GEMOC - Department of EPS - Faculty of Science - Macquarie University




Remagnetisation of Strata of the New England Fold Belt, Australia, during the Permian

M.A. LACKIE (Division of Environmental & Life Sciences, Macquarie University, Sydney NSW 2109, Australia; ph: 61-2-8508377;; P. W. Schmidt (CSIRO Division of Exploration Geoscience, PO Box 136, North Ryde NSW 2113, Australia; ph: 61-2-8878873; e-mail:

Palaeomagnetic studies of strata within the southern New England Fold Belt, Australia, show evidence of a northward migration of remagnetisation of the Belt during the Permian. In some rock units, such as the Kiah Limestone a syn-deformational remagnetisation is observed, while in others a pre-folding magnetisation is observed. The results from the Carboniferous Seaham Formation show a positive fold test (Irving, 1966, J.G.R. 71, 6025-6051) indicating that magnetisation occurred prefolding. The pole position of the Seaham Formation suggests an Early Permian age of magnetisation. Data from the Permian Werrie Basalt also shows that magnetisation has occurred prior to significant tilting of the flow units. Data from the Late Devonian/Early Carboniferous Kiah Limestone in the south of the Belt are interpreted to indicate that magnetisation occurred just before, or at the beginning of folding. This is similar to the Werrie Basalt result. However, in the central part of the Tamworth Belt the Kiah Limestone results indicate that magnetisation was acquired at various stages of folding. We hypothesise that remagnetisation and deformation both began in the southeast, although the remagnetisation predated the deformation. Both remagnetisation and deformation progressed northwest but at different rates. Deformation and remagnetisation were contemporaneous in strata in the northwest.

Palaeomagnetism, Magnetic Petrophysics and Magnetic Signature of the Mount Leyshon Gold Deposit, Queensland

M.A. LACKIE (Division of Environmental & Life Sciences, Macquarie University, NSW, 2109, Australia, ph: 61-2-8508377;;D.A. Clark (CSIRO Division of Exploration and Mining, North Ryde, NSW 2113, Australia; ph: 61-2-8878872;; D.H. French (CSIRO Division of Exploration and Mining, North Ryde, NSW 2113, Australia; ph: 61-2-8878693; e-mail:

The world-class Mount Leyshon gold deposit is associated with a Permian igneous complex comprising intrusive breccias, porphyry plugs and several generations of dykes. A complex negative magnetic anomaly, with amplitudes from -500nT to -2000nT occurs adjacent to the mineralised complex. The stable remanent magnetisation of all phases in the igneous complex and of altered country rock (metasediments, granite and dolerite dykes) is directed south and steep down, consistent with acquisition at 280Ma, during the Kiaman reversed superchron. Felsic phases of the complex and rocks with intense phyllic alteration are weakly magnetic and have Q values less than unity, whereas more mafic phases and rocks affected by early potassic (biotite-magnetite) alteration have moderate to high susceptibilities and Q values greater than unity. The source of the Mount Leyshon magnetic anomaly can be modelled as a vertical cylinder at a depth of 500-600 m with several apophyses or sheet-like magnetic zones at shallower depths. Relatively intense remanence of reversed polarity is required to account for the form of the anomaly. The evidence of a close spatial and temporal relationship between the Mount Leyshon complex and the source of the magnetic anomaly suggests that there is a genetic relationship between the gold mineralisation and the magnetic body at depth. Further investigations are under way to distinguish between alternative explanations of the magnetic anomaly: a mafic intrusive plug or a pipe-like alteration zone with secondary magnetite.


4-D Lithosphere Mapping:Constructing Stratigraphic Sections of the Lower Crust and Upper Mantle in Space and Time

Suzanne Y. O'Reilly1 and W.L. Griffin2 1.

Centre for Petrology and Lithospheric Studies, Division of Environmental & Life Sciences, Macquarie University, Sydney, 2109, Australia 2. CSIRO Division of Exploration and Mining, North Ryde, NSW, 2113, Australia

