SUMMARY The review is mainly concerned with Carboniferous non‐marine Anthracosiidae and Myalinidae, of which only the shells are known, and with certain unspecialized non‐byssate suspension‐feeding bivalves which had smooth shells and burrowed shallowly. Limited experimental evidence and observation of living bivalves suggest that in certain Recent siphonate species and in some members of the non‐siphonate Anthracosiidae the shape of the shell was functionally related to movement through the sediment in the same way. Predicted optimum shapes of shells for downward burrowing and for upward near‐vertical movement in sands and silts were apparently realized in the Anthracosiidae, which constituted a series of highly variable opportunistic assemblages. It is stressed, however, that the shape of the shell always appears to be a compromise between several functional requirements. In both the early Anthracosiidae and in several analogous Recent marine genera, orientation of the long axis of the shell was the same for downward burrowing and for upward pushing, that is near the vertical, with posterior end upward. Invasion of the Pennine late‐Namurian delta took place when marine bivalves pushed upward, thereby avoiding sedimentation from delta lobes moving seaward relatively swiftly. The evolution of Carbonicola occurred at about this time (the Marsdenian Age) when the bivalves acquired a smooth elongate shell of ‘streamlined’ form, having a hinge plate with swellings and depressions on it (later to evolve into teeth). All these features tend to characterize the active shallow burrowers of today. Entry into soft‐bottom eutrophic conditions of fresh water is characterized in several unionids by increase in height/length (H/L or w/m) ratio of shell, in anterior end/length (A/L) and in obesity (T/L) (see Fig. 2, centre). These changes also took place in established faunas of Carbonicola characteristic of richly carbonaceous shales, in faunas of supposed Anthraconaia in more carbonaceous sediments of mid‐and late‐Carboniferous times in the U.S.A. and in Anthraconauta of the British late Carboniferous (Westphalian C and D). The genus Anthracosphaerium epitomizes the culmination of these trends in the Anthracosiidae, and species of the genus were probably epifaunal or shallowly infaunal active burrowers on soft bottoms in Westphalian upper A and B time. Two contrasting patterns of growth characterize the shells of the widely variable unionid Margaritifera margaritifera. In the first, dorsal arching of the shell, with straightening and reflexion of the ventral margin, provides increased weight but decreased ligamental strength. In the second, in which the ‘hinge line’ tends to remain straight while the dorsal margin becomes more rounded and obesity increases, there is increased metabolic efficiency for active surface movement. The maintenance of these trends within the species, which may be regarded as secondarily opportunistic, affords a means of insurance for survival within the highly variable environments of fresh water. The same trends are recognizable in established faunas of Carbonicola, where it is likely that they performed the same function, as well as in Mesozoic and Cenozoic Unionidae. The functional explanation outlined in paragraph (6) may be extended to provide an ecological meaning for Ortmann's ‘Law of Stream Distribution’, which states that obesity of unionid shells increases downstream. This applies broadly, within a fairly wide range of variation, a fact which again suggests ‘insurance’ of faunas against the variable hazards of fresh‐water habitats. In bivalves having considerable thickness of shell in relation to their size, and having strong umbonal development, specific gravity of the living bivalve is correlated with H/L and T/L ratios of the shell, as in the venerid Venerupis rhomboides. In this species, differences in the specific gravities of the bivalves, as well as their shape, appear to be functionally related to shallowly infaunal burrowing in different substrates. The conclusions of paragraphs 6 and 8 provide a functional explanation, in terms of selection, for the palaeoecological ‘law’ of Eagar (1973), which is applicable to established faunas of Carbonicola in mid‐Carboniferous time, and relates variational trends in two main groups of shells primarily to increases in the relative water velocities of the palaeoenvironments. Where the growth of relatively unspecialized bivalve shells has been measured, allometric relationships have usually been found in H‐L and T‐L scatters. Logarithmic lines have two inflexions and linear scatters a sigmoidal form. A similar pattern of allometric growth has been found in both Carbonicola (H‐L) and Anthraconauta (m‐w). These patterns appear to be related to the optimum requirements of water‐borne larvae, the initial byssal phase of settlement, when ability to burrow quickly is essential, and the main period of growth and activity. It is herein suggested that the final second inflexion, which indicates a falling off of gain in H/L and T/L ratios, may be a genetically controlled modification of the growth pattern which counteracts the operation of the ‘cube‐square rule’ (of Thompson, 1942) and prolongs productive life. Patterns of relative growth of the shell may be significantly modified by conditions of the habitat; both T/L and H/L ratios may be increased, with general reduction in size, in the less ‘favourable’ habitats. Both these ratios have been similarly modified, the one in the ‘natural laboratory’ of a lake formed by the damming up of streams, and the other in transplant experiments with living Venerupis. In both these latter cases, phenotypic changes took place in the same direction as those expected on the basis of natural selection. Direct response to environmental factors cannot therefore be ruled out as an agent in similar changes noted in Carbonicola and supposed Anthraconaia in paragraphs (5) and (9) and may have been operative in those of paragraphs (7) and (8).

