The origin of ultramafic and ultrabasic rocks

Tectonophysics - Elsevier Publishing Company, Amsterdam Printed in The Netherlands

THE: ORIGIN OF uLTHAMAFIC

AND ULTRABASIC HOCKS

PETER J. \VYLLIE Department (Received

of Geophysical March 6,

Sciences,

The University

of Chicago,

Chicago,

111. (U.S.A.)

1969)

SUMMARY

Ultramafic rocks are classified in terms of their field associations and tectonic environment. Eleven associations are distinguished, some with subdivisions. The major features of each association are described in turn, and it is emphasized that the variety of field and petrographic associations indicates that a variety of processes is involved in the origin and emplacement of these rocks. For petrogenetic discussion the-associations are considered in three groups: (1) layered, stratiform and other intrusions involving gabbro or diabase together with accumulations or concentrations of mafic minerals; (2) the alkalic rocks of stable continental regions; including kimberlites, mica peridotite, members of ring complexes, and ultrabasic lava flows; (3)the several peridotite-serpentinite associations of the erogenic belts that have been classified together as alpine-type intrusions. The petrogenesis of ultramafic nodules in alkali basalts and kimberlites is also discussed. The petrogenesis of these rocks is concerned with the source of the material, its variation in physical state and temperature from source to position of intrusion, and its post-intrusion history. The study of ultramafic rocks and nodules is one approach towards determination of the composition and mineralogy of the upper mantle, INTRODUCTION

The terms “ultramafic” and “ultrabasic” relate, respectively, to mineralogical and chemical classifications of rocks, but both have been used rather loosely. Ultramafic rocks are those with color indices of more than 70, and ultrabasic rocks are those containing less than 45% SiOz (Williams et al., 1954). Most ultramafic rocks are also ultrabasic and most ultrabasic rocks are also ultramafic, but there are exceptions, and therefore both terms should be retained. Estimates of the composition and mineralogy of the upper mantle have been based on information from extraterrestrial sources and from terrestrial sources. Meteorite analogies and the application of other extraterrestrial information to the origin of the earth and the mantle have been discussed recently by Ringwood (1966a,) and by Harris et al. (1967). The extraterrestrial approaches lead to the conclusion that the mantle is composed Tectonophysics,

7 (5-6)

(1969) 437-455

437

of ultrabasic material, and the physical properties of the mantle, dotermined by geophysical techniques, are compatible with this ~~)I~~~us~oI~. Terrestrial approaches to the problem are thus concerned with investigation of ultramafic rocks exposed at the surface of the earth. Ultramafic rocks could possibly have been derived directly from the mantle, or indirectly from other mantle-derived material such as basaltic magma. The composition of basaltic magma also places limits on the composition oi the ultrabasic material of the mantle, assuming that the magma is derived by fractional fusion of the mantle. The petrological problem, therefore, is to determine which ultramafic rocks, if any, represent mantle material, and to deduce the processes which led to their emplacement into the crust, References in this paper are made frequently to review articles in a recent book on ultramafic rocks (Wyllie, 1967a) which provides a more comprehensive bibliography.

ULTKAIWAFIC

AND ULTRABASIC

ROCK ASSOCIATIONS

There are many different rock types in the ultramafic clan and these occur in a variety of field and petrographic associations. We may expect, therefore, that there is a variety of processes involved in their origin and emplacement. Although interpretation of processes ma2 require extensive study, rock associations are more easily distinguished; but the distinctions have not always been made. Hess (1955, p.394) drew attention to this with the following statement: “Gross errors have probably resulted from applying conclusions drawn from facts relafed to mica peridotites to alpine peridotites and vice versa. If mica peridotite had been called humptydumptyite, these probably would not have arisen”. This section presents an outline classification of ultramafic and ultrabasic rock associations as a Framework for petrogenetic discussion. Eleven associations are listed. There is overlap between some of the associations, and gradation between some of the subdivisions. This working classification will need modification as we learn more about the geology of the rocks.

This association has been widely recognized, and most petrologists agree that the layered ultramafic rocks of the association developed in floored chambers as superposed accumulations of crystals formed during the fractional crystallization of a parent gabbroic magma of tholeiitic type. A recent symposium volume (Fisher et al., 1963) described several layered intrusions, and Turner and Verhoogen (1960) reviewed the petrology of the association, There appear to be two extreme types, but gradational types may exist: (a) The large differentiated gabbroic lopoliths or stratiform intrusions which may be thousands of square miles in area and just a few miles thick. These occur in regions not affected by contemporaneous orogeny (Jackson, 1967; Irvine and Smith, 1967). (b) The small funnel-shaped intrusions occurring in a volcanic association, of which the type example is the Skaergaard intrusion described in detail by Wager and Deer (1939) and Wager and Brown (1968).

Group 2: ultramafic

rocks

ill differentiated

basic

sills

and in minor

intrusions

Concentration of mafic minerals in basic sills, dykes, sheets and laccoliths may produce ultramafic rocks. Two associations are recognized, and their characteristics have been reviewed by Drever and Johnston (1967a): (a) The alkalic diabase (or teschenite, or theralite)-picrite association. (b) The tholeiitic diabase-picrite association. A third group of rocks may merit separate classification as: (c) Picritic minor intrusions. These are olivine-rich rocks of tholeiitic type which have characteristics distinguishing them from picrites of differentiated sills (Drever and Johnston, 1967b). Group 3: concentrically

zo?zed dllllite-pevidotite

associatio?z

Among the ultramafic rock types occurring in erogenic belts are the small cylindrical bodies (half a mile or so in diameter) recently recognized in Alaska (Noble and Taylor, 1960; Taylor and Noble, 1960). They have features distinguishing them from the Alpine type of ultramafic intrusion, as discussed by Thayer (1960, 1967) and Taylor (1967a). Most of the bodies are hornblende pyroxenites or hornblendites, but if they contain dunites a range of ultramafic rock types forms a crude zonal arrangement. A core of dunite is surrounded by successive shells of peridotite, olivine pyroxenite, magnetite pyroxenite, and hornblende pyroxenite. Gravitational layering is present in some of the complexes indicating the operation of crystal fractionation (Irvine, 1967). However, the rocks are readily distinguished from the stratiform peridotite-gabbro complexes by their structure, mineralogy and chemistry. The ultramafic complexes are usually emplaced in gabbro that has been strongly saussuritized near the ultramafic contacts. GYO@ 4: Alpine-type

