The Foulden Site of Special Scientific Interest is one of the few Cementstone Group localities that yields significant fauna and flora. Excavations in 1980 and 1981 removed a 1·3 m2 slab of the Fish Bed for laboratory study of the vertical and spatial distribution of the biota. Layer by layer analysis revealed an almost mutually exclusive relationship between the vertical distribution of the palaeoniscoid fishes and malacostracan crustaceans, as well as a horizon crowded with juvenile acanthodians. Elongate elements of this biota showed slight preferred orientations at that prolific horizon. Some 27 m of strata including the Fish Bed, Plant Bed and Shell Bed are recorded in detail, and their biota noted. Interim results of work on the main groups collected are summarised in associated papers in this part of Transactions. Brief reports are given here on the euryhaline marine bivalve mollusc Modiolus latus (Portlock), a rare durophagous bradyodont shark tooth and an Eogyrinus-like amphibian scute.

The Berwickshire Cementstone Group was deposited on a coastal plain crossed by meandering rivers flowing S into a marine gulf, the Northumberland basin. Sedimentation was mainly by overbank flooding onto an alluvial plain which largely dried out between floods. The resulting sediments include poorly-stratified mudstones and siltstones and crevasse splay sheet sandstones. The depositional environment of the cementstones is not clear. Lacustrine and tidal flat origins are considered. Semi-permanent floodplain lakes were very rare, but in one the uniquely fossiliferous Foulden Fish Bed was deposited. This lake was shallow and filled by repeated influxes of sediment carried from rivers by sheetfloods. On entering the lake, these sheetfloods may have become density underflows from which thin, graded siltstones or sandstones were deposited. These floods carried plant debris into the lake and rapidly buried the remains of the lake fauna. The salinity of the lake cannot be determined, but the input of both fresh and saline waters can be envisaged. There is no evidence that the lake was permanently eutrophic. The mortality of the animals may have been due to changes in water chemistry produced by hot dry spells of weather followed by storms and floods.

Anatomically preserved plants are recorded from calcareous nodules and include Lepidodendron calamopsoides Long which demonstrate for the first time the presence of anisotomous branching, in addition to the normal isotomous branching and details of the stele and the cortex seen in longitudinal section. The nodules also contain Setispora subpaleocristatus (Alvin) Spinner and Lagenicula crassiaculeata Zerndt megaspores. There is an abundant compression flora. Compressions and petrifactions of Tristichia ovensi Long may not represent the same plant and may not be connected to Stamnostoma huttonense Long. We suggest that the compressions of T. ovensi were of a pteridospermous plant with unlaminated fronds bearing small uncupulate seeds and lateral clusters of microsporangiate organs of a new type.

Plant miospore assemblages obtained from eight beds of the Foulden sequence, ranging from Bed 1 to Bed 17, are basically similar to each other in stratigraphical significance. The presence of Convolutispora circumvallata, Crassispora trychera and Schopfltes claviger and the absence of Lycospora pusilla suggest assignment to the Schopfites claviger-Auroraspora macra (CM) Biozone. This indicates the Courceyan Stage of the British Dinantian and the late Tournaisian (Tn3) of the Belgian Series. Comparable miospore assemblages have been described from Cementstone Group localities elsewhere in Berwickshire, East Lothian, Ayrshire and northern England. Preservation, and lack of marine palynomorphs and scolecodonts, suggest non-marine deposition. The presence of vitrinite fragments and megaspores as well as miospore tetrads suggests little transport before deposition.

A lower Carboniferous limulacean Rolfeia fouldenensis gen. et sp. nov. is described. It is regarded as a sister group of Paleolimulus and distinguished from it by the presence of fixed opisthosomal spines and by the more primitive outline of the opisthosoma. Both characters could have been derived from belinuroid ancestors, and both were retained in the advanced belinurids and euproopids. The view that limulaceans are more closely related to certain primitive belinuraceans rather than descended from the euproopaceans is thus supported by the new evidence. Although Rolfeia was probably still capable of coaption, it is postulated that, like other limulaceans, it had adopted burial as a form of defence. It is suggested that the primary adaptive function of the fixed opisthosomal spines was to support the body over a soft substrate although they may also have served a defensive role.

The occurrence at Foulden of the eurypterid Cyrtoctenus peachi Størmer and Waterston is confirmed. The eoscorpiid Trachyscorpio squarrosus Kjellesvig-Waering 1984 has been described from Foulden and disjecta membra of Gigantoscorpio cf. willsi Størmer discovered in the recent excavation are described. The type of preservation and adaptive features of the chelicerates would suggest that the eurypterid and scorpion fragments are allochthonous and the xiphosurans probably autochthonous.

The Tealliocaris and Crangopsis previously recorded from Foulden are shown to be Bairdops elegans and Belotelson traquairi, respectively. This locality thus yields the oldest Known Bairdops. The two species are confined to the lower Carboniferous sedimentary basin S of the Southern Uplands massif. Elsewhere, these genera are thought to characterise the nearshore marine fauna. The reciprocal relationship noted between malacostracans and actinopterygians may therefore be a function of fluctuating salinity.

A unique specimen of an unusual elongate segmented compression fossil, Polylurida aenigmatica n. gen., n. sp., is described and illustrated. The bilaterally-symmetrical, dorsoventral structure (c. 6·5 cm long × 0·8 cm maximum width) narrows towards both ends and is subdivided into approximately 19 similar tapering segments, showing no evidence of tagmosis, appendages or an internal gut. The segments are covered with a substantial papillose organic cuticle, each segment bearing towards one end a pair of transverse calcareous bars. These bars are perforated by simple canals filled with strands of dark organic material which traverse their whole thickness and flare out superficially to form pores on both their lower and upper surfaces. Internal horizontal lamination of the bar material suggests an episodic accretionary pattern of skeletal growth. The bars appear to represent internal skeletal structures and are restricted to one side (dorsal or ventral) of the body, where the adjacent cuticle is locally thickened to form a corresponding series of discrete paired densely papillose transverse bands. Although Polyurida may be an unusual segmented worm, its singular suite of characters and lack of data prevent assignment to any higher taxon.

