Alkaline Igneous Rocks
Geological Society Special Publications Series Editor
K. C 0 E
GEOLOGICAL
SOCIETY SPECIAL ...
336 downloads
2430 Views
56MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Alkaline Igneous Rocks
Geological Society Special Publications Series Editor
K. C 0 E
GEOLOGICAL
SOCIETY SPECIAL PUBLICATION
Alkaline Igneous Rocks E D I T E D BY
J. G. F I T T O N & B. G. J. U P T O N Grant Institute of Geology University of Edinburgh Edinburgh EH9 3JW
1987 Published for The Geological Society by Blackwell Scientific Publications OXFORD LONDON
EDINBURGH
BOSTON PALO ALTO MELBOURNE
N O 30
Published by Blackwell Scientific Publications Editorial offices: Osney Mead, Oxford OX2 0EL 8 John Street, London WC1N 2ES 23 Ainslie Place, Edinburgh EH3 6AJ 52 Beacon Street, Boston Massachusetts 02108, USA 667 Lytton Avenue, Palo Alto California 94301, USA 107 Barry Street, Carlton Victoria 3053, Australia
DISTRIBUTORS USA and Canada Blackwell Scientific Publications Inc PO Box 50009, Palo Alto California 94303 Australia Blackwell Scientific Publications (Australia) Pty Ltd 107 Barry Street, Carlton, Victoria 3053
9 1987 The Geological Society. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by The Geological Society for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that a base fee of $02.00 per copy is paid directly to CCC, 27 Congress Street, Salem, MA 01970, USA. 0305-8719/87 $02.00
British Library Cataloguing in Publication Data
First published 1987
ISBN0-632-01616-7
Typeset, printed and bound in Great Britain by William Clowes Limited, Beccles and London
Library of Congress Cataloging-in-Publication Data
Alkaline igneous rocks.--(Geological Society special publications, ISSN 0305-8719) 1. Alkalic igneous rocks I. Fitton, J.G. II. Upton, B. G. J. III. Series 552'. 1 QE462.A4
Alkaline igneous rocks. (Geological Society special publication; no. 30) Bibliography: p. Includes index. 1. Alkalic igneous rocks. I. Fitton, J. G. II. Upton, B. G.J. III. Geological Society of London. IV. Series. QE462.A4A43 1987 552'. 1 86-26364 ISBN0-632-01616-7
Contents Preface Introduction: FITTON, J. G. & UPTON, B. G . J . BAILEY, D. K. Mantle metasomatism--perspective and prospect
vii ix 1
MENZIES, M. Alkaline rocks and their inclusions: a window on the Earth's interior
15
EDGAR,A. D. The genesis of alkaline magmas with emphasis on their source regions: inferences from experimental studies
29
LE Bhs, M. J. Nephelinites and carbonatites
53
TWYMAN, J. D. & GITTINS, J. Alkalic carbonatite magmas: parental or derivative ?
85
DAWSON, J. B. The kimberlite clan: relationship with olivine and leucite lamproites, and inferences for upper-mantle metasomatism
95
BERGMAN, S. C. Lamproites and other potassium-rich igneous rocks: a review of their occurrence, mineralogy and geochemistry
103
ROCK, N. M. S. The nature and origin of lamprophyres: an overview
191
CLAGUE, D. A. Hawaiian alkaline volcanism
227
WEAVER, B. L., WOOD, D. A., TARNEY, J. & JORON, J. L. Geochemistry of ocean island basalts from the South Atlantic: Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha
253
HARRIS, C. & SHEPPARD, S. M. F. Magma and fluid evolution in the lavas and associated granite xenoliths of Ascension Island
269
FITTON, J. G. The Cameroon line, West Africa: a comparison between oceanic and continental alkaline volcanism
273
BAKER, B. H. Outline of the petroIogy of the Kenya rift alkaline province
293
MACDONALD,R. Quaternary peralkaline silicic rocks and caldera volcanoes of Kenya
313
WOOLLEY, A. R. & JONES, G. C. The petrochemistry of the northern part of the Chilwa alkaline province, Malawi
335
BOWDEN, P., BLACK, R., MARTIN, R. F., IKEI E. C. KINNAIRD, J. A. & BATCHELOR, R . A . Niger-Nigerian alkaline ring complexes: a classic example of African Phanerozoic anorogenic mid-plate magmatism
357
LI/~GEOIS, J. P. & BLACK, R. Alkaline magmatism subsequent to collision in the Pan-African belt of the Adrar des Iforas (Mali)
381
FLETCHER, C. J. N. & BEDDOE-STEPHENS, B. The petrology, chemistry and crystallization history of the Velasco alkaline province, eastern Bolivia
403
BARKER, D. S. Tertiary alkaline magmatism in Trans-Pecos Texas
415
EBY, G. N. The Monteregian Hills and White Mountain alkaline igneous provinces, eastern North America
433
vi
Contents
UPTON, B. G. J. & EMELEUS, C. H. Mid-Proterozoic alkaline magmatism in southern Greenland: the Gardar province
449
LARSEN, L. M. & SORENSEN, H. The Ilimaussaq intrusion--progressive crystallization and formation of layering in an agpaitic magma
473
NIELSEN,T. F. D. Tertiary alkaline magmatism in East Greenland: a review
489
DOWNES, H. Tertiary and Quaternary volcanism in the Massif Central, France
517
KOGARKO, L. N. Alkaline rocks of the eastern part of the Baltic Shield (Kola Peninsula)
531
Index
545
Preface The papers contained in this volume were presented at a symposium held in Edinburgh in September 1984, which marked the passage of ten years since the publication of The Alkaline Rocks edited by Henning Sorensen. In organizing the symposium and compiling this volume we aimed to review recent developments in the petrology and geochemistry of alkaline igneous rocks. We have, for example, paid particular attention to work on lamprophyres and carbonatites which are rock associations of current interest not covered in Sorensen's book. Reviews of recent work on some of the classic alkaline provinces, such as East Africa, southern Greenland and the Kola Peninsula, are included together with reviews of less wellknown areas. Other papers discuss the impact of experimental, geochemical and isotopic studies on our understanding of the generation and evolution of alkaline magmas. We are indebted to the contributors for their collaboration in producing this volume and it is with sadness that we note the death, on 14 February 1986, of Brian Baker, whose pioneering field studies formed the basis for much of our knowledge of the tectonic and volcanic evolution of the East African Rift. An obituary and appreciation of his work is published in the Journal of Volcanology and Geothermal Research (28, v-vii). We are grateful to The Geological Society and The Royal Society of Edinburgh for their generous assistance with the symposium costs, to Lucian Begg and Dodie James for their help with organizing the symposium and producing this volume, and to our colleagues for the care and enthusiasm with which they reviewed the manuscripts. The efforts of Edward Wates and his staff at Blackwell Scientific Publications are also gratefully acknowledged. To all these we offer our sincere thanks. J.G.F. B.G.J.U.
vii
Introduction J. G. Fitton & B. G. J. Upton Alkaline igneous rocks may be defined as those which have higher concentrations of alkalis than can be accommodated in feldspars alone, the excess appearing as feldspathoids, sodic pyroxenes, sodic amphiboles and other alkali-rich phases. These rocks are, therefore, deficient in silica and/or alumina with respect to alkalis and will have nepheline and/or acmite in their norms. In practice the term 'alkaline' is used to encompass a wide range of igneous rocks, not all of which conform to this rigid definition. Carbonatites, for example, are certainly silica-deficient but are rarely alkali-rich. True (nepheline-normative) alkali basalts grade into hypersthene-normative transitional basalts without any obvious change in mineralogy. Since transitional basalts are often closely associated with alkali basalts in the field, they are traditionally regarded as alkaline. It is now usual practice to define alkaline igneous rocks simply in terms of their alkali ( N a 2 0 + K20) and silica contents (see, for example, Le Bas et al. 1986). We have not attempted to review the classification of alkaline igneous rocks in this volume as this has been dealt with elsewhere (e.g. Sorensen 1974; Streckeisen 1967, 1980). The only alkaline rocks not covered in previous reviews are those hydrous mafic to ultramafic hypabyssal rocks known as the lamprophyres. The present volume includes three papers on this group. A comprehensive overview of lamprophyres is given by Rock and of the sub-group of lamproites by Bergman. The relationship between lamproites and kimberlites (which arguably belong to the lamprophyres) is discussed by Dawson. Volumetrically, alkaline rocks account for less than one per cent of all igneous rocks. Despite this, their remarkable mineralogical diversity has brought them repeatedly to the attention of petrologists and mineralogists, with the result that alkaline rocks account for about half of all igneous rock names. Sorensen (1974) lists no fewer than 400 alkaline rock types. This diversity springs largely from an abundance of alkalis and deficiency in silica which together generate a large number of mineral species not stable in more silica-rich, alkali-poor magmas. However, a large part of the attention given to alkaline rocks is due to their characteristic high concentrations of incompatible or large-ion lithophile elements (LILE). These are often of more than academic interest as most of the world's resources of niobium, tantalum and the rare-earth elements are found in or around alkaline igneous rock bodies. The economic importance of alkaline igneous rocks is further enhanced by their association with economic deposits of apatite (Kogarko) and with diamonds (Dawson;
Bergman). Evidence for continental alkaline magmatism can be found as far back as the late Archaean. For example, biotites from the Poohbah Lake syenite in north-western Ontario have been dated at around 2.7 Ga (Mitchell 1976) and a similar age has been reported by Larsen et al. (1983) for the Tupertalik carbonatite in western Greenland. At the present time, alkaline magmas are erupted in all tectonic environments with the possible exception of mid-ocean ridges. Even here, though, mildly alkaline lavas are sometimes erupted from off-axis volcanoes, as in the Vestmann Islands of Iceland. Alkaline igneous rocks are found on all the continents and on islands in all the ocean basins. Their occurrences may be classified on the basis of tectonic setting into three categories; continental rift valley From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. ix-xiv.
ix
x
Introduction
magmatism, oceanic and continental intraplate magmatism without clear tectonic control, and alkaline magmatism related to subduction processes. In practice, however, this classification is not always easy to apply. Continental rift valleys provide, volumetrically, the most important occurrences of alkaline igneous rocks although continental rifting is not always accompanied by magmatism. The best known example (arguably the type example) is the East African Rift which, in the course of its long history, has yielded almost the entire spectrum of alkaline magmas. Three papers in this volume are devoted to various aspects of magmatism in the East African Rift. Baker reviews its magmatic associations with respect to tectonic development and discusses the origin of the magmas, particularly in relation to the problems presented by the strongly bimodal distribution of basic and salic lava compositions. He concludes that the salic magmas evolved from basic parental magmas by processes of crystal fractionation (cf. Bailey). Macdonald focusses attention on the peralkaline silicic central volcanoes of Kenya and also favours an origin by crystal fractionation for most of the evolved magmas. There is little evidence for contamination of the evolving magmas with ancient continental crust except in the case of the Naivasha comendites. It is not always possible to demonstrate a genetic link between basic and evolved magmas in the East African Rift, however. The Chilwa alkaline province in Malawi, at the southern end of the rift, is an essentially intrusive province in which salic rocks predominate. The scarcity of basic rocks in this part of the rift has led Woolley & Jones to suggest that the evolved magmas were produced directly by melting of metasomatised mantle and lower crust. Insight into the processes occurring at depth beneath rift valleys may be gained by studying ancient and deeply eroded examples. The Proterozoic Gardar province in SouthWest Greenland is probably the best studied of these and is reviewed by Upton & Emeleus. One of the most striking features of the province is the presence of giant dykes, up to 800 metres wide. These are dominantly basic but in places show in situ differentiation into more salic rocks. Salic magma generated in the wider portions of these dykes migrated upwards and may ultimately have accumulated to produce central complexes in which basic magma was subordinate or absent (e.g. the Ilimaussaq intrusion, Larsen & Serensen). The giant dykes, therefore, play a crucial role in understanding the relationship between basic and salic magmas in this and possibly other rift systems. The separation of continents to form ocean basins must always be preceded by a phase of continental rifting, leaving volcanic and intrusive complexes stranded along passive continental margins. The vigorous magmatism which accompanies continental separation is generally tholeiitic in character as, for example, in the Karoo and Deccan flood basalt provinces. Alkaline magmas, however, may be emplaced along the trailing continental margin during the waning phase of magmatism, long after the spreading centre is established off-shore. The Tertiary volcanic rocks exposed along the east coast of Greenland (described by Nielsen) provide an excellent example of such an alkaline province. A second major occurrence of alkaline igneous rocks is provided by intraplate magmatic provinces whose activity and siting are not subject to any obvious tectonic control. In the ocean basins such magmatism manifests itself as ocean islands which are sometimes aligned in chains with ages increasing away from the active centres, as in the Hawaiian islands. In these cases it is possible to relate the magmatism to convective plumes within the asthenosphere. The Hawaiian islands (reviewed by Clague) show clearly defined
Introduction
xi
magmatic cycles starting with alkaline magmas (represented by Loihi seamount), passing through a voluminous tholeiitic shield-building stage and returning to alkaline magmatism during the waning phases of individual volcanic centres. These cycles seem to result from movement of the oceanic plate over a rising plume of partially molten asthenosphere in which the degree of melting increases towards the centre. They may be typical of ocean islands in general and are broadly analogous to similar cycles seen in flood basalt provinces preserved on passive continental margins. Continental intraplate magmatism may also show age progressions as in the NigerNigeria province (Bowden et al.). These progressions are, however, very rare and not so clearly defined as in ocean island chains, probably because the continental lithosphere is thicker and less easily penetrated than oceanic lithosphere. The Monteregian Hills and White Mountain provinces of eastern North America (Eby), for example, show no obvious progression but their seaward extension, the New England seamounts, show a regular decrease in age eastwards. Reviews of other continental intraplate alkaline provinces are given by Fletcher & Beddoe-Stevens (Velasco province, Bolivia) and Kogarko (Kola Peninsula). The Cameroon line in West Africa (Fitton) includes both continental and oceanic alkaline volcanic centres. None of these examples shows any clear progression of ages. Some continental provinces undergo repeated alkaline magmatism in one place over long periods. For example, the Kola Peninsula (Kogarko) was the site of alkaline magmatism in the mid Proterozoic and again in the Devonian. Such examples could be the result of coincidence but are more likely the result of the repeated exploitation of zones of weakness in the lithosphere. Destructive plate boundaries provide the third tectonic setting in which alkaline igneous rocks may occur. During the life of a subduction zone the characteristic calc-alkaline magmas tend to become more potassic with time and may give way to volcanic rocks of the shoshonitic association, some members of which may contain leucite. A discussion of subduction zone processes is beyond the scope of this volume and the reader is referred to the reviews of Gill (1981) and Ewart (1982). There are, however, two circumstances under which subduction processes can lead to the generation of more 'normal' alkaline magmas. Once the descending slab has become dehydrated at depth it loses its capacity to stimulate the generation of calc-alkaline magmas but it can still cause melting in the overlying asthenosphere. This can lead to the production of alkaline magmas from the mantle above the deepest parts of subduction zones. One such example of alkaline magmas erupted under a compressive regime is provided by the Trans-Pecos province of west Texas (Barker). The alkaline rocks in this area grade south-westwards into the calc-alkaline rocks of the Sierra Madre Occidental in Mexico and Barker relates both suites to subduction of the Farallon Plate. After the cessation of subduction, relaxation of the former compressive regime often results in extension and the generation of alkaline magmas. The resulting switch from subduction-related calc-alkaline to extensional alkaline magmatism appears to be a common phenomenon. It occurred, for example, in the western U.S.A. about 17 Ma ago and in parts of Africa and Arabia at the end of the Pan-African metamorphic episode during the late Precambrian. An example from the Pan-African belt in Mali is discussed by Liegeois & Black. The origin of alkaline magmas has attracted a great deal of interest among igneous petrologists over the last ten years or so. This interest has been stimulated by two important
xii
Introduction
and characteristic features of alkaline rocks. Firstly, they contain high concentrations of LILE and yet isotopic evidence suggests that their parental magmas had a mantle source which had been depleted in these elements for a long time. Thus alkaline igneous rocks may provide useful information about enrichment and/or melting processes in the mantle. Secondly, many mafic alkaline volcanic rocks contain xenoliths inferred to have originated within the mantle (Menzies). These are often enriched in LILE when compared with concentrations expected of chondritic mantle material and sometimes contain amphiboles and micas of metasomatic origin (Bailey). Such clear evidence for the existence of metasomatically enriched mantle, coupled with the problem of extracting LILE-rich magmas from LILE-poor mantle, has led to hypotheses involving mantle metasomatism as a precursor to alkaline magmatism. These hypotheses, reviewed by Bailey, have gained popularity over the past fifteen years and are invoked by several contributors to this volume. Mantle metasomatism neatly explains many of the features of alkaline magmatism. For example, the frequent association of alkaline magmatism with areas of large-scale regional uplift is consistent with the relatively low density of metasomatized mantle. An essential feature of all models involving a metasomatized mantle source for alkaline magmas is that this source must lie in the lithosphere. This is the only part of the mantle where enriched material can remain in one place for long periods without being swept away by convection. The lithospheric mantle beneath the continents is likely to be chemically and isotopically different from that beneath the oceans. Continental lithospheric mantle is old and will have had as complex a metamorphic and magmatic history as the overlying crust. Oceanic lithosphere, on the other hand, is relatively young and probably depleted in LILE. These differences should be reflected in the compositions of continental and oceanic alkaline rocks. However, alkali basalts erupted in continental and oceanic settings are generally identical both chemically (Fitton) and isotopically (Menzies). Since enriched mantle xenoliths are only commonly found in continental regions it follows that the enriched lithospheric mantle represented by these xenoliths is not the source of most continental and oceanic alkaline magmas. An asthenospheric source is therefore implied. This is not to say that enriched continental lithospheric mantle is never involved in the generation of alkaline magmas. There is good evidence (e.g. Edgar) that pockets of ancient enriched mantle beneath cratonic regions provide the source for LILE-rich mafic and ultramafic alkaline rocks such as micaceous kimberlite (Dawson), and lamproite and other potassic igneous rocks (Bergman). It is significant that these rock types are exclusively continental. More extensive melting may involve the continental lithosphere mantle in the production of less exotic rock types such as flood tholeiite and mildly alkaline basalt. Upton & Emeleus, for example, argue for a lithospheric mantle source for the Gardar alkaline magmas. If most alkaline magmas have an ultimate source in the asthenosphere then they must share this source with unequivocally asthenosphere-derived rocks such as mid-ocean ridge basalt (MORB). The consistent isotopic differences between MORB and alkali basalts (and indeed all intraplate basalts) requires that the asthenosphere be heterogeneous. This heterogeneity may result from the entrainment of lower mantle material in deep mantle plumes as suggested for the Hawaiian island chain (Clague). The entire convecting upper mantle may also be heterogeneous on a small scale and alkaline magmas may be generated by the selective melting of LILE-enriched streaks while more extensive melting produces MORB (Fitton). Geochemical studies on ocean island basalts from the South Atlantic
Introduction
xiii
(Weaver et a/.) suggest that one component of these enriched streaks is provided by subducted ocean-floor sediment. Derivation of LILE-rich magmas from an asthenospheric source depleted in these elements requires either very small degress of partial melting or extensive crystal fractionation. There can be no doubt that many alkaline rocks are the products of extensive low pressure crystal fractionation but this cannot be true of those alkaline rocks which host mantle xenoliths. Even the most magnesian alkaline rocks, which must represent nearprimary magmas, are rich in LILE. Small-degree partial melting ( < l ~ ) is therefore required to produce such magmas. McKenzie (1985) has recently shown that the extraction of melt fractions as small as 0.2~ is not only physically possible but inevitable where the melt viscosity is low, as it probably is in the case of alkaline magmas. Experimental studies on alkaline rocks and synthetic analogues (reviewed by Edgar) provide useful constraints on the feasibility of fractional crystallization and partial melting models and on the temperatures and pressures involved. Many lines of evidence suggest that volatile components form a significant part of alkaline magmas. The development of extensive zones of metasomatised country rock (fenite) around alkaline plutons, the abundance of chlorine and fluorine in some alkaline igneous rocks, and the frequently explosive eruption of alkaline magma all point to high concentrations of volatiles. These volatile components play an important role in the evolution of alkaline magmas and yet relatively little is known about them. Constraints on their composition have been provided by fluid inclusion studies (Harris & Sheppard) and by thermodynamic considerations (Kogarko). The effects of volatile components on the evolution of alkaline magmas can be seen clearly in both intrusive and extrusive rocks. Larsen & Serensen, for example, discuss the crystallization history of the Ilimaussaq intrusion in South-West Greenland and show how the upward migration of low-density, low-viscosity volatile-rich magma delayed crystallization under the roof of the intrusion. Silicic alkaline pyroclastic deposits around central volcanoes in Kenya often show striking variations in the abundance of some incompatible elements within a single vertical section, implying compositional zonation in the magma chamber before eruption. Macdonald shows that these variations are too large to be accounted for by crystal fractionation alone and suggests that some elements have been transported to the magma chamber roof zones as complex ions in a volatile phase. Carbonatites provide perhaps the best illustration of all of the influence of volatile components on the origin and evolution of alkaline rocks. There is now a consensus that their parental magmas originate by the separation of an immiscible carbonate liquid phase from a CO2-saturated nephelinite or phonolite magma. There is, however, some disagreement over the nature and subsequent evolution of this parental carbonate magma. Le Bas argues that the parental magma is rich in alkalis and similar in composition to the natrocarbonatite lavas erupted from Oldoinyo Lengai. This magma evolves at low pressure towards the more common calcite carbonatite (s6vite) by loss of alkalis to the surrounding country rocks which are metasomatized (fenitized) as a result. Twyman & Gittins offer an alternative scheme in which s6vite magmas are parental and natrocarbonatite magmas are derived from them by crystal fractionation. Most petrologists now believe that evolved alkaline magmas are produced by the fractional crystallization of basic magma. The more highly undersaturated parental magmas represented by basanite and nephelinite will produce undersaturated derivatives
xiv
Introduction
such as phonolite and foyaite. Mildly alkaline and transitional basalt m a g m a is likely to produce trachyte and, with extreme fractionation, alkali rhyolite. The production of peralkaline acid rocks by crystal fractionation alone is likely to be an inefficient process. The operation of this process, however, is clearly demonstrated by their association with transitional basalt on some ocean islands, such as Ascension (Harris & Sheppard). Peralkaline acid rocks only occur in a b u n d a n c e in continental environments, however, and here there is often good evidence that crustal contamination has a c c o m p a n i e d crystal fractionation. Well-documented examples of the operation of crustal c o n t a m i n a t i o n in the evolution of alkaline m a g m a s are presented by several of the contributors to this volume. Downes, for example, shows that the assimilation of lower crustal granulite has affected the evolution of alkaline m a g m a s in the F r e n c h Massif Central and uses isotope data to estimate the extent of this contamination. Other examples are presented by Bowden et al. ( N i g e r - N i g e r i a granite ring complexes), Eby (Monteregian Hills and White Mountain provinces, N o r t h America), Fitton (Cameroon line, West Africa) and Fletcher & BeddoeStevens (Velasco province, Bolivia). Other authors propose the derivation of evolved alkaline m a g m a s directly from m e t a s o m a t i z e d mantle or lower crust (Bailey; Woolley &
Jones). Our understanding of the origin and evolution of alkaline m a g m a s has come a long way since the publication of Sorensen's book in 1974, largely through the acquisition of a far larger geochemical and isotopic data base. The contributions to this volume review the current state of this understanding. Emphasis has shifted from crustal to mantle processes with the recognition of mantle m e t a s o m a t i s m and its possible role as a precursor to alkaline magmatism. More recently, though, there has been a swing towards the opposite view, that mantle m e t a s o m a t i s m is c a u s e d b y alkaline m a g m a t i s m . Theoretical and experimental studies on the migration and segregation of small-degree melts seem destined to accelerate this swing. Despite these advances, however, m a n y mysteries remain unsolved and alkaline rocks will still provide a fruitful field of research for m a n y years to come, yielding further insights into the nature of mantle processes and the evolution of magmas.
References EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In Thorpe, R. S. (ed.) Andesites. Pp. 25-87. John Wiley & Sons, London. GILL,J. B. 1981. Orogenic Andesites and Plate Tectonics, 390 pp. Springer-Verlag, Berlin. LARSEN,L. M., REX,D. C. & SECHER,K. 1983. The age of carbonatites, kimberlites and lamprophyres from southern west Greenland: recurrent alkaline magmatism during 2500 million years. Lithos 16, 215-21. LE BAS,M. J., LE MAITRE, R. W., STRECKEISEN,A. & ZANETTIN, B. 1986. A chemical classification of
volcanic rocks based on the total alkali--silica diagram. J. Petrol. 27, 745-50. MCKENZIE, D. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-91. MITCHELL,R. H. 1976. Potassium-argon geochronology of the Poohbah Lake alkaline complex, northwestern Ontario. Can. J. Earth Sci. 13, 1456-9. SORENSEN,H. (ed.) 1974. The Alkaline Rocks. 622 pp. John Wiley & Sons, London. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks. N. Jb. Miner. Abh. 107, 144-240. - - , 1980. Classificationand nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Geol. Rundschau, 69, 194-207.
J. G. FITTON& B. G. J. UPTON,Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Introduction J. G. Fitton & B. G. J. Upton Alkaline igneous rocks may be defined as those which have higher concentrations of alkalis than can be accommodated in feldspars alone, the excess appearing as feldspathoids, sodic pyroxenes, sodic amphiboles and other alkali-rich phases. These rocks are, therefore, deficient in silica and/or alumina with respect to alkalis and will have nepheline and/or acmite in their norms. In practice the term 'alkaline' is used to encompass a wide range of igneous rocks, not all of which conform to this rigid definition. Carbonatites, for example, are certainly silica-deficient but are rarely alkali-rich. True (nepheline-normative) alkali basalts grade into hypersthene-normative transitional basalts without any obvious change in mineralogy. Since transitional basalts are often closely associated with alkali basalts in the field, they are traditionally regarded as alkaline. It is now usual practice to define alkaline igneous rocks simply in terms of their alkali ( N a 2 0 + K20) and silica contents (see, for example, Le Bas et al. 1986). We have not attempted to review the classification of alkaline igneous rocks in this volume as this has been dealt with elsewhere (e.g. Sorensen 1974; Streckeisen 1967, 1980). The only alkaline rocks not covered in previous reviews are those hydrous mafic to ultramafic hypabyssal rocks known as the lamprophyres. The present volume includes three papers on this group. A comprehensive overview of lamprophyres is given by Rock and of the sub-group of lamproites by Bergman. The relationship between lamproites and kimberlites (which arguably belong to the lamprophyres) is discussed by Dawson. Volumetrically, alkaline rocks account for less than one per cent of all igneous rocks. Despite this, their remarkable mineralogical diversity has brought them repeatedly to the attention of petrologists and mineralogists, with the result that alkaline rocks account for about half of all igneous rock names. Sorensen (1974) lists no fewer than 400 alkaline rock types. This diversity springs largely from an abundance of alkalis and deficiency in silica which together generate a large number of mineral species not stable in more silica-rich, alkali-poor magmas. However, a large part of the attention given to alkaline rocks is due to their characteristic high concentrations of incompatible or large-ion lithophile elements (LILE). These are often of more than academic interest as most of the world's resources of niobium, tantalum and the rare-earth elements are found in or around alkaline igneous rock bodies. The economic importance of alkaline igneous rocks is further enhanced by their association with economic deposits of apatite (Kogarko) and with diamonds (Dawson;
Bergman). Evidence for continental alkaline magmatism can be found as far back as the late Archaean. For example, biotites from the Poohbah Lake syenite in north-western Ontario have been dated at around 2.7 Ga (Mitchell 1976) and a similar age has been reported by Larsen et al. (1983) for the Tupertalik carbonatite in western Greenland. At the present time, alkaline magmas are erupted in all tectonic environments with the possible exception of mid-ocean ridges. Even here, though, mildly alkaline lavas are sometimes erupted from off-axis volcanoes, as in the Vestmann Islands of Iceland. Alkaline igneous rocks are found on all the continents and on islands in all the ocean basins. Their occurrences may be classified on the basis of tectonic setting into three categories; continental rift valley From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. ix-xiv.
ix
x
Introduction
magmatism, oceanic and continental intraplate magmatism without clear tectonic control, and alkaline magmatism related to subduction processes. In practice, however, this classification is not always easy to apply. Continental rift valleys provide, volumetrically, the most important occurrences of alkaline igneous rocks although continental rifting is not always accompanied by magmatism. The best known example (arguably the type example) is the East African Rift which, in the course of its long history, has yielded almost the entire spectrum of alkaline magmas. Three papers in this volume are devoted to various aspects of magmatism in the East African Rift. Baker reviews its magmatic associations with respect to tectonic development and discusses the origin of the magmas, particularly in relation to the problems presented by the strongly bimodal distribution of basic and salic lava compositions. He concludes that the salic magmas evolved from basic parental magmas by processes of crystal fractionation (cf. Bailey). Macdonald focusses attention on the peralkaline silicic central volcanoes of Kenya and also favours an origin by crystal fractionation for most of the evolved magmas. There is little evidence for contamination of the evolving magmas with ancient continental crust except in the case of the Naivasha comendites. It is not always possible to demonstrate a genetic link between basic and evolved magmas in the East African Rift, however. The Chilwa alkaline province in Malawi, at the southern end of the rift, is an essentially intrusive province in which salic rocks predominate. The scarcity of basic rocks in this part of the rift has led Woolley & Jones to suggest that the evolved magmas were produced directly by melting of metasomatised mantle and lower crust. Insight into the processes occurring at depth beneath rift valleys may be gained by studying ancient and deeply eroded examples. The Proterozoic Gardar province in SouthWest Greenland is probably the best studied of these and is reviewed by Upton & Emeleus. One of the most striking features of the province is the presence of giant dykes, up to 800 metres wide. These are dominantly basic but in places show in situ differentiation into more salic rocks. Salic magma generated in the wider portions of these dykes migrated upwards and may ultimately have accumulated to produce central complexes in which basic magma was subordinate or absent (e.g. the Ilimaussaq intrusion, Larsen & Serensen). The giant dykes, therefore, play a crucial role in understanding the relationship between basic and salic magmas in this and possibly other rift systems. The separation of continents to form ocean basins must always be preceded by a phase of continental rifting, leaving volcanic and intrusive complexes stranded along passive continental margins. The vigorous magmatism which accompanies continental separation is generally tholeiitic in character as, for example, in the Karoo and Deccan flood basalt provinces. Alkaline magmas, however, may be emplaced along the trailing continental margin during the waning phase of magmatism, long after the spreading centre is established off-shore. The Tertiary volcanic rocks exposed along the east coast of Greenland (described by Nielsen) provide an excellent example of such an alkaline province. A second major occurrence of alkaline igneous rocks is provided by intraplate magmatic provinces whose activity and siting are not subject to any obvious tectonic control. In the ocean basins such magmatism manifests itself as ocean islands which are sometimes aligned in chains with ages increasing away from the active centres, as in the Hawaiian islands. In these cases it is possible to relate the magmatism to convective plumes within the asthenosphere. The Hawaiian islands (reviewed by Clague) show clearly defined
Introduction
xi
magmatic cycles starting with alkaline magmas (represented by Loihi seamount), passing through a voluminous tholeiitic shield-building stage and returning to alkaline magmatism during the waning phases of individual volcanic centres. These cycles seem to result from movement of the oceanic plate over a rising plume of partially molten asthenosphere in which the degree of melting increases towards the centre. They may be typical of ocean islands in general and are broadly analogous to similar cycles seen in flood basalt provinces preserved on passive continental margins. Continental intraplate magmatism may also show age progressions as in the NigerNigeria province (Bowden et al.). These progressions are, however, very rare and not so clearly defined as in ocean island chains, probably because the continental lithosphere is thicker and less easily penetrated than oceanic lithosphere. The Monteregian Hills and White Mountain provinces of eastern North America (Eby), for example, show no obvious progression but their seaward extension, the New England seamounts, show a regular decrease in age eastwards. Reviews of other continental intraplate alkaline provinces are given by Fletcher & Beddoe-Stevens (Velasco province, Bolivia) and Kogarko (Kola Peninsula). The Cameroon line in West Africa (Fitton) includes both continental and oceanic alkaline volcanic centres. None of these examples shows any clear progression of ages. Some continental provinces undergo repeated alkaline magmatism in one place over long periods. For example, the Kola Peninsula (Kogarko) was the site of alkaline magmatism in the mid Proterozoic and again in the Devonian. Such examples could be the result of coincidence but are more likely the result of the repeated exploitation of zones of weakness in the lithosphere. Destructive plate boundaries provide the third tectonic setting in which alkaline igneous rocks may occur. During the life of a subduction zone the characteristic calc-alkaline magmas tend to become more potassic with time and may give way to volcanic rocks of the shoshonitic association, some members of which may contain leucite. A discussion of subduction zone processes is beyond the scope of this volume and the reader is referred to the reviews of Gill (1981) and Ewart (1982). There are, however, two circumstances under which subduction processes can lead to the generation of more 'normal' alkaline magmas. Once the descending slab has become dehydrated at depth it loses its capacity to stimulate the generation of calc-alkaline magmas but it can still cause melting in the overlying asthenosphere. This can lead to the production of alkaline magmas from the mantle above the deepest parts of subduction zones. One such example of alkaline magmas erupted under a compressive regime is provided by the Trans-Pecos province of west Texas (Barker). The alkaline rocks in this area grade south-westwards into the calc-alkaline rocks of the Sierra Madre Occidental in Mexico and Barker relates both suites to subduction of the Farallon Plate. After the cessation of subduction, relaxation of the former compressive regime often results in extension and the generation of alkaline magmas. The resulting switch from subduction-related calc-alkaline to extensional alkaline magmatism appears to be a common phenomenon. It occurred, for example, in the western U.S.A. about 17 Ma ago and in parts of Africa and Arabia at the end of the Pan-African metamorphic episode during the late Precambrian. An example from the Pan-African belt in Mali is discussed by Liegeois & Black. The origin of alkaline magmas has attracted a great deal of interest among igneous petrologists over the last ten years or so. This interest has been stimulated by two important
xii
Introduction
and characteristic features of alkaline rocks. Firstly, they contain high concentrations of LILE and yet isotopic evidence suggests that their parental magmas had a mantle source which had been depleted in these elements for a long time. Thus alkaline igneous rocks may provide useful information about enrichment and/or melting processes in the mantle. Secondly, many mafic alkaline volcanic rocks contain xenoliths inferred to have originated within the mantle (Menzies). These are often enriched in LILE when compared with concentrations expected of chondritic mantle material and sometimes contain amphiboles and micas of metasomatic origin (Bailey). Such clear evidence for the existence of metasomatically enriched mantle, coupled with the problem of extracting LILE-rich magmas from LILE-poor mantle, has led to hypotheses involving mantle metasomatism as a precursor to alkaline magmatism. These hypotheses, reviewed by Bailey, have gained popularity over the past fifteen years and are invoked by several contributors to this volume. Mantle metasomatism neatly explains many of the features of alkaline magmatism. For example, the frequent association of alkaline magmatism with areas of large-scale regional uplift is consistent with the relatively low density of metasomatized mantle. An essential feature of all models involving a metasomatized mantle source for alkaline magmas is that this source must lie in the lithosphere. This is the only part of the mantle where enriched material can remain in one place for long periods without being swept away by convection. The lithospheric mantle beneath the continents is likely to be chemically and isotopically different from that beneath the oceans. Continental lithospheric mantle is old and will have had as complex a metamorphic and magmatic history as the overlying crust. Oceanic lithosphere, on the other hand, is relatively young and probably depleted in LILE. These differences should be reflected in the compositions of continental and oceanic alkaline rocks. However, alkali basalts erupted in continental and oceanic settings are generally identical both chemically (Fitton) and isotopically (Menzies). Since enriched mantle xenoliths are only commonly found in continental regions it follows that the enriched lithospheric mantle represented by these xenoliths is not the source of most continental and oceanic alkaline magmas. An asthenospheric source is therefore implied. This is not to say that enriched continental lithospheric mantle is never involved in the generation of alkaline magmas. There is good evidence (e.g. Edgar) that pockets of ancient enriched mantle beneath cratonic regions provide the source for LILE-rich mafic and ultramafic alkaline rocks such as micaceous kimberlite (Dawson), and lamproite and other potassic igneous rocks (Bergman). It is significant that these rock types are exclusively continental. More extensive melting may involve the continental lithosphere mantle in the production of less exotic rock types such as flood tholeiite and mildly alkaline basalt. Upton & Emeleus, for example, argue for a lithospheric mantle source for the Gardar alkaline magmas. If most alkaline magmas have an ultimate source in the asthenosphere then they must share this source with unequivocally asthenosphere-derived rocks such as mid-ocean ridge basalt (MORB). The consistent isotopic differences between MORB and alkali basalts (and indeed all intraplate basalts) requires that the asthenosphere be heterogeneous. This heterogeneity may result from the entrainment of lower mantle material in deep mantle plumes as suggested for the Hawaiian island chain (Clague). The entire convecting upper mantle may also be heterogeneous on a small scale and alkaline magmas may be generated by the selective melting of LILE-enriched streaks while more extensive melting produces MORB (Fitton). Geochemical studies on ocean island basalts from the South Atlantic
Introduction
xiii
(Weaver et a/.) suggest that one component of these enriched streaks is provided by subducted ocean-floor sediment. Derivation of LILE-rich magmas from an asthenospheric source depleted in these elements requires either very small degress of partial melting or extensive crystal fractionation. There can be no doubt that many alkaline rocks are the products of extensive low pressure crystal fractionation but this cannot be true of those alkaline rocks which host mantle xenoliths. Even the most magnesian alkaline rocks, which must represent nearprimary magmas, are rich in LILE. Small-degree partial melting ( < l ~ ) is therefore required to produce such magmas. McKenzie (1985) has recently shown that the extraction of melt fractions as small as 0.2~ is not only physically possible but inevitable where the melt viscosity is low, as it probably is in the case of alkaline magmas. Experimental studies on alkaline rocks and synthetic analogues (reviewed by Edgar) provide useful constraints on the feasibility of fractional crystallization and partial melting models and on the temperatures and pressures involved. Many lines of evidence suggest that volatile components form a significant part of alkaline magmas. The development of extensive zones of metasomatised country rock (fenite) around alkaline plutons, the abundance of chlorine and fluorine in some alkaline igneous rocks, and the frequently explosive eruption of alkaline magma all point to high concentrations of volatiles. These volatile components play an important role in the evolution of alkaline magmas and yet relatively little is known about them. Constraints on their composition have been provided by fluid inclusion studies (Harris & Sheppard) and by thermodynamic considerations (Kogarko). The effects of volatile components on the evolution of alkaline magmas can be seen clearly in both intrusive and extrusive rocks. Larsen & Serensen, for example, discuss the crystallization history of the Ilimaussaq intrusion in South-West Greenland and show how the upward migration of low-density, low-viscosity volatile-rich magma delayed crystallization under the roof of the intrusion. Silicic alkaline pyroclastic deposits around central volcanoes in Kenya often show striking variations in the abundance of some incompatible elements within a single vertical section, implying compositional zonation in the magma chamber before eruption. Macdonald shows that these variations are too large to be accounted for by crystal fractionation alone and suggests that some elements have been transported to the magma chamber roof zones as complex ions in a volatile phase. Carbonatites provide perhaps the best illustration of all of the influence of volatile components on the origin and evolution of alkaline rocks. There is now a consensus that their parental magmas originate by the separation of an immiscible carbonate liquid phase from a CO2-saturated nephelinite or phonolite magma. There is, however, some disagreement over the nature and subsequent evolution of this parental carbonate magma. Le Bas argues that the parental magma is rich in alkalis and similar in composition to the natrocarbonatite lavas erupted from Oldoinyo Lengai. This magma evolves at low pressure towards the more common calcite carbonatite (s6vite) by loss of alkalis to the surrounding country rocks which are metasomatized (fenitized) as a result. Twyman & Gittins offer an alternative scheme in which s6vite magmas are parental and natrocarbonatite magmas are derived from them by crystal fractionation. Most petrologists now believe that evolved alkaline magmas are produced by the fractional crystallization of basic magma. The more highly undersaturated parental magmas represented by basanite and nephelinite will produce undersaturated derivatives
xiv
Introduction
such as phonolite and foyaite. Mildly alkaline and transitional basalt m a g m a is likely to produce trachyte and, with extreme fractionation, alkali rhyolite. The production of peralkaline acid rocks by crystal fractionation alone is likely to be an inefficient process. The operation of this process, however, is clearly demonstrated by their association with transitional basalt on some ocean islands, such as Ascension (Harris & Sheppard). Peralkaline acid rocks only occur in a b u n d a n c e in continental environments, however, and here there is often good evidence that crustal contamination has a c c o m p a n i e d crystal fractionation. Well-documented examples of the operation of crustal c o n t a m i n a t i o n in the evolution of alkaline m a g m a s are presented by several of the contributors to this volume. Downes, for example, shows that the assimilation of lower crustal granulite has affected the evolution of alkaline m a g m a s in the F r e n c h Massif Central and uses isotope data to estimate the extent of this contamination. Other examples are presented by Bowden et al. ( N i g e r - N i g e r i a granite ring complexes), Eby (Monteregian Hills and White Mountain provinces, N o r t h America), Fitton (Cameroon line, West Africa) and Fletcher & BeddoeStevens (Velasco province, Bolivia). Other authors propose the derivation of evolved alkaline m a g m a s directly from m e t a s o m a t i z e d mantle or lower crust (Bailey; Woolley &
Jones). Our understanding of the origin and evolution of alkaline m a g m a s has come a long way since the publication of Sorensen's book in 1974, largely through the acquisition of a far larger geochemical and isotopic data base. The contributions to this volume review the current state of this understanding. Emphasis has shifted from crustal to mantle processes with the recognition of mantle m e t a s o m a t i s m and its possible role as a precursor to alkaline magmatism. More recently, though, there has been a swing towards the opposite view, that mantle m e t a s o m a t i s m is c a u s e d b y alkaline m a g m a t i s m . Theoretical and experimental studies on the migration and segregation of small-degree melts seem destined to accelerate this swing. Despite these advances, however, m a n y mysteries remain unsolved and alkaline rocks will still provide a fruitful field of research for m a n y years to come, yielding further insights into the nature of mantle processes and the evolution of magmas.
References EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In Thorpe, R. S. (ed.) Andesites. Pp. 25-87. John Wiley & Sons, London. GILL,J. B. 1981. Orogenic Andesites and Plate Tectonics, 390 pp. Springer-Verlag, Berlin. LARSEN,L. M., REX,D. C. & SECHER,K. 1983. The age of carbonatites, kimberlites and lamprophyres from southern west Greenland: recurrent alkaline magmatism during 2500 million years. Lithos 16, 215-21. LE BAS,M. J., LE MAITRE, R. W., STRECKEISEN,A. & ZANETTIN, B. 1986. A chemical classification of
volcanic rocks based on the total alkali--silica diagram. J. Petrol. 27, 745-50. MCKENZIE, D. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-91. MITCHELL,R. H. 1976. Potassium-argon geochronology of the Poohbah Lake alkaline complex, northwestern Ontario. Can. J. Earth Sci. 13, 1456-9. SORENSEN,H. (ed.) 1974. The Alkaline Rocks. 622 pp. John Wiley & Sons, London. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks. N. Jb. Miner. Abh. 107, 144-240. - - , 1980. Classificationand nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Geol. Rundschau, 69, 194-207.
J. G. FITTON& B. G. J. UPTON,Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Mantle metasomatism
perspective and prospect
D. K. Bailey S U M M A R Y : Mantle replenishment in lithophile elements has been discerned in the patterns of trace elements and isotopes in lavas. One replenishment process is identified as metasomatic replacement, seen in ultramafic xenoliths brought up in high-velocity alkaline eruptions. Thus alkaline magmatism provides the best primafacie evidence of metasomatism and open-system conditions in the upper mantle. The list of added lithophile elements includes the following: H, C, F, Na, A1, P, S, C1, K, Ca, Ti, Fe, Rb, Y, Zr, Nb, Ba and rare earths. Some metasomatism may be due to wall-rock alteration near magma bodies, but the evidence for metasomatism prior to melting opens the possibility that the process is a precursor to alkaline magmatism, giving the necessary source enrichment in lithophile elements. In some igneous provinces the metasomafism is widespread, intensive and pervasive; in others it appears as veining of variable intensity. Metasomatism as a largescale process is best indicated by the widespread distribution of alkaline magmatism in space and time: volatile flux through the lithosphere would then be the necessary precursor of metasomatism and magmatism. Volatile activity, metasomatism and melt enrichment clearly widen the scope for mafic magma generation in the mantle, but some long-standing problems of the alkaline associations (and indeed the calc-alkaline) also call for re-examination in terms of volatile activity in the lithosphere mantle. These include the diversity of magmas, the generation of large felsic volumes, and composition gaps in magma series. Experiments show that felsic minerals are stable to 30 kb, indicating the possibility of felsic-melt generation in the upper part of the mantle. A combination of volatile flux and melt percolation along geotherms that intersect the solidus at depths of less than 80 km would lead to enrichmentand metasomatism, providing distinct mantle sources for felsic magmas. Initial (or residual) melts from such a region, as distinct from those from greater depths, would be constrained by equilibria involving felsic minerals. Thus an igneous cycle could generate a bimodal association, with felsic melts forming in the upper regime and the mafic melts originating at depths below the range of felsic-mineral stabilities. Such a magma system is consistent with the observed eruptive characteristics, explains the typical ultramafic nodule and megacryst suites in the alkali olivine-basalt association and is free of difficulties with relative volumes of melts, with eruption timing and with rapid changes in erupted compositions.
Perspective From its Greek origins the term metasomatism should imply a 'change of body' or a change of substance. It must indicate a chemical change in a pre-existing rock or mineral, and traditionally has been used to make the distinction from metamorphism ('change of form') indicating reconstitution without chemical introduction. Unfortunately, early definitions of metasomatism are not rigorous, but common usage applies to cases where material has been transferred through a vapour or fluid without melting. More recently, in reference to the mantle, the usage has widened considerably, sometimes referring to simple melt infiltration, and even unspecified enrichment processes of a pre-existing mantle composition. Obviously the latter usages are not justified but there is a boundary problem. W h e n melt is introduced into a rock it may react with its surroundings and metasomatism may correctly describe the alteration process in the wall-rocks;
this borderline condition is perhaps part of the reason for some of the current ambiguity. There is at present a need for general agreement about terminology and it would certainly help to retain scientific precision if 'metasomatism' could be restricted to those cases where there is petrographic evidence of replacement of a pre-existing rock. Certainly the term should not be used to describe simple melt injection, producing a hybrid mixed rock, nor when there is no petrographic evidence that there has been replacement of an earlier mineralogy. In other words, the term should be applied only when there is unequivocal evidence of the previous substance--otherwise, to say that a rock has changed its chemistry is supposition. If there is a case for chemical introduction but the process is unclear or unknown, it would be better, and more straightforward, to use the term enrichment. Evidence for metasomatism in the mantle has
From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 1-13.
2
D.K.
been steadily accumulating during the past 10 years (Harte et al. 1975; Lloyd & Bailey 1975) and the topic has provoked such interest that some surveys and reviews have already appeared (e.g. Boettcher & O'Neil 1980, Introduction and tabulation; Bailey 1982). The cited examples are relatively succinct and in some ways complementary, and may be taken as constituting a comprehensive starting point for the following discussion which aims to sketch an overall perspective to the subject. Occurrence and distribution Mantle samples showing evidence of metasomatism are provided in ultramafic nodules carried to the surface by high-speed volcanic eruptions. Some peridotite massifs (e.g. Lherz) contain veins of less refractory minerals, while others also contain evenly distributed minerals such as phlogopite and amphibole (e.g. Finero). These are sometimes cited as examples of 'enriched' and even metasomatized mantle but the evidence that such minerals have been introduced from an external source while the peridotite was part of the mantle has still to be established. Even in volcanically derived mantle fragments there may still be problems in distinguishing reaction products produced after incorporation in the magma; various criteria can be applied, however, and it is clear that in many cases minerals were introduced into the peridotite prior to eruption (Bailey 1982). Mantle nodules showing signs of metasomatism are brought to the surface exclusively by alkaline activity, especially in ultramafic and mafic eruptions. These range from kimberlites, which are characteristically fragmental, through to magmas in the alkali olivine-basalt association (basanites through to mugearites) which may occasionally carry ultramafic nodules. Hence, in terms of distribution, samples may be found in any parts of the stable plates, both oceanic and continental, where there has been alkaline igneous activity; such activity is usually associated with uplift and dislocation, in other words with some kind of lesion in the lithosphere. Compositions Metasomatism is recognized when there is replacement of a pre-existing peridotite mineralogy, new minerals being characterized by their content of mobile and volatile elements, as listed in Table 1. Introduction of the listed minerals into garnet or spinel peridotite (generally taken as representing average mantle composition) signifies mantle enrichment in lithophile ele-
Bailey TABLE 1. Lithophile-element-bearing m&erals in peridotite xenoliths which either may show metasomatic replacement of pre-existing peridotite mineralogy or are associated with metasomatism a (Major) b (Minor)
biotite, amphibole, clinopyroxene, carbonate phosphate, titanates, oxides, sulphides
Possible candidates for higher-level metasomatic minerals c (< 25 kb) feldspathoids, alkali feldspars ments. Most descriptions of mantle metasomatism deduce the introduction of some or all of the following elements: H, C, F, Na, A1, P, S, C1, K, Ca, Ti, Fe, Rb, Y, Zr, Nb, Ba, rare earths. If metasomatism were wholly a replacement process then there should be subtraction of equivalent material, balancing the introduction of the above elements. As yet this aspect has received little attention and this may be because the fabric of the rocks showing metasomatism is characteristically veined, with metasomatism proceeding along grain boundaries and fissures: hence the introduced material could be largely accommodated by an increase in rock volume, with redistribution rather than removal of the displaced elements. A parallel case may be seen in alkaline metasomatism in the crust (fenitization) which is characteristically marked by the introduction of new minerals along a close network of cracks. Timing In many cases the indications are that the metasomatic introduction of new minerals, although it must have preceded eruption, is part of the cycle of igneous activity, and there is isotopic harmony between the nodules and the magmas. In some instances there is evidence of an earlier metasomatic event (Erlank & Shimizu 1977; Menzies & Murthy 1980a) and there are cases of possible complex metasomatism and/or enrichment dating as far back as 3 Ga (Erlank et al. 1980; Menzies & Murthy 1980b). No doubt the picture will become still more complex as new data become available; this may be expected because alkaline activity has often been repeated through the same segment of lithosphere (Bailey 1977) and even if metasomatism were nothing more than a minor part of igneous activity there should be samples showing complex histories. Conditions of metasomatism Using a combination of experimentally determined mineral stabilities and solidi it is possible
Mantle metasomatism
3
to put limits on the pressures and temperatures of metasomatism, and hence on the conditions in the source mantle. Figs 1 and 2 summarize the conditions for stability of phlogopite, amphibole, clinopyroxene and felsic minerals in the mantle, and by reference to Fig. 4 the relative stability of carbonates can readily be envisaged. Obviously the diagrams must be generalizations because the individual mineral stabilities will depend on bulk composition (mineral and rock), coexistence with other metasomatic minerals, and the presence or absence of vapour, fluid or melt. In spite of these reservations the diagrams offer useful limits and indicate the following relationships. 1 Amphibole is stable near the peridotite solidus only in the upper part of the mantle, its lower boundary corresponding approximately to that of spinel and showing broad equivalence to the lower stability boundaries of feldspars and feldspathoids and the upper stability boundary of carbonate (see Figs 1, 2 and 4). 2 Phlogopite stability extends to higher temperatures and greater depths than amphibole, its
FIG. 2. P-T relationships of mineral stabilities and solidi, and geotherms. Solidi: as in Fig. 1 with the addition of WE, the solidus in the system KA1SiO4MgO-SiO2-H20-CO2 (Wendlandt & Eggler 1980). WE is used to provide a measure of phlogopite stability in peridotite mantle; its extrapolation to intersect KS at point P indicates a depth limit for phlogopite in the presence of melt (this is similar to the boundary given by Wyllie (1979)). Geotherms: as in Fig. 1. Mineral stabilities: C, coesite: Ks, kalsilite; AB, albite + nepheline; (AB), albite (jadeite + quartz); OR, K-feldspar; SC, solvus crest for alkali feldspars. stability limit at the solidus approximately coinciding with the inflexion region in some kimberlite geotherms (about 180km). Phlogopitebearing nodules or lavas from greater depths would therefore seem to be ruled out (Bailey 1986). 3 Clinopyroxene stability can extend up to the vapour-absent solidus and this mineral is arguably the most important but least appreciated metasomatic mineral.
FIG. 1. P-Trelationships of solidi, and geotherms. Solidi : PSD, peridotite vapour-absent solidus; KS, kimberlite solidus (Eggler & Wendlandt 1978); OE, peridotite solidus in the presence of H20 and CO2 (limited) (Olafsson & Eggler 1983). For convenience the stability region of spinel peridotite, with or without amphibole, is shown here rather than in Fig. 2. D is the diamond stability boundary (Kennedy & Kennedy 1976). Geotherms: S and O, shield and ocean (Clark & Ringwood 1964); 180, oceanic lithosphere 180 Ma old (Sclater et al. 1980); 30, oceanic lithosphere 30 Ma old (Oldenburg 1981). If geotherms should converge more rapidly in the depth range around 200 km as proposed by Tozer (1967) the geometry of the melt systems would change but the principles would remain the same.
From the above it may be deduced that carbonate and clinopyroxene may appear in any metasomatic regime, but the geological and experimental evidence suggests that high-pressure, low-temperature conditions (e.g. kimberlitic) favour carbonates while low-pressure hightemperature conditions (basanitic) favour clinopyroxene (see Fig. 5). Consideration of the magmas and their nodule suites also points to the fact that ultramafic alkaline magmas such as kimberlites, melilitites and nephelinites are characterized by phlogopite (in melts and xenoliths) while other magmas such as some nephelinites, but especially basanites, are characterized by xenoliths containing amphibole. Basanites in particular are likely to carry ultramafic nodules containing both spinel and amphibole, which
4
D.K. Bailey
may be an indicator of a generally shallower source depth for the nodules compared with ultramafic melts.
Outstanding issues While the reality of mantle metasomatism seems to have found wide acceptance, some differences may be perceived concerning such questions as the following: 1 Which comes first, the magmatism or the metasomatism? 2 Is metasomatism always a local phenomenon (in the vicinity of alkaline intrusions) or can it develop regionally? 3 What is the source of materials erupted at the surface, e.g. are they from the lithosphere or the asthenosphere ? The last question seems irresolvable until the nature of the asthenosphere can be unequivocally defined. At the moment, the ascription of a rock to an asthenosphere source (on the basis of selected aspects of its trace-element or isotope chemistry) is a convoluted way of saying that it has some attributes of mid-ocean ridge basalt; it would be wholly inappropriate to digress into this minefield. The first two questions are related, and controversy on these issues seems likely to be fruitless. Obviously, samples showing recognizable metasomatism can do so only on a local scale, and because volcanically transported fragments of metasomatized mantle must be small the evidence for metasomatism on a large scale must be sought in other ways. Appropriate lines of evidence, indicating large-scale metasomatism, have been described elsewhere (Bailey 1982) but may be summarized as follows: (a) an ultramafic nodule population dominated by metasomatized fragments throughout an igneous province; (b) a range of samples showing various degrees of metasomatism through to completely transformed nodules (alkali clinopyroxenites); and (c) the chemical equivalence of extensively metasomatized nodules and erupted melt compositions. Much of the cogency of the case for large-scale metasomatism is lost if discussion is allowed to focus just on laboratory data: it is then easy to lose sight of the wider geological perspective, where alkali- and volatile-rich magmatism has to be seen in the context of the geothermal and tectonic conditions of stable plates (Bailey 1983). This larger-scale (geological) evidence must be taken together with that of metasomatism on a small scale if we are to retain a balanced view. Both kinds of evidence exist (O'Reilly & Griffin 1984; Wilshire 1984) and they are not mutually exclusive. No amount of evidence or
argument in favour of one can, of itself, falsify the case for the other. Good petrographic evidence of local metasomatism is obviously immune to rebuttal, and no evidence has yet emerged to reject the case for regional development of the process; until such evidence emerges it is important that the argument should not become polarized (as in some older geological disputes) lest we end up with a controversy without roots in nature. The dilemma of the relationships between melting and metasomatism has been highlighted by Hawkesworth et al. (1984), who distinguish two enrichment processes, one typified by subsolidus metasomatism in kimberlite nodules and the other typified by injection of small-volume melts (typically associated with basanitic magmatism). These are essentially expressions of the effects of pressure and temperature, and it will be shown that they may be explained by the interplay of geothermal gradient, mantle solidus and mineral stabilities. It will be seen there that subsolidus metasomatism can also play a vital role in basanitic magmatism. Although it is valuable to identify the chemical characteristics of the two types of enrichment (Hawkesworth et al. 1984) it should be remembered that there is a spectrum of alkaline magmatism between kimberlites and basanites, and a spectrum of enrichment processes may be expected (see Fig. 5 and Bailey 1986).
Geological factors If all the observed metasomatism were a localized consequence of alkaline igneous activity, then the quantities of metasomatized mantle and the petrological significance of the process would be essentially trivial. It is when the geological picture is seen in total that such a conclusion becomes most suspect. Association with alkaline magmatism signifies the high activities of alkalis, mobile elements and volatiles, and the high-velocity eruptions that bring the samples to the surface emphasize the role of gases in this type of magmatism. When the tectonic framework, especially the connection between alkaline magmatism and lesions in stable lithosphere (and the repetition through time), is taken into account, it is difficult to escape the conclusion that we are observing a process in which volatiles must play a vital role. Alkali rich means volatile rich and, although there may be room for debate about whether the metasomatism is a precursor to magmatism in any particular instance, there can be little doubt that volatile migration may have far reaching effects without the necessary intervention of magma. One of the most graphic examples is in kimberlite peridotite
Mantle metasomatism nodules containing phlogopite: these were sampled from points on the geotherm well below the kimberlite solidus (Bailey 1982, Fig. 3) and hence well outside conditions where any melt could exist. Certainly there are strong grounds for supposing that volatile influx is a natural precursor to both metasomatism and alkaline magmatism.
Prospect Given an outline of present concepts of mantle metasomatism it becomes fruitful to look beyond and to consider other consequences of volatile movement through the mantle. One aspect, of immediate relevance to a volume on Alkaline Rocks, has been largely neglected in discussion of mantle metasomatism: this concerns the broader spectrum embracing not just mafic and ultramafic compositions but felsic alkaline magmas. At what has been described as the limiting case of cratonic magmatism (Bailey 1980a) there is kimberlite with no felsic associates, and closely related must be the lamproites and melilitites with very limited, if not uncertain, felsic connections. These and the ultramafic lamprophyres share other common features, such as style of eruption, that require special consideration in terms of petrogenesis; they have been discussed elsewhere (Bailey 1986). The more voluminous expressions of alkaline mafic magmatism are essentially basanitic (having some connections with nephelinites) grading through to transitional basalts, and these typically form major volcano complexes and have associated felsic magmas. Such alkali basalt magmas show evidence of mantle enrichment and characteristically may carry mantle nodules, which themselves show signs of metasomatism and/or enrichment. Their mantle source is hardly in doubt, and the possibility that this source may have undergone metasomatism raises the question whether the felsic magmas could be part of the same process. Some of the problems of alkaline felsic magmatism, such as regional magma development and comparative volumes, could be resolved in the context of regional metasomatic processes (Bailey 1972, 1974) and there have been no data on relevant mineral stabilities to exclude the possibility of sources in the upper mantle (Bailey 1976); more recent data lend credence to this possibility and allow a more penetrating look at the question. Before embarking on an examination of the more recent experimental information it is useful to look over the case in geological terms. The popular view for a long time has been that
5
felsic alkaline rocks are products of differentiation from mafic parents at relatively shallow (usually crustal) depths. Processes such as fractional crystallization are still currently in vogue in spite of many difficulties (see for instance Yoder 1979; Bailey 1981) and discussions of felsic magmas in particular seem impervious to contrary evidence. When the various pieces of evidence are assembled together, however, the case for low-pressure differentiation as a universal mechanism for producing felsic magmas can be seen to be without foundation. Essentially, evidence contrary to continuous differentiation at low pressure takes two forms: (1) felsic and intermediate volcanics carry ultramafic nodules; and (2) there are substantial composition and volume discrepancies in the igneous products of a given complex or province. As will be seen later, the first could be considered as the evidence from the high-energy ranges of the magmatic system and the second from the lower-energy ranges. Nodules
In addition to the dramatic examples of felsic lavas carrying ultramafic nodules (Wright 1966, 1969), mugearites and hawaiites are commonly reported as the hosts in ultramafic-nodule localities (see Boettcher & O'Neil (1980) for an informative listing). High-speed eruption is essential, and these are unequivocal examples of the existence of felsic and intermediate melts in the mantle: it is clear that low pressure differentiation is not essential for the formation of magmas in the range hawaiite-trachyte-phonolite. In fact, the composition of these melts may be a direct indication of physical and chemical conditions in the mantle source. Composition gaps
Daly (1925, 1927) first drew attention to the basalt-trachyte bimodality of oceanic lavas, a finding later supported by Barth (Barth et al. 1939) and statistically verified by Chayes (1963) who suggested that this raised doubts about the evolution of trachyte from basalt by continuous differentiation. Quite naturally, this suggestion provoked spirited opposition from advocates of fractional crystallization; the case was reviewed by Yoder (1973) who concluded that the gap was real and that this was not a realistic outcome of continuous fractional crystallization. More recent studies (e.g. Zielinski 1975) still invoke this mechanism, however, without apparent question, and so it is permissible to cite some additional geological facts.
6
D.K. Bailey
1 In deeply exposed continental sections revealing syenite plutons, comparable alkaline mafic and intermediate plutons are rare or absent. 2 Zoned plutons are rare, and where known are relatively small and generally considered to be composite (e.g. Monteregian Hills). 3 In keeping with 1 and 2, the common nodules in felsic volcanics are typically cognate, e.g. syenite in trachyte, peralkaline granite in comendite. To the above it may be added that the typical nodules in basanites are peridotite and gabbro. Hence it is hardly reasonable to argue that the basalt-trachyte bimodality in the ocean basins is an artefact of exposure level, eruption characteristics or sampling bias: the continental plutonic bodies and the nodule suites confirm the scarcity of intermediate magmas. Perhaps the most vivid evidence of magmatic bimodality lies on Graciosa in the Azores. Here there have been alternating eruptions of felsic and mafic pyroclastics, without detectable time breaks, requiring the simultaneous coexistence of contrasting magmas within the one volcanic system (Maund & Bailey 1982; Maund, 1985). This is simply a graphic example of the contemporaneous eruption of trachyte and basalt on oceanic islands generally, to which Daly called attention so long ago. Other composition gaps are present in other magma associations, such as basanite-phonolite and nephelinite-nephelinitic phonolite, but it may be relevant that even among the felsic rocks themselves there seem to be marked distribution maxima, as between phonolites, trachytes and rhyolites (and the plutonic equivalents). These gaps, too, militate against continuous differentiation, tending to favour partial melting or at least multiprocess origins (see Bailey (1976) for a full discussion).
Volumediscrepancies Another major problem for continuous differentiation at low pressure lies in the volume relations of the magmas in many provinces. The usual image of a large alkali basalt centre with small spines and flows of trachyte may be reasonable for some oceanic volcanoes, but it is a dangerous generalization, even for the oceans, e.g. Azores and Canaries. On the continents it is probably the exception rather than the rule: here again we have deep sections providing plutonic evidence. In major syenite provinces, such as Kola and Malawi, the basic and intermediate rocks required by continuous differentiation are not merely insufficient--they are lacking altogether.
Large syenite plutons (and large monotonic trachyte and phonolite volcanoes, as in the Kenya rift) are further evidence against low-pressure evolution of felsic magmas from a basaltic parent. Indeed, to suppose such differentiation has occurred when there is no sign of the earlier stages is to assume that syenite can form in only one way. In the volcanic regime the volume discrepancy can be seen in the Miocene-Recent activity of the East African rift zone where the calculated volumes of felsic and mafic volcanics are approximately equal (Williams 1972) and there have been regional floods of phonolite and trachyte. The latter are difficult to provide for in any process of magma generation but pose an acute problem for high-level differentiation (Bailey 1974, 1978). It should be said, too, that detailed studies of particular trachyte and pantellerite suites have not only failed to find any links with basalt (Bailey 1978) but have even revealed distinctive differences between felsic volcanoes erupting contemporaneously in the same province. In the area around Lake Naivasha in Kenya, there are Holocene phonolite, trachyte, pantellerite, comendite and basalt eruptions from overlapping centres that provide not only a series of major composition gaps but also distinctive major and trace-element patterns that are inexplicable in terms of continuous differentiation (Bailey & Macdonald, 1987). Clearly, there is commonality in that the magmas are part of a cycle of igneous activity, but the need is for a multisource/ multiprocess system of magma genesis. Yoder's solution (1973) to the dilemma of the Daly gap was the production of two contrasting magmas by fractional melting, along lines indicated by Presnall (1969). This envisages a source composition with two invariant points which by continuous melt extraction during an igneous cycle yields two contrasting magmas. Melt generation from two distinct invariant points would clearly alleviate some of the problems of alkaline magmatism and would explain composition gaps, but the generation of large volumes of felsic magma from a peridotite mantle source remains a difficulty, and a new problem is introduced of irreversibility and timing. Either the source volume would have to keep changing or the liquid from the lower-temperature invariant point could be erupted only at the start of a melting cycle. These difficulties could be solved through replenishment of felsic components by metasomatism or other forms of enrichment. At the very least therefore Yoder's solution requires an open system with melting pulses acting on periodically enriched sources (cycles of melting and enrichment). Even so the volume problem remains, and
Mantle metasomatism an additional factor is introduced by alternating mafic-felsic volcanism, exemplified by Graciosa, where there is effectively simultaneous eruption of felsic-mafic magmas poor in phenocrysts. Such activity implies the need for two (or more) distinctly different source compositions to be melted and tapped during the same igneous cycle. Geological evidence indicates that felsic sources richer than normal peridotite are feasible in the mantle and recent experimental evidence confirms that appropriate mineral stabilities extend into the upper part of the mantle. When these are taken together with eruption characteristics, depth indicators in the nodule suites and evidence of metasomatism, it becomes possible to propose a scheme of magma genesis that can reconcile all the seemingly conflicting observations and give an integrated pattern to the activity.
Metasomatism and the development of separate felsic sources Peridotites containing minor amounts of plagioclase, trachytes containing peridotite nodules, and sanidine-coesite eclogite nodules are some of the evidence indicating that felsic melts and sub-solidus felsic mineralogies are possible in the mantle, and it has been clear for some years that there is no a priori reason to suppose otherwise (Bailey 1976). In view of the problems with felsic magmatism, it is appropriate to update the evidence on mineral stabilities and then to enquire whether and how two sources could develop, and how magmas could be generated and erupted. In Figs 1 and 2 some of the relevant stability fields are plotted. Obviously the bulk mantle composition must influence some of these, but broadly speaking felsic minerals could exist to depths of about 100 kin, the exact limit depending also on the prevailing temperature. The special importance of the stability ranges shown in Figs 1 and 2 lies in the fact that the upper mantle within the felsic stability zone has the potential to yield melts of felsic character at a relatively low-temperature invariant point. No such invariant melting point is possible at greater depths, so that the first melts (from a phase assemblage containing no felsic minerals) must of necessity be different, and in all probability distinctly mafic. Thus the possibility of two distinct sources with completely different first melts is evident. Could their potential for producing contrasting magmas be established and enhanced by processes of enrichment and/or metasomatism ? Starting from the critical observation that high
7
volatile activity is a hall-mark of alkaline magnatism, it has been possible to develop a hypothesis relating magmatic variations to melting and metasomatism by volatile flux along different geothermal gradients (Bailey 1970, 1980a). Further consideration of the effects of the processes along low geothermal gradients has been given elsewhere (Bailey 1984, 1986). For the generation of felsic melts, it is necessary to look to activity along steeper geotherms more appropriate to oceanic conditions (or those away from continental craton nuclei). On steeper geotherms, such as those that would cross the solidus at high levels in the mantle (in the felsic stability region), volatiles could migrate along the geotherm only if the channelways were lined (Olafsson & Eggler 1983); melting is then more likely. When the possibility of flux melting along steeper geotherms was discussed previously (Bailey 1980a, 1983) attention was focussed on the eruption of melt through the overlying lithosphere with increasing departure of melt temperature from that of the wall-rocks, i.e. an intrusion, or melt diapir, rising as a detached thermal anomaly. In recent years, however, the concept of migration of very small percentage melts has found increasing favour (Walker et al. 1978; Waft 1980; Stolper et al. 1982). Essentially the melt is envisaged as percolating through the mantle, in which case it should maintain thermal equilibrium with its path, and in the simplest case could be envisaged as migrating along the geothermal gradient. Where the melt is generated by volatile flux, it would effectively extend the path of flux migration. Gradual percolation over a long distance would mean that the melt must also maintain phase equilibrium with the mineralogy through which it moves: most alkaline ultramafic melts are of alkali clinopyroxenite composition (Lloyd & Bailey 1975; Lloyd 1981) and could percolate only along channelways akin to the source mineralogy. Either there must be pervasive distribution of all the required phases or the movement must be restricted to channelways lined with these phases. As the initial melting is seen here as resulting from volatile influx, the pathways are more likely to be enriched by interaction with the melt, which may be contrasted with the scavenging process of melt percolation envisaged by Fitton & Dunlop (1985). In the higher section of the path the percolating melt will, in any case, start to reapproach the solidus and will then progressively precipitate solids such as clinopyroxene and, closer to the solidus, minerals such as amphibole, phlogopite, carbonate, feldspar and feldspathoids. Consequently the melt channels will be further lined or plated with minerals containing lithophile ele-
8
D.K. Bailey
ments. This may lead to occlusion of the percolation holes, requiring changes of course for subsequent percolating melt and thus extending and intensifying an enriched zone in the mantle. At the same time the temperature will be raised by the release of latent heat of fusion so that continuing percolation will act as an effective means of heat transfer and gradually steepen the geothermal gradient. Under conditions of 'equilibrium' percolation, melt will be used up where the geotherm passes back through the solidus, leaving only residual volatiles to continue along the sub-solidus path. These may be expected to produce an intense zone of metasomatism in the immediately overlying lithosphere segment indicated in Fig. 3. Melts or volatiles migrating along steep geotherms will not encounter the stability regions of either phlogopite or amphibole (see Fig. 2) so that most of the trapping of lithophile elements would be in clinopyroxene and/or carbonate until the solidus is approached. Thus, most of the mobile constituents such as Na, K, OH, F and C1 will be concentrated into any surviving lowtemperature melts and fluids, leading to enrichment in felsic minerals and amphibole at nearsolidus conditions. During the early phase of a new igneous cycle along steeper geotherms, therefore, there would be enrichment and meta-
800
IO00
1200
T ~
-,o
5o-
~.:/..-..: ~ u s
I 0 0 --
. ..........
I
"~" ~.~ : ~ ACCUMULATION " 0.710) isotope ratios as a means of identification of possible source regions (Powell & Bell 1974; Rock 1976). Any similarity between isotope ratios measured in alkaline rocks and mid-ocean ridge basalts (MORBs) or ocean island basalts (OIBs) was taken as unequivocal proof that the particular rock in question was mantle derived (e.g. East African rift). Similarly, any alkaline rock with an isotopic composition similar to crustal values was assumed to be a crustal derivative or to have experienced crustal contamination (e.g. Kimberley, Australia). For example, Rock (1976) commented that in certain alkaline rocks the Sr isotopic ratios were 'so high (>0.710) that a significant contribution from crustal material was inescapable'. It is implicit in these models that interaction with a 'crustal' component occurred after magma
formation. More recently, alkaline rocks with radiogenic Sr (and non-radiogenic Nd) isotopic ratios have been interpreted as extracts from mantle source regions where addition of a crustal component to the mantle occurred prior to magma production (Chase 1981; Hofmann & White 1982; Zindler & Jagoutz 1987) by recycling processes involving subduction and/or delamination of aged sub-continental mantle (McKenzie & O'Nions 1983). Alkaline rocks may include components from the following:
1 Lower mantle." below the boundary layer (about 670 kin). 2 Upper mantle." oceanic lithosphere; continental lithosphere; convecting asthenosphere. 3 Recycled components." oceanic and continental crust and mantle. In this review isotopic data pertinent to inclusion-bearing alkaline rocks and their cargo of mantle inclusions erupted in ocean basins, continental rift valleys and stable cratonic regions will be used to construct an Earth model. Earth models will be considered where the structural elements that form the upper mantle below both oceans and continents are lithosphere (crust and mantle) and asthenosphere. These in turn are underlain by the lower mantle below the 670 km boundary layer.
Alkaline rocks and their mantle inclusions Random fragments of mantle are transported to the surface in alkaline rocks. The inclusions entrained in alkaline rocks range from garnet
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 15-27.
15
M. Menzies
[6
and spinel peridotites to mica clinopyroxenites, glimmerites, apatite-amphibole pyroxenites and mica-rutile-ilmenite-diopside (MARID) rocks. It is useful to think of the variety of inclusions as disrupted mantle pegmatites viz. pieces of mantle wall-rock (i.e. peridotites) and fragmented conduits of silicate melts or aqueous fluids (i.e. pyroxenites, glimmerites etc.). Analogues can be found in inclusion suites for the various melts or fluids that have migrated through the mantle. A m p h i b o l e - a p a t i t e - p y r o x e n ites found as veins in spinel lherzolite inclusions are believed to be a product of silicate-melt migration (Hawkesworth et al. 1984; Menzies et al. 1985c). A by-product of this magmatism is the Fe-Ti metasomatism observed in wall-rock peridotites. In other words, upwelling basanitic or nephelinitic melts may have crystallized in the mantle as these vein networks. It can be shown that their isotopic and trace-element characteristics are consistent with such an interpretation (Irving and Frey 1984; Roden et al. 1984). M A R I D rocks (Kramers et al. 1983; Smith 1983) found in kimberlite inclusion suites are believed to be portions of the conduit lining left after passage of a hydrous fluid enriched with incompatible elements. Glimmerites and micaceous pyroxenites (Erlank et al. 1987; Menzies et al. 1987) may also represent disrupted vein systems formed by crystallization of LIL- (LIL stands for large
05132 ~ N I V A K ~ " 1 0
o.51~L__~ 0.702
~
Ocean basins Before any discussion of the isotopic variability of alkaline basaltic and kimberlitic rocks erupted in continental regions, it would seem logical to review the data on alkaline rocks from ocean basins (Fig. 1). Although not strictly oceanic, the Nunivak data are included in Fig. 1. Alkaline rocks associated with disruption and entrainment of fragments of oceanic mantle have a very narrow range of Sr and Nd isotopic ratios (i.e. STSr/86Sr < 0.706 and 143Nd/144Nd > 0.51270 (Fig. 1, Table 1)). These magmas are clearly derived from mantle depleted in the light REE for a considerable period of time. A concomitant depletion in Rb is also apparent, but in the case of Tahiti and Malaita the magmas have slightly higher SVSr/86Sr ratios. This could result from (a) mixing of MORB mantle with a recycled radiogenic component or (b) influx of a high Rb/Sr fluid.
0.513211
j-5
|
I
TAHITII|
|A / ~10
|
I
07.02
i
I
0.704
I
0.706
Inclusions
Host Magmas
0.513211_
'
0.51~4t-,
0.706
0.704
(~
ion lithophile elements) and LREE-enriched hydrous fluids in the mantle (LREE stands for light rare-earth elements). Mantle inclusions provide a very important record, not only of ancient depletion events (more than 103 Ma ago) related to crust formation, but more importantly of the passage of silicate melts and hydrous fluids from the mantle to the surface.
3 0.5132~,inlcl "-
~
i
HAWAllli
+10
-5 "O 0.5128 z
Z-O 9-
0.5128
_
9
0.5124 I 0.702 87Sr/86S
I
I 0.704
I
I 0.706
o
0.5124
-I 0.702
1
I 0704
I
I 0.706
-5
-o Z
uo
r
FIG. 1. Nd versus Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basaltic rocks from oceanic basins. Note the following: (i) the host magmas are derived from mantle depleted in Rb and light REE; (ii) the inclusions exhibit a greater range of Nd and Sr isotopes than the host magmas: (iii) the high 8VSr/a6Srratios in the Malaita and Tahiti inclusions. The vertical and horizontal reference lines in this and other diagrams define the generally accepted values for undifferentiated or chondritic mantle or bulk earth (87Sr/S6Sr= 0.7045 and 143Nd/144Nd=0.51264). Data are taken from several sources. Nunivak : Menzies & Murthy (1980a); Roden et al. (1984a). Malaita: Bielski-Zyskind et al. (1984). Tahiti: Vidal et al. (1984); Menzies & Hawkesworth (1987). Hawaii: Chert & Frey (1983).
Alkaline rocks and their inclusions
I7
TABLE 1.87Sr/S6Sr, 143Nd/144 Nd, R b / S r a n d S m / N d ratios f o r inclusion-bearing alkaline, kimberlitic a n d lamproitic m a g m a s
Locality
87Rb/86Sr
esra
/;Ndb
147Sm/
Ref
144Nd Afar (Assab) - 17 to - 10.0 -Africa (Tanzania) - 14.2 -Alaska (Nunivak) - 2 8 . 4 to - 17 0.049-0.121 Antarctica Foster Crater - 13.8 to - 11.4 - Ross Island - 19.2 0.089-0.165 Arizona Geronimo - 23.4 to - 17 0.118-0.240 San Carlos - 22.7 0.091 Australia Kiama - 7.1 0.051-0.164 Kimberley + 85 to + 218.6 0.593-9.855 Victoria - 28.4 to + 4.3 -California (Sierran Province) c + 4.3 to + 34.1 0.056-1.28 Fiji - 13.4 -France (Massif Central) - 14.2 to 0 0.153-1.225 French Polynesia (Tahiti) d + 7.1 to + 34.1 -Hawaii (Honolulu Series) - 18.5 to - 17.03 0.015-0.103 Scotland (Midland Valley and N W High- - 4.3 to + 25.6 0.077-0.203 lands) Solomon Islands (Malaita) + 1.4 to + 2.8 0.059-0.183 South Africa (various localities) Group I kimberlites - 11.4 to + 14.2 0.038-1.52 Group II kimberlites + 41.2 to + 90.8 0.140-0.786 South Yemen (Ataq) - 17.0 to - 2.8 0.028-0.692 Mantle xenoliths Keel e - 3 5 . 5 t o +1867.0-Asthenosphere -35.5to+21.3 --
+4.7 to +7.2 + 1.4 + 4 . 9 t o +7.8
-0.096 0.119-0.133
+3.3
--
to
+4.7
1 2 3
--
--
4
+ 4 . 9 t o +8.0 +7.6
0.101-0.126 0.130
5 6
- 0 . 6 t o +0.4 - 16.2 to - 8 . 0 - 4 . 7 to +5.0 - 9 . 6 to +2.9
0.116-0.120 0.065-0.082 0.112-0.119 0.103-0.140
+3.7
--
+ 7 . 0 t o +7.6 - 6 . 6 t o +3.1
0.120-0.168 0.068-0.108
7 8 9 10 11 12 13 14 15
+ l . 9 t o +2.9
0.102-0.121
16
- 4 . 1 to +1.0 - 12.9 to - 8.8 - 2 . 3 to +8.0
0.085-0.098 0.062-0.087 0.100-0.130
17 18
- 2 4 . 2 t o +14.8 -4.7to+14.8
---
19
+2.1 to +4.1
0.087-0.119
-
--
1.9
to
+2.5
"es~ represents the present-day 87Sr/86Sr ratio assuming a bulk earth s 7Sr/86Sr of 0.7045. beyo represents the present-day 143Nd/144Nd ratio assuming a bulk earth 143Nd/144Nd of 0.51264. cRange in Rb/Sr and Sm/Nd given for the Sierran Province lavas are for basaltic rocks that are not necessarily host to the xenoliths. The isotope data and trace-element data are for different samples. aTahitian samples are not the host magmas to the xenoliths. eThe range given for the 'keel' inclusions excludes inclusions found in diamonds (Richardson et al. 1984). 1 2 3 4 5 6 7 8 9 10
Betton & Civetta 1984; Cohen et al. 1984; Menzies & Murthy 1980a; Roden et al. 1984a; Stuckless & Ericksen 1976; Menzies et al 1985b; Menzies et al. 1985a; Evans & Nash unpublished data; Zindler & Jagoutz 1987; Wass & Rogers 1980; Menzies & Wass 1983; McCulloch et al. 1983 ; Frey & Green 1974; McDonough & McCulloch 1985; Van Kooten etal 1985;
T h e spinel a n d g a r n e t p e r i d o t i t e s a n d p y r o x e n i t e s e n t r a i n e d b y t h e s e m a g m a s (Fig. 1) h a v e a s i m i l a r r a n g e o f Sr a n d N d isotopes e x c e p t t h a t t h e i n c l u s i o n s do r e c o r d s o m e h i g h e r 87Sr/S6Sr ratios (e.g. M a l a i t a a n d Tahiti). It is v e r y difficult to state u n e q u i v o c a l l y t h a t t h e c h e m i c a l c h a r a c t e r istics o f t h e inclusions are n o t a n a d j u n c t o f
11 12 13 14 15 16 17 18 19
Gill 1984; Chauvel & Jahn 1984; Downes 1984; White & Hofmann 1982; Clague & Frey 1982; Chen & Frey 1983; Stille et al. 1983; Thirlwall 1982; Menzies & Halliday 1984; Bielski-Zyskind et al. 1984; Smith 1983; Menzies & Murthy 1980a; Menzies et al. 1987, and references cited therein; Menzies & Hawkesworth 1987.
m a g m a t i c processes, i.e. r e p e a t e d cycles o f m a g m a t i s m in t h e m a n t l e . R e g a r d l e s s o f this d i l e m m a , the basalt and inclusion data reveal mantle r e g i o n s d e p l e t e d in t h e light R E E a n d v a r i a b l y d e p l e t e d or e n r i c h e d in Rb. T h e latter e n r i c h m e n t m u s t be a r e l a t i v e l y r e c e n t p h e n o m e n o n (less t h a n 0.2 x 103 M a ago) w i t h i n t h e l i f e t i m e o f t h e
18
M. Menzies I
I
0.5132
I
I
I
t
/
MIDDLE EASTt!10 (Jd~ ~L'T-----":~ ~?-al ,.? --ETHIOPIA
I..~--~ I iFRANCE i I&4+1o /
0.5132~.~ , ~ I~-
~
GERMANY|
"
0.5128 0.5124 1 0.702
I
0.704
,'
l
o.51241-,
I
,
0.702
0,706 I I I -]40
I ~
F,
,
0.704 ~
'
I '
, 4-5
0.706 ' |
sw u
~ 0,5124[1
0.702
l
0704
0.5130
l
; 1-s
0706
TANZANIA
1
L I
I
-
0.5124 [-1 I 0.702 5
,o
-25 t
0.750
t
0.800
I ~ I 0704
2 I ~ -5 0.706
Magmas
-15 -20
~
II I
0.700
0 . 5 1 2 8 ~
[ 1_0
. . . .
F 05110t
o
L
0.850
-30
(j
Inclusions z
%9
87Sr/86Sr
FIG. 2. Nd versus Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basalts from rifted continental regions. Note the following : (i) the continental host magmas are derived from source regions not unlike the oceanic magmas in Fig. 1 ; (ii) in the majority of cases the host magmas and inclusions are derived from different mantle source regions; (iii) in the case of the African rift valley, fragments of heterogeneous mantle have been entrained. Data are taken from several sources. Middle East-Ethiopia: Menzies & Murthy (1980a); Betton & Civetta (1984); Menzies & Hawkesworth (1987). France and Germany: Stosch et al. (1980); Chauvel & Jahn (1983); Downes (1984). Eastern Australia: Menzies & Wass (1983). SW U.S.A. : Jagoutz et al. (1980); Stosch et al. (1980); Menzies & Hawkesworth (1987); Menzies et al. (1985a); Van Kooten et al. (1985); Zindler & Jagoutz (1987). ocean basin. It can be deduced from these data that the source regions of alkaline magmas and of the overlying mantle as represented by the inclusions are isotopically quite similar. Exposures of peridotite in the ocean basins (Roden et al. 1984b) exhibit isotopic variability that falls within the limits of the data in Fig. 1. Roden et al. (1984b) believe that the St Paul's rocks contain evidence of a recent (less than 0.2 • 103 Ma ago) trace-element enrichment related to alkaline magmatism.
