1 If: Palaeontology VOLUME 29 • PART 3 SEPTEMBER 1986 Published by The Palaeontological Association London Price £21-50 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to promote research in palaeontology and its allied sciences. COUNCIL 1986-1987 President. Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Vice-Presidents: Dr. M. G. Bassett, Department of Geology, National Museum of Wales, Cardiff CFl 3NP Dr. D. E. G. Briggs, Department of Geology, University of Bristol, Bristol BS8 IRJ Treasurer: Dr. M. Romano, Department of Geology, University of Sheffield, Sheffield SI 3JD Membership Treasurer: Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Birmingham B4 7ET Institutional Membership Treasurer: Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 5HN Secretary: Dr. P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA Circular Reporter: Dr. D. J. Siveter, Department of Geology, University of Hull, Hull HU6 7RX Marketing Manager: Dr. V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 IRJ Public Relations Officer: Dr. M. J. Benton, Department of Geology, The Queen’s University of Belfast, Belfast BT5 6FB Editors Dr. P. R. Crowther, City of Bristol Museum and Art Gallery, Bristol BS8 IRL Dr. D. Edwards, Department of Plant Science, University College, Cardiff CFl IXL Dr. L. B. Halstead, Department of Geology, University of Reading, Reading RG6 2AB Dr. T. J. Palmer, Department of Geology, University College of Wales, Aberystwyth SY23 2AX Dr. C. R. C. Paul, Department of Geology, University of Liverpool, Liverpool L69 3BX Dr. P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL Other Members Dr. H. A. Armstrong, Newcastle upon Tyne Professor B. M. Funnell, Norwich Dr. M. E. CoLLiNSON, London (plus one vacancy) Overseas Representatives Australia: Professor B. D. Webby, Department of Geology, The University, Sydney, N.S.W., 2006 Canada: Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta Japan : Dr. I. Hayami. University Museum, University of Tokyo, Hongo 7-3-1, Bunkyo-Ku, Tokyo New Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt U.S.A.: Dr. R. J. Cuffey, Department of Geology, Pennsylvania State University, Pennsylvania 16802 Professor A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66045 Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403 South America: Dr. O, A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108. Venezuela MEMBERSHIP Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1986 are; Institutional membership Ordinary membership Student membership Retired membership £45-00 (U.S. $68) £21-00 (U.S. $32) £11-50 (U.S. $18) £10-50 (U.S. $16) There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 5HN. Student members are persons receiving full-time instruction at educational institutions recognized by the Council. On first applying for membership, an application form should be obtained from the Membership Treasurer, Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Aston Triangle, Birmingham B4 7ET. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer. All members who join for 1986 will receive Palaeontology, Volume 29, Parts 1-4. All back numbers are still in print and may be ordered from Marston Book Services, P.O. Box 87, Oxford 0X4 ILB, England, at £21-50 (U.S. $33) per part (post free). Cover: The chitinozoan Ancyrochitina onniensis Jenkins 1967 from the Late Caradoc, Onnian of the Onny River, Shropshire. The specimen measures 130 /im in length. Dr. W. A. M. Jenkins provided the photomicrograph. THE COMMUNITY STRUCTURE OF THE MIDDLE CAMBRIAN PHYLLOPOD BED (BURGESS SHALE) by S. CONWAY MORRIS Abstract. The palaeoecology and taphonomy of the Middle Cambrian Phyllopod Bed fauna (Burgess Shale, British Columbia) is described. Examination of over 30 000 slabs of shale and more than 65 000 specimens, many of them showing soft-bodied preservation, provides estimates of numbers of individuals and biovolumes of approximately 100 genera belonging to twelve major groups. Life habits are diverse with a sessile and vagrant infauna and epifauna, together with a nektobenthos, being recognized; components of a separate pelagic community are also present. Trophic analysis documents deposit feeders, suspension feeders, predators, and scavengers, and reconstructs the trophic nucleus and a feeding web. Possible niche structure of various ecological categories is discussed in the context of dominance diversity curves. Most distributions are log-normal, but for epifaunal vagrant deposit (collector) feeders a geometrical distribution may support the hypothesis of niche pre- emption. Comparisons are drawn between the community structure of the Phyllopod Bed biota and typical Cambrian shelly faunas. In isolation the shelly component of the Phyllopod Bed has a typical Cambrian aspect, but it accounts for only c. 14% of genera and perhaps as little as 2% of individuals alive at the time of burial. Synecological pronouncements based on normal Cambrian assemblages are suspect, and the likely importance of predation is emphasized. The wider implications of this study include comparisons with younger Palaeozoic deeper-water communities in an attempt to trace the evolution of ecological analogues through time. This exercise is conducted in terms of broad categories of carnivores, suspension and deposit feeders. Some groups such as sponges, ostracodes, and trilobites persist, but probably changed in relative importance. In other cases there is evidence for ecological replacement; one example could be the rise of nautiloids and eunicid polychaetes as carnivores and scavengers. U p to 70% or more of species and individuals in a modern marine macrobiota may be effectively soft bodied (taken here to include lightly skeletized species) and possess a minimal fossilization potential (Johnson 1964; Lawrence 1968; Macdonald 1976). An upsurge in taphonomic studies has reinforced earlier conclusions that the fossil record is inevitably a seriously biased sample of original biotic diversity (e.g. Lawrence 1968; Schopf 1978; Boucot 1981; and many others). Almost all studies of fossil assemblages are restricted to a time-averaged shelly component (see Walker and Bambach 1971) that in most cases probably bears little comparison with the original living community, although attempts exist to estimate original standing crops (Stanton et al. 1981). Trace fossils may provide an indication of soft-part diversity, but incorporation of such data is fraught with problems. A plurality of traces may result from the varying behaviour of one species, or two or more unrelated species (even of different phyla) may make indistinguishable traces. Many animals, including soft- bodied species, may not produce preservable traces (see e.g. Hertweck 1972), while information may be duplicated when shelly species create traces. In the absence of soft-part preservation many fossil community studies are literally ‘fleshed-out’ with a variety of uniformitarian assumptions based on observations amongst modern faunas; the validity of such exercises has been succinctly discussed by Scott (1963). In view of these difficulties, why have those Lagerstatten with extensive soft-part preservation not been used as a direct source of otherwise missing information? Their comparative neglect from the viewpoint of community and palaeoecological analysis (see Cisne 1913b and Schram 1979 for exceptions) presumably lies in their elevation to a unique status, isolated from the mainstream of study of fossil associations. Just suppose that the extraordinary conditions of preservation in the Phyllopod Bed had failed to materialize, so that only shelly fossils remained (Conway Morris 1981). The precise composition of (Palaeontology, Vol. 29, Part 3, 1986, pp. 423-467.| 424 PALAEONTOLOGY, VOLUME 29 this assemblage would vary according to taphonomic conditions, but in this quiet muddy environment the fauna might be expected to include trilobites, brachiopods, monoplacophorans, hyolithids, rare echinoderms, and sponges, the latter two groups mostly as scattered ossicles and spicules, respectively. This assemblage is little different in terms of major groups from many other Cambrian faunas, and as noted below many of the shelly genera have a broad geographical distribution; in the absence of soft-bodied fossils the Phyllopod Bed fauna would have no special claim for attention (see also Conway Morris and Robison 1982). Nevertheless, this is not meant to imply that the Burgess Shale is typical of all communities, and there is evidence for it perhaps being a conservative fauna (Conway Morris and Robison 1986). Prevailing palaeo-oceanographic condi- tions may also have been a factor in favouring certain faunal assemblages and promoting (or retarding) an environment conducive to soft-part preservation with cold, poorly oxygenated water providing propitious conditions. The main purpose, therefore, of this paper is to explain how study of the Phyllopod Bed soft- bodied fauna yields palaeoecological insights into a Cambrian community that are not available from the shelly assemblage. This is because the latter accounts for little of the taxonomic diversity and a trivial percentage in terms of individuals. With regard to taxonomic composition the fauna is dominated by arthropods but, unlike the great majority of other Cambrian assemblages, trilobites are an inconspicuous component. Study of ecological categories according to position relative to substrate and feeding type, in terms of numbers of individuals and biovolumes, permits a far more extensive characterization of a Cambrian community than has hitherto been possible; the role of predators is given particular attention. Some aspects of the synecology of the Phyllopod Bed biota have been briefly reviewed elsewhere (Conway Morris 19796), but that introductory analysis was both brief and based only on numbers of genera. The data for this study were collected mostly in 1979, and only a summary of the major conclusions has been published (Conway Morris 1981). THE SETTING OF THE PHYLLOPOD BED FAUNA Introductory remarks The redescription of the Burgess Shale (Stephen Formation) biota was initiated by Whittington (1971fl, 6), following the reopening of the Walcott and Raymond Quarries by the Geological Survey of Canada in 1966 and 1967. Recent review papers have summarized our enhanced understanding of ; this biota and its place in Cambrian life (Conway Morris \919a, b, 1982; Conway Morris and Whittington 1979, 1985; Whittington 1980a, 1981a, 1982). In one of these reviews Conway Morris (1979a) suggested that within the Phyllopod Bed, exposed only in the Walcott Quarry and the source of the great majority of specimens from the Burgess Shale, a single benthic community was recogniz- able. This was termed the Marrella-Ottoia community after the abundant epifaunal arthropod and infaunal priapulid worm respectively. In addition rare pelagic species were taken to represent the Amiskwia-Odontogriphus community, members of which were trapped only infrequently by the turbidity currents responsible for transporting and burying the benthic community. In the strati- graphically higher Raymond Quarry, a benthic assemblage distinct from the Marrella-Ottoia com- munity of the Phyllopod Bed was identified. New discoveries of soft-bodied faunas in the vicinity of the Burgess Shale include examples from Mount Field and Mount Stephen with a marked similarity to the Raymond Quarry fauna (Collins et al. 1983) and lend credence to the notion that this is indeed a distinct community. Apart from scattered specimens recovered from float on Mount Field no direct counterpart to the Marrella Ottoia community has yet been discovered beyond the Walcott Quarry (Collins et al. 1983). It is significant, however, that with few exceptions the soft-bodied taxa found in these adjacent localities also occur in the Phyllopod Bed, albeit in very varied proportions. Thus, while the Phyllopod Bed fauna was probably only one of several benthic communities (possibly intergrading) in the area, its gross faunal composition is representative of the overall faunal diversity of this basin. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 425 Evolutionary setting The evolutionary importance of this fauna should be considered in the context of early metazoan evolution. Whether or not metazoans first appeared far back in the Precambrian (e.g. Kauffman and Steidtmann 1981; Kauffman and Fiirsich 1983; Runnegar 1982), the available fossil record points to a rapid increase in metazoan diversity, exponential at the taxonomic level of order (Sepkoski 1978), reflecting a series of adaptive radiations that originated in the Vendian. This climb in diversity levelled out during the mid-Cambrian to give an apparent 'evolutionary lull’, with a limited rise in diversity over about 1 5 Ma (c. 540-525 Ma), prior to a further episode of major diversification during the late Cambrian and Ordovician (Sepkoski 1978, 1979, 1981a, h). The age of the Burgess Shale (c. 530 Ma) means that it lies within this 'lull’, and so is well placed to show the relative richness of Cambrian life in terms of both ecologies and biotic diversity as an end-result of these early adaptive radiations. Cambrian shelly marine faunas had a distinct identity that stands in contrast to those of the later Palaeozoic (Sepkoski 1979, 1981a, h, 1984; but see Ludvigsen and Westrop 1983), and it is important to determine whether the soft-bodied component showed a comparable distinction (see below). Regional setting Restoration of possible Cambrian palaeo-continental configurations place British Columbia within the tropical zone at about 15° N. (Kanasewich et al. 1978; Scotese et al. 1979). A near-equatorial position for the Burgess Shale is thus significant if latitudinal diversity gradients existed, because faunas such as that in the Phyllopod Bed may have approached maximum diversity for this particular deeper-water biofacies. Much of the Phyllopod Bed fauna appears to have been benthic, inhabiting basinal muds and silts deposited beside a precipitous algal reef that was ultimately overwhelmed by clastic deposition (Mcllreath 1977; Aitken and Mcllreath 1984; see also Surlyk and Hurst 1984, fig. 4 for a remarkably similar occurrence in the Silurian of North Greenland). The sheer reef (Cathedral escarpment) marks unusually sharply the boundary between the median carbonate and outer detrital sedimentary belts that encircled the North American craton (e.g. Palmer 1972, 1974). Elsewhere in North America a pronounced change in slope across this boundary is sometimes evident from slumps and other mass- flow deposits containing shallow-water clasts, typically carbonates, reef and algal debris (e.g. Reinhardt 1977; James 1981; Kepper 1981; Read and Pfeil 1983), but subsequent tectonic disturbance has generally obscured the precise nature of the margin. The overall extent of the basin in which the Burgess Shale was deposited is uncertain. General palaeogeographic considerations, however, suggest that the basin and its faunas faced the open sea and would have been accessible to faunal migration. Much of the shelly fauna of the Phyllopod Bed has a wide geographic distribution and none of these genera is endemic to this horizon or the Burgess Shale. For example, amongst the brachiopods and trilobites practically all the genera represented are widespread within North America, and in a number of cases (e.g. Oryctocephalus) have an even more extensive distribution (see also Jell 1974). Thus the position of the Burgess Shale, with its indigenous deep-water faunas, argues for a cosmopolitan aspect. Moreover, the abundance of apparently pelagic agnostoid and eodiscoid trilobites is also consistent with open access. In particular Ptychagnostus praecurrens has an enormous geographical range (Robison 1982; Bednarczyk 1984) which suggests that potentially the Phyllopod Bed was open to migration from distant parts of the earth. Evidence is also growing for the existence of faunas broadly comparable to the Burgess Shale elsewhere in the western Cordillera (Utah, Idaho), representing a present-day geographical separation of about 1300 km and a temporal range through a substantial portion of the Middle Cambrian and in some cases probably the upper Lower Cambrian (see Robison 1984). Of approximately forty genera of soft-bodied arthropods, sponges, priapulids, annelids, medusoids, incertae sedis, and algae, over 70% are also known from the Burgess Shale (Conway Morris and Robison 1986). Information on soft-bodied faunas from elsewhere in the Middle Cambrian is still restricted, but occurrences in Spain (Linan 1978; Conway Morris and Robison 1986) and China (Resser 1929) hint at their former distribution. 426 PALAEONTOLOGY, VOLUME 29 TAPHONOMY OF THE PHYLLOPOD BED Introduction The events that are believed to have led to the preservation of the Phyllopod Bed biota are sum- marized in text-fig. 1 . The benthic fauna was carried and buried catastrophically by turbidity currents (Whittington 1971a; Piper 1972), and it is conceivable that biological activity (Hecker 1982) ora high organic content (Keller 1982) contributed to sediment instability. Transport leads to a distinction between the pre-slide environment, where the organisms lived and apparently flourished, and the post-slide environment where inimical conditions helped to ensure the astonishing preservation. The post-slide environment is represented by the Phyllopod Bed, now exposed in the Walcott Quarry. The Phyllopod Bed is located about 20 m from the reef escarpment, and its sequence of graded beds conceivably accumulated in a local sea-floor depression. There is no evidence that the muds bypassed the carbonate shelf via the reef top (Aitken 1971; Mcllreath 1977), but there are various clues as to the location of the pre-slide environment in the basin. It appears that many and perhaps all the other Resistant hard parts Pre-slide environment Transport of fauna Deposition in graded beds I Compaction and lithification => Quarried rock o ? Collector bias o Palaeoecological analysis Post-slide environment TEXT-FIG. 1. Taphonomy of the Phyllopod Bed (Burgess Shale). The benthic biota inhabiting the pre-slide environment were transported periodically downslope into a presumed anoxic zone, the post-slide environment. The sediment flows carried both living individuals and empty exuviae and shells. The latter would have had a variable residence time in the sediment had they remained in the pre-slide environment; soft parts, however, would have been consumed almost immediately. Pelagic elements were also incorporated, via several possible routes. Subsequent to burial, poorly understood factors that must have included the restriction of microbial decay led to the exceptional preservation. See text for further details. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 427 prolifically fossiliferous localities are located adjacent to the reef scarp, with the abundance of fossils declining sharply away from the reef (Mcllreath 1975; Collins et al. 1983). If both the pre- and post- slide environments were beside the reef the turbidity currents presumably flowed downslope parallel and adjacent to the reef. Scanty palaeocurrent data from the Phyllopod Bed (Piper 1972) support this hypothesis. The pre-slide environment There is circumstantial evidence that the pre-slide environment may have been loeated near the present day Mount Field. A phase of basin infilling immediately prior to the aceumulation of the Phyllopod Bed produced deposition of carbonate flows along a considerable length of the reef-toe; the flows extended some distance away from the Cathedral escarpment and now form the ‘Boundary Limestone Unit’ (cf. Kepper 1981). In the Mount Field area, however, there is an area of non- deposition adjacent to the escarpment. Mcllreath (1977) speculated that this might have been due to either an island on the reef top deflecting carbonate debris as it moved over the rim into the basin, or a local topographic high on the basin floor. Accepting the latter alternative, maintenance of relief until Phyllopod Bed time could have provided a suitable source area from which the turbidity currents originated. The distance between the proposed location of the pre-slide environment in the vicinity of Mount Field and the Walcott Quarry (post-slide environment) is c. 0-9-1 -8 km. This estimate lies within the range of limiting values calculated by another method, which was based on the separation in depth between the pre-slide (assumed to lie in the photic zone) and post-slide environ- ments, combined with varying values for the angle of slope between the two environments (Conway Morris 1979c). Some algae, especially Marpolia often occur densely strewn on bedding planes with very few metazoans and they may not have grown in direct association with other elements of the pre-slide environment. However, examples of repetitive association, such as that between the algae Morania and polychaete Burgessochaeta (Conway Morris 1979c), indicate that the pre-slide environment lay within the photic zone, perhaps at a depth of less than 100 m (but see Littler et al. 1985). Contrary to Rhoads and Morse (1971), who suggested that the diversity and presence of organisms with hard parts is indicative of well-oxygenated eonditions, Savrda et al. (1984) and Thompson et al. (1985) have shown that such groups occur in modern dysaerobic environments. Although taxonomically very different from Recent faunas, such may conceivably also have been the case in the Phyllopod Bed, and is consistent with the hypothesized proximity of anoxic conditions (see below). Even if the pre-slide environment was dysaerobic, the exceptional conditions promoting soft-part preservation presumably lay in the deeper post-slide environment, and the sea-floor of the pre-slide environment had no unusual preservational properties so that upon death the soft parts of any benthic organisms would have been destroyed quickly by microbial decay and scavenging. In reconstructing the original community it is assumed, therefore, that specimens with soft parts were alive at the time of slumping, because any corpse in the pre-slide environment would have been rapidly eonsumed. Parts of some animals, however, were more resistant to decay; most significant were the shelly skeletons (trilobites, brachiopods, monoplacophorans, hyolithids) which presumably were largely immune to destruction and had a prolonged residence time in the sediment. There are also species with lightly skeletized parts, especially arthropod carapaces, that persisted after the disappearance of the associated soft tissues. The ratio between intact individuals and those without soft parts in various species (Table 1) is presumably controlled largely by the relative resistance to decay of the lightly skeletized parts. To a lesser extent the ratio will also reflect the species’ original abundanee (see also Briggs 1978) and, in the ease of arthropods, the number of moults. For example, the abundant arthropod Marrella splendens consists almost solely of entire specimens, and the rarity of the exuviae (head shields; usually cephalic and lateral spines, more rarely separated) presumably reflects a delicate constitution with a limited resistance to decay and hence a short residence time in the sediment. At the other extreme the arthropods Tuzoia spp., Proboscicaris, Isoxys, Hurdia, and MoUisonia are known only from their carapaees; this could result from the combination of a relatively tough exoskeleton and original rarity, so that by chance no living specimens were preserved with their soft parts. Other species with 428 PALAEONTOLOGY, VOLUME 29 TABLE 1 . Abundance and percentage of specimens in eleven genera with lightly skeletized carapaces, headshields, or tubes, lacking their associated soft parts. Totals are estimates based on assumptions of disassociated parts and counterparts (see text). Genus Total with soft parts Total without soft parts Percentage without soft parts Canadaspis 4719 4050 46-2 Hurdia 0 128 1000 Isoxys 0 452 1000 Marrella 15092 40 0-3 Mollisonia 0 16 1000 Naraoia 129 168 56-6 Prohoscicaris 0 62 1000 Selkirkia 190 958 83-5 ‘S.’ gracilis 1 37 97-4 Sidnevia 177 132 42-7 Tuzoia 0 37 1000 a more equitable division between preservational types (Table 1) could represent both moderate original abundance combined with resistance to decay. This observation concerning the post-mortem accumulation of skeletal parts, especially shelly remains, is important because it has biased numerical proportions in the fauna and it is necessary to subtract these vacated hard parts when estimating the total number of organisms alive at the time of slumping, i.e. standing crop. This opportunity to estimate directly the standing crop of a fossil community is normally denied to palaeoecologists, although Kranz (1977) has presented a model for such estimates. Pelagic elements Additions to the fauna appear to have come from the pelagic zone. Presumably they include agnostoid and eodiscoid trilobites, groups that are usually interpreted as predominantly pelagic (Jell 1975; Robison 1972, 1975). There is an apparent anomaly in the occurrence of five agnostoids lodged beneath the cephalic doublure of a specimen of the predatory arthropod Sidneyia inexpectans, an animal regarded as benthic (Bruton 1981, pi. 13, fig. 99). Some evidence, however, is possibly consistent with the distribution of certain agnostoids being controlled by deeper, cooler water rather than a pelagic existence per se (Conway Morris and Rushton, in press). The occurrence of the agnostoid Ptychagnostus praecurrens and eodiscoid Pagelia hootes in the Phyllopod Bed may reflect the introduction of colder deep waters, perhaps induced by a rise in the thermocline. The Phyllopod Bed fauna could include elements more typical of deeper waters that were able to migrate into relatively shallow water given appropriate conditions. Whatever doubts surround the life habits of some agnostoid and eodiscoid trilobites, there are also rare soft-bodied species with apparently pelagic adaptations. Presumably, the pelagic animals sank to the sea-bed of the pre- or post-slide environment, or were trapped by one of the turbidity currents (text-fig. 1). Transport and preferred orientation The frequency of slumping is not known, although if each graded unit of the Phyllopod Bed represents a separate slump the estimated total of fifty presumably corresponds to this number of events. If the proposed location of the pre-slide environment on a topographic high adjacent to the present Mount Field is correct, then according to Mcllreath’s (1977, Fig. 1) palaeogeographic reconstruction this would indicate a sea-floor area of perhaps 1 km^ from which slumping sampled CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYTLOPOD BED 429 a number of sea-floor sites. Data for this study are taken from the entire Phyllopod Bed and so represent a time-averaged sample of unknown duration (?several decades or centuries). The presence of an infauna suggests that the bulk of the benthic community was present at the inception of slumping, rather than being swept up by flows descending from a more remote area. The fine grain size (silt-clay) of the sediment and apparently short distance of transport suggest that the erosive power of the turbidity currents was limited. Although some epifauna, nektobenthos, and stray pelagic organisms may have been incorporated downslope, any infauna probably remained in situ. In contrast, where long distance transport of shallow-water faunas occurred, as is inferred from the turbidites of the Pliocene Capistrano Formation of California for example, there was also erosion of deep water infauna (Kern and Wicander 1974). The fauna was deposited in graded beds with some shelly remains accumulating near the base of a turbidite. The monoplacophoran Scenella may form locally dense accumulations with its breviconic shell lying concave side up (Walcott 1912; cf. Piper 1 972), a feature also noted in other turbidites (e.g. Middleton 1967). Orientation of other concave-convex fossils has not been studied extensively and is compounded by uncertainty regarding the way-up of collected slabs. Preliminary data, however, indicate that small trilobites have a consistent preference to be one way up (assumed dorsal-to- ventral ratio is 1 : 2-3-3-9), whereas the larger Olenoides serrutus has a more or less equal distribution. In contrast, Cisne (19736) recorded no way-up preference in the small Ordovician trilobite Triarthrus entombed in the microturbidite of Beecher’s Trilobite Bed. Soft-bodied animals were deposited in a wide variety of orientations, including some individuals with their longitudinal axes either steeply or vertically inclined to the bedding (Whittington 1971«), while compressed worms such as Pikaia (chordate) and Louisella (priapulid) may be folded or otherwise contorted (Conway Morris \977a, 1979a). The reduction of the Phyllopod Bed collections to thousands of unoriented sawn slabs, each with only one or a few specimens, has removed a potentially valuable source of information regarding non-random distributions, but some evidence exists for preferred orientation of elongate specimens (cf. Cisne 19736). In vacated tubes of Selkirkia Columbia (priapulid) there is evidence of bimodal and more rarely unimodal distributions (text-fig. 2), although the samples are small. Amongst the agnostoid and eodiscoid trilobites (text-fig. 3) a t^st indicates that their distributions are not significantly dififerent from uniform. This may also be true for one sample of the larger trilobite Olenoides (un-numbered USNM slab), but in USNM 189800 the distribution appears to be bimodal (text-fig. 3). The post-slide environment The presence in the post-slide environment of anoxic conditions, perhaps due to hydrogen sulphide extending above the sediment-water interface (text-fig. 1), is inferred from the absence of either 200204 TEXT-FIG. 2. Rose diagrams showing varying extent of preferred orientation in the tubicolous priapulid Selkirkia Columbia on Phyllopod Bed (Burgess Shale) slabs; true North not known. USNM specimen numbers given; azi- muth direction towards anterior aper- ture; no soft parts present. 200460 200398 (n = 6) 201328 430 PALAEONTOLOGY, VOLUME 29 scavenging or bioturbation. Rare pyrite-filled structures (text-fig. 4a) may represent gas-generated hollows in the sediment (see Cloud 1960; Martens 1976) rather than metazoan burrows. To judge by the general restriction of prolific faunas to the basinal margins beside the escarpment (Mcllreath 1974, 1977), anoxia may have been a regional feature of the deeper waters of the basin. In addition to anoxic conditions, rapid burial was presumably an important factor in promoting exceptional 198882 81141 (n = l9) TEXT-FIG. 3. Rose diagrams showing varying extent of preferred orientation in trilobites on Phyllopod Bed (Burgess Shale) slabs; true North not known. USNM un-numbered slab, left-hand rose, Olenoides serratus, azimuth direction towards anterior (5 individuals dorsal-up, 3 ventral-up, 7 or possibly 8 with soft parts); right-hand rose, agnostoids, no azimuth distinction between anterior and posterior (52 individuals dorsal-up, 23 ventral-up (black), 36 complete, 39 incomplete). USNM 189800, left- hand rose, O. serratus, azimuth direction towards anterior (4 individuals dorsal-up, 5 ventral-up, 8 or possibly 9 with soft parts); right-hand rose, agnostoids, no azimuth distinction between anterior and posterior (46 individuals dorsal-up, 18 ventral-up (black), 37 complete, 27 incomplete). USNM 198882, Pagetia bootes, azimuth direction towards anterior (32 individuals dorsal-up, 8 ventral-up, 1 lateral). GSC 81141 ( = 35-5 cm above base of Phyllopod Bed), agnostoids, no azimuth distinction between anterior and posterior (15 individuals dorsal-up, 4 ventral-up (black), 18 complete, 1 incomplete). CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 431 preservation (cf. e.g. Goldring and Stephenson 1972; Liddell 1975). Experiments and observations on the depth of overburden necessary to immobilize various benthic taxa in modern marine environments (e.g. Schafer 1972; Kranz 1974; Nichols et al. 1978) and some ancient settings (e.g. Hardy and Broadhurst 1978) may only be broadly indicative considering the taxonomic distinctive- ness of the Phyllopod Bed fauna. Presumably all sessile organisms, especially sponges and echinoderms, and an appreciable portion of the vagrant epifauna had limited escape abilities after burial, but the absence of escape structure associated with the priapulids that were evidently effective burrowers (Conway Morris 1 977a) suggests that the fauna was either stunned or even dead on arrival at the post-slide environment. The latter condition is indicated by an apparent lack of response to anoxic conditions; some modern annelids and arthropods combat an anoxic milieu by entering metabolic stasis and coiling tightly (Dean et al. 1964), but such a feature is infrequent amongst Phyllopod Bed specimens. Reasons for death are speculative, but if the pre- and post-slide environments were separated by a temperature boundary then the fauna may have perished owing to thermal shock. In this context it may be significant that elevated water temperatures have been recorded during the introduction of a storm-generated turbidite into deeper water (Dengler et al. 1984). Osmotic shock promoted by transport through an equivalent pycnocline generated by salinity differences may be considered less likely because soft-bodied specimens show neither conspicuous ballooning nor wrinkling attributable to osmosis-induced shrinkage, although such features have been documented in other exceptional fossil faunas (e.g. Barthel 1970; Muller 1985). Living metazoans appear to have been absent from the post-slide environment, but the post-slide environment was not sterile as anaerobic bacteria were presumably present and responsible for the incipient rotting noted in many specimens (e.g. Conway Morris \977a, 1979c). A conspicuous feature of decay in some species is a dark organic patch beside the specimen; in some specimens of Marrella this stain has been observed to extend at least 20 mm from the body, suggesting that during its diffusion or flow in the early stages of decay, the surrounding sediment was wet and had a high porosity. Some Recent marine animals (decapod crustaceans) exude a brownish fluid when subjected to pressure changes (Menzies and Wilson 1961 ), but the presumably modest change in water pressure (7-2 kg/cm^ for a 70 m vertical displacement: see Conway Morris 1979c) between the pre- and post-slide environments would probably have been insufficient either to kill the fauna or expel body contents. Decay was only preliminary, and one or more factors evidently intervened before it was far advanced. In modern marine sediments bacterial populations, including anaerobics, often decline markedly beneath the sediment-water interface (e.g. Zobell 1938, 1942; Rittenberg 1940; Oppenheimer 1960; Marty 1981). Progressive burial beneath a series of graded beds may have diminished bacterial numbers so drastically as to render their activity on buried organic matter ineffective. More speculatively, appeals could be made to abrupt decrease in water content (early lithification) or changes in salinity (K. M. Towe, pers. comm.) to render anaerobic bacteria inoperative. Diagenesis Whatever mechanism was responsible for the termination of decay the buried organisms were pre- sumably reduced to carbon-rich remains relatively shortly after burial. The soft parts of fossils, how- ever, do not appear to be carbonaceous, but consist of aluminosilicate films (Conway Morris 1 977a). Reflective areas of these films are said to be composed of muscovite mica (K. M. Towe, pers. comm.). The exoskeleton of the trilobite Olenoides has been replaced by various minerals including chlorite and mica (Conway Morris and Pye, unpublished; see also Whittington 19806), and other shelly remains that were originally calcareous and soft-bodied material both appear to have a broadly similar composition. The time and rate of transition of the fossils to phyllosilicates is speculative, but may have occurred at depth under fairly elevated temperatures and pressures, coincident with the pronounced compaction noted by various authors (Rasetti 1966; Whittington 1975a, 6; Robison 1982). Pyrite occurs in association with many specimens, and is usually in the form of framboids (Conway Morris 1985a); pyrite formation is probably linked with the activity of anaerobic bacteria (Bubela 432 PALAEONTOLOGY, VOLUME 29 TOXT-FiG. 4. Radiographs of Phyllopod Bed (Burgess Shale) fossils showing pyritization. a, UM 1320, irregular pyritic structure, possibly representing original gas generated hollows, x 1-7; b, GSC 45368, edrioasteroid Walcoltidiscus sp. in aboral view, x2-7; c, d, GSC 45369, an eocrinoid, probably Gogial radiata (see Sprinkle 1973), x2 0, c, part with ridged plates and incomplete stem, d, counterpart with enigmatic branching structure associated with lower region of calyx; c, GSC 45370, isolated thoracic segments of trilobites, x 1-8; /, GSC 78452, problematical organism, possibly an echinoderm, x 1-8; g, GSC 78453, thoracic trilobite segment, x 10. GSC, Geological Survey of Canada, Ottawa; UM, University of Montana, Missoula. and Cloud 1983). In some cases, however, the pyrite has replaced the fossil rather than formed a framboidal coating. This partial replacement appears to have been largely restricted to hard parts, including those of echinoderms (text-fig. 4h-d), trilobites (text-fig. 4e, g), problematical remains (text- fig. 4/), and sponge spicules (see Walcott 1920). One echinodenn (text-fig. 4c) which appears to be referable to Gogial radiata (see Sprinkle 1973) is of particular interest. The part is moderately well preserved, with an incomplete stem (if comparisons with GP. radiata are valid) and remains of the radiating ridged plates. The counterpart is more poorly preserved, although there is a fairly clear indication that some of the calyx plates have epispires. More enigmatic is a prominent branching structure, adjacent to the stem, that could be interpreted as the brachioles. It is located beside the proximal calyx, however, and folding of the brachioles into this position is unlikely; unless the brachioles became detached, chance superposition with some other structure is perhaps more likely. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 433 Concluding remarks Exceptional fossil preservation gradually dies out above the Phyllopod Bed, and this decline could be linked to the cessation of conditions of rapid burial by turbidity currents. The final stage in the taphonomic history, that of quarrying and collection, has been summarized elsewhere (Whittington 1971a, 1980a; Conway Morris 1982). Measurements of the dimensions of the Walcott Quarry indicate that c. 600 m^ of Phyllopod Bed were removed. Discarded material forms a thick talus on the slopes below the quarry and, together with debris from Walcott’s camp, has provided an additional source of specimens to later collectors (Collins 1978). The extent of collection bias is not certain. In the USNM the large size and relative inferiority of some of the specimens suggests Walcott was not unduly selective. Collecting by the Geological Survey of Canada appears to have been reasonably thorough. SOURCES OF DATA Locations of specimens The data of specimen counts on the collected slabs come from several sources. By far the most important is the enormous collection made by Walcott during five seasons (1910-1913, 1917) of field-work and now stored in the National Museum of Natural History, formerly the United States National Museum (USNM), Smithsonian Institution. Phyllopod Bed specimens are usually labelled 35k, while the much smaller number of specimens (c. 465) from a higher excavation (apparently the same as Raymond Quarry: see Whittington 1971a) are labelled 35k/ 10 or 35k/ 1. When the labels are absent the provenance can usually be determined by the lithology. Two seasons (1966, 1967) of excavations by the Geological Survey of Canada (GSC) produced a smaller sample, approximately 1 1 % of the estimated USNM total. A number of museums and university departments possess small Burgess Shale holdings (Appendix 1 ), the majority of which were either exchanged or purchased from the Smithsonian and are mostly included here with the USNM totals. In 1975 the Royal Ontario Museum, Toronto (ROM), collected fossils from the talus discarded by previous expeditions. With the exception of a few undoubted Phyllopod Bed specimens this material is excluded because the exact provenance is not certain. For the purposes of this paper specimens from all levels of the Phyllopod Bed are considered together. This is because, although the GSC recorded the vertical distribution of specimens, the measured divisions usually encompass a number of turbidite horizons, while Walcott (1912) offered only a few meagre hints at an early stage of his collecting. Thus, while the GSC data show abundance of various species to vary through the Phyllopod Bed (see Conway Morris 1985a), the ideal case of assessing separately the contents of each slump is not feasible and the fauna is a time-averaged assemblage (Walker and Bambach 1971). As noted above, however, there appears to have been only one benthic community in the pre-slide environment and, while an individual slump would be unlikely to trap representatives of the entire fauna, multiple slumping of various portions of the sea- floor increased the probability of assessing its overall diversity. Such successive temporal sampling is thus presumably analogous to the sampling of a modern benthic community by a series of sediment grabs or a towed sledge. Problems in estimating numbers of individuals Burgess Shale specimens adhere to both sides of the split rock, so giving parts and counterparts. It is important that part and counterpart be kept together because the level of splitting through the specimens often varies and may confine certain features to either part or counterpart (e.g. Whittington 1975a; Conway Morris 1979c, 1985a). Walcott apparently failed to appreciate this, and in the USNM collections c. 95% of specimens are presently disassociated. Part and counterpart may be stored in separate drawers. Alternatively, one side either may never have been collected or was even sent to another institution. Using several species where the totals of parts only and parts plus counterparts are known, it is estimated that in the USNM collections approximately 75% of specimens are unassociated parts while the other 25% are disassociated parts and counterparts. 434 PALAEONTOLOGY, VOLUME 29 Independent support for these estimates comes from a comparison in the tubicolous priapulid Selkirkia of the relative proportion of vacated tubes to those with associated soft parts (Conway Morris 1977a). In the GSC collections the proportion of specimens with soft parts is known to be c. 21 % with reasonable confidence because a careful attempt was made to find all parts and counter- parts. In the USNM sample of this priapulid, that is about six times larger, the same proportion is calculated on the basis of the figures given above to be c. 19%. During the GSC collecting a point was made to keep parts and counterparts together and c. 25 % of specimens are presently associated. Although some counterparts were never collected, some disassociation has occurred in the GSC material. In the majority (66 %) of cases where only the part is known it is supposed that the counterpart is not available, but in the remaining 34% of specimens inadvertent disassociation is assumed. If these correcting factors, which are used as the basis of this paper’s analysis, are accepted, the original proportion of associated USNM specimens was c. 19%, whereas in the GSC sample the figure is approximately 42%. The Phyllopod Bed collections used in this study are located on a total of about 33 520 slabs, of which about 3400 consist of associated parts and counterparts. Using the correcting factors given above for disassociated parts and counterparts in the USNM and GSC collections, the original total of slabs (parts only, parts and counterparts) is believed to be approximately 29 700 slabs. I estimate to have inspected directly c. 91 of the slabs, so that any bias should be consistent. Data for the remaining 3% of slabs were supplied from specimen lists of various institutions’ holdings. Each slab (both sides) had a note made of its specimens (average 2-5 specimens/slab), both for counts of individuals and intra- or interspecific associations. On large slabs crowded with specimens only estimates were made, especially for Canadaspis perfecta that occurs as ‘herds’ of intermingled soft- part specimens and isolated carapaces (Briggs 1978), and dense tangled knots that, although poorly preserved, appear to represent specimens of the worm "Ottoia tenuis. Specimens concealed entirely within slabs escape notice, and even radiography would fail to reveal buried specimens whose silicate films have minimal density contrast with the surrounding matrix. The thinness of most slabs, however, combined with preferential splitting through fossil films (Conway Morris 1979c) suggests that this is not a serious source of error. Numbers of individuals It is calculated, therefore, that the available sample consists of about 73 300 specimens (of which the USNM component accounts for c. 89%), distributed as follows: animals 87-9%, algae 1 1-3%, and indeterminate material 0-8%. The algal (see Walcott 1919; Satterthwaite 1976) count is very approximate as each slab with abundant coverings, especially of Marpolia spissa and Morania eonfluens, was registered as a single specimen. The total for the indeterminate category is also only a rough guide. These two categories are not considered in any further detail; the remainder of this paper assesses the animal component and in particular that fraction judged to have been alive at the time of slumping (estimated 40368 specimens). As noted above, it is necessary to distinguish between individuals alive at the time of transport and those dead specimens, or parts thereof, that persisted in the sediment owing to a resistant skeleton. The live; dead ratio within shelly species of modern communities has received wide attention (e.g. Macdonald 1976) because of its palaeoecological applications. The Phyllopod Bed fauna offers a most unusual opportunity for at least partial inspection of this problem. In some species with hard parts it is possible to distinguish live from dead specimens, even when soft tissue is not apparent. In Hyolithes, helens and operculum attached to the conical shell are good evidence for vitality. Mantle setae protruding from the valves of Micromitra are similarly interpreted as evidence of the specimens being alive during transport. Such criteria, however, are not available in the monoplacophorans, represented almost entirely by Seenella, and most of the brachiopods. The proportion of live to dead individuals is arbitrarily taken as 5% for monoplacophorans (same as Hyolithes) and 14% for brachiopods (same as Micromitra). Biovolumes In addition to specimen counts, biovolumes were calculated for each genus that had representatives CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 435 living at the time of transport. Estimates were made by reference to an individual of average length, and then multiplied by the specimen total to give the population’s biovolume. Simple formulae for standard shapes were used. Calculation of biovolumes in cylindrical worms such as priapulids was straightforward, but in more complex organisms (e.g. an arthropod) the body was treated as a number of separate units. In these calculations no allowance (as biomass) was made for skeletal hard parts, but serious discrepancies are unlikely because groups with hard parts were not an important part of the original community (see below). Not surprisingly, despite the wealth of palaeocommunity studies, use of biovolume measurements in palaeoecology has been rather limited (e.g. Walker 1972a, b\ Fiirsich and Wendt 1977; Stanton et al. 1981; Dzik 1979; Wiedman 1983, 1985) given the nature of the material (see Duff 1975). Nevertheless, they correspond approximately to original biomass and so provide a crude estimate of relative production and standing crop. These biovolumes reflect in some way the energy expended in growth. As Stanton et al. (1981; see also Powell and Stanton 1985) pointed out, however, when assessing biovolumes of carnivorous gastropods in an Eocene assemblage, no account is taken of energy expended in maintenance (respiration, excretion); these authors’ attempts to estimate a time-related cumulative biomass (as biovolume), while praiseworthy, were based on a number of assumptions that make any application to this study practically impossible. In any event, biovolumes may be a better guide to relative ecological importance than numbers of individuals (see also Staff et al. 1985) because of size variation between species, from a few mm (e.g. Lecytluoscopa) to over 300 mm long (e.g. Auomalocaris, Tegopelte), although in his study of a modern community Sanders ( 1 960) considered numbers to be more reliable than biomass estimates because only a few, large, randomly distributed animals may alter radically the measured values. The ratio between smallest and largest calculated biovolumes in average-sized specimens is several orders of magnitude. Comparing biovolumes between the major groups, however, is a very questionable exereise because of the range of body plans involved; sponges and cnidarians, with extensive water- and mesoglea-filled interiors respectively; priapulids, with spacious body cavities; and arthropods, with haemocoels and almost invariably at least a lightly skeletized exoskeleton. Intragroup comparisons of biovolumes, however, may be more valid. ANALYSIS OF THE FAUNA Numbers of genera, individuals, and bio volumes Genera. Several estimates are available of the composition of the entire Burgess Shale fauna in terms of numbers of genera, most of which are monospecific, within the principal groups (e.g. Conway Morris 1979a, b\ Conway Morris and Whittington 1979; Whittington 1980a). The present com- pilation (text-fig. 5) refers to the Phyllopod Bed only and excludes the few genera (Carnarvonia (Walcott 1912), Priscansermarinus (Collins and Rudkin 1981), Scolecofurca (Conway Morris 1 977a)) restricted to beds above the Walcott Quarry. Values given for all genera in text-fig. 5a differ little from earlier estimates. Text-fig. 5b shows the relative numbers of genera that had living representatives at the time of transport. It excludes those genera, mostly arthropods, known only from exuviae or other resistant parts lacking soft tissues. Apart from the decreased importance of the arthropods (especially trilobites) when only living genera are considered, comparisons of generic abundance in the two histograms (text-fig. 5a, b) show a broadly similar order of importance. It should be emphasized that only c. 14% of the genera with preserved soft parts have shelly skeletons (exeluding sponges, that would disaggregate into widely dispersed spicules) capable of being fossilized in normal circum- stances, although this value rises to approximately 20% when genera known only from exuviae or empty shells are included. An absence of soft-part preservation is particularly conspicuous in the polymeroid trilobites. Only four genera have soft parts recorded (Kootenia, Elrathina, Elimaniella, Olenoides), and praetically all (97%) are specimens of Olenoides. Judging by these samples the soft parts in the other trilobites are unlikely to have been entirely concealed; the erratic distribution of soft-part preservation may be due in part either to original rarity of most trilobite genera com- bined with a prolonged residenee time of the empty exoskeletons in the pre-slide sediments, or their 436 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 5. Relative percentages of the Phyllopod Bed (Burgess Shale) fauna in terms of numbers of genera (mostly monospecific) within each major group, a, genera with living representatives at time of burial plus genera represented by vacated shells and exuviae; h, genera with living representatives only. Stippled portion of arthropod bar refers to proportion of trilobites. post-mortem introduction from elsewhere, which would reduce even further their importance in the standing crop of the pre-slide community. The former proposal, however, does not readily explain the high ratio (c. 0-8 : 1 0) between living specimens and unexpectedly rare intact exuviae of Olenoides, unless the exoskeleton of this large trilobite was particularly susceptible to destruction by premoult demineralization or by benthic scavengers. In comparison, Cisne (1973a) recorded a similar live to dead ratio in specimens of the Ordovician trilobite Triarthrus eatoni from the celebrated Beecher’s Trilobite Bed, New York. His census included, however, fragmentary material and, in addition, practically all the protaspids and meraspids were represented only by exuviae. Accordingly, Cisne (\91hb) regarded almost 99% of the holaspid specimens as being alive immediately before burial. An analogous problem exists in the Devonian (Emsian) Hunsriickschiefer of West Germany, where soft-part preservation occurs in a number of trilobites including Phacops and "Asteropyge', but not in Parahomalonotus. In this instance Brassel and Bergstrom (1978) suggested that the potential for soft-part preservation was controlled by the position relative to the sediment-water interface. The infaunal Parahomalonotus is believed to have been capable of emerging subsequent to rapid burial, unlike the epifaunal Phacops and 'Asteropyge\ while if individuals of the former genus died in their burrows the openness of their domiciles would promote rapid decay. However, in Phacops and ' Asteropyge soft-parts are known only in c. 10% of individuals. Individuals. The relative importance of the principal groups in terms of numbers of individuals is shown in text-fig. 6a, b, again distinguishing between all specimens and only those individuals alive during transport. Arthropods predominate, while the important role of hemichordates results from the estimated abundance (see sources of data) of ‘D/ra/a’ tcawA (Walcott 19116) which is unrelated to the type species O. prolifica (Priapulida) (Conway Morris 1977a). ‘<9.’ tenuis has a bulbous anterior CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 437 TEXT-FIG. 6. Relative percentages of the Phyllopod Bed (Burgess Shale) fauna in terms of numbers of individuals and biovolumes of the major groups, a, individuals with living representatives at time of burial plus specimens represented by vacated shells and exuviae; b, individuals with living repre- sentatives only; c, calculated total biovolumes of each group. Stippled portions of arthropod bars refer to proportion of trilobites, and in h and c have been exaggerated for the purposes of clarity (see text). attached to an elongate trunk by a narrow stalk, an appearance which is reminiscent of hemichordate enteropneusts (unpublished observations). In the living fauna (text-fig. 6b) the remaining groups are relatively insignificant with none accounting for more than 4-5% of the total. A striking feature emerges in comparing the shelly components (trilobites, molluscs, brachiopods, very rare echino- derms): in text-fig. 6a they are fairly conspicuous, whereas in text-fig. 6h these groups are 438 PALAEONTOLOGY, VOLUME 29 insignificant. Their exact proportions in the living community are dependent on various assumptions (see sources of data), but our notion of the Phyllopod Bed eommunity and its ecology, and by implication some other Cambrian faunas, would be seriously warped if it had to be based solely on shelly fossils. Thus, if these figures are taken to be correct, only c. 2% of individuals (excluding sponges) alive at the time of burial had hard parts capable of fossilization in normal taphonomic circumstances. Even if it were assumed that half the shelly component assigned to the ‘dead’ category was actually alive at the time of burial (a rather optimistic assumption given the likely residenee time of empty shells and exuviae in the sediment), then their proportion of the standing crop in terms of individuals would be only about 12%. Such estimates take on special significance amongst the trilobites, as this group dominates most Cambrian assemblages. In the Phyllopod Bed community trilobites with robust skeletons (excluding Naraoia and Tegopelte: see Whittington 1977, 1985) account for less than 0-5% of all arthropod individuals alive at the time of burial (text-fig. 6b) and a correspondingly smaller fraction of the entire fauna. Even if estimates of trilobites alive at the time of burial are taken to include all intact exuviae (excluding presumed pelagic agnostoids and eodiscoids) the proportion of trilobites within the arthropod assemblage would fall short of 5%. Biovolumes. Text-fig. 6c depicts the relative biovolumes of the major groups. Arthropods maintain their pre-eminence, but in parallel with numbers of individuals the estimated percentage biovolume of trilobites within the arthropods is a trivial 1 %. The small size of ‘O.’ tenuis greatly reduces the significance of the hemichordates in comparison with numbers of individuals. The sponges and priapulids account for an appreciable fraction of the total biovolume as both are fairly numerous and large. The importance of the echinoderms is due to the medusoid-like Eldonia. Following Durham (1974) this organism is presently regarded as a holothurian, although a detailed redescription may well support Paul and Smith’s (1984, p. 496) reservations regarding its echinoderm affinities. Various comparisons between the major groups should be noted (text-fig. 6). For example, apart from the recurrent importance of arthropods, it is worth emphasizing that although the Miscellanea (which includes a number of bizarre species that if found today would probably be regarded as new phyla: see Conway Morris 19856) and chordates have attraeted wide interest, they are of rather minor significance in terms of individuals (text-fig. 66) and biovolume (text-fig. 6c). Comparisons of individuals and biovolumes. Three groups (arthropods, priapulids, polychaetes) are analysed in more detail with respeet to numbers of individuals and biovolumes to see whether these alternative methods of assessing relative importance give eompatible results (text-fig. 7). Intragroup comparisons of biovolume may be reasonably valid because of overall similarity in body architecture. In the priapulids (text-fig. la) and polychaetes (text-fig. 76) the distribution of numbers of individuals and biovolumes is fairly concordant, whereas wide discrepancies occur in the arthropods (text-fig. 7c) where large, relatively uncommon genera, e.g. Sidneyia (estimated 177 individuals with associated soft parts), outweigh in terms of biovolume far more numerous but smaller forms such as Marrella and Burgessia (estimated 15092 and 2158 individuals). Ecology: life habits Introduction. The ecology of the Phyllopod Bed fauna is assessed here in terms of life habits and trophic groups, both of which are divisible into a number of eategories (Table 2) based on standard schemes of palaeoecological classificatiop (e.g. Scott 1972, 1978; Walker 1972a; Walker and Bambach 1974; West 1977). With a few exceptions, such as symbiotic associations (?commensalism) between brachiopods and sponges (Conway Morris 1977a, 1982; Whittington 1980a) and possibly Wiwa.xia (Conway Morris 1985a), direct evidence of life habits (position relative to the water- sediment interface, tracks, burrows, and other traces) has been destroyed as a result of transport (text-fig. 1). Each genus is assigned a life habit, and those of some groups can be inferred with reasonable assurance. For example, all sponges and brachiopods are regarded as sessile epifauna even though some Cambrian inarticulate brachiopods were evidently infaunal (Pemberton and Kobluk 1978). CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 439 TEXT-FIG. 7. Comparison of numbers of individuals alive at the time of burial and biovolume in genera of three major groups in the Phyllopod Bed (Burgess Shale), a, priapulids; h, polychaetes; c, arthropods. 440 PALAEONTOLOGY. VOLUME 29 TABLE 2. Ecological classification used for the Phyllopod Bed fauna. Life habit Code Feeding type Code Infaunal, sessile IS Suspension, undifferentiated SU Infaunal, vagrant IV Suspension, high level (c. > 10 mm) su“ Epifaunal, sessile ES Suspension, low level (c. < 10 mm) su^ Epifaunal, vagrant EV Deposit, collector DC Nektobenthic NK Deposit, swallower DS Pelagic, floater PE Carnivores/scavengers PC Pelagic, swimmer PS Unknown UN Unknown UN Benthos. Considerable uneertainty surrounds the life habits of the arthropods, which are mostly classified as vagrant epifauna (see also Briggs and Whittington 1985). The extent of infaunal activities amongst the arthropods is conjectural and has only been briefly discussed for a few forms, e.g. Burgessia (Hughes 1975). Other epifaunal animals evidently included Wiwaxia as an adult, although juveniles may have been infaunal (Conway Morris 1985r/), and some of the genera of Miscellanea. This latter group includes Hallucigenia which appears to have supported its body and walked on seven pairs of sharply pointed stilt-like appendages that lacked ‘snow-shoe’ adaptations (see Thayer 1975) and which are comparable, for instance, to the distal rosettes of spines found in arthropods that patrolled the soft substrates of the Devonian Hunsriickschiefer (e.g. Seilacher 1962; Stiirmer and Bergstrom 1978). While the phyletic relationships of Hallucigenia are obscure (Conway Morris 19776), ecological analogues apparently exist. They include the tripod fish (Bathypterois and its relatives) that rests on the sea-bed with the aid of enormously elongate spines arising from the tail and pectoral fins; pycnogonoids; and elasipod holothurians, e.g. Scotoplanes, that support their bodies on the tips of their tube-feet (e.g. Barham et al. 1967; Heezen and Hollister 1971; Herring and Clark 1971; Lemche et al. 1976). Hallucigenia was presumably very close to neutral buoyancy, and locomotion over muddy substrates must have been feasible unless there was a deep layer of flocculent material. On the basis of its calculated biomass, M. LaBarbera (pers. comm.) has suggested that Hallucigenia would only be stable in areas of very low current velocity (< 2-5 m/s) unless it had additional means, e.g. its tentacles, of securing itself. In this context, it is interesting to note that elasipod holothurians are swept off their podia and tumbled about by the movement of a passing bathyscape, indicating that they too have only very slight negative buoyancy (Barham et al. 1967). Amongst the infauna, all the priapulids apparently burrowed but some species may have made at least occasional excursions over the sea-bed in the same manner as modern priapulids (Menzies et al. 1973, fig. 7-15f). Furthermore, the priapulid Louisella may have been more sedentary. Its dorsoventrally compressed body with two rows of gill papillae along much of its trunk finds an analogy with some eunicid polychaetes which occupy horizontal burrows and drive respiratory currents over the gills by dorso ventral undulations of the body (Clark 1964). The proboscis scalids of Louisella are unusual amongst priapulids in being papillate, and by secreting mucus they could have shored up the walls as the animal moved along its burrow. Nothing is known about either the relative depths occupied by various infauna or the maximum depth of burrowing. The abundance of large priapulids would indicate, however, that bioturbation was pronounced (cf. Miller and Byers 1984). Nektobentlios. Nektobenthic animals are identified on one or more criteria; adaptations consistent with a swimming ability, relative abundance indicating a near-bottom habit within the range of the turbidity flows, and inferred feeding habits inconsistent with an entirely pelagic mode of life. Examples of nektobenthos appear to include the polychaete Canadia (Conway Morris 1979c), with large paddle-like neurosetae, and the laterally compressed chordate Pikaia. Eldonia has been depicted as pelagic (Durham 1974) and, while its abundance at one horizon (Walcott 1911a; Whittington 1971a) could have resulted from the trapping of a shoal, this medusiform echinoderm is CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 441 provisionally regarded as nektobenthic. The two species of the swimming arthropod Perspicaris are rare, but as they were possibly deposit feeders (Briggs 1977) a pelagic existence presumably can be excluded. Odaraia was probably another nektobenthic arthropod (Briggs 1981), while among the miscellaneous category both Opabinia (Whittington 1975^?) and Anomalocaris (Whittington and Briggs 1985) appear to have been well adapted for swimming above the substrate. The pelagic com- ponent (Conway Morris 1979a, b, c\ Whittington 1981^) apparently includes trilobites (agnostoid and eodiscoid: but see above), and rare soft-bodied animals which may have had a gelatinous composition (Amiskwia, Odontogriphus), streamlined shape with fins (Nectocaris), or prominent natatory organs {Insolicorypba, Sarotrocercus). Conclusion. Text-fig. 8 depicts the relative proportions of life habits, excluding pelagic animals which presumably represent an incomplete sample from this ecological zone, in terms of numbers of genera, individuals, and biovolume. The histograms also show the most important group in each category. The general importance of the epifauna is clear, with the vagrant and sessile divisions being dominated by arthropods and sponges respectively. 'Ottoia tenuis is prominent in the sessile infauna and is inferred to have had a life habit similar to most modern enteropneusts. The vagrant infauna is dominated by priapulids. The important role of the echinoderms in the nektobenthos in terms of individuals and biovolume is due to Eldonia (see Durham 1974). HABITAT TEXT-FIG. 8. Relative percentages in the Phyllopod Bed (Burgess Shale) of major categories of life habits (NK, nekto- benthic; EV, epifaunal vagrant; ES, epi- faunal sessile; IV, infaunal vagrant; IS, infaunal sessile) in terms of genera, bio- volume, and individuals alive at the time of burial. The most important group within each category is also shown. 442 PALAEONTOLOGY, VOLUME 29 Ecology: trophic groups Introduction. Notwithstanding the soft-part preservation, there are several problems in assessing feeding habits. The extent to which opportunistic feeding (see Cadee 1984) took place is not known and some Phyllopod Bed species may have had catholic diets. It is unfortunate that relatively few species have identifiable gut contents, e.g. hyolithids in Ottoia\ trilobites, brachiopods, and hyolithids in Sidneyia. Accordingly, the structure of the food-collecting apparatus is used as an additional source of evidence. In most cases, therefore, feeding habits are more speculative than life habits and for a substantial fraction of the fauna they remain unknown, a recurrent problem in palaeoecology (e.g. Scott 1978). Principal feeding categories. The inferred distribution of feeding types with respect to numbers of genera, individuals, and biovolume, and the most important group within each category, is shown in text-fig. 9. Briggs and Whittington (1985) have recently provided a more detailed review of arthropod feeding habits. In text-fig. 9 no distinction is made between high- and low-level suspension feeders, although such information has been presented graphically elsewhere (Conway Morris \919h). Amongst suspension feeders, sponges predominate in terms of number of genera and biovolume, but the abundance of ‘C.’ tenuis is reflected in the count of individuals. Deposit (collector) feeders are another important trophic group, with arthropods predominating. Nothing is known of the type of particulate matter (see Johnson 1974) utilized by the various deposit feeders, and such information may not be readily available because of diagenetic alterations. Difficulties in distinguishing scavenging from carnivorous habits in some forms means that they are best taken as a single trophic group. In terms of numbers of genera and individuals this category is not especially notable, but the biovolume of carnivores/scavengers is significant, mostly because of the inclusion of the large FEEDING HABITS ARTHROPODS HEMICHORDATES MOLLUSCS PRIAPULIDS SPONGES TEXT-FIG. 9. Relative percentages in the Phyllopod Bed (Burgess Shale) of major trophic categories (DC, deposit, collector; DS, deposit, swallower; SU, suspension; PC, carnivore or scavenger) in terms of genera, biovolume, and individuals alive at the time of burial. The most important group within each category is also shown. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 443 arthropod Sidneyia which has gut contents and a feeding apparatus consistent with predatory habits ( Bruton 1 98 1 ). Its importance may have been overestimated because intact exuviae are common and some specimens interpreted as being alive during transport may belong to this category. Nevertheless, Sidneyia was probably the most significant predator in the fauna. Together with other carnivorous arthropods (e.g. Olenoides, Whittington 1975/); Naraoia, Whittington 1977), priapulids (e.g. Ottoia, Conway Morris 1977a), and incertae sedis such as Anonialocaris (Whittington and Briggs 1985), the abundance of Sidneyia emphasizes the role of this trophic group in the Phyllopod Bed fauna. With regard to predation in the Phyllopod Bed community the example of hyolithids in the alimentary tract of Ottoia (Conway Morris 1985a) is especially interesting. First, apart from their location, the ingested shells appear to show no evidence of having suffered predation (e.g. scratch marks, crushing), and once returned to the sediment would presumably be indistinguishable from those that perished by more mundane causes. Secondly, a comparison between the number of hyolithids recorded in the guts of Ottoia and Sidneyia, as against those comprising the free-living standing crop (see sources of data), is indicative of high levels of predation (Conway Morris 1985a). Furthermore, apparent defensive adaptations, such as the spines of Whvaxia, and more circum- stantial indications, e.g. frequent breakage of the slender tubes of Tubulella into short sections, may be added to the roster of evidence. Possible evidence for scavenging (in the post-slide environment) by ostracodes on Marrella was discussed by Conway Morris (1977a). The relative abundance of this association in the Marrella population is very small (c. 0- 1 5 %), but the specificity of the ostracodes’ location between the median and lateral cephalic spines is so consistent that it may be interpreted better as a symbiotic association formed in the pre-slide environment. Trophic nucleus. This concept, which is based on Neyman’s (1967) criterion of the dominant species that together comprise at least 80% of a fauna’s biomass, has been used in a number of palaeo- ecological studies, with the number of individuals necessarily substituting for unavailable data on biomass (e.g. Rhoads et a!. 1972; Duff 1975; Fiirsich 1977; Fiirsich and Wendt 1977). Here the TROPHIC NUCLEUS TEXT-FIG. 10. The trophic nucleus of the Phyllopod Bed (Burgess Shale) fauna in terms of biovolume and numbers of individuals alive at the time of burial. Abbreviations for ecological categories given in Table 2. ARTHROPOD m CNIDARIAN I C | ECHINODERM m HEMICHORDATE | H | PRIAPULID I P I SPONGE I S I INDIVIDUALS 444 PALAEONTOLOGY, VOLUME 29 nucleus is identified both with respect to number of individuals and biovolume, the rank order repre- senting the five and ten most abundant genera respectively within the benthic community (text- fig. 10). The trophic nucleus is more compact when considered in terms of individuals. In each case the dominant species is a deposit (collector) feeding arthropod, but the ranking does not show a close match between the two categories. Thus two of the genera, "O' tenuis and Burgessia, which rank second and fourth in the trophic nucleus of individuals, have low total biovolumes that exclude them from the latter trophic nucleus. In either case the make up of the trophic nucleus demonstrates that the Phyllopod Bed community was dominated by relatively few species, with c. 10% of the benthic species (ranked 1-9 in abundance out of 93) accounting for c. 91 % of individuals and 82% of bio- volume of the entire benthic fauna. This situation is comparable to many modern communities (e.g. Sanders 1960), but direct comparisons between the trophic nucleus of the Phyllopod Bed and ‘normal’ fossil assemblages is unrealistic because, as Johnson ( 1 964) stressed, the marked dominance of few shelly species in the latter assemblages is likely to be an artefact of preservation rather than a reflection of original biotic proportions. Turpaeva (1957) drew attention to the widespread, although not invariable, alternation of trophic types in declining species rank within modern benthic communities from the boreal and arctic seas around the USSR; such alternation being linked to division of resources and diminution in feeding competition. Turpaeva’s (1957) emphasis on organisms belonging to lower trophic levels, which are often well skeletized (e.g. Schopf 1978; Nicol 1979, 1982), has found considerable palaeontological application (e.g. Rhoads et al. 1972; Walker 1972a, 6; Duff 1975; Fiirsich 1977; Fursich and Wendt 1977; see also Cisne 19736). This analysis follows Duff (1975) in incorporating a wider than usual range of feeding types. In terms of biovolume, no adjacent species within the trophic nucleus share a feeding habit, although the trophic category of Helmetia is not known. In terms of individuals, however, the arthropods Canadaspis and Burgessia, which rank third and fourth in the trophic nucleus, are both interpreted as having the same feeding habit. Whether there was competition for the same resource is debatable. These arthropods differ considerably in size, and the anatomy of their feeding limbs is not especially similar (Hughes 1975; Briggs 1978f It may also be significant that these arthropods appear to show evidence for negative association, in that although both arthropods are very abundant (estimated 2158 specimens of Burgessia, 47 1 9 of Canadaspis alive at time of burial), co-associations on sawn slabs are rare. Trophic web. Data on feeding types have been used to reconstruct a tentative trophic web (text- fig. 1 1 ), with the importance of major taxonomic groups being assessed in terms of numbers of individuals and biovolume. Direct evidence for primary producers is available from macroscopic benthic algae (Walcott 1919), while G. Wood (pers. comm., 1982) reported an acritarch assemblage dominated by sphaeromorphs and small acanthomorphs. The importance of primary consumers is evident, with suspension feeders presumably capturing phytoplankton and suspended detritus, while deposit feeders are assumed to have exploited bacteria (and other microbial forms), benthic algae, and detritus. A higher trophic level of carnivores is also identified and its significance in the fauna is apparent, especially with respect to biovolume. There is no direct evidence that carnivores occupied more than the primary level, and the apparent absence of secondary or tertiary levels gives the trophic web as reconstructed a relatively simple structure. Establishing whether this simplicity is original is important if meaningful comparisons are eventually to be drawn with trophic webs of younger fossil assemblages and living communities. The restriction of carnivores to a single level and apparent absence of one or a few top carnivores may well be because of incomplete information. Estimates of the efficiency of energy transfer between primary consumers and carnivores are uncertain (see below), but there is no reason to think that relatively inefficient mechanisms of either capture or physiology placed a prohibitive energetic constraint on the existence of top carnivores. More speculatively, appeal could be made to the lower levels of productivity that have been inferred by some authors for the Lower Palaeozoic (e.g. Calef and Bambach 1 973) constraining the length of food chains (see also Tappan 1970), and this topic is returned to below. The few known examples of predatory genera identified in text-fig. 1 1 serve to emphasize both the CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 445 Primary producers PHYTOPLANKTON benthic ALGAE Primary consumers Carnivores HEMICHORDATES ' 1 Suspended detritus ECHINODERMS PRIAPULIDS JZH , '^^Ottoia ARTHROPODS XX Benthic detritus / Bacteria - MOLLUSCS f HVQLITHID POLYCHAETES, MISCELLANEOUS CNIDARIANS UNKNOWN TROPHIC GROUPS POLYCHAETES G MI3CEL! ANf'^US , , J^PRIAPULIDS fsio. OF INDIVIDUALS [J2500 BIOVOLUME Lj 2500 cm3 CARNIVORES PRIMARY CONSUMERS SCALES INDIVIDUALS I \ 6000 BIOVOLUME i i 1200cm^ TEXT-FIG. 1 1 . Tentative reconstruction of the trophic web in the Phyllopod Bed (Burgess Shale) community in terms of numbers of individuals alive at the time of burial and biovolume (stippled). Genera known to be carnivores on the basis of gut contents {Ottoia, Sidneyia) or intimate association with supposed food (Aysheaia) are depicted, but must represent only an incomplete part of the transfer of energy to the higher trophic level. Cannibalism has been documented in Ottoia. See text for further discussion. scarcity of data and a resulting inability to reconstruct this web in more detail. Evidence for predation by Ottoia, which appears also to have been cannibalistic (Conway Morris 1977a), and Sidneyia was alluded to above. Frequent association between Aysheaia and several species of sponge suggests that it fed on these organisms. Comparison between individuals and biovolume totals of primary consumers (deposit and suspension feeders taken together, see Stanton and Nelson 1980; cf. Scott 1972; Fiirsich and Wendt 1977) and carnivores allows approximate estimates of the degree of 446 PALAEONTOLOGY, VOLUME 29 energetic efficiency operating between these two trophic levels: in terms of individuals the percentage of predators in comparison with deposit and suspension feeders is c. 7%, while for biovolume the equivalent value is much higher at 52 %. The lower value is comparable to some other estimates of the efficiency in energy flow between marine trophic levels (c. 10-12%; Stanton and Nelson 1 980; see also Powell and Stanton 1985), but it is based on individuals with wide size ranges and, by implication, calorific values. The presumably closer relationship between biovolume and original biomass might suggest that the value of 52% is slightly more accurate, assuming that biomass of all species is pro- portional to the same energetic equivalents. The tenuous connection between biovolume and bio- mass, however, together with problems in calculating accurately biovolume of a population and questionable assumptions, such as the apparent absence of second (and third) levels of carnivores, suggests that this higher figure is an overestimate. Given the problems inherent in palaeoecology it is not surprising that relatively few workers (e.g. Kapp 1975; Fiirsich and Wendt 1977; Stanton and Nelson 1980) have attempted to reconstruct the trophic web of a given marine fossil assemblage. Text-fig. 1 1 depicts what is the most complete trophic reconstruction so far compiled for a Lower Palaeozoic marine community, yet it is sobering to consider that, despite extensive soft-part preservation, neither a detailed reconstruction of the trophic web with reliable indications of feeding habits of individual species, nor a well-constrained estimate of the efficiency of energetic transfer is available for the Phyllopod Bed community. The relative importance of combinations of life habit and feeding type with respect to numbers of genera and individuals and biovolume (Table 3) shows that, in terms of number of individuals, ecological category EVDC (epifaunal, vagrant, deposit (collector) feeder; see Table 2) is the most important, with ISSU (infaunal, sessile, suspension feeder) rating a low second. With regard to biovolume, EVDC is again significant, but other notable categories include NKSU (nektobenthic. TABLE 3. Relative importance of various ecological categories in the Phyllopod Bed fauna (see Table 2 for explanation of abbreviations). Pelagic and benthic components are treated separately. Excluded combinations include PSDC, PSDS, PFDC, and PFDS; unknown but theoretically possible combinations include NKDC, NKDS, EVSU, ESDC, ESDS, IVSU, ISDS, and ISPC. Ecological category Number of genera Per cent individuals Per cent biovolume Pelagic: PSSU 2 80 00 39-70 PSPC 2 13-30 6-40 PFSU 1 6-70 53-90 Benthic: NKSU 3 2-08 15-34 NKPC 3 0-56 1-81 NKUN 1 0-06 0-02 EVDC 17 59-06 26-87 EVDS 3 1-05 0-05 EVPC 8 1-32 16-39 EVUN 10 0-96 6-43 ESSU” 24 3-87 16-81 ESSU^- 9 0-73 0-03 ESPC 1 0-22 3-30 ESUN 1 <0-01 <0-01 IVDC 3 0-19 0-22 IVDS 1 0-03 <0-01 IVPC 4 4-45 10-00 IVUN 2 0-04 0-27 ISDC 1 0-94 0-08 ISSU 1 23-01 1-04 UNUN 1 1-42 1-34 CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 447 suspension feeder), EVPC (epifaunal, vagrant, carnivore/scavenger), ESSU^ (epifaunal, sessile, high level suspension feeder), and IVPC (infaunal, vagrant, carnivore/scavenger). Dominance diversity and possible niche parlitioning Introduction. Dominance-diversity curves were prepared for a number of ecological categories in the Phyllopod Bed fauna. Such curves have found wide ecological application (e.g. Whittaker 1965, 1972, 1975; May 1975, 1981) and, despite problems of time-averaging and preservation bias in fossil assemblages, this technique has been employed on occasion by palaeontologists (e.g. McBride 1976; Schwartz et al. 1977; McGhee 1981) and in sub-fossil assemblages by palaeolimnologists (e.g. Goulden 1966, 1969; Deevy 1969; but see Smol 1981), while Driscoll and Swanson ( 1973) considered potential palaeontological applications with particular regard to epifaunal assemblages. Much of the interest in dominance-diversity curves has been concerned with their depiction of resource subdivision and so their possible use as indicators of niche partitioning. Three principal distributions have been noted; particularly useful discussions regarding them have been provided by Whittaker (1972) and May (1975). Two of the distributions relate to models that describe resource appropria- tion in terms of ‘random niche boundaries’ (often referred to as the broken-stick model; not considered further here, see below) and ‘niche pre-emption’. The third, the log-normal distribution, traditionally has been interpreted as a mathematical artefact arising from the Central Limit Theorem and resulting from the interaction of many more or less independent factors that reflect a wide number of variables, most typically involving a large number of species with little resource overlap. More recently, however, biological explanations of a log-normal distribution have been offered. Sugihara (1980) proposed that this distribution was a consequence of the segmental division of niche space in a hierarchically structured community, while Ugland and Gray (1982; see also Shaw et al. 1983) proposed an alternative explanation and linked log-normality with community equilibrium. A geometric distribution is usually linked to the niche pre-emption model, in which the dominant species takes a given fraction (k) of a resource, the next dominant species a fraction (k) of the remainder of the resource, and so on. Results. The dominance-diversity curves for four simple ecological categories (IVPC, EVPC, EVDC, and ESSU”; see Table 2), all epifaunal vagrant animals (EV/DC-l-DS-l-PC-l-UN), and the entire benthic fauna (i.e. excluding PS + PE), are depicted in text-fig. 12, while the relevant statistics are given in Table 4. Each graph was prepared in the standard manner by plotting the abundance, in this case according to both numbers of individuals and total biovolume, against the rank abundance (1-/7, most to least abundant). Visual inspection of the curves is liable to be misleading, and they were analyzed statistically for goodness-of-fit (square correlation, r^) for each of the three principal distributions mentioned above (Table 4). These statistics and calculations of n diversity (the widely used Shannon- Wiener index ^ Pi^ogjPi, where p, is the proportion i = 1 of the /th species) and evenness (e = //'/log 5, where 5 is the total number of species) were kindly computed by J. J. Sepkoski (University of Chicago). The test statistics also allowed the possibility of further sampling by the hypothetical addition of ‘undiscovered’ species. This is the equivalent procedure to drawing the ‘veil-line’ of a log-normal distribution to the right to reveal rarer species (Preston 1962). The large samples from the Phyllopod Bed suggests that the differences between the observed and original diversity may have been relatively small, and that the hypothetical addition of a large number of undiscovered species to each category may be unwarranted. In no case did the broken-stick (random niche boundary) model provide a good fit to the dominance-diversity data. In the majority of cases the distribution in terms of both numbers of individuals and biovolumes was best explained by a log-normal distribution (text-fig. 12; Table 4). In the case of the larger categories of all epifaunal vagrant animals and the entire benthic fauna such a distribution may well be an artefact reflecting their heterogeneous ecological composition, but for smaller categories the alternative biological explanations of Sugihara (1980) and Ugland and Gray (1982), in principle, are feasible. In a few cases a geometric distribution provided the best-fit, and 448 PALAEONTOLOGY, VOLUME 29 TABLE 4. Distribution (log-normal or geometric) that best explains dominance-diversity curves in seven ecological categories or portions of the Phyllopod Bed (Burgess Shale), with calculated goodness of fit. Where hypothetical addition (up to six) of taxa alters the distribution, this is indicated. The geometric distribution remains unaltered if further taxa are added. Values of diversity (Shannon-Wiener H') and evenness (e) are also given for each category or portion (see text for further details). Category Individuals Biovolume IVPC Log-normal (0-9791) H' = 0-587,6 = 0-424 + 2 -f4 Geometric (0-9698) Log-normal (0-9808) Log-normal (0-9834) H' = 0-405, 6 = 0-292 EVPC + 2 + 6 Geometric (0-971 1) Log-normal (0-9776) Log-normal (0-9763) H'= 1-767,6 = 0-850 Log-normal (0-9620) H' = 0-287,6 = 0-138 EVDC Geometric (0-9669) H'= 1-103,6 = 0-407 Log-normal (0-9397) H' = 0-764, 6 = 0-282 ESSU” + 6 Log-normal (0-9765) Geometric (0-9673) H' = 2-149, 6 = 0-676 + 6 Log-normal (0-9058) Geometric (0-8581 ) H' = 1-982,6 = 0-624 EV/DC-^DS-ePC-eUN Log-normal (0-9378) H'= 1-343,6 = 0-367 Log-normal (0-9873) H'= 1-675,6 = 0-457 Entire fauna, excluding PS-rPF Log-normal (0-9549) H' = 2-094, 6 = 0-462 Log-normal (0-9891) H' = 2-624, 6 = 0-579 Fauna, hard parts Geometric (0-9144) H' = 1-843,6 = 0-699 Log-normal (0-8743) H' = 0-635,6 = 0-240 therefore might conceivably reflect niche pre-emption (Table 4). In categories IVPC (biovolume) and EVPC (individuals), however, hypothetical addition of species resulted in the distribution being better explained by log-normality. Conversely, further hypothetical sampling in category ESSU” (individuals and biovolumes) resulted in a geometric distribution providing a better fit than a log- normal distribution. Finally, in category EVDC (individuals) a geometric distribution was maintained in the face of further sampling. The dominance-diversity data for those elements of the fauna with hard parts (excluding sponges, but including rare benthic echinoderms) that were alive at the time of burial were also assessed (Table 4), as this distribution is relevant to that portion of the standing crop that would be expected to fossilize and provide a Typical’ assemblage (see also Easker 1 976). The figures are based on major assumptions (see sources of data), and a geometric distribution providing the best fit in terms of individuals is unlikely to be of any biological significance. Conclusion. These results broadly accord with a more extensive survey of dominance-diversity in fossil assemblages by Schwartz et al. (1977), who noted the predominance of log-normal distributions. Nevertheless, it seems possible that in at least ecological category EVDC partitioning resources followed the niche pre-emption model. In a study of Upper Cambrian communities McBride (1976) portrayed dominance-diversity curves of polymeroid trilobites, reconstructed as TEXT-FIG. 1 2. Dominance-diversity curves for four ecological categories (IVPC, EVPC, EVDC, ESSU^) and two major portions (EV/DC-l-DS-l-PC-l-UN; entire benthic fauna, excluding PS 4- PE) of the Phyllopod Bed (Burgess Shale) fauna in terms of individuals alive at the time of burial (filled circles) and biovolume (open circles). See Table 2 for explanation of abbreviations; see Table 4 for preferred distribution (log-normal or geometric), goodness of fit (U), diversity (H'), and equitability (e). log, abundance log, abundance log, abundance CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 449 IVPC (n=4) o o <9 2 3 4 Rank D O — I n EVD C ( n = 15 ) « • o 5 10 Rank 15 10 20 Rank 450 PALAEONTOLOGY, VOLUME 29 epifaunal vagrant deposit feeders, and interpreted his data as probably reflecting niche pre-emption. Re-analysis of his data, however, does not support a geometric distribution (J. J. Sepkoski, pers. comm.), although McBride’s (1976, p. 147) comments on the likelihood of severe competition and niche partitioning amongst deposit feeders may still hold. COMPARISONS BETWEEN THE PHYLLOPOD BED AND OTHER COMMUNITIES Cambrian communities Introduction. The ecological data from the Phyllopod Bed fauna provide a unique insight into the structure of a relatively deep-water Cambrian community. Useful comparisons with other Cambrian assemblages are dependent on the observation that, had the Phyllopod Bed fauna failed to experience those conditions conducive for soft-part preservation, then (as noted above) the impoverished shelly fauna would be indistinguishable in general taxonomic outline from most other Cambrian assemblages. The minor role of shelly taxa in the ecological spectrum of the Phyllopod Bed, especially with respect to numbers of individuals, means that synecological pronouncements based on ‘typical’ Cambrian faunas are likely to be insecure. Detailed comparisons with other fossil assemblages of Cambrian age are not straightforward however, both on account of a general absence of soft-bodied forms and the inhabited range of marine environments. Ecological comparisons. On the basis of shelly faunas, Cambrian marine life has been characterized as ecologically generalized, with broad niches and a relatively simple trophic structure dominated by detritus and low-level suspension feeders. In general, such communities are identified as having wide environmental tolerances that are reflected in broad distributions and diffuse boundaries (e.g. •Valentine 1973; Sepkoski 1979; Sepkoski and Sheehan 1983). Information on the overall distribution of the Phyllopod Bed fauna and its relative distinctiveness in comparison with other benthic communities in the off-reef Stephen Formation is difficult to assess on present evidence, although the recurrent nature of a Raymond Quarry-like biota is some evidence for community identity (Collins et al. 1983). Ecologically, however, the Phyllopod Bed community does not entirely correspond with the characterization of Cambrian life as based on the shelly component. It is indeed true that detritus and suspension feeders predominate (text-figs. 9-11), but the possibility of niche partitioning in at least one group of deposit feeders (text-fig. 12; Table 4) may point to a hitherto unsuspected ecological complexity. With regard to epifaunal suspension feeders, their distribution and history of tiering is reviewed by Ausich and Bottjer ( 1 982). The existence in the Phyllopod Bed of a low level with mostly brachiopods and rare edrioasteroids, and higher levels dominated by sponges together with much rarer pelmatozoan echinoderms (eocrinoids and primitive ?crinoids) corresponds well with Ausich and Bottjer’s (1982) descriptions, but the maximum height above the substrate may have been twice as high (26 cm) as they reported (see Conway Morris 19796). In terms of trophic analysis, however, nowhere is the contrast between the Phyllopod Bed fauna and other Cambrian assemblages more apparent than on the role of predation (text-figs. 9 and 11). The presence of a significant proportion of predators in the Phyllopod Bed suggests that trophic descriptions of other Cambrian communities have been biased severely by fossilization potential. On the basis of shelly faunas, macrophagous predators have been inferred to be either rare or even absent, especially in the Lower Cambrian (e.g. Glaessner 1972; Valentine 1973, 1975). In one of the very few Cambrian community studies, McBride’s (1976) analysis of open shelf and ‘shoaling’ shelf biofacies in the Upper Cambrian Dunderhergia Zone of western USA, which was based on shelly faunas, made no mention of carnivores or scavengers. Bambach (1983) claimed in his synthesis of marine guilds and ecologies through geological time that Cambrian faunas have few carnivores, most of which were referred to vaguely as infaunal ‘polychaetes’ (his fig. 3). Hutchinson (1961; see also Matthew 1891; Bengtson 1977) perceptively suggested, however, that the absence of Cambrian predators may be an artefact arising from their having a minimal fossilization potential, a trait that continues into present- day marine communities (Schopf 1978; Nicol 1982). The abundance of effectively soft-bodied CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 451 carnivores in the Phyllopod Bed combined with various other lines of evidence (Conway Morris I985u; Conway Morris and Jenkins 1985) amply confirms Hutchinson’s insights. Even apparent carnivores with hard parts, such as the trilobite Olenoides (Whittington 1975^, 1980/r), would not be readily attributable to this trophic group in the absence of exceptional preservation. This is because, paradoxically, the feeding appendages have a minimal preservation potential, and the exoskeleton is unlikely to provide direct evidence on the style of feeding. Regarding the other major ecological category discussed above, that of life habits (Table 2), direct comparisons between their relative proportions in the Phyllopod Bed (text-fig. 8) and other Cambrian communities are not straightforward, both because of severe taphonomic bias in ‘normal’ shelly faunas and a lack of published studies. One instance where a comparison may have some validity is in the slightly younger Wheeler Formation of Utah, where the frequency distribution of the principal shelly groups is statistically similar to the equivalent component from the Burgess Shale (Conway Morris and Robison 1982). From one prolifically fossiliferous horizon in the House Range, which has yielded some exceptionally preserved material, Robison (1985) commented on the absence of an infauna and the predominance of a vagrant epifauna. In the overlying Marjum Formation Rogers (1984) speculated on the likelihood of some communities being dominated by epifaunal sponges. In a broader survey of Upper Cambrian communities McBride (1976) portrayed a far less diverse range of life habits, although in this instance taphonomic factors must be a source of severe bias. Younger Palaeozoic communities Introduction. Taxonomic differences aside, it may transpire that the community structure of the Phyllopod Bed fauna was not fundamentally diflferent from that of many younger Palaeozoic soft- bottom faunas (see also Bretsky 1969) in that patterns of resource supply and distribution presumably did not change radically until at least the late Silurian, when the introduction of abundant terrestrial plant debris may have influenced community structure (Bambach 1977; see also Calef and Bambach 1973). Calef and Bambach (1973) argued that in the absence of a pre-Silurian terrestrial flora and its associated influence on rates of chemical weathering, nutrient supply into the shallow seas was drastically diminished while ‘Deep basin benthos was generally excluded for lack of food’. While nutrient supply may have been lower during the early Palaeozoic, the diversity and abundance (reflecting biomass) of Phyllopod Bed organisms casts some doubt on Calef and Bambach’s (1973) general thesis. Moreover, deep-water Cambrian (Hales Limestone) faunas, possibly living in excess of 1600 m (Cook and Taylor 1977), also demonstrate that a deep basin benthos was present and very widespread. Admittedly the trilobite faunas are not diverse (Taylor 1976; Taylor and Cook 1976) and the scarcity of burrowing (Taylor and Cook 1976; Cook and Taylor 1977) indicates a restricted infauna, but the abundance of sponge spicules suggests the former presence of suspension feeders, a mode of life that is energetically marginal (Jorgensen 1966). Discussion of nutrient levels in Cambrian oceans and their comparison with younger systems is perhaps more one of degree than kind. Palaeo-oceanographic factors must have been significant, and productivity of the Phyllopod Bed fauna, for example, may have been favoured by proximity to the reef and more generally seasonal (winter) upwelling (see Parrish 1982, fig. 5). Evolution of ecological categories. Various attempts have been made to study trophic replacement in marine communities through geological time (e.g. Ager 1976; McGhee 1981). The tracing of ecological analogues of deeper-water Cambrian communities, as typified by the Phyllopod Bed fauna, through time is complicated by specialization (niche partitioning) and associated proliferation of taxa, especially in the Ordovician, and the longer-term decline of major groups and the rise of others. At least two approaches, however, hold some promise. First, there have been repeated suggestions that oxygen levels in early Palaeozoic oceans were considerably less than those of the present-day. With particular reference to the Phyllopod Bed community, there is evidence to suggest that it was adjacent to anoxic water and, given such proximity, may itself have been within a dysaerobic zone (see Rhoads and Morse 1971). In describing metazoan faunas from a Canadian fjord subject to low oxygen levels, T unnicliflfe (1981) suggested that they might provide a useful analogue to 452 PALAEONTOLOGY, VOLUME 29 early Palaeozoic faunas. Evidence of at least episodic states of low oxygenation in early Palaeozoic oceans is known (Leggett et al. 1981) and a specific study through time of metazoan communities inhabiting poorly oxygenated environments, for which various criteria exist (see Savrda et al. 1984), may prove fruitful. Byers (1979) has provided initial pointers in his study of the ichno-faunas of dysaerobic basins through geological time, and Savrda el al. (1984) have reviewed critically some applications arising from this and related work. How effectively sueh ecological analogues can be pursued through geological time, however, is not yet clear. Irrespective of similarities or differences in biotic and ecological composition, it is, of course, no coincidence that the likelihood of soft-part preservation would be greatly enhanced in such dysaerobic-anoxic environments, especially given the prevalence of fine-grained sediment accumula- tion and opportunities for catastrophic burial, such as that provided by sediment put into suspension by storms or turbidity currents. A more general method of tracing ecological analogues lies in an examination of deeper-water faunas through geological time. It has been claimed that there is an off-shore displacement of benthic communities, corresponding to the three great ‘evolutionary faunas’ of the Phanerozoic (Cambrian, Palaeozoic, Modern) through time, with the relatively archaic ‘Cambrian fauna’ migrating into deeper-water during post-Cambrian times (Jablonski et al. 1983; Sepkoski and Sheehan 1983; see also Axelrod 1958; Berry 1972; Rowland et al. 1984). According to this hypothesis, younger equivalents to the Phyllopod Bed community may be best sought in deeper-water facies (but see Bretsky and Klofak 1985). Of the relevant Upper Cambrian and Ordovician communities from North America listed by Sepkoski and Sheehan (1983), only Beecher’s Trilobite Bed in the Frankfort Shale (Caradoc) shows soft-part preservation (Cisne 1973Z?, 1981). Elsewhere diversity information is limited to shelly species, so comparisons are not straightforward. The following attempt to document the broadest patterns of ecological replacement in deeper- water faunas is based on the simple trophic groups of deposit feeders, suspension feeders, and carnivores. In Beecher’s Trilobite Bed deposit feeders appear to have included trilobites and ostracodes, and these two groups figure prominently in the almost coeval Triarthrus eommunity of the Trenton Group (Titus and Cameron 1976). Trilobites, and more especially ostracodes, may represent the ecological descendants of the extraordinarily diverse arthropod fauna, dominated by deposit feeders (text-fig. 9), of the Phyllopod Bed. This transition may have occurred by the Upper Cambrian in that at least some relatively deep-water sediments (Orsten) had a diverse micro-arthropod fauna, including ostracodes and small crustaceans (e.g. Muller 1979, 1981, 1982, 1983, 1985). However, as Muller stressed, much of the contrast between these Upper Cambrian arthropod faunas and those of the Burgess Shale could be a result of different modes of preservation, with selective phosphatization of small fossils in the Upper Cambrian Orsten. The phylogenetic relationships between these Middle and Upper Cambrian arthropods are obscure (Muller 1983), and it is therefore the more interesting to note a further example of apparent ecological continuity between the marrellomorph arthropod Mimetaster from the Lower Devonian Hunsriickschiefer (Birenheide 1971; Stfirmer and Bergstrom 1976) and the apparently related Phyllopod Bed Marrella (Whittington \91\b). Other deposit feeders, at least locally, were the molluscs. In the Cambrian these were predominantly monoplaco- phorans, while in some younger communities (especially Silurian) ecological descendants may have included infaunal bivalves with rare gastropods (and monoplacophorans) also ecologieally replacing in part the trilobites (e.g. Hurst and Watkins 1981). Turning to low-level suspension feeders in deep-water Ordovician facies, the brachiopods were presumably little different ecologically from those of the Phyllopod Bed and include both inarticulates and articulates. High-level suspension feeders include pelmatozoan echinoderms (especially crinoids), locally dendroid graptolites and sponges, but the latter group may be under- represented by methodological oversight and taphonomic faetors. Taylor and Cook (1976; Cook and Taylor 1977) noted an abundance of sponge spicules in deep-water Upper Cambrian deposits, and other comparisons between these deep-water faunas and the Burgess Shale were drawn above. Examples of exceptional preservation, such as the celebrated deep-water sponges from the late Devonian of the Appalachians (Hall and Clarke 1 900; Clarke 1918; Caster 1939) probably reflect the CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 453 Suspension feeders Deposit feeders Carnivores (& scavengers) Palaeozoic (Ordovician-Devonian) Sponges (especially lass sponges) Crinoids * Ectoprocts Brachiopods > (esp articulates) Ostracodes Trilobites Petalonamids, pennatuloids Ostracodes “Non-tr Monoplacophorans lobitic" * Primitive arthropods & \ Trilobites Nautiloids Fish Eunicid polychaetes <1/1 Vi "Non-trilobitic" arthropods (& inceriae sedls) TEXT-FIG. 13. Simplified diagram showing possible descent of ecological analogues of suspension feeders, deposit feeders, and carnivores in the early Phanerozoic deeper-water communities; see text for sources of data and further discussion. Solid arrows indicate descent within the same group, although not necessarily with direct phylogenetic links; dotted arrows indicate an indirect relationship. Size of lettering indicates relative importance. spread of cold-water conditions (McGhee 1982), but similar sponge faunas may formerly have been widespread in deep-water communities with a low preservation potential. The ecological equivalents of the epifaunal predators of the Phyllopod Bed are less easily identifiable but may include fish, nautiloids, and perhaps some trilobites. No unequivocal analogue to the infaunal priapulid predators of the Phyllopod Bed appears to have been recognized but, if jawed polychaetes represent their ecological descendants (Conway Morris \911a), dispersed scolecodonts should occur in these later deep-water facies. Text-fig. 13 summarizes in a highly simplified fashion the proposed evolution of ecological analogues in deeper-water communities from the Cambrian to the Devonian. In some cases ecological descent is linked to phylogenetic descent, but more often the successors are not closely related to the earlier forms. The ecological replacements depicted here may have arisen by either opportunistic or competitive displacement. This diagram is tentative, but differs somewhat from earlier depictions of Palaeozoic marine community evolution (e.g. Bretsky 1969; Berry 1974; Hurst and Watkins 1981; Lockley 1983) in emphasizing the more synecological aspects, albeit at a simplified level, and by drawing on information from soft-bodied Lagerstatten (Burgess Shale, Beecher’s Trilobite Bed (Cisne 1973a, b), Hunsruckschiefer (Sturmer et al. 1980)). More effective comparisons of deep-water ecologies during the Palaeozoic will be enhanced by as yet unavailable community studies of appropriate Lagerstatten. Concluding remarks. If the assumption that the origins of the ‘Cambrian fauna’ ultimately lay in very shallow water (Axelrod 1958; Sepkoski and Sheehan 1983) has any validity, then extrapolation of Phyllopod Bed (Burgess Shale) ecological equivalents to older faunas presumably should concentrate on shallow-water assemblages of earliest Cambrian (Tommotian) or even Ediacaran age (text-fig. 13). Ecological comparisons are hindered by major environmental differences, e.g. turbulence, oxygen, substrates, and possibly salinity. Nevertheless, it may be significant that shallow- 454 PALAEONTOLOGY. VOLUME 29 water Tommotian faunas contain abundant halkieriids, known from their dispersed sclerites, descendants of which occur in the deeper-water Phyllopod Bed (Bengtson and Conway Morris 1984; Conway Morris 1985a). In addition, an indication of the ecologies of the pre-skeletal Ediacaran fauna, some of which were apparently deep-water, is given (text-fig. 1 3). These faunas were apparently dominated by suspension feeders, which in the case of cnidarians conceivably were microcarnivores on protoctistans, metazoan larvae, and other plankton. Deposit feeders and perhaps grazers were evidently represented by worms such as Dickinsonia (Wade 1972), the primitive ?arthropods, and probably some of the makers of the abundant trace fossils. Macrophagous predators and scavengers appear to have been absent (Glaessner 1984). Apparently only one reconstruction of a feeding web for an Ediacaran assemblage has been attempted (Goldring 1972), and this makes possibly unwarranted assumptions regarding the presence of predatory groups (as nemerteans). The available evidence points to a simple trophic structure in Ediacaran communities. The inference that microbial decomposers to some extent usurped the role of higher trophic levels, however, may be dilficult to reconcile with the widespread survival of soft parts. Recent communities Modern dysaerobic communities do not appear to have a Palaeozoic aspect (Savrda et al. 1984; Thompson et al. 1985), but it may be significant that, as reconstructed, the Phyllopod Bed fauna has a number of functional and ecological similarities with modern deep-sea faunas. Moreover, a number of authors (e.g. Hickman 1981; Thayer 1983) have commented on ecological (not usually phylo- genetic) similarities between Palaeozoic faunas and modern assemblages living in the deep sea as well as shallower waters around Antarctica. More specifically I wish to stress that detailed ecological investigations of certain modern refugia (for a recent discussion see Thayer 1983) may enhance our understanding of the make-up of ancient communities by way of analogy. For example, modern deep-water ahermatypic mounds may include substantial numbers of articulate brachiopods and stalked crinoids, an assemblage with a remarkably Palaeozoic aspect (Heckel 1 974) that if transferred to shallow (and readily accessible) waters probably would be hailed as an outstanding relic. Concluding observations The preceding attempt to trace ecological analogues in relatively deep-water faunas through geological time may be widened to a more general documentation of increasing ecological complexity during the early metazoan radiations. This topic has already received attention from studies of changing trace fossil diversities (e.g. Alpert 1977; Brasier 1979), tiering of sessile epifauna and infauna (Ausich and Bottjer 1982), and specific groups such as echinoderms (Sprinkle 1980). The contrasts between the ecology of Ediacaran and Cambrian (e.g. Phyllopod Bed) communities are striking and text-fig. 13 serves to stress the radical innovations that must have accompanied the appearance of numerous bodyplans. The appearance of carnivores which evidently occurred in the Lower Cambrian (see Conway Morris 1985a; Conway Morris and Jenkins 1985) is the most obvious trophic innovation, and has a direct bearing on the long standing debate on whether hard parts evolved as a protective adaptation. Niche differentiation in suspension feeders included tiering (Conway Morris 1979fi; Ausich and Bottjer 1982), and possibly different size/type selection of food particles amongst the various tentacular and sieve organs of established and newly evolving bodyplans. Amongst deposit feeders, the possibility of niche pre-emption in one ecological category (EVDC, epifaunal, vagrant, deposit (collector) feeder) of the Phyllopod Bed fauna may indicate that these resources were apportioned. It remains to be seen whether niche differentiation proceeded at different rates in different ecological categories (see also Bambach 1983). Styles of deposit feeding may have shown no radical change for hundreds of millions of years (Levinton and Bambach 1975), but amongst suspension feeders the development of tiering and increased vertical range is indicative of continuing specialization (Ausich and Bottjer 1982). In the case of predation, although the existence of CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 455 Cambrian predators is now firmly established, it seems plausible that the degree of sophistication in styles of predation (search and attack) and deterrence in comparison with younger Palaeozoic faunas (Signor and Brett 1984) was substantially less. The mode and tempo by which the ‘ecological barrel’ was filled will be understood better when Lower Cambrian communities, including those of soft-bodied Lagerstatten such as the Kinzers Formation (Pennsylvania), are better documented. Preparation of dominance-diversity curves through geological time (see McGhee 1981 for preliminary analysis of changes in oflf-shore and near- shore communities during a 5 Ma span (Frasnian, Devonian)) may provide further insights into ecological diversification of both entire communities and particular ecological categories. Analogies may exist with some patterns documented in short-term ecological successions (e.g. Whittaker 1972; Bazzaz 1975; see also Kempton and Taylor 1976). Conceivably the earliest stages in the filling of the Phanerozoic ‘ecological barrel’ may have entailed minimal competition in marine environments (Conway Morris and Fritz 1984), and niche structure may show a non-overlapping and non- contiguous nature. THE ROLE OE EOSSIL LAGERSTATTEN The frequency of fossil Lagerstatten The Phyllopod Bed fauna provides such a wealth of insights into the ecology and diversity of Middle Cambrian life that it is appropriate to inquire whether similar soft-bodied Lagerstatten occur with sufficient frequency during the Phanerozoic to provide ultimately a more comprehensive view of the palaeoecology and community structure of former life. As Table 5 demonstrates the Phyllopod Bed is indeed only one of many exceptional deposits, but this list is based on a number of important qualifications. Only the four most diverse Ediacaran faunas from the Late Precambrian are included, although there are numerous other localities (e.g. Canada (Northwest Territories and British Columbia), China, England, Wales, Iran, Morocco, Siberia, Sweden, Ukraine, USA (North and South Carolina)) where certain elements of the Ediacaran biota have been found. The wide- spread occurrence of the entirely soft-bodied Ediacaran fossils is linked to an apparent absence of scavengers and predators combined with a restricted degree of burrowing and other bioturbation; such conditions favour extensive soft-part preservation (Sepkoski 1979; Glaessner 1984). With the TABLE 5. List of principal Lagerstatten with soft-part preservation, excluding pre-Ediacaran examples (e.g. Gunflint Chert). Deposit Age Locality Environment of depositioi Permafrost, frozen Pleistocene Siberia, USSR Tundra mammoths and other mammals Calico Mountains nodules Miocene California, USA Fluvial, lacustrine (Barstow Eormation) Monterey Eormation Miocene California, USA Marine Oeningen Molasse Miocene Lake Constance. Switzerland/Germany Lacustrine, fluvial Amber Oligocene Baltic region, Dominican Republic Terrestrial Elorissant Beds Oligocene Colorado, USA Lacustrine Quercy Phosphorites Eocene-Oligocene Massif Central, France Terrestrial Geiseltal Lignite Eocene Halle, East Germany Terrestrial, swamp Messel Oil Shales Eocene Darmstadt, West Germany Lacustrine Monte Bolca Pish Beds Eocene Monte Bolca, Italy Marine Green River Eormation Eocene Wyoming, Utah, Colorado, USA Lacustrine Sahel Alma Fish Beds Cretaceous Lebanon Marine 456 PALAEONTOLOGY, VOLUME 29 TABLE 5 (cont.) Deposit Age Locality Environment of deposition Amber Cretaceous Lebanon, Canada Terrestrial Solnhofen Limestone Jurassic Bavaria, West Germany Marine Cerin Limestone Jurassic Jura, France Marine Posidonia Shale (Holzmaden) Jurassic Wiirttemberg, West Germany Marine Selcifero Lombardo Jurassic Osteno, Italy Marine Voltzia Sandstone Triassic Northern Vosges, France Marine, deltaic Mecca and Logan Black Carboniferous Indiana, USA Marine Shales Mazon Creek nodules Carboniferous Illinois, USA Terrestrial/marine Montceau-les-Mines Carboniferous Autun, France Terrestrial/marine concretions and shales Granton Sandstones Carboniferous Granton, Scotland Marine Bear Gulch Limestone Carboniferous Montana, USA Marine Bearsden (Top Hosie Carboniferous Glasgow, Scotland Marine Limestone) Gilboa mudstones Devonian New York, USA Terrestrial Hunsriick Slate Devonian Rhineland, West Germany Marine Rhynie Chert Devonian Aberdeenshire, Scotland Terrestrial Cleveland Shale Devonian Ohio, USA Marine Lesmahagow Inlier Silurian Lanarkshire, Scotland Marine Eurypterid Dolomite Silurian Saaremma, Estonia Marine (Rootsikula Stage) Waukesha Dolomite Silurian Wisconsin, USA Marine (Brandon Bridge) Beecher’s Trilobite Bed Ordovician New York, USA Marine (Frankfort Shale) Anthraconite (Orsten) Cambrian Sweden Marine Burgess Shale (Stephen Cambrian British Columbia Marine Formation) Spence Shale Cambrian Utah, USA Marine Wheeler Formation Cambrian Utah, USA Marine Emu Bay Shale Cambrian Kangaroo Island, Australia Marine Kinzers Formation Cambrian Pennsylvania, USA Marine Pound Quartzite Vendian Flinders Ranges, Australia Marine Valdai Series Vendian Onega Peninsula, USSR Marine Nama Group Vendian Namibia Marine Conception Group Vendian (?late Riphean) Avalon Peninsula, Newfoundland Marine post-Vendian rise of scavengers and predators (see above) and greater churning of the sediment, conditions for preservation of soft-bodied and lightly skeletized species with a generally minimal fossilization potential became more localized (Sepkoski 1979). Amongst the post-Ediacaran faunas the continuum in preservational quality that exists in sedimentary rocks means that there are no clear- cut criteria to determine whether a given fossil fauna or flora should be regarded as an exceptional Lagerstiitte. For instance, there are many additional horizons, e.g. various ‘insect beds’ or ‘fish beds’, which could qualify for inclusion in Table 5. Even the Lagerstatten in Table 5 do not possess equal status. It is generally agreed that amongst the most significant are the Solnhofen Limestone, Mazon Creek (Francis Creek Shale) ironstone nodules, Hunsriickschiefer, and Burgess Shale. To what CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 457 extent their pre-eminence is due to greater diversity and related perfection of preservation, extent of collecting, ease of geographical access, and attendant scientific publicity is uncertain. The principal purpose of Table 5, however, is not to present an exhaustive compilation, but to emphasize the abundance of these Lagerstatten through geological time. Are Lagerstatten representative? Incorporating data from Lagerstatten into the mainstream of palaeoecological research, that otherwise relies almost wholly on hard-part remains, may have advantages but is not without its pitfalls. A more systematic approach to the study of Lagerstatten in the context of the continuum of surrounding biotas (Schram 1979; Conway Morris 1981) is overdue. Nevertheless, it would be naive to imagine that these Lagerstatten can be used as a simple panacea in palaeoecological analysis. Soft- part preservation frequently entails anoxic or hypersaline conditions, thereby excluding scavengers and perhaps diminishing rates of microbial decay. The necessary proximity of a benthic biota, perhaps adapted to unusual environmental conditions, or even restriction of the biota to a pelagic element occupying the aerated top of a water column, means that the preserved assemblage cannot be representative in comparison with those inhabiting ‘normal’ environments. Moreover, a biota occupying areas where catastrophic burial is favoured need not necessarily be representative (see also Brett and Eckert 1982). Perusal of the literature on Lagerstatten makes the above discussion abundantly clear (see Whittington and Conway Morris 1985). There are repeated references to an absence or rarity of otherwise abundant groups, e.g. stenohaline brachiopods, trilobites, echino- derms, a limited benthic component with emphasis on pelagic elements (e.g. fish), or preservational bias towards organisms of small size (e.g. amber, Orsten), or limited escape abilities following catastrophic burial. On the other hand some Lagerstatten do appear to represent a more or less normal assemblage and, as was argued above, the likelihood that the Phyllopod Bed community falls into this category suggests that inferences based on its ecology may have a wider application. Future research A neglected role of Lagerstatten is their possible use in assessing diversity changes through the Phanerozoic (Cisne 1974). Concerning this, Raup (1972) considered that Lagerstatten would have a largely disruptive effect on calculations by adding ‘noise’ to the available data, and specific instances of this were documented by Hoffman and Ghiold (1985). Sepkoski (1981fi) indicated that when considering the total of marine metazoan families through the Phanerozoic, information derived from Lagerstatten is of little significance. Nevertheless, about one fifth of the marine metazoan clades are known only from three major Palaeozoic Lagerstatten (Burgess Shale, Hunsriickschiefer, Mazon Creek) (Sepkoski 198 Ifi). The final area that deserves comment is the ignorance surrounding the diagenetic conditions leading to exceptional preservation. While the overall mode of preservation is evident (e.g. pyrite, siderite, carbonaceous residues, phosphates, silica or silicates), the ultrastructure of replacement is poorly known. Thus, Cisne (1981) commented on the erratic extent of pyritization in trilobites with appendages from Beecher’s Trilobite Bed (Frankfort Shale), and the controlling factors are not well understood. The underlying physico-chemical conditions that led to special preservation are obscure, but it is likely that in some cases bacteria (e.g. Wuttke 1983u) may have had a significant intermediate role. The unusual conditions make laboratory replication difficult and pioneer studies by workers such as Hecht (1933), Zangerl and Richardson (1963), Oehler and Schopf (1971), Zangerl ( 1971 ), Leo and Barghoorn (1976), Francis et at. (1978), and Wuttke (19836) need further development. Work in progress on the ultrastructure of soft-part preservation may provide new insights. CONCLUSIONS On the basis of a shelly assemblage little different from those that characterize many Cambrian faunas, the soft-bodied biota of the Phyllopod Bed (Burgess Shale) is regarded as reasonably representative, at least with regard to moderately deep-water communities living in the Cordilleran region. The shelly assemblage in the Phyllopod Bed accounts for no more than 20% of genera, and 458 PALAEONTOLOGY, VOLUME 29 perhaps as little as 2% (excluding sponges) in terms of individuals. Accordingly, the soft-bodied component is important in reconstructing the ecology of this community and, by implication, certain other Cambrian faunas. The analysis of the fauna is undertaken in terms of taxa (number of genera, mostly monospecific), individuals, and biovolumes. Ecological categories are defined according to position relative to substrate and degree of mobility, and feeding type. Of particular importance in the benthic assemblage are epifaunal vagrant deposit feeders (mostly arthropods), infaunal vagrant carnivores/scavengers (priapulid worms), epifaunal sessile suspension feeders (almost entirely sponges and brachiopods), and infaunal sessile suspension feeders (probable hemichordates). The Phyllopod Bed fauna shows unequivocally that the fundamental trophic structure of marine metazoan life was established early in its evolution (see Hutchinson 1959, 1961; Bretsky 1969). Highlights of the trophic analysis include: a, the identification of a trophic nucleus (text-fig. 10) dominated by relatively few taxa, although the composition varies according to whether the nucleus is considered in terms of either individuals or biovolumes; 6, the recognition of the importance of carnivores and scavengers; and c, the reconstruction of a trophic web (text-fig. 1 1 ) and crude estimates of energetic efficiency between the levels of primary and secondary consumers. Possible niche structure of various ecological categories is addressed via dominance-diversity curves (text- fig. 12). Most show a log-normal distribution which may have no simple ecological explanation. However, examples of a geometric distribution, such as in epifaunal vagrant deposit (collector) feeders (EVDC), could be consistent with the hypothesis of niche pre-emption. An attempt is made to trace ecological analogues of the Phyllopod Bed community in deeper-water assemblages through geological time (Cambrian-Devonian) with particular reference to the trophic categories of carnivores and scavengers, deposit feeders, and suspension feeders. Taxonomic membership of such a category may show continuity for a protracted period; for example, the apparent persistence of trilobites and ostracodes as deposit feeders. In other cases, however, innovation in the form of new groups either insinuating or perhaps replacing more archaic forms appears to have occurred. The appearance of carnivores and scavengers, such as nautiloids and eunicid polychaetes, may be an example. How far the evolution of such ecological units can be usefully traced through time is uncertain. If claims that the oxygen levels of Palaeozoic oceans were lower than today have any validity, then examination of dysaerobic faunas through geological time might provide insights into the evolution of ecological units. It must be admitted, however, that modern dysaerobic faunas (e.g. Savrda et al. 1984; Thompson et al. 1985) do not appear to have any striking ecological similarities with Palaeozoic assemblages. The paper concludes by calling attention to the frequency of soft-bodied Lagerstatten (Table 5) and emphasizes their role as potential sources of palaeoecological information. The extent to which such exceptional biotas can be regarded as representative in comparison with coeval shelly assemblages requires examination, as do the physico-chemical conditions responsible for promoting such remarkable preservation in the geological record. Acknowledgements. I thank H. B. Whittington, C. P. Hughes (Cambridge), D. E. G. Briggs (Bristol), and D. L. Bruton (Oslo) for useful discussions and generously providing information, and the many individuals who supplied specimen lists or granted access to Burgess Shale collections, especially F. J. Collier (Washington, DC). S. Bengtson (Uppsala), H. B. Whittington, and two anonymous referees reviewed the manuscript and made many helpful suggestions. J. J. Sepkoski (Chicago) kindly computed the best-fits for the dominance -diversity curves, while G. Wood (Houston) and K. M. Towe (Washington, DC) made available unpublished informa- tion on acritarchs and preservation, respectively. Radiographic photographs were prepared and printed by D. Newling, other figures were drafted by J. Taylor, H. Boxall, A. Prouse, and S. Ripper. Drafts of this manu- script were typed by Monica Siewertz, Beryl West, Mary Askham, and Sandra Last. This is Cambridge Earth Sciences Publication 636. REFERENCES ACER, D. V. 1976. The nature of the fossil record. Proc. Geol. Ass. 87, 131-159. AiTKEN, J. D. 1971. Control of Lower Paleozoic sedimentary facies by the Kicking Horse Rim, southern Rocky Mountains, Canada. Bull. Can. Petrol. Geol. 19, 557-569. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 459 AiTKEN, J. D. and MciLREATH, I. A. 1984. The Cathedral Reef Escarpment, a Cambrian great wall with humble origins. Geos, Ottawa, 13, 17-19. ALPERT, s. p. 1977. Trace fossils and the basal Cambrian boundary. In crimes, t. p. and harper, j. c. (eds.). Trace Fossils 2. Geol. J. Special Issue, 9, 1-18. AUSICH, w. I. and bottjer, d. j. 1982. Tiering in suspension-feeding communities on soft substrata throughout the Phanerozoic. Science, Wash. 216, 173-174. AXELROD, D. I. 1958. Early Cambrian marine fauna. Science, N. Y. 128, 7-9. BAMBACH, R. K. 1977. Species richness in marine benthic habitats through the Phanerozoic. Paleobiol. 3, 1 52- 1 67. 1983. Ecospace utilization and guilds in marine communities through the Phanerozoic. In tevesz, m. j. s. and MCCALL, p. l. (eds.). Biotic interactions in Recent and fossil benthic communities, 719-746. Plenum Press, New York. BARHAM, E. G., AYER, N. J. and BOYCE, R. E. 1967. Macrobcnthos of the San Diego Trough: photographic census and observations from bathyscaphe, Trieste. Deep-Sea Res. 14, 111-1%A. BARTHEL, K. w. 1970. On the deposition of the Soinhofen lithographic limestone (Lower Tithonian, Bavaria, Germany). Neues Jb. Geol. Paldont. Abb. 135, 1-18. BAZZAZ, F. A. 1975. Plant species diversity in old-held successional ecosystems in southern Illinois. Ecology, Brooklyn, 56, 485-488. BEDNARCZYK, w. 1984. Biostratigraphy of the Cambrian deposits in the Leba area. Acta. geol. pol. 34, 95-1 10. BENGTSON, s. 1977. Aspects of problematic fossils in the early Palaeozoic. Acta Univ. Upsaliensis, 415, 1-71. and CONWAY morris, s. 1984. A comparative study of Lower Cambrian Halkieria and Middle Cambrian Wiwaxia. Lethaia, 17, 307-329. BERRY, w. B. N. 1972. Early Ordovician bathyurid province lithofacies, biofacies, and correlations— their relationships to a proto-Atlantic Ocean. Ibid. 5, 69-83. 1974. Types of early Paleozoic faunal replacements in North America: their relationship to environmental change. J. Geol. 82, 371-382. BIRENHEIDE, R. 1971. Beobachtungen am ‘Scheinstern’ Mimetaster aus dem Hunsriick-Schiefer. Senckenberg. leth. 52,11-91. BOUCOT, A. J. 1981. Principles of benthic marine paleoecology, 463 pp. Academic Press, New York. BRASiER, M. D. 1979. The Cambrian radiation event. In house, m. r. (ed.). The origin of major invertebrate groups. Syst. ^55. Spec. Vol. 12, 103-159. BRASSEL, G. and BERGSTROM, J. 1978. Der Trilobit Parahomalonotus (Koch 1883), ein Homalonotidae aus dem unterdevonischen Hunsriickschiefer. Geol. Jb. Hessen, 106, 5-1 1. BRETSKY, p. w. 1969. Evolution of Paleozoic benthic marine invertebrate communities. Palaeogeogr. Palaeoclimat. Palaeoecol. 6, 45-59. and KLOFAK, s. M. 1985. Margin to craton expansion of Late Ordovician benthic marine invertebrates. Science, Wash. 221, 1469-1471. BRETT, c. E. and ECKERT, J. D. 1 982. Palaeoecology of a well-preserved crinoid colony from the Silurian Rochester Shale in Ontario. Contr. Life Sci. Div. R. Ont. Mus. 131, 1-20. BRIGGS, D. E. G. 1977. Bivalved arthropods from the Cambrian Burgess Shale of British Columbia. Palaeontology, 20, 595-621. 1978. The morphology, mode of life, and affinities of Canadaspis perfecta (Crustacea: Phyllocarida), Middle Cambrian, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B281, 439-487. -1981. The arthropod Odaraia alata Walcott, Middle Cambrian, Burgess Shale, British Columbia. Ibid. B291, 541-584. and WHITTINGTON, H. B. 1985. Modes of life of arthropods from the Burgess Shale, British Columbia. Trans. R. Soc. Edinb. 76, 149-160. BRUTON, D. L. 1981. The arthropod Sidnevia inexpectans. Middle Cambrian, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B295, 619-656. BUBELA, B. and CLOUD, p. 1983. Sulphide mineralisation of microbial cells. BMR J. Aust. Geol. Geophvs. 8, 355-357. BYERS, c. w. 1979. Biogenic structures of black shale paleoenvironments. Postilla, 174, 1-43. CADfe, G. c. 1984. 'Opportunistic feeding’, a serious pitfall in trophic structure analysis of (paleo)faunas. Lethaia, 17, 289-292. CALEF, c. E. and bambach, r. k. 1973. Low nutrient levels in Lower Paleozoic (Cambrian-Silurian) oceans. Abstr. Progm geol. Soc. Am. 5, 565. CASTER, K. E. 1939. Siliceous sponges from Mississippian and Devonian strata of the Penn-York embayment. J. Paleont. 13, 1-20. 460 PALAEONTOLOGY, VOLUME 29 CISNE, J. L. 1973a. Life history of an Ordovician trilobite Triarthrus ealoni. Ecology, Brooklyn, 54, 135-142. 1973/). Beecher’s Trilobite Bed revisited: ecology of an Ordovician deepwater fauna. Postilla, 160, 1-25. 1974. Evolution of the world fauna of aquatic free-living arthropods. Evolution, 28, 337-366. - — 1981. Triarthrus eatoni (Trilobita): anatomy of its exoskeletal, skeletal, skeletomuscular, and digestive systems. Palaeontogr. am. 9(53), 1 142. CLARK, R. B. 1964. Dynamics in metazoan evolution. The origin of the coelom and segments, 313 pp. Clarendon Press, Oxford. CLARKE, J. M. 1918. Devonian glass sponges. Bid!. N.Y. St. Mus. 196, 177-186. CLOUD, p. E. 1960. Gas as a sedimentary and diagenetic agent. Am. J. Sci. 258A, 35-48. COLLINS, D. 1978. A palaeontologist’s paradise. The 1975 ROM expedition to the 540,000,000-year-old Burgess Shale. Rotunda, 11, 12-19. BRIGGS, D. E. G. and CONWAY MORRIS, s. 1983. New Burgess Shale fossil sites reveal Middle Cambrian faunal complex. Science, Wash. 222, 163-167. and RUDKIN, D. M. 1981. Priscansermarinus barnetti, a probable lepadomorph barnacle from the Middle Cambrian Burgess Shale of British Columbia. J. Paleont. 55, 1006-1015. CONWAY MORRIS, s. \911a. Fossil priapulid worms. Spec. Pap. Palaeont. 20, i-iv, 1-95. — 19776. A new metazoan from the Cambrian Burgess Shale of British Columbia. Palaeontology, 20, 623-640. 1979a. The Burgess Shale. In fairbridge, r. w. and jablonski, d. (eds.). Encyclopedia of paleontology, 153-160. Dowden, Hutchinson and Ross, Stroudsburg, Pa. 19796. The Burgess Shale (Middle Cambrian) fauna. A. Rev. Ecol. Syst. 10, 327-349. 1979c. Middle Cambrian polychaetes from the Burgess Shale of British Columbia. Phil. Trans. R. Soc. B285, 221 -21 A. 1981. The Burgess Shale fauna as a mid-Cambrian community. U.S. geol. Surv. Open File Rep. 81-743, 47-49. (ed.). 1982. Atlas of the Burgess Shale, 31 pp. Palaeontological Association, London. 1985a. The Middle Cambrian metazoan Wiwaxia corrugata (Matthew) from the Burgess Shale and Ogygopsis Shale, British Columbia, Canada. Phil. Trans. R. Soc. B307, 507-586. 19856. Cambrian Lagerstatten: their distribution and significance. Ibid. B311, 49-65. and FRITZ, w. h. 1984. Lapworthella fdigrana n.sp. (incertae sedis) from the Lower Cambrian of the Cassiar Mountains, northern British Columbia, Canada, with comments on possible levels of competition in the early Cambrian. Palaeont. Z. 58, 197-209. and JENKINS, R. J. F. 1985. Healed injuries in early Cambrian trilobites from South Australia. Alcheringa, 9, 167-177. and ROBISON, r. a. 1982. The enigmatic medusoid Peytoia and a comparison of some Cambrian biotas. J. Paleont. 56, 1 16-122. 1986. Middle Cambrian priapulids and other soft-bodied fossils from Utah and Spain. Paleont. Contr. Univ. Kans. 117, 1-22. and RUSHTON, a. w. a. (in press). Precambrian to Tremadoc biotas in the Caledonides. Spec. Pubis geol. Soc. Loud. and WHITTINGTON, H. B. 1979. The animals of the Burgess Shale. Sclent. Am. 241, 122-133. 1985. Fossils of the Burgess Shale, a national treasure in Yoho National Park, British Columbia. Misc. Rep. geol. Surv. Canada, 43, 1-31. i COOK, H. E. and taylor, m. e. 1977. Comparison of continental slope and shelf environments in the Upper j Cambrian and lowest Ordovician of Nevada. Spec. Pubis Soc. econ. Paleont. Miner. Tulsa, 25, 51-81. DEAN, D., RANKIN, J. s. and HOFFMANN, E. 1964. A note on the survival of polychaetes and amphipods in stored jars j of sediments. J. Paleont. 38, 608-609. DEEVEY, E. s. 1969. Specific diversity in fossil assemblages. Brook haven Symp. Biol. 22, 224-241. DENGLER, A. T., WILDE, p., NODA, E. K. and NORMARK, w. R. 1984. Turbidity currents generated by Hurricane Iwa. Geo-Marine Letters, 4, 5-11. DRISCOLL, E. G. and swanson, r. a. 1973. Diversity and structure of epifaunal communities on mollusc valves. Buzzards Bay, Massachusetts. Palaeogeogr. Palaeoclimat. Palaeoecol. 14, 229-247. DUFF, K. L. 1975. Palaeoecology of a bituminous shale— the Lower Oxford Clay of central England. Palaeontology, 18, 443-482. DURHAM, J. w. 1974. Systematic position of Eldonia ludwigi Walcott. J. Paleont. 48, 750-755. DZiK, J. 1979. Some terebratulid populations from the Lower Kimmeridgian of Poland and their relations to the biotic environment. Acta, palaeont. pol. 24, 473-492. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 461 FRANCIS, s., MARGULis, L. and BARGHOORN, E. s. 1978. On the experimental silicification of microorganisms. II. On the time of appearance of eukaryotic organisms in the fossil record. Precamh. Res. 6, 65-100. FURSiCH, F. T. 1977. Corallian (Upper Jurassic) marine benthic associations from England and Normandy. Palaeontology, 20, 337-385. and WENDT, J. 1977. Biostratinomy and paleoecology of the Cassian Formation (Triassic) of the southern Alps. Palaeogeogr. Palaeocliinal. Palaeoecol. 22, 257-323. GLAESSNER, M. F. 1972. Precambrian palaeozoology. In jones, j. b. and mcgowran, b. (eds.). Stratigraphic problems of the later Precambrian and early Cambrian. Spec. Pap. Cent. Precamh. Res. Univ. Adelaide, 1, 43-52. 1984. The dawn of animal life. A hiohistorical study, xi + 244. Cambridge University Press, Cambridge. GOLDRING, R. 1972. Evolution in environments. In gass, i. g., smith, p. j. and wilson, r. c. l. (eds.). Understanding the Earth, 156-161. Artemis Press, Horsham. - and STEPHENSON, D. G. 1972. The depositional environment of three Starfish Beds. Neues Jh. Geol. Paldont. Mh. HIO, 611-624. GOULDEN, c. E. 1966. La Aguada de Santa Ana Vieja: an interpretative study of the cladoceran microfossils. Arch. Hydrohiol. 62, 373-404. 1969. Temporal changes in diversity. Brookhaven Symp. Biol. 22, 96-102. HALL, J. and CLARKE, J. M. 1900. Paleozoic reticulate sponges constituting the family Dictospongidae. Mem. N. Y. St. Mus. nat. Hist. 20, 1-197. HARDY, p. G. and broadhurst, f. m. 1978. Refugee communities of Carhonicola. Lethaia, 11, 175-178. HECHT, F. 1933. Der Verbleib der organischen Substanz der Tiere bei meersicher Einbettung. Senckenhergiana, 15, 165-249. HECKEL, p. H. 1974. Carbonate buildups in the geologic record: a review. Spec. Pubis Soc. econ. Paleont. Miner. Tulsa, 18, 90-154. hecker, b. 1982. Possible benthic fauna and slope instability relationships. In saxon, s. and nieuwenhuis, j. k. (eds.). Marine slides and other mass movements, 335-347. Plenum Press, New York. heezen, b. c. and hollister, c. d. 1971. The face of the deep, 659 pp. Oxford University Press, Oxford. herring, p. j. and clark, m. r. 1971. Deep oceans, 320 pp. Arthur Barker, London. HERTWECK, G. 1972. Georgia coastal region, Sapelo Island, U.S.A. Sedimentology and biology. V. Distribution and environmental significance of lebensspuren and in situ skeletal remains. Senckenherg. marit. 4, 125-167. HICKMAN, c. s. 1981. Giants, dwarfs and bizarre constructional patterns in the marine fossil record. Abstr. Progm geol. Soc. Am. 13, 473. HOFFMAN, A. and GHiOLD, J. 1985. Randomness in the pattern of ‘mass extinctions' and ‘waves of origination’. Geol. Mag. 122, 1-4. HUGHES, c. p. 1975. Redescription of Burgessia bella from the Middle Cambrian Burgess Shale, British Columbia. Fossils Strata, 4, 415-435. HURST, J. M. and watkins, r. 1981. Lower Paleozoic clastic, level-bottom community organization and evolution based on Caradoc and Ludlow comparisons. In gray, j., boucot, a. j. and berry, w. b. n. (eds.). Communities of the past, 69-100. Hutchinson Ross, Stroudsburg, Pa. HUTCHINSON, G. E. 1959. Homage to Santa Rosalia or why are there so many kinds of animals. Am. Nat. 93, 145-159. 1961. The biologist poses some problems. In sears, m. (ed.). Oceanography. Pubis Am. /Iss'. Advmt Sci. 67, 85-94. JABLONSKi, D., SEPKOSKi, J. J., BOTTJER, D. J. and SHEEHAN, p. M. 1983. Onshore-ofifshore patterns in the evolution of Phanefozoic shelf communities. Science, Wash. 222, 1 123-1 125. JAMES, N. p. 1981 . Megablocks of calcified algae in the Cow Head Breccia, western Newfoundland: vestiges of a Cambro-Ordovician platform. Bulk geol. Soc. Am. 92, 799-81 1. JELL, p. A. 1974. Faunal provinces and possible planetary reconstruction of the Middle Cambrian. J. Geol. 82, 319-350. 1975. Australian Middle Cambrian eodiscoids with a review of the superfamily. Palaeontographica, A150, 1-97. JOHNSON, R. G. 1964. The community approach to paleoecology. In imbrie, j. and newell, n. (eds.). Approaches to paleoecology, 107-134. J. Wiley, New York. 1974. Particulate matter at the sediment-water interface in coastal environments. J. mar. Res. 32, 313-330. JORGENSEN, c. B. 1966. The biology of suspension feeding, 357 pp. Pergamon Press, Oxford. KANASEWICH, E. R., HAVSKOV, J. and EVANS, M, E. 1978. Plate tectonics in the Phanerozoic. Can. J. Earth Sci. 15,919-955. 462 PALAEONTOLOGY, VOLUME 29 KAPP, u. s. 1975. Paleoecology of Middle Ordovician stromatoporoid mounds in Vermont. Lethaia, 8, 195-207. KAUFFMAN, E. G. and FURSICH, F. 1983. Brooksellci canyonensis: a billion year old complex metazoan trace fossil from the Grand Canyon. Abstr. Progm geol. Soc. Am. 15, 608. and STEiDTMANN, j. R. 1981. Are these the oldest metazoan trace fossils? J. Paleont. 55, 923-947. KELLER, G. H. 1982. Organic matter and the geotechnical properties of submarine sediments. Geo-Marine Letters, 2, 191-198. KEMPTON, R. A. and TAYLOR, L. R. 1976. Modcls and statistics for species diversity. Nature, Land. 262, 818-820. KEPPER, J. c. 1981. Sedimentology of a Middle Cambrian outer shelf margin with evidence for syndepositional faulting, eastern California and western Nevada. J. sedim. Petrol. 51, 807-821. KERN, J. p. and wiCANDER, E. R. 1974. Origin of a bathymetrically displaced marine invertebrate fauna in the upper part of the Capistrano Formation (Lower Pliocene), southern California. J. Paleont. 48, 495-505. KRANZ, p. M. 1974. The anastrophic burial of bivalves and its paleoecological significance. J. Geol. 82, 237-265. 1977. A model for estimating standing crop in ancient communities. Paleohiol. 3, 415-421. LASKER, H. 1976. Effects of differential preservation on the measurement of taxonomic diversity. Ibid. 2, 84-93. LAWRENCE, D. R. 1968. Taphonomy and information losses in fossil communities. Bull. geol. Soc. Am. 79, 1315-1330. LEGGETT, J. K., MCKERROW, w. s., COCKS, L. R. M. and RICKARDS, R. B. 1981. Periodicity in the early Palaeozoic marine realm. J. geol. Soc. 138, 167 176. LEMCHE, H., HANSEN, B., MADSEN, F. J., TENDAL, o. s. and WOLFF, T. 1 976. Hadal life as analyzed from photographs. Vidensk. Meddr. dansk naturli. foren. 139, 263-336. LEO, R. F. and BARGHOORN, E. s. 1976. Silicification of wood. Bot. Mus. Leaji. Harv. Univ. 25, 1-47. LEViNTON, J. s. and BAMBACH, R. K. 1975. A Comparative study of Silurian and Recent deposit-feeding bivalve communities. Paleohiol. 1,97-124. LIDDELL, w. D. 1975. Recent crinoid biostratinomy. Abstr. Progm geol. Soc. Am. 7, 1 169. LiNAN, E. 1978. Riojaia perezi, un nuevo trilobitoidea del Cambrico medio de la Sierra de la Demanda (Espafia). Estudios geol. inst. Invest, geol. Lucas Mallada, 34, 319-321. LiTTLER, M. M., LiTTLER, D. s., BLAIR, s. M. and NORRIS, J. N. 1985. Deepest known plant life discovered on an uncharted seamount. Science, Wash. Ill,, 57-59. LOCKLEY, M. G. 1983. A review of brachiopod dominated palaeocommunities from the type Ordovician. Palaeontology, 26, 111-145. LUDViGSEN, R. and WESTROP, s. R. 1983. Trilobite biofacies of the Cambrian-Ordovician boundary interval in northern North America. Alcheringa, 7, 301-319. MCBRIDE, D. J. 1976. Outer shelf communities and trophic groups in the Upper Cambrian of the Great Basin. Geology Stud. Brigham Young Univ. 23, 139-152. MACDONALD, K. B. 1976. PaleocoiTimunities; towards some confidence limits. In scott, r. w. and west, r. r. (eds.). Structure and classification of paleocommunities, 87-106. Dowden, Hutchinson and Ross, Strouds- burg, Pa. MCGHEE, G. R. 1981. Evolutionary replacement of ecological equivalents in Late Devonian benthic marine communities. Palaeogeogr. PalaeocUmat. Palaeoecol. 34, 267-284. 1982. The Frasnian-Famennian extinction event: a preliminary analysis of Appalachian marine eco- systems. Spec. Pap. geol. Soc. Am. 190, 491-500. MciLREATH, I. A. 1974. Stratigraphic relationships at the western edge of the Middle Cambrian carbonate facies belt. Field, British Columbia. Geol. Surv. Pap. Can. 74-lA, 333-334. 1975. Stratigraphic relationships at the western edge of the Middle Cambrian carbonate facies belt. Field, British Columbia. Ibid. 75-lA, 557-558. 1977. Accumulation of a Middle Cambrian, deep-water limestone debris apron adjacent to a vertical, submarine carbonate escarpment. South Rocky Mountains, Canada. Spec. Pubis Soc. econ. Paleont. Miner. Tulsa, 25, 113-124. MARTENS, C. s. 1976. Control of methane sediment-water bubble transport by macro-infaunal irrigation in Cape Lookout Bight, North Carolina. Science, Wash. 192, 998-1000. MARTY, p. 1981. Distribution of different anaerobic bacteria in Arabian Sea sediments. Mar. Biol. 63, 277-281. MATTHEW, G. F. 1891. Illustrations of the fauna of the St John Group, no. v. Proc. Trans. R. Soc. Can. sect. 4, 8, 123-166. MAY, R. M. 1975. Patterns of species abundance and diversity. In cody, m. l. and diamond, j. m. (eds.). Ecology and evolution of communities, 81-120. Belknap Press, Cambridge, Mass. 1981. Patterns in multi-species communities. In may, r. m. (ed.). Theoretical ecology, principles and applications (2nd edition), 197-227. Blackwell Scientific, Oxford. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 463 MENZIES, R. J., GEORGE, R. Y. and ROWE, G. T. 1973. Abvssal environment and ecology of world oceans, 488 pp. Wiley-Interscience, New York. and WILSON, j. b. 1961. Preliminary field experiments on the relative importance of pressure and temperature on the penetration of marine invertebrates into the deep sea. Oikos, 12, 302-309. MIDDLETON, G. V. 1967. The orientation of concavo-convex particles deposited from experimental turbidity currents. J. sedim. Petrol. 31, 229-237. MILLER, M. F. and BYERS, c. w. 1 984. Abundant and diverse early Paleozoic infauna indicated by the stratigraphic record. Geology, Boulder, Colo. 12, 40-43. MULLER, K. J. 1979. Phosphatocopine ostracodes with preserved appendages from the Upper Cambrian of Sweden. Lethaia, 12, 1-27. 1981. Arthropods with phosphatized soft parts from the Upper Cambrian ‘Orsten’ of Sweden. U.S. geol. Surv. Open-File Rep. 81-743, 147-151. 1982. Flesslandona imisidcata sp. nov. (Ostracoda) with phosphatised appendages from Upper Cambrian ‘Orsten’ of Sweden. In bate, r. h., robinson, e. and sheppard, l. (eds.). A research manual of fossil and recent ostracods, 276-304. Ellis Horwood, Chichester. 1983. Crustacea with preserved soft-parts from the Upper Cambrian of Sweden. Lethaia, 16, 93-109. 1985. Exceptional preservation in calcareous nodules. Phil. Trans. R. Soc. B311, 67-73. neyman, a. a. 1967. Eimits to the application of the ‘trophic group' concept in benthic studies. Oceanol. Acad. Sci. USSR, 7, 149-155. NICHOLS, J. A., ROWE, G. T., CLIFFORD, c. H. and YOUNG, R. A. 1978. In situ experiments on the burial of marine invertebrates. J. sedim. Petrol. 48, 419-425. NICOL, D. 1979. A survey of suspension-feeding animals. Fla Sclent. 42, 177-182. 1982. Some characteristics of extinct major groups of animals. Ibid. 45, 76-80. OEHLER, J. H. and SCHOPF, J. w. 1971. Artificial microfossils: experimental studies of permineralization of blue- green algae in silica. Science, N. Y. 174, 1229-1231 . oppenheimer, c. h. 1960. Bacterial activity in sediments of shallow marine bays. Geochim. cosmochim. Acta, 19, 244-260. palmer, a. r. 1972. Problems of Cambrian biogeography. Int. geol. Congr. 24, sect. 7, 310-315. 1974. Search for the Cambrian world. Am. Sclent. 62, 216-224. PARRISH, J. T. 1982. Upwelling and petroleum source beds, with reference to Paleozoic. Bull. Am. Ass. Petrol. Geol. 66, 750-774. PAUL, c. R. c. and smith, a. b. 1984. The early radiation and phylogeny of echinoderms. Biol. Rev. 59, 443-481 . PEMBERTON, s. G. and KOBLUK, D. R. 1978. Oldest known brachiopod burrow: the Tower Cambrian of Eabrador. Can. J. Earth Sci. 15, 1385-1389. PIPER, D. J. w. 1972. Sediments of the Middle Cambrian Burgess Shale, Canada. Lethaia, 5, 169-175. POWELL, E. N. and stanton, r. j. 1985. Estimating biomass and energy flow of molluscs in palaeo-communities. Palaeontology, 28, 1-34. PRESTON, F. w. 1962. The canonical distribution of commonness and rarity. Ecology, Brooklyn, 43, 185-215, 410-432. RASETTi, F. 1966. Revision of the North American species of the Cambrian trilobite genus Pagetia. J. Paleont. 40, 502-511. RAUP, D. M. 1972. Taxonomic diversity during the Phanerozoic. Science, N.Y. 177, 1065-1071. READ, J. F. and PFEiL, R. w. 1983. Fabrics of allochthonous reefal blocks. Shady Dolomite (Lower to Middle Cambrian), Virginia Appalachians. J. sedim. Petrol. 53, 761-778. REINHARDT, J. 1977. Cambrian off-shelf sedimentation, central Appalachians. Spec. Pubis Soc. econ. Paleont. Miner. Tulsa, 25, 83-1 12. RESSER, c. E. 1929. New Lower and Middle Cambrian Crustacea. Proc. U.S. natn. Mus. 76, 1-18. RHOADS, D. c. and MORSE, J. w. 1971. Evolutionary and ecologic significance of oxygen-deficient marine basins. Lethaia, 4, 413-428. SPEDEN, I. G. and WAAGE, K. M. 1972. Trophic group analysis of Upper Cretaceous (Maestrichtian) bivalve assemblages from South Dakota. Bull. Am. Ass. Petrol. Geol. 56, 1 100-1 1 13. RiTTENBERG, s. c. 1 940. Bacteriological analysis of some long cores of marine sediments. J. mar. Res. 3, 191-201. ROBISON, R. A. 1972. Mode of life of agnostoid trilobites. Int. geol. Congr. 24, sect. 7, 33-40. 1975. Species diversity among agnostoid trilobites. Fossils Strata, 4, 219-226. 1982. Some Middle Cambrian agnostoid trilobites from western North America. J. Paleont. 56, 132-160. 1984. New occurrences of the unusual trilobite Naraoia from the Cambrian of Idaho and Utah. Paleont. Contr. Univ. Kans. 112, 1-8. 464 PALAEONTOLOGY, VOLUME 29 ROBISON, R. A. 1985. Affinities of y4v.?/^^’a^fl(Onychophora) with description ofa new Cambrian species. J. Paleont. 59, 226-235. ROGERS, J. c. 1984. Depositional environments and paleoecology of two quarry sites in the Middle Cambrian Marjum and Wheeler Formations, ffouse Range, Utah. Geology Stud. Brigham Young Univ. 31,97-115. ROWLAND, s. M., KANiM, N. R., PARKTR, M. s., PETERSON, c. G. and VAN VACTOR, s. s. 1984. Onshore-offshore patterns of Cambrian shelly taxa in the western United States. Abstr. Progm geol. Soc. Am. 16, 640. RUNNEGAR, B. 1982. A molecular-clock date for the origin of the animal phyla. Lethaia, 15, 199-205. SANDERS, H. L. 1960. Benthic studies. III. The structure of the soft-bottom community. Limtiol. Oceanogr. 5, 138-153. SATTERTHWAiTE, D. F. 1976. Paleobiology and paleoecology of Middle Cambrian algae from western North America. Ph.D. thesis (unpublished). University of California, Los Angeles. SAVRDA, c. E., BOTTJER, D. J. and GORSLiNE, D. s. 1984. Development of a comprehensive oxygen-deficient marine biofacies model: Evidence from Santa Monica, San Pedro, and Santa Barbara Basins, California continental borderland. BuU. Am. /l55. Petrol. Geol. 68, 1 179-1 192. SCHAFER, w. 1972. Ecology and palaeoecology of marine environment.^, 568 pp. Oliver and Boyd, Edinburgh. SCHOPF, T. }. M. 1978. Fossilization potential of an intertidal fauna; Friday ffarbor, Washington. Paleobiol. 4, 261-270. SCHRAM, F. R. 1979. The Mazon Creek biotas in the context of a Carboniferous faunal continuum. In nitecki, M. H. (ed.). Mazon Creek fossils, 159-190. Academic Press, New York. SCHWARTZ, D. A., SEPKOSKi, J. J. and BAMBACH, R. K. 1977. Specics-abundance distributions in Phanerozoic marine communities. Abstr. Progm geol. Soc. Am. 9, 316. scoTESE, c. R., BAMBACH, R. K., BARTON, c., VAN DER voo, R. and ZIEGLER, A. M. 1979. Paleozoic base maps. J. Geol. 87,217-277. SCOTT, G. H. 1963. Uniformitarianism, the uniformity of nature, and paleoecology. N.Z. Jl Geol. Geophvs. 6, 510-527. SCOTT, R. w. 1972. Preliminary ecological classihcation of ancient benthic communities. Int. geol. Congr. 24, sect. 7, 103-110. 1978. Approaches to trophic analysis of paleocommunities. Lethaia, 11, 1-14. SEiLACHER, A. 1962. Form und Funktion des Trilobiten-Daktylus. Palaeont. Z. 36, 218-227. SEPKOSKI, J. J. 1978. A kinetic model of Phanerozoic taxonomic diversity. I. Analysis of marine orders. Paleobiol. 4, 223-251. 1979. A kinetic model of Phanerozoic taxonomic diversity. II. Early Phanerozoic families and multiple equilibria. Ibid. 5, 222-251. 1981a. The uniqueness of the Cambrian fauna. U.S. geol. Surv. Open-File Rep. 81-743, 203-207. 19816. A factor analytic description of the Phanerozoic marine fossil record. Paleobiol. 7, 36-53. 1984. A kinetic model of Phanerozoic taxonomic diversity. III. Post-Paleozoic families and mass extinctions. Ibid. 10, 246-267. and SHEEHAN, p. M. 1983. Diversification, faunal change and community replacement during the Ordovician radiations. In tevesz, m. j. s. and mccall, p. l. (eds.). Biotic interactions in Recent and fossil benthic communities, 613-111 . Plenum, New York. SHAW, K. M., LAMBSHEAD, p. J. D. and PLATT, H. M. 1983. Detection of pollution-induced disturbance in marine benthic assemblages with special reference to nematodes. Mar. Ecol. Prog. Ser. 11, 195-202. SIGNOR, p. w. and brett, c. e. 1984. The mid-Paleozoic precursor to the Mesozoic marine revolution. Paleobiol. 10, 229-245. SMOL, J. p. 1981. Problems associated with the use of ‘species diversity’ in paleolimnological studies. Qiiaternarv Res.N.Y.lS,209-2\2. SPRINKLE, J. 1973. Morphology and evolution of blastozoan echinoderms. Spec. Pubis Mus. comp. zool. Harv. Univ. 283 pp. — 1980. Early diversification. In broadhead, t. w. and waters, j. a. (eds.). Echinoderms. Notes for a short course. Univ. Tennessee Dept. geol. Sci. Stud. Geol. 3, 86-93. STAFF, G., POWELL, E. N., STANTON, R. J. and CUMMINS, H. 1985. Biomass: is it a useful tool in paleocommunity reconstruction? Lethaia, 18, 209-232. STANTON, R. J. and NELSON, p. c. 1980. Reconstruction of the trophic web in paleontology: community structure in the Stone City Formation (Middle Eocene, Texas). J. Paleont. 54, 1 18-135. POWELL, E. N. and nelson, p. c. 1981. The role of carnivorous gastropods in the trophic analysis of a fossil community. Malacologia, 20, 451-469. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 465 STURMER, w. and BERGSTROM, J. 1976. The arthropods Mimetaster and Vachonisia from the Devonian Hunsriick Shale. Palaeont. Z. 50, 78-111. 1978. The arthropod Cheloniellon from the Devonian Hunsriick Shale. Ibid. 52, 57-81. SCHAARSCHMIDT, F. and MiTTMEYER, H.-G. 1980. Versteinertes Leben im Rontgenlicht. Kl. Senckenberg- Reilie, 11, 1-79. SUGIHARA, G. 1980. Minimal community structure: an explanation of species abundance patterns. Am. Nat. 116, 770-787. SURLYK, F. and hurst, j. m. 1984. The evolution of the early Paleozoic deep-water basin of North Greenland. Bull. geol. Soc. Am. 95, 131-154. TAPPAN, H. 1970. Microplankton, ecological succession and evolution. Proc. N. Am. Paleont. Couv. Chicago 1969, H, 1058-1103. TAYLOR, M. E. 1976. Indigenous and redeposited trilobites from Late Cambrian basinal environments of central Nevada. J. Paleont. 50, 668-700. and COOK, h. e. 1976. Continental shelf and slope facies in the Upper Cambrian and lowest Ordovician of Nevada. Geology Stud. Brigham Young Univ. 23, 181-214. THAYER, c. w. 1975. Morphological adaptations of benthic invertebrates to soft substrata. J. mar. Res. 33, 177-189. 1983. Sediment-mediated biological disturbance and the evolution of marine benthos. In tevesz, m. j. s. and MCCALL, p. L. (eds.). Biotic interactions in Recent and fossil benthic communities, 479-625. Plenum Press, New York. THOMPSON, J. B., MULLINS, H. T., NEWTON, c. R. and VERCOUTERE, T. L. 1985. Alternative biofacies model for dysaerobic communities. Lethaia, 18, 167-179. TITUS, R. and cameron, b. 1976. Fossil communities of the Lower Trenton Group (Middle Ordovician) of central and north-western New York State. J. Paleont. 50, 1209-1225. TUNNICLIFFE, V. 1981 . High species diversity and abundance of the epibenthic community in an oxygen-deficient basin. Nature, Land. 294, 354-356. TURPAEVA, E. p. 1957. Food interrelationships of dominant species in marine benthic biocoenoses. Trans. Inst. Oceanol. Mar. Biol. USSR Acad. Sci. 20, 137-248. [Transl. Am. Inst. Biol. Sci. Wash. 1959.] UGLAND, K. I. and GRAY, J. s. 1982. Lognormal distributions and their concept of community equilibrium. Oikos, 39, 171-178. VALENTINE, J. w. 1973. Evolutionary paleoecologv of the marine biosphere, 511 pp. Prentice-Hall, Englewood Chlfs, N. J. 1975. Adaptive strategy and the origin of grades and ground-plans. Am. Zool. 15, 391-404. WADE, M. 1972. Dickinsonia: polychaete worms from the late Precambrian Ediacara fauna. South Australia. Mem. QdMus. 16, 171-190. WALCOTT, c. D. 191 Ifl. Middle Cambrian holothurians and Medusae. Smithson, misc. Colins, 57, 41-68. 19116. Middle Cambrian annelids. Ibid. 109-144. 1912. Middle Cambrian Branchiopoda, Malacostraca, Trilobita and Merostomata. Ibid. 145-228. 1919. Middle Cambrian Algae. Ibid. 67, 217-260. 