Mineral age patterns in ca. 3700 my old rocks from West Greenland

KAr,⁴⁰Ar³⁹Ar and RbSr dates are reported for minerals from the ca. 3700 my-old Amîtsoq and Isua gneisses of the Godthaabsfjord area of West Greenland. KAr dates on biotites and hornblendes range from about 1900 to 3500 my, with hornblendes having a much narrower range (ca. 2250–2750 my) than biotites. One biotite from Isua gives an impossibly high KAr date of 4940 my.⁴⁰Ar³⁹Ar mineral dates are in close agreement with conventional KAr dates over the entire range of apparent age values. The presence of minor amounts of excess argon is observed in the hornblendes, but radiogenic and excess argon in the biotites are completely homogenised and cannot be differentiated.Rb-Sr measurements on biotites are closely concordant and show that all biotites were completely open to diffusion of radiogenic⁸⁷Sr at about 1600–1700 my. This is the first proof of a regional thermal event at this time in the Archaean of West Greenland, although similar dates are well known from the Proterozoic belts to the north and south.The evidence suggests that those KAr biotite dates greater than about 2700–2800 my result from excess radiogenic argon incorporated during a thermal event of about this age or, more probably, during the 1600–1700 my Sr isotope homogenisation event. Scatter of mineral dates below about 2700 my could also be due, at least in part, to overprinting by the 1600–1700 my event.KAr mineral dates and an Rb-Sr mineral isochron from a pegmatite associated with the last major rock-forming event in the Godthaabsfjord area, namely the Qo?rqut granite, indicate an age of emplacement of 2580 ± 30 my.     

Hydrothermal Mn-deposits of the Franciscan Assemblage,II. Isotope and trace element geochemistry,and implications for hydrothermal convection at spreading centers

Major element, trace element and Sr, Nd, Pb and O isotopic data for a Franciscan Mn-deposit suggest an origin by seafloor hydrothermal circulation. Based onQ-mode factor analysis the cherts and Mn-lenses of the Blue Jay mine formed from a combination of 4 components representing 1 biogenic, 1 hydrothermal, and 2 detrital sources. RbSr, UThPb and O isotopic systematics in the Mn-lenses were affected by input from the hydrothermal circulation of material leached from the underlying basalts. Nd isotopic compositions in both cherts and Mn-lenses are identical and within the range measured for Pacific Ocean water suggesting the REE were not mobilized by hydrothermal activity. Correlation of δ¹⁸O with SiO₂ and MnO₂ in the Mn-lenses implies the lenses formed by simple mixing of hydrothermally derived Mn-oxides with seawater and biogenic silica. δ¹⁸O of the cherts is both uniform and depleted relative to DSDP Jurassic cherts but similar to microquartz-bearing cherts of the Monterey Formation: this suggests that diagenetic activity exerted more control on oxygen isotope compositions then hydrothermal alteration or metamorphism. Finally, a well defined RbSr isochron of158 ± 5Myr was obtained for these cherts and opens the possibility of determining absolute radiometric ages for similar cherts throughout the geologic record.     

Nd isotopic systematics and chemistry of Central Australian sapphirine granulites: an example of rare earth element mobility