4-D lithosphere mapping is a methodology, based on the use of xenoliths and mineral concentrates from basaltic, kimberlitic and similar volcanics, that draws together geophysical, petrological, geochemical, tectonic and geochronological information to construct geologically realistic sections showing the detailed nature of the deep crust, lithospheric mantle, the crust/mantle boundary (CMB) and the asthenosphere/lithosphere boundary. The geophysical data provide remotely-sensed and broad-scale information on the physical responses of mantle and lower crustal materials and define large-scale domains with contrasting properties. The petrological data provide more specific information on real rock types and their distribution with depth at specific localities. The petrological data constrain the interpretation of geophysical data: the geophysical information allows lateral extrapolation between the individual lithospheric stratigraphic columns constructed from xenoliths and mineral concentrates. Repeated volcanic episodes that have occurred in the same crustal region can be used trace thermal, physical and chemical modifications of lithosphere through time. Ancient lithosphere rejuvenated by fluid infiltration (metasomatism) may source younger volcanism in response to a new thermal pulse as old cratonic keels erode and transform. Paleogeotherms derived from xenoliths or mineral concentrates for specific localities provide the essential reference framework for interpreting the structure of lithospheric sections (eg O'Reilly, 1993; O'Reilly and Griffin, 1995). They are a direct measurement of the thermal state of the lithosphere at the time of eruption of the host volcanic. Considerable information on deep processes is contained in the shape and position in P-T space of specific geotherms, especially by comparison with theoretical steady-state conductive model geotherms. Indeed the "real" shape of these model geotherms is imperfectly understood and even the most quoted versions of Pollack and Chapman (1977) have been defined using assumptions unconstrained by petrological information. This focus on geotherms as the basis for "4-D Lithospheric Mapping" also reflects the paramount role of thermal energy and thermal anomalies in Earth processes including the evolution of the crust and mantle. The (paleo-) geotherm at specific localities is used to place individual samples in their original stratigraphic position, and to give the distribution with depth of rock types and (with geochemical data) of processes such as metasomatism. These data can be combined with geophysical surveys to provide a 3-dimensional picture of the composition, structure and thermal state of the lower crust and upper mantle The continental lithosphere includes the continental crust and the relatively rigid part of the underlying upper mantle, characterised by conductive heat transport. Considerable evidence suggests that the crust and the lithospheric mantle are two parts of a closely linked system, and that the nature of this system has changed markedly through geologic time. For example, some types of mantle-derived lavas (eg komatiites) erupted only at the early stages of the Earth's history, and have not contributed to crustal formation since then. Studies of mantle-derived xenoliths and minerals in volcanic rocks (eg Griffin et al, 1995)) indicate major differences in the thickness, composition and thermal structure of the upper mantle beneath continental provinces of different crustal age and seismic response (Boyd and Mertzman, 1987; Jordan, 1988). Some mantle lithologies, such as depleted garnet harzburgites, are restricted to areas with Archaean crust (eg Schulze, 1995), and the available data show major differences in the chemical composition of the mantle beneath terranes of different ages (eg Griffin and O'Reilly, unpubl. data). Seismic tomography (eg Jordan, 1988; Helmstaedt and Gurney, 1995) has shown that thick lithospheric keels also are restricted to the continental nuclei, mostly of Archaean age. These relationships are not accidental, but appear to reflect secular changes in the processes that produce lithospheric mantle. Dating of mantle rocks and minerals (eg Zhao and McCulloch, 1993; Chen et al., 1994) strongly suggests that the formation/modification of lithospheric mantle and that of the overlying crust are complementary processes, and that in many regions the mantle lithosphere remains attached to that crust for aeons, until replaced or removed during tectonic episodes. But mantle roots also can be modified and eroded. In areas as diverse as eastern Brazil, Colorado/Wyoming and eastern China, ancient lithosphere has been removed and replaced by thinner, hotter and more fertile mantle during both rifting or collision events (eg Eggler et al., 1988). This lithosphere erosion changes both the heat budget and the composition of the subcontinental mantle, and those changes also control the style and composition of subsequent magmatism and associated mineralisation in the crust. Recognising regions where the crust and the underlying mantle have become decoupled is economically important, because such areas may become less prospective for some commodities (such as diamonds) but significantly more prospective for other commodities related to the subsequent magmatism (such as copper, molybdenum and gold). Geophysics and Mantle Samples: Most of our information on the nature of the subcontinental lithospheric mantle is derived either indirectly, through large-scale geophysical studies, or directly from mantle xenoliths carried up in kimberlites and other volcanic rocks. The geophysical data provide the only realistic means of mapping large-scale variations in mantle structure and using this information to interpret major tectonic processes. However, the geophysics can only be interpreted realistically if information is available on the rocks present at depth in the study area, or in other areas thought to have comparable lithospheric geology. That information can in principle be provided by xenoliths where available and by garnet - chromite mineral concentrates which have a wider spatial and temporal distribution. The development of in-situ trace-element analysis by the proton microprobe, and its application to the heavy-mineral concentrates (HMCs) generated by diamond exploration programs, have greatly expanded the information potentially accessible on mantle lithology, geochemistry and thermal structure (Griffin and Ryan, 1995). For example, garnet grains in equilibrium with olivine can yield a temperature based on Ni content and a minimum pressure estimate (PCr). Similarly, temperatures can be derived from the zinc content of individual chromite grains. Analysis of 30-50 garnets and chromites from one locality can yield an estimate of the local paleogeotherm (Ryan et al., 1995), and the true depth of origin of each grain can then be estimated by reference of TNi and TZn to that geotherm. The geochemical information for each grain can thus be placed in stratigraphic context. The original lithology from which each grain was derived can be interpreted by comparison of garnet and chromite compositions with those known from xenoliths, in which the paragenetic associations are apparent. Similarly, the geochemical signatures of a limited range of rock-forming processes have been recognised by study of xenoliths and the mineral inclusions in diamonds. Processes identifiable in the trace-element signatures of garnets and chromites include depletion by melt extraction, high-T melt-related metasomatism (Smith et al., 1993) and a lower-T process associated with the introduction of phlogopite (Shee et al., 1993). Mantle Sections and Geophysical Interpretation: The compilation of all this information provides a stratigraphic section of the mantle under each locality which shows the vertical distribution of some rock types, the thermal structure, the depth to a (chemically defined) lithosphere base, and the distribution of depleted and metasomatically-enriched rocks. These sections can be compared with geophysical data to interpret specific features of the geophysics; for instance, the position of the lithosphere base inferred beneath Siberia (Griffin et al, this volume) correlates well with the Lehman discontinuity derived from Russian Deep Seismic Sounding experiments across this area. The position of the geotherm, which can be derived from the HMC data, is known to affect both density and seismic velocities (Morgan, 1995).The deep high-Vp roots seen in seismic tomography images of some cratons are widely interpreted as reflecting low temperatures at depth. However, these "roots" also may reflect the higher proportion of harzburgite relative to lherzolite in Archaean lithosphere, because the higher seismic velocities could be caused by this higher MgO content, rather than simply a lower T. With this kind of knowledge, the geophysical data can be interpreted more precisely, and then used to map the regional variation in specific lithologic/geologic parameters, between the sites where xenolith/heavy-mineral data are available. Different types of mantle lithosphere can be recognised, and tectonic boundaries, reflected in different ages and types of mantle, can be mapped and interpreted. Mantle Domains: The determination of the size and nature of intraplate mantle domains is of fundamental importance in understanding the mechanisms and processes of mantle/crust evolution. Distinct regions of mantle can be mapped using trace-element and isotopic characteristics of whole-rock xenoliths and separated minerals mainly clinopyroxenes and garnets and the compositions of basaltic melts (eg, Menzies, 1990; Wilson and Downes, 1991; O'Reilly and Zhang, 1995). Available data indicate that the size, shape and geochemical characteristics of these mantle domains vary with respect to depth in the lithospheric column, with tectonic environment (eg craton, Phanerozoic fold belt, collision margin) and geographically for similar tectonic settings (eg., western Australia, southern Africa and Siberia cratons). Furthermore, large-scale mantle "events" inferred from geochemical and chronological (Nd-Sr isotopic and zircon ion-probing) methods indicate coupling with crustal tectonic episodes (eg O'Reilly and Griffin, 1988; Chen, et al, 1994). Differences in lithospheric type are also delineated by seismic, gravity and heat flow data (eg., Morgan and Gosnold, 1989) and MAGSAT imagery (Mayhew et al.,1985). Global distinction of cratonic and non-cratonic regimes has many important applications, such as understanding the distribution of diamond-bearing kimberlites. 4-D lithosphere mapping is a powerful tool in defining areas where diamonds may have been preserved or destroyed throughout Earth's geological evolution and in predicting target areas for new exploration.


Boyd, F.R. and Mertzman, S.A. 1987. In: Mysen, B.O. (Ed.) Geochem. Soc. Spec. Pub. 1, 13-24.

Chen, Y.D., O'Reilly, S.Y., Kinny, P.D. and Griffin, W.L. 1994. Lithos 32, 77-94.

Eggler, D.H., Meen, J.K., Welt, F., Dudas, F.O. , Furlong, K.P., McCallum, M.E. and Carlson, R.W. 1988. Color. School Mines Quart. 83, 25-40.

Griffin et al, 1995. This volume Griffin, W.L. and Ryan, C.G. 1995. In: Griffin, W.L. (Ed.) Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 53, in press.

Helmstaedt, H. and Gurney, J.J. 1995. In Griffin, W.L. (Ed.) Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 53, in press.

Jordan, T.H. 1988. J. Petrology Spec. Vol. 1988, 11-38. Mayhew M.A., Johnson, B.D. and Wasilewski, P.J., 1985. J.Geophys. Res., 90, 2511-2522, 1985

Morgan,P. and Gosnold,W.D., 1989. GSA Memoir 172,493-521.