1920. Middle Cambrian Spongiae. Ibid. 261-364. WALKER, K. R. 1972fl. Trophic analysis: a method for studying the function of ancient communities. J. Paleont. 46, 82-93. 19726. Community ecology of the Middle Ordovician Black River Group of New York State. Bull. geol. Soc. Am. 83, 2499-2524. and BAMBACH, R. K. 1971. The significance of fossil assemblages from fine grained sediments: time averaged communities. Abstr. Progm geol. Soc. Am. 3, 783-784. 1974. Feeding by benthic invertebrates: classification and terminology for paleoecological analysis. Lethaia, 7, 61 -IS. WEST, R. R. 1977. Organism-substrate relations: terminology for ecology and palaeoecology. Ibid. 10, 71-82. WHITTAKER, R. H. 1965. Dominance and diversity in land plant communities. Science, N. Y. 147, 250-260. 1972. Evolution and measurement of species diversity. Taxon, 21, 213-251. 1975. Communities and ecosystems (2nd edition), 385 pp. Macmillan, New York. WHITTINGTON, H. B. 1971u. The Burgess Shale: history of research and preservation of fossils. Proc. N. Am. Paleont. Conv. Chicago 1969, 1, 1 170-1201. 19716. Redescription of Marrella splendens (Trilobitoidea) from the Burgess Shale, Middle Cambrian, British Columbia. Bull. geol. Surv. Can. 209, 1 -24. 466 PALAEONTOLOGY, VOLUME 29 WHITTINGTON, H. B. 1975fl. The enigmatic animal Opabinia regalis, Middle Cambrian, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B271, 1-43. \915b. Trilobites with appendages from the Middle Cambrian, Burgess Shale, British Columbia. Fossils Strata, 4, 97-136. 1977. The Middle Cambrian trilobite Naraoia, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B280, 409-443. 1980a. The signihcance of the fauna of the Burgess Shale, Middle Cambrian, British Columbia. Proc. Geol. Ass. 91, 127-148. 1980/1. Exoskeleton, moult stage, appendage morphology, and habits of the Middle Cambrian trilobite Olenoides serratus. Palaeontology, 23, 1 7 1 -204. 1981a. Cambrian animals: their ancestors and descendants. Proc. Linn. Soc. N.S.W. 105, 79-87. 1981/?. Rare arthropods from the Burgess Shale, Middle Cambrian, British Columbia. Phil. Trans. R. Soc. B292, 329-357. 1982. The Burgess Shale and the early evolution of metazoan animals. In gallitelli, e. m. (ed.). Palaeontology, essential of historical geology, 1 1-24. S.T.E.M. Mucchi, Modena. 1985. Tegopelte gigas, a second soft-bodied trilobite from the Burgess Shale, Middle Cambrian, British Columbia. J. Paleont. 59, 1251 -1274. and BRIGGS, D. E. G. 1985. The largest Cambrian animal, Anomalocaris, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B309, 569-618. and CONWAY morris, s. (eds.). 1985. Extraordinary fossil biotas: their ecological and evolutionary significance. Ibid. B311, 1-192. wiEDMAN, L. A. 1983. Biovolumc rehashed: testable reproductivity of biovolumetric parameters as a viable data collection alternative in community reconstruction. Abstr. Progm geol. Soc. Am. 15, 248. 1985. Community paleoecological study of the Silica Shale equivalent of northeastern Indiana. J. Paleont. 59, 160-182. WUTTKE, M. 1983a. ‘Weichteil-Erhaltung’ durch lithifizierte Mikroorganismen bei mittel-eozanen Vertebraten aus den Olschiefern der ‘Grube Messel’ bei Darmstadt. Senckenberg. leth. 64, 509-527. 1983/?. Aktuopalaontologische Studien uber den Zerfall von Wirbeltieren. Teil 1. Anura. Ibid. 529-560. ZANGERL, R. 1971. On the geologic significance of perfectly preserved fossils. Proc. N. Am. Paleont. Conv. Chicago 1969, I, 1207-1222. and RICHARDSON, E. s. 1963. The paleoecological history of two Pennsylvanian black shales. Fieldiana, Geol. Mem. 4, i-xii, 1-352. ZOBELL, c. E. 1938. Studies on the bacterial flora of marine bottom sediments. J. sedim. Petrol. 8, 10-18. -- 1942. Changes produced by microorganisms in sediments after deposition. Ibid. 12, 127-136. Typescript received 5 July 1985 Revised typescript received 9 January 1986 S. CONWAY MORRIS Department of Earth Sciences University of Cambridge Downing Street Cambridge CB2 3EQ APPENDIX List of institutions with holdings of Burgess Shale material* Adelaide, Department of Geology and Geophysics (Australia) Birmingham, Department of Geology (UK) Bristol, Department of Geology (UK) Cambridge, Museum of Comparative Zoology at Elarvard (USA) Cambridge, Sedgwick Museum (UK) Cardiff, National Museum of Wales (UK) Chicago, Field Museum (USA) Cincinnati, Department of Geology (USA) Cincinnati, Museum of Natural History (USA) Colorado Springs, Department of Geology (USA) * Information regarding additional listings will be welcomed by the author. CONWAY MORRIS: COMMUNITY STRUCTURE OF PHYLLOPOD BED 467 Lancaster, North Museum, Franklin and Marshall College (USA) London, Department of Palaeontology, British Museum (Natural History) (UK) London, Department of Geology, University College (UK) Lund, Department of Geology (Sweden) Melbourne, National Museum of Victoria (Australia) Missoula, Department of Geology (USA) Montreal, Redpath Museum (Canada) New Haven, Peabody Museum (USA) New York, American Museum of Natural History (USA) Oslo, Paleontologisk Museum (Norway) Ottawa, Geological Survey of Canada (Canada) Perth, Department of Geology (Australia) Pittsburgh, Carnegie Museum (USA) Stuttgart, Staatliches Museum (Federal Republic of Germany) Sydney, Department of Geology and Geophysics (Australia) Toronto, Royal Ontario Museum (Canada)t Tubingen, Institute and Museum of Geology and Palaeontology (Federal Republic of Germany) Uppsala, Palaeontological Institute (Sweden) Washington, DC, National Museum of Natural History (USA) t Small collections distributed by R.O.M. to many institutions in Canada. PORIFERAN AFFINITIES OF MESOZOIC STROMATOPOROIDS by R. A. WOOD and j. reitner Abstract. The finding of calcite and pyrite spicule pseudomorphs of monaxon spicules in six genera of Mesozoic stromatoporoids confirms poriferan affinities for at least some representatives of this group. Previously, the systematic position has been speculated upon solely from the internal organization of the skeleton. Stromatoporoids, tabulates, and chaetetids have achieved some notoriety owing to their proposed reclassification from the Cnidaria to the Porifera on the basis of their morphological similarity to some recently discovered sponges, the sclerosponges, which possess both a calcareous and a spicular siliceous skeleton (Hartman 1969, 1979; Hartman and Goreau 1970, 1972). Evidence in the form of spicule pseudomorphs has confirmed the poriferan affinity of chaetetids (Gray 1980; Kazmierczak 1979), and possibly some tabulates (Kazmierczak 1984), but the biological standing of stromatoporoids has remained problematic (for example, see Kazmierczak and Krumbein 1983; Mori 1984; Steam 1972). MATERIAL EXAMINED Several specimens of Mesozoic stromatoporoids that contain spicule pseudomorphs have been found in the collections of the late R. G. S. Hudson, housed in the British Museum (Natural History). They are from the Upper Jurassic of the Middle East and include several holotypes and paratypes. The same feature has also been found by one of us (J. R.) in an undescribed species of stromatoporoid from the Lower Cretaceous of Spain. Table 1 summarizes spicular data and information on the age and localities of the specimens. Stratigraphical, geographic, and systematic details are given in the cited references. All the spicules are preserved as pyrite or calcite pseudomorphs, either as styles/acanthostyles or possibly as tylostyles, and are arranged intramurally within the axial zone of the skeletal elements. The pyrite pseudomorphs are found as aggregates or chains of crystals arranged along the length of the spicules similar to previously described fossil chaetetid and sclerosponge spicules (Gray 1980; Kazmierczak 1 979). These pseudomorphs are found in large numbers, towards the outer edge of the specimen where their original siliceous mineralogy has been replaced, probably as a result of the prolonged leaching by iron-rich pore-waters (PI. 35, figs. 3 and 4). The calcite pseudomorphs appear as rods of monocrystalline calcite of a higher transparency than the surrounding microstructural fibres, and are generally preserved in the central areas of the specimen. In Dehornella crustans Hudson, where both pyritic and calcitic pseudomorphs are found, the length of the pyrite type is considerably reduced, indicating that pyritization has occurred after partial dissolution of the spicules, producing imperfect replacement, especially of the distal tips. In species where both pseudomorph mineralogies are found, or where obvious corrosion has taken place, only the maximum figures are given. These figures, therefore, certainly represent reduced dimensions of the original spicules. According to the most recent classification of Mesozoic stromatoporoids (Hudson 1960), all the spiculate species belong to the Milleporellicae, except Actinostromarianina lecornpti Hudson, which belongs to the Actinostromariicae. The skeleton of the Milleporellicae consists of fascicular-fibrous microstructure, connected by thin IPalaeontology, Vol. 29, Part 3, 1986, pp. 469 473, pi. 35.| 470 PALAEONTOLOGY, VOLUME 29 tabulae of granular microstructure. There is no lamellar development. The spicules found in these have tapering points, and diverge distally, subparallel to the fibres of the columns (PI. 35, figs. 1-3). Pseudomorphs do not project into the lumina, and are found singly or in assemblages. Such isolated clusters appear to be remnants of a much denser spicule skeleton, the sporadic distribution of which is now diagenetically determined. In the Actinostromariicae there is no microstructural differentiation of the pillars and lamellae. The reticulum forms a rectangular meshwork where the elements have a granular central zone and an orthogonal fibrous outer zone. Tabulae of granular microstructure are secreted across the interskeletal spaces. In A. lecompti the pseudomorphs are long and thin. They are probably tylostyles, with possible bosses at their proximal end (PI. 35, fig. 5). The calcite and pyrite pseudomorphs occur side by side and are found both within the pillars and lamellae, where they are generally aligned subparallel to the growth axis of the skeleton. The spicules are evenly distributed, irrespective of their position relative to the skeletal elements. Pseudomorphs can be seen to project into the interskeletal spaces or to terminate abruptly at the skeletal surface due to the corrosion of the projecting length. The spicular positioning appears to form the framework for the subsequent secretion of the calcareous skeleton (PI. 35, fig. 6). CONCLUSIONS The size and distribution of the spicules and the microstructure and arrangement of the calcareous skeleton of these species of Mesozoic stromatoporoid show similarities to representatives of the sclerosponge genera Murania (Kazmierczak 1974), Astwsclera, Ceratoporella (Hartman 1969; Hartman and Goreau 1970, 1972), and Calcifihrospongia (Hartman 1979). The systematic position of the sclerosponges is open to question. Some workers (Vacelet 1970, 1985; Levi 1973) believe that the sclerosponges form a polyphyletic and unnatural group and that the members can be better placed in the pre-existing taxa of the Demospongiae. It is likely that the sphinctozoans, tabulates, and the stromatoporoids are also unnatural groupings and that the finding of spicules will enable the construction of a more meaningful classification and phylogenetic reconstruction of these convergent groups. However, spicules cannot be used as a species-specific characteristic due to the rarity of their preservation. The positive placing of at least some representatives of the stromatoporoids in the Porifera has several other consequences. The terminology, at present based on cnidarian nomenclature, will have to be revised and a classification that, where possible, incorporates spicule data needs to be developed. This will enable stromatoporoids to be studied from a biological standpoint, and allow a valid appraisal of their ecological role as locally significant faunal elements in Mesozoic carbonate buildups. Acknowledgements. R. A. W. would like to thank her supervisors Dr P. W. Skelton, Dr B. R. Rosen, and Dr S. Conway Morris for their help and encouragement. We wish to thank Dr J. Vacelet, Dr N. W. Rogers, and Dr J. Kazmierczak for their comments, and Mr C. Shute for photographic support. R. A. W. conducted this work under the tenure of a NERC studentship which is gratefully acknowledged. . EXPLANATION OF PLATE 35 Figs. 1-6. Spiculated Mesozoic stromatoporoids; transmitted light photomicrographs of thin sections. 1, Dehornella crustans (Hudson), H5170c. Calcite pseudomorphs of style/acanthostyle spicule arranged sub- parallel to the microstructural fibres of the calcitic skeleton, x 100. 2, Dehornella n. sp., H5478a. Calcite pseudomorphs of style/acanthostyle spicules arranged in ?lattice. x 175. 3, Parastromatopora libani (Hudson), H4789. Large numbers of pyrite pseudomorphs in outer leached zone, x 10. 4, detail of individual pseudo- morphs of specimen shown in fig. 3. x 600. 5, Actinostromarianina lecompti (Hudson), H4608a. Pyrite pseudo- morphs of ?tylostyle spicules, x 330. 6, Spicule framework of specimen shown in fig. 5. x 50. PLATE 35 WOOD and REITNER, spiculated Mesozoic stromatoporoids H = holotype; P = paratype; Fasc. Fib. = Fascicular Fibrous; s.m. = subparallel to microstructural fibres; s.s. = subparallel to growth axis of skeleton. Specimen numbers refer to the R. G. S. Hudson Collection housed in the British Museum (Natural History). WOOD AND REITNER: PORIFERAN AFFINITIES OF MESOZOIC STROM ATOPOROIDS 473 REFERENCES GRAY, D. I. 1980. Spicule pseudomorphs in a new Palaeozoic chaetetid and its sclerosponge affinities. Palaeontology, 23, 803-820. HARTMAN, w. D. 1969. New genera and species of coralline sponges from Jamaica. Postilla, 137, 1-39. 1979. A new sclerosponge from the Bahamas and its relationship to Mesozoic stromatoporoids. In levi, c. and BOURY-ESNAULT, N. (eds.). Biologic des Spongiares. Coll. Internat. chi C.N.R.S. 291, 467 474. and GOREAU, t. e. 1970. Jamaican coralline sponges: their morphology, ecology and fossil representatives. Symp. Zool. Soc. Land. 25, 205-243. 1972. Ceratoporella (Porifera; Sclerospongiae) and the chaetetid ‘corals’. Trans. Conn. Acad. Arts. Sci. 44, 133-148. HUDSON, R. G. s. 1954. Jurassic stromatoporoids from the Lebanon. J. Paleont. 28, 657-661. 1955«. Tethyan Hydroids of the Family Milleporiidae. Ibid. 30, 714-730. 19556. Sequanian stromatoporoids from South-west Arabia. Notes Mem. Moyen-Orient, 6, 225-240. 1960. The Tethyan Jurassic stromatoporoids, Stromatoporoina, Dehornella, and Astroporina. Palaeon- tology, 2, 80 99. KAZMIERCZAK, j. 1974. Lower Cretaceous sclerosponges from the Slovakian Tatra Mountains. Ibid. 17, 341-347. 1979. Sclerosponge nature of chaetetids as evidenced by spiculated Chaetopsis favei (Denninger 1906) from the Barremian of Crimea. Neues Jh. Geol. Palaeont., Mli. 2, 97-108. 1984. Favositid tabulates: evidence for poriferan affinity. Science, 225, 835-837. and KRUMBEiN, w. E. 1983. Identification of calcified coccoid cyanobacteria forming stromatoporoid stromatolites. Lethaia, 16, 207-213. LEVI, c. 1973. Systematique de la classe des Demospongia (Demosponges). In grasse, p. p. (ed.). Traite de Zoologie, vol. 3, 577-632. Masson et Cie., Paris. MORI, K. 1984. Comparison of skeletal Structures among Stromatoporoids, Sclerosponges and corals. Palaeontographica Am. 54, 354-358. STEARN, c. 1972. The relationship of the stromatoporoids to the sclerosponges. Lethaia, 5, 369-388. vacelet, j. 1970. Les eponges pharetronid actuelles. Symp. Zool. Soc. Land. 25, 189-204. 1985. Coralline sponges and the evolution of the Porifera, 113. In conway morris, s., george, j. p., GIBSON, R. and platt, h. m. (eds.). The origins and relationships of lower invertebrates. Syst. Assoc. Spec. Piihl. 28, 390 pp. Oxford University Press. Typescript received 26 March 1985 Revised typescript received 22 June 1985 R. A. WOOD Department of Earth Sciences The Open University Walton Hall Milton Keynes MK7 6AA United Kingdom J. REITNER Institut fiir Palaontologie Schwendenerstrasse 8 1000 Berlin 33 West Germany CONTRASTING LIFESTYLES IN LOWER JURASSIC CRINOIDS: A COMPARISON OE BENTHIC AND PSEUDOPELAGIC ISOCRINIDA by MICHAEL J. SIMMS Abstract. Lower Jurassic pentacrinitids have been regarded either as pseudopelagic in habit, living suspended beneath floating objects such as driftwood, or as strictly benthic, living much like all known representatives of their sister group the Isocrinidae. The taphonomy of Lower Jurassic pentacrinitids differs significantly from that of contemporary isocrinids in their environments of preservation, extent of disarticulation, occurrence in debris accumulations, frequency of association with driftwood, and the size and position of the driftwood relative to the crinoids. Unlike contemporary Isocrinidae, Lower Jurassic pentacrinitids have a low overall population density, wide geographical distribution, and rapid growth to reproductive maturity, indicated by the growth lines on brachials and by the high filtration efliciency of the endotomous arm branching. A high fecundity is implied by the large size of adult pentacrinitids and the high concentration of larval attachment discs on driftwood. These taphonomic, palaeobiological and morphological features of pentacrinitids are consistent with those which might be anticipated for pseudopelagic organisms suspended beneath temporarily floating objects of limited availability and subject to wide dispersal. The evolutionary stasis shown by the pentacrinitid Seirocrinus subangularis (Miller), which persisted from the Carixian into the mid-Toarcian, is remarkable considering the profound faunal changes which occurred in the early Toarcian and suggests that this species was not influenced by changes in the benthic environment. Although Lower Jurassic pentacrinitids are here considered essentially as obligate pseudoplankton, there is some evidence that occasional specimens may have survived as benthic crinoids after sinking into a favourable environment. More than twenty species of crinoid are known from the British Lower Jurassic, most of which can be assigned to two closely related families, the Isocrinidae Gislen and the Pentacrinitidae Gray. These two families differ significantly not only in their morphology but also in the way in which they are commonly preserved, suggesting some fundamental difference in their respective modes of life. The Isocrinidae includes many living species and recent observations of isocrinids in their natural environment has added greatly to our understanding of how they live (Conan et al. 1981; Macurda and Meyer 1974, 1983; Rasmussen 1977). All living and fossil Isocrinidae are believed to be strictly benthic, rheophilic crinoids attaching to sea-floor debris or directly to the substratum by means of the cirri which arise from the stem at regular intervals. The Pentacrinitidae have no living representatives and are most commonly preserved as intact groups associated with coalified driftwood in organic-rich shales and mudstones. William Buckland (1836), 150 years ago, noted the occurrence of Pentacrinitesfossilis Blumenbach beneath lenses of coalified driftwood in the Sinemurian of the Dorset coast and proposed that they had lived suspended beneath floating driftwood. More recently Seilacher et al. (1968) and Haude ( 1 980) have supported this view and put forward further evidence suggesting that pentacrinitids had a pseudopelagic mode of life. This hypothesis has been challenged by others (Abel 1927; Rasmussen 1977; Kauffman 1981) who interpret pentacrinitids as being strictly benthic, like the contemporary isocrinids. This paper sets out to examine the evidence for and against a pseudopelagic mode of life in Lower Jurassic pentacrinitids. (Palaeontology, Vol. 29, Part 3, 1986, pp. 475-493, pi. 36.) 476 PALAEONTOLOGY, VOLUME 29 APPROACH AND METHODOLOGY In trying to reconstruct the mode of life of fossil crinoids there are three principal lines of evidence that can be used; taphonomy, functional morphology, and palaeobiology. Although the style of preservation generally reflects the way in which an animal met its death, it can also give insight into how the animal was living prior to this event. A series of taphonomic paradigms have been constructed for both benthic and pseudopelagic crinoids against which the taphonomic history of fossil isocrinids and pentacrinitids can be compared. This represents a modification of Rudwick’s (1964) paradigm approach in which one or more functions are postulated for a given structure, each function then being used to define an abstract mechanical model called a paradigm. The paradigm which most closely approximates to the actual structure will then indicate the most probable function for that structure. In the case of the taphonomic paradigm, certain taphonomic features should diflfer considerably between benthic and pseudopelagic crinoids. It should, therefore, prove possible to determine whether pentacrinitids were benthic or pseudopelagic on the basis of the paradigm to which their actual taphonomic history most closely approximates. In addition, certain palaeobio- logical factors are considered in relation to a pseudopelagic lifestyle in crinoids. The functional significance of those morphological features unique to pentacrinitids are also assessed in the light of the two contrasting habits. Much previous work on the mode of life of Lower Jurassic pentacrinitids has tended to concentrate on the highly specialized Seirocriinis suhangularis (Miller) from the Toarcian Posidonienschiefer (Seilacher et al. 1968; Kauffman 1981). In the present work I have concentrated largely on the less advanced P. fossilis Blumenbach from the Sinemurian of the Dorset coast. Numerous specimens from this site can be found in virtually every museum with a geological collection, but are seldom associated with accurate location data. Lang and Spath ( 1 926) described the Black Ven Marls (Sinemurian) of the Dorset coast in considerable detail but perpetuated the idea of previous authors that the specimens of P. fossilis were restricted to a single impersistent horizon, the ‘Pentacrinite Bed’, about 4 ft above the Upper Flatstones on Black Ven, west of Charmouth. P. fossilis has not been seen in situ on Black Ven during the course of this study but has been observed in situ on several occasions on Stonebarrow, to the east of Charmouth. This has revealed that P. fossilis occurs most abundantly between 2 and 4 m above the Stonebarrow Flatstones but is not restricted to particular horizons within this 2 m interval. Observations by local collectors suggest that specimens may occur sporadically somewhat lower in the succession; a short distance below the Stonebarrow Flatstones (S. Barnsley, pers. comm.) and just above the Birchi Tabular (Lang Coll., BMNH E26451-26458, E51525-51526, E51528). The occurrence of P. fossilis at more than one horizon on Stonebarrow has also been reported by Jackson (1966). The following abbreviations are used: BMNH, British Museum (Natural History); OUM, Oxford University Museum; WARMS, Warwick Museum. EVIDENCE FROM TAPHONOMY Five taphonomic paradigms have been constructed in which one might expect benthic and pseudopelagic crinoids to differ. These are listed in Table 1 and are discussed individually below. Paradigm 1, Facies distribution Whereas benthic crinoids might be expected to occur in a restricted range of environments usually associated with other benthic organisms, pseudopelagic crinoids should be facies independent. Representatives of the Isocrinidae are frequent elements of normal benthic faunas in facies ranging from mudstones through silts and sands to coarse oolites. Individual species, however, occupy a much more restricted facies range. In Britain, Balanocrinus quiaiosensis Loriol only occurs in association with abundant shell debris on muddy substrata, whereas Chladocrinus scalaris (Goldfuss) is rare in such an environment and grows to only half the size of those in a siltier facies, where it is much more abundant. Isocrinids are virtually unknown in finely laminated shale facies. I have encountered only two such specimens from the British Lower Jurassic (WARMS G992I and OUM J3231). Both are SIMMS: LIFESTYLES OF LOWER JURASSIC CRINOIDS 477 TABLE I. Taphonomic paradigms for benthic and pseudopelagic crinoids. Benthic crinoids. Pseudopelagic crinoids. 1. Facies dependent. 2. F ully articulated preservation rare and due to death by sudden burial. 3. Moderate to high population densities contribute signihcantly to benthic shell beds. 4. Sometimes associated with driftwood; attached to side or upper surface. 5. Crinoid size unrelated to driftwood size. Facies independent. Fully articulated preservation common due to sinking into anoxic environments. Low overall population densities do not contribute to shell beds. Very often associated with driftwood; attached to side or lower surface. Crinoid size proportional to driftwood size. exceptionally well-preserved specimens of C. psilonoti (Quenstedt) from an unrecorded horizon in the Hettangian or Lower Sinemurian of Dorset. In contrast most specimens of Pentacrinites and Seirocrinus from the Lower Jurassic have been obtained from organic-rich shales or mudstones. In Dorset, P. fossilis is primarily confined to the lower part of the Black Ven Marls, in the obtiisum Subzone (bed 84 of Lang and Spath 1926), where specimens occur in dark, organic-rich shales and mudstones. This part of the succession is rich in ammonites but, except for one or two very thin horizons containing small specimens of Plagiostoma, tt lacks any significant benthic fauna. This suggests that the environment was anoxic. Very small (up to 1 mm) specimens of Protocardia, Liostrea, and Grammatodon are often abundant throughout bed 84 and are interpreted as bivalve spat which settled in an unfavourable environment and quickly died. Large adult individuals of Oxytoma, Gervillia, and Cuneigervillia are also encountered in these deposits but all are byssate forms which may have fallen from floating objects. Attachment of byssate bivalves to floating driftwood is well documented (Seilacher 1982) and many specimens demon- strating this are known from the obtiisum Subzone of Dorset (e.g. BMNH LL 18767). Thus if Pentacrinites fossilis is interpreted as a benthic crinoid, as proposed by Rasmussen (1977), then it represents the only normally developed benthic organism in an otherwise barren environment. Similarly P. dichotonnis (M^Coy) from the Toarcian, S. siibangidaris (MiWer) from the Pliensbachian to Toarcian, and an undescribed species of Pentacrinites from the Hettangian to Lower Sinemurian are all found most frequently in organic-rich deposits such as the Bituminous Shales of Yorkshire and the Posidonienschiefer of Southern Germany. However, occasional specimens have been found in facies containing a reasonably well-developed benthic fauna. A large group of well-preserved individuals of Pentacrinites sp. nov. from the bucklandi Zone of Keynsham, Avon (BMNH E25102) is preserved in a muddy limestone containing ammonites {Arietites sp. and ‘lEpamnwnites) and rhynchonellid brachiopods. Material from a similar facies is also known from Gloucestershire (BMNH El 22). The significance of these specimens is discussed under paradigm 2. The high proportion of specimens recorded from organic-rich facies and the great rarity of those from higher energy environments does not, however, suggest that Lower Jurassic pentacrinitids had an ecological preference for such hostile environments. This apparently unequal distribution is more likely to reflect the much greater potential for preservation of articulated specimens in anoxic environments. In addition, the low population density characteristic of pentacrinitids, discussed under paradigm 3, renders disarticulated ossicles, such as would be found in non-anoxic facies, considerably rarer than ossicles from benthic species. Finally, a very considerable collecting bias inevitably operates here since isolated ossicles are far less likely to be noticed, or collected, than intact specimens. The wide range of environments in which pentacrinitids occur suggests they were not confined to specific benthic habitats and is compatible with their having lived pseudopelagically. 478 PALAEONTOLOGY, VOLUME 29 Paradigm 2, Extent of disarticulation Preservation of fully articulated benthic crinoids is rare and generally due to death by sudden burial. Pseudopelagic crinoids can also be preserved intact but primarily in anoxic environments into which they sank. The benthic Isocrinidae are prone to disarticulate rapidly through bacterial decay of soft tissues, current action, and the activities of scavengers and other vagile benthos in normal marine environments (Meyer 1971a; Liddell 1975). Fully articulated specimens are comparatively rare and are usually the result of death through burial by a rapid influx of sediment to a depth sufficient to prevent subsequent bioturbation. Rosenkranz (1971) has described such an occurrence in the Hettangian of southern Germany and there are numerous other examples in the British Lower Jurassic. In such cases both the upper and lower surfaces of specimens are equally well preserved. Sediment generally envelops the individuals and frequently lies between the arms on the lower and the upper surface and between different individuals (PI. 36, fig. 2). Unlike isocrinids the Pentacrinitidae are usually found intact. As discussed in paradigm 1 this can occur when the crinoid sinks into an anoxic environment. In such cases burial generally occurs some time after death of the crinoids and is a gradual process, so that the upper surface of the specimens is exposed to the action of currents and occasional vagile organisms which may subsequently move across the surface. Such a style of preservation is very common in specimens of P.fossilis from the Dorset coast and has also been reported for specimens of S. subangularis from the Posidonienschiefer (Kauffman 1981). Many specimens from Dorset show little or no disruption of the lower surface (text-fig. la) but invariably show some degree of disarticulation of the upper surface, often with winnowing and size-sorting of the ossicles clearly representing the action of currents for some time after the death of the crinoids (text-fig. \h). Other specimens also show significant disruption of the lower surface in the form of resting and crawling traces of vagile benthic organisms, possibly Crustacea, which penetrate through to the lower surface as sharply defined patches of disarticulated ossicles (PI. 36, figs. 1 and 3). Such disruption without total disarticulation may be due to specimens sinking into a poorly oxygenated environment with a very sparse benthic fauna. Alternatively, a brief return to oxygenated conditions may occur before the crinoids have been buried to a sufficient depth to prevent subsequent bioturbation. This may be the case in one specimen (BMNH E69600) in which EXPLANATION OH PLATE 36 Fig. 1. Disruption of arms and pinnules by crawling traces on lower surface oPPentacrinites Bed’, Sinemurian, obtusiim Subzone, Black Ven, Charmouth, Dorset. BMNH E69605, x 1. Fig. 2. Intact preservation of Balanocriims gracilis (Charlesworth) with specimens enveloped by sediment. Domerian, stokesi Subzone, Robin’s Wood Hill, Gloucester. BMNH E69629, x 0-8. Fig. 3. Vertical section through lenticle of "Pentacrinites Bed’ and surrounding limestone. The crinoid remains are cemented by diagenetic calcite overgrowths, with virtually no interstitial sediment. Disruption caused by an unknown organism is visible towards the right of the specimen. Sinemurian, obtusiim Subzone, Black Ven, Charmouth, Dorset. BMNH E69607, x 11. Fig. 4. Endotomous branching of arms in Pentacrinites fossilis Blumenbach. Lower Lias, Lyme Regis, Dorset. BMNH E50579, xO-5. Fig. 5. Isotomous branching of arms in the isocrinid Chladocrimis robustus (Wright). Carixian, daveoi Zone, Mickleton Tunnel, Gloucestershire. BMNH E1498, xO-65. Fig. 6. Short-stemmed specimen of P. fossilis Blumenbach attached to lower surface of coalified driftwood by recurved cirri. The line marks the edge of the driftwood which is here obscured by a layer of ‘beef’, visible in the upper part of the figure. Sinemurian, obtusiim Subzone, bed 84, Stonebarrow, Charmouth, Dorset. BMNH E69600. x 0-6. Fig. 7. Variation along the stem of P. fossilis Blumenbach. Cirriferous proximal stems are visible in the upper part of the figure. In the lower part is a bundle of stems from a more distal region with short, widely spaced cirri. Sinemurian, obtusiim Subzone, Stonebarrow, Charmouth, Dorset. BMNH E69603, x 1. Fig. 8. Symplexial articulation on aboral face of cirrinodal of P. fossilis Blumenbach. Sinemurian, obtusiim Subzone, bed 84, Stonebarrow foreshore, Charmouth, Dorset. BMNH E69606, x 3-6. PLATE 36 SIMMS, Lower Jurassic crinoids 480 PALAEONTOLOGY, VOLUME 29 TEXT-i'iG. hi. intact preservation of Pentacrinites fossilis Blumenbach on lower surface of'Pentacrinites Bed’. Sinemurian, ohtusiim Subzone, Black Ven, Charmouth, Dorset. BMNH E69604, xO-5, h, disarticulation and winnowing of stems and brachials on upper surface of same specimen. a thin shelly layer with Phigiostoimi and Liostrea is separated from a bioturbated group of crinoids beneath by a few millimetres of dark, organic-rich shale. An alternative possibility for these disrupted areas is that they represent the effects of degassing during decomposition. Disarticulation on the upper surface of all specimens from Dorset shows that decomposition of soft tissues occurred to some extent even in the most anoxic environment, yet the features under discussion are seen only in certain specimens. This explanation thus seems unlikely. Further evidence that the intact preservation in pentacrinitids is due to the anoxicity of the environment, rather than death by burial, is seen from the general absence of any interstitial sediment within groups of these crinoids. Specimens of Pentacrinites fossilis from the Dorset coast and elsewhere are characteristically preserved as thin limestone lenticles produced by the diagenetic overgrowth of the calcite of the skeleton into the interstices between the ossicles (PI. 36, fig. 3). Sediment inclusions, where present, are always very minor suggesting a slow rate of sediment accumulation. Although Lower Jurassic pentacrinitids are usually preserved intact in anoxic environments, one group of individuals of Pentacrinites sp. nov., from the hucklandi Zone of Keynsham, Avon (BMNH E25102), appears to have been killed and preserved by a rapid influx of coarse sediment. The specimen was collected in the early nineteenth century (Parkinson 1808; Townsend 1813) and no field observations are available. However, the upper surface shows negligible disarticulation and a significant interstitial sediment component is present, thus resembling the style of preservation characteristic of benthic isocrinids. This specimen is of considerable importance since it suggests that pentacrinitids were quite capable of surviving as benthic crinoids if they sank into a reasonably favourable environment. Paradigm 3, Debris accumulations Benthic crinoids frequently attain moderate to high population densities and may contribute significantly to benthic shell beds. Pseudopelagic crinoids should have a low overall population density and so cannot contribute significantly to benthic shell beds. Crinoidal limestones and other accumulations of crinoid debris are common features of many Lower Jurassic sequences but are composed only of isocrinid debris. Pentacrinitids are only represented as rare isolated ossicles in such accumulations and this suggests they existed in much SIMMS: LIFESTYLES OF LOWER JURASSIC CRINOIDS 481 lower population densities than contemporaneous benthic crinoids. One possible factor causing this may be the limited availability of suitable floating objects for attachment, such as driftwood. Another possible cause has been doeumented by Schafer (1972, p. 122) for the pseudopelagic cirripede, Lepas. Such a mode of life encourages a wide geographical distribution but diminishes the density of populations. After death the skeletal parts of Lepas fall individually to the sea-floor as the soft tissues deeay and are thus spread over a wide area and never contribute significantly to shell beds on the sea-floor. A very similar situation seems to prevail in pentacrinitids, with isolated ossicles being rare and widely scattered. Paradigm 4, Association with driftwood Association of benthic crinoids with sunken driftwood is occasional and fortuitous, with the crinoids attaching to the upper surface of the driftwood. Pseudopelagic crinoids are often associated with driftwood and invariably attach to the lower surface of the driftwood. Most Lower Jurassic isocrinids must have adopted a posture like that portrayed by Rasmussen (1977, fig. 2) with the stem anchored directly to the substratum by the cirri, much as in extant isocrinids (Conan et cd. 1981; Macurda and Mayer 1974, 1983). Available sea-floor debris, including driftwood could also have been utilized for anchorage. Inevitably in such cases the crinoids could only have attached to the side or top of the driftwood and are not found beneath it (BMNH E69601). Lower Jurassic pentacrinitids are found associated with coalified driftwood with much greater frequency. In all well-documented examples of P. fossilis from the Dorset coast found in association with driftwood, the crinoids have been found to lie at least partly beneath the driftwood, as noted by several previous collectors (Buckland 1836; Jackson 1966). The consistent occurrence of coalified driftwood overlying the crinoids is very difficult to explain plausibly without invoking a pseudo- pelagic mode of life. Pseudopelagic crinoids will have lived suspended beneath a floating log and will therefore have been trapped beneath it when the log eventually sank to the sea-floor (text-fig. 2a, h). Haude ( 1 980) considered that increasing hydrostatic pressure as the log sank would have decreased its buoyancy and thus accelerated its rate of descent so that it would have come to settle on the bottom before the crinoids. This seems a very plausible hypothesis but the evidence from specimens on the Dorset coast does not support it. Isolated individuals or small groups of pentacrinitids are not always found associated with driftwood. It is probable that such specimens became accidentally detached from the driftwood whilst it was still afloat and sank to the bottom where they were preserved due to anoxic conditions. Further strong evidence for attachment of pentacrinitids to floating driftwood is provided by the distribution of larval attachment discs of crinoids on a piece of coalified driftwood (text-fig. 2c) recently discovered on the Dorset coast. In this specimen, as in many others from the Black Ven Marls, the driftwood has been compressed without any significant lateral distortion (Briggs and Williams 1981) so that the lower surface represents the lower half of the floating log and the upper surfaee the upper half. The lower surface of this particular specimen is obscured by undisturbed adult specimens of P. fossilis but the marginal 2 cm of the upper surface is covered with numerous very small attaehment discs of larval crinoids, almost certainly belonging to P. fossilis. The remainder of the upper surface is devoid of any larval discs except for a small group about 7 cm from the driftwood margin. The specimen as preserved has a maximum width of 12 cm but clearly represents part of a mueh larger log more than 20 cm in diameter. The straight edge of the specimen represents the side of the floating log and the absence of a significant number of attachment discs towards the centre, representing the upper part of the log, strongly suggests that it was colonized whilst floating at the surface with the uppermost part projecting above the water. Initial colonization would be on the lower surface but as this became crowded with the rapidly growing crinoids and the driftwood sank lower in the water further.settlement of larval crinoids would be restricted to higher, less favourable areas exposed to greater turbulence. However, a large area of the upper surface would remain uneolonized as long as the driftwood was afloat because of its exposed position at or close to the water surface. Upon sinking to the sea-floor this area would become accessible to settling crinoid larvae. However, it is evident that no further settlement took place once the driftwood had reached 482 PALAEONTOLOGY, VOLUME 29 a TEXT-FIG. 2. Distribution of crinoids on floating drift- wood. a, diagrammatic reconstruction of crinoid distribu- tion on floating driftwood reflecting colonization by successive generations of larvae on progressively higher parts of the driftwood, h, cross-sectional appearance of crinoids and driftwood after burial and compression. The larger crinoids are seen as a white layer extending from beneath the coalified driftwood whilst the last generation of larvae to settle prior to sinking of the driftwood are found along the margins of the upper surface of the com- pressed driftwood, c, distribution of larval attachment discs (dots) of Pentacrinites fossilis Blumenbach on the upper surface of a fragment of compressed and coalified driftwood. The hatched area in the lower part of the figure represents the larger crinoids extending from beneath the driftwood. Sinemurian, obtusum Subzone, Black Ven, Charmouth, Dorset. BMNH E69602. b SIMMS: LIFESTYLES OF LOWER JURASSIC CRINOIDS 483 TEXT-FIG. 3. Plot of maximum columnal diameter/driftwood diameter for Isocrinidae and Pen- tacrinitidae associated with drift- wood. Pentacrinitid size is correlated with driftwood size, whereas that of Isocrinids is not. Columnal diameter is used as a size index because of the fragmentary nature of most isocrinid material. Based on museum material and observation of specimens in situ. 12 '§11. 10- 0) <♦«# 9- Q Tagelus divisus (Laguna Madre) Z 50 u. 45 - 1 0 1 5 2 0 2.5 >20mm 1 0 1,5 2,0 2,5 Tellina tampaensis (Laguna Madre) 150 1 0 1 5 2 0 2,5 > 3 0 MEAN SIZE (mm) ■ HIGH POINT □ LOW POINT TEXT-FIG. 8. Comparison of the size-frequency distribution in the death assemblage just after an input pulse (the abundance high, see text-fig. 1 ) with the size-frequency distribution after taphonomic decay (the abundance low, see text-fig. 1) for species with highly positively-skewed distributions. Littoridina barretti (Copano Bay) 10 1 0 1 5 2 0 2,5 3,0 Littoridina sphinctostoma (Copano Bay) 1 0 1 5 2.0 2 5 3 0 >35 MEAN SIZE (mm) ■ HIGH POINT 0 LOW POINT TEXT-FIG. 9. Comparison of the size-frequency distribution of species in the death assemblage just after an input pulse (the abundance high, see text-fig. 1) with the size-frequency distribution after taphonomic decay (the abundance low, see text-fig. 1). CUMMINS £T/4L.: MOLLUSC TAPHONOMY IN TEXAS BAYS 503 changed significantly by the addition of individuals and their subsequent decay (P ^ 010, Chi- square or Kolmogorov-Smirnov two-sample two-sided tests) (Table 1). Survivorship curves for species in the death assemblage were calculated and plotted on log paper to approximate age-specific survivorship as suggested by Thayer (1977) (text-fig. 11). Most are concave (Type II of Deevey 1947) or linear (i.e. constant mortality— Type I). A few such as M. lateralis and Rangia cimeata are sigmoidal. Two, T. pleheius and Maconia mitclielli, are convex (Type III). The survivorship curves show that species were grouped by the cluster analysis primarily by the proportion of juveniles in the population. Species also can be grouped from the pattern of the adult survivorship by visual inspection of the survivorship curves. These groupings were similar to those derived by the cluster analysis except that R. cimeata could be grouped with taxa with moderately positively skewed size-frequency distributions and Anomalocardia auheriana with other species from Laguna Madre with highly positively skewed size-frequency distributions. In addition, bimodality was not sufficiently intense to produce an obvious plateau in the survivorship curve for any of the three bimodal species. Q LU H (J LU O o OT _J < 3 9 > Q Z Ll_ o cc LU LU D Z Mulinia lateralis (Laguna Madre) Diastoma vanum (Laguna Madre) 1.0 15 2,0 >2.5 1 0 1 5 2 0-3 0 ’35 Mulinia lateralis (Copano Bay) MEAN SIZE Crepidula convexa (Laguna Madre) (mm) ■ HIGH POINT □ LOW POINT TEXT-FIG. 10. Comparison of the size-frequency distributions of species in the death assemblage just after an input pulse (the abundance high, see text-fig. 1) with the size-frequency distribution after taphonomic decay (the abundance low, see text-fig. 1 ). At both sites, about half of the individuals that died during the study (as calculated from size at death data from the living community) were ^ 30% of the maximum size collected (text-fig. 12; tables 2 and 3). By comparison, in the death assemblage at the Copano Bay site, nearly two-thirds of the individuals collected were in this size range but only 26% of the individuals were as large at Laguna Madre (text-fig. 12). Thus, the two sites differed dramatically in the individual’s average size in the death assemblage; most individuals were smaller at the Laguna Madre site, yet the average size 504 PALAEONTOLOGY, VOLUME 29 at death was similar. In contrast, biomass was distributed similarly at both sites. Individuals ^ 30% of the maximum size collected accounted for over 80% of the biomass in both death assemblages and among the individuals that died during the study. TABLE I, Comparisons of the size-frequency distribution for the calculated input pulse into the death assemblage (see methods section) with the total death assem- blage and comparisons of the size-frequency distribution of species prior to tapho- nomic loss (as judged by an abundance peak in the death assemblage following an input pulse) with the size-frequency distribution following taphonomic loss (as judged by a numerical low in the death assemblage). All tests used chi-square except those marked by * in which the Kolmogorov-Smirnov two-sample test was used. Calculated input pulse vs Total death assemblage Pre-taphonomic loss vs Post-taphonomic loss Laguna Madre Diastonui varium P < 0 001 P < 0 001 Crepidida convexa P > 0-20* P > 0-20* Tagelus divisus P > 0-20 P > 0-20* Laevicardium morloni 0 025 < P < 0 05 0 005 < P < 0 010 Tellimi tampaensis P < 0 001 0 01 < P < 0 025 Mysella plamdata 0 05 < P < 010 P > 0-20* Midinia lateralis P < 0 001 P > 0-20* Acteocina canaliculata P < 0 001 P > 0-20* Odostoina cf. teres P < 0 001 — Copano Bay Littoridina barretti P < 0 001 0 05 < P < 01* Tagelus plebeius P < 0 001 — Macuma mitchelli P < 0 001 — Rangia cuneata P < 0 001 — Mulinia lateralis 0 005 < P < 0 01 P > 0-20* Littoridina sphinctostoma P < 0 001 0 01 < P < 0 025 However, considerable variability was present from one species to another in the number of individuals in the death assemblage ^ 30 % of the maximum size collected. Less than 10 % of the individuals of most species with large input pulses were this large (Table 4). Acteocina canaliculata was the atypical member of this group. Over 69 % of its individuals were in this size range. Species with small pulses or without input had proportionately more large individuals in the death assemblage (Tables 5 and 6). Furthermore, on the average, for the latter two groups, proportionally more large individuals were present at the Copano Bay site than at the Laguna Madre site. For example, 52 % of the individuals of species at Copano Bay with mortality < 10 individuals per m^ over the study were ^ 30 % of the maximum size collected; the corresponding percentage at Laguna Madre was only 16T %. Eighty per cent of the individuals of species with mortality between 10 and 100 individuals per m^ over the study were of this size at the Copano Bay site, yet only 34 % were as large at Laguna Madre. In contrast, over 80 % of the biomass was in individuals ^ 30 % of the maximum size collected in nearly all species in both death assemblages (Tables 4-6). Three abundant species (Laevicardium mortoni, C. cancellata, and B. exiistus) in the death assem- blage at the Copano Bay site were represented only by juveniles, but adults were present in all NUMBER SURVIVING NUMBER SURVIVING CUMMINS £r^L.; MOLLUSC TAPHONOMY IN TEXAS BAYS 505 PERCENT MAXIMUM SIZE PERCENT MAXIMUM SIZE 0 PERCENT MAXIMUM SIZE 90 TEXT-FIG. II. Survivorship curves for taxa in the death assemblage, plotted on a log-log scale. 506 PALAEONTOLOGY, VOLUME 29 CO CO 100 < _) o 80 LU 1- N z CO 60 LU o z cr LU CL cc LU CQ 40 20 D Z >30%Mx.Sz. >50%Mx.Sz. > 80% Mx.Sz PERCENT MAXIMUM SIZE ■ COPANO BAY □ LAGUNA MADRE h- Z LU o oc LU CL PERCENT MAXIMUM SIZE 100 >30% Mx. Sz. >50% Mx. Sz, >80% Mx. Sz. PERCENT MAXIMUM SIZE ■ COPANO BAY □ LAGUNA MADRE TEXT-FIG. 12. The average percentage of individuals and biomass of all species in the death assemblage reaching the indicated percentage of the maximum size collected. abundant species in the death assemblage at Laguna Madre (Table 2). Most individuals (over 90 %) of most species were juveniles in both death assemblages. Of the common species, only three at Copano Bay, A. canaliciilata, T. pleheius, and M. mitchelli, had > 45 % of the individuals of adult size. Only one species, again A. canaUculata, had such a high proportion of adults at Laguna Madre. As expected, those species with negatively skewed size-frequency distributions had pro- portionally more adults. Species with moderately or highly positively skewed size-frequency distri- butions, however, differed little in the proportion of adults present (except A. canalicidata). CUMMINS £T /4L.: MOLLUSC TAPHONOMY IN TEXAS BAYS 507 TABLE 2. The observed maximum size, known maximum size, and the percentage of individuals reaching reproductive maturity for species from Copano Bay and Laguna Madre. Maximum size collected in mm Known maximum size in mm Maximum size collected as a percentage of known maximum size Percentage of individuals reaching reproductive maturity Copano Bay Laevicardiiim mortoni 7-96 23-0 26-0 0-0 Acteocina canaliculata 5.12 5.0 100-2 45-0 Cliione cimcellata 9-92 38-0 26-0 0-0 Brachidontes exustus 8-96 20-7 43-0 0-0 Tageltis pie he ins 65 02 85-0 76-0 65-1 Macoma mitchelli 22-05 29-0 76-0 75-8 Raiigia cuneata 57-00 80-0 71-0 0-6 Miiliiiid lateralis 10-88 16-2 67-0 9-7 Laguna Madre Laevicardium mortoni 14-10 23-0 61-0 4-0 A cteocina canaliculata 5-12 5-0 100-2 48-0 Tagelus di visas 36-80 30-0 119-5 3-3 Anonudocardia auberiana 10-88 16-0 67-0 1-0 Diastonia variant 6-80 6-0 113-0 28-8 Crepidala convexa 9-92 12-0 82-7 6-1 Tellina tampaensis 16-25 20-0 81-0 2-8 Midinia lateralis 12-90 16-2 79-6 18-8 TABLE 3. Comparison of the percentage of individuals and biomass that reached the indicated percentage of maximum size collected (from Table 2) in the size-frequency distribution of living individuals that died during the study (i.e. the cumulative size-frequency distribution of individuals added to the death assemblage during the study). Percentage of individuals Percentage of biomass ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size ^ 30 % Max. size ^ 50% Max. size ^ 80 % Max. size Copano Bay Malinia lateralis 80-7 25-3 8-4 97-8 73-0 44-4 Rangia caneata 0-9 0-6 0-0 45-6 42-9 0-0 Macoma mitchelli 79-4 40-0 9-7 99-2 83-5 36-2 Average 53-7 22-0 6-0 80-9 66-5 26-7 Laguna Madre Tagelus di visas 4-6 2-8 0-0 81-4 74-7 0-0 Laevicardium mortoni 14-9 4-3 0-0 91 3 67-6 0-0 Crepidala convexa 50-0 22-2 5-6 94-8 75-7 57-6 Tellina tampaensis 32-3 8-2 0-5 84-2 47-0 8-4 Malinia lateralis 84-4 42-2 13-3 99-5 85-9 52-5 Acteocina canaliculata 92-4 46-0 11-1 98-7 79-0 33-7 Average 46-4 20-9 5-2 89-9 68-8 33-2 508 PALAEONTOLOGY, VOLUME 29 TABLE 4. The percentage of individuals and biomass in the death assemblage reaching the indicated percentage of maximum size collected for species whose calculated input into the death assemblage over the study period was > 100 individuals per m^. Percentage of individuals Percentage of biomass ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size Copano Bay Rangia cimeata 140 100 4-4 96-5 69-9 26-5 Laguna Madre Tagelus clivisiis 3-8 2-9 1-3 91-7 89-5 62-4 Laevicardiwn mortoui 90 2-9 0-4 90-9 69-4 27-6 Tellina tampaensis 6-4 3-2 0-8 82-9 74-7 40-6 Acteocina canalicukita 69-5 27-7 5-6 95-3 69-9 26-5 Average 22-2 9-2 20 92-2 75-9 36-7 TABLE 5. The percentage of individuals and biomass in the death assemblage reaching the indicated percentage of maximum size collected of species whose calculated input into the death assemblage over the study period was < 10 individuals per m^. Percentage of individuals Percentage of biomass ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size Copano Bay Acteocina canalicukita 87-4 39-7 1-4 970 670 6 1 Laevicardiiim mortoni 3 0 4-8 1-8 85-6 53-5 320 Brachidontes exustus 54-0 210 4-9 96- 1 76-2 37-7 Chione cancellata 35-5 4-4 11 86-3 43-6 25-0 Average 51-7 17-5 2-3 91-3 60- 1 25-2 Laguna Madre Anomalocardia auberiana 16-1 2-5 0-5 79-5 43-3 17-4 Table 7 shows the fragment to whole ratios for each site. Species at the Laguna Madre site had far fewer fragments per whole shell than did the species at the Copano Bay site. DISCUSSION Size-frequency distributions are an important tool of ecologic study. Comparison of a temporal sequence of size-frequency distributions provides data on recruitment, growth, mortality, and life history strategy of species in the community. Size-frequency distributions also are the primary tool utilized by the palaeontologist in the analysis of a species’ population structure. A temporal sequence cannot be collected, however. Consequently, the sole datum provided by a size-frequency distri- bution of a fossil species normally is the size at death of the preserved individuals, from which all CUMMINS £T MOLLUSC TAPHONOMY IN TEXAS BAYS 509 TABLE 6. The percentage of individuals and biomass in the death assemblage reaching the indicated percentage of maximum size collected for species whose calculated input into the death assemblage over the study period was between 10 and 100 individuals perm^. Percentage of individuals Percentage of biomass ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size ^ 30 % Max. size ^ 50 % Max. size ^ 80 % Max. size Copano Bay Mnlinia lateralis 62-3 20-7 5-0 96-9 74-2 34-6 Tagehis pleheius 970 91-5 6-3 99-9 98-4 16-7 Macoma mitchelli 80-6 65-8 17-7 991 96- 1 47-2 Average 80-0 59-3 9-7 98-8 89-6 32-8 Laguna Madre Mnlinia lateralis 27-0 20-6 7-7 96-2 92-4 57-6 Diastorna varinni 36-7 28-8 0-8 84-7 75-7 5-4 Crepiclnla conve.xa 36-9 7-4 0-6 82-7 41-0 6-3 Average 33-5 18-9 30 87-9 69-7 23-1 TABLE 7. Fragment to whole shell ratios for species at Copano Bay and Laguna Madre. Copano Bay Laguna Madre Acteocina canaliculata M2 Acteocina canaliculata 0-32 Littoridina harretti 0-95 A nomalocardia anheriana 0-49 Diastorna varinni 6-20 Diastorna varinni 0-49 Brachidontes exits tns 11-60 Mysella planulata 0-05 Chione cancellata 2-53 Tellina tampaensis 0-42 Tagehis pleheius 2-23 Crepidula convexa 0-32 Mnlinia lateralis 3-35 Mnlinia lateralis 0-42 Littoridina sphinctostoma 2-37 Laevicardium mortoni 0-42 Laevicardium mortoni 3-99 inferences about population dynamics must be derived. Thus, an understanding of the effects of time averaging and taphonomy on the size-frequency distribution is crucial. Size-selective preservation Each cohort of a species produces its own size-specific mortality pattern. Some of these indivi- duals are not preserved and the remainder are mixed with individuals of previous generations to form the final death assemblage (Craig and Oertel 1966; Hallam 1967, 1972). The size-specific mortality pattern of one or a few cohorts of most species in the living community at both sites compared poorly with the size-frequency distributions of species in the death assemblages. For example, no large M. mitchelli were added to the death assemblage during the study, proportionally more adult A. canaliculata were added at the Laguna Madre site than were already present in the death assemblage, and proportionally fewer large Mnlinia lateralis were added to the death as- semblage at the Copano Bay site than were already present. Either survivorship during our study differed considerably from previous years, or taphonomic processes altered the size-frequency 510 PALAEONTOLOGY, VOLUME 29 distribution in the death assemblage (Kurten 1964; Olson 1957). The distinction between the two is crucial. If variation in survivorship is more important, then the death assemblage preserves a time- averaged picture of the population dynamics of a species. If size-specific taphonomy is responsible, then its population dynamics cannot be reconstructed from the size-frequency distribution. If size-selective preservation occurred, then small shells should decay faster (Olson 1957, Hallam 1967). We compared the size-frequency distribution just after the addition of a pulse (at the maximum point of numerical abundance in the death assemblage) with the size-frequency distri- bution after taphonomic loss (at the point of lowest abundance in the death assemblage). The distributions prior to taphonomic loss were significantly different from the distributions after taphonomic loss in about one-half of the cases. Unfortunately, in our study, instantaneous mortality did not occur, so that larger individuals died later. Taphonomic decay was so rapid (text-fig. 1; Cummins et al. 19866) that considerable decay of small shells occurred before the larger shells were added. Thus, the data suggest but do not prove size-selective taphonomic loss. Nevertheless, in all but one case, proportionally more large shells were present, yet the number of small shells added by mortality was large in comparison to the number already present and the number of large shells added was very small. To explain this fact by temporal differences in survivorship would require the extraordinary coincidence of all species at both sites having unusually poor survivorship simul- taneously during this study. This is unlikely and provides a strong suggestion of the importance of size-selective taphonomy. The proportion of targe individuals in the death assemblage of species having different population levels in the living community provides additional evidence for size-selective taphonomy. Species having < 100 individuals per m^ added to the death assemblage during the study consistently had fewer small individuals in the death assemblage than other species. A. canaliciilata at Laguna Madre, a species with unusually good survivorship (Powell et al. 1984), was the only exception. At Laguna Madre, 16-37% of all individuals of these species were ^ 30 % of the maximum size collected, whereas only 4-9 % were of similar size in species with mortality > 100 individuals per m^. The respective numbers for Copano Bay were 30-97 % and 14 %. There are only three possible reasons for such a dramatic difference. Either, u, in the past survivorship was unusually high at both sites only in species which coincidentally had little input into the death assemblage during our study, or, 6, adults and juveniles live in separate habitats only for those species which coincidentally had low or no input into the death assemblage during our study, or, c, size-specific taphonomic loss occurred. The importance of adult migration for some species is well known (e.g. Beukema 1973; Werner 1956), however the species of interest here recruit primarily by larval settlement (e.g. Fraser 1967; Moore and Lopez 1969; Powell et al. 1984) so that adult migration is an unlikely cause. Temporal variability in the survivorship of cohorts is well described (e.g. Hughes 1980; Brosseau 1978), however, the coincidence of good survivorship at both sites in so many species, but only in those species that occurred rarely or not at all in the living community during this study, is unlikely. Nearly all studies substantially underestimate juvenile mortality, frequently by 80 % or more (Powell et al. 1984). Even with such substantial errors, however, few data support low juvenile mortality in gastropod and bivalve molluscs (Cadee 1982a). Almost invariably juvenile mortality is high (in addition to Cadee 1982a, we checked Dare 1976; Muus 1973; Holland and Dean 1977; Brosseau 1978; Coe and Fitch 1950; Phillips 1981; Seager 1982; Schmidt and Warme 1969; Yamada 1982; and others referenced herein and in Powell and Cummins 1985; see also Doherty 1979 and Cadee 19826 for brachiopods). At our two sites, all molluscan species that settled during 1981-1983 had high juvenile mortality (Powell et ai 1984). Consequently, most size-frequency distributions of molluscs in death assemblages produced by in situ mortality should be highly positively skewed in the absence of taphonomic processes (see also Craig and Oertel 1966; Hallam 1972). Certainly the available data on molluscan population dynamics imply that the number of species present in this study whose size-frequency distribution might otherwise suggest good survivorship would be unprecedented in bay environments like Copano Bay and Laguna Madre. Juvenile CUMMINS £T ,-4 L : MOLLUSC TAPHONOMY IN TEXAS BAYS 511 mortality normally is very high in these habitats. The proportion of individuals in the smaller size classes differs from that expected from typical molluscan population dynamics except in species which had juveniles added to the death assemblage during the study. Apparently, without continual input of juveniles, highly positively skewed size-frequency distributions would be absent from both sites because juveniles are poorly preserved. Two abundant species, L. mortoni and M. lateralis, were common at both sites. Survivorship for both species was better at Laguna Madre. Proportionally more adults were collected in the death assemblage at this site. Even so, in both cases, proportionally more of the individuals collected were in the larger size classes (even though they were not adults) at the Copano Bay site. Thus, survivor- ship of both species was better at Laguna Madre, but their size-frequency distribution was more negatively skewed at the Copano Bay site. Clearly, variations in survivorship are an unlikely explanation for the observed differences in survivorship curves and size-frequency distributions of those two species. Consequently, although the data do not prove unequivocally that size-specific taphonomy occurred, all data strongly suggest the importance of this process in determining the shape of size-frequency distributions in both death assemblages. In nearly all cases the largest size classes became proportionally more important because more individuals in the smallest size classes decayed away. Site-specific preservation The paucity of shell material in the smallest size classes at Copano Bay is typical of many fossil and death assemblages (Dodd et al. 1985; Broadhurst 1964; Brookfield 1973; Stewart 1981). There are two possible explanations for the lack of small shells in the fossil record which may also apply to Copano Bay (Broadhurst 1964; Craig and Oertel 1966; Boucot 1953). 1 . The mortality rate could be much higher for the smallest size classes at Laguna Madre than at Copano Bay. Poor survivorship of the smallest living individuals would result in a buildup of individuals in the smallest size classes in the death assemblage in spite of decay if periodic settlement were predictable. Observations and inferences as previously discussed do not support this hypo- thesis. Moreover, potential predators were observed more frequently at the Copano Bay site. The Copano Bay site also is more environmentally unstable. In fact, more taxa were present exclusively as juveniles at the Copano Bay site than at the Laguna Madre site. Thus, differential survivorship is an unlikely explanation. 2. Taphonomic processes may be different in kind or extent. Size-specific taphonomic loss may be more extreme at the Copano Bay site or the more constant input of individuals into the death assemblage at the Laguna Madre site might continually maintain a considerable number of small shells in the death assemblage in spite of high taphonomic loss. Although rates of taphonomic loss were similar for most species at both sites (Cummins et al. 19866), two species, R. cuneata and Macoma mitchelli, decayed so rapidly at the Copano Bay site that no decay rate could be measured. Such was not the case at the Laguna Madre site. Fragmentation was much higher at the Copano Bay site as well. Taxa with little or no input into the death assemblage during the study had proportionally more large individuals at the Copano Bay site. On the average, there was more time between input pulses into the death assemblage at the Copano Bay site because species set more regularly, in greater numbers, at the Laguna Madre site. Consequently, there was more time for taphonomic loss between pulses at the Copano Bay site, but more smaller shells were added to the death assemblage at the Laguna Madre site. Thus, variation in taphonomic loss plus differential rates of input of small shells offer the best explanation for the difference in size-frequency distri- butions between the two sites. Most taphonomic processes can be size selective. Smaller shells may be more easily transported than larger shells (Boucot 1953; Boucot et al. 1958, but see Trewin 1973). At both sites, most shells ^ L5 cm in size were transported, at least locally, but larger shells were transported less (Cummins et al. 1986a). Predation may be size selective (Dare 1976; Vermeij 1980). If the preyed upon individuals were broken or removed from the area, the size classes involved would be 512 PALAEONTOLOGY, VOLUME 29 under-represented in the death assemblage. Dissolution and abrasion may be more effective in certain size classes than in others (Flessa and Brown 1983; Chave 1964). Apparently such processes as dissolution and shell breakage must be more intense in Copano Bay, because physical transportation afl'ected the spatial distribution of shells more at Laguna Madre (Cummins et al. 1986n). Modality and skewness of frequency distributions The word 'mode’ has been used in several ways by biologists and geologists. The statistical definition of a mode is the measurement which occurs with the highest frequency. In contrast, in fisheries biology, the term is used frequently to designate the size class of highest frequency for a cohort (i.e. a modal class). Many cohorts may be present in a population at any given time yielding a multimodal distribution (Abramson 1971 ). Such a size-frequency distribution may be visually multimodal if the cohorts are well separated. Frequently, however, they are not, and modes are recognized by fitting normal curves to a size-frequency distribution based on the assumption that the size distribution of organisms, within cohorts, will be normally distributed. In contrast, in geology and palaeontology, modes usually are visually distinctive abundance peaks in a size-frequency distribution (e.g. Sheldon 1965). Size-frequency distributions of our species were of four types, the first three being unimodal; highly positively skewed (text-fig. 4), moderately positively-skewed (text-fig. 5), negatively skewed (text-fig. 3), and bimodal (text-fig. 6) (with the highest mode highly or moderately postively skewed). Highly postively, moderately positively, and negatively skewed distributions primarily demonstrate the relative effectiveness of size-selective taphonomy and the temporal pattern of larval settlement. No species without substantial and frequent larval settlement had a highly positively skewed size- frequency distribution. In spite of the considerable emphasis on normal or bell-shaped distributions (e.g. Hallam 1972), none of our species had a normal distribution. We attempted to fit the observed distribution pattern to a normal distribution pattern using the NORMSEP computer program without success. We suggest that normal distributions are produced very infrequently in death assemblages from bay environments. Unimodal distributions Most distributions were unimodal. The time-averaged death assemblage contains a cumulation of shells produced by yearly patterns of size-specific mortality, each year potentially being different from all others, and taphonomic decay. Nevertheless, each size-frequency distribution should be unimodal initially and highly positively skewed. To check this assumption, we calculated the expected size-frequency distribution for input pulses from size at death data available in the literature (e.g. Brousseau 1978; Phillips 1981; Yamada 1982; Coe and Fitch 1950) and from our data. All were unimodal except Gemma gemma (Jackson 1968) and Mulinia lateralis, which was bimodal at both our sites (see also Cadee 1982n). Thus, the size-frequency distribution of shells obtained from the death of a single cohort usually is unimodal. In addition, each size-frequency distribution should be subjected to taphonomic processes all of which, if size-selective, remove relatively more individuals from the smaller size classes of the size-frequency distribution. Consequently each taphonomically altered distribution should be unimodal. On the average the cumulation of many such distributions should yield a unimodal distribution, because each individual mode will not be sufficiently different from all others to cause the distribution of modes to be anything but unimodal. Thus, unimodality can be considered the standard condition for size-frequency distributions in time-averaged death assemblages. Bimodality In textural analysis of sediment the number of modes present often is indicative of the degree of physical sorting which has taken place. Sediments with multimodal grain size distributions are not at equilibrium with the physical environment (Singer and Anderson 1984; Curray 1960; Sonu 1972). CUMMINS £T, 4 L.: MOLLUSC TAPHONOMY IN TEXAS BAYS 513 Most bimodal or multimodal size-frequency distributions in palaeontology can probably be ascribed to an analogous condition of disequilibrium, and should be rare. Olson (1957) found bimodal or multimodal distributions in only 15 % of the cases he examined. Only three species had bimodal distributions in our study. Bimodality could arise for several reasons. 1. In a living population, distinct modes are frequently visible when several cohorts are present. Catastrophic mortality could preserve a cohort’s size-frequency distribution intact (Sheldon 1965). Thus, census populations might be modelled as a series of normal curves. Catastrophic mortality in which a series of cohorts died simultaneously did not occur at our two sites during the study; nevertheless, most size-frequency distributions could be modelled by the summation of two or more normal curves (using NORMSEP) as if several cohorts were present. This clearly is false. Thus, a series of normal curves can be fitted to the time-averaged mortality pattern derived from normal cohort mortality in the living community and the mere presence of such a fit cannot be used as an indicator of catastrophic death. 2. If the input into the death assemblage from the living community is visibly multimodal, one would expect the resulting death assemblage also to be multimodal. For this to occur, survivorship would have to be very low for the very young and old and very high for the middle size classes, or sexual dimorphism must be present. Our species were not sexually dimorphic enough to produce a bimodal distribution. Little evidence for bimodal mortality was available in the literature for single cohorts (see previous references). Only M. lateralis had a bimodal mortality pattern at our two sites during the study, but more than one cohort was included in this analysis. 3. If shell material is transported into the habitat, then the size-frequency distribution of the transported assemblage would be dependent upon the size-frequency distribution of the source material and the strength of the transporting event. The transporting event might cause the mixing of size-frequency distributions from two different sources which were spatially separated or from one source at two different time periods when the size-frequency distributions were different. For instance, Diastoma varium at Laguna Madre was transported into the area on floating seagrass blades (Powell et al. 1982). One transporting event may have contained a majority of juvenile specimens whereas the other contained older individuals. The resulting size-frequency distribution for D. varium is characterised by visibly distinct modes. At Copano Bay, Bracludontes exustiis has a bimodal size-frequency distribution. This bivalve generally is found attached to oyster shells and almost certainly was allochthonous. A scenario similar to that of D. varium would explain its bimodal distribution. 4. A very large input pulse with a size-frequency distribution different from that already present in the death assemblage, limited time-averaging (i.e. the summation of only a few cohorts), or the addition of juveniles to an assemblage after size-selective taphonomy had removed nearly all of the small individuals, might insert a secondary mode into the size-frequency distribution. At Copano Bay the size-frequency distribution of Tagelus pleheius was bimodal during the time of maximum juvenile mortality but before complete taphonomic loss of the small individuals had occurred. Many bimodal or multimodal size-frequency distributions within one fossil assemblage (with no evidence for physical transportation) might suggest that a catastrophic event resulted in the preservation of the size structure of a living population. On the other hand, a few bimodal size- frequency distributions, as found at both our sites, might suggest either allochthonous input into the assemblage, the presence of species with wildly fluctuating survivorship from cohort to cohort or an unusual population dynamics yielding a bimodal mortality pattern. Discrimination between these three might be difficult without corroborating evidence from other sources. All, however, represent a condition of disequilibrium from the standard unimodal condition. Biomass Usually, a small number of large individuals contribute most of the biomass present for a species. At Laguna Madre, for example, over 88 % of the total biomass is found in individuals ^ 30 % of 514 PALAEONTOLOGY, VOLUME 29 the maximum size collected despite the fact that only 27 % of the total number of individuals are in this size range. Biomass is a more effective statement of palaeo-community structure than numerical abundance because larger specimens have a better chance of being preserved (Staff et al. 1985, 1986). Consequently, the total biomass of a cohort is probably better preserved than is the cohort’s numerical abundance. Adult abundance Similarly, adults should be preserved better than juveniles (Cummins et al. 19866). The propor- tion of the total number of individuals that were adults was quite variable among the species in both death assemblages. For instance, only 1 % of the Anonudocardia auberiana and only 6 % of the D. variuni reached reproductive maturity but over 45 % of the Acteocina canaliculata and over 65 % of the T. plebeius and Maconia mitchelli were adults. Four explanations for this variability are possible. 1. Species which died before reaching reproductive maturity might indicate settlement in a marginal habitat even if they are numerically abundant. Laevicardium mortoni, for example, reached reproductive size at the Laguna Madre site, but not at the Copano Bay site. L. mortoni is a species characteristic of hypersaline lagoons (Parker 1959). Copano Bay is a marginal habitat for this species. 2. Transported individuals may all be small. At both sites, smaller individuals were transported more effectively than larger ones (Cummins et al. 1986a). The almost complete absence of adult B. exustus at the Copano Bay site might be explained similarly. 3. Size-selective taphonomy may have removed the smaller size classes. On the average, gastro- pods had more adults than bivalves. Gastropod survivorship may be generally higher or taphonomic loss may be greater for adult bivalves. Except for D. varium, the fragment to whole shell ratios tended to be lower for gastropods (i.e. more whole shells were present). Consequently, gastropods may be preserved better. 4. Survivorship may have been high. The one species with better than average survivorship, A. canaliculata, also had a higher percentage of adults at both sites. Nevertheless, the number of adult A. canaliculata present in the death assemblage is much higher than expected even from this good survivorship. Less than 30 % of the living individuals reached ^ 30 % of maximum size (Powell et al. 1984), but nearly one-half of all individuals in the death assemblage were adults. Thus, the two primary factors affecting the number of adults in the death assemblages in Copano Bay and Laguna Madre are low survivorship primarily due to settlement in suboptimal habitat and better preservation of adults. Interestingly, adults are preserved better regardless of size. That is, size-specific taphonomy may be explained at least partially by juvenile-specific taphonomy. Adult A. canaliculata are no larger than juvenile Tellina tampaensis. Adult Mulinia lateralis are about the size of juvenile Tagelus plebeius. Nevertheless, in both cases, adults were preserved better. Perhaps changes in shell mineralogy, mode of death, or life and death position oft'er explanations. CONCLUSIONS AND RECOMMENDATIONS i' Size-frequency distributions and survivorship curves have been used to study differences in popu- ij lation dynamics between species (Thayer 1977; Richards and Bambach 1975; Stanton et al. 1981). Size-frequency distributions which are positively skewed yielding survivorship curves of Type I and | II might, for example, indicate high juvenile mortality and /--selection (Surlyk 1974; Alexander 'j 1977). Size-frequency distributions which are strongly negatively skewed yielding survivorship | curves of Type III or sigmoidal in shape might indicate Ai-selection, bet-hedging, or the transport of smaller individuals away from the site of death by physical processes (Stearns 1976; Craig and | Hallam 1963; Boucot 1953). In order for size-frequency distributions and survivorship curves to be 'i more than a function of differential taphonomy between size classes, however, all size classes must CUMMINS £T/1L.: MOLLUSC TAPHONOMY IN TEXAS BAYS 515 be affected at an equal rate. Unfortunately, as Samtleben (1973) found for Mytihis edulis and Shimoyaina (1985) for Umhonium moniliferum, most size-frequency distributions are shaped pri- marily by taphonomic processes. Few distributions from either of our sites could be explained based on the population dynamics of the species. Taphonomic processes consistently offered the better explanation. Therefore, the interpretation of a species’ population dynamics, inferences on r- and A^-selection and the like may not be accurate in many cases. On the other hand, the data suggest that inferences based on certain portions of the size-frequency distribution may be more accurate. The use of adult numbers and biomass are examples. The real problem is the poor preservation of juveniles (Olson 1957). The study of population dynamics, if it is to be successful, must be based primarily on the study of the larger size classes which typically are the adult individuals. Here, good data on a species’ success in the community may be available. In contrast to the bleak prospects for use of the entire size-frequency distribution to assess population dynamics, the data suggest that the size-frequency distribution is an outstanding reposi- tory of data about taphonomic processes. Both species- and site-specific effects may be preserved in the fossil record. To be useful, however, a better understanding of the biologic and taphonomic factors causing site- and species-specific differences is required. Moreover, the process by which time averaging cumulates individual cohorts into an overall size frequency is poorly understood. Does the final product represent a long-term average or is it dominated by the infrequent recruit- ment and mortality of unusually large cohorts? Nevertheless, all evidence points to the ultimate value of the size-frequency distribution as an indicator of taphonomic processes because it is more or less independent of other lines of evidence typically used. In effect, the initial belief in the 1950s of the value of the size-frequency distribution for assessing transportation would appear to be closer to the primary data provided by it than its more recent use to assess a species’ population dynamics. Ackfwwledgenients. We wish to thank A. Logan, D. Davies, M. Harris, and D. Miller for the assistance provided in the sampling effort. We thank Dr T. Bright for the use of his coastal residence during the sampling periods, and C. Lackey for preparing the tables and typing the manuscript. Mr J. Barrack and an anonymous reviewer provided helpful suggestions that improved the manuscript. Research funds were provided through National Science Foundation grants no. EAR-8021 164 and EAR-8302339 to R. S. and E. P. and a Sea Grant marine fellowship to H. C. REEERENCES ABRAMSON, N. J. 1971. Computer programs for fish stock assessment. FAO Fisheries Tech. Paper No. 101. ALEXANDER, R. 1977. Gi'owth moiphology and ecology of Paleozoic and Mesozoic opportunistic species of brachiopods from Idaho-Utah. 7. Paleonlol. S\, 1133-1149. ANDREWS, J. 1977. Shells and shores of Texas, 365 pp. University of Texas Press, Austin, Texas. BEUKEMA, J. J. 1973. Migration and secondary spatfall of Macoma balthica (L.) in the western part of the Wadden Sea. Nedi. J. Zool. 23, 356-357. BOUCOT, A. J. 1953. Life and death assemblages among fossils. Am. J. Sci. 251, 25-40. BRACE, w. and demar, r. 1958. Distribution of brachiopod and pelecypod shells by currents. J. Sediment. Petrol. 28, 321-332. broadhurst, f. m. 1964. Some aspects of the paleoecology of non-marine faunas and rates of sedimentation in the Lancashire Coal Measures. Am. J. Sci. 262, 858-869. BROOKFIELD, M. E. 1973. The life and death of Torquirhynchia inconstans (Brachiopoda, Upper Jurassic) in England. Palaeogeogr. Palaeoclimatol. Palaeoecol. 13, 241-259. BROSSEAU, D. J. 1978. Population dynamics of the soft shell clam Mya arenaria. Mar. Biol. (Berl.), 50, 63 71. BROWN, L. F. JR., BREWTON, J. L., MCGOWEN, J. H., EVANS, T. J., FISHER, w. L. and GROAT, c. G. 1976. Environmental geologic atlas of the Texas coastal zone— Corpus Christi area, 123 pp. Bureau of Economic Geology, University of Texas at Austin, Austin, Texas. CADEE, G. c. 1982u. Eow juvenile mortality in fossil brachiopods, some comments. Publ. Versl. preth. Inst. Sea Res. 1982-1983, 1-29. 516 PALAEONTOLOGY, VOLUME 29 CADEE, G. c. 1982/;. Population dynamics in brachiopods. Lelliaia, 15, 65-66. CALNAN, T. R. 1980. Molluscan distribution in Copano Bay, Texas, 76 pp. Bureau of Economic Geology, University of Texas at Austin, Austin, Texas. CHAVE, K. E. 1964. Skeletal durability and preservation. In imbrie, j. and Newell, w., (eds.). Approaches to Paleoecology, 377-387. John Wiley and Sons, Inc., New York. COE, w. R. and FITCH, j. E. 1950. Population studies, local growth rates and reproduction of the pismo clam {Tivela stidtorwn). J. Mar. Res. 9, 188-210. CRAIG, G. Y. and HALLAM, A. 1963. Size frequency and growth-ring analysis of Mytilus edulis and Cardium edide, and their palaeoecological significance. Palaeontology, 6, 731-750. and OERTEL, G. 1966. Deterministic models of living and fossil populations of animals. Quart. J. geol. Soc. Land. 122, 315-355. CUMMINS, H., POWELL, E. N., NEWTON, H. J., STANTON, R. J., JR. and STAFF, G. 1986fl. Asscssing transportation by the covariance of species with comments on contagious and random distributions. Lethaia, 19, 1 -22. STANTON. R. J. JR., and STAFF, G. 1986/;. The rate of taphonomic loss in modern benthic habitats: how much of the potentially preservable community is preserved? Palaeogeogr. Palaeoclimatol. Palaeoecol. 52,291-320. CURRAY, J. R. 1960. Sediments and history of holocene transgression, continental shelf, northwest Gulf of Mexico. In shepard, f. p., phleger, f. b. and van andel, t. h. (eds.). Recent sediments, northwest Gulf of Mexico, 221-266. The American Association of Petroleum Geologists, Tulsa, Oklahoma. DARE, p. J. 1976. Settlement, growth, and production of the mussel Mytilus edulis L., in Morecambe Bay, England. Fish. Invest. Ser. 2, 28, 1-25. DEEVEY, E. s. 1947. Life tables for natural populations of animals. Quart. Rev. Biol. 22, 283-314. DODD, J. R., ALEXANDER, R. R. and STANTON, R. J., JR. 1985. Population dynamics in Dendraster, Merria- master, and Anadara from the Neogene of the Kettleman Hills, California. Palaeogeogr. Pcdaeoclimatol. Palaeoecol. 52, 61-76. DOHERTY, p. J. 1979. A demographic study of a subtidal population of the New Zealand articulate brachiopod Terehratella inconspicua. Mar. Biol. 52, 331-342. FENCHEL, T. 1975. Character displacement and coexistence in mudsnails (Hydrobiidae) Oecologia, 20, 19-32. FLESSA, K. w. and BROWN, T. J. 1983. Selective solution of macro-invertebrate calcareous hard parts: a laboratory study. Lethaia, 16, 193-205. FRASER, T. H. 1967. Contributions to the biology of Tagelus divisus (Tellinacea; Pelecypoda) in Biscayne Bay, Elorida. Bull. Mar. Scl. 17, 111-132. HALLAM, A. 1961. Brachiopod life assemblages from the marlstone rock-bed of Leicestershire. Palaeontology, 4, 653-659. 1967. The interpretation of size frequency distributions in molluscan death assemblages. Ibid. 10, 25-42. 1972. Models involving population dynamics. In schopf, t. j. m. (ed.). Models in paleobiology, 62-80. Freeman, Cooper and Company, San Francisco, California. HEDGEPETH, J. w. 1947. The Laguna Madre of Texas. Twelfth North American Wildlife Conference Trans, pp. 364-380. HOFFMAN, A. 1976. Mortality patterns of some gastropods from the Badenian (Miocene) Korytnica Clays, Poland. Neues Jarhb. Geol. Paleontol Abh. 152, 293-306. HOLLAND, A. F. and DEAN, J. M. 1977. The biology of the stout razor clam Tagelus plebeius: II. Some aspects of the population dynamics. Chesapeake Sci. 18, 188-196. HUGHES, R. N. 1980. Population dynamics, growth, and reproductive rates of Littorina nigrolineata Gray from a moderately sheltered locality in north Wales. //. Exp. Mar. Biol. Ecol. 44, 21 1-228. JACKSON, J. B. 1968. Bivalves: spatial and size-frequency distribution of two intertidal species. Science, 161, 479-480. KURTEN, B. 1964. Population structure in paleoecology. In imbrie, j. and Newell, n. (eds.). Approaches to paleoecology, 91-106. John Wiley and Sons, Inc., New York. LADD, H. s., HEDGEPETH, J. w. and POST, R. 1957. Environments and facies of existing bays on the central Texas coast. Geol. Soc. Am. Mem. 67, 599-640. LEViNTON, J. s. and bambach, r. k. 1970. Some ecological aspects of bivalve mortality patterns. Am. J. Sci. 268, 97-112. MCGOWEN, J. H., PROCTOR, C. V., JR., BROWN, L. F., JR., EVANS, T. J., FISHER, W. L. and GROAT, C. G. 1976. Environmental geologic atlas of the Texas coast zone— -Port Lavaca area, 107 pp. Bureau of Economic Geology, University of Texas at Austin, Austin, Texas. CUMMINS £T/1L.: MOLLUSC TAPHONOMY IN TEXAS BAYS 517 MANCiNi, E. A. 1978. Origin of the Grayson micromorph fauna (Upper Cretaceous) of north-central Texas. J. Paleontol. 52, 1294-1314. MOORE, H. B. and lopez, n. n. 1969. The ecology of Chione cancellata. Bull. Mar. Sci. 19, 131 -148. MUNDRY, E. 1972. On the resolution of mixed frequency distributions into normal components. Math. Geol. 4, 55-60. Muus, K. 1973. Settling, growth and mortality of young bivalves in the Oresund. Ophelia, 12, 79-1 16. ODE, H. 1975-1979. Distribution records of the marine Mollusca in the northwest Gulf of Mexico. Texas Conchologist, 11, 84-98; 12, 10-52, 40-56, 79-94, 108-124; 15,45-56. OLSON, E. c. 1957. Size frequency distributions in samples of extinct organisms. J. Geol. 65, 309-333. PARKER, R. H. 1959. Macro-invertebrate assemblages of central Texas coastal bays and Laguna Madre. Bull. Am. Assoc. Pet. Geol. 43, 2100-2166. PHILLIPS, D. w. 1981. Life-history features of the marine intertidal limpet Notoacmea scutum (Gastropoda) in central California. Mar. Biol. 64, 95-103. POWELL, E. N. and cummins, h. 1985. Are molluscan maximum life spans determined by long-term cycles in benthic communities? Oecologia, 67, 177-182. STANTON, R. J., JR. and STAFF, G. 1984. Estimation of the size of molluscan larval settlement using the death assemblage. Estuarine Coastal Shelf Sci. 18, 367-384. STANTON, R. J., JR., CUMMINS, H. and STAFF, G. 1982. Temporal fluctuations in bay environments— the death assemblage as a key to the past. In davis, j. r. (ed.). Proceedings of the symposium on recent benthological investigations in Texas and adjacent states, 203-232. Texas Academy of Science, Austin, Texas. — 1985. Estimating biomass and energy flow of molluscs in palaeo-communities. Palaeontologv, 28, 1-34. RICHARDS, R. p. and BAMBACH, R. K. 1975. Population dynamics of some Paleozoic brachiopods and their paleoecological significance. J. Paleontol. 49, 775-798. SAMTLEBEN, c. 1973. Grdssenverteilungen von Populationen, Totengemeinschaften und Klappenan- sammlungen der Muschel Mytilus edulis L. Meyniana, 23, 69-92. SCHMIDT, R. and warme, j. 1969. Population characteristics of Protothaca staminea (Conrad) from Mugu Lagoon, California. Veliger, 12, 193-199. SEAGER, J. R. 1982. Population dynamics of the Antarctic opisthobranch Philline gibba Strebel. J. Exp. Mar. Biol. Ecol. 60, 163-179. SHELDON, R. w. 1965. Fossil communities with multi-modal size frequency distributions. Nature, 206, 1336- 1338. SHiMOYAMA, s. 1985. Size-frequeucy distribution of living populations and dead shell assemblages in a marine intertidal sand snail, Umbonium (Suchiwn) moniliferum (Lamarck), and their palaeoecological significance. Palaeogeogr. Palaeoclimatal. Palaeoecol. 49, 327-353. SINGER, J. K. and ANDERSON, J. B., 1984. Use of total grain size distributions to define bed erosion and transport for poorly sorted sediment undergoing simulated bioerosion. Mar. Geol. 57, 335-359. SONU, c. J. 1972. Bimodal composition and cyclic characteristics of beach sediment in continuously changing profiles. J. Sediment. Petrol. 42, 852-857. STAFF, G., POWELL, E. N., STANTON, R. J., JR. and CUMMINS, H. 1985. Biomass: Is it a useful tool in paleocommunity reconstruction? Lethaia, 18, 209-232. STANTON, R. J., JR., POWELL, E. N. and CUMMINS, H. 1986. Time-averaging, taphonomy and the community concept in paleoecology: a lesson from death assemblages in Texas bays. Geol. Soc. Am. Bull. 97, 428-443. STANTON, R. J., JR., POWELL, E. N. and NELSON, p. c. 1981. The role of carnivorous gastropods in the trophic analysis of a fossil community. Malacologia, 20, 451-469. STEARNS, s. c. 1976. Life-history tactics: a review of ideas. Quart. Rev. Biol. 51, 3-47. STEWART, I. R. 1981. Population structure of articulate brachiopod species from soft and hard substrates. N. Z. J. Zool. 8, 197-207. SURLYK, F. 1974. Life habit, feeding mechanism and population structure of the Cretaceous brachiopod genus Aemula. Paleogeogr. Palaeoclimatol. Palaeoecol. 15, 185-203. THAYER, c. w. 1977. Recruitment, growth, and mortality of a living articulate brachiopod, with implications for the interpretation of survivorship curves. Paleobiology, 3, 98-109. TREWiN, N. H. 1973. Sorting of bivalve populations in a beach environment. Geol. J. 8, 307-316. VERMEij, G. J. 1980. Drilling predation in a population of the edible bivalve Anadara granosa (Arcidae). Nautilus, 94, 123-125. 518 PALAEONTOLOGY, VOLUME 29 VERMEij, G. J., ziPSER, E. and DUDLEY, E. c. 1980. Predation in time and space: peeling and drilling in terebrid gastropods. Paleobiology, 6, 352-364. WERNER, B. 1956. Uber die Winterwanderung von Arenicola marina L. (Polychaeta Sedentaria). Helgo. Wiss. Meeresimters. 5, 353-378. YAMADA, s. B. 1982. Growth and longevity of the mud snail Batillaria attramentaria. Mar. Biol. 67, 187-192. H. CUMMINS* and e. n. powell Department of Oceanography Texas A & M University College Station, TX 77843 USA R. J. STANTON, JR. and G. STAFF Department of Geology Texas A & M University College Station, TX 77843 USA *Present address Department of Geology Muskinjum College New Concord Typescript received 9 May 1985 Ohio 43762 Revised typescript received 19 November 1985 USA A TETRAPOD TRACKWAY FROM THE CARBONIFEROUS OF NORTHERN CHILE by c. M. BELL and m. j. boyd Abstract. A tetrapod trackway, comprising ten footprints, is described from the Carboniferous rocks of the Chinches Formation, northern Chile. The trackway was apparently made by an amphibian. It is the first fossil referable to a tetrapod to be reported from the Carboniferous of any of the present-day continental areas which once made up the Palaeozoic continent of Gondwanaland. Its discovery strengthens the argument that the southward drift and consequent cooling climate of Carboniferous Gondwanaland need not necessarily have resulted in the complete extinction of the indigenous tetrapod stocks (known from the Devonian rocks of the continent) but may simply have restricted their distribution. We report below the occurrence of a series of footprints made by a tetrapod vertebrate in Carboni- ferous rocks of the Chinches Formation, in the Andes of northern Chile. The location (26° 58' S., 68° 55' W.) is in Quebrada Colorado, 8 km east of Salar de Maricunga and to the north of the road between Copiapo and Tinogasta (text-fig. 1). Tetrapod fossils of Carboniferous age have previously been described only from Europe and North America, and have not hitherto been recorded from any of the present-day continental areas which once comprised the Palaeozoic continent of Gondwanaland. The trackway from Chile is thus of considerable palaeobiogeographical importance and clearly merits description, despite the fact that circumstances have so far prevented either the specimen or a cast of it from being collected. STRATIGRAPHIC AND PALAEOENVIRONMENTAL SETTING The Palaeozoic sedimentary rocks between 26° and 28° S. in northern Chile occur in two geographi- cally distinct areas. Those in the coastal region (Las Tortolas Formation) are deep-sea basin-plain turbidites (Bell 1982), which were deformed in an accretionary wedge produced by north-east directed subduction, possibly during Carboniferous times (Bell 1 984). These sediments are separated from those in the east (Chinches Formation) by a graben containing Mesozoic and Cenozoic sedimentary, volcanic, and plutonic rocks. The strata of the Chinches Formation, which in Que- brada Colorado comprise siltstones with minor proportions of mudstone and very fine-grained arkosic arenite, are much less deformed than those in the coastal region and have been subjected to only very low-grade metamorphism. Correlation between the isolated exposures (see text-fig. 1b) is based on stratigraphic and lithologic comparisons. The strata containing the footprints, in Quebrada Colorado east of Salar de Maricunga, are lithologically, sedimentologically, stratigraphically, and structurally similar to the more extensive exposures to the south-west of Salar de Maricunga (including Quebrada Chinches). There is therefore little doubt that they form part of the same succession. The Chinches Formation in Quebrada Colorado and elsewhere is unconformably overlain by rhyodacites and ignimbrites of the Pantanoso Formation and intruded by granitoids. In some places the Pantanoso Formation overlies the granitoids; in others it is intruded by them. The Pantanoso Formation (equivalent to the Choiyoi Formation of western Argentina) is usually designated as Permian to Triassic in age (Mercado 1982; Coira et al. 1982), although Sepulveda and Naranjo (1982) suggested an earlier, Carboniferous, age for at least part of the succession. Their suggestion was based on the unconformably overlying Las Represas Formation, which contains ammonites of [Palaeontology, Vol. 29, Part 3, 1986, pp. 519-526.] 520 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1 . Location of trackway, a, map of northern Chile and adjacent countries. Arrow indicates trackway locality; b, map of area around Salar de Maricunga. Exposures of Chinches Formation stippled. Arrow indicates trackway locality. Lower Permian age. The only absolute ages available for the granitoids intruded into the Chinches Formation are 287 + 4 Ma and 260 + 5 Ma (K-Ar biotite), suggesting an Upper Carboniferous to Lower Permian age for the granitoids themselves and supporting the suggestion of a Carboniferous age for the Chinches Formation. Palaeoniscoid fish scales from the Quebrada Chinches are regarded by Professor D. L. Dineley (Department of Geology, University of Bristol) as being probably early Carboniferous in age (pers. comm, to senior author). However, in view of the K-Ar dates, noted above, for the granitoids intruded into the Chinches Formation, and of the paucity of the biostrati- graphic evidence, we do not think it justifiable at this stage to date the strata containing the tetrapod trackway any more precisely than as Carboniferous. Possible depositional environments for the Chinches Formation include tidal flats (suggested by Sepulveda and Naranjo (1982) mainly on the basis of the presence of stromatolites), lagoonal or lacustrine. Modern mud-dominated tidal flat and lagoonal deposits do exhibit facies similar to those in the Chinches Formation but differ significantly in the presence of channels and bioturbation, and of interstratified alluvial or aeolian deposits in the tidal sequences (Fisk 1959; Thompson 1975; Elliott 1978). A lacustrine origin is favoured for the Chinches Formation, based on: 1. The presence of only a sparse fauna and flora (the former including bivalve molluscs but no purely marine invertebrate taxa) in a shallow-water environment which was probably well- oxygenated (the sand grains have rims of iron oxide and the rare plant remains are preserved only as impressions). 2. The presence of sedimentary structures and lithologies characteristic of lacustrine depositional facies (see Galloway and Hobday 1983; Link and Osborne 1978; Picard and High 1972; Sturm and BELL AND BOYD: CARBONIFEROUS TETRAPOD TRACKWAY FROM CHILE 521 Matter 1978). These include parallel laminated deep-water facies, shallow wave-dominated facies and lake shore facies. 3. The presence of thin stromatolitic and pisolithic horizons in a predominantly fine-grained pelitic sequence (see Picard and High 1972; Sanders 1968; Schafer and Stapf 1978; Surdam and Wolfbauer 1975). DESCRIPTION The terminology employed for the trackway measurements in the following description is that proposed by Peabody (1948). The trackway, as preserved, measures M m in length and comprises ten positive imprints (two are poorly preserved but the remaining eight are shown in text-fig. 2) on a smooth and gently undulating bedding plane of very fine-grained sandstone. The five clearest consecutive footprints are shown in text-fig. 3 and also, with measurements, in text-fig. 4. The width of the trackway at this point is 18-8 cm. The stride is 29-5 cm and the pace angulation 101°. Two differing measurements of pace (17-9 and 20-2 cm) may be made on the basis of the pes imprints present in this, the only section of the trackway sufficiently well preserved for proper analysis, giving a mean value of 19-05 cm. The manus, which like the pes is turned forward so that its mid-line is roughly parallel to that of the trackway itself, appears to bear five digits. The digits are of moderate length and show no obvious signs of webbing. None of the pes imprints certainly shows the presence of more than four digits, but it is probable that five were actually present in the animal responsible for the trackway. TEXT-FIG. 2. Section of trackway, showing eight footprints ( x 0-25). Tape measure graduated in centimetres. 522 PALAEONTOLOGY. VOLUME 29 Although some Palaeozoic tetrapods (such as the temnospondyl and microsaur amphibians) pos- sessed fewer than five digits in the manus, five was the number commonly present in the pes. The appearance of most of the pes imprints in the trackway, in which the visible impressions of the digits are more or less widely separated from one another at their proximal ends, suggests the presence of partial webbing which depressed the substrate so that only the free part of each digit normally left a clear trace. The trackway exhibits no evidence of a tail drag. TEXT-KiG. 3. Detail of trackway, showing five footprints (xO-55 approx.). Tape measure graduated in centimetres. Using the method outlined by Baird (1952, pp. 833-834) for determining the approximate distance between the pectoral and pelvic girdles of a trackway-making amphibian or reptile, a body length (excluding head and tail) of about 26 cm may be estimated for the animal which made the trackway that forms the subject of the present study. However, as pointed out by Baird, the method makes no allowance for the curvature of the vertebral column during locomotion, so the distance calculated is likely to be slightly less than the actual gleno-acetabular distance. That the trackway was made by an amphibian rather than by a reptile is suggested by the absence of claw and epidermal scale impressions and (although less strongly) by the apparent presence of webbing between the digits of the pes. However, more detailed consideration of the relationships of the tetrapod responsible for the Quebrada Colorado trackway (and a more detailed description of the trackway itself) must await collection of the specimen. BELL AND BOYD; CARBONIFEROUS TETRA POD TRACKWAY FROM CHILE 523 ST 29-5 cm TEXT-FIG. 4. Detail of trackway, showing measurements. Dotted lines indicate restored areas. Abbreviations: W width; ST— stride; PA— pace angulation; m manus; p— pes. 524 PALAEONTOLOGY, VOLUME 29 DISCUSSION As noted above, the trackway from Quebrada Colorado represents the first recorded evidence of Carboniferous tetrapods from anywhere outside Europe and North America. During the Devonian and Carboniferous the latter two regions together formed a large part of the continent of Laurasia, which also included Greenland and Russia west of the Ural Mountains. South America, however, was a constituent of the great southern continent of Gondwanaland, as also were Africa, Antarctica, Australia, and India (e.g. Tarling 1980; Johnson 1980). Although no tetrapod fossils have hitherto been described from the Carboniferous of any of the present-day geographical areas which once made up Gondwanaland, amphibians of Devonian age have been reported from Brazil and south- east Australia (Gondwanaland) as well as from East Greenland (Laurasia). The Brazilian material consists of only a single footprint, described by Leonardi (1983) as Notopus petri. It is, however, very well preserved and is apparently of late Givetian or early Frasnian age; Notopus may thus be the earliest known tetrapod. The described Australian specimens comprise three trackways from the Frasnian of Victoria (Warren and Wakefield 1972) and an isolated lower jaw (Metaxygmithus) from the late Frasnian or early Famennian of New South Wales (Campbell and Bell 1977). The ichthyostegalian amphibians from the late Famennian of East Greenland are, however, represented by relatively abundant material, and three genera (Ichrliyostega, Ichthyostegopsis, and Acanthostega) have been described (Save-Soderbergh 1932; Jarvik 1952). Ichthyostegalian affinities have been suggested for the Australian Devonian amphibians and also, although with less certainty, for that from Brazil. The evidence at present available is not adequate to determine with any sureness whether the area of origin of tetrapods lay in Devonian Gondwanaland or in Devonian Laurasia (no tetrapod fossils at all have been described from the Devonian or Carboniferous of Angaraland, the third major Upper Palaeozoic continent (Johnson 1981)). However, the fact that Notopus may be of late Givetian (Middle Devonian) age and hence significantly older than Metaxygmithus and the foot- prints from Victoria, which are in turn older than the Greenland ichthyostegalians, does lend support to the suggestion made by Panchen (1977) and Janvier (1978) that the first tetrapods arose in Gondwanaland. Migration of terrestrial and freshwater tetrapods from Gondwanaland to Laurasia may well have been possible in the Upper Devonian. As Johnson (1980, 1981 ) has pointed out, similarities in the Upper Devonian structural history, stratigraphy and freshwater fish faunas of south-east Australia and East Greenland suggest that these two areas were in close proximity at this time. During the Carboniferous, Laurasia appears to have continued the northward drift it had exhi- bited during the Devonian but nonetheless remained largely within the tropics, the tetrapods apparently migrating southward to remain within the tropical belt (Johnson 1981; Panchen 1973, 1977). In contrast, Gondwanaland drifted southward and by early Upper Carboniferous times this drift, and the accompanying rotation of the continent, had brought south-east Australia close enough to the South Pole for the area to suffer sea-level glaciation (Johnson 1981). Johnson (1981) has suggested that the southward drift and cooling climate of Carboniferous Gondwanaland must have compelled the indigenous tetrapods to migrate northward and eventually resulted in their extinction on that continent. This view received support from the absence (until now) of any records of post-Devonian tetrapods from Gondwanaland prior to its early Permian collision with Laurasia, which event allowed colonization from the latter continent. The trackway described in the present paper is obviously of considerable interest as the first reported non-Laurasian Carboniferous tetrapod specimen, and as the latest known specimen which can be regarded with some confidence as representing the original, indigenous, tetrapod fauna of Gondwanaland. However, it inevitably also raises the question of whether the indigenous tetrapods of the latter continent did indeed all become extinct between the Upper Devonian and the end of the Carboniferous, as argued by Johnson. Carroll and Gaskill (1978, p. 196) have suggested that the Carboniferous cooling of Gondwanaland may merely have restricted the distribution of the indigenous tetrapods, and that the lack (at the time they wrote) of records of Carboniferous BELL AND BOYD: CARBONIFEROUS TETRA POD TRACKWAY FROM CHILE 525 tetrapods from the modern areas which once made up the continent probably had more to do with geological and modern cultural patterns than with the ecology and biogeography of Carboniferous times. This argument is clearly strengthened by the discovery of the Quebrada Colorado trackway; the question, however, can finally be resolved only by further field-work in the dispersed fragments of Gondwanaland. Achwwledgements. The discovery of the trackway was made by one of the authors (C. M, B.) in association with Dr M. Suarez of the Servicio Nacional de Geologia y Mineria, Chile. We thank the latter gentleman and the Servicio Nacional de Geologia y Mineria for providing facilities for the field-work. Our thanks are also due to Professor D. L. Dineley (Department of Geology, University of Bristol), Dr L. B. Halstead (Department of Geology, University of Reading), and Dr R. Riding (University of Cardiff) for their kind assistance with the identification of fossils. The work of C. M. B. was supported by a Royal Society Scientific Investigation Grant and by Natural Environment Research Council grant GR3/4962. The maps and photographs in the present paper were prepared by C. M. B., as was the text section dealing with the stratigraphy and palaeoenvironment; the remainder of the paper was the responsibility of M. J. B. REFERENCES BAIRD, D. 1952. Revision of the Pennsylvanian and Permian footprints Linmopus, Allopus and Baropus. J. Paleoui. 26, 832-840. BELL, c. M. 1982. The Lower Palaeozoic metasedimentary basement of the coastal ranges of Chile between 25° 30' and 27° S. Rev. geol. Chile, 17, 21-29. 1984. Deformation produced by the subduction of a Palaeozoic turbidite sequence in northern Chile. J. geol. Soc. Loud. 141, 339-347. CAMPBELL, K. s. w. and BELL, M. w. 1977. A primitive amphibian from the Late Devonian of New South Wales. Alcherifigci, 1, 369-381. CARROLL, R. L. and GASKiLL, p. 1978. The Order Microsauria. Me?u. Amer. phil. Soc. 126, 1-211. coiRA, B., DAVIDSON, J., MPODOZis, c. and RAMOS, V. 1982. Tectonic and magmatic evolution of the Andes of Northern Argentina and Chile. Earth Sci. Rev. 18, 303-332. ELLIOTT, T. 1978. Clastic shorelines. In reading, h. g. (ed.). Sedimentary environments and facies, 143-177. Blackwell, Oxford. FISK, H. N. 1959. Padre Island and the Laguna Madre Flats, coastal South Texas. National Academy of Science- National Research Council, Second Geography Conference, 103-151. GALLOWAY, w. E. and HOBDAY, D. K. 1983. Terrigenous clastic depositional systems, 423 pp. Springer- Verlag. JANVIER, p. 1978. Vertebres devoniens de nouveaux gisements du Moyen-Orient. Ann. Soc. geol. N. 97, 373- 382. JARVIK, E. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Meddr Gronland, 14, 1 -90. JOHNSON, G. A. L. 1980. Carboiiiferous geography and terrestrial migration routes, 39-54. In panchen, a. l. (ed.). The terrestrial environment and the origin of land vertebrates, xii+633 pp. Systematics Association Special Volume No. 15. Academic Press, London. 1981. Geographical evolution from Laurasia to Pangaea. Proc. Yorks, geol. Soc. 43, 221-252. leonardi, g. 1983. Notopus petri nov. gen., nov. sp.: une empreinte d’amphibien du Devonien au Parana (Bresil). Geobios, 16, 233-239. link, m. h. and osborne, r. h. 1978. Lacustrine facies in the Pliocene Ridge Basin Group; Ridge Basin, California. In matter, a. and tucker, m. e. (eds.). Modern and ancient lake sediments, 169 187. Spec. Pubis int. Ass. Sediment. No. 2. Blackwell, Oxford. MERCADO, M. w. 1982. Hoja Laguna de! Negro Francisco, Region de Atacama. Carta Geologica de Chile 1:100,000 No. 56. Servicio Nacional de Geologia y Mineria, Chile. panchen, a. l. 1973. Carboniferous tetrapods. In hallam, a. (ed.). Atlas of Palaeobiogeography, 117-125. Elsevier, Amsterdam. 1977. Geographical and ecological distribution of the earliest tetrapods. In hecht, m. k., goody, p. c. and hecht, b. m. (eds.). Major patterns in vertebrate evolution, 723-738. Plenum, New York. PEABODY, F. E. 1948. Reptile and amphibian trackways from the Lower Triassic Moenkopi Formation of Arizona and Utah. Univ. Calif. Dept. geol. Sci. Bull. 27, 295-468. PICARD, M. D. and high, l. r. 1972. Criteria for recognising lacustrine rocks. In rigby, j. k. and hamblin. 526 PALAEONTOLOGY, VOLUME 29 w. K. (eds.). Recognition of ancient sedimentary environments, 108-145. Society of Economic Paleontologists and Mineralogists Special Publication No. 16. SANDERS, J. E. 1968. Stratigraphy and primary sedimentary structures of fine-grained, well-bedded strata, inferred lake deposits. Upper Triassic, central and southern Connecticut. Geol. Soc. Am. Spec. Pap. 106, 265-305. SAVE-SODERBERGH, G. 1932. Preliminary note on Devonian stegocephalians from East Greenland. Meddr Gr on I and, 94 (7), 1-107. SCHAFER, A. and STAFF, K. R. G. 1978. Permian Saar-Nahe Basin and Recent Lake Constance (Germany): two environments of lacustrine algal carbonates. In matter, a. and tucker, m. e. (eds.). Modern and ancient lake sediments, 83-107. Spec. Pubis int. Ass. Sediment. No. 2. Blackwell, Oxford. SEPULVEDA, p. and NARANJO, J. A. 1982. Hoja Carrera Pinto, Region de Atacama. Carta Geologica de Chile 1:100,000 No. 53. Servicio Nacional de Geologia y Mineria, Chile. STURM. M. and matter, a. 1978. Turbidites and varves in Lake Brienz (Switzerland): deposition of clastic detritus by density currents. In matter, a. and tucker, m. e. (eds.). Modern and ancient lake sediments, 147-168. Spec. Pubis int. Ass. Sediment. No. 2. Blackwell, Oxford. SURDAM, R. c. and WOLFBAUER, c. A. 1975. Green River Formation, Wyoming: a playa-lake complex. Bidl. geol. Soc. Amer. 86, 335-345. tarling, d. h. 1980. Upper Palaeozoic continental distributions based on palaeomagnetic studies, 11-37. In PANCHEN, A. L. (ed.). The terrestrial environment and the origin of land vertebrates, xii-l-633 pp. Systematics Association Special Volume No. 15. Academic Press, London. i HOMP.soN, R. w. 1975. Tidal flat sediments of the Colorado River delta, north-western Gulf of California. In GiNSBURG. R. N. (ed.). Tidal deposits: a casebook of Recent e.xamples and fossil counterparts, 57-65. Springer- Verlag, Berlin. WARREN, J. w. and wakefteld, n. a. 1972. Trackways of tetrapod vertebrates from the Upper Devonian of Victoria, Australia. Nature, Land. 238, 469-470. C. MICHAEL BELL Department of Geography and Geology The College of St Paul and St Mary The Park Cheltenham Gloucestershire GL50 2RH MICHAEL J. BOYD Department of Natural History Kingston upon Hull Museum Queen Victoria Square Kingston upon Hull North Humberside HUl 3DX and Department of Geology University of Reading Reading Typescript received 16 January 1985 Berkshire Revised typescript received 12 December 1985 RG6 2AB SILURIAN ENCRINURID TRILOBITES EROM GOTLAND AND DALARNA, SWEDEN by LARS RAMSKOLD Abstract. Trilobites of the family Encrinuridae are described from the Silurian of Gotland and Dalarna, Sweden. Twelve species (nine named, of which four are new) are assigned to two genera. Three subgenera of Encrimims are recognized, including E. (Aiistralunis) subgen. nov. and E. {Nmiewus) subgen. nov. BalEoma is redefined by excluding species here included in E. (Niic/etiriis). A directional, morphological trend is shown between several populations of E. (Encrinums) nuicrowus Schmidt, 1859. E. {E.) schmu/ti Mannil, 1968 is a junior subjective synonym of E. (E.) schisticola Tornquist, 1884. Balizonui ohtiisus (Angelin, 1851) may be a widespread species, possibly occurring also in Britain, Estonia, Czechoslovakia, Podolia, and Canada. New species described are E. (E.) iutersitiis, E. (E.) Jcirkam/eri, E. (E.) misiitus, and E. (£.) odvaldensis. One of the earliest records of a trilobite from Sweden is Linnaeus’ ( 1759) figure and description of a pygidium of Encrinums, in all probability from Gotland, that he included under the general name Entomolilhus paradoxus. Nearly sixty years later Wahlenberg (1818) erected the species Entomostracites punctatus, which has since become one of the most commonly cited Silurian trilobite species. Other Swedish Silurian encrinurids were described subsequently by Angelin (1851), Schmidt (1859), and Tornquist (1884). More recently Tripp (1962) revised some species of Encrinums from Gotland. Encrinurids in general have received considerable attention lately, with several important works on taxonomy, morphology, and phylogeny (Evitt and Tripp 1977; Temple and Tripp 1979; Strusz 1980). This paper is a study of Silurian encrinurids from Gotland and Dalarna. Two additional species from Gotland are being described by C. Magnus (Bergen, Norway), and a further species from the Llandovery of Ostergotland is currently being studied (Ramskold and Bassett, in prep.). During the course of this work it became necessary to study all descriptions of Silurian encrinurids, and to examine the major British and Australian collections of encrinurids. This has led to the recognition of three subgenera of Encrinums and a re-evaluation of Balizonui, with bearings on other encrinurine genera. The stratigraphical scheme used for the Gotland sequence (text-fig. 1) is that of Hede (1921, 1925); subunits are those of Laufeld (1974u). Hemse Marl, northwestern part, and Hemse Marl, southeastern part are abbreviated Hemse Marl NW and Hemse Marl SE, respectively. It must be emphasized that the stratigraphical distribution of the species as plotted in text-fig. 1 does not necessarily reflect the chronological succession of the taxa, due to diachroneity of most units. The stratigraphical terms for Dalarna follow Waern (1960). NOTES ON MORPHOLOGY General terminology. The terminology used here mainly follows the Treatise on Invertebrate Paleon- tology, and Evitt and Tripp (1977). The terms ‘4L’ and ‘PL’ were defined by Howells (1982) and ‘eye socle’ by Shaw and Ormiston (1964). Following Tripp et al. (1977) the term ‘tubercle’ is used in its conventional general sense rather than sensu Miller (1976) since a precise distinction requires thin sections. ‘Rachis’ and ‘rachial furrow’ are preferred to the slightly misleading ‘axis’ and ‘axial furrow’, as is ‘labral plate’ to the anatomically incorrect ‘hypostome’. Use of the terms ‘anterome- dian depression’, ‘sagittal band’, and ‘sagittal groove’ follow Strusz (1980). Temple and Tripp (1979) IPalaeontology, Vol. 29, Part 3, 1986, pp. 527 575, pi. 37-49.) 528 PALAEONTOLOGY. VOLUME 29 SUNDRE HAMRA BURGSVIK Q 3 O o LU EKE Upper Lower Marl, top e Marl, SE part undifferentiated U HEMSE d c Marl, NW part b a KLINTEBERG Marl f e d c b undifferentiated a Upper MULDE undifferentiated Lower HALLA SLITE Siltstone P. gotlandicus beds 9 Marl, undifferentiated f Marl, NW part e d o E o < s E ^ o o LL C ** . <0 Q. o 5 « 3 c 2 C5.-C 2 O 3 Cl) c C 2 iid ■5 uj uj <0 c 0) 5 m -Q . o « 3 <0 V. 3 3 3 -S ^ C O Uj C ^ . UJ u C Ui Uj T Uj Uj c b a c u c Uj TOFTA SW facies Uj d HOGKLINT c b a U VISBY ■? o 3 V. 3 O k. o 3 E v> 3 k. 3 3 C OQ LL L VISBY TEXT-FIG. I. Occurrence of Encrinuridae in Gotland. Solid squares represent specimens assigned definitely to a taxon and open squares represent compared forms. A square with a question mark indicates that the horizon is uncertain. The stratigraphical column is a practical way of illustrating the distribution of the species within the mapped units, but is not necessarily a reflection of the chronological appearance and disappearance of various taxa. Diagram modified from Laufeld (I974u, p. 124). RAMSKOLD: SILURIAN ENC R I N U RI D TR 1 LO B 1 TES 529 TEXT-HG. 2. The circumocular tubercles CT1-CT4. Tubercle CT3 is sometimes double, or displaced posterolaterally, and has not been used in the measurements (Table 1). Drawing based mainly on Encri- mirus (Encrinurus) punclatus Form A (Ar51797). redefined several other terms as a basis for their attribute list, and these are adopted here. The mucro is taken as beginning where the curved slope posterior to the end of rachis meets the flat (in lateral view) surface of the most posterior pygidial part. Certain important features of encrinurid morphology are discussed in some detail below. 1. Glabellar tubercle formula. A row of six tubercle-pairs along the glabellar mid-line is already present in the Middle Ordovician Encrimtrokles neuter and E. imcatus described by Evitt and Tripp (1977). The gross morphology of these species suggests a position close to the ancestry of Encrinurus s.l. This basal tubercle pattern is retained in E. (Encrinurus) and E. (Nucleurus), and is present, though less clear, in the later offshoots Balizoma and Erammia. The pattern is obliterated or lost in E. (Australurus). In Encrinurus much taxonomic emphasis has been placed on the arrangement of the glabellar tubercles, as expressed in the system elaborated by Tripp (1957, 1962) and further refined by Strusz (1980). The Gotland material (and British; P. D. Lane, pers. comm.) shows that the detailed pattern is different at almost every locality, and that most of these small differences cannot be of importance at levels above populations or subspecies. For this reason complete tubercle formulae are of limited use and are not included in the diagnoses of the species described here, although particular tubercles may be of diagnostic importance. 2. Circumocular tubercles. In E. (E.) punctatus and allied species the eye socle is surrounded by a more or less distinct ring of prominent tubercles. On the fixed cheek four major tubercles, here called CT1-CT4 (see text-fig. 2), form part of the circle, usually together with several additional minor tubercles. CTl is the torular tubercle and CT2 the postocular tubercle of Evitt and Tripp (1977). The relative positions of CT1-CT4 and the distinctness of the part of the circle on the free cheek are useful diagnostic features. The ring itself probably had little functional significance, being partly at least an expression of density of general tuberculation rather than of a particular function of the tubercles surrounding the eye. Individual tubercles, however, may have had a specialized function (cf. torular tubercle in Evitt and Tripp, 1977). 3. Main tubercle row on free cheek border. In all species of E. (Encrinurus) there is a row of prominent tubercles on the free cheek border along the lateral border furrow. Anteriorly there is a commonly smaller tubercle further from the lateral border furrow, close to the anterior furrow. This is followed by a row of six large tubercles close to the border furrow, and a posterior, smaller tubercle set further from the border furrow. Posterior to this tubercle, and closer to the border furrow, there begins a row of four to six tubercles along the sutural edge of the free cheek border; these tubercles are progressively smaller posteriorly. This tubercle arrangement is consistent throughout the subgenus, but may be subdued and almost lost, as in E. (E.) stubblefieldi. A similar, but less regular tubercle pattern is already present in Enerinuroides neuter and E. uncatus, and also in Balizoma. In Encrinurus (Encrinurus) the expression of the main tubercle row is a useful diagnostic criterion. 4. RjP ratio in the pygidium. R = number of rachial rings anterior to the intersection (as defined by Temple and Tripp 1979, fig. 3) of the last, not distally converging, pair of pleural furrows and the rachial furrows. P = number of pleural ribs anterior to the same pair of pleural furrows. The 530 PALAEONTOLOGY. VOLUME 29 Q O' Q. o I Q !D >- > o Q z < concomitans FRA MM I A ^ testosteron arctica aff. hyperboreus of Thomas rossica E. sp. F of Lane ■ cf. arcticus of Bolton hyperboreus erraticus obtusus Form C obtusus Form A Eke Beds obtusus Form A Hemse Beds obtusus Form B rosensteinae FRAGISCUTUM glebalis rhytium variolaris E. sp. of Tripp et al. B. sp. of Holloway dakon dimitrovi indianensis PER R YUS globosus severnensis mareki inexpectatus rumbaensis mullochensis diabolus abyssalis selistensis kiltsiensis rotundus OO r- 3) 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 R/p TEXT-FIG. 3. R/P ratios for certain encrinurids. The species are arranged approximately chronologically, but there are probably several errors due to poor data. Each supraspecihc taxon has a limited range of the R/P ratio and a restricted vertical distribution. Authorships of the species are given in the discussions of Balizoma and Emrinurus (Nucleurus). R/P ratio defines the spacing of the rachial rings relative to the pleural ribs. The ratio eliminates the problem of making accurate counts of the often indistinct posteriormost rings. R/P ratios are very stable in encrinurid genera and may therefore aid in supraspecific grouping. They are easy to calculate, not influenced by distortion, identical on internal and external moulds, and confined to the most commonly preserved part of the exoskeleton. An R/P ratio analysis for particular encrinu- rid species is given in text-fig. 3. 5. Terminal rib notation. A numbering system is used here for the terminal pleural ribs. Counting from the anterior rib, the number of the last rib (AO is given, and a coefficient indicates whether this rib is paired (N~) or single (N^\ it may be ridge- or knob-shaped). The number of pleural ribs is not dependent significantly on pygidial size (Howells 1982, text-fig. 4; text-fig. 8 herein), although there may sometimes be a tendency for more ribs in large specimens. The system can be used for all encrinurids, but is best suited for non-mucronate forms. 6. Axial tubercles are a conspicuous feature in many encrinurid pygidia and were mentioned and figured as early as 1759 by Linnaeus. The positions of such tubercles were studied by Norton (1895), and Rosenstein’s (1941) detailed work showed the strict regularity of the tubercle distribution in Estonian material. In Best’s (1961) comprehensive study of a large collection, tubercle distribution RAMSKOLD: SILURIAN ENCRINURID TRILOBITES 531 patterns were interpreted as indicating a response to changing environmental conditions, though Perry and Chatterton (1979) preferred a hypothesis of immigrating races. It is now common to include a modal distribution pattern for these tubercles in descriptions when possible (Campbell 1967; Mannil 1978; Snajdr 1978; Perry and Chatterton 1979). More extensive information has sometimes been added (Mannil 1968; Tripp el al. 1977; Howells 1982) and a detailed notation system has been elaborated (Strusz 1980). From these studies it is clear that the number of blank rachial rings between those bearing tubercles is fairly constant in each species, and that the entire ‘set’ of tubercles often has two or more alternative positions along the sagittal line. As a somewhat simplified comment, Balizoma has a tubercle on every second ring, Fragiscutum and E. {Nucleurus) on every third, and E. (Encrinurus) on every fourth. This tubercle distribution obviously correlates strongly with the number of rachial rings and, in turn, with the R/P ratio. A pygidium with numerous, short (sag.) rings ( = high R/P ratio) will have several blank rings between the tubercle- bearing rings to allow an inter-tubercle distance comparable to that of a pygidium with few, long rings ( = low R/P ratio), where less than every second ring is blank. This correlation between distance and position of the tubercles seems to indicate a functional significance of the tubercle arrangement. However, it also means that the taxonomic value of the modal distribution may be limited, since the arrangement reflects dependence on these functional requirements (i.e. a certain distance between the tubercles) rather than an obligate interrelationship between a tubercle and a particular ring. It is likely that the tubercle distribution, as expressed in ring positions, is controlled to a large extent by the number of rachial rings (which depends on the sagittal length of the rings), and that this latter feature is of more primary taxonomic importance. SYSTEMATIC PALAEONTOLOGY The specimens are housed in Naturhistoriska Riksmuseet, Stockholm (prefixed Ar), in the Type Collection of the Geological Survey of Sweden (SGU), and in the Palaeontological Institute, University of Lund (LO). All specimens, except in SEM photos, were painted with matt black opaque and coated lightly with ammonium chloride prior to photography. Terms for the orien- tations used during photography follow Whittington and Evitt (1954, p. 1 1) Most specimens with numbers beginning Ar51 and Ar52 were collected by the author; other collectors are indicated in the plate explanations when known. Eamily ENCRINURIDAE Angelin, 1854 Diagnosis. See Strusz 1980, p. 7. Subfamily encrinurinae Angelin, 1854 Diagnosis. See Strusz 1980, p. 7. Discussion. The origin of the encrinurids discussed in this study is somewhere near the American Trentonian Encrinuroides neuter and E. uncatus described by Evitt and Tripp (1977). No certain descendant species are as yet known from the uppermost Ordovician. This apparently small group underwent prolific radiation in the lower and middle Llandovery as three main lineages (the variolaris, punctatus, and niitchelli plexa of Strusz 1980). These lineages are regarded as constituting the genus Encrinurus, and are here given subgeneric status. Two of these subgenera, E. (Encrinurus) ( = ''punctatus') and E. (Australurus) ( = dnitcliel/i'), remained relatively stable through the Silurian while one, E. (Nucleurus) ( = 'variolaris'), proliferated further and gave rise to Balizoma, Fragiscu- tum, and Frammia. Gcuux ENCRINURUS Emmrich, 1844 Type species. Entomostracites punctatus Wahlenberg, 1818 (ICZN Opinion 537, 1959, emended 1977), from the Hdgklint and Slite Beds (Wenlock) of Gotland. Sweden. 532 PALAEONTOLOGY, VOLUME 29 Other species. A list of assigned species is given under each of the three subgenera recognized here. A few additional species apparently belong to Encrinums, but are too poorly known to be assigned to any particular subgenus. These are: E. calgach Lamont, 1948; E. creher Maksimova, 1962; E. donenjalensis Balashova, 19686,' E. dris (Lamont, 1948); E. henshawensis (Lamont, 1978); E. laosensis Patte, 1929; E. marianne Maksimova, 1975; E. penkillensis (Lamont, 1978); E. ploeckensis von Gaertner, 1930a; E. shelvensis Whittard, 1938; and possibly E. nodai Kobayashi and Hamada 1974. A distinct group is formed by three late Llandovery species; the American E. americanus Vogdes, 1886, and the two Estonian E. pilisiverensis Rosenstein, 1941 and E. quinquecostatus Mannil, 1958. Only the pygidia are known of these species, but they are so different from other encrinurids that they will certainly form a new genus when better known. A few species included by Strusz (1980) in Encrhmms are regarded here as being outside the range of this genus (though reassignments are outside the scope of this study): Encrumroides meijiangensis Chang, 1974; E. meitanensis Chang, 1974; Encrimirus melzensis Krueger, 1971; E. rialpensis von Gaertner, 19306,' E. simplicicu- lus Talent, 1965. E.7 newlandensis (Lamont, 1978) of Howells (1982) does not belong to Encrinurus. Diagnosis. Glabella with regular and more or less coarse tuberculation and short (tr.) lateral glabellar furrows not deeply impressed across glabella. Preglabellar furrow faint to obscure axially. 2L-4L and PL tuberculiform, IL may be considerably reduced. Glabellar lobes and tubercles on fixed cheek along rachial furrow equal in number, set in alternating or opposing positions. Anterior cranidial border widening abaxially, with one row of tubercles. Eyes pedunculate or stalked, op- posite 2L to 3L. Rhynchos ends only slightly posterior to labral suture or protrudes beyond this. Pygidium subtriangular, at least eight pleurae and fourteen rachial rings; axial tubercles and sagittal band or groove present, may be faint. Remarks. Strusz’s (1980) diagnosis has been slightly amended, mainly to adjust for the exclusion of species since included in Balizoma. Subgenus encrinurus (encrinurus) Emmrich, 1844 Type species. As for genus. Other species. E. baiticus Mannil, 1978; E. caplanensis Northrop, 1939; E. confusevarus Howells, 1982; E. deomenos Tripp, 1962; E. egani Miller, 1880; E. expansus Haswell, 1865; E. hagshawensis Lamont, 1965; E. intersitus sp. nov.; E. jarkanderi sp. nov.; Cryptonymus laevis Angelin, 1851; E. macroiirus Schmidt, 1859; E. nasutus sp. nov.; E. odvaldensis sp. nov.; E. oimiensis Whittard, 1938; E. ornatus Hall and Whitfield, 1875; Zetluis pagei Haswell, 1865; E. reflexus Raymond, 1916; E. ruhnuensis Mannil, 1978; E. schisticola Tornquist, 1884 (including the junior subjective synonym E. schmidti Mannil, 1968); E. stateratus Howells, 1982; E. .•stiibbiefieldi Jx'xpp, 1962; E. triangidus Mannil, 1977; Asaphus tubercidatus Buckland, 1836. Two more species, E. brevispinosus Haas, 1968 and E. squarrosus Howells, 1982, differ in detail from the above species, and can only be assigned questionably to the nominate subgenus. Diagnosis. Glabella relatively long and narrow. Anterior border of cranidium with eight to fourteen tubercles, usually eight to ten. Tubercles on fixed cheek along rachial furrow equal in size to EXPLANATION OF PLATE 37 Figs. 1-9. Encrinurus {Encrinurus) punctatus (Wahlenberg, 1818), Form A. Hogklint Beds, unit a (1, 5, 7, 8), unit b (2-4), Slite Beds, Slite Marl (6, 9). Ireviken 1(1,5), Kopparsvik (8), Lickershamn (3, 7), Valve 2 (6, 9), Vattenfallsprofilen 1, 19'90-20-25 m a.s.l. (2, 4). lu-c, Ar51784, anterior, dorsal, and oblique lateral views of cranidium, x 3. 2, SGU 5045, exterior view of free cheek, x4 (coll. G. Liljevall 1908). 3u, 6, Ar51797, views normal and parallel to length axis of eye of incomplete cephalon, note circumocular tubercles, x 3 (coll. Amelang 1978). 4«, 6, Ar49774, dorsal and anterior views of enrolled specimen, x 3 (coll. V. Jaanusson 1975), 5a, 6, Ar5 1785, lateral and exterior views of labral plate, x 6. 6, Ar52293, dorsal view of incomplete cranidium, x 2. 7, Ar52288, dorsal view of cranidium with part of thorax, x 2 (coll. U. Samuelson 1984). 8a, 6, Ar52287, dorsal and posterior views of cranidium, x2. 9, Ar52294, exterior view of free cheek, x 3. PLATE 37 RAMSKOLD, Encrinurus {Encrinurus) pimctatus 534 PALAEONTOLOGY, VOLUME 29 Other fixed cheek tubercles. IL small, ridge-shaped, set well below 2L, occasionally tuberculiform abaxially. Genal angle usually with long spine, sometimes short, rarely almost lost in adults. Labral plate with wide rhynchos, level with or protruding in front of labral suture, posterior border 19- 23 % of labral length. Thorax commonly with axial spine on tenth segment. Pygidium fairly long and narrow, mucronate; mucro formed by eighth and following pleural segments. Usually 8^-9' pleural pairs, posterior pleurae not merging into ‘loop’. Remarks. The great majority of species referred to this subgenus are characterized by their spinosity (genal and thoracic spine and mucro). The earliest species (?lower-middle Llandovery) are less spinose and show considerable similarities to E. (Nucleiirus) subgen. nov., probably indicating that these two lineages split near the Ordovician/Silurian boundary. The number of free pleurae in the pygidium was already fixed by middle Llandovery times, being seven in all assigned species. E. squarrosus from the lower Llandovery of Scotland is assigned only questionably to E. (Encrinurus) since it, together with other differences, has a variable number of free pleurae and practically no mucro; this may indicate the condition before the number of free pleurae was fixed. A small group of upper Llandovery-lower Wenlock species (E. laevis, E. triangulus, and an undescribed Gotland species) differ from the main line by the reduced genal spines, subdued tuberculation on the free cheek, low eyes, and low R/P ratio ( < 2 0 as opposed to 2-3-3 0 for most E {E.) species). They are not, however, sufficiently different to form a separate subgenus as presently known. Encrinurus (Encrinurus) punctatus (Wahlenberg, 1818) Plate 37, figs. 1-9; Plate 38, figs. 1-15 v* 1818 Entomostracites punctatus Wahlenberg, p. 32, pi. 2, fig. 1* [non fig. 1]. v. 1827 Calymene pimctata\ Dalman, p. 233 (48), pi. 2, fig. 2a, h. V. 1837 Calymene punctata', Hisinger, p. 12, pi. 1, fig. 9 [copy Dalman 1827, pi. 2, fig. 2a\ ? 1847 Encrinurus punctatus Wahlenb.; Hawle and Corda, p. 91, pi. 5, fig. 55 [Gotland specimen]. ? 1851 Cryptonymus punctatus Wahl.; Angelin, p. 3, pi. 4, figs. 4-8. V. 1901 Encrinurus punctatus Wahlenberg; Lindstrom, p. 56, pi. 4, figs. 4-9, 12, 13. .1941 Encrinurus punctatus (Wahlenberg) 1821; Rosenstein, p. 53, pi. 1, figs. 1-11 [non pi. 2, fig. 4 = E. (E.) inacrourus Schmidt, 1859]. [With synonymy list.] ? 1954 Encrinurus punctatus (Wahlenberg) 1821; Balashova, p. 41, pi. 24, figs. 7-9. V. 1956 Encrinurus punctatus (Wahlenberg, 1821 ); Tripp and Whittard, p. 259, pi. 3, figs. 1 and 2. V. 1962 Encrinurus punctatus (Wahlenberg); Tripp, p. 461, pi. 65, figs. 9-1 1, 13, 14; pi. 66, figs. 2 and 3; pi. 67, figs. 5-8; pi. 68, figs. 7, ?8, 10. V. 1962 Encrinurus inacrourus Schmidt; Tripp, p. 469, pi. 65, fig. 2 only. V. 1962 Encrinurus tuberculatus (Buckland); Tripp, p. 467, pi. 65, figs. 5-8; pi. 66, figs. 4-6, 8-11 [non fig. 7]; pi. 67, figs. ?9 and ?10; pi. 68, fig. 4 [non figs. 5 and 6]. [With synonymy list.] EXPLANATION OF PLATE 38 Figs. 1-5. Encrinurus (Encrinurus) punctatus (Wahlenberg, 1818), Form A. All Fldgklint Beds, unit a. Halls Huk 1 (2), Ireviken 1 (3-5), Kopparsvik (1). 1, Ar52285, dorsal view of pygidium, x 6. 2a, b, Ar52289, posterior and dorsal views of small pygidium, x 11. 3, Ar52283, dorsal view of strongly tuberculated pygidium, x 4. 4. Ar52284, dorsal view of pygidium, x 4. 5a, b, Ar52282, dorsal and lateral views of pygidium, x 4. Figs. 6-15. E. (E.) punctatus (Wahlenberg, 1818), Form C. All Slite Beds, Slite Marl. Follingbo 7 (14), Follingbo 8 (15), Slite ‘a’ (11), Snackarve (6), Valbytte 1 (8-10, 12), Valbytte 3 (7), Valleviken 1 (13). 6, Ar52298, dorsal view of incomplete exoskeleton, glabellar tuberculation displaced leftward, x 1-5. la-c, Ar51786, anterior, oblique anterolateral, and dorsal views of incompletely enrolled specimen, a, x 3-5, b, c, x3. 8, Ar52290, dorsal view of incomplete large cranidium, x 2. 9, Ar52291, dorsal view of pygidium, x4. 10, Ar51787, exterior view of enrolled specimen, x 3. II, Ar30643, dorsal view of pygidium, note large tubercles and unusually wide mucro, x2-5 (coll. G. Lindstrom). 12, Ar52292, dorsal view of small pygidium, x 10. 13u, b, Ar52295, lateral and exterior views of labral plate, x4. 14, Ar52297, exterior view of free cheek, x 3. 15, Ar5 1791, dorsal view of incomplete exoskeleton, x4. PLATE 38 RAMSKOLD, Encrinurus (Encrinurus) punctatus 536 PALAEONTOLOGY. VOLUME 29 non 1962 Encrinurus punctatus (Wahlenberg, 1821); Maksimova, p. 155, pi. 18, figs. 4 and 5 [With synonymy list]. . 1972 Encrinurus (E.) cf. punctatus 1; Schrank, p. 38, pi. 10, fig. 12. . 1973 Encrinurus tuberculaius (Buckland 1836); Clarkson and Henry, p. 118, figs. 12, 16a-d. . 1974 Encrinurus punctatus: Stemvers — van Bemmel, p.l4, fig. 24. ? 1975 Encrinurus punctatus (Wahlenberg), 1821; Balashova, p. 1 1 1, pi. 1, fig. 23. . 1978 Encrinurus punctatus (Wahlenberg); Mannil, p. 109, pi. 1, figs. 1-7; pi. 2, figs. 1-6. [With synonymy list.] V. 1981 Encrinurus tuberculatus (Buckland, 1836); Thomas, p. 64, pi. 18, figs. 1 and 3. . 19826 Encrinurus punctatus (Wahlenberg, 1821); Mannil, p. 53, pi. 4, figs. 1 -3 [copy Mannil 1978]. Remarks. This species has been mentioned countless times in literature, and the synonymy list includes only the most important references (further synonymy accepted here is found in the lists indicated above). Lectotype. Specimen in the Palaeontological Institution, University of Uppsala, no. PIU G1200, a pygidium figured by Wahlenberg 1818, pi. 2, fig. 1*, designated as lectotype and figured by Tripp and Whittard 1956, pi. 3, figs. 1 and 2; from an unknown locality on Gotland, probably Hogklint Beds. Additional material. All material or localities of this common species cannot be listed here, but some locality data are included in ‘Remarks on distribution’. Norwegian material previously referred to this species has not been studied here. The stratigraphical ranges of the three forms here included in punctatus are as follows; Form A (the ‘type’ form). On Gotland this form first occurs together with E. (E.) laevis in beds belonging to the top of the Upper Visby Beds or base of the Hogklint Beds. It is very common in the Hogklint Beds, units a-d, and it ranges into the oldest part of the Slite Marl. Form A is also common in the British Wenlock. Form B. This is the most common trilobite in the upper part of the Jaani Stage on Saaremaa, Estonia, and a large collection has been studied. Form C. Known only from the Slite Beds on Gotland; from the northwestern part of the Slite Marl to the ‘Pentamerus gothlandicus Beds’. E. (E.) punctatus Form C is the most common trilobite at all localities in this part of the Gotland sequence. Diagnosis. Glabella not depressed below cheeks posteriorly. Anterior cranidial border with eight or ten tubercles, rarely one additional central. IL small, not tuberculate abaxially, tubercle-pair I-l very rare. Tubercles on field of free cheek prominent. Pygidium with twenty-five to thirty-two rachial rings. R/P ratio 2-5-3-0. Mucro about one sixth to one half length of rachis. Description. The distinguishing features in the cephalon of the three different forms included here in E. (E.) punctatus are listed in Table 1. The thorax and pygidium are very similar in the different forms; a comparison of certain pygidial features is given in text-fig. 4. Discussion. E. (E.) punctatus, as conceived here, is an extremely variable species. However, variation within each of the three morphological forms is small, and there is only slight or no overlap between the forms (Table 1). In spite of this it is uncertain if a splitting of punctatus into three formal subspecies or species would better reflect the original, biological relationships, and the practical advantages gained by a splitting are regarded here as too small to warrant such a step. The British Wenlock E. (E.) tuberculatus is clearly synonymous with punctatus (though a neotype has not yet been designated for tuberculatus). Most major British collections have been studied, and the material belongs to E. {E.) punctatus Form A. This is so with most of the specimens figured as E. tuberculatus by Tripp (1962), but one specimen (pi. 66, fig. 7) has prominent, tuberculiform IL and is not conspecific with the other material; this is also the case with one of the pygidia (Tripp 1962, pi. 68, fig. 5). All Estonian Jaani material figured by Rosenstein (1941) and Mannil (1978) belongs to E. {E.) punctatus Form B, as does a large collection in Riksmuseum from the Jaani Stage of Saaremaa. Form B is coeval with one or both of Forms A and C, but a precise correlation is difficult. Morphologically, Form B is intermediate between Forms A and C. The pygidium from the Podolian Lower Wenlock figured by Balashova (1975) may well belong to Form A, but eannot be firmly assigned until a Podolian cranidium is figured. TABLE 1. Comparison of the main distinguishing features of the three forms of Encrinwus (Encrinurus) pimctatiis. RAMSKOLD: SILURIAN ENCRINURID TRI LOBITES 537 538 PALAEONTOLOGY. VOLUME 29 1. Number of pleurae Form A Ireviken 1 . Number of rachial rings 3. Rings anterior to eighth pleural rib 6 4 2 0 Form B Jaani church Form C Valbytte 1 9I 92 ,q1 25 26 27 28 29 30 31 32 33 34 35 17 18 19 20 21 8 6 4 2 0 8 6 4 2 0 TEXT-UG. 4. Comparison of certain pygidial features of the three forms of Encrinurus (Encrimirus) pimctatus. The number of rachial rings anterior to the eighth pleural pair (column 3) is the same as that used for calculating the R/P ratio. Encrinurus (Eucriuiirus) iutersitus sp. nov. Plate 39. figs. 1 -9; Plate 42, figs. ?12 and ?13; Plate 45, figs. ?12, 13, ?14, ?15 V. 1962 Encrimirus nuicrouriis Schmidt; Tripp, pi. 67, fig. 1; pi. 68, fig. 2 [other figured specimens = E. ( E. ) nuicrourus] V. 1978 Encrinurus nuicrourus Schmidt; Mannil, pi. 2, fig. 9 [other figured specimens = E. (E.) mcicrou- riis]. Name. Latin iutersitus, in intermediate position; referring to the geographical and stratigraphical position between E. (E.) macrourus and E. (E.) stubblefieldi. Holotype. Ar49041, broken cranidium with five thoracic segments (PI. 39, fig. 9), from Likmide, 1, Hemse parish, Hemse Marl NW. EXPLANATION OF PLATE 39 Figs. 1 “9. Encrinurus (Encrinurus) intersitus sp. nov. Hemse Beds, Hemse Marl NW (1, 3, 4, 9), Hemse Marl SE (2, 5-8). Hulte 3 (2, 5-8), Likmide 1 (4, 9), Lukse 2 (3), field 100 m north-east of Vakten 1(1). \a-d, Ar52477*, dorsal, oblique anterolateral, anterior, and lateral views of complete exoskeleton, x4. 2, Ar52476*, dorsal view of pygidium, note aberrant short, rounded mucro, x3. 3, Ar52479*, exterior view of slightly worn pygidium, x 2-5. 4, Ar51779*, dorsal view of pygidium, x 3 (coll. C. Pleijel 1974). 5 and 6, exterior view of free cheek. 5, Ar52472*, x 2-3. 6, Ar52473*, x 3. la, b, Ar52474*, oblique anterolateral and dorsal views of incomplete cranidium, x 3. 8u, h, Ar52475*, dorsal and lateral views of pygidium, x 3. 9, Ar49041, dorsal view of holotype cranidium with part of thorax, x 2 (coll. C. Pleijel 1974). * Paratypes. PLATE 39 RAMSKOLD, Encrinurus (Encrinurus) intersitus 540 PALAEONTOLOGY, VOLUME 29 Paratypes. Apart t'rom the figured specimens there are large collections in Riksmuseum and SGU, and all material cannot be listed, but specimens in Riksmuseum are labelled as paratypes. Localities. This species is common at several localities in the Hemse Marl (text-fig. 9), although material from the northern and eastern localities is slightly different from the main type, and cannot be assigned confidently to this species. Positively identified specimens come from the following localities: Hemse Marl NW; Hablingbo parish -Lilia Hallvards 6, Lilia Hallvards 7, Lukse I, Lukse 2, field at the road junction 100 m NNE of Vakten 1; Havdhem parish— Hemmungs 1, Kvinnegarda; Hemse parish— Likmide I. Hemse Marl SE; Burs parish — Vastlaus 1; Hemse parish — Hulte 3. Eocalities with questionably assigned specimens: Klinteberg Beds, Klinteberg Marl: Gerum parish— Ajmunde 1. Hemse Marl NW: Fardhem parish— Gardarve 1, Gerete 1, 250 m south of Gardsby; Einde parish— Amlings 1; Lye parish— well by the road at the southern Medebys farm; Silte parish— Mastermyr I. Hemse Marl SE and Hemse Beds, upper part: Burs parish— Hagvide 3; Einde parish — Rangsarve I; Eojsta parish — Ase 1, Klints 1; Nar parish— Nyan 3, Ondarve 1; Rone parish— Sigdes kanal (probably = Sigdes 1). Diagnosis. Glabella weakly convex, set well below cheeks posteriorly. Anterior cranidial border with nine tubercles, rarely eight. IL large, usually with tubercle abaxially. Tubercle-pair I-l typically present. Eyes set wide apart. Genal spines and spine on tenth thoracic segment very long and stout. Tubercles on field of free cheek small. Pygidium with twenty-one to twenty-five rachial rings. R/P ratio 2-0-2-3. Mucro at least as long as rachis. Discussion. This species is characterized by the following features in addition to those listed in the diagnosis: the wide cephalon, width:length ratio 2-8:l-3 0:l, the large and densely pitted field of free cheek, and the distinct tips of the seventh pleural pair proximally on the mucro. This is an easily identified species in the southwestern localities, where intersitus replaces macrou- rus without overlap. The species can be followed towards the north-east through the localities Likmide 1, Hulte 3, and Vastlaus 1. North of this occurs a slightly different form in beds that are probably slightly older. This form, best studied in material from Ajmunde 1 (PI. 42, fig. 12), Amlings 1, and Gardarve 1 (PI. 42, fig. 13) differs from the southwestern material mainly in the more elongated pygidium with shorter mucro. Although the pygidium, and to some extent the free cheek, is actually closer to E. (E.) macrourus (‘Snoder type’; text-fig. 5c) than to intersitus, the cranidium is clearly of intersitus type. At present the form is best referred to as E. {E.) cf. intersitus. Specimens from localities further north-east, in probably slightly younger beds (Ase 1, Klints 1), again have long mucro, but are poorly preserved. The material from eastern localities (Hagvide 3, Nyan 3, Ondarve 1, Sigdes kanal) is incomplete but most similar to inter situs among Gotland Encrinurus, and is referred provisionally to that species. E. (E.) intersitus sp. nov. differs from the youngest form of E. {E.) macrourus (text-fig. 5e) by the stout genal and thoracic axial spine, the shape and tuberculation of the free cheek, fewer EXPLANATION OF PLATE 40 Figs. I, 3-13. Encrinurus (Encrinurus) jarkanderi sp. nov. All Hemse Beds, unit b. Djaupviksudden 1 (3, 12), Gyle 1 (8). ‘Kraklingbo’ (11), Vidfalle 1 (1, 4, 6, 7, 9, 10), just south of 2 in point 22-5 near Kraklingbo (5), east of road 0-9 km north of Kraklingbo (13). 1«, b, Ar52327, oblique anterolateral and dorsal views of holotype cranidium, x 3-5. 3, EO 5700*, dorsal view of incomplete cranidium, x 3 (coll. J. E. Hede). 4, Ar52330*, dorsal view of pygidium, x 6. 5, SGU 5046*, exterior view of free cheek border, x 4 (coll. J. E. Hede 1919). 6, SGU 5047*, dorsal view of cranidium, x3 (coll. H. Munthe 1908). 7. Ar52329*, anterior view of thoracic segment, x 5 (coll. L.-I. Jarkander 1984). 8, SGU 5048*, oblique lateral view of pygidium, x6 (coll. J. E. Hede 1922). 9, Ar52328*, exterior view of free cheek, x3-5. lOa-c, Ar52331*, dorsal, posterior, and lateral views of pygidium, x4 (coll. L.-I. Jarkander 1984). 11, Ar30731*, dorsal view of pygidium, x 3. 12 and 13 exterior view of free cheek. 12, SGU 5049*, x3-5 (coll. J. E. Hede 1923). 13. SGU 5050*, x4. Fig. 2. Balizotna ohtusus (Angelin, 1851), Form B. Ar52341, dorsolateral view of complete, distorted specimen, Mastermyr 1, Hemse Beds, Hemse Marl NW, x 2. * Paratypes. PLATE 40 RAMSKOLD, Encrinurus, Balizoma 542 PALAEONTOLOGY. VOLUME 29 rachial rings in the pygidium, and a markedly larger size. E. (£’.) stuhhlefieldi is similar to intersitus in several features, especially in the shape of the glabella, the short and wide cranidium, the stout genal and thoracic axial spine, and the long mucro. The labral plate of both species is apparently similar, but since the two species commonly occur together, isolated labral plates (PI. 45, figs. 12, 14, 15) are difficult to assign to either species. E. {E.) stubblefieldi differs from intersitus in the subdued tuberculation, the shape of the free cheek, and the pygidium with fewer rachial rings and posterior pleural ridges that do not end in distinct tips on the pygidial margin. Encrinurus (Encrinurus) jarkanderi sp. nov. Plate 40, figs. 1, 3-13 Name. After Dr Lars-Ivar Jarkander who collected well-preserved material of this species. Holotype. Ar52327, cranidium (PI. 40, fig. 1), from Vidfalle 1, Kraklingbo parish, Hemse Beds, unit b. Paratype.s. All material is from Hemse Beds, unit b. Ala parish— Gyle 1 (SGU 5048). Kraklingbo parish — Djaupviksudden 1 (Ar52333-52334, LO 5700, SGU 5049), Hagrummet 1 (SGU 5057), Osterby 1 (SGU 5052), Vidfalle 1 (Ar52328-52332, SGU 5047), east of road 0-9 km north of Kraklingbo church (SGU 5050), just south of 2 in point 22.5 near Kraklingbo (SGU 5046), ‘Kraklingbo’ (Ar30731). One specimen (Ar30562) from Gutenviks, Ostergarn parish, may be from Hemse Beds unit c. In total five cranidia, four free cheeks and seven pygidia. Diagnosis. Usually nine, sometimes ten tubercles on anterior cranidial margin. IL small, no tubercle abaxially. Tubercle-pair 1-1 absent. Eyes set wide apart. Tubercles on field of free cheek prominent, precranidial lobe low, elongate, main tubercle row on border prominent. Pygidium with 24-29 rachial rings. R/P ratio 2-3-2-7. Mucro very small, 5-10 % length of rachis. Description. Only characteristic features are described. IL continuing as a narrow band across glabella (PI. 40, figs. \b and 6). Two tubercles anteriorly between eye and rachial furrow. At least three large tubercles at base of genal spine. Lateral border of free cheek with main tubercle row almost straight, tubercles very large, posterior row along suture with about five distinct tubercles. Precranidial lobe with about eleven tubercles roughly arranged in two to three oblique rows. Rostral plate not known, rostral suture short (dorsoventrally). Thorax known from isolated segments; at least one stout axial spine, presumably on tenth segment. Pygidium highly vaulted, usually 9' pleurae, occasionally 10'. Tips of sixth and seventh pleural ribs weakly defined on margin. Interpleural furrows wide (exsag.), anterior pleural ridge distinct in anterior five furrows. Discussion. This species is most easily recognized by the characteristic free cheek tuberculation combined with anterolaterally set eyes and a short-tipped pygidium. These features also distinguish E. {E.) jarkanderi from the otherwise rather similar E. (E.) niacrourus and E. (£.) nasutus (see these species below). E. (E.) jarkanderi is unusual in its subgenus in having IL and 2L that can be traced across the glabella, defined by faint, elongate depressions. These are, however, too weak to be regarded as truly continuous IS and 2S. EXPLANATION OF PLATE 41 Figs. 1-10. Encrinurus (Encrinurus) niacrourus Schmidt, 1859. Mulde Beds, lower part (4, 5, 8), Hemse Beds, Hemse Marl NW (1-3, 6, 7, 9, 10). Blahiill 1 (4, 5, 8), Eske 1 (9), Smissarve 1 (3), Snoder 1 (6, 10), Snoder 2 (1, 2), Urgude 3 (7). 1 and 2, dorsal view of glabella of complete specimen. 1, Ar52463, note tubercle iii-0, X 5. 2, Ar52464, x 6-5 3u, h. Ar52452, dorsal and anterolateral views of enrolled specimen, x 3. 4, Ar52456, exterior view of free cheek, x4. 5, Ar52459, dorsal view of pygidium, x4. 6, Ar52461, ventral view of labral plate of complete specimen, x4. 7u-c, Ar51789, dorsal, anterior, and lateral views of complete specimen, note indistinct sagittal band, a, c, x 3, 6, x 5. 8, Ar52455, dorsal view of cranidium, x4. 9a-c, Ar51790, exterior, anterior, and lateral views of complete specimen, a. c, x2-5, h, x 3. 10, Ar52462, dorsal view of large incomplete pygidium, x 4. PLATE 41 RAMSKOLD, Encrinurus (Encrinunis) macrourus 544 PALAEONTOLOGY, VOLUME 29 E. (E.) jarkauderi is known with certainty only from Hemse Beds unit b, which must be synchronous with part of the Hemse Marl NW. However, trilobite faunas from these two units are very different and cannot be correlated precisely with each other. Encrinurus (Encrinurus) macrourus Schmidt, 1859 Plate 41, figs. HIO; Plate 42, figs. 1-11; text-fig. 5 * 1859 Encrinurus punctatus var. macrourus Schmidt, p. 438. .1941 Encrinurus punctatus (Wahlenberg) 1821; Rosenstein, text-fig. 4a, pi. 2, fig. 4-46. V. 1962 Encrinurus macrourus Schmidt; Tripp, p. 469, pi. 65, figs. 1, 3, 4; pi. 66, fig. \a-c: pi. 67, figs. 2- 4; pi. 68, figs. 1, 3, ?9; non pi. 65, fig. 2[= E. {E.) punctatus]\ pi. 67, fig. 1; pi. 68, fig. 2[= E. (E.) cf. intersitus sp. nov.]. . 1972 Encrinurus (E.) cf. punctatus 2; Schrank, p. 38, pi. 11, fig. 4 [figs. 1, 2, 7 = paratypes of E. (E.) ruhnuensis Mannil, 1978; figs. 3, 5, 6 indeterminable]. non 1972 Encrinurus (E.) punctatus macrourus Schmidt, 1859; Schrank, p. 42, pi. 12, fig. 6 [? = E. (E.) ruhnuensis Mannil, 1978], fig. 7 [? = E. (E.) punctatus], non 1972 Encrinurus (E.) cf. punctatus macrourus Schmidt, 1859; Schrank, p. 43, pi. 13, figs. 1, 2 [= £. (E.) balticus MannW, 1978]. 1977 E. macrourus Schmidt, 1859; Schrank, p. 112. V. 1978 Encrinurus macrourus Schmidt; Mannil, pi. 2, figs. 7, 8, non fig. 9 [= E. (E.) intersitus sp. nov.]. 1980 Encrinurus macrourus Schmidt, 1859; Strusz, p. 56, text-fig. 15 [but note that the cephalon is drawn from specimen of Schrank 1972, pi. 12, fig. 6,2 = E. (E.) ruhnuensis Mannil, 1978]. 1982 Encrinurus macrourus Schmidt; Alberti et al., p. 32. 1982a Encrinurus macrourus', Mannil, p. 65. Lectotype. Specimen in the Geological Institute, Tallinn, Estonia, no. Tr 1905, incomplete thorax and pygidium from Petesvik, Hablingbo parish, Hemse Marl NW, selected and figured by Tripp 1962, pi. 67, fig. 4; refigured Mannil 1978, pi. 2, fig. 7. Remarks. Through the courtesy of Dr. Reet Mannil in Tallinn I have been able to study Schmidt’s Petesvik material (syntypes), and all specimens but one (see below) are conspecific. Schmidt (1859) also listed E. punctatus var. macrourus from several localities in the Hemse Beds apart from Petesvik. Pygidia with long mucro from these other localities do not belong to macrourus but to E. [E.) stuhhlefieldi or E. (E.) intersitus sp. nov. Schmidt’s taxon was not used until Tripp (1962) revived it and selected a Petesvik specimen as lectotype. The original Petesvik locality is now inaccessible for recollecting. However, well-preserved material from nearby Lilia Hallvards 4, at the southern end EXPLANATION OF PLATE 42 Figs. 1-11. Encrinurus (Encrinurus) macrourus Schmidt, 1859. Mulde Beds, lower part (1-3), undifferentiated (6), Hemse Beds, Hemse Marl NW (4, 5, 7-1 1). Blahall 1 (1-3), Djupviksvagen 1 (6), Hagsarve 4(11), Lilia Hallvards 4 (7, 9), Petesvik (4), Smissarve 1 (8), Urgude 3 (5, 10). 1, Ar52460, exterior view of small pygidium, xll. 2-4, exterior view of labral plate. 2, Ar52458, x 5. 3, Ar52457, x4. 4, Ar30438, complete specimen, x 5 (coll. G. Lindstrom). 5a, b, Ar52451, exterior and posterior views of pygidium, note the very short ‘mucro’, x 4. 6a, b, Ar52454, dorsal and lateral views of aberrant pygidium, with one extra pair of pleural ribs, x 6. la, b, Ar52350, complete specimen with pygidium and cephalon in dorsal views, x 3. 8 and 9, dorsal view of pygidium. 