 1

1

2

FeO(t)/RgO

Fig. 3. Variations of SiO2, TiO2 and FeO(t) (total iron as FeO) contents with FeO(t)/MgO ratio in the sheeted dykes (crosses) and pillow lavas (dots) of Betts Cove ophiolite, Newfoundland. The rocks can be divided into three subgroups 1, 2 and 3 representing typical calc-alkalic, mildly callc-alkalicand tholeiitic fields respectively [ 1 ]. In the FeO(t) and TiO2 plots, only the average trend lines for each subgroup are shown. The arrows indicate the trends of fractionation. Open circle shows plot of the average of six pillow lava samples from Baie Verte ophiolite [6]. The area enclosed by broken line indicates the field of seven samples from the upper level of the pillow lava member of the Betts Cove ophiolite that are petrochemically different from the rest. CA = calc-alkaline and TH = tholeiite field.

the ophiolite so the magma type must have changed at this level. This tholeiitic trend was maintained during subsequent evolution since the overlying two non-ophiolitic pillow lavas of the Snooks Arm Group (Fig. 1, column 2) show remarkable petrochemical similarities to the seven tholeiitic samples. Such changes in magma types are also reported from the volcanic terranes o f the New Hebrides [41], Fiji [42]

and Troodos [36]. Such trends can be explained by a change from a water-undersaturated to a watersaturated environment, the latter leading to an increase in the degree of partial melting [36] or by a return from subarc-mantle to physiochemical conditions that are unaffected by subduction [44]. The implications of this are discussed in the last section o f the paper.

99 4.3. Trace element geochemistry

5. Discussion

Since most of the major elements concerned can undergo migration during metamorphism and weathering, Pearce and Cann [43,44] and Pearce [35] have proposed very effective and elaborate variation diagrams involving much less mobile elements such as Ti, Zr, Cr, Y and rare earths for distinguishing low-potash island arc tholeiites, calc-alkaline basalts and ocean floor basalts from one another. The trace element contents of the dykes and/or pillow lavas of Ballantrae, Othrys, Troodos and Newfoundland ophiolites are presented here in a series of such diagrams (Fig. 4). Ti-Zr plots of six volcanic rocks of the Ballantrae ophiolite, Scotland [46] also point to an island arc basaltic composition (Fig. 4I). Nearly all of the 17 mafic rocks from the Othrys ophiolite in Greece show an island arc to calc-alkaline composition on the Ti-Zr diagram (Fig. 4I). Likewise, the Lower Pillow Lavas of Troodos ophiolite are now recognized as island arc basalts in the Ti-Cr plots ([35]; Fig. 4J).

The terms "spreading ridge" and "island arc" for the two tectonic environments under discussion unfortunately and erroneously are often taken to imply mutually exclusive settings. Thus one does not generally consider the possibility of spreading near an island arc. Admittedly, there is very little doubt that the ophiolite suites of orogenic belts represent remnants of oceanic crust-mantle sequences and that most of them, especially those possessing sheeted dykes, were formed through lateral spreading. On the other hand field, stratigraphic and geochemical data on many well-developed ophiolites around the world consistently point to their close spatial association with island arcs. Accordingly, to reconcile the evidence, the most logical environments for the origin of such ophiolites would be marginal ocean basins which, as understood from modern analogues [8], evolve in the immediate vicinity of island arcs and widen through spreading along highly diffused zones. The crust of such marginal basins would be similar to

BAY OF ISLANDS

~IING~ c BIGHT

A

C

25, OOO[-

BETTS COVE E

LUSIIS P,IGrlT G

I

I ,l::l LJ~,t r a ¢ ' ( p l o t o f o sJml,les) O {~t h r y s • h-oodo~

20,O00Ii . . c:. . .c.a.l.c.- a. l.k., . . b. a. s. a l tIts s & d: o c e a ~ i f l o o r b a s a l t s Ti (ppm.)

BALLANTRAE-OTHRYS-TROODOS

15, (~OO~-

I

10,000~a | ~ B I,~~ e d ~

0 50 I00 ISO 200 z r (lq,,n.)

B

D

i0,000 e.mo•

Ti

L

H

J

.(,:.



(l'l'm.)

1,000

100 10

100

1,0DO0

L

J

Fig. 4. Trace element plots of sheeted dyke and pillow lava samples of the Bay of Island (A, B), Mings Bight (C, D), Betts Cove (E, F), Lush's Bight (G, H), Ballantrae (I), Othrys (I), and Troodos (I, J). The first four are taken from a compilation by Strong [36], except Betts Cove for which the data was obtained by the present authors. The plots for Othrys and Ballantrae have been taken from Pearce and Cann [43] and Wilkinson and Cann [46] respectively. The sources for the data on Troodos are Smewing et al. [45] (Ti-Zr) and Pearce [35] (Ti-Cr). The Ti-Zr and Ti Cr diagrams are after Pearce and Cann [43] and Pearce [35] respectively.

