The Stac Fada Member, which forms part of the Stoer Group, contains the only definitive evidence of volcanic activity within the thick Torridonian succession. At the type area in the Bay of Stoer, four volcanic-rich sandy mudstone units are recognised in the Stac Fada. An irregular erosional contact at the base of the main Stac Fada unit truncates two thinner layers of similar volcanic-rich material that were formerly interpreted as intrusions. Directional sedimentary structures associated with the Stac Fada mass flows, and the distribution and thickness of an accretionary or armoured lapilli tuff, support the existence of at least two volcanic sources. Geochemical evidence suggests that the thick Stac Fada unit at the Bay of Stoer and comparable units to the SW were derived from the same volcanic centre. The chemical composititon of volcanic fragments in the main Stac Fada unit indicates that their potassic nature is not a primary feature but is due to alteration and subsequent metasomatism.

Changes in Holocene (Flandrian) relative sea levels and coastal geomorphology in the lower Cree valley and estuary, SW Scotland, are inferred from detailed morphological and stratigraphical investigations. A graph of relative sea level changes is proposed for the area. Rising relative sea levels during the early Holocene were interrupted at c. 8300–8600 14C years B.P.(c. 9400–9900 calibrated years B.P.), when an extensive estuarine surface was reached at c. −1 m O.D., after which a fluctuating rise culminated at c. 6100–6500 14C B.P. (c. 7000–7500 calibrated years B.P.) in a prominent shoreline and associated estuarine surface measured at 7·7–10·3 m O.D. A subsequent fall in relative sea level was followed by a rise to a shoreline at 7·8–10·1 m O.D., exceeding or reoccupying the earlier shoreline over much of the area after c. 5000 14C B.P. (c. 5,800 calibrated years B.P.), before relative sea level fell to a later shoreline, reached after c. 2900 14C B.P. (c. 3100 calibrated years B.P.) at 5·5–8·0 m O.D., following which relative sea levels fell, ultimately reaching present levels. During these changes, a particular feature of the coastline was the development of a number of barrier systems. The relative sea level changes identified are compared with changes elsewhere in SW Scotland and their wider context is briefly considered.

The Rosroe Formation comprises a series of Lower Ordovician (Llanvirn) conglomerates and sandstones, that lies on the southern limb of the South Mayo Trough, within the Iapetus Suture Zone of western Ireland. Trace element chemistry of granite boulders within the formation indicates a continental, rather than a volcanic arc character that can be correlated to latest Precambrian granites within the Dalradian Metamorphic Block, part of the deformed Laurentian margin. A minority of the clasts may correlate with syn-collisional granites, similar to, but older than, the Oughterard Granite of Connemara. Pb isotope compositions of K-feldspar grains within the sandstones, measured by both ion microprobe and conventional mass spectrometry, show a clear Laurentian affinity, albeit with greater source variability in the sand grains compared to a limited range in the proximal boulders. Palaeo-current indicators demonstrate dominant derivation from the NE, with a significant axial E–W flow. We propose that the Rosroe Formation records unroofing of a rapidly exhuming Dalradian metamorphic belt in North Mayo, following extensional collapse of the Grampian Orogen starting at ˜468 Ma, with minor input from a southerly arc source. The lack of metamorphic input from the S until deposition of the Derryeeny Conglomerate argues that the Connemara terrane was not positioned S of South Mayo Trough through strike-slip faulting until after the end of Rosroe sedimentation (460–443 Ma).