Lower Proterozoic sapphirine-bearing and associated granulites from Central Australia exhibit the greatest range of present-day¹⁴³Nd/¹⁴⁴Nd ratios (∈_Nd(O)= ?26.5 to +112.3) yet reported for rocks believed to be cogenetic. The Nd isotopic data and REE abundances of these rocks demonstrate extreme fractionation of the rare earths during the formation of stratiform CuPbZn sulfide deposits with which they are closely associated. Field relationships, petrography and chemistry of the sapphirine granulites suggest that their protoliths comprised chlorite-rich rocks which were generated by hydrothermal alteration of a range of rock types prior to metamorphism; calculations employing REE abundances of the sapphirine granulites and associated rocks, combined with bulk solid-fluid distribution coefficient data yield high fluid/rock ratios, consistent with a pre-metamorphic hydrothermal origin for the unusual REE patterns. The SmNd data for these rocks define an age of 1760±75Ma, which is significantly younger than the crust formation age of the terrain ( 2070±125Ma) but indistinguishable from the RbSr whole rock age for granulite facies metamorphism (1790±35Ma). These data are interpreted in terms of major hydrothermal fractionation of the rare earths shortly (perhaps tens of millions of years) before granulite facies metamorphism, followed by redistribution of Nd isotopes or small fractionations of the Sm/Nd ratio during the granulite facies event, and possibly also during intense retrogression which reset RbSr whole rock and UPb zircon and monazite systematics at about 1700 Ma.     

The effect of initial230Th disequilibrium on young UPb ages: the Makalu case,Himalaya

Comparative UPb dating of zircon, xenotime and monazite from two different samples of the Himalayan “Makalu” granite shows the two U decay series to be in disequilibrium, particularly in monazite. This disequilibrium is due to excess or deficit amounts of radiogenic²⁰⁶Pb which originate from an excess or deficit of²³⁰Th, respectively, occurring initially in the mineral. Such an initial disequilibrium is caused by UTh fractionation between the crystallising mineral and the magma. Therefore, the UPb ages of Th-rich minerals such as monazite (and allanite) have to be corrected for excess²⁰⁶Pb due to excess²³⁰Th, whereas Th-poor minerals such as zircon and xenotime require a correction for a deficit of²⁰⁶Pb due to deficiency of²³⁰Th. The extent of this correction depends on the degree of ThU fractionation and on the age of the rock. For the two monazite populations analysed here, these excess amounts of²⁰⁶Pb were, with reference to the amount of radiogenic²⁰⁶Pb, 8–10% and 15–20% respectively, and less than 1% for zircon and xenotime. The varying degrees of Th enrichment relative to U in monazite show that the ThU partition coefficients for this mineral are not constant within a single granite. Furthermore, for monazite there is evidence for excess amounts of radiogenic²⁰⁷Pb originating from the decay of initial excess²³¹Pa, also enriched during crystal growth.The very low Th/U ratios of 0.196 and 0.167, determined for thetwo whole rocks from which the minerals have been extracted, substantiate the view that granite formation is a fundamental mechanism for ThU fractionation in continental crust.The different ages of 21.9 ± 0.2m.y. and24.0 ± 0.4m.y., obtained by averaging the corrected²³⁸U²⁰⁶Pb ages of the monazites, suggest that the apparently homogeneous Makalu granite was generated over a period of at least 2 m.y.     

U–Th–Pb and Rb–Sr systematics of Apollo 17 boulder 7 from the North Massif of the Taurus-Littrow Valley

Portions of highland breccia boulder 7 collected during the Apollo 17 mission were studied using UThPb and RbSr systematics. A RbSr internal isochron age of3.89 ± 0.08b.y. with an initial⁸⁷Sr/⁸⁶Sr of0.69926 ± 0.00008 was obtained for clast 1 (77135,57) (a troctolitic microbreccia). A troctolitic portion of microbreccia clast 77215,37 yielded a UPb internal isochron of3.8 ± 0.2b.y. and an initial²⁰⁶Pb/²⁰⁷Pb of 0.69. These internal isochron age are interpreted as reflecting metamorphic events, probably related to impacts, which reset RbSr and UPb mineral systems of older rocks.Six portions of boulder 7 were analyzed for U, Th, and Pb as whole rocks. Two chemical groups appear to be defined by the U, Th, and Pb concentration data. Chemical group A is characterized by U, Th, and Pb concentrations and²³⁸U/²⁰⁴Pb values which are higher than those of group B. Group A rocks have typical²³²Th/²³⁸U ratios of ～ 3.85, whereas-group B rocks have unusually high Th/U values of ～ 4.1.Whole-rock UPb and PbPb ages are nearly concordant. Two events appear to be reflected in these data — one at ～ 4.4 b.y. and one at ～ 4.5 b.y. The chemical groupings show no correlation with documented ages. The old ages of ～ 4.4 b.y. and ～ 4.5 b.y. may, like the younger ～ 4.0 b.y. ages, be related to basin excavation events.     

