Little is known about the physical properties of carbonatite magmas, making it difficult to predict dynamic behaviour in carbonatite-bearing magmatic systems. The viscosity of calcium-rich carbonatite magma is approximately estimated from molten salt data to be 0.1 Pa s at 700–800°C, while density is estimated at 2.3−2.5 × 103 kg m−3. The corresponding values for natrocarbonatite are 0.01 Pa s and 2.0−2.1 × 103 kg m−3. It is thus possible for carbonatite to be negatively buoyant with respect to some silicate magmas. The surface tension in air of carbonatite magmas is estimated at 0.25 and 0.21 N m−1 for Ca-carbonatite and natrocarbonatite respectively. Knowledge of the interfacial tension between carbonatite and silicate liquids is critical before the formation and behaviour of silicate–carbonatite emulsions can be properly understood. Interfacial tension is constrained to < 0.09 N m−1 by the application of multiphase drop theory to experimentally-produced textures, and this value receives some support from geological observations. The mechanics of extraction from layered carbonatite-silicate magma chambers are briefly examined using the recommended density and viscosity values and the equations of Blake & Ivey (1986); the degree of eruptive mingling is dependent on which liquid was uppermost in the chamber.

The c. 15000 km2, c. 2 Ga Kunene complex of southern Angola and northern Namibia is one of the world's largest occurrences of anorthositic rocks, rivalled only perhaps by the Lac St Jean massif of the Grenville Province in Quebec. We report here the results of a detailed field and laboratory study of a 100 km2 area in the northern part of the complex. Coarse grained (av. 1–2 cm) anorthosite and leucotroctolite are the predominant rock types, although colour index varies between 0 and 50, averaging about 10. Ultramafic rocks are totally absent. Typical mineralogy is: plagioclase (An57–76) + olivine(Fo64–71) + orthopyroxene(En65–74) + clinopyroxene ± Fe—Ti oxides ± apatite. Textures are dominantly massive, although weak, impersistent, magmatic lamination with nearvertical dip and unsystematic strike is also present. Other magmatic features include plagioclase and orthopyroxene megacrysts, block structure, mortar texture, and anorthositic dykes. Metamorphic effects are minimal or absent. All of these attributes are similar to those found in typical massif-type anorthosites. Since comparable features are present over large areas elsewhere in the complex, we suggest that Kunene should be considered analogous to a large, composite, massif-type anorthositic intrusive complex, rather than to a large, single or composite mafic layered intrusion such as Bushveld, as has been previously suggested or assumed. This interpretation is supported by satellite imagery, which shows the outlines of several individual anorthositic plutons, one of which clearly encompasses our field area.

Very thick shallow water sedimentary sequences were deposited in Neoproterozoic time along the future margins of Laurentia. On the eastern margin these include the Eleonore Bay and Hecla Hoek sequences of Greenland and Svalbard; these are described and their geotectonic context briefly reviewed. They present both geotectonic and geodynamic problems: why did subsidence continue for some 300 Ma prior to the opening of Iapetus, and how could 15–20 km of sediment be accommodated in an ensialic environment?

Prolonged slow stretching appears to have affected the eastern margin of Laurentia while the western (Cordilleran) margin progressed through the rift-drift transition as the Pacific opened. It is proposed that expansion of the Pacific was associated with both the convergence of Proto-Gondwanan continental terranes during the Pan-African orogeny (the extended SWEAT hypothesis) and also the maintenance of very slow extension rates on the future Iapetus margin. The strain-hardening effect of slow stretching may have been inhibited by a continuous basin-fill of juvenile heat-producing Grenville detritus. The onset of subduction in the Pacific freed up this margin; major rifting took place between East Greenland and possibly the Tornquist margin of Baltica in Vendian time, followed by the opening of northern Iapetus.

A detailed carbon- and oxygen-isotope stratigraphy has been generated from Upper Cretaceous coastal Chalk sections in southern England (East Kent; Culver Cliff, Isle of Wight; Eastbourne and Seaford Head, Sussex; Norfolk Coast) and the British Geological Survey (BGS) Trunch borehole, Norfolk. Data are also presented from a section through the Scaglia facies exposed near Gubbio, Italian Apennines. Wherever possible the sampling interval has been one metre or less. Both the Chalk and Scaglia carbon-isotopic curves show minor positive excursions in the mid-Cenomanian, mid- and high Turonian, basal Coniacian and highest Santonian–lowest Campanian; there is a negative excursion high in the Campanian in Chalk sections that span that interval. The well-documented Cenomanian–Turonian boundary ‘spike’ is also well displayed, as is a broad positive excursion centred on the upper Coniacian. A number of these positive excursions correlate with records of organic-carbon-rich deposition in the Atlantic Ocean and elsewhere. The remarkable similarity in the carbon-isotope curves from England and Italy enables cross-referencing of macrofossil and microfossil zones and pinpoints considerable discrepancy in the relative positions of the Turonian, Coniacian and Santonian stages.

