Ever since the landmark paper by England and Richardson (1977) it has been clear that metamorphic rocks are not merely passively advected towards the surface during unroofing, but respond in a dynamic and interactive way to the mechanisms and rates of exhumation. England and Richardson based their modelling upon erosion-driven exhumation dominated by fluvial processes in mid-latitude climates. Since then a variety of mechanisms has been identified, mostly emphasizing tectonic mechanisms including extensional tectonics (e.g. Platt, 1996), tectonic extrusion (Chemenda et al., 1996; Hynes et al., 1996; Thompson et al., 1997) and ‘gravity currents’ (Wernicke and Getty, 1997). It has also become apparent that exhumation itself may be strongly influenced by metamorphic processes. For example, changes in petrophysical properties such as density may dramatically affect the buoyancy of rock masses and hence alter the driving forces for exhumation (Bousquet et al., 1997). The evolution of fluids or melts may create weak zones in the crust and enhance ductile or brittle deformation and shear zone development (e.g. Austrheim, 1998).

Most unroofing mechanisms invoked for the exhumation of blueschist-plus-eclogite terranes, including corner-flow and extensional collapse of the orogenic wedge, predict steep unroofing paths for the deeply-buried rocks and are applicable only to unroofing from depths within the crust. Many high-P and ultrahigh-P rocks of continental affinity are derived from greater depths than this. Their lack of warming during unroofing, together with indications that they may rest directly on less deeply buried equivalents, are suggestive of shallow unroofing paths similar to those for the subduction-channel model. They are interpreted to have been emplaced by the upward extrusion of coherent slices of continental crust, bounded below by thrust faults and above by normal faults, with unroofing paths essentially reversing the original burial paths.

Where continental crust has been subducted into the mantle, upward extrusion is probably driven largely by buoyancy forces, although examples of upward extrusion without subduction into the mantle indicate that buoyancy forces may not be essential. Two features in addition to buoyancy may promote upward extrusion. Slab breakoff may reduce the pull from the descending slab, and subduction-zone geometry may change as a continental margin is dragged into the subduction zone. Both features may promote the extrusion of continental crust at precisely the time at which it has been partially subducted.

A close spatial relationship between a lateral ramp and a lobate zone of extruded high-P rocks in the Mesoproterozoic Grenvillian orogen indicates that lateral ramps may be important in localizing extrusion. Lateral ramps disturb the two-dimensional flow, with channelling of material into the region of the lateral ramp as it is extruded. Many exhumed ultrahigh-P terrains are associated with jogs in the trends of orogenic fronts that may reflect the presence of lateral ramps at depth. Ultrahigh-P rocks may be expected to be concentrated at such jogs, and may record the channelling in their deformation history.

At time-scales of 102 to 105 years, erosion by rivers, landslides and glaciers can exceed 5 mm/y. Sustained denudation at these rates is sufficient to account for many of the rapid rates of unloading or cooling that are revealed by geobarometric or thermochronologic studies. Because feedbacks exist among many surface processes, determinations of erosion rates on a few geomorphic processes can be adequate to estimate mean rates across an entire landscape. Rates of fluvial and glacial incision exert a dominant control on landscape lowering, because these rates set the local base level and modulate the flux from adjacent hillslopes. When rates of deformation are sufficiently rapid and sustained, a collisional orogen approaches a dynamic equilibrium or topographic steady state. Due to variations in erosion rates as a function of Late Cenozoic climate changes, such a steady state should be defined at time-scales longer than one climate cycle. During dynamic equilibrium, the geomorphic system displays a predictable configuration of interacting rivers, hillslopes and glaciated process zones. A dynamic equilibrium appears to prevail near Nanga Parbat, Pakistan, and the Southern Alps of New Zealand, where spatial variations in geomorphic erosion rates mimic variations in bedrock cooling rates at time-scales of 106 y.

We review the causes, mechanisms and consequences of crustal anatexis during the exhumation of metamorphic terranes, from a petrological perspective. During both prograde and retrograde metamorphism, limited influx of free hydrous fluids may result in small volumes of very hydrous melts, which cannot ascend far (if at all) before reaching their solidus. If thermal conditions for dehydration melting are attained in fertile micaceous crustal layers, much larger volumes of water-undersaturated granitic magmas may result, especially where limited external fluid influx raises water activities above those that may be buffered by dehydrating hydrous phases. Magmas have specific trace element characteristics depending on the reaction which formed them which, combined with accessory phase thermometry, may enable the (P-T) conditions of melting to be ascertained. Small volume-fraction magmas will typically remain as in situ migmatites unless their extraction is assisted by deformation. In turn, deformation will be focused in weaker partially molten zones, so that water-undersaturated magmas may often be mobilized. Once segregated, their ascent is limited by the rate of dyke propagation, and they may reach shallow levels (<2 kbar) before crystallizing. The complex interplay between deformation and melting is exemplified by the Miocene evolution of the central Himalaya, where thrust and normal faulting, melting and exhumation were all simultaneously active processes which were linked by feedback relations. In the Nanga Parbat Massif of the western Himalaya, rapid post-Miocene denudation and vigorous fluid flux enabled rocks to experience more than one episode of melting simultaneously, at different levels of the same exhuming crustal section.