Composition of the core,II. Effect of high pressure on solubility of FeO in molten iron

An experimental and theoretical investigation of the effect of pressure on the solubility of FeO in molten iron has been carried out. Analyses of shock-wave compression data on iron oxides combined with measurements of the FeO bond length in “metallic” oxides suggest that the partial molar volume of FeO(V^*) dissolved in molten iron is substantially smaller than that of molten wüstite. Hence the effect of high pressure should be to increase the solubility of FeO in molten iron at a given temperature. This inference is confirmed by an experimental investigation of the effect of pressure on the position of the FeFeO eutectic. Thermodynamic calculations based on these experiments yield an estimate forV^* which is in reasonable agreement with the theoretical estimates. The experimental value ofV^* is used to calculate the effect of high pressure upon the FeFeO phase diagram. Solubility of FeO in molten iron increases sharply with pressure, the liquid immiscibility region contracts and disappears around 20 GPa and it is predicted that the FeFeO phase diagram should resemble a simple eutectic system above about 20 GPa. Analogous calculations predict that the solubility of FeO in molten iron in equilibrium with magnesiowüstite (Mg_0.8Fe_0.2)O at 2500°C increase from 14 mol.%(P = 0) to above 25 mol.% at 20 GPa. If the core formed by segregation of metallic iron originally dispersed throughout the earth, it seems inevitable that it would dissolved large amounts of FeO, thereby accounting for the observation that the density of the outer core is substantially smaller than that of pure iron under correspondingP, T conditions.     

RbSr dating of the Yatsushiro granite and gneiss,Kyushu, Japan

RbSr measurements on the Yatsushiro granite and gneiss, which had been considered stratigraphically to be of possible Precambrian age, are reported. The whole rock isochron for the granite gives an age of 352 ± 8 my with a low initial⁸⁷Sr/⁸⁶Sr ratio of 0.7037 ± 0.0006. Data for constituent minerals of the granites are dispersed irregularly around the whole rock isochron (possibly by later tectonic events). For the gneiss, a metamorphic event around 410 my is indicated by the muscovite RbSr ages. The present results do not support the possibility that the Yatsushiro granite and gneiss are Precambrian in age.     

Morphology versus UPb systematics in zircon: a high-resolution isotopic study of a zircon population from a Variscan dike in the Central Alps

UPb isotopic measurements on individual zircon crystals combined with morphological analyses permit the identification of three distinct components within the zircon population of the Saedelhorn diorite, a Variscan dike from the western Gotthard (Central Alps, Switzerland): (i) 94% of the grains in the zircon population are elongate crystals with pronounced skeletal morphology indicative of rapid growth from a supercooled melt. (ii) 5% of the population consist of turbid, mostly subhedral zircons frequently showing D-type morphology (classification according to Pupin and Turco [1]) and elevated uranium contents compared to the skeletal variety. Single-crystal and multi-grain UPb isotopic data of group (i) and (ii) zircons define an intrusion age of 293 +5/−4 m.y. for the dike. (iii) Rare, transparent zircon crystals (<1% of the zircon population) yield apparent UPb ages in the range of 370–490 m.y. and display morphological and isotopic characteristics closely resembling those of a Caledonian orthogneiss intruded by the dike. This implies presence of assimilated wall-rock components in the macroscopically homogeneous dike sample.A comparison of the data obtained by conventional analysis of multi-grain zircon fractions and those obtained by grain-by-grain analysis demonstrates that age resolution is considerably improved by single-crystal UPb dating. Furthermore, quantitative identification of zircon components assimilated by the ascending magma along its path to the present level of exposure is feasible by the latter technique. Since it is likely that such zircon grains are common to a broad variety of magma types, valuable information on age and composition of crustal layers not accessible to direct observation may readily become available by application of precise micro-analytical techniques.Low initial¹⁴³Nd/¹⁴⁴Nd(ε_Nd = −2.7) at the time of intrusion of the Saedelhorn dike requires the magma to be derived partially or totally from a crustal source. For this crustal precursor, a model age of 1050 m.y. (T_DM) is obtained, indicating that Proterozoic crust was involved in the petrogenesis of the Variscan intrusives of the Gotthard area.     