/wvidotite-se@e?ztbzite

association

This association includes large and small bodies distributed along deformed mountain chains and island arcs, usually along with gabbros or basic volcanic rocks. Many of the rocks have had a complex history, with the early stages being obscured by deformation and metamorphism. They have characteristics distinguishing them from other ultramafic associations occurring in erogenic zones (Benson, 1926; Thayer, 1960, 1967). The different times of emplacement of Alpine-type intrusions (preorogenic, synorogenic and postorogenic) produce rocks of distinct mineral facies which provides a basis for their classification (O’Hara, 1967a). Tectonic migration of serpentinite masses along fault zones can lead eventually to surface extrusion of serpentinite breccias by plastic flow (Dickinson, 1966). The following groups of intrusions are distinguished mainly on the basis of their present environment: (a) Serpentinites and peridotites occurring in strongly deformed but peripheral low grage metamorphic zones, with greenschists, glaucophane schists, and even unmetamorphosed greywackes and lavas (De Roever, 1957). This includes the ophiolite association discussed by Miyashiro (1966), Thayer (1967) and S0rensen (1967). Tectonophysics,

7 (5-6)

(1969) 437-455

439

(b) Serpentinites and peridotites occurring in regionally metamorphosed rocks of greenschist, epidote-amphibolite, and amphibolitc facies. An example is the Apalachian ultramafic belt discussed by Jahns (1967) and Lapham (1967). (c) Peridotites and garnet peridotites associated with regionally metamorphosed rocks of upper amphibofite, granulite or eclogite facies, which are apparently restricted to rocks of Europe metamorphased during the Caledonian orogeny (O’Hara, 196713; Lappin, 1967). (d) A distinctive group of high-temperature peridotite intrusions, of intermediate size (5-15 miles in diameter), with well defined dynamotherma1 metamorphic aureoles. These have been reviewed by Green (1967) and discussed by Thayer (1967). (e) Some metamorphic and metasomatic rocks in group 11 are appropriately considered in the Alpine association (Sgrensen, 1967).

The main crustal layer beneath the oceans may be serpentinized peridotite, like that dredged from fault scarps on the Mid-Atlantic Ridge, from the Puerto Rico Trench, and collected from St. Peter’s and St. Paul’s Rocks on the Mid-Atlantic Ridge. Hess has reviewed this association in the volume edited by Burk (1964) which includes studies on the serpentinite cores obtained from the A.M.S.O.C. drill hole near Mayaguez, Puerto Rico. Dietz (1963) suggested that tectonic emplacement of oceanic serpentinites into overlying deformed sediments of a continental rise would produce Alpine-type intrusions (group 4).

The “granitic” batholiths include rock types ranging in composition from ultrabasic to acid. There are at least three associations including ultrabasic rocks, In the first two, basic and ultrabasic members are usually concentrated as masses within the stocks forming satellites to the major complex: (a) Peridotite-pyroxenite-hornblendite-tonalite-granodiorite-granite. (bf Hornblendite-diorite-monzonite-syenite, the hornblende-rich varieties making up the appinite suite. (c) Ultrabasic lamprophyre dikes occurring as members of dike swarms associated with granitic or granodioritic complexes, and often converging on individual intrusive centres. Joplin (1959) reviewed the occurrence and origin of some of these basic and ultrabasic bodies. Ultrabasic lamprophyres also occur in association with alkalic ring complexes, group 7c, and dikes or sheets of alnoite (and other lamprophyres) may be distributed on a regional scale; compare group 8b.

Alkalic

ring complexes

containing

ultrabasic

rocks

are usually

circular

or elliptical in plan, with evidence for cylindrical or funnel shape in vertical section, They are usually a few miles in diameter, although they range from small diatremes to masses many miles across. The rock types may be arranged in concentric fashion, and many of them exhibit gravity layering. The complexes occur in stable or fractured Continental regions, following linear trends, and they are often directly associated with alkalic volcanic provinces. Three series of rocks encountered in these complexes include ultrabasic rocks: (a) The subsiliceous alkalic series containing normative but not necessarily modal nepheline: peridotite-cyroxenite-biotite-gabbro-dioritesyenite; (b) The peralkalie series including pyroxenites, biotitites, rocks of the ijolite family, melilite-rich rocks of the uncompahgrite family, and carbonatites;, (c) Alnoite and kimberlite dykes and sheets associated with some of the carbonatite complexes in group b. Upton (1967) and Gold (1967) reviewed the rocks of this association, and Von Eckermann (1967) discussed the rocks in group c. The place of carbonatites in this association has been described in recent books by Tuttle and Gittins (1966), and Heinrich (1966).

GrouP 8: kimberlites Kimberlites are rare rocks occurring in small diatremes, dikes, veins and sills in stable or fractured continental regions. Their distribution appears to be related to deep-seated tectonics with linear trends. Davidson (1967a) drew attention to the geographic association of kimberlites with alkalic ultrabasic complexes (group 7) in the petrographic province of the Maimecha-Kotui district in Russia. Three associations may be considered: (a) Clusters of pipes or diatremes, which tend to form elongate chains. These include the diamond-bearing kimberlites. Dawson (1967a) reviewed the geology and world distribution of kimberlites, and Davidson (1967a) reviewed kimberlites in Russia. (b) Dikes and sheets of kimberlite (mica-peridotite) or alnoite distributed on a regional scale. The suite in eastern North America extending along the stable area to the west of the Appalachian erogenic belt was described by Watson (1967). (c) Dikes and sheets of alnoite and kimberlite associated with carbonatite complexes. Dawson (1967a) referred to these as “central complex kimberlites”, and they were listed previously under group 7c. Grotlp 9: ultrabasic

lams

Ultrabasic lavas are listed apart from their plutonic equivalents and associates because the existence and nature of ultrabasic lavas is of critical importance in petrogenesis (cf., Bowen, 1928): (a) Ultrabasic or ultramafic lavas can be formed by enrichment of basic magma in mafic crystals, either by crystal settling before or after extrusion, or by flow differentiation during extrusion. Richter and Murata Tectonophysics, 7 (5-6) (1969) 437-455

441

(1966) discussed olivine collcentl.atioI~ in the iavas of Hawaii, Simkin (1967 j mentioned flows with features su ggesting flow diiiereilti~~ti~j~~, and Mathews et al. (1964) showed that gravity concentration of olivine can occur in basaltic pillows. The ultrabasic lavas of Cyprus described by Gass (1958) are believed to represent olivine-enriched basic magma. Rocks of the ophiulite suite (group 4a) may represent submarine lava flows cnrichcd in maiic minerals. (b) Alkalic ultrabasic lavas occur in the same cratogenic environment as the alkaline ultrabasic complexes of group 7. In his review oi the sodic, alkalic igneous suites of eastern Uganda, King (1965) noted that the ultrabasic lavas in the mclanephelirlite-17ephclinite series correspond to the plutonic series pyroxenite-ijolite family-carbonatite oi group 7%. He proposed that the alkali carbonatite lava flows observed by DaWSon (1962) on Oldoinyo Lengai reprcscnted the final residuum, (*orres~~jildi~~~ to the carbonatites of the plutonic series. Potassic ultrabasic lavas ~(~I~taijiit~g leucitc, melilitc, and rarely kalsilitc, are represented by the uganditemafurite-katungitc series of the Tore-Ankole volcanic province oi Uganda, whose petrology and petrogenesis werc~ discussed by Holmes (1950). The Congo volcano of Mount Nylragongo in the Virunga volcanic field, is composed of both sodic and potassic ultrabasic lavas containing various proportions of nepheline, leucite, and melilitt: along with their mafic minerals. Sahama (1962) revieWed the petrology (Ii this cc~mplcs volcano. (c)An effusive equivalent oi kimbcrlitc:. m(:imet’hitc, is known in the U.S.S.R. both in lava-form and as a tuii, or lava-trreccia. Meimechites were described briefly by Dawson (196 7a) and by Davidson (1967a).