Menzies, M.A., 1990. In Continental Mantle, M.a.Menzies, (Ed), pp 67-86, OUP, Oxford.

Morgan, P. 1995. In Griffin, W.L. (Ed.) Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 53, in press.

O'Reilly, S.Y., 1993.Heat pulses of the Earth: the volcano/ lithosphere/ asthenosphere connection through time. Plenary Address, IAVCEI General Assembly, September 1993: Abstr p82.

O'Reilly, S.Y., and Griffin, W.L., 1988. Geochim. Cosmochim. Acta, 52, 433-447.

O'Reilly, S.Y. and Griffin, W.L. 1995. 4-D Lithosphere Mapping. Tectonophysics, in press

O'Reilly and Zhang, 1995. Contrib. Mineral. Petr., in press

Pollack, H.N. and Chapman, D.S., 1977. Tectonophysics, 38, 279-296

Ryan, C.G., Griffin, W.L. and Pearson, N.J. 1995. Contr. Mineral. Petrol., submitted

Schulze, D.J. 1995. In: H.O.A. Meyer and O.H. Leonardos (eds). Kimberlites, Related Rocks and Mantle Xenoliths. CPRM Spec. Publ. 1A/93, 327-335.

Shee, S.R., Wyatt, B.A. and Griffin, W.L. 1993. IAVCEI 1993 Gen. Assemb. Abst., p98.

Smith, D., Griffin, W.L. and Ryan, C.G. 1993. Geochim. Cosmochim. Acta 57, 605-613.

Wilson, M. and Downes, H., 1991. J. Petrology, 32, 811-849

Zhao, J. and McCulloch, M. 1993. Geology 21, 463-466.


Lithosphere Evolution in the Archangelsk Kimberlite Province

Sablukov1, L., Sablukov1, S., Griffin2, W.L., O'Reilly3, S.Y., Ryan2, C.G., Win2, T.T. and Grib4, V. 1. TsNIGRI, Moscow, Russia 2. CSIRO Div. of Exploration and Mining, Box 136, North Ryde 2113, Australia 3. Division of Environmental & Life Sciences, Macquarie University, Sydney 2109, Australia 4. Archangelsk Geological Enterprise, Leningrad, Russia

Major- and trace-element compositions of diamond indicator minerals (satellite minerals) from kimberlites have been used to map lateral and vertical variations in the composition and thermal state of the mantle beneath the Zimni Bereg area of the Archangelsk kimberlite province. The study includes the diamondiferous kimberlites of the Al-series in the Zolotitsa Field, and the diamond-poor to barren kimberlites of the Fe-Ti series in the Pachuga field and the An-734 group, ca. 50 km to the east. Major element compositions have been determined by EMP, and trace elements by proton microprobe; ca.1000 grains of Cr-pyrope garnet and 300 grains of chromite from 12 kimberlites have been analysed. Temperatures have been estimated for each garnet grain using the empirical Nickel Thermometer (TNi) and for each chromite using the Zinc Thermometer (TZn) (Griffin et al., 1994; Griffin and Ryan, 1995; Ryan et al., 1995). Garnet Geotherms (Ryan et al., 1995; this conference) have been determined for each area; individual grains of garnet and chromite have then been placed in stratigraphic context by referring TNi and TZn to the relevant local paleogeotherm. The two areas show striking differences in geotherm and lithospheric structure at the time of kimberlite intrusion (mid-Paleozoic). The mantle beneath the Zolotitsa field had a relatively cool paleogeotherm, lying near a 37 mW/m2 conductive model between 800-1100°C. In contrast, the mantle sampled by the Fe-Al kimberlites had a geotherm significantly steeper than a conductive model. The base of the lithosphere, as defined by the maximum depth of Y-depleted garnets (Ryan et al., 1995), lay at _180 km depth beneath both areas, but the temperature at this depth was _1100°C beneath the Zolotitsa field and _1300°C beneath the areas intruded by Fe-Ti kimberlites. There are also marked differences in the extent of metasomatic processes beneath the two areas. The lithosphere beneath the Zolotitsa field is relatively depleted; phlogopite-related metasomatism is prominent at depths of 125-150 km, affecting up to 50% of the volume, but melt-related metasomatism is minor and essentially restricted to depths >160 km. Beneath the areas intruded by Fe-Ti kimberlites, depleted garnets make up _30% of the total, and melt-related metasomatism affects _50% of the mantle volume over the entire depth range sampled. The abundance of subcalcic harzburgitic ("G10") garnets is similar in the diamond-rich kimberlites of the Zolotitsa field and in some low-grade to barren kimberlites of the Fe-Ti series. This reflects the similar abundance (<30%) and stratigraphic distribution of harzburgite, which extends over depths of 130-180 km beneath both areas. The differences in diamond grade between the two kimberlite series reflect both metasomatism and the thermal structure of the lithosphere. Beneath the Zolotitsa field, the diamond stability field in the lithosphere extends from 130-180 km, a depth range that encompasses most of the harzburgitic rocks. Beneath the Fe-Ti kimberlite fields, the diamond stability field extends only from ca 150-180 km, and many of the harzburgitic rocks lie in the graphite stability field. Equally importantly, the lithosphere beneath the Fe-Ti kimberlite fields has been strongly affected by asthenospheric, presumably oxidising, metasomatism. Empirical evidence from many kimberlite fields worldwide indicates that this style of metasomatism is destructive to diamonds. The thinning, heating and metasomatism of the lithosphere beneath the Fe-Ti kimberlite fields is attributed to the intrusion of asthenosphere-derived magmas; this intrusion may be responsible for some of the major domal structure associated with the Archangelsk kimberlite province (Kaminsky et al. 1995). The heat input from these magmas resulted in a progressive steepening of the geotherm with depth, indicating that heat transport was at least partly non-conductive. Chromites are moderately abundant in the kimberlites of the Zolotitsa field, where they occur to depths of _160 km, but are absent in most of the Fe-Ti series kimberlites. The similarities in the rock types and stratigraphy beneath the two areas, derived from the analysis of garnet concentrates, suggest that chromite originally was present in the rocks beneath both areas. Its rarity in the Fe-Ti series kimberlites therefore is ascribed to the effects of the intense metasomatism of the lithosphere. The Solokha kimberlite of the Pachuga field contains a range of chromites with moderate to high Cr contents, but low Mg and high Ti. They show a wide range of TZn which is not reflected in the TNi distribution of the garnets from the same kimberlite, and these chromites are interpreted as largely a magmatic population. High-Mg, Cr ilmenites are abundant in the concentrates from the kimberlites of the Fe-Ti series. Although their compositions normally would be regarded as favourable for diamond preservation, they occur in barren or very low-grade kimberlites. These ilmenite suites show good magmatic fractionation trends (cf. Griffin et al., this conf.), and probably are related to the asthenospheric melts that caused the metasomatism of the mantle and the elevated geotherm. The presence of these ilmenites, and of strongly metasomatised peridotite, in the finely comminuted mantle material of the kimberlite, may account for the distinctive high-Fe,Ti nature of the kimberlites of the Pachuga and An-734 kimberlites. The traditional use of "G10" garnets, chromites and picroilmenites to estimate diamond grade can give misleading results in this region, but the methods described here (Griffin and Ryan, 1995), based on both major- and trace-element data, appear to give a reliable basis for prioritisation of exploration targets and kimberlite fields, both in this area and further south.