Fluids can play a direct role in exhumation by influencing exhumation mechanisms and the driving processes for these mechanisms. In addition, the process of exhumation leads to the development of fluid-related features that in themselves may not drive exhumation. Fluids involved in exhumation are generally derived from dehydration reactions occurring during decompression, but at shallower crustal levels may also involve the introduction of exotic fluids. The composition of fluids attending exhumation are generally saline – CO2 mixtures, but N2, CH4, H2O mixtures have also been recorded. Studies of fluid features related to exhumation have found that fluids may contribute to density changes and the initiation of partial melting during decompression, as well as the development of extensive vein systems. However, the preservation of geochemical signatures related to fluid processes occurring prior to high-P and ultrahigh-P metamorphism indicates that large-scale pervasive fluid flow systems, in general, do not operate at any stage during the exhumation history. Large-scale channelled fluid flow may have operated in detachment faults and shear zones related to exhumation, and this requires further study. The most significant role of fluids during exhumation appears to be their controlling influence on the preservation of high-P or ultrahigh-P rocks.

The exhumation history of mountain belts can be derived from radiometric dating of detrital mineral grains in proximal and distal post- and synorogenic sediments. The application of single-crystal dating techniques avoids the averaging effect that characterizes multi-grain and whole-rock techniques and allows the identification of populations of grains with distinct thermal histories. Of the major single crystal dating methods available, 40Ar/39Ar dating of detrital K-bearing minerals, in particular white mica, is perhaps the most versatile and widely applied technique. For a closure temperature of Ar of 350–400°C, muscovite 40Ar/39Ar ages record the time a rock mass passed through 8–10 km beneath actively eroding mountain belts. Detrital muscovit ages eroded from orogenic mountain belts have been used extensively to identify the provenance of sediments from source regions with distinct thermal histories, determine the history and rate of exhumation of the source region, and provide an upper limit on the sediment age. Here I review the principles of 40Ar/39Ar dating of detrital muscovite and illustrate the method with examples showing how the provenance and the thermal history of sediment source regions derived from such studies can be used to constrain the exhumation and tectonic history of orogenic belts.

The new minerals manganvesuviante and tweddillite, both formed by hydrothermal alteration of primary manganese ores, are described from the Kalahari manganese fields (Republic of South Africa). In addition, single-crystal X-ray structure refinements of both new minerals are presented.

Manganvesuvianite is a tetragonal vesuvianite mineral with the simplified formula Ca19Mn3+(Al,Mn3+,Fe3+)10(Mg,Mn2+)2Si18O69(OH)9, characterized by Mn3+ occupying the five-coordinated position (square pyramid). The crystals simple prismatic forms: {100}, {110} terminated by {101} and exhibit deep maroon red colour. With polarized light the crystals are strongly pleochroic, yellowish parallel to E and dark red to lilac parallel to O.

Tweddillite is an epidote-group mineral (space group P21/m, a = 8.932(5), b = 5.698(4), c = 10.310(5) Å, β = 114.56(4), V = 477.3(8) Å3) with the simplified formula CaSr(Mn3+,Fe3+)2Al[Si3O12](OH), closely related to strontiopiemontite. The difference between strontiopiemontite and tweddillite is the concentration of octahedral Mn3+. Strontiopiemontite has Mn3+ mainly on the M3 site whereas tweddillite has Mn3+ with minor Fe3+ on M3 and M1. Tweddillite forms aggregates of very thin dark red {001} blades characterized by striking pleochroism. The crystals appear dark red parallel to b and orange-yellow parallel to a. Perpendicular to (001) the blades appear magenta to red.

Apatite samples from rare-metal mineralization were investigated by a combination of cathodoluminescence (CL) microscopy and spectroscopy, microchemical analysis and trace element analysis. Internal structures revealed by CL can be related to variations in the crystal chemistry and may sometimes reflect changes in the composition of the mineralizing fluids.

Apatite from mineralization related to alkaline rocks and carbonatites (Type 1) typically exhibits relatively homogeneous blue and lilac/violet CL colours due to activation by trace quantities of rare earth element ions (Ce3+, Eu2+, Sm3+, Dy3+ and Nd3+). These results correlate with determined trace element abundances, which show strong light rare earth element (LREE) enrichment for this type of apatite. However, a simple quantitative correlation between emission intensities of REE3+/2+ and analysed element concentrations was not found.

