Granites of the Batholith of Central Patagonia were emplaced into late Precambrian metamorphic basement rocks and Palaeozoic orthogneisses (here dated by Rb–Sr whole-rock errorchrons at 346 ±35 Ma and 267 ±27 Ma), and are in fault-contact with younger volcanic rocks, mostly andesites. Two main suites of granites are recognised: both are much younger than their previously-supposed Late Palaeozoic age. The Gastre Suite is composed predominantly of hornblende-biotite granodiorite and monzogranite, often slightly foliated, and has yielded a Rb-Sr whole-rock isochron age of 220 ±3 Ma. The Lipetrén Suite includes biotite monzogranite and leucogranite, grading into quartz-feldspar porphyries and felsites, and has been dated at 208 ±1 Ma. A minor granodioritic unit yields an apparently Middle Jurassic age of 172 ±15 Ma. Textural evidence and hornblende geobarometry confirm that these are high-level sub-volcanic intrusions. Metaluminous compositions are common in the Gastre Suite, but are subordinate to highly siliceous (>70% SiO2) and peraluminous varieties in the Lipetrén Suite. Despite this compositional bias, the granites are almost entirely calcalkaline and I-type, and have volcanic-arc rather than intraplate or collisional trace element characteristics. Initial 87Sr/86Sr ratios of 0·706 and –2·5 are also relatively primitive and are thought to indicate a juvenile crustal contribution (Nd “depleted mantle” model ages are less than 1,000 Ma).

The Triassic “Gondwana” magmatic episode is thus not an expression of Permo-Triassic collision of an allochthonous Patagonian terrane with the rest of southern S America, but may be related to the initial stages of supercontinent rifting.

The ore-element associations of granite-related ore deposits in the eastern Australian Palaeozoic fold belts can be related to the inferred relative oxidation state, halogen content and degree of fractional crystallisation within the associated granite suites. Sn mineralisation is associated with both S- and I-type granites that are reduced and have undergone fractional crystallisation. Cu and Au are associated with magnetite- and/or sphene-bearing, oxidised, intermediate I-type suites. Mo is associated with similar granites that are more fractionated and oxidised. W is associated with a variety of granite types and shows little dependence on inferred magma redox state. The observed ore deposit-granite type distribution in eastern Australia, and the behaviour of ore elements during fractionation, is consistent with models of ore element sequestering by sulphides and Fe-Ti phases (e.g. pyrrhotite, ilmenite, sphene, magnetite) whose stability is nominally fO2-dependent. Fractional crystallisation acts to amplify this process through the progressive removal of compatible elements and the concentration of incompatible elements into decreasing melt volumes. The halogen content is also important. S-type granites are poorer in Cl than I-types. Cl decreases and F increases in both S- and I-type granites with fractional crystallisation. Low Cl contents combined with low magma fO2 in themselves seem to provide an adequate explanation for the rarity of Mo, Cu, Pb and Zn type mineralisation with S-type granites. Although such properties of granite suites seem adequately to predict the associated ore-element assemblage to be expected in associated mineral deposits, additional factors determine whether or not there is associated economic mineralisation.

Size and composition (bulk metal ratios) of magmatic hydrothermal mineral deposits are affected by a number of chemical and physical processes including the nature of the source region and mode of emplacement. At shallow levels, rising plumes of vapour bubbles + melt, and the advection of water through interconnected vapour bubbles, allows access of the magmatic aqueous phase to the upper reaches of a magma chamber. These processes are operative at shallow levels where low water solubility and high molar volume for water make these processes more efficient.

Partitioning experiments suggest that oxygen fugacity-dependent crystal/melt partitioning of ore metals leads to different efficiencies of removal of Cu, W, and Mo from silicate melts into ore-forming aqueous fluids. For example, the Mo/W ratio in magmatic hydrothermal deposits should increase as the oxygen fugacity of the magma increases. Further, Cu should behave as a crystal-compatible element in H2O-undersaturated, sulfide-saturated felsic magmas with fO2 NNO + 1 due to the strong partitioning of Cu from the melt into pyrrhotite.