Griffin, W.L. and Ryan, C.G. (1995) Trace elements in indicator minerals: Area selection and target evaluation in diamond exploration. In W.L. Griffin, Ed., Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 52 (in press).

Griffin, W.L., Ryan, C.G., Gurney, J.J., Sobolev, N.V. and Win, T.T. (1994) Chromite macrocrysts in kimberlites and lamproites: geochemistry and origin. In H.O.A. Meyer and O.H. Leonardos (eds). Kimberlites, Related Rocks and Mantle Xenoliths. CPRM Spec. Publ. 1A/93, pp. 366-377.

Griffin, W.L., Kaminsky, F.V., Ryan, C.G., O'Reilly, S.Y., Win, T.T. and Ilupin, I.P. (1995) Thermal state and composition of the lithospheric mantle beneath the Daldyn kimberlite field, Yakutia. Tectonophysics (in press).

Kaminsky, F.V, Feldman, A.A., Varlamov, V.A., Boyko, A.N., Olofinsky, L.N., Shofman, I.L. and Vaganov, V.I., (1995) Prognositcation of primary diamond deposits. In W.L. Griffin, Ed., Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 52 (in press).

Ryan, C.G., Griffin, W.L. and Pearson, N., 1994. Garnet Geotherms: a technique for derivation of P-T data from Cr-pyrope garnets. Jour. Geophys. Res. (submitted).


Mapping the Siberian Lithosphere with Garnets and Spinels

Griffin, W.L.1, Kaminsky, F.2, O'Reilly, S.Y.3, Ryan, C.G.1 and Sobolev, N.V.4 1. CSIRO Division of Exploration and Mining, Box 136, North Ryde, NSW 2113, Australia; 2. Inst. of Diamonds, Moscow, Russia; 3. Macquarie Univ., Sydney 2109, Australia; 4. Russian Acad. of Sciences, Novosibirsk, Russia

Trace- and major-element analyses of garnets and chromites from kimberlites and other volcanic rocks can be used to study the thermal and compositional structure of lithospheric mantle in space and time (Griffin and Ryan, 1995). Because concentrates have a very high information content, and are easier to obtain and faster and cheaper to study than appropriate xenolith suites, this type of lithospheric mapping can be done on a large scale, making it especially suitable for integration with regional geophysical data. We have carried out an extensive study of concentrates from Paleozoic (340-390 Ma) kimberlites in the Daldyn field of the Siberian platform. Major elements have been determined by EMP, and trace elements by proton microprobe, for >800 garnets and ca 100 chromites. These data have been used to determine the thermal state and the compositional structure of the lithosphere beneath this important diamond province. The Garnet Geotherm (Ryan et al., 1995; this conference) lies near a 35 mW/m2 conductive model up to T_1100°C, in agreement with some xenolith data (Pokhilenko et al., 1991; Griffin et al. 1995). The geotherm apparently is kinked or stepped near 1200°C (Boyd, 1984), and this T coincides with the disappearance of depleted garnets and the appearance of a pronounced "asthenospheric" signature typical of high-T sheared peridotite xenoliths. This chemical and thermal discontinuity lies near 210 km depth, and coincides with the seismically determined Lehman Discontinuity in this area; this depth therefore is interpreted as the base of the lithosphere. The coincidence of the Paleozoic lithosphere-asthenosphere boundary with the present-day seismic boundary suggests that the lower lithosphere has changed little between the Paleozoic and the present. The Nickel Temperature (TNi) of each garnet and the Zinc Temperature (TZn) of each chromite can be referred to the Garnet Geotherm to place the grains into a stratigraphic context (Ryan et al., 1995; this volume). Their parent rock types can be determined by comparison of their chemistry with that of the corresponding minerals in xenoliths. Similarly, the action of different types of metasomatic fluids on the original rocks can be recognised by specific trace-element "fingerprints", and these processes also may be assigned to specific depth ranges. Beneath the Daldyn field, harzburgites (including the "megacrystalline dunites") are concentrated in the depth range 140-190 km, making up 30-60% of the peridotitic rock types (assuming all lithologies contribute similar amounts of garnet). This harzburgite-rich zone is underlain by depleted lherzolites, and then by lherzolites with "asthenospheric" chemical signatures (reduced Cr and Mg, elevated Ti, Zr, Y and Ga). Similar data from the main pipes of the Alakit field and from the Malo-Botuobiya field indicate geotherms of _35 mW/m2. Both the lithosphere and the harzburgite zone may be somewhat thinner in these areas (lithosphere-asthenosphere boundary at ca 200 km and 180 km, respectively). In the Malo-Botuobiya region, available evidence suggests that the lherzolitic portion of the lithospheric mantle is unusually depleted; despite this, harzburgites are less abundant, and less strongly stratified. Data from several Devonian kimberlite fields along a north-south traverse from Mirny to the lower Olenek River indicate that both the geotherm and the base of the lithosphere rise northward from the Daldyn area, with the lithosphere thinning to ca 120 km in the lower Olenek region. The base of the harzburgite-rich layer also rises, while the top stays near 140 km, and this layer apparently pinches out in the region of the upper Olenek River. Depleted lherzolites also are much less common in the northern part of the Platform. The lithosphere beneath the northern areas is more typical of Proterozoic areas (Protons) than of Archean cratonic mantle, although this area is shown as an Archon on some compilations. The northward thinning and compositional variation of the lithosphere is defined by kimberlites with ages in the 350-450 Ma age range, extending into the Middle Olenek region, and appears to rpresent a synchronous lateral variation. However, the northernmost kimberlites sampled in this study (in the Kuoika field) are Jurassic-Cretaceous in age, and the still higher geotherm and much thinner lithosphere in this area may reflect post-Paleozoic processes, perhaps related to the opening of the Arctic Ocean. If this is the case, then older kimberlites in this area may reflect a thicker lithosphere.


Boyd, F.R. (1984) A Siberian geotherm based on lherzolite zenoliths from the Udachnaya kimberlite, U.S.S.R.. Geology, 12: 528-530.

Griffin, W.L. and Ryan, C.G. (1995) Trace elements in indicator minerals: Area selection and target evaluation in diamond exploration. In W.L. Griffin, Ed., Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 52 (in press).