Lavas and diatremes contain a variety of inclusions or nodules, including ultrabasic types, which may be cognate (cumulative) or accidental (exotic) xenoliths. These ultrabasic xenoliths occur in iour environments: (a) Many alkali olivine-basalts contain nodules of peridotite (spinellherzolitc), garnet peridotite, and eclogite but only three examples are known of inclusions in tholeiitic basalts. Forbes and Kuno (1967) reviewed the world-wide distribution and composition of the inclusions and their host basalts, and Harris et al. (1967) considered their average composition and source. (b) Alkalic basic and ultrabasic lavas of group 9b contain inclusions of peridotite, pyroxenite, biotite-pyroxenite, and biotite-horilblelldite (Upton, 1967). (c) Kimberlite diatremes are crowded with xenoliths, including peridotites, garnet pyroxenites, and eclogites. Scapolite may occur in the xenoliths, Recent accounts of these xenoliths have been presented by Davidson (196’7b), and Harris et al. (1967). (d) Ultrabasic or ultramafic inclusions alSO occur in peridotites in minor and major intrusions. Drever and Johnston (196713) described peridotite xenoliths in the picrite dikes of Skye. Irvin [1967) described inclusions of peridotite in the layered ultrabasic rocks of Duke Island (group 3). The processes involved in the formation and location of the inclusions in these intrusions are almost certainly of local significance.

Group 11: metamorphic

and metasomatic

ultrabasic

and ultramafic

rocks

Any of the preceding types of ultrabasic or ultramafic rocks could become involved in regional metamorphism, Some of the Alpine-type intrusions have had a complex history, and their origin may be obscured by deformation and recrystallization. Other ultratiafic rocks in schists, amphibolites, and hornblende gneisses of the erogenic zones may be formed by metamorphic differentiation or by metasomatic processes, as described by Splrensen (1967).

PETROGENESIS

OF THE ROCK

ASSOCIATIONS

Important factors to be considered in petrogenesis include the source of material, its variation in physical state and temperature between its source and its position of intrusion, and its post-intrusion history. It is generally agreed that the source of ultramafic rocks is the upper mantle. The only ultramafic rocks composed of crustal material are those formed by metasomatic processes. Possible physical states of mantle-derived material, and the processes operating during their uprise into and through the crust, are summarized in Table I. The more complex the post-intrusion history of an ultramafic rock, the more difficult is its petrogenetic interpretation. At one extreme are the stratiform gabbro-peridotite complexes that crystallized slowly under stable conditions and were not subsequently metamorphosed. At the other extreme are those peridotites of the Alpine suite that have been subjected to several episodes of deformation and metamorphism, serpentinization, and tectonic transportation. Metasomatism or explosive activity arising from the late concentration of alkalis and volatiles in alkalic ultramafic complexes also tends to destroy the evidence of the conditions during and immediately after emplacement. For petrogenetic discussions, it is convenient to consider three groups of associations, and to examine them in the following sequence: (1) Layered, stratiform and other intrusions involving gabbro or diabase together with crystal cumulates or concentrations; (2) the alkalic rocks including kimberlites, mica peridotites, members of ring complexes, and ultrabasic lava flows; (3) the several peridotite-serpentinite associations of the erogenic belts, grouped together as Alpine-type intrusions. Most members of group (2) are distinguished from the other groups by their chemistry, mineralogy, and tectonic setting, Criteria for distinguishing group (1) from group (3) have been proposed by Hess (1938, 1955) and Thayer (1960, 1967). However, recent literature indicates that several associations occur in the Alpine tectonic environment, and some Alpine-type intrusions are not readily distinguished from layered rocks of group (1). PETROGENESIS

OF STRATIFORM

INTRUSIONS

AND DIFFERENTIATED

BASIC

SILLS

There is general agreement that the parent magma of the stratiform Tectonophysics, 7 (5-6)

(1969)

437-.&Z

4-I::

intrusions in group (1) is tholeiitic basaltic magma derived by partial fusiorl of the upper mantle, and that gravity settling of early-formed crystals is the dominant process (Jackson, 1967). Irvine and Smith (1967) concluded that gravitational settling of mineral grains in plutonic basaltic magma is almost inevitable. Other processes cause the cyclic features of the gravity layering. Periodic convection of a single Cell appears satisfactory for the Skaergaard intrusion, but in larger bodies such as the Bushveld intrusion a complex system of convection cells could exist. For the Muskox intrusion, Irvine and Smith (1967) concluded that the cyclic units were produced by repeated intrusions of fresh basaltic liquid flowing through the whole width of the magma chamber, the “old” magma being pushed on and erupted through volcanic fissures. Jackson (1967) emphasized that similarities between layered intrusions had been obscured by differences in terminology used by different authors. He offered a standard nomenclature based on sedimentary terminology adapted for description of magmatic sediments. The ultramafic rocks produced in basic sills, dykes, sheets and laccoliths by concentration of mafic minerals provide striking examples of differentiation, and they have been widely cited as examples rii gravity settling of crystals. More detailed work oi recent years indicates that this simple explanation is inadequate (Drever and Johnston, 1967a). It now appears that gravity effects are superimposed on the dominant process of flow differentiation of basaltic liquid containing suspended olivine crystals, possibly in tholeiitic sills as well as in those with alkalis afiinities (Simkin, 1967; Bhattacharji, 1967). The crystals could have been derived dirrctly from the mantle, or from an olivine cumulate formed by gravity difierentiation during a halt in a deeper reservoir. The crystals trould also have been precipitated during ascent of picritic or basaltic liquid. Since Bowen’s (1928) study, the small picrite and peridotite sills and dykes of Skye have been considered as examples of the intrusion of a crystal mush lubricated by basaltic liquid, but Drcver and Johnston (1967b) concluded that these and similar rocks in Greenland were emplaced as liquid picritic magmas containing suspended olivine crystals. The experimental studies oi Ito and Kennedy (1967) and Green and Ringwood (1967) indicate that liquid picritic magmas should form in the upper mantle, although special conditions would be required for these to reach near-surface conditions without precipitating their olivine.

