5' f '•Ki « » UBurvnrvluo om 1 1 noUI>< IMIN IINoillUMUN I^UI. ^ 5 CO UJ t: uj o “ 2 _ >3iyvaan libraries Smithsonian institution NoiiniiiSNi nvinoshiuns saiavaan lib “ z: f- _ z 3 I — h~ in z NSTITUTION NOliniliSNl NVINOSHimS S3iavaan“LlBRARIES”SMITHS0NIAN~INSTITUTl0N NOI Z W z ^ 2 ... t/J I in O z I . 51 in :/WA^W in -z in z to » z (/) “ z 53iavaan libraries smithsonian institution NoiiniiiSNi nvinoshiii^s saiavaan lie ^ Z ^ V • z to •< .3) o 5 o NSTITUTION NOIiniliSNl“'NVINOSHilWS SBIdVaan LI B RAR I ES^SMITHS0NIAN“'|NSTITUTI0N^N0I Z [I ^ •“ 2 [I 2 to to Vx>«*w - J3 lava an libraries smithsonian iNSTiTUTi0N^N0iiniiiSNi“NviN0SHiiiAis^S3iavaan lib 2 > to 2 > to Z to Z NSTITUTION NOIiniliSNl_NVlNOSHillNS S3iavaan libraries smithsonian INSTITUTION NOI ^ z '\ ^ to ™ to - o • ,.>^ _ xa;A2>^ o “• o -« 2 j 2 j 2 mavaan libraries smithsonian institution NoiiniiiSNj nvinoshiiiais siiavaan lie “ ? •“ Z 2 00 ;:o > m x!ivDC;>' ^ m xj2tot>' ^ m ' S m to — to s (/> \ ? to NSTITUTION NOliniliSNl NVINOSHIIINS S3iaVaaiT LIBRARIES SMITHSONIAN INSTITUTION NOI X s ii) ^ > 2 > '■^siOiiJASX' s > 53 I ava a LIBRARIES smithsonian INSTITUTION NOliniliSNl NVINOSHims‘^S3iavaan^LIB “ to = to := ._ . to 5522^ . .to NViNuaMiiKva sjiavaai i LiDnMr\ic.o omii nauiMini'i n>io 1 1 1 u i iuin imwixi iaixoimi ivviixuonxi»»o . (/) — w = (/7 Z \ ^ UJ ^ UJ 7k uj xSsTtrD^ ,n V UJ % “ /S o o Z ' _1 z SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHilSNS S3IUVHan LIBRARIES SMITHSONIAN II m ><£L£i>^ 71 m m ^ m w) := — (/> X ~ (/) _ w 1 NViNOSHims S3iyvaan libraries Smithsonian institution NoiiniiiSNi nvinoshiisais s Z w Z , , .• W Z .V- W z ^ •^y — K xiasr/ — w' ± "y/iilW — — > 2 Z 2 2 > >‘^SMITHS0NIAN institution NOliniliSNI NVINOSHlI^VS^^SaiBVaan^LIBRARlES SMITHSONIAN_ll w 5 w 5 , 5 z UJ ,n ^ . . y - m 5 5 5 “ V- o l“NVIN0SHimS^S3 I avaa n LI B RAR I ES^SMITHSONIAN^INSTITUTION^NOlinillSNI^NVINOSHlUAIS^S > te£ 3?) 2 MryA.-> (p.^ ? (§t^, ^1S°) > \^Ms. t 73 ■ 73 73 IT| ^ XjVAStj}?^ |T1 ^ rn w ■■ ■ = w — S o> _ w ) SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIIAIS S3iavaail LIBRARIES SMITHSONIAN II z \ w z CO z •■» w 2: < »^ 2 S ^ < S ^ S i i ” I ^ | I I %v I . NVINOSHIIIAIS S3 1 a vy an libraries SMITHSONIAN institution NOlinillSNI NVINOSHimS s =;•■» I S s; ^ ^ < z J z — _J Z _1 3 SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS S3iavaan LIBRARIES SMITHSONIAN I z r- z f~ . z ^ ^ *“ O ,^ti^VAr\ “ j\sSX' . O > 73 ^ z rn ^ m I NVINOSHIIIAIS S3iavaan LIBRARIES^^SMITHSONIAN'lNSTITUTION^NOlinillSNrNVINOSHlIlMS S Z w Z W . CO Z V.. w z w < 2 ,.< xVX... 2 «f . 2 .< _ > 2 ^ > CO Z CO * Z z CO z 3 SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIIAIS S3iavaan LIBRARIES SMITHSONIAN ^ Z - ^ ~ CO LO UJ UJ W .v.c^'N ^ CO ^ ii« A VOLUME 33 • PARTI FEBRUARY 1990 Published by The Palaeontological Association • London Price £30 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to promote research in palaeontology and its allied sciences. COUNCIL 1989-1990 President '. Dr J. D. Hudson. Department of Geology, University of Leicester, Leicester LEI 7RH Vice-Presidents'. Dr M. Romano, Department of Geology, University of Sheffield, Sheffield S3 7HF Dr P. W. Skelton. Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA Treasurer: Dr M. E. Collinson, Department of Biology, King’s College, London W8 7AH Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne NEl 7RU Institutional Membership Treasurer : Dr A. W. Owen, Department of Geology and Applied Geology, University of Glasgow, Glasgow G12 8QQ Secretary : Dr P. Wallace, The Croft Barn, Church Street, East Hendred, Oxon 0X12 SLA Circular Reporter : Dr D. Palmer, Department of Geology, Trinity College, Dublin 2 Marketing Manager : Dr C. R. Hill, Department of Palaeontology, British Museum (Natural History), London SW7 5BD Public Relations Officer : Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 IRJ Editors Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 IRJ Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester, Manchester M13 9PL Dr D. Edwards, Department of Geology, University of Wales College of Cardiff, Cardiff CFl 3YE Dr C. R. C. Paul. Department of Earth Sciences, University of Liverpool. Liverpool L69 3BX Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL Dr P. D. Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD Other Members Dr J. A. Crame, Cambridge Dr G. B. Curry, Glasgow Dr E. A. Jarzembowski, Brighton Dr R. A. Spicer, Oxford Overseas Representatives Argentina : Dr M. O. Mancenido, Division Paleozoologia invertebrados, Facultad de Ciencias Naturales y Museo, Paseo del Bosque, 1900 La Plata. Australia: Dr K. J. McNamara, Western Australian Museum, Francis Street, Perth, Western Australia 6000. Canada: Professor S. H. Williams, Department of Earth Sciences, Memorial University, St John’s, Newfoundland AIB 3X5. China: Dr Chang Mee-mann, Institute of Vertebrate Palaeontology and Paleoanthropology, Academia Sinica, P.O. Box 643, Beijing. Dr Rong Jia-yu, Nanjing Institute of Geology and Palaeontology, Chi-Ming-Ssu, Nanjing. France: Dr J.-L. Henry, Institut de Geologic, Universite de Rennes, Campus de Beaulieu, Avenue du General Leclerc, 35042 Rennes Cedex. Iberia : Prof. F. Alvarez, Departamento de Geologia, Universidad de Oviedo, C/. Jesus Arias de Velasco, s/n. 33005 Oviedo. Spain. Japan : Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Tokyo. New Zealand: Dr R. A. Cooper, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt. Scandinavia: Dr R. Bromley, Fredskovvej 4, 2840 Holte, Denmark. U.S.A. : Prof. A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66044. Prof. N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403. Prof. M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. IE. Germany : Prof. F. T. Fursich, Institut fiir Palaontologie, Universitat, D8700 Wurzburg, Pliecherwall 1 MEMBERSHIP Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1990 are: Institutional membership . . . £60 00 (U.S. $108) Student membership . . . . £1 1-50 (U.S. $20) Ordinary membership . £28-00 (U.S. $50) Retired membership .... £14 00 (U.S. $25) There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen, Department of Geology and Applied Geology, The University, Glasgow G12 8QQ. 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 H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne NEl 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer. All members who join for 1990 will receive Palaeontology, Volume 33, Parts 1^. Enquiries concerning back numbers should be directed to the Marketing Manager. Non-members may subscribe, and also obtain back issues up to 3 years old, at cover price through Basil Blackwell Ltd, Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford 0X2 ODT, England. For issues before Vol. 29 contact the Marketing Manager. Cover: Framboidal pyrite within lumen of a tracheid of the early land plant, Grosslingia breconensis, from the Eower Old Red Sandstone of South Wales, x 700. See P. Kenrick and D. Edwards. Botanical Journal of the Linnean Society, 97, 95-123. PRESERVATION OF SOFT-BODIED AND OTHER ORGANISMS BY BIOIM M U RATION - A REVIEW by P. D. TAYLOR Abstract. Bioimmuration is a poorly-known mode of preservation which results from orgd'nip yvergrowth of sessile organisms. Soft-bodied organisms (and organisms with lightly mineralize'd^Skeletons) can be preserved if overgrown by other organisms possessing mineralized skeletons. These bioimmure'd fossils are visible on the attachment areas of the overgrowing organisms as natural moulds which sometimes become filled by diagenetic mineral growth to give a natural cast. Three types of bioimmuration are distinguished: substratum bioimmuration, epibiont bioimmuration and bioclaustration. Common bioimmuring organisms include oysters and oyster-like cemented bivalves, serpulid worms and cyclostome bryozoans. Among organisms found preserved by bioimmuration are algae, marine angiosperms, hydroids and ctenostome bryozoans. Most documented finds of bioimmured fossils are from the post-Palaeozoic, especially the late Cretaceous, reflecting both the stratigraphical distribution of potential bioimmuring organisms and concentration of research effort. Future finds of bioimmured fossils offer considerable scope for adding to our knowledge of the fossil history and ecological contributions of soft-bodied organisms living on firm or hard substrata. sediments containing fossils of soft-bodied marine organisms (i.e. organisms lacking mineralized skeletons) provide invaluable windows into parts of the biosphere which are normally unavailable to the palaeontologist. Several such deposits, described as fossil Lagerstatten, have been intensively studied, notably the Burgess Shale, Hiinsruck Shale and Solnhofen Limestone (e.g. see papers in Whittington and Conway Morris 1985). Instrumental in the preservation of these soft-bodied biotas were rather special taphonomic factors generally involving rapid burial in fine-grained sediments and inhibition of organic decay. Because these circumstances have occurred rather infrequently during the Phanerozoic, soft-bodied Lagerstatten are comparatively rare in the marine record. In contrast, bioimmuration is a process of organic overgrowth which routinely results in the preservation of sessile soft-bodied organisms in unexceptional palaeoenvironments. Very few studies have been undertaken of bioimmuration, and even the existence of this important mode of preservation is poorly known among palaeontologists. Vialov ( 1961 ) first recognized immuration as a mode of fossilization. He distinguished two types of immuration: lithoimmuration (e.g. envelopment of snakes by calcareous tufa during hibernation) and bioimmuration (e.g. overgrowth of balanid barnacles by oysters). Only Voigt (1956, 1966, 1968a, 6, 1972, 1973a, 6, 1977, 1979, 1980, 1981, 1983; Voigt and Harmelin 1986) has published extensively on bioimmured fossils in a series of mostly German language papers, mainly dealing with bioimmured fossils from the late Cretaceous Chalk-Tuflf of Maastricht. The present paper sets out to review the process of bioimmuration and the variety of soft-bodied organisms which have been found as bioimmured fossils. Specimen repositories', figured specimens are in the collections of the British Museum (Natural History) (abbreviated BMNH) and the Voigt Collection, Universitiit Hamburg (VH). TYPES AND PROCESSES OF BIOIMMURATION To ‘immure’ means to imprison. Sessile organisms are routinely bioimmured when partially or completely overgrown by other organisms. If the overgrowing organism possesses a mineralized skeleton, then the basal surface of this skeleton may carry a preservable replicate in negative relief IPalaeontology, Vol. 33, Part I, 1990, pp. 1-17, 2 pls.| © The Palaeontological Association I PAL 3.1 2 PALAEONTOLOGY, VOLUME 33 of the upper surface of the overgrown organism. Such natural moulds are bioimmurations of soft- bodied organisms. For example, many species of present-day animals and plants live as epiphytes on the fronds or stipes of marine plants. The attached undersides of these epiphytes are often adpressed very closely to their plant substrata, thereby creating a natural mould which is revealed when the epiphyte is detached from its substratum. Similarly, epiphytes sharing the same plant substratum are often found to overgrow one another. This too may result in a natural mould of the overgrown (or bioimmured) epiphyte on the base of the overgrowing (or bioimmuring) epiphyte. It is useful to distinguish three principal types of bioimmuration : (1) substratum hioimmuration in which the organism being overgrown is the substratum for the bioimmuring organism (text-fig. 1a); (2) epibiont bioimmuration in which the overgrowing and overgrown organism share the same substratum (text-fig. 1b); (3) bioclaustration in which the substratum is a living organism and is itself is responsible for overgrowing the bioimmured organism (text-fig. Ic). Substratum bioimmurations are two component systems (substratum organism + bioimmuring organism). For example, an impression of an algal frond on the attachment area of an oyster which lived as an algal epiphyte is a substratum bioimmuration. Epibiont bioimmurations are three component systems (substratum -l- bioimmuring encruster -I- bioimmured encruster). If two bryo- zoans encrust the same bivalve shell and one overgrows the other, then the mould of the overgrown bryozoan on the base of the overgrowing bryozoan is an epibiont bioimmuration (note that the term epibiont is here used in accordance with the recommendation of West 1977 for an organism living on rather than within a substratum; there is no implication regarding the nature of the substratum). Bioclaustrations are a special type of bioimmuration with two components (combined sub- stratum/bioimmuring organism + bioimmured organism). An example of bioclaustration is the embedment of a sponge growing on the surface of a coral by continued growth of the coral. Substratum bioimmurations These are a consequence of fouling of a biotic substratum, i.e. the settlement of a larva directly onto the surface of the substratum. Such substrata include organisms alive at the time of fouling and also dead, especially skeletal, remains. Although fouling of the shell can be advantageous in some bivalves which are less likely to be victims of predators when fouled (e.g. Pitcher and Butler 1987), fouling is very often disadvantageous to living organisms, e.g. the photosynthetic output of seaweeds may drop as a result of fouling (Cancino et al. 1987), and organisms fouling mussels can increase drag and therefore the probability of dislodgement (Witman and Suchanek 1984). Marine organisms frequently possess defences (e.g. antibiotic surface secretions, see Dyrynda 1986) which enable them to resist being fouled. Nevertheless, fouling is of widespread occurrence, especially on the older parts of marine plants, and away from the actively feeding regions of marine animals. Marine plants, colonial coelenterates, crinoids and other organisms with arborescent morphologies may be particularly favoured substrata because they elevate the fouling organism above the sea-bed where mortality caused by sedimentation, predation and competition may be more severe (Jackson 1979). Epibiont bioimmurations These result from lateral overgrowth in which one encruster encroaches a second encruster along the surface of their shared substratum and succeeds in overgrowing the edges of the second encruster. Overgrowth may entail a living organism growing onto the surface of a dead organism, or it may involve two living organisms in competition for substratum space. Spatial competition can be intense among encrusters colonizing firm or hard substrata, and a wide variety of physical and chemical methods are utilized by living encrusters both to assist in overgrowing competitors and to defend against being overgrown by them (see Buss 1986). Despite these, interspecific overgrowths can occur with great frequency. The effect on the overgrown organism of substantial overgrowth is usually death or, if the overgrown organism is a colonial animal, partial mortality, i.e. death of some of the zooids within the colony but not the entire colony (see Jackson and Hughes 1985). Some TAYLOR: BlOl M M U RATION 3 A V A A A A A A AAAAAAAAAAAAAAAA /TA a a vAAAAAAAAAAAAAAAAAAAAAAAAAA vAAAAAAAAAAAAAAAAAAAAAAAAAA vAAAAAAAAAAAAAAAAAAAAAAAAAA lAAAAAAAAAAAAAAAAAAAAAAAAAA vAAAAAAAAAAAAAAAAAAAAAAAAAA vAAAAAAAAAAAAAAAAAAAAAAAAAA iAAAAAAAAAAAAAAAAAAAAAAAAAA B TEXT-FIG. 1. Diagrammatic vertical sections depicting the three types of bioimmuration and their formation. A, substratum bioimmuration; bioimmuration is revealed on the underside of the overgrowing organism (coarse stipple) following loss of the substratum (chevron ornament), b, epibiont bioimmuration; bioimmuration is revealed on underside of overgrowing organism (coarse stipple) following loss of the epibiont (fine stipple) and its substratum (chevron ornament), c, bioclaustration ; bioimmuration becomes visible when the embedding organism (coarse stipple), which formed both the substratum and overgrowing organism, is fractured to reveal the mould of the epibiont (fine stipple) within. I-’ 4 PALAEONTOLOGY, VOLUME 33 bryozoans are now known to survive in a dormant state pending removal of the covering organism (Todd and Turner 1988). Bioclaustrations The term bioclaustration was introduced by Palmer and Wilson (1988) to describe the process of embedment of a soft-bodied infesting organism by the skeletal growth of a host organism. The result of embedment is a pseudoboring, often mistaken for a true boring. Palmer and Wilson (1988, p. 940) regarded bioclaustration as distinct from bioimmuration because bioclaustration is ‘a response to an interaction that is of one partner’s seeking’ whereas bioimmuration ‘demonstrates chance competition for space’. However, such distinction depends upon the inference of biological processes (host selection and competition) which in most cases cannot be made with sufficient confidence, and are not always mutually exclusive. When bioclaustration is viewed in terms of the resulting pattern (text-fig. Ic), its close relationship to other types of bioimmuration becomes clear. Furthermore, to exclude from the definition of bioclaustration the embedment of organisms with hard skeletons seems inappropriate if bioclaustration is to be viewed as a category of bioimmuration; Vialov (1961) included organisms with mineralized skeletons in his original concept of bioimmuration. Rugose or tabulate corals embedded within the coenostea of stromatoporoids with which they intergrew during life (see Kershaw 1987) are, for example, here regarded as bioclaustrations. Opportunities for bioimmuration by fouling, overgrowth and embedment are frequent on hard and firm marine substrata at the present day, and there is no evidence that they have been any less so throughout much of the Phanerozoic. ‘Skeletal overgrowths’ between organisms with mineralized skeletons have often been recorded among ancient hard substratum assemblages (e.g. Taylor 1979, 1984; Liddell and Brett 1982). PRESERVATIONAL STATES Most bioimmured fossils are preserved in negative relief as external moulds (PI. 1, figs. 1, 5-6; PI. 2, figs. 1 and 2, 4 and 5; text-fig. 2). These moulds become visible only after detachment of the bioimmuring organism from its substratum. If the substratum is perishable, detachment can occur before burial or shortly afterwards; if it is aragonitic, detachment often follows diagenetic shell dissolution. Oysters and many other bioimmuring organisms may remain firmly cemented to calcitic EXPLANATION OF PLATE 1 Fig. 1. Arachnidium smitliii (Phillips), BMNH D 57495, a soft-bodied ctenostome bryozoan preserved as an epibiont mould bioimmuration on the attachment area of Gryphaea \ note partial collapse of zooid in centre, Villers-sur-mer, Normandy, Oxfordian, x45. Fig. 2. Andriopora major Larwood, BMNH D 58095, a calcified cribrimorph bryozoan showing 3 pairs of pore chambers (arrowed) which are invisible in conventionally preserved specimens but are seen in this bioimmured zooid on the attachment area of Pycnodonte vesiculare, Weybourne, Norfolk, Weybourne Chalk, Campanian, x 86. Figs. 3 and 4. Hippothoa flagellum (Manzoni), BMNH 1988.12.1.1, a calcified ascophoran bryozoan preserved on the attachment area of an overgrowing oyster which encrusted a plastic pipe, Piran, Adriatic Sea, Recent. 3, zooid (growth direction top left to bottom right) with a distal and two lateral buds, x 90. 4, detail of bud origins showing oyster shell seemingly filling the narrow gap between the slightly raised points of origin of the buds and the substratum, x 230. Figs. 5 and 6. Bioimmured soft-bodied organisms on the attachment area of the bivalve Pycnodonte vesiculare, BMNH H 5501, Thanet, Kent, Santonian, 5, distorted bioimmuration of organism with cuspate margins which has collapsed in the direction of overgrowth (top to bottom), x 15. 6, Eisenackiella thanetensis Taylor, the erect stem of a probable hydroid pushed flat against the substratum during overgrowth, x 25. All illustrations are back-scattered electron micrographs of uncoated specimens. PLATE I TAYLOR, bioimmuration 6 PALAEONTOLOGY, VOLUME 33 substrata (e.g. other oysters, sedimentary hardgrounds) during fossilization. Therefore, the organisms they overgrew are seldom revealed but may on occasions be visible as hummocks and irregularities on the inner side of the attached valve. Sometimes the mould is filled during diagenesis by calcite or pyrite giving an external cast of the bioimmured organism (see Taylor 1990, text-fig. 1). Cast bioimmurations may become visible through fracturing or exfoliation of the encrusting organism from its substratum. The fidelity of casting can be very high, especially when the casting material is pyrite (Taylor 1990, pi. 2, figs. 1 and 2). Whereas epibiont bioimmurations are found as both moulds and casts, substratum bioimmurations are preserved only as moulds. Each bioimmuration is formed gradually during the period of progressive overgrowth by the bioimmuring organism. The shorter this period of time, the more likely it is that the process will be completed before significant deterioration of the overgrown organism. Therefore, bioimmurations of the highest fidelity should occur when the bioimmuring organism grows rapidly and/or when the bioimmured organism is small and is consequently overgrown quickly. Organisms which retain their shape during overgrowth should provide the most easily distinguishable bioimmurations. Conversely, flaccid organisms may become flattened and distorted during overgrowth, and can be difficult to identify when bioimmured. The ability of colonial animals such as hydrozoans and bryozoans to sustain partial mortality favours their preservation by bioimmuration because while certain zooids are being overgrowth, other zooids remain alive and continue to sustain the colony. Large organisms fouled by small epibionts may be affected insignificantly by overgrowth, suffer little deterioration and therefore yield good quality substratum bioimmurations. The time required for the formation of a bioimmuration obviously depends on the growth rate of the bioimmuring organism. Rather than being instantaneous ‘snap-shots’, bioimmurations are summations over time of the morphology of the substratum and its epibionts as they were progressively covered by the advancing growing edge of the bioimmuring organism. This has two implications. First, the morphology of a bioimmured organism need not necessarily correspond to its appearance at any one time during its life. This may be especially true for large and/or colonial organisms which do not suffer mortality immediately on commencement of overgrowth. For example, the single specimen of bioimmured hydroid Eisenackiella thanetensis described by Taylor (1988) may have been a large colony or a narrow strip-like colony which inhabited the substratum just in front of the bioimmuring bivalve Pycnodonte vesiculare and advanced with growth of the bivalve. Secondly, bioimmurations can document temporal successional changes in the organisms living on firm or hard substrata; the early growth stages of the bioimmuring organism overgrow epibionts recruited during early stages of ecological succession, the later formed parts overgrow epibionts recruited during later successional stages. This offers a potentially useful way of studying short-term ecological succession in fossil material. Erect components of soft-bodied organisms are generally pushed over during overgrowth and are flattened against the substratum, resting in an orientation parallel to the growth direction of the bioimmuring organism. A good example of this is found in the weakly-calcified cheilostome bryozoan Aetea which has been described as a bioimmuration from the Pliocene and Recent by Voigt (1983). The zooids of Aetea each have an adnate proximal part and a tall erect distal ‘peristome’. Adnate parts of the zooid remain in position during bioimmuration, but erect peristomes are pushed over so that they lie flat and parallel to the growth direction of the bioimmuring oysters and to one another. Similarly (PI. 1, fig. 6), erect stems of the late Cretaceous hydroid Eisenackiella thanetensis were flattened against the substratum by the overgrowing bivalve Pycnodonte vesiculare. Interpretation of bioimmured fossils must take into account such distortions. Structures formed by pushing over can be distinguished by their orientation which is parallel to the local growth direction of the bioimmuring organism. BIOIMMURING ORGANISMS Potential bioimmuring organisms comprise a taxonomically diverse variety of encrusting animals with mineralized skeletons. These include attached foraminifers, sponges, corals, serpulid TAYLOR: BIOI M M U RATION 7 polychaetes, cemented brachiopods, bryozoans, oysters and oyster-like bivalves (see Nicol 1978). However, only serpulids, cyclostome bryozoans and oysters have been described frequently as bioimmuring organisms, although bioimmuration has also been recorded involving foraminifers, sponges and trepostome bryozoans. An obvious property required of a bioimmuring organism is that it should be capable of overgrowing fellow epibionts or of fouling organic substrata. Barnacles seem rarely to yield epibiont bioimmurations, possibly because they tend to prise competitors off the substratum rather than overgrowing them. Encrusters with sheet-like morphologies (as opposed to runner-like, ramifying organisms) are generally adept at overgrowing competitors for substratum space. These are highly likely to result in bioimmuration. The probability of an encruster encountering an epibiont, and the likelihood of successful overgrowth occurring should each increase with the size of the overgrowing organism. Furthermore, large organisms provide larger sampling areas. Therefore, large encrusters with large attachment areas are more likely to be found with bioimmurations of epibionts than are smaller encrusters. As noted on page 6, rapid overgrowth should produce bioimmurations of the highest quality. The fastest shell growth rate known in a present day oyster occurs in Crassostrea cuttackensis (Smith and Newton) living in Madras Harbour. Shells may grow 0-27-0-62 mm per day (see Stenzel 1971, p. N1014). An oyster shell growing at this rate could, for example, completely overgrow an average bryozoan zooid in a matter of one or a few days, presumably before any marked deterioration in the condition of the zooid. No data appear to be available on growth rates in cyclostome bryozoans, but some Recent sheet-like cheilostomes with a similar colony morphology grow at rates of 30- 1 10 mm per year (Jackson and Coates 1986, p. 9), i.e. 0-08-0.30 mm per day. Although these rates would not be expected to yield such good quality bioimmurations as those of the fastest growing oysters, small organisms might still be overgrown in a matter of days. The basal calcified skeletons of encrusting animals are not usually in direct contact with the substratum but are separated from it by an organic layer which contains the cement causing adhesion of the encruster, and onto which the calcified layers are seeded. However, because these organic layers tend to be exceedingly thin, their presence seems to have little effect on the fidelity of the bioimmuration impressed on the calcified skeleton. For example, the periostacum of oysters, secreted by glands in folds of the mantle lobes, is described as ‘very thin’ by Stenzel (1971, p. N977; see also Carricker, Palmer and Prezant 1980), while the initial thickness of the cuticle in some cheilostome bryozoans is only a few microns (see Ryland 1976, p. 295). Because the organic basal layer in some encrusters is very thin, the calcified layer is able to enter and mould extremely confined recesses on the substratum and any overgrown epibionts. A Recent specimen from the Adriatic Sea provides a good illustration (PI. 1, figs. 3 and 4). Here an oyster, once attached to a plastic pipe, has bioimmured a runner-like colony of the cheilostome bryozoan Hippothoa flagellum (Manzoni). In H. flagellum, new zooids originate as buds from the pore windows of parent zooids. The pore windows are apparently situated a little above substratum level, giving a very low ‘arch’ before the bud regains the substratum during distal growth. The fact that this arch has apparently been filled by calcite of the oyster shell in the illustrated specimen (leaving parent and daughter zooids apparently separated) demonstrates that extremely small-scale aspects of morphology are capable of being moulded by oysters. BIOIMMURED SOFT-BODIED ORGANISMS The overwhelming majority of bioimmured soft-bodied organisms have been described from the Jurassic and Cretaceous. This reflects both the focus of Professor E. Voigt’s research, especially on the type Maastrichtian, as well as the undoubted abundance of good bioimmuring organisms, such as cemented bivalves, serpulids and cyclostome bryozoans, in the Jurassic and Cretaceous of north- west Europe. Furthermore, the abundance of aragonitic substrata, subsequently dissolved to reveal the attachment surfaces of their encrusters, also increases in the post-Palaeozoic. Very few bioimmurations of soft-bodied organisms have been described from the Palaeozoic and there is clearly much potential for further discoveries. 8 PALAEONTOLOGY, VOLUME 33 Soft-bodied or poorly-mineralized organisms belonging to the following taxonomic groups have been found as bioimmurations : 1. Algae. Voigt (1956, 1966, 1973a) has described many examples of bryozoans from the Maastrichtian Chalk-Tutf of Maastricht which were apparently attached to the stems and leaves of macroalgae. In a sample of 1800 bryozoans with intact colony bases, 79% of colonies were inferred to have been algal epiphytes (Voigt 1973a). The substratum bioimmurations of these algae often have smooth and glossy surfaces, and in some cases it seems possible that the epiphytic bryozoan did not make close contact with its algal substratum (Voigt 1973a described some Recent epiphytic bryozoans having strut-like outgrowths at the bases of their colonies). Putative algal stem bioimmurations are recognizable as cylindrical hollows around which the bryozoan colony was wrapped (Voigt 1956, pi. 1, figs. 1-4, text-figs. 2 and 3). Some of these colonies have tubular ereet branches, a colony growth-form described as cavariiform (however, not all cavariiform bryozoans were algal epiphytes: examples in which the hollow is partitioned by skeletal walls cannot have grown around algae). Specimens of the worm Spirorbis with concave basal parts (Voigt 1956, text- fig. 4) were also apparently epiphytes of algal stems. Bryozoans such as the cyclostome Actinopora disticha (v. Hagenow) sometimes bioimmured algal fronds which are preserved as narrow tubular voids around which the colony is wrapped (Voigt 1956, pi. 2, figs. 9 and 10). The Recent chlorophytacean Codium bursa (L.) from the Mediterranean is a cushion-shaped alga which often supports a fauna of epiphytic bryozoans (33 species) growing on cryptic surfaces beneath the overhanging edge of the plant (Voigt and Harmelin 1986). The mammilate surface of the alga, consisting of the ends of the utriculi, is replicated on the undersides of epiphytes such as the cyclostome Tubulipora plumosa Harmer (text-fig. 2d). Very similar patterns have been found by Voigt and Harmelin on the encrusting bases of the fossil cyclostomes Osculipora tetragona (Michelin) from the Middle Cenomanian of Le Mans (France), and O. houzeaui Pergens, O. tnmcata (Goldfuss) and Reteporidea lichenoides (Goldfuss) from the Chalk-Tuff of the Maastricht region (Voigt and Harmelin 1986, pi. 2, figs. 4 and 5, pi. 3, figs. 9, 12 and 13). These are interpreted as bioimmurations of a Codium-Mke alga. An Upper Jurassic dasycladacean alga, Goniolina geometrica (Roemer), from West Germany, is preserved on the cementation areas of small oysters as bioimmured impressions of the regular hexagonal surface pattern (Voigt and Harmelin 1986, pi. 4, figs. 15 and 16). A rare example of bioimmuration in the Palaeozoic is provided by a specimen from an Ordovician erratic boulder from Gotland. Hillmer and Schallreuter (1987, fig. 3i-j) figure the bioimmuration of a putative alga with a Goniolina-Mkt surface overgrown by a cryptostome bryozoan. 2. Marine angiosperms. The Chalk-Tuff of Maastricht contains bioimmured seagrass leaves which were described by Voigt (1956, 1966). Bioimmuring organisms include the oyster Exogyra, the cyclostomes Actinopora disticha and Lichenopora sp., and the foraminifer Planorbidinella cretae (Marsson). On their basal surfaces are moulded the patterns of epidermal cells and veins of seagrass leaves (see Voigt 1956, pi. 4, figs. 1 and 2). It seems possible that some of these leaves are from the seagrass Thalassocharis bosqueti (Debey ex Miquel) which occurs commonly as silicified axes and EXPLANATION OF PLATE 2 Figs. 1 and 2. Ventriculitid sponge preserved as a substratum mould bioimmuration on the underside of the cemented bivalve Pycnodonte vesiculare, BMNH S 10250, England, Chalk. 1, general view, x 14. 2, detail of sponge surface, x 23. Fig. 3. Xenomorph of trigoniid bivalve on the unattached, right valve of a Gryphaea, BMNH 24065, Weymouth, Jurassic, Oxford Clay, xO-8. Figs. 4 and 5. Substratum bioimmurations of unknown identity visible on the attachment areas of cemented bivalves. 4, BMNH Z 1062, Le Mans, Sarthe, Cenomanian, Sables du Perche, x 8. 5, BMNH D 32168, Bognor, Sussex, Eocene, London Clay, x 28. Figs. 2, 4 and 5 are back-scattered electron micrographs of uncoated specimens. PLATE 2 TAYLOR, bioimmuration 10 PALAEONTOLOGY, VOLUME 33 rootlets, sometimes bryozoan encrusted, at Kunrade 30 km ESE of Maastricht (Voigt and Domke 1955; Voigt 1973o). 3. Protists. Ernst (1985) reported the occurrence of cylindrical tubes, up to 1-5 mm long and 01 0-0- 15 mm in diameter, within zooecia of the cheilostome bryozoans Onychocella piriformis and 0. cyclostoma from the Maastricht Chalk-Tuff. The tubes were apparently secreted by the bryozoans in response to the presence of an infesting organism and are thus bioimmurations of the bioclaustration type. Ernst regarded the infesting organism as probably a folliculinid ciliate. 4. Poriferans. The spiculate surface of a putative monactinellid sponge bioimmured by a cyclostome is described by Voigt (1966, pi. 34, figs. 1 and 2). 5. Hydrozoans. Although very abundant members of present-day hard and firm substratum communities, hydrozoan cnidarians (excepting a few well-calcified groups such as milleporids and stylasterines) have a meagre fossil record, and many putative body fossils of hydrozoans from the Palaeozoic require confirmation (Hill and Wells 1956). Scrutton (1975) described the hydroid Protidophila gestroi Rovereto preserved by bioclaustration. This Middle Jurassic to Pliocene species occurs in association with serpulid worms, having become embedded in the tubes as they grew. Epibiont bioimmurations of thecate hydroids have been described by Voigt (19736) and Taylor (1988). Hydrallmania graptolithiformis Voigt, 19736, preserved as a mould on the underside of the cyclostome Actinopora disticha from the Maastricht Chalk-Tuff, is the only known fossil of the Sertulariidae, a family of hydroids which is common at the present day. The Maastrichtian species closely resembles some Recent species of Hydrallmania in having an imbricate arrangement of hydrothecae along one side of the branch. Of more problematical affinity is Eisenackiella thanetensis Taylor, 1988, based on a single specimen bioimmured by the bivalve Pycnodonte vesiculare (Lamarck) from the Santonian of Kent. The colony appears to have an adnate system of hydrothecae-bearing stolons which gives rise to a series of erect stems also bearing hydrothecae (PI. 1, fig. 6). Both the upright distal parts of the stolonal hydrothecae, and the erect stems have been flattened in the direction of growth of the bioimmuring bivalve. Stolonal morphology recalls that of certain Recent Lafoeidae, whereas stem morphology is reminiscent of the Family Sertulariidae. E. thanetensis is possibly a compound organism resulting from chance juxtaposition of stolon forming and an erect hydroid species. Voigt (1966, fig. 1) illustrates an unidentified bioimmured Maastrichtian hydroid which he compares with the living species Syncoryne sarsi Loven. 6. Octocoral cmthozoans. One of the few bioimmurations recognized from the Palaeozoic is of an inferred gorgonian octocoral. Plumalina conservata was described by Glinski (1956) from the Middle Devonian of the Eifel. It is represented by a single pinnate specimen bioimmured by a trepostome bryozoan determined as Heterotrypa sp. 7. Ctenostome bryozoans. The Ctenostomata are an exclusively soft-bodied, primitive, para- phyletic order of bryozoans. A minority of species bore into calcareous substrata and are found as trace fossils, and encrusting species preserved by bioimmuration are not uncommon in the Mesozoic (see Taylor 1990). Although ctenostime classification is somewhat contentious, it has been traditional to distinguish two groups of ctenostomes: the Stolonifera in which the autozooids are linked by a stolonal system comprising kenozooids, and the Carnosa in which stolons are wanting. Voigt (1966, 1972, 1979) has described bioimmured stoloniferan species assigned to one extinct and two extant genera. Stolonicella Voigt, 1966 is a probable ctenostome known only as bioimmurations. The colony consists of a stolonal system bearing erect autozooids at intervals (text- fig. 2a and B), and resembles the living ctenostome Avenella fusca Dalyell. Zooids often possess a fine transverse ornament. Three species of Stolonicella occur in the Chalk-Tuff of Maastricht (S. schindewolfi Voigt, 1966; S. filosa Voigt, 1966; S. hillmeri Voigt, 1979) bioimmured by bryozoans and oysters, and one in the Turonian Greensand of Miilheim-Broich, West Germany {S. westfalica Voigt, 1966) bioimmured by Ostrea (Lopha) semiplana Sowerby and small Exogyra. Voigt ( 1979, pi. 3, figs. 1 and 2) illustrates an unusual colony of S. filosa which apparently encircled a colony of the cyclostome Stellocavea franccpiana d'Orbigny before being overgrown by the cyclostome. The extant ctenostome genus Amathia is characterized by erect stolons which bear biserial clusters of autozooids (see Chimonides 1987). A. immurata Voigt, 1972 from the Maastricht Chalk-Tuff is TAYLOR: BIOIMMURATION 1 1 TEXT-FIG. 2. Bioimmured organisms preserved as moulds on the undersides of overgrowing organisms, a and B. VH 8611, the probable ctenostome bryozoan Stolonicella sp., Blom Quarry, near Maastricht. Upper Maastrichtian. A, zooids arising from a stolon, x 12. b, detail of zooids and stolon x 33. c, ribbon-like thallus of the alga Fosliella inexspectata Voigt, VH 9494, Blom Quarry, Maastricht, Upper Maastrichtian, x 60. d, VH 10522, bioimmuration of the chlorophytacean alga Codiwn bursa (L.) on the underside of the cyclostome bryozoan Tuhulipora phimosa Harmer, Mediterranean, Recent, x 20. Electron micrographs kindly provided by Professor E. Voigt. the only known fossil species, owing its preservation to bioimmuration by the cyclostome Idmidronea macilenta (v. Hagenow). Colonies of the Recent A. cornuta Lamouroux bioimmured by the foraminifer Acervuliiia adhaerem (Schultze) were shown by Voigt (1972) to have an appearance very like the Maastrichtian fossil. Another extent stoloniferan genus. Buskin, is represented by two fossil species preserved as bioimmurations (Voigt 1979): B. inexpectata Voigt from the Upper Maastrichtian of Curfs preserved on the base of the cyclostome IDitaxia with which it shared an algal substratum, and B. hac/iii \o\gl from the Pliocene of Puget sur FArgens (France) bioimmured by a serpulid tube. Several Jurassic and Cretaceous bioimmured carnosan ctenostome species have been described, all referred to the extant family Arachnidiidae (see Taylor 1990). Their encrusting colonies are composed of uniserial chains of zooids which ramify across the substratum (PI. 1, fig. 1 ). The zooids have a pyriform outline shape and sometimes possess a long and narrow proximal portion (cauda). Arachnidium hramiesi Voigt, 1968r/ is founded on a single specimen from the Lower Cretaceous (Barremian) of Hoheneggelsen (West Germany). The colony is preserved as a cast bioimmuration which was revealed when the bioimmuring serpulid, Proliserpula (Proliserpu/a) bucculenta Regenhardt, was stripped away from the substratum, a guard of the belemnite Oxyteuthis 12 PALAEONTOLOGY, VOLUME 33 hrunsvicensis v. Stromb. A similar species, Arachnidium jurassicum Voigt, but with smaller zooids, was subsequently described by Voigt (1977) from the Middle Jurassic (Aalenian) of Goslar (West Germany). As in zl. hrandesi, the ctenostome is preserved as a cast bioimmuration on the guard of a belemnite (Megateuthis). In this case, however, the bioimmuring organism is an oyster. Taylor (1978) recognized that the type specimens of two nineteenth-century species of Jurassic bryozoans previously regarded as belonging to the cyclostome genus Stomatopora Bronn were in foct cast bioimmurations of Arachnidium. The first of these had been described by Philips (1829) as Cellaria smithii and originated from the Middle Jurassic (‘Cornbrash’; probably Callovian) of Scarborough (Yorkshire). The holotype of Arachnidium smithii (Phillips) is a cast bioimmuration attached to the bivalve Cardium citrinoideum Phillips, and was possibly bioimmured by an oyster though little remains of the bioimmuring organism. The second species, Stomatopora phillipsii Vine, 1892, placed in synonymy with A. smithii, is represented by a cast bioimmuration, overgrown by an oyster, attached to the brachiopod Ohovothyris from the Middle Jurassic (Bathonian, Cornbrash) of Thrapston (Northamptonshire). Three new species of arachnidiids from the Jurassic are described by Taylor (1990) who emphasizes the relative abundance of these bioimmured ctenostomes in late Jurassic deposits. Finally, Voigt (1980) described Arachnidium longicauda from the Chalk-Tuff of Maastricht. This species is represented by a mould bioimmuration of a colony originally attached to an algal leaf and overgrown by the cyclostome TruncatuUpora. Many of the zooids have extremely long caudae which serve to distinguish A. longicauda from previously described species of Arachnidium. 8. Incertae sedis. Very many bioimmurations are taxonomically indeterminate. This may be a consequence of distortion of the bioimmured organism during overgrowth, insufficient mor- phological characters for identihcation, lack of knowledge of the appearance when bioimmured of comparative living organisms, or a combination of these factors. Two particular morphotypes of incertae sedis occur commonly among epibiont bioimmurations : threads and mounds. Bioimmurations of narrow thread-like structures are often encountered on the attachment areas of Mesozoic to Recent oysters and bryozoans. The threads may ramify and cover substantial areas of substratum. Several different groups of organisms are potentially responsible, including stoloniferan ctenostome bryozoans (zooids of Recent stoloniferans are sometimes deciduous, their loss leaving an undiagnostic stolonal system), hydroids, filamentous fungi and algae. Mound-shaped bioimmurations are commonly distorted as a result of overgrowth of a semiflaccid organism. Only the basal outline shape of the organism may be preserved as a potential distinguishing character. For example, an Upper Cretaceous mound bioimmuration (PI. 1, fig. 5) occurring with the holotype specimen of Eisenackiella thanetensis, was evidently a soft-bodied organism with a scalloped edge which was pushed over and severely distorted during overgrowth by Pycnodonte vesiculare. This and similar mound bioimmurations may possibly be zoanthid or actiniid cnidarians, ascidiaceans or sponges. Two substratum bioimmurations of unknown identity are shown in PI. 2, figs. 4 and 5. Both occur on the attachment areas of cemented bivalves. The Cretaceous example (PI. 2, fig. 4) overgrew a substratum of matted fibres, and the Eocene example (PI. 2, fig. 5) a substratum with a pattern of conjugate ridges. Rohr and Boucot (1989) have recently described a substratum bioimmuration preserved by individuals of the oyster Lopha ramicola Beurlen from the Upper Cretaceous of Brazil. These oysters bioimmured stem-like structures, about 1 cm in diameter, which were covered with closely- spaced circular nodules arranged in a spiral pattern. The bioimmuration is replicated in positive relief on the unattached valves by xenomorphism. Rohr and Boucot regard the bioimmured organism as of unknown affinity, but make comparisons with gorgonacean octocorals and axes of plants (e.g. gymnosperm stems bearing the traces of leaf scars). The bioclaustrated ichnotaxon Catellocaula vallata Palmer and Wilson, 1988, embedded in trepostome bryozoan colonies from the Upper Ordovician of the Cincinnati area of the USA, is another organism of unknown affinity. The fossil consists of a series of 2 mm wide radiating tunnels connecting pits in the surface of the host bryozoan colony. Palmer and Wilson interpret the TAYLOR; BIOIMMURATION 13 organism as a stoloniferous colony, possibly a hydroid but more probably a colonial ascidiacian tunicate. Circular-parabolic pits are commonly found in fossil echinoderms, particularly Palaeozoic crinoids (see Brett 1985 and references therein). They were apparently produced by a combination of boring and embedment (i.e. bioclaustration). Brett introduced the ichnogenus Tremichmts for such structures and regarded them as the work of a sessile, host-selective, probably filter-feeding epibiont. BIOIMMURED SKELETAL ORGANISMS Although the most interesting bioimmured fossils are undoubtedly those of soft-bodied organisms, bioimmurations of organisms with mineralized skeletons may also be valuable in certain circumstances. This is true if the skeletons are normally disarticulated, suffer from diagenetic dissolution, or cover only part of the external surface of the organism. Aragonitic shells Many fossil assemblages lack molluscs with diagenetically unstable shells of aragonite. For example, aragonitic molluscs are generally absent from the Aptian Faringdon Sponge Gravel of Oxfordshire. However, some aragonitic gastropods at Faringdon are preserved as substratum bioimmurations formed on the undersurfaces of cyclostome bryozoans and neuroporid sponges which fouled the gastropod shells. These natural moulds accurately replicate details of shell ornamentation and permit taxonomic identification of the gastropods (R. J. Cleevely, in prep.). Similar moulds of mollusc shells occur in the Bathonian of Normandy (T. J. Palmer, pers. comm. 1988). Celleporid bryozoans from the Neogene occasionally bioclaustrate small solitary corals (Pouyet 1978). Whereas the bryozoans are calcitic, the corals are aragonitic and their skeletons tend to be lost during diagenesis. However, their past presence can be indicated by horn-shaped cavities remaining in the surface of the host bryozoan colony (e.g. m material from the Pliocene Coralline Crag of Suffolk). Numerous examples have been described of oysters bioimmuring lost substrata such as ammonites and other aragonitic molluscs (see Stenzel 1971 and references therein). Most accounts focus on the positive relief replica of the aragonitic mollusc carried by the free valve of the oyster (PI. 2, fig. 3) as a result of the two valves maintaining a constant separation during growth across the mollusc shell, a process termed xenomorphism by Stenzel. However, it is the cemented valve which plays the primary role in the preservation of these substratum bioimmurations. To use a photographic analogy, the cemented valve captures the image as a negative, while the free valve makes a positive print from the negative. Xenomorphic impressions on the free right valves of oysters are rarely as sharp as bioimmurations on the attached left valves. Therefore, they are of less value in identifying the overgrown organism. Exposed soft tissues Not all organisms with mineralized skeletons have their entire external surface covered by hard material. Sponges have a spicular skeleton enveloped during life by soft parts. Hexactinellid sponges of the Family Ventriculitidae (see Reid 1962) are sometimes bioimmured by cemented bivalves (notably Pycnodonte vesiculurc) in the late Cretaceous Chalk of England. These bioimmurations reveal the original surface morphology of the sponge including the ostia (PI. 2, figs. 1 and 2). As early as 1847, Toulmin Smith (p. 89) observed bioimmurations made by oysters which had grown on the surface of ventriculitids. He used their structure to testify to ‘the firmness of the texture of the body and to its noncontractility, as well as to its durability’. This enabled him to assert that structures visible in conventionally preserved ventriculitids were not artefacts resulting from post- mortem distortion. Frontal membrane morphology in bioimmured cheilostomes has been described by Voigt (1968, 14 PALAEONTOLOGY, VOLUME 33 \919a) and Voigt and Ernst (1985). In Taeiiioporina aracimoidea (Goldfuss), the cuticle is ornamented by numerous small projections and pores (Voigt 1968/?, pi. 4, figs. 3 and 4). The Maastrichtian onychocellid cheilostome Nudonychocella nuda Voigt and Ernst, 1985, has greatly reduced cryptocystal frontal wall calcification in post-ancestrular zooids giving it a resemblance to a membraniporimorph. A serpulid bioimmured colony (Voigt and Ernst 1985, pi. 2, fig. 5) confirms the onychocellid affinities of the species, and shows clearly the position of operculum and orifice. Lightly mineralized skeletons The corallinacean alga Fosliella Howe has weakly calcified thalli which are unknown as body fossils. However, Voigt (1981) has described a new species of this genus from the Maastrichtian of Maastricht and Kunrade which is preserved only by bioimmuration. F. inexspectata Voigt has narrow, ribbon-like thalli with files of cells arranged in transverse bands (text-fig. 2c). Cover cells, germination discs and conceptacles are preserved. The alga was an epiphyte of macroalgae and was overgrown by the cyclostome Tnmcatulipora. Previous notions of Fosliella being a primitive member of the Corallinaceae are substantiated by this early occurrence of the genus. Although all cheilostome bryozoans have mineralized skeletons, mineralization can be very slight and certain living groups are unknown as body fossils. Hence bioimmuration provides a valuable opportunity for fossilization of lightly calcified cheilostomes. A new cheilostome from the Maastrict Chalk-Tuff was described by Voigt (1966) as Taeniocellaria setifera. The delicate erect colony, with long setose vibraculae, was flattened during overgrowth by an individual of Exogyra and is preserved as a mould bioimmuration on the attachment area of the oyster along with Stolonicella schindewolfi, and small cyclostomes, bivalves and Vermetus. Such excellent preservation of this fragile colony suggested to Voigt (1979a) that overgrowth occurred rapidly, possibly during the life-time of the colony. The affinities of T. setifera within the Cheilostomata are obscure, but the orifice with sinus (see reconstruction in Voigt 1966, fig. 3) suggests that the species is an ascophoran. Laterotecatia pseudamathia Voigt, 1979a is a hippothoid ascophoran from the Maastrichtian which is known only from bioimmurations. Zooids of L. pseudamathia, named because of its resemblance to bioimmured Amathia immurata, are arranged in characteristic transverse rows. Colonies were algal epiphytes bioimmured by organisms such as the cyclostome Stellocavea francquana. Despite frequent citation as a fossil, the extant cheilostome genus Aetea Lamouroux had no certain fossil record until Voigt (1983) described bioimmured specimens from the Pliocene. This weakly calcified anascan possesses zooids with an adnate proximal part from which there arises an erect tubular part containing the frontal membrane and orifice. Fossil examples (identified as Aetea sp., A. tnmcata (Landsborough) and A. trimcata pygmaea Hincks) from Crete and southern France are preserved as a result of bioimmuration by oysters. Their appearance is very similar to that of Recent Mediterranean specimens of Aetea bioimmured by oysters, the foraminifer Miniacina miniacea Pallas, and the cheilostome Watersipora cucculata Busk. Chitinous exoskeletons of crustaceans commonly disintegrate before burial and fail to fossilize. A bioimmured example of the lobster Liniiparus preserved has been described by Bishop (1981) from the late Cretaceous Ripley Formation of Mississippi. The carapace of Linuparus was overgrown by the oyster Exogyra costata Say probably after the death of the lobster because living lobsters groom themselves to remove epizoans. Using growth lines on the oyster shell. Bishop estimated that complete overgrowth of the carapace took almost a year, and therefore that the carapace remained intact for at least this length of time. This period far exceeds the four weeks quoted by Schafer (1972) for decapod cuticle to lose its strength in the North Sea at the present day. Exposure of hidden undersides Bioimmuration in conjunction with substratum loss, especially dissolution of aragonitic substrata, can also be useful in revealing the undersides of encrusters with mineralized skeletons which are normally juxtaposed with the substratum and hidden from view. For example, adnate cheilostome TAYLOR: BIOIMMURATION 15 bryozoans are usually very firmly cemented to their substrata when found as fossils, and the colony underside is not accessible for study. However, the undersides of colonies overgrown by bivalves and other bioimmuring organisms may become visible after detachment from their substratum. These include species with colonies which are too fragile to survive intact unless held together by the bioimmuring organism. Bioimmurations of zooids of Andriopora major reveal the existence of pore chambers (PI. 1, fig. 2), not recorded from conventionally preserved material (Larwood 1962). DISCUSSION Bioimmuration is a preservational process which can preserve soft-bodied organisms, as well as lightly mineralized or unmineralized components of organisms possessing hard skeletons. Clearly, therefore, it offers considerable potential for adding to our knowledge of biotas of the past. This potential has been little explored. In particular, very few bioimmured fossils have been reported from the Palaeozoic. Potential bioimmuring organisms in the Palaeozoic include cemented articulate and inarticulate brachiopods, sheet-like trepostome and cystoporate bryozoans, cornulitids, corals and stromatoporoids. Future research directions should include: 1. A concerted and systematic search for bioimmurations. Many examples of bioimmured fossils doubtless remain unrecognized in existing collections. Oysters with targe attachment areas are a particularly fruitful source of bioimmurations in the Mesozoic and Cenozoic. The development of techniques to separate oysters and other encrusters from their substrata, thereby exposing bioimmured organisms, would surely increase the probability of making such finds. 2. Study of the details of overgrowth processes in present-day hard substratum communities, and the taphonomy of the organisms being overgrown. Nothing is currently known regarding the relative preservation potentials during bioimmuration of different organisms. Fossil finds suggest that runner-like encrusters provide the most easily preservable (and readily recognizable) subjects for bioimmuration, but this supposition requires testing by reference to modern bioimmurations. Organisms preserved by bioimmuration are usually sessile inhabitants of firm or hard substrata. Fossil assemblages of firm or hard substrata are especially good subjects for palaeoecological studies because their constituent fossils are demonstrably in situ, thereby retaining their original spatial relationships to one another and to the substratum. Fouling and overgrowth interactions between organisms, and interactions between organisms and their substratum (e.g. patterns of spatial recruitment, see Bishop 1988) can be recorded with minimal interpretive assumptions. This additional information on biotic interactions, taken in conjunction with evidence of soft-bodied organisms preserved by bioimmuration, should permit inferences to be made for hard substrate which are beyond those normally possible in palaeoecological studies. Acknowledgements. Professors E. Voigt (Hamburg) and A. Boucot (Corvallis), and Dr T. J. Palmer (Aberystwyth) kindly commented on a draft manuscript. I am especially grateful to Professor Voigt for encouraging my interest in bioimmuration, for supplying the illustrations used in text-fig. 2, and the specimen depicted in PI. 1, figs. 3 and 4. REFERENCES BISHOP, G. A. 1981 . The lobster Linuparus preserved as an attachment scar on the oyster E.xogyra costata. Ripley Formation (Fate Cretaceous), Union County, Mississippi. Mississippi Geology, 2, 2-5. BISHOP, j. D. D. 1988. Disarticulated bivalve shells as substrates for encrustation by the bryozoan Cribrilina puncturata in the Plio-Pleistocene Red Crag of eastern England. Palaeontology, 31, 237-253. BRETT, c. E. 1985. Tremichnus: a new ichnogenus of circular-parabolic pits in fossil echinoderms. Journal of Paleontology, 59, 625-635. BUSS, L. w. 1986. Competition and community organization on hard substrates in the sea. 517-536. In DIAMOND, J. and CASE, T. J. (eds.). Community ecology. Harper and Row, New York. 16 PALAEONTOLOGY, VOLUME 33 CANCiNO, J. M., MUNOZ, j., MUNOZ, M. and ORELLANA, M. c. 1987. Effects of the bryozoan Membranipora tuberculata (Bose.) on the photosynthesis and growth of Gelidium rex Santelices et Abbott. Journal of Experimental Marine Biology and Ecology, 113, 105-112. CARRICKER, M, R., PALMER, R. E. and PREZANT, R. s. 1980. Functional ultramorphology of the dissoconch valves of the oyster Crassostrea virginica. Proceedings of the National Shellfislieries Association, 70, 139-183. CHiMONiDES, p. J. 1987. Notes on some species of the genus Amathia (Bryozoa, Ctenostomata). Bulletin of the British Museum {Natural History) (Zoology), 52, 307-358. DYRYNDA, p. E. J. 1986. Defensive strategies of modular organisms. Philosophical Transactions of the Royal Society of London, Series B, 313, 227-243. ERNST, H. 1985. Biomuration [m] of folliculinids in Upper Cretaceous cheilostome Bryozoa. 