8, Ar52453, x4. 9, Ar52349, mucro broken, x 6. 10, Ar52450, dorsal view of pygidium and part of thorax of complete specimen, x4. 11, Ar52449, exterior view of free cheek, visual surface lost, x 4. Figs. 12 and 13. E. (E.) cf. intersitus sp. nov. Dorsal view of pygidium. 12, Ar52465, Ajmunde 1, Klinteberg Beds, Klinteberg Marl, x4. 13, Ar52466, incomplete, Gardarve 1, Hemse Beds, Hemse Marl NW, x4. Figs. 14 and 15. E. (E.) nasutus sp. nov. Eke Beds, upper part; Lau Backar 1. 14, Ar51792*, latex cast of cephalon and an additional cranidium, x 2-5. 15, Ar51793*, dorsal view of cranidial fragment, note the very sparse tuberculation, x 3. * Paratypes. PLATE 42 RAMSKOLD, Encrinurus {Encrinurus) 546 PALAEONTOLOGY, VOLUME 29 of Petesvik, is morphologically identical to Schmidt’s specimens. Two more of Schmidt’s specimens have been figured; a cranidium (Mannil 1978, pi. 2, fig. 8) and a pygidium, possibly not from Petesvik (Mannil 1978, pi. 2, fig. 9); the latter apparently belongs to E. (E.) intersitus sp. nov. Additional material. Localities: Miilde Beds, lower part: Eksta parish- Blahiill 1; Klinte parish — Viirsande 1. Undin'erentiated: Eksta parish — Djupviksvagen 1, Nordervik (635350 163845); Frojel parish — Mulde Tegel- briik 1. Upper part: Frojel parish — Haugkiintar 1; Klinte parish— Eoggarve 2. Klinteberg Beds, lower part: Loggarve 2. Hemse Beds, Hemse Marl NW: Hablingbo parish — Lilia Hallvards 4, 5, Petesvik; Levide parish — Levide 1; Silte parish- -Mastermyr 1, Mickels I, Smissarve 1, Snoder 2; Sproge parish— Eske 1, Hagsarve 2, 4, Snoder I, Urgude 3, 4. Localities with too few or poor specimens to permit definite assignment: Mulde Beds, upper part: Frojel parish -Diipps 1, 2. Klinteberg Beds, lower-middle part: Klinte parish— Klinteberget. Diagnosis. Anterior cranidial border usually with nine, sometimes eight or ten tubercles. 1 L small, no tubercle abaxially, tubercles I-l and iii-0 rare. Precranidial lobe low, field of free cheek with prominent tubercles. Slender axial spine or tubercle on tenth thoracic segment. Pygidium with seventeen to twenty-nine rachial rings. R/P ratio 2-3-2-9. Description. Only features that may be of diagnostic value are described. Glabella weakly inflated, set below cheeks posteriorly. Anterior border of cranidium usually with nine tubercles (60-70 %), sometimes eight (c. 25 %), rarely ten ( < 10 %). Tubercle iii-0 present in less than 10 % of specimens. Tubercle 1-0 fairly common (c. 35 %). Tubercle row VI invariably present. IL ridge-shaped, sometimes almost reaching sagittal line. Eye pedunculate, height of stalk varying from equal to height of visual surface (Hemse Beds, PI. 41, figs. 3 and 9) to 1-5 times that height (Mulde Beds, PI. 41, fig. 4), eye-stalk only slightly constricted at base. Eye set close to rachial furrow, anteriorly one tubercle between eye and rachial furrow, distance to posterior border furrow less than or equal to diameter of eye (visual surface). Rostral plate (PL 41, figs. 3, 7, 9) two to three times as high as wide, with a tubercle centrally on upper half, lower half slightly protruding. Axial tubercle on seventh thoracic segment. Hemse Beds specimens sometimes with only a tubercle on tenth segment (PI. 42, fig. 7), usually with a short, slender spine (PI. 41. figs. 7 and 9); Mulde Beds specimens appear to have a slightly stouter spine. Pygidium (text-fig. 5) usually with 9^ pleurae. Posterolateral tip of sixth pleura reaches as far back as end of rachis or longer, tips of all seven pleurae distinctly defined on margin. Mucro varying from (uniquely) a rounded knob (PI. 42, fig. 5) to flat subtriangular (all Mulde Beds, most Hemse Beds specimens) TEXT-FIG. 5. Chronologically arranged pygidia oi Encrinurus (Encrinnrus) macrourus. The lettering is the same as in text-figs. 6 and 9. All reconstructions are based on specimens of about equal size, a, Mulde Beds, Blahall 1. B, Hemse Marl NW, Urgude 3. c, Hemse Marl NW, Snoder 1 and 2. D, Hemse Marl NW, Smissarve 1. E, Hemse marl NW, Lilia Hallvards 4. RAMSKOLD: SILURIAN ENC R I N U R I D T R I LO B ITES 547 rachial rings 30- 20- . , ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ , , 0.3 0.5 0.7 0.9 1.1 1.3 1.5 TEXT-HG. 6, Diagram showing progressive change in the pygidium of Encrinwus (Eucrimous) macrowiis towards fewer rachial rings and longer mucro. The scale on the abscissa is the ratio between the sagittal length of the rachis (excluding articulating half-ring) and the postrachial part of the pygidium (most of which is the mucro). The letters a~e indicate the localities of text-ligs. 5 and 9. Pygidia of less than 7 mm total length have not been included. to lancet-like, subtriangular in section (large Hemse Beds specimens, PI. 41, figs. 9 and 10) to as long as rachis (PI. 42, figs. 7 and 9). One small pygidium (PI. 42, tig. 6) has eight pairs of pleurae reaching margin, similar to the specimen figured by Schrank (1972, pi. 11, fig. 6); this is considered abnormal. Trends in variation. During the long range of this species some changes in morphology can be followed. The oldest (Mulde Beds) specimens have highly stalked eyes, deep rachial furrows, a fairly stout axial spine, about twenty-seven to thirty rachial rings in the pygidium, 9' -10' pleurae, and a triangular mucro (text-fig. 5a). In the lower Hemse Marl NW, specimens still have deep rachial furrows, but lower eyes, a very slender axial spine, about twenty-two to twenty-five rachial rings, 9' pleurae, and a mucro ranging from triangular (lowermost Hemse Marl NW; text-fig. 5b) to lancet-shaped (slightly younger; text-fig. 5c). The type material (and Lilia Hallvards 4; text-fig. 5e) is the youngest of the species. It is characterized by shallow rachial furrows, an axial tubercle instead of spine, only seventeen to nineteen rachial rings, and 8' pleurae that are markedly extended posteriorly, so that the posterolateral tip of the fifth pleura reaches as far back as the end of rachis. From this a long mucro is produced; the seventh pleural pair reaches half its length, and the total length exceeds that of the pygidial rachis. The eye height is similar to low Hemse Marl NW material. There is thus a general trend in macroiirus towards lower eyes, shallower rachial furrows, reduction of the thoracic axial spine, fewer rachial rings and pleurae in the pygidium, and a longer mucro. Discussion. E. (E.) macroiirus is one of the smallest Gotland Encrinurus, with a total length for extended specimens usually 25-30 mm. Very few specimens reach 40 mm, but a fragmentary pygidium (PI. 41, fig. 10) indicates a total length of 49-50 mm. The small size was one criterion used by Tripp (1962) to distinguish macroiirus from punctatus. Other difterences from piinctatiis include the less inflated glabella, the smaller precranidial lobe, the rarity of tubercle iii-0 ( < 10 % compared to > 90 %), and the common occurrence of a median tubercle on the anterior cranidial border. Pygidia of the Gotland form E. {E.) cf. intersitus sp. nov. (PI. 42, figs. 12 and 13) show considerable similarities to some Hemse Beds specimens of macroiirus. These pygidia are, however, associated with a type of cranidium much more similar to intersitus than to macroiirus, being very wide and short and with densely tuberculate fixed cheeks. Better material is needed to clarify the relationship between this form, intersitus and macroiirus. The middle Wenlock E. {E.) ruhniiensis Mannil, 1978 from Estonian cores also closely resembles macroiirus. It differs by its short, weak genal spines and tubercle iii-0 combined with eight tubercles on the anterior cranidial border, a pattern that has not been observed in macroiirus. 548 PALAEONTOLOGY, VOLUME 29 Encrinurus {Encrinurus) nasutus sp. nov. Plate 42, figs. 1 4 and 1 5; Plate 43, figs. 1-11 . 1967 Enaimirus punctatus Wahlenberg 1821; Hucke, pi. 28, fig. 8, non fig. 7 [= Balizoma obtusus (Angelin, 1851)]. . 1972 Encrinurus (E.) cf. punctatus 3; Schrank, p. 39, pi. 11, figs. 9 and 10, non fig. 8 [= Balizoma obtusus (Angelin, 1851)]; pi. 12, figs. 1-3; text-fig. 3. Name. Latin nasutus, 'with nose’; referring to the nose-like profile of the rostral plate in situ. Holotype. Ar30558, incomplete extended exoskeleton, Plate 43, fig. \a-c, from Gannor 1, Lau parish. Eke Beds, lower part (local unit d of Hede in Munthe et al. 1925). Paratypes. From the type locality: Ar30557, Ar30528-30529. From the upper Eke Beds at Lau Backar 1, Lau parish: Ar51781, Ar5 1792-5 1795, Ar52480-52485, and a large number of unnumbered, fragmentary specimens. Incomplete material (Ar52440-52447, Ar52467) from the Hemse Beds unit e, at Millklint/ Torsburgen (Gammelgarn/Kraklingbo parishes) almost certainly also belongs to this species. Material from German Ludlow erratics described by Schrank (1972) as E. (E.) punctatus 3 is conspecific with E. (E.) nasutus sp. nov. Diagnosis. Anterior cranidial border with nine or occasionally eight or ten tubercles. IL large, with tubercle abaxially. Tubercle-pair I-l commonly present. Occipital ring and thoracic segments with numerous distinct tubercles. Field of free cheek with small tubercles, precranidial lobe low, very elongate. Rostral plate with strongly protruding lower part. Pygidium with twenty-six to twenty-nine rachial rings. R/P ratio 2-3-2-6. Mucro lancet-shaped, about half length of rachis. Description. Only characteristic features are described. IL commonly traceable across glabella, tubercle I-O or I-l present in over 50%. Tubercle ii-0 or ii-1 present in 100%. Anterior cranidial border usually with nine tubercles (five specimens), occasionally eight or ten (one specimen each). Fixed cheek densely tuberculate adaxially and posterior to eye. One or two small 'tubercles’ between PL and cranidial margin. Occipital ring with one central and two flanking tubercles. Precranidial lobe with about eleven tubercles. Labral plate with rhynchos bluntly subtriangular anteriorly. Weak axial tubercles on most thoracic segments, best developed on third, fifth, and seventh; tenth segment with a stout axial spine. A strong lateral tubercle on several segments. Pygidium with 9' pleurae. No variation observed in shape of mucro. Discussion. This species is readily distinguished from all other Gotland Encrinurus by the combi- nation of tuberculiform IL and lancet-shaped mucro. A British cranidium figured by Tripp (1962, pi. 66, fig. 7) as E. tuhercuiatus (= E. {E.) punctatus here) appears indistinguishable from nasutus, but more complete material is needed for a firm assignment. E. (E.) nasutus is also known from German upper Ludlow erratics (E. (E.) cf. punctatus 3 of Schrank 1972). EXPLANATION OF PLATE 43 Figs. 1-11. Encrinurus (Encrinurus) nasutus sp. nov. Eke Beds, lower part (1, 2, 7), upper part (3-6, 8-11). Gannor 1 (1, 7), Lau Backar 1 (3-6, 8-11), 'Lau Kanaf (2). \a-c, Ar30558, dorsal, lateral, and anterior views of almost complete holotype, a, b, x2-5, c, x4. 2a, b, Ar30528*, dorsal and oblique anterolateral views of enrolled specimen, x 3. 3, Ar52482*, exterior view of labral plate, x4. 4n, 6, Ar5 1781*, anterior and dorsal views of incomplete cephalon, a, x 5, b, x 3. 5a, h, Ar52480*, anterior and dorsal views of enrolled specimen, a, x6, b, x 4, only eight tubercles on anterior cranidial border. 6, Ar52485*, dorsal view of pygidium, x 6. 7, Ar30557*, dorsal view of cranidium and thorax, x2-5. 8 and 9, dorsal view of pygidium. 8. Ar52483*, x 5. 9, Ar52484*, x4. 10, Ar5 1795*, ventral view of pygidium, x 4. ll,Ar52481*, exterior view of free cheek, x 5. * Paratypes. PLATE 43 RAMSKOLD, Encrinums (Encrinurus) nasutus 550 PALAEONTOLOGY, VOLUME 29 Encrimirus (Encriminis) odvaldensis sp. nov. Plate 44, tigs. 1-13 Name. From the type locality. Holotype. Ar52313, cranidium (PI. 44, fig. 10), from Odvalds 1, Klinte parish, Slite Beds, ‘Pentamerus gothlan- dicus Beds’. Paratypes. From the ‘Pentamerus gothlandicus Beds’, Klinte parish-Odvalds 1 (Ar523 14-52323), Robbjans 1 (Brl31177), Robbjans 2 (Ar52300-52312), Svarvare 1 (Ar52324-52326), drainage ditch in C. Smitterberg’s field (Ar30546). From field south-west of Gannarve, Frojel parish ( Ar30547), uncertain horizon. Three pygidia (Ar30671, Ar30687-30688) from ?Mulde Beds at Djupvik, Eksta parish, also belong to this species. In total four cranidia, five free cheeks, four labral plates, and twenty pygidia. Diagnosis. Glabella sparsely tuberculate; tubercles iv-0, v-0, VI-0 typical. Anterior cranidial border with eight tubercles. IL small, not tuberculate, tubercle-pair I-l absent. Eye fairly highly stalked. Prominent tubercles on field of free cheek. Pygidium with twenty to twenty-two rachial rings. R/P ratio 2-4-2-6. Mucro equal in length to rachis. Seventh pleural pair extends well on to the mucro. All tuberculation coarse. Discussion. This species is characterized by the reduced glabellar tuberculation, highly stalked eyes, and wide pygidium with long mucro. The pygidia are uniform in shape, number of rings and ribs, and length of mucro. Of the Gotland species it is most similar to E. {E.) sp. A, which is poorly known, but differs at least in the more elongate pygidium with more rachial rings. E. {E.) sp. A is probably only slightly older than odvaldensis. The Estonian Wenlock E. (E.) balticus Mannil, 1978 closely resembles odvaldensis, but differs in the more elongate glabella (of macrourusjinlersitus shape) with well-developed IL carrying tubercle pair I-l, the less coarse tuberculation, larger number of tubercles on field of free cheek, and the pygidium with narrower rachis, seventh pleural pair set even farther posteriorly, and more slender, longer mucro (see also Schrank 1972, pi. 13, figs. 1 and 2; specimens included by Mannil in balticus). E. (E.) odvaldensis is also remarkably similar to the American Wenlock E. (E.) egani Miller, 1880 (revised by Holloway 1980), especially in the reduced glabellar tuberculation, but in most other features as well. However, egani was interpreted as dimorphic by Holloway ( 1 980), has much higher eye stalks, and the base of the mucro is raised above the plane of the pleural tips. E. (E.) egani is even closer to E. {E.) sp. A, with which it is compared in the discussion of that species. E. (E.) odvaldensis is known with certainty only from the ‘Pentamerus gothlandicus Beds’, immediately below the Slite Siltstone, but may continue into beds following directly on the Slite Siltstone near the western coast of Gotland. This poorly known interval of the Gotland sequence includes the boundary of the Slite/Mulde Beds. The specimen from south-west of Gannarve and those from Djupvik probably originate from these strata. The distribution is otherwise restricted to EXPLANATION OF PLATE 44 Figs. 1-13. Encrinurits (Encrimiru.s) odvaldensis sp. nov. Slite Beds, ‘Pentamerus gothlandicus Beds’ (1-6, 8- 13), ?Mulde Beds, lowest part (7). Odvalds 1 (4, 5, 9, 10), Robbjans 2 (1, 2, 8, II, 13), Svarvare 1 (3, 6, 12), field south-west of Gannarve (7). 1«, /?, Ar52304*, lateral and dorsal views of pygidium, x 4. 2, Ar52303*, exterior view of labral plate, x 5-5. 3u, h, Ar52326*, posterior and dorsal SEM views of last meraspid stage pygidium, x35. 4 and 5, exterior view of free cheek. 4, Ar523 14*, x 3. 5, Ar52315*, x 8. 6u, b, Ar52325*, dorsal and lateral views of pygidium, x 3 (coll. Amelang 1982). 7, Ar30547*, dorsal view of pygidium, x 5. 8, Ar52301*, exterior view of labral plate, x 5. 9, Ar523 16*, dorsal view of pygidium, x3. lOu, Ar52313, dorsal and oblique anterolateral views of holotype cranidium, x 3. 11, Ar52302*, exterior view of labral plate, x4. 12, Ar52324*, anterolateral view of incomplete cephalon, x 3. 13o, 6, Ar52300*, oblique anterolateral and dorsal views of incomplete cranidium, x 3. * Paratypes. PLATE 44 RAMSKOLD, Encrinurus {Encrinurus) odvaldensis 552 PALAEONTOLOGY, VOLUME 29 the southwestern part of the Tentamerus gothlandicus Beds’, where this species is fairly common, and the only encrinurid; in the north-east these beds yield E. (E.) punctatus Form C. Encriminis {Encriminis) schisticola Tornquist, 1884 Plate 45, figs. 1-9 V* 1884 Encrimiru,s schisticola Tornquist, p. 23, p\. 1, figs. 15-17. . 1968 Encrimirus schmidti Miinnil, p. 273, pi. 1, figs. I -5; pi. 2, figs. 1-5. Lectotype. Selected here; LO 573 T, incomplete cranidium and thorax, figured Tornquist 1884, pi. 1, fig. 15, refigured here Plate 45, fig. 4, from the Retiolites shale (Kullatorp Stage, upper Llandovery), Styggforsen, Dalarna. Remarks. Tornquist (1884, pi. 1, fig. 16) also figured a pygidium (LO 574 t) from the counterpart slab of the lectotype, possibly belonging to the same individual. The third specimen he figured is a pygidium (SGU 4187), also from the type locality. Other material. In Sweden this species is known from the Retiolites shale at Styggforsen (Ar24675-24676) and Nittsjo in Dalarna (Ar24637-24638, Ar24643-24671, LO 5701-5705, and unnumbered material). In total eight cranidia, one labral plate, four free cheeks, eighteen pygidia, and numerous thoracic segments. A well- preserved pygidium (SGU 5056) from 100-2-101 -3 m in the core from the boring at Visby Cement factory is indistinguishable from E. (E.) schisticola. The age is probably upper Llandovery (Dalip Sethi, pers. comm.). In Estonia this species is known from the Adavere Stage (Upper Llandovery), Konovere River, Latikiila. Diagnosis. See Mannil 1968, p. 278 (diagnosis of E. schmidti). Description. Only some minor details can be added to the comprehensive description of E. schmidti given by Mannil (1968). As noted by Mannil, there is a weak but distinct axial tubercle on each of the fifth, seventh, and tenth thoracic segments. This arrangement is likely to be homologous with the axial spines and tubercles present in later E. (Encriminis) on the same segments. The number of rachial rings can be counted only in four Swedish pygidia (apart from the Visby core specimen which has 20 rings); these have 18, 18, 17, and 17 rings, compared with 19 or occasionally 20 in Estonian material. R/P ratio 2 0-2-3. 9' pleurae; last pair of furrows visible only in well-preserved specimens. Anterior pleural bands distinct. Discussion. The Estonian specimens of E. schmidti Mannil, 1968 agree completely morphologically with the Swedish material on which Tornquist ( 1 884) erected E. schisticola. E. schmidti is accordingly regarded here as a junior subjective synonym of E. (E.) schisticola. This species is easily identified, and cannot be mistaken for any other described species. Mannil (1968) recognized a morphological group of E. (E.) schmidti, including also E. kiltsiensis and E. rumbaensis, both described by Rosen- stein ( 1941 ). Such a group is not recognized here, and the two latter species are referred to a different subgenus, E. (Nucleurus), from E. (E.) schmidti. EXPLANATION OF PLATE 45 Eigs. 1-9. Encrimirus (Encrimirus) schisticola Tornquist, 1884. All from Retiolites shale, Dalarna. Nittsjo (1- 3, 5-9), Styggforsen (4). Figs. 1-5, 7 coll. S. L. Tornquist. 1, LO 5701, dorsal view of pygidium, x 3. 2 and 3, dorsal view of cranidium. 2, LO 5702, incomplete x 3. 3, LO 5703, x4. 4, LO 573T, exterior view of lectotype incomplete cranidium and thorax, x3. 5, LO 5704, dorsal view of pygidium, x3. 6, Ar24664, dorsal view of small cranidium, x 6. 7, LO 5705, exterior view of free cheek, x 4. 8, Ar24658, exterior view of incomplete labral plate, x 4. 9, Ar24668, exterior view of small free cheek, x 6. Figs. 10 and 1 1. £. (E.) sp. C? Both from Hemse Marl NW, Urgude 4. 10, Ar52281, dorsal view of almost complete exoskeleton, x4. 1 1 , Ar52280, exterior view of free cheek lacking visual surface, x2.5. Figs. 12, 14, 15. E. (E.) intersitus sp. nov. or E. (E.) stiibblefieldi Tripp, 1962. All from the Hemse Marl SE or top. Gannor 3(15), Vastlaus I (14), unknown locality no. I (12). Exterior view of labral plate. 12, Ar52299, x4. 14,Ar5l769, x 4 (coll. G. Holm). 1 5, Ar51777, x4. Fig. 13. E. (E.) intersitus sp. nov. Ar51780*, exterior view of labral plate from Likmide I, Hemse Marl NW, X 4 (coll. C. Pleijel 1974). * Paratype. PLATE 45 ^ V “ ■ • T ■ y i 1 1 A f' i 1 'i t 1 1 JyV / -V 1 RAMSKOLD, Encrinurus (Encrinurus) 554 PALAEONTOLOGY, VOLUME 29 Encrinurus (Encrimo us) stubhlefieldi Tripp, 1962 Plate 45, tigs. ? 1 2, ? 1 4, ? 1 5; Plate 46, figs. 1 - 1 4 V* 1962 Encrinurus stiibblefielcli Tnpp, p. 471, pi. 65, fig. 12, pi. 67, figs. 14 and 15; pi. 68, fig. 1 1. non 1968« Encrinurus stubhlefieldi Tx'\pp\ Balashova, p. 116, pi. 3 fig. 12. . 1972 Encrinurus (E.) stubhlefieldi Tripp, 1962; Schrank, p. 42, pi. 12, figs. 4 and 5 [with synonymy list]. Holotype. Specimen in the British Geological Survey, Nottingham, no. GSM 36846, internal mould of crani- dium, from the ‘Upper Ludlow shales’, WhitclilT, Shropshire, England, figured Tripp 1962, pi. 65, fig. 12; pi. 67, fig. 15. Material and localities. Outside England this species has been described from German Ludlow erratics (Schrank 1972). It is also present in the Podolian Ludlow (see below). On Gotland stubhlefieldi is restricted to the Hemse Marl SE, belonging to the younger part of the Hemse Beds (lower/middle Ludlow). Material: One fairly complete exoskeleton and numerous (minimum twenty each) cranidia, free cheeks, labral plates, and pygidia. Localities: all Hemse Beds, Hemse Marl SE (and top?): Alva parish— Alva kanal, c. 100 m south-east of Overostris farm; Burs parish — Hagvide 3, Djuptrask, Vastlaus 1, Kama kanal, drainage ditch south of Kalmans; Hablingbo parish— Leisungs 1; Hemse parish— Hulte 3; Lau parish— Botvide 1, Gannor 3, Gogs 1, Hallsarve 1, Tuten I, ditch 500 m north-east of Lau church; Nar parish— Nyan 4; Nas parish— Klasard 1, vaktard 2; Rone parish— Sigdes kanal. Diagnosis. Cranidium very wide, short. Glabella only weakly convex, depressed posteriorly. Rachial furrows wide. 1 L well developed, non-tuberculate. Eyes very low. Genal spines very stout and long. Glabella lacking tubercle row VI, eight tubercles on anterior border, PL set far from margin. Free cheek with faint tuberculation, border furrow shallow. Pygidium with sixteen to twenty-four rachial rings, pleural tips poorly defined on lateral margin. R/P ratio 1-7-2-2. Mucro at least as long as rachis. All tuberculation of low relief. Description. Cranidium with width to sagittal length 3-3:1 -3-5:1. Faint sagittal depression present. IS and 2S can be traced across glabella as very faint depressions. Tubercles on fixed cheek along rachial furrow small, arrangement opposite lateral glabellar lobes indistinct. The six central tubercles on anterior cranidial border set in contact with margin, PL set more posteriorly. Tubercle row I sometimes absent (PI. 46, fig. 9). Fixed cheek between palpebral lobe and rachial furrow wide (tr.), densely tuberculated and pitted. Genal spines at least as long as glabella. Free cheek subtriangular, lateral margin only gently convex, border furrow meeting anterior branch of facial suture at about midlength. Border faintly granular (PI. 46, fig. 10). Rostral suture short, as long as height of visual surface. Rostral plate unknown. Labral plate (PI. 45, figs. 712, 714, 715; PI. 46, fig. 13) similar to E. (E.) intersitus sp. nov., with rhynchos rounded or truncated (this may be abnormal) anteriorly. Thorax of eleven segments. Stout axial spine on tenth segment. Presence of axial spines or tubercles anterior to this not known. Pygidium with evenly curved lateral margin, with very weak bulges of pleural tips. Pleural EXPLANATION OF PLATE 46 Figs. 1-14. Encrinurus (Encrinurus) stuhblefieldi'Ynpp, 1962. All Hemse Marl SE. Alva kanal (10), Botvide 1 (9), Gannor 3 (2, 4, 1 1-14), Hulte 3 (5), Klasard I (3), Vaktard 2 (8), Vastlaus 1 (7), unknown locality no. 1 (1,6). Ifl, b, Ar52468, dorsal and anterior views of cranidium, x 2. 2, Ar51774, dorsal view of incomplete cranidium, x 2-5. 3 and 4, dorsal view of pygidium. 3, SGU 5053, x 5 (coll. C. Bergman 1984). 4, Ar52470, incomplete, x 3. 5u, b, Ar52469, lateral and dorsal views of pygidium, x4. 6. Ar52471, ventral view of cranidium, x 3. 7, Ar51771, dorsal view of pygidium, x4 (coll. G. Holm 1904). 8, SGU 5054-5055, exterior view of two small free cheeks, x5 (coll. S. Laufeld 1984). 9, Ar51782, dorsal view of small cranidium, x4. 10-12, exterior view of free cheek. 10, Ar30566, x 3 (coll. G. Lindstrom). ll,Ar51778, X 3. 12, Ar51775, x 3. 13, Ar5 1776, exterior view of labral plate, x 3 (truncated rhynchos abnormal?). 14, Ar51783, dorsal view of incomplete pygidium, mucro short, abnormal, x 3. PLATE 46 RAMSKOLD, Encrinurus {Encrinurus) stubblefieldi 556 PALAEONTOLOGY, VOLUME 29 furrows often with both rib furrow and interpleural furrow distinct. Mucro about twice as wide (tr.) as thick (dorsoventrally) at base. About three faint axial tubercles. Discussion. All British specimens examined are moulds, but distinctive enough to allow the Gotland material to be assigned to stuhhiefieldi without doubt. The testiferous specimens described by Schrank ( 1972) from German Ludlow erratics are indistinguishable from Gotland material. E. (E.) stuhblefieldi is also present in the Podolian Ludlow. Specimens kindly lent by Dr R. M. Owens can be confidently assigned to slubblefieldi. The material is from the Grinchuk Suite of the Malinovtsy Horizon, localities 6 and 7 of the stratigraphical excursion in Tsegelnyuk et al. (1983) (see also Balizoma obtusus below). The Podolian Pfidoli pygidium figured by Balashova (1968r/) as E. stubblefieldi lacks a sagittal band, and is not closely related to stubblefieldi, or to E. (E.) nasutus sp. nov. (= E. cf. pimctatiis sp. 3 of Schrank 1972, who referred Balashova’s specimen to that taxon). E. (E.) intersitus sp. nov. resembles E. (E.) stubblefieldi in several features. These include the wide cephalon, strong 1 L, glabellar shape, laterally set eyes, stout genal spines, very similar labral plate, long axial spine on the tenth thoracic segment, long mucro, and low R/P ratio. The shape of the free cheek and the tuberculation is, however, very different. The pattern of distribution (see text- fig. 9 and ‘Remarks on distribution’), together with the morphological similarities, suggest that stubblefieldi may be directly related to intersitus. No other described species appears to be closely related to stubblefieldi. Encrinurus (Encrinurus) sp. A. Plate 47, figs. 1, 2, 5, ?7; text-fig. 7 Material and localities. All specimens are from the Slite Beds, Slite Marl, undifferentiated but close to the transition Slite Marl/unit g. Boge parish— Mojner 3 (Ar52277-52279); Endre parish — Bolarve 1 (Ar51949), Hajdungs 1 (SGU 5051 and unnumbered SGU material); Follingbo parish— Follingbo 6 (Ar52494). In total ten pygidia and possibly one free cheek. Description. The material differs from E. (E.) pimctatus Form C in the following features: pygidial rachis wider, lateral margin almost straight in dorsal view; in lateral view curved upward at the sixth and seventh segments, base of mucro at a level above pygidial margin. Mucro circular in section, length unknown but at least half the length of rachis. R/P ratio 2T-2-2. Doublure narrow. The free cheek possibly belonging to this species (PI. 47, fig. 7) has a highly stalked eye and a precranidial lobe with sparse, very large tubercles. Large pygidia give an estimated total length for an extended exoskeleton of just over 70 mm. Discussion. This form occurs together with E. (E.) punctatus Form C at all the above localities, in proportions from 1:1 to LIO. Although dimorphism has been proposed for a related species, E. (E.) egatii Miller, 1880, this seems unlikely here. First, E. (E.) punctatus Form A and Form B are both monomorphic; secondly, E. {E.) sp. B is likely to be derived from E. {E.) sp. A, and the two forms appear to be a short side-branch from E. {E.) pimctatus. EXPLANATION OF PLATE 47 Figs. 1, 2, 5, ?7. Encrinurus (Encrinurus) sp. A. All Slite Beds, Slite Marl. Bolarve 1 (2), Hajdungs 1 (5), Mojner 3(1,7). 1 and 2, lateral and dorsal views of pygidium. la, b, Ar52279, x 5. 2a, b, Ar51949, x 2. 5, SGU 5051, dorsal view of pygidium, x3 (coll. G. Liljevall 1908). 7. Ar52277, exterior view of incomplete free cheek, x 3. Figs. 3, 6, ?8. E. (E.) sp. B. All Slite Beds, ?Slite Marl/unit g, Slitebrottet 2 (8), Slite ‘a’ c (3), unknown locality (6). 3n, b, Ar30749, dorsal and lateral views of pygidium, x3. ba d, Ar30713, lateral, posterior, dorsal, and ventral views of pygidium, x 2.5. 8, Ar52276, exterior view of free cheek lacking visual surface, x 2-5 (coll. Amelang 1983). Fig. Aa-c. E. (E.) sp. C. Ar51788, incomplete, partly disarticulated, enrolled specimen, Urgude 3, Hemse Marl NW; dorsal, oblique anterolateral, and exterior views, note thoracic axial spine, x 2-5. PLATE 47 RAMSKOLD, Encrimirus (Encrinurus) 558 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 7. A, B, dorsal and lateral views of reconstructed pygidia. a, Encrimirus (Eiicr munis) sp. A (based mainly on Ar52279), x 4. b, E. (E.) sp. B (based mainly on Ar3071 3 and Ar30749), x 3. E. (E.) sp. A closely resembles the American Wenlock E. (E.) egani Miller, 1880, redescribed by Holloway (1980). Similarities include the highly stalked eye and the long mucro with its base raised dorsally. However, E. (E.) sp. A is known too incompletely to be compared accurately with egani. Encrimirus (Encrimirus) sp. B Plate 47, figs. 3, 6, ?8; text-fig. 7 Material and localities. All specimens are from the Slite Beds, probably unit g, but possibly Slite Marl near the transition to unit g. Othem parish— slite ‘a’ c (Ar30749-30750); unknown locality (Ar30713). A free cheek (Ar52276), from Slitebrottet 2, Othem parish, possibly belongs to this species. Description. The three pygidia differ from E. (E.) sp. A mainly in having the base of the mucro set high above the level of the lateral pygidial margin. The seventh pleural pair is curved backwards to a horizontal position. The mucro itself is not preserved, but from its base it is clear that it must have been directed upwards at 45° or more. R/P ratio 2- 1-2-2. The ventral part of the border (PI. 47, fig. 6d) is very narrow. The free cheek probably belonging to this species (PI. 47, fig. 8) has a very highly stalked eye. Discussion. This species is probably slightly younger than E. (E.) sp. A; the occurrence in unit g of the Slite Beds is suggested by the stylolitic structures in one pygidium (PI. 47, fig. 6); such structures are common in unit g but rare in the Slite Marl in this area. The material shows a further development of the characteristic features of E. (E.) sp. A; the raised mucro and the highly stalked eye. It is likely that E. (E.) sp. B is a direct descendant from E. (E.) sp. A, but since both species are so incompletely known, new material must be collected to confirm this suggestion. E. (E.) sp. B is not as close to E. (E.) egani Miller, 1880 as E. (E.) sp. A (see discussion of the latter species), but the three species are more similar to each other than to any other Encrimirus, with the possible exception of E. (E.) odvaldensis sp. nov. Highly stalked eyes are common in trilobites, though unusual in Encrimirus outside the above group. A mucro directed upwards is not otherwise known in Encrimirus, but is present in some Phacopida, e.g. Chattiaspis (see Hammann 1974, pi. 7, fig. 107/)). Encrimirus (Encrimirus) sp. C Plate 45, figs. ?10 and ?1 1; Plate 47, fig. 4 Material and localities. Ar51788, cranidium with ten thoracic segments, from Urgude 3, Sproge parish, Hemse Beds, Hemse Marl NW. Two more specimens (Ar52280, free cheek; Ar52282, almost complete exoskeleton) from Urgude 4 are assigned questionably. Description. Differs from E. (E.) macroiiriis in the following features: cranidial tubercles very large, anterior tubercles higher than wide. Eight tubercles on anterior cranidial border. Tubercle pair VI- 1 absent. Tubercle RAMSKOLD: SILURIAN ENCRINURID TRILOBITES 559 iv-0 instead of tubercle pair iv-1. Seventh thoracic segment broken axially indicating axial spine, on tenth segment is a very long, stout spine; both segments with a tubercle above fulcrum (PI. 47, fig. 4c). Additional differences in the two questionably assigned specimens are: eyes high, with stalk (on free cheek) over 1 -5 times height of visual surface (PI. 45, fig. II); pygidium with broken mucro circular in section, possibly indicating long mucro. Discussion. These specimens represent a fundamental taxonomic problem. If the two questionably assigned specimens were the only known they would, perhaps questionably, be referred to E. {E.) macroiirus, which is common at both Urgude 3 and 4. However, the third specimen differs more from niacfourus than most species within E. (Encrinunis) differ from each other, and the two first specimens are definitely closer to this specimen than to macrourus. The possibility that the specimens are extreme, but not abnormal, variations within macrourus, cannot be excluded, but if this view is followed in encrinurid taxonomy, very few species can be defined accurately. At least until new material shows otherwise, E. (E.) sp. C is regarded as an independent species. No described Encrinunis is particularly close to E. (E.) sp. C, but some features (stalked eyes, long mucro?) are shared with E. {E.) spp. A and B of this paper. Subgenus Encrinurus (australurus) subgen. nov. Name. Latin australis, southern, and Greek aura, tail; referring to the numerous Australian species included. Type species. E. mitchelli Foerste, 1888, from the Yarwood Siltstone Member of the Black Bog Shale (Leintwar- dinian; Strusz 1980, p. 24), Yass Basin, New South Wales, and Coppins Crossing, Canberra, Australia. Other species. See Strusz 1980, p. 20 (E. mitchelli species-group). Diagnosis. IL well developed, only slightly smaller than 2L-4L. IS separate from occipital furrow. Preglabellar furrow always present but usually faint. Anterior border of cranidium with ten to eighteen tubercles, usually about fourteen. Four large tubercles on fixed cheek along rachial furrow. Eyes low, never stalked. Genal angles rounded or angular, never with spine. Labral plate with short posterior border. No thoracic spines. Pygidium large for genus, non-mucronate, with narrow rachis and 10--15* pleurae. Remarks. This group was discussed comprehensively by Strusz (1980), and although he did not recognize it as a distinct subgenus of Encrinurus, his definition and discussion cover all important aspects of E. {Australurus). This subgenus is geographically restricted to Australia, south-east Asia, and Japan. One species is upper Llandovery in age, the remainder are Wenlock and Ludlow. E. {Australurus) is compared with E. {Nucleurus) subgen. nov. under the latter. Subgenus encrinurus (nucleurus) subgen. nov. Name. Latin nucleus, core, and Greek oura, tail; referring to the central morphological position of this subgenus. Type species. E. ahyssalis Mannil, 1977 from the Kolka core, northwestern Latvia, Raikkiila Stage (middle Llandovery); known also from Gustavsvik near Motala, Ostergotland, south central Sweden, Klubbudden Stage (middle-upper Llandovery). Other species. E. anticostiensis Twenhofel, 1928; E. diaholus Tripp et al., 1977; E. elegantulus Billings, 1866; E. ine.xpectatiis Snajdr, 1978; E. kiltsiensis Rosenstein, 1941; E. mareki Snajdr, 1975; E. mullochensis Reed, 1931; E. palmrei Mannil, 1958; E. rotundus Mannil, 1977; E. rumhaensis Rosenstein, 1941; E. selistensis Mannil, 1977. These species are all Llandovery in age. The poorly known E. hypoleprus Steam, 1956 and E. nereus Hall, 1867 (both Wenlock) show similarities to the above group (i.a. in the R/P ratio), and are questionably assigned to E. {Nucleurus). Diagnosis. Glabella relatively short and wide. Anterior cranidial border with nine to fourteen tubercles, usually 10-12. Tubercles on fixed cheek along rachial furrow equal in size or slightly larger than other fixed cheek tubercles. IL usually small, ridge-shaped, set well below 2L, never tuberculate abaxially. Genal angle rounded or with a tiny spine. Labral plate with wide rhynchos 560 PALAEONTOLOGY, VOLUME 29 not reaching level anteriorly with labral suture, posterior border of small to medium length. No thoracic axial spines. Pygidium short and wide, non-mucronate. 7^-10^ pleural ribs. Posterior pleurae commonly merged to form a ‘loop’. R/P ratio 1-8-2-5. Discussion. Most of the species assigned to E. {Nucleiirus) were referred to Fragiscutum by Holloway (1980). Although there are certain similarities, especially in the pygidium, Fragiscutum is very different in the cephalon, particulary in the free cheek, and has only ten thoracic segments. The differences are regarded here as fundamental, and Fragiscutum is restricted to F. rhytium and F. glehalis. It is, however, not unlikely that E. (Nucleiirus) gave rise to Fragiscutum, or at least that the two groups share a common ancestry around the Ordovician/Silurian boundary. E. (Nucleiirus) differs from F. (Encrimiriis) mainly in consisting of non-mucronate species with a rather short, wide glabella, a short rhynchos and a wide, short pygidium. The subgenus is dis- tinguished from E. (Australiiriis) mainly by the small 1 L, the wide and short rather than elongated pygidium, and the 7^-10' pleurae compared to 10^-15‘. There is practically no overlap in time between the two subgenera, and they are completely separated geographically. It is therefore, at least at present, difficult to determine their relation to each other. Only the type species of this subgenus is known from Sweden, E. (Nucleiirus) ahyssalis Mannil, 1977; this will be described by Ramskold and Bassett (in prep.). Genus balizoma Holloway, 1980 Type species. Calymeiie variotaris Brongniart, 1822, from the Much Wenlock Lime.stone Formation (upper Wenlock) and lower Ludlow, Dudley, West Midlands, England; by original designation. Other species. E. hyperhoreiis Thomas in Thomas and Narbonne, 1979; E. iinlianensis Kindle and Breger, 1904; Cryptonynnis ohtusus Angelin, 1851 (including the junior subjective synonym E. ohtusiis erraticiis Schrank, 1972, and the possible synonyms E. diniitrovi Perry and Chatterton, 1979, E. rosensteinae Tripp et al., 1977, and Balizoma ilakon Snajdr, 1983); E. sp. of Tripp et al. 1977; B. sp. of Holloway 1980; E. aff. hyperhoreiis of Thomas in Thomas and Narbonne, 1979 (cranidium and free cheek only, the pygidium may belong to Eranvnia), and E. sp. cf. E. (Frammia) arcticus of Bolton 1965. Cronnis transiens Barrande, 1852 (see von Gaertner I930u, pi. 25, hgs. 9-10) has been studied from Bohemian material in British collections; this material belongs to Balizoma. E. siibvariolaris Munster, 1840 (see von Gaertner 1930«, pi. 25, fig. 6) probably also belongs to Balizoma. E. spp. F and G of Lane (1979) are poorly known but show certain similarities to Balizoma. Diagnosis. Tubercles ii-1, II-Or, 1, iii-0 form a pentagon with tubercle ITOr, when present, in centre. Tubercles on 2L, 3L, 4L, and PL commonly displaced adaxially. Anterior border of cranidium with ten to twelve tubercles. Genal angle typically rounded, tiny spine may be present. Field of cheek thickly covered with tubercles. Pygidium with seven to fifteen rings and six to twelve pleurae; one to two fewer pleurae than rings, R/P ratio M-L4. Sagittal groove present, rachial rings commonly with a pair of tubercles flanking the groove, several axial tubercles. Tip non-mucronate, rounded in both dorsal and lateral views. Discussion. In addition to the diagnostic features listed above, all except the type species have non- stalked eyes and a very low field of free cheek. Some cephalic features considered diagnostic for Balizoma by Holloway (1980) are also found in other taxa within the "variolaris plexus’ (of Strusz 1980), e.g. in Fragiscutum Whittington and Campbell, 1967, so the diagnostic emphasis must be on pygidial features, as also noted by Holloway. A very useful criterion is the R/P ratio (defined under ‘Notes on morphology’ above) of 1 • 1 -1 -4, but no single feature is completely restricted to Balizoma, which is therefore distinguished by the unique combination of characters. Thomas (1981, p. 64) preferred not to use Balizoma although ‘accepting that generic status may eventually be justified’, while Snajdr (1983) recognized the genus and assigned certain Bohemian species to it. Balizoma is conceived here in a more restricted sense than by either Holloway or Snajdr. Species assigned by them to Balizoma but excluded here are: E. inexpectatiis Snajdr, 1978; RAMSKOLD; SILURIAN EN C R I N U R I D T R I LO B ITES 561 E. mareki Snaidr, 1975; E. nereus Hall, 1867; E. suhvariolaris concomitcms Pfibyl and Vanek, 1962; E. teslosteron Snajdr, 1981; and E. tuherculifrons Weller, 1907. All these species have a much higher R/P ratio than Balizoma. The very similar Bohemian Pfidoli concomitcms and testosicron may be derived from Balizomo, but they also show similarities to Fragiscutum, and their position is uncer- tain. The American Wenlock species nereus and tuberciilifrons are both in need of revision. E. tuherculifrons was referred hesitantly to Encrinuroides by Strusz (1980); photographs of the type specimens support that view. Encrinurus nereus seems to be close to the large group of middle and upper Llandovery species within the "variolaris plexus’ here formalized into E. (Nucleurus) subgen. nov. Balizoma and Eragiscufum are apparently related closely to E. (Nucleurus) and could have their origins in lower Llandovery species similar to E. rotundus Mannil, 1977, the earliest of the species assigned to E. (Nucleurus). That species and several others here included in E. (Nucleurus) were referred to Eragiscutum by Holloway (1980). There are certainly similarities, but at present it seems best to restrict Eragiscutum to E. rhytium Whittington and Campbell, 1967 and F. glehalis Campbell, 1967 (see discussion of E. (Nucleurus) above. The late Llandovery-early Wenlock Perryus Gass and Mikulic, 1982, including P. severnensis Gass and Mikulic, 1982 (= E. sp. I of Norford 1981) and possibly E. glohosus Makisimova, 1962, is more distantly related to the other species discussed above, but may be closer to the "variolaris plexus’ than to the Encrinuroides j Cromus line which was suggested by Gass and Mikulic (1982). Two members of the "variolaris plexus’ are known from Sweden: B. ohtusus (Angelin, 1851) from Gotland, and E. (Nucleurus) ahyssalis Mannil, 1977 from Ostergotland (will be described in Ramskold and Bassett (in prep.)). Balizoma ohtusus (Angelin, 1851) Plate 40, tig. 2; Plate 48, tigs. 1-14; Plate 49, tigs. 110 * 1851 V non 1901 v. 1972 . 1972 . 1972 v? 1977 ? 1979 ? 1983 Cryptonynius ohtusus Angelin, p. 3, pi. 4, hg. 9. Encrinurus ohtusus'. Lindstrom, p. 56, pi. 4, figs. 14 and 15 [= E. (E.) intersitus sp. nov. or E. (E.) stuhblefieldiTu^yp. 1962]. Encrinurus (Franimia) o. ohtusus (Angelin, 1851); Schrank, p. 45, pi. 13, fig. 3 [with synonymy list]. Ecrinurus (Eramniia) ohtusus erraticus Schrank, p. 45, pi. 13, tigs. 4-7; pi. 14, figs. 1-5 [with synonymy list], Encrinurus cf punctatus 3; Schrank, pi. 11, fig. 8, non figs. 9 and 10; pi. 12, figs. I 3 [= E. (E.) nasutus sp. nov.]. Encrinurus rosensteinae Tripp et ah. p. 860, pi. 1 5, figs. 1 1 3; text-fig. 3c. Encrinurus dimitrovi Perry and Chatterton, p. 589, pi. 72, figs. 1-3; pi. 73, tigs. 1-17, 29-31; pi. 74, figs. 1-14, 18-23, 30-35 [with synonymy list]. Balizoma dakon Snajdr, p. 175, pi. I, fig. 2. Neotype. Ar30309, almost complete exoskeleton from Hassle, Ostergarn parish, Hemse Beds, unit c (or d), selected and figured by Schrank 1972, pi. 13, fig. 3-36. Additional material. See under each form and ‘Discussion’ below. Diagnosis. Anterior border of cranidium with ten to twelve tubercles. I L non-tuberculate. One or two tubercles between palpebral lobe and axial furrow. Field of free cheek with seven to thirteen tubercles. Pygidium with width to length L2 ; 1-1 -6 ; 1. Ten to fourteen axial rings, 8^-1 H pleurae. Axial tuberculation varying from tubercles on most rings to on about every second. Remarks. The three slightly different Gotland forms are treated here as one species and are diagnosed as such. The resulting diagnosis is not precise enough to separate ohtusus from dimitrovi, rosen- steinae, and dakon, and the possible synonymy of these species with ohtusus is discussed below. Description. Eonn A: the ‘type form’. Restricted to the northeastern limestone areas in the Hemse Beds, and the Eke Beds outlier at Lau Backar. The neotype is from the Ostergarn area, Angelin's type locality, where only Form A is found. 562 PALAEONTOLOGY, VOLUME 29 Material and localities. Five nearly complete exoskeletons, five free cheeks, and over thirty pygidia (all from Hemse Beds), plus two glabellae, one free cheek, and about ten pygidia from Lau Backar 1. Hemse Beds, unit b ore: Ostergarn parish— ditch 1 km north-west of Katthammarsvik (probably = Angmans 1). Hemse Beds, units c and d: Ardre parish — Botvalde, Kaupungs 3, Rudvier 1; Kraklingbo parish-field east of road 200 m north of Osterby; Ostergarn parish -Briinnklint, Grogarnshuvud 1, Gutenviks I, Hammarudden 1, Kuppen, Kyrkberget, ditch at Herrvik. Hemse Beds, upper part: Gammelgarn parish — Herrgardsklint 1, Millklint (or Torsburgen). Eke Beds, upper part: Lau parish— Lau Backar 1. A few poorly preserved specimens from the following more southwesterly localities may also belong to Lorm A: Garde parish — large ditch 1 km south- east of Garde church. Lye parish -Mannegarda. Fardhem parish— Sandarve kulle (all Hemse Beds, upper part). Comparative description. Density and number of glabellar tubercles variable (compare PI. 48, fig. 3 with Schrank 1972, pi. 13, fig. 3). Eye small, occupying less than half the length (exsag.) of field of cheek (PI. 48, fig. 6a). Part of fixed cheek between posterior border furrow and posterior branch of facial suture as wide as or wider (exsag.) than posterior border. Lield of free cheek with eight to thirteen tubercles. Lateral border furrow meets posterior branch of facial suture closer to margin than to eye. Genal angle with tiny spine in small specimens, rounded in larger ones. Posterior borders and occipital ring each with several faint, perforate tubercles (PI. 48, fig. 4); all thoracic segments with perforations in corresponding places. Pygidium with width to length 1-3-1 -4 : 1, 10'-11‘ pleurae, rarely up to 12'. Eleven to fourteen rachial rings, usually thirteen recognizable, occasionally there is a terminal piece composed of several ‘segments’ (PI. 48, fig. 9c). Pleural regions high dorsoventrally (see posterior views PI. 48). Dorsoposterior part evenly sloping in lateral view, posteriormost point usually well below last rachial ring. Posteriormost paired pleural furrows form a three- quarter circle open upwards, not reaching pygidial margin (PI. 48, figs. 2fi, lb, 8fi, 9h, lOfi). Pleurae rather flat-topped, the plane formed by each ridge only slightly inclined to give a very weak imbricating appearance of the pleural areas. Pleural furrows usually very narrow. Axial tuberculation very variable, from a distinct tubercle on most rings to tubercles on about every second ring starting from fourth or fifth ring; there are commonly small, faint tubercles also on many ‘blank’ rings. Rachial rings and pleurae with faint tubercles or perforations in corresponding positions to those on thorax. Form B. Restricted to the ‘marl’ west and south of the southwestern outcrops of the Hemse Beds limestone area. Material and localities. Two nearly complete exoskeletons, two complete and four incomplete cranidia, four free cheeks, and about fifty pygidia. Klinteberg Beds, Klinteberg Marl: Gerum parish— Ajmunde 1. Hemse Beds, Hemse Marl NW: Lardhem parish— Gardarve 1, Gerete 1, Burge sawmill; Hablingbo parish— Lilia Hallvards 5; Hemse parish— east of road south-east of Niksarve (in Lardhem parish); Silte parish— Mastermyr 1. Smissarve 1. Hemse Marl SE: Burs parish— Rone drainage ditch at Vastlaus (probably = Viistlaus 1); Hemse parish -Hulte 3, Hemse church. Two pygidia from Loggarve 2 (Klinte parish, Mulde Beds, upper- most part) apparently belong to Lorm B; this stratigraphically aberrant occurrence is discussed under ‘Remarks on distribution’ below. EXPLANATION OF PLATE 48 Ligs. 1 and 11. Balizoma obtusns (Angelin, 1851) Lorm A (or C?). Both from upper Eke Beds, Lau Backar 1. 1, Ar52346, exterior view of labral plate, x6. 1 1, Ar52345, dorsal view of incomplete cranidium, x 4. Figs. 2-10. B. obtusus (Angelin, 1851) Form A. Hemse Beds, unit c (2, 3, ?4, 5, ?6, ?8, 10), Eke Beds, upper part (7, 9). Briinnklint (5), Lau Backar 1 (7, 9), ‘Ostergarn’ (2, 3, 10), ditch 1 km north-west of Katthammarsvik (4, 8), unknown locality (6). 2a-c, Ar30356, lateral, posterior, and dorsal views of pygidium, x 3. 3, Ar30359, dorsal view of cranidium of almost complete specimen, (note rudimentary genal ‘spine’), x4. 4, Ar51772, exterior view of free cheek, x 5. 5, Ar30352, dorsal view of fairly complete specimen, x4 (coll. G. Lindstrom). 6u, b, Ar52344, dorsal and lateral views of complete exoskeleton x4. la, b, Ar51768, dorsal and posterior views of pygidium x 5. Sa-c, Ar51773, dorsal, posterior, and lateral views of pygidium, x4. 9a-c, Ar51767, dorsal and posterior views and detail of rachis of pygidium, a, h, x4, c, x 12. 10r;-c, Ar30339, dorsal, posterior, and lateral views of pygidium, x 5. Ligs. 12- 14. B. obtusus (Angelin, 1851) Form C. All from upper Eke Beds, Lau Backar 1. \2a-c, Ar51768, dorsal, posterior, and ventral views of largest pygidium, x4. 13 and 14, dorsal view of pygidium. 13, Ar52347, x 6. 14, Ar52348, x 6. PLATE 48 RAMSKOLD, Balizoma obtusus 564 PALAEONTOLOGY, VOLUME 29 Comparative description. Glabella similar to Form A. Eye of two types; either small (similar to Form A) or large (see ‘Discussion’ below). Part of fixed cheek between posterior border furrow and posterior branch of facial suture usually narrower (exsag.) than posterior border. Field of free cheek with seven to eleven tubercles (PI. 49, figs. 2 and 7). Lateral border furrow meets posterior branch of facial suture closer to eye than to margin. Tubercles and perforations on occipital ring and posterior border as in Form A. Pygidium with width to length 1 -2-1 -4 ; I . Usually 10' pleurae (51 %), range 9' -10^. Ten to eleven rachial rings, one large specimen has twelve rings. Pleural regions lower than in Form A; most specimens are. however, slightly distorted. Dorsoposterior part in lateral view commonly angular and bent down just posterior to the last rachial ring, posteriormost point commonly at this Ilexure, overhanging postrachial pleural region. Posteriormost paired pleural furrows as in Form A. The plane formed by the outer surface of each pleural rib is strongly inclined anteroventrally from the general surface plane of the pygidium. Pleural furrows fairly wide. Axial tubercles on every second ring on average, commonly starting on fourth ring. Weak tubercles and perforations present on rachial rings and pleurae at least to some extent. Form C. Only known from pygidia from the upper Eke Beds at Lau Backar I, Lau parish, where Form A is also found (more commonly). About five pygidia, of smaller size than most of Form A and Form B. A very small incomplete, enrolled specimen possibly belongs to Form C. Two glabellae, one free cheek, and two labial plates from Lau Backar 1 are assigned to Form A, but the possibility that the two forms found together have similar cephala cannot be excluded, and these specimens may belong to either one of the forms. Comparative description. Pygidium with width to length I -6 ; 1,9' pleurae, eleven rachial rings. Pleural regions low dorsoventrally; possibly lower than in Form B, but most specimens are slightly distorted. Dorsoposterior part gently sloping in lateral view, posteriormost point at border flexure. Posteriormost paired pleural furrows subparallel, reaching posterior margin (PI. 48, hg. \2h). Outer surface of each pleura gently convex (exsag.). Pleural furrows wide. Axial tubercles, weak pleural and rachial tubercles, and perforations similar to Form A. Discussion. It is probably realistic to assume that the three Gotland forms of B. obtusus really represent more than one biological species. Flowever, in spite of large samples and excellent preser- vation, the extensive variation in morphology makes the forms difficult to separate, and Form C is known only from pygidia. Splitting into several species or subspecies is therefore undesirable, at least for the present. Accordingly, the three forms are treated here as morphological forms of obtusus, but to avoid loss of information, they are described and figured as separate entities. Form B is probably slightly older than the Hemse Beds material of Form A, although some overlap in time cannot be ruled out. The differences between the ’limestone’ specimens, Form A, and the ‘marl’ material. Form B, are small but real. Form A specimens are generally better preserved, which may account for part of the differences in pygidial tuberculation, since surface details in Form B are more difficult to observe. However, there are at least two definite differences: 1, the number of pygidial pleurae is clearly different (Table 2). Size has very weak or no correlation with the number of pleurae (text-fig. 8), and the mean size of the two forms is about the same. 2, the pygidial pleural regions have a greater height in Form A, resulting in a more vaulted pygidium than the rather low Form B. The size of the eyes and the field of cheek are also different, but more difficult to evaluate. In Form A the size of the eyes and field of cheek does not vary, while there are EXPLANATION OF PLATE 49 Figs. 1-10. Balizoma obtusus (Angelin, 1851) Form B. Hemse Beds, Hemse Marl NW (2-5, 7), Hemse Marl SE (1, 6, 8-10). Gardarve 1 (2), Hemse church (9), Hulte 3 (1, 6), Mastermyr 1 (4, 5), Viistlaus I (8, 10), Visne myrs kanal (7), unknown locality (3). \a-c, Ar52335, oblique anterolateral, lateral, and dorsal views of cranidium, x 3. 2, Ar52340, exterior view of free cheek, x 3. 3f/-r/, Ar30381, anterior, lateral, and dorsal views and pygidium of enrolled specimen, x 5. 4, Ar52342, dorsal view of large incomplete cranidium, x 3 (coll. Amelang 1982). 5u-c, Ar52343, posterior, lateral, and dorsal views of pygidium of almost complete specimen, x3 (coll. Amelang 1982). 6a, b, Ar52337, posterior and dorsal views of pygidium, x 3-5. 7, Ar52338, exterior view of free cheek, x 6. 8u-c, Ar31488, dorsal, lateral, and ventral views of pygidium, X 4 (coll. G. Holm 1904). 9u-c, Ar303 15, dorsal, lateral, and posterior views of pygidium, x4. 10, Ar51770, dorsal view of pygidium, x 4 (coll. G. Holm 1904). PLATE 49 RAMSKOLD, Balizoma ohtusus 566 PALAEONTOLOGY, VOLUME 29 TABLE 2. Number of pygidial pleurae in the different forms of Balizoma obtusus. The most important localities are listed separately. A two-sided t-test of significance of the mean values has been used to determine differences between certain groups (9‘ is given the value 9, 9^ = 9-5 etc.). The two Hemse Beds forms (II and IV) give a t = 7-71***^ which confirms their dissimilarity in this particular character. Within Form A, Hemse Beds (II) and Eke Beds (III) specimens differ slightly (t = 2-29*fi and the Eke Beds material is actually closer to Form B (IV) in this feature (t = 1-70). Within Form B, Hulte 3 specimens commonly have 10' pleurae (PI. 49, fig. 6) and Mastermyr 1 specimens usually have 9^ (PI. 49, fig. 5), but this difference is not statistically significant (t = 1-65). The diagram obviously suggests that Form C (VII) is simply one extreme of the variation range of the other Lau Backar 1 material (III). However, this solution seems to be in conflict with other morphological differences, although most of the Form C material is distorted and incomplete, and the evidence is not conclusive. (One, two, and three asterisks indicate significance at the 5 %, 1 %, and 01 % level, respectively.) 9' 92 10' 102 11' IF 12' m s N I Form A total 5 12 9 9 2 1 10-42 0-62 38 II Hemse Beds — — 5 8 7 2 1 10-69 0-54 23 III Lau Backar 1 — 5 7 1 2 -- - 10-00 0-50 15 IV Form B total 2 13 18 2 — — — 9-79 0-35 35 V Hulte 3 — 2 12 1 — — — 9-97 0-23 15 VI Mastermyr 1 — 8 3 1 — — — 9-75 0-45 12 VII Form C 4 — — — — — — 9-00 0-00 4 two types in Form B; one with very large eye and small field of cheek (PI. 49, figs. 3 and 7), and one with smaller, taller eyes and relatively larger area of field of cheek (PI. 40, fig. 2; PI. 49, fig. 2). Form C is distinguished mainly by features in the posterior part of the pygidium, and may possibly be one end of the morphological variation within Form A (see Table 2). Several species and subspecies with a morphology very close to obtusus have been described recently. Schrank ( 1 972) erected Encrinurus {Frammia) obtusus erraticus for Ludlow specimens from German erratics (originating from the Baltic region). These specimens (regarded as a separate species by Tripp et al. 1977 and by following workers) fall well within the morphological range of the Gotland obtusus Form A material, and erraticus is treated here as a junior subjective synonym of obtusus. Specimens assignable to obtusus are also known from the Podolian Ludlow. The material, collected and kindly lent by Dr R. M. Owens, consists of about ten pygidia and two free cheeks from the Grinchuk Suite of the Malinovtsy Horizon, excursion localities 6 and 7 of Tsegelnyuk et al. (1983). Of seven pygidia there are five with 9- pleurae, and one each with 10* and 10^ (compare with Mastermyr 1 in Table 2). The available material is indistinguishable from obtusus Form B. One similar pygidium is from the Sokol Suite, locality 5. The British Ludlow specimens described by Tripp et al. (1977) as E. rosensteinae are more difficult to compare with the Gotland material. The holotype pygidium of rosensteinae has been studied and is slightly distorted, but very similar to obtusus, being somewhat intermediate between Form A and Form B, except for being more elongate, as are the other figured specimens apart from the complete exoskeleton (Tripp et al. 1977, pi. 115, figs. 1, 2, 7). This specimen, and other unfigured British material (e.g. BM I. 6197; see Tripp et al. 1977, p. 861), is similar to Gotland Form B specimens, and can be confidently assigned to obtusus. It is clear that the British material consists of two slightly different types. The only significant difference between rosensteinae (holotype and other elongate pygidia) and obtusus is the slightly higher R/P ratio in the former (confirmed in British material kindly loaned by Mr Stephen Tunnicliff, Nottingham). The possible synonymy of rosen- steinae and obtusus can be confirmed only by further study of the British material. E. dimitrovi Perry and Chatterton, 1979 (including E. (Fragiscutum) sp. of Perry and Chatterton 1977) from the Canadian upper Wenlock or lower Ludlow is even closer to obtusus than rosensteinae. Synonymizing of the two species here is only prevented by the difficulty in assigning the Canadian material to either one of Form A or Form B. Pygidia of dimitrovi have a pleural height intermediate RAMSKOLD: SILURIAN ENC RIN U RI D T R I LO B ITES 567 pleurae 10^ o 10'- o oo oooooo oo o 9^ o o V— 1 ' 1 ' 1 1 1 ' 1 ' 1 ' 1 ' 1 5 6 7 8 9 10 11 12 mm TEXT-FIG. 8. Number of pygidial pleurae plotted against pygidial length (articulating half-ring excluded) in BaHzoma obtusus Form B. There is no correlation between size and number of pleurae. All specimens from Hulte 3 (Ar52337, Ar52423-52436). between Form A and Form B, while cheeks and eyes are more similar to Form A. Canadian material similar to obtusus, from Avalanche Lake, is currently under study by Dr Brian Chatterton, Edmonton. The material seems to show weak dimorphism in the pygidium, with one form longer, higher and with slightly wider pleural furrows (Chatterton, pers. comm.). The only figured pygidium of B. dakon Snajdr, 1983 from the Bohemian Ludlow is indistinguishable from Form A of obtusus, but no other parts have yet been figured, and the short description of dakon indicates that there may be differences in the cephalon. The Estonian Ludlow material mentioned by Nieszkowski (1859, p. 377), Schmidt (1859, p. 448; 1881, p. 224), Twenhofel (1916, p. 330), and Mannil (1982u, p. 65) has not been figured, and the assignment to obtusus cannot be conhrmed. The American Wenlock E. indianensis Kindle and Breger, 1904 is remarkably similar to obtusus (photographs of the type material kindly supplied by Dr Philip Lane, Keele; specimen figured by Levi-Setti 1975 not conspecihc). The cranidium has the same proportions, size and position of tubercles and lateral glabellar lobes, eye position and shape of the genal angle. The holotype pygidium has 1 1 ' pleurae, hfteen rachial rings and a sagittal groove with axial tubercles on most rings, and differs from obtusus only in the more elongate shape, resembling rosensteimie. However, there may be additional differences since indianensis is preserved as internal moulds, and the free cheek is unknown. Differences between B. variolaris and B. obtusus were listed by Tripp et al. (1977, table 5), and although obtusus is interpreted in a wider sense here, the differences are obvious. B. Iiyperboreus (Thomas in Thomas and Narbonne 1979) has a pygidium similar to obtusus Form A, but is distinguished by a tuberculate IL, denser glabellar tuberculation, and eyes set further from the rachial furrows. REMARKS ON DISTRIBUTION Upper Visby Beds. At Halls Huk 1 E. (E.) punctatus Form A occurs on the same bedding planes as E. {E.) laevis (Angelin, 1851; this species will be described separately elsewhere). The exact level of this co-occurrence is uncertain, but very close to the Upper Visby Beds/Hogklint Beds boundary. The first appearance of punctatus, and the last of laevis, is therefore either in the top Upper Visby Beds or lowest Hogklint Beds. Hogklint Beds. E. {E.) punctatus Form A is abundant in all units of this division. Tofta Beds. No encrinurids have been collected in the Tofta Beds. Slite Beds. From units a to e only poor material is known, apparently belonging to E. (E.) punctatus. In the oldest, northwestern area of the Slite marl, E. (E.) punctatus Form A makes its last appearance (Valve 2). In the remaining part of the Slite Marl E. [E.) punctatus Form C is abundant. E. (E.) sp. A occurs together with, but less commonly than, E. (E.) punctatus Form C in Slite Marl areas bordering to unit g (Follingbo 6, Hajdungs 1, Bolarve 1, Mojner 3). It is probable that E. {E.) sp. B 568 PALAEONTOLOGY, VOLUME 29 occurs slightly higher stratigraphically in the transition from Slite Marl to unit g. In the youngest, southwestern area of the Slite Marl, in the ‘Pentamerus gothlandicus Beds’, E. (E.) punctatus is replaced by E. {E.) odvaldensis (Odvalds 1, Svarvare 1, Robbjans 1, 2). Halla Beds. Poor material from unit b (Horsne 5) seems to be close to E. {E.) macrounis. Midde Beds. E. (E.) macrourus makes its first appearance in the lower part (Blahall 1), and ranges through the division. In the uppermost Mulde Beds at Loggarve 2 B. ohtiisus occurs. This locality, on the edge of the Klinteberg Beds limestone area, may represent an environment similar to the near-reef Hemse Beds, where B. ohtusus is present. The absence of ohtusus in the Klinteberg Beds (and Mulde Beds) between Loggarve 2 and the Klinteberg Marl (Ajmunde 1) may thus simply reflect the absence of exposed near-reef strata in this area. No trilobites have been found in the southeastern area of the Mulde Beds. Klinteberg Beds. Well-preserved trilobites are known only from Ajmunde 1 in the southwestern, marly part, originally mapped as Klinteberg limestone by Hede (1927); termed Klinteberg Marl by Laufeld (I974u). .leppsson (1983) regarded this locality as belonging to the Hemse Marl NW on conodont evidence, a view supported by the trilobite distribution. At Ajmunde 1 E. (E.) cf. intersitus is abundant, and co-occurs with the less common B. ohtusus. Hemse Beds. This division consists of two, partly synchronous, main facies types; first, the northeast- ern outcrops of reefs and thick-bedded limestones; secondly, the more argillaceous, thin-bedded facies in the west and south (see text-fig. 9), divided into Hemse Marl NW and Hemse Marl SE. The faunal succession along the western coast is: 1. E. {E.) macrourus occurs north-west of a straight line approximately from Vakten (south of Petesvik) to Fardhem church. It co-occurs with E. (E.) sp. C at Urgude 3 and ?4. 2. Co-occurrence of macrourus and B. ohtusus close to the above line (Lilia Hallvards 5, Smissarve 1, Mastermyr I ). 3. Replacement without overlap by E. (E.) intersitus (the possible co-occurrence of macrourus and intersitus at Mastermyr 1 is probably due to the relatively large sequence exposed, 2 or 3 m, which probably covers the level where the shift takes place). 4. Replacement by E. {E.) stuhhiefieldi, found at Klasard 1 and Vaktard 2. This occurrence agrees with the view of Jeppsson (1983, p. 133) that these localities represent a westerly continuation of the Hemse Marl SE. In the uppermost part of the Hemse Beds in this area, at Bodudd 1, encrinurids are apparently absent. Inland of the above succession E. {E.) cf. intersitus co-occurs with B. ohtusus in the Hemse Marl NW (Gardarve 1, Gerete 1, Amlings 1). E. (E.) cf. intersitus is also present in the strata immediately underlying the southwestern outcrops of the limestone area at Rangsarve 1, Ase 1, and Klints 1. According to Jeppsson (1983) at least the first of these localities should be referred to the Hemse Marl SE. The localities are situated at an elevation of 40 m and the southeasterly dip of the strata indicates an approximate time-equivalence with the more low-lying main area of the Hemse Marl SE. In the main area of the Hemse Marl SE E. (E.) stuhhiefieldi is abundant until disappearing at, or slightly below, the Eke Beds boundary. It first co-occurs with intersitus and B. ohtusus (Hulte 3, Vastlaus 1) in ratios of about 3:3: 1. Specimens referred to here as cf. intersitus co-occur with stuhhiefieldi at some localities in beds close to the Eke Beds boundary (within 1 or 2 m at most; Hagvide 3, Nyan 3, Ondarve 1, Sigdes kanal). E. (E.) jarkanderi is restricted to unit b in the northernmost outcrops of the Hemse Beds. B. ohtusus is characteristic of unit c of the limestone sequence. In the uppermost part of this sequence, unit e, specimens probably belonging to E. {E.) nasutus occur. Eke Beds. Encrinurids are as yet only known from the Lau outlier. At Gannor 1 E. (E.) nasutus occurs both in the reefs and in layer d (of Hede in Munthe et al. 1925). The species continues into the upper Eke Beds at Lau Backar 1, where it is the commonest trilobite and co-occurs with B. ohtusus. No encrinurids have been encountered in younger strata on Gotland. RAMSKOLD: SILURIAN ENCRINURIDTRILO BITES 569 TKXT-FiG. 9. Distribution of encrinurids in the Mulde, Klinteberg, and Hemse Beds, a-e indicate localities yielding the specimens with corresponding letters in text-figs. 5 and 6. The boundaries are those of Hede (1942). 570 PALAEONTOLOGY, VOLUME 29 Discussion. It is clear that the species have a markedly restricted vertical distribution, even more so when the different forms (‘populations’) are considered. This reflects either a restricted temporal span or facies dependence (the latter has been discussed extensively by e.g. Chlupac 1977 and Lane 1972). The frequent horizontal and vertical facies changes in the Gotland sequence would strongly restrict the distribution of facies dependent species. When major facies changes are absent, as on the western coast from the lower Mulde Beds to well into the Hemse Beds, a series of populations forms a morphological dine through the sequence {E. (E.) inacrourus). Such a dine makes the definition of the range of a particular species or chronosubspecies within the morphological series impossible or arbitrary. Acknowledgements. I am grateful to Professor V. Jaanusson for advice and encouragement throughout this study, and to Dr M. G. Bassett for fruitful discussions at several stages. Preliminary drafts of the manuscript were read critically also by Drs B. D. E. Chatterton, D. J. Siveter, and R. Tripp. Loans of specimens were arranged by Dr S. Laufeld (Museum of the Geological Survey of Sweden), Ms K. Lindholm (Department of Historical Geology and Palaeontology, University of Lund), and Mr S. Tunniclitf (British Geological Survey). Access to Australian collections was arranged by Drs K. S. W. Campbell (Australian National University, Can- berra), P. A. Jell (National Museum of Victoria, Melbourne), A. Ritchie (The Australian Museum, Sydney), and B. D. Webby (University of Sydney). Mr W. Amelang kindly donated specimens from his private collection. Dr P. D. Lane supplied photographs of American type material, and Dr R. M. Owens loaned Podolian material. Text-fig. 7 was made by Mr L. Andersson. Field-work was sponsored by the Swedish Museum of Natural His- tory. Grants from Langman’s Fund and Erik Stensio Palaeozoology Fund defrayed some publication costs. APPENDIX Locality data Locality names followed by numbers are defined according to the system of reference localities introduced by Laufeld (1974h) for Gotland, and are described in that study or in Larsson (1979), Jeppsson (1982), Cherns (1983), and Ramskold (1983, 1984). Other localities are defined by twelve figure grid references according to the Swedish National Grid (Rikets Nat) system whenever possible. Nineteen new localities are described here. HAGSARVE 1,634947 164391 (CJ 3048 4872), Sproge parish, c. 4150 m north-west of Silte church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Ditch exposure immediately south of the road junction. Hemse Beds, Hemse Marl NW. HAGSARVE 2, 634861 164415 (CJ 3066 4787), Sproge parish, c. 3400 m north-west of Silte church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Shallow ditch west of the road, running south-west from the road and along the side of the wood. Hagsarve 2 comprises the entire length of the ditch (c. 100 m). Hemse Beds, Hemse Marl NW. HAGSARVE 4, 634842 1 64430 (CJ 3080 4765), Sproge parish, c. 3200 m north-west of Silte church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Ditch exposure 0-50 m east of road/ditch intersection 900 m south-south-east of the west house at Hagsarve. Hemse Beds, Hemse Marl NW. HU1.TE 3, 634725 165655 (CJ 4294 4557), Hemse parish, c. 1750 m south-east of Hemse church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Ditch section c. 150 m north-north-west of main road. Comprises from the corner of the ditch and 100 m toward south-west. Material excavated from the ditch is dumped 100 m south of the road at 634697 165651. Hemse Beds, Hemse Marl SE. RAUPUNGS 3, 636358 167350 (CJ 6106 6054), Ardre parish, c. 1725 m south-south-west of Ardre church. Topographical map sheet 6 J Roma SV. Geological map sheet Aa 170 Katthammarsvik. Section immediately west of new road across Kaupungsklint, 0-100 m from junction with road to Ardre abandoned church (‘odekyrka’). Hemse Beds, units c and d. KLiNTS 1, 635622 165540 (CJ 4243 5457), Lojsta parish, c. 850 m south of Lojsta church. Topographical map sheet 6 J Roma SV. Geological map sheet Aa 164 Hemse. Temporary exposure during well construction at Klints. The excavated material is dumped 250 m north of Klints, at 635653 165537 (CJ 4245 5491). Hemse Beds, upper? part, and Hemse Marl SE? RAMSKOLD: SILURIAN ENCRINURID TRILOBITES 571 LEViDE I, 635305 164753 (CJ 3440 5201 ), Levide parish, c. 940 m west-south-west of Levide church. Topogra- phical map sheet 6 I Visby SO. Geological map sheet Aa 164 Hemse. Shallow ditch running north-north-west from the road, north of the road. The ditch is c. 30 m long. Levide 1 is located just north-east of the road junction. Hemse Beds, Hemse Marl NW. LiLLA HALLVARDS 5, 634202 1 64400 (CJ 3002 4131 ), Hablingbo parish, c. 4850 m west-south-west of Hablingbo church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Temporary exposure 5 m north of road, in excavation parallel to road, 400-600 m east-north-east of road junction at Lilia Hallvards. Hemse Beds, Hemse Marl NW. LILLA HALLVARDS 6, 634221 164445 (CJ 3048 4146), Hablingbo parish, c. 41 50 m west-south-west of Hablingbo church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Small exposure immediately south of the road to Lilia Hallvards in ditch perpendicular to road, in the small glade, c. 2500 m west-south-west of junction with main road. Hemse Beds, Hemse Marl NW. LILLA HALLVARDS 7, 634278 164562 (CJ 3168 4194), Hablingbo parish, c. 2950 m west of Hablingbo church. Topographical map sheet 5 1 Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Temporary exposure in pit for a metal post, immediately south of road to Lilia Hallvards, c-. 1225 m west- south-west of junction with main road. Hemse Beds, Hemse Marl NW. LUKSE 2, 634287 164745 (CJ 3354 4190), Hablingbo parish, c. 1 100 m west of Hablingbo church. Topographical map sheet 5 I Hoburgen NO & 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Exposure in ditch running north-east to south-west, partly along the woods edge, 0-100 m from intersection with main ditch running south-east to north-west. Hemse Beds, Hemse Marl NW. MASTERMYR I, 634828 164873 (CJ 3520 4720), Silte parish, c. 2575 m north-east of Silte church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Drainage ditch section from the intersection of the smaller ditch from north-east, and 400 m toward south-east to the wooden bridge 475 m north-west of point 14.30. Hemse Beds, Hemse Marl NW. MiCKELS 1. 634590 164641 (CJ 3270 4501), Silte parish, c. 750 m south-west of Silte church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Exposure 5-25 m south of road in ditch running north-south, c. 250 m east of road junction at Hallvide. Hemse Beds, Hemse Marl NW. ODVALDS 1, 636442 164518 (CJ 3291 6352), Klinte parish, c. 875 m west-north-west ofKlinte church. Topogra- phical map sheet 6 I Visby SO. Geological map sheet Aa 160 Klintehamn. Temporary exposure in excavation during construction of road running north-north-east from the main road c. 1 50 m west-south-west of the house at Odvalds north of the road. The excavation continued c. 100 m toward north-north-east and then 50 m west-north-west. Slite Beds, Slite Marl, ‘Pentamerus gothlandicus Beds’. ONDARVE 1, 634917 167234 (CJ 5877 4626), Niir parish, c. 3175 m south-east of Niir church. Topographical map sheet 5 1 Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 156 Ronehamn. Low section in ditch running north-west to south-east along west side of road, 0-100 m from road/ditch intersection c. 1275 m east-south-east of the windmill at Frigges. Hemse Beds, top, and Eke Beds, base. RUDViER 1, 636260 167312 (CJ 6060 5960), Ardre parish, c. 2700 m south-south-west of Ardre church. Topographical map sheet 6 J Roma SV. Geological map sheet Aa 1 70 Katthammarsvik. Large quarry, marked on the topographical map, c. 200 m west of the main road where it passes Kaupungs. The boundary between the units c and d is exposed in the south wall of the quarry. Hemse Beds, units c and d. References: Mori 1970, loc. 116. SMISSARVE 1, 634517 164437 (CJ 3065 4445), Silte parish, c. 2800 m west-south-west of Silte church. Topographi- cal map sheet 5 I Hoburgen NO and 5J Hemse NV. Geological map sheet Aa 164 Hemse. The southeastern termination (at intersection between the wood and the field) of ditch running along, and south of, the road, 1600 m from the road junction at Siglajvs. Hemse Beds, Hemse Marl NW. URGUDE 3, 635003 164375 (CJ 3038 4931), Sproge parish, c. 1425 m west of Sproge church. Topographical map sheets 6 I Visby SO and 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Exposure in ditch perpendicular to road, from road to main ditch running east-north-east to west-south-west 250 m north of road. Urgude 3 includes exposure in ditch on east side of road to Snoder, 0-200 m from junction with Sproge-Lindarve road. Hemse Beds, Hemse Marl NW. 572 PALAEONTOLOGY, VOLUME 29 URGUDB 4, 634996 164358 (CJ 3018 4925), Sproge parish, c. 1600 m west of Sproge church. Topographical map sheet 5 I Hoburgen NO and 5 J Hemse NV. Geological map sheet Aa 164 Hemse. Exposure in ditch perpendicular to road, 0-50 m south of road, along west side of small wood, c. 600 m east of the windmill at Kruse. Hemse Beds, Hemse Marl NW. REFERENCES ALUERTi. H., ALBERTI, L., MANNIL, R. and TOMCZYKOWA, E. 1982. Correlation of the Proetus signatus level (upper Silurian) in Gotland, Poland and the East Baltic area. Eesti NSV Tend. Akad. Toim., Geol. 31, 29-32. ANGELIN, N. p. 1851. Palaeoiilologia Svecica. E Icoiiograpliia crustaceorum fonnationis trcmsitionis. Holmiae. Ease. 1, 1-24, pis. 1-24. 1854. Palcieontologia Scuudinavica. 1: Crustacea fonnationis transitionis. Lund. Ease. 2, 21-92, pis. 25- 41. BALASHOVA, E. A. 1954. In ALiKHOVA, T. N., BALASHOVA, E. A. and BALASHOV, z. G. Polevoy atlas kliaraktemykli konipleksov fanny otiozheniy ordovika i gotlandiya yuzhnoy cluisti Litovskoy SSR. Trudy VSEGEI, 43 pp. Moskva. 1968u. Trilobity skal’skogo i borshchovskogo gorizontov Podolii. In: Siliiriysko-devonskaya fauna Podo- lii, 95-122, 3 pis. Leningrad (Leningradskogo Universiteta). [In Russian.] 19686. Novye predstaviteli otryada Polymera nekotorykh rayonov SSSR. In markovskiy, b. p. (ed.). Novye vidy drevnikh rasteniy i hespozvokochnykh SSSR, vyp 2, cluisf vtoraya, 194-210. VSEGEI. Moskva. [In Russian.] 1975. Trilobity kitaygorodskogo gorizonta Podolii. Vopr. paleont. 1 (Fauna pogranitsnykh ordoviksko- siluriyskikh otiozheniy Podolii), 102-123, 2 pis. [In Russian.] BARRANDE, J. 1852. Systcnie Silurien du centre de la Bolienie. C’’’’ par tie. Recherclies paleontologic/ues, I. Crustaces, Trilohites, xxx + 935 pp, 51 pis. Prague and Paris. BEST, R. V. 1961. Intraspecific variation in Encrinurus ornatus. J. Paleont. 35, 1029-1040, pi. 124. BILLINGS, E. 1866. Catalogue of the Silurian fossils of the island of Anticosti, with descriptions of some new genera and species. Geol. surv. Can. 1-93. BOLTON, T. E. 1965. Trilobites from Upper Silurian rocks of the Canadian Arctic Archipelago: Encrinurus (Frannnia) and Henuarges. Bull. geol. Surv. Can. 134, 1-14. BRONGNIART, A. 1822. Histoire naturelle des Crustaces fo.ssiles, sous les rapports zoologicpies et geologicpies, savoir les Trilohites par Ale.xandre Brongniart . Les Crustaces proprement dits par A.-G. Desmarest, vii+ 154 pp., 1 1 pis. Paris. BUCKLAND, w. 1836. Geology and nnneralogy considered with reference to natural theology, I, xvi + 618 pp., II, vii + 129 pp., 69 pis. London. CAMPBELL, K. s. w. 1967. Trilobites of the Henryhouse Formation (Silurian) in Oklahoma. Bull. Okla. geol. Surv. 115, 1 68, pis. 1 19. CHANG, w. T. 1974. [Ordovician and Silurian trilobites in; A handhook of the Stratigraphy and Palaeontology in southwest China], 173-187. Ed. by Nanking Inst. Geol. Paleont., Academia Sinica. Academic and Science Press, Peking. CHERNS, L. 1983. The Hemse-Eke boundary. Facies relationships in the Ludlow Series of Gotland, Sweden. Sver. geol. Unders. C 800, 45 pp. CHLUPAC, I. 1977. The phacopid trilobites of the Silurian and Devonian of Czechoslovakia. Rozpr. Ustfed. list. Geol. 43, 172 pp., 32 pis. CLARKSON, E. N. K. and HENRY, J.-L. 1973. Structures coaptatives et enroulement chez quelques Trilobites ordoviciens et siluriens. Lethaia, 6, 105-132. DALMAN, J. w. 1827. Om Palaeaderna, eller de sa kallade Trilobiterna. K. svenska Veten.sk. -Akad. Handl. [for 1826], 113-152, 226-294, pis. 1-6. EMMRiCH, H. E. 1844. Zur Nciturgeschiclite der Trilohiten, 28 pp., 1 pi. Meiningen. EViTT, w. R. and tripp, r. p. 1977. Silicihed Middle Ordovician trilobites from the families Encrinuridae and Staurocephalidae. Palaeontographica (A), 157, 109-174, pis. 1-24. FOERSTE, A. E. 1888. Notes on Paleozoic fossils. Bull. Sci. Lahs Denison Univ. 2, 89-1 10 [tide Strusz 1980]. GAERTNER, H. R. VON. 1 930r/. Silurisclie und tiefunderdevonische Trilobiten und Brachiopoden aus den Zentral- karnischen Alpen. Jh. preuss. geol. Landesanst. 51, 188-252, pis. 24-26. RAMSKOLD; SILURIAN ENCRINURID TRILOBITES 573 1930/>. Obersilurische Faunen aus den spanischen Pyrenaen. Nadir. Ges. IVisx. Gottingen (Mat.-Phy.s., Geol. Min.), 179-188 [translated 1935; Fauna del Siliiric superior en els Pirineus. Bull. Inst. cat. Hist. nat. 35, 110 122], GASS, K. c. and mikulic, d. g. 1982. Observations on the Attawapiskat Formation (Silurian) trilobites of Ontario with description of a new encrinurine. Can. J. Earth Sd. 19, 589-596, pi. 1. HAAS, w. 1968. Trilobiten aus dem Silur und Devon von Bithynien (NW-Tiirkei). Palaeontographica (A), 130, 60-207, pis. 26-37. HALL, J. 1867. Account of some new or little known species of fossils from rocks of the age of the Niagara Group. Ann. Rep. N. Y. St. Cah. nat. Hist. 20, 305-401 [fide Strusz 1980]. — and WHITFIELD, R. p. 1 875. Descriptions of invertebrate fossils, mainly from the Silurian System, Paleonto- logy of Ohio, 2, 65-179. HAMMANN, w. 1974. Phacopina und Cheirurina (Trilobita) aus dem Ordovizium von Spanien. Senckenherg. Lc//l 55, 1 - 1 5 1 , pis. 1 12. HASWELL, G. c. 1865. On the Silurian f ormation in the Pentlaiul Hills, 48 pp., 4 pis. Edinburgh [fide Howells 1982]. HAWLE, I. and CORDA, A. J. c. 1847. Prodrotn einer Monographie cler hohtnisehen Trilobiten, 176 pp., 7 pis. Prague. [Also 1848, Abh. k. bdhni. Ges. Wiss. 5, 1 17-292, pis. 1-7.] HEDE, J. E. 1921. Gottlands silurstratigrali. Sver. geol. Uiulers. C 305, 100 pp. 1925. Beskrivning av Gotlands silurlager. In munthe, h., hede, j. e. and von post, l. Gotlands geologi. Ibid. C 331, 130 pp. 1927. Berggrunden (Silursystemet). In munthe, h., hede, j. e. and von post, l. Beskrivning till kartbladet Hemse. Ibid. Aa 164, 15-56. 1942. On the correlation of the Silurian of Gotland. MetUldanden Lunds geol. -miner. Instn, 101, 25 pp. HisiNGER, w. 1837. Lethaea Svecica sen petreficata Svedae, ieonibus et eharacteribus illustrata, 124 pp., 36 pis. Holmiae. HOLLOWAY, D. J. 1980. Middle Silurian trilobites from Arkansas and Oklahoma, U.S.A. Palaeontographica (A), 170, 1-85, pis. 1-20. HOWELLS, Y. 1982. Scottish Silurian trilobites. Palaeonlogr. Soc. [Monogr.] 561, 76 pp., 15 pis. HUCKE, K.. 1967. Einfiihrung in die Gesehiebeforschung. Verlag Nederlandse Geologische Vereiniging, Oldenzaal. JEPPSSON, J. 1982. Third European Conodont Symposium (EGOS 111). Guide to excursion. Publ. Inst. Min., Paleont. Qiiat. Geol. Univ. Lund, 239, 32 pp. 1983. Silurian conodont faunas from Gotland. Eossils and Strata. 15, 121-144. KINDLE, E. M. and BREGER, c. L. 1904. Part 11. Paleontology. In kindle, e. m. The stratigraphy and paleontology of the Niagara of Northern Indiana. Ann. Rep. Dep. Geol. nat. Res. Indiana 28, 428-486 [fide Strusz 1980]. KOBAYASHi, T. aiid HAMADA, T. 1974. Silurian trilobites of Japan in comparison with Asian, Pacific and other faunas. Spec. Pap. Palaeont. Soc. Japan, 18, 155 pp, 12 pis. KRUEGER, H.-H. 1971. Encriiiuriden aus ordovizischen Geschieben. Geologic, 20, 1 132-1 169, 8 pis. LAMONT, A. 1948. Scottish dragons. Quarry Mgrs J. 31, 531-535. 1965. Gala-Tarannon trilobites and an ostracod from the Hagshaw Hills, Lanarkshire. Scottish J. Sd. 1, 33-45, pi. 5. 1978. Pentlandian miscellany: Mollusca, Trilobita, etc. ibid. 1, 245-299, pis. 25-32. LANE, p. D. 1972. New trilobites from the Silurian of north-east Greenland, with a note on trilobite faunas in pure limestones. Palaeontology, 15, 336-364. 1979. Llandovery trilobites from Washington Land, North Greenland. Bull. Gronl. geol. Unders. 131, 37 pp., 6 pis. LARSSON, K. 1979. Silurian tentaculitids from Gotland and Scania. Eos.sds and Strata, II, 180 pp. laufeld, s. 1974u. Silurian Chitinozoa from Gotland. Ibid. 5, 130 pp. 19746. Reference localities for palaeontology and geology in the Silurian of Gotland. Sver. geol. Unders. C705, 172 pp. levi-setti, r. 1975. Trilobites: a photographic atlas, 213 pp. Univ. Chicago Press. LiNDSTROM, G. 1885. Lorteckning pa Gotlands siluriska crustaceer. Ofvers. K. Vetensk.-Akad. Ebrh. (6), 37- 100, pis. 12-16. 1901. Researches on the visual organs of the Trilobites. K. sveiiska Veten.sk. -,4kad. Handl. 34, 1-86, 6 pis. LINNAEUS, c. 1759. Petrilicatet Entomolithus paradoxus, sadant, som det finnes uti Hans Excellence, Riks- Radets Hogvalborne Herr Grefve c. G. tessins Samling. K. Vetensk.-Akad. Handl. 20, 19-24, pis. 1 and 2. 574 PALAEONTOLOGY, VOLUME 29 MAKSIMOVA, z. A. 1962. Trilobity ordovika i silura sibirskoy platformy. VSEGEl, 76, 1-214, pis. 1-18. 1975. Trilobity. In MiiNNt-R, v. v. (ed.). Kharakteristika fanny pognmichnykh shev Silura i Devona tsen- tral'nogo Kazakhstana [Materialy po geologii tsentral’nogo Kazakhstana 12]. 1 19-133. Moskva. MANNIL, RALF M. 1958. Trilobity semeystv Cheiruridae i Encrinuridae iz Estonii. Eesti NSV Tead. Akad. Geol. Inst. Uurim. 3, 165-212, pis. 1-6 [with English summary]. MANNIL, REET. 1968. Encfinwiis schmidti sp. n. (Trilobita) iz Llandoveri Estonii. Eesti NSV Tead. Akad. Toiin., Keem. Geol. 17, 273-278, pis. 1 and 2 [with English summary]. 1977. Novye enkrinuridy (Trilobita) Llandoveri Pribaltiki. Ibid. 26, 46-56, pis. I and 2 [with English summary]. - 1978. Trilobity vidovoy gruppy Encrinurus pimctatiis v Wenloke Pribaltiki. Ibid. 27, 108-115, pis. 1-4 [with English summary]. 1982u. Wenlock and Late Silurian trilobite associations of the East Baltic area and their stratigraphical value. In kaljo, d. and klaamann, e. (eds.). Ecostratigraphy of the East Baltic Silurian. [Akad. Nauk Est. SSR, Inst. Geol.] 63-70. Tallinn. 19826. Trilobite communities (Wenlock, East Baltic). In kaljo, d. and klaamann, e. (eds.). Communities and hiozones in the Baltic Silurian. [Akad. Nauk Est. SSR. Inst. Geol.] 51-62, pis. 3-6. Tallinn. MILLER, J. 1976. The sensory fields and life mode of Phacops rana (Green, 1832) (Trilobita). Trans. R. Soc. Edinh. 69, 337-367, pis. 1-4. MILLER, s. A. 1880. Description of two new species from the Niagara Group and five from the Keokuk Group. J. Cincinn. Soc. nat. Hist. 2, 254-259 [fide Strusz 1980]. MORI, K. 1970. Stromatoporoids from the Silurian of Gotland. Part 2. Stockholm Contrih. Geol. 22, 1-152, 30 pis. MUNTHE, H., HEDE, J. E. and VON POST, L. 1925. Beskriviiing till kartbladet Ronehamn. Sver. geol. Unders. Aa 156, 96 pp. MUNSTER, G. GRAF zu. 1840. Beitidge zur Petrefakten-kunde 3, 132 pp, 20 pis. Bayreuth [fide Temple and Tripp 1979]. NiESZKOWSKi. J. 1859. Zusatze zur Monographie der Trilobiten der Ostseeprovinzen, nebst der Beschreibung einiger neuen obersilurischen Crustaceen. Tr(7;/v77«/w'6. L/V-, Efct-A'wr/unr/.s'. Ser. 1, 2, 345-384, pis. 1 and 2. NORFORD, B. s. 1981. The trilobite fauna of the Silurian Attawapiskat Formation, northern Ontario and northern Manitoba. Bull. geol. Surv. Can. 327, 15 pp., 1 1 pis. NORTHROP, s. A. 1939. Paleoiitology and stratigraphy of the Silurian rocks of the Port Daniel-Black Cape region, Gaspe. Spec. Pap. geol. Soc. Am. 21, 302 pp. NORTON, w. H. 1895. Variation in the position of the nodes on the axial segments of pygidium of a species of Encrinurus. Iowa Acad. Sci. 5, 3, 79-81 [fide Best 1961]. PATTE, E. 1929. Description de fossiles paleozoi'ques et mesozoiques recueillis par Mm. Dussault er Fromaget en extreme-orient. Bull. Serv. geol. Indochine, 18 [fide Strusz 1980]. PERRY, D. G. and CHATTERTON, B. D. E. 1977. Silurian (Wenlockian) trilobites from Baille-Hamilton Island, Canadian Arctic Archipelago. Can. J. Earth Sci. 14, 285-317, 7 pis. and 1979. Wenlock trilobites and brachiopods from the Mackenzie Mountains, north-western Canada. Palaeontology, 22, 569-607, pis. 68-76. PRiBYL, A. and VANEK, J. 1962. Die Trilobiten-Fauna aus dem bohmischen Obersilur (Budnanium und Lochko- vium) und ihre biostratigraphische Bedeutung. Sh. Ndr. Muz. Praze, (B) 18, 25-46, 4 pis. RAMSKOLD, L. 1983. Silurian cheirurid trilobites from Gotland. Palaeontology, 26, 175-210, pis. 19-28. 1984. Silurian odontopleurid trilobites from Gotland. Ibid. 27, 239-264, pis. 26-31. and BASSETT, M. G. (in prep.). A middle Llandovery shelly fauna from Ostergotland, Sweden. RAYMOND, p. E. 1916. New and old Silurian trilobites from southeastern Wisconsin, with notes on the genera of the Illaenidae. Bull. Mus. comp. Zool. Harvard Univ. 60 ( 1), 1-41, 4 pis. REED, F. R. c. 1931. The Lower Palaeozoic trilobites of the Girvan District, Ayrshire. Supplement no. 2. Palaeontogr. Soc. [Monogr.], 30 pp. ROSENSTEIN, E. 1941. Die Encrinurus-Arten des Estlandischen Silurs. Tartu Ulik. Geol.-Inst. Toim. 62, 49-77, pis. 1-4. SCHMIDT, F. 1859. Beitrag zur Geologic der Insel Gotland, nebst einigen Bemerkungen fiber die untersilurische Formation des Festlandes von Schweden und die Heimath der norddeutschen silurischen Geschiebe. Archiv Naturk. Liv-, Ehst-Kurlands, Ser. I, 2, 403-464. 1881. Revision der ostbaltischen silurischen Trilobiten, nebst geognostischer Ubersicht des ostbaltischen Silurgebiets. Abt. 1. Phacopiden, Cheiruriden und Encrinuriden. Zap. Imp. Akad. Nauk [MNn. Acad. imp. Sci. St.-Petersb.] Ser. 7, 30, 1-238, pis. 1-16. RAMSKOLD: SILURIAN ENCRINURID TRILOBITES 575 SCHRANK, E. 1972. Proctacca, Encrinuridae und Phacopina (Trilobita) aiis silurischen Geschieben. Geologic, 21 (Beih. 76), 1-117,21 pis. 1977. Zur Paliiobiogeographie silurischer Trilobiten. N. Jh. Geol. Pcilaoiit. Abh. 155, 108-136. SHAW, F. c. and ormiston, a. r. 1964. The eye socle of trilobites. J. Paleom. 38, 1001 -1002. SNAJDR, M. 1975. New Trilobita from the Llandovery at Hyskov in the Beroiin area, central Bohemia. Vest. Ustr. list. geol. 50, 311-316, 2 pis. 1978. The Llandoverian trilobites from Hyskov (Barrandian area). Shor. geol. Veil., R. P. 21, 7-47, 12 pis. 1981. New Silurian and Devonian trilobites (Barrandian, Czechoslovakia). Vest. Ustr. list. geol. 56, 301- 303, 2 pis. 1983. New Silurian trilobites from Bohemia. Ibid. 58, 175 178, 2 pis. STEARN, c. w. 1956. Stratigraphy and palaeontology of the Interlake group and Stonewall formation of southern Manitoba. Mem. geol. Siirv. Can. 281, 162 pp., 16 pis. STEMVERS-, VAN BEMMEL, J. 1974. Geologie van Gotland. Geti, 7, 1, 3-7, 10-14. STRUSZ, D. L. 1980. The Encrinuridae and related trilobite families, with a description of Silurian species from southeastern Australia. Palaeoiitograpliica (A), 168, 1-68, pis. 1-6. TALENT, J. A. 1965. The Silurian and early Devonian faunas of the Heathcote district, Victoria. Mem. geol. Stirv. Viet. 26 (for 1964), 1-55. TEMPLE, J. T. and TRIPP, R. p. 1979. An investigation of the Encrinuridae (Trilobita) by numerical taxonomic methods. Trans. Roy. Soc. Eclin. 70, 223-250. THOMAS, A. T. 1981. British Wenlock trilobites. Part 2. Palaeontogr. Soc. [Monogr.], 559, 57-99, pis. 15-25. and NARBONNE, G. M. 1979. Silurian trilobites from arctic Canada. Geol. Mag. 116, 1 19, 5 pis. TORNQUIST, s. L. 1884. Uiidersokningar 5fver Siljansomradets trilobitfauna. Sver. geol. Unders. C 66, 101 pp., 3 pis. TRIPP, R. p. 1957. The trilobite Encrinurus nndtisegmentatns (Por{\oc\s) and allied Middle and Upper Ordovician species. Palaeontology, 1, 60-72, pis. 1 1 and 12. 1962. The Silurian trilobite Encrinurus punctatus (Wahlenberg) and allied species. Ibid. 5, 460-477, pis. 65-68. — TEMPLE, J. T. and GASS, K. c. 1977. The Silurian trilobite Encrinurus variolaris and allied species, with notes on Erammia. Ibid. 20, 847-867, pis. 113-115. and WHiTTARD, w. F. 1956. Proposed use of the plenary powers (a) to designate type species in harmony with accustomed usage for the genera 'Encrinurus'' Emmrich, 1844, and 'Odontochile' Hawle & Corda, 1847, and (b) to validate the specific name 'punctatus' Wahlenberg, 1821, as published in the combination ' Entomostracites punctatus' (class Trilobita). Bull. Zool. NomencL 12, 259-263, pi. 3. TSEGELNYUK, P. D. et cd. 1983. The Silurian of Podolia. The guide to excursion, 224 pp. Naukova Dumka, Kiev. TWENHOFEL, w. H. 1916. Expedition to the Baltic provinces of Russia and Scandinavia, 1914. Part 2. The Silurian and high Ordovician strata of Esthonia, Russia and their facies. Bull. Mus. comp. Zool. Harvard Univ. 56,4, 290-340, pis. 1-3. — - 1928. Geology of Anticosti Island. Mem. geol. Surv. Can. 154, 481 pp. VOGDES, A. w. 1886. Description of a new crustacean from the Clinton group of Georgia, with remarks upon others, 5 pp. New York. WAERN, B. 1960. On the Middle Llandovery of Dalarna. Intermit. Geol. Congress, Rep. 21st Session, Norden, 7, 126-133. WAHLENBERG, G. 1 8 1 8. Petrificata Telluiis Svecanae. Tctu R. Suc. SC. C/wu/. 8 [for 1821], 1-116, pis. 1-4. WELLER, s. 1907. The paleontology of the Niagaran Limestone in the Chicago Area. The Trilobita. Bull. Chicago Acad. Sci. 4 (2), 163-281, pis. 16-25. WHITTARD, w. F. 1938. The Upper Valentian trilobite fauna of Shropshire. Ann. Mag. nat. Hist. (11), 1, 85- 140, pis. 2-5. WHITTINGTON, H. B. and CAMPBELL, K. s. w. 1967. Silicified Silurian trilobites from Maine. Bull. Mus. comp. Zool. Harvard Univ. 135, 9, 447-483, 19 pis. and EVITT, w. r. 1954. Silicified Middle Ordovician trilobites. Mem. geol. Soc. Amer. 59, 1-137, pis. 1-33. LARS RAMSKOLD Avd. for Paleozoologi Naturhistoriska Riksmuseet Box 50007 S-104 05 Stockholm, Sweden Typescript received 28 January 1985 Revised typescript 20 November 1985 I '^1 6 THE FIRST TERTIARY SCLEROSPONGE FROM THE AMERICAS by EDWARD C. WILSON Abstract. Diplochaetetes mexicumts sp. nov. (Porifera: Sclerospongiae) is erected for specimens of fossil coralline sponges collected from the El Cien Formation of Late Oligocene and Early Miocene age in eastern Baja California Sur, Mexico. The genus was previously known only by its type species, D. Weissermel, 1913 from the Eocene of Namibia (South West Africa). This is the first record of sclerosponges in the Tertiary of the Americas. In 1940, J. W. Durham discovered a Tertiary marine section near Punta San Telmo, Baja California Sur, Mexico, from which he subsequently (Durham 1950) erected a new species of the bivalve Auadara and reported a few other marine mollusks. The same locality yielded the marine mammal Cornwallius (Vanderhoof 1942). In a correlative section 20 km south of this locality. I collected the fossils which form the basis of this report. REGIONAL GEOLOGY AND STRATIGRAPHY The geology of the isolated area east of the Sierra de la Giganta of Baja California Sur has been little studied. The sea cliflfs that extend for 90 km north of Punta La Laguna (text-fig. 1 ) show beds that are arched in a broad, gentle anticline trending roughly east-west with northern and southern limbs that dip at about 5 degrees and are broken in places by normal faulting of minor displacement. Moderately metamorphosed sediments at the core of the anticline are exposed in places for 12 km between Punta San Telmo southward to Roca Moreno just oft' Punta del Cobre. They are known as the San Telmo Formation and considered to be of Mesozoic age. The strata overlying the San Telmo Formation are unmetamorphosed. The lowest unit is about 100 m thick and composed of cross-bedded, red sandstones of uncertain age that may correlate with the Upper Oligocene Salto Formation which crops out further north (McFall 1968; Gastil et al. 1979; Hausback 1 984). The marine Tepetate Formation of Cretaceous to Eocene age occupies this position in sections on both sides of the peninsula further south so there is also a possibility that the redbeds may, at least in part, be a terrestrial equivalent of this unit. The superjacent unit is composed of a basal conglomerate and about 100 m of marine sandstones, shales, phosphorites, porcellanites, limestones, conglomerates, and tuft's of Late Oligocene and Early Miocene age. It has been named the San Gregorio Eormation by various workers, including Durham (1950) and Hausback (1984). Others have referred to it as the Monterey, Monterrey, or Monte-Rey Eormation because of a reputed similarity to the younger Monterey Formation of central and southern California. Its unique lithology, however, led Applegate (in press) to name it the El Cien Formation. It is the source of phosphates mined in large quantities at San Juan de la Costa on the coast 45 km north-west of La Paz and formerly mined near El Cien on the Transpeninsular Highway (Mexico I) 75 km north-west of La Paz. This formation yielded the material described here. Above this unit lie more than 1 000 m of Miocene marine, terrestrial sedimentary, and volcanoclastic rocks of the San Isidro and Comondu Formations that form the bulk of the Sierra de la Giganta. Pliocene marine sedimentary rocks have not been recognized in the area, but there are marine terrace deposits of Pleistocene age. (Palaeontology, Vol. 29, Part 3, 1986, pp. 577-583, pi. 50.| 578 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1 . Index map (left) showing general location of study area (rectangle) in Baja California Sur, and more detailed map (right) indicating type localities of Diplochaetetes mexicana n. sp. (LACMIP loc. 6349) and of Anadara vanderhoofi Durham, 1950 (UCMP loc. A3595). Tembabiche (map, right) is the only settlement in the study area other than scattered ranches. Its small airfield is indicated by the dashed-line rectangle. WILSON: TERTIARY SCLEROSPONGE FROM MEXICO 579 LOCALITIES The locality (LACMIP loc. 6349) which yielded the paratypes is in the south-west quarter of section 09-82 as shown on the Estados Unidos Mexicanos, Direccion General de Geografia, Carta Topografica Los Burros G12D41, Baja California Sur (1978, 1:50000). It is 11-5 km south of Tembabiche and about 400 m inland from the sandy beach at the mouth of the first wash north of Arroyo San Jose on a low ridge forming its north boundary (text-fig. 1 ). Cursory examination of this locality served to place the sponge bed in a regional stratigraphic framework. Two metres stratigraphically below it is a well-indurated limestone coquina of Anadara vanderhoofiDarh'dm, 1950 (LACMIP loc. 6348). The type locality of this bivalve (UCMP loc. A3595) is in Durham’s section near Punta San Telmo (text-fig. 1 ). Applegate and Wilson (1976) proposed a biostratigraphic correlation using occurrences of this species at El Cien and the type locality, which seems corroborated by subsequent field observations. It seems likely, therefore, that the occurrence of the species below the sponge bed represents the same stratigraphic position in the El Cien Formation. If so, the sponge bed lies between beds of known Late Oligocene and Early Miocene ages. An age refinement awaits more precise placement of the Oligocene-Miocene boundary within the formation. A single specimen of the sponge was collected by Applegate in 1978 from the main quarry of Roca Phosphorica Mexicana at San Juan de la Costa, on the coast 45 km north-west of La Paz (text-fig. 1 ). I have designated this specimen the holotype because of its superior preservation. Applegate reports that it came from the same stratigraphic position in the El Cien Formation as the specimens found by me further north. The locality has since been destroyed by the strip mining operation. SYSTEMATIC PALAEONTOLOGY It has been nearly 20 years since Hartman and Goreau ( 1 966) reported that some living sponges secrete skeletons so similar to those of fossils then classified in the phylum Coelenterata that they should be reassigned to the Porifera. The history and significance of this and related reassignments recently has been summarized succinctly by Basile et at. (1984), Fagerstrom (1984), and others and need not be repeated here. Even recently, however (Hill 1981, p. 520), Diplochaetetes Weissermel, 1913 has been placed among the tabulate corals although Hartman and Goreau (1972, p. 138) have earlier recognized the relationship of the genus to the sclerosponges. Morphological terminology follows de Laubenfels ( 1 955), with modifications for coralline sponges by Steam (1984). Repositories of type specimens and locality numbers are denoted by the following institutional acronyms: IGM (Instituto de Geologia, Universidad Nacional Autonoma de Mexico), LACMIP (Invertebrate Paleonto- logy Section, Natural History Museum of Los Angeles County), and UCMP (University of California Museum of Paleontology, Berkeley). Phylum PORIFERA Class SCLEROSPONGIAE Hartman and Goreau, 1970 Order tabulospongida Hartman and Goreau, 1975 Family acanthochaetetidae Fisher, 1970 ( = tabulospongidae Mori, 1976) Genus diplochaetetes Weissermel, 1913 Type species. Diplochaetetes longituhiis Weissermel, 1913, from the Eocene of Namibia. Diagnosis. Diplochaetetes is a coralline sponge of large domal growth form with polygonal to rounded tabulated calicles that lack septa and spinules, a completely aspicular skeleton, lamellar ealicle walls that are fused where calieles touch and unfused where they do not, tabulae distally coneave and complete and grouped and ungrouped but not zoned, an axial and bipartite increase, and a poorly known fine-wall microstructure. 580 PALAEONTOLOGY. VOLUME 29 Diplochaetetes mexicanus sp. nov. Plate 50, figs. 1-4 Diagnosis. A species of Diplochaetetes characterized by great numbers of tabulae (as many as 1 5 per mm) clumped in the calicles of some specimens. Etymology. The species is named after Mexico, country of origin. Type material and Locality. IGM holotype 3948, IGM paratypes 3949-3955, LACMIP paratype 7195. Two thin sections and eighteen polished sections from the holotype and seven thin sections and sixty-six polished sections from the eight paratypes were studied. Upper Oligocene or Lower Miocene, El Cien Formation; San Juan de la Costa (holotype) and 1 1-5 km south ofTembabiche (paratypes), both Baja California Sur, Mexico. External description. Growth form domal, large, diameter up to 60 cm; calicles polygonal giving cerioid appearance; external surfaces not well preserved. Transverse section description. Calicles 1-5 to 2-0 mm in diameter, generally polygonal externally, circular internally, not everywhere in contact in all neighbouring calicles (sediment in ‘corners’); septa and spinules not present; tabulae absent or represented by one or more circular shapes in some calicles; walls lamellar, fused where neighbouring calicles touch, separate elsewhere. Longitudinal section description. Calicles long, straight to gently curved (rarely broadly vermiform), 1-5 to 2 0 mm in diameter; septa and spinules not present; tabulae generally complete, concave upward, not zoned in neighbouring calicles, generally grouped 1 to 3 per mm, with clumps 5 to 20 mm apart, densely grouped in some calicles, 4 to 1 5 per mm for vertical distances as much as 10 mm, with intervening distances between dense groups without tabulae or less densely grouped as in other calicles; wall lamellar, preservation poor. Discussion. The densely clumped tabulae of the new species distinguish it from D. longituhus Weissermel, 1913, the type and only other known species, from the Eocene (Siesser 1977) of Namibia. Weissermel (1913) had eleven specimens in the type lot and later received another thirteen which he described (Weissermel 1926), so it seems unlikely that his sample was too small to detect densely grouped tabulae if they had been present in his species. His type specimens are apparently lost and 1 have been unable to obtain topotypes. It is unfortunate that the wall microstructure of the type species was not described minutely but only mentioned as being lamellar. The Mexican specimens have Si02 intergrown between the wall lamellae as well as filling the calicles and the original microstructure cannot be observed with assurance. Abundant microfractures caused by the Si02 emplacement are widespread in most calicles. Although external surfaces of none of the specimens of D. mexicanus were preserved adequately to show the presence or absence of such features as astrorhizae, Weissermel (1926, pi. 35, figs. 1 and 2) illustrated specimens of D. longitubus with well-preserved external surfaces that show that these features are not present in the type species. In the LACMIP paratype of D. mexicanus there is an opening suggestive of an osculum, but it was the only one encountered and may represent growth around another organism. The specimens seen by me at LACMIP locality 6349 appeared to have been in place and in growth position. They resembled coralla of Chaetetes that I have seen in Pennsylvanian formations of eastern Nevada. Large domal heads of D. mexicanus are abundant and upright, closely spaced, but not touching. EXPLANATION OF PLATE 50 Figs. I -4. Diplochaetetes mexicanus sp. nov., El Cien Formation, Oligocene-Miocene, Mexico, x 10. 1 -2, IGM holotype 3948. 1, transverse section. 2, longitudinal section showing densely grouped tabulae. 3-4, FACMIP paratype 7195. 3, transverse section. 4, longitudinal section. PLATE 50 WILSON, Dip/ochaetetes mexicanus sp. nov. 582 PALAEONTOLOGY, VOLUME 29 Some specimens of D. mexicanus exhibit sinuous calicles in places either as the result of upward growth around the edges of the domes or as the result of reoriented upward growth after the specimen had been tilted. In this respect, they somewhat resembled D. longituhus var. vennicularis Weissermel, 1926, also from the Eocene of Namibia. I have no record of the field appearance of the holotype from the locality at San Juan de La Costa. D. mexicanus occurs in a formation that has bone beds, including some articulated specimens, probable turbidites, and other indicators of at least moderately deep water. Its presence does not necessitate a shallow-water palaeoenvironmental interpretation, although most recent work on sclerosponges has emphasized their occurrences as part of the shallow cryptofauna of tropical reefs. Hartman and Goreau (1970) reported living sclerosponges to depths of 92 m on outer reef slopes off Jamaica. The holotype of the sclerosponge Ceratoporella nicholsonii (Hickson 1911) was dredged off Cuba at a depth of 100 fathoms (Hickson 1911). Sheldon (1982) postulated that the phosphate deposits of Baja California Sur were deposited by extensive upwelling of phosphate-rich waters brought up to replace surface waters moved away by off-shore trade winds. Since D. mexicanus is associated intimately with these deposits, it follows that it inhabited cool water. The rarity of tropical corals and large tropical mollusks in the fauna further suggests cool temperatures. Acknowiedgements. I wish to thank Shelton P. Applegate of the Instituto de Geologia, Mexico City, for arranging for me to accompany him in October 1983 on the field trip to the Punta San Telmo area where I found the specimens described here. The Los Angeles County Museum of Natural History Foundation provided funds. REFERENCES APPLEGATE, s. p. In press. The El Cien Formation, strata of Late Oligocene and Early Miocene age in Baja California Sur. Univ. Nal. Anton. Mexico, Inst. GeoL, Revista, 6. and WILSON, e. c. 1976. Correlation of Upper Oligocene or Lower Miocene sections at San Telmo Point and Arroyo Guadalupe, Baja California Sur, Mexico, and a possible new phosphate source. Latinoamericano Congreso de Geologica III, Reswnenes, Mexico, p. 6. BASiLE, L. L., CUFFEY, R. J. and KOSiCH, D. F. 1984. Sclerosponges, pharetronids, and sphinctozoans (relict cryptic hard-bodied Porifera) in the modern reefs of Enewetak Atoll. J. Paleont. 58, 636-650. DURHAM, J. w. 1950. Megascopic paleontology and marine stratigraphy. Mem. Geol. Soc. Am. 43, 1-216. FAGERSTROM, J. A. 1984. The paleobiology of sclerosponges, stromatoporoids, chaetetids, archaeocyathids and non-spicular calcareous sponges. Introduction. Palaeontogr. Am. 54, 303-304. FISCHER, J. c. 1 970. Revision et essai de classification des Chaetetida (Cnidaria) post-Paleozoiques. Ann. Paleont. 56,151-220. GASTiL, G., KRUMMENACHER, D. and MINCH, J. 1979. The record of Cenozoic volcanism around the Gulf of California. Bull. Geol. Soc. Am. 90, 839-857. HARTMAN, w. D. and GOREAU, T. F. 1966. Ceratoporella, a living sponge with stromatoporoid affinities. Am. Zool. 6, 563-564. 1970. Jamaican coralline sponges: their morphology, ecology and fossil relatives. Symp. zool. Soc. bond. 25, 205-243. 1972. Ceratoporella (Porifera: Sclerospongiae) and the chaetetid ‘corals’. Trans. Conn. Acad. Arts Sci. 44, 133-148. 1975. A Pacific tabulate sponge, living representative of a new order of sclerosponges. Postilla, 167, 1-14. HAUSBACK, B. p. 1984. Cenozoic volcanic and tectonic evolution of Baja California Sur, Mexico. In frizzell, V. A. (ed.). Geology of the Baja California peninsula. Pac. Sec., Soc. Econ. Paleont. Mineralog. 39, 219-236. HICKSON, s. J. 1911. On Ceratopora, the type of a new family of Alcyonaria. Proc. R. Soc. B84, 195-200. HILL, D. 1981. Rugosa and Tabulata. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology. Part F. Coelenterata, FI -762. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. LAUBENFELS, M. w. DE. 1955. Poiifera. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology. Part E. Archaeocyatha and Porifera, E21-122. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. WILSON: TERTIARY SCLEROSPONGE FROM MEXICO 583 MCFALL, c. c. 1968. Reconnaissance geology of the Concepcion Bay area, Baja California, Mexico. Stanford Univ. Pub. Geol. Sci. 10, 1-25. MORI, K. 1976. A new recent sclerosponge from Ngargol, Palau Islands and its fossil relatives. Sci. Rep. Tohoku Univ. 2 (Geol.), 46, 1-9. SHELDON, R. p. 1982. Phosphate rock. Sci. Am. 246, 45-51. SIESSER, w. G. 1977. Upper Eocene age of marine sediments at Bogenfels, South West Africa, based on calcareous nannofossils. X. W. Afr. Geol. Surv. Bull. 60, 1'i-IA. STEARN, c. w. 1984. Growth forms and macrostructural elements of the coralline sponges. Palaeontogr. Am. 54, 315-325. VANDERHOOF, V. L. 1942. Occurrence of the Tertiary marine mammal Cornwallius in Lower California. Am. J. Sci. 240, 298-301. WEISSERMEL, w. 1913. Tabulaten und Hydrozoen. Beit. geol. Erforsch. dt. Schutzgeb. 5, 84- 111. 1926. Neues uber Tabulaten, Hydrozoen und eine Hexakoralle aus dem Tertiar der Bogenfelser Diamantenfelder. In kaiser, h. Die Diamantemvuste Sudwest-Afrika.s, 2, 88-106. Berlin. Typescript received 15 April 1985 Revised typescript received 12 July 1985 EDWARD C. WILSON Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 . , n ’ li, > :'i " ; ■ - f. ■ i 1 PALAEOECOLOGY OF SILURIAN CYCLOCRINITID ALGAE by STEVEN C. BEADLE and MARKES E. JOHNSON Abstract. Cyclocrinitids are intriguing macrofossils, commonly considered calcareous green algae, which are locally abundant in certain Ordovician and Silurian deposits. Silurian cyclocrinitids were ecologically and morphologically similar to dasycladacean algae. Cyclocrinitids and dasyclads commonly coexisted, and showed similar patterns of radiation and decline in response to early Palaeozoic environmental changes. Locally, Silurian cyclocrinitids showed considerable variation in adult thallus size; populations in certain environments had significantly smaller thalli than neighbouring populations. This variation was probably due to differences in light intensity, which is an important control on the growth of Recent algae. Cyclocrinitids should, there- fore, be useful locally as relative depth indicators: populations with small thalli should indicate deeper, darker waters. In all cases the relative depths indicated by cyclocrinitid size variation are the same as those indicated by the invertebrate communities. This supports the hypothesis that Lower Silurian invertebrates were largely zoned by factors related to water depth. Communities containing cyclocrinitids probably existed at depths of less than 100 m. The cyclocrinitids are a small group of problematical organisms which lived in shallow-marine environments from mid-Ordovician to early Silurian times. Although generally rare as fossils, they may be locally common in association with more familiar Palaeozoic forms such as brachiopods and corals. In a few places they reach such high densities that they outnumber all other fossils combined. Cyclocrinitids are useful guide fossils for a number of European and North American rock units, and several marker beds have been named for Cyclocrinites or a related genus. The systematic placement of cyclocrinitids is still debated today, nearly 150 years after they were first described. This study focuses instead on cyclocrinitid ecology, particularly on their responses to environmental change. Their distribution was affected by global environmental events, and they were also sensitive to local variations. This makes them useful as palaeoenvironmental indicators; specifically, they can be used to test the proposed palaeoenvironmental significance of the associated Silurian invertebrate communities. MORPHOLOGY Cyclocrinitids were essentially hollow organisms (text-fig. 1 ). They were attached to the substrate by a central axis, which supported lateral branches that radiated outwards in all directions. The branches expanded at their distal tips to form thick lateral heads, which were hexagonal in outline and packed tightly together in a honeycomb-like fashion. Thus an ovoid or spherical thallus was formed which completely enclosed the main axis and the lateral branches. Calcium carbonate encrusted various parts of the thallus. All Silurian cyclocrinitids had generally spherical thalli which were calcified about the interior surface of the lateral heads. Consequently, only the lateral heads are usually preserved, but a few speeimens with internal structures are known (text-fig. 2a-e; see also Elliott 1972; Nitecki and Johnson 1978). Silurian cyclocrinitids may be preserved in their original spherical shape, usually as internal moulds (text-fig. 2f-l), or they may be flattened to varying degrees after burial. The size of intact spherical thalli can be measured in three different ways. The most direct method is to measure the thallus diameter (text-fig. la). Eurthermore, individual lateral heads increase in size along with the total diameter, so that lateral head width (text-fig. 17>) or thickness (text-fig. Ic) can I Palaeontology, Vol. 29, Part 3, 1986, pp. 585-601.) 586 PALAEONTOLOGY, VOLUME 29 lateral branch lateral head mam axis TEXT-FIG. 1 . Reconstructed morphology of Cyclocrinites dactioloides. Dimensions: a = thalliis diameter; b = lateral head width; c = lateral head thickness. Modified after Nitecki (1972, fig. 12). also be used as size indicators. It is usually impossible to reconstruct the original diameters of flattened thalli, but lateral head width can be used as a relative size indicator instead. AFFINITIES Cyclocrinitids were first described by nineteenth-century invertebrate palaeontologists, who variously considered them foraminiferans, sponges, corals, bryozoans, gastropod eggs, cystoids, or tunicates. Eventually it became clear that they actually resembled calcareous green algae, specifically dasyclads such as Bornetella and Neomeris (Nitecki 1970; Nitecki and Johnson 1978). The Family Dasycladaceae is a distinctive clade of the Division Chlorophyta (green algae). Dasyclads are characterized by a central stem with primary branches radiating outwards in regular whorls (J. H. Johnson 1961a). In some forms, the branches dilate at the tips to form hexagonal heads, which are closely packed together to form a faceted outer cortex. Many Recent forms are encrusted by calcium carbonate (as aragonite). Cyclocrinitids shared this general morphology, and Pia (1927) placed them within the Dasycladaceae as the Tribe Cyclocriniteae (= Cyclocrineae). This view has since gained wide acceptance (Osgood and Fischer 1960; Riding 1977). More recently, Nitecki (1970, 1972, 1976) has allied cyclocrinitids with the even more problematical receptaculitids. It is now thought that receptaculitids are calcareous algae of uncertain position (Rietschel 1969; Rietschel and Nitecki 1984). If cyclocrinitids are truly related to this group, then they too should be considered problematic calcareous algae (Fischer and Nitecki 1982). However, other receptaculitid workers believe that the two groups are unrelated (Campbell et al. 1974; Rietschel 1969, 1977). Recent studies of dasyclad evolution continue to place cyclocrinitids within the Dasycladaceae (Elliott 1972; Herak et al. 1977). STRATIGRAPHIC RANGE The earliest known cyclocrinitids are of middle Ordovician age. They apparently expanded rapidly, since middle and late Ordovician cyclocrinitids are widely distributed in North America, BEADLE AND JOHNSON; SILURIAN CYCLOCRINITID ALGAE 587 TEXT-FIG. 2. Representative cyclocrinitid algae from the Lower Silurian of Iowa and Norway, a, b, Cyclocrinites dactioloides Owen. Portion of a single specimen from near the top of the Farmers Creek Member of the Hopkinton Dolomite (loc. 79), FMNH UC 59064. See also Nitecki and Johnson 1978. a, apical view showing radial arrangement of lateral branches, x 3-25. b, lateral view showing lateral branches, x 3-25. C-E, C. dactioloides Owen. Hollow cast of a single specimen from near the top of the Farmers Creek Member of the Hopkinton Dolomite (loc. 79), FMNH UC 59430. c, interior view of thallus, x 1. d, enlarged view of main axis remnant (box in c), x 2. e, enlarged view of lateral heads, showing the distal tips of the lateral branches (the rest of the branches are not preserved), x 2. f, C.faviis Salter. Internal mould of immature thallus ( = Mastopora sp. of Kiaer 1908) from the upper part of the Leangen Member of the Solvik Formation in Asker, Norway, FMNH PP 34246, x 1. g-i, C. dactioloides Owen. Three thalli from the upper part of the Farmers Creek Member of the Hopkinton Dolomite (loc. 79), FMNH UC 59442, 59443, and 59458 respectively, x 1 . All belong to a pentamerid community dominated by Harpidium maquoketa. j-l, C. dactioloides Owen. Three thalli from the lower part of the Farmers Creek Member of the Hopkinton Dolomite, FMNH UC 59237, 59293 (loc. 28), and 62733 (loc. 49) respectively, x 1 . All belong to a stricklandiid community dominated by Stricklaitdia laevis ( = S. lens ultima). northern Europe, and even the central Himalayas (text-fig. 3r/). Three genera and at least eight species are currently recognized (J. H. Johnson 1961/?; Nitecki 1970 and references therein). Present evidence suggests that the global range of cyclocrinitids began to shrink during the late Ordovician. By early Llandovery time, only two species, Cyclocrinites favus and C. gregarius, remained, and they had a much more restricted geographic distribution (text-fig. 3b). They were 588 PALAEONTOLOGY. VOLUME 29 a Middle/Late Ordovician TEXT-FIG. 3. Palaeogeographic distribution of cyclocrinitids. Solid circles = fossil localities, dotted line = inferred range. A = Avalonia, B = Baltica, C = China, G = Gondwana, K = Kazakhstania, L = Laurentia, S = Siberia, a. Middle and Late Ordovician. Genera present in- clude Cyclocriniles ( = Mastopora, Nichdites, Pasceolus), Coelosphaeridiiim, and Apidium. North American localities from Nitecki (1970); European localities from Johnson (1961b) and Hoeg (1961); Asian localities from Reed (1912). Ashgill base map from Scotese (1984, fig. 3). b. Early Llan- dovery (Rhuddanian-lower Aeronian). Cyclo- crinifes faviis and C. gregarius are the only species present; they may be synonymous. See text-hg. 5 for locality references. Ludlow base map from Scotese (1984, hg. 4). c. Late Llandovery (upper Aeronian Telychian). C. dactiohides is the only species present. No later cyclocrinitids are known. See text-fig. 5 for locality references. Ludlow base map from Scotese (1984, hg. 4). succeeded in the late Llandovery by C. dactiohides, the last known cyclocrinitid, which is found only in the American Midwest (text-fig. 3c). The group was apparently extinct by the start of Wenlock time. Benthic calcareous algae generally reached peak diversities in the early to mid-Ordovician and subsequently declined through the late Ordovician and early Silurian (Riding 1984). This may have been due to glacio-eustatic sea-level changes or climatic deterioration. At any rate, both cyclocrinitids and dasyclads conform closely to this general pattern (text-fig. 4). The appearance of cyclocrinitids in the mid-Ordovician coincided with a major dasyclad radiation (Chuvashov and Riding 1984, text-fig. 7). These early dasyclads (Chuvashov and Riding’s Assemblage I) subsequently declined, and they disappeared from the fossil record at about the same time as the cyclocrinitids. Calcareous dasyclads reappeared in the Carboniferous and rediversified (Chuvashov and Riding’s Assemblage II). Some of the late Palaeozoic dasyclads have been placed within the Cyclocriniteae (Pia 1927; Wood 1942), but such an assignment is probably incorrect. The history of the receptaculitids is quite different. They appeared in the early Ordovician and showed little change in abundance until late Devonian time (Chuvashov and Riding 1984, text-fig. 3). BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 589 PALAEOECOLOGY Cyclocrinitids were associated with typical Palaeozoic marine faunas, both along continental margins and in epicontinental seas. Palaeogeographic reconstructions indicate that most of them lived within 20° north and south of the equator (text-fig. 3); a few exceptions occur as far south as 45°. Palaeozoic reconstructions are still tentative, but the generally equatorial distribution is clear. Cyclocrinitids were encrusted by calcium carbonate, and presumably produced carbonate sediment after death. They were not restricted to carbonate environments, however, and they are abundant in certain siltstones and shales as well as in limestones and dolostones. € DASYCLADALES CYCLOCRINITEAE RECEPTACULITALES TEXT-FIG. 4. Comparative stratigraphic ranges of the Dasycladales, Cyclo- criniteae, and Receptaculitales in the Palaeozoic (after Chuvashov and Riding 1984. text-figs. 3 and 7). The main constraints on the local distribution of cyclocrinitids were probably light intensity and water energy. Cyclocrinitids were probably calcareous algae, and as such they would have required at least a certain minimum photon flux density for net photosynthesis. However, the intensity of submarine daylight is attenuated rapidly in even the clearest oceanic water; the attenuation is still more pronounced if there is any turbidity (Liining 1981). This factor must have prevented them from occupying waters below a critical depth. Silurian cyclocrinitids were probably quite sensitive to water energy as well. Their thalli were hollow, weakly calcified, and almost certainly rather fragile. They were commonly flattened during burial, and it seems unlikely that they could have withstood persistent turbulence. Furthermore, cyclocrinitids on soft substrates probably attached themselves to small solid objects such as pebbles, coral fragments or shells, much as modern dasyclads do. Any movement or resuspension of the substrate would have been potentially disastrous for the attached thallus. Therefore, Silurian cyclocrinitids probably avoided shallow, unprotected waters. Their fossils occur primarily in quiet-water deposits that formed below normal wave base. Storm-generated currents occasionally penetrate into deeper, normally quiet waters. The Silurian rocks of Iowa and Norway show the vulnerability of cyclocrinitids to such currents. Some beds contain large aggregations of well-sorted, spherical cyclocrinitid thalli, often packed tightly together in a single layer (Nitecki and Johnson 1978, fig. 4; Mork and Worsley 1980, fig. 5). These thalli were apparently detached, rolled about, and swept together by currents. However, post-mortem transport of cyclocrinitids was probably uncommon, since their thalli were relatively large and remained intact after death. Most calcareous green algae disintegrate after death, and the resulting fragments may be transported widely (Riding 1975). Cyclocrinitids were ecologically very similar to modern and ancient dasyclad algae. Dasyclads are also primarily equatorial, although Dasycladus and Acetahiilaria occur in the Gulf of Trieste at over 45° N. (Cinelli 1979). Recent dasyclads are usually, but not exclusively, associated with marine carbonate environments; dasyclad floras presently flourish on terrigenous sediments throughout the Mediterranean and along the main islands of Japan (Cinelli 1979; Arasaki and Shihira-Ishikawa 1979). Furthermore, Recent dasyclads also avoid rough water, and are usually found only below 590 PALAEONTOLOGY, VOLUME 29 wave base or in protected pools and lagoons (Wray 1977). A detailed study of fossil dasyclads by Elliott (1968) noted that they were extremely rare or absent in reef and shoal environments, but abundant in calm lagoons or coastal bays. Early dasyclads such as Vermiporella, Rhabdoporella, and Dasyporella are commonly associated with cyclocrinitids in Ordovician and Silurian limestones (Kiaer 1920; Hoeg 1961; J. H. Johnson 19616; Gauthier-Coulloudon and Mamet 1981). The receptaculitids were ecologically rather different. They were significant Palaeozoic reef- builders, and some may have been resistant to high water energies (Rietschel 1969; Nitecki 1972). Cyclocrinitids and receptaculitids rarely co-occur. CYCLOCRINITIDS AS DEPTH INDICATORS Cyclocrinitids were morphologically and ecologically very similar to Recent dasyclads, which suggests that they probably occupied similar depths. Great caution must be used in this approach (Riding 1975). It has often been assumed that Recent dasyclads are restricted to shallow water of less than 10 12 m depth (J. H. Johnson 1961<3; Konishi and Epis 1962), but in fact many have much greater ranges. In the Mediterranean, Dasycladus ranges down to 90 m (Edelstein 1964) and Acetabularia occurs at 30-40 m (Funk 1955). In the Caribbean, Acicularia has been dredged from 73 m, Dasycladus from 55 m, and Neomeris from 50 m (Taylor 1960). Divers have found Acetabularia and Neomeris on fore-reef slopes in Jamaica at depths over 30 m (Goreau and Goreau 1973, figs. 21 and 22). However, all of these forms are most abundant in much shallower water. These records suggest that any palaeocommunity containing cyclocrinitids most likely existed at depths no greater than 100 m, and probably less. Other studies of such late Ordovician and early Silurian communities have produced similar figures (Cocks and McKerrow 1984). An alternative approach is to use cyclocrinitids as indicators of relative (not absolute) depth. As noted earlier, light intensity is attenuated rapidly with increasing water depth, and its spectral composition changes as well. The morphologies of photosynthetic organisms often vary in response to the ambient light conditions. Such ecophenotypic variation is particularly well known in colonial reef corals, which normally require photosynthetic algae as symbionts (Graus and Macintyre 1976). In general, most algal growth is reduced or arrested at low photon flux densities, resulting in small, stunted adult plants (review in Norton et al. 1981 ). Recent calcareous green algae, both udoteaceans and dasycladaceans, are no exceptions. Udoteaceans such as Udotea and Halimeda often show abnormal, reduced growth when cultured under fluorescent lights; such lights are approximately equal in intensity to tropical waters of 40-50 m depth (Colinvaux et al. 1965; Hillis-Colinvaux 1980, pp. 164-165). Among dasyclads, only Acetabularia has been studied in detail. Its growth and morphogenesis are largely controlled by the intensity and wavelength of the ambient light (review in Puiseux-Dao 1970). Deep-water specimens of A. acetabulum from the Mediterranean have whorls of unusually small diameter (Cinelli 1979). Silurian cyclocrinitids are associated with a variety of substrates, sedimentary features, and invertebrate faunas; they apparently ranged throughout many distinct environments. However, neighbouring populations from different environments generally show marked differences in thallus size. The close resemblances between cyclocrinitids and dasyclad algae suggest that this size variation was probably due to differences in ambient light intensity. Therefore, cyclocrinitid size variation should be useful locally as a relative depth indicator. Populations with smaller thalli should indicate deeper, darker waters, while those with larger thalli should indicate shallower, brighter waters. CYCLOCRINITIDS FROM THE LLANDOVERY SERIES Early attempts to classify cyclocrinitids led to a profusion of genera and species, often based on preservational differences. The resulting confusion has been largely dispelled by Nitecki’s (1970, 1972) revision of the North American forms. The present study confirms Nitecki’s view that there were only two or three species of cyclocrinitids present during the Silurian; these are restricted to the Llandovery Series (text-fig. 5). BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 591 TEXT-FIG. 5. The stratigraphic distribution of Silurian Cyclocrinites. Hatched bars = common occurrence, dotted bars = rare occurrence. Llandovery Stages after Holland (1985). Wales: 1 = Gasworks Mudstone, Haverford ‘Stage’. Localities: Strahan et al. (1914). Correlation: Ziegler et al. (1974). Scotland: 2 = Mulloch Hill Formation. 3 = Glenwells Shale. 4 = Newlands Formation. Localities: Peach and Horne (1899), Cocks and Toghill (1973). Correlation: Ziegler et al. ( 1974). Norway: 5 = Solvik Formation. 6 = Rytteraker Formation. Localities: Kiaer (1908), Mork and Worsley (1980), Baarli (pers. comm.). Correlation: Baarli (1985). Anticosti Island: 7 = Becscie Formation. 8 = Gun River Formation. Locali- ties: Billings (1866), Twenhofel (1928). Correlation: Johnson et al. (1981). North-west Illinois: 9 = Kankakee Formation ( = Blanding Formation). 10 = Farmers Creek Member, Hopkinton Dolomite. Localities: Nitecki (1970, 1972). Correlation: M. E. Johnson (1983). Northern Michigan: 1 1 = Lower Laminated Beds, Schoolcraft Formation. Locality and Correlation: Johnson and Campbell (1980). Eastern Iowa: 12 = Farmers Creek Member, Hopkinton Dolomite. Eocalities and Correlation: M. E. Johnson (1983). North-east Illinois-South-east Wisconsin: 1 3 = Waukesha or Racine Eorma- tions. Localities: Whitfield (1882), Nitecki ( 1970). Correlation: Willman (1973), M. E. Johnson (1983). 592 PALAEONTOLOGY, VOLUME 29 Two species of Llandovery cyclocrinitids are now recognized in North America: Cyclocrinites gregarius in the lower Llandovery, and C. dactioloides in the upper Llandovery. The two may also be distinguished by the relative width of the lateral heads; their thalli reached similar dimensions, but the lateral heads of C. gregarius are considerably smaller. The North American species C. gregarius may be a junior synonym of the contemporary European species C. favus (Kiaer 1908; Nitecki 1970, p. 100). The thalli of C. favus also had relatively small lateral heads (compare text-fig. 2f and 2i). All Llandovery cyclocrinitids are associated with diverse invertebrate faunas, usually dominated by brachiopods. Several distinct Llandovery communities have been recognized, and their palaeo- environmental significance has been studied in detail (Ziegler 1965; Ziegler et al. 1968; Cocks and McKerrow 1984). These communities were zoned in bands parallel to the shoreline, and so it is thought that their distribution was controlled by factors related to water depth. The original model (for Wales) placed the Lingula community in the shallowest water, followed by the Eocoelia, Pentamerus, Stricklandia, and Clorinda communities with increasing depth and distance from shore. In other areas, different communities may substitute for these, particularly in shallow water (Berry and Boucot 1970; Ziegler et al. 1974; M. E. Johnson 1977). Llandovery invertebrate communities are commonly used to indicate relative water depths. Cyclocrinitid size variation provides an independent measure of relative depth and can be used to test the value of this approach. Llandovery cyclocrinitids are common in Iowa, Anticosti Island, TABLE 1. Size variation in Llandovery cyclocrinitid populations a, C. dactioloides. Farmers Creek Member, Hopkinton Dolomite, eastern Iowa Index Lower («) Upper (n) Change Thallus diameter (cm) 1-94(24) 2-78(36) +43-3%* * Lateral head width (cm) 0-18(24) 0-26(36) +44-4%* Lateral head thickness (cm) 0-14(11) 0-24(15) +71-4%* 6, C. gregarius (syntypes and hypotype). Becscie and Gun River Formations, Anticosti Island. Data from Bolton (1966), Nitecki (1970, table 4) Index Becscie (n) Gun River (/?) Change Thallus diameter (cm) 1-90(21) 2-86(8) +50-5%* c, C. favus. Leangen Member, Solvik Formation, Bjorkoya Index Lower (n) Upper (n) Change Lateral head width (cm) 0-19(61) 0-21(402) +10-5%* d, C. favus. Leangen Member, Solvik Formation, Malmoya Index Shales (?i) Siltstones («) Change Lateral head width (cm) 0-22(67) 0-24(39) +9-1 %t Means compared with one-tailed Mest. * Increase significant at 99% confidence level, t Increase significant at 95% confidence level. BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 593 Great Britain, and Norway (text-fig. 5); the following sections examine the evidence in each of these areas. Data base. The Iowan data came from sixty exceptionally well-preserved thalli of Cyclocrmites dactioloides which were collected from the Farmers Creek Member of the Hopkinton Dolomite in eastern Iowa (text-fig. 6). They are deposited at the Field Museum of Natural History in Chicago, Illinois. The Norwegian data came from field measurements of 463 thalli of C. favus in the Leangen Member of the Solvik Formation on Bjorkoya, and from 106 thalli in the Padda Member of the Solvik on Malmoya (text-fig. 8). Thirteen immature C. favus thalli ( = Mastopora sp. of Kiaer 1908) were collected from the upper Leangen Member in Asker, and deposited at the Field Museum. The data on C. gregarius from Anticosti Island and C. favus from Great Britain came from previously published species descriptions. All numerical data for this study have been deposited with the British Library, Boston Spa, Wetherby, Yorkshire L523 7BQ, UK, as a supplementary publication No. SUP 14025 (24 pages). The data are summarized in Table 1. Iowa. The Farmers Creek Member of the Hopkinton Dolomite crops out in eastern Iowa and northwestern Illinois (text-fig. 6). It contains C. ( = Cerionites) dactioloides in such abundance that it was formerly called the "Cyclocrinites Beds’ (M. E. Johnson 1983). The unit contains well-preserved dolomitized fossils, but sedimentary structures are generally absent. The lower third of the Farmers Creek is dominated by the deep-water brachiopod Stricklandia laevis ( = 5. lens ultima). Fragile fossils such as reteporiform bryozoans and crinoid calyxes are often preserved intact, suggesting low-energy, deep-water conditions. The upper Farmers Creek, however, contains a very ditferent fossil assemblage, dominated by the shallower water brachiopod Harpidium maquoketa, a close relative of Pentamerus. There is evidence of occasional current activity, probably storm-generated: the brachiopods are TEXT-FIG. 6. Stratigraphic and geographic distribution of Cyclocrinites dactioloides populations in eastern Iowa. Llandovery Stages after Holland (1985). 594 PALAEONTOLOGY, VOLUME 29 sometimes truncated by erosional surfaces, and the cyclocrinitids may be preserved in large, swept-together assemblages (Nitecki and Johnson 1978; M. E. Johnson 1977, 1980). The faunal evidence indicates that the upper Farmers Creek was deposited in signihcantly shallower water than the lower. The cyclocrinitids are generally preserved as internal moulds, and many have some complete lateral heads as well. The specimens from the upper Farmers Creek are much larger than those from the lower, by all size measures (Table la). The size ranges of the two populations are generally distinct, with little overlap (text-fig. 7). Some representative thalli from the two populations are illustrated (compare text-fig. 2g-i with 2j-l). The overlying Picture Rock Member of the Hopkinton Dolomite contains a very shallow coral- stromatoporoid community (thought to be the depth equivalent of the Eocoelia community). The tabulate corals and stromatoporoids here are usually flattened or lenticular, presumably in response to persistently high water energies. Until recently we believed that C. dactioloides was excluded from this community; however, Brian Witzke (pers. comm, 1984) has found three thalli in a drill core through an apparent coral community from the Picture Rock or the top of the Farmers Creek. The largest had an estimated diameter of 4-8-5-2 cm, which is larger than any of the thalli examined in this study or by Nitecki (1970, table 5). The discovery of such an unusually large thallus is consistent with its location in unusually shallow water. Anticosti Island. Cyclocrinitids are common in the Becscie and Gun River Formations of Anticosti Island, Quebec. These units consist predominantly of unaltered limestones with thin shaly partings, and the fossils and sedimentary structures are generally well preserved. The Becscie fauna is dominated by the pentamerid Virgiana, often in life position. The deeper water brachiopods Stricklandia and Clorinda occur rarely. More common are corals and stromatoporoids, especially 0.30 - a 0.25 - E o J3 as a> 0.20 - 0.1 5 - X X X • X X • XX X X X X XX XX XX)* X XX X 0.10- 0.05 — 1 .00 X • Pentamerid-associated population (N = 36) X Stricklandiid-associated population (N=24) 0.835 “I 1 1 1 1 1 I 1 I 1.50 2.00 2.50 3,00 3.50 Maximum thallus diameter (cm) TEXT-FIG. 7a, b. Size variation in Cyclocrinites dactioloides populations from eastern Iowa. The pentamerid-associated population is from the upper Farmers Creek Member, while the stricklandiid- associated population is from the lower Farmers Creek Member. Text-fig. lb is on facing page. BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 595 at the top of the unit where they may form small bioherms, although brachiopods are still more abundant (Twenhofel 1928; Bolton 1981; Petryk 1981). Ripple marks are common, but larger bedforms are absent. Intraclasts of granule to pebble size are reported as well; these are mostly flat and may be imbricated (Petryk 1981). Gun River deposition was initiated by a transgression, which Johnson et u/. ( 1 98 1 ) place at approximately an Aj position. Deep-water conditions apparently continued in eastern Anticosti, but in the west shallowing soon resumed, and in general the Gun River in this area resembles the underlying Becscie. However, differences in the fauna and the sedimentary structures suggest significantly shallower water conditions. Pentamerids are much less conspicuous: Virgiana occurs only rarely in the lower Gun River, and is replaced by stunted Pentamems ohlongiis (variety 'juvenalis'). The deep-water brachiopods Stricklandia and Clorinda are completely absent, but the generally shallow-water Brachyprion is abundant (Twenhofel 1928; Bolton 1972), Corals are also abundant, more so than in the Becscie, and locally they form common bioherms and biostromes (Twenhofel 1928; Petryk 1981). Ripple marking is common, including interference ripple marks. The upper Gun River also contains megaripples with wavelengths of 3-5 6 0 m and amplitudes of 15- 20 cm. Channel-fill deposits 1 m wide and 10-15 cm are known as well. Intraclasts are larger than in the Becscie, up to large cobble size (Barnes et al. 1981; Petryk 1981 ). Anticosti cyclocrinitids usually occur as complete, but often deformed, internal moulds. The size variation between the Becscie and Gun River populations is so great that they were long considerd separate species. They were first described by Billings (1866) as Pasceolus gregarius and P. intermedins respectively, but Twenhofel (1928) reassigned them to Cyclocrinites (their descriptions are reprinted in Nitecki 1970). The only morphological differences are thallus size and lateral head width; C. intermedins is larger in both respects. Nitecki’s measurements of the type fossils (1970, table 4) show that the mean thallus diameter of the C. inter- medins specimens is much greater than that of the C. gregarius specimens, and the size ranges do not even 596 PALAEONTOLOGY. VOLUME 29 overlap (Table 1^). Lateral head width apparently grew at a relatively slow rate, and their size variation in this regard is not as great. Nitecki considered the two species synonymous and assigned them both to C. gregarius. We agree that it is unreasonable to establish separate species based solely on size difference; the variation is much more probably ecophenotypic. Great Britain. Llandovery cyclocrinitids are known from the Craighead Inlier of the Girvan district, Ayrshire, Scotland, and at Haverfordwest in Pembrokeshire, Wales. These two populations existed at the same time, but they lived in very different environments. British cyclocrinitids were carefully examined by Salter (1851), who described them as NiduUtes favus (reprinted in Nitecki 1970). They were later redescribed as Mastopora fava, and then as C. favus (Currie and Edwards 1942; Nitecki 1970, p. 100). Salter’s specimens came from 'Mulloch Quarry’ ( = Mulloch Hill Formation) in Scotland and from ‘Haverfordwest’ ( = Gasworks Mudstone) in Wales. The specimens consist mostly of flattened thalli, but lateral head width can be used as an indicator of relative thallus size. Salter noted that the lateral heads of the Scottish specimens were ‘a line wide’ ( = tV or 0-21 cm), while those of the Welsh specimens were ‘something less’. The Mulloch Hill quarries contain well-preserved, diverse fossil communities, with conspicuous C. ( = NiduUtes) favus (Peach and Horne 1899). The classic quarry faunas, in sandstones with interbedded shales and siltstones, were examined by Cocks and Toghill (1973), who considered them all variants of the shallow-water Cryptothyrella community. This is considered a precursor to the Eocoelia community (Ziegler et al. 1974). The Haverfordwest material is from the Gasworks Mudstone of the Haverford ‘Stage’. In the type section, described in detail by Strahan et al. (1914, pp. 90-91 ), the deep-water brachiopod Stricklandia) = Stricklandinia) lens is clearly dominant. Cyclocrinites ( = NiduUtes) favus is common in association with Stricklandia. Both the Mulloch Hill and the Gasworks Mudstone indicate an Aj to A4 position (Ziegler et al. 1974). The faunal evidence suggests that the Gasworks Mudstone represented a deeper water environment, and the cyclocrinitids from this unit have smaller lateral heads than those from the shallow-water environ- ments of the Mulloch Hill. It would also be interesting to compare the Mulloch Hill cyclocrinitids with those of the overlying Newlands Formation, which contains C. favus ( = M. fava) in a deep-water community with Clorinda and Stricklandia (Cocks and Toghill 1973). There are presently no data available on the Newlands forms. Norway. Cyclocrinitids are common in the Solvik Formation of Oslofjorden in southeastern Norway (text- fig. 8). The Solvik ( = ‘Etage’ 6 of Kiaer 1908) consists primarily of shale with siltstone and limestone interbeds; the fauna and sedimentary structures suggest generally shallowing conditions throughout the unit (Johnson and Worsley 1982; Baarli 1985). The palaeoshoreline lay to the west and trended south-south-west to north-north- east. In relatively proximal areas (e.g. Asker), cyclocrinitids occur throughout most of the Solvik, but in more distal areas (e.g. Malmoya) they are found only near the top. Cyclocrinitids are usually preserved as flattened thalli which are conspecific with the British form Cyclocrinites favus (= M. fava of Kiaer 1908, 1920). As in Britain, lateral head width can be used as a measure of relative thallus size. Cyclocrinitids are most abundant on the island of Bjorkoya, in west-central Oslofjorden north of Holmestrand, where they occur with Stricklandia on shale bedding planes. Only the uppermost 24-5 m of the Solvik (the Leangen Member of Baarli, 1985) are exposed; Mork and Worsley (1980) found that this section represented gradually shallowing environments. The basal beds, representing the deepest conditions, have more siltstone, less shale and limestone, and fewer cyclocrinitids than the overlying beds (Mork and Worsley 1980, hg. 3). The mean lateral head width of C. favus thalli from the basal 4 m of section is significantly less than that of thalli from the upper beds (Table Ic). The difference is small (about 10%) but statistically significant due to the large number of samples. The lateral heads of C. favus grew slowly, and it is likely that a 10% increase in lateral head width reflects a somewhat larger increase in thallus diameter, probably 15-20%. The Malmoya section lies south of Oslo in Bunnefjorden. The upper Solvik (Padda Member of Worsley et al. 1 982) contains common cyclocrinitids, both in the predominant shales and in thin siltstone interbeds. An earlier study of Malmoya brachiopods by Worsley (1971) found that those in the shales were generally deposited in place, while those in the siltstone interbeds had been transported down from shallower, more proximal environments. The mean lateral head width of C. favus thalli from the shales is significantly less than that of the transported thalli in the siltstones (Table Id). The difference is about the same as that between the two populations on Bjorkoya. The size differences among the cyclocrinitids of Bjorkoya and Malmoya are relatively small, and this suggests that the depth differences were small as well. The faunal evidence also indicates that the depth changes were BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 597 TEXT-FIG. 8. Stratigraphic and geographic distribution of Cyclucrinites faviis populations in Oslofjorden, Norway. B = Bjorkoya, M = Malmoya. Solid bars = abundant C. favus, dotted lines = uncommon C.fuviis. Llandovery Stages after Holland (1985). relatively minor: all of the sampled cyclocrinitid populations discussed above occur with generally similar Strickkmdia communities. Cyclocrinitids are also known from the Asker area where they are preserved in two different ways. Typical llattened C. favus thalli occur, but small spherical internal moulds are common as well (text-tig. 2f). Kiaer (1908 p. 327; 1920) referred to these as M. sp. but never formally described them. These thalli are far smaller than typical C. favus thalli, and they probably represent immature individuals. Similar small spherical thalli occur rarely in the Mulloch Hill Formation of Scotland (Currie and Edwards 1942); these have internal structures which closely resemble those of immature Boruetella capitata. a Recent dasyclad (Elliott 1972). Such small thalli were especially vulnerable to currents, and they often occur in large, well-sorted groups. One block from Avlos in Baerum contains some 135 thalli packed tightly together in a single layer (Mork and Worsley 1980, hg. 5). The moulds consist of mud and bioclastic debris, and their preservation suggests that they were detached, transported along muddy bottoms, and swept together. Such groups are found only in proximal environments, notably Asker but also Bjorkoya. The Solvik is overlain by the Rytteraker Formation, which is primarily limestone with Pentamerus, corals, and stromatoporoids. It was apparently formed in shallower, high-energy water (Johnson and Worsley 1982). Cyclocrinitids are conspicuously absent from this unit, although a few specimens are known from the lower Rytteraker on Bjorkoya and Malmoya (Mork and Worsley 1980; Baarli, pers, comm.). CONCLUSIONS Silurian cyclocrinitids showed considerable intraspecific variation in adult thallus size. This variation was apparently ecophenotypic; populations in certain environments had significantly smaller thalli 598 PALAEONTOLOGY, VOLUME 29 than neighbouring populations. Cyclocrinitids are commonly considered calcareous green algae, and they were morphologically and ecologically similar to dasycladaceans. This suggests that the observed size variation was caused by differences in light intensity, since such differences have marked effects on the growth of Recent algae. Cyclocrinitid populations should therefore be useful locally as relative depth indicators; those with small thalli should represent deeper, darker environments, while those with large thalli should indicate shallower, brighter environments. In all cases the relative depths indicated by cyclocrinitid size variation are the same as those indicated by the invertebrate communities and sedimentary features. In Iowa, Anticosti Island, and Great Britain, large changes in cyclocrinitid thallus size are associated with obvious changes in the composition of the associated invertebrate communities. On Bjorkoya and Malmoya in Norway the size change is much less, and the faunal changes are minor as well. The distribution of Llandovery cyclocrinitid algae reinforces the hypothesis that Llandovery invertebrates were largely zoned by factors related to water depth. The absolute depths inhabited by these communities are more difficult to define. Boucot (1975 fig. 1 5) considered the Strickkmdia community ( = Benthic Assemblage 4) to lie below the photic zone, but the abundance of cyclocrinitids in the Strickkmdia communities of Iowa, Wales, and Norway suggests otherwise. Cyclocrinitids are even known from a Clorinda-Stricklaiidia community in Scotland (Cocks and Toghill 1973, p. 215). By analogy with Recent dasyclads, it is probable that these cyclocrinitids lived at depths of less than 100 m. Acknowledgements. This study began as S.C. B.s bachelor’s thesis (Williams College). S.C. B. is primarily responsible for the Norwegian data, and M.E. J., for the Iowan data. The staff of the Paleontologisk Museum (University of Oslo) generously offered us advice and field equipment; we are especially grateful to Gudveig Baarli and David Worsley. The Field Museum in Chicago loaned the Iowan material. We are indebted to Brian Witzke for additional information on Iowan cyclocrinitids, and to William Fox, John MacFadyen, and Katherine Ayer for general criticism and assistance. We thank Robert Riding for his comments on an earlier draft. Acknowledgement is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of S. C. B.s field work in Norway during the summer of 1982. M. F. J. is grateful for concurrent support in Norway from a Fulbright-Hays Research Fellowship and the Royal Norwegian Council for Scientific and Industrial Research. RFFFRFNCFS ARASAKi, s. and SHiHiRA-iSHiiCAWA, I. 1979. Distribution and ecology of Acetabiilaria in Japan. In bonotto, s., KEFELi, V. and PUISEUX-DAO, s. (eds.). Developmental Biology of Acetabiilaria, 15-18. Flsevier/North-Holland Biomedical Press, Amsterdam. BAARLI, B. G. (1985). The stratigraphy and sedimentology of the early Llandovery Solvik Formation in the Central Oslo Region, Norway. Norsk geol. Tidsskr. 65, 229-249. BARNES, c. R., PETRYK, A. A. and BOLTON, T. E. 1981. Aiiticosti Island, Quebec. In lesperance, p. j. (ed.). Siibcommission on Silurian Stratigraphy, Ordovician- Silurian Boundary working group. Field meeting, Anticosti-Gaspe, Quebec 1981, I: Guidebook, 1-24. BERRY, w. B. N. and BOUCOT, A. J. 1970. Correlation of the North American Silurian rocks. Spec. Pap. geol. Soc. Am. 102, 289 pp. BILLINGS, E. 1866. Catalogues of the Silurian fossils of the Island of Anticosti, with descriptions of some new genera and species. Spec. Rep. geol. Surv. Can. 93 pp. BOLTON, T. E. 1966. Catalogue of type invertebrate fossils of the Geological Survey of Canada. III. Geol. Surv. Can. Dept. Mines and Tech. Surveys, 203 pp. 1972. Geological map and notes on the Ordovician and Silurian litho- and biostratigraphy, Anticosti Island, Quebec. Geol. Surv. Pap. Can. 71-19, 45 pp. 1981. Ordovician and Silurian biostratigraphy, Anticosti Island, Quebec. In lesperance, p. j. (ed.). Sub- commission on Silurian Stratigraphy, Ordovician-Silurian Boundary working group. Field meeting, Anticosti- GaspC Quebec 1981, II: Stratigraphy and Paleontology, 41-59. BOUCOT, A. J. 1975. Evolution and Extinction Rate Controls, All pp. Flsevier, Amsterdam. CAMPBELL, K. s. w., HOLLOWAY, D. J. and SMITH, w. D. 1974. A new receptaculitid genus, Hexabactron, and the relationships of the Receptaculitaceae. Palaeontographica, AI46, 52-77. BEADLE AND JOHNSON: SILURIAN C YC LOG R I NITI D ALGAE 599 CHUVASHOV, B. and riding, r. 1984. Principal floras of Palaeozoic marine calcareous algae. Palaeontology 27, 487-500. ciNELLi, F. 1979. Acetabularia acetahiilum (L.) Silva, Acetahukiria parvula Solms-Laubach and Dasycladus vennicularis (Scopoli) (Chlorophyta, Dasycladaceae): ecology and distribution in the Mediterranean Sea. In BONOTTO, s., KEFELi, V. and PUISEUX-DAO, s. (eds.). Developmental Biology of Acetabularia, 3-14. Elsevier/ North Holland Biomedical Press, Amsterdam. COCKS, L. R. M. and mckerrow, w. s. 1984. Review of the distribution of the commoner animals in lower Silurian marine benthic communities. Palaeontology, 27, 663-670. and TOGHiLL, p. 1973. The biostratigraphy of the Silurian rocks of the Girvan District, Scotland. Jl geol. Soc. Land. 129, 209-243. COLINVAUX, L. H., WILBUR, K. M. and WATABE, N. 1965. Tropical marine algae: growth in laboratory culture. J. P by col. 1, 69-78. CURRIE, E. D. and Edwards, w. n. 1942. Dasycladaceous algae from the Girvan area. J! geol. Soc. bond. 98, 235-240. edelstein, t. 1964. On the sublittoral algae of the Haifa Bay area. Vie Milieu, 15, 177-212 ELLIOTT, G. F. 1968. Permian to Palaeocene calcareous algae (Dasycladaceae) of the Middle East. Bull. Br. Mils. nat. Hist. (Geol.) Suppl. 4, 1 1 1 pp. 1972. Lower Palaeozoic green algae from southern Scotland, and their evolutionary signiflcance. Ibid. 22, 355-375. FISHER, d. c. and nitecki, m. h. 1982. Receptaculitids are problematic Paleozoic algae. J. Paleont. 56, Suppl. to No. 2, 10. FUNK, G. 1955. Beitrage zur Kenntnis der Meeresalgen von Neapel zugleich mikrophotographischer Atlas. Pubbl. Staz. zool. Napoli 25, Suppl., 166 pp. GAUTHIER-COULLOUDON, L. and MAMET, B. L. 1981. Algues et sedimentation carbonatee dans I’Ordovicien et le Silurien de file d’Anticosti. In lesperance, p. j. (ed.). Subcommission on Silurian Stratigraphy, Ordovician- Silurian Boundary working group. Field meeting, Anticosti-Gaspe, Quebec 1981, II: Stratigraphy and Paleontology, 87-90. GOREAU, T. F. and GOREAU, N. I. 1973. The ecology of Jamaican coral reefs. II. Geomorphology, zonation, and sedimentary phases. Bull. mar. Sci. 23, 399-464. GRAUS, R. R. and macintyre, I. G. 1976. Light control of growth form in colonial reef corals: computer simulation. Science, Washington, 193, 895-897. HERAK, M., KOCHANSKY-DEViDE, V. and Gusic, I. 1977. The development of the dasyclad Algae through the ages. In FLUGEL, E. (ed.). Fossil Algae, Recent Results and Developments, 143-153. Springer-Verlag, Berlin. HILLIS-COLINVAUX, L. 1980. Ecology and taxonomy of Halimeda: primary producer of coral reefs. Adv. mar. Biol. 17, 1-327. H0EG, o. A. 1961. Ordovician algae in Norway. Colo. Sch. Mines Q. 56, no. 2, 1 101. HOLLAND, c. H. 1985. Series and stages of the Silurian System. Episodes, 8, 101 -103. JOHNSON, J. H. 1961«. Limestone-budding Algae and Algal Limestones, 297 pp. Publ. Colo. Sch. Mines, Golden. 1961/). Review of Ordovician algae. Colo. Sch. Mines Q. 56, I -101 . JOHNSON, M. E. 1977. Succession and replacement in the development of Silurian brachiopod populations. Lethaia, 10, 83-93. 1980. Paleoecological structure in Early Silurian platform seas of the North American Midcontinent. Palaeogeogr. Palaeoclimat. Palaeoecol. 30, 191216. 1983. New member names for the Lower Silurian Hopkinton Dolomite of Eastern Iowa. Proc. Iowa Acad. Sci. 90, 13-18. and CAMPBELL, G. T. 1980. Recurrent carbonate environments in the Lower Silurian of Northern Michigan and their inter-regional correlation. J. Paleont. 54, 1041 1057. COCKS, L. R. M. and copper, p. 1981. Late Ordovician-early Silurian fluctuations in sea level from eastern Anticosti Island, Quebec. Lethaia, 14, 73-82. and WORSLEY, d. 1982. Chronology and depositional environments of Early Silurian sequences in the Oslo region. In worsley, d. (ed.). lUGS Subcommission on Silurian Stratigraphy. Field meeting, Oslo region 1982. Palaeontological Contributions from the University of Oslo no. 278, 149-160. KiAER, J. 1908. Das Obersilur im Kristianiagebiete: eine stratigraphisch-faunistische Untersuchung. Skr. Vidensk Selsk. Christiania, Mat. Nat. Kl. 1906 II, 596 pp. — 1920. Oversigt over kalkalgefloraene i Norges Ordovicium og Silur. Norsk geol. Tidsskr. 6, 1 1 3-142. KONiscHi, K. and epis, r. c. 1962. Some early Cretaceous calcareous algae from Cochise County, Arizona. Micropaleontology, 8, 67-76. 