The oxygen-isotope values of the various Chalk sections, although showing different absolute values that are presumably diagenesis-dependent, show nonetheless a consistent trend. The East Kent section, which is very poorly lithified, indicates a warming up to the Cenomanian–Turonian boundary interval, then cooling thereafter. Regional organic-carbon burial, documented for this period, is credited with causing drawdown of CO2 and initiating climatic deterioration (inverse greenhouse effect). Data from other parts of the world are consistent with the hypothesis that the Cenomanian–Turonian temperature optimum was a global phenomenon and that this interval represents a major turning point in the climatic history of the earth.

The Themed Fault marks the southernmost border of the Early Mesozoic passive continental margin of north Sinai. This 200-km long fault transects the northern part of the Tih Plateau that supposedly occupies a tectonically stable area. Post-Middle Eocene–pre-Early Miocene rejuvenation of this fault proceeded by right-lateral wrenching and represents a newly recognized phase of deformation in the history of north and central Sinai. The minimum estimate for the strike-slip movement on this fault is about 300–750 m. To the north of the Themed Fault is a narrow fault belt (Sinai hinge belt) that marks the boundary between a tectonically unstable crustal block to the north (the north Sinai fold belt area) and a tectonically stable crustal block to the south, the main part of the Tih plateau area.

Four phases of dextral wrenching rejuvenated the faults of the Early Mesozoic passive continental margin in northern Egypt; one of them affected the Themed Fault. The oldest (Dl) deformation is early Late Senonian and is related to the closure of Neotethys and the Eastern Mediterranean basin. The D1 deformation proceeded by pure wrenching in the north Western Desert of Egypt. In contrast, it proceeded by transpression in north Sinai due to the irregular plate boundary and the relationship of this boundary to the slip vectors. D2 deformation (post-Middle Eocene–pre-Early Miocene) is clear in the Themed Fault area although reported herein for the first time; it is related to continued closure of the Eastern Mediterranean basin and proceeded by pure wrenching. D3 deformation (Late Oligocene–Early Miocene) proceeded by divergent wrenching in the north Eastern Desert and is kinematically related to the transfer of slip from the nearby faults of the Suez rift. D4 deformation (post-Early Miocene to Recent) affected the Sinai hinge belt by pure wrenching and is probably related to the left-lateral slip on the Dead Sea Transform and the related drag of the eastern edges of the fault blocks of this hinge belt. Recent seismic activity in the Sinai hinge belt perhaps indicates that the D4 deformation has continued to the present time, although morphological expression of recent tectonic movement is lacking. In contrast, the Themed Fault is seismically quiet at present.

After a long period of subaerial fissure-fed extrusions of basaltic magmas (∼ 12 to > 3 Ma) volcanic activity was then concentrated in the central part of Tenerife. Phonolitic magma chambers formed and a central volcanic complex was constructed (the Las Canadas edifice). The formation of a large depression (the Las Canadas caldera) truncated the top of the edifice. The active twin strato-cones Teide—Pico Viejo are sited in this depression. The history of the Las Canadas caldera and edifice are established from stratigraphy, geochronology (K—Ar dates) and volcanological studies. Two different groups are recognized, separated by a major unconformity. The Lower Group is dated at 2 to 3 Ma and includes the products of several volcanic centres, which together represent several cycles. The Upper Group ranges from 1.56 to 0.17 Ma and includes three different formations representing three long-term (∼ 100 to 300 Ka) volcanic cycles. The periods of dormancy between each formation were of ∼ 120 to 250 Ka duration. The Las Canadas caldera is a multicyclic caldera which formed over the period 1.18–0.17 Ma. Each cycle of activity represented by a formation culminated in caldera collapse which affected different sectors of the Las Canadas edifice. Geological observation and geochronology support an origin by collapse into a magma chamber. The minimum volume of pyroclastic ejecta is substantially greater than the present caldera depression volume (45 km3), but approaches the inferred volume of the original caldera depression (> 140 km3). After the formation of the caldera, sector collapses could also occur at the northern flank of the volcano causing the disappearance of the northern side of the caldera wall.