345. In nielsen, c. and LARWOOD, o. p. (eds.). Bryozoa: Ordovician to Recent, Olsen and Olsen, Fredensborg, 364 pp. GLiNSKi, A. 1956. Plumalina conservata n. sp. (Gorgonaria) aus dem Mittel-Devon der Eifel. Senckenbergiana Lethaea, 37, 53-57. HILL, D. and wells, j. w. 1956. Hydroida and Spongiomorphida. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part E. Coelenterata, F81-F89. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. HiLLMER, G. and SCHALLREUTER, R. 1987. Ordovician bryozoans from erratic boulders of northern Germany and Sweden. 113-119. In ross, j. r. p. (ed.). Bryozoa: present and past. Western Washington University, Bellingham, 333 pp. JACKSON, J. B. c. 1979. Morphological strategies of sessile animals. 499-555. In larwood, g. and rosen, b. r. (eds.). Biology and systematics of colonial organisms. Academic Press, London, 589 pp. and COATES, a. g. 1986. Life cycles and evolution of clonal (modular) animals. Philosophical Transactions of the Royal Society of London, Series B, 313, 7-22. and HUGHES, T. p. 1985. Adaptive strategies of coral-reef invertebrates. American Scientist, 73, 265-274. KERSHAW, s. 1987. Stromatoporoid-coral intergrowths in a Silurian biostrome. Lethaia, 20, 371-380. LARWOOD, G, p. 1962. The morphology and systematics of some Cretaceous cribrimorph Polyzoa (Pelmatoporinae). Bulletin of the British Museum {Natural History) (Geology), 6, 1-285. LIDDELL, w. D. and BRETT, c. E. 1982. Skeletal overgrowths among epizoans from the Silurian (Wenlockian) Waldron Shale. Paleobiology, 8, 67-78. NICOL, D. 1978. Shell-cemented pelecypods. Elorida Scientist, 41, 39-41. PALMER, T. J. and WILSON, M. A. 1988. Parasitism of Ordovician bryozoans and the origin of pseudoborings. Palaeontology, 31, 939-949. PHILLIPS, J. 1829. Illustrations of the geology of Yorkshire. Part I. The Yorkshire Coast. York. 192 pp. PITCHER, c. R. and butler, a. j. 1987. Predation by asteroids, escape response, and morphometries of scallops with epizoic sponges. Journal of Experimental Marine Biology and Ecology, 112, 233-249. POUYET, S. 1978. L’association bryozoaires-cnidaires. Reunion Annuelle des Sciences de la Terre, 6, 319. REID, R. E. H. 1962. Notes on hexactinellid sponges -IV. Nine Cretaceous Lychniscosa. Annals and magazine of Natural History (3), 5, 33^5. ROHR, D. M. and BOUCOT, A. 1989. Zenomorphism, bioimmuration, and biologic substrates: an example from the Cretaceous of Brazil. Lethaia, 22, 213-215. RYLAND, J. s. 1976. Physiology and ecology of marine bryozoans. Advances in Marine Biology, 14, 285^43. SCHAFER, w. 1972. Ecology and palaeoecology of marine University of Chicago Press, Chicago. 568 pp. SCRUTTON, c. T. 1975. Hydroid-serpulid symbiosis in the Mesozoic and Tertiary. Palaeontology, 18, 255-274. STENZEL, H. B. 1971. Oysters. In moore, r. c. and teichert, c. (eds.). Treatise on invertebrate paleontology. Part N. Bivalvia 3, N953-N1224. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. TAYLOR, p. D. 1978. A Jurassic ctenostome bryozoan from Yorkshire. Proceedings of the Yorkshire Geological Society, 42, 211-216. 1979. Palaeoecology of the encrusting epifauna of some British Jurassic bivalves. Palaeogeography, Palaeoclimatology, Palaeoecology, 28, 241-262. 1984. Adaptations for spatial competition and utilization in Silurian encrusting bryozoans. Special Papers in Palaeontology, 32, 197-210. 1988. A probable thecate hydroid from the Upper Cretaceous of southern Lngland preserved by bioimmuration. Paldontologisches Zeitschrift, 62, 167-174. 1990. Bioimmured ctenostomes from the Jurassic and the origin of the cheilostome Bryozoa. Palaeontology, 33, 19-34. TAYLOR: BIOIM M U RATION 17 TODD, c. D and TURNER, s. J. 1988. Ecology of intertidal and sublittoral cryptic epifaunal assemblages. II. Nonlethal overgrowth of encrusting bryozoans by colonial ascidians. Journal of Experimental Marine Biology and Ecology, 115, 113-126. TOULMIN SMITH, J. 1847. On the Ventriculidae of the Chalk; including the description of peculiar characters of structure observed in their tissues. Annals and Magazine of Natural History, 20, 73-97. VIALOV, o. s. 1961. Phenomena of vital immuration in nature. Dopovidi Akademf N auk Ukravin skoi RSR, 11, 1510-1512. VINE, G. R. 1892. Notes on the Polyzoa, Stomatopora and Proboscina groups, from the Cornbrash ofThrapston, Northamptonshire. Proceedings of the Yorkshire Geological Society, 12, 247-258. VOIGT, E. 1956. Der Nachweis des Phytals durch Epizoen als Kriterium der Tiefe vorzeitlicher Meere. Sonderdruck aus der Geologischen Rundschau, 45, 97-119. 1966. Die Erhaltung verganglicher Organismen durch Abformung infolge Inkrustation durch sessile Tiere. Neues Jahrbuch fiir Geologic und Paldontologie Abhandlungen, 125, 401^22. 1968a. Eine fossile Art von Arachnidium (Bryozoa, Ctenostomata) in der Unteren Kreide Nord- deutschland. Ibid, 132, 87-96. 19686. Uber Immuration bei fossilen Bryozoen dargestellt an Funden aus der Oberen Kreide. Nachrichten der Akademie der Wissenschaften in Gottingen. II. Mathematisch-Physikalische Klasse, 1968, 47-63. 1972. Amathia immurata n. sp., ein durch Bioimmuration erhaltenes ctenostomes Bryozoon aus der Maastrichter Tuffkreide. Paldontologisches Zeitschrift, 46, 87-92. 1973a. Environmental conditions of bryozoan ecology of the hardground biotope of the Maastrichtian Tuff-Chalk, near Maastricht (Netherlands). 185-197. In larwood, g. p. (ed.). Living and fossil Bryozoa. Academic Press, Eondon, 634 pp. 19736. Hydralhnania graptolithiformis n. sp., eine durch Biomuration erhaltene Sertulariidae (Hydroz.) aus der Maastrichter Tuffkreide. Paldontologisches Zeitschrift, 47, 25-31. 1977. Arachnidium jurassicum n. sp. (Bryoz. Ctenostomata) aus dem mittleren Dogger von Goslar am Harz. Neues Jahrbuch fiir Geologic und Paldontologie Abhandlungen, 153, 170-179. 1979. The preservation of slightly or non-calcified fossil Bryozoa (Ctenostomata and Cheilostomata) by bioimmuration. 541-564. In larwood, g. p. and abbott, m. b. (eds.). Advances in bryozoology. Academic Press, London. 1980. Arachnidium longicauda n. sp. (Bryozoa Ctenostomata) aus der Maastrichter Tuffkreide (Ob. Kreide, Maastrichtium). Neues Jahrbuch fiir Geologic und Paldontologie Monatschefte, 1980, 738-746. 1981. Erster fossilier Nachweis des Algen-Genus Fosliella Howe, 1920 (Corallinaceae; Rhodophyceae) in der Maastrichter un Kunrader Kreide (Maastrichtium, Oberkreide). Facies, 5, 265-282. 1983. Fossilerhaltung durch Bioimmuration: Aetea (Lamouroux) (Bryozoa Cheilostomata) aus dem Mittelmeerraum (Pliozan und rezent). Ibid, 9, 285-310. and domke, w. 1955. Thalassocharis bosqueti Debey ex Miquel, ein structurell erhaltenes Seegras aus der holliindischen Kreide. Mitteilungen aus dem Geologischen Staatinstitut in Hamburg, 24, 87-102. and ERNST, H. 1985. Regressive Astogenese bei Nudonychocella n.g. n.sp. und anderen Bryozoen aus der Tuffkreide von Maastricht. Paldontologisches Zeitschrift, 59, 57-73. AND HARMELiN, j.-G. 1986. Erster mutmaBlicher fissiler Nachweis des Chlorophyceengenus Codium in der Oberkreide. Senckenbergiana Maritima, 18, 253-273. WEST, R. R. 1977. Organism-substrate relations; terminology for ecology and palaeoecology. Lethaia, 10, 71-82. WHITTINGTON, H. B. and CONWAY MORRIS, s. 1985. Extraordinary fossil biotas; their ecological and evolutionary significance. Philosophical Transactions of the Royal Society, Series B, 311, 1-191. wiTMAN, J. D. and SUCHANEK, T. H. 1984. Mussels in flow: drag and dislodgement by epizoans. Marine Ecology Progress Series, 16, 259-268. Typescript received 15 December 1988 Revised typescript received 9 February 1989 P. D. TAYLOR Department of Palaeontology British Museum (Natural History) London SW7 5BD PA I O / (ST'i- BIOIMMURED CTENOSTOMES FROM THE JURASSIC AND THE ORIGIN OF THE CHEILOSTOME BRYOZOA by P. D. TAYLOR Abstract. Soft-bodied ctenostome bryozoans, preserved as bioimmurations following overgrowth by encrusting organisms with hard skeletons, are described from the Middle and Upper Jurassic of England and Normandy. They include one new genus, C ardour admidiiim, and three new species, C. haiitai, C. voigti and Arachnoidella ahiisensis. These and new bioimmured specimens of Arachnidium smilhii (Phillips) show fine- scale details of original morphology as well as artefacts caused by partial collapse of zooids during overgrowth. The D-shaped zooidal orifices present in Cardoaradmidium probably indicate that the zooids were operculate. Opercula are one of the two apomorphies of the closely-related cheilostomes, and therefore the Oxfordian Cardoaradmidium is placed in the stem-group of the Cheilostomata, which first appear in the Tithonian. CTENOSTOMES are unique among marine bryozoan orders in lacking a calcified skeleton. Accordingly, they are less well-represented in the fossil record than groups such as trepostomes, fenestrates, cyclostomes and cheilostomes, more familiar to palaeontologists. Fossil ctenostomes, however, can be preserved as borings (see Pohowsky 1978) and as bioimmurations, i.e. natural moulds and casts formed as a result of overgrowth by organisms with mineralized skeletons (see Taylor 1990). Boring ctenostomes, first known from the Upper Ordovician, are generally regarded as a specialized group or groups. In contrast, bioimmured fossil ctenostomes include several uniserial encrusting species which have many features apparently primitive among marine Bryozoa. During recent years, bioimmured ctenostomes have been discovered with increasing frequency in the Jurassic and Cretaceous, mainly through the work of Voigt ( 1966, 1968, 1972, 1977, 1979, 1980). The present paper utilizes an uncoated SEM technique (Taylor 1986) to describe some finds and redescribe earlier finds from the Jurassic. The quality of preservation of some of the new specimens exceeds that previously known and has revealed hitherto unknown morphological features. Artefacts formed during the process of bioimmuration process are also illustrated. Jurassic bioimmured ctenostomes have special significance because they almost certainly include representatives of the stem-group of the order Cheilostomata. Cheilostomes, the dominant bryozoans of present-day faunas, first appear in deposits of latest Jurassic age. Comparative anatomical studies have led to the belief that the ancestor of the cheilostomes was a ctenostome or, more strictly, that living ctenostomes are the primitive sister-group of living cheilostomes. The description here of a new bioimmured Upper Jurassic ctenostome which apparently possesses one of the two apomorphic features characteristic of the Cheilostomata supports this hypothesis of cheilostome origins and gives the sequence of appearance of the two apomorphies. SYSTEMATIC PAEAEONTOLOGY Material. Specimen repositories are abbreviated as follows: BMNH, British Museum (Natural History); YM, Yorkshire Museum, York; VH, Voigt Collection, Universitat Hamburg. Order ctenostomata Busk, 1852 Suborder carnosa Gray, 1841 Family arachnidiidae Hincks, 1877 Genus arachnidium Hincks, 1859 I Palaeontology, Vol. 33, Part I, 1990, pp. 19-34, 3 pls,| © The Palaeontological Association 20 PALAEONTOLOGY, VOLUME 33 Type species. Arachnidium hippothooides Hincks, 1859. Discussion. Arachnidium Hincks is founded on a comparatively rare present-day type species which has been redescribed recently by Hayward (1985, p. 78). D’Hondt (1983) included five living species in the genus, all possessing uniserial branching colonies with zooids pyriform or ovoidal in outline shape. In common with many ctenostomes, the surface morphology of species of Arachnidium is comparatively simple, and descriptions of living species generally include only a small number of external morphological characters (notably size, shape and ornamentation of the zooids). These characters are equally available in bioimmured fossil Arachnidium. Not available in fossil Arachnidium are aspects of polypide morphology (e.g. tentacle number) which have been used in species descriptions. The difference between Arachnidium and Arachnoidella d’Hondt, 1983 (see p. 26) is rather slight, the latter possessing rather longer peristomes and sometimes developing lateral crenulations close to the substratum. However, pending further study of living species, the two genera are retained. Range. Jurassic (Aalenian) to Recent. Arachnidium smithii (Phillips, 1829) Plate 1, figs. 1-6; Plate 2, figs. 1 and 2; text-figs. 1, 2a, b, 3a, b. 1829 Cellaria smithii Phillips, p. 143. pi. 7, fig. 8. 1875 Hippothoa smithii (Phillips); Phillips, p. 242, pi. 7, fig. 8. 1892 Stomatopora phiUipsii Vine, p. 250, pi. 12, figs. 1-3. 1894 Stomatopora smithi (Phillips); Gregory, p. 58, fig. 1. 1895 Stomatopora smithi (Phillips); Gregory, p. 226. 1896 Stomatopora smithi (Phillips); Gregory, p. 56, fig. 8. 1907 Stomatopora phillipsi Vine; Lang, fig. 6. 1911 Corynotrypa smithi (Phillips); Bassler, p. 521. fig. 22. 1935 Stomatopora smithi (Phillips); Melmore, p. 1, test-figs. 1 and 2. 1977 Arachnidium jurassicum Voigt, p. 172, figs. 1^. 1978 Arachnidium smithii (Phillips); Taylor, p. 214, pi. 7, figs. 1^. 1980 Arachnidium smithii (Phillips); Voigt, fig. 4a. 1980 Arachnidium jurassicum Voigt; Voigt, fig. 4b. Holotype. YM 78, Cornbrash, Scarborough, Yorkshire. This specimen (PI. 1, figs. 1 and 3; text-fig. 1), a cast bioimmuration attached to the type specimen of the bivalve Cardium citrinoideum Phillips, is almost certainly of Callovian age (macrocephalus Zone). Other material. BMNH D 31 144, Cornbrash, Thrapston, Northamptonshire; the holotype of Stomatopora phillipsi Vine, 1892; a cast bioimmuration attached to the brachiopod Obovothyris and undoubtedly originating from the Bathonian (discus Zone). BMNH D 53156, Bathonian, Bradford Clay (discus Zone), Cirencester, Gloucestershire, J. P. Woodward Collection; a mould bioimmuration on the attachment area of EXPLANATION OF PLATE 1 Figs. 1-6. Arachnidium smithii (Phillips, 1829). 1, 3, holotype, YM 78, natural cast bioimmuration, Scarborough, Cornbrash (Callovian). 1, crowded zooids with overgrowing branches, x28. 3, partly collapsed zooid (with well-preserved orifice) overgrown by another zooid, x 100. 2, 4, BMNH D 57497, natural cast bioimmuration. South Ferriby, Kimmeridgian. 2, collapsed zooids cast in pyrite, x 50. 4, calcite cast zooids emerging from beneath the cover of the bioimmuring bivalve (bottom), x 16. 5 and 6, BMNH D 57492, mould bioimmuration, Villers-sur-mer, Oxfordian. 5, uniserial chain of zooids x 35. 6, orifice, x240. All illustrations are back-scattered electron micrographs of uncoated specimens. PLATE I TAYLOR, Arachnidium PALAEONTOLOGY, VOLUME 33 TEXT-HG. I. Arachnidium smithii (Phillips. 1829), hololype, YM 78, Scarborough, Cornbrash (Callovian). Extensive colony preserved as a natural cast attached to the surface of a bivalve shell. Back-scattered scanning electron micrograph of an uncoated specimen, x 19. a small Praee.xogyra. BMNH D 57492, 4-5, D 58002, Oxfordian, float probably from the ‘Couches a Myopliorella hudlestoni et Lopha gregaria' (plicatilis Zone), Vaches-Noires, Villers-sur-mer, Normandy, collected by P. D. Taylor, 1985; apart from D 58002. all are mould bioimmurations on the attachment areas of gryphaeate oysters; D 58002 is a partially exposed, partly infilled cast bioimmuration. BMNH D 57496, Kimmeridgian, Bed M8 of Birkelund et al. (1983) (mid nmtahilis Zone), Westbury, Wiltshire, collected by P. Wignall, 1986; mould bioimmuration. partly infilled by pyrite, on the attachment area of a small oyster. BMNH D 57497, Kimmeridgian, Bed 9 of Birkelund and Calloman (1985) (lower haylei Zons), South Ferriby, Humberside, collected by M. Simms, 1986; a cast bioimmuration (PI. 1, figs. 2 and 4; PI. 2, figs. 1-2), partly calcitic and partly pyritic, attached to Delloideum delta (Sowerby). The holotype of Arachnidium Jiirassicum Voigt, 1977, unavailable for study during 1987, is VH 1899. Middle Dogger (Po/v/’/nn«-Schichten, Dogger 7), Goslar, West Germany; this is a cast bioimmuration attached to a guard of Metateuthis. Description. Colony adnate, consisting of ramifying, uniserial branches of zooids (text-figs. 1 and 2). Branches often gently curved, new branches arising with variable frequency, sometimes in pairs but sometimes singly, by distolateral budding at an angle of between about 45° and 90° to the parent branch. Crowding of zooids EXPLANATION OF PLATE 2 Figs. 1 and 2. Arachnidium ,S7?;/7/u; (Phillips. 1829), BMNH D 57497, South Ferriby, Kimmeridgian. 1, pyritic cast of collapsed zooid with wrinkled frontal membrane, x 100. 2, orifice of the same zooid, x270. Figs. 3-6. Arachuoidella ahuseusis sp. nov., holotype, BMNH D 57637, South Ferriby. Kimmeridgian. 3, crowded zooids moulded on the underside of the bioimmuring bivalve, x 1 5. 4, astogenetic increase in zooid length along a branch originating as a lateral bud from the zooid on the left, x 28. 5, mould of zooid showing marginal processes; note lack of visible orifice which is located on a distally-directed peristome hidden in the shadow at the distal end of the zooid, x 93. 6, marginal processes in zooid preserved as a calcitic cast x 1 36. All illustrations are back-scattered electron micrographs of uncoated specimens. PLATE 2 TAYLOR, Araclwidiuni, Arachnoidella 24 PALAEONTOLOGY. VOLUME 33 TEXT-FIG. 2. Mould bioimmuration of Arachnidiiun smithii (Phillips, 1829) visible on the large attachment area of the bivalve Gryphaea from the Oxfordian of Villers-sur-mer, BMNH D 57492. A, general view of the bioimmuring Gryphaea, x L2 B, attachment area showing prominent branch of zooids running from top to bottom, and dense settlement of the foraminifer Nuheculinella on the younger, peripheral parts of the attachment area, x2-l. may occur in colonies with a high frequency of branching, and can be accompanied by branch overgrowths (PI. 1, fig. 3) and/or abutment of growing branches against existing branches. In early astogeny (visible only in D 57495), the ancestrula apparently buds proximal and distal periancestrular zooids which initiate two primary colony branches growing in opposite directions. A brief primary zone of astogenetic change is marked by an increase in zooid size with generation. Putative ancestrula about 0-23 mm long by 0-17 mm wide. Autozooids moderately pyriform in frontal outline shape, narrow proximally. achieving maximum width about or a little distally of mid-length, and with rounded distal ends. Length and width of autozooids is very variable between colonies, measured length ranging from 0 47 to I 00 mm, width from 0-25 to 0 45 mm, with TEXT-FIG. 3. Zooid morphology in Jurassic arachnidiids. A and B, Arachnidium smithii (Phillips, 1829) showing variation in zooid size (A is based on YM 78, B on BMNH D 57497). C, Aracimoidella abusemis sp. nov. D, Cardoaracimidiiim bantai sp. nov. E, C. voigti sp. nov. zooids generally about twice as long as wide. Frontal wall gently convex, often preserved with flattened lateral margins flanking a raised median area which includes the orifice. Longitudinal wrinkles and folds (PI. 1, fig. 2; PI. 2, fig. 1) developed on frontal walls of zooids cast by pyrite in BMNH D 57497. In BMNH D 58002, the casting mineral near an autozooidal orifice contains shallow pits which may perhaps represent original pits on the frontal wall. Orifice simple, subterminal, located opposite or a little distal to origins of lateral buds, subcircular to transversely elliptical in shape (PI. 1, fig. 6), about 0 04-0 08 mm in diameter, occasionally with a slightly raised rim. An eccentrically perforated structure partly occludes the orifice of one zooid of BMNH D 57497 (PI. 2, fig. 2). Kenozooids may develop in regions of crowding as small subtriangular-shaped buds (?aborted autozooids) lacking an orifice (text-fig. I). Dimensions {mm)- X autozooid length (range) X autozooid width (range) YM 78 (holotype) 0-59 0-31 (0-50-0-68) (0-26-0-36) BMNH D 31144 0-52 0-30 (0-47-0-59) (0-26-0-36) VH 1899 {fide Voigt 1977) 0-72 0-33 (0-63-0-84) (0-25-0-41) BMNH D 57492 0-74 0-32 (0-63-0-90) (0-27-0-36) BMNH D 57494 0-66 0-38 (0-54-0-74) (0-30-0-45) BMNH D 57495 0-62 0-32 (0-53-0-71) (0-30-0-38) BMNH D 57496 0-76 0-38 (0-62-0-90) (0-32-0-42) BMNH D 57497 0-88 0-40 (0-80-1.00) (0-38-0-45) 26 PALAEONTOLOGY, VOLUME 33 Discussion. The present concept of Araclinidium smithii (Phillips) encompasses colonies exhibiting a wide range of variation in autozooid length and width (compare text-figs. 3 a and b). It seems possible that A. smithii may represent a species complex. However, the overlapping dimensions of colonies from different stratigraphical horizons prohibit convenient splitting into two or more putative species. Furthermore, there are no obvious differences in zooid shape, budding pattern etc. which might be used for this purpose. Accordingly, Stomatopora phiUipsii Vine and Arachnidium jurussicum Voigt are taken into synonymy with A. smithii. It should be noted that the ctenostome nature of A. smithii and S. phiUipsii were unknown when Voigt (1977) erected A. jurassicum ; neither species was revised until Taylor (1978) and both were presumed to be cyclostomes of the common Jurassic genus Stomatopora Bronn. The Barremain species A. hrandesi Voigt strongly resembles A. smithii but has considerably larger zooids; according to Voigt (1968), zooid length ranges from 1-65 to 1-75 mm. The appearance of the zooids within colonies of A. smithii may vary according to the extent of their collapse during overgrowth. Uncollapsed zooids have evenly convex frontal walls (PI. 1, fig. 5), whereas partially collapsed zooids generally have flattened lateral margins flanking a raised median area (PI. 1, fig. 3). The pattern of collapse undoubtedly reflects some aspect of original zooidal morphology. Dried zooids of the Recent species Arachnoidea annosciae figured by d’Hondt and Geraci (1976, fig. 4) show a similar collapsed structure. Banta (1975, fig. 22), in a drawing of bioimmured A. hrandesi, labels the margins of the zooids as ‘gymnocyst’ and the median area as ‘opesium’, suggesting an organization like that of anascan cheilostomes in which the frontal membrane is attached to rigid lateral walls and stretches over the opesium (see Taylor 1981). The parietal muscle (whose contraction depresses the frontal membrane and brings about polypide eversion) were possibly attached to the frontal membrane along the well-defined lines between flattened lateral margins and raised median area. Wrinkling and folding of the frontal wall in pyrite casts of zooids in BMNH D 57497 (PI. 1, fig. 2; PI. 2, fig. 1 ) is a further indication of partial collapse during overgrowth. Transverse contraction of the zooid resulted in the relatively non-elastic cuticle being thrown into a series of folds and wrinkles running subparallel to the length of the zooid. Stratigraphical range. Aalenian (polyplocus Zone) to Kimmeridgian (baylei Zone). Genus arachnoidella d’Hondt, 1983 Type species. Arachnoidea annosciae d’Hondt and Geraci, 1976. Discussion. Arachnoidella was originally proposed by d’Hondt ( 1983) as a subspecies of Arachnoidea Moore, 1903. However, Gordon (1986) elevated Arachnoidella to genus rank because the type species of Arachnoidea (A. raylankesteri Moore) is a freshwater species in which the zooids are interconnected by anastomosing filaments, absent in marine Arachnoidella. As noted on p. 20, the distinction between Arachnoidella and Arachnidium is not great, and the former may eventually prove to be a junior subjective synonym of the latter. D’Hondt (1983) recognized eight Recent species of Arachnoidella. The genus has not been previously recorded as a fossil. Range. Jurassic (Kimmeridgian) to Recent. Arachnoidella ahusensis sp. nov. Plate 2, figs. 3-6; text-fig. 3c; text-fig. 4 Holotype. BMNH D 57637, Kimmeridgian, Bed 9 of Birkelund and Calloman (1985) (lower haylei Zone), South Ferriby, Humberside, collected by P. D. Taylor, 1987; predominantly a mould bioimmuration (PI. 2, figs. 3-5) but with a few zooids cast by calcite-(PI. 2, fig. 6) on the attachment area of the bivalve Deltoideum delta (Sowerby). Paratype. BMNH D 57602, details as for holotype; a cast bioimmuration (text-fig. 4) partly exposed by abrasion of the thin overgrowing organism which encrusts a bivalve shell fragment. TAYLOR: CTENOSTOME BRYOZOANS 27 TliXT-HG. 4. Anichnoidella ahusensis sp. nov., partly collapsed zooids preserved as matural casts, BMNH D 57602. South Ferriby, Kimmeridgian. A. zooid with distally-directed peristome exposed by abrasion of the bioimmuring organism, x 162. B, distal part of frontal membrane of another zooid showing minute pores which are absent from the peristome base at the top of the figure; note presenee of 3 sets of fibres in eorroded frontal membrane left of centre, x 890. Back-scattered scanning electron micrographs of an uncoatcd specimen. Derivalioii of name. From Abus. Roman name for the River Humber which is close to the type locality of South Ferriby. Diagnosis. Arachnoidella with small zooids having about 30 marginal processes; orifice situated terminally on a distally orientated peristome; caudae account for half or more of total zooid length and increase in length during early branch astogeny. Description. Colony adnate, consisting of branches of uniserially-arranged zooids. New branches arise as lateral buds and diverge from the parent branch at an angle averaging about 60°. Frequent branch ramification results in areas of zooid crowding and branch overgrowth (PI. 2, fig. 3). Ancestrula unknown. Secondary zones of astogenetic change oecur in the early parts of new branches: caudal length increases progressively for the first three or four generations of zooids (PI. 2, fig. 4). Autozooids pyriform with a cauda of variable length, generally accounting for half or more of the total length of the zooid, succeeded distally by a longitudinally elliptical dilated frontal wall. Autozooid length variable within colonies, observed range 0-42-F26mm (x = ()-92mm); width ranging from 0-2I-0-27 mm (x = 0-25 mm). Distal frontal wall gently convex (PI. 2, fig. 5) or flattened (text-fig. 4a) as a result of overgrowth, and bearing minute pores which are absent from the cauda and peristome (text-fig 4b). Lateral buds arise a little distally of the level of maximum width on the distal frontal wall. Marginal processes visible around the distal Irontal walls of some zooids. numbering about 30 per zooid (PI. 2, figs. 5 and 6). Orifice situated at the extreme 28 PALAEONTOLOGY, VOLUME 33 distal end of the zooid on a peristome (text-fig 4a) which is directed distally and is about 0-05 mm wide at its base. Discussion. This species is distinguished from previously described Jurassic arachnidiids by the extreme distal location of the orifice. Secondary zones of astogenetic change, manifested by progressive increase in zooid length along new branches, are considerably better developed than in other species. In comparison with Recent species of Arachnoidella, A. ahusensis zooids have fewer marginal processes than A. annosciae (c. 30 versus 60) but more than all the other species tabulated by d’Hondt (1983). The preservation of the paratype (text-fig. 4) deserves comment. Naturally-cast zooids are partially visible beneath the thin basal layer of the overgrowing organism. The casting is of a very high quality, and shows clearly the presence of minute pores on the frontal wall and their absence on the cauda and peristome. Although the identity of the casting material is unknown, an abraded area of the cast of one zooid has a fibrous structure (text-fig. 4b). Fibres are arranged in 3 conjugate sets orientated at 120° to one another. One of the sets exactly parallels the long axis of the zooid. This suggests that the fibres may reflect some aspect of the original structure of the zooidal frontal wall. Possibly they are pseudomorphs of one of the organic components of the zooid body wall. Stratigraphical range. Kimmeridgian (baylei Zone). Genus cardoarachnidium gen. nov. Type species. Cardoarachnidium bantai sp. nov. Derivation of name. Cardo-, Latin for hinge, in reference to the apparent presence of a hinged operculum. Diagnosis. Arachnidiidae in which the autozooidal orifice is D-shaped with a straight proximal edge and a curved distal/lateral edge. Discussion. This new genus is established for two new species of Jurassic bioimmured ctenostomes which diflfer from species of Arachnidium and Arachnidiella in having a D-shaped autozooidal orifice suggesting the presence of a hinged operculum during life. Opercula are generally regarded as being absent in ctenostome bryozoans. However, Banta (1975) notes that structures ‘virtually indistinguishable from opercula’ occur in the living ctenostome genera Flustrellidra, Elzerina and Hislopia, and d’Hondt (1983) describes the lip of the orifice in Haywardozoon as ‘simulating an operculum’. An important difference between Cardoarachnidium and these Recent genera is that the operculum of Cardoarachnidium is a well-defined, univalved structure, whereas those of the Recent genera are bivalved - they are commonly described as bilabiate - with a distal hinged flap as well as a generally larger proximal hinged flap. Although the phylogenetic significance of the inferred operculum in Cardoarachnidium is discussed fully on page 32, it is appropriate to note here that its presence allows Cardoarachnidium to be placed within the stem-group of the Cheilostomata. If correct, Cardoarachnidium is more closely related (in cladistic terms) to the Cheilostomata than it is to other genera of the Arachnidiidae. The latter family as understood here and by previous authors is almost certainly paraphyletic, and demands future division into its constituent monophyletic clades. Until this is accomplished Cardoarachnidium is assigned to the Arachnidiidae. Cardoarachnidium bantai sp. nov. Plate 3, figs. 1-3; text-fig. 3d Holotype. BMNH D 57499 (a), Sandsfoot Clay (sandy top). Upper Oxfordian (serratum or regidare Zones), below Sandsfoot Castle, Weymouth, Dorset, collected by W. J. Kennedy, N. J. Morris and C. P. Palmer, 1971. TAYLOR: CTENOSTOME BRYOZOANS 29 CHEILOSTOMATA TEXT-FIG. 5. Simplified cladogram showing inferred phylogenetic relationships between genera of arachnidiid ctenostomes present in the Jurassic and the Cheilostomata. Cardoarachnidium is depicted as a stem-group cheilostome because it apparently possesses one (opercula) but not the second (calcification) apomorphic feature of the crown-group cheilostomes. The trichotomy between Aracimidiiim, Aradmoidella and Cardoarachnidium -I- the Cheilostomata is unresolved. A mould bioimmuration on the attachment scar of the bivalve Deltoideum delta (Sowerby), and intergrown with the holotype of Cardoarachnidium voigti sp. nov. (see p. 30). Derivation of name. In recognition of the bryozoological research of Dr W. C. Banta (The American University, Washington). Diagnosis. Cardoarachnidium with pyriform zooids lacking significant caudae; orifice located subterminally. Description. Colony adnate, consisting of uniserial branches, often slightly curved, from which daughter branches arise sporadically as lateral buds and diverge at angles of 60-90° to the parent branch (PI. 3, fig. 1). Ancestrula budding a proximal periancestrular zooid only; no distal periancestrular zooid visible (PI. 3, fig. 2). Ancestrula of similar shape but smaller than the periancestrular zooid and later zooids, with a length of 0-30 mm, width 017 mm and D-shaped orifice 0 06 mm wide. Autozooids moderately pyriform in outline shape, narrow proximally, rounded distally, achieving maximum width rather distal of mid-length. Length of autozooid about 2-5 times the width, length averaging 0-76 mm (range 0-63-0-93 mm), width averaging 0-32 mm (range 0-27-0-39 mm). Frontal wall gently convex, sometimes with slightly flattened lateral margins. Close to the orifice in one autozooid, the basal layer of the bioimmuring bivalve is penetrated by small pores which may represent a negative impression of an originally spinose or pustulose frontal wall and operculum. Orifice when visible is D-shaped (PI. 3, fig. 3), wider than long, small, about 0-05 X 0.08 mm in size, situated more-or-less opposite sites of lateral bud origin. The straight proximal edge is less well-defined than the crescentic lateral/distal edge which, together with the shape of the orifice, suggests that an operculum was hinged on the proximal edge of the orifice. In many zooids the orifice cannot be seen, possibly because of poor quality bioimmuration. Discussion. Colony-form and autozooid shape and size are very similar to A. sniithii, although the length: width ratio of the autozooids is a little greater in C. hantai. However, the D-shaped orifice serves to distinguish C. hantai from A. sinithii. Stratigraphical range. Oxfordian (serratum or regulare Zones). 30 PALAEONTOLOGY, VOLUME 33 C anloarachnidium voigti sp. nov. (Plate 3, figs. 4-6; text-fig. 3e) Holotype. BMNH D 57499 (b), Sandsfoot Clay (sandy top). Upper Oxfordian {serratiim or regulare Zones), below Sandsfoot Castle, Weymouth, Dorset, collected by W. .1. Kennedy, N. .1. Morris and C. P. Palmer, 1971. A mould bioimmuration on the attachment scar of the bivalve Deltoideum dcdra (Sowerby), and intergrown with the holotype of C. Inintai sp. nov. (see p. 28). Derivation of name. In honour of Professor E. Voigt (Universitat Hamburg). Diagnosis. Cardoaraclinidium with slender, caudate zooids; orifice located terminally. Description. Colony adnate, consisting of uniserial branches from which daughter branches arise as lateral buds at an angle of about 45-90° to the parent branch (PI. 3, fig. 4). Branch ramification occurs frequently, most zooids producing two lateral buds. Growing branches generally terminate on meeting established branches, giving an anastomosing network in some areas of the colony. Ancestrula unknown. Autozooids slender (PI. 3, fig. 5), 0-74-1 -65 mm long (x 1-20 mm), a narrow proximal cauda accounting for a third to three quarters of their length and succeeded distally by a dilated part the shape of a longitudinally elongate ellipse 0-32-0-45 mm long (x 0-37 mm) by 0-20-0-29 mm wide (x 0.22 mm). Distal frontal wall gently convex or flat-topped; ornamentation not apparent. Laterally-budded daughter zooids often rather shorter than their parental zooids, but clearly-defined secondary zones of astogenetic change not obvious. Lateral budding loci situated a little distally of mid-length on the distal frontal wall. Orifice terminal (PI. 3, fig. 6), small, slightly raised, a poorly-defined D-shape, about 0-07 mm wide. Discussion. This species is immediately distinguished from most other Jurassic bioimmured ctenostomes by the long length of the caudae. In this feature it most closely resembles the Maastrichtian Arachnidiiim longicaiida Voigt, 1980 which has a roughly circular distal frontal wall, and subcircular orifice situated much further proximally than that of C. voigti. Although the caudae of ArachnoideUa ahusensis (see p. 26) may also be long, the orifice of this species is more terminally situated and the zooids often possess marginal processes. None of the autozooidal orifices are as well-preserved as those of C. hantai which shares the same substratum. Nevertheless, C. voigti also appears to have a D-shaped orifice suggesting the presence of an operculum hinged on the straight proximal edge of the orifice. This D-shape is apparently not an artefact caused by the overgrowing organism pushing over short peristomes in a distal direction ; if it were the D-shape would be better developed in zooids orientated parallel to the growth direction of the overgrowing organism, which is not the case. Stratigraphical range. Oxfordian (serratiun or regulare Zones). DISCUSSION Present-day bryozoan faunas are dominated by species belonging to the order Cheilostomata. For example, in a survey of seven regional faunas (Taylor 1981), the percentage contribution of cheilostome species ranged from 69 to 91. However, cheilostomes are a geologically young order. EXPLANATION OF PLATE 3 Ligs. 1-3. Cardoaraclinidium hantai sp. nov., holotype, BMNH D 57499 (a), mould bioimmuration, Weymouth. Oxfordian. I, uniserial branch, x 14. 2, ancestrula (lower left) with proximal ancestrular bud (centre) and its distal bud, x 55. 3, D-shaped orifice of proximal periancestrular zooid, x 15. Ligs. 4-6. Cardoaraclinidium voigti sp. nov., holotype, BMNH D 57499 (b), mould bioimmuration, Weymouth, Oxfordian. 4, x 16. 5. caudate zooids, x 38. 6, distal part of zooid showing terminal orifice, x 120, All illustrations are back-scattered electron micrographs of uncoated specimens. PLATE 3 TAYLOR, Cardoaraclinidiuni 32 PALAEONTOLOGY, VOLUME 33 first appearing at the end of the Jurassic (Pohowsky 1973) and not diversifying appreciably until the late Cretaceous (see Taylor 1988). Several hypotheses have been proposed to account for the phylogenetic origin of cheilostomes (Banta 1975). Cheilostome ancestry has been sought in the Fenestrata (Ulrich 1890), Cyclostomata (Dzik 1975) and Ctenostomata (e.g. Banta 1975; Cheetham and Cook in Boardman et al. 1983). There are severe difficulties with hypotheses which propose an origin from either of the first two orders which are stenolaemate bryozoans. Similarities do exist between some cheilostome species and fenestrates (which may share box-shaped zooecia), and between other cheilostome species and stomatoporid or corynotrypid cyclostomes (which can both have uniserial colonies with pyriform zooecia), but these are best explained by homoplasy in the first case (Tavener-Smith 1971) and plesiomorphy in the second. Comparative anatomical studies of living bryozoans support the theory of a ctenostome ancestry for the cheilostomes, or more strictly, that ‘ctenostomes’ (a primitive, paraphyletic grouping of marine bryozoans) include the sister- group of cheilostomes. As Banta (1975) points out, there are extremely few differences between many carnosan ctenostomes and primitive cheilostomes. The Cheilostomata are very probably a monophyletic clade which can be distinguished from their sister group ctenostomes by two apomorphic characters; calcification of parts of the zooid body walls, and the possession of an operculum to close the orifice on retraction on the tentacles. All living cheilostomes have some degree of calcification, although in certain genera (e.g. Memhranipora. Fliistra) this is slight. Most cheilostomes possess opercula; absence of the operculum in the feeding zooids of a few genera (e.g. Bugula) is undoubtedly a result of their secondary loss, as shown by the presence in the same colonies of polymorphic zooids (avicularia) which retain opercula. Extinct cheilostomes for which adequate information is available also possess the two apomorphies. Indeed, they are present even in the oldest known cheilostome, Pyriporopsis portlandensis from the Tithonian (Portlandian) of southern England, which has thickly-calcified zooid vertical walls, and in which the past presence of a non-calcified operculum can be inferred confidently from impressions on the closure plates of degenerated zooids (Pohowsky 1973; Banta 1975; Taylor 1987). Therefore, current knowledge of living and fossil cheilostomes provides no information on the order of appearance of the two apomorphies; did calcification predate opercula or vice-versa? The new genus Cardoarachnidiiim is important because it apparently has one of the two apomorphies of the Cheilostomata. The simple D-shaped opercula of Cardoarachnidiiim closely resemble the opercula of primitive anascan cheilostomes. However, Cardoarachnidiiim clearly did not have calcified zooids and therefore lacked the second apomorphy of the Cheilostomata; bioimmured colonies show no trace of calcified zooid walls, and distortion of the zooids during overgrowth, typical of that seen in other bioimmured ctenostomes, implies that the zooids were soft- bodied. The stem-group concept in phylogenetics has been explained recently by Jefferies et al. (1987). Stem- groups are the paraphyletic ancestral groupings of extinct taxa remaining when the crown groups have been subtracted from the total group. They are distinguished by possessing some but not all of the apomorphies that separate any two monophyletic clades with extant representatives (crown groups). By this criterion, Cardoarachnidiiim can be included in the stem-group of the Cheilostomata. This inferred relationship is expressed in the cladogram depicted in text-fig. 5. Stratigraphical sequence is consistent with the cladogram; arachnidiids date back at least to the Aalenian, Cardoarachnidiiim occurs in the Oxfordian, and the first cheilostome in the Tithonian. The discovery and phylogenetic placement of Cardoarachnidiiim solves the problem of the order in which the two apomorphies of cheilostomes appeared; opercula apparently predate calcification. As the most crownward known representative of the stem-group of the Cheilostomata, Cardoarachnidiiim is a useful outgroup for inferring character polarities (plesiomophic versus apomorphic) during studies of phylogenetic relationships within the cheilostomes. Unfortunately, however, relatively few morphological characters are available in bioimmured Cardoarachnidiiim and these do not, of course, include characters of the calcified skeleton which are generally emphasized during studies of cheilostomes. Nevertheless, Cardoarachnidiiim does suggest the following plesiomorphic character states in cheilostomes: uniserial pattern of colony growth with TAYLOR: CTENOSTOME BRYOZOANS 33 the potential for each zooid to produce a distal and two lateral buds; pyriform zooid shape; and the budding of a proximal periancestrular zooid from the ancestrula. Pyriporopsis, Pyripora and Herpetopora among primitive ‘malacostegan ’ cheilostomes (see Taylor 1987) retain all of these characters, although there is a tendency towards pluriserial growth in Pyriporopsis. Knowledge of Jurassic bryozoans is strongly geographically constrained; very few Jurassic bryozoans have been described outside Europe, and modern studies are mostly concerned with faunas from France and southern England. Within this limited geographical region, however, an interesting pattern of stratigraphical distribution is apparent. Rich cyclostome bryozoan faunas occur locally in the Middle Jurassic (e.g. Gloucestershire Aalenian, Normandy Bathonian), with assemblages containing encrusting and erect tubuliporine and cerioporine cyclostomes. Few ctenostomes are evident, either as borers in the abundant shelly substrata, or as encrusters preserved by bioimmuration. Upper Jurassic bryozoan faunas from southern England and France contain assemblages of encrusting tubuloporine cyclostomes greatly reduced in diversity and abundance. However, ctenostomes are much more prominent than in the Middle Jurassic. They include shell borers as well as the bioimmured encrusters described herein. Considering the fairly low probability of preservation of soft-bodied arachnidiids, their abundance and diversity in the late Jurassic epicontinental sea of north-west Europe may have been substantial, quite possibly exceeding arachnidiid abundance and diversity at the present-day. The origin of the Cheilostomata should be viewed within the context of this time of comparative arachnidiid prevalence. Acknowledgements . I am grateful to M. Simms and P. Wignall for donating specimens for study, and B. Pyrah for arranging a loan from Yorkshire Museum. Professor E. Voigt generously provided discussion, encouragement and hospitality. REFERENCES BANTA, w. c. 1975. Origin and early evolution of cheilostome Bryozoa. Documents des Lahoratorie de Geologie de la Facidte des Sciences de Lyon (hors serie), 3 (2), 565-582. BASSLER, R. s. 1911. Coryiiotrypa, a new genus of tubuliporoid Bryozoa. Proceeding, of the United States National Museum, 39, 497-527. BiRKELUND, T. and CALLOMAN, J .H. c. 1985. The Kimmeridgian ammonite faunas of Milne Land central East Greenland. Gronlands Geologiske Undersogelse Bulletin, 153, 1-56. , , CLAUSEN, c. K., NOHR HANSEN, H. and SAUNAS, I. 1983. The Lower Kimmeridge Clay at Westbury, Wiltshire, England. Proceedings of the Geologists Association, 94, 289-309. BOARDMAN, R. S., CHEETHAM, A. H., BLARE, D. B., UTGAARD, J., KARKLINS, O. L., COOK, P. L., SANDBERG, P. A., LUTAUD, G. and wood, t. s. 1983. Bryozoa. Revised. Volume 1. In robinson, r. a. (ed.). Treatise on invertebrate paleontology. Part G. 1-625. Geological Society of America and University of Kansas Press, Boulder, Colarado and Lawrence, Kansas. BUSK, G. 1852. An account of the Polyzoa and Sertularian Zoophytes collected in the voyage of the Rattlesnake on the coast of Australia and the Louisiade Archipelago. In macgillivray, j. Narrative of the Voyage of H.M.S. Rattlesnake, commanded by the late Captain Owen Stanley, during the years 1846-1850, 1, 343^02. Boone, London. DZiK, J. 1975. The origin and early phylogeny of early cheilostomatous Bryozoa. Acta Paleontologica Polonica, 20, 395-423. GORDON, D. p. 1986. The marine fauna of New Zealand: Bryozoa: Gymnolaemata (Ctenostomata and Cheilostomata Anasca) from the western South Island Continental Shelf and Slope. New Zealand Oceanographic Institute Memoir, 95, 1-121. GRAY, J. E. 1841 . Synopsis of the contents of the British Museum. 43rd edition. British Museum, London, 382 pp. GREGORY, J. G. 1894. Catalogue of the Jurassic Bryozoa in the York Museum. Annual Report of the Yorkshire Philosophical Society, 1893, 58-61. 1895. A revision of the Jurassic Bryozoa. Part 1 . The genus Stomatopora. Annals and Magazine of Natural History (6), 15, 223-228. 1896. Catalogue of the fossil Bryozoa in the Department of Geology, British Museum (Natural History). The Jurassic Bryozoa. British Museum (Natural History), London, 239 pp. HAYWARD, p. J. 1985. Ctenostome bryozoans. Synopses of the British Fauna (New Series), 33, 1-169. 3 PAL 33 34 PALAEONTOLOGY, VOLUME 33 HiNCKS, T. 1859. On some new and interesting forms of British Zoophytes. Report of the British Association for the Advancement of Science, 28, 128. 1877. On British Polyzoa. Part 1. Annals and Magazine of Matured History (4), 20, 212-218. HONDT, J. -L. d' 1983. Tabular keys for identification of the Recent ctenostomatous Bryozoa. Memories de rinstitut oceanographiqiie, Monaco, 14, 1-134. and GERACi, s. 1976. Arachnoidea annosciae n.sp., nouveau bryozoaire ctenostomc Arachnidiidae du Golfe de Genes. Estratto Dagli Anncdi Del Museo Civico Di Storia Natural Di Genova, 81, 139-152. JEFFERIES. R. p. s., LEWIS, M. and DONOVAN, s. K. 1987. Protocvstites menevensis - a stem-group chordate (Cornuta) from the Middle Cambrian of South Wales. Palaeontology, 30, 429-484. LANG, w. D. 1907. The evolution of Stomatopora dichotomoides (d’Orbigny). Geological Magazine (new series, 5), 4, 20-24. MELMORE, s. 1935. Notc On Stomatopora sniithi (Phillips). Proceedings of the Yorkshire Philosophical Society, 1934, 1-2. MOORE, J. E. s. 1903. The Tanganyika problem. Hurst and Blacknett, London, xxiii4-371 pp. PHILLIPS, J. 1829. Illustrations of the geology of Yorkshire. Part 1. The Yorkshire Coast. York, 192 pp. (3rd edition, 1875, London). POHOWSKY, R. A. 1973. A Jurassic cheilostome from England. In larwood, g. p. (ed.) Living and fossil Bryozoa, 447^61. Academic Press, London, 634 pp. 1978. The boring ctenostomate Bryozoa: taxonomy and paleobiology based on cavities in calcareous substrata. Bulletins of American Paleonotology, 73, 1-192. TAVENER-SMiTH, R. 1971. Polyporci stenostonuf. a Carboniferous bryozoan with cheilostomatous features. Palaeontology, 14, 178-187. TAYLOR, p. D. 1978. A Jurassic ctenostome bryozoan from Yorkshire. Proceedings of the Yorkshire Geological SoaV/v, 42, 21 1-216. 1981. Functional morphology and evolutionary significance of differing modes of tentacle eversion in marine bryozoans. In larwood, g. p. and nielsen, c. (eds) Recent and fossil Bryozoa, 235-247. Olsen and Olsen, Fredensborg, 334 pp. 1986. Scanning electron microscopy of uncoated fossils. Palaeontology, 29, 685-690. 1987. Skeletal morphology of malacostegan grade cheilostome Bryozoa. In ross, j. r. p. (ed.) Bryozoa: present and past, 269-276. Western Washington University, Bellingham, 333 pp. 1988. Major radiation of cheilostome bryozoans: triggered by the evolution of an new larval type? Historical Biology, 1, 45-64. 1990. Preservation of soft-bodied and other organisms by bioimmuration. Palaeontology, 33, 1-17. ULRICH, E. o. 1890. Paleozoic Bryozoa. Illinois Geological Survey, 8, 283-688. VINE, G. R. 1892. Notes on the Polyzoa, Stomatopora and Prohoscina groups, from the Cornbrash of Thrapston, Northamptonshire. Proceedings of the Yorkshire Geological Society, 12, 247-258. VOIGT, E. 1966. Die Erhaltung verganglicher Organismen durch Abformung infolge Inkrustation durch sessile Tiere. Neues Jahrhuch fur Geologic imd Paldontologie Abheinlungen, 125, 401-422. 1968. Eine fossile Art von Araclmidium (Bryozoa, Ctenostomata) in der Unteren Kreide Norddeutschland. Ibid., 132, 87-96. 1972. Amathia immurata n.sp., ein durch Bioimmuration erhaltenes ctenostomes Bryozoon aus der Maastrichter Tuffkreide. Paldontologisches Zeitchrift, 46, 87-92. 1977. Araclmidium jurassicum n. sp. (Bryoz. Ctenostomata) aus dem mittleren Dogger von Goslar am Harz. Neues Jahrbuch fiir Geologic und Pcddontologie Abheinlungen, 153, 170-179. 1979. The preservation of slightly or non-calcified fossil Bryozoa (Ctenostomata and Cheilostomata) by bioimmuration. In larwood, g. p. and abbott, m. b. (eds) Advances in bryozoology, 541-564. Academic Press, London, 639 pp. 1980. Araclmidium longicaiida n. sp. (Bryozoa Ctenostomata) aus der Maastrichter Tutfkreide (Ob. Kreide, Maastrichtium). Neues Jahrbuch fiir Geologic und Paldontologie Monatschefte, 1980, 738-746. P. D. TAYLOR Department of Palaeontology British Museum (Natural History) London SW7 5BD Typescript received 15 December 1988 Revised typescript received 13 January 1989 PALAEOPERIDINIUM CRETACEUM: A BRACKISH- WATER PERIDINIINEAN DINOFL AGELLATE FROM THE EARLY CRETACEOUS by I. C. HARDING Abstract. New SEM observations of topotype material of Palaeoperidiniwn cretacewn from the early Cretaceous of Alberta, Canada have allowed a complete description of this poorly known morphotype, including tabulation and plate overlap patterns. The features displayed by P. cretacewn show that, unlike other members of the genus, this morphotype cannot be interpreted as an extrathecal ‘exophragm'. The life-cycle stage represented by P. cretacewn is discussed in relation to the life-cycle of modern dinoflagellates. It is suggested that P. cretacewn may represent a hitherto unknown fossilizable zygotic stage in the dinoflagellate life-cycle. This paper details the unusual morphology displayed by Pctlaeoperidinium cretacewn Pocock ( 1962) as revealed by the scanning electron microscope (SEM). This morphology is interpreted in the light of recent discoveries regarding this genus and our present knowledge of the dinoflagellate life-cycle. The genus Palaeoperidinium was created by Deflandre (1935, p. 227) to accommodate those dinoflagellate fossils which displayed a similar morphology to that of the modern dinoflagellate genus Peridiniwn (now Protoperidiniwn), but whose tabulation was insufficiently known to allow allocation to an existing genus. No type species was designated and the genus became a ‘waiting genus’ for incompletely known morphotypes (Sarjeant 1967, p. 242). Pcdaeoperidinium pyrophoriini, originally described by Ehrenberg (1838) and allocated to the modern genus Peridiniwn, was selected as the type species of the genus when it was formally emended by Sarjeant (1967). Interest in the genus Palaeoperidiniwn has been stimulated by the meticulous study of late Cretaceous P. pyrophorwn (Gocht and Netzel 1976). Utilizing the superior resolution capabilities of the SEM, Gocht and Netzel showed that the prominent sculptural elements of P. pyrophorwn are located on the internal surface of the fossil and not the external surface as in other known fossil peridiniineans. Previous observations had shown that the dinoflagellate fossil record consisted of intrathecal cysts. Furthermore, tabulation patterns are known to occur on the internal surfaces of modern dinocyst wall layers (cf. Peridiniwn Umhatwn Evitt and Wall (1968), which led Gocht and Netzel to explain cyst formation and morphology in P. pyrophorwn in terms of genetic control by the cell protoplast. However, whereas it does seem reasonable to explain the endocystal tabulation of Peridiniwn Umhatwn in terms of genetic control (the possibility of the theca acting as a ‘template’ being precluded), it is unlikely that such a mechanism can be invoked for P. pyrophorwn Evitt (1985 pp. 205-207) reasoned that P. pyrophorwn is unlikely to represent a conventional intrathecal cyst because its single wall layer often contains a two-walled internal body which itself could be interpreted as an intrathecal cyst. He also argued that phenetic influences on the ontogeny of an individual mature theca (represented by pandasutural areas) were unlikely to be found on a body wall independent of the theca. In the case of P. pyrophorwn, the most elegant interpretation of the negative relief patterns on the internal cyst surface is that they were formed by direct contact with the external surface of the parent theca. Thus, P. pyrophorwn was taken to represent a resistant (sporopollenin?) extrathecal structure I Palaeontology, Vol. 33, Part I, 1990, pp. 35-48, 3 pls.