Xenacanthids were a very successful group of elasmobranchs that ranged from the Lower Carboniferous to the Upper Triassic. The history of discovery of the xenacanthids, which is closely connected with the history of coal prospecting in England, began with the finding of the type specimen of Xenacanthus laevissimus in the Westphalian B of the West Midlands. In this first review of British Carboniferous xenacanthids, the number of taxa, mainly erected during Victorian times, is reduced to 14 species distributed among six genera. Determinable remains are recorded from at least 96 localities in the British Isles. Unique characteristics of the Dinantian Diplodoselache suggest that the lineage to which this taxon belongs marks a dead end in xenacanthid evolution. This investigation also shows that the Pendleian Dicentrodus, formerly described as Cladodus, belongs to the xenacanthids. The occurrence of Orthacanthus cf. kounoviensis in the Pennines, also known from the German Saar-Nahe basin, the Saale depression and from Bohemia, indicates a faunal exchange between these intramontainous basins during the Carboniferous. The genus Triodus is identified from British deposits for the first time. A cladistic analysis of the xenacanthids suggests that they evolved from phoebodontid elasmobranchs. This analysis also confirms separation of the Middle Devonian Antarctilamna from a relationship with xenacanthid sharks.

The dermal skull roof of Acanthostega gunnari Jarvik is described in detail, giving new information on the anatomy, and new reconstructions of the skull in dorsal, lateral and occipital views, as well as sections through the skull at two points. The types of suture are compared with those of temnospondyls and discosauriscids, comparable animals in terms of habit and skull shape. The skull-bones of Acanthostega are much more closely integrated with each other than those of the two later groups. However, regions of relatively less firmly and more firmly sutured regions are comparable in position. This may be connected with having the lateral lines accommodated in tubes in Acanthostega rather than grooves, or some other property of the dermal bone in the later groups that reduced the need for stronger integration. The post-orbital region of the skull (skull table) is the most strongly integrated region, possibly connected with the need for attachment of occipital musculature. The skull table and dentition are the most labile of cranial features among Devonian tetrapods, whereas the snout, cheek, palate and infradentary regions remain conservative. These are for the most part the lateral-line-bearing regions of the skull. The specimens of Acanthostega found at one site on Stensiö Bjerg may form part of a year-group that had perhaps come together for spawning.

The Eocene sparid fauna (Teleostei: Percoidei) from Monte Bolca, Italy and from the London Clay, U.K. is revised based on re-examination of the type material and phylogenetic analyses of primarily osteological data. Two phylogenetic analyses, one of the Eocene taxa and a combined analysis of fossil and extant taxa, were performed. The addition of fossils to the extant data greatly increased numbers of most parsimonious trees, destabilising and obscuring basal relationships within the Sparidae. Combination of the data from fossil and extant data also affected relationships among the fossil taxa, changing some from those recovered using fossil data alone and destabilising others. Successive approximations character weighting supported the inclusion of the Eocene taxa within a monophyletic Sparidae. The genus Sparnodus, as previously conceived, is paraphyletic and is partitioned to remove the paraphyly. Five monotypic genera are recognised, including three new genera, Abromasta, Ellaserrata and Pseudosparnodus. Inclusion of the fossils in the phylogenetic analysis implies a minimum age of origin for the Sparidae of 55 Ma with most Recent sparid fauna in place no later than the Miocene, and provides further evidence that the diversification of feeding strategies occurred early on in the evolutionary history of the group.

Tertiary–Recent magmatism in the Kenya Rift Valley was initiated c. 35 Ma, in the northern part of Kenya. Initiation of magmatism then migrated southwards, reaching northern Tanzania by 5–8 Ma. This progression was accompanied by a change in the nature of the lithosphere, from rocks of the Panafrican Mozambique mobile belt through reworked craton margin to rigid, Archaean craton. Magma volumes and the geochemistry of mafic volcanic rocks indicate that magmatism has resulted from the interaction with the lithosphere of melts and/or fluids from one or more mantle plumes. Whilst the plume(s) may have been characterised by an ocean island basalt-type component, the chemical signature of this component has everywhere been heavily overprinted by heterogeneous lithospheric mantle. Primary mafic melts have fractionated over a wide range of crustal pressures to generate suites resulting in trachytic (silica-saturated and-undersaturated) and phonolitic residua. Various Neogene trachytic and phonolitic flood sequences may alternatively have resulted from volatile-induced partial melting of underplated mafic rocks. High-level partial melting has generated peralkaline rhyolites in the south–central rift. Kenyan magmatism may, at some future stage, show an increasing plume signature, perhaps associated ultimately with continental break-up.