Cycling of oxidised, hydrated, sulfidised and Cl-enriched oceanic crust into mantle can give rise to magmas that contain S but are oxidised (≥NNO). The combination of high oxidation state, relatively hydrous but shallow conditions and a high Cl/H2O ratio leads to saturation with respect to H2O early during crystallisation, and loss of a large proportion of magmatic Cu to the aqueous phase. Ores formed from these oxidised magmas also possess high Mo/W ratios due to the effect of oxygen fugacity on the sequestering of Mo vs W.

In less oxidised magmas, Cu and Mo are partitioned into sulfides and Ti-bearing phases, respectively, resulting in lower efficiencies of removal of Cu and Mo from melts into aqueous fluids. Further, the partitioning of W into crystallising phases is reduced, producing a more efficient removal of W into ore-forming fluids. This ultimately leads to mineral deposits with higher W/(Mo + Cu) ratios relative to deposits associated with more oxidised systems. Silicic, high-F magmas with fO2 = NNO can be found in tensional environments (e.g. rocks associated with the Climax-type deposits of the Colorado Mineral Belt). High HF/H2O activity ratios in the source regions yield melts that evolve an aqueous phase late during crystallisation, leading to relatively low ratios of compatible/incompatible elements in the melt at H2O saturation.

Silicic and minor intermediate and mafic pyroclastics, lavas, and dykes occupy a NW-trending zone through the Whitsunday, Cumberland and Northumberland Island groups, and locally areas on the adjacent mainland, over a distance of more than 300 km along the central Queensland coast. K-Ar and Rb-Sr data indicate an age range of 95–132 Ma, with the main activity approximately between 105–120 Ma; there is, however, evidence for easterly increasing ages. Comagmatic granites, some clearly intrusive into the volcanics, occur together with two localised areas of Triassic potassic granites (229 Ma), that form the immediate basement.

The volcanics are dominantly rhyolitic to dacitic lithic ignimbrites, with intercalated surge and bedded tuffs, accretionary lapilli tuffs, and lag deposits. Associated rock types include isolated rhyolitic and dacitic domes, and volumetrically minor andesite and rare basalt flows. The sequence is cut by abundant dykes, especially in the northern region and adjacent mainland, ranging from dolerite through andesite, dacite and rhyolite. Dyke orientations show maxima between NW-NNE. Isotope data, similarities in petrography and mineralogy, and alteration patterns all suggest dyke intrusion to be broadly contemporaneous with volcanism. The thickness of the volcanics is unconstrained, although in the Whitsunday area, minimum thicknesses of >1 km are inferred. Eruptive centres are believed to occur throughout the region, and include at least two areas of caldera-style collapse. The sequences are thus considered as predominantly intracaldera.

The phenocryst mineralogy is similar to modern “orogenic” volcanics. Phases include plagioclase, augite, hypersthene (uralitised), magnetite, ilmenite, with less common hornblende, and even rarer quartz, sanidine, and biotite. Fe-enriched compositions only develop in some high-silica rhyolites. The granites range from quartz diorite to granite s.s., and some contain spectacular concentrations of partially disaggregated dioritic inclusions.

Chemically, the suite ranges continuously from basalt to high-silica rhyolite, with calc-alkali to high-K affinities, and geochemical signatures similar to modern subduction-related magmas. Only the high-silica rhyolites and granites exhibit evidence of extensive fractional crystallisation (e.g. pronounced Eu anomalies). Variation within the suite can only satisfactorily be modelled in terms of two component mixing, with superimposed crystal fractionation. Nd and Sr isotope compositions are relatively coherent, with εNd + 2·2 to +7·3, and ISr (calculated at 110 and 115 Ma) 0·7031-0·7044. These are relatively primitive, and imply mantle and/or newly accreted crustal magma sources.

