I-type granites are produced by partial melting of older igneous rocks that are metaluminous and hence have not undergone any significant amount of chemical weathering. In the Lachlan Fold Belt of southeastern Australia and the Caledonian Fold Belt of Britain and Ireland there was a major magmatic event close to 400 Ma ago involving a massive introduction of heat into the crust. In both areas, that Caledonian-age event produced large volumes of I-type granite and related volcanic rocks. Granites of these two areas are not identical in character but they do show many similarities and are markedly different from many of the granites found in Mesozoic and younger fold belts. These younger, dominantly tonalitic, granites have compositions similar to those of the more felsic volcanic rocks forming at the present time above subduction zones. The Palaeozoic granites show little evidence of such a direct relationship to subduction. Within both the Caledonian and Lachlan belts there are some granites with a composition close to the younger tonalites. A particularly interesting case is that of the Tuross Head Tonalite of the Lachlan Fold Belt, which can be shown to have formed from slightly older source rocks by a process that we refer to as remagmatisation which has caused no significant change in composition. Since remagmatisation has reproduced the former source composition in the younger rocks, the wrong inference would result from the use of that composition to deduce the tectonic conditions at the time of formation of the tonalite. Granites, particularly the more mafic ones, will generally have compositions reflecting the compositions of their source rocks, and attempts to use granite compositions to reconstruct the tectonic environment at the time of formation of the granite may be looking instead at an older event. This is probably also the case for some andesites formed at continental margins.

Several arguments can be presented in favour of a general model for the production of I-type granite sources by underplating the crust, so that the source rocks are infracrustal. Such sources may contain a component of subducted sediments with the consequence that some of the compositional characteristics of sedimentary rocks may be present in I-type source rocks and in the granites derived from them. The small bodies of mafic granite and gabbro associated with island arc volcanism have an origin that can be related to the partial melting of subducted oceanic crust or of mantle material overlying such slabs and can be referred to as M-type. These rocks have compositions indistinguishable from those of the related volcanic rocks, except for a small component of cumulative material. The tonalitic I-type granites characteristic of the Cordillera are probably derived from such M-type rocks of basaltic to andesitic composition, which had been underplated beneath the crust. Some of the more mafic tonalites of the Caledonian-age fold belts may also have had a similar origin. More commonly, however, the plutonic rocks of the older belts are granodioritic and these probably represent the products of partial melting of older tonalitic I-type source rocks in the deep crust, these having compositions and origins analogous to the tonalites of the Cordillera. In this way, multiple episodes of partial melting, accompanied by fractionation of the magmas, can produce quite felsic rocks from original source rocks in the mantle or mantle wedge. These are essential processes in the evolution of the crust, since the first stages in this process produce new crust and the later magmatic events redistribute this material vertically without the addition of significant amounts of new crust.

Ring complex granites of the Peruvian Batholith are tabular bodies with flat roofs emplaced by cauldron subsidence. Marginal precursory ring dykes extend upwards above roof level and a typical intrusion is “H”-shaped in cross-section. Advance of magma by repeated subsidence would give a ladder-shaped profile for such intrusions above the brittle-ductile transition. Close relationships exist between intrusion geometry, emplacement process and petrogenetic evolution. Initially a granodioritic magma chamber lay beneath the present erosion level, trapping a rising mass of dioritic magma. Expansion of granodioritic liquid resulted in the injection of microgranite and tuffisite cone sheets accompanied by roof uplift within a ring fault. Next, during subsidence within the ring fault, liquids from deeper levels in the underlying chamber rose by stoping along the outer margin of the fault to form a ring dyke. Prior to intrusion, this liquid was vertically zoned from rhyodacite downwards to diorite and these liquids became partially mixed during emplacement. Finally, granodioritic magma rose to the present level by subsidence of a roof slab bounded by the ring fault. The precursory ring structures preserve evidence of significant but transient events in the underlying chamber. Liquid differentiation may be significant in the evolution of many large plutons.