Forty-six samples, taken from the Upper Ordovician and lower Silurian strata of the Girvan district, Midland Valley of Scotland, yield moderately well-preserved and diverse assemblages of chitinozoans. The area was chosen to study the composition of the chitinozoan assemblages, rarely described before in Scotland, which lay at the south-eastern margin of the Laurentia palaeocontinent. These assemblages are compared with those of the same age on other palaeocontinents, representing a critical time when both the Iapetus and Tornquist oceans were closing, thus forming the Caledonide and associated orogens. It is possible to correlate the chitinozoan occurrences at Girvan with those of other parts of Laurentia (Quebec area), other palaeocontinents such as Baltoscandia (Sweden, Estonia), Avalonia (Wales, Brabant Massif) and some parts of Northern Gondwana (Saudi Arabia). For the lower Silurian, it is possible to correlate with the global chitinozoan biozonation, and a calibration between the chitinozoan occurrences and the graptolite biozonation of Girvan is established. Several levels within the Balclatchie, South Shore, Three Mile, Shalloch, Wood Burn, Lauchlan, Drumyork Flags and other formations are accurately dated.

Although only about 20m thick, the Andersö Shale contains one of the most diverse, if not the most diverse, late Middle-early Late Ordovician faunas known in Baltoscandia. It includes more than 20 trilobite species, more than 20 species of other shelly fossils, about 10 graptolite species, and about 20 conodont species. Based on its lithology, its geographical position near the foreland basin margin, and the presence of trilobites of the raphiophorid association and conodonts of the Periodon-Pygodus biofacies, this formation is interpreted to represent an outer shelf-upper slope (ramp) deposit laid down in moderately deep water. The co-occurrence of some widespread and biostratigraphically diagnostic conodonts, graptolites and trilobites makes it possible directly to compare distribution patterns of these fossils, establish ties between graptolite and conodont zones, and correlate the formation with units elsewhere in Europe, North America and China. Stratigraphically and faunally, the Andersö Shale is of particular interest in straddling the Middle-Upper Ordovician Series boundary as this boundary is recognised in the new global classification of the Ordovician System.

The Rhizodontida (Pisces: Sarcopterygii) is a clade of predatory fishes from the Upper Devonian (Aztecia; ?Givetian of Antarctica) through to the Upper Carboniferous (Strepsodus; Moscovian of northern Europe and North America). They form the most basal plesion within the tetrapod stem-lineage. The mandibles were dominated by large symphysial tusks on the dentary. Not much else is known of the mandibles in primitive rhizodontids. However, later forms show several derived characters: the mandible is very deep dorsoventrally and narrow linguolabially; the coronoid fangs bear only a single fang and no other dentition; the Meckelian element was unossified, leaving the adductor fossa unfloored by bone; the prearticular produced a large dorsal process lingual to the adductor fossa, presumably for muscle attachment. These and other characters are discussed in the context of the evolution of the tetrapod stem-group.

The mandible appears to have been split into two functional units, one comprising the firmly sutured prearticular, coronoids and dentary, the other comprising the firmly sutured infradentaries. The connection between the two units was weak, suggesting a longitudinal intramandibular hinge. The possibility that this acted as a ‘torsion grip’ during feeding is discussed.

A new crustacean, Castracollis wilsonae is described from a loose block of the Early Devonian Rhynie chert, found in the vicinity of Rhynie, Aberdeenshire, Scotland. It differs markedly from Lepidocaris rhyniensis, Scourfield 1926, the only other crustacean found in the Rhynie chert.