The two end-members proposed are within-plate tholeiitic melt, and ?low-silica rhyolitic melts generated by partial fusion of Permian (to ?Carboniferous) arc and arc basement. The arc-like geochemistry is thus considered to be source inherited. The tectonic setting for Cretaceous volcanism is correlated with updoming and basin rifting during the early stages of continental breakup, culminating in the opening of the Tasman Basin. Cretaceous volcanism is also recognised in the Maryborough Basin (S Queensland), the Lord Howe Rise, and New Caledonia, indicating the regional extent of volcanism associated with the complex breakup of the eastern Australasian continent margin.

Geochemical and isotopic data obtained from three bimodal basalt-rhyolite fields in west-central Arizona point to significant variation in the sources of the siliceous products despite petrological similarities between them. The studied fields, which straddle the boundary between the Basin and Range and Colorado Plateau physiographic provinces, include Castaneda Hills (CH) in the Basin and Range, Kaiser Spring (KS) in the Transition Zone, and Mount Floyd (MF) in the Colorado Plateau. Two types of rhyolite (high-silica [HSR] and low-silica [LSR]) occur as lavas in the KS and MF fields, whereas all analysed CH samples are HSR. These lavas generally post-date low-angle extension of the Basin and Range crust and become generally younger towards the Colorado Plateau. Our isotopic data illustrate that the CH rhyolites require a source with comparatively radiogenic Sr, that most rhyolites require a source with comparatively unradiogenic 206Pb/204Pb, and that the MF-HSR require a source with comparatively radiogenic Nd and 206Pb/204Pb. The isotopic data clearly indicate that the western Arizona rhyolites contain a large crustal component. Importantly, the Pb isotopic compositions of the rhyolites illustrate the transition between two crustal provinces identified by other workers on the basis of Nd and Pb isotopic studies of Proterozoic granites. The Sr, Nd, and Pb isotopic compositions of our rhyolites can be modelled by mixing a basaltic component with three hypothetical crustal end-members which themselves may be formed by mixing two crustal components. One crustal component, which is characterised by unradiogenic Pb and Sr, may be similar to the mafic gneiss xenoliths of Tule Tank on the Colorado Plateau. The second crustal component, which has more radiogenic Sr and Pb, may be similar to the Proterozoic Fenner Gneiss of southeastern California. This conceptual model indicates an apparent increase in the contribution of the depleted Tule Tank source from the Basin and Range onto the Colorado Plateau and accommodates the notion that there is an underlying consistency to the petrogeneses of the western Arizona rhyolites as suggested by their chemistry and mineralogy. Although the CH and KS lavas require only one crustal end-member, the MF-HSR and LSR require two isotopically distinct sources. This difference may be attributed to the homogenisation of diverse crustal lithologies by ductile stretching during crustal extension.

One of the most outstanding apparent examples in N America of a forcibly emplaced pluton is the Papoose Flat Pluton of eastern California. Sideways expansion of this granitic pluton, during emplacement into a series of Cambrian shelf strata, has been regarded by early workers as resulting in the observed intense crystal plastic deformation of the pluton's mylonitic border facies and surrounding country rocks. This deformation is evidenced by up to 90% thinning of individual stratigraphic layers within the pluton's metamorphic aureole, although such intense penetrative deformation of the country rocks is not observed outside the aureole.

Previously published quartz c-axis fabrics associated with this deformation (and presented on projection planes oriented perpendicular to lineation) were interpreted as being symmetrical with respect to foliation and lineation, implying almost coaxial deformation histories. Such fabrics could be interpreted as indicating that the pluton evolved by “ballooning” as a result of new magma being intruded into its core during emplacement. However, a major problem with applying the strict ballooning model to the Papoose Flat Pluton is that while oblate strains would be expected to develop in association with a ballooning mechanism, the mylonitic rocks of this elongate WNW-ESE-trending pluton and its aureole are characterised by both a strongly developed foliation, which is concordant with the pluton's margin, and an intense, NW-SE trending, shallow plunging stretching lineation.