The Peninsular Ranges Batholith of southern and Baja California is the largest segment of a Cretaceous magmatic arc that was once continuous from northern California to southern Baja California. In this batholith, the emplacement of igneous rocks took place during a single sequence of magmatic activity, unlike many of the other components of the Cordilleran batholiths which formed during successive separate magmatic episodes. Detailed radiometric dating has shown that it is a composite of two batholiths. A western batholith, which was more heterogeneous in composition, formed as a static magmatic arc between 140 and 105 Ma and was intrusive in part into related volcanic rocks. The eastern batholith formed as a laterally transgressing arc which moved away from those older rocks between 105 and 80 Ma, intruding metasedimentary rocks. Rocks of the batholith range from undersaturated gabbros through to felsic granites, but tonalite is the most abundant rock throughout. Perhaps better than elsewhere in the Cordillera, the batholith shows beautifully developed asymmetries in chemical and isotopic properties. The main gradients in chemical composition from W to E are found among the trace elements, with Ba, Sr, Nb and the light rare earth elements increasing by more than a factor of two, and P, Rb, Pb, Th, Zn and Ga showing smaller increases. Mg and the transition metals decrease strongly towards the E, with Sc, V and Cu falling to less than half of their value in the most westerly rocks. Oxygen becomes very systematically more enriched in18O from W to E and the Sr, Nd and Pb isotopic systems change progressively from mantle values in the W to a more evolved character on the eastern side of the batholith. In detail the petrogenesis of the Peninsular Ranges Batholith is not completely understood, but many general aspects of the origin are clear. The exposed rocks, particularly in the western batholith, closely resemble those of present day island arcs, although the most typical and average tonalitic composition is distinctly more felsic than the mean quartz diorite or mafic andesite composition of arcs. Chemical and isotopic properties of the western part of the batholith indicate that it formed as the root of a primitive island arc on oceanic lithosphere at a convergent plate margin. Further E, the plutonic rocks appear to have been derived by partial melting from deeper sources of broadly basaltic composition at subcrustal levels. The compositional systematics of the batholith do not reflect a simple mixing of various end-members but are a reflection of the differing character of the source regions laterally and vertically away from the pre-Cretaceous continental margin.

The genesis of subduction-related magmas in the Andean region of South America and the Antarctic Peninsula is considered in relation to the Palaeozoic to Cenozoic granitoids belts which are thought to parallel palaeo-coastlines. Their Sr-Nd isotope systematics show a wide range of initial compositions (87Sr/86Sr0 0·7038 to >0·710; εNd, +4 to –10) requiring material input from both depleted mantle and continental crust. In local transects there are consistent trends with time of emplacement, from enriched (crustal) to depleted (mantle) sources, regardless of the sense of migration of magmatism (towards or away from the continent). These trends represent mixing between mantle-derived material and anatectic melts of the lower crust: in each case the crustal end-member reflects the age and isotopic composition of the local deep crustal basement (Precambrian in the easternmost Andes, Palaeozoic in the W and in the Antarctic Peninsula). The depleted end-member could be derived by melting within the subducted oceanic crust, the overlying mantle or previously crystallised mafic underplating. One of the most important factors controlling the mixing process is the angle of subduction, resulting in magma generation under variable tectonic conditions.

The pursuit of a comprehensive theory for the origin and evolution of granitoids is hindered by our incomplete understanding of the nature of the source and the mechanisms by which the magma is segregated and transported. This paper is a collection of three largely independent and necessarily incomplete perspectives on these outstanding issues. Lower to mid-crustal regions, which contain the principal source material for granitoid magmas, are highly heterogeneous. Consideration of available transfer mechanisms suggests that (1) this heterogeneity survives all foreseeable lower crustal processes; (2) closure is on very different scales for different chemical systems (e.g. Pb, Nd, Sr and O isotopes); in almost all cases, however, closure scale is much smaller than the scale of magma extraction zones for plutons; and (3) pluton-wide homogenisation of magmas by diffusion is precluded by low diffusivities in felsic melts. Thus, granitoid magmas begin life as aggregates of small, isolated chemical domains; homogenisation occurs only through (and on the scale of) effective stirring by convection. Because of variability in local conditions as well as in bulk composition, crustal regions undergoing anatexis must be patchworks with variable melt fractions and melt compositions. The way in which magma is extracted from and coalesces with this patchwork exerts a critical influence on the nature of granitoid magmas. Decoupling and unusual coupling of compositional parameters and isotopic heterogeneity within plutons are to be expected in crust-derived granitoids and do not require contamination. Granites image their sources, but these sources are ill-defined and do not correspond to simple, easily-recognised materials. Extent and patterns of heterogeneity remaining in crystallised plutons may be effective indicators of the ascent process.