600 PALAEONTOLOGY, VOLUME 29 LUNiNG, K. 1981. Light. In lobban, c. s. and Wynne, m. j. (eds.). The Biology of Seaweeds, 326-355. University of California Press, Berkeley. MORK, A. and WORSLEY, D. 1980. The environmental significance of algae in the middle Llandovery succession of the central Oslo Region. Lethaia, 13, 339-346. NiTECKi, M. H. 1970. North American cyclocrinitid algae. Fieldiana, Geol. 21, 182 pp. 1972. North American Silurian receptaculitid algae. Ibid. 28, 108 pp. 1976. The systematic position of cyclocrinitids. Geol. Soc. Am. Abstracts with Programs, 8, 500-501. and JOHNSON, M. E. 1978. Internal structures of Cyclocrinites dactioloides, a receptaculitid alga from the Lower Silurian of Iowa. Fieldiana, Geol. 39, 1-15. NORTON, T. A., MATHiESON, A. c. and NEUSHUL, M. 1981. Morphology and environment. In lobban, c. s. and WYNNE, M. J. (eds.). The Biology of Seaweeds, 421-451. University of California Press, Berkeley. OSGOOD, R. G. and eischer, a. g. 1960. Structure and preservation of Mastopora pyriformis, an Ordovician dasycladacean alga. J. Paleont. 34, 896-902. PEACH, B. N. and HORNE, J. 1899. The Silurian rocks of Britain. Vol. I: Scotland. Mem. geol. Stirv. U.K. 749 pp. PETRYK, A. A. 1981 . Stratigraphy, sedimentology and paleogeography of the Upper Ordovician-Lower Silurian of Anticosti Island, Quebec. In lesperance, p. j. (ed.). Suhcommission on Silurian Stratigraphy, Ordovician- Silurian Boundary working group. Field meeting, Anticosti-Gaspe, Quebec 1981, II: Stratigraphy and Paleontology, 11-39. PIA, J. 1927. Abteilung: Thallophyta. In hirmer, m. Handbuch der Paldobotanik, Bd. I, 31-136. Oldenbourg, Miinchen. PUISEUX-DAO, s. 1970. Acetabularia and Cell Biology, 162 pp. Springer- Verlag, New York. REED, E. R. c. 1912. Ordovician and Silurian fossils from the central Himalayas. Mem. geol. Surv. India Palaeontol. indica, ser. 15, vol. 7, no. 2, 168 pp. RIDING, R. 1975. Girvanella and other algae as depth indicators. Lethaia, 8, 173-179. 1977. Problems of affinity in Palaeozoic calcareous algae. In elugel, e. (ed.). Fossil Algae, Recent Results and Developments, 202-21 1. Springer-Verlag, Berlin. 1984. Sea-level changes and the evolution of benthic marine calcareous algae during the Palaeozoic. Jl geol. Soc. Loud. 141, 547-553. RiETSCHEL, s. 1969. Die Receptaculiten. Senckenberg. leth. 50, 465-517. 1977. Receptaculitids are calcareous algae but no dasyclads. In elugel, e. (ed.). Fossil Algae, Recent Results and Developments, 212-214. Springer-Verlag, Berlin. and NiTECKi, M. H. 1984. Ordovician recaptaculitid algae from Burma. Palaeontology, 27, 415-420. SALTER, J. w. 1851. List of some of the Silurian fossils of Ayrshire. In Murchison, r. i. On the Silurian rocks of the South of Scotland. Jl geol. Soc. Lond. 7, 170-178. SCOTESE, c. R. 1984. An introduction to this volume: Paleozoic paleomagnetism and the assembly of Pangea. In VAN DER voo, R., SCOTESE, c. R. and BONHOMMET, N. (eds.). Plate Reconstruction from Paleozoic Paleo- magnetism, 1-10. Geodynamics Series, 12. American Geophysical Union, Washington, DC. STRAHAN, A., CANTRiLL, T. c., DIXON, E. E. L., THOMAS, H. H. and JONES, o. T. 1914. The geology of the South Wales coalfield, Pt. XI, the country around Haverfordwest. Mem. geol. Surv. England and Wales, sheet 228, new series, 262 pp. TAYLOR, w. R. 1960. Marine Algae of the Eastern Tropical and Subtropical Coasts of the Americas, 870 pp. University of Michigan Press, Ann Arbor. TWENHOEEL, w. H. 1928. Geology of Anticosti Island. Mem. geol. Surv. Can. 154, 481 pp. WHITFIELD, R. p. 1882. Description of Fossils. In Geology of Wisconsin, Survey of 1873-1879, IV, pt. Ill, pp. 161-363. wiLLMAN, H. B. 1973. Rock Stratigraphy of the Silurian System in Northeastern and Northwestern Illinois. Illinois State geol. Surv. Circ. 479, 55 pp. WOOD, A. 1942. The algal nature of the genus Koninckopora Lee; its occurrence in Canada and Western Europe. Jl geol. Soc. Lond. 98, 205-222. WORSLEY, D. 1971. Faunal anticipation in the Lower Llandovery of the Oslo Region, Norway. Norsk geol. Tidsskr. 51, 161-167. AARHUS, N., BASSETT, M. G., HOWE, M. p. A., MORK, A. and OLAUSSEN, s. 1982. The Silurian succession of the Oslo Region. Nor. Geol. Unders. 384, 1-57. WRAY, J. L. 1977. Calcareous Algae, 185 pp. Elsevier, Amsterdam. ZIEGLER, A. M. 1965. Silurian marine communities and their environmental significance. Nature, Lond. 207, 270-272. BEADLE AND JOHNSON: SILURIAN CYCLOCRINITID ALGAE 601 COCKS, L. R. M. and bambach, r. k. 1968. The composition and structure of Lower Silurian marine communities. Lethaia, 1, 1-27. RICKARDS, R. B. and MCKERROW, w. s. 1974. Correlation of the Silurian rocks of the British Isles. Spec. Pap. geol. Soc. Am. 154, 154 pp. STEVEN C. BEADLE Department of Earth and Planetary Sciences The Johns Hopkins University Baltimore Maryland 21218 Typescript received 24 April 1985 Revised typescript received 31 October 1985 MARKES E. JOHNSON Department of Geology Williams College Williamstown Massachusetts 01267 V «<. , I ) i:i 5= i’ I •« A NEW ANTHRACOSAUR AMPHIBIAN FROM THE CARBONIFEROUS OF SCOTLAND by T. R. SMITHSON Abstract. The anthracosaur amphibian Proterogyrinus pancheni, sp. nov., from localities in the Namurian of the Scottish Carboniferous, is the earliest known member of the Embolomeri to be described from Europe. It closely resembles the North American form P. scheelei, but is distinguished by differences in morphology of the dentition and vertebrae. Functional explanations for the presence of the large Meckelian fenestrae in the embolomere mandible are reconsidered, but it is suggested that they had no specific function and represent incomplete ossification of the mesial surface of the lower jaw. Recent discussions of anthracosaur systematics are reviewed. The newly proposed schemes are supported by very few synapomorphies and it is concluded that a solution to the problem of anthracosaur phytogeny will be found only within the framework of a larger study of the interrelationships of early tetrapods. In 1980 I published a preliminary account of the tetrapod fauna in a richly fossiliferous bone bed from the Dora opencast site, near Cowdenbeath, Fife, Scotland (Smithson 1980a). Since then, much of this material has been fully described (Smithson 1980/), 1985a; Panchen 1985), but still awaiting description are the specimens I placed in the anthracosaur assemblages 3 and 4 (Smithson 1980a, p. 420). The two assemblages include bones from different parts of the skeleton, and in 1980 I was uncertain whether they represented separate species or were partial skeletons of the same species. Each assemblage was found to share a number of derived features with a North American anthra- cosaur Proterogyrinus scheelei Romer, but at the time, only preliminary descriptions of Protero- gyrinus had been published (Romer 1970; Hotton 1970) and the significance of these similarities could not be assessed. P. scheelei was found in the Upper Mississippian Bickett Shale at Greer, West Virginia (Romer 1970). The Bickett Shale forms part of the Bluefield Formation which is equivalent in age to the uppermost Visean (Lower Carboniferous) and the lowermost Namurian (Upper Carboniferous) of Europe (see Panchen 1970, table 3). The fossiliferous deposits at Greer are therefore roughly contemporaneous with those at the two principal Scottish Namurian tetrapod localities, Dora and Loanhead (Smithson 1985/)). Following the recent complete description of P. 5c/)cc/c/ (Holmes 1984), it is now clear that all the material in assemblages 3 and 4 from Dora belongs to the same species. It is described here as a new species of Proterogyrinus and it represents the earliest known member of the Embolomeri to be found in the Carboniferous of Europe. MATERIALS AND METHODS The specimens used in the deseription of P. pancheni sp. nov. are listed below. The following abbreviations are used for institutions housing the material: BM(NH), Department of Palaeonto- logy, British Museum (Natural History), London; NUZ, Department of Zoology, University of Newcastle upon Tyne; RSM, Department of Geology, Royal Scottish Museum, Edinburgh. The grid reference (GR) given for specimens from Dora refers to the site map published elsewhere (Smithson 1985a, text-fig. 2). Dora Bone Bed, Dora opencast site, near Cowdenbeath, Fife Region. BM(NH) R9940 GR 88 Incomplete interclavicle. NUZ 75.1 1.2 GRX12 Skull table and interorbital region. [Palaeontology, Vol. 29, Part 3, 1986, pp. 603-628.) 604 PALAEONTOLOGY, VOLUME 29 NUZ 77.1.6 GR WIO Left lacrimal. NUZ 77.3.8 GRXIO Right intertemporal. NUZ 76.10.17 GR D51 Incomplete intercentrum. NUZ 77.2.17 GR WIO Cervical neural arch. NUZ 78.3.33 GRC32 Caudal neural arch. NUZ 77.2.19 GRXll Incomplete left clavicle. NUZ 77.2.20 GR UlO Incomplete right humerus. RSM GY 1977.46.33 GR89 Right mandible, trunk vertebra, and incomplete rib RSM GY 1975.48.49 GR - Trunk vertebra. RSM GY 1977.46.35 GR D30 Trunk vertebra. RSM GY 1978.4.21 GR - Trunk vertebra. RSM GY 1983.9.1 GR L66/67 Incomplete trunk vertebra. RSM GY 1976.19.48 GR K20 ?Sacral vertebra. RSM GY 1977.46.34 GRD30 Pleurocentrum. RSM GY 1976.19.49 GR Intercentrum. RSM GY 1983.9.2 GR H39 Haemal arch. Burghlee Ironstone (Rumbles Ironstone, see Smithson 1985/r), Burghlee colliery, Loanhead, Lothian Region. BM(NH) R3960 Incomplete maxilla. BM(NH) R4085 Right femur. Shale overlying South Parrot Coal, Niddrie colliery, Niddrie, Lothian Region. RSM GY 1893.135.84 Pleurocentrum. Dora specimens from grid squares U, W, X were included in assemblage 3 (Smithson 1980^), the remaining material formed assemblage 4. The Dora Bone Bed was prepared initially by the hot water technique developed by Mr Stanley Wood and the author (see Boyd and Turner 1980, p. 20). Further preparation of specimens was by mounted needles and an industrial airbrasive machine using sodium bicarbonate powder. Most specimen drawings were prepared using a camera lucida, but the illustrations of the skull table and lower jaw were based on photographs. SYSTEMATIC DESCRIPTION Order anthracosauria Suborder anthracosauroideae Infraorder embolomeri Family proterogyrinidae Romer, 1970 Diagnosis. As for Proterogyrinus. Genus proterogyrinus Romer, 1970 Type species. P. scheelei Romer, 1970 Diagnosis. (Based on information in Holmes 1984 and author’s studies.) Primitive embolomeres probably growing to about 1 -5 m in length. Skull structure similar to that of Palaeoherpeton and Pholiderpeton but with a relatively shorter antorbital region. Median suture posterior to the pineal on a ridge flanked on either side by a depression. Kinetic junction extends the length of the skull table as in Eoherpeton. Biramous tabular horn. Jugal exposed on ventral skull margin. Maxilla makes no sutural contact with premaxilla. Laterosphenoid region of braincase unossified. Thirty-two presacral vertebrae. Atlas pleurocentrum incompletely ossified ventrally as well as dorsally. More posterior vertebrae gastrocentrous with disc-shaped pleurocentra and crescentic intercentra. Tall neural spines with areas of origin and insertion of median dorsal axial musculature well defined. Appendicular skeleton similar to that of Archeria, but with two foramina piercing the puboischiadic plate, instead of the single obturator foramen, and a poorly developed iliac blade. Four centralia in tarsus, the most proximal (fourth) being partly fused to tibiale. SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 605 Proterogyrinus pancheni sp. nov. Diagnosis. As for genus, plus: parietal-postparietal suture on a ridge. Dentary with spaces for at least thirty-two sharply pointed, incurved, and strongly hooked teeth. Coronoid series covered entirely with denticles. Pleurocentra in trunk region well ossified and fused dorsally, small notochordal foramen. Hololype. RSM GY 1977.46.33. Incomplete right ramus of lower jaw, trunk neural arch with articulated pleurocentrum, and incomplete rib. Referred material. See Materials and Methods. Type Horizon and Locality. Dora Bone Bed. Localized seatrock, beneath a coal seam below the Lochgelly Blackband Ironstone, upper part of the Limestone Coal Group (Namurian A, Upper Carboniferous), Dora opencast site, Cowdenbeath, Fife Region, Scotland. Distribution. Lothian and Fife Regions, Scotland. Range. Limestone Coal Group (Dora Bone Bed) to Upper Limestone Group (South Parrot Coal Shale). Ej zone of Namurian A (Upper Carboniferous). Derivation of specific name. After Dr Alec Panchen in recognition of his considerable work on Carboniferous anthracosaurs. Description. The skull of P. pancheni is represented by the skull table and interorbital region NUZ 75.1 1 .X illustrated by Andrews et at. (1977, text-fig. 4), a displaced right intertemporal NUZ 77.3.8, a left lacrimal NUZ 77.1.6, and an incomplete right ramus of the lower jaw RSM GY 1977.46.33. Skull table. The overall shape and pattern of the skull roofing bones of P. pancheni (text-fig. 1a, b) is very similar to that of P. scheelei (Holmes 1984), and only brief notes on the general arrangement are given. The dorsal surface of the bones is ornamented with an irregular series of shallow pits and grooves which undercut the bone surface. This ornamentation is more strongly developed than in the contemporaneous form Eoherpeton (Smithson 19856), and most closely resembles that in the Coal Measure embolomeres Palaeoherpeton (Panchen 1964) and Pholiderpeton (Panchen 1972). (I have accepted Dr Jennifer Clack’s (nee Agnew) conclusion, following her review of British Coal Measure embolomeres, that Eogyrinus is the junior synonym of Pholiderpeton (Agnew 1984).) All the material from the Westphalian B previously attributed by Panchen (1972) to Eogyrinus attheyi is here referred to as P. attheyi). Lateral line sulci are absent. As preserved the skull table and interorbital region is 95 mm long from the anterior edge of the frontals to the posterior edge of the postparietals, and 57 mm wide in the region of the supra- temporals. The bones have been displaced following post-mortem compression and collecting disturbance. The right intertemporal bone is missing from the skull table but was recovered from the bone bed approximately 0-5 m to the left of it. The ventral surface of NUZ. 75.1 1.2 is exposed and permits detailed comparison with the ventral surface of the skull table and interorbital region of Palaeoherpeton (Panchen 1964, text-fig. 3). The most notable features of the dorsal surface of NUZ 75.1 1.2 are the ridged midline sutures of the parietals (behind the pineal foramen) and postparietals, and the ridged transverse sutures between the parietals and postparietals. Ridged midline sutures are also present in Proterogyrinus scheelei, an undescribed specimen from Point Edward, Nova Scotia and Pteroplax (Holmes, 1984), but ridged transverse sutures are absent. Ridged sutures have not been observed in other anthracosaurs and are here regarded as an autapomorphy (uniquely derived feature) of the Proterogyrinidae. In Proterogyrinus scheelei and the Point Edward specimen the dorsal margin of the pineal foramen also bears a raised rim, but this is absent in P. pancheni. In Carboniferous anthracosaurs the tabular bone exhibits two diagnostic features of the skull roof: a connection with the parietal bone and a tabular horn. In NUZ 75.1 1.2 the dorsal surface of the tabular is overlapped by the parietals anteriorly and the postparietals mesially. The postparietals are themselves overlapped by the parietals and this arrangement influences the relationship of the tabular and parietal on the ventral surface of the skull table. The characteristic tabular-parietal suture is PALAEONTOLOGY, VOLUME 29 606 TEXT-FIG. 1. Proterogyrinus pancheni sp. nov., skull table and interorbital region restored. A, dorsal view, b, ventral view, c, posterior view. Natural size. Broken bone hatched. Abbreviations: f, frontal; it, intertemporal; p, parietal; pf, postfrontal; pp, postparietal; st, supratemporal; t, tabular. evident only on the left side (ventral right) where the anterolateral corner of the tabular meets the parietal. On the right a narrow process of the postparietal extends laterally between the tabular and parietal to contact the supratemporal (text-fig. 1b). This difference in the pattern of bones on the dorsal and ventral surfaces of the skull table has not been observed previously in anthracosaurs. The tabular horns are incomplete on both sides of the skull table, but on the left only the lateral edge of the horn is damaged. It appears that a blade-like posterior process of the type present in SMITHSON: NEW CARBONIFEROUS ANTH R ACOS AU R 607 Palaeoherpeton (Panchen 1964) and Pteroplax (Panchen 1970) was not developed in Proterogyrinus pancheni. In P. scheelei the tabular horn is biramous with a short accessory process projecting from the ventrolateral edge of the tabular (Holmes 1984). Because this region is damaged in NUZ 75.1 1 .2 it is not possible to determine whether the horn was similarly biramous in P. pancheni. The morphology of the ventral surface of the tabulars in NUZ 75.1 1.2 is very similar to that in Palaeoherpeton (Panchen 1964, text-fig. 3). A deep flange projects from the ventromesial edge of the tabular and forms a flat unornamented surface, which is oriented both posteroventrally and ventro- laterally. Its mesial edge is unfinished and forms part of a surface of attachment for the otic capsules which extend around the entire anteroventral edge of the tabular. In addition, a stout process projects from the centre of the bone which is unfinished anteromesially and forms a second area of attachment for the otic capsules. Between the flange and the process is a deep groove which Panchen (1964, p. 602) suggested carried a vena capitis lateralis, but which Holmes (1984, p. 457) has argued carried the vena capitis dorsalis. In front of the tabulars the ventral surface of the parietals bears a pair of shallow V-shaped depressions, one on either side of the convex median portion of the skull table. Each recess extends forward to a point level with the posterior edge of the pineal foramen and they appear to mark the area of attachment of the otic capsules. A similar impression is present on the ventral surface of Palaeoherpeton (Panchen 1964, text-fig. 3) and also Loxomma acutirhinus (Beaumont 1977, text- fig. 2), although in neither does it extend forward to the pineal foramen. The ventral surface of each postparietal is excavated to form a short narrow transverse furrow medial to the postparietal-tabular suture. These furrows are absent in Palaeoherpeton and their function is unknown. Behind and lateral to the pineal foramen a pair of short, ill-defined ridges run forward on either side of the pineal towards the anterior edge of the parietal. The ridge on the right is more strongly developed than that on the left. After a break in the ridge system between the orbit margins it con- tinues forward on the lateral edge of the frontal bones. Here the ridge pair is strongly developed and forms the boundaries of a shallow trough on the undersurface of the frontals (text-fig. 1b). This ridge system almost certainly delimits the area of attachment between the sphenethmoid region of the braincase and skull roof. These ridges are absent in Palaeoherpeton but in Proterogyrinus scheelei, Seymouria, and primitive reptiles, for example, Eocaptorhinus (Heaton 1979), they support the thin walls of a Y-shaped sphenethmoid. A short distance behind the anterior end of the frontals the lateral edges of the bones are excavated to form a pair of shallow concavities bounded mesially by the ridges for the sphenethmoid. The anterior end of each concavity is pierced by a small foramen. Their function is unknown. At the lateral edge of the skull table the ventral surface of the supratemporal is excavated to form a shallow groove which extends along most of the length of the bone. In the complete skull it would have been occupied by the dorsal edge of the squamosal to form the characteristic ‘kinetic’ joint between the skull table and cheek. The groove is incised with a number of deep pits, not seen in P. scheelei, which probably served as areas of attachment for connective tissue which stabilised the joint (Panchen 1964). A shallow concavity on the lateral edge of the squamosal may also have formed part of this joint. Further posteriorly, the lateral edge of the skull table is strongly concave. This region is traditionally taken to represent the anterodorsal margin of an otic notch. However, a number of authors (e.g. Smithson 1982, 1985u; Clack 1983; Holmes 1984) have suggested recently that Carboniferous anthracosaurs lacked a tympanum, and this region may be a vestige of the spiracular cleft (Smithson 1982; Panchen 1985). Lacrimal. An incomplete bone NUZ 77.1.6 recovered a short distance behind the skull table closely resembles the lacrimal of Pholiderpeton (Panchen 1972, text-fig. 6). It resembles no other bone in the anthracosaur skull or mandible. Its external surface is gently convex and exhibits the characteristic pitted ornament of anthracosaur dermal roofing bones. The thickened ventral edge of the bone is deeply incised with grooves which probably represent the area of contact with the maxilla. The posteroventral portion of the bone is missing but the external surface of the posterodorsal region is excavated to form a V-shaped area, exhibiting numerous ridges and grooves, which probably formed 608 PALAEONTOLOGY, VOLUME 29 an area of sutural overlap with the prefrontal or jugal. The dorsal margin of the bone is incomplete and I could not determine whether the lacrimal was included or excluded from the anterior border of the orbit. Similarly, I was unable to judge whether the anterior end of the bone formed part of the border of the external naris. The internal surface of the lacrimal is divided by a braided longitudinal ridge which extends forward from the posterior end of the bone. Above the ridge the bone surface is mostly smooth although at the posterior end a small area of sutural overlap is developed. Below the ridge the surface is traversed by a second oblique deeply pitted ridge which probably had a sutural connection with either the maxilla or a lateral palatal bone. Lower jaw. The lower jaw of Proterogyrinus pancheni is represented by the type specimen. It is incomplete anteriorly and the external surface is badly fractured and incomplete in the region overlying the adductor fossa. The surface of the bone has been removed during preparation and very little detail of the ornamentation or the possible course of the mandibular lateral line canal can be distinguished. The internal surface of the anterior two thirds of the jaw was exposed by the author. The mesial wall of the jaw behind the anterior border of the anterior Meckelian fenestra has collapsed into the Meckelian space and has also been pushed forward, in part, between the presplenial and dentary. In addition the whole ramus has been laterally compressed forcing the coronoid series over the mesial shelf of the dentary. With one exception all the teeth in the dentary are incomplete. In its general morphology the lower jaw is similar to those of the P. scheelei (Holmes 1984), Pholiderpeton (Panchen 1970, 1972; Agnew 1984), Eobaphetes (Panchen 1977), and Neopteroplax (Romer 1963). It exhibits the characteristic surangular crest and has a pair of moderately large Meckelian fenestrae on the mesial surface of the jaw. As preserved it is 144 mm long, reaches a maximum depth of 45 mm beneath the surangular crest and tapers to a depth of 1 5 mm at its incom- plete anterior end. In front of the adductor fossa the jaw is approximately parallel sided in dorsal view. There is no apparent curvature in this region and it is reasonable to assume that when articu- lated with its antimere and two rami described a V as in Pholiderpeton (Panchen 1972, text-fig. 12). An outline drawing of the external surface of RSM GY 1977.46.33, as preserved, is given (text- fig. 3a) showing the course of discernable sutures. The extent of surface damage is shown in text-fig. 2. The exposed mesial surface of the jaw is fully illustrated (text-fig. 3c). The restorations (text-fig. 4a, b) of the internal and external views of the jaw have been modified from those published elsewhere (Smithson 1980a, text-fig. 7), in particular with respect to the course of the sutures on the posterior half of mesial surface and the size of the Meckelian fenestrae. The dentary accounts for more than half of the total length of the dorsal margin of the jaw, as preserved. Its convex lateral surface has broad overlapping sutures with the infradentary bones (presplenial, postsplenial, angular, and surangular) and it reaches its maximum depth above the postsplenial. Projecting from the mesial surface of the dentary, a short distance below its dorsal edge, is a horizontal tooth bearing shelf. It appears to maintain an almost constant width along its length although the posterior region is concealed by intractable matrix. The shelf is slightly broader than the bases of the teeth it supports. Further mesially it is excavated to form a shallow recess which is normally occupied by the ventrolateral portion of the coronoid series. This recess is visible in the anterior portion of the jaw where the coronoids have been displaced by lateral compression post- mortem. Details of the dentition are discussed separately. The coronoid series forms an almost horizontal roof to the Meckelian space but the sutures between individual coronoid bones could not be traced. The floor of the Meckelian space comprises two splenials and the angular. They wrap around the ventral edge of the jaw forming a distinct angle between the gently convex external surface and the almost vertical internal surface. The presplenial forms a considerable portion of the anteromesial surface of the jaw. It contacts the coronoid series dorsally and forms the anterior border of the anterior Meckelian fenestra posteriorly. Unfortunately, I cannot determine if it contacted the prearticular. In the jaw of primitive tetrapods, for example Iclitliyostega (Jarvik, 1980), Metaxygnathus (Campbell and Bell 1977), and Doragnathus (Smithson 19806), the prearticular has a broad contact with the presplenial. In the jaw of embolomeres. SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 609 TEXT-HG. 2. Proterogyrinus panchem sp. nov. Right mandible, trunk vertebra, and rib, RSM GY 1977.46.33, lateral view. Natural size. 610 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 3. Proterogyrinus pancheni sp. nov., right mandible, RSM GY 1977.46.33. a, lateral view. B, lateral (internal) view showing dorsal edge of adductor fossa, c, mesial view. Natural size. Broken bone hatched, eroded external surface of mandible light mechanical stipple, matrix heavy mechanical stipple. Abbreviations: ang, angular; art, articular; cor, coronoid; d, dentary; Meek, fen, Meckelian fenestra; pra, prearticular; ps.f, postsymphysial foramen; psp, postsplenial; sp, splenial. however, for example Pholiderpeton (Panchen 1970, 1972), Eobaphetes, and Anthracosaurus (Panchen 1977), the prearticular is prevented from contacting the presplenial by the anterior coronoid. Among the anthracosaurs this feature appears to be unique to embolomeres, although Agnew (1984) has described Pholiderpeton as having an arrangement similar to that found in primitive tetrapods with a short suture between the prearticular and presplenial. SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 611 Anteriorly the internal surface of the presplenial is gently concave and pierced by two postsymphysial foramina. The posterior end of the bone is strongly concave and in my earlier restoration of the mesial surface of the lower jaw of Proterogyrinus pancheni (Smithson 1980a) I interpreted the whole of this concave region as representing the anterior border of the anterior Meckelian fenestra. I now think this is incorrect. The dorsal rim of this concavity appears on closer examination to be a broken surface originally attached to bone pushed between the dentary and presplenial, rather than the border of a fenestra. As a result the restored height of the anterior Meckelian fenestra has been reduced. However, the ventral margin of the fenestra is preserved and its length in the two restorations remains unchanged. The postsplenial forms the posteroventral and anteroventral borders of the anterior and posterior Meckelian fenestrae respectively. The two fenestrae were presumably divided by a projection of the prearticular which contacted the postsplenial a short distance above the ventral margin of the jaw. This region is not preserved in RSM GY 1977.46.33 although the sutural surface on the postsplenial is visible (text-fig. 3c). The angular forms the posteroventral border of the posterior Meckelian fenestra. The bone has cracked along the ventral margin of the jaw and in my previous restoration 1 took this crack to represent the angular prearticular suture. This is incorrect and the true line of contact is more dorsally placed (text-fig. 4b). Fracturing of the bones overlying the adductor fossa has obliterated the angular-surangular suture. However, I have assumed that parts of both bones are present in RSM GY 1977.46.33, and that the suture between them extends across the external surface of the jaw approximately midway between its dorsal and ventral edges (text-fig. 4a). Behind the tooth row the dorsal margin of the surangular rises steeply to form a strongly convex surangular crest similar to that of Pholiderpeton (Panchen 1972, text-fig. 11; Agnew 1984). This region is obscured in most specimens of Proterogyrinus scheelei by the overlying bones of the cheek. The base of the crest immediately behind the tooth row is thickened and excavated to form a short shallow groove. The edge of the crest is champered and forms a sharp dorsal margin of the jaw. It was not possible to determine whether the bone was thickened below the dorsal margin as in Pholiderpeton (Panchen 1972, p. 306) or whether the surangular wrapped around the posterior edge of the jaw to contact the prearticular. The prearticular is very poorly preserved. It occupied a position on the mesial surface of the jaw between the infradentary series and the coronoid bones and formed a part of, or possibly all, the dorsal border of the Meckelian fenestrae. In addition it formed the concave mesial edge of the adductor fossa which remained attached to the matrix when the anterior portion of the jaw was removed (text-fig. 3b). The bone in this region is considerably thickened and probably acted as an area of insertion for the adductor muscles as in Eoherpeton (Smithson 1980a, 1985a). The articular is embraced on its lateral surface by the surangular and overlaps the prearticular mesially. In dorsal view the articular surface forms a strongly concave arc with the internal corner of the articulation projecting in front of its external corner. It is not saddle-shaped as in Pholiderpeton (Panchen 1972, p. 308) and presumably formed a simple hinge joint with the quadrate. The precondyloid process is incomplete but the postcondyloid process is well defined and forms the posterior edge of the jaw; no retroarticular process is developed and the prominent boss present on the articular of Proterogyrinus scheelei is absent. Dentition. With the exception of a single tooth, all the marginal teeth in the type specimen of P. pancheni are damaged. Judging from the incomplete portions preserved they appear to have been of a uniform size along most of the tooth row becoming smaller towards its posterior end. The bases of the teeth are oval in outline with their long axes oriented at right angles to the jaw margin. In lateral view the complete tooth is strongly hooked back (text-fig. 4d) and in anterior view gently incurved (text-fig. 4c). A clearly defined sharp ridge runs up the apical half of the anterior surface of the tooth and passes over the apex to run down on to the posterior surface. This type of tooth is unlike that of any other described anthracosaur and represents the most diagnostic feature of the type specimen. 612 PALAEONTOLOGY, VOLUME 29 A 30 mm I I L 3 mm TEXT-FIG. 4. Proterogyrinus pancheni sp. nov. a, b, right mandible restored. A, lateral view, b, mesial view. Approx, x |. c, d, marginal tooth, c, anterior view. D, lateral view, x 7. Abbreviations: see text-fig. 3. There are spaces for thirty-two teeth in the preserved portion of the dentary and labyrinthine infold- ing of the enamel is restricted to the region below the dorsal alveolar margin of the jaw. The coronoid bones are covered with a shagreen of very small denticles. A small patch of denticles is also present on the dorsal rim of the presplenial between the posterior postsymphisial foramen and the anterior margin of the anterior Meckelian fenestra, and on the prearticular immediately in front of the adductor fossa. The expanse of denticles on the coronoid series suggests a similarly extensive cover on the pterygoids of P. pancheni. In most amphibia the area of denticle covering on the coronoid bones is coincident with that present on the pterygoids for example Pholiderpeton (Panchen SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 613 1972, cf. text-figs. 7 and 12) and Greererpeton (Smithson 1982, cf. text-figs. 11 and 19). A similar situation is present in osteolepiform fishes, although here it is the prearticular rather than the coronoids which exhibit denticle covering, for example Eusthenopleroii ( Jarvik 1 980, cf. text-figs. 1 24 and 125). Somewhat unusually, the lower jaw referred to Anthracosanrus by Panchen (1977) exhibits denticles on the three coronoids while the palatal bones of the two skulls are bare. Remarks. One of the characteristics of the lower jaw of embolomerous anthracosaurs is the pair of large Meckelian fenestrae on the mesial surface of each ramus. A. russelli exhibits a slight variation of this pattern in that the two fenestrae have coalesced to form a single, large oval opening (Panchen 1981). Large Meckelian fenestrae are also present in two other groups of early tetrapods, colosteid temnospondyls (Smithson 1982) and the Diadectomorpha {sensu Heaton 1980). In the latter two groups the fenestra is similar to that found in Authracosaunis and represents a considerable opening on the mesial surface of the jaw. The arrangement of bones around the Meckelian fenestrae dilfer in the three groups. Embolomeres and colosteids retain both splenial bones while the diadectomorphs retain only one. The colosteids and diadectomorphs maintain the plesiomorphic connection between the prearticular and pre- splenial (cf. Doragnafhus, Smithson 1980/? and Iclitliyostega, Jarvik 1980) while in some embolomeres the anterior coronoid extends posteroventrally between the presplenial and prearticular precluding a connection (see above). These differences in the pattern of bones surrounding the enlarged Meckelian fenestrae in colosteids, diadectomorphs, and embolomeres suggests that the fenestrae developed convergently in the three groups. The function of the fenestrae has been discussed by Panchen (1972) for anthracosaurs and by Heaton (1980) for diadectomorphs. Both agree that they developed, in part, in association with the intermandibular musculature. This proposal was based partially on the assumption that the presence of a retroarticular process in ‘more advanced labyrinthodonts’ is a derived condition, and its absence in embolomeres is primitive (Panchen 1972, p. 309). However, recent descriptions of the mandibles of Metaxygnathus (Campbell and Bell 1977), Doragnathiis (Smithson 1980/7), and Ichthyostega (Jarvik 1980) have demonstrated that a retroarticular process is present in the earliest amphibia, and its absence in embolomeres need not necessarily be primitive. In addition, colosteids (Smithson 1982) and diadectomorphs (Heaton 1980) retain the retroarticular process suggesting that enlarged intermandibular muscles are not directly related to its absence. Heaton (1980) suggested that the large fenestrae in diadectomorphs were also correlated with enlarged intermandibular muscles. He suggested that the vacuity was a space which accommodated contracting muscles and proposed that its function was analogous with that of the external mandibular foramen of crocodiles. In fishes, for example Amia (Allis 1897), and many tetrapods (Romer and Parsons 1977) the intermandibular muscles form a thin sheet of tissue between the jaws. If these muscles were enlarged in diadectomorphs, as suggested by Heaton, contraction would almost certainly result in an increase in muscle volume between the jaws rather than in the immediate vicinity of their origins. It therefore seems unlikely that Meckelian fenestrae developed to accommodate contracting muscles. A further explanation which may account for the presence of enlarged Meckelian fenestrae in colosteids, diadectomorphs, and embolomeres is afforded by reference to the jaws of lepisosteid fishes (text-fig. 5). In Lepisosteus platyrhiuus (Nelson 1973) a large area of Meckel’s cartilage is exposed on the mesial surface of the jaw as a result of incomplete ossification of the surrounding dermal bones. In a related form Atractosteus spatula (Wiley 1976) Meckel’s cartilage is also exposed to a lesser degree. Removal of Meckel’s cartilage from the jaws of both fishes would leave fenestrae analogous to those in the jaws of embolomeres, diadectomorphs, and colosteids and it is possible that rather than having a specific function the enlarged Meckelian fenestrae in some early tetrapods represent persistant incomplete ossification of the mesial surface. Axial skeleton. Twelve vertebral elements and one incomplete rib from the Dora Bone Bed are referred to Proterogyrinus paneheni, together with an isolated pleurocentrum from Niddrie. Six pleurocentra and associated neural arches were recovered but successive vertebrae were never found 614 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 5. Mandibles of lepisosteid fishes and Palaeozoic tetrapods. a, Lepisosteus platyrhinus (after Nelson 1973). B, Atraclosteus spatula (after Wiley 1976). c, Proterogyrinus pancheni sp. nov. D, Anthracosaurus russelli (after Panchen 1977). e, Tseajaia campi (after Moss 1972). f, Greererpeton hurkemorani (after Smithson 1982). SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 615 in articulation. The vertebrae are very similar to those of P. scheelei. They are gastrocentrous with short intereentra and well-ossified pleurocentra. All parts of the vertebral column are represented by the material from Dora. Pleurocentrum. The description of the pleurocentrum is based on an isolated specimen from Cowdenbeath RSM GY 1977.46.34 (text-fig. 6f-h). In most respects it is very similar to the pleurocentra of the embolomeres Pholiderpeton (Panchen 1966; Agnew 1984) and Pteroplax (Boyd 1 980). It is subcircular in anterior view, deeply amphicoelous, and perforated by a small notochordal 30 mm TEXT-FIG. 6. Proterogyrinus panclieni sp. nov., vertebrae, a-e, trunk vertebra, RSM GY 1975.48.49. a, anterior view, b, posterior view, c, right lateral view, d, sagittal section through pleurocentrum. e, ventral view, f-h, pleurocentrum, RSM GY 1977.46.34. f, anterior view. G, posterior view, h, right lateral view, i, sacral/postsacral vertebra, RSM GY 1976. 19.48, left lateral view. Natural size. Broken bone hatched. Abbreviations: M, area of attachment of axial musculature; M.isp,interspinalis muscle; M. it, intertransversarii muscle; M.ob. cap. mag, obliquus capitis magnus muscle; M.rect.cap.p, rectus capitus posterior muscle; M.sp.c, spinalis capitus muscle; M.sp.d, spinalis dorsi muscle; M.ssp, semispinalis muscle; M.ssp.c, semispinalis capitus muscle; s-n.c, supra-neural canal; v.k, ventral keel. 616 PALAEONTOLOGY, VOLUME 29 foramen. A periosteal bone covering is restricted to the ventrolateral portion of the centrum which is excavated over most of its length to form a concave rim to the bone. The ventromedial region is not excavated and forms a pronounced ventral keel which is most clearly seen on RSM GY 1975.48.49 (text-fig. 6c). A pair of steeply inclined, concave facets, which articulated with the neural arch, form the dorsolateral corners of the anterior face of the centrum. They bound the gently convex floor of the neural canal which is fully ossified, as in Plwliderpeton and Pteroplax, and lacks the suture present in Proterogyrinus scheelei (Holmes 1984). The centrum is more deeply amphicoelous posteriorly and is pierced by the notochordal canal slightly above its centre. In RSM GY 1977.46.34 the diameter of the anterior surface is somewhat smaller than the posterior surface suggesting that in a fully articulated skeleton it occupied a position immediately in front of the larger trunk vertebrae. The centra with attached neural arches and the larger, isolated specimen from Niddrie differ from RSM GY 1977.46.34 in a number of respects. They are oval rather than circular in outline and most taper ventrally (text-fig. 6a, b). In addition the most complete vertebra RSM GY 1975.48.49 (text- fig. 6a-e) is relatively much longer than any other attributed to P. pancheni. The pleurocentrum of RSM GY 1976.19.48 (text-fig. 6i) is similar in most details to RSM GY 1 977.46.34 but directly below the neural arch facets it bears a pair of concave facets for the capitulum of the ribs. In P. scheelei the last presacral, the sacral, and the first three postsacral pleurocentra exhibit rib facets. It is reasonable to assume that RMS GY 1976.19.48 occupied a similar position in the skeleton of P. pancheni. Intercentrum. Two trunk intercentra and one haemal arch from Dora are referred to P. pancheni (text-fig. 7). They are smaller than the intercentra referred to Eoherpeton (Smithson 1985a), not so strongly horseshoe shaped and lack the concave facets which cup the anterior surface of the pedicel of the neural arch. The trunk intercentrum is crescent shaped in anterior view and when articulated with a pleurocentrum of similar diameter it extends dorsally to the ventral edge of the neural arch facets. In lateral view it narrows dorsally to form a bony wedge. A periosteal covering is restricted to the ventral surface of the intercentrum and a narrow groove on the lateral edge. A faeet for the eapitulum of the rib is not developed on the ossified portion of the bone. Projecting from the ventrolateral corner of RSM GY 1976.19.49 is a small tubercle, restricted to one face of the centrum where it forms a con- tinuation of the articulating surface (text-fig. 7a). In NUZ 76.10.17 this tubercle forms a well-defined ridge along the length of the ventral surface (text-fig. 7c) and presumably formed the lateral margin of a groove for the dorsal aorta. The haemal arch RSM GY 1983.9.2 (text-fig. 7e, f) is almost complete and lacks only the ventral tip 30 mm ■ ' ■ ■ TEXT-FIG. 7. Proterogyrinus pancheni sp. nov., vertebrae, a, b, intercentrum, RSM GY 1976.19.49. a, ?anterior view, b, lateral view, c, D, intercentrum, NUZ 76.10.17. c, ?anterior view. D, lateral view, e, f, haemal arch, RSM GY 1983.9.2. e, lateral view, f, anterior view. Natural size. Broken bone hatched. SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 617 and the thin anterior and posterior edges of the haemal spine. It has been crushed slightly but the overall form of the bone is readily determined. The dorsal articulating surface corresponds very closely with the trunk intercentrum RSM GY 1976.19.46 and fused to it is a chevron-shaped arch which surrounded the haemal canal. The canal is oval in anterior view, its long axis oriented vertically. The arch tapers ventrally to form a narrow spine which is inclined posteriorly at an angle of 35° to the vertical. This is similar to that of the anterior haemal arches of P. scheelei (Holmes 1984, text-fig. 22). Cervical neural arch. A neural arch NUZ 77.2. 17 (text-fig. 8a-c) is considered to be from the cervical region of P.pancheni on the basis of scars which are here interpreted as areas of origin and insertion of occipital muscles. It differs markedly from the atlas and axis neural arches of P. scheelei (Holmes 1984, text-fig. 21) and it therefore probably formed the first undifferentiated cervical vertebra, the most posterior of the three vertebrae from which occipital muscles originated (see Olson 1936). Unfortunately the third and fourth neural arches of P. scheelei have not been described and it is not possible to confirm the interpretation of NUZ 77.2.17 by direct comparison. The neural arch is reasonably well preserved but the right pedicel and prezygapophysis are missing. It is 21 mm long between the zygapophyses and the neural spine has a maximum height of 18 mm. This compares with an average length of 23 mm and height of 37 mm in the trunk vertebrae attributed to P. pancheni. The angle of inclination of the prezygapophysis is 25° which compares with 25° in the trunk vertebrae and c. 50° in the caudal vertebra attributed to P. pancheni. The pedicel accounts for less than half the total length of the neural arch. Its internal and anterior surfaces form a continuous surface of unfinished bone which articulated with the pleurocentrum M.sp.c 30 mm j I I TEXT-FIG. 8. Proterogyrirms pancheni sp. nov., vertebrae. A-c, cervical neural arch, NUZ 77.2.17. A, anterior view, b, posterior view, c, left lateral view, o-f, caudal neural arch, NUZ 78.3.33. d, anterior view, e, posterior view, f, right lateral view. Natural size. Broken bone hatched. Abbreviations: see text-fig. 6. 618 PALAEONTOLOGY, VOLUME 29 posteromesially and the intercentrum anteriorly. The ventrolateral edge of the pedicel is also unfinished and forms the articulation for the tubercular rib head. In lateral view this articulating surface is oriented at an angle of 38° to the horizontal. The neural spine is straight sided in lateral view, gently inclined posteriorly, and with a forward sloping dorsal edge. Between the zygapophyses the base of the spine is pierced by a supraneural canal. The muscle scars are clearly differentiated and their arrangement differs markedly from that found on the posterior trunk vertebrae (see below). The interpretation of the scars follows Olson’s (1936) analysis of the dorsal axial musculature in early tetrapods. The posterior region of the lateral surface of the neural spine, above the level of the supraneural canal, is excavated to form a shallow rugose recess whieh is separated from the more anterior portion of the spine by a strongly defined ridge. The morphology of this region is very similar to the posterodorsal corner of the axis neural spine in Eryops illustrated by Olson (1936, text-fig. 8g). It represents the area of insertion of M. spinalis cervicus and M. semispinalis cervicus. A similar, but less extensive, recessed area is present on the anterior region of the left lateral surfaee of NUZ 77.2.17. In Eryops this represents the area of origin of M. rectus capitis posterior. Between the recessed areas on the anterior and posterior edges of the spine the median region is heavily striated. It is not developed into the characteristic median ridge for M. semispinalis present on the trunk vertebrae. This region in Eryops represents the area of insertion of M. obliquus capitis magnus. The occipital muscles M. rectus capitis posterior and M. obliquus capitis magnus are restricted to the first three cervical vertebrae (Olson 1936, p. 299) where they form with M. obliquus capitis superior and M. obliquus capitis inferior a series of short muscles passing forward to insert on the occipital region of the skull. Trunk neural arch. Eight trunk neural arches from Dora are attributed to P. pancheni. The most complete specimen RSM GY 1975.48.49 (text-fig. 6a-c) forms the basis of the description. It closely resembles the trunk neural arches of P. scheelei (Holmes 1984, text-fig. 22), with neural spines considerably taller than those of Pholiderpeton and Pteroplax, and clearly defined areas of attachment of the median dorsal axial musculature. The pedicels account for only half the total length of the neural arch. They bear two pairs of articulating surfaces: a large pair on the posterolateral surface which articulate with the pleurocentrum and a smaller pair on the anterior surface, directly beneath the prezygapophyses, which presumably articulated with a cartilaginous extension of the intercentrum. In RSM GY 1975.48.49 the transverse processes are very short and only the posterior portion projects laterally beyond the neural arch facets. Its dorsal edge is steeply inclined ventrolaterally and its lateral margin lies well below the dorsal edge of the prezygapophyses. The articulation for the tuberculum of the rib is a strap of unfinished bone oriented anteroventrally, at an angle of approximately 45° to the horizontal, and also slightly anteromesially. The posterior surface of the transverse process/pedicel is excavated to form a well-defined recess which probably marks the origin of M. intertransversarii. In the less well-preserved specimen RSM GY 1977.46.35 the transverse process is more pronounced and forms a distinct lateral projection from the body of the pedicel. The facet for the tuberculum is less steeply inclined anteroventrally than that of RSM GY 1975.48.49 and appears to project wholely laterally. Extending posterodorsally from the posterior edge of the pedicel is a strong ridge which buttresses the postzygapophysis. A similar ridge extends anteroventrally from the base of the neural spine to buttress the prezygapophysis. Between them a deep fossa probably marks the origin of M. spinalis dorsi. A distinct origin of M. spinalis dorsi has not been described on the vertebrae of Coal Measure embolomeres but it is present in Proterogyrinus scheelei and is a characteristic of the vertebrae of Diadectes, pelycosaurs, and other amniotes (Olson 1936). The prezygaphyses are closely spaced, project slightly in front of the facets for the intercentrum, and are inclined ventromesially and slightly posteroventrally. The postzygapophyses are correspond- ingly inclined dorsolaterally and very slightly anterodorsally, allowing a certain degree of rotation and dorso-ventral bending. Between the zygapophyses the base of the neural spine is pierced by a SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 619 small supraneural canal for the dorsal ligament. The spine is approximately one and a half times the length of the neural arch with a gently convex anterior margin and concave posterior margin. The lateral surfaces are roughly divided into two equal halves by a median ridge. On the right side of the spine of RSM GY 1975.45.49 the posterior half is the largest while on the left the anterior half is the largest. The median ridge marks the origin and insertion of M. semispinalis and the areas on either side of it mark the origin (anterior) and insertion (posterior) of M. interspinalis (Olson 1936). Such a clear differentiation of areas of muscle attachment have not been observed on the vertebrae of Coal Measure embolomeres but they are well defined on the vertebrae of P. scheelei (Holmes 1984, text-fig. 22). Specimen RSM GY 1977.46.33, preserved on the same block as the lower jaw of P. pancheni, is very similar to RSM GY 1975.48.49. The neural arch of the sacral or immediately post-sacral vertebrae RSM GY 1976.19.48 is also similar but the median ridge on the neural spine and the fossa for M. spinalis dorsi are less well defined. In RSM GY 1983.9.1 the supraneural canal is slightly more extensive dorsally and the base of the spine somewhat narrower. Caudal neural arch. A neural arch NUZ 78.3.33 (text-fig. 8d-f) is thought to be from the tail of P. pancheni on the basis of steeply inclined zygapophyses and the absence of rib facets. Although the specimen has been crushed and the neural spine is missing, it is clearly different from the trunk neural arch RSM GY 1975.48.49. The angle of inclination of the postzygapophyses, c. 50°, is more than twice that measured from the trunk neural arches. In P. scheelei there appears to be a progressive increase in the angle of inclination along the tail. The zygapophyses of caudal vertebrae 5 and 1 1 are inclined at 26° and 45° respectively (Holmes 1984, p. 477). If a similar trend occurred in P. pancheni NUZ 78.3.33 would have occupied a position between caudal vertebra 13 and 20. The pedicels lack the normal transverse processes with articulating surfaces for the ribs, but projecting anteroventrally from their bases are a pair of flanges. Unfortunately they are incomplete on both sides, that of the left being most fully preserved. They are very thin with delicate lateral processes and completely covered in periosteal bone. Their form is mostly clearly appreciated from text-fig. 8d-f. Ribs. The incomplete rib preserved with the lower jaw of P. pancheni is the only specimen from Dora referred to this species with confidence. It is 43 mm long and considerably smaller than the ribs from Dora attributed to Eoherpeton (Smithson 1985r/). The tubercular rib head is preserved but the capitulum is missing. The estimated span of the rib head is approximately half that of the ribs of Eoherpeton. Appendicular skeleton. Clavicle. An incomplete left clavicle NUZ 77.2.19 (text-fig. 9a-c) was recovered from the grid square immediately adjacent to the skull table NUZ 75.1 1.2. The external surface of the ventral plate of the clavicle is ornamented with the undercutting pits and grooves which characterizes anthracosaur dermal skull roofing bones. It diflFers from all other clavicles recovered from the Dora Bone Bed which exhibit the characteristic pit and ridge ornament of temnospondyls. The clavicle of P. pancheni resembles closely those of Archeria (Romer 1957) and P. scheelei (Holmes 1980, 1984). It is divided into two regions; a ventral plate which overlaps the interclavicle mesially, and a dorsally directed shaft which overlaps the cleithrum dorsomesially and the scapulocoracoid ventromesially. The ventral plate is incomplete and its outline could not be determined. The shaft is almost entire but has been slightly crushed. The ventral plate is relatively thick anteriorly and tapers to a thin lamina posteriorly. Its sharp anterior edge extends dorsally on to the shaft where it divides to form two ridges. One ridge forms the leading edge of the clavicular shaft while the other, more posterior ridge, extends across the dorsolateral region of the ventral plate and separates the ornamented external surface from a highly striated area which Holmes (1980) suggested may mark an area of origin of the ventral throat musculature. At the ventral end of this area is a small tubercle. Running down the internal surface of the clavicular shaft is a broad ridge which fades out anteromesially as it meets the ventral plate. 620 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 9. Proterogyrinus pancheni sp. nov. a-c, left clavicle, NUZ 77.2. 19. A, ventral view. B, anterior view, c, dorsal view, d, interclavicle, BM(NH) R9940, ventral view. Natural size. Broken bone hatched. Behind the ridge a deep recess is scarred with pits which probably mark areas of ligamentous attachment. The recess was probably occupied by the anteroventral edge of the scapulocoracoid. Interclavkie. An interclavicle, BM(NH) R9940 (text-fig. 9d) was recovered from the grid square adjacent that which yielded the lower jaw of P. pancheni and associated vertebra and rib. It is preserved in ventral (external) view on a small block of bone bed. The incomplete posterior portion of the interclavicle has been displaced anterolaterally and slightly overridden the ventral surface. Its anterior edge was damaged during preparation and the left lateral corner of the area of clavicular overlap is missing. As preserved the interclavicle is 53 0 mm long and has a maximum width of 72-5 mm; the estimated maximum width of the intact specimen was c. 88 0 mm. BM(NH) R9940 is similar to the interclavicles of P. scheelei (Holmes 1980, text-fig. 3) and that attributed by Panchen (1972, text-fig. 13) to Pholiderpeton but which Agnew (1984) suggests might pertain to Pteroplax. The anterior edge is gently convex and lacks the anterior process of Archeria ( Romer 1957, text-fig. 1 ). The concave lateral edges taper posteriorly and possibly continued to form a modest parasternal process. The areas of clavicular overlap are well developed on the dorsolateral margins of the interclavicle. They are separated from the ornamented median portion by a shallow step which accommodates the thickness of the clavicular plate and allows the external surfaces of the clavicles and interclavicle to form a plane surface when articulated. Anteriorly the step is excavated to form a deep groove which was probably occupied by a ridge on the anterior edge of the clavicle. In SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 621 BM(NH) R9940 the mesial edge of the overlap area is sinuous but is oriented principally anteromesially. Anteriorly the two sides are separated by a thickened rugosity 10 mm wide. The median portion of the interclavicle is ornamented with small shallow pits which slightly undercut the bone surface. Towards the margins of the exposed surface the pits are drawn out into shallow furrows and the anterior end of the interclavicle is finely pitted. The areas of clavicular overlap are lightly striated. Humerus. An incomplete right humerus, NUZ 77.2.20 (text-fig. 10a, b) was found at Dora in the same general area as the skull table NUZ 75.11 .2. It was damaged during collection and unfortunately the TEXT-FIG. 10. Proterogyrinus pancheni sp. nov. a, b, right humerus, NUZ 77.2.20. A, dorsal view. B, ventral view. Natural size. Broken bone hatched. c-F, P. scheelei right humerus (after Holmes 1980, text-fig. 6, reversed). Position occupied by preserved portion of the P. pancheni humerus represented by mechanical stipple, c, dorsal view in plane of distal dorsal surface, d, dorsal view in plane of proximal dorsal surface, e, anterior view, f, ventral view in plane of distal ventral surface. Not to scale. Abbreviation: dp.c, deltopectoral crest. 622 PALAEONTOLOGY, VOLUME 29 broken pieces were not recovered. The entepicondyle and ectepicondyle are missing and the proximal end of the bone is incomplete. However, despite this, the specimen is readily identified as a humerus and the preserved portion resembles that of Proterogyrinus scheelei (Holmes 1980, text-fig. 6). The most diagnostic feature of NUZ 77.2.20 is a short deltopectoral crest (text-fig. 10b). Its gently convex unfinished surface faces anteroventrally and lacks the deep pit present on the humerus of Eoherpeton (SmWhson 1 9856t, text-fig. 25). The crest continues distally as a narrow, sinuous, unfinished edge which forms the anterior margin of the humerus. This region is very similar to that in P. scheelei and lacks the small tubercle described as a supinator process on the humerus of Eoherpeton. No other diagnostic feature of the anthracosaur humerus is preserved on NUZ 77.2.20. The position the preserved portion occupied in the intact bone is outlined on the illustration of the humerus of P. scheelei (text-fig. 10c-f). Pennir. In 1914 Watson described a right femur, BM(NH) R4085, which originally formed part of a collection of Scottish Lower Carboniferous lungfish made by Dr R. H. Traquair and later acquired by the British Museum of Natural History. No data accompanied the specimen, but on the basis of a small piece of matrix Watson (1914) attributed it to the Loanhead No 2 Ironstone. Unfortunately this cannot be verified. Until 1954 the entire Carboniferous Limestone Series in Scotland was included in the Lower Carboniferous (Currie 1954) and thus the pre-Coal Measure localities, for example, Burghlee, Burdiehouse, Gilmerton, Loanhead, and Niddrie, were regarded as Lower Carboniferous. The matrix associated with the femur is insufficient to allow a full appraisal of its lithology, and comparison with that from other ‘Lower Carboniferous’ localities neither cor- roborates nor refutes the view that the specimen was collected at Loanhead. However, as it may have been found in pre-Coal Measures rocks, possibly at Loanhead, and is well preserved and yields important data on the morphology of the femur of early tetrapods, it is here fully described for the first time. Watson erected the name Papposaurus traqiiairi for the femur and in his description noted its likeness to those of early reptiles. Subsequently White (1939) remarked on its similarity with the femur of Archeria and recently Holmes (1984) emphasized its close resemblance with that of Proterogyrinus scheelei. It is possible, therefore, that the specimen collected by Traquair is a femur of the anthracosaur here named P.pancheni. If it could be demonstrated that the femora of Papposaurus and Proterogyrinus scheelei were sufficiently similar to indicate the two taxa were members of the same genus Proterogyrinus would become a junior synonym of Papposaurus. However, the absence of features shared uniquely by the femora of Proterogyrinus scheelei and Papposaurus prevents such a demonstration. This paucity of valid diagnostic features and the uncertainty surrounding the provenance of the Papposaurus femur, suggests that the name P. traquairi should be considered a nomen vanum (Simpson 1945) and refer to the type specimen only. BM(NH) R4085 (text-fig. 1 1 ) is well preserved and the fully ossified proximal and distal condyles suggest it formed part of a mature adult. It is 69 mm long and less robust than the femora of Archeria (Romer 1957, text-fig. 8) and Proterogyrinus scheelei (Holmes 1984, text-fig. 33). The proximal end of the femur has a gently rounded parabolic outline in dorsal view. The articulating surface forms a dorsally convex strap of unfinished bone which extends along the anterior edge towards the internal trochanter. The medial portion of the articulation is deeper than the anterior and posterior margins and the whole surface is oriented ventrolaterally. Above the condyle the posterior edge of the dorsal surface bears a large, well-defined muscle scar, which probably marks the insertion of M. ischio- trochantericus. The adductor fossa is a deep oval recess bounded, in part, by the unfinished proximal articulating surface. It reaches its maximum depth behind the anterior edge of the condyle forming a prominent ridge which bears the internal and fourth trochanters. The internal trochanter lies mid way between the proximal edge of the femur and the fourth trochanter. It is less well defined than that on the femora of Archeria and forms a continuous unfinished surface with the proximal condyle. Below it the fourth trochanter forms a modest rugosity which passes over the anterior edge of the femur and is visible in all aspects except distal view. Extending distally from the trochanter is a prominent, but SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 623 C D TEXT-FIG. II. Papposaunis traquairi Watson, right femur, BM(NH) R4085. a, anterior view. B, ventral view, c, posterior view, d, dorsal view. E, distal view. F, proximal view. Natural size. Abbreviations: fib, fibial condyle; int.troch, internal trochanter; tib, tibial condyle; 4th troch, 4th trochanter; M.ischio, insertion of ischiotrochantericus muscle. slightly eroded, adductor crest. At its proximal end the crest is oriented posterodistally but below the adductor fossa it curves forward and runs down toward the intercondylar ridge. It occupies a more anterior position than that in Archeria, but unlike Ichthyostega (Jarvik 1980, text-fig. 162) and Eoherpeton (Smithson 1985r/, text-fig. 29) where the crest forms part of the anterior edge of the femur, it is not visible in dorsal view. The distal articulating surface is divided into anterior and posterior condyles by a deep intercondylar groove dorsally and a prominent intercondylar ridge ventrally, which together give the distal end of the femur a distinctly V-shaped outline in lateral (distal) view. The posterior condyle is the smaller of the two articulations and forms a gently convex, triangular surface which articulated with the fibula. In front of the condyle, on the ventral surface of the femur is a deep, subcircular depression, similar to that in Eoherpeton and more pronounced than that in Archeria. It lacks the rugosities for joint ligaments and flexor muscles present in Eoherpeton. Above the fibular condyle the dorsal surface is slightly scarred and probably marks the origin of M. extensordigitorum communis. There is no clear division on the articulating surface between the tibial and fibular condyles. As a proportion of the entire distal articulation the tibial condyle appears to be relatively smaller than that in Archeria or Eoherpeton suggesting a relatively smaller tibia in Papposaunis. The condyle lies, in part, beneath the intercondylar groove and is an anteriorly expanded strap which is wholely oriented ventromesially. In front of it, on the ventral surface of the femur, is a shallow recess, the popliteal space which is slightly less extensive than in Archeria but notably larger than in Eoherpeton. 624 PALAEONTOLOGY, VOLUME 29 DISCUSSION During the past eighteen months, three detailed reviews of anthracosaur systematics have been published (Holmes 1984; Panchen 1985; Smithson 1985a). Each followed a study of new material from sites in the Visean and Namurian of Scotland and North America, and, with the exception of Panchen’s review which included discussion of ideas presented by Smithson, each author wrote without detailed reference to his colleagues unpublished work. The three hypotheses of relationships proposed by Holmes, Panchen, and Smithson are illustrated in text-fig. 12. It is clear from the three cladograms that the authors accept the traditional view that the Carboniferous anthracosaur taxa, the Eoherpetontidae, Gephyrostegidae, and Embolomeri, are closely related, although differ as to how they are interrelated, but that they disagree on the closeness of relationships of the Palaeostegalia (Crassigyrimis) and the Seymouriamorpha to Carboniferous anthracosaurs. Holmes’s scheme (text-fig. 12a) is an extension of that proposed by Panchen in 1980 but which he has now rejected (Panchen 1985). The additional characters Holmes has used to support his scheme are nevertheless important and have a direct bearing on the validity of the hypotheses proposed by Panchen and Smithson. The principal difference between this scheme and those of Panchen and Smithson is the position of the Seymouriamorpha. A sister-group relationship between the Gephyrostegoidea (Gephyrostegidae and Eoherpetontidae) and Seymouriamorpha was proposed first by Panchen (1980). It was based on the view that Eoherpeton retained posttemporal fossae and had an incipient otic tube (Panchen 1975) and that among anthracosaurs these features were shared uniquely with the seymouriamorphs. However, on the basis of new material from Dora, I have shown (Smithson 1985a) that the braincase of Eoherpeton closely resembles that of embolomeres and lacks both the otic tube and posttemporal fossae. There appear to be no other derived characters which support a sister-group relationship between Eoherpeton and the seymouriamorphs and this part of Holmes’s scheme must be rejected. Panchen (text-fig. 12b) and I (text-fig. 12c) both agree that the Eoherpetontidae, Gephyrostegidae, and Embolomeri form a monophyletic group, which I named the Anthracosauroideae (Smithson 1985a). In addition I accepted the traditional view that the Anthracosauroideae and the Seymouria- morpha are sister-groups and together comprise the Anthracosauria. This relationship was based on the view that the tabular-parietal suture on a skull roof including all three bones of the temporal series, the tabular, supratemporal, and intertemporal, was a feature uniquely shared by the two groups (Smithson 1 985a). (In a recent review of the systematic position of intasuchid temnospondyls, Gubin (1984) corrected Konzhukova’s (1956) interpretation of the pattern of their skull roofing bones (see Heaton 1980). Konzhukova had shown a tabular-parietal suture like that of anthracosaurs in both species of intasuchid Intasuchus silvicola and Syndyodosuchus tetricus. However, this was in error. Intasuchids exhibit the supratemporal-postparietal suture present in all temnospondyls (Gubin 1984, text-figs. 1 and 2).) Holmes (1984) has suggested that the absence of postemporal fossae and the presence of tabular horns may also be synapomorphies uniting taxa traditionally placed in the Anthracosauria, but this is disputed by Panchen (see below). Prior to Holmes’s suggestion, I had assumed that the absence of posttemporal fossae was a synapomorphy of the Anthracosauroideae (Smithson 1985a) but I now accept Holmes’s (1984) interpretation of the occiput of the seymouriamorphs Karpinskiosaurus and Seynwuria, and agree with him that posttemporal fossae are probably absent in all seymouriamorphs. Panchen’s scheme (text-fig. 12b) breaks completely with tradition. He rejects the idea that the Anthracosauria is a monophyletic group, and suggests that ‘the tabular-parietal contact uniting anthracosauroids and seymouriamorphs must be a case of parallelism or homplasy’ (Panchen 1985, p. 555). Instead, he proposes that the anthracosauroids are most closely related to Crassigyrimis and supports this relationship with four characters he considers to be uniquely shared by the two groups: a, the dermal ornament on the skull roof and dermal pectoral girdle; h, tabular horns; c, the lack of posttemporal fossae; d, the histology of the teeth. Panchen contrasts the dermal ornament of loxommatids and temnospondyls which he defines as ‘a SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 625 A B C TEXT-FIG. 12. The phylogeny of the Anthracosauria. A, after Holmes (1984). b, after Panchen (1985). c, after Smithson ( 1985«). 626 PALAEONTOLOGY, VOLUME 29 raised honeycomb pattern which becomes extended in zones of intensive growth’ with that of anthracosauroids and Crassigyrims, which he defines as ‘less regular than the temnospondyl type and . . . consists of rounded pits more widely separated by less sharp ridges’ (Panchen 1985, p. 551). However, the polarity of this character has not been established and there is no reason to assume that the ornament of anthracosauroids and Crassigyrinus is derived with respect to that of loxommatids and temnospondyls. The ornament of Ichthyostega is different again being largely pustular, while that of sarcopterygian fishes is extremely variable but none has dermal ornament like that of early tetrapods. Thus, this character must be regarded as equivocal until polarity can be established and should not be used to support a relationship between anthracosauroids and Crassigyrinus. Modified tabular horns have been described by Holmes (1984) in seymouriamorphs and he has also reported the absence of posttemporal fossae in this group. These two features are therefore not unique to anthracosauroids and Crassigyrinus but may define a group which also includes the seymouriamorphs (see below). Crassigyrinus shares a unique tooth histology with two Coal Measure embolomeres Anthraco- saurus and Pholiderpeton (Panchen 1985). In primitive tetrapods, for example Ichthyostega, the histology of the palatal tusks sectioned at or near the junction of the root and the crown, shows a characteristic zigzag folding of the dentine into the open pulp cavity, with most of the angles bearing short side branches. In temnospondyls, there is a progressive reduction in the number of side branches and the most derived condition is found in Mastodonsaurus where side branches are absent. This derived condition is also found in Crassigyrinus, Anthracosaurus, Pholiderpeton, and probably Archeria (Panchen 1985). In addition, the tusks of Crassigyrinus, Anthracosaurus, and Pholiderpeton show a second feature which has not been described in the teeth of other early tetrapods. Panchen (1985, p. 529) notes that ‘between each successive pair of radial infoldings of primary dentine . . . there is a wedge-shaped zone of “dark dentine” with densely-packed tubules oriented radially’. This condition is unique to these three taxa and may, as Panchen suggests, be a shared derived character uniting Crassigyrinus and anthracosauroids. However, as noted by Panchen, its distribution among other anthracosauroids, particularly the smaller forms (Crassigyrinus, Anthracosaurus, and Pholi- derpeton are among the largest known Carboniferous tetrapods, with skull lengths greater than 30 cm) is not known. Consequently this synapomorphy must also be regarded as equivocal until its distribution within the Anthracosauroideae is established. Thus in summary, it is clear that both the traditional view of the Anthracosauria, as a group embracing the Carboniferous taxa, Eoherpetontidae, Gephyrostegidae, and Embolomeri, and the Seymouriamorpha, and the new scheme proposed by Panchen, are supported by very few synapomorphies. The traditional view is supported by a single character which Panchen suggests may have developed independently in anthracosauroids and seymouriomorphs, while Panchen’s new scheme is supported by two characters which are regarded here as equivocal either because the polarity of the character state has not been established or the extent of the character within the group is not known. Additional work is now necessary to demonstrate the validity of these characters. This should be done within the framework of a larger study of the interrelationships of early tetrapods. Currently, there is some measure of agreement between Holmes, Panchen, and Smithson that Crassigyrinus, anthracosauroids, and seymouriamorphs form part of a ‘reptiliomorph’ clade which also includes loxommatids, diadectomorphs, and amniotes. However, the intrinsic relationships of the members of this clade are disputed (see Holmes 1984; Smithson 1985a). It is hoped that new loxommatid material, including postcrania, from the Coal Measures of Lancashire being prepared by Dr Angela Milner at the British Museum (Natural History), and new anthracosauroid material from the Lower Carboniferous of West Lothian, Scotland (Wood et ah 1985), will provide new data on which to reappraise these relationships and test the competing hypotheses discussed here. Until then, we must accept that for Palaeozoic tetrapods, as for most other groups of fossil vertebrates, a generally agreed phylogeny and classification remains to be proposed. Acknowledgements. The descriptive section of this paper formed part of a thesis submitted for the Degree of Ph.D. in the University of Newcastle upon Tyne. The work began while I was a Junior Research Associate SMITHSON: NEW CARBONIFEROUS ANTHRACOSAUR 627 financed by a Natural Environment Research Council grant (Number GR3/2983) awarded to Dr A. L. Panchen, my research supervisor, to whom I express my special thanks for suggesting the topic and for his help and encouragement throughout the project. I am grateful to Dr A. C. Milner at the British Museum (Natural History) and Dr S. M. Andrews and Dr R. L. Paton at the Royal Scottish Museum for permission to examine and borrow specimens in their care and to Mr S. P. Wood for helping to establish the grid position of specimens from the Dora site. Dr R. Holmes kindly gave me free access to his material of Proterogyrinus scheelei during my stay at the Redpath Museum in 1980 and I thank him together with Drs J. A. Clack, A. R. Milner, and A. L. Panchen for helpful discussion of the problem of anthracosaur systematics. Dr J. A. Clack kindly read and commented on the manuscript which was typed by Mrs E. B. Kinghorn. REEERENCES AGNEW, J. A. 1984. Pholiderpetou scutigerum Huxley, an amphibian from the Yorkshire Coal Measures. Ph.D. thesis, University of Newcastle upon Tyne. ALLIS, E. p. 1897. The cranial muscles and cranial and first spinal nerves in Amia calva. J. Morph. 12, 487-809. ANDREWS, s. M., BROWNE, M. A. E., PANCHEN, A. L. and WOOD, s. p. 1977. Discovery of amphibians in the Namurian (Upper Carboniferous) of Eife. Nature, Land. 265, 529-532. BEAUMONT, E. H. 1977. Cranial morphology of the Loxommatidae (Amphibia: Labyrinthodontia). Phil. Trans. R. Soc. Land. (B), 280, 29-101. BOYD, M. J. 1980. The axial skeleton of the Carboniferous amphibian Pteroplax cornutus. Palaeontology 23, 273-285. and TURNER, s. 1980. Catalogue of the Carboniferous amphibians in the Hancock Museum, Newcastle upon Tyne. Trans. Nat. Hist. Soc. Nor thumb. 46, 5-24. CAMPBELL, K. s. w. and BELL, M. w. 1977. A primitive amphibian from the Devonian of New South Wales. Alcheringa, 1, 369-381. CLACK, J. A. 1983. The stapes of the Coal Measures embolomere Pholiderpetou scutigerum Huxley (Amphibia: Anthracosauria) and otic evolution in early tetrapods. Zool. J. Linn. Soc. 79, 121-148. CURRIE, E. 1954. Scottish Carboniferous goniatites. Trans. R. Soc. Edinh. 62, 527-602. GUBIN, YU. M. 1984. The systematic position of the intasuchids. Paleont. J. 18, 115-1 18. HEATON, M. J. 1979. Cranial morphology of primitive captorhinid reptiles from the late Pennsylvanian and early Permian, Oklahoma and Texas. Bull. Oklahoma Geol. Surv. 127, 1 -84. 1980. The Cotylosauria: a reconsideration of a group of archaic tetrapods. In panchen, a. l. (ed.). The terrestrial environment and the origin of land vertebrates, 497-551. Academic Press, London. HOLMES, R. 1980. Proterogvrinus scheelei and the early evolution of the labyrinthodont pectoral limb. Ibid. 351-376. 1984. The Carboniferous amphibian Proterogvrinus scheelei Romer, and the early evolution of tetrapods. Phil. Trans. R. Soc. Lond. (B), 306, 43 1 -527. HOTTON, N. 1970. Mauchchunkia bassa gen. et sp. nov. an anthracosaur (Amphibia: Labyrinthodontia) from the Upper Mississippian. Kirtlandia, 12, 1-38. JARVIK, e. 1980. Basic structure and evolution of vertebrates (2 vols). Academic Press, London. KONZHUKOVA, YE. D. 1956. The early Permian Inta fauna of the Northern Ural forelands. Trudy. Paleont. Inst. 62, 5-55. MOSS, J. L. 1972. The morphology and phylogenetic relationships of the Lower Permian tetrapod Tseajaia campi Vaughn (Amphibia: Seymouriamorpha). Univ. Calif. Publ. Geol. Sci. 98, 1-63. NELSON, G. J. 1973. Relationships of clupeomorphs, with remarks on the structure of the lower jaw in fishes. In GREENWOOD, p. H., MILES, R. s. and PATTERSON, c. (cds.). Interrelationships of fishes, 333-349. Academic Press, London. OLSON, E. c. 1936. The dorsal axial musculature of certain primitive Permian tetrapods. J. Morph. 59, 265-31 1 . PANCHEN, A. L. 1 964. The Cranial morphology of two Coal Measure anthracosaurs. Phil. Trans. R. Soc. Lond. (B), 247, 593-637. 1966. The axial skeleton of the labyrinthodont Eogyrinus attheyi. J. Zool. 150, 199-222. 1970. Teil 5a, Anthracosauria. Handbuch der Paldoherpetologie. Lischer, Stuttgart. 1972. The skull and skeleton of Eogyrinus attheyi Watson (Amphibia: Labyrinthodontia). Phil. Trans. R. Soc. Lond. (B), 263, 279-326. 1975. A new genus and species of anthracosaur amphibian from the Lower Carboniferous of Scotland and the status of Pholidogaster pisciformis Huxley. Ibid. 269, 581-640. 628 PALAEONTOLOGY, VOLUME 29 PANCHEN, A. L. 1977- On Aiitliracosaurus russelli Huxley (Amphibia; Labyrinthodontia) and the family Anthra- cosauridae. Ibid. 279, 447-512. 1980. The origin and relationships of anthracosaur Amphibia from the late Palaeozoic. In panchen, a. l. (ed.). The terrestrial environment and the origin of land vertebrates, 319-350. Academic Press, London. — 1981. A jaw ramus of the Coal Measure amphibian Anthracosaiirus from Northumberland. Palaeontology, 24, 85-92. 1985. On the amphibian Crassigyriniis scoticus Watson from the Carboniferous of Scotland. Phil. Trans. R. 5oc. Lorn/. (B), 309, 505-568. ROMER, A. s. 1957. The appendicular skeleton of the Permian embolomerous amphibian Archeria. Contr. Miis. Geol. Univ. Mich. 13, 1()3-159. 1963. The larger embolomerous amphibians of the American Carboniferous. Bull. Mas. comp. Zool. Harv. 128,415-454. 1970. A new anthracosaurian labyrinthodont Proterogyrinus scheelei from the Lower Carboniferous. Kirtlandia, 10, 1-16. and PARSONS, t. s. 1977. The vertebrate body (5th edn.). Saunders, Philadelphia. SIMPSON, G. G. 1945. The principles of classification and the classification of mammals. Bull. Am. Mus. nat. Hist. 85, 1-350. SMITHSON, T. R. 1980«. An early tetrapod fauna from the Namurian of Scotland. In panchen, a. l. (ed.). The terrestrial environment and the origin of land vertebrates, 407-438. Academic Press, London. 19806. A new labyrinthodont amphibian from the Carboniferous of Scotland. Palaeontology, 23, 91 5-923. 1982. The cranial morphology of Greererpeton burkemorani Romer (Amphibia: Temnospondyli). Zook J. Linn. Soc. 76, 29-90. 1985fl. The morphology and relationships of the Carboniferous amphibian Eoherpeton watsoni Panchen. Ibid. 85, 317-410. 19856. Scottish Carboniferous amphibian localities. Scott. J. Geol. 21, 123-142. WATSON, D. s. M. 1914. On a femur of reptilian type from the Lower Carboniferous of Scotland. Geol. Mag. (6) 1, 347-348. WHITE, T. E. 1939. Osteology of Seymouria baylorensis Broili. Bull. Mus. comp. Zook Harv. 85, 325-409. WILEY, E. o. 1976. The phylogeny and biogeography of fossil and Recent gars (Actinopterygii; Lepisosteidae). Misc. Pubk Mus. nat. Hist. Univ. Kans. 64, 1-111. WOOD, s. p., PANCHEN, A. L. and SMITHSON, T. R. 1 985. A terrestrial fauna from the Scottish Lower Carboniferous. Nature, Land. 314, 355-356. T. R. SMITHSON Department of Zoology The University Newcastle upon Tyne NEl 7RU Typescript received 31 May 1985 A NEW IDENTITY FOR THE SILURIAN ARTHROPOD N EC ROG A M M A R U S by PAUL A. SELDEN Abstract. Restudy of the enigmatic arthropod Necrogammams sahveyi Woodward, hitherto considered to be a crustacean or myriapod, reveals that it is the infracapitulum (fused labrum and palpal coxae) and palp of a large but unspecified pterygotid eurypterid. I N the quest for the oldest representative of a taxon the fossil record commonly presents a number of poorly preserved specimens of doubtful assignation. Such is the case with the myriapods (Almond 1985), but one contender, Necrogammarus sahveyi Woodward, can now be rejected with certainty. N. sahveyi was collected by Mr Humphrey Salwey in Church Hill quarry at Leintwardine, Herefordshire. The fauna there occurs in a channel-fill of Lower Leintwardine Beds (middle Ludlow) cut down into Middle Elton Beds (see Whitaker 1962), hence the age of lower Ludlow assigned to the specimen by its first describers, Huxley and Salter ( 1 859, pp. 25, 97, pi. 1 3, fig. 17). They considered the specimen to be a crustacean and not a Pterygotus, but since it had been ‘accidentally introduced into this plate’ (Ibid., p. 97, referring to pi. 13, fig. 17) they probably thought it was eurypterid in the first instance. Woodward (1870) redescribed the specimen as an amphipod crustacean and named and sketched it. Peach ( 1 899) did not see the specimen but refigured it and referred it to the Diplopoda because it appeared to show diplosegments and a uniramous limb. In the Treatise it was mentioned under Crustacea Peracarida incertae sedis (Hessler 1969), but was discussed in connection with myriapods again, and refigured, by Rolfe (1980). Both Rolfe (1980) and Almond (1985) preferred to regard Necrogammarus as a possible aquatic relative of the myriapods. No photograph of Necrogammarus has hitherto been published, but when Almond showed a colour slide of it at the Palaeontological Association’s annual meeting in Cambridge, 1984, I immediately recognized it as part of a pterygotid eurypterid. Investigation of the feeding mechanism of Erettopterus hilohus (Salter) (unpublished except for the biomechanics of the chela system: Selden 1984) had revealed that this species, and probably other pterygotids, possesses only three pairs of slender walking legs (limbs III, IV and V) and that limb II is a small palp. The palpal coxae and labrum are fused into a discrete plate; such a structure occurs in other chelicerates (e.g. Ricinulei: van der Hammen 1979) and is termed the infracapitulum. As in other Chelicerata the infracapitulum is intimately associated with the chelicerae in Erettopterus, but the arrangement is complex, not yet fully understood, and is not described herein. It is obvious, however, from a comparison of text-figs. 1 and 2 that Necrogammarus is the infracapitulum and attached palp (of one side only) of a large pterygotid eurypterid. Two pterygotids are known from Church Hill: P. areuatus Salter and E. marstoni Kjellesvig- Waering, as well as numerous pterygotid fragments (Kjellesvig-Waering 1961). It is not possible to assign Necrogammarus to either of these species since they are diagnosed solely on features of chelicerae, coxal gnathobase, and metastoma. It may belong to one of these species, or be a new form; in any case, it was evidently a large animal. Acknowledgements. I thank R. A. Fortey, British Museum (Natural History) and A. W. A. Rushton, British Geological Survey, for the loan of specimens in their care. (Palaeontology, Vol. 29, Part 3, 1986, pp. 629-631.] 630 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1. Infracapitulum and palp ol'pterygotid eurypterids. a, Necrogammams salweyi Woodward, British Museum (Natural History) In 43786, holotype; Church Hill quarry, Leintwardine, Hereford; Lower Leintwardine Beds, middle Ludlow, x I . b, Erettoplerus hilohus (Salter), British Geological Survey GSM 59651; Logan Water, Lesmahagow, Scotland; Priesthill Group, Llandovery/Wenlock, x 2. TEXT-FIG. 2. Explanatory drawings for text-fig. 1 . A, Necrogammanis salweyi Woodward. B, Erettopterus hilohus (Salter). For explanation of homologous terms see Selden (1981). SELDEN: SILURIAN ARTHROPOD 631 REFERENCES ALMOND, J. E. 1985. The Siluro-Devonian fossil record of the Myriapoda. Phil. Trcm.s. R. Soc. B309, HI -Til . HAMMEN, L. VAN DER. 1979. Comparative studies on Chelicerata I. The Cryptognomae (Ricinulei, Architarbi and Anactinotrichida). Zool. Verh. Leiden, 174, 1-62. HESSLER, R. R. 1969. Peracarida. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part R. Arthropoda 4(1), R360-R393. Geological Society of America and University of Kansas Press, Boulder, Colorado, and Lawrence, Kansas. HUXLEY, T. H. and SALTER, J. w. 1859. On the anatomy and affinities of the genus Pterygotus and description of new species of Pterygotus. Mem. geol. Surv. U.K., Monogr. I, 1-105, 16 pis. KJELLESViG-WAERiNG, E. N. 1961. The SiluHan Eurypterida of the Welsh Borderland. J. Paleont. 35, 789-835. PEACH, B. N. 1899. On some new myriapods from the Palaeozoic rocks of Scotland. Proc. R. phvs. Soc. Edinh. 14, 113-126. ROLFE, w. D. I. 1980. Early invertebrate terrestrial faunas. In panchen, a. l. (ed.). The terrestrial environment and the origin of land vertebrates, 117-157. Systematics Association Special Volume No. 15. Academic Press, London and New York. SELDEN, p. A. 1981. Functional morphology of the prosoma of Baltoeurypterus tetragonophthalmus (Fischer) (Chelicerata: Eurypterida). Trans. R. Soc. Edinb. 72, 9-48. 1984. Autecology of Silurian eurypterids. In bassett, m. g. and lawson, j. d. (eds.). Aiitecology of Silurian organisms. Spec. Pap. Palaeont. 32, 39-54. WHITAKER, J. H. MCD. 1962. The geology of the area around Leintwardine, Herefordshire. Q. Jl geol. Soc. Land. 118, 319-351. WOODWARD, H. 1870. On Necrogammarus salweyi (H. Woodward), an amphipodous crustacean from the Lower Ludlow of Leintwardine. Trans. Woolhope Nat. Eld Club, 1870, 271-272, 1 pi. Typescript received 13 September 1985 Revised typescript received 6 November 1985 PAUL A. SELDEN Department of Extra-Mural Studies University of Manchester Manchester M13 9PL ®I' 'Vsil .1 ■*«. l/»' . V ■ . 'I- •e, ,# r i, &■ j^' ./iS ,J 4 NOTES FOR AUTHORS The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform in style to those already published in this journal, and should be sent to Dr. Dianne Edwards, Department of Plant Science, University College, P.O. Box 78, Cardiff CFl IXL, who will supply detailed instructions for authors on request (these are published in Palaeontology 1985, 28, pp. 793-800). Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology. SPECIAL PAPERS IN PALAEONTOLOGY In addition to publishing Palaeontology the Association also publishes Special Papers In Palaeontology. Members may subscribe to this by writing to the Membership Treasurer: the subscription rate for 1986 is £33 (LI.S. S53) for Institutional Members, and £16-50 (U.S. $27) for Ordinary and Student Members. A single copy of each Special Paper is available to Ordinary and Student Members only, for their personal use, at a discount of 25“„ below the listed prices. Non-members may obtain copies, but at the listed prices, from Marston Book Services, P.O. Box 87, Oxford 0X4 1 LB, England. RECENT PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS Special Papers in Palaeontology Numbers 1 -19 are still in print and are available (post free) together with those listed below: 20. (for 1977): Fossil Priapulid Worms, hy s. conway morris. 155 pp., 99 text-figs., 30 plates. Price £16 (U.S. $24). 21. (for 1978): Devonian Ammonoids from the Appalachians and their bearing on International Zonation and Correlation, hy M. R. HOUSE. 70 pp., 12 text-figs., 10 plates. Price £12 (U.S. $18). 22. (for 1978, published 1979): Curation of Palaeontological Collections. A joint Colloquium of the Palaeontological Association and Geological Curators Group. Edited by m. g. bassett. 279 pp., 53 text-figs. Price £25 (U.S. $38). 23. (for 1979): The Devonian System. A Palaeontological Association International Symposium. Edited by m. r. house, c. t. SCRUTTON and m. o. bassett. 353 pp., 102 text-figs., 1 plate. Price £30 (U.S. $45). 24. (for 1980): Dinoflagellate Cysts and Acritarchs from the Eocene of Southern England, hy j. p. bujak, c. downie, g. l. EATON and G. L. WILLIAMS. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23). 25. (for 1980): Stereom Microstructure of the Echinoid Test, hr A. b. smith. 81 pp., 20 text-figs., 23 plates. Price £15 (U.S. $23). 26. (for 1981): The Fine Structure of Graptolite Periderm, by P. r. crowther. 1 19 pp., 37 text-figs., 20 plates. Price £25 (U.S. $38). 27. (for 1981): Late Devonian Acritarchs from the Carnarvon Basin, Western Australia, hy G, playford and R. s. uring. 78 pp., 10 text-figs., 19 plates. Price £15 (U.S. $23). 28. (for 1982): The Mammal Fauna of the Early Middle Pleistocene cavern infill site of Westbury-sub-Mendip, Somerset, hy M. j. bishop. 108 pp., 47 text-figs., 6 plates. Price £25 (U.S. $38). 29. (for 1982): Fossil Cichlid Fish of Africa, hy j. a. h. van couvering. 103 pp., 35 text-figs., 10 plates. Price £30 (U.S. $45). 30. (for 1983): Trilobites and other early Arthropods. Edited by d. e. g. briggs and p. d. lane. 276 pp., 64 text-figs., 38 plates. Price £40 (U.S. $60). 31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, M- w. j. KENNEDY. 160 pp.. 42 text-figs., 33 plates. Price £25 (U.S. $38). 32. (for 1984): Autecology of Silurian organisms. Edited by M. G. bassett and j. d. lawson. 295 pp.. 75 text-figs.. 13 plates. Price £40 (U.S. $60). 33. (for 1985): Evolutionary Case Histories from the Fossil Record. Edited by j. c. w. cope and p. w. skelton. 202 pp., 80 text- figs., 4 plates. Price £30 (U.S. $45). 34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by o. i. Nikiforova, T. L. modzalevskaya and m. g. bassett. 66 pp., 6 text-figs., \6 plates. Price £10 (U.S. $15). Field Guides to Fossils 1. (1983): Fossil Plants of the London Clay, hy m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12). Other Publications 1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30). 1985. Atlas of Invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the Palaeontological Association, xiii-l-241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95. © The Palaeontological Association, 1986 Palaeontology VOLUME 29 • PART 3 CONTENTS Tlic comnuinily slriicUirc orUic Middle Cambrian Pliyllopod Bed ( liurgess Shale) S. ( ONWA'*' MOKKIS 423 l\)riferan allinities of Meso/oie slromatoporoids R. A. WOOD and J. Ri:i TNi:R 469 Conlrasting lil'eslyles in LowerJurassieerinoids: a eomparison orbenlhieand pseudopelagie Isoerinida MK'llAlil, J. SIMMS 475 The si/e-l'ieqneney dislribulion in palaeoeeology: elTeels of laplK)nomie proeesses during fornialion ol' mollusean death assemblages in Texas bays 11. (■ u M M I NS, i;. N. no Wl- 1. 1, R. .1. s 1 A N 1 ON, .1 R. aiid SI A 1- 1- 495 A letrapod traekway from the Carboniferous ofnorlhern Chile ( . M. m i l, and ,i. rovd 519 Silurian enerinurid Irilobites from (jolland and Dalarna, Sweden l.ARS RAMSK()1,1) 527 The (irst Tertiary selerosponge from the Amerieas l.DWARl) t. Wll.SON 577 Palaeoeeology of Siltirian eyeloerinitid algae sii:vi:n c. movDi.r: and marki:s i;. .ioiinson 5S5 A new anthraeosaur amphibian from the Carboniferous ol' Seotland r. R. SMITHSON 603 A new identity for the Silurian arthropod Nccro^amniarn.s RAtll. A. si: TORN 629 l‘riiili’(l in (iival /iriUiiii cU the [Inivcrsily h'inliiif’ lloii.se. Oxford hy David Slaiiford. Primer lo die Dniver.sily I.S.SN 00.^1 Palaeontology VOLUME 29 • PART 4 DECEMBER 1986 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to promote research in palaeontology and its allied sciences. COUNCIL 1986-1987 President'. Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Vice-Presidents'. Dr. M. G. Bassett, Department of Geology, National Museum of Wales, Cardiff CFl 3NP Dr. D. E. G. Briggs, Department of Geology, University of Bristol, Bristol BS8 IRJ Treasurer. Dr. M. Romano, Department of Geology, University of Sheffield, Sheffield SI 3JD Membership Treasurer. Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Birmingham B4 7ET Institutional Membership Treasurer: Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 5HN Secretary: Dr. P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA Circular Reporter: Dr. D. J. Siveter, Department of Geology, University of Hull, Hull HU6 7RX Marketing Manager: Dr. V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 IRJ Public Relations Officer: Dr. M. J. Benton, Department of Geology, The Queen's University of Belfast, Belfast BT5 6FB Editors Dr. P. R. Crowther, City of Bristol Museum and Art Gallery, Bristol BS8 IRL Dr. D. Edwards, Department of Plant Science, University College, Cardiff CFl IXL Dr. L. B. Halstead. Department of Geology, University of Reading, Reading RG6 2AB Dr. T. J. Palmer, Department of Geology, University College of Wales, Aberystwyth SY23 2AX Dr. C. R. C. Paul, Department of Geology, University of Liverpool, Liverpool L69 3BX Dr. P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL Other Members Dr. H. A. Armstrong, Newcastle upon Tyne Professor B. M. Funnell, Norwich Dr. M. E. CoLLiNSON, London Dr. P. D. Taylor, London Overseas Representatives Australia: Professor B. D. Webby, Department of Geology, The University, Sydney, N.S.W., 2006 Canada: Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta Japan : Dr. I. Hayami. University Museum, University of Tokyo, Hongo 7-3-1, Bunkyo-Ku, Tokyo Ne'w Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt U.S.A.: Dr. R. J. Cuffey, Department of Geology, Pennsylvania State University, Pennsylvania 16802 Professor A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66045 Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403 South America: Dr. O. A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108. Venezuela MEMBERSHIP Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1986 are: Institutional membership Ordinary membership Student membership Retired membership £45 00(U.S. S68) £2100(U.S. $32) £11-50 (U.S. $18) £10-50 (U.S. $16) There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 5HN. Student members are persons receiving full-time instruction at educational institutions recognized by the Council. On first applying for membership, an application form should be obtained from the Membership Treasurer, Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Aston Triangle, Birmingham B4 7ET. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer. All members who join for 1986 will receive Palaeontology, Volume 29, Parts U4. All back numbers are still in print and may be ordered from Marston Book Services, P.O. Box 87, Oxford 0X4 ILB, England, at £21-50 (U.S. $33) per part (post free). Cover: The chitinozoan Ancyrochitina onniensis Jenkins 1967 from the Late Caradoc, Onnian of the Onny River, Shropshire. The specimen measures 130 pm in length. Dr. W. A. M. Jenkins provided the photomicrograph. \ / SEP 2 9 1992 A PHYLOGENETIC CLASSIFICATION OF THE GRAPTOLOIDS by RICHARD A. FORTEY and ROGER A. COOPER Abstract. Graptolite classification has traditionally been based upon grade groups reflecting general levels of evolutionary complexity. This has been acknowledged as unsatisfactory and the present classification, used widely in the Western world, is a hybrid between this system and what are claimed to be ‘natural’ groups. A phylogenetic classification, in which taxa are based upon common ancestry, produces a more objective classification with taxa that are diagnosable. We attempt a reclassification of the planktic graptolites (excluding retiolitids) using phylogenetic methods, and employ a cladistic representation of relationships to portray character distribution and decide taxonomic levels. The Graptoloidea are defined as a monophyletic group by retention of a nematophorous sicula in the adult. Hence the Anisograptidae are transferred from the Den- droidea to the Graptoloidea. The higher Graptoloidea fall into two natural suborders: the Virgellina suborder nov. and the Dichograptina Lapworth, which replace the four existing suborders. The Virgellina includes mostly scandent graptolites of monograptid, diplograptid, dicranograptid, nemagraptid, and phyllograptid groups. These are united by having a virgellar spine on the sicula. Recognition of superfamilies within the Virgellina is based upon proximal structure and development type. Seven groups at family level are recognized and it is only within these groups that thecal morphology — long used as the primary criterion for diplograptid genera — is likely to be useful for distinguishing genera. Within the Suborder Dichograptina are two super- families; the Glossograptacea is characterized by having isograptid symmetry and contains the families Isograptidae Harris, Pseudisograptidae Cooper and Ni, Corynoididae Bulman, and Glossograptidae Lapworth; the Dichograptacea contains three families, of which the Sigmagraptidae Cooper and Fortey is distin- guished by its slender and asymmetrical proximal end, the Sinograptidae by prothecal morphology, while the Dichograptidae is subdivided by the progressive delay and suppression of dichotomies and some generic synonymy is recommended. Although our scheme embraces all graptoloids, problems remain with determining the interrelationships between some higher order taxa, and with the pendent graptoloids. The Anisograptidae is retained as a paraphyletic group. Most modern graptolite workers would like to think that they classify their organisms phylogenetic- ally. Classifications are, however, inherited from earlier workers where different concepts may have prevailed. We cannot necessarily apply our phylogenetic criteria to the available type species of genera, for example, and existing taxa may reflect judgements about whether a group is ‘different enough’ to be worth recognition at family level, rather then basing its status upon its common ancestry. It is perhaps surprising that there has been relatively little recent discussion about the higher level classification of the graptolites. Most Western workers have followed the Treatise on Invertebrate Paleontology in its two editions (Bulman 1955, 1970), and such unity of usage is useful as a convention, although Bulman himself clearly had reservations about whether the high-level classification was correct. New family level taxa were introduced by Mu (1950, 1974), but the phylogenetic basis of these is not clear. This paper reviews some of the problems of high-level classification, and in particular the status and relationships of the Family Anisograptidae and subordinal taxa in the Graptoloidea. In our analyses we have used a cladistic method for presenting the distribution of characters within and between groups. This method has several advantages: it makes clear what characters are used to define groups which are considered to have descended from a common ancestor; it identifies likely polyphyletic groups; and it points to areas where it is profitable to seek new information. We should emphasize that a cladogram is not an evolutionary tree: in many cases it is possible to determine [Palaeontology, Vol. 29, Part 4, 1986, pp. 631-654.) 632 PALAEONTOLOGY, VOLUME 29 the actual succession of graptoloid species from stratigraphic criteria, which can also be used as part of the evidence for classification. Cladograms are, however, extremely useful for examining the logic of classification, for determining the taxonomic status of subgroups, and for identifying monophyletic groups. The principle of phylogenetic classification we follow is that taxa, be they orders, families, or genera, should as far as possible be monophyletic— that is, they include all the descendants of a single common ancestor. Nineteenth-century classification tended to be based upon a grade of organization of the rhabdosome that was ambiguously phylogenetic. The past fifty years has seen the breaking up of some of these grade groups. The Glossograptidae, for example, had been recognized as different from other scandent biserial graptoloids, and at present has subordinal status (Glossograptina) with two families and a few included genera. While this makes it clear that the glossograptids have little to do, phylogenetibally speaking, with the diplograptids, it contributes nothing to our understanding of the phylogenetic position of the glossograptids within the Grapto- loidea as a whole. As we discuss below, the glossograptids are better classified with related groups which share a common ancestor: the high level given to the group in the present classification obscures their relationships. We discuss most groups of planktic graptolites here, excluding only those with reduced periderm (Retiolitidae) which pose detailed problems beyond the scope of this paper. CHARACTERS OF USE IN HIGH-LEVEE CLASSIFICATION Almost all taxonomists, other than pure pheneticists, recognize that it is unprofitable to treat all characters as equal. Some characters operate at a more fundamental level than others and are considered more appropriate in the recognition of high-level groups. We believe that all graptolite workers would concur, for example, that the final number of stipes in a many-stiped graptoloid is significant at species level but scarcely at a higher level, and stipe characters (thecal spacing, stipe width and the like) are regarded similarly. Recently (Cooper and Fortey 1982, 1983) we have emphasized aspects of sicula morphology and proximal structure in the discrimination of higher taxa, following the lead given by Bulman’s (1936) classic paper; other authors (Finney 1980; Kearsley 1982) have also implicitly adopted such an approach. Graptoloids sharing distinctive small growth stages were regarded as having descended from a common ancestor, and hence as belonging to a monophyletic group capable of being charaeterized taxonomically. In the course of describing the proximal end structure of early Ordovi- cian graptoloids we redefined several aspects of terminology for development type. It was shown that the isograptid development mode (with thl^ dicalycal) was the primitive mode for the graptoloids, as opposed to the apparently simpler artus mode (with thl’ dicalycal) which had been regarded as ‘ancestral’ on erroneous morphological grounds. Phyllograptid development was described for the first time, and the polyphyletic origin of quadriserial scandent rhabdosomes was recognized. This paper adopts the concepts and terminology introduced by Cooper and Fortey (1982, 1983), but there is no opportunity here to review all the evidence, and the reader is referred to our earlier works for details. One way of justifying such an approach is to consider the development of the rhabdosome during astogeny as a series of ‘instructions’ which have to be followed after commencement of development of the metasicula (see Table 1). Characters of high taxonomic importance are those which are initiated early in the sequence: a change to a single one of these may alter the developmental programme for the rest of the rhabdosome. Conversely, characters of lesser taxonomic moment occur at the end of the development sequence: a change in one of these is effected by a genomic change which leaves the rest of the development untouched. In this category belong the thecal and stipe characteristics, and those relating to the timing and suppression of a distal dichotomy. If our list is approximately correct, it shows that the greater number of such ‘instructions’ operate on the first few thecae of the colony, which in turn implies that structures at the proximal end are a priori of greater taxonomic significance. On the other hand, quite small changes higher in the list can, in FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 633 TABLE I . Scheme for progressive series of growth decisions which have to be made during the growth of the graptolite colony after the prosicular stage. 1. Initiation (or not) of virgella. 2. Point of origin and formation of initial bud. 3. Direction of growth of first theca (down, across, upward). 4. Direction of growth of distal part of sicula (straight, bend). 5. Growth of first crossing canal. 6. Growth (or not) of ‘counterbalancing’ lip on sicula. 7. Distal attitude of first theca (to form a pair with the sicula, or extend distally, etc.). 8. Angle subtended between thl' and thl^ (acute, obtuse, greater than 180°). 9. Apertural form of first theca (flared, rectimarginate, denticle etc., usually repeated along subsequent stipes in dichograptoids). 10. Decision about proximal dicalycal theca (thl ‘, thl^, th2', or later, or suppress). 11. Decision about number of thecae before second dichotomy. 12. Second dichotomy (repeats proximal end). 13. Decision about angle between daughter stipes. 14. Do subsequent dichotomies simply repeat proximal pattern, or are extra instructions introduced (e.g. change angle, introduce additional thecae between branching events, or suppress)? 15. Decision about thecal form to determine stipe characters (three parameters sufficient for graptoloids with simple thecae, see Fortey 1983). 16. Thecal elaboration along stipe (spines, lappets, curvature, and the like). theory at least, make a considerable difference to the aspect of the colony as a whole. Hence there is a greater chance for the parallelisms that are such a striking aspect of graptoloid history to be initiated by similar developmental switches late in astogeny— and these can only be recognized as such by examination of the more fundamental proximal development. It follows that higher level monophyletic units within the Graptoloidea are best recognized by development mode and structure of the proximal end. Species with matching proximal end structure are best classified together within a higher taxon. This reveals their common ancestry. The level at which a taxon is recognized is not well characterized by its overall ‘distinctiveness’ — however superficially usable this may seem —but by its common ancestry. While an entirely monophyletic classification is the aim, we admit that there are many uncertain- ties, mainly deriving from a lack of morphological information on species designated as types of higher groups; for some time at least it will be necessary to live with some large undivided groups and some paraphyletic high-level taxa. We show inferred relationship in cladograms (text-figs. 1-3, 7-10) and a tentative scheme showing the inferred phylogeny of the major groups discussed here is given in text-fig. 1 1. The implications of phylogenetic classification for the definition of the Grapto- loidea as a whole are considered first. In the following discussion the terms two, three, and four primary stipes refer to biradiate, triradiate, and quadriradiate rhabdosomes respectively. The primary stipes are in fact formed from successive dicalycal thecae in the proximal development of the rhabdosome; they are not produced simultaneously. In a biradiate rhabdosome, for example, the apertures of the first thecal pair lie in the first pair of stipes (first order stipes) and dicalycal thecae are separated along the stipe by a normal theca (text-fig. 4c), whereas in triradiate and quadriradiate rhabdosomes the apertural regions of one or both primary thecae lie in a second order stipe (text-fig. 4d, e). Derived characters discussed in the text are cross-referenced to the cladograms by numbers in bold face. CLASSIFICATION OF THE ANISOGRAPTIDAE The Family Anisograptidae has been traditionally classified in the Order Dendroidea. The intermedi- ate grade of organization of the anisograptids has been recognized at least since Bulman (1949, 634 PALAEONTOLOGY, VOLUME 29 ORDER DENDROIDEA ORDER GRAPTOLOIDEA - TEXT-FIG. 1. Diagram of relationships of anisograptids, assuming that loss of bithecae (9) happened only once, showing the monophyletic Graptoloidea as a definable taxon. Dashes show derived character states; circles show retained primitive character states. Derived characters are: 1, presence of rooting sys- tem; 2, sicula retains nema in adult colony; 3, bilaterally symmetrical colony; 4, dichotomies more or less regular; 5, reduction series in number of primary stipes; 6, three or fewer primary stipes (triradiate or biradiate); 7, two primary stipes (biradiate); 8, ultimate number of dichotomies fixed; 9, loss of bithecae; 10, perfect radial habit; 11, pendent habit; 12, three primary stipes (triradiate); 13, three or more orders of dichotomy; 14, suppression of distal dichotomies; 15, large conical rhabdosome; 16, stipe reduction, loss of dissepiments; 17, three or more orders of dichotomy; 18, as 14; 19, large spreading colony (three orders of dichotomy or more); 20, dichotomies delayed; and 21, only one dichotomy. Retained primitive characters are: a, four primary stipes; b, dissepi- ments; and c, retention of bithecae. p. 533) described them as ‘transitional between the Dendroidea and Graptoloidea’. Bulman then I and subsequently emphasized the seminal role of Dictyonema (now Rhahdinopora) flahelliforme in I anisograptid evolution. Indeed he stated that it was ‘tempting to claim D. flahelliforme as the first ^ ancestral graptoloid’, but baulked at so doing because it would ‘create problems in diagnosis and definition’. So it has remained. Equally, the ‘close and intricate’ (ibid., p. 534) relationship between ! the Graptoloidea and the anisograptids has become an accepted part of the broad view of graptolite evolution. The diagnosis of the Anisograptidae given by Bulman in 1950 when he established the family emphasized the ‘dendroid structure’ of the stipes, referring to the presence of bithecae. This is, of course, a retained primitive character of the Graptolithina as a whole, and as such can scarcely be used as diagnostic. The same applies to the basically isograptid development type (Cooper and Fortey 1983). In text-fig. I we show a cladistic representation of the character distribution within anisograptid genera. This figure is not based on stratigraphic considerations but primarily on a hierarchy of primitive to derived character states, although these are usually mirrored by strati- graphic sequence. Thus, the quadriradiate condition is considered primitive, and three (6) and two (7) primary stipes (triradiate and biradiate) as derived character states (text-fig. 4), because sup- pression of dichotomies is regarded as a derived condition. In addition, the earliest anisograptids are unique in having two successive dicalycal thecae to produce the quadriradiate condition; FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 635 suppression of one of these to produce a normal dichograptid-like division on one side of the rhab- dosome is a derived condition in triradiate forms (Cooper and Fortey 1983, fig. 8). The form at the base of the cladogram is that with the greatest number of retained primitive characters (marked as circles), including four primary stipes, retention of dissepiments, and irregular dichotomies. This is the subspecies Radiograplus rosieranus flexibilis described recently from the base of the Ordovician in western Newfoundland (Fortey in Fortey et al. 1982). Recent interpretations of early graptolite phylogeny (ibid.; Erdtmann 1982) based on field occurrences have emphasized the secondary deri- vation of the pendent Rhabdinopora rhabdosome, and this is also how it appears on the cladogram. R.flabelliformis is not necessarily \hefons et origo of all subsequent planktic graptolites as tradition- ally claimed, presumably from its widespread geographic occurrence in many ropk sections at the base of the graptolitic sequence. The derived character linking all the anisograptids with higher graptoloids is the presence of a nematophorous sicula. In text-fig. 1 the assumption is made that loss of bithecae (9) occurred only once (shown as the autapomorphy of ‘all other graptoloids’). If this were the case it would at least be possible to define the Graptoloidea as used at present as a monophyletic group, but even here there are problems. Character 7— two primary stipes with the inclusion of at least one pair of thecal apertures before the first dichotomy (text-fig. 4e) — is also shared between two genera of anisograptids {Clonograptus and Adelograptus) and the higher graptoloids. Why should we not define the Graptoloidea as a monophyletic group having this character rather than on the loss of bithecae? If the cladogram is at all correct it is clearly preferable to extend the Graptoloidea to include all the anisograptids. TEXT-FIG. 2. Diagram of relationships of anisograptids, this time assuming that loss of bithecae (9) is polyphy- letic (compare text-fig. 1). Explanation of symbols as in text-fig. 1 regarding retained primitive characters and derived characters 1-12. Derived characters 13-25 are: 13, large spreading colony (three or more orders of dichotomy); 14, dichotomies after dl delayed; 15, three or more orders of dichotomy; 16, suppression of distal dichotomies; 17, large conical rhabdosome; 18, stipe reduction; 19, loss of dissepiments; 20, small size, fewer than three orders of dichotomy; 21, two orders of dichotomy (four or rarely three stipes); 22, three or more orders of dichotomy; 23, as 16; 24, lax branching, sparse dichotomies; and 25, only one dichotomy. 636 PALAEONTOLOGY, VOLUME 29 defined by the derived character of retaining a nematophorous sicula in the adult. Otherwise there is simply no rational way of separating the anisograptids from the graptoloids other than one of traditional usage. Several authors (e.g. Bulman 1960, 1970; Erdtmann 1982) have contended that the loss of bithecae did not occur once, as shown by its treatment as a derived character in text-fig. 1, but several times. If this is the case, then the closest relatives of each such graptoloid lacking bithecae is an anisograptid with bithecae. A cladogram showing the distribution of characters suggested by such a hypothesis is shown on text-fig. 2. This has to incorporate several dichograptid genera with their supposedly closest related anisograptid neighbours. The classification is now even more difficult than that suggested by text-fig. 1. It becomes impossible to define the Graptoloidea as a monophyletic group if its bounds are to be taken as at present. The stippled area on text-fig. 3 shows how the Order at present defined comes out as (at least) triphyletic. The alternative, to extend the Graptoloidea to include all anisograptids to the right of Rhabdinopora, would leave three genera (Radiograptus, Staurograptus, and Aletograptus) as stem-group graptoloids definable, if at all, only by combinations of primitive characters. Furthermore, the notions of what should or should not be included within a monophyletic Graptoloidea will depend greatly on opinions about whether or not Pendeograptus and Pseudobryograptus share a common ancestor with Bryograptus, questions which are not yet resolved. None of the relationships between bithecae-bearing and graptoloid-like genera shown on text-fig. 2 are yet proved from stratigraphic or morphological grounds, so that the simpler cladogram is still a possibility. Whichever cladogram proves to represent better the phylogenetic relationships, the implications for classification are the same: genera usually referred to the Anisograptidae are best included within the Graptoloidea. The Order then becomes definable by ‘presence’ characters (especially Graptoloidea without bithecae TEXT-FiG. 3. Diagram of relationships from text-fig. 2 showing how the present concept of Graptoloidea (stippled) is polyphyletic; this is avoided by taking the phylogenetic definition proposed in this paper. FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 637 nematophorous sicula (2)) rather than an ‘absence’ character (loss of bithecae) which is in contention as polyphyletic. The Order Graptoloidea can then be diagnosed thus: Graptolithina retaining nematophorous sicula in adult, and with bilaterally symmetrical colonies, secondarily lost in Monograptacea. Remarks on some of the characters and taxa 1 . Radiograptus. The type species R. rosieranus rosieranus has three primary stipes (12) according to Bulman (1950). The earliest species R. r. flexibilis Fortey in Fortey et ai, 1982 has four primary stipes— and it is this form which is shown as the sister group of the rest of the Graptoloidea on the diagrams. The type species may be derived directly from that with four primary stipes, or is conceivably independently derived, e.g. from Rhabdinopora. It is not shown on the diagrams because of these uncertainties. 2. Chitinized stolon. This is another primitive character which has been described from several anisograptids (e.g. Rhabdinopora, Clonograptus), and whose general presence in primitive grapto- loids is perhaps likely, but it is not widely enough described to be marked on the diagrams. 3. Rhabdinopora. Some members of the flabelliformis group have been shown to possess four primary stipes (Legrand 1974). These forms, strictly speaking, should be separated from the rest on the cladograms as the sister group of all those taxa to the right of Rhabdinopora s.s. The strati- graphical evidence for primary stipe reduction is so convincing, however, that they are included within the Rhabdinopora j Bryograptus segment, and the derived character (3) is taken to include the tendency for primary stipe reduction within accepted monophyletic groups to the right on the cladograms. TEXT-FIG. 4. Primitive and derived conditions in anisograptid proximal ends, a, thecal diagram of the primitive quadriradiate condition (‘four primary stipes’), showing successive dicalycal thecae (shaded); bithecae omitted following Cooper and Fortey (1983). b, the derived triradiate condition (‘three primary stipes’) with loss of one dicalycal theca. c-E, rhabdosome proximal ends showing position of apertures of the first thecal pair with respect to the stipes of the first and second orders in the c, biradiate, d, triradiate, and e, quadriradiate condition. 638 PALAEONTOLOGY, VOLUME 29 4. Pendent habit (11). For a recent view on the geological relationships of Rhabdinopora and Bryograptus see Erdtmann (1982) and Fortey in Fortey et al. (1982). Erdtmann’s Subfamily Rhab- dinoporinae is based primarily on the single derived character of pendent habit. Its usefulness will depend on whether the Pseudobryograptus and Pendeograptus link is confirmed; if it is, a phylo- genetic taxon Rhabdinoporinae will have to include these genera. 5. Two primary stipes (7). All graptoloids above the acquisition of this character have, or were derived from, forms with two primary stipes. The characteristic feature is that a pair of thecal apertures open out on the Tunicle’ prior to dichotomy(ies) (see text-fig. 4). This feature is also characteristic of all branched dichograptids. Status of the Anisograptidae If it is accepted that the Anisograptidae are to be classified within the Graptoloidea as we advocate, there remains the question of the status of this family, based as it is primarily on the retention of the primitive character of bithecae, with the addition of other primitive characters (e.g. dissepiments, irregular dichotomies) in certain early species. There now exist four subfamilies proposed within the group: Staurograptinae Mu, 1974; Anisograptinae Bulman, 1950 {sensu Mu 1974); Adelograpti- nae Mu, 1974; and Rhabdinoporinae Erdtmann, 1982. The first three are essentially formaliza- tions of the main groupings shown on the cladogram (text-fig. 1), based on the number of primary stipes (four, three, and two respectively). The Rhabdinoporinae has to be based on pendent habit because the other character claimed by Erdtmann (presence of dissepiments) is again a primitive character. It is doubtful whether any purpose is served by recognizing these subfamilies. The Staurograptinae and the Anisograptinae include only the nominate genera, together with one other genus each which are paedomorphic derivatives from the nominate genus (Aletograptus and "Triograptus’’ respectively)— and it could be argued that these are not worth generic rank. The use of Rhabdino- porinae depends on whether or not the genera Rhabdinopora and Bryograptus belong to the same phylogenetic unit as Pseudobryograptus and Pendeograptus, which is not proven. Stratigraphic evidence may also prove Anisograptus to have been derived from Staurograptus, in which case separate subfamilies would be redundant. Finally, if either of the cladograms is correct then the Adelograptinae (including Clonograptus, Adelograptus, and Kiaerograptus) is a paraphyletic group; this is not acceptable unless the Subfamily is classified with the rest of the graptoloids (see above), which Mu (1974) does not propose. Subfamily division in the Anisograptidae either seems to introduce excessive splitting, or is taxonomically ambiguous. For the moment there seems to be no alternative but to retain the Anisograptidae as a paraphyletic taxon characterized by the retention of bithecae. If the polyphyletic view for the loss of bithecae (as shown in text-figs. 2 and 3) is supported by further work, at least the inclusion of the Anisograptidae within the Graptoloidea allows the whole group to be classified as monophyletic. The detailed relationship of the anisograptids to the high-level taxa in the higher graptoloids lacking bithecae (discussed below) is certainly still unclear. CLASSIFICATION OF GRAPTOLOIDS OTHER THAN THE ANISOGRAPTIDAE As with the anisograptids there are many problems and unanswered questions to be faced before a detailed classification of the remaining graptoloids can be attempted. Here we suggest a grouping based primarily on the structure of the sicula and proximal part of the rhabdosome (text-figs. 7- 10). One problem at the outset is the polyphyletic nature of many genera. Hence it is not possible simply to list all genera and show their phylogenetic relationships on a cladogram if we are aiming towards a monophyletic classification. For this reason we have listed only those genera (appropriately qualified where necessary), species groups, subfamilies, and families for which there are good grounds to assume monophyly. Diplograptid genera are particularly difficult, based as they are primarily upon thecal morphology; the present genera are widely acknowledged to be FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 639 polyphyletic. Bulman’s analogy of diplograptid evolution— a bundle of fibres (lineages) interwoven into several parallel ropes (genera) — implies that individual lineages pass from one thecal form genus to another and back again. Bulman (1963, p. 414) wrote: ‘Admittedly there has so far been little attempt to trace phylogeny, and future opportunities to carry out more detailed morphological and phylogenetic investigations will undoubtedly introduce taxonomic problems.’ It is exactly these taxonomic problems that we are faced with here. The species groups shown on the cladograms are discussed briefly in the following sections, but we are conscious of many problems that remain before our graptoloid classification can be regarded as definitive. Our aim at the moment is to try to identify morphological characters that are useful in recognizing high-level monophyletic groups and to trace this distri- bution throughout as many sub-groups as possible. Our main attention is given to the dichograptids, but we include diplograptids in some cladograms (text-figs. 9 and 10) to show some characters useful for establishing relationships within this latter group, and to show how they relate to Dichograptina. Lack of morphological information about some groups means that they are perforce left undivided. The ‘species groups’ listed in the clado- grams are not necessarily of equal rank. We do not recognize a taxon exactly equivalent to the Suborder Diplograptina as used by Bulman (1970), so we refer to the biserial graptolites of this group informally as diplograptoids in the following discussion. Primitive characters shared with anisograptids. There are a number of characters shared with the anisograptids, presumed to be primitive, and therefore of no direct use in recognizing groups within non-anisograptid graptoloids. These include: 1, isograptid development type; 2, simple thecal form; 3, rhabdosome with biradiate origin; 4, bilaterally symmetrical rhabdosome; 5, regular number of dichotomies in rhabdosome; 6, dicalycal thecae separated on the stipes by one or more normal thecae (i.e. not successive); 7, capacity for numerous dichotomies in the rhabdosome; and 8, stipe length unrestricted. Suborders Dichograptina and Virgellina At the highest level we recognize two major groups within the non-anisograptid graptoloids, based upon the presence or absence of a virgella in the sicula (text-fig. 5a-f). The Suborder Virgellina is proposed here for those graptoloids having such a spine (67); it is shown as a monophyletic group on text-figs. 7 and 8. At first glance this feature may not seem a very significant one on which to define a major division. It is, however, developed on the metasicula, that part of the colony directly developing from the larval prosicula, which is presumed to be sexually produced (all subsequent development of the colony being by asexual budding). Furthermore, the virgella is not merely a late stage apertural modification, but is present in earlier metasicular development, as shown on growth stages (Bulman 1932, pi. 2, figs. 1-20) and by deflection of growth lines (text-fig. 5a, b). It is in fact one of the earliest ‘decisions’ required by the developing colony (Table 1). In general among graptoloids the sicula is not highly variable in morphology: elaboration of the aperture is known in Corynites (Kozlowski 1956) and many monograptids, but in the bulk of the graptoloids variation is largely confined to overall size (other than dorsal apertural processes produced at a late stage in metasicular development). The presence or absence of a virgella is one of the very few structural features of sicular morphology, if not the only one, that can be recognized and readily used in taxonomy, at least in our present stage of knowledge. Its presence serves to unite the whole group of diplograptids s.l., monograptids, dicellograptids, and nemagraptids with the one group of ‘dichograptoids’ that has a diplograptid-like proximal end (Cooper and Fortey 1983, fig. 5)— the phyllograptids— while at the same time excluding the glossograptids, now thought to be derived from Pseudisograptus (Cooper and Ni 1986). It is significant that the development and structure of the virgella is apparently closely similar in Phyllograptidae and the remaining virgellinids (cf. text-fig. 5a, b). The change in origin of thl ‘ from the antivirgellar side (in Phyllograp- tidae) to the virgellar side (in the remaining virgellinids) is not accompanied by a change in the virgella itself — another reason why we believe it to be monophyletically derived and of high taxonomic value. 640 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 5. The virgella and analogous structures, a, Xiphograplus formosus formosiis (Bulman), par- tially grown metasicula showing early formation of virgella (after Skevington 1965, fig. 19). b, Dicrano- graptus nicholsoni Hopkinson, showing early formation of virgella (after Bulman 1944, text-fig. 20). c, D, Climacograptus typicalis Hall, partially grown (c) and fully grown (d) metasicula showing late development of dorsal spines (after Bulman 1932, pi. 1, figs. 1, 4, respectively), e, Corynoides cf. gracilis Hopkinson, distal end of mature metasicula showing lamelliform ventral process (after Bulman 1947, text-fig. 39/n). f, Pseudisograptus geniculatus (Skevington), complete but immature sicula showing ex- tended ventral process (after Skevington 1965, fig. 52). Scale bars represent 0-2 mm. The change in side of origin of thl‘ is likely to have accompanied the change from a prosicular to a metasicular origin (discussed below). The presence of a virgella spine may or may not be a character with particular adaptive importance for the developing colony, but it should be noted that it is quite legitimate to use even a relatively minor character in the recognition of a major group. The character need only signify a speciation event in the early history of an evolving plexus; if the character is retained in descendants of the first species having it, then it becomes of taxonomic importance regardless of theories about its function. The taxonomic level which the character serves to define is really a matter of the diversity and size of the subsequent speciation— in the case of the Virgellina more than half the known graptoloids. Our grouping of forms with a virgella differs fundamentally from previous high-level groupings, hence we coin a new subordinal name— Virgellina. The nearest existing concept for this group is possibly Axonophora Freeh, 1897 (as reviewed by Mu 1950, 1974), which, however, excluded the Nemagraptidae, Dicranograptidae, and Phyllograptidae. For the other suborder (discussed below) we use Dichograptina Lapworth, 1880, in order to avoid conceptual confusion with the Didymo- graptina Lapworth, 1880, which was used in the Treatise (Bulman 1970) to include the Nemagrap- tidae, Dicranograptidae, and Phyllograptidae— all families which are included within the Virgellina here. There are outstanding problems in the relationships between the Dichograptina (in which we include the Glossograptacea and the Dichograptacea) and the Virgellina; this is indicated by an FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 641 unresolved trichotomy on text-figs. 7-10. A strictly cladistic treatment would accord Glossograp- tacea and Dichograptacea subordinal status like that of Virgellina in view of this trichotomy; we prefer for the moment to unite these two superfamilies within a Suborder Dichograptina, even though this is based upon a shared primitive (absence of virgella) character rather than a derived character, at least until more is known about the phylogenetic history of these graptolites. Discussion of characters Isograptid symmetry (55). The Dichograptina is divided into two major groups. The superfamily Glossograptacea is characterized by having isograptid symmetry, i.e. the sicula and thl' form a symmetrical pair (text-fig. 6c, d), and the axis of rhabdosome symmetry passes between them (Cooper 1973). The primitive condition is found in the remaining Dichograptina; the axis of rhabdosome symmetry passes through the sicula itself, and the sicula and thl' form an asymmetrical pair. Development type. Development type refers to the thecal budding arrangement at the first dichotomy in the rhabdosome (Cooper and Fortey 1983). The primitive condition shared with dendroids and early graptoloids is the isograptid development type; this type prevails in the great majority of Dichograptina and in the Family Phyllograptidae of the Virgellina. It has been shown that in species whose thecal budding pattern is known at distal dichotomies (i.e. at formation of branches) the pattern replicates the isograptid development type (Cooper and Fortey 1983). Thus only one development type is employed throughout the entire development of the rhabdosome. Deviation from the isograptid type goes in one of two directions. First, the dicalycal theca may appear earlier in the sequence and may be thl' instead of thl". This produces what has been termed the artus development type, and is found in some pendent didymograptids of Arenig to Llanvirn age, together with odd species like Kinnegraptus kinnekullensis and Oncograptus iipsilon. Secondly, the dicalycal theca may appear later in astogeny and be th2‘ (72) or a later theca, collectively referred to as diplograptid development type by Cooper and Fortey (1982). Further work is needed to characterize the various development types found in this group, but the delayed dicalycal theca serves as a useful apomorphic character to define superfamilies Diplograptacea (and Dicranograptacea). In Monograptidae the initial dichotomy is lost and the character has no chance of expression, but we know that Glyptograptus s.s., which has it, is the sister group of monograptids. Proximal structural types. We rely heavily on proximal structural types in defining major groups within the non-anisograptid graptoloids. These types are discussed and described in the outline of the proposed classification below and are illustrated in text-fig. 6. The earliest diplograptaceans, the "Glyptograptus' austrodentatus group of species (which differs from the type species G. ta?nariscus), have a strong downward component of growth in their early thecae (thl', thC, and th2'), giving the proximal end a square symmetrical outline. Particularly in later diplograptids, thC and th2' have a less prominent downward component of growth giving the proximal region a more pointed outline. A morphological transition between one and the other was claimed by Bulman (1936, pp. 88-93, text-fig. 29), who later named and described them as the streptoblastic and prosoblastic type respectively (Bulman 1963). Commonly associated with the streptoblastic type is the left-handed origin and strongly sigmoidal growth path of thF. The streptoblastic type is found in Dicaulograptus hystrix, Gymnograptus ret hides, Pseudoclimaco- graptus spp., as well as the austrodentatus group. The prosoblastic type is typified by such forms as Orthograptus apiculatus (described by Bulman 1946), O. gracilis (sensu Bulman 1932, and Amplexograptus maxwelli (sensu Walker 1953). A more refined classification of proximal structural types is applied to all diplograptoids by Mitchell (in press), and incorporates features of growth and construction such as the formation of a cowl during the early development of thC and its daughter th2', and the formation of a ventral plate which grows up and eventually joins the cowl to form the second ‘crossing canal’. Nine proximal patterns (termed A— I) are recognized and fully explained by Mitchell (in press). The terms streptoblastic and prosoblastic therefore no longer seem useful, and in our cladograms 642 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 6. Proximal structural types used in classification, a, anisograptid listed as Clonograptus tenellus? by Hutt (1973, fig. 6c). b, Tetragraptiis {Pendeograptus) fruticosus Hall (after Cooper and Fortey 1982, fig. 30/). c, Pseiidisograptus geniculatiis (Skevington) (after Skevington 1965, fig. 536). D, Isograptus ^gihbendus' (Nicholson semu Bulman 1932, pi. 8, fig. 2). e, PhyUograptus typus Hall (after Cooper and Fortey 1982, fig. 71/). F. Sigmagraptus praecursor Ruedemann (after Cooper and Fortey 1982, fig. 61c). G, Tetragraptiis (Tetragraptus) pliyllograptoides triiimphans Cooper and Fortey (1982, fig. 22b). H, 'Glyptograptus' dentatus (Brongniart) (after Skevington 1965, fig. 61a). Scale bars represent 0-5 mm. FORTEY AND COOPER: GRAPTOLOID CLASSIEICATION 643 (text-figs. 7 and 8) we employ Mitehell’s proximal patterns instead, and follow his mapped distribu- tion of these among the diplograptoids. Cladia (73). Cladia are branehes formed at the apertural margin of the sicula or a theca; they differ fundamentally in their mode of origin from the branches produced by dichotomous division. They are known in the Cyrtograptidae and have recently been found in the Nemagraptidae (Finney 1985). The capacity for production of cladia presumably appears with the change in development type from isograptid to diplograptid. If this inference is correct it can be predicted that branching in the Dicranograptidae (e.g. in rare specimens of Leptograptiis, and in Ordosograptus) is cladial rather than dichotomous. Biseriality and mojioseriality. The primitive state for the Graptoloidea is monoseriality. Among the Dichograptina scandency was attained at least six times independently if the relationships shown in text-figs. 7 and 8 are correct (independent lineages would be found for Pseudophyllograplus, Pliyllograptus, Pseudotrigonograptiis, Skiagraptus, Cardiograptm, and the Glossograptidae), pro- ducing biserial, triserial, or quadriserial forms. Among the Virgellina it is possible that scandent biserial rhabdosomes were also attained more than once, but our cladograms and inferred phylo- genetic tree (text-fig. 11) suggests that this was not so (74). If stratigraphic evidence is accepted that the ‘G.’ austrodentatus group is ancestral to the Diplograptacea, then it follows that biseriality is an apomorphic character for the whole group, and that monoseriality has been secondarily acquired in the nemagraptids and dicranograptids (86) and, of course, in the Monograptacea (79). Prosicidar and metasicular origin of thl' (70). A prosicular origin of thl ' is primitive, being shared by the Anisograptidae (Hutt 1974) and the early Dichograptina and Virgellina (Cooper and Fortey 1982). Unfortunately, preservation is rarely good enough to determine the point of origin of the first theca, but a prosicular origin seems to be more general among the Dichograptina, whereas a metasicular origin is general (with apparently no exceptions) among the Dicranograptacea and Diplograptacea (70). A prosicular origin is retained in the Phyllograptidae. A metasicular origin is regarded as a synapomorphy for the Monograptacea and the Diplograptacea. Its rare appearance in certain Dichograptina (e.g. Didymograptus artus) is, therefore, considered to be a parallelism. Stipe reduction and dichotomy delay. Stipe reduction has long been regarded as an important ‘trend' in graptolite evolution. More fundamental, in terms of the rhabdosome development programme, is the number of orders of dichotomy present, since rhabdosomes with, for example, three orders of dichotomy may have five, six, seven, or eight terminal stipes (e.g. Dichograptus octobrachiatus). We therefore discuss the branching dichograptinids in terms of the pattern, distribution, and number of orders of dichotomy. The first order is termed dl , the second order d2, and so on for ease of reference. There is no doubt that the capacity for many orders of dichotomy, as expressed in the multiramous dichograptinids, is a primitive character shared with the anisograptids. It is present in the sister group of the higher Graptoloidea {Clonograptus; text-fig. 1). Furthermore, this assumption is supported by at least one stratigraphically based lineage known from several sections around the world, this being Pendeograptus fruticosus in which first four, then three, then two stiped forms appear in upward sequence. Recognition of the primitive nature of multistiped rhabdosomes is important for establishing relationships among the Dichograptacea. In the primitive state dichoto- mies are consecutive (defined under branching styles) at least for the first several orders, and this state is modified by delaying a particular dichotomy, for example d3, such that the stipe segment between d2 and d3 contains several thecae instead of just one. The least modification of the primitive state is a delay in distal dichotomies (36). Suppression of a dichotomy can be regarded as an indefinite delay, and more fundamental changes to the rhabdosome are introduced by the more proximal position of a suppressed or delayed dichotomy (29, 30, 34-44, 46, 49-51, and 54). The ultimate state is the suppression of dichotomy dl, resulting in the single-stiped rhabdosome of Azygograptus. 644 PALAEONTOLOGY, VOLUME 29 Branching style. Terms to denote styles of branching in a rhabdosome were given by Cooper and Fortey (1982, table 2). In progressive branching, each of the two daughter stipes resulting from a dichotomy divides again, whereas in monoprogressive branching (52, 53) one stipe remains undivided (e.g. Sigmagraptiis, Pterograptus, and Brachiograptiis). Dichotomies can be consecutive, when di- calycal thecae are separated along the stipe by a normal ‘unicalycal’ theca, or delayed, when two or more thecae separate dicalycal thecae. Monoprogressive branching usually follows one or more orders of progressive branching in the proximal regions of the same rhabdosome, either in a regular ordered manner, as in Goniograptus, or in an irregular manner, as in Clonograptus or Zygograptus, where occasionally stipes of a relatively early order remain undivided while their sister stipes go on to several further orders of dichotomy. Distal dichotomies of multistiped rhabdosomes, particularly those with progressive branching, are commonly somewhat delayed; this is necessary to avoid overcrowding of stipes. A single early dichotomy may be delayed, as in Zygograptus (35), or all dichotomies after one, two, or three orders of consecutive dichotomy may be delayed, as in Mimograptus (40), Temnograptus (38), and Ortliodichograptus (37) respectively. No distinction is made here, so far as branching type is concerned, between ‘lateral’ and ‘dichoto- mous’ branching (see Bulman 1970, pp. V82-85). All branching is produced by dichotomy (also referred to as dichotomous division or stipe division) and the difference between lateral and dichoto- mous branching is the result of the directions of growth of the two daughter stipes. The distinction has little influence on subsequent decisions in a development programme and is unlikely to have much taxonomic importance. CLASSIFICATION OF THE DICHOGRAPTINA Dichograptacea The major group of dichograptinids is the Superfamily Dichograptacea (text-fig. 7), which lacks isograptid symmetry. Four subfamilies can be distinguished on the basis of proximal end structure. In the Subfamily Sigmagraptinae Cooper and Fortey, 1982 (text-fig. 6f) the sicula and early thecae are slender, especially the prothecal segment of th2', and thl' and thP diverge from the sicula at different levels; these features can be referred to as the sigmagraptine proximal end (47). Relationships within the subfamily are based on branching characteristics: forms with monopro- gressive branching (52, 53) have one or two initial progressive (normal) dichotomies, e.g. Sigma- graptus (?synonym of Thamnograptus Hall) and Goniograptus respectively, and in the case of Yushanograptus the second dichotomy is delayed (54); Laxograptus has only progressive dichoto- mies, but all are irregularly delayed (50); Acrograptus can be regarded as a sigmagraptine in which dichotomy d2 has been suppressed, and could thus be more closely related to either the Laxograptus or Sigmagraptus clades. The Subfamily Tetragraptinae Mu, 1950 includes forms with a massive structure on the reverse side of the rhabdosome formed by the crossing canals thl“, th2', and the proximal portion of th2^. This proximal end can be referred to as a T. serra proximal type (28) and is displayed by T. (T.) serra, T. {T.) reclinatus, and T. (T.) pliyllograptoides (see Cooper and Fortey 1982, fig. 22; text-fig. 6g herein). Reclined stipes and suppression of distal dichotomies appear to accompany the feature. The T. {T.) serra species group, T. pliyllograptoides, Pseudophyllograptus cor, and the remaining Pseudophyllograptus spp. all belong within this group; the relationships among them have been discussed by Cooper and Lindholm (1983) and are shown on text-fig. 7. Dichograptus solidus may represent a member of the group in which dichotomy d3 has not been suppressed; it is suppressed in all others. The Subfamily Dichograptinae Lapworth, 1880 retains a primitive, generalized dichograptid proximal end and is subdivided on the pattern and distribution of dichotomies. The group is characterized by the regularity of its branching, at least until delayed dichotomies are introduced at higher levels. At the lowest level a delay in dichotomy d2 (35) results in the rhabdosome of FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 645 Dichograptacea TEXT-FIG. 7. Cladogram of relationships of the main species groups included within the Superfamily Dicho- graptacea. Suggested groupings at subfamily and family level are indicated. All groupings are monophyletic. Derived characters are; 26, ultimate number of orders of dichotomy in the rhabdosome limited (generally not exceeding seven); 27, reclined stipes; 28, serra-iypQ proximal end; 29, dichotomy d4 suppressed; 30, dichotomy d3 suppressed; 31, partially scandent; 32, fully scandent; 33, proximal gape; 34, branching pattern regular (lost where dichotomies become delayed); 35, dichotomy d2 delayed; 36, dichotomies after d4 delayed; 37, dichotomies after d3 delayed; 38, dichotomies after d2 delayed; 39, dichotomy d3 suppressed; 40, dichotomies after dl delayed; 41, dichotomy d2 suppressed; 42, dichotomy dl suppressed; 43, dichotomy d4 suppressed; 44, five or more stipes attained by proximal dichotomy; 45, prothecal folds; 46, dichotomy d4 suppressed; 47, sigmagraptine proximal end; 48, all dichotomies delayed; 49, dichotomy d4 suppressed; 50, dichotomies irregularly delayed; 51, dichotomy d2 suppressed; 52, monoprogressive branching after dichotomy d2; 53, monoprogressive branching after dichotomy dl; and 54, dichotomy d2 delayed. Primitive retained characters are: d, prosicular origin of thl '; e, progressive (normal) branching pattern; f, generalized dichograptid proximal end; g, isograptid development type; h, capacity for large number of dichotomies; i, lack virgella; and j, asymmetry of sicula and thl '. 646 PALAEONTOLOGY, VOLUME 29 Zygograptiis. A progressive delay in dichotomy, starting from distal and working towards proximal dichotomies, characterizes the multiramous genera Ortiwdichograptus, Temnograptus, and Mimo- gniptus (37, 38, and 40). Suppression of dichotomies can also be traced progressively from distal to proximal through the classical series Loganograptus-Dichograptus octobrachiatus, tetragraptids of the quadrihrachiatus and approximatus groups, and extensiform didymograptids (39, 41, 43, and 44). It becomes clear that on the basis of these branching characteristics there is little justification for maintaining all the members of the Temnograptus group (i.e. Temnograptus Nicholson, Schizo- graptus Nicholson, Trochograptus Holm, Holograptus Holm, and probably Calamograptus Clark) as distinct genera; Temnograptus (or Schizograptus) takes priority. Two minor groups are also distinguished on text-fig 7. The sinograptids are characterized by the exaggerated development of prothecal folds (45), and relatively few stipes. Little is known of their proximal structure, but it appears to be of a generalized dichograptid type; hence their position on the cladogram is not certain. They could, for example, form a sister group with that containing Dichograptus octobrachiatus', this would assume that suppression of dichotomy d4 is a more primi- tive character than the acquisition of prothecal folds. They are tentatively shown as a separate sister group (Subfamily Sinograptinae Mu, 1957) to the three main groups above. A second minor group comprises the Arenig species referred to Clonograptus, such as C. persistens and C. magnificus. Little is known about the morphology of these forms; they retain primitive (clonograptid) branching characteristics but bithecae have not been seen. They may prove to be a distinct group lacking the regularity of branching and limited numbers of orders of dichotomy that characterize the four main groups; on the other hand, they may prove to be nothing more than stratigraphically late Clonograptus. They are not shown on the cladogram because of these uncertainties. Another group of species not shown is the pendent plexus {Didymograptus s.s., Pendeograptus, and Pseudobryograp- tus). Didymograptus is known on stratigraphic grounds to have had two evolutionary ‘bursts’, distinguished by Cooper and Fortey (1982) as the subgenera Didymograptus (Didymograptus) with artus development and Didymograptus (Didymograptellus) with isograptid development. The main problem in establishing their relationships is in assessing whether or not the pendent habit is of phylogenetic significance. If it is a derived character for the whole group, then the suppressed dichotomy d2 in Didymograptus s.l. has been independently derived and its sister group is Pendeo- graptus. As discussed above, Pendeograptus and Pseudobryograptus could in turn relate to the anisograptid group Rhabdinoporinae. We are not certain whether this is the case, however, and if the pendent habit was acquired more than once the closest relatives of pendent Didymograptus might be found in dichograptid groups such as Expansograptus or Corymbograptus. We leave the group unassigned because of these uncertainties. Finally, we note that the relationships among the four dichograptacean subfamilies are far from clear, and the relative order of the advanced characters which distinguish them is uncertain. In particular, Dichograptinae is retained on the basis of primitive, rather than derived characters, and its relationship to Tetragraptinae may be subject to change when more is discovered about the detailed structure of these early graptoloids. Glossograptacea The acquisition of isograptid symmetry (55) (isograptid proximal type; text-fig. 6c, d) appears to be accompanied by a reclined attitude of stipes and to characterize a broad group within the Dichograptina, the Superfamily Glossograptacea Lapworth, 1873 (text-fig. 8). The pseudisograptids have been described and discussed by Cooper and Ni (1986) and shown to share several features with the Glossograptidae, including partially developed pseudopericalycal (defined by Cooper and Ni) proximal structure. Fully developed pseudopericalycal structure (65) is confined to the Glosso- graptidae. The pseudisograptids and glossograptids were regarded as sister groups, and the Suborder Glossograptina Jaanusson was therefore downgraded to subfamily level; the Subfamily Pseudiso- graptinae was accepted as a paraphyletic stem group. The Corynoidinae were paedomorphically derived (66) from a glossograptid (either pseudisograptine or glossograptine) and share the extended ventral apertural processes with that group. Together the three subfamilies comprise the family FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 647 Glossograptacea 1 Pseudotrigonograptidae Isograptidae Glossograptidae 1 ‘ ' Pseudotrigonograptinae Isograptinae 7 / Pseudisograptinae Corynoldinae Glossograptinae 7 / 7 / 7 TEXT-FIG. 8. Cladogram of relationships of the main species groups included within the Superfamily Glossograptacea. Suggested groupings at the subfamily, family, and superfamily levels are indicated. All groupings are monophyletic except for the Pseudisograptinae which is retained as a paraphyletic stem group following Cooper and Ni (1986). Derived characters are: 55, isograptid symmetry of sicula and thl*; 56, dichotomy d2 suppressed; 57, scandency; 58, presence of manubrium; 59, high initial angle of inclination of thecae; 60, well-developed manubrium; 61, smoothly rounded ventral margin; 62, sicula and thl' project ventrally; 63, partial or complete scandency; 64, advanced manubriate thecal form (see Cooper and Ni 1986); 65, pseudopericalycal proximal structure (see Cooper and Ni 1986); and 66, paedomorphism. Retained primitive characters are: k, low initial angle of inclination of thecae. Glossograptidae. Cooper and Ni regarded the subfamily Isograptinae sensu Cooper and Fortey, 1982 to be the sister group of the Glossograptidae, but the isograptines may be diphyletic (Cooper 1973; Cooper and Fortey 1982). The species group containing I. primuhis Harris, I. scandens Cooper and Fortey, and the Swedish I. ^gibberulus' {sensu Bulman 1932) is thought to have been independently derived and is shown separately on text-fig. 8. Although the I. vicloriae group is thought likely to be more closely related to the Glossograptidae than to the I. prumdus group, we know of no characters (other than the primitive one of low initial thecal inclination) that unite the Isograptinae and the Glossograptidae without also including the I. primulus group; these are therefore shown as an unresolved trichotomy on text-fig. 8. CLASSIFICATION OF THE VIRGELLINA All members of the Virgellina carry a virgellar spine on the sicula (67). The suborder is divided in the first instance on the structure and development of the proximal end (text-figs. 9 and 10). The 648 PALAEONTOLOGY, VOLUME 29 primitive isograptid development type is retained by only one group, the stem group Phyllograpti- dae, which also retains the prosicular origin of thl ‘ and simple dichograptid thecae. Phyllograptus and Xiphograptus are the only two genera included with certainty, together with an unnamed reclined tetragraptimorph from Victoria (Vandenberg in Vandenberg and Stewart 1984). This no doubt reflects the lack of knowledge of sicular structure in the large majority of forms currently included in Dichograptacea, and the full list of Phyllograptidae is likely to prove larger. Tentatively included in the family is the Abrograptidae Mu, 1958, in which the periderm has become so reduced that it is no longer possible to interpret such features as development type and proximal structure; Diplograpfacea I VIRGELLINA Stem group Dicranograptacea Phyllograptidae Suborder Superfamily Glyptogr- Diplograplidae Virgellina, assuming that the right-handed origin of thl^ in the Nemagraptidae is a primitive character shared with the Dichograptina. Suggested groupings at family and superfamily level are shown; all are monophyletic except for the Dicranograptidae. Derived characters are: 67, virgella present; 68, dicho- tomy d3 suppressed; 69, dichotomy d2 suppressed; 70, metasicular origin of th'; 71, sigmoidally curved thecae; 72, th2' dicalycal; 73, capacity for cladia generation; 74, biserial; 75, simple sicula with dorsal sinus; 76, left handed origin of thl^; 77, complication of sicular apertural margin; 78, asymmetrical proximal end; 79, dichotomy dl suppressed; 80, sinus origin of thl', 81, cladia present (or potentially so); 82, type B proximal pattern; 83, symmetrical proximal end; 84, type A proximal end; 85, type C proximal end; 86, partial or complete monoseriality; 87, reduction of periderm; and 88, scandency. Retained primitive characters are: k, thl^ dicalycal; 1, prosicular origin of thl'; and m, dichograptid thecae. FORTEY AND COOPER; GRAPTOLOID CLASSIFICATION 649 VIRGELLINA Diplograptacea Suborder stem group Phyllograptidae Superfamily r 1 Glyptogr- Diplograptidae Monograptidae aptidae Orthograplidae Dicranograptidae Phyllograptidae Family TEXT-FIG. 10. Same as text-fig. 9 except that the Nemagraptidae and Dicranograptidae form a group sharing partial or complete monoseriality. This arrangement has the advantage of comprising only monophyletic taxa and being consistent with stratigraphy. a virgellar spine, however, is apparently present. Finney (1980) has suggested that the biserial form Reteograptus belongs within the family. ^ Azygograptiis' incurviis Ekstrbm (Finney 1980, text-figs. 9 and 10) also bears a virgella and should be included. In all the remaining Virgellina an isograptid development type has been modified by a delay in the dicalycal theca which, instead of being thl“ becomes th2* (72), th2", or a later theca. The loss of a capacity of (normal) isograptid branching and, instead, the acquisition of a potential for cladia generation appears to accompany this change. The approach to the problem of diplograptoid classification adopted by Kearsley (1982, 1985) and Mitchell (in press) is entirely consistent with the principles of graptolite classification set out here. The group is subdivided on features of the proximal end, particularly sicular structure, proximal symmetry, and pattern of proximal construction, and we largely follow Mitchell (in press) to whom reference should be made for explanation and definition of his nine constructional patterns A-I and synapomorphies. Two major groups can be recognized on the basis of proximal symmetry. First, a symmetrical proximal end (83) is shared by two families, Diplograptidae and Orthograptidae. The Orthograp- tidae includes, in addition to Orthograptiis s.s., 'CUniacograptus' typicaUs and related species, and 650 PALAEONTOLOGY, VOLUME 29 "Glyptograptus'' austrodenlatus. It is distinguished by having a type A proximal pattern (84), whieh becomes modified to types G or F in some species. The Diplograptidae includes Diplograptus s.s., Pseudoclimacograptus s.s., Dicaidograptus and related species. It is distinguished by having a type C proximal pattern (85), modified to types D or E in some species. The second major group has an asymmetrical proximal end (78) and comprises two families. The Family Monograptidae includes the monograptids and cyrtograptids, here grouped in one family because the only basis for distinc- tion between them is the presence of cladia (81)— a primitive character at this level— in the ‘Cyrto- graptidae’. The family is defined by having dichotomy dl suppressed (79) and thl' arising from a sinus in the developing metasicula rather than through a resorption foramen (80). The strong upward direction of growth of thl' links it with the second family, the (paraphyletic) Glyptograpti- dae, which includes Glyptograptus s.s. and "Climacograptiis' of brevis type, and has type B proximal pattern (82), modified to types H or I in some species. Insertion of the Nemagraptidae and the Dicranograptidae in the cladogram presents some prob- lems. Finney (1985) established that the Nemagraptidae have th2' dicalycal, and that branching is by the formation of cladia rather than by dichotomous division. He showed that Leptograptus has dicellograptid type thecae and proximal structure, and transferred the genus to the Dicranograpti- dae. We suggest two alternatives. In the first (text-fig. 9) the right-handed origin of thl^ in the Nemagraptidae is regarded as a primitive character shared with the Dichograptina, and the Nema- graptidae is accordingly regarded as a sister group to the Diplograptacea and Dicranograptidae. The Dicranograptidae shares with the Nemagraptidae a ‘simple’ dorsal apertural margin on the sicula, but shares with the Diplograptacea a left-handed origin of thP (76); it is therefore a sister group of the Diplograptacea. The Nemagraptidae and Dicranograptidae are grouped on the clado- gram as the paraphyletic Superfamily Dicranograptacea, which is considered preferable to each being assigned to a separate and monotypic superfamily. This arrangement, however, conflicts with the stratigraphic order of appearance of the characters (text-fig. 11); dorsal spines (or other modifications) on the sicula aperture (77) and left-handed origin of thl^ precede a simple sicular aperture (in nemagraptids) in the stratigraphic record by a gap equivalent to the whole of the Llanvirn Series. In the other alternative (text-fig. 1 0) the simple sicula and partial or complete monoseriality define a group comprising the Nemagraptidae and the Dicranograptidae. This group is a sister group to both the Diplograptidae and the Pseudoclimacograptidae in an unresolved trichotomy; all three, however, share a symmetrical proximal end (83). This arrangement results in a monophyletic Dicranograptidae (including the Nemagraptidae) and is consistent with stratigraphy if the simple sicula is regarded as a derived character state. At the superfamily level a revised definition of the Diplograptacea is suggested to include the Family Monograptidae (which comprises both monograptids and cyrtograptids). To distinguish the Monograptidae at superfamily level leaves a paraphyletic Diplograptacea— hence we regard it as in principle more satisfactory to recognize an enlarged, monophyletic Diplograptacea as shown. This grouping of forms with a virgella cuts across the traditional grouping based on thecal type. Hence fonns with glyptograptid thecae are present in each of the groups discussed above, as are forms with climacograptid thecae. Thecal form may well prove useful in recognizing phylogenetically based genera within each group, but it is of little use in discerning the major groups themselves. We have made no attempt to include the Retiolitidae in our classification because current investigations into their phylogenetic relationships would make their inclusion premature. For example, a discussion of relationships of the Retiolitidae and some minor diplograptid families, such as the Lasiograptidae and the Peiragraptidae, is given by Mitchell (in press). OUTSTANDING PROBLEMS 1 . The relationship between Virgellina, Glossograptacea, and Dichograptacea needs to be clari- fied. A careful search for forms with a virgella in the early Arenig, for example, might reveal further characters, in an ancestral or early species, useful for linking the Virgellina with one of FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 651 TEXT-FIG. 11. Phyletic tree of the main families and subfamilies of the Order Graptoloidea (excluding the pendent dichograptids and retiolitids) derived from phylogenetic relationships (as shown in cladograms, text- hgs. 1 -3, 7-10) and stratigraphic data. the other groups. At present no more than two orders of dichotomy are known in any Virgellina species and it is possible that the ancestral species lay within one of the pauciramous dichograptid groups. 