Previously published fabrics from the Papoose Flat Pluton and its metamorphic aureole have been rotated on to a projection plane oriented parallel to lineation and perpendicular to foliation. Examination of the fabrics in this projection plane has revealed that they are in fact dominantly asymmetric, and that a constant sense of asymmetry is detected across the pluton, suggesting a consistent (top-to-the-SE) shear-sense. This new interpretation is strongly supported by microstructural and petrofabric analysis of additional L-S tectonites collected, during recent fieldwork, from both the aureole and quartz veins within the pluton's gneissic border facies. Thus mylonite formation around the Papoose Flat Pluton could have involved large-scale consistently oriented translation and associated shearing, rather than passive “blister-like” coaxial deformation associated with pluton ballooning. It should be noted that mylonitic deformation is restricted to the western half of the pluton, features indicative of a more “permitted” emplacement mechanism being found in the eastern portion of the pluton.

The detected top-to-the-SE shear-sense could be interpreted as indicating that the granitic material forming the western part of the pluton was forcibly intruded in a northwestward direction from the pluton source as a nearly solidified wedge beneath a static cover of sedimentary rocks. Alternatively, the detected shear sense could also be interpreted as indicating SE-directed thrusting of the cover rocks over the underlying pluton, the western margin of the pluton suffering intense mylonitic deformation, while the eastern margin was located in a “stress-shadow” region. If this alternative interpretation is correct, then the deformation temperatures indicated by the pattern of quartz c-axis fabrics dictate that thrusting must either be synchronous with pluton emplacement, or at least have commenced during the early stages of pluton cooling.

A review of granite emplacement mechanisms in transcurrent, extensional and contractional (thrust sense) shear zones reveals that in all three tectonic settings the plutons have been constructed by multiple granite sheeting parallel to the shear zone walls and deformation fabrics. The sheets and plutons are non-Andersonian in type and were emplaced obliquely to the principal stress directions. Their shape and orientation is more likely to reflect the exploitation of faults and shear zones which were active during emplacement. Sheeting (dyking) is therefore also likely to be the mechanism of ascent along fault zones in the crust.

The Great Tonalite Sill (GTS) of southeastern Alaska and British Columbia (Brew & Ford 1981; Himmelberg et al. 1991) is one of the most remarkable intrusive bodies in the world: it extends for more than 800 km along strike and yet is only some 25 km or less in width. It consists of a belt of broadly tonalitic sheet-like plutons striking NW–SE and dipping steeply NE, and has been dated between 55 Ma and 81 Ma (J. L. Wooden, written communication to D. A. Brew, April 1990) (late Cretaceous to early Tertiary). The sill (it is steeply inclined and rather more like a “dyke”) is emplaced along the extreme western margin of the Coast Plutonic and Metamorphic Complex (CPMC), the high grade core of the Western Cordillera. The CPMC forms the western part of a group of tectonostratigraphic terranes including Stikine and Cache Creek, collectively known as the Intermontane Superterrane (Rubin et al. 1990). To the W of the GTS, rocks of the Insular Superterrane, including the Alexander and Wrangellia terranes and the Gravina belt, form generally lower metamorphic grade assemblages. The boundary between these two superterranes is obscure but it may lie close to, or be coincident with, the trace of the GTS.