The efforts of materials scientists in characterising the nature and evolution of solid-phase interconnectivity in partially-molten materials may offer some insights into crustal magmatic processes. In particular, the rheological properties of partially-molten crustal rocks are probably strongly affected by the contiguity of the solid grains in the system (i.e. the fraction of their surface area that is shared with other grains). Theory and experimental data for simple alloy systems reveal that contiguity depends principally upon melt fraction and upon the characteristic wetting angle (θ) of the system. Measured θ's in granitoids (∼50° on average) imply contiguities as high as ∼0·2 for melt fractions of 0·5 or greater. This value in turn suggests that, at least under static conditions, a continuous skeleton of solid grains is maintained to quite high degrees of melting in the crust. Consequently, regions consisting of 50% or more of melt can, in principle, maintain not only high yield strength, but also high viscosity (provided the strain rate is sufficiently low to avoid disrupting contiguity).

Despite the fact that on some time scale the continuous solid skeleton of a partially-molten region resists deformation, it is itself subject to textural evolution that could lead to the upward migration of melt. Occasional detachment of grains from the skeleton and subsequent “microsettling” within the partially-molten column may lead eventually to compaction of the solid (without plastic deformation) and net upward displacement of melt.

Proposed granite transport mechanisms are discussed, although several are viewed as having historical interest only. In the absence of tectonic transport, diapirism appears to be the most compelling of these processes. However, considerable diversity exists in the literature regarding a pivotal requirement for this mechanism. Structural studies have tended to conclude that the granite diapir must be highly crystallised in order to ascend, whereas results of physical modelling yield contradictory results. For ascent to occur in these models, the magmas must be sufficiently fluid to allow convective circulation. Indeed, heat loss associated with diapirism is so efficient as to be a significant restriction on overall ascent. The resolution of these contrasting views appears to be that they reflect different phases of the ascent/emplacement continuum. Understanding the emplacement history of a southeastern Australian pluton allows assessment, via the diapir model, of the flow properties of the rock within the deformation aureole. Results suggest rock viscosities about an order of magnitude lower than those predicted by laboratory experiments, perhaps reflecting difficulties in reproducing natural conditions in the laboratory.

The zoned McMurry Meadows Pluton has been examined for REE and trace element variations in hornblende, sphene, apatite, allanite and zircon. Mineral separates (17), were analysed by INAA from four granitoids spanning the compositional range of the pluton (60%–75% SiO2). All of the phases examined exhibit significant compositional variations, with sphene having the largest changes in chondrite normalised REE patterns. Compositional variations in these minerals are related to paragenetic sequence and, as the whole rocks become more evolved, increases in partition coefficients; not subsolidus re-equilibration. Hornblende is only a dominant site for REE in granites where sphene is a later crystallising phase, otherwise allanite (LREE only) and sphene are the dominant whole rock sites for REE. Zircon and apatite normally contribute less than 10% each to the whole rock abundance of REE. Three zircon samples contain only small compositional differences and are interpreted as having crystallised from the bulk magma prior to differentiation. Zr variation in the pluton is nonlinear, first increasing and then decreasing with whole rock fractionation. A simple process, analogous to “restite unmixing” applied to the Zr variation, defines a bulk magma composition of about 63% SiO2, before differentiation of the zoned pluton. The modelled bulk magma need only have evolved by about 2·5% silica and still have produced the entire compositional range (60–75% SiO2) observed in the pluton.

S-type granites have properties that are a result of their derivation from sedimentary source rocks. Slightly more than half of the granites exposed in the Lachlan Fold Belt of southeastern Australia are of this type. These S-type rocks occur in all environments ranging from an association with migmatites and high grade regional metamorphic rocks, through an occurrence as large batholiths, to those occurring as related volcanic rocks. The association with high grade metamorphic rocks is uncommon. Most of the S-type granites were derived from deeper parts of the crust and emplaced at higher levels; hence their study provides insights into the nature of that deeper crust. Only source rocks that contain enough of the granite-forming elements (Si, Al, Na and K) to provide substantial quantities of melt can produce magmas and there is therefore a fertile window in the composition of these sedimentary rocks corresponding to feldspathic greywacke, from which granite magmas may be formed.