100 that of the major oceans in many respects. However, we disagree with Miyashiro's [1 ] suggestion that the Troodos ophiolite represents a volcano within an island arc. This and a score of other ophiolite suites probably originated adjacent to an island arc rather than in it. Thus, the newly created crust of a marginal basin would lie directly on the mantle rather than upon a remnant of another oceanic crust, a requirement [5] to be met if the ophiolites were to form in the island arc. As the marginal basin grew wider, the petrochemistry of its crust would gradually have changed in character from island arc to ocean floor basalt. This is substantiated by the fact that lavas and dykes dredged from marginal basins have chemical characteristics similar to basalts of major oceanic ridges [47]. Despite their narrow widths, the crustal rocks of the marginal basins could be compositionally more diverse than those of major oceanic crust because of factors such as the interaction between the downgoing oceanic plate and the adjacent active continental margin (e.g. development of rifts above the subduction zone), the possible role of sediments that are subducted into the trench, the greater heat retention caused by the insulating effect of the abnormally thick sedimentary cover(s) above the volcanic layer. The ophiolite suites would, therefore, comprise a wide range of rock types, the tectonic environments of which could seem to be characteristic of either major oceanic or marginal oceanic crust. For example, the variation from tholeiitic island arc basalts to tholeiitic ocean floor basalts is observed in the ophiolites of Troodos, Bay of Islands and Lush's Bight (Fig. 4); in the Bay of Islands "off-axis" alkali lavas are also present [48]. The Betts Cove ophiolite contains komatiitic [39] as well as calc-alkaline dykes and pillow lavas, the latter capped by tholeiitic ocean floor basalts at the upper stratigraphic level. Mixed volcanics including komatiites [6], calc-alkaline and tholeiitic lavas [37] are also known from the Mings Bight-Bale Verte ophiolite. The existence of komatiite-like rocks in Paleozoic ophiolite suites [6,39] and closely associated volcanic assemblages [49] could possibly be of great importance in evaluating the tectonic regimes of ophiolite genesis. We hedge only because quench textures have not yet been recognized in all the Paleozoic lavas of komatiitic chemical affinities. True komatiites are

confined to Archean terranes that have a closer affinity with island arc settings than mid-oceanic ridge environments [40]. Other suggested sites of komatiite genesis are grabens formed through rifting of continental blocks [50]. We contend that rift and island arc environments could be accommodated within a single continually evolving tectonic environment. The komatiite-bearing ophiolites probably originated through the distension of a continental edge thereby initiating the formation of a graben into which such high-Mg lavas were emplaced and, as the island arc began to build upon the break-away continental fragment (or the newly formed ocean floor?) and as the graben was gradually replaced by a marginal ocean basin, calc-alkaline, tholeiitic and other lavas began to form. We join Moores [3] in emphasizing that the petrochemical data, especially reliance on major elements alone, must be weighed against the structural and stratigraphic evidence in unravelling the tectonic setting of ophiolite genesis. The common occurrence of volcaniclastic rock (in some cases of andesitic composition) stratigraphically above the ophiolites (Fig. 1) is an extremely important factor since it implies that the source of the debris was most probably the nearby island arcs. The accumulation of thick volcaniclastic formations, punctuated by repetitive eruptions of pillow lavas as at Betts Cove (Snooks Arm Group), Lush's Bight (Western Arm Group) and Baie Verte, is much more plausible in an island arcmarginal basin setting than at a mid-oceanic ridge. The recurrence of lava eruptions was perhaps triggered by the intermittent subduction of the major ocean floor. The complex "reverse" setting of sheeted dykes above pillow lavas (e.g. Moretons Harbour, Newfoundland; Fig. 1, column 8) is perhaps another manifestation of the complex extrusive-intrusivepyroclastic activities taking place in the rather narrow marginal basins and adjacent island arcs. In summary, although we do not rule out the possibility that ophiolites could originate in more than one tectonic setting, the structural, stratigraphic and trace element data we have summarized show that the most logical site for the origin of many ophiolites around the world was in the Western Pacific-type marginal seas [7,8] bordered by island arcs on one side and continental margins on the other.

