Metamorphic rocks of undoubted crustal origin have been described in recent years, principally from Mediterranean collision zones that have been subjected to PT conditions along very low geothermal gradients (∼ 7°C/km) and have reached pressures up to 30 kbar. MgAl-rich metapelites develop particularly diagnostic high-pressure minerals and mineral assemblages that have been and are being studied experimentally in model systems involving the components K2O, MgO, Al2O3, TiO2, SiO2, P2O5, and H2O up to pressures of 50 kbar and temperatures of 1000°C.

In the present review the following synthetic phases and phase assemblages are discussed, emphasizing their water-pressure-temperature stability fields (approximated in parentheses here), their reaction relationships, and their known or potential occurrences in metamorphic rocks. Sudoite (0 to ∼ 12 kbar, 150? to 380°C) occurs in very low-grade metapelites. Mg-carpholite (∼ 7 to ∼ 45 kbar, ∼ 200 to 600°C) is found in subducted metabauxites, metapelites, and related quartz veins. Mg-chloritoid (18 to 45 kbar?; 400 to 760°C) has not been found in nature as pure or nearly pure end-member; it requires silica-deficient environments. Yoderite, known in nature only from a single talc-kyanite schist occurrence, has only a small stability field (9 to 18 kbar?, 700 to 870°C?), cannot coexist with quartz, but may be stabilized by Fe3+. Pyrope (∼ 15 to at least 50 kbar, ∼ 700°C to melting), with or without relic coesite inclusions, occurs spectacularly in quartzites. Mg-staurolite (∼ 14 to some 90 kbar?, 700 to 1000°C), recently discovered as inclusions in pyrope, requires silica-deficiency. MgMgAl-pumpellyite is a new synthetic phase in which Mg totally replaces Ca of normal pumpellyite; because of its very high-pressure, low-temperature stability (∼ 37 to at least 55 kbar, < 400 to 780°C) it may not form within our globe. Ellenbergerite, the new high-pressure mineral forming inclusions in pyrope, apparently exhibits a rather composition-dependent stability with Ti-ellenbergerite, requiring higher pressures (> 20 kbar) than P-bearing, Ti-free members; a pure hydrous Mg-phosphate with ellenbergerite structure was synthesized at 10 kbar. Phengites, the widespread MgSi-substituted muscovites, require increasingly high water pressures (up to ∼ 20 kbar) for higher degrees of substitution, but the Al-celadonite end-member is not stable under any conditions; the compositions of phengites coexisting with limiting assemblages such as phlogopite, K-feldspar, and an SiO2 phase are useful geobarometers. The common assemblage Mg-chlorite + Al2SiO5 (mainly kyanite) has an extensive stability field ranging from near zero to 31 kbar at temperatures varying from some 320 to ∼ 760°C depending on pressure. The whiteschist assemblage talc + kyanite (6 to ∼ 45 kbar, 550 to 810°C) plays an important role in collision zone metamorphism as it forms from the greenschist assemblage chlorite + quartz at low grades but is also known to break down into pyrope + coesite at the highest grade observed thus far. The assemblage talc-phengite (11 to at least 35 kbar, 300? to 820°C depending on pressure), on the other hand, is well known from subducted metapelites. At pressures of 15–20 kbar and temperatures of 400–650°C a very K,Mg-rich, siliceous fluid forms as a consequence of the mutual reaction of the minerals K-feldspar and phlogopite (biotite) which are very common in crustal rocks including granites. Such fluids are bound to cause metasomatism in neighbouring mantle rocks which, upon subsequent increase of temperature, produce post-collisional ultrapotassic, lamproitic melts.

The surfaces of quartz crystals that were partially dissolved in superheated, H2O-saturated rhyolite melt are covered with hemispherical embayments; each embayment is judged to have formed where a gas bubble in the melt approached the crystal. ‘Flux-line attack’ and ‘upward-drilling’ of the refractory lining of glass tanks are analogous processes. As a bubble nears a dissolving solid it enters a compositional boundary layer in the melt, resulting in melt of variable surface tension surrounding the bubble. This unstable situation results in small-scale convection of the melt about the bubble (Marangoni convection) which can cause locally enhanced dissolution rate of the solid. It is suggested that this mechanism could cause round embayments to form in quartz phenocrysts in acid volcanic and sub-volcanic rocks. Criteria by which embayed phenocrysts formed by dissolution can be distinguished from those formed by unstable growth are reviewed briefly.