| © The Palaeontological Association 36 PALAEONTOLOGY, VOLUME 33 (Evitt 1985, p. 206), subsequently termed an ‘exophragm’ (Evitt el al. 1987). The presumed relationships of exophragm, thecal and cyst wall layers can be seen in text-fig. 1. Evitt et al. (1987) have subsequently examined P. hasilum and two as yet undescribed early Cretaceous species and concluded that they too seem to be exophragms. Modern dinoflagellates are not known to produce fossilizable extrathecal wall layers, but some species do form a gelatinous extrathecal sheath (Evitt 1985, p. 206) indicating a mechanism whereby such a layer might be deposited. Rm ► Tp ► Rm ► Tp IB Tp ► Rm ► Ex Th Pe En TEXT-FIG. 1. Cross-sectional wall layer relationships in a hypothetical dinoflagellate. Ex = 'exophragm', Th = thecal plate. Pm = periphragm, En = endo- phragm. The first three wall layers are formed closely appressed to one another. The following features are represented in various guises on several of the wall layers: Rm = reticular murus, Tp = trichocyst pore, IB = intercalary growth band, Cy = cell cytoplasm. This cross-section is a composite reconstruction, the 'exophragm' (Ex) is based on that of Palaeoperi- dinium pyrophonmi, the thecal plates (Th) on modern Peridinium. the periphragm (Pe) on Subtilisphaera terrula and the endophragm (En) on modern Per- idiniion limbatwn (note parasutural groove on in- ternal surface of endophragm). The plasmalemma, vesicular membranes and pellicle are omitted. N.B., all wall layers shown would not be present at one time. Palaeoperidinium cretaceum was first described by Pocock (1962) from the Quartz Sand Member of the early Cretaceous Lower Mannville Formation (Imperial McMurray Testhole 6, Alberta, Canada). Davey (1970) made P. cretaceum the type species of his new genus Astrocysta after studying Albian specimens from the International Yarbo No. 17 borehole, east of Regina, Saskatchewan. Subsequent authors have treated Astrocysta as a junior synonym of Palaeoperidinium (Lentin and Williams 1976), but Bujak and Davies (1983, p. 134) suggest the retention of Astrocysta to accommodate conventional intrathecal cysts presently allocated to Palaeoperidinium. More recently the type specimen of P. cretaceum has been re-examined by Jansonius (1986). The present study was prompted by the discovery of abundant specimens of P. cretaceum, from 130 ft (39-62 m) in the IGS (now British Geological Survey) Hunstanton Borehole in Norfolk. These specimens were observed during an SEM-based biostratigraphical project on the dinocysts of the western European Barremian (Harding in press). Topotype material of both P. cretaceum and "Astrocysta cretacecP have been made available to enable an SEM study of this morphotype. MATERIAL AND METHODS The topotype material was supplied as a wet residue, prepared following standard palynological techniques, but without oxidation or alkali treatment, in order to keep degradation of this thin-walled taxon to a minimum. Little obscuring amorphous organic matter was present. Other samples mentioned in this paper have been oxidized. For SEM observation, specimens were either strew-mounted or individually picked via micropipette from an aqueous solution on to squares of X-ray film. This minimizes the loss of morphological definition experienced with standard SEM preparation techniques (e.g. Hughes et al. 1979). The specimens were found to adhere well to the moistened emulsion, with no sinking of palynomorphs into the mounting medium. The film squares, most bearing scribed reference grids, were cemented on to aluminium pin-stubs with HARDING; CRETACEOUS PE R I D I N 1 1 N E AN DINOFLAGELL ATE 37 cyanoacrylate adhesive, and sputter-coated with gold. Observations were made using a Philips 50 IB SEM; permanent micrographic records were made on 70 mm Ilford FP4 film. Residues were stained using Saffranin 0, but specimens of P. cretacewn showed no coloration. This negative staining reaction has been observed in many types of thin-walled dinocysts, raising the question as to whether they are of the same composition as other ‘sporopollenin’ dinocysts. Lentin (pers. comm.) suggested that P. cretacewn might have a wall composed of chitinous material («-acetylglucosamine). In order to test this possibility a chitin-specific stain was added to the residue. This stain consisted of iodine, potassium iodide and hydrated calcium chloride in distilled water, a positive test turning chitin red-violet (Lillie 1965, p. 503). The test proved negative. This result is perhaps unsurprising, given that hot hydrochloric acid is used in the processing of a palynological residue and that chitinous arthropod exoskeletal cuticle will dissolve in such a medium. The precise composition of the wall of P. cretacewn therefore remains unclear. The following descriptions are based on the examination of the topotype material and use conventional dinocyst terminology (see Evit 1985), although the prefix ‘para-’ is not used (Norris 1978). Illustrated specimens are held in the Sedgwick Museum, Cambridge, England. PALAEONTOLOGY AND AGE OF THE SAMPLES In Hunstanton borehole sample HUN 130 was unique amongst the European samples in yielding abundant specimens of P. cretacewn. No more than ten specimens were recorded in total from all of the other localities studied: Speeton (Yorks.), Alford (Lines.), Warlingham (Surrey) and Gott (West Germany) (Harding in press). The microfloral assemblages isolated from HUN 130 suggested that this sample represents a much nearer-shore environment than the other samples in which P. cretacewn only rarely occurred. After a light microscope (LM) count of 200 palynomorphs (slide X413/3) the microfloral composition of the sample was found to be as follows (%): saccate gymnosperm pollen, 39; Classopollis, < 1; Eiicoiwniidites, 1; large trilete fern spores, 10; small trilete fern spores, 18; angiosperm pollen (Afropollis), 2; P. cretacewn, 3; other dinocysts, 10; foram. linings, 1; fungal bodies, 2; unidentified ‘simple sacs’, 13. The single species of angiosperm pollen present in this sample represents a new species of the genus Afropollis (Penny 1989). Twenty-six dinocyst taxa were identified in the sample, most of a very low numerical abundance. The limited dinocyst assemblage and the abundance of terrestrial palynomorphs suggest a strong terrestrial input into a marginal area of deposition - probably a brackish-water environment. The sample is of late Barremian age. In contrast, the microfloral assemblage in the topotype sample (LM slide JANSONIUS 640/4) gave the following composition after a count of 200 palynomorphs (%): saccate gymnosperm pollen, 14; large trilete fern spores, 2; small trilete fern spores, 6; angiosperm pollen, < 1; P. cretacewn, 6; other dinocysts, 7; unidentified ‘simple sacs’, 64. This sample has a low-diversity dinocyst flora in which ten taxa have been identified at least to generic level. The dinocysts are dominated by P. cretacewn with ceratioids, some similar to those described by Bint (1986), forming the next largest group. The large percentage of unidentified ‘simple sacs’ precludes making accurate environmental statements. However, the remaining microflora and especially the nature of the dinocyst assemblage (see Bint 1986) suggest a low- salinity (at most brackish) origin for the deposition of this sample. This sample was originally dated as Cretaceous by Pocock (1962), but a more precise estimate of Aptian or Albian age was given by Jansonius (1986). The latter age is favoured here from the presence of tricolpate angiosperm pollen with bimodal luminal sizes and by the ceratioid dinocysts. The sample from the International Yarbo borehole No. 17 (count of 200 palynomorphs from slide CH238/3) yielded a microfloral assemblage as follows (%): saccate gymnosperm pollen, 4; small trilete fern spores, 7; " Astrocysta" cretacea, 6; Ovoidiniutn ostium, 16; other dinocysts, 17; unidentified palynomorphs, 50 (poorly preserved). Ten species of dinocyst have so far been recorded. Identified terrestrial palynomorphs are rare, with a far greater percentage of dinocysts present in the assemblage. This leads to the conclusion that this sample was deposited in an environment in which a more normal marine salinity prevailed than in those discussed above. However, as will be seen later, the dinoflagellate from this sample 38 PALAEONTOLOGY, VOLUME 33 which Davey (1970) described under the new genus Astrocysta, is not the same morphotype as that from the type sample of P. cretaceum. Thus, environmental interpretations derived from the International Yarbo sample are not of significance in assessing the life environment of P. cretaceum. The age of the sample is late Albian (Davey 1970). From the analyses of the first two study samples it appears that P. cretaceum occurs in residues isolated from samples deposited in restricted salinity environments. The distribution of this morphotype in samples from restricted environments is also corroborated by some samples from offshore Canada where this species comprises 100% of the dinocyst flora (Lentin pers. comm.). It is perhaps unwise to suggest that P. cretaceum was a freshwater species. It does seem reasonable, however, to interpret P. cretaceum as being formed by a dinoflagellate which was found in low- salinity marginal environments. The presence of marine dinocysts alongside P. cretaceum may be due to taphonomic transport of P. cretaceum into these areas (i.e. they are allochthonous), or that the morphotype was euryhaline. MORPHOLOGY OF P A LA EO P E R I D I N I U M CRETACEUM The most striking feature of P. cretaceum is its diaphanous nature. The body wall is only ~0 2 /, lacks multiplication of ventral ornament and has ventrolateral and siphonal tubercles that are clavate rather than nodate. The multiplication of ventral ribbing on juveniles of the older form of 104 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 12. Acanthoceras amphibolum Morrow, 1935. TMM 19801, the holotype of Acanthoceras alvaradoense Moreman, 1942, from the basal part of the Eagle Ford Group 6-4 km (4 miles) south of Alvarado, Johnson County. Natural size. Acanthoceras amphiholum prompted Cobban (1987/)) to assign the species to Ciinningtoniceras Collignon, 1937 which has this type of ornament. The present authors now believe that there are separate lineages of Cwmingtoniceras and Acanthoceras, and that amphiholum, although not typical, is best assigned to Acanthoceras. A. amphiholum evolved from A. hellense. They differ most obviously in the much earlier loss of differentiated inner and outer ventrolateral tubercles in A. amphiholum, where a massive horn develops, and the equally early loss of the siphonal tubercles. Ammonites loevianus White, 1877 (p. 201, pi. 19, fig. 1) the original of which is shown as PI. 3, figs. 4 and 5, is an Acanthoceras amphiholum, from Ojo de los Cuervas, New Mexico. The name has never been used since White’s original account and should be suppressed under the plenary powers of the International Commission on Zoological Nomenclature to whom application has been made. Occurrence. Acanthoceras amphiholum zone. Trans-Pecos and central Texas, New Mexico, Kansas, Colorado, Wyoming, South Dakota, Montana. Middle Cenomanian of Japan and Nigeria. TEXT-FIG. 13. Acanthoceras amphiholwn Morrow, 1935. USNM 420220, from the basal part of the Eagle Ford Group, 1-6 km (1 mile) north of Lillian and west of the Lillian-Retta road, Johnson County, Reduced xO-8. Genus calycoceras Hyatt, 1903, p. 589 (= Metacalycoceras Spath, 1926«, p. 83; ICZN rejected name no. 1265) Type species. ICZN Opinion 557, 1959, name no. 1633; Ammonites navicularis Mantell, 1822 (p. 198, pi. 22, fig. 5). Subgenus newboldiceras Thomel, 1972, p. 105. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES Type species. Acanthoceras newholdi Kossmat, 1897, p. 5(112), pi, 1(12), figs. 2, 3; pi. 3(14), fig. 2, by original designation. 106 PALAEONTOLOGY, VOLUME 33 Calycoceras {Newholdiceras) sp. Plate 1 1, figs. 3 and 4 1928 Acanthoceras n. sp. 2 (aff. tiirneri C. A. White); Adkins, p. 246, pi. 30, figs. 3 and 4. 1942 Acanthoceras aff. turneri White, Adkins; Moreman, p. 202. 1959 Acanthoceras n. sp. aff. turneri White, Adkins; Matsumoto, p. 84, text-fig. 37. Material. TMM 34032, from the basal shell bed of the Bluebonnet Member, Bird Creek, 6 4 km (4 miles) east- north-east of Belton, Bell County. Acanthoceras bellense zone. Discussion. What appears to be the only Newholdiceras known from the Gulf Coast region is a fragmentary specimen 110 mm in diameter. There are 12-13 ribs per half whorl, all of them long and extending from the umbilical shoulder, showing weak differentiation into feebly bullate and non-bullate, with blunt inner and clavate outer ventrolateral tubercles, the venter crossed by broad, bar-like ribs with a trace of a siphonal davits. It most closely resembles C. (N.) newholdi (Kossmat, 1897, p. 5(112), pi. 1(12), figs. 2 and 3; pi. 3(14), fig. 2), but has a much lower rib density, 12-13 TEXT-FIG. 14. Acanthoceras amphiholum Morrow, 1935. USNM 420231, from the basal part of the Eagle Ford Group, 1-6 km (1 mile) north of Lillian and west of the Lillian-Retta road, Johnson County. Reduced xO-8. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 107 per half whorl versus 20 in the type material at comparable diameters. The Texas specimen is left in open nomenclature at this time. Occunence. As under material. Genus conlinoceras Cobban and Scott, 1973, p. 61 Type species. Calycoceras (Conlinoceras) gilberti Cobban and Scott, 1973 (p. 61, pi. 1 ; pi. 2, figs. 5-9, 13-18; pi. 3, figs. 5-7, 11; text-figs. 23 and 24). Discussion. Conlinoceras was originally proposed as a subgenus of Calycoceras. The inner whorls of the type species are, however, quite distinct from those of the early subgenera of Calycoceras, rather being identical in style of ornament (notably the clavate ventral and ventrolateral tuberculation) and whorl section with Acanthoceras rhotomagense (e.g. PI. 7, figs. 1-13, 23-25). Some middle growth stages of Conlinoceras tarrantense (Adkins, 1928) were actually placed in the synonymy of A. rhotomagense by Kennedy and Hancock (1970). In contrast, the outer whorls, with their rounded section, loss of tubercles and alternately long and short distant ribs (PI. 5, figs. 1 and 2) are homoeomorphous with those of Calycoceras (Proeucalycoceras) Thomel, 1972. On the evidence of the inner whorls, we regard Conlinoceras as an independent genus allied to Acanthoceras and endemic to the US Western Interior and Gulf Coast Region. Conlinoceras tarrantense (Adkins, 1928) Plate 5, figs. 1-5; Plate 6, figs. 7-12; Plate 7, figs. 1-13, 15, 23-25; text-figs. 5D, 6D, 15-17. 1893 Buchiceras inaequiplicatus Shumard (probably in part); Cragin, p. 233. 71927 Acanthoceras rotomagense (Defrance); Scott, p. 617, pi. 2, figs. 1 and 2. 1927 Acanthoceras aff. rhotomagense (Defrance); Moreman, p. 92, pi. 13, fig. 1. 1928 Metacalycoceras (?) tarrantense Adkins, p. 241, pi. 28, fig. 3; pi. 29, fig. 1. 1928 Acanthoceras wintoni Adkins, p. 243, pi. 25, figs. 2 and 3. 1942 Acanthoceras wintoni Adkins; Moreman, p. 202. 1942 Acanthoceras inaequiplicatum (Adkins); Moreman, p. 201, pi. 32, fig. 2. 1951 Acanthoceras tarrantense (Adkins); Adkins and Lozo, pi. 2, fig. 2. 1953c/ Acanthoceras tarrantense (Adkins); Stephenson, p. 198, pi. 45, figs. 9 and 10; pi. 46, figs. 2-4. 1953/7 Acanthoceras tarrantense nitidum Stephenson, p. 199, pi. 50, figs. 5 and 6. 1953/7 Acanthoceras wintoni Adkins; Stephenson, p. 200, pi. 45, figs. 7 and 8; pi. 46, fig. 1 ; pi. 47, figs. 1 and 2. 1970 Acanthoceras wintoni Adkins; Kennedy and Hancock, p. 487. 1973 Calycoceras (Conlinoceras) tarrantense (Adkins); Cobban and Scott, p. 62. 1973 Acanthoceras adkinsi Stephenson; Cobban and Scott, p. 62. 1977/7 Calycoceras (Conlinoceras) tarrantense (Adkins); Cobban, p. 22, pi. 3, fig. 9; pi. 4, figs. 1-3, 6. 19776 Calycoceras (Conlinoceras) tarrantense (Adkins); Cobban, p. 219, fig. 21. 1978 Conlinoceras tarrantense (Adkins, 1928); Young and Powell, pi. 4, figs. I, 5. Types. Holotype of C. tarrantense is TMM 2424, from the Tarrant Formation ‘2 miles (not more than 1-5 miles) east of Tarrant Station, Tarrant County’. The holotype of C. tarrantense nitidum Stephenson, 1953/7 is USNM 105965, from the Tarrant Formation 14-5 km (9 miles) north of Arlington, Tarrant County. The holotype of Acanthoceras adkinsi Stephenson, 1953/7 is USNM 105968, from the Tarrant Formation, on a branch north of Chicago, Rock Island and Pacific Railroad, 1-6 km (1 mile) west of Dallas County Line, Tarrant County. The holotype of Acanthoceras wintoni Adkins, 1928 is TMM 2426, from the Tarrant Formation on Big Bear Creek, 4-8 km (3 miles) north-east of Tarrant Station, Tarrant County. All Conlinoceras tarrantense zone. Name of species. Under the Rules of the International Commission on Zoological Nomenclature, tarrantense and adkinsi of Adkins are deemed to have been published simultaneously. As first revising authors Cobban and Scott (1973, p. 62) selected the name tarrantense. 108 PALAEONTOLOGY, VOLUME 33 Dimensions D Wb Wh Wb:Wh U USNM 420232 c 13-2 (100) 7-5 (56-8) 6-1 (46-2) 1-23 2-1 (15-9) USNM 420233 c 21-2 (100) 10-7 (50-5) 10-1 (47-6) 1-06 — (— ) USNM 420234 c 25 0 (100) 14-2 (56-8) 12-4 (49-6) 1-15 4-2 (16-8) USNM 420235 c 27-2 (100) 16-8 (61-8) 13-0 (47-8) 1-29 5-3 (19-5) USNM 420236 c 29-5 (100) 16-2 (54-9) 15-6(52-9) 1-04 4-9 (16-6) USNM 420237 c 29-7 (100) 17-0(59-9) 15-0(50-5) 1-13 — (— ) USNM 420238 c 32 0 (100) 16-5 (51-6) 16-5 (51-6) 1-0 5-9 (18-4) USNM 420239 c 40-4(100) 21-0 (50-7) 19-9 (49-3) 1-06 8-9 (22-0) USNM 420240 c 47-2 (100) — (— ) 23-3 (49-4) — 9-6 (20-3) USNM 420241 c 49-8 (100) 260 (52-2) 23-6 (47-4) 1-13 10-7 (21-5) USNM 420242 c 56-1 (100) 23-4(41-7) 24-9 (44-4) 0-93 11-6 (20-7) USNM 420243 c 60-8 (100) 33-2 (54-6) 26-4 (43-4) 1-26 14-2(23-4) USNM 420244 c 63-7 (100) — (— ) 28-5 (44-7) — 14-7 (23-1) USNM 420245 64-7 (100) 30-8 (47-6) 31-4(48-5) 0-98 12-6(19-5) USNM 420246 66-2 (100) 30-9 (46-7) 28-9 (43-7) 1-07 13-4 (20-2) USNM 420247 c 83-6 (100) 42-7 (51-5) 39-8 (47-6) 1-07 19-1 (22-8) USNM 420248 c 101-2 (100) 49-7 (49-1) 45-4 (44-9) 1-09 27-4 (27-1) USNM 420249 125-5 (100) 59-2 (47-2) 56-5 (45-0) 1-05 32-4 (25-8) USNM 420250 146-0 (100) 60-0 (41-1) 62-0 (42-5) 0-97 41-2 (28-2) USNM 420251 157-0 (100) 67-0 (42-7) 66-5 (42-4) 1-0 38-0 (24-2) USNM 420252 205-0 (100) 80-5 (39-3) 90-0 (43-9) 0-89 61-0 (29-8) Material. More than 100 specimens from the Lewisville Member of the Woodbine Formation, in the Dallas-Fort Worth area and elsewhere in north-central Texas in the USNM, USGS, JPC and OUM Collections, not listed separately. Description. The typical shell form and ornament of the middle growth stages are already present from 12 mm diameter and extend to approximately 100 mm diameter. Coiling is evolute, the umbilicus comprising 16-20% of diameter, the figure increasing through ontogeny. The whorl section varies from compressed to depressed (Wb:Wh varies from 1 •23-0-98). The umbilicus is quite shallow, with a rounded umbilical wall on moulds, and a broadly rounded umbilical shoulder. The whorls are quadrate in intercostal section, with the greatest breadth below mid-flank, the sides and venter flattened, the ventrolateral shoulders broadly rounded. The costal section is trapezoidal in compressed forms, polygonal in depressed ones, with the greatest breadth at the umbilical bullae. Robust individuals have as few as 17 ribs per whorl, gracile ones up to 22. Primary and secondary ribs alternate very regularly. Primaries arise at the umbilical seam and strengthen into a weak to strong umbilical bulla which migrates outwards and declines in strength through ontogeny. In gracile specimens it may disappear at an early stage; in robust specimens it persists. Secondary ribs arise low on the flank. All ribs are straight and recti- to feebly prorsiradiate, and develop conical pointed inner ventrolateral tubercles. A low broad rib connects these to strong, long inner ventrolateral clavi, linked across the venter by a broad, low rib to significantly weaker siphonal clavi. This ornament undergoes progressive modification in middle growth. The umbilical bullae decline at a variable rate, and migrate outwards to the inner flank. The ribs broaden and coarsen, the inner ventrolateral tubercles decline and ultimately disappear, the outer ventrolateral coarsening and persisting. The siphonal clavi also decline so that, by the beginning of the adult whorl there are none or only feeble umbilical bullae, strong ventrolateral clavi and none or only a trace of a weak siphonal clavus. The umbilical ratio increases progressively on the outer whorl, and the umbilical wall flattens markedly, so that the ribs appear to arise on the inner flank. They change from straight to feebly convex in many specimens, while the ventral ribbing EXPLANATION OF PLATE 5 Figs. 1-5. Conlinoceras tarran tense (Adkins, 1928). 1-3, USNM 420253, from 2-4 km (L5 miles) east of Euless, south bank of tributary to Big Bear Creek, Tarrant County. Note anaptychus in 2 and 3. 4 and 5, USNM 420240. from USGS locality 22614, 6-4 km (4 miles) south-east of Euless, Tarrant County. Both specimens are from the Tarrant Eormation, Conlinoceras tarrantense zone. Figs. 3-5 are x 1 ; figures 1 and 2 are reduced xO-75. PLATE 5 KENNEDY and COBBAN, Conlinoceras no PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 15. Conlinoceras tanantense (Adkins, 1928). TMM 2426, the holotype of Acanthoceras wituoni Adkins, 1928, from the Tarrant Formation, Big Bear Creek, near Dallas County line, Tarrant County. Natural size. strengthens on the body chamber so that the costal ventral profile changes from concave to tabulate with a strong bar-like rib crossing the venter, into which the ribs are progressively assimilated. The last few ribs of the adult body chamber weaken somewhat and crowd, and the venter becomes rounded. Specimens regarded as macroconchs (USNM 420252) are adult at up to 200 mm diameter. Adult microconchs are 150 mm diameter or less. Suture line with broad, bifid E/L narrower L and broad L/U^. One adult, USNM 420253, preserves part of the jaw apparatus in the body chamber (PI. 5, figs. 1 and 2). The shell is 185 mm in diameter, the lower jaw 40 mm long at the symphysis, with fine concentric growth lines as well as radial striations, most conspicuous on the lateral and outer edges. It is preserved as a blackened film and is unmineralized. Discussion. The abundant material before us is highly variable, demonstrating the holotype of Metacalycocerasl tarrantense Adkins, 1928 (p. 241, pi. 28, fig. 3; pi. 29, fig. 1 ; see Stephenson 1953r/, pi. 45, figs. 9 and 10; pi. 46, fig. 2 for better photographs) to be an incomplete macroconch of a rather average morphology, while the holotype of Acanthoceras wintoni Adkins, 1928 (p. 243, pi. 25, figs. 2 and 3) is merely more compressed and feebly, if as distantly ribbed (text-fig. 15). Acanthoceras adkinsi Stephenson, 1953a (p. 200, pi. 47, figs. 3 and 4) is no more than a juvenile of the gracile form KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES TEXT-FIG. 16. Conlinoceras tarrantense (Adkins, 1928). USNM 105968, the holotype of Acanihoceras adkinsi Stephenson, 1953a, from the Tarrant Formation, branch north of Chicago, Rock Island and Pacific railroad near Dorothy Siding, Tarrant County. Natural size. with dense ribbing (text-fig. 16). Acanihoceras tarrantense nitidum Stephenson, 1953a (p. 199. pi. 50, figs. 5 and 6) is another variant with well-diflferentiated tubercles on the inner whorls which link to the robust individuals shown here as PI. 6, figs. 7-12; Pi. 7, figs. 23-25. Conlinoceras tarrantense and C. gilherti are closely allied. They differ chiefly in that C. gilberti has much higher, stronger ribs. 'Acanihoceras' harciisi Jones, 1938 (p. 117, pi. 6, figs. 2, 3, 8, 9) is a much smaller species that also comes from the Tarrant Formation of north central Texas. The two differ in the much more spinose inner whorls of barcusi which have a rounded rather than flattened venter, persistence of all tubercles to the middle of the adult body chamber and rounded venter at maturity. Text-fig. 17 shows a remarkable pathological specimen of C. tarrantense with rursiradiate ribbing and no tubercles on the outer whorl. This specimen is the basis for the occurrence of Paracalycoceras in Texas cited in the Treatise (Wright 1957). Occurrence. Conlinoceras tarrantense zone, Tarrant Formation of central Texas; Oak Canyon Member and Cubero Tongue of Dakota Sandstone in west-central New Mexico. Conlinoceras sp. Text-fig. 18 Discussion. A species of Conlinoceras is represented by a fragment from the Acanihoceras bellense zone on Bird Creek in Bell County. USNM 420255 corresponds to the coarse ribbed variants of C. tarrantense, with striking differentiation into primary and secondary ribs. 112 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 17. Conlinoceras tarrantense (Adkins, 1928). USNM 420254, from the Tarrant Formation, tributary to Big Bear Creek, 2-4 km (L5 miles) east of Euless, Tarrant County. This pathological specimen is the basis for the Treatise record of Paracalycoceras in Texas. Reduced x 0-9. EXPLANATION OF PLATE 6 Figs. 1-6. Paraconlinoceras leonense (Adkins, 1928). I and 2, USNM 420260; 3 and 4, USNM 420261, both from uses locality 13577, branch south of Belton-Temple road, I 6 km (1 mile) east of Midway Church, Bell County. 5 and 6, the holotype, TMM .34051, from near the Belton-Temple Highway, Bell County. All specimens are from the basal shell bed of the Bluebonnet Member, Acanthoceras bellense zone. Figs. 7-12. Conlinoceras tarrantense (Adkins, 1928). 7 and 8, USNM 420244; 9 and 10, USNM 420241 ; 1 1 and 12, USNM 420248, all from USGS locality 22614, 6-4 km (4 miles) south-east of Euless, Tarrant County; all from the Tarrant Formation, Conlinoceras tarrantense zone. All figures are x 1 . PLATE 6 KENNEDY and COBBAN, Paraconliuoceras, Conliuoceras 114 PALAEONTOLOGY, VOLUME 33 Genus paraconlinoceras nov. Type species. Eucalycoceras leonense Adkins, 1928, p. 240, pi. 28, fig. 1 ; pi. 29, fig. 3. Diagnosis. Moderately small ammonites with narrow, high ribs that cross the venter; acanthoceratine nuclei bearing long and short ribs with umbilical bullae, conical inner ventrolateral and nodate to clavate outer ventrolateral and siphonal tubercles; inner ventrolaterals decline in middle growth, all tubercles except umbilical ones lost on later parts of body chamber where venter rounds; umbilical bullae decline and disappear at adult aperture. Discussion. Paraconlinoceras microconchs are adult at 50^60 mm, macroconchs at 90 mm. Nuclei of P. leonense are identical in style and shape of ornament to slightly older Conlinoceras tarrantense (text-fig. 19) while sutures are identical in style (compare text-figs. 6D and 20F). Conlinoceras tarrantense reach maturity at 1 50 mm or less in microconchs, 200 mm in macroconchs. The acanthoceratine stage persists in Conlinoceras tarrantense to a size where Paraconlinoceras leonense are adult (PI. 9, figs. 26, 27, 29, 30). If adult phragmocones are compared, Conlinoceras tarrantense has passed from a stage of alternately long and short ribs to one in which all ribs are long and variably bullate, have lost their inner ventrolateral and, commonly, siphonal tubercles. Ribs are broad and blunt, rather than sharp and narrow. Adult body chambers are utterly distinct. On the basis of comparable nuclei and stratigraphic occurrence, Paraconlinoceras leonense is regarded as a possible derivative of Conlinoceras, via " Acanthoceras' barcusi Jones, 1938, which is also referred to the new genus. Paraconlinoceras is a homoeomorph of Gentoniceras Thomel, 1972. They differ in the acanthoceratine nuclei of the former which have markedly clavate outer ventrolateral and siphonal tubercles, not seen in Gentoniceras. Occurrence. Middle Cenomanian Conlinoceras tarrantense zone of Texas, Acanthoceras bellense zone of Texas and Wyoming, Conlinoceras gilberti zone of Colorado. Paraconlinoceras barcusi (Jones, 1938) PI. 8, figs. 1-7 1938 Acanthoceras barcusi Jones, p. 117, pi. 6, figs. 2, 3, 8, 9. 1951 Acanthoceras barcusi Jones; Adkins and Lozo, pi. 2, fig. 4. 1953fl Acanthoceras barcusi Jones; Stephenson, p. 203, pi. 44, figs. 9-1 1. Type. Holotype is UMM 16543, the original of Jones (1938, pi. 6, fig. 2) from the ‘basal Eagle Ford’ ofTarrant County, Texas. Material. USNM 420264-420267 without precise locality data; USNM 105977 from Big Bear Creek, Dallas County, all from the Tarrant Formation, middle Cenomanian Conlinoceras tarrantense zone. EXPLANATION OF PLATE 7 Figs. 1-13, 15, 23-25. Conlinoceras tarrantense (Adkins, 1928). 1 and 2, USNM 420237; 3-5, USNM 420234; 6, 10, 13. USNM 420235; 7-9, USNM 420238; 11 and 12, 15, USNM 420233; 23-25, USNM 420243, all from 2 4 km (L5 miles) east of Euless, south bank of tributary to Big Bear Creek, Tarrant County; Tarrant Formation, Conlinoceras tarrantense zone. Figs. 14, 16, 17-22. Paraconlinoceras leonense (Adkins, 1928). 14, 16, USNM 420257; 17 and 18, USNM 420256; 19 and 20, USNM 420262, from USGS Locality 11845; 21 and 22, USNM 420263, all from the basal shell bed of the Bluebonnet Member, Bird Creek, 6 4 (4 miles) east-north-east of Belton, Bell County. Acanthoceras bellense zone. All figures are x 1. PLATE 7 KENNEDY and COBBAN, Conlinoceras, Paraconlinoceras 116 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 18. Conlinoceras sp. USNM 420255, from the basal shell bed of the Bluebonnet Member on Bird Creek, Bell County. Natural size. Dimensions D Wb Wh Wb:Wh U USNM 420264 c 52-0(100) 28-9 (55-6) 21-7 (41-7) 1-33 — {— ) ic 48-5 (100) 22-9 (47-2) 19-0(39-2) 1-21 — (— ) USNM 105977 c 58-7 (100) — (— ) 22-0 (37-5) — 19-7 (33-6) ic 55-5 (100) — (— ) 20-0 (36-0) — — (— ) USNM 420265 c 64-2 (100) 30-4 (47-4) 25-0 (38-9) 1-2 23-2 (36-1) ic 60-2 (100) 26-5 (44-0) 22-3 (37-0) 1-19 — (—) Description. Coiling evolute with U = 33-36% of diameter, quite deep with subvertical wall in early growth becoming rounded at maturity. Umbilical shoulder broadly rounded. Costal whorl section depressed (Wb: Wh ratio 1-2-1 ■33), polygonal, with greatest breadth at umbilical bullae. Intercostal section depressed trapezoidal (Wb:Wh ratio M9-1-21) with broadly rounded inner flanks, flattened convergent outer, broadly rounded ventrolateral shoulders and somewhat flattened venter. There are 12-13 primary ribs per whorl in middle growth. These arise at the umbilical seam and develop into strong, distant coarse ribs with a strong, pointed bulla perched on the umbilical shoulder. These give rise to strong rursiradiate ribs which alternate regularly with secondary ribs that arise low on the flank. These strengthen to match the development of the primaries. EXPLANATION OF PLATE 8 Figs. 1-7. Paraconlinoceras barciisi (Jones, 1938). 1, 5, 6, the holotype, UMM 16543, from the ‘Basal Eagle Ford’ (e.g. Tarrant Formation) of Tarrant County. 2, 7, USNM 105977, from 2-4 km (L5 miles) east of Euless, on south bank of Big Bear Creek, Tarrant County. 3 and 4, USNM 420265, without precise locality data but from the Tarrant Eormation of Tarrant County. All Conlinoceras tarranlense zone. All figures are x I. PLATE 8 KENNEDY and COBBAN, ParacouUnoceras PALAEONTOLOGY, VOLUME 33 and all ribs bear a sharp inner ventrolateral tubercle linked by a strong rib to clavate outer ventrolateral and siphonal tubercles. This ornament persists to the beginning of the adult body chamber, where there is a progressive outward migration and decline in strength of the umbilical bullae, weakening of the ventrolateral and siphonal tubercles. Specimens that we take to be adult microconchs are 59-63 mm in diameter, and show marked egression of the umbilical seam with the last few ribs before the aperture weakened and crowded, with no or very weak tubercles and an evenly rounded costal whorl section. The holotype and USNM 420265 are over 80 mm in diameter and appear to be incomplete macroconchs. Suture (Jones 1938, pi. 6, fig. 3) with broad, bifid E/L and narrow L. Occurrence. Conlinocercis tarrantense zone. Tarrant Formation of north central Texas only. Paraconluwceras leonense (Adkins, 1928) Plate 6, figs. 1-6; Plate 7, figs. 14, 16-22; Plate 9, figs. 26, 27, 29, 30; text-figs. 19A-H. 1928 Eucalycoceras leonense Adkins, p. 240, pi. 28, fig. 1 ; pi. 29, fig. 3. 1928 Metacalycoceras ? sp. 2; Adkins, p. 242, pi. 28, fig. 2; pi, 29, fig. 2. 1942 Eucalycoceras leonense Adkins; Moreman, p. 207. 1969 Eucalycoceras (Proeucalycoceras) leonense Adkins, Thomel, p. 650. 1972 Eucalycoceras {Proeucalycoceras) leonense Adkins; Thomel, p. 81. 1973 Calycoceras leonense (Adkins); Cobban and Scott, p. 60, pi. 3, figs. 1-4. 19876 Calycoceras (Gentoniceras) leonense (Adkins); Cobban, p. 4, pi. 1, figs. 1-17, text-fig. 2. Holotype. TMM 34051, the original of Adkins (1928, p. 240, pi. 28, fig. 1 ; pi. 29, fig. 1), from the basal shell bed of the Bluebonnet Member of the Lake Waco Formation of the Eagle Ford Group near the old Belton- Temple Highway, Bell County, Acanthoceras bellense zone. Material. More than 100 specimens and fragments from the same horizon as the holotype: USGS localities 1 1845, 13577 and 19554, J. P. Conlin and OUM collections, on Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County. Middle Cenomanian Acanthoceras bellense zone. Dimensions D Wb Wh Wb;Wh U USNM 420256 c 49-7 (100) 21-5 (43-3) 180(36-2) 1-2 17-5 (35-2) (microconch) USNM 420257 c 53-0 (100) 22-5 (42-4) 20-8 (39-2) 1-08 15-5 (29-2) (juvenile) USNM 420258 c 61-0(100) 27-5 (45-1) 26-5 (43-4) 1-04 20-5 (33-6) (microconch) USNM 420259 c 90-0 (100) — (~) 30-5 (-) 32-8 (36-4) (macroconch) c 80-5 (100) 31-2 (38-8) 28-0 (34-8) 1-14 27-2 (33-8) EXPLANATION OF PLATE 9 Figs. 1-25, 28. Plesiacanthoceratoides vetula (Cobban, 19876). 1-4, USNM 420291; 5-8, USNM 420292; 10-12, USNM 420290; 13-16, USNM 420287; 17-20, USNM 420289, all from a loose concretion at USGS Mesozoic locality D12626, basal part of the Eagle Ford Group, 8-9 km (5-5 miles) north-east of Mansfield, Tarrant County. 21 and 22. paratype USNM 388194; 23-25, paratype USNM 388195; 28, paratype USNM 388197, all from USGS Mesozoic locality D5900, Belle Fourche Shale, Old Woman anticline (south-west of the Black Hills, head of Elm Creek, in W| sec. 14, T.36N, R.62W, Niobrara County, Wyoming. All Acanthoceras aniphiboluin zone. Figs. 26 and 27, 29 and 30. Paraconluwceras leonense (Adkins, 1928). 26 and 27, USNM 420258; 29 and 30, USNM 420259, from USGS locality 13577, basal shell bed of Bluebonnet Member, Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County. Figs. 31 and 32. Acanthoceras bellense Adkins, 1928. The holotype, TMM 3034, from the same horizon and locality as the originals of figs. 26 and 27. Figures 21-25, 28 are x2; the remainder are x 1. PLATE 9 KENNEDY and COBBAN, Plesiacanthoceratoides, Paraconlinoceras, Acanthoceras 120 PALAEONTOLOGY, VOLUME 33 G H TEXT-FIG. 19. Paraconlmoceras leonense (Adkins, 1928). a and B, USNM 388089; C-E, USNM 388087; F, external suture of USNM 388087; G and h, USNM 388088. All are from the Belle Fourche Shale at USGS Mesozoic locality D884I, west of Osage in the SWl/4 sec. 8, T. 46 N., R. 63 W., Weston County, Wyoming. A-E. G and H are natural size. Description. The species appears to be markedly dimorphic; 2 complete microconchs are 50 and 61 mm diameter, two macroconchs 90 and 92 mm diameter. The early growth stages are not shown by the Texas material. In middle growth as far as the beginning of the adult body chamber, the coiling is evolute, with U = 27-33 % of diameter, shallow, with a rounded wall. Intercostal whorl section depressed, reniform, the greatest breadth well below mid-flank. Costal section trapezoidal-polygonal. There are generally 12-14 primary ribs per whorl. They arise at the umbilical seam, strengthen into sharp bullae on the umbilical shoulder and give rise to sharp, distant, prorsiradiate primary ribs. At the smallest diameters visible these bear small inner ventrolateral tubercles that are no wider than the rib, linked by a strong rib to a clavate outer ventrolateral and a somewhat weaker siphonal tubercle. At small sizes ribs may arise in pairs from the umbilical bullae, but in general the primaries alternate regularly with secondaries inserted at or below mid- flank. The latter strengthen to equal the primaries at the ventrolateral shoulder, where they develop a full complement of tubercles. The inner ventrolateral tubercles decline around the outer whorl of the phragmocone, and the beginning of the body chamber bears outer ventrolateral and siphonal tubercles only. The umbilical seam of the mature body chamber egresses markedly, to give U = up to 37%. Umbilical bullae migrate out to the inner flank, before declining and disappearing just before the adult aperture. The outer ventrolateral and siphonals persist to the last quarter whorl before the adult aperture, thereafter they decline. The last few ribs before the aperture are weakened, and extend to the umbilical shoulder without bullae. Suture with broad, symmetrically bifid E/L, narrow bifid L and broad L/U,. Discussion. The Texas material does not show the inner whorls. Study of well-preserved material from the Belle Fourche Shale at USGS Mesozoic locality D8841 in Weston County, Wyoming (text- fig. 19) shows juveniles to have strikingly Acanthoceras-Wko, nuclei, with a polygonal whorl section in costal section and markedly clavate outer ventrolateral clavi. This is quite different from inner whorls of Gentoniceras Thomel, 1972, which the species resembles at maturity. We conclude that P. leonense is a homoeomorph of Gentoniceras, its inner whorls pointing to an origin in slightly older Paraconlinoceras harcusi which has comparable inner whorls (PI. 8, figs. 1-7) and outer whorls with high, narrow ribs. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 121 Occurrence. Middle Cenomanian Acanthocercis bellense zone of central Texas and eastern Wyoming. Conlinoceras gilberti zone of southeastern Colorado. Genus cunningtoniceras Collignon, 1937, p. 64 (40) (? = Giierangericeras Thomel, 1972) Type species. Ammonites cunningtoni Sharpe, 1855 (p. 35, pi. 15, fig. 2). Discussion. See Wright and Kennedy (1987, p. 193). Cunningtoniceras lonsdalei (Adkins, 1928) Plate 12, figs. 1-3, 8; text-figs. 20-22 1928 Acanthoceras lonsdalei Adkins, p. 244, pi. 26, fig. 5; pi. 27, fig. 3. 1942 Acanthoceras lonsdalei Adkins; Moreman, p. 204. 1955 Euomphaloceras lonsdalei (Adkins); Stephenson, p. 62 (pars), pi. 6. figs. 6-8, non 9-20. 1963 Euomphaloceras lonsdalei (Adkins); Wright, p. 609, pi. 87, fig. 2; pi. 88, fig. 1 ; pi. 89, fig. 2. 71973 Euomphaloceras cf. lonsdalei (Adkins); Cobban and Scott, p. 71, pi. 5, figs. 1, 2, 4. 1978 Euomphaloceras lonsdalei (Adkins); Young and Powell, pi. 5, fig. 1 only (non 7, = Acanthoceras bellense). 1987 Cunningtoniceras lonsdalei (Adkins); Wright and Kennedy, p. 194, text-fig. 80. Holotype. TMM 2410, the original of Adkins (1928, p. 244, pi. 26, fig. 5, pi. 27, fig. 3) by original designation. From the Bluebonnet Member of the Lake Waco Formation of the Eagle Ford Group on the Belton-Temple road. Bell County. Middle Cenomanian Acanthoceras bellense zone. Material. USNM 108831a-b (originals of Stephenson 1955, pi. 6, figs. 6-8), USNM 420268^20272, TMM 1069, WSA 12830, all from the same horizon as the holotype on Bird Creek, 6 4 km (4 miles) east-north-east of Belton, Bell County. Dimensions USNM 108831b c USNM 420269 c TMM 2410 c D Wb 5F5 (100) 32-1 52-3 (100) 28-9 79-2 (100) 43-5 410 Wh (62-3) 24-0 (46-6) (55-3) 24-0 (45-9) (54-9) 35-5 (44-8) (— ) 36-1 (— ) Wb:Wh U F34 12-3 (23-9) 1-20 12-9 (24-7) F23 22-0 (27-7) 114 Description. Up to 100 mm: coiling very evolute, umbilicus comprises up to 28% of diameter. Whorl section depressed, rounded-trapezoidal in intercostal section (Wb:Wh 114), with greatest breadth outside umbilical shoulder. Inner flanks rounded, outer flanks flattened, convergent, ventrolateral shoulders broadly rounded, venter somewhat flattened. Costal section very depressed, polygonal, with greatest breadth at umbilical bullae. Distant primary ribs, 10 per half whorl, arise at the umbilical seam. They strengthen across the umbilical wall and develop into strong umbilical bullae perched just outside the umbilical shoulder. These give rise to broad, straight, prorsiradiate ribs, somewhat effaced on the outer flank, where they connect to a strong conical inner ventrolateral horn, A low broad rib connects to a clavate inner ventrolateral tubercle, accompanied by one or two weaker ribs connecting to generally weaker inner ventrolaterals, all ribs extending to evenly developed siphonal tubercles. There are occasional intercalated ribs with outer ventrolateral and siphonal tubercles so that the 10 or so primary ribs per half whorl correspond to 20 or more ventral ribs. An adult specimen, USNM 420271 (text-fig. 20), has inner whorls identical to those of the holotype and is mature at an estimated intercostal diameter of just over 200 mm. On the early body chamber there are distant ribs with flared umbilical bullae and a strong conical ventrolateral horn without a trace of an outer ventrolateral clavus, the horns linked across the venter by a pair of looped riblets on a broad low rib, the costal section being markedly concave. The ventral rib strengthens markedly at the end of the body chamber and is strong and bar-like, the costal profile loosing the deep ventral concavity. Discussion. There is some variation in the ventral ribbing, as in other Cunningtoniceras species (Wright and Kennedy 1987). Cunningtoniceras lonsdalei was derived from Acanthoceras bellense by stabilization of multiple ventral ribbing; it is linked to A. bellense by a common style of flank ornament, whorl proportions, the presence of a marked siphonal ridge and distinctive bar-like rib 122 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 20. Cwmingtoniceras lomdalei (Adkins, 1928). USNM 420271, an adult specimen from the basal shell bed of the Bluebonnet Member on Bird Creek, Bell County. Reduced x 0-5. at the adult aperture. Zaborski (1985) regarded C. lonsdalei as a subspecies of C. cuwiingtoni, but the latter is a derivative of Acanthoceras rhotomageuse. The two species differ in that C. cuimiugtom is much more depressed, has far fewer ribs, massive inner ventrolateral horns, obvious looping of ventral ribs and utterly different adult ornament. Stephenson (1955) confused C. lonsdalei and C. jolmsonanum (Stephenson, 1955); differences are outlined below; features of other Cimningtoniceras species are reviewed by Wright and Kennedy (1987) and are not repeated here. Occurrence. Acanthoceras bellense zone, central Texas. Middle Cenomanian of Bathurst Island, northern Australia. There is a doubtful record from the Acanthoceras muldoonense zone of SE Colorado. EXPLANATION OF PLATE 10 Figs. 1 17. Cimningtoniceras johnsonanwn (Stephenson, 1955). 1-3, USNM 420275, from east of old Alvarado-Grandview highway, 6-4 km (4 miles) south of Alvarado, Johnson County; 4, 8, USNM 420276; 5-7, USNM 420273; 15-17, USNM 420278, from concretions in cotton field 1-6 km (1 mile) north of Lillian, Johnson County; 9-1 1, USNM 420274; 12-14, USNM 420277, from 0-3 km (0-2 mile) south of Mountain Creek, 61 km (3-8 miles) north-east of the town square in Alvarado, Johnson County. All specimens from the basal part of the Eagle Ford Group, Acanthoceras ainphiholian zone. All figures are x I. PLATE 10 pi ^ f y ,y \ , ]p, i -W r. .., nI 1 .) ■' ' KENNEDY and COBBAN, Cimniiigtoniceras 124 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 21. Cwmingtonkeras lonsdalei (Adkins, 1928). The holotype, TMM 2410, from the basal shell bed of the Bluebonnet Member on the Belton-Temple road. Bell County. Natural size. Cunningtoniceras johnsonanum (Stephenson, 1955) Plate 10, figs. 1-17; Plate 11, figs. 1 and 2. 1955 Acanthoceras jolmsonanum Stephenson, p. 58, pi. 4, figs. 14-17. 1955 Eiiomphaloceras lonsdalei (Adkins); Stephenson, p. 62 {pars), pi. 6, figs. 9-20 only. Type. The holotype is USNM 108846, from USGS locality 14583, north facing slope of Mountain Creek Valley, 4 km (2-5 miles) north-north-east of Alvarado, Johnson County. Basal part of the Eagle Ford Group, Acanthoceras amphiholwn zone. Material. Numerous specimens from the same horizon as the holotype at the following localities: gully in field L6km (I mile) north of Lillian, west of Lillian-Retta road; scattered concretions east of the old Alvarado-Grandview highway, 6-4 km (4 miles) south of Alvarado; hillside 0-3 km (0-2 mile) south of Mountain Creek and west of secondary road 6T km (3 8 miles) airline north-east of the town square in Alvarado, all in Johnson County and all e.x Conlin Collection. USGS Mesozoic locality D12626, 8-9 km (5-5 miles) north-east of Mansfield, Tarrant County. All of the above are from concretions in the basal part of the Eagle Ford Group, Acanthoceras amphibolum zone. EXPLANATION OF PLATE II Figs. 1 and 2. Cunningtoniceras johnsonanum (Stephenson, 1955). The holotype, USNM 108846, from USGS locality 14583, north-facing slope on Mountain Creek Valley, 4 km (2-5 miles) north-north-east of Alvarado, Johnson County, Acanthoceras amphibolum zone. Figs. 3 and 4. Calycoceras (Newboldiceras) sp. TMM 34032, from the basal shell bed of the Bluebonnet Member, Belton-Temple Highway, Bell County. Acanthoceras bellense zone. All figures are x I . PLATE 1 1 ^ -r ,- '5i‘'^ /• A/: -^’ '*' miiii 'lA'Wf^ ', KENNEDY and COBBAN, Cuiiningtoiuceras, Calycoceras 126 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 22. Cwiningtoniceras lonsdulei (Adkins, 1928). TMM W. S. Adkins Collection 12830, from the basal shell bed of the Bluebonnet Member on Bird Creek, Bell County. Reduced x0 95. EXPLANATION OF PLATE 12 Figs. 1-3, 8. Cwiningtoniceras lonsdalei (Adkins, 1928). 1 and 2, USNM 420268; 3, 8, USNM 420269, from the basal shell bed of the Bluebonnet Member, Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County. Acanthoceras bellense zone. Figs. 4, 7. Acanthoceras bellense Adkins. 1928. USNM 420215, horizon and locality as for the originals of figs. 1-3, 8. Figs. 5, 6, II. Turrilites (Twrilites) dearingi Stephenson, 19530. 5, USNM 420314; 6, USNM 420315, casts of specimens in the Gerry Kienzien Collection (Dallas, Texas), from roadcut on east side of Texas Flighway 360, F9 km (F2 miles) south of bridge over Trinity River. 1 1, paratype USNM 105957, from a branch of Big Bear Creek, 2-4 km (F5 miles) east of Euless, both in Tarrant County, Tarrant Formation, Conlinoceras tarrantense zone. Fig. 9. Plesiacanthoceras bellsanwn (Stephenson, 1953o). USNM 105984, from the Templeton Member of the Woodbine Formation 4-3 km (2-7 miles) north of Bells, Grayson County. Fig. 10. Twrilites (Twrilites) aciitiis Passy, 1832. USNM 420301, from the same horizon and locality as the originals of figs. 1-3, 8. Figs. 12-14. Sciponocerasl sp. TMM 2425, from the base of the Bluebonnet Member, 0-8 km (0-5 mile) south- east of Round Rock, Williamson County. Acanthoceras bellense zone. Figure 9 is x2; the remainder x I. PLATE 12 KENNEDY and COBBAN, Cenomanian ammonites 128 PALAEONTOLOGY, VOLUME 33 Dimensions D Wb Wh Wb:Wh U USNM 420273 c 26-8 (100) 15 0 (56 0) 12-2 (45-5) 1-23 6-5 (24-3) USNM 420274 c 310 (100) 16-5 (53-2) 13-2 (42-6) 1-25 8-0 (25-8) USNM 420275 c 50-3 (100) 28-5 (56-7) 20-6 (40-9) 1-38 13-2 (26-2) ic 46-5 (100) 24-5 (52-7) 19-3 (41-5) 1-27 Description. Coiling is moderately evolute with approximately 30% of the previous whorl covered. Umbilicus of moderate depth, comprising 24-26% of diameter, with subvertical wall. Whorls vary from very depressed (intercostal Wb:Wh ratio 1-27) to equidimensional. The intercostal whorl section is trapezoidal, the costal section polygonal, with the greatest breadth at the umbilical bulla. There are generally 12-14 primary ribs per whorl. They arise at the umbilical seam, strengthen across the umbilical wall and shoulder and develop weak to strong umbilical bullae, strength of ornament varying from weak to coarse between individuals. In robust individuals (USNM 420270, 420272, 420274) straight prorsiradiate ribs broaden and weaken somewhat as they pass across the flanks only to strengthen into a pronounced inner ventrolateral horn. A broad rib leads to strong outer ventrolateral and siphonal clavi. At small diameters, shorter ribs intercalate low on the flank and bear a full complement of tubercles, with the inner ventrolateral weaker than on the primary ribs. With increasing diameter this tubercle effaces to leave short ventral ribs with inner ventrolateral and siphonal clavi only, the short ribs sometimes linking to the inner ventrolateral horns of the primary ribs, there being twice as many ventral as flank ribs. In gracile individuals (USNM 420279), there may be up to 2 intercalated ribs with outer ventrolateral and siphonal clavi between primaries, so that there are 3 times as many ventral as flank ribs. Most specimens are less than 70 mm in diameter, and closely resemble the holotype (PI. 1 1, figs. 1 and 2). This appears to be a microconch, showing effacement of tuberculation and rounding of venter not seen in our material. The latter includes larger fragments in which inner and outer ventrolateral tubercles are differentiated to whorl heights of up to 35 mm, with multiple ventral ribs and tubercles (USNM 420274). We presume these differences to be within the limits of intraspecific variation. Suture not seen. Discussion. Stephenson (1955 p. 62) confused this species with the older Cimningtoniceras lonsdalei. Adkin’s species has a rounded, rather than polygonal section, more intercalated ribs, with tubercles often much weaker than on the primaries. C . johnsonanum may well be descended from C. lonsdalei. Co-occurring Acanthoceras amphiholum amphibolum variants may be superficially similar to C. johnsonanum. They have constricted innermost whorls, however, are compressed with distant ribs and show early loss of inner ventrolateral clavi leaving a single ventrolateral horn, with a siphonal ridge and weak siphonal clavi (PI. 4, figs. 1-14). Adults, with striking umbilical bullae and ventrolateral horns as the only strong tuberculation are immediately distinct (text-figs. 13 and 14). Occurrence. Acanthoceras amphiholum zone of central Texas only. Cimningtoniceras inerme (Pervinquiere, 1907) 1855 Ammonites susse.xiensis Mantell; Sharpe, p. 34, pi. 15, fig. 1. 1907 Acanthoceras cunningtoni var. inermis Pervinquiere, p. 277. 1953fl Acanthocerasl euiessanum Stephenson, p. 201, pi. 47, fig. 5; pi. 48, figs. 3 and 4. 1987 Cimningtoniceras inerme (Pervinquiere, 1907); Wright and Kennedy, p. 194, pi. 52, fig. 1 ; pi. 53, fig. 6; text-figs. 74 and 75, 79 (with full synonymy). EXPLANATION OF PLATE 13 Figs. 1-12. Tarrantoceras sellardsi (Adkins, 1928). 1-3, USNM 400759; 4 and 5, USNM 400769; 10 and 11, USNM 400770, from concretions in cotton field 1-6 km (1 mile) north of Lillian, west of the Lillian-Retta Road. Johnson County. 6-8, USNM 400760; 9, 12, USNM 420284, from USGS Mesozoic locality D12626, 8-9 km (5-5 miles) north-east of Mansfield, Tarrant County. All specimens are from the basal part of the Eagle Ford Group, Acanthoceras amphibolum zone. All figures are x 1. PLATE 13 KENNEDY and COBBAN, Tarrantoceras 130 PALAEONTOLOGY, VOLUME 33 Discitssioti. Wright (1963) and Kennedy (1971) both regarded Acanthoceras ? eulessanum Stephenson, 1953a as a synonym of Cwmingtoniceras cimningtoni. Wright and Kennedy (1987, p. 204) pointed out that it had strong flank ribs and compared it to C. inerme. It differs from specimens of the latter species (from the English Chalk) only in its lower rib density (13-14 per whorl vs. 17-20), and it is in this respect transitional to C. cimningtoni. We regard it as no more than a variant of C. inerme. Occurrence. Conlinoceras tarrcmtense zone, Tarrant Formation of north central Texas. Where precisely dated in western Europe, it occurs at the top of the lower, Turrilites costatus subzone of the Acanthoceras rhotomagense zone in England and the German Federal Republic. It also occurs in the middle Cenomanian of France and Japan. Genus tarrantoceras Stephenson, 1955 Type species. Tarrantoceras rotatile Stephenson, 1955 (p. 59, pi. 5, figs. 1-10) by original designation (= Mantelliceras sellardsi Adkins, 1928, p. 239, pi. 25, fig. 1; pi. 26, fig. 1). Diagnosis. Small, macroconchs adult at 90-100 mm, microconchs adult at 60 mm or less. Evolute, compressed, early whorls with umbilical bullae, inner and outer ventrolateral and siphonal clavi on flexuous primary ribs separated by several secondaries; all but umbilical bullae decline or disappear at maturity. Suture simple, with broad bifid E/L and shallow bifid L. Discussion. Tarrantoceras is a homoeomorph of certain Eucalycoceras species. The types of the two genera are distinct enough, while species such as Eucalycoceras rowei (Spath, 19266) (see Kennedy 1971, p. 83, pi. 49, figs. 2-7; pi. 50, figs. 3-7) can be distinguished by the greater complexity of the suture with a long, narrow L, and umbilical bullae that project into the umbilicus. The inner whorls of Tarrantoceras have far more pronounced ventral clavi. Cooper (1978) believed Sumitomoceras to be a synonym of Tarrantoceras ', Wright and Kennedy (1981) and Kennedy (1988) regarded Sumitomoceras as a subgenus of Tarrantoceras. Subsequent work supports separation ; the very early loss of siphonal tubercles in Sumitomoceras, presence of constrictions and suture with deep L are distinctive ; the last feature suggests Sumitomoceras is allied to the Old World acanthoceratines, not those of the New World. Occurrence. Middle and low upper Cenomanian of the US Western Interior and Angola. Tarrantoceras sellardsi (Adkins, 1928) Plate 13, figs. 1-12; Plate 14, figs. 1-16, 19, 20, 25, 29, 30; text-figs. 6A, 23A and B. 1928 Mantelliceras sellardsi Adkins, p. 239, pi. 25, fig. 1; pi. 26, fig. 4. 1942 Mantelliceras sellardsi Adkins; Moreman, p. 207. 1955 Tarrantoceras rotatile Stephenson, p. 59, pi. 5, figs. 1-10. 1955 Tarrantoceras stantoni Stephenson, p. 60, pi. 5, figs. 11-21. 1955 Tarrantoceras litlianense Stephenson, p. 60, pi. 5, figs. 22-27. 1971 Eucalycoceras sellardsi (Adkins); Kennedy, p. 84. 1973 Tarrantoceras rotatile Stephenson; Cobban and Scott, p. 64, pi. 10, figs. 1-11; text-fig. 25. 1977a Tarantoceras rotatile Stephenson; Cobban, p. 23, pi. 6, figs. 8-10, 28 and 29; pi. 11, figs. 7 and 8, 11-16; pi. 12, figs. 13 and 14; text-fig. 4. 19776 Tarrantoceras rotatile Stephenson; Cobban, p. 219, figs. 3N and O, 4G. 1978 Utaturiceras ? sellardsi (Adkins); Young and Powell, p. xxv, 18. 1978 Tarrantoceras rotatile Stephenson; Cooper, p. 92, text-fig. 20. 1984 Tarrantoceras sellardsi (Adkins); Cobban, p. 78. 1986 Tarrantoceras sellardsi (Adkins); Cobban, p. 78, figs. 3C and D. Types. The holotype is TMM 34048, the original of Adkins (1928, pi. 25, fig. 1; pi. 26, fig. 1) from the Bluebonnet Member of the Lake Waco Formation of the Eagle Ford Group, 2-4 km (F5 miles) south-east of KENNEDY AND COBBAN:TEXAN CENOMANIAN AMMONITES 131 E TEXT-FIG. 23. External sutures, a and b, Tarrantoceras sellardsi (Adkins, 1928), USNM 400772, 400760. c, Plesiacanthoceras bellsamim (Stephenson, 1953a), USNM 105983. Round Rock, Williamson County, by original designation. The holotype of T. rotatile is USNM 1 1740, the original of Stephenson (1955, pi. 5, figs. 1^) from Walnut Creek, 7-6 km (4-75 miles) east-north-east of Mansfield, Tarrant County. The holotype of T. stantoni is the original of Stephenson (1955, pi. 5. figs. 11-13); the holotype of T. liUianense is USNM 108841, the original of Stephenson (1955, pi. 5, figs. 22 and 23), both specimens from gully in field E6 km (1 mile) north of Lillian, Johnson County, both basal Eagle Ford Group. All of the above are from the Acanthoceras amphibolum zone. Material. Numerous specimens from USGS Mesozoic locality 12626, roadside 8 9 km (5-5 miles) north-east of Mansfield, Tarrant County, in a loose concretion; from USGS locality 24510 from a concretion in a crop field 1-6 km (1 mile) north of Lillian, Burleston, Johnson County; from a concretion on hillside 0-3 km (0-2 miles) south of Mountain Creek, 61 km (3-8 miles) north-east of town square in Alvarado; from USGS locality 14591, 2-3 m (7-5 feet) above base of Cloice Member of the Lake Waco Formation, abandoned brickpit on Cloice Branch, 1 -3 km (0-8 mile) east of South Bosque, McLennan County. Also numerous specimens as OUM KT40008-9, concretions from Walnut Creek, 7-6 km (4-75 miles) east of Mansfield, Tarrant County; from USGS Mesozoic locality D9502, concretions in field just east of gravel road, 0-5 km (0-3 mile) north-north-west of Lillian; from USGS Mesozoic locality D96 ‘Eagle Ford Shale’ 61 m (20 feet) above base, concretion 6-7 km (41 5 miles) northeast of centre of Alvarado, Johnson County. All lower part of Eagle Ford Group, Acanthoceras amphibolum zone. Dimensions D Wb Wh Wb:Wh U USNM 420280 c 23 0 (100) 90(391) 90 (39 1) 10 7-6 (33-0) USNM 420281 c 23 0 (100) 9-5 (41-3) 8-4 (36-5) 113 8-7 (37-8) USNM 420282 c 29 0 (100) 11-3 (38-9) 11 1 (39-0) 102 11 1 (38-3) USNM 400759 c 51-7 (100) 17-8 (34-4) 18-0 (34-8) 0-98 18-2 (35-2) USNM 400760 c 710 (100) 25-0 (35-2) 26-7 (37-6) 0-94 23-3 (32-8) USNM 400770 c 89 0 (100) 26-5 (29-8) 32-0 (36 0) 0-83 32-9 (37-0) Description. The species is markedly dimorphic. A near-complete microconch is 57-5 mm in diameter (PI. 1 3, figs. 4 and 5), complete macroconchs are 90 mm in diameter (PI. 13, figs. 10 and 1 1), and a fragment (PI. 13, figs. 9, 12) with a whorl height of 33-5 mm suggests a macroconch diameter of nearly 100 mm. 132 PALAEONTOLOGY, VOLUME 33 Juveniles are variable. Coiling is very evolute, with U = 33-38%, the umbilicus shallow, with a flattened umbilical wall and broadly rounded shoulder. The intercostal whorl section is compressed trapezoidal, with flattened convergent flanks, broadly rounded ventrolateral shoulders and flattened venter. The costal section varies from slightly compressed to slightly depressed, with greatest breadth at the umbilical bullae, and polygonal. There are generally 10-13 strong umbilical bullae per whorl, perched on the umbilical shoulder, and connected to the umbilical seam by a low broad rib. Additional non-bullate ribs extend to the umbilical shoulder. The ribs are strong, straight, narrow, prorsiradiate and wider than the interspaces. Single intercalated ribs arise around the middle of the flank and strengthen to match the primaries by the ventrolateral shoulder, where all bear a small, sharp inner ventrolateral tubercle that is the same width as the rib. A somewhat broadened rib extends forwards across the ventrolateral shoulder to spinose to feebly clavate inner ventrolateral clavi ; a low broad transverse rib extends across the venter and bears a siphonal clavus, weaker than the outer ventrolateral. As size increases the whorls generally become more compressed and denser- ribbed, with coarsely ornamented variants with persistent tubercles having as few as 30 ribs and weakly ornamented variants having as many as 48 ribs per whorl, the inner ventrolateral tubercles weakening markedly. On the adult body chamber of both macro- and microconchs the umbilical seam egresses and the coiling becomes progressively more evolute. The ribs flex back, becoming coarser and rectiradiate, sometimes bunching at bullae. The inner ventrolateral tubercles efface at the beginning of the body chamber, the umbilical bullae efface towards the aperture. The siphonal tubercle declines to give a flattened venter while at the adult aperture ribbing weakens and the venter rounds. Suture simple, with broad, asymmetrically bifid E/L, narrow, little-incised L and broad L/Uj (text-figs. 6A, 23A and B). Discussion. Juveniles are highly variable, from hypernodose to feebly tuberculate (PI. 14, figs. 1-1 1, 25, 29). This variation persists into middle growth, with the holotype of T. stantoni a coarsely ribbed and tuberculate individual, the holotypes of T. sellardsi and T. rotatile with subdued tuberculation and the holotype of T. lillianense intermediate between the two. Tarrantoceras multicostatum Stephenson, 1955 (p. 61, pi. 6, figs. 21-23) comes from a different locality than the types of T. sellardsi (and its synonyms), and is kept separate here, with no great confidence. It has an estimated 60 ribs per whorl, a rounded whorl section and early loss of tubercles. An as yet undescribed Tarrantoceras species is present in the Calycoceras canitaurinum zone in Trans-Pecos Texas, EXPLANATION OF PLATE 14 Figs. 1-16, 19, 20, 25, 29, 30. Tarrantoceras sellardsi (Adkins, 1928). 1-3, USNM 400762; 4-6, USNM 400767; 10 and 1 1, USNM 400766, from USGS Mesozoic locality D12626, 8-9 km (5-5 miles) north-east of Mansfield, Tarrant County. 7 and 8, USNM 400763, from hillside 0 3 km (0-2 mile) south of Mountain Creek, 61 km (3-8 miles) north-east of the town square in Alvarado, Johnson County. 12 and 13, USNM 400764, from USGS locality 11740, left bank of Walnut Creek, 7-6 km (4-75 miles) east-north-east of Mansfield, Tarrant County. 14-16, USNM 400765, from USGS locality DlOl 13, L6 km (1 mile) north of Lillian, west of Lillian-Retta road, Johnson county. 25, 29, USNM 420283, from 0-3 km (0-2 mile) south of Mountain Creek, 61 km (3-8 miles) north-east of the town square in Alvarado, Johnson County. All specimens from the basal part of the Eagle Ford Group. 30, the holotype, TMM 34048, from the Bluebonnet Member, 2-25 km (1-5 miles) south-east of Round Rock, Williamson County. 19 is USNM 420285; 20 is USNM 420286, both from USGS Mesozoic locality 14591 at brick pit on Cloice Branch, 1-3 km (0-8 mile) east of South Bosque, McLennan County. All specimens are from the Acantiweeras amphiholum zone. Figs. 21-24, 26-28. Tarrantoceras ciispiduin (Stephenson, 1953u). 