101 Acknowledgements We are grateful to W.R.A. Baragar, G.H. Gale, Edward Ghent, Roger Laurent, A k i h o Miyashiro, J.A. Pearce and Mavis S t o u t for helpful criticisms o f the first draft o f this manuscript and to Dr. G h e n t and t w o a n o n y m o u s reviewers for further guidance. Thanks are due to J a n e t Korbus, J u d y D o b r y m a n and Marion Benson for preparation of the manuscript. Chemical analyses o f Betts Cove rocks were kindly made b y Jaan Vahtra, Gertrude Andrews and David Press o f Memorial University o f N e w f o u n d l a n d , through the good offices o f David F. Strong. This w o r k has been supported by grants f r o m the National Research Council o f Canada (to E.R.W.N.) and f r o m Northeastern Illinois University's C o m m i t t e e on Organized Research (to H.D.U.).

References 1 A. Miyashiro, The Troodos ophiolitic complex was probably formed in an island arc, Earth Planet. Sci. Lett. 19 (1973) 218. 2 A. Miyashiro, Origin of Troodos and other ophiolites: a reply to Hynes, Earth Planet. Sci. Lett. 25 (1975) 217. 3 E.M. Moores, Discussion of "Origin of Troodos and their ophiolites: a reply to Hynes", by Akiho Miyashiro, Earth Planet. Sci. Lett. 25 (1975) 223. 4 P.J. Smith, Disagreements over Troodos, Nature 255 (1975) 192. 5 W.R. Church and R.A. Coish, Oceanic versus island arc origin of ophiolites, Earth Planet. Sci. Lett. 31 (1976) 8. 6 W.S.F. Kidd, The Bale Verte lineament, Newfoundland: ophiolite complex floor and mafic volcanic fill of small Ordovician marginal basin, in: Island Arcs, Deep Sea Trenches and Back-Arc Basins, M. Talwani and W. Pitman, eds. (American Geophysical Union, Washington, D.C., 1977) 407. 7 G.H. Packman and D.A. Falvey, An hypothesis for the formation of marginal seas in the western Pacific, Tectonophysics 11 (1971) 79. 8 D.E. Karig, Origin and development of marginal basins in the western Pacific, J. Geophys. Res. 76 (1971) 2542. 9 N. Rast, D.E. Rast and M.J. Kennedy, Plate tectonic significance of repeated tectonic trends in eastern Canada, Nature 258 (1975) 61. 10 I.W.D. Dalziel, M.J. de Wit and K.F. Palmer, Fossil marginal basin in the southern Andes, Nature 250 (1974) 291. 11 R. Laurent, Occurrences and origin of the ophiolites of southern Quebec, northern Appalachians, Can. J. Earth Sci. 12 (1975) 443. 12 W.H. Poole, J.R. Beland and R.K. Wanless, Minimum age

13

14

15

16

17

18

19

20

21

22

23

24

25 26

27

28 29

30

of Middle Ordovician rocks in southern Quebec, Geol. Soc. Am. Bull. 74 (1963) 1063. E.M. Moores, Petrology and structure of the Vourinos ophiolitic complex of northern Greece, Geol. Soc. Am., Spec. Paper 118 (1969) 74 pp. E.D. Jackson, H.W. Green and E.M. Moores, The Vourinos ophiolite, Greece: cyclic units of lineated cumulates overlying harzburgite tectonite, Geol. Soc. Am. Bull. 86 (1975) 390. T.P. Thayer, Some implications of sheeted dyke swarms in ophiolitic complexes, in: International Symposium on Ophiolites in the Earth's Crust, 1973 (Academii Nauk, Moscow) in press. T.P. Thayer and G.R. Himmelberg, Rock succession in the alpine-type mafic complex at Canyon Mountain, Oregon, 23rd Int. Geol. Gongr., Prague 1 (1968) 175. A.W. Snoke, A thrust plate of ophiolitic rocks in the Preston Peak area, Klamath Mountains, California, Geol. Soc. Am. Bull. 88 (1977) 1641. H.D. Upadhyay, J.F. Dewey and E.R.W. Neale, The Betts Cove ophiolite complex: Appalachian oceanic crust and mantle, Geol. Assoc. Can. Proc. 24 (197l) 27. B.E. Marten, Stratigraphy of volcanic rocks in the Western Arm area of the central Newfoundland Appalachians, Geol. Assoc. Can. Proc. 24 (1971) 73. D.F. Strong, Lush's Bight and Roberts Arm Groups of central Newfoundland: possible juxtaposed oceanic and island-arc volcanic suites, Geol. Soc. Am. Bull. 84 (1973) 3917. R. Lamarche, Ophiolites of southern Quebec, in: The Ancient Oceanic Lithosphere, Dep. Energy, Mines and Resources Publ. 42 (1972) 65. P. St. Julien and C. Hubert, Evolution of the Taconian orogen in the Quebec Appalachians, Am. J. Sci. 275A (1975) 337. R. Laurent and Y. H~bert, Features of submarine volcanism in ophiolites from the Quebec Appalachians, Geol. Assoc. Can., Spec. Paper 16 (1978). J.B. Hersey, Sedimentary basins of the Mediterranean Sea, in: Submarine Geology and Geophysics, Colston Papers, W.F. Whittard, ed. (Butterworths, London, 1965) 75. G.G. Shor, Structure of the Bering Sea and the Aleutian ridge, Mar. Geol. 1 (1964) 213. J.F. Dewey and J.M. Bird, Origin and emplacement of the ophiolite suite: Appalachian ophiolites in Newfoundland, J. Geophys. Res. 76 (1971) 3179. R.E. Norman and D.F. Strong, The geology and geochemistry of ophiolitic rocks exposed at Mings Bight, Newfoundland, Can. J. Earth Sci. 12 (1975) 777. A.H.G. Mitchell, Facies of an early Miocene arc, Malekula island, New Hebrides, Sedimentology 14 (1970) 201. A. Ewart and W.B. Bryan, Petrography and geochemistry of the igneous rocks from Eua, Tonga Islands, Geol. Soc. Am. Bull. 83 (1972) 3281. B.F. Kean and D.F. Strong, Geochemical evolution of an Ordovician island arc of the central Newfoundland Appa-