Griffin, W.L., Ryan, C.G., Gurney, J.J., Sobolev, N.V. and Win, T.T. (1994) Chromite macrocrysts in kimberlites and lamproites: geochemistry and origin. In H.O.A. Meyer and O.H. Leonardos (eds). Kimberlites, Related Rocks and Mantle Xenoliths. CPRM Spec. Publ. 1A/93, pp. 366-377.

Griffin, W.L., Kaminsky, F.V., Ryan, C.G., O'Reilly, S.Y., Win, T.T. and Ilupin, I.P. (1995) Thermal state and composition of the lithospheric mantle beneath the Daldyn kimberlite field, Yakutia. Tectonophysics (in press).

Pokhilenko, N.P., Pearson, G., Boyd, F.R. and Sobolev, N.V. (1991) Megacrystalline dunites and peridotites: Hosts for Siberian diamonds. Ann. Rept. Geophys. Lab., Carnegie Institutions, Washington D.C., 11-18.

Ryan, C.G., Griffin, W.L. and Pearson, N., 1994. Garnet Geotherms: a technique for derivation of P-T data from Cr-pyrope garnets. Jour. Geophys. Res. (submitted).


Trace Elements in Indicator Minerals: Area Selection and Target Evaluation in Diamond Exploration

Griffin, W.L. and Ryan, C.G. CSIRO Div. of Exploration and Mining, Box 136, North Ryde 2113, Australia

Early recognition and rejection of uneconomic prospects is essential to an economically rational diamond exploration program. Some powerful new techniques for prospect evaluation have been developed by the CSIRO, based on the trace-element capabilities of the proton microprobe (Griffin and Ryan, 1995). On the larger scale, these techniques also can contribute to the process of area selection. The nickel content of chrome-pyrope garnet equilibrated with mantle olivine increases with temperature; because the Ni content of mantle olivine is large and relatively constant (2900-360 ppm), the Ni content of the garnet gives a temperature estimate without prior knowledge of the coexisting olivine's composition. This "Ni thermometer" can be used to measure the distribution of equilibration temperatures (TNi) in garnet concentrates from exploration targets such as kimberlites and lamproites. A "Cr barometer", based on the partitioning of Cr between garnet and orthopyroxene in equilibrium with chromite, gives a minimum estimate of pressure (PCr) for each grain. By combining TNi and PCr, the position of the local paleogeotherm (the "Garnet Geotherm") can be derived from garnet - chromite concentrates (see Ryan et al., 1995, and this volume), and the depth of origin of each garnet grain can be determined by referral of its TNi to the derived geotherm. For our purposes, the lithosphere is defined as the depth to which depleted Y_ 10 ppm) garnets are found. The depth to the base of this chemically-defined lithosphere can be derived from plots of Y content vs. TNi, and where xenolith data are available, this depth is seen to correspond to the "kink" or "step" in xenolith P-T estimates. This depth therefore is regarded as the point where conduction is no longer the dominant mechanism of heat transfer, and suggests that the chemical and thermal definitions of the lithosphere are roughly coindident beneath many cratonic areas. Assuming that most macrodiamonds are derived from the lithosphere, the "Diamond Window" is defined as the range of TNi between the intersection of the geotherm with the diamond-graphite equilibrium curve and the base of the lithosphere. Diamond-rich pipes contain a large proportion of garnets with TNi in the diamond window, while diamond-poor pipes typically contain a high proportion of garnets with lower TNi, reflecting greater sampling of mantle within the graphite field. Many weakly diamondiferous and barren pipes also contain abundant garnets with high Zr, Ti and Y contents, reflecting metasomatic processes in the mantle. These relations show that the diamond content of a kimberlite is determined firstly by the extent to which it has sampled mantle within the diamond window, and secondarily by the previous depletion/metasomatism history of that volume of the mantle. The observed correlation (within some Archons) between subcalcic "G10" garnets and diamond grade reflects a stratification of the lithosphere in these regions, in which most harzburgitic rocks lie at depths within the diamond stability field. However, in some regions these rocks occur at shallow depths, while in others magmatic processes have raised the geotherm and brought the harzburgites into the graphite stability field (cf. Sablukov et al., this volume). "G10" garnets therefore are only useful as a guide to diamond prospectivity where they can be shown by Ni thermometry to be derived from within the diamond stability field. An empirical Zn thermometer (Griffin et al., 1994; Ryan et al., 1995 and this volume) allows temperature estimates for single chromite grains. The relation between Cr/Cr+Al and TZn allows a rough estimate of the local paleogeotherm, which can be used to verify the Garnet Geotherm. TZn, used in concert with statistical discriminants based on major- and trace elements, also helps to identify high-Cr chromites ("diamond-inclusion chromites") that are derived from crustal source rocks. A combined measure (G) of TNi distribution, rock type proportions and metasomatism in garnet concentrates shows a strong correlation with diamond grade, and can be used to predict the maximum probable grade of an exploration target. Significant deviations from this correlation are shown by some pipes, such as Sloan and Jwaneng, that contain a high proportion of eclogitic diamonds; in these cases the predicted grades are lower than the actual ones. Pipes with unusually high proportions of small diamonds, such as Roberts Victor, deviate in the opposite direction; the G estimator predicts a higher ("geological") grade than the relevant ("commercial") one. Despite these problems, the technique is a useful tool in prioritising exploration targets, including drainage samples, for further work. Areas with elevated geotherms are inherently less prospective for diamonds, since the geotherm enters the diamond stability field only within the deepest part of the lithosphere, or not at all. Determination of the Garnet Geotherm therefore is immediately useful in the area selection process for diamond exploration, and this can be done on the basis of early heavy-mineral sampling, before any primary source rocks have been found. With more data, stratigraphic sections showing the thickness and thermal and compositional structure of the lithosphere can be constructed from garnet and chromite concentrates. These sections can be compared with those from well-known areas of different age and diamond prospectivity, to evaluate the probability that economic diamond deposits will be found within a region.


Griffin, W.L. and Ryan, C.G. (1995) Trace elements in indicator minerals: Area selection and target evaluation in diamond exploration. In W.L. Griffin, Ed., Diamond Exploration: Into the 21st Century. Jour. Geochem. Explor. 52 (in press).

Griffin, W.L., Ryan, C.G., Gurney, J.J., Sobolev, N.V. and Win, T.T., 1994. Chromite macrocrysts in kimberlites and lamproites: geochemistry and origin. In H.O.A. Meyer and O.H. Leonardos (eds). Kimberlites, Related Rocks and Mantle Xenoliths. CPRM Spec. Publ. 1A/93, pp. 366-377.

Ryan, C.G., Griffin, W.L. and Pearson, N., 1994. Garnet Geotherms: a technique for derivation of P-T data from Cr-pyrope garnets. Jour. Geophys. Res. (submitted).