Inherited isotope systems and the source region pre-history of early Caledonian granites in the Dalradian Series of Scotland

RbSr and UPb isotope analyses are reported for two pre-metamorphic Caledonian granites which intrude Dalradian rocks in the Central Highlands of Scotland. These data indicate that the origin of the granitic magmas involved partial fusion of old crustal material.UPb systems of zircon size and magnetic fractions from the Ben Vuirich granite are strongly discordant. However, U/Pb isotopic ratios precisely define a chord which intersects concordia at 514_?7⁺⁶ m.y. and 1316_?25⁺²⁶ m.y. Geological constraints suggest that the lower intersection records the post-F₂, pre-M₃ emplacement age of the granite. The upper intersection reflects the presence of old zircon xenocrysts incorporated into the granite magma without complete isotopic resetting. The ultimate source of these xenocrysts is probably a metamorphic basement complex which formed about 1320 m.y. ago, but the immediate source region of the granites could have been Dalradian sediments derived therefrom.RbSr whole-rock systems of the Ben Vuirich granite are also strongly discordant, although 8 out of 13 data points scatter about an “errorchron” of 564 ± 24 m.y. with an initial⁸⁷Sr/⁸⁶Sr ratio of about 0.716. This is interpreted as a spurious result due to incomplete homogenization of Sr isotopes in the source region during partial fusion. Initial⁸⁷Sr/⁸⁶Sr ratios at the time of emplacement indicated by the zircon data ranged from 0.7173 to 0.7191. Whole-rock samples from the Dunfallandy Hill granite have Rb/Sr ratios 2–3 times higher than those from Ben Vuirich and define a reasonably good isochron age of 491 ± 15 m.y. with an initial⁸⁷Sr/⁸⁶Sr of 0.7185 ± 0.0008. This may date granite emplacement or subsequent resetting of the high Rb/Sr rocks during Caledonian metamorphism. RbSr systematics indicate that the crustal source regions of these and other Caledonian granites separated from the upper mantle at least ca. 800 m.y. ago and probably ca. 1300 m.y. ago, thus confirming the interpretation of the upper intersection age of the zircon UPb data.     

Interpretation of a discordant KAr age pattern (Capo Vaticano,Calabria)

KAr age determination on whole rocks, biotites, quartz-feldspar separates and pegmatitic muscovites from a small quartz dioritic stock give a complex discordant age pattern. KAr dates from whole rocks and mineral separates define a single 116 my isochron with positive intercept, whereas muscovites from pegmatites fit a 180 my isochron with a probable negative intercept.Both ages are younger than the probable crystallization age of the stock (around 300 my), indicating a complex post-crystallization history. The fit of different mineral phases and whole rocks to a single isochron with positive intercept suggests that a thermal event caused rehomogenization of Ar among different mineral phases.     