The material comprises complete individuals up to 8 mm long and fragmentary remains. The head is normally damaged and detached from the body, and damage to the anterior of the trunk suggests the specimens are exuviae. The head appears domed with a labrum, robust mandibles, and long biramous antennae. A few specimens exhibit a probable detached cephalo-thoracicshield. The trunk is multi-segmented, comprising similar ring-like somites. The thorax variably comprises up to 26 segments. The anterior 11 segments possess similar long, phyllopodous appendages, the remainder variably possess 10 to 15 phyllopodous appendages with a disposition of one per somite. The abdomen comprises up to 28 apodous segments. The posterior of the body comprises a telson with two furcal rami.

The crustacean is most probably a calmanostracan branchiopod. It occurs associated with Lepidocaris, charophytes, cyanobacteria and coprolites within a ‘clotted’ chert texture, indicating subaqueous deposition, most probably in a temporary freshwater pool in an area of surficial hydrothermal activity.

Graphic correlation using graptolites and conodonts provides a high-resolution timescale for correlating from coastal to deep oceanic sections and, thereby, also a detailed record of the sequence of changes during the Mulde Secundo-Secundo Event. That interval includes sedimentary facies otherwise unknown in older Wenlock to early Ludlow strata on Gotland. The identified sequence of changes includes a detailed record of, in order: two extinctions (Datum points 1 and 1·5); widespread deposition of carbon-rich sediments extensive enough to cause a δ13C increase of c. 4.8‰, the onset, maximum and end of a sea-level fall and rise of at least 16 m during 30 kyr; a third extinction (Datum 2); a disaster fauna; and a slow faunal recovery. Thus, a secondary result of the event was a weakened greenhouse effect triggering a glaciation: the Gannarve Glaciation (new term). The order of changes proves that regression did not cause the extinctions. Faunal and sea-level changes, as well as the sedimentary succession, fit well with predictions based on an oceanic model. Extinctions were primarily caused by a severe drop in primary planktonic productivity, causing starvation among planktonic larvae in non-coastal settings. The Grötlingbo Bentonite (new term), the thickest in the Wenlock of Gotland, was deposited across the basin shortly after Datum 2. Temporal resolution is high enough to permit some comparison with Quaternary glaciations.

Abundant microvertebrate remains from the Siberian Platform are described as early acanthodians. All are preserved with both excellent morphology and histology. They are assigned to a new order, Tchunacanthida, with two new families, Lenacanthidae and Tchunacanthidae. These comprise two new genera, Lenacanthus and Tchunacanthus with type species L. priscus sp. nov. and T. obruchevisp.nov. The evidence from the morphology and histology is that they are the most ancient acanthodian scales so far found. The total collection of vertebrate material from the Irkutsk amphitheatre is described, together with their geological distribution, geographical range and systematic palaeontology. Head scales, tesserae of three morphotypes, transitional scales and body scales have been found. All comprise morphological sets as determined by comparison of morphology using SEM and of histology using Nomarsky interference optics. Observations of growth were possible from details of concentric lines on the crown and also from incremental layers seen in both horizontal and vertical sections. Starting from the primordial scale, consecutive layers are added, coronally, laterally, and deep in the corium onto the base; these occurred simultaneously in both crown and base. A type of areal-superpositional growth occurred in some body scales but in other scales there was little superpositional growth. It is concluded that Tchunacanthus possessed very solid body armour with tightly joined scales covering a large area of the body, and scale bases deeply set in the corium. Both tissues of the scales, dentine and bone, feature enclosed cells. This character is regarded as primitive within acanthodians, as in derived forms both tissues are acellular. Acanthodian scales are one of many examples of transformation from cellular to acellular tissue in evolution.