Some of the direct and indirect petrogenetic links between kimberlites, alkalic ultrabasic rocks, and carbonatites (Von Eckermann, 1967) are as follows: the alkalic ultrabasic rocks of ring complexes occur in the same tectonic environment as kimberlites, and both are related to deep-seated fracture systems; the ring complexes often contain carbonatite intrusions; carbonatites are often associated with “central-complex kimberlites”; many kimberlites contain much carbonate which may be primary. Parent magmas proposed for the ultrabasic rocks of group (2) include undersaturated alkalic basalts and alkalic ultrabasic liquid magmas, evidence for the latter including the existence of alkalic ultrabasic lavas. Although the upper mantle is usually considered to be the source of these magmas, 444

‘fcctonophysics,

i (5-G) (ltr(it)) 43-456

contamination with crustal material is often invoked to explain their unusual chemistry (Turner and Verhoogen, 1960, pp.249, 396). Reaction of a primary earbonatite magma from the mantle with “granitic” crustal rocks to form some alkalic ultrabasic rocks including kimberlites, has also been proposed but not generally accepted (Holmes, 1950; Dawson, 1967b). The processes involved in the formation of the gabbro-peridotite associations of group (1) operate also in these plutonic alkalic associations, with gravity settling of crystals and convective circulation being dominant in the early stages of crystallization (Upton, 1967). Now that flow differentiation has been recognized as an important igneous process, its effects will probably be discovered in the alkalic associations. The petrogenesis of the alkalic ultrabasic rocks is further complicated by additional processes involving the concentration of alkalis and volatiles in residual liquids. Retention of these components facilitates differentiation in the liquid phase by diffusion of materials along temperature and concentration gradients; produces extreme differentiates including the volatile-rich, low-silica ultrabasic rocks such as carbonatite and okaite (Gold, 1967); and leads to metasomatism (fenitization) and recrystallization that may obliterate previous textures as well as forming metasomatic ultrabasic rocks (Upton, 1967). Explosive exsolution of volatile components from basic or ultrabasic magmas causes eruption of fragmental volcanic rocks with complementary crystalline residues remaining in some volcanic conduits. It also causes brecciation of previously crystallized rocks and the formation of diatremes. These effects of volatile components are well illustrated by kimberlites: the deep-seated massive rock crystallizes from an ultrabasic magma, and higher level rocks are emplaced in diametres as fluidized solid-gas systems, xenocrysts and xenoliths of waI1 rocks being mixed with the fragmented and usually greatly altered kimberlite (Dawson, 1967a). Some kimberlites and mica peridotites may be emplaced as crystal aggregates transported by a carbonatite magma and gases at temperatures of 600*-700°C (Watson, 1967; Franz and Wyllie, 1967). Dawson (1967b) reviewed the origin of kimberlites. I’ETROGENESISOF THE PERIDOTITE-SERPENTINITEASSOCIATIONOF ALPINE OROGENICBELTS The petrogenesis of these rocks is more complex because it involves metamorphic processes; indeed, Den Tex (1965) discussed them in terms of their metamorphic lineages rather than their igneous origin. The problems of interpretation introduced by metamorphism, deformation, and reintrusion have contributed to what Hess (1955, p.391) termed a “magnificent argument” between the field geologists, who concluded that many alpine-type ultramafic bodies were emplaced as fluid ultrabasic magmas, and the experimental petrologists, who denied that such materials could be liquid magmas at any reasonable temperatures. The recent history of this argument in the light of new experimental data and changing field interpretations has been described by Wyllie (1967b), and its resolution in favor of solid emplacement was discussed by Hess (1966, p.5-6). The hypothesis of solid intrusion for the emplacement of ultramafic rocks in Alpine belts is now widely accepted (De Roever, 1957; Ragan, 1967;

Jahns, 1967), and the methods of structural petrology have helped to unravei the processes involved in their emplacement by correlation of deformation patterns in the ultramafic rocks with the regional erogenic history (Lapham, 196’7; Lappin, 1967). Recent experiments on deformation have provided valuable information on the mechanism of plastic flow of periciotite minerals (Raleigh, 1967), and the experimentally observed weakening and embrittlement of serpentinite at dehydration provides a satisfactory explanation for the field conclusion that many serpentinites were emplaced by block tectonic transport. The controversy concerning ultrabasic magmas tended to ignore the gabbros and basic volcanic rocks almost invariably associated with the peridotites and serpentinites, or to relate them to a different stage of the erogenic and magmatic activity. Thayer (1960, 1967) discussed and deplored the conceptual divorce of the ultramafic and mafic rocks, and he listed six criteria that characterize the intrusive peridotite-gabbro complexes of alpine type. Reconsideration of the petrogenesis of ultramafic and mafic rocks together (Miyashiro, 1966) has contributed to the formulation or revival of several hypotheses. Current petrogenetic hypotheses are Summarized in Table I. Emplacement temperatures of the rocks at successive stages of formation and transportation are considered in four ranges, with column widths approximately in scale to the temperature intervals. The hypotheses are arranged in four groups. Varying degrees of fusion of the mantle may produce basic liquids, or a range of ultrabasic liquid compositions. The hypotheses involving ultrabasic liquid magmas form group 1. Hypotheses involving the formation of ultramafic rocks by concentration of mafic minerals from a basic liquid form group 2. Group 3 includes the hypotheses stating that ultramafic rocks, or peridotite-gabbro associations, constitute representative samples of a portion of the mantle. The processes involved in stage III represent deformationandmetamorphism of the ultramafic rocks produced in stages I and II. Hypothesis 4 refers to the ultramafic rocks that have become an integral part of a terrane of regional metamorphism.

The zoned ultramafic complexes of Alaska constitute a distinctive group of very high temperature intrusions with pronounced metamorphic and metasomatic effects at their margins. Taylor’s (1967a) review concluded that they were formed by the successive intrusion of liquid ultrabasic magmas of different compositions, all with high contents of FeO, CaO, and Hz0 (see Presnall, 1966). This is hypothesis 1A. Onuki (1965, 1966) described ultramafic rocks from Japan which are petrographically similar to the Alaskan rocks, and he concluded that intrusion and differentiation of a parent ultrabasic liquid magma produced the “ultramafic rock series” formed from essentially liquid magmas, crystal mushes, and residual basic magmas. This is hypothesis 1B.