Apatite from P-rich altered granites, greisens, pegmatites and veins from Sn-W deposits (Type 2) shows strong Mn2+-activated yellow-greenish CL, partially with distinct oscillatory zoning. Variations in the intensity of the Mn2+-activated CL emission can be related either to varying Mn/Fe ratios (quenching of Mn activated CL by Fe) or to self-quenching effects in zones with high Mn contents (>2.0 wt.%). The REE distribution patterns of apatite reflect the specific geological position of each sample and may serve as a “tracer” for the REE behaviour within the ore system. Although the REE contents are sometimes as high as several hundred parts per million, the spectral CL measurements do not exhibit typical REE emission lines because of dominance of the Mn emission. In these samples, REE-activated luminescence is only detectable by time-resolved laser-induced luminescence spectroscopy.

Both types of apatite (Type 1 in the core and Type 2 in the rim) were found in single crystals from the Be deposit Ermakovka (Transbaikalia). This finding proves the existence of two stages of mineralization within this deposit.

Verbeekite, ideally PdSe2, monoclinic with space-group choices C2/m, C2 or Cm; a = 6.659(7), b = 4.124(5), c = 4.438(6) Å, ß = 92.76(3)°, V = 121.7(4) Å3; a:b:c = 1.6147:1:1.0761, Z = 2, is a new, very rare, primary mineral, intimately associated with secondary oosterboschite {(Pd,Cu)7Se5}, from the Musonoi Cu-Co-Mn-U mine, near Kolwezi, Shaba Province, Democratic Republic of Congo. Additional associated minerals are Cu- and Pd-bearing trogtalite {(Co,Cu,Pd)Se2}, Se-bearing digenite and Se-bearing covellite. The strongest five lines of the X-ray powder-diffraction pattern {d in Å (I) (hkl)} are: 4.423(30)(001), 3.496 (30)(110), 2.718(100)(111), 1.955(50)(310 and 1.896(50)(12). The mineral has also been identified, as a single anhedral 25 µm-sized grain, from Hope's Nose, Torquay, Devon, England where it is associated with native gold, chrisstanleyite Ag2Pd3Se4, oosterboschite(?), unnamed Pd2HgSe3 and cerussite. At Musonoi, altered verbeekite grains do not exceed 200 µm in size and are anhedral, black, with a black streak and a metallic lustre. The mineral is opaque, brittle, has an uneven fracture, and lacks discernible cleavage. The VHN5 ranges 490–610, mean 550 kp/mm2 (2 indentations), roughly approximating a Mohs' hardness of 5Ý. Dcalc. = 7.211 g/cm3 for the ideal formula. Electron-microprobe analyses (mean of 4 spot analyses) yielded Pd 39.6, Cu 0.5, Se 58.8, total 98.9 wt.%. The empirical formula is (Pd0.99Cu0.02)σ1.01Se1.99, based on Pd+Cu+Se = 3. In plane-polarized reflected light, the mineral is a nondescript grey and is neither pleochroic nor perceptibly bireflectant. Anisotropy is moderate with rotation tints in varying shades of brown. Reflectance spectra and colour values are tabulated. The name honours Dr Théodore Verbeek (1927–1991) who was the first geoscientist to study the Musonoi palladium mineralization in the Democratic Republic of Congo (1955–1967) and who co-discovered this new mineral phase.