Two new broad-cupped camerate crinoids are reported from the Upper Ordovician (Ashgill, Rawtheyan) Lady Burn Starfish Beds of the Girvan district, Strathclyde, SW Scotland. These represent a significant addition to the limited crinoid diversity known from the Ordovician of the British Isles. The 20-armed diplobathrid Eodimerocrinites littlewoodi gen. et sp. nov. is similar to Silurian Dimerocrinites Phillips, but lacks median ray ridges and has a distinct calyx sculpture. Camerate sp. indet. differs from E. littlewoodi in the sculpture of the dorsal cup and the arrangement of the interbrachial plates, but it preserves insufficient information for it to be further classified with confidence.

New specimens of the diplobathrid gen. et sp. nov. cf Botting from the Middle Ordovician (Llanvirn, Abereiddian) of central Wales, that illustrate hitherto unknown morphological features, including a distal rhizoidal holdfast and a geniculate proximal column, suggest that this species was a rheophilic filter feeder. One specimen is one of the most complete fossil crinoids known from the Ordovician of the British Isles.

A section of the Vitabäck Clays at Eriksdal in southern Sweden was sampled for vertebrate fossils. Large bulk samples were collected from three horizons, including two coquina beds, VC3 and VC11, and a silty clay bed, VC7. Shark teeth are very common and constitute the main portion of the vertebrate material discussed herein. The selachian tooth faunas are almost exclusively represented by hybodonts, although a single tooth from a neoselachian shark, Squatinidae indet., was recorded from one horizon (VC3). Hybodont species identified from the Vitabäck Clay samples include Egertonodus basanus, Hybodus parvidens and Parvodus rugianus. Hybodont remains, other than teeth, include five morphotypes of placoid scales, incomplete cephalic spines and fragmentary fin spines.

Other fossil groups represented in the sieved residues from the bulk samples include bivalves, gastropods and bony fish. Together with the selachians, they indicate fluctuating palaeosalinities in the area. The lower coquina bed, VC3, includes taxa indicating mesohaline conditions while the composition of the fauna in the other coquina bed, VC11, suggests oligohaline settings. In bed VC7, the presence of amphibian remains and the rarity of selachian fossils indicate an even lower salinity. Palynomorphs from the basal part of the section, immediately below bed VC3, indicate an earliest Cretaceous (Berriasian) age.

In late Ordovician and early Silurian times, the Girvan district lay in a shelf marinesetting on the margin of Laurentia, on the northern side of the Iapetus Ocean. The Lower Palaeozoic rocks of the Girvan district, and their shelly and graptolitic fossil fauna, were systematically described by Lapworth in 1882 and have formed an important research resource ever since. They provide valuable evidence for the depositional environment and geological setting of Girvan during the early Palaeozoic, in both regional and wider contexts, and demonstrate the long-recognised close affinity with contemporaneous Laurentian faunas. However, by late Ordovician and into Silurian times, the earlier Iapetus oceanic barrier to faunal migration had largely gone and there is good correlation between contemporaneous marine fauna throughout the British Isles and Scandinavia. Despite much recent research in the area, including resurvey work by the British Geological Survey, no comprehensive review of Silurian lithostratigraphy at Girvan has been published since the revision by Cocks and Toghill in 1973. The present review of the Silurian rocks addresses this need and complements the recently published (Fortey et al. 2000) revision of the underlying Ordovician rocks, thus bringing the entire Girvan Lower Palaeozoic succession up to modern standards of nomenclature.