Neoproterozoic-early Cambrian successions in Iberia are reexamined. A gradual transition across the Neoproterozoic-Cambrian boundary is present in Central Iberia, whereas in the Cantabrian region and the Iberian Chains Lower Cambrian arenaceous successions rest with profound angular unconformity on Neoproterozoic turbidites. In Central Iberia, the Neoproterozoic sedimentary succession is referred to the informal Domo Extremeño group, representing mostly basinal facies, and the overlying Rio Huso group consisting of slope deposits and proximal turbidites that grade into shallower marine deposits. The latter is inferred to represent distal slope to outer platform depositional conditions and contains widespread carbonate olistostromic units. The position of the Precambrian-Cambrian boundary is within the Pusa shale of the Rio Huso group and can be correlated at the regional level by the occurrence of trace fossils, acritarchs, and in particular the abundant shelly metazoan Cloudina. The succession also yielded megascopic carbonaceous fossils, such as vendotaenids and Beltanelioides? sp. ind., and Sabellidites. Contrary to former interpretations assuming transport of older platform carbonates from the Ibor region into ‘younger’ olistostromic beds of the Rio Huso group, we interpret sedimentary and fossil evidence to suggest that shallower platform deposits of the informal Ibor group were penecontemporaneously incorporated in the olistostromic lower part of the Rio Huso group. Hence, the olistostromes are not believed to mark a major erosive unconformity. Based on the ichnofossil record and recent U-Pb age determinations, we argue that a proposed disconformity between Lower and Upper ‘Alcudian’ strata is neither regional nor does it mark a significant hiatus. Neoproterozoic-early Cambrian deposition in Central Iberia can be accommodated in a model that implies a generalized stretching of the crust during an extensional event which closely followed the Cadomian phase of the Pan-African Orogeny and which eventually could have included transcurrent components. An extensional phase with transcurrent components during the deposition of the Ibor and lower Rio Huso groups is regarded as a probable cause of widespread ponding resulting in the juxtaposition of platform and basinal successions, eventually leading to anoxic conditions in Pusa shale deposition times. A possible cause for repeated collapse events developing olistostromes and intra-sequential folding could be sought in this tectonic context.

A single component, natural remanent magnetization (NRM) is carried largely by pseudosingle domain magnetite in the Cretaceous Lower Chalk and Red Chalk of eastern England. The Red Chalk also records the same direction in haematite. Most of the ferro-magnetic minerals occur as primary clastic or early diagenetic grains. A stable remanence component is resistant to demagnetization, and is carried by both magnetite and haematite. Nevertheless, it has a steep inclination close to the present Earth's field and it is too steep for the previously reported palaeolatitude of these rocks at the time of sedimentation. A postglacial slump breccia scatters the ChRM but also provides some evidence of viscous, partial magnetic overprinting during slumping. Despite its resistance to thermal and alternating field demagnetization the characteristic remanent magnetization (ChRM) is probably a young Bruhnes epoch viscous remanent remagnetization (VRM).

The Desert de les Palmes area, in the southeast Iberian Chain, belongs to a Mesozoic NE–SW high which separated the early Cretaceous basins of the Maestrat and Aliaga-Penyagolosa from the little Orpesa basin. Its structure is characterized by the development of a system of NE–SW to ENE–WSW extensional listric faults detached in a shallow upper crustal level (1.7–2.2 km), mostly affecting the pre-Upper Cretaceous rocks. These faults record two well-differentiated rifting periods: (1) a first late Triassic–early Jurassic rifting period that divided the Desert de les Palmes high in several blocks; (2) a second early Cretaceous rifting period, only developed in the eastern margin of the Desert de les Palmes high, which was related to the opening of the Maestrat, Aliaga-Penyagolosa and Orpesa basins. Based on the comparison of the main features of this Mesozoic structure with an analysis of the structural and subsidence data already known in the neighbouring Mesozoic basins (Maestrat, Aliaga-Penyagolosa and Columbrets), a geodynamic scenario for the crustal evolution of the eastern Iberian Chain is also suggested. This involves four evolutionary stages: (1) Triassic rift (late Permian–Hettangian); (2) early and middle Jurassic postrift (Sinemurian–Oxfordian); (3) late Jurassic and early Cretaceous rift (Kimmeridgian–middle Albian), which includes a short Hauterivian postrift period; and (4) late Cretaceous postrift (late Albian–Maastrichtian).