Continental rift valleys Inclusion-bearing alkaline rocks erupted in the major rift valleys of the world have present-day Sr and Nd isotopic ratios not unlike those of the inclusion-bearing alkaline rocks from ocean basins. 87Sr/S6Sr ranges from 0.7027 to 0.7055, and 143Nd/144Nd ranges from 0.51305 to 0.51245 (Table 1). Interestingly, however, the marked disposition of the oceanic data towards high 87Sr/ S6Sr ratios is not evident in the continental
alkaline rocks. In fact, these data are more inclined to plot towards lower 143Nd/144Nd ratios (Fig. 2). Again, the alkaline host magmas are derived from a mantle source depleted in Rb and the light REE for a considerable period of time. Modifications to this mantle source have clearly occurred and account for the scatter of the data. For example, alkaline rocks from the Middle East (Betton & Civetta 1984; Menzies & Hawkesworth 1987) and SW U.S.A. (Van Kooten et al. 1985) plot towards more radiogenic 87Sr/86Sr ratios, a fact that indicates input of Rb into the source region or mixing of end-members with diverse 8VSr/86Sr ratios. The variety of mafic and ultramafic inclusions entrained in alkaline magmas erupted in rift valleys exposes a complex mantle structure that is at first glance extremely heterogenous for both Sr and Nd isotopes and is thus distinct from oceanic mantle. It is apparent, however, that the majority of inclusions have 87Sr/86Sr ratios ranging from 0.702 to 0.7065 and t43Nd/t44Nd ratios ranging from 0.51330 to 0.51240. Considering the paucity of data from ocean basins and the prolific number of isotopic
Alkaline
rocks and their inclusions
analyses from continental rift valleys, these data are only marginally different from oceanic-basin data. However, isotopic data from Tanzania appears to complicate matters (Fig. 2) since the inclusions display an extreme range in Sr and Nd isotopic composition. Nothing of this type has been seen in the ocean basins or in most of the continental rift valleys (compare Figs 1 and 2). These extremely radiogenic Sr and non-radiogenic Nd isotopic ratios record ancient (more than 103 Ma ago) trace-element enrichments involving addition of the light R E E and/or Rb to the mantle. With the passage of time, the isotopic ratios have adjusted so as to record the modified
0.5120
~
/
19
trace-element ratios. It is unlikely that such old trace-element enrichment events would be preserved in actively convecting asthenosphere, A mantle layer with fundamentally different isotopic characteristics clearly exists beneath the continents and not beneath the ocean basins. Furthermore, the similarity between oceanic and continental rift valley alkaline basalts points to a similar mantle source at depth for the host magmas. While most of the inclusions record an old depletion in both Rb and the light REE not unlike the oceanic mantle, certain fragments of 'anomalous' enriched mantle are occasionally entrained within continental rift valley magmas.
~-10 0'5125I
0,110
'
1o
0.702 Q704 0.706 0.708 0.710
0.700 0.710 0.720
( •Host Magmas 0.5132 0.5130 0.5128 0.5126
!
~
I
~
0.5124
0.740
( ~ Inclusions
I
~ t
0.730
I
...
-,10
I
SOUTH
AFRICA
-+~
o ~
' ,NOLUS,O,S
/
o.5122 o.5120 t Z ,r,-~., 0.5118 ~- I "O z
OY02
-
GROUP II~ K,MBERL,TES garnet inclusions in diamonds I
I
I
I
0.706
0.710
0.714
0.718
-lO
-15 "t~ Z
87Sr / 86 Sr
FIG. 3. Nd and Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basalts and kimberlites from other continental regions. Note the following: (i) the host alkaline magmas from Scotland and Antarctica and the basaltic kimberlites (Group 1) from S Africa have Sr and Nd isotopic compositions not unlike OIBs; (ii) the mantle inclusions display, in most instances, a greater range of Sr and Nd isotopic ratios than the host magmas; (iii) a greater proportion of the inclusions than were entrained in oceanic basalts are from regions of the mantle enriched in Rb and light REE for a considerable time period; (iv) micaceous kimberlites (Group II) have Sr and Nd isotopic ratios distinctly different from those of OIB and similar to some of the inclusions. Data are taken from several sources. Scotland: Menzies & Halliday (1984); Halliday et al. (1985). Antarctica: Stuckless & Ericksen (1976); Kyle et al. (1985); Menzies et al. (1985b). South Africa: Menzies & Murthy (1980b); Kramers et al. (1981 ; 1983); Erlank et al. (1982, 1987); Smith (1983); Menzies et al. (1987).
M. Menzies
20
Although the source regions of the host magmas may be similar, clearly the potential exists in rift valleys for magmas to entrain fragments of mantle that does not exist under ocean basins.
istries. The range in isotopic composition is extreme, particularly with respect to Sr isotopic ratios (Fig. 3). These isotopic data record a timeintegrated response to (a) major partial melting or depletion events (similar to that observed in the alkaline volcanic rocks erupted in the oceans and rift valleys) and (b) ancient enrichment or metasomatic events caused by infiltration of silicate melts and aqueous fluids that have not as yet been reported from the ocean basins.
Other continental regions Alkaline basalts, lamproites, lamprophyres and kimberlites intrude regions of old crust. The alkaline basalts have Sr and Nd isotopic ratios consistent with derivation from either enriched (e.g. Scotland) or depleted (e.g. Antarctica) portions of the Earth's mantle (Fig. 3). Moreover, most of these continental alkaline basalts and basaltic kimberlites (Group I) are isotopically indistinguishable from ocean island alkaline basalts or rift valley basalts (Table 1). In contrast, micaceous kimberlites (Group II) and lamproites have more radiogenic Sr and less radiogenic Nd isotopic ratios than anything erupted in the ocean basins or rift valleys. In the case of the Group II kimberlites and lamproites we must invoke a source region distinctly different from that of alkaline rocks erupted in the ocean basins or continental rift valleys. It is perhaps to be expected that the petrographically and chemically diverse mafic and ultramafic inclusions entrained from below regions of old crust would display heterogeneous isotopic chemI
I
Asthenosphere-lithosphere The available host magma and inclusion data can now be integrated into some kind of Earth model. To assist us, the Sr and Nd isotopic data for continental and oceanic alkaline basaltic rocks (excluding Group II kimberlites) are compared with available data from several ocean islands and MORB (Fig. 4). The first point to be made is that the source is identical (Fig. 4) regardless of whether the rocks are erupted in continents or ocean basins. Their source must therefore lie not in the lithosphere but in or below the convecting asthenosphere since (a) the oceanic lithosphere is relatively sterile and not capable of producing large volumes of alkaline magma and (b) the I
I
I
+15
0.5134 MORB 0.5132
+1o
~Nd
0.5130 4"
A
+5
V
0.5128
erguelen It
0.5126 9
A9 ~9
9
"-~
,O-
0.5124 I
0.702
0.703
-5
I
0.704
0.705
0.706
87Sr/86Sr
FIG. 4. Nd versus Sr isotopic variation in alkaline volcanic rocks and basaltic kimberlites. Comparative basalt data are taken from Dosso & Murthy (1980), Cohen & O'Nions (1982), Chen & Frey (1983) and Stille et al. (1983). Other sources of data are: o, San Quintin, Mexico (De Paolo 1978); , San Carlos, Arizona (Zindler & Jagoutz 1987); o, Geronimo, Arizona (Menzies et al. 1985a); O, Pisgah Crater, California (De Paolo 1978); ~ Nunivak, Alaska (Menzies & Murthy 1980a); ,t, Ataq, South Yemen (Menzies & Murthy 1980a); 4, Kiama, New South Wales (Menzies & Wass 1983); ~, Afar, Ethiopia (Betton & Civetta 1984); in, Massif Central, France (Chauvel & J ahn 1984); ~ , Malaita, Solomon Islands (Bielski-Zyskind et al. 1984); 4,, Victoria, Australia (McDonough & McCulloch 1985); ~, Sierran Province, California (Van Kooten et al. 1985); ~, Ross Island and vicinity, Antarctica (Menzies et al. 1985b); V, Tanzania (Cohen et al. 1984); v, Tahiti (White & Hofmann 1982); v, Hawaii (Stille et al. 1983); , , Scotland (Menzies & Halliday 1984; Halliday et al. 1985); *, South Africa (Smith 1983).
Alkaline rocks and their inclusions continental lithosphere is heterogeneous and differs markedly from the oceanic lithosphere. The second point to note concerns the similarity between mantle sampled by volcanic rocks erupted in ocean basins and continental rift valleys (summarized in Fig. 5). Essentially this fragmented and entrained lithosphere has a prehistory not unlike the source regions of alkaline magmas ultimately responsible for the transport of inclusions to the surface. This can be interpreted to mean that lithosphere has been rehomogenized by impingement of asthenospheric melts. Since rift valleys are relatively young (less than 0.2 x 10a Ma) any trace-element enrichment locked into the lithosphere would have had insufficient time to evolve isotopically or would have been obliterated by recent rehomogenization of the isotope systems. The third point has relevance to the presence
+1!
+10 I
+5 I E~
r ANTARCTICA
Na20 and those with high Na20 contents, e.g. olivine melilite nephelinites, require very small degrees of partial melting. The genesis of these highly alkaline SiO2-undersaturated magmas is much more likely to be the result of direct partial melting than of fractionation processes as these magmas have a primitive chemistry which precludes even moderate fractionation and they are in equilibrium with a pyrolitic mineral assemblage (cf Ringwood 1975). Because these magmas have much greater abundances of incompatible, major and trace elements than are present in less alkaline types, very low degrees of partial melting seem to be required (cf Kay & Gast 1973). Nevertheless the problem of segregation of these very small amounts of partial melt and their subsequent ascent from 100 km or greater is formidable (see Yoder 1976). This difficulty is even greater if, as Ringwood (1975, p. 153) suggests, less than 1% partial melting of a hydrous source may be required for the generation of olivine nephelinite, olivine melilitite and kimberlite magmas. More recently McKenzie (1984) has developed models based on equations involving
34
A. D. Edgar
the movement of melt and matrix in partialmelts. Based on his calculations, very low degrees of partial melting (less than 3~) at depths commonly accepted for alkali basalts can readily allow melts to ascend to the surface. For highly alkaline magmas derived by direct partial melting of a mantle source the grid shown in Fig. 4 may not be as valuable as for less alkaline types. The grid is based on a pyrolite (lherzolite) source in which H20 is the only volatile component. Recently the concept that alkaline magmas are partial melts of mantle sources metasomatized to produce regions of quite different compositions from that of pyrolite and the importance of volatiles other than H20 in such source regions has been widely discussed ( c f Bailey 1982). Much of the remainder of this review is devoted to experiments pertinent to these concepts.
Recent experimental studies on alkaline mafic-ultramafic magmas Alkaline olivine basalts
The Basaltic Volcanism Study Project (1981, chapter 3) defines alkali olivine basalts as those with not more than 5~ normative nepheline and lists (Appendix 3.1, pp. 614-15) 11 experimental studies carried out on natural or synthetic samples up to 1978. The pressures varied from 1 atm to 33 kb under dry to H20-excess conditions, many with controlled or reasonably estimated oxygen fugacity (fO2) and temperatures from nearliquidus to solidus. Experiments performed at low pressures are not pertinent to the problems of the generation of alkali magmas under mantle conditions, but are important in establishing the sequence of mineral crystallization particularly near the liquidus and possible parental magma compositions (see Thompson 1973). These experiments have also been important in determining the effects offO2 on the stabilities of minerals in alkali olivine basalts (see Helz 1973, 1976). Unfortunately, fewer data are available for the effects offO2 on alkali magmas than are available for other basalts. Thompson (1974) studied a primitive alkali olivine basalt composition from Skye, Scotland (Mg number, 65; ne, 2.84~), between 8 and 31 kb at fOz>iron-wustite (IW) buffer under dry conditions. Because of its high Mg value it is unlikely that this alkali olivine basalt could be derived from a tholeiitic parental magma by fractionation; it is more likely to represent a pristine melt as suggested by Green (1970a) for similar alkali olivine basalts. Near-liquidus rela-
tionships in these experiments are very complicated at about 17kb (see Thompson 1974) suggesting that liquids must be saturated in olivine + clinopyroxene_ pigeonite. On the basis of these results Thompson proposed that this alkali olivine basalt was either a direct partial melt of an Fe-rich lherzolite at about 50km depth or a partial melt of a more Mg-rich lherzolite from deeper sources which underwent some olivine fractionation. Frey et al. (1978) estimated that 11~-15~ partial melting of garnet-free lherzolite can produce an alkali olivine basalt magma. Experiments without volatiles give no information on the roles of amphiboles and micas on the genesis of alkali olivine basalt. Allen et al. (1975) studied a Hualalai, Hawaii, alkali olivine basalt under pressure conditions similar to those used by Thompson (1974) but with excess H20 and f O 2 approximating those of the hematitemagnetite (HM), nickel-nickel oxide (NNO) and magnetite-wustite (MW) buffers. This composition has a comparable Mg number (67) but a lower ne (0.20~) than the sample used by Thompson (1974). Their experiments are shown in Fig. 5 which indicates that under HzO-excess conditions amphibole + olivine + clinopyroxene are near-liquidus phases around 13 kb, olivine is a liquidus phase up to 25 kb and garnet is present between 18 and 23 kb (Fig. 5(a)). The main differences between the results of the dry experiments (Thompson 1974) and the H20-excess conditions used by Allen et al. (1975) are the presence of amphibole and absence of pigeonite in the HzO-excess experiments (Fig. 5(a)). Allen et al. (1975) found that a n f O 2 corresponding to the N NO buffer produced amphibole at higher temperatures relative to the HM and MW buffer conditions (Fig. 5(b)). The absence oforthopyroxene in the experiments reported by Allen et al. places some constraints on the potential of orthopyroxene fractionation in the genesis of these basalts under H20-saturated conditions. The experiments support the concept that andesitic magmas are derived from basaltic magmas by amphibole-liquid fractionation processes as amphibole is stable at temperatures greater than the H20-saturated liquidus of andesite. Thus fractionation of amphibole from basaltic liquids during partial melting or fractional crystallization processes would lead to SiO 2 enrichment. Melting at depths corresponding to pressures greater than 23 kb, where amphibole is no longer stable (Fig. 5), would enrich liquids in Na20 and H20. Figure 6 shows the results of dry experiments performed by Takahashi (1980) on a primitive alkali olivine basalt from Oki-Dogo Island, Japan (Mg number, 63; ne, 3.40~), under low-fO2
Genesis o f a l k a l i n e m a g m a s
[--,
| I"
O I + C p x + G a + Ru + O p
25 ~-rA'mphibo'0"~ "
~
9 f, . / LOlivine Out] "4="
35
, ~
,
,
Y+
x 20
13_
10
,, ,,'
800
900
,
+ Cpx
o
__ J
1000 1100 800
900
1000
1100
Temperature, ~ (b)
(o)
FIG. 5. Phase relations of an alkali olivine basalt from Hawaii : (a) under H20-saturated conditions; (b) effects of variousfO2 on amphibole stability. Abbreviations are as in Fig. 2 and as follows : Cpx, clinopyroxene; Am, amphibole; Ga, garnet; Ru, rutile; Op, opaque; N-NO, nickel-nickeloxide buffer assemblage; M-H, magnetite-hematite buffer assemblage ; M-W, magnetite-wustite buffer assemblage. Liquid(s) coexist with all assemblages. (Modified from Allen et al. 1975.) conditions within the wustite stability field. This basalt, which contains spinel lherzolite nodules, gives results similar to those obtained by Thompson (1974) with orthopyroxene+clinopyroxene on the liquidus at 14-15kb and olivine+orthopyroxene (pigeonite) + clinopyroxene + liquid between 13.5 and 14 kb. This suggests that the
i
I
I
magma could have coexisted with the latter assemblage at depths corresponding to this pressure. This assemblage, along with spinel, also occurs at an isobaric invariant point within the D i - A n - F o tetrahedron between 7 and 20 kb (Presnall et al. 1978, 1979). Kushiro (1973), Presnall et al. (1978, 1979) and Mysen & Kushiro
i
i
I
i
1500
ou 1 4 0 0
Ga§
.
OL+Opx
~
+ Cpx+L t
d
OL+Cpx+L
\Opx+Cpx+L
/
_ /
r 1300 n
E
r
1200
1100
j
~
OL§
0
n 5
n 10
n 15 Pressure,
n 20 kb
~ 25
I 30
35
FIG. 6. Phase relations in an alkali olivine basalt from Japan (abbreviations as in Figs 2 and 5). (Modified from Takahashi 1980.)
36
A. D. E d g a r
(1977) suggest that liquids in more complex natural systems become nepheline normative between 10 and 20 kb. All these results suggest that the alkali olivine basalt lay on the SiOz-poor side of the A b - F o - D i plane (Fig. 1) at 14 kb. Based on these experiments Takahashi (1980) proposed that alkali olivine basalt magmas can be partial melts oflherzolite at depths corresponding to 14 kb provided that H20 and CO2 are unimportant in the melting relations, or that they were generated at greater depths but remained in equilibrium with lherzolite in the upper mantle. Alkali olivine basalts and systems of corresponding composition have also been investigated experimentally at crustal pressures under a variety of conditions. Duke (1974) studied an alkali olivine basalt (Mg number, 57.3; he, 1.63~) from the Azores at 1 atm and withfO2 determined using CO2 and C O + C O 2 mixtures. He found that oxidation increased the stability of spinel and clinopyroxene but decreased that of olivine and plagioclase, promoted crystallization of magnesioferrite-rich spinel and acmitic pyroxene, and caused nephetine-normative liquids to become SiO2 saturated or oversaturated. Using the system CaO-MgO-AlzO3-SiO2-Na20-H2 O at 5 kb to model various basalt types, Cawthorn (1976) found that nepheline-normative liquids could crystallize amphibole___ olivine + Ca-rich clinopyroxene to produce quartz-normative residual liquids of andesitic composition.
Green (1973a) investigated a lherzolite-bearing olivine basanite composition from Mount Laura, Victoria, with 10% added olivine (Mg number, 70; ne, 10.5%) between 5 and 36 kb and over a range of H20-undersaturated liquid conditions. As shown in Fig. 7, liquids from this composition are saturated with olivine + orthopyroxene + clinopyroxene+garnet only under very limited conditions corresponding to 25-30kb, 12001300~ and 2 - 7 ~ H20. Based on these experiments, Green (1973a) concluded that basanites could be generated from a lherzolite mantle with 0.2-0;4% H20 by 6 ~ partial melting at a depth of around 100 km. He suggested that the excess olivine fractionated from the liquid prior to the incorporation of lherzolitic mantle xenoliths, presumably during ascent. Using a kaersutite eclogite nodule from Kakanui, Hawaii, of composition corresponding to olivine basanite (Mg number, 63; ne, 6.4~), Merrill & Wyllie (1975) showed that orthopyroxene+ clinopyroxene + garnet coexisted at 20 kb under HzO-excess conditions, and at greater than 30 kb with progressively lower H=O contents. With the addition of normative olivine to their bulk composition, they found a more extensive field of olivine + orthopyroxene + clinopyroxene+ garnet in equilibrium with liquid than that reported by Green (1973a) and they also found that kaersutite was stable on the H20-undersaturated liquidus between 12 and 21 kb for liquids 40
Basanites and nephelinites Basic rocks more alkaline than alkali olivine basalts are generally defined as those with more than 5% normative nepheline. They may also contain normative leucite and a few have normative kalsilite. These magmas have primitive Mg numbers and other chemical features indicative of an origin by direct partial melting from a mantle source. Some are extremely rich in K and other incompatible and large-ion-lithophile (LIL) elements. Because of the many problems involved in the genesis of these rocks, including the fact that many cannot apparently be derived by partial melting of a (normal) lherzolitic mantle source, a large number of experiments on these magma compositions have been carried out in recent years. Those experiments particularly applicable to the nature of the source regions are concentrated on in this section. Over a decade ago it was recognized that highly alkaline magmas could not be derived by partial melting of volatilefree mantle peridotite but that some of these magmas could represent partial melts of an HzOand CO2-bearing peridotitic mantle (see Basaltic Volcanism Study Project 1981, pp. 529-30).
Ol
/
,/ go+'~px,/ I
I
~1 t 'IX
l +wx/ l + opx /
I go
~',,~ + ol l
I
go
/o,/// /op,+.l
:~ 20
!
,/ I g ~
/
/
/
/
/
/
J
/
/ /
l,,
II
I/ II
I jq.~
/
/
~'~
~ ~
Olivine/
/
,~
\1
!9
/ O........ livine
\//
o
L 1100
,
J
"-.I 1300
/ ,
. 1500
TEMPERATURE, ~ FIG. 7. Phase relations in an olivine basanite + 10~ olivine under dry to HzO-saturated conditions: - - -, percentage HzO added. Abbreviations are as given in Figs 2 and 5. (After Green 1973b).
Genesis of alkaline magmas containing at least 2-3% H20. From this Merrill & Wyllie (1975) inferred that basanites with kaersutite megacrysts may contain 2-3% H20 at depths corresponding to these pressures. Arculus (1975) investigated basanites from Mount Shadwell, Victoria (Mg number 68; ne, 11.92%), and from Grenada, Lesser Antilles (Mg number, 74; ne, 7.84%), under dry conditions between 10 and 35 kb. Owing to the absence of liquidus assemblages compatible with a dry peridotitic source (notably orthopyroxene), he concluded that basanites could not be derived from such a source unless H20 and CO2 were present; the former stabilizes olivine relative to pyroxene, and the latter stabilizes orthopyroxene relative to olivine and clinopyroxene (Eggler 1974). Arculus (1975) also concluded that oxides of polyvalent cations such as TiO2 and P205, commonly abundant in basanites, might lower the olivine stability field by 2-3 kb. Thus the absence of volatiles and such polyvalent cations in previously proposed models (e.g. O'Hara & Yoder 1967; Kushiro & Yoder 1974) suggests that such models might be inapplicable for natural systems of basanitic affinities. Evolved alkali basalts Evolved alkali basalts with lower Mg numbers and less normative nepheline than the primitive basanites and nephelinites may be produced from these more primitive magmas by various fractionation mechanisms. Bultitude & Green (1971) suggested that many basanites and nephelinites might be products of olivine fractionation of lowSiO2 picritic magmas. Irving & Green (1972) demonstrated that kaersutite was a near-liquidus mineral at 14 kb from a nepheline mugearite composition with about 3.5% H20 which contained lherzolite nodules. These studies indicate that fractionation of olivine+kaersutite from basanitic magma could produce nepheline hawaiites and nepheline mugearites at depths around 40-45 km. Although experimental data are lacking, it seems likely that further fractionation of clinopyroxene+biotite could produce nepheline benmoreites and phonolites (cf Irving & Price 1981). Olivine-rich nephelinites and melilitites In experiments with olivine nephelinites and comparable compositions up to 36 kb and under dry to H20-excess conditions (Bultitude & Green 1968, 1971; Allen et al. 1975; Merrill & Wyllie 1975) it was shown that such magmas could not be in equilibrium with a lherzolitic mantle source. In all cases clinopyroxene rather than orthopyrox-
37
ene appeared on the liquidus. Bultitude & Green (1971) indicated the improbability that tholeiitic picrite or alkaline picritic magmas are related to olivine and olivine melilite nephelinites by fractionation processes, thus implying that these magmas are primary. Brey & Green (1975, 1976, 1977) and Brey (1976, 1978) conducted a very detailed study of olivine melilitite compositions to investigate the effects of H20:CO2 ratio a n d f O 2 up to 40 kb on the genesis of these magmas. These authors used an olivine melilitite from Laughing Jack Marsh, Tasmania (Mg number, 74; ne, 17.2%; lc, 6.6%), and a synthetic Ca- and Mg-rich olivine melilitite (Mg number, 75; ks, 4.4%). In addition to analyses of the solid phases, the H20 and CO2 contents of the liquids were also determined. Using the Laughing Jack Marsh sample Brey & Green (1975) found no combination of pressure, temperature and percentage H20 conditions under which this olivine melilitite could be in equilibrium with a pyrolite mantle assemblage. In the presence of H20 the solubility of CO2 is high in the olivine melilitite liquid, and it suppresses olivine and clinopyroxene and promotes orthopyroxene and garnet as liquidus or near-liquidus phases. These results, using an external HM buffer assemblage, are summarized in the 30 kb section shown in Fig. 8. With increasing COz (decreasing HzO/(H20 + CO2)) the liquidus temperature rises almost 200~ between 0.5 H20/ !
!
! 1400
? e
\
13oo
0
ox
§
0
9
~ 9
0.'25 0.~5
- 1200
0
0175
C02
38wt7.
T*C
H20 H20 * C02 mole fraction in c h a r g e
1100 H20 20wt7.
Fla. 8. Phase relations in an olivine melilitite showing variations in the liquidus phases at 30 kb and at f O 2 = HM buffer with varying HzO/(HzO + C02). Liquidus and near-liquidus phases are indicated above the liquidus: O ; garnet; +, orthopyroxene; 0, clinopyroxene; o, all liquid. (Modified from Brey & Green 1975.)
38
A.D.
(H20 + CO2) and CO2-only experiments. Based on these results, Brey & Green (1975) inferred that olivine melilitite magmas could be derived at 30kb and 1150-1200~ by less than 5~ equilibrium partial melting of pyrolite containing 0.5-1.0 H20/(H:O+CO2). They suggested low fO2 values (less than the N N O buffer) as most appropriate. Brey & Green (1977) refined the results of their earlier study by systematic investigation of the liquidus and near-liquidus relations with varying H 2 0 / ( H 2 0 + C O 2 ) and f O 2 conditions. The resuits shown in Fig. 9 represent near-liquidus phases at 30 kb andfO2 = HM buffer. Of particular importance is the narrow field of clinopyroxene separating the garnet+orthopyroxene+ clinopyroxene field from the olivine + clinopyroxene field for varying CO2 and H20 contents. Experiments at 27 kb with a composition containing 0.92 H20/(H20 + CO2) closely approached a near-liquidus olivine + orthopyroxene + clinopyroxene+garnet assemblage, compatible with equilibration of the olivine melilitite magma with a pyrolite source. Brey & Green (1977) proposed that the additon of 5% olivine (implying fractionation of a similar amount of olivine from the primary magma) would produce this four-phase assemblage. They also determined that liquidus and near-liquidus assemblages were not appreciably different at more realistic mantlefO2 values corresponding to those less than the N N O buffer assemblage. Based on these results, Brey & Green (1977) proposed that olivine melilitite magmas could be derived at depths corresponding to 27 kb at 1160~ by partial melting of a pvrolite mantle source containing both H20 and CO2 equivalent to 6 - 7 ~ CO2 and 7-8~o H20 dissolved in the melt. Brey (1978) proposed that olivine melilitites which were more SiO2-undersaturated than the ,
/ O~
4Q
O~. /"'f I"j1~
20
Laughing Jack Marsh sample might be derived at greater depths and with higher CO2 in the source region. Experiments at 30-35 kb with higher CO2 contents using a very SiO2-undersaturated synthetic olivine melilitite composition produced olivine+clinopyroxene as liquidus phases. At 30 kb this assemblage is clearly separated from the garnet + orthopyroxene field by an orthopyroxene field. He found that garnet stability increased at lower H20/(H20+CO2) and suggested that highly SiO2-undersaturated olivine melilitites could be derived at deeper levels (corresponding to 35 kb) by partial melting of a more CO2-rich pyrolite source. Although Brey & Green (1975, 1977) and Brey (1978) demonstrated the role of CO2 in addition to H20 for the genesis of olivine melilitite magmas, Eggler (1974, 1978), Huang & Wyllie (1974), Wyllie & Huang (1975, 1976) and Wendlandt & Eggler (1980a, b) proposed that CO2 was an essential component in the source regions for mafic SiO2-undersaturated magmas. Reviews of the experimental evidence for the role of CO2 are given by Eggler & Holloway (1977) and Wyllie (1977, 1979). Eggler (1978) studied joins in the system Na20-CaO-A1203-MgO-SiO2CO2 up to 30 kb and showed that orthopyroxene stability increased relative to that of olivine, thus producing liquids containing less SiO 2 than those formed in the absence of CO2. He suggested that nephelinites and melilite nephelinites could be partial melts of a peridotitic source at 50-90 km provided that CO2 was the predominant volatile component. If the total volatile contents (H20 + CO2) of the mantle source region are greater than 0.5 wt.% all variations between H20-saturated and CO2-saturated melts can occur, whereas with low total vapour concentrations in the source regions the compositions of the vapour a r e restricted and hence solidus temperatures and near-solidus melt compositions are also restricted. As an example of the role the amount of H20 and CO2 plays in the composition of melts. Eggler (1978) proposed that partial melts of an amphibole-peridotite in the presence of both H20 and CO2 (not exceeding 0.37 wt.%) will produce a nephelinitic melt composition at 15 kb.
/"
~i~ 20
Olivine Melilitite
,,
Edgar
Kimberlites "
ga + opx
i 40
ga + opx + cpx C pX ol § c p x ol 60 H20
FIG. 9. Phase relations in the system olivine melilititeH20-CO 2 at 30 kb andfO 2= HM buffer. Abbreviations are as in Fig. 5. (Modified from Brey & Green 1977).
The role of CO2 in the genesis of magmas is particularly relevant to kimberlites and carbonatites. Eggler & Wendlandt (1979) determined phase relations in a kimberlite with variable H20 and CO2 contents at 30 and 55 kb. At the higher pressure they found that kimberlite-like liquids coexisted with a lherzolite assemblage in experiments with 5% H20 and 5% CO2, but kimberlites
Genesis of alkaline magmas
could not be derived from a lherzolite source at the lower pressure. At 30 kb phlogopite and dolomite were sub-solidus phases in experiments with HzO-rich buffered vapour, whereas magnesite solid solution was a sub-solidus mineral at 55 kb. On the basis of phase relationships in the peridotite-COz-H20 system, particularly those determined by Eggler (1978), Wyllie (1977, 1978, 1979) and Ellis & Wyllie (1980), Wyllie (1980) suggested that kimberlite originated by partial melting of peridotitic mantle near 260 km in which solidus temperatures were reduced by upward migration of volatiles. This melting causes a density inversion and results in a diapiric uprise under adiabatic conditions with subsequent crystallization of the partially melted diapiric material beginning at depths of 10080 kin. Leucitites and olivine leucitites With the exception of kimberlites, all the alkali magmas so far described have N a z O > K 2 0 . Mafic-ultramafic rocks with K z O > N a 2 0 are essentially leucitites and olivine leucitites, but with a plethora of varietal names. Most of these rocks fall within the broad category of lamproites (cf. Bergman 1987). They commonly have normative leucite but may be SiO2-saturated to oversaturated and contain normative quartz. Problems involved in the genesis of these often highly LIL-element-rich rocks have been summarized by Bell & Powell (1969). Edgar et al. (1976) have pointed out some of the improbabilities of the older theories given by Bell & Powell. These theories fall into three categories. Firstly, differentiation of basaltic or peridotitic magma with fractionation of olivine (Wade & Prider 1940), eclogite and olivine (Holmes 1932), or eclogite (O'Hara & Yoder 1967). Secondly, assimilation of magmas of such diverse compositions as granite, carbonatite, nephelinite and monchiquite by a wide variety of crustal materials (cf Daly 1910; Bowen 1928; Williams 1936; Holmes & Harwood 1937; Larsen 1940; Holmes 1950, 1965; Turner & Verhoogen 1960). And thirdly, partial melting of biotites and hornblendes (Waters 1955) and kimberlite (Harris, quoted by Bell & Powell 1969). Any hypothesis involving basaltic magma seems unlikely, as in most occurrences of rocks of the leucititic kindred basalts are sparse or absent. Even if basaltic material was involved, very extreme degrees of crustal contamination or assimilation would be required to account for their high incompatibleelement content and to maintain their high MgO and CaO values. The theories involving partial melting of some types of mantle material are now widely accepted by most workers. Recent wide-
39
spread interest in these rocks has been prompted by the possibility of greater heterogeneity in the mantle source regions than previously proposed and by the current popularity of the process of mantle metasomatism (cf Bailey et al. 1980; Bailey 1982). In general, the similarities in high Mg number, low SiO2 content and high volatile contents of olivine leucitites, olivine basanites, olivine melilitites, olivine nephelinites and kimberlites suggest that these magmas are primary rather than derivative and hence can only be evolved from a mantle source containing H20 and CO2. Cundari & O'Hara (1976) conducted dry experiments on a mafic leucitite from New South Wales, Australia (Mg number, 79; lc, 0H; ne, 9.08%). They found clinopyroxene as the only liquidus phase between 20 and 40 kb and olivine at lower pressures. Absence of the assemblage olivine + orthopyroxene+ clinopyroxene _+garnet_+ spinel as near-liquidus assemblages precludes this magma's being derived by partial melting of dry lherzolite. Cundari & O'Hara (1976) concluded that this rock may have formed by partial melting of an olivine pyroxenite source containing essential phlogopite. Absence of normative leucite or quartz in this composition sets it apart from other compositions. Thompson (1977) conducted experiments under similar conditions on an apparently less-primitive leucitite from the Alban Hills, Italy (Mg number, 63; lc, 32.5%; ne 13.6%), with the same results. High-pressure experiments using H20 and H20 + CO2 have been carried out for compositions ranging from potassic intermediate to ultramafic from SW Uganda (Edgar et al. 1976, 1980; Ryabchikov & Green 1978; Arima & Edgar 1983a), from the Leucite Hills, Wyoming (Barton & Hamilton 1979, 1982), and from the W Kimberley area, W Australia (Arima & Edgar 1983b). All these compositions have primitive Mg numbers (70-78) but widely varying degrees of K enrichment (K/(K + Na) mol = 0.41-0.86). In Fig. 10 these rocks are plotted on the Mg2SiO4 (Fo)-KA1SiO~ (Ks)-SiO2 system which indicates the variable degrees of SiO2 saturation. The three compositions from SW Uganda (olivine ugandite, katungite and biotite mafurite) and the madupite from the Leucite Hills are all SiO2 undersaturated, whereas the orendite from the same locality and the wolgidite from W Kimberley are SiO2 saturated and slightly oversaturated respectively. The SiO2-undersaturated group shows distinctly different liquidus to near-liquidus phases from the group which is SiO 2 saturated to oversaturated. The primary phase fields at 3 kb PH2o (Luth 1967) and 28 kb Pmo (Green et al. 1984) in this system are also included in Fig. 10. With
40
A . D. E d g a r Oi
Sil wt.% FIG. 10. K-rich mafic-ultramafic rocks (O), for which experimental data are available, plotted as CIPW normative minerals in the system Mg2SiO4-KA1SiO4-SiO2 together with phase relations at PHzo of 3 and 28 kb (Luth 1967; Green et al. 1984)" HI, phlogopite composition. Abbreviations are as in Figs t, 2 and 3" Hy, hypersthene. Ks
Lc
So
increasing pressure the fields of phlogopite and enstatite expand relative to olivine. Thus partial melting of a mantle source with H20 as the only volatile phase results in liquids which may crystallize these minerals. With H20 as the only volatile phase neither orthopyroxene nor garnet crystallize from any of the SiO2-undersaturated compositions. Thus these magmas cannot have equilibrated with a normal lherzolitic mantle source. Figure 11 is a schematic diagram of near-liquidus phase assemblages for these compositions with H20 contents between 3 and 15 wt.~, representing vapourabsent conditions above about 10 kb. Clinopyroxene___phlogopite are the commonest near-liquidus minerals at high pressures and olivine+ clinopyroxene are the commonest at lower pressures. Based on these near-liquidus phase assemblages, possible source regions for these magmas range from a slightly K-enriched wehrlite for the least K-rich olivine ugandite magma to a phlogopite-rich clinopyroxenite for the highly K-rich mafurite magma (Fig. 11). The appearance of
phlogopite as a near-liquidus assemblage is directly correlated with the degree of K enrichment (Edgar & Arima 1983). Figure 12 shows near-liquidus assemblages for the SiO2-saturated (orendite) and SiO2-oversaturated (wolgidite) compositions in which only H20 has been added. Orthopyroxene is a common near-liquidus mineral for these compositions. In the orendite the near-liquidus assemblage of olivine + orthopyroxene + clinopyroxene + garnet + phlogopite (trace) at 30 kb suggests that this magma could be derived from a phlogopitebearing garnet lherzolite containing small amounts of H20. A phlogopite orthopyroxenite and a phlogopite harzburgite seem the likeliest sources for the wolgidite composition at high and low pressures respectively. The nature of the source materials and the depth of generation depend on the amount of H20 in the source. Ryabchikov & Green (1978) studied the effects of CO2 and H20 at 30kb on the mafurite composition used by Edgar et al. (1976). Figure 13 shows that, with total volatiles equivalent to 5
Genesis of alkaline magmas
40-
Madupite OI-ugandite Katungite Mafurite (Barton 8~ (Edgar e/o/. (Arima 8~ (Edgar elo/. Hamilton 1980) 1980) Edgar 1983a) 1976)
41
OI-ugandite (Edgareta~. 1980)
Katungite Mafurite (Arima & (Edgar et o/. Edgar 1983a) 1976) -140
T
T
35
t
30-
s_HlightlyK-enriched wehrlite (ol+cpx)
K-bearing clinopyroxenite (cpx) a_ 2 5 -
I
3%
-
RI---#-~ h g~ phlog_o_pite clinopyroxenite clinopyroxenite (cpx+phl) (cpx+ phi)
p_hloggRitewehrlite (cpx + ol + phi)
phlogopitewehrlite (cpx +ol+phl)
20-
, ]
P_h I._9_og_opit e___: clin.~_opyroxenite (cpx+phl)
t
120
slightly K-enriched wehrlite (ol+ cpx)
(ol only) p_hl og_op_it_eewehrlite (ol+cpx+phl)
I
~
(ol only)
l
5%
4
H20
p_hlogo!it.__eewehrlite (ol.cpxephl)
I
added
l
p_hl og_opit_.ee wehrlile (cpx + ol + phi) 15%
I00
80
I
60 I
FIG. 11. Schematic diagram showing near-liquidus phases for SiO2-undersaturated K-rich magmas in which H20 is the only added volatile. The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets. wt.% H20 , olivine+clinopyroxene are liquidus phases at CO2/(CO 2 Jr H 2 0 ) mol < 0.50, whereas orthopyroxene stability increases above this value. For CO2/(CO2 + H20) mol between 0 and 0.50 phlogopite is stable but clinopyroxene rapidly becomes unstable at higher ratios. For CO2/(CO2+H20 ) equivalent to 15 wt.% H20 , phlogopite and clinopyroxene stability decrease at low ratios, with orthopyroxene as the liquidus phase over most of the range of CO2/(C0 2 Jr H20) values. These relations and the absence of garnet suggest that a lherzolitic source may be possible for this highly K-rich composition. Figure 14 is a schematic diagram of three SiO 2undersaturated SW Ugandan lavas at CO2/ (CO2+H20) mo1=0.75 (equivalent to 15 wt.% H20 added). Only the low-K ugandite with orthopyroxene + clinopyroxene_+ garnet 4- spinel as near-liquidus phases from 22 to 40 kb can be derived from a lherzolitic source under these conditions. The katungite liquidus relationships suggest a garnet clinopyroxenite or a clinopyroxenite as the likeliest source compositions. Although phlogopite does not occur in these highCO2 experiments, the source regions must be K enriched in order to produce the K enrichment in the magmas provided that these are direct mantle melts. No experiments with high CO2 have been carried out on the SiO2-saturated to SiO 2oversaturated orendite or wolgidite.
Wendlandt & Eggler (1980a, b) modelled the genesis of potassic magmas based on melting relations in the systems KAISiO4-Mg2SiO4SiO2, KA1SiO~-MgO-SiO2-CO2 and KA1SiO4MgO-SiO2-H20-CO2 as simple mantle sources at pressures up to 50 kb. In these systems they were able to determine a large number of reactions producing a wide variety of possible liquids ranging from simplified tholeiites to carbonatites. On the basis of previous studies they suggested that the absence of Ca (as CaMgSi206) and Na (as NaA!SiO4 and NaA1Si3Os) would not appreciably alter their conclusions. The expansion of the primary phase fields of enstatite and sanidine in the KA1SiO~-Mg2SiO,~SiO2 system with increasing pressure is illustrated by the schematic diagrams given by Wendlandt & Eggler (1980a, Fig. 6). This results in liquids becoming progressively enriched in K20 and depleted in SiO2 at pressures ranging from 5 kb to greater than 30 kb. In the KA1SiO4-MgOSIO2-CO2 system, phase relations are similar but occur at lower pressures than in the CO2-free system. The trends of liquids with increasing pressure are schematically shown in Fig. 15 which is based on the variation in the extent of the primary solid phase fields with pressure in the systems KA1SiO4-MgO-SiO2-CO2 and KA1SiO4-Mg2SiO4-SiO 2 (Wendlandt & Eggler 1980a). The approximate pressures at which the
A. D. Edgar
42 Orendite {Barton 8 Hamilton 1982)
Wolgidite (Arima 8 Edgar !983b)
40 -f
-140
35
120 ph Iogopite
orthopyroxenite (opx+phl)
p_h I__qgo_pite ?garnet- Iherzolite {ol+opx+cpx+gt+ph)
30
Io0 T s s
25
ph I__0_ogo_pit3 orthgpy_roxenite (opx+phl)
p hl_#_ogop~t__e harzbuEgit__e (oi+ opx + phl)
K3
H. p_ ogoplte harzburqit._ee (opx +ol + phi)
80
20
(ol Only) I te p.hi ogopj harzburglte (o1+ opx+ phi) 5%
1'2%
60 15%
H20 added FIG. 12. Schematic diagram showing near-liquidus phases for SiO2-saturated and SiO2-oversaturated K-rich magmas in which H20 is the only added volatile. The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets. 1400
m
L
Iq
o
1400 1 30 kb (equiv to 15 wt% H20)
.+v/
J+FTa 1300
,,'
(bq~//
1300 -
,,
r
OI
,p
/
-....--
/ /
L O
/
~ ~ +
Xtols
/ /
1200
_
/
1200
-
,., C~4\\
-
L+Xtals
-% \\ \ \ \,o. \ \ \-o \%
30 kb {equiv. to 5 wf% H20) I100 H20 5 wt~
I
0.25
I
0-5
I
0.75
C02/(C%+ H20} mole fraction in charge
CO2 12.2 wt~
I100 H20 15 w1%
I
0.25
\
I
0"5
I
0.75
C02/(C02-1- H20) mole fraction in charge
C% 36.7 wt%
F]G. 13. Phase relations in the biotite mafurite system with varying C02/(H20 --kC O 2 ) contents equivalent to 5 and 15 wt.% H20 at 30 kb. Abbreviations are as in Figs 2-4. (Modified from Ryabchikov & Green 1978.)
Genesis of alkaline magmas Katungite (Arima a Edgar 1983a)
OI-ugandite (Edgar et 0/ 1980
40
t
:55 -
Mafurite (Ryabchikova Green 1978)
I
garnetIherzolite ? (cpx + opx + ga)
43
i
' ' garnet' clinopyroxenite (cpx + ga)
-
140
'
120
,r
E
v (D
50
Iherzolite (cpx + opx + ol)
clinopyroxenite (cpx)
r
ioo
d3
13._
25
_
garnet or spinel Iherzolite (opx + cpx + ol + ga + sp) 80
wehrlite (ol +cpx)
20 COz H201
27.4 3-5
27.5 3.8 % CO2 and H20 added
27.5 5"8
FIG. 14. Schematic diagram showing near-liquidus phases for SiO2-undersaturated K-rich magmas with CO2/(CO2 + H20) mol = 0.75 (equivalent to 15 wt.% HzO). The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets.
Mag Pressure (kb) of appearance of liquids Vol absent 19 34-5 --
CO2 satd. 14 27.5 > 29
Liquids Lc normative Ks normative Mag normative Carbonatitic
/
/
/
/
~.// 7
/
7
// "
~
/
/
/ 9
/
/
/~hl
/
Olivine tholeiitic
~/
/
/
/ //
/ /
/
/
Quartz.. tholeiitic
\
/
7
/.. Ks
/ Lc [ Sa (Ks normative] Alkali basaltic / Lc normative}
Qz
FIG. 15. Trends of liquids with increasing pressure in the systems KAISiO4-MgzSiO4-SiO2 and KA1SiO,-MgOSIO2-H20-CO2 (modified from Wendlandt & Eggler 1980a). Abbreviations are as in Figs 1 and 3 ; Mag, normative magnesite. The inset shows the normative nature of liquids at increasing pressures (as indicated by the line on the diagram) under volatile-free and CO2-saturated conditions.
44
A. D. Edgar
compositions of the residual liquids on partial melting become leucite, kalsilite and magnesite normative are shown in the inset to Fig. 15. At approximately 14 kb for CO2-saturated conditions and 19 kb for volatile-free conditions, liquids in equilibrium with sanidine, enstatite and forsterite are leucite normative and change from tholeiitic to alkali basaltic with increasing pressure. At greater than 27.5 kb under CO2saturated conditions and greater than 34.5 kb in the volatile-free system, sanidine is no longer stable and melts become extremely K20 rich and SiOz poor, yielding kalsilite-normative liquids. At pressures above 29 kb in the KA1SiO4-MgOSIO2-CO2 system melts are actually magnesite normative. Based on these results Wendlandt & Eggler (1980a) concluded that quartz tholeiitic liquids cannot fractionate to alkali basaltic ones. Liquids generated at high pressures with some fractionation could result in increasingly SiO2poor residual liquids. In contrast, fractionation at lower pressures could result in SiO2 enrichment. Wendlandt & Eggler's (1980a) experiments have been used by Kuehner et al. (1981) to model the Leucite Hills, Wyoming, magmas which show this trend of SiO2 enrichment. Although it is theoretically possible that leucite-normative rocks could be products of fractionation rather than direct partial melting, Wendlandt & Eggler (1980a) consider this to be unlikely. Using the KA1SiO4-MgO-SiO2-CO2-H20 system in which phlogopite is an additional phase, Wendlandt & Eggler (1980b) modelled the melting behaviour of phlogopite peridotite and its role in the genesis of K-rich magmas, kimberlites and carbonatites. Peridotite with small amounts of H20 + CO2 undergoes univariant melting in which H20 and CO2 vapour compositions are buffered by phlogopite and magnesite respectively. Such buffering is pressure dependent and hence the compositions of liquids produced by various reactions depend on the pressure of melting. With increasing pressure up to 30 kb, melting of a phlogopite peridotite with insufficient CO2 and H20 to hydrate all the potential phlogopite involves four reactions between combinations of phlogopite, enstatite, sanidine (or kalsilite), forsterite, liquid and vapour. As the pressure increases up to 30 kb the liquids change in composition progressively from quartz to enstatite to leucite to kalsilite normative. These relations are comparable with those for the KA1SiO4-Mg2SiO4-SiO2 system (Wendlandt & Eggler 1980a). Above 30 kb magnesite becomes stable and melting occurs by two reactions involving combinations of phlogopite, enstatite, magnesite, forsterite, kalsilite, liquid and vapour. The liquids produced are carbonatitic. With
increasing pressure phlogopite has decreased hypersolidus stability and at 50 kb ceases to be a solidus phase. Thus between 30 and 50kb, melting of phlogopite peridotite will produce increasingly K20-rich SiO2-poor liquids.
Mantle metasomatism Experimental studies and other data indicate that as the alkalinity of mafic to ultramafic magmas increases there is a decreasing possibility that these magmas are partial melts of a 'normal' pyrolitic model mantle composition. Although there are a number of mechanisms whereby local heterogeneities can occur in the mantle source, the concept of heterogeneities being caused by mantle metasomatism (cf. Lloyd & Bailey 1975; Bailey 1982) is now widely accepted. Experimental modelling of metasomatic processes is extremely difficult, particularly at mantle depths where very little is known of either the composition or the speciation of the fluids involved. Among the many elements considered to be responsible for metasomatism (Bailey 1982), K and Na are very important. Ryabchikov & Boettcher (1980) showed that the solubility of potassium (as K/O) in aqueous fluids increased about sixfold between 11 and 30 kb in the range 1050-1100~ Their values, ranging from 4 to 25 g K20 per 100 g H20, indicated that the solubility of K:O is sufficiently high to produce radical changes in the mantle. Ryabchikov et al. (1982) determined that the amount of sodium silicate leached by hydrous fluids at constant temperature from omphacitic pyroxenes increases with decreasing pressure. On the basis of both sets of experiments, Ryabchikov et al. (1982) proposed that K/Na fractionation in the upper mantle might produce K-rich metasomatism at deeper levels and Na-rich metasomatism at shallower levels which, on partial melting, would result in K-rich magmas generated at greater depths than Na-rich magmas. Edgar & Arima (1984; unpublished data) have carried out preliminary metasomatism-modelling experiments involving the system pyroliteK2Oaqueou s and pyrolite-(K20+Na20)aq ..... at 2 0 - 3 0 kb and between 850 and 950~ In these experiments K20 , as dried K2CO3, was dissolved in deionized H20. Na20 was added as Na2CO 3. This resulted in an aqueous solution containing negligible amounts of dissolved CO2. Figure 16(a) shows that at 30 kb an increased concentration of K20 in solution produces an initial increase in the proportions of phlogopite relative to orthopyroxene, clinopyroxene and olivine. Although
Genesis of alkaline magmas not in sufficient quantity to be observed in the Xray diffraction patterns (Fig. 16) garnet (at 30 kb) and spinel (at 20 kb) decrease with increasing K20 as observed optically. At about 3.8 g K20 per 10 g of solution, liquid appears at both 850 and 950~ indicating a decrease in the amount of phlogopite with K partitioning preferentially into the liquid. These results are in accordance with those of Takahashi & Kushiro (1983). Edgar & Arima (1984) were unable to analyse their run products using the electron microprobe because of the fine grain size, but they speculated that at 30 kb the formation of phlogopite involved a reaction between olivine, orthopyroxene and garnet in the approximate proportions ofpyrolitetype mantle at 30 kb (Wyllie 1975) with the K20 in solution to produce a product with less garnet and olivine and more orthopyroxene. Clinopyroxene is probably not involved in the formation of phlogopite. These results are in accordance with the observed decrease in the olivine/orthopyroxene and olivine/clinopyroxene intensities (Fig. 16(a)). In contrast with the results at 30 kb, at 20 kb an amphibole of pargasitic composition as analysed by microprobe shows an inverse relationship to phlogopite which remains the dominant 'metasomatic' product (Fig. 16(b)). On the basis of the relative proportions of phlogopite to the other mineral phases (Fig. 16(b)) and the ratios of olivine to orthopyroxene, olivine to clinopyroxene and orthopyroxene to clinopyroxene (Fig. 16(c)), the formation of phlogopite at 20 kb and at K20 concentrations where amphibole is no longer present (more than 2.0 g K20 per 10 g of solution) may involve a complex reaction between olivine, orthopyroxene, clinopyroxene, amphibole and spinel, together with K20 in solution, to form a product containing a higher amount of olivine and clinopyroxene and lower amounts of orthopyroxene relative to the reactants. A few preliminary experiments at 20 kb and 950~ in which both Na and K have been added in various proportions equivalent to 3.8 g total alkali per 10 g of solution show that, with increasing Na relative to K, the field of amphibole stability increases as phlogopite decreases. Although vastly simplified, these experiments indicate that alkali metasomatism can readily occur. Nodules with mantle mineralogy found in alkali basalts, kimberlites and related rocks often contain phlogopite and amphibole and range in overall composition from lherzolite to pyroxenite. Lloyd & Bailey (1975) and Lloyd (1981) studied nodules mainly of phlogopite clinopyroxenite composition from the K-rich volcanic province of SW Uganda and proposed that they might represent metasomatized mantle sources for these
45
lavas. Alternatively these nodules could be depleted residual source materials. Lloyd et al. (1985) determined the compositions of liquids produced from varying degrees of partial melting at 20 and 30 kb of an average nodule with no added volatiles (based on analyses of 84 nodules from this area). Figure 17 shows the compositions of liquids produced by various degrees of partial melting at 30 kb together with the coexisting minerals. At approximately 25% partial melting of this nodule, the liquid corresponds closely with that of an average katungite composition found in SW Uganda. Although this is a higher degree of partial melting than generally considered likely for the production of such K-rich magmas, it seems feasible in the light of the upwarped mantle isotherms, postulated in rifted areas such as SW Uganda (Bailey 1970), caused by mantle degassing (Bailey 1980), and in view of the high proportion of radioactive elements associated with these magmas.
Felsic alkaline rocks Although experiments pertinent to felsic alkaline rocks have been less numerous than those carried out on mafic varieties in the past decade, studies have been undertaken on liquid immiscibility, on the role of C1, F and P20 5 in alkaline magmas and on the problems of the genesis of pseudoleucites and primary analcites.
Liquid immiscibility Roedder (1979) reviewed the role of experimental work on the concept of liquid immiscibility as a viable process in the genesis of igneous rocks. Only studies applicable to alkaline rocks are considered here, most of which concern silicatecarbonatitic magmas. Freestone (1978) showed that the liquid immiscibility field in the fayalite-leucite-silica system was extended to more K-rich compositions by the addition of 1 mol ~ P205 and 3 mol ~ TiO2. He suggested that this expansion supports the liquidunmixing origin suggested by ocelli textures in syenites and sheets within alkali gabbro intrusions. The role of liquid immiscibility on the genesis of carbonatites was considered by Hamilton et al. (1979) and Freestone & Hamilton (1980). These workers used lavas ranging in composition from felsic nephelinites to phonolites comparable in composition with those in the Oldoinyo Lengai volcano, Tanzania. These were mixed with synthetic Na- and Ca-rich carbonatites. Experiments at 0.7-7.6 kb and 900-1250~ showed that
46
A. D. Edgar 8
_- q~p
/
6
"~ ~
o
o=~=I=ua 2 0
-
@
o
k
,~ ~
~ H
o
1.5~ I'0
o
u H
4
\ Iamph 2
o
oI
0
1"5~~I'0
9
-~---I
H
0.5 8
4
q,t-fl~p
4
6 n
"~
3 2 I
0
v
I'0
2.0
3-0
4-0
I
~
6.0
5.0
1
i0
2. 0
50
l
4-0
g K20/IO g sOIn.
g K 2 0 / I O g soln.
(a)
(b)
I 5.0
I 6,0
3-0 -
2o' I.O
o
u
-6
1.5
0"5
2"0 I-0 I ~ - -
0
a-
1
I-O
o
@ I
2.0
I
3.0
I
4-0
g K20/IO g soln. (C)
I
5-0
I
6-0
FIG. 16. Results of experiments in the pyrolite-K20 aqueous solution system plotted as ratios of X-ray diffraction intensities (I) of minerals against K 20 concentrations: (a) runs at 30 kb, 850~ ( 0 ) , and at 30 kb, 950~ (O); (b) runs at 20 kb, 950~ showing variations in the ratios of phlogopite and amphibole relative to other minerals; (c) runs at 20 kb, 950~ showing variations in the ratios of olivine, orthopyroxene and clinopyroxene. Abbreviations are as in Figs 2 and 3 ; q, quench phlogopite; p, primary phlogopite. (After Edgar & Arima 1984.)
Genesis of alkaline magmas
~_ApI
(not used in Freestone & Hamilton's experiments) liquid immiscibility could account for the carbonate-rich groundmass in kimberlites and the latestage associated carbonatite dykes and diapirs.
Phi C p x ~
Im. . . .
37
Halogen-bearing
I0
~
8-
u
/
-
e
-
systems
Hards (1976) determined the distribution of certain elements between the melt and vapour phases in granite and nepheline syenite compositions at 1 kb between 680 and 850~ in the presence of C1 and F. He found that C1 was more strongly concentrated in the vapour phase in granitic melts than it was in nepheline syenite melts, whereas F was concentrated preferentially in the liquid phase for both compositions and only entered the vapour phase after crystallization
O
~
6-
NazO
0 ~- ~ 9
e
O~
xcr..-"
if
0
47
J
w,o,o
II
L
I
i
I
~
I
20 40 60 80 Degree of melting (wt.%)
i
I
I00
FIG. 17. Compositions of liquids with varying degrees of partial melting of an average clinopyroxenite nodule from SW Uganda at 30 kb : - - , determined compositions; - - - , calculated compositions based on analyses of minerals. The phases present are given at the top of the figure. Cpx, clinopyroxene: Phi, phlogopite: Ilm, ilmenite: Ap, apatite. (After Lloyd et al. 1985)
Si% + AiaO3
CaO (a)
Na20
the miscibility gap expands with increasing Pco2 and decreasing temperature (Fig. 18). Results at 0.7 kb and l l00~ are plotted on the N a 2 0 (SiO2+AlzO3)-CaO system (Fig. 18(a)). Under the miscibility gap inincreasing P t o t a l = P c o 2 , creases and there is a rotation of tie-lines resulting in more CaO-rich liquids (Fig. 18(b)). From the compositions of the melts, Freestone & Hamilton (1980) concluded that the natrocarbonatites from Oldoinyo Lengai separated from a phonolitic rather than a nephelinitic magma. Because the miscibility gap closes away from the Na20-rich compositions (Fig. 18), exsolution of a carbonatitic melt is more likely in salic than in mafic silicate magmas. Freestone & Hamilton's (1980) results do not support the idea that kimberlite sills and dykes form by an immiscibility process at crustal pressures. Bedson & Hamilton (1981), however, showed that in the presence of H 2 0
SiOa + Ala03
CaO
(b) FIG. 18. Liquid immiscibility in the system NazOCaO-(A1203 + SiO2) at (a) 0.7 kb and 1100~ and (b) 1100~ and varying pressures Pc% = Ptotal. (After Freestone & Hamilton 1980.)
A. D. E d g a r
48
was completed. Sodium was the most common element in the vapour phase of both compositions. These experiments may explain the presence of Cl-bearing sodalite in SiO2-undersaturated rocks but the absence of Cl-bearing minerals from granitic rocks. On the basis of Hard's experiments the presence of F-bearing minerals in both types of rocks is to be expected. On the basis of experiments at 1 kb in the system NaA1SiO4 (Ne)-SiO2-NaC1-H20, Barker (1976) found that in the SiO2-undersaturated portion of the system NaC1 stabilized sodalite whereas in the SiO2-oversaturated part it has no effect on phase relations because of immiscibility between NaCl-rich liquid and SiO2 melt. These results suggest that the high C1 content of feldspathoidal rocks is not a cause of their SiO 2 undersaturation and that such rocks are unlikely to be produced by anatexis or assimilation of halite-bearing sediments (Appleyard 1974; Ayrton 1974). For the SiO2-undersaturated part of the same system, Binsted (1981) showed that sodalite might exist as a primary mineral between about 25% and 60% Ne in the pseudobinary ( A b + Ne)gs(NaC1)5 join (Binsted 1981, Fig. 13). He found that melts crystallizing sodalite follow a non-linear trend (Binsted 1981, Fig. 14) owing to partial replenishment of NaCI in the melt from the fluid reservoir. With increasing NaC1, nepheline will crystallize as the second phase following albite. These relationships may be important in the genesis of agpaitic rocks such as the lujavrites in Ilimaussaq, Greenland, where sodalite is a primary mineral.
sial for many years. Taylor & Mackenzie (1975) studied differences in leucite morphologies with different cooling methods at 2 kb Pnz o in the system NaA1SiO4-KAISiO4-SiOz-H20. As shown in Fig. 19 the compositions of the leucites produced appeared to depend on their morphologies. From the common occurrence of natural pseudoleucites along the leucite-analcite join, Taylor & MacKenzie (1975) suggested that metastable Na-rich leucites are produced by extremely rapid alkali-ion exchange rather than by the breakdown of high-temperature leucites or by leucite breaking down to nepheline+ feldspar at the reaction point in this system.
Primary versus secondary origin of analcites Experimental studies bearing on the problem of the origin of analcite phenocrysts in a few rare volcanic rocks have been carried out by Roux & Hamilton (1976) in the system NaA1SiO4KA1SiO4-SiO2. They found that analcite coexisted with liquid above about 5 kb PH2o and 600~ supporting the concept of primary analcites. In contrast, Gupta & Fyfe (1975) found a very rapid reaction between leucite and saturated NaCI solutions at 1 kb Pn2 o at low temperatures. This supported the idea that these natural analcite phenocrysts could be transformed from leucites in a low-temperature hydrothermal environment. The plausibility of both of these mechanisms occurring in nature has been discussed by Edgar (1984).
Other experiments on felsic alkaline rocks Pseudoleucites The origin of pseudoleucites has been controver-
Dolfi et al. (1978) and Ruddock & Hamilton (1978) investigated the KA1Si206 (Lc)-
Zoned
c Ne
"x,~~Ks Greatly exsolved
leucite
Slightly ( I11r-='_lib IIIII
exsolved I--'l
leucite
~
FIG. 19. Recalculated microprobe analyses and morphologies of leucites crystallized from a bulk composition (+) crystallized in part of the system NaA1SiO4-KA1SiO4-SiO2-H20. Lines join spots probed on leucite crystals to compositions in the system. (After Taylor & Mackenzie 1975.)
Genesis of alkaline magmas CaMgSi20 6 (Di) and KA1Si206-CaMgSi20 6S i O 2 - H 2 0 - C O 2 systems respectively. In the former system, Dolfi et al. showed that increasing pressure up to 12 kb increased the primary diopside field at the expense of leucite. In experiments at 4 kb in the system KA1Si20 6CaMgSizO6-H20, Ruddock & Hamilton (1978, Fig. 19) indicated that with increasing pressure the fields of leucite and quartz contract while those of diopside and sanidine expand. Phlogopite is restricted to compositions with low diopside contents. On the basis of these results they explain the common phenocryst assemblage of diopside and phlogopite along with groundmass quartz and potash feldspar in many lamprophyres. High-pressure experiments up to 30 kb on the composition Lc60DisQ35 with H 2 0 and CO2 (Ruddock & Hamilton 1978, Fig. 19) confirmed these conclusions. These experiments also suggested that minettes are derived by partial melting of the mantle and that primary carbonate minerals in these rocks could have formed at mantle pressures. In order to determine the effect of CaA12Si2Os
49
(An) on phase relations in the system NaA1SiO4KA1SiO4-SiOz-H20, Norris & MacKenzie (1976) and Whiteley (1981) have determined phase relations at 1 kb PH20 with 3, 5 and 10 wt.% CaA12SizOs. With increasing An contents, the leucite and alkali feldspar fields are reduced as the plagioclase field expands. For the 3% An plane, Norris & MacKenzie (1976, Fig. 30) located a eutectic between nepheline, K-rich feldspar and Na-rich feldspar. For the 5% and 10% An planes (Norris & MacKenzie 1976, Figs 31 and 32) the projected phase volumes of Kfeldspar and Na-feldspar cut the leucite phase volume at progressively more K-rich compositions as the K-rich feldspar field contracts. In the absence of a Ca-bearing mafic mineral, such as diopside, these systems more closely approximate felsic magmas.
ACKNOWLEDGMENTS: I am grateful to M. Arima and A. Kolisnik for their assistance in the preparation of the diagrams for this paper. The Natural Science and Engineering Research Council of Canada provided financial support.
References ALLEN,J. C., BOETTCHER,A. L. & MARLAND,G. 1975. Amphiboles in andesite and basalt: I. Stability as a function of P - T - f 0 2 . Am. Mineral. 60, 1069-85. APPLEYARD,E. C. 1974. Syn-orogenic igneous alkaline rocks of eastern Ontario and northern Norway. Lithos, 7, 147-60. ARCULUS, R. J. 1975. Melting behaviour of two basanites in the range 10-35 kbar and the effect of TiO 2 on the olivine-diopside reactions at high pressures. Yearb. Carnegie Inst. Washington, 512-15. ARIMA, M. & EDGAR, A. D. 1983a. High-pressure experimental studies on a katungite and their bearing on the genesis of some potassium-rich magmas of the west branch of the African rift. J. Petrol. 24, 166-87. & - - 1983b. A high pressure experimental study of a magnesian-rich leucite-lamproite from the West Kimberley area, Australia: petrogenetic implications. Contrib. Mineral. Petrol. 84, 228-34. AYRTON, R. 1974. Rifts, evaporites and the origin of certain alkaline rocks. Geol. Rundsch. 63, 430-50. BAILEY, D. K. 1970. Volatile flux, heat focussing and the generation of magma. Geol. J. Spec. Issue 2, 177-86. 1974. Experimental petrology relating to oversaturated peralkaline volcanics, a review. Bull. Volcanol. 38, 637-52. Applications of experiments to alkaline rocks. In: BAILEY,D. K. & MACDONALD,R. (eds) The Evolution of the Crystalline Rocks, pp. 419-69. Academic Press, New York. 1980. Volatile flux, geotherms, and the generation 74,
-
-
-
-
- - 1 9 7 6 .
-
-
of the kimberlite-carbonatite-alkaline magma spectrum. Mineral. Mag. 43, 695-9. - 1982. Mantle metasomatism--continuing chemical change within the earth. Nature, Lond. 296, 525-30. --,TARNEY, J. & DUNHAM, K. C. (eds) 1980. Evidence for chemical heterogeneity in the earth's mantle. Phil Trans. R. Soc. Lond. Ser. A, 297, 134493. BARKER, D. S. 1976. Phase relations in the system NaA1SiO4-SiO2-NaC1-H20 at 400-800~ and 1 kilobar, and petrologic implications. J. Geol. 84, 77-106. - - 1983. Igneous Rocks, 417 pp. Prentice-Hall, Englewood Cliffs, NJ. BARTON, M. & HAMILTON,D. L. 1979. The melting relationships of a madupite from the Leucite Hills, Wyoming, to 30 kb. Contrib. Mineral. Petrol. 69, 133-42. & - - 1 9 8 2 . Water undersaturated melting experiments bearing upon the origin of potassiumrich magmas. Mineral. Mag. 45, 267-78. BASALTIC VOLCANISMSTUDY PROJECT, 1981. Basaltic Volcanism on the Terrestrial Planets, 1286 pp. Pergamon Press, Oxford. BEDSON, P & HAMILTON, n. L. 1981. Kimberlites, carbonatites and liquid immiscibility. Prog. Exp. Petrol., NERC Publ. Set. D, 29-32. BELL, K. & POWELL, J. L. 1969. Strontium isotopic studies of the Burunga and Toro-Ankole regions, east and central equatorial Africa. J. Petrol. 10, 536-72. BERGMAN,S. C. 1987. Lamproites and other potassium-
5o
A. D. Edgar
rich igneous rocks: a review of their occurrence, mineralogy and geochemistry. In: FITTON, G. J. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 103-90. BINSTED, N. 1981. The system Ab-Ne-NaC1-HzO. Prog. Exp. Petrol., NERC Publ. Ser. D, 34-35. BOWEN, N. L. 1922. The reaction principle in petrogenesis. J. Geol. 30, 177-98. --1928. Evolution of the Igneous Rocks, 251 pp. Dover, New York. BREY, G. 1976. CO2 solubility and solubility mechanisms in silicate melts at high pressures. Contrib. Mineral. Petrol. 57, 217-21. 1978. Origin of olivine melilitites--chemical and experimental constraints. J. Volcanol. geotherm. Res. 3, 61-88. & GREEN, D. H. 1975. The role of CO2 in the genesis of olivine melilitite. Contrib. Mineral. Petrol. 49, 93-103. & -1976. Solubility of COz in olivine melilitite at high pressures and the role of CO2 in the earth's upper mantle. Contrib. Mineral. Petrol. 55, 217-30. & -1977. Systematic study of liquidus phase relations in olivine melilitite + H20 + CO 2 at high pressures and petrogenesis of an olivine melilitite magma. Contrib. Mineral. Petrol. 61,141-62. BULTITUDE, R. J. & GREEN, D. H. 1968. Experimental study at high pressures on the origin of olivine nephelinite and olivine melilite nephelinite magmas. Earth planet. Sci. Lett. 3, 325-37. & -1971. Experimental study of crystal-liquid relationships at high pressures in olivine nephelinite and basanite compositions. J. Petrol. 12, 12147. CAWTI-IORN, R. G. 1976. Melting relations in part of the system CaO-MgO-A1203-SiO2-Na20-H20 under 5 kb pressure. J. Petrol. 17, 44-72. CUNDARI, A. & O'HARA, M. J. 1976. Experimental study at atmospheric and high pressures of a mafic leucitite from New South Wales, Australia. Prog. Exp. Petrol., N.E.R.C. Publ. Ser. D, 260-2. DALY, R. A. 1910. Origin of the alkaline rocks. Geol. Soc. Am. Bull. 21, 87-115. DOLFI, O., HAMILTON,D. H. & TREGILA, R. 1978. The system KA1Si206-CaMgSi206 at 4 and 12 kilobars. Prog. Exp. Petrol., NERC Publ. Ser. D, 21-2. DUKE, J. M. 1974. The effect of oxidation on the crystallization of an alkali basalt from the Azores. J. Geol. 82, 524-8. EDGAR, A. D. 1974. Experimental studies. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 355-89. Wiley, New York. 1979. Mineral chemistry and petrogenesi: of an ultrapotassic-ultramafic volcanic rock. Contrib. Mineral. Petrol. 71, 171-5. 1984. Chemistry, occurrence and paragenesis of feldspathoids: a review. In: BROWN, W. L. (ed.) Feldspars and Feldspathoids, pp. 501-32. Reidel, Dordrecht. & ARIMA, M. 1983. Conditions of phlogopite crystallization in ultrapotassic volcanic rocks. Mineral. Mag. 47, 11-19. -& -1984. Experimental studies on K-metasomatism of a model pyrolite mantle and their -
-
-
-
-
-
-
-
-
-
bearing on the genesis of ultrapotassic magmas. Proc. 27th Int. Geological Cong. on Petrology (Igneous and Metamorphic Rocks), Vol. 9, 509-41. V N U Science Press, Utrecht. CONDLIFFE,E., BARNETT,R. L. & SHIRRAN, R. J. 1980. An experimental study of an olivine ugandite magma and mechanisms for the formation of its K-enriched derivatives. J. Petrol. 21,475-97. - - , GREEN, D. H. & HIBBERSON, W. O. 1976. Experimental petrology of a highly potassic magma. J. Petrol. 17, 339-56. EGGLER, D. H. 1974. Effects of CO2 on the melting of peridotite. Yearb. Carnegie Inst. Washington, 73, 215-24. 1978. The effect of CO2 upon partial melting of peridotite in the system Na20-CaO-A1203-MgOSIO2-CO2 to 35 kb, with an analysis of melting in a peridotite-H20-CO2 system. Am. J. Sci. 278, 305-43. --& HOLLOWAY, J. R. 1977. Partial melting of peridotite in the presence of H20 and CO2: principles and review. Oreg. Dep. Geol. mineral Industries Bull. 96, 15-36. • WENDLANDT, R. F. 1979. Experimental studies of the relationship between kimberlite magmas and partial melting of peridotite. In: BOYD, F. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds: Their Geology, Petrology and Geochemistry, pp. 330-8. American Geophysical Union, Washington, DC. EL-GORESY, A. & YODER, H. S. 1974. Natural and synthetic melilite compositions. Yearb. Carnegie Inst. Washington, 73, 359-71. ELLIS, D. E. & WYLLIE, P. J. 1980. Phase relations and their petrological implications in the system MgOSIO2-H20-CO2 at pressures up to 100 kbar. Am. Mineral. 65, 540-56. FREESTONE, I. C. 1978. Liquid immiscibility in alkalirich magmas. Chem. Geol. 23, 115-23. --& HAMILTON, D. L. 1980. The role of liquid immisciblity in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from southeastern Australia utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. GREEN, D. H. 1969, The origin of basaltic and nephelinitic magmas in the earth's mantle. Tectonophysics, 7, 409-22. 1970a. A review of experimental evidence on the origin of basaltic and nephelinitic magmas. Phys. Earth planet. Inter. 3, 221-35. The origin of basaltic and nephelinitic magmas. Trans. Leicester Lit. Phil. Soc. 44, 26-54. 1971. Composition of basaltic magmas as indicators of conditions of origin: applications to oceanic volcanism. Phil. Trans. R. Soc. Lond. Ser. A, 268, 707-25. 1973a. Conditions of melting of basanite magma from garnet peridotite. Earth planet. Sci. Lett. 17, 456-65. Experimental melting studies on model -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
9
7
0
b
.
1
9
7
3
b
.
Genesis of alkaline magmas upper mantle compositions at high pressure under both water-saturated and water-undersaturated conditions. Earth planet. Sci. Lett. 19, 37-53. - - , GUPTA, A. K. & TAYLOR, W. R. 1984. Experimental studies on the effects of H20 and CO2 on liquidus relations in Fo-Ks-SiO2 and F o - N e SiO2 at 28 kb. Abstracts, Workshop on Experimental Geochemistry, Monash University, Melbourne. HmBERSON, W. O. 1970. Experimental duplication of conditions of precipitation of high pressure phenocrysts in a basaltic magma. Phys. Earth planet. Inter. 3, 247-54. & RINGWOOD, A. E. 1967. The genesis of basaltic magmas. Contrib. Mineral. Petrol. 15, 103-90. GUPTA, A. K. & FYFE, W. S. 1975. Leucite survival: the alteration to analcime. Can. Mineral. 13, 3613. HAMILTON, D. L., FREESTONE, I. C., DAWSON, J. B. & DONALDSON, O. H. 1979. Origin of carbonatites by liquid immiscibility. Nature, Lond. 279, 52-4. HARDS, N. 1976. Distribution of elements between the fluid phase and silicate melt phase of granites and nepheline syenites. Prog. Exp. Petrol., NERC Publ. Ser. D, 88-90. HELZ, R. T. 1973. Phase relations of basalts in their melting range at PH2o=5 kb as a function of oxygen fugacity. Part I. Mafic phases. J. Petrol. 14, 249-302. 1976. Phase relations of basalts in their melting ranges at PH2o = 5 kb. Part II. Melt compositions. J. Petrol. 17, 139-93. HOLMES, A. 1932. The origin of igneous rocks. Geol. Mag. 69, 543-58. 1950. Petrogenesis of katungite and its associates. Am. Mineral. 35, 772-92. -1965. Principles of Physical Geology (2nd edn), 623 pp. Ronald Press, New York. --& HARWOOD, F. 1937. The petrology of the volcanic rocks of Bufumbira. Mere. Geol. Surv. Uganda, 3 (2), 1-300. HUANG, W. L. & WYLLIE, P. J. 1974. Eutectic between wollastonite I1 and calcite contrasted with thermal barrier in i g O - S i O 2 - C O 2 at 30 kilobars with emphasis to kimberlite-carbonatite petrogenesis. Earth planet. Sci. Lett. 24, 305-10. IRVING, A. J. & GREEN, D. H. 1972. Experimental study of phase relationships in a high pressure mugearitic basalt as a function of water content. Geol. Soc. Am. Meet. Abstracts with Programs 4, 550-1. - - - & PRICE, R. C. 1981. Geochemistry and evolution of high pressure phonolitic rocks from Nigeria, Australia, Eastern Germany and New Zealand. Geochim. cosmochim. Acta, 45, 1309-20. ITO, K. & KENNEDY, G. C. 1968. Melting and phase relations in the plane tholeiite-lherzolite-nepheline basanite to 40 kilobars with geological implications. Contrib. Mineral. Petrol. 19, 177-211. KAY, R. W. & GAST,P. W. 1973. The rare earth content and origin of alkali rich basalts. J. Geol. 81, 65382. KOGARKO, L. N. & ROMANCHEV, B. P. 1982. Phase equilibria in alkaline melts. Zap. vses. mineral. Ova. 8, 167-182 (in Russian). -
-
-
-
-
-
-
-
&
51
KUEHNER, S. M., EDGAR, A. D. & ARIMA, M. 1981. Petrogenesis of the ultrapotassic rocks from the Leucite Hills, Wyoming. Am. Mineral. 66, 663-77. KUSHIRO, I. 1973. Origin of some magmas in oceanic and circum-oceanic regions. Tectonophysics, 17, 211-22. -& YODER, n . S. 1974. Formation of eclogite from garnet lherzolite: liquidus relations in a portion of the system MgSiO3-CaSiO3-A1203 at high pressures. Yearb. Carnegie Inst. Washington, 73, 266-9. LARSEN, E. S. 1940. Petrographic province of central Montana. Geol. Soc. Am. Bull. 51,887-948. LLOYD, F. E. 1981. Upper mantle metasomatism beneath a continental rift: Clinopyroxenes in alkalic mafic lava and nodules from South West Uganda. Mineral. Mag. 44, 315-24. & BAILEY, D. K. 1975. Light element metasomatism of the continental mantle: the evidence and the consequences. Phys. Chem. Earth, 9, 389-416. , ARIMA,M. & EDGAR, A. D. 1985. Partial melting of a phlogopite-clinopyroxenite nodule from south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift volcanics. Contrib. Mineral. Petrol., 91, 321-9. LUTH, W. C. 1967. Studies in the system of KA1SiO4Mg2SiO4-SiO2-H20. I. Inferred phase relations and petrologic applications. J. Petrol. 8, 372-416. MCKENZlE, D. P. 1984. The generation and compaction of partially molten rock. J. Petrol. 25, 713-65. MERRILL, R. B. & IRVING, A. J. 1977. Chemistry and phase relations of an orthopyroxene-bearing transitional alkalic basalt (abstract). Los, 58, 526. --& WYLLIE, P. J. 1975. Kaersutite and kaersutite eclogite from Kakanui, New Zealand--water excess and water deficient melting relations to 30 kilobars. Geol. Soc. Am. Bull. 86, 555-70. MYSEN, B. O. & KUSHIRO, I. 1977. Compositional variations of coexisting phases with degree of melting of peridotite in the upper mantle. Am. Mineral. 62, 843-65. NORRIS, G. & MACKENZIE,W. S. 1976. Phase relations in the system NaA1SiO4-KAISiO4-CaA12Si2OsSiO2 at PH2o = 1 kb. Prog. Exp. Petrol., NERC Publ. Set. D, 79-81. O'HARA, M. J. 1968. The bearing of phase equilibria studies on synthetic and natural systems on the origin and evolution of basic and ultrabasic rocks. Earth Sci. Rev. 4, 69-133. & BIGGAR, G. M. 1969. Diopside-spinel equilibria, anorthite and forsterite reaction relationships in silica-poor liquids in the system C a - M g O A1203-SiO 2 at atmospheric pressure and their bearing on the genesis of melilitites and nephelinites. Am. J. Sci. 267A, 364-90. YODER, H. S. 1967. Formation and fractionation of basic magmas at high pressures. Scott. J. Geol. 3, 67-117. PRESNALL, D. C., DIXON, J. R., O'DONNELL, T. H. & DIXON, S. A. 1979. Generation of mid-ocean ridge tholeiites. J. Petrol. 20, 3-35. , DIXON, S. A., DIXON, J. R., O'DONNELL, T. H., BRENNER, N. J., SCHROCK, R. L. & DYCUS, D. W. 1978. Liquidus phase relations on the join diopside-forsterite-anorthite from 1 atm to 20 kbar; -
-
-
-
-
-
&
52
A. D. Edgar
their bearing on the generation and crystallization of basaltic magma. Contrib. Mineral. Petrol. 66, 203-220. RINGWOOD, A. E. 1975. Composition and Petrology of the Earth's Mantle, 618 pp. McGraw-Hill, New York. ROEDDER, E. 1979. Silicate liquid immiscibility in magmas. In: YODER, H. S. (ed.) The Evolution of the Igneous Rocks: Fiftieth Anniversary Perspectives, pp. 15-58. Princeton University Press, Princeton, NJ. Roux, J. & HAMILTON, D. L. 1976. Primary igneous analcite--an experimental study. J. Petrol. 17, 244-57. RUDDOCK, D. I. & HAMILTON, D. L. 1978. The system KA1SizO6-CaMgSi206-SiO2-H20 at 4 kilobars. Prog. Exp. Petrol., NERC Publ. Ser. D, 25-31. RYABCHIKOV, I. D. & BOETTCHER,A. L. 1980. Experimental evidence at high pressure for potassic metasomatism in the mantle of the earth. Am. Mineral. 65, 915-19. & GREEN, D. H. 1978. The role of carbon dioxide in the petrogenesis of highly potassic magmas. In: Problems of Earth's Crust and Upper Mantle, Trudy Instituta Geologii GeofizikL So An SSR 403, Nauka, Novosibirsk (in Russian). , SCHREYER, W. & ABRAHAM, K. 1982. Compositions of aqueous fluids in equilibrium with pyroxenes and olivines at mantle pressures and temperatures. Contrib. Mineral. Petrol. 79, 80-4. SCHAIRER, J. F. & YODER, H. S. 1970. Critical planes and flow sheet for a portion of the system CaOA1203-SiO 2. Yearb. Carnegie lnst. Washington, 68, 202-14. , TILLEY, C. E. & BROWN, G. M. 1969. The join nepheline-diopside-anorthite and its relation to alkali basalt fractionation. Yearb. Carnegie Inst. Washington, 66, 467-71. , YAGI, K. & YODER, H. S. 1962. The system nepheline-diopside. Yearb. Carnegie lnst. Washington, 61, 96-8. SORENSON, H. (ed.). 1974. The Alkaline Rocks, 622 pp. Wiley, New York. TAKAHASHI, E. 1980. Melting relations of an alkali olivine basalt to 30 kbar, and their bearing on the origin of alkali basalt magmas. Yearb. Carnegie Inst. Washington, 79, 271-6. - & KUSHIRO, I. 1983. Melting of a dry peridotite at high pressures and basalt magma genesis. Am. Mineral. 68, 859-79. TAYLOR, D. & MACKENZIE,W. S. 1975. A contribution to the pseudoleucite problem. Contrib. Mineral. Petrol. 49, 321-33. THOMPSON, R. N. 1972. The one atmosphere melting patterns of some basaltic volcanic series. Am. J. Sci. 272, 901-32. - - 1 9 7 3 . One atmosphere melting behaviour and nomenclature of terrestrial lavas. Contrib. Mineral. Petrol. 41, 197-204. - 1974. Primary basalts and magma genesis. I. Skye, North-West Scotland. Contrib. Mineral. Petrol. 45, 317-41.
Primary basalts and magma genesis. III. Alban Hills, Roman comagmatic province, central Italy. Contrib. Mineral. Petrol. 60, 91-108. TURNER, F. J. & VERHOOGEN, J. 1960. lgneous and Metamorphic Petrology, 694 pp. McGraw-Hill, New York. WADE, A. & PRIDER, R. T. 1940. The leucite-bearing rocks of the West Kimberley area, Western Australia. Q.J. geol. Soc., Lond. 96, 39-98. WATERS, A. C. 1955. Volcanic rocks and the tectonic cycle. Geol. Soc. Am. Spec. Pap. 63, 703-722. WENDLANDT, R. F. & EGGLER, D. H. 1980a. The origins of potassic magmas. 1. Melting relations in the systems KA1SiO~-Mg2SiO4-SiO2 and KA1SiO4-MgO-SiO2-CO2 to 30 kilobars. Am. J. Sci. 280, 385-420. - & 1980b. The origins of potassic magmas. 2. Stability of phlogopite in natural spinel lherzolite and in the system KA1SiO4-MgO-SiOz-H20CO2 at high pressures and temperatures. Am. J. Sci. 280, 421-58. WIalTELEY, C. 1981. Phase relations in the system NaA1SiO4-KA1SiO4-CaA12Si2Os-SiO2-HzO at 1 kb. Prog. Exp. Petrol., NERC Publ. Ser. D, 33-4. WILLIAMS, H. 1936. Pliocene volcanoes of the NavajoHopi county. Geol. Soc. Am. Bull. 47, 111-72. WYLLIE, P. J. 1975. The earth's mantle. Sci. Am. March 1975. - - 1 9 7 7 . Mantle fluid compositions buffered by carbonates in peridotite-CO2-H20. J. Geol. 8 5 , 187-207. - - 1 9 7 8 . Mantle fluid compositions buffered in peridotite-H20--CO2 by carbonates, amphibole and phlogopite. J. Geol. 86, 687-713. - - 1 9 7 9 . Magmas and volatile components. Am. Mineral. 64, 469-500. - 1980. The origin of kimberlite. J. geophys. Res. 8 5 , 6902-10. - & HUANG, W. L. 1975. Peridotite, kimberlite, and carbonatite relations in the system CaO-MgOSiO2-COz. Geology, 3, 621-4. - - & 1976. Carbonation and melting reactions in the system CaO-MgO-SiO2-CO2 at mantle pressures with geophysical and petrological applications. Contrib. Mineral. Petrol. 54, 79-107. YODER, H. S. 1975. Relationship of melilite-bearing rocks to kimberlite: a preliminary report on the system akermanite-CO2. Phys. Chem. Earth, 9, 883-94. -1976. Generation of Basaltic Magmas, 265 pp. National Academy of Sciences. Washington, DC. - 1979. Melilite-bearing rocks and related lamprophyres. In." YODER, H. S. (ed.) The Evolution of Igneous Rocks. Fiftieth Anniversary Perspectives, pp. 391-411. Princeton University Press, Princeton, NJ. - & KUSrtIRO, I. 1972. Composition of residual liquids in the nepheline-diopside system. Yearb. Carnegie Inst. Washington, 71,413-16. ZYRYANOV, V. N. 1981. Phase Correspondence in the Systems of Feldspars and Feldspathoids, 217 pp. Nauka, Moscow (in Russian). - - 1 9 7 7 .
A. D. EDGAR, Department of Geology, University of Western Ontario, London, Ontario, N6A 5B7, Canada.
Nephelinites and carbonatites M. J. Le Bas S U M M A R Y : Carbonatites found in strongly alkaline intra-plate petrographic volcanic provinces are associated with olivine-poor nephelinites and with phonolites. Olivine-rich nephelinites occur in basanitic and alkali basalt provinces, normally without carbonatites. These igneous provinces are marked by epeirogenic crustal uplift. Nephelinites, ijolites and carbonatites form discrete magmatic events within individual complexes and correspond to the silicate-carbonate conjugate immiscible liquids observed in the laboratory. Carbonate liquids of variable alkali content can separate from both nephelinitic and phonolitic liquids. The silicate liquids give rise to pyroxenites, ijolites, nepheline syenites and nephelinitic pyroclast-rich strato-volcanoes. The carbonate liquids lose alkalis and fractionate to s6vite, alvikite and ferrocarbonatite, each with increasing incompatible-element content. Further fractionation of carbonatite magma can produce mineralizing fluids rich in rare-earth elements, F, Ba, U and Th. Dolomite carbonatite forms only in the deeper parts of carbonatitic complexes, perhaps at depths greater than 2 km. Explosive carbonatite volcanism can occur giving widespread carbonate tufts, rarely with lavas. Fenitization characterizes ijolite-carbonatite intrusive complexes. Syenitic fenites containing alkali feldspars, sometimes perthitic, are formed as aureoles 500 m wide around ijolites in granitic terranes, with nepheline syenite commonly formed as a contact-reaction rock. The feldspar-rich syenitic fenite aureoles which develop around the early carbonatites usually contain pure orthoclase or pure albite.
