Palaeozoic volcanism in the Avalon Terrane of northern Nova Scotia occurred during three time intervals: Cambrian–early Ordovician, late Ordovician–early Silurian and middle–late Devonian. In the Meguma Terrane of southern Nova Scotia, Palaeozoic volcanism is limited to the middle Ordovician. Geochemical data show that most of these volcanic rocks are bimodal, within-plate suites. Initial εNd signatures range from +5.4 to −1.9 in the rhyolites and +6.8 to +2.7 in the basalts, a difference attributable to the absence or presence, respectively, of a significant crustal component. The data and regional tectonic settings of the Avalon and Meguma terranes suggest that the volcanism was generated in three different within-plate settings: (1) Cambrian–early Ordovician volcanism related to thermal decay of late Proterozoic arc magmatism during transtensional deformation; (2) middle Ordovician–early Silurian volcanism during sinistral telescoping between Laurentia and Gondwana where extensional bends in the Appalachians produced rifting; and (3) Devonian volcanism resulting from lithospheric delamination during dextral transpression and telescoping. In each setting, active faults served as conduits for the magmas. Nd isotopic data indicate that the source of the Palaeozoic felsic volcanic rocks is isotopically indistinguishable beneath southern and northern Nova Scotia and did not substantially change with time. This crustal source appears to have separated from the mantle during the Proterozoic, a conclusion consistent with the hypothesis that the Palaeozoic rocks in Nova Scotia were deposited upon a late Proterozoic oceanic–cratonic volcanic arc terrane. The Nd data, when combined with published faunal, palaeomagnetic and U–Pb isotopic data, suggest that the Avalon Terrane was peripheral to Gondwana off northwestern South America during Neoproterozoic and early Palaeozoic times.

The Agrio Formation of the Neuquén Basin, Argentina, contains an extensive sequence of ammonite faunas, most of which are monogeneric. Detailed collecting through 15 sections across the basin has facilitated a major revision of the ammonite zonation. Formerly embracing four broad zones, the Agrio Formation is now divided into nine zones, the lowest four of which are divided into a total of 11 subzones. The new zonation provides a standard against which other South American faunas can be compared. The degree of subdivision now achieved is comparable to that for the ‘standard’ sequences of the West Mediterranean region. Although it is not possible to correlate the two regions in detail, the occurrence of some widely distributed genera (Olcostephanus, Karakaschiceras, Oosterella, Spitidiscus and Crioceratites) at well-defined levels in the Neuquén Basin provides some crucial links. Thus the approximate positions of the Lower/Upper Valanginian, Valanginian/Hauterivian and Lower/Upper Hauterivian boundaries can be determined.

The stratigraphy of the Siktefjellet and Red Bay groups of northwestern Spitsbergen is revised in the light of new fieldwork. An early Early Devonian (Lochkovian) age is indicated by the rich fossil vertebrate faunas of the Red Bay Group. The underlying Siktefjellet Group is likely to be Late Silurian in age, although there is no direct fossil evidence of this. The most abundant sediment type in the two groups is fine-grained fluvial sandstone, although conglomerates and lacustrine siltstones also occur. The conglomerates that predominate in the lowest parts of both groups imply the presence of fault scarps at the surface. There has been much debate about the importance of strike-slip faulting in the area, particularly in relation to the major north–south trending Raudfjorden and Breibogen faults. Between these faults, the Biskayerhuken–Holtedahlfonna terrane contains a zone for which the deliberately genetic name ‘Siktefjellet Strike-slip Zone’ is proposed. This zone is characterized by evidence for a succession of events that, taken together, suggest a history of continuing, repeated strike-slip. These events were: (1) the tectonism and metamorphism of the basement (Krossfjorden Group) and the emplacement of the high-grade Richarddalen Complex, (2) development of crush zones in the basement, (3) deposition of the Siktefjellet Group in two distinct basins (perhaps of pull-apart origin), (4) deposition of the Red Bay Group (first as discrete scarp-related conglomerates, and then as part of a more stable fluvial basin), (5) localized compressional deformation of the Siktefjellet and Red Bay groups.