Decrespignyite-(Y) is a new copper yttrium rare earth carbonate chloride hydrate from the Paratoo copper mine, near Yunta, Olary district, South Australia. Decrespignyite-(Y) occurs as blue crusts, coatings and fillings in thin fissures on the slatey country rock. Individual pseudohexagonal platelets are typically 10–50 µm in maximum dimension and are often curved. Associated minerals include caysichite-(Y), donnayite-(Y), malachite and kamphaugite-(Y). Electron microprobe and CHN analyses gave: Y2O3 42.2; La2O3 0.1; Pr2O3 0.1; Nd2O3 1.3; Sm2O3 1.0; Gd2O3 4; Tb2O3 0.4; Dy2O3 3.7; Ho2O3 2.6; Er2O3 2.5; CaO 0.5; CuO 10.9; Cl 3.0; CO2 19.8; H2O 10.8, yielding an empirical formula of (Y3.08Gd0.22Dy0.16Ho0.11Er0.10Nd0.06Sm0.05Tb0.02La0.02Pr0.01Ca0.08)∑3.91Cu1.12(CO3)3.70-Cl0.7(OH)5.79·2.4H2O. The simplified formula is (Y,REE)4Cu(CO3)4Cl(OH)5·2H2O. The mineral is royal blue to turquoise-blue in colour, transparent, with a pearly to vitreous lustre and a pale blue streak. No cleavage was observed but the morphology suggests that cleavage would be on [010]. The Mohs' hardness is estimated to be 4. The strongest lines in the X-ray powder pattern are {dobs (Iobs) (hkl)} 22.79 (30) (010); 7.463 (30) (001); 7.086 (50) (011); 6.241 (100) (021); 4.216 (30) (12); 3.530 (40) (022); 3.336 (30) (032); 2.143 (30) (222, 01). The powder diffraction pattern was indexed on a monoclinic cell with a = 8.899(6), b = 22.77(2), c = 8.589(6)Å, β = 120.06(5)°, V = 1506.3(7) Å3 and Z = 4. The structure of the new mineral could not be determined but powder diffraction data indicate the space group is P2, Pm or P2/m. The measured density is 3.64(2) g/cm3 and the calculated density is 3.645 g/cm3. Decrespignyite-(Y) is biaxial negative with α = 1.604(4) and γ = 1.638(3) with β very close to γ pleochroism is medium strong; X very pale bluish, Y and Z bluish (with greenish tint). Decrespignyite-(Y) is a supergene mineral which precipitated from mildly basic carbonated ground waters. The mineral is named after Robert Champion de Crespigny, a prominent figure in the Australian mining industry and chancellor of the University of Adelaide.

The volume proportions of the mineral phases in two strongly deformed olivine-orthopyroxene rocks have been quantified by whole-pattern stripping of fixed geometry X-ray powder diffraction data. The results were compared with the phase proportions as determined by Rietveld refinement of time-of-flight neutron powder diffraction data, and were shown to be in excellent agreement. The X-ray technique not only provides a very rapid and cost-effective method of determining phase proportions, but it also circumvents several of the problems associated with obtaining this information by image analysis. Moreover, the technique is particularly advantageous in strongly textured rocks or in rocks that contain significant residual strains. As such it offers a powerful technique for analysing the mineralogical composition of fine-grained and/or deformed experimental run products, which makes it of considerable potential for monitoring in situ the progress of mineral reactions during laboratory experiments.

The compositions of fluid inclusions hosted in emerald and quartz (the Maria deposit, Mozambique) were studied using microthermometric and Raman microprobe techniques. The fluid inclusions in the emerald contain fluids within the Na-Ca-Mg-(HCO3)−-(CO3)2−-Cl-H2O system saturated in carbonic acid brines. Nahcolite is a main daughter solid phase within the fluid inclusions. The mean nahcolite and NaCl contents are 25 and 5 wt.%, respectively. Mg-calcite, magnesite, calcite and aragonite are also identified as daughter phases in the fluid inclusions. Many fluid inclusions show necked-down appearance. Groups of nahcolite crystals often make up ∼50 vol.% of necked-down inclusions. It seems that zones of fluid inclusions with numerous birefringent solid phases are distinctive of the Maria emerald deposit. The likely conditions of emerald growth are 400 < T < 500°C and 3 < P < 4 kbar.

Tetrahedrite fahlores from the Coeur d'Alene mining district (Idaho) have been found to be enriched in Ag by the Ag–Cu exchange reaction

which occurred during cooling following galena mineralization. This solid–state reaction resulted in quantitative removal of Ag (in a AgSbS2 component) from galena and development of bournonite coronas on fahlore grains. The reaction produced a distinct population of high-Ag fahlores found in galena-rich samples and accounts for all of the bournonite mineralization. The most argentian of these high-Ag fahlores examined in this study (molar Ag/(Ag+Cu) = 0.303±0.011 and 0.336±0.011) are found in samples which achieved saturation with respect to other Ag-sulfosalts, namely pyrargyrite and polybasite and diaphorite, respectively. Multiple lines of evidence for the Ag-Cu exchange reaction are presented in this paper which applies mass-balance constraints and a thermodynamic database for sulfides/sulfosalts to microprobe analyses and textural observations, and to bulk production data. This solid-state reaction may explain why previous district studies have been unable to demonstrate convincingly primary fahlore zoning. Based on the Ag/(Ag+Cu) of fahlores coexisting with other Agsulfosalts and Fe-Zn partitioning between fahlore and sphalerite, we estimate that fahlore compositions were frozen in by 175°C. Examining the composition data for other Ag tetrahedrite fahlores found in the Coeur d'Alene district and elsewhere, we conclude that the thermodynamic database provides an accurate description of phase equilibria, except possibly for very Fe-rich systems where further studies of the Fe-Zn partitioning between fahlore and Fe-rich (Zn,Fe)S phases and of the thermodynamics of Fe-rich (Zn,Fe)S phases are warranted.