Carbonatites achieved geological respectability only about 30 years ago, largely through the impact of work in Scandinavia and Africa. Br6gger's work (1921), at Fen in S Norway, which was strongly disputed by Bowen (1924) but later re-established by Saether (1957), demonstrated that calcite-rich rocks could have all the characteristics of intrusive magmatic bodies and that they alkali metasomatize (fenitize) their host rocks. In Sweden, von Eckermann (1948) conducted a meticulous study of the Aln6 complex. He maintained that alkali carbonatite magma was the prime intrusive material at Aln6, that the associated alkaline silicate rocks were all metasomatic products of fenitization and that the calcite-rich carbonatite was the residue of the intrusive carbonate magma. Dixey et al. (1935) were the first to identify carbonatites in Africa. Garson re-investigated the Malawi complexes of Chilwa, Tundulu, Kangankunde, Songwe and others, and wrote a series of memoirs culminating in a valuable summary account (Garson 1965). The association of carbonatites with nephelinitic volcanism was first recognized by King (1949) at the Napak complex in Uganda. Tuttle & Gittins (1966) and Heinrich (1966) wrote comprehensive accounts of intrusive carbonatites, but extrusive carbonatites have only recently become widely appreciated although they were first identified in 1960 by Dawson (1962).
Strongly alkaline igneous rocks are notorious for their diversity of rock type and abundance of names. Some names, such as ijolite and urtite, are necessary, but many can be discarded. The names used here are based on the work of the IUGS Subcommission on the Systematics of Igneous Rocks (Streckeisen 1976, 1978). Carbonatites are defined as igneous carbonate rocks with more than 50% modal carbonate minerals. If the carbonate is calcite, the rock is a calcite carbonatite which is called a s6vite if it is coarse grained with adcumulate texture or a micro-s6vite if it is medium or fine grained. Beforsite is equivalent to dolomite carbonatite. A less common variety of carbonatite is alvikite, which typically has a texture of small calcite rhombs packed together and is enriched in incompatible elements such as rare-earth elements (REE), Ba, Mn, Zn and commonly Nb. The boundary between s6vite and alvikite is still uncertain but may be set at 0.4% MnO, 1500 ppm Ba and 2000 ppm REE. The calcite of alvikite is slightly ferroan. Ferrocarbonatite is a ferroan calcite to ankeritic calcite carbonatite even more strongly enriched in some or all of REE, Ba, Mn, Fe, Zn, F and U, with possible lower limits of 1.0% MnO, 5000 ppm REE and 5000 ppm Ba. This rock type is very rare, but it is the main source for mineralization of these elements in carbonatites. Natrocarbonatite is carbonatite lava. It is fine grained, extrusive and composed of nyerereite ((Na,K)2Ca(CO3)2) and gregoryite
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 53-83.
53
54
M. J. Le Bas
((Na,K)2CO3). Some natrocarbonatites are richer in Ca and have calcite in place of gregoryite. The purpose of this contribution is to give a petrological synthesis of the nephelinite-carbonatite complexes, at both plutonic and volcanic levels. It begins with the rise of the primary magma through the uppermost mantle, then considers the processes suffered by the magma when it stops in reservoirs at or near the base of the crust and finally traces the passage of the different magmas, carbonate and silicate, through the crust to the ultimate products at the surface. Kimberlitic carbonatites are not included because they have recently been reviewed by Dawson (1980) and because they are different from nephelinitic carbonatites in their primary lack of alkalis, particularly Na. This lack is expressed by the absence of fenitization around the kimberlitic carbonatites. However, the zonation of the E African and E Siberian ultraalkaline provinces, recently compared by Le Bas (1986), indicates that these two types of carbonatite are genetically related in their mantle sources.
Distribution Although nephelinite-carbonatite complexes are rare, being only a fraction of 1~ of all igneous rocks, their distribution is worldwide (Fig. 1) with ages back to the early Proterozoic. Nephelinite-carbonatite complexes occur in intra-plate environments, both continental and, rarely, oceanic. For instance they occur in the Canary 150 r
I
120 I
I
90 ~
89
60 ~f",./
I
30 T
I
Islands and the Cape Verde Islands, where there is a clear regional association with ocean island basalts. Petrographically there is no distinction between the complexes that occur singly, for example as at Kaiserstuhl in S Germany, and those in groups such as the complexes composing the 20 Cenozoic carbonatitic volcanic centres of the inner portion of the E African province (Le Bas 1980). The E African province (Fig. 2) is zoned, with kimberlites in a central core about 300 km across. Carbonatites, olivine-poor nephelinites and ijolites, but not basalts, comprise the next inner concentric zone 100-300 km wide, and at the edge is a carbonatite-free outer zone ranging in width from 100 to 300 km which is composed of basanites, alkali basalts and some olivine-rich nephelinites (Fig. 3). The large Palaeozoic-Mesozoic E Siberian province is similarly zoned (Dawson 1980; Le Bas 1986). In the Angola province, or half-province if S Brazil is rejoined to it, the distribution is different. There the alkali basaltic centres cluster near the coast line, with carbonatitic complexes further inland and kimberlite diatremes furthest inland. Namibia is similar (Prins 1981). The scale of these distributions suggests deep-seated mantle processes. They are not related to plate motion or plume traces. These alkaline and ultra-alkaline provinces are also marked by epeirogenic uplift (Le Bas 1971, 1980; Baker 1987) which produces broad swells and plateau uplift both continental (e.g. the Rhine upwarp and the East African plateau) and oceanic (e.g. Cape Verde Rise). These uplifts
0 f
I
I
30 C~-/I
60 I
I
90 I
r~.....~,r~J
120 ]
150 [
I
180 I
9
b
~
9
"
';~.
~
FIG. ]. World distribution of nephelinite-carbonatRe complexes: I I , r215
or province of comple•
[
Nephelinites and carbonatites "J,.
SUDAN
..""
".... ...... .'" ~ ,~v
"9 ~ X \"
/
fl
''..
"-
55
"''..
A
ETHIOPI/~
A'"
'/ U "~'.
.....
'^~ "~/
-.""~/ "~ANDA u~
/ U ' ~ IIAI / ." U~ I " 1 ~ As~" ,I i
. /.. /."~
0*
-
/
K
i
;1:/\. - u3-'~."\ -r
~
\\
-/:/
",
KENYA
u
__-/~.'\- " :.
/
I 2 ~ i
TANZANIA
9
~
'
"q'/~'L I I I 3;"~I
"".
IIi1
A " ""
~
~ I~])
Ill
,
\\ \ I' 3\~ j
\
-/
~ ~'~ ~
/
. . . . ,' ,,I~176
,J
' 3 ""---
\
K
;-
-
'u\
,,
\\
L~"~
u)
I
\
x
" . . .A
K'
i
"'
(
\" L~ 250 km "
]'~
\
40~' t
FIG. 2. Map of the E African Alkaline Province (or Lake Victoria Province) showing the zonation ( - - - ) : 1, central zone of kimberlite diatremes (K); 2, inner zone of ultra-alkaline complexes (U); 3, outer zone of alkali basaltic volcanism (A). Radial lines indicate the edge of the E African uplift. Bold lines indicate rift faults. The faults show no apparent relation to the zonation. The inset square marks the area shown in Fig. 3. (RW, Rwanda; BU, Burundi.)
evidently relate to the same deep-mantle processes as those giving rise to the ultra-alkaline magmatism (Bailey 1987). Such up-arching produces tension in the upper crust which is quite independent of the motion of the whole tectonic plate. The two motions, the first vertical, covering an equidimensionat area and emanating from the mantle associated with ultra-alkaline magmatism, and the second essentially horizontal and related to plate motion, are interpreted to be distinct and not necessarily related (Le Bas 1986). It is argued that ultra-alkaline magmatism and regional uplift are not necessary initial stages of the Wilson cycle which describes the break-up of continents followed by sea-floor spreading.
Nephelinite magma There are nephelinites and nephelinites, the two varieties being distinguished by the properties given in Tables 1-3. Those defined as Group I
nephelinites are olivine rich, and are typically associated with alkali basalt and basanite volcanic provinces. Group II nephelinites are rarer and comprise olivine-poor nephelinites which occur with carbonatitic and ijolitic complexes (Fig. 3). The olivine-poor nephelinites commonly have abundant euhedral clinopyroxene phenocrysts, whereas the few olivine phenocrysts often show resorbed crystal margins, sometimes even reaction rims, indicating that olivine was no longer in equilibrium. The abundance of pyroxene usually means that these rocks are metanephelinites rather than nephelinites, and many could be better described as clinopyroxenephyric melanephelinites. Aphanitic varieties are rare. Olivine-poor nephelinite can fractionate to olivine-free nephelinite and to phonolite (Table 1). These phonolites are often glassy, commonly making abundant dykes and plugs, and Lippard (1973) distinguished them (his Gwasi-type phonolites) from the phonolites in and around the
56
M. J. Le Bas 134o +
--
+
r
t 3 8~
4-
ETHIOPIA
+ + +
,~
+
~1 40 ~
4- •
LODWAR
~t~
MOROTO
z
--2 0
NAPAK
~
KENYA
C e n o z o i c l a v a s and pyroclastic deposits
x +
NYAMBENI
I&UISHU
~ Kavirondo ~ M T .
~ A
KENYA
Olivine-poor volcanic A Olivine-rich volcanic
ABERDARES
nephelinitic centres nephelinitic centres
AL
Mixed nephelinitic volcanic centres
SJNGIR I
~
~
NAIROBI Z o n e of o v e r l a p of p r o v i n c e s
Rift
faults
SHOMBOLE
( L.N~tro~ OLDOINYO LENGA :::::K?IMASI
KILIMANJARO
SADIMAN 9 100 L
1
i !
I
ESSIMINGOR 4o
TANZANIA 134~
~
HANANG
~
1380
FIG. 3. Map of Kenya and adjacent areas showing the present distribution of the Group II olivine-poor nephelinitic (and carbonatitic) volcanic centres corresponding to a portion of the inner zone shown in Fig. 2, and the Group I olivine-rich nephelinitic (non-carbonatitic) volcanic centres corresponding to the Kenyan region of the outer zone shown in Fig. 2. Between these two zones lies a zone of mixed nephelinitic types resulting from an overlap or mixing of the inner- and outer-zone magmas.
Kenya rift valley (his Plateau and Kenya types). The latter are derived from basanitic and basaltic sources, and form extensive lavas evidently of low viscosity. The phonolites derived from the nephelinites are rich in Sr (more than 1000 ppm) and Ba (more than 100 ppm, and some more than 1000 ppm), and are richer in total alkalis (about 14%) relative to silica (about 53%) than those derived from plagioclase-bearing m a g m a s w h e n both Sr and Ba contents are less than 100 ppm.
The Miocene lavas of the Kisingiri volcano in W Kenya provide a good example of the normal fractionation trend from olivine-poor nephelinites and they are not dissimilar from those at other centres nearby except in detail. The early stages of fractionation are dominated by the precipitation of m u c h clinopyroxene plus a few oxides and some olivine at first (Fig. 4), producing only a m i n i m a l rise of Si in the m a g m a with fractionation. On some occasions Si falls, partic-
Nephelinites and carbonatites
-
~ e ~6
~ --'lk
57
\
,," /
~.
~09o'~0
9 9
10 : /
I
~
2.. 3
\ 1 phonolites
/
0
e ~.Nx :
l nephelinites
t-~
I
t
I
10
I
t
t
i
wt
I
20 %
t
I
t
t
I
30
I
I
[
40
MgO
FIG. 4. CaO-MgO plot of fields for olivine-rich nephelinites (Group I) and olivine-poor nephelinites or melanephelinites (Group II), and their mutual fractionation product of ordinary nephelinites and phonolites. • is the composition of olivine fractionated; + - - + are successive compositions of clinopyroxenes fractionated from melanephelinites, nephelinites and phonolites. Dotted lines are paths of fractionation, and to the right of the fields are crystal extracts with the pyroxene-to-olivine ratio indicated. (Data from Le Bas (1978), Varne (1968), Saggerson (1970), Spencer (1969), Strong (1972) and EAGRU data file (unpublished.)
ularly if no oxides are precipitated, and melilite nephelinite magma can form. Some data are given in Table 1, and further data can be obtained from the U K - I G B A data base which can be accessed via BGS (London) or the World Data Center-A (Colorado). They show that the production of phonolite from olivine-poor nephelinite is accompanied by a fall in Ca, Mg, total Fe, P and sometimes Ti, with a rise in Si, A1, Na, K and the ratios Fe/Mg and K/Na. The fractionation of trace elements from olivine-poor nephelinites can be shown on twoelement plots, but is best seen on primordial mantle-normalized plots (Figs 5 and 6). Figure 5 is a plot of four samples from Kisingiri (data in Table 1) and shows a marked relative increase in the incompatible elements Th, Nb and Zr, with depletion in Sr, P, Ti and V; the latter is pulled out by the precipitation of pyroxene, apatite and any titanomagnetite. The increase in the ratio Zr/Ti appears to be a sensitive index of fractionation, and a similar rotation of the Zr-Ti tie-line (i.e. Zr/Ti increases) is also seen from the trace data from other complexes. Figure 6 compares melanephelinites of carbon-
atitic complexes of different ages in E Africa and reveals that, while the high-field-strength elements Nd, P, Zr, Ti and Y remain closely similar, the others vary slightly, particularly the largeion-lithophile (LIL) elements. The younger complexes of Homa Mountain and Elgon both show a slight enrichment in Ba and Th, and Ba mineralization is seen at Homa (Le Bas 1977, p. 237). If data ranges were marked on Fig. 6(a), the values of Rb, Sr, Nb and REE would show overlap, except in the case of Homa Mountain where all the data consistently show higher REE and lower K values. This is not instrument bias since all the data were obtained on the X-ray fluorescence (XRF) spectrometers at both Edinburgh and Leicester University. If the Kisingiri and N a p a k averages plotted in Fig. 6(a) are taken as normal for nephelinitic rocks associated with carbonatites (but whether these are nephelinites from before or after possible liquid-immiscible separation from carbonate melt is not yet certain), they should be chemically comparable with other carbonatitic nephelinites. Appropriate data are sparse, but they do compare closely with the Triassic nephel-
58
M. J. Le Bas
TABLE 1. Selected nephelinitic lavas from carbonatitic and other provinces Elgon, WKenya
Kisingiri, W Kenya
Napak, EUganda
RR659 RR585 RR457 U1041
SUN6 SUN89 SUN21
ElK8
ElK6 ElK7
54.73 0.47 21.37 5.13 0.23 0.21 2.70 9.74 5.72 0.16
43.50 2.03 10.30 15.03 0.18 9.18 16.94 1.26 1.68 0.58
42.84 2.77 11.47 15.46 0.29 7.16 15.33 2.76 1.27 0.99
42.20 44.12 2.48 2.61 12.04 13.89 1 3 . 9 7 14.11 0.27 0.25 6.22 5.03 1 5 . 1 7 11.42 3.93 5.52 2.54 2.78 0.88 0.95
101.27 100.46
100.68
97.76
97.23
100.35
99.68
100.68
82 92 333 31 6 749 1098 46 49 88 38 135 18
19 121 291 54 7 1251 1167 67 69 132 50 184 25
12 133 260 55 8 1212 1267 73 74 133 56 206 26
29 117 286 60 15 1857 1946 163 96 154 57 271 36
29 114 211 45 12 1211 910 111 92 158 61 193 25
30 121 218 58 12 1128 684 123 94 157 57 214 28
Major elements (wt.%) SiO 2 TiO 2 AI20 3 Fe203 T MnO MgO CaO Na20 K20 P205
Total
41.99 3.08 12.94 15.22 0.24 7.43 13.65 3.46 1.58 0.89
45.42 49.73 2.56 1.95 1 4 . 8 8 16.72 13.46 10.62 0.23 0.22 4.39 2.36 9.17 6.75 7.26 8.35 2.80 3.82 0.93 0.75
100.48 101.10
40.38 42.56 2.52 2.48 1 1 . 6 8 13.62 1 6 . 2 7 14.48 0.25 0.23 6.27 5.52 1 2 . 5 3 10.76 4.42 4.58 2.33 2.50 0.81 0.79
Trace elements (ppm) Ni Zn V Rb Th Ba Sr Nb La Ce Nd Zr Y
47 114 323 40 12 1862 3135 119 105 194 79 284 32
25 119 238 63 11 1161 1342 111 92 174 70 317 29
9 129 143 81 15 1401 1448 132 95 161 132 362 30
4 168 52 141 56 1139 808 336 74 135 40 816 24
RR659, aphanitic melanephelinite lava, top of Kaniamwia, Kisingiri (Leicester XRF); RR585, nephelinite lava, Rangwe, Kisingiri (Leicester XRF); RR457, nephelinite dyke, with katophorite, Kiawindu, Kisingiri (Leicester XRF); U 1041, phonolite lava, trachytic texture, Nyamaji, Homa Bay (Leicester XRF); SUN6, pyroxene-phyric melanephelinite lava, Napak (wet analysis by R. C. Tyler 1966; traces Edinburgh XRF); SUN89, melanephelinite lava, Napak (Leicester XRF); SUN21, aphanitic melanephelinite lava, Okathim (Edinburgh XRF); ELK8, melanephelinite lava above Kitum Cave, E flanks Mount Elgon (Leicester XRF); ELK6, melanephelinite lava, Kitum Cave, E flanks Mount Elgon (Leicester XRF); ELK7, nephelinite lava, Kitum Cave, E flanks Mount Elgon TABLE 2. Worm average nephelinites
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 P205 H20
Group I (olivine rich) (wt.% (SD))
Group II (olivine poor) (wt.~(SD))
39.8 (1.5) 3.0 (0.3) 11.3 (0.9) 5.1 (1.9) 7.9 (1.2) 0.2 (0.01) 12.3 (1.8) 12.8 (1.3) 3.4 (0.3) 1.1 (0.3) 0.9 (0.6) 2.1 (0.6)
41.5 (2.0) 2.8 (0.3) 11.4 (1.6) 7.4 (1.5) 6.0 (1.0) 0.2 (0.01) 8.1 (2.0) 13.2 (3.2) 3.7 (1.9) 1.8 (1.5) 0.8 (0.6) 3.0 (1.1)
99.9 n=81
99.9 n--97
After Le Bas 1978.
inites of the Karoo province with w h i c h carbonatites are associated (Bristow & Saggerson 1983). The Miocene nephelinites of the southern portion (south of the equator) of the K e n y a rift valley, particularly those in the Kishalduga region, are almost identical (cf. Figs 6(a) and 6(b), and have olivine contents m i d w a y between the rich and poor types defined above (Kishalduga data from R. Tarzey, personal communication). Miocene nephelinites of N K e n y a are m a r k e d l y poorer in K and the more incompatible trace elements. N o olivine-free nephelinites have yet been observed in the northern rift. The few data that are available from the olivine-rich nephelinites of the Massif Central in France and of Greenland, Hawaii and the Comores Islands are all closely similar to the carbonatitic nephelinites, apart from variable K / T h ratios and Rb and K contents.
Nephelinites and carbonatites
Homa Mt., W. Kenya
N Kenya rift
S Kenya rift
59
HF566
HC42
11.108
8.428
15.591
17.548
Massif central 42466
Hawaii
Comores
65KAPAAI 1
Moh24
45.17 2.94 11.16 12.06 0.13 6.48 11.81 2.84 1.75 0.79
53.28 0.59 21.14 4.23 0.16 0.48 3.33 9.97 4.82 0.18
41.71 2.23 11.52 12.22 0.18 12.72 12.64 3.06 0.77 0.41
42.84 2.13 12.53 11.74 0.18 10.64 12.01 3.99 1.27 0.49
42.78 4.05 10.09 15.13 0.20 9.13 15.79 1.62 1.16 0.73
43.46 3.73 9.82 15.49 0.21 8.54 14.67 2.81 1.36 0.68
43.61 3.62 8.80 11.08 0.15 13.44 11.45 2.38 3.66 0.82
40.15 2.55 11.45 14.40 0.22 12.77 12.10 3.51 0.77 0.80
41.37 2.71 11.49 13.89 0.24 11.03 12.10 3.39 1.56 1.00
95.12
98.18
97.45
97.81
100.69
100.76
99.01
98.67
98.78
37 113 350 46 10 4668 1213 105 92 170 71 249 25
4 120 62 104 33 1224 1153 143 103 151 37 490 26
214 78 317 9 7 551 604 69 56 108 39 147 23
180 83 297 17 10 697 685 80 73 135 49 166 24
135 109 357 45 13 623 915 128 105 202 85 345 33
121 114 344 47 13 651 1075 116 96 183 74 335 29
309 80 -78 9 824 764 93 79 174 73 409 23
340 107 290 20 6 780 1050 -53 110 56 171 --
168 127 -47 -1250 880 ----257 --
(Leicester XRF); HF566, pyroxene-phyric melanephelinite plug, N flanks Homa Mountain (Edinburgh XRF); HC42, feldspar-phyric phonolite plug, S flanks Homa Mountain (Edinburgh and Durham XRF); 11.108, olivine nephelinite lava, Kasorogol, S Turkana (Edinburgh XRF); 8.428, olivine nephelinite, Nasaken, S. Turkana (Edinburgh XRF); 15.591, olivine nephelinite, S. Nakuru (Edinburgh XRF); 17.548, olivine nephelinite, Kishalduga, Narok (Leicester XRF); 42466, olivine nephelinite, Cantal, France (Downes 1984); 65APAAll, olivine nephelinite, Honolulu (Clague & Frey 1982); MOH24, olivine nephelinite, Moheli Island, Comores (Strong 1972). T h e similarities of the trace-element signatures in Fig. 6 and Table 1 indicate broadly similar chemical sources for the m a g m a s in the mantle, w h e t h e r or not carbonatites are associated. A c h a r a c t e r c o m m o n to all nephelinites is the e n r i c h m e n t in K , Th, N b and the light R E E w h i c h points to a m a n t l e source rich in the heatproducing elements, although the K / T h ratio does vary, indicating a variable m a n t l e source for these components. A further study of the detailed variation in nephelinitic and other primitive m a g m a s across the zoned alkaline igneous province of E Africa, c o m p a r i n g early Cenozoic with later Cenozoic m a n t l e m a g m a t i c products, is being carried out at present by R. Tarzey. A similar investigation of the Cape V e r d e m a g m a tism is being u n d e r t a k e n by N. Hodgson. The olivine-rich and olivine-poor nephelinites, in terms of their m a j o r - e l e m e n t chemistry, could
be derived from e a c h other, but their geographical separation by some h u n d r e d s of kilometres into the c o n t e m p o r a n e o u s outer and inner zones of the E A f r i c a n province shows that distinct processes occurred in the two cases.
Magma plumbing These two types of nephelinites are distinct chemically and physically (Table 3). The olivinerich nephelinites have high Mg and N i contents a n d few or no fractionation products. T h e y build small volcanic cones with an occasional abund a n c e of lherzolite nodules, w h i c h m a y be t a k e n to indicate that these m a g m a s c a m e directly from the m a n t l e with few or no intervening crustal reservoirs. In strong contrast, the olivine-poor nephelin-
M. J. Le Bas
6o
~
U
1041
RR 4 5 7
-----+
10 3
"
..
."
~
"...
102
o
i ...,.
101
KISINGIRI
LAVAS
W. K E N Y A
100
I
Rb
I
I
I
1
I
I
I
I
I
K Th Ba Sr Nb La Ce Nd
I
P
1
Zr Ti
I
Y
TI
V
I -
Zn
FIG. 5. Primordial mantle (PM) normalized plot of four selected lavas from the Kisingiri strato-volcano (Table 1). Note the decreasing coherence of distribution, particularly of the high-field-strength elements during fractionation. The PM values of Wood et al. (1979) were used; V was taken as 84 ppm.
ites make large central strato-volcanoes and almost never carry mantle-type xenoliths. They exhibit strong fractionation to phonolite and ijolite, which suggests that magma produced in the mantle must have stopped in a magma reservoir during its upward passage to the surface. This is corroborated by the low Mg number and Ni content of magmas that reach the surface. The question arises: why should one type of nephelinite consistently stop on its rise through the continental lithosphere, and the other not? Trace elements do not support the argument of mantle chemical heterogeneity as the cause. Therefore, either the physical conditions of passage of the two differ, i.e. one portion of the lithosphere is easier to penetrate than another, or the physical properties of the two types of magma must differ. No lateral variation of the lithosphere is apparent in any of the provinces and therefore the second alternative must be examined. The density of primitive nephelinite magma would be about 2.9-3.0 g cm- 3, whilst the density contrast across Moho from mantle to continental crust is about 3.4-2.8 g cm-3. Thus a nephelinite magma would be expected to stop rising in the lower continental crust, but perhaps not in the lower oceanic crust (density about 3.0 g cm-3). Why then do nodule-bearing nephelinites pass rapidly up through continental crust without stopping ? The answer may lie in the volatile content. Nephelinites associated with carbonatites by liquid immiscibility must be carbonated nephelinite magmas and hence have a high C O 2 / ( H 2 0 -kCO2) volatile fraction. If, however, the COz/ ( H 2 0 + C O 2 ) ratio is lower, the nephelinite magma would be more hydrous and less viscous
500
I.
/ I "t,,, /
~o
200
\\
/_
,R,
,?-~
i~
,
-
I00 o >o
q
d
50
1:;7 ;;2~ 20
-~0 x)\\_
i
]
I
L
i
I
i
I
b K Th Ba Sr Nb La Ce Nd P Zr Ti
Y
9 S Kenya rift
L-~'XX
[] N Kenya rift
\
-
'~
o Kisingiri (Miocene) I
-
s
9 Napak (Miocene)
I
[]
I0
Rb K Th Ba Sr Nb La Ce Nd P Zr Ti
Y
FIG. 6. (a) Normalized plot of average trace-element abundances for melanephelinites from four Tertiary carbonatitic volcanic complexes in E Africa (Fig. 3); (b) normalized plot of averaged values of trace elements in olivine nephelinites from the northern and southern sectors of the Kenya rift valley.
Nephelinites and carbonatites
6i
TABLE 3. Contrasts between nephelinites of Groups I and H Character
Principal phenocrysts Clinopyroxene MgO content of rock Mg no. (100Mg/(Mg+Fe2+)) at 100 K/(Na + K) at Associated basalts Chemical fractionation Volcanic structure Percentage pyroclastic Plutonic equivalents Mantle-type xenoliths Examples
Group I (olivine-rich nephelinites) Olivine and clinopyroxene Lilac and Ti rich > 10 wt.% 74 18 Abundant Very slight Small parasitic Moderate Rare Common Kenya Rift Comores Islands Hawaiian Islands Texas Massif Central
and would perhaps be able to rise through the lithosphere without stopping. Some dynamic fractionation during ascent may also have taken place. Now that the style of nephelinitic magmatism associated with carbonatites has been distinguished from other nephelinitic developments, the term nephelinitic volcanism in the remainder of this contribution will refer to the olivine-poor nephelinites and their derivatives. During or before retention in a magma reservoir, olivine was precipitated and hence largely lost to the remaining nephelinitic magma. Resorption and reaction rims on remaining olivine phenocrysts indicate that they were no longer stable in the magma. Augite was also being precipitated in abundance. It seems that the initial precipitation of olivine increased the already high CO2 content in the magma, which eventually destabilized any remaining olivine and widened the pyroxene stability field. Cumulate bodies of dunite within carbonatite complexes are rare (e.g. Kovdor, Kola, and Shawa, Zimbabwe), whilst alkali pyroxenites occur in most carbonatitic complexes.
Pyroxene fractionation Well over half the extrusive products of nephelinite volcanism are pyroclastic, indicating a volatile-rich source. Among the products in W Kenya are numerous megacrysts of clinopyroxene up to 3 cm long. The chemistry of these and other clinopyroxenes in nephelinitic rocks is revealing
Group II (olivine-poornephelinites (melanephelinites)) Clinopyroxene and rarely olivine Pale brown or greenish and Ti poor < 10 wt.% 67 24 None Strong Large central High Frequent Very rare W Kenya E Uganda Malawi Siberia Amba Dongar, India
in the deciphering of the upward passage of the magma. Table 4 gives the analyses of a range of pyroxenes as xenocrysts, phenocrysts and groundmass phases from melanephelinites, nephelinites and phonolites. Most of them have low Ti and A1 contents which, if they came from groundmass phases, would be at variance with the higher Ti and A1 contents normally encountered in strongly alkaline rocks (Thompson 1974). Instead, as pointed out by Thompson, the low Ti and AI appear to indicate a high temperature of formation. The high Ca content of these magmas would also favour partition of the Ti into perovskite or sphene, which are both quite abundant as microphenocrysts in these rocks. Until analysed by microprobe, the presence of Cr-diopside among the xenocrysts and phenocrysts in the melanephelinites was not appreciated (Table 4, column 1). Cr-diopside is typical of pyroxenes from mantle-type lherzolites, and sometimes forms whole xenocrysts. These xenocrysts can also be present as colourless to palegreen cores to more normal pyroxene compositions (Table 4, columns 4a and 4b) and fall in the mantle pressure field for chromiferous pyroxenes described by Onuma & Tohara (1984) where Na > A1vI. The presence of Cr-diopside is thought to indicate that the nephelinitic magma does record a history of passage through upper-mantle lherzolite. Olivine, with more than 90% Fo, has not been recognized among the xenocrysts or phenocrysts in the lavas, and the few olivines present range from 86% to 81% Fo. Olivine with Fo > 90% does occur in the cumulates, however.
62 TABLE 4.
M. J. Le Bas
Analyses of clinopyroxenes from melanephelinites, nephelinites and phonolites 2a
2b
53.08 0.62 0.58 . 3.19 0.11 16.16 24.31 0.39 0.00 0.14 98.58
51.67 0.87 1.61 . 6.73 0.13 13.58 23.80 0.66 0.00 0.15 99.20
51.25(0.55) 1.06 (0.19) 1.84 (0.26) . . 4.63 (0.28) 0.05 (0.02) 14.83 (0.34) 24.08 (0.09) 0.41 (0.03) 0.02 (0.00) 0.28 (0.07) 98.45
51.94 49.95 0.71 1.95 1.53 3.09
Total
53.20(0.45) 0.45 (0.07) 1.00 (0.14) . . 2.93 (0.13) 0.06 (0.01) 16.85 (0.18) 23.43 (0.37) 0.48 (0.07) 0.00 0.86 (0.27) 99.33
48.51 1.56 4.49 3.19 4.34 5.44 4.81 0.05 0.05 0.14 15.12 14.02 13.58 23.81 24.55 22.40 0.47 0.40 0.96 0.01 0.02 0.00 0.44 0.01 . 98.42 99.47 99.64
46.55 47.80 50.45 50.41 2.19 1.37 0.80 0.82 5.04 4.82 2.30 2.27 4.01 -12.80 13.51 7.19 17.00 8.45 8.27 0.30 0.65 0.55 0.52 10.29 6.05 4.87 4.53 22.07 19.20 13.98 13.43 1.22 2.79 5.08 5.89 0.18 -0.00 0.34 . . . . 99.04 99.68 99.28 99.99
Si Ti A1 Cr Fe 3+ Fe 2+ Mn Mg Ca Na K Ca Mg Fe* Na Mg Fe z+ + M n Mg/(Mg + Fe*)
1.956 0.013 0.046 0.025 . 0.090 0.002 0.924 0.923 0.034 . 47.6 47.7 4.7 3.4 91.2 5.4 91
1.971 0.017 0.025 0.004 . 0.099 0.003 0.894 0.968 0.028 . 49.3 45.5 5.2 2.8 90.0 7.2 90
1.940 0.025 0.071 0.005 . 0.211 0.004 0.760 0.958 0.048 . 49.6 39.3 11.1 5.0 79.1 15.9 80
1.923 0.030 0.081 0.008 . . 0.145 0.002 0.830 0.968 0.030 . . 49.8 42.7 7.5 3.1 86.9 10.0 85
1.943 1.867 1.817 0.020 0.055 0.044 0.068 0.136 0.198 0.013 . . . 0.090 0.136 0.170 0.151 0.002 0.002 0.004 0.843 0.781 0.758 0.954 0.983 0.899 0.034 0.029 0.070 . 49.3 50.8 47.3 43.6 40.3 39.9 7.1 8.9 12.9 3.5 3.2 7.8 86.5 85.9 84.9 10.0 10.9 7.3 86 82 76
1.786 0.063 0.228 . 0.116 0.231 0.010 0.589 0.907 0.091 0.009 48.9 31.8 19.3 10.9 70.5 18.6 62
1
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20
Cr203
.
.
3
4a
4b
5
6
7
8
9
1.871 1.949 1.942 0.040 0.023 0.024 0.222 0.105 0.103 . . -0.372 0.392 0.557 0.273 0.266 0.022 0.018 0.017 0.353 0.280 0.260 0.805 0.579 0.554 0.212 0.381 0.440 --0.017 46.4 38.0 37.2 20.3 18.4 17.5 33.3 43.6 45.3 23.6 40.0 46.8 39.4 29.4 27.7 37.0 30.6 25.5 38 30 28
1, Cr-diopside xenocryst from strongly porphyritic olivine melanephelinite lava, Kisingiri volcano, W Kenya (RR710) (mean of six wavelength-dispersive microprobe analyses (WDMAs); standard deviation to 2tr given in parentheses); 2a, core of pale-green zoned pyroxene in block of pyroxenite ejected from Oldoinyo Lengai volcano, N Tanzania (24407) (energy-dispersive microprobe analysis (EDMA)); 2b rim of 2a; 3, euhedral megacryst (2 cm) of black diopside in agglomerate between melanephelinite lavas of Kisingiri volcano, Kenya (K79) (mean of four WDMAs; standard deviation given); 4a, core of colourless zoned Cr-diopside in phenocryst in normaltype melanephelinite lava, lower N flanks of Kisingiri volcano, Kenya (U16) (WDMA); 4b rim of 4a; 5, bulk pyroxene from fine-grained melanephelinite lava, S Ruri, W Kenya (S197) (wet analysis by H. Lloyd); 6, bulk pyroxene from nephelinite lava, S. Ruri ($33) (wet analysis by H. Lloyd); 7, Aegirine-augite phenocryst in feldspar-phyric phonolite lava of Kisingiri volcano, Wasaki Peninsula, W Kenya (U410) (EDMA); 8, Aegirineaugite microphenocrysts in glassy phonolite, volcanic plug, S Ruri (S 17) (wet analysis by H. Lloyd); 9, Aegirineaugite microphenocrysts in porphyritic phonolite, volcanic plug, S Ruri ($29/1) (wet analysis by H. Lloyd). Fe*, total F e + Mn. Titaniferous lilac-coloured Al-augite phenocrysts are exceedingly rare in these nephelinites, but are a b u n d a n t in the olivine-rich nephelinites and have been interpreted (Fodor et al. 1982) as being precipitated at pressures in the vicinity of 1520 kb during passage t h r o u g h the u p p e r mantle. Since Cr-diopside occurs but Ti-Al-augite is rare in the olivine-poor nephelinitic rocks, it is inferred that the latter were not precipitated during the passage of the p a r e n t m a g m a t h r o u g h the mantle. Similar chromiferous diopsides in pyroxenite (e.g. Table 4, c o l u m n 2a) are inter-
preted as cumulates, w h i c h further suggests that a c c u m u l a t i o n in a deep reservoir did occur. T h e more n o r m a l pyroxene phenocrysts in the m e l a n e p h e l i n i t e are a b u n d a n t low-Ti low-A1 augites (Table 4, columns 2b and 3) w h i c h zone out to more aluminous augites (Table 4, columns 4b and 5). They also show m o d e r a t e Fe enrichment, indicative of fractionation in a m a g m a chamber. Large euhedral black glistening megacrysts showing signs of resorption on the crystal faces also occur. T h e y have the same composition as
Nephelinites and carbonatites the phenocrysts (Table 4, column 3) and hence presumably have the same source. Evidently they were transported directly to the surface without stopping. The rare presence of pargasitic amphibole in some E African pyroclastic rocks (Varne 1968; Le Bas 1977) may be related to fractionation at mantle pressures as proposed by Varne, but they may also be the product of strongly hydrous conditions obtaining at only slightly lower pressures in a deep crustal magma chamber and the consequent explosive eruption of this magma to the surface. If this interpretation of the magmatic plumbing is correct, then the subvolcanic coarse-grained bodies of ijolite exposed by erosion at carbonatitic centres are products of nephelinitic magma rising from the lower crustal reservoir which stopped and crystallized. Therefore, the subvolcanic bodies would not mark the position of the main magma chambers feeding the volcanic structures above. Crystal fractionation of the melanephelinitic magma in the lower crustal reservoir produced nephelinitic and finally phonolitic liquids with increasingly Fe-rich sodic pyroxenes (Table 4, columns 6-9) as plotted in Fig. 7. Compositions richer in acmite than shown in
63
Fig. 7 do occur, but only rarely, and are confined to needles growing in late-stage fractionates. Also significant, with a bearing on the plumbing, is that some pyroxenes have green cores and show reverse zoning. They are quite common in melanephelinites and nephelinites. Table 5 lists analyses of three such zoned pyroxenes which have cores richer in Fe and Na than their colourless rims. The green cores correspond to pyroxenes from phonolites, while the rims correspond to the nephelinitic and melanephelinitic hosts. These give evidence for magma mixing, whereby pyroxene phenocrysts from phonolite are mixed into incoming magma to produce the reverse-zoned pyroxenes which can then be carried to the surface. Such a process necessitates a magma reservoir as proposed above. Analysis 3 in Table 5 further shows that mixing did not always lead to eruption, but that cumulative pyroxenites could be produced. Similar magma mixing revealed by green cores with reverse zoning is well known in many parts of the world. Brooks & Prinzlau (1978) give an excellent summary covering such mixed products from Antarctica, Africa, Europe, America and Greenland.
Crustal intrusive processes
\
~/
\
ee/
...:,.'.. Fe 2+ + Mn
FIG. 7. Na-(Fe 2+ + Mn)-Mg at.% plot of pyroxenes taken from Table 4 together with unpublished energy dispersive microprobe data on further E African samples: II, Cr-diopside from RR710 (Table 4, column 1); A, megacryst from K79 (Table 4, column 3); + - - +, zoned core-to-rim pale-green pyroxene in pyroxenite (Table 4, columns 2a and 2b); O, phenocryst and groundmass pyroxenes (the tie-line with the arrow indicates zoning); - - - , pyroxene trend for the sequence pyroxenite to ijolite to nepheline syenite in the sub-volcanic complexes of E Uganda (Tyler & King 1967) and W Kenya (Le Bas 1977, and unpublished data). The path from melanephelinite to nephelinite to phonolite corresponds to fractionation beginning with lowfO2.
Recent geochemical modelling by the author using major-element, trace-element and isotopic data shows that mixing of phonolite with melanephelinite magma can also produce lavas of mugearitic and trachyandesitic compositions. Such lavas are rare, but have been recognized in the nephelinitic volcanoes of E Africa (King 1949; Le Bas 1977). In the absence of any basalts in these volcanoes, they had been thought to be products of crustal contamination of melanephelinite magma. Another feature indicative of magma residing in a continental crustal chamber for long periods is that, the more extreme is the phonolitic fractionate from a nephelinite parent, the greater is the STSr/86Sr ratio. In W Kenya the 8781"/8651" initial ratios of melanephelinites are usually between 0.7032 and 0.7038 (Norry et al. 1980), whilst phonolites and trachyphonolites have values of 0.7046 and 0.7053 respectively (Le Bas 1977, p. 140). These higher values in the fractionates are interpreted as the product of crustal contamination. The density-contrast conditions cited above for arresting the upward passage of nephelinitic magmas near the mantle-crust boundary hold true only for continental crust. They cannot be applied to oceanic islands such as the Canary and
M . J. Le Bas
64
TABLE 5. Analyses of clinopyroxenes with green cores and reverse zoning 1
Core
2
Rim
SiOz TiOz A1203 FeO T MnO MgO CaO NazO CrzO 3 Total
51.12 0.52 1.96 15.50 0.36 8.47 19.44 2.21 . 99.58
50.80 1.50 2.19 6.50 0.12 14.36 23.42 0.45 . . 99.34
Si Ti A1 Cr Fe r Mn Mg Ca Na Na Mg Fe 2+ +Mn Mg/(Mg + Fe*)
1.972 1.903 0.015 0.042 0.089 0.098 . . . 0.500 0.204 0.012 0.004 0.487 0.802 0.804 0.940 0.165 0.033 16.2 3.4 47.8 82.3 36.0 14.3 57 85
3
Core
Rim
Core
Rim
47.56 1.53 5.12 17.90 0.39 5.32 19.98 1.88 . 99.68
48.32 2.15 4.67 7.58 0.12 12.53 22.58 1.02
51.26 0.37 1.07 15.69 0.35 8.31 20.36 2.09 0.16 99.66
50.43 0.87 2.19 10.11 0.22 11.45 22.88 0.76 0.19 99.10
1.987 0.011 0.049 0.005 0.509 0.011 0.480 0.845 0.157 15.8 48.2 36.0 57
1.924 0.025 0.098 0.006 0.323 0.007 0.651 0.935 0.056 5.9 67.9 26.2 72
98.97
1.866 0.045 0.237
1.830 0.061 0.208
0.587 0.013 0.311 0.840 0.143 15.4 33.5 51.1 40
0.240 0.004 0.707 0.916 0.075 8.3 78.5 13.2 86
.
1, Fine-grained nephelinite lava with occasional phenocrysts of pyroxene, mostly with green cores and colourless rims, showing reversed zoning, Kisingiri volcano, Kenya (K61) (EDMA); 2, hybrid fine-grained lava with scattered large phenocrysts of phonolitic pyroxene rimmed by groundmass composition pyroxene of melanephelinite host, NE flanks of Kisingiri volcano (U926) (EDMA); 3, apatite pyroxenite block from Oldoinyo Lengai volcano, N Tanzania (24420) containing subhedral pyroxenes with green cores and pale-green margins (EDMA). Cape Verde Islands, where nephelinites and carbonatites occur and where the distinction between the olivine-rich and olivine-poor nephelinites is not so clear. Green pyroxene cores exist, and so do mantle nodules. If a deep crustal reservoir is lacking, then fractionation must take place during ascent or in high-level chambers such as those now occupied by the subvolcanic plutonic complexes where fractionation and segregation can be seen to have occurred. A final question may be asked concerning the stage and depth at which magma chambers are formed in continental or oceanic crust. What relation do these feeders have to the onset of liquid immiscibility which, as will be shown below, affected the carbonated nephelinite magma and permitted the formation of conjugate silicate and carbonate magmas? If the unmixing of liquids is an exothermic process, its onset together with loss of carbonate fraction from the silicate liquid would temporarily raise the liquidus temperatures of the magmas until each had precipitated its more refractory phases, mainly
olivine and/or pyroxene +oxides and a p a t i t e + oxides respectively, until the remaining liquids were again buoyant. In summary, the following magma plumbing is envisaged. A carbonated olivine-rich magma is derived by partial melting of the asthenosphere. This rises through the mantle, gathering lherzolite fragments from the walls and perhaps precipitating some olivine. On passing into the base of the crust, the magma stops temporarily in a reservoir where olivine is precipitated, C02 builds up and the pyroxene stability field expands. The limit for liquid immiscibility is reached and carbonatitic magma separates. Pyroxene crystallizes to yield pyroxenite cumulates and the magma differentiates to phonolite. Influxes of new primitive magma mix with the phonolite producing reversezoned pyroxenes and, in extreme cases, new melts of mugearitic composition. All these magmas can rise directly to the surface and erupt as lavas, but some stop at subvolcanic levels and crystallize there, forming small plutons of ijolite and nepheline syenite.
Nephelinites and carbonatites
However, kimberlites do seem to produce a residual carbonatitic fraction in some cases, such as at Benfontein, S Africa (Dawson & Hawthorne 1973). Gittins (1978) likewise describes some carbonatitic and ultramafic rocks from the Cargill complex, N Ontario, as related by simple magmatic fractionation. The third possibility is that carbonatite and nephelinite magmas can be derived as independent melts in the asthenosphere, a view upheld by Woolley & Jones (1987), but, as Treiman & Essene (1983) point out, carbonatites emanating from partial melting of mantle peridotite would be Na-poor. Derivation of a carbonate melt from the mantle has been shown by the experimental work of Eggler (1978) and Wyllie (1978, 1980). They show that at about 30 kb in a phlogopitebearing mantle with a high CO2/H20 ratio the initial melt could be carbonatitic and associated with kimberlitic or melilitic magmas. Wyllie further showed that at lower CO2/H20 ratios the upper asthenosphere can produce an alkaline ultrabasic silicate partial melt from which an immiscible alkali-rich carbonate liquid would probably separate.
Liquid immiscibility The process of the separation of two immiscible liquids is considered to be central to the interpretation of the evolution of carbonatitic magmatism. Therefore liquid immiscibility must be examined to see whether it best accords with the petrological evidence. First, however, the alternatives for the production of carbonatite magma should be briefly considered, with the assumption that carbonatites are magmatic and not hydrothermal metasomatic deposits (Bulakh & IskozDolinina 1978). There are three possible styles of origin. First, the carbonatite magma is a product of melting of limestone in the crust. This was upheld by Shand (1949) among others, but experimental and geochemical work on synthetic systems, mainly by Wyllie (well summarized in his 1978 paper), has demonstrated the igneous nature of carbonatites. The second possibility is that carbonatite is the product by fractional crystallization of nephelinitic and ijolitic magma (King 1965). The discreteness of many ijolite and carbonatite intrusions and the almost complete lack of mineralogical and chemical continuity between the silicate and carbonate magmas is not in keeping with carbonatite as a residual magmatic fraction from nephelinite magma, a point discussed by Prins (1978). Volcanic rocks midway between nephelinite and carbonatite are unknown except for a few pyroclastic examples. Plutonic rocks midway between ijolite and carbonatite are uncommon, and bimodality can be demonstrated in most complexes. In the few complexes where transitional rock types occur, the structural and textural relations of these rocks in both the field and the laboratory usually show that ijolitic rocks have been invaded by carbonatite, producing hybrids. 900
L1
65
Experimental evidence In their study of possible liquid immiscibility, Koster van Groos & Wyllie (1966, 1968, 1973) observed liquid immiscibility along the join NaA1Si3Os-Na2CO3 above 870~ and at 1 kb or more, but not at atmospheric pressure (Fig. 8). Introducing water into the system (Koster van Groos & Wyllie 1968) Caused the field of liquid immiscibility to shrink. At lower temperatures (about 700-750~ albite is joined by cancrinite on the liquidus. The accompanying H20-COz vapour contained Na20 and SiO2, and such vapours are considered to be the synthetic
[
L 2----~ LI+L2
T~ 800
NO+L2
NC+LI+L 2
_
~
NC+L~
700 Albite
I
20
40 WT %
I 6o
I 8o
Na2CO3
FIG. 8. Highly simplified phase diagram of the join Ab-NC with 10% H20 at 1 kb pressure. The fields of cancrinite and vapour have been omitted. L1 and L2 are immiscibleliquids. (After Koster van Groos & WyUie 1973.)
66
M. J. Le Bas
equivalents of N a - K fenitizing fluids known to emanate from carbonatite magmas. In 1973 Koster van Groos & Wyllie added plagioclase (Abs0 and Abs0) to the runs and observed liquid immiscibility along both joins between the plagioclase and Na2CO3. Cancrinite lay on the solidus and the carbonate liquid was richer in Ca than the silicate liquid which was dominantly albite and nepheline normative. Watkinson & Wyllie (1971) studied the comparable system NaA1SiO4-CaCO3-H20 at 1 kb but found no liquid immiscibility along the join nephelinecalcite. Had higher pressures been used, the experiments of Freestone & Hamilton (1980) indicate that liquid immiscibility would have been observed. Besides revealing the extent of liquid immiscibility in certain silicate-carbonate systems, these experiments showed (Fig. 8) that the silicate melt could have moderate quantities of dissolved carbonates, the proportion varying with water and Ca contents. In contrast, the carbonate melt usually had only about 5~ dissolved silicates. The proportion of silicate in carbonate melts and carbonate in silicate melts decreases as the temperature reduces. These distributions of one component dissolved in another correspond closely to those seen in natural ijolites and carbonatites. Calcite is sometimes observed as a late-stage trace constituent in ijolites, urtites and nepheline syenites, and fluid inclusions trapped in apatites in ijolites also reveal the presence of carbonates early in the history of crystallization of those melts (Rankin & Le Bas 1974a). Introducing K into the phase system likewise permits liquid immiscibility (Wendlandt & Harrison 1979). P and Ti also enhance the occurrence of liquid immiscibilityby increasing the polymerization of the melt (Freestone 1978; Mysen e t al. 1981). P is particularly relevant to carbonatites which are often rich in apatite. Because previous experiments did not contain a pyroxene component, Verwoerd (1978) carried out melting experiments at 2 kb on compositions along the join synthetic ijolite (5acmite +3 nepheline + 2 Na-disilicate) and an Na-Ca-natrocarbonatite end-member in order to obtain synethtic melts close to those of natural compositions. The results confirmed the liquid immiscibility shown by Koster van Groos & Wyllie. Freestone & Hamilton (1980) furthered the study of liquid immiscibility by conducting experiments over the range 0.7-7.6 kb and 9001250~ Using natural nephelinite and phonolite starting materials from Oldoinyo Lengai, together with a simplified synthetic natrocarbonatite, they showed that liquid immiscibility occurs between relatively polymerized and depolymerized liquids
and that both nephelinite and phonolite magmas are immiscible with alkali-rich carbonate melts (Fig. 9). Increasing the pressure widened the immiscibilityfield and increased the partitioning of Mg into the carbonate fraction but decreased that of K. Decreasing the temperature also widened the immiscibilitygap. Most nephelinites and phonolites of carbonatite complexes plot near the solvus for 7.6 kb at 1100~ (Fig. 9). These conditions are interpreted as those closest to the natural ones. Bedson (1983) took the experiments further, extending the pressure range to 25 kb. Amongst various distribution coefficients measured, he found that the rare earths were approximately equally distributed between the silicate and carbonate liquids, but that Ta, and more particularly Hf, were enriched in the silicate liquid. Hence immiscible silicate melts coexisting with carbonate melts should have high Hf/Ta ratios, a feature which would not be consistent with crystal fractionation of carbonate melt from silicate (Bedson 1984). Not enough data exist to test this hypothesis at present. Some early experiments by Brey & Green (1976) suggested the possibility that liquid immiscibility might occur between olivine melilitite and carbonatite, but later work (Brey 1978) showed that this was not the case. Similarly, no evidence has been found that kimberlite and carbonatite are related by liquid immiscibility. Natural evidence
At Fen and Aln6 the ijolitic and carbonatitic components do not crop out as discrete intrusions. Instead, the early ijolites and associated fenites are intruded and refenitized by carbonatites. Since the intrusion of carbonatite is frequently A
NazO + K20
IX iquids
~/ /
II
0.7/ ~.> / / /,i L . ~ "
Si02 + AI205
r~o~_
~ i
wt. %
/SOl
X
\\ \
CaO
FIG. 9. Liquid immiscibilityplot showing silicate fields I-VIII (described in the text p. 68) and the natrocarbonatite field IX around the two-liquid boundaries for 0.7, 3.0 and 7.6 kb at 1100~
67
Nephelinites and carbonatites accompanied by extensive brecciation, the ijolites are fragmented and at Fen, where the original extent of the ijolite intrusion was at least 2 km across, relatively unaltered ijolitic and melteigitic rocks are now present only in a small area (less than 1 km across) around Melteig farm in the SW corner of the complex. The rest are permeated and veined by carbonatite, giving the impression that a complete spectrum of compositions from pure silicate to pure carbonate rocks exists with the carbonate as a late-stage component. Tveitasite, hollaite, kasenite and ringite are local names given to these mixed rocks, but since they are hybrid rocks the names are of little value. In some complexes the ijolite and carbonatite intrusions do not penetrate each other but form discrete intrusions. The type locality of ijolite at Iivaara in NE Finland (Lehijarvi 1960) is one such discrete 2 km wide plug, with no signs of carbonatite at the surface. Carbonatites can also form discrete intrusions, as in the 5 km diameter pipe at Sokli on the west flank of the Kola ultraalkaline province (Vartiainen 1980). The field relations in E Africa show distinct bimodality of the intrusions (Le Bas 1977). Most intrusions combine discrete ijolitic and carbonatitic components with few or no hybrids. The clearest example is the Usaki complex (Le Bas 1977) which is an ijolite ring complex 3 km across with micro-ijolite marginal facies but with no development of carbonatite or calcite ijolite (i.e. an ijolite with more than 10% calcite). A few samples do show late-stage interstitial calcite (about 1%) as well as a few calcite veins, and this calcite is interpreted as being the residue of the few per cent of carbonate remaining dissolved in the silicate melt after the immiscible separation of the carbonate melt. The carbonatite partner of this ijolite is the Wasaki subvolcanic complex which occurs 5 km away to the north. The bimodality is particularly well marked on the rare occasions when extrusive volcanic products are preserved. At Oldoinyo Lengai the lavas are either silicate (nephelinite to phonolite) or carbonate. Parallels of this bimodal behaviour are seen in the extrusive products of all other known volcanic structures, even the newly discovered structure on the Cape Verde Islands (Silva et al. 1981). Fluid inclusions provide another line of evidence that liquid immiscibility can separate carbonate and silicate magmas. During heating and homogenization experiments on fluid inclusions contained in the apatite crystals of E African ijolites, a sequence was observed showing liquid immiscibility actually taking place in the heating stage under the microscope (Rankin & Le Bas 1974b). At 960~ a homogeneous liquid
filled the negative-crystal-shaped cavity of the fluid inclusion, but at 950~ small globules appeared and by 800~ the globules had solidified to silicate glass. At 575~ the remaining liquid solidified and instantaneously crystallized to a strongly birefringent mesh of elongated carbonate crystals. Further evidence of immiscibility comes from fluid inclusions in apatites within ijolite, which have bulk compositions close to natrocarbonatite (Le Bas & Aspden 1981). The ijolite-inclusion tie-line (Fig. 10, Ij-IC) attests that the relation is one of liquid immiscibility. It is interpreted that nephelinite magma was constrained in its crystallization path by being on the solvus and that, as it progressively crystallized, so it lost by immiscibility more of its small content of dissolved carbonate. Since the change in bulk chemical composition of the silicate melt as it migrated along the solvus was governed by loss of Ca (precipitation of diopsidic aegirine-augite and apatite), it moved towards urtite (Fig. 10, Ur). During this process the carbonate liquid became richer in Ca (Fig. 10, UC). Hence it is recognized that not only do droplets ofcarbonatite liquid separate from crystallizing nephelinite magma but also the droplets change composition with fractionation of that magma. One criterion for liquid immiscibility is that, not only must the two liquids be in equilibrium with each other, but so also must any additional phase. Several s6vitic carbonatites contain olivine, e.g. Aln6, Jacupiranga, Kaiserstuhl, Mbeya, Oka and Palaborwa, and, where measured, they lie in the range Fo95_ss. This range is similar to those in co-magmatic olivine ijolites, olivinites and olivine-bearing nephelinites. Such composiA
84
Na20 + K20
t, NCb i
~iIc
/-.// / S[O2+ AI203
'JIj
/ ,liquid
\ CoO
FIG. 10. Plot as Fig. 9 showing joins of silicate host and carbonate fluid inclusions for ijolite (Ij-IC) and urtite (Ur-UC) compared with the solvus for 3 kb at 1100~ NCbl and NCb2 are the span of natrocarbonatite analyses from Oldoinyo Lengai.
68
M. J. Le Bas
tions are in keeping with the proposition that the carbonate and silicate magmas could be related by immiscibility. Liquid immiscibility is also demonstrated by cryptic variation in the apatites in carbonatites and ijolitic rocks. Apatites precipitating from ensuing fractions of silicate magmas take different paths from those in fractionating carbonatite magmas, although all commence at a composition common to both silicate and carbonate hosts (Le Bas & Handley 1979). If the ijolites and carbonatites were related by fractional crystallization, a solid-solution phase crystallizing throughout would be expected to show a continuous variation throughout. Lapin & Vartiainen (1983), in their study of apatites from Finnish orbicular carbonatites, also reached this conclusion.
Relation between magmas and the system (Si + AI)-(Na + K)-Ca The carbonate-saturated system (Si + A1)-(Na + K)-Ca reported by Freestone & Hamilton (1980) accounts for 75%-85~ of the components of ultra-alkaline rocks and is therefore suitable for discussion here. When the compositions of common igneous rocks as well as ultra-alkaline rocks are plotted on this phase system using solvus relations at 1100~ (Fig. 9), a significant distribution is seen. This temperature is chosen because melting experiments by Piotrowski & Edgar (1970) on W Kenyan nephelinites, ijolites and nepheline syenites showed that liquidus temperatures lay in the range 1100-1200~ at 1 arm pressure, with solidus temperatures of about 950-1000~ and liquid immiscibility operates only above solidus temperatures. Figure 9 shows that, apart from some unusual rock types (fields IIa and IIb), most compositions are confined to the (Si + A1) corner with fields II and III lining up along an apparent boundary parallel but lower down in the diagram than the 7.6 kb solvus. Assuming an approximately regular interval between the solvus isobars, the apparent boundary can be taken at about 10 kb. The first significant feature is that it is the nephelinites and ijolites (field II) and phonolites and nepheline syenites (field III) which abut the solvus, and that all the other common rock compositions plot away from the solvus. This distribution supports the proposition that, if experimentally determined tie-lines are used (Freestone & Hamilton 1980), only these strongly alkaline compositions are possible candidates for liquids which could be conjugate with C a - N a - K carbonate liquids (field IX).
A second feature is that olivine nephelinites, melanephelinites and melilite nephelinites (field I) do not apparently lie on the solvus, although they could do so at higher pressures or lower temperatures, both of which enlarge the twoliquid field. The latter case of a lower temperature is petrologically unrealistic, but having mafic magmas of field I on the solvus at higher pressure is quite possible in the conditions envisaged for magma chambers deep in the crust. Even if these nephelinitic compositions were not on the solvus, they could become so after slight fractionation to field II compositions. This distinction perhaps explains why melanephelinites are associated with carbonatites but olivine nephelinites, which seldom fractionate, do not have accompanying carbonatites, it being assumed that the nephelinites were carbonated in the first place. It is further possible that if an olivine nephelinite rose high into the crust and then did fractionate, the magma would not impinge on the solvus and carbonatite could be produced by fractional crystallization. Syenites (field IV) also plot away from the proposed solvus whilst nepheline syenites plot on it, and this relation corresponds to the known absence and presence respectively of carbonatites with these rocks in the field. A third feature is that melilitites (field V) and kimberlites (field VI) lie well away from the proposed solvus, in keeping with the known geology of melilitite and kimberlite occurrences. Likewise basalts, andesites and rhyolites (fields VII and VIII) cannot produce carbonatites by liquid immiscibility. Fields IIa (alnoite) and IIb (urtite) are interesting exceptions. Alnoite is a frequent accessory magma in carbonatitic complexes and is itself commonly strongly carbonated. It is suggested that alnoite could be the fractionated product of olivine-melilite nephelinite which had penetrated high into the crust and from which the late-stage carbonate-rich residue had not separated. Urtite (field IIb) is rare even when compared with the ijolite in which it occurs, usually as minor intrusions at relatively high levels in the ijolitic plutons. It is proposed that urtite can result from liquid immiscibility taking place in an ijolite pluton at mid-crustal levels. The low Ca content of urtite compared with ijolite (Fig. 9) is consistent with loss of a C a - N a rich carbonate fraction by liquid immiscibility. This is further corroborated by the composition of the carbonate fluid inclusions trapped in urtite which are relatively Ca-rich compared with those trapped in ijolite, as shown in Fig. 10. Many of these possibilities must remain speculative until the important constituents Fe, Ti, P, F and oxygen fugacity are incorporated into the system.
69
Nephelinites and carbonatites
TABLE 6. Analyses of typical carbonatites (wt.~)
Carbonatites Despite the fact that nearly all the carbonatites of the world are calcite or dolomite-rich, the primary and parental carbonatitic magmas produced by liquid immiscibility are considered to be strongly alkaline, akin to natrocarbonatite. Twyman & Gittins (1987), however, do not consider natrocarbonatite to be either primary or produced by liquid immiscibility. The close agreement between the experimental work described above leaves no doubt that liquid immiscibility must be significant and that the primary composition is rich in Na + K. The fact that fluid inclusions trapped in early apatites in carbonatites are alkali-rich (Rankin 1975, 1977) lends further support. Most carbonatitic complexes display a wide variety of carbonatites. Field work invariably shows that the earliest intrusions are s6vites, with later alvikites and then ferrocarbonatites. N a - K fenitization is usually confined to the s6vite. Alvikites rarely show fenitization and ferrocarbonatites never show it. S6vites occasionally form diatremes 1-3 km in diameter. More usually they are tens or hundreds of metres across, while alvikites and ferrocarbonatites occur in swarms of dykes and veins.
Natrocarbonatite
1
SiO2 TiO2
AI203 Fe203 MnO MgO CaO Na20 K20 P205 CO2 F C1 SO3 SrO BaO REE
2
3
4
0.05 0.88 0 . 1 6 6 . 1 2 0.01 0.18 0 . 0 7 0 . 6 8 0.11 0.37 0 . 1 7 1.31 0.41 2.62 4.04 7.55 0.48 0.39 0.41 0 . 7 5 0.48 0.31 0.67 12.75 14.43 53.60 51.20 29.03 33.89 0.09 0 . 2 5 0.14 8.39 0.03 0.01 0.79 0.93 3.18 1.52 2 . 6 6 30.53 38.38 39.50 37.03 0.09 2.71 0.06 - . . 3.81 Trace . -0.89 2.88 - 1.35 0.23 0 . 1 0 0.01 0.08 0 . 1 7 0.11 1.26 0.05 0.3 ND c.0.1
5
6
3.24 0.83 0 . 0 0 0.07 0.20 0.65 11.50 11.00 5.18 5.53 10.74 0.36 25.85 43.60 -0.05 -0.06 1.27 0.42 32.62 30.42 --. 0.49 - 0 . 7 3 0.07 2.48 >4.0 2.82 c. 1.5
All analyses sum to 100 __+1.7 (LOI included). 1, Natrocarbonatite lava, 1960 eruption, Oldoinyo Lengai, N Tanzania (Gittins & McKie 1980); 2, s6vite dyke, Tundulu, Malawi (Garson 1965, p. 24); 3, alvikite cone sheet, Homa Mountain, W Kenya (Le Bas 1977, p. 317 (HC629) traces by C. Barber, unpublished); 4, dolomite carbonatite (beforsite dyke), Aln6, Sweden (von Eckermann 1948, p. 122); 5, ferrocarbonatite (high Mg), Kangankunde, Malawi (Garson 1965, p. 54); 6, ferrocarbonatite (low Mg), Homa Mountain, W Kenya (Le Bas 1977, p. 318 (HF15) traces by C. Barber, unpublished).
magma
The chemical composition of parental natrocarbonatite is still uncertain. Table 6, column 1, gives the composition of natrocarbonatite extruded from Oldoinyo Lengai in 1960. This is the only known instance of fresh carbonatite lava, and its composition (CC, 30 wt.~; NC, 58 wt.%; KC, 12 wt.~o) when plotted on Fig. 9 (field IX) indicates that it is the conjugate melt with phonolite magma, as interpreted by Freestone & Hamilton (1980). The conjugate carbonate melt with nephelinite would, according to them, be more calcic, but no natural examples are known except perhaps at Kaiserstuhl (Keller 1981). O and C isotopes confirm the igneous nature of natrocarbonatite and indicate a temperature of crystallization in the range 400-800~ (Suwa et aL 1975). Natrocarbonatite has been described in detail by Dawson (1962), Du Bois et al. (1963), Cooper et al. (1975), McKie & Frankis (1977) and Gittins & McKie (1980). It is porphyritic with large clear platy crystals of nyerereite ((Na,K)2Ca(CO3) 2 solid solution) and large brownish rounded crystals of gregoryite (Na2CO3 with some Ca + K substitution for Na) in a glassy carbonate matrix containing quench needles of nyerereite.
Table 7 gives a modal analysis of a sample collected by N. J. Guest from the 1960 eruption. Older examples of carbonatitic lavas or their pyroclastic equivalents have recently been discovered. Deans & Roberts (1984) described carbonatite tufts and lava clasts from the Miocene volcano at Tinderet in Kenya in which flowaligned calcified nyerereite crystals were identified. At one time they were thought to be pseudomorphs after melilite, and although relict nyerereite has never been found (nor will it be since it is water soluble) further textural study leaves little doubt that the platy cavernous pseudomorphs, largely composed of myriads of small calcite crystals, are after nyerereite. Such pseudomorphs are now also known in lavas at Kerimasi in N Tanzania, and suspected at Homa Mountain and the Ruri Hills, W Kenya, at Kruidfontein, S Africa, and in the Cape Verde Islands. The platy pseudomorphs at Kerimasi volcano described by Mariano & Roeder (1983) were identified on textural features by Hay (1983) as former nyerereite crystals. They coexist with fresh tablet-shaped calcite phenocrysts. Although both pseudomorphs and phenocrysts are com-
7o
M. J. Le Bas
TABLE 7. Modes ofnatrocarbonatites (volume ~ ) Oldoinyo Kerimasi, Lengai, Tanzania Tanzania 82-8-15A2
Kerimasi, Tanzania 82-8-14C6
Kerimasi, Tanzania 83-7-26A
Tinderet, Kenya L. KIP 1
Tinderet, Kenya U. KIP 3
Kaiserstuhl, S Germany, and Fort Portal, W Uganda
A few Abundant
9.3 2.7
Very few Abundant
Abundant None
Phenocrysts Calcite Nyerereite Gregoryite
Nil 38.1 48.0
29.7 67.0 1
7.6 46.0 __
Groundmass Calcite Nyerereite Opaques Apatite Perovskite Pyrochlore
-6.02 1.1 ----
2.8 -0.3 0.2 Trace --
46.33 Trace Trace ---
Interstitial glassa
6.8
--
--
c.50 m
7.1 79.3 0.7 0.4
m
0.5 m
1, not identified in thin section; 2, may include groundmass gregoryite; 3, Cc" N y ~ 1 1 4, includes quench nyerereite. posed of calcite, electron diffraction microprobe analysis (Table 8) distinguishes the typical igneous calcite with high Sr from that in the pseudomorphs which have low Sr but detectable Mg and Na. This distinction, the interpretation that the pseudomorphs cannot be after calcite, the relations known from the C a - N a - K carbonate system (Fig. 11) and the probable liquid immiscibility relations shown in Fig. 9 all confirm the identification. The modal variation of natrocarbonatites (Table 7) shows that these lavas can have a wide range of compositions, assuming that the phenocryst phases are complete and cognate. By using the phase data of the synthetic C a - N a - K carbonate system at I kb (Fig. 11) determined by Cooper et al. (1975), it can be seen that the natrocarbonatite of Oldoinyo Lengai plots in field A on Fig. 11, but the Kerimasi and Tinderet
lavas are more calcic, since both nyerereite and calcite are precipitated, and plot in the field B. Clasts from the Cape Verde Islands also appear to plot in field B. The occurrences of carbonatite lava at Kaiserstuhl in S Germany (Keller 1981) and at Fort Portal in W Uganda (in samples kindly provided by P. H. Nixon) show phenocrysts of calcite alone. This cannot easily be reconciled with the two-liquid field system shown in Fig. 9 and the liquidus surfaces of Fig. 11 unless the lavas were relatively N a poor. It is concluded that a wide variety of natrocarbonatite magmas can exist. Some occur with gregoryite and nyerereite phenocrysts, indicating compositions appropriate to immiscible liquids conjugate with phonolite magma, as at Oldoinyo Lengai. Some have nyerereite phenocrysts alone or with calcite (Table 7), appropriate to immiscible liquids conjugate with nephelinite magma.
TABLE 8. Microprobe analytical data on Kerismasi natrocarbonatites Brown pseudomorphs
Clear calcite crystals 1
MgO CaO S~ Na20
BD 55.1• 0.7• BD
2
BD 54.8• 0.7• BD
3
BD 51.5• 0.6• BD
4
0.8• 48.3• BD 0.5•
5
0.4• 50.6• BD BD
1 and 2, Large phenocrysts in lava (samples 82-8-14C6 and 83-7-26A respectively); 3, groundmass quench prismatic crystal (sample 83-7-26A); 4 and 5, two large pseudomorphs replaced partly by granules of calcite and partly by voids (with 1 and 2 respectively). Each analysis is the mean of four points with standard deviation to 2tr. The energy dispersive microprobe analyses were performed using the Cambridge Microscan Mk.5 at the University of Leicester. BD = below detection limit.
Nephelinites and carbonatites
/
/
NazCOs
/ K2Ca(CO3)2
Gry/(~A"XNyr/~ \\ / ~ ) ~ ~/ .) CaCOs No2Co(C03)2 Na2C%(C05)3
FIG. 11. CC-NC-KC carbonatite system showing field of natrocarbonatites from Oldoinyo Lengai (A) and from Kerimasi and Tinderet (B). Gry and Nyr are plotted at the natural compositions of gregoryite and nyerereite minerals. NY is the end-member nyerereite Na2Ca(CO3)2. FC is fairchildite K2Ca(CO3)2. (After Cooper et al. 1975).
Natrocarbonatites contain few minerals apart from carbonates. The few include oxide and phosphate minerals, but no silicates apart from rare biotite. Deans & Roberts (1984) report 15 wt.~o SiO2 and up to 5~o modal silicates, often biotite, in the Tinderet lavas. Bedson (1983) too usually found 1-5 wt.~ SiO z in his experimental carbonate charges. A120 3 was also present but was even more weakly partitioned into the carbonate melt. Biotite is the most frequent silicate mineral encountered, but only in accessory proportions. In some s6vitic carbonatites, mica and other silicates, particularly mildly sodic augite sometimes rimmed by aegirine, become abundant, and these rocks are the metacarbonatites. Vartiainen (1980) described a collar 2 km wide of metacarbonatite (average SiO2 content, 25 wt.~) around the s6vite core 2 km in diameter at Sokli. Armbrustmacher (1979) described petrographically similar rocks in Colorado, and Robbins & Tysseland (1983) have described silicocarbonatites (metacarbonatites) derived from gabbros at Pollen in N Norway. Metacarbonatites are interpreted to be 'reaction rocks' between magmatic carbonatite and country rock; the resulting 'mixed rock' is often up to 70~o calcite. Similar reactions, but on a smaller scale, are known in nearly all s6vites, and the products are often incorporated as xenoliths in s6vite intrusions. The experimentally determined distribution coefficients of elements between immiscible car-
71
bonate and silicate melts (Freestone & Hamilton 1980; Bedson 1983) help explain some of the observed variability. P partitions strongly into the carbonate melt, and hence nephelinites are often poor in apatite while the early s6vites are rich in apatite. Carbonatite lavas, however, always have less apatite than predicted by the above. Apatite is a dense mineral (specific gravity, 3.1) whereas carbonatite magma has a very low density of 2.2 g cm- 3 (Nesbitt & Kelly 1977), and therefore it is not surprising that apatite attains eruption less commonly. The partition of Mg between silicate and carbonate melts depends mainly on Pco2. At 7.6 kb Mg appears to partition preferentially into the carbonate (the data are poor), but at low pressure (1-3 kb) it partitions into the silicate. The distribution coefficient of K is never far from unity, unlike Na and Ca which both partition strongly into the carbonate melt. This causes a fall in the Na/K ratio of the silicate magma after the separation of carbonatite. Sr also partitions preferentially into the carbonate melt, particularly in the lower-temperature range 600-700~ but the Ca/Sr ratio remains unaltered in all melts (Koster van Groos 1975). Greater pressures increase the partition of REE into carbonate melt, with more light REE than heavy REE. This gives the steep REE pattern typical of carbonatites, but the steepness can vary considerably depending on the temperature at the time of separation (Bedson 1984). Intrusive carbonatites
The intrusion of carbonatite in pipe-like structures usually follows that of the silicate rocks (Fig. 12). The commonest carbonatite is s6vite showing adcumulate texture, and it is normally composed of 90~-100}/o calcite (Table 6). Dolomite may occur instead, particularly in older and more deeply eroded complexes, in keeping with the fact that dolomite is not stable at depths less than about 2 km at carbonatite magmatic temperatures (600-800~ Periclase occurs in the Amba Dongar and Kerimasi shallow-seated carbonatites. Common accessory minerals in carbonatites are fluorapatite, magnetite, biotite, pyrochlore and olivine, and some less-common minerals are aegirine-augite, arfvedsonite, K-feldspar, pyrite, zircon, baddeleyite, perovskite, sphene, quartz, fluorite, baryte, bastnaesite, parisite and monazite. The last six minerals occur mainly in the late-stage carbonatites, the ferrocarbonatites, while the remainder can usually be found in the earlier and more abundant carbonatites. There is a noticeable lack of sodic minerals
M. J. Le Bas
72 0
~,
.
4 0 0 0 t_ |
L
3 0 0 0 [I
NAPAK
,
5 [
-...--~--J~--~
+ + + +'+
metres above sea-level
/ ~ o~\~
~\c 9~J . ~'G'
, ,~o\~~ ~ \ \ ~
.~,~'~ /I --oba b~e-~ ~ '
2ooo t - _ ~ 1000~+
, l .Km
+ -r+
\
~o.~;' . . . . . . . . . . 9
Vole~
I "] -
. . . . . . .
~
UI~ A / t-IIM
~..
LOKUPOI
~eg~,,z,~" ~~ .,e o4 s u b -
-~ .
.-.n,~ ~o~
"oo @
_I
~"
J ""
~
"--.. " - ~
J
- q
+ + + + + /
Fenitized
~.
1/
basement
/ Ijolite
~ Carbonatite
~, Basement
FIG. 12. Cross-section of Napak nephelinite-carbonatite volcanic complex in E Uganda. The structure is typical of such complexes. (After King 1949.) and few potassic ones, considering the abundance of Na and K in parental natrocarbonatite. The Na and K are lost by metasomatic reaction of the early carbonatites with country rock to form fenites. Therefore the carbonatites seen cropping out are not products of the complete magma less any volatiles but have also lost their original complement of alkalis (Table 6). In the case of adcumulate s6vites the loss could have accompanied the escape of the intercumulus material. The loss of alkalis unfortunately masks the original alkali content of the parental natrocarbonatite, and therefore it is not possible to decide whether the carbonate immiscibly separated from a nephelinite or a phonolite melt. However, in some complexes (e.g. Ruri Hills, W Kenya) there is a close temporal association of phonolite and carbonatite intrusions, which could thus be pairs of conjugate immiscible liquids. The immiscible separation of carbonate melt from silicate melt takes place at l l00-1000~ This is well above the liquidus temperature for the carbonate component of the melt, but not above that for olivine and apatite. Both olivine and prismatic apatite therefore precipitate from carbonatite magma and under these conditions form cumulates at the base of the intrusions such as those occurring in Kola. Not only are olivine and apatite much denser (specific gravity of 3.3 and 3.1 respectively) than carbonatite magma (specific gravity, 2.2), but carbonatite magma has an unusually low viscosity of 5x 10-2 P (Treiman & Schedl 1983) permitting efficient gravity settling. If oxidizing conditions prevail, magnetite may also precipitate and accumulate. The carbonate liquidus is encountered at 500600~ This temperature is gauged from homogenization studies on fluid inclusions trapped in apatites (Rankin 1975, 1977). These apatites are not the prismatic crystals which accumulated from the early phase of cooling but are the large ovoid apatite crystals which characterize s6vite. This temperature is in general agreement with the liquidus temperatures determined by Wyllie and others for carbonatite magma. The low
temperatures are also marked by fenitization of the thermal aureole around the carbonatite, in which low albite, orthoclase, microcline, phlogopite and arfvedsonite commonly occur. With cooling of a carbonatite intrusion, fractionation of s6vite to alvikite to ferrocarbonatite takes place, with a final small-scale development of calcite carbonatite which lithologically looks like s6vite but is devoid of the usual accessory mineralogy characteristic of normal s6vite. This sequence is observed in E Africa (Le Bas 1977), in the U.S.S.R (Kapustin 1981), the Chilwa Province of Malawi (Garson 1965) and elsewhere. A variant on this occurs when the carbonatite is dolomitic. In this case the sequence is s6vite to dolomite carbonatite to ferrocarbonatite. These sequences accord with Wyllie's (1965) experiments. Olivine (Foss_90), rarely monticellite and very rarely melilite are restricted to the earlier and deeper calcite carbonatites (s6vites). Apatite and pyrochlore can crystallize throughout the sequence of carbonatites but are most abundant in the early s6vites, with apatite sometimes being so abundant that it forms apatite rock. Such apatite deposits have been mined. Apatite rock also occurs as disrupted schlieren in s6vite and dolomite carbonatite where they are dragged and folded by the magmatic flow. These apatites are almost completely deficient in Sr and REE and can be matched with microfragments of similar textured apatite rock caught up in the nephelinitic lavas. The main apatite in the s6vites occurs as large slightly resorbed crystals which have much the same contents of Sr and La as the apatite phenocrysts in nephelinitic lavas, as is appropriate to the postulated liquid-immiscible relation. Lapin & Vartiainen (1983) similarly describe olivine, apatite and magnetite which have fractionated from carbonatite. Biotite can occur in all intrusive carbonatites, as can pyroxenes and amphiboles, but they are more frequent in the early s6vites and dolomitic carbonatiteso Biotite flakes, often with margins oxidized to opaque oxides, together with magne-
Nephelinites and carbonatites tite phenocrysts can also occur in carbonatite lavas such as those near Fort Portal in SW Uganda. Feldspar, usually potassic but sometimes sodic, occurs in the margins of s6vitic and dolomitic carbonatites and is derived from the enveloping feldspathic fenite. Many but not all s6vites carry 1%-2~ of euhedral magnetite microphenocrysts which often coexist with biotite, apatite and pyrochlore. The magnetite is Ti poor with usually only 1-2 wt.% TiO2, and at Jacupiranga, Brazil, the magnetites in the s6vites are zoned with decreasing contents of Ti towards the rims. MnO and MgO, both of which usually show contents of less than 1 wt.~ and less than 4 wt.~ respectively, also decrease with the zoning. Magnetites in the later intrusive phases of carbonatite have compositions comparable with these outer zones and are also pure magnetite (Prins 1972; Gaspar & Wyllie 1983a). Ilmenite is rare in carbonatites and has very variable Mg, Mn and Nb contents. At Jacupiranga, Brazil, where it coexists with magnetite in the reaction zone betweenjacupirangite (titaniferous pyroxenite) and later s6vite, Gaspar & Wyllie (1983b) have calculated an equilibration temperature of 570-595~ with oxygen fugacities of 10- lS'5-10-19.5 atm, which are similar to the values estimated by Prins (1972) for magnetites from a variety of African carbonatites (540-575~ 10- 23-10- 24 atm). Alvikites are medium grained compared with the coarse-grained s6vites and are less voluminous, although in some complexes they appear to be very abundant because they form numerous but thin minor dykes and cone-sheets which comprise the majority of the outcrop. The mineralogy of alvikites is similar to that of s6vite, with a tendency for alvikite to have fewer accessory oxide and silicate minerals. The principal difference lies in the calcite which, in alvikite, was slightly ferroan when it crystallized. Probe analysis of the calcite often fails to reveal this ferroan character, evidently because the iron has unmixed from the calcite structure as a result of late-stage processes to which all carbonatites are prone. In such cases the iron is visible only as minute black specks scattered throughout the carbonatite quite separate from any magnetite microphenocrysts, and as brown staining along the crystal boundaries of the calcites. Alvikites are also marked by slightly greater contents of Mn, Ba and REE (Table 6). These elements appear to be in solid solution in the carbonate minerals. The apatites in alvikites are also rich in REE as well as Sr. True alvikites usually have a pale brownish colour, reflecting the Mn and Fe contents. Some, however, are white, like s6vite, and these commonly do not
73
contain as much Mn, Fe, Ba and REE as true alvikites and are in fact micros~vites. The distinction in the field is not easy as the brown coloration is often obscured by secondary processes, but trace-element analysis of the rock or probe analysis of the apatite can reveal the distinction. Much of the economic wealth of carbonatites lies in the ferrocarbonatites. The primary carbonate mineral of ferrocarbonatite is ankeritic, but with weathering much of it breaks down to secondary calcite or dolomite rimmed by iron oxides. Ferrocarbonatite occurs in only minor quantities in most complexes, and appears as thin brown dykes and veins with chilled margins and cross-cutting relations against earlier carbonatites. There are many good examples in E Africa, Malawi, Aln6 and the Cape Verde Islands. In other cases ferrocarbonatites form diffuse veins and patches apparently replacing s6vites, and are considered to be a late-stage development of the magmatic ferrocarbonatite. This residue of carbonatite magma appears to have become so rich in volatiles that it no longer behaved as a melt but as a volatile-rich fluid. Brown ferrocarbonatitic material can be seen progressively penetrating coarse calcite and dolomite carbonatites, migrating along crystal boundaries and gradually replacing the calcite and dolomite. The replacement process can go to completion, and if it begins along a fracture, as it usually does, the end-product can be a parallel-sided zone a metre or so wide which can look rather like a crosscutting dyke. This process can be seen replacing s/Svite in Kenya, Malawi and Fen. At Newania, India, and Sarfartoq, W Greenland, it replaces dolomite carbonatite. The apatites in the ferrocarbonatites also provide evidence for the replacement process. The cryptic variation shown by apatite is usually recorded by increases in REE, which reach a maximum in the ferrocarbonatites, but often the apatite in ferrocarbonatite lies compositionally in the alvikite and sometimes s6vite fields. This is consistent with the interpretation that many ferrocarbonatites were formerly alvikites subsequently transformed and mineralized. Mineralization
Four stages of mineralization can be recognized, all following on the magmatic Fe-enrichment (and Mn-enrichment) stage of ferrocarbonatite development. The first stage of mineralization is the production of REE-rich carbonatites from residual Frich magmas (Moiler et al. 1980; Samoylov & Smirnova 1980; Viladkar & Dulski 1986). The
74
M. J. Le Bas
REE minerals segregate in pockets and fractures along sheet-like ferrocarbonatite bodies. The minerals, principally bastnaesite ((Ce,La)FCO3), appear as pale-brown streaks within the darkerbrown to black ferrocarbonatite, and show strong light REE enrichment with La/Yb ratios of 1001000. U - T h radioactivity and associated fluorite or baryte are usually indications of the presence of REE minerals. Nearly all carbonatite complexes show this mineralization to some extent. The largest is at Mountain Pass, California, where Sr-rich carbonatites have more than 10% bastnaesite and other REE fluorcarbonate minerals cut by barytebearing and silicified radioactive veins (Olsen et al. 1954). Experiments by Jones & Wyllie (1983) suggest that the bastnaesite could have been precipitated from a low-temperature carbonatite magma (perhaps about 500~ or less). The Kangankunde carbonatite in Malawi has a similar REE mineralization but with Sr > Ba. Following the formation of the replacement ferrocarbonatites is the second-stage mineralization associated with late-stage fluids and the precipitation of fluorite at lower temperatures (100-200~ (Roedder 1973). Massive fluorite usually forms along fractures, particularly in the roof zones of carbonatitic complexes. It is variously purple, yellow or colourless, and can also occur as scattered crystals in the early carbonatites adjacent to the fractures along which the fluids passed. Evidently, F-rich fluids penetrated into the wall-rocks. At Kruidfontein near Johannesburg and at Amba Dongar, India, the F-rich fluids rose to the top of the carbonatite complex and entered and mineralized the fractured rocks capping the complex. At Kruidfontein the fluids penetrated up into bedded carbonatitic tufts sitting in a volcanic caldera structure and formed fluorite-rich horizons interbedded with the tufts. The third stage is the formation of baryte. It is present in nearly all carbonatite complexes and follows closely on the formation of fluorite with which it commonly occurs but in much smaller quantities. Local concentrations do occur and have been mined (e.g. at Aln6, Sweden). The Ba, F and REE are all mantle derived. They are not products of assimilation of continental crust, as it is sometimes considered for the Curich mineralization at the Palaborwa carbonatite in S Africa, because exactly similar Ba, F and REE mineralization is known in the Cape Verde Islands which are seated directly on oceanic crust (Le Bas 1984). The fourth stage of mineralization is marked by the formation of U - T h minerals together with pervasive silicification. This mineralization is
never very extensive and is characterized by low U/Th ratios which makes some deposits uneconomic (e.g. Loe-Shilman, N W Pakistan (Jan et al. 1981)). This is often followed by late-stage calcite veining and reprecipitation of Fe oxides, usually hematite. The Fe is derived from ferrocarbonatites penetrated by hydrothermal fluids. If the fluids become locally all-pervasive, assisted by incursions of groundwaters, then large hematite deposits can form. The Rodberg iron deposit at Fen, S Norway, is an example of this process (Andersen 1984). These Fe deposits are distinct from the magnetite deposits such as that at Kovdor, Karelia, U.S.S.R., which is a product of primary magmatic precipitation.