21-24, holotype, USNM 105974; 26-28 paratype, USNM 105975, both from USGS locality 18971, gullies south of old Sherman highway, 4-5 km (2 8 miles) east of Whitesboro, Grayson County, Templeton Member, Plesiacanthoceras wyomingense zone. Figs. 17 and 18. Cimningtoniceras tonsdalei (Adkins, 1928). USNM 420270, from the basal shell bed of the Bluebonnet Member, 6-4 km (4 miles) east-north-east of Belton, Bell County. Acanthoceras bellense zone. All figures are x 1 . PLATE 14 KENNEDY and COBBAN, Tarrantoceras, Ciumingtoniceras 134 PALAEONTOLOGY, VOLUME 33 Wyoming and western South Dakota. It differs from T. sellardsi in the persistent strong ventrolateral and siphonal clavi. Tarrantoceras cuspidium (Stephenson, 1953o) (p. 202, pi. 50, figs. 1^) from the P. wyomingense zone fauna of the Templeton Member of the Woodbine Formation is discussed further below; it is much more coarsely ribbed and tuberculate than any T. sellardsi seen. Of the four species of Tarrantoceras described (Collignon 1967) from the mid-Cenomanian of the Tarfaya Basin in Morocco, three co-occur, and represent no more than a single variable species, for which we here select the name T. wrighti. They are stouter than T. sellardsi, with coarser, blunter ribbing and tuberculation ; persistent ventrolateral and siphonal tuberculation plus ribs that arise in pairs from bullae suggest they belong to some other genus. The Tarrantoceras cf. rotatile of Collingnon (1967, p. 30, pi. 16, fig. 2) is generically indeterminate from the figure and does not belong to T. sellardsi. Occurrence. Acanthoceras amphiboliim zone. Central and Trans-Pecos Texas, many localities in New Mexico, Colorado and, rarely, eastern Wyoming. Tarrantoceras midticostatum Stephenson, 1955 1955 Tarrantoceras midticostatum Stephenson, p. 61, pi. 6, figs. 21-23. Discussion. This species is carefully described and well illustrated by Stephenson (1955). None of the more than 100 specimens of T. sellardsi seen matches the holotype and paratype of T. multicostatum and they are in consequence kept separate here. Occurrence. Basal part of the Eagle Ford Group, Acanthoceras amphibolum zone, 3 6 km (2-25) miles north-east of Mansfield, Tarrant County only. Tarrantoceras cuspidiim (Stephenson, 1953a) 1953fl 1971 1980 non 1980 Plate 14, figs. 21-24, 26-28 Acanthoceras cuspidum Stephenson, p. 202, pi. 50, figs. 1-4. Protacanthoceras cuspidum (Stephenson); Kennedy, p. 122. "Acanthoceras' cuspidum Stephenson; Wright and Kennedy, p. 99, figs. 56, 59C. "Acanthoceras' aff. cuspidum Stephenson; Wright and Kennedy, p. 100, figs. 55, 59a and b. Types. Holotype is USNM 105947, by original designation; a paratype is USNM 18971, both from gullies south of the old Sherman road, 4-5 km (2-8 miles) east of Whitesboro, Grayson County. Paratype USNM 14902 is from a bluff south of the Missouri-Kansas-Texas Railroad, L6km (I mile) north and 2-9 km ( 1 -85 miles) east of Sadler, Grayson County. All are from the Templeton Member of the Woodbine Formation, Plesiacanthoceras wyomingense zone. Discussion. This species is carefully described and well figured by Stephenson (1955). Of interest are the affinities of the species. Wright and Kennedy (1980) drew attention to the distinctive asymmetry of the ventrolateral clavi, a feature common to several North American acanthoceratines such as Plesiacanthoceras and Dunveganoceras. We have since noted remarkable similarities in this feature, as well as style of ribbing, to the most ornate variants of T. sellardsi (compare PI. 14, figs. 21-24, 26-28 and PI. 14, figs. 25, 29) such that we place cuspidum in Tarrantoceras. Occurrence. As for types. Genus plesiacanthoceras Haas, 1964 (= Paracanthoceras Haas, 1963, p. 2; non Furon, 1935, p. 59) Type species. By original designation: Metoicoceras wyomingensis Reagan, 1924 (p. 181, pi. 19, figs. I and 2). KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 135 Diagnosis. Large, markedly dimorphic. Inner whorls with polygonal whorl section; bullate primaries alternate with secondaries, all ribs bearing strong conical inner and clavate outer ventrolateral and siphonal tubercles with markedly asymmetric profile; intercalated ventral ribs transiently present in some. Secondary ribs disappear after an early stage, siphonal clavi disappear at a progressively earlier ontogenetic stage in stratigraphically younger species. Inner and outer ventrolateral tubercles fuse into prominent horns in middle and late growth stages. Final rib flattened and bar-like over venter. Suture with broad E/L and L. Discussion. The type species, P. wyoniingense is the last member of a Western Interior lineage that can be traced back to Conlinoceras tarrantense via Plesiacanthoceras niuldoonense (Cobban and Scott, 1973). We also refer Mammites bellsanus Stephenson, 1953a to Plesiacanthoceras. All show variable ontogenetic development in early stages, but are linked by progressively earlier loss of siphonal tubercles, progressively earlier acquisition of primary ribs only and development of a ventrolateral horn at an early ontogenetic stage. Plesiacanthoceras niuldoonense is represented by abundant well-preserved material from the Frontier formation at USGS Mesozoic locality D9801 (south of Lone Bear Road in SWl /4 Sec. 13, T.42N., R.82 W., Johnson County, Wyoming) and USGS Mesozoic locality 23459 (near head of Fisher Draw, about 4-5 km (2 7 miles) south of Kaycee in Sec. 25, T.43 N., R. 82 W., Johnson County, Wyoming). These show inner whorls like those of Conlinoceras gilberti and C. tarrantense, with alternately long and short ribs, conical inner and clavate outer ventrolateral and siphonal tubercles to a diameter of around 25 mm (PI. 16, figs. 1-6), beyond which the outer ventrolateral and siphonal tubercles decline, disappearing by 30-40 mm diameter. The ribbing pattern modifies so that all ribs are long, with an umbilical bulla that migrates out to an inner flank position and inner ventrolateral tubercles that strengthen into a clavate upward-directed ventrolateral horn to give a quite distinctive whorl section. Ribbing declines on the outer whorl and tubercles dominate; there may be looping of riblets between ventrolateral horns, while the adult aperture shows the development of a terminal flared ventral rib. In Plesiacanthoceras wyoniingense the timing of ontogenetic development of tubercles is again modified, the siphonal row disappears by 10 mm or so, secondary ribs by 10 mm (PI. 16, figs. 7 and 8), but inner and outer ventrolateral tubercles are present and well-difTerentiated (PI. 16, figs. 1 1-15) to a diameter of 110 mm in microconchs and 160-170 in macroconchs, beyond which the outer ventrolaterals decline and a massive inner ventrolateral horn develops, projecting outwards and above the venter. Adult P. wyoniingense develop the same high, bar-like rib at the adult aperture as do ancestral P. niuldoonense . P. wyoniingense is the largest species of the lineage, and retains the ‘juvenile' characters of well- dilTerentiated inner and outer ventrolateral tubercles to a diameter where all P. niuldoonense are horned and adult. It is thus a hypermorphic giant. In spite of this, continuity of characters links the lineage into a single generic grouping in our view; resemblance to Acanthoceras and Cunningtoniceras in part reflects remote ancestry; in part evolutionary convergence. Old World Acanthoceras and Cunningtoniceras typically have a broad E/L and a narrow L. The New World taxa also have a broad E/L, but L is broad. Occurrence. Middle and upper Cenomanian of the US Western Interior and Gulf Coast. Plesiacanthoceras bellsanum (Stephenson, 1953a) Plate 2, figs. 4-8; Plate 12, fig. 9, text-fig. 23C 1953a Mammites bellsanus Stephenson, p. 204 {pars), pi. 49, fig. 3; pi. 51, figs. 8-11. 1971 Mammites ? bellsanus Stephenson; Kennedy, p. 122. Types. The holotype is USNM 105983, paratypes are USNM 105984-6, Templeton Member of the Woodbine Formation, Plesiacanthoceras wyoniingense zone, branch of Cornelius Creek, 4-3 km (2-7 miles) north 5° east 136 PALAEONTOLOGY. VOLUME 33 of Bells, Grayson County. One of the paratypes, USNM 105986, is a Metoicocems latoventer Stephenson, I953fl. Dimensions D Wb Wh Wb:Wh U USNM 105983 88-7 (100) 38-6 (43-5) 42-7 (48-1) 0-9 20 0 (22-5) USNM 105985 75-5 (100) 33-3 (44- 1) 38-3 (50-7) 0-86 12-2 (16-2) Description. USNM 105984 shows the earliest growth stage: up to a diameter of 6 mm approximately the shell is globose, smooth, non-tuberculate, bearing only low irregular folds and constrictions (PI. 12, fig. 9). A fragment at a diameter of 27 mm shows a feeble siphonal clavus, as in Plesiacanthoceras wyomingense (PI. 16, figs. 7 and 8). In middle growth coiling is fairly evolute, with 46 % of the previous whorl covered. Whorl section compressed, with maximum breadth at umbilical bulla. Ornament consists of alternately long and short ribs. The former arise at elongate umbilical bullae, 1 1 per whorl, are rursiradiate, weakened at mid-flank but strengthened into prominent rounded inner ventrolateral tubercles, connected by a broad swelling to strongly clavate outer ventrolaterals with a markedly asymmetric profile. Most ribs are long at the smallest diameter visible; as size increases shorter ribs arise around mid-flank and strengthen into inner and outer ventrolateral tubercles that match those on the long ribs. There are no siphonal tubercles. USNM 105985 is in a more robust shell, but has weaker umbilical bullae; USNM 105986 is a Mefoicoceras latoventer-, it is very compressed, and has very strong umbilical bullae. Suture with broad bifid E/L and U.^, L narrower (text-fig. 23C). Discussion. P. hellsanum is easily separated from Acanthoceras amphiholum by the very early loss of siphonal tubercles and persistence of alternately long and short ribs into middle growth plus separation of inner and outer ventrolateral tubercles to a large size. It is close to P. wyomingense , which has very flat sides, larger outer ventrolateral clavi in middle growth and enormous finger-like ventrolateral horns when adult (PI. 16, figs. 7 and 8, 11-15). Occurrence. Plesiacanthoceras wyomingense zone of north-central Texas only. Genus plesiacanthoceratoides nov. Type species. Protacanthoceras vetida Cobban, 19876 (p. 21, pi. 10, figs. 1-28; text-fig. 16). Middle Cenomanian Acanthoceras amphiholum zone Belle Fourche Shale of the Black Hills area of eastern Wyoming and basal Eagle Ford Group of north central Texas. Diagnosis. Progenic dwarf. Macroconchs adult at 30 mm or less diameter, microconchs about 60% of diameter of corresponding macroconchs. Whorls subquadrate, middle growth stages with primary ribs with umbilical bullae or not alternating with shorter intercalatories, all ribs with conical inner and clavate outer ventrolateral and siphonal tubercles, constrictions sometimes present. Adult body chambers show crowding and strengthening of ventral ribs and decline of ventral tubercles. Suture simple with broad, little-incised E/L, narrower L and small U.^. Discussion. Plesiacanthoceratoides is a homoeomorph of Protacanthoceras Spath, 1923, but whereas the latter is a progenic dwarf derivative of Acanthoceras rhotomagense, the former is derived from Plesiacanthoceras. Five successive species/subspecies of Plesiacanthoceratoides occur in the U.S. Western Interior. The oldest species appears in association with Plesiacanthoceras muldoonense. Plesiacanthoceratoides vetula (Cobban, 19876), described below, from the lower part of the Acanthoceras arnphibolum zone is succeeded by two new subspecies of a third species in the upper part of the A. amphiholum zone and a fourth in the Plesiacanthoceras wyomingense zone. Occurrence. Middle and low upper Cenomanian. North-central Texas, Wyoming and Montana. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 137 TEXT-FIG. 24. External sutures, a and B, Metoicoceras latoventer Stephenson, 1953o, USNM 105998, 106000. c and D, M. swallovi (Shumard, 1860), USNM 105991, 105993. E, Plesiacanthoceratoides vetula (Cobban, 19876), USNM 388189. f, Metoicoceras geslinianum (d’Orbigny, 1850), USNM 22938. Plesiacanthoceratoides vetula (Cobban, 19876) Plate 9, figs. 1-25, 28; text-fig. 24E. Types. Holotype is USNM 388189, paratypes USNM 388190-7, from the middle Cenomanian Acanthoceras amphiholum zone fauna of the Belle Fourche Shale at USGS Mesozoic locality D5900, on the Old Woman anticline south-west of the Black Hills, head of Elm Creek in Wl/2 sec. 14, T. 36 N, R. 62 W., Niobrara County, Wyoming. Material. Figured specimens USNM 420287 to 420292, from the concretions in the lower part of the Eagle Ford Group at USGS Mesozoic locality D12626, 8-9 km (5-2 miles) northeast of Mansfield. Tarrant County, Texas. A. amphiboliim zone. Twelve unfigured specimens, USNM 420293, from the same locality. 138 PALAEONTOLOGY, VOLUME 33 Dimensions Macroconchs D Wb Wh Wb:Wh U USNM 420287 c 19-3 (100) 10-2 (52-8) 9-4 (48-7) 109 2-9 (15-0) USNM 420288 c 19 0 (100) 9-6 (50-5) 8-9 (46-8) 108 2-6 (13-7) USNM 420289 Microconch c 20-3 (100) 10-5 (51-7) 9-4 (46-3) IT2 3-4 (16-7) USNM 420290 c 13-8 (100) 8-0 (57-9) 6-7 (48-5) 119 L6 (11-6) Description. Markedly dimorphic, macroconchs 20 mm in diameter, microconchs two thirds diameter of macroconchs. Coiling very involute with tiny, deep umbilicus. Whorl section depressed, quadrate in intercostal section, polygonal in costal section, with greatest breadth at umbilical bullae when present and below mid-flank when not. Earliest ornamented stages bear 14 ribs per whorl, limited to the outer flank, with conical inner, clavate outer ventrolateral and siphonal tubercles. As size increases, ribs extend down the flank and are irregularly long and short, long ribs extending to the umbilical shoulder, where they may develop feeble bullae, with in adults, up to 18 ribs of which 6 or 7 are bullate. Periodic broad, deep constrictions are present on the phragmocone, flanked by ribs, while there are occasional non-tuberculate ribs in early and middle growth. Adult body chambers show a crowding of ribs and strengthening ventrally to produce a marked ventral chevron; whereas tubercles dominate ribs on the phragmocone, the reverse is true on the last part of the body chamber. There is a marked ventral lappet at the adult aperture. Suture line very simple with broad bifid E/L, narrow E and little incised Uj. Discussion. Middle growth stages with constrictions are very close to those of juvenile A. amphiholum amphiholum. Mature body chamber ornament of P. vetula and disparate size immediately distinguishes the two. Occurrence. A. amphiholum zone of north-central Texas and Wyoming. Subfamily mammitinae Hyatt, 1900 p. 588 ( = Buchiceratinae Hyatt, 1903, p. 26; Metoicoceratidae Hyatt, 1903, p. 115; Fallotitinae Wiedmann, 1960, p. 741) Genus metoicoceras Hyatt, 1903, p. 115 Tvpe species. By subsequent designation by Shinier and Shrock 1944, p. 591: Ammonites swallovi Shumard, 1860 (p. 591). Metoicoceras swallovi (Shumard, 1860) Text-fig. 24C and D. 1860 Ammonites swallovi Shumard, p. 591. 1953a Metoicoceras swallovi (Shumard); Stephenson, p. 207, pi. 51, figs. 1-3; pi. 52, figs. 1-5 (with full synonymy). 1953a Metoicoceras swallovi macrum Stephenson, p. 209, pi. 51, figs. 4-7. Discussion. Stephenson described all of the surviving material of this species in detail as well as the early uncertainties surrounding it. His variety macrum, described as more compressed and with feebler ornament than the nominate subspecies, is regarded as a strict synonym. As already noted (p. 81 ) there are no longer outcrops at the type locality and the exact age of the species is unknown. All that can be said is that the suture with its little-incised elements (text-fig. 24C and D) suggests it may be from quite a high horizon in the Cenomanian, and is very distinct from that of M. latoventer (text-fig. 24A and B), but closer to that of M. geslinianum (d’Orbigny, 1850) (text-fig. 24F). KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 139 Occurrence. Templeton Member of Woodbine Formation in Lamar County (see Stephenson 1953a p. 209 for details). Metoicoceras latoventer Stephenson, 1953a Plate 17, figs. 1 and 2; text-fig. 24A and B 1953a Metoicoceras latoventer Stephenson, p. 209, pi. 53, figs. 1-9; pi. 54, figs. 9-11. ?1953a Metoicoceras crassicostae Stephenson, p. 210, pi. 58, figs. 6-8. 1953a Maniniitesl hellsanus Stephenson, p. 204 (pars), non pi. 49, fig. 3; pi. 51, figs. 8-11. Types. The holotype of M. latoventer is TMM 2574, from the Templeton Member of the Woodbine Formation, 6-4 km (4 miles) east of Whitesboro; one paratype is in the same collection. Paratypes USNM 105998-105601, 106002a-j are from gullies just south of the old Sherman Highway, 4-5 km (2-8 miles) east of the centre of Whitesboro, also from the Templeton. The holotype of Metoicoceras crassicostae is USNM 106003 from the Templeton Member on Cornelius Creek, 4-3 km (2-7 miles) north 5° east of Bells in Grayson County. All Plesiacanthoceras wyomingense zone. Material. OUM KT3926-7, 3929-3936, from the Templeton Member of the Woodbine Formation, gullies just south of old Sherman Highway, 4-5 km (2-8 miles) east by south of the centre of Whitesboro, Grayson County. Plesiacanthoceras wyomingense zone. USNM 105986, a paratype of Mammitesl bellsanus Stephenson, 1953a, is from the same horizon and locality as the holotype of Metoicoceras crassicostae (see above). Dimensions D Wb Wh Wb:Wh U USNM 105998 12-3 (100) 7-0 (56-9) 5-4 (43-9) F30 2-2 (17-9) USNM 106001 240(100) 13-4 (55-8) 10-9 (45-4) F23 4-9 (20-4) USNM 106002 c 650(100) 27-7 (42-6) 30-7 (41-2) 0-9 15-8 (24-3) USNM 106000 76-3 (100) 310 (40-6) 33-4 (43-8) 0-93 18-9 (24-8) Holotype of 94-5 (100) 37-8 (40 0) 40-8 (43-2) 0-93 25-1 (26-6) M. crassicostae USNM 100603 Discussion. Stephenson (1953a) provides a careful account of this species, and his description is not repeated here. The holotype of M. latoventer is an adult, as in the largest paratype figured by Stephenson (1953a, pi. 53, figs. 8 and 9), 113-120 mm in diameter. There are no significant dififerences between these specimens and the holotype (and only known specimen) of M. crassicostae (PI. 17, figs. 3 and 4), a near complete adult 94-5 mm in diameter, and we suspect but cannot prove them conspecific. M. crassicostae, now dated as from the Texas equivalents of the US Western Interior Plesiacanthoceras wyomingense zone, is the oldest species of the genus. It, like Metoicoceras praecox Haas, 1 949 (p. 15, pis. 5-7 ; text-figs. 5-9) of the Calycoceras canitaurinum zone, differs from all other species of the genus in the presence of feeble siphonal tubercles in early growth, up to a diameter of 20-25 mm in M. latoventer and 28-34 mm in M. praecox. Whereas M. praecox has the compressed whorl section and low broad ribs seen in later Metoicoceras species (including the type), M. latoventer has an only slightly compressed whorl even when adult, with persistent inner ventrolateral tubercles to the beginning of the outer whorl whereas these are lost at an early ontogenetic stage in later species and even lost in some variants beyond a diameter of 20 or so millimetres. This, plus the presence of a siphonal tubercle in youth suggests the origin of latoventer, and hence Metoicoceras, lay in some contemporary acanthoceratine such as Plesiacanthoceras, where there are parallel trends of progressively earlier loss of siphonal clavi and differentiated inner and outer ventrolateral tubercles. Thomelites Wright and Kennedy, 1973, suggested as a possible ancestor to Metoicoceras (e.g. Wright and Kennedy 1981, p. 40) may thus be no more than a convergent compressed acanthoceratine and not a close ally or ancestor. Occurrence. As for types. 140 PALAEONTOLOGY, VOLUME 33 Suborder ancyloceratina Wiedmann, 1966, p. 54 Superfamily turrilitaceae Gill, 1871, p. 3 Family hamitidae Gill, 1871, p. 3 Genus hamites Parkinson, 1811, p. 145 ( — Tomeutoceras Hyatt, 1900, p. 586 (objective synonym); Stomohamites Breistroffer, 1940 p. 85; Hamitella Breistroffer, 1947 p. 100 (84), twm. nov. pro. Helicoceras d’Orbigny, 1842 p. 611, non Koenig, 1825 p. 19) Type species. Hamites attenuatus J. Sowerby, 1814 (p. 137, pi. 61, figs. 4 and 5) by the subsequent designation of Diener (1925, p. 88). Hamite.s cimarronensis (Kauffman and Powell, 1977) Plate 15, figs. 11, 13, 15, 17, 19-21 1953a Hamites ? sp. Stephenson, p. 197. 1977 Stomohamites simplex cimarronensis Kauffman and Powell, p. 97, pi. 9, figs. 1, 3 and 4; text-figs. 5 and 6. Type. Holotype is USNM 167160, the original of Kauffman and Powell (1977, pi. 9, fig. 1) from USGS Mesozoic Locality 30235 in Cimarron County, Oklahoma, It is from the Hartland Member of the Graneros Shale, of middle Cenomanian age. explanation of plate 15 Figs. 1, 3, 5, 6, 10, 12, 16, 18, 22, 23, 25. Turrilites (Turrilites) acutus Passy, 1832, including transitional forms to T. (T.) costatus Lamarck, 1822 (figs. 10, 23). 1 is TMM 21055; 3 is USNM 420302; 6 is USNM 420305, 10 is TMM 35359; 16 is USNM 420303, 23 is USNM 420308; 25 is USNM 420304, from the basal shell bed of the Bluebonnet Member, Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County. Acantlwceras hellense zone. 5 is USNM 420311 ; 12 is USNM 420310; 18 is USNM 420312; 22 is USNM 420309, all from USGS Mesozoic locality D12626, roadside 8-9 km (5-5 miles) north-east of Mansfield, Tarrant County, basal part of the Eagle Ford Group, Acanthoceras amphibolum zone. Figs. 2, 4, 9, 14. Ostlingoceras (Ostlingtoceras) davisen.se Young, 1958. Specimens are from USGS Locality 14598, temporary exposure 0-8 km (0-5 mile) east of South Bosque, near railroad, McLennan County. Pepper Shale (inferred); Forhesiceras brwidrettei zone. Figs. 7 and 8. Anisoceras cf plicatile (J. Sowerby, 1819), USNM 420299, 0 3 km (0-2 mile) south of Mountain Creek, 61 km (3-8 miles) north-east of town square in Alvarado, Johnson County. Basal part of the Eagle Ford Group, Acanthoceras amphibolum zone. Figs. 11, 13, 15, 17, 19-21. Hamites cimarronensis (Kauffman and Powell, 1977). 11, USNM 420297; 17, USNM 420298, from USGS Mesozoic locality D9502, concretion in field 0 5 km (0-3 mile) north-west of Lillian, Johnson County, basal part of the Eagle Ford Group, Acanthoceras amphibolum zone. 13, 19, USNM 424123, from USGS locality 14591, abandoned brickpit on Cloice Branch, near Waco, McLennan County. 15, USNM 420296, from the Moody Hills opposite Baggett station about 7-2 km (4-5 miles) south of McGregor, McLennan County. Both Acanthoceras amphibolum zone. 20 and 21, USNM 420294, from Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County. Acanthoceras hellense zone. Fig. 24. Turrilites (Turrilites) dearingi Stephenson, 1953u. USNM 420316, from roadcut on east side of Texas highway 360, L9 km (L2 miles) south of bridge over Trinity River, Tarrant County. Tarrant Formation, Conlinoceras tarrantense zone. Figures 20 and 21 are x2; the remainder are x 1. PLATE 15 ^ r» i -if 3'.' Mi&J iVcsV f^h. KENNEDY and COBBAN, Cenomanian ammonites 142 PALAEONTOLOGY, VOLUME 33 Material. USNM 420294 and 420295, from the Bluebonnet Member of the Lake Waco Formation of the Eagle Ford Group, Bird Creek, Bell County; USNM 420296 from the same horizon near Waco, Acanthoceras bellense zone. USNM 420297 and 420298 from the basal part of the Eagle Ford Group, Acanthoceras amphibolwn zone, USGS Mesozoic locality D9502, concretion in field 0-5 km (0-3 mile) NW of Lillian, Johnson County. A specimen from USGS Locality 14591, abandoned brickpit on Cloice Branch near Waco, McLennan County. OUM KT3937-3939 from the Templeton Member of the Woodbine Formation, Plesiaccmthoceras wyomingense zone, gullies just south of old Sherman Highway, 4-5 km (2-8 miles) east by south of Whitesboro, Grayson County. Discussion. Kauffman and Powell (1977) based their Stomohamites simplex cimarronensis on a series of crushed fragments. They differentiated it from H. simplex of d’Orbigny (1842, p. 550, pi. 134, figs. 12-14) (see Kennedy and Juignet 1983, p. 13, figs. 15a-d; 17a-w; 36J; 37v and w) because it had four shafts rather than three, and was larger. The complete adult form of H. simplex is unknown to us, whereas the reconstruction given by Kauffman and Powell (1977, text-fig. 5) shows a smaller shell than is represented by English specimens of H. simplex (Kennedy 1971, pi. 1, figs. 1-8). Instead, H. cimarronensis differs from H. simplex in its dense ribbing, with a rib index of up to 8 at apparent whorl heights of 11-28 mm. In H. simplex, the index ranges from 4-5 to 7. Occurrence. Acanthoceras bellense, Acanthoceras amphibolwn and P. wyomingense zones in central Texas. Similar fragments are widespread in this interval in the US Western Interior north as far as Wyoming. Family anisoceratidae Hyatt, 1900, p. 587 (= Algeritidae Spath, 1925, p. 190) Genus anisoceras Pictet, 1854, p. 705 Type species. By original designation: Hamites saussureanus Pictet, 1847, p. 374, pi. 13, figs. 1^. Anisoceras cf. plicatile (J. Sowerby, 1819) Plate 15, figs. 7 and 8 compare : 1819 Anisoceras plicatile J. Sowerby, p. 281, pi. 234, fig. I. 1983 Anisoceras plicatile (J. Sowerby); Kennedy and Juignet, p. 25, fig. 16a-m, p and q; 19a-e; 341, m (with synonymy). Material. USNM 420299^20300, from the basal part of the Eagle Ford Group, Acanthoceras amphibolwn zone, concretion on hillside 0-3 km (0-2 mile) south of Mountain Creek and west of secondary road, 61 km (3-8 miles) north-east of town square in Alvarado, Johnson County. Discussion. Whorl section is slightly compressed oval with a rib index of 6. Rounded lateral tubercles are linked by groups of 2 or 3 ribs to larger, rounded ventrolateral tubercles, linked over EXPLANATION OF PLATE 16 Figs. 1-6, 9. 10. Plesiacanthoceras muldoonense (Cobban and Scott, 1973). 1-3, USNM 3881 17; 4-6, USNM 3881 14; 9 and 10, USNM 388121, from calcareous siltstone concretions in the Frontier Formation at USGS Mesozoic locality 23459, near head of Fisher Draw about 4-5 km (2-7 miles) south of Kaycee in sec. 25, T. 43 N., R. 82 W., Johnson County, Wyoming. Figs. 7, 8, 11-15. Plesiacanthoceras wyomingense (Reagan. 1924). 7 and 8, USNM 388164; 11-15, USNM 388165, all from the Belle Fourche Shale at USGS Mesozoic locality 22871, about 9-6 km (6 miles) north- east of Alzada in the SEj sec. 6, T. 95 N., R. 59 E., Carter County, Montana. All figures are x I . PLATE 16 KENNEDY and COBBAN, Plesiacarithoceras 144 PALAEONTOLOGY, VOLUME 33 the venter by groups of 2-3 ribs with 2-3 intercalatories between. With such small fragments confident identification is impossible. Occurrence. As for material. Family baculitidae Gill, 1871, p. 3 Genus sciponoceras Hyatt, 1894, p. 578 (= Cyrtochihis Meek, 1876, p. 392 (non Jakowlew, 1875, p. 252); Cyrtocliilella Strand, 1929, p. 8). Type species. By original designation: Hamites baculoide Mantell, 1822 (p. 123, pi. 23, figs. 6 and 7). Sciponocerasl sp. Plate 12, figs. 12-14 Material. TMM 2425, from the base of the Bluebonnet Member of the Lake Waco Formation of the Eagle Ford Group at TMM locality 245-T-24, about 0-8 km (0-5 mile) south-east of Round Rock, Williamson County. Acanthoceras bellense zone. Description. Specimen is wholly septate with a maximum whorl height of 21 -5 mm, and retains extensive areas of recrystallized shell. Whorl section compressed oval with whorl breadth to height ratio 0-8. Dorsum somewhat flattened, flanks very broadly rounded, venter rounded, only slightly narrower than dorsum. Shell surface ornamented by closely and evenly spaced growth lines and riblets. These are somewhat effaced on the dorsum, markedly concave on the dorsolateral area but markedly prorsiradiate and straight on the laterovental region, intersecting the line of the venter at 18°, strengthening, and crossing the venter in a linguoid peak. This same ornament is present on the internal mould. There are no constrictions visible. Imperfectly exposed sutures have rectangular, bifid elements. Discussion. This remarkable species differs from all other described Cenomanian Baculitidae in the absence of constrictions, the presence of which differentiate Sciponoceras from Baculites. The latter is known from the Turonian onwards, and we doubt that the present specimen suffices to extend the range of the genus back for nearly half a stage. In consequence we refer it to Sciponocerasl sp. Occurrence. As for material. Family turrilitidae Gill, 1871, p. 3 (= Pseudhelicoceratinae Breistroffer, 1953, p. 1350) Genus and Subgenus ostlingoceras Hyatt, 1900, p. 587 Type species. Turrilites puzosianus d’Orbigny, 1842 (p. 587, pi. 123, figs. 1 and 2) by original designation. Ostlingoceras (Ostlingoceras) brandi Young, 1958. 1958 Ostlingoceras brandi Young, p. 287, pi. 40, figs. 4 and 5, 7; text-fig. 1 n. 1959 Ostlingoceras brandi Young; Young, p. 37, pi. 8, figs. 2, 7. 1965 Ostlingoceras (Ostlingoceras) brandi Young; Clark; p. 37, pi. 8, figs. 2, 7. Type. Holotype is TMM 10281 from the base of the Boquillas Formation on the NE flank of the Davis Mountains, Jeff Davis County, Texas. Forbesiceras brundrettei zone. Material. One specimen from USGS locality 14592, old brickpit on Cloice Branch, F3 km (0-8 mile) east of South Bosque, McLennan County. Probably from the Pepper Shale; Forbesiceras brundrettei zone. Discussion. Specimen is an external mould of a single whorl. It shows low, even, rounded, oblique ribs, without tubercles on the outer whorl face, weakening markedly towards the base of the whorl. The lower surface bears much finer, markedly prorsiradiate riblets and growth striae. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 145 Occurrence. F. bnmdrettei zone. In addition to the present record it occurs at Gold Hill and Chispa Summit as well as the type locality, in Trans-Pecos Texas. Ostlingoceras (Ostlingoceras) daviseme Young, 1958 Plate 1 5, figs. 2, 4, 9, 14 1958 Ostlingoceras davisen.se Young, p. 289, pi. 39, figs. 29, 34. 1965 Ostlingoceras [Ostlingoceras] davisense Young; Clark, p. 36, pi. 8, figs. 1, 3. Type. Holotype is TMM 10286, from the base of the Boquillas Formation on the north-east flank of the Davis Mountains, Jeff Davis County, Texas. Material. Three specimens from USGS Locality 14598, temporary exposure 0 8 km (0-5 mile) east of South Bosque, near railroad, McLennan County. F. hrundrettei zone. Pepper Shale (inferred). Discussion. Specimens are external moulds only. Ornament consists of low, oblique, prorsiradiate ribs with a weak tubercle on the upper third of the outer whorl face, a second two thirds down the face and two closely spaced tubercles at the lower whorl suture, the lowermost concealed below the suture. The species is in some respects transitional to Mariella. Occurrence. F. hrundrettei zone. In addition to the present record the species occurs at Gold Hill and Chispa Summit as well as the type locality in Trans-Pecos Texas. Genus and Subgenus turrilites Lamarck, 1801, p. 102 (= Euturrilites Breistroflfer, 1953, p. 1351; Tiirhinites Dubourdieu, 1953, p. 42 non Martin 1809, pi. 38). Type species. Turrilites costatus Lamarck, 1801 (p. 102) by original designation. Turrilites [Turrilites) acutus Passy, 1832 Plate 12, fig. 10; Plate 15, figs. I, 3, 5, 6, 10, 12, 16, 18, 22, 23, 25. 1832 Turrilites acutus Passy, p. 334, pi. 16, figs. 3 and 4. 1977a Turrilites acutus Passy, Cobban, p. 22, pi. 4, figs. 4 and 5. 19776 Turrilites acutus Passy; Cobban, figs. 2i, 2k. 1983 Turrilites [Turrilites) acutus Passy; Kennedy and Juignet, p. 51 (with synonymy). 1985 Turrilites [Turrilites) acutus Passy; Atabekian, p. 77, pi. 28, figs. 5-13; pi. 29, figs. 1-10; pi. 30, figs. 1-1 1 (with synonymy). Lectotype. The original of Passy (1832, pi. 16, fig. 3) designated by Juignet and Kennedy (1976, p. 65), and from the middle Cenomanian Rouen Fossil Bed of Cote Ste Catherine, Rouen, Seine-Maritime, France. It is in the collections of the Sorbonne, now in the Universite Paris VI (Pierre et Marie Curie), Paris. Material. Numerous specimens; USNM 420309^20313, from USGS Mesozoic locality D 1 2626, roadside 8'9 km (5-5 mile) north-east of Mansfield, Tarrant County, in a loose concretion. One specimen, USNM 420308, from USGS Mesozoic locality D9502, concretions in field just east of gravel road, 0-5 km (0-3 mile) north- north-west of Lillian, Johnson County. Six specimens USNM 420307, from concretion on hillside 0-3 km (0-2 mile) south of Mountain Creek, 61 km (3 8 miles) north-east of the town square in Alvarado, Johnson County. All are from the basal part of the Eagle Ford Group, Acanthoceras amphiholum zone. USNM 420301^20306 and OUM KT2058 are from the basal shell bed of the Bluebonnet Member of the Lake Waco Formation on Bird Creek, 6-4 km (4 miles) east-north-east of Belton, Bell County, Acanthoceras bellense zone. Discussion. The specimens from the A. amphiholum zone are very typical representatives of this well- known species. Most of those from the A. bellense zone are equally unexceptional, but a few (e.g. PI. 15, fig. 23) are transitional to T. costatus. Clark (1965) illustrated and described specimens from 10 PAL .O 146 PALAEONTOLOGY, VOLUME 33 what he spoke of as Tarrant Formation on Pepper (e.g. Bird) Creek that are actually from the basal shell bed of the Bluebonnet Member. The original of his pi. 20, figs. 1, 2 and 7 seem to be transitions from T. acutus to T. costatus (e.g. PI. 15, figs. 1, 10, 23 herein). The same author’s T. scheuchzerianus (1965, pi. 20, fig. 6) is no more than a worn example of this passage form, as are others cited by him on p. 54. Occurrence. Turrilites acutus first appears in the middle of the middle Cenomanian, where it is widespread and common; it ranges to the lower part of the upper Cenomanian, where it is generally rare. It is known from western and eastern Europe, the USSR, North Africa, Nigeria, Angola, Zululand, Madagascar and Mozambique. In the United States there are records from California, Texas, New Mexico and Colorado; in the Western Interior and Texas it ranges from the Acanthoceras hellense zone to the A. amphibolum zone, being particularly widespread in the latter. Turrilites (Turrilites) dearingi Stephenson, 1953cf Plate 12, figs. 5 and 6, 1 1 ; Plate 15, fig. 24 1953a Turrilites dearingi Stephenson, p. 197, pi. 44, figs. 6-8. 1965 Turrilites (Turrilites) dearingi Stephenson; Clark, p. 55, pi. 20, fig. 4. 1971 Turrilites dearingi Stephenson; Kennedy, p. 31. 1976 Turrilites dearingi Stephenson; Juignet and Kennedy, p. 65. Types. Holotype is USNM 105956, paratype JPC 4134 (cast is OUM KT6105 and USNM 105957), both from the Tarrant Formation at USGS locality 20788, small branch of Big Bear Creek, 2-4 km (1-5 miles) east of Euless, Tarrant County. Conlinoceras tarrantense zone. Material. USNM 420136, plus casts of two specimens in the Gerry Kienzlen Collection (Dallas, Texas) (casts USNM 420314, 420315; OUM KT 6106-7), all from the same horizon as the types, roadcut on east side of Texas Highway 360, 1-9 km (1-2 miles) south of bridge over Trinity River, Tarrant County. Description. Coiling sinistral; apical angle 21° approximately; whorls in close contact. Intercostal section shows concealed upper surface of whorls markedly concave, outer face evenly convex, lower face convex. Costal section shows outer whorl face with strikingly concave upper, middle and lower sections. There are 13-18 ribs per whorl. They arise at the crenulated upper whorl suture and are feebly prorsiradiate, strengthening into a strong pointed tubercle at the junction of upper and middle sectors of the outer whorl face. A low, broad prorsiradiate rib connects to a smaller, feebly clavate pointed tubercle at the Junction of the middle and lower sections of the outer whorl face, and a further broad prorsiradiate rib connects to a slightly smaller tubercle of similar shape to those in the second row. This third row of tubercles is housed in a marked notch in the suture between the whorls, such that all three rows of tubercles are visible. The lower whorl face bears faint ribs that correspond to grooves in the upper whorl face of the succeeding whorl. Some specimens show a faint spiral ridge linking the lower two rows of tubercles. Sutures not seen. Discussion. Stephenson (1953a) compared this species only with the Lower Cenomanian Turrilites bosquensis Adkins, 1920 (a Wintonia) and Clark (1965) with Turrilites morrisii of Sharpe (1857) (= carcitanensis Matheron, 1842) (a Neostlingoceras). Instead, it is a close ally of Turrilites acutus EXPLANATION OF PLATE 17 Figs. 1 and 2. Metoicoceras latoventer Stephenson, 1953a. TMM 21677, a paratype from the Templeton Member 4 miles east of Whitesboro, Grayson County. Figs. 3 and 4. Metoicoceras crassicostae Stephenson, 1953a. The holotype USNM 106003, from the Templeton Member on a branch of Cornelius Creek 4-4 km (2-75 miles) north 5° east of Bells, Grayson County. All figures are natural size. PLATE 17 148 PALAEONTOLOGY, VOLUME 33 Passy, 1832 (p. 334, atlas, p. 7, pi. 16, figs. 3 and 4) from which it differs only in that the lowest row of tubercles is only slightly smaller than the second row and is much more conspicuous, not being eoncealed by the succeeding whorl. In the lectotype of T. acutus and topotypes this tubercle is far less conspicuous, although European T. acutus show the same range of rib density. Kennedy (1971, p. 31) thought T. dearingi might be an aberrant T. acutus, while Juignet and Kennedy (1976, p. 65) regarded it as a synonym of T. acutus. Of the five specimens before us, the types and the large Conlin specimen (PI. 15, fig. 24) show the prominent lowest row of tubercles to advantage, and if this is a consistent difference, it suggests that T. dearingi is probably a local subspecies of T. acutus. It is kept separate here, but whether treated as a species or subspecies of T. acutus is of no importance. Occurrence. Conlinoceras tarrantense zone. Tarrant Formation of north-east Texas only. Acknowledgements . We thank J. M. Hancock and R. Parish for assistance in the field, the staff of the Geological Collections, University Museum, and the Department of Earth Sciences, Oxford, for technical assistance. One of us (WJK) gratefully acknowledges the financial assistance of the Natural Environment Research Council, Royal Society and Astor Fund (Oxford). REFERENCES ADKINS, w. s. 1920. The Weno and Pawpaw Formations of the Texas Comanchean. University of Texas Bulletin, 1856, 1-77, pis. 1-11. (1918 imprint). 1928. Handbook of Texas Cretaceous fossils. University of Texas Bulletin, 2838, 385 pp., 37 pis. 1933. The geology of Texas. Part 2. The Mesozoic systems in Texas. University of Texas Bulletin, 3232, 239-518 (1932 imprint). and Lozo, f. E. 1951 . Stratigraphy of the Woodbine and Eagle Ford, Waco area, Texas. In lozo, f. e. and PERKINS, B. F. (eds.). The Woodbine and adjacent strata of the Waco area of central Texas, a Symposium. Fondren Science Series, 4, 101-169, pis. 1-6. ARKHANGUELSKY, A. D. 1916. The Upper Cretaceous Molluscs of Turkestan. Part I. Trudy Geologicheskago komiteta, 152, 1-57, pis. 1-8. [In Russian.] ATABEKiAN, A. A. 1985. Turrilitids of the late Albian and Cenomanian of the southern part of the USSR. Interdepartmental Stratigraphic Committee of the USSR, Transactions, 14, 112 pp., 34 pis. [In Russian.] BARRANDE, J. 1867. Systenie Silurien du Centre de la Boheme, I: Recherches paleontologicjues. 2 (Cephalopodes), pt. 1. Chez Vauteur, Prague and Paris, 712 pp., 36 pis. BASSE, E. 1952. Ammonoides. In piveteau, j. (ed.). Trade de Paleontologie, 2, 522-555; 581-688, 24 pis. Masson, Paris. BAYLE, E. 1878. Fossiles principaux des terrains. Explication de la Carte Geologique de France, 4(1) (Atlas), 158 pis. Service de la Carte Geologique detaillee, Paris. BERGQUiST, H. R. 1949. Geology of the Woodbine Formation of Cooke, Grayson and Fannin Counties, Texas. United States Geological Survey, Oil and Gas Investigation, Preliminary Map, 98. BREiSTROFFER, M. 1940. Revision des ammonites du Vraconien de Salazac (Gard) et considerations generales sur ce sous-etage albien. Travaux du Laboratoire de Geologie de la Faculte des Sciences de FUniversite de Grenoble, 22 (1938-39), 17-171 (1-101 in separates). 1947. Sur les zones d’ammonites dans f Albien de France et d’Angleterre. Travaux du Laboratoire de Geologie de la Faculte des Sciences de FUniversite de Grenoble, 26, 17-104 (1-88 in separates). 1953. L’evolution des Turrilitides Albiens et Cenomaniens. Compte Rendu Hebdomadaire des Seances de TAcademie des Sciences, Paris (D), 237, 1349-1351. BRONGNiART, A. 1822. Sur Quelques terrains de Craie hors du Bassin de Paris, 80-101. In cuvier, g. and BRONGNIART, A. Description geologique des environs de Paris, 3rd edn. G. Dufour et D’Ocagne, Paris, Amsterdam, 428pp. BROWN, c. w. and pierce, r. l. 1962. Palynologic correlations in Cretaceous Eagle Ford Group, northeast Texas. Bulletin of the American Association of Petroleum Geologists, 46, 2133-2147. CASEY, R. 1960. Hengestites, a new genus of Gault ammonites. Palaeontology, 2, 200-209, pi. 29. CLARK, D. L. 1965. Heteromorph ammonoids from the Albian and Cenomanian of Texas and adjacent areas. Geological Society of America Memoir, 95, 99 pp., 24 pis. COBBAN, w. A. 1961. The ammonite family Binneyitidae Reeside in the Western Interior of the United States. Journal of Paleontology, 35, 737-758, pis. 87-89. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 149 1972. New and little-known ammonites from the Upper Cretaceous (Cenomanian and Turonian) of the Western Interior of the United States. United States Geological Survey Professional Paper, 699, 24 pp., 18 pis. (1971 imprint). \911a. Characteristic marine molluscan fossils from the Dakota Sandstone and intertongued Mancos Shale, west-central New Mexico. United States Geological Survey Professional Paper, 1009, 30 pp., 29 pis. \911b. Fossil mollusks of the Dakota Sandstone and intertongued Mancos Shale of west-central New Mexico. In. San Juan Basin III. New Mexico Geological Society Guidebook, 28, 213-220. 1983. Molluscan fossil record from the northeastern part of the Upper Cretaceous seaway, western interior. United States Geological Survey Professional Paper, 1271, 24 pp., 15 pis. 1984. Mid-Cretaceous ammonite zones. Western Interior, United States. Bulletin of the Geological Society of Denmark, 33, 71-89. 1986. Upper Cretaceous molluscan record from Lincoln County, New Mexico. Pp. 77-89. In. Southwestern Section of AAPG Transactions and Guidebook of 1986 Convention. New Mexico Bureau of Mines and Mineral Resources, Socorro. 1987a. The Upper Cretaceous (Cenomanian) ammonites Metengonoceras dumbli (Cragin) and M. acutum (Hyatt). United States Geological Survey Bulletin, 1690, C1-C7, 3 pis. 1987/). Some Middle Cenomanian (Upper Cretaceous) acanthoceratid ammonites from the Western Interior of the United States. United States Geological Survey Professional Paper, 1445, 28 pp., 13 pis. , HOOK, s. c. and Kennedy, w. j. 1989 Upper Cretaceous rocks and ammonite faunas of southwestern New Mexico. New Mexico Bureau of Mines and Mineral Resources Memoir, 45, 137 pp. and SCOTT, G. R. 1973. Stratigraphy and ammonite fauna of the Graneros Shale and Greenhorn Limestone near Pueblo, Colorado. United States Geological Survey Professional Paper, 645, 108 pp., 41 pis. (1972 imprint). COLLIGNON, M. 1928-1929. Paleontologie de Madagascar, XV. Les cephalopodes du Cenomanien pyriteux de Diego-Suarez. Annales de Paleontologie, 17 (1928), 139-160 ( 1-24), pis. 15-19(1-5); 18(1929), 1-56(25-80), pis. 1-2 (6-7). 1937. Ammonites Cenomaniennes du sud-ouest de Madagascar. Annales Geologicptes du Service des Mines de Madagascar. 8, 29-72, pis. 1-11. 1956. Ammonites neocretacces du Menabe (Madagascar). IV - Les Phylloceratidae. V - Les Gaudry- ceratidae. VI - Les Tetragonitidae. Annales Geologicpies du Service des Mines de Madagascar, 23, 1-106, pis. 1-11. 1967. Les cephalopodes cretaces du bassin cotier de Tarfaya. Notes et Memoires. Service des Mines et de la Carte Geologique du Maroc, 175, 7-148, 35 pis. (1966 imprint). COOPER, M. R. 1978. Uppermost Cenomanian-basal Turonian ammonites from Salinas, Angola. Annals of the South African Museum, 75, (5), 51-152. CRAGIN, E. w. 1893. A contribution to the invertebrate paleontology of the Texas Cretaceous. Texas Geological Survey, 4th Annual Report (1892), 139-246, pis. 24-46. CRICK, G. C. 1907. Cretaceous fossils of Natal. In anderson, w. Third and fined report of the Geological Survey of Natal and Zululand. 161-250, pis. 10-15. West, Newman and Co., London. DIENER, c. 1925. Ammonoidea neocretacea. Fossilium Catalogus. (1: Animalia), 29, 244 pp. DODGE, c. 1969. Stratigraphic nomenclature of the Woodbine Formation, Tarrant County, Texas. Texas Journal of Science, 21, 43-62. DOUVILLE, H. 1879. (Note accompagnant le presentation de f Atlas de t. iv de I’explication de la carte geologique de France de E. Bayle et R. Zeiller). Bulletin de la Societe Geologicpie de France, 7 (3), 91-92. 1890. Sur la classification des ceratites de la Craie. Bulletin de la Societe Geologique de France, 18 (3), 275-292. DUBOURDiEU, G. 1953. Ammonites nouvelles des Monts du Mellegue. Bulletin du Service de la Carte Geologique cT Algerie, Premiere Serie, Paleontologie, 16, 76 pp., 4 pis. EORBES, E. 1846. Report on the fossil Invertebrata from Southern India, collected by Mr Kaye and Mr Cunliffe. Transactions of the Geological Society of London, (2), 7, 97-174, pis. 7-19. EURON, R. 1935. Le Cretace et le Tertiaire du Sahara soudanais. (Soudan, Niger, Tchad). Archives du Museum National cFHistoire Naturelle, Paris, (6), 13, 1-96, pis. 1-7. GILL, T. 1871. Arrangement of the Families of Mollusks. Smithsonian Miscellaneous Collections, 121, xvi + 49 pp. GROUssouvRE, A. de 1894. Recherches sur la craie superieure. 2. Paleontologie. Les ammonites de la craie superieure. Memoires du Service de la Carte Geologicpie Detaillee de France, 264 pp., 39 pis. (1893 imprint). HAAS, o. 1949. Acanthoceratid Ammonoidea from near Greybull, Wyoming. Bulletin of the American Museum of Natural History, 93, (1), 1-39, pis. 1-5. 150 PALAEONTOLOGY, VOLUME 33 1963. Paracanthoceras wyomingense (Reagan) from the Western Interior of the United States and from Alberta (Ammonoidea). American Museum Novi fates, 2151, 19 pp. 1964. Plesiacanthoceras, new name for Paracanthoceras Haas, 1963 non Furon, 1935. Journal of Paleontology, 38, 610. HATTIN, D. E. 1965. Upper Cretaceous stratigraphy, paleontology, and paleoecology of western Kansas. State Geological Survey of Kansas, field conference guidebook for the Geological Society of America. 69 pp. 1977. Upper Cretaceous stratigraphy, paleontology and paleoecology of western Kansas. Mountain Geologist, 14, 175-218. and SIEMERS, c. 1978. Guidebook. Upper Cretaceous stratigraphy and depositional environments of western Kansas. 102 pp. Kansas Geological Survey, Lawrence, Kansas. HAUG, E. 1898. Etudes sur les Goniatites. Menioires de la Societe Geologitjue de France. Paleontologie, 7, Mem., 18, 112 pp. HILL, R. T. 1902. Geology and geography of the Black and Grand Prairies, Texas. U.S. Geological Survey 2 1st Annual Report, 666 pp. (1901 imprint). HOOK, s. c. and cobban, w. a. 1983. Mid-Cretaceous molluscan sequence at Gold Hill, Jefif Davis County, Texas, with comparison to New Mexico. New Mexico Bureau of Mines and Mineral Resources Circular, 185, 48-54. HOWARTH, M. K. 1966. A mid-Turoiiian ammonite fauna from the Moqamedes Desert, Angola. Garcia de Orta (Lisboa), 14, 217-228, 3 pis. HYATT, A. 1889. Genesis of the Arietidae. Smithsonian Contributions to Knowledge, 673, xi + 238 pp., 14 pis. — 1894. Phylogeny of an acquired characteristic. Proceedings of the American Philosophical Society, 32, 349-647, pis. 1-14. 1900. Cephalopoda. 502-604. In zittel, k. a. von 1896-1900, Textbook of Palaeontology, Transl. EASTMAN, c. R. Macmillan, London and New York, 1903. Pseudoceratites of the Cretaceous. United States Geological Survey Monograph, 44, 351 pp., 47 pis. JAKOWLEW, B. 1875. Hemiptera and Homoptera of the Russian fauna. Bulletin de la Societe Imperiale des Naturalistes de Moscou, 49, 248-285. [In Russian.] JONES, T. s. 1938. Geology of Sierra de la Pena and paleontology of the Indidura Formation, Coahuila, Mexico, Bulletin of the Geological Society of America, 49, 69-150, pis. 113. JLiiGNET, p. and KENNEDY, w. J. 1976. Fauncs d’ammonites et biostratigraphie comparee du Cenomanien du nord-ouest de la France (Normandie) et du sud de I’Angleterre. Bulletin Trimestriel de la Societe Geologkpie de Normandie et Amis du Museum du Havre, 63, 1-193, 34 pis. KAUFFMAN, E. G., COBBAN, w. A. and EiCHER, D. L. 1978. Albian through Lower Coniacian strata, biostratigraphy and principal events, western interior United States. Annales du Museum cT Histoire Naturelle de Nice, 4, xxiii, 1-52, 17 pis. (1976 imprint). , HATTIN, D. E. and POWELL, J. D. 1977. Stratigraphic, paleontologie, and paleoenvironmental analysis of Upper Cretaceous rocks of Cimarron County, northwestern Oklahoma. Geological Society of America Memoir, 149, 149 pp., 12 pis. and POWELL, j. d. 1977. In kauffman, e. g., hattin, d. e. and powell, j. d. 1977, q.v. KENNEDY, w. J. 1971, Cenomanian ammonites from southern England. Special Papers in Palaeontology, 8, v-F 133 pp., 64 pis. 1988. Late Cenomanian and Turonian ammonite faunas from north-east and central Texas. Special Papers in Palaeontology, 39, 131 pp., 24 pis. , COBBAN, w. A. and hook, s.c. 1988. Middle Cenomanian (Upper Cretaceous) molluscan fauna from the base of the Boquillas Formation, Cerro de Muleros, Dona Ana County, New Mexico. Bulletin of the New Mexico Bureau of Mines and Mineral Resources, 114, 35^4. and HANCOCK, J. m. 1970. Ammonites of the genus Acanthoceras from the Cenomanian of Rouen, France. Palaeontology, 13, 462-490, pis. 88-97. and JLIIGNET, p. 1975. Presence du genre Anagaudryceras (Ammonoidea) dans le Cenomanien de Haute- Normandie. Comptes Rendu Sommaire des Sciences de la Societe Geologique de France (1975), 77-79. and 1983. A revision of the ammonite faunas of the type Cenomanian. I. Introduction, Ancyloceratina. Cretaceous Research, 4, 3-83. and 1984. A revision of the ammonite faunas of the type Cenomanian. 2. The families Binneyitidae, Desmoceratidae, Engonoceratidae, Placenticeratidae, Hoplitidae, Schloenbachiidae, Lyelliceratidae and Forbesiceratidae. Cretaceous Research 5, 93-161. , and HANCOCK, J. m. 1981. Upper Cenomanian ammonites from Anjou and Vendee, western France. Palaeontology, 24, 25-84, pis. 3-17. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 151 KOENIG, c. D. E. 1820-1825. leones Fossiliwn Sectiles. 4 pp., 19 pis., London. KOSSMAT, F. 1895-1898. Untersuchungen iiber die Sudindische Kreideformation. Beitrdge zur Paldontologie imd Geologie Oslerreich- Ungarens imd des Orients^ 9, ( 1 895) : 97—203 (1-107), pis. 1 5-25 (1-11); 11,(1 897) ; 1^6 (108-153), pis. 1-8 (12-19); 11. (1898): 89-152 (154-217), pis. 14-19 (20-25). KULMANN, J. and wiEDMANN, J. 1970. Significance of sutures in phylogeny of Ammonoidea. Paleontological Contributions, University of Kansas, 44, 1-32. LAMARCK, J. p. B. de M. de. 1801. Systeme des Animaiix sans vertebres... The author, Deterville, Paris, vii + 432 pp. LAPORTE, L. F. 1968. Ancient Environments. Englewood Cliffs, Prentice-Hall, x+115 pp. LOEBLiCH, A. R. Jr. 1946. Foraminifera from the Pepper Shale of Texas. Journal of Paleontologv, 20, 130-139, pi. 22. MANTELL, G. A. 1822. The Fossils of the South Downs: or illustrations of the geology of Sussex. Lupton Relfe, London, xvi + 327 pp., 42 pis. MARCiNOWSKi, R. 1979. Altemaccmthoceras subgen. nov. (Ammonoidea) and some remarks on other Cenomanian representatives of the genus Acanthoceras Neumayr, 1875. Acta Geologica Polonica, 29, 59-65, 1 Pl- MARTIN, w. 1809. Figures and Descriptions of Petrifactions collected in Derbyshire. D. Lyon, Wigan, etc., vii + 28 pis. MATHERON, p. 1842. Catalogue methodique et descriptif des corps organises fossiles du departement des Bouches- du- Rhone et lieitx circonvoisins. Marseille, 269 pp., 41 pis. MATSUMOTO, T. 1938. A biostratigraphic study on the Cretaceous deposits of the Naibuchi Valley, South Karahuto. Proceedings of the Imperial Academy of Japan, 14, 190-194. 1959. Upper Cretaceous ammonites of California. Part II. Memoirs of the Faculty of Science, Kyushu University, Series D, Geology, 1, 172 pp., 41 pis. 1960. On some type ammonites from the Gulf Coast Cretaceous. Science Reports of the Faculty of Science, Kyushu University, Geology, 5, 36-49. [In Japanese]. 1972. In MATSUMOTO, T., MURAMOTO, T. and TAKAHASKI, T. A new gaudryceratine ammonite from the Cenomanian of Hokkaido. Memoirs of the Faculty of Science, Kyushu University, Series D, Geology, 21, 207-215, pl. 33. , MURAMOTO, T. and TAKAHSHi, I. 1969. Selected acanthoceratids from Hokkaido. Memoirs of the Faculty of Science, Kyushu University, Series D, Geology, 19, 251-296, pis. 25-38. and OBATA, 1. 1963. A monograph of the Baculitidae from Japan. Memoirs of the Faculty of Science, Kyushu University, Series D., Geology, 13, 1-116, pis. 1-27. and 1966. An acanthoceratid ammonite from Sakhalin. Bulletin of the National Science Museum, Tokyo, 9, 43-52, pis. 1-4. MEEK, F. B. 1876. A report on the invertebrate Cretaceous and Tertiary fossils of the upper Missouri country. In HAYDEN, F. V. Report of the United States Geological Survey of the Territories, 9, lxiv-l-629 pp., 45 pis. MEREWETHER, E. A., COBBAN, w. A. and CAVANAUGH, E. T. 1979. Frontier Formation and equivalent rocks in eastern Wyoming. Mountain Geologist, 16, 67-101, 4 pis. MOREMAN, w. L. 1927. Fossil zones of the Eagle Ford of North Texas. Journal of Paleontology, 1, 89-101, pis. 13-16. 1933. In ADKINS, w. s., pp. 425^26, q.v. 1942. Paleontology of the Eagle Ford group of north and central Texas. Journal of Paleontology, 16, 192-220, pis. 31-34. MORROW, w. L. 1935. Cephalopods from the Upper Cretaceous of Kansas. Journal of Paleontology 9, 463^73, pis. 49-53. NEUMAYR, M. 1875. Die Ammoniten der Kreide und die Systematik der Ammonitiden. Zeitschrift der Deutschen Geologischen gesellschaft, 27, 854—942. NORTON, G. H. 1965. Surface geology of Dallas County. Pp. 40-125. In. The Geology of Dallas County. 21 1 pp. Dallas Geological Society, Dallas. OLIVER, w. B. 1971. Depositional systems in the Woodbine Formation (Upper Cretaceous), northeast Texas. Bureau of Economic Geology, University of Texas, Report of Investigation, 73, 28 pp. ORBiGNY, A. d’. 1840-1842. Paleontologie frangaise : Terrains cretaces. 1. Cephalopodes. 1-120 (1840); 121-430 (1841); 431-662 (1842); 148-1-3 pis., Masson, Paris. 1850. Prodrome de Paleontologie stratigraphique universelle des animaux mollusques et rayonnes. Vol. 2. Masson, Paris, 428 pp. PARKINSON, J. 1811. Organic Remains of a Eormer World. 3. Sherwood, Neily and Jones, London, xvi -f-479 pp., 22 pis. 152 PALAEONTOLOGY, VOLUME 33 PASSY, A. 1832. Description geologicjue du departement de la Seine- Inferieiire. Nicetas Periaux, Rouen, xvi + 371 pp., 20 pis. + 1 coloured map. PERViNQUiERE, L. 1907. Etudes de paleontologie tunisienne. 1. Cephalopodes des terrains secondaires. Carte Geologique de Tunisie, v + 438 pp., 27 pis. De Rudeval, Paris. PESSAGNO, E. 1969a. Upper Cretaceous stratigraphy of the western Gulf Coast area, Mexico, Texas, and Arkansas. Geological Society oj America Memoir, 111, xiii+ 139 pp., 60 pis. 1969/). Cenomanian-Turonian (Eaglefordian) stratigraphy in the western Gulf Coast plain area. Proceedings of the First International Conference on Planktonic Microfossils, Geneva, 1, 509-525. PICTET, F. J. 1847. In PICTET, F. j. and ROUx, w. 1847-54. Description des mollusques fossiles qui se trouvent dans les Gres Verts des environs de Geneve. Memoires de la Societe de Physique et d'Histoire Naturelle de Geneve, 11 (1847), 257^12; 12 (1849), 21-151; 13 (1852), 73-173; 14 (1854), 279-341; 51 pis. 1854. In PICTET, F. J. and ROUX, w. 1847-1854. q.v. and RENEViER, E. 1866. Notice geologiques et paleontologiques sur les Alpes Vaudoises et les regions environnantes. IV. Cephalopodes de Cheville. Bulletin de la Societe Vaudoise des Sciences Naturelles, 9, 93-114, pis. 3-5. POWELL, J. D. 1968. Woodbine-Eagle Ford Transition, Tarrant Member. Pp. 27^3 In Field Trip Guidebook. Second Annual Meeting, South Central Section, Geological Society of American Inc., University of Texas, Arlington. REAGAN, A. B. 1924. Cretacic Mollusca of the Pacific Slope. Pan-American Geologist, 41, 179-190, pis. 18-21. REESiDE, J. B. 1928. Cephalopods from the lower part of the Cody Shale of Oregon Basin, Wyoming. United States Geological Survey Professional Paper, 150-A, 1-19, pis. 1-8. SCFILOTHEIM, E. v. VON 1820. Die Petrefaktenkiinde aufihrem jetzigen Standpunkte durch die Beschreihung seiner Sammlung ... Becker’schen Buchhandlung, Gotha, lxii + 437 pp. SCOTT, G. 1927. Etudes stratigraphiques et paleontologiques sur les terrains cretaces du Texas. Travaux du Lahoratoire de Geologie de lUniversite de Grenoble, 14, 77-298, 3 pis. SHARPE, D. 1853-57. Description of the fossil remains of Mollusca found in the chalk of England. I. Cephalopoda. Monograph of the Palaeontographical Society, 68 pp., 27 pis. 1-26, pis. 1-10 (1853); 27-36, pis. 11-16 (1855); 37-68, pis. 17-27, (1857). SHiMER, H. w. and SHROCK, R. R. 1944. Index Fossils of North America. 837 pp. John Wiley and Sons, New York. SHIMIZU, S. 1934. Ammonites. In shimizu, S. and obata, t. Cephalopoda. Iwanami's lecture series of Geology and Palaeontology. 137 pp., Tokyo. [In Japanese.] 1935. The Upper Cretaceous cephalopods of Japan. Part 1. Journal of the Shanghai Science Institute, sec. 2,2,159-226. SHUMARD, B.F. 1860. Descriptions of new Cretaceous fossils from Texas. Transactions of the Academy of Science of St Louis, 1, 590-610. SILVER, B. A. 1963. The Bluebonnet Member, Lake Waco Formation (Upper Cretaceous), Central Texas... a lagoonal deposit. Baylor Geological Studies Bulletin, 4, 46 pp. SOWERBY, J. 1812-1822. The Mineral Conchology of Great Britain. 1, pis. 1-9 (1812), pis. 10-44 (1813), pis. 45-78 (1814), pis. 79-102 (1815); 2, pis. 103-114 (1815), pis. 115-150 (1816), pis. 151-186 (1817), pis. 187-203 (1818); 3, pis. 204-221 (1818), pis. 222-253 ( 1819), pis. 254-271 (1820), pis. 272-306 (1821); 4, pis. 307-318 (1821), pis. 319-383 (1822). The author, London. SOWERBY, J. DE c. 1823-1846. The Mineral Conchology of Great Britain (continued). 4, pis. 384-407 (1823); 5, pis. 408^43 (1823), pis. 444-485 (1824), pis. 486-503 (1825); 6, pis. 504-544 (1826); pis. 545-580 (1827); pis. 581-597 (1828); pis. 598-609 (1829); 7, pis. 610-618 (1840); pis. 619-623 (1841); pis. 624-628 (1843); pis. 629-643 (1844), pis. 644-648 (1846). The author, London. SPATH, L. F. 1922a. On the Senonian ammonite fauna of Pondoland. Transactions of the Royal Society of South 4/nca, 10, 113-147, pis. 5-9. 