Isotopic chemostratigraphy has proven successful in the correlation of carbonate-rich Neoproterozoic successions. In successions dominated by siliciclastic rocks, chemostratigraphy can be problematic, but if thin carbonates punctuate siliciclastic strata, useful isotopic data may be obtained. The upper Pocatello Formation and lower Brigham Group of southeastern Idaho provide an opportunity to assess the potential and limitations of isotopic chemostratigraphy in overwhelmingly siliciclastic successions. The 5000 m thick succession consists predominantly of siliciclastic lithologies, with only three intervals that contain thin intercalated carbonates. Its depositional age is only broadly constrained by existing biostratigraphic, sequence stratigraphic and geochronometric data. The lowermost carbonates include a cap dolomite atop diamictites and volcanic rocks of the Pocatello Formation. The δ13C values of these carbonates are distinctly negative ( −5 to −3), similar to carbonates that overlie Neoproterozoic glaciogenic rocks worldwide. Stratigraphically higher carbonates record a major positive δ13C excursion to values as high as +8.8 within the carbonate member of the Caddy Canyon Quartzite. The magnitude of this excursion is consistent with post-Sturtian secular variation recorded elsewhere in the North American Cordillera, Australia, Svalbard, Brazil and Namibia, and exceeds the magnitude of any post-Varanger δ13C excursion documented to date. In most samples, Sr-isotopic abundances have been altered by diagenesis and greenschist facies metamorphism, but a least-altered value of approximately 0.7076 supports a post-Sturtian and pre-Marinoan/Varanger age for upper Pocatello and lower Brigham rocks that lie above the Pocatello diamictite. Thus, even though available chemostratigraphic data are limited, they corroborate correlations of Pocatello Formation diamictites and overlying units with Sturtian glaciogenic rocks and immediately post-Sturtian successions in western North America and elsewhere.

New calmoniids from the Lower Devonian Talacasto Formation in the Precordillera of San Juan, Argentina, extend the stratigraphic and geographic ranges of a clade including Bouleia Kozlowski, 1923 and Parabouleia Eldredge, 1972. The new genus Talacastops accommodates the Lochkovian T. zarelae sp.nov. from the Talacasto Formation and a closely related species from western Bolivia (Talacastops sp.nov. A). The diagnosis of Parabouleia is broadened to include P. eldredgei sp.nov., from Lochkovian strata in the lower part of the Talacasto Formation. Calmoniids from below the Scaphiocoelia Assemblage Zone display morphological disparity that rivals later occurrences, and do not conform to a model of gradual transformation of an acastomorph ancestor. Stratigraphic range extensions based on correction for ghost lineages imply a high diversity within Calmoniidae very early in the Devonian.

A single slab of silicified silty and cross-bedded mudstone from the Katkoyeh Formation, east-central Iran yields Yutagraptus cf. mantuanus Riva which indicates a late Arenig age. The slab also bears Dictyonema ghodsiae sp. nov., regarded as the only unequivocally planktonic dendroid from post-Tremadoc strata. The sicula is surrounded by at least twelve vesicular bodies which constitute a raft of floats. These are the only certain floats yet identified in the Order Dendroidea, and one of very few cases known in the Class Graptolithina.

Four sections from the mid-Atdabanian to lowest Toyonian (middle Cambrian) along the Lena River of Siberia were sampled for carbon isotope stratigraphy. These show a mainly heavy but highly oscillatory δ13C signature for the Atdabanian to mid-Botomian interval, coincident with the major phase of invertebrate innovation. A prolonged interval of negative δ13C followed until late Toyonian times, coincident with Botomian-Toyonian mass extinctions. Eleven carbon isotope cycles are identified through the lower Cambrian, which should now be tested for their utility in global correlation and relationship to bioevents in the Cambrian explosion.

Several rock stratigraphic successions, metamorphosed and non-metamorphosed, are found to be similar and/or identical with each other across the so-called ‘Main Zagros Thrust’. Stratigraphic successions form thin allochthonous sheets carried from northeast to southwest by numerous low-angle thrust faults of either ductile to brittle-ductile type or brittle type. Similarities in lithic and faunal characteristics of the stratigraphic units and in the style of structural deformation across the ‘Main Zagros Thrust’ imply that either the suture between the Afro-Arabian and Iranian lithospheric plates is not located in the Nahavand region or, if it is, it must be buried under several thrust sheets.