In this paper, three contrasting S-type granite suites of the Lachlan Fold Belt are discussed. Firstly, the Cooma Granodiorite occurs within a regional metamorphic complex and is associated with migmatites. It has isotopic and chemical features matching those of the widespread Ordovician sediments that occur in the fold belt. Secondly, the S-type granites of the Bullenbalong Suite are found as voluminous contact-aureole and subvolcanic granites, with volcanic equivalents. These granites are all cordierite-bearing and have low Na2O, CaO and Sr, high Ni, strongly negative εNd and high 87Sr/86Sr, all indicative of S-type character. However, the values of these parameters are not as extreme as for the Cooma Granodiorite. Evidence is discussed to show that these granites were derived from a less mature, unexposed, deeper and older sedimentary source. Other hypotheses such as basalt mixing are discussed and can be ruled out. The Strathbogie Suite granites are more felsic but all are cordierite-bearing and have chemical and other features indicative of an immature sedimentary source. They are closely associated with cordierite-bearing volcanic rocks. The more felsic nature of the suite results in part from crystal fractionation. It is suggested that the magma may have entered this “crystal fractionation” stage of evolution because it was a slightly higher temperature magma produced from an even less mature sediment than the Bullenbalong Suite. The production of these S-type magmas is discussed in terms of vapour-absent melting of metagreywackes involving both muscovite and biotite. The production of a magma in this way is consistent with the low H2O contents and geological setting of S-type granites and volcanic rocks in the Lachlan Fold Belt.

This paper reviews the petrogenesis of Himalayan leucogranites (HHγ) on the basis of field, petrological and geochemical data collected over the last fifteen years. HHγ are intruded at the top of the 2 to 8km-thick High Himalayan Crystallines (HHC). These are metamorphosed (Ky to Sill) and present much evidence of partial melting. During the MCT thrusting, the already metamorphosed HHC were thrust on top of the weakly metamorphosed Midland Formations, inducing the main phase of Himalayan metamorphism. The genesis of HHγ and North Himalaya leucogranites (NHγ) associates thrusting along the MCT, propagation of inverted metamorphism, liberation of large quantities of fluid in the Midlands, and partial melting of the HHC.

The restricted compositions of the granites are close to minimum melt compositions; variations in the alkali ratio probably relate to the variable amount of B, F and H2O. The HHγ were issued from the migmatitic zone around 700°C and 800 MPa., and still emplaced some 10 to 15 km below the surface. This syn- to late-tectonic emplacement of the leucogranites lasted for more than 10 Ma according to isotopic ages (25 to 14 Ma).

O, Rb–Sr, Nd–Sm and Pb isotope studies corroborate the unambiguous filiation between the HHC and the leucogranites in central Nepal. They also imply that the plutons are generated as numerous poorly mixed batches of magma produced preferentially in specific zones of the source rock. δD values may be explained by infiltration of water from the Midlands in the melting zone, and/or by water degassing during crystallisation. The positive covariations between Sr-, Nd- and O-isotope ratios relate to the variations in the original sediment composition of the source gneisses. Whereas trace element characteristics often date back to the anatectic process, limited magmatic differentiation is recorded by the biotite. These granites are typical crustal products, keeping track of some of the pre-Himalayan evolution together with that of their own origin.

The Miocene-Pliocene Macusani ash-flow tuffs and glasses from SE Peru are a rare example of two-mica felsic peraluminous volcanic rocks. They outcrop in three separate tectonic basins of the Cordillera Oriental in the Central Andes. In the Macusani field, the rocks are characterised by andalusite and muscovite phenocrysts. Compositions are both very felsic and peraluminous, similar to two-mica granites. Field relations, age differences and isotopic heterogeneities suggest that several distinct magma batches were involved. Two separate magmatic stages are recognised: (1) partial melting and evolution at or near the source region, and (2) shallow-level crystallisation and eruption. Magma genesis involved partial melting of metapelitic materials, with internally controlled. High heat flux, rapid heating, elevated temperatures and F-rich compositions were essential for the production of these mobile, H2O-undersaturated magmas. Chemical variations between the erupted products can be attributed to different degrees either of partial melting in somewhat variable source materials and (or) of fractional crystallisation at shallow levels. We discuss some important differences between the magmatic evolution of the Macusani Volcanics and of Hercynian and Himalayan two-mica granites.

Processes involved in the formation and evolution of melts within the lower crustal mafic granulites are considered with reference to mafic migmatites from late Proterozoic (1200-1000 Ma) granulites of the Rauer Group, East Antarctica. Metaluminous dioritic and noritic leucocratic veins on scales of 1 cm to 1 m show agmatitic, stromatic and schlieren structures. These possible melts are compositionally distinct from charnockitic and enderbitic orthogneisses, which show intrusive contacts with the migmatites in areas of low strain.