Fenitization The products of fenitization are mineralogically syenitic, with the result that they are sometimes misidentified as igneous syenites. The chief points of distinction lie in the texture and structure. Syenitic fenites are often heterogeneous and show low-temperature deformation or shattering. With increasing metamorphic temperatures, schlieren and mechanical flow textures develop, and next to the igneous contact the fenites become coarse grained. The effects of fenitization are more extreme than those of ordinary contact metamorphism because the metasomatism causes strong chemical reactions which can totally obscure the original composition of the rock. As a result there are many theories of fenitic processes (King & Sutherland 1960; Woolley 1982). The key lies in the realization that fenitization is not a single process but is often multiple. The variables involved are (a) the geochemistry of the magma causing the metasomatism, (b) the nature of the fluids leaving the magma, (c) the depth profile of the magma (different reactions develop at different depths), (d) the porosity and structure of the country rocks being fenitized, (e) the mineralogy and geochemistry of the rocks being fenitized and (0 any water-rock reaction.
Syenitic fenites generated by ijolite Fenites are not seen around nephelinitic volcanic pipes and dykes, nor around xenoliths of country rock trapped in nephelinitic magma. Evidently time is insufficient for identifiable metasomatism to take place. However, fenites develop strongly around larger intrusive bodies of nephelinite magma that fractionate and crystallize to ijolite. Fenitic zones can often be identified around ijolite plutons, and Fig. 13 gives the typical
Nephelinites and carbonatites ORIGINAL MINE RA~.,~400
300
METRES 200
I
!
75 0
100 |
CONTACT
GRANODIORITE ZONE IT
ZONE I
=
ZONE 111"
ZONE
I
Quartz ~
Ae~i,rine- ou~lite Mognetite ]lmenite Oligoclose
'
Atbite (An3-s) i
Microcline
Orthoctose
_.
Biotite ,,, ~ Alk(:Ui FeldSpQr
Mesoperthlte & . . . . . o~tipe'rthite iNephetine . . . . .
.........
,
!
~
Pate brown biotite ,,I
; .
.
.
Phlogopite . . . .
.
W
I
I
i_j
tO Hornblende
~, s.0dic a..~hibote (,~Y,)
Sodic ;Qmphibote ' WoliQstonite ~"
~ SHATTERIN~L. w-
STRONG MORTAR
FENIT/ZEO
GRANULATION TEXTURES
GRANOOtoRITE
COMPLETE
& "-
GRANULATION
$YENtTh
FIG. 13. The mineralogical changes across the fenite zones surrounding ijolite in granodiorite. (After Rubie, in Le Bas 1977, p. 72.)
mineralogical changes that take place across the zones. The pyroxene grades from a weakly sodic aegirine-augite near the contact to almost pure aegirine in zone 1, and can form 5%-50~ of the rock. The feldspar near the contact is new hypersolvus alkali feldspar commonly in the range Orgo_6oAb60_40, but in the outer zones the original sub-solvus feldspars of the granitic basement survive. The hypersolvus character of the feldspars and sub-solidus texture of the fenites suggest fenitization temperatures of about 800~ Ijolite intrusion causes uplift of the country rocks. At one time it was thought that the uplift was related to the metasomatic transfer of material from magma to fenite, but Rubie (1982) has shown that in W Kenya the transfer of material, mainly SiOz, was from country rock to magma, and that granodiorite suffered a volume loss of up to 20~ near the contact. This results in the formation of feldspathized ijolite (petrographically equivalent to nepheline syenite) at the margin of ijolite intrusions (Fig. 14). Aluminium appears to have been almost immobile during fenitization, and the K / N a ratio
decreases slightly away from the contact with falling temperature (Rubie & Gunter 1983). Other studies have been based on the assumption that oxygen is immobile, a condition which approximates to a constant-volume fenitization (McKie 1966), but Appleyard & Woolley (1979) based their study of fenitization on a constant-specificgravity model which suggested, at the Borrolan syenite in N W Scotland and the Sokli carbonatite in N Finland, that modest volume increases had taken place. Carbonate 9minerals are rare in fenites around ijolites unless later carbonatite intrusions are present. Evidently the metasomatic solutions causing the fenitization were not carbonate bearing.
Syenitic fenites generated by carbonatite N a - K fenites around carbonatites are much more varied than those around ijolites. Only s6vites, dolomitic carbonatites and rarely micros6vites show fenitization. Alvikites and ferrocarbonatites almost never show fenitization, and this is
76
M. J. Le Bas +
+ +
+
BASEMENT IttLIIII IIIIIIIII1,, Nepheline-poor Nep~;~i;;-po, o:
_jr_
.....
+
--- =
,,
11111IT,iF-_--
x
lllrlll
-
--
•
IIII rr ;i-~-
Nepheline-ll
,,
x
•
'I[II[EIIJV-s'
+
" "
x "
inward d ffus on of Si during fenitization
Ii'/
~X_--~lllrelative
-.~
from less ldesilicified
~- '~1, basement
X
~/N,,~
/,, I
;^~J I
]l~lll(
~
~o.-o,,,o,,e
~'1
x
+
~
+
screen of chilled m c r o - j o te prevents diffusion of Si
X
-
-Jr-
Ix|/i!/
x
aegirine-rich
to
--~-_(!1 'jmore d e s i l i c i f i e d
rich fenite /
+ Jlil~
downward
-', =lllt~'"""~176o, ~,
lJ
I11111
from b a s e m e n t
FIG. 14. Schematic cross-section of ijolite intrusion with chilled margin and influx of silica during fenitization, and the development of feldspathized ijolite (mineralogicallyequivalent to nepheline syenite but texturally distinct) particularly along the upper parts of the margins. taken to indicate that early carbonatite magmas are alkali-rich but that these alkalis are lost during the early stages of carbonatite fractiona1 ~
0 + K20
/ -.6"D~ /
i' ;,,!gO + Fe20T
~k
~
A A
A
9
x X CaO
wt% FIG. 15. Carbonatite plot showing the path of
fractionation of carbonatite magma from natrocarbonatite (#) to s0vite (O) and alvikites ( x ) to beforsites and ferrocarbonatites (A): P, Fort Portal lava; T, estimated Tinderet lava; K, estimated Kerimasi lava (Table 7). (After Le Bas 1981.)
tion (Fig. 15). In some cases even the final phases of s0vite intrusion fail to cause fenitization. Evidently by that stage the s6vite magma has lost its alkalis, and instead could only brecciate the fenites formed earlier. This is contrary to the proposal that the concentration of alkalis increases during carbonatite fractionation (Twyman & Gittins 1987). Depth profiles are important. Fenitization around the topmost parts of a s6vite is strongly potassic, usually with the production of almost pure K-feldspar rock. Early records of this came from Malawi (Dixey et al. 1935), Songwe, Tanzania (Brown 1964), Kaiserstuhl, S Germany (Sutherland 1967) and Zambia (Bailey 1966). At Koga in Swat, N Pakistan, K-feldspar rock (plus rare aegirine) about 50 m thick completely caps a s0vite plug 500m across (Mian, personal communication). The K-feldspar is usually orthoclase Or87_9s, whereupon the rocks are termed orthoclasites (Sutherland 1965). Sometimes the K-rich feldspar is microcline, and Heinrich & Moore (1970) named such cases 'burnt rock' as they had that appearance. Commonly, the only other phase present is 1%-2~ of limonitic iron oxides, with calcite absent. The high K content is character-
Nephelinites and carbonatites istic of these fenites, whereas the K-feldspar developed in feldspathized ijolites and ijolitic fenites is less potassic (Or90) or almost pure orthoclase (Or>85), whereas the feldspar in the syenitic fenites around ijolite and syenite is an alkali feldspar Ab60_40 Or4o_60. The compositions reflect the low temperature of carbonatite fenitization (below 550~ in contrast with the higher temperatures of the silicate fenitization (more than 700~ Albitite also occurs as fragments brought up by carbonatitic volcanic breccias, e.g. Amba Dongar, India (Deans et al. 1972), and NyamajiWasaki, W Kenya (Le Bas 1977), indicating albite-rich fenites at depth whilst K-rich fenites are exposed at the surface. A similar zonation occurs around the Novopoltavskoye carbonatite, Azov, where inner zone fenites are albitic but the outer ones are microcline-rich (Kapustin 1982). The N a - K zonation is in accord with Orville's (1963) experiments on alkali-ion exchange between fluid and feldspar at falling low tempera-
77
tures, and finds support from Woolley's worldwide investigations (1982). S6vite and alvikite carry very little F, but the natrocarbonatitic magmas from which they were derived were probably F rich and C1 rich. Natrocarbonatite lava has 1.49-2.93 wt.% F and 0.77-4.15 wt.% C1 (Gittins & McKie 1980). Some minerals in carbonatites also show that they crystallized from halogen-rich magma, and Dawson & Fuge (1980) estimate that F > C1 was usual. In addition to apatite, biotite and Na-amphibole, which are the most important F-bearing minerals, clinohumite (Mg(F,OH)2.4MgzSiO4) has been reported at Cargill in Ontario, Jacupiranga in Brazil, Palaborwa in S Africa, Sokli in Finland and Gardiner in E Greenland (Gittins 1978). Even the residual nephelinitic glasses from Oldoinyo Lengai volcano have appreciable halogen contents (0.25-0.69 wt.% C1), and the F/C1 ratio for the whole rock is 1.6 (Donaldson & Dawson 1978). The F and C1 are evidently mantle derived, probably from phlogopite rather than apatite (Edgar & Arima 1985). However, at Oka, Quebec, Treiman & Essene (1984) calculate that the fluorine fugacity was low (10-44). If carbonatite magma were rich in halogens, particularly F, then all fenitic fluids emanating from it would also be F rich as would the residual fluids. Evidence for the former is seen in the fluorapatites and fluoramphiboles developed in fenites, and for the latter in the fluorite late-stage mineralization. Yardley (1985) also finds evidence from apatites that F might be important in metamorphic fluids. It is considered probable that a fluoride species is the main solvent of the fenitizing fluids.
Conclusions In this assessment of the main features of carbonatite-nephelinite volcanism, the following salient points can be made. The starting point is carbonated nephelinitic magma generated from an enriched mantle source such as that described by Cohen et al. (1984). Such magma rises through the lithosphere, zone refining as it rises, until it stops in the lower crust. Fractionation at that stage would permit olivine and pyroxene cumulates to form. At the same time the melt would fractionate to phonolite, and liquid immiscibility could take place forming a series of conjugate carbonatite magmas (Fig. 16). Nephelinitic magma is dense (about 2.7 g cm-3) and hence is not seen where the crust is composed of thick low-density unaltered sediments. Where nephelinitic magma has the buoyancy to penetrate the crust it can make
78
M. J. Le Bas
MELANEPHELINITE NEPHELINITEPHONOLITE LAVAS&PYROCLASTICS
PSEUDO-
DYKES
Rap,d
t /
IJOLITE
,
Ropid ascent
" "
I --.~
K.FEN!TE~A~ vK~TEE/
FenilizatiOn
IPER,DOTITE
/ /
FERROC,~BONATIT~
/I.SYE
I Slow / aster, I
I
LATE- STAGE CARB~NATITE
/
I:k/1 t oo.o
oscenl
o, .....
"~ "
TRACHYTE
URTITE & WOLL-URTITE LATE- STAGE ~11 ALKALINE FLUIDS/' CALCITE 1~ / IJOLITE -
SYE.,TES \
CARBONATED BRECCIASTUFFS & AGGLOMERATES
CARBONAIIELAVA & PYROCLASTC IS
.a
PQchonohon Low volahle ~ .
silicote melt
Volobie-rJch fluid
I lower crustal magma chamber
HELINITE I LIQUID FRACTIONATES I TO PHONOLITE I IMMISCIBILITY I
CUMULATES
? BASE OF CRUST
CARBONATED NEPHELINITIC MAGMA
PARTIAL MELT OF ENRICHED MANTLE SOURCE
FIG. 16. Schematic flow chart for the evolution of nephelinitic and carbonatitic magmas. (After Le Bas 1977, Fig. 24.2.) pyroxenite-ijolite-urtite plutons, with the ijolite perhaps developing an outer zone and capping of feldspathic ijolite (Fig. 14) as a result of the inward diffusion of silica from the country rocks undergoing fenitization. The country rocks are metasomatized to syenitic fenites with alkali feldspar and Na-pyriboles (Fig. 17). Although natrocarbonatite magmas of various (Na + K)/Ca ratios can be produced by the liquid
immiscibility, depending on whether the conjugate silicate melt is nephelinite or phonolite, the alkali loss to fenitization obscures the differences. All carbonatites follow similar paths producing s6vites, dolomite carbonatites, alvikites and ferrocarbonatites with REE enrichment, with the final production of residual mineralizing fluids which precipitate fluorite, baryte and U - T h minerals. F-rich alkali-bearing fluids cause feni-
Nephelinites and carbonatites
79
J
Cone-sheets~ plones of brecciation end K |enitizetion
'.?3
:"
+ +
/ ~_
.
I/0tite~, ~ + 1 ~ - I * +N'. ;, § i..I/."~ 4-
O,L
__.~.___,
- IN Cl i . r'." .I
FIG. 17. Schematic cross-section across an idealized nephelinite-carbonatite volcanic complex showing the normal structure and sequence of intrusion, the zonal fenitization around the ijolite, the deeper-level albitization (Na), the nearer-surface potassic feldspathization (K) accompanied by brecciation (A) and the emplacement of s6vite (CI), the cone-sheet fracture system of the alvikites (C2) and ferrocarbonatites (C3), and the final development of late-stage carbonate veins (C4) and a mineralized capping above the intrusive centre. (After Le Bas 1977, Fig. 23.1.) tization. Fluids e m a n a t i n g from early carbonatites cause albitization with albite-rich fenites +_ Na-pyriboles being developed at deeper levels around the carbonatite pluton. K-rich feldspathic fenites form towards the top and cap the s6vitic intrusions. They are often a c c o m p a n i e d by
brecciation and are usually followed by an F, Ba and U - T h mineralization. ACKNOWLEDGMENTS" I acknowledge most sincerely the many colleagues who have helped me reach these conclusions.
References ANDERSEN,T. 1984. Secondary processes in carbonatites: petrology of 'rodberg' (hematite-calcitedolomite carbonatite) in the Fen central complex, Telemark (South Norway). Lithos, 17, 227-45. APPLEYARD, E. C. & WOOLLEY,A. R. 1979. Fenitization: an example of the problems of characterizing mass transfer and volume changes. Chem. Geol. 26, 1-15. ARMBRUSTMACHER,T. J. 1979. Replacement and primary carbonatites from the Wet Mountains area, Fremont and Custer Counties, Colorado. Econ. Geol. 74, 888-901.
BAILEY, D. K. (1966) Carbonatite volcanoes and shallow intrusions in Zambia. In: TUTTLE,O. F. & GITTINS, J. (eds) Carbonatites, pp. 127-154. Wiley, New York. --1987. Mantle metasomatism--perspective and prospect. In: FtTTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 1-13. BAKER, B. Iq. 1987. Outline of the petrology of the Kenya rift alkaline province. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 293-311.
8o
M. J. Le Bas
BEDSON, P. 1983. The origin of the carbonatites and their relation to other rocks by liquid immiscibility. Ph.D Thesis, University of Manchester (unpublished). - 1984. Rare earth element distributions between immiscible silicate and carbonate liquids. Prog. Exp. Petrol. NERC 6th Rep., pp. 12-19. BOWEN, N. L. 1924. The Fen Area in Telemark, Norway. Am. J. Sci. 8, 1-11. BREY, G. 1978. Origin of olivine melilitites--chemical and experimental constraints. J. Volcanol. geotherm. Res. 3, 61-88. - - & GREEN, D. H. 1976. Solubility of CO2 in olivine melilitite at high pressures and role of COz in the earth's upper mantle. Contrib. Mineral. Petrol. 5 5 , 217-30. BRISTOW, J. W. & SAGGERSON, E. P. 1983. A general account of Karoo vulcanicity in southern Africa. Geol. Rundsch. 72, 1015-60. BROGGER, W. G. 1921. Die Eruptivgesteine des Kristianiagebietes. IV. Das Fengebeit in Telemark, Norwegen. Vidensk. Skr. 1. Mat-Naturv. Klasse, 1920 (9). BROOKS, C. K. & PRINZLAU,I. 1978. Magma mixing in mafic alkaline volcanic rocks: the evidence from relict phenocryst phases and other inclusions. J. Volcanol. geotherm. Res. 4, 315-31. BROWN, P. E. 1964. The Songwe scarp carbonatite and associated feldspathization in the Mbeya Range, Tanganyika. Q.J. geol. Soc. 120, 223-40. BULAKH, A . G . & ISKOZ-DOLININA, I. P. 1978. Origin of carbonatites in the Central'nyy massif, Turig peninsula. Int. geol. Rev. 20, 822-30. CLAGUE, D. A. & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. COETZEE, G. L. 1963. Carbonatites of the Karema depression, western Tanganyika. Trans. geol. Soc. S. Afr. 66, 283-333. COHEN, R. S., O'NIONS, R. K. & DAWSON, J. B. 1984. Isotope geochemistry of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth planet. Sci. Lett. 68, 209-20. COOPER, A, F., GxYrINS, J. & TUYrLE, O. F. 1975. The system Na2CO2-K2CO3-CaCO 3 at 1 kilobar and its significance in carbonatite petrogenesis. Am. J. Sci. 275, 534-60. DAWSON, J. B. 1962. The geology of Oldoinyo Lengai. Bull. Volcanol. 24, 349-87. --1980. Kimberlites and their Xenoliths, Springer, Berlin. --& FUGE, R. 1980. Halogen content of some African primary carbonatites. Lithos, 13, 139-43. -& HAWTHORNE,J. B. 1973. Magmatic sedimentation and carbonatitic differentiation in kimberlite sills at Benfontein, South Africa. J. geol. Soc. Lond. 12, 61-85. DEANS, T. & ROBERTS, B. 1984. Carbonatite tufts and lava clasts of the Tinderet foothills, Western Kenya: a study of calcified natrocarbonatites. J. geol. Soe. Lond. 141,563-80. - - , SUKHESWALA,R. N., SETHNA,S. F. & VILADKAR,
S. G. 1972. Metasomatic feldspar rock (potash fenites) asssociated with the fluorite deposits and carbonatites of Amba Dongar, Gujarat, India. Trans. Inst. rain. Metall. Sect. B, 81, 1-9. DIXEY, F., SMITH, W. C. & BISSET, C. B. 1935. The Chilwa Series of Southern Nyasaland. Geol. Surv. Nyasaland, Bull. 5 (revised 1955). DONALDSON, C. H. & DAWSON, J. B. 1978. Skeletal crystallization and residual glass compositions in a cellular alkalic pyroxenitic nodule from Oldoinyo Lengai. Contrib. Mineral. Petrol. 67, 139-49. DOWNES, H. 1984. Sr and Nd isotope geochemistry of coexisting alkaline magma series. Cantal, Massif Central, France. Earth planet. Sci. Lett. 69, 32134. DE BOlS, C. G. B., FURST, J., GUEST, N. J. & JENNINGS, G. J. 1963. Fresh natrocarbonatite lava from Oldoinyo L'Engai. Nature, Lond. 197, 445-6. VON ECKERMANN,H. 1948. The alkaline district of Alno Island. Sver. Geol. Unders. Ser. Ca, No. 36. EDGAR, A. D. & ARIMA, M. 1985. Fluorine and chlorine contents of phlogopites crystallized from ultrapotassic rock compositions in high pressure experiments : implication for halogen reservoirs in source regions. Am. Mineral. 70, 529-36. EGGLER, D. H. 1978. The effect of CO2 upon partial melting of peridotite in the system N a 2 0 - C a O A1203-MgO-SiO2-CO2 to 35 kb, with an analysis of melting in a peridotite-H20-CO2 system. Am. J. Sci. 278, 305-43. FODOR, R. V., BAUER, G. R. & KEIL, K. 1982. Ultramafic inclusions and megacrysts in olivine nephelinite. Aitutaki Island. Cook Islands. N.Z. J. Geol. Geophys. 25, 67-76. FREESTONE, I. 1978. Liquid immiscibility in alkaline magmas. Chem. Geol. 23, 115-23. & HAMILTON, D. L. 1980. The role of liquid immiscibility in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. GARSON, M. S. 1965. Carbonatites in Southern Malawi. Geol. Surv. Malawi, Bull. 15. GASPAR, J. C. & WYLLIE, P. J. 1983a. Magnetite in the carbonatites from the Jacupiranga complex, Brazil. Am. Mineral. 68, 195-213. - & -1983b. Ilmenite (high Mg, Mn, Nb) in the carbonatites from the Jacupiranga complex, Brazil. Am. Mineral. 68, 960-71. GITTINS, J. 1978. Some observations on the present status of carbonatite studies. Proc. 1st Int. Symp. Carbonatites, Ministerio das Minas e Energia, Departamento Nacional da Produ~ao Mineral, Polos de Caldas, Minas Gerais, Brazil, pp. 10715. -& McKm, D. 1980. Alkalic carbonatite magmas: Oldoinyo Lengai and its wider applicability. Lithos. 13, 213-15. , ALLEN, C. R. & COOPER, A. F. 1975. Phlogopitization of pyroxenite; its bearing on the composition of carbonatite magmas. Geol. Mag. 11,503-7. HAGGERTY, S. E. & MARIANO, A. N. 1983. Strontianloparite and strontio-cherkinite: two new minerals in rheomorphic fenites from the Parana Basin
Nephelinites and carbonatites carbonatites, South America. Contrib. Mineral. Petrol. 84, 365-81. HAY, R. L. 1983. Natrocarbonatite tephra of Kerimasi volcano, Tanzania. Geology, 11,599-602. HEINRICH, E. WM. 1966. The Geology of Carbonatites, Rand McNally, Chicago, I1. --& MOORE, D. G. 1970. Metasomatic potash feldspar rocks associated with igneous alkaline complexes. Can. Mineral. 10, 571-84. JAN, M. Q., KAMAL,M. & QURESHI,A. 1981. Petrography of the Loe Shilman carbonatite complex, Khyber Agency. Geol. Bull. Univ. Peshawar, 14, 29-43. JONES, A. P. & WYLLIE, P. J. 1983. Low-temperature glass quenched from a synthetic, rare earth carbonatite: implications for the origin of the Mountain Pass deposit, California. Econ. Geol. 78, 1721-3. KAPUSTIN, Y. L. 1981. Mineralogy of Carbonatites. Amerind, New Delhi. --1982. Fenitization around carbonatite bodies. Dokl. Akad Nauk SSSR, 249, 126-8. KELLER, J. 1981. Carbonatite volcanism in the Kaiserstuhl alkaline complex: evidence for highly fluid carbonatite melts at the Earth's surface. J. Volcanol, geotherm. Res. 9, 423-31. KING, B. C. 1949. The Napak area of southern Karamoja, Uganda. Geol. Surv. Uganda, Mem. V. --1965. Petrogenesis of the alkaline igneous rock suites of the volcanic and intrusive centres of eastern Uganda. J. Petrol. 6, 67-100. --& SUTHERLAND, D. S. 1960. Alkaline rocks of eastern and southern Africa. Sci. Prog., Lond. 48, 298-321,504-24, 709-20. KOSTER VAN GROOS, A. F. 1975. The distribution of strontium between coexisting silicate and carbonate liquids at elevated pressures and temperatures. Geochim. cosmochim. Acta, 39, 27-34. -d~ WYLLIE, P. J. 1966. Liquid immiscibility in the system Na20-A1203-SiO2-CO2 at pressures to 1 kilobar. Am. J. Sci. 26, 234-55. --& ~ 1968. Liquid immiscibility in the join NaA1Si3Oa-Na2CO3-H20 and its bearing on the genesis of carbonatites. Am. J. Sci. 266, 932-67. --&--1973. Liquid immiscibility in the join NaA1Si308-CaAI2Si2Os-Na2CO3-H20. Am. J. Sci. 273, 465-87. LAPIN, A. V. & VARTIAINEN,H. 1983. Orbicular and spherulitic carbonatites from Sokli and Vuoriyarvi. Lithos, 16, 53-60. LE BAS, M. J. 1971. Per-alkaline volcanism, crustal swelling and rifting. Nature, Lond., Phys. Sci. 230, 85-7. - - 1 9 7 7 . Carbonatite-Nephelinite Volcanism, Wiley, New York. --1978. Nephelinite volcanism at plate interiors. Bull. Volcanol. 41,459-62. --1980. The East African Cenozoic magmatic province. Acad. Naz. Lincei, Atti Convegni Lincei, 47, 111-22. --1981. Carbonatite magmas. Mineral. Mag. 44,
133-40. Oceanic carbonatites. In: KORNPROBST, J.
--1984.
81
(ed.) Kimberlites 1 : Kimberlites and Related Rocks, pp. 169-78. Elsevier, Amsterdam. -1986. Ultra-alkaline magmatism with or without rifting. Tectonophysics (Special Issue on Continental and Oceanic Rift Zones), in press. --& ASPDEN, J. A. 1981. The comparability of carbonatitic fluid inclusions in ijolites with natrocarbonatite lava. Bull. Volcanol. 44, 429-38. --& HANDLEY, C. 1979. Variations in apatite composition in ijolitic and carbonatitic igneous rocks. Nature, Lond. 27, 54-6. LEHIJARVI, M. 1960. The alkaline district of Iivaara, Kuusamo, Finland. Bull. Comm. Geol. Fin. 185. LIPPARD, S. J. 1973. The petrology of phonolites from the Kenya Rift. Lithos, 6, 207-34. McKIE, D. 1966. Fenitization. In: TUTTLE, O. F. & GITTINS, J. (eds) Carbonatites, pp. 261-94. Wiley, New York. - - & FRANKIS,E. J. 1977. Nyerereite: a new volcanic carbonate mineral from Oldoinyo Lengai, Tanzania. Z. Kristallogr. 145, 73-95. MARIANO, A. N. & ROEDER, P. L. 1983. Kerimasi: a neglected carbonatite volcano. J. Geol. 91,449-55. MOLLER, P., MORTEANI,G. & SHLEY, F. 1980. Discussion of REE distribution patterns of carbonatites and alkaline rocks. Lithos, 13, 171-9. MYSEN, B., RYERSON, F. J. & VIRGO, D. 1981. The structural role of phosphorus in silicate melts. Am. Mineral. 66, 106-17. NESmTT, B. E. & KELLY, W. C. 1977. Magmatic and hydrothermal inclusions in carbonatite of the Magnet Cove complex, Arkansas. Contrib. Mineral. Petrol. 63, 271-94. NORRY, M. J., TRUCKLE, P. H., LIPPARD, S. J., HAWKESWORTH, C. J., WEAVER, S. D. & MARRINER, G. F. 1980. Isotopic and trace element evidence from lavas, bearing on mantle heterogeneity beneath Kenya. Phil. Trans. R. Soc. Lond., Ser. A, 297, 259-71. OLSEN, J. C., SHAW, D. R., PRAY, L. C. & SHARP, W. N. 1954. Rare-earth mineral deposits of the Mountain Pass District, San Bernardino County, California. U.S. Geol. Surv. Prof Paper 261. ONUMA, K. & TOHARA, T. 1984. The join CaMgSi206CaA12SiO6-CaCrA1SiO6 with special reference to chrome clinopyroxene and chrome spinel. Lithos, 17, 289-98. ORVILLE, P. M. 1963. Alkali ion exchange between vapour and feldspar phases. Am. J. Sci. 261,20137. PIOTROWSKI, J. M. & EDGAR, A. D. 1970. Melting relations of undersaturated alkaline rocks from S. Greenland compared to those of Africa and Canada. Medd. Gronl. 181 (9). PRINS, P. 1972. Composition of magnetite from carbonatites. Lithos, 5, 227-40. s 1978. Geochemical aspects of the alkaline carbonatite complexes of Damaraland, South West Africa. Proc. 1st Int. Symp. on Carbonatite, Ministerio das Minas e Energia, Departmento Nacional da Produ~ao Mineral, Polos de Caldas, Minas Gerais, Brazil, pp. 179-88. 1981. The geochemical evolution of the alkaline and carbonatite complexes of the Damaraland
82
M. J. Le Bas
igneous Province, South West Africa. Annen. Univ. Stellenbosch, Ser. A1 (Geol.) 3, 145-278. RANKIN, A. H. 1975. Fluid inclusion studies in apatite from carbonatites of the Wasaki area of western Kenya. Lithos, 8, 123-36. -1977. Fluid-inclusion evidence for the formation conditions of apatite from the Tororo carbonatite complex of eastern Uganda. Mineral. Mag. 41, 155-64.
LE BAS, M. J. 1974a. Nahcolite (NaHCO3) in inclusions in apatites from some E. African ijolites and carbonatites. Mineral. Mag. 39, 563-70. - & -1974b. Liquid immiscibility between silicate and carbonate melts in naturally-occurring ijolite magma. Nature, Lond. 2 5 0 , 206-9. ROBBINS, B. &. TYSSELAND, M. 1983. The geology, geochemistry and origin of ultrabasic fenites associated with the Pollen carbonatite (Finnmark, Norway). Chem. Geol. 40, 65-95. ROEDDER, E. 1973. Fluid inclusions from the fluorite deposits associated with carbonatite at Amba Dongar, India and Okorusu, South West Africa. Trans. Inst. Min. Metall., Sect. B, 82, 35-9. RUBIE, D. C. 1982. Mass transfer and volume change during alkali metasomatism at Kisingiri, western Kenya. Lithos, 15, 99-109. - & GUNTER, W. D. 1983. The role of speciation in alkaline igneous fluids during fenite metasomatism. Contrib. Mineral. Petrol. 82, 165-75. SAErHER, E. 1957. The alkaline rock province of the Fen area in southern Norway. Kgl. norske Videns. Selsk. Skr. 1957 (1). SAGGERSON, E. P. 1970. The structural control and genesis of alkaline rocks in central Kenya. Bull. Volcanol. 34, 38-76. SAMOVLOV,V. S. & SMIRNOVA,Ye. V. 1980. Rare earth behaviour in carbonatite formation and the origin of carbonatites. Geochem. Int. 17, 140-52. SnAND, S. J. 1949. Eruptive Rocks. Thomas Murby, London. SILVA, L. C., LE BAS, M. J. & ROBERTSON, A. H. F. 1981. An oceanic carbonatite volcano on Santiago. Cape Verde Islands. Nature, Lond. 294, 664-5. SPENCER, A. B. 1969. Alkalic igneous rocks of the Balcones Province, Texas. J. Petrol. 10, 272-306. STRECKEISEN,A. 1976. To each plutonic rock its proper name. Earth Sci. Rev. 12, 1-33. -1978. IUGS Subcommission on the systematics of igneous rocks; classification and nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Neues Jb. Mineral. Abh. 134 (i), 114. STRONG, D. F. 1972. Petrology of the island of Moheli, western Indian Ocean. Geol. Soc. Am. Bull. 83, 389-406. SUTHERLAND,D. S. 1965. Nomenclature of the potassicfeldspathic rocks associated with carbonatites. Geol. Soc. Am. Bull. 76, 1409-12. -1967. A note on the occurrence of potassium-rich trachytes in the Kaiserstuhl carbonatite complex, West Germany. Mineral. Mag. 36, 334-41. SUWA, K., OANA, S., WADA, H. & OSAKI, S. 1975. - - &
Isotope geochemistry and petrology of African carbonatites. Phys. Chem. Earth, 9, 735-45. THOMPSON,R. N. 1974. Some high-pressure pyroxenes. Mineral. Mag. 39, 768-87. TREIMAN, A. H. & ESSENE, E. J. 1983. Mantle eclogite and carbonate as sources of sodic carbonatites and alkalic magmas. Nature, Lond. 302, 700-3. --&--1984. A periclase-dolomite-calcite carbonatite from the Oka complex, Quebec, and its calculated volatile composition. Contrib. Mineral. Petrol. 85, 149-57. - - & SCHEOL, A. 1983. Properties of carbonatite magma and processes in carbonatite magma chambers. J. geol. 91,437-47. TUTTLE, O. F. & GITTINS, J. 1966. Carbonatites. Wiley, New York. TWYMAN, J. D. & GITTINS,J. 1987. Alkalic carbonatite magmas: parental or derivative. In: FITI'ON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 85-101. TYLER, R. C. & KING, B. C. 1967. The pyroxenes of the alkaline igneous complexes of eastern Uganda. Mineral. Mag. 36, 5-21. VARNE, R. 1968. The petrology of Moroto Mountain, eastern Uganda, and the origin of nephelinites. J. Petrol. 9, 169-90. VARTIAINEN, H. 1980. The petrology, mineralogy and petrochemistry of the Sokli carbonatite massif, northern Finland. Geol. Surv. Finland, Bull. 313. VERWOERD, W. J. 1978. Liquid immiscibility and the carbonatite-ijolite relationship: preliminary data on the join NaFea+Si206-CaCO3 and related compositions. Ann. Rep. geophys. Lab. Wash. 77, 767-74. VILADKAR,S. G. & DULSKI,P. 1986. Rare earth element abundances in carbonatites, alkaline rocks and fenites of the Amba Dongar complex, Gujarat, India. Neues Jb. Mineral. 1986, H.1, 37-48. WATKINSON,D. H. & WYLLIE, P. J. 1971. Experimental study of the composition join NaAISiO4-CaCO3H20 and the genesis of alkalic rock-carbonatite complexes. J. Petrol. 12, 357-78. WENDLANDT, R. F. & HARRISON, W. J. 1979. Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapour: results and implications for the formation of light rare earthenriched rocks. Contrib. Mineral. Petrol. 69, 40919. WOOD, D . A . , JORON, J.-L., TREUIL, M., NORRY, M. J. & TARN~Y, J. 1979. Elemental and Sr isotope variations in basic lavas from Iceland and surrounding ocean floor. Contrib. Mineral. Petrol. 70, 319-39. WOOLLEY, m. R. 1982. A discussion of carbonatite evolution and nomenclature, and the generation of sodic and potassic fenites. Mineral. Mag. 46, 1317. - & JONES, G. C. 1987. The petrochemistry of the northern part of the Chilwa Alkaline Province, Malawi. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline lgneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 335-55.
Nephelinites and carbonatites WYLLIE, P. J. 1965. Melting relationships in the system CaO-MgO-CO2-HzO, with petrological applications. J. Petrol. 6, 101-23. -1978. Silicate-carbonate systems with bearing on the origin and crystallization of carb0natites. Proc. 1st Int. Syrup. on Carbonatites, Ministerio das Minas e Energia, Departamento Nacional da
83
Produqao Mineral, Pogos de Caldas, Minas Gerais, Brazil, pp. 61-78. - - - 1980. The origin of kimberlites. J. geophys. Res. 85, 6902-10. YARDLEY,B. 1985. Apatite composition and fugacities of HF and HC1 in metamorphic fluids. Mineral. Mag. 49, 77-9.
M. J. LE Bks, Department of Geology, University of Leicester, Leicester, U,K.
Alkalic carbonatite magmas- parental or derivative? James D. Twyman & John Gittins S U M M A R Y : During the 1950s and 1960s the Tanzanian volcano Oldoinyo Lengai erupted a highly alkalic lava composed almost exclusively of the N a - K - C a carbonate minerals nyerereite and gregoryite and containing about 32~ Na20 and 7% K20. For many years the lava was considered a petrological curiosity but gradually the idea developed that alkalic carbonatite magmas might develop generally during the evolution of commoner carbonatite magmas. In 1981 Le Bas, followed in 1982 by Woolley, ascribed to the Oldoinyo Lengai magma a parental status and erected a scheme whereby other commoner carbonatite rock types are derived from it; the alkalic carbonatite magma is one of two derived by immiscible separation from a nephelinitic magma, the other being ijolitic. The carbonatite liquid has 500~ of superheat and loses alkalis progressively as the magma cools to its liquidus temperature of 400~176 by which stage it has become a calcitic-dolomitic liquid. This liquid then differentiates to produce the commoner types of carbonatite rocks. In rejecting such a scheme we argue as follows: such a scheme would require the magma to be water saturated at the moment of immiscibility, and progressive loss of water and alkalis would induce crystallization thus preventing the formation of a calcite-dolomite magma; the 500~ of superheat is a consequence of the design of the Freestone & Hamilton experiments and does not exist; carbonatite magmas generally begin to crystallize at temperatures greater than 900~ rare-earth-element compositions do not permit derivation of alkalic carbonatite magma from nephelinite magmas by liquid immiscibility; the mafic silicate mineralogy of carbonatites demonstrates alkali enrichment rather than depletion. We propose that the commonest parental magma of carbonatites is a mildly alkalic olivine s6vite composition and that most carbonatite rock types are derived from it by fractional crystallization, allowing the development of cumulates that are well lubricated by intercumulus liquid and capable of much subsequent movement, deformation and re-intrusion as a crystal mush. As this parental magma crystallizes, continued fractionation of alkalideficient and anhydrous minerals increases both its alkali and water contents until water saturation is reached. This in turn causes the development of an aqueous fluid which limits the alkali content of the magma. The composition of the magma at which water saturation develops is controlled by the rate of rise of the magma, and therefore aqueous fluids with various N a : K ratios can develop and control the type of fenitization that occurs. Under plutonic conditions alkalis in excess of the amount that can be fixed as silicate minerals are carried away as fenitizing fluids. Alkali loss does therefore occur, and through the medium of an aqueous fluid, but calcite and/or dolomite crystallize throughout the magmatic cooling history rather than simply at the low-temperature end of this history. The alkalic carbonatite magma of Oldoinyo Lengai type also develops through fractional crystallization of the same alkali-poor olivine s6vite magma, but under essentially dry conditions where the magma is kept liquid by alkalis and halogens, and alkali loss is prevented by the absence of an aqueous phase. Alkalic carbonatite magmas are therefore late differentiates of a more normal mildly-alkalic olivine s6vite magma developed under low water fugacity. They are of very small volume and require long quiescence to develop. They are not to be considered parental to other carbonatite rock types.
Introduction The problem of the c h a r a c t e r and origin of carbonatite m a g m a s is at once tantalizing and enigmatic. For m a n y years few geologists believed in the existence of m a g m a t i c c a r b o n a t e rocks at all, largely because of the high melting t e m p e r a t u r e of CaCO3 and failure to appreciate the effect of H 2 0 , F a n d alkali carbonates in lowering that temperature. T h e melting studies p e r f o r m e d by Wyllie and Tuttle a n d their coworkers b e t w e e n 1957 and the 1970s gradually o v e r c a m e these conceptual difficulties and the
petrological world was won over to the idea of carbonatite m a g m a s . Petrologists in the U. S.S.R., w h e r e there has always been a great propensity for m e t a s o m a t i s m , r e m a i n e d resistant to the idea of carbonatite m a g m a s , but they have n o w j o i n e d the igneous fold. Once the initial barrier of carbonate melting at reasonable t e m p e r a t u r e and pressure was overcome, the course of petrogenetic thought was g o v e r n e d largely by the melting studies of Wyllie and his students at the P e n n s y l v a n i a State U n i v e r s i t y and later at the U n i v e r s i t y of Chicago. T h e d e v e l o p m e n t of calcitic and dolomitic liquids
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 85-94.
85
86
J. D . T w y m a n & J. G i t t i n s
from various ultramafic ultrabasic compositions over a wide range of crustal and mantle pressures was pursued in this work. It must be remembered that this experimental programme was not designed solely to elucidate carbonatite genesis, but was at least as much concerned with the development of kimberlite magmas and magmas that might be parental to the basalts of the Earth's surface and upper crust. Consequently, Na and K did not figure prominently in these experimental systems. The general pattern emerged that calcitic and dolomitic liquids, given adequate H20 and CO2 fugacities, could develop from certain SiO2undersaturated peridotitic compositions such as might commonly be expected in the Earth's mantle. Consequently, they were thought to offer a probable explanation of how carbonatite magmas are formed. However, at the same time some petrologists were suggesting that these experimentally derived liquids were more applicable to the rather special type of carbonatite that is associated with kimberlites and the carbonate that so frequently pervades kimberlite intrusions, and that they might have relatively little relevance to the carbonatites of the classic carbonatitealkalic igneous rock association. Almost at the beginning of the melting studies an interesting turn of events was introduced when in 1960 the Tanzanian volcano Oldoinyo Lengai erupted its unusual N a - C a - K carbonate lava at temperatures too low to be incandescent. Earlier eruptions had, of course, been noted in the 1950s by Guest (1956), but were so unusual that the significance was not widely recognized. The initial tendency, with a few exceptions, was to consider it as a petrological curiosity of little application to the broader problems of carbonatite genesis, but it gradually assumed greater importance in the minds of certain petrologists. Dawson (1964) had originally suggested its importance and he was later supported by Cooper et al. (1975) and Gittins et al. (1975). In each of these papers it was argued that alkalic carbonatite magmas probably develop during the evolution of a carbonatite-igneous rock complex but that in most cases the alkalis are lost to aqueous or aqueous-halide fluids that may be involved in fenitization. The residues were thought to be normal calcitic and dolomitic carbonatites. In general it was felt that there was overwhelming evidence that most carbonatite magmas could not just be calcitic or dolomitic because of the frequent presence within the rocks of sodic pyroxenes, phlogopite-biotite and sodi-potassic amphibole, as well as fenitization adjacent to dykes and other intrusions; extreme examples were the presence of several centimetres of
phlogopite-rich rock in pyroxenite adjacent to carbonatite dykes. The concept emerged, and has now been accepted more widely, that carbonatite rocks do not necessarily represent the composition of the magma from which they crystallized. Le Bas (1981) argued the case that a highly alkalic magma of Oldoinyo Lengai type is parental to the full range of carbonatite rock types. His thesis was that carbonated nephelinite magma separates immiscibly into an ijolitic silicate liquid and an alkalic carbonate liquid at a temperature of 1000~176 The silicate liquid is said to be at its liquidus but the carbonate liquid is at least 500~ above its liquidus, with crystallization not commencing until 400~ 600~ The pressure at which the immiscible separation occurs was not stated but the context seems to imply lower-crustal rather than mantle conditions. The scheme envisaged by Le Bas involves the formation of a progressively lesssodic carbonate magma as alkalis are lost, presumably, although not clearly stated, during cooling through the 500~ superheat range, followed by crystallization differentiation that forms calcitic carbonatites at first and then increasingly more dolomitic and ankeritic carbonatites as the magma cools from its 400~ 600~ liquidus. Le Bas further suggested that the proportion of Ca: Na: K in the alkalic carbonatite magma may be a function of the temperature and pressure at which immiscible separation occurred, a conclusion drawn from the experimental work of Freestone & Hamilton (1980) and which he referred to as 'dynamic un-mixing'. The variable proportion of Na-rich to K-rich fenites was attributed to possible temperature,controlled variations in relative Na and K mobility, which Gittins et al. (1975) had also invoked. The following year Woolley (1982), while supporting the evolutionary concept of the formation of calcitic, dolomitic and ankeritic carbonatites from alkalic magma through alkali loss and fractional crystallization, introduced a modification by suggesting that the aqueous fluid becomes progressively more potassic. In this way he sought to explain the superposition of potassic fenitization on early sodic fenitization observed in some carbonatites. Thus the highly alkalic carbonatite magma of Oldoinyo Lengai was promoted from a stage in the evolution of carbonatite complexes (Cooper et al. 1975) to a parental magma derived immiscibly from a carbonated nephelinite magma. Purportedly, it gives rise, through progressive loss of alkalis followed by fractional crystallization, to the full diversity of carbonatite rock types. It is this parental status and proposed differentiation scheme with which we disagree. We
87
Alkalic carbonatite magmas argue our case by examining progressively the evidence of an immiscible origin, the evidence for or against alkali depletion, rare-earth-element (REE) data, the problems of generating and then abstracting an aqueous fluid phase and problems of magma temperatures.
Critique of the parental-status scheme The Freestone & Hamilton experiments Freestone & Hamilton (1980) established that molten nephelinite and phonolite are each immiscible with molten alkalic carbonatite lava. What was done was the following: a nephelinite and an alkalic carbonatite lava, both of which had erupted from the volcano Oldoinyo Lengai, were mixed together; the mixture was heated to a temperature above the liquidus of both rocks and then quenched in order to establish texturally that two immiscible liquids had coexisted. The experiments were repeated with phonolite and produced similar results. It was concluded that the silicate and carbonate liquids separated in the volcano from a common parent, for which phonolite was preferred to nephelinite. However, the experiments do not prove this. They establish that the two liquids are immiscible at 1 kb but not that they came into being together. There is a weakness in the design of the experiments. If the alkalic carbonatite is a highly evolved rock with a low liquidus temperature (about 500~ as we shall attempt to show, then melting it with a silicate rock whose liquidus is about 1000~ is bound to produce liquids with sharply divergent liquidi. The superheat is a consequence of the experimental design, and the argument is dangerously circular. Two rocks from the same volcano, but which may have been derived by fractional crystallization of two entirely different magmas, are melted together and found to produce immiscible liquids with liquidi 500~ apart. This is held to prove that the two rocks represent the compositions of magmas that were derived immiscibly from a common parent. The experiments prove what is possible but not that the possible happened. It is analogous to calculations that show a reaction to be thermodynamically feasible but not whether it is kinetically possible. We agree that alkalic carbonatite and silicate magmas are mutually immiscible, but we believe that the immiscibility occurred a long way back in the petrogenetic history, and in the mantle.
Alkali depletion or alkali enrichment? The arguments surrounding the alkalic carbonatite question have become rather theoretical with insufficient attention being paid to real rocks. Natural carbonatites can settle the question of whether alkali depletion occurs, for if it does the associated silicate minerals ought to reflect this depletion. Amphiboles from three calcitic and dolomitic carbonatite complexes are shown in Fig. 1. The direction of amphibole evolution is established by compositional zonation and is, from core to rim, an increase in Na + Fe accompanied by a decrease in Ca + Mg. In short, it is the reverse of what is required by the parentalstatus scheme. Rare-earth-element data It is possible to calculate the REE distribution in a carbonate liquid that would be in equilibrium with the Oldoinyo Lengai phonolite by using the partition coefficients published by Wendlandt & Harrison (1979). The result is shown in Fig. 2 where it is compared with the measured patterns of the alkalic carbonatite lava and both the nephelinite and the phonolite. Even after allowing for the fact that the nephelinite and phonolite analyses are not of the highest quality there is a considerable discrepancy between the calculated carbonatite pattern and the measured pattern, 0"7 [ " OLIVINE CARBONATITES l + CLINOHUM,CARBONATITES/
t( ,
~
l-Z+. ~ \
0.5
Co
9 ~%t
0.4
No+K 0.5 0.2 0.1 ol
"'s
~
,
I
l I0
,
I , 20 30
Mg/Fe FIG. 1. Compositions of amphiboles from the Argor and Goldray carbonatites, Ontario, Canada (Twyman 1983) and the Safartoq carbonatite, Greenland (Secher & Larsen 1980). The arrows indicate the direction of compositional change from core to rim. The petrographic varieties represented are those with amphibole alone, amphibole with titanian clinohumite and amphibole with olivine. The general trend of amphibole evolution is enrichment in Na, K and Fe with depletion in Ca and Mg.
88
J. D. Twyman & J. Gittins IO3
\.
,o'
-
--__-~ I I LO Ce
I Nd
I I 1 SmEu Tb
II YbLu
FIG. 2. Chondrite-normalized REE distribution for Oldoinyo Lengai lavas. The calculated composition of the alkalic carbonatite ( - - - - - - ) is based on the distribution coefficients of Wendtlandt & Harrison (1979) for a carbonatite in equilibrium with nephelinite and a phonolite whose measured values are taken from Gerasimovsky et al. (1972). The measured values of the alkalic carbonatite are taken from Twyman (1983) ( - - . . - - ) and Philpotts et al. (1972) (----). and the divergence increases towards the heavy REE. It has been argued already that the melting experiments performed by Freestone & Hamilton (1980) offer no proof that alkalic carbonatite magma was derived from nephelinite magma. The REE data provide further support for this view. The problem of generating an aqueous fluid from the alkalic carbonatite m a g m a
Progressive alkali loss is supposed to have occurred through 'exchange processes [with the] enveloping country rock' and 'by reaction' (Le Bas, 1981, pp. 138, 139). The aqueous fluid that has been invoked by other workers (cited earlier) seems to be inherent in the proposed alkali loss, and so such a fluid must have been generated from the carbonate liquid in some manner. That, in turn, requires the liquid to have reached water saturation. The available possibilities are that the carbonate liquid was saturated at the moment of immiscible separation from its alleged nephelinitic parent, or that it reached saturation during cooling. It is unlikely that the carbonate liquid reached saturation during cooling. The commonest process by which an aqueous fluid is developed from a magma is the progressive crystallization of anhydrous minerals which increases the concentration of water in the remaining liquid. Although the minerals of the Oldoinyo Lengai carbonatite lava are anhydrous, this process is not available
because we are asked to believe that the magma cooled through 500~ of superheat. It was therefore above its liquidus and so could not have been crystallizing. It is possible for a magma to reach water saturation as it rises to a regime of lower pressure but only if the solubility of water in it is relatively low. Solubility data for alkalic carbonatite magmas are incomplete, but it is probable that the solubility is very high and that saturation would not be reached in this way. The remaining alternative, that the carbonate liquid was water saturated at the moment of immiscibility, is also fraught with difficulty. This would require the ijolitic liquid as well as the alkalic carbonatite liquid to be saturated. Yet ijolites, being anhydrous rocks, appear to have crystallized from fairly dry magmas. Furthermore, nephelinitic magmas have fairly low water contents and so, for the alkalic carbonate magma to reach saturation, the available water would have to partition almost exclusively into the carbonate liquid. However, available data on water solubility in nepheline-bearing liquids suggest that the solubility of water in ijolitic magma would not be inconsiderable. The problem of generating an aqueous fluid within the constraints of the parental-status scheme appears to be insuperable, yet without such a fluid the scheme does not work. The effect of continuous removal of water and alkalis from a m a g m a
Suppose that an aqueous fluid could be generated in some manner. What would be the effect of its continued removal from the alkalic carbonatite magma? It has been well established experimentally, through the studies of Wyllie, Gittins and their coworkers, that calcitic and dolomitic liquids are possible primarily through the grace of water and alkalis. If either, or both, are removed, crystallization is the inevitable result. This is especially so since the solubility of water in calcitic and dolomitic liquids is markedly less than in sodi-potassic carbonate liquids. In short the proposed method by which the magma composition changes would be fatal to the magma's continued existence. In addition, under the proposed magmatic evolutionary scheme there is no reason why the Oldoinyo Lengai alkalic carbonatite lava should have erupted at all. And if it did, it ought to have been a vigorous eruption as its aqueous phase escaped and the remaining water in the saturated magma boiled offexplosively. In contrast, the 1960 eruption was quiescent and composed of anhydrous lavas. If the proposed alkali loss was not through the medium of an aqueous fluid but rather by wall-
89
Alkalic carbonatite magmas rock reaction, the same arguments apply. Reaction of such a magma with the enclosing silicate rocks must, inevitably, lead rapidly to crystallization.
Magma temperatures and crystallization temperatures The parental-status scheme has the crystallization of calcitic and dolomitic magmas commencing at 400~176 and is in accord with much popular thinking about the temperature ranges over which carbonatite magmas crystallize. Unfortunately, there is a good deal of misleading thought on this subject. Low temperatures (below 500~ are overemphasized, largely through misuse of the calcite-dolomite geothermometer. The fact that temperatures obtained from measurements of the amount of Mg substituting Ca in calcite that coexists with dolomite are minimum temperatures is often overlooked. If dolomite has exsolved from the calcite electron microprobe analysis cannot generally be used, and even if the calcite is homogeneous it cannot be assumed that there has been no diffusion of Ca and Mg between calcite and dolomite after crystallization, particularly if a fluid phase is present. Published crystallization temperatures are rarely over 550~ and many are as low as 200~ Yet it can be shown that in at least one carbonatite complex (Argor, N Ontario, Canada) a carbonate mineral began to crystallize at a temperature higher than 885~ The bulk composition of coarsely exsolved single crystals (74 CACO322 MgCO3-4 FeCO3) gives an equilibrium temperature of 885~ indicating that the liquidus temperature of the carbonatite magma was at least that high. Recrystallization and re-equilibration during cooling render accurate geothermometry in carbonatites very difficult and have often caused overemphasis on the 500~ range (Gittins 1979). Argor offers proof of the existence of a carbonatite magma at not less than 900~ and we suggest that this is an approximate liquidus temperature for many carbonatite magmas. It is, of course, in sharp contrast with the 400~ 600~ liquidus temperature required by Le Bas for the magma from which calcitic and dolomitic carbonatites crystallized.
The volume requirements of the alkali-depletion scheme A very general idea of the volume of material that would have to be removed from an alkalic carbonatite magma to produce a calcitic-dolomitic carbonatite can be obtained from a simple
subtraction diagram. This does not require knowledge of the fluid phase compositions since we can assume that everything in the fluid came originally from the magma. In Fig. 3 the data are the compositions of the Oldoinyo Lengai alkalic carbonatite lava, the solid composition of the subtracted phase and the average carbonatite of Gold (1966) which, despite the limited value of averages from that period, suffices for the purpose. The composition of the subtracted phase is taken at FeO = 0, and the diagram indicates that 95~ of the alkalic carbonatite magma must be removed to convert it into normal carbonatite. If the calculations are repeated for a subtracted phase with SiO 2 -- 0, the proportion is still 90%. If we are to assume that all normal calciticdolomitic carbonatites are the residue remaining after the removal of 9 0 ~ - 9 5 ~ of the mass of a parental alkalic carbonatite magma the volume requirements become unrealistically enormous, even if it were possible to have a magma still in existence after abstracting so much water and alkalis. Again, if wall-rock reaction rather than fluid involvement is responsible for the alkali loss the proportion of carbonatite to metasomatic encircling rock would have to be vastly different from what is generally observed.
Origin of alkalic carbonatite magma We have attempted to show that alkalic carbonatite magma is not parental to the more normal w t~ < -r
laJ
w
R.~o
o
----Z
Z
4 0 " ""~:
~"
OXIDE IO
0
I 2
5
4
5
6
S iO 2
FIG. 3. Subtraction diagram showing the volume and composition of a fluid phase that must be removed from an alkalic carbonatite to produce a normal carbonatite. See text for further discussion.
J.D. Twyman & J. Gittins
90
calcitic-dolomitic carbonatites. Yet it clearly exists. Where, then, does it come from? We believe that it can be generated by fractional crystallization of a mildly-alkalic olivine s6vite magma which we propose as a fairly common carbonatite parental magma. The arguments in support of such a composition are to be published separately but are based on the magnesian composition of carbonatite olivines (most commonly F095_85), which implies generation of the magma from a very mafic source, and upon the partition coefficients derived by Freestone & Hamilton (1980) from their melting studies. Although we do not believe that the results of their experiments demonstrate what they set out to prove, the data can be used to estimate the N a 2 0 and K 2 0 content of carbonatite liquids in equilibrium with immiscible silicate liquids of
various compositions. We believe that the olivine s6vite magma is the result of immiscible separation from an olivine nephelinite magma at a pressure of approximately 27 kb. The partition coefficients for nephelinite give N a 2 0 + K 2 0 18~, but a more primitive melanephelinite gives a value of about 8%. Our conclusion is that the commonest parent magma that generates the diversity of carbonatite rocks is an olivine s6vite with about 8% N a 2 0 + K20. A computer-modelling scheme has been employed to test whether an Oldoinyo Lengai alkalic carbonatite lava can be produced by fractional crystallization of this mildly-alkalic olivine s6vite magma. The compositions used are an olivine s6vite from the Argor intrusion (Ontario, Canada) with additional alkalis to bring N a 2 0 + K 2 0 up to 8.1~, and the mineral compositions from
TABLE 1. Results of the computer modelling of the differentiation of carbonatite magmas
SiO2 A12Oa TiOz FeO MgO MnO CaO Na20 K20 PEO5 CO2 Total
Weighting factor
Parent olivine s6vite
5.0 1.0 1.0 20.0 5.0 0.1 10.0 50.0 10.0 50.0 5.0
6.56 10.66 1.09 10.66 3.68 0.25 34.50 5.74 2.41 5.92 28.17 100.00
Target alkalitic carbonatite 0.66 0.33 0.11 0.33 1.34 0.16 17.80 33.86 8.18 1.09 36.34 100.00
Calculated target 0.66 0.33 0.12 0.33 1.34 0.14 17.85 33.84 8.30 1.09 36.66 100.44
Difference (%) 0.0 0.0 1.0 0.0 0.2 - 14.6 0.3 - 0.1 1.5 0.0 0.9
Minerals removed ct SiO2 A1203 TiO 2 FeO MgO MnO CaO NazO
1.77 0.68 0.36 51.07
ilm
51.70 48.30
mag
apt
3.05 96.84 0.11 57.10
dt
58.39
42.69 9.93 0.45 16.50 20.05 0.08 0.09
9.52 14.00 1.65 29.34
2.04 8.16 1.74
P2Os Total Solution
bt
0.03 12.57 17.07
K20 CO 2
amph
10.21
42.90 46.13
45.50
100.00
100.00
100.00
100.00
-44.67
-1.57
-6.80
-13.40
100.00 -3.78
100.00 -9.95
ct, calcite; ilm, ilmenite; mag, magnetite; apt, apatite; amph, amphibole; bt, biotite; dt, dolomite. 83.87~ crystallization of parent olivine s6vite will generate a residual alkalic carbonatite.
100.00 -3.70
Alkalic carbonatite magmas the Argot carbonatites. The computer program is a modification by Dr. M. P. Gorton (University of Toronto) of the linear-programming and leastsquares method P E T M I X due to Wright & Doherty (1970). It is essentially a reiterative regression analysis that minimizes the squares of the residuals. The results of the calculation are given in Table 1, and in summary it can be said that 84~ crystallization can generate a lava of Oldoinyo Lengai type by removal of minerals that would form a rock composed of 53% calcite, 4 ~ dolomite, 16~ apatite, 8 ~ magnetite, 2 ~ ilmenite, 12~ biotite and 5 ~ amphibole. This will be recognized as a common type of s6vite. We do not claim a definitive status for the calculation but we believe that it does demonstrate that alkalic liquid can be derived through crystal fractionation of mildly-alkalic olivine s6vite magma. Further support for this concept comes from the chemical compositions of the amphiboles and phlogopite-biotites from the Argor intrusion. Throughout most of the crystallization of the magma there is very little change of A1 relative to Mg/Fe in amphibole (Fig. 4) or of Ti relative to Mg/Fe in biotite (Fig. 5), but in a small number of highly evolved rocks there is a substantial increase amounting to tenfold for A1 and thirteenfold for Ti. Both elements can be treated as trace elements and the curves are strikingly similar to trace-element enrichment curves for liquid predicted by Rayleigh fractionation with bulk crystal-liquid partition coefficients of less than unity. These enrichments can be used to estimate the amount of crystallization required to produce them: about 9 0 ~ - 9 2 % They offer further support for the concept that alkalic carbonatite magmas are produced by fractional crystallization.
An alternative petrogenetic scheme We propose that the alkalic carbonatite magmas known to have erupted from Oldoinyo Lengai, and possibly from other volcanoes (Dawson 1964; Hay & O'Neil 1983, Deans & Roberts 1984; Hay 1984), are of small volume and are derived from a more primitive parental olivine s6vite magma (with about 8 ~ alkalis) by fractional crystallization and gravity separation. Differentiation of this magma can follow more than one course. Development
of normal
carbonatite
The general course of differentiation, whether wet or dry, is toward increasingly alkalic compositions, and this is portrayed by the silicate minerals (principally the amphiboles). Initial magma temperatures are high (close to 1000~
91 -I
0.5
-!0 9 8 7
0.4 AI 0.:5
~
6
5 2-64 CAI
0.2 0.1 0 0
i'. "A~
5 2 " .,,v~- I J t t I I t I J 0 I 2:54 5 6 78 9
Mg/Fe FIG. 4. Compositions of amphiboles from the Argor carbonatite showing variation of A1 with Mg/Fe. Note the dramatic increase in AI as Mg/Fe falls below a value of 2. 15
0.30 0.25 0.20
7.5
CTi o
5
CTi
"
Ti 0.15 0.10 ~o
0.05 0
9
i
i
I
2
9
~ oo
i 5
I .
i
i
4
5
:o
i 6
9
~
.25
I
7
Mg/Fe
FIG. 5. Compositions of biotites from the Argor carbonatite showing variation of Ti with Mg/Fe. Note the dramatic increase in Ti as Mg/Fe fails below 2. and are similar to those of the associated silicate magmas. The alkali content of about 8 ~ and the fairly low water content are sufficient to keep it liquid at high pressure. (A study of the system calcite-dolomite with sodium and potassium carbonate at 1 kb (Beckett 1986) has established that at 8 ~ N a 2 0 the system would have extensive liquid at 900 ~ C.) Crystallization of dominantly anhydrous minerals and, initially, minerals of low alkali content (calcite, dolomite, apatite and Fe-Ti oxide) drives up both the alkali and the water contents which become increasingly concentrated in a diminishing volume of magma and progressively reduce its viscosity. At some stage the magma will become water saturated and an aqueous fluid will separate carrying with it most of the alkalis that have not already been fixed as silicates. The stage at which this occurs will be strongly dependent on the pressure and hence upon the rate at which
92
J. D. Twyman & J. Gittins
the magma has risen through the crust. Thus it is possible for separation of the aqueous phase to occur at various degrees of alkali enrichment according to the rate of magma rise. For example, a rapidly rising magma will not have sufficient time for crystal fractionation to produce a high degree of alkali enrichment before water saturation occurs and so the escaping fluid will carry off very little N a 2 0 + K 2 0 . In contrast, a magma that has risen slowly will have fractionated to an advanced degree, generating a high water content, and so its aqueous fluid will be very alkalic and a powerful fenitizing agent. The wide variations in the extent of fenitization associated with carbonatite complexes can be explained in this way. Woolley's (1982) suggestion that the relative proportion of sodic to potassic fenitization is depth dependent also fits into this scheme. In this scheme magmas of varied alkalic enrichment are possible as a result of fractional crystallization, the degree of enrichment being limited only by the stage at which water saturation develops and an aqueous fluid phase separates carrying off alkalis in excess of the amount that can be fixed as silicates. We suggest that, under plutonic conditions, the process is prevented, by continuous alkali loss, from producing a magma of the composition of the Oldoinyo Lengai lava.
Development of alkalic carbonatites How then does the extreme alkalic carbonatite magma develop? The discussion so far has focussed on two factors that serve to reduce the liquidus temperature of our proposed parental carbonatite magma: increasing water content and increasing alkali content. However, water has the further effect of removing alkalis once water saturation is reached, and so it seems unlikely that the extreme alkalic composition can be reached in a hydrous magma. We pointed out earlier that the aqueous fluid phase will control the alkali content of the magma. Furthermore, the mineralogy of the Oldoinyo Lengai lavas shows that the magma was very dry when erupted. The answer to the dilemma seems to be that the extreme alkalic magma is produced by fractional crystallization under fairly dry conditions. We have already noted that an olivine s6vite with about 8% alkalis will be largely liquid at 900~ but factors other than alkali content can be invoked to reduce the liquidus in such a magma. The Oldoinyo Lengai lava contains as much as 8% F and as much as 4% C1. F is not at all uncommon in carbonatite complexes as indicated by fluorite (see yon Eckermann (1948) on Aln6) and the extensive substitution of F for OH in many carbonatite amphiboles and micas. F
and C1, together with the alkalis, would maintain the parental carbonatite magma in a liquid state throughout its crystallization history. Increase in alkali content, through continued fractionation of alkali-free minerals and a very small percentage of alkali-bearing minerals, is able to continue uninterrupted by alkali loss since an aqueous fluid phase does not develop. A continuum of magmatic compositions from the initial 8% alkalis to the Oldoinyo-Lengai-type extreme of about 37% is possible, but the latter can be reached only if the magma is dry. In this scheme the majority of carbonatite rock types would be formed by cumulus processes as crystals sink through a very fluid magma. They are not cumulates in the passive sense of, for example, gabbro cumulates. The magmatic plumbing system in a carbonatite complex is probably much more dynamic so that the cumulates, with their lubricating interstitial liquid, are readily disturbed, re-intruded and deformed by pulses of new magma. Much of the chaotic intrusive structure of carbonatite complexes can probably be explained in this way.
Conclusions Alkalic carbonatite magma of the type erupted as lava in the Tanzanian volcano Oldoinyo Lengai is not a parental liquid from which other carbonatite magmas and rock types are derived. Rather, it is a late-stage liquid derived through fractional crystallization of an olivine s6vite magma containing about 8% N a z O + K 2 0 . In the parental-status scheme nephelinite magma divides into two immiscible liquids, of which one is ijolitic and the other is an alkalic carbonatite composition with about 500~ of superheat. During cooling, the latter continuously loses alkalis through the medium of an escaping aqueous fluid phase or by wall-rock reaction, and it changes composition until it becomes a calciticdolomitic magma which differentiates further to give rise to a wide variety of residual products. We have argued that, although the details are not brought out in the original paper, the scheme requires that the alkalic carbonatite liquid must be water saturated from the moment of its creation and that progressive loss of water and alkalis will induce crystallization rather than development of a calcitic-dolomitic magma. We have argued that the scheme is not supported by the geothermometry of carbonatite rocks or by REE studies and that mineral zonation in carbonatite amphiboles implies alkali enrichment in the magma rather than alkali depletion.
Alkalic carbonatite magmas The 500~ of superheat is a consequence of experimental design rather than a logical deduction from unconstrained experiments, and does not exist. We propose an alternative petrogenetic scheme in which the parent magma, from which carbonatite rock types are derived, is a mildly-alkalic olivine sSvite composition with about 8 ~ total alkalis. This magma probably originates by immiscible separation from an olivine nephelinite magma deep within the mantle, a proposal that makes less extreme demands on element and component partitioning than does the parentalstatus scheme. During its subsequent rise into and through the crust it undergoes fractional crystallization in which anhydrous and alkalipoor minerals sink, leaving the residual liquid progressively more enriched in alkalis and water. Water saturation is eventually reached but this can occur over a wide range of compositions and pressures governed by the rate at which the magma rises and the time available for differentiation. Rapid rise with little time for fractionation will generate saturation at a relatively low alkali content. Slow rise will generate water saturation at a much later stage and a higher alkali content because of the much higher solubility of water in alkalic carbonatite liquids than in calcitic-dolomitic liquids. The alkali content of the rocks is limited by the amount that can be fixed as silicates and this in turn is limited by the activities of Si and AI, both of which are
93
very low. Separation of an aqueous fluid phase when water saturation is reached controls the alkali content of the remaining magma, and is itself controlled by the solubility of alkali carbonates and halides in supercritical water and by the rate at which anhydrous alkali-poor minerals continue to crystallize. Thus in most plutonic settings the crystallizing magma will lose alkalis at some stage when water saturation is reached; this may be a longcontinued process providing a source of fenitizing fluids. The pre- and post-saturation history of the magma will probably control the N a : K ratio of the fluid and have a profound bearing on the type of fenitization that ensues. A major difference from the parental-status scheme is that the water content of the magma gradually increases through fractional crystallization until saturation is reached, rather than escaping continuously from an already alkalic liquid that is merely cooling but not crystallizing. If the parental olivine s6vite magma has exceptionally low water and high halogen contents, alkali enrichment might reach a very advanced stage and even develop the extreme composition of the Oldoinyo Lengai lava. This will be possible because the magma does not reach water saturation and so alkalis are able to accumulate continuously without being abstracted by an aqueous fluid. Alkalic carbonatite magmas are late derivatives and not primary liquids.
References BECKETT, M. F. 1986. Phase relations in alkali-bearing dolomite carbonatites; effect of alkalinity and fluorine content on the solubility of pyrochlore in carbonatite magma. M.Sc Thesis, University of Toronto, Canada (unpublished). COOPER, A. F., GITTINS,J. & TUTTLE,O. F. 1975. The system NazCO3-KzCO3-CaCO 3 at 1 kilobar and its significance in carbonatite petrogenesis. Am. J. Sci. 275, 534-60. DawsoN, J. B. 1964. Reactivity of the cations in carbonatite magmas. Geol. Assoc. Can. Proc. 15, 103-13. DEANS, T. & ROBERTS,B. 1984. Carbonatite tufts and lava clasts of the Tinderet foothills, western Kenya: a study of calcified natrocarbonatites. J. geol. Soc. Lond. 141, 563-80. VONECKERMANN,H. 1948. The alkaline district of Aln5 Island. Sver. geol. Unders., Ser. C, 36. FREESTONE,I. C. & HAMILTON,D. L. 1980. The role of liquid immiscibility in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. GERASIMOVSKY,V. I., ALASHOW,Yu. A. & KARPUSHINA, V. A. 1972. Geochemistry of the rare earth
elements in the extrusive rocks of the East African rift zones. Geochem. Int. 1972, 305-19. GITTINS, J. 1979. Problems inherent in the application of calcite-dolomite geothermometry to carbonatites. Contrib. Mineral. Petrol. 69, 1-4. - - , ALLEN, C. R. & COOPER,A. F. 1975. Phlogopitization of pyroxenite; its bearing on the composition of carbonatite magmas. Geol. Mag. 112, 5037. GOLD, D. P. 1966. The average and typical chemical composition ofcarbonatites. Proc. 4th Gen. Meeting Inter. Mineral. Assoc., 1964, India, 83-9. GUEST, N. J. 1956. The volcanic activity of Oldoinyo L'Engai, 1954. Tanganyika geol. Surv. Rec. 1954, 4, 58-9. HAY, R. L. 1984. Natrocarbonatite tephra of the volcano Kerimasi, northern Tanzania. Geology, 11, 599-602. -& O'NEm, J. R. 1983. Carbonatite tufts in the Laetolil Beds of Tanzania and the Kaiserstuhl in Germany. Contrib. Mineral. Petrol. 82, 403-6. LE BAS, M. J. 1981. Carbonatite magmas. Mineral. Mag. 44, 133-40.
94
J. D. Twyman & J. Gittins
PHILPOTTS, J. A., SCHNETZLER,C. C. & THOMAS,H. H. 1972. Petrogenetic implications of some new geochemical data on eclogitic and ultrabasic inclusions. Geochim. cosmochim. Acta, 36, 113166. SECHER, K. & LARSEN, L. M. 1980. Geology and mineralogy of the Sarfartoq Carbonatite Complex, southern West Greenland. Lithos, 13, 199-212. TWVMAN, J. 1983. The generation, crystallization, and differentiation of carbonatite magmas: evidence from the Argor and Cargill complexes, Ontario. PhD Thesis, University of Toronto (unpublished).
WENDLANDT, R. F. & HARRISON, W. J. 1979. Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapor: results and implications for the formation of light rare earthenriched rocks. Contrib. Mineral. Petrol. 69, 40419. WOOLLEY, A. R. 1982. A discussion of carbonatite evolution and nomenclature, and the generation of sodic and potassic fenites. Mineral. Mag. 46, 13-7. WRIGHT, T. L. & DOHERTY, P. C. 1970. A linear programming and least squares computer method for solving petrologic mixing problems. Bull. Geol. Soc. Am. 81, 1995-2008.
JAMES D. TWYMAN* & JOHN GITTINS, Department of Geology, University of Toronto, Toronto M5S 1A1, Canada. * Present address: Arco Oil and Gas Company, Lafayette, LA, U.S.A.
The kimberlite clan: relationship with olivine and leucite lamproites, and inferences for upper-mantle metasomatism J. B. Dawson S U M M A R Y : Kimberlites are rare ultrabasic potassic low-volume melts that originate in the diamond stability field of the upper mantle. They are petrographically complex, and wide
mineralogical and chemical variations suggest that they should be regarded as a group or clan rather than a single narrowly-defined rock type. In S African kimberlites there is a major division between highly-micaceous (Group II) kimberlites that are mineralogically, chemically and isotopically distinct from poorly-micaceous (Group I) kimberlites. Nd and Sr isotope studies of diamondiferous olivine and leucite lamproites from NW Australia indicate that they are isotopically similar to Group II kimberlites but are distinct from Group I. New bulk rock analyses of Group II kimberlites indicate closer chemical similarities with olivine lamproites than previously supposed. and carbonatisation. Kimberlite commonly contains inclusions of upper mantle derived ultramafic rocks. Variable quantities of crusKimberlites are rare ultrabasic potassic igneous tal xenoliths and xenocrysts may also be rocks occupying small vents, sills and dykes. present. Kimberlite may contain diamond but They are petrographically complex because of only as a very rare constituent.' variations in texture (diatreme facies versus The number of qualifications (my italics) in this hypabyssal facies (Dawson & Hawthorne 1970)), definition serves to illustrate the petrographic variability and particularly the range of mineral variations in xenolith and xenocryst content, variable content and proportions of macrocrysts, species developed in the matrix. The mineralogy and variations in the mineralogy of the fine- is also complicated by the fact that some species grained matrix, all of which combine to give a (e.g. calcite and apatite) that are relatively wide range of chemical composition (Dawson common in hypabyssal-facies kimberlites have 1967, 1980). These variations, particularly in not developed to the same extent in diatrememineralogy, have created difficulties in defining facies kimberlites, presumably owing to volatile kimberlite, as exemplified by the recent definition l o s s during diatreme formation; also, certain of kimberlite by Clement et al. (1984): species such as monticellite are extremely suscep'Kimberlite is a volatile-rich potassic, ultratible to calcitization under high Pco2 conditions. basic igneous rock which occurs in small Despite these complications it can be seen that volcanic pipes, dykes and sills. It has a individual kimberlites, although not necessarily distinctively inequigranular texture resulting containing identical matrix minerals, do possess from the presence of macrocrysts set in a some permutation of the minerals listed by finer-grained matrix. This matrix contains, as Clement et al. (1984), and their similar and prominent primary phenocrystal and/or overlapping mineralogies suggest that they should groundmass constituents, olivine and several be regarded as a group or clan of closely-related of the following minerals: phlogopite, carbonrock types rather than a single tightly-defined ate (commonly calcite), serpentine, clinopyrock type. roxene (commonly diopside), monticellite, There is nonetheless a fundamental division apatite, spinels, perovskite and ilmenite. The within the kimberlite clan, recognized as such by macrocrysts are anhedral, mantle-derived, Wagner (1914), between 'lamprophyric' kimberferromagnesian minerals which include olilites having a highly micaceous matrix, and the vine, phlogopite, picroilmenite, chromian spiso-called 'basaltic' variety containing little or no nel, magnesian garnet, clinopyroxene mica in the matrix. (The term 'basaltic' is quite (commonly chromium diopside), and orthopyinappropriate and Dawson (1980) has proposed roxene (commonly enstatite). Olivine is exthat it should be dropped.) In addition to the tremely abundant relative to the other major difference in the phlogopite content of the macrocysts, all of which are not necessarily matrices of these two types of kimberlite, there present. The macrocrysts and relatively-earlyare certain other differences. formed matrix minerals are commonly altered 1 In the case of S Africa, the micaceous by deuteric processes, mainly serpentinisation
Varieties of kimberlite
From:FITTON, J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 95-101.
95
J. B. Dawson
96
kimberlites were intruded 120-200 Ma ago which contrasts with the non-micaceous varieties intruded 80-90 Ma ago. 2 The micaceous kimberlites tend to contain considerable amounts of matrix calcite and apatite. 3 The micaceous kimberlites generally also contain groundmass diopside, which is quite rare in the non-micaceous types (Dawson et al. 1977). 4 Although macrocrystal picro-ilmenite is common in the non-micaceous kimberlite, it is relatively rare in micaceous varieties. 5 The micaceous kimberlites are occasionally accompanied by other high-K minor intrusives, as at the Helam Mine in the Swartruggens area of the western Transvaal where an E-W-trending dyke-swarm contains both diamondiferous kimberlite and non-diamondiferous sanidine-nepheline-lamprophyre (the so-called 'Male' dyke) (Skinner & Scott Smith 1979). Another major difference between the micaceous and non-micaceous kimberlites in S Africa has recently been recognized on isotopic grounds (Smith 1983). The two groups are quite distinct by reason of their U, Pb, Sm, Nd and Sr contents
and isotopic ratios, which are interpreted as reflecting substantial chemical differences in the mantle source rocks for the two groups. The nonmicaceous kimberlites (denoted Group I by Smith) are derived from undifferentiated to slightly depleted sources relative to bulk earth (Sr and Nd systematics) but with high U/Pb ratios. In contrast, the micaceous kimberlites (referred to as Group II by Smith) are derived from an enriched source with high Rb/Sr and Nd/Sm ratios but with low U/Pb ratios; it was suggested by Smith that this enrichment event took place in excess of 103 Ma ago. The isotopic differences between the Group I and Group II kimberlites are well illustrated on a Sr versus Nd isotope plot (Fig. 1) on which the Helam Mine lamprophyre dyke, referred to above, falls close to the micaceous kimberlites. In summary the term 'kimberlite' in S Africa has been used to cover rocks that are distinct petrographically and is 9 and which, furthermore, were intruded at different times. Their common factors are that they are both products of low-volume ultrabasic upper-mantle melts originating in the diamond stability field of the Earth's upper mantle in the late-Jurassic~Sr
-50
0
50
100
150
200
I
]
I
I
i
i
I~MORB I
[] Ohvme " ' lamproltes "
\ ' ~ \ DEPLETED MANTLE (DM) '~'~
9 Leucite lamproites
~ ,
~" Helam lamprophyre
" ..... . a/ ,J,h oc~' K~.~ ~3". ~ o ~"N~. 9'BULTFONTEIN / \ " 1.5 mol.% includes WKB, ARG, BOB, SB, MBAY, GSB, HP and HOL. Interestingly, the NWI, COR, PEN, LH and MAP suites also share the feature of occurring in a young palaeoorogenic zone and are most probably depleted in TiO2 (as well as Nb etc.) because of a TiOzdepleted source, a feature shared by orogenic andesites and related rocks. The extreme enrichments in TiO 2 displayed by the WKB, CHE, BOB and SB suites must represent TiOz-enriched source rocks.
Normative composition The average CIPW normative compositions for most lamproite suites are given in Table 7, bearing in mind the limitations of the CIPW normative calculations in describing phlogopiteand amphibole-rich rocks. The average lamproite consists of nearly equal amounts (by weight) of salic and femic normative minerals. Almost half (43%) of the 295 lamproites that have been included in this data set are quartz normative; in fact, all suites for which more than five rocks have been analysed contain at least one quartznormative rock. Only eight of the 21 suites contain rocks with normative leucite, and normative K-metasilicate is present in small amounts because of the Al-depleted character of lamproites. All suites but two contain at least a few rocks with normative feldspathoids. Lamproites as a group are characterized by significant normative orthoclase (16-59 wt.%, average 35 wt.%) and femic minerals (33-81 wt.%, average 51 wt.%), but relatively low normative albite (012 wt.%, average 5 wt.%) and anorthite (0-6 wt.%, average 1 wt.%) relative to nearly all other igneous rocks. As with most peralkaline rocks, almost all lamproites contain normative acmite (average 2.5 wt.%). Compared with minettes, lamproites contain similar average amounts of normative quartz, orthoclase and nepheline but significantly lower albite and anorthite, slightly lower clinopyroxene, and slightly higher orthopyroxene, acmite, olivine and leucite. Lamproites as a group are somewhat more enriched in femic minerals than minettes. Despite the common occurrence of m o d a l leucite in lamproites, many lamproite magmas tend to differentiate towards quartz-normative residua. This is based on the natural evidence of
I49
the normative compositions of interstitial glasses and whole-rock samples of glassy lamproites (e.g. GSB (Sheraton & Cundari 1980), LH (Kuehner et al. 1981) and WKB (Wade & Prider 1940; Prider 1982)). In fact many leucite-bearing lamproites are SiO 2 saturated or oversaturated. This differentiation trend can be explained by a small amount of phlogopite fractionation and additionally by the experimental work of Luth (1967) in the system kalsilite-forsterite-silicawater which demonstrated that the phlogopite liquidus surface bridges the forsterite-kalsilite thermal divide at PHzo > 1 kb.
Trace-element chemistry A summary of the trace-element geochemistry of lamproites, compared with that of kimberlites, lamprophyres and alkali basalts, is given in Table 6 and Fig. 23(a). All trace elements show an extreme degree of variance with coefficients of variation typically more than 50% and often more than 100%. If these extreme variations are ignored, some general conclusions can be drawn. On average, lamproites have the highest incompatible-element contents (e.g. Rb, Sr, Ba, U, Zr and rare-earth elements (REE)) and alkali basalts the lowest, whereas kimberlites have the highest compatible-element contents (e.g. Ni, Cr, Co, Sc, Zn and Cu) and alkali basalts generally the lowest. Lamproites are more enriched in both compatible and incompatible elements compared with other ultrapotassic rocks. With respect to the average K/Rb ratios, the various rock groups are ranked as follows: alkali basalts (350), nondiamondiferous lamproites (220), calc-alkaline and alkaline lamprophyres (180), kimberlites (160) and diamondiferous lamproites (85). The average REE contents of lamproites, lamprophyres and kimberlites all overlap to some degree (Table 6 and Fig. 23(a)), and all are markedly light REE enriched and heavy REE depleted compared with alkali basalts. Although a large degree of overlap exists, average lamproite REE contents generally exceed those of kimberlites. The average ratios of light REE to heavy REE (La/Lu)on are ranked as follows: lamproite (100), kimberlite (90), ultramafic lamprophyre (47), alkaline and calc-alkaline lamprophyre (27) and alkali basalt (11), showing that lamproites contain the smallest quantities of heavy REE (relative to light REE) compared with these other rock types. Since the degree of elemental variation for these rock types is so great, overlap is generally observed in comparing the trace-element contents of the various rock groups. Nevertheless, the
S. C. Bergman
15 ~
2.0]
t
DIAMONDIFEROUS SUITES 9 AVE, KIMBERLITE (550) 9 AVE. LAMPROPHYRE (800) ",9 AVE. ALKALI BASALT (4230)
WKB
1.0
"PRA
o
1'o
1
~
(o)
i
30
,o
K20 1 Na20 (moles)
15 -] I / / | / ,
--.,"-9 9 ,,~ f-I O
K20 1 AI203 = 2 J J
10 1
W
K
B
~
DIAMONDIFEROUS SUITES AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) AVE. LAMPROITE AMPHIBOLE (97)
K20/AI203 = 1
=
0
5
(b)
10 AI203 (mole %)
15.
I
DIAMONDIFEROUS SUITES 9 AVE. KIMBERLITE (550) 9 AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) I-I AVE. LAMPROITE PHLOGOPITE (229) O AVE. LAMPROITE AMPHIBOLE (97)
.:..< .;-..-----
g
PRA
;o ( C )
1'5
CaO (mole %)
COR SB HP KAM GSB MBAY PEN NWl PPN MAP HOL LH ENO CHE NSW
. . . . . . . . . . .
............
............ ...............
.~'/-...--t/.-~q,~ -~',--~,'l ---J, .v~,~vvVV~NV.
Lamproites and K-rich igneous rocks gO-
,
151
ARG
PRA
9 .
~ Z BOB
80-
o r,
70-
E 3 l=
eo-
50-
40-
9 9 ~, I-1 O
DIAMONDIFEROUS AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE, ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) AVE. LAMPROITE AMPHIBOLE (97)
;o
3O
(d)
WKBf
"::. . . . . " .... . .....-'/ - ......... .--
~
;o
;o
SIO 2 (mole %)
109 9 zx 17
WKB
DIAMONDIFEROUS SUITES AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) E. LAMPROITE AMPHIBOLE (97)
COR SB HP YAM GSB MBAY PEN NWI PPN MAP HOL LH ENO CHE NSW
~
~
..............
-----~ ~ - - - r ,'----r ---.-.-"~r
_o
'...
"~
;
,~ _
..-~---~
>-~--53
0
( e )
K20 (mole %)
FIG. 22. Plots illustrating the compositional space occupied by various lamproite suites, with average rock and mineral compositions plotted for comparison (see text and Tables 2 and 3 for abbreviations): (a) KzO/A1203 versus K20/Na20; (b) K20 versus A1203; (c) K20 versus CaO; (d) Mg number versus SiOz; (e) TiO2 versus K20.
trace elements that are most useful in distinguishing lamproites from kimberlites are as follows: (a) Rb, Sr, Ba, Zr and the K/Rb ratio; (b) Ni, Cr, Co and the Ni/Cr ratio. Kimberlites generally have Rb < 150 ppm, Sr < 1200 ppm, Ba < 3000 ppm, Z r < 8 0 0 ppm and K/Rb 150 ppm, Sr> 1200 ppm, Ba > 3000 ppm, Zr > 800 ppm and K/ R b > 150. With respect to the compatible elements, kimberlites generally have Ni > 500 ppm, Cr>800 ppm, Co>50 ppm and Ni/Cr>0.7, whereas lamproites generally have Ni < 500 ppm, Cr 2 6 kb and Pi_i2o= Ptotal" Madupites probably represent partial melts of phlogopite-pyroxenite or phlogopite-olivinepyroxenite assemblages (Barton & Hamilton 1979). Barton & Hamilton's experiments indicate that the dominant liquidus or near-liquidus phases are leucite, olivine, orthopyroxene, clinopyroxene and garnet at P < 3 0 kb. They also concluded that the peralkalinity of these ultrapotassic magmas could reflect either primary source rock compositions or the selective melting of phlogopite + pyroxene. They postulated that the association of low-SiO: madupites with highSiO2 orendites at LH results from variations in H20 and CO2 and local mineral assemblages in the upper mantle source region. It should be noted that A. Edgar (personal communication, 1985) questions the validity of the phase relationships determined by Barton because of Fe partitioning in platinum capsules. Sobolev et al.