The provenance and crustal residence ages of British sediments in relation to palaeogeographic reconstructions

24 SmNd isotope analyses of fine-grained Phanerozoic and modern clastic sediments from Britain and Quebec are presented. In combination with published data, they have been used to calculate “crustal residence ages” (t_CR) and to assess the provenance of the British sedimentary mass. Sediments now preserved on either side of the suture formed by the closure of the Iapetus Ocean ca. 400 Ma ago were derived from isotopically distinct source regions. Sediments north of the suture are characterised by1.7 < t_CR < 2.8 Ga, whereas those to the south exhibit a smaller range and an average value of 1.6 Ga. The northern and southern source regions were most probably Laurentia and Gondwana respectively. It seems likely that a third source, perhaps Baltica, provided the lower Palaeozoic Southern Uplands succession. Sediments deposited in southern Britain after the closure of Iapetus were derived mainly from the recycling of older sediments. The tectonic rearrangements which occurred during the Phanerozoic are not reflected in the SmNd isotopic structure of the southern British sedimentary mass, suggesting that the Caledonian and Hercynian orogenies, and even the Grenville orogeny, involved minimal accretion of new mantle-derived material into the British and adjacent continental crust. SmNd analyses of fine-grained clastic sediments provide a powerful sedimentological tool for elucidating palaeogeography, clastic source areas, sediment recycling and maturity, and some aspects of sediment transport.     

A Nd and Sr isotopic study of the Trinity peridotite; implications for mantle evolution

Field evidence indicates that the Trinity peridotite was partially melted during its rise as a part of the upwelling convecting mantle at a spreading center. A SmNd mineral isochron for a plagioclase lherzolite yields an age,T = 427 ± 32 Ma and initialε_Nd = + 10.4 ? 0.4 which is distinctly higher than that expected for typical depleted mantle at this time. This age is interpreted as the time of crystallization of trapped melt in the plagioclase lherzoliteP-T field. This time of crystallization probably represents the time when the massif was incorporated as a part of the oceanic lithosphere. The SmNd model age of the plagioclase lherzolite totalrock isT_CHUR^Nd = 3.4 AE. This suggests that the Trinity peridotite was derived from a mantle that was depleted rather early in earth history. The peridotite contains many generations of pyroxenite dikes and some microgabbro dikes. We report data for two dikes that clearly crosscut the main metamorphic fabric of the peridotite. A microgabbro dike yields a SmNd mineral isochron age ofT = 435 ± 21 Ma andε_Nd = + 6.7 ? 0.3. A pyroxenite dike yields an initialε_Nd = + 7.3 ± 0.4. The initialε_Nd values for the pyroxenite and gabbro dikes are fairly similar to those for the depleted mantle at this time and are distinct from the lherzolite—demonstrating that they are not genetically related. RbSr data do not give any coherent pattern. However, some bounds can be put on initial Sr values ofε_Sr ? ?21 for the plagioclase lherzolite andε_Sr ? ?8.7 for the microgabbro dike. It is plausible that the dikes represent cumulates left behind from island arc magmas that rose through the the oceanic lithosphere within the vicinity of a subduction zone. Major and trace elements and SmNd isotopic data indicate a multiple stage history for the Trinity peridotite; a small melt fraction was extracted from an undepleted source ～ 3.4 AE or more ago to produce the proto-lherzolite; a large fraction of melt (～ 12 to 23%) was extracted from the proto-lherzolite to produce the present rock; the lherzolite was then crosscut by dikes from average depleted mantle ～ 0.44 AE ago. The data are compatible with the depleted mantle source being formed very early in earth history. Although most available data indicate that the depleted upper mantle has been relatively well stirred through time, the Trinity data suggest that very ancient Nd isotopic values are preserved and thus chemical and physical heteorgeneities are sometimes preserved in the depleted source of mid-ocean ridge basalts as well as the oceanic lithosphere which they intrude.     