Tectonophysics, 7 (5-6) (1969) 437-435

(1965)

Den Tes

Dietz (1%:;)

(1960)

&?SS

(1967)

Green (1967) Thayer

Miyashiro (1966)

Challis (1.965a)

(1967)

Green

from

man&F

--..--_

fusion

of current

If ==high temperatures solidus and serpentine

flow,

I II III

within crystallization interval stability; L =. 10% temperatures

of anhydrousbasalt; in serpentine stability

I II III

I II III erosion exposing mantle

I II III

I;

I II III

\

~~~.~~~~~~~~~

tectonic block transportation, and solid

~erpe~~~~~~~tion, I

plastic

III

Stage

plastic t’lo~ of layered complex

submarine extrusion and , differentiation

removal of basic magma above cumulate

reintrusion of cumulate mush

and tectonic emplacement in overlying sediments

uplift

depression heating of serpentinite

tectonic

and serpentinites

downward movement convection

?

mush intrusion with flow layering

-_--_

reintrusion of crystal mushes

II

stages

differentiation in magma chambers (?)

differentiation forming stratiform intrusions

@XWity

differentiation of parent magma

intrusions

SUW.?t?klSiV~

.~.in mantle Successive ------I

mobilization by partial fusion

partial

of peridotites

mantle pcridatitc or serpentinized layer :I of oeeanjc crust

regionaL metamorphism

representative mantle sample

mush with basic liquid

peridotite

basic iiquid magma, \vith OS without suspended crystals

rocks Process

complete (?\ or partial fusion

of Alpine ultramafic

ultrabasic liquid magma, with or without suspended crystals

Material

for the origin

IVH -= very high tetnpcraturi~s abo\re basalt liquidus; iM = medium temperatures bet\vccn anhydrous basalt

~-

C

I3

A

c

2 B

A

(1965)

Onuki

A

B

Taylor (1967a)

1

advocate

number

3

4

hypotheses

Hypothesis

Current

TABLE I

field.

VHHML

--.* .

Temperature ranges1

Hypotheses

of gvozil,

2,

Table I

Green (1967) recently reviewed the distinctive group of peridotites with high temperature aureoles, and he considered their source to be either deep-seatedcumulate or the upper mantle, and their mode of intrusion to be as a crystal mush followed by diapiric flow after complete crystallization. The alternatives correspond to hypothesis 2A, or to hypothesis 3A (with gabbro removed). Thayer (1967) remains unconvinced that the intrusions were unusually hot when intruded. O’Hara (1967~) has developed a I’- T grid based on compositions of coexisting pyroxenes, and representatives of these rocks plotted on the grid give estimated conditions of equilibration at temperatures usually greater than l,lOO°C, and at pressures corresponding to depths ranging from near-surface to 50 or 60 km; the conditions correspond closely to the crystallization interval of anhydrous basaltic liquid (Green and Ringwood, 196’7; Cohen et al., 1967). Challis (1965a, b) reported pyroxene granulite facies metamorphism at the unfaulted contact of the Red Hills ultramafic intrusion of New Zealand, which has long been classified as a typical low-temperature Alpine-type intrusion, and she interpreted the ultramafic rocks in terms of hypothesis 2B. Several other Alpine peridotites have recently been interpreted as gravity-stratified crystal cumulates, following hypotheses 2A or 2B (O’Hara, 1967b, c). Smith (1958) suggested that there is a continuous series of peridotite-gabbro associations between the stratiform types and the Alpine types, Several recent proposals have connected volcanic activity with stratiform and other gravity-layered intrusions (Brown, 1956; Challis, 1965a; Irvine and Smith, 1967; Irvine, 1967; Taylor, 1967a). Comagmatic lava flows could thus be associated with Alpine-type ultramafic rocks formed from ultrabasic magmas (hypothesis l), or from crystal concentrates derived from intrusive basic magmas or stratiform intrusions (hypotheses 2A and 2B). The ophiolite hypothesis, 2C, interprets the whole assemblage as massive, differentiated submarine lava flows ranging in composition from basic to ultrabasic, together with some intrusive rocks. If liquid ultrabasic flows have occurred, then there should be a hypothesis lC, directly comparable with 2C. There appears to be a revival of interest in the hypothesis (Maxwell and Azzaroli, 1963; Moores et al., 1966; Miyashiro, 1966), although Thayer (1967) presented reasons for concluding that ophiolite complexes are essentially identical with normal, intrusive Alpine complexes.

Thayer (1960, 1967) discussed evidence supporting hypothesis 3A, with the anhydrous crystal mushes from the upper mantle (perhaps already gravity stratified), developing flow layering, and entraining masses of chromitite; associated quartz diorite and granophyre complete the calcalkaline “Alpine mafic magma stem”. According to Table I, the only Alpine peridotites not genetically related to gabbroic rocks are the peridotites derived directly from the mantle (De Roever, 1957) in the solid state, or from partly serpentinized rocks forming layer 3 of the oceanic crust (Hess, 1962). Hess (1960, p.235) suggested that the serpentinites of Puerto Rico represented uplifted oceanic crust, which “may be altered mantle -44s

Tectonophysics,

7 (k-6)

(1969)

4:1’7-453

rocks exposed at the surface”, This is hypothesis 3B. A spreading ocean floor could cause tectonic incorporation of layer 3 serpentinite within sediments of the continental rise (Dietz, 1963). This is hypothesis 3C. Hypotlzeses of group 4, Table I O’Hara’s (1967c) pyroxene grid indicates that Alpine ultramafic rocks have equilibrated under a wide range of conditions, and that some high temperature ultramafic rocks have been re-intruded upwards through several kilometres. Hypothesis 4 implies that a peridotite of any origin becomes an alpine type if it is involved in regional metamorphism with the associated processes of deformation being represented by stage III, and O’Hara (196’7c) suggested that the only feature these rocks have in common is an Alpine setting that causes tectonic transport and re-intrusion. Sgrensen (1967) reviewed evidence that peridotite and garnet peridotite, when enclosed in amphibolite, are the products of metamorphic differentiation in zones of stress concentration.