Geochemistry of Magnesian Ilenite Megacrysts from Southern African Kimberlites

W.L. Griffin1, R.O. Moore2, C.G. Ryan1, J.J. Gurney3 and T.T. Win1 1. CSIRO Div. of Exploration and Mining, Box 136, NSW 2113, Australia 2. BHP .Diamonds, 1697 Powick Road, Kelowna, B.C., Canada V1X 4LI 3. Dept of Geochemistry, Univ. of Cape Town, Rondebosch 7700, South Africa

Ilmenite megacryst suites from 20 kimberlites in southern Africa have been analyzed for trace elements (Ni, Zn, Cu, Ga, Nb ,Ta , Zr, Hf) using the proton microprobe; major and minor elements have been analyzed by electron probe. These data, combined with those from other suites (Moore et al. 1992; Griffin et al. 1991, 1994), show that ilmenite macrocrysts in kimberlites represent a minor phase, crystallizing late in the fractionation history of mafic magmas at mantle depths. Nb typically behaves as an incompatible element in the magma throughout the crystallization history, and the Nb content of ilmenite serves as an index of the degree of fractional crystallization. Observations on the chemistry of ilmenites intergrown with other megacryst phases (Moore et al. 1992; Griffin et al. 1991) allow interpretation of the changing composition of the ilmenite in terms of crystallisation history. Plots of other trace- and major elements in the ilmenites against their Nb content typically define smooth curves, with breaks in slope corresponding to changes in the assemblage of minerals crystallizing together with ilmenite. For example, Ni decreases with increasing Nb when ilmenite crystallises together with mafic silicates such as olivine, pyroxenes and garnet, but increases together with Nb when the coexisting phases are low in Ni (phlogopite, zircon). Coprecipitation of ilmenite with zircon leads to buffering of the Zr content of ilmenite at a near-constant value as Nb increases. The onset of phlogopite crystallisation leads to a decrease in the Ga content of the coexisting ilmenite. Relations between major- and trace elements, previously regarded as evidence against a fractional-crystallization origin for the ilmenites, are shown to be consistent with such an origin. Ilmenite suites from different kimberlites show broadly similar crystallization histories, but differ in detail. In general the sequence of phases crystallising together with ilmenite is: Px+Gar - Mg-Oliv ' Phlog - Zirc ' Zirc - Fe-oliv -Phlog. Two groups of kimberlites may be defined, on the basis of the maximum Zr content of ilmenite reached during fractionation. One group, defined by essentially constant Zr (ca 500-700 ppm) with increasing Nb (to _1%), is geographically restricted to the Kimberley area and Uintjiesberg, and this trend is similar to those seen in ilmenites from peridotite xenoliths. All other suites show a positive correlation between Zr and Nb through most of the fractionation sequence, and the ilmenites in each kimberlite appear to have crystallized from a single batch of magma. These magmas were broadly similar, and modelling based on limited relevant distribution-coefficient data suggests that they were alkali-picritic in composition. Small differences in initial magma composition have been magnified by extreme fractional crystallization, to produce the distinctive characters of ilmenite suites from individual kimberlites. These observations suggest a genetic relation between the megacryst magma and the host kimberlite in each pipe, but the nature of this relation is not clear. In general the Hem content of ilmenites increases with increasing Nb content. Ilmenite suites from most significantly diamondiferous kimberlites have _10% Hem on average; weakly diamondiferous kimberlites may have either low or high average Hem. Barren pipes on-craton generally contain high-Hem ilmenite suites, while those off-craton may contain low-Hem ilmenites. Ilmenite suites from significantly diamondiferous pipes tend to be dominated by high-Mg, Cr ilmenites, while poorly diamondiferous on-craton kimberlites typically have lower-Mg chromites with either low or high Cr. Barren pipes in off-craton situations show a full range of ilmenite patterns, including some with very high Mg and Cr. Because the trace-element patterns of ilmenites from different kimberlites are commonly distinctive, even within small areas, they may be used to characterize the sources of ilmenites found in loam and drainage samples, to recognise the local vs distant distribution of the sources, and to conduct inventories of individual drainages during diamond exploration.


Griffin, W.L., Ryan, C.G. and Schulze, D.J. (1991) Ilmenite and silicate megacrysts from Hamilton Branch: Trace element geochemistry and fractional crystallisation. Ext. Abst. 5th Intl. Kimb. Conf., Brazil, 148-150.

Griffin, W.L., O'Reilly, S.Y., Ryan, C.G. and Waldman, M.A. (1994) Indicator minerals from Prairie Creek and Twin Knobs lamproites: relation to diamond grade. In: H.O.A. Meyer and O.H. Leonardos (eds). Diamonds: Characterization, Genesis and Exploration. CPRM Spec. Publ. 1B/93, 302-311.

Moore, R.O., Griffin, W.L., Gurney, J.J., Ryan, C.G., Cousens, D.R., Sie, S.H. and Suter, G.F. (1992) Trace element geochemistry of ilmenite megacrysts from the Monastery kimberlite, South Africa. Lithos 29, 1-18.


Plagioclase-Bearing Lherzolite Xenoliths in Alkali Basalts from Hamar-Daban, Southern Baikal Region, Russia

Ionov1, D.A., O'Reilly1, S.Y., Ashchepkov2, I.V. 1. Centre for Petrology and Lithospheric Strudies, Division of Environmental & Life Sciences, Macquarie University, Sydney, N.S.W. 2109, Australia 2. United Institute of Geology, Mineralogy and Geophysics, Novosibirsk 630090, Russia