1922/?. On Cretaceous Ammonoidea from Angola, collected by Professor J. W. Gregory, D. Sc., F.R.S. Transactions of the Royal Society of Edinburgh, 53, 91-160, 4 pis. 1923. On the ammonite horizons of the Gault and contiguous deposits. Summary of Progress of the Geological Survey of Great Britain for 1922, 139-149. 1924. On a new ammonite (Engonoceras iris n.sp.) from the Gault of Folkestone. Annals and Magazine of Natural History, 14 (9), 504-508. 1925. On Upper Albian Ammonoidea from Portuguese East Africa, with an appendix on Upper Cretaceous ammonites from Maputoland. Annals of the Transvaal Museum, 11, 179-200, pis. 28-37. 1926a. On new ammonites from the English Chalk. Geological Magazine, 63, 77-83, table. 1926/?. On the zones of the Cenomanian and the uppermost Albian. Proceedings of the Geologists' Association, 37, 420^32. KENNEDY AND COBBAN: TEXAN CENOMANIAN AMMONITES 153 1927. Revision of the Jurassic Cephalopod Fauna of Kachh (Cutch). Memoirs of the Geological Survey of India, Palaeontologica Indica, (NS), 9, mem. no. 2, part I, I-7I, pis. I-I7. STEPHENSON, L. w. 1929. Unconformities in Upper Cretaceous series of Texas. Bulletin of the American Association of Petroleum Geologists, 10, 1323-1334. 1953a. Larger invertebrate fossils of the Woodbine Formation (Cenomanian) of Texas. United States Geological Survey Professional Paper, 242, iv + 226 pp., 59 pis (1952 imprint). — 1953/5. Mollusks from the Pepper Shale member of the Woodbine Formation, McLennan County, Texas. United States Geological Survey Professional Paper, 243-E, 57-67, pi. 13. 1955. Basal Eagle Ford fauna (Cenomanian) in Johnson and Tarrant counties, Texas. United States Geological Survey Professional Paper 247-C, 53-67, pis. 4-7. STOLiczKA, F. 1863-1866. The fossil cephalopoda of southern India. Ammonitidae with revision of the Nautilidae etc. Memoirs of the Geological Survey of India. Palaeontologica Indica, 3, (1), 41-56, pis. 26-31 (1863); (2-5), 57-106, ph. 32-54 (1864); (6-9), ’l 07-1 54, pis. 55-80 (1865); 155-216, pis. 81-94 (1866). STRAND, E. 1929. Zoological and palaeontological nomenclatorial notes. Latvifas Universitates Raksti, 20, 3-29. THOMEL, G. 1969. Reflexions sur les genres Eucalycoceras et Protacanthoceras (Ammonoidea). Compte Rendu Hehdomadaire des Seances, Paris, (D), 268, 649-652. 1972. Les Acanthoceratidae Cenomaniens des chaines subalpines meridionales. Memoires de la Societe Geologkpie de France (N.S.), 116, 204 pp., 88 pis. WEDEKIND, R. 1916. Uber Lobus, Suturallobus und Inzision. Zentralhlatt fur Mineralogie, Geologie und Paldontologie, 1916, 185-195. WHITE, c. A. 1877. Report upon the invertebrate fossils collected in portions of Nevada, Utah, Colorado, New Mexico and Arizona, by parties of expeditions of 1871, 1872, 1873 and 1874, with descriptions of new species. U.S. Geographical and Geological Explorations and Surveys west of the 100th Meridian Report, 4, ( I ), 219 pp., 21 pis. wiEDMANN, J. 1960. Le Cretace Superieure de I’Espagne et du Portugal et ses cephalopodes. Compte Rendu du Congres des Societes Savantes, Dijon, 1959: Colloque sur le Cretace superieur franyais, 709-764, 8 pis. 1966. Stammesgeschichte und System den posttriadischen Ammonoideen; ein Uberblick. Neiies Jahrhuch fur Geologie und Paldontologie Ahhandlungen, 125, 49-79, pis. 1-2; 127, 13-81, pis. 3-6. WRIGHT, c. w. 1952. A classihcation of the Cretaceous ammonites. Journal of Pcdeontology, 26, 213-222. 1955. Notes on Cretaceous ammonites. II. The phytogeny of the Desmocerataceae and the Hoplitaceae. Annals and Magazine of Natural History, 8 (12), 561-575. 1957. In MOORE, R. c. (ed.). Treatise on Invertebrate Paleontology. Part L. Mollusca 4, Cephalopoda Ammonoidea. Geological Society of America and University of Kansas Press, New York and Lawrence, xxii + 490 pp. 1963. Cretaceous ammonites from Bathurst Island, Northern Australia. Palaeontology, 6, 597-614, pis. 81-89. and KENNEDY, w. J. 1973. Paldontologie systematique. In juignet, p., Kennedy, w. j. and wright, c. w. La limite Cenomanien-Turonien dans la region du Mans (Sarthe): stratigraphie et paldontologie. Annales de Paleontologie (Invertdbres), 59, 207-242, 3 pis. and 1980. Origin, evolution and systematics of the dwarf acanthoceratid Protacanthoceras Spath, 1923 (Cretaceous Ammonoidea). Bulletin of the British Museum (Natural History), Geology, 34, 65-107. and 1981. The Ammonoidea of the Plenus Marls and the Middle Chalk. Monograph of the Palaeontographical Society, 148 pp., 32 pis. and 1984. The Ammonoidea of the Lower Chalk. Monograph of the Palaeontographical Society, Part 1, pp. 1-126, pis. 1^0. and 1987. The Ammonoidea of the Lower Chalk. Monograph of the Palaeontographical Society, Part 2. pp. 127-218, pis. 41-55. — and WRIGHT, E. V. 1951. A survey of the fossil Cephalopoda of the Chalk of Great Britain. Monograph of the Palaeontographical Society, 41 pp. YOUNG, K. 1958. Cenomanian (Cretaceous) ammonites from Trans- Pecos Texas. Journal of Pcdeontology, 32, 171-182, pis. 27-29. 1959. Index fossils of the Trans-Pecos area. Geology of the Val Verde Basin and field trip Guidebook, West Texas Geological Society, pp. 79-84. 1986. The Albian-Cenomanian (Lower Cretaceous-Upper Cretaceous) boundary in Texas and northern Mexico. Journal of Pcdeontology, 60, 1212-1219. and POWELL, j. d. 1978. Late Albian-Turonian correlations in Texas and Mexico. Annales du Museum d'Histoire Naturelle de Nice, 4, xxv, 1-36, 9 pis. (1976 imprint). 154 PALAEONTOLOGY, VOLUME 33 ZABORSKi, p. M. p. 1985. Upper Cretaceous ammonites from the Calabar region, southeast Nigeria. Bulletin of the British Museum (Natural History) (Geology), 39, U72. ziTTEL, K. A. 1884. Handbuch der Palaeontologie ... Nhi. 1, 2, (Lief 3), Cephalopoda, 329-522. R. Oldenbourg, Munich and Leipzig. 1895. Grwuhiige der Palaeontologie (Palaeozoologie). R. Oldenbourg, Munich and Leipzig, vii + 972 pp. W. J. KENNEDY Geological Collections University Museum Parks Road Oxford 0X1 3PW UK W. A. COBBAN US Geological Survey Paleontology and Stratigraphy Branch Box 25046, Mail Stop 919 Denver, Colorado 80225 USA Typescript received 5 August 1988 Revised typescript received 26 January 1989 THE ACTINOPTERYGIAN FISH PROHALECITES FROM THE TRIASSIC OF NORTHERN ITALY hy ANDREA TINTORI Abstract. The bony fish Prohalecites is redescribed from new well-preserved material from the locality of Ca’ del Prate (northern Italy), dated close to the Ladinian-Carnian boundary. A few poorly preserved specimens from the type locality, Perledo (Ladinian), have also been restiidied. The specimens represent several ontogenetic stages as evidenced by vertebral column development, and it is concluded that in structure Prohalecites is intermediate between the Parasemionotidae and Dapedium plus the Pholidophoridae, being closer to the last two. In fact Prohalecites, though similar to some of the Parasemionotidae in the dermal skull covering, has a splint-like quadratojugal, similar in shape and position to that of the Pholidophoridae, but not fused to the quadrate (as is the case for Dapedium), and ural neural arches approaching the uroneural condition of the Pholidophoridae. New finds in the Kalkschieferzone (top member of the Meride Kalk) near Ca’ del Prate (Viggiii, Varese) offer an opportunity for a revision of the genus Prohalecites Deecke 1889, so far known only from the Ladinian of Perledo (Como). New stratigraphical data (Gaetani et cti, in prep.) point to an uppermost Ladinian to lowermost Carnian age, which is somewhat older than previously thought (Tintori et al. 1985). The Kalkschieferzone is characterized by more or less marly limestone, often in thin laminated layers. The depositional environment was marine, but probably influenced by continental areas; this hypothesis is supported by the presence of the conchostracan crustacean Palaeolimnadia, a fresh-water dweller (Tintori, in press). The body parts and eggs of these organisms are often well preserved because of the total lack of oxygen in the fossilization environment. Prohalecites has not been found in the Besano-Monte San Giorgio Scisti Ittiolitici di Besano (Grenzbitumenzone) (Tintori and Renesto 1983), which includes the Anisian-Ladinian boundary. Unfortunately most of the original material used by Bellotti (1857), Deecke (1889) and De Alessandri (1910) has been lost or destroyed during the last World War. Furthermore, no material has been collected in the Calcare di Perledo-Varenna (Calcare di Perledo in Tintori et al. 1985) for at least fifty years since the cessation of quarrying. MATERIAL So far only the topmost part of the Ca' del Prate horizon (now thought to be the basal part of the Kalkschieferzone of the Meride Kalk) has been extensively studied. Most of the Prohalecites specimens come from only two bedding planes. Those from the lower, paler bedding planes (a few cm below the upper darker one) are generally smaller. This does not represent a taxonomic difference but rather a mass mortality event which affected a school of juvenile specimens, perhaps in a different season (summer?) from the later event which caused the mortality at the upper level. A great number of Palaeolimnadia has been found on a bedding plane similar to the upper one, and it is hypothesized that this upper deposition may have occurred during the rainy season (autumn/winter?), the Conchostraca having been transported into the Ca’ del Prate marine environment by river flooding from a nearby island (Tintori, in press). On both surfaces the small fishes occur at an average of about one in 100 cm" and, even though most of them show more or less the same orientation, a few specimens are randomly scattered. The I Palaeontology, Vol. 33, Part 1, 1990, pp. 155-174, I pl.| © The Palaeontological As.sociation 156 PALAEONTOLOGY, VOLUME 33 alignment may be due to a weak bottom current, but this seems unlikely since the fishes are always complete and their bones articulated. Most of the larger fishes, as well as some of the smaller, have the skull crushed dorso-ventrally, showing either the skull-roof or the gular region with the jaws and sometimes part of the snout. This kind of preservation is related to the very wide head of Prohalecites. The body is usually preserved in lateral view, but occasionally it is irregularly twisted. Thus the sea bottom must have had very low energy currents and a very high sedimentary rate to cause rapid burial of the dead fishes by the calcareous mud. Anoxic conditions were also present (see above). SYSTEMATIC PALAEONTOLOGY Subclass ACTiNOPTERYGii Cope 1871 Infraclass neopterygii Regan 1923 Genus prohalecites Deecke, 1889 PROHALECITES PORROI (Bellotti, 1957) Plate 1 ; Text-figures 1-9 1857 Pholklophonis porro Bellotti, p, 430. 1853-1860 Pholidophorus porro Bellotti; Costa, p. 65, pi. 5, fig. 9-9b. 1866 Pholidophorus porro Bellotti; Kner, p. 185. 1889 Prohalecites porro (Bellotti); Deecke, p. 125, pi. 7, figs. 5-7. 1895 Prohalecites (?) porro (Bellotti); Woodward, p, 489. 1910 Prohalecites porroi (Bellotti); De Alessandri, p. 137, pi. 9, figs. 4-5. Diagnosis (emended). Very small naked fish. Rostral followed by broad contiguous nasals. Maxilla short, no supramaxilla, quadratojugal splint-like. Preopercular made from two bones, the dorsal one being tube-like. Five infraorbitals. Parietals sometimes fused. Unpaired median extrascapular often present. Vertebral segments about 33 with hemichordacentra. Epineurals and supradorsal present. Vertebral column diplospondylous in the caudal region; unpaired median neural spines; ural-neural arches as primitive uroneurals; ural chordacentra ; a few urodermals present. Type specimens. The original material described by Bellotti (1857) was destroyed during the last World War. However I do not think it necessary to designate a neotype, the species being easily recognizable and the only one in the genus. The following material can be considered as topotypes : eleven specimens of which one is from the Curioni collection in the Museum of the Geological Survey of Italy in Rome (no catalogue number), seven are from the Ruppel collection in the Senckenberg Museum in Frankfurt am Main (SM P1239a,b; P1245-7- 8; P1251-4; P1262), one is from the Palaontologisches Institut und Museum der Universitat in Zurich (PIMUZ AI-551 ), and two are from Costa's collection in the Museo di Paleontologia dell'Universita di Napoli (MPUN M 172-3-4; M174, being the counterpart of M173). There are no accurate locality data with these specimens, although they are probably from the quarries in the middle to upper part of the Calcare di Perledo-Varenna, The available specimens are small and poorly preserved, some of them being only counterparts. Latex peels have been made, but are uninformative. Nevertheless, the fishes’ position is interesting: they are often in lateral view but, as in several Ca’ del Prate specimens, the entire skull-roof is visible. No single bone shape is detectable, owing to a peculiar kind of EXPLANATION OF PLATE 1 Prohalecites porroi (Bellotti 1857). Scale bars, if not otherwise stated, 10 mm. I, two specimens (MCSNIO P370/I-2, 38 and 36 mm s.l.) on one of the two major fossiliferous surfaces (the darker one), both dorso- ventrally crushed. 2, mature specimen (MCSNIO P349a, 30 mm s.l., see also text-fig. 7B) with skull in lateral view; note thoracic hemichordacentra and stout paired neural arches as well as pleural ribs articulating with parapophyses. 3, mature specimen (MCSNIO P373/I, 41 mm s.l., see also text-figs. 4A, 7C); note thoracic hemichordacentra with no intercalaries. 4, young specimen (MCSNIO P341 /I, 31 mm, see also text-fig. 7A) with small hemichordacentra only in the middle of the caudal region; skull shows inner surfaces of the roofing bones and of the left lower jaw as well as external surface of the right side bones. 5, young specimen (MCSNIO P376. 23 mm s.l.) with no hemichordacentra (scale bar, 5 mm). PLATE 1 TINTORI, Proluilecites 158 PALAEONTOLOGY, VOLUME 33 preservation in which the original bone is usually no longer present: only a rough natural mould shows the general shape of the fish. Other material. The 334 specimens stored in the Museo Civico di Storia Naturale di Induno Olona, Varese (MCSNIO P328 to P416). Three more specimens from Ca’ del Prate, but labelled as from Besano, are in the British Museum (Natural History) collection (BMNH P.19471-3; C. Bender Collection, purchased in 1935). The new specimens considered in this paper were prepared mainly with dilute acetic acid, but mechanical techniques were used on occasion. Most of the observations concern a few dozen specimens. Horizon and locality. The topotypes are from near Perledo (Como, northern Italy), most probably from the upper part of the Calcare di Perledo-Varenna (Scisti di Perledo auct.\ Upper Ladinian, Middle Triassic). The other material is from the Kalkschieferzone (upper member of the Meride Kalk) near Ca’ del Prate (Viggiii, Varese, northern Italy), DESCRIPTION Skull ami lower jaw The nasals (text-figs. 2,3) are joined along their whole length; the posterior nostril must have opened on the lateral side of the nasal where a notch is present, while the anterior one presumably opened along the anterior margin. The rostral (text-figs. 2^) contains the ethmoid commissure and there is a lacuna in the bony cover of the snout where the supraorbital sensory canal may have joined the ethmoid commissure itself. The true position of the antorbital (text-fig. 2) is not clear: it probably touched the corresponding nasal but not the rostral. The skull roof is very wide in the orbital region and the frontals (text-figs. 2,3) are very broad posteriorly. The parietals (text-figs. 2,3) are sometimes fused, giving rise to a large shield posterior to the frontals. The parietal pit-lines are seen as grooves, lacking the thin ganoine layer which elsewhere covers these bones. The dermopterotic (text-figs. 2,3) is trapezoidal. The sensory canal branches at about the posterior third to connect with the preopercular sensory canal. In at least one specimen the dermopterotics seem to have fused to the adjoining roofing bones. The extrascapulars (text-figs. 2,3) are unusual: three to four bones carry the temporal commissure. Between the two lateral bones sometimes there is a third, narrow, unpaired element with two symmetrically arranged pores. Paired median extrascapulars are present in several other specimens. Posterior to the antorbital and to the postero-lateral corner of the nasal there are three supraorbitals (text- figs. 2,3), the first of which is somewhat larger than the other two. The supraorbitals are followed by the dermosphenotic and the infraorbital series comprising five elements (text-figs. 2,3). The two most dorsal infraorbitals bear up to three denticles on their posterior margins. Only one suborbital (text-figs. 2,3) the upper, is known with certainty: it completely covers the uppermost part of the preopercular. Traces of a second suborbital have been seen only in MCSNIO P362 from Ca’ del Prate and in MPUN Ml 73 from Perledo, but from the configuration of the cheek we can infer that a second suborbital was probably present. The maxilla (text-figs. 2,4) is about half as long as the lower jaw, ending free below the middle of the orbit. The whole oral margin bears about 20 teeth and it is thickened, especially in the central part. No traces of a supramaxilla have been detected. The teeth are very long and conical : the dentition is remarkably powerful for such a small fish. The premaxilla (text-figs. 2,4) is triangular: the oral margin bears seven or eight teeth similar to those of the maxilla. A stout nasal process is present lying under the rostral and possibly reaching the nasal. Both the maxilla and the premaxilla are ornamented with flecks of ganoine. The lower jaw (text-figs. 2,4) bears a very high coronoid process. The dentary is the largest bone, with about 20 teeth. The sensory canal ran only in the dentary, probably being free for a short distance between the hind tip of the dentary itself and the ventral preopercular. Dorsal to the angular is the surangular which shows a very small exposed area, much as in Amia, and it is seen in only one specimen (MCSNIO P370/3). A notch is present in the lower half of the posterior margin which presumably received the articulatory head of the quadrate. However, it is not clear whether the notch is a true articulatory notch or is like the feature found in the same position in Amia. The true articulation pattern is not detectable. The opercular (text-figs. 2-4) is very broad. The subopercular is subtriangular and the interopercular is small and trapezoidal. All the opercular bones are ornamented with small tile-like ganoine flecks. The branchiostegal rays (text-figs. 2,4) are at least ten in number, gently decreasing in size forwards. The TINTORI: TRIASSIC FISH P ROH ALECITES 159 TEXT-HG. 1 . Frohcilecites porroi (Bellotti 1857). A, head of a 33 mm long specimen (MCSNIO P328/ 1 , sec also text-fig. 3B); note the fused parietals and the unpaired median cxtrascapular. B, head of a 35 mm long specimen (MCSNIO P379) witli fused parietals. C, head of a mature fragmentary specimen (MCSNIO P353); note the S-shaped left dorsal preopcrcular and the two right lateral line scales behind the supracleithrum. D, ventral view of a 35 mm long specimen (MCSNIO P370/3). E, caudal tin of a 30 mm long specimen (MCSNIO P335, see also text-fig. 9B). F, caudal fin of a 33 mm long specimen (MCSNIO P41 1 ). Scale bars, 2 mm. Lengths of fishes quoted arc standard lengths (s.l.). 160 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. Prohalecites porroi. Restoration of the skull. A, lateral view; B, dermal skull roof. Length of skull, c. 10 mm. pa esci three most dorsal ones follow the usual pattern, with the ventral edge overlapping the preceding ray. This pattern reverses at the fourth ray so that, starting from the fifth, the dorsal margin overlaps the following one. This arrangement is described by Zambelli (1975, 1978, 1981 ) for all the Norian pholidophorids from Northern Italy. The dorsal preopercular (text-figs. 2-A) is a very slender bone, gently bent forwards ventrally. Its ventral region is partially exposed and has two or three very short, backwardly-directed pegs, enveloping the branches of the sensory canal. The ventral preopercular lies just behind the lower jaw articulation and in front of the interopercular. It is a small triangular bone bearing the connection between the mandibular and the preopercular sensory canals. A pore is present at the postero-ventral corner of the bone, at the end of a branch leaving the main canal where it bends sharply upwards. A bone in a similar position, also bearing part of the preopercular sensory canal, is present in Cleithrolepis and was named quadratojugal by Wade (1935) and Hutchinson (1973). In Cleithrolepis, however, the bone is close to the posterior end of the maxilla, which is as long as the lower jaw. Later, Wade (1941) named this bone the 'second preopercular’. Furthermore, Daget (1964) pointed out that the quadratojugal always lies close to the posterior end of the maxilla and in front of the preopercular, having no sensory canal. The quadratojugal (sensu Daget 1964) is present in a few actinopterygians (Patterson 1973, p. 249; Gardiner 1984), sometimes bearing traces of the vertical pit-line (Pteronisculus, Nielsen 1942; Canobius ramsayi, Westoll 1944). Patterson (1973) seems to agree with the interpretation of Daget (1964), not citing Cleithrolepis in his list. I also consider this bone as a ventral preopercular, both because of the presence of the sensory canal and its position relative to the very short maxilla. So far, Prohalecites is the only fish having a similar bone associated with a short maxilla. The parasphenoid (text-fig. 5) has a median rounded keel and two lateral wings. Between the keel and each wing there is a groove as in Hiilettia (Schaeffer and Patterson 1984). The ascending processes arise at the level of the buccohypophysial canal opening, and have a small stout basipterygoid process anteriorly. The posterior portion of the parasphenoid is a flat lamina which is partly seen in only one specimen. No traces of dentition can be seen. In a few specimens otoliths are visible: in MCSNIO P391 /I (s.l. 24 mm) four otoliths are present just behind the parasphenoid. The anterior two are smaller and may be utricular (lapilli) while the posterior are somewhat bigger and are considered as saccular (sagittae). Otoliths are visible in small specimens; this may simply be due TINTORI: TRIASSIC FISH PROHALECITES 161 TEXT-FIG. 3. Prohalecites porroi. Skull bones as preserved in A, MCSNIO P362, s.l. 35 mm; B, MCSNIO P328/1, s.l. 33 mm (see also text-fig. la); C, MCSNIO P377, s.l. 35 mm. Scale bar, 2 mm. to the thickness of the bones in the juvenile stage: they break up easily above the hard masses of the otoliths themselves. The hyomandibular (text-figs. 4,5) is large and powerful and perforated by the hyomandibular nerve. The quadrate (text-figs. 2,4) has a stout articular head, which is buttressed by the anterior end of the quadratojugal (sensu Patterson 1973). The quadratojugal is applied to the posterior margin of the quadrate, 1 1 PAL .V! 162 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 4. Prohalecites porroi. Lower jaw and associated preopercular bones as preserved in A, MCSNIO P373/1, s.l. 41 mm (see also pi. 1, fig. 3, text-fig. 7C); B. MCSNIO P373/2, s.l. 37 mm; C, MCSNIO P338, s.l. 33 mm. Scale bars, 2 mm. TEXT-FIG. 5. Prohalecites porroi. Restoration of A, the parasphenoid (dorsal view); B, the hyomandibular (side view). •’M opr but it does not fuse to it ; the splint-like bone is very similar to that of Dapediwn and Lepidotes (Patterson 1973, p. 293). The symplectic (text-figs. 2,4) is a triangular bone lying on the inner side between the quadrate and the quadratojugal. It is much larger than the quadratojugal and extends upwards much beyond the quadrate, though not reaching the hyomandibular. So far, no trace of a condyle for articulation with the lower jaw has been detected. However, the proximal tip lies close to the quadrate articular head. The whole palate is ossified and sutures between the different bones are hard to detect. Some teeth may be present in the anterior region. All the bones, except the quadratojugal, show a cancellous structure. TINTORI: TRIASSIC FISH PROHALECITES 163 The ceratohyals are never clearly visible, but possibly there is a small ceratohyal followed by a larger epihyal. The hypohyals are stout and heavily ossified and must have been only weakly tied to each other. Girdles In the pectoral girdle the scapulocoracoid is a large plate with a pronounced antero-ventral process. The scapular foramen, close to the antero-dorsal corner, is small and round. On the inner side, a smaller foramen is present ventral to the scapular. In the coracoid region, very close to the posterior margin, there is a large elongated fenestra. At least four elongated, strong pectoral radials are visible which are enlarged distally where they articulate with the lepidotrichia. The cleithrum (text-fig. 2) is strongly convex and the anterior region is much larger than the posterior one. The external surface shows the same ornamentation as the supracleithrum, i.e. elongated ganoine flecks more or less parallel to the posterior edge. The supracleithrum (text-figs. 2,3) is narrow and elongated, somewhat wider in the upper region where the sensory canal crossed the whole bone. Posterior to the supracleithrum there are two scales bearing the lateral line. The postcleithrum (text-fig. 2) is elongated dorso-ventrally, with a gently rounded posterior edge. Its outer surface is smooth. Axial skeleton The vertebral segments (text-figs. 6-8) number about 33, 20 or 21 of which are in the caudal region. The vertebral centra consist of crescentic hemichordacentra throughout the length of the body. Usually the dorsal and ventral hemicentra do not meet though they are opposite to each other in the caudal region. However, in a couple of specimens a few centra are ring-shaped, showing that fusion has occurred between the two opposing hemicentra. In the anterior trunk region only dorsal precentra, probably related to cartilaginous intercalaries, and ventral centra with parapophyses, are present. Fully mature specimens show neural arches bearing hemicentra from the lOth-llth vertebral segment. The whole preural part of the caudal region is diplospondylous. Although precentra become larger and larger back to the middle of the caudal region they do not reach the size of centra. Hemicentra are not present in juvenile specimens : recently Schaeffer and Patterson (1984) described a similar situation for Hulettia americana and Todiltia schoewei, confirming what Patterson (1973) wrote about Eiiriconnus' hemichordacentra. In Prohalecites, chordacentra commence in the anterior caudal region in specimens of about 30 mm standard length. Initially, both ventral and dorsal centra appear in that region. Then ventral centra develop anteriorly (but parapophyses are not firmly fixed to them at this stage) and also posteriorly in the ural region. Finally, the dorsal centra reach their full extent and precentra appear. The latter are largest in the mid caudal region. The size of the precentra gives information about two possible growth gradients of the hemichordacentra, as already suggested by Schaeffer and Patterson (1984) for Hulettia: backwards in the trunk region and centrifugally from the middle caudal region. That this is the usual pattern is confirmed by observations on Norian Pholidopleuridae now being made by the author. The neural arches and spines are paired from the first to the I4th~16th segment; then median spines arc present. In young individuals, the paired arches and spines are rectangular; in adults they become thinner and more elongated. Supraneurals are present from the first vertebral segment back to the second neural spines beneath the dorsal fin radials. Their proximal ends lie between the distal ends of the paired spines. Paired elongated bean-like bones flank the dorsal tip of the supraneurals, at least in the anterior trunk region. They are in line with the distal parts of the dorsal fin radials (sec below). Unfortunately, these bones have been seen only in one of the largest specimen (MCSNIO P41 3). Similar bones are also known in Cleithrolepis (Wade 1935, fig. 25, and p. 54; Patterson pers. comm.), though here they do not have the one-to-one relationship with the supraneurals. Epineurals are well developed as posterolaterally directed outgrowths from the neural arches. They reach their maximum length between the 7th and 10th abdominal neural arches. A small forwardly- directed process is set halfway along the anterior edge of each abdominal neural arch. At the same level, there are rod-like thickenings on the medial surfaces of the arches that may be considered to be supradorsals. These thickenings, and the anterior process, mark the position in life of the longitudinal ligament, as in Amici and Salmo (Jollie 1973). Neural and haemal arches become tightly bound to supporting hemicentra in mature specimens. There are large parapophyses beneath the notochord in the abdominal region, which bear long slender pleural ribs. Posterior to the 13th or 14th vertebra there are rather expanded haemal arches, bearing long haemal spines, which are much enlarged in the last three or four pre-ural vertebrae. 164 PALAEONTOLOGY, VOLUME 33 \: TEXT-FIG. 6. Prohalecites porroi. Axial skeleton and fins as preserved in A, MCSNIO P400, s.l. 19 mm; and B, MCNSIO P392, s.l. 27 mm. Scale bars, 2 mm. No intercalaries have been observed except a few paired interventrals in the middle of the caudal region. However, external surfaces of the dorsal precentra are never completely exposed. Paired fins The pectoral fins (text-figs. 6-8) are large, their length being usually somewhat less than the head length. Each fin consists of 10 to 12 lepidotrichia preceded by two spines. The leading lepidotrich bears slender fringing fulcra. The proximal segment is very long and the more distal segments somewhat shorter. Each lepidotrich bifurcates only once or twice and always well beyond the commencement of segmentation. The insertion of the pectoral fins is ventral, very close to the mid-line. The pelvic fins (text-figs. 6-8) are small, with only six to eight long lepidotrichia and two short anterior spines in each. Slender fringing fulcra are present on the leading lepidotrich. The structure of the lepidotrichia is similar to that of the pectoral fin. A pair of long slender bones is the only endoskeletal support of the fins. Each bone has an asymmetrically enlarged distal end. TINTORI: TRIASSIC FISH PROHALECITES 165 TEXT-FIG. 7. Prohalecites porroi. Axial skeleton and fins as preserved in A, MCSNIO P341/I, s.l. 31 mm (see also pi. 1, fig. 4); B, MCSNIO P349a, s.l. 30 mm (see also pi. I, fig. 2); C, MCSNIO P373/I, s.l. 41 mm. (see also pi. I, fig. 3, text-fig. 4A). Scale bars, 5 mm. Unpaired fins The dorsal fin (text-figs. 6-8) is inserted about half way between the skull roof and the beginning of the caudal fin. The fin is short: 10 to 12 lepidotrichia are preceded by an oval median scale and four or five basal fulcra. The first lepidotrich is unbranched and short, bearing only one or two fringing fulcra. All of the main lepidotrichia branch once; their proximal segment is very long and usually there are two or three more segments before the branching. The general shape of the fin is triangular. The radials are equal in number to the lepidotrichia and the first is much larger than the others and also supports the basal fulcra. Each radial is composed of a long slender proximal part and a couple of very short distal bones which are close to the articulation with the lepidotrichia. The anal fin (text-figs. 6-8) originates a little more posteriorly than the dorsal. Its shape and size are also comparable, with 12 lepidotrichia, a few basal fulcra and a median scale. Fringing fulcra are borne by the first two lepidotrichia, the second of which is the longest in the fin. 166 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 8. Prolialecites porroi. A, MCSNIO P4I3 as preserved, s.l. 43 mm; scale bar, 10 mm; B, restoration of the axial skeleton. The caudal skeleton (text-hgs. 6-8) is sometimes well exposed but, owing to the small size of the fishes, some structures, such as the ural neural arches, are difficult to interpret. The striking similarity to the caudal skeleton of the Pholidophoridae (Patterson 1968) helps to determine where the ural structures begin. In a few specimens, a small anterior process is visible at the base of the haemal spine; easier to detect is a change in the anterior outline of the arches from convex to concave or straight. Both these features, together with the upward flexure of the posterior outline of the haemal spines, occur on the fifth elongated haemal spine, which is therefore considered to be the first hypural. The complete hypural series is composed of only six or seven elements. The first two or three, together with the last four haemal spines, support the lower lobe of the fin. The notochord was calcified even in the ural region, but usually only ventral hemicentra are present, as far as the sixth ural centrum. The last dorsal hemicentrum may be the first preural or the first ural. In a couple of specimens two hypurals are borne by a single hemicentrum, much larger than the others; the first and the second urals in one specimen, the second and the third in the other. In at least one of these specimens there is one more double hemicentrum in the caudal region but in a dorsal position; it bears two pairs of neural arches which are fused into a single median neural spine. Furthermore, in this specimen, a ring-shaped centrum is found in the middle of the caudal region. In the light of these facts the occasional fusion of the ural ventral hemicentra is considered an individual malformation rather than an indication of relationship (Patterson 1973; SchaelTer and Patterson 1984). Again, as in the Pholidophoridae s.s., the last three preural neural spines, which remain unpaired, gradually decrease in length. Posterior to them there are three or four long epurals. There are six or seven ural neural arches in one-to-one correspondence with the hypurals. They are always preserved in lateral view so that it is impossible to determine whether or not they lack a median spine. The first two or three are rather similar to each other and, though smaller, do not differ in shape from the last preural neural arches. The more posterior ural neural arches are small, longitudinally elongated and close to each other, giving rise to a continuous cover over the neural canal and approaching or even touching the hypural bases. Though similar to those of pholidophorids in their general aspect, these ural neural arches are surely more primitive. Dapedium too has similar structures, but its preural neural arches are paired (Patterson 1973). On the other hand, the modified arches are hollowed anteriorly to receive the posterior edge of the preceding arch, as illustrated in Pholidophorus hecliei by Patterson (1968), and they also closely resemble the last three ural neural arches in TINTORI: TRIASSIC FISH PROHALECITES 167 B TEXT-FIG. 9. Prohalecites porroi. Caudal fin as preserved in A, MCSNIO P33I, s.l. unknown; B, MCSNIO P335, s.l. 30 mm (see also text-fig. IE). Urodermal shaded. Scale bars, 2 mm. Leptolepis coryphaenoides (Patterson 1968). I therefore consider these ural neural arches to be uroneurals, comparable to those of Dapedium and of the Pholidophoridae s.s. (Patterson 1968, 1973). The caudal fin is moderately forked and almost symmetrical in its outline. There are 14 to 18 principal lepidotrichia, seven or eight in the lower lobe and seven to ten in the upper. The lowermost and perhaps the uppermost two principal lepidotrichia are unbranched: all the others branch once or twice. In the lower lobe, the leading ray bears fringing fulcra in its distal part, and is preceded by three or four unbranched but segmented shorter rays, which also bear a few fringing fulcra. Usually, four basal fulcra are present; they are preceded by a median scale with a short anterior process. This latter scale is considered to be a caudal scute. In the upper lobe, the proximal ends of the lepidotrichia become more and more asymmetrical upward, so that long, downwardly bent processes overlap the complete series of the upper hypurals. Along the upper margin of the fin there is a caudal scute, larger than the ventral one, and ten basal fulcra followed by slender fringing fulcra on the uppermost lepidotrich. The anterior tips of the basal fulcra lack ganoine, but the remainder of their surface is enamelled, even where covered by the preceding fulcrum. Squamution Most of the body of Prohalecites was naked or covered by very thin scales which have left no trace. In several specimens, however, the body outline is shown by small carbonized patches, much closer to each other in the abdominal region. The only known scales are ganoine-covered ; for example, the two at the beginning of the lateral line just posterior to the supracleithrum, or the median ones in front of the unpaired fins. At the base 168 PALAEONTOLOGY, VOLUME 33 of the upper lobe of the tail there is an elongated patch of eight to ten scales, arranged in a long row of six to eight scales with one or two more rounded ones dorsally. They are clearly homologous to the urodermals of some pholidophorids (Patterson 1968). The scales of the main row show more or less the same shape as the posterior uroneurals, but each slightly overlaps the following one. The two upper urodermals are overlapped by the epaxial basal fulcra. The whole urodermal scale patch covers the proximal ends of the upper lobe lepidotrichia, but seems to lie just posterior to the last uroneural and hypural; thus the uroneurals form a continuous series with the urodermals, except that the two are distinguished by the marginal overlapping of the urodermals. TAXONOMIC REMARKS Relationships of Prohalecites No detailed anatomical description of Prohalecites has been made since that of De Alessandri (1910), but the genus was often mentioned or listed in papers concerning Triassic fish faunas. Brough (1939, p. 107) considered Prohalecites as a possible sub-holostean while both Griffith (1977, p. 81) and Patterson (1981, p. 217) tentatively ascribed the genus to primitive teleosts, perhaps in the light of its original designation as ^ Pholiclophorus'. Even now, after a much more detailed description, the taxonomic position of Prohalecites is difficult to define owing to the presence of both advanced and primitive characters. In many respects (especially the axial skeleton), it seems to fit rather well in the Pholidophoridae s.s. Outside the Pholidophoridae, chordacentra are known in the Pholidopleuridae and Caturidae (Patterson 1973), Ophiopsidae (Bartram 1975), Hiilettia, and immature Todiltia (Schaeffer and Patterson 1984), as well as in several other groups. Furthermore, they seem to be present also in other undescribed genera from the Italian Upper Triassic (pers. obs.). The caudal endoskeleton shows more or less the same organization as in pholidophorids, with ural neural arches slightly modified and approaching the uroneural stage. The dermal skull also shows a striking resemblance to the pholidophorids in the position and shape of the quadratojugal, even though this bone is not yet fused with the quadrate in Prohalecites. The infraorbitals are also very similar in shape apart from the postero-ventral one. Many other characters, however, are remarkably different : for example the snout pattern has large contiguous nasals, the preopercular is double and shows a tube-like dorsal part, the maxilla is very short, and there are no supramaxillae. Thus, Prohalecites cannot be placed in the Pholidophoridae sensu Nybelin (1966), or even as emended by Zambelli (1981), or sensu Patterson (1973). The dermal skull pattern of Prohalecites is much like that of advanced Parasemionotidae (sensu Patterson 1973), genera such as Promecosomina and especially Paracentrophorus and Phaidrosoma (Griffith 1977), but Parasemionotidae have no vertebral centra (Patterson 1973), while Paracentrophorus has no fringing fulcra (Gardiner 1960). Prohalecites also shows some resemblance to the Caturidae, but their quadrate and symplectic articulation with the lower jaw is so far unknown in Prohalecites. The very short maxilla, without any supramaxilla, is here considered a derived character which might have formed either by the shortening of a long toothed maxilla and the concomitant loss of the supramaxilla, or directly from a maxilla not yet provided with a supramaxilla. This opinion is in contrast with that of Schaeffer and Patterson (1984) who consider the lack of supramaxillae as a primitive character in Hulettia, as in all the other chondrosteans with long maxillae. In some Parasemionotidae, Promecosomina and Phaidrosoma for instance, the supramaxilla is also absent and the maxilla is rather short. Among Triassic fishes, the Semionotidae with grinding dentition also show a short maxilla, but this is usually toothless. The caudal endoskeleton of Prohalecites is rather similar to that of Dapedium (Patterson 1973, fig. 27), which is, in my opinion, close to Pholidophorus hechei (Patterson 1968, fig. 5). In many other respects, apart from the large contiguous nasals in Dapedium and some other genera, the narrow but single opercular and the splint-like quadratojugal in Lepidotes and Dapedium (Patterson 1973), the Semionotidae are quite different from Prohalecites. The heavy scale covering, lack of chordacentra (but Tetragonolepis apparently does have them; Patterson 1973, TINTORI: TRIASSIC FISH PROHALEC/TES 169 p. 294), small mouth, and body more or less elevated are in contrast with the characters of Prohalecites. Since no trace of scales is preserved in any specimen, the body of Prohalecites is considered naked. This helps little in the search for possible relatives; among the inferred ancestors, none shows thin or absent scales, but scales are wanting in a few Pholidophoridae and Caturidae. This lack of a scale covering led Woodward ( 1895) to consider Prohalecites porroi all as immature specimens. However, De Alessandri (1910) pointed out that, though the smallest specimens may well be young, the large ones show adult characters, especially in body proportions. Though De Alessandri’s arguments are perhaps superficial, the ontogenetic development of the vertebral column described herein proves without doubt that Prohalecites porroi represents a naked species of about 40 mm standard length. Moreover, fishes other than Prohalecites are uncommon in the Ca' del Prate beds, comprising less that 10% of the total collected specimens from about ten other species. None of the latter shows any tendency towards a reduction in scale covering. Less important characters, such as fusion of the parietals and the presence of a median extrascapular, are found in Amia, Sinamia, and Ikechaoamia (Patterson 1973; Stensio 1935; Su Te- tsao 1973; Zhang Mi-man and Zhang Hong 1980; Jain 1985). However, this character alone is not sufficient to prove a relationship, because fusion of the parietals has occurred several times in different groups, including the Pholidophoridae themselves (Zambelli 1975, 1978), while a median extrascapular is common in many stem-group neopterygians. In Prohalecites, a ventral preopercular separates the interopercular from the hind edge of the lower jaw as in Macrosemiidae. This was considered a unique specialization of that family by Bartram (1977). Apart from this last character, and the partial fusion between the quadratojugal and quadrate, macrosemiids and Prohalecites are very different in other features. Recently, Schaeffer and Patterson (1984) gave a detailed description of Hulettia americana, a mid- Jurassic fish, which is rather similar to Prohalecites in many features. Major differences between the two genera are the position of the rostral, which is anterior to the totally contiguous nasals in Prohalecites, and the caudal endoskeleton. In this latter, Prohalecites shows hemichordacentra and ossified ural neural arches, which can be considered as uroneurals seusa Patterson (1973); none of these characters is present in Hulettia. The fact that Prohalecites is a neopterygian is proved by the presence of several characters among those listed by Patterson (1973), such as the reduced body lobe of the tail, dorsal and anal fin radials equal in number to their lepidotrichia, premaxilla with nasal process, coronoid process of the lower jaw, suspensorium vertical, tube-like dorsal preopercular and loss of clavicles. However, problems arise when trying to evaluate closer relationships. Similarities between Prohalecites and the Parasemionotidae and Caturidae for example are not substantiated and there is no articulation between the symplectic and the lower jaw in Prohalecites itself. This articulation, however, is also not recorded in Parasemionotidae s.l. such as Paracentrophorus, Promecosomiua, and Phaidrosoma which are the closest to Prohalecites at least in the skull-bone pattern. On the other hand, the presence of chordacentra in Prohalecites makes it difficult to put all these genera together. What relationships exist with the Pholidophoridae is difficult to determine: very close similarities are seen in the caudal endoskeleton and in a few characters of the dermal skull. That the quadratojugal is not completely fused with the quadrate is primitive with respect to the pholidophorids, but the juxtaposition of these elements is similar. Unfortunately, there is a lack of neurocranial information for Prohalecites. It is also worth considering the similarities between Prohalecites and Dapedium. Once more, the caudal skeleton as well the lower jaw joint are very similar in the two. On the other hand, the similarities of Prohalecites both with Dapedium and with the Pholidophoridae is consistent with Olsen’s (1984) view that Dapedium was the closest relative of Pholidophorus. In conclusion, Prohalecites is clearly more advanced than the Parasemionotidae and somewhat more primitive than Dapedium and the Pholidophoridae; moreover, it is certainly closer to the last two than to the former. Nevertheless, I think it better to leave Prohalecites as a Neopterygian incertae sedis because its characters do not perfectly fit in any of these cited groups. 170 PALAEONTOLOGY, VOLUME 33 Species of Prohalecites Other problems arise at the species level because of the different preservation of the Perledo and Ca’ del Prate specimens. A comparison between the two groups of specimens show that they are very similar in most of the known characters. According to De Alessandri (1910) the standard length range is 20^5 mm, comparable with the 19^3 mm range of my specimens. Fin positions and lepidotrichia number are also similar in each, except for the caudal fin. In the latter, De Alessandri recorded 25-26 segmented rays with six more ‘anterior rays, small and shortening backwards’. Four of the Ladinian specimens (SM pi 247, 54, 62 and MPUN Ml 74), however, show a smaller caudal fin with less than 20 principal rays, as in the Ca’ del Frate specimens; the others are too poorly preserved to count the lepidotrichia. Hemichordacentra are found in both groups, ranging in number from 32 (De Alessandri 1910) to 33-35 in the Ca’ del Frate specimens. The scales also have the same distribution, few in number and only at the base of the upper lobe of the caudal fin. The two most remarkable differences are the opercular shape, triangular in the Ladinian specimens (De Alessandri 1910) but rectangular in the Carnian, and the absence of teeth in the Ladinian specimens (De Alessandri 1910) compared with the well-toothed mouth of the new specimens. However, I have been able to prepare specimen MPUN Ml 73, the only Perledo one with bone preserved : teeth are present at least on the lower jaw. Considering that two of the supposed differences resulting from De Alessandri’s description proved to be untrue, we may have doubts regarding the other character. The Perledo specimens can easily be misinterpreted owing to their poor preservation. On the basis of these considerations I include the new Carnian material in the existing Prohalecites species, P. porroi (Bellotti 1857), at least until new or better preserved material from Perledo, or from coeval beds, is found. So far, P. porroi is still the only species of Prohalecites because " Pholiclophoriis' microlepidotus Kner 1866 is very different; it has cycloid scales, a large caudal fin with about 40 lepidotrichia, including few epaxial rays, no vertebral centra, and no pelvic fins (pers. obs.). Therefore, it cannot be related to Prohalecites porroi, even if Kner ( 1866) himself thought the two species very close to each other. On the other hand, Deecke (1889), in proposing the new genus Prohalecites for Pholidophorus porroi, also noticed some differences, which suggested that Kner’s species should not be included. GENERAL REMARKS Nybelin (1966, 1974) placed great emphasis on the preopercular structure in establishing phylogenetic relationships, and subsequently the reduction of the dorsal limb of the preopercular has been used for hypothesizing relationships between Amia, gars, and teleosts (Olsen 1984). The preopercular in the most advanced Parasemionotidae (seusii Patterson 1973), such as Paracerttrophorus, Promecosomina, and Phaidrosoma, is very narrow, much like the dorsal preopercular in Prohalecites, but with no posterior branches of the sensory canal (Gardiner 1960). Lehman (1952) and Lehman et al. (1959) postulated that this narrow preopercular might well have been produced in some Paraseminotidae (Thomasinotus, Stensionotus, and Jacohulus) by separation of the suborbitals (fragmentation), thus losing the area in front of the sensory canal. A few very short branches of the sensory canal are found in Prohalecites, while in the Pholidophoridae they are longer and surrounded by laminar bone, and the posterior region is expanded. The ventral preopercular of Prohalecites has one long branch of the sensory canal, but it is expanded posteriorly so that it could bear additional branches. In Leptolepis normandica (Nybelin 1974), as well as in some Recent teleosts such as the salmon, the preopercular comprises two bones; a ventral compound one, with canal elements attached to a laminar base early in ontogeny, and a small dorsal tubular bone called the suprapreopercle (Jollie 1984). The dorsal preopercular of Prohalecites is not strictly homologous to the suprapreopercle, since it is usually associated with three or four neuromasts, whereas the suprapreopercle has none (Jollie 1984). However, both the dorsal preopercular and the suprapreopercle are simple canal TINTORI: TRIASSIC FISH PROH ALECITES 171 M TEXT-FIG. 10. Possible evolution of the preopercular bone(s). A, Phaidrosoma lunzensis, after Griffith 1977, tig. 23; B. Prohalecites porror, C, Pholidoctenus serianus, after Zambelli 1978, fig. I ; D, Panipholidophorus nyhelini, after Zambelli 1975, fig. I; E, Pholidorhynchodon malzannii, after Zambelli 1980, fig. I; F. Pholidophorus latiuscuhis gervasuttii, after Zambelli 1980, fig. I; G, Pholidophorus bechei, after Nybelin 1966, fig. 16; H, Pholidolepis dorsetensis, after Nybelin 1966, fig. 16; 1, Proleptolepis furcata, after Nybelin 1974, fig. 17k; L, Leptolepis normandica, after Nybelin 1974, fig. 2e; M, Leptolepidcs sprattifonuis, after Nybelin 1974, fig. 30a. bones. This may be interpreted as a tendency in neopterygians to have more than one bone along the preopercular sensory canal, but with only the ventral part as a compound (tubular plus laminar) bone. The Prohalecites ventral preopercular resembles in shape the ventral region of the preopercular in a few pholidophorids, such as Pholidophorus latiusculus, P. hechei, and Pholidolepis dorsetensis (Nybelin 1966, p. 428), and Pholidorhynchodon /na/ca/tmY (Zambelli 1981). Furthermore, in another pholidophorid, Pholidoctenus serianus (Zambelli 1978), the preopercular is double, its antero- ventral part bearing only a very short sensory canal. Zambelli (1978) noted that the two preopercular bones in Pholidoctenus are very similar in shape to the single preopercular of other Triassic pholidophorids, implying that the genus was derived from genera more advanced in other characters by splitting of the preopercular. Zambelli (1986) wrote that the anterior preopercular of Pholidoctenus was lost in the main pholidophorid lineage during the Jurassic, leaving the quadrate uncovered by dermal bone. In my opinion, this is incorrect, because in other pholidophorid genera the sensory canal ran ventral to the notch which Zambelli (1978, fig. 4) interpreted as the point of separation of the parts of the preopercular. Since, in Parasemionotidae, the preopercular is 172 PALAEONTOLOGY, VOLUME 33 presumed to have lost the sensory canal-free anterior region, it seems improbable that part of the sensory canal itself appeared again on a suborbital-like bone. I think that PIrolidoctenus (which is a primitive genus because its nasals are contiguous for their whole length) could more easily have achieved its preopercular structure by ventral growth of the principal bone, while the primitively separate ventral bone shifts forward, losing most of its sensory canal. A further step was the fusion of the two bones (as is presumed to have occurred in the other Triassic genera) resulting in the absence of the sensory canal in the area anterior to the antero-ventral notch. The groove on the inner surface, starting from this notch, is here interpreted as a trace of fusion and not as an early stage of splitting (Zambelli 1978, 1986). Accordingly, I think that Nybelin’s assumption (1966, p. 429) about the primitiveness of the preopercular sensory canal position on the bone is incorrect. In more advanced Parasemionotidae, the narrow stage of the preopercular had already been reached, and in Prohalecites, as we have seen, it is just a tube of bone round the sensory canal. Acquisition of the inflated postero-ventral region, together with the long posterior branches of the sensory canal, could be achieved by the formation of a ventral preopercular as in Prohalecites and the subsequent fusion of these two preopercular bones. This fusion may have occurred more than once, giving two distinct patterns. The first is seen in Pholidoctenus and most of the other Triassic pholidophorids (text-fig. IOC, D, E) where the ventral preopercular is presumed to have fused along the antero-ventral edge of the dorsal preopercular. The second pattern is thought to have developed from a more simple dorso-ventral fusion between the two bones, which often leaves a posterior notch, as in Pholidophorus latiuscidus (text- fig. lOF) and a few Jurassic pholidophorids (Pholidophorus bechei and Pholidolepis dorse tensis as well as in Proleptolepis for instance; text-fig. lOG, H, I). If Prohalecites was ancestral to the Pholidophoridae, then a preopercular such as that of Pholidoctenus is primitive compared to that of the other late Triassic pholidophorids (with a deep antero-ventral notch), and Pholidoctenus and Pholidorhynchodon could not have been ancestral to the main Lower Lias pholidophorid to leptolepid lineage. Nybelin had already noticed this fact (1966, fig. 16) that all the Lias species are derived from unknown or hypothetical ancestral forms. Recently, Zambelli (1986) also wrote that no Upper Triassic genus of his new subfamily Pholidophorinae (Pholidophorus, Parapholidophorus, Pholidoctenus, and Pholidorhynchodon) could be directly ancestral to teleosts, even if he supposed that Pholidophorus had to be the closest relative to Lias Pholidophoridae. Finally, concerning the shape of the preopercular, Aniia is like the Parasemionotidae, while gars, in which the ventral branch is well developed and L-shaped, are most like the leptolepids. If relationships between Prohalecites and the pholidophorids are to be strengthened, then this preopercular character could be of interest in relation to the different hypotheses (Patterson 1973; Olsen 1984) for gars, Aniia and teleost relationships. Acknowledgements. Thanks are due to F. Stojaspal, Geologische Bundesanstalt, Wien; G. Plodowski, Forschungsinstitut Senckenberg, Frankfurt am Main; K. A. Hunermann, Palaontologisches Institut und Museum der Universitat, Zurich; F. Angelelli, Servizio Geologico di State, Roma, and G. Bonaduce, Museo Paleontologico dell’Universita, Napoli, for the loan of specimens. I am particularly indebted to C. Patterson, British Museum (N.H.), London, for criticizing the manuscript: his advice greatly improved the paper. Also R. Nursall (University of Alberta, Edmonton) and an anonymous referee helped me in improving the manuscript. The field work was done by E. Bigi, G. F. Crugnola, and G. L. Danini for the Induno Olona Museo Civico di Storia Naturale, which gave financial support. S. Tuscano rediscovered the locality of Ca' del Frate after decades of obscurity. I am grateful to Professor C. Rossi Ronchetti for access to her M.P.I. 40% grant ‘Triassico Alpi Meridionali e Mediterraneo ’. Final drawings by C. Ferliga; photographs by G. Chiodi. Abbreviations af, anal fin; afr, anal fin radial; ang, angular; ant, antorbital; asp, ascending process of parasphenoid ; bb, ‘bean’ bone; bf, basal fulcra; bp, basipterygoid process of parasphenoid; br, branchiostegal ray; ch, hemichordacentrum ; chu, ural hemichordacentrum ; cl, cleithrum; cs, caudal scute; df, dorsal fin; dfr, dorsal fin radial; dfs, dorsal fin scale; dn, dentary; dpt, dermopterotic; dsp, dermosphenotic; en, epineurals; ep. TINTORI: TRIASSIC FISH PROHALECITES 173 epurals; escl, lateral extrascapular; escm, median extrascapular; fr, frontal; hs, haemal spine; hym, hyomandibular; hypl, first hypural; i, intercalary; io, infraorbital; iop, interopercular; Iw, lateral wing of parasphenoid ; mx, maxilla; na, nasal; neu, paired neural arch and spine; ns, median neural spine; op, opercular; opr, opercular process of hyomandibular; pa, parietal; pf, pectoral fin; pci, postcleithrum ; pmx, premaxilla; popd, dorsal preopercular; popv, ventral preopercular; pp, parapophysis; ppl, pelvic plate; pt, posttemporal; q, quadrate; q^j, quadratojugal ; r, pleural ribs; ro, rostral; s, lateral line scale; sang, surangular; sbo, suborbital; scl, supracleithrum; sd, supradorsal; sn, supraneural; so, supraorbital; sop, subopercular; sy, symplectic; ud, urodermal; un, uroneural. REFERENCES BARTRAM, A. w. H. 1975. The holostean fish genus Ophiopsis Agassiz. Zoological Journal of the Lirmean Society 56, 183-205. 