2. The relationship between members of the pendent ‘dichograptids’, including Pendeograptus, the Didymograptelliis group, and the Didymograptus (nnirchisoni) group is quite uncertain, as is the question of whether all these members constitute a monophyletic group. Is the pendent habit a feature shared with the pendent anisograptids, or has it been independently derived? If one or all of these forms are derived from an anisograptid such as Bryograptus, and are therefore more closely related to it than to other Dichograptina, they should be classified in a separate group which includes Bryograptus and Rhabdirwpora. 3. There is a considerable morphological gap between the stem group Phyllograptidae and the remaining members of the Virgellina, across which several new characters are acquired, i.e. a change of the dicalycal theca from thl^ to th2', suppression of dichotomy d2, change to metasicular origin of thl', change to sigmoidal curvature of thecae, and acquisition of the potential for cladia generation. Our cladograms (text-figs. 9 and 10) suggest that eventually a form possessing some, but not all, of these characters might be found. 4. Regrouping of diplograptoid species into groups with comparable structure (see Mitchell, in press). Thecal morphology may well be useful for recognition of genera within these groups, but as they stand at present, most diplograptoid genera are strongly polyphyletic. Elucidation of the relationship between these groups based on proximal structure requires a careful survey of the available data on proximal morphology. 652 PALAEONTOLOGY, VOLUME 29 5. Further study is needed to clarify the relationships of groups with reduced periderm— the Retiolitidae and Abrograptidae. SUMMARY OF PROPOSED CLASSIFICATION Listed below is the proposed classification of the Graptoloidea together with characters (or character states) useful for diagnosis. A phyletic tree showing the stratigraphic ranges of taxa is given in text- fig. 11. Order graptoloidea; graptolites in which the nema is retained in the adult stage. Suborder dichograptina Lapworth, 1873; graptoloids lacking bithecae and virgella. Superfamily dichograptacea Lapworth, 1873 (emend.); dichograptinids lacking isograptid symmetry, number of orders of dichotomy in rhabdosome limited. Family dichograptidae Lapworth, 1873 (emend.); dichograptaceans lacking prothecal folds and sigmagraptine proximal end. Subfamily dichograptinae Lapworth, 1873 (emend.); generalized dichograptid proximal end, branching pattern regular (but lost when dichotomies become delayed). Subfamily tetragraptinae Mu, 1950 (emend.); proximal region of ^crra-type, reclined stipes, dichotomy d4 suppressed. Family sinograptidae Mu, 1957; characters of subfamily. Subfamily sinograptinae Mu, 1957; pronounced prothecal folds. Family sigmagraptidae Cooper and Fortey, 1982; characters of subfamily. Subfamily sigmagraptinae Cooper and Fortey, 1982; sigmagraptine proximal region. Superfamily glossograptacea Lapworth, 1873 (emend.); isograptid symmetry. Family glossograptidae Lapworth, 1873 (emend.); presence of manubrium. Subfamily glossograptinae Lapworth, 1873; pseudopericalycal or pericalycal proximal structure. Subfamily corynoididae Bulman, 1944; development arrested at th3, sicula and first theca elongated. Subfamily pseudisograptinae Cooper and Ni, 1986; paraphyletic, includes forms with manubrium but lacking pseudopericalycal structure. Family isograptidae Harris, 1933 {sensu Cooper and Fortey 1982); characters of subfamily Subfamily isograptinae Harris, 1933 {sensu Cooper and Fortey 1982); biramous, reclined, thecae with initially low angle of inclination. Family nov. (includes the ‘/.’ primulus group of species); biramous, reclined, thecae with initially high angle of inclination. Family pseudotrigonograptidae nov.; scandent, triserial or quadriserial. (Type species; P. ensiformis. ) Suborder virgellina nov.; graptoloids with virgella. Superfamily diplograptacea Lapworth, 1873; dorsal margin of sicula bears spines. Family monograptidae Lapworth, 1873 (incorporates Cyrtograptidae); thl' arises via sinus in metasicula, monoserial. Family glyptograptidae Mitchell, in press (includes Glyptograptus s.s., Climacograptus brevis group); type B proximal pattern (or modified to types H or I). Family diplograptidae Lapworth, 1873 (emend.) (includes Ortliograptus and Diplograptus S.S., Climacograptus typicalis group, ‘’Glyptograptus' austrodentatus group); type A proxi- mal pattern (or modified to types G or F). Family dicranograptidae Lapworth, 1873 (emend.) (includes Nemagraptidae Lapworth, 1 873); simple sicula with dorsal sinus, partial or complete monoseriality. [In the alternative cladogram, text-fig. 9, the family alone comprises the Superfamily Dicranograptacea.] Family pseudoclimacograptidae Mitchell, in press (includes Pseudoclimacograptus s.s., Dicaulograptus)’. type C proximal pattern (or modified to types D or E). FORTEY AND COOPER: GRAPTOLOID CLASSIFICATION 653 Stem group phyllograptidae: characters of family. Family phyllograptidae Lapworth, 1873 (emend. Cooper and Fortey 1982): isograptid development type, simple (dichograptid) thecal form, prosicular origin of thl‘. Suborder not assigned Family anisograptidae Bulman, 1950: paraphyletic group, sicula retains nema in adult stage, bithecae present, rhabdosome more or less bilaterally symmetrical, and quadriradi- ate, triradiate or biradiate. GRAPTOLOIDEA incertae sedis Group containing the pendent dichograptids. Acknowledgements. We thank Drs S. Finney, C. E. Mitchell, and Mr A. Kearsley for access to their work prior to publication, and the British Council for financial assistance to Cooper. REFERENCES BULMAN, o. M. B. 1931. South American graptolites, with special reference to the Nordenskiold collection. Ark. Zool. 22A, 1-111, 12 pis. 1932. On the graptolites prepared by Elolm. Part 1. Ibid. 24A, 1-46, 9 pis. 1936. On the graptolites prepared by Holm. Part 6. Ibid. 28A, 1-107, 4 pis. 1944-1947. Monograph of Caradoc (Balclatchie) graptolites from limestones in Eaggan Burn, Ayrshire. Palaeontogr. Soc. [Monogr.], 78 pp., 10 pis. 1949. The anatomy and classification of the graptolites. Int. Congr. Zook 13 (9), 529-535. 1950. Graptolites from the Dictyonema Shales of Quebec. Q. Jl geol. Soc. Lond. 106, 63-99, pis. 4-8. 1953. Some graptolites from the Ogygiocaris Series of the Oslo District. Ark. Miner. Geol. 1, 509-518, 2 pis. 1955. Treatise on invertebrate paleontology. Part V. Graptolitbina, xvii+101 pp. Geological Society of America and University of Kansas Press, New York and Eawrence, Kansas. 1960. Some morphologically intermediate genera in graptolite phytogeny. Int. geol. Congr. 21 (22), 65-70. 1963fl. The evolution and classification of the Graptoloidea. J. geol. Soc. 119, 401-418. 1963Z?. On Glvptograptiis dentatus (Brongniart) and some allied species. Palaeontology, 6, 665-689, pis. 96, 97. 1970. Treatise on invertebrate paleontology. Part V. Graptolitbina (revised), xxxii+163 pp. Geological Society of America and University of Kansas Press, New York and Eawrence, Kansas. COOPER, R. A. 1973. Taxonomy and evolution of Isograptus Moberg in Australasia. Palaeontology, 16, 45-115. and FORTEY, R. A. 1982. The Ordovician graptolites of Spitsbergen. Bull. Br. Mas. nat. Hist. (Geol.), 36, 157-302, 6 pis. 1983. Development of the graptoloid rhabdosome. Alcberinga, 7, 201-221. and LiNDHOLM, K. 1985. The phylogenetic relationships of the graptolites Tetragraptiis pbyllograptoides and Pseudopbyllograptus cor. Geol. For. Stockb. Forb. 106, 279-291. and Ni YUNAN. 1986. The manubriate isograptids. Palaeontology, 29, 313-363, pis. 24-27. CROWTHER, p. R. 1981. The fine structure of graptolite periderm. Spec. Pap. Palaeont. 26, 1-119. ERDTMANN, B.-D. 1982. A reorganisation and proposed phylogenetic classification of planktic Tremadoc (early Ordovician) dendroid graptolites. Norsk Geol. Tiddskr. 62, 121-144. FINNEY, s. c. 1985. Nemagraptid graptolites from the Middle Ordovician Athens Shale, Alabama. J. Paleont. 59, 1100-1137. FORTEY, R. A. 1983. Geometrical constraints in the construction of graptolite stipes. Paleobiology, 9, 1 16-125. Landing, e. and skevington, d. 1982. Cambrian-Ordovician boundary sections in the Cow Head Group, western Newfoundland. Pp. 95-129. In bassett, m. g. and dean, w. t. (eds.). The Cambrian-Ordovician boundary: sections, fossil distributions and correlations. National Museum of Wales, Cardiff. FRECH, F. 1897. Letbaea geognostica', Theil 1, Letbaea Palaeozoica, 1 Bd., Graptolitbiden, pp. 544-684. Schwei- zerbach, Stuttgart. HARRIS, w. J. 1933. Isograptus caduceus and its allies in Victoria. Proc. R. Soc. Viet. 46, 79-1 14, pi. 1. 654 PALAEONTOLOGY, VOLUME 29 HARRIS, w. J. and thomas, d. e. 1942. Clonograptus pervelatiis sp. nov. and Goniograptus macer, T. S. Hall and some related forms. Min. geol. J. 2, 365-366. HUTT, J. 1974. The development of Clonograptus tenellus and Adelograptus hunnebergensis. Lethaia, 7, 79-92. JONES, w. D. V. and rickards, r. b. 1967 Diplograptus penna Hopkinson 1869, and its bearing on vesicular structures. PaUiont. Z. 41, 173-185. KEARSLEY, A. T. 1982. The morphology and systematics of species of Orthograptus Lapworth. Newsl. Graptolite Wkg Grp Int. Palaeont. Ass. 3, 10-13. 1985. A new phylogeny of diplograptine graptoloids, and their classification based on proximal and thecal construction. Ibid. 6, 8-22. KOZLOwsRi, R. 1956. Nouvelles observations sur les Corynoididae (Graptolithina). Acta palaeont. pol. 1, 259-269. LAPWORTH, c. 1873. On an improved classification of the Rhabdophora. Geol. Mag. 10, 500-504, 555-560. 1880. On new British graptolites. Ann. Mag. nat. Hist. ser. 5, 5, 149-177, pis. 4, 5. MITCHELL, c. E. (in press). Classification of diplograptid graptolites. Palaeontology, 30. MU, A. T. 1950. On the evolution and classification of graptolites. Geol. Rev., Beijing, 15, 171-183. 1957. Some new or little known graptolites from the Ningkuo Shale. Acta palaeont. sin. 5, 369-437, pis. 1-8. 1958. Ahrograptus, a new graptolite genus from the Hulo Shale. Ibid. 6, 259-265, 1 pi. 1974. Evolution, classification and distribution of Graptoloidea and Graptodendroids. Scientia sin. 17 (2), 227-238. SKEViNGTON, D. 1965. Graptolites from the Ontikan Limestones (Ordovician) of Oland, Sweden. II. Grapto- loidea and Graptovermida. Bull. geol. fnstn Univ. Upsala, 43, 1-74. SKWARKO, s. 1967. Some Ordovician graptolites from the Canning Basin, western Australia. 1 . On the structure of Didvmograptus artus Elies & Wood. Bull. Bur. Miner. Resour. Geol. Geopliys. Aust. 92, 171-183, pis. 21-23.' STRACHAN, I. 1985. The significance of the proximal end of Crvptograptus tricornis (Carruthers) (Graptolithina). Cco/. Mug. 122, 151-155. VANDENBERG, A. H. M and STEWART, I. R. 1983. Excursioii to Devilbend Quarry and Enoch’s Point. Newsl. Spec. Grp Palaeont. Biostratigr. geol. Soc. Aust. 12, fig. 9. WALKER, M. 1953. The development of a diplograptid from the Platteville Limestone. Geol. Mag. 90, 1-16. RICHARD A. FORTEY British Museum (Natural History) Cromwell Road London SW7 5BD ROGER A. COOPER Typescript received 11 December 1985 Revised typescript received 12 March 1986 New Zealand Geological Survey PO Box 30368 Lower Hutt New Zealand RANDOMNESS AND DIVERSIFICATION IN THE PHANEROZOIC: A SIMULATION by ANTONI HOFFMAN and EUGENE J. FENSTER Abstract. The essential features of the pattern of marine animal diversification in the Phanerozoic are reproduced by a stochastic, quasi-random simulation. The model assumes only (i) random variation in the origination and extinction of lineages and (ii) occurrence of two extraordinary events: ‘Late Cambro- Ordovician radiation’ and ‘Late Permo-Triassic extinction’. It can mimic the shape of the curve of global diversity as well as the pattern of so-called evolutionary faunas. The pattern and process of change in global taxonomic diversity during the Phanerozoic have long been among the focal points of palaeobiological research. Sepkoski (1978, 1979, 1984) and Sepkoski and Sheehan (1983) developed a multiphase logistic model to account for the empirical pattern of marine animal diversification (Sepkoski et al. 1981; Sepkoski 1982). The model assumes diversity dependence of the processes of origination and extinction of taxa, which should bring the biosphere to a dynamic equilibrium. It envisages three large groups of marine animal classes, or evolutionary faunas, as identified by factor analysis of their diversity histories (Sepkoski 1981). Each evolutionary fauna has its own characteristic parameters of diversity dependence, but each fauna also responds to the total diversity and thus interacts with the other two faunas. When this purely biological model is complemented with mass extinctions as extrinsic perturbations, it fits the empirical pattern quite well. A similar model has also been developed for plant diversification (Knoll et al. 1979, 1984; Niklas et al. 1980, 1983). Kitchell and Carr (1985), in turn, proposed a nonequilibrium model of diversification. Their model is also based on the assumption of converse diversity dependence of the rates of origination and extinction within each evolutionary fauna, but the system is continually kept away from equilibrium by evolutionary innovations and mass extinctions. This model also fits the empirical pattern. The assumption of diversity dependence, however, was questioned by Hoffman (1981, 1985a, b), Cracraft (1982, 1985), and Walker and Valentine (1984), on both theoretical and empirical grounds, and Hoffman (in press a, b) suggested a double random walk model of diversification. This model assumes that the average probabilities of speciation and species extinction per time interval behave as two random walks, i.e. vary at random and independently from each other; the average prob- ability of speciation over the entire Phanerozoic, however, exceeds the average probability of species extinction. Given all the biases of the fossil record (Raup 1972; Signor 1982), this double random walk model cannot be rejected as a null hypothesis when tested against the empirical pattern of global diversity of marine animal families (Hoffman in press a). In this paper, we wish to address the question of whether this double random walk model can account for both the shape of the curve of global diversity and the pattern of three evolutionary faunas. SIMULATION The essential features of the empirical pattern of marine animal diversity in the Phanerozoic are: a relatively low diversity in the Cambrian, a rapid increase in the Late Cambrian-Ordovician and fluctuating diversity through the remainder of the Palaeozoic, a drop in diversity in the Late Permian- Triassic, and an increasing diversity through the Mesozoic and Cainozoic (Sepkoski et al. 1981). IPalaeontology, Vol. 29, Part 4, 1986, pp. 655-663.| 656 PALAEONTOLOGY, VOLUME 29 0 10 20 30 40 50 TIME TEXT-FIG. 1. An example of random clade generation by Raup and Schopf (1978), as employed in our simulation. The number of lineages at each point in time gives the measure of clade diversity. Time is expressed in computer steps. Given the biases of the fossil record, it is doubtful if a more detailed description of this pattern would be robust. The pattern is a net result of changes in diversity of the three evolutionary faunas (Sepkoski 1981, 1984). The Cambrian fauna achieved maximum diversity in the Middle and Late Cambrian and then gradually declined. The Palaeozoic fauna appeared in the Cambrian, rapidly proliferated in the Early and Middle Ordovician, fluctuated without any clear trend through the remainder of the Palaeozoic, suffered heavily from the Late Permo-Triassic extinction, and again fluctuated but at a much lower level in the Mesozoic and Cainozoic. The Modern fauna also appeared in the Cambrian but it continually increased in diversity throughout the Phanerozoic. It is this total pattern that needs to be reproduced by any model of diversification. We considered seventy-five random clades (like the one illustrated in text-fig. 1). Fifty clades were taken from Raup and Schopf (1978), who generated them under the assumption that at each of the fifty time steps each lineage has a 01 probability of branching, a OT probability of being terminated, and a 0-8 probability of persisting without any change. With use of a table of random numbers, twenty-five additional clades were produced by random coupling of these clades into pairs. Sixty of these seventy-five clades were then distributed at random along the fifty-step time axis. Thus, some long and thick clades could be reduced to minute clades if they happened to be located very close to the end of the time axis. This procedure eliminated any possible intercorrelation between the clades produced by random coupling and the original clades. The continual addition of new clades ensured that, although the probabilities of lineage origination and termination in each clade were equal and constant, the average probability of lineage origination in the entire system was greater than the average probability of lineage termination. This system represents an example of the double random walk model. Biologically it implies that the average probabilities of speciation and species extinction are each determined by a myriad of biotic and abiotic factors and, therefore, change at random from one time step to another. Not surprisingly, the lineage number (standing diversity) in the system increases without showing the characteristic features of the empirical pattern of marine animal diversity. The picture changes substantially, however, when this model is supplemented with two extraordinary events representing a kind of intervention function. Fifteen randomly chosen clades were added to the system at time HOFFMAN AND FENSTER: PHANEROZOIC DIVERSIFICATION 657 "€/e” TEXT-FIG. 2. A simulated system of randomly distributed random clades affected by the ‘Late Cambro-Ordovician radiation’ and ‘Late Permo- Triassic extinction’ (see text for explanation). Time is expressed in fifty artificial steps. steps 14, 15, and 16— to model the Late Cambro-Ordovician radiation; ten randomly chosen clades were terminated at time step 33— to model the Late Permo-Triassic extinction. One such simulation is presented in text-fig. 2. This quasi-random system corresponds to the double random walk model perturbed by two events representative of a different class of random phenomena. This is illustrated by the plot of per lineage rates of origination and extinction per time step (text-fig. 3), which shows two random 658 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 3. Per lineage origination (open circles) and extinction (closed circles) of lineages per time step in the system illustrated in text-fig. 2. TIME TEXT-FIG. 4. Total diversity (left-hand scale) of the system illustrated in text-fig. 2 (solid line) compared to the empirical pattern (dashed line) of marine animal family diversity (right-hand scale). Time is expressed in fifty artificial steps. DIVERSITY DIVERSITY HOFFMAN AND FENSTER: PHANEROZOIC DIVERSIFICATION 659 TEXT-FIG. 5. Total diversity of the system in six other simulations, additional to that illustrated in text-fig. 4. 660 PALAEONTOLOGY, VOLUME 29 walks, each of them with an outlier. The outliers represent historical events extrinsic to the system- mass extinctions in the analyses of Sepkoski (1984) and Kitchell and Carr (1985). Total diversity of this system is shown in text-fig. 4. It is compared to the actual empirical pattern of marine animal diversity which, for the sake of comparability, is plotted here to the same scale and against an artificial time scale of fifty steps, rather than against absolute time in millions of years. Apart from the timing and magnitude of the ‘Late Cambro-Ordovician radiation’ and ‘Late Permo-Triassic extinction— which we determined arbitrarily, without much effort to tune them to the actual events — the empirical pattern does not appear to be clearly different from the simulated one. If these events were shifted five to seven steps back in time, as is needed to fit the actual events more precisely, the two patterns would match even better. Our other simulations produce, in fact, fairly similar patterns of total diversity (text-fig. 5). The difference between the shape of the empirical curve shown by the dashed line in text-fig. 4 and the pattern shown, for instance, by Sepkoski (1984, fig. 1) reflects the effect of time scale. Our simulation assumes constant time steps. Sepkoski’s empirical pattern is here plotted against the same time scale. We thus regard the absolute duration of the actual geological time intervals employed for compilation of the empirical data as one of the factors which contribute to the randomness of the entire system. Alternatively, the results of our simulation could be plotted against an absolute time scale; the pattern generated by the simulation would then also conform to the empirical pattern. We conducted factor analyses of the system illustrated in text-fig. 2. In Q-mode analysis, which had also been used by Flessa and Imbrie (1973) and Sepkoski (1981), the first three factors accounted for 61-2%, 16-0%, and 7-8% of the total variance respectively (85% jointly). In R-mode analysis, the first three factors accounted for 32%, 16-2%, and 11 -2% of the total variance respectively (about 59% jointly). All clades were classified into three groups according to their loadings in R-mode TEXT-FIG. 6. Scattergram of clade loadings on the first two factors of the system illustrated in text-fig. 2. Differential graphic symbols denote three groups of clades which are comparable to ‘evolutionary faunas’. HOFFMAN AND FENSTER: PHANEROZOIC DIVERSIFICATION 661 "C/e" TIME TEXT-FIG. 7. Total diversities of three groups of clades identified by factor analysis of the system illustrated in text-fig. 2. analysis (text-fig. 6), and the total diversities of these three groups of clades were plotted against time (text-fig. 7). Clearly, the smallest group includes the clades which attained their maximum diversity very early in the history of the system, generally prior to the ‘Late Cambro-Ordovician radiation’. The second group comprises the clades which expanded prior to, and suffered from, the ‘Late Permo-Triassic extinction’. The largest group includes the clades which were not affected by the ‘Late Permo- Triassic extinction’ and expanded chiefly after that event. This pattern strongly resembles the pattern of three evolutionary faunas identified by factor analysis of the actual data on marine animal diversity. Our groups of clades, however, do not show any diversity dependence and do not interact with each other. R-mode factor analysis of our other simulations resulted in the first three factors accounting for 40-60% of the total variance (Q-mode analysis generally accounted for 20-30% more), and the grouping of clades was also primarily determined by the relationship of individual clades to the ‘radiation’ and ‘mass extinction’ events. DISCUSSION These results suggest that the essential features of the empirical pattern of marine animal diversifica- tion in the Phanerozoic can be mimicked by a double random walk model perturbed by one ‘radiation’ and one ‘mass extinction’ event. The difference between our simulation and Sepkoski’s (1981) factor analysis of the actual data is remarkably small. In Q-mode analysis, the first three factors extracted from our quasi-random data account for approximately 85 % of the total variance. The first three factors of the actual data on Phanerozoic diversification account for about 91 % of the total variance. This small disparity may not suggest any major difference in the degree of nonrandomness between our simulation and family diversity patterns in the Phanerozoic. 662 PALAEONTOLOGY, VOLUME 29 In our simulation, we may encounter a scaling problem because our randomly generated clades are often considerably smaller than the actual classes analysed by Sepkoski (1981). A similar problem was pointed out by Stanley et al. (1981) who discussed the effect of unrealistic probability distributions and clade sizes on the shape of the random clades generated by Gould et al. (1977) under the assumption of equilibrium diversity. This effect, however, does not appertain to our results for two reasons: first, the clades we considered had been generated by Raup and Schopf (1978) without the equilibrium constraint; and second, the structure revealed by factor analysis in our system of randomly distributed clades depends primarily upon the extrinsic perturbations rather than upon details of the shape of individual clades. Of course, the fact that our simulation can reproduce the essential features of the empirical pattern does not demonstrate the validity of the double random walk model of diversification. It shows only that this stochastic model may provide an adequate explanation and, therefore, that analysis of these features of the pattern cannot point to the process of diversification. Much more detailed pattern analysis would be needed to this end. Given all the biases of the fossil record, however, and the uncertainty about the absolute calibration of the geological time scale, a detailed pattern analysis might give more weight to statistical noise than to real evolutionary phenomena. The empirical pattern of marine animal diversification in the Phanerozoic, and the underlying pattern of three evolutionary faunas, may reflect a myriad of independent biological factors, environ- mental events, and biases of the fossil record rather than an orderly process. Acknowledgements. We thank Jack Sepkoski for his critical comments on an earlier version of the manuscript. This study was completed during Hoffman’s tenure of NSF grant BSR 84-13605. REFERENCES CRACRAFT, J. 1982. A nonequilibrium theory for the rate-control of speciation and extinction and the origin of macroevolutionary patterns. Syst. Zool. 31, 348-365. 1985. Biological diversification and its causes. Ann. Missouri hot. Gdn, 72, 794-822. FLESSA, K. w. and IMBRIE, J. 1973. Evolutionary pulsations: evidence from Phanerozoic diversity patterns. In TARLING, E. H. and RUNCORN, s. K. (eds.). Implications of continental drift to the earth sciences, 247-285. Academic Press, London. GOULD, s. J., RAUP, D. M., SEPKOSKI, J. J., SCHOPF, T. J. M. and siMBERLOFF, D. s. 1977. The shape of evolution: a comparison of real and random clades. Paleohiol. 3, 23-40. HOFFMAN, A. 1981. Stochastic versus deterministic approach to palaeontology: the question of scaling or metaphysics? Neues Jb. Geol. Palaont. Abh. 162, 80-96. 1985«. Biotic diversification in the Phanerozoic: diversity independence. Palaeontology, 28, 387-391. 19856. Island biogeography and paleobiology: in search for evolutionary equilibria. Biol. Rev. 60, 455-471. (in press a). Neutral model of Phanerozoic diversification: implications for macroevolution. Neues Jb. Geol. Palaont. Abh. (in press b). Neutral model of taxonomic diversification in the Phanerozoic: a methodological discussion. In NiTECKi, M. H. and hoffman, a. (eds.). Neutral models in biology. KiTCHELL, J. A. and CARR, T. R. 1985. Nonequilibrium model of diversification: faunal turnover dynamics. In VALENTINE, J. w. (cd.). Phatierozoic diversity patterns: profiles in macroevolution, 277-309. Princeton Univer- sity Press, Princeton. KNOLL, A. H., NiKLAS, K. J. and TiFFNEY, B. H. 1979. Phanerozoic land plant diversity in North America. Science, N.Y. 206, 1400-1402. GENSEL, p. G. and TIFFNEY, B. H. 1984. Character diversification and patterns of evolution in early vascular plants. Paleobiol. 10, 34-47. NIKLAS, K. J., TIFFNEY, B. H. and KNOLL, A. H. 1980. Apparent changes in the diversity of fossil plants: a preliminary assessment. Evolut. Biot. 12, 1-89. 1983. Patterns in vascular land plant diversification: a statistical analysis at the species level. Nature, Land. 303, 614-616. RAUP, D. M. 1972. Taxonomic diversity during the Phanerozoic. Science, N.Y. 177, 1065-1071. HOFFMAN AND FENSTER: PHANEROZOIC DIVERSIFICATION 663 and SCHOPF, t. j. m. 1978. Stochastic models in paleontology: a primer. Workshop on 'Species as Particles in Space and Time’, Smithsonian Institution, Washington. SEPKOSKi, j. j. 1978. A kinetic model of Phanerozoic taxonomic diversity. I. Analysis of marine orders. Paleobiol. 4, 223-251. 1979. A kinetic model of Phanerozoic taxonomic diversity. II. Early Phanerozoic families and multiple equilibria. Ibid. 5, 222-251. 1981. A factor analytic description of the Phanerozoic marine fossil record. Ibid. 7, 36-53. 1982. A compendium of fossil marine families. Milwaukee Publ. Mas. Geol. Biol. 51, 1-125. 1984. A kinetic model of Phanerozoic taxonomic diversity. III. Post-Paleozoic families and mass extinc- tions. Paleobiol. 10, 246-267. BAMBACH, R. K., RAUP, D. M. and VALENTINE, J. w. 1981. Phanerozoic marine diversity and the fossil record. Nature, Land. 293, 435-437. and SHEEHAN, p. M. 1983. Diversification, faunal change, and community replacement during the Ordovi- cian radiations. In tevesz, m. j. s. and mccall, p. l. (eds.). Biotic interactions in Recent and fossil benthic communities, 673-717. Plenum Press, New York. SIGNOR, p. w. 1978. Species richness in the Phanerozoic: an investigation of sampling effects. Paleobiol. 4, 394- 406. STANLEY, s. M., SIGNOR, p. w., LiDGARD, s. and KARR, A. F. 1981. Natural clades differ from ‘random’ clades: simulations and analyses. Ibid. 7, 1 15-126. WALKER, T. D. and VALENTINE, J. w. 1984. Equilibrium models of evolutionary species diversity and the number of empty niches. Am. Nat. 124, 887-899. ANTONI HOFFMAN Lamont-Doherty Geological Observatory Columbia University Palisades, New York 10964 U.S.A. (Present address; Wiejska 14 m. 8 PL-00-490 Warszawa Poland) Typescript received 20 January 1986 Revised typescript received 4 March 1986 EUGENE J. FENSTER Queens College Biology Department City University of New York Flushing, New York 1 1367 U.S.A. GROWTH RINGS IN CRETACEOUS AND TERTIARY WOOD FROM ANTARCTICA AND THEIR PALAEOCLIMATIC IMPLICATIONS by JANE E. FRANCIS Abstract. Although the Antarctic Peninsula now has a glacial climate, during the Cretaceous and early Tertiary it was sufficiently warm for forests to thrive, even at palaeolatitudes of 59°-62° S. The forests grew on an emergent volcanic arc and the wood was subsequently buried in fluvial and basinal sediments on the margins of the back-arc basin. The forests were composed mainly of podocarp and araucarian conifers. By the late Cretaceous, angiosperm trees were also present, particularly Nothofagus, forming the characteristic forest association of the southern hemisphere today. The growth rings in the fossil wood are wide and extremely uniform, indicating that the environment was very favourable for tree growth. By comparison with living forest trees with similar growth characteristics, a warm to cool-temperate climate is proposed for the Antarctic Peninsula in the Cretaceous and early Tertiary. Features of fossil floral assemblages and sedimentary rocks are also indicative of this type of climate. An increase in the level of atmospheric carbon dioxide is considered the most likely cause of the warm polar climate at this stage. Although the Antarctic Peninsula (text-fig. 1) now has a glacial climate during the Cretaceous and early Tertiary it was sufficiently warm and temperate to allow the growth of fairly diverse plant communities. These were principally composed of conifers, cycadophytes, and ferns, with angiosperms appearing in the late Cretaceous (Dusen 1908; Gothan 1908; Halle 1913; Plumstead 1962; Jefferson 1982, 1983). Structurally preserved wood is not only an important component of these floral assemblages (Gothan 1908; Jefferson 1982) but is also present in both marine and non- marine sediments in which other plant material is absent. Certain aspects of the flora, such as the overall composition, leaf morphology and size, can give some indication of the tolerance of the plants to climatic extremes, particularly temperatures. However, the growth rings in the fossil wood preserve a more detailed record of the environment from which not only temperature constraints but also seasonal variations in rainfall can be deter- mined in greater detail. Growth ring analyses have been used to evaluate changing geological climates on a broad scale (Chaloner and Creber 1973; Creber and Chaloner 1984n, h) and to investigate individual fossil forest environments, e.g. Cretaceous forests of Antarctica (Jefferson 1982) and the Upper Jurassic/Lower Cretaceous forests of southern England (Francis 1984, in press). In this project a collection of Cretaceous and early Tertiary fossil wood from various localities around the northern part of the Antarctic Peninsula was studied to determine the botanical composition of the fossil forests and the nature of the palaeoclimate. The results of the growth ring analyses are presented and their palaeoenvironmental significance discussed. This complements a similar study by Jefferson (1982, 1983) on higher latitude early Cretaceous fossil forests from Alexander Island. FOSSIL FORESTS AT HIGH LATITUDES The fossil remains of plant assemblages that grew very near to both the North and South poles in the Cretaceous and early Tertiary provide some of the main evidence that polar climates were then much warmer than at present. Their presence has provoked much discussion since plant growth at IPalaeontology, Vol. 29, Part 4, 1986, pp. 665-684, pi. 51. | 666 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1 . Map of the northern Antarctic Peninsula region showing localities, a, location of Byers Peninsula, Livingston Island (see inset C) and the northern part of the Antarctic Peninsula, enlarged in inset B. b, fossil wood localities in the northern Antarctic Peninsula region. CV = Cross Valley, SV = Sharp Valley, WB = Whisky Bay. • D.8218.12 = fossil locality and number, c, location of fossil localities on Byers Peninsula, western Livingston Island, South Shetland Islands. such high latitudes today is restricted by very low temperatures (Billings 1964; Grime 1979) and to some extent the low angle of the sunlight of the polar light regime (Wolfe 1980). To provide a more even annual distribution of light, a reduction in the tilt of the earth’s axis of rotation has been proposed for Cretaceous and early Tertiary times (Wolfe 1978, 1980; Douglas and Williams 1982; Jefferson 1982, 1983). There are, however, several arguments against such a proposal (Donn 1982; Axelrod 1984; Creber and Chaloner 1985). In particular, the clearly marked growth rings in fossil wood from the south polar region (Jefferson 1982; Douglas and Williams 1982; this study) and the Arctic region (Smiley 1967; Schweitzer 1980) are evidence that the climate had well-defined seasons, whereas the reduced axial tilt would produce a seasonless, more uniform environment. Furthermore, no plausible physical mechanism for actually changing the axial tilt has been proposed (Creber and Chaloner 1985). The climate model simulations of Barron (1984) have shown that a reduced obliquity would in fact result in lower polar temperatures due to a decrease in mean annual insolation at high latitudes. That the plants (and animals) were able to adapt physiologically to long periods of winter darkness was presented as an alternative solution by Axelrod (1984) and Creber and Chaloner FRANCIS: CRETACEOUS AND TERTIARY WOOD 667 (1984o, b, 1985). Although the annual distribution is not uniform, the total light energy flux for polar regions today is equivalent to that at more temperate latitudes. Thus the high-latitude trees may well have been photoperiodic ecotypes (as some are today), specifically adapted to growing at faster rates in short seasons with long hours of daylight (Vaartaja 1962; Axelrod 1984; Creber and Chaloner 1985). Other adaptions include suitable tree densities and conical growth habits to inter- cept the maximum amount of solar radiation with minimum mutual shading (Creber and Chaloner 1984fl, h\ Creber, in press). Axelrod (1984, p. 128) suggests that the trees ‘survived the winter darkness in polar regions by a general metabolic shutdown’. The warm Cretaceous/Tertiary polar temperatures, as indicated by the plants and animals, were part of the warm equable climate with a low global temperature gradient that was characteristic of the Mesozoic (brakes 1979; Hallam 1985). The cause of the polar warming has been investigated using climatic simulation models based on Cretaceous continental distribution (Barron et al. 1981; Barron and Washington 1982, 1984; Barron 1983). Their results indicated that the proximity of the continental masses would result in an overall global warming with a warmer low latitude belt, correlating well with evidence from fossil tree-ring data collated by Creber and Chaloner (1984a, h, 1985). However, the polar regions would have remained cold. An increase in the CO2 level of the atmosphere was proposed by Berner et al. (1983) and Barron and Washington (1984) to be the additional mechanism required to explain the palaeoclimatic data. Climate models based on modern palaeogeography illustrate that an increase in atmospheric CO2 would result in an increase in the poleward transport of latent heat. This would cause a reduction in the equator-pole temperature gradient, much like that proposed for the Mesozoic (Manabe and Wetherald 1975, 1980). Further- more, the temperatures at the poles would increase as would the rates of evaporation and precipi- tation. Since CO2 is important for photosynthesis, higher levels would have also raised photosynthetic rates and increased plant productivity. As shown by Creber and Chaloner (1985, tables IV and V), tree species in particular show large increases in productivity under higher levels of CO2. Changes in growth rates exhibited by modern tree-rings were also attributed to recent changes in the level of atmospheric CO2 by LaMarche et al. (1984). Thus higher levels of atmospheric CO2 in the Cretaceous and early Tertiary would not only have raised temperatures at high latitudes to levels allowing plant growth, but plant productivity would also have been enhanced, possibly accounting for the wide growth rings in the fossil wood described here. ANTARCTIC FOSSIL WOOD Specimens used in this growth ring analysis are housed in the collections of the British Antarctic Survey, Cambridge and the British Museum (Natural History). The wood was collected in past field expeditions including Operation Tabarin (1943-1945), Falkland Islands Dependencies Survey 1945-1947 (collections mainly by W. N. Croft), and during more recent field surveys by the British Antarctic Survey. The wood ranges from early Cretaceous (Barremian) from the South Shetland Islands (text-fig. 1) to Eocene/early Oligocene from Seymour Island. Specimen numbers, locality, and stratigraphic information are given in text-fig. 1 and Table 1. Both conifer and angiosperm woods are represented, the angiosperms occurring in late Cretaceous (Campan- ian of the Naze) and younger rocks. The conifers are mainly of podocarp and araucarian affinities, forming assemblages similar to the typical southern hemisphere conifer forests of today. The most common fossil angiosperm wood is that of Nothofagm (southern beech), the living relatives of which are evergreen and deciduous broad-leaved beeches found today in South America and Australasia (Darlington 1965). The fossil wood specimens range in size from small twigs to large sections over 25 cm in diameter. However, these represent both small branches with central pith and outer bark still attached and rings of small radius of curvature, and also sections of outer parts of large trunks. Most of the larger samples consist only of the heartwood (the inner secondary xylem) so the maximum size of the trunk is unknown. An estimate of the size of the trees from which the samples originated can be obtained from the degree of curvature of the growth rings. Many of the specimens have almost ‘flat’ rings (PI. 51, figs. 2, 5) indicating that they came from outer parts of large trees. In particular, most angiosperm woods have rings of large radius of curvature from trees with trunks or branches at least 30-40 cm in diameter. These were clearly 668 PALAEONTOLOGY, VOLUME 29 TABLE 1 . Locality and stratigraphic information for fossil wood samples used for tree-ring analysis. Localities are shown in text-fig. 1. Locality and sample infor- Lithology/Environment of deposition Probable age mation intra-arc: Byers Peninsula, western Livingston Island, South Shetland Islands: P.1728.9 (Crame and Far- quharson 1984) and BR.80.8 from tuff of Vol- canic Member south of Chester Cone. P.245.38 from surface on west coast, probably from Volcanic Member. Mount Flora, Hope Bay, Trinity Peninsula: D.20.3, D.20.4, D.20.10, D.20.16, D.48.2, D.48.3, D.48.5, D.48.7. Tower Peak, Trinity Pen- insula: R.1334.12 BACK-ARC: Western James Ross Is- land: D. 83 11.5 from Sharp Val- ley. D. 833 1.4 from south of Whisky Bay. Lachman Crags and the Naze, James Ross Island: Lachman Crags. D.42I, 4774, 4776, 8731. The Naze. D.90, D.87.3, 5056, 5059, 5060. Intra-arc sequence of deep marine mudstones (Mudstone Member), marine volcaniclastics (Mixed Marine Member), and terrestrial pyroclastic rocks (Volcanic Member), form- ing the Byers Formation. A basinal environ- ment envisaged by Smellie et a], (1980), with pyroclastic material from local vents being deposited in small lakes and over forested land. Main wood-bearing strata belong to Volcanic Member, which also contains leaf impressions. Fossil tree stump preserved in growth position. An extensive flora preserved within the Mount Flora Formation, Botany Bay Group (Farquharson, 1984). Non-marine conglom- erates and mudstones deposited on alluvial fans and braided river plains on emergent arc. Plant remains buried in fine sediment in ephemeral lakes but grew locally, as shown by coals, rootlet horizons, and an in situ tree stump at Camp Hill (text-fig. 1) (Farquhar- son 1982, 1984). Tower Peak Formation, Botany Bay Group. Sedimentary setting and age similar to that at Hope Bay (Farquharson 1984). Lower Cretaceous rocks originated as domi- nantly coarse-grained volcaniclastic detritus deposited across faulted margins of back- arc basin (Bibby 1966; Farquharson 1984; Ineson 1985). Divided by Ineson et al. (1986) into Gustav Group for lower coarser units and Marambio Group for upper finer units (= Snow Hill Island Series of Bibby 1966). Calcified wood present throughout Gustav Group, having originated from forests on volcanic arc to the west. Sedimentary strata consist of poorly lithified sands with intercalated clays and nodules of calcareous, glauconitic sandstone (Bibby 1966). Much fossil wood collected from within sandy clays and nodules at both localities by Croft (1947). Angiosperm leaves also found at Lachman Crags. Strata belong to lower part of Snow Hill Island Series (Bibby 1966) = Lopez de Bertodano Forma- tion, Marambio Group of Rinaldi (1982). Volcanic Member flora con- sidered ‘Wealden’ in age (Araya and Herve 1966), possibly Bar- remian (Hernandez and Azca- rate 1971). A ?Valanginian- Barremian-Hauterivian age was proposed by Askin (1983n) on palynological evidence. (K-Ar dating of intrusive rocks gives a minimum age of Cenomanian, Smellie et al. 1984.) Flora of ferns, cycadophytes and conifers originally con- sidered of Middle Jurassic age (Nathorst 1904; Halle 1913) or early Jurassic (Rao 1953). Latest floral and stratigraphic evidence suggests early Cretaceous age (Stipanicic and Bonetti 1970; Taylor et al. 1979; Farquharson 1983, 1984). D.83 1 1 .5. Upper part of Kotick Point Formation of Aptian/ Albian age (Ineson et al. 1986). D. 8331. 4. Base of Marambio Group (lower part of Snow Hill Island Series) of late Cretaceous age, probably post Santonian. Ammonite faunas from both localities considered to be upper Lower to Middle Campanian in age (Spath 1953; Howarth 1958, 1966), even possibly Maastrichtian (Henderson and McNamara 1985). Crame (1983) proposed pre-Campan- ian age for lower part on basis of Inoceranms. Palynological evidence suggests a ?Senonian- Campanian age for the Naze (Askin 19836). FRANCIS: CRETACEOUS AND TERTIARY WOOD 669 Locality and sample infor- Lithology/Environment of deposition Probable age mation Cape Lamb, Vega Island: D.8212.12 Sedimentary sequence at Cape Lamb similar to that at Lachman Crags and on the Naze (Bibby 1966). Seymour Island: D.495 from Cross Valley (Sobral Formation). D.494, D.509, and 4388 from Cross Valley Forma- tion. D.502 and D.8321.4 from northern part of island, in La Meseta Formation. In the southern part of the island large logs and plant remains occur within concretion- ary sandstones and siltstones of the upper part of the Lopez de Bertodano Formation (= Older Seymour Island Beds of Andersson 1906). Many leaf assemblages and logs, some up to 1 metre diameter (Zinsmeister 1982) in the Tertiary Sobral and Cross Valley Formations in south-east and Cross Valley. Strata represent distributary channel fills and interdistributary flats of delta, with river systems draining vegetated highlands to west (Elliot et al. 1975). The overlying La Meseta Formation represents shallowing deltaic- lagoonal environment as back-arc basin gradually filled (Elliot and Trautman 1982). Askin (1983/?) proposed a late Campanian-?early Maastrich- tian age and suggests strata here slightly younger than those at the Naze. Lopez de Bertodano Formation is late Campanian-Maastrich- tian in age (Rinaldi et al. 1978; Elliot and Trautman 1982; Zinsmeister 1982); upper part may even be Danian (Askin 1985). A Palaeocene age given for Cross Valley Formation by microfossils (Hall 1977; Wrenn 1985). La Meseta Formation con- sidered to be late Eocene-?early Oligocene age (Zinsmeister and Camacho 1982; Hall 1977). well-grown trees rather than weedy shrubs, the habit proposed for some of the early angiosperms that colonized unstable floodplain environments (Retallack and Dilcher 1981). Preservation. Fossil wood from the South Islands and Trinity Peninsula (including Hope Bay and Tower Peak) is preserved by chalcedony (text-fig. 1). The lignin retained within the cell walls is dark brown or black and displays only a homogeneous amorphous structure, suggesting that the wood has been subject to some degree of maturation or thermal metamorphism. At these localities the wood is buried within volcaniclastic sediments. Thus the most likely sources of the permineralizing silica is from weathered volcanic minerals, a common source for the silicification of wood (Murata 1940). In contrast, fossil wood from James Ross, Vega, and Seymour Islands (text-fig. I ) is calcified. The state of perservation is excellent; the very fine ultrastructure of the cells is still apparent in most cases, particularly in the angiosperm wood. This calcified wood has been extensively bored by bivalves (boring type Teredolites; Kelly and Bromley 1984). The borings were subsequently filled with the surrounding sediment or crystalline calcite cement (PI. 51, fig- L text-fig. 2). The wood was incorporated in the marine clastic and deltaic sediments of the back-arc basin so the calcite for permineralization most likely has a marine source. These two types of mineralization show an interesting relationship to the palaeoenvironmental setting. The silicified material is associated with the terrestrial volcanic environment on the emergent arc and the calcified wood found in the marine back-arc basin. A more refined analysis of this distribution may prove that the study of mineral types in fossil wood could be a useful tool for locating the arc/basin margin. GEOLOGICAL SETTING During the Cretaceous and early Tertiary an emergent and periodically active volcanic arc was present in the region now occupied by the Antarctic Peninsula (Thomson et al. 1983; Farquharson 1983) . Volcaniclastic sediments eroded from this arc gradually accumulated across the fault-con- trolled margins of a subsiding back-arc basin to the east (Farquharson 1983; Farquharson et al. 1984) . The sedimentary environments range from a non-marine fluvial setting across the arc (the Botany Bay Group, Farquharson 1984) to deeper marine clastic sediments within the basin (the Gustav and Marambio Groups of Ineson et al. in press). As the basin progressively filled, a deltaic 670 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 2. Calcified fossil wood from The Naze, James Ross Island, a, 5059. Wood with calcitized and sediment-filled borings, x IT. h, D.87.3. Wood with a portion of the surrounding volcaniclastic sedimentary concretion still attached on the upper surface, x 0-8. TEXT-FIG. 3. Position of Antarctica at selected times during the Cretaceous and early Tertiary, based on Lawyer el at. 1985. 135 Ma = Valanginian, 1 19 Ma = Aptian, 84 Ma = Campanian, 64 Ma = Palaeocene. FRANCIS: CRETACEOUS AND TERTIARY WOOD 671 environment developed during the early Tertiary in the region of Seymour Island to the east (text- fig. 1) (Elliot and Trautman 1982; Zinsmeister 1982). Stratigraphic details are given in Table 1. The most recent palaeogeographic reconstructions by Lawver et al. (1985) for selected times during the Cretaceous and early Tertiary place the northern tip of the Antarctic Peninsula at palaeolatitudes ranging from approximately 59° to 62° S. (text-fig. 3). The actual position of the peninsula in the reconstruction of Gondwanaland is still uncertain, particularly in relation to east Antarctica, but it remains in approximately this latitudinal range in most models (Dalziel and Elliot 1982, and references therein). GROWTH RING ANALYSIS Growth rings in the fossil wood were measured from polished slabs in transverse section viewed under a dissecting microscope, from acetate peels of acid-etched surfaces, and from petrographic thin-sections. The rings were measured along a radial line to obtain as long a ring series as preservation permitted, although readings often had to be continued along adjacent radii to avoid patches of poor cellular preservation. If the TABLE 2a. Results of growth ring analysis. (Wood types A = angiosperm, C = conifer.) Locality Specimen number Number of rings Mean ring width (mm) Maximum ring width (mm) Minimum ring width (mm) Mean sensitivity Wood type TERTIARY Seymour Island: D.502 23 L32 3-3 0-4 0-207 A D. 8321.4 11 5-70 8-4 2-5 0-214 A D.494 14 2-58 4-6 20 0-173 A D.495 27 0-52 0-7 0-4 0-139 C 4388 16 2-51 6-3 L2 0-267 C D.509 13 0-84 1-4 0-4 0-371 C CRETACEOUS Vega Island: D. 8218. 12 6 7-50 90 6-2 0-123 A The Naze: 5056 19 L37 2-4 0-7 0-177 A 5059 31 0-78 1-6 0-4 0-188 C 5060 8 2-29 2-9 1-2 0-205 C D.87.3 26 0-80 1-4 0-4 0-192 C D.90 19 L82 3-8 11 0-284 A Lachman Crags: D.421 56 0-61 L7 01 0-357 C 4774 60 L45 2-5 0-8 0-187 C 4776 1 1 2-33 3-6 T9 0-146 C 8731 9 5-28 7-8 3-6 0-181 C Sharp Valley: D. 831 1.5 12 117 1-4 10 0-125 C Whisky Bay: D. 8331.4 17 L84 2-4 11 0-199 C Hope Bay: D.20.3 15 2-71 3-4 L9 0-146 C D.20.4 12 4-92 6-7 3-6 0-178 C D.20.10 20 3-26 4-3 L9 0-137 C D.20.16 35 1-47 2-7 0-9 0-146 C D.48.2 17 3-77 4-8 1-6 0-145 C D.48.3 7 3-23 3-8 2-6 0-146 C D.48.5 20 1-48 L9 10 0-148 C D.48.7 20 1-46 2-0 TO 0-152 C Tower Peak: R. 1334. 12 9 5-57 6-9 4-0 0-218 C Byers Peninsula: BR.80.8 12 1-35 1-8 TO 0-204 C P.245.38 14 0-99 1-8 0-6 0-239 C P.1728.9 29 118 L9 0-6 0-189 C 672 PALAEONTOLOGY, VOLUME 29 TABLE 2b. Summary of growth ring analyses. Number of ring series Mean ring width Range of mean ring widths Mean sensitivity Range of mean sensitivities Total ring series 30 2-30 0-52-7-50 0-206 0-123-0-371 Tertiary 6 2-25 0-52-5-70 0-228 0-139-0-371 Cretaceous 24 2-35 0-61-7-50 0-184 0-123-0-357 centre of the branch was not present, the radius of curvature of the rings was noted to estimate whether the wood specimen was part of a small branch or from the outer part of a large trunk. The absolute ring widths were measured as a record of annual growing conditions. From these some of the standard statistical parameters used to describe modern ring characteristics (Fritts 1976) were calculated, but also taking into account the problems associated with fossil wood, such as the preservation of only a limited number of rings from an unknown part of the tree. The most imformative parameter was the variability in width of the rings from year to year. This was calculated in terms of the annual sensitivity (AS), i.e. the difference in width between a pair of consecutive rings divided by their average width (Creber 1977), which illustrates the degree of variability between years. The average of these values for each tree, the mean sensitivity (MS) (Fritts 1976), gives an indication of the tree’s response to variable factors of the climate that may have influenced its growth. The mean sensitivity is calculated using the formula MS 1 z x, + i +x, where .v is the ring width, n is the number of rings, and t is the year number of each ring. Values of mean sensitivity range from 0 where there is no variation from year to year, to a theoretical maximum of 2 representing the greatest variation. An arbitrary value of 0-3 is taken to separate ‘complacent’ trees that have grown under a favourable uniform climate (MS < 0-3) from those ‘sensitive’ to limiting factors of climate (MS > 0-3) (Fritts 1976). In selected rings with good cellular preservation, individual cell diameters along radial files were measured in order to classify the rings according to the scheme proposed for fossil wood by Creber and Chaloner (1984/), p. 371). Results A total of 30 ring series were obtained ranging in length from 7-60 rings, though series of 15-20 rings were most common. Some wood samples had the central pith present surrounded by rings of small radius of curvature representing the initial growth of the trees. However, the large radius of curvature of the rings in most specimens (PI. 51, fig. 2) illustrates that they are from the outer sections of large trunks, areas where environmental influences have a stronger signal than the inherent growth trends. Individual rings in the conifers are characterized by a relatively wide zone of large, thin-walled earlywood cells terminated by only four to five thick-walled cells of the latewood (text-fig. 4). The angiosperms also exhibit clear ring boundaries, marked by a decrease in fibre diameter and termi- nated by two to three very narrow fibres. The position of the largest vessels at the beginning of each ring in the ring-porous angiosperms also makes the rings very conspicuous (PI. 51, fig. 3). Plots of EXPLANATION OF PLATE 51 Figs. 1-6. Growth rings in Antarctic fossil wood. 1, D. 832 1.4 (angiosperm wood with borings), Seymour Island, X 1-7. 2, D.509 (conifer), Seymour Island, x 13-2. 3, D.90 (angiosperm). The Naze, James Ross Island, x8-8. 4, D. 833 1.4 (conifer). South of Whisky Bay, James Ross Island, x 51. 5, D.48.5 (conifer), Hope Bay, Trinity Peninsula, x 4 0. 6, D.48.7 (conifer), Hope Bay, Trinity Peninsula, x 3-3. PLATE 51 FRANCIS, Antarctic fossil wood 674 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 4. Growth rings in fossil conifer wood, showing well-defined ring boundaries (EW = early- wood, LW = latewood). a, 5059. The Naze, James Ross Island, x28. b, D. 8331. 4. South of Whisky Bay, James Ross Island, x 22. c, D.509. Cross Valley, Seymour Island, x 80. variations in cell dimensions throughout selected rings are shown in text-fig. 5. The curves show a very slight decrease in cell radial diameters throughout the earlywood portion of the ring followed by an abrupt transition to the narrow latewood cells at the end. These curves are characteristic of types D and E under the classification scheme for fossil growth rings proposed by Creber and Chaloner (1984/)), interpreted as representing a relatively uniform growing season followed by a terminal event due to a cessation or retardation of cambial activity. False (intra-annual) rings, which can be caused by frost or drought (Fritts 1976), and scars due to insect attack or fire, were not observed. The growth rings in the wood are very prominent (PI. 51, figs. 1-6). The average ring widths were 2-25 mm for the Tertiary woods, 2-35 mm for the Cretaceous specimens, and 2-30 mm for the whole collection of wood (Table 2b). These values indicate that, on average, growth rates were quite fast. However, a notable feature is the presence of very wide rings (e.g. 9 00 mm in D.8218.12, 8-4 mm in D. 832 1.4) in a large number of specimens. This indicates that the trees had the potential for extremely high growth rates in a very favourable environment. Some of the widest rings occur in the angiosperms. In contrast, a few specimens have consistently very narrow rings. Specimens 5059 and D.87.3 from the Naze, James Ross Island, consist of central portions of trunks so the narrow rings around the pith may simply represent initial development of the tree. However, specimens D.495 and D.509 (Cross Valley, Seymour Island) are parts of large trunks and their narrow rings may well be the result of a climatic change. These samples are Palaeocene in age and possibly reflect climatic deterioration at the Cretaceous/Tertiary boundary, although clearly many more samples are re- quired to test this theory. Whatever their width, the most conspicuous feature of the growth patterns in all the Cretaceous and Tertiary woods is the uniformity in width of the rings from year to year (PI. 51; text-fig. 6). This is reflected in their consistent mean sensitivity values. The averages are 0-206 for all the trees, 0-228 for the Tertiary, and 0-184 for Cretaceous trees. Of the thirty samples, twenty-eight had MS values less than 0-3. Most were between 0-100 and 0-199 (75 %) and some were between 0-200 and 0-299 (21 %). These trees are thus overwhelmingly ‘complacent’ and their uniform growth reflects very constant and favourable growing environments in which no single climatic factor adversely affected tree growth. Histograms of the annual sensitivity values for each tree (text-fig. 7) also FRANCIS: CRETACEOUS AND TERTIARY WOOD 675 E TEXT-FIG. 5. Graphs of variation in cell radial diameter across selected growth rings. The point at which the cumulative sum curve finally turns to zero is used to determine the earlywood (EW)/latewood (LW) boundary. illustrate that the variation in growth was very low since a high proportion of the AS values are less than 0-3, the dominant values lying between 0 and 0 099. However, each tree has a few AS values of the ‘sensitive’ kind reflecting occasional periods of more marked climatic influence. INTERPRETATION OE THE RINGS The presence of well-marked growth rings in these Antarctic woods illustrates that their growth environment was characterized by well-defined seasons. The pattern of cell dimensions throughout the rings suggests that conditions during the growing season were very suitable for cell division and expansion (Creber and Chaloner 1984/?), resulting in the large and numerous cells of the earlywood zones. However, conditions changed markedly at the end of the growing season and cell production was severely retarded. A narrow latewood zone within a ring can be the result of water shortage (Creber and Chaloner 1984/?). However, this seems an unlikely cause in such a elastic-dominated sedimentary setting. Alternatively, the growing season may have been abruptly terminated by the low light levels during the onset of the winter season. It is possible that the type of cell pattern in the rings seen here (text-fig. 5) may be a genetic 676 PALAEONTOLOGY, VOLUME 29 RING NUMBER TEXT-FIG. 6. Plots of growth ring sequences. The first 6 graphs are of rings from the northern Antarctic Peninsula region showing fairly uniform growth. In contrast, the graph of KG. 1702.6 from Alexander Island (Jefferson 1982) shows variable growth from year to year. characteristic and not so significant in terms of climatic influence. A large number of the Antarctic fossil woods are conifers of araucarian or podocarpacean affinity; these types may have annual cell production somewhat independent of environmental influence (Creber and Chaloner 1984fi). However, analyses of their ring series are still significant and living members of these families have been used for dendrochronological work in South America (LaMarche et al. \979a, b; Hughes et al. 1982). However, the same ring types D and E are present in other types of Antarctic conifers and to some extent in the angiosperms, implying that these ring patterns truly reflect environmental influence. The uniform width of the rings is the most revealing signal of the climate. Such trees with low mean sensitivity values of ‘complacent’ type grow today in stable forest environments where no single factor controls growth. Fritts et al. (1965), Fritts (1976), and LaMarche (1974) found that trees of this type grew in forest interiors where fairly uniform environments were established. This contrasts with trees growing nearer the forest borders and near the climatically determined limits of their distribution in which certain limiting factors, such as low temperatures at high altitudes or low rainfall at the lower forest border, have a very pronounced effect on the growth. This in turn is reflected in their growth patterns. The absence of false or partial rings in the Antarctic fossil wood FRANCIS: CRETACEOUS AND TERTIARY WOOD 677 Annual Sensitivity increasing sensitivity ► TEXT-FIG. 7. Histograms showing the amount of annual variation in the growth patterns of selected fossil trees. All these trees exhibit predominantly complacent growth. (Arrow indicates position of mean sensitivity.) is consistent with the observations of Fritts el al. (1965) that these ring types are absent in forest interior trees. DISCUSSION Comparison with modern forest growth. The forest trees that grew in the Antarctic Peninsula region in the Cretaceous and early Tertiary clearly enjoyed very favourable growing conditions in a mild, temperate environment. Such growth is not apparent in trees at similar latitudes (59°-62° S.) today. In the southern hemisphere the most southerly forests grow in Magellanic (southernmost) Chile, up to about 56° S., approximately 500 miles north of the Antarctic Peninsula (Skottsberg 1960; Godley 1960; Darlington 1965; Young 1972; Crow 1975). The characteristic vegetation is very dense evergreen rainforest dominated by Nothofagus betuloides, and with only a few other conifer and angiosperm genera. This evergreen forest is limited to the fjord-like topography of the western slopes of the Andes where the rainfall is extremely high (up to 7500 mm/yr in the mountains; Young 1972) and distributed evely throughout the year. Both summers and winters are cool. The mean annual temperature is about 6 °C and the annual range only about 4 °C (Young 1972). The restriction to plants (and animals) is not therefore extreme cold but the lack of warmth for growth in summer, particularly for deciduous trees (Darlington 1965). The N. hetidoides rainforest trees grow to large sizes (up to 30 m in height and 2 m in trunk diameter) at slow rates. Some are estimated to be 1500-2000 years old (Young 1972). An average increase in trunk diameter of 1-42 mm/year (mean ring width of 0-71 mm) was recorded for a selection of these trees. Average ring widths of 0-37 mm and 100 mm were recorded from the conifer Pilgerodendron uvifera and the evergreen angiosperm Drimys winteri respectively (Young 1972). These trees grow at such slow rates due to the continuously cold and wet environment. To 678 PALAEONTOLOGY, VOLUME 29 the north and east of the rainforest region a more seasonal climate with lower rainfall prevails. A transitional type of drier, less dense forest becomes dominant, characterized by deciduous species of Notliofagus (e.g. N. pumilio, N. antarctica). LaMarche et al. (\919a) have obtained tree-ring chronologies mainly from Araucaria araucana in this area. Such trees from within latitudes 37°- 43° S. in Argentina, have mean ring widths ranging from 0-38 to 3-33 mm, although the majority grow at rates of less than 1 mm/year. Values of mean sensitivity were also very low, ranging from 012 to 0-23, illustrating that growth was very uniform from year to year. In Chile, between latitudes 32° and 40° S., mean ring widths of between 0-40 to 1-75 mm were recorded (LaMarche et al. 1 979/>; Hughes ct <7/. 1982). The composition of these South American forests (and similar forests in Australasia, Godley 1960) is very similar to that of the Antarctic Peninsula fossil assemblages, as indicated by the fossil remains of Notliofagus, and members of the Podocarpaceae and Araucariaceae (Gothan 1908; Dusen 1908; Halle, 1913; Florin 1940; Cranwell 1959, 1969; Barton 1964; Orlando 1964; Plumstead 1962; Askin 1983«, h; Torres 1984). However, compared with the growth rings in the Antarctic fossil wood, those in the living trees are much narrower, recording considerably slower growth rates resulting from the lower temperatures of the cold temperate climate. However, most trees exhibit the same degree of ring uniformity as the fossils (Young 1972; LaMarche et al. 1979a, b), probably due to the constant rainfall throughout the year. The present latitudinal limit of tree growth in the south is much lower than that in the northern hemisphere due to the influence of the extensive ice sheet over the Antarctic continent. In the north, the tree line lies at approximately 66° N., almost coincident with the 10 °C isotherm for July (Creber and Chaloner 1984a, h). The composition of the northern forests at latitudes corresponding to those in Chile differs in that they consist mainly of conifers, rather than evergreen angiosperms. The conifers grow under a more seasonal climate where growth is restricted to warmer summer months. Data compiled by Creber and Chaloner (19846, Table XII, p. 429) illustrate that some angiosperms between 48°-61° N. have average growth rates of 0-5-T6 mm/yr. At even higher latitudes, beyond 75°-70° N., the ring widths are exceedingly small (e.g. 0-032 mm) in trees from east Greenland, 73° N. However, conifers are still able to produce relatively wide rings up to about 70° N. Poleward of the tree-line, tree growth is limited by very low temperatures which adversely affect the metabolic processes, resulting in such low productivity that almost no wood is produced (Billings 1964; Grime 1979; Creber and Chaloner 1984a, b, 1985). The tree-rings in the Antarctic fossil wood are more comparable with those in living trees in warm/cool-temperate forests in Australasia. The growth of selected examples of the podocarp Phyllocladus in southern New Zealand and Tasmania was considered analogous to that of the Cretaceous Alexander Island trees by Jefferson (1982, 1983); both have fairly large rings and high mean sensitivities. Some of these living trees also have low, complacent mean sensitivities more like the fossil trees from the northern Antarctic Peninsula. In Australia (25°-43° S.) LaMarche et al. (\919d ) recorded mean ring widths of 0-3 1-1 -22 mm and mean sensitivities of 0- 12-0-47, mainly in trees of the Podocarpaceae and Cupressaceae. Further north, in Queensland (17° S.), Callitris conifers are able to produce very wide rings (3-5 mm, average 2-5 mm) (Ash 1983). The rings are also very uniform in width. The climate here is much warmer and has a more tropical aspect and, although the rainfall is very high (1700-2700 mm/yr), a distinct dry season causes the formation of growth rings. Compared to the wide and uniform growth rings in the northern Antarctic Peninsula fossil wood, the early Cretaceous trees on Alexander Island had wide but very variable ring widths. Jefferson (1982, 1983) recorded very high mean sensitivity values in his fossils, characteristic of trees growing at the limits of their ecological range in marginal environments where climate strongly influences growth. Since these trees were located at higher latitudes (65-75° KrS.) in the Cretaceous than those on the northern part of the peninsula, it may be that the Alexander Island forests represent the marginal limits of the Antarctic Peninsula forests. The Cretaceous jearly Tertiary climate of the Antarctic Peninsula. The growth patterns in fossil wood illustrate that the Antarctic climate was clearly not like that of today but was much warmer and FRANCIS: CRETACEOUS AND TERTIARY WOOD 679 more comparable to warm temperate environments (Thomson 1982). Previous interpretations of fossil plant collections have reached the same conclusions. The assemblage of conifers, ferns, and cycadophytes of the early Cretaceous Hope Bay flora was considered by Halle (1913) to be part of a global Mesozoic flora (although Florin (1940) and Plumstead (1962, 1964) argued that it was rather more localized to the southern hemisphere) of warm temperate or subtropical nature. A similar climate prevailed during the growth of the ferns, conifers, and angiosperms of the Seymour Island flora (Dusen 1908). However, a rather cooler climate was envisaged by Askin (19836) for the Campanian palynomorphs from the Naze and Vega Island. The size and morphology of angiosperm leaves have also been used as criteria for determining climatic tolerance. Large leaves ( > 10 cm long) of Nothofagiis type from the Tertiary of Alexander Island were considered to be indicative of warm-temperature conditions (Thomson and Burn 1977), as were Tertiary leaves from Adelaide Island (Jefferson 1980). Mid-Tertiary plant assemblages from King George Island, South Shetland Islands (Orlando 1964), which are composed primarily of Nothofagus and Podocarpaceae, are comparable with cool/warm-teniperate rainforests of Australia on the basis of leaf morphology (Zastawniak et al. 1985). However, Barton (1964) considered that the small size of the leaves was more like that of the trees in Chilean rainforests where the climate is of cool-temperate type. The early Cretaceous (Barremian-Albian) fossil forests of Alexander Island were also considered analogous to Australian rainforests, with regard to both the wide but variable growth rings in the wood and the floral characteristics. The leaves are exceptionally large and their margins entire, typical of the ‘tropical rainforest’ types of Wolfe (1971) and those found in modern Australasian temperate rainforests (Jefferson 1983). The large lamina size suggests that the leaves would not have been tolerant of frosts. They do not exhibit xeromorphic adaptations so it would appear that water was probably in plentiful supply. Palaeoclimatic interpretations of the associated sedimentary strata support plant evidence. A warm and periodically wet climate within a temperate environment was considered by Taylor et al. (1979) to be necessary to produce the clastic sedimentation within the deltaic environments of the late Jurassic/early Cretaceous rocks of Alexander Island. Similarly, the presence of braided stream, debris flow, and sheet-flood deposits within the Botany Bay Group also implied a climate with moderate to high rainfall causing periodic flooding (Farquharson 1984). Sea water temperatures around the Antarctic were also much higher than today. Oxygen isotope data and palaeotempera- tures from benthonic and planktonic Tertiary Foraminifera show a decline in temperatures through- out the Tertiary. Bottom water temperatures of about 16°C and surface water temperatures of 19 °C were recorded for the late Palaeocene (Shackleton and Kennett 1975; Kennett 1977). However, stable isotope measurements and species distributional analysis of Foraminifera, which once in- habited a mid-outer shelf environment and are now preserved within the Cretaceous sediments of Seymour Island, indicate much cooler bottom water temperatures of 6°-8 °C and surface water temperatures of approximately 12 °C at the Campanian/Maastrichtian boundary (Barrera and Huber 1985). Records of the species diversity and their stratigraphic distribution demonstrate long- term stability of cool-temperate conditions through the Campanian/Maastrichtian in this area. The climate of the northern part of the Antarctic Peninsula region appears, therefore, to have been of a warm/cool-temperate type during the Cretaceous and early Tertiary. Conditions for forest growth were very favourable and no single climatic factor appeared to be limiting tree growth. By comparison with warm/cool-temperate forests today, a heavy rainfall of perhaps 1000-2000 mm/yr is envisaged for this area. The qualitative palaeoclimate models of Parrish et al. (1982) also predicted a belt of high rainfall along the western margins of southern South America and the Antarctic Peninsula during the late Mesozoic. The models show a slight decrease in rainfall (though still high) from west to east across the peninsula region, presumably due to the oceanic influence to the west and the orographic effect across the highland arc area. This region was also located within an area of persistent low pressure, according to the models of Parrish and Curtis (1982). Enhanced rainfall at high latitudes was also predicted in models of increased atmospheric CO2 levels, the most plausible mechanism for the elevated temperatures in the Mesozoic (Manabe and Wetherald 1980). 680 PALAEONTOLOGY, VOLUME 29 Organic productivity of the forests. Although the wood collection studied here consists only of isolated samples, it is apparent that fossil wood is preserved throughout Cretaceous and early Tertiary strata in the Antarctic Peninsula region, evidence that the emergent arc was extensively forested for the duration of this long period. Furthermore, some of the trees are preserved in their original growth positions (Smellie et al. 1980; Farquharson 1982) and it may be possible, with further field measurements, to obtain important information about forest density, tree habit, and forest productivity similar to that obtained by Jefferson (1982, 1983) and Creber and Chaloner (1984fl, b, 1985) from the in situ Cretaceous forests on Alexander Island. The nature of the growth rings in the fossil wood from the Antarctic Peninsula suggests that these trees had the potential for high productivity (Creber, in press). Consequently, the trees may well have supplied a large quantity of organic material to the adjacent marine basins. During the same period in south-eastern Australia, cool-temperate rainforests were supplying organic debris to the Gippsland Basin, forming the Victorian Brown Coals, which later served as a source of hydrocarbons. In particular, coniferous vegetation (southern hemisphere types, principally Araucariaceae) provided large quantities of hydrogen-rich, exinite macerals (mainly from leaf cuticles and resin) for the generation of oil (Shanmugan 1985). Although there are large quantities of wood and organic material within the Cretaceous and early Tertiary proximal marine sediments of the Antarctic Peninsula back-arc basin, they were buried within predominantly oxygen-rich environments. There may have been greater potential for the preservation of organic matter as a source for hydrocarbons under anoxic conditions in more distal facies to the east (J. R. Ineson, pers. comm.), although this area is as yet unexplored. Acknowledgements. I am grateful for provision of research facilities by the British Antarctic Survey and to many colleagues there for valuable discussion and technical support. REFERENCES ANDERSSON, J. G. 1906. On the geology of Graham Land. Bull. geol. Instn. Univ. Uppsala, 7, 19-71. ARAYA, R. A. and HERVE, F. A. 1966. Estudio geomorfologico y geologico en las Islas Shetland del Sur, Antarctica. Publnes, Inst, antarct. Chil. 8, 1 -76. ASKiN, R. A. 1983a. Tithonian (uppermost Jurassic)-Barremian (Lower Cretaceous) spores, pollen and micro- plankton from the South Shetland Islands, Antarctica. In Oliver, r. l., james, p. r. and jago, j. b. (eds.). Antarctic Earth Science, 295-291 . Australian Academy of Science, Canberra. 19836. Campanian palynomorphs from James Ross and Vega Islands, Antarctic Peninsula. Antarct. Jl U.S. 18 (Review), 63-64. 1985. Gradual palynological change at the Cretaceous/Tertiary boundary in Antarctica. Workshop on Cenozoic geology of the southern high latitudes. Sixth Gondwana Symposium, Ohio (Abstracts, p. 1). ASH, J. 1983. Tree rings in tropical Callitris macleayana F. Muell. Aust. J. Bot. 31, 277-281. AXELROD, D. I. 1984. An interpretation of Cretaceous and Tertiary biota in polar regions. Palaeogeogr. Palaeoclimatol. Palaeocol. 45, 105-147. BARRERA, E. and HUBER, B. 1985. Seymour Island Campanian-Maastrichtian climate as inferred from foramini- feral isotope ratios and distributional analysis. Workshop on Cenozoic geology of the southern high latitudes. Sixth Gondwana Symposium, Ohio (Abstracts, p. 2). BARRON, E. J. 1983. A warm, equable Cretaceous: the nature of the problem. Earth Sci. Rev. 19, 305-338. 1984. Climatic implications of the variable obliquity explanation of Cretaceous-Paleogene high latitude floras. Geology, 112, 595-598. and WASHINGTON, w. M. 1982. Cretaceous climate: a comparison of atmospheric simulations with the geologic record. Palaeogeogr. Palaeoclimatol. Palaeoecol. 40, 103-133. 1984. The role of geographic variables in explaining palaeoclimates: results from Cretaceous climate model sensitivity studies. J. geophys. Res. 89, 1267 1279. THOMPSON, s. L. and SCHNEIDER, s. H. 1981. All ice-ffee Cretaceous? Results from climate model simula- tions. Science, Washington, 212, 501-508. BARTON, c. M. 1964. Significance of the Tertiary fossil floras of King George Island, South Shetland Islands. In ADIE, R. J. (ed.). Antarctic geology, 603-608. North-Holland Publishing Company, Amsterdam. FRANCIS: CRETACEOUS AND TERTIARY WOOD 681 BERNER, R. A., LASAGA, A. c. and GARRELS, R. M. 1983. The carbonate-silicatc geochemical cycle and its effect on atmospheric carbon dioxide over the past 100 million years. Am. J. Sci. 283, 641-683. BiBBY, j. s. 1966. The stratigraphy of part of north-east Graham Land and the James Ross Island group. Scient. Rep. Br. Antarct. Surv. 53, 1-37. BILLINGS, w. D. 1964. Plants and the ecosystem, 154 pp. MacMillan, London. CHALONER, w. G. and CREBER, G. T. 1973. Growth rings in fossil woods as evidence of past climates. In tarling, D. H. and RUNCORN, s. K. (eds.). Implications of contmental drift to the earth sciences, 425-437. Academic Press, London. crame, j. a. 1983. Cretaceous inoceramid bivalves from Antarctica. In Oliver, r. l., james, p. r. and i. b. (eds.). Antarctic Earth Science, 298-302. Australian Academy of Science, Canberra. and FARQUHARSON, G. w. 1984. Stratigraphical and sedimentological studies on Byers Peninsula, western Livingston Island (South Shetland Islands). British Antarctic Survey Field Report R/83-84/G9, 21 pp. (Unpublished.) cranwell, l. m. 1959. Fossil pollen from Seymour Island, Antarctica. Nature, Land. 184, 1782-1785. 1969. Antarctic and circum-Antarctic palynological contributions. Antarct. Jl U.S. 4, 197-198. CREBER, G. T. 1977. Tree rings: a natural data-storage system. Biol. Rev. 52, 349-383. (in press). Tree growth at very high latitudes in the Permian and the Mesozoic. Naturalia Monspel. and CHALONER, w. G. 1984fl. Climatic indications from growth rings in fossil woods. In brenchley, p. j. (ed.). Fossils and climate. 49-77. John Wiley, Chichester. 1984h. Environmental signihcance of wood structure in living and fossil trees. Bot. Rev. 50, 357- 448. 1985. Tree growth in the Mesozoic and Early Tertiary and the reconstruction of palaeoclimates. Palaeogeogr. Palaeoclimatol. Palaeoecol. 52, 35-60. CROFT, w. N. 1947. Geological reports for the year ending January 1947. Base D, Hope Bay. Falkland Islands Dependencies Survey. Report E89/47, 182 pp. (Unpublished.) CROW, G. A. 1975. Vegetation of Isla de Los Estados, Argentina. Antarct. Jl U.S. 10, 81-85. DALZiEL, I. w. D. and ELLIOT, D. H. 1982. West Antarctica: Problem child of Gondwanaland. Tectonics, 1, 3-19. DARLINGTON, p. J. 1965. Biogeograpliy of the southern end of the world, 236 pp. Harvard University Press, Massachusetts. DONN, w. L. 1982. The enigma of high-latitude paleoclimate. Palaeogeogr. Palaeoclimatol. Palaeoecol. 40, 199-212. DOUGLAS, J. G. and WILLIAMS, G. E. 1982. Southern polar forests: the early Cretaceous floras of Victoria and their palaeoclimatic significance. Ibid. 39, 171-185. DUSEN, p. 1908. iiber die Tertiare Flora der Seymour-Insel. Wiss. Ergebn. schwed. Siidpolarexped. 3, 27 pp. ELLIOT, D. H. and TRAUTMAN, T. A. 1982. Lower Tertiary strata on Seymour Island, Antarctic Peninsula. In CRADDOCK, c. (ed.). Antarctic Geoscience, 287-297. University of Wisconsin Press, Madison. RINALDI, c., ZINSMEISTER, W. J., TRAUTMAN, T. A., BRYANT, W. A. and DEL VALLE, R. 1975. Geological investigations on Seymour Island, Antarctic Peninsula. Antarct. Jl U.S. 10, 182-186. FARQUHARSON, G. w. 1982. Late Mesozoic sedimentation in the northern Antarctic Peninsula and its relation- ship to the southern Andes. Jl geol. Soc. Fond. 139, 721-728. 1983. Evolution of late Mesozoic sedimentary basins in the northern Antarctic Peninsula. In Oliver, R. L., JAMES, p. r. and jago, j. b. (eds.). Antarctic Earth Science, 323-327. Australian Academy of Science, Canberra. 1984. Late Mesozoic non-marine conglomeratic sequences of northern Antarctic Peninsula (the Botany Bay Group). Bull. Br. Antarct. Surv. 65, 1-32. HAMER, R. D. and INESON, J. R. 1984. Proximal volcaniclastic sedimentation in a Cretaceous back-arc basin, northern Antarctic Peninsula. In kokelaar, b. p. and howells, m. f. (eds.). Marginal basin geology, 219-229. The Geological Society and Blackwells, London. FLORIN, R. 1940. The Tertiary fossil conifers of south Chile and their phytogeographical significance. K. svenska VetenskAkad. Handl., Ser. 3, 19, 107 pp. FRAKES, L. A. 1979. Climates through geologic time, 310 pp. Elsevier, Amsterdam. FRANCIS, J. E. 1984. The seasonal environmental of the Purbeck (Upper Jurassic) fossil forests. Palaeogeogr. Palaeoclimatol. Palaeoecol. 48, 285-307. (in press). The palaeoclimatic significance of growth rings in Late Jurassic/Early Cretaceous fossil wood from southern England. In ward, r. (ed.). British Archaeological Reports. Tree-Ring Symposium, London. FRiTTS, h. c. 1976. Tree rings and climate, 576 pp. Academic Press, London. 682 PALAEONTOLOGY, VOLUME 29 FRiTTS, H. c., SMITH, D. G., CARDis, J. w. and BUDELSKY, c. A. 1965. Tree-ring characteristics along a vegetation gradient in northern Arizona. Ecology, 46, 393-401. GODLEY, E. J. 1960. The botany of southern Chile in relation to New Zealand and the subantarctic. Proc. R. Soc. Loud. B 152, 457-475. GOTHAN, w. 1908. Die fossilen Holzer von der Seymour- und Snow Hill-Insel. Wiss. Ergehn. schwed. Siidpolar- exped. 3, 33 pp. GRIME, J. p. 1979. Plant strategies and vegetation processes, 222 pp. John Wiley, Chichester. HALL, s. A. 1977. Cretaceous and Tertiary dinoflagellates from Seymour Island, Antarctica. Nature, Lond. 267, 239-241. HALLAM, A. 1985. A rcvicw of Mesozoic climates. Jl geol. Soc. Lond. 142, 433-445. HALLE, T. G. 1913. The Mesozoic flora of Graham Land. Wiss. Ergehn. schwed. Siidpolarexped. 3, 123 pp. HENDERSON, R. A. and MCNAMARA, K. J. 1985. Maastdchtian non-heteromorph ammonites from the Miria Formation, Western Australia. Palaeontology, 28, 35-88. HERNANDEZ, p. J. and AZCARATE, V. 1971. Estudio paleobotanico preliminar sobre restos de una tafoflora de la Peninsula Byers (Cerro Negro), Isla Livingston, Islas Shetland del Sur, Antarctica. Ser. dent. Inst, antart. Chil. 2, 15-50. HOWARTH, M. K. 1958. Upper Jurassic and Cretaceous ammonite faunas of Alexander Land and Graham Land. Sci. Rep. Falkid. Isl. Depend. Siirv. 21, 16 pp. — 1966. Ammonites from the Upper Cretaceous of the James Ross Island group. Bull. Br. Antarct. Surv. 10, 55-69. HUGHES, M. K., KELLY, p. M., PILCHER, J. R. and LAMARCHE, v. c. 1982. Climate from tree rings, 223 pp. Cambridge University Press, Cambridge. INESON, J. R. 1985. Submarine glide blocks from the Lower Cretaceous of the Antarctic Peninsula. Sedimen- tology, 32, 659-670. CRAME, J. A. and THOMSON, M. R. A. (in press). Lithostratigraphy of the Cretaceous strata of West James Ross Island, Antarctica. Cret. Res. JEFFERSON, T. H. 1980. Angiosperm fossils in supposed Jurassic volcanogenic shales, Antarctica. Nature, Lond. 285, 157-158. 1982. The Early Cretaceous fossil forests of Alexander Island, Antarctica. Palaeontology, 25, 681-708. 1983. Palaeoclimatic significance of some Mesozoic Antarctic fossil forests. In Oliver, r. l., james, p. r. and JAGO, J. B. (eds.). Antarctic Earth Science, 593-598. Australian Academy of Science, Canberra. KELLY, s. R. A. and BROMLEY, R. G. 1984. Ichnological nomenclature of clavate borings. Palaeontology, 27, 793-807. KENNETT, J. p. 1977. Cenozoic evolution of antarctic glaciation, the circum-antarctic ocean, and their impact on global paleoceanography. J. Geophys. Res. 82, 3843-3860. LAMARCHE, V. c. 1974. Paleoclimatic inferences from long tree-ring records. Science, Washington, 183, 1043- 1048. HOLMES, R. L., DUNWIDDIE, p. w. and DREW, L. G. 1979. Tree-ring chronologies of the Southern Hemisphere. (a)- 1 Argentina, (6). 2 Chile, (c). 3 New Zealand, (d). 4 Australia. Chronology series V. Laboratory of Tree-Ring Research, Arizona. GRAYBiLL, D. A., FRITTS, H. c. and ROSE, M. R. 1984. Increasing atmospheric carbon dioxide: tree ring evidence for growth enhancement in natural vegetation. Science, Washington, 225, 1019-1021. LAWYER, L. A., SCHLATER, J. G. and MEiNKE, L. 1985. Mesozoic and Cenozoic reconstructions of the South Atlantic. Tectonophysics, 114, 233-254. MANABE, s. and WETHERALD, R. T. 1975. The effects of doubling the CO2 concentration on the climate of a general circulation model. J. Atmos. Sci. 32, 3-15. 1980. On the distribution of climatic change resulting from an increase in CO2 content of the atmosphere. Ibid. 37, 99-118. MURATA, K. 1940. Volcanic ash as a source of silica for the silification of wood. Am. J. Sci. 238, 586-596. NATHORST, A. G. 1904. Sur la flore fossile des regions antarctiques. C.R. Acad. Sci., Paris, 138, 1447-1450. ORLANDO, H. A. 1964. The fossil flora of the surroundings of Ardley Peninsula [Ardley Island], 25 de Mayo Island [King George Island], South Shetland Islands. In adie, r. j. (ed.). Antarctic geology, 629-636. North Holland Publishing Company, Amsterdam. PARRISH, J. T. and CURTIS, R. L. 1982. Atmospheric circulation, upwelling, and organic-rich rocks in the Mesozoic and Cenozoic eras. Palaeogeogr. Palaeoclimatol. Palaeoecol. 40, 31-66. ZIEGLER, A. M. and SCOTESE, c. R. 1982. Rainfall patterns and the distribution of coals and evaporites in the Mesozoic and Cenozoic. Ibid. 67-102. FRANCIS: CRETACEOUS AND TERTIARY WOOD 683 PLUMSTEAD, E. p. 1962. Gcology. 2. Fossil floras of Antarctica. Scient. Rep. Transcmtarctic Expect. 9, 154 pp. 1964. Palaeobotany of Antarctica. In adie, r. j. (ed.). Antarctic geology, 637-654. North Holland Publish- ing Company, Amsterdam. RAO, A. R. 1953. Some observations on the Rajmahal flora. Palaeobotanist, 2, 25-28. RETALLACK, G. and DiLCHER, D. L. 1981. A coastal hypothesis for the dispersal and rise to dominance of flowering plants. In niklas, k. j. (ed.). Paleobotany, paleoeeology and evolution, 27-77. Praeger Publishers, New York. RINALDI, c. A. 1982. The Upper Cretaceous in the James Ross Island Group. In craddoc'k, c. (ed.). Antarctic Geoscience, 281 286. University of Wisconsin Press, Madison. MASSABiE, A., MORELLi, J., ROSENMAN, H. L. and DEL VALLE, R. 1978. Geologia de la Isla Vice Comodoro Marambio. Contrib. dent. Inst. Ant. Argentina. 217, 5-43. SCHWEITZER, H. J. 1980. Environment and climate in the early Tertiary of Spitsbergen. Palaeogeogr. Palaeo- climatol. Palaeoecol. 30, 297-31 1. SHACKLETON, N. J. and KENNETT, J. p. 1975. Paleotempeiature history of the Cenozoic and the initiation of Antarctic glaciation: oxygen and carbon isotope analysis in DSDP sites 277, 279 and 381. Initial Reports of the Deep Sea Drilling Project, 29, 743-755. US Government Printing Oflice, Washington, D.C. SHANMUGAN, G. 1985. Significance of coniferous rain forests and related organic matter in generating commer- cial quantities of oil, Gippsland Basin, Australia. Bull. Am. Ass. Petrol. Geol. 69, 1241-1254. SKOTTSBERG, c. 1960. Remarks on the plant geography of the southern cold temperate zone. Proc. R. Soc. Lomt. B 152, 447-457. SMELLiE, J. L., DAVIES, R. E. s. and THOMSON, M. R. A. 1980. Geology of a Mesozoic intra-arc sequence on Byers Peninsula, Livingston Island, South Shetland Islands. Bull. Br. Antarct. Surv. 50, 55-76. PANKHURST, R. J., THOMSON, M. R. A. and DAVIES, R. E. s. 1984. The geology of the South Shetland Islands: VI. Stratigraphy, geochemistry and evolution. Brit. Antarct. Surv. Sci. Rep. No. 87, 85 pp. SMILEY, c. 1967. Paleoclimatic interpretation of some Mesozoic floral sequences. Bull. Am. Ass. Petrol. Geol. 51, 849-863. SPATH, L. F. 1953. The Upper Cretaceous cephalopod fauna of Graham Land. Sci. Rep. Falklcl. Isl. Depend. Surv. 3, 60 pp. STiPANicic, p. N. and bonetti, m. i. r. 1970. Posiciones estratigraflcas y edades de las principales flora jurasicas argentinas. II. Floras doggerianas y malmicas. Ameghiniana, 7, 101-118. TAYLOR, B. J., THOMSON, M. R. A. and WILLEY, L. E. 1979. The geology of the Ablation Point -Keystone Cliff's area, Alexander Island, Scient. Rep. Br. Antarct. Surv. 82, 1-65. THOMSON, M. R. A. 1982. Mesozoic paleogeography of West Antarctica. In craddock, c. (ed.). Antarctic Geoscience, 331-337. University of Wisconsin Press, Madison. and BURN, R. w. 1977. Angiosperm fossils from latitude 70° S. Nature, Loud. 269, 139-141. PANKHURST, R. J. and CLARKSON, p. D. 1983. The Antarctic Peninsula -a Late Mesozoic-Cenozoic Arc (Review). In Oliver, r. l., james, p. r. and jago, j. b. (eds.). Antarctic Earth Science, 289-294. Cambridge University Press, Cambridge. TORRES, G. T. 1984. Identification de madera fosil del Tertiario de la Isla Rey Jorge, Islas Shetland del sur, Antartica. Ill Congreso Latinoamericano de paleontologia, Mexico, 555-565. VAARTAJA, o. 1962. Ecotypic variation in photoperiodism of trees with special reference to Pinus resinosa and Thuja occidentalis. Can. J. Bot. 40, 849-856. WOLFE, J. A. 1971. Tertiary climatic fluctuations and methods of analysis of Tertiary floras. Palaeogeogr. PalaeocUmatol. Palaeoecol. 9, 21-51 . 1978. A palaeobotanical interpretation of Tertiary climates in the Northern Hemisphere. Am. Scient. 66, 694-703. 1980. Tertiary climates and floristic relationships at high latitudes in the Northern Hemisphere. Palaeo- geogr. PalaeocUmatol. Palaeoecol. 30, 313-323. WRENN, J. H. 1985. Paleogene dinoflagellate cyst biostratigraphy, Seymour Island, Antarctica. Workshop on Cenozoic geology of the southern high latitudes. Sixth Gondwana Symposium (Abstracts, p. 36). YOUNG, s. B. 1972. Subantarctic rain forest of Magellanic Chile: distribution, composition, and age and growth rate studies of common forest trees. Antarctic Research Series, 20, 307-322. ZASTAWNiAK, E., WRONA, R., GAZDZiCKi, A. and BiRKENMAJER, K. 1985, Plant remains from the top part of the Point Hennequin Group (Upper Oligocene), King George Island (South Shetland Islands, Antarctica). Stadia geol. pol. 81, 143-164. ziNSMEiSTER, w. J. 1982. Review of the Upper Cretaceous- Lower Tertiary sequence of Seymour Island, Antarctica. Jl geol. Soc. Land. 139, 779-785. 684 PALAEONTOLOGY, VOLUME 29 ziNSMEiSTER and CAMACHO, H. H. 1982. Late Eocene (to possibly earliest Oligocene) molluscan fauna of the La Meseta Formation of Seymour Island, Antarctic Peninsula. In craddock, c. (ed.). Antarctic Geoscience, 299-304. University of Wisconsin Press, Madison. JANE E. FRANCIS British Antarctic Survey Natural Environment Research Council High Cross, Madingley Rd Cambridge CB3 OET Present address: Department of Geology and Geophysics The University of Adelaide Box 498, G.P.O. Adelaide Typescript received 19 December 1985, accepted 5 February 1986 South Australia, 5001 SCANNING ELECTRON MICROSCOPY OF UNCOATED FOSSILS by P. D. TAYLOR Abstract. The necessity of coating fossils with a conductor prior to scanning electron microscopy is avoided using a system in which backscattered electron images are formed of specimens maintained under a relatively low vacuum in an ‘environmental chamber’. Resolution and other image characteristics at low magnifications ( < 500 X ) generally compare favourably with conventional secondary electron images of coated specimens. Charging artefacts are reduced, edge effect is eliminated, and the backscattered electron image appears flatter than a conventional secondary electron image. As well as minimizing sample preparation time, the system is valuable in allowing scanning of fossils for which coating is either undesirable (e.g. type specimens) or difficult (e.g. large specimens). Commercial availability of the scanning electron microscope (SEM) during the past twenty years has revolutionized studies of fine-scale morphology. Applications of the SEM in palaeontology have been widespread. The SEM is a routine tool in micropalaeontological studies and during studies of skeletal ultrastructures in macrofossils. Conventional SEM techniques necessitate mount- ing fossils onto stubs and coating the surface with a conducting material, usually gold or gold- palladium. Although techniques are available for removal of these metallic coatings (Hansen 1968), they can be time-consuming and hazardous to the specimen. Therefore, once applied, coatings are usually looked upon as permanent. Restudy of coated specimens using an optical microscope is difficult because of the high reflectance from the specimen surface. It is common curatorial practice to discourage or even prohibit coating of important specimens such as types, thereby excluding their study with a conventional SEM. Five years experience has now been gained at the British Museum (Natural History) with a system for scanning uncoated specimens which uses an environmental chamber in conjunction with a backscattered electron detector (Buchanan 1983). The availability of this system is not widely known despite its major advantage in permitting scanning of types, etc. without alteration. The objectives of this paper are to describe the principles of operation of this system, compare images obtained with those of conventional SEM images, and discuss some applications. PRINCIPLES To understand the operation of the system for uncoated specimens it is necessary to give a brief account of some of the principles of scanning electron microscopy. A modern text such as Goldstein et al. (1981) should be consulted for details. SEM images are formed by scanning a narrow beam of electrons across the surface of a specimen, collecting and processing the emitted electrons, and displaying them on a cathode ray tube using a visual raster which is synchronized with the beam scan. When an electron beam strikes a specimen, a complex interaction takes place and several kinds of emission occur. For the purpose of scanning electron microscopy the most important emissions are secondary electrons (SE) and backscattered electrons ( BSE). SEs are shallow, low-energy emissions resulting from inelastic events which transfer energy from the electron beam to the specimen. BSEs are deeper, high-energy emissions resulting from elastic events during which there is no energy transfer between beam and specimen. Usually two to five times more BSEs are emitted than SEs (Buchanan 1983). Conventional SEM images IPalaeontology, Vol. 29, Part 4, 1986, pp. 685-690, pi. 52.| 686 PALAEONTOLOGY, VOLUME 29 A B TEXT-FIG. 1. Diagrams showing the essential features of a SEM: a, operating conventially using secondary electrons for scanning coated specimens: b, adapted for scanning uncoated specimens using backscattered electrons. VP= vacuum pump; SE= secondary electron; BSE= backscattered electron. comprise mostly SEs which are attracted to a detector, generally an Everhart-Thornley scintillator- photomultiplier, positively charged and located to the side of the specimen (text-fig. 1a). Although some BSEs are also detected, high resolution BSE imaging requires a special detector. The system described here for uncoated specimens uses a scintillator BSE detector. Unlike the SE detector, this detector is uncharged and is positioned directly above the specimen (text-fig. 1b). Conventional SEM requires that the specimen chamber as well as the microscope column be maintained at a high vacuum (about 10^'* torr) which prevents electrical discharge and interference with the electron beam by air molecules. The system for scanning uncoated specimens also operates with a high vacuum in the microscope column but has a separate vacuum pump for the specimen chamber (‘environmental chamber’) which is held at a relatively low vacuum (about 10~'-10“^ torr). The electron beam passes through a 200 micron aperture (text-fig. 1b) which is sufficiently small to allow the differential vacuum between column and chamber to be maintained. The strong positive charge of SE detectors prohibits their use in conjunction with an environmental chamber in which the low vacuum would cause electrical discharge. Non-conducting specimens, for example, the great majority of fossils, must normally be coated with a conducting material prior to scanning in order to prevent charging and enhance electron emission. Uncoated specimens irradiated by an electron beam accumulate a negative charge which can cause image distortion or at best bright spots of enhanced emission on the image. Coating allows this charge to run to earth via the specimen stub. Satisfactory images of uncoated specimens TAYLOR; UNCOATED FOSSILS 687 (Howden and Ling 1974) may sometimes be obtained by using a low beam voltage (c.5KV) but resolution is usually poor. In the system for scanning uncoated specimens residual air molecules present in the specimen chamber dissipate charge on the specimen by ionization. Argon gas can be introduced into the chamber to assist this process. The specimen need not be grounded via the stub and coating is therefore unnecessary. IMAGE CHARACTERISTICS There are considerable differences between conventional SE images of coated specimens and BSE images of uncoated specimens. These differences must be appreciated when interpreting morphology from the microscope screen or from micrographs. Therefore a specimen of the cheilostome bryozoan Metrarabdotos moniliferum (Milne Edwards) from the Pliocene Coralline Crag of Suffolk was selected for a comparative study. The specimen was first examined uncoated using BSEs, and then coated with gold-palladium and re-examined using SEs. Micrographs obtained by the two methods at three magnifications ( x 30, x 150, x 550) are shown in Plate 52 (see Cheetham (1968) for optical micrographs of M. moniliferum). The beam voltage used for both series was 15 KV, specimen working distance was the same, and brightness and contrast level equivalent. At low magnification resolution appears to be about the same for the uncoated as the conventional coated image. However, resolution tends to become noticeably inferior to that of conventional images at magnifications above about 500 x . Uncoated BSE images in excess of IK x have rarely proven satisfactory with the system in operation at the BM(NH). SE images are considerably more three-dimensional than BSE images. This is well-illustrated by comparing hgs. 1 and 2 of Plate 52. Zooecial frontal walls appear relatively flat in the BSE image but ridged and elevated in the SE image. This important difference in image characteristics must be taken into account when interpreting morphology from the SEM. BSE images of coated specimens are similarly flat. The cause of the difference may be the relative locations of BSE and SE detectors (text-fig. 1); BSE detectors are positioned directly above the specimen whereas SE detectors are positioned laterally to the specimen and receive more electrons from the side of the specimen that is closer. Specimen edges and protuberances (e.g. spines) produce high levels of emission in SE images. This ‘edge effect’ is absent from BSE images. For example, compare the fractured left-hand edge of the specimen in Plate 52, figs. 1 and 2. Edge effect is particularly pronounced in SE images of specimens on a black background. The bright circumference of specimens in SE images contributes greatly to the aesthetic appeal of scanning electron micrographs. However, lack of edge effect in BSE images means that they are closer to the appearance of specimens viewed with an optical microscope. Electron emission from cavities or depressions in specimen surfaces is greater in SE images than BSE images. For example, the areolar pores are brighter in the SE image shown in Plate 52, fig. 4 than the corresponding BSE image of Plate 52, fig. 3. The uncoated BSE image may be regarded as superior in lacking artificially high levels of pore brightness that would not be seen with an optical microscope. Additionally, charge accumulation within pores where coating may be inadequate is a problem of many coated SE images (e.g. PI. 52, fig. 6). This is invariably absent from BSE images (e.g. pi. 52, fig. 5). Whereas the relative brightness of SE images depends mostly on specimen relief, that of BSE images is determined also by the elemental composition of the specimen. This high atomic number contrast has been utilized extensively in BSE studies of clay mineralogy (e.g. Pye and Krinsley 1983) and may also be valuable in some palaeontological contexts, for example, to enhance the distinction between fossils embedded in a matrix of a different composition. However, high atomic number contrast can be disadvantageous in emphasizing the presence of adherent particles of dirt, sediment, or glue (e.g. Taylor 1984, fig. 1b). Most BSEs are emitted at high angles to the specimen surface; the number of BSEs emitted at low take-off angles diminishes rapidly. Therefore with the BSE detector positioned directly above 688 PALAEONTOLOGY, VOLUME 29 the specimen, most BSEs are detected from subhorizontal surfaces of the specimen (i.e. surfaces perpendicular to the electron beam) and few from subvertical surfaces (i.e. surfaces parallel to the beam). Relatively flat, untilted specimens produce the most satisfactory images whereas images of specimens of high relief or tilted specimens can be unsatisfactory. Perspective views and stereo pairs can be less successful of uncoated specimens using BSE imaging. DISCUSSION There are many applications of BSE imaging of uncoated specimens. Most importantly it allows scanning of specimens for which coating is not permitted or is deemed undesirable. Eor example, type specimens can be examined with the SEM without alteration or damage. Increasing use of fine scale morphological features in taxonomic discrimination means that SEM study of types is becom- ing crucial to species characterization in some groups. Eor example, SEM study of the uncoated holotype of the Cretaceous cheilostome bryozoan Charixa vetmensis Lang revealed that the spine bases supposedly diagnostic of the species were not present (Taylor, in press). Even when optical microscopy is capable of resolving the detailed morphology of type specimens it may be impossible to record these details adequately by photo-micrography in which depth of field is a problem. SEM micrographs provide an obvious solution to difficulties of illustration. Although several SEMs are equipped with large specimen chambers, specimens much larger than stub diameter (12 mm) can be impossible to scan because of difficulties in applying an adequate coating to large specimens. Inadequate coating commonly leads to charging artefacts. Uncoated specimens over 10 cm in diameter have been scanned successfully using the system in operation at the BM(NH). Removing the need to coat specimens not only reduces sample preparation time but also elimi- nates curatorial problems associated with coated specimens glued permanently to a stub. In the uncoated system, cleaned and dried fossils are simply fixed temporarily onto a stub using plasticine or a similar mounting medium. They can be removed from the stub immediately after scanning; no special provision need be made for storing stub-mounted fossils. Eragile specimens on their original mounts (e.g. card-mounted specimens) can be scanned in-situ without risking the damage that removal may entail. Savings in time become especially important when specimens need to be repeatedly coated and scanned after periods of treatment (e.g. etching). A wide taxonomic range of specimens of differing chemical composition have been scanned successfully using the system in operation at the BM(NH). These include fossils of calcitic bryozoans, goethite-encrusted bryozoans and silicified bryozoans (Taylor and Curry 1985), latex casts and shale impressions of fossil plants (Hill et al. 1985), phosphatic problematica (Taylor 1984), fossil ostracodes (Neale 1985), and Recent spicular forminifers (Bronnimann and Whittaker 1983, pi. 4), fish teeth (Greenwood 1983) and insects (Day 1984). Acknowledgements. S. H. Barnes, P. J. Chimonides, D. Claugher, and P. L. Cook kindly read the manuscript, and C. R. Hill and J. E. Whittaker provided useful discussions. EXPLANATION OF PLATE 52 Figs. 16. Scanning electron micrographs of the cheilostome bryozoan Metrarabdotos nioniliferum (Milne Edwards) from the Pliocene Coralline Crag of Gedgrove, Suffolk; British Museum (Natural History) D54322. 1, 3, and 5 are backscattered electron images of the specimen uncoated: 2, 4, and 6 are conventional secondary electron images of the specimen coated with Au-Pd. 1, 2, colony surface; the frontal walls of the autozooecia appear more convex in the SE image, and enhanced emission (edge effect) causes the fractured edge of the specimen (lower left) to appear bright, x 30. 3, 4, autozooecial orifice flanked by adventitious avicularia; note the increased brightness of the areolar pores in the SE image relative to the BSE image, X 150. 5, 6, adventitious avicularium (left) and areolar pores; the BSE image has poorer resolution but lacks the charging artefact present in the lowermost areolar pore of the SE image, x 550. PLATE 52 TAYLOR, SEM study of uncoated fossils 690 PALAEONTOLOGY, VOLUME 29 REFERENCES BRONNiMANN, p. and WHITTAKER, j. E. 1983. Zcwiiiettiii n. gen., a spicular-walled remaneicid (Foraminiferida, Trochamminacea) from the Indian and South Atlantic Oceans with remarks on the origin of the spicules. Revue de Paleohiologie 2, 13-33. BUCHANAN, R. 1983. SEM examination of nonconducting specimens. American Laboratory, April 1983. CHEETHAM, A. H. 1968. Morphology and systematics of the bryozoan genus Metrarahdotos. Smithson. Misc. Collect. 153, 1-122. DAY, M. c. 1984. The enigmatic genus Heterogyna Nagy (Hymenoptera: Sphecidae; Heterogyninae). Syst. Entom. 9, 293-307. GOLDSTEIN, J. I., NEWBURY, D. E., ECHLIN, p., JOY, D. c., FiORi, c. and LIFSHIN, E. 1981. Scanning electron microscopy and X-ray microanalysis, xiii + 673 pp. Plenum Press, New York. GREENWOOD, p. H. 1983. Oil M acroplewodus, Chilotilapia (Teleostei, Cichlidae), and the interrelationships of African cichlid species flocks. Bull. Br. Mus. nat. Hist. {Zool.) 45, 209-231. HANSEN, H. J. 1968. A technique for removing gold from plated calcareous microfossils. Micropaleontologv 14, 499-500. HILL, c. R., WAGNER, R. H. and EL-KHAYAL, A. A. 1985. Qasimia gen. nov., an early Marattia-Wkt fern from the Permian of Saudi Arabia. Scr. geol. 79, 1-50. HOWDEN, H. F. and LING, L. E. c. 1974. Low-magnification study of uncoated specimens. In hayat, m. a. (ed.) Principles and techniques of scanning electron microscopy. Volume 1. Biological Applications. 273 pp. Van Nostrand Reinhold Company, New York. NEALE, J. w. 1985. On Pokornyella mersondaviesi (Latham). Stereo-Atlas Ostracod Shells, 12, 123-126. PYE, K. and KRiNSLEY, D. H. 1983. Inter-layered clay stacks in Jurassic shales. Nature, 304, 618-620. TAYLOR, p. D. 1984. Marcusodictyon Bassler from the Lower Ordovician of Estonia: not the earliest bryozoan but a phosphatic problematicum. Alcheringa, 8, 177-186. (in press). Charixa Lang and Spinicharixa gen. nov., cheilostome bryozoans from the Lower Cretaceous. Bull. Br. Mus. nat. Hist. (Geol.), 40, and CURRY, G. b. 1985. The earliest known fenestrate bryozoan, with a short review of Lower Ordovician Bryozoa. Palaeontology, 28, 147-158. P. D. TAYLOR Department of Palaeontology British Museum (Natural History) Typescript received 15 November 1985 London SW7 5BD APPENDIX Instrumentation. The equipment used at the BM(NH) for scanning uncoated specimens consists of an ISI 60A SEM fitted with an ETPSEMRA Robinson Detector (for BSE) and a CFAS unit (charge free anticontamina- tion system which provides the low vacuum environmental chamber for the specimen). These are marketed by Expo-SEM, Moat Farm, Church Road, Milden, Ipswich, Suffolk IP7 7AF. The Robinson Detector is available for most makes of SEM, and the possibility of developing CFAS units for SEMs other than the ISI are being explored (A. J. Ditheridge, pers. comm. May 1985). COELOBITES AND SPATIAL REFUGES IN A LOWER CRETACEOUS COBBLE-DWELLING HARDGROUND FAUNA by MARK A. WILSON Abstract. A diverse hardground fauna of encrusters, nestlers and borers has been found on calcareous cobbles in the Lower Cretaceous (Upper Aptian) Faringdon Sponge Gravel of south-central England. The bulk of the fauna consists of coelobites that inhabited the vacated borings of bivalves. These coelobites often clustered near the cavity openings, apparently to escape epifaunal predators and physical abrasion, and to avoid sedimentary infilling of the cavity interiors. The few encrusting species common on the outside surfaces are robust and apparently adapted to abrasion resistance. Some species, notably the serpulid Glomerula gordialis, had a growth strategy that exploited the advantages of cavity-dwelling as juveniles and the resources available on the outside cobble surfaces as adults. The fauna thus shows adaptations to life in cavities and to existence on a mobile hardground in a high energy environment. The increase in hardground boring in the Upper Palaeozoic and Mesozoic may have caused a general increase in hardground faunal diversity by providing more niche space for coelobites. Coelobites, or cavity-dwelling organisms (Ginsburg and Schroeder 1973), form discrete com- munities that first appear in Lower Cambrian rocks and are common today (Kobluk 1981fl, Choi 1984). Epizoans on pebble, cobble, or boulder substrates (‘mobile hardgrounds’) also range from the Lower Palaeozoic (at least as early as the Ordovician) to the Recent (Wilson 1985, Osman 1977). Organisms from both types of communities usually experience high levels of competition, especially for living space. When these organisms are occasionally fossilized in life positions, they become valuable tools for reconstructing ancient ecological systems and for charting evolution in well-defined niches. An extraordinarily diverse fauna of encrusting, nestling and boring invertebrates has been found preserved on and in heavily bored calcareous claystone and siltstone cobbles from the Faringdon Sponge Gravels (Lower Greensand; Lower Cretaceous) at Faringdon, England. The fossils are in life positions within the cavities and on the outer cobble surfaces, and so show adaptations spanning the coelobite and mobile hardground niches. Broadly, the cavities were spatial refuges from the physical and biological rigours inherent in living on a cobble in a highly disturbed gravel environ- ment. LOCALITY AND STRATIGRAPHY The fossiliferous cobbles were collected from the ‘Red Gravel’ of the Faringdon Sponge Gravel in an interval between 5-5 and 8-5 metres above the disconformable contact with Oxfordian limestones of the Corallian Group, in the Wicklesham gravel pit at the south-east edge of Faringdon, Oxfordshire, England (national grid reference SU 293943; lat. 51° 41' 20" N, long. 1° 34' 45" W; text-fig. 1). Faringdon fossils have attracted scientific attention for centuries. The first published reference to them is found in the museum catalogue by Lhwyd (1699). Mantell (1838, 1844) discussed the fossils in some detail, but it was Austen (1850) who produced the first comprehensive study of the Faringdon Sponge Gravel. He (Austen 1850, p. 459) was also the first to note the bored cobbles, referring to them as '. . . fragments of secondary calcareous rocks, much eaten out by perforating animals’. Meyer (1864) established the internal stratigraphy of the Sponge Gravel, naming the lowermost unit the ‘Yellow gravel’ and the overlying ferru- ginous sediments (where the fossiliferous cobbles used in this study are found) the ‘Red gravel’. Melville [Palaeontology, Vol. 29, Part 4, 1986, pp. 691-703, pi. 53.) 692 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1. Location of the Wicklesham Pit, Faringdon, where the fossiliferous cobbles were collected. The solid pattern is the outcrop of the Lower Greensand in southeastern England (after Casey 1961). (1941, p. 11) proposed using the term ‘Sponge Gravel’ for both the Yellow and Red gravels. This usage has been followed by all subsequent authors. The Faringdon Sponge Gravel is placed in the Parahoplites nutfieldensis Zone of the Upper Aptian (Casey 1961). PALAEOENVIRONMENT The fossil-bearing cobbles are scattered throughout the top three metres of the gravels. They are interspersed with sand and gravel and are not confined to discrete layers. They also do not show any preferential orientation. The cobbles are spherical to oblate and range in size from 2 to 10 cm in the longest dimension, with the most common size at approximately 8 cm. They are composed of fine-grained sediments ranging from calcareous clay to coarse silt in a calcareous matrix, with a few specimens of oolitic limestone. All of the studied cobbles have a carbonate content of at least 40 % by weight. The presence of encrusting organisms and borings on all surfaces, and the occurrence of exterior encrusters that were abraded and then regenerated or had their space reoccupied, shows that these cobbles were rolled (and subjected to abrasion) during inhabitation. WILSON: CRETACEOUS COBBLE-DWELLERS 693 In addition, there are a variety of encrusted cobbles composed of either quartzite or rhyolite, although the latter is rare. There are also bored and encrusted phosphatic fragments of reworked Jurassic ammonites. Only the calcareous cobbles are considered in this study because they have an approximately constant composition and are consistently bored. One calcareous claystone cobble contained within its matrix the ammonite Prorasenia howerhanki (kindly identified by H. G. Owen). This indicates that at least some of the cobbles were derived from the lower Kimmeridge Clay or Upper Oxfordian clays and siltstones, which had long been suspected (Arkell 1947). In the most thorough palaeoenvironmental analysis of the Faringdon Sponge Gravel, Krantz (1972) concluded that these sediments were deposited in narrow channels during an Early Creta- ceous transgression. The gravel was continually reworked in its upper layers, probably by tidal currents. Bridges ( 1 982) presents a palaeogeographic reconstruction of southeastern England during the late Upper Aptian showing the probable sediment sources for the Faringdon Sponge Gravel and related units. METHODS Three hundred cobbles were collected from the top three metres of the Faringdon Sponge Gravel. Some were removed from the gravel pit wall, but most were collected loose on a shelf below the interval excavated by quarrymen. One hundred cobbles were randomly selected and washed with water and detergent to loosen the sand and gravel filling the borings. Each cobble in this collection was then carefully broken apart and the identity and location of the encrusting species recorded. Location on the eobble was listed as one of four areas: (1) outer surface, (2) interior surface within 2 mm of the boring opening, (3) interior surface deeper than 2 mm from the boring opening, and (4) inside-to-outside, for those forms that grew from within a boring onto the outside surface. Because of the probable loss of encrusting specimens during the breaking process, only the presence, and not the numerical density, of a species could be recorded from each cobble. Nestling species were recorded only if their shells were trapped within a boring in such a way that they could not have been washed in. The remaining 200 cobbles were washed and broken apart in the same manner as the first 100, but only to search for rare species that may not have been recorded in the numerical census. Representative specimens of the bryozoans are deposited in the British Museum (Natural History) collections (BMNH D55420-38), as are the nestling bivalves (BMNH LL3 1824-31) and the fora- miniferans (BMNH P5 1734-9). Samples of the cobbles and the remaining species are deposited in the Oxford University Museum (OUMK 37751-93). RESULTS Erom the entire eollection of 300 cobbles, 37 encrusting, 5 nestling and 1 boring species were recorded, along with 8 ichnospecies (Table 1). Table 1 also contains the results of the census of the random sample of 100 cobbles. It should be noted that the results in Table 1 are for statistical comparison only. Some species that in this study were ‘entirely inside’ are found on the exteriors of some cobbles housed in the British Museum (Natural History). Table 2 is derived from Table 1 and shows the ranked percentage of the enerusting species recorded inside the borings, excluding those species found on less than 5 % of the cobbles. THE FAUNA Ichnofossils. The most common ichnofossils, homes for most of the enerusters, are the clavate Gastrochaenolites borings recently described by Kelly and Bromley (1984). Over 90% of these borings are referred to G. lapidicus (PI. 53, fig. 4). A few G. lapidicus specimens still possess a calcareous lining, and fewer still contain poorly preserved bivalves of the Subfamily Lithophaginae. 694 PALAEONTOLOGY, VOLUME 29 ■ir; ^ (y Ci. (V> s ° a y o S "i CU C c O {3ij C H n ^ O X) (U ■§ g'S o o ao 00 ^ o &i:j ^ r- c/5 b i> 03 CJ ^ "O o I o3 5 O O c/5 C c3 OJ L— C^_ • Ui U c/5 3 5^ C/5 ^ ^ u 2 S oj t« X) 3 ■-= O 5 ^ ^ G-. c O a S "o — ^ "d o o> -,2 "O c/5 O ^ C/5 Ui D fC 'u ^ (U a s c/5 I ^ D. w •- ^ O CQ ^ < D H 'O O O t-H c/5 O ia cu X X o X O o ZJ ^ Ciij rsj 5 c S X o .dl a N > c3 c u X o ^ cd 03 X (U cd >0^ "I S ^ s: Q> -2 ^ ? Cj « s g •2 a 5 < o . C >1 O c bO M i; ^ Pp 'a ^ § 2 J3 S ’ t3 a ^ C< CO Kl s ! .§ P 'X Co X I :r- O : ^ c/^ £P Si, .<>5 I s'2 Q aj bo "» S a •a § § "o o' o iiP < c c & S C/5 g ^ bc-^ >> X u (U O C/!i u i < -a 8 ^ 2 ^ X 'Z < C oi CQ •T3 X) O ^ X C) W S « -H o g a V to ^ « 5 o, S -2 ^ ~ aj p S S Cj C5 <1, e o tH CQ oi^ 'o § s ’s ot i'l O, (ij I s ?u >' 0) E -5 £ o CQ >-^ W l| (U S 2 4 1 < 00 bX) bX) .a .a X X c o 00 00 < CP CU <:w (U (U £• s- bfi bX) 5 5 C C ^ O O K X Xr^ oo c/:i Si TABLE 1. Species recorded from the calcareous claystone cobbles of the Karingdon Sponge Gravel. See text for complete descriptions. The statistical columns contain the results of the detailed census of 100 cobbles. The numbers represent the percentage of cobbles in or on which the particular species was found. Those specimens recorded 'near opening’ were within 2 mm of the outer lip of a boring. Those listed as ‘inside to outside' extended from a boring onto the outer cobble surface. Cobble-Dwelling Species Life Habit Exterior Interior Inside to Near Deep Outside Opening Inside i-okaminifera: Family Placopsilinidae, sp. A Encrusting, agglutinated 2 22 Acruliammiria sp. A Encrusting, agglutinated Bdclloidina cf. B. vincenionnemis Hoflccr Encrusting, agglutinated BuUopora sp. A Encrusting, calcareous 8 Nuhoctilinella cf. N. higoli Cushman Encrusting, calcareous porifera: Ncuropora Iwmisphcrica Canu & Bassler Encrusting, lamellar 14 1 Barroisia o/ioj/fwno.wi.v (Mantell) Erect, encrusting base 2 1 Coryiiellu foraminosa (Goldfuss) Erect, encrusting base 2 7 iiryozoa: 'Stoniaiopora calvpao' (d'Orbigny) Encrusting, runner-type 1 6 Siomaiopora sp. A Encrusting, runner-type 7 26 32 2 S. sp B Encrusting, runner-type 1 S. sp. C Encrusting, runner-type 1 ' Proboscinu coarciaia Canu & Bassler Encrusting, ribbon-type 4 7 15 1 P.' cornucopia (d’Orbigny) Encrusting, ribbon-type P ' parvuUi Canu & Bassler Encrusting, ribbon-type 'P ' sp. A Encrusting, ribbon-type 3 1 10 ■p: sp. B Encrusting, ribbon-type 3 1 1 'Bcrcnicea' grandipora Canu & Bassler Encrusting, sheet-like 3 4 B ' orbifera (Canu & Bassler) Encrusting, sheet-like 13 23 17 3 B.‘ spissa (Gregory) Encrusting, sheet-like 1 'B.' hainiei (sensu Gregory 1 899) Encrusting, sheet-like 3 2 B.' sp. A Encrusting, sheet-like 1 2 Ccriopora collis (d’Orbigny) Encrusting, mound-likc Afiilticrescis manimilosa Canu & Bassler Encrusting, mound-like Discosparsa fecunda (Vine) Encrusting, mound-likc 10 3 1 Tbtilopora virgulosa Gregory Encrusting, mound-like 6 11 1 Idmonea deniiculaia (Canu & Bassler) Encrusting, ribbon-type 2 2 1 Repioclatisa bagenowi (Sharpe) Encrusting, sheet-like 15 1 1 Semimuhicavea sp. A Encrusting, mound-like 1 ■ Mcliccritites semiclausa' Gregory Erect, encrusting base Afeliccriiites hainieana d’Orbigny Erect, encrusting base ANNELIDA: Flticiicularia sbarpei Ware Encrusting 2 4 1 2 Proliserpula faringdonensis Ware Encrusting Propomtiioceros gracilis Ware Encrusting 5 4 4 4 Glomerula gordialis (Schlotheim) Encrusting, partly erect 19 23 25 hivalvia; Subfamily Mytilinac, sp. A Nestling Subfamily Lithophaginae. sp. A Boring Exogvra sp. Encrusting 6 3 Lopha diluviana (Linnc) Encrusting BRACHIOPODA. Cvcloihyris depressa (J. de C. Sowerby) Nestling C. lepidu Owen Nestling Gcmniarcula aurea Elliott Nestling Praelongilbyris praeloiigiforma Middlemiss Nestling ichnofossils: Macandropolydora sulcans Voigt Gasirochaenoliles lapidiau Kelly & Bromley G. duniformis Kelly & Bromley C- lurbinaius Kelly & Bromley Sponge Boring A Sponge Boring B Trypaniies sp. A Trypanites sp. B 696 PALAEONTOLOGY, VOLUME 29 TABLE 2. Percentage of encrusting species found inside the borings (excluding those species found in less than 5 % of the cobbles). Based on the data from Table 2. Placopsilinid sp. A Bullopora sp. A "Stomatopora calypso' ' Proboscina' sp. B G/omerula gordialis Stomatopora sp. A 'Proboscina' coarc tat a 'Proboscina' sp. A Corynella foraminosa Flucticularia sharpei 'Berenicea' orbifera Propomatoceros gracilis Tholopora virgulosa Exogyra sp. Discosparsa fecunda Reptoclausa hagenowi Neuropora liemispherica 100% 100% 1 VfV/Q 100% 100% 100% Entirely inside 29V Usually outside 90% 85% 79% 78% Usually inside These were apparently the original borers. Sponge borings A and B are shallow systems of ramifying tubes, possibly produced by the holdfasts of larger poriferans (PI. 53, fig. 7). Trypanites sp. A is a relatively straight, cylindrical tube with a diameter of approximately 3 mm. Trypanites sp. B is a curved cylindrical tube, approximately 5 mm. in diameter, with an expanded chamber at its terminus (PI. 53, fig. 6). Maeandropolydora sulcans Voigt, a wandering cylindrical tube between 1 and 2 mm. in diameter (PI. 53, fig. 2), was recently redescribed by Bromley and d’ Alessandro (1983). Forantiniferida. One new genus and three new species of adherent foraminiferans from this fauna are described in Wilson (1986). Porifera. Neuropora liemispherica (a sclerosponge; see Kazmierczak and Hillmer 1974) is the most common poriferan. Barroisia anastomosans and Corynella foraminosa are found as juveniles. Bryozoa. All of the lettered species in Table 1 are new and will be described in a taxonomic paper on the Faringdon bryozoans by L. J. Pitt and P. D. Taylor. Those taxa in quotation marks are species noted by Canu and Bassler (1926); their correct classification has not yet been determined. The bryozoan life-habits are taken from the colony shapes discussed by Taylor (1984). Figs. 1-10. Fauna or cobbles in the Faringdon Sponge Gravel. 1, Flucticularia sharpei Ware extending from within a boring (Gastrochaenolites iapidicus Kelly and Bromley) to the cobble surface (centre), Glomerula gordialis (Schlotheim), and another F. sharpei within a boring (lower right), OUM K. 37753, x 1. 2, Maeandropolydora sulcans Voigt boring, OUMK 37751, x 0-7. 3, Stomatopora sp. A (left) and 'Proboscina' sp. B within a G. Iapidicus boring, BMNH D55422, x 15. 4, Complete cobble with G. Iapidicus borings, OUMK 37757, x 0-6. 5, Cyclothyris lepida Owen nestling within a G. Iapidicus boring, OUMK 37752, x 2. 6, Trypanites sp. B in a broken cobble, OUMK 37756, x 1. 7, Sponge boring A, OUMK 37755, x 1. 8, Bivalve of the Subfamily Mytilinae (sp. A) nestling between the valves of another mytilinan within a G. iapidicus boring, BMNH LL31824, x 8. 9, Reptoclausa hagenowi (Sharpe) encrusting the exterior of a cobble, BMNH D55420, x 0-8. 10, G. Iapidicus boring interior with 'Proboscina' sp. B (centre), Glomerula gordialis (lower left), and Stomatopora sp. A (upper right), BMNH D55421, x 2. EXPLANATION OF PLATE 53 PLATE 53 10 WILSON, Lower Cretaceous cobble-dwelling fauna 698 PALAEONTOLOGY, VOLUME 29 Annelida. The serpulid tubes are classified according to the scheme proposed by Ware (1975). Bivalvia. The most common nestler is a small, well-ornamented mytilid that apparently fits no described genus (PI. 53, fig. 8). Holder (1972) found similar bivalves nestling in borings excavated in Jurassic belemnite rostra. A few rare specimens of a smooth-shelled boring bivalve were found in situ, but their poor preservation precluded generic identification. The attachment scars of Exogyra and Lopha are common on the cobble outer surfaces, as would be expected, but some whole shells were also found attached to the interiors of the borings. Brachiopoda. Nestling brachiopods are common in the borings, but they are difficult to extricate as whole shells. The rhynchonellid Cyclothyris (PI. 53, fig. 5) is the most abundant, followed by the terebratulids Gernmarcula and Praelongithyris. FACTORS CONTROLLING THE DISTRIBUTION OF COELOBITE AND COBBLE-DWEEEl NG EAUNA Modern coelobites Marine coelobite faunas in the Recent are physically controlled by relative levels of light and food. Garrett el al. (1971, p. 657) recognized three cavity-dwelling assemblages controlled by light availability. The ‘open’, ‘gloomy’, and ‘dark’ communities show the expected decrease in algal abundance with decreasing light, but they also show concurrent increases in bryozoan and encrusting foraminiferan abundance. The bryozoans and foraminiferans may be benefiting from the lower level of space competition with algae in the darker cavity recesses. Food resources are a function of the size of the cavity and the number and type of openings into it. A large cavity will contain a relatively large amount of suspended food. Small cavities contain smaller amounts of food and thus the ‘Konsumationszeit’ (the time in which the fauna filters the water content of a cavity; Reidl 1966) is shorter in smaller cavities. The larger and more numerous the connections between the cavity and the open water, the more quickly these food resources can be replaced. Reidl (1966) showed that modern cavity-dwellers are often distributed with the passive filter-feeders clustered near cavity openings and the active (and more efficient) filter- feeders in the recesses where the water is less turbulent and partially depleted of its original food content. Interspecific competition is a well-known phenomenon in modern coelobite faunas. Jackson (1977), Choi (1984) and others have shown that in general, solitary encrusters are usually the pioneers on a new substrate, but they are later outcompeted for living space by colonial organisms. Predation, while not rare, is uncommon in coelobite faunas (Jackson 1977). Fossil coelobites Fossil cavity-dwelling assemblages show evidence that food and light availability were also basic controls of their distribution. The known record of Palaeozoic coelobite faunas was summarized by Kobluk (1981a), and the most diverse Mesozoic coelobite fauna was described by Palmer and Fiirsich (1974). Although there appears to be no evidence of algal growth. Palmer and Fiirsich (1974) proposed that the serpulids and bryozoans in their Jurassic crevice-dwelling fauna were abundant because low levels of light excluded algae. They also showed a zonation in which active filter-feeders dominated the crevice roof fauna, which would have had the lowest ambient levels of suspended food. An additional physical factor that limits fossil coelobite faunas is the sedimentary infilling of the cavities (Kobluk 1981a, 1985). The influx of sediment would kill the encrusting organisms by either burying them or restricting the flow of water in the cavities. The biological factors that controlled fossil coelobite faunas have been difficult to demonstrate. Direct evidence of interspecific competition among these organisms has not been shown. Holder (1972), Palmer and Fiirsich (1974), Kobluk and James (1979), and Kobluk (1985) have shown some WILSON: CRETACEOUS COBBLE-DWELLERS 699 indistinct successions in fossil coelobite faunas that can be alternatively interpreted as overgrowth patterns resulting from interspecific competition. Modern cobble-dwellers Recent marine epifaunal cobble-dwelling communities are primarily controlled by physical dis- turbance. Osman (1977) showed that the encrusting organisms are continually subject to local catastrophes when the cobble is overturned or abraded. This high level of disturbance produces diverse communities in which the top competitors for space are prevented from occupying the entire surface of the substrate. When disturbanee frequeneies are low, modern epifaunal cobble-dwelling faunas experience the same high levels of interspecific competition as the coelobite faunas, with the same resulting pattern of colonial dominance over solitary forms. The competition on these open substrates, though, may be lessened by predation, which may act as a ‘biological disturbance’ to reduce the abundance of the top space competitor (Paine 1974). Fossil cobble-dwellers Wilson (1985) showed that an epifaunal bryozoan, edrioasteroid, and crinoid community inhabiting Ordovician cobbles was subjected to the same physical disturbances and levels of interspecific competition noted by workers in the Recent. The resulting pattern of high diversity on disturbed substrates was in agreement with the conclusions of Osman (1977) and the dominance of colonial organisms over solitary on undisturbed cobbles matched Jackson’s (1977) interpretations. FACTORS CONTROLLING THE DISTRIBUTION OF THE FARINGDON FAUNA The cobbles of the Faringdon Sponge Gravel provided two major habitats: the exterior surfaces and the interiors of the borings (text-fig. 2). The cobble fauna was thus controlled by a combination of physical and biological factors associated with cavity-dwelling and epifaunal encrustation. Exterior encrusters Those forms usually found on the exterior surfaces of the Faringdon cobbles are characteristically robust, heavy-shelled species (Table 2). The two most common exterior species, the bryozoan Reptoclausa hagenowi (PI. 53, fig. 9) and the sclerosponge Neuropora liemisplierica, produce low sheet-like colonies with large areas of attachment and no erect or protruding branches. R. hagenowi has an especially interesting morphology that combines concentrated zooids on radiating ridges with a large attachment surface. This form may have been advantageous in the high-energy gravel environment, since most of the zooids on ridge flanks would have been protected by ridge crests from direct impacts when the cobbles were rolled, and the sheet-like colony base would form a strong attachment surface. There is undoubtedly some preservational bias behind the distribution of exterior species because robust fossils will survive the abrasive conditions longer than the more delicate forms. Abraded specimens are indeed common, but so are unworn encrusters. If other species had frequently inhabited the outside surfaces, more of their fossils would be expected to have survived post-mortem abrasion. Interior encrusters {coelobites) Every species in this fauna was found at least once on the inside of a boring, and most were found there predominantly (Table 2). Those encrusters that always occurred inside, including the juvenile stages of Glomerula gordialis, are characterized by relatively fragile skeletons and comparatively weak attachments. The bryozoans in this ‘entirely inside’ group do not include any sheet-like forms, but several runner-like and ribbon-like species are present (PI. 53, figs. 3, 10). Most of the interiors of borings were probably provided with through-flowing water currents 700 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 2. Reconstruction of the cobble-dwelling community at Faringdon. a, encrust- ing foraminiferan; b, Stomatopora (bryozoan); c, Glomeriila gordialis (serpulid); D, ‘Bere- nicea (bryozoan); E, Reptoclausa hagenowi (bryozoan); f, Exogyra (oyster); G, nestling mytilid (bivalve); H, Trypanites sp. A (ichnofossil); i, GasIrochaenoHtes lapidicus (ichno- fossil). because of the overlapping and interconnected cavities. Food resources would have thus been nearly as high inside the borings as in the outside water, although they may have been somewhat reduced by the higher concentration of filter-feeders. The coelobites would have been limited by light availability and sediment influx. The deep interiors of the borings would have only received light reflected down the passageways, and so would be classified as ‘gloomy’ in the modern work of Garrett et al. (1971). Sediment infilling of the cavities, though, would have been a far greater limitation. All of the borings eventually filled with sand, and this was undoubtedly the factor that killed most of the preserved specimens. Siliciclastic sediment frequently entered the cavities during the period of inhabitation as shown by the number of encrusters overgrowing geopetal accumulations of sand and gravel. There is little direct evidence of space competition preserved in the Faringdon fauna. Only twenty- six encounters between species (where one is superimposed on another) were recorded in the entire collection of 300 cobbles. In none of these encounters was there morphologic evidence of interference competition. There was also no consistent pattern in these overgrowths. However, these cobbles were certainly colonized by soft-bodied organisms that left no record. Tunica tes and sponges, for example, routinely outcompete bryozoans and serpulids for living space on modern substrates (Jackson 1977; Osman 1977; Choi 1984). Interspecific competition was probably an important factor on the Faringdon cobble surfaces and within the borings, but its scale cannot even be estimated (see Rasmussen and Brett 1985). The Faringdon coelobites could have thus exploited the bivalve borings as spatial refuges from the abrasive conditions in the high-energy gravel, and as probable refuges from predation and WILSON: CRETACEOUS COBBLE-DWELLERS 701 competition on the exterior surfaces. The limitations of this coelobite life habit were reduced levels of food and light and the dangers of sedimentary infilling. There were two methods by which the coelobites mitigated the limitations of cavity-dwelling. By living near the cavity openings (Table 1), some organisms retained the refuge from predators and abrasion, yet had greater resources of light and food-bearing currents than those forms deep in the borings. They would also survive the sedimentary infilling process considerably longer. Most of the coelobites in this study were found near the cavity openings, with the notable exceptions of species of the bryozoan ^ Proboscina and the agglutinated foraminiferan placopsilinid sp. A. A second method of escaping the limitations of coelobitic life was to live on both the interior and the exterior (PI. 53, fig. 1). The convoluted shell of the serpulid G. gordialis is very commonly found starting as a contorted, thin-shelled tube inside a boring, and then growing out of the cavity onto the external cobble surface as a fairly robust, thick tube. This could be either a growth strategy adapted to the physical and biological rigours of epifaunal cobble-dwelling or a method of escaping progressive sedimentary infilling of the boring. The serpulids may have originally settled in this orientation to take full advantage of the food-bearing water currents, in a similar fashion to the cornulitids discussed by Schumann (1967). Occasionally other encrusters also show this inside- to-outside growth pattern, although at such a low frequency that it was probably due to luck rather than design. IMPLICATIONS FOR HARDGROUND DIVERSITY THROUGH TIME Mesozoic coelobite fauna have been shown to be more diverse than their Palaeozoic counterparts (Palmer 1982), although later work has narrowed that difference (Kobluk 1980, 198H/, 19816, 1981c, 1985). With the exception of the work by Holder (1972) and a brief mention by Voigt (1973), previous analyses of fossil coelobites have not considered the fauna within borings. It has been demonstrated in this study and in work on modern communities (R. C. Evans (1949), J. W. Evans (1967), and Warme (1970)) that vacant bivalve borings can be important niches for encrusting organisms. Since hardground boring increased dramatically in the Upper Palaeozoic and Mesozoic (Palmer 1982) as part of the general trend of bivalve infaunalization (Stanley 1977), we can predict that hardground diversity (including coelobites) will show a parallel increase. The comparatively high diversity of the Cretaceous cobble-dwelling hardground fauna from Faringdon clearly results from the availability of empty borings as a habitat for encrusting species. This evidence supports the hypothesis that higher frequency of boring was a contributing factor to the increase in hardground diversity. Acknowledgements. This work was completed during a research leave spent in the Department of Earth Sciences, University of Oxford. I am grateful for the generous support from Oxford and The College of Wooster. P. D. Taylor of the BMNH was especially helpful with his advice and criticism. I also thank W. S. McKerrow, T. J. Palmer, W. J. Kennedy, G. M. Wilson, H. P. Powell, J. E. Whittaker, N. Morris, L. J. Pitt, F. A. Middlemiss, and H. G. Owen. Technical support was given by C. Fagg, C. Grainger, A. Fowler, and R. McAvoy. REFERENCES ARKELL, w. J. 1947. The geology of Oxford, 267 pp. Clarendon Press, Oxford. AUSTEN, R. A. c. 1850. On the age and position of the fossiliferous sands and gravels of Faringdon. Q. Jl geol. Soc. Lond. 6, 454-478. BRIDGES, P. H. 1982. Ancient offshore tidal deposits. In Stride, A. H. (ed.). Offshore tidal sands, 172-192. Chapman and Hall, London. BROMLEY, R. G. and d’alessandro, a. 1983. Bioerosion in the Pleistocene of southern Italy: Ichnogenera Caulostrepsis and Maeandropolydora. Riv. It. Paleont. Strat. 89, 283-309. CANU, F. and bassler, r. s. 1926. Studies on the cyclostomatous Bryozoa. Proc. U.S. Natl. Museum, 67, 1-124. 702 PALAEONTOLOGY, VOLUME 29 CASEY, R. 1961. The stratigraphical palaeontology of the Lower Greensand. Palaeontology, 3, 487-621. CHOI, D. R. 1984. Ecological succession of reef cavity-dwellers (coelobites) in coral rubble. Bull. Mar. Science, 35, 72-79. EVANS, J. w. 1967. Relationship between Penitella penita (Conrad 1837) and other organisms of the rocky shore. Veliger, 10, 148. EVANS, R. c. 1949. The intertidal ecology of rocky shores in south Pembrokeshire. J. Ecol. 37, 120-139. GARRETT, p., SMITH, D. L., WILSON, A. o. and PATRiQUiN, D. 1971. Physiography, ecology, and sediments of two Bermuda patch reefs. J. Geol. 79, 647-668. GiNSBURG, R. N. and SCHROEDER, J. H. 1973. Growth and submarine fossilization of algal cup reefs, Bermuda. Sedimentology, 20, 575-614. HOLDER, H. 1972. Endo- und Epizoen von Belemniten-Rostren {Megateuthis) im nordwestdeutschen Bajocium (Mittlerer Jura). Paldont. Z. 46, 199-220. JACKSON, J. B. c. 1977. Competition on marine hard substrata: the adaptive significance of solitary and colonial strategies. Atner. Naturalist, 111, 1A3-161. KAZMiERCZAK, J. and HiLLMER, G. 1974. Sclerosponge nature of the lower Hauterivian ‘bryozoan’ Neuropora pustulosa (Roemer 1839) from western Germany. Acta palaeontol. Polonica, 29, 443-453. KELLY, s. R. A. and BROMLEY, R. G. 1984. Ichnological nomenclature of clavate borings. Palaeontology, 27, 793-807. KOBLUK, D. R. 1980. Upper Ordovician (Richmondian) cavity-dwelling (coelobiontic) organisms from southern Ontario. Can. J. Earth Sci. 17, 1616- 1627. 198L/. The record of early cavity-dwelling (coelobiontic) organisms in the Paleozoic. Ibid. 18, 181-190. 19816. Earliest cavity-dwelling organisms (coelobionts). Lower Cambrian Poleta Eormation, Nevada. Ibid. 669-679. 1981c. Cavity-dwelling biota in Middle Ordovician (Chazy) bryozoan mounds from Quebec. Ibid. 42-54. 1985. Biota preserved within cavities in Cambrian Epiphyton mounds, upper Shady Dolomite, south- western Virginia. J. Paleont. 59, 1 158-1 172. and JAMES, N. p. 1 979. Cavity-dwelling organisms in Lower Cambrian patch reefs from southern Labrador. Lethaia, 12, 193-218. KRANTZ, R. 1972. Die Sponge-Gravels von Earingdon (England). Neues Jahrh. Geol. Paldontol. Abh. 140, 207-231. LHWYD, E. 1699. Lithophylacii Britannici Ichnographia, 139 pp. London. MANTELL, G. 1838. The wonders of geology, volume 2, 312 pp. London. 1844. The medals of creation, volume 2, 559 pp. London. MELVILLE, R. v. 1941. The Lower Greensand of Earingdon, Berks. MSc. thesis (unpubl.). University of Reading. MEYER, c. J. A. 1864. Three days at Farringdon. Geologist, 7, 5-1 1. OSMAN, R. w. 1977. The establishment and development of a marine epifaunal community. Ecol. Monogr. 47, 37-63. PAINE, R. T. 1974. Intertidal community structure: experimental studies on the relationship between a dominant competitor and its principal predator. Oecologia, 15, 93-120. PALMER, T. J. 1982. Cambrian to Cretaceous changes in hardground communities. Lethaia, 15, 309-323. and fOrsich, f. t. 1974. The ecology of a Middle Jurassic hardground and crevice fauna. Palaeontology, 17, 507-524. RASMUSSEN, K. A. and BRETT, c. E. 1985. Taphonomy of Holocene cryptic biotas from St Croix, Virgin Islands: Information loss and preservational biases. Geology, 13, 551-553. REiDL, R. 1966. Biologic der Meereshdhlen, 636 pp. Hamburg and Berlin. SCHUMANN, D. 1967. Die Lebenweise von Mucrospirifer Grabau, 1931 (Brachiopoda). Palaeogeography, Palaeoclimatol., Palaeoecol. 3, 381-392. STANLEY, s. M. 1977. Trends, rates and patterns of evolution in the Bivalvia. In hallam, a. (ed.). Patterns of evolution, 209-250. Elsevier, Amsterdam. TAYLOR, p. D. 1984. Adaptations for spatial competition and utilization in Silurian encrusting bryozoans. pp. 197-210 In BASSETT, m. g. and lawson, j. d. (eds.). Autecology of Silurian organisms. Spec. Pap. Palaeont. 32, 1-295. VOIGT, E. 1973. Environmental conditions of bryozoan ecology of the hardground biotope of the Maastrichtian tuff-chalk, near Maastricht (Netherlands). In larwood, g. p. (ed.). Living and fossil Bryozoa, 185-197. Academic Press, London. WARE, s. 1975. British Lower Greensand Serpulidae. Palaeontology, 18, 93-1 16. WILSON: CRETACEOUS COBBLE-DWELLERS 703 WARME, j. E. 1970. Traces and significance of marine rock borers. In crimes, t. p. and harper, j. c. (eds.). Trace fossils. Geol. J. Spec. Issue, 3, 515-526. WILSON, M. A. 1985. Disturbance and ecologic succession in an Upper Ordovician cobble-dwelling hardground community. Science, 228, 575-577. 1986. New adherent foraminiferans from the Lower Cretaceous (Aptian) of south-central England. J. Micropcilaeontol. MARK. A. WILSON Typescript received 21 October 1985 Revised typescript received 2 December 1985 Department of Geology The College of Wooster Wooster, Ohio 44691 USA •fV» t; 5 j :'v ;i '■'■■• / 4 ;> i.'.'i.>».:.,-^ If ,7 .1 • •:• V. '■■-.=• . 4 > 4 • > * ‘ fs-.", T.' u (. I y • *, s j£ Y/ f 4. r.'*’ SILICIFIED TRILOBITES OF THE FAMILY ASAPHIDAE FROM THE MIDDLE ORDOVICIAN OF VIRGINIA by R. p. TRIPP and w. r. evitt Abstract. Protaspis, meraspis, and holaspis stages of Isotehis giselae sp. nov. from the Edinburg Formation (Middle Ordovician) of Virginia are described; diagnostic features of the dorsal shield are the absence of a distinct lateral depression, and the short, strongly elevated palpebral lobes. Four moults during the protaspis period are dehned by changes in the hypostomes. Isotelus spp. A-E, from the lower and upper Lincolnshire, Oranda, and Martinsburg Formations (Middle Ordovician) of Virginia are based on less complete material. Nahannia sp. is described from the Lincolnshire Formation. The first description of silicified meraspis and holaspis Isotelus material from the Martinsburg Formation of Virginia was provided by Whittington (1941). Evitt (1961) gave a detailed account of the protaspis ontogeny of Isotelus, based on material from the lower Lincolnshire, Edinburg, and Martinsburg Formations. Hu (1975) described Isotelus protaspides (as Remopleuricles), meraspides, and holaspides from the Edinburg Eormation. In this paper, six species of Isotelus and one of Nahcumia are described from five Middle Ordovician horizons near Strasburg, Virginia (Table 1), with emphasis on the distinction between generic and specific developmental characters. The differences between the species are greatest at the protaspis stage. Mode of occurrence, preservation, and techniques employed are as described by Whittington and Evitt (1954, p. 5). Localities are as described by Whittington ( 1 959, p. 380) and Evitt ( 1 953, p. 34; 1 96 1 , p. 987). THE DEVELOPMENT OF ISOTELUS Protaspides Evitt (1961, pp. 988-990) described the earliest and latest stages of the protaspis period, and illustrated intermediate stages; his text-figs. 1-3 are partly based on more than one species. Four moults in the protaspis period are defined by coordinated and progressive changes affecting the TABLE I. Comparative frequency of occurrence of species described; VC, very common (more than 100 specimens); c, common (10-100 specimens); r, rare ( 1 -9 specimens). Formation Lincolnshire lower upper Edinburg Oranda Martinsburg Isotelus sp. A sp. B giselae sp. C sp. D sp. E protaspides c c vc r r c meraspides c c VC r c r holaspides r r c r c r Nahannia sp. sp. meraspides r r holaspides — '■ [Palaeontology, Vol. 29, Part 4, 1986, pp. 705-724, pis. 54-57. | 706 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1. A, Isotelus spp., length (mm)/frequency histograms of protaspis dorsal shields from the lower and upper Lincolnshire, Edinburg, and Martinsburg forma- tions. B, I. giselae sp. nov., width (tr.)/frequency distribution of protaspis hypostomes. Stages 1 -4, from the Edinburg Formation, Iocs. 3 and 4 (Evitt Collection). ten hypostomal characters listed on Table 2. In summary. Stage 1 is distinguished by the fused hypostomal suture, stumpy lateral and posteromedian spines, and rounded outline; Stage 2 by the abaxially open hypostomal suture and longer spines; Stage 3 by the tripartite open hypostomal suture, longer and more divergent lateral spines, and more oval outline; and Stage 4 by the bipartite open hypostomal suture, the posteromedian notch, and elongated posterior (fourth) lateral spines parallel to and matching the posteromedian spine. The hypostome became relatively smaller compared with the size of the dorsal shield in successive stages, with the result that the smallest hypostomes of each stage are the same size, but there is an increase in size range and average size (text-fig. 1b). Changes in the librigenae are described in the systematic descriptions below. HYPOSTOMAL ANTERIOR NECK ANTERIOR PROCESS OF DOUBLURE- MIDDLE BODY SECOND LATERAL SPiNES LATERAL AREA DOUBLURE POSTERIOR posteromedian EMBAYMENT SPINE TEXT-FIG. 2. Terminology applied to parts of the Isotelus protaspis hypostome (dorsal view of Stage 4). TRIPP AND EVITT: ASAPHIDTRILOBITES 707 Dorsal shields can only be distinguished as early or late, based on size, development of the glabella, and the course of the facial suture; the only present/absent feature is the bifid posteromedian cusp in the late stage (coaptative with the studs at the tips of the librigenae), but this is seen only in well- preserved specimens. There is no discontinuity in the size/frequency charts (text-fig. 1a), and the stages evidently overlapped in size. Early and late protaspis dorsal shields from the lower and upper Lincolnshire and Martinsburg Formations are compared with those from the Edinburg Formation on Table 3. All are strongly convex, and possess pairs of anterior and posterior spines. The differences between protaspides of the different species are considerable, affecting size, outline, convexity, strength of axial furrows, and length and position of the anterior and posterior spines, and are greater than at any subsequent stage. A few protaspis hypostomes are known from the lower Lincolnshire and Martinsburg Forma- tions; these conform to the four stages of the Edinburg material in the features summarized above, but with differences in other characters, such as the claw-like lateral spines of the Stage 1 hypostome from the Lincolnshire Formation (PI. 54, fig. 1) and the deep doublure of the Stage 4 hypostome from the Martinsburg Formation (PI. 54, fig. 20). Early and late stages of protaspis dorsal shields and ventral plates were recognized in I. parvirugosus Chatterton and Ludvigsen, 1976, by Chatterton (1980, pi. 2, figs. 1-11, 16, 30; pi. 3, figs. 5-14; text- fig. 3d, e). The general development of the ventral plate is comparable to that of the Edinburg Formation material, albeit with strong specific characters. The protaspis dorsal shield differs mark- edly from all the species described herein in lacking marginal spines, and librigenae differ in outline. We have examined the protaspides from the Eden Formation, western Covington, Kentucky, illustrated by Hu (1971, pi. 26, figs. 17-21; text-fig. 57a-d); these differ markedly from known Isotelus protaspides in their weak convexity and in other features. Gap in the ontogenetic series Evitt (1961, p. 991) described in detail the striking difference in gross morphology between prota- spides and meraspides of Isotelus; his observations have been confirmed in I. parvirugosus by Chatterton (1980). The largest protaspis from the Edinburg Formation is 113 mm long; the degree zero meraspis must have been about 2-0 mm long (PI. 56, fig. 17; text-fig. 4). This gap between protaspis and meraspis is exceptionally great amongst trilobites. Evitt suggested two possible explanations; Chatterton (1980, p. 17) supported the first, that the young asaphid passed through a metamorphosis during which greatly accelerated development was incompletely recorded in moulted exoskeletons, commenting: ‘The gap could be either caused by a period of rapid growth during this metamorphosis, or a longer period without forming a new exoskeleton during ecdysis.’ Evitt drew attention to the resemblance between asaphid and remopleuridid protaspides, but demonstrated that the discontinuity was much greater in the asaphid (Isotelus) ontogeny. Meraspides Differences between meraspides described herein are summarized on Table 3. The following changes occur in all known ontogenies of Isotelus. 1, cranidium: in the smallest cranidia the glabella is narrow; preglabellar, axial, and occipital furrows are strongly defined; with growth the glabella widens, the occipital furrow is effaced, and the preglabellar furrow becomes shallower; in the smallest cranidia LI is well defined, more strongly adaxially than abaxially; in successive moults the axial furrow deepens and SI dies out, LI being almost completely fused in the effaced glabella; the palpebral lobe becomes shorter, the anterior extremity shifting further back; both anterior and posterior branches of the facial suture become more divergent. 2, librigena: in the smallest specimens the vincular socket and panderian opening are absent; the socket starts to develop at an estimated cephalic length of 4-5 mm; the inclination becomes steeper, the eye becomes shorter, and the doublure becomes more narrow. 3, hypostome: becomes wider in proportions, and the posterome- dian notch becomes longer; the shoulder spine becomes less prominent and is finally lost. 4, transitory pygidium: becomes more elongate in proportions; the posteromedian notch, 10% the length of the transitory pygidium in the smallest specimens, is steadily reduced until lost at a pygidial length of about 2-3 mm. 708 PALAEONTOLOGY, VOLUME 29 Superimposed on this shared pattern of development, the following are the main characters which differ between species. 1, cranidium: proportions; width of preglabellar field; presence of an anteromedian ridge; size, position, and slope of the palpebral lobe. 2, librigena: width of border; degree to which the border furrow is retained; length of genal spines at comparable sizes. 3, hypostome: length and outline of the posteromedian notch; development of the posterior wing; sculpture and extent of the finely striate, bevelled inner slope of the doublure of the posterior forL 4, transitory pygidium: proportions; convexity; strength of furrows. In addition, the size at which changes in morphology take place varies between species (Table 3). The morphological developments outlined above were completed at an estimated skeletal length of 9 0 mm in /. giselae; this compares with a skeletal length of 9-43 mm for the smallest recorded holaspis of /. gigas Dekay (Whittington 1957, p. 445). The sagittal length of the dorsal shield of I. giselae as represented by the largest known part (PI. 57, fig. 16) is estimated at 100 mm., an increase in size of 130-fold throughout life. SYSTEMATIC PALAEONTOLOGY Repositories of specimens. BMNH, British Museum (Natural History), London; USNM, National Museum of Natural History, Smithsonian Institution, Washington, D.C.; Geological Museum, University of Cincinnati, Ohio; Hunterian Museum, Glasgow University. Locality data. Locality numbers are those of localities described by Whittington (1959) and Evitt (1953, 1961). Family asaphidae Burmeister, 1843 Subfamily isotelinae Angelin, 1854 Genus isotelus Dekay, 1824 Type species. I. gigas Dekay, 1824, Sherman Fall Formation, Trenton Falls, New York, U.S.A. Discussion. Formerly all the species described below would have been referred to Homotelus. Raymond (1920, p. 285) founded the genus Homotelus to include species which differed from typical Isotelus in lacking the concave lateral depression; he selected H. ulriclii Raymond, 1920, as type, a species in which the cephalic marginal depression is effaced mesially but is present laterally (Whit- tington 1950, p. 552, pi. 73, fig. 5). Whittington, De Mott (1963, p. 80), and Shaw (1968, p. 62) agreed on the unreliability of the partial absence of the lateral depression as a generic EXPLANATION OF PLATE 54 Figs. 1-8. Isotelus protaspis dorsal shields. 1 and 2, 1, sp. A, lower Lincolnshire Formation, loc. 1; 1, It 20000, Stage 1, with incomplete ventral plate adhering to inner surface, ventral view, x 55; 2, It 17276, late stage, ventral view, x 25. 3 and 4, 1, sp. B, upper Lincolnshire Formation, loc. la: 3, It 17286, early stage, ventral view, X 50; 4, It 17284, late stage, ventral view, x 25. 5 and 6, /. giselae sp. nov., Edinburg Formation: 5, It 17319, loc. 2, Stage 2, with incomplete ventral plate adhering to inner surface, ventral view, x 50; 6, It 17308, loc. 2, late stage, ventral view, x25. 7 and 8, /. sp. E, Martinsburg Formation, loc. 9: 7, It 17350, early stage, dorsal view, x 50; 8, It 17351, late stage, ventral view, x 25. Figs. 9-22. Isotelus protaspis ventral plates. 9-17, 7. giselae sp. nov., Edinburg Formation, loc. 3 except fig. 17 (loc. 4): 9, It 17328, Stage 1, ventral view, x 50; 10, It 17325, Stage 1, dorsal view, x 25; 11, It 17337, Stage 2, ventral view, x 25; 12, It 17326, Stage 2, ventral view, x 50; 13, It 17324, Stage 3, ventral view, x 50; 14, It 17315, Stage 3, dorsal view, x 50; 15, It 17322, Stage 4, ventral view, x 50; 16, It 17336, Stage 4, dorsal view, X 50; 17, It 17331, Stage 4, oblique ventral view (note ventral stud at tip of librigena), x 50. 18-20, 1. sp. E, Martinsburg Formation, loc. 10 except fig. 20 (loc. 12): 18, It 17354, Stage 3, dorsal view, x 50; 19, It 1 7355, Stage 4, oblique ventral view, x 50; 20, It 17361, Stage 4, oblique ventral view showing deep doublure, resembling a lateral border, x 135. 21 and 22, 7. sp. A, lower Lincolnshire Formation, loc. 1, specimen lost, ventral and left lateral views, x 36. Figs. 1, 21, 22, light micrographs by W. R. E. of whitened specimens; all others SEM micrographs. All BMNH specimens except figs. 2 1 , 22. PLATE 54 TRIPP and EVITT, Isotelus 710 PALAEONTOLOGY, VOLUME 29 criterion, and suggest the abandonment of Homotelus\ the evidence of the species deseribed herein supports this view. Isotelus giselae sp. nov. Plate 54, figs. 5, 6, 9-17; Plate 56, figs. 9-17; Plate 57, figs. 10-13, 15-18; text-figs. 3, 4, 5a-c, 6a 1942 Homotelus simplex (Raymond and Narraway); Butts, pi. 101, figs. 27, 28. 1961 asaphid indet., Evitt, p. 988, pi. 1 17, figs. 5-16, 19-21; pi. 1 18, figs. 1-34, 37-39. 1975 Remopleurides caelatus Whittington; Hu, pi. 2, figs. 3, 4, 6, 7, 20; pi. 3, figs. 21, 22, text-fig. Ic-E, G. 1975 Isotelus sp., Hu, p. 41, pi. 4, figs. 12-15, 20-23, 25, 26, 28-33; text-fig. 3h, m, n, p-s. Derivation of name. After Mrs. W. R. Evitt who, in sorting fine residues, was the first to suspect the true association of parts of the asaphid protaspis. Diagnosis. Lateral depression of exoskeleton indistinct, absent mesially. Minimum width of cranid- ium (at anterior extremities of palpebral lobes) 65-70 % sagittal length, and situated opposite midlength of cranidium. Palpebral lobe strongly elevated. Librigenal spine absent in full-grown cephala. Hypostome strongly sculptured. Holotype. USNM 398481 (inner layer of silicified cranidium, text-fig. 5a-c), locality 2 (Crabill Farm), Stras- burg, Virginia, USA; Lantz Mill facies, Edinburg Formation, Middle Ordovician. Collected by Mr Y. Kirk- patrick-Howat. The dimensions of the holotype cranidium are: length, 20-8 mm; minimum width (at anterior extremity of palpebral lobe), 14-6 mm; maximum preocular width, 16T mm; estimated width across palpebral lobes, 210 mm; estimated width at posterior margin, 30 0 mm. Other material. See Plate explanations for specimen details. Locs. 2-7, 14, and 0-6 km SE of Willow Grove Station, 4-8 km SW of Woodstock; Edinburg Formation, Middle Ordovician. Description. Cranidium gently convex, minimum width (at anterior extremity of palpebral lobe) 65-70 % sagittal length and 80-90 % maximum preocular width, situated opposite midlength of cranidium. Glabella fused with occipital ring, narrowing forwards to 65 % of its posterior width opposite anterior extremity of palpebral lobe, widening and becoming undefined anteriorly. Lateral glabellar lobes and furrows faintly impressed on holotype cranidium (in which only the inner layer of quartz is preserved): SI short and strongly oblique, S2 and S3 short transverse. Shallow abaxial preglabellar and occipital furrows distinguishable on holotype, plus a small posteromedian tubercle. Axial furrow broad and shallow posteriorly, dying out anterior to palpebral lobe. Basal apodemes small, rounded, 50 % maximum width of cranidium apart. Fixigena broad (tr.) posteriorly, exceptionally so on holotype, in which width between sutures at posterior margin estimated at 140 % length (sag.) of cranidium. Palpebral lobe 10 % cranidial length, posterior extremities slightly further apart than anterior; lobe strongly elevated, as tall as long (exsag.), sloping inwards with increasing steepness. Posterior border furrow effaced. Anterior branches of facial sutures diverge strongly outwards and forwards from eyes to 90 % width across palpebral lobes, opposite 30-35 % sagittal length of cranidium from front, thence curving inwards to midpoint intramarginally. Posterior branch curves gently outwards and backwards, at 50° to sagittal axis. External surface smooth or faintly pitted. Doublure of glabella comparatively short, narrowing out abaxially. Doublure of posterior border absent except for triangular area laterally. Largest librigena BMNH It 18810 (PI. 57, fig. 15), appropriate to a cephalon 26 mm in estimated length, slopes steeply outwards with even convexity; lateral border not demarcated. Eye not preserved. Subocular furrow broad and shallow. Lateral border absent. Genal angle narrowly rounded. Median connective suture open. External surface smooth or faintly pitted. Doublure 35 % cephalic length, horizontal, flattened; inner margin broadly embayed mesially to accommodate hypostome, forming cusp opposite anterior extremity of eye, thence curving inwards and backwards and bending inwards posteriorly to meet lateral doublure of posterior border of cranidium. Lateral depression narrows and deepens backwards, terminating abruptly in a vincular socket at 20 % estimated cephalic length anterior to genal angle. Panderian process represented by a thinning in test, often incompletely silicified, a short distance posterior and adaxial to vincular socket. Terrace lines strong, closely spaced, running parallel to adjacent margin, absent in vincular socket. Adult librigenae (PI. 57, fig. 11) up to 22 mm in estimated cephalic length, retain slender genal spine estimated at about 35 % length (sag.) of cephalon. Largest hypostome BMNH It 18814 (PI. 57, figs. 12, 13) 9-7 mm long (exsag.), as long as wide, weakly TRIPP AND EVITT:ASAPHIDTRILO BITES 71 1 convex. Anterior margin transverse mesially, curving downwards and backwards at subtrapezoidal anterior wing. Lateral margin strongly rounded. Middle body undefined, without independent convexity. Macula smooth, depressed, gently swollen, transversely oval, at 30 % length from front, 50 % maximum width of hypostome apart, rarely preserved. Median notch 45 % length (exsag.) of hypostome. Posterior fork narrows steadily backwards, tips 50 % maximum width of hypostome apart. Posteriorly, inner margin of fork diverges backwards at 20° to midline, slightly less divergent at front. Ventral surface with terrace lines as follows: wavy, transverse lines on middle body; fourteen or fifteen longitudinal lines abaxially; lateral seven or eight extend successively further beyond macula, running parallel to lateral margin; nine or ten lines parallel to apex of median notch; posteriorly and adaxially widely spaced faint lines branch to form faint network. Anterior margin of doublure transverse for 45 % width of hypostome just anterior to median notch, bending forwards sharply and extending forward to anterior wing abaxially. Posterior wing process just anterior to bend projects laterally and dorsally as a tongue-shaped projection. Doublure beneath fork swollen to form keel curving forwards from tip and slightly outwards, dying out anteriorly; inner slope flat, with about fifty faint, closely spaced, straight, parallel striae running aslant bevelled face, curving forwards near crest of keel; outer slope convex, with a few widely spaced raised lines. Number of thoracic segments unknown. Axis 40-45 % width of thorax, convex transversely. Axial furrow shallow. Pleural lobe downturned at fulcrum; articulating facet strong, marked off by faint, oblique furrow. Pleural furrow shallow, running obliquely from inner anterior corner, dying out at fulcrum. Articulating half ring short, ring furrow shallow; apodeme at abaxial extremity. Axial doublure about half length of ring. Pleural doublure about 50 % width to fulcrum for most of length, inner margin sinuously longitudinal. Terrace lines parallel to margin and closely spaced adaxially, oblique and widely spaced for most of width. Panderian TABLE 2. Isotelus giselae sp. nov. Edinburg Formation. Stages in the development of the protaspis librigena and hypostome. Stage 1 Stage 2 Stage 3 Stage 4 Librigena connective suture fused fused partly fused open ventral stud at tip absent absent present present (width between extremities) 40 20 5 5 % (maximum width across plate) Hypostome width (mm) 0-43 -0-49 0-43-0-53 0-43-0-57 043-0-59 mean width (mm) 0-45 0-48 0-49 0-50 number of specimens 7 23 131 94 (width) % (maximum width 55 55 50 50 across ventral plate) hypostomal suture fused partly fused tripartite bipartite (width of neck) % (width of 70 60 55 50 hypostome) middle body large narrower smaller small, convex hindmost lateral spine very short short, divergent long, divergent long, longitudinal posteromedian projection blunt stumpy spine short spine long, steep spine posteromedian outline bowed backwards convex backwards rounded notched height shallow shallow deep deep anterior process on doublure absent absent small long thickness of test thin thin thick thick Illustrations Evitt 1961, pi. 117, fig. 19 Evitt 1961, pi. 118, figs. 18,22-24 25, 26 27-29 21, 30-34, 37-39 Hu 1975, pi. 2, fig. 20 this paper, text-fig. 3 a B c D this paper, PI. 54, figs. 9, 10 11, 12 13, 14 15-17 712 PALAEONTOLOGY, VOLUME 29 opening adaxial to midwidth and anterior to midlength, covered by backwardly facing cowl. Doublure widens abruptly at back, extending to 75 % width to fulcrum, narrowing out adaxially. Largest pygidium BMNH It 18813 (PI. 57, fig. 16) 68 mm wide, incomplete posteriorly, but approximately 60 % as long as wide, gently convex. Axis extremely ill defined, 30 % maximum width of pygidium anteriorly. Axial furrow faint and broad, running inwards and backwards at 20° to midline. Articulating half ring and furrow short. Pleural lobe evenly and gently convex, with no change of convexity at border. Anterolateral facet strongly developed and extending for half width of pleural lobe. First pleural furrow deep and broad. A B 0-5 mm TEXT-FIG. 3. Isotelus giselue sp. nov., Edinburg Formation, Iocs. 3 and 4, protaspis ventral plates, a. Stage 1, hypostomal suture entirely fused, b. Stage 2, hypostomal suture fused mesially, functional laterally, c. Stage 3, hypostomal suture tripartite and functional, d. Stage 4, hypostomal suture bipartite. EXPLANATION OF PLATE 55 Figs. 1-32. Isoteliis sp. A, lower Lincolnshire Formation, loc. 1. 1-13, holaspides: 1 and 2, USNM 258078, cranidium, dorsal and left lateral views, x 8; 3, USNM 258079, right librigena, x 8; 4, 5, 9, USNM 258080, pygidium, left lateral, dorsal, and ventral views, x 8; 6-8, USNM 258081, thoracic segment, dorsal, posterior, and right lateral views, x 6-6; 10-13, USNM 258082, thoracic segment, dorsal, posterior, right lateral, and ventral views, x 6-4. 14-32, meraspides; 14-22, USNM 258083-258090, cranidia showing progressive effacement of anteromedian ridge and the preglabellar, occipital, and SI furrows, all x 10; 23- 25, USNM 258091, right librigena, right lateral, dorsal, and ventral views, x 10; 26, USNM 258092, hypostome, x 24; 27 and 28, USNM 258093, transitory pygidium, x 14; 29-32, USNM 258094, transitory pygidium, posterior, right lateral, and dorsal views, x 12, ventral view, x 13. Figs. 33-35. Nahannia sp. 33, BMNH It Mill, lower Lincolnshire Formation, loc. 1, transitory pygidium, X 10. 34 and 35, upper Lincolnshire Formation, loc. \a, small pygidia: 34, BMNH It 17304, x 10; 35, BMNH It 17305, x 10. Figs. 1-32, light micrographs by W. R. E.; figs. 33-35, SEM micrographs. PLATE 55 TRIPP and EVITT, Isotelus, Nahannia 714 PALAEONTOLOGY, VOLUME 29 marking off a swollen first half-pleura. Six short, unfurrowed pleural ribs, dying out before midwidth of pleural lobe; rib furrows successively shorter and fainter towards back. External surface smooth. Doublure of uniform width, 25 % length of pygidium at sagittal axis; anterior margin simply rounded, not embayed mesially. Elongate process at anterolateral corner (fitting with vincular notch of librigena). Terrace lines faint, running subparallel to margin. Smaller pygidia relatively narrower, 65 % as long as wide, axis 80 % length (Hu 1975, pi. 4, figs. 30, 32,33). Description of prolaspis ventral plates. Stage 1 (PI. 54, figs. 9, 10). Librigenae and hypostome fused; median connective and hypostomal sutures absent. Width of librigena at midlength 10 % width of ventral plate. Librigena narrows steadily backwards, distance between tips 40 % width across ventral plate. Hypostome 0-43-0-49 mm wide (all hypostomal width measurements exclude lateral spines), 55 % maximum width of ventral plate. Neck continuous with librigena, 70 % width of hypostome. Middle body large with slight independent convexity. Lateral area crescentic in outline, sloping slightly inwards. Lour outwardly directed lateral spines, length one to three times basal width, subequal in size. Posteromedian margin projects back- wards in blunt, broad-based marginal projection. Doublure shallow, almost vertical, continuous with narrow doublure of librigena; anterior dorsal process absent. Test thin. Surface smooth. Stage 2 (PI. 54, figs. 1 1 and 12). Differs from Stage 1 as follows. Librigenae and hypostome fused mesially; median connective suture absent; hypostomal suture open laterally, curving outwards and backwards. Distance between tips of librigenae 20 % maximum width of ventral plate. Hypostome 0-43-0-53 mm wide, neck 60 % width of hypostome. Middle body smaller and more strongly defined. Lateral spines more erect and more splayed, length four to five times basal width. Posteromedian outline rounded; posteromedian spine as wide as long, projecting outwards and upwards at base of neck. Stage 3 (PI. 54, figs. 13 and 14). Median connective suture partially fused. Distance between tips of librigenae 5 % maximum width of ventral plate. Small rounded stud at tip of librigena projects ventrally (fitting with bifid process on posteromedian margin of dorsal shield). Hypostomal suture open (occasionally secondarily fused), tripartite, transverse mesially, concave forwards laterally. Hypostome 0-43-0-57 mm wide, 50 % maximum width of ventral plate. Neck 55 % width of hypostome. Middle body narrower. Lateral area slopes more strongly inwards. Lateral spines more erect, more splayed, hindmost (fourth from front) divergent. Posteromedian spine longer. Doublure deeper, narrowing out at base of neck where dorsal projection larger. Test thicker. Stage 4 (PI. 54, figs. 15-17). Median connective suture open. Hypostomal suture bipartite and monocuspid. Ventral stud at tip of librigena larger (PI. 54, fig. 17). Hypostome 0'43-0-59 mm wide, 50 % maximum width of ventral plate. Neck 50 % width of hypostome. Middle body smaller and more convex. Posteromedian spine longer, more slender, and sloping forwards more steeply. Fourth lateral spine parallel and equal to posteromedian spine in slope and length. Posterior outline of doublure notched for 25 % maximum width of hypostome; anterolateral projection longer and stouter. Description of protaspis dorsal shields. Early stages (1 and 2). Dorsal shield BMNH It 17319 (PI. 54, fig. 5) length 0'75-c.0'85 mm, almost as long as wide, strongly convex, recurved for about 40 % length (sag.); anterior margin simple, convex forwards; posterior margin monocuspid forwards. Glabella short, with slight independent convexity. Axial furrow indistinct, but represented by pit at 10 % length of protaspis from front. Anterior spine horizontal, 25 % length of protaspis, tips at 95 % width of protaspis apart, slender, diverging forwards at 35° to midline. Posterior spine directed longitudinally, 25 % length of protaspis, sloping ventrally at 40° to plane of anterior spine, tips at 30 % width of protaspis from midline. Librigena ventral except for small anterolateral surface, without eye. EXPLANATION OF PLATE 56 Figs. 1-8. Isotelus sp. B, upper Lincolnshire Formation, loc. la; meraspides except fig. 3. 1-3, cranidia: 1, It 17219, X 10; 2, It 17291, x 10; 3, It 17292, small holaspis, x 10. 4 and 5, right and left librigenae; 4, It 17306, x 10; 5, It 17293, x 10. 6 and 7, hypostomes, dorsal views: 6, It 17297, x25; 7, It 17298, x25. 8, It 17300, transitory pygidium, x 25. Figs. 9-17. I. giselae sp. nov., Edinburg Formation, loc. 3, except fig. 14 (loc. 4), meraspides. 9-11, cranidia: 9, It 17339, X 25; 10, It 17344, x 25; 1 1, It 17343, x 10. 12, It 17349, left librigena, x 25. 13-15, hypostomes: 13, It 17340, dorsal view, x 50; 14, It 17341, x 50; 15, It 17338, ventral view, x25. 16, It 17347, transitory pygidium, x 10. 17, It 17346, disassociated exoskeleton, probably degree zero, x 25. All SEM micrographs of BMNH specimens. PLATE 56 TRIPP and EVITT, Isotelus 716 PALAEONTOLOGY, VOLUME 29 A B TEXT-FIG. 4. Reconstruction of degree zero meraspis of Isotelus giselae sp. nov., Edin- burg Formation, loc. 3. a, dorsal view, b, ventral view. TEXT-FIG. 5. A-c, Isoleliis giselcw sp. nov., Edinburg Formation, loc. 2; USNM 398481 holotype cranidium, dorsal, posterior, and right lateral views, x 2. D, /. sp. B, upper Lincolnshire Formation, loc. la; USNM 398482, cranidium, x 1-5. Photographs prepared by the National Museum of Natural History, Washington. Late stages (3 and 4). Dorsal shield (PI. 54, hg. 6; PI. 57, fig. 18) length c.0-80-113 mm, wider than early stage anteriorly, deeper posteriorly, anterior margin and outline more sinuous due to stronger development of glabella and facial suture; posterior margin with bifid median cusp (fitting with stud at tip of librigena). Glabella 20 % length of protaspis, with moderate independent convexity. Axial furrow shallow, deepening towards front, but shallowing near anterior margin. Anterior spine 30 % length of protaspis, slender, tips 1 10 % width of protaspis apart, diverging at 45° to midline. Posterior spine 30 % length of protaspis, directed downwards and slightly outwards at 70° to plane of anterior spine. Dorsal area of librigena larger, indenting anterolateral margin of dorsal shield. Description of meraspis. Smallest cranidium BMNH It 17339 (PI. 56, fig. 9) 11 mm long, minimum width 70 % sagittal length and 95 % maximum preocular width. Glabella circumscribed, gently convex, waisted near midlength, widening posteriorly, maximum preocular width 55 % length. Occipital ring more than 10 % length and 45 % preocular width of cranidium, bowed backwards. Occipital furrow broad and shallow. Axial furrow shallow anteriorly, deepening backwards. Preglabellar field 10 % length of cranidium. Fixigena weakly convex; LI with slight independent convexity, 20 % length of cranidium. SI much deeper than axial furrow alongside LI. Palpebral lobe 35 % cranidial length, strongly rounded in outline, anterior extremity at 40 % length from front. Posterior border furrow shallow, continuous with lateral border furrow. Anterior branch of facial suture diverges forwards before curving inwards to midline. Posterior branch curves outwards and backwards. Doublure of occipital ring extends half way to occipital furrow; doublure of posterior border absent adaxially, developing abaxially. Cranidium BMNH It 17344 (PI. 56, fig. 10) 1-7 mm long, minimum width 60 % sagittal length and 80 % maximum preocular width. Glabella convex, clearly defined, narrowing backwards. Occipital TRIPP AND EVITT: ASAPHID TRILOBITES 717 ring 60 % preocular width of cranidium. LI and SI distinct, but less well defined. Palpebral lobe 35 % cranidial length, anterior extremity opposite midlength of cranidium. Indistinct preoccipital tubercle slightly off-centre, present only on this specimen. Cranidium BMNH It 17343 (PI. 56, fig. 11)31 mm long, preocular width 85 % palpebral width. Glabella defined anteriorly only by change in convexity. Axial furrow deepens backwards. Occipital ring 65 % preocular width of cranidium. Occipital furrow faint. SI effaced. Palpebral lobe 25 % length (sag.) of cranidium. Smallest librigena BMNH It 17349 (PI. 56, fig. 12) 11 mm estimated cephalic length, wide, sloping gently outwards. Eye rounded in outline, not marked off from cheek, convex. Lateral border uniformly narrow, half width of field opposite eye, continuous with posterior border and genal spine. Genal spine slender, pointed, 45 % estimated length of cephalon. Doublure 20 % estimated length of cephalon mesially, wide, inner margin curving backwards and inwards, forming a cusp, thence curving outwards and backwards, bending strongly inwards posteriorly to abut against abaxial doublure of posterior border of cranidium; vincular socket and panderian opening absent. With increase in size librigena slopes more steeply outwards, narrower; lateral border furrow obsolete. Vincular socket develops slowly between lengths 4-5 mm and 6 0 mm. Smallest hypostome BMNH It 17340 (PI. 56, fig. 13) 0-65 mm long (exsag.), 85 % as wide as long. Anterior margin broadly rounded. Middle body weakly swollen; middle furrow at 35 % length from front, running for a short distance inwards and backwards, more distinct on internal than on external surface; posterior lobe triangular, with slight independent convexity, 30 % width of hypostome. Lateral border runs longitudinally; shoulder forms prominent outwardly directed spine opposite midlength of hypostome. Posterior forks narrow steadily to acute points, 60 % width of hypostome apart. Median notch 30 % length (exsag.) of hypostome. Anterior margin of doublure convex forwards anterior to median notch, convex backwards at base of fork, curving obliquely forwards to anterior wing; posterolateral process absent; bevelled inner slope of fork striate. Hypostome BMNH It 17338 (PI. 56, fig. 15) 1-4 mm long (exsag.), middle body and posterior lobe undefined, greatest width at anterior wing; shoulder spine less protuberant, cuspid. Median notch 35 % length (exsag.) of hypostome. Smallest transitory pygidium 0-9 mm long, length 65 % width, convex, posteromedian notch 10 % length of pygidium. Axis 30 % maximum width and 85 % length, narrowing slowly backwards. Pleural lobe convex adaxially, with a concave area widening towards back abaxially. Eight segments, increasingly ill-defined towards back. Rib furrows broader but shorter than inter-rib furrows. Doublure extends in an even curve to tip of axis, lying close beneath dorsal surface; terrace lines faint and widely spaced. Transitory pygidium BMNH It 17347 (PI. 56, fig. 16) 1-9 mm long (sag.), length 80 % width, posteromedian notch small, adaxial area of pleural lobe more strongly convex, five segments clearly defined. With increase in size to 2-3 mm, posteromedian notch dies out, and segmentation is effaced. Discussion. All silicified material from the Edinburg Formation appears to belong to a single new species of Isotelus. The numerous protaspides from various localities conform closely with the figured specimens. The variation in meraspis cranidia is limited, e.g. a posteromedian tubercle is present on one specimen (BMNH It 17344; PI. 56, fig. 10). Larger cranidia are closely similar in linear measurements, except that the holotype is comparatively broad. Librigenae are consistent in character, subject to changes with growth. All hypostomes are identical, even as regards the fine detail of the sculpture, and we regard this as weighty evidence. The single large incomplete pygidium BMNH It 18813 (PL 57, fig. 16), 68 mm in width, appears to have been shorter in proportions than smaller specimens; this difference might be attributable to growth, although Whittington (1957, p. 445, figs. 25-27) found that the proportions of I. gigas altered little with development. The crack-out cranidium and pygidium illustrated by Butts (1942, pi. 101, figs. 27 and 28), from the Edinburg Formation, near Strasburg, appear to be conspecific with the silicified material, the only difference being that the axis of the pygidium is more strongly outlined. Raymond (1920, p. 288; 1925, p. 88, pi. 4, figs. 1-3) described Homotehis elongatus from the Echinosphaerites and Nidulites Beds of the Edinburg Formation. I. giselae differs in having the palpebral lobe shorter (exsag.) and more strongly elevated, and the pygidium smooth, not pitted. Wilson (1947, p. 22, pi. 1, fig. 5) figured an extended dorsal shield from the Hull Beds at Chaudiere Falls, Hull, Quebec, as //.? elongatus Raymond. This specimen differs from both /. elongatus and I. giselae in the more elongate and more convex pygidium. I. giselae bears a general resemblance to /. simplex (Rayond and Narraway 1910, p. 51, pi. 16, figs. 6-8; De Mott 1963, pi. 4, figs. 1-22) from the Platteville Group, but difl^ers conspicuously in the elevated palpebral lobes. 718 PALAEONTOLOGY, VOLUME 29 /. parvirugosus Chatterton and Ludvigsen (1976, p. 21, pi. 2, figs. 1-42) from the Esbataottine Formation, has similarly elevated palpebral lobes, but the presence oflateral and posterior cranidial borders precludes any possibility of a close relationship to I. giselae. An occipital and three pairs of lateral glabellar furrows, comparable to those in our holotype, are distinguishable in I. parvirugosus. I. giselae is similar to the type species, I. gigas Dekay (1824, p. 176, pi. 12, fig. 1; pi. 13, fig. 1), in eye position, in the absence of genal spines in the adult, and in hypostomal construction (Ross 1967, pi. 1, figs. 2, 6-9). I. giselae differs markedly in that the lateral depression is effaced, the cranidium expands more strongly anteriorly, the fixigena are much wider posteriorly, the eyes are elevated, and the pygidium is much wider. Isotelus sp. A Plate 54, figs. 1, 2, 21, 22; Plate 55, figs. 1-32 1950 hypostome indet., Evitt, pi. 2, figs, \la-d. 1961 asaphid indet., Evitt, pi. 117, figs. 1-4; pi. 1 18, figs. 35, 36. Material. See plate explanations for specimen details. Loc. 1, lower Lincolnshire Eormation, Middle Ordovician. Description. Differs from /. giselae as follows: palpebral lobes not elevated; hypostome (Evitt 1961, pi. 118, fig. 36) with tips of forks 55 % (cf. 50 %) maximum width of hypostome apart and longitudinal (cf. transverse) raised lines on the middle body. The smallest protaspis, BMNH It 20000, is a dorsal shield 0-5 mm in length, with the hypostome (fused to part of the librigena) adhering to the inner surface (PI. 54, fig. 1 ). The hypostome shows the distinctive features of Stage 1 as described in the Edinburg Formation material, particularly the short lateral spines, stumpy posteromedian spine, and broad neck fused with the librigenae. Interestingly, it differs significantly and recognizably in the short, claw-like, first and second lateral spines. A single hypostome with the posteromedian notch of Stage 4 is illustrated (PI. 54, figs. 21, 22). Dorsal shields of late protaspides differ from /. giselae in size range (Table 3), longer axial furrows, and shorter marginal spines, placed closer together. A comparison of meraspides with other species described herein is summarized in Table 3. The series of meraspis cranidia illustrated (PI. 55, figs. 14-22) clearly demonstrates the gradual effacement of the pregla- bellar, occipital, and SI furrows, and fusion of El with the rest of the glabella. The most distinctive feature is the strong anteromedian ridge in small meraspis cranidia, which is gradually effaced with growth. This ridge is much longer in some specimens (PI. 55, figs. 16 and 17) than in others, but we consider this to be only a variational distinction. Discussion. The cranidium not quite 4 0 mm in sagittal length (PI. 55, figs. 1 and 2) is by far the largest, though fragments indicate larger individuals. Breakage was due to mechanical stresses before preservation, and was not the result of the recovery process— all the lower Lincolnshire trilobites show similar signs of fragmentation, the cause of which is not known. Not so in the Edinburg material, in which extremely careful etching has led to the recovery of a few large and articulated specimens. EXPLANATION OF PLATE 57 Figs. 1-8, 14. Isotelus sp. D, Martinsburg Formation, loc. lOu; meraspides except fig. 3. 1-3, cranidia: 1, It 17377, X 25; 2, It 17375, x 10; 3, It 17374, small holaspis, x 5. 4, It 17382, right librigena, ventral view showing absence of vincular notch, x 10. 5, It 17372, hypostome, ventral view, x 10. 6 and 7, transitory pygidia: 6, It 17379, x 25; 7, It 17384, x 10. 8, It 17385, right librigena, ventral view, x 6. 14, It 17366, hypostome, ventral view, x 5. Fig. 9, /. sp. B, upper Lincolnshire Formation, loc. la; It 19068, cranidium, ventral view, x 5. Figs. 10 13, 15-18. I. giselae sp. nov., Edinburg Formation, loc. 3. 10, It 1881 1, cranidium, x 2. 11, It 18812, largest librigena retaining genal spine, ventral view, x 2. 12 and 13, It 18814, hypostome, ventral and dorsal views, X 3. 15, It 18810, adult right librigena, x 2. 16, It 18813, large pygidium, x 1. 17, It 17345, small hypostome, dorsal view showing finely striated, bevelled inner slope of doublure of fork, x 30. 18, It 17310, protaspis dorsal shield, ventral view, late stage, x 30. All SEM micrographs of BMNH specimens. PLATE 57 TRIPP and EVITT, Isotelus 720 PALAEONTOLOGY, VOLUME 29 TABLE 3. Isotehis spp. from Virginia. Comparison of protaspides and meraspides. Averages of early and late stages are separated by a solidus. Species A B giselae D Formation Lincolnshire Edinburg Martinsburg lower upper Protaspis range in dorsal length (mm) 0-5/1 0 0-6/1-0 0-75/1-13 0-75/1-05 late protaspis: (length of axial furrow) % 35 20 20 50 (length) late protaspis: (length of spines) % (length) 10 20 25 20 (width between tips of anterior spines) % 55/75 95/55 95/110 75/80 (protaspis width) (length of ventral aperture) % (length of 65 60 60 60 protaspis) sculpture smooth smooth smooth pitted Meraspis cephalon range in length (mm) 1 -4/2-0 -/2-2 1-1/3-1 1-5/4-5 (width across palpebral lobe) % (length of 85/90 95/90 115/90 -195 cranidium) (maximum preocular width) % (pal- 85/95 80/100 75/85 -175 pebral width) (basal width of glabella) % (preocular 65/70 60/70 45/65 45/70 width) (length of anterior extremity of palpebral 45/50 -/50 40/50 50/50 lobe from front of glabella) % (length of glabella) preglabellar ridge strong slight strong moderate slope of palpebral lobe moderate slight strong moderate (length of genal spine) % (estimated 55/45 47/50 45/- - cephalic length) (length of median suture) % (estimated 35/- -/35 301- 30/- length of cephalon) Hypostome (sagittal length) % (exsagittal length) 60 55 75 65 Transitory pygidium (sagittal length) % (width) 60/60 65/80 60 5 (anterior width of axis) % (maximum 30/40 30/- 30/35 25/- width of pygidium) sagittal length to which posteromedian 1-6 1-8 1-9 ? notch retained sculpture smooth smooth smooth pitted Isotelus sp. B Plate 54, figs. 3 and 4; Plate 56, figs. 1-8; Plate 57, fig. 9; text-fig. 5d Material. See Plate explanations for specimen details. Loc. 1 A, upper Lincolnshire Formation, Middle Ordovician. Description. Differs from I. giselae as follows: largest cranidium (text-fig. 5d), 29 mm in sagittal length, minimum width 80 % sagittal length, situated posterior to midlength of cranidium; surface finely pitted; hypostome with broader apex to notch and reduced posterior wings. Protaspides differ in their shorter marginal spines, and much more broadly rounded and swollen posterior area. Meraspis cranidia differ in the TRIPP AND EVITT: ASAPH ID TRILOBITES 721 broader glabella and stronger convexity; palpebral lobe somewhat elevated in a late meraspis cranidium, BMNH It 17291 (PI. 56, fig. 2). Shoulder spine of meraspis hypostome further forward; transitory pygidium more weakly segmented. Discussion. The adult cranidium closely resembles /. giselcie (text-fig. 5a-c) but the course of the anterior branch of the facial suture is less angular. There is a great difiference between the protaspides of the two species, and meraspis parts are readily distinguishable (Table 3). Isotelus sp. C Text-fig. 6b Material. Loc. 8, Oranda Formation, Middle Ordovician. Discussion. Adult hypostomes are the only parts sufficiently well preserved for comparison, and they establish the distinctness of this species (text-fig. 6); they differ from those of the other species described in being more elongate, width 75 % length (exsag.); the posteromedian notch is 40 % the length (exsag.) of the hypostome, and the tips of the forks are 60 % the maximum width of the hypostome apart. Protaspides and meraspides are poorly preserved. TEXT-riG. 6. Reconstructions of adult Isotelus hypostomes (left, ventral view; right, dorsal view). A, 7. gi.selae sp. nov., Edinburg Formation, b, 7. sp. C, Oranda Formation, c, 7. sp. D, Martinsburg Formation. D, 7. sp. E, Martinsburg Eormation. Isotelus sp. D Plate 57, figs. 1-8, 14; text-fig. 6c 1961 asaphid indet., Evitt, pi. 117, figs. 17, 18,22,23. Material. See Plate explanation for specimen details. Loc. lOr/ (Evitt 1961, p. 987), Martinsburg Formation, Middle Ordovician. Description. As in the Oranda material, the only adult part adequately preserved for comparison is the hypostome, which differs from that of 7. giselae in its less convex lateral outline, smaller anterior and posterior wings, and light sculpture. Late protaspides 0-9 mm long (sag.) illustrated by Evitt (1961, p. 117, figs. 17 and 18) differ from other species in the long axial furrows and the traces of pygidial segmentation. Furthermore, 722 PALAEONTOLOGY, VOLUME 29 the spines are extremely short. Small meraspis cranidia differ from I. giselae in the presence of the anteromedian ridge; they resemble /. sp. A in this respect, though the ridge is fainter. Discussion. This species most closely resembles /. sp. A in the narrow early meraspis glabella and the anteromedian ridge. It is possible that the material described by Whittington (1941, p. 512, pi. 75, figs. 27-45, 47) is conspecific but the axis of the transitory pygidium is less strongly defined posteriorly, and the anterior margin of the hypostomal doublure is transverse, not bowed forwards, in I. giselae. As Evitt (1961, p. 988) has mentioned, the only other genera of trilobites occurring at this locality are a calymenid and a cheirurid, leaving Isotehis as the only candidate for the protaspides assigned to it. Isotehis sp. E Plate 54, figs. 7, 8, 18-20; text-fig. 6d Material. See Plate explanation for specimen details. Loc. 9, 10, 11, 12, and 2 of Evitt (1953, p. 34). Description. Adult hypostomes differ from those of /. sp. D in the larger anterior and posterior wings, and in the stronger keel on the doublure of the posterior fork. Early protaspides differ from other species described in being more triangular and strongly pitted, a sculpture which is retained in some degree throughout development. The late protaspis is broadly rounded posteriorly, the glabella is much shorter than in the foregoing Martinsburg Eormation species, and there is no trace of segmentation. Protaspis hypostomes agreeing with Stages 2, 3, and 4 of the Edinburg Eormation material occur; the late stages differ in the deeper doublure, giving the lateral spines a gun-turret appearance (PI. 54, fig. 20). Meraspis parts of this species are not clearly distinguishable from the foregoing species. Genus nahannia Chatterton and Ludvigsen, 1976 Type species. Nahannia liiimilisiilcata Chatterton and Ludvigsen, 1976, Esbataottine Formation, District of Mackenzie, Canada. Nahannia sp. Plate 55, figs. 33-35; text-fig. 7 Material. See Plate and text-figure explanations for specimen details. Loc. 1, lower Lincolnshire Formation. Loc. 1«, upper Lincolnshire Formation, Middle Ordovician. Description. Cranidia L5-3-8 mm long with glabella and occipital ring faintly and decreasingly defined, narrowing opposite anterior extremity of palpebral lobe. Minimum width 60-70 % sagittal length, and 80- 90 % preocular width, situated opposite midlength of cranidium. Palpebral lobe 30 % length of cranidium, horizontal. Anterior branches of facial suture diverge to 85 % palpebral width, thence curving to midline. Posterior branch curves backwards and outwards, cutting posterior margin at a steep angle. Hypostome BMNH It 19063 (text-fig. 7e), the largest part found, 6-8 mm long (exsag.), 75 % long as wide, weakly convex. Anterior margin gently concave forwards, curving downwards and backwards at anterior wing. Lateral margin indented posterior to wing, strongly rounded in outline anteriorly, weakly so for most of length. Middle body undefined. Macula eroded, situated opposite lateral indentation at about midwidth of hypostome. Inner margin of fork curved gently inwards. Doublure extends forward from fork to anterior wing laterally. Anterior margin transverse for less than half width of hypostome anterior to median embay- ment, curving obliquely and sinuously forwards and outwards to anterior wing. Longitudinal keel on fork curves sigmoidally forwards from tip, dying out anteriorly; inner slope flat, bevelled, bearing faint, closely spaced, straight, parallel raised lines slanting forwards and inwards. Transitory pygidia incorporating one thoracic segment (PI. 55, figs. 33 and 34; text-fig. 7g) 1 •9-2-5 mm long, as long as wide; inner area with gentle independent convexity, outer area slightly concave. Axis bowed forwards for 45 % anterior width, narrowing steadily to apex at 70 % length of pygidium from front, indistinctly defined by stronger convexity. Posterior margin broadly rounded. Articulating half ring not seen. Articulating facet distinct but small. Doublure 30 % length (sag.) of pygidium, anterior margin slightly embayed for 5 % width; terrace lines closely spaced, faint, gently convergent. Small holaspis pygidium BMNH It 17305 (PI. 55, fig. 35) 2-8 mm long, 60 % as long as wide, axis 40 % maximum width, defined anteriorly only. Inner area moderately convex, with a well-marked concave depression setting off a weakly convex abaxial area. Doublure TRIPP AND EVITT: ASAPHID TRILOBITES 723 TEXT-FIG. 7. Nahanuia sp. a-f, upper Lincolnshire Formation, loc. \a. a-d, meraspis cranidia; a. It 17288, x25; B, It 19056, x 20; c, It 19057, x 18; d. It 17290, x 10. e. It 19063, hypostome, ventral view, x4-5. F, It 17305, small pygidium, ventral view (dorsal view, PI. 55, fig. 35), x 13. G, lower Lincolnshire Formation, loc. 1; It 17277, transitory pygidium, ventral view (dorsal view, PI. 55, fig. 33), x 16. All BMNH specimens. 25 % length (sag.) of pygidium, anterior margin slightly embayed mesially; terrace lines faint, oblique. Dorsal surface smooth externally. Discussion. The adult hypostome described above differs from Isotelus and corresponds with Nahan- nia in being relatively wider, and in the deeper posteromedian embayment, which exceeds 50 % of the sagittal length of the pygidium. It differs from the hypostome of the type species, N. humilisulcata Chatterton and Ludvigsen (1976, p. 25, pi. 3, figs. 26-29, 32, 33, 36, 37) from the Chazyan part of the Esbataottine Formation, in its greater width, and in the tips of the forks being closer together. The meraspis cranidia resemble N. humilisulcata and differ from Isotelus in the wide glabella and long palpebral lobes; the distinctive feature of adult Naliamiia cranidia, that the preocular width exceeds the maximum palpebral width, is not evident in the meraspis. Lincolnshire Formation juvenile cranidia differ from N. humilisulcata in the narrower preocular width, and more forward position of the palpebral lobe; pygidia differ in their weaker convexity and absence of segmentation. Chatterton and Ludvigsen listed other occurrences of Nahannia in North America, ranging up to the Richmondian. Acknowledgements. We are most grateful to the following for helpful suggestions: B. D. L. Chatterton, R. A. Fortey, R. Ludvigsen, and J. T. Temple. We thank the staff of the SEM Department, British Museum (Natural History) for preparing the micrographs. Finally, we offer our special thanks to the anonymous referee for constructive recommendations. REFERENCES ANGELIN, N. p. 1854. PalaeontologicQ Scandinavica 1: Crustacea formationis transitionis. Ease. 2, i-ix, 21-92, pis. 25-51. Lund. BURMEISTER, H. 1843. Die Organisation der Trilobiten. viii-l- 147 pp., 6 pis. Berlin. 724 PALAEONTOLOGY, VOLUME 29 BUTTS, c. 1941. Geology of the Appalachian Valley in Virginia. Va. geol. siirv. Bull. 52 (2), 1-271, pis. 73-95. CHATTERTON, B. D. E. 1980. Ontogenetic studies of Middle Ordovician trilobites from the Esbataottine Forma- tion, Mackenzie Mountains, Canada. Palaeontographica, A171, 1-74, pis. 1-19. and LUDViGSEN, R. 1976. Silicified Middle Ordovician trilobites from the South Nahanni River area, District of Mackenzie, Canada. Ibid. A154, 1-106, pis. 1-22. DELAY, j. E. 1824. Observations on the structure of trilobites and description of an apparently new genus. Ann. Lyceum not. Hist. 1, 174-189, pis. 12, 13. DE MOTT, L. L. 1963. Middle Ordovician trilobites of the upper Mississippi Valley. Ph.D. thesis (unpubl.). Harvard University. EVITT, w. R. 1950. Trilobites from the Lower Lincolnshire Limestone near Strasburg, Shenandoah County, Virginia. Ph.D. thesis (unpubl.), Johns Hopkins University. 1953. Observations on the trilobite Cerunis. J. Paleont. 27, 33-48, pis. 5-8. 1961. Early ontogeny in the trilobite family Asaphidae. Ibid. 35, 986-995, pis. 1 17 and 1 18. HU, CHUNG-HUNG. 1971. Ontogeny and sexual dimorphism of Lower Paleozoic Trilobita. Palaeontogr. am. 7, 1-155, pis. 7-26. 1975. Ontogenies of two species of silicified trilobites from the Middle Ordovician, Virginia. Trans. Proc. palaeont Soc. Japan, (n.s.), 97, 32-47, pis. 2-4. RAYMOND, p. E. 1913. Notes on some new and old trilobites in the Victoria Memorial Museum. Bull. Victoria meml Mus. 1, 33-80, pis. 3-6. 1920. Some new Ordovician trilobites. Bull. Mus. comp. Zool. Harv. 64, 273-296. 1925. Some trilobites of the lower Middle Ordovician of eastern North America. Ibid. 67, 1-180, pis. 1-10. — and NARRAWAY, J. E. 1910. Notes on Ordovician trilobites III. Asaphids from the Lowville and Black River. Ann. Carneg. Mus. 7, 35-45, pis. 15 and 16. ROSS, J. R. 1967. Calymenid and other Ordovician trilobites from Kentucky and Ohio. Prof. Pap. U.S. geol. Surv. 583B, B1-B18, pis. 1-5. SHAW, F. c. 1968. Early Middle Ordovician Chazy trilobites of New York. Mem. N.Y. St. Mus. Sci. Serv. 17, 1 163, pis. 1-24. WHITTINGTON, H. B. 1941. Silicified Trenton trilobites. J. Paleont. 15, 492 522, pis. 72-75. 1950. Sixteen Ordovician genotype trilobites. Ibid. 24, 531-565, pis. 68-75. — 1957. The ontogeny of trilobites. Biol. Rev. 32, 401-469. 1959. Silicified Middle Ordovician trilobites: Remopleurididae, Trinucleidae, Raphiophoridae, Endy- mionidae. Bull. Mus. comp. Zool. Harv. 121, 371-496, pis. 1-36. and EVITT, w. r. 1954. Silicified Middle Ordovician trilobites. Mem. geol. Soc. Am. 59, 1-137, pis. 1-33. WILSON, A. E. 1947. Trilobites of the Ottawa Formation of the Ottawa St. Lawrence Lowland. Bull. geol. Surv. Can. 9, 1-86, pis. 1 10. R. P. TRIPP British Museum (Natural History) Cromwell Road London SW7 5BD Typescript received 12 June 1985 Revised typescript received 16 December 1985 W. R. EVITT Stanford University Stanford California, U.S. A. INTERNAL MOULD MARKINGS IN A CRETACEOUS AMMONITE FROM NIGERIA by P. M. P. ZABORSKI Abstract. Linear and concentric markings are described from steinkerns of the Upper Cretaceous ammonite Paravascoceras from Nigeria. The markings are intimately associated with the lobules of the suture lines. They probably record adapical projections of a preseptal prismatic zone of shell material which were secreted by those parts of the mantle corresponding to the lobules during mantle translocation. The posterior mantle margin was under muscular attachment at these points during translocation and subsequent septal secretion. Translocation itself probably took place in two stages: an initial phase of rapid mantle movement during which liquid entering the new chamber space was derived from the main body tissues; and a second, longer, phase of gradual translocation during which existing cameral liquid was transferred to the new chamber space. Various markings are known from ammonite internal moulds. Most are confined to the body chamber and have generally been interpreted as the sites of muscle attachments. The following four types of marking occur repeatedly: 1. A pair of symmetrically disposed tongue-shaped or circular impressions in the posterodorsal part of the body chamber (see Crick 1898, Jones 1961, Jordan 1968, Kennedy and Cobban 1976), usually interpreted as the scars of retractor muscles similar to those in Nautilus (see Mutvei 1957). Jordan (1968) and Bayer (1970) described an associated dark band running around the umbilical shoulders which they interpreted as marking the successive positions of these muscles during growth. 2. A small lobate or circular structure in the posteroventral part of the body chamber (see Jones 1961, Jordan 1968). Mutvei and Reyment (1973) suggested that a gill retractor muscle may have been attached here. 3. An annular elevation or elevations, often indicated by coloration, which encircles the adapical part of the body chamber (see Crick 1898, Jordan 1968). This feature has generally been compared with the annular elevation in Nautilus (see Mutvei 1957, 1964) which is the site for attachment of the longitudinal mantle muscles and the subepithehal muscles. 4. A pair of large, tongue-shaped lateral indentations in the body chamber and, less commonly, also on the phragmocone (see Jordan 1968), which are of uncertain origin. In addition, Palframan (1969) described large, regularly shaped areas of coloration in the body chambers of Hecticoceras, again of uncertain significance. Phragmocone markings are usually confined to the siphonal region (see, for example, Grandjean 1910, Neaverson 1927, H51der 1954, Birkelund 1965, Henderson 1984). They occur in two main forms, fine longitudinal grooves and ridges, the ‘Haftstreifen’ of Holder (1954), and concentric striae which mirror the outlines of the sutural lobules, the 'Schleppstreifen’ of Holder (1954). PRESENT MATERIAE The material described here comes from limestones in the uppermost Cenomanian to Lower Turonian Gongila Formation at Ashaka Quarry in north-east Nigeria (see Wozny and Kogbe 1983). A number of ammonite genera (Vascoceras, Nigericeras, Paravascoceras, Paramammites, Thomasites, Pseudotissotia, and Wrightoceras) occur here in extraordinary abundance, all as internal moulds in a similar state of preservation. Although vague indications are present in several of these genera, however, it is only in Paravascoceras that the markings described below can be positively identified. [Palaeontology, Vol. 29, Part 4, 1986, pp. 725-738.) 726 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 1. Two successive sutures in Paravascoceras cauvini (Chudeau) showing associated linear and concen- tric markings. In this specimen (BM (NH) C. 90403) (see also text-fig. 3d) the linear markings are expressed in the form of faint grooves and the concentric markings as roughened or pitted areas on the mould surface. The ‘gutter’ is the mould of a postseptal prismatic zone of shell material (see Henderson 1984). Such a ‘gutter’ occurs frequently in many genera from Ashaka. ZABORSKI: INTERNAL MARKINGS IN AMMONITE 727 The specimens concerned are weakly ribbed, compressed to slightly depressed variants of P. cauvini (Chudeau 1909) (see also Schobel 1975 for review and synonymy). They have been deposited in the Department of Palaeontology, British Museum (Natural History), London (register numbers C. 90402-9). The most prominent markings displayed by these specimens are linear, spiral features (text-figs. 1-3) preserved in a number of ways: as bands of lighter or darker coloration; as very faint grooves; or as strips of surviving shell material. They occur upon the flanks, the venter, and the umbilical wall. Some extend the full distance between successive sutures (text-fig. 2) but many disappear adapically in a broad, annular area of coloration immediately adoral of the previous suture line (text-fig. 3t/). Others can be traced only a short distance adapical of the suture line with which they are associated. In no case do these markings continue from chamber to chamber. Adorally they terminate at the lobules of the suture lines. Unlike forms of external spiral ornament, therefore, these linear features do not form regular, continuous spirals. Rather, each chamber contains its own association of these markings within which individual features frequently deviate slightly from a true spiral course. They do not reflect an external ornament such as strigation which might be displayed upon composite internal moulds (see McAlester 1962 for an account of this TEXT-FIG. 2. Two successive sutures in Paravascoceras cauvini (Chudeau) showing linear and concentric markings, expressed in this specimen (BM (NH) C. 90404) as areas of lighter coloration on the mould surface. TEXT-FIG. 3. Linear markings in Paravascoceras cm/v/n/iChudeau). a-c, specimen BM (NH) C. 90402; n, ventral view ( X 2) also showing annular bands of coloration adoral of each suture line; /), c, ventral and lateral views ( X 1 ), In this specimen the markings are expressed in the form of bands of darker coloration or as strips of surviving shell material, r/, specimen BM (NH) C. 90403, in which the markings are expressed in the form of bands of darker coloration or as grooves on the mould surface (see also text-fig. 1 ). Both specimens uncoated PALAEONTOLOGY. VOLUME 29 type of mould). Markings closely similar to those in Paravascoceras, however, have been described as ‘Schleppstreifen der Lobenlinie’ in the Triassic ceratites Koiiinckites and Clypeoceras by John (1909) and in the Lower Cretaceous ammonite Polyptychites by Vogel (1959, p. 509, text-fig. 14). Unfortunately, none of the present specimens has such markings preserved at the junction of the body chamber and phragmocone. In accordance with the explanation of their origin given below, however, they might be expected to extend into the non-adult body chamber for a distance not ZABORSKI: INTERNAL MARKINGS IN AMMONITE 729 exceeding the length of one phragmocone chamber. In particularly well-preserved specimens these linear features are associated with a series of concentric markings reflecting the outlines of the sutural lobules (text-figs. 1 and 2). The concentric markings are most obvious when the linear features are expressed in the form of grooves but they are always faint and frequently visible only with strong, oblique lighting. They generally take the form of roughened or pitted areas on the mould surfaces, but may also be indicated by coloration, and are usually best developed close to the suture lines. Similar markings (‘Schleppstreifen’) have been described from the siphonal region in several Jurassic and Cretaceous ammonites by Hdlder (1954) and Birkelund (1965, p. 36). In rare Nigerian specimens these concentric markings may be contiguous with faint traces reflecting the outlines of the folioles, most commonly close to the suture lines but also around the umbilical area. One specimen (BM (NH) C. 90409) shows a series of fine, darkly coloured bands upon the umbilical wall which reflect the entire outline of the suture (text-fig. 4). There are approximately fourteen successive bands within one chamber in this specimen. These markings are similar to the ‘Pseudolobenlinie’ described in Koninckites by Bayer (1977«, p. 327, fig. 15) and to the same features described by John (1909) in Koninckites and Clypeoceras. Numerous specimens from Ashaka, belonging to several genera, show a relatively broad, darkly coloured band along the siphonal line. This feature is commonly present even when no other markings whatsoever are apparent. It is uncertain, therefore, whether it is related in any way to the markings described above. This siphonal marking is closely similar to the ‘Dunkles Sipho-Band’ described by Jordan (1968) and interpreted as marking the successive positions of a posteroventral muscle scar. Bayer (1974), however, believed that this dark band represented the area of attachment of siphuncle to shell, while Vogel (1959) thought a similar feature in Polyptychites to be a particularly prominent example of ‘Schleppstreifen’. umbilical shoulder umbilical seam . 5 mm TEXT-FIG. 4. Two successive sutures upon the umbilical wall in Paravascoceras cauvini (Chudeau) showing a series of darkly coloured bands reflecting the outline of the suture (specimen BM (NH) C. 90409). 730 PALAEONTOLOGY, VOLUME 29 INTERPRETATION The linear markings described here in Paravascoceras are not continuous from chamber to chamber. The structures they reflect, therefore, appear to have formed in close association with the individual chamber in which they occur. As mentioned above, these markings may be preserved as surviving shell material, or as grooves upon the steinkerns which must also record the original presence of strips of shell material in contact with the inner shell wall and terminating at each septum. Henderson (1984, p. 477) described a band of prismatic shell material (the ‘preseptal prismatic zone’) preceding the septum proper in Scipouoceras glaessiieri Wright. Although narrow and sharply defined along the folioles, this band of shell material is broader in the lobules and, here, extends adapically for a considerable distance in finger-like projections (see Henderson 1984, text-fig. 9a). It is probable that the linear markings in Paravascoceras record similar projections of a preseptal prismatic zone. Their associated concentric markings probably represent growth lines originally present upon these projections. If this interpretation is correct, these strips of shell material would have been secreted by those parts of the posterior mantle surface corresponding to the sutural lobules during mantle translocation. Holder (1954, p. 374) and Vogel (1959, p. 509), following the proposal of John (1909, p. 35), also suggested that mantle translocation was responsible for the presence of ‘Schleppstreifen’. In this context, it is instructive to examine chamber formation in the extant cephalopods Nautilus and Spirilla. Chaiiiher formation in Nautilus Chamber formation and buoyancy control during this process in Nautilus have been described by Collins et al. (1980), Ward et al. (1981), and Ward and Chamberlain (1983). Chamber formation takes between 70 and 120 days in captive N. inacroinplialiis and between 85 and 132 days in captive N. poinpiliiis. In order to maintain buoyancy throughout its growth cycle, Nautilus retains liquid in the adoral chambers of the phragmocone which serves as a reserve of ballast. This liquid is slowly expelled to compensate for mass added to the animal through shell and soft tissue growth. The rate of liquid removal is extremely slow; Ward et al. (1981) showed that in aquarium-based N. niacromplialus complete evacuation of a chamber takes as long as 135 days. Chamberlain and Moore (1982) found the constraining factor to be not the permeability of the siphuncular tube, but the rate of osmotic pumping. In fact, only in Sepia among living shelled cephalopods has short- term buoyancy control been observed; density changes in excess of 1 % can be effected in about six hours by varying the liquid : gas ratio of the cuttlebone (Denton and Gilpin-Brown 1973, p. 49). Ward (1979), Ward and Martin (1978), and Ward and Greenwald ( 1981 ) rejected the generally held view that Nautilus is able to vary its short-term buoyancy state in the same way; its rate of liquid exchange is far too slow for this. They noted that the animal maintains a very slight negative buoyancy and thought that it makes any vertical migrations by active swimming using the hypo- nome. Chamberlain’s (1981) calculation that powered swimming would be a much more efficient method of effecting vertical migration than buoyancy adjustment fully supports this conclusion. Cameral liquid is essential, therefore, only to the growing animal; fully grown adults contain virtually none at all (Ward 1979, Collins et al. 1980), a slightly larger body chamber compensating for the extra buoyancy resulting from this. Translocation of the mantle during chamber formation is far more rapid than apertural growth. Only maximum estimates are available for the duration of translocation. As Henderson (1984, p. 475) noted, a maximum of 6 days can be inferred for one individual of N. inacroinplialiis from the account given by Ward et al. (1981). In N. poinpiliiis translocation phases are completed within maximum periods of 10-20 days (Ward and Chamberlain 1983). Actual translocation rates may be much faster than these timings suggest. In any event, it is known that translocation takes up a maximum of only 10 % of the entire chamber formation period. The migrating mantle leaves behind it a liquid which performs two functions: to support the newly forming septum until its calcification is well advanced; and to provide a reserve of ballast to be expelled during future growth at the aperture. This liquid is not sea water; although approximately isosmotic with sea water, its composition is different (Denton and Gilpin-Brown 1966). The liquid ZABORSKl; INTERNAL MARKINGS IN AMMONITE 731 must be derived from within the animal’s soft tissues. There is some, inconclusive, evidence that translocation in N. pompilius is associated with unusual weight changes, perhaps caused by flooding of the new chamber space (Ward and Chamberlain 1983). This, if true, would suggest that expelled body fluid is rapidly replaced. Given the very low rates at which cameral liquid can be expelled, it is difficult to see how such weight changes would not result in buoyancy maintenance problems. A delay in the replacement of this body fluid might also have the result of bringing about a deflation of the body, thereby effecting the shortening of the body chamber which accompanies translocation (see text-fig. 5). Whatever the truth of this matter, there is a great disparity between rates of apertural growth and cameral liquid removal on the one hand, and rates of translocation and flooding of new chamber spaces on the other. This disparity requires that new cameral liquid be derived from the main body tissues. Chamber formation in Spirilla In Spirilla the shell is internal, there being no living chamber as such. The process of chamber formation has been described by Denton and Gilpin-Brown (1971) and Denton (1974). Each chamber space develops slowly and, as it does so, is filled with a clear liquid secreted by the body tissues which is isosmotic with sea water and the body fluids. In order to maintain buoyancy during shell growth, liquid is removed from the preceding chambers of the shell at the same time as the newly forming chamber space is being flooded. During evacuation of a chamber dissolved salts are first removed from the cameral liquid. The first liquid itself is only removed when the concentration of dissolved salts has fallen to about one-fifth that of sea water. The proportion of the overall cameral space occupied by liquid, however, remains much the same throughout the chamber formation cycle (see text-fig. 5). Chamber formation in Paravascoceras If the linear markings described here in Paravascoceras record the presence of adapical projections of a preseptal prismatic zone, then secretion of these structures and, consequently, mantle trans- location, would appear to have been a gradual and incremental process. Translocation would not have been ‘rapid’ as in Nautilus, but, as in Spirula, a chamber space would have developed slowly. If this was so, then the mechanism of buoyancy control must also have differed from that in Nautilus. A gradual translocation would have required that, as in Spirula, cameral liquid was removed from completed chambers at the same time as the newly forming chamber space was being flooded. The liquid entering a new chamber could have been transmitted either directly from the posterior mantle surface or, alternatively, from the siphuncular tissue. As noted above, the osmolar- ity of the cameral liquid varies in successive chambers of the Spirula shell. This is also the case in Nautilus (Ward 1979). Its cameral liquid is initially isosmotic with sea water but the osmolarity is reduced during the early stages of chamber emptying, that is, prior to decoupling of the siphuncle and cameral liquid. After decoupling, however, the osmolarity increases. In view of these osmolarity changes in both Spirula and Nautilus, it is likely that a similar phenomenon prevailed in Paravasco- ceras. Liquid entering a new chamber space would have been isosmotic with the body fluids, thus precluding osmotic interchange with the posterior mantle surface. Liquid leaving the phragmocone, however, would have been hyposmotic to the body fluids. Transfer of liquid through the siphuncle from the phragmocone to a new chamber would, therefore, have required that parts of the siphun- cular epithelium were absorbing liquid of low osmolarity while, simultaneously, the adoral part of the siphuncle was secreting liquid of higher osmolarity. It is not known whether any liquid is transferred to a new chamber through the siphuncle in Spirula. Denton and Gilpin-Brown (1971, pp. 370-371), however, thought it possible that the concentration of solutes might vary along its siphuncular epithelium. Diamond and Bossert (1968) suggested that the direction of water flow across a pumping epithelium would be reversible according to the direction of solute transport. As regards Nautilus, it is known that, if made artificially more buoyant, liquid can be returned to the camerae, its osmolarity, furthermore, corresponding to that of the liquid previously being removed from those camerae (Ward and Greenwald 1981). These factors suggest that transfer of liquid from 732 PALAEONTOLOGY, VOLUME 29 the phragmocone to a new chamber through the siphuncle may have been possible in Paravasco- ceras. During initial release of the mantle from the septum, however, expulsion of liquid from the main body tissues, directly from the posterior mantle surface, might have been necessary for the following reasons; 1. To maintain cameral liquid levels, as the chambers normally become increasingly voluminous during growth. 2. To bring about a partial deflation of the lobes, thereby aiding release of the posterior mantle surface from the septum. In Paravascoceras, as in the majority of ammonoids, the necks of the sutural lobes are constricted. Septal architecture, therefore, requires that some distortion of the soft tissues must have taken place upon mantle release. The septal surface in Paravascoceras is relatively simple and any lobe deflation would have needed only to be slight. It may, nevertheless, be significant that the linear markings described here frequently terminate adapically in a broad, annular band of coloration immediately adoral of the previous suture line (text-fig. 3u). If this colour band can also be taken as indicating the posterior mantle imprint, it would suggest a general loss of posterior mantle shape immediately upon release from the septum. 3. To produce a rapid, but short-lived, phase of mantle translocation which would have been necessary if apertural growth continued while the posterior mantle surface was stationary during septal secretion. This would have been the case even if, as suggested by Doguzhayeva (1982), apertural growth in ammonoids was slower during septal secretion. After this initial phase trans- location would have been much slower. It would, in fact, have kept pace with apertural growth. If chamber emptying in Paravascoceras proceeded as in Spirula, it may have been during septal secretion, with the posterior mantle stationary, that dissolved salts were removed from the chamber being completed. Clearly, the maximum possible rate of cameral liquid removal is a constraint upon the rate of apertural growth in Nautilus, but not upon the rate of mantle translocation. In Paravascoceras, however, the rate of cameral liquid redistribution would have been correlated with both rates of apertural growth and the largely synchronous process of mantle translocation. It may be mentioned that Paravascoceras was a shallow water genus of Tethyan distribution. In Nigeria it was a denizen of the trans-Saharan epeiric seaway which flooded northern Nigeria during the late Cenomanian and early Turonian. Water depths here are estimated as having averaged only 20-30 m (Fetters 1978, Reyment 1980). On theoretical grounds, a mechanism of rapid buoyancy control would seem to have been superfluous to Paravascoceras. It is likely, therefore, that cameral liquid pumping rates were primarily matched with growth rates and were too slow for short-term buoyancy adjustment. The gradual translocation phase envisaged here for Paravascoceras requires that the posterior mantle surface was unsupported by a septum for lengthy periods. In this case, a method of anchoring the posterior mantle surface during translocation would have been advantageous. The linear and concentric markings described here are intimately associated with the sutural lobules. The corre- sponding parts of the posterior mantle margin would, therefore, appear to have provided these attachment points, though specimens such as that shown in text-fig. 4 suggest that, in places, attachment was along a more continuous line. The linear markings in Paravascoceras are thought to represent adapical projections of a preseptal prismatic zone. Henderson (1984) suggested that this zone in Sciponoceras served as a temporary muscular attachment site for the perimeter of the posterior mantle surface immediately prior to secretion of the septum proper. He believed a second zone of prismatic shell, the ‘postseptal prismatic zone’, encircling the adoral face of the septum, provided a subsequent attachment surface until such time as translocation recommenced. Because, however, he regarded the septal periphery as the principal muscle attachment site in ammonites, Henderson (1984) suggested that bodily functions dependent upon the longitudinal musculature would have had to be suspended during translocation. He therefore concluded that translocation was accomplished very rapidly, perhaps taking only a few hours. Such a rapid translocation rate would imply either a mechanism of buoyancy control like that in Nautilus, or impossibly rapid rates of apertural growth and cameral liquid redistribution. The evidence provided by Paravasco- ZABORSKI: INTERNAL MARKINGS IN AMMONITE 733 ceras suggests that the adapical projections of the preseptal prismatic zone were secreted in in- crements by the mantle lobules during gradual translocation. The lobules would, therefore, have provided a series of temporary, presumably muscular, attachments throughout this lengthy period of mantle movement. The tenacity of such a form of attachment is, however, questionable. This suggests that the posterior mantle periphery was not the only, or even the principal, site of attach- ment. As Henderson (1984, p. 480) himself conceded, the various structures interpreted by Jordan (1968) as muscle scars may represent additional zones of attachment. Kulicki (1979) noted the metameric nature of the umbilical and ventral muscle(?) scars and took this fact to suggest a ‘stepwise’, rapid translocation pattern. Jordan ( 1968), however, found that these scars are frequently connected by darkly coloured bands which he regarded as marking the track of the relevant muscle attachments during translocation. The clarity of the scars at certain points within each chamber could be a reflection of the location of the muscles(?) during septal secretion when the mantle was stationary. Seilacher (1973, 1975) and Westermann (1975c) have also attributed particular physiological significance to the lobe endings in ammonoids. Seilacher (1973, 1975) pointed out that the shape of the ammonoid suture could be explained by envisaging the existence of a number of ‘tie-points’ located at the tips of the lobules. The posterior mantle, initially with a planar form, was thought to have become strongly attached to the shell wall at successive ‘tie-points’ after translocation. Septal shape resulted from ‘pull-off’ or radial tension acting upon the septal mantle. Westermann (1975c) proposed that the posterior mantle surface in ammonoids consisted of an aponeurosis-like structure which retained much of the septal shape during translocation. While accepting that those parts of the mantle corresponding to the saddles, folioles, and associated flutes might have suffered some distortion during translocation, he believed mantle shape would have been maintained along the periphery of the lobes. Anteriorly the body would have been fastened at an annular elevation like that in Nautilus. During translocation the posterior aponeurosis would have slid along the shell wall until it reached the position of the new septum. Here it would have reaffixed itself, initially at the tips of the lobe incisions. Pressure of cameral liquid, transmitted adorally from the phragmocone, would have been a major factor causing reflation of the saddles, rather than ‘pull-off’. This would also have produced adorally convex septa and prochoanitic septal necks. Finally, fixing of the entire mantle margin would have occurred, tightening in a radial direction extending the septal flutes, with septal secretion completing chamber formation. Westermann (1975c, text-fig. 7) (see also Ward and Westermann 1976) figured a suture in Glyptoxoceras in which the positions of the hypothetical ‘tie-points’ were normal but the lobules and folioles were reversed. This peculiarity was explained by envisaging a reversal of the direction of pressure applied to the posterior mantle surface during septal secretion. Bayer (1978), although suggesting a slow, ‘creeping’, method of translocation, denied the existence of ‘tie-points’, at least in the ontogenetic sense. He believed the reversed lobules of Westermann (1975c) to be distorted folioles resulting from the aberrant shaping of a planar posterior mantle surface. The specimens of Paravascoceras described here, however, provide evidence that posterior mantle shape was largely maintained during the greater part of translocation. Bayer (1977rt) regarded the retention of posterior mantle shape during translocation as an exceptional circumstance. He found the ‘Pseudolobenlinie’ in Kouinckites to be associated with fine ridges on the inner shell wall but regarded these as the result of abortive septal secretion following incomplete fastening of the septal mantle to the shell wall. John (1909), however, proposed that the various forms of ‘Schleppstreifen’ were connected with mantle growth and movement, more especially, with posterior muscle attachments concentrated in the lobes. He expressed surprise that soft tissues could leave such clear markings, but, as suggested here for Paravascoceras, these features probably mark the sites of myostracum. The factor controlling chamber size in Paravascoceras remains uncertain. It may have been genetic, as suggested by Bayer (19776) for cephalopods, related to the need for shell support by a septum, related to cameral liquid pressure, or perhaps related to siphuncular length. In regard to the last of these, Kulicki (1979) suggested that the connecting rings formed prior to translocation in ammonoids. This is not the case in Nautilus where calcification of the connecting ring is coincident 734 PALAEONTOLOGY, VOLUME 29 SPIRULA TKXT-FiG. 5. Median sections showing chamber formation cycles and cameral liquid distributions in Nautilus, Paravascoceras, and Spirula. Nautilus: a, immediately prior to translocation. B, during translocation of the mantle, c, during secretion of the new septum thus completing the formation of chamber ‘z’. Paravascoceras: A, as septal secretion completed the formation of chamber ‘y’. Dissolved salts were probably removed from the liquid Hlling chamber ‘y’ at this stage, while some liquid may have been expelled from chamber 'x’ to compensate for mass added through continuing apertural growth, b, at commencement of translocation with possible expulsion of body liquid into the newly forming chamber space, c, during the ZABORSKI: INTERNAL MARKINGS IN AMMONITE 735 with septal secretion (Ward et al. 1981). Connecting rings have been described from the body chamber in ammonoids. Although inferred by Kulicki (1979) for Kosmoceras and Quendstedtoceras, however, such a feature has been reported only from phylloceratids (Drushchits and Doguzhayeva 1974; Kulicki 1979; Westermann 1982) and, in some cases, the connecting rings project adorally well in excess of a single chamber’s length. Consequent upon his proposal of a rapid translocation rate in ammonites, Henderson (1984) suggested that the soft tissues of the siphuncle were preformed by a process of mantle invagination. The siphuncular mantle itself was thought to have been responsible for the secretion of the connecting ring following translocation. Rapid mantle movement is envisaged here only during the initial phase of translocation in Paravascoceras. Although partial preformation of the siphuncle may have been necessary as a consequence of this, the adoral part of the siphuncle could have formed during the latter, gradual phase of translocation. CONCLUSIONS A two-stage process of mantle translocation is proposed for Paravascoceras. An initial phase of rapid movement, resembling that in Nautdus, is suggested, during which the new chamber space was Hooded with liquid derived from the main body tissues. The major translocation phase, however, was a gradual process during which mantle movement kept pace with apertural growth and the new chamber space was flooded with liquid derived from the phragmocone. In this part of its chamber formation cycle Paravascoceras showed closer similarities to Spirida. The growth strategies em- ployed by Nautilus, Paravascoceras, and Spirida are shown in text-fig. 5. If other ammonoids employed the same growth strategy as Paravascoceras, it is possible that the relative importance of the two translocation phases varied from group to group. This might also have been the case at different ontogenetic stages within the same species. It is of interest that in their early ontogenetic stages ammonoids frequently show a morphology reminiscent of Nautilus, with retrochoanitic septal necks, adorally concave septa, and median siphuncles (see, for example, Spath 1950; Drushchits and Khiami 1970; Kulicki 1979). Nautilus has a very slow growth rate, at least in its post-juvenile stages (Ward et al. 1981; Ward and Chamberlain 1983; Cochran and Landman 1984). It is, however, questionable whether the growth mechanism proposed for Paravascoceras was adopted to allow faster growth. Although Nautilus spends a much greater proportion of its chamber formation period actively involved in septal secretion than is suggested for Paravascoceras, it is unclear whether its overall growth strategy is necessarily linked with a slow growth rate. The rate of apertural growth in Nautilus is constrained by the rate at which ballast in the form of cameral liquid can be expelled. The same constraint, however, would apply if it showed a method of buoyancy control like that proposed for Paravasco- cm/.v; the rate of cameral liquid redistribution would also be correlated with apertural growth rates. A more elficient osmotic pump would, alone, be sufficient to remove this constraint gradual and lengthy phase of mantle translocation when cameral liquid was pumped from the phragmocone into the newly forming chamber space, d, at completion of translocation when septal secretion completed the formation of chamber ‘z’. Levels of cameral liquid in Parava.scoceras are assumed to have been of the same order as those suggested for ammonoids by Westermann (I975u, h), that is, filling the last one to three chambers of the phragmocone. Spirilla: A, after completion of chamber ‘y’- At this stage dissolved salts are being removed from the liquid in this chamber. B, as the walls of chamber ‘z’ begin to form, c, later during the formation of chamber 'z' as liquid isosmotic with sea water fills the new chamber space and cameral liquid hyposmotic to sea water is removed from chambers ‘x’ and ‘y • d, at completion of chamber ‘z’. Based on Denton and Gilpin-Brown (1971) and Denton ( 1 974). Notice that the length of the body chamber and the total amount of cameral liquid vary markedly in Nautilus. In Paravascoceras the length of the body chamber varies but little and, as in Spirilla, the proportion of the cameral space occupied by liquid also remains approximately constant. 736 PALAEONTOLOGY, VOLUME 29 upon the growth rate in Nautilus but pumping rates are probably only one of a number of constrain- ing factors. The chamber formation cycle in Spirula, nevertheless, certainly proceeds much more rapidly than in Nautilus. Mature animals may have a shell comprising more than thirty chambers (Denton and Gilpin-Brown 1971). Clarke (1970) estimated that Spinila is sexually mature at an age of 12-15 months and has a total life span of only 18-20 months. As for ammonoids, Doguzhayeva (1982), on the evidence of assumed daily growth bands, estimated that a new chamber was added about every 14 days. Landman (1983) questioned this figure, and it should be mentioned that Saunders (1983) found that widely varying periods of time are represented by growth lines in Nautilus. It is, however, of interest that one specimen of Paravascoceras shows approximately fourteen colour bands mirroring the sutural outline within one chamber (text-fig. 4). These bands seem to record posterior mantle movement. Westermann (1975c), on the other hand, suggested that an adorally directed camera! liquid pressure may have played a significant role in the formation of adorally convex and fluted septa. Such a morphology could result if the posterior mantle margin was affixed to the shell wall at the lobules, as seems to have been the case in Paravascoceras. An adorally directed cameral liquid pressure requires a gradual translocation process. Acknowledgements. I am grateful to Dr M. K. Howarth and Mr D. Phillips for assistance in several ways. Special thanks are due to Dr H. G. Owen for valuable discussions and a review of the manuscript. Field work was supported by a University of florin Senate Research Grant. Photographs were provided by the British Museum (Natural ITistory) Photographic Unit. REFERENCES BAYER, u. 1970. Anomalien bei Ammoniten des Aaleniums und Bajociums und ihre Beziehung zur Lebensweise. Neiies Jh. Geol. Paldont. Ahli. 135, 19-41. 1974. Die Runzelschicht— ein Eeichbaulelement der Ammonitenschale. Paldont. Z. 48, 6-15. \911a. Cephalopoden-Septen. Teil 1: Konstruktionsmorphologie des Ammoniten-Septums. Neiies Jb. Geol. Paldont. Abh. 154, 290-366. \911b. Cephalopoden-Septen. Teil 2; Regelsmechanismen im Gehause- und Septenbau der Ammoniten. Ibid. 155, 162-215. 1978. The impossibility of inverted suture lines in ammonites. Lethaia 11, 307-313. BiRKELUND, T. 1965. Ammonites from the Upper Cretaceous of west Greenland. Medd. Gronland 179 (7), 1-192, pis. 1-48. CHAMBERLAIN, J. A. 1981. Hydromechanical design of fossil cephalopods. In house, m. r. and senior, j. r. (eds.). The Ammonoidea. Spec. Vol. Syst. Assoc. 18, 289-336. and MOORE, w. a. 1982. Rupture strength and flow rate of Nautilus siphuncular tube. Paleobiology 8, 408-425. CHUDEAU, R. 1909. Ammonites du Damergou (Sahara meridionel). Bull. Soc. geol. Fr. (4) 9, 65-71, pis. 1-3. CLARKE, M. R. 1970. Growth and development of Spirula spirula. J. mar. biol. Ass. U.K. 50, 53-64. COCHRAN, J. K. and LANDMAN, N. H. 1984. Radioactive determination of the growth rate of Nautilus in nature. Nature 308, 725-727. COLLINS, D., WARD, p. D. and WESTERMANN, E. G. 1980. Eunction of cameral water in Nautilus. Paleobiology 6, 168-172. CRICK, G. c. 1898. On the muscular attachment of the animal to its shell in some fossil Cephalopoda (Ammo- noidea). Trans. Linn. Soc. (2) 7, 71 114, pis. 17-20. DENTON, E. J. 1974. On buoyancy and the lives of modern and fossil cephalopods. Proc. roy. Soc. Land. (B) 185, 273-299. and GILPIN-BROWN, J. B. 1966. On the buoyancy of the pearly Nautilus. J. mar. biol. Ass. U.K. 46, 723- 759. 1971. Eurther observations on the buoyancy of Spirula. Ibid. 51, 363-373. 1973. Flotation mechanisms in modern and fossil cephalopods. Adv. mar. Biol. 11, 197-268. DIAMOND, J. M. and BOSSERT, w. H. 1968. Functional consequences of ultra-structural geometry in ‘backwards' fluid-transporting epithelia. J. Cell. Biol. 37, 694-702. ZABORSKI: INTERNAL MARKINGS IN AMMONITE 737 DOGUZHAYEVA, L. 1982. Rhythms of ammonoid shell secretion. Lethaia 15, 385-394. DRUSHCHITS, V. V. and DOGUZHAYEVA, L. A. 1974. On certain features of morphogenesis of Phylloceratidae and Lytoceratidae (Ammonoidea). Paleont. Zli. 1974, 42-53. [In Russian.] and KHiAMi, N. 1970. Structure of the septa, protoconch walls and initial whorls in early Cretaceous ammonites. Ibid., 1970, 35-47. [In Russian.] GRANDJEAN, F. 1910. Le siphon des ammonites et des belemnites. Bull. Sac. geol. Fr. (4) 10, 496-519. HENDERSON, R. A. 1984. A muscle attachment proposal for septal function in Mesozoic ammonites. Palaeon- tology 27, 461-486, pis. 48-49. HOLDER, H. 1954. Uber die Sipho-Anheftung bei Ammoniten. Neiies Jh. Geol. Paldont. Mh. 1954, 372-379. JOHN, R. 1909. iJher die Lehensweise und Organisation der Ammoniten. Inaugural Dissertation, University of Tubingen. 53 pp., 1 pi. Stuttgart. JONES, D. L. 1961. Muscle attachment impressions in a Cretaceous ammonite. J. Paleont. 35, 502-504, pi. 71. JORDAN, R. 1968. Zur Anatomie mesozoischer Ammoniten nach den Strukturelementen der Gehause- Innenwald. Beih. geol. Jalirh. 77, 1-64, pis. 1-10. KENNEDY, w. J. and COBBAN, w. A. 1976. Aspects of ammonite biology, biogeography and biostratigraphy. Spec. Pap. Palaeont. 17, 1-94, pis. 1-11. KULiCKi, c. 1979. The ammonite shell; its structure, development and biological significance. Palaeont. Pol. 39,97-142, pis. 24-48. LANDMAN, N. H. 1983. Ammonoid growth rhythms. Lethaia 18, 248. MCALESTER, A. L. 1962. Mode of preservation in early Paleozoic pelecypods and the morphologic and ecologic significance. J. Paleont. 36, 69-73, pi. 16. MUTVEi, H. 1957. On the relations of the principal muscles to the shell in Nautilus and some fossil nautiloids. Ark. Miner. Geol. (2) 10, 219-254, 20 pis. 1964. Remarks on the anatomy of recent and fossil Cephalopoda, with descriptions of the minute shell structure of belemnoids. Stockh. Contr. Geol. 11, 79-102. and REYMENT, R. A. 1973. Buoyancy control and siphuncle function in ammonoids. Palaeontology 16, 623-636. NEA VERSON, E. 1927. The attachment of the ammonite siphuncle. Proc. Lpool. geol. Soc. 14, 65-77, pi. 11. PALERAMAN, D. E. B. 1969. Taxoiiomy of sexual dimorphism in ammonites: morphogenetic evidence in Hectico- ceras briglitii (Pratt). In westermann, g. e. g. (ed.). Se.xual dimorphism in fossil metazoa and taxonomic implications, 126 152. Schweizerbart’sche Verlagsbuchhandlung, Stuttgart. piiTTERS, s. w. 1978. Stratigraphic evolution of the Benue Trough and its implications for the Upper Cretaceous paleogeography of west Africa. J. Geol. 86, 31 1 -322. REYMENT, R. A. 1980. Biogeography of the Saharan Cretaceous and Paleocene epicontinental transgressions. Cret. Res. 1, 299-327. SAUNDERS, w. B. 1983. Natural rates of growth and longevity of Nautilus helauensis. Paleobiology 9, 280- 288. SCHOBEL, J. 1975. Ammoniten der Familie Vascoceratidae aus dem Unterturon des Damergou-Gebietes, Republique du Niger. Spec. Vol. palaeont. Instn. Univ. Uppsala 3, 1 - 136, 6 pis. SEILACHER, A. 1973. Fabricational noise in adaptive morphology. System. Zook 22, 451-465. -- 1975. Mechanische Simulation und funktionelle Evolution des Ammoniten-Septums. Paldont. Z. 49, 268-286. SPATH, L. E. 1950. The study of ammonites in thin, median sections. Geol. Mag. 87, 77-84. VOGEL, K.-p. 1959. Zwergwuchs bei Polyptychiten (Ammonoidea). Geol. Jb. 76, 469-540, 4 pis. WARD, p. D. 1979. Cameral liquid in Nautilus and ammonites. Paleobiology 5, 40-49. and CHAMBERLAIN, J. 1983. Radiographic observation of chamber formation in Nautilus pompilius. Nature 304, 57-59. and GREENWALD, L. 1981. Chamber refilling in Nautilus. J. mar. biol. Ass. U.K. 62, 469-475. — and MAGNIER, Y. 1981. The chamber formation cycle in Nautilus macromphahis. Paleobiology 7, 481 493. and MARTIN, A. w. 1978. On the buoyancy of the pearly Nautilus. J. exp. ZooL 205, 5-12. and WESTERMANN, G. E. G. 1976. Sutural inversion in a heteromorph ammonite and its implication for septal formation. Lethaia 9, 357-361 . WESTERMANN, G. E. G. 1975a. Remarks on Mutvei and Reyment's hypothesis regarding ammonoid phragmo- cones. Palaeontology 18, 437-438. 1975/). Architecture and buoyancy of simple cephalopod phragmocones and remarks on ammonites. Paldont. Z. 49, 221-234. 738 PALAEONTOLOGY, VOLUME 29 WESTERMANN, G. E. G. 1975c. Model for origin, function and fabrication of fluted cephalopod septa. Ibid. 49, 235-253. 1 982. The connecting rings of Nautilus and Mesozoic ammonites: implications for ammonoid bathymetry. Lethaia 15, 373-384. wozNY, E. and kogbe, c. a. 1983. Further evidence of marine Cenomanian, Lower Turonian and Maastrichtian in the upper Benue Basin of Nigeria (west Africa). Cret. Res. 4, 95-99. Typescript received 4 September 1985 Revised typescript received 15 February 1986 P. M. P. ZABORSKI Department of Geology University of Maiduguri PMB 1069 Maiduguri, Nigeria ISOLATED ALGAL CYSTS IN THE PALAEOCENE-LOWER EOCENE OF KURDISTAN by GRAHAM F. ELLIOTT Abstract. Individual spherical bodies from a richly algal deposit in the Palaeocene-Lower Eocene of Kurdi- stan are described in detail and interpreted as detached cysts from a dasycladalean alga believed to resemble Halicoryne (Cenozoic-Recent). The fossils are named Sedalanelhi segonzacae gen. et sp. nov. The Tethyan Palaeocene-Lower Eocene of Iraqi Kurdistan is locally richly algal, the microfloras showing a profusion of calcareous algae, both reds and greens. Amongst the latter the Dasycladales are represented by numerous species referable to fourteen genera (Elliott 1968, 1978; plus Acropo- rella sp., new record). The richest sample in my experience was that from near Sedalan, Sulaimaniya (details in Elliott 1978), a rock packed with fossils mostly algal, set in clear calcite. Amongst these are numerous examples of spherical bodies, described below, and now recognized as probably detached algal cysts. MATERIAL This consists of numerous examples seen in thin-section. The majority are uniformly near-circular, with thin outer wall of dark crystalline calcite, and infilling of clear calcite, or calcite and pyrite. These exclusively near-circular outlines, irrespective of diameter, indicate that these objects are near-spherical. There is, however, a minority of individuals in which the wall shows an irregular, ‘crushed’, outline. Of the uncrushed circles, those of diameter from 015 mm up to about 0-36 mm mostly show an outer wall of blurred or fuzzy appearance, of which it is difficult to measure the thickness. They are interpreted as non-equatorial circles, with consequent oblique cuts of the wall structure. Those from 0-36 to 0-45 mm show clearly a wall of about 0 006 mm thickness, and are interpreted as full or near equatorial cuts. Some show a slight flattening of the circle on one side (text-fig. 1a). One showed an external coating of calcite on one hemisphere only, interpreted as a diagenetic feature. The infillings are usually of clear calcite, sometimes showing zones of difFerent calcite growth, peripheral and central, and sometimes including small areas of pyrite, or in one instance, a pyritic core. DISCUSSION These objects are interpreted as single thin-walled dasycladalean reproductive cysts, freed from the plants of origin. In those living dasycladaleans which calcify, the cysts are formed in reproductive branches or chambers, and are usually indicated by individual cavities in a calcified body, occurring fossil as part of the main plant if this is heavily calcified; e.g. as in Cyniopolia. Often, however, the calcified reproductive part, with its component cysts, is found separately, as in the familiar micro- fossil Acicutaria. Groups of cysts with their individual calcifications fused may also occur. Adjacent cyst-formation in these groups may lead to crowding and deformation, with rounded-angular out- lines and pyriform or tubular morphologies. None of the irregular Sedalan cysts (text-fig. 1b) show this; their outlines all appear due to pressure and crushing, and they are interpreted as uniformly post-mortem in origin. They appear to have been normal spheres, now showing post-mortem damage. I Palaeontology, Vol. 29, Part 4, 1986, pp. 739-741. | 740 PALAEONTOLOGY, VOLUME 29 In the living Halicoryne, calcification is very light, except around the cysts. Valet (1969) has studied and figured these (along with the other living dasycladaleans) in great detail. In H. spicata (Kiitzing) Solms-Laubach the cysts are cemented in groups by their calcification, but in H. wrightii they are individually calcified, but separate within the uncalcified reproductive chamber. The flat- tened side of the Sedalan cysts probably indicates the operculum or mechanism by which the reproductive contents were eventually shed, as known to occur in these living species. Calcification is a feature of the adult growth-stages of dasycladaleans; it is especially characteristic of, and likely to originate around, their reproductive bodies, and may be confined to them in some genera. Valet and Segonzac (1969) have studied the fossil record of Halicoryne; species are known from the Palaeocene-Lower Eocene of the south of France, and from the Czechoslovakian Miocene. Both are represented by aggregated cysts of the H. spicata type. In the French Palaeocene-Eocene species, the cysts are smaller in size and thicker-walled (cyst diameter 0-10 mm, wall-thickness 0 015-0 020 mm). The Miocene H. morelleti (Pokorny) Valet and Segonzac reaches the size of the Sedalan species, but is also thicker-walled (cyst diameter 0-20-040 mm, wall-thickness 0-030- 0 035 mm). Summarizing, the Sedalan spheres are interpreted as reproductive cysts from an extinct, largely uncalcified, dasycladalean alga. From living analogy, this could have been a Halicoryne sp. or a plant very similar. The cysts developed free in the reproductive chambers, and were probably the only structures of the plant to show calcification. A Halicoryne sp. is known from Tethyan rocks of the same age and similar facies in France, and the Miocene H. morelleti had cysts of a similar size. The wall-thickness of the Sedalan cysts is exceptionally thin, but does not otherwise conflict with what is known of the incidence of light calcification in dasycladaleans. The presumed operculum- site matches that known in the living Halicoryne spp. (Valet 1969). It remains only to consider whether to name these spheres. Hollow calcareous-walled spheres A B TEXT-FiG. 1a, b. Secialanella segonzacae gen. et sp. nov. Thin-section examples from the Kolosh Formation (Palaeocene-Lower Eocene) of Sedalan, Sulaimaniya, north-east Iraq. BM (NH) V. 41606, x 100. 1a. Holo- type, complete individual with portions of other individuals, one differently calcified. 1b. Crushed example. ELLIOTT: TERTIARY ALGAL CYSTS 741 occur at many stratigraphic levels from the Lower Palaeozoic onwards, and the term calcisphere is often used for this type of microfossil. It seems likely that they are of varied origins. However, Rupp (1966) drew attention to the similarity of recent acetabularian cysts to fossil calcispheres. Marszalek (1975), dealing with reproductive cysts of a species identified as the living Acelahiilaha antillana (Solms-Laubach) Egarod, studied the wall ultrastructure, and reviewed the environmental conditions and current taxonomy of these plants. Clearly some fossil calcispheres are probably of acetabularian algal origin. In view of the largely algal flora associated in the Kurdistan samples, and the occurrence elsewhere of Halicoryne at the same stratigraphic horizon, I think it desirable to name these objects. Whether this name will ever become a synonym of Halicoryne depends on future discoveries, if any, of cysts associated with plants. SYSTEMATIC PALAEONTOLOGY Sedalanella gen. nov. Type-species. S. segonzacae sp. nov.: Palaeocene-Lower Eocene of Iraqi Kurdistan. Derivation of name. After the type-locality of Sedalan, Sulainianiya, north-eastern Iraq. Diagnosis. Calcareous, thin-walled, near-spherical hollow body, slightly flattened on one hemi- sphere. Sedalanella segonzacae sp. nov. Text-fig. 1 Derivation of name. In tribute to Mme G. Segonzac for her work on algae of the French Pyrenean Palaeocene- Lower Eocene. Holotype. V. 41606 (text-fig. Ia); British Museum (Natural History), Department of Palaeontology. Horizon and locality. Palaeocene-Lower Eocene Kolosh Formation; L5 km south-west of Sedalan, 48 km north-west of Sulaimaniya, north-eastern Iraq. Diagnosis. As for genus; diameter of hollow body 0-30-0-45 mm; wall-thickness 0-006 mm. Other material. Numerous examples from the same horizon and locality, thin-sections V. 32492, V.41606, V.51232. REFERENCES ELLIOTT, G. F. 1968. Permian to Palaeocene calcareous algae (Dasycladaceae) of the Middle East. Bull. Br. Mus. nat. Hist. (Geol.), Suppl. 4, 1 - 1 1 1 . 1978. A new dasycladacean alga from the Palaeocene of Kurdistan. Palaeontology. 21, 687-691. MARSZALEK, D. s. 1975. Calcisphere ultrastructure and skeletal aragonite from the alga Acetabiilaria antillana. J. sedim. Petrol. 45, 266-27 1 . RUPP, A. w. 1966. Origin, structure and environmental significance of Recent and fossil calcispheres. GeoL Soc. Amer., Annual meeting. San Francisco 1966, Abstracts. 186. VALET, G. 1969. Contribution a I’etude des Dasycladales. 2. Cytologic et reproduction. 3. Revision systema- tique. Nova Hedwigia. 17, 551-644. and SEGONZAC, G. 1969. Les genres Chalmasia et Halicorvne (Algues acetabulariacees). Bull. Soc. geol. Fr. (7) 11, 124-127. Typescript received 5 September 1985 Revised typescript received 20 December 1985 G. F. ELLIOTT Dept, of Palaeontology British Museum (Natural History) Cromwell Road, London, SW7 5BD _.„i-s .3-VL, ,■ ■' *L.« ^ ' .'V' ..* ■ .► ■■i* •nj- . 4# i- ■ '-*\ V • » ORDOVICIAN TRILOBITES FROM CHEDAO, GANSU PROVINCE, NORTH-WEST CHINA by ZHOU ZHiYi and w. x. dean Abstract Two trilobite assemblages of middle Llandeilo to basal Caradoc and of Caradoc age are described from the type section of the Chedao Formation at Chedao, Huanxian County, Gansu Province, north-west China. They include twenty-nine taxa, of which five are new species: Phorocephala quadrata, Penispis ohscura, Microparia (Qiuidrapvge) chedaoensis, Ischyrophynia? zidqiangi, and Hammatocneinis ohsoletus. The compo- sition of both assemblages approaches that of the Nileid Association and suggests that the Huanxian area formed part of the western slope of the North China carbonate platform. The presence of genera such as Peraspis, Lyrapyge, and Cyphoniscus indicates that some genera once thought to have a short stratigraphic range persisted much longer when appropriate environments were available. The Huanxian area from which the trilobites now described were obtained lies in eastern Gansu Province, eastern part of north-west China (text-fig. 1). The importance of Huanxian to Chinese Ordovician geology lies in its palaeogeographic position between the shallow-water shelf of the North China, or Sino-Korean, platform to the east and the deep basin of the Qilian Trough or Geosyncline to the west, terms used by Huang et al. (1977) and Jen et al. (1980). Lai et al. (1982) included the area in the western marginal belt of the North China carbonate platform. The macrofossil assemblages may be interpreted either as intermediate between those of basin and shelf or, as suggested by Lu (in Lu et al. 1976), of transitional type. Although the existence of Ordovician strata in the Huanxian area has been known for over twenty-five years (Lu 1959), only a few cephalopods (Chang 1962) and two trilobites (Nileus hiumxianensis Zhou and Psendostygina lepida Zhou, both in Zhou et al. 1982) have been described. LOCALITIES AND STRATIGRAPHY The well-preserved, but disarticulated, trilobites described here were collected from the measured section at Shixiezi, 8 km north-west of Chedao, Huanxian County (text-fig. 1); the base of the section is faulted against Cretaceous rocks, and the top is overlain by Recent loess deposits. The Ordovician rocks consist mainly of limestones of different colours (text-fig. 2), the purplish beds in the upper part of the section being termed Chedao Formation by Lin (in Lai el al., 1982, p. 66). For present purposes, and on the basis of unpublished information from Fei Anqi and Liu Pingjun of the Chanqing Bureau of Petroleum Prospecting, we use Chedao Formation for the whole succession, divided into thirteen numbered, informal beds, pending a formal revision of the stratigraphic terminology. Macrofossils were obtained only from Beds 4 and 12, and no conodonts were found. Macrofossil.s from Bed 4 Trilobites include: Birnumites sp. indet., Bidbaspi.s sp. indet., Cyphoid.scu.s cf. sociali.s Salter, 1853, Geragnostii.s aff. longicoUis (Raymond 1925), Glaphurina sp., Ilammatocnemi.s ka}diugen.sis Zhang, 1981, PI. oh.soletm sp. nov., Lonchodomas names Zhou in Zhou et al. 1982, Microparia (Quadrapyge) chedaoeiisis sp. nov., Peraspis ohscura sp. nov., Rorringtonia sp., Slmnuirdia tarinniensis Zhang, 1981, Stenopareia sp. indet. and Telepinna sp. Of these Peraspis ohscura outnumbers every other species, and Stenopareia is represented by only one abraded cranidium. The single graptolite taxon found in association with the trilobites was identified by Ni Yunan (Nanjing Institute of Geology and Palaeontology) as Dicellograptus sextans exilis Elies and Wood, 1904, indicative of the Nemagraptus gracilis Biozone, now considered correlative with the middle and upper Llandeilo Series and IPalaeontology, Vol. 29, Part 4, 1986, pp. 743-786, pis. 58-65.) 744 PALAEONTOLOGY, VOLUME 29 basal Caradoc Series (Williams el al. 1972). Microparia (Qiiadrapyge) is known from Caradoc strata of Taojiang, Hunan, considered as part of the slope area of the Yangtze Platform in the Ordovician, and probably also from southern Scotland (J. K. Ingham, pers. comm.). Rorringtonia occurs in the Llandeilo to early Caradoc of England (Shropshire), Wales, and Norway. Lonchodomcis nanus is known from the topmost Pingliang Formation (Nemagraptus gracilis Biozone) of Ningxia, north-west China. None of these records contradicts the present graptolite evidence. MacrofossUs from Bed 12 The trilobite fauna is rich in Nileus huanxianensis Zhou and Pseudostygina lepida Zhou both in Zhou el al. 1982. Other taxa include: Arthrorhachis cf. tarda (Barrande 1846), Birmanites aff. asiaticiis (Petrunina in Repina el al. 1975), Cyclopyge cf. recurva Lu in Wang el al., 1962, Ovalocephalus kelleri Koroleva, 1959, Ischyrophymal zhiqiangi sp. nov., Lichas aff. laeiniatus (Wahlenberg 1818), Lyrapygel gaoluoensis (Zhou in Zhou el al. 1977), Paraphillipsinella glohosa Lu in Lu and Chang 1974, Paratiresias tiirkestanicus Petrunina in Repina et al. 1975, Phorocephala quadrata sp. nov., Stenopareia alT. howmanni (Salter 1848), Telephina convexa Lu, 1975 and Xenocybel sp. TEXT-FIG. 1. Outline maps showing location of the described measured section. Associated cephalopods, identified by Zou Xiping (Nanjing Institute of Geology and Palaeontology) as species of Sinoceras and Miclielinoceras, constitute a fauna that, according to Lai et al. ( 1982, pp. 66, 229) is identical with the Sinoceras chinense assemblage of the Pagoda Formation in the Yangtze region, and the presence of the trilobites Cyclopyge recurva and Paraphillipsinella glohosa in the Pagoda Formation supports the evidence from the cephalopods. A Caradoc age for the Pagoda Formation has been suggested on the basis of both faunas and regional stratigraphic context (Lu 1959; Chang 1960; Lu et al. 1976; Lai et al. 1982; Zheng etal. 1982). ZHOU AND DEAN: ORDOVICIAN TRILOBITES 745 Fm Bed No LITHOLOGY Recent loess deposits < S a: o u. O < Q 111 Z o 13 12 1 1 I - I 3ZHn Q I ° T / . / / / 2.2m purplish red oomicrite 2.5m purplish red medium-bedded biomicrite 3.7m pale grey & grey medium-bedded dolomitic micrite 1.7m pale grey thick-bedded nodular micrite 0.6m brownish grey thick-bedded calcarenite 1.0m pale grey dolomitic mottled micrite 2.0m pale grey medium-bedded micrite 2.6m greenish grey medium-bedded micrite 31.26m grey massive ferruginous calcirudite 10 Metres 1.75m grey thin-bedded micrite interbedded with purplish-grey & yellowish green dolomitic micrite 2.0m grey massive calcirudite 0.9m yellowish-green mudstone intercalated with pale grey calcirudite 2.0m grey massive calcirudite Faulted against Cretaceous rocks TEXT-FIG. 2. Measured section of the type Chedao Formation, showing bed numbers referred to in the text. RELATIONSHIPS OF THE TRILOBITES The composition of both trilobite assemblages approximates to that of the Nileid Association proposed by Fortey {\915h) on the basis of the Spitzbergen succession; this suggests that the Huanxian area formed part of the (now) western slope of the North China carbonate platform. The Huanxian trilobites, including all listed from Beds 4 and 12, may be divided into three groups: 1. Widespread or pelagic genera that occurred in environments from shallow platform to slope. These include Phorocephala, Birmanites, Cyclopyge, Lyrapyge, Telephina and H animat ocneniis. 746 PALAEONTOLOGY, VOLUME 29 2. Genera such as Nileus, Peraspis, Bidhaspis, Microparia (Quaclrapyge) and Shionardia (s.s.) that are restricted to, or found predominantly in, the Nileid Association or occur in equivalent slope facies. Several species from Bed 4 are of typical North American type and closely resemble forms from the Middle Table Head Formation (Whittington 1965) of western Newfoundland; they include Geragnostus aff. longicollis, Shumardia tarimuensis (closely allied to S. grcmulosa, type species of the genus) and Peraspis obscura. According to Fortey (19756, p. 344) the Table Head trilobite assem- blage is similar to that of the Nileid Association. In Shropshire Rorringtonia occurs in dark, graptolitic shales with, inter al., Cnemidopyge, Selenopeltis, OgygiocarelUy and Spirantyx (Whittard 1966, pp. 299-302, 306). In central Wales the genus is known from dark, graptolitic mudstones with various cyclopygids, Geragnostus, Barrandia, Ogygiocarella, and Ogyginus\ these associations suggest a deep water, offshore environment (Owens 1981, p. 91). 3. Genera that may have originated from nearby carbonate mounds (or ‘reefs’) include Cyphonis- cus, Paratiresias, Glaphurina, and Ischyropliymal Isocolid trilobites are more typical of a carbonate mud-mound facies (Dean 1971, p. 52), as are Glaphurina and Ischyrophyma (Fortey 19756; Mikulic 1980). Stenopareia and probably Lichas may also be mound-derived allochthonous forms, judging by faunas described from Ashgill carbonate mounds in central Sweden (Warburg 1925, 1939), northern England and eastern Ireland (Dean 1971, 1974, 1978). Although the trilobite assemblages from Beds 4 and 12 are both indicative of a slope facies, that of Bed 4 suggests deeper water than that of Bed 12. Only 14% of trilobites from the graptolite- bearing strata of Bed 4 belong to what Fortey (19756.) termed the Illaenid Association, which includes the carbonate mound facies, but in Bed 12 the corresponding figure is 33%. Trilobites from Bed 12 are closely related to late Caradoc to Ashgill faunas in Central Asia that include the following forms: Arthrorhachis cf. tarda, Birmanites aff. asiaticus, Xenocyhel sp., Lichas aff. laciniatus, Ovalocephalus kelleri, and Paratiresias turkestanicus. The Bed 12 assemblage also bears a great resemblance to the Caradoc to early Ashgill faunas of the Yangtze Platform, as indicated by the following, or closely allied species: Arthrorhachis cf. tarda, Lyrapygel gaoluoensis, Cyclopyge cf. recurva, Paraphillipsi- nella globosa, and Te/ephina convexa. The assemblage from Bed 4 exhibits mixed affinities and includes several trilobites of North American type (see above), though Birmanites, Bidbaspis, and Hammatocnemis indicate Asiatic affinities. One of the results of the present work is additional evidence that many Ordovician trilobite genera had long stratigraphic ranges, and that some once thought to be very short-ranging persisted much longer when the appropriate environments were available. Examples are Arthrorhachis, Birmanites, Geragnostus, Nileus, Phorocephala, and Shumardia, all of which persisted from the Tremadoc to the Ashgill, and Hammatocnemis and Telephina, which ranged from Arenig to Ashgill. Cyphoniscus, once regarded as a typical Ashgill isocolid, occurs at our section in the Llandeilo or lowest Caradoc (at one locality in eastern Canada it is known also from lowest Silurian beds; Dean 1972). Lyrapyge, founded on material from the Arenig of Spitsbergen, is here recorded from the Caradoc; and Peraspis, recorded previously from the Arenig to Llanvirn, has its range extended into the Caradoc on the basis of P. obscura, its youngest known species, in Bed 4. The long range of trilobite genera found predominantly in carbonate mound facies is related to persistence of the tropical shelf-edge habitat (Fortey 19806). A particular problem of Ordovician trilobite generic assemblages concerns autochthonous el- ements of the Nileid Association, particularly in rocks of Arenig-Llanvirn age such as the Table Head Formation of western Newfoundland. The subject was discussed by Ross (1970) in an account of the problems involved in defining the Whiterock Stage in western USA. Ross and Ingham (1970), after finding a trilobite generic assemblage of ‘Whiterock’ type in lowest Caradoc strata in the Girvan area, Scotland southern, noted the very widespread occurrence of what they termed the ‘Toquima-Table Head Faunal Realm’ (named for areas in Nevada and western Newfoundland). Many of the genera in their ‘faunal realm’ had extended ranges but were adapted to a particular environment and palaeogeographic setting, peripheral to Laurentia in the Ordovician and close to ZHOU AND DEAN: ORDOVICIAN TRILOBITES 747 the transition from miogeosynclinal to eugeosynclinal facies; distribution and migration of the trilobites would follow dispersal of the appropriate facies belts. Further evidence of this concept is provided by the Chedao Formation trilobites, so that some genera documented particularly from the early Ordovician persist as late as the Caradoc, while others predate genera thought at one time to be characteristic of the Ashgill. SYSTEMATIC PALAEONTOLOGY Terminology is essentially that Harrington et al. (in Moore 1959, pp. 0117-0126); additions are lunette (Whittington 1954, p. 139) for a crescent-shaped area in the axial furrows of some illaenids; eye socle (Shaw and Ormiston 1964); and baccula (Opik 1967, p. 53). Taxa are arranged in Treatise order (Moore 1959), additions and modifications are as follows: Scharyiinae are included in the Aulacopleuridae (Thomas and Owens 1978), Phillipsinellidae are considered to be related to the Styginidae (Bruton 1976; Lane and Thomas 1983). Metagnostidae is used instead of Geragnostidae, following Lortey ( 1980u). The Hammatocnemidae Kielan, 1960 is regarded as an independent family following Apollonov (1974) and Lu (1975); a more recent study (Lu and Zhou 1979) shows that it is closer to the cheirurids than to the phacopids. Specimens described and listed in this account are deposited at institutions designated by the following abbreviations: NI, Nanjing Institute of Geology and Palaeontology, Academia Sinica; XTR, Regional Geo- logical Survey Team of Xinjiang; XITr, Xi’an Institute of Geology and Mineral Resources; II IV, Yichang Institute of Geology and Mineral Resources; IGG, Institute of Geology and Geophysics, Siberian Branch of the Academy of Sciences of the U.S.S.R. Figured specimens were first blackened with ink and then coated with magnesium oxide before being photographed. Family metagnostidae Jaekel, 1909 Subfamily metagnostinae Jaekel, 1909 Genus arthrorhachis Hawle and Corda, 1847 Type species. Battus tardus Barrande, 1846, from the Kraluv Dvur Formation (Ashgill) of Liomysl, near Zdice, Czechoslovakia. Remarks. We follow Fortey (1980rt) in restricting the genus Trinodus to the holotype of the type species, T. agnostif orniis M^Coy, 1846. Arthrorhachis cf. /arr/u (Barrande, 1846) Plate 58, figs. 3, 4 Figured specimens. One cephalon (NI 80590) and one pygidium (NI 80591 ) from Bed 12. Description and remarks. Arthrorhachis tarda and allied forms have been recorded from the Ashgill of Bohemia (Barrande 1852), Poland (Kielan 1960), South Wales (Dean 1971), North Wales (Whittington 1968), northern England (Ingham 1970), Kazakhstan (Apollonov 1974), and Uzbekistan (Abdullaev 1972), and from the late Caradoc to Ashgill of Norway (Owen and Bruton 1980). The lectotype of A. tarda, refigured by Pek (1977, pi. 8, fig. 2), and topotypes described by Whittington (1950) are comparable with the present form but the latter has a wider (tr.) glabella, in which respect it is similar to specimens described by Whittington (1968) and by Dean (1971). Traces of five pairs of muscle scars on the glabella posterior to the glabellar furrow are comparable with those on a well-preserved specimen of A. danica brevis Fortey (1980u, pi. 2, fig. 10). Patterns of glabellar furrows and muscle scars are not known in A. tarda but traces on a cranidium figured by Whittington (1968, pi. 29, fig. 10) seem compatible with those of the present form. As Owen and Bruton (1980) concluded, species of Arthrorhachis (as Trinodus) from the late Ordovician of Europe need to be revised, as do species from the Caradoc to Ashgill of China and central Asia. In specimens of A. cf. tarda from the Ashgill of Ireland (Dean 1971, p. 7) the length of the glabella varied from 63 % to 70 % that of the cephalon, while the pygidial axis was 38 % to 50 % of the pygidial length. In A. aff. tardus from the Upper Chasmops Limestone of Norway (Owen and Bruton 1980) the corresponding figures were 55 % and 55%. Trinodus subtardus Petrunina in Repina et al. 1975, Trinodus cylindricus Chen in Li et al. 1975, Trinodus corrugatus Chen in Li et ai 1983 and Trinodus carinatus Ju in Qiu et al. 1983 fall within the range of 748 PALAEONTOLOGY, VOLUME 29 variation above, and other characters are also similar to those of A. tarda. All these species are allied to, and may even be referable to, the type species of Arthrorhachis. Genus geragnostus Howell, 1935 Type species. Agnostus sidenhladhi Linnarsson, 1869 from the late Tremadoc (Apatokephalus serratus Zone) of Mossebo, Vastergotland, Sweden (seeTjernvik 1956, pi. 1, figs. 5 and 6). Geragnostus alT. longicollis (Raymond 1925) Plate 58, figs. 1, 2, 5 Figwed specimens. Two cephala (NI 80587, 80588) and one pygidium (NI 80589) from Bed 4. Remarks. The cephalon strongly resembles that of Arthrorhachis tarda (Barrande) but shows an additional tiny basal glabellar tubercle and a more flattened border. The pygidial axis occupies 70 % of the pygidial length and has a tiny median tubercle near the tip. The posterior axial segment, with length about half that of the axis, is much longer than in Arthrorhachis tarda and the Chinese form is referred to Geragnostus rather than Arthrorhachis. As pointed out by Dean (1966) and Fortey (1980u) features such as relative length of pygidial axis are insufficient to separate these genera and Dean (1966) suggested that Geragnostus may prove eventually to be a junior subjective synonym of Arthrorhachis (as Trinodus). The present form recalls G. longicollis (Raymond 1925, p. 12, pi. 1, fig. 5; Whittington 1965, p. 301, pi. 1, figs. 112, 14, 16, 17) from the Table Head Formation (approx, lower Llanvirn) of western Newfoundland, and an allied form from the north-western Yukon described by Dean (1973, p. 2, pi. l,figs. 1-6). G. c/umv Whittington ( 1963, p. 28, pi. l,figs. 1-17), from the Llanvirn of Lower Head, western Newfoundland, and G. symmetricus Zhou {in Zhou et cd. 1982, p. 215, pi. 57, figs. 2 and 3) from the Llanvirn of north-west China differ only in the more narrowly rounded tip of the pygidial axis and the narrower cephalic and pygidial border furrows. These three species are very alike and, as discussed by Whittington (1965, pp. 301, 302) and Zhou (in Zhou et cd. 1982, p. 216), distinguishable only on the basis of minor characters. The well-rounded anterolateral eephalic and posterolateral pygidial angles suggests that the present form is closest to G. longicollis. The type species of Geragnostella Kobayashi, 1939, Agnostus tullhergi Novak, 1883 from the Sarka Fm. (Llanvirn) of Bohemia and redescribed by Pek (1977), exhibits an Arthrorhachis-Gerag- nostus type of glabella with indications of a V-shaped glabellar furrow in front of the median tubercle. Its pygidium is, as noted by Pek, characterized by the long, narrow posterior axial ring but agrees more or less with that of the present form. The length of the pygidial axis is variable in agnostids (Fortey 1980u, p. 26) and the relative length of the posterior axial ring may vary even within a species, as Whittington (1963, p. 29) pointed out for Geragnostus clusus. The great similarity between Geragnostella tullhergi and our speeies supports the view (Dean 1966, p. 273) that Geragno- stella may be a junior subjective synonym of Geragnostus. EXPLANATION OF PLATE 58 Figs. 1, 2, 5. Geragnostus alT. longicollis (Raymond, 1925). Bed 4. 1, pygidium, NI 80587, x 10. 2, cephalon, NI 80588, xIO. 3, cephalon, NI 80589, x 6. Figs. 3 and 4. Arthrorhachis cf. tarda (Barrande, 1846). Bed 12. 3, cephalon, NI 80590, x 12. 4, pygidium, NI 80591, X 12. Figs. 6-9, 13. Slnmiardia tarimuensis Zhang, 1981. Bed 4. 6, cranidium, NI 80592, x 12. 7, pygidium, NI 80593, xl2. 8, cranidium, NI 80594, x 12. 9, pygidium, NI 80595, x 12. 1 3, pygidium, NI 80596, x 12. Figs. 10, 12, 14-19. Pliorocephala quadrata sp. nov. Bed 12. 10, 14, pygidium; dorsal and right lateral views, holotype, NI 80597, x 5. 12, juvenile cranidium, paratype, NI 80598, x 12. 15 and 16, cranidium, dorsal and right lateral views, paratype, NI 80599, x 8. 17, pygidium, paratype, NI 80600, x 6. 18, cranidium, paratype, NI 80601, x 8. 19, cranidium, paratype, NI 80602, x 10. Fig. 1 1. Telephina convexa Lu, 1975. Bed 12. Cranidium, NI 80603, x 5. PLATE 58 ZHOU and DEAN, Ordovician trilobites 750 PALAEONTOLOGY, VOLUME 29 Family shumardiidae Lake, 1907 Genus shumardia Billings, 1862 Type species. Shumardia granulosa Billings, 1862 from the Shumardia Limestone (lower Llanvirn) of Levis, Quebec, Canada. Shumardia tarimuensis Thding, 1981 Plate 58, figs. 6-9, 13 1981 Shumardia tarimuensis Zhang, p. 163, pi. 61, figs. 7 and 8. Diagnosis. Species of Shumardia without posterolateral glabellar lobe. Pygidium stout with narrow (tr.) axis. Hoiotype. XTR 131, incomplete dorsal shield (Zhang 1981, pi. 61, fig- 7) from the Saergan Formation (Llanvirn to lower Caradoc) of Kanling, Keping County, Zinjiang, China. Figured specimens. Two cranidia (NI 80592, 80594) and three pygidia (NI 80593, 80595, 80596) from Bed 4. Description. Cranidium twice as broad as long, semicircular in outline, gently convex, declined slightly both anteriorly and laterally. Glabella with outline like arrow-head occupies two-thirds length and one-quarter posterior width of cranidium; it is convex, highest at midpoint of posterior margin, from which it declines forwards. Anterior part of glabella V-shaped frontally, with pair of prominent, tear-shaped anterolateral lobes, the overall width of which is about one-and-a-half times basal glabellar width; posterior half of glabella tapers forwards into deep furrows at base of anterolateral lobes. Occipital ring transversely convex with length (sag.) one-fifth that of cranidium; its posterior margin is bowed backwards immediately behind posteriorly placed median tubercle; occipital furrow deep. Axial furrows deep and broad beside posterior half of glabella but deep and narrow anteriorly. Preglabellar field narrow (sag.) with deep mesial notch. Posterior border widens (exsag.) distally; posterior border furrow is narrow proximally and dies out distally. Pygidium triangular in outline, 1 -2 times wider than long. Conical axis occupies little more than half length of pygidium and one-third its breadth measured across anterior margin; there are four axial rings and small rounded terminal piece, all delimited by deep ring furrows, in addition to articulating half-ring. Axial furrows deep and wide. Pleural lobes exhibit up to five pairs interpleural furrows that decrease in size posteriorly; first pair extend near facets but fifth pair appear only as short indentations. No pleural furrows visible except first pair, which are deep and transverse. Surface granular except for pygidial border, which is smooth. Remarks. Although Conophrys Callaway, 1877, type species C. saiopiensis Callaway from the Upper Tremadoc of Shropshire, and Shumardia are sometimes regarded as possibly distinct (e.g. Dean 1973, p. 8) they have also been considered congeneric (Poulsen in Moore 1959, p. 0245). Judging from the long triangular pygidium of S. tarimuensis we consider it to be a typical Shumardia, and species with similar pygidia include not only the type species S. granulosa but also S. dicksoni Moberg and Segerberg, 1906 and S. lacrima Koroleva, 1964. Of these the Chinese species bears the greatest resemblance to S. granulosa (see Whittington 1965, p. 327, pi. 16, figs. 1-17) but differs in the absence of posterolateral glabellar lobes, the less acutely pointed front of the glabella, and the relatively shorter pygidium with narrower axis. Family komaspididae Kobayashi, 1935 Genus phorocephala Lu in Lu et al., 1965 Type species. Phorocephala typa Lu in Lu et al. 1965, from the Siliangssu Formation (upper Arenig) of Liangshan, Shaanxi Province, China. Remarks. Tripp (1976, p. 423) suggested that Carrickia Tripp, 1965, type species C. pelagia Tripp, 1965 from the Caradoc of Scotland, is a junior synonym of Phorocephala. We tentatively consider both genera as synonymous, following Zhou, Yin, and Tripp (1984, p. 23). On the basis of stratigraphic occurrence and the presence or absence of a preglabellar field, described species of Phorocephala fall into two groups: ZHOU AND DEAN: ORDOVICIAN TRILO BITES 751 1. Upper Tremadoc to Llandeilo species with preglabellar field. These include; P. similis Lu, 1975; P. genalata Lu, 1975; P. shizipuensis Yin in Yin and Lee 1978; P. gracilis Zhou in Zhou et al. 1982; Bathyurus mansuyi Reed, 1917; Leiostegiuml cf. mansuyi (Reed) of Weber 1948; P. typa Lu in Lu et al. 1965; Carrickia setoni Shaw, 1968; Goniophrys hreviceps (Billings 1865); Garrick ia sp. 1 of Ross 1972, p. 29, pi. 10, fig. 19; undetermined genus and species A of Ross 1951, pi. 18, figs. 21, 23, 24; and, questionably, Bathyurus hreviceps Billings, 1865. 2. Caradoc to Ashgill species without preglabellar field. These include; Carrickia pelagia Tripp, 1965; C. athleta Dean, 1971; C. pinguimitra Chatterton and Ludvigsen, 1976; C. ulugtana Petrunina in Repina et al., 1975; P. quadrata sp. nov.; Phorocephala sp. of Owen and Bruton 1980; and gen. indet. of Abdullaev ( 1 972, pi. 46, figs. 9-11). In earlier species such as P. genalata Lu (from the middle Arenig of Hubei Province, China) and the unnamed species of Ross (1951, from letter Zone F of the Canadian Series in north-eastern Utah, USA) the glabella is relatively smaller and the palpebral lobes shorter than in younger species. Morphological changes that take place during the ontogeny of Phorocephala, as exemplified by P. pinguimitra and described by Chatterton (1980, p. 31), include the medial disappearance of the preglabellar field, and the increase in relative size of glabella (tr. and sag.) and palpebral lobes (exsag.). Similar changes would have taken place during the phylogeny of Phorocephala. Phorocephala quadrata sp. nov. Plate 58, figs. 10, 12, 14 -19 Derivation of name. G&tm—qiiadratus, a, urn, square, referring to the shape of the combined glabella and occipital ring. Holotype. Pygidium, NI 80597 (PI. 58, figs. 10 and 14) from Bed 12. Paratypes. Four cranidia (NI 80598, 80599, 80601, 80602) and one pygidium (NI 80600) from Bed 12. Diagnosis. Phorocephala species with cranidium about half as long as wide, and transversely straight anterior border; glabella large, subquadrate, its width about half that of cranidium. Pygidium as wide as long, composed of three segments; axis trapezoidal in cross-section. Description. Cranidium about twice as wide as long, trapezoidal in plan, widest across base. Glabella convex, subquadrate, defined by deep axial furrows, rounded anterolaterally, its width about half that of cranidium. Occipital ring wider (tr.) than base of glabella, with length (sag.) less than one-fifth that of cranidium, gently convex posteriorly in plan, becoming slightly shorter (exsag.) abaxially. Occipital furrow deep, broad, straight. Anterior border straight, well defined by deep anterior border furrow, uniformly narrow (sag.), upturned. Posterior border narrows (exsag.) laterally, bounded by deep, broad posterior border furrow. Preglabellar field absent in mature cranidia. Fixigenae wide posteriorly, narrow forwards, moderately convex, declined abaxially. Palpebral lobes separated from fixigenae by narrow, deep palpebral furrows and extend forwards from just in front of posterior border furrow almost to anterior border furrow; they are widest (tr.) opposite centre of glabella, converge gently forwards, at the same time becoming steadily narrower, and posterior portions are strongly convex abaxially. Anterior branches of facial suture very short, strongly convergent frontally; pos- terior branches short, divergent posteriorly. Surface of exoskeleton densely covered with pits that are circular except on occipital ring, where transversely elongated. Two pairs muscle scars represented by smooth, unpitted patches on external surface; anterior pair (2p) triangular, narrow adaxially, located on anterolateral margin and extend adaxially backwards; posterior (Ip) pair subrectangular, widen (exsag.) abruptly adaxially, begin in axial furrows at points just in front of line through centre of glabella, and run subparallel to 2p muscle scars; in large specimens the Ip scars deepen and coalesce with pair of apodemal pits (at adaxial extremities). A juvenile cranidium (PI. 58, fig. 12) has a narrow (sag.) preglabellar field and the glabella is subspherical. Pygidium subelliptical, as long as wide, strongly convex transversely. High, tapered axis three-quarters as wide as long, and occupies about two-thirds anterior pygidial width; it consists of two axial rings, terminal piece, and articulating half-ring. Axis is subtrapezoidal in cross-section, divided into three bands by two closely spaced, subparallel, longitudinal ridges. Medial part is flat; the lateral bands are steeply declined abaxially; and in large specimens two longitudinal rows of nodes are arranged along the ridges. First two axial rings uniformly long (sag.), arched forwards in plan, well defined by narrow ring furrows; third ring visible, though faintly demarcated, on large specimens, and tip of axis slopes down posteriorly to merge 752 PALAEONTOLOGY, VOLUME 29 smoothly with postaxial area. Axial furrows narrow, weaker posteriorly. Pleural lobes include three pairs of pleurae that extend posteriorly, but only anterior two pairs well defined. Three pairs deep, short (tr.) pleural furrows die out at midpoints of pleurae, each of which is divided into raised anterior and depressed posterior pleural bands. Three pairs of interpleural furrows present; anterior two pairs distinct, narrow, and reach pygidial margin; posterior pair wide, shallow, and disappear before attaining tip of axis. Articulating facets broad (exsag.), and decline from fulcra sited near front ends of axial furrows. Pygidial border apparently absent, but may be represented by narrow strip at margin. Surface of exoskeleton densely granulate. Remarks. The new species generally resembles P. pinguimitra (Chatterton and Ludvigsen 1976, p. 44, pi. 17, figs. 1-50) from the Chazy of north-west Canada but in the latter the anterior border of the cranidium is arched forwards in plan, the pygidium is wider (tr.), with only two segments, and the pygidial axis is shorter (sag.), semicircular in cross-section, with sigmoidal ring furrows. P. athleta (Dean 1971, p. 48, pi. 23, figs. 2-9; pi. 24, figs. 1-4, 6, 8, 11), from the Ashgill of Ireland, is distinguished from P. quadrata by its frontally rounded glabella, the anterior border that is curved backwards abaxially, and the wider pygidium with two segments and a stouter axis that is semi- circular in cross-section. Family TELEPHiNiDAE Marek, 1952 Genus telephina Marek, 1952 Type species. Telepinis fractiis Barrande, 1852, from the Kraluv Dvur Formation (Ashgill) and Nucice Beds (late Caradoc) of Bohemia. Telephina convexa Lu, 1975 Plate 58; fig. 11; Plate 59, figs. 1 and 4 1975 Telephina (Telephina) convexa Lu, p. 294, pi. 2, figs. 25 and 26. 1977 Telephina (Telephina) convexa Lu; Zhou et al., p. 190, pi. 56, fig. 6. 1978 Telephina (Telephina) convexa Lu; Xia, p. 158, pi. 28, fig. 9. 1983 Telephina (Telephina) convexa Lu; Qiu et ah, p. 170, pi. 57, fig. 14. Diagnosis. Telephina species with strongly angulate palpebral lobes. Glabella and occipital ring covered with web-like raised lines and scattered coarse granules on external surface, and coarse granules on exfoliated surface. Holotype. Cranidium, NI 16419, figured Lu (1975, pi. 2, figs. 25 and 26), from the Linhsiang Formation (early Ashgill) of Fenxiang, Yichang, Hubei Province, eastern China. Figured specimens. Two cranidia (NI 80603, 80604) from Bed 12. Remarks. The smaller cranidium (PI. 58, fig. 1 1) differs from the larger in that the preglabellar area is narrower and the glabella has a pair of shallow impressions. In the light of morphological changes during the ontogeny of T. hicuspis (Angelin 1854) demonstrated by Nikolaisen (1963, pp. 365-366), we consider differences between our two specimens to be due to differences in size. The external surface is covered with raised lines in a multiple web-like pattern; individual lines radiate from coarse granules on the glabella and occipital ring and from fine and scattered coarse granules on the fixigenae and palpebral lobes. The exfoliated surface shows low granules that are well marked on glabella and occipital ring. Anterior border and frontal spines exhibit raised lines parallel to their anterior margins. The pattern of ornamentation is very similar to that of T. americana (Billings 1865) (see Whittington 1965, p. 367, pi. 37; pi. 38, figs. 7-9, 1 1) and of T. sp. of Whittington (1965, p. 371, pi. 38, figs. 1-6, 18), both from the Table Head Formation (Llanvirn) of western Newfoundland, but the Canadian species have fainter granules on the internal mould. The holotype of T. convexa is an internal mould and the surface sculpture is not preserved. The occipital tubercle and small spine seen on the holotype are not preserved on the present material, but otherwise there is good agreement. As noted by Lu (1975) the American species T. gelasinosa (Ulrich in Butts 1926) (see Ulrich 1930, p. 26, pi. 7, figs. 12-14), from the middle Ordovician of Alabama, may be closely related to T. convexa but differs in the rounded anterolateral angles of the cranidium and the densely granular surface of the exoskeleton. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 753 T. convexa is very similar to T. nikolaiseni Apollonov (1974, p. 14, figs. 10, 11, 14) from the Ashgill of Kazakhstan, and the two may prove synonymous, but specimens of the Russian species are too poorly preserved to show the surface ornamentation. According to Apollonov’s description the eranidium of T. nikolaiseni is eovered with what he called ‘capillary’, but it is difficult to see whether or not granules are present and what the pattern of striae is. We refer the Chinese specimens to T. convexa until further material of T. nikolaiseni is available. Telephina sp. Plate 59, figs. 2 and 5 Figured specimen. Cranidium NI 80605, from Bed 4. Remarks. This form may represent a new species but the available material is insufficient. The specimen has a pair of extraordinarily thick, tear-like cephalic frontal spines. In this respect it somewhat resembles T. convexa but differences include: the more tapered glabella with much denser granules on the surface; the more rounded anterolateral corners of the cranidium; and the three pairs of granulose glabellar muscle scars, the anterior two pairs of which are discrete instead of confluent. Similar features are found in the American Caradoc species T. gelasinosa (Ulrich in Butts 1926) but the latter has a pair of more slender frontal spines and its palpebral lobes are semicircular in plan. In the outline of the glabella and size of the fixigenae the Chinese specimen is comparable with T. angulata Yi (1957, p. 554, pi. 3, fig. 4) from the Yangtze region and T. versa Nikolaisen (1963, p. 379, pi. 3, figs. 5-10) from Scandinavia, both of Llandeilo age, but in the two last-named the glabella is relatively shorter (sag.) (length — 0-8 of its width, compared with 0-88), there are more scattered granules on glabella and occipital ring, and the frontal spines are smaller. T. versa also has less angulate palpebral lobes than T. angulata, and in this respect seems closer to our specimen. Family glaphuridae Hupe, 1953 Genus glaphurina Ulrich, 1930 Type species. Glaphurina larnottensis Ulrich, 1930, from the Upper Chazy (approx. Llandeilo) of Isle La Motte, Vermont, USA. Glaphurina sp. Plate 59, fig. 8 Figured specimen. Cranidium, NI 80612, from Bed 4. Remarks. The single, imperfect cranidium closely resembles G. larnottensis Ulrich (1930, p. 45, pi. 8, figs. 14-16; Shaw 1968, p. 29, pi. 8, figs. 10-15; pi. 9, figs. 1-3, 5-8) in the following charac- ters: wide cranidium, broadly and evenly rounded anteriorly; deeply incised longitudinal posterior glabellar furrow continuous with shallow, oblique anterior furrow; deep axial furrows; narrow (sag.), steeply upturned anterior border; divergent anterior branches of facial suture; large posterior area of fixigenae; and surface densely covered with tubereles of two sizes. It differs in having a very narrow (sag.) preglabellar field like that of, for example, the Llanvirn species G. granulosa Zhou in Zhou et al. (1982, p. 254, pi. 63, fig. 15u, b) from north-west China and the early Caradoc species G. dulanensis Weber, 1948 (see Chugaeva 1958, p. 75, pi. 8, figs. 13-15; non Weber 1948, p. 55, pi. 8, figs. 22 and 23) from Kazakhstan. Family ASAPHiDAEBurmeister, 1843 Subfamily asaphinae Burmeister, 1843 Genus BiRMANiTES Sheng, 1934 Type species. Ogygites birmanicus Reed, 1915, from the Hwe Mawng Beds (lower Ordovician) of Hwe Mawng and Hpakhi, northern Shan State, Burma. 754 PALAEONTOLOGY, VOLUME 29 Remarks. Opsimasaphus Kielan, 1960 differs from Pseudohasilicus Reed, 1931 mainly in the longer preglabellar area, and the narrower axis and pygidial doublure (cf. Kielan 1 960, p. 75). These charac- ters, in turn, agree well with those of Birmanites, of which we consider Opsimasaphus a junior subjec- tive synonym. Both genera may eventually prove synonymous with Ogygites Tromelin and Lebesconte, 1876, as discussed by Zhou, Yin and Tripp (1984, p. 17). Birmanites aff. asiaticus (Petrunina in Repina et al. 1 975) Plate 59, figs. 7, 10, 1 1, 13-15 Figured specimens. Two cranidia (NI 80607, 8061 1 ) and three pygidia (NI 80608-8061 1) from Bed 12. Description. Cranidium of low convexity. Glabella subparallel-sided, broadly rounded frontally, three-quarters as long as wide; Ip furrows directed inwards and backwards, distinct anteriorly; Ip lobes triangular; 2p furrows faint, short, close to Ip abaxially, directed anteriorly; prominent median tubercle close behind occipital furrow; median ridge in front of frontal glabellar lobe. Occipital ring slightly convex transversely, with straight posterior margin; occipital furrow shallow. Axial furrows wide and shallow. Narrow, elongate bacculae sited opposite Ip lobes. Palpebral lobes semicircular in outline, narrow (tr.), weakly defined by obscure palpebral furrows; length one-fifth that of glabella, their anterior extremities opposite adaxial ends of 1 p furrows and at one-third cranidial length from posterior margin. Preglabellar area slightly depressed, turned up frontally into weakly defined anterior border, its length one-third that of glabella. Anterior branches of facial suture diverge forwards for short distance, then curve forwards to anterior border furrow, and finally turn inwards to meet medially at obtuse angle; posterior branches sinuous and divergent. Posterior areas of fixigenae narrow (exsag.) with uniformly wide border, well defined by border furrow. Pygidium semielliptical in outline, about three-fifths as long as wide; anterolateral facets occupy half width pleural lobes. Convex axis, one-third to one-fourth length and one-fourth width of pygidium, tapers backwards, divided into seven to eight axial rings and small rounded terminal piece; transverse ring furrows shallow on external surface, distinct on internal mould; axial furrows broad. Pleural lobes flat with five pairs ribs defined by pleural furrows, only first pair of which is deeply incised. Border furrow obscure. Doublure extends to midlength of pygidium and occupies half width of pleural lobes anteriorly. Surface covered with irregular, fine, raised lines, subparallel to margins, and densely arranged only on pre- glabellar area and that part of pygidial border above the doublure. Remarks. The cranidium is characterized by a preglabellar area that is relatively short (sag.) compared with most species in the genus, but is close to those of Birmanites latus (Angelin 1851, p. 14, pi. 10; Kielan 1960, p. 78, pi. 6, figs. 1 and 2; pi. 7, fig. 3; pi. 8, fig. 4) from the Red Tretaspis Mudstones (Ashgill) of Vastergdtland, Sweden and B. asiaticus (Petrunina in Repina et al. 1975, p. 186, pi. 32, figs. 1-9) from the Kielanella-Tretaspis Beds (Ashgill) of southern Tian-Shan. The pygidium of B. latus is much shorter (sag.) than in the present form; that of B. asiaticus is comparable but slightly longer and has seven instead of five pairs of pleural ribs. Birmanites sp. indet. Plate 59, figs. 9 and 1 2 Figured specimens. One hypostoma (NI 80613) and one pygidium (NI 80614) from Bed 4. EXPLANATION OF PLATE 59 Figs. 1,4. Telephina convexahu, 1975. Bed 12. Cranidium, dorsal and anterior views, NI 80604, x4. Figs. 2, 5. Telephina sp. Bed 4. Cranidium, dorsal and anterior views, NI 80605, x 4. Figs. 3, 6. Nileus Inianxianensis Zhou in Zhou et al. 1 982. Bed 1 2. Right librigena and attached cephalic doublure, NI 80606, x2. Figs. 7, 10, 11, 13-15, Birmanites aff. asiaticus (Petrunina in Repina et al. 1975). Bed 12. 7, cranidium, NI 80607, X 1-5. 10, pygidium, NI 80608, x 2. 1 1, pygidium, NI 80609, x4. 13, pygidium, NI 80610, x 1. 14 and 15, cranidium, dorsal and left lateral views, NI 8061 1, x 2. Fig. 8. Glaphurina sp. Bed 4. Cranidium, NI 80612, x 8. Figs. 9, 12. B//7;7nn;7f.s sp. indet. Bed 4. 9, hypostoma, NI 80613, x 5. 12, pygidium, NI 80614, x4. PLATE 59 ZHOU and DEAN, Ordovician trilobites 756 PALAEONTOLOGY, VOLUME 29 Remarks. The well-preserved hypostoma has the middle body poorly defined, with middle furrow deeply incised abaxially; the posterior margin is forked, and the whole surface is covered with densely arranged, transverse terrace lines. The pygidium has a narrow, convex axis and slightly depressed border; it is semicircular in outline with nine pairs of pleural ribs. The only comparable described species is B.yangtzeensis Lu ( 1975, p. 321, pi. 8, figs. 9-13; pi. 9, figs. 1-5) from the Miaopo Formation ( Llandeilo) of western H ubei, China, which has a relatively broader pygidial border and only six pairs of pleural ribs. Family NiLEiDAE Angelin, 1854 Genus NiLEUS Dalman, 1827 Type species. Nileus armadillo Dalman, 1 827, probably from the E.xpaiisus Limestone (Arenig; see Tjernvik 1956, p. 208) of Husbyfjol, Ostergotland, Sweden. Nileus huanxianensis Zhou in Zhou et al. 1982 Plate 59, figs. 3 and 6; Plate 60, figs. 1 -6, 8, 1 1 1982 Nileus huanxianensis Zhou in Zhou et al., p. 266, pi. 66, fig. 9. Diagnosis. Species of Nileus with cranidium as long as wide, rounded anteriorly. Median glabellar tubercle placed anterior to line between posterior ends of palpebral lobes. Pygidium with shallow, broad border furrow. Surface of cranidium smooth; pygidium covered with densely arranged fine ridges subparallel to margin. Holotype. XI Tr-139, cranidium, figured Zhou (in Zhou el al. 1982, pi. 66, fig. 9) from Bed 12 of the Chedao Formation, Chedao, Huanxian County, Gansu Province, China. Figured specimens. Two cranidia (NI 80615, 80622), two hypostomata (NI 80618, 80619), one right librigena with cephalic doublure attached (NI 80606), one thoracic segment (NI 80621), two pygidia (NI 80616, 80620) and one pygidial doublure (NI 80617) from Bed 12. Description. Cranidium of low convexity, as long as wide, broadly rounded anteriorly. Median glabellar tubercle situated in front of line joining posterior ends of palpebral lobes. Weak medial glabellar ridge visible only on internal mould. Palpebral lobes semielliptical with length about two-fifths that of cranidium, their posterior ends located one-sixth cranidial length from posterior margin of cranidium; palpebral furrows faint. Axial furrows weak. Anterior branches of facial suture diverge forwards at 60° to 90°; posterior branches diverge strongly backwards. Posterior areas of fixigenae very small, triangular, abaxially declined. Librigenae narrow (tr.) with no visible border; genal fields gently convex transversely; genal angles obtusely rounded; eye socles narrow, separated from genal fields by deep, broad furrows; eyes large, crescentic. Doublure widest (sag. ) frontally, where it equals about one-third cephalic length, narrows sharply abaxially and is gently convex, upturned anteriorly and posterolaterally beneath the librigenae. Posteromedial part of doublure is depressed to form transverse groove immediately in front of transversely straight hypostomal suture. Rostral plate weakly defined by non- functional connective suture; outline trapezoidal, widest (tr. ) frontally, where equals about two-fifths of cephalic width. Hypostoma four-fifths as long as wide, flat, or weakly convex. Middle body ill defined, wider than long, with pair of depressed, elongate maculae sited on lateral margins opposite centre of hypostoma. Anterior wings small, triangular. Border gently convex, narrows anteriorly and posteriorly, and is tripartite posteriorly; posterior margin with pointed, broadly triangular median process; lateral and posterior borders well defined by deep, wide furrows and have low peripheral rim. Thoracic segment (PI. 60, fig. 8) with large articulating half ring well defined by deep, broad, transverse articu- lating furrow. Axial ring convex, with breadth (tr.) three-quarters that of segment overall. Axial furrows faint. Pleurae narrow (tr.) with rounded tips; triangular facets well developed outside fulcra that are sited close to axial furrows. M oderately convex, unsegmented pygidium is subsemicircular in outline, length three-fifths the breadth . Large triangular axis occupies three-fifths anterior width and more than three-fifths length of pygidium, and is weakly defined by very shallow, broad axial furrows. Pleural lobes decline gently towards wide, shallow border ZHOU AND DEAN: ORDOVICIAN TRILOBITES 757 furrow; border indistinct, narrow (sag.), slightly convex. Doublure broad, deeply indented around tip of axis, and inner margin on either side is broadly curved, convex forwards. Surface of cranidium, librigenae and thoracic axial rings smooth. Doublure, hypostoma, and thoracic pleurae covered with terrace lines. Dorsal surface of pygidium has densely crowded, fine ridges subparallel to margins; ridges are slightly thicker at anterolateral corners of pygidium. Remarks. Nileus Imanxianensis differs from N. exarmatus ohsoletiis Chang and Fan (1960, p. 1 10, pi. 12, figs. 13 and 14; pi. 3, figs. 1 and 2; text-fig. 1 1) from the Llanvirn of Qaidam, Qinghai Province, China in having a longer cranidium with anterior margin more strongly convex forwards, and smaller palpebral lobes. The pygidium of the Qinghai form is twice as broad as long with coarser but scattered transverse ridges on the surface. The Arenig subspecies N. glazialis costatus Fortey (1975n, p. 41, pi. 10; pi. 16, fig. 8) from Spitsbergen agrees in many respects with the present species, but differences shown by the former include: the slightly shorter (sag. ) cranidium, steeply declined anteriorly; the less divergent posterior branches of the facial suture; the narrower (sag.) cephalic doublure; and the ab- sence of a pygidial border. Both forms have quite different surface sculpture; the cranidium of N. glazialis costatus is finely punctate and the pleural lobes of the pygidium are covered with distinctive, deep grooves, only a few of which extend across the axis. The Kazakhstan species N. t eugriensisV^ do&r ( 1 948, p. 48, pi. 8, figs. 1 -3, 5 and 6), from the Karakan horizon (Llandeilo), resembles N. luianxianensis in the outline of the pygidium and the size of its axis; but the cranidium has the median tubercle sited further back, and other cranidial features are comparable with those of N. exarmatus ohsoletus. Genus PERASPis Whittington, 1965 Type species. Niobe lineolata Raymond, 1925, from the middle Table Head Formation (Llanvirn), Aguathuna, Port au Port Peninsula, southwestern Newfoundland. Peraspis ohscura sp. nov. Plate 60, figs. 7, 9, 10, 12-15; Plate 61, fig. 1 Derivation of name. Latin obscurus, a, ion, obscure, referring to the indistinct pygidial border furrow. Holotype. Cranidium, N1 80628 (PI. 60, fig. 1 3), from Bed 4. Paratvpes. Two cranidia (NI 80630, 80631a), four pygidia (N1 80623, 80624, 80625, 80629) and two hypostomata (NI 80626, 80627) from Bed 4. Diagnosis. Species of Peraspis with wide pygidial axis that occupies less than two-fifths the pygidial width anteriorly. Pygidial border poorly defined. Cranidium without anterior border. Axial furrows straight and parallel opposite palpebral lobes. Flypostoma transverse with well-defined middle body and wide (tr.) lateral border. Description. Cranidium of low convexity, as long as wide or slightly wider than long, rounded anteriorly. Gently convex glabella widest across anterolateral angles, highest along sagittal line but not carinate; tiny median tu- bercle sited behind transverse line between posterior ends of palpebral lobes. Axial furrows shallow on external surface, deep on internal moulds, parallel and straight between palpebral lobes, obscure between posterior ends of palpebral lobes and posterior cranidial margin, where they end at a pair of pits. Palpebral lobes long (0-4 0-45 cranidial length), semicircular in outline, narrow (tr.), faintly defined by weak palpebral furrows in testaceous specimens; transverse line joining posterior ends cuts sagittal line at about one-fifth of cranidial length from posterior cranidial margin. Palpebral areas of fixigenae slightly convex with width half length of palpebral lobes; posterior areas small, triangular, wider than long (exsag.). Anterior branches of facial suture diverge forwards at 90° to 1 20°; posterior branches diverge strongly backwards. Hypostoma transversely elliptical in outline, four-fifths as long as wide, broadly rounded anteriorly. Middle body convex, oval, longer than wide, well defined by deep, wide lateral furrows; pair of deep middle furrows curve gently backwards from lateral margins of middle body at points one-third its length from rear. Anterior wings small, triangular. Lateral borders gently convex, their width two-fifths that of middle body, narrowing forwards to end near posterior corners of anterior wings; margins rounded. Posterior border narrow (sag.). 758 PALAEONTOLOGY, VOLUME 29 about half width (tr.) of lateral border; posterior margin tripartite with obtusely triangular median point. More or less uniformly narrow rim runs around lateral and posterior borders. Surface covered with coarse, transverse terrace lines. Pygidium weakly vaulted, subsemicircular in plan with length about 0-7 the width. Convex, funnel-shaped axis occupies less than two-fifths anterior width and less than two-thirds length of pygidium; ring furrows almost obsolete. Axial furrows broad, entire. Pleural lobes unfurrowed; facets subtriangular, their width (tr.) two-thirds the anterior width of each pleural lobe. Border indistinct. Surface shows a few raised lines at anterolateral angles and near lateral and posterior margins. Doublure narrow (sag.), about one-third pygidial length at sagittal line; inner margin parallel to that of pygidium. Plate 60, fig. 14 shows an exfoliated transitory pygidium of Peraspis obscura. Remarks. The hypostoma of the new species is of typical N ileus type, as in N. armadillo (Dalman 1 827) (see Poulsen in Moore 1959, fig. 267, la), N. limbatus Brogger, 1882 (see Tjernvik, 1956, pi. 2, fig. 12), N. exarmatus Tjernvik (1956, pi. 2, fig. 16) and N. platys stigmatus Schrank ( 1972, pi. 8, fig. 5). How- ever, its glabella is longer than in Nileus and the median glabellar tubercle is sited further back; the pygidium has a narrow (tr.), funnel-shaped axis and narrower (sag.) doublure with inner margin parallel to that of the pygidium. These features are found in Peraspis (Fortey 1975u, pp. 34, 35), to which genus we prefer to assign the species. Compared with Peraspis lineolata (Raymond 1925) (see Whittington 1965, p. 364, pis. 34; 35; 36, figs. 11 and 12) the present species differs in: absence of anterior cranidial border; narrower (tr.) posterior areas of fixigenae; slightly wider glabella with median tubercle sited further forwards; straight instead of sinuous axial furrows; effaced pygidium with stouter axis; much more faintly defined pygidial border; and much better-defined middle body of hypostoma. P. omega Fortey (1975u, p. 49, pi. 20, figs. 1-11), from the late Arenig to Llanvirn of Spitzbergen, is also comparable with the new species in outline and size of cranidium and glabella. On the other hand the hypostoma and pygidium of P. omega are more similar to those of the type species, and its median glabellar tubercle and palpebral lobes are sited further forwards, so that the posterior areas of the fixigenae are longer (exsag.). Family cyclopygidae Raymond, 1925 Genus cyclopyge Hawle and Corda, 1847 Type species. Egle rediviva Barrande, 1846, from the Cernin Beds (Caradoc) of Trubin, near Beroun, Czecho- slovakia. Cyclopyge cf. recurva Lu in Wang et al., 1962 Plate 61, figs. 2-5, 7-9, 14 Figured specimens. Three cranidia (NI 80632, 80633, 80634), one rostral plate with attached librigenae (NI 80635), and one pygidium (NI 80636), all from Bed 12. Description. Glabella strongly convex (tr.), parabolic in outline, gently tapered forwards, strongly curved down anteriorly and overhangs preglabellar furrow. Ip glabellar furrows shallow, except for deep, pit-like middle portion, directed inwards and backwards, situated one-third of glabellar length from posterior margin; EXPLANATION OF PLATE 60 Figs. 1-6, 8, 11. Nileus huanxianensis Zhou in Zhou et al. 1982. Bed 12. 1, cranidium, NI 80615, x 3. 2, pygidium, NI 80616, x 3. 3, pygidial doublure, NI 80617, x 3. 4, hypostoma, NI 80618, x 5. 5, hypostoma, NI 80619, X 6. 6, pygidium, NI 80620, x 5. 8, thoracic segment, NI 80621, x 3. 1 1, cranidium, NI 80622, X 2. Figs. 7, 9, 10, 12-15. Peraspis obscura sp. nov. Bed 4. 7, transitory pygidium (Degree 6), paratype, NI 80623, X 6. 9, transitory pygidium (Degree 6), paratype, NI 80624, x 6. 10, pygidium, paratype, NI 80625, x4. 12, two hypostomata, paratypes, NI 80626 (above), 80627, x 8. 13, cranidium, holotype, NI 80628, x 8. 14, transitory pygidium (Degree 4), paratype, NI 80629, x 8. 