In a recent issue of this journal, Crampin (1998) suggested that analysing seismic shear-wave splitting along appropriate ray paths can be used to monitor the build-up of stress before earthquakes. If the source and recording geometry are suitable, this procedure, known as stress-forecasting, allows the approximate time and magnitude, but not the location, of future large earthquakes to be estimated.

Lower Cambrian shallow-water carbonates enclosing microbial structures are documented for the first time from the upper Lastours Member of the Montagne Noire (southern France). Microbial organisms constructed self-supported boundstones resulting in the formation of exclusively microbial-accreted buildups, which exhibit three main types of non-skeletal microbialites: planar stratiform stromatolites, dome-shaped stromatolites and nonlaminated (thrombolitic) biostromes. In addition, thrombolitic boundstones display four distinct microbial microstructures: clotted and Renalcis-like forms, branching bushy forms, clusters of unbranching straight filaments and crustose forms.

The upper member of the Lastours Formation records an upward transition from a shalydominant open shelf to a protected shelf environment bounded by a surface representing a major subaerial exposure. Initially, at the inception of the highstand systems tract, flat stratiform stromatolites formed on open sea subtidal shaly substrates, while stacked domal stromatolites developed in peritidal areas which record subaerial exposure. In contrast, prograding shoal barriers of the transgressive systems tract favoured the establishment of thrombolitic boundstones in protected (back-shoal) environments.

The Devono-Carboniferous chondrichthyan Stethacanthus is best known for its striking 'spine-brush' complex. Here we provide the first detailed description of the neurocranium, using several well-preserved specimens from Bearsden, Scotland (Serpukhovian; Pendleian E1). Significant features include an otico-occipital region of about the same length as the ethmosphenoid division; short but distinct lateral otic processes; a short, broad endolymphatic fossa; a persistent otico-occipital fissure; Y-shaped basicranial canals indicating origin of the lateral aortae anterior to the occipital level; and minimal suborbital shelves limited to small, ethmoidally located outgrowths. Using placoderm, osteichthyan and acanthodian neurocrania for outgroup comparison, PAUP analysis of a matrix of 23 neurocranial characters produced the following branching sequence: (Helodus (Stethacanthus, Cobelodus (Tamiobatis (Xenacanthus (Cladodus wildungensis (Tristychius, Hybodus)))))). The location of Cladoselache in this analysis is uncertain and probably results from missing data and a morphologically restricted character set. Plesiomorphic conditions for chondrichthyan braincases are found to include the presence of a persistent otico-occipital fissure, projection of the median dorsal aorta anterior to the occipital level, and absence of suborbital shelves. These results may have important implications with regard to primitive gnathostome interrelationships, and especially the phylogenetic position of placoderms.

Interferometry, a technique widely used in physics to study radiation fields, has been applied to the investigation of dowsing. Pairs of horizontal linear components, typically some metres in length and a metre or so apart, and parallel in a vertical plane, produce patterns of lines that can be mapped by dowsing rods. The lines are parallel to each other and to the components, and are equally spaced. The spacing of the lines depends on the length, separation and composition of the components. The existence of the patterns shows that the dowsing field is wave radiation. Since there is no detectable difference between patterns produced by pairs of components made of good electrical conductors such as copper, and those produced by good insulators such as uPVC, it is concluded that the radiation field is not electromagnetic. Furthermore, aluminium and tin behave differently from copper and steel. The spacing of the lines has been found to depend on time of year. It increases rapidly in late April and falls rapidly in late November; in between these times it appears to be relatively stable except for an isolated maximum in early March. It shows a high degree of repeatability from year to year.