1977. The Macrosemiidae, a Mesozoic family of holostean fishes. Bulletin of the British Museum (Natural History), Geology Series, 29, 127-234 BELLOTTi, c. 1857. Descrizione di alcune nuove specie di pcsci fossili di Perledo e di altre localita lombarde. 419-432. In stopani, a. (ed.). Stucli geologici sulla Lombardia. Editore Turati, Milano BROUGH, J. 1939. The Triassic fishes of Besano, Lombardy. British Museum (Natural Elistory), London, 1 17 pp. costa, g. o. 1853-1860. Ittiologia fossile italiana. Napoli, 67 pp. DAJET, J. 1964. Le crane des teleosteens. Memoires du Museum nationale cfHistoire naturelle, A 32, 163-342. DE ALESSANDRi, G. 1910. Studi sui pesci triassici della Lombardia. Memorie della Societa Italiana di Science Naturali, 3, 1-145. DEECKE w. 1889. Ueber Fischer aus vershiedenen Horizonten der Trias. Palaeontographica, 35, 97-138. GARDINER, B. G. 1960. A revision of certain actinopterygian and coelacanth fishes, chiefly from the Lower Lias. Bulletin of the British Museum (Natural History), Geology Series, 4, 241-384. 1984. The relationships of the palaeoniscid fishes, a review based on new specimens of Mimia and Moythomasia from the Upper Devonian of Western Australia. Bulletin of the British Museum (Natural history). Geology Series, 37, 173-428. GRIFFITH, J. 1977. The Upper Triassic fishes from Polzberg bei Lunz, Austria. Zoological Journal of the Linnean Society, 60, 1-93. HUTCHINSON, p. 1973. A revision of the redfieldiiform and perleidiform fishes from the Triassic of Bekker's Kraal (S. Africa) and Brookvale (New South Wales). Bulletin of the British Museum (Natural History), Geology Series, 22, 233-254. JAIN, s. L. 1985. Variability of dermal bones and other parameters in the skull of Amia calva. Zoological Journal of the Linnean Society, 84, 385-395. JOLLIE, M. 1973. Chordate morphology. R. E. Krieger, Huntington, 478 pp. 1984. Development of the head skeleton and pectoral girdle of salmons, with a note on the scales. Canadian Journal of Zoology, 62, 1757-1778. KNER, R. 1866. Die Fische der bituminosen Schiefer von Raibl in Karnthen. Sitzungsberichte der Akademie der Wissenschaften in Wien, 53, 152-197. LEHMAN, J. p. 1952. Etude complementaire des poissons de I’Eotrias de Madagascar. Kongliga svenska vetenskapsakademies Handlingar, 2, 1-202. CHATEAU, c., LAURAiN, M. and NAUCHE, M. 1959. Paleoiitologic de Madagascar. XXVIH. Les poissons de la Sakamena moyenne. Annales de paleontologie, 45, 175-219. NIELSEN, E. 1942. Studies on Triassic fishes from East Greenland. E Glaucolepis and Boreosonms. Meddelelser om Gronland, 138, 1-394. NYBELiN, o. 1963. Zur Morphologie und Terminologie des Schwanzskelettes der Actinopterygier. Arkiv for Zoologi, 15, 485-516. 1966. On certain Triassic and Liassic representatives of the family Pholidophoridae s. str. Bulletin of the British Museum (Natural History), Geology Series, 11, 353^32. 1974. A revision of the leptolepid fishes. Acta Regiae Societatis scientiarum et litterarum gothoburgensis (Zoologica), 9, 1-202. OLSEN, p. R. 1984. The skull and pectoral girdle of the parasemionotid fish Watsonulus eugnathoides from the Early Triassic Sakamena Group of Madagascar, with comments on the relationships of the holostean fishes. Journal of Vertebrate Paleontology, 4, 481^99. PATTERSON, c. 1968. The caudal skeleton in Lower Liassic pholidophorid fishes. Bulletin of the British Museum (Natural History), Geology Series, 16, 203-239. 174 PALAEONTOLOGY, VOLUME 33 1973. Interrelationships of holosteans. 233-305. In greenwood, p. h., miles, r. s. & Patterson, c. (eds), Interrelationships of fishes. Academic Press, London. 1981. Agassiz, Darwin, Huxley, and the fossil record of teleost fishes. Bulletin of the British Museum (Natural History), Geology Series, 35, 213-224. SCHAEFFER, B. and PATTERSON, c. 1984. Jurassic fishes from the Western United States, with comments on Jurassic fish distribution. American Museum Novitates, 2796, 1-86. STENSio, E. A. 1935. Sinamia zdanskyi, a new amiid from the Lower Cretaceous of Shantung, China. Paleontologia Sinica, C3, 1^8. su TE-TSAO 1973. A new Sinamia (S. huananensis, sp. nov.) from the Upper Jurassic of Southern Anhui. Vertehrata Palasiatica, 11, 149-153. TiNTORi. A. and RENESTO, s. 1983. The Macrosemiidae (Pisces, Actinopterygii) from the Upper Triassic of Lombardy (N. Italy). Rivista italiana di Paleontologia e Stratigrafia, 89, 209-222. MUSCio, G. and nardon, s. 1985. The Triassic fossil fishes localities in Italy. Rivista italiana di Paleontologia e Stratigrafia, 91, 197-210. WADE, R. T. 1935. The Triassic Fishes of Brookvale, New South Wales. British Museum (Natural History), London, 1 10 pp. 1941. Australian Triassic Fishes. I. The Triassic fishes of St. Peter's, Sydney, New South Wales. II. The relationships of the Australian Triassic fishes to each other and to other bony fishes. Journal and Proceedings of the Royal Society of New South Wcdes, 74, 377-396. WESTOLL, T. s. 1944. The Haplolepidae, a new family of Late Carboniferous bony fishes. Bulletin of the American Museum of Nat wed History, 83, 1-121. WOODWARD, A. s. 1895. Catalogue of the fossil fishes in the British Museum (Natural History). British Museum (Natural History), London, 544 pp. ZAMBELLi, R. 1975. Note sui Pholidophoriformes. I. Parapholidophorus nybelini gen. n. sp. n. Istituto Lombardo Accademia di Scienze e Lettere, B109, 3-49. 1978. Note sui Pholidophoriformes. II. Pholidoctenus serianus gen. n. sp. n. Rendiconti Accademia Nazionale delle Scienze detta Dei, XL, 3, 101-124. 1981. Note sui Pholidophoriformes. IV. contributo: Pholidorhynchodon malzannii gen. n. sp. n. Rivista Museo civico Scienze naturali “£. Caffi", 2 (1980), 129-159. 1986. note sui Pholidophoriformes. VI. contributo; Pholidophorinae subfamiglia nuova del Triassico Superiore. Rivista Museo civico Scienze naturali “£. Caffi", 10, 1-32. ZHANG MI-MAN & ZHANG HONG 1980. Discovery of Ikechaoamia from South China. Vertebrata Palasiatica, 18, 89-93. ANDREA TINTORI Dip. Scienze della Terra Via Mangiagalli 34 Typescript received 5 June 1988. 20133 Milano, Italy Revised typescript received 25 January 1989 THE CLASSIFICATION, ORIGIN AND PHYTOGENY OF THECIDEIDINE BRACHIOPODS by P. G. BAKER Abstract. Recent studies on ontogeny and shell microstructure enable recognition of the inherent weakness of basing thecideidine taxonomy and phylogeny on morphological characteristics of separated brachial valves. Some previous assumptions are invalidated. The more robust components of lophophore supports are useful in determining relationships, but the form of the brachial lobes, because of their fragility, is less easy to establish, and, without the supporting evidence from sectioned complete shells, their value as taxonomic indicators is questionable. General evolutionary trends may be established through increasing complexity of lophophore supports, but for a clear understanding of thecideidine phylogeny detailed investigation of ontogeny and shell microstructure is required. Neotenous origin and the masking efl'ects of convergent evolution have combined to obscure the line of thecideidine descent. However, shell microstructural evidence now points clearly to a spiriferacean ancestor. The high degree of external morphological similarity makes it essential to consider evidence compiled from studies of morphology, ontogeny and shell microstructure. A revised taxonomy assigning the Thecideidina, Thecospiracea and Bactryniidae to the Spiriferida is proposed. Thecideidine brachiopods have the dubious distinction not only of having previously been assigned to three articulate orders within the Brachiopoda but to the Mollusca also. Although some of the important early contributions must be mentioned, this paper is not an attempt to chronicle the many publications on thecideidine brachiopods. Much of the content of studies prior to 1965 was synthesized by Pajaud: his monograph (Pajaud 1970) is especially useful in providing fuller details of earlier works on the taxonomy and systematics of the group. Williams’s (1973) review of the origin of the thecideidines provided an important summary of previous opinions about the systematic position of the group. The essential purpose of this review, therefore, is to consider the status of arguments advanced in the early 1970s in the light of further developments of the past twenty years. Although obviously interconnected, investigations relevant to this paper can broadly be grouped into studies aimed at the elucidation of taxonomy, ancestry and evolution using morphological, ontogenetic or microstructural evidence. This arrangement broadly reflects the chronological order of the major landmarks in the study of the group and thus serves as a useful framework around which to order the content of the paper. THECIDEIDINE TAXONOMY Although shells had been described earlier (Faujas 1798; Schlotheim 1813), the earliest use of ‘thecidean’ apparently dates from the introduction of the term by Defrance (1822) to refer to the distinctive morphological characters of representatives of a newly designated genus Thecidea. It was, however, another eighteen years before the group emerged (Gray 1840) as a taxon of family rank containing six species, all assigned to Thecidea. After the establishment of the Thecideidae Gray, 1840, a series of classic descriptive works followed, notably those of Davidson (1851, 1854, 1874, 1876), D’Orbigny (1847), Eudes-Deslongchamps (1853), Lacaze-Duthiers (1861) and Moore (1854). These studies, although varying the generic spelling between Thecidea, Thecidium and Thecideum, introduced many new species. At about this time the expanding family attracted the interest of systematists (e.g. Dali 1870). Munier-Chalmas (1880, 1887) began the task of dilTerentiating the taxa at generic level and, in view of the small size of many of the representatives. I Palaeontology, Vol. 33, Part 1, 1990. pp. I75-I9I.| © The Palaeontological Association 176 PALAEONTOLOGY, VOLUME 33 it is perhaps a tribute to his observation and interpretative skills that the genera he created remain unmodified to the present day. At higher level, Waagen (1882) assigned the thecideans to the Terebratulacea, whereas Schuchert (1896) transferred them to the Strophomenacea. With only minor adjustments (Rollier 1915; Thomson 1915), this steady state persisted for half a century. However, beginning with the combination (Termier and Termier 1949) of thecideids and lyttoniids into a new superfamily Thecideacea, the next twenty years saw more changes than had been witnessed during the previous one hundred and fifty. TEXT-FIG. 1. Drawings to show the internal morphology, typical of the principal types of thecideidine brachial valve. A, monoseptate form cf. Moorellina with blade-like median septum, brachial bridge broken, b, median septum with well-developed sinus cf. PraelacazeUa. c, polyseptate form cf. Mimikonstantia, bridge broken, lateral septa damaged, d-f, transverse sections, x-y, through brachial valves A-C respectively. Abbreviations: rt. /•./., auxiliary resorption lobe; b., border; hr., bridge; h.c., brachial cavity; b.l., brachial lobe; b.t., brachial tubercle; c.p., cardinal process; cl.s., dental socket;/., flange; i.s.r., inner socket ridge ; y., jugum; l.a.s., lateral adductor muscle scar; l.g., lophophore groove; l.s., lateral septum; m.s., median septum; r., ramulus; s., sinus; s.p.r., sub peripheral rim; s.r., serration rib; /., tubercle; v.c., visceral cavity. Scale bar represents 0-5 mm. Elliott (1948) recognized two groups of thecideidines : (i) a thecideiform group, in which numerous septa (text-fig. Ic, f), supporting a multilobed lophophore, arose either from the valve margin or as lateral branches of a median septum; (ii) a thecidelliniform group, characterized by a single median septum and a bilobed lophophore (text-fig. 1a). In the latter group, Elliott (1948, p. 26) detected the incomplete record of a continuous line of descent, Thecidella (Lias) — Bifolium (Cretaceous) — Thecidellina (Recent). Subsequently, Elliott (1953) assigned all the monoseptate forms to the new subfamily Thecidellininae. He conceded that the classification of the heterochronous thecideiform branches was more difficult, as most of the stocks, with the exception of Eolacazella — Lacazella, were divergent. He recognized that Davidsonella was technically thecidelliniform, but argued that the very long brachial lobes showed the same functional attainment as the later thecideiform ptycholophe and included the genus, together with Eudesella, Thecidiopsis, Thecidea, Eolacazella, Lacazella and Vermicidotliecidea, in the subfamily Thecideinae. Later, Elliott (1958) considered that intermediate forms between thecideidines and strophomenides or terebratulides were unlikely to be identified and proposed the elevation of the Thecideacea, as understood by Termier and Termier (1949), to subordinal rank to emphasize their distinctiveness. Pajaud’s ( 1 970) criticism of Elliott’s classification was unjustified. His assertion that Elliott regarded Thecidella, Moorellina, Bifolium and Thecidellina as the trunk of a phyletic tree from which the ptycholophous forms branched was incorrect. Elliott clearly understood that two plexi of descent were involved, one rectilinear (Thecidellininae) and the other discontinuous (Thecideinae). BAKER: THECIDEIDINE BRACHIOPODS 177 Although incorrect in several respects, e.g. the systematic position of Thecidella and Lacazella, Elliott’s reasoning has stood the test of time (and an enlarged database) remarkably well and in it can be identified the root of all modern classification of the group. However, an important oversight on Elliott’s part was his failure to recognize that in genera such as Thecidella the median septum was divided by the development of a trough-like depression (text-fig. 1b) or sinus (Baker and Laurie 1978, p. 564). Backhaus equated the median septum with ascending lophophore supporting elements. He coined the terms (Backhaus 1959, p. 12) apparatus ascendens apertus to describe the divided median septum (text-fig. 1e) as in Lacazella, and apparatus ascendens clausiis to describe the blade-like median septum (text-fig. Id) as in Thecidellina. Backhaus, like Elliott, also perceived two groups (= tribes of Backhaus 1959) but, unlike Elliott, noted that the Praelacazella species showed a progressive passage between the Thecidella species of the Lias and the Lacazella species of the Tertiary. A further problem stemmed from Elliott’s (1948) account of the ontogeny of Bifolium farmgdonense . Subsequent studies (Baker and Laurie 1978) showed that Elliott had unknowingly combined the ontogeny of the thecidellinin B. faringdonense with that of the lacazellin Neothecidella parviserrata. The earlier failure to recognize this mixed assemblage had led Backhaus (1959), Pajaud (1966t?) and Smirnova and Pajaud (1968) to assign forms with an ‘open’ ascending apparatus to Bifolium. Worse, the authors shared a common view that a juvenile ‘closed’ ascending apparatus (i.e. undivided median septum) could give way to an open (divided median septum) form in the adult, thus paving the way for important misconceptions about thecideidine phylogeny. It was only later that Pajaud and Smirnova (1971) showed that the form of the median septum is established very early and remains unaltered throughout ontogeny (text-fig. 2). They removed the ‘open’ Bifolium lacazelliforme types to Praelacazella. Having established a marker (Pajaud 1963), and having published a series of short communications on preliminary notes and problems (Pajaud 1966u, 19666), new genera (Pajaud 1966c, 1966(7; Pajaud and Glazewski 1964; Pajaud and Patrulius 1964; Termier, Termier and Pajaud 1966), mutation (Pajaud 1968u), neoteny (Pajaud 19686), and ontogeny (Smirnova and Pajaud 1968), Pajaud then embarked on the monumental task of monographing all known thecideidines. The monograph (Pajaud 1970) continued to include the Permian Cooperina Termier, Termier and Pajaud, 1966, despite the clear indication (Cooper and Grant 1969, p. 18) that Cooperina should be regarded as a productidine assigned to the Strophalosiacea. In view of the, then current, controversy surrounding thecideidine ancestry, one can sympathize with the authors’ eagerness (Termier, Termier and Pajaud 1966) to reveal to the scientific community the first Palaeozoic thecideidine with such obvious links with the Strophomenida. Cooper and Grant were not to be denied, however, and had, understandably, retained the finest specimens in their own collections. The description and illustration of this material (Cooper and Grant 1975) closely following the systematic dismantling (Dagis 1973, p. 367; Williams 1973, p. 470) of Pajaud’s (1970) arguments, dispelled any further doubt about the genus’s productidine identity, although reassigned (Cooper and Grant 1975) to the Aulostegacea. Pajaud’s reluctance to abandon the identification of Cooperina as a thecideidine (Pajaud 1974; Patrulius and Pajaud 1974) succeeded only in casting a shadow over a study (Pajaud 1970) which remains the most comprehensive statement on matters other than thecideidine shell microstructure. Pajaud’s (1970) proposal to elevate the Thecideidina to a taxon of ordinal rank received little support from other workers, probably because it would have contributed nothing towards a better understanding of thecideidine systematics, but would merely have frozen the uncertainty within the broader framework of ordinal relationships. Pajaud (1970, p. 74) constructed a taxonomy based on the philosophy that the recognition of genera should be based principally on the morphology of the brachial system. By comparing plans of the brachial structure he was able to identify what appeared to constitute natural groups with a high level of internal coherence. The scheme worked reasonably well up to subfamily level, but encountered difficulties when the phyletic relationship between subfamilies was considered. The principal source of the problems lay in the belief that polyseptate genera such as Eudesella arose from monoseptate genera such as Moorellina by simple mutation, that monoseptate juveniles of Boscpietella-iy^t somehow metamorphosed into polyseptate adults of Thecidiopsis-iy'pQ, and that Thecidellina-\ike 12 PAL 178 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. Drawings to show juvenile and adult phases in the brachial valve ontogeny of representative species of monoseptate (apparatus ascendens clausiis), b-j, and polyseptate (apparatus ascendens apertus), K.-S, thecideidine genera, a, locational diagram, b, c, Moorellina granulosa (Moore), d, e, Pachytnoorellina dundriensis (Rollier). F, G, Rioultina ornata (Moore). H, J, Bifoliuin faringdonense (Davidson). K, L. Thecidella rustica (Davidson), m, n, Neothecidella parviserrata Baker and Laurie. P, Q. Mimikonstantia sculpta Baker and Elston. R, s, Thecidiopsis tetragona (Roemer). Abbreviations: a.rd., auxiliary resorption lobe; h.c., brachial cavity; i.b.c., intra-brachial cavity; s., sinus. Scale bar represents 0-5 mm. morphology arose neotenously from Backhausina. It is now accepted (O. Nekvasilova, pers. comm. 1985) that the monoseptate specimens figured as juveniles of Thecidiopsis hohemica (Nekvasilova 1964, pi. 11, figs. 1^) are in fact wrongly assigned to that species. Conversely, the detailed study of the ontogeny of Thecidiopsis tetragona revealed (Smirnova 1969) that brachial valves as little as 2-5 mm long already showed the development of lateral septa (text-fig. If) and confirmed that Pajaud was incorrect in the belief that the early ontogenetic development of Thecidiopsis passed through an auriform (entire median septum and auriform brachial lobes) phase. Although critical of the systematic schemes of Elliott and Backhaus, two thinly disguised groups ( = clans) also emerged in Pajaud’s (1970) classification. Pajaud’s idea of a loose grouping into six subfamilies failed to appeal to Smirnova (1972, 1984) who, on the basis of detailed studies of ontogeny and comparative morphology, decided (Smirnova 1984, p. 109) that the fundamental shortcoming of all BAKER: THECIDEIDINE BRACHIOPODS 179 existing classifications was the underrating of the importance of the type of lophophore. Smirnova, in resurrecting Elliott's (1965) family grouping, ascertained that the thecidellinid forms with their schizolophous lophophore (text-fig. 2a-j) differed sharply in ontogeny from the thecideid forms (text-fig. 2k-s) with their ptycholophous or thecidiolophous (Pajaud 1970) lophophore. Members of the Thecidellinidae Elliott, 1958 are characterized by a schizolophous lophophore and an undivided median septum which remained a stable character through the Mesozoic to Cenozoic history of the group. The taxon embraces the subfamilies Moorellininae Pajaud, 1966 and Thecidellininae Elliott, 1958. In the Thecideidae Gray, 1840, a complexly divided thecidiolophous lophophore is inherent, but it always originates as a concave plate dividing into separate lobes. The taxon embraces the subfamilies Thecideinae Gray, 1840, Davidsonellinae Pajaud, 1968, and Lacazellinae Backhaus, 1959. Arising from the more precise understanding of the various ontogenies, Smirnova (1972) recommended a certain amount of inter-subfamilial rearrangement of taxa, namely that EiideseUa and Koustantia be transferred from the Moorellininae, as understood by Pajaud (1970), to the Thecideinae, and that the monoseptate genera Bifoliwn and Bosquetella be removed from the Thecideinae. Bosquetella was reassigned to the Moorellininae, Bifoliwn, along with Rioultinu and Elliottina, being reassigned to the Thecidellininae. The only other important subsequent move (Smirnova 1984) was the transfer of Glazewskia from the Lacazellinae to the Thecideinae. Discovery of new sources of material enabled the restudy (Baker 1983) of the minute Enallothecidea pygniaea (Moore). The absence of a median septum and the incomplete sub- peripheral rim placed the genus close to the most primitive thecideidines. Other discoveries (Baker and Elston 1984) revealed that Eudesella was not the sole Early Jurassic polyseptate representative. Detailed study of the ontogeny of Mimikonstantia Baker and Elston, 1984 showed that, in its basic expression, it differed little from the development pattern described for Thecidiopsis by Smirnova (1969). The conclusion that Mimikonstantia was also related to Koustantia greatly strengthened Smirnova’s (1972) argument for reassignment of Eudesella and Koustantia to the Thecideinae. THECIDEIDINE ANCESTRY A major problem posed by the study of the Thecideidina has always concerned the affinities of the suborder. The group has been identified as having originated from various articulate groups, with the Terebratulida (Elliott 1965), Spiriferida (Williams 1968, 1973) and, in particular, several strophomenide taxa (Rudwick 1968; Baker 1970; Pajaud 1970; Grant 1972; Dagis 1973) emerging as the main contenders. Clearly, their neotenous origin and the influence of palingenetic and proterogenetic processes have clouded the image of the line of thecideidine descent. Ideas in the late 1960s and early 1970s had been hampered by uncertainty regarding the Triassic spire-bearing Thecospira. Williams (1968, p. 48) revised an earlier opinion that the genus should be identified as a davidsoniacean (Williams 1953, p. 12) in favour of reassignment to the Spiriferida. This proposal was contested by Rudwick (1968, 1970), Baker (1970) and Dagis (1973), who favoured a strophomenide affinity for the genus. Williams, however, remained unconvinced by any counter arguments and reiterated (Williams 1973, p. 475) his earlier view that Thecospira should be regarded as a spiriferide, a view supported by Mackinnon (1974). Additionally, it was shown by Elolder (1975) that the brachial supports of complex thecideids were anatomically different from strophalosiacean productidines. In rejecting the views of other authors, Williams (1973, p. 441) declared that any attempt to identify the ancestor of the thecideidines must take account of the shell microstructure and the likelihood that the thecideidines arose neotenously or paedomorphically. After refuting the arguments in favour of a strophomenide ancestor, Williams was less certain about choosing between spiriferides and terebratulides as the probable ancestral stock. However, after citing cementation in Thecospira and recalling differences in the structure of the mantle edge in thecideidines and terebratulides, Williams (1973, p. 475) finally emerged in favour of descent from a punctate spiriferide. The earlier identification of the Permian Cooperina as a thecideidine (Termier, Termier and Pajaud 1966) had sparked oft' similar controversy (Cooper and Grant 1969; Dagis 1973) before its productidine identity was finally confirmed (Cooper and Grant 1975). 180 PALAEONTOLOGY, VOLUME 33 An important development, however, had been the demonstration (Dagis 1973) that the shell microstructure of thecospirids is practically the same as that of early thecideidines and that, in the Thecospiracea, the hungarithecids were probably ancestral to the thecideidines. Dagis perceived a similarity between plectambonitacean and thecospiracean shell microstructure but, because of the time gap, rejected any idea of a genetic link. However, after evaluation of other characters he considered that, although morphological resemblance to the davidsoniaceans may have been largely convergent, the thecospiraceans were descended from strophomenide ancestors which, by lineage therefore, were also ancestral to thecideidines. Williams's (1973) meticulously detailed defence of his argument for spiriferide affinity left counter arguments difficult to sustain. The single weakness in Williams’s thesis was his inability to demonstrate, other than at general level, any microstructure in the shells of spiriferides and thecideidines which represented unequivocal evidence of a genetic relationship. Because of the profound influence of neoteny in thecideidine evolution (Elliott 1953; Pajaud 1970; Williams 1973, 1984), I concluded (Baker 1984) that study of the characters of potentially ancestral adult shells would be unlikely to provide the key to the identification of the thecideidine ancestor. Also, within the Thecideidina the effects of neotenous suppression of some shell secretion processes were so dramatic that it became difficult to identify characters of phylogenetic significance. Eventually, cyrtomatodont teeth (Jaanusson 1971), secondary fibrous shell, and tubercles came to be recognized as characters which were sufficiently stable to survive all but the most drastic changes affecting the shell microstructure of later representatives of the group. It was perceived that such stable characters should be a feature of at least juveniles of the ancestral stock. The results of examination of the shell microstructure of juvenile representatives of potentially ancestral stock (Baker 1984) confirmed Williams’s (1973) opinion and were positive enough to make his arguments in favour of spiriferide affinity virtually unassailable. The circle was closed, therefore, on the earlier demonstration (Dagis 1973) that the shell microstructure of thecospiraceans was almost identical with the microstructure of the majority of Early Jurassic thecideidines, since the confirmation of a genetic link between the thecideidines and spiriferides (Baker 1984) also established a genetic relationship for the thecospiraceans and spiriferides. The identification of the thecideidine tubercle as a structural homologue of the spiriferacean denticle pointed to a spiriferacean, rather than the suessiacean ancestor envisaged by Williams. Contrary to the opinion of Smirnova (1984, p. 1 15), I have never considered that the thecideidines might be descended from a terebratulide ancestor. THECIDEIDINE PHYLOGENY Elliott (1948, 1953) may be credited with the first attempt to elucidate the phylogeny of the, by then, numerous described species. Elliott envisaged, within the suborder, a palingenetic progression from a schizolophous monoseptate form to a ptycholophous polyseptate condition. He considered that the monoseptate genera like Bifolium and Thecidellina, appearing later, represented the heterochronous expression of the palingenetic trend. Backhaus (1959) rejected Elliott’s views, making the important observation that thecideidine lophophore supports developed from the two basic, apparatus ascendens clausus and apparatus ascendens apertus, patterns. On this basis, Backhaus conceived the idea of two phyletic groups (tribes). He seems to have extrapolated from a thorough knowledge of Cretaceous thecideidines to a much more tenuous understanding of pre- and post-Cretaceous representatives of the suborder. Thus, his proposed phylogeny identifying Moorellma, despite its stratigraphic range, as a juvenile Eudesella, Bosquetella as a juvenile Thecidiopsis, and Thecidellina as the juvenile of some undiscovered ptycholophous adult, was manifestly suspect. Rudwick’s study of food-gathering mechanisms (Rudwick 1968) brought the problematical Triassic genera Bactrynium Emmrich, 1855, and Thecospira Zugmayer, 1880 into the picture, introducing the concept of the ‘functional zone’ as a measure of the phyletic relationship between the Thecideacea, Davidsoniacea and Lyttoniacea. Rudwick (1968, p. 353) included Bactrynium in the Thecideacea, arguing that the strophic hinge and articulation was quite unlike the aberrant BAKER. THECIDEIDINE BRACHIOPODS structures of the lyttoniaceans and that the lophophore lay in lobed grooves like those found in polyseptate thecideidines. Work on a very large collection of 2700 Early Cretaceous thecideidine shells from Valanginian and Hauterivian bioherms in the Crimea (Smirnova 1969) enabled detailed study of the ontogeny of species of the thecidellinid genera Bifoliiim and Bosquetella, and the thecideid genera Thecidiopsis and Praelacazella. Close similarities in development suggested (Smirnova 1969, p. 64) that Moorellina was probably ancestral to Bosquetella. More important, the earliest stages in the development of both Praelacazella and Thecidiopsis were characterized by a broad median septum with a central sinus. In Praelacazella the structure was retained throughout all growth stages, whereas in Thecidiopsis it was quickly translated, though the development of a split in the median septum near the hinge-line, into the precursor of a polyseptate condition. The distinctiveness of the thecidellinid and thecideid ontogenetic development patterns provided interesting confirmation of Backhaus’s (1959) idea of two phyletic groups. Additionally, within the groups, lineages showing parallel development were beginning to appear. Moorellina — Bosquetella and Elliottina — Bi- folium— Thecidellina lineages were identified in the Thecidellinidae, and Thecidella — Prae- lacazella— Lacazella and Eudesella — Thecidiopsis — Glazewskia lineages were recognized in the Thecideidae. Smirnova’s important contribution probably appeared too late to allow Pajaud (1970) the opportunity for comment. He was unenthusiastic about Rudwick’s (1968) views on thecideidine phyletic relations, in which he saw the resurrection of Elliott’s ideas. However, Pajaud (1970, p. 79) did feel able to support the location of Bactrynium near to the Thecideidae in the Thecideacea. He appreciated the phyletic significance of Backhaus’s two tribes and, with some modification of Backhaus’s ideas, introduced a new phylogeny based on the concept of Lacazella and Thecidellina ‘clans’, embracing five subfamilies. Despite the weight of evidence against it, the Permian Cooperina was still identified at the base of the main trunk from which the two branches separated. Pajaud considered that the Lacazella clan, including the Davisonellinae and Lacazellinae, arose from a Davidsonella-type ancestor and that the evolution from the Lias to Recent followed a rectilinear pattern. The Thecidellina clan which included the Moorellininae, Thecideinae and Thecidellininae was also believed to have arisen from a Davidsonella-iypQ ancestor. The evolution of the latter group was more complicated, and mutation was invoked (Pajaud 1968u) to explain the recurrent appearance of Jurassic and Cretaceous polyseptate forms, whereas neoteny was invoked (Pajaud 19686) to account for the return to a monoseptate condition in the Tertiary. Through his comprehensive survey of characteristically thecideidine features, Williams (1973, p. 466) was able to identify a range of unifying characters in strong contrast to the profound evolutionary changes suffered by the shell microstructure and lophophore supports. In consideration of the ptycholophe as opposed to the thecidiolophe (Pajaud 1970, p. 33), Williams concluded that both conditions could be regarded as equipotential adult elaborations of an immature schizolophe and may, therefore, have recurred many times in thecideidine history. Unfortunately, although ignoring Cooperina, his chart showing the phyletic variation in thecideidine shell microstructure (Williams 1973, fig. 100, p. 468) was based on Pajaud’s phylogenetic reconstruction (Pajaud 1970, fig. 31, p. 82) and was, therefore, constrained by the same misconceptions which characterized Pajaud’s phylogeny. In a more recent contribution, Smirnova (1984) has shown that the early juveniles of the Lower Cretaceous Bifolium mica are very similar to adults of the Middle Jurassic Rioultina and Elliottina which, in their adult morphology approach Bifolium and Thecidellina. On this basis, Smirnova detected a genetic relationship and rejected Pajaud’s (1970) view that Thecidellina arose neotenously from a Cretaceous thecidein which, instead of having a blade-like median septum, would be characterised by a concave median septum, quickly opening out to form lobes. It is now clear (author’s unpublished work) that the adult Bifolium faringdonense has canopied brachial lobes like Thecidellina, offering further support for Smirnova’s view. Although the development of the polyseptate condition from a concave triangular plate (Smirnova 1984) is a unifying character in the Thecideidae, representatives of the Thecideinae are characterized by lobes which are differentiated in a lateral direction, whereas representatives of the Lacazellinae are characterized by lobes which 182 PALAEONTOLOGY, VOLUME 33 MIDDLE JURASSIC 1 1 LOWER CRETACEOUS 1 1 TERTIARY RECENT schizolophous, monoseptate 1 1 ptycholophous, essentially polyseptate exterior brachial valve interior pedicle valve exterior granular calcite shell fibrous secondary shell acicular calcite shell schizolophous, monoseptate C E exterior brachial valve interior 2^^ pedicle valve exterior G TEXT-HG. 3. Diagrammatic representation of the clironological succession of the principal events in the evolution of thecideidine shell structure. Posterior and postero-lateral sectors of valve, left, anterior and antero-lateral sectors of valve, right. Horizontal lines indicate continuous layer, diagonal lines indicate restricted distribution, a, Moorellina, continuous fibrous secondary shell layer in both valves. B, Mimikonstantia, partially suppressed fibrous secondary shell, c, Pacliymoorellina, partially suppressed fibrous secondary shell and introduction of acicular calcite tracts. Thecidiopsis, D, more strongly suppressed fibrous secondary shell, acicular calcite well-developed, e-g, fibrous secondary shell vestigial, restricted mainly to teeth and inner socket ridges, acicular calcite often well-represented but the shell is composed principally of granular calcite. e, Bifoliim, f, Lacazella. G, Thecidellina. No lineage is implied. are differentiated in a front to rear direction. In this respect, Pajaud’s (1970) assignment of Bactryniwn to a position near the Thecideidae seems reasonable. The two subfamilies of the Thecideidae appear to have showed parallel evolutionary development, expressed through a small number of lophophore lobes in the Lower Jurassic, becoming increasingly complex during the Upper Jurassic and Lower Cretaceous, with the maximum complexity reached simultaneously in both groups during the Upper Cretaceous. Changes at the Cretaceous/Palaeocene boundary led to the extinction of the specialized forms. The survival of thecideidines was attributed (Smirnova 1984) to the existence of ‘primitive’ forms able to adapt to life in various conditions and continue the existence of simply-organized genera in modern basins. Smirnova also studied the evolution of thecideidine shell microstructure and reached essentially the same conclusion as Williams (1973, 1984), namely that the continuous fibrous secondary lining (text-fig. 3a) characteristic of early Jurassic shells had been reduced to vestigial patches on the teeth and sockets (text-fig. 3e) by the Early Cretaceous. Williams (1973, p. 469) envisaged that suppression of the fibrous secondary layer was accomplished rapidly, and placed the onset of the trend in late Jurassic time. Williams (1984, p. 739) regarded the suppression of fibrous secondary shell as an expression of neoteny. Smirnova (1984) was able to identify three stages to the suppression process which affected the brachial and pedicle valves differentially. Jurassic representatives were found to have a fibrous secondary layer in both valves. In the Lower Cretaceous, Berriasian to Hauterivian species showed a reduction (text- BAKER: THECIDEIDINE BRACHIOPODS 1H3 fig. 3d) in the fibrous secondary layer in the brachial valve. The suppression of fibrous secondary shell was then extended to the pedicle valve also (text-fig. 3e-g), so that from the Barreniian to the Present both valves were characterized by the complete reduction of the fibrous secondary layer, the end product being a granular calcite shell, with or without acicular calcite aggregations, in which the occurrence of fibrous secondary shell, if present at all, was restricted to the teeth and inner socket ridges. Study of a newly-discovered basal Middle Jurassic species (Baker and Elston 1984) cast some doubt on opinions about the timing of the onset of the evolutionary changes in shell microstructure. The shell microstructure of Mimikoiistantia revealed that the onset of the neotenous suppression of fibrous secondary shell could be traced back at least as far as the beginning of the Middle Jurassic. Also, the shell microstructure resembles that of the Lower Cretaceous species Thecidiopsis tetragona and T. lata. The conclusion that Mimikoiistantia and Thecidiopsis were phylogenetically linked (Baker and Elston 1984) offered indirect support for Smirnova’s belief in a phylogenetic link between Jurassic and Cretaceous polyseptate forms. Consideration of the mechanical requirement for the multiplication of lateral septa suggested a possible sequence (Baker and Elston 1984, fig. 5, p. 790) in the development of a thecidiolophous form from a ptycholophous ancestor and thus, by implication, linked Mimikonstaiitia, Thecidiopsis, Backliaiisiiia and Tliecidea, again supporting Smirnova’s (1984, fig. 64, p. 110) ideas. The evidence indicated that Pajaud’s (1970) tentative derivation of Konstantia and Thecidiopsis from monoseptate Rioidtina stock was no longer tenable as Mimikoiistantia pre-dates Rioidtina. Recent studies (Baker 1989) also enable the origin of Thecidiopsis to be traced back to basal Middle Jurassic roots, although analysis of the shell microstructure of a newly designated genus indicates that both Mimikoiistantia and Eudesella are slightly diverged from the main line of descent. DISCUSSION Ideas about thecideidine taxonomy, origin and phylogeny have been developed through consideration of morphological, ontogenetic and, more recently, shell microstructural evidence. The review would be incomplete without consideration of the value of the contribution made by each of these aspects. The value of morphology Because of their external morphological similarity and because their abnormally wide gape facilitates post-mortem liberation of brachial valves, the taxonomy of thecideidine brachiopods has traditionally relied heavily on the internal morphology of separated pedicle and brachial valves. Particular attention has always been paid to the skeletal supports for the lophophore. Although septa are usually sufficiently robust to provide reliable evidence of their location and general form, other structures, especially brachial and interbrachial lobes, are often fragile and finely sculptured and are almost invariably broken in separated brachial valves. Reservations about the reliability of the evidence as seen in separated brachial valves were expressed as early as the beginning of this century (Upton 1905, p. 91). Nekvasilova, after careful study of the Lower Turonian Thecidiopsis (T.) hohemica imperfecta, reached the conclusion (Nekvasilova 1967, p. 130) that, in thecideidines, determination based on the so-called brachial ridges as seen in detached brachial valves was quite inadequate, since such ridges may be a relic of structures whose complete shape may only be studied through sectioning of complete shells. She remarked on the similarity between the reconstructed brachial lobes of the Lower Turonian specimens and those of the Recent Thecidellina hlochmaimi Dali. Work on the ontogenetic development of Moorelliiia granulosa (Moore) showed that the brachial tubercles were the broken remains of much more elaborate structures (Baker 1969) which overarched the intra-brachial cavities. Perforate canopies have subsequently been identified (research in progress) in the Aalenian M. dtihia and the Aptian Bifoliiim faringdonense . Clearly therefore, the development of the brachial lobes in the Moorellina — Thecidellina plexus of descent, as envisaged by Smirnova (1984), conformed to a more coherent pattern than became apparent from the study of separated brachial valves. Similarly, the radially disposed septa (Baker and Elston 184 PALAEONTOLOGY, VOLUME 33 1984) of the polyseptate Mimikonstantia sculpta are almost invariably broken in separated brachial valves. Over-reliance on morphological evidence has been responsible for some of the most heated controversy surrounding the probable origin and systematic position of the Thecideidina. Termier, Termier and Pajaud’s (1966) introduction of Cooperina as the first Palaeozoic thecideidine is a case in point. The authors’ selective recognition of ‘thecideidine’ morphological characteristics prompted the omission of the obviously non-thecideidine aspects of the genus. Their arguments were systematically dismantled by Cooper and Grant (1969, p. 18) and finally refuted (Cooper and Grant 1975) through the description and figuring of the superb Cooperina specimens to which they had access. A similar selective approach was required to enable Grant (1972) to ‘force’ the conclusion (see Holder 1975) of a genetic relationship between thecideidines and strophalosiacean productidines such as Falafer. Rudwick (1968) recognized that in polyseptate thecideidines such as Thecidiopsis the primary lophophore lobe is that furthest from the mid-line in a postero-lateral position, and that in Bactrynium the primary lobe is located close to the median septum in an anterior position. In this respect the antero-postero extension of the lophophore lobes of Bactrynium resembles the development pattern of the lacazellins rather than the lateral extension pattern seen in the thecideins. Rudwick’s error lay in the assumption that, in polyseptate thecideidines, growth without shell resorption was able to translate the juvenile arrangement into the adult complement of lateral septa. Study of the development of septa in relation to shell growth in polyseptate thecideacean species (Baker and Elston 1984) has shown that as the brachial valves increase in size, the zones of maximum growth acceleration (Baker 1970) become increasingly separated from the median septum. Therefore, in a shell which is increasing in width more rapidly than it is increasing in length, lateral migration of septa relative to the principal growth vectors is required, and a precisely- controlled process of shell accretion and resorption is necessary for this to be accomplished. On the other hand, if the brachial valve is increasing in length more rapidly than it is increasing in width, as in Bactrynium, zones of maximum growth acceleration will remain aligned essentially parallel with the median septum, and the antero-postero development pattern described by Rudwick (1968) will represent the optimum for the circumstances appertaining. This implies, therefore, that septal (and lophophore lobe) development patterns are a strategic response to shell growth requirements and present no real obstacle to the postulation of a genetic relationship between Bactrynium and thecideidines. The observation, in addition to the more general morphological considerations, that the shell succession in Bactrynium included a normally developed fibrous secondary layer and impersistent tubercles (Williams 1973, p. 475) further substantiated the view that the genus might reasonably be included in the Thecideidina. Although Rudwick was able to accept convergent evolution as the explanation for the morphological similarity between Bactrynium and lyttoniacean genera, he firmly resisted the idea (Rudwick 1968, p. 329) that the morphological similarity between Tliecospira and the davidsoniaceans could be similarly explained. By concentrating on cementation, lobed brachial grooves, pseudodeltidium and absence of pedicle foramen as prime evidence, Rudwick was able to assign the thecideidines (including Bactrynium) to the Davidsoniacea along with Thecospira. Williams (1973) was critical of Rudwick’s selective approach and, as subsequent studies have shown (Baker 1984), convergent evolution in reef-associated faunas is probably common. Also, brachial grooves are not characteristic of all early thecideidines (Baker and Elston 1984). Cementation and a pseudodeltidium are also not as exclusive as Rudwick believed (Cooper and Grant 1975). Additionally, there is some evidence that very early thecideidines may have possessed a transient apical pedicle foramen (Baker 1983) and that, initially at least, the pseudodeltidium was located apically in the delthyrium. The accumulated evidence, reinforced by the discovery that the thecideidine tubercle should be regarded as homologous with the spiriferide denticle rather than the strophomenide taleola (Baker 1984) led to abandonment of the notion that the ancestors of the thecideidines were to be found among the strophomenides. BAKER: THECIDEIDINE BRACHIOPODS 185 The value of ontogeny Although some of the palingenetic and neotenous processes thought to have been operative during thecideidine evolution subsequently proved to be questionable, it is probable that as many issues have been resolved through the interpretation of ontogeny as have been clouded by mis- interpretation of morphology. It is unfortunate, therefore, that one of the earliest detailed studies (Elliott 1948) combined events in the ontogenies of Bifoliwn (text-fig. 2h, J) and NeothecideUa (text-fig. 2m, n) into a single sequence, resulting in considerable taxonomic confusion until the error was noted and rectified (Baker and Laurie 1978). Similarly, it is now clear that the rioultinid (undivided median septum and auriform brachial lobes) brachial valves (text-fig. 2g) figured as juveniles of Thecidiopsis bohemica (Nekvasilova 1964, pi. II, figs. 1-4) were wrongly identified (O. Nekvasilova, pers. comm. 1985). After a detailed study of available ontogenetic records, Pajaud (1970) concluded that the evolution of the lophophore supports followed a more complex pattern than the palingenetic (Elliott 1953, p. 698) or neotenously-induced (Backhaus 1959, p. 77) progressions previously favoured. The real key to understanding phyletic relationships, however, was provided by Smirnova’s (1969, 1984) correlation of the development of the lophophore supports in the various groups. The very large collections of material from the Crimea enabled considerable progress to be made in the detailed interpretation of the ontogeny (Smirnova 1969) of Thecidiopsis, Praelacazella, Bosquetella and Bifolium. The discovery that the ontogeny of Thecidiopsis (text-fig. 2r, s) did not pass through a rioultinid phase (Smirnova 1969) came too late, however, to prevent Pajaud (1970) from deriving Thecidellina neotenously from Thecidiopsis stock. But, it was not until much later, that Smirnova formally rejected (Smirnova 1984) the citation of monoseptal forms such as Bosquetella as initial stages in the evolution of Thecidiopsis tetragona (Backhaus 1959) and T. bohemica (Nekvasilova 1964). Smirnova considered thecidellinin ontogenetic changes as being exemplified by the ontogenetic development of Bifolium mica. The ontogeny of Bifolium was traced via genetic links with Elliottina and Thecidellina and Bosquetella with Moorellina, thereby vindicating Elliott’s remarkably perceptive observation (Elliott 1948) that although some species may have become extinct there is little doubt that most of the monoseptate, schizolophous species known, represent the broken record of a continuous series of thecidellinins from the Mesozoic to the present day. With allowance for a tachygenetic element, the correlation of the mechanics of shell growth with septal development pattern (Baker and Elston 1984) reveals a remarkable similarity between the early ontogeny of Mimikonstantia and Thecidiopsis, even down to the thickened triangular structure from which the lateral septa develop (text-fig. 2p, q). According to Smirnova (1984) this triangular element could be correlated with the primitive divided median septum of some davidsonellins which was considered to link the Davidsonellinae with the Thecideinae and Lacazellinae in the Thecideidae. Evidence from ontogenetic studies has proved useful in the resolution of other taxonomic problems. Because Pajaud (19666) had queried the validity of the species designation, Barczyk (1970, p. 653) was uncertain about the status of specimens of Moorellina septata (Moore) from the Upper Jurassic of Poland. Study of juveniles of a newly- designated Middle Jurassic genus (Baker 1989) shows that the specimens of M. septata queried by Pajaud as juveniles of M. dundriensis (Pajaud 19666) in no way correspond to the latter. The value of shell microstructure The use of shell microstructure as an indicator of thecideidine relationships entered the arena relatively late, and for a variety of reasons (Grant 1972, p. 244; Williams 1973, p. 441 ), its potential value continued to be underestimated (Smirnova 1984). With the exception of three investigations of Lacazella shell microstructure (Davidson 1887, Oehlert 1887; Thomson 1927), thecideidine shell microstructure remained virtually unknown until Elliott’s (1953, 1955) studies. The early studies were hampered by the difficulty of preparing sections and certain resolution deficiencies of optical microscopes. Also, by chance selection for study (Williams 1955) of a species in which fibrous secondary shell had been almost completely suppressed, it was not discovered that, in the majority of thecideidines, the shell microstructure differs in brachial and pedicle valves. Although Elliott 186 PALAEONTOLOGY, VOLUME 33 (1953, p. 695) observed ‘the dark-coloured elements of typical pseudopunctation are seen against clearer lamellar shell, but never so clearly as good examples of pseudopunctation in certain Palaeozoic brachiopods’, neither the resolution of the instruments nor the current state of knowledge enabled the observers to differentiate between taleolae and tubercle cores. Resolution problems were subsequently eliminated (Williams 1968) with the advent of scanning electron microscopy facilities, and the problems of section preparation were considerably alleviated by the use of cold-setting transparent embedding resins and the development of techniques (Baker 1970) for serially sectioning the very small shells at intervals of approximately 20 //m. The demonstration that the shell (text-fig. 3a) of the Middle Jurassic Moorellina granulosa was lined by a continuous layer of fibrous secondary shell (Baker 1970) caused Williams (1973) to modify his earlier conclusion (Williams 1968) that the shell microstructure (text-fig. 3f) of Lacazella was typical of the thecideidine model. Appreciation of the critical importance of precise location and orientation of section (Baker 1970) no doubt informed the ensuing study, of thecospiracean shell microstructure (Dagis 1973) the very detailed investigation of the Recent thecideidines Thecidellina barret ti (Davidson) and Lacazella niediterranea (Risso) by Williams (1973), and studies of the shell microstructure of Lower Cretaceous species (Smirnova 1979, 1984). Even after the thecideidine structures had been identified as the cores of tubercles (Baker 1970), the pseudopunctation signal was still so strong that it was easy to continue to regard them as being homologous with taleolae and, through association, perceive a relationship with strophomenides. Although concluding that thecospiraceans were closer to thecideidines than any other group, Dagis’s opinion (1973) about the systematic position of the thecospiraceans and their relationship with thecideidines was also coloured by the strength of the historical association of both groups with the Strophomenida. It is unfortunate that, after demonstrating the genetic relationship, he then went on to parallel in thecospiraceans my misconception (Baker 1970) of the thecideidine tubercle as structurally homologous with the strophomenide taleola. Williams ( 1973), prompted by the discovery of fibrous secondary shell in early Middle Jurassic representatives of the group, undertook a critical survey of the shell microstructure of the majority of the described thecideidine genera. This study established that a continuous lining of fibrous secondary shell was the standard Lower to Middle Jurassic condition (text-fig. 3a) and the indications were that the onset of its neotenous suppression was not elTected until the Upper Jurassic or Early Cretaceous (text-fig. 3d, e). Smirnova (1979), in pursuit of Williams’s (1973) idea of a Late Jurassic to Early Cretaceous date for fibrous secondary shell suppression, studied the shell microstructure of three Lower Cretaceous species, Thecidiopsis tetragona (Roemer), Thecidiopsis lata Smirnova and Praelacazella valangiensis (de Loriol) in an attempt to find out if the change was abrupt or gradual. The differences in the microstructure of T. tetragona and T. lata and the similarity of P. valangiensis to both, helped to convince her that microstructure was of limited taxonomic value, but useful in dating the important steps in the structural evolution of the shell. This conviction was strengthened by a later study (Smirnova 1984) in which she deduced that, although the structural changes of the shell proceeded steadily, the disappearance of the fibrous secondary layer was quick and probably occurred in the first half of the Lower Cretaceous. Smirnova concluded that the degree of plasticity of microstructural changes in the thecideidine shell rendered shell microstructure almost valueless in the comparison of Mesozoic thecideidines and Triassic thecospirids with ancient Palaeozoic groups, and virtually’ useless for drawing conclusions about their hypothetical relationship. She felt that shell microstructure was only of value when its use was confined to the study of successive groups in time. The situation was envisaged as being most complicated from the Upper Cretaceous to Recent, as a consequence of the relative uniformity of shell microstructure (Smirnova 1984), when, the extreme difficulty of determining the relationship between layers of granular and acicular calcite rendered it of little use for systematization purposes. Work by Baker and Elston (1984) on newly-discovered Middle Jurassic material demonstrated, contrary to the view of Williams (1973) and Smirnova ( 1979, 1984), that although the fibrous secondary layer was still continuous, the trend towards its suppression (text-fig. 3b) was already established by the early Middle Jurassic. Subsequent investigations (Baker 1989) revealed that Miinikonstantia sculpta was not the only species aflfected. BAKER: THECIDEIDINE BRACHIOPODS 187 In the contemporaneous Pachymoorellina dundriensis, fibrous secondary shell had already disappeared (text-fig. 3c) from the anterior and antero-lateral sectors of both valves. CONCLUSIONS The conjectural nature of some of the identifications of brachial lobe morphology has created taxonomic and phylogenetic problems. Until about the mid-1960s, interpretation was made more difficult because the size of most thecideacean representatives of the suborder fell outside the optimum resolution range for both conventional photography and reflected light photomicro- graphy, so that the quality of plate figures was modest by current standards. Study of the detail of thecideidine morphology, and especially their shell microstructure was revolutionized by the advent of the scanning electron microscope. Considerable emphasis has been placed on the relative value of shell microstructure as a distinctive character. Although general shell fabric may not be diagnostic, structures such as tubercle cores, from a unifying point of view, show a remarkable continuity of expression throughout the history of the group. Also, it was only the demonstration that the thecideidine tubercle was probably the structural homologue of the spiriferacean denticle, rather than the strophomenide taleola, that enabled the link between thecideidines and strophomenides to be severed with confidence. The thecideidines represent a group whose classification is aided by the recognition of genera showing a high level of coherence from a genetic relationship point of view. Clearly, much remains to be discovered about the stratigraphic and geographical range of this still comparatively little-known group. Also, much more work is required on the detailed mapping of the various successions of shell microstructure, which is emerging as a taxonomic indicator of far greater potential than has hitherto been appreciated. However, in my opinion, sufficient information is now available to enable a reliable taxonomic and phylogenetic framework to be established. RECOMMENDATIONS Revised classification Excluding the aulostegacean Cooperina, for the reasons given by Cooper and Grant (1969, 1975), Dagis (1973) and Williams (1973), the Thecideidina, as understood by Pajaud (1970, pp. 82-83) included twenty two genera distributed among five subfamilies. Apart from some rearrangement of genera and the resurrection of older family grouping (Smirnova 1984), Pajaud's framework has persisted, largely unaltered, to the present time. A further, subfamily Enallothecideinae Baker, 1983 and four additional genera Enallothecidea Baker, 1983, Mimikonstantia Baker and Elston, 1984, Pajaudina Logan, 1988 and Pachymoorellina Baker, 1989 may now be added. The general organization and shell microstructure of Enallothecidea is similar to that of early juvenile moorellinins prior to the full differentiation of the median septum. It is proposed, therefore, to reassign the Enallothecideinae to the Thecidellinidae. In view of the perceived relationship between Bactryniiini and thecideidines, and the weight of evidence about the genetic relationship between thecospiraceans and thecideidines and their affinity with spiriferides a revised classification is proposed as follows: Order spiriferida Waagen, 1883 Suborder thecideidina Elliott, 1958 Superfamily thecospiroidea Bittner, 1890 Eamily thecospiridae Bittner, 1890 Genus thecospira Zugmayer, 1880 Family thecospirellidae Dagis, 1973 Genus thecospirella Bittner, 1900 188 PALAEONTOLOGY, VOLUME 33 Family hungarithecidae Dagis, 1973 Genus hungaritheca Dagis, 1973 Superfamily thecideoidea Gray, 1840 Family thecidellinidae Elliott, 1958 Subfamily enallothecideinae Baker, 1983 Genus enallothecidea Baker, 1983 Subfamily moorellininae Pajaud, 1966 Genera moorellina Elliott, 1953; pachymoorellina Baker, 1989; bosquetella Smirnova, 1969 Subfamily thecidellininae Elliott, 1953 Genera rioultina Pajaud, 1966; bifolium Elliott, 1948; thecidellina Thomson, 1915 Family bactryniidae Williams, 1965 Genus bactrynium Emmrich, 1855 Family thecideidae Gray, 1840 Subfamily davidsonellinae Pajaud, 1966 Genera davidsonella Munier-Chalmas, 1880; agerinella Patrulius, 1964 Subfamily lacazellinae Backhaus, 1959 Genera thecidella Oehlert, 1887; neothecidella Pajaud, 1970; parabifolium Pajaud, 1966; praelacazella Smirnova, 1969; vermiculothecidea Elliott, 1953; danella Pajaud, 1966; eolacazella Elliott, 1953; lacazella Munier-Chalmas, 1880; pajaudina Logan, 1988 Subfamily thecideinae Gray, 1840 Genera eudesella Munier-Chalmas, 1880; mimikonstantia Baker and Elston, 1984; KONSTANTiA Pajaud, 1970; thecidiopsis Oehlert, 1887; backhausina Pajaud, 1966; parathecidea Backhaus, 1959; thecidea Defrance, 1822; glazewskia Pajaud, 1964 Elliot tina Pajaud, 1963 is not included in the revised classification. The genus was never strongly placed and was soon relegated to sub-generic rank by Pajaud himself (Pajaud 19666). The whole concept of the erection of a genus on the basis of the width of the ventral interarea is highly suspect in attached forms, where the morphology of the pedicle valve is so strongly influenced by the size and shape of the surface to which attachment is effected. It is rather surprising, therefore, that a new subfamily Elliottininae was proposed (Pajaud and Smirnova 1971) to include the resurrected Elliottina and also Rioultina and Bifolium. The Elliottininae was subsequently abandoned (Smirnova 1984), with Rioultina, Bifolium, and Elliottina being restored to the Thecidellininae. The superfamily suffix -ACEA enjoys such a wide usage in current terminology that it would have caused considerable confusion to have changed it in the body of the paper. Flowever, in compliance with the ICZN recommendation 29a (Ride et al. 1985, p. 55) -OIDEA is added to the superfamily stem as the preferred suffix in the revised taxonomy. REFERENCES backhaus, e. 1959. Monographic der cretacischen Thecideidae (Brachiopoden). Mitteihmgen aus dem Geologischen Staatsinstitut in Hamburg, 28, 5-90. baker, p. g. 1969. The ontogeny of the Ihecideacean brachiopod Moorellina granulosa (Moore) from the Middle Jurassic of England. Palaeontology, 12, 388-399. 