No abstract

Reconstructions based on biogeography, palaeomagnetism and facies distributions indicate that, in later Palaeozoic time, there were no wide oceans separating the major continents. During the Silurian and Early Devonian time, many oceans became narrower so that only the less mobile animals and plants remained district. There were several continental collisions: the Tornquist Sea (between Baltica and Avalonia) closed in Late Ordovician time, the Iapetus Ocean (between Laurentia and the newly merged continents of Baltica and Avalonia) closed in Silurian time, and the Rheic Ocean (between Avalonia and Gondwana and the separate parts of the Armorican Terrane Assemblage) closed (at least partially) towards the end of Early Devonian time. Each of these closures was reflected by migrations of non-marine plants and animals as well as by contemporary deformation. New maps, based on palaeomagnetic and faunal data, indicate that Gondwana was close to Laurussia during the Devonian and Carboniferous periods, with fragments of Bohemia and other parts of the Armorican Terrane Assemblage interspersed between. It follows that, after Early Devonian time, the Variscan oceans of central Europe can never have been very wide. The tectonic evolution of Europe during Devonian and Carboniferous time was thus more comparable with the present-day Mediterranean Sea than with the Pacific Ocean.

Fossils from a new Upper Carboniferous LagerstaÈ tte of Late Westphalian A (Langsettian) age at Bickershaw, Lancashire, UK, are reported and an account of the locality is presented. In common with many Upper Carboniferous LagerstaÈ tten, the fossils are preserved within siderite nodules hosted by coal seam roof shales. The fossils include a diverse assemblage of plants, nonmarine bivalves, crustaceans, insects, arthropleurid fragments, a euthycarcinoid, xiphosurans, a whip scorpion, scorpion cuticle fragments, ®sh scales and coprolites. The arthropods, with a greater diversity and abundance of aquatic taxa, and the non-marine bivalves suggest a brackish,¯uvio-deltaic setting probably above the Haigh Yard Coal. However, tabulation of the relative abundances of fossils is of limited palaeoecological value, since the data represent a combination of the original community structure, each groups' preservation potential and mixing of material from dierent stratigraphic horizons on the tip.

Studies in the variation of the Anthracosiidæ (= Carbonicolidæ) in the South Wales Coalfield by Davies and Trueman (1927) and in the Scottish Coalfields by Leitch (1936, 1940, 1941) have suggested that environmental influence on shell form is confined to changes in thickness of shell and in obesity. Davies and Trueman quoted H. R. Wakefield, who remarked that to-day among the Unionidæ these characters appear to vary according to the station of the shells in a single lake. Leitch referred to the work of Ortmann (1920) on the relation between obesity and distribution of river mussels in a large drainage basin in the U.S.A. The writer, in the course of very detailed work on the most highly variable faunas of the Anthracosiidæ known in the British Coal Measures, found evidence suggesting that certain other changes in shell shape and dimensions may also accompany environmental changes, as far as past environments may be reflected in the lithology of the rocks containing the shells.