The lower jaw of the Devonian tetrapod Acanthostega is described for the first time. Redescriptions are provided for the lower jaws of the elpistostegid Panderichthys, the Devonian tetrapods Elginerpeton, Obruchevichthys, Metaxygnathus, Ventastega and Ichthyostega, and the Carboniferous tetrapods Crassigyrinus, Megalocephalus and Gephyrostegus. The character distri- butions thus revealed differ considerably from previous accounts, particularly in the wide distribution of certain primitive characters. Meckelian ossification in the middle part of the jaw is widespread among Devonian tetrapods, being demonstrably absent only in Acanthostega. Among Carboniferous tetrapods, a tooth-bearing parasymphysial plate is shown to be present in Crassigyrinus and Megalocephalus (having already been demonstrated by other authors in Whatcheeria and Greererpeton). A phylogenetic analysis of 26 early tetrapods including all the aforementioned genera, scored for 51 lower jaw characters, produces at least 2,500 equally parsimonious trees. However, the lack of resolution lies largely in a big top end polychotomy containing anthracosaurs, temnospondyls, seymouriamorphs, microsaurs and a nectridean-amniote clade. Below this polycho- tomy, which may correspond approximately to the tetrapod crown group, there is a well-resolved stem-group containing, in descending order, Megalocephalus, Greererpeton, Crassigyrinus, (jaws associated with) Tulerpeton, Whatcheeria, Acanthostega, Metaxygnathus, Ichthyostega, Ventastega and Metaxygnathus (unresolved), an Elginerpeton-Obruchevichthys clade, and Panderichthys. This conflicts with recently published phylogenies by Coates and Lebedev & Coates, which place Tulerpeton and all post-Devonian tetrapods in the amphibian or amniote branches of the tetrapod crown group.

A new temnospondyl, Thabanchuia oomie gen. et sp. nov., is described from three skulls and associated postcranial elements from the Early Triassic Lystrosaurus Zone of South Africa. T. oomie is the best-preserved member of the Tupilakosauridae, a temnospondyl taxon previously known only from disarticulated material from Russia and Greenland, but shown here to include Kourerpeton bradyi, an articulated specimen with no locality data but presumed to be from North America. Tupilakosaurids are the only Mesozoic survivors of the dvinosaurian radiation, the sister group to the archegosaurs of the Permian plus the stereospondyl clade. It is postulated that their occurrence in Russia, Greenland and North America represents a reinvasion from Gondwana, where they survived the Late Permian extinctions along with the stem of the stereospondyl clade. T. oomie is probably immature but, like other members of the dvinosaurian clade, was fully aquatic, had ossified ceratobranchials and most likely an increased number of vertebrae and reduced limbs. Like Tupilakosaurus, T. oomie had diplospondylous centra.

The tholeiitic, columnar (Staffa type) lavas found at, and near, the base of the Tertiary lava succession on the Isle of Mull contrast markedly with overlying transitional tholeiiticalkalic Mull Plateau Group lavas, and as a result previous workers have classed the Staffa type lavas as a distinct group. The Staffa type lavas are among the most contaminated in the British Tertiary Igneous Province and have equilibrated and fractionated at shallow depths within the crust (relative to the overlying Plateau Group lavas), so obscuring the chemical identity of their parental magma. However, Staffa type lavas which are more basic than examples hitherto reported help in the elucidation of the parental magma. It is demonstrated that the Staffa type lavas were derived from a parental magma which was compositionally similar to that of Plateau Group lavas. The primary reasons for the distinctive features of the Staffa type are related to post-mantle melt generation processes, i.e. to extensive crystal fractionation and to crustal contamination at shallow crustal depths. The lavas are compositionally and temporally unrelated to the Hebridean tholeiitic magma types, the Central Mull Tholeiite type and the Skye Preshal More type.

The Bail Hill Volcanic Group (Caradoc) represents the largest, single volcanic complex exposed within the Ordovician turbidite succession of the Northern Belt in the Southern Uplands of Scotland. The group comprises a heterogeneous sequence of submarine lavas, volcaniclastic and intrusive rocks (up to 2 km thick), and crops out in a small area (c. 4 km2) around Bail Hill, north of Sanquhar. The Cat Cleuch Formation (older) is dominated by a sequence of autobrecciated basaltic lavas which contain large, zoned diopsidic clinopyroxene. The overlying Peat Rig Formation comprises a more mixed sequence of plagioclase-amphibole-phyric lavas, volcaniclastic rocks and contemporaneous volcaniclastic sedimentary rocks. The Cat Cleuch and lower part of the Peat Rig formations are cut by a vent breccia, the Bught Craig vent breccia, which formed part of the feeder to the upper part of the Peat Rig Formation. The Bail Hill volcanic rocks are alkaline in character, ranging from alkali basalt to trachyandesite in composition, possessing trace element characteristics and enrichment patterns typical of oceanic within-plate basalts. The Bail Hill Volcanic Group, although geochemically distinct, forms part of a mixed assemblage of tholeiitic and alkaline withinplate lavas within the Southern Uplands which are of broadly similar age, some of which are intercalated within the greywacke sandstone sequence. This assemblage clearly indicates that a period of extension and within-plate volcanism occurred during the early stages of the development of the Southern Uplands sedimentary basin.