L a m p r o i t e s a n d K-rich igneous rocks (1975) studied the low-P near-liquidus phase equilibria of a LH wyomingite and found that diopside is the sole liquidus phase at temperatures around 1320~ and P = 1 b despite the fact that phlogopite phenocrysts are common and that the 1 b solidus temperature is about 1000~ Arima & Edgar (1983) studied the hydrous (+ COO liquidus and sub-liquidus phase relations of a wolgidite (from the WKB, Mount North) to 40 kb. They found that olivine (at P < 15-24 kb) and orthopyroxene (P > 15-24 kb) occur on the liquidus; phlogopite, rutile, clinopyroxene, armalcolite and priderite follow at lower P and T. Rutile reacts with phlogopite and liquid to produce priderite at P < 15 kb and T < 1010~ and armalcolite reacts to form priderite at T< 1010~ and P < 15 kb; rutile is the high-pressure phase and armalcolite is the high-temperature phase. These phases (priderite, armalcolite and rutile) were only found in runs containing added H20 (13 wt.%) and were absent from runs with low H20 contents (3 wt.~). Their results indicate that wolgidite magma most probably represents the partial melt of a source mantle containing phlogopite, rutile, olivine and orthopyroxene. It is unlikely that a wolgidite-composition melt would be derived from the partial melting of a simple garnet- or spinel-bearing lherzolite mantle, in contrast with the derivation of typical alkali basalts from such a mantle. Other ultrapotassic rock compositions
Edgar et al. (1976, 1979), Edgar (1979), Edgar & Arima (1983), Ryabchikov & Green (1978), Arima & Edgar (1983) and Edgar & Condliffe (1978) investigated the phase behaviour of several TAN rocks (biotite ugandite, biotite mafurite and katungite) in the presence of H20 + CO2 to 40 kb. They found that these melts never equilibrate with an orthopyroxene- or garnet-bearing assemblage and that only diopside, olivine, ilmenite and phlogopite occur on the liquidus, even at high pressure. These three magmas cannot be related to each other by fractionation alone or by partial melting of a single mantle source region. The three magma compositions could be derived by partial melting, at various depths, of clinopyroxenite or peridotite sources with varying degrees of enrichment in K where low H 2 0 / C O 2 ratios produce low enrichments in K. Their experimental findings favour metasomatism of a peridotitic mantle source as a prerequisite for the production of K-rich magmas. Ryabchikov & Green (1978), however, concluded that biotite mafurite, olivine leucitite and ugandite liquids could be produced by the partial melting in the presence of H20 and CO2 of a
I65
lherzolite source mantle locally enriched in phlogopite. Phlogopite occurs on the liquidus at high H20/CO 2 ratios (Xco2=0-0.25), but becomes unstable at higher CO2 contents (Xco2 > 0.25). Cundari & O'Hara (1976) studied the phase equilibria of a leucitite from NSW under anhydrous conditions to 40 kb and determined that garnet did not exist with enstatite above 35 kb, indicating that orthopyroxene did not occur in an anhydrous source peridotite. Thompson (1977) experimentally studied a clinopyroxene leucitite from the RP and found that the phenocryst assemblage could be duplicated at 14 kb and 1260~ indicating a mantle derivation. He postulated that the experimental data were consistent with the view that the apparent crustal contamination (suggested by SrO isotope data) was due to the partial fusion of subducted ocean-floor sediments within the upper mantle. Other experimental work on potassic and ultrapotassic rocks includes that of Dolfi et al. (1976), Lloyd et al. (1985), and Esperanca & Holloway (1985). Compositions in the system N a 2 0 - K 2 0 - M g O AI203-SiO2-H20-CO 2
Schairer & Bowen (1938) studied the anhydrous system KA1Si206-CaMgSi206-SiO2 at 1 atm and established the phase equilibria pertinent to Fe-free K-rich alkaline magmas. They found an extremely large leucite field in the leucitediopside-silica ternary system. However, more recent work on the hydrous analogue of this system (Ruddock & Hamilton 1978a) demonstrates that the leucite field is extremely compressed, which explains the general absence of leucite in minettes. Luth (1967) determined the phase equilibria in the system MgzSiO 4KA1SiO4-SiOz-H20. Bravo & O'Hara (1975) studied the partial melting of a synthetic phlogopite-bearing garnet and spinel lherzolite (CO2 free) and found that partial melts of these solids at 15 and 30 kb had extremely low K20 contents (1-4 wt.~) and were quartz- and/or hypersthene-normative, These liquids share an SiO2-rich compositional feature with lamproites, although they are relatively depleted in K20 relative to lamproites. Modreski & Boettcher (1972, 1973) investigated the stability of phlogopite in a model system and found that it is unstable under oceanic geothermal conditions of P > 25 kb but may persist in sub-continental geothermal conditions to P > 50 kb. Wendlandt (1977a, b) and Wendlandt & Eggler (1980a, b, c) investigated the stability and melting behaviour of sanidine and phlogopite in olivine-
I66
S. C. Bergman
bearing assemblages. They found that sanidine + forsterite breaks down to kalsilite+enstatite along a P-Tlocus including 20 kb at 1000~ and 30 kb at 1300~ indicating that the latter assemblage will be stable along most geotherms (Wendlandt & Eggler 1980c). The melting behaviour of phlogopite in peridotite was found to be controlled by the equilibrium mineral assemblage as well as by the composition of the volatile phase (H20 versus CO2). At P < 1 4 kb the liquid compositions were quartz or enstatite normative, but became leucite to kalsilite normative at higher pressures in the presence of CO2 + H20 vapour. At pressures above 30 kb in the presence of magnesite, phlogopite peridotite melts to produce carbonatitic liquids which become progressively enriched in K20 and SiO2 as pressure increases. Phlogopite is not stable in a peridotite assemblage above 50 kb. Ryabchikov & Boettcher (1981) investigated the solubility of K in an aqueous fluid in equilibrium with phlogopite-pyrope-forsterite gels at 10-30 kb and 1050-1100~ and found that water is capable of dissolving 4-25% K20 and that solute KzO/AI203 ratios are close to unity. Ruddock & Hamilton (1978a) experimentally studied the system leucite-diopside-quartz-H20 to 4 kb and confirmed that, under hydrous conditions, diopside and phlogopite will be the first two liquidus phases followed by sanidine and quartz. This study nicely explains the petrography of minettes and other lamprophyres (however, see Rock (1984, Fig. 8) for an alternative interpretation). Ruddock & Hamilton (1978b) determined the stability of carbonate in an ultrapotassic mafic assemblage and found that a minette composition liquid is saturated in olivine, garnet, orthopyroxene and clinopyroxene at pressures above 15-20 kb and is saturated in carbonate at pressures above 20-30 kb. Schneider (1982) and Schneider & Eggler (1984) determined the high P - T (15-20 kb and 750-1100~ solubility of major-element oxides in H20 and H20-CO2 fluids in equilibrium with jadeite-, amphibole- and phlogopite-peridotite and several individual minerals. They found that the fluid's solute contents were in the range 0.518 wt.% and increased with decreasing CO2 content. H20 or H20 + CO2 fluids in equilibrium with phlogopite-peridotite were found to be peraluminous whereas H20 at- CO2 fluids in equilibrium with amphibole- or jadeite-peridotite ranged from mildly peraluminous to strongly peralkaline in their solute compositions. Solute K/Na ratios were found to be controlled by H20 / CO2 as well as by P, T and the bulk composition of the solid, whereby K/Na increases with increasing CO2/H20 ratio of the fluid.
Petrogenesis A wide variety of models have been advocated for the evolution of K-rich rocks; the reader is referred to previous reviews by Bell & Powell (1969), Sahama (1974) and Gupta & Yagi (1980). Most hypotheses can be placed in one of three broad groups. 1 Partial melting of mantle material that was metasomatically (or otherwise, e.g. by plutonism) enriched in phlogopite and other LIL-elementenriched minor phases (e.g. apatite, zircon, sphene etc.); the melt may be subject to variable amounts of fractionation but does not undergo substantial assimilation of crustal material (Waters 1955; Yagi & Matsumoto 1966; Kay & Gast 1973; Boettcher et al. 1975; Beswick 1976; Gupta et al. 1976; Van Kooten 1980). 2 Assimilation of continental crustal material by 'normal' mantle-derived alkali-basaltic, carbonatitic or alkali-ultramafic melts (Daly 1910, 1933; Shand 1931; Rittman 1933; Larsen 1940; Williams 1936; Holmes 1950; Powell & Bell 1970; Taylor & Turi 1976; Turi & Taylor 1976; Taylor et al. 1984). 3 Evolution involving processes such as zone refining, fractional resorption, gaseous transport etc. in otherwise typical mantle-derived melts (Bowen 1928; Saether 1950; Kennedy 1955; Harris 1957; Prider 1960; Marinelli & Mittempergher 1966; Fuster et al. 1967; Kogarko 1980; Kogarko et al. 1968; Kushiro & Aoki 1968; Harris & Middlemost 1969; Stewart 1979; Wones 1979; Ryabchikov et al. 1982). Sahama (1974) observed that researchers tend to desire hypotheses that attempt to explain the petrogenesis of K-rich rocks in general; however, the geochemical distinctions between lamproites and kamafugitic as well as other K-rich rocks do not permit such a luxury. The tectonic-geological and geochemical distinctions between lamproites, kamafugites, shoshonites and other Krich rocks clearly support this view. The last statement in Sahama's (1974)review of K-rich rocks speculates on a relationship between kimberlites, lamproites and kamafug o ites. Gupta & Yagi (1980) additionally recognized the importance of this speculation. The last 10 years have witnessed major discoveries of diamondiferous lamproites and other developments which make this speculation a reasonable, if not undeniable, view. Although kimberlites and olivine lamproites evolve differently, they possess many similarities which have led several workers to suggest a petrogenetic model for lamproites
Lamproites and K-rich igneous rocks whereby kimberlite magma is differentiated to produce a more potassic SiO2-rich lamproitekamafugite magma (Holmes 1932; Wade & Prider 1940; Scott 1981). Although more data are required to establish an unambiguous petrogenetic picture for lamproites, the data now available make it possible to constrain their petrogenesis to a much greater degree than was possible as little as 10 years ago. In addition, many aspects of the petrogenesis of lamproites have important implications for a variety of aspects of mantle evolution and magma genesis. Lamproites are mantle-derived melts and do not simply result from the melting of recycled continental crust. This is supported by (a) the presence of diamonds in olivine lamproites, (b) the presence of mantle-derived peridotite xenoliths (albeit much rarer than their abundance in alkali basalts and kimberlites) in leucite, phlogopite and olivine lamproites, (c) a primitive (Mg number 74 + 9) mafic-ultramafic major-element composition, (d) an enrichment in the refractory trace elements Co, Cr, Ni and Sc, and (e) the mantle-type pressures and temperatures (P = 2030 kb and T= 1100-1300~ calculated by performing thermodynamic isoactivity calculations (Nicholls & Carmichael 1972; Carmichael et al. 1974, 1977; Barton & Wood 1976) in which lamproite magmas are equilibrated with various source-mantle assemblages. In contrast with being enriched in LIL elements, lamproite magmas possess major-element compositional features typical of partial melts of a refractory (i.e. Mg number, 88-92, diopsidedepleted, harzburgite ?) mantle : they are depleted in the incompatible elements Na, A1 and Ca compared with alkali basalts and other primary melts considered to represent low degrees (less than 10%) of partial melting of a lherzolitecomposition mantle. A depleted diopside-poor source is further supported by the high-P-T experimental studies on lamproite compositions which suggest a non-lherzolite mantle source for lamproite magmas (see above) and also by the characteristically refractory harzburgite and dunite xenolith population that occurs rarely in several lamproite suites (e.g. WKB, LH, MAP). However, there are other ways of interpreting this depletion in certain major elements. Lamproites are phlogopite-rich rocks and many whole-rock analyses of lamproites approach but are more calcic and siliceous than the bulk composition ofphlogopite. Therefore it is possible that lamproites represent the selective fractional fusion of mantle phlogopite and other LILelement enriched phases (e.g. apatite, zircon etc.) diluted by small amounts of peridotite compo-
167
nents (clinopyroxene, orthopyroxene and olivine). Since phlogopite is severely depleted in Na and Ca and moderately depleted in AI compared with basaltic melts, this model nicely explains the bulk chemistry of lamproites. It is further supported by the fact that phlogopite is not expected to survive extensive amounts of partial melting, most of it being consumed at extremely low degrees of melting. The ultrapotassic but variable K/Na ratios of lamproites can be interpreted in a number of ways. K can be enriched relative to Na as a result of (a) source-mantle characteristics, i.e. precipitation of phlogopite (Na absent) via metasomatism in a Na-poor refractory peridotite, (b) fractionation of high K/Na phases (e.g. phlogopite, leucite) and (c) selective volatile-phase transfer of K relative to Na. The experimental work by Schneider & Eggler (1984) and Schneider (1982) contributes important data with which to judge the effectiveness of the latter process, i.e. K/Na in the fluid varies with CO2/H20. However, much more work on it is required to understand vapour-phase transfer of alkalis properly. Since this feature is a local as well as a general feature of lamproites, it most probably results from a combination of all these processes (see discussion in Sahama (1974, p. 106).) Like alkali basalts and lamprophyres, the volatile budgets of leucite-lamproite and phlogopite-lamproite magmas are water-dominated (relative to CO2), whereas those of kimberlites (and olivine-lamproites) possess higher and subequal H20 and CO2 contents. The H20 contents of kimberlites indicated in Tables 4 and 5 are maximum values for the magmas, and the true values are most probably a factor of 2 lower owing to the ubiquitous and often secondary serpentinization that characterizes kimberlite groundmass and xenocrysts. As mentioned above this HzO-rich character provides an explanation for the sherbet-glass shape of lamproite pipes. It is also consistent with the SiO2-rich and variable composition of lamproites. Experimental work by Mysen & Boettcher (1975a, b) and Eggler (1977) has shown that partial melting of a peridotite with a high H20/(CO2 + H 2 0 ) ratio will produce a relatively SiO2-rich melt, eliminating the need to invoke fractionation of low-SiOz mafic minerals to produce SiO2-rich lamproites. Slight variations in the source H 2 0 / ( C O 2 -k-H 2 0 ) ratio would produce variations in the SiO2 contents of partial melts (e.g. Bachinski & Scott 1979). Therefore, whereas kimberlites and alkali olivine basalts are SiO 2 undersaturated and are most probably derived from a mantle source with a low H20/(CO 2 -k-H20) ratio, lamproite sources represent the complementary volatile situation
I68
S. C. B e r g m a n
(high H20/(CO 2+H20)) in addition to other contrasts, i.e. LIL-element enrichments. Of all the alkaline rocks, lamproites are consistently among the most enriched in LILelements. They are also characterized by enrichments in higher-series elements relative to lowerseries elements in the same group (e.g. lower K/ Rb and Sr/Ba) compared with virtually all other rock types. It is now clear that this feature is a mantle phenomenon and does not require crustal contamination for its explanation. The highly variable concentrations of LIL-elements in lamproites is easily explained by a heterogeneous distribution of LIL-element-enriched phases in the source mantle in addition to volatile-phase transfer processes acting during magma ascent. These minerals, most importantly phlogopite but also including other LIL-element-enriched phases such as apatite, monazite, sphene, rutile and zircon, originated from the process of mantle metasomatism, most recently reviewed by Bailey (1982) and Dawson (1984). These phases could be precipitated or otherwise produced from any number of sources, including (a) fluids ascending from a down-going slab which hybridize the overlying mantle wedge, (b) fluids emanating from an ascending plume, (c) 'background' deepmantle fluids ascending throughout the upper mantle and (d) silicate melts of variable origins. As mentioned above, this metasomatism must have been fairly old (10-103 Ma) to generate the 87Sr/86Sr and 143Nd/144Nd ratios of lamproites. The age of this lamproite-source metasomatism contrasts with that suggested for the source regions of alkali basalts and related melts. Since H20 is suggested as the most important 'lamproite volatile', the mantle metasomatism associated with lamproite source regions may be distinct from that associated with alkali basalts (CO2-rich?). If we accept the view that olivine lamproites and leucite lamproites (or phlogopite lamproites) form a cognate petrological continuum, a view that is supported by the time-space-composition relationships present within the WKB and PRA suites, several implications arise. The compositional variations in these suites indicate the existence of differentiation processes more complex and less understood than simple crystal fractionation. Also, the preservation of diamonds within the WKB and PRA suite diatremes indicates that these magmas explosively ascend to the surface from depths of greater than 150 km, i.e. within the diamond stability field in the upper mantle because lamproite melts, which are extremely hydrous, would otherwise oxidize the diamonds. Although this feature is shared by kimberlites, many lamproite diamonds display a
greater degree of graphitization and rounding than those from kimberlites; these features may indicate that lamproite magmas ascend more slowly than kimberlites. Much more work on the stability of diamond in lamproite magmas is required to acquire a better understanding of the ascent characteristics of lamproites. Another implication resulting from the mantle derivation of lamproites involves the radiogenic isotope systems Sm-Nd and Rb-Sr. Some researchers have recently interpreted the isotopic data of K-rich rocks as indicating the operation of continental crust contamination (e.g. TAN (Vollmer & Norry 1983b)) or the mixing of a depleted mantle component (similar to MORB: eNa = + 10 and eSr 40) with an enriched mantle component (eNd= -- 16 and ~Sr= + 240, e.g. WKB (McCulloch et al. 1983a, b)). However, the isotopic data can also be interpreted in the context of a single-mantle reservoir that is heterogeneous but enriched in its Rb/Sr and Nd/ Sm ratios relative to the bulk Earth. The heterogeneities could be related to partial melting or plutonic processes. Small differences in these ratios can produce the time-integrated range in 143Nd/l*4Nd and 87Sr/86Sr ratios observed for lamproites. Regardless of the specific model used to interpret the isotopic data (as all are nonunique), an enriched mantle source is required by all. This 'enriched' source is also represented by Group II kimberlites (Smith 1983a, b) and the diopsides from kimberlite xenoliths (Menzies & Murthy 1980a, b; Basu & Tatsumoto 1980). However, the major-element chemistry of lamproites is characterized by a depletion in some of the major elements that are the first to enter partial melts (Ca, AI and Na) and an enrichment in refractory elements (Mg and Ni) so that the source probably experienced a partial-melting event prior to the enrichment in Nd/Sm and Rb/ Sr (see above). When this is viewed in the light of the Pb isotope data obtained by Fraser et al. (1985) and Nelson et al. (1986) (see above), it is seen that lamproite mantle sources must have been subjected to depletion events followed by enrichment events a long time ago ((1-3)x 103 Ma). This model has interesting implications for Earth history because not only does it suggest the enriched complement to the depleted mantle source represented by MORB and nearly all alkali basalts (except suites such as Kerguelen and Tristan da Cunha), it also supports the existence of Precambrian partial-melting episodes in the mantle and suggests a three-stage model of petrogenesis. If the more recent lamproite suites that occur overlying fossil Benioff zones indicate the general tectonic environment of lamproites, then it is possible that lamproites represent =
- -
Lamproites and K-rich igneous rocks partial melts of sub-continental lithosphereasthenosphere that was (a) initially depleted by the formation of MORB at a spreading centre, (b) moved beneath a continent and (c) later hybridized perhaps by fluids and melts emanating from a down-going slab. All this must have taken place sufficiently long ago to permit the development of time-integrated high 87Sr/86Sr and low 143Nd/14aNd ratios. The most recent melting event that led to the surface emplacement of lamproite magmas could be triggered or enhanced by a number of mechanisms, including transient hot-spots (randomly contacting the unique lamproite-source hybridized mantle; unlike ocean island chains, however, lamproites do not generally form time-space-related linear trends), the build-up of radiogenic heat (due to sources enriched in K, Th and U) and the tectonic conditions of the overlying crust (involving deep fractures). It should be noted that variations of this three-stage model of depletion-enrichmentmelting have been suggested by various workers (Best 1975; Beswick 1976; Thompson et al. 1984). When the lamproite sensu stricto and other ultrapotassic rock localities are plotted on a map with the continents in the positions of 180-200 Ma ago (see Fig. 26), a zonal pattern of the distribution of these K-rich rocks emerges. Two broad belts are evident: one belt 3000 km long includes the Indian, W Australian and Antarctic occurrences, and the other more diffuse belt includes the N American, European and W African localities. Since the mantle source regions
9
~ 7
of lamproites are evidently very old, this pre-drift zonation is perhaps expected and not only permits the possible prediction of other lamproite localities but also has important implications for mantle evolution. Is it possible that the southern hemisphere lamproite belt (representing 1.2x 103 Ma of magmatic activity) delineates a Precambrian (Archaean?) zone of recycling of crustal material (fossil Benioff zone)? While much speculation has filtered into this discussion, it is hoped that a coherent picture of lamproite petrogenesis has emerged. The lamproite revolution is only in its infancy and future studies of these exotic rocks will surely clarify many of the issues only briefly discussed here.
Kimberlite--lamproite relationships Dawson (1987) summarizes the similarities in geochemistry and mineralogy between olivine lamproites and the more micaceous kimberlites (including those in Group II of Smith (1983a, b)) and suggests that olivine lamproites may, in fact, be members of the Group II kimberlites. Smith (1983b) demonstrated Nd-Sr-Pb isotopic similarity between Group II kimberlites and olivine lamproites from PRA and WKB and suggested further that they are one and the same, a view that is additionally held by the present author. Smith (1983b) also suggested that lamproites (as well as Group II kimberlites) were derived by the partial melting of sub-continental lithosphere,
~
~'/~'~ ~
169
LAMPROITES 8~ II ULTRKAsPOTASSIC~/
~'~"~7
9 KIMBERLITES j
EAnt;
~.
\\ \
4-
"')l.".
-
\
,,
4-
FIG. 26. Distribution oflamproites, ultrapotassic rocks and kirnberlites on a map showing a reconstruction of the continents 180-200 Ma ago (Owen 1983). Lamproites fall along two broad belts.
I7O
S. C. Bergman
whereas Group I kimberlites were derived by the partial melting of sub-continental lithosphere or underlying asthenosphere. The fact that MARID-suite xenoliths occur in kimberlites, indicating lamproite-type plutonic-metasomatic processes in sub-continental lithosphere encountered by the host ascending kimberlite magmas, is consistent with both lamproite and kimberlite melts existing in similar areas in the upper mantle.
Relationships between lamproites and MARID and other mica-rich xenoliths The similarity between the geochemistry and mineralogy of lamproites on the one hand and the M A R I D glimmerite and PKP (amphiboleperidotite) suites of xenoliths found in kimberlites on the other suggests some type of relationship. MARID xenoliths, which are also known as 'mantle pegmatites' consist of mica, amphibole, rutile, ilmenite and diopside (Dawson & Smith 1977); in addition, compositionally-similar phases (Ti-mica, K-richterite) are observed as metasomatic alteration products of many peridotite xenoliths from kimberlites (Erlank 1970, 1973; Erlank & Rickard 1977; Erlank et al. 1982; Jones & Smith 1985). Haggerty (1983) and Haggerty et al. (1983) discussed K - B a - C r titanates from metasomatized peridotite xenoliths from S Africa; these phases are not unlike the priderite-jeppeite phases from lamproites. Jones et al. (1985) presented isotopic data for xenoliths from Kimberley, S Africa, which indicated that glimmerite xenoliths were disrupted pegmatitic segregations of Group I kimberlites (terminology of Smith (1983a, b)), whereas M A R I D and PKP xenoliths formed by magmatic-metasomatic processes involving both Group I and Group II components. However, trace-element and isotopic studies on metasomatized peridotite and M A R I D xenoliths from the Bultfontein kimberlite mine, Kimberley, S Africa (Kramers et al. 1983) have demonstrated that the S r - N d - P b isotopic features are consistent with mixing of a metasomatic fluid from a slightly depleted mantle source with mantle material showing a timeintegrated enrichment in incompatible elements. Nevertheless, the dominant (albeit extremely rare) amphibole in kimberlites or in the replacement zones in their included-mantle xenoliths is the K-rich richterite (slightly depleted in Ti relative to lamproite K-richterites) which is the characteristic (and nearly ubiquitous) amphibole in lamproites. The subtle compositional differ-
ences between K-richterites of lamproite, M A R I D xenoliths and metasomatized peridotite xenoliths can be explained by a variable buffering capacity of the host mantle peridotite which may modify the original amphibole composition. The presence of K-richterite, combined with the Tiand phlogopite-rich nature of M A R I D suite xenoliths, further suggests that lamproite-like fluids were the agents of patent metasomatism of kimberlite xenoliths; lamproite melts may be responsible for the precipitation of M A R I D xenoliths. However, much remains to be explored in the M A R I D and metasomatized xenoliths before the extent of this genetic relationship is fully realized. Interestingly, the nature of metasomatic phases (mica and amphibole) in lherzolite and harzburgite xenoliths from alkali basalts (higher A1, and kaersutite or pargasite rather than richterite (Boettcher & O'Neil 1980; Menzies & Murthy 1980b)) substantiates a distinct nature to the mantle source regions of MARIDsuite xenoliths and metasomatized alkali basalt xenoliths.
Conclusions Lamproites are recognized as a coherent but variable and exotic petrographic and chemical group of K-rich igneous rocks which border on and share certain petrogenetic aspects with the alkali basalt, kimberlite, lamprophyre and other K-rich rock clans. Lamproites should not be included in the lamprophyre clan. Lamproites can be effectively distinguished from these other rock types on the basis of mineralogy, mineral chemistry and whole-rock major, trace-element and isotope chemistry. Lamproites are unparalleled by other rock types with respect to their compositions. Known lamproites occur in 21 suites on six continents. Lamproites occur closer to the margins of continents whereas kimberlites are most abundant nearer the craton cores; lamproites often intrude crust that overlies fossil Benioff zones. Lamproites are partial melts of a metasomatized (i.e. phlogopite-, apatite-bearing) but depleted (in Na, AI, Ca) source-mantle peridotite (harzburgite). A three-stage model (depletion-enrichment-melting), at least, is required to explain the evolution of the source regions of lamproite magmas. Lamproites differentiate by processes including crystal fractionation, fractional resorption and volatile-phase transfer, among others. ACKNOWLEDGMENTS: This research was supported by the Anaconda Minerals Company, a division of the Atlantic Richfield
Lamproites and K-rich igneous rocks Corporation, in conjunction with their diamond exploration programme in the U.S. and Kalimantan. I thank R. W. Knostman, I. Gemuts and L. G. Krol for their continued support. Preprints supplied by R. Mitchell, B. Scott-Smith, D. Colchester, J. V. Smith, P. H. Nixon, P. Gregory, D. Velde, R. Vollmer, C. Hawkesworth and M. Menzies are gratefully appreciated. These workers, as well as L. G. Krol, N. R. Baker, M. Skinner, D. Velde, S. Bachinski, P. Berendson, M. Bickford, K. Collerson, F. Albarede, K. Fraser and F. Dodge, participated in many fruitful discussions which served to improve the content of this report. Two days with D. and B. Velde in Ayron proved enlightening. Pat Bickford generously supplied unpublished mineral chemical data on the Hills Pond rocks, Peter Nixon provided samples of the Murcia-Almeria lamproites, Pieter Berendsen provided samples of the Hills Pond lamproites and B. Scott-Smith supplied samples of the Holsteinsborg lamproites. I thank R. Baker, D. Dunn, J. Crann and W. Turner for logistical support and assistance in the field in the U.S.A., Indonesia and Australia. S. Self greatly assisted in the interpretation of the piperno and other pyroclastic rock textures. L. Noodles and Maggie clarified many confusing concepts. An exceedingly constructive review by Sharon Bachinski was invaluable; D. Velde, L. Krol and J. G. Fitton additionally provided useful criticisms on an earlier draft. I thank J. G. Fitton and B. G. J. Upton for only requiring a 25% reduction in the length of the original manuscript. I thank G. W. DeArmond, G. McEntire, D. Schraeder, L. Auvermann and I. Martinez for performing many arduous literature searches and for obtaining many of the obscure references cited in the text. I thank J. Toney and D. J. Henry for much invaluable help in the microprobe work. The rock and mineral chemistry data bases were compiled with the able assistance of V. Mount, S. R. Yang, F. Stiffand E. Kinsel to whom I express sincere thanks. The manuscript was typed by S. Epperson and the graphics were professionally produced by N. Murray.
PRA
SB
(ALL) (CSN) (DSV) (FOR) (GAL) (HM)
(KNX) (NHB)
(SEW) (SFC) (TB) (CHIN)
171
Bolivar 1977; Gogineni et al. 1978; Williams 1891; Scott-Smith & Skinner 1984a, b; Bergman, unpublished data Matson 1960; Velde 1975; Bergman, unpublished data Hawkins 1976 Van Kooten 1980 Dodge & Moore 1981 Ross 1926b Allen et al. 1975 Weed & Pirsson 1896; Pirsson 1905; Osborne & Roberts 1931 ; Wolff 1938; Hurlbut & Griggs 1939; Buie 1941; Tappe 1966; Witkind 1969, 1973; Woods 1975; Bergman, unpublished data Bastin 1906 Williams 1936; Lewis 1973; Schmitt et al. 1974; Rogers et al. 1982 Miller 1972 Templeman-Kluitt 1969 Cross 1906 Arculus & Smith 1979; Schulze & Helmstaedt 1979
Australia WKB
ARG (NSW)
Wade & Proder 1940; Prider 1960, 1982; Atkinson et al. 1984a; Jaques et al. 1984b; Nixon etal. 1984; Bergman, unpublished data Atkinson et al. 1984a Cundari 1973
Europe Osann 1906; Washington 1917; Parga Pondal 1935; Fuster & Pedro 1953; Borley 1967; Fuster et al. 1967; Nixon etal. 1984; Venturelli etal. 1984a Dal Piaz et al. 1979; Venturelli et al. NWI 1984b Hall 1982 PEN Tidmarsh 1932; Knill 1969 HLM Velde 1967 COR (TUSC) Gallo 1984 (SUNN) Furnes et al. 1982 N6mec 1972 (BOH) Duda & Schminke 1978 (LAC) MAP
APPENDIX 1 Sources of whole-rock major and trace-element analytical data in Tables 4-7 and associated figs 20-22
N America EVD HOL HP KAM LH
Libby 1975 Scott 1979 Franks et al. 1971 ; Merrill et al. 1977 Best et al. 1968 ; Bergman, unpublished data Carmichael 1967; Kuehner 1980, Kuehner et al. 1981; Schultz & Cross 1912; Yagi & Matsumoto 1966; Bergman, unpublished data
Africa BOB PPS
Bardet 1973 (Swartruggens only) Skinner & Scott 1979; Bergman, unpublished data
172 (TAN) (AZZ)
S. C.
Bergman
Holmes & Harwood 1937; Holmes 1950; Higazy 1954; Sahama 1974; Mitchell & Bell 1976 Vila et al. 1974
Antarctica
GSB Sheraton & Cundari 1980 MBAY, PP Sheraton & England 1980 Asia and Indonesia
COC Lacroix 1933a, b CHE Bergman & Baker 1984 GDW S a r k a r e t a l . 1980; G u p t a e t a l . 1983 (BAL) Kostyuk 1983 (KAJ) Brouwer 1909 Abbreviations are given in Tables 2 and 3 and in the text. Average compositions
The major- and trace-element data for lamproite sensu stricto are cited above; the major-element data for lamprophyres, kimberlites and alkali basalts were compiled on a computerized data bank that is a combination of the Mutchler et al. (1973) PETROS data bank and a literature compilation by Bergman (File DIROC). Individual references can be obtained by writing to Bergman. Trace-element data for kimberlites and alkali basalts are taken from Wedepohl & Muramatsu (1979) and data for lamprophyres are taken from Rock (1984, 1986). The estimated composition of the primitive mantle comes from Taylor & McLennan (1981) or Mason (1979).
leucite, A analcime, J priderite and jeppeite, A amphibole, I ilmenite, G garnet (suite abbreviations are as given in appendix 3). ARG
Atkinson et al. 1984a (G, H, C); Scott-Smith & Skinner 1984b (P) GSB Sheraton & Cundari 1980 (I, S, H, O, P, C) HOL Scott 1981 (A, S, O, P, C) HP Merrill et al. 1977 (A, P, C); P. Bickford, personal communication 1984 (A, C, P, S); Mitchell 1985 (P, S) LH Carmichael 1967 (A, S, O, P, C, J, F, L); Barton 1979 (L, F, C, P); Kuehner 1980 (A, S, O, P, C); Barton & van Bergen 1981 (H, O, P, C); Kuehner et al. 1981 (A, S, O, P, C); Mitchell 1985 (S, J) MAP Borley 1967 (P, C); Carmichael 1967 (P, F, O, C, A); Fustor et al. 1967 (P); Lopez Ruiz & Rodriguez Badiola 1980 (A, O, P, C); Venturelli et al. 1984a (A, H, S, P, P, C); Mitchell 1985 (P, F, I) MBAY, PP Sheraton & England 1980 (A, P, I,
F) PEN PRA
SB
WKB
APPENDIX 2 Sources of mineral chemical data of Table 10 and Figs 23-25
The mineral chemical data used in these figures and table have been compiled in a computerized data file DIMIC that comprises over 137,000 analyses of phases from kimberlites, lamprophyres, lamproites, alkali basalts and their included xenoliths and megacrysts (S. C. Bergman, unpublished). Since it is unreasonable to tabulate all the data sources here, a copy of the sources can be obtained by writing to Bergman. Only sources containing lamproite mineral chemical data will be summarized. The abbreviations are as follows: P phlogopite, O olivine, C clinopyroxene, H orthopyroxene, S spinel, F sanidine, L
Hall 1982 (A, P, F) Lewis et al. 1976 (G, S, O, P, C); Gogineni et al. 1978 (A, G, S, O, P, C); Mitchell & Lewis 1983 (A, C, P, J); Scott-Smith & Skinner 1984a, c (A, S, O, P, C); Mitchell 1985 (J) Velde 1975 (A, S, P, C); Mitchell 1985 (P, O, N, F, S); Henry & Bergman, unpublished data (A, P, C,O, S) Carmichael 1967 (C, P, A, F, L, J); Mitchell 1981 (P); Mitchell & Lewis 1983 (A, P, C): Atkinson et al. 1984a (G, H, S, C); Jaques et al. 1984a, b (A, G, H, S, O, C); Pryce et al. 1984 (J); Scott-Smith & Skinner 1984b (A, P, S); Mitchell 1985 (P, A, C, L, S,J)
APPENDIX 3 Alphabetical listing of rock suites discussed in the paper
ARG (AZZ) (BAL) BOB (BOH) CHE
Argyle, W Australia Azzaba, Algeria, Africa Baikal Rift-Aldan Shield, U.S.S.R. Bobi and Seguela, Ivory Coast, Africa Bohemian Massif, Czechoslovakia Chelima, Andhra Pradesh, India
Lamproites and K-rich igneous rocks (CHN) (CJ)
(coc) (COL) COR (CSN) DSV EVD (FOR) (GDW) GSB HLM (HM)
HOL HP KAM (KNX) (LAC) LH LUA MAP MBAY (NHB)
(NSW) NWI
Channel Islands, U.K. Campos de Jordao, Brazil Coc Pia, N Vietnam Colima Graben, Mexico Sisco, Corsica, France Central Sierra Nevada, California Deep Spring Valley, California Enoree Vermiculite, S Carolina Fortification Dyke, Colorado Gondwana Coalfields, India Gaussberg, Antarctica Holmeade Farm, U.K. Highwood Mountains, Montana Holsteinsborg, W Greenland Hills Pond, Kansas Kamas, U t a h Knox County, Maine Laacher See Province, Germany Leucite Hills, Wyoming Luangwa Graben, Zambia Murcia-Almeria province, Spain Mount Bayliss, Antarctica N a v a j o - H o p i Buttes, Arizona Lake Cargelligo Area, NSW, Australia N W Italy
PAM PEN PIS PD PP PPS PRA (RP) (SAB) (SAC) SB (SEW) (SRE) (SUNN) (TAN) (TB) (TUSC) WKB
173
Pamir, U.S.S.R. Pendennis, U.K. Orciatico, Pisa, Italy Piedade, Brazil Priestly Peak, Antarctica Postmasburg, Pneil etc. S Africa Prairie Creek, Arkansas Roman Province, Italy Santo Antonia da Barra, Brazil Sacramento, Brazil Smoky Butte, Montana Seward Peninsula, Alaska Srednogorie, Bulgaria Sunnfjord, Norway Toro-Ankole, Birunga, U g a n d a Two Buttes, Colorado Tuscany, Italy W Kimberley, Fitzroy Basin, W Australia
Those suites in parentheses are not lamproites but merely K-rich rocks; those not in parentheses conform to the present definition of lamproite sensu stricto
References ALLEN, D. G., PANTELEYEV,A. & ARMSTRONG,A. T. 1975. Porphyry copper deposits of the Alkalic Suite. Can. Inst. Min. Spec. 15, 402-14. ANDERSON, O. L., BREED, W. J., BREED, C. S., CHIDESTER, A. H., MCGETCHIN, T. R. & SHOEMAKER, E. M. 1970. Field guide to kimberlitebearing diatremes of northern Arizona-southern Utah including stops at Hopi Buttes, Monument Valley and Grand Canyon. Int. Syrup. on Mechanical Properties and Processes of the Mantle, June 2428, 1970, Flagstaff', Arizona, 52 pp. ANDREWS, J. R. & EMELEUS, C. 1971. Preliminary account of kimberlite intrusions from the Frederikshab district SW Greenland. Geol. Surv. Greenland Rep. 31, 26 pp. AOKI, K. I. & KURASAWA,H. 1984. Sr-isotope study of the tephrite series from Nyamuragira volcano, Zaire. Geochem. J. 18, 95-100. APPARADHANULU, K. 1962. Minette and riebeckitekalisyenite dykes in some Upper Cuddapah rocks, Kurnool district, Andhra Pradesh. Rec. geol. Surv. India, 94(2), 303-4. APPLETON, J. D. 1972. Petrogenesis of potassium-rich lavas from the Roccamonfina volcano, Roman Region, Italy. J. Petrol. 13, 425-56. ARCULUS,R. J. & SMITH,D. 1979. Eclogite, pyroxenite and amphibolite inclusions in the Sullivan Buttes latite, Chino Alley, Yavapai County, Arizona. In: BOYD F. R. & MEYER, H. O. A. (eds) The Mantle Sample: Inclusions in Kimberlites and Other Volcanics, Proc. 2nd Int. Kimberlite Conf. Vol. 2. pp. 309-
17. American Geophysical Union, Washington, DC. ARIMA, M. & EDGAR, A. D. 1980. Stability of wadeite (Zr2KzSi6018) under upper mantle conditions: petrological implications. Contrib. Mineral. Petrol. 72, 191-5. & 1981. Substitution mechanisms and solubility of titanium in phlogopites from rocks of probable mantle origin. Contrib. Mineral. Petrol. 77, 288-95. -& - - 1983. High pressure experimental studies of a katungite and their bearing on the genesis of some potassium-rich magmas of the West Branch of the African Rift. J. Petrol. 24, 166-87. ATKINSON, W. J. 1982. Diamond discoveries in the Kimberley, Western Australia. Trans. Inst. Min. Metall. 91A, 135-7. - - , HUGHES,F. E. & SMITH,C. B. 1982. A review of the kimberlitic rocks of Western Australia. Terra Cognita, 2, 204. & -1983. A review of the kimberlitic rocks of Western Australia. Proc 6th Australian Geology ConS, pp. 284-5. Geological Society of Australia, Sydney. & -1984a. A review of the kimberlitic rocks of western Australia. In: KORNPROaST, J. (ed.) Kimberlites I. Kimberlites and Related Rocks, pp. 195-225. Elsevier, Amsterdam. - - , SMITH,C. B. & BOXER,G. L. 1984b. The discovery and evaluation of the Ellendale and Argyle Lamproite diamond deposits, Kimberley region,
x74
S. C. Bergman
Western Australia. Soc. Min. Eng. A I M E Preprint 84-384. BACHINSKI, S. W. & SCOTT, R. B. 1979. Rare-earth and other trace element contents and the origin of minettes (mica-lamprophyres). Geochim. cosmochim. Acta, 43, 93-100. & SIMVSON, E. L. 1984. Ti-phlogopites of the Shaw's Cove minette: a comparison with micas of other lamprophyres, potassic rocks kimberlites and mantle xenoliths. Am. Mineral. 69, 41-56. BAGSHAW, A. N., DORAN, B. H., WHITE, A. H. & WILLIS, m. C. 1977. Crystal structure of a natural K-bearing hexatitanate isostructural with K2Ti6OI 3. Aust. J. Chem. 30, 1195-200. BAILEY, D. K. 1974. Continental rifting and alkaline magmatism. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 148-59. Wiley, New York. 1982. Mantle metasomatism--continuing chemical change within the Earth. Nature, Lond. 296, 525-30. BALASEBRAHMANYAN,M. N., MURTY, M. K., PAUL, D. K. & SARKAR, 1978. Potassium-argon ages of Indian kimberlites. J. Geol. Soc. India, 19(12), 5845. BANERJEE, S. 1953. Petrology of the lamprophyres and associated rocks of Raniganj coal field. Indian Mineral. J. 1, 9-29. BARBERI, F., BIZOUARD, H., CAPALDI, G., EERRARA, G., GASPARINI, P., INNOCENTI, F., JORON, J. L., LAMBRET, B., TREUIL, M. & ALLEGRE, C. 1978. Age and nature of basalts from the Tyrrhenian Abyssal Plain. Initial Rep. Deep Sea Drilling Project, 42 (Part 1), 509-14. -& INNOCENTI, F. 1967. Le rocce Selagitiche di arciatico e Montecatini in Val di Cecina. Atti Soc. Tosc. Pisa Sci. Nat. Mem. P. V., Ser. A, 74, 139-80. , FERRARA, G., KELLER, J. & VILLARI, U 974. Evolution of Aeolian arc volcanism (southern Tyrrhenian Sea). Earth planet. Sci. Lett. 21, 26976. --, SANTACROCE, R. & VARET, J. 1982. Chemical aspects of rift magmatism. In : PALMASONG. (ed.) Continental and Oceanic Rifts, pp. 223-58. American Geophysical Union, Washington, DC. BARBERI, M., PENTA, A. & TURI, B. 1975. O and Sr isotope ratios in some ejecta from the Alban Hills volcanic area, Roman comagmatic region. Contrib. Mineral. Petrol. 51, 127-33. BARDET, M. G. 1973. Geologie du Diamant, Vol. 1. Editions BRGM no. 83, Paris. 1974. Geologie du Diamant, Vol. 2. Editions BRGM no. 83, Paris. 1977. Geologie du Diamant, Vol. 3. Editions BRGM no. 83, Paris. & VACHETTE,M. 1966. Determination of the ages of kimberlites of West Africa. 3rd Symp. on African Geology, Rep. GCLO 88, pp. 15-17. Commonwealth Geological Liaison Office. BAREN, F. A. VAN, 1948. On the petrology of the volcanic area of the Geonoeng Moeria (Java). Meded. Alg. Proefstn. Landb. Bultenzorg, Java, 60. BARTON, M. 1976. Melting relations of some ultrapotassic volcanic rocks. Prog. Exp. Petrol., NERC Publ. Ser. D, Vol. 6, pp. 91-4. -
-
-
-
-
-
1979. A comparative study of some minerals occurring in the potassium-rich alkaline rocks of the Leucite Hills, Wyoming, the Vico Volcano, western Italy, and the Toro-Ankole Region, Uganda. Neues Jb. Mineral. Abh., 137(2), 113-34. & VANBERGEN, M. J. 1981. Green clinopyroxenes and associated phases in a potassium-rich lava from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 77, 101-14. & HAMILTON,D. L. 1978. Water-saturated melting relations to 5 kilobars of three Leucite Hills lavas. Contrib. Mineral. Petrol. 66, 41-9. & -1979. The melting relationships of a madupite from the Leucite Hills, Wyoming, to 30 kb. Contrib. Mineral. Petrol. 69, 133-42. &-1982. Water saturated melting experiments bearing upon the origin of potassium-rich magmas. Mineral. Mag. 45, 267-78. & WOOD, B. J. 1976. Equilibration of orendite with lherzolitic mineral assemblages: a combined experimental/theoretical approach. Prog. Exp. Petrol., NERC Publ., Ser. D, Vol. 6, pp. 90-1. BASTIN, E. S. 1906. Some unusual rocks from Maine, J. Geol. 14, 173-87. BASE, m. R. & TATSUMOTO, M. 1980. Nd-isotopes in selected mantle-derived rocks and minerals and their implications for mantle evolution. Contrib. Mineral. Petrol. 75, 43-54. BEGER, P. J. 1923. Der chemismus der lamprophyre. In: NIGGLI, P. (ed.) Gesteins-undMineralProvinzen, Vol. 1. pp. 448-51. Verlag Gebruder Borntraeger, Berlin. BELL, K. & POWELL, J. L. 1969. Strontium isotopic studies of alkalic rocks: the potassium-rich lavas of the Birunga and Toro-Ankole regions, east and central equatorial Africa. J. Petrol. 10, 536-72. & -1970. Strontium isotopic studies of alkalic rocks: the alkalic complexes of eastern Uganda. Geol. Soc. Am. Bull. 81, 3481-90. BELLON, H. 1981. Chronologique radiometrique (KAr) des manifestations magmatiques autour de la Mediterran6e occidentale entre 33 Ma et 1 Ma Proc. Int. Conf. Urbino University, Italy, pp. 34160. -& LETOUSEY, J. 1977. Volcanism related to plate tectonics in the Western and Eastern Mediterranean. In: BYE DUVAL B. & MONTADERTL. (eds) Int. Syrup. on Mediterranean Basins, pp. 165-84. Editions Techniq, Paris. BERENDSEN, P., CULLERS, R. L., MANSKER, W. L. & COLE, G. P. 1985. Late Cretaceous kimberlite and lamproite occurrences in E. Kansas, U.S.A. Geol. Soc. Am. Abstr. Progm. 17, 151. BERGMAN, S. C. & BAKER, N. R. 1984. A new look at the Proterozoic dikes from Chelima, Andhra Pradesh, India: diamondiferous lamproites? Geol. Soc. Am. Abstr. Progm. 16(6), 444. & -The Chelima dikes, Andhra Pradesh, India: diamondiferous lamproites? Submitted to Nature, Lond. - - , DUNN, D. P. & KROL, L. G. 1985. Petrology and geochemistry of the Linhaisai Minette, Karamu River, central Kalimantan. Abstr. Geol. Assoc. Can. Progm. 10, A4. -
-
-
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks BEST, M. G. 1975. Migration of hydrous fluids in the upper mantle and potassium variation in calcalkaline rocks. Geology, 3, 429-32. --, HENAGE, L. F. & ADAMS, J. A. S. 1968. Mica peridotite, wyomingite, and associated potassic igneous rocks in northeastern Utah. Am. Mineral. 53, 1041-8.
BESWICK,A. E. 1976. K and Rb relations in basalts and other mantle derived materials. Is phlogopite the key? Geochim. cosmochim. Acta, 32, 1167-84. BICKFORD, M. E., MOSE, D. E., WETHERILL,G. W. & FRANKS, P. C. 1971. Metamorphism of Precambrian granitic xenoliths in a mica periodite at Rose Dome, Woodson County, Kansas. Part 1, Rb-Sr isotopic studies. Geol. Soc. Am. Bull. 82, 2863-8. BLAXLAND, A. B. 1976a. Geochronology and isotope geochemistry of the Gardar Alkaline complex, South Greenland. PhD Thesis, University of Edinburgh (unpublished). -1976b. Rb-Sr isotopic evidence for the age and origin of the Ivigtut granite body and associated cryolite body, South Greenland. Econ. Geol. 71, 864-9. BOCCALETTI, M., MANETTI, P., PECCERILLO, A. & STANISHEVA-VASSILEVA,G. 1978. Late Cretaceous high potassium volcanism in E. Srednogorie, Bulgaria. Geol. Soc. Am. Bull. 89, 439-47. BOETTCHER, A . L . , MYSEN, B . O . & MODRESKI, P. J.
1975. Melting in the mantle: phase relationships in natural and synthetic peridotite-H20 and peridotite-HzO-CO2 systems at high pressures. Phys. Chem. Earth, 9, 855-68. - & O'NEIL, J. R. 1980. Stable isotope, chemical and petrographic studies of high-P amphiboles and micas: evidence for metasomatism in the mantle source regions of alkali basalts and kimberlites. Am. J. Sci. 280A, 594-621. BOLIVAR,S. L. 1977. Geochemistry of the Prairie Creek, Arkansas and Elliott County, Kentucky intrusions. PhD Thesis, University of New Mexico (unpublished). - 1982. The Prairie Creek kimberlite, Arkansas. In: MCFARLAND, J. D., III (ed.) Contributions to the Geology of Arkansas, State of Arkansas Geological Commission Misc. Publ. 18, pp. 1-21. - 1984. An overview of the Prairie Creek intrusion, Arkansas. Soc. Min. Eng. A I M E Fall Meet., October 1984, Preprint 84-346. - & BROOKINS,D. G. 1979. Geophysical and Rb-Sr study of the Prairie Creek, AR kimberlite. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds." their Geology, Petrology and Geochemistry, Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 289-99. American Geophysical Union, Washington, DC. BORLEY, G. D. 1967. Potassium-rich volcanic rocks from Southern Spain. Mineral. Mag. 36, 364--79. BORSI, S., FERRARA, G. & TONGIORGI, E. 1967. Determinazioni con il metoda K/Ar delle rocce magmatiche della Toscana. Boll. Soc Geol. Italy, 86, 403-10. BOUTWELL,J. M. 1912. Geology and ore deposits of the Park City District, Utah. U. S. Geol. Surv. Prof Pap. 77, 231.
175
BOWEN, C. F. 1915. Possibilities ofoil in the Porcupine Dome, Rosebud County, Montana. U. S. Geol. Surv. Bull. 621, 61-70. BOWEN, N. L. 1928. Evolution of the Igneous Rocks. Princeton University Press, Princeton, NJ. BRADLEY, W. H. 1964. Geology of Green River Formation and associated Eocene rocks in SW Wyoming and adjacent parts of Colorado and Utah. U.S. Geol. Surv. Prof. Pap. 496A, A1-A86. BRANNER,J. C. & BRACKETT,R. N. 1889. The peridotite of Pike County, Arkansas. Am. J. Sci. 38, 50-9. BRAVO, M. S. & O'HARA, M. J. 1975. Partial melting of phlogopite-bearing synthetic and garnet-lherzolites. Phys. Chem. Earth, 9, 845-54. BRIQUEU, L., BOUGAULT, H. & JORON, J. L. 1984. Quantification of Nb, Ta, Ti and V anomalies in magmas associated with subduction zones: petrogenetic implications. Earth planet. Sci. Lett. 68, 297-308. BROOKINS, D. G., DELLA VALLE, R. S. & BOLIVAR, S. L. 1979. Significance of uranium abundance in United States kimberlites: In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes, and Diamonds: Their Geology, Petrology and Geochemistry: Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 280-8. American Geophysical Union, Washington, DC. BROOKS, C. K., FAWCETT, J. J., GITTINS, J. G. & RUCKLIDGE, J. C. 1981. The Batbjerg complex, east Greenland: a unique ultrapotassic Caledonian intrusion. Can. J. Earth Sci. 18, 274-85. , NOE-NYGGARD,A., REX, D. C. & ROENSBO,J. C. 1978. An occurrence of ultrapotassic dikes in the neighborhood of Holsteinsborg, West Greenland. Bull. geol. Soc. Denmark, 27, 1-8. BROUWER, H. A. 1909. Glimmerleucitbazalt van OostBorneo. Versl. Kon. Akad. Wetensch. Amsterdam wis. Nat. Afd. 18, 85-91. BUCKING, H. 1899. Ber. Naturforsch. Ges. Freiburg, i ( 1 1 ) , 78. Bum, B. F. 1941. Igneous rocks of the Highwood Mountains, Montana, III, Dike and related intrusives. Geol. Soc. Am. Bull. 52, 1754-807. - & STEWART,O. F. 1954. Origin of vermiculite at Tigerville, S. Carolina. Geol. Soc. Am. Bull. 65, 1356-7. BUREAUOF MINERALRESOURCES,GEOLOGYAND GEOPHYSICS, 1976. Geological Map of Australia (1/2,500,000 scale, 4 sheets), Canberra, A.C.T. CARABALLO, J. M. 1975. Geoquimica de las rocas lamproiticas espanolas. PhD Thesis, University of Madrid (unpublished). CARMICHAEL, I. S. E. 1967. The mineralogy and petrology of the volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 15, 2466. , NICHOLLS, J., SPERA, F. J., WOOD, B. J. & NELSON, S. A. 1977. High temperature properties of silicate liquids: application to the equilibration and ascent of basic magma. Phil. Trans. R. Soc. Lond. Ser. A, 286, 373-431. - - , TURNER, F. S. & VERHOOGEN, J. 1974. Igneous Petrology, McGraw-Hill, New York. CARR, S. G. & OLLIVER,J. G. 1980. Olivine basalt dyke
I76
S. C. Bergman
near Terowie. Qt. Geol. Notes Geol. Surv. S. Aust. 73, 2-7. CARRARO, F. & FERRAR, G. 1968. Alpine 'Tonalite' at Miagliano, Biella (zona-diorite-Kinzigitica). Schweiz Mineral. Petrogr. Mitt. 48, 75-8. CHATTERJEE, S. C. 1974. Petrography of the Igneous and Metamorphic Rocks of India, 391 pp. Macmillan, Bombay. CI-IAYES, F. 1976. Independent exploration of the database and reduction repertoire of the petrographic information system. Carnegie Instn. Wash. Yb. 75, 791. CIVETTA,L., INNOCENTI, F., MANETTI,P., PECCERILLO, A. &POLI, G. 1981. Geochemical characteristics of potassic volcanics from Mts. Ernici (southern Latium, Italy). Contrib. Mineral. Petrol. 78, 37-47. , ORSI, G. & SCANDONE, G. 1978. Eastwards migration of the Tuscan anatectic magmatism due to anticlockwise rotation of the Appenhines, Nature, Lond. 276, 604-6. CLARK, F. W. & W~HINGTON, H. S. 1924. The composition of the earth's crust. Prof. Pap. U.S.G.S. 127. CLEMENT, C. R., SKINNER, E. M. W. & SCOTT, B. H. 1977. Kimberlite redefined. Extended Abstr. 2nd Kimberlite Conf., Santa Fe, NM. COLCHESTER, D. M. 1982. Geology and petrology of some kimberlites near Terowie, S. Australia. M. Sc. Thesis, New South Wales Institute of Technology (unpublished). -1983. The geology and petrology of some kimberlites near Terowie, S. Australia, Prof. 6th Australian Geology Cony., pp. 282-3. Geological Society of Australia, Sydney. COLLERSON,K. D. & MCCULLOCH,M. T. 1982. Nd and Sr isotope geochemistry of leucite-bearing lavas from Gaussberg, Eastern Antarctica. J. geol. Soc. Aust. 30. fONDlE, K. C. 1976. Plate tectonics and crustal evolution. Pergamon Press, Inc., New York. CORNELISSEN, A. K. & VERWOERD, W. J. 1975. The Bushmanland kimberlites and related rocks. Phys. Chem. Earth, 9, 71-80. COSGROVE,M. E. 1972. Geochemistry of the potassiumrich Permian volcanic rocks of Devonshire, England. Contrib. Mineral. Petrol. 36, 155-70. Cox, K. G., HAWKESWORTH,C. J., O'NIONS, R. K. & APPLETON, J. D. 1976. Isotopic evidence for the derivation of some Roman region volcanics from anomalously enriched mantle. Contrib. Mineral. Petrol. 56, 173-80. CRAWFORD, A. R. 1969. India, Ceylon and Pakistan: new age data and comparisons with Australia. Nature, Lond. 223, 380-4. -& COMPSTON, W. 1970. Age of the Vindhyan Systems of Peninsular India. Q. J. Geol. Soc. Lond. 125, 351-71. & -1973. An age of the Cuddapah and Kurnool systems, southern India. J. geol. Soc. Aust. -
-
19,
4.
CRITTENDEN, M. D. & KIRSTLER, R. W. 1966. Isotopic dating of intrusive rocks in the Cottonwood area, Utah. Geol. Soc. Am. Spec. Pap. 101,298-9. CROSS, C. W. 1897. The igneous rocks of the Leucite
Hills and Pilot Butte, Wyoming. Am. J. Sci. 4, 1 1 5 - 4 1 .
1906. Prowersose (syenitic lamprophyre) from Two Buttes, Colorado. J. Geol. 14, 165-72. CULLERS, R. L., RAMAKRISHNAN,S., BERENDSEN, P. & GRIFFIN, T. 1985. Geochemistry and petrogenesis of lamproites, late Cretaceous age, Woodson County, Kansas, U.S.A. Geochim. cosmochim. Acta. 49, 1388-402. CUNDARI, A. 1973. Petrology of the leucite-bearing lavas in New South Wales. J. geol. Soc. Aust. 20, 465-92. & FERGUSON, A. K. 1982. Significance of the pyroxene chemistry from leucite-bearing and related assemblages. Tschermaks Mineral. Petrogr. Mitt. 30, 189-204. & O'HARA, M. J. 1976. Experimental study at atmospheric and high pressures of a mafic leucitite from New South Wales, Australia. Prog. Exp. Petrol, 3rd Rep. NERC Pub., Ser. D, 6, 260-1. --, RENARD, J. G. R. & GLEADOW, A. J. W. 1978. Uranium-potassium relationship and apatite fission track ages for a differentiated leucite suite from New South Wales (Australia). Chem. Geol. 22, 11-20. CURRIE, K. L., CURTIS, L. W. & GITTINS, J. 1975. Petrology of the Red Wine alkaline complexes Central Labrador and a comparison with the Ilimaussaq complex, SW Greenland. Geol. Surv. Can. Pap. 75(1A), 271-80. DAL PIAZ, G. V., HUNZIKER, J. C. & MARTINOTTI,G. 1973. Excursion to the Sesia zone of the Schw. mineral-petrograph gesellschaft, September 30th to October 3rd, 1973. Schweiz Mineral. Petrogr. Mitt. 53, 447-90. --, VENTURELLI, G. & SCOLARI, A. 1979. Calfalkaline to ultrapotassic postcollisional volcanic activity in the internal northwestern Alps. Mem. Sci. Geol. 32. DALY, R. A. 1910. Origin of alkaline rocks. Geol. Soc. Am. Bull. 21, 87-118. 1914. Igneous Rocks and their Origin, 1st edn. McGraw-Hill, New York. -1933. Igneous Rocks and the Depths of the Earth, 598 pp. Hafner, New York. DAWSON, J. B. 1967a. A review of the geology of kimberlite. In: WYLLIE, P. J. (ed.) Ultramafic and Related Rocks, pp. 241-51. Wiley, New York. 1967b. Geochemistry and origin of kimberlite. In : WYLLIE P. J. (ed.) Ultramafic and Related Rocks, pp. 269-78. Wiley, New York. 1970. The structural setting of African kimberlite magmatism. In: CLIFFORD, T. N. • GASS, I. G. (eds) African Magmatism and Tectonics, pp. 32135. Oliver & Boyd, Edinburgh. -1979. New aspects of diamond geology. In: FIELDS, J. E. (ed.) Diamonds, pp. 539-54. Academic Press, New York. 1980. Kimberlites and their Xenoliths, 252 pp. Springer New York. 1984. Contrasting types of upper mantle metasomatism? In: KORNPROBST, J. (ed.)Kimberlites H." The Mantle and Crust-Mantle Relationships pp. 289-94. Elsevier, Amsterdam.
-
-
-
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks 1987. The kimberlite clan: relationship to olivine and leucite lamproites, and inferences for uppermantle metasomatism. In: LITTON,J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 95-102. & SMITH, J. V. 1977. The MARID (micaamphibole-rutile-ilmenite-diopside) suite of xenoliths in kimberlite. Geochim. cosmochim. Acta, 41, 309-23. DE MARCO, L. 1959. Su alcuni filoni lamprofirici radioattivi del complesso Sesia-Lanzo. Stud. Rich. Div. Geomin. Cnen, 1, 1-30. DE ROEVER, W. P. 1975. Belts with scattered leucite occurrences indicate direction of dip of coeval subduction zones. Petrologie, 1, 37-42. DERRICK, G. M. & GELLATLY,D. C. 1972. New leucite lamproites from the west Kimberley, western Australia. Bur. Min. Resources, Aust., Bull. 125, 103-19. DODGE, F. C. W. & MOORE, J. G. 1981. Late Cenozoic volcanic rocks of the southern Sierra Nevada, California: II. Geochemistry. Geol. Soc. Am. Bull. 92, 1670-761. DOLFI, D., FARINATO,R., TRIGILA, R. & HAMILTON,D. L. 1976. Some experimental and thermodynamic evidences bearing on the origin of potassic volcanics from central Italy. Proc. Int. Congress Thermal Waters, Geothermal Energy and Volcanism of the Mediterranean Area, Vol. 3 Natl. Tech. Univ., Athens, Greece. pp. 43-59. DORT, W. 1972. Late Cenozoic volcanism in Antarctica. In: ADIE, R. J. (ed.) Antarctic Geol. Geophys., pp. 645-52. Universitetsforlaget, Oslo. DRYGALSKI, E. VON 1904. The German Antarctic Expedition. Geogr. J. 24, 129-52. DUBEAU, M. I. & EDGAR, A. D. 1985. The stability of priderite up to 30 kb: inferences on its potential as a reservoir for K, Ba and Ti in mantle sources for lamproitic magmas. Geol. Assoc. Can. Progm. Abstr. 10, A15. DUDA, A. & SCHMINKE' H. V. 1978. Quaternary basanites, melilite nephelinites and tephrites from the Laacher See area (Germany). Neues Jb. Mineral. Abh. 132, 1-33. EDGAR, A. D. 1979. Mineral chemistry and petrogenesis ofultrapotassic-ultramafic volcanic rocks. Contrib. Mineral. Petrol. 71, 171-5. - 1983. Relationship of ultrapotassic magmatism in the western U.S.A. to the Yellowstone plume. Neues Jb. Mineral. Abh. 147, 35-46. & ARIMA, M. 1981. Geochemistry of three potassium-rich lavas from the west branch of the African rift: inferences on their genesis. Neues Jb. Mineral. Monatsh. 539-52. - & -1983. Conditions of phlogopite crystallization in ultrapotassic volcanic rocks. Mineral. Mag. 47, 11-19. - & CONDLIFFE, E. 1978. Derivation of K-rich ultramafic magmas from a peridotitic mantle source. Nature, Lond. 174, 639-40. , , BARNETT,R. L. & SHIRRAN, R. J. 1979. An experimental study of an olivine ugandite magma and mechanisms for the formation of its Kenriched derivatives. J. Petrol. 21,475-97. - -
--,
177
GREEN, D. H. & HIBBERSON, W. O. 1976. Experimental petrology of a highly potassic magma. J. Petrol. 17, 339-56. EGGLER, D. H. 1977. The principle of the zone of invariant vapor composition: an example in the system CaO-MgO-SiO2-COz-H20 and implications for the mantle solidus. Carnegie Instn. Wash. Yb. 76, 428-33. ERLANK, A. J. 1970. Distribution of potassium in mafic and ultramafic nodules. Carnegie Instn. Wash. Yb. 68, 433. 1973. Kimberlite potassium richterite and the distribution of potassium in the upper mantle. Int. Conf. on Kimberlites, Cape Town, Extended Abstracts, pp. 103-6. Univ. Capetown Rondebosch. & RICHARD,R. S. 1977. Potassic richterite bearing peridotites from kimberlite and the evidence they provide for upper mantle metasomatism. Extended Abstr., 2nd Kimberlite Conf., Santa Fe, NM. - - - , ALLSOPP,H . L . , H A W K E S W O R T H , C . J . & MENZIES, M. A. 1982. Chemical and isotopic characterization of upper mantle metasomatism of peridotite nodules from the Bulfontein Kimberlite. Terra Cognita, 2, 261-3. ESCHER, A. 1970. Tectonic~Geological map of Greenland (1/2,500,000 scale), Geological Survey of Greenland, Copenhagen. - • WATTERSON,J. 1973. Kimbedites and associated rocks in the Holsteinsborg-Sondre Stromfjord Region, Central West Greenland. Gronland Geol. Unders. Rapp. 55. ESKOVA, E. M. & KAZAKOVA, M. E. 1955. Shcherbakovite--a new mineral. Am. Mineral. 40, 788. ESPERANCA, S. & HOLLOWAY, J. R. 1985. An experimental study of a mafic minette from Buell park, Arizona. Geol. Assoc. Can. Progm. Abstr. 10, A17. EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In: THORPE, R. S. (ed.) Andesites, pp. 25-95. Wiley, New York. - & LEMAITRE, R. W. 1980. Some regional compositional differences within Tertiary-Recent orogenic magmas. In: LEMAITRE, R. W. & CUNDARI, A. (eds) Chemical Characterization o f Tectonic Province. Chem. Geol. 30, 257-83. FALLOT, P. & JEREMINE, E. 1932. Remarque sur une vari6t6 nouvelle de jumillite et sur l'extension des laves de ce groupe. C. R. Congr. Natl. Soc. Savantes 1929, 1-13. FARQUHARSON,R. A. 1920. Petrologic work. Annu. Rep. geol. Surv. W. Aust. 1919, 42. -1922. Petrologic work. Annu. Rep Geol. Surv. W. Aust. 1921, 56. FAURE, G. 1977. Principles of Isotope Geology, 464 pp. Wiley, New York. -& POWELL, J. L. 1972. Strontium Isotope Geology, 188 pp. Springer, New York. FERAUD, J., FORNARI, M., GEFFROY, J. & LENCK, P. 1977. Min~ralisations arseni6es et ophiolites: le filon a realgar et stibine de Matra et sa place dans le district a Sb-As-Hg de la Corse alpine. Bull. B.R.G.M. 2, II, 91-112. FERGUSON, A. K. ~r CUNDARI, A. 1982. Feldspar
I78
S. C. Bergman
crystallization trends in leucite-bearing and related assemblages. Contrib. Mineral. Petrol. 18, 212-8. FERMOSO, M. L. 1967a. Composicion quimica de las sanidinas de las rocas lamproiticas espanolas. Estud. Geol. 23, 29-30. 1967b. E1 diopsido de las rocas volcanicas de Jumilla. Estud. Geol. 23, 31-3. FERNANDEZ, S . & HERNANDEZ-PACHECO, A. 1972. Lamproitic rocks of Cabezo Negro, Zeneta, Murcia. Estud. Geol. Inst. Invest. Geol. Lucas Mallada, 28, 267-76. FERRARA, G., LAURENZI, M. A., TAYLOR, H. P., TONARINI, S. & TURI, B. 1985. An oxygen and strontium isotope study of the Alban Hills volcanic complex, Italy. Earth planet. Sci. Left. 75, 1328. , MARTINEZ, M. P., TONARINI,S. & TURI, B. 1980. Oxygen and strontium isotopic correlation in volcanic rocks from Mr. Vulsini district (N. Latium, Italy). Int. Conf. on Geochronology, Cosmochronology and Isotopic Geochemistry, Extended Abstracts, pp. 95-6. U.S. Geol. Survey, Washington. FITZGERALD, W. V. 1907. Reports on portions of the Kimberleys (1905-6). W. Australia Parliamentary Paper 19. FODEN, J. D. 1983. The petrology of the calcalkaline lavas of Rindjani Volcano, East Sunda Arc; a model for island arc petrogenesis. J. Petrol. 24, 98130. FOLEY, S. F. 1982. Mineralogy, geochemistry, petrogenesis and structural relationships of the Aillik Bay alkaline intrusive suite, Labrador, Canada. M. Sc. Thesis, Memorial University Newfoundland (unpublished). - 1984. Liquid immiscibility and melt segregation in alkaline lamprophyres from Labrador. Lithos, 17, 127-37. - 1985. The oxidation state of lamproitic magmas. Tschermaks Mineral. Petrol. Mitt., 34, 217-38. , TAYLOR, W. R. & GREEN, D. H. 1986a. The role of fluorine and oxygen fugacity in the genesis of the ultrapotassic rocks. Contrib. Mineral. Petrol., 94, 183-92. , TAYLOR,W. R. & GREEN, D. H. 1986b. The effect of fluorine on phase relationships in the system KA1SiO4-Mg2SiO4-SiO2 at 28 kbar and the solution mechanism of fluorine in silicate melts. Contrib. Mineral. Petrol. 93, 46-55. , VENTURELLI,G., GREEN, D. H. & TOSCANI, L. in review. The ultrapotassic rocks: characteristics, classification and constraints for petrogenetic models. Fox, C. S. 1930. The Jharia Coalfield. Mem. Geol. Surv. India 56, 255. FRANKS, P. C. 1959. Pectolite in mica peridotite, Woodson County, Kansas. Am. Mineral. 44, 10826. 1966. Ozark Precambrian-Paleozoic relations: discussion of igneous rocks exposed in E. Kansas. Bull. Am. Assoc. Petrol. Geol. 50, 1035-42. --, BICKFORD, M. E. & WAGNER, H. C. 1971. Metamorphism of Precambrian granitic xenoliths in a mica peridotite at Rose Dome, Woodson
County Kansas. Part 2, Petrologic and mineralogic studies. Geol. Soc. Am. Bull. 82, 2869-90. FRANTSESSON,E. V. 1968. The Petrology of the Kimberlites, 194 pp. (Translated by D. A. Brown, Australian National University, Canberra, 1970.) FRASER, K. J., HAWKESWORTH, C. J., ERLANK, A. J., MITCHELL, R. H. & SCOTT-SMITH,B. H. 1985. Sr, Nd and Pb isotope and minor element geochemistry of lamproites and kimberlites. Earth Planet. Sci. Lett. 76, 57-70. FRIEND, C. R. L. & JANARDHAN, A. S. 1984. Mineral chemistry of coexisting phases from shonkinitic rocks, Salem, Tamil Nadu, India, Mineral. Mag. 48, 181-93. FROMAGET,J. 1933. Sur la presence de roches intrusives alkalines dans la zone charri6e neotriasique des plateaux calcaires du Tonkin occidental. C. R. Acad. Sci. Paris, 197, 651-3. FURNES, H., MITCHELL, J. G., ROBINS, B., RYAN, P. & SKJERLIE, F. J. 1982. Petrography and geochemistry of peralkaline, ultrapotassic syenite dykes of ' Middle Permian age, Sunnfjord, W. Norway. Norsk. Geol. Tidsskr. 62, 147-59. FUSTER, J. M. 1956. Las erupciones delleniticas del Torciario superior de la Fosa de Vera (provincia de Almeria). Estud. Geol. 20, 299-314. - & DE PEDRO, F. 1953. Estudio petrologico de las rocas volcanicas lamproiticas de Cabezo Maria (Almeria). Estud. Geol. Inst. Lucas Mallada Invest. Geol. Publ. 20, pp. 477-508. - & GASTESl, P. 1964. Estudio petrologico de las rocas lamproiticas de Barqueros (provincia de Murcia). Estud. Geol. 20, 299-331. , SAGREDO, J. & FERMOSO, M. L. 1967. Las rocas lamproiticas del S.E. de Espana. Estud. Geol. 23, 35-69. , IBARROLA, E. & LOBATO, M. P. 1954. Analysis quimacos de rocas espanolas publicados hosta 1952. C.S.LC. Monogr. Inst. Lucas Mallada Invest. Geol. 14. GALLAGHER, M. J. 1963. Lamprophyre dykes from Argyll. Mineral. Mag. 33, 415-30. GALLO, F. 1984. The Kamafugitic rocks of San Venanzo and Cuppaello, Central Italy, Neues .lb. Mineral. Monatsh. 115, 198-210. GARLICK, G. D. 1966. Oxygen isotope fractionation in igneous rocks. Earth planet. Sci. Lett. 1,361-8. GARLICK, H. J. 1979. Australian diamond prospects. The story so far. Indust. Miner. (Feb.) 137, 1729. GATES, O. 1959. Breccia pipes in the Shoshone ranges, Nevada. Econ. Geol. 54, 790-815. GEOLOGICAL SURVEY OF ITALY, 1961. Carta Geologico D'Italia (1/1,000,000 scale, 2 sheets), Rome. GEST, D. E. & MCBIRNEY, A. R. 1977. Genetic relations of shoshonitic and absarokitic magmas, Absaroka Mountains, Wyoming. J. Volcanol. geotherm. Res. 6, 85-104. GHOSE, C. 1949. A petrochemical study oflamprophyres and associated intrusive rocks of the Jharia coalfield. Quart. J. Geol. Min. Metall. Soc. India, 21, 133-47. GIARDINI, A. A. & MELTON, C. E. 1975a. The nature of
Lamproites and K-rich igneous rocks cloud-like inclusions in two Arkansas diamonds. Am. Mineral. 60, 931-3. -& -1975b. Chemical data on a colorless Arkansas diamond and its black amorphous CFe-Ni-S inclusions. Am. Mineral. 60, 934-6. & -1975c. Gases released from natural and synthetic diamonds by crushing under high vacuum at 200~ and their significance to diamond genesis. Fortschr. Mineral. 52, 455-64. , HURST, V. J., MELTON, C. E. & STORMER, J. C. 1974. Biotite as a primary inclusion in diamond: its nature and significance, Am. Mineral. 59, 7839. GILBERT, G. K. 1896. Laccolites in S. E. Colorado. J. Geol. 4, 816-25. GILL, J. B. 1981. Orogenic andesites and plate tectonics. Springer-Verlag, New York. GITTINS, J. 1979. The feldspathoidal alkaline rocks. In: YODER, H. S., Jr. (ed.) The Evolution of the Igneous Rocks, Fiftieth Anniversary Perspectives, pp. 35190. Princeton University Press, Princeton, NJ. , FAWCETT, J. J., BROOKS, C. K. & RUCKLIDGE, J. C. 1980. Intergrowths of nepheline-K-feldspar and kalsilite-K-feldspar: a re-examination of the pseudo-leucite problem. Contrib. Mineral. Petrol. 73, 119-26. , MACINTYRE, R. M. & YORK, D. 1967. The ages of carbonatite complexes in eastern Canada. Can. J: Earth Sci. 4, 651-5. GLAZNER, A. F. 1979. Geochemistry of ultrapotassic metasomatism in the Sleeping Beauty volcanic area, central Mojave Desert, California. Geol. Soc. Am. Abstr. Progm. 11, 79-80. & STORK, A. L. 1976. Ultrapotassic, silicarich Tertiary volcanic necks from the Mojave Desert, California. Geol. Soc. Am. Abstr. Progm. 8, 886. GOGINENI, S. V., MELTON, C. E. & GIARDINI, A. A. 1978. Some petrological aspects of the Prairie Creek diamond-bearing kimberlite diatreme, Arkansas. Contrib. Mineral. Petrol. 66, 251-61. GOLDSCHMIDT, V. M. 1933. Grundlagen der quantitativen Geochemie. Fortschr. Mineral. Krist. Petrog. 112-56. GOLES, G. G. 1967. Trace elements in ultramafic rocks. In: WYLLIE, P. J. (ed.) Ultramafic and related rocks, J. Wiley and Sons, New York, 352-62. GRAPES, R. H. 1975. Petrology of the Blue Mountain Complex, Marlborough, N . Z . J . Petrol. 16, 371428. GREGORY, G. P. 1969. Geochemical dispersion patterns related to kimberlite intrusives in N. America. PhD Thesis, Royal School of Mines, Imperial College, London. (Unpublished) - - & TOOMS,J. S. 1969. Geochemical prospecting for kimberlite. Q. J. Colorado School Mines, 64, 265305. GROZDANOV, L. 1979. Chemistry and genesis of biotite from lamproites in the vicinity of the village of Svidnya, district of Sofia (Bulgaria), Dokl. Bolg. Akad. Nauk, 32, 341-4. GRUBB, P. L. C. 1965. Undersaturated potassic lavas and hypabyssal intrusives in N. Johore. Geol. Mag. 102, 338-46. 1 7 ,
179
Gt2MBEL, C. W. VON, 1874. Die Paliiolithischen Eruptivgesteine des Fichtelgebirges, Franz, Munich. GUPTA, A. K. & YAGI, K. 1980. Petrology and Genesis of Leucite-bearing Rocks, Springer, New York. - - , YAGI, K., HARIYA, Y., ONUMA, K. 1976. Experimental investigation of some synthetic leucite rocks under water pressures. Proc. Jap. Acad. Sci. 52, 469-72. --, LEMAITRE, R. W., HAUKKA, M. T. & YAGI, K. 1983. Geochemical studies on the carbonated apatite glimmerites from Damodar Valley, India. Proc. Jap. Acad. Sci. 59B, 113-16. HAGGERTY, S. E. 1976. Opaque mineral oxides in terrestrial igneous rocks. In: RUMBLE, D. (ed.) Oxides, Mineral. Soc. Am. Short Course Notes, Vol. 3, pp. 101-300. 1983. The mineral chemistry of new titanates from the Jagersfontein Kimberlite, S. Africa: implications for metasomatism in the upper mantle. Geochim. cosmochim. Acta, 47, 1833-854. - - , SMYTH, J. R., ERLANK, A. J., R/CHARD, R. S. & DANCHIN, R. V. 1983. Lindsleyite (Ba) and mathiasite (K): two new chromium-titanates in the crichtonite series from the upper mantle. Am. Mineral. 68, 494-505. HALL, A. 1974. West Cornwall, Geol. Assoc. Guide Book 19. Banham & Co., Colchester, England. 39 PP. 1982. The Pendennis peralkaline minette. Mineral. MaR. 45, 257-66. HALL, A. E. & SMITH, C. B. 1985. Lamproite diamonds: are they different? In: GLOVER, J. & HARRIS,P. G. (ed.) Kimberlite Occurrence and Origin, Univ. W Australia Publ. 8, pp. 167-211. HARGRAVES, R. B. & ONSTOTT,T. C. 1980. Paleomagnetic results from some southern African kimberlites, and their tectonic significance. J. geophys. Res. 85, 3587-97. HARRIS, P. G. 1957. Zone refining and the origin of potassic basalts. Geochim. cosmochim. Acta, 12, 195-208. 1984. Kimberlite volcanism. In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin, Univ. I41.Australia Publ. 8, pp. 125-42. & MIDDLEMOST, E. m. 1969. The evolution of kimberlites. Lithos, 3, 77-88. HAWKESWORTH, C. J., ERLANK, A. J., MARSH, J. S., MENZlES, M. A. & VAN CALSTEREN, P. 1983. Evolution of tile continental lithosphere: evidence from volcanics and xenoliths in southern Africa. In: HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 11139. Shiva, Orpington. - - , FRASER,K. J. & ROGERS, N. W. 1986. Kimberlites and lamproites: extreme products of mantle enrichment processes. Spec. Pub. Geol. Soc. S. Afr. in press. HAWKINS, D. W. 1976. Emplacement, petrology and geochemistry of ultrabasic to basic intrusives at Allik Bay, Labrador. M. Sc. Thesis, Memorial University Newfoundland, St. Johns (unpublished). HAWTHORNE, J. B. 1975. Model of a kimberlite pipe. Phys. Chem. Earth, 9, 1-16. -
-
-
-
-
-
-
-
I
8
0
S. C.
Bergman
HEALD, K. C. 1926. Geology of the Ingomar anticline, Treasure and Rosebud counties, Montana. U.S. Geol. Survey Bull. 786, 1-37. HEARN, B. C. Jr. 1968. Diatremes with kimberlitic affinities in north-central Montana. Science 159, 622-5. 1979. Preliminary map o f diatremes and alkalic ultramatic intrusions, Missouri River Breaks and vicinity, North Central Montana. U.S. Geol Surv. Open File Rep. 79-1128. & MCGEE, E. S. 1982. Garnets in Montana diatremes: a key to prospecting for kimberlites. U.S. Geol. Surv. Open File Rep. 82-722. & -1984. Garnet peridotites from Williams kimberlites, north-central Montana, U.S.A. In: KORNPROBST, J. (ed.) Kimberlites II: The Mantle and Crust-Mantle Relationship, pp. 55-70. Elsevier, Amsterdam. HEINRICH, E. W. 1966. The Geology of Carbonatites, 555 pp. Rand McNally, New York. 555 pp. & MOORE, D. G. Jr. 1970. Metasomatic potash feldspar rocks associated with igneous alkalic complexes. Can. Mineral. 10, 571-84. HELMSTAEDT,H. & GURNEY, J. J. 1984. Kimberlites of southern Africa--are they related to subduction processes ? In: KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 425-34. Elsevier, Amsterdam. HENAGE, L. F. 1972. A definitive study of the origin of lamproites. M.S. Thesis, University of Oregon (unpublished). HERNANDEZ-PACHECO, F. 1935. Estudio fisiogral~hico y geologico del territorio comprendido entre Hellin y Cieza. An. Univ. Madrid (Cienc.), 4, 1-38. 1965. Una richterita potasica de rocas volcanicas alcalinas, Sierra de las Cabras (Albacete). Estud. Geol. 20, 265-70. HERTOGEN, J., LOPEZ-RUIZ, J., RODRIGUEZ-BADIOLA, E., DEMAIFFE, D. & WEIS, D. 1985. Petrogenesis of ultrapotassic rocks from S.E. Spain: trace elements and Sr-Pb isotopes. Geol. Assoc. Can. Progm. Abstr. 10, A26. HIGAZY, A. 1954. Trace elements of volcanic and ultrabasic potassic rocks of south-western Uganda and adjoining part of Belgian Congo. Geol. Soc. Am. Bull. 65, 39-70. HINTZE, L. F. 1980. Geological map of Utah (1/500,000 scale, 2 sheets). Utah Geol. Mineral. Surv., Salt Lake City. HOBSON, B. 1892. On the basalts and andesites of Devonshire, known as 'feldspathic traps'. Q. J. Geol. Soc. 48, 496-507. HOEFS, J. & WEDEPOHL, K . H . 1968. Strontium isotope studies on young volcanic rocks from Germany and Italy. Contrib. Mineral. Petrol. 19, 328-38. HOFMANN, A. W., WHITE, W. M. & WHITFORD, D. J. 1978. Geochemical constraints on mantle models: the case for a layered mantle. Yb. Carnegie lnstn. Wash. Yb. 77, 54-61. HOLM, P. M. & MUNKSGAARD, N. C. 1982. Evidence for mantle metasomatism: an oxygen and strontium isotope study of the Vulsinian district, central Italy. Earth planet. Sci. Lett. 60, 376-88.