Rock deformation and metamorphism can interact at the mechanistic level in the following ways: (a) facilitation of cataclasis through the release of high-pressure fluid during dehydration and decarbonation reactions; (b) facilitation of intracrystalline plasticity through the stresses induced during solid-state volume changes; (c) enhanced deformability through the transient existence of fine-grained reaction products; (d) modification of chemical potential gradients driving diffusion if a reaction can occur along the diffusion path; (f) changes in the resistance to intracrystalline plasticity through the effect of reaction-induced changes, pore fluid pressure and chemistry on point defect chemistry of the solid phases.

Examples of experimental studies of each of these types of interaction are described. Special experimental problems arise through: (i) the effects of solid phase and pore space volume changes, and their effects on pore fluid pressure and measured strain; (ii) the effects of such microstructural changes on the determination of flow law parameters; and (iii) in many instances the need for very long duration deformation experiments if reaction kinetics are sluggish.

There is an outstanding need for experimental studies of the effects of non-hydrostatic stress on the conditions for the onset of metamorphic reactions and phase transformations, as a basis for understanding some classes of deformation/metamorphism interaction. However, it is emphasized that the threefold classification of rock deformation mechanisms into cataclastic, crystal-plastic, and diffusive mass transfer processes, established from the study of deformation of monomineralic rocks, forms an essential framework for the understanding of deformation/metamorphism inter-relationships.

The work on liquid immiscibility in carbonate-silicate systems of Freestone and Hamilton (1980) has been extended to include alkali-poor and alkali-free compositions. Immiscibility is shown to occur on the joins albite-calcite and anorthite-calcite at 5 kbar. These results make it possible to interpret ocellar structure between calcite-rich spheroids in lamproite or kimberlite host rock as products of liquid immiscibility. The common sequence of rock types found in carbonatite complexes of melilitite-ijolite-urtite-phonolite is interpreted as being the result of both fractional crystallization and liquid fractionation, the corresponding carbonatite composition changing from nearly pure CaCO3 (±MgCO3) progressively to natrocarbonate. A carbonate melt cooling in isolation will suffer crystal fractionation, the residual liquid producing the rarer ferrocarbonatites, etc., whilst the crystal accumulate of calcite (dolomite) plus other phases such as magnetite, apatite, baryte, pyrochlore, etc., are the raw material for the coarse-grained intrusive carbonatites commonly found in ring complexes.

Glasses quenched from synthetic REE-carbonatite liquids in recent Tuttle bomb experiments require fast cooling rates. To study these glasses further, an internally-heated miniclave has been developed to increase quench rates and extend the operating P-T range of standard Tuttle bombs. The Haskel miniclave can achieve simultaneous P and T of up to 1200°C at 4 kbar for small diameter (1–2 mm) samples. Gas (Argon) pressure is supplied by a small intensifier unit and heating by an internal platinum-wound furnace. Because of the relatively small thermal mass, run temperatures can be reached within a few minutes and quench rates approach those of solid media apparatus.

A comparison of R.P.M. Union and Amandelbult Sections reveals close geochemical and stratigraphic correlations, but the sequence at the latter is more complex. Mafic members at Union Section are consistently more magnesian than their equivalents at Amandelbult Section and (where available data allow comparisons) than at Rustenburg Section. This, taken together with regional patterns of progressive attenuation and elimination of leucocratic rocks beneath harzburgitic layers, identifies Union Section as proximally located with reference to an irruptive centre from which primitive liquids were injected along the floor/supernatant liquid interface. Successive injections led to thermal erosion of the floor, causing ‘dimpling’ and ‘potholing’ on a local scale, and elimination of noritic-anorthositic layers on a regional scale. Such erosion of the floors beneath new inputs of magmatic liquid is likely to be initiated by only partial remelting, mainly of lower-temperature phases (sodic rims to zoned plagioclase grains, and Fe-enriched pyroxenes) within intercumulus space. It is argued that resultant contamination of the primitive liquids may have led to direct superposition of anorthosites (accompanied by rare troctolite) upon harzburgite, without intervening pyroxenite members, as in the Pseudoreef Multicyclic Unit.