Dolomite is common in parts of the Chalk Group of North Sea petroleum reservoirs. The stratigraphic distribution, microtextures, and stable isotope ratios of dolomite vary both within and between four of the North Sea chalk fields examined in this study, the Eldfisk and Tor of the Norwegian Sector, the Dan Field of the Danish Sector, and the Machar Field of the United Kingdom Sector, indicating that the conditions favourable for dolomitization occurred at different times and places in the North Sea Central Graben. Dolomitization in all of the examined fields occurred late during diagenesis, after significant compactional grain breakage. Dolomite precipitation in the Eldfisk, Dan, and Machar fields occurred in modified sea water, which had Σ18O values between the assumed Cretaceous/Tertiary seawater value −1 and +8‰ (SMOW). The enrichment in 18O was probably the product of calcite recrystallization at elevated temperatures in a low water/rock ratio system. The timing of dolomitization varied with respect to organic diagenesis; dolomite precipitation in the Eldfisk Field coincided with bacterial methanogenesis whereas dolomite precipitation in the Machar Field probably coincided with bacterial sulphate reduction. The magnesium in the dolomite may have been derived from the neomorphism of high magnesium calcite to low magnesium calcite.

Simple trace fossils are reported from three localities in central Spain within a monotonous succession of shale and greywacke and classical turbidites attributed to the Neoproterozoic (Riphean) Lower Alcudian megaunit. The Lower Alcudian strata are believed to have a complex tectonic history including deformation during the Pan-African Orogeny. Chronostratigraphic control is not available for this structurally complex succession and fossil evidence is sparse. Here we report on a low-diversity ichnofossil association including the ichnospecies (isp.) Gordia marina Emmons, 1844, Gordia isp., G. aff. arcuata Książkiewicz, 1977. The stratigraphically inferred Proterozoic age of the rocks, added to the present ichnofossil association, suggests that colonization of deeper waters started before the early Cambrian. We further consider that the plausible oxygen requirements of marine invertebrate(s) which produced the present ichnofossil taxa may well have exceeded the oxygen levels in a largely dysaerobic environment. A circulation model proposed by one of us (Palacios, 1989), implying the presence of extensive upwelling and descending oxygen-rich waters, helps to explain the continuous colonization of deep-water settings in Lower and Upper Alcudian strata. We argue that the ichnofaunal record of strata older than 650 Ma is both rare and inconclusive. Moreover, the present ichnofauna, being similar to ichnofaunal associations in environmentally comparable Upper Alcudian strata, is largely documented from Neoproterozoic (Vendian) and early Palaeozoic associations elsewhere. We conclude that the new ichnofossil evidence is consistent with a Vendian or younger age for the Lower Alcudian megaunit. This view is also consistent with a recently published maximum deposition age of 565 Ma inferred from U–Pb datings of detrital zircons from the supposedly much older (Middle Riphean) Tentudia Group or ‘Serie Negra’, that appears roughly time-equivalent with Lower Alcudian strata. It thus may appear that basin formation and flysch deposition in central Iberia encompass a minor segment of Neoproterozoic time, being probably limited to the Vendian. By default, formerly postulated Pan-African deformation of Neoproterozoic Lower Alcudian strata appears implausible. We further conclude that the complex pre-Phanerozoic structural history of Alcudian strata is probably restricted to Cadomian and younger deformation. Former structural models need substantial revision.

Quartz veins and vein-breccias in a greywacke-shale sequence of ?Carboniferous-Triassic age were previously regarded as mesothermal silicified fault breccias, and related to an adjacent Eocene granodiorite pluton. New mapping of vein assemblages and textures, and their structural and cross-cutting relationships, demonstrates that the steeply dipping, sheeted, epithermal-textured vein array was hydraulic in origin and possibly Cretaceous in age. The main vein and breccia swarm trends for 14 km NNE along-strike and 2 km across-strike, cutting large irregular areas of silicified and brecciated sandstone, and patchy areas of pyritic, propylitic and K-feldspar alteration. Angular vein fabrics and hydraulic disruption textures indicate wedging by hydrothermal solutions, hydraulic rupture, brecciation and fragment transport, followed by open-space precipitation, in veins generally < 15 cm thick and breccias up to a few metres thick. Hydrothermal quartz, chlorite, calcite and chalcedony predominate, with variable amounts of chalcopyrite, galena, sphalerite and pyrite. Epidote, arsenopyrite, K-feldspar and andradite garnet are conspicuous in places. Breccias were pre-and syn-mineralization, whereas mineral precipitation was pre-, syn- and post-breccia formation. Hydrothermal activity was simultaneous with extensional faulting, striking NNE, and accompanied by intrusion of dacitic dykes. There followed conjugate shearing on east- and ESE-striking faults, intrusion of high-level tonalite stocks, and several phases of basaltic andesite dyke intrusion. These hypabyssal rocks were probably coeval with the Antarctic Peninsula Volcanic Group dominating Livingston Island, dated between 130 and 75 Ma. Minor copper and iron sulphide-bearing veins occur in adjacent volcanic and hypabyssal intrusive rocks. The Hurd Peninsula veins may, therefore, form part of a volcanic-epithermal hydrothermal system (adularia-sericite-quartz type), of Cretaceous age, rather than a porphyry-related system of Eocene age.