Late Ordovician volcanic rocks of the English Lake District typically have a magnetic remanence dominated by a single characteristic component. Previous investigations have interpreted this remanence as both of primary (pre-folding) and secondary origin. Palaeomagnetic field tests have been conducted on (a) andesite blocks from an autobrecciated lava top, (b) andesite blocks in mass-flow breccias, and (c) fault-blocks tilted during Ordovician caldera collapse to establish the time of remanence acquisition. All three tests show that the lavas retain a magnetization acquired during initial cooling: magnetizations of the breccias are coherent within clasts and random between clasts, whilst magnetizations of the tilted fault blocks converge with better that 95% confidence when corrected for the effects of caldera collapse. In contrast, the volcaniclastic sedimentary and pyroclastic rocks possess an Ordovician secondary remanence acquired after strata had been tilted by volcano-tectonic subsidence. A distributed sample of 65 andesite and basalt sheets through the Borrowdale Volcanic Group has a mean remanence direction D/I=341.9/−48.9° (α95=4.0°) yielding a positive fold test and a palaeomagnetic pole at 12.7°E, 4.3°S (dp/dm=3.5/5.3°). A progressive steepening of the palaeofield direction is recorded during emplacement of the Borrowdale Volcanic Group (∼I=−39° to I=−51°) which continued into the interval of volcanotectonic overprinting (I=−62°); the equivalent motion of Eastern Avalonia is ∼20° into higher southerly latitudes.

Both the Eycott and Borrowdale volcanic groups exhibit uniform normal polarity throughout. Correlation with the geomagnetic time scale for the Ordovician restores the broad correlation between the two groups by constraining their emplacement and partial overprinting to a single long normal polarity chron occupying the Nemagraptus gracilis and earlier part of the Diplograptus multidens biozones (late Llandeilo and early Caradoc). All the volcanism, therefore, occurred within a period of no more than ∼5 Ma. The palaeomagnetic evidence confirms that the Borrowdale Volcanic Group was affected by both syn-volcanic deformation (caldera collapse) and regional compressive deformation prior to deposition of the (late Ordovician–Silurian) Windermere Supergroup. The succession of primary and secondary Ordovician palaeomagnetic poles from the Lake District inlier defines an anticlockwise apparent polar wander (APW) loop with the apex correlating with ‘soft’ closure of the Iapetus Ocean and late Ordovician deformation. The APW paths from Avalonia and Baltica converge at this point as subduction ceased and the arc subsided beneath the sea after mid-Caradoc times.

Fore-arc magmatic sequences associated with high Mg number andesite lavas unconformably overlie LeMay Group accretionary complex in Alexander Island. High-resolution 40Ar/39Ar, U–Pb zircon, fission track and K–Ar ages demonstrate that subduction-related fore-arc magmatism migrated northwards along the length of Alexander Island between c. 80 Ma and c. 46 Ma. The magmatic rocks represent a third of the western margin of the Antarctic Peninsula magmatic arc and are critical to the understanding of the final phase of subduction along the southern Antarctic Peninsula margin. The onset of late Cretaceous magmatism is recorded by poorly exposed volcanic rocks on Monteverdi Peninsula (79.7±2.5 Ma). In central and northern Alexander Island, the magmatic rocks can be distinguished by the proportion, range and types of lithofacies present, and by the periods of magmatism represented. The volcanic rocks of the Colbert Mountains range in age from c. 69–62 Ma and are dominated by large volume dacitic and rhyolitic crystal-rich ignimbrites interpreted as caldera-fill deposits. Elgar Uplands sequences range in age from c. 55–50 Ma, and contain approximately equal proportions of pyroclastic deposits and less evolved (basaltic-andesite and andesite) lavas including high Mg number andesite lavas near the base of three sequences. The volcanic rocks of Finlandia Foothills probably represent the youngest calc-alkaline units on Alexander Island (48±2 Ma). The sequence is lithologically similar to the Elgar Uplands and also contains high Mg number andesite lavas, but it is dominated by polymict conglomerates, with minor lavas, which were deposited in a graben associated with regional extension. Plutonic rocks exposed in the Rouen Mountains, adjacent to the Elgar Uplands, yielded a U–Pb age of 56±3 Ma which is in discordance with a previously published Rb–Sr age (46±3 Ma), probably due to hydrothermal perturbation of the Rb–Sr system. Northwards migration of magmatism was caused by the progressive collision and subduction of three ridge segments prior to the previously reported ridge crest–trench collisions that occurred c. 20–30 Ma later and following which subduction ceased.

The catastrophic slope failures and landslides that occur in the final stages of lateral collapses of volcanoes destroy much of the evidence for precursory deformation and the early stages of the collapses concerned. Aborted or incomplete collapse structures, although rare, are rich sources of information on these stages of development of catastrophic collapses. The San Andres fault system, on the volcanic island of El Hierro, is a relatively young (between about 545 and about 261–176 ka old) but inactive lateral collapse structure. It appears to represent an aborted giant landslide. It is developed along the flank of a steep-sided volcanic rift zone, and is bounded by a discrete strike-slip fault zone at the up-rift end, closest to the centre of the island. This geometry differs markedly from that of collapse structures on stratovolcanoes but bears some similarities to that of active fault systems on Hawaii. Although the fault system has undergone little erosion, cataclasites which formed close to the palaeosurface are well exposed. These cataclasites are amongst the first fault rocks to be described from volcano lateral collapse structures and include the only pseudotachylytes to have been identified in such structures to date. Their development at unusually shallow depths is attributed to large movements on the fault in a single event, the inferred aborted landslide, and a lack of pressurized pore water. The absence of pressurized fluids in the slumping block may have caused the San Andres fault system to cease moving, rather than develop into a giant volcanic landslide. The recognition that the San Andres fault system is inactive greatly reduces the estimated volcanic hazard associated with El Hierro. However, the lack of evidence for precursory deformation prior to the aborted landslide event is disturbing as it implies that giant lateral collapses can occur on steep-sided oceanic islands with little warning.