-
-
-
-
-
-
-
-
-
-
HOLMES, A. 1932. The origin of igneous rocks. Geol. Mag. 69, 543-58. 1950. Petrogenesis of katungite and its associates. Am. Mineral. 35, 772-92. & HARWOOD,H. F. 1932. Petrology of the volcanic fields east and southeast of Ruwenzori, Uganda. Q. J. Geol. Soc. 88, 370-442. -& -1937. The volcanic area of Bufumbira. Pt II. Mere. Geol. Surv. Uganda, 3. HOLUB, F. V. 1977. Petrology of inclusions as a key to petrogenesis of the Durbachitic rocks from Czechoslovakia. Tschermaks Mineral. Petrogr. Mitt. 24, 133-50. HUNZIKER, J. C. 1974. Rb-Sr age determination and the alpine tectonic history of the western Alps. Mem. Ist. Geol. Min. Univ. Padova, 31, 1-54. HURLBUT, C. S. & GRIGGS, D. 1939. Igneous rocks of the Highwood Mountains, Montana: Part I, the Laccoliths. Geol. Soc. Am. Bull. 50, 1043-112. HURLEY, P. M., FAIRBAIRN, H. W. & PINSON, W. H. 1966. Rb-Sr isotopic evidence in the origin of potash-rich lavas of western Italy. Earth planet. Sci. Lett. 1,301-6. INNOCENTI, F., MANETTI,P., MAZZUOLI,R., PASQUARE, G. & VILLARI,L. 1982. Anatolia and NW Iran. In: THORPE, R. S. (ed.) Andesites, pp. 327-49. Wiley, London. JANSE, A. J. A. 1984. Kimberlites--where and when. In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin. A Basis for Conceptual Models in Exploration. Univ. W. Australia Publ. 8, pp. 19-62. Perth, W. Australia. JAQUES, A. L. 1983. The ultrapotassic rocks of the West Kimberley region, W. A. : products of deep-seated intraplate volcanism. Proc. 6th Geological Society of Australia Cony., Vol 9. Geological Society of Australia, Sydney. -& FOLEY, S. F. 1985. The origin of Al-rich spinel inclusions in leucite from the leucite lamproites of western Australia. Am. Mineral., 70, 1143-50. GREGORY, G. P., LEWIS, J. D. & FERGUSON, J. 1982. The ultra-potassic rocks of the W. Kimberley region, W. Australia and a new class of diamondiferous kimberlite. Terra Cognita, 2, 251-2. --, LEWIS, J. D., GREGORY, G. P., FERGUSON, J., SMITH, C. B., CHAPPELL, B. W. & MCCULLOCH, M. T. 1983. The ultrapotassic, diamond-bearing rocks of the West Kimberley region, Western Australia. Proc. 6th Geological Society of Australia Cony., Vol. 9, pp. 286-7. Sydney. & 1984a. The , , , , , diamond bearing ultrapotassic (lamproitic) rocks of the West Kimberley region, Western Australia. In : KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 225-54. Elsevier, Amsterdam. -
-
-
-
-
-
& SMITH, C. B. 1986. T h e k i m b e r l i t e s a n d
lamproites of Western Australia. Geol. Surv. Western Australia Bull. 132, Perth, W.A. --, WEBB, A. W., FANNING, C. M., BLACK, L. P., PIDGEON, R. T., FERGUSON, J., SMITH, C. B. & GREGORY, G. P. 1984b. The age of the diamond bearing pipes and associated leucite lamproites of
Lamproites and K-rich igneous rocks the west kimberley region, western Australia. Bur. Mineral. Res. Aust. Geol. Geophys. 9, 1-7. JENSEN, H. I. 1908. The distribution, origin and relationships of alkaline rocks. Proc. Linn. Soc. N.S.W. 33, 491-588. JEREMINE, E. & FALLOT, P. 1929. Sur la pr6sence d'une vari6t6 de jumillite aux environs de Calasparra (Murcia). C. R. Acad. Sci. Paris, 188, 800. JEWETT, J. M. 1964. Geologic map of Kansas (1/ 500,000 scale). State Geol. Surv. Kansas. JOHANSSEN,A. 1931. A Descriptive petrography of Igneous Rocks, Vol. 1, University of Chicago Press, Chicago, IL. -1933. A Descriptive Petrography of Igneous Rocks, Vol. 2, University of Chicago Press, Chicago, IL. -1935. A Descriptive Petrography of Igneous Rocks, Vol. 3, University of Chicago Press, Chicago, IL. -1939. A Descriptive Petrography of Igneous Rocks, Vol. 4, University of Chicago Press, Chicago, IL. JOHNSTONR. H. 1959. Geology of the Northern Leucite Hills, Sweetwater County, Wyoming. M. A. Thesis, University of Wyoming, Laramie, WY (unpublished). JONES, A. P. & SMITH, J. V. 1983. Petrological significance of mineral chemistry in the Agathla Peak and the Thumb minettes, Navajo volcanic field. J. Geol. 91,643-56. - & 1985. Phlogopite and associated minerals from Permian minettes in Devon, S. England. Bull. Geol. Soc. Finland, 57, 89-102. JOPLIN, G. A. 1965. The problem of the potash-rich basaltic rocks. Mineral. Mag. 34, 266-75. - 1968. The shoshonite association: a review. J. Geol. Soc. Aust. 15, 275-94. KAPLAN, G., FAURE, D., ELLOY, R. & HEILAMMER,R. 1967. Contribution ~ l'6tude de l'origine des lamproites. Bull. Centre Rech. PA VSNPA, l, 1539. KAY, R. W. & GAST, P. W. 1973. The rare earth content and origin of alkali-rich basalts. J. Geol. 81, 65382. KAY, S. M., KAY, R. W., HANGAS, J. & SNEDDEN, Z. 1978. Crustal xenoliths from potassic lavas, Leucite hills, Wyoming. Geol. Soc. Am. Abstr. Progm. 7, 432. KELLER, J. 1983. Potassic lavas in the orogenic volcanism of the Mediterranean area. J. Volc. Geoth. Res. 1 8 , 321-35. KEMP, J. F. 1897. The Leucite Hills of Wyoming. Geol. Soc. Am. Bull. 8, 169-82. - - & KNIGHT, W. C. 1903. Leucite Hills of Wyoming. Geol. Soc. Am. Bull. 14, 305-36. KENNEDY, G. C. 1955. Some aspects of the role of water in rock melts. Geol. Soc. Am. Spec. Pap. 62, 489503. KNIGHT, G. L. & LANDES,K. K. 1932. Kansas laccoliths. J. Geol. 40, 1-15. KNILL, D. C. 1969. The Permian igneous rocks of Devon. Bull. Geol. Surv. Gt. Br. 29, 115-38. -1982. Permian volcanism in Southwestern England, In : SUTHERLAND,D. S. (ed.) Igneous Rocks of the British Isles, pp. 329-32. Wiley, London. KNOPF, D. 1970. Les Kirnberlites et les Roches Apparentes de Cote d'Ivoire. Sodemi, Abidjan.
181
KOGARKO, L. N. 1980. Significance of chemical bond polarity in the geochemistry of magmatism. Geokhimiya, 9, 1286-97. , KRIGMAN, L. D. & SHARUDILO, N. S. 1968. Experimental investigations of the effect of the alkalinity of silicate melts on separation of fluorine into the gas phase. Geokhimiya, 8, 948-56. KOSTYUK, V. P. 1983. The potassic alkalic magmatism of the Barkal Aldan Belt. Soy. Geol. Geophys. 24, 31-38. KRAMERS, Ji D., RODDICK, J. C. i . & DAWSON, J. B. 1983. Trace element and isotope studies on veined, metasomatic and " M A R I D " xenoliths from Bultfontein, S. Africa. Earth planet Sci. Lett. 65, 90106. KRANCK, E. 1953. Bedrock geology of the seaboard of Labrador between Domino and Hopedale. Bull. Can. Geol. Surv. 26. KRAVCHENKO, S. M., VLASOVA,E. V. & PINEVICH, V. S. 1960. Batisite, a new mineral, Dokl. Akad. Nauk S.S.S.R. 133, 657-8. KRESTEN, P., PRINTZLAU,1., REX, D., VARTIAINEN,H. & WOOLLEY, A. 1977. New ages of carbonatic and alkaline ultramafic rocks from Sweden and Finland. Geol. Fbren Stockh. Fbrh. 99, 62-65. KRUMMENACHER, D. & EVERNDEN, J. F. 1960. D6termination d'gtge isotopique fautes sur quelques roches des Alpes par le m~thode K-At. Schweiz. Mineral. Petrogr. Mitt. 40, 267-77. KUEHNER, S. M. 1980. Petrogenesis of ultrapotassic rocks, Leucite Hills, Wyoming. M. Sc. Thesis, University of Western Ontario, London, Ontario (unpublished). - - , EDGAR, A. D. & ARIMA, M. 1981. Petrogenesis of the ultrapotassic rocks from the Leucite Hills, Wyoming. Am. Mineral. 66, 663-7. KUSHIRO, I. & AOKI, K. 1968. Origin of some eclogite inclusions in kimberlite. Am. Mineral. 53, 1347-67. KYSER, T. K., O'NEIL, J. R. & CARMICHAEL,I. S. E. 1981. Oxygen isotope thermometry of basic lavas and mantle nodules. Contrib. Mineral. Petrol. 77, 11-23.
& - 1982. Genetic relations among basic lavas and ultramafic nodules: Evidence from oxygen isotope compositions. Contrib. Mineral. Petrol. 81, 88-102. LACROIX, A. 1926. La syst6matique des roches leucitiques: les types de la famille syenitique. C. R. Acad. Sci. Paris, 182, 597-601. 1933a. Contribution /t la connaissance de la composition chimique et mineralogique des roches 6ruptives de l'Indochine. Bull. Serv. Geol. Indochine, 20, 3, 1-208. - 1933b. Les roches 6ruptivespotassiquesleucitiques ou non du Tonkin occidental. C. R. Acad. Sci. Paris, 197, 625-7. LANGFORD, R. E. 1973. A study of the origin of Arkansas diamonds by mass spectrometry. PhD Thesis, University of Georgia, Athens, GA (unpublished). LANGWORTHY,A. P. & BLACK,L. P. 1978. The Mordor complex: a highly differentiated potassic intrusion with kimberlitic affinities in central Australia. Contrib. Mineral. Petrol. 67, 51-62.
I82
S. C.
Bergman
LARSEN, E. S. 1940. Petrographic province of central Montana. Geol. Soc. Am. Bull. 51,887-948. , HURLBUT, C. S., BURGESS, C. H. & BULL, B. F. 1941. Igneous rocks of Highwood Mountains, Montana. Geol. Soc. Am. Bull. 52, 1857-68. LARSEN, L. M., REX, D. C. & SECHER, K. 1983. Age of carbonatites, kimberlites and lamprophyres from southern W. Greenland: recurrent alkaline magmatism during 2500 million years. Lithos, 16, 21521. LARSEN, W. N. 1954. Precambrian geology of the W. Uinta Mountains, Utah, M.S. Thesis, University of Utah, Salt Lake City, UT (unpublished). LASSERRE, J. L. 1976. Amphibolites and alkaline gneisses in the midland formations of Nepal; petrography, geochemistry-geodynamic involvements. In: Jest, C. (ed.) Himalaya ; sciences de la terre, Fr. Cent. Natl. Rech. Sci., Colloq. Int. 268, 2, 213-36. LEES, G. J. 1974. Petrochemistry of the mica lamprophyres (minettes) of Jersey (CI). Proc. Ussher Soc. 3, 149-55. LEMAITRE, R. W. 1976. The chemical variability of some common igneous rocks. J. Petrol. 17, 589637. LEWIS, H. C. 1887. On a diamondiferous peridotite and the genesis of the diamond. Geol. Mag. 4, 22-4. LEWIS, P. H. 1973. Mica pyroxenite inclusions in limburgite, Hopi Buttes Volcanic field, Arizona. unpubl. M. Sci. Thesis, Brigham Young Univ., Utah. LEWIS, R. D. 1977. Mineralogy, petrology and geophysical aspects of Prairie, Creek kimberlite, near Murfreesboro, Arkansas. M. S. Thesis, Department of Geology, Purdue University (unpublished). - - , MEYER, H. O. A., BOLIVAR,S. L. & BROOKINS,D. B. 1976. Mineralogy of the diamond bearing "kimberlite", Murfreesboro, Arkansas LOS, 57, 761 (abstract only). LIBBY, S. C. 1975. Origin of K-ultramafic rocks in the Enoree 'vermiculite' districts, S. Carolina. PhD Thesis, Pennsylvania State University (unpublished). LLOYD, F. E., ARIMA, M. & EDGAR, A. D. 1985. Partial melting of a phlogopite clinopyroxenite nodule from south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift volcanics. Contrib. Mineral. Petrol. 91,321-9. LOPEZ RUIZ, J. & RODRIGUEZ BADIOLA, E. 1980. La region volcanica Neogena del Sureste de Espana. Estud. Geol. 36, 5-63. LOVE, J. D., WEITZ, J. L. & HOSE, R. K. 1955. Geological map of Wyoming (1/500,000 scale) U.S. Geol. Surv., Washington. LUHR, J. F. & CARMICHAEL,I. S. E. 1980, The Colima volcanic complex, Mexico I. Post caldera andesites from Volcan Colima. Contrib. Mineral. Petrol. 71, 343-72. -& -1981. The Colima volcanic complex, Mexico: I1. Late Quaternary cinder cones. Contrib. Mineral. Petrol. 76, 127-47. LUTH, W. C. 1967. Studies in the system KA1SiO4Mg2SiO,-SiO2-H20. I: Inferred phase relations
and petrologic applications. J. Petrol. 8, 372416. MACDOWELL, F. W. 1966. Potassium-Argon dating of Cordilleran intrusives. PhD Thesis, Columbia University, New York (unpublished). MCCANDLESS,T. E. 1982. The mineralogy, morphology and chemistry of detrital minerals of a kimberlitic and eclogitic nature, Green River Basin, Wyoming. M. S. Thesis. University of Utah, Salt Lake City, UT (unpublished). - 1984. Detrital minerals of mantle origin in Green River Basin, Wyoming. Soc. Min. Eng. AIME, Preprint 84-395. MCCLURE, W. C. 1963. Geology of the Enoree area: S. Carolina. M.Sc. Thesis, University of South Carolina, Columbia, SC (unpublished). MCCULLOCH, M. T., JAQUES, A. L., NELSON, D. R. & LEWIS, J. D. 1983a. Nd and Sr isotopes in kimberlites and lamproites from W. Australia: an enriched mantle origin. Nature, Lond. 302, 400-3. , NELSON, D. R., JAQUES, A. L. & LEWIS, J. O. 1983b. Nd and Sr isotopic systematics in kimberlites and lamproites from West Kimberley, Western Australia. Proc. 6th Geological Society o f Australia Cony., Vol. 9, pp. 287-8. Geological Society of Australia, Sydney. McKIE, O. 1966. Fenitization. In: TUTTLE, C. F. & GITTINS, J. (eds) Carbonatites, pp. 261-94. WileyInterscience, New York. MADIGAN, R. 1983. Diamond exploration in Australia. Indiaqua, 35, 27-38. MAKHERJEE, K. K. 1961. Petrology of lamprophyres of the Bokaro coalfield. Qt. J. Geol. Min. Metall. Soc. India, 33, 69-87. MAKSlMOV, E. P. 1982. The Mesozoic magmatism of the Aldan shield as an indicator of its tectonic regime. Soy. Geol. Geophys. 23, 9-14. MANSON, V. 1967. Geochemistry of basaltic rocks: major elements. In : HESS, H. H., & POLDERVAART, A. (eds) Basalts. Wiley, New York, 215-70. MARINELLI, G. & MITTEMPERGHER, M. 1966. On the genesis of some magmas of typical Mediterranean (potassic) suite. Bull. Volcanol. 29, 113-40. MARTHUR, S. M. & SINGH, H. N. 1971. Petrology of the Majgawan pipe rock: Geol. Surv. India Misc. Publ. 19, 78-85. MARVIN, R. F., HEARN, B. C., JR., MEHNERT, H. H., NAESER, C. W., ZARTMAN,R. E. & LINDSEY,D. A. 1980. late Cretaceous-Paleocene igneous activity in north-central Montana. Isochron/West, 29, 525. MASON, B. 1967. Meteorites. Am. Sci. 55, 429-55. - 1977. Elemental distribution in minerals from the Wolgidee Hills intrusion, Western Australia. N. Z. DSIR Bull. 218, 114-20. -1979. Cosmochemistry, Part I, Meteorites. Prof. Pap. U.S.G.S. 440B-1. MATSON, R. E. 1960. Petrography and petrology of Smoky Butte Intrusives, Garfield County Montana. M.S. thesis, Montana State University, Missoula, MT (unpublished). MEAKINS, A. 1983. Geology and genesis of the Argyle alluvial diamond deposits, Kimberley region, W. Australia. In." DAVY, R., BUTTE, C. R. M. &
Lamproites and K-rich igneous rocks BALLINGER,T. A. (eds) Geochemical Exploration in Arid and Deeply Weathered Environments. Geochemistry and Genesis of Ore Deposits Associated with Weathering, Aust. Reg. Meet., Assoc. Expl. Geochem. Perth, Abstracts volume. 54-6. MELTON, C. E. & GIARDINI, A. A. 1976. Experimental evidence that oxygen is the principal impurity in natural diamond. Nature, Lond. 263, 309--10. - & -1980. The isotopic composition of Ar included in an Arkansas diamond and its significance. Geophys. Res. Lett. 7, 461-4. --, Salotti, C. A. & GIARDINI, A. A. 1972. The observation of N2, H20, CO2, CH4 and Ar as impurities in natural diamonds. Am. Mineral. 57, 1518-23. MENZIES, M. & MURTHY, V. R. 1980a. Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature, Lond. 283, 634-6. - - & - 1980b. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts: implications for local heterogeneities in metasomatically veined mantle. Earth planet. Sci. Lett. 46, 323-34. MERRILL, R . B . , B I C K F O R D , M . E . & I R V I N G , A . J . 1977. The Hills Pond Peridotite, Woodson County, Kansas: A richterite-bearing Cretaceous intrusive with kimberlitic affinities. Extended Abstr. 2nd Kimberlite Conf., Santa Fe, NM. MESEGUER PARDO, J. 1924. Estudio de los yacimlentos de azufre de las provincias de Murcia y Albacete. Bol. Inst. Geol. Esp. 45, 131-214. MEYER, H. O. A. 1976. Kimberlites of the continental United States: a review. J. Geol. 84, 377-403. & MITCHELL, R. H. 1985. Sapphire-bearing lamprophyre from Yogo Gulch, Montana. Geol. Assoc. Can. progm, abstr. 10, A39. --, LEWIS, R. D., BOLIVAR, S. & BROOKINS, D. G. 1977. Prairie Creek Kimberlite, Murfreesboro, Pike County, Arkansas. Extended Abstr. 2nd Int. Kimberlite Conf., Santa Fe, NM. MIDDLEMOST,E. A. K. 1975. The basalt clan. Earth Sci. Rev. 11,337-64. MILLER,J. A. & MOHR, P. A. 1964. K - A r measurements on the granites and some associated rocks from Southwest England. Geol. J. Liverpool, 4, 105-26. MILLER, K. C. 1982. Paleomagnetism of Indian Kimberlites. B. A. Thesis, Princeton University, Princeton, NJ (unpublished). MILLER, T. P. 1970. Petrology of the plutonic rocks of west-central Alaska. U. S. Geol. Surv. Open File Rep. .545. - 1972. Potassium-rich alkaline intrusive rocks of Western Alaska. Geol. Soc. Amer. Bull. 83, 211128. MINEYEVA, I. G. 1972. Principal geochemical characteristics of potassic alkalic rocks. Geochem. Int. 9, 173-9. MISER, H. D. & PURDUE, A. H. 1929. Geology of the De Queen and Caddo Gap quandrangle, Arkansas. U.S. Geol. Survey Bull. 808, 195. - & Ross, C. S. 1923. Diamond-bearing peridotite in Pike County, Arkansas. U. S. Geol. Survey Bull. 735--1,279-322. MIrCHtELL, R. H. 1981. Titaniferous phlogopites from
183
the leucite lamproites of the W. Kimberley Area, W. Australia. Contrib. Mineral. Petrol. 76, 243-51. - 1985. A review of the mineralogy of lamproites. Trans. Geol. Soc. SA, 88, 411-437. - & BELL, K. 1976. Rare earth element geochemistry of potassic lavas from the Birunga and ToroAnkole regions of Uganda, Africa. Contrib. Mineral. Petrol. 58, 293-303. - & HAWKESWORTH, C. J. 1984. Geochemistry of potassic lavas from Smoky Butte, Montana. Geol. Soc. Am. Abstr. Progm. 16, 597. - & Lewis, R. D. 1983. Priderite-bearing xenoliths from the Prairie Creek mica peridotite, Arkansas. Can. Mineral. 21, 59-64. MODRESKI, P. J. & BOETTCHER,A. L. 1972. The stability of phlogopite + enstatite at high pressures: a model for micas in the interior of the earth. Am. J. Sci. 272, 852--69. -& 1973. Phase relationships of phlogopite in the system K 2 0 - M g O - C a O - A 1203-SIO2-H20 to 35 kilobars : a better model for micas in the interior of the earth. Am. J. Sci. 273, 385~414. MOODY, C. L. 1949. Mesozoic igneous rocks of Northern Gulf Coast plain, Bull. Am. Assoc. Petrol. Geol. 33, 1410-28. MORRIS, E. C. 1953. Geology of the Big Piney area, Summit County, Utah. M.S. Thesis, University of Utah, Salt Lake City, UT (unpublished). MORRISON, G. W. 1980. Characteristics and tectonic setting of the shoshonite rock association. Lithos, 13, 97-108. MULLEN, E. D. & MURPHY, S. 1985. Mineralogy and petrology of lamprophyric and carbonatite dikes, central Arkansas. Geol. Soc. Am. Abstr. Progm. 17, 185. - & PETTY, W. B. 1985. Petrologic relations among syenites and lamprophyric rocks, Arkansas alkalic Province. Geol. Assoc. Can. Progm. Abstr. 10, A42. MURTY, M. K. 1980. Tectonics and geochemistry of the diamondiferous kimberlites of the Jungle Valley, Merzapur District, U.P.--an evolutionary model. Proc. Conf. on Diamonds, Geological Survey of India, Bombay. MURTY, Y. G. K., RAO, M. G., MISRA, R. C. & AJIT KUMAR REDDY, T. 1980. Kimberlite diatremes of Andhra Pradesh--their assessment and search for concealed bodies. Prof. Conf. on Diamonds, Geological Survey of India, Bombay. MUTCI-ILER, E. E., ROUGHON, D. J. & LAVIN, O. P. 1976. PETROS: a data bank of major-element chemical analyses of igneous rocks for research and teaching. Comput. Geol. Sci. 2, 51-7. MYSEN B. O. & BOETTCHER,A. L., 1975a. Melting of a hydrous mantle: I. Phase relations of natural peridotite at high pressures and temperatures with controlled activities of water, carbon dioxide, and hydrogen. J. Petrol. 16, 520--48. - & -1975b. Melting of a hydrous mantle: II. Geochemistry of crystals and liquids formed by anatexis of mantle peridotite at high pressures and high temperatures as a function of controlled activities of water, hydrogen, and carbon dioxide. J. Petrol. 16, 549-93. NELSON, D. R., MCCULLOCH,M. T. & SUN, S.-S. 1986.
184
S. C. Bergman
The origins of ultrapotassic rocks as inferred from Sr, Nd and Pb isotopes. Geochim. cosmochim. Acta, 5 0 , 231-46. N~MEC, D. 1972. Micas of the lamprophyres of the Bohemian Massif Neues Jb. Mineral. Abh. 117, 196-216. 1973. Amphibole der lamprophyrischen und lamproiden Gestine in Ostfeil der Bohmischen Masse. Cas. Mineral. Geol. 18, 31-46. -1974. Petrochemistry of the dyke rocks of the Central Bohemian Pluton. Neues J. Mineral. Monatsh, 193-209. - 1985. Ba and its carriers in dyke rocks of the minette series. Geol. Assoc. Canada Progm. Abstr. 10, A43. NICHOLLS, J. & CARMICHAEL, I. S. E. 1972. The equilibrium temperature and pressure of various lava types with spinel- and garnet-peridotites. Am. Mineral. 57, 941-59. NIELSEN, B. L., 1976. Economic minerals diamonds, olivine, garnets in Greenland. In. ESCHER, A. & WATT, W. S. (eds) Geology of Greenland, pp. 4816. Geological Survey of Greenland, Copenhagen. NIGGLI, P. 1923. Gesteins- undMineralprovinzen. Gebruder Borntraeger, Berlin. NIXON, P. H., THIRWALL, M. F. & BUCKLEY F. 1982. Kimberlite-lamproite consanguinity. Terra Cognita 2, 252-54. , , -& DAVIES, C. J. 1984. Spanish and Western Australian lamproites: aspects of whole rock geochemistry. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 285-96. Elsevier, Amsterdam. NOBEL, F . A . , A N D R I E S S E N , P . A . M . , H E B E D A , E . H . PRIEM, H. N.A. & RONDEEL, H. E. 1981. Isotopic dating of the post-Alpine Neogene volcanism in the Betic Cordilleras, southern Spain. Geol. Mijnbouw, 60, 209-14. NOCKOLDS, S. R. 1940. Petrology of rocks from Queen Mary Land. Australian Antarctic Expedition 191114, Rep. Ser. A, IV, Part 2, pp. 15-86. - 1954. Average chemical composition of some igneous rocks. Geol. Soc. Am. Bull. 65, 1007-32. NOE-NYGAARD, A. & RAMBERG, H. 1956. Geological reconnaisance map of the country between latitudes 69~ and 63~ West-Greenland, (1/500,000 scale). Geological Survey of Greenland, Copenhagen. -& -1961. Geological reconnaissance map of the country between latitudes 69~ and 63~ W. Greenland. Meddr. Gronland, 123, 3. NORRISH, K. 1951. Priderite, a new mineral from the leucite-lamproites of the West Kimberley area, Western Australia. Mineral Mag. 24, 496-501. NOUGIER, J. 1970. Contribution ~ l'6tude g6ologique et g6omorphologique des iles Kerguelen (Territoire des Terres Australes et Antarctiques Fran~aises) Com. Natl. Fr. Rech. Antarct., 27, 696 pp. O'BRIEN, H. E., IRVING, A. J. & MCCALLUM,I. S. 1985. Complex zoning of clinopyroxene in shonkinites and mafic phonolites, Highwood Mtns. Montana: evidence for periodic mixing with a K-rich basanitic magma. Geol. Soc. Am. Abstr. Progm. 17, 187.
OGDEN, P. R., Jr. 1979. The geology, major element geochemistry and petrogenesis of the Leucite Hills volcanic rocks, Wyoming. PhD Thesis, University of Wyoming, Laramie, WY (unpublished). --, SPERR, J. T., GUNTER, W. D. & PAJARI, G. E., JR. 1978. Morphology of a recent ultrapotassic volcanic field, Leucite Hills, SW Wyoming. Geol. Soc. Am. Abstr. Progm. 10, 140. --, VOLLMER, R. & SCHILLING, J. G. 1980. Leucite Hills revisited. LOS, 61,412. OSANN, A. 1889. Beitrage Zue Kentniss der Eruptivegesteine des Cabo de Gata I. Z. dch. Geol. Ges. 41, 297-311. 1906. Uber einige Alkaligestine aus Spanien, Festchritte, Rosenbusch, Stuttgart. OSBORNE, F. F. & ROBERTS, E. J. 1931. Differentiation in the Shonkin Sag laccolith, Montana. Am. J. Sci. 5th ser. 22, 331-53. OWEN H. G. 1983. Atlas of continental displacement, 200 million years to the present. Cambridge Univ. Press, Cambridge. OYAWOVE, M. O. 1976. The origin and evolution of the potassic rocks of SW Nigeria. 25 Int. Geological Congr. Abstr., Vol. 21, 10A, pp 424-5. PAGEL, M. & LETERRIER, J. 1980. The subalkaline potassic magmatism of the Ballous Massif (S. Vosges, France); shoshonitic affinity. Lithos 13, 110. PARGA PONDAL, I. 1935. Quisismo de las manifestaclones magmaticas Cenozoicas de la Peninsula Iberica. Trab. Mus. Nac. C. Nat. Ser. Geol. 39. PAUL, O. K., REX, D. C. & HARRIS, P. G. 1975. Chemical characteristics and K-Ar ages of Indian kimberlites. Geol. Soc. Am. Bull. 86, 364-6. PELLICER, M. J. 1973. The petrology and geochemistry of new lamproite rocks near Aljorra, Murcia, Spain. Eustud. Geol. Inst. Invest. Geol. Lucas Mallade, 29, 99-106. PHILLIPI, E. 1912. Geologische Beschreibung des Gassbergs. Deutsche Siidpolar Expedition 19011903. Geogr. Geol. l l , 47-71. PIRSSON, L. V. 1905. Peti-ography and geology of the igneous rocks of the Highwood Mountains, Montana. Bull. U.S. Geol. Surv. Bull. 237. POLANSKI, A. 1949. The alkaline rocks of the EastEuropean Plateau. Bull. Soc. Amis. Sci. Lett. Poznan, Ser. B, 10,119-84. POLDERVAART,A. 1955. Chemistry of the Earth's crust. Geol. Soc. Am. Spec. Pap. 62, 119-44. POWELL, J. L. & BELL, K. 1970. Strontium isotopic studies of alkalic rocks. Localities from Australia, Spain, and the Western United States. Contrib. Mineral. Petrol. 27, 1-10. - & 1974. Isotopic composition of Sr in alkalic rocks. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp 412-21. Wiley, New York. PRIDER, R. T. 1939. Some minerals from the leuciterich rocks of the West Kimberley area, Western Australia. Mineral. Mag. 25, 373-87. - 1960. The leucite lamproites of the Fitzroy Basin, Western Australia. J. Geol. Soc. Aust. 6, 71-118. - 1965. Noonkanbahite, a potassic batisite from the lamproites of Western Australia. Mineral. Mag. 34, 403-5.
185
Lamproites and K-rich igneous rocks 1982. A glassy lamproite from the West Kimberley area, Western Australia. Mineral. Nag. 45, 27982. - & COLE, W. F. 1942. the alteration products of olivine and leucite in the leucite-lamproites from the West Kimberley area, Western Australia. Am. Mineral. 27, 373-84. PRIMAL, L. 1963. Etude g6ologique et m6tallog6nique de la partie m6ridional du cap Corse. PhD Thesis, University of Paris (unpublished). PROSTKA, H. J. 1973. Hybrid origin of the absarokiteshoshonite-banakite series, Absaroka volcanic field, Wyoming. Geol. Soc. Am. Bull. 84, 697-702. PRYCE, M. W., HODGE, L.C. & CRIDDLE, A. J. 1984. Jeppeite, a new K-Ba-Fe titanate from Walgidee Hills, Western Australia. Mineral. Nag. 48, 2636. RADULESCU, D. P. 1966. Rhyolites and secondary ultrapotassic rocks in the subsequent Neogene volcanism from the E. Carpathians. Bull. Volcanol. 29, 425-34. RAVlCH, M. G. & KRYLOV, A. H. 1964. Absolute ages of rocks from E. Antarctica. In." ADIE, R. J. (ed.) Antarctic Geology, pp. 578-89. North-Holland, Amsterdam. REINISCH, R. 1912. Petrographische-Beschreibungder Gaussberg-Gesteins. Deutsch Sfidpolar-Expedition 1901-03. Geogr. Geol. 11, 73-8. RIOU, R., DuPuY, C. & DOSTAL, J. 1981. Geochemistry of co-existing alkaline and calc-alkaline volcanic rocks from Northern Azerbaijan (N.W. Iran). J. Volc. Geoth, Res. 11,253-75. RITTMANN,A. 1933. Die geologisch bedingte Evolution und Differentiation des Somma-Vesuvius magma. Z. Vulkanol. 5, 8-94. -1951. Magmatic character and tectonic position of the Indonesian volcanoes. Nomenclature of volcanic rocks. Bull. Volcanol. 12, 75-109. ROBINS, B., FURNES, H. & RYAN, P. 1983. A new occurrence of dalyite. Mineral. Nag. 47, 93-4. ROBINSON, E. C. 1976. Geochemistry of lamprophyric rocks of the eastern Ouachita Mtns, Arkansas. M.S. Thesis, University of Arkansas, Fayetteville, AR (unpublished). ROCK, N. M. S. 1976. The comparative Sr isotopic composition of alkaline rocks; new data from South Portugal and East Africa. Contrib. Mineral. Petrol. 56, 205-28. 1977. The nature and origin of lamprophyres: some definitions, distinctions and derivations. Earth Sci. Rev. 13, 123-69. - 1984. Nature and origin of calc-alkaline lamprophyres: minettes, vogesites, kersantites, spessartites. Trans. R. Soc. Edinburgh, 74, 193-227. 1986. The nature and origin of lamprophyres: alnoites and allied rocks (ultramafic lamprophyres). J. Petrol. 27, 155-96. - 1987. The nature and origin of lamprophyres : an overview. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 191-226. RODEN, M. F. 1981. Origin of coexisting minette and ultramafic breccia, Navajo volcanic field. Contrib. Mineral. Petrol. 77, 195-206. - -
---,
SMITH, D. & McDOWELL, F. W. 1979. Age and extent of potassic volcanism on the Colorado Plateau. Earth planet. Sci. Lett. 43, 279-84. ROEDDER, E. 1984. Fluid inclusions. Rev. Mineral. 12, 506-511. ROGERS, N. W. BACHINSKI, S. W. HENDERSON, P. & PARRY, S. J. 1982. Origin of potash-rich basic lamprophyres; trace element data from Arizona minettes. Earth planet. Sci. Lett. 57, 305-12. , HAWKESWORTH,
C.
J., PARKER,
R.
J. & MARSH,
J. S. 1985. The geochemistry of potassic lavas from Vulsini, central Italy and implications for mantle enrichment processes beneath the Roman region. Contrib. Mineral. Petrol. 90, 244-57. RONOV, A. B. & YAROSHEVSKY,A. A. 1969. Chemical composition of the Earth's Crust. Am. Geophys. Union Monogr. 13, 37-57. ROSENBUSCH, H. 1877. Mikroskopische Physiographie, 1st edn. Schweizerbart, Stuttgart. 1887. Mikroskopische Physiograpie, 2nd edn. Schweizerbart, Stuttgart. RosI, M. 1980. The island of Stromboli. Soc. Ital. Mineral. Petrol., Rend. 36, 345-68. Ross, C. P., ANDREWS, D. A. & WITKIND, I. J. 1955. Geological map of Montana (1/500,000 scale). Montana Bureau of Mines and Geology, Helena. Ross, C. S. 1926a. Nephelite-hauynite alnoite from Winnett Montana. Am. J. Sci. 11,218-27. -1926b. A Colorado lamprophyre of the verite type, Am. J. Sci. 12, 217-29. - - , MISER, H. O. & STEPHENSON,L. W. 1929. Water laid volcanic rocks of early Upper Cretaceous age in SW Arkansas, SE Oklahoma and NE Texas. U.S. Geol. Surv. Prof Pap. 154F, 175-202. ROWELL, W. F. & EDGAR, A. D. 1983. Cenozoic potassium-rich mafic volcanism in the western U.S.A.: its relationship to deep subduction. J. Geol. 91,338-41. RoY, B. C. 1962. Geological Map of India (1/5,000,000 scale). Geol. Surv. India. RoY CHAUDHURY,M. K. 1973. Geological and mineral map of Andhra Pradesh (1/2,250,000 scale). Geol. Surv. India. RUDDOCK, D. I. & HAMILTON,D. L. 1978a. Stability of carbonate in a simple potassium-rich rock model. Prog. Exp. Petrol. N E R C Publ., Ser. D, 11, 2831. -& -1978b. The system KA1Si206-CaMgSi206-SiO 2 at 4 kb. Prog. Exp. Petrol. NERCPubl., Ser. D, 11, 25-27. RYABCHIKOV, I. D. & BOETTCHER,A. L. 1981. Experimental evidence at high pressure for potassic metasomatism in the mantle of the earth. Am. Mineral. 65, 915-9. - & GREEN, D. H. 1978. The role of CO2 in the petrogenesis of highly potassic magmas. Tr. Inst. Geol. Geofiz. Akad. Nauk S.S.S.R. No 403, 49-64. --, SCHREYER, W. & ABRAHAM, K. 1982. Composition of aqueous fluids in equilibrium with pyroxenes and olivines at mantle pressures and temperatures. Contrib. Mineral. Petrol. 79, 80-4. SAETHER, E. 1950. On the genesis of peralkaline rock provinces. 18th Int. Geology Cortgr., London, Part II, pp. 123-30.
186
S. C. Bergman
SAHAMA, Th. G. 1974. Potassium-rich alkaline rocks. In : SORENSEN,H. (ed.) The Alkaline Rocks, pp. 96109, Wiley, London. SALTERS, V. J. M., & BARTON, M. 1985. The geochemistry of ultrapotassic lavas from The Leucite Hills, Wyoming. LOS, 66, I 109. SAN MIGUEL DE LA CAMERA,M. 1935. Una erupcion de jumillita en la Sierra de las Cabras (Albacete). Bol. R. Soc. Esp. Hist. Nat. 35, 147-54. , ALMELA, A. & FUSTER, J. M. 1952. sobre un volcan de veritas recientemente descubierto en el Miocene de Barqueros (Murcia). Estud. Geol. 7, 411-29. SANYAL,S. P. 1964. Petrology of certain lamprophyres from the Jharia Coalfield Bihar, with a discussion on the differenttiation of the Sudamdih Sill. Misc. Publ. Geol. Surv. India, 8, 27-44. SARKAR, A., PAUL, D. K., BALASUBRAHMANYAN,M. N. & SENGUPTA, N. R. 1980. Lamprophyres from Indian Gondwanas; K - A r ages and chemistry. J. Geol. Soc. India, 21, 188-93. SARMA, B. S. P. 1983. A report on ground magnetic survey over Chelima dyke (Cuddapah basin). Nat. Geophys. Res. Inst., Hyderabad, India, Tech. Rep. 83-210. SATHE, R. V. & S. S. 1975. Petrogenesis oflamprophyres of Mt. Girnar, Saurahtra. Q. J. Geol. Min. Metall. Soc. India, 46, 61-7. SATYANARAYANA RAO, P. & PHADTRE, P. N. 1966. Kimberlite pipe rocks of Wajrakarur Area, Anantapur district, Andhra Pradesh. J. Geol. Soc. India, 7, 118-23. SCHAIRER, J. F. & BOWEN, N. L. 1938. The system leucite-diopside-silica. Am. J. Sci. 235-A, 289-309. SCHEURING, B., AHRENDT, H., HUNZIKER, J. C. & ZINGG, A. 1974. Paleobotanical and Geochronological evidence for the Alpine Age of the metamorphism in the Sesia Zone. Geol. Rundsch. 63, 305-26. SCHMITT, H. H., SWANN, G. A. & SMITH, D. 1974. The Buell Park kimberlite pipe, N. E. Arizona. In: KARLSTRUM, T., SWANN, G. A. & EASTWOOD, R. L. (eds) Geology of Northern Arizona, pp. 672-98. Northern Arizona Univ., Flagstaff. SCHNEIDER, M. E. 1982. Major element compositions of silicates dissolved in supercritical fluids at mantle conditions: implications for mantle metasomatism. M.S. Thesis, Pennsylvania State University (unpublished). & EGGLER, D. N. 1984. Compositions of fluids in equilibrium with peridotite: implications for alkaline magmatism-metasomatism. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 383-94. Elsevier, Amsterdam. SCHULTZ, A. R. & CROSS, W. 1912. Potash-bearing rocks of the Leucite Hills, Sweetwater County, Wyoming. U.S. Geol. Surv. Bull. 512, 1-39. SCHULZE, D. J. & HELMSTAEDT, H. 1979. Garnet pyroxenite and eclogite xenoliths from the Sullivan Buttes latite, Chino Valley, Arizona. In : BOYD, F. R. & MEYER, H. O. A. (eds) The mantle sample: Inclusions in kimberlites and other volcanics. Proc. 2nd Int. Kimberlite Conf., Vol. 2. pp. 318-29. American Geophysical Union, Washington, DC. -
-
--,
SMITH, J. V., & N~MEC, D. 1985. Mica chemistry of lamprophyres from the Bohemian Massif, Czechoslovakia. Neues. Jahrb. Mineral. Abh. 152, 321-34. SCOTT, B. H. 1977. Petrogenesis of kimberlites and associated potassic lamprophyres from central West Greenland. PhD Thesis, University of Edinburgh (unpublished). ,1979. Petrogenesis of kimberlites and associated potassic lamprophyres from central West Greenland. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds. Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 190-205. American Geophysical Union, Washington, DC. 1981. Kimberlite and lamproite dykes from Holsteinsborg, W. Greenland. Meddr Gronland, 4, 3-24. SCOTT-SMITH,B. H. & SKINNER, E. M. W. 1982. A new look at Prairie Creek, Arkansas. Terra Cognita, 2, 210. & -1984a. A new look at Prairie Creek, Arkansas. In: KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 255-84. Elsevier, Amsterdam. & -1984b. Diamondiferous lamproites. J. Geol. 92, 433-8. --- & -1984c. Kimberlite and American Mines, near Prairie Creek Arkansas. In: KORNPROBST,J. (ed.) Kimberlites III: Documents Ann. Sci. Univ. Clermont-Ferrand II 74, 27-36. SCULL, B. J. 1959. The age of mineralization in the Ouachita Mountains of Arkansas and Oklahoma. Syrup. on the Geology of the Ouachita Mountains, pp. 62-9. Dallas Geological Society and Ardmore Geological Society. SEARLE, M. P. 1984. Alkaline peridotite pyroxenite and gabbroic intrusions in the Oman Mountains, Arabia. Can. J. Earth Sci. 21,396-406. SEN, S. N. & NARASHIMARAO 1970. Chelima dykes. Proc. 2nd Syrup. on Upper Mantle, Project Session 5, Combined Geosurveys in Dharwar and Cuddapah Basin, Hyderabad, India, pp. 435-9. SHAFIQULLAH,M., DAMON, P. E. & LYNCH, D. J. 1976. Ultrapotassic trachytes at Pichaco Peak in the southern Basin and Range province, Arizona. Geol. Soc. Am. Abstr. Progm. 8, 628. SHAND, S. J. 1927. Eruptive Rocks, l st edn. Thomas Murby, London. -1931. The granite-syenite-limestone complex of Palabora, Transvaal. Trans. Geol. Soc. S. Aft. 67, 81-92. SHERATON, J. W. 1983. Geochemistry of mafic igneous rocks of the northern Prince Charles Mountains, Antarctica. J. Geol. Soc. AuNt. 30, 295-300. & CUNDARI, A. 1980. Leucitites from Gaussberg, Antarctica. Contrib. Mineral. Petrol. 71,417-27. & ENGLAND, R. N. 1980. Highly potassic mafic dykes from Antarctica. J. Geol. Soc. AuNt. 27, 12935. , OFFE, L. A., TINGLY, R. J. & ELLIS, D. J. 1980. Enderbyland, Antarctica-an unusual Precambrian high-grade metamorphic terrain. J. Geol. Soc. AuNt. 27, 1-18. -
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks SIMPSON, E. S. 1925. Contributions to the mineralogy of Western Australia. Series I. J. R. Soc. W. Aust. 12, 57-66. SKEATS, E. W. & RICHARDS, H. C. 1926. Report of the alkaline rocks of Australia and New Zealand Committee. Aust. Assoc. Adv. Sci. 18, 36-45. SKINNER, E. M. W. & CLEMENT, C. R. 1979. Mineralogical classification of Southern African kimberlites. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds: Their Geology, Petrology and Geochemistry. Proc 2nd Int. Kimberlite Conf., Vol. 1, pp. 129-39. American Geophysical Union, Washington, DC. -& SCOTT, B. H. 1979. Petrography, mineralogy and geochemistry of kimberlite and associated lamprophyre dykes near Swartruggens, W. Transvaal, RSA. Kimberlite Symp. 11, Cambridge. SMITH, C. B. 1983a. Pb, Sr and Nd isotopic evidence for sources of Southern African Cretaceous kimberlites. Nature, Lond. 304, 51-4. -1983b. Rb-Sr, U - P b and S m - N d isotopic studies of kimberlite and selected mantle-derived xenoliths. PhD Thesis, University of Witwaterstrand, Johannesburg (unpublished). -1984a. The genesis of the diamond deposits of the West Kimberley, W. A. In: PURCELL, P. G. (ed.) The Canning Basin, WA. Proc. Geological Society of Australia and Petrology Exploration Society of Australia Symp., Perth, pp. 463-74. -1984b. What is a kimberlite? In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin: a Basisfor Conceptual Models in Exploration. Univ. W. Australia Publ. 8, pp. 1-18. SMITHSON, S. B. 1959. The geology of the southeastern Leucite Hills, Sweetwater County, Wyoming. M.A. thesis, Univ. Wyoming, Laramie (unpublished). SMYTH, C. H. MR. 1927. The genesis of alkaline rocks, Am. Phil. Soc. Proc. 66, 535-80. SOBOLEV, V. S. 1970. On the genesis of leucite-bearing rocks. Dokl. Akad. Nauk S.S.S.R., Earth Sci. Sect, 194, 166-8. , BAZAROVA,T. J. & YAGI, K. 1975. Crystallization temperature of wyomingite from Leucite hills. Contrib. Mineral. Petrol. 49, 301-8. SOLOVlEV, D. A. 1972. Platform magmatic formations of East Antarctica. In: ADIE, R. J. (ed.) Antarctic Geology and Geophysics, pp. 531-8. Universitetsforlaget, Oslo. SORENSEN, H. 1960. On agpaitic rocks. In: OFTEDAHL, C. & HJELMQVIST, S. (eds) Proc. 21st Int. Geol. Congr., Part 13, Petrographic Provinces, lgneous and Metamorphic Rocks, Det Berlingske Bogtrykkeri, Copenhagen pp. 319-27. STAUDIGEL,H. • ZINDLER, A. 1978. Nd and Sr isotope compositions of potassic volcanics from the East Eifel, Germany: implications for mantle source regions. Geol. Soc. Am. Abstr. Prog. 10, 497. STEELE, K. F. & WAGNER, G. H. 1979. Relationship of the Murfreesboro kimberlite and other igneous rocks of Arkansas, USA. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes, and Diamonds. Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp.
187
393-9. American Geophysical Union, Washington, DC. STEFANINI, G. 1934. I1 complesso eruttivo di orciatico e Montecatini inprovincia di Pisa. Atti Soc. Tosc. Sci. Nat. Mem. 44, 224-300. STEFANOVA, M. & BOYADJIEVA, R. 1975. On the geochemistry of Nb and Ta in K-alkaline lamproitic rocks from the village of Svidnja, district of Sofia. Geocham. Min. Petrol. 1, 16-30. STEPHENSON, P. J. 1972. Geochemistry of some Heard Island igneous rocks. In: ADIE, R. J. (ed.) Antarctic Geology and Geophysics, pp. 793-801. Universitetsforlaget, Oslo. STEWART, D. B. 1979. The formation of siliceous potassic glassy rocks. In: YODER, H. S. (ed.) The Evolution of the lgneous Rocks: 50th Anniversary Perspectives, pp. 339-50. Princeton University Press, Princeton, NJ. STEWART, O. F. 1949. Origin and occurrence of vermiculite at Tigerville, S. Carolina. M S Thesis, University of South Carolina, Columbia, SC (unpublished). STONE, C. G. & STERLING, P. J. 1964. Relationship of igneous activity to mineral deposits in Arkansas. State of Arkansas-Ark. Geol. Com., Little Rock. Ar. STRACKE, K. J., FERGUSON, J. & BLACK, L. P. 1979. Structural setting of kimberlites in Southeastern Australia. ln: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds, Proc. 2nd Kimberlite Conf., Vol. 1, pp. 71-91. Am. Geophys. Union, Washington. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks--final report of an enquiry. Neues Jb. Mineral. Abh. 107, 144-214. 1980. Classification and nomenclature of volcanic rocks, carbonatites and melilitic rocks. IUGS subcommission on the systematics of igneous rocks. Geol. Rundsch. 69, 194-207. SVISERO, D. P., HASUI, Y. & DRUMMOND, D. 1979a. Geologia de kimberlitos de Alto Paranaiba, Minas Gerais (Geology of kimberlites of the Upper Paranaiba, Minas Gerais). Min. Metall. 42, 34-38. --, MEYER, H. O. A. & HSIAO-MING TSAI 1979b. Kimberlites in Brazil: an initial report. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites Diatremes, and Diamonds: Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 92-100. American Geophysical Union, Washington, DC. - - , HARALYI,N. L. E. & HASUI, Y. 1984. A note on the geology of some Brazilian kimberlites. J. Geol. 92, 331-8. TAPPE, J. 1966. The chemistry, petrology and structure of the Big Timber complex, Crazy Mountains, Montana. PhD. Thesis, Univ. Cincinnatti (unpubl.). TAYLOR, H. P. 1968. The oxygen isotope geochemistry of igneous rocks. Contrib. Mineral. Petrol. 19, 171. --, GIANETTI, B. & TURI, B. 1979. Oxygen isotope geochemistry of the potassic igneous rocks from the Roccamonfina volcano, Roman comagmatic region. Earth planet. Sci. Lett. 46, 81-106. & TURI, B. 1976. High 180 igneous rocks from the -
-
-
-
I 8 8
S. C.
Bergman
Tuscan magmatic province, Italy. Contrib. Mineral. Petrol. 55, 34-54. & CUNDARI, A. 1984. 180/160 and the chemical relationships in K-rich volcanic rocks from Australia, E. Africa, Antarctica and San Venanzo, Cupaello, Italy. Earth planet. Sci. Lett. 69, 263-76. TAYLOR, S. R. 1979. Lunar and terrestrial potassium and uranium abundances: implications of the fission hypothesis. Proc. lOth Lunar Sci. Conf., 2017-30. -& MCLENNAN, S. M. 1981. The composition and evolution of the continental crust: rare earth element evidence from sedimentary rocks. Phil. Trans. R. Soc. Lond. A301, 381-405. TEMPLEMAN-KLUIT, D. J. 1969. Re-examination of pseudoleucite from Spotted Fawn Creek, West Central Yukon. Can. J. Earth Sci. 6, 55-62. TERZAGHI, R. D. 1935. The origin of the potash-rich rocks. Am. J. Sci. 29, 369. THIBAUT, G., RAOULCARUBA,M. M. & TURCO, G. 1972. Etude exp6rimentale du domaine de formation de la dalyite K2ZrSi6015 et essai de corr61ations p&rologiques. C. R. Acad. Sci. Paris, 274, 792-5. THOENEN, J. R., HILL, R. S., HOWE, E. G. & RUNKE, S. M. 1949. Investigation of the Prairie Creek diamond area, Pike County, Arkansas. U.S. Bur. Mines, Rep. Invest. 45-9. THOMPSON, R. N. 1977. Primary basalts and magma genesis 111. Alban Hills, Roman Comagmatic province, central Italy. Contrib. Mineral. Petrol. 60, 91-108. - 1985. Asthenospheric source of Ugandan ultrapotassic magma. J. Geol. 93, 603-8. , MORRISON,M. A., HENDRY, G. L. & PARRY, S. J. 1984. An assessment of the relative roles of crust and mantle in magma genesis: an elemental approach. Phil. Trans. R. Soc. Lond. Ser. A, 310, 549-90. THY, P. 1982. Richterite-arfvedsonite-riebeckite-actinolite assemblage from MARID dykes associated with ultrapotassic magmatic activity in central West Greenland. Terra Cognita, 2, 247-9. - 1985. Contrasting crystallization trends in ultrapotassic lamproites from central W. Greenland. Geol. Assoc. Canada Progm. Abstr. 10, A63. TIDMARSH, W. G. 1932. The Permian lavas of Devon. Q. J. Geol. Soc. Lond. 8 8 , 712-3. TIKHONENKOV,I. P., KUKHARCHIK,M. V. & PYATENKO, Yu. A. 1960. Wadeite from the Khibiny Massif and the conditions of its formation. Dokl. Akad. Nauk. S.S.S.R. 134, 920-3. TINGEY, R. J. 1981. Geological investigations in Antarctica 1968-1969: The Pryce Bay-Amery Ice Shelf-Prince Charles Mountains area. Aust. Bur. Miner. Resources Geol. Geophys. Rep. 1981/34. - - , McDOUGALL, I. & GLEADOW, A. J. W. 1983. The age and mode of formation of Gaussberg, Antarctica. J. geol. Soc. Aust. 30, 241-24. TRAIL, D. S. 1963. The 1961 geological reconnaissance in the Southern Prince Charles Mountains, Antarctica. Rec. Bur. Miner. Resources Geol. Geophys. Aust. 1963-155 (unpublished). TROGER, W. E. 1935. Spezielle petrographie der Erup-
tivgesteine, Deutschen Mineralogischen Gesellschaft, Berlin. TRONNES, R. G., EDGAR, A. D. & ARIMA, M. 1985. A high pressure-high temperature study of TiO2 solubility in Mg-rich phlogopite: implications to phlogopite chemistry. Geochim. cosmochim. Acta, 49, 2323-9. TUREKIAN, K. K. & WEDEPOHL, K. H. 1961. Distribution of the elements in some major units of the earth's crust. Geol. Soc. Am. Bull. 72, 175-91. TURI, B. & TAYLOR,H. P. 1976. Oxygen isotope studies of potassic volcanic rocks of the Roman Province, central Italy. Contrib. Mineral. Petrol. 55, 1-31. TURNER, F. J. & VERHOOGEN, J. 1960. Igneous and Metamorphic Petrology, pp. 235-56. TUTTLE, O. F. & GITTINS, J. 1966. Carbonatites. WileyInterscience, New York. TWENHOFEL, W. H. & BREMER, W. H. 1928. An extension of the Rose Dome intrusives Kansas. Bull. Am. Assoc. Petrol. Geol. 12, 757-62. ULBRICH, H. H. G. J. & GOMES, C. B. 1981. Alkaline rocks from Continental Brazil. Earth Sci. Rev. 17, 135-54. UNESCO 1976. Geological Worm Atlas., CHOUBERT, G. & FAURE- MURET, A. (compilers), UNESCO, Paris. UPTON, B. G. J. & THOMAS, J. E. 1973. Precambrian potassic ultramafic rocks: South Greenland. J. Petrol. 14, 509-34. USSING, N. V. 1912. Geology of the country around Julianehfib, Greenland. Meddr. Gronland. Geosci. 38, 1 146. VAN KOOTEN, G. K. 1980. Mineralogy, petrology and geochemistry of an ultrapotassic basaltic suite, Central Sierra Nevada, California, U . S . A . J . Petrol. 21, 651-84. - 1981. Pb and Sr systematics of ultrapotassic and basaltic rocks from the Central Sierra Nevada, California. Contrib. Mineral. Petrol. 76, 378-85. - & PECK, D. L. 1977. Highly potassic phonolite of Pliocene age near Merced Peak, central Sierra Nevada, California. Geol. Soc. Am. Abstr. Progm. 9, 1208-9. VEEVERS, J. J. 1981. Australian rifted margins. In: SCRUTTON,R. A. (ed.)DynamicsofPassive Margins. Geodynam. Ser. Vol. 6, Amer. Geophys. Union Washington D.C. pp. 72-89. VELDE, D. 1967. Sur un lamprophyre hyperalcalin potassique : la minette de Sisco (ile de Corse). Bull. Soc. Fr. Minbral. Cristallogr. 90, 214-23. 1968. A new occurrence of priderite. Mineral. Mag. 36 867-70. 1969a. Micas from lamprophyres; kersantites, minettes and lamproites. Bull. Soc. Fr. Min&al. Cristallogr. 92, 203-23. - 1969b. Minettes et kersantites. Une contribution fi l'&ude des lamprophyres. Th&e de Doctorat d'Etat, Facult6 des Sciences, Univ Paris VI (unpublished). 1971. A note on an analcite-bearing lamproite from Devonshire. Geol. Mag. 108, 201-4. -1975. Armalcolite-Ti-phlogopite-diopside-analcite-bearing lamproites from Smoky Butte, Garfield County, Montana. Am. Mineral. 60, 566-73.
Lamproites and K-rich igneous rocks VENTURELLI, G., BALESTRA, G., TOSCANI, L. & CAPEDRI, S. 1985. The ultrapotassic rocks and their geological setting. Geol. Assoc. Can. Progm. Abstr. 10, A65. - - , CAPEDRI, S., DI BATTISTINI,G., CRAWFORD, A., KOGARKO, L. N. 8s CELESTINI, S. 1984a. The ultrapotassic rocks from SE Spain. Lithos, 17, 3754. - - , THORPE, R. S., DAL PIAZ, G. V., DEE MORO, A. & POTTS,P. J. 1984b. Petrogenesis ofcalc-alkaline, shoshonitic and associated ultrapotassic Oligocene volcanic rocks from the north western Alps, Italy. Contrib. Mineral. Petrol. 86, 209-20. VERWOERD, W. J. 1966. Fenitization of basic igneous rocks. In: TVTTLE, C. F. & GITTINS, J. (eds) Carbonatites, pp. 295-307. Wiley-Interscience, New York. VILA, J. M., HERNANDEZ, J. 8s VELDE, D. 1974. Sur la pr6scence d'un ilion de roche lamproitique (trachyte potassique fi olivine) recoupant le flysch de type (Guerrouch entre Azzaba (ex-Jemmapes) et Hammam-Meskoutine dans l'Est du Constantinois (Algerie)). C. R. Acad. Sei. Paris, Skr. D, 278, 2589-91. VOLLMER, R. 1976. Rb-Sr and U-Th-Pb systematics of alkaline rocks; the alkaline rocks from Italy. Geochim. cosmochim. Acta, 40, 283-95. 1977. Isotopic evidence for genetic relation between acid and alkaline rocks in Italy. Contrib. Mineral. Petrol. 60, 109-18. 8s HAWKESWORTH, C. J. 1980. Lead isotopic composition of the potassic rocks from Roccamonfina (South Italy). Earth planet. Sci. Lett 47, 91101. 8s NORRY, M. H. 1983a. Unusual isotopic variations in Nyiragongo nephelinites. Nature, Lond. 301, 141-3. & -1983b. Possible origin of K-rich volcanic rocks from Virunga, East Africa, by metasomatism of continental crustal material: Pb, Nd & Sr isotopic evidence. Earth planet. Sci. Lett. 64, 37486. - - - , OGDEN, P. & SCHILLING, J. G. 1981. Nd isotope geochemistry of Leucite Hills ultrapotassic volcanism. EOS, 62, 431. , KINGSLEY, R. H. 8s WAGGONER, D. G. 19'84. Nd and Sr isotopes in ultrapotassic volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 87, 359-68. VYALOV, O. S. & SOBOLEV, V. S. 1959. Gaussberg, Antarctica. Int. Geol. Rev. 1, 30-40. WADE, A. 1937. The geological succession in the west Kimberley district of western Australia. Rep. Aust. Assoc. Adv. Sci. 73, 93. & PRIDER, R. T. 1940. The leucite-bearing rocks of the West Kimberley area, Western Australia. Q. J. geol. Soc. Lond. 96, 39-98. WAGNER, C 8s VELDE, D. 1986a. The mineralogy of Krichterite-bearing lamproites. Am. Mineral., 71, 17-37. & -1986b. Lamproites in N. Vietnam? A reexamination of cocites. J. Geol. 94, 770-6. WAGNER, P. A. 1914. The diamond fields of Southern Africa. Transvaal Leader, Johannesburg. -
-
-
-
189
WAGNER, H. C. 1954. Geology of the Fredonia Quadrangle Kansas. U.S. Geol. Surv. Map GQ49. WALDMAN, M. A., MCCANDLESS, T. E. & DUMMETT, H. T. 1985. Geology and mineralogy of the Twin Knobs # 1 lamproite, Pike County, Arkansas. Geol. Soc. Am. Abstr. Progm. 17, 196. WALKER, K. R., 8s MOND, A. 1971. Mica lamprophyre (alnoite) from Radok Lake, Prince Charles Mountains, Antarctica. Aust. Bur. Miner. Resources Geol. Geophys. Ree. 1971/108 (unpublished report). WASHINGTON,H. S. 1917. Chemical analysis of igneous rocks, published from 1884-1913 inclusive. Prof. Pap. U.S. Geol. Surv. 99. WATERS, A. C. 1955. Volcanic rocks and the tectonic cycle. Geol. Soc. Am. Spec. Pap. 62, 703-22. WATKINSON, D. H. & CHAO, G. Y. 1973. Shortite in kimberlite from the Upper Canada gold mine, Ontario. J. Geol. 81, 229-33. WEDEPOHL, K. H. 8s MURAMATSU, Y. 1979. The chemical composition of kimberlites compared with the average composition of three basaltic magma types. In: BOYD, F. R. & MEYER, H. O. A. (eds) Diatremes and Diamonds: Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 300-12. American Geophysical Union, Washington, DC. WEED, W. H. & PIRSSON, L. V. 1895. Highwood Mountains of Montana. Geol. Soc. Am. Bull. 6, 389-422. 8s - 1896. Missourite, a new leucite rock from the Highwood Mountains, Montana. Am. J. Sci. 2, 315-23. WELLMAN, P. 1973. Early Miocene potassium-argon age for the Fitzroy lamproites of Western Australia. J. Geol. Soc. Aust. 19, 471-4. , CUNDARI,A. 8s McDOUGALL, J. 1970. Potassiumargon ages for leucite-bearing rocks from New South Wales, Australia. Proc. R. Soc. N.S. W. 103, 103-7. WENDLANDT, R. F. 1977a. The system KA1SiO4-SiO4SIO4-CO2: phase relations involving forsterite, enstatite, sanidine, kalsilite, leucite, vapor, and liquid to 30 kbar. Carnegie Instn. Wash. Yb. 76, 435-41. 1977b. The system K20-MgO-A1203-SiO2H 2 0 - C O 2 : stability of phlogopite as a function of vapor composition at high pressures and temperatures. Carnegie Instn. Wash. Yb. 76, 441. 8s EGGLER, D. H. 1980a. The origins of potassic magmas: I. Melting relations in the systems KA1SiO4-Mg2SiO~-SiO2 and KA1SiO4-MgOSiO2-CO 2 to 30 kb. Am. J. Sci. 280, 385-420. & -1980b. The origin of potassic magmas. II. Stability of phlogopite in natural spinel lherzolite and in the system KA1SiO4-MgO-SiO2H20-COz at high P and T. Am. J. Sci. 280, 421-45. & -1980c. Stability of sanidine+forsterite and its bearing on the genesis of potassic magmas and the distibution of potassium in the upper mantle. Earth planet. Sci. Lett. 51,215-20. WICKHAM, J., ROEDER, D. & BRIGGS, G. 1976. Plate tectonics models for the Ouachita foldbelt. Geology, 4, 173-6. -
-
-
-
-
-
-
-
S. C. Bergman
190
WILLIAMS, H. 1936. Pliocene volcanoes of the NavajoHopi country. Geol. Soc. Am. Bull. 47, 111-72. , TURNER, F. J. & GILBERT, C. M. 1954. Petrography: an introduction to the study of rocks in thin sections. Freeman, San Francisco. WILLIAMS, J. F. 1891. The igneous rocks of Arkansas. Geol. Sum. Arkansas, Report II. WILSON, A. N. 1982. Diamonds, from Birth to Eternity, Gemolog. Institute of America, Santa Monica, CA. WIMMENAUER, W. 1973. Lamprophyre, semilamprophyre and anchibasaltic dike rocks. Fortschr. Mineral. 51, 3-67. -1974. The alkaline province of central Europe and France. In: SORENSEN,H. (ed.) The Alkaline Rocks, pp. 238-71. Wiley, New York. WITKIND, I. J. 1969. Clinophyroxenes from acid, intermediate and basic rocks, Little Belt Mountains, Montana. Am. Mineral. 54, 118-38. - 1973. Igneous rocks and related mineral deposits of the Barker Quadrangle, Little Belt Mountains, Montana. Prof. Pap. U.S. Geol. Surv. 752, 1-58. WOLFF, J. E. 1938. Igneous rocks of the Crazy Mountains, Montana. Geol. Soc. Am. Bull. 49, 1569-628.
WONES, D. P. 1979. The fractional resorption of complex minerals and the formation of strongly femic alkaline rocks. In: YODER, H. S. (ed.) The Evolution of the lgneous Rocks: 50th Anniversary Perspectives, pp. 413-22. Princeton University Press, Princeton, NJ. WOODS, M. J. 1975. Textural and geochemical features
of the Highwood Mountains volcanics, central Montana. PhD Thesis, University of Montana (unpublished). WYOMING STATESURVEYGEOLOGICALMAP 1955. YAGI, K. & MATSUMOTO,H. 1966. Note on the leucitebearing rocks of Leucite Hills, Wyoming, U.S.A. J. Fac. Sci. Hokkaido Univ., 13, 301-12. - - , ISHIKAWA,H. & KOJIMA, M. 1975. Petrology of a lamprophyre sheet in Tanegashima Island, Kagosshima Perfecture, Japan. J. Jap. Assoc. Min. Petrol. Econ. Geol. 70, 213-24. YARZA, R. A. 1895. Roca eruptiva de Fortuna (provincia de Murcia). Bol. Comm. Mapa Geol. Espana, 20, 349-53. ZARTMAN, R. E. 1977. Geochronology of some alkalic rock provinces in eastern and central United States. Ann. Rev. Earth planet. Sci. 5, 257-86. - - , BROCK, M. R., HEYLE, A. V. & THOMAS, H. H. 1967. K-Ar and Rb-Sr ages of some alkalic intrusive rocks from central and eastern US. Am. J. Sci. 265, 848-70. ZHURAVLEYA, L. N., YURKINA, K. V. & RYABEVA,Y. G. 1978. Priderite, first find in the USSR. Dokl. Akad. Nauk S.S.S.R., Earth Sci. 239, 141-3. ZINGG, A., HUNZIKER, J. C., FREY, M. & AHRENDT, H. 1976. Age and degree of metamorphism of the Canavese Zone and the sedimentary cover of the Sesia Zone. Schweiz. Mineral. Petrogr. Mitt. 56, 361-75. ZIRKEL, F. 1870. Untersuchungen fiber die Mikroskopische Zusammensetzung und Struktur der Basaltgesteine, Bonn.
STEVENC. BERGMAN,ARCO Resources Technology, Exploration and Production Research Center, 2300 W. Piano Pkwy, Piano, Tx 75075, U.S.A.
The nature and origin of lamprophyres: an overview Nicholas M. S. Rock S U M MA R Y: Lamprophyres are 'alkaline rocks' because they carry high alkalis at a given percentage of SiOz, together with one or more of normative ne, lc or ac, modal foids, and NaK-Ti-rich amphiboles or pyroxenes. They reach higher whole-rock H20,CO2, Sr and Ba contents than other silicate igneous rocks. Contents of related element subsets in amphiboles (Ti, Ba), K-feldspars (Ba, Fe3), phlogopites (Ti, Ba, Fe 3) and pyroxenes (Ti, A1, Fe 3) include among the highest values known in nature for these minerals. 'Primitive' minerals (diopside, forsterite) commonly coexist with 'evolved' minerals (albite + orthoclase, quartz). Four welldefined 'branches' of the lamprophyre 'clan' have distinctive compositions: calc-alkaline (shoshonitic) lamprophyres (minettes etc.), alone among lamprophyres, have mixed alkalinecalc-alkaline affinities; alkaline lamprophyres (camptonites etc.) are basanitic to nephelinitic and, alone among lamprophyres, usually have Na >>K; ultramafic lamprophyres (aln/Sites etc.) are the most Si poor and Ca rich of silicate igneous rocks, and grade into carbonatites; lamproites (orendites etc.) are uniquely rich in K, Rb, Ba, Th, Mg, Cr and Ni at mainly 'andesitic' SiO2 contents, and carry a suite of diagnostic minerals (wadeite etc.). Each branch comprises at least four rock types which resemble each other much more than rock types of other branches; however, some rock types can be grouped into slightly distinct "families' within one branch (e.g. phlogopitic and madupitic lamproites). A case can be made for including kimberlites as a fifth branch of the lamprophyre clan. Synoptic plots and tables, based on some 5000 major, trace-element and mineral analyses, are presented, to aid identification and classification. Lamprophyres are far more common than generally stated, occurring worldwide in more tectonic settings than many other alkaline rocks and throughout the geological record. They may approximate intratelluric magma compositions. Nearly all represent primitive magmas, and many represent primary magmas. Some represent parental magmas to a wide range of hydrous alkaline intrusive suites: calc-alkaline lamprophyres to potassic pyroxenite-diorite-shonkinite-syenite-granite plutons (Cortlandt); alkaline lamprophyres to hornblendic gabbro-syenite plutons (Monteregian Hills); and ultramafic lamprophyres to ijolite-carbonatite complexes (Fen).
Introduction
Nomenclature
Lamprophyres are the only 'alkaline rocks' covered neither by Sorensen (1974) nor yet in a specialized monograph. Despite Streckeisen's rationalized nomenclature of 1979 and many petrologists' awakening recognition of their global significance, they remain to many others a curiosity (Seeliger 1975) and continue to be treated at best as an 'uncomfortable afterthought' (Hughes 1982) in textbooks. The term 'lamprophyric' often conceals meanings as vague but varied as 'altered', 'biotite-rich' and 'exotic' (Heinrich & Alexander 1975) or even 'not studied in detail'! Yet it is one of few rock names intrinsically conveying its meaning (Greek lampros, porphyros, glistening porphyry or purple rock)--compare locality-based names such as kimberlite, and terms such as granophyre and keratophyre which, sophistically, do not imply porphyritic texture. Furthermore, lamprophyres can often be identified in the field (unlike many fine-grained rocks), and they are among the most widespread of alkaline rocks. This overview is part of a continuing attempt to give them their rightful place in the literature.
Previous confusion perhaps attained its nadir with the Arrow Peak dyke, Montana, variously called 'orthoclase-camptonite' (Rosenbusch 1897), 'minette' (Pirsson 1905), 'leucite-monchiquite' (Beger 1923), 'diopside-lamprophyre' (Knopf 1936) and even 'mafic phonolite' (Buie 1941). A hierarchical classification, expanded in the following ways from Streckeisen (1979), is used here (Fig. 1). 1 As in Rock (1986), Streckeisen's 'melilitic lamprophyres' are renamed 'ultramafic lamprophyres'. 2 As in Rock (1977, 1981), lamproites (the leucite-lamproites of some authors) are regarded as a fourth 'branch' of the lamprophyre 'clan'; they share all the characteristics defined below, and have almost invariably been described as 'lamprophyric' (from Niggli (1923) to Jaques et al. (1984)); the name lamproite, of course, intrinsically implies close affinity with lamprophyres. 3 As in Hughes (1982), and as elaborated in Rock (1987) kimberlites are regarded as a fifth
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline lgneous Rocks, Geological Society Special Publication No. 30, pp. 191-226.
I9I
N. M. S. Rock
I92
m
~
m
~
9
~
g
9
~ ~ ~ "r~ ~ .=.~ ~ ,
Z ~'~
~ ~
._
9
~ -
m-.~
~oo
,-~ ~ . ~
~ 9"-'
9~ - ~ ~ x~ ~_~ ~ ~ o ~.~
oo
~
~
~-
9~
-~-~
~
,-~
.~ ~
o~
~9
~7~: ~ ,~
.~_.~ -~ ~:~.~
,.~
~.
~
~
~l~ F.~ .,,-~
,,.~ ~- ,..~ .c~
.
~.~ o~ ~
~ ~'~
"o .~
O, Ns=O 2 * 3 Ac>O, Ns>O < 1 * < 1 Ac>O, Ks>O * * 0 Ac >O, Ns> O, Ks >O * * 0 Wo>O O 10 7 0 Ru > 0 3 6 0
293 43 21 UML > CAL > AL. 3 The light REE thus increase both with increasing K / N a and K-feldspar/plagioclase ratios (cf. Cullers & Graf 1984). 4 Different lamproite families define quite distinct fields.
t
_/
9
I / .~
,Jumillite lamproites ,~~
.....
/
~
Miscellaneous petrological characteristics of the whole lamprophyre clan Xenolith and megacryst/macrocryst assemblages (Table 12)
'
/a
l
20-
t m Camp o ' es I Spessartites ~x-tr,, ~ , ~ M L " ~ ' , Z / / * ~ " X ~ / , ~ ._ F ....~"o . ~ : ~ ' i . . o . . . x . ~ " _ ~.~Fitzroyite /vy..~ lamproites I01- ".'O~.~..o.~..~l:~" i~.' / O
/ 10
Plag~
,,,
20
40
~
__
IO0
200
,
Lu/Yb
FIG. 9. Comparison of lamprophyre REE spectra with those of kimberlites and carbonatites. Symbols as in Fig 3. Note the log-log scales. Sources as in Fig. 8. The majority fields are labelled. Kimberlite field after Nixon et al. (1980).
Lamprophyres characteristically carry abundant crustal xenoliths, which sometimes appear to sample inferred subcrop faithfully (Read et al. 1926). Many lamprophyres, moreover, grade into (or are associated with) breccias, ~zarrying abundant country-rock material (Nixon et al. 1980; Rock 1983). All four lamprophyre branches have also furnished spectacular mantle xenolith suites: CAL, Navajo, U.S.A. (Ehrenberg 1979); AL, Scottish Highlands (Rock 1983); UML, Malaita, Solomons (Dawson et al. 1978); LL, W Kimberley, Australia (Atkinson et al. 1984). The number of recent reports suggests that lamprophyres may even carry proportionately more mantle material than basalts, taking into account relative hostrock abundances. For example, monchiquites have yielded some of the most varied and abundant mantle material among Scottish Permo-Carboniferous alkali basaltic intrusions (Rock 1983; Upton et al. 1983). Curiously,
TABLE 11. Combined Sr and N d &otopic data for lamproites Lamproite source W Australia W Australia Leucite Hills
Reference McCulloch et al. (1983) Nixon et al. (1984) Vollmer et al. (1984)
143Nd/144Nd 0.5110-0.5114 0.5119-0.5121 0.5118-0.5121
87Sr/86Sr 0.7104-0.7187 No data 0.7053-0.7061
Nature of lamprophyres 8-
213
Calc-alkaline lamprophyres (28)
4 I
4--
6 2 3 16
5
21
12-
Alkaline lamprophyres (27)
8 12 9
8-
F.~Tl
I0 13 I4 I
!5
4-
16 -8
II
23
12-
Ultramafic lamprophyres
20
(27)
21
8-
17 18
23
4-
22 I9
25 -
~n
12
8--
Lamproifes (99)
t-
25
424 I
0"705
I
I
0"715
26
177 0"725
e7Sr/e6Sro
whereas the mantle xenolith (?diamond) contents of LL and UML imply depths of origin greater than other igneous rocks (except kimberlites), their contents of leucite and melilite imply crystallization at among the l o w e s t pressures. Varied macrocryst suites are also characteristic (Table 12). Zoned ovoid Ti-phlogopite megacrysts ('eggs') up to 3 cm long (Le Cheminant & Le Cheminant 1985) appear so far to be unique to lamprophyres. Opinions differ as to the cognate or xenocrystic origin of such phases (see references in Table 12).
FIG. 10. Distribution of initial 87Sr/S6Sr ratios in lamprophyres. The numbers in brackets indicate the total numbers of values (isochron intercepts are taken as single values). The numbers within the histograms identify the data sources and localities as follows. Calc-alkaline lamprophyres: 1 kersantites, Massif Central, France (Cantagrel et al. 1970);2 kersantite and spessartite, Sines, Portugal (Canilho 1971); 3 minettes, Spanish Peaks, Colorado (Jahn et al. 1979); 4 minettes, Navajo, U.S.A. (Powell & Bell 1970; Roden 1981); 5 spessartites & kersantites, Cortlandt, U.S.A. (Ratcliffe et al. 1983); 6 minettes, Lake Huron, Canada (Van Schmus 1971); 7 'lamprophyre', probably lamproite, NW Alps (Dal Piaz et al. 1979). Alkaline lamprophyres: 8 camptonites and monchiquites, Monchique, Portugal (Rock 1976); 9 camptonites, Spanish Peaks, U.S.A. (Jahn et al. 1979); 10 monchiquite, Magnet Cove, U.S.A. (Powell et al. 1966); 11 camptonites and monchiquites, New England, U.S.A. (McHone et al. 1976); 12 'lamprophyre' (camptonite), Girnar, India (Paul et al. 1977); 13 camptonite, Predazzo, Italy (Lucchini & Mezzetti 1969); 14 monchiquite, Pyrenees, Spain (Vitrac-Michard et al. 1977); 15 sannaite, Jombo, Kenya (Rock 1976); 16 monchiquites, Hopi, U.S.A. (Powell & Bell 1970). Ultramafic lamprophyres: 17 alnoite, Oka, Canada (Powell et al. 1966); 18 alnoitic rock, Rangwa, Kenya (Rock 1976); 19 damtjernites, Fen, Norway (Griffin & Taylor 1975); 20 aillikite, Ytteroy, Norway (Priem et al. 1968);21 alnoites, Montana, U.S.A. (Powell & Bell 1970); 22 alnoites, Alno, Sweden (Brueckner & Rex 1980); 23 aillikites and alnoites, Frederikshaab, Greenland (Hansen 1981). Lamproites: 24 SE Spain (Powell & Bell 1970); 25 Leucite Hills, U.S.A. (Powell & Bell 1970; Vollmer et al. 1984); 26 W Kimberley, Australia (Powell & Bell 1970); McCulloch et al. 1983; Jaques et al. 1984).
Globular structures
Felsic 'globular structures' (Phillips 1973) ('amygdales', 'ocelli', 'patches', 'varioles' etc. of other workers) are found in all lamprophyre types (summaries in Popov (1972) and Rock (1977, 1984, 1986)). Although similar structures occur in basalts, they are much less widespread. In Scotland, for example, they distinguish PermoCarboniferous AL from Tertiary and Devonian basalts (Bailey et al. 1924). Globular structures differ from their host rock in their lower colour
2 I4
N.M.S.
Rock
TABLE 12. G e n e r a l i z e d a b u n d a n c e o f x e n o l i t h i c a n d m a c r o c r y s t i c / m e g a c r y s t i c m a t e r i a l in the d i f f e r e n t lamprophyre branches
Calc-alkaline lamprophyres 1
Alkaline lamprophyres 2
Ultramafic lamprophyres 3
Lamproites 4
( X )6 --( x )6
X X
X X
X
x x x
x x x
x x x x
x x x x x x x x x x x
x x x x x x x x x
9 x x x x x x x
? x x x x x x x
--
( x )9
?8
x x
x x x
x x --
x x x --
x x x x
x x x x --
(x) -x x (x )
x x
x ? x
Mantle-type ultramafic xenoliths s
Lherzolites, harzburgites Wehrlites, websterites Phlogopite-hornblende-rich rocks Eclogites/garnet-peridotites Other xenolithic material
Al-rich restites Miscellaneous (meta-)sediments Basalt, gabbro, amphibolite Granitoids Deep crustal granulites Country rocks Megacrysts and xenocrysts 7 Diamond Spinel, ilmenite Orthopyro xene, olivine Garnet ( M g - C r )
Kaersutite 1~ Apatite 1~ Clinopyroxene 1~ Sanidine-anorthoclase l~
-x x --
x x x x
x x x x
Quartz, plagioclase Corundum, kyanite, staurolite, sillimanite, A l - F e garnet
x x x x
x x x
x x , widespread; x , recorded; ( x ), very rare; - - , unknown; ?, unconfirmed. 1For sources see table 5 of Rock (1984); additional information in Kramer & Seifert (1984) and Jaques & Perkin (1984). 2Sources: Carstens 1958; Bayrakov 1964; Yeremenko & Shvakova 1969; Dobretsov & Dobretsova 1969; Azambre 1970; Brooks & Rucklidge 1973; Val'ter & Yeremenko 1974; Brooks & Platt 1975; Chapman 1975; Wallace 1975; Janse 1977; Rock 1977, 1983; Faerseth 1978; Leavy 1980; Larsen 1981; Praegel 1981; Mitchell & Janse 1982; Upton et al. 1983. 3 For sources see table 8 of Rock (1986). 4Sources: Carmichael 1967; Kay et al. 1978; Barton & Van Bergen 1981 ; Nixon et al. 1984; Jaques et al. 1984; Bergman 1987. 6 Navajo province, U.S.A. only (e.g. Ehrenberg 1979). vThose generally attributable to the disaggregation of xenoliths are in italics. 8 Aln6, Isle Bizard diatremes (Rock 1986). 9 Wandagee (W. Australia) only currently known example. 10 Probably includes both cognate megacrysts (phenocrysts) and xenocrysts.
index. M a n y o f t h e m g r a d e f r o m s h a r p l y d e f i n e d s u b s p h e r i c a l ocelli, w i t h d e l i m i t i n g t a n g e n t i a l biotite, to v a g u e p a t c h e s , b a r e l y d i f f e r e n t i a t e d f r o m t h e i r host ( R o c k 1979, 1983). The mineral compositions of such structures are q u i t e d i s t i n c t in t h e f o u r l a m p r o p h y r e b r a n c h e s ( T a b l e 1). G i v e n g o o d e x p e r i m e n t a l e v i d e n c e for s i l i c a t e - c a r b o n a t e liquid i m m i s c i bility ( H a m i l t o n et al. 1979), t h e c a r b o n a t e - r i c h s t r u c t u r e s in U M L p r o b a b l y r e p r e s e n t i m m i s c i b l e c a r b o n a t e d r o p l e t s ( ? c a r b o n a t i t e differentiate).
O p i n i o n s differ m u c h m o r e s h a r p l y o v e r t h e f e l d s p a t h i c s t r u c t u r e s in C A L a n d A L , h o w e v e r , with some workers advocating silicate-silicate liquid i m m i s c i b i l i t y ( P h i l p o t t s 1972, 1976,; E b y 1980) a n d o t h e r s a d v o c a t i n g s e g r e g a t i o n o f latestage liquid into vesicles ( C o o p e r 1979; F o l e y 1984; M a u g e r 1984). A p r o f u s i o n o f g e n e t i c suggestions exists for c a r b o n a t e - f e l d s p a r - c h l o r i t e - q u a r t z s t r u c t u r e s in "CAL a n d A L ( R o c k 1984), b u t i d e n t i f i c a t i o n as a m y g d a l e s or varioles has m u c h to c o m m e n d it ( J a f f e 1952; Z i m m e r l e
Nature of lamprophyres 1977). No assessment of these views can be made at least until microprobe and experimental data are available.
Heteromorphism CAL and AL rock types (Fig. l) are heteromorphic, as indicated by the coexistence of nearly isochemical rock types within single dykes, and by both local and global compositional merging (Rock 1977, 1984). Trivial heteromorphism, in which one rock type is merely a chilled version of another, is also found between holocrystalline and glassy lamprophyres (camptonites-hyalomonchiquites, orendites-wyomingites, jumillites-verites) (Rock 1977; Vollmer et aL 1984). Nevertheless, significant heterogeneities within the different branches also occur (Rock 1984, 1986). Heteromorphism is characteristic of lamprophyres, but it is not universal.
Global distribution Distribution through geological time Because they are partly characterized by texture and 'mobile'-element contents, lamprophyres are unlikely to survive even low-grade metamorphism. The older and/or more foliated they are, the more they may be confused with, say, microdior, ites (as in Johnson & Dalziel 1966). Some Precambrian 'lamprophyres' (Watson 1983 and personal communication) and many 'meta-lamprophyres' (Steiner 1984) are very doubtful. Hence the detailed distribution of lamprophyre magmatism through geological time may never be completely defined. Nevertheless, available age data show that lamprophyres are widely distributed from early Precambrian to Recent times (Table 13). As with kimberlites and carbonatites, certain times appear to have been particularly favourable for their emplacement: for example, the PermoTriassic and Jurassic for AL, and the Caledonian and Hercynian orogenies for CAL. LL alone appear to be restricted to a few age bands (Table 13). Some regional lamprophyre dyke-suites span periods of as much as 100 Ma in several distinct tectono-magmatic events (Baxter & Mitchell 1984).
Volumetric abundance Although individual intrusions are of small volume, numerous regional lamprophyre swarms, with thousands of dykes covering many thou-
215
sands of square kilometres, may be volumetrically equivalent to substantial plutons (Rock 1983). Watson (1984) estimated that lamprophyres and associated felsic minor intrusions make up 15% of Scottish Late Caledonian igneous rocks (i.e. several batholith equivalents), and crustal extensions reach 50% locally (Rock et al. 1985, 1986, 1987; Barnes et al. 1986). Although quantitative comparisons are not yet possible, lamprophyres probably have an aggregate global bulk considerably greater than kimberlites, carbonatites or leucite-bearing rocks and at least equal to nepheline syenites. Within lamprophyres, the order of abundance is almost certainly CAL AL >>UML g LL.
Range of tectonic settings (Table 2) Lamprophyres, like other alkaline rocks, are most abundant in continental rifts, failed arms of triple junctions (Brooks & Platt 1975) and on some oceanic islands (De Almeida 1955, 1961 ; Mitchell-Thom~ 1976). However, they also occur widely in orogenic zones and their peripheries (Himalayas (Viterbo & Zanettin 1959); Alps (Dal Piaz et al.1979); Pyrenees (Barrab~ 1952; Rock 1982); Caledonides (Richey 1939); Hercynides (Knill 1982)), in island arcs (Japan (Yagi et al. 1975); Solomons (Nixon et al. 1980)), in passive to destructive continental margins (W U.S.A. (Snaveley & Wagner 1961)), in anomalous uplifted fragments o f oceanic crust (Gorringe Bank (Cornen 1981, 1982); Bermuda (Aumento & Ade-Hall 1973; Aumento et aL 1974)) and near major transcurrent faults (Alpine Fault, New Zealand (Cooper 1979)). Collectively, therefore, lamprophyres are associated not only with intra-plate magmatism, but also with divergent, convergent and even passive plate-margin magmatism.
Petrogenetic significance of lamprophyres Throughout this section, the words primitive, primary and parental are used in the sense of Rhodes (1981). Previous petrogenetic theories for lamprophyres are legion (Rock 1977, 1984, 1986; Bergman 1987). This section concentrates on (1) unifying petrogenetic characteristics of the whole clan and (2) the increasingly recognized possibility that lamprophyres represent primary-mantle melts--by analogy with widely accepted models for alkaline basic to ultrabasic rocks (Frey et al. 1978). Since 'an essential requirement for any primary basalt candidate is that the bulk compo-
216
N. M. S. Rock
.o
~
o e0
0
""
..
o =
~ ~o ~
oov
0 ,.~ ~ ~ . ~
~
,...~
"~ 9
.~:~
"-".~..
~q=l
"
n:~r.~
x x x
x
~5:~zzz
>~_< *
x*
*
x
x
T'T"
x
~,-~l
x x
x
x
~
x
x
"-"
'~ 0 ~ - ' ' - " ~.~ :~,"~
x
x x
x
"
x x
"-"
~
~"
~ ~
~
~ ,~'~,.~
~
x
x
x x
.~
x
x
~
~
~
.~
x
x
~
~
~
~ "~
x
~
~I
"*"~
x
x~'x
~
~D
o
,,-, x:~ ~ C ~
~.
~ ~
, ~ r ~ I= . m
~
*.