Phosphorus contents in cumulus rocks occurring close to the level of apatite appearance in the basic rocks of the Bushveld Complex, South Africa, provide a method of calculating the proportion of intercumulus component in these rocks. Previous experimental studies have accurately constrained the phosphorus content of magmas when apatite becomes stable. The ratio of the phosphorus content in the cumulates immediately below the appearance of apatite to this liquid composition defines the proportion of trapped liquid.

Application of this method to rocks from the uppermost mafic rocks of the Bushveld Complex leads to the conclusion that there is from 1 to 6 per cent intercumulus component. Many of these rocks are multiphase cumulates and in such rocks estimation of intercumulus component from textural criteria is difficult.

If crystals grow In situ on the floor of the magma chamber such small proportions of interstitial component can be produced without appealing to excessive diffusion and circulation of magma through an unconsolidated crystal pile. The geometry of the intrusion as well as its size might have a major influence on the proportion of the liquid ultimately solidifying within a cumulus rock.

Amphiboles in dykes from the dominantly silica-undersaturated Monteregian series range from pargasitic megacrysts and xenocrysts to kaersutitic, pargasitic and hastingsitic phenocrysts, groundmass prisms and reaction rims. Amphiboles in dykes and plutons from the silica-oversaturated White Mountain Magma Series range from kaersutite, through hornblende, hastingsite and edenite, to sodic-calcic and sodic varieties. This contrast between the amphiboles from the two series is probably a reflection of differing melt silica activity and is a useful petrologic discriminant. In most cases, pargasitic amphibole megacrysts from Monteregian monchiquites reflect the Mg numbers of their host rocks and are considered cognate. The megacrysts are lower in Ti and higher in Mg and Al than their phenocryst mantles and rims. This is probably a result of higher pressures of formation.

Mg-Mn amphibole (tirodite), with or without pyroxmangite in the total absence of pyroxenes and high-calcic pyroxenoids, occurs in the Mn silicate rocks of the Sausar Group, India. The rocks were metamorphosed to amphibolite facies condition (T ∼ 650°C, P ∼ 6 kbar). Tirodite-pyroxmangite pairs developed in both carbonate-free and rhodochrosite-bearing assemblages. Also tirodite coexists with either kutnahorite or manganoan calcite in the absence of pyroxmangite. Mineral reactions inferred from modal abundances and compositions of the phases indicate stabilization of the amphibole alone from a bivalent cation-bearing residual unbuffered XCO2 system with XMn < 0.3. On the other hand, tirodite-pyroxmangite pairs appeared in unbuffered low to intermediate XCO2 assemblages with XMn > 0.35. Pyroxenes and high-calcic pyroxenoids did not appear in the present situation, though they occur elsewhere in rocks with broadly similar contents of immobile components. Closely associated assemblages of diverse mineralogy suggest that the XMn and XCO2, rather than the physical conditions of metamorphism, are the decisive factors in promoting the observed phase assemblages.

Chevkinite/perrierite crystallized from a series of REE-enriched compositions, ranging from basalt through andesite to rhyolite at pressures from 7.5 to 20 kbar and temperatures from 900–1050°C. The Al2O3 content increased with increasing pressure, but no other consistent composition change was observed with pressure and temperature variation in this range. The REE partition pattern consistently favoured light over heavy REE, with a regular decrease in partition coefficients (D) from La through to Lu. D values increase with decreasing temperature and decreasing REE content (i.e. Henry's Law does not apply, as REE are essential structural constituents of chevkinite/perrierite), but appear unaffected by pressure and fo2. Chevkinite/perrierite may be a significant fractionating phase in evolving silicic magmas, or a residual phase from low degrees of partial melting of granulites. In both these circumstances it will exert a major control on the REE content of derivative liquids.