Sub-silicic to silicic pitchstones are widespread throughout the British Tertiary Igneous Province (BTIP), with examples being found at all the major igneous centres. Both highly porphyritic and almost completely aphyric varieties occur, and take the form of sills, dykes and lava flows. Here we present previously unreported mineral chemistry data on phenocryst and microcrystallite populations from a number of pitchstones from throughout the BTIP. Phenocryst assemblages are completely anhydrous, comprising mixtures of plagioclase, sanidine, fayalite, orthopyroxene, pigeonite, ferroaugite, ferrohedenbergite and quartz. Microcrystallite assemblages are also diverse, consisting of sanidine, ferrohedenbergite, fayalite and, occasionally, almost pure end-member ferrosilite, as well as hydrous phases such as ferrohornblende and biotite. Textural and mineral chemistry observations support interpretations derived from whole-rock and residual glass major element analyses, together with whole-rock trace element and the available Sr-Nd-Pb isotope data, that the Tertiary pitchstones of Scotland are either the products of intimate mixing between a range of basaltic magmas with hydrous crustal melts, or were formed by the crustal contamination of basaltic magmas.

Triarthrinid trilobites are described from the Middle Ordovician Andersön Shale and Örån Shale of Jämtland, central Sweden. The faunas include at least five species, belonging to two genera, Porterfieldia Cooper 1953 and Triarthrus Green 1832. Porterfieldia is represented by a single species, Porterfieldia humilis (Hadding 1913), while at least four species belong to Triarthrus: T. freji Thorslund, 1940, T. jemtlandicus Linnarsson, 1875, T. latissimus nom. nov. and T. linnarssoni Thorslund, 1940. P. humilis is found only in the lower Andersön Shale, equivalent to the Hustedograptus teretiusculus Zone. T. freji, T. jemtlandicus and T. latissimus occur frequently in the upper part of the Anderson Shale, which comprises the Nemagraptus gracilis Zone, and probably also the lower part of the Diplograptus multidens Zone. T. linnarssoni occurs throughout the Örån Shale (mainly Dicranograptus clingani Zone). T. jemtlandicus and T. aff. linnarssoni are also sporadically present in the lower and middle parts of the Oran Shale. Almost all growth stages of T. latissimus, from the anaprotaspid stage and onwards, are represented and the ontogeny is described. The early stages are, however, comparatively uncommon.

A new chasmataspid (Chelicerata, Chasmataspida) is described as Forfarella mitchelli gen. et sp. nov. from the Lower Devonian of Scotland. This species is problematic in that it was catalogued under the name Forfarella mitchelli Waterston (no date), but was never formally described. Furthermore, its type locality is uncertain. Chasmataspids are a poorly known taxon within Chelicerata which appear to be closely related to eurypterids. F. mitchelli is only the fourth chasmataspid species to be described. Though poorly preserved, it shows affinities with other Devonian chasmataspids and supports the hypothesis that this group may have been more geographically widespread in the Devonian than has previously been recognised.

Specimens of Kerygmachela kierkegaardi Budd are described, from the Lower Cambrian Sirius Passet fauna of N Greenland. The cephalic region is characterised by a pair of stout unsegmented appendages each bearing long spinose processes, and an anterior mouth. The trunk shows alternating rows of tubercles and transverse annulations along the axis, to which are attached 11 pairs of gill-bearing lateral lobes and lobopodous limbs. The caudal region is small, and bears two long tail spines. There is some evidence for circular musculature arranged around the trunk and a dorsal, longitudinal sinus, and several details of the muscular pharynx have been preserved.