X
X
X
X
~
~"~
X
,--,
~ n~
~
~"
0~ 4%) from the main volcanic centres of the Cameroon line. Number of analyses: Pagalu, 16; S~o Tom6, 77; Principe, 28; Bioko, 23; Etinde, 15; Mount Cameroon, 24; Manengouba, 15 ; Bambouto and Oku, 22; Ngaound6r6, 27; Biu, 35. Concentrations have been normalized to chondrite and primitive mantle abundances (Sun 1980). The spread in Th data is due to analytical uncertainty. chemical variation between centres it is clear that there are no systematic differences between the oceanic and continental suites. The basic volcanic rocks in each of the centres could have been generated by melting very similar mantle sources. Smaller degrees of melting followed by extensive crystal fractionation can account for the higher incompatible-element concentrations in the Etinde nephelinite lavas. The remarkable similarity between the oceanic and continental basic volcanic rocks is illustrated in Fig. 9 in which the average concentrations of incompatible elements in all the samples (except the Etinde nephelinites) from the two sectors are compared. The two suites of basic rocks are also isotopically indistinguishable (Fitton & Dunlop 1985). These data imply that the oceanic and continental basalts had very similar mantle sources. Since it is most unlikely that the ancient lithosphere mantle beneath the continental sector is chemically and isotopically similar to the young lithosphere mantle beneath the oceanic sector, it follows that lithospheric mantle was not the source of the Cameroon line basalts. The lack of any consistent migration of volcanism with time
285
rules out a source below the 670 km discontinuity. If the Cameroon line were the product of a deepmantle plume it would require that lower-mantle convection has kept pace exactly, in both velocity and direction, with the movement of the African plate over the past 65 Ma. Therefore the convecting upper mantle is the only plausible mantle source. The convecting upper mantle (of which the asthenosphere is the upper part) is also the source of MORB. Fitton & Dunlop (1985) proposed that the LILE-rich Cameroon line basalts could be derived from the LILE-depleted MORB source by small-degree (about 0.2%) melting using bulk partition coefficients consistent with experimentally determined values for mantle silicate phases. They suggested that the convecting upper mantle, although LILE depleted in its bulk composition, contains LILE-rich streaks. The presence of an ancient enriched component in the asthenosphere is necessary to account for the small but consistent isotopic differences between MORB and intraplate basalts. Such heterogeneities could be produced by the migration of magma through the asthenosphere (cf Hawkesworth et al. 1984). They may also be produced by the return of ocean crust (including ocean islands) to the asthenosphere during subduction. Whatever their origin, heterogeneities in the asthenosphere will be drawn out into streaks by convection (Hoffman & McKenzie 1985). Partial melts from these streaks will be selectively incorporated into smalldegree melts, whereas larger-degree melting will tend to homogenize the source and result in magmas (MORB) reflecting the bulk composition of the asthenosphere. T
,
T
CAMEROON
,
LINE
c
n3 lO0
so o
o Oceanic sector mean
lo
I
I
y
I
I
Rib Ba I~ NIb La cle Sir NId P Zr
1
Ti
I
Y
Fie. 9. Normalized abundance patterns (Sun 1980) for
incompatible-elementconcentrations in average continental (134 analyses) and oceanic (144 analyses) basic volcanic rocks from the Cameroon line. Data from the Etinde nephelinites are not included in the averages. (After Fitton & Dunlop 1985).
J. G. F i t t o n
286
from the asthenosphere, the mantle part of the lithosphere and the crust. Radiogenic isotope studies can constrain the possible magma sources in some cases but cannot always distinguish unambiguously between contributions from continental crust and enriched lithosphere mantle. In the case of the Cameroon line we can be confident that the basic magmas have not interacted to any significant extent with either the mantle or crustal parts of the lithosphere. This provides a unique opportunity of studying the effects of environment on the subsequent evolution of two essentially identical sets of basalt magmas during storage in the oceanic and continental lithosphere respectively. The compositional range of the Cameroon line volcanic rock samples collected from all the main centres is illustrated on an alkali-silica diagram in Fig. 10. Despite the similarity of the oceanic and continental basalts, the intermediate and evolved rocks from the two sectors show a clear divergence. With very few exceptions, the oceanic magmas evolve to phonolite and the continental magmas to rhyolite. The two samples of oceanic quartz trachyte (Fig. 10) were both collected from the Ilhbu das Cabras off S~o Tomb; the two continental phonolite samples are from plugs on the Ngaoundbrb Plateau. Apart from these four samples, the separation of oceanic and continental salic rocks is complete. This is further illustrated in Fig. 11 in which the CIPW norms of the salic rocks have been projected into the residua system (quartz-nepheline-kalsilite).
McKenzie (1984, 1985) has recently derived a set of equations describing the compaction of partially molten rock and concluded that melt will begin to flow at very small degrees of melting (less than 0.570) provided that its viscosity is low. The degrees of melting required to produce LILErich intra-plate basalts from the MORB source should no longer be regarded as impossibly small. If the Cameroon line basic magmas were generated by small-degree melting of the MORB source then the same must also be true of other intra-plate basic magmas. This has been discussed by Fitton & James (1986) who compared incompatible-element concentrations in a wide range of intra-plate volcanic rocks with calculated concentrations in liquids generated by variable degrees of equilibrium melting of the MORB source. Partition coefficients required by the Cameroon line data were used in the calculations. The comparisons are encouraging and suggest that ocean island and rift valley magmas may share a common asthenosphere source. Metasomatic enrichment of the source is not necessary.
Origin of the salic rocks The extent to which contamination of magmas by crustal rocks influences their composition and evolution has been debated for many years (see, for example, Moorbath et al. 1984). Magmas in continental provinces may receive contributions
16
L
-s 9
14
9
o
s
:
~ 9
12 0 Z + 10 0 Z
s
eo 9
O ,
0
:3
9
9
9
9O 9
~
t
o
O0
9 O
9 A
8
A 9
%6
0 o ~ 00
0 0
9 9
9
0 " "
0
9
~
9
9
9
9 09
9
9
9
4
A ~-'~3~:
~ 2
0
0
4
,0/
w
~ /oo /
4t5
9 Etinde nephelinites 9 Other continental sector volcanic rocks 9 Oceanic sector volcanic rocks 510
L
55
~ 60
J 65
t 70
J 75
Weight % SiO 2
FIG. 10. Alkali-silicadiagram for the Cameroon line volcanic rocks. The line separating Hawaiian tholeiitic and alkaline rocks (Macdonald & Katsura 1964) is shown for comparison. (From Fitton 1983.)
The Cameroon line
287
Q
9 Continental sector o Oceanic sector s from granulite xenoliths
Ne
Weight %
Ks
FIG. 11. Normative compositions of evolved (more than 80%) normative salic components) volcanic rocks from the Cameroon line projected into the residua system. Additional data from Manengouba (supplied by C. A. Hirst) are included. Phase boundaries (P(H20)= 5 kb) are taken from Hamilton & MacKenzie (1965). The oceanic rocks occupy the thermal valley between the trachyte and phonolite minima, whereas the continental rocks plot in the corresponding valley down to the granite minimum. This separation into undersaturated oceanic and oversaturated continental salic rocks must be related to the environment in which the magmas evolved. The basaltic parental magmas in both sectors of the Cameroon line are transitional to strongly-alkaline in character (Figs 7 and 10). Low-pressure crystal fractionation alone should therefore have produced salic rocks ranging from phonolite to trachyte, with only small amounts of oversaturated magma (as in the oceanic sector). Progressive crustal contamination of the fractionating magmas provides the simplest explanation for the dominance of oversaturated salic rocks in the continental sector. Evidence for crustal contamination is provided by the occurrence of partly digested granulite xenoliths in a basalt lava flow from Bambouto, near to the village of Babadjou. The xenoliths are partially melted and contain patches of fresh glass in a residual mineral assemblage of quartz, plagioclase, orthopyroxene and minor-alkali feldspar. Electron microprobe analyses of the glass
show it to have a composition close to the minimum on the quartz-alkali feldspar cotectic (Fig. 11). Contamination of the continental magmas with crustal rocks or their partial melts could allow the magmas to cross the low-pressure thermal divide and evolve towards rhyolite. Dunlop's (1983) strontium isotope study of the Cameroon line volcanic rocks provides further evidence for crustal contamination. Although the oceanic and continental basalts are generally indistinguishable in their isotope ratios, the distribution of 87Sr/86Sr in the continental basalt samples is skewed slightly towards higher values (Fitton & Dunlop 1985). The contaminated basalt from Bambouto has the highest initial 875r/86Sr ratio (0.70403) of all the continental basic lavas analysed. It is in the evolved rocks, however, that crustal contamination can be most clearly demonstrated. The trachytes and rhyolites from Bambouto, Oku and the Mandara Mountains have initial 87Sr/86Sr values ranging up to 0.715. For comparison, two of the granulite xenoliths from Bambouto have age-corrected 875r/86Sr ratios of 0.71021 and 0.72088 (Dunlop 1983). A possible alternative to crustal contamination as a means of crossing the low-pressure thermal
J. G. F i t t o n
288
divide is provided by fractional crystallization of hornblende-bearing assemblages. Cawthorn et al. (1973), for example, suggested that removal of silica-poor ne-normative hornblende from undersaturated basic magma could lead to the evolution of silica-oversaturated residual liquids. This mechanism cannot apply, however, to the Cameroon line magmas since hornblende is seldom found as a phenocryst phase in the continental volcanic rocks. Curiously, though, hornblende phenocrysts are abundant in the intermediate lavas and even in some of the basalts in the oceanic sector. The abundance of hornblende in the oceanic lavas and its scarcity in the continental rocks is puzzling since the basic magmas in the two sectors were compositionally identical (Fig. 7). The most important factor controlling amphibole stability in basaltic magma is the partial pressure of water vapour (Allen & Boettcher 1978). It is possible that magma stored beneath the islands of the oceanic sector had easier access to water than did the continental magma reservoirs, and that seepage of small amounts of sea-water into the oceanic magma reservoirs stimulated the crystallization of hornblende. The middle and late stages in the evolution of magmas on S~.o Tom6 are dominated by fractionation of hornblende-bearing assemblages (Fig. 5). This is clearly demonstrated by the abundance of apparently cognate hornblende-rich cumulate
blocks. The operation of this process in the oceanic magmas, but not in those of the continent, is well illustrated in a plot of Y against Zr (Fig. 12). Y is a compatible element in amphiboles (Pearce and Norry 1979), and so hornblende removal would cause the concentration of Y to fall. Zr, which is incompatible with all the observed phenocryst phases, is a useful index of fractionation. Figure 11 shows that the Y concentration rises with fractionation in the continental lavas, where hornblende is seldom present, but falls in the intermediate and evolved oceanic sector rocks. The hornblende-bearing cumulate blocks from S~o Tomb are relatively rich in Y. Fractional crystallization involving the removal of this cumulate material from the basic to intermediate magmas (cf. Fig. 5) would increase the Zr/Y ratios in the residual liquids along a line such as that on Fig. 12. This line was constructed by applying the Rayleigh fractionation equation to the average oceanic-sector basic lava composition (from Fitton & Dunlop 1985). Bulk partition coefficients for Zr and Y were taken as the ratio of their average concentrations in the cumulate blocks to the average in the oceanic-sector basic lavas. The degrees of fractionation indicated by the numbers on this line depend heavily upon the concentrations of Zr and Y in the crystal extract and serve only to show that the array of oceanicsector data points could be generated by crystal
/
300
/ 9 Etinde nephelinites
200
/ / /
9 Other continental s e c t o r volcanic r o c k s 9 O c e a n i c s e c t o r volcanic r o c k s
if
/
H o r n b l e n d e - r i c h cumulate b l o c k s
/
..../ ....
/
.//"
loo
j'/
.........,
,/
1.-E
-
-/
.............
9
Q
d'/O ....
~.// /
o
.. .....
o
.......---
o
--b-___.Z~ .@ ~ ./ ~-_:' o o y_.~; oo c~o-_ _--- ~ ~9 / / / -"-~ L /
3
~f
// L,//
10100
........ 9 9 /$~O
LI 9
9 2
20
9
1 " ~'
....
;.%
/ 9 6,1.9 i 9 1 4 .9~ k ~ C z"
_
9
9 9
~,
/1.
-'/
50
,~/~ O ///z~,~. ~
9
/z
from S&o Tome
>-
///"
200
.-
~..// I
I
I
300
400
500
pp.m
I
I
L
I
d
1000
2 O0
Zr
FIG. 12. Y and Zr contents of the Cameroon line volcanic rocks and cognate xenoliths. The fractionalcrystallization path has been calculated by subtracting the average xenolith composition from the average oceanic-sector basic volcanic rock composition. The percentages indicate the amount of crystallization.
The Cameroon line f r a c t i o n a t i o n o f h o r n b l e n d e - b e a r i n g assemblages. T h e r e is no e v i d e n c e t h a t r e m o v a l o f hornb l e n d e has c a u s e d the oceanic-sector m a g m a s to b e c o m e less u n d e r s a t u r a t e d . Virtually all these m a g m a s h a v e e v o l v e d t o w a r d s the p h o n o l i t e m i n i m u m (Figs 10 a n d 11) w h i c h shows clearly t h a t crystal f r a c t i o n a t i o n i n v o l v i n g h o r n b l e n d e is not an efficient m e c h a n i s m for crossing the lowpressure t h e r m a l divide.
289
ACKNOWLEDGMENTS.Field work in Cameroon, Nigeria and the Gulf of Guinea was financed by research grants from the U.K. Natural Environment Research Council and The Royal Society. The author is indebted to D. J. Hughes for his assistance with the field work, to H. M. Dunlop, C. A. Hirst and R. M. Maclntyre for permission to use their unpublished data, and to D. James and G. R. Angell for assistance with the chemical analyses. D. James, B. G. J. Upton and P. Vincent are thanked for their constructive comments on an earlier draft of this paper.
References ALLEN, J. C. & BOETTCHER,A. L. 1978. Amphiboles in andesite and basalt: II. Stability as a function of P - T - f H 2 0 - f O2. Am. Mineral. 63, 1074-87. BAILEY, D. K. 1982. Mantle metasomatism--continuing chemical change within the Earth. Nature, Lond. 296, 525-30. 1987. Mantle metasomatism--perspective and prospect. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, 1-13. BOWDEN, P., BLACK, R., MARTIN, R. F., IKE, E. C., KINNAIRD, J. A. & BATCHELOR,R. A. 1987. NigerNigerian alkaline ring complexes: a classic example of African Phanerozoic anorogenic mid-plate magmatism. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, 357-379. CANTAGREL, J.-M., JAMOND, C. (~; LASSERRE,M. 1978. Le magmatism alkalin de la ligne du Cameroun au Tertiaire inf6rieur: donn6es g6ochronologiques K--Ar. C. R. Somm. Soc. g~ol. Ft. 6, 300-3, CARTER, J. D., BARBER, W. d~; TAIT, E. A. 1963. The geology of parts of Adamawa, Bauchi and Bornu Provinces in northeastern Nigeria. Bull. geol. Surv. Nigeria, 30. CAWTHORN, R. G., CURRAN, E. B. & ARCULUS, R. J. 1973. A petrogenetic model for the origin of the calc-alkaline suite of Grenada, Lesser Antilles. J. Petrol. 14, 327-37. CHAPUT, M., LOMBARD,J., LORMAND,J. & MICHEL, H. 1954. Granites et traces d'btain dans le Nord Cameroun. Bull. Soc. gbol. Ft. 4, 373-93. CLAGUE, D. A. & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. CORNEN, G. & MAURY, R. C. 1980. Petrology of the volcanic island of Annobon, Gulf of Guinea. Mar. Geol. 36, 253-67. DI~RUELLE, B., MOREAU, C. & NSIFA, E. N. 1983. Sur la r6cente 6ruption du Mont Cameroun (16 octobre-12 novembre 1982). C. R. Acad. Sci. Paris, 296, 807-12. DUNLOP, H. M. 1983. Strontium isotope geochemistry
and potassium-argon studies on volcanic rocks from the Cameroon line, West Africa. PhD Thesis, University of Edinburgh (unpublished). -& FITTON, J. G. 1979. A K - A r and Sr-isotopic study of the volcanic island of Principe, West Africa--evidence for mantle heterogeneity beneath the Gulf of Guinea. Contrib. Mineral. Petrol. 71, 125-31. EMERY, K. O. & UCHUPI, E. 1984. The Geology of the Atlantic Ocean, Springer, New York. ESCH, E. 1901. Der Vulcan Etinde in Kamerun und seine Gesteine (I). Sitzungsber. Akad. WiNs. Berlin, 277-99. FITTON, J. G. 1980. The Benue trough and Cameroon line--a migrating rift system in West Africa. Earth planet. Sci. Lett. 51, 132-8. -1983. Active versus passive rifting: evidence from the West African rift system. Tectonophysics, 94, 473-81. - - & DUNLOP, H. M. 1985. The Cameroon line, West Africa, and its bearing on the origin of oceanic and continental alkali basalt. Earth planet. Sci. Lett. 72, 23-38. -& HUGHES, D. J. 1977. Petrochemistry of the volcanic rocks of the island of Principe, Gulf of Guinea. Contrib. Mineral. Petrol. 64, 257-72. & -1981. Strontian melilite in a nephelinite lava from Etinde, Cameroon. Mineral. Mag. 44, 261-4. t~ JAMES, D. 1986. Basic volcanism associated with intraplate linear features. Phil. Trans. R. Soc. Lond., Ser. A, 317, 253-66. - - , KILBURN, C. R. J., THIRLWALL,M. F. & HUGHES, D. J. 1983. 1982 eruption of Mt. Cameroon, West Africa. Nature, Lond. 306, 327-32. FREETH, S. J. 1979. Deformation of the African plate as a consequence of membrane stress domains generated by post-Jurassic drift. Earth planet. Sci. Lett. 45, 93-104. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from southeastern Australia utilising geochemical and experimental petrological data. J. Petrol. 19, 463-513. -
-
-
-
29o
J. G. Fitton
FRISCH, T. & WRIGHT, J. B. 1971. Chemical composition of high-pressure megacrysts from Nigerian Cenozoic lavas. Neues Jb. Mineral. Monatsh. 289304. GAZEL, J. 1956. Carte Gdologique du Cameroun au 1/1,000,000. Direction des Mines et de la G6ologie du Cameroun, Paris. - - , LASSERRE, M., LIMASSET,J.-C. & VACHETTE, M. 1963. Ages absolus des massifs granitiques ultimes et de la min~ralisation en &ain du Cameroun central. C. R. Acad. Sci. Paris, 256, 2875-8. GI~ZE, B. 1943. G6ographie physique et g~ologie du Cameroun occidental. Mem. Mus. natl Hist. nat. Paris, Nouv. Skr. 17, 1-272. -1953. Les volcans du Cameroun occidental. Bull. Volcanol. 13, 63-92. GORINI, M. A. & BRYAN, G. M. 1976. The tectonic fabric of the Equatorial Atlantic and adjoining continental margins: Gulf of Guinea to northeastern Brazil. An. Acad. Bras. Ci~nc. 48 (Suppl.), 10119. GOUHIER, J., NOUGIER, J. & NOUGIER, D. 1974. Contribution ~t l'&ude volcanologique du Cameroun ("Ligne du Cameroun"--Adamaoua). Annls Fac. Sci. Cameroun, 3-48. GRANT, N. K., REX, D. C. & FREETH, S. J. 1972. Potassium-argon ages and strontium isotope ratio measurements from volcanic rocks in northeastern Nigeria. Contrib. Mineral. Petrol. 35, 277-92. GREEN, D. H. & RINGWOOD, A. E. 1967. The genesis of basalt magmas. Contrib. Mineral. Petrol. 15, 103-90. GRUNAU, H. R., LEHNER, P., CLEINTUAR, M. R., ALLENBACH, P. & BAKKER, G. 1975. New radiometric ages and seismic data from Fuerteventura (Canary Islands), Maio (Cape Verde Islands) and S~,o Tom6 (Gulf of Guinea). In: BORRADAILE, G. J., RITSEMA,A. R., RONDEEL, H. E. & SIMON, O. J. (eds) Progress in Geodynamics, pp. 90-118. North-Holland, Amsterdam. HAMILTON, D. L, & MACKENZIE, W. S. 1965. Phase equilibrium studies in the system NaA1SiO,, (nepheline)-KA1SiO4 (kalsilite)-SiO2-H20. Mineral. Mag. 34, 214-31. HARRIS, P. G. 1957. Zone refining and the origin of potassic basalts. Geochim. cosmochim. Acta, 12, 195-208. HAWKESWORTH, C. J., ROGERS, N. W., VAN CALSTEREN, P. W. C. & MENZlES, M. A. 1984. Mantle enrichment processes. Nature, Lond. 311,331-5. HEDBERG, J. D. 1968. A geological analysis of the Cameroun Trend. PhD Thesis, Princeton University (unpublished). HOFFMAN, N. R. A. & MCKENZIE, D. P. 1985. The destruction of geochemical heterogeneities by differential fluid motions during mantle convection. Geophys. J. R. Astron. Soc. 82, 163-206. JACQUEMIN, H., SHEPPARD, S. M. F. & VIDAL, P. 1982. Isotopic geochemistry (O, Sr, Pb) of the Golda Zuelva and Mboutou anDrogenic complexes, North Cameroon: mantle origin with evidence for crustal contamination. Earth planet. Sci. Lett. 61, 97-111.
LASSERRE, M. 1978. Mise au point sur les granit6ides dits "ultimes" du Cameroun: gisement, p6trographie et g+ochronologie. Bull. Bur. Rech. gdol. min., Paris, Ser. 2, Sect IV, no. 2, pp. 143-59. LIOTARD, J. M., DUPUY, C., DOSTAL,J. & CORNEN, G. 1982. Geochemistry of the volcanic island of Annobon, Gulf of Guinea. Chem. Geol. 35, 11528. LLOYD, F. E. & BAILEY, D. K. 1975. Light element metasomatism of the continental mantle: the evidence and the consequences. Phys. Chem. Earth, 9, 389-416. MACDONALD, G. A. & KATSURA, T. 1964. Chemical composition of Hawaiian lavas. J. Petrol. 5, 82133. MCKENZIE, D. 1984. The generation and compaction of partially molten rock. J. Petrol. 25, 713-765. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-9 I. MENZIES, M. A. & MURTHY, V. R. 1980. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts : implications for local heterogeneities in metasomatically veined mantle. Earth planet. Sci. Lett. 46, 323-34. MITCHELL-THOMi~, R. C. 1970. Geology of the South Atlantic Islands. Gebruder Borntraeger, Berlin. MOORBATH, S., THOMPSON, R. N. & OXBURGH, E. R. (eds) 1984. The Relative Contributions of Mantle, Oceanic Crust and Continental Crust to Magma Genesis, 342 pp. The Royal Society, London. OKEKE, P. I. 1980. Petrology of igneous and metamorphic rocks in the area around Gwoza, northeast Nigeria. M. Phil. Thesis, University of Edinburgh (unpublished). PARSONS, I., BROWN, W. L. & JACQUEMIN, H. 1986. Mineral chemistry and crystallization conditions of the Mboutou layered gabbro-syenite-granite complex, North Cameroon. J. Petrol. 27, 13051330. PEARCE, J. A. & NORRY, M. J. 1979. Petrogenetic implications of Ti, Zr, Y, and Nb variations in volcanic rocks. Contrib. Mineral. Petrol. 69, 33-47. PIPER, J. D. A. & RICHARDSON, A. 1972. The palaeomagnetism of the Gulf of Guinea volcanic province, West Africa. Geophys. J. R. Astron. Soc. 29, 147-71. PLATT, R. G. & EDGAR, A. D. 1972. The system nepheline-diopside-sanidine and its significance in the genesis of melilite and olivine bearing alkaline rocks. J. Geol. 80, 224-36. SEAN (Scientific Event Alert Network) 1985. Bulletin, 10(8), 3-4. -1986. Bulletin, 11(8), 2-5. SmUET, J.-C. & MASCLE, J. 1978. Plate kinematic implications of Atlantic equatorial fracture zone trends. J. geophys. Res. 83, 3401-21. SMITH, A. G., HURLEY, A. M. & BRIDEN, J. C. 1981. Phanerozoic Palaeocontinental World Maps, 102 pp. Cambridge University Press, Cambridge. SUN, S.-S. 1980. Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Phil. Trans. R. Soc. Lond., Ser. A, 297, 409-45. -
-
The Cameroon line TCHOUA, F. 1972. Sur la formation des calderas des Monts Bambouto (Cameroun). C. R. Acad. Sci. Paris, 274, 799-801. 1973. Sur l'existence d'une phase initiale ignimbritique dans le volcanisme des Monts Bambouto (Cameroun). C. R. Acad. Sci. Paris, 276, 2863-6. TYRRELL, G. W. 1934. Petrographical notes on rocks from the Gulf of Guinea. Geol. Mag. 71, 16-23. VAN HOUTEN, F. B. 1983. Sirte Basin, north-central Libya: Cretaceous 'rifting above a fixed mantle hotspot? Geology 11, 115-8. -
-
2 91
VINCENT, P. 1968. Attribution au Cr6tac~ de conglom6rats m6tamorphiques de l'Adamaoua (Cameroun). Annls Fac. Sci. Cameroun, 1, 69-76. WINCHESTER,S. 1984. The nastiest place on Earth? The Sunday Times Magazine, Jan. 29, 1984, pp. 30-37. WRIGHT, J. n. 1970. High pressure phases in Nigerian Cenozoic lavas: distribution and setting. Bull. VolcanoL 34, 833-47. WRIGHT, T. L. 1984. Origin of Hawaiian tholeiite: a metasomatic model. J. geophys. Res. 89, 3233-52.
J. G. FITTON, Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Outline of the petrology of the Kenya rift alkaline province B. H. Baker S U M M A R Y : The Kenyan and N Tanzanian volcanic province contains sodic alkaline rocks ranging from melilitites and melanephelinites to transitional alkali basalts and their differentiates. Individual volcanoes display three principal magmatic suites: (i) nephelinitic; (ii) alkali basaltic; (iii) transitional basaltic. However, some large volcanoes contain more than one of these suites implying that parental magmas of variable alkalinity were available at certain times and places. A general decrease in alkalinity with time is detectable in the rift zone and for any time period there was a tendency for the least alkaline magmas to be erupted within the central and deepest part of the rift zone. Compositional variation within the suites was largely controlled by low-pressure crystalliquid fractionation. Extended fractionation produced salic differentiates. Liquid fractionation caused upward segregation of phonolitic and trachytic magmas which were erupted in preference to more mafic magmas. Isotopic data suggest that crustal contamination did not occur on a large scale.
Introduction The upper Cenozoic alkaline igneous province associated with the eastern branch of the E African rift system in Kenya and northern Tanzania contains nearly the whole spectrum of sodic alkaline igneous compositions, and includes unusually large proportions of intermediate and silicic rocks. The volcanic and tectonic evolution of the rift valley during the last 30 Ma is now moderately well understood, providing an opportunity to assess interrelations between volcanism and tectonism in time and space, to summarize the petrology and geochemistry of the province and to make inferences concerning petrogenetic processes. This review is an outline of the characteristics of the province, and is biased towards areas and suites familiar to the author. Much of the region is known only from reconnaissance mapping and study of thin sections, supported by a few rock analyses. Few volcanic suites have been analysed systematically for a wide range of geochemical elements. The geochemical data used to prepare variation diagrams include the majority of wellstudied suites of alkaline rocks in Kenya, including some unpublished data. Apart from questions concerning the petrogenesis of individual volcanic series, the major problems of the province concern conditions under which parental magmas were formed, and the characteristics of the mantle source. The fact that there are some systematic variations in the alkalinity of lavas in space and time invites consideration of the tectonic effects of rifting. The bimodality of transitionally alkaline suites and large volumes of salic rocks raise questions concerning low-pressure differentiation processes
and the degree to which crustal contamination is involved. In view of recent progress in understanding physical processes in magma chambers the opportunity is taken to comment on how such processes might influence petrogenesis and eruptive mechanisms.
General characteristics of the province The alkaline volcanic province in Kenya and northern Tanzania is associated with the development of the eastern branch of the E African rift system (Fig. 1). Volcanism began about 30 Ma ago in northern Turkana and extended progressively southward, becoming more voluminous 16 Ma ago and continuing to the present (Baker et al. 1971; Williams 1978; Williams & Truckle 1980). In Kenya and northern Tanzania about 220 000 km 3 of rock were erupted (Williams 1982), of which 68% are mafic rocks (King 1978). This amounts to a dense rock equivalent eruption rate of about 0.006 km 3 per year, but the parental magma production rate must have been at least an order of magnitude greater than this. Tectonic evolution of the rift began with development of a shallow basin in the Turkana region in the N sometime in the early Miocene, in which some 35 000 km 3 of alkali basalt lavas were erupted (Fig. 1). By 15 Ma BP faults had begun to develop on the western side of this depression, and alkali basalts and voluminous phonolites were erupted in a developing halfgraben (Fig. 2, upper part). By 7 Ma BP a major fault extended along the
From: Fir'roy, J. G. & UVTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 293-311.
293
294
B. H. Baker
FXG. 1. Geological map of the alkaline igneous province of Kenya and northern Tanzania (bold lines are major faults): 1, nephelinite-carbonatite suite; 2, Miocene alkali basalts; 3, Miocene P-phonolites; 4, mixed association volcanoes; 5, Pliocene alkali basalts; 6, Pliocene to Recent transitional basalts and trachytes; 7, Pliocene to Recent alkali basalts of the E rift flank (clusters of vents shown as dots). Abbreviations are volcano and formation names referred to in the text. western side of the rift, which was a half-graben bounded by a monoclinal flexure on its eastern side. At this time there was a change to less alkaline volcanism characterized by the building of basalt-trachyte shield volcanoes, although more alkaline central volcanism continued at intervals in the Kavirondo rift, Uganda borderland and N Tanzania. By 4 Ma BP major faults were developing on the eastern side of the rift giving rise to a deeping graben structure in which basalt-trachyte volcanism continued to predominate. It is convenient to divide the tectonic development into pre-rift (3012MaBP), half-graben ( 1 2 - 4 M a BP) and graben stages of development (Fig. 2). The elevation of the flanking plateaux of the
rift valley is due in part to local areas of residual highlands composed of Precambrian rocks, to the accumulation of lavas that periodically filled the rift and overflowed its flanks, and to late uplifts which reached a maximum of 1700 m in the central region (Saggerson & Baker 1965; Williams et al. 1983). The thickness of volcanic rock accumulated in the central section of the rift valley is at least 3 km, and it can be inferred that the sub-volcanic surface is below sea level along much of the rift, being deepest in its central section. Volcanism of the pre-rift stage was dominantly of many small basaltic shields in the Baringo and Turkana regions, with scattered contemporaneous nephelinite-carbonatite central volcanism in the K e n y a - U g a n d a borderland (King et al. 1972) and the Kavirondo rift valley (Le Bas 1977). During the half-graben stage alkali basalts and voluminous phonolites were erupted, but by 7 Ma BP the character of volcanism changed to a transitionally alkaline character marked by the growth of many bimodal basalt-trachyte volcanoes, many of them being calderas with ash-flow tufts (Webb & Weaver 1975 ; Williams 1978). During Plio-Pleistocene times the eastern flank of the rift was the site of widespread fissure volcanism that built chains of many monogenetic alkali basalt cones, such as the Chyulu, Nyambeni and Hurri ranges, and isolated low basaltic shields such as the Marsabit volcano (Fig. 1, Ch, Ny, Hu and Ma). On or near the eastern edge of the rift large central volcanoes composed of a great variety of lavas formed Mount Kenya and Kilimanjaro (Ke and Ki). Within the rift contemporaneous Quaternary volcanism consisted of voluminous eruptions of trachyte and alkali rhyolite, and the building of a chain of basalttrachyte caldera volcanoes (Fig. 1, Su, Lo, Mi, Si and Em).
Associations of alkaline igneous rocks The Kenyan sodic igneous rocks have been divided into suites of contrasting alkalinity, consisting of a strongly alkaline suite of nephelinites, basanites, alkali basalts, tephrites and phonolites, and a weakly alkaline suite of olivine basalts, trachytes, trachyphonolites and alkali rhyolites (Saggerson & Williams 1964; Williams 1969a, 1970, 1972; King 1970; Saggerson 1970; King et al. 1972; Baker et al. 1978; Williams & Truckle 1980). These suites are characterized by normative ne greater and less than 5%, and within the rift zone are broadly represented by Miocene and post-Miocene volcanism (Williams 1972).
Petrology of the Kenya rift A-NORTH My
W flank
o
295
SECTOR
--------- R i f t
,,
2
~
E
flank
=~,~--~~
4 6:
a:
~,~__~_._t_r.__~__=~:~-~'~-~.It i::::::::::::..~.
II i'
10-
12: 14: ]6: 18: 20: 22: 24:
"l
o
2: 4"6:
".... I "'~;b
=====================[
depression
J Ii
My
half graben
B-SOUTH W
=
SECTOR Rift
l
~
~ ~ ~ :
AI
~-~', b --'-,-'_---_"-~_~--7~L.-':~Zb ~ i i ~ i ~ ~ ~:= tr 1
b'~l~l~i~A
A
E
I t,
_tr
~'~..'1"...
igraDell ] half -
j~graben
8: io:
12:
unexposed basanitoid rocks ? . . . . . . . . . . . 7 " . ~ Z Z Z z ~ " ..........
I
Nephelin ite-pho no[ ite-(carbonat ite) Basanite/atk.basalt- phonotite
~
Mixed associations
Basatt - t rachyte-rhyolite/t rachyphonoii te b-basaqite or basalt; ph-phonotite; tr-trachyte; tph-tr&chyphono[ite I periods of faulting FIG. 2. Diagrammatic summary of the ages and extent of lavas of different associations and their relation to phases of rift faulting in the northern and southern sectors of the rift.
Many varied associations can be recognized among the more strongly alkaline rocks (Miyashiro 1978), and it is convenient to divide the strongly alkaline suite further on the character of its differentiated rocks into a nephelinitic suite associated with sphene-bearing phonolites and carbonatites and an alkali olivine basalt-basanite suite associated with sphene-free phonolites. Many volcanoes consist of one or other of these suites, but some large central volcanoes contain
representatives of both and are referred to as 'mixed' association volcanoes. By using association in space and time as an indicator of petrogenetic relationship, the province can be divided into rocks that fall into four types of volcanic association. l Melilitites, melanephelinites, ankaratrites, nephelinites, and G-type phonolites characterized by high Sr and Ba (Table 1 ; Table 2, analyses
296
B. H. Baker
TABLE 1. Nephelinite--phonolite suite 1
2
3
4
5
6
Major elements (wt.~)
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO NazO K20 P2Os
43.0 3 9 . 0 4 9 . 2 5 5 . 4 40.5 43.8 3.1 3.7 1.3 0.5 2.4 2.5 11.2 8.3 18.3 2 0 . 8 15.4 12.9 8.2 6.9 5..0 3.5 6.7 7.1 6.3 7.1 2.9 1.5 3.4 5.2 0.2 0.2 0.2 0.2 0.2 0.3 7.8 9.7 1.7 0.5 3.4 6.0 14.2 19.3 7.2 2.9 16.6 12.7 5.9 3.4 2.4 9.1 9.2 7.2 2.8 1.8 1.9 4.6 5.5 3.1 0.7 0.8 1.2 0.4 0.1 0.9
FeO* Mg~
13.7 50.3
13.3 7.5 4.7 9.5 5 6 . 9 2 9 . 3 14.7 40.1
11.5 48.2
Alkaline norm in we~htpercentvo~ti&~ee~
Ap Mt
Sp Pv Di Geh Akr Wo Ac O1 Or Ab Ne Ks Cg
3.0 7.5 4.8 22.9 10.3
5.5 7.2 19.1 19.7
4.5 1.0 5.9 1.6 2.1 1.3 1.5 3 8 . 4 10.4 5.9
0.3 2.0 0.6
0.5 0.6
0.3
3.8 4.3
4.4
2.5 4.3 1.2 17.0 0.4 0.2 3.7
5.8 13.7 11.6 27.8 36.1 6 3 . 9 50.3 68.4 1.3 0.1 0.1 2.2
2.1 4.8 2.9 23.8 0.3
0.7 1.9 7.7 55.9
Average analyses from Kavirondo rift valley and eastern Uganda (Le Bas 1977). The number of analyses is given in parentheses:l, melanephelinites (43); 2, melilitites (6); 3, nephelinites (14); 4, G-phonolites (55); 5, ijolites (41); 6, estimated parent magma. * Total Fe as FeO t Cation ratio Mg/(Mg + Fe) :~From Le Bas 1973. 1-4). Older volcanoes of this type may have exposed sub-volcanic intrusions composed of ijolite, nepheline syenite and carbonatite. 2 Limburgites, basanites, alkali basalts, tephrites, and P-type phonolites characterized by low Sr and Ba (Table 2, analyses 7 and 8; Table 3, analyses 4 and 7). 3 Transitional basalts, ferrobasalts, scarce mugearites and benmoreites, abundant trachytes, and alkali rhyolites (Table 3, analyses 1-3, 9 and 10; Table 4). Locally the differentiated rocks are trachyphonolites, and in a few suites oversaturated and undersaturated silicic rocks occur on the same volcano.
4 Mixed association central volcanoes which contain usually two of the suites mentioned above, the suites being erupted either sequentially or in alternating phases (Figs 1 and 2). The fields of the first three suites are shown on a Q L M diagram (Fig. 3(a)) in which a continuum of compositions between the transitional (T-ba) and alkali basalt (A-ba) suites, and a gap between the alkali basalt and nephelinitic (M-ne) suites at mafic and intermediate compositions can be seen. The silicic rocks cover the entire range from the oversaturated (Rhy) to the undersaturated (GPh) minima. The trend lines for individual volcanic series are shown in Fig. 3(b). The transitional basalt series exhibits a marked Daly gap, and is associated with voluminous trachytes (TR) and comenditic and pantelleritic rhyolites (RH). Some of the trachytic series contain both rhyolitic and trachyphonolite (T-ph) lavas, suggesting that oversaturated and undersaturated trends can develop from a saturated trachyte stem. Trends within the alkali olivine basalt-basanite suite are more variable, but all trend toward, or end in, the field of P-phonolites (PP). In these suites intermediate rocks such as tephrites and rhomb porphyries are less well represented than the end-member basalts and phonolites. In the strongly-alkaline series, melilitites (ME) and melanephelinites (NE) are usually accompanied by nephelinites and intrusive members of the ijolite series. Strongly-alkaline phonolites (GP) with ne greater than 15% are present in subordinate amounts. There is a tendency for the Daly gap to be most marked in the least alkaline series, becoming absent in the most alkaline series. The phonolitic end-members of nephelinitic and basanitic suites are scarcely distinguishable on the basis of major-element compositions, since both suites converge on the phonolite minimum, but are readily identified by their contrasting Sr, Ba and light rare-earth element (REE) contents (see below).
Distribution and volcanological characteristics of the suites The nephelinite-carbonatite association is represented by large central volcanoes composed largely of pyroclastic rocks and debris flows, exemplified by Mounts Elgon, Napak, Meru, Hanang and O1 Doinyo Lengai. They range in age from 22 Ma to recent, and are found on the N W flank of the rift in the K e n y a - U g a n d a borderland (Davies 1952; King 1965), along the Kavirondo branch of the rift and in southern
Petrology of the Kenya rift
297
TABLE 2. Strongly- and mildly-alkaline suites 1
Major elements (wt.~) SiO2 41.4 TiO2 2.5 Al103 9.2 Fe203 5.9 FeO 6.6 MnO 0.2 MgO 13.6 CaO 13.5 Na20 3.4 K20 1.4 H2O+ 1.8 PzOs 0.5 Trace elements Sc Ni Co Zr Ta Th Rb Sr Ba La Ce Nd Sm Eu Tb Yb
(ppm) 40 240 36 244 6.5 8.2 42 ND ND 63 145 53 10.5 2.5 ND 2.3
2
3
4
5
6
7
8
41.5 2.7 11.9 6.6 5.7 0.2 6.1 14.6 5.7 2.6 1.8 0.9
46.2 1.6 18.6 6.0 3.5 0.2 2.3 7.3 9.3 4.2 1.0 0.5
51.7 0.9 19.3 3.9 2.4 0.2 1.1 4.1 8.9 4.6 2.6 0.3
45.8 2.5 14.0 4.9 8.1 0.3 8.0 10.8 2.9 1.2 1.0 0.5
61.5 0.7 13.7 4.5 4.5 0.3 0.5 1.3 6.8 4.9 1.8 0.3
48.6 3.4 15.1 5.8 5.5 0.2 4.2 8.5 4.1 2.4 ND 0.8
44.3 2.9 12.1 3.3 10.0 0.2 11.3 10.6 2.9 1.1 ND 0.6
16 38 38 263 7.2 8.2 63 1160 915 68 117 52 10.9 2.5 ND 2.5
3.5 17.7 6.0 703 10.0 17.3 116 2260 1650 110 179 53 8.5 2.2 1.2 2.2
3.5 11.5 12.0 840 15.0 24.6 130 1190 1240 93 163 54 10 2.2 ND 2.2
23 135 51 291 3.2 3.8 33 735 440 31 61 30 6.3 2.0 ND 2.0
ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND
14 44 34 396 4.6 9.8 52 1073 882 78 ND 69 13.1 3.7 1.2 2.3
23 ND 66 ND ND ND ND ND ND 52 ND ND 9.8 ND ND ND
3.1
U
ND
Y
29
31
29
3.5 28
0.8 26
ND ND
ND 38
ND ND
FeO* Mg
11.9 67.1
11.6 48.3
8.8 31.8
5.9 24.1
12.6 53.2
7.9 9.3
10.7 41.1
13.0 60.8
1.8
1, Olivine melanephelinites; 2, melanephelinites; 3, nephelinites; 4, phonolites; 5, olivine basalts; 6, trachytes; 7, average Olokisalie basalt (Henage 1976); 8, Average Chyulu basalt (Goles 1975). Nos 1-6, average analyses of samples from Kenya, eastern Uganda and northern Tanzania (Beloussov et al. 1974). ND, no data. Kenya and northern Tanzania (King et al. 1972; Le Bas 1977; Figs 1 and 4). They have been divided into those having olivine-poor melanephelinite and carbonatite (Moroto (Mo) Yelele (Ye), Napak (Na), K a d a m (Ka), Elgon (El), Kisingiri, Wasaki and the early phase of Tinderet (Ti)), and those having olivine-rich nephelinites and basanites without exposed carbonatite (Ketumbeine, Moroto, Yelele and the later phase of Tinderet (Le Bas 1977)). O1 Doinyo Lengai (Fig. 1, O1) is unique in having erupted natrocarbonatite ash and lava in historical times (Dawson 1962). Kerimasi is a
similar but inactive volcano. Evidence of natrocarbonatite pyroclastic eruptions has also been found at Tinderet (Deans & Roberts 1984). The Kenya-Tanzania border region is characterized by many recent explosion craters which erupted carbonate tufts rich in xenoliths (Dawson 1964; Dawson & Powell 1969). Mildly-alkaline basanites and alkali basalts are widely distributed as low small-volume shields in the Turkana depression (Fig. 4), and as discrete series of lavas ranging from 16 to 6 Ma in age both on the flanks and in the northern part of the Gregory rift (Elgeyo basalts). They are associated
B. H. Baker
298
0
0
/ .... ,
~
/
/
'"
-/ ~TI ~ - ~ - - _ 2 ~ - - \ - - - ~
(a) (b) FIG. 3. QLM synoptic diagrams of data from the Kenya alkaline province: (a) fields of all analyses and commonly used rock names; (b) trends of series and volcanoes (ME, melilitites; NE, nephelinites; GP, G-type phonolites; AB, alkali basalts; PP, P-type phonolites; TP, trachyphonolites; TB, transitionally-alkalinebasalts; TR, trachytes: RH, alkali rhyolites). with large volumes of P-type phonolites in the Kamasia region, which reach an aggregate thickness of over 2000 m and a volume of about 4 0 0 0 0 k m 3 (Lippard 1973a and b; Williams 1982). Similar suites are found associated with more strongly-alkaline rocks on Kilimanjaro and other mixed association volcanoes (Figs 2 and 4). In the north-central rift basalts of intermediate alkalinity are found as formations intercalated in a dominantly phonolitic succession, and they show a tendency to become less alkaline with
~-,.~ti:-:i-~:!t-: ,38o ]
t t ~A
A
A
:i
,t
/
9
&A o ~"~ 2- ~ f A A A~
AL
Nephetinite-phonotite central volcanoes ~asanite and
phono[[te shields Quaternary basaLts Mixed ass central volcanoes
2~-
Basalt-trachyte trachyte-rhyotite,& trachyphono[ite shields FauLts
II
200 km l
FIG. 4. Map showing the distribution of major eruptive centres of different associations. Alignment of symbols indicates eruption on fissures.
decreasing age (Truckle 1977, 1980; Koyaguchi 1984). Alkali basalts with a different mode of eruption are widely distributed on the eastern rift flank, and form volcanic ranges composed of many monogenetic cinder cones aligned along fissures. On the SE and E flank they form the Quaternary Chyulu (Saggerson 1963; Goles 1975) and Nyambeni ranges (Mason 1955). On the NE flank similar lavas form the Hurri Hills and Marsabit shield (Fig. 1). Differentiated lavas are scarce or absent, and the magmas have risen rapidly from depth with little low-pressure fractionation. The more extensive basalt lava fields in the NE consist of a thin cover of flows of at least two ages. Those of the Kaisut region (Fig. 1, Ks) are slightly older than those farther NE, but they are all probably of Pleistocene age. Little is known of their composition; they appear to be alkali olivine basalts comparable with the Chyulu lavas. These more strongly-alkaline associations are commonly grouped together on variation diagrams to distinguish them from transitionally-alkaline suites associated with over-saturated differentiates (Saggerson 1970; King & Chapman 1972; Williams 1972; Truckle 1980). Transitionally-alkaline rocks consist of mildlyalkaline olivine basalts, ferrobasalts, very scarce mugearites and benmoreites with abundant feldspar phenocrysts, abundant trachytes and generally minor alkali rhyolites. These e ruptives were erupted within the Gregory rift and within a halfgraben in northern Tanzania, and are everywhere less than 7 Ma in age (Webb & Weaver 1975; Baker & Mitchell 1976; Truckle 1980). The ratio of mafic to silicic rocks varies widely.
Petrology of the Kenya rift
299
TABLE 3. Average analyses o f basalts, phonolites and trachytes 2
3
4
5
6
7
8
9
10
H20PzO 5
45.6 3.8 13.9 2.7 11.2 0.3 5.3 9.9 3.5 1.0 0.2 ND 0.9
46.9 2.7 15.5 4.5 7.5 0.2 5.7 10.8 2.5 0.9 1.8 ND 0.7
59.8 0.7 17.6 5.9 3.1 0.3 0.4 1.7 6.1 4.8 1.6 ND 0.2
55.0 0.5 19.9 2.4 2.3 0.3 0.5 1.4 8.4 5.9 3.2 0.7 0.I
58.2 0.7 16.6 4.0 3.1 0.3 0.6 1.5 6.9 5.2 2.0 1.5 0.I
54.5 0.4 21.2 4.0 0.2 0.2 0.3 1.1 9.4 5.9 2.3 1.0
43.8 2.9 14.2 13.3 ND 0.2 7.3 10.9 3.9 1.5 ND ND 0.6
45.6 2.4 15.6 12.6 ND 0.2 6.9 10.4 3.2 1.3 ND ND 0.6
46.1 2.4 16.0 12.8 ND 0.2 7.1 10.7 3.0 0.8 ND ND 0.4
47.2 3.1 15.1 13.9 ND 0.2 5.7 10.2 3.8 1.0 ND ND 0.7
FeO* Mg
13.7 50.0
11.5 46.9
8.4 8.4
4.4 17.3
6.7 13.0
3.8 13.6
11.9 52.1
11.3 52.0
11.5 52.5
12.5 44.6
1.2 4.1 2.1 9.0 10.0 17.4 27.4
1.2 3.4 2.0 7.6 19.9 25.2 19.2
0.9 3.4 2.0 4.7 23.6 28.5 18.0
1.5 4.3 2.2 5.8 27.7 21.6 20.1
13.4 15.4
15.6 5.8
16.6 2.2
12.8 4.0
1
Major elements (wt.~) SiO2 TiOz A1203 Fe203 FeO MnO MgO CaO Na20 K20
H2O+
Molecular norm (FeO/(Fe203 + FeO) adjusted to 0.85 for all samples) Ap I1 Mt Or Ab An Di Hy O1 Ne Ac
2.0 5.6 3.3 5.9 29.8 20.0 20.0 11.6 1.8
1.5 3.9 2.8 5.7 23.7 29.3 17.4 11.5 4.2
0.4 1.0 1.9 28.4 54.5 6.2 0.8
0.1 0.7 0.5 34.7 26.8
0.1 0.9 1.0 31.2 46.8
0.1 0.6 33.8 22.9
5.2
5.6
4.3
6.9 0.1
2.7 27.7 1.5
4.5 8.2 1.6
2.2 33.7 2.3
1, Average basalt, Emuruangogolak caldera volcano (Weaver 1977); 2, average basalt, Silali caldera volcano (McCall & Hornung 1972); 3, average trachyte, Silali caldera (McCall & Hornung 1972); 4, average plateau Ptype phonolite, alkali basalt suite (Lippard 1973a); 5, average Kenya-type phonolite (Lippard 1973a); 6, average Gwasi G-type phonolite, nephelinite suite (Lippard 1973a); 7, average of 21 Miocene alkali basalts, N Kenya rift (Norry et al. 1980); 8, average of 32 Pliocene alkali basalts, N Kenya rift (Norry et al. 1980); 9, average of 12 Pleistocene basalts, N Kenya rift (Norry et al. 1980); 10, average of 10 Quaternary transitional basalts, N Kenya rift (Norry et al. 1980). ND, no data. Some are basaltic shields with or without extensive flat-lying salic lavas; others are cones and shields c o m p o s e d largely of trachytic lavas and ash-flow tufts on w h i c h late-basaltic satellite eruptions took place from fissures. All the transitionally-alkaline series show m a r k e d D a l y gaps b e t w e e n 4 8 ~ and 5 8 ~ SiO2 (e.g. E m u r u a n g o g o l a k ( W e a v e r 1977)), and, in some, mafic lavas are entirely absent. R a r e occurrences of b e n m o r e i t e lavas are of unusual thickness a n d extent, and are associated with fluid peralkaline trachytes that cover m u c h of the floor of the central and southern part of the rift (Baker & Mitchell 1976). The trachytes range from m e t a l u m i n o u s to peralkaline and are associated with c o m e n d i t i c and pantelleritic rhyolites w h i c h show extreme e n r i c h m e n t of incompatible
trace elements. Large volumes of trachytic ashflow tufts were erupted in the N a i v a s h a - N a k u r u central sector of the rift at about 4 M a BP, and probably c a m e from major caldera structures that were buried u n d e r later lavas. Locally, trachytes and comendites form zones of H o l o c e n e plugdomes in the E b u r r u - N a i v a s h a sector of the central rift and are associated with a geothermal area. T h e c h a i n of Q u a t e r n a r y caldera volcanoes along the axis of the deepest part of the G r e g o r y rift characterizes b a s a l t - t r a c h y t e b i m o d a l volc a n i s m (Scott 1980; Williams et al. 1984). One r e p r e s e n t a t i v e - - M e n e n g a i - - h a s p r o v e d to contain compositionally stratified ash-flow tufts (Leat et al. 1984), and it is likely that m a n y others will be found. Several central volcanoes, including some of
B. H. Baker
300
TABLE 4. Quaternary transitionally-alkaline rocks 1
2
3
4
5
6
7
8
47.5 2.6 14.7 13.9 0.2 5.9 11.6 2.9 1.0 0.5
58.5 1.5 15.9 7.1 0.2 2.2 4.7 5.2 3.9 0.4
61.8 1.0 14.2 7.1 0.3 0.5 1.8 6.2 5.2 0.2
64.8 0.8 13.3 7.3 0.3 0.4 1.2 5.9 5.2 0.1
67.0 0.7 13.7 5.4 0.2 0.4 0.9 5.6 5.2 0.1
60.5 0.7 14.7 8.6 0.4 0.5 1.3 6.9 5.1 0.1
57.1 0.9 16.1 9.0 0.4 0.9 2.6 7.6 5.0 0.2
(ppm) 37 86 45 116 3.0 1.5 2.5 14 453 390 22 20 44 5.4 2.0 0.8 2.3 0.4
33 69 46 140 3.5 1.8 3.1 20 447 472 33 31 39 7.0 2.2 0.9 2.8 0.5
14 ND 17 336 8.9 4.9 9.5 67 282 950 82 126 61 11.2 3.2 1.2 4.3 0.8
7 ND 3 676 15.4 9.6 17.4 116 48 347 135 187 91 16.8 3.3 2.3 8.2 1.4
4 ND 1 1144 28.0 18.4 32.1 159 13 167 185 307 115 21.6 2.6 2.9 10.0 1.8
4 ND 2 1286 36.6 18.7 41.1 181 37 362 203 302 116 24.1 2.7 3.2 12.5 2.0
3 ND 1 1043 26.8 24.5 34.3 166 14 148 218 318 123 24.3 3.5 3.7 ND 2.4
3 ND 3 961 10.2 23.7 35.4 170 48 315 190 293 101 20.6 3.2 2.9 ND 1.9
U Y
ND ND
0.8 32
ND 49
3.3 105
7.5 107
9.8 ND
6.1 131
6.9 111
FeO* Mg
11.4 50.1
12.5 45.8
6.4 38.5
6.4 13.2
6.5 10.5
4.9 12.5
7.7 10.4
8.1 16.7
Major elements (wt.~) SiO2 47.6 TiO2 2.0 A1203 14.8 Fe203 12.7 MnO 0.2 MgO 6.4 CaO 11.5 Na20 2.7 KzO 0.8 P2Os 0.3 Trace elements Sc Ni Co Zr Hf Ta Th Rb Sr Ba La Ce Nd Sm Eu Tb Yb Lu
1, Average of 31 transitionally alkaline Quaternary basalts from the southern rift (Baker, unpublished data); 2, average of 10 O1 Tepesi basalts (Baker et al. 1977); 3, average of 3 benmoreites (Baker et al. 1977); 4, average of 40 trachytes (Plateau and Magadi trachytes) (Baker et al. 1977); 5, average of 6 alkali rhyolites (Baker, unpublished data); 6, average of 39 trachyrhyolites and alkali rhyolites (Limuru trachytes) (Baker, unpublished data); 7, average of 15 trachytes, Suswa pre-caldera shield lavas (Baker, unpublished data); 8, average of 31 trachyphonolites, Suswa post-caldera lavas (Baker, unpublished data). ND, no data. the very high cones such as K i l i m a n j a r o and M o u n t K e n y a , contain at least two of the preceding suites, and are referred to as m i x e d central volcanoes (Fig. 4). K i l i m a n j a r o (Williams 1969b) and Meru (Beloussov et al. 1974) contain nephelinites and alkali basalts and their differentiates, whereas M o u n t K e n y a (Baker 1967), Olorgesailie ( H e n a g e 1976) and the Sattima and Kipipiri centres of the A b e r d a r e R a n g e consist of mildly-alkaline and transitional suites (Shackleton 1945). In the K a v i r o n d o rift and the U g a n d a borderland, M o r o t o (Varne 1968) and T i n d e r e t contain both strongly- and mildly-alkaline suites.
These mixed volcanoes tend to be large steepsided central volcanoes located on the rift flanks, but Olorgesailie is a well-studied example that developed within the Gregory rift ( H e n a g e 1976). Differentiated rocks are subordinate on such volcanoes, but M o u n t K e n y a is an exception; its m a i n cone is constructed of P-type phonolites, succeeded by flank eruptions of alkali basalt a n d trachyphonolite, and later by transitional basalts and trachytes of the I t h a n g u n i centre (Baker 1967). These volcanoes testify to the c o n t e m p o r aneous availability of mafic m a g m a s of a wide range of alkalinity.
P e t r o l o g y o f the K e n y a rift
Petrology and geochemistry Nephelinite--carbonatite suite The compositional fields of several series of the more alkaline rocks is shown on an R1-R2 variation diagram (De la Roche 1980) in Fig. 5. The most strongly alkaline of these is the nephelinite-carbonatite suite (Fig. 5, NE). The suite ranges from melilitites, melanephelinites and ankaratrites to phonolites, and is associated with intrusive complexes of ijolite, nepheline syenite and carbonatite. Nyamaji (NY) in the Kavirondo rift is slightly less alkaline and contains a greater proportion of phonolites (Le Bas 1977). Mount Elgon (EG) on the Uganda border and the Kishalduga shield (KD) on the western side of the southern rift also display trends intermediate between melanephelinite and basanite. Moroto contains distinct nephelinitic and alkali basalt series (Fig. 5, MOa and MOb) (Varne 1968). Representative average analyses of some of these series are given in Table 1 and Table 2, analyses 1-4. The mafic rocks are characterized by nepheline and pyroxene phenocrysts, with or without olivine and/or melilite. Perovskite is present in some lavas, and the phonolites are characterized by accessory sphene, high Sr and Ba, and declining
301
light REE with differentiation. Calcic plagioclase is generally absent in the mode and norm of this suite (Table 1), which accounts for the high Sr and Ba contents of the phonolites. Notable features of the strongly-alkaline suite are the unbroken range of compositions and absence of a Daly gap, subordinate volumes of differentiated rocks, abundance of pyroclastic deposits and a tendency to form large central volcanic piles outside the graben sector of the rift. Le Bas (1977) proposed that the parental magma for the nephelinitic members of the suite is carbonated melanephelinite (Table 1, analysis 6), which gave rise to a carbonatite fraction by liquid immiscibility. The silicate fraction evolved to nephelinite and G-phonolite by fractionation of olivine, pyroxene and sphene. The high contents of incompatible elements are ascribed to a combination of low degrees of partial melting, high pressures and high volatile contents in the source mantle. The phonolitic end-members are distinguished by high Sr and Ba contents owing to the lack of plagioclase fractionation, and are referred to as Gwasi or G-phonolites (Lippard 1973a). Such phonolites are present in subordinate proportions, in contrast with the P-phonolites of the alkali basalt suite (see below). The characteristics of the nephelinite-carbonatite suite are furtherdescribed by Le Bas (1987).
R2
Basanitic and alkali basaltic suite
r --~,
/
,, /
f/~
1--~
NE~-~mel f MO.a I
I
. o/./I.//.
Ii
/ ~:~~
.-2
/
.
" /
. i " .- ? ~ ~ 89
..?A~
(/-. I
-1000
0
I
1000
!
R1
2000
FIG. 5. R1-R2 variation diagram (De La Roche 1980) showing the fields of some strongly-alkaline series : NE, average analyses of the nephelinite-carbonatite suite of the Kavirondo rift (Le Bas 1977) (mel, melilitites; mne, melanephelinites; ij, ijolites; ne, nephelinites; ph, phonolites); NY, Nyamaji suite (Le Bas 1977); MO, Moroto volcano (Varne 1968)(a, nephelinitic series; b, alkali basalt-phonolite series); EG, Elgon volcanics; KD, Kishalduga formation (Crossley & Knight 1981); ab, tb, fields of alkali and transitional basalts; *, points for average compositions given in Table 1.
Basanites and alkali basalts with normative ne greater than 5% are spatially associated with tephrites, ne-mugearites, and exceptionally large volumes of P-phonolite within the Turkana depression and on the flanks and within the northern part of the Gregory rift (Fig. 4). The older mafic series have not been adequately dated but range from 30 Ma in northern Turkana, where they are locally tholeiitic (Bellieni et al. 1981), to 7 Ma, spanning a time range from long before rifting began to the time of the first largescale rift faulting (Baker et al. 1971 ; Williams & Truckle 1980). The mafic lavas frequently carry olivine and augite phenocrysts; plagioclase is less common, and nepheline occurs in the groundmass or is replaced by sodalite. The average Miocene alkali basalt composition is shown in Fig. 6 (Mb), together with the fields of the Miocene Noroyan and Saimo basalts (NO + SO) and associated Pphonolites (Pp). The late Miocene Kaparaina basalts (KA) are less alkaline, and the fields of transitionally-alkaline lavas are shown for comparison (TB, Tr, Tp and Rh). Phonolite lavas cap large areas of the rift
302
B. H. Baker
R2
~
."--.TB~v/
Ch
2000~ .-~ . NO+' " ~ ~ Mo
"9
j,/.y
~#JKn
1000
,~%J-:J ~ I -1000
"
"--"
1000 I
2000 I
R1
FIG. 6. R 1 - R 2 variation diagram of Miocene alkali basalts and phonolites, and of the post-Miocene transitionally-alkalinesuite : NO + SO, Noroyan and Saimo basalts (15 Ma); KA, Kaparaina basalts (5 Ma); Pp, Miocene P-type phonolites; Mb, average Miocene alkali basalt; Pb, average Pliocene alkali basalt; Qb, average Quaternary transitional basalt (Norry et al. 1980) (see Table 3). Average Quaternary basalt of the southern rift (Table 4): TB, Tr, Tp and Rh, fields of transitional basalts, trachyte, trachyphonolites and alkali rhyolites from the southern rift (Baker et al. 1977, and unpublished data) (see Table 4); Kn, late Oligocene tholeiitic and transitional basalts of northern Turkana (Bellieni et al. 1981).
Such occurrences suggest that the basalts and phonolites are genetically related, and imply contrasts in eruptive mechanisms between the rift-floor and rift-flank environments. Other occurrences of series of intermediate alkalinity are in the form of chains of fissurealigned multicentre monogenetic cones and lavas comprising the Chyulu (Fig. 7, CH; Table 2, analysis 8), Nyambeni and Hurri ranges on the eastern rift flank, and large expanses of lavas with maars and small cones in NE Kenya. These fissure-fed lavas show little compositional variation and have only insignificant volumes of differentiated lavas (Saggerson 1963, 1968; Goles 1975). Somewhat more alkaline Miocene lavas in the southern rift are much less voluminous, and consist of localized small limburgite and nephelinite shields (Fig. 5, KD) which are overlain by phonolitic lavas (Crossley 1979; Crossley & Knight 1981).
R2 KI
CH
~SI
2000
NG
~,~
KO
shoulders (Fig. 1), and a greatly expanded succession (up to 2.5 km thick) is found within the north-central part of the Gregory rift, ranging from 16 to 9 Ma in age (Lippard 1973b; Williams & Truckle 1980; Williams e t al. 1983). Phonolite flows up to 270 m thick reach volumes of 300 km 3, and the Yatta phonolite can be traced down the east flank of the rift over a distance of 290 km (Fig. 1, Ya). The phonolites contain alkali feldspar and nepheline phenocrysts, with ferroaugite, apatite and biotite microphenocrysts, and show little variation. They have a similar bulk composition (Table 3, analysis 4) to phonolites of the nephelinite-carbonatite suite (Table 3, analysis 6), but have less Ba and Sr, and light REE abundances which increase with differentiation (Lippard 1973a). Intermediate rocks are scarce among the more voluminous representatives of the suite within the rift, but associations which show gradations from alkali basalt to tephrite, rhomb porphyry and phonolite occur on Mount Kenya and Kilimanjaro on the eastern rift flank, and are described separately below as mixed associations.
...... .... ~ I -1000
KEA .
_ 0
.
0
t
L
.
"
. . . . 1000
2000
. . . .
R1
FIG. 7. R l-R2 variation diagram of Quaternary mixed association volcanoes, Chyulu basalts E of the rift and transitionally-alkalineseries in the southern Kenya rift valley: KI, Kilimanjaro (Williams 1969b); KE, Mount Kenya (Baker 1967, and unpublished data); CH, alkali basalts of the Chyulu range (Goles 1975) (Table 2, analysis 8); SI, NG and KO, Singaraini, Kirikiti and O1 Tepesi transitionally-alkalinebasalts of the southern rift (Baker et al. 1977, and unpublished data) (Table 4, analysis 1); PT, Plateau and Magadi trachytes (Baker et al. 1977; Crossley and Knight 1981) (Table 4, analysis 4); LT, Limuru trachytes (Baker, unpublished data) (Table 4, analysis 6); ME, Menengai (Leat et al. 1984); SU, Suswa volcanics (a, pre-caldera trachytes; b, post-caldera trachyphonolites (Nash et al. 1969; Baker, unpublished data) (Table 4, analyses 7 and 8); HA, Hannington trachyphonolites (Griffiths & Gibson 1980).
P e t r o l o g y o f the K e n y a rift
Transitionally-alkaline series As already noted, the great bulk of transitionallyalkaline basalts, and of the trachytes, trachyphonolites and alkali rhyolites with which they are associated, are less than 7 Ma in age and were erupted within a developing graben. Such rocks are found in strongly bimodal shield volcanoes (Webb & Weaver 1975) as expanses of flood lavas (Baker & Mitchell 1976) and as a chain of Quaternary caldera volcanoes along the axis of the Gregory rift (Macdonald et al. 1970; Leat et al. 1984; Williams et al. 1984). The compositional fields of some of these series are presented in Fig. 7, which shows the pronounced Daly gap between the basalts (SI, NG and KO) and the salic lavas (LT, PT, HA, SU and ME). The trend straddles the critical line of silica saturation, but the salic rocks trend to either the oversaturated (LT) or the undersaturated (SU and HA) side, or extend on both sides (PT and ME). This raises several problems of the transitionally-alkaline suites: the significance of the Daly gap, the occurrence of apparently related rocks of contrasted silica saturation on the same volcano and the occurrence of series that consist largely or entirely of salic rocks. The transitionally-alkaline basalts range from olivine-augite basalts to ferrobasalts with titanomagnetite microphenocrysts (Table 3, analyses 1, 2, 9 and 10; Table 4, analyses 1 and 2). Intermediate lavas are exceedingly rare and include thick benmoreite flows with abundant plagioclase phenocrysts (Baker et al. 1977) (Table 4, analysis 3). Trachyte lavas are well represented and in many series range from mafic trachytes to peralkaline rhyolites (Fig. 7, PT, LT; Table 4, analyses 4-6). Suswa caldera (SU) and the Hannington lavas (HA) exhibit a compositional range from trachyte to trachyphonolite (Nash et al. 1969; Griffiths & Gibson 1980) (Fig. 7; Table 4, analyses 7 and 8). Among the other Quaternary trachytic caldera volcanoes of the axial part of the Gregory graben all but two have pre-caldera shields composed of trachytic and basaltic lavas, overlain by postcaldera trachytic lavas and ash-flow tufts. On several of these volcanoes post-caldera eruptions include basalts erupted from fissures (Williams et al. 1984), testifying to the coexistence of marie and salic magmas throughout much of their eruptive life. The salic rocks on these volcanoes range from metaluminous to peralkaline (Macdonald 1974), and Emuruangogolak (Weaver 1977) shows contrasts in trace-element geochemistry between eruptive phases. Suswa built a pre-caldera shield of trachytic lavas, but post-caldera lavas are
30 3
phonolitic and less differentiated (Fig. 7; Table 4, analyses 7 and 8). A detailed study of Menengai has shown that two major trachytic ash-flow tufts were erupted from a compositionally stratified magma chamber, and the petrogenesis of its trachytic eruptives has been ascribed to a combination of fractional crystallization, magma mixing and unspecified processes of 'liquid-state' differentiation (Leat et al. 1984). The peralkaline salic rocks tend to expel sodium-rich fluid on crystallization, with the result that only obsidians provide true magma compositions. It has been shown that most trace elements are not greatly affected by these processes (Baker & Henage 1977); nevertheless, this effect creates difficulties in mass-balance modelling of differentiation processes.
Mixed volcanic associations Several Quaternary central volcanoes contain rocks of a range of alkalinities; the compositional fields of the Mount Kenya suite (Baker 1967, and unpublished data) and Kilimanjaro (Williams 1978) are shown in Fig. 7 (KI and KE). Both volcanoes were built on the eastern rift flank, and their compositions contrast markedly with those of contemporaneous Quaternary lavas erupted within the rift graben. Mount Kenya consists of a large pile of Pphonolite and rhomb porphyry lavas, with latestage fissure-fed alkali basalts and trachyphonolites erupted on its flanks. Subsequent parasitic volcanism erupted mixed comendite-basalt ashflows and built a trachytic lava cone (Baker 1967; Rock 1976a). The main cone of the volcano contains the alkali basalt-P-phonolite composition range, and satellite eruptions were of basalt, trachyte and rhyolite (not shown in Fig. 7, but equal to the NG and PT fields). Several alkali rhyolite dykes contain xenoliths of quenched basalt representing partial mixing of the two magmas. Kilimanjaro is composed of rocks of a wide range of alkalinity, such as nephelinites, ankaratrites, alkali basalts and phonolites (Fig. 7, KI) (Williams 1969b). Less-well-known volcanoes that also contain mixed associations are Meru, Tinderet and Sattima, and all are large central cones situated outside the Gregory rift (Williams 1969a). One central volcano which contains the whole alkaline spectrum from nephelinites to hypersthene normative basalts is Olorgesailie (Henage 1976), which is a 3.2 Ma old volcano on the floor of the southern Gregory rift. The eruptive sequence shows four stages of contrasting alkalinity: 1, alkali basalt, mugearite and benmoreite;
304
B. H. Baker
2, transitional basalt (hypersthene normative) and mugearite; 3, weakly-alkaline benmoreite, trachyte and trachyphonolite; 4, mixed alkali basalt, nephelinite and G-phonolite. This suggests episodic changes in the nature of the magma being supplied to the volcano. Differentiation within each of the stages can be modelled by fractionation of phenocrysts (Henage 1976). Mixed association volcanoes are of a wide age range and tend to build large cones, and most are situated on the rift flanks. The occurrence of some within the rift, but probably pre-dating its formation, suggests that tapping of magmas of contrasted alkalinities was favoured around the margins of the rift and on its flanks.
Trace-element characterization of volcanic suites and series Few systematic studies of the trace-element geochemistry of volcanic series have been performed. Many of the existing data have been used to address two questions. First, are the mafic and salic members of bimodal transitionallyalkaline series co-magmatic? Second, how were the salic rocks generated (Sceal & Weaver 1971 ; Weaver et al. 1971 ; Weaver 1977)? Contrasts between suites of differing alkalinity are shown by the differing behaviour of Rb, Sr and Ba (Fig. 8). The alkali basalt-phonolite trend (AB-PHO) has a small depletion of Sr and Ba by comparison with the transitionally-alkaline trend
Rb
Sr
Ba
FIG. 8. Ternary variation diagram showing contrasts in the trends of Rb enrichment and the degree of Sr and Ba depletion in rocks of the alkali basaltphonolite ( ; AB, PHO) and transitional basalttrachyte-rhyolite suite ( - - - ; TB, TRA, RHY). The range of the Laacher See phonolitic tephra (LZ) is shown for comparison (W6rner et al. 1983).
Hf
la FIG. 9. Ternary variation diagram showing differences in the Ta/Hf ratio for selected alkaline suites: OS, Olokisalie (Henage 1976); PT, O1 Tepesi basalts and Plateau trachytes (Baker et al. 1977); L, Limuru trachytes and alkali rhyolites (Baker, unpublished data); K, alkali basalts and P-phonolites of Mount Kenya (Baker, unpublished data). A selection of ranges of compositionallyzoned calc-alkaline magmas (MZ, KZ, T, BZ) is shown for comparison (Baker & McBirney 1985) together with the range of the Laacher See tephra (LZ) (W6rner et al. 1983).
in which Sr and Ba were greatly depleted by fractionation of plagioclase and alkali feldspar ( T B - T R A - R H Y ; see Table 4, analyses 2-5). The incompatible trace elements in the alkaline series (Zr, Hf, Nb, Ta, Th, U, Rb and the light REE) are greatly enriched in salic rocks and exhibit linear covariance. This has been ascribed to their having low bulk distribution coefficients owing to the absence of accessory phases that could deplete such elements (Sceal & Weaver 1971; Weaver et al. 1972; Barberi et al. 1975; Weaver 1977; Baker et al. 1977, 1978). Zr/Nb ratios for each volcanic series tend to be constant and have distinctive values. In bimodal series this has been used to argue that the mafic and silicic members are co-magmatic and related by crystal-liquid fractionation (Weaver et al. 1972), but this conclusion is permissive rather than required because the contents of these elements in basaltic lavas have not been determined with sufficient precision. Nevertheless it is necessary to try to demonstrate a co-magmatic relationship in a series of rocks before petrogenetic modelling is performed. The characteristics outlined above extend to Ta, Hf, Th and the light REE except Eu (Baker et al. 1977) and are similar to those of comparable oceanic series (Baker et aL 1978). The proportions of Hf, Ta and Th in several Kenyan suites vary little (Fig. 9, PT, O, L and K); Th is the most
Petrology of the Kenya rift incompatible element owing to its combination of high charge and large ionic radius. The suites show distinctive Hf/Ta ratios, suggesting comagmatic relationships over a wide range of compositions. The comparable behaviour of a selection of calc-alkaline compositionally stratified magma bodies suggests that the petrogenetic processes were similar (Baker & McBirney 1985) (Fig. 9, MZ, KZ, T, L and BZ). In contrast the zoned Laacher See phonolitic tephra shows Ta depletion owing to the crystallization of sphene (W6rner e t al. 1983; Wolff & Storey 1984) 9 A study of the Emuruangogolak Quaternary caldera volcano showed that trachytes erupted between phases of explosive activity and caldera collapse have distinctive Zr/Nb ratios which are thought to be the effects of volatile degassing (Weaver 1977). Much more complex contrasts in trace-element variance were found in the eruptive phases of Menengai caldera volcano (Leat e t al. 1984) and were ascribed to a variety of processes including fractionation, magma mixing and 'liquid-state' processes. Among the more strongly-alkaline phonolites the linear covariance of incompatible trace elements tends to break down owing to the presence of minor phases that incorporate such elements. Lippard (1973a) has shown that the light REE are depleted in Gwasi-type phonolites whereas they are enriched in P-type phonolites and (Kenya-type) trachyphonolites. Similar effects, including depletion of Nb and Ta, have been noted in compositionally stratified phonolitic tufts and ascribed to fractionation of one or another of sphene, apatite, allanite, perovskite or
z z m
" J"
~ "'75-';-/"
~#"J t
# I /
/ [
0 5 miles O ~ - ~ k m
j /
FIG. 1. The northern part of the Chilwa alkaline igneous province (mainly after Bloomfield 1965, Plate XII).
Petrochemistry of the Chilwa alkaline province erate-choked vents occur (Bloomfield 1965, p. 130; Garson 1965b, p. 91)and areas of metasomatism and alteration of the nepheline syenites are present. Xenoliths of metamorphosed basic volcanic rocks occur in all four complexes and there are a few in-faulted masses up to a kilometre across in Chikala and Chaone. Junguni is an isolated intrusion of nepheline syenite lying to the N of Mongolowe (Bloomfield 1965, p. 111). Contacts with the basement are obscured by Recent sediments, but are exposed in some small hills NE of the main intrusion where the country rocks show the effects of fenitization. Chilwa Island is a multiple carbonatite complex and was the first carbonatite to be recognized in Africa (Dixey et al. 1937). It has been described in detail by Garson & Smith (1958). The present account is not directly concerned with the carbonatites, but rather with the associated plugs and dykes which consist of alkaline silicate rocks including nephelinite, nepheline syenite and aln6ite. The Kangankunde carbonatite (Garson 1965a) lies to the W of the area shown on Fig. 1, but does not have associated silicate rocks although some dykes could be carbonatized nephelinite or aln6ite (Garson 1965a, pp. 45-8). Dykes of microgranite, s61vsbergite, microsyenite and phonolite are widespread (Bloomfield 1965, p. 133 and Plate XII); the s61vsbergites represent the most north-easterly part of a swarm 100 km long extending from Salambidwe on the Mozambique border. Phonolite dykes form a swarm W of Chinduzi (Fig. 1) and are found in the vicinity of, and cutting, all the nepheline syenite complexes.
Petrography Granite, quartz syenite and syenite
337
from the main complex over much of its length by a screen of basement gneiss (Fig. 1). Granites also occur on the western and northern margins but here appear to grade inwards into quartz syenite and syenite. Malosa contains a more heterogeneous suite of syenitic and granitic rocks than Zomba. The syenite of the central plug of Zomba is a coarse rock of perthite, sometimes containing a little quartz, hedenbergitic pyroxene, often zoned to green rims and sometimes mantled by amphibole, variably altered fayalite, a little reddish biotite, opaque oxides and accessory apatite, zircon and pyrite. Malosa rocks are similar but are free from fayalite, while, particularly in the N close to the contact with the nepheline syenites, a variety containing aegirine and arfvedsonite occurs. The rocks of the inner syenitic ring of Zomba consist of variably perthitic feldspar phenocrysts in a groundmass of alkali feldspar, a little quartz, hedenbergitic pyroxene, usually rimmed by amphibole which also forms independent poikilitic crystals, a little altered fayalite and accessories. The granites are generally coarse rocks consisting of tabular perthites, with interstitial quartz, aegirine-(aegirine-augite), riebeckitic and arfvedsonitic amphibole, fayalite or its alteration products and accessories including biotite and pyrochlore. Dyke rocks include aegirine-aenigmatite microgranite and alkali feldspar-quartz porphyries with a felsitic groundmass. Mpyupyu
The rocks of Mpyupyu Hill are hornblendebiotite syenites, sometimes with a little quartz. Perthite forms large very irregular plates. Brown hornblende contains rare pyroxene cores, while biotite mantles amphibole. Pyrochlore-bearing aegirine-riebeckite syenite outcrops just S of the Hill (Garson 1960, p. 32).
Zomba-Malosa
The field relationships and petrography of the Zomba-Malosa complex have been described in some detail by Bloomfield (1965, pp. 96-107). The following brief petrographic account is based on rocks collected by Woolley together with sections made from specimens originally collected by Bloomfield. Bloomfield (1965, p. 96) divided the Zomba part of the complex into a central syenitic plug, an inner syenitic ring and an outer granite ring. The outer ring is a welldefined peripheral ring-dyke some 100 m wide which can be followed for about 25 km along the S and E boundaries of the complex; it is separated
Saturated and oversaturated dykes A regional swarm of trachyte and quartz trachyte dykes, generally referred to in the Malawi geological literature as s61vsbergites, extends approximately from the vicinity of ZombaMalosa to the Salambidwe complex to the SW. They were described by Bloomfield (1965, p. 138) and in references cited therein. They are almost invariably slightly altered with turbid subhedral short prismatic alkali feldspars, often aligned; interstitial quartz varies modally from zero to 15%, and ubiquitous opaque oxides pseudomorph
338
A. R. Woolley & G. C. Jones
mafic minerals. Bloomfield (1965, p. 139) identified riebeckite and aegirine in a few specimens. Some varieties are rich in brown biotite.
Nepheline syenite and syenite Chikala, Chaone, Mongolowe and Chinduzi Chikala, which was mapped by Garson (1960) and Stillman & Cox (1960), Chaone and Mongolowe, which were mapped by Vail & Monkman (1960) and Vail & Mallick (1965), and Chinduzi, which was mapped by Bloomfield (1965), all consist of rocks described as perthosite, pulaskite and foyaite or microfoyaite. There was some slight variation in the definitions of these rock types as used by the above workers but essentially the perthosites are rocks rich in alkali feldspar (perthite) with up to 5% quartz, the pulaskites contain up to 5}/o nepheline and 5%-50~o mafic minerals, and the foyaites contain 5%-50% feldspathoids. The present study has confirmed this range of rock types and indicated a continuous transition from rocks with a little quartz to nepheline syenites with more than 30% nepheline. On Chaone and Mongolowe the field relationships suggest that the perthosites were emplaced first, followed by the pulaskites and then foyaites. Perthosite is the least abundant rock type, occurring principally along the eastern margin of Chikala and the southern margins of Chaone and Mongolowe. It is a coarse rock of large plates of perthite, sometimes a little quartz, a greenish amphibole with more rarely an arfvedsonitic type, often with cores of aegirine-augite, and a little brown biotite. The pulaskites and foyaites vary from coarse rocks with feldspar plates over 1 cm in diameter and patches of mafic minerals of a similar size to finer-grained more homogeneous varieties. The feldspar is invariably perthitic displaying a wide range of complex exsolution and replacement textures. Nepheline is usually fresh with only minor alteration to cancrinite. The nepheline may be interstitial or form short stout subhedral prisms and rather rounded patches enclosed in feldspar. Interstitial sodalite occurs in some of the foyaites, and in Chikala may develop as approximately circular patches, poikilitically including myriads of small feldspars. In Chikala, Chaone and Mongolowe amphibole, usually katophorite (Woolley & Platt 1986), is the principal mafic phase, but sodic pyroxenes and brown biotite are also abundant. Most of the rocks of Chinduzi are aegirinebiotite-nepheline syenites, but amphibole occurs in about one-third of the samples studied. Arfvedsonitic and riebeckitic amphiboles are
found in western Mongolowe, just N of Malosa, but elsewhere the amphibole is a katophorite. Although mica, pyroxene and amphibole are generally clearly primary crystallizing phases, textures commonly indicate some secondary recrystallization, invariably with mica forming at the expense of pyroxene. An opaque oxide phase is ubiquitous, sphene and apatite are abundant, and a little zircon is sometimes present.
Junguni These nepheline syenites contain the most feldspathoid of all the rocks of the area. Perthitic alkali feldspar forms subhedral prisms amongst which nepheline may be interstitial or form stout fresh prisms. Aegirine is abundant, biotite occurs in moderate amounts and amphibole is rare. Sodalite, referred to by Bloomfield (1965, p. 122) as analcite, forms rounded masses, interstitial areas and, in a few rocks, rectangular areas within feldspar prisms which are clearly secondary after nepheline. Sodalite forms between 80 and 90 vol.% of two specimens. Accessories include opaque oxides, sphene, carbonate, apatite and pyrite. Pegmatitic nepheline syenites on Kadongosi Hill, just to the NE of Junguni, include large blades of alkali amphibole, fluorite and several unidentified accessories; Bloomfield (1965, p. 122) described eudialyte from these rocks.
Chilwa Island Nepheline syenite forms a small plug about 100 m in diameter and several sheets cutting s6vite on the NW side of the Chilwa Island carbonatite plug, as well as several microfoyaite dykes in fenitized basement rocks. The nepheline syenites are relatively homogeneous medium-grained rocks comprising alkali feldspar, nepheline, aegirine-(aegirine-augite), a little brown biotite and accessories. The microfoyaites are similar except for the porphyritic nature of the nepheline and the presence of melanite at one locality.
Phonolite dykes The phonolites are texturally and mineralogically rather variable. Alkali feldspar, nepheline, aegirine, alkali amphibole and biotite may occur as phenocrysts. The groundmass is generally a felted mass of prismatic alkali feldspar and nepheline, usually with numerous tiny aegirine needles, while biotite is widespread. Zeolite, probably natrolite, is prominent in some samples and possible sodalite may occur.
Petrochemistry of the Chilwa alkaline province Ijolite, nephelinite, camptonite and alnOite These rock types form dykes on Chilwa Island, usually within fenitized basement rocks but also cutting carbonatite. Ijolite forms one dyke-like mass in K-rich fenites and is characteristically extremely heterogeneous. It consists principally of nepheline, aegirine-augite, melanite and wollastonite, the last sometimes as stellate clusters several centimetres in diameter. K-feldspar is present in some samples and is probably the result of late feldspathization. The nephelinites consist of phenocrysts of olivine and pyroxene in a dense groundmass of pinkish augite prisms, abundant opaque oxides, a little biotite and turbid interstitial material probably composed of zeolite, sericite and possibly alkali feldspar. Only one specimen containing fresh nepheline has been found. Rocks described as camptonites by Garson & Smith (1958, p. 79) occur as concentric dykes in the basement and as a plug within the s6vite. The rock of the plug comprises phenocrysts of olivine, augite and a red-brown amphibole in a groundmass of pyroxene, amphibole, opaque oxides, sodic plagioclase and an unidentified turbid material. A sample from a dyke contains scarce phenocrysts ofaugite and olivine in a groundmass of pyroxene, olivine, reddish-brown amphibole, an opaque phase, probable feldspar and zeolite. Aln6ite, from a dyke 300 m long cutting s6vite, consists of 50% carbonate with numerous phenocrysts of biotite, reddish amphibole, pyroxene, apatite and an opaque phase.
Metavolcanic rocks A collection of metamorphosed basic lavas was made from outcrops on the NW margin of Chaone, which is faulted against basement gneisses but cut by nepheline syenite. These are compact medium-grained melanocratic rocks which are sometimes vaguely foliated and may be cut by a few white veinlets. The dominant minerals are a pale-green pyroxene and a hornblendic amphibole, either of which may be concentrated in monomineralic layers. An opaque phase is abundant and biotite is found in some samples. Nepheline is the dominant felsic mineral with alkali feldspar present in more mesocratic rocks; both may form narrow veinlets.
Petrochemistry 196 rock samples have been analysed for this study. There are 14 published analyses for the area, some of which include a few trace-element
339
data, but these were determined spectrographically. Approximately two-thirds of the newly analysed samples were collected by Woolley, and about 50 were acquired from the collection of the Geological Survey of Malawi. Most of the latter were originally collected by K. Bloomfield over Zomba and Malosa and from widely distributed phonolite dykes, and a few by M. S. Garson in the eastern part of Chikala. Four of the Chikala samples analysed come from the University of Leeds and were originally collected by C. J. Stillman and K. G. Cox. Twelve of the analysed specimens are from the collection of the British Museum (Natural History), the majority of these having been collected by Garson during his mapping of Chilwa Island. Although some of the specimens acquired from these sources were rather small, the material collected specifically for this study consisted of large pieces, generally several kilograms in weight. This material, after initial reduction by screw press and jaw crusher, was subdivided using a sample splitter. A portion was further reduced in a tungsten carbide disc mill to pass 100 gm mesh for analysis. The rocks were analysed as follows. Si, Ti, A1, Fe, Mn, Mg, Ca, Na, K, P, Nb, Zr, Y and La were determined by X-ray fluorescence (XRF) using lithium metaborate discs, and C1 was determined using pressed powder discs. Calibration was with international standard rocks; a few of the samples were analysed gravimetrically and used as standards for the remainder. Li, Be, V, Cr, Ni, Cu, Zn, Rb, Sr, Cs, Ba and Pb were determined by atomic absorption spectroscopy (AAS) in HF-H3BO3 solutions; calibration in this case was performed with synthetic mixures checked against standard rocks. H20 + and CO2 were determined by means of a C H N elemental analyser, FeO was determined by titration after dissolution in HF-H2SO4 and F - was determined by pyrohydrolysis with colorimetric or ionselective electrode finish. A selection of analyses is presented in Tables 1-4, but copies of all the analyses are obtainable from the authors and a full data set has been deposited in the Mineralogy Library of the British Museum (Natural History).