The thermodynamic relations embodied in the Quasicrystalline Model of Burnham and Nekvasil (1986), as recently extended by the author, have been used to quantitatively assess the feldspar-quartz liquidus relations in two I-type (Jindabyne and Moruya) and two S-type (Bullenbalong and Dalgety) suites of Australian granites, using analytical data provided by B. W. Chappell and co-workers. Among the more notable results obtained from these calculations at a constant pressure of 5·0kbar and = 0·30 (≍2·8 wt% H2O), for purposes of comparison, are that: (1) felsic melts of remarkably uniform, but distinctive composition can be extracted from each suite, leaving solid residues in amounts up to 65 mol%; (2) all melts from both S-type suites have two feldspars plus quartz on their liquidii, whereas both I-type suites have only plagioclase plus quartz on their liquidii; (3) the total solid residue ranges from 27-63% in the Jindabyne suite, from 15–62% in the Moruya suite, from 30–65% in the Bullenbalong suite, and from 27–65% in the Dalgety suite; (4) liquidus temperatures of the S-type Bullenbalong and Dalgety melts are similar (856° and 860°C), reflecting similar feldspar compositions of An53, Or75 and An60, Or77, respectively; (5) liquidus temperatures of the I-type Jindabyne and Moruya melts, however, are distinctly different (950° and 894°C), reflecting correspondingly different plagioclase compositions of An80 and An52; (6) the calculated liquidus plagioclase composition throughout a given suite is very uniform (±1%) and amounts to as much as 46% of the total rock; and (7) these calculated liquidus and residual plagioclase compositions are also the same, within the uncertainty of measurement, as those of the plagioclase crystal-cores determined optically by A. J. R. White. The only plausible explanation for this remarkable consanguinity in plagioclase liquidus, residue, and crystal-core compositions, hence liquidus temperatures, is that the bulk of the residue is restite, in accordance with the model of White and Chappell (1977). This explanation is corroborated by the very systematic variations in the amounts of individual restite minerals with respect to total restite contents. Accordingly, those members of each suite that contain more than 60% total restite probably closely represent the bulk composition of the source rock, which is dioritic or andesitic for the Jindabyne suite, tonalitic or dacitic for the Moruya suite, pelitic metagreywacke for the Bullenbalong suite, and feldspathic metagreywacke for the Dalgety suite. As a corollary, those members with less than 60% restite must have undergone melt-restite segregation (unmixing), probably during ascent and emplacement.

Understanding the controls on the evolution of natural feldspars is greatly assisted by coupling experimental determinations of feldspar/melt equilibria with thermodynamic modelling and the calculation of crystallisation paths. Such a combined approach permits the evaluation of the influence of intensive and extensive variables on feldspar compositions. Feldspar compositional paths are influenced only to a minor degree by pressure. The presence of H2O or other melt component incompatible in feldspar has a more major effect, not only in increasing the temperature interval over which feldspar crystallises, but also in decreasing the amount of Ab enrichment of the feldspar which occurs during crystallisation. The amount and behaviour of H2O in the magma has a pronounced influence on feldspar compositions when two feldspars are stable. Under conditions of high bulk H2O content or constant activity of H2O, plagioclase and alkali feldspar compositions evolve by increasing Ab content similar to the behaviour manifested in the simple binary systems. If the bulk H2O content is low, however, and the H2O content increases during crystallisation, plagioclase evolves by increasing Ab content until alkali feldspar is stabilised. From that point on, plagioclase compositions change mainly by decreasing Or content, while alkali feldspar evolves mainly by increasing Or content.

Liquidus phase relationships at H2O-saturated and -undersaturated conditions and 2 kbar in the systems Qz-Or-Ab (SiO2-KAlSi3O8-NaAlSi3O8), Qz-Or-Ab-Al2O3, and subsystems are compared and discussed. In the peraluminous systems (i.e. when melts are saturated with respect to mullite) the liquidus temperatures are lowered by 40-55°C for compositions in the quartz primary field and by 15-25°C for cotectic compositions. The composition of the Qz-Ab eutectic and of the minimum are slightly shifted towards more Qz-rich compositions (minimum composition at P(H2O) = 2 kbar in the system Qz-Or-Ab-A12O3, saturated with respect to mullite: Qz40Or23Ab37). In melts saturated with sillimanite or mullite, the effect of high Al content may be lower for the Qz-Or than for the Qz-Ab eutectic.