Important field relationships include the following:

(a) Leucocratic veins contain plagioclase and rare quartz, coarse subhedral to euhedral orthopyroxene, ilmenite and apatite. Finer (2 cm) veins and layers are richer in mafic phases than larger (2-10 cm) veins.

(b) Selvedges or melanosomes are developed between the larger melt areas and enclosing mafic gneisses. These melanosomes consist of garnet, orthopyroxene, plagioclase and biotite and are apatite-rich.

(c) Pyroxene granulite palaeosomes typically display bleached zones (1-2 cm) adjacent to selvedges and veins, in which the modal proportion of clinopyroxene diminishes in favour of orthopyroxene.

Geochemical and petrological studies demonstrate that localised or near-localised partial melting of the mafic granulites occurred during decompression from 8-9 kb to 7 kbar at a minimum temperature of 800-850°C. Geochemical mass balance calculations using measured vein, selvedge and palaeosome compositions indicate that near-closed system melting behaviour is likely for a large number of major and trace elements, but LILE behaviour is affected by the introduction of biotite probably associated with late stage fluids. Minor- and rare-earth element modelling predicts similar percentages of melting to those observed in the field, but yields reasonable results only when garnet is included as a minor residual phase. HREE concentrations in melanosomes do not show expected enrichments, probably as a result of later subsolidus changes including the breakdown of garnet during decompression.

This study demonstrates that migmatites may form through the near-localised partial melting of basic lithologies within the granulite facies. The exact role of fluids in this case cannot be determined but melting is interpreted to be vapour-undersaturated. This process may be important in the production of volumetrically significant amounts of dioritic to tonalitic calc-alkaline magmas.

One-dimensional modelling of the thermal history of a sialic crust thickened by multiple overstack thrusting of upper crustal material shows that anatexis is likely. Both the uplift rate and the length of the incubation period between end of tectonism and start of uplift are important controls on the amount and temperature of the melt. Heat of fusion does not significantly affect the long-term thermal structure of the crust if the melt is not extracted because only a small fraction of conductive heat is converted to latent heat, though short-term thermal effects of latent heat can be locally important.

Model results show that commonly <15% of mantle heat flux is converted to latent heat; even during peak melting in the most productive models, less than half of incremental mantle flux is converted. The results have obvious implications on the acceptability of proposed heat sources for crustal anatexis. Fusion could retard crustal temperature rise by nearly 100°C, but the system would recover except for situations of very rapid uplift. Understanding of the thermal evolution of a burial-uplift system requires knowledge not only of the timing of anatexis but of the pooling and movement of the magma, as well as the duration and nature of the incubation period; we are poorly equipped to measure these events. The model predicts that the characteristic time for anatexis in a thickened sialic crust is several tens of millions of years, comparable to the time lapse between orogenies; in making geological interpretations of magmatism, this time lag must be considered.

When basaltic magma is emplaced into continental crust, melting and generation of granitic magma can occur. We present experimental and theoretical investigations of the fluid dynamical and heat transfer processes at the roof and floor of a basaltic sill in which the wall rocks melt. At the floor, relatively low density crustal melt rises and mixes into the overlying magma, which would form hybrid andesitic magma. Below the roof the low-density melt forms a stable layer with negligible mixing between it and the underlying hotter, denser magma. Our calculations applied to basaltic sills in hot crust predict that sills from 10-1500 m thick require only 2-200 years to solidify, during which time large volumes of overlying layers of convecting silicic magma are formed. These time scales are very short compared with the lifetimes of large silicic magma systems of around 106 years, and also with the time scale of 107 years for thermal relaxation of the continental crust. An important feature of the process is that crystallisation and melting occur simultaneously, though in different spots of the source region. The granitic magmas formed are thus a mixture of igneous phenocrysts and lesser amounts of restite crystals. Several features of either plutonic or volcanic silicic systems can be explained without requiring large, high-level, long-lived magma chambers.

This paper is a structural and tectonic approach to the emplacement and deformation of granitoids. The main methods available in structural geology are briefly reviewed and this emphasises that (a) a wealth of data, particularly strain and shear sense, which pertain to these problems, can be determined in and around plutons; (b) given the nature, unlike many other crustal rock types, of granites to crystallise from isotropic through weakly anisotropic crystal suspension fluids, that deformation which has occurred in these states may not be well preserved; and (c) it is entirely possible, using this methodology, to separate deformation resulting from externally originating tectonic stresses from that which is associated with internal magma-related stresses. It is also recommended that the genetically-based Cloosian classification of granite fabrics and structures into “primary” (magmatic flow/magmatic flow current) and “secondary”, be abandoned and that a more observationally-based approach which classifies granite deformation fabrics and structures according to their time of occurrence relative to the crystallisation state of the congealing magma, be adopted (i.e. pre-full crystallisation deformation and crystal plastic strain deformation).