Lherzolite xenoliths in Miocene to Pleistocene basalts from five sites in the Hamar-Daban range in southern Siberia (Ionov et al., submitted) provide sampling of the mantle close to the axis of the Baikal rift. The xenoliths from the Tumusun and Margasan vocanic centres and two lava flows in the Margasan river valley (Fig. 1) are spinel lherzolites that commonly have foliated fabrics and spongy rims around clinopyroxene (cpx), and many contain accessory feldspar. The feldspar occurs in reaction zones adjacent to resorbed spinel and orthopyroxene (where it appears to have been formed by the reaction: spl + opx + cpx = fs + ol) and less commonly as thin, irregular veins. The temperature range for the Hamar-Daban xenolith suite (950-1010°C and a single value of 880°C) is more restricted than those for spinel peridotite xenoliths from other occurrences in the Baikal area. The feldspar-bearing lherzolites yield T estimates similar to or slightly lower than feldspar-free ones. Xenoliths from a lava outcrop near Slyudyanka at Lake Baikal (Fig. 1) yield high T values (_1120°C) and show mineral zoning indicating a major heating event. No amphibole or mica was found in the samples studied. The majority of the Hamar-Daban lherzolites are fertile and cpx-rich, as for most other occurrences in the Baikal region. Trace element compositions of selected xenoliths and their clinopyroxenes were determined by ICP-MS, INAA and proton microprobe. All xenoliths analysed are enriched in alkalies (Fig. 2). Most xenoliths and their clinopyroxenes have contents of heavy REE, Sr and Y common for fertile or moderately depleted mantle peridotites but are depleted in light REE, Nb, Ta, Th and Ba and show minor negative Ti-Zr-Hf anomalies (Fig. 2). Few are moderately enriched in LREE, Sr, Th and U. Sr-Nd isotope compositions of cpx indicate long-term depletion (probably about 2 Ga) in incompatible elements similar to unmetasomatised xenoliths from other occurrences south and east of Lake Baikal (Ionov et al., 1992). The feldspars have variable compositions but are generally alkali-rich, some are extremely rich in potassium. Bulk-rock enrichment in Na, K and Rb suggests the presence of about 1% of feldspar, an estimate consistent with petrographic observations. The formation of feldspar and of spongy aggregates after clinopyroxene, and the enrichment in alkalies appear to be recent phenomena related to infiltration of an alkali-rich, H2O-poor fluid into spinel peridotites. Estimates of crustal thicknesses in the region (40-45 km) indicate pressures of about 12 kbar just below the crust/mantle boundary, i.e. higher than the upper limit of 10-11 kbar for plagioclase stability inferred for fertile peridotites at ~1000°C (Green and Hibberson, 1970). The alkali-rich compositions of the feldspars appear to be responsible for their stability in the uppermost mantle beneath Hamar-Daban. It can be concluded based on the limited number of mantle xenoliths studied that the mantle in northwestern Hamar-Daban (between the Margasan and Tumusun volcanic centres and lower Margasan river) has consistent characteristics for at least 50 km across at the depths sampled. Typical of this mantle domain is the fine grain size and foliation of the peridotites, fertile or moderately depleted modal and major element compositions, the absence (or extreme paucity) of volatile-bearing minerals and the ubiquitous presence of intergranular feldspar and spongy clinopyroxene. Equilibration temperatures of the xenoliths from northwestern Hamar-Daban are not higher than for those from other occurrences in the region located further from the rift and are well below temperatures expected for mantle diapirs (Logatchev and Zorin 1987). Our data do not indicate any significant differences between the younger Pleistocene (1-4 m.y.) and the older (probably "pre-rift") Miocene xenolith suites. Therefore, the northwestern Hamar-Daban area appears to be beyond the zone (probably restricted to the rift axis) where the recent active rifting has significantly affected the lithospheric mantle. This study and other work on mantle xenoliths in the Baikal region (Ionov et al., 1992; Kiselev and Popov, 1992) found no evidence that the "anomalous mantle" beneath the BRZ defined by low P-wave velocities immediately beneath the Moho (7.7-7.8 km/sec) consists of partially melted rocks.

Ionov, D.A., O'Reilly, S.Y., and Ashchepkov, I.V. (1995) Plagioclase-bearing lherzolite xenoliths in alkali basalts from Hamar-Daban, southern Baikal region, Russia (submitted to Contrib. Mineral. Petrol.).

Ionov, D.A., Kramm, U., and Stosch, H.-G. (1992) Evolution of the upper mantle beneath the southern Baikal rift zone: a Sr-Nd isotope study of xenoliths from the Bartoy volcanoes. Contrib. Mineral. Petrol., 111, 235-247. Green, D.H., and Hibberson, W. (1970)

The instability of plagioclase in peridotite at high pressure. Lithos, 3, 209-221 Logatchev, N.A., Zorin, Y.A. (1987)

Evidence and causes of the two-stage development of the Baikal rift. Tectonophysics, 143, 225-234. Kiselev, A.I., Popov, A.M. (1992)

Asthenospheric diapir beneath the Baikal rift: petrological constraints. Tectonophysics, 208, 287-295.



Volatile-bearing Minerals and Lithophile Trace Elements in the Upper Mantle

D.A. Ionov, S.Y. O'Reilly (Division of Environmental & Life Sciences, Macquarie University, N.S.W. 2109, Australia); W.L. Griffin (CSIRO, Box 136, North Ryde, N.S.W. 2113, Australia); A.W. Hofmann (MPI f-r Chemie, P.f. 3060, Mainz 55020, Germany); C. Dupuy

Amphibole (n=11), mica (n=16), apatite (n=9) and clinopyroxene (n=19) from mantle xenoliths in alkali basalts from Siberia, Mongolia and Spitsbergen were analyzed for Rb, Sr, Y, Zr, Nb, Th, U, Pb, La, Ce by proton microprobe. These and additional ICP-MS data on mineral separates help to define the residence, variation ranges and elemental ratios for these elements in metasomatized mantle. Mica is the major host of Rb and Ba and has high Rb/Sr (0.13-2.1) and Ba/La ratios. Amphibole, by comparison, has low Rb/Sr ratios (usually _0.03). Apatite is a major host for Th, U, Cl, Br and Sr in the mantle, but some apatites are low in Th, U, Br. The Th/U ratio in apatites analysed ranges from 0.7 to 3.1 and is lower than that inferred for the primitive mantle. Clinopyroxene appears to be the dominant host for Sr and Zr in spinel lherzolites with disseminated amphibole and mica. Trace element partition coefficients between amphibole, mica, apatite and clinopyroxene of natural peridotites are estimated. The content of Nb in amphibole and mica is much higher than in clinopyroxene and varies from _2 ppm to 870 ppm. Vein amphibole and mica commonly have much higher Nb and Zr contents than those disseminated in peridotites. It is suggested that (Zr,Nb)-rich amphibole and mica crystallize from silicate melts (or silica-rich fluids) whereas those disseminated in wall-rock peridotite may precipitate from percolating silica-poor, late-stage fluids that have much lower contents of the HFSE. Amphibole, mica and apatite are all products of mantle metasomatism but have different abundances of some indicative trace elements and elemental ratios. Their presence and modal abundances in mantle rocks or large-scale domains may result in distinctive trace element signatures. Long-term evolution of such domains may yield distinct radiogenic isotope ratios that can be inherited by mantle-derived magmas. The long-term presence of mica is required to create lithospheric mantle reservoirs enriched in radiogenic Sr as the Rb/Sr ratio in mantle amphibole is commonly below bulk earth value. Apatite may be responsible for creating mantle reservoirs with high U/Pb ratios, but such a source may have a Th/U ratio lower than the bulk earth value. The presence of amphibole and mica in the residue at low degrees of partial melting of metasomatised mantle may result in negative HFSE, K, Rb anomalies in the melts derived. The negative Nb-Ta anomaly in arc volcanics may be related to retention of these elements by amphibole and mica in the metasomatised mantle above the subducting slab.

Geochemistry of Metasomatism Adjacent to Hydrous Dykes in the Lherz Peridotite Massif

E. McPherson1, I. J. Parkinson1, M. F. Thirlwall2, M. A. Menzies2, J. L. Bodinier3, G. Bussod4 & A. Woodland5. 1 Dept. of Geology, Australian National University, A.C.T. 0200, Australia. 2 Dept. of Geology, Royal Holloway University of London, Egham, Surrey TW20 0EX, UK. 3 Centre Geol. et Geophys., USTL, Place E. Bataillon, F-34060, France. 4 Bayerisches Geoinstitut, W-8580 Bayreuth, FRG. Min. Intsitut., Universitdt Heidelberg, D-69120, Heidelberg, Germany.