THE PETROGENESISOF ULTRAMAFICNODULES Current interest in the nodules of garnet peridotite, peridotite, and eclogite enclosed in alkali basalts and kimberlites lies in the hope that they will provide information about the composition and mineralogy of the earth’s mantle. The nodules have been interpreted as representative of primary mantle material; as residual mantle material after extraction of basalt; as xenoliths picked up within the earth’s crust; and as crystal concentrations from primary basaltic magma formed as bottom cumulates in temporary reservoirs, or formed marginally during upward flow. It is likely that all of these processes produce ultramafic nodules, Forbes and Kuno (196’7) presented a world-wide review of peridotite inclusions and their basaltic host rocks, and noted that spinel-lherzolite is most frequently reported. Their origin a ears to be related to the genesis of basaltic magma within the mantle. Sr J-? /Sr86 ratios confirm that the inclusions are genetically related to the basalts carrying them (Hurley, 1967), and indicate that there is no contemporary relationship with Alpine-type intrusions. Their low K/Rb ratios (Murthy and Stueber, 1967) and O’H&a’s (1967c) pyroxene grid, indicate that they equilibrated in the uppermost mantle, within the crystallization interval of anhydrous basaltic liquids. Their residual character (Goles, 1967; Murthy and Stueber, 1967) indicates that they do not represent a source of basaltic liquids. In contrast, the eclogite and garnet peridotite nodules in kimberlites are a potential source of basaltic liquids. O’Hara’s (1967c) pyroxene grid indicates that these equilibrated in the mantle at depths of 100-140 km on the thermal gradient for shield regions. In a recent review of the chemistry of ultramafic rocks and nodules, Harris et al. (1967) proposed that the nodules represent mantle fragments from a pyroxene-peridotite layer and deeper garnet-peridotite. The variable

Tectonophysics, 7 (5-6) (1969) 437-455

449

composition of pyroxene-peridotite nodules corresponds to various stages of depletion of basaltic components, the original mantle pyroxene-peridotite probably being chemically equivalent to the deeper garnet-peridotite. They pointed out that the few garnet-peridotite nodules so far analyzed may notbe representative. Although the chemical variation of garnet-peridotite nodules is inadequately known, a wide variation of mineralogical composition is established, and this variation encompasses rock types developed in high grade regional metamorphism (Davidson, 1967b). Even if these rucks are mantle fragments, Davidson’s comment demands attention: “If theso inclusions represent the upper mantle, then the latter is clearly of so varied a composition that petrogenetic studies based on a single rock-type such as garnet-peridotite must be of questionable value”.

GEOCHEMISTRY

OF

ULTRAMAFIC

ROCKS

In view of the difficulties of interpretation iol many ultramafic rocks in Alpine regions, geochemical data may prove useful for distinguishing among the various associations. But first it is necessary to characterize the different associations in chemical terms, and this has not been completed, even for the major elements, Some comparative data are available, Thayer (1967) used A-F-M diagrams to illustrate chemical trends for specific Alpine intrusive complexes, and for Alpine differentiation in general. Bowes et al. (1966) used a similar diagram to compare the chemical trends of Alpine-type rocks, rocks of the Bushveld intrusion, and the ultrabasic and basic gneisses in the Lewisian of Scotland; t,hey also compared these three groups of rocks in terms of Ca, Fe, and Mg. O’Hara (1966), discussing these trends, emphasized the importance of examining the geochemical sequence in successive layers. Only recently have advances in technique made possible the accurate measurement of trace element concentrations in olivine-rich rocks. Reviews of trace element abundances confirm that they may be useful as genetic guides (Goles, 1967; Murthy and Stueber, 196’7; Hurley, 1967). The extremely low concentrations of Na, K, Rb, Sr, and a low value for K/Rb in the Alpine ultramafic rocks indicate that they are residual in nature. The SrsY/Sr 86 values indicate no contemporary relationship between alpinetype ultramafic rocks and other materials such as stratiform intrusions, modern basalts, and ultramafic nodules in basalt, which suggests that the Alpineultramafic rocks can represent neither the residual refractory residue of mantle material from which basalts have been derived, nor the ultramafic portions of stratiform intrusions. On the other hand, the 018/016 ratios of minerals in basalts and gabbros are so similar to those in Alpinetype ultramafic rocks that they must be genetically related at some stage in earth history (Taylor, 1967b). The main conclusion arising from geochemical data is that the Alpine ultramafic rocks are derived from a portion of the mantle that attained its residual character during an early period of mantle differentiation (Murthy and Stueber, 1967). This implies that of the current hypotheses shown in Table I, only 3B and 3C appear to be capable of providing suitable material, because all others involve gabbroic rocks or magmas. However, I believe

that geochemists will admit that it is too soon to rule out hypothesis 1 and 2, because the available data are so sparse. There is a sampling problem that may be severe. For example, the Th and U contents of the Twin Sisters dunite (Goles, 1967) show marked variation from one specimen to another, and the concentrations of Rb, Sr, and K in ultramafic rocks are very sensitive to the distribution of small proportions of amphibole. Goles emphasized that what is urgently needed is a detailed investigation of the trace element contents of a single Alpine-type body, preferably one that has not been subjected to extreme deformation and metamorphism, showing the distribution of trace elements among the minerals within the body and in different parts of the same body. This should be compared with a similar study for a typical stratiform intrusion* If the petrogenesis of ultramafic rocks in Alpine and other associations is to be worked out, then both the hammer and black-box approaches must be followed. The geologist needs the information that can only be obtained by the use of sophisticated chemical techniques, and the geochemist can apply his data with maximum effect only if his samples can be located within an adequate petrological framework, which requires a detailed field map of the ultramafic association and the country rocks.