102 lachians, Am. J. Sci. 275 (1975) 98. 31 D.F. Strong and J.G. Payne, Early Paleozoic volcanism and metamorphism of the Moreton's Harbour area, Newfoundland, Can. J. Earth Sci. 10 (1973) 1363. 32 George Einarsson, personal communication, 1977. 33 E.M. Moores and F.J. Vine, The Troodos Massif Cyprus, and other ophiolites as oceanic crust: evaluations and implications, Philos. Trans. R. Soc. London, Ser. A, 268 (1971) 443. 34 H.L. Davies, Peridotite-gabbro-basalt complex in eastern Papua: an overtbrust plate of oceanic mantle and crust, Bur. Miner. Resour., Geol. Geophys. Aust., Bull. 128 (1971) 48 pp. 35 J.A. Pearce, Basalt geochemistry used to investigate past tectonic environments on Cyprus, Tectonophysics 25 (1975) 41. 36 D.F. Strong, Volcanic regimes of the Newfoundland Appalachians, Geol. Assoc. Can., Spec. Paper 16 (1977) 61. 37 T.N. lrvine and W.R.A. Baragar, A guide to the chemical classification of the common volcanic rocks, Can. J. Earth Sci. 8 (1971) 513. 38 M.G. Best, Differentiation of calc-alkaline magmas, Oreg., Dep. Geol. Miner. Ind., Bull. 65 (1969). 39 H.D. Upadhyay, Phanerozoic peridotitic and pyroxenitic komatiites from Newfoundland, Science (1979) in press. 40 C.K. Brooks and S.R. Hart, On the significance of komatiite, Geology 2 (1974) 107. 41 A.H.G. Mitchell and A.J. Warden, Geological evolution of New Hebrides island arc, J. Geol. Soc. London 127 (1971) 501.

42 J.B. Gill, From island arc to oceanic islands: Fiji, southwestern Pacific, Geology 4 (1976) 123. 43 J.A. Pearce and J.R. Cann, Ophiolite origin investigated by discriminant analysis using Ti, Zr, and Y, Earth Planet. Sci. Lett. 12 (1971) 339. 44 J.A. Pearce and J.R. Cann, Tectonic setting of basic volcanic rocks determined using trace element analyses, Earth Planet. Sci. Lett. 19 (1973) 290. 45 J.D. Smewing, K.O. Simonian and I.G. Gass, Metabasalts from the Troodos Massif, Cyprus: genetic implication deduced from petrography and trace element geochemistry, Contrib. Mineral. Petrol. 51 (1975)49. 46 J.M. Wilkinson and J.R. Cann, Trace element and tectonic relationships of basaltic rocks in the Ballantrae igneous complex, Ayrshire, Geol. Mag. 111 (1974) 35. 47 S.R. Hart, W.E. Glassley and D.E. Karig, Basalts and seafloor spreading behind the Mariana island arc, Earth Planet. Sci. Lett. 15 (1972) 12. 48 D.F. Strong, An off-axis alkali volcanic suite associated with the Bay of Islands ophiolites, Newfoundland, Earth Planet. Sci. Lett. 21 (1974) 301. 49 G.H. Gale, Paleozoic basaltic komatiite and ocean-floor basalts from northeastern Newfoundland, Earth Planet. Sci. Lett. 18 (1973) 22. 50 R.W. Nesbitt and S. Sun, Geochemistry of Archean spinifex-textured peridotites and magnesian and lowmagnesian tholeiites, Earth Planet. Sci. Lett. 31 (1976) 433.