Examples of recent, structurally based, studies of emplacement mechanisms of plutons within tectonic settings are described and these show that, in general, space for magma can be created by the combination of tectonically-created cavities and internal magma-related buoyancy. This occurs in both transcurrent and extensional tectonic settings and there is no reason to doubt that it can happen in compressive-contractional regimes. It is concluded that transient and permanent space creation, such as may be exploited by available magmas, is a typical feature of the tectonically stressed and deforming lithosphere and this, in combination with the natural buoyancy and ascending tendency of magmas, can generate the varied emplacement mechanisms of granites.

Field associations (voluminous ash flow deposits, rhyolitic stocks and dykes, ring complexes), evidence of repeated influxes of mafic magma, and thermal constraints indicate that many high-level silicic plutons (magma chambers) acted as open systems for considerable parts of their history. The long thermal lifetime, as well as other evidence from the volcanic record, suggests that some such systems reached a quasi-steady state in which magma input was balanced by magma output for times longer than those required for crystallisation. Reconstruction of the evolution of large, long-lived caldera-forming systems, such as that of the Jemez Mountains, New Mexico, indicates that many chambers have lost a highly fractionated silicic cap, in some cases cyclically. Crystallised plutons may contain no obvious record of this evolutionary phase.

Geochemical data from silicic ash flow deposits can be used to reconstruct the volcanic stage of pluton development. Many silicic systems, especially of alkaline affinity, apparently pass from a stage in which melt evolution is dominated by crystal-liquid processes to one in which other processes may also contribute to differentiation. Apparently, the transition is most readily achieved in volatile-rich, alkaline silicic systems emplaced in complex, ancient sialic crust of the cratons. Once established, the preservation of highly fractionated caps on magma chambers requires a balance between thermal input and cooling-induced crystallisation. If heat enters the system too quickly, the cap may get stirred into the dominant magma volume by convection. If heat input is too slow, the magma body will crystallise inward from the margins, and the plutonic-consolidation stage will begin.

Structural and topographic relief along the eastern margin of the Rio Grande rift, northern New Mexico, provides a remarkable cross-section through the 26-Ma Questa caldera and cogenetic volcanic and plutonic rocks of the Latir field. Exposed levels increase in depth from mid-Tertiary depositional surfaces in northern parts of the igneous complex to plutonic rocks originally at 3–5 km depths in the S. Erosional remnants of an ash-flow sheet of weakly peralkaline rhyolite (Amalia Tuff) and andesitic to dacitic precursor lavas, disrupted by rift-related faults, are preserved as far as 45 km beyond their sources at the Questa caldera. Broadly comagmatic 26 Ma batholithic granitic rocks, exposed over an area of 20 by 35 km, range from mesozonal granodiorite to epizonal porphyritic granite and aplite; shallower and more silicic phases are mostly within the caldera. Compositionally and texturally distinct granites define resurgent intrusions within the caldera and discontinuous ring dikes along its margins; a batholithic mass of granodiorite extends 20 km S of the caldera and locally grades vertically to granite below its flat-lying roof. A negative Bouguer gravity anomaly (15–20 mgal), which encloses exposed granitic rocks and coincides with boundaries of the Questa caldera, defines boundaries of the shallow batholith, emplaced low in the volcanic sequence and in underlying Precambrian rocks. Palaeomagnetic pole positions indicate that successively crystallised granitic plutons cooled through Curie temperatures during the time of caldera formation, initial regional extension, and rotational tilting of the volcanic rocks. Isotopic ages for most intrusions are indistinguishable from the volcanic rocks. These relations indicate that the batholithic complex broadly represents the source magma for the volcanic rocks, into which the Questa caldera collapsed, and that the magma was largely liquid during regional tectonic disruption.