The combination of characters found in Kerygmachela allows it to be allied with the lobopods, represented in the extant fauna by the onychophorans, tardigrades, and possibly the pentastomids, and in the Cambrian fossil record by a morphologically diverse set of taxa, some of which are not assignable to the extant groupings. It also shares important characters with the previously problematic Burgess Shale forms Opabinia regalis Walcott and AnomalocarisWhiteaves, and the Sirius Passet form Pambdelurion Budd. These taxa together form a paraphyletic group at the base of the clade of biramous arthropods. The position of the so-called ‘Uniramia’ remains unclear. It can be demonstrated from the reconstruction of the arthropod stem-group that full arthropod segmentation has a different derivation from that of the annelids. In line with other recent analyses, this suggests that the ‘Articulata’ of Cuvier should be dismantled, and the arthropods considered to be a group of protostomes which are phylogenetically distinct from the classic spiralians. Arthropod affinities may rather lie with the other moulting animals, in the so-called ‘Ecdysozoa’.

The Pliocene Hipparion from La Gloria 4, La Calera, Orrios 1 and 3, Villalba Alta Río 1, Villalba Alta (Teruel basin) and Layna (Soria province), all in Spain, and Dorkovo (Sofia, Bulgaria), are described and tested by multivariate analysis. These Spanish and Dorkovo forms are compared with the specimens from Pavlodar (large-sized) and Kalmakpai (both in Kazakhstan), Çalta (large-sized, Turkey), the Roussillon (France) and Alcoy (Spain). The faunal assemblages indicate an Early and Late Alfambrian age (Ruscinian, MN14 and MN15) for the Spanish Hipparion and Early Alfambrian (MN14) for Dorkovo specimens. Hipparion remains from Pliocene Spanish localities show a very similar structure. They are referred to Hipparion fissurae, except Orrios 1 assigned to Hipparion aff. H. fissurae and Villalba Alta classified as Hipparion cf. H. fissurae. Despite certain relationships between Hipparion from Layna and Pavlodar, the multivariate analysis confirms the independence of both forms: H. fissurae in Layna, and H. longipes in Pavlodar. The specimens from Kalmakpai present many features close or intermediate to Layna and Pavlodar forms. Finally, Hipparion from Dorkovo is determined as Hipparion cf. H. crassum. These remains have a great morphological resemblance to H. crassum from Roussillon.

Conventional rock classification diagrams do not distinguish the variety of peraluminous rock series. Moreover, peraluminous granite types have not been clearly discriminated in recent revisions. The study of several peraluminous series in different intracontinental orogenic belts reveals that four distinct groups can be defined. Using an A-B diagram, these four groups are: (1) highly peraluminous granitoids (hP) characterised by high A values and typified by an increase in peraluminosity toward the most mafic varieties; (2) moderately peraluminous granitoids (mP) which occupy the intermediate field and generally show increasing peraluminosity towards the most felsic varieties; (3) low peraluminous granitoids (IP) which plot in the lowest part of the peraluminous field defining negative slope trends; (4) highly felsic peraluminous granites (fP) with poorly defined variation trends.

In intracontinental orogenic belts, the genesis of peraluminous granitic series is favoured by the abundance of fertile crustal protoliths, mainly metapelites, metaigneous rocks and metagreywackes. The difficulty of attaining temperatures in excess of 950°C at lower crustal levels during the tectonothermal evolution of thickened crust, inhibits the partial melting of more basic sources. Although the physical parameters of the melting process influence their chemical and mineralogical characteristics, source rock composition ultimately determines the degree of peraluminosity of the granitic series.

Both textural maturity and structure acquired by gravels on beaches are largely a response to the interaction between beach surface and wave-backwash energy. The turbulent energy driving the processes of particle rejection andselection at the surface of growing gravel sheets is partly controlled by the grain size of the sheet itself, so the process is to a large extent self-regulating. Beach surfaces are seen to comprise many discrete sheets of gravel competing for growth at their boundaries, but each characterised by a uniform or uniformly gradational texture.