Granite, quartz syenite and syenite Analyses of 43 rock samples from Zomba-Malosa and five from Mpyupyu have been made, and a further six are available in the literature (Dixey et al. 1937; Bloomfield 1965), although all the latter are from the southern end of Zomba Mountain. A selection of new analyses and their normative compositions is given in Table 1. Of
340
A. R. Woolley & G. C. Jones
TABLE 1. A n a l y s e s o f granites, quartz syenites a n d syenites f r o m the Z o m b a - M a l o s a c o m p l e x a n d
M p y u p y u , a n d o f a sblvsbergite d y k e 1
2
10
11
12
13
69.6 0.45 13.0 3.39 1.24 0.08 0.19 0.63 5.4 4.82 0.46 0.32 0.03 0.09 0.72
56.9 1.67 15.1 2.22 7.50 0.25 1.90 4.58 4.6 3.95 0.54 0.06 0.65 0.15 0.44
61.6 0.77 15.8 2.34 4.62 0.25 0.42 1.25 6.1 5.83 0.36 0.08 0.07 0.06 0.36
62.1 0.51 18.4 0.80 2.75 0.11 0.51 2.44 5.9 5.18 0.24 0.11 0.17 0.07 0.43
99.98 100.61 100.46 100.06 100.40 100.02 100.05 100.42 100.51
99.91
99.72 100.28
3
4
5
6
7
8
9
14
Major elements (wt. ~ ) SiO 2 TiO 2 AI203 Fe203 FeO MnO MgO CaO Na20 K.,O H20* H_,OPzO5 CO, Others Total
70.6 0.29 13.7 1.61 2.23 0.13 0.18 1.06 4.4 5.21 0.33 0.10 0.04 0.10 0.52
72.0 0.25 13.2 1.47 1.79 0.09 0.06 0.45 5.0 4.88 0.31 0.14 0.02 0.13 0.42
100.50 100.21
73.3 0.22 13.3 1.02 1.61 0.09 -0.55 4.7 4.50 0.24 0.08 0.04 0.04 0.29
72.0 0.28 10.4 6.39 1.93 0.15 0.10 0.48 3.9 4.33 0.25 0.03 0.02 0.07 0.28
73.3 0.14 12.5 2.35 1.04 0.08 0.07 0.26 4.8 4.74 0.27 0.16 0.02 0.13 0.60
74.4 0.01 14.9 0.10 0.04 0.01 0.02 -4.9 4.67 0.50 0.17 0.02 0.15 0.17
69.0 0.31 14.7 1.21 2.35 0.10 0.11 0.83 5.1 5.56 0.28 0.11 0.03 0.06 0.65
62.0 0.99 15.8 1.77 4.64 0.17 0.94 2.73 4.7 5.07 0.41 -0.30 0.08 0.42
63.3 0.78 15.7 1.32 4.49 0.20 0.58 2.27 4.8 5.46 0.36 0.07 0.24 0.06 0.42
60.1 0.51 18.1 6.33 -0.27 0.24 0.46 5.8 5.86 1.46 0.44 0.31 0.10 0.30
Trace elements (ppm) F CI Be Cr Li Nb Ni Cu Zn V Zr Y Sr Ba Rb
CI P W norms Q c or ab an ac di hy mt hm il ap ce H2 O+ H:O-
3010 1800 5
1290 200 5
1260 220 2
2570 420 10
610 930 6
760 160 4
1240 100 4
1200 610 3
1030 810 3
3860 230 18
1290 570 2
1610 580 8
570 730 2
650 3
14 100
8 100
6 60
13 --
33 100
9 200
12 100
14 70
7 70
58 500
12 65
35 190
7 42
7 215
. 210 . 2200 105 15 370 180
-. 100 50 16 220 580
110 . 3600 65 6 -150
. 115 . 585 75 200 1065 95
145
200
545 80 100 1085 85
2200 220 19 60 350
14 150 55 500 70 310 1260 50
230 -515 100 35 260 170
28.1 1.8 27.6 41.3 -. ---0.1
15.5 . 32.9 43.2 0.8
7.7 . 32.3 40.6 5.2
19.9 . 28.5 40.0 -5.3 1.8 -2,3 . 0.9 0.1 0.2 0.5 0.3
13 95
. 310
--
--
130 . 1000 110 28 160 170
100 . 1800 50 6 -120
850 75 8 30 90
100 50 10 80 210
22.1 . 30.8 37.4 2.2
23.2 27.0 . . . 28.8 26.6 40.7 39.8 -1.9
33.2 . 25.6 29.4 -2.8 0.5 -5.9
--
1.9 2.0 2.3
1.4
1.1 1.8 1.4
.
.
--
0.3 1.8 1.5
--
--
--
0.6 0.1 0.2 0.3 0.1
0.5 0.1 0.3 0.3 0.1
0.4 0.1 0.1 0.2 0.1
.
1.4
0.5 0.1 0.2 0.3 --
27.2 28.0 37.9 -2.7 0.3 0.7 2.1 --
0.3 0.1 0.3 0.3 0.2
.
0.1 0.3 0.5 0.2
.
. 2.5 2.0 1.8 . 0.6 0.1 0.1 0.3 0.1
6.9 . 30.0 39.8 7.1 . 5.1 5.5 2.6 . 1.9 0.7 0.2 0.4 --
.
3.6 5.8 1.9 .
. t.5 0.6 0.1 0.4 0.1
--
2.0 23.4 39.2 8.8
34.5 48.8 --
--
7.4 10.6 3.2 . 3.2 1.5 0.3 0.5 0.1
2.7
--
--
46 -310 27 590 1950 60
98 -580 77 120 720 142
0.2 30.6 49.9 8.4
3.9 2.3 34.6 48.7 --
--
--
4.7 3.3 2.1
1.8 4.1 1.2
1.5 0.2 0.1 0.4 0.1
1.0 0.4 0.2 0.2 0.1
.
-0.5 -6.3 0.6 0.7 0.1 1.5 0.4
1, Amphibole-pyroxenegranite, Zomba (1980, PI9, 16); 2, amphibole-pyroxene granite, Zomba (1937, 232, 10); 3. riebeckiteaegirine granite, Zomba (1980, PI8, 15); 4, riebeckite-aegirine granite, Malosa (1980, PI9, 30); 5, aegirine-riebeckite granite, Malosa (1980, P29, 28); 6, microgranite (dyke), Malosa (1980, P19, 31); 7, amphibole-pyroxene-quartz syenite, Zomba (1980, P18, 13); 8, pyroxene-amphibole-biotite-quartz syenite, Zomba (1980, P19, 1); 9, amphibole-pyroxene-quartz syenite, Zomba (1980, P18, 20); 10, riebeckite-pyroxene-quartz syenite, Malosa (1980, P19, 33); 11, pyroxene-amphibole-fayalite syenite, Malosa (1980, P19, 18); 12, amphibole syenite, Malosa (1980, P19, 11); 13, amphibole-biotite-quartz syenite, Mpyupyu, (1980, P26, 6); 14, s61vsbergite, Chilwa Island (1957, 1056, 172). - - , element below the detection limits, which are as follows: F, 10 ppm; C1, 10 ppm; Be, 1 ppm; Cr, 5 p p m ; Li, 1 ppm; Nb, 10 ppm; Ni, 25 ppm; Cu, 10 ppm; Zn, 1 ppm; V, 50 ppm; Zr, 10 p p m ; Y, 10 ppm; Sr, 5 ppm; Ba, 25 p p m ; Rb, 5 ppm. Analysts, G. C. Jones and V. K. Din. The numbers are British Museum (Natural History) Rock Collection numbers.
Petrochemistry of the Chilwa alkaline province the newly analysed samples 14 have more than 20% normative quartz and are therefore granites, taking the boundary as recommended by Streckeisen (1976) although modal mineralogy is the basis of that system. Of the remaining 34 rocks, 27 are quartz normative, four contain a little nepheline and three are exactly silica saturated. Two of the ne-normative rocks and two of the saturated examples come from the Kasupe area at the extreme northern end of Malosa, close to the contact with the nepheline syenites of Mongolowe, while one of the other ne-normative rocks comes from a small isolated mass of syenite near the NW end of Malosa, again near Mongolowe. There appears to have been some alteration of the rocks at the northern end of Malosa caused by the emplacement of the Mongolowe complex, and this has affected concentrations of several elements; this must be taken into account when considering the significance of the chemical data. There is a distinct bimodality in silica content, with no analyses in the range 64.6%-68.8% SiO2 (Fig. 2). These rocks are also illustrated by the Qz-Ne-Ks diagram of Fig. 3, in which they plot in the 'low-temperature trough' between the granite and feldspar minima at PH2o=I kb (Tuttle & Bowen 1958). 18 of the newly analysed rocks prove to be chemically peralkaline, as indicated by the presence of acmite in the norm. Half the granites fall into this category as well as one-third of the syenites. However, all the latter come from the Kasupe area at the northern end of Malosa which, as mentioned above, is considered to have been affected by the intrusion of Mongolowe. One possible effect of this intrusion would be the introduction of sodium. The variation in the peralkalinity of the suite is illustrated by the SiO2-A1203-(Na20+K20) diagram (Bailey & Macdonald 1969) of Fig. 4. Differentiation at Zomba-Malosa was from syenites to granites and the trend illustrated by Fig. 4 could be generated by fractionation of alkali feldspar, which is invariably slightly peraluminous. The peralkaline granites are chemically very similar to comendites, which lie in the SiO2 oversaturated part of Fig. 4 between the two lines radiating from the Qz apex; the broken line separates comendites from pantellerites according to Macdonald & Bailey (1973, Fig. 6). However, there was certainly some alkali loss from the Zomba-Malosa intrusion during and after emplacement, which is clearly shown at the petrologically similar Mulanje Complex by peripheral fenitization. It is inferred, therefore, that the primary syenite magma probably had an original composition slightly more alkaline than the centre of the syenite cluster in Fig. 4, although it was not quite
341
peralkaline. The syenite differentiated towards more peralkaline compositions, almost certainly by alkali feldspar fractionation (the plagioclase effect) as no other crystallizing phase would enhance the alkali-to-alumina ratio of the liquid. The highly evolved nature of most of the rocks of the province precludes the use of variation diagrams based on Mg- and Fe-dominated differentiation indices. The greatest variation is in SiO2 and this proves to be the most useful plotting index. A range of oxides is plotted against SiO2 in Fig. 2, and most of the Zomba-Malosa and Mpyupyu rocks define simple coherent linear trends which can be interpreted as reflecting a liquid line of descent. Even Ba, which encompasses a great range in the syenites, produces a welldefined trend to low values in the granites, presumably controlled by alkali feldspar and possibly mica fractionation. One group ofsyenites plots markedly off the trends for some elements, e.g. CaO. The majority of these rocks are from the northern end of Malosa and are thought to have been affected by emplacement of the Mongolowe complex. The Ba: Sr ratio is about 8 : 1 and is reasonably linear (Fig. 5). F is generally greater than C1 with some high F:C1 ratios (Fig. 6). For comparison, the shaded area on Fig. 6 is for peralkaline oversaturated glassy lavas from East Africa (Bailey 1977, Fig. 1l). Clearly, there will have been considerable loss of halogens from the Zomba-Malosa rocks but, allowing for this, the patterns are comparable and suggest the early fixing of F and CI, presumably in amphiboles and micas.
Oversaturated dykes Specimens from four s61vsbergite dykes have been analysed, one analysis being given in Table 1. Generally they prove to have similar compositions to the syenite and quartz syenites of ZombaMalosa, but they are rather lower in Mg and much poorer in Ca, plotting in Fig. 2 with the Kasupe-area rocks. It is not clear whether this difference is due to the minor alteration which appears to have affected all the dykes or to separation of mafic phases during emplacement, or whether it is fundamental, indicating a quite separate magma batch.
Nepheline syenite and syenite
ChinduzL Mongolowe, Chaoneand Chikala A selection of analyses and norms of rocks from the E-W-trending nepheline syenite complexes
A. R. Woolley & G. C. Jones
342 f ./~,:
CO2
6
/
4
:: ..9
2
i
:
~ 9
L ',
-
.........~.~. .... ~,~
::~..." x
+~
o
,
/ 10 ~
...."'~'
Bn
,~..." ::..
8
:n 6 MgO
x "~, 9 o
. . . . . .
~
...... - ~
~ Sy
d~,~...P. ~mU'. . _
.,,.+ " ,,,D NS
x
;~':~-&~-"-----" 9 Syenite-granite
~
Oversaturated syenites- C h i k a l a Nepheline syenite - Chikala, C h a o n e , Mongolowe and Chinduzi o Nepheline syenite - J u n g u n i + Phonolite dykes 9 Metavolcanics
.... o
~
: ~ "~N
9
"
.m
~ .
i. Bn
,,
x CaO
......... - Zomba and Matosa
,~ S y e n i t e - Mpyupyu 4, Solvsbergite dykes
9
8
r
Chilwa Island Nepheline syenite e
x
Microfoyaite
t
Average ijolite
-
Nephelinite Camptonite
9 Average analyses (Le Maitre, 1976). N, nephelinite;
~
NS, nepheline syenite; 9
"-~
9. .
"
40
.,. %Sy
50
60 Si02(Wt
Bn, basanite; Sy, Syenite ~
Average sodalite
70
%)
Fio. 2. Plots of a selection of major oxides (wt. %) and barium (ppm) against SiO2. The relatively coherent trends of the Zomba-Malosa-Mpyupyu syenites and granites, the Junguni nepheline syenites and the phonolite dykes are indicated by broken lines. The Junguni trend is extrapolated to an average sodalite composition (Deer et al. 1963, Table 36). The areas occupied by the nephelinites, camptonites and nepheline syenites of Chilwa Island and of the metamorphosed basic volcanics are outlined by dotted lines. Several average compositions are plotted for comparison (Le Maitre 1976). For detailed discussion see text. is given in Table 2. A total of 23 from Chinduzi, 16 from Mongolowe, 14 from Chaone and 26 from Chikala have been analysed. Seven of the analysed rocks have normative quartz, of which five come from the early perthosite sheets of Chikala and one from a thin strip of syenite along the southern contact of Mongolowe. The seventh quartz-normative rock is from the southern contact of Chinduzi and is thought to be a fenite. The five SiO2-oversaturated Chikala rocks have similar compositions to the Zomba syenites as illustrated, for instance, by the plot of SiO2 against CaO (Fig. 2). The fact that Chikala is the
oldest of the E-W-trending nepheline syenite complexes, which are younger than ZombaMalosa, suggests that emplacement of the nepheline syenites was preceded by intrusion of a late pulse of Zomba-Malosa-type syenite. Thus, as illustrated by Fig. 4, the oversaturated perthosites are the only rocks which are common to both the nepheline syenite and syenite-granite complexes. There appears to be a general increase in the degree of undersaturation from Chikala westwards to Chinduzi (Fig. 3). The rocks mapped as pulaskites tend to be less SiO2-undersaturated than the foyaites. However, there is an overlap
Petrochemistry of the Chilwa alkaline province 1
,
1
i
,
,
i
n
i
,
,
,4
u
~l
i
i
f
,
343 ;
,
,
2800 2000 Ba (p.p.m.) 1200
.9
alA
..
9" i :...... . . :.~ ~ ~: ..... -.. ~i..
9
:
drA.
! 9
[ .
'.
. nn!
.......~:~ . . . . . .
400
... .
.. '.
-. ".
x
..
.:'
..
..
o
I I | "..~ l
x
o
. o
~...~ ,~:.~
v..
...
...
* I ~ 9 I 9149 9 9 . ~-,& 9 ~'e 9 ,~*.P'A ~
~ ...
.k ~
..:
~
9
.-~"
9 9 ~,A 9149 9
.r
o
Oli
9
,
o
I
t]
e~l~
~,
A
9
9
"~.~ ,,.
24 o
+
........
20
9
AI203
~NS
~
~e
+ +~s~.~;..
........~ . L
+
" g l ~ { ~ , ~ ". 9
16 : , ,... N, " " m ",,~ ....
~
.
,~-~.~*~
9
~ Bn
' .l""~"
~
-..,
.
x
9
9
, ..
"~
~
9
". ,..
o,,
22
'~..,
",o,
x
14 +,.+ ~, +9
Na20 10
~ .........................+ .....~ , I "
laD
:~:~ .. "'..1 . " ,... 9
J
~
4"O
NS"
,~.-..~,=~a~'~*'~.,
.
~
~'~7,--
-~
9
"~ ' ~:...,
9
: Z ' ~ & ' ~ , ;. o
"~
~ Bn
~
"
9
L
I
I
I
L
I
I
50
I
60
I
I
I
I
t
I
I
I
70
SiO2(Wt %)
and a frequency diagram of normative ne values does not show distinct populations. There is also a marked increase in the degree of peralkalinity from E to W. Only one analysed sample from Chikala has normative ac, but seven from Chaone, 10 from Mongolowe and 21 from Chinduzi contain ac, as illustrated by the peralkaline diagram (Fig. 4, inset). Plots of individual oxides and elements against SiO2 (Fig. 2) show rather vague trends. There is a fall in CaO, for instance, with increasing SiO2, but it is on quite a different trend from that of Zomba-Malosa; the same applies to MgO, NazO and A1203. There are minor variations for particular rock types and between complexes. For example, the more nepheline-rich foyaites and the rocks of Chinduzi with significant sodalite tend to have higher values of Na20 than the other complexes, while Chinduzi rocks are generally lower in TiO2, MgO and P205 and have
higher Fe203 : FeO ratios. The evidence for highlevel fractionation is thus sparse, although there was some variation between magma batches. If a decrease in MgO and C a 9 is taken to indicate the direction of evolution then an average leastevolved rock, at the present level of erosion, has about 58% SiO2, 1% MgO, 2% CaO and 8%Na20, and the main fractionation trend was towards rocks richer in SiO2 and with less modal nepheline. There are some rocks, however, particularly in Chinduzi, which are enriched in soda and are chemically similar to rocks from Junguni. These may have been involved in a different process which concentrated soda, as will be discussed later. There is some coherence in the Ba and Sr values (Fig. 5) and F and C1 values (Fig. 6). Chinduzi has high Ba:Sr ratios and a well-defined trend. Chaone and Mongolowe are somewhat variable with a higher Ba: Sr ratio for pulaskites,
344
A. R. Woolley & G. C. Jones Qz
Qz
Qz
Qz
Qz
Qz
A
Ne
.-.
9 ~o,:
Syenite-granite Syenite Nepheline syenite
: * §
Oversaturated syenite S01vsbergite Phonolite
,A- Nepheline syenite x Microfoyaite 9 Ijolite
Ks
FIG. 3. Rocks of the northern part of the Chilwa province, excluding basic rocks (i.e. metamorphosed basic volcanics, nephelinites and camptonites), plotted in terms of Q z - N e - K s (wt. %). For full key to symbols see Fig. 2. and Chikala seems to follow several trends, although significantly the five quartz-normative samples define a trend with the highest Ba:Sr ratio; this is almost identical with the ZombaMalosa trend, helping to confirm the correlation already suggested. In all the complexes F exceeds CI initially, but they evolve, particularly in Chinduzi, to rocks in which F is less than CI and with C1 values of over 10 000 ppm in two rocks, reflecting the presence of abundant sodalite. In comparison with average nepheline syenites (Le Maitre 1976) the nepheline syenites of Chikala, Chaone, Mongolowe and Chinduzi (Fig. 2) are generally 2 % - 5 ~ higher in SiO, and
somewhat lower in Na20, A1203 and CaO, while Chinduzi also has lower tenors of MgO, TiO2 and P20 5. These differences result in a much lower average normative ne value of 12.2 compared with 21.8 for Le Maitre's average nepheline syenite. This is balanced by higher normative albite in the Malawi rocks. Junguni
18 rock samples from Junguni have been analysed, and a selection is given in Table 3. All are strongly SiO2-undersaturated with normative ne ranging from 14 to 67, and as a whole they
SiO2 Chikala
Chinduzi
1I I !
I AI203.3SIO
/
}
~a i
i
....... xO ~ /~ ~,
,7/
" Zomba-Malosa
v/'
o Junguni
a20 § \K20).
i 9b r
9
1I 1
i [ I
Mongolowe chaone
§ Dykes 9 Metavolcanics
I ~o
FIo. 4. Rocks of the northern part of the Chilwa province plotted in terms of SiO2-AI203 (Na20 + K20) (wt. %). Comendites plot between the full and broken lines radiating from SiO~ and pantellerites to the right of the broken line according to Macdonald & Bailey (1973, Fig. 6).
Petrochemistry o f the Chilwa alkaline province :7omba-I~lalosa ~A~
9
cL 500
~~ #.--
_~_i__~, . . . . . z-- 9 9149 /
9
' 9 -*~*".
9 ,...........
_0 9. . . .
- ,../ ~
Chikala
~
.
., . . . . o-
-o
,#
.
.
345 .
500
.
Chaone - Mongolowe
9
9
~,A
Chinduzi
200 ~ v
T/~/"
9
/
2000
~1000 E d.
x
+
/
500
/
1500
.-->,
E r1 d.
i
v
lOOO -m-.
E .~' 500
o
.~of'~o~
~
dunguni
o
9
500
~
500
i
1000 1500 Ba(p.pm.)
,
J
2000
i
i
500
a
1500 Ba(p.p.m.)
L
Chilwa Island & dykes i
i
2500
FIG. 5. Plots of Ba against Sr (ppm). Clear linear trends are indicated by broken lines. Symbols as for Fig. 2. For discussion see text.
constitute the most undersaturated rocks in the northern part of the Chilwa Province as illustrated by the Qz-Ne-Ks diagram (Fig. 3). Two rocks have no feldspar at all in the norm and leucite occurs in three norms, although this does not reflect high values of K 2 0 - - i n fact these rocks are the poorest in potash of all the nepheline syenites--but demonstrates extreme undersaturation. Some rocks have exceptionally high values of Na20, two containing more than 20%, and the same rocks are highly enriched in C1, with up to 4.5%, reflecting their high modal sodalite. The trend of sodium enrichment clearly correlates with the presence of sodalite, as can be seen from the average sodalite composition (Deer et al. 1963, Table 36) plotted in Fig. 2. This also accounts for the unusual trends of the Junguni rocks in all the SiO2-based plots (Fig. 2). For instance, MgO, CaO, FeO and TiO2 decrease and AI203 and N a 2 0 increase with decreasing SiO2, which is the reverse of the normal situation and reflects a distinctive petrogenesis. All the Junguni rocks are ac normative and Fig. 4 illustrates both the range of SiO2 undersaturation and the peralkalinity of the rocks, which increases as they become more undersaturated because of the higher content of sodalite. The main cluster of Junguni rocks in Fig. 4 corresponds closely to the most undersaturated Chinduzi rocks, so that the increase in SiO2 undersaturation and peralkalinity from Chikala
through Chaone and Mongolowe to Chinduzi continues into Junguni. The average Ba:Sr ratio in the Junguni rocks is only marginally above unity (Fig. 5), which is lower than for the other nepheline syenite complexes. There is therefore some correlation of this ratio with the degree of saturation, ZombaMalosa having the highest and Junguni the lowest ratio. Chinduzi (Fig. 5) does not follow this trend. The amount of F in the Junguni rocks is generally similar to that in the other nepheline syenites but C1 is usually higher, sometimes exceptionally so (Table 3; Fig. 6). There is a sequence, correlating with the degree of saturation and peralkalinity, from Zomba-Malosa through the nepheline syenites: Zomba-Malosa Chikala-ChaoneMongolowe Chinduzi Junguni
F > C1 in nearly all rocks F>C1 in the majority of rocks C1 > F in the rationl4:9 C I > F in the ratio 14:4 but in some rocks in excess of 100:1
Chilwa Island Four nepheline syenites from the small plug on Chilwa Island together with two microfoyaites from dykes have been analysed (Table 3). The microfoyaites, and one in particular, prove to be
346
A. R. Woolley & G. C. Jones Phonolite oot ot
Junguni & Chinduzi
dykes 6000
+
+
d..
//
+
,,._..
/
/
/
12000
/
/
4-
10000
/
/
o
/
/
~' 4000 0..
/
/
+
2000 ~/ //4r +
+
8000 +
/ g"
//
+
++
~
++
+
E d_
Chikala Chaone
6000
6000 E
Mongolowe"
o
/
c c
o
/
,..--.
4000
,/
d. d. .._....
/
// ///
/
(.3
/i
9
..,..
4000
//
0 0
/ /
o
2000
/
9, ;
o~
2000 9
~"o
,,
/ / 9
,/..N:-4 E d..
9 yr
-
2000
....
ZombaMalosa ....
,
d.
---- 1000
9
A*,
e/./
o o
9
,,
o
/
Chilwa Island &
/
metavolcanics
9
2000
4000
F(p.p.m.)
9
2000
E d. 1ooo d
x
o x
2000
p~
0
4000
F(p.p.m.)
FIG. 6. Plots of F against CI. Symbols as for Fig. 2. For discussion see text. rather higher in MgO, CaO and total Fe than the nepheline syenites and may have involved some accumulation of mafic minerals, so that the two groups are sometimes separated on variation diagrams. The discussion is therefore concentrated on the nepheline syenites which, for clarity, are outlined in Fig. 2. The Chilwa Island nepheline syenites are quite distinct from those discussed above, being much poorer in SiO 2 (Fig. 2) with only the more sodalite-rich rocks of Junguni being m o r e S i O 2 depleted. Because of their much lower SiO 2 values they occupy a separate area of Fig. 2 and can also be distinguished by their different concentrations of other oxides and elements, e.g. CaO. Whereas the nepheline syenites of the main complexes are less undersaturated than the
average nepheline syenite of Le Maitre (1976), those from Chilwa Island are more undersaturated with an average normative n e value of 28. The different chemistry of the Chilwa Island nepheline syenites is considered to be of considerable petrogenetic significance, as is discussed later, and is emphasized by the Ba:Sr ratios and CO2 values. The nepheline syenites of all the complexes from Chikala to Junguni contain more Ba than Sr, and this is the normal relationship in such rocks. The Chilwa Island nepheline syenites, however, have more Sr than Ba, usually very much more (Fig. 5), which is a characteristic of carbonatite complexes. Similarly, CO2 is negligible in all the nepheline syenites, syenites and granites so far considered (Fig. 2), but is relatively abundant in all the silicate rocks that have been
Petrochemistry o f the Chilwa alkaline province
347
TABLE 2. Analysesof nepheline syenites, syenitesand quartz syenitesfrom Chinduzi, Mongolowe, Chaone
and Chikala 15
16
17
18
19
20
21
22
23
24
25
26
27
28
Major elements (wt. %) 58.7 SiO 2 0.23 TiO2 20.5 A1203 3.12 Fe203 FeO 1.55 MnO 0.12 0.07 MgO CaO 0.23 9.0 Na20 5.36 K20 0.42 H20 + 0.01 HzOPzOs 0.11 CO2 Others 0.43
60.0 0.76 17.2 2.17 2.84 0.16 0.66 1.88 7.4 5.54 0.44 0.08 0.18 0~12 0.54
60.1 0.31 17.6 2.81 3.15 0.22 0.25 1.27 8.1 5.21 0.38 0.04 0.06 -0.40
56.3 0.20 20.9 1.88 3.02 0.16 0.13 0.92 9.9 5.73 0.43 0.05 0.02 -0.93
59.4 0.25 20.3 2.14 1.91 0.12 0.10 0.66 9.2 5.60 0.26 --0.11 0.47
57.8 0.72 20.1 2.23 1.63 0.20 0.51 1.14 9.3 5.55 0.32 0.09 0.09 0.13 0.31
60.1 1.05 17.6 1.32 2.26 0.16 1.16 1.87 6.2 6.68 0.53 0.06 0.36 -0.57
60.3 0.95 17.9 1.64 1.77 0.24 1.00 1.36 7.0 5.48 0.78 0.14 0.17 0.29 0.53
57.2 0.73 19.7 1.59 1.97 0.20 0.84 1.60 8.8 5.70 0.47 0.14 0.25 0.15 0.55
59.2 0.74 19.4 1.46 2.17 0.13 0.82 1.93 7.2 5.64 0.46 0.06 0.21 0.14 0.26
58.4 0.80 19.0 1.53 2.73 0.17 1.05 2.39 6.8 5.35 0.41 0.02 0.28 0.15 0.42
57.3 0.78 20.1 1.15 2.89 0.17 0.96 2.05 7.8 5.66 0.43 0.06 0.28 0.23 0.50
59.3 1.08 17.3 2.36 2.65 0.19 1.16 2.81 6.0 4.60 0.30 0.11 0.36 0.48 1.22
64.0 0.63 16.5 1.22 3.54 0.19 0.60 1.83 5.2 5.40 0.18 0.10 0.17 0.14 0.28
Total
99.97
99.90 100.57 100.52 100.12
99.92
99.55
99.89
99.82
99.50 100.36
99.92
99.98
1650 270 1 10 15 60
1570 110 9
1690 1640 8
710 260 1
1480 170 5 . 11 90
1340 600 5
13 30
1570 2440 2 10 5 105
- -
99.85
Trace elements (ppm) 410 F C1 3790 Be 5 Cr Li 13 Nb 205 Ni Cu . Zn 105 V Zr 100 Y 60 Sr 25 Ba 95 Rb 230 CIP W norms Q C or
ab an ?le ac
di
hy ol mt hm it ap CC
H2O+ H20-
. 31.7 44.7 0.2 16.5 - -
0.6 . - -
4.5 0.4 - -
- -
0.4 - -
1800 1360 2 . 21 85 .
1580 690 1860 10 780 2 1 . . . 21 7 100 95 .
990 4190 4
- -
37 220
. 101
.
98
71
75
85
10 88
. 145
310 45 160 2070 83
465 45 50 490 131
215 40 20 50 121
165 45 15 75 145
155 45 60 190 87
235 45 280 3100 104
1155 60 160 1480 100
.
.
.
.
32.7 42.7 . 8.1 4.4 6.3 . 1.6 1.0 1.4 0.4 0 . 3
0.4 0 . 1
.
- -
20 275
. 30.8 42.7
.
33.9 24.4 . 27.7 5.4 3.9
. 10.2 6.2 5.2
.
.
.
.
2.3 1.0
2.9 --
0.6 0.1
0.4 0.1
- -
0.4 - -
- -
0.4 0 . 1
. 33.1 39.1 . 18.5 3.0 2.9 . 0.9 1.6 0.5 - -
- -
0.3 - -
32.8 32.6 21.6 5.5 3.6 . 1.0 0.5 1.4 0.2 0 . 3
0.3 0.1
.
39.5 41.4 0.5 6.0
7
.
. 32.4 51.2 1.2 4.4
93
.
. 49
820 55 340 870 138
230 35 380 820 72
.
1.4 1.9
1.6 2.4
33.7 31.1 -20.9 4.3 4.5 . 1.9 0.2
2.0 0.9
1.8 0.4 0.7 0.8 0.1
1.4 0.6 0.3 0.5 0.1
- -
- -
5.2 .
2.1 .
- -
0.5 0.l
.
3 100
1250 360 4 . 13 110 .
1450 990 5 . . 19 135 .
18 90
58
58
62
16 88
485 40 450 1400 132
670 40 340 t250 165
690 60 450 8300 71
200 70 150 900 133
1.1
7.2
27.2 50.8 6.7 --
31.9 44.0 5.7 --
1.5 3.6
1.2 5.7
. 33.3 43.3 4.0 9.6 . 2.8 . 1.8 2.1
31.6 43.2 5.5 7.8 . 3.0
33.5 35.9 3.1 16.3
2.9 2.2
3.1 1.7
3.4
1.8
1.4 0.5 0.3 0.5 0.1
1.5 0.7 0.3 0.4
1.5 0.7 0.5 0.4 0. l
2.1 0.9 1.1 0.3 0.1
1.2 0.4 0.3 0.2 0.1
.
. 3.2
- -
- -
15, Biotite-amphibole-aegirine-nepheline syenite, Chinduzi (1980, P20, 6); 16 aegirine-augite-amphibole-nepheline syenite, Chinduzi (1980, P20, 17); 17, aegirine-biotite amphibole-sodalite syenite, Chinduzi (1980, P20, 21); 18, amphibole biotitenepheline-sodalite syenite, Chinduzi (1980, P20, 38); 19, amphibole-aegirine-biotite-nepheline syenite, Mongolowe (1980,P21, 2); 20, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 32); 21, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 32); 21, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 33); 22, pyroxene biotite-amphibole syenite, Chaone (1980, P21, 27); 23, amphibole-biotite-nepheline syenite, Chaone (1980, P21, 13); 24, amphibole-biotite-nepheline syenite, Chaone (1980, P21, 26); 25, p y r o x e n e - a m p h i b o l e - b i o t i t e - n e p h e l i n e syenite, Chikala (1980, P22, 5); 26, a m p h i b o l e pyroxene-nepheline syenite, Chikala (1980, P22, 16); 27, pyroxene-amphibole-biotite syenite, Chikala (1980, P22, 7) 28, amphibole-quartz syenite, Chikala (1980, P22, 3). - - , as for Table 1. Analysts, G. C. Jones and V. K. Din.
348
A. R. Woolley & G. C. Jones
T A B L E 3. A n a l y s e s o f nepheline a n d sodalite syenites f r o m J u n g u n i a n d nepheline syenites, microfoyaite,
average ijolite, olivine nephelinite, camptonite a n d aln6ite f r o m Chilwa I s l a n d 29
30
31
32
33
34
35
36
37
38
39
40
41
42
13.6 1.98 3.1 4.06 8.20 0.27 6.97 28.88 0.5 1.91 0.39 0.13 5.42 24.28 1.04
Major elements (wt. %) SiO 2 TiO 2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 H20 + H20 P_,O5 CO2 Others
56.4 0.44 21.9 1.4 1.33 0.13 0.22 0.79 9.5 6.19 0.37 0.04 0.06 0.28 0.53
53.7 0.37 21.0 2.4 0.55 0.14 0.19 1.21 11.6 4.71 1.56 0.15 0.04 1.62 0.59
56.7 0.50 21.2 2.7 0.67 0.26 0.33 1.46 9.3 5.24 0.87 -0.07 0.01 0.47
56.1 0.58 20.7 2.50 0.87 0.25 0.43 1.66 10.3 4.96 0.41 0.08 0.11 0.20 0.53
48.8 0.28 24.3 1.3 0.79 0.10 0.01 0.33 17.2 2.71 0.28 0.09 0.06 0.34 2.82
43.1 0.28 26.2 1.9 0.81 0.15 -0.27 20.5 1.95 0.41 0.09 0.03 0.19 3.67
46.0 1.35 20.4 3.36 2.57 0.40 1.47 5.86 7.9 5.75 1.50 0.15 0.62 1.14 0.82
49.8 0.39 19.0 3.79 1.55 0.29 0.12 3.03 8.7 5.78 3.78 0.27 0.04 2.28 0.67
48.0 2.20 12.3 5.32 4.03 0.46 2.52 10.48 5.9 3.55 1.62 0.10 1.27 1.23 0.73
42.3 0.65 14.7 4.90 1.19 0.33 0.83 20.82 5.5 3.95 0.71 0.10 0.82 2.99 0.45
39.0 4.75 10.9 7.11 7.14 0.24 10.14 11.96 1.2 2.79 2.06 0.42 0.95 0.28 0.83
36.6 4.28 10.9 6.69 7.57 0.28 5.20 12.34 4.2 1.40 3.14 0.29 2.57 3.80 0.60
41.5 3.12 12.4 4.23 8.56 0.17 9.35 10.50 2.6 2.04 1.76 0.23 0.84 1.93 0.43
Total
99.58
99.83
99.78
99.68
99.4l
99.55
99.29
99.49
99.71
100.24
99.77
99.86
99.66 100.73
160 840 34730 44060 6 11 . 22 23 136 174 . . . . . . lll 54 77 . . . 982 488 1115 41 23 19 666 161 88 750 100 100 168 100 112
1470 510 4 6 13 187 . . 120 . 250 92 2988 2130 142
310 770 5 -15 270 .
1500 160 6 7 8 367
910 965 3 6 13 75
16 165
12 64 380 420 93 710 610 69
1800 70 2 274 8 40 54 44 125 380 22 84 2983 1150 57
2150 480 3 19 24 68 -17 154 160 135 84 1593 1260 48
1350 10 -247 9 36 65 21 101 200 89 87 1260 690 40
8.3 25.0 6.8
12.1 15.5 16.1 . 3.5
Trace elements (ppm) F C1 Be Cr Li Nb Ni Cu Zn V Zr Y Sr Ba Rb
CIP W norms C or ab an lc ne ac ns di wo ol mt hm il ap cc magn chaly H2 O+ H20-
440 5050 5 . 17 72 . . 52 . 404 17 226 100 119
680 1300 2870 570 6 8 . . . 4 17 215 198 . . . . . . 6 115 . . . 956 1179 41 48 710 558 600 600 102 200
. 35.4 26.7 . . 26.8 3.7
. 27.8 24.8 . . . . 31.7 6.9
- -
1.5 -1.2 0.2 -
-
0.8 0.1 0.6 - -
- -
0.4 - -
. 31.0 33.5
1 . 6
-----
-
-0.1 2.1 0.4 1 . 1
1.6 0 . 2
.
. . 23.5 1.7
1390 1470 11 . 26 165
. 29.3 27.0 . . 28.1 7.2
- -
1.8 1.9 -1.6 1.1 1.0 0.2 -. - -
0.9 -
-
-
-
4.4 0.4 --. 1.1 0.3 0.5 . -
-
0.4 0.1
. 16.0 15.4 .
. 5.5 --
4.8 50.9 66.9 3.8 5.5 7.4 10.2 ----0.6 1.0 --. . . 0.5 0.5 0.1 0.1 0.5 0.4 . . . 0.4 . 0.3 0.4 0.1 0.1
. 34.0 4.1 3.2 -34.0 -. 8.2 1.7 -4.9 . 2.9 1.5 2.6 . . 1.5 0.2
.
124 . 1191 96 2456 480 188
1676 87 1920 730 102
. 34.2 18.3 --25.6 7.1 . 0.3 -1.4 1.9 . 0.7 0.1 5.2 . .
. 21.0 20.0 --12.8 5.7 .
. 17.0 6.1 -4.9 4.2 3.0 2.8 .
. 3.8 0.3
. 10.8 -3.8 9.8 25.2
. 1.6 0.1
. 4.5 29.0 -3.0 2.8 1.2 1.9 6.8 . . 0.7 0.1
. 16.5 0.1 16.1 . 5.4 -. 27.4 -8.8 10.0 0.2 9.0 2.2 0.6 . . 2.0 0.4
.
.
. 5.7 -. 10.8 -6.9 9.7 -
0.1 11.3 4.4 ----
.
-
14.6
--
17.0 6.1
4.4 5.9
- -
8.1 6.1 8.6
-
-
5.9 2.0 4.4
3.8 12.8 38.8 13.8
1.7 0.2
0.2 0.1
. .
2720 290 6 72 40 212 26 33 228 260 125 87 4639 1400 45
. 3.0 2.9
29, Biotite-aegirine-nepheline-cancrinite syenite, Junguni (1980, P25, 2); 30, aegirine-nepheline-cancrinite syenite, Junguni (1980, P25, 8); 31, aegirine-nepheline syenite, Junguni (1980, P25, 17); 32, aegirine-amphibole-nepheline syenite, Junguni (1980, P25, 21); 33, sodalite-biotite-nepheline syenite, Junguni (1980, P25, 15); 34, sodalite-aegirine syenite, Junguni (1980, P25, 18); 35, nepheline syenite, Chilwa Island (1957, 1056, 133); 36, nepheline syenite, Chilwa Island (1957, 1056, 136); 37, microfoyaite, Chilwa Island (1957, 1056, 134); 38, melanite-wollastonite ijolite (average of four analyses), Chilwa Island; 39, olivine nephelinite, Chilwa Island (1957, 1056, 161); 40, olivine nephelinite, Chilwa Island (1957, 1056, 157); 41, camptonite, Chilwa Island (1957, 1056, 182); 42, aln6ite, Chilwa Island (1968, P37, 168). - - , as for Table 1. Analysts, G. C. Jones and V. K. Din.
Petrochemistry of the Chilwa alkaline province analysed from Chilwa Island. It must be stressed that the CO2 is not the result of carbonatite veining or replacement; most of the silicate rocks post-date the carbonatite and textural evidence suggests that the carbonate is a primary crystallizing phase.
Phonolite dykes Samples from 20 phonolite dykes distributed widely over the northern part of the Chilwa province have been analysed and a selection is given in Table 4. Only three specimens were collected by the present authors, 17 being originally collected by K. Bloomfield. Most of the phonolites are strongly SiO2 undersaturated, but are comparable with the main group of Junguni nepheline syenites and those of Chilwa Island (Fig. 3). They are also strongly peralkaline (Fig. 4), which correlates with their Na20 values and Fe 3 + :Fe 2+ ratios, which are higher than for the main nepheline syenites. The dykes have SiO2 values spanning the gap between the main cluster of nepheline syenites and those of Chilwa Island (Fig. 2), and for most elements they define coherent and 'normal' trends, e.g. for CaO, MgO and Ba. Although the dykes, like the main syenites, have exceptionally low concentrations of Cu, Ni and V, they are distinctly enriched in Nb, Be, Rb and Zr, elements that tend also to be enriched in the nepheline syenites of Chilwa Island. The most evolved dykes, those with the lowest MgO, CaO etc., have similar compositions to some of the main group of Junguni rocks and certain of the foyaites of the main complexes, while the least evolved dyke rocks overlap with the nepheline syenites of Chilwa Island for some elements although not for others. However, the CO2 values and Ba:Sr ratios clearly point to a correlation of the phonolites with the Chilwa Island nepheline syenites. They contain a moderate amount of CO2 (Fig. 2) which is otherwise a unique feature of the Chilwa Island rocks, and Sr is greater than Ba for many of the dykes which, as already discussed, is also characteristic of Chilwa Island rocks. Therefore, although the phonolites are widely distributed over the whole province and some of the analysed samples come from more than 50 km west of Chilwa Island, their compositions strongly suggest that they are consanguineous with the Chilwa Island centre rather than with the large nepheline syenite complexes.
Nephelinite, camptonite, ijolite and alnOite Four olivine nephelinites, two camptonites, four ijolites and an aln6ite from Chilwa Island have
349
been analysed and some results are given in Table 3. All four analysed specimes of ijolite come from the same intrusion to the N W of the carbonatite plug. One specimen was relatively homogeneous but the others are extremely heterogeneous and in part pegmatitic, so that very large samples were crushed and quartered. However, the analyses are still rather variable so an average of the four is given in Table 3, and this value is plotted on the variation diagrams although all four are shown in Fig. 3. The very high CaO in the ijolites, one of which contains 27% CaO, is not in carbonate but in wollastonite, melanite and pyroxene. The first two minerals appear to have crystallized very late, and some metasomatism may have been involved. The suite of rocks as a whole is chemically distinct from all the other rocks of the province, with the sole exception of the metavolcanics. The clearest distinction is the lower silica, where the nephelinites all have less than 40% SiOz and the camptonites only slightly more. These rocks are also lower in alkalis and A1203, Zr, Rb and Nb but contain more CaO, MgO, FeO, TiO2, V, Ni and Cu; this is clearly shown in Fig. 2. Sr exceeds Ba for all Chilwa Island rocks (Fig. 5). The camptonites have compositions similar to the average nephelinite of Le Maitre (1976) (Fig. 2), while the olivine nephelinites are similar to the melanephelinites of Homa Bay, Kenya (Le Bas 1977, Appendix 2). Compared with Homa Bay, the Malawi rocks are somewhat poorer in Na20; in three samples K20 is greater than Na20, and two samples contain substantial CO2. Only three of the nephelinites contain normative ne and this in small amounts (5.4-5.7). Normative alkali feldspar is abundant, but this value is probably inflated by the high potassium held in modal biotite and possibly in groundmass zeolite. A thorough understanding of the chemistry of the nephelinites and camptonites must await an electron microprobe study of their mineralogy, but they are the most basic silicate rocks of the province and, together with the metavolcanics, clearly represent potential parental magmas. In this regard it is noteworthy that feasible fractionation trends for the nephelinites, perhaps by separation of olivine and later pyroxene, extrapolate towards the Chilwa Island nepheline syenites rather than the nepheline syenites of the large complexes; this is illustrated, for instance, by CaO and N a 2 0 in Fig. 2. The analysis of aln6ite (Table 3) confirms the trend towards carbonatite, the norm indicating more than 50% carbonate and over 12% apatite. The very high concentrations of Sr and Ba and moderately high Nb are consistent with this tendency.
350
A. R. Woolley & G. C. Jones
T A B L E 4. A n a l y s e s o f p h o n o l i t e d y k e s a n d m e t a m o r p h o s e d 43
nephelinites from Chaone
44
45
46
47
Majorelements(wt.%) SiO 2 57.8 TiO: 0.13 AlzO3 17.6 Fe203 4.86 FeO 1.68 MnO 0.36 MgO 0.11 CaO 1.06 Na20 9.3 K20 4.89 H20 + 1.03 H20-P.,Os 0.02 CO, 0.25 Others 0.85
55.6 -20.0 4.61 0.13 0.27 0.06 0.56 11.6 4.37 0.70 0.01 0.05 1.04 0.51
51.4 0.37 21.8 0.77 3.00 0.26 0.34 1.61 7.6 8.57 2.07 0.04 0.04 1.38 0.89
59.4 0.14 17.9 4.11 2.20 0.29 0.10 1.15 8.1 4.93 1.09 0.04 0.04 0.14 0.56
52.4 1.34 17.5 5.47 1.60 0.50 1.13 1.31 10.2 5.42 0.86 0.08 0.44 0.98 1.38
40.2 3.62 14.3 5.49 9.38 0.27 6.63 10.72 4.7 1.45 1.09 0.02 1.07 0.35 0.51
41.5 2.60 13.7 4.21 7.71 0.19 7.76 14.09 4.8 1.72 0.67 0.02 0.51 0.33 0.53
42.4 3.31 14.8 4.67 9.05 0.20 5.86 9.95 4.9 2.49 1.01 0.04 0.66 0.32 0.66
Total
99.51
100.14
100.19
100.61
99.80
100.34
100.32
2060 1360 9 . 19 100 . . 130 . 600 20 2242 2230 270
2890 230 10
3770 4830 19
41 300
71 600
2660 40 1 20 21 85
4040 130 7 385 4 45 85 45 109 250 245 30 910 130 40
99.94
Trace e&ments ( p p m ) F 4180 C1 262." Be 13 Cr . Li 50 Nb 325 Ni . Cu . Zn 170 V . Zr 1190 Y 165 Sr 8 Ba 80 Rb 260 CIPWnorms C or ab an lc ne nc ac di wo ol cs mt hm il ap cc H2 O+
H20-
-28.9 36.6 -. 14.5 0.6 10.6 4.5 -0.3 . 1.8 -0.3 0.1 . 1.0 --
940 390 20 .
. 13 200 .
.
.
. 160
.
. 2400 30 16 50 150 -25.8 36.7 --
.
. 22.7 2.5 6.6 0.3 0.9 --
.
. 1.3 1.4 -0.1
.
. 0.7 --
0.4 50.7 6.4 7.7 . 22.5 3.3 ---4.3 . 1.1 . 0.7 0.1 . 2.1 --
48
.
.
49
.
. 200
200
1700 60 65 -240
2800 75 1065 670 250
30 149 350 310 35 1160 830 13
.
. 29.1 49.3 --
.
.
. 8.6 7.1 13.7
32.0 19.8 --
.
21.7 2.4 12.8 2.9 . 1.6
--10.8 8.0 22.0 --37.3
.
17.6 --24.4 .
.
4.9 0.3 0.1
1.5 . 2.5 1.0
1.1 --
0.8 0.1
.
.
8.0 .
71 136 300 385 35 940 1490 55
14.7 4.5 10.9 20.1 --26.1
.
8.6
.
.
3830 300 3 25 2 75
.
. 8.3 0.3 2.1 3.5 0.6 --
50
5.9 2.3 6.1
7.0 6.8
4.9 1.2 0.8 0.7 --
6.3 1.6 0.7 1.0 --
. 6.9 2.5 0.8 1.0 --
43, F e l d s p a r p h y r i c p y r o x e n e - b i o t i t e - a m p h i b o l e p h o n o l i t e , N E side o f C h i n d u z i (1980, P20, 1); 44, a e g i r i n e - c a n c r i n i t e p h o n o l i t e , S E o f Z o m b a (1980, P29, 1); 45, c a n c r i n i t e - b i o t i t e p h o n o l i t e , 4 k m N o f C h i n d u z i (1980, P I 9 , 24); 46, a e g i r i n e - a m p h i b o l e - b i o t i t e p h o n o l i t e , S W o f C h i n d u z i (1980, P29, 16); 47, n a t r o l i t e p h o n o l i t e , E o f M a l o s a (1980, P29, 27); 48, a m p h i b o l e - p y r o x e n e - a l k a l i f e l d s p a r - n e p h e l i n e rock, N W side o f C h a o n e (1980, P21, 40); 49, p y r o x e n e a m p h i b o l e - a l k a l i f e l d s p a r - n e p h e l i n e rock, N W side o f C h a o n e (1980, P21, 42); 50, p y r o x e n e - a m p h i b o l e - b i o t i t e - a l k a l i f e l d s p a r rock, N W side o f C h a o n e (1980, P21, 44). - - , as for T a b l e 1. Analysts, G . C. Jones a n d V. K . D i n .
Petrochemistry of the Chilwa alkal&e province Metavolcanic rocks
Eight specimens of metavolcanic rocks from the Chaone locality have been analysed (Table 4), care being taken to select samples with a minimum of veining by nepheline syenite; only two samples have very small veinlets. The resulting analyses lie very close to the average nephelinite of Le Maitre (1976) (Fig. 2) with normative n e plus leucite values of 14-30 and feldspar from 11 to 31. An analysed sample of metamorphosed basic rock from Chikala (Stillman & Cox 1960) was considered by them to be close to an average nepheline tephrite but their sample, with 6.5% normative ne, was much more felsic (normative feldspar 59%) than those analysed here. It is not certain how closely the present analyses reflect the original rock compositions, but on most variation diagrams they cluster relatively tightly (Figs. 2 and 6). On the SiOz plots (Fig. 2) they trend towards the Chilwa Island nepheline syenites, but such trends could equally well represent primary differences. However, if the metavolcanics had been substantially altered then their original compositions would have been even more basic than nephelinite, which seems unlikely. The metavolcanics are close in composition to the olivine nephelinites and camptonites of Chilwa Island, differing only a little in CO2, K20, F and Zr. The chemical evidence indicates strongly that the metavolcanics were originally nephelinite lavas, and thus also represent potential parental magma in any petrogenetic model for the province.
Discussion The Chilwa alkaline igneous province is remarkable for the extreme range of rocks it encompasses, from carbonatite to granite with many intermediate rock types. Individually, some of these rock types pose petrogenetic problems that have engaged petrologists for decades, but when such relatively exotic rocks are concentrated in one province understanding their origin becomes even more intractable. The wide range of rock types of the Tundulu carbonatite complex was considered by Garson (1962) to have evolved from kimberlite or mica-peridotite primary magmas by combinations of crystal fractionation, fenitization and rheomorphism, including the production of trachytes and phonolites by the mobilization of feldspathic and nephelinized fenites. It has also been suggested that the large intrusions of nepheline syenite represent mobilized nepheline fenites (Garson & Walshaw 1969), and these workers speculate that the syenite-
351
granite intrusions were generated by the rheomorphism of fenites formed around deep-seated carbonatite and ijolite bodies. Woolley & Garson (1970) considered that fractionation of primary alkali olivine basalt magma generated at intermediate depths, together with nephelinite-carbonatite magmas at greater depths within the mantle, could account for the province as a whole. Again, rheomorphism of fenites and nepheline fenites was suggested to have produced certain rock types. A similar scheme was erected by Woolley (1969) but called on primary nephelinitic, phonolitic and carbonatitic magmas. The syenites, quartz syenites and granites of Zomba-Malosa and Mpyupyu define coherent trends on most geochemical plots, which strongly suggest that crystal-liquid fractionation controlled a descent from slightly SiO2 oversaturated syenite to granite, with a concomitant increase in peralkalinity. The granite plots define simpler trends than the syenites but the scatter in the syenites seems to be mainly attributable to rocks at the northern end of Malosa which have undergone some metasomatic alteration associated with the emplacement of the Mongolowe nepheline syenite. Fractionation of alkali feldspar, fayalite, pyroxene, amphibole and an opaque phase is consistent with the chemistry of the series. I n contrast, the nepheline syenites and syenites of Chikala, Chaone, Mongolowe and Chinduzi are chemically similar and show only moderate alkali enrichment, particularly in Chinduzi, and a rather muted trend towards more SiO2-rich syenites. Although the nepheline syenites could be produced by fractionation of alkali feldspar from a slightly undersaturated syenitic magma driving the liquid towards the phonolite residuum, the great preponderance of nepheline syenites and the correlation of higher CaO, MgO etc. with greater SiO2 undersaturation suggests that the nepheline syenites are primary and that the few syenites are differentiates. There are slight differences between each complex, and an overall increase in SiO2 undersaturation and peralkalinity from E to W, but the indications are that these highly evolved rocks were emplaced as such and that little further evolution took place at their present level. Junguni is an exception to this and certainly underwent SiO2 impoverishment and Na and C1 enrichment leading to sodalite-rich rocks of unusual composition. The textural evidence of replacement indicates that this was a high-level effect which probably took place at the present level of intrusion, and probably involved Na concentration at the top of a magma column, perhaps by C1 streaming with the Na transported as gaseous C1 complexes.
352
A. R. Woolley & G. C. Jones
Experimental work on the system A b - N e - N aC1H20 has shown that NaCl is strongly partitioned into the vapour phase (Binsted 1981). Syenites are present in both the nepheline syenite and granite complexes, but they are chemically distinct. The only syenites that occur out of context, from the point of view of field relationships, are the quartz syenites of Chikala whose compositions match those of the syenites associated with the granites. Although there is an overlap on some chemical diagrams between the syenites of Zomba-Malosa and the syenites and nepheline syenites of the nepheline syenite complexes, there are other chemical indications that they are distinct, and this distinction forms a barrier which it is not easy to circumvent. On the SiOz-CaO plot (Fig. 2), for instance, not only is there a gap between the two rock populations but the Zomba-Malosa syenites are actually higher in CaO as well as SiO2, which poses considerable difficulties in deriving one group of rocks from the other by the crystal fractionation of an acceptable mineral phase or phases. The chemical differences between these two groups of rocks are therefore, such, as to require two parental magmas. It has been shown that the nepheline syenites of Chilwa Island, as well as being in a rather different geological context, have very different compositions from the nepheline syenites of the large complexes. They do not lie on common trends with the latter but show affinities with the spatially associated nephelinites and camptonites. Again, therefore, separation in the field correlates with chemical differences which appear to be fundamental and necessitate the postulation of a third 'primary' magma. The correlation of the phonolite dykes with the Chilwa Island centre rather than with the large nepheline syenite complexes was unexpected, but suggests that the centre played a somewhat larger role than suggested by its present small area. Although they are volumetrically insignificant, the nephelinites and camptonites of Chilwa Island must inevitably have an important place in any petrogenetic scheme because of their potential parental role. This is supported by the complete absence of basalts from the province. Karoo dolerites are widespread and the Karoo lavas which are now confined to the extreme south of Malawi probably formerly extended further north, but these rocks are 20-30 Ma older than the earliest Chilwa rocks (Woolley & Garson 1970, Table 1) and are tholeiitic (Woolley et al. 1979; Macdonald et al. 1983). The remnants of metamorphosed lavas found on Chikala, Chaone and Mongolowe must also occupy an important place in any scheme of petrogenesis, especially
now that those of Chaone have been shown to be nephelinitic and rather similar to the nephelinites of Chilwa Island. It seems very probable that the nephelinitic lavas were extruded from the Chilwa Island centre which, together with the correlation of the phonolite dykes with Chilwa Island, suggests that this was originally a volcano of some considerable magnitude. The chemistry and field associations of the rocks of the northern part of the Chilwa province indicate that three suites of rocks are present: (1) syenite-quartz syenite-granite, (2) nepheline syenite-syenite and (3) nephelinite-carbonatitenepheline syenite. If undoubted high-level fractionation is taken into account, the presence of three parental magmas is indicated, with compositions which are essentially trachytic (slightly oversaturated), phonolitic and nephelinitic-carbonatitic. The rocks representing these three magma series are shown in Fig. 7. Only number three of the postulated magmas is basic, and there is no indication within the province of the former presence of other less evolved basic rocks from which the nepheline syenites, syenites and granites could reasonably be expected to have derived. There are no basic plugs, or basic areas within the main complexes and, most significantly, there are no basic dykes apart from the Karoo dykes which, as already explained, cannot play a direct role in petrogenesis. If alkali basalt or some more alkaline basic rock had been available some evidence of it would be expected. There may, of course, be large basic bodies at depth beneath the province, but if such bodies were present they would need to be of enormous size to generate, by crystal fractionation, the large nepheline syenite and granite-syenite complexes of the northern part of the province, and the problem would be compounded in the south by the huge peralkaline granite-syenite pluton of Mulanje which covers nearly 640 km 2. Unfortunately, no geophysical work has yet been undertaken which might indicate the likely presence or absence of such bodies, but the lack of any surface manifestation of alkali basaltic rocks is thought to indicate that they are very unlikely to be present. The southern part of the Chilwa province includes the same range of rock types as the northern, and will be the subject of a future petrochemical study. There are a few published analyses of rocks of the southern area, notably of the syenites and granites of Mulanje Mountain (Garson & Walshaw 1969), which are variably peralkaline rocks similar to those of ZombaMalosa. Of more significance, perhaps, are the two nephelinites, foyaite and phonolite from the Tundulu carbonatite complex (Garson 1962),
Petrochemistry of the Chilwa alkaline province plotted in Fig. 7, which are clearly similar to comparable rocks from Chilwa Island (Fig. 2), as is also a foyaite from the small Nkalonje carbonatite centre. A nepheline syenite from the large Mauze intrusion, in contrast, is richer in SiO 2 and more comparable with the nepheline syenites from the large undersaturated plutons of the northern part of the province. Although the data are few it seems that the important petrogenetic distinction between a nephelinite (carbonatite)-nepheline syenite series and a syenitenepheline syenite series can also be made in the southern part of the province, while the oversaturated series represented by Zomba-Malosa is to be found in Mulanje. Since the 1960s the concept of mantle metasomatism has become more generally accepted and, combined with the notion of lithosphere focussing as developed in particular by Bailey (1964, 1980), is particularly well suited to explaining the genesis of the Chilwa rocks, bearing in mind the exceptionally broad range of rock types involved, their overwhelming felsic character and the paucity of basic rocks. Further, the Chilwa province is situated on a large dome, as are other alkaline provinces in Africa, and there is a close
353
temporal association of doming, rifting and igneous activity. It is considered that the dome lies above an area of low-density mantle produced by metasomatism over a long period of time, and probably initiated by the Karoo igneous event at least 50 Ma, and probably much longer, before the earliest manifestation of alkaline igneous activity in the Chilwa province. There is evidence of faulting in the area long before Chilwa times (Dixey et al. 1937; Dixey 1956) and this may be related to the deep structures that initially helped to localize the metasomatism. The wedge of anomalous mantle produced in this way is thought to have been enriched in phlogopite, carbonate and probably apatite, passing upwards into more amphibole-rich rocks. The metasomatism progressed upwards with time until even the lower crust was affected and rocks probably barely distinguishable from fenites were generated. The wedge ofmetasomatized mantle is thus envisaged as being zoned mineralogically and extending vertically at least from the base of the lithosphere to the lower crust. The dome eventually fractured (rifted) with consequent release of volatiles and upward
+
14
+
§
§
le §
~.1.§
12
§
+
10
CaO
+§ §
8
6
§
e2 4 +
§
§
I
e3
~1%
o
2 o i
4()
o§~ § ~|5
o o,
~
ee
o~;.o.-~,~j.
"-' "
§ § ~%'og~+o o . o , Agadez series 800 ~
Continental i ntercalaire
700 600 5OO A'I" r
massif
U at Arlit
-
9 i i9. . -:::.:.:::
: -i.
I
-
0
Permian
400
200 U at Madaouela~
k
Izegouanda series
~176
Tagora series
Lower Marine,Carbonifepicontinental erous
~Talak series
FIG. 8. Schematic stratigraphic column for sedimentary deposits adjacent to the Air massif, showing the location of uranium deposits : horizontal shading, shales/siltstones; fine stipple, sandstones; diagonal etching, sandstones with current bedding;., conglomerates: v, volcanogenic sandstones with analcime.
In G a l l basin AGADES.="
FIG. 7. Geological sketch map of the basins to the west of the Air massif. Tin Mersoi basin, Upper Palaeozoic sedimentary rocks; Arlit (Izegouda series), Permian; Agades series, Trias to Jurassic; In Gall basin, Cretaceous 9The thick black lines represent postdepositional faults, related to continued movement of the Air block, providing 'traps' for the retention of uranium in carbonaceous sedimentary horizons. (From Bowden et al. 1981.) 200-60 ~ towards the centre of the complex giving the impression of a funnel-shaped array. The structure is rather like a lopolith with the feeder zone offset towards the S. Away from the border zone the texture coarsens with modally less olivine and clinopyroxene. Towards the centre anorthosite contains plagioclase crystals up to 48 cm long. Some rhythmic layering is developed near the margins. Interstitial minerals in the main anorthosite zone are brown to green amphibole, titanomagnetite with coronas of secondary biotite, minor olivine and, rarely, primary clinopyroxene (Table 2). In the coarsertextured anorthosites, titanomagnetite and ilmenite constitute about 10~ of the whole rock. However, one distinctive feature of the Ofoud
basic series is the presence of numerous layers and lenses of titanomagnetite (Karche & Moreau 1978) up to 1 0 m thick intercalated with the anorthosites. These ores contain up to 50~ modal euhedral cumulus olivine. The attitude of these mafic titanomagnetite-enriched layers confirms the overall inward dip of the layered leucogabbroanorthosite units (Fig. 10). Major-element compositions for both Ofoud and Adrar Bous are given in Table 3. TABLE 1. Anorthosites a n d associated rocks (ATr, Niger) Sub-volcanic centres
Adrar Bous, Tamgak, Ofoud, Taguei, Meugueur-Meugueur, Abontorok, Iskou Palaeozoic age 470 412 Ma Funnel-shaped With oriented and inclined intrusions plagioclase crystals near border zones; dips of 20~ ~towards centre Petrographic Gabbros, leucogabbros, association anorthosites, monzogabbros, monzoanorthosites, ferrosyenites, syenites, alkaline and peralkaline granites Enclaves of leucotroctolite; leuconorite in leucogabbro Titanomagnetite lenses in anorthosites and leucogabbros Data from Bowden et al. (1976), Karche & Vachette (1976), Karche & Moreau (1977), Moreau et al. (1978), Leger (1980) and Husch & Moreau (1982).
Niger-Nigerian alkaline ring complexes
FIG. 9. Geological sketch maps of four principal anorthosite centres Air, Niger: 1, basic rocks (gabbros, leucogabbros); 2, syenite-granite series; 3, monzoanorthosite. (From Husch & Moreau 1982.)
365
A similar concentric array of aligned platy plagioclase crystals with an inward-dipping foliation (about 60 ~ can be seen at Abontorok, a circular plug about 2.5 km in diameter (Fig. 11). Within the basal series there is an obvious variation in modal content from the outside towards the centre (Fig. 12). With increasing distance from the centre, the amount of plagioclase diminishes while olivine and opaque minerals increase. The plagioclase crystals become more equidimensional and alignment less pronounced. At about 100 m from the contact the rock contains approximately 65~-70% modal plagioclase. Within 5 m of the contact the grain size decreases dramatically. Cross-cutting the leucogabbro-anorthosite unit are dykes of syenite and alkaline granite, while the central part of the small massif is invaded by syenite breccia. In the leucogabbro margins and anorthositic centre, plagioclase crystals are compositionally zoned from a core of An60 to a rim of An40. Furthermore, the proportions of olivine, clinopyroxene and titanomagnetite vary considerably. This tiny massif of Abontorok convincingly demonstrates the evidence for low-pressure fractional crystallization of anorthositic liquids in the Air sub-volcanic massifs. The apparent centripetal fractionation witnessed at Abontorok and Ofoud suggests that flow differentiation may have been a dominant factor in their development.
FIG. 10. Distance view of Ofoud massif with syenite-granite peak. The lower slopes and middle ground contain inwardly-dipping anorthosite-titanomagnetite units of the Ofoud complex. The foreground shows coarse-grained anorthosite at the southern edge of the Ofoud complex.
366
P. Bowden et al. TABLE 2. Modal analyses of anorthosites and related rocks from Adrar Bous and
Ofoud, Agr, Niger vol. %
1
Quartz Alkali feldspar Plagioclase Clinopyroxene Olivine Amphibole Opaques Biotite Others
2
--62.5 7.2 29.4 -0.9 ---
--59.9 32.2 6.5 -1.4 ---
3
4
5
6
--68.3 14.9 14.3 -1.2 --
0.7 1.2 93.5 -0.3 3.0 1.0 0.3
-0.7 94.4 0.4 -3.1 0.4 --
1.0 5.4 81.8 --8.6 3.2 --
1.3
--
1.0
--
1, 2, 3, Adrar Bous; 4, 5, 6, Ofoud. Details from Husch & Moreau (1982). Major-element analyses in Table 3. TABLE 3. Major-element analyses of anorthosites
and related rocks from Adrar Bous and Ofoud, A~r, Niger wt. %
SiO2
1
2
3
4
5
6
Fe203 FeO MgO CaO Na20 K20 TiO2 P205 LOI
46.33 49.05 47.81 52.32 52.75 49.87 17.73 15.74 20.29 24.61 25.82 23.14 0.89 2.19 0.97 1.85 0.33 0.86 10.32 7.09 6.65 2.99 2.53 5.03 0.16 0.15 0.11 0.08 0.04 0.06 9.75 13.55 11.46 9.30 9.62 9.04 2.59 2.47 2.53 4.51 4.95 4.14 0.15 0.16 0.11 0.80 0.69 0.50 0.61 0.86 0.26 1.03 0.56 1.16 0.17 0.09 0.01 0.07 0.15 0.24 0.60 -0.53 0.97 1.66 2.38
Total
98.90 98.92 98.92 99.63 99.76 99.30
AI203
c o m p o s i t i o n s are p l o t t e d in Fig. 13 as Niggli p a r a m e t e r s . Syenite a n d g r a n i t e c o m p o s i t i o n s for N i g e r are i n c l u d e d in this d i a g r a m as well as t h e t r e n d for the N i g e r i a n ring c o m p l e x e s .
1,2,3, Adrar Bous; 4,5,6, Ofoud. LOI, loss on ignition. Details in Husch & Moreau (1982). Modal analyses in Table 2. F l o w sorting o f crystals at the initial e m p l a c e m e n t stage m u s t h a v e b e e n a i d e d by o t h e r m e c h a n i s m s o f m i n e r a l layering on cooling. M a j o r - e l e m e n t c o m p o s i t i o n s o f the various rock types f o u n d at A b o n t o r o k are p r e s e n t e d in Table 4, a n d c a n be c o m p a r e d w i t h c o r r e s p o n d i n g d a t a for A d r a r Bous a n d O f o u d in Table 3. T h e s e m a j o r - e l e m e n t
Petrogenesis
T h e r e is little d o u b t t h a t t h e e v o l u t i o n o f a n o r t h o s i t i c c o m p l e x e s is the result o f a n o r o g e n i c m a g m a t i s m (Emslie 1978) in b o t h the P r e c a m b r i a n a n d the P h a n e r o z o i c . H o w e v e r , the n a t u r e o f the p a r e n t m a g m a a n d t h e m e c h a n i s m s by w h i c h c o n c e n t r a t i o n s o f plagioclase c a n o c c u r are the subject of c o n t r o v e r s y (Emslie 1973, 1978; D e W a a r d et al. 1974; R y d e r 1974). V a r i o u s c o m p o n e n t s h a v e b e e n p r o p o s e d for the c o m p o sition of the p a r e n t a l m a g m a o f a n o r t h o s i t i c rocks a n d the possible link w i t h the associated acid rock types (Fig. 13). Possible p a r e n t a l m a g m a c o m p o s i t i o n s include l e u c o g a b b r o (Budd i n g t o n 1969; S i m m o n s & H a n s o n 1978), g a b b r o (Emslie 1978), m o n z o g a b b r o ( D e W a a r d et al. 1974; D u c h e s n e et al. 1974) a n d diorite ( P h i l p o t t s 1966; G r e e n 1969). A c c o r d i n g to Leger (1980, 1985) the p e t r o g e n e t i c e v o l u t i o n at the I s k o u m a s s i f (Air, Niger) s u p p o r t s the ideas of M o r s e (1979a, b) w h o suggests t h a t a basaltic m a g m a could evolve t o w a r d s a l u m i n a - r i c h c o m p o s i t i o n by f r a c t i o n a t i o n o f m e t a s t a b l e olivine. E
W
0 3 km I
I
FiG. 11. Interpretative cross-section of Abontorok : 1, Pan-African domain of Tuareg shield ; 2, gabbro border facies; 3, leucogabbro; 4, gabbroic anorthosite; 5, anorthosite ; 6, syenite; 7, granite and granite porphyry; 8, central syenitic breccia. (From Moreau 1982.)
367
Niger-Nigerian alkaline ring complexes
Major-element analyses of the gabbroanorthosite series, Abontorok, Agr, Niger
TABLE 4.
ABONTOROK
"4" 15
[\
I ~,
~,ol -\\
',,,, cpx
01 .5
centre
I 05
I 1
km
Fza. 12. Modal distribution of major mineral phases across the Abontorok complex, Air, Niger: plg, plagioclase; ol, olivine ; cpx, clinopyroxene; op, opaque minerals. (From Husch & Moreau 1982.)
wt. ~
Abl
Ab4
Ab9
Abl0
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO N%O K20 P205 LOI
42.74 5.32 16.26 2.18 14.21 0.21 5.92 7.23 3.37 0.91 0.79 1.09
45.96 3.62 16.88 3.23 10.53 0.20 3.75 8.91 3.82 1.31 1.14 0.23
53.05 0.79 25.24 0.68 2.33 0.04 0.54 9.74 4.32 0.81 0.42 0.99
54.08 0.72 25.25 0.64 2.64 0.05 0.64 9.64 4.30 1.32 0.26 0.77
Total
99.63
100.29
99.21
100.60
Abl, ferrogabbro; Ab4, anorthositic gabbro; Ab9, Abl0, anorthosite. Taken from Moreau (1982).
c+fm
t
III FII I" t 0 I I OIBm,,II II I I I IB~A I II
i
11211
tI i i
/
9
~,-A n 50 Ab50 /
1(13
w a
CO Oi 0
I o
, t n
alc
t
i i
al
FIG. 13. Niggli diagram describing the major-element variations in the Air ring complexes: squares and triangles, gabbro-anorthosite series; circles and diamonds, syenite-granite series; broken lines, trend of Nigerian anorogenic complexes. Analyses compiled from many sources, and summarized by Husch & Moreau (1982) and Leger (1985).
368
P. B o w d e n et al.
At Iskou and the other Niger anorthositic centres (Table 1) Leger (1985) has identified three distinct stages of magmatic evolution from the variations recorded in the clinopyroxene compositions (Fig. 14). The first stage is indicated by the occurrence of leuconorite cumulates and orthopyroxene relics included in olivine in the leucotroctolites. The leuconorites probably developed during crystallization at high pressure (around 10-11 kb (Green & Ringwood 1976)). Such cumulates were generated during a shortlived residence time at the crust-mantle boundary. Such sub-crustal magma chambers may have been the source which fed the important Meugueur-Meugueur ring dyke. The Meugueur-Meugueur structure is a large ring dyke 65 km in diameter which is somewhat eccentric in its relation to the Ofoud massif. The width of the dyke varies between 200 and 400 m. Its occurrence is noted topographically by erosional hollows bordered by the Precambrian basement or the sub-volcanic centres which it dissects. Compositionally the ring dyke is variable with an abundance of enclaves of leucotroctolite, leuconorite, peridotite, leucogabbro and rare ferros_yenite set in a matrix of olivine monzogabbro (30-40 vol. ~ olivine, plagioclase An 6~176 and rare clinopyroxene) with olivinerich concentrates towards the centre and more plagioclase-rich zones near the margins. Even along its perimeter the Meugueur-Meugueur ring shows further textural and compositional contrasts as well as a variable-enclave suite. For example, magmatic layering with an almost
/ / ~
Fo"
k
/ --'-.
Ar
, "~------
k'~-. ~
vertical dip is apparent in the southern part. In contrast, the N E quadrant contains a dark picritic microgabbro, anorthositic enclaves, plagioclase xenocrysts and a border containing syenitic to granitic margins. The overall impression of this remarkable structure is a steep-sided composite ring dyke (Table 5) confined to a deeplypenetrating ring fracture through the continental lithosphere with enclaves providing a remarkable record of periods of crystallization in sub-crustal and crustal magma chambers. There is some comparison between the feeder-dyke system which fed the Muskox layered intrusion in Labrador and the Meugueur-Meugueur ring dyke, although the layered gabbro-anorthosite suites noted at the present erosional level in Air are of much smaller dimensions than their Precambrian counterparts elsewhere. The first stage of deep-level mantle crystallization was followed by the formation of anorthositic cumulates in a crustal magma chamber. A complicated fractionation process is envisaged with sedimentation of ferromagnesian minerals at the base of the magma chamber and plagioclase near the roof. This model follows that proposed by Morse (1979a, b) for the Kiglapait layered intrusion from Labrador, and at the same time helps to explain the enrichment of Fe in the magma and the presence of anorthositic enclaves in the Iskou complex and the Meugueur-Meugueur ring dyke. Leger (1980, 1985) then finally considered the formation of the sequence of leucogabbro-monzogabbros and ferrosyenites as the third stage occurring close to the level of .1
\
Ac
:: ~4
.,'----o---*----I\
....___-,--\
______ ...i,,~., "\""~ ~"5.
~
,7
og
=
-- = ~
- ~ F:a
FIG. 14. Compositional variations in olivines and pyroxenes from the Iskou complex, Niger: -/t, leuconorite; V, leucotroetolite; A, anorthosite; II, basic and [-], intermediate rocks; 0, olivine-pyroxene syenite; I ) in contrast with the surrounding Precambrian granites which are calc-alkaline. It should also be noted at this stage that, despite the field and petrographical distinction between the biotitehornblende granites and the riebeckite granites, there is no significant chemical difference between them although the latter exhibit minor acmite in the norm. The large-scale desilication or sodium enrichment that characterizes other fenitized complexes (McKie 1966) is absent.
C. J. N. Fletcher & B. Beddoe-Stephens
406
0 O< O0 O0
~
0
m
oo "~t" O0 , ~
c',l~ er~
on
o~'V= r~
. . . .
~
.
-
-
.
~
-~o-
--N
o o
K
~-~
c~
1"4
~N m ~
.
.~
.~
.~
.
.m
.
. ~7 ~
9
. ~
. ~
. ~
.
. ~
9
~N
t'N
..
.
~)
.'~ ~
~.~ N
m
~
kJ .
.
~
.
~
.
.
~
.
.
.
r
r
O~ I ~
r
I~ 0
~
r
--
on ,-- m
.
x O t'xl A
~
~
--~ o n
--
.-~ t"q
~o
~
c~
4 wt. %) lavas and hypabyssal intrusions of the Gardar province: Eriksfjord formation lavas; A, mid-Gardar dykes; O, late-Gardar dykes of Nunarssuit-Isortoq; O, late-Gardar dykes of the Tugtut6q-Ilimaussaq-Nunataq zone; m, ultramafic lamprophyres (early to late Gardar). The composition fields are those suggested by Cox et al. (1979). The Hawaiian alkali-tholeiitic basalt divide is indicated (--.--) for visual reference.
-
458
B. G. J. Upton & C. H. Emeleus
(1.75-2.40) which set these Mid-Proterozoic Gardar magmas apart from the earlier Proterozoic (Ketilidian) basalts of southern Greenland and from the Phanerozoic basalts exemplified by the Mesozoic 'coast-marginal' dykes and the early Tertiary flood basalts of both the E and W Greenland provinces. The A12Og/CaO ratios confer high contents of relatively sodic plagioclase (both normatively and modally) and concomitantly low contents of clinopyroxene. Typically, the chilled rocks are plagioclase phyric or plagioclase-olivine phyric, and the coarsePb
Rb
Ba
K
ND
grained products are troctolitic. Dykes in the Nain area of Labrador intruded in the interval 1450-1040 Ma have Fe-rich troctolitic compositions while possessing alkaline to transitional normative characteristics (Wiebe 1985). Clearly the broad characteristics of the Gardar basic magmas are shared by their time equivalents in Labrador. The absence of highly magnesian basalts has already been noted. Each of the group means has a Mg number (Mg = 100 x Mg/(Mg + Fe") atomic where Fe" is calculated assuming that Fe203/ La
Ce
,
400 •"
~ /A\ ...f"ll'~.\ ,," ,/ \ ~
Sr
T
Nd
,
P
r
Zr
T~
'
i
T
i
i
i
Y
'~ O G ; C "[ug'utoq 9 u 9 O t h e r E N E d y k e s T u g t u t 6 q zone 9 ENE dykes N u n a r s s u i t - I s o r t 6 q
2OO
100
5O
2O
Late Gardar i
i
i
i
20(
i
9
I
i
i
i
'i
d y k e s BDOs v
10(
Rock Chonc]~rite 5O
2O
Mid Gardar I
i
i
20i /
J 9 x x
I
I
i
I
9 Ilimaussaq v o l c a n i c 9 Ulukasik .....
I
I
I
I
I
I
I Nd
I P
I Zr
I Ti
I Y
member
lO(
5o
Early Ga rclar i Pb
I Rb
i Ba
I K
I Nb
I La
I Ce
I Sr
FIG. 6. Chondrite-normalized incompatible-element plots for averaged groups of Gardar basaltic-hawaiitic rocks. Early Gardar: group 1 (Mussartfit member), group 2 (Ulukasik member) and group 3 (Ilimaussaq member). Mid-Gardar : group 4 (BD0 dykes) and group 5 (younger ENE and NE dykes of the Ivigtut area). Late Gardar: group 6 (Nunarssuit-Isort6q dykes), group 7 (chilled margins, older giant-dyke complex, Tugtut6q), group 8 (chilled margins, younger giant-dyke complex, Tugtut6q) and group 10 (other ENE basic dykes of the Tugtut6q swarm). Group numbers correspond to columns in Table 1.
Mid-Proterozoic magmatism in southern Greenland FeO=0.15) of less than 50. Compositions of olivine phenocrysts are more Fe rich than Foso and typically more so than Fovo. Whereas the Gardar basalt-hawaiite magmas were characterized, as a whole, by a combination of high A1203/CaO and low Mg numbers, these features are most accentuated in (i) the early~ Gardar lavas and (ii) the late-Gardar dykes of the Tugtut6q-Ilimaussaq zone (and their ENE extensions). A general geochemical affinity between the various groups is brought out in chondritenormalized incompatible-element plots (Fig. 6; chondrite values from Sun (1980)). All tend to show high values for the most highly incompatible elements Pb, Rb and especially Ba. Almost all show low values of Nb (relative to neighbouring elements in the plot), the mid-Gardar Ivigtut dykes providing the only exception. A P peak is an obvious feature in the early-Gardar lavas and the late-Gardar Tugtut6q-Ilimaussaq dykes, and is weakly developed in the mean of the Nunarssuit-Isort6q dykes. The incompatible-element patterns of the early- and late-Gardar basic magmas, with Ba and K peaks and Nb troughs, are of interest. Thompson et al. (1983) have noted that Nb troughs are commonly developed in continental flood basalts, whereas strongly-alkaline basalts more generally display relative Nb enrichment. Ti is generally low (e.g. in comparison with average oceanic alkaline basalt (Sun 1980)) throughout the Gardar basalts. Whereas these
%
7
6
9
o
o
5
o
o
~
Oo
9
o
o
TiO 2 o A
~
o ,-..J'~§ ~-~ §
cP ~ ~~ o
31 , , ~ _ A , ~
1 -o Iio
o
o 2
o
o I
I 1
I
I 2
1
I 3
P~O s
I
I 4
I
I 5
[
%
FIG. 7. TiO 2 versus PzOsplot for Gardar basic dykes and lavas: +, Eriksfjord formation lavas; A, midGardar dykes; O, late-Gardar dykes of the Nunarssuit-]sort6q zone; 9 late-Gardar dykes of the Tugtut6q-Ilimaussaq-Nunataq zone.
459
ppm
1500 o
o
Sr
o
o
o
o
o
8
100o
+
~ o%~ 80
o
~ ~
o8
o oo 0 o o~oOOOOo % ~
o
"~ +%o o ~ 9 ~o~oo ~o oo o .++e + o+11oO +i' + t ,
+Ao
o
A
o
o
+
+
+
+
+.;
,,if+
+ i
,
10100
l
20100
I
3000
Bo
FIG. 8. Sr versus Ba plot for Gardar basic dykes and lavas. Symbols as for Fig. 7. geochemical features could have resulted from assimilation of crustal rocks, it is more likely that they reflect primitive characteristics. Despite overall similarities, the basic magmas of the Tugtut6q-Ilimaussaq zone dykes and their ENE continuations into the Nunataq region are geochemically distinct from all the earlier-Gardar basaltic rocks as well as from those of the Nunarssuit-Isort6q zone. Firstly, they tend to be more alkaline (Fig. 5), a feature largely due to comparatively high K20 and low SiO2 contents. They also tend to be richer in P, Ba, St, light rareearth elements (light REE) and Nb as is shown in Figs 7-10. Figure 7 shows the higher P2Os/ TiO2 ratios and higher P2Os absolute values. Figure 8 shows that, despite broad data scatter, the Tugtut6q-Ilimaussaq-Nunataq basic dykes tend to be richer in Ba and St, whereas Fig. 9 demonstrates the tendency for these dykes to be enriched in light REE (exemplified by La) and to have higher La/Y ratios. The most striking discriminant, however, is the Zr versus Nb plot (Fig. 10). Early- and mid-Gardar basalts and hawaiites and those of Nunarssuit-Isort6q show strong Z r - N b correlation with an average Zr/Nb ratio of about 18, whereas the Tugtut6q-Ilimaussaq-Nunataq dykes are distinctly richer in Nb with Zr/Nb ratios generally between 3 and 7. The fact that the major-element compositions of the Tugtut6q-Ilimaussaq-Nunataq basalts and hawaiites are otherwise similar to those of the other groups (e.g. similar Mg numbers) and that such relatively-incompatible elements as Ti and Zr are not notably concentrated suggests that the enrichment in P etc. did not stem from advanced crystal fractionation but indicates more
B. G. J. Upton & C. H. Emeleus
460 ppm
o
o o Oo
9 A
o
9 o
+~
+9
o
oo
Ii 9 c §
A+ c
~,~+ ~
+
+
~
~ 9 o_o
I.J_
0o a,,
a~
o u P I''
+%'~ ,,. I.
1
+#+~ J,,"
|
~ 0
o
,..
+l~
o+ .--
I',,,.-
l--
ql~
1
,.-
~
-o
o CD
0
Y3
121[3
iT- I
I ~ i ~ .,,..-.
~
o ~
0
#'~"
,i,.-
0+ "~
0
s
I~
n~
0
,,,.-
~
~
g
~~
%0
I
I
~
;> ~
;
o
Massif Central, France
523
50O 200
200
100
100
9
50
50
20
20
._
c
+oJ"
9
kD
~176
x
O to x
10
c?
5
o
in 14;
llm ill I
9 I
I
I
La Ce
I
Nd
i
I
I
Srn Eu G d T b Dy
i
1
Er
Yb
/
o-o. / /
9"
16-
/
o ~ 9
/
/
/
/
/
/
~-O23
Io,
~
//
9
] 0 Nephelinites 9 Basalts / Basanites 0 Xenocryst-rich
basalts
[]I n t e r m e d i a t e
9 tl
-,5
i
.~
D.I. 50
FIG. 5. REE patterns for two phonolites from Cantal
(i, 0 ) and pumiceous glass from Mont-Dore rhyolite (11). The distribution coefficientsfor REE between sphene and trachytes are also shown (line A from Villement et al. (1980b) and line B from Downes 0983)).
7~
,,oo~-
800
~ ~ 2 ~ 2 M 2 d d 2 ~ d
~ R ~ - ~ ~ ~ o ~m
~~
~
_
~
-