The litho- and biostratigraphical subdivisions of the Tremadocian-Arenigian succession of the southern Montagne Noire are hereby revised. The Lower Ordovician diversity patterns are estimated through statistical analysis of 27 families, 64 genera and 132 species of trilobites across nine interval zones; from bottom to top: the Proteuloma geinitzi, Shumardia (Conophrys) pusilla, Euloma filacovi, Taihungshania miqueli, Taihungshania shui landyranensis, Colpocoryphe maynardensis, Neseuretus (Neseuretus) arenosus, Apatokephalus incisus and Hangchungolithus primitivus zones. Maximum trilobite diversity occurred within the E. filacovi and A. incisus zones, alternating with two sharp declines (S. (C.) pusilla and T.shui landyranensis zones), in both cases drastically related to transgressive-regressive trends. Trilobites attained their highest diversity in muddy outer-platform settings, decreasing in fine-grained sandstones and siltstones of shoreface environments, where trilobites could still be numerically abundant. Four major trilobite turnovers are recognised at the bottom of the P. geinitzi, S.(C.)pusilla, E. filacovi and T. shui landyranensis zones, in which survivors from previous intervals are not reported at specific level, and important generic replacements took place. Despite the relative incompleteness of the Lower Ordovician trilobite record, envisaged after analysing the high proportion of monotypic taxa and discontinuous ranges of some families, the trilobite record is representative enough to estimate diversity patterns, although a better understanding of the palaeoenvironmental control will greatly enhance biodiversity resolution.

A revised reconstruction of the dermal skull roof of Acanthostega gunnari from the Upper Devonian of East Greenland is provided, incorporating modifications consequent upon further study of the specimens and in the light of new information from the Upper Devonian genus Ventastega from Latvia. The new reconstruction features a midline gap between the frontal and nasal bones and median rostrals. A revised profile for the skull and a rendition of the dermal ornament is also provided.

The nominal species Mosasaurus ivoensis from the latest early Campanian of the Kristianstad Basin in southern Sweden, is redescribed and assigned to the tylosaurine genus Tylosaurus on the basis of its dental and vertebral morphology. A partial skeleton (KUVP 1024) from the late Coniacian to earliest Campanian Smoky Hill Chalk Member of the Niobrara Formation in western Kansas, USA, was previously referred to “M”. ivoensis. Nevertheless, its marginal teeth are markedly different, both in size and morphology, from those of topotypic T. ivoensis.

Examination of type specimens and topotypic material of the nominal tylosaurines Hainosaurus pembinensis from the late early Campanian of Manitoba, Canada, H. gaudryi from the late Santonian or early Campanian of northwestern France, and H. lonzeensis from the Coniacian or Santonian of Belgium, indicates that all three may be Tylosaurus.

The utility of isolated tooth-crowns in mosasaur taxonomy has been hampered by the often poor quality of the published illustrations of these fossils in combination with poor stratigraphic control. All Swedish remains of T. ivoensis, including 172 marginal teeth, 6 pterygoid teeth, several jawbone fragments and 12 vertebrae, were collected from a narrow stratigraphic interval corresponding to the highest biozone in the German eight-fold division of the early Campanian, providing the first good insight into the intraspecific dental variation in a tylosaurine mosasaur.

It has been known for some years that parallel horizontal structures separated in a vertical plane produce patterns of equally spaced parallel lines that can be detected and mapped by dowsing detector rods, indicating that the patterns are produced by interaction between the structures and the dowsing field, whatever that is. Consequently the arrangement is called a dowsing interferometer and the lines are referred to as interference fringes. The patterns are produced equally well by electrical insulators as by electrical conductors, and the inference that the field is not electromagnetic was confirmed by experiments carried out in an electromagnetically shielded laboratory. Measurements of interferometer fringe spacings made irregularly from 1991 to 1996, analysed retrospectively, were found to be dependent on time of year, decreasing relatively suddenly by a factor of about three in November and increasing again in April. An isolated rise and fall in early March was also found. These results were confirmed by more frequent measurements in 1997 and published the following year. This paper reports a co-operative programme of interferometry between Scotland and New Zealand from 1997 to 2001. The annual pattern of the fringe spacing is clearly defined, with rapid changes between 2m and 6m in November and in April, and an isolated event in early March; but remarkably the pattern in the southern hemisphere is inverted with respect to that in the north. There are differences of a few days in the timings of the events in the north and south, and the amplitude of the March event is increasing.

This paper is a continuation of the work described in Transactions of the Royal Society of Edinburgh: Earth Sciences89 (1998) 1–9.