There are two main types of gravel sheet: (1) selection pavements which occur on low beach slopes, showing little areal grain-size or grain-shape variation, and (2) turbulence shadows which occur on steeper slopes and produce assemblages of clasts which may show perfect lateral shape and size gradation.

The clasts which make up these various gravel sheets are termed ‘clast assemblages’, and such assemblages are the fundamental units from which beaches are constructed. Clast assemblages are classified in terms of their textural maturity—the degree to which they exhibit uniformity in clast size and shape. In beach sections they are, either singly or in combination, bounded by planes of discontinuity (bedding planes) to form beds.

Repeated combinations of either clast assemblages or beds in a genetic association are regarded as sedimentary structures, many of which are diagnostic of the gravel beach environment. Growth of beaches involves the stacking of sedimentary structures, and four growth patterns have been identified. The beach structure is, therefore, classified in a hierarchy comprising clast assemblage, bed, structures and growth form, and it is the growth pattern which may be related to tidal range. P.ecognition of the processes which generate beach gravels through the structure of the gravels permits an analysis of the internal structure of major gravel bodies such as barrier beaches. It adds another set of criteria which may further lead to an understanding of the processes responsible for the generation and evolution of these large gravelforms.

Three types of gravel lithosomes have been identified. (1) Regressive barrier bars which form a series of gravel ridges separated by lagoonal deposits. Barriers are built initially by swash bars which grow in size and coarsen through time to become wave-resistant forms. They form as a response to times when the sediment, unable to be evenly distributed and sorted on the beach surface, forms a discrete bar seaward of the active beach. This is the result of a punctuated or continuously high sediment supply compared with the wave energy available to disperse the sediment, or a falling sea level which shifts the locus of sediment accretion. (2) In contrast, regressive (prograding) gravel sheets form as a response to a continuous supply of sediment to the beach surface, allowing it to build seaward by the uniform accretion of sediment which is sorted and retained on its surface. (3) Gravel sheets produced in transgression are characterised by an abundance of spherical clasts and are often overlapped by the sand beaches which occur seaward of them.

A new occurrence of the Hirnantia brachiopod fauna is documented from the Criadero Quartzite of Almadén, Ciudad Real Province, Spain. This unit is the regional development of a largely unfossiliferous sandy facies that frequently overlies the typical Late Ordovician diamictitic glaciomarine formations in the Iberian Peninsula and the Armorican Massif. The new occurrence establishes palaeontologically the latest Ashgill age of the quartzite, at least for its lowest horizons, and adds new data on a fauna that, although widespread, has been very rarely documented from peri-Gondwanan Europe. The new collection contains only Hirnantia sagittifera and Plectothyrella crassicosta chauveli. The subspecific status of the latter and its inclusion within Plectothyrella crassicosta is discussed herein, based on the continuous variation in rib thickness of several samples of both forms. The extremely low diversity and the occurrence of the key form P. c. chauveli, are both typical of the Bani Province that developed on the subpolar margins of Gondwana. This contrasts with other occurrences of the Hirnantia Fauna in peri-Gondwanan Europe, such as those from Sardinia and the Carnic Alps, which are characteristic of the more temperate Kosov Province.

Self-organised criticality of the crust appears to make deterministic earthquake prediction of time, place and magnitude of individual large earthquakes inherently impossible. This closes one line of approach to mitigating earthquake hazards. This paper suggests that a viable alternative to earthquake prediction is monitoring the build-up of stress before a large earthquake can occur. A new understanding of rock deformation allows stress changes to be monitored with seismic shear-wave splitting (seismic birefringence). With a suitable monitoring installation, this would allow the stochastic proximity of impending earthquakes to be recognised so that earthquakes could be forecast in the sense of recognising that crustal deformation was preparing for a large earthquake. Such stress-forecasting is not prediction, but, in many circumstances, a possible forecast crescendo of increasing urgency is exactly what is needed to best mitigate hazard to life and property.