Volcanic and plutonic magmas (1) changed from early high-K calc-alkaline to alkalic prior to caldera eruptions; (2) differentiated to a weakly peralkaline rhyolite and equivalent acmiteartvedsonite granite cap (underlain by calc-alkaline granite) when the caldera formed at 26·5 Ma; then (3) reverted to calc-alkaline compositions. Concentrations of alkalis and minor elements such as Rb, Th, U, Nb, Zr, and Y reached maxima at the caldera stage. The volcanic rocks constitute intermittently quenched samples of upper parts of Questa magma bodies at early stages of crystallisation; in contrast, the comagmatic granitic rocks preserve an integrated record of protracted crystallisation of the magmatic residue as eruptions diminished. Multiple differentiation processes were active during evolution of the Questa magmatic system: crystal fractionation, replenishment by mantle and lower crustal melts of varying chemical and isotopic character, mixing of evolved with more primitive magmas, upper crustal assimilation, and perhaps volatile-transfer processes. As a result, an evolving batholithic cluster of coalesced magma chambers generated diverse assemblages of broadly cogenetic rocks within a few million years. Evolution of the Questa magmatic system and similar high-level Tertiary granitic batholiths nearby in the southern Rocky Mountains provides broad insights into magmatic processes in continental regions such as the overall shapes of batholiths, time and compositional relations between cogenetic volcanic and plutonic rocks, density equilibration of magmas with country rocks, and thermal evolution of continental crust.

The 6845 ± 50 BP climactic eruption of Mount Mazama discharged 47 ± 9 km3 of vertically zoned calc-alkaline magma, affording a virtually complete section through the chamber. Evidence for two andesitic parents with different trace-element (particularly Sr) and water contents is preserved in the ejecta. Prior to eruption, a dominant volume of rhyodacite was underlain successively by high-Sr andesite, high-Sr crystal mushes, and low-Sr crystal mushes. Intergranular liquids in the high-Sr magmas were probably richer in water than those in the low-Sr magmas. Thermal continuity throughout the ejecta favours eruption from a single, zoned reservoir. Insight into chamber development is given by preclimactic rhyodacitic lavas and tephra erupted between about 30,000 BP and the climactic eruption. The oldest of these lavas, contaminated derivatives of low-Sr magma, contain crystal-poor magmatic inclusions of low-Sr andesite; the youngest has inclusions of high-Sr andesite and, like rhyodacitic pumice in the climactic ejecta, is hybrid magma containing an admixed high-Sr component. A model for steady-state growth of the chamber is inferred whereby repeated recharge, first by low-Sr then high-Sr andesite (± basalt), builds up a cumulate succession, while derivative liquid fractionates convectively, segregates, and mixes with an incrementally growing silicic volume. The magma chamber at Mount Mazama may provide insight into the evolution of some granitoid plutons.

Jahns and Burnham (1969) proposed that the internal evolution of zoned granitic pegmatites could be explained by crystallisation from water-saturated melts which evolved to produce systems with a melt plus a separate aqueous fluid. Examination of microthermometric properties, chemical compositions and gas contents of solid-liquid-vapour inclusions from a number of the zones of the Tanco rare element granitic pegmatite places constraints on fluid evolution within the framework of the crystallisation history of the pegmatite, and contributes to an examination of the Jahns–Burnham proposal.

Initial crystallisation at Tanco was from the wall rock inwards, producing the relatively unfractionated wall zone (potassium feldspar–quartz-albite-muscovite). Textural evidence, and an upward increase in the level of geochemical fractionation, indicate that much, but not all, of the subsequent crystallisation of the pegmatite was from the base upwards. Inclusions trapped by wall zone and metasomatic wall rock tourmaline indicate that the pegmatite was intruded as a 2 phase alumino-silicate melt/fluid mixture at ∼720°C, with an initial fluid composition of ∼98mol.% H2O (containing 2 equiv. mo1% NaCl) and <2mol% CO2 (containing <5 equiv. mo1% CH4). These observations indicate that both melt and fluid were present from the start of crystallisation (Jahns & Burnham 1969), but show that CO2 and dissolved salts were important additional components of the fluid phase. The bulk of the pegmatite then crystallised in the range 600-470°C from melts and fluids with continued low levels of CO2 (2-3mol.%) and approximately constant salinity (∼7 equiv. wt.% NaCl dissolved in the aqueous phase). Crystal-rich inclusions, which may represent trapped alumino-silicate melts, are present throughout pegmatite crystallisation down to temperatures as low as ∼262°C. The final stages of crystallisation resulted in the formation of the beryl fringe at 291 ± 33°C and the lower part of the quartz zone at 262 ± 29°C. By the later stages the fluid had cooled through an H2O-CO2– dissolved salt solvus resulting in H2O-CO2 phase separation. Gas chromatographic analysis of the fluid components in the vug quartz, beryl fringe and lower part of the quartz zone shows that the inclusions contain H2O, CO2, CH4, N2, CO, Ar, and trace C2H6 in the beryl fringe. Measured CH4:CO2 ratios of 0·0060 (±0·0015) for the beryl fringe (twenty crushes on five samples) and 0·0042 (±0.0021) for the quartz zone (thirty crushes on six samples) yield fO2 estimates of 1×10−36 and 2 × 10−38, respectively, which are just above QFM at these temperatures.