The late Caledonian Galway Batholith is cut by two major faults which divide it into three separate areas: the western, central and eastern blocks. The upthrown and more deeply eroded central block is bounded by these faults, in the west by the north–northeast trending Shannawona Fault and in the east by the north–northwest trending Barna Fault. We present new granite field relations from part of the central block (Inveran sector) which are fundamental in establishing the order and style of emplacement for the granites of the central block and the batholith as a whole. Unequivocal field evidence from the Inveran sector indicates upward movement of early central block granites which then became the solid roof rocks to subsequent intrusions. In the case of the Knock Granite these earlier intrusions were block stoped. We use this field evidence to review the geology of the central block in a 200 km2 area that incorporates the previously mapped Costelloe and Spiddal areas. Sharp intrusive contacts are a predominant feature of this sector of the central block and are in marked contrast to the gradational contacts recorded elsewhere in the batholith. Whereas juxtaposition of plutons in the western block occurred as the granites were partly crystallized, the central block reveals earlier, deeper level granites that were consolidated by the time they were intruded by late-stage higher level granites.

A scanning electron microscope based cathodoluminescence technique utilizing a novel collector system reveals complex internal heterogeneities within granitic quartz grains. The technique overcomes the low intensity and limited variation in cathodoluminescence generated by quartz, which hamper conventional cathodoluminescence analysis. Detailed images of zoning patterns in quartz are comparable to those observed in minerals such as feldspar, and attributed to a combination of progressive growth, boundary layer effects and mineral–melt disequilibria produced during fluctuations in melt composition and temperature during the crystallization interval. We attribute such mineral–melt disequilibria to open system, mixing behaviour in the granite plutons sampled.

A very large pterosaur cervical vertebra is described from the Upper Maastrichtian deposits of Mérigon, in the foothills of the French Pyrenees. It resembles the vertebrae of Quetzalcoatlus, from the Maastrichtian of Texas, more than those of Arambourgiania, from the Maastrichtian of Jordan. The estimated wing span of the Mérigon pterosaur is close to 9 m, which makes it one of the largest known flying creatures. Giant pterosaurs still had a wide geographical distribution at the end of Maastrichtian time, which is not suggestive of a declining group, although it is difficult to obtain an accurate estimate of taxonomic diversity of terminal Cretaceous pterosaurs on the basis of available data.

Miocene subtidal sandwave deposits in southwest Japan were influenced by periodic flow and steady flow. The sandwave deposits can be divided into five units, based on lithofacies and thickness. In order of accretion, unit 1 consists of unidirectional sand bedforms without mud drapes, unit 2 of unidirectional sand bedforms with thin, discontinuous mud drapes, unit 3 of bidirectional sand bedforms with thin continuous mud drapes, and units 4 and 5 of relatively thinner and smaller bidirectional sand bedforms with continuous mud drapes. The thickness of units 1 to 3 increase progressively to 2.6 m, and units 4 to 5 subsequently decrease from 2.0 to 1.0 m. Variations between the units are due to differing combinations of periodic and steady flow velocities. Palaeoflow velocity is estimated from grain size and unit thickness. Depth-mean velocities of steady flow components gradually decrease from 0.72 ms−1 to 0.16 ms−1 with unit accumulation.

B. J. Bluck & J. K. Ingham comment: The status of the fossiliferous rocks at the Highland Border (the Highland Border Complex) has been a matter of speculation for over a century. The question has been whether the Highland Border Complex has affinities with the Dalradian block or constitutes a separate block, a sliver or slivers of which have been caught up along the Highland Boundary Fault.

M. J. M. Turnbull & S. Moorbath comment: The paper of Yang et al. (1996) describes a novel dating approach, using the Sm–Nd method on phosphatic material, to determine the age of the Precambrian–Cambrian boundary. The dates reported are of a precision which, in light of the relatively small range of 147Sm/144 Nd ratios (0.09–0.17), is surprising. We have recalculated the published data using the ISOPLOT program of Ludwig (1991) and this is shown below in Table 1. These results cast considerable doubt on the statistical analysis of Yang et al. (1996). While this technique may have utility with Precambrian successions it would appear to be of limited use for the high precision dating required for Phanerozoic stratigraphic correlation.