15, cranidium, paratype, NI 80630, x 8. PLATE 60 ZHOU and DEAN, Nileus, Peraspis 760 PALAEONTOLOGY, VOLUME 29 circular swellings located between deep, pit-like part of Ip furrows and posterior end of cranidium; median glabellar tubercle in-line with distal ends of Ip furrows. Palpebral lobes narrow, band-like, well defined by deeply incised axial furrows, and meet in front of glabella at about 120°. Eyes separated by longitudinal groove that narrows upwards and alongside which there are only four eye facets for each eye. Rostral plate triangular in outline, one fourth as long as wide, with posterior margin slightly convex backwards and posterolateral angles produced to form narrow band that extends along ventral margin of eyes; median furrow broad, triangular, narrows forwards where confluent with groove between eyes; surface covered with transverse raised lines. Pygidium subsemicircular in plan, its dorsal surface declined posteriorly. Axis strongly convex, slightly tapered backwards, broadly rounded posteriorly, and occupies about one-third anterior width and half overall length of pygidium; two axial rings and terminal piece are indicated by ornamentation of transverse raised lines, and narrow (sag.) articulating half-ring is raised and gently arched forwards. Axial furrows deep, wide. Postaxial field crossed by sagittal groove that reaches border furrow. Weakly convex pleural lobes carry pair of articulating facets and seven pairs transverse, raised lines on surface; anterior four pairs extend inwards from lateral margins to cross axis and meet, while three posterior pairs become faint towards axis. Border flat with width (sag.) about one-fifth pygidial length, narrows anteriorly and is covered with a few additional raised lines between the pairs of transverse lines noted above. Remarks. The holotype of C. recurva Lu (1975, p. 377, pi. 30, figs. 8-12), from the Pagoda Formation (Caradoc) of southern Shaanxi, is poorly preserved but indistinguishable from the cranidia in our collection. However, librigenae of C. recurva are unknown and only transitory pygidia have been found. Recently published species of Cyclopyge from the Caradoc to Ashgill of north-west China and central and southern Tian-Shan have cranidia similar to that of C. recurva but different pygidia. They include: C. beishanensis Zhou (in Zhou et al. 1982, p. 268, pi. 67, figs. 5 and 6); C. oculus Abdullaev (1970, see Petrunina in Repina et al. 1975, p. 189, pi. 33, figs. 1-9, 11, 12, 14, 18); C. abdullaevi Petrunina (in Repina et al. 1975, p. 187, pi. 31, figs. 11-14, 16, 23); and C. binodosa Zhang (1981, p. 191, pi. 72, figs. 1-3). Of these C. binodosa Zhang (not to be confused with Aeglina binodosa Salter, 1859, later placed in Cyclopyge and finally made type species of Pricyclopyge R. and E. Richter, 1954) is very similar to the present form and may be conspecific with it, but the evidence is as yet insufficient. The pygidium of C. cf. recurva resembles that of C.vigilans (Cooper and Kindle 1936, p. 367, pi. 52, figs. 36, 39, 41-51), from the Ashgill of eastern Canada, in the outline, wide border, and pre- sence of a postaxial depression; the only difference is the bluntly pointed posterior margin of C. vigilans. The glabella of C. cf. recurva is narrower than that of C. vigilans but comparable with that of C. mirabilis (Forbes MS in Salter, 1853) (assigned to Phylacops by Whittard 1952, pi. 32, figs. 6-8) from the Portraine Limestone (Ashgill) of eastern Ireland; the last-named has the glabellar furrows deeply incised throughout their length and its eyes are much broader mesially in anterior view. Genus microparia Hawle and Corda, 1847 Type species. Microparia speciosa Hawle and Corda, 1847, from the Kraluv Dvur Beds (Ashgill) of Kraluv Dvur, Czechoslovakia. Subgenus quadrapyge Zhou in Zhou et al., 1977 Type species. Microparia (Quadrapyge) latilimbata Zhou in Zhou et at., 1977, from the Modaoxi Formation (Caradoc) of Modaoxi, Taojiang County, Hunan Province, China. Diagnosis. Dorsal shield elliptical in plan. Glabella parabolic in outline without glabellar furrows; median tubercle present. Eyes long, separated by frontal glabellar tongue. Thorax of five segments. Pygidium rectangular, well segmented on exfoliated surface; axis entire, rounded posteriorly; well developed border narrows anteriorly; border furrow deep, broad. Remarks. This diagnosis is based on both the type species M. (Q.) latilimbata Zhou (in Zhou et al. 1977, p. 230, pi. 69, figs. 14-16) and M. (Q.) chedaoensis sp. nov. described below. The cephalic characters agree well with those of Microparia (Microparia) but the pygidium of the latter is ZHOU AND DEAN: ORDOVICIAN TRILOBITES 761 semicircular in outline; the axis is triangular and faintly defined; the pleural lobes are almost unfurrowed; and the border furrow is shallower. Microparia {Quadrapyge) cliedaoensis sp. nov. Plate 61, figs. 6, 10, 11, 13, 16 Name. After Chedao, a small town near the measured section. Holotype. Pygidium, N1 80640 (PI. 61, fig. 16), from Bed 4. Paratypes. Two pygidia (NI 80631b, 80639) and two cranidia (NI 80637, 80638), all from Bed 4. Diagnosis. Microparia (Quadrapyge) species with median tubercle situated in centre of glabella when viewed dorsally. Pygidial axis narrow (tr.); width (sag.) of border about one-fifth sagittal length of pygidium. Description. Cranidium longer than wide. Glabella parabolic in outline, slightly tapered forwards, strongly curved down anteriorly, and fused posteriorly with occipital ring; median tubercle located in centre of glabella when viewed dorsally; anterior tongue short (sag.), curved down frontally where extends further forwards than anterior ends of palpebral lobes; glabellar furrows absent. Axial furrows deep beside palpebral lobes, shallow and curve inwards from posterior ends of palpebral lobes to posterior margin of glabella. Palpebral lobes ridge-like and narrow slightly forwards. Anterior branches of facial suture converge gently forwards; posterior branches diverge slightly backwards. Rostral suture short (tr.), transversely straight. Posterior areas of fixigenae small, triangular, low, and decline abaxially. Pygidium rectangular in outline, its length two-thirds the width. Axis tapers backwards and is broadly rounded posteriorly; it occupies half length (sag.) and more than one-third anterior width of pygidium and is faintly defined by shallow axial furrows. In testaceous specimens axis includes a ridge-like articulating half- ring, well-defined first axial ring, and large terminal piece; pleural lobes decline slightly towards border furrow and are unfurrowed except for first pair pleural furrows, which demarcate convex, narrow (exsag.) pair anterior half ribs. In internal moulds the axis is well defined, with five rings and small terminal piece, and the pleural lobes exhibit six broadly furrowed pleurae. Border low, flat, with posterior margin slightly bent down and bluntly pointed; width (sag.) about one-fifth pygidial length at sagittal line but quickly narrows (tr.) forwards; inner margin semi-circular, well defined by deep, broad border furrow. Surface of cranidium covered with widely spaced raised lines near and subparallel to anterior and posterior margins. Pygidium has transverse raised lines on posterior border and postaxial part of pleural fields. Pleural lobes have paired raised lines running parallel to pleural furrows; some raised lines join across axis, and most reach lateral margins of pygidium. Doublure ornamented with terrace lines. Remarks. The new species resembles M. (Q.) latilimhata in many respects, but the latter differs in its wider (tr.) pygidial axis (anterior width two-fifths to one-half that of pygidium) and wider border (about one-third length of pygidium at sagittal line). Known cephala of the type species are flattened and precise comparison is difficult, but the glabella seems much wider than in the new species and the median tubercle is located in front of centre of the glabella. Family STYGINIDAE Vogdes, 1890 Genus pseudostygina Zhou in Zhou et al., 1982 Type species. Pseudostygina lepida Zhou in Zhou et al., 1982, from the upper part of the Chedao Formation (Caradoc) of Chedao, Huanxian County, Gansu Province, China. Diagnosis. Styginid trilobites with well-defined, convex anterior and lateral cephalic borders. Glabellar and occipital furrows absent. Genal angles rounded. Hypostoma with strongly convex, divided middle body, its oval outline strongly convex forwards; anterior wings large, subquadrate; lateral and posterior borders convex. Pygidial axis short (sag.), parallel-sided, rounded posteriorly; pleural lobes unfurrowed; border faintly defined. Remarks. The narrow (tr.) cranidium, with well-defined glabella that expands steadily forwards, suggests that the genus is a typical styginid. Pseudostygina differs from Stygina Salter, 1853 in its 762 PALAEONTOLOGY, VOLUME 29 lack of an occipital furrow, the rounded genal angles, the shorter (sag.) pygidial axis, the absence of first pygidial pleural furrows, and the more weakly defined pygidial border. The preglabellar furrow is obscure in the lectotype of the type species of Stygina, S. latifrons (Portlock 1843) (see Whittington 1950, pi. 172, figs. 1-3) and only faint in certain other specimens (Whittington 1950, pi. 72, figs. 4 and 5; Skjeseth 1955, pi. 1, figs. 3-5, 7 and 8). The anterior pits in the axial furrows of Stygina, emphasised by Skjeseth, are lacking in Psenclostygina; in addition the cephalic border of Stygina is flat or upturned. Raymondaspis Pribyl, 1948 has a well-defined anterior border as in Pseudostygina, and the shape and size of cranidium and glabella in both genera are comparable. Raymondaspis is distinguished by the wider (tr.) pygidium with longer (sag.), triangular axis, the presence of librigenal spines, and the prominent occipital furrow, pygidial border furrow and first pygidial pleural furrows. The hypostoma of Pseudostygina is similar in shape to that of other styginids but is characterized by the convex lateral and posterior borders, the better-defined middle body with maculae on its lateral margins, and the anterior margin, which is more strongly convex forwards medially. These features are also typical of the hypostoma in scutelluids, but in its large, subquadrate anterior wings and elongate, oval middle body the hypostoma of Pseudostygina differs from that of scutelluids but resembles that of illaenids. On the other hand a pair of anterior pits on the cranidium, which is characteristic of styginids (Lane and Thomas 1983), is not found in Pseudostygina, and this may suggest a further similarity to the illaenids. In the general features of its cephalon and pygidium Pseudostygina also resembles illaenid species such as Illaeniis spitiensis Reed, 1912, but the type species, P. lepida, has the glabella well defined laterally as well as frontally, a prominent anterior border, and a parallel-sided pygidial axis with rounded tip. On the whole Pseudostygina is a typical styginid but shares some important characters with illaenids, a conclusion that favours a close relationship between styginids and illaenids, as recently discussed by Lane and Thomas (1983). Pseudostygina lepida Zhou in Zhou et al., 1982 Plate 61, figs. 12, 15, 17-20; Plate 62, figs. 1-5 1982 Pseudostygina lepida Zhou in Zhou et al., p. 273, pi. 68, figs. 5 and 6. Diagnosis. As for genus. Holotype. Cranidium (XI Trl59), figured Zhou (in Zhou et at. 1982, pi. 68, figs. 5a, h), from Bed 12 of the Chedao Formation, Chedao, Huanxian County, Gansu Province, China. Figured specimens. Three cranidia (NI 80641, 80643, 80647), two hypostomata (NI 80648, 80650), one left librigena (NI 80645) and three pygidia (NI 80642, 80644, 80649), all from Bed 12. EXPLANATION OF PLATE 61 Fig. 1. Peraspis obscura sp. nov. Bed 4. Paratype cranidium, NI 80631a, with associated paratype pygidium, NI 80631b, of Microparia (Quadrapyge) chedaoensis sp. nov., x 6. Figs. 2-5, 7-9, 14. Cyclopyge cf. recurva Lu, 1962. Bed 12. 2, cranidium, NI 80632, x 5. 3, 4, 7, cranidium, dorsal, right lateral and anterior views, NI 80633, x 6. 5, cranidium, NI 80634, x 8. 8, 9, rostral plate with eyes attached, ventral and anterior views, NI 80635, x 5. 14, pygidium, NI 80636, x 8. Figs. 6, 10, 11, 13, 16. Microparia (Quadrapyge) chedaoensis sp. nov. Bed 4. 6, cranidium, paratype, NI 80637, x 6. 10, 11, cranidium, dorsal and right lateral views, paratype, NI 80638, x 10. 13, pygidium, paratype, NI 80639, X 10. 16, pygidium, holotype, NI 80640, x 10. Figs. 12, 15, 17-20. Pseudostygina lepida Zhou in Zhou et al. 1982. Bed 12. 12, small cranidium, NI 80641, X 6. 15, pygidium, NI 80642, x 3. 17, 18, cranidium, right lateral and dorsal views, NI 80643, x 3. 19, small pygidium, NI 80644, x 8. 20, left librigena, NI 80645, x 5. Fig. 21. Stenopareia alf. /rmrmonn/ (Salter, 1848). Bed 12. Cranidium NI 80646, x4. PLATE 61 ZHOU and DEAN, Ordovician trilobites 764 PALAEONTOLOGY, VOLUME 29 Description. Cranidium broadly rounded frontally, rounded anterolaterally, gently convex in lateral view, longer than wide, widest across base. Glabella moderately convex, expands forwards from opposite palpebral lobes, where its width (tr.) is about three-fifths the maximum, attained near anterior margin; its outline is broadly rounded anteriorly, rounded anterolaterally, and there are no glabellar furrows. An oval anterior muscle scar on the exfoliated surface of a cranidium (PI. 62, fig. 1) and a pair of smooth patches on the glabellar flanks opposite the palpebral lobes of a well-preserved specimen (PI. 61, fig. 12) provide evidence of two subparallel rows muscle scars. Occipital furrow absent. Axial furrows wide, deep posteriorly but gradually shallow and narrow forwards. Palpebral lobes small, semicircular, length about one-sixth that of cranidium, strongly convex, standing high above fixigenae and glabella and situated well back. Anterior branches of facial suture long, gently divergent, running forwards in abaxially convex curves; very short posterior branches extend abaxially backwards. Preglabellar field absent. Anterior border narrow, ridge-like, uniform in width (sag.), well defined by distinct anterior border furrow. Fixigenae narrow (tr.); low, narrow (exsag.), triangular posterior areas include pair subcircular alae located adjacent to axial furrows. Librigenae with rounded genal angles; genal fields broad, subquadrate, vaulted; eye socles narrow, vertical, bounded by deep, broad furrows; reniform eyes stand high above genal fields, and visual surface is crowded with very small facets. Lateral border narrow, upturned; posterior border weakly convex, markedly wider (exsag.) abaxially; border furrows broad. Cephalic doublure steeply declined, densely covered with terrace lines. Hypostoma shield-shaped, arched forwards medially, rounded posteriorly. Anterior wings large, subquad- rate. Middle body four-fifths as long as wide, oval, well defined by deep, broad border furrows; it is divided into a large, strongly convex anterior lobe, length (sag.) three-quarters that of middle body, and a small, weakly convex posterior lobe; transverse median furrow is deep and wide laterally, shallow mesially, and slightly curved forwards; transversely oval maculae prominent, sited at anterolateral corners of posterior lobe. Anterior border narrow (sag.), flat; lateral borders subparallel anteriorly, converge opposite median furrow, and then confluent with wide (sag.), strongly convex posterior border. Large pygidium semielliptical in outline, with length three-quarters width; small specimen sub-semicircular (PI. 61, fig. 19), three-fifths as long as wide; anterolateral facets triangular, their width (tr.) one-eighth that of pygidium. Short (sag.) parallel-sided axis occupies one-third length and slightly more than one-quarter width of pygidium; it is rounded posteriorly and unfurrowed except for faint first ring furrow on external surface. Axial furrows deep, broad, but shallow posteriorly, and there is a post-axial ridge. A pair of oval muscle scars close to posterior part of axial furrows is visible, particularly on exfoliated surfaces. Pleural lobes unfurrowed, their dorsal surface declined both abaxially and towards anterior part axial furrows; border furrow indistinct. Doublure with evenly curved inner margin, narrowing (tr.) forwards; width (sag.) at sagittal line less than one-third pygidial length. Surface of cranidium covered with densely grouped transverse raised lines in small specimen (PI. 61, fig. 12); large specimens have only a few similar raised lines on anterior part of glabella and on palpebral areas of fixigenae. Genal fields of librigenae have anastomosing ridges subparallel to lateral margins; posterior border ornamented with scattered anastomosing ridges subparallel to posterior border furrow. Hypostoma has coarse ridges, subparallel to lateral and posterior margins, on anterior lobe of middle body, and on lateral and posterior borders. In most examples of the pygidium, the frontal portion carries three pairs widely spaced, transverse ridges, subparallel to anterior margin. EXPLANATION OF PLATE 62 Figs. 1-5. Pseudostygina lepida Zhou in Zhou el al. 1982. Bed 12. 1, cranidium, NI 80647, x 4. 2, hypostoma, NI 80648, X 6. 3, 4, pygidium, dorsal and left lateral views, NI 80649, x 4. 5, hypostoma, NI 80650, x 8. Figs. 6-11, 16. Stenopareia aff. howmanni (Salter, 1848). Bed 12. 6, pygidium, NI 80651, x 2. 7, pygidium, NI 80652, X 1-5. 8, 9, small cranidium, dorsal and right lateral views, NI 80653, x4. 10, pygidium showing doublure, NI 80654, x2. 1 1, small cranidium, NI 80655, x4. 16, transitory pygidium, NI 80656, x 12. Fig. 12. Stenopareia sp. indet. Bed 4. Cranidium, NI 80657, x 2. Figs. 13-15. ParaphiUipsinella globosa Lu in Lu & Chang 1974. Bed 12. 13, 14, cranidium, dorsal and left lateral views, NI 80658, x 12. 15, cranidium, NI 80659, x 12. Figs. 17 and 18. Xenocybel sp. Bed 12. 17, cranidium, NI 80660, x 10. 18, pygidium, NI 80661, x 10. Fig. 19. Rorringtonia sp. Bed 4. Cranidium, NI 80662, x 10. Fig. 20. Ischyrophyma'l zhiqiangi sp. nov. Bed 12. Cranidium, paratype, NI 80663, x 10. PLATE 62 ZHOU and DEAN, Ordovician trilobites 766 PALAEONTOLOGY, VOLUME 29 Family illaenidae Hawle and Corda, 1847 Genus stenopareia Holm, 1886 Type species. Illaenus linnarssoni Holm, 1882 from the Boda Limestone (Ashgill) of the Siljan district, Dalarne, Sweden. Stenopareia howmanni {Sa\\.cr, 1848) Plate 6 1 , fig. 2 1 ; Plate 62, figs. 6-11,16 Figured specimens. Three cranidia (NI 80646, 80653, 80655) and four pygidia (NI 80651, 80652, 80654, 80656) from Bed 12. Description. Cranidium wider than long, broadly and evenly rounded frontally in plan. Glabella slightly less convex posteriorly, its basal width twice that of each palpebral area. Broad, shallow axial furrows converge forwards until level with middle of palpebral lobes and then diverge slightly forwards, dying out in front of line joining mid-points of palpebral lobes. Palpebral lobes about one-sixth length of cranidium, gently convex abaxially, placed far back, so that the posterior end of each lobe is about half its own length from the posterior margin. Surface of internal mould shows pair of elliptical lunettes (Whittington 1954, p. 139) opposite palpebral lobes, small posteriorly situated median glabellar tubercle, weakly defined median glabellar ridge, four pairs faint muscle scars that decrease in size forwards, and broad (exsag.) posterior border furrow. Pygidium semielliptical in outline, truncated anterolaterally by weakly defined facets. Axis occupies half anterior width of pygidium, but is otherwise poorly defined. Doublure U-shaped, narrow (tr.) anteriorly, wide (sag.) posteriorly where it occupies about half pygidial length; inner margin of the doublure gently convex forwards medially; its surface is ornamented with faint terrace lines. A small transitory pygidium (PI. 62, fig. 16) is semicircular in outline, length less than three quarters its width, steeply curved down to the margins; the well defined axis occupies 0-4 width and 0-7 length of the pygidium. Surface of anterior third of largest cranidia carries terrace lines subparallel to anterior margin; similar lines occur in fixigenae of smallest cranidium (PI. 62, fig. 8). Remarks. As far as we are aware, only three other species of Stenopareia have a pygidial doublure similar to that of the present form. They are: Stenopareia bowmanni (Salter 1848) (see Whittard 1961, p. 217, pi. 31, figs. 1, 2; Price 1974, p. 842, pi. 112, figs. 1-8, 9?) from the Ashgill of South Wales; Stenopareia miaopoensis Lu (1975, p. 387, pi. 32, fig. 13; pi. 33, figs. 1-3) from the lower Miaopo Formation (Llandeilo) of western Hubei, China; and Stenopareia^ sp. of Dean (1978, pi. 50, figs. 4, 6, 8) from the Chair of Kildare Limestone (Ashgill) of eastern Ireland. In the two last species the inner margin of the doublure is much more convex forwards at and near the sagittal line than in S. bowmanni and the present Chinese form. In shape of cranidium and pygidium, and in size of glabella our material resembles S. bowmanni but differs in having larger palpebral lobes; the lectotype of S. bowmanni is somewhat crushed and further comparison is impracticable. Stenopareia sp. indet. Plate 62, fig. 12 Figured specimen. Cranidium NI 80657, from Bed 4. Remarks. This poorly preserved specimen is inadequate for confident generic assignment, but the posteriorly situated small palpebral lobes suggest that it may best be assigned to Stenopareia. Family phillipsinellidae Whittington, 1950 Genus paraphillipsinella Lu in Lu and Chang, 1974 Type species. Paraphillipsinella globosa Lu in Lu and Chang, 1974 from the Pagoda Formation (Caradoc), Chenkou, Sichuan Province, China. Diagnosis. Phillipsinellid trilobites with glabella separated into a swollen, spherical frontal lobe and a cylindrical posterior lobe that tapers forwards slightly. Four pairs of pit-like lateral glabellar lobes on flanks of posterior glabellar lobe. Eyes small, almost spherical. Anterior border absent. Rostral ZHOU AND DEAN: ORDOVICIAN TRILOBITES 767 plate large, trapezoidal, wider than long, Pygidium faintly furrowed, without border or border furrow. Remarks. The diagnosis is based on new material in addition to well-preserved specimens described by Ji (1982) and Ju (in Qiu et al. 1983). The genus differs from Phillipsinelki Novak, 1885 mainly in the more swollen frontal glabellar lobe, the lack of an anterior border, and the smaller, spherical eyes. Protophillipsinella Chen in Li et al. 1975 is indistinguishable from Paraphillipsiuella, as noted by Xia (1978, p. 176), Lu and Zhou (1981, p. 14) and Ji (1982), and is considered a junior subjective synonym. Species of Paraphillipsiuella have been recorded from the Caradoc to early Ashgill of the Yangtze region by Lu (in Lu and Chang 1974), Chen (in Li et al. 1975, as Protophillipsinella), Zhou (in Zhou et al. 1977), Sheng (1974, as Phillipsinella), Xia (1978), Ji (1982) and Ju (in Qiu et al. 1983), but ranges of variation are not available for most of them because of limited material. La and Zhou (1981, p. 14) believed that Paraphillipsiuella included only two species: P. glohosa, with subcircular, and P. nanjiangensis Lu in Lu and Chang, 1974 with transversely oval frontal glabellar lobe. Paraphillipsiuella glohosa Lu in Lu and Chang, 1974 Plate 62, figs. 13 15 1974 Paraphillipsinella glohosa Lu in Lu and Chang, p. 133, pi. 53, figs. 8 and 9. 1974 Phillipsinella sp., Sheng, p. 77, pi. 2, figs. 14- 16. 1975 Protophillipsinella typa Chen in Li et al., p. 156, pi. 20, fig. 5a, h. 1975 Protophillip.sinella ciirvusa Chen in Li et al., p. 157, pi. 20, lig. 9a, h. 1977 Paraphillipsinella hubeiensis Zhou in Zhou et al., p. 241, pi. 73, figs. 5 and 6. 1978 Paraphillipsinella glohosa Lu; Xia, p. 175, pi. 34, figs. 9-11. non 1983 Paraphillipsinella Iniheiensis Zhou; Ju in Qiu et al., p. 228, pi. 76, figs. 8 and 9 ( = Paraphillipsi- nella nanjiangensis Lu in Lu and Chang, 1974) Holotype. Cranidium, figured Lu (in Lu and Chang 1974, pi, 53, figs. 8 and 9), from the Pagoda Formation (Caradoc), Chenkou, Sichuan Province, China. Figured speeimens. Two cranidia (NI 80658, 80659) from Bed 12. Description. Cranidium wider than long. Glabella divided into anterior and posterior lobes; anterior lobe subspherical, wider than long, three-fifths as long as cranidium and almost as wide as cranidium across base, strongly curved down anteriorly and laterally but only gently posteriorly; posterior lobe low, moderately tapered forwards, transversely convex with basal width about two-fifths that of frontal lobe. Four pairs faint, pit-like glabellar furrows situated on margins of flanks of glabella; 4p pair located between frontal and posterior lobes. Occipital ring higher than posterior lobe and slightly wider (tr.) than base of glabella; occipital furrow shallow. Axial furrows deep, broad. Fixigenae gently convex, declined laterally. Palpebral lobes small, sited opposite anterior part of posterior lobe of glabella. Posterior border Hat, widens abaxially, separated from fixigena by shallow border furrow that is broader (exsag.) in exfoliated specimens. Anterior branches of facial suture at first subparallel, turn inwards anteriorly and extend along frontal margin of glabella; posterior branches short, slightly divergent backwards. Surface of exoskeleton densely pitted. Remarks. The present specimens agree well with the holotype and those described as Phillip- sinella sp. by Sheng (1974). The holotypes of Protophillipsinella typa Chen and P. curvusa Chen from the Pagoda Formation of southern Saanxi, and of Paraphillipsinella hubeiensis Zhou from the Linhsiang Formation of Xuan, western Hubei agree with Paraphillipsinella glohosa in both morphological characters and ratio of length : width of the frontal glabellar lobe of the glabella to the posterior lobe. All three species were therefore considered synonymous by Ju (in Qiu et al. 1982, p. 59). Specimens recorded as P. hubeiensis Zhou by Ju (in Qiu et al. 1982) from the Huangnekang Forma- tion (early Ashgill) of western Zhejiang, eastern China, have the frontal glabellar lobe transversely oval in outline, a feature differing from the type species but comparable with P. nanjiangensis Lu (in Lu and Chang 1974, p. 133, pi. 53, fig. 10). 768 PALAEONTOLOGY, VOLUME 29 Family proetidae Salter, 1 864 Subfamily proetinae Salter, 1 864 Genus XENOCYBE Owens, 1973 Type species. Xetiocyhe micrommata Owens, 1973, from the Tretaspis Series, Stage 5a (Ashgill), of Holmen- skjaeret, Oslo region, Norway. Xenocybel sp. Plate 62, figs. 1 7 and 1 8 Figured specimens. Cranidium (NI 80660) and pygidium (NI 80661 ) from Bed 12. Remarks. The cranidium, imperfectly preserved, has rounded, subtriangular, inflated glabella, narrow (sag.), depressed preglabellar field, and the strongly convex anterior border is semicircular in cross section. Position of posterior end of anterior branches of facial suture close to axial furrows suggests that palpebral area of fixigenae was very narrow. The specimen is generally comparable with the holotype of Xenocybe micrommata Owens (1973, fig. 14g) but differs in its shorter glabella and fainter 2p and 3p furrows, features that in turn resemble those of the holotype of the type species of Panarchaeogonus, P. parvus Opik, 1937, refigured by Owens (1979, fig. 4 a-d). The pygidium is characterized by virtual lack of interpleural furrows and border furrow, but in its broad axis and transverse outline it is more like Pauarchaeogonus (See Owens 1979, fig. 4n) than Xenocybe (see Lane and Owens 1982, pi. 3, figs. 1 and 2). Trigonoproetus triquetrus XpoWonov (1974, p. 38, pi. 19, figs. 1-3) from the Ashgill of Kazakhstan may be conspecific with the Chinese species but its pygidium is unknown. Trigonoproetus Apollonov, 1974 was considered synonymous with Xenocybe by Lane and Owens (1982f Family aulacopleuridae Angelin, 1854 Subfamily scharyiinae Osmolska, 1957 Genus rorringtonia Whittard, 1966 Type species. Rorringtonia flabelliformis Whittard, 1966 from the Rorrington Beds, Shelve inlier, Shropshire, England. Although the Rorrington Beds extend from about middle Llandeilo to lowest Caradoc, Whittard’s (1966, table 5) record of R. flabelliformis from only the lower half of the unit suggests that the species is of Llandeilo age in its type area. Rorringtonia sp. Plate 62, fig. 19 Figured specimen. Cranidium (NI 80662) from Bed 4. Description. Cranidium gently convex, its outline slightly arched forwards frontally. Glabella subtriangular, length about two-thirds that of cranidium, broadly rounded anteriorly, as long as wide, slightly convex transversely and declined forwards; outline tapers forwards and its lateral margins are broadly curved, abaxially convex. Three pairs narrow, distinct, lateral glabellar furrows decrease in size posteriorly; Ip furrows curve adaxially, defining Ip lobes whose length is more than one third that of glabella; 2p furrows run parallel to abaxial part of Ip furrows and end opposite centre of glabella; 3p furrows run slightly forwards adaxially; suggestion of 4p furrow on left side of glabella may be result of crushing. Occipital ring narrows (exsag.) laterally; occipital furrow narrow (sag.), transversely straight medially, turns backwards distally. Axial furrows narrow, deep. Palpebral lobes reniform, slightly convex abaxially, sited close to glabella; their length is one- third that of glabella and they end posteriorly opposite abaxial ends of Ip furrows. Anterior branches of facial suture subparallel; posterior branches straight, running obliquely backwards and outwards. Preglabellar area wide (sag.), equal to one-quarter length of cranidium, slightly vaulted, with narrow (sag.), convex anterior border well defined by shallow border furrow. Posterior areas of fixigenae each have outline of right-angled triangle, with width one-quarter that of base of cranidium; narrow posterior border furrow separates posterior border that is narrow (exsag.) proximally and widens distally. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 769 Remarks. Species of Rorringtonia have been reported from the Llandeilo and low Caradoc of Britain and Norway (Whittard 1966; Owens 1973; Hughes 1979; Owens 1981). Of all these species, the present form is closest to R. kennedyi Owens (1981, p. 91, pi. 1, figs, a-d), from the Llandeilo of Llandrindod Wells, central Wales, but differs in having a longer glabella and cranidium, and narrower glabellar furrows. Madygenia snavis Petrunina (in Repina et al. 1975, p. 230, pi. 45, figs. 6 and 7, 12-1 7), from southern Tyan-Shan, type species of Madygenia, generally resembles the present cranidium. Minor differences include the more strongly declined preglabellar area and more weakly defined anterior border of the Russian species. The diagnostic characters of Madygenia fall within the range of variation in Rorringtonia and the two are probably congeneric, though a decision must await further information on the thoracic and pygidial characters of Madygenia. ?Family dimeropygidae Hupe, 1953 Genus ischyrophyma Whittington, 1963 Type species. Ischyrophyma tiiberculata Whittington, 1963, from a white limestone boulder in a conglomerate (latest Arenig-basal Llanvirn) of the Cow Head Group, Lower Head, western Newfoundland, Canada. Ischyrophyma? zhiqiangi sp. nov. Plate 62, fig. 20; Plate 63, figs. 1-7, 12 Name. After Zhou Zhiqiang, of the Xian Institute of Geology and Mineral Resources, who did much to help the authors in this project. Holotype. Cranidium, NI 80665 (PI. 63, figs. 2, 7), from Bed 12. Paralypes. Three cranidia (NI 80663, 80664, 80666) and four pygidia (NI 80667, 80668, 80669, 80670), all from Bed 12. Diagnosis. A possible Ischyrophyma species with relatively large Ip glabellar lobes. Glabella con- stricted in front of anterior ends of Ip furrows. Pygidium semicircular in plan, with well-defined conical axis and broad border. Surface finely granulate. Description. Glabella strongly convex transversely, steeply turned down frontally to overhang slightly the preglabellar furrow; glabella about as long as wide, subparallel-sided, broadly rounded anteriorly, its outline slightly constricted in front of anterior ends of Ip furrows. Two pairs distinct glabellar furrows present: Ip deep, broad, narrowing posteriorly, connected by shallow depressions to occipital furrow and axial furrows, and separating pair of subquadrate, inflated glabellar lobes that occupies more than one-quarter width and one-third to one-quarter length of glabella; 2p furrows short, directed slightly backwards adaxially, and show as smooth patches on external surface but as shallow depressions on internal mould, opposite anterior ends of palpebral lobes. Occipital ring uniformly wide (sag.) medially, where equals one-sixth glabellar length, but narrows (exsag.) sharply towards axial furrows from points opposite posterior end of Ip furrows; there is a prominent median tubercle, and posterior margin of cranidium curves gently forwards laterally. Occipital furrow deep, broad (exsag.), overhung laterally by Ip lobes. Axial furrows deeply incised. Preglabellar field absent at sagittal line. Anterior border upturned, low, almost uniformly narrow (sag.). Palpebral lobes semicircular in plan, small (about one-fifth length of cranidium), and end posteriorly opposite centre of cranidium and close to axial furrows. Anterior branches of facial suture diverge slightly forwards; posterior branches run backwards and slightly outwards. Posterior areas of fixigenae triangular, abaxially declined, their length and width nearly same as for Ip lobes. Posterior border low, narrow (exsag.). Pygidium semicircular in plan, three-fifths as long as wide. Strongly convex, conical axis rounded posteriorly and occupies two-fifths frontal width of pygidium; it comprises a large articulating half-ring, five axial rings, and a terminal piece, all separated by curved ring furrows that shallow medially. Axial furrows deep, broad. Pleural lobes gently convex with three pairs furrowed pleurae; pleural furrows shallow except those of first pair, which are relatively deep, shallow abaxially, and reach pygidial margin in some specimens. Border broad. 770 PALAEONTOLOGY, VOLUME 29 flat, well defined by shallow, wide border furrow; width (tr., sag.) of border uniform, equal to one-quarter pygidial length. Surface of cranidium and pygidium covered with dense, fine granules. Remarks. In addition to the type species, the following have been referred to Ischyrophyma: I. tiimida Whittington, 1965, from the Table Head Formation (Llanvirn) of western Newfoundland; I. marmorea Dean, 1970 from a limestone of probable Arenig age in north-eastern Newfoundland; I. deserta (Billings 1865) (see Dean 1970, pi. 1, figs. 4, 8, 12) from a boulder of early Ordovician limestone in Quebec; and probably, as noted by Dean (1970), the Swedish Arenig species Glaphurinal iusolita Tjernvik, 1956. Ischyrophymal borealis Fortey, 1980, from the upper Val- hallfonna Formation (Arenig) of Spitzbergen, was reassigned to Ischyrotoma Raymond, 1925 by Bruton (1983, p. 218). In its inflated, strongly down-curved glabella the new species resembles I. tumida Whittington (1965, p. 339, pi. 19, figs. 6-12, 15) rather than the type species 7. tiiherculata Whittington (1963, p. 48, pi. 8, figs. 1-10). /. tumida differs in: smaller lateral glabellar lobes; parallel-sided glabella, which does not narrow in front of abaxial ends of Ip furrows; narrower, ridge-like anterior border; and more scattered but coarser granules on surface of exoskeleton. The only pygidium so far assigned to Ischyrophyma is that figured as I. marmorea Dean (1970, pi. 2, fig. 2), which is well segmented, with broad axis and wide border, and is comparable with that of the present species. The cranidium of I. marmorea resembles that of /.? zhiqiangi in the following respects: outline of glabella and occipital ring; small palpebral lobes close to axial furrows; upturned, uniformly wide (sag.) anterior border; and posterior branches of facial suture, which are straight and extend obliquely backwards (see holotype in Dean 1970, pi. 1, figs. 2, 11, 13). I. marmorea is distinguished by: presence of four pairs lateral glabellar furrows; lack of an occipital tubercle; ornamentation of more scattered, coarser tubercles; longer Ip lobes; and less inflated glabella. The Newfoundland specimens are much larger than ours, and somewhat distorted, which may account for the two last differences. Unfortunately the pygidium of Ischyrophyma tuherculata has not been found but the holotype (Whittington 1963, pi. 8, figs. 1-3, 5), an enrolled specimen comprising cephalon and seven thoracic segments, has ample space for additional thoracic segments and a pygidium like that of 7. marmorea and the new species. But the question remains open and the pygidium of Ischyrophyma could alternatively be a tiny one, composed of a single segment, as claimed by Fortey (1980a, p. 68). If the latter is the case, the pygidia of 7. marmorea and 7.? zhiqiangi may be incorrectly assigned, or both species may have to be excluded from Ischyrophyma. A further possibility is that Ischyrophyma is a junior subjective synonym of Celmus Angelin, 1854, as suggested by Bruton (1983, p. 218). Fortey (1980a, p. 68) was inclined to consider 7. marmorea a glaphurid, even though the posterior branches of the facial suture are straight and run abaxially backwards. The pygidium in our collection, if correctly assigned, suggests that the new species may be related to the proetids. For the present we follow Whittington (1963) in referring Ischyrophyma questionably to the Dimeropygidae, pending a more complete knowledge of its type species. EXPLANATION OF PLATE 63 Figs. \~1,\2. 1 schyrophymal zhiqiangi ip. nov.^edM. 1, cranidium, paratype, NI 80664, x 8. 2, 7, cranidium, dorsal and right lateral views, holotype, NI 80665, x 8. 3, cranidium, paratype, NI 80666, x 10. 4, pygidium, paratype, NI 80667, x 10. 5, pygidium, paratype, NI 80668, x 8. 6, pygidium, paratype, NI 80669, X 8. 12, pygidium, paratype, NI 80670, x 10. Figs. 8, 9, 11, 13. Lonchodomas nanus Zhou in Zhou et al. 1982. Bed 4. 8, cranidium, NI 80671, x 8. 9, 1 1, cranidium, dorsal and right lateral views, NI 80672, x 8. 13, cranidium, NI 80673, x 8. Fig. 10. Bulbaspis sp. Bed 4. Immature cranidium, NI 80674, x 12. PLATE 63 ZHOU and DEAN, Ischyrophymal, Lonchodomas, Bulbaspis 772 PALAEONTOLOGY, VOLUME 29 Family raphiophoridae Angelin, 1854 Genus lonchodomas Angelin, 1 854 Type species. Ampyx rostratus Sars, 1835, from the Ampyx Limestone (late Llandeilo or early Caradoc) of Bygdoy, Oslo, Norway. Lotichodomas nanus Zhou in Zhou et al. 1982 Plate 63, figs. 8, 9, 11, 13 1982 Lonchodomas nanus Zhou in Zhou et al., p. 280, pi. 69, figs. 1 1 and 12. Diagnosis. Species of Lonchodomas with stout, weakly carinate glabella. Median glabellar spine with deeply incised median groove. Holotype. Cranidium XITr-173, figured Zhou (in Zhou et al. 1982, pi. 69, fig. 12), from the topmost Pingliang Formation (early Caradoc), Shijiezigou, Guyuan County, Ningxia, China. Figured specimens. Three cranidia, NI 80671-80673, from Bed 4. Description. Cranidium triangular in outline with length (excluding glabellar spine) less than half maximum width. Glabella weakly carinate, transversely convex, rhombic in plan, extending for up to half its length in front of fixigenae; maximum width across middle of glabella equals 0-7 to 0-85 length and about 0-65 basal width of glabella. Median glabellar spine subquadrate in section, with deeply incised median groove; three closely spaced pairs suboval muscle scars on glabellar flanks extend abaxially to axial furrows. Occipital ring convex, higher than rest of cranidium, and gently curved backwards; occipital furrow distinct, curved forwards laterally. Axial furrows broad, slightly convex laterally, efl'aced posteriorly where they are indicated by change in slope between posterior border and occipital ring; pair of small, shallow, anterior fossulae sited close to anterolateral corners of glabella. Triangular fixigenae about three-quarters as long as wide, strongly declined anteriorly, gently abaxially, with pair small, circular muscle scars located posteriorly, opposite fulcra. Posterior border wide (exsag.), flat, confluent with occipital ring, bluntly pointed backwards at fulcra sited at its mid- points, from which it narrows sharply abaxially; posterior border furrow mostly deep, narrow (exsag.), but effaced near axial furrows. Facial suture curves in broad arcs along anterior margins of fixigenae, turns sharply inwards and slightly backwards at posterolateral corners of cranidium to cross posterior border obliquely and end near fulcra. Small cranidium (PI. 63, fig. 8) has narrow (exsag.) posterior border, with fulcra placed nearer abaxial extremities of fixigenae, and fixigenae are longer, but otherwise agrees well with larger specimen. Surface covered with broadly spaced terrace lines near, and subparallel to, anterior margin of cranidium. Remarks. Lonchodomas nanus strongly resembles L. hlackstonensis Legg (1976, p. 16, pi. 6, figs. 28, 33) from the Goldwyer Formation (Llanvirn) of Blackstone, Canning Basin, Western Australia, in the outline of the weakly carinate glabella; but the latter species has a shorter (sag.) glabella with median spine that lacks a median groove and is probably rounded in cross section. The Llandeilo or early Caradoc species L. paenepennatus Ross (1970, p. 88, pi. 16, figs. 23-27; pi. 17, figs. 1 and 2), from the top part of the Antelope Valley Limestone, and L. retrolatus Ross (1967, p. D24, pi. 7, figs. 22-28), from the lower Eureka Quartzite of Nevada, USA are similar in many respects to L. nanus but have a much narrower (tr.) glabella. The surface ornamentation of the American species is unknown, and the pygidium of the Chinese species has not been found, so further comparison is impossible. The early Caradoc species L. tecturmasi (Weber 1932, p. 6, pi. 4, fig. 43; 1948, p. 18, pi. 2, figs. 20-22, 26; Chugaeva 1958, p. 32, pi. 2, figs. 3-5) from Kazakhstan resembles the two Nevada species in the outline of cranidium and glabella, but the latter is strongly carinate and thus easily distinguished from that of L. nanus. The holotype of L. nanus has a slightly narrower (tr.) glabella and better-defined occipital ring than the present material from Chedao, but is slightly crushed and much larger. Genus bulbaspis Chugaeva, 1958 Type species. Ampyx hulbifer Weber, 1932, from the Djebagly ‘horizon’ (Llandeilo) of the Djebagly Range, southern Kazakhstan, U.S.S.R. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 773 Bulhaspis sp. indet. Plate 63, fig. 10 Figured specimen. Cranidium, NI 80674, from Bed 4. Remarks. Cranidium has strongly carinate, narrow (tr.) glabella, subrhombic in outline; fixigenae wide (tr.), gently arched forwards; posterior border straight, ridge-like. All these features suggest that the specimen is possibly referable to Bulhaspis. One of the latter’s most important characters is the presence of a median glabellar bulb but, as noted by Zhang (1981, p. 202), juvenile specimens show only a median spine as in Lonchodomas. The present specimen, about 3 mm wide and 1 -6 mm long, is apparently immature and the tip of the glabella is not preserved. Species of Bulhaspis from Llanvirn to Caradoc in age have been reported from Kazakhstan (Chugaeva 1958; see also Weber 1932, 1948), Inner Mongolia, China (Lu in Lu el al. 1976), north- west China (Zhou in Zhou et al. 1982; Zhang 1981), southern Tasmania, Australia (Burrett et al. 1983) and possibly Langkawi Island, Malaysia (Kobayashi and Hamada 1978). Among them Bulhaspis ovulum (Weber 1948, p. 15, pi. 2, figs. 6-10) from the Kopalin and Karakan horizons (Llanvirn-early Llandeilo) is closely comparable with our specimen in the outline of the glabella but the fixigenae are longer; the glabella is widest opposite its midpoint in the Chinese specimen, but at a point one third of the glabellar length (excluding bulb) from its anterior end in B. ovulum. Family hammatocnemidae Kielan, 1960 Genus hammatocnemis Kielan, 1960 Type species. Hammatocnemis letrasulcatus Kielan, 1960, from the Stuurocephaliis clavifrons Zone (Ashgill), Brzezinki, Poland. Remarks. The oldest known member of the genus (Lu and Zhou 1979) is believed to be Hammatoc- nemis primitivus primitivus Lu, 1975, from the Arenig of the Yangtze region of China. On the basis of the structure of the preoccipital ring, the remaining, younger species fall into two groups; 1. species with a pair of isolated preoccipital (Ip) lobes but without an intervening median ‘ring’, found only in the Yangtze region, China; and 2. species with entire median preoccipital ring between the Ip lateral glabellar lobes, widely distributed in the Inner Mongolia region of north China, northwest China, Uzbekistan and Kazakhstan, southern U.S.S.R., and Poland. Group 1 includes: H. huayinshanensis Lu in Lu and Zhang 1974; L. yangtzeensis Lu in Lu and Zhang 1974 (= H. nanzhengensis Zhou in Li et al. 1975 = H. cf. pagodus Chen of Li et al. 1975); FI. decorosus Lu in Lu and Chang 1974 (= H. tetrasiilcatus Sheng, 1964 non Kielan 1960 = H. liangshanensis Chen in Li et al. 1975, and probably = H. sinensis Han, 1980); H. ovatus Sheng, 1964 (described first as H. tetrasulcatus ovatus — H. orientalis Chen in Li et al. 1975 = H. pagodus Chen in Li et al. 1975 = //.? tudilingensis Chen in Li et al. 1975 = H. hexianensis Q. Z. Zhang in Qiu et cd. 1983); and H. longicervix Zhou in Lu et cd. 1976. Group 2 includes: H. primitivus extraneus Lu and Zhou, 1979; H. intermedins Lu and Zhou, 1979; H. tetrasulcatus Kielan, 1960; H. glohosus Abdullaev, 1972; H. kanlingensis Zhang, 1981; and H. ohsoletus sp. nov. (described below). Hammatocnemis ohsoletus sp. nov. Plate 64, figs. 1-4, 8, 9, 1 1, 12; Plate 65, fig. 1 1 Holotype. Cranidium, NI 80675 (PI. 64, figs. 1 and 2), from Bed 4. Paratypes. Five cranidia (NI 80676, 80677, 80678, 80679, 80702) and two pygidia (NI 80680, 80681) from Bed 4. Name. Latin ohsoletus, a, um, obsolete, referring to the lack of deep 3p and 4p glabellar furrows. 774 PALAEONTOLOGY, VOLUME 29 Diagnosis. Hammatocnemis species with 3p and 4p lateral glabellar furrows faintly defined. Pre- occipital segment wide (sag.) with 1 p lateral glabellar lobes weakly defined. Glabellar surface smooth or with scattered small granules. Description. Glabella gently convex, divided into small, parallel-sided preoccipital ring and large anterior lobe that extends well forwards and is broadly rounded frontally. Preoccipital ring flat, low, its length (sag.) half that of occipital ring; it widens (exsag.) slightly abaxially to form pair of weakly convex, poorly defined, subquadrate Ip lobes. Preoccipital (or Ip transglabellar) furrow transverse in direction, deeply incised laterally, shallow medially. Anterior lobe of glabella (i.e. portion of glabella in front of Ip furrows) nearly as long as wide, widest across anterolateral angles, where 1-6- 1 -8 times width (tr.) of occipital ring. Three equidistant pairs lateral glabellar furrows decrease in length anteriorly; 2p pair narrow, deep, and run slightly forwards adaxially; 3p pair faint, parallel to 2p; 4p pair very faint, directed slightly forwards adaxially and sited opposite anterior ends of palpebral lobes. Occipital ring about three times as wide (tr. ) as long (sag.) and wider (tr.) than base of glabella; its anterior margin is curved forwards medially, where length (sag.) about 0-22 that of glabella, and straight posterior margin turns forwards near axial furrows; there is a tiny median tubercle. Occipital furrow deep, becoming slightly shallower medially. Axial furrows deep, broad. Narrow (tr.) pal- pebral lobes, well defined by deep palpebral furrows and elevated above fixigenae, located between 2p and 4p glabellar furrows. Palpebral areas of fixigenae triangular; subrectangular posterior areas decline steeply abaxially from fulcra and occupy one-third basal width and one-quarter median length of cranidium; posterior border gently convex, widens (exsag.) abaxially, bounded by deep, wide posterior border furrow. Anterior branches of facial suture run along axial furrows before turning adaxially; posterior branches curve strongly back abaxially. Pygidium two-and-a-half times as wide as long, widest posteriorly, where posterior margin curved, slightly concave. Low, triangular axis has frontal breadth about one-third that of pygidium and extends to posterior margin of pygidium; there are four transversely rectangular axial rings and rounded terminal piece in addition to articulating half-ring. Ring furrows deep, arched forwards; axial furrows distinct, narrow. Laterally declined pleural lobes comprise four pleurae that are separated by distinct interpleural furrows and turn backwards abaxially; tips of first three pleurae extend slightly beyond posterior pygidial margin, but only those of first pleura form free points. Surface of cranidium and pygidium smooth or covered with scattered granules. Remarks. The species appears to occupy an isolated position within the genus. It resembles older species such as the Llanvirn Hammatocnemis intermedins Lu and Zhou (1979, p. 427, pi. 2, fig. 9; text-fig. 6) in having the glabellar outline less constricted at the 2p furrows, so that the width there is nearly the same as that of the preoccipital (Ip) segment; and the 2p to 4p glabellar furrows are longer than those of younger species. The pygidium closely resembles that of younger species such as H. tetrasulcatus Kielan (1960, p. 141, pi. 25, fig. 3; pi. 26, figs. 2, 4; pi. 27, figs. 6-8), but the anterior three pleurae of H. ohsoletus extend further backwards, beyond the posterior pygidial margin, as seen in older species such as H. primitivus primitivus Lu (1975, p. 231, pi. 45, figs. 4-14) and H. primitivus extraneus Lu and Zhou (1979, p. 426, pi. 1, figs. 1-13; pi. 2, figs. 1-8), though the second and third pleurae do not end in free points in H. ohsoletus. EXPLANATION OF PLATE 64 Figs. I 4, 8, 9, II, 12. Hammatocnemis ohsoletus sp. nov. Bed 4. 1 and 2, cranidium, dorsal and right lateral views, hololype, NI 80675, x 8. 3, cranidium, paratype, NI 80676, x 6. 4, cranidium, paratype, NI 80677, x8. 8, cranidium, paratype, NI 80678, x 6. 9, cranidium, paratype, NI 80679, x 6. 1 1, pygidium, paratype, NI 80680, X 10. 12, pygidium, paratype, NI 80681, x 6. Figs. 5-7, 10. Hammatocnemis kanlingensis Zhang, 1981. Bed 4. 5 and 6, cranidium, right lateral and dorsal views, N I 80682, x8. 7, cranidium, NI 80683, x 5. 10, pygidium, NI 80684, x4. Figs. 13-15. Hammatocnemis A'c/Zer; (Koroleva, 1959). Bed 12. 13 and 14, cranidium, right lateral and dorsal views, N I 80685, x 10. 15, cranidium, NI 80686, x8. PLATE 64 ZHOU and DEAN, Hammatocnemis 776 PALAEONTOLOGY, VOLUME 29 Hammatocnemis kanlingensis Zhang, 1 98 1 Plate 64, figs. 5^7, 10 1979 Hammatocnemis tetrasulcatus Kielan; Lu and Zhou, p. 428, pi. 2, figs. 10, 11, non Kielan, 1960. 1981 Hammatocnemis kanlingensis Zhang, p. 209, pi. 77, figs. 5-7. Holotype. Cranidium, XTR 344, figured Zhang (1981, pi. 77, fig. 7), from the Qilang Formation of Kanling, Keping county, Xinjiang, China. Figured specimens. Two cranidia (NI 80682, 80683) and one pygidium (NI 80684), from Bed 4. Remarks. The species is generally comparable with H. tetrasulcatus Kielan, 1960 but the anterior lobe of the glabella is longer (sag.), oval in plan, and less expanded forwards. H. kanlingensis also resembles H. ovatus Sheng (1964, p. 560, pi. 2, figs. 2a-c\ Lu 1975, p. 441, pi. 45, figs. 1-3), from the Caradoc to early Ashgill of the Yangtze region, from which it differs in having the preoccipital ( = Ip) lateral glabellar lobes linked by a prominent median ring. Specimens from the Ashgill of central Asia determined as H. tetrasulcatus by Apollonov (1974, p. 62, pi. 13, figs. 1-8; pi. 14, figs. 1-6) and by Petrunina (in Repina et al. 1975, p. 219, pi. 46, figs. 1-3, 6-14) differ from the Polish species in the shape of the anterior lobe of the glabella and may be synonymous with H. kanlingensis. Genus ovalocephalus Koroleva, 1959 Type species. Ovalocephalus kelleri Koroleva, 1959, from the late Caradoc of Kazakhstan and Uzbekistan. Remarks. One of the diagnostic features of Ovalocephalus appears to be the lack of 3p and 4p lateral glabellar furrows, but judging from the present material assigned to the type species, O. kelleri, 3p and 4p are represented by small pits beside the axial furrows on the external surface or by noticeable muscle scars on the exfoliated surface, as seen also in an Ashgill specimen from Kazakhstan figured by Apollonov (1974, pi. 13, fig. 9). Both the length and the degree of development of the anterior lateral glabellar furrows vary in Hammatocnemis. According to Lu and Zhou (1979, p. 423) one of the evolutionary trends affecting the glabella is the shortening of the anterior glabellar furrows; similar effacement of the corresponding furrows can be seen in, for example, H. obsoletus sp. nov. and seems to be of no more than specific importance. Other diagnostic features of Ovalocephalus are the presence of genal spines (seen also in our material) and the sub-rounded anterior margin of the glabella. However, genal spines such as are developed in juvenile specimens of Hammatocnemis (see Kielan 1960, pi. 25, fig. 3; Lu and Zhou 1979, pi. 3, fig. 10) exist also in some mature cranidia of the genus, for example, H. longicervix Zhou (in Lu et al. 1976, p. 75, pi. 13, fig. 13), though they are shorter. The shape of the glabella in Hammatocnemis varies both between and within species. Present material indicates that the frontal glabellar lobe of Ovalocephalus is oval to rounded- rhombic in outline, and is more or less comparable with some specimens of Hammatocnemis ovatus Sheng (see Lu 1975, pi. 45, figs. 2 and 3). All the pygidia in the present collection agree with that of Hammatocnemis. It seems likely that Hammatocnemis may eventually be considered a junior subjec- tive synonym of Ovalocephalus. Ovalocephalus kelleri Koroleva, 1959 Plate 64, figs. 13-15; Plate 65, figs. 5, 13 1959 Ovalocephalus kelleri Koroleva, p. 1316, text-fig. 3. 1972 Ovalocephalus kelleri Koroleva; Abdullaev, p. 110, pi. 45, fig. 1. 1974 Ovalocephalus kelleri Koroleva; Apollonov, p. 65, pi. 1 3, fig. 9. 1975 Ovalocephalus kelleri Koroleva; Petrunina in Repina et al., p. 220, pi. 46, figs. 1 5 and 16. Holotype. Cranidium, figured Koroleva (1959, text-fig. 3u), from a limestone of late Caradoc age in northern Kazakhstan. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 111 Figured specimens. Three cranidia (NI 80685, 80686, 80699) and one pygidium (NI 80700) from Bed 12. Remarks. Cranidia in the present collection are indistinguishable from the holotype. This species is characterized by the longer anterior lobe of the glabella, the effaced 3p and 4p furrows, and the presence of a pair of prominent genal spines. The outline of the anterior glabellar lobe varies from oval (PI. 64, fig. 15) to rounded— rhombic (PI. 64, figs. 13 and 14; PI. 65, fig. 5), and the anterior glabellar margin is strongly convex forwards. Family encrinuridae Angelin, 1854 Subfamily cybelinae Holliday, 1942 Genus lyrapyge Fortey, 1980 Type species. Lyrapyge ehriosus Fortey, 1980, from the middle part of the Olenidsletta Member (middle Arenig), Valhallfonna Formation, northern Ny Friesland, Spitsbergen. Lyrapygel gaoluoensis (Zhou in Zhou el al. 1977) Plate 65, figs. 1, 2, 9, 14 1975 Atractopyge sp. Lu, p. 445, pi. 46, fig. 1. 1977 Atractopyge gaoluoensis Zhou in Zhou et al., p. 260, pi. 79, figs. \a, h, and 2. 1978 Atractopyge gaoluoensis Zhou; Xia, p. 182, pi. 36, figs. 13-15. 1981 Atractopyge gaoluoensis Zhou; Lu and Zhou, p. 19, pi. 3, fig. 9. 1983 Atractopyge .xiangnanensis Q. Z. Zhang in Qiu et al., p. 242, pi. 83, fig. 3. 71983 Atractopyge gaoluoensis Zhou; Q. Z. Zhang in Qiu et al., p. 242, pi. 83, fig. 4. Holotype. Dorsal shield, II IV 70198, figured Zhou in Zhou et al. (1977, pi. 79, fig. 2), from the Linhsiang Formation (early Ashgill), Gaoluo, Xuan County, Hupei Province, China. Figured specimens. Two cranidia (NI 80691, 80692) and two pygidia (NI 80693, 80694) from Bed 12. Description. Glabella transversely convex, subparallel-sided posteriorly, expands strongly from adaxial extremities of palpebral ridges into a wide frontal lobe that is highest posteriorly and declines gently forwards. Maximum glabellar width (between anterolateral corners of glabella) is about twice the basal glabellar width and equal to the glabellar length. Three pairs short, deep, pit-like lateral glabellar furrows are equispaced on flanks of posterior lobe of glabella. 3p furrows bifurcate adaxially and shallow abaxially, their faint anterior branches located beside distal ends of eye ridges. Distinct, deep occipital furrow curves forwards abaxially. Occipital ring lenticular in plan, its length (sag.) one seventh that of cranidium. Axial furrows deep posteriorly, slightly shallower anteriorly, with pair of anterior pits sited in line with centre of frontal glabellar lobe. Narrow (sag.), upturned anterior border narrows laterally and is well defined medially by anterior border furrow. Conspicuous palpebral ridges run obliquely backwards abaxially. Anterior branches of facial suture converge forwards and turn inwards. Pygidium subtriangular in plan, about 1-4 times as wide as long. Strongly convex, conical axis occupies about one-quarter anterior breadth of pygidium and comprises five well-defined axial rings and a terminal piece; ring furrows deep laterally, shallow medially. Axial furrow deep, broad. Pleural regions consist of pair anterior half ribs, three pairs ribs, and a triangular post-axial piece. Ribs of first pair are transverse adaxially but curve backwards and slightly inwards abaxially; remaining ribs curve backwards and inwards; pleural furrow short (tr.), about half length (tr.) of pleurae; deep, long interpleural furrows separate ribs into two bands, of which the anterior bands are highly convex, widen to become spatulate posteriorly, and terminate in free points. Surface finely granular except for fixigenae, which are densely pitted. Remarks. The species resembles Lyrapyge ehriosus Fortey (1980a, p. 100, pi. 23, figs. 10-14; pi. 24, figs. 1-9) rather than Atractopyge verrucosa (Dalman), type species of Atrctctopyge, the holotype of which, from the Ashgill of south Wales, was re-figured by Dean (1974, p. 97, text-fig. 4a, h). L.l gaoluoensis differs from L. ehriosus as follows: absence of anteromedian, longitudinal glabellar furrow; absence of preglabellar furrow, though its abaxial extremities are indicated by the anterior pits; much narrower (tr.) posterior lobe of glabella; much shorter lateral glabellar furrows; three instead of four pairs pleural ribs; pygidial spines spatulate. 778 PALAEONTOLOGY, VOLUME 29 Family lichidae Hawle and Corda, 1847 Subfamily lichinae Hawle and Corda, 1847 Genus Lidias Dalman, 1827 Type species. Entomostracites laciniatus Wahlenberg, 1818, from the Dalmanitina Beds (Ashgill) of Bestorp, Mosseberg, Sweden. Lidias (Wahlenberg, 1818) Plate 65, figs. 15-18 Figured specimens. Two cranidia (NI 80688, 80690), one hypostoma (NI 80689) and one pygidiinn (NI 80687), all from Bed 12. Remarks. Cranidium has eentral glabellar lobe expanded gradually forwards, its neck narrower than the mean width of, and set lower than, the lateral glabellar lobes. According to Warburg (1925, p. 306) these features suggest L. laciniatus rather than L. ajfiiiis Angelin, 1854. L. laciniatus has been recorded from the Ashgill to lowest Llandovery of Sweden (Warburg 1925, 1939), northern England (Reed 1896, as Lidias conformis var. keisleyensis Reed; Temple 1969) and eastern Ireland (Dean 1974). Specimens from different areas exhibit slightly different surface sculpture. Surface of the present cranidium, covered with densely grouped, small, low granules of different sizes, is comparable with a specimen figured by Temple (1969, pi. 2, figs. 1-9) but differs from those figured by Warburg (1939, pi. 9, fig. 3«, b) and Dean (1974, pi. 34, figs. 2, 3, 9), which show larger but sparsely distributed granules. The Chinese hypostoma resembles that of a Silurian specimen (Temple 1969, pi. 3, figs. 1-3) in the branched middle furrow, but the posterior branches turn adaxially behind the maculae and meet in a distinct transverse furrow. The associated small pygidium from China is shorter (sag.), more broadly rounded posteriorly, with an additional pair of tiny, closely spaced posterior spines but otherwise agrees well with the holotype (Wahlenberg 1818, pi. 2, fig. 2*; Warburg 1925, text-fig. 20; 1939, pi. 9, fig. 1; Temple 1969, pi. 3, fig. 5). However, the specimen, only 2-3 mm wide and 1-4 mm long, is much smaller than any previously recorded pygidium of L. laciniatus. Family isocolidae Angelin, 1854 Genus cyphoniscus Salter, 1853 Type species. Cyphoniscus socialis Salter, 1853, from the Chair of Kildare Limestone (Ashgill) of Kildare, eastern Ireland. Cyphoniscus cf. socialis Salter, 1853 Plate 65, figs. 8, 12 Figured specimen. Cranidium, NI 80701, from Bed 4. EXPLANATION OF PLATE 65 Figs. 1, 2, 9, 14. Lyrapygel gaoluensis (Zhou in Zhou et al. 1977). Bed 12. I, cranidium, NI 80691, x 5. 2, cranidium, NI 80692, x 8. 9, pygidium, NI 80693, x 6. 14, pygidium, NI 80694, x 10. Figs. 3, 4, 6, 7, 10. Paratiresias turkestanicus Petrunina in Repina et al. 1975. Bed 12. 3, cranidium, NI 80695, X 8. 4, cranidium, NI 80696, x 6. 6 and 7, cranidium, dorsal and right lateral views, NI 80697, x 6. 10, cranidium, NI 80698, x 6. Figs. 5, 13. Ovalocephalus kelleri Koroleva, 1959. Bed 12. 5, cranidium, NI 80699, x 5. 13, pygidium, NI 80700, X 10. Figs. 8, 12. Cvphoniscus cf. socialis Salter, 1853. Bed 4. Cranidium, right lateral and dorsal views, NI 80701, X 10. Fig. 1 1. Hammatocnemis ohsoletus sp. nov. Bed 4. Small cranidium, paratype, NI 80702, x 12. Figs. 15-18. Lidias alT. laciniatus (Wahlenberg, 1818). Bed 12. 15, juvenile pygidium, NI 80687, x 12. 16, cranidium, NI 80688, x 5. 17, hypostoma, NI 80689, x 12. 18, cranidium, NI 80690, x 6. PLATE 65 ZHOU and DEAN, Ordovician trilobites 780 PALAEONTOLOGY, VOLUME 29 Remarks. The single cranidium has a proportionately longer glabella and the palpebral lobes are situated slightly further forwards, but otherwise agrees well with the lectotype and other cranidia from the Chair of Kildare Limestone (Salter 1853, pi. 9, figs. 1, 3, 4, 6, 7; Whittington 1956, pi. 130, figs. 1-9, 11; Dean 1971, pi. 20, figs. 11, 13; pi. 21, figs. 1, 4-6). C. socialis has also been recorded from the Ashgill of Uzbekistan (Abdullaev 1972), Kazakhstan (Apollonov 1974), and southern Tyan Shan (Petrunina in Repina et al. 1975). Cranidia figured by both Petrunina (in Repina et al. 1975, pi. 48, figs. 17-20) and Apollonov (1974, pi. 17, figs. 4, 8) show the glabella more elongated than that of the lectotype, and in this respect agree with our specimen. Other cranidia figured by Apollonov (1974, pi. 17, figs. 6, 7, 9) compare closely with those from the type locality and may indicate intraspecific variation in the ratio of glabellar length : width. If so, the present specimen would be referable to C. socialis sensu stricto, but for the time being it is left under open nomencla- ture. Genus paratiresias Petrunina in Repina et al. 1975 Type species. Paratiresias turkestauicus Petrunina in Repina et al. 1975, from the Kielanella-Tretaspis Beds of Ulugtay District, Turkestan and Alai Ridges (southern Tyan Shan), Uzbek SSR. Remarks. The diagnosis given by Petrunina (in Repina et al. 1975, p. 227) is as follows (free translation from the Russian): ‘Cranidium elongated-trapezoid in outline with entire anterior border arched forwards. Glabella evenly convex, expanding forwards. Axial furrows distinct. Preglabellar field absent. Anterior border quite wide, convex, evenly curved. Anterior border shallow. Fixigena narrow, strongly sloping downwards. Palpebral lobe small, forwardly located. Posterior border strongly widening abaxially. Anterior branches of facial sutures weakly divergent. Surface covered with fine lines.’ Tiresias M‘Coy, 1846 and Holdenia Cooper, 1953 are considered synonymous (Dean 1962, p. 342), but as the former is preoccupied by Tiresias Stephens, 1833 (see Sherborn 1931, p. 6528) Holdenia should be used instead. Characteristic features of Paratiresias that differ from those of Holdenia include: glabella less expanded forwards; narrower fixigenae; presence of well-defined anterior border; anterior branches of facial suture divergent. Cranidia described below are essentially indistinguishable from the holotype of the type species but are better preserved and provide additional detail. The anterior border furrow is, in fact, deep and broad as shown in the original material (Petrunina in Repina et al. 1975, pi. 47, figs. 1, 4-7). As in Holdenia there is a median occipital tubercle and three pairs of smooth patches on the flanks of the glabella represent glabellar furrows. The Bertillon pattern ornamentation of the exoskeletal surface is comparable with that of Holdenia (see Dean 1962, pi. 49, figs. 1-8), but the raised lines are more densely grouped and there are no granules on the intervening spaces. Paratiresias turkestanicus Petrunina in Repina et al. 1975 Plate 65, figs. 3, 4, 6, 7, 10 1975 Paratiresias turkestanicus Petrunina in Repina et al. p. Ill, pi. 47, figs. 1, 4-7. Figured specimens. Four cranidia (NI 80695-80698) from Bed 12. Description. Cranidium trapezoidal in plan, longer than wide, broadly rounded anteriorly. Glabella strongly convex transversely, gently rounded in profile; its outline is suboval, transversely straight posteriorly, broadly rounded anteriorly, gently expanded forwards, about four-fifths as long as wide with maximum width developed at one third length of glabella from its anterior margin. Three pairs lateral glabellar furrows represented by smooth, suboval patches on flanks of glabella; 3p pair directed forwards and sited opposite palpebral lobes; 2p pair transverse, opposite centre of glabella; Ip pair subparallel to 2p pair and widen (exsag.) abaxially. Occipital ring carries a tiny median tubercle, has sagittal length about one-sixth that of glabella, and shortens (exsag.) and curves forwards abaxially. Occipital furrow deeply incised. Axial furrows deep and wide, and become still wider posteriorly. Fixigenae steeply declined, sub-triangular in plan, basal width one- fifth that of cranidium and three times anterior width. Palpebral lobes small, very narrow, raised, abaxially curved, their length (exsag.) equal to that of anterior border; anterior ends of palpebral lobes just reach ZHOU AND DEAN: ORDOVICIAN TRILOBITES 781 anterior border furrow. Anterior branches of facial suture very short; posterior branches long, straight, extending across posterior part of lateral border to cut posterior border distally. Anterior border convex, semicircular in cross-section, slightly narrower (sag.) medially, and well dehned by deep, broad (sag.) anterior border furrow. Posterior border widens (exsag.) laterally. Surface covered with scattered granules; anterior border, glabella and occipital ring carry Bertillon pattern of anastomosing raised lines subparallel to glabellar margins. Similar anastomosing raised lines are more densely grouped on the fixigenae, subparallel to the lateral margins of the cranidium. Acknowledgements. We wish to express our deep indebtedness to Zhou Zhiqiang, Xi’an Institute of Geology and Mineral Resources, and to Liu Pingjun and Fei Anqi, Changqing Bureau of Petroleum Prospecting, for providing most of the material on which this paper is based; the columnar section of the Chedao Formation used here (text-fig. 2) is based on their unpublished data. The work was carried out in the Department of Geology, University College, Cardiff, during a visit by Zhou Zhiyi sponsored by the Royal Society and the Academia Sinica. We thank Zhou Zhiqiang, R. M. Owens, S. F. Morris and R. A. Fortey for helpful discussions, and J. Harris, Hu Shangqing, Wen Meijing, and Yang Ronqing for technical assistance. REFERENCES ABDULLAEV, R. N. 1972. Trilobites of the Upper Ordovician of Bukantan. 103-126. In masymov, a. s. and ABDULLAEV, R. N. (eds.). New data on the fauna of the Palaeozoic and Mesozoic of Uzbekistan. Akad. Nauk. Uzb. SSR Inst. Geol. Geophys., FAN, Tashkent, 1-142. [In Russian.] ANGELIN, N. p. 1851. Palaeontologia Suecica I. Iconographia Crustaceonon formationis transitionis, Fasc I, 1-24. Samson and Wallin, Lund. 1854. Palaeontologia Scandinavica I. Crustacea Formationis transitionis, Fasc 2, 21-92. Samson and Wallin, Lund. APOLLONOV, M. K. 1974. AsligHI trUobites of Kazakhstan. Izd. ‘Nauka’ Kazakhskoi SSR, Alma-Ata, 1-136. [In Russian.] BARRANDE, J. 1846. Notice preliminaire sur le Systane sihtrien et les Trilobites de Boheme, vi +97, pp. Leipsic. 1852. Systeme silurien du centre de la Boheme. /ere partie. Recherches paleontologiques. Vol. 1. Crustaces, Trilobites, xxx + 935 pp. Prague and Paris. BILLINGS, E. 1861-1865. Palaeozoic Fossils. Volume 1. Containing descriptions and figures of new or little known species of organic remains from the Silurian rocks, 426 pp. Montreal. BROGGER, w. c. 1882. Die silurischen Etagen 2 und 3 im Kristiania-gebiet und auf Eker, ihre Gliederung, Fossilien, Schichten-storungen und Contactmetamorfosen. Univ.-Programm (Christiania), 1-376. BRUTON, D. L. 1976. The trilobite genus Phillipsinella from the Ordovician of Scandinavia and Great Britain. Palaeontology 19, 699-18. 1983. The morphology of Celmus (Trilobita) and its classification, 213-219. In briggs, d. e. g. and lane, p. D. (eds.). Trilobites and other early arthropods. Spec. Pap. Palaeont. 30, 276 pp. BURMEISTER, H. 1843. Die Organisation der Trilobiten. 147 pp. Berlin. BURRETT, c., STAIT, B. and LAURIE, J. 1983. Trilobites and microfossils from the Middle Ordovician of Surprise Bay, southern Tasmania, Australia. Mem. Ass. Australas. Palaeontols. 1, 177-193. BUTTS, c. 1926. The Paleozoic rocks. Geology of Alabama. Spec. Rep. geol. Surv. Ala. 14, 41-230. CALLAWAY, c. 1877. On a new area of Upper Cambrian rocks in South Shropshire, with a description of a new fauna. Q. Jl. geol. Soc. Fond. 33, 652-672. CHANG RiDONG. 1962. Several nautiloid species from the Middle Ordovician of Huanxian, Giansu. Acta palaeont. sin. 10 (4), 514-523. [In Chinese with Russian summary.] CHANG WENTANG. 1962. Ordovician of China. Science Press, Beijing, 161 pp. [In Chinese.] CHANG WENTANG and FAN, c. s. 1960. Class Trilobita of the Ordovician and Silurian periods of the Chilian Mountains. In Geological Gazetteer of the Chilian Mountains 4 (1), 83-148. Science Press, Beijing. [In Chinese.] CHATTERTON, B. D. E. 1980. Ontogcnetic studies of Middle Ordovician trilobites from the Esbataottine Forma- tion, Mackenzie Mountains, Canada. Palaeontographica 171, 1-74. and LUDViGSEN, R. 1976. Silicified Middle Ordovician trilobites from the South Nahanni River area. District of Mackenzie, Canada. Palaeontographica A154, 1-106. CHUGAEVA, M. N. 1958. The Ordovician of Kazakhstan. Ill: The Ordovician trilobites of the Chu-Ili Mountains. Trudy geol. Inst. Acad. Nauk S.S.S.R. 9, 5-138. [In Russian.] 782 PALAEONTOLOGY, VOLUME 29 COOPHR, B. N. 1953. Trilobites from the Lower Champlainian formations of the Appalachian Valley. Mem. geol. Soc. Am. 55, 1-69. COOPER, G. A. and kindle, c. h. 1 936. New brachiopods and trilobites from the Upper Ordovician of Perce, Quebec. J. Paleont. 10, 348-372. DALMAN, J. w. 1827. Om Palaeaderna, eller de sa kallade Trilobiterna. K. Svenska Vetensk Akad. Hcmdl. 1, 113-152, 226-294. DEAN, w. T. 1962. The Ordovician trilobite genus Tiresias M‘Coy, 1846. Palaeontology, 5, 340-343. 1966. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the neighbouring district of the Montagne Noire, south-western France. Bull. Br. Mas. nat. Hist. (Geol.), 12, 247-353. 1 970. Lower Ordovician trilobites from the vicinity of South Catcher Pond, northeastern Newfoundland. Geol. Surv. Pap. Can. 70-44, 1-11. 1971. The trilobites of the Chair of Kildare Limestone (Upper Ordovician) of eastern Ireland. I. Palaeon- togr. Soc. [Monogr.] 1-60. 1972. The isocolid trilobites Cyphoniscus Salter 1853 and Effnaspis gen. nov. in the Appalachian region of Canada and U.S.A. Can. J. Earth Sci. 9, 415-421. 1973. Ordovician trilobites from the Keele Range, northwestern Yukon Territory. Bull. geol. Surv. Can. 223, 1-43. 1974. The trilobites of the Chair of Kildare Limestone (Upper Ordovician) of eastern Ireland. 2. Palaeon- togr. Soc. [Monogr.], 61-98. 1978. The trilobites of the Chair of Kildare Limestone (Upper Ordovician) of eastern Ireland. 3. Palaeon- togr. Soc. [Monogr.], 99-129. ELLES, G. L. and WOOD, E. M. T. 1904. A monograph of British Graptolites. 4. Palaeontogr. Soc. [Monogr.], liii- Ixxii, 135-180. FORTEY, R. A. 1975u. The Ordovician trilobites of Spitsbergen. II. Asaphidae, Nileidae, Raphiophoridae and Telephinidae of the Valhallfonna Formation. Norsk Polarinst. Skr. 163, 1-207. 19756. Early Ordovician trilobite communities. Eossils and Strata, 4, 339-360. 1980u. The Ordovician trilobites of Spitsbergen. III. Remaining trilobites of the Valhallfonna Formation. Norsk Polarinst. Skr. 171, 1-163. 19806. Generic longevity in Lower Ordovician trilobites: relation to environment. Paleohiologv 6, 24-31. HAN NAIREN. 1980. On the exoskeleton of Hammatocnemis. Geol. Rev. 26, 183-189. [In Chinese.] HAWLE, J. and CORDA, A. J. c. 1847. Prodrom einer Monographie der bohmischen Trilobiten. 176 pp. J. G. Calve, Prague, HOLLIDAY, s. 1942. Ordovician trilobites from Nevada. J. Paleont. 16, 471-478. HOLM, G. 1882. De svenske arten af Trilobitslagtet Illaenus (Dalman). K. Svenska Vetensk Akad. Handl. 7, 1-148. 1886. Illaeniden. Revision der Ostbaltischen silurischen Trilobiten von Fr. Schmidt. Abt. III. Mem. Acad. Sci. St. Petersb. 33, 1-173. HOWELL, B. F. 1935. Cambrian and Ordovician trilobites from Herault, southern France. J. Paleont. 9, 222- 238. HUANG, T. K., JEN CHiSHUN, JIANG CHUNFA, CHANG CHiMENG and XU ZHiQiN. 1977. An Outline of the tectonic characteristics of China. Acta geol. sin., 1972, 1 17-135. [In Chinese with English abstract.] HUGHES, c. p. 1979. The Ordovician trilobite faunas of the Builth-Llandrindod inlier, central Wales. Bull. Br. Mus. nat. Hist. (Geol.), 32, 109-181. HUPE, p. 1953. Classification des trilobites. Annls Paleont. 39, 61-168. INGHAM, J. K. 1970. The Upper Ordovician trilobites from the Cautley and Dent districts of Westmorland and Yorkshire. 1. Palaeontogr. Soc. [Monogr.], 1-58. JAEKEL, o. 1909. Ueber die Agnostiden. Z. dt. geol. Ges. 61, 380-401. JEN CHISHUN, JING CHUNFA, ZHANG ZHENGKUN and QIN DEYU. 1 980. The geotectonic evolution of China. 1 24 pp. Geological Publishing House, Beijing. [In Chinese.] Ji ZAILIANG. 1982. Stratigraphical distribution and range of the trilobite family Phillipsinellidae in South China. Bulk Inst. geol. Chinese Acad. Sci. 6, 57-68. [In Chinese with English abstract.] KiELAN, z. 1960, Upper Ordovician trilobites from Poland and some related forms from Bohemia and Scandi- navia. Palaeont. Pol. 11, 1-198. KOBAYASHi, T. 1935. The Cambro-Ordovician formations and faunas of South Chosen. Palaeont., Pt. 3. Cambrian faunas of South Chosen with special study on the Cambrian trilobite genera and families. J. Fac. Sci. Imp. Univ. Tokyo, 4, 49-344. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 783 and HAMADA, T. 1978. Upper Ordovician trilobites from the Langkawi Island, Malaysia. Geol. Palaeont. SEAsia. 19, U27. KOROLEVA, M. N. 1959. Ncw spccics of trilobites from the Middle and Upper Ordovician of northern Kazakh- stan. Dokl. Akad. Nauk SSSR 124, 1313-1316. [In Russian.] 1964. New Middle Ordovician trilobites Shumardia from northern Kazakhstan. Paleont. Zh. 1964 (1), 71-75. [In Russian.] LAI CAIGEN, WANG XIAOFENG, YI YONGEN, AN TAIXIANG, ZHANG WENTANG, QIU HONGRONG, CHEN TINGEN, EU KUN, LIN BAOYU, LIU DiYONG, and YANG JiNGZHi. 1982. The Ordovician System of China. Stratigraphy of China. No. 5. 297 pp. Geol. Publishing House, Beijing. [In Chinese.] LAKE, p. 1907. A monograph of the British Cambrian trilobites. 2. Palaeontogr. Soc. [Monogr.], 29-48. LANE, p. D. and OWENS, R. M. 1982. Silurian trilobites from Kap Schuchert, Washington Land, western North Greenland. Rapp. Gronland geol. Unders. 108,41-69 and THOMAS, a. t. 1983. A review of the trilobite Suborder Scutelluina, 141 -160. In briggs, d. e. g. and LANE, p. D. (eds.) Trilobites and other early arthropods. Spec. Pap. Palaeont. 30, 276 pp. legg, d. p. 1976. Ordovician trilobites and graptolites from the Canning Basin, Western Australia. Geologica et Palaeontologica 10, 1-58. LI YAOXi, SONG LiSHENG, ZHOU ZHIQIANG and YANG JiNGYAO 1975. Stmtigrapkical Gazetteer of Lower Palaeozoic, western Dahashan. 372 pp. Geological Publishing House, Beijing. [In Chinese.] LINNARSSON, J. G. o. 1869. Om Vestergotlands Cambriska och Siluriska Aflagringar. K. Svenska Vetensk- Akad.Handl. 8(2), 1-89. LU YANHAO. 1959. Subdivision and correlation of the Ordovician rocks of South China. Proceedings of Special Summaries of Geological Fundamental data of China 2, 1113. Geological Press, Beijing. [In Chinese with English summary.] 1962. Middle Ordovician index trilobites. 52-53. In wang yu (editor): Handbook of the Index Fossils of the Yangtze region. Science Press, Beijing. [In Chinese.] 1975. Ordovician trilobite faunas of central and southwestern China. Palaeont. sin. 11,1 -484. and CHANG WENTANG. 1974. Ordovician trilobites. 124-136. In A handbook of Stratigraphy and Palaeon- tology in southwest China. Science Press, Beijing. [In Chinese.] CHU CHAOLiNG, CHiEN YIYUAN and HSIANG LEEWEN. 1965, Trilobites of China. 766 pp. Science Press, Beijing [In Chinese.] — ZHOU ZHiYi, CHEN JUNYUAN, YU WEN, CHEN XU and XU HANKUi. 1976. Ordovician biostrati- graphy and palaeogeography of China. Mem. Nanjing Inst. Geol. and Palaeont. 7, 1 83 [In Chinese.] and ZHOU zhiyi. 1979. Systematic position and phylogeny of Hanvnatocneniis (Trilobita). Acta Palaeont. sin. 18, 415-434. [In Chinese with English summary.] 1981. Early Upper Ordovician trilobites from the Nanjing Hills. Bull. Nanjing Inst. Geol. Palaeont., Acad. sin. 3, 1 -27. [In Chinese with English abstract ] MAREK, L. 1952. Contribution to the stratigraphy and fauna of the uppermost part of the Kraluv Dvur Shales (Ashgillian). Sb. Ustred. Ust. Geol. Paleont. 19, 429-455. m‘coy, f. 1846. A synopsis of the Silurian fossils of Ireland. 72 pp. McGlashand and Gill, Dublin. MiKULic, u. G. 1980. Trilobites in Paleozoic carbonate buildups. Lethaia, 14, 45-56. MOBERG, J. c. and segerberg, c. o. 1906. Bidrag till kannedomen om Ceratopyge-VQgiowQn med Sarskild hansyn till dess utrecking i Fogelsangstrakten. Lunds Univ. Arsskr. ne 2 (7), 1-113. MOORE, R. c. (ed.) 1959. Treatise on Invertebrate Paleontology, Pt.O, Arthropoda I. xix -I- 560 pp. The Geological Society of America and the University of Kansas Press, Lawrence, Kansas. NiKOLAiSEN, F. 1963. The Middle Ordovician of the Oslo region, Norway. 14. The trilobite family Telephinidae. Norsk geol. Tidsskr. 43, 345-399. NOVAK, o. 1883. Zur Kenntnis der bohmischen Trilobiten. Beitr. Paldont. Geol. ost.-Ung. 3, 23-63. 1885. Studien an Hypostomen bohmischen Trilobiten. III. Sbir. K. bbhm. Ges. Wiss. Math.-nat. Kl., Jg 1884, 581-587. OPiK, A. 1937. Trilobiten aus Estland. Acta Comment. Univ. tartu. (A). 32 (3), 1 163. 1967. The Mindyallan Eauna of north-western Queensland. Bull. Bur. Miner. Resour. Geol. Geophvs. di«7. 74, 1-404. OSMOLSKA, H. 1957. Trilobites from the Couvinian of Wydryszow (Holy Cross Mountains, Poland). Acta palaeont. pol. 2, 53-77. OWEN, A. and bruton, d. l. 1980. Late Caradoc-early Ashgill trilobites of the central Oslo region, Norway. Paleont. Contr. Univ. Oslo 245, 63 pp. OWENS, R. M. 1973. Ordovician Proetidae (Trilobita) from Scandinavia. Norsk geol. Tidsskr. 53, 117-181. 784 PALAEONTOLOGY, VOLUME 29 OWENS, R. M. 1979. The trilobite genera Panarchaeogonus Opik, Ishergia Warburg and Cyamops gen. nov. from the Ordovician of Balto-Scandia and the British Isles. Norsk geol. Tidsskr. 58, 199-219. 1981. The Ordovician proetacean trilobite Rorringtoiiia. Geol. Mag. 118, 89-94. PEK, I. 1977. Agnostid trilobites of the central Bohemian Ordovician. Sb. geol. Ved. paleont. 19, 7-44. PORTLOCK, J. E. 1 843. Report on the geology of the county of Londonderry, and of parts of Tyrone and Fermanagh. xxi + 784 pp. Dublin (Milliken, Hodges and Smith) and London (Longman, Brown, Green, and Long- mans). PRICE, D. 1974. Trilobites from the Sholeshook Limestone (Ashgill) of South Wales. Palaeontology, 17, 841- 868. QIU HONGAN, LU YANHAO, ZHU ZHAOLING, BI DECHANG, LIN TIANRUl, ZHOU ZHIYl, ZHANG QUANZHONG, QIAN YIYUAN, JU TIANYIN, HAN NAIREN, and WEI xiuzHE, 1983. Tiilobita. In Palaeontological Atlas of east China. I. 28-254. Geological Publishing House, Beijing. [In Chinese.] RAYMOND, p. E. 1925. Some trilobites of the lower Middle Ordovician of eastern North America. Bull. Mus. Comp. Zool. Harv. 67, 1 1 80. REED, F. R. c. 1896. The fauna of the Keisley Limestone. Part 1. Q. Jl. geol. Soc. Lond. 52, 407-437. 1912. Ordovician and Silurian fossils from the central Himalayas. Mem. geol. Surv. India Palaeont. indica (15)7(2), I 168. 1915. Supplementary memoir on new Ordovician and Silurian fossils from the Northern Shan States. Ibid., ns6(1) 1-123. 1917. Ordovician and Silurian fossils from Yun-nan. Ibid., 6 (3), 1-84. REPINA, L. N., PETRUNINA, Z. E. and HAJRULLINA, T. I. 1975. Trilobita. 100-351. In REPINA, L. N., YASKOVICH, B. V., AKSARINA, N. A., PETRUNINA, Z. E., PONIKLENKO, I. A., RUBANOV, D. A., BOLGOVA, G. V., GOLIKOV, A. N., HAJRULLINA, T. I. and POSOKHOVA, M. M. Stratigraphy and fauna of the Lower Palaeozoic of the southern submontane belt of Turkestan and the Alai ridges (southern Tien-Shan). Trudy Inst. Geol. Geophys. Sib. Otd. 278, 352 pp. [In Russian.] ROSS, R. J. 1951. Stratigraphy of the Garden City Formation in northeastern Utah, and its trilobite faunas. Bull. Peabody Mus. nat. Hist. 6, vi+ 161 pp. 1967. Some Middle Ordovician brachiopods and trilobites from the Basin Ranges, western United States. Prof. Pap. U.S. geol. Surv. 523-D, D1-D43. 1970. Ordovician brachiopods, trilobites and stratigraphy in eastern and central Nevada. Prof. Pap. U.S. geol. Surv. 639, 1-103. 1972. Fossils from the Ordovician bioherm at Meiklejohn Peak, Nevada. Ibid., 685, 1-47. and INGHAM, J. K. 1970. Distribution of the Toquima-Table Head (Middle Ordovician Whiterock) Faunal Realm in the northern hemisphere. Bull. Geol. Soc. Amer. 81, 393-408. SALTER, J. w. 1848, In PHILLIPS, J. and salter, j. w. ‘Palaeontological Appendix to Professor John Phillips’ memoir on the Malvern Hills, compared with the Palaeozoic Districts of Abberley, &c.’ Mem. geol. Surv. Gr. Br. 2, 331-386. 1853. Figures and descriptions illustrative of British organic remains. Mem. geol. Surv. U.K., Decade 7, 78 pp. — 1864. A monograph of the British trilobites from the Cambrian, Silurian and Devonian formations. 1. Palaeontogr. Soc. [Monogr.] 1-80. SARS, M. 1835. Ueber einige neue oder unvollstandig bekannte Trilobiten. Isis, Jena, Jahrg. 1835, 4. SCHRANK, E. 1972. Nileus-Nxt&n (Trilobita) aus Geschieben des Tremadoc bis tieferen Caradoc. Ber. dt. Ges. geol. fViss., A, Geol. Paldont. 17, 351-375. SHAW, F. c. 1968. Early Middle Ordovician Chazy trilobites of New York. Mem. N.Y. St. Mus. Sci. Service, 17, 1-163. and ORMiSTON, a. r. 1964. The eye socle of trilobites. J. Paleont. 38, 1001-1002. SHENG, s. F. 1934. Lower Ordovician trilobite faunas of Chekiang. Palaeont. sin. (B). 3 (1), 1-19. 1964. Upper Ordovician trilobite faunas of Szechuan-Kweichow with special discussion on the classifica- tion and boundaries of the Upper Ordovician. Acta palaeont. sin. 12, 553-563. [In Chinese with English summary.] 1974. Ordovician trilobites from western Yunnan and its stratigraphical significance. In Subdivision and correlation of the Ordovician System in China. T-153 Geological Publishing House, Beijing. 1-153. [In Chinese.] SHERBORN, c. D. 1931. Index AnimaHum 1801-1850. Part 26, Index T-Trichoscelia. 6359-6582. British Museum, London. ZHOU AND DEAN: ORDOVICIAN TRILOBITES 785 SKJESETH, s. 1955. The Middle Ordovician of the Oslo region, Norway. 5. The trilobite family Styginidae. Norsk geol. Tidsskr. 35, 9-28. TEMPLE, j. T. 1969. Lower Llandovery (Silurian) trilobites from Keisley, Westmorland. Bull. Br. Mus. nut. Hist. (Geol.), 18, 197-230. THOMAS, A. T. and OWENS, R. M. 1978. A review of the trilobite family Aulacopleuridae. Palaeonlologv, 21, 65- 81. TJERNVIK, T. E. 1956. On the early Ordovician of Sweden. Stratigraphy and fauna. Bull. geol. Inst. Univ. Uppsala, 36, 107-284. TRIPP, R. p. 1965. Trilobites from the Albany Division (Ordovician) of the Girvan district, Ayrshire. Palaeon- tology 8, 577-603. 1976. Trilobites from the basal superstes mudstones (Ordovician) at Aldons Quarry, near Girvan, Ayr- shire. Trans. R. Soc. Edinh. 69, 369-423. TROMELiN, G. DE and LEBESCONTE, p. 1876. Essai d’un catalogue raisonne des fossiles siluriens des departements de Maine-et-Loire, de la Loire-Inferieure et du Morbihan, avec des observations sur les terrains paleozoiques de rOuest de la France. C. r. Ass. fr. Avanc. ScL, session, Nantes (1875), 606-661. ULRICH, E. o. 1930. Ordovician trilobites of the family Telephidae and concerned stratigraphic correlations. Proc. U.S. natl. Mus. 76(21), I 101. voGDES, A. w. 1890. A bibliography of Palaeozoic Crustacea from 1698 to 1889, including a list of North American species and a systematic arrangement of genera. Bull. U.S. geol. Surv. 63, 1 1 77. WAHLENBERG, G. 1818. Petrifacta Telluris Svecanae. Nova Acta R. Soc. Sclent, upsal. 8, 1-116, 295-291 . WARBURG, E. 1925. The trilobites of the Leptaena Limestone in Dalarne. Bull. geol. Inst. Univ. Uppsala, 17, 1 -446. 1939. The Swedish Ordovician and Lower Silurian Lichidae. K. svenska Veten.sk Akad. Handl. 17 (4), 1-162. WEBER, v. N. 1932. Trilobites of Turkestan. Izd. Uses. Geol.-Razv. Oh'ed. NKTP. iv+ 157 pp. [In Russian with English summary.] 1948. Trilobites of the Silurian beds. No. 1. Lower Silurian trilobites. Mon. Palaeontol. U.S.S.R. 69 (1), 1-110. (In Russian.) WHITTARD, w. E. 1952. Cyclopygid trilobites from Girvan and a note on Boheniilla. Bull. Br. Mus. nat. Hist. (Geol.), 1, 305-324. 1961. The Ordovician trilobites of the Shelve inlier, west Shropshire. 6. Palaeontogr. Soc. [Monogr.] 197-228. 1966. Ibid. 8. 265-306. WHITTINGTON, H. B. 1950. Sixteen Ordovician genotype trilobites. J. Paleont. 24, 531-565. 1954. Ordovician trilobites from Silliman’s Fossil Mount. In miller, a. k., youngquist, w. and collin- SON, c. Ordovician cephalopod fauna of Baffin Island. Mem. geol. Soc. Amer. 62, 1 19-149. 1956. The trilobite family Isocolidae. J. Paleont. 30, 1 193-1 198. 1963. Middle Ordovician trilobites from Lower Head, western Newfoundland. Bull. Mus. comp. Zool. Harv. 129, 1-118. 1965. Trilobites of the Ordovician Table Head Formation, western Newfoundland. Ibid., 132, 275-441. 1968. A monograph of the Ordovician trilobites of the Bala area, Merioneth. 4. Palaeontogr. Soc. [Monogr.] 93- 1 38. williams, a., STRACHAN, L, BASSETT, D. A., DEAN, W. T., INGHAM, J. K., WRIGHT, A. D. and WHITTINGTON, H. B. 1972. A correlation of Ordovician rocks in the British Isles. Geol. Soc. Loml. Spec. Rept 3, I 74. xiA SHUFANG. 1978. Ordovician trilobites. 157-185. In Sudan to Permian stratigraphy and paleontology of East Yangtze Gorge area. Geological Publishing House, Beijing. [In Chinese.] Yi YONGEN. 1957. The Caradocian trilobite fauna from the Yangtze Gorges. Acta Palaeont. sin. 5 (4), 527- 560. [In Chinese with English summary.] YIN GONGZHENG and LEE SHANZi. 1978. Trilobita. 385-595 in Atlas of Palaeontology of southwest China, Guizhou Province. Geological Publishing House, Beijing. [In Chinese.] ZENG QINGLUAN, NI SHIZHAO, XU GUANGHONG, ZHOU TIANMEI, WANG XIAOFENG, LI ZHIHONG, LAI CAIGEN and XIANG LIWEN. 1983. Subdivision and correlation on the Ordovician in the eastern Yangtze Gorges. Bull. Yichang Inst. Geol. Mineral Resources Chinese Acad. geol. Sci. 10, 1 -56. ZHANG TAIRONG. 1981. Trilobita. 134-213. In Palaeontological Atlas of northwest China. Xinjiang (1). Geo- logical Publishing House, Beijing. [In Chinese.] ZHOU TIANMEI, LIU YiREN, MENG xiANSONG and SUN ZHENHUA. 1977. Trilobita. In Atlas of Palaeontology of centred and south China. Geological Publishing House, Beijing. 140-266. [In Chinese.] 786 PALAEONTOLOGY, VOLUME 29 ZHOU ZHIQIANG, LEE JINGSEN and QU xiNGGUO. 1982. Trilobita. 215-460. In Palaeontological Atlas of northwest China: Shaanxi, Gansu and Ningxia Volume, Part 1, Pre-Cambrian to Early Palaeozoic. Geological Publishing House, Beijing. [In Chinese.] ZHOU ZHiYi, YIN GONGZHENG and TRIPP, R. p. 1984. Trilobites from the Ordovician Shihtzupu Formation, Zunyi, Guizhou Province, China. Trans. R. Soc. Edinh. 75, 13-36. ZHOU ZHIYI Institute of Geology and Palaeontology Academia Sinica Chi-Ming-Ssu Nanjing China Typescript received 19 September 1985 Revised typescript received 15 April 1986 W. T. DEAN Department of Geology University College CardilfCFl IXL THE DISAPPEARING PEEL TECHNIQUE: AN IMPROVED METHOD FOR STUDYING PERMINERALIZED PLANT TISSUES by JOHN HOLMES Cllld JOELLE LOPEZ Abstract. A new technique is described for the preparation of sections of anatomically preserved plant fossils starting from the simple and well-known cellulose acetate peel method. Improvements of the results obtained by the peel method are described whereby use of very dilute acid allows extraction and observation in planar view of cell walls that are almost always destroyed by traditional methods. The Disappearing Peel Technique has been especially developed for histological investigations of the delicate cell walls of extra-xylary vascular tissue. Fern phloem from a Carboniferous coal ball is used to demonstrate this new method. It includes transfer of the peel to Araldite, dissolving the peel in acetone, complete demineralization of the plant cells remaining on the Araldite and allows optical examination of cell walls under immersion oil. Artefacts in coverslip and peel as well as residual carbonate content of the peel that may simulate cell wall sculpturing are thereby eliminated. Furthermore SEM and TEM observation of the same cell walls previously examined by the light microscope is possible. This technique will be applied to evolutionary studies of vascular tissues in Palaeozoic ferns. The techniques described here were developed while the authors were initiating studies on the vascular tissues of Palaeozoic ferns (Anachoropteridaceae, Botryopteridaceae, Psalixochlaenaceae) from British coal balls with a view to elucidating their phylogeny. Fertile structures of these essentially Carboniferous ferns are extremely rare and the number of taxa known to date, although confirming the leptosporangiate character of these ferns, is too small to fill in the gaps in our knowledge of their phylogeny. Morphological studies have revealed a wide range in anatomy, branching patterns, and habit (see references in Taylor 1981, pp. 525-527; Stewart 1983, pp. 193- 194). Even so, phylogenetic trends within the group are not clear and their possible ancestry in Devonian plants and the extent of their connection with modern ferns remains unknown. These ferns are generally simple protostelic plants and possess only a small number of variable characters compared with gymnosperms or angiosperms. In the absence of fertile parts classification tends to be based on the xylem anatomy of leaves. One variable and widely available character that has only recently begun to be studied in detail and that may be of phylogenetic value in fossil ferns is the sculpturing of the cell walls of the vascular tissue, notably the phloem (Smoot 1979, 1985; Smoot and Taylor 1978, 1984). The techniques presented in this paper were designed as a result of repeated failures by the authors to view successfully the phloem tissues in the stem of an anatomically preserved (coal ball) Palaeozoic filicalean fern in optical microscopy (OM), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The morphology of this fern, Psalixochlaena cylimirica (Will) Holden, is known (Holmes 1977, 1981) but recently attempts to characterize the phloem tissues were fraught with artefacts in all three modes of observation, i.e. OM, SEM, and TEM. The word artefact is used here in the sense of phenomena that are not of biological origin or that are created during the preparation of the material. Additionally it was discovered that the use of traditional strengths of hydrochloric acid (1-6 %) used in the preparation of cellulose acetate peels (Joy et al. 1956) is almost totally destructive of delicate cell walls lying parallel to the coal ball surface. During optical microscopy examination of peels prepared using 6 % HCl and containing thin walled phloem cells, this destruction led to false observations caused by cell wall debris and [Palaeontology, Vol. 29, Part 4, 1986, pp. 787-808, pis. 66-70.| 788 PALAEONTOLOGY, VOLUME 29 the presence of calcite embedded within the peel. The new Disappearing Peel Technique has been devised not only to eliminate artefacts, but to totally isolate delicate cell walls from their mineral matrix with very dilute HCl whilst constantly providing them with a support to prevent collapse. Starting from the cellulose acetate peel method, this technique also enables the rapid preparation and transfer of the same group of cells to either of the three types of study, OM, SEM or TEM, with ease and a minimum of destruction. METHODS Initial methods of investigation, the peel method, and artefacts (a) Optical microscopy. The traditional peel technique (Joy et cd. 1956) consists in etching with HCl a cut and polished surface of a carbonate rock containing plant fossils (text-fig. 1a). HE is used for silica-preserved fossils. The cell walls then stand proud of the surface which is wetted with acetone. A sheet of cellulose acetate is applied which is softened by the acetone and embeds the cell walls. When dry, the sheet is peeled off (text-fig. Ic), hence the name ‘peel’ and, after washing in HCl to remove adhering crystals (text-fig. Id, e), the peel is mounted in Canada balsam on a slide under a coverslip. There are many disadvantages to this traditional method. Observation of the acid-etching process under the binocular microscope revealed that HCl as dilute as 1 % was extremely destructive of delicate cell walls lying parallel to the coal ball surface. Face view observation of the sculpturing of these walls in OM and SEM is essential to the type of study being undertaken. Those cell walls lying at the surface are destroyed by effervescence as the acid penetrates beneath them and reacts with the calcite crystals filling the cell (text-fig. 1a, b). If the lower wall from the opposite side of the cell is pulled out by the peel (text-fig. Ic) then it is destroyed by the effervescence of its own calcite content as the peel is demineralized (text-fig. Id, e). Therefore, only the cell outline remains to be seen in the peel. The result, in the stelar tissues of P. cylindrica, is that only a few tracheid walls will be visible in face view because their scalariform thickening confers the mechanical strength necessary to resist the effervescence. This destructive process was observed on large xylem cells; most phloem EXPLANATION OF PLATE 66 Fig. 1. SEM. Circular to oval configurations 1-5-6 0 /;m in cellulose acetate peel produced by wetting with acetone, x 1,000. Fig. 2. SEM. Regular scattering of OT pm spots on coverslip, x 14,400. Figs. 3-11. Psalixochlaena cylindrica. 3, OM, phase contrast. Apparent pores O-S-l-S /im in wall of phloem cell, X 2,000. 4, OM, normal light. Apparent pores forming sieve areas in phloem cell wall. Effect probably produced by calcite in peel, x 2,000. 5, SEM. Contents of phloem parenchyma cell, x 7,200. 6, OM. Contents of phloem parenchyma cell. Note resemblance to sieve areas in cell wall, x400. 7, SEM. Etched surface of calcite viewed in fault in xylem zone. Note numerous pits (arrows) in surface, x 3,600. 8, SEM. Phloem parenchyma cell viewed on etched surface of coal ball. Apart from outline of cell wall (w), distinction between organic matter and residual calcite is not clear, x 1,080. 9, SEM. View inside sieve cell on etched surface of coal ball. Pitted wall in face view cannot be identified with certainty as calcite or as cell wall, X 1,200. 10, TEM. Vertical section through xylem wall showing empty bars (arrows) of scalariform thickening, x 2,700. 1 1, SEM, idem fig. 10, x 1,200. Fig. 12. TEM. Section through gold layer put on Araldite by sputter coater showing lump of gold (arrow) 0-4 pm across, x 6,800. Fig. 13. TEM. Section through wall of phloem cell showing false pores (arrows) of 0-4 pm probably caused by stretching of Araldite. Note their barrel shape which recalls that of plasmodesma-derived pores, x 10,000. Figs. 14-18. OM, normal light. Photographs taken inside a phloem cell. The arrowed area appears as a particle or group of particles (figs. 14, 15, 18) or as a hole (figs. 16 and 17) depending on the plane of focus. Opening (fig. 18) or closing (fig. 14) of the diaphragm totally alters the aspect of the field of view, x 3,000. PLATE 66 HOLMES and LOPEZ, Psalixoclilaena 790 PALAEONTOLOGY, VOLUME 29 cells are too small to be observed under the binocular microscope but the almost total lack of any of their walls lying in face view in the peel suggests that they are destroyed due to their more fragile nature. As is demonstrated by Plate 70, fig. 5, a good deal of cell wall is ripped from the matrix by the peel and projects unprotected, but held in place by its calcite content, from the lower surface of the peel. The habit of some palaeobotanists of scrubbing their peels in acid is therefore highly destructive of longitudinally orientated walls. If the peel is not demineralized the lower cell wall is masked by its own calcite content in OM. This cell wall destruction, combined with artificial optical phenomena described here caused the authors to observe and photograph false phloem cell walls for some considerable time. Peels containing longitudinal sections of P. cyliudrica phloem were mounted in Canada balsam under a cover slip and viewed with a Leitz Orthoplan microscope in both normal and polarized light and a Leitz Ortholux for phase contrast. Plate 66, fig. 3 is typical of some of the pictures of ‘cell wall’ obtained in phase contrast and Plate 66, fig. 4 in normal light, where a scattering of circular pores about 0-75-2 0 /^m in diameter appears to occur in the wall in planar view. Several obstacles arose to the interpretation of pores within cell walls; some circular objects proved to be dust particles in the microscope since they occurred in the same position in every picture. SEM examination of surfaces of coverslips and of cellulose acetate sheet that had been previously wetted with acetone revealed a regular scattering of 01 /im particles on the glass surface (PI. 66, fig. 2) and an irregular occurrence of circular patterns from 1-5 to 6 0 j.i\n in the cellulose acetate (PI. 66, fig. 1 ). Further photographs from peels mounted in Canada balsam are shown in Plate 66, figs. 14-18 and Plate 67, figs. 2, 3, 5, 7. They illustrate several other artefacts likely to be interpreted as sculpturing of cell walls. These artificial phenomena often occur where sieve cells are expected to be found. Plate 67, fig. 2 shows an apparent wall pattern viewed under polarized light that closely resembles that of the leptoid walls of Polytrichum commune viewed under the same conditions (PI. 67, fig. 1). In the photograph (PI. 67, fig. 2) there is in fact no cell wall present in face view and the ‘sculpturing’ effect may result from diffraction and interference patterns around calcite particles in the peel accentuated by over-use of the diaphragm. Plate 67, fig. 3 illustrates what was initially interpreted as a tangential sieve cell wall with a sieve area and other isolated pores. When these two examples were re-photographed under varying optical conditions it proved impossible to reproduce an identical negative. These artificial images are due not only to calcite embedded in the peel, but also to combinations of the many variable conditions of optical photography, e.g. condenser level, light intensity, polarization, diaphragm, etc. The peel shown in Plate 67, fig. 3 proved to be EXPLANATION OF PLATE 67 Fig. 1. OM, polarized light. Polytrichiim commune leptoid wall. Re-photographed from specimen figured by Hebant(1964, fig. 6), x 2,000.' Figs. 2-9. Psalixochlaena cylindrica. 2, OM, polarized light. View inside sieve cell in longitudinal section. No cell wall is present here, note similar aspect to fig. 1, x 2,000. 3, OM. Tangential longitudinal section through sieve cell. Appearance of sieve area (large arrow) and other isolated pores (arrows) is due to calcite content of peel, x 600. 4, OM. Same peel as fig. 3 after 24 hrs. in 1 % HCl. Note disappearance of apparent perforated wall, x 600. 5, OM. Contact area between two sieve cells, note apparent perforated nature of lateral walls (arrows), x 600. 6, SEM. Exactly the same field of view as fig. 5 after preparation of section by Disappearing Peel Technique but without demineralization. The calcite content of the cells masks details of cell walls and creates in fig. 5 the impression of perforated lateral walls, x 600. 7, OM. Contact zone between two sieve cells with apparent wall sculpturing imitating sieve areas, x 1,000. 8, OM. Same field as fig. 7 after treatment by Disappearing Peel Technique, apparent wall sculpturing was due to calcite and has been removed by FfCI. Note that although cell wall is shown to be present by SEM in fig. 9, it is too thin to be observable here, x 1,000. 9, SEM. View of area where false wall sculpturing due to calcite has disappeared in fig. 8. Note delicate cell wall with pores (arrows) about 10 //m in diameter, x 3,000. PLATE 67 HOLMES and LOPEZ, Polytrichum, Psalixochlaena 792 PALAEONTOLOGY, VOLUME 29 incompletely demineralized. It was removed from the slide and left overnight in 1 % HCl. When re- photographed (PI. 67, fig. 4), the ‘perforated wall’ had disappeared, demonstrating that the illusion of a cell wall had been produced by calcite filling the intracellular space. One factor probably responsible for the production of artefacts by diffraction and interference in the etch-peel technique is the white powder of some 300-500 i.im depth that forms on the coal ball surface after etching with HCl. Despite placing peels in HCl to remove the adhering crystals pulled off with the peel, some of this powder is actually incorporated within the peel (text-fig. 1e) and is protected when the peel is demineralized in acid. This was demonstrated by scraping the undersur- face of a demineralized peel, and treating the scrapings with HCl where they were seen to effervesce. The formation of this powder seems to be accompanied by an increase in volume as it is often level with or above cell walls exposed by etching prior to making a peel. Comparison of Plate 67, figs. 5, 6 shows the extent to which this powder masks details of cell walls. Analysis by X-ray diffraction showed that this white powder retains a crystalline structure and may simply be partially dissolved CaC03. Its composition is not clear, a small calcium peak occurred with a larger peak for an unidentified component. Tests for chlorides were negative. A priori this powder is another source of diffraction in light microscopy and may create the impression of pores and mask details of true cell walls. Plate 67, figs. 5, 7 illustrate phloem end walls which, in OM, appear to have some kind of sculpturing. Application of the Disappearing Peel Technique to these specimens revealed the artificial nature of these wall patterns. Experiments in ‘peeling’ the surfaces of extant leaves showed that cellulose acetate perfectly moulds the smallest detail of cells and stomata and that this is best seen in phase contrast micro- scopy. It is possible, therefore, that when a peel is removed from a coal ball surface it has moulded not only the carbonate surface but cell walls that have remained in the matrix. The apparent perforations seen in phase contrast in Plate 66, fig. 3 and in normal light in Plate 66, fig. 4 may be due not only to patterns in the coverslip (PI. 66, fig. 2) and circular areas (PI. 66, fig. 1) or calcite powder in the peel but also to moulding of the carbonate surface or cell wall by the acetate sheet. Work on extant Filicales (Esau 1969; Hebant 1969; Lamoureux 1961; Liberman-Maxe 1968, 1978; Maxe 1964) has shown that these ferns have wall perforations of 0-2- T5 /(m or more rarely up to 4 0 pm in their metaphloem cell walls. One logically expects to find similar perforations in the phloem walls of a Palaeozoic filicalean. However, this size range of pores is exactly that of the potentially interfering objects in peels and coverslips described above. The permanent calcite content of the peel is another factor that may create the appearance of pores by optical effects. Around the 0-5 pm order of magnitude and below (the limit of resolution of OM) a particle and a hole of the same size give the same diffraction pattern and turn from dark to light on changing focus. Addition- ally, use of the diaphragm can totally change the image at high magnification. Plate 66, figs. 14-18 are five photographs of the same field within a phloem cell showing an area which may appear as a particle (PI. 66, figs. 14, 15, 18) or a hole (PI. 66, figs. 16, 17) depending on the level of focus. This area totally changes in appearance with different diaphragm values (compare PI. 66, figs. 14 and 18) and it was impossible to judge, at such high magnification, whether the area represented a pore or whether indeed cell wall was present. Several interfering factors (dust, particles on coverslip, patterns and calcite in peel) are present in traditional preparations of longitudinal peel sections through permineralized tissues. We have found that these are likely to create interference and imitate a pitted cell wall, specifically in areas where cell wall is not, in fact, present in planar view. Cellulose acetate sheet, previously wetted with acetone, develops circular areas which do not appear to be bubbles (PI. 66, fig. 1). All these factors, and the alteration of image obtained by changes in diaphragm closure and focusing, caused the authors to feel that little faith could be placed in high-magnification optical micrographs such as those referenced in this section where apparently perforations are present in a cell wall. It was concluded that not only is it difficult to detect if cell wall is really present but cellulose acetate sheet is not a good medium through which to detect minute pores in thin cell walls seen in planar view. HOLMES AND LOPEZ: PERM INERALIZED PLANT TECHNIQUES 793 (h) Scanning electron microscopy. Stems of the fern P. cylindrica were examined on both broken and sawn etched surfaces in longitudinal section. The specimens figured here received 250-300 A of gold in a Polaron E5000 sputter coater and were examined with a Jeol JSM 35 Scanning Electron Microscope. Although the phloem zone is easy to locate external to the xylem, the photographs proved difficult to interpret for the following reasons. First, the different cell layers are not as easy to identify as they are in a series of peel sections viewed in OM; secondly, with the JSM 35 used by the authors there is no method of knowing whether one is viewing carbonate or organic matter in the image provided by the secondary emission of electrons. The phloem cells under investigation generally contained a mass of irregularities and projections which probably correspond to the white powder mentioned above which remains after etching (PI. 66, fig. 8). Although this powder dissolves in ffCl, further etching to reach a stage where empty cells were visible caused the collapse of these cells. Smooth areas of calcite in a fault in the coal ball where no powder was present proved to contain a scattering of pits 0'2-0-8 / V CVn N \ N n » s s \ s ✓ ////// s. \ V \ \ \ \ V V \ \, \ \ > .NNNNNN' Coal ball H PEEL ^ ^ ^ E f J 796 PALAEONTOLOGY, VOLUME 29 The investigator must use rip peels, one micron peels, or deeper etch peels in an order that obtains the maximum amount of information from cells exposed at any one moment. The Disappearing Peel Technique Slides must be prepared in the following manner. A portion of cellulose acetate sheet, slightly larger than the peel that is to be mounted, is placed on a glass slide wetted with acetone. When the acetate sheet dries, it adheres to the slide merely by the exclusion of air. The acetate is then covered with a thin layer of commercial Araldite (manufactured by Ciba-Geigy and sold in two tubes, epoxy-resin and hardener) which must spread beyond the edges of the acetate sheet on to the glass. The glue is left to harden for at least 12 hrs. at 55 °C (text-fig. 2a). The acetate sheet, which is simply in contact with and does not firmly adhere to the glass, serves to liberate chosen portions of the final preparation when cut out for TEM. The layer of Araldite provides a rigid support for the organic matter during the following stages. A peel containing the fossil section is glued, with the same type of Araldite, rough side down on to a slide prepared as above (text-figs, li, 2b). It is first placed smooth side down on a heavily greased glass plate. A small amount of Araldite is spread over the rough, now uppermost, surface of the peel where the organic matter is embedded, and also over the prepared area of the slide. Both are heated for 5 min. at 55 °C. This renders the Araldite quite fluid. The slide surface with the liquid Araldite on it is turned over on to the peel on the glass plate and a lead weight placed on top. The preparation is polymerized at 55 °C for 45 min. after which time the slide is easily removed from the glass plate. The timing is critical as after longer periods the slide may stick to the glass despite the grease. It is therefore advisable to predetermine the minimum time taken for the glue to become hard but of a rubbery consistency. Excess grease is removed from the peel with xylol and polymerization is continued overnight at 55 °C. This type of mounting is no more arduous than mounting under a coverslip in balsam and the peel can be examined by transmitted light at this stage. Any blurring caused by irregularities in the peel surface can be eliminated by applying a smear of immersion oil. In the above preparation, commercial Araldite is used for its resistance to the acetone and acid treatments that now follow and provides a suitably transparent support for OM by transmitted light. The slide is placed, peel side sloping downwards, in an acetone bath with a magnetic stirrer for 15 min. and during this time the cellulose acetate dissolves completely (text-fig. 2c). The organic matter remains undisturbed and is now exposed on the Araldite which supports its lower surface (text-figs. 1j, 2d). Demineralization is now effected to remove the calcite powder previously imprisoned in the peel. The disappearance of a white hue from the section indicates this change. Eor longitudinal sections where small cells filled with calcite are likely to have been removed from the coal ball, a 1-2 hrs. bath of 0 05 % HCl is recommended to avoid breakage of cells and their contents by effervescence. The exact demineralization time can be determined by an examination under polarized light. Treatment with HCl seems sufficient to remove the mineral matter for coal ball plants. (a) Optical microscopy. The hollow left by the dissolving of the acetate peel is filled with a drop of immersion oil (text-fig. 2e). A slight vacuum may be necessary for a few minutes if air bubbles TEXT-FIG. 2. Applications of the Disappearing Peel Technique to OM, SEM, and TEM. a, b, peel with cells is glued with Araldite on to a slide prepared as described in text, c, peel is dissolved by 15 min. agitated bath in acetone, d, the dissolution of the peel leaves the cells exposed in a hollow in the Araldite. e, optical microscopy is performed after filling the hollow with immersion oil. Immersion objectives dip directly into the oil. f, after removal of immersion oil by acetone (as in c) the slide is placed on SEM stub and gold coated. Note the carbon bridge. G, in preparation for TEM studies, the cells are embedded by filling the hollow with Araldite. H and I, a selected portion of the slide is dissected out and glued on to an Araldite stub for sectioning with ultramicrotome. HOLMES AND LOPEZ: PERMINERALIZED PLANT TECHNIQUES 797 ULTRAMICROTOME H 798 PALAEONTOLOGY, VOLUME 29 remain inside entire cells. Wall sculpturing of stelar cells may now be photographed without interference from peel or coverslip. Oil immersion objectives may dip directly into the oil covering the preparation (text-fig. 2e). This is particularly useful as it solves the often-encountered problem where the combined thickness of coverslip and peel prevents focusing of some x 100 objectives. At this stage of the preparation the investigator has the choice of performing either SEM or TEM on groups of cells that appear of interest under the optical microscope. {b) Scanning electron microscopy. The immersion oil is removed from the slide by washing in acetone for 2 min. (as in text-fig. 2c) and the end of the slide bearing the specimen is broken off after scoring with a diamond and mounted on an SEM stub with double-sided tape. The observation chamber of the JSM 35 conveniently housed the 25 x 25 mm slide bearing the preparation. Two bridges of conducting carbon glue must be made between the upper surface of the slide and the stub before metal coating (text-fig. 2f). The greatest advantage of this type of preparation is obtained by giving the coal ball surface a very high polish (as for a one micron peel) before etching and making the peel. In observing longitudinal sections not only can one see cell walls in planar view but the vertical cell walls are seen neatly sectioned by the polishing with their pores cut at different levels (PI. 69, figs. 5, 6). The SEM picture thus combines the advantages of depth of field with a section view of walls rising towards the observer. As already pointed out, when etched pieces of coal ball containing fossil tissue are examined by SEM, it is difficult, if not impossible, to distinguish organic matter from its mineral support (PI. 66, figs. 8, 9). The same problem exists when the organic matter from a peel is transferred to its Araldite support which also contains many pits. This is probably because the liquid acetate sheet moulds the calcite surface where pits occur (PI. 66, fig. 7) and in our new technique the fiuid Araldite then moulds the peel. A control SEM observation of the Araldite surface before and after acetone and acid treatment revealed it to be perfectly smooth in both cases. The transfer of these pits to the Araldite by moulding is one confusing artefact that we have not been able to eliminate. This highlights the fact that a careful observation of the tissues is first necessary by a series of light micrographs at different levels of focus (PI. 68, figs. 3, 4). This not only ascertains the extent of cell walls but an optical map of overlapping photographs helps rapidly locate a given area under SEM. Arrows can also be engraved in the Araldite as an aid to location. In a radial section through a stem it is almost impossible to obtain a face view of walls of all cells comprising the phloem zone in order to compare the sculpturing of the different types of cell wall. This information can be more easily obtained by peeling well-preserved axes sectioned transverse- obliquely at 20-30° and etched to a depth of 15-20 yum after fine polishing. After transfer to EXPLANATION OF PLATE 68 Figs. I -7. Radial wall of sieve cell of Psalixochlaena cylindrica with numerous pores. Rip peel prepared by the Disappearing Peel Technique. 1, OM. Areas where cell wall are present appear dark. Plane of focus is low causing pores at top and bottom of picture to appear white. CA = calcite, x 700. 2, OM. Higher plane of focus causing all pores to appear dark. Note false pores (FP) at bottom of picture where no wall is present. Figs. 1 and 2 are not demineralized, note zones of calcite (CA) imitating perforated wall, x 700. 3, OM. Demineralized section. Full extent of perforation is visible as plane of focus, intermediate between figs. 1 and 2, causes almost all pores to appear as white patches, x700. 4, OM. Demineralized section in same plane of focus as fig. 1 . Arrow shows 0-5 /nu pore (magnified under SEM in fig. 7) in focus as white patch, X 700. 5, SEM of same cell shows up relief, notably bulging part of wall in lower part of picture which is probably part of end contact wall, x 700. 6, OM. Detail of uppermost portion of cell showing dark patches on wall. These are not focusable as white and without SEM confirmation (fig. 7) cannot be identified as minute pores or particles, x 2,000. 7, SEM. Same field as fig. 6. Some but not all of dark patches correspond to particles on the cell wall. Note 0-5 ^m pore (arrow) focused as white patch at top of fig. 4, x 2,000. CA PLATE 68 HOLMES and LOPEZ, Psalixochlaena 800 PALAEONTOLOGY, VOLUME 29 Araldite and complete demineralization, the stem is viewed under SEM where the walls are seen rising obliquely from the support. By this preparation one can observe almost in face view, on one side of the stem the inner, and on the other side of the stem the outer tangential walls of all tissues with the rising radial walls seen in section view (PL 69, fig. 6). At 90° to this axis both sides of radial walls can be examined in face view with tangential walls in section view. This type of preparation is perhaps the richest in information with the limitation that only about 40-50 /m lengths of wall are visible and any possible patterns of alternating smooth and pitted wall can only be viewed over greater lengths in longitudinal section. The thickness of this type of preparation renders prior optical examination uninformative. (c) Transmission electron microscopy. After optical examination and photography of areas of interest, the immersion oil is washed away from the slide with acetone and replaced, after drying, by a drop of embedding quality Araldite (M Araldite) containing hardener and accelerator (text- fig. 2g; composition: Araldite M-lOml, Hardener HY 964-10 ml. Dibutyl phthalate-1 ml. Well mix all three before adding Accelerator DY 064-0-6 ml). The slide is put under a — 1 bar vacuum overnight and then hardened at 55 °C for two and a half days. The embedding of a thickness of only one or two cell walls is rapid, and eliminates the need for several baths of progressively less dilute Araldite solutions. If no further use is immediately made of the slide, this stage represents a good permanent preparation of a demineralized section on which optical photography can be performed. Small portions of the preparation about 1 mm x 2 mm containing cell wall to be microtomed are dissected out. The layer of cellulose acetate sheet in contact with the slide (text-fig. 2h) immediately liberates the dissected portion which is then glued on to the end of an Araldite stub with the required orientation (text-fig. 2i). If the slide contains a tangential section, then several cells of the same tissue can fairly easily be dissected out. For radial sections as in the case of the sieve cell shown Plate 70, fig. 2, it may be desired to section one cell only. A microscope is then set up horizontally to view the stub in the microtome vice and to compare the cells in the dissected portion with optical micrographs taken previously. It is thus possible to spot the tissue from which the ultra-thin sections originate when the knife passes through the chosen cell. RESULTS A lesson in artefacts The authors’ conclusion is that, contrary to xylem walls where well-defined scalariform pitting is present, the identification of a thin phloem cell wall perforated by minute pores is highly problema- tical. When treated with the Disappearing Peel Technique and demineralized, the end wall in Plate 67, fig. 7 loses its artificial sculpturing (PI. 67, fig. 8), and examination of the same field by SEM reveals a partly intact cell wall with scattered perforations to the order of 1-0 /tm in diameter (PI. 67, fig. 9). It is important to point out here that this peel was not washed in HCl following removal from the coal ball surface: had this been done the delicate cell wall rendered visible under SEM by the Disappearing Peel Technique would have been destroyed by the effervescence of the same calcite content causing artificial wall sculpturing. These transverse crystalline structures across the end walls have been observed several times and only in the sieve cell position. It has been suggested to us (J. Moret, pers. comm.) that this may be due to surface tension or gravitational effects. The cell in Plate 67, fig. 3 was not treated by our new technique but was merely washed in 1 % HCl for 24 hrs. It is seen to have no tangential wall after acid treatment (PI. 67, fig. 4), and the entire aspect of a pitted wall in this figure was created by a calcite infill. By washing the peel in acid, any cell wall that might have been present behind the calcite has been destroyed by effervescence of the latter. Thus an important aspect of the Disappearing Peel Technique is to protect these cell walls that project from the lower surface of the peel by providing them with an Araldite support on a slide as soon as they are removed from the coal ball (text-fig. li). HOLMES AND LOPEZ: PERMINERALIZED PLANT TECHNIQUES 801 Plate 67, fig. 5 shows a terminal contact wall between two phloem cells. The side walls appear to contain scattered circular 1 0 /im perforations. This field, viewed by SEM after the Disappearing Peel Technique but without demineralization, shows the upper layer of calcite powder picked up by the peel (PI. 67, fig. 6). Many depressions occur in the carbonate which are responsible for the apparently perforate wall seen in optical view (PI. 67, fig. 5). In view of these results, optical micrographs (PI. 66, figs. 3, 4; PI. 67, figs. 2, 3, 5, 7) from mounted peels and scanning electron micrographs from etched coal ball blocks (PI. 66, fig. 9), where groups of pores or sieve areas appear to be present in a cell wall, must be accepted with great caution. There is doubt not only as to whether real pores are being viewed but also as to whether organic cell wall is actually present. Finally, the first pores the authors viewed with TEM were very convincing indeed (PI. 66, fig. 13), but also difficult to interpret as this type of preparation was being viewed for the first time. It was only the successful viewing of real cell walls in planar view (PI. 68, figs. 1 -4; PI. 70, fig. 7), either from rip peels or from peels made after very dilute acid treatment and their preparation by the Disappearing Peel Technique, that finally revealed the artificiality of the above examples. Botanical results Plate 68, figs. 1-5 shows in face view the radial wall of a sieve cell picked up by a rip peel. The wall appears dark in light microscopy and the 0-5-1 -5 / TEXT-FIG. 4. Schematic diagrams of cup plating in cladid inadunate crinoids. A, "Cyathocrinites' ramosus (Schlotheim). b, Cyathocrinites, based on the type species C. planus Miller, 1821. Basals and infrabasals unshaded; radials black (with unshaded articular facets); plates of the anal series stippled; X, anal X plate. Shading scheme after Moore (1962). 814 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 5. 'Cyathocrinites' nimosus (Schlotheim), UCG C26a, camera lucida drawings of the dorsal cup. a, lateral view with CD interray central, b, lateral view with A ray central, c, base with lumen surrounded by infrabasals central and A ray top. d, arti- cular facet of A ray radial. The only available cup of ‘C.’ ramosus most closely resembles that of an advanced poteriocrinine inadunate but is not sufficiently well preserved to enable a generic identification. The precise taxonomic position of ‘C.’ ramosus is therefore indeterminate. "Cyathocrinites'' ra/?;o5;/x (Schlotheim, 1816-1817) Plates 71 and 72; text-figs. 4a, 5, 6, 9 1 8 1 6 1817 Eucrinites ramosus Schlotheim, p. 20, pi. 2, fig. 8; pi. 3. 1820 Eucrinites ramosus Schlotheim; Schlotheim, p. 20, pi. 2, fig. 8; pi. 3, fig. 19fl, h. 1848 Cyathocrinus ramosus (Schlotheim); Geinitz, p. 16, pi. 7, figs. 3-6. 1848 Eucrinites planus Miller; Howse, p. 259. 1850 Cyathocrinus ramosus Schlotheim; King, pp. 50-52, pi. 6, figs. 15-20. 1861 Cyathocrinus ramosus (Schlotheim); Geinitz, p. 109, pi. 20, figs. 10-14. 1866 Cyathocrinus ramosus (Schlotheim); Geinitz, p. 62, pi. 4, fig. 19. 1894 Cyathocrinus ramosus (Schlotheim); Netschajew, p. 116, pi. 1, figs 4, 5. 1898 Cyathocrinus ramosus (Schlotheim); Spandel, p. 28, pi. 12, figs. 1-4, 8-20. 1913 Cyathocrinus ramosus (Schlotheim); Trechmann, p. 215. 1943 Cyathocrinites ramosus Schlotheim; Bassler and Moodey, p. 120. 1943 Cyathocrinites (?) ramosus (Schlotheim); Bassler and Moodey, p. 395. 1943 Cyathocrinus ramosus (Schlotheim); Trechmann, p. 343. 1948 Cvfl//;oa7>?r« (Schlotheim); Branson, p. 189. 1967 crinoid columnals; Smith and Francis, p. 178. 1977 Cyathocrinites ramosus (Schlotheim); Pattison, p. 44. 1980 Cyathocrinites sp.; Pettigrew, p. 18. Material, horizon, and localities (text-fig. I). King’s original specimens are numbered UCG C26a-g. The unique dorsal cup is UCG C26a; all other specimens are pluricolumnals. UCG ?C27 is a poorly preserved pluricolumnal for which a definite identification is impossible. All of this material comes from Tunstall Hill, Sunderland, Tyne and Wear (NGR NZ391544, approximately). Column material in the collection of Sunderland Museum (SuM) comes from three localities: SuM B2769 (fifty-four specimens) from Ford Quarry (NGR NZ362572); SuM B2770 (106 specimens) from Tunstall Hill; and SuM B2772 (twenty-two specimens) TEXT-FIG. 6. "Cyathocrinites" ramosus (Schlotheim), camera lucida drawings of pluricolumnals and dissociated ossicles from Tunstall Hill (a-e) and Beacon Hill (f-j). a, b, BMNH E70126, lateral views of a pluricolumnal from opposite sides, showing wedge-shaped tertinternodals which contribute to the curvature of the specimen and cirrus scars arranged in three rows (indicated by arrows), although no cirrinodal bears three scars, c, o, BMNH E70127, a pluricolumnal from the mesistele; c, lateral view, showing regular arrangment of columnals (N3231323N) and nudinodals; d, articular facet and convex latus of a nodal, e, BMNH E70128, part of a pluricolumnal, with a cirrus arising from a cirrinodal at a slight angle (noditaxis arrangement N212). f, BMNH E70138, articular facet of a radial plate (cf. text-fig. 5d). g, h, BMNH E70130, E70I32 respectively, brachial ossicles, i, BMNH E70129, CD interray basal. J, BMNH E70134, pluricolumnal with strongly curved cirrus and (apparently) tall cirral ossicles. 816 PALAEONTOLOGY. VOLUME 29 from Humbledon Hill (NGR NZ380552). New material has been collected by Hollingworth and Donovan from Tunstall Hill (BMNH E70I26-E70I28) and Beacon Hill railway cutting near Seaham (BMNH E70I29-E70172; NGR NZ44I455). All of these localities are within the Eord Formation reef facies (see above), Zechstein Cycle I, Upper Permian. We have been unable to trace Schlotheim’s original material. Diagnosis. A species of cladid inadunate crinoid with a broad, bowl-shaped calyx which tapers to a relatively narrow base. Infrabasals and basals pentagonal, except for the basal in the CD interray, which is hexagonal. Radials broader than high, elongate heptagonal in outline, with elongate arm articulation facets. Anal X and the right proximal plate of the anal tube incorporated in the CD interray of the calyx, the latter supported by the C-ray radial and X. Arms unknown but brachials were broad proximally, becoming narrower and more U-shaped distally. Column circular to pen- tagonal in outline with a central, pentagonal axial canal with pentastellate jugula. Columnals have marginal crenularia and circular areolae. Stem xenomorphic, with a proxistele composed of numerous low columnals, a mesistele of taller columnals which lack cirri and a cirriferous dististele with a maximum of three cirrus scars per cirrinodal. Attachment was by cirriferous runner. Description. The arms are unknown apart from dissociated brachials. King (1850, p. 51) noted, ‘I have not yet succeeded in procuring any specimens of the arms or branches of Cyathocrinus ramosus’, though single joints have now and then occurred to me’. Two poorly preserved brachials have been collected from Beacon Hill by Donovan (text-fig. 6c, h). These are approximately semicircular in outline with a V- or U-shaped adoral food groove and apparently synostosial articulation (probably due to poor preservation). Such brachials are not suited to articulate on the radial facet (text-figs. 5d, 6f) and are, therefore, derived from a more distal part of the arm. The IBrj would have been somewhat broader, with a symplexial fulcral ridge parallel to the oral surface and an elliptical lumen aboral to this ridge. The lumen position indicates that the flexure of this articulation was principally away from the oral surface. The pattern of arm branching is unknown. The calyx is broad, conical, and tapers towards the base (text-figs. 4a, 5, 6f, i); dicyclic. Five elongate, pentagonal infrabasal plates (text-figs. 4a, 5c) which were presumably visible in lateral view (in UCG C26a all five infrabasals are damaged, possibly due to mechanical damage after the cup had been glued to a board for display). Infrabasal plates and angles of the small, pentagonal lumen are radial in position. The stem articulation facet at the base of the cup is unknown. Five basal plates with slightly convex lateral surfaces (text-fig. 5a, b). Basal in CD interray hexagonal (text-figs. 5a, 6i), other basals pentagonal. Each basal supported by the two adjacent infrabasals (text-figs. 4a, 5a-c). Five broad, heptagonal radial plates, all slightly convex. Arm facets broad, ‘banana-like’, with a narrow, central, adoral food groove. A synarthrial articulation ridge lies parallel to the long axis of each facet but slightly aboral in position. An elliptical lumen is situated aborally to the fulcral ridge. Two short ridges, one at each end of the facet, are perpendicular to the fulcral ridge (text-figs. 5d, 6f). Lateral surfaces of radials and basals unsculptured. Anal X approximately pentagonal. Small, tetragonal right proximal plate of anal tube supported by X and C-ray radial. Stem fragments of this species are common fossils at certain localities and, following examination of over 300 specimens, it is confidently recognized that the column was zenomorphic and divided into three distinct regions (text-fig. 9). The proxistele is composed of very numerous low columnals of varying diameter. SuM B2770/11 (PI. 72, fig. 1) is a pluricolumnal from the proxistele formed of more than fifteen ossicles, either EXPLANATION OF PLATE 7 1 Figs. 1-8. 'Cyathocrinites' ramosus (Schlotheim). Tunstall Hill, Sunderland (except fig. 7); Ford Formation reef facies, Zechstein Cycle 1, Upper Permian. Scanning electron micrographs of features of the stem. I and 2, SuM B2770/9; 1, articular facet with a well-preserved pentastellate jugulum within the pentagonal lumen, X 11-5; 2, enlargement of the lumen to show the jugulum, x 23. 3, SuM B2770/18, cirrus scar with well- preserved crenularium and perilumen in depressed areola, x 23. 4, SuM B2770/13, pluricolumnal with disc-like juvenile attachment!?) extending over three columnals (including a cirrinodal, left), x 11-5. 5, SuM B2770/23, nodal with incipient cirrus scar, x 48. 6, SuM B2770/27, angled cirrus scars on cirrinodal, the upper scar retaining a wedge-shaped cirral ossicle (cf. PI. 72, fig. 5), x 11 -5. 7, SuM B2772/8, Humbledon Hill, Sunderland; cirral ossicle encroaching onto internodal adjacent to cirrinodal (note that the second cirrus scar is perpendicular, not angled, to the long axis of the column), x 9. 8, SuM B2770/20, pluri- columnal which is slightly curved due to a wedge-shaped columnal (centre), x 11-5. PLATE 71 DONOVAN et al., ‘‘Cyathocrinites' 818 PALAEONTOLOGY, VOLUME 29 N3231323 or N434243414342434(??). The nodals in this region are non-cirriferous. Latera convex, with either a ‘knobbly’ sculpture or unsculptured. Some columnals of the proxistele have raised rims which surround the articular facet. SuM B2770/19 (PI. 72, fig. 2) is from a more distal part of the proxistele, with a general increase in nodal height. The mesistele is composed of taller columnals than the proxistele, with planar or convex, unsculptured latera, and a columnal arrangement N3231323 (text-hg. 6c; possibly also Nl, N212 in some stems). Nodals in this region do not bear cirri (at least proximally) but what appear to be incipient cirrus scars are apparent on some nodals as depressed, circular grooves (PI. 71, fig. 5) or small pores (canaliculi). The ossicle enclosed by the circular groove probably develops into the distal primary cirral ossicle (Donovan 1984). The dististele is similar to the mesistele but nodals are usually cirriferous. Nodals in this region bear from zero to three cirrus scars (PI. 71, figs. 3, 4, 6, 7; PI. 72, figs. 3-5; text-fig. 6a, b, e, j). Cirrinodal height varies within single pluricolumnals. Cirrus scars are generally at 72° to each other and arranged in columns on the latera (arrowed in text-fig. 6b). Cirrus scars rarely extend onto adjacent internodals and may be flush with the latera, depressed within a raised cone (PI. 71, fig. 3), or angled towards the long axis of the stem (PI. 71, fig. 6; the direction of angling is consistent within pluricolumnals but it is not known if this was towards or away from the crown; both orientations are found in isocrinids but are not mixed in any species). Curvature of the dististele is sometimes aided by wedge-shaped tert- and quartinternodals (PI. 71, fig. 8; text-fig. 6a, b). Columnals are either circular (sometimes slightly elliptical, due to poor preservation in the majority of examples) or pentagonal with rounded angles. Latera are planar or convex. Articular facets are either circular or rounded pentagonal in outline (PI. 71, fig. 1; PI. 72, figs. 6-9). The lumen is central, pentagonal (sometimes appearing circular although this is due to poor preservation in many, if not all examples; text-fig. 6d), with the angles of the column and axial canal coincident. In exceptionally well-preserved specimens a pentastellate jugulum is seen, produced by claustra which slope towards the centre of the axial canal (PI. 71, figs. 1 and 2). The lumen is often surrounded by a narrow perilumen of irregular ridges and grooves (PI. 71, figs. 1 and 2) which in turn lies within a slightly depressed, circular areola (PI. 71, fig. 1; PI. 72, figs. 6-9; text-fig. 6d). Articulation is symplexial with marginal, radial crenularia showing a limited range of morphological variation. At first crenulae are short, peglike, and unbranched (PI. 72, fig. 6; text-fig. 6d). On a few specimens fine, paired ridges are seen to extend into the areola (PI. 72, fig. 7), which is perhaps a prelude to the culmina becoming longer and extending towards the lumen (PI. 72, fig. 8). These culmina show some slight bifurcation at the circumference of the facet. Further growth of the culmina, possibly related to an increase in columnal diameter, leads to increased bifurcation and implantation (PI. 72, fig. 9; Moore 1939, p. 184, fig. 4) until separation of branched crenulae occurs. In addition, some of the columnal-columnal articulations are curved (e.g. SuM 2769/4, where KFI(N) varies from T9 mm to 2 0 mm). Bivariate analysis of columnals and pluricolumnals from the collection of Sunderland Museum (text-fig. 7) enabled calculation of the following equations which define the nodal in ‘C.’ ramosus (symbols explained in the caption to text-fig. 7): FD = 6 0KH — 7-9 = 7-4LD— 1-6 = 0. 2Cn — 2 0. Cirri arise from the nodals (cirrinodals) of the dististele (and possibly also the distal part of the mesistele). The distribution of cirrus scars is discussed above. The articular facet sculpture of cirrus scars and cirral ossicles is similar to that of columnals. However, cirrus scars are sometimes concave and the crenulae of both scars and cirral ossicles do not bifurcate (PI. 71, figs. 3, 6, 7). Cirral ossicles are also much smaller than columnals. The axial canal outline in cirri is indistinct but may be circular. In rare examples the cirral ossicle EXPLANATION OF PLATE 72 Figs. 1-9. 'Cyathocrinites' ramosus (Schlotheim). Tunstall Hill, Sunderland; Ford Formation reef facies, Zechstein Cycle 1, Upper Permian. Scanning electron micrographs of features of the stem. 1-5, pluricolum- nals in lateral view; 1, SuM B2770/1 1, part of the proxistele (note that columnals are generally low, with obvious symplexial articulations between columnals and no cirrus scars on nodals), x 9-5; 2, SuM B2770/19, a pluricolumnal from a slightly more distal part of the stem, x 12; 3, SuM B2770/48, part of the dististele encrusted by the bryozoan DyscriteUa sp., x 12; 4, SuM B2770/4, a more robust pluricolumnal from a dististele, x 10; 5, SuM B2770/27, part of the dististele, including a cirrinodal with three cirrus scars (cf. PI. 71, fig. 6), X 8. 6-9, sequence of articular facets showing the probable pattern of crenularium growth; 6, SuM B2770/23, marginal crenularium of short, peg-like culmina, x 14; 7, SuM B2770/79, fine paired ridges extend from the culmina towards the lumen, x 12-5; 8, SuM B2770/18, lengthening of culmina towards lumen, x 10; 9, SuM B2770/4, facet with branching and secondary culmina, x 10. PLATE 72 DONOVAN et at., "Cyathocrinites 820 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 7. Graphs plotted from measurements oVCyathocrinites' ramosus (Schlotheim) taken from specimens in Sunderland Museum, a, facet diameter (FD) against nodal height (KH); FD = 6 0KH-7'9. b, FD against lumen diameter (LD); FD = 7-4LD-l -6. c, FD against number of crenulae per facet (Cn); FD = 0-2Cn-2 0. Lines of best fit determined by Bartlett’s method (Fryer 1966). adjacent to the cirrus scar is wedge-shaped (PI. 71, fig. 6) and very infrequently encroaches on the adjacent internodal (PI. 71, fig. 7). Cirri vary from straight (text-fig. 6e) to highly curved (text-fig. 6j); the latter example seems to have very tall cirral ossicles but they are usually broader than high and homeomorphic (?). A cirrus on BMNFl E70138 is 45 mm long but incomplete, approximately straight but slightly sinuous; for comparison, the longest pluricolumnal (on BMNH E70128) is 68 mm long. The attachment disc of one juvenile has been identified (PI. 71, fig. 4). Discussion. Although only a single complete dorsal cup of ‘C.’ ramosus is known from the Upper Permian of north-east England, pluricolumnals and columnals are plentiful. Dissociated plates from the crown are also known. The constancy of morphology shown by stem fragments indicates that only a single species of crinoid is present and, in consequence, all cup, arm, and column material are related with confidence. Peculiarities exist but these are rare and can be explained as being due to poor preservation. For example, two columnals on BMNH E70139 have a subsemicircular section which is reminiscent of the stem in myelodactylid inadunates. However, weathering has emphasized the growth lines, which are concentric, circular, and truncated against the flattened surface. It is concluded that these columnals were originally circular in outline but have had one side flattened, probably by weathering or pressure solution. Both are possibly derived from the same pluricolumnal. ‘C.’ ramosus differs from a modern isocrinid in not having regularly spaced, synostosially articulat- ing columnals (cf. Donovan 1984). In modern isocrinids synostoses occur between each nodal of the dististele and the adjacent distal internodal, called the infranodal (Breimer 1978, p. T24, fig. 1 1). These are the preferred zones of autotomy in the isocrinid stem (Emson and Wilkie 1980, pp. 200-201). There are no preferred autotomy surfaces apparent in the stem of ‘C.’ ramosus. Indeed, there is no direct evidence that Palaeozoic crinoids were able to autotomize. However, any functional examination that we make of an ancient crinoid must be prejudiced by our knowledge of recent, stemmed crinoids, particularly isocrinids. There are two conflicting conclusions that may be drawn from the absence of definite autotomy surfaces in the stem of ‘C.’ ramosus. Obviously, it may be DONOVAN, HOLLINGWORTH AND VELTKAMP: BRITISH PERMIAN CRINOID 821 that the stem of ‘C.’ ramosus was incapable of autotomy. On the basis of our interpretation of certain structures discussed below, however, it is tentatively proposed that the capability for self mutilation in ‘C’ ramosus was evenly distributed throughout the column, or perhaps just within the dististele. This is problematic in that it requires every columnal to have the ability to adopt the role of terminal ossicle, should autotomy occur at its distal facet. In modern isocrinids such as Neocrinus decorus this only applies to nodal ossicles, which have an axial canal unlike other columnals in the stem (Donovan 1984, p. 836, pi. 74, fig. 3). This canal is constricted by spicules of calcite which presumably grow rapidly after autotomy to seal the distal lumen. Grimmer, et al. (1985, p. 44) have also proposed that the spicules act as anchors for the soft tissues of the axial canal, preventing them from being torn out during autotomy. We propose that the pentastellate jugula in ‘C.’ ramosus (PI. 71, figs. 1 and 2) may have acted in a similar manner to the spicules of isocrinids. The jugulum could have fulfilled an anchoring function for soft tissue while being able to rapidly infill by precipitation of calcite. It is intuitively obvious that a pentastellate canal could become infilled more rapidly than a pentagonal canal of identical area. The frequency of pentastellate jugula in columnals of ‘C.’ ramosus is unknown but is certainly not a feature which is confined to nodals. It is perhaps too delicate a structure for frequent preservation and is probably often obscured by sediment. We postulate that every columnal in the dististele, and possibly also the mesistele, had a pentastellate jugulum. It is recognized that the soft tissues of the axial canal would also need to show the same frequency of ‘segmentation’ as the columnals of the stem in order to make autotomy possible at every columnal-columnal articulation. Bivariate analyses of stem material were made using the collections from Sunderland Museum. These columnals were generally well preserved and come from three quarries which represent the reef core lithology. (The collection of material from Beacon Hill, now in the BMNH, is large but preservation is poorer, with pluricolumnals often broken or partly obscured.) The Sunderland Museum material has also been used in an analysis of pluricolumnal length (text-fig. 8). Grimmer et al. (1985) recognized three ligament types in the isocrinid stem: intercolumnal ligaments at synostosial articulations; intercolumnal ligaments at symplexial articulations, which only insert at the crenulae; and peripheral through-going ligaments, which are limited to the areola and include about a dozen columnals each. By analogy, the latter two ligament types may have been present in ‘C.’ ramosus; it might be possible to determine the length of the peripheral through-going ligaments in this species (if present) by examination of the number of columnals per pluricolumnal, assuming that these long fibres had some influence on the pattern of post-mortem stem disarticulation. Independent support for this hypothesis is not given by the stereom microstructure of the areola, which is at best very poorly preserved. Simple bar graphs of number of specimens against ossicles per pluricolumnal have been plotted for Tunstall Hill, Ford Quarry, and Humbledon Hill, both separately and combined (text-fig. 8). Unfortunately it is not known what form collecting bias takes in these samples, although more single, dissociated columnals might be expected, as well as some longer pluricolumnals. For example, BMNH E70126 (text-fig. 6a, b) has twenty-three columnals. Also, slightly different environmental conditions, rates of burial, and rates of cementation at these localities probably had some effect. Examination of the graphs of Tunstall Hill, Ford Quarry, and the combined localities (Humbledon Hill is ignored as a separate sample because of the small number of specimens) indicates a decrease in pluricolumnal number after the seven to nine columnal region; this may be due to peripheral through-going ligaments reaching such a maximum length. However, some mechanical constraint on pluricolumnal length cannot be discounted. Although the calculated ‘ligament lengths’ correspond approximately to internoditaxes 3231323 and noditaxes N3231323, other seven to nine ossicle pluricolumnals are also common. Epifaunal elements associated with pluricolumnals of ‘C.’ ramosus include the bryozoan Dyscri- telUi sp. (PI. 72, fig. 3) and a possible juvenile attachment structure (PI. 71, fig. 4) but no borings. There is no evidence that these encrustations occurred during the life of the crinoid. Indeed, confirmation of post-mortem encrustation is shown by SuM B2770/68, in which Dyscritella sp. has grown over an articular facet. Some plates on the calyx UCG C26a show signs of possible encrus- tation by epizoans. 822 j Complete pluricolumnals PALAEONTOLOGY, VOLUME 29 : ; Pluricolumnals with attached fragments columnal(s) of TUNSTALL HILL Ossicles/pluricolumnal Ossicles per pluricolumnal FORD QUARRY COMBINED LOCALITIES TEXT-HG. 8. Bar graphs showing the variation in pluricolumnal length, with respect to number of columnals, for specimens of 'Cyathocrinites' ramosus (Schlotheim) in Sunderland Museum. Number of specimens against ossicles per pluricolumnal are plotted for Tunstall Hill (N = 102, mode = 7, mean = 6), Humbledon Hill (N = 21, mode= I , mean = 5), Ford Quarry (N = 50, mode = 2, mean = 4), and combined localities (N = 173, mode = 7, mean = 5). In the absence of a crown, the arm facets of radial plates are of particular importance in deducing the relationship of the arms to the cup. The radial arm facets (text-figs. 5d, 6f) of the British Permian species each have a longitudinal synarthrial articulation ridge which would have permitted an arm to ‘rock’ about an axis parallel to a tangent to the circumference of the cup in the radial position. The axial canal of the arm facet does not lie in the centre of the articular ridge but between the ridge and the aboral margin. This would probably have favoured articulation in the aboral direction (i.e. away from the oral surface). Thus, the feeding orientation of the crown may have been with the arms fanned out. The interpretation of form and function of ‘C.’ ramosus adopted herein has been largely deter- mined by using modern isocrinids as a model. An attempt has been made to restore the possible life attitude of the column of ‘C.’ ramosus (text-fig. 9) by functional interpretation and by comparison with living crinoids (e.g. Macurda and Meyer 1974, 1983). The proxistele, composed of numerous low columnals, was the most flexible part of the stem and would have enabled the crown to change orientation in response to changes in current direction (PI. 72, figs. 1 and 2). Here there are two important factors to recognize. First, the life habit of ‘C.’ ramosus in the high energy environment of a reef strongly implies that it was a rheophilic feeder (although the presence of abundant fenestrate bryozoans possibly suggests that energy levels in the reef environment were not always very high). This is reflected by the distribution of ‘C.’ ramosus both laterally and vertically within different reef facies. Secondly, the crenularia of ‘C.’ ramosus were radial, not petalloid as in iso- crinids, so it was equally flexible in all directions through 360°. This was probably advantageous in making slight adjustments to changes in current direction. The mesistele lacked cirri, at least proximally, and probably functioned mainly to elevate the crown above the substrate, i.e. the height of elevation was approximately proportional to the length of the mesistele. Attachment was by the DONOVAN, HOLLINGWORTH AND VELTKAMP: BRITISH PERMIAN CRINOID 823 TEXT-FIG. 9. Tentative restoration of the stem and crown oi 'Cyathocriuites' ramosu.s (Schlotlieim) in life position, by analogy with modern isocrinids. dististele, which was adapted as a cirriferous runner (compare Rasmussen 1977, fig. 2, with text-fig. 9 herein). Acknowledgements. The authors gratefully acknowledge the help of Dr D. A. T. Harper (University College, Galway) and Mr T. H. Pettigrew (Sunderland Museum) for arranging the loan of specimens. Dr David Southwood helped with some of the field collecting. N.T.J.H. gratefully acknowledges his NERC research studentship. We thank Mr A Reed for drawing text-figs. 1-3. This paper was improved following constructive comments by an anonymous referee. REFERENCES BASSi.ER, R. s. and MOODEY, M. w. 1943. Bibliographic and faunal index of Palaeozoic pelmatozoan echinoderms. Spec. Pap. geol. Soc. Am. 45, 734 pp. BRANSON, c. c. 1948. Bibliographic index of Permian invertebrates. Mem. geol. Soc. Am. 26, 1049 pp. BREiMER, A. 1978. General morphology. Recent crinoids. In moore, r. c. and teichert, c. (eds.). Treatise on Invertebrate Paleontology. Part T. Echinodermata 2(1), T9 T58. Geological Society of America and Univer- sity of Kansas Press, New York and Lawrence, Kansas. DONOVAN, s. K. 1984. Stem morphology of the Recent crinoid Chladocrimts (Neocrinus) decorus. Palaeontologv, 27,825-841. EMSON, R. G. and wilkie, I. c. 1980. Fission and autotomy in echinoderms. A. Rev. Oceanogr. Mar. Biol. 18, 155-250. FRYER, H. c. 1966. Concepts and methods of e.xperimental statistics. Allyn and Bacon, Rockleigh, New Jersey. GEiNiTZ, H. B. 1848. Die Versteinerungen des Zeclisteingehirges and Rothliegenden oder des Permischen Systemes in Sachsen. 26 pp. Arnoldische Buchhandlung, Dresden and Leipzig. 1861. Dyas oder die Zechsteinformation and das Rothliegende, pp. 107-1 24. Leipzig. 1866. Carbonformation und Dyas in Nebraska. Verh. K. Leap. Carol, dt. Akad. Naturforscher, 33, xiiT-91 pp. GRIMMER, j. c., HOLLAND, N. D. and HAYAMi, I. 1985. Fine structure of the stalk of an isocrinid sea lily (Metacrinus rotundu.s) (Echinodermata, Crinoidea). Zoomorph. 105, 39-50. HASZELDiNE, R. s. 1984. Carboniferous North Atlantic palaeogeography: stratigraphic evidence for rifting, not megashear or subduction. Geol. Mag. 121, 443-463. 824 PALAEONTOLOGY, VOLUME 29 HOWSE, R. 1848. A catalogue of the fossils of the Permian system of the counties of Northumberland and Durham, drawn up at the request of the Tyneside Naturalists’ Field Club. Tram. Tyneside Nat. Fid Cl. 1, 219-264. KING, w. 1850. Monograph of the Permian fossils of England. Palaeontogr. Sac. [Monogr.], 258 pp. MACURDA, D. B., JR. and MEYER, D. L. 1974. The feeding posture of modern stalked crinoids (Echinodermata). Nature, Loud. 247, 394-396. 1983. Sea lilies and feather stars. Am. Sclent. 71, 354-365. MILLER, j. s. 1821. A natural history of the Crinoidea or lily-shaped animals, with observation on the genera Asteria, Euryale, Comatida, and Marsupites, 150 pp. Bryan and Co., Bristol. MOORE, R. c. 1939 (for 1938). The use of fragmentary crinoidal remains in stratigraphic paleontology. J. sclent. Labs. Denison Univ. 33, 165-250. 1962. Ray structures of some inadunate crinoids. Paleont. Contr. Univ. Kans., Echinodermata Art. 5, 1-47. — JEFFORDS, R. M. and MILLER, T. H. 1968. Morphological features of crinoid columns. Ibid. 8, 1-30. LANE, N. G. and STRiMPLE, H. L. 1978. Order Cladida Moore and Laudon, 1943. In moore, r. c. and TEICHERT, c. (eds.). Treatise on Invertebrate Paleontology. Part T. Echinodermata 2{2), T578-T759. Geo- logical Society of America and University of Kansas Press, New York and Lawrence, Kansas. and LAUDON, l. r. 1943. Evolution and classification of Palaeozoic crinoids. Spec. Pap. geol. Soc. Am. 46, 1 53 pp. NETSCHAJEW, A. w. 1894. Die Fauna der Permablagerungen des ostiches Teiles des europaischen Russlands. Obshch. Estestovoi. Trudy Soc. Nat. Kazan Mem. 34, 1-44. PATTisoN, J. 1977. Catalogue of the type, figured and cited specimens in the King collection of Permian fossils. Bull. geol. Surv. Gt Br. 62, 33-44. PETTIGREW, T. H. 1980. Geology. In dunn, t. c. (ed.). The Magnesian Limestone of Durham County. Durham County Conservation Trust, 6-26. RASMUSSEN, H. w. 1977. Function and attachment of the stem in Isocrinidae and Pentacrinitidae: review and interpretation. Lethaia, 10, 51-57. RHODES, F. H. t. 1967. Permo-Triassic extinction. In harland, w. b. et al. (eds.). The fossil record, 57-76. Geological Society, London. SCHLOTHEIM, E. F. VON 1816-1817. Beytrage zur Naturgeschichte der Versteinerungen. Geogn. Hinsicht, 13-36. 1820. Die Petrefactenkunde auf ihrem jetzigen Standpunkte durch die Beschreibung seiner Sammiung versteinerter und fossiler Uberreste des Thier-und Pfianzenreichs der Vorwelt erldutert, 437 pp. Beckersche Buchhandhmg, Gotha. SMITH, D. B. 1980. The evolution of the English Zechstein basin. Contr. Sedimentol. 9, 7-34. 1981. The Magnesian Limestone (Upper Permian) reef complex of northeastern England. Spec. Pubis Soc. econ. Paleont. Miner. Tulsa, 30, 187-202. et al. (in press). A revised nomenclature for Upper Permian strata in eastern England. Spec. Rep. geol. Soc. Land. 20. and FRANCIS, e. a. 1967. Geology of the country between Durham and West Hartlepool. Mem. geol. Surv. Gt Br. xiii + 354 pp. SPANDEL, E. 1898. Die Echinodermen des deutschen Zechsteins. Abh. naturhist. Ges. Niirnberg, 1 1, 17-45. TAYLOR, J. c. M. 1984. Late Permian-Zechstein. In glennie, k. w. (ed.). Introduction to the petroleum geology of the North Sea, 236 pp. Blackwell, Oxford. TUCKER, M. E. and hollingworth. n. t. j. (in press). The Upper Permian reef (EZl) of north east England: diagenesis in a marine to evaporitic setting. In schroeder, j. and purser, b. (eds.). Reef diagenesis. Springer Verlag, Berlin. TRECHMANN, c. T. 1913. Oil a mass of anhydrite in the Magnesian Limestone of Hartlepool and on the Permian of south-east Durham. Q. J. geol. Soc. 69, 184-218. 1943. On some new Permian fossils from the Magnesian Limestone near Sunderland. Ibid. 100, 333- 354. UBAGHS, G. 1978. Skeletal morphology of fossil crinoids. In moore, r. c. and teichert, c. (eds.). Treatise on Invertebrate Paleontology. Part T. Echinodermata 2(1), T58-T216. Geological Society of America and University of Kansas Press, New York and Lawrence, Kansas. wachsmuth, c. and springer, f. 1885. Revision of the Palaeocrinoidea, pt. Ill, sec. 1. Discussion of the classification and relations of the Brachiate crinoids, and conclusion of the generic descriptions. Proc. Acad, nat. Sci. Philad. 1885, 225-364. DONOVAN, HOLLINGWORTH AND VELTKAMP: BRITISH PERMIAN CRINOID 825 WALKER, K. R. and ALBERSTADT, L. p. 1975. Ecologic successioii as an aspect of structure in fossil communities. Paleohiol. 1, 238-257. WEBSTER, G. D. 1974. Crinoid pluricolumnal noditaxis patterns. J. Paleont. 48, 1283-1288. WRIGHT, j. 1950. A monograph of the British Carboniferous Crinoidea. Palaeoritogr. Soc. [Monogr.], 1(1), XXX + 24 pp. STEPHEN K. DONOVAN Department of Geology University of the West Indies Mona, Kingston 7 Jamaica NEVILLE T. J. HOLLINGWORTH Department of Geology University of Durham South Road Durham DHl 3LE Typescript received 18 November 1985 Revised typescript received 26 Eebruary 1986 CORNELIS J. VELTKAMP Department of Botany University of Liverpool PO Box 147 Liverpool L69 3BX 1 I f ll r I i* __ S iv.# EVOLUTION OF THE EARLIEST SMOOTH SPIRE-BEARING ATRYPOIDS (BRACHIOPOD A: LISSATRYPIDAE, ORDOVICIAN -SI LU RIAN) by PAUL COPPER Abstract. Calcified lophophore supports present in the oldest of the spire-bearing brachiopods, the Lissatrypi- dae which range in age from middle Ordovician (Caradoc) through middle Silurian (Wenlock) time, demon- strate complex evolutionary patterns and early divergence. The smooth-shelled brachiopods Protozyga, Idiospira, and Cyclospira, which had medially or dorso-medially directed spiralia, evolved from the primitive early Caradoc atrypoid Manespira n. gen., which had a spiralium of less than one whorl and a whole or partial jugum. Early divergence produced four separate lineages, the Protozyginae n. subfam., Septatrypinae, Cyclospirinae, and Zygospiridae (which led to the ribbed atrypoids). Each of these groups developed distinctive brachidia, probably a reflection of early experimentation in filter-feeding strategies. In late Ordovician and early Silurian time, the ribbed Atrypidae, which had evolved from the Zygospiridae, expanded rapidly to dominate the shallow benthos of tropical seas: they accomplished this by perfecting large shell sizes, increasing the number of whorls in the spiralia, orienting the spiralia dorsally and separating the jugum into discrete processes. The smooth-shelled atrypoids, Lissatrypidae, usually played a secondary role, though they were prominent in the Silurian. Two new developments took place in the Silurian: evolution of the thicker-shelled Lissatrypmae (by mid-Llandovery), and the unique Glassiinae, with medially directed spiralia (by Wenlock time). Both of these declined during the Devonian, but the Glassiinae were the last survivors of the smooth atrypoids in late Devonian (Frasnian) time, with a new genus, Peratos (type species P. arrectus n. sp.), ending the lineage. Some one hundred years ago Thomas Davidson (188k/) suggested that there were only four spire- bearing brachiopod families: these were based on the common genera Spirifer, Athyris, Nucleospira, and Atrypa. Except for the Nucleospiridae, which are now submerged within the family Athyridae, Davidson’s categories have stood the century test of time very well, even though the families now have higher taxonomic rank. The Treatise (Boucot el al. 1965) has broadly treated the spire-bearing brachiopods as a monophyletic stock under the order Spiriferida. With the discovery of dorsally directed spiralia in the ‘orthid’ Tuvaella by Vladimirskaya (1972), in the ’strophomenid’ Davidsonia by Garcia-Alcalde (1973) and Copper (1978), and in the ‘rhynchonellid’ Hircinisca by Havlicek and Plodowski (1974), the Tuvaellidae and at least some Davidsoniidae and ‘rhynchonellids’ may be further added to the spire-bearers. Rzhonsnitskaya (1960), Rudwick (1970), and more recently Wright (1979), have supported a diphyletic view of the spire-bearers, with the Spiriferidae derived from an orthid stock during the late Ashgill or early Llandovery and the Atrypida (including, for these authors, the Atrypa and Athyris groups) evolving from rhynchonellids during Llandeilo time. This paper is an attempt to clarify the basic internal structures, particularly the brachidia (crura, jugum, and spiralia) which supported the lophophore, of the earliest smooth spire-bearing brachio- pods. It is a follow-up of an earlier paper (Copper 1977) which dealt with the oldest ribbed spire- bearers, the Zygospiridae. The information on which this paper is based is the re-examination of all types and topotypes of the known Ordovician and most of the Silurian type species of smooth atrypoid spire-bearing genera, i.e. those with medially or dorsally directed spiralia. Reconstructions of the internal morphology are based on the technique of serial sectioning using rapid-drying 0-004 mm thick, Acetobutyratfolie peels (made by Bayer G. m. b. H., Leverkusen, Germany). These peels were mounted under glass IPalaeontologv, Vol. 29, Part 4, 1986, pp. 827 866, pis. 73-75.) 