1970. The growth and shell microstructure of the thecideacean brachiopod Moorellina granulosa (Moore) from the Middle Jurassic of England. Palaeontology, 13, 76-99. 1983. The diminutive thecideidine brachiopod Enallothecidea pygmaea (Moore) from the Middle Jurassic of England. Palaeontology, 26, 663-669. 1984. New evidence of a spiriferide ancestor for the Thecideidina (Brachiopoda). Palaeontology, 27, 857-866. BAKER: THECIDEIDINE BRACHIOPODS 189 1989. Evaluation of a thecideidine brachiopod from the Middle Jurassic of the Cotswolds, England. Palaeontology, 32, 55-68. and ELSTON, d. g. 1984. A new polyseptate thecideacean brachiopod from the Middle Jurassic of the Cotswolds, England. Palaeontology, 27, 777-791. and LAURIE, K. 1978. Revision of the Aptian thecideidine brachiopods of the Faringdon Sponge Gravels. Palaeontology, 21, 555-570. BARCZYK, w. 1970. Some representatives of the family Thecideidae (Brachiopoda) from the Upper Jurassic of Poland. Acta geologica polonica, 20, 647-655. BITTNER, A. 1890. Die Brachiopoden der alpinen Trias. Ahluindlungen der Geologischen Bundesanstalt, Wien. 14, 1-325. 1900. Brachiopoden aus der Trias der Bakonyer Waldes. Residtate der Wissenschaftlichen Erforsclnmg des Balatonsees. Wien. Paldontologie Ergdnzung 7, Pt 1, 1-70. COOPER, G. A. and grant, r. e. 1969. New Permian brachiopods from West Texas. Smithsonian Contributions to Paleobiology, 1, 17-18. 1975. Permian brachiopods of West Texas, III. Smithsonian Contributions to Paleobiology, 19, 795-1921. DAGis, A. s. 1973. Ultrastructure of thecospirid shells and their position in brachiopod systematics. Paleontological Journal, 6, 359-369. DALE, w. H. 1870. A revision of the Terebratulidac and the Lingulidae. American Journal of Conchologv, 6, 88-168. DAVIDSON, T. 1851-1855. Monograph of the British fossil Brachiopoda: Tertiary, Cretaceous, oolitic and liasic species (British oolitic and liasic Brachiopoda, 1851; British Cretaceous Brachiopoda, 1852; General introduction. On the classification of Brachiopoda, 1854; Appendix 1854; Supplementary appendix, 1855). Palaeontographical Society [Monograph] . 1, 1-376. 1874-1882. Monograph of the British fossil Brachiopoda: Tertiary, Cretaceous, Jurassic, Permian and Carboniferous supplements (Supplement to the British Cretaceous Brachiopoda, 1874; Supplement to the British Cretaceous Brachiopoda, 1876). Palaeontographical Society [Monograph] . 4, 1-383. 1886-1888. A monograph of Recent Brachiopoda; Series 2. Transactions of the Linnean Society of London. 4, 1-248. defrance, m. 1822. Thecidea. 114. In ferussac, a. de (1821-1822) Tableaux systematiques des animaux mollusces classes en families naturelles, suivis cf un prodrome general pour tons les molhisces terrestres on fluviatiles, vivants on fossiles. Volume 27. Arthus Bertrand, Paris, 157 pp. ELLIOTT, G. F. 1948. Palingenesis in Thecidea (Brachiopoda). Annals and Magazine of Natural History, (12), 1, 1-30. 1953. The classification of the thecidean brachiopods. Annals and Magazine of Natural History, (12), 6, 693-701. 1955. Shell-structure of thecidean brachiopods. Nature, London, 175, 1124. 1958. Classification of thecidean brachiopods. Journal of Paleontology, 32, 373. 1965. Order uncertain - Thecideidina. H857-H862. In moore, r. c. (ed ). Treatise on invertebrate paleontology. Part H. Brachiopoda. Geological Society of America and University of Kansas Press, New York and Lawrence, Kansas, 927 pp. EMMRiCH, H. 1855. Notiz fiber den Alpenkalk der Lienzer Gegend. Jahrbuch der Geologischen Bundesanstalt, Wien, 6, 444-M50. EUDES-DESLONGCHAMPS, E. 1853. Memoire sur les genres Leptaena et Thecidea des terrains Jurassiques du Calvados. Memoires de la Societe linneenne de Normandie, 9, 213-250. FAUJAS DE SAINT-FOND, B. 1798. Histoire naturelle de la Montague Saint-Pierre de Maestricht. H. J. Jansen, Paris, 263 pp. GRANT, R. E. 1972. The lophophore and feeding mechanism of the Productidina (Brachiopoda). Journal of Paleontology, 46, 213-249. GRAY, J. E. 1 840. Brachiopoda. 151. In, Synopsis of the contents of the British Museum, (42nd edn). G. Woodfall, London, 370 pp. HOLDER, H. 1975. Uber Geschichte und Stand der Thecideen-Forschung (Thecideida, Brachiopoda articulata). Mitteilungen aus dem Geologischen und paldontologischen Institut der Universitdt Hamburg, 44, 133-152. JAANUS.SON, V. 1971. Evolution of the brachiopod hinge. Smithsonianian Contributions to Paleobiology, 3, 33-J6. LACAZE-DUTHIERS, H. 1861 . Histoire naturelle des brachiopodes vivants de la Mediterranee. Annales des sciences naturelles, Zoologie, 15, 259-330. 190 PALAEONTOLOGY, VOLUME 33 LOGAN, A. 1988. A new thecideid genus and species (Brachiopoda, Recent) from the southeast North Atlantic. Journal of Paleontology. 62, 546-551. MACKINNON, D. I. 1974. The shell structure of spiriferide Brachiopoda. Bulletin of the British Museum (Natural History) (Geology), 25, 189-261. MOORE, c. 1854. On new Brachiopoda from the Inferior Oolite of Dundry, etc. Proceedings of the Somersetshire Archaeological and Matured History Society, 5, 107-128. MUNiER-CHALMAS, E. p. 1880. Note sommaire sur les genres de la famille des Thecideidae. Bulletin de la Societe geologique de France. (3), 8, 279-280. 1887. Note sommaire sur les genres Koninckella. Cadomella. Thecidella and Lacazella. Bulletin de la Societe geologique de France, (3), 15, 34-40. NEKVASiLOVA, o. 1964. Thecidcidae (Brachiopoda) der bohmischen Kreide. Shornik geologikych ved Praze, Paleontologie, 3, 119-162. 1967. Thecidiopsis ( Thecidiopsis ) hohemica imperfecta n. subsp. (Brachiopoda) from the Upper Cretaceous of Bohemia. Shornik geologikych ved Praze. Paleontologie, 9, 115-136. OEHLERT, D. p. 1887. Brachiopodes, (Appendix). 1208-1209. In fischer, p. Manuel de conchyliologie et de paleontologie. F. Savy, Paris, 1369 pp. ORBIGNY, A. d’ 1847. Paleotitologie fran(,'cuse . Terrains Cretaces 4: Brachiopoda. Arthus Bertrand, Paris, 390 pp. PAJAUD, D. 1963. Note sur les Thecidcidae (brachiopodes) jurassiques. Bulletin de la Societe geologicpie de France. (7), 5, 995-1000. 1966a. Note preliminaire a la classification des thecidees (brachiopodes). Bulletin de la SociHe geologique de France. (1). 8, 615-620. 1966/). Problemes relatifs a la determination des especes chez le Moorellininae (Thecideidae, brachiopodes). Bulletin de la Societe geologique de France. (7), 8, 630-637. 1966c. Deux genres nouveaux de thecidees (brachiopodes): Parahifolium (P. priscum n.sp.) et Danella (D. fragilis n.sp.). Compte rendu sommaire des seances de la Societe geologique de France. 5, 70-71. 1966c/. Sur deux nouvelles formes de thecidees (brachiopodes), dont le genre nouveau Backhausina (B. rugosa (d’Orb.)). Compte rendu sommaire des seances de la Societe geologicpie de France. 6, 124-125. 1968a. Importance des mutations chez les thecidees (brachiopodes). Compte rendu hehdomadaire des seances de 1' Academic des sciences. Paris. Series D. 267, 43^5. 19686. La neotenie chez les thecidees (brachiopodes). Compte rendu hebdomadaire des seances de ! Academic des sciences. Paris. Series D. 267, 156-159. 1970. Monographic des thecidees (brachiopodes). Memoires de la Societe geologique de France. 112, 1-349. 1974. Biostratigraphie des thecidees (brachiopodes) donnees recentes ou inedites. Geobios. 7, 323-344. and GLAZEWSKi, K. 1964. Remarques sur les Thecideidae (brachiopodes) tertiaires. Sur une nouvelle forme du Miocene rhodanien: Glazewskia demarequi nov. gen. nov. sp. Bulletin de la Societe geologicpie de France. (7), 6, 258-261. and PATRULius, D. 1964. Sur la presence de trois especes de Thecideidae (brachiopodes) en Roumanie. Bulletin de la Societe geologique de France. (7), 6, 579-585. and SMIRNOVA, T. N. 1971. Systematics of brachiopods of the suborder Thecideidina. Paleontological Journal. 5, 191-196. PATRULIUS, D. 1964. See pajaud, d. and patrulius, d. 1964. and pajaud, d. 1974. Presence de thecidees et de thecospires dans les depots detritiques du Norien des Monts Apuseni (Roumanie). Documents des Lahoratoires de geologic de la Facidte des Sciences de Lyon. 62, 129-135. ride, w. d. l., sabrosky, c. w., bernardi, g. and melville, r. v. (eds.). 1985. International code of zoological nomenclature (3rd edn). University of California Press, London, Berkeley and Los Angeles, 338 pp. ROLLiER, L. 1915. Synopsis des spirobranches (brachiopodes) Jurassiques Celto-souabes. .Abhandlungen der Schweizerischen paldontologischen Gesellschaft. 41, 46-56. RUDWiCK. M. j. s. 1968. The feeding mechanisms and affinities of the Triassic brachiopods Thecospirci Zugmayer and Bactryniiim Emmrich. Palaeontology. 11, 329-360. 1970. Living and fossil brachiopods. Hutchinson, London, 199 pp. SCHLOTHEIM, B. VON 1813. Beitrage zur Naturgeschichte der Versteinerungen in geognostischer Hinsicht. In LEONHARD, c. C. Taschenhucli fiir die gesammte Mineralogie 1. Joh, Christ. Hermannschen Buchhandung., Frankfurt am Main, 134 pp. SCHUCHERT, c. 1896. Brachiopoda. 291-343. In zittel, k. a. Textbook of pcdaeontology 1 (1st edn). Macmillan and Co, London, 706 pp. BAKER: THECIDEIDINE BRACHIOPODS 191 SMIRNOVA, T. N. 1969. Ontogeny and phylogeny of Early Cretaceous brachiopods of the suborder Thecideidina. Paleontological Journal, 3, 64-78. 1972. Thecideidina. 1 13-127. In, Rannemelovye hrakhiopody Kryina i Severnogo Kavkaza. Nauka Press, Moscow, 140 pp. [In Russian]. 1979. Shell microstructure in Early Cretaceous thecidean brachiopods. Paleontological Journal, 13, 339-344. 1984. [Early Cretaceous brachiopods {morphology, systematics, phylogeny and their significance in biostratigraphy and palaeozoogeography )]. Nauka Press, Moscow, 199 pp. [In Russian]. and PAJAUD, D. 1968. Contribution a la connaissance des thecidees (brachiopodes) du Cretace d’Europe. Bulletin de la Societe geologitpie de France, 10, 138 147. TERMtER, H. and TERMIER G. 1949. Sur la classification des brachiopodes. Bulletin de la Societe cfhistoire naturelle de P Afritpw du Nord. 40, 51-63. and PAJAUD, D. 1966. Decouverte d’une thccidee dans le Pcrmien du Texas. Coinpte rendu hebdoinadaire des seances de P Academie des sciences, Paris, Series D, 263, 332-335. THOMSON, J. A. 1915. On a new genus and species of the Thecidiinae. Geological Magazine, 2, 461-464. — - 1927. Brachiopod morphology and genera ( Recent and Tertiary). New Zealand Board of Sciences and Art. Manual No. 7. Dominion Museum, Wellington, 338 pp. UPTON, c. 1905. Some Cotteswold Brachiopoda. Proceedings of the Naturalists' Field Club, 15, 82-92. WAAGEN, w. 1882-1885. Salt Range fossils. Productus-limestone fossils (fasc. 2). Brachiopoda. Memoirs of the Geological Survey of India. Palaeontologia indica, (13), 1, 329-770. WILI.IAMS, A. 1953. The classification of the strophomenoid brachiopods. Journal of the Washington Academy of Sciences, 43, 1-13. 1955. Shell-structure of the brachiopod Facazella mediterraneum (Risso). Nature, Fondon, 175, 1123. 1968. Evolution of the shell structure of articulate brachiopods. Special Papers in Palaeontology, 2, 1-55. 1973. The secretion and structural evolution of the shell of thecideidine brachiopods. Philosophical Transactions of the Royal Society of Fondon, Series B, 264, 439^78. 1984. Eophophorates. 728-745. In bereiter-hahn, j., matoltsy, a. g. and Richards, k. s. (eds.). Biology of the integument. I Invertebrates. Springer-Verlag, Berlin, Heidelberg, New York, Tokyo, 800 pp. ZUGMAYER, H. 1880. EIntersuchungen liber rhlitische Brachiopodcn. Beitrdge zur Paldontologie and Geologic Ostereich-Ungarns and des Orients, 1, 1^2. P. G. BAKER Department of Geology Derbyshire College of Higher Education Manuscript submitted 5 October 1988 Kedleston Road, Revised manuscript submitted 21 February 1989 Derby DE3 1GB, UK " A ■ ; i ■ 0: 1 I TEUTHID CEPHALOPODS FROM THE LOWER JURASSIC OF YORKSHIRE by PETER DOYLE Abstract. Specimens of four teuthid species, Loligosepia aalensis (Zieten), Jeletzkyteuthis agassizi (Deslongchamps), Teudopsis schuebleri (Quenstedt) and T. siihcostata (Munster), are described from the Whitby Mudstone Formation (Toarcian) of North Yorkshire. Jeletzkyteuthis nom. nov. is erected to replace Loliginites Quenstedt which is nomcnclaturally invalid. This teuthid fauna is found to be similar to that of southern England, Normandy and southern Germany. The initial phase of fossil teuthid research took place during the first half of the nineteenth century. Most of the species’ names still in use were first applied by workers such as Zieten (1830-32), Deslongchamps (1835), Buckland (1836), Quenstedt (1839, 1845^9), d'Orbigny (1842, 1845), and Munster (1843). Many of these authors directly compared their fossil specimens with the ‘pens’ of recent squid, and the genus Loligo was used for some of the named species (e.g. Zieten 1830-32; Buckland 1836). Recently, interest in fossil teuthids has been rekindled by a number of works from the Tfibingen school (e.g. Reitner and Engeser 1981, 1982; Engeser and Reitner 1983, 1985, 1986; Riegraf and Ziigel 1984). However, despite the encouragement provided by Jeletzky (1966, p. 42) and Donovan (1977), little has been written on British teuthids in recent years. In this paper, I redescribe the teuthid specimens discussed by Simpson (1855, 1884) and Tate and Blake (1876) (see also Crick 1922, p. 288) from Yorkshire, figuring the type specimens of Simpson’s species for the first time. Well-preserved teuthid specimens are rare due to their fragility and so new material is not readily available, requiring the re-examination of older collections. Fortunately, both Simpson and Blake carefully localized their material, and matrix lithologies allow reasonable stratigraphical control. Interest in fossil teuthids in Britain probably began with Buckland (1836). He figured several specimens from the Lower Jurassic of Dorset which he assigned to Loligo, suggesting that they were similar to Zieten’s (1830-32) Loligo aalensis. Voltz (1840) was later to reassign these forms to a new species, Belopeltis hucklandi. This is one of the oldest known species of Loligosepia (Sinemurian- Pliensbachian : Jeletzky 1966, p. 42), although earlier forms are known from Germany (Reitner 1978; Reitner and Engeser 1981). The majority of British nineteenth century discoveries were from southern England, as shown by Morris (1854, p. 303), Moore (1867) and Smithe (1877). Smithe (1877, p. 400) described a new species, Beloteuthis glevensis, from the sands overlying the Marlstone Rock Bed in North Gloucestershire which may represent the basal Toarcian (see Howarth 1980). Engeser and Phillips (1986) and Engeser (1988) have interpreted this as a species of Teudopsis. A revision of teuthids from the Fish and Insect beds (Toarcian) of southwest England was given by Crick (1921) who recorded two species, Geoteuthis agassizi (Deslongchamps) and Teuthopsis [5/c] hnmelli [^/c] (Deslongchamps). These species may now be assigned to Loligosepia aalensis (Zieten) and Teudopsis bunelii Deslongchamps, respectively. A third species from the same horizon in Gloucestershire, Teudopsis siihcostata (Miinster), may be added to this list (BMNH C.5252). Moore (1867, p. 183) indicated that these teuthids were extremely common in southern England, and was able to suggest that they formed the stable diet of the contemporary ichthyosaurs, an idea supported by Pollard (1968) from the examination of ichthyosaur gastric masses. Simpson (1855, 1884) was the first author to describe teuthids from Yorkshire. He erected four new species. Sepia ohtusalis, S. cuspidata, S. haustruni and S. inconiposita. Blake (in Tate and Blake IPalaeonloIogy, Vol. 33. Part 1, 1990, pp. 193-207. | © The Palaeontological Association 13 PAL 33 194 PALAEONTOLOGY, VOLUME 33 1876) re-examined Simpson’s material, assigning 5. cuspidata to the genus Teudopsis, and considering the nominal species S. ohtusalis and S. haustnim junior synonyms of Geoteuthis coriaceus (Quenstedt). The types of these species are preserved in the Whitby Museum (WM). The fourth species, S incomposita, is not mentioned in the WM catalogue, but Blake {in Tate and Blake 1876, p. 257) suggested that these specimens were in the York Museum collections. In any case, this species is actually based on fragments of fish vertebrae (Tate and Blake 1 876, p. 257) and is therefore not dealt with here. Blake’s specimens of this fish {Gyros tens miriahilis) are certainly in the York Museum (Pyrah 1979, p. 417). Blake added two more species with his own Beloteuthis leckenbyi (now in the Sedgwick Museum), and a single specimen of Beloteuthis subcostata Munster (now in the British Museum (Natural History)) (see Crick 1922). These Yorkshire specimens were all collected from the Jet Rock (= Upper Lias Division 6 of Simpson, 1884) which is equivalent to the Jet Rock sensu stricto {exaratum Subzone) as defined by Howarth (1962, p. 386), and the lower part of the Jet Rock Member of Powell (1984). This original published information (Simpson 1855, 1884; Blake in Tate and Blake, 1876) is confirmed by the examination of the fossil matrices. SYSTEMATIC PALAEONTOLOGY All specimens are housed in Whitby Museum (WM), the British Museum (Natural History) (BMNH) and the Sedgwick Museum, Cambridge (SM). The descriptions given below are based purely on Yorkshire material and so little can be added to the original diagnoses. Most of the type specimens of the species described below are preserved in the Geologisches und Palaontologisches Institiit, Tubingen (GPIT), West Germany. The terms used below are discussed by Jeletzky (1966) (text-fig. 1). Recently, Engeser (1986, 1988) and Berthold and Engeser (1987) have suggested a revised taxonomy of the Coleoidea based on phylogenetic systematics. Jeletzky’s (1966) less controversial classification is used below, however, as discussion of higher taxa is beyond the scope of this study, and as taxa of low rank remain unaffected. The synonymy lists follow the convention of Matthews (1973). Subclass COLEOIDEA Bather, 1888 Order teuthida Naef, 1916 Suborder loligosepiina Jeletzky, 1965 Eamily loligosepiidae Van Regteren Altena, 1949 (= Belopeltidae Naef, 1921) Genus loligosepia Quenstedt, 1839 (= Belopeltis Voltz, 1840, (objective synonym); Palaeosepia Theodori, 1844, Geoteuthis Miinster, 1843 (subjective synonyms)) Type species. Loligo aalensis Zieten, 1832, by monotypy. Diagnosis. Large Loligosepiidae with gladius comprising relatively broad, smooth median field with central well-defined median keel and parallel striations; narrow hyperbolar fields with markedly anterior-concave growth lines; wings of parabolic form and with pointed anterior ends, generally extending for not less than one-third of the length of the median field. Large ink-sac present. Range. This genus is known primarily from the Lower Jurassic (Sinemurian-Toarcian) of Europe (Jeletzky 1966; Reitner and Engeser 1981 ). However, Reitner (1978) has reported this genus from the Upper Triassic of Bavaria, although the specimens illustrated by him have markedly reduced wings. Remarks. On the basis of the reconstructions given by Naef (1922, p. 129), specimens of Loligosepia may be distinguished from the similar genus Parabelopeltis Naef which has smaller hyperbolar zones with less anteriorly concave growth lines, and from Jeletzkyteuthis nom. nov. which has a narrower, elongate gladius with a narrow median field (text-fig. 1). DOYLE: JURASSIC TEUTHID CEPHALOPODS 195 MK HZ I MA LA MK HZ I MA LA MK MK TEXT-FIG. 1. Diagramatic reconstructions of the four teuthid species under discussion, not to scale. Abbreviations: HZ, hyperbolar zone; LA, lateral asymptote; MA, median asymptote; MF, median field; MK, median keel; W, wing. 13-2 196 PALAEONTOLOGY, VOLUME 33 Loligosepia aalensis (Zieten, 1832) Text-figs. 1, 2a-c, 3b, c *. 1832 Loligo Aalensis Zieten, p. 34, pi. XXV, fig. 4. .1832 Loligo Bollensis Zieten, p. 34, pi. XXV, fig. 5. [Lectotype, here selected. Lower Toarcian, Posidonienschiefer, Boll, Southern Germany. Original in the GPIT]. non 1832 Loligo Bollensis Zieten, p. 49, pi. XXXVII, fig. 1. [=Teudopsis sclmehleri (Quenstedt)]. . 1849 Lolipnites Bollensis Zieten; Quenstedt, p. 508, pi. 32, figs. 11-13; pi. 33, figs. 1-5. [pi. 33, fig. 1 is Zieten’s pi. XXV, fig. 5 specimen refigured]. V. 1855 Sepia haustrum Simpson, p. 21. V. 1876 Geoteiuhis coriaceus Quenstedt; Blake (in Tate and Blake), p. 313, pi. IV, fig. 1. V. 1884 Sepia haustrum Simpson; Simpson, p. 19. 1920 Geoteuthis Bollensis (Schiibler) Zieten; Bulow-Trummer, p. 253. [Full early synonymy], . 1921 Geoteuthis agassizi (Deslongchamps) ; Crick, p. 251, pi. a. . 1922 Geoteuthis coriaceus (Quenstedt); Crick, p. 288. . 1922 Belopeltis Aalensis (Zieten); Naef, p. 125, text-fig. Alb. 1949 Loligosepia aalensis (Zieten); Van Regteren Altena, p. 58. . 1981 Loligosepia aalensis (Zieten); Reitner and Engeser, p. 427, text-fig. 2. 1984 Loligosepia aalensis (Zieten); Riegraf et al., p. 37. . 1988 Loligosepia aalensis (Schiibler in Zieten); Engeser, p. 9. [Full synonymy]. Type specimen. Holotype, the original of Zieten (1832, p. 34, pi. XXV, fig. 4), from the Lower Toarcian of Aalen, southern Germany. Original in the GPIT. Material. Nine specimens: WM 2, 3, 5; BMNH C.651, C.2699, C.2698, C.37529, C.46828 and C. 12047 (the original of Tate and Blake 1876, pi. IV, fig. 1), all from the Jet Rock Member, Whitby Mudstone Formation (falciferum Zone), of Whitby, North Yorkshire. Notes on Simpson's material. Simpson’s (1855) nominal species Sepia haustrum is a junior subjective synonym of Loligosepia aalensis. Two specimens are available in the Whitby Museum labelled Sepia haustrum in Simpson’s handwriting, namely WM 2 and 3 (text-fig. 2a, b). Specimen WM 2 is here designated lectotype (text-fig. 2b). Both specimens are preserved in concretions and are from Simpson’s Upper Lias 6 division, equivalent to the Jet Rock sensu stricto of Howarth (1962). Diagnoses. See Quenstedt (1849, p. 508; ' Loliginites Bollensis'), Naef (1922, p. 25; 'Belopeltis aalensis') and Reitner and Engeser (1981, p. 427). Description. Several specimens of this species have been recovered from the Toarcian of Yorkshire. The majority (BMNH C.651, C.2698, C.2699, C. 12047, C.46828) are crushed and generally unrepresentative of the actual in-life gladius. However, two of the specimens (WM 5, text-fig. 2c; BMNH C.37529; text-fig. 3b, c) preserved in concretions are more representative of the original form of the shell. When flattened, the gladius has a rounded, obtuse posterior with fan-like anterior (text-fig. 2b). In concretions, the posterior is acute with the wings extended into the vertical plane to form a small conus (e.g. BMNH C.37529; text-fig. 3b). The shell comprises a large (length 140 mm max.) gladius with a regular anteriorly-diverging, broad, median field with an apical angle of 18°. The median field is separated from the narrow hyperbolar zones by sharply- defined median asymptotes which are marked by a narrow, well-defined groove which expands anteriorly. Transverse growth lines are discernible on the lateral parts of the median field. A sharp median keel, bounded by grooves of an equivalent width, is accommodated in the central third of the median field. This third displays no growth lines, but has striations parallel to the keel. The hyperbolar zones are each up to one-third of the width of median field, and display anterior-concave growth lines bounded by sharp asymptotes. The wings are parabolic in shape with correspondingly-formed growth lines. These are deflected sharply to the posterior at the lateral asymptotes. The wings are present along most of the preserved length of the specimens. Remarks. Reitner and Engeser (1981) have recently discussed the differential diagnosis of this species within the genus Loligosepia. It can be distinguished from the coexisting species DOYLE: JURASSIC TEUTHID CEPHALOPODS 197 TEXT-HG. 2. Loligosepia aalensis (Zietcn), ventral views xl. a. WM 3, paralectotype of Sepia luiustnon Simpson, n, WM 2, lectotype of Sepia haustrum Simpson, c, WM5. 198 PALAEONTOLOGY, VOLUME 33 Jeletzkyteuthis agassizi (Deslongchamps) by its broad median field and well-developed median keel. Riegraf (1987) recently described a large isolated ink-sac from southern Germany that he interpreted as belonging to Loligosepia aalemis. Genus jeletzkyteuthis nom. nov. {nom. nov. for LoUginites Quenstedt, 1849) Type species. Teudopsis agassizi Deslongchamps, 1835. Derivation of name. In memory of Dr J. A. Jeletzky, 1915-1988. Diagnosis. Large Loligosepiidae with smooth, narrow and elongate gladius comprising a narrow median field with median keel; hyperbolar zones equivalent in width to the median field with anterior-concave growth lines; wings of parabolic form similar to those of Loligosepia. Ink-sac present. Range. Lower Jurassic (Toarcian) of Europe (Naef 1922). Remarks. Jeletzkyteuthis is erected here as a replacement name for LoUginites Quenstedt, 1849, which Engeser (1988, p. 50) found was unavailable for the Principle of Priority under Article 20 of the International Code of Zoological Nomenclature, being originally used for fossil species of the Recent genus Loligo (Quenstedt 1849). Engeser (1988) further suggested that the type species of LoUginites, L. coriaceus Quenstedt ( = Teudopsis agassizi Deslongchamps), was actually a plesioteuthid of the genus Romaniteuthis Fischer and Riou, 1982. However, forms formerly attributed to the genus LoUginites have much greater affinity to the Loligosepiidae than Plesioteuthididae, having well-developed wings, and relatively broad median fields with simple keels. Romaniteuthis is distinguished by its much reduced wings and conus, and its rounded, well- developed, keel (Fischer and Riou 1982; Riegraf and Zugel 1984). In consequence Jeletzkyteuthis nom. nov. is erected here for those narrow, elongate Loligosepiidae characterized by the species Jeletzkyteuthis agassizi (Deslongchamps) (see Naef 1922, text-fig. 47 and text-fig. 1 herein). Jeletzkyteuthis agassizi (Deslongchamps, 1835) M835 . 1849 V.1855 non 1876 v.1884 . 1920 non 1921 . 1922 1984 .1988 Text-figs. 1 and 3a Teudopsis Agassizi Deslongchamps, p. 72, pi. 5, fig. 15. LoUginites coriaceus Quenstedt, p. 512, pi. 34, figs. 5-8. Sepia obtusalis Simpson, p. 20. Geoteuthis coriaceus Quenstedt; Blake (in Tate and Blake), p. 313, pi. IV, fig. 1 [ = Loligosepia aalensis]. Sepia obtusalis Simpson; Simpson, p. 18. Geoteuthis coriacea Quenstedt; Biilow-Trummer, p. 253. Geoteuthis agassizi Deslongchamps; Crick, p. 251, pi. a [= Loligosepia aalensis (Zieten)]. LoUginites coriaceus Quenstedt; Naef, p. 130, text-fig. 47c. LoUginites agassizi (Deslongchamps); Riegraf et al., p. 37. Romaniteuthis agassizi (Elides- Deslongchamps); Engeser, p. 51. [Eull synonymy]. Type specimen. Holotype, the original of Deslongchamps (1835, p. 72, pi. 5, fig. 15), Toarcian, Curcy, Normandy, northern France. It is not known whether Deslongchamps’s specimen still exists. However, Quenstedt’s (1846-49) specimens of LoUginites coriaceus are preserved in GPIT. Material. Three specimens, WM 1, BMNH 83685, C.3654, Lower Toarcian, Whitby Mudstone Formation, Jet Rock Member, (falciferum Zone), Whitby, North Yorkshire. DOYLE: JURASSIC TEUTHID CEPHALOPODS 199 Notes on Simpson s specimens. Simpson’s (1855) nominal species Sepia obtusalis, is a junior subjective synonym of Jeletzkyteuthis agassizi. Two specimens bear this name in the Whitby Museum Catalogue. Specimen WM 1 has a distinct well-preserved ink-sac, and is undoubtedly the specimen referred to by Simpson (1855, p. 20). This specimen is here designated lectotype of Sepia obtusalis (text-fig. 3a). It is preserved in bituminous shale and is undoubtedly from the Jet Rock Member. A second specimen, WM 5, bears no label, but is noted as Sepia obtusalis in the Whitby Museum catalogue. This is actually a representative of the species Loligosepia aalensis (Zieten) (text-fig. 2c; see above). Diagnoses. See Quenstedt (1849, p. 512; Loliginites coriaceus) and Naef (1922, p. 130; Loliginites coriaceus). Description. Few specimens of this species have been recovered from the Yorkshire Toarcian. The single WM specimen (text-fig. 3a) is fragmentary and worn, but comprises a posterior portion (length 50 mm) of a gladius slightly offset from a large, elongate ink-sac. The gladius fragment is flattened and narrow with a maximum preserved width of 19 mm. A central median field commences with an acute apex and remains narrow for its total length. The median field is slightly crushed and worn, but a narrow median keel can be discerned in its mid-part. The greater part of the width of the specimen is taken by the wings which are each up to three times the width of the median field. Preservation of this specimen is such that no growth lines can be discerned on the wings, and the hyperbolar zones cannot be identified, although the lateral asymptotes can be seen. Slightly oblique from the gladius is an elongate, flask-shaped ink-sac with a total length of 75 mm. Specimen BMNFI 83685 is rather better preserved, and com.prises a narrow, elongate gladius 180 mm long. The median field is largely unornamented with a weak median keel. The hyperbolar zones are so narrow in its posterior as to be barely discernable; in the anterior they display anterior concave growth lines. The wings are three times as wide as the median field, extending for at least one half of the length of the gladius, and possess parabolic growth lines. Remarks. These specimens are representative of the species Jeletzkyteuthis agassizi. Simpson’s specimen, WM 1 , is very close to Quenstedt’s ( 1 849, pi. 34, fig. 5). The specimen illustrated by Blake (in Tate and Blake 1878, pi. IV, fig. 1 : BMNH C. 12047) is more properly assigned to Loligosepia aalensis (Zieten) as it possesses a broader and more ornamented median field. Jeletzkyteuthis agassizi is clearly distinguished from Loligosepia aalensis (Zieten) by possessing a narrow, elongate and less ornamented gladius as illustrated by Naef (1922, text-fig. 47). Suborder mesoteuthina Naef, 1921 Family palaeololiginidae Naef, 1921 Genus teudopsis Deslongchamps, 1835 (= Beloteuthis Munster, 1843 (Subjective synonym)) Type species. Teudopsis bunelii Deslongchamps, 1835, by subsequent designation (Woodward 1851-56, p. 69). Diagnosis. Small to large, rhomboid to pyriform Palaeololiginidae comprising spoon-like gladius with posterior blade-like extension and rounded elliptical wings; median field restricted to well- developed keel or mid rib; hyperbolar zones developed as weak deflection of the growth lines. Range. Lower Jurassic (Toarcian) to Middle Jurassic (Callovian) of Europe (Naef 1922; Fischer and Riou 1982). The species Teudopsis brodiei Caruthers reported from the Upper Jurassic of Dorset has recently been redescribed as an indeterminate plant fragment by Engeser and Phillips (1986). Remarks. The nominal genus Beloteuthis Munster is a junior subjective synonym of Teudopsis Deslongchamps. Van Regteren Altena (1949) subsequently designated the species Loligo bollensis Zieten, 1832 as type of Beloteuthis. This is not without complication, however, as the nominal genera Geoteuthis Munster, 1843 and Palaeosepia Theodori, 1844 (junior subjective synonyms of Loligosepia Quenstedt) also share this type species. Although actually based on a different syntype 200 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 3. A, Jeletzkyteuthis agassizi (Deslongchamps), WM 1, x 1. Ventral view of gladius and ink-sac. Lectotype of Sepia obtusalis Simpson, b, c, Loligosepia aalensis (Zieten), BMNH C. 37529, x 1. b, lateral view showing wings in lateral plane, c, ventral view. of Loligo bollensis, they are junior objective synonyms of Beloteuthis. Fortunately, as already discussed, Teudopsis is the senior synonym. In addition, the lectotype of Loligo hollensis Zieten selected above is a specimen of Loligosepia aalensis (Zieten), thus making L. hollensis a junior subjective synonym of the same (see below). Teudopsis was first recorded from Britain by Moore (1867, p. 303) and from Yorkshire, by Blake {in Tate and Blake 1876, p. 314). Crick ( 1921 ) described the type species Teudopsis hnmelli [i'/c] from the ‘Saurian and Fish Bed’ (Lower Toarcian, /h/q/erwm Zone) of Alderton Hill in Gloucestershire. DOYLE: JURASSIC TEUTHID CEPHALOPODS 201 This species is unknown in Yorkshire, but a single crushed specimen has been recovered from the Lower Toarcian (falcifemm Zone) of Lincolnshire (BMNH C. 46971). Teudopsis schuebleri (Quenstedt, 1849) non 1832 .1832 *.1843 .1845 . 1845 . 1849 V. 1855 .1858 V. 1876 V.1884 1920 .1922 . 1922 1949 pars 1988 Text-figs. 1 and 4a, c Loligo Bollensis Zieten, p. 34, pi. XXV, fig. 4. [ = Loligosepia aalensis (Zieten)]. Loligo Bollensis Zieten, p. 49, pi. XXXVII, fig. 1. Loligo Schiibleri Quenstedt, p. 254. Teudopsis ampullaris Munster; d’Orbigny, p. 156, pi. 14, figs. 1, 2. Teudopsis Bollensis Zieten; d’Orbigny, p. 187, pi. 14, fig. 3. Loliginites Sclnihleri Quenstedt; Quenstedt, p 499, pi. 32, figs. 14, 15. Sepia cuspidata Simpson, p. 21. Loliginites Schiibleri Quenstedt; Quenstedt, p. 243, pi. 34, fig. 9. Teudopsis cuspidatus Simpson; Blake (in Tate and Blake), p. 314, pi. IV, fig. 3. Sepia cuspidata Simpson; Simpson, p. 19. Beloteuthis Sclnihleri Zieten; Biilow-Trummer, p. 260. [Full early synonymy]. Teudopsis cuspidatus (Simpson); Crick, p. 288. Beloteuthis Bollensis (Zieten); Naef, p. 144, text-fig. 536. Teudopsis schiibleri (Quenstedt); Van Regteren Altena, p. 60. Teudopsis hunelii Eudes-Deslongchamps; Engeser, p. 77. [Includes T. schuebleri in synonymy]. Type specimen. Lectotype, here designated, the original of Quenstedt (1849, p. 499, pi. 32, fig. 15), from the Toarcian Posidonienschiefer of Holzmaden, southern Germany. The original is housed in the GPIT. Material. Two specimens, WM 4 and SM J. 3501 3 (the original of Blake in Tate and Blake 1876, pi. IV, fig. 3), from the Whitby Mudstone Formation, Jet Rock Member {falciferum Zone), of Whitby, North Yorkshire. Notes on Simpson's specimens. Simpson’s (1855) species Sepia cuspidata is a junior subjective synonym of Teudopsis schuebleri. Two specimens exist in the Whitby Museum that bear the label Sepia cuspidata. The first, WM 4 (text-fig. 4a), bears a label in Simpson’s handwriting that reads 'Sepia cuspidata, UL, Whitby’. This specimen is preserved in bituminous shale typical of the Jet Rock Member, and Simpson (1884, p. 19) recorded this species from his Upper Lias division 6 equivalent to the Jet Rock sensu stricto of Howarth (1962). This specimen is here designated lectotype. The second specimen, WM 682 (text-fig. 4b) is also labelled Sepia cuspidata, but close examination reveals that it is actually a skull of the fish Saurorhynchus hrevirostris (Woodward) similar to that figured by Woodward (1899, text-fig. 1). Diagnoses. See Quenstedt (1849, p. 499; Loliginites Schiibleri) and Naef (1922, p. 144; Teudopsis bollensis). Description. The Whitby Museum specimen (WM 4) consists of an elongate, conical, leaf-shaped gladius 120 mm long, with the posterior-most portion missing (text-fig. 4a). The gladius is flattened, but it is clear that the conus would have been spoon-like. The wings are represented by a slight lateral expansion in the posterior area with arcuate growth lines. Hyperbolar zones bounded by indistinct asymptotes are just discernible, and the growth lines are only slightly flexuous rather than notably anterior-concave. A relatively broad median keel, slightly displaced by compression in the posterior, is present for the length of the gladius and expands anteriorly to a maximum width of 4 mm. The keel is surrounded by an anterior blade-like extension with a posterior angle of divergence of 29°. Growth lines are discernible on the anterior extension, and mirror its leading edge. The Sedgwick Museum specimen (SM J. 350 13) is a juvenile with a maximum length of 76 mm, preserved in three dimensions in a pyrite-skinned concretion (text-fig. 4c). The wings extend anteriorly for 34 mm from the spoon-like conus, and hyperbolar zones similar to the other specimen are seen on this example (text-fig. 4c). The median keel is well-developed with a relief of approx. I mm above the gladius, and it expands anteriorly to a width of F5 mm. The anterior blade-like extension consists of a platform containing the keel which has a border with a sub-relief of < 1 mm (text-fig. 4c). The anterior extension diverges posteriorly at an angle of 21°, and displays growth lines similar to those preserved on the WM example. 202 PALAEONTOLOGY, VOLUME 33 Remarks. Zieten (1832, pi. XXXVII, fig. 1) figured a specimen of this species from the Posidonienschiefer of southern Germany under the name Loligo Bollensis. Earlier in his monograph Zieten had used this name for a Loligosepia-Wkt form now considered identical to his Loligo aalensis of the same work. The latter specimen is designated lectotype of Loligo Bollensis above, leaving the former specimen, a different species, without a name. Quenstedt (1839, p. 163, footnote) was aware of these problems, and later (Quenstedt 1843, p. 254) erected the name Loliginites Schiibleri for the Teudopsis-Wkt specimen of Loligo Bollensis. Engeser (1988) considered the nominal species Teudopsis schuebleri a junior synonym of Teudopsis bimelii Deslongchamps. However, T. schuebleri can be readily distinguished from T. bunelii (syntype BMNH 74009, original of Deslongchamps 1835, pi. 3, fig. 3) by its elongate wing area, and its sharp, blade-like anterior extension which has a less regular form in T. bunelii. Both species can be distinghished from T. subcostata (Munster) in possessing a smaller form and less rhomboid shape (text-fig. 1). Teudopsis subcostata (Munster, 1843) Text-figs. 1 and 5a-c *. 1843 Beloteuthis subcostata Munster, p. 61, pi. V, fig. 2; pi. VI, fig. 2. . 1845 Beloteuthis subcostata Munster; d’Orbigny, p. 364, pi. 22, figs. 1-3. . 1849 Loliginites subcostatus Munster; Quenstedt, p. 501, pi. 32, figs. 7, 8. v. 1876 Beloteuthis subcostatus Munster; Blake (in Tate and Blake), p. 313. V. 1876 Beloteuthis Leckenbyi Blake (in Tate and Blake), p. 314, pi. IV, fig. 2. 1920 Beloteuthis subcostata Munster; Biilow-Trummer, p. 261. [Full early synonymy]. . 1922 Beloteuthis subcostata Munster; Naef, p. 142, text-fig. 53a. Type specimen. Lectotype, here designated, the original of Munster (1843, pi. 5, fig. 2) from the Toarcian Posidonienschiefer of Ohmden, southern Germany. The original of this specimen is believed to have been destroyed during the Second World War (W. Riegraf, pers. comm., 1988). However, Quenstedt’s (1846-49) specimens from the same area are available in the GPIT for neotype selection. Material. Two specimens from the Toarcian Jet Rock Member (Whitby Mudstone Formation, falciferum Zone) of the environs of Whitby, North Yorkshire. The first, BMNH C. 12046 (text-fig. 5a), was recorded by Blake (in Tate and Blake 1878, p. 314) from the Serpentinus Beds of Kettleness. Only the Jet Rock sensu stricto of Howarth (1962) (falciferum Zone, e.xaratum Subzone) is exposed on the foreshore there. The second, SM J. 35012 (text-fig. 5b, c) is the holotype of Beloteuthis Leckenbyi Blake. It was recorded by Blake (in Tate and Blake 1876, p. 314) as also from the Serpentinus Beds, and the specimen bears the label ‘Whitby’. It is preserved in a small pyrite concretion or dogger, typical of the Jet Rock sensu stricto (Howarth 1962). Diagnoses. See Munster (1843, p. 61; Beloteuthis subcostata), d’Orbigny (1845, p. 364; Beloteuthis subcostata), Quenstedt (1849, p. 501; Loliginites subcostatus) and Naef (1922, p. 142; Beloteuthis subcostata). Description. The Yorkshire material comprises two specimens. The first, BMNH C. 12046 (text-fig. 5a), is a large (length 240 mm), almost complete gladius of rhomboid shape, and is flattened. The second, SM J. 35012 (text-fig. 5b, c), is smaller (length 100 mm), probably a juvenile, and is preserved in three dimensions. The larger specimen (BMNH C. 12046) is flattened with most of the gladius preserved, but with the wings present only as impressions. The wings are represented as small lateral extensions of the rhomb, and are present for almost half of the length of the gladius. Growth lines curve posteriorly, changing direction sharply at the lateral asymptotes, becoming almost straight in the hyperbolar zones and extending anteriorly to the median keel with a divergent angle of 54°. A broad, triangular anterior extension is present, with slightly curving leading edges (text-fig. 5a). A very broad median keel, reaching 10 mm at its widest, is central to the extension, and the surface of the central portion of the gladius is ornamented by lesser ridges and depressions that gradually diverge from the posterior (text-fig. 5a). The smaller specimen (SM J. 3501 2), has a similar morphology to the larger specimen, and is obviously a juvenile of the same species. It is preserved in three dimensions, with the hyperbolar zones falling as the DOYLE: JURASSIC TEUTHID CEPHALOPODS 203 TEXT-FIG. 4. A, Teitdopsis schuehleri (Quenstedt), WM 4, x 1. Ventral view, posterior part missing. Lectotype of Sepia cuspidata Simpson, b, Saworhynchus brevirostris (Woodward), WM 682, x 1. View of skull in same orientation as 4a. This specimen was labelled Sepia cuspidata in the WM catalogue, c, Teudopsis schuehleri (Quenstedt), SM J.350I3, x 1. Ventral view. curvature between the laterally extended wings and the main body of the gladius (text-fig. 5b). The median keel has a relief of 2 mm above the rest of the gladius (text-fig. 5c). The leading edges of the anterior extension of this specimen are curved giving a spatulate appearance. This obviously misled Blake {in Tate and Blake 1876), who used this specimen as holotype of his new species Beloteuthis leckenbyi. However, that this spatulate form is an artifact of the preservation is indicated by the growth lines on the surface of the gladius which reveal the typical rhomboid form of Teudopsis suheostata (text-fig. 5b). Remarks. This species is very clearly distinguished from both Teudopsis bunelii Deslongchamps and Teudopsis schuehleri (Quenstedt) by its larger size and regular rhomboid form (text-fig. 1). 204 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 5. Teudopsis subcostata (Munster), a, BMNH C. 12056, xO-75. Ventral view, b, c, SM J. 35012, holotype of Beloteuthis leckenbyi Blake, x 1. b, ventral view, c, left lateral view. DOYLE: JURASSIC TEUTHID CEPHALOPODS 205 INTER-REGIONAL COMPARISONS Inter-regional comparisons of fossil teuthid faunas are difficult because of their fragility and hence scarcity in the fossil record. Preservation of such delicate features as ink-sacs requires immediate burial (see Lyell 1878, p. 350) and/or a lack of scavengers. Thus, many teuthid finds are from deposits renowned for their exceptionally preserved faunas (fossil Lagerstatten; see Seilacher et al. 1985 for a review). Fortunately, the Toarcian was a period of widespread anoxia in the shelf-seas of Europe (e.g. Riegraf et al. 1984) and elsewhere. Close similarities exist between the Yorkshire and southern England teuthids and those of Normandy and southern Germany (e.g. Deslongchamps 1835; d’Orbigny 1842, 1845; Quenstedt 1846^9; Riegraf et al. 1984, p. 36) in these adjacent shelf seas. Local differences do occur however, with, for example, the apparent absence of Teudopsis hunelii from Yorkshire, and similarly the absence of Jeletzkyteuthis agassizi from southern England. However, these are most probably artifactual differences, a result of the imperfect teuthid record. Nagy (1958) described a specimen of Teudopsis from the Lower Jurassic of Hungary which Engeser (1988) subsequently referred to T. hunelii. However, Nagy’s figures illustrate a specimen too poorly preserved to be reliably identified to specific level. In addition to the species described above, Riegraf et al. (1984) listed Lioteuthis prohlematica Naef (unknown outside Germany), Geopeltis emarginata (Voltz) and Parahelopeltis flexuosa (Munster). None of these species are known to occur in Britain. Riegraf et al. (1984) also listed the plesioteuthids Paraplesioteutliis sagitata (Miinster) and P. liastata (Munster). Again, neither of these teuthids are known in Britain, but Hall (1985) has recorded the latter species from the Toarcian bituminous shales of Alberta, Canada. This is the only Lower Jurassic squid recorded outside Europe. Acknowledgments. I thank Mr A. A. Berends and Mr P. Thornton (Whitby Museum), Dr M. K. Howarth (British Museum (Natural History)) and Dr D. Price (Sedgwick Museum) for permission to examine and borrow specimens in their care. Discussions with Mr D. Phillips, Dr M. K. Howarth (BMNH) and Dr T. Engeser (Universitat Hamburg) proved fruitful. Dr W. Riegraf (Munster) kindly advised on the status of the type specimens. Dr C. Patterson (BMNH) kindly identified the specimen of Saurorhynchus hrevirostris. The photographs were taken by Mr Harry Taylor of the BMNH photographic studio. This work was completed during a period of employment with the British Antarctic Survey, Cambridge. REEERENCES BATHER, F. A. 1888. Professor Blake and shell growth in Cephalopoda. Annals and Magazine of Natural History. (1888), 421^27. BERTHOLD, T. and ENGESER, T. 1987. Phylogeiietic analysis and systematization of Cephalopoda. Verhandhmgen des Naturwissenschaftlichen Vereins in Hamburg. N.F.. 29, 187-220. BUCKLAND, w. 1836. Geologv and mineralogy considered with reference to natural theology. Treatise IV. The Bridgewater Treatises on the power. wVdom and goodness of God as manifested in creation. William Pickering, London. 1, 618 pp. 2, 129 pp. BULOW-TRUMMER, E. V. 1920. FossHium Catalogus 1: Animalia. Pars II; Cephalopoda Dihranchiata. Junk, Berlin, 313 pp. CRICK, G. c. 1921. On some dibranchiate Cephalopoda from the Upper Lias of Gloucestershire. Proceedings of the Cotteswold Naturalists' Field Club. 20, 249-256. 1922. Notes on some specimens of Cephalopoda figures in Tate and Blake’s ‘Yorkshire Lias’. Naturalist. Hull. IWipm. 273-288. DESLONGCHAMPS, E. 1835. Mcmoirc sur les Teudopsides, animaux fossiles, voisins des Calmars. Bulletin de la Societe linneenne de Normandie. 5, 68-78. DONOVAN, D. T. 1977. Evolution of the dibranchiate Cephalopoda. Symposia of the Zoological Society of London. 38, 15^8. ENGESER, T. 1986. Beschreibung einer wenig bekannten und einer neuen Coleoiden-Art (Vampyromorphoidea, Cephalopoda) aus den Untertithonium von Solnhofen und Eichstatt (Bayern). Archaeopteryx. (1986), 27-35. 206 PALAEONTOLOGY, VOLUME 33 1988. Fossilium Calalogus 1: Animalia. Pars 130: Vampyromorpha (“Fossile Teuthiden ’’ ) ■ Kugler, Amsterdam, 167 pp. ENGESER, T. and PHILLIPS, D. 1986 Redescription of two specimens previously recorded as fossil teuthids (Coleoidea, Cephalopoda). Bulletin of the British Museum (Natural History) (Geology), 40, 259-264. and REiTNER, J. 1983. Geoteuthinus muensteri (d’Orbigny, 1845) aus dem Untertithonium von Daitung und Arnsberg (Bayern). Stuttgarter Beitrdge zur Naturkunde aus dem Staatlichen museum fiir Naturkunde in Stuttgart, B, 92, 1-12. 1985. Teuthiden aus dem Unterapt (‘Tock') von Helgoland (Schleswig-Holstein, Norddeutschland). Paldeontologische Zeitschrift, 59, 245-260. 1986. Coleoidenreste aus der Oberkreide des Libanon im Staatlichen Museum fiir Naturkunde in Stuttgart. Stuttgarter Beitrdge zur Naturkunde aus dem Staatlichen Museum fiir Naturkunde in Stuttgart, B, 124, 1 15 pp. FISCHER, J. and Riou, B. 1982. Les Teuthoides (Cephalopoda, Dibranchiata) du Callovien inferieur du la Voulte- Sur-Rhone (Ardeche, France). Annales de Paleontologie, 68, 295-325. HALL, R. 1985. Paraplesioteuthis hastata (Munster), the first teuthid squid recorded from the Jurassic of North America. Journal of Paleontology, 59, 870-874. HOWARTH, M. K. 1962. The Jet Rock Series and the Alum Shale Series of the Yorkshire Coast. Proceedings of the Yorkshire Geological Society, 33, 381^22. 1980. The Toarcian age of the upper part of the Maristone Rock Bed of England. Palaeontology, 23, 637-656. JELETZKY, J. A. 1965. Taxonomy and phylogeny of fossil Coleoidea (= Dibranchiata). Geological Survey of Canada Papers, 65-2, 72-76. 1966. Comparative morphology, phylogeny and classification of fossil Coleoidea. Paleontological Contributions. University of Kansas. Mollusca, 7, 1-162. LYELL, c. 1878. The student's elements of geology. (3rd edn.). John Murray, London, 672 pp. MATTHEWS, S. C. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology, 16, 713-719. MOORE, c. 1867. On the Middle and Upper Lias of the South West of England. Proceedings of the Somersetshire Archaeological and Natural History Society, 13, 117-197. MORRIS, J. 1854. A catalogue of British fossils. (2nd edn.). Privately published, London, 372 pp. MUNSTER, G. 1843. Die schalenlose Cephalopoden im unteren Jura, den Lias Scieferen von Franken und Schwaben. Beitrage zur Petrefactenkunde, 6, 51-11 . NAEF, A. 1916. Systematische Ubersicht der mediterranen Cephalopoden. Puhblicazioni della Stazione zoologica di Napoli, 1, 11-19. 1921. Das System der dibranchiaten Cephalopoden und die mediterranen Arten derselben. Mitteihmgen aus der Zoologischen Station zu Neapel, 22, 527-542. 1922. Die fossilien Tintenfische. Fine paldozoologische monographie. Fischer, Jena, 322 pp. NAGY, I. z. 1958. Teudopsis subacuta n. sp. A Mecseki Liaszbol. Fdldtani Kozldny, 88, 240-242. ORBIGNY, A. d’ 1842. Paleontologie fraiifaise. Terrains jurassiques. Tome 1. Cepludopodes. Masson, Paris, 642 PP- 1845. Mollusques, vivants et fossiles. Gide. Paris. 605 pp. PHILLIPS, J. 1871. The Geology of Oxford and the Valley of the Thames. Clarendon Press, Oxford, 523 pp. POLLARD, J. E. 1968. The gastric contents of an ichthyosaur from the Lower Lias of Lyme Regis. Palaeontology, 11, 376-388. POWELL, J. H. 1984. Lithostratigraphical nomenclature of the Lias Group in the Yorkshire Basin. Proceedings of the Yorkshire Geological Society, 45, 51-57. PYRAH, B. 1979. Catalogue of type and figured fossils in the Yorkshire Museum: Part 4. Proceedings of the Yorkshire Geological Society, 42, 415-437. QUENSTEDT, F. A. 1839. LoUgo Bollensis ist kein Belemniten-Organ. Neues Jahrbuch fiir Mineralogie, Geognosie, Geologie und Petrefaktenkunde, (1839), 156-157. 1843. Das Fldzgebirge Wiirtembergs. Laupp, Tubingen, 560 pp. 1846-49. Petrefacktenkimde Deutschlands, 1. Die Cephalopoden. Fries, Tubingen, 581 pp. — 1858. Der Jura. Laupp, Tubingen, 842 pp. REGTEREN ALTENA, c. o. VAN 1949. Teyler’s Museum. Systematic catalogue of the palaeontological collection. Sixth supplement (Teuthoidea). Archives du Musee Teyler (3), 10, 53-62. REITNER. J. 1978. Ein Teuthiden-Rest aus dem Obernor (Kossener Schicten) der Lahnes-Niedernachuilde dei Garmisch-Partenkirchen (Bayern). Paldontologische Zeitschrift, 52, 205-212. DOYLE: JURASSIC TEUTHID CEPHALOPODS 207 and ENGESER, T. 1981. Eine neiie Teuthiden-Art aus dem unteren Sineniurium (Lias alpha 3 ‘Olschiefer’) von Dusslingen bei Tubingen (Baden-Wiirttemberg). Neues Jahrhuch fur Geologie imd Paldontulogie. Monatschefte, (1981), 425-430. 1982. Teulhiden aus dem Barreme der Insel Maio (Kapverdische Inseln). Paldontologische Zeitschrift, 56, 209-216. RiEGRAF, w. 1987. On Lower and Upper Jurassic dibranchiate cephalopods from Germany and England. Paldontologische Zeitsclirift, 61, 261-272. and ZUGEL, p. 1984. Ronumiteuthis? ommastrephiformis n. sp. (Cephalopoda, Dibranchiata) aus dem Ornatenlon (Mittlerer Jura, Callovium, Sudwestdeutschland). Neues Jahrbuch fur Geologie uiul Pal- dontologie. Monatshefte, (1984), 61 1-618. , WERNER, G. and LORCHER, E. 1984. Der Posklonienschiefer. Biostratigraphie, Fauna unci Fazies des siidwestdeutschen Untertoarciums (Liase). Enke, Stuttgart, 195 pp. SEiLACHER, A., REiF, w.-E. and WESTPHAL, E. 1985. Sedimentological, ecological and temporal patterns of fossil Lagerstatten. Philosophical Transactions of the Royal Society, Series B, 311, 5-24. SIMPSON, M. 1855. The fossils of the Yorkshire Lias. Privately published, Whitby, 149 pp. 1884. The fossils of the Yorkshire Lias. (2nd edn.). Privately published, Whitby, 256 pp. SMiTHE, E. 1877. On the Middle Lias of North Gloucestershire. The Spinatus Zone. Proceedings of the Cotteswold Naturalists' Field Club, 6, 349-405. TATE, R. and BLAKE, J. E. 1876. The Yorkshire Lias. Van Voorst, London. 475 pp. TEIEODORI, c. 1844. Uber iene im Lias von Banz vorkommende Sepia-Schulpe. Neues Jahrbuch fur Minercdogie, Geologic unci Paldontologie, (1844), 666-675. VOLTZ, M. 1840. Observations sur les Belemnites en general et sur les Belopeltis. Bulletin cle la Societe geologicpce de France (1), 11, 39^8. WOODWARD, A. s. 1899. On the fossil fishes of the Upper Lias of Whitby. Part IV Proceedings of the Yorkshire Geological and Polytechnic Society N.S., 13, 455^72. WOODWARD, s. p. 1851-56. Manual of the Mollusca. Weale, London, 488 pp. ziETEN, c. H. v. 1830-32. Die Versteinerungen Wiirttembergs. Expedition des Werkes unserer Zeit, Stuttgart, 102 pp. PETER DOYLE Nature Conservancy Council Typescript received 30 November 1988 Northminster House Revised typescript received 17 Lebruary 1989 Peterborough PEI lUA, UK Note added in proof. Hall and Neuman (1989) have recorded a new species of Teudopsis, close to T. schuebleri, from the lower Jurassic { falciferum Zone) of Alberta, Canada, suggesting close faunal links with Europe. HALL, R. L. and NEUMAN, A. G. 1989. Teuclopsis caclonunensis, a new teuthid squid from the Toarcian (Lower Jurassic) of Alberta. Journal of Paleontology. 