Observations on pyritized graptolites have revealed a variety of pyrite morphotypes, several of which often coexist within the same fossil steinkern. The commonest forms of pyrite are those consisting of size-sorted microcrysts, which show degrees of ordering ranging from well-ordered framboids to a more homogeneous, unordered groundmass. Larger, euhedral pyrite crystals may be scattered throughout the microcrystalline groundmass, or be the dominant pyrite form in themselves. ‘Floriform’ pyrite frequently overgrows earlier framboids and may act to mould the inner periderm microstructure of the graptolite. Overpyrite occurs either as a primary ‘overspill’ of internal pyrite, or as a separate, later phase.

All of the diagenetic pyrite has light sulphur isotope compositions, indicating formation during early diagenesis; isotopic evidence of the relative timing of pyrite generations matches the morphological paragenesis. Under most conditions, available iron appears not to be a limiting factor in pyrite formation, with reactive organic matter only being limiting at the lowest concentrations encountered in these sediments. The availability of organic matter does, however, exert a control on the timing of fossil pyritization. At one locality, a change in diagenetic conditions has allowed for the formation of a morphologically late generation of pyrite with isotopically light and therefore apparently early diagenetic signatures.

Many marine fossil concentrations are considered the result of episodic sedimentological events, and in particular those due to storms. Most storm or tempestite concentrations are identified as autochthonous or parautochthonous assemblages created by a variety of winnowing processes within shallow water environments. In contrast, samples described here from both a ‘shelf’ and a ‘basinal’ setting within the Ludlow (Upper Silurian) succession of the Welsh Basin reveal the presence of a biota transported by tempestite activity into a setting dominated by a more offshore biota. Tempestite horizons from within an ‘outer shelf’ mud dominated setting include shelly lenses with a transported fauna abounding in gastropods, tentaculitids and atrypid brachiopods, the background sediment being rich in graptolites, cephalopods and small strophomenid brachiopods. Within the ‘basinal’ area, distal tempestites range from minor siltstone layers to thicker bioclastic limestone lenses. The siltstones are largely graptolitic (dominated by Bohemograptus), with some small brachiopods, whilst Saetograptus colonus is the only common graptolite in the limestones, which also contain a fauna of broken brachiopods and bryozoa. The transport of assemblages distally into a variety of settings represents a potential source of error in palaeoecological analysis. Transported assemblages may, however, provide evidence of the composition of both benthic and pelagic shallower water faunas no longer known in situ.

Thin, subplanar or parallel-sided syntectonic quartz veins from the Hodgkinson Province, North Queensland, Australia contain mica inclusions that are identical to wallrock mica textures. These ghost textures can only be produced by incomplete replacement of the wallrock by quartz. The planar form of the veins indicates that there was a planar control on their formation. We propose that the veins formed around syntectonic fractures that localized silica micrometasomatism of the fracture wall. The formation of the veins can be explained by a model which involves fracture of the wallrock, reaction between the fracture wall and the ambient metamorphic fluid and eventual sealing of the fracture by precipitation from the fluid. The new vein is then a locus of further fracture. Replacement of wallrock micas occurs via a two-stage process where biotite is replaced by muscovite, which is in turn replaced by quartz. We propose the term crack–reaction to describe the resulting cyclic process because it can be compared with the crack–seal model of vein formation. Crack–seal and crack–reaction are different only in the relative amounts of metasomatism and precipitation that occur subsequent to fracture and can be envisaged as end members of a more general vein-forming process.