CONCLUDING REMARKS

As more information is obtained about the upper mantle the more complex and heterogeneous it appears to be. The study of appropriate terrestrial rocks is one approach for estimating the composition and mineralogy of the mantle, and the compositions so derived have been tabulated and compared with the results of other approaches by Hess (1964) and by Harris et al. (1967). Other reviews are included in Gaskell’s (1967) volume on the upper mantle. Petrological and geochemical studies of ultramafic rocks and nodules undoubtedly will continue to provide data for more refined speculations, but we must remind ourselves occasionally that it is a long extrapolation to the upper mantle in most locations, and that themethod of multiple working hypotheses has always worked well in the earth sciences.

ACKNOWLEDGEMENT

from

This review the National

developed from research Science Foundation.

supported

by Grant

GA-1289

REFERENCES Benson,

W.N., 1926. The tectonic conditions accompanying the intrusion of basic and ultrabasic igneous rocks. Natl. Acad. Sci. Mem., 19: l-90. Bhattacharji, S., 1967. Scale model experiments on flowage differentiation in sills. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.69-70. Bowen, N.L., 1928. The Evolution of the Igrz?eous Rocks. Princeton Univ. Press: Princeton, 332 pp. Bowes, D.R., Wright, A.E. and Park, R.G., 1966. Origin of ultrabasic and basic masses in the Lewisian. Geol. Mag., 103: 280-283.

Tectonophysics,

7 (5-6)

(1969) 437-455

431

ti.M., 1956. The layertvl ultrabasic lw~ks 01 Rhum, lnnc~i~ tlc~ill~icl~~>. l)llll. 1 I‘;,,, Roy. Sot. London: %I,. 13: 240: l-5::. Burli, C.A., 1964. A study 01 serlwntinitc. tllc, A.M.S.0.c’. r’ol’C’ IlOl~~ 11(‘Bi’ iva~aguc/., Puerto Rico. Katl. Acad. Sci. Natl. Iics. Council l’ubl.. llY\. 1;; 111,. Challis. (;.A., 196.h. The ol,lgir 01 St,\\ %(‘;iiantl ultl~amafic~ intl.izsi0rl.i. .J. I’~~t1~ol.. Brown,

6: 322-x4. Challis,

G.A., l!J63b. IIigh-tenll)eratul,~, c:cJntact nletan~o~l)l~is~~~ at tlw 1lt.d llills ultramaiic intrllsion Wxirau Valln~ Kc?\ %eahld. .I. l~etlol.. ii: ::!JZ-41!1. Cohen. L.lI., Ito, K. and Kc~nncdy. G.(!.. 136i. Melting and lkec, rc,latiorlh in an anhydrous basalt to 40 kbars. Am. J. Ski., 265: L73->15. Davidson, C.F.; 1967a. Thv Itimberlites oi the l’.S.S.R. In: l’.J. Wyllic, (I*;dltor), Irltramafiv and Related 1~0~1~s. Wilrsy. Xc\\ YOrk. S.Y., 11p.2.;l-Xl. Davidson, C.F., l!JG’ib. ‘I‘hv so-called “cognatv sfnoliths” of Iiimlx~rlitc~. In: 1l.J. Wylliv (Editor), l.Jltramafica and Relatc:d Rocks. LVilcy, Nc3~ York, N.Y _. I)f).::12-3%. Dawson. J 13. l!Xi Y. ‘> The geology Of Oldoin -0 1,cngai. Hull. Vol~anol.. 24: :i 49-387. Da\\ son, J:U. 1 1967a. A review 01 the geol 2gy of ” klmbrrlitc~. In: J’.J. 12’~Iii,, (Editor). Ultramal’ic and Related Rocks. Wiley , New York, iY.Y., I~fl.Z4-221. Damson. J.Is., 1%7b. Geochemistry and ol,igin 01 Itimbei~lite. III: 1’.J. CV~.llic (EditOr), l.:ltramal’ic and Related Rovks. C’v’ilcy,. Ncn Yo~~ic, K.Y.. ~~~~.26!)-27,~. Den ‘I‘cx, E., lY(i.7. Metamorphic IineagcTs 01 orogcniv ~)lutonism. (;eoi. P.iijnbouw: 44(4): lU5-1;:“.

Drever. H.I. hnd Johnston,

Il., 19tiia. ‘The, ultrabasic lwies in s~mt sills and shcels. In: P.J. Wyllic (Editor). Ultramal’iv and Related Ro~~lts. iL’iI(,y. New Yo~,lc, N.Y., pp.“l-ci:~. Drever, lI.1. and Johnston, R.. 19ti7b. l’ivritiv niinol, intl‘usiOns. 111: 1j.J. Wyllie (Editor)? rltramaiic and Related Rocks. LVilcs-. Nc\v York, N.Y., 1)11.71-X2. Ffshcr, D.J., I~‘ruvh. A.J.. IlUrlbut, C.S., ‘l’illcy, C.E. and IIcini~ich; 1c.W. 1%:;. Symposium On layr%rc>d intrusion. filincral. Sot. i-\nl.. Sl)c‘c’.I’apvlx. 1: l- 1::‘1. Forbes, 11.11. and Kuno. II.. 1967. I’tlridotitc inclusions and t)asaitic host IwckS. In: I'.J. IV\-Ilic, (EditOr). I’ltramafic and Relatfcl Rwks. LL’iltav. N:(,\\ Yol~k. S.Y..

(;aslwll. ‘1‘.1:. 1!16;, Thv Earth’s nlanilr,. Acatlcmic l’l~c~ss. London. 5U11 l,p. C;ass. I.(;.. l!).jS. rltrabasic I,illO\\ lavas Iron7 (‘yprus, Geol. Msg.. 9.7: 241-231. (:old, D.l’., 1967. Alkaline ultrabasic ro(.ks in the Muntrcal area. Quebec. In: I’.
1.7 2

D.H.. 1’367. High tvmpcraturc l~eridutitc intrusions. In: 1j.J. \Vyllie (I+;ditor). Ultramalic~ and Related Rovks. Wile) i Neu York, S.S.. 1~11.2lil--22t!. 01 basaltic magmas. (Iontrib. D.11. and RingxowI. A.E., 1967. Thv gcncsis Mineral. f’ctrol., 1,;: 103-190.

‘I‘c~ctonophysic~s.

7 (:-ti)

(19ti’J) 4:‘,7-4.?5

project: progress report. Bull. Gcol. SOC. Am., H.H., 1960. Caribbean research 71: 235-240. Hess, H.H., 1962. History of ocean basins. In: A.E.J. Engle, H.L. James and B.L. Leonard (Editor), Petrologic Studies - A Volume in Honor of A.F. Buddington. Geol. Sot. Am., New York, N.Y., pp.599-620. Hess, H.H., 1964. The oceanic crust, the upper mantle and the Mayaguez serpentinized peridotite. In: C.A. Burk (Editor), A Study of Serpentinite - Natl. Acad. Sci. Natl. Res. Council Publ., 1188: 169-175. Hess, H.H., 1966. Caribbean geological investigations. Geol. Sot. Am. Mem., 9X: l-310. Holmes, A., 1950. Petrogenesis of katungitc and its associates. Am. Mineralogist, 35. 772-792. relationships in the differentiation of the mantle. Hurley, P.M., 1967. 87Rb-87Sr In: P.J. Wgllic (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., p1,.::72-375. Irvine, T.S., 1967. The Duke Island ultramafic complex, southeastern Alaska. In: 1l.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y..

Hess,

pp.&-97. Irvine,