The depression of the liquidus temperatures may be partly related to the higher H2O solubility in melts saturated with respect to mullite. The solubility of H2O in a melt with a composition of Qz28Or34Ab38 at 2 kbar and 800°C is 5·77 ± 0·15 wt% H2O and 6·36 ± 0·30 wt% H2O in a melt with the same Qz/Or/Ab proportions but saturated with respect to mullite.

The effect of high Al contents on the Mg and Fe contents of Ca-free granite melts was investigated at 775°C-3 kbar (melts coexisting with phlogopite), and at 820°C-2 kbar (melts coexisting with biotite and spinel), under NNO buffer conditions. Less than 0·15 wt% MgO is incorporated in subaluminous melts coexisting with phlogopite, whereas peraluminous melts (2·9 wt% normative corundum) contain 0·6–0·7 wt% MgO. A similar behaviour of the MgO content is observed for melts coexisting with biotite. In contrast, no significant effect of high Al contents on the FeO content of melt coexisting with biotite was observed. This suggests that the Fe/Mg ratio may be significantly lower in peraluminous than in subaluminous granitic melts.

Melting experiments with plagioclases were performed in the systems Ab-An, Ab-An-H2O, Qz-Ab-An-H2O, and Qz-Or-Ab-An-H2O-CO2. The experimental products were analysed by electron microprobe, and the kinetics of the reactions were studied qualitatively.

Melting of the plagioclase in the system Ab-An (P = 1 atm, T = 1420°C) is very fast in the first minutes but becomes slower with increasing run duration and is incomplete even after 1000 hours. The Ab/An fractionation between melt and residual plagioclase is similar to that described by Bowen (1913).

Melting kinetics of plagioclase in the system Ab-An-H2O ( = 5 kbar, T = 1000°C) is controlled by the diffusion of water into the plagioclase structure. Melting is especially fast parallel to the a-axis. The experimental products show separation of melts and crystals.

In the tonalite system Qz-Ab-An-H2O, equilibrium melting could be observed down to 830°C ( = 2 kbar) but not a lower temperatures. The kinetics of the reaction is enhanced by deficiency or excess of alumina in the aluminosilicate melt surrounding the plagioclase crystals. The fractionation of Ab and An between melt and plagioclase crystals is more pronounced in the presence of quartz than in the Ab-An-H2O system. The ratio An/An + Ab is approximately 0·35 in the melt and 0·85 in the coexisting plagioclase T = 880°C).

In the haplogranodiorite system Qz-Or-Ab-An-H2O–CO2, melting reactions were performed at P = 0·5 kbar, T = 880°C, and of approximately 0·5. It is assumed that near equilibrium compositions of melt and coexisting residual plagioclase could be obtained in long duration runs (run time = 60 days). The distribution of Ab and An between melt and minerals is similar to that observed in the tonalite system. The partial melt coexisting with an An-rich plagioclase and Or-rich K-feldspar is relatively poor in An.

A two-parameter Margules method has been developed for modelling the thermodynamic mixing properties and hypersolidus phase relations of albite-water (ab-w) melts. The method is based largely on phase-equilibrium and calorimetric data that are either currently available or readily obtainable; P-V-T data are not required.

The new modelling method has been applied to calculate: (1) thermodynamic mixing properties for ab-w melts at 2-5 kbar, 815°C; and (2) hypersolidus phase relations for the ab-w system at 2·5 kbar. Results are uniformly positive—activity-composition relations are similar to those predicted by equations based on P-V-T data, excess enthalpies are in good agreement with data obtained from calorimetry, and calculated phase relations are fully compatible with phase-equilibrium data for ab-w melts acquired at 2·5 kbar.

It is concluded that the Margules method for thermodynamic modelling of ab-w melts is a practical alternative to the P-V-T modelling method.