Oxygen isotope data are very useful in determining the source rocks of granitic magmas, particularly when used in combination with Sr, Pb, and Nd isotope studies. For example, unusually high δ18O values in magmas (δ18O> +8) require the involvement of some precursor parent material that at some time in the past resided on or near the Earth's surface, either as sedimentary rocks or as weathered or hydrothermally altered rocks. The isotopic systematics which are preserved in the Mesozoic and Cenozoic batholiths of western North America can be explained by grand-scale mixing of three broadly defined end-members: (1) oceanic island-arc magmas derived from a “depleted” (MORB-type?) source in the upper mantle (δ18O c. +6 and 87Sr/86Sr c. 0·703); (2) a high-18O (c. +13 to +17) source with a very uniform 87Sr/86Sr (c. 0·708 to 0·712), derived mainly from eugeosynclinal volcanogenic sediments and (or) hydrothermally altered basalts; and (3) a much more heterogeneous source (87Sr/86Sr c. 0·706 to 0·750, or higher) with a high δ18O (c. +9 to +15) where derived from supracrustal metasedimentary rocks and a much lower δ18O (c. +7 to +9) where derived from the lower continental crust of the craton. These end-members were successively dominant from W to E, respectively, within three elongate N–S geographic zones that can be mapped from Mexico all the way N to Idaho.

18O/16O studies (together with D/H analyses) can, however, play a more important and certainly a unique role in determining the origins of the aqueous fluids involved in the formation of granitic and rhyolitic magmas. Fluid-rock interaction effects are most clear-cut when low-18O, low-D meteoric waters are involved in the isotopic exchange and melting processes, but the effects of other waters such as seawater (with a relatively high δD c. 0) can also be recognised. Because of these hydrothermal processes, rocks that ultimately undergo partial melting may exhibit isotopic signatures considerably different from those that they started with. We discuss three broad classes of potential source materials of such “hydrothermal-anatectic” granitic magmas, based mainly on water/rock (w/r), temperature (T), and the length of time (t) that fluid-rock interaction proceeds: (Type 1) epizonal systems with a wide variation in whole-rock δ18O and extreme 18O/16O disequilibrium among coexisting minerals (e.g. quartz and feldspar); (Type 2) deeper-seated and (or) longer-lived systems, also with a wide spectrum of whole-rock δ18O, but with equilibrated 18O/16O ratios among coexisting minerals; (Type 3) thoroughly homogenised and equilibrated systems with relatively uniform δ18O in all lithologies. Low-18O magmas formed by melting of rocks altered in a Type 2 or a Type 3 meteoric-hydrothermal system are the only kinds of “hydrothermal-anatectic” granitic magmas that are readily recognisable in the geological record. Analogous effects produced by other kinds of aqueous fluids may, however, be quite common, particularly in areas of extensional tectonics and large-scale rifting. The greatly enhanced permeabilities in such fractured terranes make possible the deep convective circulation of ground waters and sedimentary pore fluids. The nature and origin of low-18O magmas in the Yellowstone volcanic field and the Seychelles Islands are briefly reviewed in light of these concepts, as is the development of high-D, peraluminous magmas in the Hercynian of the Pyrenees.

Granites and their associated extrusive rocks are formed in large volumes whenever the continental crust is heated by rising hot mantle, or thickened by collision processes. The complexity of rocks of the granite family is related to the complexity of the continental crust itself and the complexity of processes which lead to thermal perturbations. The light continental crust acts as a density filter which screens out heavy mantle magmas and leads to complex underplating and magma mixing processes. Perhaps the primary cause of crustal melting is the deep recycling of volatiles which are fixed in the oceanic crust before subduction. Modern studies of subduction and collision processes show the large scale and complexity of processes which modify old continental crust.