Interpretation of Nd,Sr and Pb isotope data from Archean migmatites in Lofoten-Vesterålen,Norway

Sm-Nd data for the Archean granulite and amphibolite facies migmatites of Langøy and Hinnøy in Vesterålen are presented which indicate that their protoliths formed ～2.6 AE ago. Rubidium and U loss during a granulite facies metamorphism at ～1.8 AE caused serious disturbance of total-rock U-Pb and Rb-Sr systems. Therefore these systems do not provide any precise age information for the granulite facies migmatites. For the amphibolite facies migmatites of Vesterålen both SmNd, RbSr and PbPb total-rock systems give model ages of ～2.6 AE. The results on both granulite and amphibolite facies rocks are thus in agreement. Previous interpretations based on PbPb data, which indicated an age of 3.41 AE for the Archean terrane of Vesterålen, are not valid.One SmNd model age from the granulite facies migmatites at Moskenesøy in Lofoten indicates that the protoliths of these migmatites formed ～2.0 AE ago and are thus not related to the Vesterålen migmatites.     

Clinopyroxenes in Mesozoic pillow lavas from the French Alps: influence of cooling rate on compositional trends

The clinopyroxene compositions determined in four spilite samples from the French Alps show trends with high Al and Ti contents. Two samples from the center and outer zones of a single pillow have different Al/Ti ratios as well as different evolution trends in the CaFeMg system. Two samples collected from two different pillows in another outcrop also show different Al and Ti contents.These variations can be correlated with texture and grain size of the rocks and can be seen to follow trends predicted from recent cooling rate experiments. Analyses of relict clinopyroxenes from a glaucophane-lawsonite-bearing pillow show that these metastable clinopyroxenes can survive an episode of low-temperature, high-pressure metamorphism.     

The chronology of the Nakhla achondritic meteorite

A well-defined internal RbSr isochron has been determined for the unbrecciated Nakhla achondrite, defining an age T = (1.24 ± 0.01) AE and an initial ⁸⁷Sr/⁸⁶Sr ratio of 0.70254 ± 0.00003. The excellence of fit to the isochron shows that a very thorough Sr isotopic equilibration took place at this time between all phases of the meteorite. Recent petrological observations indicate that the age most probably represents the time of an igneous crystallization event on the parent body of Nakhla. The age shows that we may no longer assume that major chemical and physical differentiation processes took place only in early solar system history. The RbSr systematics are shown not to require late formation of the Nakhla parent body. Geochemical similarities between Nakhla and the earth are discussed.     

The relationship between major element chemistry and tectonic environment of basic and intermediate volcanic rocks

This work reports the preliminary results of a study of the relationship between tectonic environment and major element chemistry of volcanic rocks. Using a file of 8400 analyses of rocks of varying ages and geographic distribution it has been found that a simple ternary plot of MgOFeO(total)Al₂O₃ may be used to distinguish five tectonic environments: ocean ridge and floor, ocean island, continental, orogenic, and spreading center island. Subalkaline “basaltic-andesites” (51–56% SiO₂ calculated anhydrous) are used in this study. Alkaline rocks do not generally show the simple patterns of tholeiitic or subalkaline rocks. The ocean island suite is bimodal with one mode tholeiitic and the other alkaline. A bimodal age/chemistry relationship in the continental suite is tentatively related to different conditions during and after active breaking up of continents (rifting and drifting stages).Most of the Archean rocks in our file generally do not fall in the orogenic field. It is unlikely that modern island arc models can be generally applied to Archean greenstone belts. Locally, however, there may be Archean rocks analogous in some respect to modern calc-alkali rocks.     

SmNd isotopic systematics of Enderby Land granulites and evidence for the redistribution of Sm and Nd during metamorphism