Rare garnet phenocrysts and garnet-bearing xenoliths occur in high-silica, metaluminous to peraluminous andesites and dacites (and their high-level intrusive quartz diorite equivalents) from a Miocene calc-alkaline province in Northland, New Zealand. These garnets are among the most Ca-rich (17–28 mol% grossular) garnets of igneous origin so far recorded in calc-alkaline suite rocks. Associated minerals are dominant hornblende and plagioclase and minor augite, occurring as phenocrysts in xenoliths and as inclusions in the garnet. This mineralogy points to the I-type character of the garnet-bearing host magma compositions, and contrasts this garnet occurrence with the more frequently recorded grossular-poor (3–10 mol%) garnets with hypersthene, plagioclase, biotite and cordierite, found in S-type volcanic and intrusive host rocks.

Detailed experimental work on a glass prepared from one of the garnet-bearing dacites closely constrains the conditions under which the natural phenocryst and xenolith mineral assemblages formed. This work was conducted over a pressure-temperature range of 8–20 kbar, 800–1050°C with 3–10 wt% of added H2O, defining overall phase relationships for these conditions. Importantly, amphibole only appears at temperatures of 900°C or less and clinopyroxene at >900°C (with 3wt% H2O). Orthopyroxene occurs with garnet at lower pressure (∼15 kbar with 3wt% H2O; ∼>10kbar with 5wt% H2O). Absence of orthopyroxene from the natural garnet-bearing assemblages indicates pressures above these limits during crystallisation. Plagioclase is markedly suppressed (with respect to temperature) with increasing H2O content, and for pressures of 10–15 kbar, the maximum H2O content possible in the magma with retention of clinopyroxene and plagioclase together (as evident in xenoliths) is 5–6 wt%. Finally, the lack of quartz in any of the xenoliths suggests magma H2O content higher than 3% (where quartz appears with amphibole at 900°C), since the quartz liquidus temperature decreases with increasing H2O content, and with decreasing pressure. In experiments with 5wt% H2O, a quartz-free field of crystallisation of garnet-clinopyroxene-amphibole-plagioclase occurs between 10 and 15 kbar and temperatures between 850 and 900°C. In addition, detailed experimentally-determined garnet compositional trends, together with ferromagnesian mineral compositional data for specific experiments with 5 wt% H2O added and run at 10-13 kbar and ∼900°C, suggest that the natural assemblages formed at these conditions. This implies that the parental dacitic magma must have been derived at mantle depths (the Northland crust is ∼25 km thick), and any basaltic or basaltic andesite precursor must have contained ∼2–3 wt% H2O.

The unique nature of the Northland volcanics and high-level intrusives, preserving evidence of relatively grossular-rich garnet fractionation in the high-pressure crystallisation history of an originally mantle-derived magma, is attributed to a combination of unusually hydrous conditions in the source region, complex tectonic history involving obduction and subduction, possible incorporation of crustal slivers in a mantle-crust interaction zone, and relatively thin (∼25 km) crust.

Temporal and spatial variations in the Nd isotopic compositions of Tertiary caldera-forming rhyolite tuffs, and Cretaceous and Tertiary granites of the western U.S.A. are used as a basis for a model that accounts for the observed proportions of crustal versus mantle contributions to silicic magmas in terms of two parameters: the ambient crustal temperature and the rate of supply of basaltic magma from the mantle. The crustal contribution to silicic igneous rocks is measured in terms of the Neodymium Crustal Index (NCI). The relationships between crustal temperature, basalt supply and NCI are quantified using a model of a magma chamber undergoing continuous recharge, wall-rock assimilation and fractional crystallisation. From the model, a critical value of the ratio of basalt recharge-to-assimilation, (r/a)c, is deduced, which increases with decreasing crustal temperature. The r/a value must exceed (r/a)c to allow the volume of differentiated magma to increase, a prerequisite for developing large volumes of silicic magma. Strongly peraluminous (or S-type) magmas (NCI = 0·8–1), form under conditions of high crustal temperature and low basalt supply. Metaluminous or I-type granites form over a wide range of conditions (NCI = 0·1–1), generally where basalt supply is substantial. In individual long-lived volcanic centres, the large-volume high-silica ignimbrites are associated with the highest r/a and lowest NCI values, indicating that these magmas are typically differentiates of mantle-derived basaltic parents.