T.N. and Smith, C.H.; 1967. The ultramafic rocks of the Muskox intrusion? Northlvest Territories, Canada. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York; N.Y., pp.38-49. Ito, K. and Kennedy, G.C., 1967. Melting and phase relations in a natural peridotite to 40 kbar. Am. J. Sci., 265: 319-338. Jackson, E.D., 1967. Ultramafic cumulates in the Stillwater, Great Dyke and Bushveld Intrusions. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.20-38. Jahns, R.H., 1967. Serpentinites of the Roxbury District, Vermont. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, Ne\v York, N.Y., pp.137-160. Joplin, G.A., 1959. On the origin and occurrence of basic bodies associated with discordant bathyliths. Geol. Mag., 96: 361-373. King, B.C., 1965. Petrogenesis of the alkaline igneous rock suites of the \,olcanic and intrusive centres of eastern Uganda. J. Petrol., 6: 67-100. Lapham, D.M., 1967. The tectonic history of multiply deformed serpentinite in the piedmont of Pennsylvania. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.174-183. Lappin, M.A., 1967. Structural and petrofabric studies of the dunites of Almklovdalen, Nordfjord, Norway. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.183-190. Mathews, W.H., Thorarinsson, S. and Church, N.B., 1964. Gravitative settling of olivine in pillows of an Icelandic basalt. Am. J. Sci., 262: 1036-1040. Max\vell, J.C. and Azzaroli, A., 1963. Submarine extrusion of ultramafic magma. Geol. Sot. Am., Spec. Papers, 73: 20;1-204. Miyashiro, A., 1966. Some aspects of peridotite and serpcntinite in erogenic belts. Japan. J. Geol., Geogr., Trans., 37: 45-61. Moores, E.M., Hess, H.H. and Maxvvell, J.C., 1966. Petrology of an ultramafic-mafic (ophiolite) sequence. (abstr.). Progr. Geol. Sot. Am., Ann. Meeting, San Francisco, 1966 (abstract). Murthy, V.R. and Stueber, A.M., 1967. Potassium-rubidium ratios in mantle derived rocks. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.376-380. Noble,

J.A. and Taylor Jr., H.P., 1960. Correlation of the ultramafic complexes of southeastern Alaska with those of other parts of North America and the world. Intern. Gcol. Congr., 21st, Copenhagen, 1960, Rept. Session, Norden, 13: 188-197. O’Hara, M.J., 1966. Origin of ultrabasic and basic gneiss masses in the Lewisian. Geol. Mag., 103: 284. G’IIara, M.J., 1967a. Mineral facies in ultrabasic rocks. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.7-18.

Tectonophysics,

7 (5-6)

(1969) 437-455

453

O’Hara,

M.J., 1967~. Llinc>ral paragenc>sc’s in ultrabasic lwvks. In: I’.J. W!,llic (Er1itol.i Ultramafir and Related Rocks. LVilczy> Ntw York: N.Y., p~1,;1!~::---10::, Onuki, H., 1363. I’c~troc’hemical resc~arch on the IIoroman and Miyanlol,i ultramallc, intrusi\.es, northern Japan. Sci. Rq)t. Tohol;u Univ.. Third &I,.. 9: 217-27ti. (1nul;i, II., lD(i6. On tiw il,on- rich pcbl,idotitcxs in the Sanbagawa mc~tamoI@ic belt 01 the Iianto hlountainh. ,I. Jal)an. Assoc~. Mineral. I’c,trol. Econ. kol.. .i5: ::!1- 1;. I’JX?Silall. D.It., 1966. ‘J‘hr joint I’orstr:rit~,-diol)sitlt~-i,~on oxide and its bcaaring on tlw c.rystalli%atiw of’ basaltic and ultl~amafic magmas. Am. J. kki.! X.4: i.5:&80!). Ragan, D.M., 1967. The I‘\\ in Sisters duni&,, Washingt:ton. Jn: l’..l. \Vyllic, (Editor). Llltramalic and Rc~latc~d Roc:l;s. \Vil<,x, NC’LL.York, F.Y., pl~.ltiO-lti;. Raleigh, C.H.. 19ti7. l 1 Kcw York, K.Y.: lq,.l!Jl-199. Richter, L).If. and Murata, K.J., 19fX. I’ctlq,ral)hy 01 the lavas 01 the> I!I.59-19ciU csruption of’ Kilauca Volcano, Ilawaii. I:.S.. (;col. Surv., 1’1~~1~~ss. l’apers, 537-D: 12 pp. Ringwood, .A.E.. 1966a. The ehcmical composition and origin of thv Earth. Jn: i’.M. IIul,ley (Editor). Advances in Earth Sciences. M.I.T. I’I,~ss. Cambridge, Mass., lq’.%-::X. Ring\vwd, A.E., 1966b. Mincxra:ogy 01’ the mantle. In: P.M. IIurley (Editor), Advances in Earth Scienws. M.I.T. Press. Cambridge, Mass,, pp.%-:;9!J. Sahama, Th.G., 1962. l’&rology of Mount Nyiragongo. Trans. Edinburgh (k:ol. Sw., 19: 1-2s. Simkin. T., 1967. E’lon differentiation in the picritic sills of North Skycl. In: l’.,J. Wyllie (Editor). Lltramafic and Related Rocks. Wiley, Nc\\ York, N.Y., 1’1’.6-!-ti9. Smith, C.Il., 19.>h. Bay of Islands igneuus complex, western Neafoundland. Geol. Surv. Can. Mvm. ~ 2!)0? 1::2 pp. S$renson, II.. 1967. Metamorphic and metasomatic processes in the lormation of ultramaliv rocks. In: 1l.J. Wyllie (Editor). Ultramafic and Related Rocks. Wiltby. Ne\\ York. S.T.; pp.204-212. Taylor Jr., 11.1). 1967a. The zoned ultramafic complexes of Southcastcrn Alaska. In: l’.J. L\:)-lliv (Editor); Ultramafic and Rclatctl Rocks. Wiley, Nr\v York, N.T., pp.97- 121. Taylor ,Jr., IJ.1’. , 196ib. Stable isotope studies of ultramafic rocks and meteorites. In: 1l.J. Wyllie (Editor), Ultramal’ic and Related Rocks. Wiley, Nc>w York, N.Y.,

pp.:j(i2-Ai”. Taylor Thaycr.

Thay~r.

Turner, Tuttle.

ol the ultramafic complexes in south,Jr.. 11.1’. and Noble>. J.A.. 1960. Orlg;n eastc1.n Alaska. hitcsrn. Gcol. Congr., 21st, Copenhagen, 13: 173-187. ‘T.1’. . 1960. Some c.ritical differences bet%eeen alpine-type: and stratiform peridotitc-gabbro caoml)lexes. Intern. Geol. Con&r., 21st, Copenhagen. Rept. Session, Norden, 13: 247-239. T.1’. 1967. Chcmival and structural relations of ultramalic and feldspathic r(~cl
,291pp. Upton. U.G.J.I 1967. Allcalinc Rclatod Rocks. \~on Eckc~rmann. II., In: I’..J. Wyllic pp.::02-312.

434

pyroxenites. Wiley, Kcw York, 19ti7. A comparison (Editor). liltramafic

In: P.J. Wyllic (Editor’). Ultramafic and N.Y., pp.281-288. of Swedish, African and Russian Itimberlitcs. and Rc,lated Rocks. Wiley, New York, N.Y..

Tectonophysics,

7 (5-6)

(1969)

-437-43.;

Wager, L.R. and Brown, G.M., 1967. Layered Igneous Rocks. W.H. Freeman, San Francisco, Calif., 588 pp. Wager, L.R. and Deer, W.A., 1939. Geological investigations in East Greenland, Pt III. The petrology of the Skaergaard intrusion, Kangerdlugssuaq, east Greenland. Medd. Grrdnland, 10514): l-352. Watson, K.D., 1967. Kimberlites of eastern North America. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y ., pp.312-323. Williams, H., Turner, F.J. and Gilbert, C.M., 1954. Petrography. W.H. Freeman, San Francisco, Calif., 406 pp. Wyllie, P.J., 1967a. Ultramafic and Related Rocks. Wiley, New York, N.Y., 464~~. Wyllie, P.J., 1967b. Review. In: P.J. Wyllie (Editor), Ultramafic and Related Rocks. Wiley, New York, N.Y., pp.403-416.

Tectonophysics,

7 (5-6) (1969) 437-455

455