828 PALAEONTOLOGY, VOLUME 29 and projected with standard 35 mm slide projectors on white paper, and then copied on tracing paper. Peel details are of high resolution and can be photographed under a microscope. The traced images were then measured into the plane parallel to the shell commissure (or at right angles to the angle of sectioning) using the plane of bilateral symmetry as a guide. A ventral three-dimensional view of the brachial valve and the brachidia was then reconstructed by tracking points from the section plane into the commissural plane (technique developed by Copper 1967). Lateral view reconstructions were also made, using the curve of the plane of symmetry of the brachial valve as the base line (using plaster casts of the sectioned specimens), and normally drawing only the right spiralium as viewed into the dorsal valve. Such reconstructions are accurate to 0-5 mm at a magnification of x 12, x 16, or x 20. STRATIGRAPHIC FIRST APPEARANCES Cooper (1956, 1976) suggested that the earliest spire-bearing species (still un-named) occurs in the Crown Point Limestone, which has a Conodont Fauna (CF) 6 (Pygochis anserinus to P. serrus. Sweet and Bergstrom 1976), graptolite Zone 10 (late Glyptograptus cf. G. teretiuscuhis. Berry 1976), or an Ashby, pre-Blackriver brachiopod correlation (Cooper 1976, p. 177). In the British succession the Crown Point would probably be of upper Lower Llandeilo age (text-fig. 1). Unfortunately the internal structure in this un-named and undescribed species is unknown, and material is so scarce that the few specimens in question may well be camerellids or smooth rhynchonellids, from which they are difficult to distinguish (see Raymond 1911). Cooper (1956) reported ribbed zygospirids ( = Auazyga) and smooth spire-bearers ( = Manespira this paper) from CF7 formations such as the Lebanon of Tennessee and Little Oak of Alabama. Thus by very early Caradoc time (Costonian), left, ribbed on the right. The ‘x’ in the lineage marks the occurrence of the type species of the genus. The ranges of conodont and graptolite zones are shown for comparison. COPPER: EVOLUTION OF THE LISSATRYPIDAE 829 at least two stocks co-existed, both very small shells less than 5 mm wide, one including Anazyga, the other Manespira. By Blackriver time (Soudleyan, Conodont Fauna 8: Sweet and Bergstrom 1976), in the Napanee Limestone of eastern Ontario and north-eastern New York, the spire- bearers Protozyga exigua, Idiospira panderi, and Anazyga cf. A. recurvirostra co-occur in the same stratigraphic horizons though not necessarily at the same localities or in the same bed (text-fig. 1). This strongly suggests that, if North America represents the centre of early spire-bearer evolution (as appears likely at present), first diversification of the spire-bearers occurred by or during CF7 (Harnagian) time. Glassiinae. WENLOCK LLANDOVERY ASHGILL CARADOC Septatrypinae Lissatrypinae Gjgssig Idiospira Manespira Protozyga Protozyginae TEXT-FIG. 2. Evolutionary trends in the brachidia of the smooth atrypoid spire-bearers in the Ordovician- Silurian. Note the modifications of the spiralium orientation and jugum with time. By CF8 time (Harnagian-Soudleyan), Protozyga had evolved from Manespira by extending the simple whorls, expanding the jugum, and widening the spiralial blade (see text-figs. 2, 5). The ribbed Anazyga continued, not giving rise to true Zygospira and Catazyga until the late Caradoc (Onnian). Lastly, Idiospira showed the most rapid evolution: a complex spiralium with many whorls and a jugum which had moved posteriorly as early as CF8 time. Cyclospira also appeared in late Caradoc time (CFIO, Actonian-Onnian), possibly from Manespira by complete loss of a jugum, or even possibly independently by never developing a jugum at all, since it may not be surprising to find Manespira without the jugum. Earlier so-called Cyclospira reported from ‘Wilderness’ or Porterfield age rocks by Cooper (1956) have here been referred to Protozyga or Manespira, since typical bisulcate morphology is more common in Ashgill time. In Ontario and New York Cyclospira appears suddenly in the late Trenton (Shermanian, Actonian) and disappears equally quickly above that horizon. At the same time as Cyclospira in Ontario there are first appearances of internally more complex genera like Zygospira (e.g. Z. raymondi Foerste) and Catazyga (e.g. C. ftlistriata Sproule). CFIO equivalent sedimentary rocks of the interior of eastern North America therefore 830 PALAEONTOLOGY, VOLUME 29 mark a renewed phase of spire-bearer innovation, with at least four co-occurring genera of smooth and ribbed atrypoids. The last cycle of Ordovician spire-bearer evolution is marked in the Gamachian Ellis Bay Forma- tion of Anticosti (CF13 = Hirnantian, and possibly partly Rawtheyan) with the sudden appearance of abundant Eospirigerina and Hindella, precursors of the Silurian Atrypidae and Meristellidae respectively. These taxa probably represent immigrants from the north-west USSR or central Asia, moving eastwards across the Baltic and USSR (Jaanusson 1979). Corroboration for this may lie with the discovery of possibly the oldest, late Caradoc, athyroids {Weibeia and Apheathyris, with laterally directed spiralia), recently described from North China by Fu (1982). Eospirigerina makes a first appearance in Eurasia in the Dicellograpim complanatus graptolite or low Amor phognat bus ordovicicns CF12 zone, in rocks of probably early to mid-Ashgill (Pusgillian-Cautleyan) age. This immigration may account for the lack of ‘transitional’ North American species leading to Eospiri- gerina and Hindella in the late Ordovician of eastern North America. Thus, although the subtropical to tropical carbonate shelf environments of North America appear to have been the ancestral ‘home’ of the atrypoids in Flandeilo to Caradoc time, the more ‘advanced’ Silurian Atrypidae which replaced them seem to have originated elsewhere. Finally there are the Silurian first appearances of new stocks of smooth atrypoids. The first of these was the appearance of the Fissatrypinae in late Rhuddanian (A3) time: this group was characterized by generally thicker shells, separated jugal processes and solid teeth without dental cavities. A second development was the introduction in late Flandovery or early Wenlock time of the subfamily Glassiinae, a unique group in which the spiralia were directed towards the shell centre (earlier Glassia have been recorded but not confirmed). This spiralium orientation seems to have evolved by neoteny. Quite independently, the Spirifer group, sensu restricto, made its first appearance in middle Llandovery (Idwian) time. Three genera appear almost synchronously in China: Yingwuspirifer, Eospirifer, and Striispirifer (Rong and Yang 1978). On Anticosti Island, Eospirifer makes its first appearance in the upper 50 m of the Jupiter Fm. of Telychian (C4-5) age. In north-west Europe, the athyroid brachiopod group, the Dayiinae, with very simple juga and spiralia, appeared in Ludlow time. PHYTOGENY AND ONTOGENY OF THE SPIRALIA The atrypoid brachiopods are typified by having spiralia that are medially, dorsomedially or dorsally directed. Spiralia are attached to the dorsal shell hinge by means of crura. The athyroids and spiriferoids, on the other hand, had laterally or ventrally directed spiralia, i.e. ‘inside-out’ opposites of the atrypoid spiralia. Manespira n. gen., the oldest spire-bearer known, had a spiralium which consisted of less than one complete whorl coiled in the plane of symmetry, with the tip of the spiralium twisted in a medial direction. The spiralium here is essentially an elongated crus. The jugum may be complete or incomplete (text-figs. 2, 6). Usually the rhynchonellids have been identified as the ancestral group (e.g. Rudwick 1960), by virtue of their earlier appearance in Llanvirn time, shell similarities, hinge, and crura. An alternate ancestral possibility is the Llanvirn or Llandeilo pentamerids, e.g. camerellids, by loss of the cruralium, elongation of the crura and development of the delthyrium. The camerellids possess a smooth shell: in fact, old museum collections sometimes list early atrypoids like Idiopira and Protozyga under the name Camerella, accentuating the morphologic similarity. Early work on the ontogenetic development of spiralia, within species such as Anazyga recurvi- rostra (Hall) by Beecher and Schuchert (1893), had great influence in suggesting phyletic affinities of the spire and loop-bearing brachiopods. They tried to demonstrate that zygospirid brachidia were initiated as anteriorly fused, bladed crura which gradually diverged but remained connected by development of a jugum between them. The early growth pattern is at least superficially similar to loop construction in the terebratulids. The Beecher-Schuchert model was taken over by Williams and Wright (1961) and Williams and Hurst (1977), who speculated that a ‘centronelliform loop’ COPPER: EVOLUTION OF THE LISSATRYPIDAE 831 was developed by the atrypoid Manespira elongata (Cooper) from the Bromide Fm. (Caradoc) of Oklahoma. However, the terebratulid loop appears to have been derived entirely independently during the early Devonian since no post-Caradoc Protozyga or Manespira have ever been described, leaving a time gap of more than 60 million years between these earliest atrypoids and the oldest terebratulids. There is no direct evidence to link the development of the spiralia and jugum in atrypoids with the development of the loop in terebratulids. The jugum as a single, complete structure was abandoned and replaced by separated jugal processes in the Silurian atrypoids, but its retention in all the athyroid and spiriferoids stocks suggests that if the loop developed from the jugum, a possible source of terebratulids lies with smooth athyroids. But, the terebratulids could also have evolved from pentamerids, which lacked a calcified lophophore support. Samtleben ( 1 972), by examining detailed micro-structure of the spiralia and loops using scanning electron microscopy, has convincingly demonstrated that the jugum of spire-bearers is unrelated to the terebratulid loop. Beecher and Schuchert (1893) were also the first to suggest that there was a fundamental distinc- tion between the atrypoids, ‘with spirals between the first descending branches of the lamellae’, and the Spiriferidae, Nucleospiridae, and Athyridae, where ‘the primary lamellae are between the spirals’. In the atrypoids the crura sweep laterally, changing to the lamellae of the spiralia, with the spiralia coiled towards the centre or dorso-medially between them. In the athyroids and spiriferoids the crura stay close together in the shell middle, changing into spiralial lamellae which are then directed laterally or latero-ventrally. In several early athyroids the spiralia are not connected by solid calcite ribbons to the spiralia, but by interlocking ‘hooks’. The Ordovician smooth (and ribbed) atrypoids appear to have had continuous calcite growth from crura to spiralia, though the junction between the two is sharply geniculate (text-figs. 2, 4, 6). The role of the jugum is not clear in the atrypoids. The oldest spire-bearer, Manespira, lacked a complete jugum in its adult stages. In Protozyga and Anazyga, its descendants, and in other Ordovician atrypoids (except Eospirigerina) the jugum was a complete, single piece, either U-shaped or W-shaped. In all atrypoids, the jugum seems to have evolved its morphology independent of the main trends in the evolution of the spiralia. For example, the conversion of a one-piece jugum to two jugal processes occurred in several unrelated lineages. Another factor in jugum evolution is the change with time in disposition of the jugum. Older Ordovician genera show a jugum in an anterior position (e.g. the Protozyginae). Later smooth spire-bearers have a posterior position for the jugum or jugal processes (text-fig. 2). Moreover, the general position shifted as well from a dorsal location to a more ventral location in Silurian taxa (text-fig. 3). It seems possible therefore that the jugal complex played no direct lophophore-supporting role, unlike the terebratulid loop. Instead it may have served to support the mouth parts (its central location in the brachial valve supports this idea), or to have acted as a holdfast for muscles to manipulate or support the spiralia (many jugal processes were provided by spines and these may have served to hold the spiralia in place). In all the main atrypoid lineages, whether ribbed or smooth stocks, there were parallel trends in the development of the brachidia (jugum and spiralia). These were as follows: 1. Increasing the amounts of internal shell space occupied by the spiralia, and increasing numbers of revolutions of the spiralia. By Silurian and certainly Devonian time (for which shells wider than 60 mm are known), spiralia came to occupy as much as 80 to 90 % of the internal shell cavity. 2. Rotation of the spiralial cones from medial to dorso-medial to nearly dorsal orientation (with the exception of the Cyclospirinae and Glassiinae which adopted a medial orientation). This dorsal orientation allowed the apices of the spiralial cones to grow as large as dorsal valve convexity permitted. 3. Relocation of the jugum from a dorso-anterior position to ventro-posterior position (the Cyclospirinae having no jugum). The position of the jugal complex was usually at or near the apices of the spiralia and this shift may thus have reflected exhalant water canalization. 4. Conversion of a one-piece jugum to separated jugal processes for most atrypoids by Silurian time (with the exception of the Zygospiridae, which were ‘living fossils’ in the early Silurian, retaining a jugum). This may have avoided breakage of the jugum in higher-energy conditions. 832 PALAEONTOLOGY, VOLUME 29 TAXONOMY The taxonomy of the smooth spire-bearers is far from being settled in a completely satisfactory manner. This is particularly true for the Atrypida, with medially or dorsally directed spiralia, and the Athyrida, with laterally directed spiralia (as defined herein). The nature and direction of the spiralia and the location of the jugal complex and nature of the crura to which these were attached, must have been critical to the filter-feeding process. Ancillary shell morphology, i.e. attachment of the shell with or without a pedicle, shell shape, ribbing, and anterior fold, was also related directly or indirectly to filter-feeding efficiency. These features are thereby of prime taxonomic importance in these groups. Rzhonsnitskaya (1960) separated the Atrypa and Athyris groups respectively under a new order Atrypida, and the Athyracea under incertae ordinis, and later she suggested that these two groups were monophyletic (Rzhonsnitskaya 1964). Rzhonsnitskaya also assigned the Coelospiracea and Dayiacea to the Atrypida, a practice followed by Boucot et al. (1965) and in nearly all modern literature. It can be demonstrated that both the Coelospira group (including Bifida and Kayseria) and the Dayia group have spiralia and juga like normal athyroids and should thus be assigned to the Athyrida (Copper 1973). Another vexing question is the relationship of the true spiriferoids to the atrypoids and athyroids. Boucot et al. (1965) and Williams and Hurst (1977) have defended the conventional view, first promoted by Davidson in 1881, that all the spire-bearers are related and hence can be grouped under one order, the Spiriferida. Rzhonsnitskaya (1960), Rudwick (1970) and more recently Wright (1979) have advocated a bipartite or tripartite division into separate orders, the Atrypida and Spiriferida (the former including the Athyrida). Wright (1979) specifically selected the orthid sub- family Platystrophiinae, as the probable ancestral line of the true spiriferids, the transition occurring in late Ashgill or early Llandovery time. The requisite parallel conclusion is, of course, that spiralia developed independently in both the athyrids and true spiriferids. Wright argued this point of view on two grounds: the similar strophic hinge and shell microsculpture. Rong and Yang (1978) have shown that Yingwuspirifer appears just before Eospirifer in the middle Llandovery. The micro- sculpture of these two oldest spirifers is very finely striated or micro-costellate, so fine that the shells often appear to be smooth. They apparently lack a micropustulose ornament and coarse plications and wide shelf-like area of Platystrophia, though Wright (1979) stated this was present in some Eospirifer. It is primarily the later spirifers that begin to resemble Platystrophia. The problem of ancestry therefore still seems unresolved, yet I would agree with Wright that the Eospirifer group, the oldest spirifers, were probably derived independently from the atrypoids and athyroids. The definition of the order Atrypida used in this paper specifically excludes the Coelospiracea and Dayiacea. The Athyrida are treated as a distinct order characterized by laterally directed single or double spiralia, which are normally derived from centrally located, anteriorly directed crura, and a simple to highly complex, one-piece jugum enclosed within a non-strophie shell with narrow hinge. Genus level and subfamily level criteria in the Atrypida are considered to be the nature of the shell surface (ribbed, smooth, etc.), convexity, flexure of the anterior commissure, the beak area (foramen, interarea), and hinge structures (teeth, socket plates). Muscle attachment scars yield ambiguous information at present. Muscle scars, vascular canals and gonadal pits are very faint or absent in all early atrypoids. SYSTEMATIC PALAEONTOLOGY Order atrypida Rzhonsnitskaya, I960 (emend.) Diagnosis. Smooth or ribbed brachiopods, usually with small interareas and beaks, primarily non- strophic shells, medially to dorsally directed spiralia located between divergent crura and primary lamellae; jugum or jugal processes dorsal to ventral, some lacking jugal complex. COPPER: EVOLUTION OF THE LISSATRYPI DAE 833 Family lissatrypidae Twenhofel, 1914 Diagnosis. Smooth atrypoid shells with medially, dorso-medially, or dorsally directed spiralia, jugum or jugal processes or without a jugum. Subfamilies included. Protozyginae n.subfam., Lissatrypinae Twenhofel, 1914, Glassiinae Schuchert and Levene, 1929, Septatrypinae Kozlowski, 1929, Cyclospirinae Schuchert, 1913. Names discarded as synonymous. Subfamily Atrypellinae Poulsen, 1943 (= Lissatrypinae), Atrypopsinae Poulsen, 1943 (= Septatrypinae), Aulidospirinae Williams, 1962(= Cyclospirinae). Remarks. The late Silurian genus Aiistralina Clarke, 1913 from Argentina is not a senior synonym of the North American genus Lissatrypa Twenhofel, 1914 (Copper, Fliinicken and Benedetto, in preparation), as suggested by Strusz (1982a, b). Australina differs externally in its planoconvexity, having a strongly inflated pedicle valve and concave to flattened brachial valve. Lissatrypa has a biconvex to weakly dorsibiconvex shell, as well as having distinctive internal pedicle constrictions and cardinalia. An approximate Ordovician-Silurian phylogeny is given in text-figs. 2, 3, 18. TEXT-HG. 3. Nature of the jugum and spiralia in the Silurian Lissatrypidae subfamilies. Representative serial sections shown were taken at maximum spiralium size. The ends of the jugal processes, i.e. usually the jugal plates, are superimposed on these serial sections although these normally do not extend to mid-shell. Note that all Silurian smooth atrypoids have separated jugal processes but for one un-named new genus from Anticosti. All specimens to scale of x 2. 834 PALAEONTOLOGY, VOLUME 29 Subfamily protozyginae n. subfam. Diagnosis. Small, smooth or anteriorly corrugated, planoconvex to biconvex, sulcate to rectimargi- nate shell with simple, medially or vertically oriented spiralia normally of two whorls or less, and a simple, whole or partially developed anterior or central-dorsal jugum. Genera assigned. Prolozyga Hall and Clarke, 1893, Manespira n. gen. Range. Llandeilo-Caradoc. Distribution. North America, north-west Europe, south-east Asia, Remarks. This subfamily includes the oldest known spire-bearing genera and is distinguished primarily by its brachidial simplicity: a very simple coil or partial spiralial whorl and a partly or fully developed anterior to dorsal jugum. A small, smooth shell (usually less than 5 mm wide), with a straight or sulcate commissure (the opposite of later atrypoids which are usually plicate), and a small, pointed beak with minute foramen characterizes external morphology. The cyclospirinids also have medially directed spiralia with few coils, but they lack a jugum and appear to have a thick ventral valve, stronger dentition, and a relatively long dorsal median septum. Genus protozyga Hall and Clarke, 1893 Type speeies. Atrypa e.xigua Hall, 1847, p. 141, pi. 33, fig. 6a-d. CF8: Caradoc, late Harnagian-early Soudleyan. Range. Blackriveran-Trentonian (early-middle Caradoc), ?Ashgill. Distribution. North America, Europe, south-east Asia. Diagnosis. Normally smooth, planoconvex-ventribiconvex, sulcate shells with small incurved beak, minute pedicle opening and deltidial plates, medially pointed or planispiral spiralia of one or more dorso-ventrally oriented coils, giving rise anteriorly to simple, broad jugum directed posteriorly. Thin shells with dental cavities and simple hinge structure, very weakly defined muscle scars, weak dorsal septum. Remarks. Since the whorls are nearly planispiral, it is difficult to determine in which direction a single coil may be oriented, except that in the type, P. exigua (text-fig. 4), the terminal part of the whorls is located towards the centre in relation to the ascending lamella. This means that for the type species the spiralia are medially directed. It is not impossible, however, that other related species may be planispiral, the difference in direction being only a matter of a millimeter. This suggests that the Protozyginae may also have been potential ancestors of the Athyrida, which are typified by laterally directed spiralia. The oldest athyroids appear, as presently known, in late Caradoc time (Fu 1982). There is a substantial time gap between the oldest Athyrida and youngest protozyginids in the stratigraphic sections of North America and western Europe, suggesting important easterly migrations in Ashgill time. Protozyga differs from Manespira in a bigger shell, in usually having a smooth, even shell surface (though this may be present in some Manespira), and in lacking corrugations on the commissure. Protozyga has a more complex spiralium of one or more revolutions (maximum observed was two), and a complete one-piece jugum. Most of the species of Protozyga have been described by Cooper (1956) from Caradoc rocks of the eastern US (Oklahoma to Virginia). For these species the spiralia are unfigured except for one. If Cooper’s species are distinctive enough to be recognized, it implies great variability in the group, all species of which fall within two conodont zones. Four of Cooper’s species of Cyclospira have been tentatively assigned to Protozyga until their spiralia can be verified. Species assigned. 1 Protozyga haydeni Reed, 1936, p. 51, pi. 4, fig. 12, ‘Hill behind Taunggyi’, southern Shan states, Burma; Caradoc. A single cast of a brachial valve is the only specimen available. Affinities doubtful. COPPER: EVOLUTION OF THE LISSATR YPIDAE 835 inward spiralium TEXT-FIG. 4. Brachidia of two specimens of Protozyga exigua (Hall, 1847) based on text-fig. 5. Note the broad ascending (primary) lamellae, centro-dorsal jugum, and simple spiralium (lateral view shows only one spiralium). Scale x 8. 1 Protozyga jingheensis Fu, 1982, p. 160, pi. 42, hg. 1 \a-d. Dongzhuang, Liquan, Shaanxi, China; Jinghe Fm., Caradoc. This species has a fold in the dorsal sulcus. Internal structure unknown. Protozyga loeblichi Cooper, 1956, pp. 679-680, pi. 140c, figs. 17-21. Bromide Fm., Pooleville Mbr (CF 7), Oklahoma; Caradoc. Species with one or two plications on the commissure. Brachidia unknown. Protozyga microscopica Cooper, 1956, p. 681, pi. 141d, figs. 22-24; pi. 14If, figs. 33-37. Lincolnshire Fm., Hogskin Mbr, Tennessee; Caradoc. Brachidia unknown. Protozyga nasiita Cooper, 1956, pp. 681 682, pi. 14Ig, figs. 38-41. Whistle Creek Fm., Virginia; Caradoc. According to Cooper, below the Lincolnshire Fm. Brachidia unknown. Cyclospira parva Cooper, 1956, p. 694, pi. 142k, figs. 49-53. Bromide Fm., Pooleville Mbr, Oklahoma; Caradoc. Cooper reported that the spiralium extended well beyond the shell centre, but the absence of a jugum, as typical of Cyclospira, was not mentioned. Cyclospira preciosa Cooper, 1956, pp. 694-695, pi. I41e, figs. 25-32. Edinburgh Fm., Virginia; Caradoc. Cooper mentions ‘descending processes of spire reaching nearly to front margin’, suggesting Protozyga. Cyclospira quadrata Cooper, 1956, pp. 695-696, pi. 141a, figs. 1-12; pi. 143b, figs. 7-11. Edinburgh Fm., Virginia; Caradoc. Externally similar to Protozyga exigua, but brachidia unknown. Protozyga rotunda Cooper, 1956, pp. 683-685, pi. 140b, figs. 10 16; pi. 140m, figs. 48-52; pi. 140i, figs. 53-58. Warden Em., Tennessee; Caradoc. It has one or two obscure plications on the flanks and may be conspecific with P. loeblichi. Brachidia with ‘a long descending branch . . . Jugum complete’ (Cooper, 1956, p. 684). Williams (1962) stated that he identified this species in Llandeilo rocks of Girvan, UK. Protozyga rotundiforinis Cooper, 1956, p. 685, pi. 140j, figs. 59-62. Lincolnshire Fm., Tennessee; Caradoc. Has one or two plications; brachidia unknown. This species may belong to Manespira and/or be a variant of P. microscopica. 836 PALAEONTOLOGY, VOLUME 29 Cyclospira sulcata Cooper, 1956, p. 696, pi. 142m, figs. 60-69. Sevier Fm., Tennessee; Caradoc. The smooth shell suggests an elongate P. exigua. Brachidia unknown. Protozyga tianzuensis Fu, 1982, p. 160, pi. 42, fig. \0a-d. Hengliang Mtn, Tianzhu, Gansu; Goulang Fm., Caradoc. Brachidia unknown. Species deleted. Protozyga gastrodesTempls, 1968, referred to Cyclospira. Protozyga muscidosa Lockley, 1980, p. 218, figs. 70-76. Bed TB19, Nod Glas, Nant Tan y Blwch, Wales; Caradoc. Illustrations indicate dental plates nearly half the shell length. This is not known from any atrypoid. Affinities questionable. No spiralia described. Protozyga perplexa Williams, 1962, referred to Cyclospira. Protozyga carrickeiisis Reed, 1917, referred to Idiospira. Protozyga profunda Cooper, 1956, p. 683, pi. 143a, figs. 1-6. Trenton Fm., St Francis de la Salle Quarry, Montreal. Examination of topotype material sent by Dr. G. A. Cooper shows that this species is finely ribbed and assignable to Anazyga, a zygospirid. ‘IProtozyga ohsoleta Foerste, 1914, p. 133, pi. 2, fig. lOa-6, 'lower part of the Millersburg Member, Cynthiana Fm.’, CFIO, Ashgill. Internal structures were not examined by Foerste, who, however, noted that the shell had a shape and ribs similar to Zygospira, where it should probably be assigned. Protozyga exigua HaW, 1847 PI. 73, figs. 1 -5; text-figs 4, 5 1847 Atrypa exigua Hall, p. 141, pi. 33, fig. 6a~d. 1893 Protozyga exigua Hall; Hall and Clarke, p. 149, figs. 137-138. 1956 Protozyga exigua Hall; Cooper, pp. 678-679, pi. 1 I9d, figs. 9-14; pi. 140f, figs. 38-42; pi. 141g, figs. 29-33. Type locality. 'Lowville and near Martinsburg, Lewis County’, New York (Hall 1847). The exact locality is unknown, and is not indicated with the type materials (Cooper 1956, pp. 675-6 and personal examination). No new localities at which this species is abundant could be found in New York. Type horizon. 'In the central part of the Trenton Limestone’ (Hall 1847, p. 141). Titus (pers. comm.) says that Hall was almost certainly mistaken here, and that he and colleagues have found it only in the lower Trenton, specifically most abundantly in the Napanee Limestone (source of the serially sectioned specimen: text-figs. 4-5), but as low as the underlying Selby Lst. Both of these limestones belong in the CL8 zone. Type specimens. A lectotype, AMNH 714-la, was selected by Cooper (1956, pi. 142g, figs. 29-33), from five syntypes in the Hall collection. EXPLANATtON OF PLATE 73 Ligs. 1-5. Protozyga exigua (Hall, 1847). Lectotype AMNH714a, imperfect specimen with faintly corrugated lateral commissure. 'Lowville, New York’; probably Napanee Limestone, CL8, Soudleyan, Caradoc, x 5. Figs. 6-20. Manespira nicolleti (Winchell and Schuchert, 1892). East of Chatfield, Minnesota; McGregor Mbr, Platteville Lm., CF7 or CF8, Harnagian-Soudleyan, Caradoc, x 5. Relatively variably corrugated species, one of the youngest survivors in the genus. 6-10, Neotype GS59146, a large adult shell, nearly smooth in early growth stages (small carbonate fragment adhering to ventral umbo); 11-15, medium-sized shell GS59I47; 16-20, small shell GS59148. Figs. 21-35. Idiospira panderi (Billings, 1859). All material from 'Paquette Rapids’ type locality, east Ontario (Ottawa Valley); high Napanee or Kings Falls Limestone, late CL8 or early CL9, Soudleyan, Caradoc, x 3. Specimens are partly or wholly silicified. 21-25, lectotype GSI149c, partly damaged specimen, one of six syntypes from Billings collection. 26-30, paralectotype GSll49b, a more elongate variety showing strong 'beeckite’ silicification. 31-35, hypotype ROM23918a, whole specimen, nearly identical to the Billings lectotype to show the distal development of the shell corrugation and the uniplicate fold. PLATE 73 COPPER, Ordovician atrypoids 838 PALAEONTOLOGY, VOLUME 29 TEXT-FIG. 5. Serial sections of two specimens of Protozyga exigua (Hall, 1847). Napanee Limestone, near Napanee, Ontario, NTS 31C/7W 44600:02930; Caradoc, Harnagian-Soudleyan. Note that spiralium connec- tion occurs near the anterior commissure. Scale x 4. Diagnosis. Small, smooth, with faint trace of ribs at commissure, ventribiconvex to planoconvex, 3 to 5 mm wide, about as wide as long, subquadrate to somewhat pentagonal shell with narrow apical angle (100°-1 10°) and sulcate margin, anteriorly rounded. Beak anacline to hypercline. Proportionally large, wide dental cavities and subhorizontally directed teeth, no pediele layers in umbonal cavity. Socket plates thin, but long, crural bases and inner socket ridges relatively promi- nent, bulbous; crura directed as dorso-ventral strong blades, broadening and curving laterally to form prominent ascending lamellae and terminating in a narrowing spiral band of one to two revolutions. Jugum W-shaped with broad central-ventral arch; jugum may be curved posteriorly (text-fig. 5). Weak dorsal septum divides adductors. Remarks. There is no fresh information on the nature of the spiralia and jugum in other species of Protozyga. The internal details of Protozyga elongata Cooper, 1956, illustrated by Williams and Wright (1961, text-fig. 4) indicate that the species elongata should be referred to Manespira on the basis of its primitive brachidia. Hall and Clarke (1893, p. 149) illustrated the primitive nature of the brachidia of Protozyga for the first time, and their illustration is essentially correct except in showing a very narrow, instead of broad and wide ascending lamella and a thin anterior jugum. This may be infraspecific variation. Weller (1903, pi. 10, figs. 27-30) illustrated a Protozyga from New Jersey which is similar to P. exigua but more elongated and with a narrower hinge angle. Titus and Cameron (1976) indicate that Protozyga exigua occurs in the Liospira and Triplesia communities that occupied a earbonate shoal to lagoonal marine facies. This was a deeper, quieter, high diversity environment ranging to shallower, higher energy zones. Protozyga is apparently nowhere abundant in New York or Ontario, representing only a very minor fraction of the com- munity. The typical lithology is a medium to thickly bedded micrite with shale partings, minor broken shell layers and occasional Solenopora algal balls. Materials. Protozyga exigua is not uncommon in the Napanee Limestone near Napanee. In New York it occurs near Lowville, Sugar River, and Port Leyden (Cooper 1956), but is rare. It characterizes the upper COPPER: EVOLUTION OF THE L I SS AT R Y PI D A E 839 Rocklandian Napanee Limestone (CF8, approximately Harnagian-Soudleyan), and is associated with two other spire-bearers, Anazyga and Idiospira. Sectioned comparative material came from a Highway 401 roadcnt about 4 km north-east of Napanee (NTS 3IC/7W 44600:02930), marked on the map as Gull River Fm. by Liberty (1971), but belonging to the Napanee Limestone, which is assigned by Liberty to the Bobcaygeon Fm. Protozyga has not been found in the northern Michigan Basin (Manitoulin) nor in Quebec. MANESPiRA n. gen. Name. Mane, Latin, dawn or morning, and spira, spire or coil. Type species. Ilallina nicolleli Winchell and Schuchert 1892, p. 293, and 1893, p. 474, pi. 34, hgs. 50-62. Platteville Fm., McGregor Mbr (CF8: Sweet and Bergstrom 1976); Minnesota. Range. Late Llandeilo-middle Caradoc (middle Soudleyan). Cooper (1956, 1976) reports the earliest possible spire-bearers from the Ashby stage (about early CF6 of Sweet and Bergstrom, 1976). This was based on two specimens, one from the Crown Point Fm. of New York and one from the Row Park Fm. of Maryland, neither of which have been figured or described or examined for spiralial configuration. This material may belong to Manespira. No further specimens have yet been discovered (pers. comm. G. A. Cooper, D. Fisher, H. J. Hofmann, T. M. Clarke, C. W. Steam). Williams (1962) identified Protozyga rotunda from the Confinis Flags at Girvan, which are of Llandeilo age. Cooper described three species of Protozyga (here assigned to Manespira) from the Mountain Lake Mbr of the Bromide Fm., slightly younger than the Crown Point Fm. and probably of upper CF6 or lower CF7 age. One of the youngest Manespira is the type species, from CF8 equivalents in Minnesota, where it is very abundant. Diagnosis. Small, ventribiconvex to biconvex, sometimes weakly ventrocarinate shells with smooth umbonal regions, smooth or corrugated anterior and lateral commissure, weakly sulcate, small anacline to hypercline beak. Internally thin shelled, small dental cavities, minute deltidial plates, delicate medially directed teeth; thin socket plates, laterally pointed crura curving dorsally into ascending lamellae, in turn giving rise anteriorly to a dorsal complete or incomplete jugum, and ventrally into a revolution of one whorl or less. Median dorsal septum absent or faint. Remarks. Manespira is distinguished from Protozyga by its possession of a primitive brachidial apparatus with a spiralium of less than one complete whorl, a dorso-anterior complete or incomplete jugum and a shell normally typified by corrugations, even leading to the appearance of a small fold in the dorsal sulcus (text-fig. 6). The calcified brachidial apparatus occupies only a small part of shell volume. Contemporaneous Anazyga are distinguished by their finely ribbed shells and spiralia with several revolutions, and probably evolved from Manespira in early Caradoc time. Suppression of jugal development and hinge modification could also have led to the evolution of Cyclospira from Manespira in late Caradoc time. Manespira is similar to the Ashgill genus Zygospira (Sulcatospira) Xu, 1979, from the Qinghai Plateau, China, in having relatively coarse corrugations or costae, but the latter is closely related to, if not a junior synonym of Zygospira (Zygospira), since it has more complex dorso-medial spiralia with four or more revolutions. Xu distinguished Sulcatospira from Zygospira by its possession of a strong dorsal fold but commonly Zygospira specimens also have a dorsal sulcus interrupted in the centre by a strong costa or corrugation which leads to the production of a central, dorsal ‘fold’ like that of Sulcatospira. A second, more enigmatic Ashgill Chinese genus is Manosia Zeng, 1983, from the Yangtze Gorge area. This is a partly smooth, partly coarsely ribbed or corrugated shell somewhat similar to Manespira in form except in showing many costae and larger shell size: Zeng compared it to triplesioids with a question mark, but the shell may be an atrypoid. Internal structure is unknown. Species assigned. Protozyga costata Cooper, 1956, pp. 676-677, pi. 142a, figs. 1-5. Bromide Fm., Mountain Lake Mbr, Oklahoma; Caradoc. Brachidia undescribed. In view of the considerable rib and shape variation seen in Manespira nicolleti, it seems possible that M. costata and M. elongata (below) are conspecific. Sweet and Bergstrom (1976) indicate that the CF6 and CF7 boundary is located within the Mountain Lake Mbr (i.e. approx, late Costonian). 840 PALAEONTOLOGY, VOLUME 29 ICycIospira diversa Reed, 1917, p. 150, pi. 24, figs. 37-43. Basal Ardwell Mudstones (mid-Caradoc), Ardmillan Braes, Girvan, Strathclyde. Brachidia unknown. Williams (1962, pi. 25, figs. 19, 25, 26) illustrated a damaged specimen with a strong sulcus. This is a doubtful Manespira. Protozyga elongata Cooper, 1956, pp. 677-678, pi. 140e, figs. 27-37; pi. 143i, figs. 41-46. Mountain Lake Mbr, Doleroides Zone, Oklahoma. Williams and Wright (1961, p. 158) illustrated a growth series of the brachidia for this species, showing a simple jugum and partial first whorl which indicate assignment to Manespira. Late CF6 or early CF7, late Chazyan-early Blackriveran; Costonian. Protozyga magnicostata Cooper, 1956, p. 680, pi. 140a, figs. 1-9. Bromide Fm., Mountain Lake Mbr; Caradoc. Brachidia unknown. Possibly a deeply ribbed variant of M. costata. Protozyga tumida Cooper, 1956, p. 687, pi. 140g, figs. 43-47; pi. 141i, figs. 46-50. Effna Fm., Virginia. The four to five corrugations on the flank indicate Manespira-, brachidia unknown. Probably CF6, late Llandeilo or earliest Caradoc. Protozyga iiniplicata Cooper, 1956, pp. 687-688, pi. 140d, figs. 22-26; pi. 141h, figs. 42-45; pi. 141k, figs. 58- 62. Benbolt Fm., Virginia; CF7, Caradoc. Brachidia unknown. No other species have apparently been described. Raymond (1911) identified "Zygospira acutirostris' from the Crown Point Fm. in the Lake Champlain area of New York, but Cooper (1956) pointed out that this is assignable to the rhynchonellid Sphenotreta. There remains the question of whether Camerella longirostris and C. varians Billings (1859) from Llandeilo or possibly Llanvirn rocks of the Mingan Islands, Quebec, are early atrypoids. Cooper (1956) assigned the former species to the triplesiid Onychoplecia and the latter to Camerella. Examination of Twenhofel’s Mingan brachiopod collections do not show evidence of spire-bearing. Manespira nicolleti (Winchell and Schuchert, 1892) PI. 73, figs. 6-19; text-figs. 6 and 7; PI. 75, figs. 3-6 1892 Hallina nicolleti Winchell and Schuchert, p. 293 (no figs., publication date cited as 1 April, 1892). 1892 Zygospira aquila Sardeson, p. 335, pi. 4, figs. 15-18 (publication date 9 April 1892). 1893 Hallina nicolleti Winchell and Schuchert, p. 474, pi. 34, figs. 59 62 (it should be noted that the publication date is shown as 1895 but a published letter at the beginning of vol. 3 states that the first copy of the volume was tendered in December 1891. Nevertheless, Hall and Clarke (1893) cite the publication date as 1893 and this is used here. 1893 Zygospira nicolleti Winchell and Schuchert; Beecher and Schuchert, pi. 10, fig. 23 (illustrates brachidia only). 1956 Protozyga nicolleti Winchell and Schuchert; Cooper, pp. 682-683, pi. 141 j, figs. 51-57. 1977 Protozyga nicolleti Winchell and Schuchert; Bretsky, Bretsky and Schaefer, fig. 9g, p. 124. T\pes. The location of the type material is unknown and presumed lost. A neotype is selected herewith: GS59046, with paraneotypes GS59047, 59048, 59068. Type locality and horizon. \ . . abundant in the upper third of the Trenton limestone at Minneapolis, Rochester and Fountain, Minnesota and Decorah, Iowa’ (Winchell and Schuchert 1893, p. 474). The caption to pi. 34, fig. 64 reads ‘Fountain, Minnesota’ which is the nominal type locality. Cooper (1956) identified the type horizon as the McGregor Mbr of the Platteville Fm., which has a CF8 correlation according to Sweet and Bergstrom (1976), but may be somewhat lower, i.e. CF7 or early Blackriveran. Bretsky et al. (1977) showed that this species is widespread in the lower part of the McGregor Mbr only (the Miiflin Unit) and is extremely abundant, forming up to 13 % of the community. The species was found very abundantly (500+ specimens) at a north side roadcut about 4 km due east of Chatfield along Hwy. 74, Minnesota, some 3 m above a basal sandstone unit. This may be designated a locus typicus restrictus. Associated lithology and fauna. M. nicolleti is abundant on micrite bedding-plane surfaces in concentrations of more than 200 per 100 cm^. Many were found attached to each other (small to larger shells), but the lime muds may also have formed incipient hardgrounds to which shells were fixed. Bretsky, Bretsky and Schaefer (1977) indicated ‘episodically agitated . . . and shallower waters’. Description. Relatively small shell, 4 to 5 mm wide (rarely 6 mm), about as wide as long, ventribiconvex, with rounded outline and hinge angle of 100-110°. Beak small, anacline, foramen apical (deltidial plates minute, not visible externally), anterior commissure sulcate with raised centre producing a distinct dorsal fold and ventral sulcus; two to four corrugations on lateral commissure, with shell smooth, rectimarginate or weakly COPPER: EVOLUTION OE THE LISSATR YPIDAE 841 Manespirg nicolleti TEXT-FIG. 6. Brachidia of Manespira nicolleti (Winchell and Schuchert, 1892), based on text-fig. 7. Note the incomplete jugum and delicate half-whorl of the spiraliuin. Scale X 10. TEXT-FIG. 7. Serial sections of Manespira nicolleti (Winchell and Schuchert, 1892). Mifflin Horizon, McGregor Mbr, Platteville Fm., east of Chatfield, Minnesota; Caradoc, Harnagian. GS79443. Scale x 5. sulcate in early growth stages. Interior of pedicle valve lacking pedicle deposits, large dental cavities and centrally pointed teeth. Crura divergent, ventro-anteriorly directed, curving dorsally to result in ascending lamella; dorsal jugum complete or incomplete, very thin, delicate spiralial half revolution in largest shells (text-hgs. 6, 7). Remarks. Sardeson (1892, pi. 4, fig. 18) showed a one-piece jugum resembling a terebratulid loop. Winchell and Schuchert (1893) illustrated specimens showing as many as seven lateral corrugations on the shell, and, internally, a complete loop-like jugum but no spiralial coil. This feature was not seen in specimens sectioned. M. nicolleti occurs in nest-like concentrations of hundreds of shells, about 60 % of which appear to be tilted at an angle, brachial valve down, shell umbo pointing into the substrate, suggesting life positions. Thus the anterior commissure pointed upwards. Subfamily septatrypinae Kozlowski, 1929 Diagnosis. Smooth or corrugated, plicate, smooth, thin-shelled atrypoids with multi-coiled, dorsally directed spiralia, a posterior jugum or jugal processes, large dental cavities, usually dental plates, thin hinge plates. 842 PALAEONTOLOGY, VOLUME 29 Remarks. In this group, the Ordovician genus Idiospira is the only genus known with a dorsally located, one-piece jugum (possibly it deserves separate subfamily status). All others, as far as those for which internal structures have been studied, have jugal processes. Internal structure of the problematic late Ordovician genus Manosia Zeng 1983 is unknown. Genus idiospira Cooper, 1956 Type species. CamereUa panderi Billings 1859, p. 302 (illustr. Billings, 1863, p. 143, fig. ISa-h), Rockland Fm., Paquette Rapids, Alumette Island, Ontario; Middle Caradoc, Blackriveran, CF8. Range. Middle Caradoc-Ashgill. Nikiforova and Modzalevskaya (1968) assigned two early Llandoverian Siberian species, Protozeuga khetaensis Nikiforova 1942 and Glassia mogoklaensis Nikiforova 1961 to Idio- spira. Specimens of these two species were sectioned, but are not Idiospira (the latter remains provisionally assigned to Glassia). In Ontario and New York Idiospira have a restricted range from the middle Cloche Island Fm. (Bobcaygeon Fm.) to Cobourg Fm., or, Napanee through Denley Limestones, and are absent in Ashgill rocks. Distribution. North America, Europe, Siberia, China. Diagnosis. Small, ovoid, biconvex-dorsibiconvex, smooth atrypoids usually with corrugated anterior commissure. Beaks anacline to nearly hypercline, interarea small, apical to expanded foramen, small deltidial plates. Internally with large dental cavities, thin dental plates, horizontal hinge plates, weak septum, crura latero-ventral and geniculated sharply at jugal connection, jugum simple, straight or weakly arched, dorsal in position. Spiralia up to six whorls, directed dorso-medially at about 40-50° to commissural plane. Remarks. Idiospira are distinguished externally from Protozyga and Cyclospira by their biconvexity, anterior corrugations and dorsal fold. Internally these genera differ radically in their spiralia and jugum. Idiospira, and a still undescribed genus of smooth atrypid from Anticosti Island related to Meifodia, are the only smooth spire-bearers known to have a dorsally located jugum. This appears to be a ‘primitive’ feature in that later taxa have ventral jugal processes (text-fig. 3). Idiospira is ‘advanced’ in the sense of having spiralia with numerous coils directed dorsally (‘primitive’ smooth spire-bearers have medially directed spiralia with very few coils). Idiospira appears to have replaced Protozyga in the late Caradoc or early Ashgill and in turn it was dominated in Ashgill time by ribbed atrypoids such as Zygospira, Catazyga and Eospirigerina. Possibly as many as seventeen described species may be assigned to Idiospira, but most species have poorly known or unknown brachidia and require verification. Species assigned. 1 Protozeuga anticostiana Twenhofel, 1914, pp. 29-31, pi. 1, figs. 8-10. ‘Macasty Bay, zone 4, of English Head Formation’ (= Vaureal Fm.), Anticosti Island, Quebec; Ashgill. ICamarella (sic) bernensis Sardeson, 1892, p. 328, pi. 4, figs. 4-6. ‘From the Camarella bed at Berne . . . Minn(esota)’, Decorah Fm.; Trentonian, late Caradoc. Identified as Parastrophina by Cooper (1956), but internals unknown. IHyattidina charletona Twenhofel, 1914, pp. 34-35, pi. 1, figs. 6-7. ‘Charleton Point, zone 3 of Charleton Formation’ ( = Vaureal Fm.), Anticosti Island; early Ashgill. lAtrypa circulus Hall, 1847, pp. 142-143, pi. 33, fig. la-c. ‘Compact black limestone ... at Middleville’, New York; Trentonian, Caradoc. '1 Rhynchonella cuneatella Davidson, 1883, p. 200, pi. 10, fig. 11. Balclatchie Conglomerate, Balclatchie, Girvan, UK; early Caradoc. Brachidia unknown; assigned by Cocks (1978). Vdiospira gansuensis Fu, 1982, p. 160, pi. 42, fig. 9a~c. Wangyao Valley, Tianzhu, Gansu; Shantai Fm., Llandoverian. Brachidia unknown. ICanierella inornata Weller, 1903, pp. 157-158, pi. 10, figs. 8-10. Jacksonburg Fm., ‘Locality 192A’, Haines- burg. New Jersey; Trentonian, late Caradoc. A relatively large species, brachidia unknown; assigned by Cooper (1956). COPPER: EVOLUTION OF THE LISSATRYPIDAE 843 Idiospira lata Su, 1977, p. 301, p. 115, fig. 17. Lower reaches of the Guangniao He [river], Nenjiang County, Heilongjian Province, north-east China; Guangniaohe Fin.; Caradoc. Brachidia unknown. Cvclospira longa Cooper, 1956, pp. 693-694, pi. 142i, figs. 39-43. Rysedorf Conglomerate, New York; late Caradoc, but possibly reworked specimens. Biconvexity and lack of dorsal sulcus suggest Idiospira', brachidia unknown. IZygospira maynei Roy, 1941, p. 103 = 103, fig. 69. ‘Sillimans Fossil Mount, Frobisher Bay . . . Richmond’, N. Canada; Caradoc. Brachidia unknown. Idiospira minor Fu, 1982, p. 161, pi. 42, fig. 6a-/?. Dongzhang, Liquan, Shaanxi; Beiguoshan Fm., Upper Ordovician (Ashgill). Brachidia unknown. IHyatlidina plicata Mitchell, 1977, p. 127, pi. 17, figs. 29-32. Bardahessiagh Fm., Pomeroy, County Tyrone, Ireland; late Caradoc. Brachidia unknown. IGlassia romingeri Hall and Clarke, 1894, p. 153, pi. 83, figs. 32 -35. ‘in a drifted boulder of Trenton Limestone near Ann Arbor, Michigan’. The specimen illustrated shows a biconvex, bisulcate shell and medially directed spiralia, unlike other Idiospira known. No similar specimens known from probable Caradocian source beds in the northern Michigan basin. Brachidia unknown. IHyattidina sulcata Williams, 1962, p. 254, pi. 25, figs. 47, 53, 58-60. Kiln Mudstones, Craighead, Girvan; late Caradoc. Brachidia unknown. Idiospira taoqupoensis Fu, 1982, p. 161, pi. 42, figs. 7-8. Taoqupo, Yao, Shaanxi; Upper Beiguoshan Fm., Ashgill. Brachidia unknown. Rhynchonella thomsoni Davidson, 1869, p. 186, pi. 24, fig. 18. Craighead Limestone, Craighead Quarry, Girvan; late Caradoc. Brachidia unknown. Idiospira warthini Cooper, 1956, pp. 692-693, pi. 140k, figs. 63-72; pi. 195g, figs. 38-41. Wappinger Fm., near Poughkeepsie, New York; Trentonian, late Caradoc, CFIO. Brachidia unknown. Vdiostrophia sp. 1 Cooper, 1956, p. 590, pi. 113a, figs. 1-5. Boulder in Mystic Conglomerate, Range 6, Lot 20, Stanbridge Township, Quebec; late Caradoc or early Ashgill (but reworked, and possibly older). Brachidia unknown. Idiospira pander! (Billings, 1859) PI. 73, figs. 21-35; PI. 74, figs. 1-5; text-figs. 8, 9 1859 Camerella panderi Billings, p. 302 (no figs.). 1863 Camerella panderi Billings; Billings, p. 143, fig. 78u, b (no description). 1893 Camerella panderi Billings; Hall and Clarke, p. 220, pi. 62, figs. 19-23. 1932 Camerella panderi Billings; Wilson p. 1 39, pi. 2, fig. 4 (Figs. 1 -3, 5 show the septalium-cruralium of Camerella volbortlu, type species of Camerella, which co-occurs with Idiospira and may be confused with it). 1946 Camerella panderi Billings; Wilson, pp. 118-1 19, pi. 11, fig. 2. 1956 Idiospira panderi (Billings); Cooper, pp. 691-692, pi. I08f, figs. 26-32. Type locality. ‘Pauquette’s Rapids’ [i’/c] (Billings 1859, p. 302) and in Billings (1863, p. 176), ‘the best specimens, however, are obtained in the bed of the river’. In an attempt to collect new topotypes, no shells were found on the east side of the Alumette outlier, nor were any fossiliferous outcrops observed on the west side of the Westmeath peninsula facing Paquette Rapids (NTS Fort Coulonge 31F/15W, 49-50: 82). Type horizon. ‘Black River limestone’ (Billings 1859, p. 302). Wilson (1946) identified the syntypes as coming from ‘Leray-Rockland’ beds, equivalent to the Kings Falls Limestone in New York, at the CF8 to 9 transition (Sweet and Bergstrom 1976), or in the Napanee Limestone (CF8). Titus and Cameron (1976) do not list Idiospira from the New York shallow marine communities, though Cooper (1956) cited its presence in the lower Trenton at Amsterdam. Strata around Paquette Rapids have common Solenopora, Foerstepliyllum, and Stromatocerium which tend to indicate a late Blackriver or early Trenton age in the region. I'ype materials. There are six syntypes in the Billings collection dated ‘1845’: these types are GSI 149, 1 \ 49a-e, only one of which corresponds to Billings’s figure, except that the shell is damaged (pi. 1, figs. 21-25). This specimen, GSI 149c, is selected as lectotype. The remaining specimens are all referable to I. panderi. All material from Paquette Rapids is partly or wholly silicified. In addition to the six syntypes, there were sixty-three 844 PALAEONTOLOGY. VOLUME 29 other topotype specimens from Paquette Rapids in the Royal Ontario Museum collection. These were variously labelled as Camerella panderi or C. volbortlii. All specimens were larger than the Billings material, most of the larger specimens having been labelled C. volhorthi, yet all specimens represented a clear growth gradient (text- fig. 9). A further collection of thirty-six specimens was made from the upper Napanee Limestone, about 1 km north of Napanee (NTS 31C/7W 44600:02920). Total 105 specimens. TEXT-FIG. 8. Brachidia of Idiospira panderi (Billings, 1859), based on text-fig. 10. Note the disposition of the jugum and spiralia. Scale x 8. EXPLANATION OF PLATE 74 Figs. 1~5. Idiospira panderi (Billings, 1859). Hypotype ROM23918b, elongated specimen with high dorsal fold (compare with PI. 73, figs. 3 1 - 35), x 3. Fig. 6. ^Triplecella diplicata' (Wilson, 1932). Holotype GS6659, single brachial valve in small limestone fragment. Lot 35, Concession 14, Charlottenburg Township near Ottawa, Ontario; Cobourg Fm., CFll, Caradoc, x 3. This species rests as a nomen diibium since only this fragment is known; very probably this is the dorsal valve of Cyclospira bisidcata (Emmons). Figs. 7-21. Cyclospira bisulcata (Emmons, 1842). Two syntypes from the Emmons collection, and hypotype from the restricted type locality selected. Syntypes from Trenton Limestone, Adams, New York and hypotype from Gulfstream, near Rodman, NY; Flillier Mbr, Cobourg Fm., CFll, Onnian, Caradoc, x 3. 7-1 1, lectotype AMNH713-la, well preserved specimen showing the double fold on the pedicle valve and bisulcate brachial valve. 12-16, paralectotype AMNH713-lb, a more quadrate specimen from the same locality. 17-21, hypotype GS59152, intermediate in shape between the two lectotypes and from Gulfstream locality. Figs. 22-31. Glassia elongala Davidson, 1882. Specimens from Gotland, 61 Visby SO ‘Djupvik 2’ 55730:41660; Mulde Marl, late Wenlock, x 3. 22-26, hypotype Brl06523, typical rounded specimen with small incurved beak. 27-31, hypotype Br 106524, specimen with more strongly developed doubly sulcate commissure. Figs. 32-36. Peratos arrectus n. gen. and sp. Eifel, Germany, MTB Dollendorf 53850;74680; Eilenberg Horizon, Freilingen Beds, late Eifelian, x 3. Hypotype GS59132, a typically sized specimen showing the development of a large interarea, deltidial plates and foramen. PLATE 74 COPPER, Ordovician, Silurian, and Devonian atrypoids 846 PALAEONTOLOGY, VOLUME 29 Diagnosis. Moderately large for the genus, slightly longer than wide, ovoid, biconvex shells with anacline-hypercline beak, mostly smooth shell surface except for one or two lateral corrugations and a broad W-shaped dorsal fold. Internally shells have a broad, flattened U-shaped, postero- central jugum and dorso-medial spiralia of three to four revolutions. Description. Shells are ovoid, elongate, somewhat globular but about 20 % of the shells are wider than long, more flattened and have less incurved beaks. Most shells have hinge angles of 85-95°, reaching maximum width anteriorly, with rounded commissure. Width peaks at 8-9 mm, lengths 9-10 mm, depths at 6-7 mm (text-fig. 9). Somewhat pointed beaks are incurved, but not completely hypercline early in growth, and the small pedicle opening and deltidial plates are usually obscured. Most shells have a broad dorsal fold with the crenulation in the middle producing a W-shape (rare shells lack this); the fold is flanked laterally by two short corrugations which begin late in shell growth (small shells being nearly smooth). TEXT-FIG. 9. Scatter diagrams and frequency curves for Idiospira panderi (Billings, 1859) based on the type specimens and topotypes from the type locality at Paquette Rapids, Ontario (GS1149, ROM30884, ROM23920). Internally, shells have no pedicle deposits, but large dental cavities, and small, hollow deltidial plates surrounding a minute foramen. Teeth are medially directed. Hinge plates are relatively thick, with a weak, striated cardinal process in the cardinal pit, bounded by a modest septum anteriorly. Small crural bases produce delicate crura which point ventrally, then laterally (text-fig. 10). The U-shaped jugum arises posteriorly and curves dorsally, ending as a flat shelf in the shell centre. Spiralial ribbons are tilted strongly to the shell middle but the cone axes point dorso-medially. Remarks. The type species, /. panderi, is distinguished from /. warthini and /. thomsoni by its shape and presence of only one corrugation in the dorsal fold. Within the known succession of ‘chronospecies’ of Idiospira in Ontario, there is a general trend towards size increase, greater number of spiralial whorls and posterior-central migration of the jugum. Differences between species are apparently gradational. COPPER: EVOLUTION OF THE LISSATRYPI DAE 847 TEXT-FIG. 10. Serial sections of Idiospira pcmderi (Billings, 1859). Upper Napanee Limestone, Napanee, Ontario, NTS 3IC/7W 444600:02930; Caradoc, CF8 to CF9. Specimen GS59080. Scale x 4. Subfamily cyclospirinae Schuchert, 1913 (emend.) Diagnosis. Small, smooth, usually ventribiconvex to planoconvex, sulcate shells with normally thick, solid teeth, medially directed spiralia, no jugum. Remarks. On the basis of broad similarity in shell morphology and apparent similarity in muscle scars to Dayia, Schuchert (1913), Schuchert and Cooper (1930) and Boucot et al. (1965) referred the genus Cyclospira to the family Dayiidae. Dayia and other Dayiinae like Protozeuga have a spiralium and jugum which is identical to those of the Hindellinae and Meristellinae. In addition the Dayiinae have laterally directed spiralia and are thus best relegated to the family Meristellidae (with nomenclatorial priority), and to the order Athyrida. Cyclospira differs from athyroid brachiopods in having medially directed spiralia, and in having the spiralia located between diverging crura. In lacking any sign of a jugum or jugal processes it also differs from both athyroids and known atrypoids (thus far the only spire-bearer without this feature). Cyclospira seems an unlikely ancestor of the Athyrida, whose next oldest stratigraphic successor in North America is the Ashgill Hindellcp since it would not only have to evert its spiralia but also ‘grow’ a jugum. Also, Fu (1982) reports a fully developed athyroid, Apheathyris, from late Caradoc strata of China. The median septum of Cyclospira is relatively long, reaching to mid-shell: this is a feature more common to some athyroids. The Cyclospirinae appear to have reached an evolutionary ‘dead-end’ in the Ordovician, and to have been an early, failed experiment in atrypoid filter-feeding strategy. Genus Cyclospira Hall and Clarke, 1893 (= Wilson, 1932; Williams, 1962) Type species. Orthis bisidcata Emmons, 1842, p. 396, fig. 4, Hillier Limestone Mbr, Adams, New York; Caradoc, CFl 1 . Range. Late Caradoc-Ashgill. The lack of information on internal structure on many described species 848 PALAEONTOLOGY, VOLUME 29 makes it difficult to distinguish sulcate protozyginids from cyclospirinids and therefore to determine the exact range of the genus. Typical Cyclospira appear only in the late Caradoc (Actonian-Onnian, or Cobourgian in N. America), but the genus is absent to rare in Ashgill rocks of New York and Ontario. The genus has been reported by Foerste (1893) and Rubel (1977) from Early Silurian rocks, but these identifications are doubtful. Diagnosis. Relatively small, smooth, ventribiconvex to planoconvex shells with a deep pedicle valve, weakly flattened, sulcate-bisulcate brachial valve and anacline to hypercline beak. Interior of shell very thick walled, solid teeth with slit-like dental cavities or dental nuclei. Small, medially oriented, weakly divergent crura, anteriorly continuous with medially pointed spiralia forming one to four coils; jugum lacking. Relatively long dorsal septum (text-fig. 1 1 ). Remarks. The lack of a jugum is unique. It can be differentiated from Manespira and Protozyga, which may have a similar shape, by the thick apical shell wall, absence of large dental cavities and its relatively long dorsal septum. Twenty-five species have been assigned to the genus, and eleven more have been referred to other genera in the past. No doubt many of these are synonymous (see list below). Brachidia have been positively identified in only three species. Species assigned. Cyclospira ahimensis Liu, Zhang and Di 1984, p. 160, pi. 2, figs. 13 14. Mt Altun, China; Caradoc. This species is unusual in apparently possessing ‘dental plates’ and strong ribs on the anterior commissure. Brachidia unknown. Cyclospira (?) barrandei Cooper, 1930, p. 281. A species based on illustrations of Barrande (1879, pi. 28, figs. 11-3-16) from Czechoslovakia; late Ordovician. Ashgill(7). Brachidia unknown. Cyclospira biloha Fu, 1982, p. 166, pi. 43, fig. Sa-d. Yueya Mountain, Ejina Qi, Inner Mongolia; Hengmanshan Fm., Middle Ordovician. Brachidia unknown. Cyclospira(l) canadensis Cooper, 1930, p. 281, pi. 2, figs. 7-8. ‘Dark shales in the road cut along Restigouche River, 3 miles above Matapedia’, Quebec, Canada; late Ordovician. Brachidia unknown. Protozyga carrickensis Reed, 1917, p. 945, pi. 24, figs. 30-31. Craighead Limestone, Craighead Quarry, Girvan, Scotland; late Caradoc (see Cocks 1978, p. 167). Brachidia unknown. Merista cymbula Davidson, 1867, p. 204, pi. 22, figs. 28-29. Hendre-hen, Wales; Ashgill (see Cocks 1978, p. 168). Brachidia unknown. Dayia cymbula var. girvanensis Reed, 1917, p. 948, pi. 24, fig. 33. Upper Drummock Group, Thraive Glen, Girvan; Rawtheyan, late Ashgill (see Cocks 1978, p. 168). Brachidia unknown. Triplecella diplicata Wilson, 1932, pp. 399-400, pi. 5, fig. 13. ‘Lower Cobourg Fm., east half of lot 35, Concession IX, Charlottenburg Township, in the creek bed’; late Caradoc. Only a single valve is known (see pi. 2). This is the type species of Triplecella. Wilson regarded the specimen as a pedicle valve, but Cooper (1956) viewed it as a brachial valve, placing it in synonymy with Cyclospira. If the specimen is indeed a pedicle valve, and this seems indeterminable because too little is preserved, then Triplecella could be a senior synonym of Idiospira. Cooper’s concept is favoured. Brachidia unknown. Cyclospira ejneqiensis Fu 1982, p. 166, pi. 43, fig. 9a-d. Florizon and locality as for C. biloba Fu, 1982. Brachidia unknown. Cyclospira elegantida Rozman, 1964, pp. 188-189, pi. 23, figs. 4-5. Darpirskii Horizon, upper Kalychanskoi Suite, Selennyakh region, left bank Taryn-Yuryakh river, north-east USSR; Caradoc. In 1968, Rozman re- assigned the species to Oligorhynclua, but evidence seems inconclusive. Brachidia unknown. Protozyga gastrodes Temple, 1968, pp. 53-55, pi. 10, figs. 1-12. ‘Hirnantian limestones and mudstones above Keisley Limestone, lane near Keisley, Cumbria’. Strong ventral convexity, anterior sulcus and flat brachial valve point to Cyclospira. Brachidia unknown. Cyclospira globosa Rozman, 1964, pp. 189-190, pi. 23, figs. 1-3. Kalychanskoi Suite, Selennyakh region, Kalychan, north-east USSR; early Ashgill (stratigraphically above C. elegantida). Brachidia unknown. Cyclospira glansfagea Cooper and Kindle, 1936, p. 359, pi. 52, figs. 1,4, 7. ‘Rare in the crystalline limestone lenses on the southwest side of Priest’s road’, Perce, Quebec; Ashgill. Brachidia unknown. Cyclospira levisulcata Roomusoks, 1964, pp. 12-13, pi. 4, figs. 6-10. Nabala Horizon, Lithuania; Harjuan, late Onnian-early Pusgillian, at Caradoc-Ashgill boundary. Brachidia unknown. COPPER: EVOLUTION OF THE LI SS AT R YPI D A E 849 Cyclospira (?) minuscula Cooper, 1930, pp. 280-281, pi. 2, ligs. 9-12. South Cove, locality F8, Perce, Quebec; Whitehead Fm., Ashgill. According to Cooper, ‘spiralia indistinct, suggesting Cyclospira . Rliynchonella mina Davidson, 1869, p. 192, pi. 24, hg. 26. Killey Bridge Fm., Pomeroy, County Tyrone, Ireland; Cautleyan, middle Ashgill (see Cocks 1978). Brachidia unknown. Camarella [av'c] owatonnensis Sardeson, 1892, p. 328, pi. 14, figs. I -3. "Caniarella bed at Owatonna, Minnesota’, probably Decorah Fm.: Trentonian, Caradoc. General shape suggests Cyclospira, but the shell is unusual in having lateral corrugations. Brachidia unknown. Dayia pentagonalis Reed, 1897, p. 74, pi. 6, figs. 5, 5a-c. Keisley Limestone, Keisley, UK; Ashgill. See Cocks 1978, p. 168. Reed mentioned spiralia similar to Dayia, i.e. laterally directed, but this is questionable. Protozyga perplexa Williams, 1962, p. 246, pi. 25, figs. 54, 55, 61, 62. Blue-grey mudstones of late Caradoc age overlying the Craighead Limestone, Craighead Quarry, Girvan (Cocks 1978). Brachidia unknown. Cyclospira schucherti Roy, 1941, pp. 103-104, fig. 70. Frobisher Bay, Baffin Island; Caradoc. ‘Spiralia are slightly introverted and nearly parallel to the vertical plane of the shell’. This indicates Cyclospira. Cyclospira shaanxiensis Fu, 1982, p. 165, pi. 43, fig. (>a-d. Baiwangxilin Gou, Jingyang, Shaanxi Prov., China; Jinghe Fm., Middle Ordovician. Spiralia apparently directed postero-laterally with 2-4 revolutions. Wrotozyga superha Cooper, 1956, p. 686, pi. I41l, figs. 63-69. Auburn Chert, Missouri; upper Caradoc. Brachidia unknown. Cyclospira tetraplicata Fu, 1982, pp. 165-166, pi. 43, fig. la-c. Hengliang Mtn., Tianzhu, Gansu, China; Goulang Fm., Middle Ordovician. Brachidia unknown. Aulidospira trippi Williams, 1962, p. 253, pi. 25, figs. 44-46, 48, 49, 52, 56, 57. Kiln Mudstones, Craighead Quarry, Girvan; late Caradoc. I have examined the type materials; the ‘shoe-lifter’ process is a sparry calcite- mudstone grain boundary reflecting partial shell infill after death (i.e. a geopetal structure). Boucot et al. (1965) mention the lack of a jugum. Cyclospira vokesi Roy, 1941, pp. 104-105, fig. 71. Locality and horizon as for C. schucherti of which it may be a population variant with weak ribs on the shell flanks. Caradoc. Brachidia unknown. Duhious species. ICyclospira circularis Rubel, 1977, pp. 212-213, pi. 1, figs. T-7. Lemmesk-Raikkiu Horizons, Estonia; middle Llandoverian. Brachidia unknown. lCyclospira{l) sparsiplicala Foerste, 1893, pp. 590-591, pi. 37a, fig. 18n-6. Hulfman’s Quarry, ‘Clinton Group’; Llandoverian. Brachidia unknown. Cyclospira hisulcata (Emmons, 1842) PI. 74, figs. 7-19; text-figs. 1113 1842 Orthis hisulcata Emmons, p. 396, figs. 4, Aa-c. 1847 Athyris hisulcata (Emmons); Hall, p. 1 39, pi. 33, fig. 3n-e. 1855 Atrypa hisulcata (Emmons); Emmons, p. 190, pi. 10, fig. ha-e. 1894 Cyclospira hisulcata (Emmons); Hall and Clarke, p. 147, pi. 54, figs. 38-40. 1937 Cyclospira hisulcata (Emmons); Kay, pi. 10 (unnumbered fig.). 1946 Cvc/osp/ra fi/sn/rata (Emmons); Wilson, pi. 11, fig- la-h. 1956 Cyclospira hisulcata (Emmons); Cooper, p. 693, pi. 142l, figs. 54-59. Type locality. ‘Adams, New York’ (Emmons 1842); this locality is no longer exposed. Kay (1933, p. 7) found the species abundantly along Gulf Stream, near Rodman, NY; this material is almost identical to the type specimens. The measured section of Kay was re-collected: C. hisulcata is abundant in a 5 cm thick bed, along a bedding plane in the creek bed, about 30 m east of the bridge over Gulf Stream, 600 m east-north-east of Rodman, NY. This is selected as the restricted type locality. Type horizon. ‘Trenton Limestone’ (Emmons 1842). Hall (1847, p. 139) mentioned ‘shaly Trenton limestone, where few brachiopods occur’. Titus (pers. comm.) placed it in a ‘shallow shelf facies’. The restricted horizon is the shaly part of the Hillier Limestone Mbr (upper Cobourg Fm.), about 10 m below the contact with the ‘black shale of the Utica Fm’ (Kay 1933, p. 7). The Cyclospira bed occurs in dark grey, thinly bedded, relatively barren micrite, and Kay (1937) stated it was also locally abundant in the underlying Steuben (Hallowell) Mbr of the Cobourg Fm. Both of these members lie within CFl 1 and are approximately equivalent to the uppermost Caradoc (Onnian), or lowermost Pleurograptus linearis zone. length 850 PALAEONTOLOGY, VOLUME 29 Cyclospirg bisulcata TEXT-FIG. 1 1 . Brachidia of Cyclospira bisulcata (Emmons, 1842), based on text-fig. 13. Note the medial direction of spiralia and absence of a jugum. Scale x 10. TEXT-FIG. 12. Scatter diagrams and frequency curves comparing Cyclospira bisulcata (shown on left, solid dots, solid lines) from Gulf Stream, NY with an undescribed sp. C (circles, dashed line) from the upper Cobourg Fm., Craigleath, Ontario which is believed to represent a stratigraphically younger form (high CFl 1), depth COPPER: EVOLUTION OF THE LISSATR YPI DAE 851 Type material. Two syntypes from the Emmons collection in the American Museum of Natural History (AMNH), New York. AMNH713-la is selected as lectotype; AMNH7I3-Ib is a paralectotype. In addition more than hfty well-preserved specimens were collected at Gulf Stream, NY; two specimens from the Escanaba River, Cornell, Michigan (ROM24219); twenty specimens from ‘Cobourg Em., Ottawa’ (ROM18913). Diagnosis. Small, slightly elongated, ventribiconvex shells with narrow hinge, inflated, angular umbo, foramen expanded into umbo, strongly incurved beak and bisulcate commissure. Internally thick-shelled, narrow, slit-like dental cavity or dental nucleus, strong, rectangular muscle pad and stout, solid teeth. Dorsally, a distinct, long, median septum (nearly one-third shell length), arcuate crural bases extending into crural blades which are trough-like anteriorly; spiralia of two or three medially directed whorls (text-fig. 1 1 ). Description. Shells relatively small for the genus (average width 6 mm, length 7 mm, depth 4-5 mm; w/l ratios 0-85-0-89 (text-fig. 12). Apical angle about 70 80°, Umbo inflated, beak hypercline, foramen normally breaking into umbo with small, wide flanks adjacent to umbo. Ventral valve strongly inflated, dorsal valve flattened; commissure bisulcate. Internally, apical ventral cavity squared, with weakly indented sides; rounded double muscle pad posteriorly, weak groove anteriorly. Teeth dorso-medial, with stubby sides, pointed ends. Dorsal cardinalia thickly reinforced, crura delicate, arched ventrally and laterally; spiralia with two or three whorls located ventrally, directed to centre (text-figs. 11, 13). TEXT-FIG. 13. Serial sections of Cydospira bisnicata (Emmons, 1842) from the designated type locality, 600 m east-north-east of Rodman, along Gulf Stream, NY; Hillier Mbr, Cobourg Em., late Caradoc, lower CFl 1. GS79442. Scale x 5. Remarks. C. bisnicata occupies a relatively intermediate position in terms of shell size, being smaller than C. schucherti Roy, 1941, whieh at 12 to 13 mm is large for the genus. It differs from other species in its bisuleate commissure, and distinctive shell shape. Evolutionary trends in the known species are not possible to establish at present. 852 PALAEONTOLOGY, VOLUME 29 Subfamily glassiinae Schuchert and Levene, 1928 Diagnosis. Relatively thick-shelled, normally biconvex, smooth atrypoids with medially directed, barrel-shaped spiralia, separated jugal processes, and incipient or fully developed dental plates. Remarks. This subfamily is unique amongst the Siluro-Devonian spire-bearers for possessing medially directed spiralia (text-fig. 14). The oldest known spire-bearing brachiopods, including Ordovician taxa such as Manespira, Protozyga and Cyclospira also developed medially directed spiralia, but this is seen as the ‘starting point’ for spiralia, and thus a primitive trend. Medial orienta- tion places an immediate constraint on expansion of the spiralia, as the growing apices would meet in the shell centre. With lateral spiralia, expansion occurs alongside lateral shell growth (as seen in Spi- rifer), and with dorsal spiralia alongside brachial valve globosity (as seen mAtrypa). Hence the Glassi- inae tended to be small-sized, rarely exceeding 15 mm in width except in Middle Devonian time. Similarly, the spiralia of Glassiinae tend to be barrel-shaped instead of conical (accommodating shell convexity), with the shell commonly being pinched (or bisulcate) in the centre, reflecting spiralium shape. The subfamily was initially erected to include only a single genus, Glassia. The subfamily status was dropped by Siehl (1962), who assigned Glassia to the Lissatrypinae, elevated to family status by Rzhonsnitskaya (1964), only to be dropped once more in the Treatise (Boucot et al. 1965), who as- signed these to the Septatrypinae. Three other genera are here assigned to the Glassiinae: Karhous Havlicek, 1985 (which has shell structure and dental plates like some Glassia), Peratos n. gen. and, questionably, the genus Holynatrypa Havlicek, 1973, for which brachidia are still undescribed (but which were stated by Havlicek to have medial spiralia). Genus Glassia Davidson, 1 88 1 (= CryptatrypaS\t\\\, \962) Type species. Atrypa ohovata Sowerby, 1 839, p. 61 8, pi. 8, figs. 8-9. ‘Mathon Lodge, west flank of Malvern Hills’, England; ‘Lower Ludlow rock’. Range. ?Middle Llandoverian, Wenlock-Ludlow, ?Pridoli. Distribution. North-west Europe, Central Asia?, China?, Australia, north-west Canada? (the only definitive Glassiinae to date are known from north-west Europe). Diagnosis. Small- to medium-sized, biconvex, smooth, rounded shells, possessing anacline-hypercline beaks, small apical or trans-apical foramen, deltidial plates and small interarea; rectimarginate or bisulcate. Internally thick-shelled, with minute posterior dental cavity filled in anteriorly to produce stout, solid teeth with an inner, buried, thin, dental plate. Dorsal valve with stout hinge plates, narrow cardinal pit, small crural bases, rapidly diverging distally feathered crura. Jugal processes ventro-posterior, arched centrally, terminating in small jugal plates which nearly touch to form an O-structure. Spiralia medial to ventro-medially directed with posterior parts of whorls trough-shaped, directed to jugal processes (text-figs. 14, 15). Remarks. Glassia is similar to its Devonian descendant, the genus Karbous Havlicek, 1985, in it biconvexity, small anacline beak and interarea, but differs internally by its absence of dental cavities and perhaps by its modest jugal plates (since these are undescribed for Karbous). Glassia differs from Middle to Late Devonian Peratos n. gen. in its incurved beak and lack of prominent dental plates and large dental cavities. There was a trend in the Glassiinae towards increasing the size of the interarea, developing an orthocline beak, and enlarging the dental cavities and the jugal plates. The genus Holynatrypa Havlicek, 1973, from Emsian-Eifelian strata of Czechoslovakia, was said to have ‘spiralia probably directed medially’. If this is correct, it must belong to the Glassiinae, but the beak is minute and shell shape resembles Lissatrypa. It is a doubtful glassiinid and probably a Lissatrypa. COPPER: EVOLUTION OF THE LISSATR YPIDAE 853 Glassig elonqata TEXT-FIG. 14. Brachidia of Glassia elongata Davidson, 1881, based on text-fig. 15. Note that crura are fibrous, jugal processes separated and spiralia curved in the centre. Scale x 8. The genus Spondylobolus M‘Coy, 1851 (type species S. craniolaris M‘Coy, 1851) is probably a senior synonym of Glassia Davidson, 1881. The name has been used only twice in the literature; firstly by Salter (1873, p. 135), who stated ‘a genus unfortunately founded in mistake; a species of Meristella, probably M. obovata being so pressed in shale . . . ’, thus recognizing its identity with Glassia and the senior synonym status of the genus Spondylobolus. This was re-asserted by Cocks (1978, p. 160). The name Spondylobolus should be suppressed because it would create substantial confusion in the literature. The Ludlovian genus Quangyuania Sheng, 1975 from China superficially resembles Glassia, but figures of the type species, Q. ovalis, indicate teeth, crura, spiralia, and a dorsal septum as in Lissatrypa, which is probably a senior synonym. Similarly, the Ludlovian genus Buceqia Havlicek, 1984, type species Terebratula obolina Barrande, 1847, is, from the serial section data provided by Havlicek, internally identical to Lissatrypa Twenhofel, 1914, and is here regarded as a junior synonym thereof. The first evidence for the unique medial orientation of the spiralia in Glassia was provided by Kunth (1865), who examined and illustrated spiralia from Glassia found in Silurian glacial erratics in North Germany (derived from Gotland: see also Gagel, 1890, pi. 1, fig. 38). This was confirmed by the work of Glass (in Davidson 188l£/, b), which led to the founding by Davidson of the genus. Glass was remarkably accurate in his work but for showing a jugum, instead of disconnected jugal processes. The oldest known, possible Glassia is G. mogoktaensis Nikiforova 1961, for which Modzalevskaya (in Nikiforova 1961) showed medial spiralia. Glassia may have evolved from either Meifodia or Lissatrypa. Some twenty species have been assigned to the genus, but for most the brachidia are unknown; there may be substantial synonymy. Four species previously assigned have been removed. 854 PALAEONTOLOGY, VOLUME 29 TBXT-FIG. 15. Serial sections of Glassia elongata Davidson, 1881 from the lower Mulde Beds, Djupvik 2, 61 Visby SO 41660:55730, Gotland, Sweden; middle Wenlockian. A dental cavity is present posteriorly, but dental plates are absent; spiralia are barrel-shaped instead of conical. GS79444. Scale x4. Species assigned. lAlrypa canaliculata Barrande, 1879, pi. 15, hgs. I, 1-4; pi. 145, hgs. III-VII. ‘Dlauha Hora, Etage E-F’, Kopanina Em.; Ludlovian {non CoUarolhyris canaliculata Venyukov, 1899). ITerehratida cingulata Meyer and Munster, 1840, pp. 77-78, pi. 14, figs. 12, I3a-b. ‘Orthoceratitenkalk von Elbersreuth’, Germany; Wenlock-Ludlow. Brachidia unknown. Atrypa compressa Sowerby, 1839, p. 629, pi. 13, fig. 5 (both views). ‘Woodside and Nash, near Presteign’; Wenlock shales. Usually regarded as a synonym of G. ohovata. Spondylobolus craniolaris M'Coy, 1851, p. 408 (figured in M'Coy 1855, p. 255, pi. 1h, figs. 4-5). ‘Black shale of Builth Bridge', Powys, UK; Wenlock Shales. A probable senior synonym of G. obovata. It should be declared a nomen nudum. Atrypa decapitata Heritsch, 1929, p. 8, ‘E2 von Bohmen (Kozel)’. This has sometimes been cited as a Glassia but the name appears invalid (no figures are known). Glassia elongata Davidson, 18816, pp. 148-149, pi. 5, figs. 3-4. ‘Railway cut between Tickwood and Farley Dingle’ (Davidson 1881«, p. 103), Wenlock Edge, Salop; Coalbrookdale Em., formerly Tickwood Beds (Sheinwoodian-Homerian: Wenlock). This is a common species in Gotland (see pi. 2, figs. 22-31; pi. 3, figs. 2, 7; text-figs. 14-15). It may be a synonym of Atrypa laevigata Kunth 1865. 'ITerebratula ephemera Barrande, 1847, p. 408, pi. 16, fig. 1 \a-e. ‘Grenze . . . unteren und mittleren Kalketage E und F in den Umgegenden von Beraun und St. Iwan’; Ludlovian-Pridolian. Brachidia unknown. ! Atrypa fugitiva Barrande, 1879, pi. 84, figs. V, 1-3; pi. 136, Figs. Ill, 1-2. ‘Dlauha Hora . . . Kolednik, -E2’; Ludlovian. Brachidia unknown. 1 Atrypa fugitiva Barrande var. depressa Vinassa de Regny, pp. 557-558, pi. 20, fig. 6«-6. Volaia, Carnic Alps, Italy, ‘Calcari con crinoidi e Retzia (= Gracianella) umbra'\ Pridolian. Brachidia unknown. ICryptatrypa glaberconcha Lenz, 1977, p. 1550, pi. 12, figs. 1-13. ‘Locality 2, approximately 6 miles east of COPPER: EVOLUTION OF THE LISSATRYPIDAE 855 Avalanche Lake, Mackenzie Mountains, 90-99 m below section top, Whittaker-Road River formations transition’; early Wenlock. Brachidia unknown. ISpirigera heimoi Heritsch, 1929, p. 34, pi. 3. figs. 145-150. ‘Rotenkalk’, Kokberg, Carnic Alps, Austria; Ludlovian. Brachidia unknown. lAiistraliiia (Ausiralina) kraitsi Strusz, 1982/), pp. 123-126, figs. 19-20. Mudstone in the Walker volcanics, Molonglo Valley west of Canberra, Australia; Wenlockian. Brachidia unknown. Biconvexity and interarea suggest Glassia but internal structures are unknown. Atrypa laevigata Kunth, 1865, p. 313, pi. 7, fig. \a-e. ‘Graptolithengestein, Tempelhof, Berlin’; Ludlovian (specimens from glacial erratics probably derived from Gotland to the north). Figures were later copied by Heidenhain ( 1 869, p. 155) and Roemer ( 1 885, pi. 9, fig. 11). Spiralia medially directed. 1 Atrypa lindstroemi Venyukov, 1899, pp. 122-123. Studenitsa Limestone, Studenitsa, Podolia; Wenlockian. Brachidia undescribed. IGIassia miiiuta Rybnikova, 1967, pp. 203-204, pi. 23, fig. 3a-f. Kholdve, Latvia; Llandoverian. May be a Lissatrypa. Brachidia unknown. IGlassia mimita Fu, 1982, pi. 42, fig. \3a-d. Shimen valley, Zhouqu, Gansu, China; Zhouqu Fm., Wenlock. Spiralia undescribed. Glassia mogoktaensis Nikiforova, 1961, pp. 250 252, pi. 52, figs. 10 16. Mogokta River, Khantaiki River basin, north-west Siberia; middle Llandoverian. Medially directed spiralia present (ibid., text-fig. 44), jugum unknown. IKarpimkia nalivkini Nikiforova, 1937, p. 23, pi. 2, figs. 4-5. Outcrop 723, western Balkhash, central asiatic USSR; Wenlockian. Brachidia unknown. Terebratala philomela Barrande, 1847, p. 387, pi. 15, fig. 7. 'Letzten Schichten unseren Kalketage E . . . und hochste Entwicklungstufen in dem untern Theil unseren mittleren Kalketage E’, no locality cited. According to Havlicek ( 1985) this species is of Wenlock age (Motol Fm.). ICryptatrypa praecordata Kulkov, 1974, pp. 69-70, pi. 23, fig. 7, pi. 24, fig. 1. South-west Altai, Gora Rosypnaya, Talyi, Chinetinsk Florizon, Yavorski Fm.; upper Llandoverian. Brachidia unknown. INiicleospira raritas Amsden, 1968, p. 157, pi. 8, fig. 8a-/;. Batesville District, Arkansas, St Clair Limestone; Ludlovian. Brachidia unknown. This occurrence, if correct, would be anomalous since no other Glassia are known from eastern North America. IGlassia rotunda Rybnikova, 1967, pp. 201-202, pi. 23, fig. 2a-/. Piltene, Latvia; Pristiograptus tumescens Zone, Ludlovian. Brachidia unknown. ICryptatrypa rotunda Lmz, 1977, pp. 1550-1552, pi. 12, figs. 14-27. Locality and horizon as for C. glaberconcha Lenz, 1977. Brachidia unknown; a homonym of G. rotunda Rybnikova. Atrypa subcompressa mut. progona Freeh, 1887, p. 121 . ‘Vorkommen in der Zone der Rliynchonella niegaera am Wolayer Thorl’; ? Pridolian. Brachidia unknown. Species deleted or dubious. Glassia drevernianni Mailleux, 1936, referred to Peratos n. gen. Cryptatrypa fabraeusi Lenz, 1970, pp. 492-493, pi. 86, figs. 11-24, 26, 28. 'Upper part of the Road River Formation, Prongs Creek, Yukon’; Pridolian. Very large dental cavities and shell form appear to indicate Septatrypa. Brachidia unknown. Glassia minuscida Dahmer, 1932, referred to Karbous. Terebratala obolina Barrande, 1847, pp. 404-405, pi. 20, fig. 16a-c. 'Kalkbanken mit Phacops glockeri und Arethusa konincku near Beroun, Czechoslovakia; Wenlockian. Havlicek (1984) has demonstrated dorsally directed spiralia and internal structure like Lissatrypa, but assigned it to a new genus, Bueeqia. Glassia paucicosta Spriestersbach and Fuchs, 1909, referable to Karbousl IGlassia ronungeri Hall and Clarke, 1894, referred to Idiospira. Glassia sulcata Siehl, 1962, referred to Peratos n. gen. IGlassia tenella Williams, 1951, pp. 114-115, pi. 5, figs. 16-18. ‘Mountain road almost half a mile SSW of Cwm Crychan, north-east of Llandovery’; Llandoverian. The cardinalia shown resemble those of Lissatrypa. Brachidia unknown. 856 PALAEONTOLOGY, VOLUME 29 Cryptatrypa triangularis Johnson, Boucot and Murphy, 1976, pp. 75-76, pi. 25, figs. 22-30; pi. 26, figs. 1-4. ‘B fauna, Pete Hanson Creek area’, Alaska; Ludlovian. Brachidia unknown. The peculiar triangular shape of this species suggests possible affinity with Eokarpinskia nalivkini (Nikiforova 1937), the type of the genus (Rzhonsnitskaya 1964), a smooth or nearly smooth brachiopod whose brachidia are unknown. E. nalivkini was described by Lenz (1970) as a Cryptatrypa. Lenz’s species is similar to Terebratula bauds Barrande, 1847, whose internal structure is still unknown. None of these are specifically referable to the subfamily Glassiinae. Glassia variabilis Whiteaves, 1904, pp. 42-43; Whiteaves 1906, pp. 273-274, pi. 26, figs. 3-5. ‘Loose blocks of limestone from or near the mouth of the Winisk River’, Hudson Bay Lowlands, Ontario; Wenlockian. Types re-examined: spiralia dorsal and specimens referred to Atrypopsis. Genus Karbous Havlicek, 1985 Range. Gedinnian (Lochkovian)-Dalejan (Lower Devonian, possibly lowest Eifelian). Distribution. North-west Europe, Urals, ?central Asia. Emended diagnosis. Small to medium-sized, smooth, ventribiconvex to planoconvex shells with rela- tively prominent incurved beak, hidden foramen, solid deltidial plates, usually rectimarginate or weakly bisulcate. Internally thick-shelled, with small, short, dental plates flanked by small, slit-like dental cavities. Dorsally thick, disjunct hinge plates. Crura and spiralia yet undescribed or figured. Remarks. The genus diflfers from Glassia in having a clearly defined dental plate and usually a narrow, slit-like dental cavity (Siehl 1962, has shown that the dental cavity may be filled in but that the dental plate is still clearly developed as a single crystal structure within the teeth). In Glassia there is a thin lining to the pedicle cavity (pi. 75, fig. 1; text-fig. 17) and an apical dental cavity lining, the precursors of dental plates in Karbous. Karbous differs from Peratos in having an incurved beak and small dental cavities. Species assigned. lAtrypa canaliculata var. dissidens Barrande, 1879, pi. 146, figs. I, 1-2. ‘Konieprus, Etage F’, Czechoslovakia; Pragian, Devonian. Brachidia unknown. ICryptatrypa curvirostris Xu, 1979, p. 374, pi. 4, figs. 22-28. Luofu, Nandan, Guangxi Province, south-west China; late Emsian-early Eifelian. Brachidia unknown. lAtrypa insocia Barrande, 1879, pi. 147, fig. 1. ‘Konieprus, Etage F’; early Devonian. Brachidia unknown. IGlassia minuscula Dahmer, 1922, pp. 287-288, pi. 15, figs. 23-25. ‘Giengelsberger Schichten’, Giengelsberg, Germany; Emsian. Brachidia unknown: appearance possibly suggests the anoplothecid athyroid Dnestrina. IGlassia paucicosta Spriestersbach and Fuchs, 1909, pp. 68-69, pi. 10, figs. 3-4. Dalhausen, Germany, Remsch- eider Schichten; upper Emsian. Brachidia unknown. Karbous truncatus Havlicek, 1985, pp. 237-238, pi. 2, figs. 3-4. Suchomasty Limestone (Dalejan; Emsian); Herget Quarry, Koneprusy, Czechoslovakia. Karbous vaneki Havlicek, 1985, p. 238, pi. 1, figs. 1-2. Zlichov Limestone, Zlichovian, Lower Devonian; Koneprusy area. Species deleted. Cryptatrypa lenticula Perry 1984, pp. 105-106, pi. 40, figs. 1-27. ‘Locality S3, East limb of Sekwi Anticline, north-north-west of Natla River, 16L5-164-6 m level below top of Delorme Em’; Zlichovian. In this posthumous manuscript Perry illustrated laterally directed spiralia which are unmistakably athyroid; the species is best assigned to Protathyris. Atrypa subcompressa Freeh 1887, pp. 726-727 (designated for figures in Barrande, 1879, pi. 85, fig. 1; pi. 1 14, fig. 4), ‘Konieprus’, Czechoslovakia; Eower Devonian. Barrande’s figures show laterally directed spiralia, and this is thereby an athyroid. Genus Peratos n. gen. Name. From the Greek, peratos, meaning the end of the line, referring to the fact that this genus is the last occurrence of the Glassiinae. Type species. Peratos arrectus n. sp. Eilenberg Horizon, Freilingen Beds, late Eifelian; Eifel, Germany. COPPER: EVOLUTION OF THE LISSATR YPIDAE 857 Range and distribution. Eifelian-Frasnian, Devonian; north-west Europe. Diagnosis. Relatively large glassiinids with a biconvex shell, prominent interarea with erect, ortho- cline beak and large, exposed pedicle foramen. Internally with long, straight, flat dental plates and very large dental cavities. Delicate crura feathered distally, short jugal processes giving rise to highly arched jugal plates producing a trough structure. Spiralia medially directed as in Glassia and possessing spines anteriorly (text-figs. 16 and 17). TEXT-FIG. 16. Brachidia of Peratos arrectus n. gen., n. sp., based on text-fig. 17. Note the large jugal plates. Scale X 8. Remarks. This genus differs externally from other members of the subfamily in having a large, orthocline beak and relatively large interarea. On the inside it has long and straight dental plates defining a large, open dental cavity behind them. Spiralia are identical to those of Glassia, i.e. barrel-shaped and medially oriented, but the jugal processes are quite different. The presence of dental plates in Devonian Cryptatrypa (here identified as Peratos) led Siehl (1962), who did not find spiralia, and Boucot et al. (1964), to place such shells in the family Septatrypinae. But the overall shell wall and hinge structures of Glassia, Karhoiis, and Peratos, and especially the peculiar spiralia (text-fig. 16), show that the three are related. In Glassia there is a thin lining to the pedicle cavity (pi. 75, fig. 1; text-fig. 17) and an apical dental cavity lining, signalling dental plates in both Karhous and Peratos. Peratos is most common in calcarenites associated with reef development in the Devonian. Silurian Glassia are more common in dark gray to black calcareous shales or micritic limestones representing offshore environments. It is possible that this represents a change in habitat in the Glassiinae from Silurian to Devonian time. There appears to be strong homeomorphy between Peratos, Devonian athyrids like Protatliyris, and centronellid brachiopods, which makes it difficult to assess described species in the literature unless the brachidia are known. About thirteen species of probable or possible Peratos have been described. 858 PALAEONTOLOGY, VOLUME 29 Species assigned. Rliynchonella beyriclii Kayser, 1872, p. 678, pi. 26, fig. 6a k. ‘Rotheisenstein’, red limestone, Brilon, Germany; late Givetian or Frasnian. Refer to Maurer (1885, pp. 192-193, pi. 8, figs. 11-15), who illustrated the medially directed spiralia, Torley (1934, p. 125, pi. 9, figs. 78-79) and Gunia (1962, p. 511, pi. 47, figs. 11- 12) for additional material. Serial sectioning by me of Brilon specimens shows medially directed spiralia and internal structures like the type species. lAtrypa canaliculata Barrande, var. 1 Maurer 1881, p. 38, pi. 2, fig. 23a-c. Flerborn, Germany, Greifensteiner- kalk; Eifelian. Brachidia unknown. lAtrypa canaliculata Barrande, var. 2 Maurer 1881, p. 39, pi. 2, fig. 2Aa-c. Locality and horizon as for var. 1. TEXT-FIG. 17. Serial sections of Peratos arrectiis n. gen., n. sp. from the Eilenberg Horizon, Freilingen Beds, MTB Dollendorf 53850 : 74680, Eifel, Germany; late Eifelian. Note the dental plates and dental cavity and unusual jugal plates. GS59133. Scale x 4. IGlassia drevennanni Mailleux, 1936, p. 25 (no figs.). ‘Les schistes de Matagne’, Couvin and Sautour, Belgium; Frasnian F3b. Brachidia unknown. Cryptatrypa bassiaca Siehl, 1962, pp. 198-199, pi. 27, figs. 5-6, pi. 37, figs. 4, 6. ‘Schurf in der Wiege, Greifenstein, Schicht 17.0 m Greifensteinerkalkes’; upper Eifelian. Brachidia unknown. ITerebratida newtoniensis Davidson, 1867, pp. 8-9, pi. 1, figs. 16 17. Lane’s or Woolborough Quarry, Newton Abbot; Middle Devonian. These are very large shells and may belong to early terebratulids. Cryptatrypa philomela minor Biernat, 1966, p. 110. Holy Cross Mountains, Poland, Skala Beds; Givetian. ITerebratida puschiana Verneuil, 1845, pp. 69-70, pi. 9, fig. lOu-e. ‘A Ulahue, entre Krapivna et Odoiefif, sur la route de Toula a Kalouga’; ?Upper Devonian. Compared by Verneuil to Glassia ohovata, but brachidia unknown. ICamarophoria rhomboidea Phillips, in Tietze 1871, p. 151, pi. 17, figs. 41, 4\a. 'Hauptkalk, Ebersdorf, Germany; Middle Devonian. Brachidia unknown. COPPER: EVOLUTION OF THE LISSATR YPI DAE 859 ITerebratula roiwulata Meyer and Munster, 1840, p. 75, pi. 14, fig. 'ia-h. 'Clymenienkalk von Schubelhammer’, Germany; Frasnian. This species was cited by Drevermann (1901, p. 168) and Gunia (1968, p. 166, pi. 7, fig. Met). Brachidia unknown. Glassia sulcata Siehl, 1962, pp. 194-195, pi. 25, figs. 2-4; pi. 37, figs. 1-3. ‘Schurf in der Wiege, SW Griefen- stein . . . Schicht 5-20 m des Greifensteinerkalkes’; upper Eifelian. Atrypa verrucula Maurer, 1881, pp. 43-44, pi. 3, fig. 9a-c. Griefensteinerkalk, Herborn, Germany; Eifelian. Serial sections by Siehl (1962) indicate the genus Peratos. Glassia whidbornei Davidson, 1882, pp. 38-39, pi. 1, figs. 10-14. ‘Middle Devonian limestone at Lummaton, England; Late Eifelian or Givetian. Brachidia medially directed. Peratos arrectus n. sp. PI. 74, figs. 32-36; PI. 75, fig. 1; text-figs. 16, 17 Name. Latin arrectus, referring to the large, upright beak. Type locality and horizon. Small roadcut in the Hillesheim Syncline, Eifel, MTB Dollendorf 53850:74680; Eilenberg Horizon, Freilingen Beds, Late Eifelian. Diagnosis. Shell medium-sized, rounded in outline, 12-15 mm wide, equidimensional or slightly longer than wide, biconvex, with orthocline beak and hinge angle averaging about 110'^. Internal structures as shown in text-figs. 16 and 17. Frasne Givet o > LLI Q Eifel Ems Prag Lochkov Pridoli < DC ID C/D Ludlow Wenlock Llandovery < o > o Q a: O Ashgill Caradoc Llandeilo GLASSIINAE LISSATRYPINAE CYCLOSPIRINAE PROTOZYGINAE I I SEPTATRYPINAE TEXT-FIG. 18. Stratigraphic distribution of known genera and subfamilies of smooth atrypoid brachiopods from Ordovician through Devonian time. The genera Nanospira, Lissatrypoidea, and Tyrothyris are taken as synonyms of Australina. Buceqia Havlicek and Quangyiiania Sheng are interpreted as probable synonyms of Lissatrypa. Cryptatatrypa is a synonym of Gla.ssia. The smooth brachiopod Manosia Zeng is a genus whose affinity is uncertain. The genus Holynatrypa is also of very uncertain status. Synonymous and dubious genera are not included in the range chart. 860 PALAEONTOLOGY, VOLUME 29 Remarks. This species is larger and less elongate than Peratos heyrichi (Kayser, 1872), of which a small collection was available for comparison, and Glassia whidhornei Davidson, 1882. Davidson was the first to figure the internal morphology of the genus correctly, i.e. the medially directed spiralia he demonstrated for P. whidhornei. However, he showed a fused jugum, which is not present in the two species examined. Peratos arrectus lacks the sulcus of Glassia sulcata (Siehl, 1962) and is substantially larger in size and with a more prominent beak than Cryptatrypa philomela minor Biernat 1966 or Cryptatrypa hassiaca Siehl, 1962. It is most similar in external form to Terebratula newtoniensis Davidson, 1867, but this is an even larger shell whose internal structure is unknown. CONCLUSIONS The diversity of the smooth atrypoids was at a peak during the Silurian, especially Wenlock- Ludlow time (text-fig. 18). Locally, smooth atrypoids such as Meifodia, Lissatrypa, Atrypopsis, Septatrypa, Atrypoidea, Austrcdina, and Glassia were abundant enough to have formed thick coqui- nas, sometimes dominating to exclude other brachiopods. Such domination was very rare in the Devonian, when the group was in retreat, probably due to competition from the ribbed and frilly atrypoids. The last surviving group of smooth atrypoids, dying out in the Frasnian, was the Glassiinae, represented by the genus Peratos, which had medially directed spiralia similar to those with which the earliest atrypoids started in the Ordovician. The spire-bearers probably first evolved about 470 Ma ago (Llandeilo), then saw relatively rapid evolution via a burst of diversification in Harnagian-Soudleyan times (CF7 to CF9). A second major burst occurred in the late Caradoc (late CFIO to CFl 1 ), some 455 Ma ago, with the expansion of the primitive smooth spire-bearers {Idiospira and Cyclospira) and ribbed spire-bearers (Zygospira and Catazyga). These earliest spire-bearers were then rapidly replaced during late Rawtheyan to Hirnantian times by the evolution of the Atrypidae, heralded by Eospirigerina. The Zygospiridae hung on as ‘living fossils’ during the latest Ordovician (CFl 3) and Llandoverian, to die out in Telychian (late Llandovery) time, except for the Tuvaellinae which persisted through to the late Silurian. It is difficult to extend precise correlations between the brachiopod, graptolite and conodont zonations of the Caradoc and Ashgill. The conodont and graptolite zones tend to be of relatively long duration, for example, when compared to the upper Devonian. For the Upper Devonian, which lasted roughly 20 Ma, more than ten conodont zones exist. This compares unfavourably with the Upper Ordovician (Ashgill), which lasted about 19 Ma, and for which only two conodont zones are known. Brachiopods have the potential for a more precise dating scheme for Caradoc to Ashgill shelf carbonates and shales. The atrypoid spire-bearers show relatively rapid evolution at the genus level, and use for correlations at the species level. At the Ordovician-Silurian boundary only two atrypoid genera, Idiospira and Cyclospira, may EXPLANATION OF PLATE 75 Fig. I. Peratos arrectus n. gen. and sp. Peel photograph of serially sectioned specimen demonstrating very thin dental plate and tissue lining the dental plate, dental cavity, deltidial plate (refer to text-fig. 17, section 1 -2 mm), X 40. Figs. 2, 7. Glassia elongata (Davidson, 1882). Peel photographs, x40. 2, showing the presence of a very thin dental plate in the apical shell portions (arrow; see text-fig. 15, section at 0-6 mm). 7, showing the hinge plate structure with teeth in place and narrow crural bases (section at 1 -4 mm). Figs. 3-6. Manespira nicolleti (Winchell and Schuchert, 1892). Peel photographs, x 40. 3, 4, show the terminal points of the incomplete jugum (see text-fig. 7, section 2-9 mm). 5, 6, show the anterior portions of the jugum adjacent to their connection with the spiralium at 3-2 mm from the apex (see text-fig. 7). Scale x 40. Note the differences in the coarse calcite crystals which surround the ventral parts of the brachidia and the fine-grained surrounding matrix. This is common in all spire-bearers and appears to reflect the fact that early sediment infill preceded decay of the soft tissues. PLATE 75 dental plate deltidial plate jugum jugum crural base '■•X. X dental cavity ! rS^lngy COPPER, atrypoid internal structures 862 PALAEONTOLOGY, VOLUME 29 have become extinct, and even these two have been reported (albeit here uncorroborated) from Silurian rocks. There is therefore no apparent major extinction of spire-bearing taxa at this boundary. The major decline and turn-over occurred during late Rawtheyan and/or early Hirnantian time, and this did not differ markedly from the late Caradoc turn-over. If late Ordovician glaciation did have an impact, it was at the CF12-CF13 (i.e. late Rawtheyan-early Hirnantian) boundary, which marks the introduction of strong ‘Silurian’ elements into the spire-bearer faunas, e.g. the appearance of Hindella (Meristellidae) and Eospirigerina (Atrypidae). It should be noted that not all smooth atrypoids were necessarily derived monophyletically. A trend towards rib elimination and smoothing of the shell surface seems to be present in some genera of the Palaferellidae (e.g. Gracianella, Prodavidsonia, Zeravshania, Eokarpinskia) and Atrypidae (e.g. Beitaia Rong and Yang 1974, a very finely ribbed clintonellinid, is almost a homeomorph of Septatrypa\ some Spinatrypa in the late Devonian have lost almost all ribs). The key to the evolu- tionary relationships of these secondarily derived ‘smooth’ genera, and other smooth brachiopods, must be internal structure, which is still unknown for more than ninety-five possible or probable species. Acknowledgements. I would like to thank the many colleagues who have replied to queries concerning the global location of type materials, type localities and horizons, literature, and other data bearing on their palaeoecology and distribution. Those who contributed were; Valdar Jaanusson, G. A. Cooper, Bob Titus, Hans Hofmann, Donald Fisher, Peter Bretsky, Colin Steam, Alwyn Williams, Tony Wright, Robin Cocks, Len Alberstadt, Ken Walker, Madis Rubel, T. L. Modzalevskaya, O. I. Nikiforova, M. A. Rzhonsnitskaya, S. V. Cherkesova, Fu Li-pu, and Rong Jia-yu. Figured materials are stored in the Geological Survey of Canada Museum (GS), American Museum of Natural History (AMNH), Royal Ontario Museum (ROM), and the Naturhistoriska Riksmuseet Stockholm (BR). Financial support for this study came from the Natural Sciences and Engineering Research Council of Canada. REFERENCES AMSDEN, T. w. 1968. Articulate brachiopods of the St Clair Limestone (Silurian), Arkansas and the Clarita Eormation (Silurian), Oklahoma. Paleont. Soc. Mem. 1, 1-117. BARRANDE, j. 1847. Ueber die Brachiopoden der silurischen Schichten von Bohmen. Natnrwissenschaftliclien Ahhandhmgen 1(1). Braumiiller und Seidel, Wien. 357-475. 1879. Systeme Silurien du centre de la Boheme, 7, Recherches paleontologiques, 5, Classe des mollusques, ordre de hrachiopodes, 226 pp. -1- atlas. Charles Bellmann, Prague. BEECHER, c. E. and SCHUCHERT, c. 1893. Development of the brachial supports in Dielasma and Zygospira. Proc. biol. Soc. Wash. 8, 71-82. BERRY, w. B. N. 1976, Aspects of correlation of North American shelly and graptolitic faunas. In bassett, M. G. (ed.). The Ordovician System, 153-169. University of Wales Press and National Museum of Wales, Cardiff. BiERNAT, G. 1966. Middle Devonian brachiopods of the Bodzentyn Syncline (Holy Cross Mountains). Palae- ontogr. polonica, 17, 1-162. BILLINGS, E. 1859. Description of a new genus of Brachiopoda, and on the genus Cyrtodonta. Can. Nat. Geologist, 4, 301-303. 1863. Paleontology. In logan, w. e. (ed.). Geology of Canada. Geol. Surv. Can. Rep. Prog. 1863, 9, 136- 197. BOUCOT, A. J., JOHNSON, J. G. and STATON, R. D. 1964. On some atrypoid, retzioid and athyridoid Brachiopoda. J. Paleont. 38, 805-822. BOUCOT, A. J., JOHNSON, J. G., PiTRAT, c. w. and STATON, R. D. 1965. Spirifcrida. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part H. Brachiopoda. H632-H728. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. bretsky, p. w., bretsky, s. s. and schaefer, p. j. 1977. Molluscan and brachiopod dominated biofacies in the Platteville Formation (Middle Ordovician), upper Mississippi Valley. Bull. geol. Soc. Denmark, 26, 1 1 5-132. CLARKE, J. M. 1913. Fosseis devonianos do Parana. Monografias Div. geol. miner. Bras. 1, 1-353. cocks, l. r. m. 1978. A review of British Lower Palaeozoic brachiopods including a synoptic revision of Davidson’s monograph. Palaeontogr. Soc. [Monogr.]. 131, 1-356. COPPER: EVOLUTION OF THE LISSATRYPIDAE 863 COOPER, G. A. 1956. Chazyan and related brachiopods. Smithson, misc. Colins. 127, 1-1024 (plus atlas). 1976. Early Middle Ordovician of United States. In bassett, m. g. (ed.). The Ordovician Systenu 1-696. University of Wales Press and National Museum of Wales, Cardilf. and KINDLE, c. H. 1936. New brachiopods and trilobites from the Upper Ordovician of Perce, Quebec. J. Paleont. 10, 348-372. COPPER, p. 1967. Brachidial structures of some Devonian atrypid brachiopods. Ibid. 41, 1 176 1 183. 1973. Bifida and Kayseria (Brachiopoda) and their affinity. Palaeontology, 16, 1 17 138. 1977. Zygospira and some related Ordovician and Silurian atrypoid brachiopods. Ibid. 20, 295-335. 1978. Devonian atrypoids from western and northern Canada. Spec. Pap. geol. /I55. Can. 18, 289- 331. 1982. Early Silurian atrypoids from Manitoulin Island and Bruce Peninsula, Ontario. J. Paleont. 56, 68- 702. DAHMER, G. 1922. Studieii fiber die Fauna des oberharzer Kahlebergsandsteins \\. Jb. preass. geol. Landesanst . 40,161-306. DAVIDSON, T. 1865. A monograph of the British Fossil Brachiopoda, The Devonian Brachiopoda. Palaeontogr. Soc. [Monogr.]. 6(1), 1-131. 1869. A monograph of the British fossil Brachiopoda, The Silurian Brachiopoda. Ibid., 7(3), 169-248. 1881f/. On genera and species of spiral-bearing Brachiopoda, from specimens developed by the Rev. Norman Glass. GeoL Mag. 8(4), 1-12. 18816. Descriptions of new Upper Silurian Brachiopoda from Shropshire. GeoL Mag. 8(4), 145-156. — 1882. Supplement to the British Devonian Brachiopoda. Palaeontogr. Soc. [Monogr.]. 5(1), 1-62. 1883. Supplement to the British Silurian Brachiopoda. Ibid. 63-134. DREVERMANN, F. 1901. Die Fauiia des oberdevonischen TullFreccie von Langenaubach bei Haiger. Jh. preiiss. geol. Landesanst. 21, 99-201 . EMMONS, E. 1842. Geology of New York, Part II, comprising the survey of the second geological district, 437 pp. White and Visscher, Albany, New York. 1855. American geology, containing a statement of the principals of the science with full illustrations of the characteristic American fossils, 251 pp. Sprague and Co., Albany, New York. FOERSTE, A. F. 1893. Fossils of the Clinton Group in Ohio and Indiana. GeoL Ohio, 7, 516-601. 1914. The Rogers Gap fauna of central Kentucky. J. Cincinn. Soc. nat. Hist. 21, 109- 148. FRECH, F. 1887. Ueber das Devon der Ostalpen, nebsl Bemerkungen fiber das Silur und einem palaontolo- gischen Abhang. Z. dt. geol. Ges. 39, 659-728. FU, Li-pu, 1982. Brachiopods. In Paleontological atlas of Northwest China, Shaanxi-Gansu-Ningxia Volume, Part I, 95-179. Xian Inst. Geol. Miner. Resources, China. GAGEL, c. 1890. Die Brachiopoden der cambrischen und silurischen Geschiebe im Diluvium der Provinzen Ost- und West Preussen. Beitr. Naturk. preuss. 6, 1 -79. GARCIA-ALCALDE, J. L. 1973. Braquiopodos Devonicos de la Cordillera Cantabrica, 5, El aparato braquial de Davidsonia Bouchard-Chantereaux 1849 (Strophomenida-Davidsoniacea). Breviora geol. astur. 17, 1 -5. GUNiA, T. 1962. Fauna of the limestone pebbles and the problem of the age of the Witoszow conglomerates (Lower Silesia). Ann. Soc. geol. Pologne 32, 493-522. 1968. On the fauna, stratigraphy and conditions of sedimentation of the Upper Devonian in the Swie- bodicze depression (Middle Sudeten). Geologica sudet. 4, 1 15-220. HALL, J. 1847. Containing descriptions of the organic remains of the Lower Division of the New York System. Rep. N.Y. St. Mus. nat. Hist. 1-338. 1852. Containing descriptions of the Lower Middle Division of the New York System. Ibid., 1-362. and CLARKE, J. M. 1894. An introduction to the study of the genera of Palaeozoic Brachiopoda. Rep. St. Geol N.Y. 13, 601-943. HAVLiCEK, V. 1973. New brachiopod genera in the Devonian of Bohemia. Vestn. listr. Ust. geol 48, 337-340. — 1984. Buceqia, a new genus of the Lissatrypidae (Brachiopoda, Silurian). Vestn. ustr. Ust. geol. 59, 109 no. 1985. Karhous, g. n. (Lissatrypidae, Brachiopoda: Devonian) in Bohemia. Vestn. ustr. Ust. geol. 60, 235- 239. and PLODOWSKi, G. 1974. Uber die systematische Stellung der Gattung Hircinisca (Brachiopoda) aus dem bohmischen Ober-Silurium. Senck. leth. 55, 229-249. HEiDENHAiN, F. 1869. Ueber graptolithenffihrende Diluvial-Geschiebe der norddeutschen Ebene. Z. dt. geol. Ges. 21, 143-182. HERITSCH, F. 1929. Faunen aus dem Silur der Ostalpen. Ahh. geol. Bundesanst., Wien. 23, 1 -183. 864 PALAEONTOLOGY, VOLUME 29 JAANUSSON, V. 1979. Ordovician. In teichert, c. (ed.). Treatise on invertebrate paleontology. Part A, A136- A166. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. JOHNSON, j. G., BOUCOT, A. j. and MURPHY, M. A. 1976. Wenlockian and Ludlovian age brachiopods from the Roberts Mountain Formation of central Nevada. Univ. Calif. Pubis geol. Sci. 115, 1-102. KAY, M. 1933. The Ordovician Trenton Group in northwestern New York; stratigraphy of the lower and upper limestone formations. Am. J. Sci. 26, 1-15. 1937. Stratigraphy of the Trenton Group. Bull. geol. Soc. Am. 48, 233-302. KAYSER, E. 1872. Studien aus dem Gebiete des rheinischen Devon, III, Die Fauna des Rotheisensteins von Brilon in Westfalen. Z. dt. geol. Ges. 24, 653-690. KULKOV, N. p. 1974. Brachiopods. In Rugosans, brachiopods and stratigraphy of the Silurian of the Altai- Sayan mountain region. Trudy Inst. Geol. Geofiz. sib. Otd. 231, 1-121. [In Russian.] KUNTH, A. 1865. Die losen Versteinerungen im Diluvium von Tempelhof bei Berlin. Z. dt. geol. Ges. 17, 311- 332. LENZ, A. c. 1970. Late Silurian brachiopods of Prongs Creek, northern Yukon. J. Paleont. 44, 480-500. \911a. Upper Silurian and Lower Devonian brachiopods of Royal Creek, Yukon, Canada. Palaeonto- graphica A 159, 111-138. 19776. Llandoverian and Wenlockian brachiopods from the Canadian Cordillera. Can. J. Earth. Sci. 14, 1521-1554. LIBERTY, B. A. 1971. Paleozoic geology of Wolfe Island, Bath, Sydenham and Gananoque map-areas, Ontario. Geol. Surv. Pap. Can. 70-35, 1-12. LIU DiYONG, ZHANG zixiN and Di QiAOLiNG. 1984. Middle Ordovician brachiopods from Mts. Kunlun and Altun. Acta Palaeont. Sin. 23, 155-169. LOCKLEY, M. G. 1980. The Caradoc faunal associations of the area between Bala and Dinas Mawddwy, north Wales. Bull. Br. Mus. nat. Hist. (Geol.) 33, 167-235. M‘cOY, F. 1851. On some new Cambro-Silurian fossils. Ann. Mag. nat. Hist. 8, 387 409. 1855. A systematic description of the British Palaeozoic fossils. In sedgwick, a. and m‘coy, f., A synopsis of the classification of the British Palaeozoic rocks, 661 pp. J. W. Parker, West Strand, London. MacKiNNON, D. I. The shell structure of spiriferide Brachiopoda. Bull. Brit. Mus. nat. Hist. (Geol.). 25, 189- 261. mailleux, e. 1936. La faune des schistes de Matagne. Mem. Inst. r. Sci. nat. Belg. 77, 1-75. MAURER, F. 1881. Palaontologischen Studien im Gebiet des rheinischen Devon, 4, Der Kalk bei Greifenstein. Neues Jb. Miner. Geol. Paldont. 1, 1-112. 1885. Die Fauna des Kalkes von Waldgirmes bei Giessen, 340 pp. Bergstrasser, Darmstadt. MENAKOVA, G. N. 1983. Zeravshania—a new Paleozoic brachiopod genus from central Tadzhikistan. Paleont. Zh. 1983(1), 66-73. MEYER, H. V. and MUNSTER, G. G. 1840. Die Versteinerungen des Uebergangskalkes mit Clymenien und Ortho- ceratiten von Oberfranken. Beitr. Petrefactenkunde, 3, 33-121. MITCHELL, w. I. 1977. The Ordovician Brachiopoda from Pomeroy, Co. Tyrone. Palaeontrog. Soc. [Monogr.]. 130, 1-138. NIKIFOROVA, o. I. 1937. Faunal characteristics of the Upper Silurian of western Pribalkhash. Tsent. nauchno- issled. geol. Razv. Inst. (TSNIGRI), 1 1-36. [In Russian.] 1941. Some brachiopods of the upper Silurian from the Kheta and Khandyga river basins (Siberia). Trudy arkt. nauchno-issled. Inst. 158, 103-120. 1961 . Paleontology. In Nikiforova, o. i. and andreeva, a. n. Stratigraphy of the Ordovician and Silurian of the Siberian Platform and its paleontological basis. Trudy Vses. nauchno-issled. geol. Inst. ( VSEGET). 1, 67-290. [In Russian.] and MODZALEVSKAYA, T. L. 1968. Some Llandoverian and Wenlockian brachiopods from the northwest part of the Siberian Platform. Uchen. Zap. nauchno-issled. Inst. geol. Arkt. 21, 50-81. [In Russian.] PARKS, w. A. 1915. Palaeozoic fossils from a region southwest of Hudson Bay. Trans, r. Can. Inst. 11, 1-95. PERRY, D. G. 1985. Brachiopoda and biostratigraphy of the Siluro-Devonian Delorme Formation in the District of Mackenzie. Contr. r. Ont. Mus. 138, 1-243. RAYMOND, p. E. 1911. The Brachiopoda and Ostracoda of the Chazy. Ann. Carneg. Mus. 7, 215-259. REED, f. r. c. 1897. The fauna of the Keisley Limestone, II. Q. Jl geol. Soc. Land. 53, 67-106. 1917. The Ordovician and Silurian Brachiopoda of the Girvan District. Trans, r. Soc. Edinb. 51, 795- 998. 1936. The lower Palaeozoic faunas of the Southern Shan states. Mem. geol. Surv. India, 21, 1-183. COPPER: EVOLUTION OF THE LISSATRYPIDAE 865 ROEMER, F. 1885. Lethaia erratica oder Aufzahlung und Beschreibung der norddeutschen Ebene vorkom- menden Diluvial-Geschiebe nordischer Sedimentar-Gesteine. Paldont. Abh. 5, 1 173. RONG JiAYU, XU HANKUi and YANG XUECHANG. 1974. Silurian brachiopods, 195-208. In Handbook of strati- graphy and palaeontology of SW China, 454 pp. Academy of Science, Nanjing. [In Chinese.] ROY, s. K. 1941. The Upper Ordovician fauna of Frobisher Bay, Baffin Island. Mem. Field. Mas. nat. Hist. 2, 1-212. ROOMUSOKS, A. 1964. Some brachiopods from the Ordovician of Estonia. Tartu Ulik. Geol.-Inst. Toim. 153, 3-28. ROZMAN, K. s. 1964. Middle and Late Ordovician brachiopods of the Selennyakh range. Trudy Inst. geol. Nauk, Mos. 106, \06-204. 1968. Brachiopods of the Selennyakh and Sette-Daban mountain ranges. In balashov, e. g. (ed.). Field atlas of the Ordovician fauna of NE USSR. 53-76. Akad. Nauk Moskva. RUBEL, M. 1977. Revision of dayiacean brachiopods from the Silurian of the NE Baltic. Geol.-Inst. Toim. 26, 211-220. RUDWiCK, M. J. s. 1960. The feeding mechanisms of spire-bearing fossil brachiopods. Geol. Mag. 97, 369-383. RYBNIKOVA, M. V. 1967. Brachiopoda. In galite, l. k. et al.. Stratigraphy, fauna, and environmental conditions of the Silurian rocks of the centra! Baltic, 304 pp. Minist. Geol. SSSR Inst. Geol. Riga. [In Russian.] RZHONSNITSKAYA, M. A. I960. Order Atrypida. In orlov, y. a. (ed.). Osnovy paleontologii: Bryozoa and Brachiopoda, 1-343. Nedra Press, Moskva. [In Russian.] 1964. On Devonian atrypids of the Kuznetsk Basin. Paleontologiya i stratigrafiya, Trudy Vses. nauchno- issled. geol. Inst. 93, 91-112. SALTER, J. w. 1873. A catalogue of the collection of Cambrian and Silurian fossils contained in the geologiccd museum of the university of Cambridge, 204 pp. Cambridge University Press, Cambridge. SAMTLEBEN, c. 1972. Fcinbau und Wachstum von Spiriferiden-Armgeriisten. Paldont. Z. 46, 20-33. SARDESON, F. W. 1892. The range and distribution of the Lower Silurian fauna of Minnesota with descriptions of some new species. Bulk Minn. Acad. nat. Sci. 3, 326-343. SCHUCHERT, c. 1894. A revised classification of the spire-bearing Brachiopoda. Am. Geol. 13, 102-107. and COOPER, G. a. 1930. Upper Ordovician and Lower Devonian stratigraphy and paleontology of Perce, Quebec. Am. J. Sci. 20, 162-288. SHENG HUAIBIEN. 1975. Silurian Atrypella from Quangyuan, Sichuan. Prof. Pap. Strat. Paleont. 2, 78-79. [In Chinese.] siEHL, A. 1962. Der Greifensteinerkalk (Eiflium, Rheinisches Schiefergebirge) und seine Brachiopodenfauna. Palaeontographica A119, 173-221. sowERBY, J. de c. 1839. Organic remains. In Murchison, r. i.. The Silurian System, 579-712. John Murray. SPRiESTERSBACH, J. and FUCHS, A. 1909. Die Fauna der Remscheiderschichten. Abhandl. preuss. geol. Landesanst. 58, 1-81. STRUSZ, D. L. 1982a. On Australina Clarke and its junior synonyms Lissatrypa, Lissatrypoidea, and Tyrothyris (Silurian-Devonian Brachiopoda). J. Aust. Geol. Geophys. 7, 12-11 . 1982fi. Wenlock brachiopods from Canberra, Australia. Alcheringa, 6, 105- 142. SU YANGZHENG. 1977. Cambro-Devonian Brachiopoda. In Paleontological atlas of northeast China, 254-327. Shenyang Institute of Geology and Mineral Resources, Geol. Publ. House, Beijing. [In Chinese.] SWEET, w. c. and Bergstrom, s. m. 1976. Conodont biostratigraphy of the Middle and Upper Ordovician of the United States midcontinent. In bassett, m. g. (ed.). The Ordovician System, 121 152. University of Wales Press and National Museum of Wales, Cardiff. TEMPLE, J. T. 1968. The lower Llandovery (Silurian) brachiopods from Keisley, Westmorland. Palaeontogr. Soc.[Monogr.].\22,\-5%. TiETZE, E. 1871. Ueber die devonischen Schichten von Ebersdorf umweit Neurode in der Grafschaft Glatz, eine geognostisch-palaontologische Monographie. Palaeontographica, 19, 103-158. TITUS, R. and cameron, b. 1976. Fossil communities of the lower Trenton Group (Middle Ordovician) of central and northwestern New York state. J. Paleont. 50, 1209-1225. TORLEY, K. 1934. Die Brachiopoden des Massenkalkes der oberen Givetstufe von Bilveringsen bei Iserlohn. Abh. senckenb. naturforsch. Ges. 43, 67-148. twenhofel, w. h. 1914. The Anticosti Island faunas. Mus. Bull. Can. geol. Surv. 3, 1-35. VENYUKOV, p. N. 1899. Die Fauna der silurischen Ablagerungen des Gouvernements Podolien. Mater. Geol. Ross. 19, 21-266. VLADIMIRSKAYA, Y. v. 1972. Systematic position and geologic range of the genus Tuvaella (Brachiopoda). Paleont. Z. 1972(1), 37-44. 866 PALAEONTOLOGY, VOLUME 29 WELLER, s. 1903. The Paleozoic faunas. Rep. geol. surv. New Jers. 3, I -462. WHiTEAVES, J. F. 1904. Preliminary list of fossils from the Silurian (Upper Silurian) rocks of the Ekwan River and Sutton Mill lakes, Keewatin. Ann. Rep. geol. Surv. Can. 14 (F), 38-59. — 1906. The fossils of the Silurian (Upper Silurian) rocks of Keewatin, Manitoba, the northeastern shore of Lake Winnipegosis and the lower Saskatchewan river. Rep. geol. Surv. Can. Palaeozoic Fossils. 8, 243- 298. WILLIAMS, A. 1951. Llandovery brachiopods from Wales with special reference to the Llandovery District. Q. Jl geol. soc. Land. 107, 85- 1 36. 1962. The Barr and Lower Ardmillan Series (Caradoc) of the Girvan District, southwest Ayrshire, with descriptions of the Brachiopoda. Mem. geol. Soc. Lond. 3, 1-267. and HURST, j. 1977. Brachiopod evolution. Dev. Paleont. Strut. 5, 79 121. and WRIGHT, A. D. 1961. The origin of the loop in articulate brachiopods. Palaeontology, 4, 149-176. WILSON, A. E. 1932u. Ordovician fossils from the region of Cornwall, Ontario. Trans, r. Soc. Can. 3(26), 373- 404. - 1932b. Palaeontological notes. Can. Fid Nat. 46, 133-140. 1946. Brachiopoda of the Ottawa Formation of the Ottawa-St. Lawrence Lowland. Bull. geol. Surv. Can. 8, 1-149. wiNCHELL, N. H. and SHUCHERT, c. 1892. Preliminary descriptions of new Brachiopoda from the Trenton and Hudson River groups of Minnesota. Am. Geol. 9, 284-294. 1893. The Lower Silurian Brachiopoda of Minnesota. Geology of Minnesota, 3, 333-474. WRIGHT, A. D. 1979. The origin of the spiriferidine brachiopods. Letliaia, 12, 29-33. XU HANKUI. 1979u. Brachiopoda. In Paleontological atlas of northwestern China. I, Qinghai, 60-112. Beijing Geological Publishing House, China. 1979b. Brachiopods from the Tangxiang Formation (Devonian) in Nandan, Guangxi. Acta Paleont. Sin. 18, 362-380. ZENG QiNGLUAN. 1983. Latest Ordovician and early Silurian faunas from the eastern Yangtze Gorges, China, with comments on the Ordovician-Silurian boundary. Bull. Yichang Inst. Geol. Min. Res. 6, 95-127. PAUL COPPER Department of Geology Laurentian University Typescript received 8 March 1984 Sudbury, Ontario Revised typescript received 5 August 1985 Canada P3E 2C6 NOTE ADDED IN PROOF After this manuscript went to press, three further papers, citing new genera, species, or revised designations, were received. These are relevant to the data presented: Terehratiila turjensis Gruenewaldt, 1854, pi. 2, hg. 8a-r/. Emsian-Eifelian beds, eastern slopes of the Urals, USSR. Mizens (1984) relegates this to the genus Holynatrypa, but serial sections provided indicate this is probably the glassiid genus Karhous Havlicek. Lissatrypella Sapelnikov and Mizens, 1982 (type Atrypa kuschvensis Chernyshev, 1893), Upper Wenlock to Pridoli, Urals and Central Asia. This is like the genus Atrypoidea but has a wide anterior fold giving a resemblance to some Septatrypa. It may be more suitable to regard this as a subgenus variant of Atrypoidea considering the variation in development of the anterior fold. Lissatrypa (Nanatrypa) Sapelnikov and Mizens, 1982 (type Atrypa canaliculatiformis Chernyshev, 1893), Wenlock -Ludlow, eastern Urals, USSR. This is a subgenus with a marked sulcus on both valves. No internal structures are shown, but there is a strong similarity to the bisinuosity seen in many Glassia, to which it may possibly be assigned if spiralia are shown to be medially directed. Aulidospira chuxianensis Liu, 1983, p. 285, pi. 93, hgs. 15-16. Tangtao Formation, Upper Ordovician, Zhe County, Anhui Province, China. Liu diagnoses this species with a convex horizontal, ventral platform. No internal illustrations are shown and this species is tentatively assigned to Cyclospira. COPPER; EVOLUTION OF THE LISSATRYPIDAE 867 REFERENCES CHERNYSHEV, T. 1893. Die Fauna des unteren Devon am Ostabhange des Ural. Trudy Geol. Korn. 4(3), 1 221. GRUENEWALDT, M. V. 1854. Ueber die Versteinerungen des silurischen Kalke von Bogosslowsk. Ein Beitrag zur Geologie des oestlichen Ural. Mem. Savants Etrang. 7, 571 620. LIU DIYONG, XU HANKUi and L IANG WANGPING. 1983. Bracliiopoda. In Paleontological atlas of east China, 254- 286. Beijing Geological Publishing House, China. [In Chinese.] MiZENS, L. I. 1984. Lower Devonian and Eifelian atrypids from the eastern slopes of the Urals. Akad. Nauk SSSR, Uralskii Nauchnyi Tsentr, 1-111. SAPELNiKOV, V. p. and MIZENS, L. I. 1982. Smooth Silurian atrypoids from the eastern slopes of the Central and Northern Urals. Ibid. 1-51. [In Russian, labelled ‘preprint’.] ERRATUM Palaeontology, Vol. 29, Part 2, 1986, p. 303: The senior author’s name should read Malinky not Malinkey. . ' '■^ ' t- i I Ti- ■’■ ' '■^m THE PALAEONTOLOGICAL ASSOCIATION Annual Report of Council for 1985 Membership and Subscriptions. Membership totalled 1,378 on 31 December 1985, a decrease of 42 over the previous year. There were 875 Ordinary Members, a decrease of 57; 68 Retired Members, an increase of 20; 1 17 Student Members, a decrease of 6; and 318 Institutional Members, an increase of 1. Total individual and institutional subscriptions to Palaeontology through Marston’s agency numbered 427, a decrease of 8. Subscriptions to Special Papers in Palaeontology numbered 1 13 individuals, a decrease of 10; and 128 Institu- tions, a decrease of 3. Standing orders through Marston’s agency for Special Papers numbered 78 for Special Paper 31 and 41 for Special Paper 32. Sales of back parts of Palaeontology via the Membership Treasurer realized £258 05. Sales of back parts of Palaeontology to Institutional Members yielded US S32.00. Sales of back numbers of Special Papers in Palaeontology to individuals yielded £1,288-58, and to Institutions, £1 3000. Sales of 12 copies of the Atlas of the Burgess Shale through the Marketing Manager yielded £128-50 plus US S90.00, and sales of Fossil Plants of the London Clay through him numbered 70 copies, yielding £447-70 and US S38.50. Sales of Atlas of Invertebrate Macrofossils under the special offer to members of the Association numbered 127, yielding £1,209. Finance. During 1985 the Association published Volume 28 of Palaeontology at an estimated cost of £52,209 (including postage and distribution). Special Papers 33 and 34 were published at a cost of £14,419. The Association is grateful to all those who made donations to offset the cost of publishing Palaeontology and Mrs Sylvester-Bradley who donated £200 to the Sylvester-Bradley Fund. Publications. Volume 28 of Palaeontology, for 1985, published in 4 parts, contained 810 pages and 91 plates. Special Papers 33 (Evolutionary case histories from the fossil record edited by J. C. W. Cope and P. W. Skelton), and 34 (Review of the Upper Silurian and Lower Devonian articulate brachiopods of Podolia by O. I. Nikiforova, T. L. Modzalevskaya, and M. G. Bassett) were published in August. Meetings. Nine meetings were held in 1985. The Association is indebted to the organizers, hosts, and field leaders of these. a. Joint meeting with the Geological Society, on ‘Fossils and Tectonics’, held on 13 February, at Burlington House, Piccadilly, London. About 80 attended the meeting, which was convened by Dr L. R. M. Cocks and Dr R. A. Fortey. Papers are to be published in the Journal of the Geological Society. b. Review Seminar on ‘Graptolites’, held on 27 February at the British Museum (Natural History). About 30 attended the meeting, convened by Dr T. J. Palmer. c. Twenty-Eighth Annual General Meeting, held in the Lecture Theatre of the Geological Society of London on 1 5 March. Professor Bruce Runnegar delivered the Annual Address on ‘Molecular Palaeontology’. The Sylvester-Bradley Award was made to Mr D. M. Blake. d. Eield Meeting, organized by the Carboniferous Group, to Gower, South Wales, led by Dr A. T. S. Ramsay, and held on 19-22 April. e. Progressive Palaeontology Meeting, held at University College, Cardiff, on 29 May. About 60 attended the meeting, which was convened by Miss Candace Brooke. /. Third International Congress of Systematic and Evolutionary Biology, included a one day symposium, jointly sponsored by the Association and the Linnean Society, on ‘Random and Directed Processes in Evolution’, held on 9 July, at the University of Sussex. The convener for the symposium on behalf of the Association was Professor A. Hallam. g. Eield Meeting on the Jurassic of Dorset, held on 27-29 September, and organized and led by Dr T. J. Palmer and Dr J. C. W. Cope. Twenty-six attended the meeting. h. Review Seminar on ‘Ancient and Recent Reefs’, held on 13 November at University College, Cardiff. About 120 attended. The convener was Dr R. Riding. 870 THE PALAEONTOLOGICAL ASSOCIATION i. The Annual Conference, held at the University College of Wales, Aberystwyth on 19-21 December comprised an open meeting followed by a field trip to the Lower Palaeozoic of central Wales and the Welsh Borderlands, led by Dr D. E. B. Bates and Dr D. Campbell. The Conference was attended by 138 people, and the Local Secretary was Dr T. J. Palmer. The President’s Award was made to Miss R. A. Wood. Council. The following members served on Council following the Annual General Meeting on 1 5 March 1985: President: Professor C. Downie; Vice-Presidents: Dr M. G. Bassett, Dr R. Riding; Treasurer: Dr M. Romano; Membership Treasurer: Dr A. T. Thomas; Secretary: Dr P. W. Skelton; Marketing Manager: Dr R. J. Aldridge; Institutional Membership Treasurer: Dr A. R. Lord (co-opted); Editors: Dr D. E. G. Briggs, Dr E. B. Halstead, Dr R. Harland. Dr P. R. Crowther, Dr T. J. Palmer, and Dr D. Edwards (co-opted); Circular Reporter: Dr D. J. Siveter; Public Relations Officer: Dr M. J. Benton; Other Members: Dr M. E. Collinson, Dr P. L. Forey, Dr A. W. Owen, Dr C. R. C. Paul, Dr A. B. Smith, and Professor T. N. Taylor. Circulars. Four Circulars, numbers 1 19-122, were distributed to Ordinary, Student, and Retired Members, and, on request, to over 90 Institutional Members. Council Activities. Council created a new post from among its membership, of Public Relations Officer, the first occupant of the post being Dr M. J. Benton. During the course of the year, the Editors produced new ‘Instructions for Authors’, published in Palaeontology, Volume 28, part 4. One important change affecting future contributors to Association publications is that authors will henceforth be requested to transfer copyright to the Association. This is intended simply to enable the Association to handle copyright matters more efficiently and effectively, and will not diminish in any way authors’ control over the use of this material. A form for copyright assignment, with explanatory notes, has been prepared by the Editors. A special offer to individual members for the Atlas of Invertebrate Macrofossils at a reduced price was successful. Participation by the Association in the I.C.S.E.B. Third meeting (see meeting/, above), with support towards the speakers’ expenses provided by the Natural Environment Research Council, produced a well-attended and successful symposium. Plans for the future include a large joint meeting with the Geological Society— the first Lyell Meeting— (2 days), on ‘Gondwana-Tethys’, from which a publication is intended, and the first publication by the Associa- tion of a microcomputer software package, designed for palaeontological work. A joint workshop/seminar meeting is also to be held with the Malacological Society. BALANCE SHEET AND ACCOUNTS EOR THE YEAR ENDING DECEMBER 1985 Balance Sheet as at 31 December 1985 1984 £ £ £ 50,459 Investments AT Cost (see schedule) Current Assets 2,003 Sundry Debtors ......... 2,000 29,630 Cash at Bank ......... 45,363 1,816 Sylvester-Bradley Fund ....... 2,029 33,449 49,392 Current Liabilities 2,337 Subscriptions Received in Advance ..... 5,81 1 Provision for cost of publication of: 13,800 Palaeontology 2^14 ........ 13,500 13,450 Special Paper 32 ........ — 2,509 Sundry Creditors ......... 3,025 2,000 Loan from Royal Society ....... 1,000 34,096 23,336 (647) £49,812 Represented by; Publications Reserve Account 44,135 Balance Brought Forward ....... 45,365 Excess of Income over Expenditure for the Year Transferred 1,230 from Income and Expenditure Account .... 24,552 45,365 Sylvester-Bradley Fund 1,912 Balance Brought Forward ....... 1,816 Donation .......... 200 104 Interest .......... 213 (200) Grant Awarded ......... (200) 1,816 Balance Carried Forward ....... 2,631 Meeting Reserve £ 48,521 26,056 £74,577 69,917 2,029 2,631 £49,812 £74,577 Income and Expenditure Account 1984 £ 41,152 26,207 9,496 1,011 1,468 779 3,917 8,947 99 INCOME Subscriptions 1985 1984 Palaeontology Sales ...... Donations Special Papers Sales Donations Burgess Shale Portfolio Sales ...... Postage/Stationery Fossil Plants of the London Cla y Sales ...... Postage/Stationery A TEAS OF Inver terra te Ma crofossils Sales Postage/Stationery Offprints Profit on Sales of I nvestments Investment Income (see schedule) Sundry Income .... £ £ 44,003 750 44,753 30,070 1,971 32,041 8,853 8,853 497 (14) 483 710 (14) 696 1,963 (163) 1,800 2,346 62 11,602 235 £93,076 £102,871 FOR THE Year Ended December 1985 EXPENDITURE 1984 £ 54,002 23,755 3,992 1,919 100 4,055 4,023 £91,846 Cost of Publication of Palaeontology Volume 28 — Part 1 Part 2 Part 3 ... Part 4 (provisional) Under Provision for Volume 27, Part 4 Cost of Publication of Special Papers No. 33 No. 34 Under Provision for No. 32 Cost of Publication of: Fossil Plants of the London Clay Cumulative Index Atlas of Invertebrate Macrofossils Warehousing of Publications Grants Cost of Circulars Preparation .... Postage .... Credit ..... Administrative Costs Postage and Stationery Editorial Expenses Meeting Expenses . Audit Fee Membership of Societies 13,127 13,244 13,302 13,256 769 10,325 4,094 175 2,677 1,352 (310) 1,059 84 2,762 250 60 £1,230 Excess of Income over Expenditure for the Year Transferred to Publications Reserve Account £ 53,698 14,594 2,093 3,719 4,215 £78,319 £24,552 Schedule of Investments and Investment Income as at 31 December 1985 Gross Income Cost for Year £12,000 13|% Exchequer Stock 1987 ........ JL 11,520 t 1,590 £1,000 9% Treasury Stock 1 992/ 1996 992 90 £1.000 9%Treasury Stock 1994 995 90 £4,000 8% Treasury Stock 2002/2006 2,192 320 £5,357 1 3i%Treasury Stock 1997 5,000 710 £3,280 1 3^% Exchequer Stock 1996 3,000 435 £2,000 Agricultural Mortgage Corporation Ltd. Debenture 1980/1985 — sold March 1985 .......... — 92 5,270 M. & G. Charifund Units ......... 4,073 1,143 10,000 New Throgmorton Trust ( 1983) p.l.c. 25p Income Shares 1,706 386 700 Clarke, Nicholls& Coombs p.l.c. 25p Shares ..... 668 63 £6,180 M.E.P.C. p.l.c. 6|% Convertible Unsecured Loan Stock 1995/2000 4,943 402 374 M.E.P.C. p.l.c. 25p Shares ......... 703 48 570 National Westminster Bank p.l.c. £1 Ordinary Shares .... 2,788 214 £10,150 Agricultural Mortgage Corporation Ltd. 7|% Debenture Stock 1991/1993 8,251 787 1,865 Imperial Group p.l.c. 25p Ordinary Shares ...... 1,730 228 6,598 Bank Interest 5,004 48,521 1 1 ,602 Market VALUEat 31 December 1985(1984— £71,437) .... £74,498 Report of the Auditor to the Members of THE Palaeontological Association In my opinion, the Accounts as set out above give a true and fair view of the state of the affairs of the Association at 3 1 December 1 985 and of its income and expenditure for the year ended on that date. March 1986 Market Harborough, Leicestershire G. R. Powell Chartered Accountant INDEX Pages 1212 are contained in Part 1; pages 213 422 in Part 2; pages 423-630 in Part 3; pages 631-873 in Part 4. Figures in Bold Type indicate plate numbers. A Acanthoscaphites verneuiliamis, 74, 16 Actinostromarianina lecompti, 35 Africa: archosaur predation, 415 Algae: Silurian cyclocrinitid, 583; cysts in Palaeocene-Lower Eocene, Kurdistan, 739 America: Cretaceous fish, 365 Ammonite: Upper Maastrichtian, France, 25; Cretaceous, 725 Anapachydiscus fresvillensis, 42, 7, 8, 9 Annual Address: Runnegar, B. Molecular Palaeontology, 1 Antarctic: review of ichthyofaunas, 113; Cretaceous and Tertiary wood, 665 Arthropod: Silurian, 627 Arthrorhachis ci. tarda. 747, 58 Ausich, W. I. Palaeoecology and history of the Calceocrini- dae (Palaeozoic Crinoidea), 85 B Baarli, B. G. A biometric re-evaluation of the Silurian brachiopod lineage Stricklandia lensjS. laevis. 187 Bacidites anceps. 58, II, 12; vertehralis. 57, 11, 12 Balanocrinu.s gracilis. 36 Balizoma ohtusus. 56 1 , 40, 48, 49 Bancroft, A. J. Ovicells in the Palaeozoic bryozoan Order Fenestrata. 155; Secondary nanozooecia in some Upper Palaeozoic fenestrate Bryozoa, 207 Beadle, S. C. and Johnson M. E. Palaeoecology of Silurian cyclocrinitid Algae, 583 Bell. C. M. and Boyd M. J. A tetrapod trackway from the Carboniferous of Northern Chile, 519 Benton. M. J. The late Triassic reptile Teratosaurus—a raui- suchian, not a dinosaur, 293 Birmanites aff. asiaticus. 754, 59; sp. indet., 754, 59 Bodacriniis columnus gen. et sp. nov., 236, 23 Bolton. T. E. See Dixon, O. A., Bolton. T. E. and Copper, P. Bosence, D. See Carthew, R. and Bosence, D. Boyd, M. J. See Bell, C. M. and Boyd, M. .1. Brachiopoda: Silurian, 187; Evolution of the Lissatrypidae, 827 Britain: crinoid. Permian. 809; see also England, Scotland, Wales, and Ireland Broadhead, T. R. See Malinky, J. M., Mapes. R. H. and Broadhead, T. R. Bryozoa: Palaeozoic, 155, 207 Bidhaspis sp. indet., 773, 63 Bumhanius gen. nov., 269 C Cambrian: Burgess Shale, 423 Carboniferous: tetrapod trackway in Chile, 519; amphibian, Scotland, 601 Carle. K. J. See Mitchell. C. E. and Carle, K. J. Carthew. R. and Bosence, D. Community preservation in Recent shell-gravels, English Channel, 243 Cliandlerichihys sirickeri gen. et sp. nov., 368 Chile: Carboniferous tetrapod trackway, 519 China: Trilobite, Ordovician, 743 Chinzei, K. Shell structure, growth, and functional mor- phology of an elongate Cretaceous oyster, 139 Chladocrinus rohustus. 36 Cocco.seris'l escanabaense, 34 Conway Morris, S. The community structure of the Middle Cambrian Phyllopod Bed (Burgess Shale). 423 Cooper, R. A. and Ni Yunan Taxonomy, phylogeny, and variability of Pseudi.sograptus Beavis, 313 Cooper. R. A. See Fortey, R. A. and Cooper. R. A. Copper. P. Evolution of the earliest smooth spire-bearing atrypoids (Brachiopoda: Lissatrypidae, Ordovician Silurian). 827 Copper, P. See Dixon, O. A., Bolton, T. E. and Copper, P. Coral: Upper Ordovician heliolitid. 391 Cretaceous: ammonite fauna, France, 25; functional mor- phology of an oyster, 139; freshwater fish. North America, 365; wood, Antarctica, 665; cobblc-dwcllers, 691 ; ammon- ite, Nigeria, 725 Crinoidea: Palaeozoic palaeoecology, 85; Ordovician, Sweden, 235; contrasting lifestyles in Lower Jurassic crin- oids, 475; Permian, 809 Cruickshank, A, R. I. Archosaur predation on an east Afri- can M iddle T riassic dicynodont. 4 1 5 Cummins. H.. Powell, E. N.. Stanton, R. J. Jr. and Staff, G. The size-frequency, distribution in palaeoecology: effects of taphonomic processes during formation of molluscan death assemblages in Texas bays, 495 Cyclopyge cf. recurva. 758, 61 Cyclospira hisidcatu. 849, 74 Cyphoniscus cf. socialis. 778, 65 'Cythathoermites' ramosus. 814, 71, 72 D Darwiniies grafordensis gen. et sp. nov., 306 Dashzeveg, D. See Russell, D. E, and Dashzeveg, D. Dean, W. T, See Zhou Zhiyi and Dean. W. T. Dehortiella crus Ians, 35; sp. nov.. 35 Diplochaeietes mexicanus sp. nov. 578, 50 Diplomocera.s cylindraceum. 51, 4, 9, 10 Dixon, O. A., Bolton, T. E. and Copper, P. Ellisites. an Upper Ordovician heliolitid coral intermediate between coccoserids and proporids, 391 Diversification: Phanerozoic, 655 Donovan. S. K. A new genus of inadunate crinoid with unique stem morphology from the Ashgill of Sweden, 235 876 INDEX Donovan, S. K., Hollingworth, N. T. J. and Veltkamp, C. J. The British Permian crinoid 'Cyathocrinites' ramosus (Schlotheim), 809 Dualilesl speleana, 31 E Eastman, J. T. See Grande, L. and Eastman, J. T. Elliott, G. E, Isolated algal cysts in the Palaeocene-Lower Eocene of Kurdistan, 739 Ellisites astomata, 32, 33, 34; glyptum, 34; Lahechioides sp. nov„ 393, 30, 31,32 Encrinurus (Encriimnis) intersitus sp. nov., 538, 39; (E.) cf. intersitus sp. nov., 42; (E.) jarkanderi sp. nov., 542, 40; (£.) macrourus, 544, 41, 42; (E.) nastus sp. nov., 548, 42, 43; (E.) punctatus, 534, 37, 38; (E.) oldvaldensis sp. nov., 550, 44; (E.) scliisticola, 552, 45; (E.) stuhhlefieldi, 554, 46; (£.) sp. A, 556, 47; (£.) sp. B, 558, 47; (£.) sp. C, 558, 45, 47 England: Triassic, sphenodontids, 165 Eocene: insectivores (Mammalia), Mongolia, 269; algal cysts, Kurdistan. 739 Europe: Jurassic, Triassic, Semionotidae, 213 Evitt, W. R. See Tripp, R. P. and Evitt, W. R. F Fenestella cf. fanata. 19 Fenster, E. J. See Hoffman, A. and Fenster, E. J. Fish: Triassic and Jurassic, Europe, 213; freshwater, Cretaceous, America, 365 Fortey, R. A. and Cooper, R. A. A phylogenetic classihca- tion of the graptoloids, 631 France: Upper Maastrichtian ammonite fauna, 25 Francis, J. E. Growth rings in Cretaceous and Tertiary wood from Antarctica and their palaeoclimatic implications, 665 Fraser. N. C. New Triassic sphenodontids from south-west England and a review of their classification, 165 Eresvillia constricta gen. et sp. nov., 62, 14 G Geragnostus alT. longicoUis, 748, 58 Glaphurina sp., 753, 59 Glassia elongata, 74, 75 Grande, L. The first articulated freshwater teleost fish from the Cretaceous of North America, 365 Grande, L. and Eastman, J. T. A review of Antarctic ichthyo- faunas in the light of new fossil discoveries, 113 Graptolites: review of Pseudisograptus, 313; nematula- rium of Pseudoclimacograptus scharenhergi, 373; phylo- genetic classification, 631 H Elammatocnemis kanlingensis, 776, 64; ohsoletus sp. nov., 773, 64, 65 Hardground: Cretaceous, 691 Hemitrypa hihernica, 19 Hoffman, A. and Fenster, E. J. Randomness and diversifica- tion in the Phanerozoic: a simulation, 655 Hollingworth, N. T. J. See Donovan, S. K., Hollingworth, N. T. J. and Veltkamp, C. J. Holmes, J. and Lopez, J. The disappearing peel technique: an improved method for studying permineralized plant tissues, 787 Hoplitoplacenticeras lasfresnayanum. 48. 9 Hoploscaphites constrictus, 64, 13, 14, 15; sp., 73, 13, 16 I Idiospira panderi, 843, 73, 74 Ischyrophymd? zhiqiandi sp. nov., 769, 62, 63 Isotelus giselae sp. nov., 710, 54, 56, 57; sp. A, 718, 54, 55; sp. B, 718, 54, 55, 57; sp, C, 721; sp. D, 721, 57; sp. E, 722, 54 J Johnson, M. E. See Beadle, S. T. and Johnson, M. E. Jurassic: fish, Europe, 213; comparison of benthic and pseudobenthic Isocrinida, 475 K Kennedy, W. J. The ammonite fauna of the Calcaire a Bacu- lites (Upper Maastrichtian) of the Cotentin Peninsula (Manche, France), 25 Konhostrea konho gen. et sp. nov., 140, 18 Kurdistan: Palaeocene-Lower Eocene algal cysts, 739 L Lichas cf. laciniatus, 778, 65 Lonchodomas nanus, 772, 63 Lopez, J. See Holmes, J. and Lopez, J. Lyrapygel gaoluoensis, 778, 65 M McCune, A. R. A revision of Semionotus (Pisces: Semiono- tidae) from the Triassic and Jurassic of Europe, 213 Malinky, J. M., Mapes, R. H. and Broadhead, T. R. New Late Palaeozoic Hyolitha (Mollusca) from Oklahoma and Texas, and their palaeoenvironmental significance, 303 Mammalia: Insectivores, Eocene, Mongolia, 269 Mandaodonites cavt ich nogen. et ichnosp. nov., 416 Manespira nicolleti gen. et sp. nov., 840, 73, 75 Mapes, R. H. See Malinky, J. M., Mapes, R. H. and Broad- head. T. R. Mesozoic: stromatoporoids, 469 Mexico: first Tertiary sclerosponge, 575 Microfossils: Precambrian, 101 Microparia (Quadrapyge) cliedaoensis sp. nov., 761, 61 Mitchell, C. E. and Carle, K. J. The nematularium of Pseudoclimacograptus scharenhergi (Lapworth) and its secretion, 373 Molecular Palaeontology: Runnegar, B. Twenty-eighth annual address, 1 Mollusca Palaeozoic from Oklahoma and Texas, 303 Mongolia: Eocene insectivores, 269 N Nahannia sp., 722, 55 Naranius infrequens gen. et sp. nov., 280 Ni Yunan. See Cooper, R. A. and Ni Yunan, 313 Nigeria: ammonite internal mould markings, 725 N ileus huanxianensis, 756, 59, 60 INDEX 877 O Oedolius perexigius gen. et sp. nov., 275 Ordovician: new crinoid from Sweden, 235; heliolitid coral, 391; asaphid trilobites, 705; trilobites, China, 743; evolu- tion of Lissatrypidae, 827 Ovalocephalus kelleri, 776, 64, 65 P Pachydiscus (Pachydiscus) gollevillensis, 28, 1, 2, 3, 4, 5, 11; (P.) jacquotti, 34, 5, 6; {P.) mokotihense, 40; (P.) sp., 38, 5 Palaeocene: algal cysts, Kurdistan, 739 Palaeoecology: Palaeozoic Crinoidea, 85 Palaeozoic: palaeoecology and history of Calceocrinidae, 85; Bryozoa, 155, 207; Mollusca, Oklahoma and Texas, 303 Paraphdlipsinella glohosa, 767, 62 Parastromatopora lihcini, 35 Paratiresias turkenslanicus, 780, 65 Pelecymala rohustus gen. et sp. nov., 170, 20 Pennirelepora spiiiosa, 158. 19; sp., 158, 19 Permian: Crinoid, 809 Peraspis obscura sp. nov., 757, 60, 61 Peratos arrectus gen. et sp. nov., 859, 74, 75 Phanerozoic: Randomness and diversification, 655 Phorocephala quadrata sp. nov., 751, 58 Plant: permineralized plant tissues, 787 Powell, E. N. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and StaflF, G. Precambrian: microfossils, 101 Protozya exigua, 836, 73 Pseudisograptus helhdus sp. nov., 348, 26; duniosus, 353, 25; jiangxiensis, 355, 25; mamibriatus harrisi ssp. nov., 331, 25; m. janus ssp. nov., 339, 25, 26, 27; m. koi ssp. nov., 332, 24, 25; m. manubriatus, 321, 24 Pseudoclimacograptus scharenbergi, 28, 29 Pseudostygina lepida, 762, 61, 62 R Ramskold, L. Silurian encrinurid trilobites from Gotland and Dalarna, Sweden, 527 Reitner, J. See Wood, R. A. and Reitner, J. Rorringtonia sp., 768, 62 Runnegar, B. Molecular Palaeontology. Twenty-eighth annual address. 1 Russell, D. E. and Dashzeveg, D. Early Eocene insectivores (Mammalia) from the People’s Republic of Mongolia, 269 S Scotland: Carboniferous amphibian, 601 Sedalanella segonzacae gen. et sp. nov., 741 Selden, P. A. A new identity for the Silurian arthropod Necrogammams, 627 Sigmala sigmalti gen. et sp. nov., 166, 20 Silurian: Stricklandia lensjS. laevis, 187; encrinurids, Sweden, 527; cyclocrinitid algae, 583; arthropod, 627; evolution of Lissatrypidae, 827 Simms, M. J. Contrasting lifestyles in Lower Jurassic crin- oids: A comparison of benthic and pseudopelagic Isocri- nida, 475 Semionotus bergeri, 221, 22 Siphonophycus inornalum, 17 Smithson, T. R. A new anthracosaur amphibian from the Carboniferous of Scotland. 601 Sphenodontids: Triassic, England, 165 Staff, G. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. Stanton, R. J. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. Stenopareia aff. bowmanni, 766, 61, 62; sp. indet., 766, 62 Stricklandia laevis, 21; lens intermedia, 21; /. lens, 21; /. prima, 21; /. progressa, 21 Stromatoporoids: poriferan affinities of, 469 Sweden: Ordovician, crinoid, 235; Silurian encrinurid trilo- bites, 527 T Taylor, P. D. Scanning electron microscopy of uncoated fossils, 685 Techniques: scanning electron microscopy of uncoated fos- sils, 685; disappearing peel technique for permineralized plant tissues, 787 Telepina convexa, 752, 58, 59; sp., 753, 59 Tertiary: sclerosponge, Mexico, 575; wood, Antarctica, 665 Tetrapod: Carboniferous, Chile, 519 Triassic: sphenodontids, England, 165; fish, Europe, 213; reptile, 393; archosaur predation, 415 Trilobites: Silurian, Sweden, 527; Asaphidae, Ordovician, 705; Ordovician, China, 743 ‘Triplecella diplicata' 74 Tripp, R. P. and Evitt, W. R. Silicified trilobites of the family Asaphidae from the Middle Ordovician of Virginia, 705 Tsaganius ambiguus gen. et sp. nov., 284 V Vellkamp, C. J. See Donovan, S. K., Hollingworth. N. J. T. and Veltkamp, C. J. W Wilson, E. C. The first Tertiary sclerosponge from the Americas, 575 Wilson, M. A. Coelobites and spatial refuges in a Lower Cretaceous cobble-dwelling hardground fauna, 691 Wood. R. A. and Reitner, J. Poriferan affinities of Mesozoic stromatoporoids, 469 X Xenocybel sp., 768. 62 Z Zaborski, P. M. P. Internal mould markings in a Cretaceous ammonite from Nigeria, 725 Zhang Zhongying. Solar cyclicity in the Precambrian micro- fossil record, 101 Zhou Zhiyi and Dean, W. T. Ordovician trilobites from Chedao, Gansu Province, north-west China, 743 r V J I VOLUME 29 Palaeontology 1986 PUBLISHED BY THE PALAEONTOLOGICAL ASSOCIATION LONDON Dates of Publication of Parts of Volume 29 Part 1, pp. 1-212, pis. 1- 21 Part 2, pp. 213-422, pis. 22-34 Part 3, pp. 423-630, pis. 35-50 Part 4, pp. 631-877, pis. 51-72 14 February 1986 27 June 1986 25 September 1986 12 December 1986 THIS VOLUME EDITED BY D. E. G. BRIGGS, P. R. CROWTHER, D. EDWARDS, L. B. HALSTEAD, R. HARLAND, T. J. PALMER, C. R. C. PAUL AND P. A. SELDEN Dates of Publication of Special Papers in Palaeontology Special Paper No. 35, 21 November 1986 Special Paper No. 36, 12 December 1986 © The Palaeontological Association, 1986 Printed in Great Britain at the University Printing House, Oxford by David Stanford Primer to the University CONTENTS Part Ausich, W. I. Palaeoecology and history of the Calceocrinidae (Palaeozoic Crinoidea) 1 Baarli, B. G. a biometric re-evaluation of the Silurian brachiopod lineage StrickUmdia lensjS. laevis 1 Bancroft, A. J. Ovicells in the Palaeozoic bryozoan Order Fenestrata 1 Bancroft, A. J. Secondary nanozooecia in some Upper Palaeozoic fenestrate Bryozoa 1 Beadle, S. C. and Johnson, M. E. Palaeoecology of Silurian cyclocrinitid algae 3 Bell, C. M. and Boyd, M. J. A tetrapod trackway from the Carboniferous of northern Chile 3 Benton, M. J. The late Triassic reptile Teratosaiirus- a rauisuchian, not a dinosaur 2 Bolton, T. E. See Dixon, O. A., Bolton, T. E. and Copper, P. Bosence, D. See Carthew, R. and Bosence, D. Boyd, M. J. See Bell, C. M. and Boyd, M. J. Broadhead, T. R. See Malinky, J. M., Mapes, R. H. and Broadhead, T. R. Carle, K. J. See Mitchell, C. E. and Carle, K. J. Carthew, R. and Bosence, D. Community preservation in Recent shell-gravels, English Channel 2 Chinzei, K. Shell structure, growth, and functional morphology of an elongate Cretaceous oyster 1 Conway Morris, S. The community structure of the Middle Cambrian Phyllopod Bed (Burgess Shale) 3 Cooper, R. A. See Fortey, R. A. and Cooper, R. A. Cooper, R. A. and Ni Yunan. Taxonomy, phylogeny, and variability of Pseudisograptus Beavis 2 Copper, P. Evolution of the earliest smooth spire-bearing atrypoids (Brachiopoda: Lissatryp- idae, Ordovician-Silurian) 4 Copper, P. See Dixon, O. A., Bolton, T. E. and Copper, P. Cruickshanr, a. R. F, Archosaur predation on an east African Middle Triassic dicynodont 2 Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. The size-frequency distri- bution in palaeoecology: effects of taphonomic processes during formation of mollusc death assemblages in Texas bays 3 Dashzeveg, D. See Russell, D. E. and Dashzeveg, D. Dean, W. T. See Zhou Zhiyi and Dean, W. T. Dixon, O. A., Bolton, T. E. and Copper, P. Ellisites, an Upper Ordovician heliolitid coral intermediate between coccoserids and proporids 2 Donovan, S. K. A new genus of inadunate crinoid with unique stem morphology from the Ashgill of Sweden 2 Donovan, S. K., Hollingworth, N. T. J. and Veltramp, C. T. The British Permian crinoid ‘'Cyathocrmites'’ ramosus (Schlotheim) 4 Eastman, J. T. See Grande, L. and Eastman, J. T. Elliott, G. F. Isolated algal cysts in the Palaeocene- Lower Eocene of Kurdistan 4 Evitt, W. R. See Tripp, R. P. and Evitt, W. R. Fenster, E. j. See Hofeman, A. and Fenster, E. J. Fortey, R. A. and Cooper, R. A. A phylogenetic classification of the graptoloids 4 Francis, J. E. Growth rings in Cretaceous and Tertiary wood from Antarctica and their palaeoclimatic implications 4 Fraser, N. C. New Triassic sphenodontids from south-west England and a review of their classification 1 Grande, L. The first articulated freshwater teleost fish from the Cretaceous of North America 2 Grande, L. and Eastman, J. T. A review of Antarctic ichthyofaunas in the light of new fossil discoveries 1 Page 85 187 155 207 583 519 293 243 139 423 313 827 415 495 391 235 809 739 631 665 165 365 113 IV CONTENTS Part Page Hoffman, A. and Fenster, E. J. Randomness and diversification in the Phanerozoic: a simulation 4 655 Hollingworth, N. T. J. See Donovan, S. K., Hollingworth, N. T. J. and Veltkamp, C. T. Holmes, J. and Lopez, J. The disappearing peel technique: an improved method for studying permineralized plant tissues 4 787 Johnson, M. E. See Beadle, S. C. and Johnson, M. E. Kennedy, W. J. The ammonite fauna of the Calcaire a BaaiUtes (Upper Maastrichtian) of the Cotentin Peninsula (Manche, France) 1 25 Lopez, J. See Holmes, J. and Lopez, J. Malinky, j. M., Mapes, R. H. and Broadhead, T. R. New late Palaeozoic Hyolitha (Mol- lusca) from Oklahoma and Texas, and their palaeoenvironmental significance 2 303 Mapes, R. H. See Malinky, J. M., Mapes, R. H. and Broadhead, T. R. McCune, a. R. a revision of Semionotus (Pisces: Semionotidae) from the Triassic and Jurassic of Europe 2 213 Mitchell, C. E. and Carle, K. J. The nematularium of PseudocUmacograptus scharenhergi (Lapworth) and its secretion. 2 373 Ni Yunan. See Cooper, R. A. and Ni Yunan Powell, E. N. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. Ramskold, L. Silurian encrinurid trilobites from Gotland and Dalarna, Sweden 3 527 Reitner, j. See Wood, R. A. and Reitner, J. Runnegar, B. Molecular palaeontology 1 1 Russell, D. E. and Dashzeveg, D. Early Eocene insectivores (Mammalia) from the People’s Republic of Mongolia 2 269 Selden, P. a. a new identity for the Silurian arthropod Necrogarnmarus 3 627 Simms, M. J. Contrasting lifestyles in Lower Jurassic crinoids: a comparison of benthic and pseudopelagic Isocrinida 3 475 S-MiTHSON, T. R. A new anthracosaur amphibian from the Carboniferous of Scotland 3 601 Staff, G. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. Stanton, R. J., Jr. See Cummins, H., Powell, E. N., Stanton, R. J. Jr. and Staff, G. Taylor, P. D. Scanning electron microscopy of uncoated fossils 4 685 Tripp, R. P. and Evitt, W. R. Silicified trilobites of the family Asaphidae from the Middle Ordovician of Virginia 4 705 Veltkamp, C. T. See Donovan, S. K., Hollingworth, N. T. J. and Veltkamp, C. T. Wilson, E. C. The first Tertiary sclerosponge from the Americas 3 575 Wilson, M. A. Coelobites and spatial refuges in a Lower Cretaceous cobble-dwelling hard- ground fauna 4 691 Wood, R. A. and Reitner, J. Poriferan alfinities of Mesozoic stromatoporoids 3 469 Zaborski, P. M. P. Internal mould markings in a Cretaceous ammonite from Nigeria 4 725 Zhang Zhongying. Solar cyclicity in the Precambrian microfossil record 1 101 Zhou Zhiyi and Dean, W. T. Ordovician trilobites from Chedao, Gansu Province, north- west China 4 743 NOTES FOR AUTHORS The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform in style to those already published in this journal, and should be sent to Dr. Dianne Edwards, Department of Plant Science, University College, P.O. Box 78, Cardiff CFl IXL, who will supply detailed instructions for authors on request (these are published in Palaeontology 1985, 28, pp. 793-800). Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology. SPECIAL PAPERS IN PALAEONTOLOGY In addition to publishing Palaeontology the Association also publishes Special Papers in Palaeontology. Members may subscribe to this by writing to the Membership Treasurer: the subscription rale for 1986 is £33 (LI.S. $53) for Institutional Members, and £16-50 (U.S. $27) for Ordinary and Student Members. A single copy of each Special Paper is available to Ordinary and Student Members only, for their personal use, at a discount of 25% below the listed prices. Non-members may obtain copies, but at the listed prices, from Marston Book Services. P.O. Box 87. Oxford 0X4 ILB, England. RECENT PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS Special Papers in Palaeontology Numbers 1-19 are still in print and are available (post free) together with those listed below; 20. (for 1977): Fossil Priapulid Worms, by s. conway morris. 155 pp., 99 te.xt-figs., 30 plates. Price £16 (U.S. $24). 21. (for 1978): Devonian Ammonoids from the Appalachians and their bearing on International Zonation and Correlation, by M. R. HOUSE. 70 pp., 12 text-figs., 10 plates. Price £12 (U.S. $18). 22. (for 1978, published 1979): Curation of Palaeontological Collections. A joint Colloquium of the Palaeontological Association and Geological Curators Group. Edited by m. g. bassett. 279 pp., 53 text-figs. Price £25 (U.S. 138). 23. (for 1979): The Devonian System. A Palaeontological Association International Symposium. Edited by m. r. house, c. t. SCRUTTON and m. g. bassett. 353 pp., 102 text-figs., 1 plate. Price £30 (U.S. $45). 24. (for 1980): Dinoflagellate Cysts and Acritarchs from the Eocene of Southern England, by j. p. bujak, c. downie, g. l. EATON anclG. L. WILLIAMS. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23). 25. (for 1980): Stereom Microstructure of the Echinoid Test, by a. b. smith. 81 pp., 20 text-figs., 23 plates. Price £15 (U.S. $23). 26. (for 1981): The Fine Structure of Graptolite Periderm, by p. r. crowther. 1 19 pp., 37 text-figs., 20 plates. Price £25 (U.S. $38). 27. (for 1981): Late Devonian Acritarchs from the Carnarvon Basin, Western Australia, by g. playford and r. s. dring. 19>pp., \Q text-figs., \9 plates. Price £15 (U.S. $23). 28. (for 1982): The Mammal Fauna of the Early Middle Pleistocene cavern infill site of Westbury-sub-Mendip, Somerset, by M. j. bishop. 108 pp., 47 text-figs., 6 plates. Price £25 (U.S. $38). 29. (for 1982): Fossil Cichlid Fish of Africa, by j. a. h. van couvering. 103 pp., 35 text-figs., 10 plates. Price £30 (U.S. $45). 30. (for 1983): Trilobites and other early Arthropods. Edited by d. e. g. briggs and p. d. lane. 276 pp., 64 text-figs., 38 plates. Price £40 (U.S. $60). 31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, by w. j. KENNEDY. 160 pp., 42 text-figs., 33 plates. Price £25 (U.S. $38). 32. (tor 1984): Autecology of Silurian organisms. Edited by m. g. bassett and j. d. lawson. 295 pp., 75 text-figs., 13 plates. Price £40 (U.S. $60). 33. (for 1985); Evolutionary Case Histories from the Fossil Record. Edited by j. c. w. cope and p. w. skelton. 202 pp., 80 text- figs., 4 plates. Price £30 (U.S. $45). 34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by o. i. Nikiforova, T. L. modzalevskaya and m. g. bassett. 66 pp., 6 text-figs., \6 plates. Price £10 (U.S. $15). Field Guides to Fossils 1. (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12). Other Publications 1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30). 1985. Atlas of invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the Palaeontological Association, xiii-l-241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95. © The Palaeontological Association, 1986 Palaeontology VOLUME 29 • PART 4 CONTENTS A phylogenetic classification of the graptoloids R. A. FOKTEY and K. A. COOPER Randomness and diversification in the Phanerozoic: a simula- tion A. HOFFMAN and E. J. FENSTER Growth rings in Cretaceous and Tertiary wood from Antarc- tica and their palaeoclimatic implications J. E. FRANCIS Scanning electron microscopy of uncoated fossils P. D. TAYLOR Coelobites and spatial refuges in a Lower Cretaceous cobble- dwelling hardground fauna M. A. WILSON Silicified trilobites of the family Asaphidae from the Middle Ordovician of Virginia R. P. TRIPP fl«c/w. R. EVITT Internal mould markings in a Cretaceous ammonite from Nigeria P. M. P. ZABORSKI Isolated algal cysts in the Palaeocene-Lower Eocene of Kurdistan G. F. ELLIOTT Ordovician trilobites from Chedao, Gansu Province, north- west China ZHOU ZHIYI and W. T. DEAN The disappearing peel technique: an improved method for studying permineralized plant tissues J. HOLMES LOPEZ The British Permian crinoid 'Cyathocrinites' ranwsus (Schlot- heim) S. K. DONOVAN, N. T. J. HOLLINGWORTH and C. T. VELTKAMP Evolution of the earliest smooth spire-bearing atrypoids (Brachiopoda: Lissatrypidae, Ordovician-Silurian) P. COPPER Primed in Great. Britain at the University Printing House, Oxford by David Stanford, Printer to the University ISSN 00 631 655 665 685 691 705 725 739 743 787 809 827 0239. MAR 00