In situ, microscale, U-Pb isotopic analyses of zircon using the SHRIMP ion microprobe demonstrate both the potential and the limitations of zircon U-Pb geochronology. Most zircons, whether from igneous or metamorphic rocks, need to be considered as mixed isotopic systems. In simple, young igneous rocks the mixing is principally between isotopically disturbed and undisturbed zircon. In polymetamorphic rocks, several generations of zircon growth can coexist, each with a different pattern of discordance. A similar situation exists for igneous rocks rich in inherited zircon, as these contain both melt-precipitated zircon and inherited components of several different ages. Microscale analysis by ion probe makes it possible to sample the record of provenance, age and metamorphic history commonly preserved within a single zircon population. It also indicates how the interpretation of conventionallymeasured bulk zircon isotopic compositions might be improved.

Using zircons taken from two granite plutons, Strontian (Caledonian, northwestern Scotland) and Kameruka (Bega Batholith, southeastern Australia), this study presents observations that have a bearing on refractory zircons as provenance indicators. Two broad textural types of refractory zircon were identified: (1) those which show simple two-stage growth histories; and (2) those which have apparently undergone repeated periods of growth, resorption, mechanical abrasion, fracturing and fracture-healing. SHRIMP U-Pb ages obtained from the Kameruka zircons indicate that the cores are the textural manifestation of inheritance. The shapes of refractory cores are not unambiguously indicative of their ultimate origin, since they may also be modified by processes that occur before and after incorporation into the magma. The cores within the two populations show a great diversity in types and styles of zoning, and in composition, implying that they have not chemically equilibrated internally, or externally with their host melt.

Using the ion microprobe SHRIMP we have analysed zircons from the Ben Vuirich, Glen Kyllachy, Inchbae and Vagastie Bridge granites from the Scottish Caledonides, in an attempt to resolve the ages of inherited zircons shown to be present in these granites by previous conventional multigrain analyses. Middle Proterozoic age components were found in inherited zircons from all four granites. Late Proterozoic (900–1,100 Ma) components have been identified in zircons from the Glen Kyllachy and Ben Vuirich granites in the Grampian Highlands. A Late Archaean age has only been detected in one zircon from the Glen Kyllachy granite. The distribution of inherited components in the granite zircon populations could reflect fundamental divisions in the age composition of granite source rocks; however, detailed assessment of this possibility must await further ion microprobe analyses on zircons from many more granites.

SHRIMP isotopic and U, Th and Pb analyses were made on successive shells of zoned zircon surrounding inherited cores from the Glen Kyllachy granite to monitor chemical changes during magmatic zircon growth. Results show that zircon shells have characteristic but significantly different Th, U and Pb concentrations. Magmatic zircon from the Vagastie Bridge granite also forms as clearly defined oscillatory zoned shells around unzoned nuclei of inherited zircon. However, the distinction between magmatic and inherited zircon in zircons from the Inchbae granite is less clear. Zircons from the Ben Vuirich granite occur as euhedral, magmatic zircons, or as rounded, subhedral, inherited zircon grains. A SHRIMP age of 597 ± 11 (2σ) Ma for euhedral magmatic zircon from this granite is identical, within the uncertainty, to the conventional multigrain zircon age of 590 ± 2 (2σ) Ma reported by Rogers et al. (1989) and confirms the conclusions of those authors that sedimentation of the Dalradian sequence took place in the Precambrian.