Measurements of¹⁴³Nd/¹⁴⁴Nd and¹⁴⁷Sm/¹⁴⁴Nd are reported for whole rocks and mineral separates from granulites of the Napier Complex at Fyfe Hills. Charnockites, leuconorites and gabbros yield a whole rock SmNd isochron age of3060 ± 160m.y. and an initial¹⁴³Nd/¹⁴⁴Nd ratio of0.50776 ± 10 (?_Nd(3060m.y.) = ?2.0 ± 1.8). The negative ?_Nd value and the presence of geologically induced dispersion in the data suggest that the isochron age does not represent the time of primary crystallization of the complex but instead indicates a time of later redistribution of Sm and Nd and partial re-equilibration of¹⁴³Nd/¹⁴⁴Nd ratios. This probably occurred during the upper granulite facies metamorphism which has also been dated at～ 3100m.y. by RbSr and UPb zircon studies [1]. Coexisting clinopyroxene, apatite and total rock fractions in two adjacent samples define an approximately linear array corresponding to an age of 2300 ± 300 m.y. This array indicates that redistribution of Sm and Nd and re-equilibration of¹⁴³Nd/¹⁴⁴Nd ratios occurred on an intermineral scale during the upper amphibolite to lower granulite facies metamorphism at～ 2450m.y.Due to the resetting of the SmNd system on both whole rock and mineral scales, the primary crystallization age of the igneous protolith is not well constrained by the present data, although it is clearly3100m.y. If it is assumed that the complex was derived initially from a depleted mantle reservoir(?_Nd(T) ? 2), evolution of the negative ?_Nd value of ?2.0 with the observed Sm/Nd ratios requires a prehistory of～ 380m.y. This implies a primary age of～ 3480m.y. However, substantially older primary ages can be inferred if the source reservoirs had?_Nd(T) > 2 and/or substantial reductions in the Sm/Nd ratio occurred in whole rocks during the granulite facies metamorphism at 3100 m.y. Such an inferred reduction in the Sm/Nd ratio may have been the result of preferential loss of Sm relative to Nd, or introduction of a low Sm/Nd fluid with?_Nd ≥ 0 during granulite facies metamorphism.     

Rare earth element geochemistry of garnet lherzolite and megacrystalline nodules from minette of the Colorado plateau province

Mineral and whole-rock REE abundances in garnet lherzolite and megacrystalline nodules from The Thumb display broad correlations with major element compositions. Lherzolites with > 12 modal % clinopyroxene plus garnet (“high-CaAl lherzolites”) have relatively flat chondrite-normalized whole-rock REE patterns. Lherzolites poor in clinopyroxene and garnet (“low-CaAl lherzolites”) have lower HREE in clinopyroxenes and garnets and higher whole-rock LREE/HREE. It is concluded that the low-CaAl lherzolites may have undergone LREE metasomatism after depletion of the major element compositions by partial melting and that much of the garnet now present was originally dissolved in aluminous orthopyroxene. The high-CaAl lherzolites may be interpreted either as primordial mantle samples or as products of equilibration with very LREE-enriched liquids. The “megacrystalline” nodules are medium- to ultracoarse-grained intergrowths and megacrysts with mineral compositions similar to discrete nodule suites in kimberlites. The REE abundances of the megacrystalline minerals are consistent with an origin as cumulates from magma with extremely fractionated REE, similar to minette or kimberlite.The patterns of correlation of REE and major elements in this inclusion suite are similar to the patterns observed in the garnet lherzolite and discrete nodule suites of southern African kimberlites. Both of the subcontinental mantle provinces represented by these suites contain three distinct petrogenetic components: refractory garnet lherzolite enriched in LREE and depleted in HREE, fertile garnet lherzolite with generally chondritic REE abundances, and a suite of ultracoarse minerals precipitated from magma with extremely fractionated REE generally similar to the host magmas.     

Problems of two-pyroxene geothermometry

It is possible to formulate a two-pyroxene (MgFeCa) geothermometer using non-ideal but disordered site models for the pyroxenes. Experimental data on any two of the following three sets are needed: (a) the site occupancy data in orthopyroxene, (b) the site occupancy data in clinopyroxene (both sets of data with varying composition and temperature), and (c) synthetic phase equilibrium data on coexisting pyroxenes. A geothermometer based on site occupancy data is very sensitive to small variations of composition and may be useless for common petrological purposes but immensely useful for studying disequilibrium and kinetic history of the rocks.