Coprolites of four distinct shapes, cylindrical, discoidal, beaded and irregular masses, are described from 11 horizons within Bed 16 of the Foulden sequence. These coprolites are composed of either phosphatic or claystone material, although they have not been analysed mineralogically or chemically. Their producers appear to have been fish, probably both nektonic predators and nekto-benthonic scavengers or detritus feeders.

Ostracods are poorly preserved and of two broad types, thick shelled paraparchitids including Shemonaella scotoburdigalensis and thin shelled podocopids, possibly species of Carbonita or Cavellina. This ostracod assemblage represents a conservative ‘carbonaceous fades’ fauna which occurs widely in plant-rich sediments in lower Carboniferous (Courceyan) sequences of northern England and southern Scotland. Other microfauna includes spirorbid worm tubes, fish scales, arthropod cuticle fragments, spat of myalinid bivalves, plant fragments and spores. No significant changes in the vertical distribution of the microfauna could be detected.

The microfauna indicates a general fresh to brackish water lacustrine or lagoonal environment possibly with access to quasi-marine conditions.

Acanthodians are represented by numerous young and poorly cranially ossified Acanthodes ovensi White, the earliest known acanthodid. Study of the small A. ovensi together with the type series shows that squamation begins posteriorly, and spreads anteriorly along an initially triangular front, as in A. bridgei Zidek. Gyracanthus sp. occurs at several horizons in the sequence. The coelacanth remains include the first articulated specimen known from Foulden, but preservation only allows an assignment to cf. Rhabdoderma. Although these taxa are distinctive, resemblances to other taxa occur only in generalised features. Hence, unlike the Foulden actinopterygian fish, the acanthodians and coelacanths cannot be used to suggest closer affinities with any other fish of this age, on which information is sparse.

The four stem-group neopterygians from Foulden are redescribed: Carboveles ovensi White, Aetheretmon valentiacum White, Strepheoschema fouldenensis White and Fouldenia ottadinica White. Carboveles is shown to be a junior synonym of Phanerosteon Traquair, Fouldenia ottadinica White is synonymised with Styracopterus fulcratus (Traquair), Strepheoschema rhodesi Moy-Thomas with S. fouldenensis White and Aetheretmon whitei Moy-Thomas with A. valentiacum White. The family Strepheoschemidae is erected to comprise Strepheoschema and Aetheretmon, characterised by paired ventral ridge scales in front of the pelvic fins. Cosmoptychius striatus (Agassiz), recorded here from Foulden for the first time, is widespread in the geologically younger Oil Shale Group. The Foulden actinopterygian fauna is identical to that of the Cementstone Group at Coomsdon Burn, Northumberland, although with two additional taxa. Styracopterus fulcratus occurs also at a further two Cementstone Group localities in Scotland.

Material of crossopterygian fishes from Foulden, Berwickshire, is important in including the first known entire specimen giving a clear picture of the body form of the Family Rhizodontidae (sensu Andrews 1973). This group is specialised in having numerous subsidiary lateral lines on the body, a ventrally expanded shoulder girdle, fins with peculiar stiffened structure, the pectorals forming greatly enlarged paddles and other fins being reduced in size. The Foulden remains fall into large and small size-ranges occurring largely at different levels in the sequence, but although these may represent growth in one taxon the available parts of the body are mostly complementary. Separate preliminary descriptions of the two size-ranges are therefore given, the small complete form being named ?Strepsodus anculonamensis sp. nov., and the large form remaining unnamed pending further study. An abridged historical account of previously named rhizodont genera (Rhizodus, Strepsodus, Sauripterus and Pycnoctenion) is given in order to explain why the new Foulden form(s) cannot at present be generically assigned more closely. A revised diagnosis of the Family Rhizodontidae leads to a discussion of the functional morphology and mode of life of fish of this type. Analogies with modern forms indicate that, like some sharks and crocodiles, they may have fed by tearing flesh off large prey, rotating or shaking against its inertia. Ways in which variously shaped rhizodont teeth (recurved, sigmoid etc.) may have functioned are suggested. Addenda deal with (1) the possible occurrence at Foulden of a large lungfish and (2) (after considering primitive and advanced rhizodont features) the renaming of Rhizodus ornatus Traquair 1878 from higher in the Scottish lower Carboniferous, as the type of a new genus, Screbinodus.

The Foulden Fish Bed was deposited in a semi-permanent brackish water lake on a floodplain between a northerly mountain range and the sea. The lake supported a rich biota mainly of swimming animals (fish and crustaceans) with a benthos of trace fossils only in the earlier and later stages of the lake's development. Xiphosurans are very rare. This biota changed according to the physical conditions of the lake. Salinity, fluctuating according to water level, was a dominant control. Early in the lake's history there was a dominant benthos of euryhaline bivalves; at the same time, large rhizodonts lived in the lake as the top carnivores. Subsequently, crustaceans were the dominant nektobenthos, in a probable brackish-water regime, the rare Bairdops preying on the common Belotelson; there was a limited infauna. Following this, palaeoniscids abound in the middle part of the sequence, perhaps due to a higher water level but crustaceans are absent at this horizon. During the last stages before final silting-up following a mass-mortality episode, biotic diversity reached a maximum with many fishes and crustaceans and some trace fossils, due to optimal salinity conditions.

Comparisons are made with other Carboniferous biotas and with the Jurassic Solnhofen limestone.