There is little consensus on the mechanisms which control fluid : rock interaction in the lithospheric mantle and hence mantle metasomatism. One place where these processes can be investigated directly is in metasomatic aureoles surrounding dykes within the continental lithospheric mantle. Dykes are common in mantle xenoliths from many localities, but the small size of xenoliths precludes evaluation of the metasomatic effects over greater length scales. Here we examine metasomatic changes in harzburgites, sampled in systematic traverses, adjacent to hydrous dykes in the Lherz peridotite massif, French Pyrenees. Whole rock harzburgite powders (as well as selected mineral separates) were analysed in order that lower pressure re-equilibration effects can be discounted. Harzburgites adjacent to amphibole-bearing dykes contain amphibole & phlogopite and have lower modal clinopyroxene and generally lower Cr# spinels than remote harzburgites. They also generally have higher MnO, REE, HFSE & Fe2O3 and lower CaO, V & Sc. Modelling indicates that harzburgites closest to the dykes are not in equilibrium with melts of the composition of hydrous dykes, but are in equilibrium with an alkali-basalt, from which the dyke minerals are thought to have segregated. Extreme variations in elemental abundance and elemental and isotopic ratios occur in <1m. No one process is able to account for all theses changes. Rather, a complex combination of precipitation of hydrous phases, reaction and equilibration of anhydrous wallrock minerals with melt from the dyke is proposed. One wallrock traverse is anomalously enriched in LREE, with low HFSE/REE ratios. For these harzburgites, interaction with melt from the amphibole-bearing dyke is unable to explain the geochemical variations. Interaction with an extremely LREE-enriched, HFSE-depleted fluid, possibly a carbonatite melt, is proposed.

Implications for the composition of melts generated in the continental lithospheric mantle are discussed. Geochemistry and Petrology of ODP Leg 125 Peridotites, Izu-Bonin-Mariana Forearc

1 Ian J. Parkinson, 2 Julian A. Pearce, 3 Matthew F. Thirlwall, 4,5 Kevin T. M. Johnson and 3 Gerry Ingram. 1 Department of Geology, Australian National University, Canberra, ACT 0200, Australia. 2 Department of Geological Sciences, Science Laboratories, South Road, Durham, DH1 3LE, UK. 3 Department of Geology, Royal Holloway University, Egham Hill, Egham, Surrey, TW20 0EX, UK. 4 Geological Institute, University of Tokyo, 7-3-1 Hongo, Tokyo 113, Japan. 5 Present Address: Bishop Museum-Natural Sciences, 1525 Bernice St., P.O. Box 19000-A, Honolulu, HI 96817-0916, USA.

ODP Leg 125 recovered a suite of serpentinised harzburgites and dunites from two serpentinite seamounts in the Izu-Bonin-Mariana forearc and they represent some of the first extant forearc peridotites. Harzburgites from both Conical Seamount in the Mariana Forearc and Torishima Forearc Seamount in the Izu-Bonin Forearc are very depleted with low modal clinopyroxene (<4%), highly refractory spinels (Cr# 0.40-0.80) and very low incompatible element contents. However, whole-rock and clinopyroxene trace element geochemistry and mineral chemistry indicate that harzburgites from the two seamounts have very different origins. Clinopyroxenes from Conical Seamount have HREE-MREE patterns similar to those from the Bouvet hotspot but at lower absolute REE contents, although the Leg 125 peridotites have elevated Ce (and Sr) contents. Whole rock patterns also record this LREE enrichment. Calculated oxygen fugacities for these harzburgites range between 1.1 log units below to 0.8 log units above the FMQ buffer and overlap the redox state of MORB peridotites. These harzburgites are interpreted as being tectonically emplaced residual mantle from the subducting Pacific Plate which have been subsequently modified by melt/aqueous fluid interaction within the mantle wedge. Dunites from Conical Seamount contain small amounts of clinopyroxene, orthopyroxene and amphibole and are LREE-enriched. Moreover, they are considerably more oxidised than the harzburgites to which they are spatially related, with calculated oxygen fugacities of 0.2 log units below to 1.2 log units above the FMQ buffer. These dunites are interpreted as being zones of melt focussing where pyroxene has been preferentially dissolved. The oxidised nature of these dunites records the passage of subduction-zone magmas allowing the surrounding harzburgites to retain their reduced oxygen fugacities. Clinopyroxenes in harzburgites from Torishima Forearc Seamount have REE patterns consistent with large degrees (>20%) of near fractional melting in the spinel field only. Whole-rock patterns bear out this assertion although some patterns also record the presence of amphibole with elevated MREE contents. LREE-enrichment is ubiquitous in both the whole-rock and clinopyroxenes. Calculated oxygen fugacities for these harzburgites range between 1.2 to 1.8 log units above the FMQ buffer similar to other subduction-zone related peridotites. These peridotites are interpreted as being refractory sub-arc mantle.

Geochemistry and Petrology of Peridotites from the Eastern Solomon Islands

1 Ian J. Parkinson and 1 Richard J. Arculus 1 Department of Geology, Australian National University, Canberra, ACT 0200, Australia.

The Solomon Islands have a complicated history of subduction zone polarity reversals, oceanic plateau collision and obduction and active arc magmatism. Associated with these complicated tectonics has been the emplacement of ultramafic rocks in fault-bound slices and within serpentinite flows from a serpentinite diapir. Petrologically these peridotites are clinopyroxene-poor lherzolites, harzburgites and pyroxenites. Many of the samples are extremely fresh and exposure is extensive and they therefore provide an unique suite of oceanic peridotites. Geochemically they span a range of depletion from moderately depleted (spinel Cr# 0.30) to depleted (spinel Cr# 0.60). Initial whole-rock trace element studies and oxygen fugacity calculations indicate that individual fault slices contain ultramafic rocks with affinities to either MORB, IAB or plume-related settings. This study will present data from two of the peridotite suites. Firstly a suite of harzburgites and pyroxenites from the island of San Jorge. Fresh harzburgite samples are LREE-enriched and moderately oxidised. Cutting these harzburgites are veins of clino- and orthopyroxenite. The pyroxenes in these veins are very depleted with Al2O3 in orthopyroxene <1.5 wt.% and <2.5 wt.% in the clinopyroxenes. These veins also contain very refractory interstitial spinel with Cr# 0.75-0.80 and high ferric iron contents (6-9 wt.% Fe2O3 and Fe+3/(Fe+3 + Al + Cr) ratios of 0.09-0.11). The composition of the pyroxenites is similar to those reported from some supra-subduction zone ophiolites and these data will be discussed within this context. Secondly ultramafic samples from the island of Choiseul range in composition from clinopyroxene-poor lherzolites to harzburgites and are extremely fresh. Whole-rock REE patterns and oxygen fugacity calculations (

Annual Report 1996