63, 324—327. J fri' Jp': n f-, Ik' gj CLUSTER ANALYSIS OF PREVIOUSLY DESCRIBED COMMUNITIES FROM THE LUDLOW OF THE WELSH BORDERLAND by PIERRE J. LESPERANCE Abstract. Previously described communities ('associations’) from the Ludlow of the Welsh Borderland were subjected to cluster analysis on a PC microcomputer. Miscellaneous absence-presence data sets derived from the published information were analysed using different combinations of clustering algorithms, distance and similarity measurements, with and without Jaccard’s coefficient of association. Variations in order of data entry produced major differences (chaotic behaviour) in clustering using unmodified data. With unmodified data, only the single linkage method showed no clustering differences with the three distance and similarity measurements applied to the whole of the data. The raw data, modified with Jaccard's coefficient, showed major improvement in clustering upon variation of order of data entry. Such improvement in non-chaotic behaviour is interpreted as the result of the lesser, finite raw data consisting of zeros and ones, as compared with the more infinite numbers generated using a coefficient of association. Nearly equally good results are obtained, however, when the unmodified data are analysed using the cosine 0 measurement. The UPGMA, single and complete linkage methods, with cosine 0, are recommended as quickly and routinely applicable to unmodified data for, at least, first approximations in community analysis. Dendrograms so generated should nonetheless be submitted to variation in order of data entry to test for chaotic behaviour. With the advent of microcomputers and powerful programs adapted to this specific hardware, treatment of data previously available only on mainframe computers is now possible within each scientist’s office or home. The purpose of this contribution is to explore the use of cluster analysis on a microcomputer well within its practical limits, but in a situation typical of most palaeoecological studies. Cluster analysis is a non-parametric statistical technique, that is, it is not based on the familiar Gaussian, or probability, bell-shaped curve. It classifies data and produces, if desired, dendrograms (i.e. tree-shaped diagrams), which are visually instructive and, hopefully, of easy interpretation of the classified data. The practical limits of the microcomputers previously alluded to, refer to the available memory. More specifically, the internal memory to hold data, or matrices, places an upper limit to these data in better equipped standard IBM-type personal microcomputers (PC) in the neighbourhood of 4-6 Mb (mega bytes) (of which 4 Mb is expanded memory), at least with the software here employed. About three-quarters of this memory is needed, in a study like this one, if a coefficient of association is calculated on a spreadsheet contained in a single file (note 1, appendix gives more details). Data storage is also a problem, but hard disks of 30 Mb are readily available, and those ten times that amount equally available, but expensive. The description of the Ludlow communities of part of the Welsh Borderland, a portion of the world standard of the upper Silurian, was first outlined by Calef and Hancock (1974), criticized by Lawson (1975), reinvestigated by Watkins (1979) and partly extended to a regional scale by Cherns (1988). Watkins (1979) is a comprehensive study of many aspects of synecology of the whole of the Ludlow in the Welsh Borderland. Although more restricted in area than Cherns (1988), published faunal lists, community tabulations, and stratigraphical coverage are more extensive in Watkins (1979) and better suited to the aims of this study. Although he named the resultant communities ‘associations’, this term was used interchangeably with 'communities’ (Watkins 1979, p. 210), which the writer prefers (but then, discussion persists as to whether these 'associations’ are IPalaeontology, Vol. 33, Part 1, 1990, pp. 209-224. | © The Palaeontological Association 14 PAL .33 210 PALAEONTOLOGY, VOLUME 33 communities: Cherns 1988, p. 488). Other aspects of Ludlow palaeoecology are treated in Watkins (1978), Watkins and Aithie (1980), Hewitt and Watkins (1980), and Mikulic and Watkins (1981), but these contributions do not present data as complete as in Watkins (1979), nor do they specifically modify the 1 979 tabulations. Consequently, Watkins ( 1 979) will be used exclusively here. Stratigraphic nomenclature of the Welsh Borderland dates back, of course, to R. I. Murchison, but modern usage rests on Holland et al. (1963), subsequently very slightly modified by Holland et al. (1980), Holland (1980) and Antia (1980). As Watkins’s (1979) distributional data rest on Holland et al. (1963), their stratigraphic nomenclature is followed. This contribution proposes to use Watkin’s (1979) data as a test case for isolating specific methods and procedures of cluster analysis. As these communities were described using ‘classical’ methods (i.e. communities are recurrent associations defined on the basis of abundance of specific taxa, with consideration commonly given to the spatial continuity of the associations and the absence of specific taxa), comparison of clustering efficiency and correctness can be assessed and judgments expressed. Clustering algorithms are numerous and their respective merits and disadvantages under specific circumstances are not obvious to the applied researcher (some theoretical aspects are covered in Milligan (1980) and Milligan and Isaac (1980)). Furthermore, many distinct coefficients of association between samples have been suggested (binary (absence- presence) ones are surveyed in Cheetham and Hazel 1979; a comprehensive survey is given in Legendre and Legendre 1983), but their respective advantages and disadvantages, again in specific circumstances, are equally far from obvious. In fact, one of the aims of this investigation was to question the necessity of the use of coefficients of association, following results and methodological uncertainties inherent to these results, previously obtained by Lesperance and Sheehan (1988). In view of these uncertainties, a pragmatic approach was best indicated ; this is detailed in the following pages. METHODS Hardware Calculations were performed on an IBM-compatible PC, equipped with an 8088 chip and mathematical coprocessor (8087). Mainboard memory was 640 Kb, with an expanded memory card of L5 Mb (only used by the Symphony software). The PC had a 30 Mb hard disk, with a tape backup of 60 Mb. An EGA card (Extended Graphics Adapter) or better, with its consequent monitor, are a requisite to produce the text-figures as presented. Mainboard memory was always sufficient, and matrices 2-4 times the size of the data here treated, requiring 188 K of RAM, have been analysed without memory shortage, although in this last case processing time increases dramatically to about \ hour. Software Statistical calculations were done using the SPSS/ PC -I- statistical package (version 2-0). Cluster analysis is available in the advanced statistics package; the optional data entry package facilitated the entry of the unmodified data from Watkins (1979), subsequent modification in a spreadsheet, and/or direct entry in the statistical programs. Jaccard’s coefficient of association (note 1, appendix) was calculated on a spreadsheet; Symphony, version L2, was used (Lotus 1-2-3 is equivalent). A file compression utility (SQZ! plus) was invaluable to manage the matrices generated to calculate Jaccard’s coefficient. Data sets The justification for the analysis of different data sets, all derived from Watkins (1979), will be presented in subsequent sections. Only a brief summary of the relations between the major data sets is presented here. Tables 15 to 20 of Watkins (1979) give detailed faunal lists of the six communities, and intermediates or variants, recognized in the Ludlow of the Welsh Borderland. The communities are LESPE RANGE: CEUSTER ANALYSIS OF LUDLOW COMMUNITIES 211 not detailed in the same fashion; as a first step, the 48 different faunal lists were included in one data set. Only taxa identified to the specific or generic level were retained for analysis. A few pelagic taxa were singled out by Watkins (1979); these were included, if only because they provide some sort of information on the physico-chemical conditions in the water column above the level-bottom communities, if indeed all the taxa so identified were pelagic (as, for instance, the case of the brachiopod Aegiria grayi: Cherns 1988, p. 486). A total of 1 12 taxa was consequently retained and, unfortunately, the Atrypa rericularis-cora\ community does not contain a generically determined coral with this procedure. The order of data entry was as presented successively in Watkins’s ( 1979) tables 15 to 20, except that three of the six cumulative faunas, of medium to high diversity from table 15, were arbitrarily entered as the last three faunas. The first data set is referred to as data set 48A. Data set 48B is identical to 48A except that five faunal lists, chosen at random (with the Symphony function of the same name) were removed and reinserted as the last five faunal lists. Data set 46A was derived from 48A by the deletion of two cumulative faunas from Watkins’ (1979) table 15: the Go2 (Glassiu obovata) fauna in the mudstone facies is detailed in table 17 and the Ml (Mesopholidoslrophia laevigata) fauna in table 18, and hence repetitive in the data set. Data set 46B was derived from 48B, with the deletion of the same two cumulative faunas. Table 16 of Watkins (1979) lists six cumulative faunas of two taphonomic categories: disturbed neighbourhood assemblages, and transported assemblages, as recognized in three different communities. The faunal content of these three communities, from specific localities and samples, is presented in subsequent tables. Hence, these six faunal lists are also repetitive in the data sets, and they were deleted to produce data set 40. Data set 40A was derived from 46A, and 40B from 46B. Data set 40C is data set 40B, with five faunal lists chosen at random, deleted, and reinserted as the last five entries. As selected faunal lists were deleted, so were the taxa occurring only in the deleted faunal lists. The raw, unmodified data were coded 0 and I (absent and present respectively) and used as such directly in the clustering. These data were subsequently used for Jaccard’s coefficient, and the same order of data entry retained (e.g. data set 46A, modified or not by Jaccard’s coefficient, has the same sequential order of entry). Justification of the use of absence-presence data is found in most discussions of cluster analysis, and need not be repeated here. The writer believes it is particularly appropriate to regional palaeoecological syntheses, to nullify local effects of species abundances. THE COMMUNITIES Watkins (1979) investigated the Ludlow of part of the Welsh Borderland, but excluded the basal part (the Lower Elton Beds) and the overlying brackish water Ludlow Bone Bed. Within this sequence, he recognized six communities which are, in ascending order, the Glassia obovata (Go) community, a succeeding transitional (tr) fauna with the one above, the Mesopholidostrophia laevigata (Ml) community, a lower phase of the Sphaerirhynchia wilsoni (IwSw) community, an Atrypa reticidaris-cova\ (AC) community, the preceeding two communities locally absent below the upper phase of the Sphaerirhynchia wilsoni (iipSw) community, the Shaleria ornatella (So) community and, uppermost, the Protochonetes ludloviensis (PI) community (table I). These communities were defined using ‘classical’ methods and, more specifically, by a graphical method known as the transect method (Watkins 1979, p. 208) on a spatial basis. Appendix 3 of Watkins (1979, tables 15-20, pp. 262-274) presents a formidable amount of distributional data, comprising but a small part of his unpublished data in the Library of the British Museum (Natural History). Visual examination of these data does not obviously reveal the community relationships, and hence its intended use as a test case for cluster analyses. Tables 15 and 16 of Watkins (1979) give 12 cumulative faunas, while tables 17-20 give examples (from specific localities and collections) from five communities, the Go fauna in the laminated shale facies, the IpSw and the AC faunas being known only by cumulative faunas. Diversity (d) and average diversity ((7) within the miscellaneous faunal lists vary widely. Cumulative faunas have d = 22 to 75. Average diversity in tables 17-20 of Watkins (1979) decreases to about half its value 212 PALAEONTOLOGY, VOLUME 33 TABLE 1. Stratigraphic relations of the benthic communities discussed within the Ludlow (modified from Watkins 1979) WHITCLIFFE PI Protochonetes ludloviensis LEINTWARDINE So Shaleria ornatella upSw upper phase Sphaerirhynchia wilsoni BRINGEWOOD AC Atrvpa reticiilaris-cord\ IpSw lower phase Sphaerirhynchia wilsoni Ml Mesopholidostrophia laevigata ELTON tr transitional fauna Go Glassia ohovata from the lower to the upper part of the Ludlow: from d= 18 9 in the Go fauna, to d= 15-7 in the Ml and S\v faunas, to r? = 8 8 in the Sw fauna and d = 7-3 in the PI fauna. Distributional data are seldom available for sequences of communities, and Watkins’s (1979) readily available publication mitigates against using his more complete unpublished data. Admittedly, these data are almost a worst case situation with such widely varying diversities (the P12C2 locality has but four taxa!) and unequal tabulation of cumulative faunas and individual localities. Nonetheless, results obtained were encouraging (see below). CLUSTERING BEH AVIOU R: CHAOTIC? Reproducibility is assuredly a prime requisite of any analytical method. It is thus particularly disturbing that, following F. Vogel, Bayer (1985, p. 98) has shown, at least geometrically, and with specific data sets, that cluster analysis is subject to chaotic processes (i.e. stochastic, random, aleatory processes, and hence the results are unreproducible) during the formation of clusters. This chaotic behaviour depends, to a great extent, on the sequence of input of the data. This serious defect of cluster analysis needs clarification before any palaeoecological application can confidently be pursued. Data sets 48, 46, and 40 were used as a means of judging this suspected chaotic behaviour. Each set was used in its unmodified form, and its modified form using Jaccard’s coefficient of association, and subjected to various clustering algorithms and distance, or similarity measurements. Cluster techniques Cluster analysis is described to varying extents by the following authors, amongst others, to which the reader may refer for fuller treatment than presented here: Sneath and Sokal (1973), Anderberg (1973), Everitt (1980), Legendre and Legendre (1983), Romesburg (1984), and Jones (1988). Q- mode analysis (between samples) and hierarchical clustering are used exclusively in this contribution. Clustering algorithms. Seven clustering algorithms (methods for combining clusters) are available in the SPSS/PC + package. These are : ( 1 ) average linkage within groups method ( WPGMA : weighted pair-group method using arithmetic averages), (2) average linkage between groups method (UPGMA : unweighted pair-group method using arithmetic averages), (3) single linkage method, (4) complete linkage method, (5) centroid method (UPGMC : unweighted pair-group centroid method), (6) median method (WPGMC : weighted pair-group centroid method), and (7) Ward’s method. Much mistrust has been expressed relative to the single linkage method (discussed in Milligan 1980; rejected by Legendre et al. 1985, p. 275 in succession studies), although it has a natural logic, while UPGMA is generally considered to be the best method (and used exclusively, for instance, by Baarli 1987) . Ward’s method is popular (and used exclusively, for instance, by Lesperance and Sheehan 1988) ; it produces, probably, the most visually appealing (and interpretable?) dendrograms (see also discussion by Romesburg 1984, pp. 134-135). LESPERANCE. CLUSTER ANALYSIS OE LUDLOW COMMUNITIES 213 Distance ami angular measurements. Before combining clusters, an assessment of the distance between the items to be clustered is calculated. Either euclidian distances, or squared euclidian distances, can be used in the clustered hyperspace. Euclidian distances obey the familiar Pythagorean relations of the hypotenuse in a triangle, while squared euclidian distances do not. Squared euclidian distance is consequently referred to as a pseudometric or a semimetric measure (Sneath and Sokal 1973, p. 121 ; Legendre and Legendre 1983, p. 194). Ward's method (Ward 1963) explicitly requires squared euclidian distances, while this same measure is preferable with the centroid and median methods (discussion in Sneath and Sokal 1973, p. 235). Reversals in cluster formation occur when euclidian distances are used with the centroid and median methods (Sneath and Sokal 1973, p. 235), but also occur with squared euclidian distances and the centroid method (Boyce 1969, p. 15). These reversals in clustering values (the agglomeration schedule printouts are available with the SPSS/PC + package) occur in both the raw and modified data, and have been observed with the three distance and similarity measures used with the centroid and median methods. Additional complications arise with the use of euclidian distances with the centroid and median methods in that the dendrograms, with unmodified data, are step-wise and very difficult to interpret. This was not observed with modified data, using euclidian distances, but was present in a few cases with the centroid method using unmodified data and the proper squared euclidian distances. Dendrograms produced using the cosine 6 measure and Ward’s method give results where the majority of the clusters combine at the lowest level of similarity, and hence are meaningless. Nonetheless, it takes little effort to try all distance or similarity measurements on the algorithms to see what happens, much as Jones (1988, p.l6) suggests, even though the mathematics may not be rigorously adhered to, and some of the resultant dendrograms may be of limited use. A third measure, allied to distance measurements, is a similarity measure known as cosine 6 (or as cos tj, or the cosine measure, either considered a shape measure or a pattern similarity measure: note 2, appendix). Imbrie and Purdy (1962) have used it in their study of bahamian carbonates and faunas. Zhang and Hofmann (1982) employed it in their study of lamina shape of Precambrian stromatolites. Ward (1985) has used the cosine 0 measure to compare disjunct variables in Cretaceous communities in Canada; other references are given by Romesburg (1984, p. 109). An a priori assumption in this study was that this measure could possibly help in grouping similar faunas differing only in diversity, but this was not borne out. Coefficients of association. Jaccard’s coefficient of association was used as representative of the numerous coefficients of association that have been suggested. It may be noted here that it can only be used on absence-presence data, and it does not take into account the absence of a taxon in both collections being compared. Eurthermore, this coefficient has long been used in ecology (Sneath and Sokal 1 973, p. 1 3 1 ), is assuredly one of the best known, and most widely used of its class ( Lesperance and Sheehan 1988), although its shortcomings have previously been pointed out (Raup and Crick 1979). Archer and Maples (1987) have also questioned the utility of Jaccard’s coefficient, based on a probabilistic (gaussian) approach (Anderberg 1973, p. 91 discusses only briefly this aspect). In any event, the use of Jaccard’s coefficient is meant as a test of raw versus modified data. Input order and clustering Even though some methods of clustering require specific measurements of distance, all data sets 48, 46 and 40 were sequentially submitted to the seven clustering algorithms, with all three distance- similarity measures taken in turn. Both the raw and modified data were submitted to the same cluster techniques, and dendrograms of the results generated. In order to judge if indeed clustering is chaotic, subjective criteria had to be devised. An obvious result of the great majority of the dendrograms generated was that the nine localities of the Ml Community (Watkins 1979, table 18), as well as the nine localities of the Go Community (Watkins 1979, table 17), with commonly the addition of the Gol {Glassia ohovata fauna in the laminated shale facies, table 15), were correctly clustered together, at various levels of similarity depending on the measure used. The level of similarity defining each individual community was found, and the 214 PALAEONTOLOGY, VOLUME 33 TABLE 2. Comparison of clusters produced upon varying order of data entry on unmodified and modified absence-presence data ds WPGMA U/M UPGMA U/M Single U/M Complete U/M Centroid U/M Median U/M Ward U/M EUCLID 48 -/I -/- I/I -/I -/I -/I -/- 46 -/I -/- I/I -/I I/I I/I -/- 40 I/I I/I I/I I/I -/I V- I/I SEUCLID 48 I/I -/“ I/I -/I -/I I/I -/- 46 I/- -h I/I -/I -/I I/I -/I 40 -/I I/I I/I I/I I/I -/I I/I COS 0 48 I/I I/I I/I I/I -/I I/- NA/NA 46 -/I I/I I/I I/I -/I -/I NA/NA 40 I/I I/I I/I I/I I/I I/I NA/NA I: clusters identical; - clusters do not contain the same collections or faunas; NA: not applicable (clusters meaningless); ds; data set; U: unmodified data; M : data modified with Jaccard’s coefficient; EUCLID: euclidian distance measurements; SEUCLID: squared euclidian distance measurements; COS 0\ cosine 0 similarity measure. Successive data sets derived from the preceding by the deletion of cumulative faunas (Go2 and Ml from data set 48 to obtain data set 46; and deletion of disturbed and transported assemblages upSw, So, and PI from data set 46 to obtain data set 40). average between the two noted. This average was projected throughout the dendrogram (i.e. a phenon line was plotted) and used to assess if clusters of the data using different orders of entry were identical. All major clusters with identical or slightly higher levels of similarity than this phenon line were compared. Those that contained the same faunas (or collections) were judged to be clustered identically. In cases where the dendrograms obtained were step-wise (chaining of Jones 1988, p. 16), judgement was less structured but followed the same general principles. Table 2 presents the results obtained upon varying the order of data entry on the raw and modified data. Obviously, chaotic clustering behaviour occurs, and is a significant characteristic to be reckoned with (text-fig. 1 is typical of the results obtained). This chaotic behaviour was not detected when using the single linkage method, but occurs with all other methods. It occurs equally abundantly whether euclidian or squared euclidian distance measurements are used, but less so when the cosine 6 similarity measure is employed. Furthermore, the modified data clearly are less prone to chaotic behaviour than the raw data. Viewed somewhat differently, the centroid, median and Ward’s methods perform no better with their mandatory squared euclidian distances. A coefficient of association is thus useful, if only to decrease chaotic behaviour, contradicting one of the first assumptions of this study. The cosine 0 similarity measure also appears almost ideal when compared to either of the other two distance measurements, particularly if the centroid, median and Ward’s methods are restricted to their proper squared euclidian distance measurements. This improvement in non-chaotic behaviour is interpreted to be the result of the lesser, finite raw data consisting of zeros and ones, as compared with the more infinite numbers generated using a coefficient of association. Ten decimals were used to compute Jaccard’s coefficient, thus greatly decreasing the chance of two points in hyperspace having the same coordinates (even though only a single quadrant, of 0 to 90°, is considered with the cosine 6 measure), thus decreasing chaotic behaviour as described by Bayer (1985). The most noteworthy results of table 2 are (a) the poor performance of the UPGMA method with euclidian and squared euclidian distances and the equally poor performance of Ward’s method with all three distance-similarity measures, (b) the excellent performance of the single linkage method with all three distanee-similarity measures, and (c) the equally excellent performance of the cosine 9 similarity measure with unmodified and modified data with Jaccard’s coefficient. Dendrograa using Average Linkage (Between Groups) Dendrograe using Average Linkage (Between LESPERANCE: CLUSTER ANALYSIS OF LUDLOW COMMUNITIES 215 *0 ^ V 4-* < m fO N ^ n P5 (0 O') fo < < < < < < < < < < *4 3 » » 3 > U (n CO (0 CO (0 N & a. 0. Q. o, — 3 3 3 3 3 0. (0 (0 < N 0) < U 3 CJ U U N CO N CM N “ MakbW (j— « — u — 0. 30.0.0.0.0.0. 3 3(ocno.o.(/i(ocno. TJ *0 T3 OflDN 0-4 cMcnco if) < o-4®(nr^coioo< MCMCM0-4^®*4-4^®^-4 -4M-HNCMCMCMCMCM << U U O 3 3 iNCMCM®W^*^*^'0^v^CN-4 — — — Q.D.00 — — 000— 0000000000 — — — — — — w-— — w - - a - -.a .a a a .a .a .a a oc50ooooc30C5i:trttrrrro*^E-< -O *0 4J 4J 4J < ® -4 CM CO ® ® ® ® ® < < < < < ^ < < < < 3 3 3 3 3 U ® ® ® ® cn CM aa.Q.Q.a — 3 3 3 3 3 0. ® ■D *0 "O < N ®0 0-4CM®®®< ® — C4I > S o -o OJj 3 3 a ? 2 o ^ c — o> f- CO X S -a o > f]j 3Xj aij 'C ^ 'rr §-< o cd cd o ^ ^ 0-2^ ' X l> u o ;»• 1 0/J I c ^ I ■*-> ’■ \ S '• D- b/j 2 O C 70 U C ^ *"[' "O ^ --5 c d> fO > - 2 ^ X i2 3 cd 3 cd f- CO •= 5 ^ — o .§ ^ -o ■O O H 70 cog rr\ O .= Cl I'S s 0.-S X ^ E 3 p _ Q. ^ P 70 ^ c o o u "O o X u 0 o> c I— P o X .c UJ U L— I f- CL O I § cd "O O 216 PALAEONTOLOGY, VOLUME 33 CLUSTERING EFFICIENCY Again, subjective criteria of clustering efficiency (in relation to the previously described communities) must be devised to judge cluster analyses against Watkins’s ( 1979) communities. The make-up of the various data sets, as well as an overview of the communities, have been previously given, but additional considerations are essential to understand the criteria used. Data set 48 presents an unrealistic situation of cluster analysis in which cumulative faunas and, partly, their constituent parts are considered in a single data set. Nevertheless, it does offer extremes in diversity and an opportunity of testing chaotic behaviour in cluster analysis. Efficiency of clustering was extended to this data set for the sake of completeness. Data set 46 contains the results of a major conclusion submitted by Watkins (1979). He has submitted that transported assemblages are similar in content of epifaunal species to adjacent disturbed neighbourhood assemblages, that the basic community integrity of the transported assemblages is maintained (Watkins 1979, pp. 207-208), and that there was no significant difference between the two. Data set 46 is more realistic than 48, but nonetheless suffers from the improbability of including cumulative taphonomic assemblages with individual collections and, partly, their cumulative faunas. Even so, the taphonomically separated cumulative faunas (Watkins 1979, table 16) do give more comprehensive cumulative faunas than the sum of the individual collections cited in succeeding tables. Data set 40 excludes these taphonomically separated faunas and is, with data set 46. not ideal as some sort of standard. Data sets 46 and 40 must consequently be used, at least as data for tests of clustering efficiency with sets of differing diversity. TABLE 3. Clustering efficiency as judged from comparison with previously described communities or faunas of Watkins (1979) ds WPGMA U/M UPGMA U/M Single U/M Complete U/M Centroid U/M Median U/M Ward U/M EUCLID 48A NC/8 NC/NC 1/6 NC/8 NC/6 NC/NC NC/NC 46B NC/8 NC/NC 1/6 NC/6 NC/6 NC/NC NC/NC 40A NC/5 NC/NC 1/5 NC/5 NC/5 NC/NC NC/NC SEUCLID 48A NC/NC NC/NC 1/6 NC/8 NC/8 NC/8 NC/NC 46B NC/NC NC/NC 2/7 NC/7 NC/7 NC/7 NC/NC 40A NC/NC NC/NC 1/5 NC/5 NC/5 NC/5 NC/NC COS 0 48A NC/8 7/6 4/6 7/9 NC/7 NC/NC NA/NA 46B NC/7 6/8 6/7 8/8 NC/6 NC/NC NA/NA 40A NC/5 4/5 4/5 4/5 NC/4 NC/NC NA/NA NA: not applicable (clusters meaningless); NC : not considered; ds: data set, maximum possible scores on 48A = 12, 46B = 1 1 and 40A = 8; U; score on unmodified data; M: score on data modified with Jaccard’s coefficient ; EUCLID : euclidian distance measurements; SEUCLID : squared euclidian distance measurements; COS 0 '.cosine 0 similarity measure. See table 2 for explanation of the derivation of the data sets, and the text for explanation of A and B suffixes. Keeping in mind the nature of the clustered data, and the fact that high diversity collections will cluster together before clustering with their constituent parts, or other low diversity collections (notes 3 and 4, appendix), it is possible to define a priori clustering results. A total of 12 sets of circumstances applicable to data set 48 can be envisioned (1 1 on data set 46 and 8 on data set 40); these are numbered (1) to (12) below. Clustering efficiency can be judged with the requirement that individual collections from specific localities assigned to specific communities (Watkins 1979, tables 1 7-20) be individually recognized ; these are ( 1 ) the Go medium diversity community, (2) the Ml low Dendrogras ustng Single Linkage Dendrograe using Single Linkage LESPERANCE: CLUSTER ANALYSIS OF LUDLOW COMMUNITIES 217 CQ •O ^ < N U cocnwwtn-eewtft'D-^^v'oetn n aaaaOiO o & a.o o-* — o au — i- — — a — — — j D 3 3 3cncn 3 3incna.a.(n — ®©©eo©© NeM NecsjN-H-H(N -H-N Mar>N©©«4®*40 ^ e © e UJ © © V 4J © © 4J O N ^ < © © © ©©©©©© < < < < < < ^ I » s )U< ©MNNNNNCNJNWWWWWNe o.- — — — — — — — ao.ao.0.— o 3a.a.a.a.©a.o.CL 3 3 3 3 so.© © 0) N©<©0©© M NNNNN*^N < SflBfiSCQffl© 0 o------- ©©zzzzzzz 9 ^ © o < N ■HN ©B bUWUblUJU ——0000000 ZZ©0000(313 © © ^9 <*j © «^ UJ Ui 3 3 U 3 ^-^©©©•i^ « 00— ao.oou&«* (300. 3 3U©< — TEXT-FIG. 2. Unretouched computer printouts demonstrating increase in clustering efficiency with data modified with Jaccard s coefhcient. otherwise as text-fig. 1. A, using unmodified absence-presence data, data set 46B. squared euclidian distances and the single linkage algorithm. B. same as A. but data set modified with Jaccard’s coefficient. Dendrograa ualnf Coaplete Llnkaga Dendrograa using Coapir'^ Linkage 218 PALAEONTOLOGY, VOLUME 33 QQ ^ ^ 4>l V V ^ V < IT) cn M - O cn <<<<< < aONO^csin(om< X1SS3UUXUUUUUU33 << o.a.aa.a— — a» — — — — — Qbaoo — — 000— OOOOOOOOOOQ.O 0 3 3 3 30.au 30.0.0.0.0.0. 3 3t0(no.o.io(ncno.oooooooooo — <• 0) of CO (0 N N S S (B CD < < < <• N ® in o < N N cvj N CM CD OD (B 09 OQ < < < < < CErZZESEZ ^ T} 4.* 4J 4.* V TJ < lO (»5 CM - ♦ p (ocowcnco Titj'o <<<<< ppN < oi> 93 << ♦ m o N ^ o cn 00 fo (H < < < < < » s » s » s •0 “O TJ ®®N o 226 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 1. Map showing fossil wood localities. Triangles: Nanushuk Group (Parrish and Spicer 1988ir/). Circles: Kogosukruk Tongue (this paper). Arrow indicates position of the Campanian Maastrichtian boundary (at Sentinel Hill) with Campanian rocks occurring southward of this point. mire environment (Spicer and Parrish 1987). Plant megafossils generally are found in the overbank mudstones and lacustrine (pond) claystones, although the wood specimens were collected from channel sandstones and levee deposits as well. None of the wood was found in life position. On the basis of pollen, Frederiksen (1986) provisionally drew the Campanian-Maastrichtian boundary within the Kogosukruk Tongue just north of Sentinel Hill. Thus, thirteen of our wood specimens (71 to 80-5; Table I) are probably Late? Campanian age whereas the remaining two (272-1 and 336-1) are Maastrichtian age. PREVIOUS WORK We recently reported an analysis of growth rings in woods from middle Cretaceous rocks of the Nanushuk Group (Chandler Formation) of the North Slope of Alaska (U.S.A.; Parrish and Spicer 1988r/). These rings showed the following: (1) moderate to high variability in ring width, suggesting variable growth conditions from year to year; (2) wide growth rings with rare false rings, suggesting favourable conditions during the growing seasons; and (3) very narrow late wood, suggesting abrupt cessation of otherwise rapid growth. We interpreted these data to mean the following: 1 . The Nanushuk Group trees grew in a shifting sedimentological and hydrological environment. This conclusion is consistent with the fluvial setting interpreted from the rocks. EXPLANATION OF PLATE I Figs. 1^. Xenoxylon latiporosum. I, specimen 272-1, transverse section, x 177. 2, specimen 272.1, tangential longitudinal section, x 177. 3, specimen 336-1, tangential longitudinal section, x 177. 4, specimen 336-1, radial longitudinal section, x 177. PLATE 1 SPICER and PARRISH, Xenoxylon 228 PALAEONTOLOGY. VOLUME 33 TEXT-FIG. 2. Generalized Cretaceous stratigraphy of the Colville River region (modihed from Brosge and Whittington 1966). Shading: marine units. SW NE 2. Conditions during each growing season were favourable for rapid and continuous growth. This is consistent with equable climate during the growing season. The high-latitude positions of northern Alaska during the Late Cretaceous (77°-85°N; Smith et al. 1981; Ziegler et al. 1983) would have resulted in continuous daylight during the summer, and the fossil leaf flora supports the interpretation that climate was relatively warm (Spicer and Parrish 1986). 3. Growth ceased abruptly at the end of the growing season, consistent with the rapid change in photoperiod at high latitudes. Seasonality of light also is supported by the leaves (Spicer and Parrish 1986), which show that all plants were deciduous (angiosperms, taxodiaceous conifers, cycadophytes, ginkgophytes), could die back every winter (sphenophytes, ferns), or could become dormant (cupressaceous conifer). DESCRIPTION OF THE WOODS Xenoxylon latiporosum (Cramer) Gothan (PI. 1, figs. 1^; PI. 5, fig. 3) Transverse section. Secondary wood consists of tracheids. Early wood tracheid lumina typically are 900 pvX in cross-sectional area. Resin canals apparently are absent. Radial longitudinal section. Bordered pits are uniseriate, contiguous, oval, and 25 /mi wide by 15-18 /mi high, with apertures 5 x 7-5 /mi. Tracheids have numerous septa (resin plates?), typically 13-25 /im apart. Cross-field pits are fenestriforni. Ray cells are 25 /mi high. EXPLANATION OF PLATE 2 Figs. 1 and 2. Taxon A. 1, specimen 7-1, transverse section, x 177. 2, specimen 71, radial longitudinal section, xl77. Figs. 3-5. Taxon B. 3, specimen 46T. tangential longitudinal section, x 177. 4, specimen 46T, radial longitudinal section, x 177. 5, specimen 80-1, transverse section, x 177. PLATE 2 SPICER and PARRISH, Cretaceous wood 230 PALAEONTOLOGY, VOLUME 33 Transverse longitmlinal section. Rays are uniseriate, very rarely biseriate, and generally short (usually less than 22 cells high) and 15 /mi wide. Taxon A (PI. 2, figs. 1 and 2) Transverse section. Secondary wood consists of tracheids and parenchymatous rays only. Resin canals are absent. Cross-sectional area of early wood tracheid lumina is typically 550 Radial longitudinal section. Axial tracheids only are present; no axial parenchyma is visible. Bordered pits are irregularly uni- and biseriate, not always contiguous, and typically 12 /mi in diameter, with apertures 3^ /m diameter. Cross-field pits are 6-8 /mi in diameter. Ray cells are 20-24 /mi wide in a vertical direction and 70- 190 /m long, with mostly vertical or sometimes oblique walls. Transverse longitudinal section. Rays are uniseriate, typically 5-20 cells high. Areas of ray cell lumina typically measure 400 /mi- in vertical section. Taxon 5 (PI. 2, figs. 3-5) Transverse section. Secondary wood consists of tracheids and parenchymatous rays only. Resin canals arc absent. Early wood tracheid cell lumina typically are 2000-2300 //m- in cross-section area; overall cell dimensions are 20-25 //m diameter. Radial longitudinal section. Tracheid bordered pits are biseriate, opposite, contiguous, and 20 / 20 is common). Ray cells are 12-15 /im wide. Highly multiseriate areas occur rarely. Taxon D (PI. 4, figs. 1-5) Transverse section. Secondary wood consists of tracheids; resin canals and parenchyma apparently are lacking. Early wood tracheid cross-sectional areas are variable, between 750 /mi’ and 2900 /ml'^ but usually about 2300 /mi“. Radicd longitudinal .section. Bordered pits are small, 12-18 /rm in diameter, with apertures 4-6 /mi in diameter, and are numerous, irregularly distributed, and often isolated. Cross-field pits are poorly preserved, and may be circular, with several per cross-field area. Ray eells are approximately 20 /mi high. EXPLANATION OF PLATE 3 Figs. I — 4. Taxon C. 1, specimen 46-8, transverse section showing false ring in early wood x 177. 2, specimen 46-8, radial longitudinal section, x 177. 3, specimen 46-2, tangential longitudinal section, x 177. 4, specimen 80.5, radial longitudinal section, x 177. PLATE 3 SPICER and PARRISH, Cretaceous wood 232 PALAEONTOLOGY, VOLUME 33 Transverse longitudinal section. Rays are mostly uniseriate, but may be bi- or multiseriate in places, 5-30 cells high, 1 5 /an wide. Taxon £ (PI. 5, figs. 1 and 2) Transverse section. Secondary wood is composed of tracheids and ray parenchyma only. Resin canals apparently are absent. Early wood tracheid lumina are typically 1500-2000 /;m'^ in cross-sectional area. Radial longitudinal section. Bordered pits are uniseriate and contiguous, with borders typically 15 /;m in diameter and apertures 5 //m in diameter. Cross-field pits are not preserved. Ray parenchyma cells are typically 25 //m high. Ray tracheids are 20 //m high. Axial parenchyma is abundant and associated with the rays. Transverse longitudinal section. Rays are uniseriate, typically 30 /mi wide. Axial parenchyma cells are typically 100 /nn X 45 /mi. METHODS Methods of growth-ring analysis have been explained in detail elsewhere (Fritts 1976; Creber 1977 ; Parrish and Spicer 1988«). so only a brief summary will be included here. Characteristics of growth rings that are useful for studying the climatic signal in pre-Quaternary woods are (I ) ring width, (2) interannual variability in ring width, (3) proportion of late wood to early wood, and (4) presence or absence of false rings. Growth rings wider than about 0-5 cm are regarded as indicative of favourable conditions during the growing season, that is, enough light, water and warmth to permit rapid and continuous growth. However, because ring width is also dependent on other factors, such as genetics, no quantitative climatic information may be drawn from this parameter, and it is most useful for comparing woods from diflerent times and/or localities, Interannual variability in ring width is termed 'mean sensitivity’ and quantified using the equation M.S. 1 n-\ ! = n-\ y (-1 2(-V.i-T) ■V+1 + -V where .v, is width of ring t and .v,^^ is the width of the adjacent younger ring. Woods with mean sensitivities less than 0-3 are termed ‘coniplacement ’ and are interpreted to have grown under conditions that were stable from year to year. Woods with mean sensitivities greater than 0-3 are termed ‘sensitive’, suggesting variable conditions from year to year. Sensitive trees of a particular taxon live at the edges of the range of that taxon (e.g. LaMarche 1974; Kay 1978), and the sensitivity is generally linked to climatic effects, although other factors, such as waterlogging of roots, also can affect sensitivity (Fritts 1976). The proportion of late wood to early wood which, like ring width, is qualitative and most useful in a comparative sense, can reflect the nature of seasonality. A high proportion of late wood to early wood is typical of temperate-forest trees, where growing conditions gradually become less favourable as the summer wanes. Narrow late wood, on the other hand, is suggestive of rapid cessation of growth owing to abrupt change in growing conditions such as light (Parrish and Spicer 1988n) or water (Francis 1984). False rings are formed during temporary slowing or cessation of growth during the growing season. False rings indicate that the tree grew under conditions that became temporarily inimical, owing, for example, to fire, drought, freezing or insect attack. The climatic significance of false rings, must, therefore, be supported with other, such as sedimentological, evidence. Preservation of the woods was generally good. Only one specimen, 80-3, showed crushing of the growth rings during compaction. Interestingly, the crushing was not in the early wood, as observed by Jefferson (1982) and EXPLANATION OF PLATE 4 Figs. 1-5. Taxon D. 1, specimen 80-4, radial longitudinal section, x 177. 2, specimen 46-9, radial longitudinal section, x 177. 3, specimen 80-4, tangential longitudinal section, x 177. 4, specimen 46-9, transverse section, X 177. 5, specimen 46-5, transverse section, x 177. PLATE 4 SPICER and PARRISH, Cretaceous wood 234 PALAEONTOLOGY, VOLUME 33 TABLE 1. Wood samples providing data on growth rings and thickness of late wood. Sample number Name Thickness of late wood-number of cells or percentage of total ring width 272-1 Xenoxylon latiporoswn 2-4 cells 336-1 Xenoxylon laliporosum 1-5 cells 7-1 Taxon A 5-12 cells 46-1 Taxon B 2-A cells 80-1 Taxon B 3-9 cells 46-2 Taxon C 2-5 cells 46-8 Taxon C 6-26 cells 80-3 Taxon C not measured (see text) 80-5 Taxon C not measured (see text) 46-5 Taxon D 1-10 46-9 Taxon D up to 34%; 6-23 cells 80-4 Taxon D 5-8 cells 46-7 Taxon E up to 58 % 46-13 Taxon E 1-8 cells 46-4 unident., branch (see text) up to 77 % Parrish and Spicer ( 1988a), but in the late wood, making measurement of the late wood impossible (Table I ). However, overall the average diminution of ring width in this specimen was only about 10%; corrected values were used in the statistics (Tables 2 and 3). RESULTS Growth rings in fifteen specimens were measured from thin sections and/or polished blocks (Table 2). The blocks provide longer ring sequences on which to perform the statistics, so in the discussion below and in Table 3, we use measurements taken from polished blocks in preference to those taken from thin sections. However, both data sets for the relevant samples are presented in Table 2. The longest ring-width series, one from each taxon, are presented in text-figs. 3 and 4. Raw measurements are available from J.T.P. on request; the material is lodged with R.A.S. at Oxford University Museum. All of the growth-ring characteristics reported below were observed in woods collected from widely separated localities. Mean ring widths in woods from the Kogosukruk Group ranged from 0 39 mm to 3-67 mm, with a mean of T76 mm. The narrowest ring measured was 0T4mm (specimen 46-4). The widest ‘normal’ growth ring measured was 5-88 mm (specimen 46-5). This was the innermost ring of the specimen. Ring width normally decreases as the tree ages, although it should be noted that, in this particular specimen, the adjacent and all subsequent rings were much narrower. A growth ring 13-6 mm wide was measured in specimen 46T, and was the first ring in a five-year sequence of wood generated in response to injury. The injury is apparent as a longitudinal scar in the trunk, around which the rings grew. The response, as indicated by the contrast with the normal rings, was dramatic (Table 4), but typical. Mean sensitivities of the Kogosukruk woods were OT 0-0-77, with a mean of 0-40. EXPLANATION OF PLATE 5 Figs. 1 and 2. Taxon E. 1, specimen 46-7, radial longitudinal section, x 177. 2, specimen 46-7, tangential longitudinal section, x 177. Fig. 3. Xenoxylon latiporosiim, specimen 272-1, radial longitudinal section, x 177. PLATE 5 SPICER and PARRISH, Cretaceous wood 236 PALAEONTOLOGY, VOLUME 33 TABLE 2. Data on growth-ring characteristics of woods from the Kogosukruk Tongue of the Prince Creek Formation. Samples are in order by taxon (see Table 1). Sample Number of rings Mean ring width (mm) Variance Mean sensitivity 272-1 15 2-23 1-00 0-52 336-1 31 1-46 1-31 0-53 7-1 7 1-94 0-87 0-77 46-1 20 0-87 0-03 0-12 46-1* 30 0-77 0-02 0-15 80- It 46 1-14 (1-02) 0-24 0-41 46-2 13 3-04 0-82 0-10 46-8 6 1-18 0-31 0-49 80-3t 6 3-67 (3-51) 0-75 0-32 80-5 47 0-53 0-18 0-42 46-5: transect 1** 9 1-35 0-99 0-66 transect 2, (2 thin sections) 17 1-86 1-62 0-48 46-9 14 2-87 1-30 0-39 46-9* 15 3-32 1-56 0-34 80-4 6 3-94 0-29 0-17 80-4* 16 2-96 0-41 0-25 46-7 13 0-97 0-25 0-48 46-13 10 0-98 0-23 0-39 46-4 35 0-39 0-03 0-32 * Measured from block ** This portion sustained injury during growth; the measurements are provided for information only and not included in statistics. t Number m parentheses is uncorrected for crushing (see text). TABLE 3. Comparison of growth-ring characteristics of woods from the Kogosukruk Tongue and the Nanushuk Group. For samples measured from blocks and thin sections, the measurements from the blocks are used in the combined statistics. Kogosukruk woods Nanushuk woods Ring width (mm) Range 0-14-5-88 0-4-12-9 Range of means 0-39-3-67 11^-9 Means of means 1-76 2-81 Mean sensitivity Range 0-10-0-77 0-28-0-76 Mean 0-40 0-44 Late wood Number of cells 1- > 30 1-15 Ratio to early wood max. 0-83 max. 0-30 False rings abundant multiple false rings per growth ring rare ( 1 sample), 1 false ring per growth ring Ring Width (mm) SPICER AND PARRISH: ALASKAN CRETACEOUS WOODS 237 TABLE 4. Growth response to injury in specimen 46- 1 . Only the last few normal rings of a 31 -ring sequence are listed. Ring type Ring width (mm) normal 0-9 normal 10 normal 08 normal 0-7 response 1st year 13-6 2nd year 111* 3rd year 11-6* 4th year 6-6 5th year 5-2 bark * False ring present. Ring Number TEXT-FIG. 3. Ring-width series for specimens 71, 46-1 (block), 46-7, and 80-4 (block). Locations in the series where rings were unmeasurable arc marked by arrows. Note that vertical scales are not equivalent. 238 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 4. Ring-width series for specimens 46-4, 80-5, and 336-1. Locations in the series where rings were immeasurable are marked by arrows. Note that vertical scales are not equivalent. Late wood in growth rings from woods of the Kogosukruk Tongue was rarely narrower than three cells and, where counted, ranged as high as twenty-six cells (Table 1 ). In many rings, counting late wood cells was not possible because the cells diminished in size gradually over a substantial width of the ring; all the late wood in specimen 80-5 was of this nature. Where a distinct zone of late wood was present, its width was as much as 83% of the total ring width. The thickness of late wood was independent of the total width of the ring. False rings were observed in nine of the specimens of wood from the Kogosukruk Tongue (specimens 71, 46-4, 46-5, 46-7, 46-9, 80-1, 80-5, 272-1, 336-1). As many as four false rings were observed in a single growth ring (specimen 46-5); two false rings per growth ring were common (specimens 7-1, 46-5, 46-7, 80-5, 336-1). No difference in growth-ring characteristics exists between Xenoxylon latiporosum, which occurred in the Maastrichtian part of the section, and the other taxa, which occurred in the Campanian part. SPICER AND PARRISH: ALASKAN CRETACEOUS WOODS 239 INTERPRETATION AND COMPARISON WITH WOODS FROM THE NANUSHUK GROUP Ring width and mean sensitivity Growth-ring characteristics of woods of the Kogosukruk Tongue are compared with those of the Nanushuk Group in Table 3. Growth rings in woods from the Nanushuk Group were wider, with means ranging Tl-4-9mm, against 0-39-3-67 mm in woods from the Kogosukruk Tongue. This difference is statistically significant at P-^O-Ol. The narrowest ring measured from Nanushuk woods was 0-4 mm, versus 0T4 mm in the Kogosukruk woods, and the widest rings were 12-9 mm versus 5-88 mm. Woods from both the Nanushuk Group and Kogosukruk Tongue were sensitive (Table 3). Mean sensitivities in woods from the Nanushuk Group were 0-28-0-76 with a mean of 0-44, against 01 0-0 77 and a mean of 0-40 in woods of the Kogosukruk Tongue. This difference in means is statistically significant at P ^ 0-05. Although climate is most important in determining mean sensitivity, it is not the only factor, and we (Parrish and Spicer 1988c?) interpreted the sensitivity of the Nanushuk woods to a shifting sedimentological and hydrological environment, rather than to climate. The woods generally lacked other characteristics, such as significant late wood and false rings, that would have indicated a stronger climatic effect on the growth of the trees. The sensitivity of most of the Kogosukruk woods also could be due to variations in sedimentology and hydrology, as the Kogosukruk Tongue and Chandler Formations (the unit of the Nanushuk Group from which the woods were collected) were deposited in similar environments. However, the Kogosukruk woods exhibit additional characteristics that indicate a somewhat severer climate than that encountered by the woods from the Nanushuk Group. Nevertheless, many of the woods in the Kogosukruk Tongue were complacent, suggesting that, although climate might have been severer overall, the interannual variability was not great. Late wood and false rings The major difference between the two sets of woods was in the amount of late wood and the number of false rings. Late wood in woods from the Nanushuk Group was rarely wider than three cells, with a maximum of fifteen cells in one ring. By contrast, late wood in woods from the Kogosukruk Tongue was usually wider than three cells and, because the rings were narrower, constituted a substantially higher proportion of the wood than in woods from the Nanushuk Group. Growth of the Kogosukruk trees did not cease abruptly, as it did in the Nanushuk trees, but rather slowed during the latter part of each growing season, much as occurs in temperate-region trees today. Thus, the cessation of growth in the Kogosukruk woods was not due just to light, as we interpreted from the Nanushuk woods, but also was influenced by temperature. The resiliency lent to the Kogosukruk woods by the high proportion of late wood is probably partially responsible for the generally good preservation; about half the samples collected were well enough preserved to be useful for analysis. Crushing was much more prevalent in woods from the Nanushuk Group and, indeed, only a fraction of the samples collected from the Nanushuk (seven of forty-five) were sufliciently well preserved to permit growth-ring analysis (Parrish and Spicer 1988??). False rings were abundant in the woods from the Kogosukruk Tongue, whereas only one specimen of wood from the Nanushuk Group had false rings. Multiple false rings within a single growth ring were not observed in woods from the Nanushuk Group. We tentatively rule out insect attack as the cause of false rings in the Kogosukruk woods because we found no evidence for insect attack in either the woods or the leaves, and because many growth rings have more than one false ring. Insects in seasonal climates tend to have rigid life cycles and attacks on trees by a given species of insect will occur during a relatively constrained time period. Thus, if the false rings were formed during the stress of insect attack, several species of insects would have to have been involved. No evidence for even temporary drought has been found in Kogosukruk sediments; indeed, the system 240 PALAEONTOLOGY, VOLUME 33 was very wet (Phillips 1987). Fire and freezing both are plausible explanations for growth disruption in Kogosukruk woods. Charcoal is abundant in Kogosukruk Tongue (indeed, the presence or absence of charcoal can be useful in distinguishing these rocks in core; J. T. Parrish and R. A. Spicer, unpublished data). However, freezing is an equally likely explanation, given the low temperatures suggested by the thick late wood and the vegetational physiognomy (Parrish and Spicer 1988/d- Vegetation The fossil leaf flora changed dramatically between the Nanushuk Group and the Kogosukruk Tongue (Spicer and Parrish 1987; Parrish and Spicer 1988d). The flora of the Nanushuk Group is very diverse, including sixty-seven forms of angiosperm leaves; several taxa each of ferns and conifer leaves and cones; and ginkgophytes, sphenophytes and cycadophytes (Spicer and Parrish 1986). By Kogosukruk time, total diversity of megafossils other than wood was ten forms, including the sphenophyte Equisetites, two ferns, two conifer leaf forms, a fruit, two angiosperm leaf forms and two types of small seed (Spicer and Parrish 1987; Parrish and Spicer 19886). Quantitative estimates of mean annual temperature, derived from angiosperm leaf-margin analysis (Wolfe 1979), were 10°C for the latest Albian and Cenomanian and 13°C for the Coniacian (Parrish and Spicer 19886). The angiosperm megaflora was too depauperate in the Kogosukruk Tongue for leaf-margin analysis, but we interpreted the drastic drop in diversity as indicative of cooling. Based on the overall physiognomy of the flora, we estimated the mean annual temperature to be 2-6°C. In such a climate, cold snaps seem likely (Parrish et al. 1987). Frederiksen et al. (1988) reported a relatively high diversity of angiosperm pollen from the Kogosukruk Tongue. The discrepancy between the diversity of the megaflora and that of the palynoflora would appear to indicate that the angiosperm component of the vegetation was principally herbaceous. This is consistent with the cooler and more variable climate indicated by the woods. Such a climate would favour opportunistic taxa with annual life cycles. The question of whether freezing occurred is critical to understanding the presence of dinosaurs in the Kogosukruk Tongue (Clemens 1985; Brouwers et al. 1987; Parrish et al. 1987; Paul 1988). The palaeobotanical data suggest that winter temperatures were likely to have been close to freezing. In addition, the presence of glendonites, which form in seawater near freezing, in older and younger marine sediments of the North Slope (Kemper 1987) suggests that the Arctic Ocean was likely to have been cool throughout the Cretaceous. Although the vegetation suggests cool temperatures, the morphology of the tracheids in the woods may provide evidence against prolonged freezing. Tracheid cross-sectional areas are related to exposure to water stress. Where water stress is experienced, thick tracheid walls (and therefore small lumina) are necessary to prevent tracheid collapse as tension builds up in the water column. Conversely thin-walled tracheids and large lumina are correlated with stress-free environments. Although some Kogosukruk early wood tracheid cross-sectional areas are as small as 550 //m'^ (specimen 7-1) most are greater than 2000 //m^. This figure is comparable to modern conifers growing in highly mesic environments where water stress is never experienced (Carlquist 1975). Thus, it is unlikely that the Kogosukruk trees experienced freezing of the root zone, which could have induced severe water stress, particularly during the early spring growth. It appears that periglacial conditions were not experienced at sea level, even at northern latitudes greater than 80° N. Acknowledgments. The authors gratefully acknowledge support from British Petroleum, the U.S. Geological Survey, Goldsmiths’ College, University of London, and NATO Grant for International Cooperation in Research RG. 84/0646. SPICER AND PARRISH: ALASKAN CRETACEOUS WOODS 241 REFERENCES ARNOLD, c. A. 1952. Silicified plant remains from the Mesozoic and Tertiary of western North America. II. Some fossil woods from northern Alaska. Papers of the Michigan Academy of Sciences, 38, 9-20. BROSCtE, w. p. and Whittington, c. l. 1966. Geology of the Umiat-Maybe Creek region, Alaska. U.S. Geological Survey Professional Paper, 303-H, 501-638. BROUWERS, E. M. 1988. Late Maestrichtian and Danian faunas from northern Alaska: Reconstructions of environment and biogeography. Abstracts and Programs of the Geological Society of America, 20, A371. CLEMENS, w. A., SPICER, R. A., ACER, T. A., CARTER, L. D. and SLiTER, w. V. 1987. Dinosaurs on the North Slope, Alaska: High latitude latest Cretaceous environments. Science, 237, 1608 1610. CARLQUiST, s. 1975. Ecologicul strategies of xylem evolution. University of California Press, Berkeley, 259 pp. CLEMENS, w. A. 1985. Evaluation of the first Late Cretaceous vertebrate local fauna discovered on the North Slope, Alaska: Report on 1985 field work and preliminary analysis of collections, 36 pp. Unpublished report to the U.S. Geological Survey. CREBER. G. T. 1977. Tree rings: a natural data-storage system. Biological Reviews, 52, 349-383. and CHALLONER, w. G. 1984. Climatic indications from growth rings in fossil woods. 49-74. In BRENCHLEY, p. J. (ed.). FossHs and climate, John Wiley and Sons, New York. 1985. Tree growth in the Mesozoic and Early Tertiary and the reconstruction of palaeoclimates. Palaeogeography, Palaeoclimatology, Palaeoecology, 52, 35-60. 1987. The contribution of growth ring studies to the reconstruction of past climates. 37-67. In ward, R. G. w. (ed.). Applications of tree-ring studies. British Antarctic Research International Series, 333, ERANCis, j. E. 1984. The seasonal environment of the Purbeck (Upper Jurassic) fossil forests. Palaeogeography, Palaeoclimatology, Palaeocology, 48, 285-307. 1986. Growth rings in Cretaceous and Tertiary wood from Antarctica and their palaeoclimatic implications. Palaeontology, 29, 665-684. FREDERIKSEN, N. o. 1986. Reconnaissancc biostratigraphy of Expressipollis and Oculata pollen groups in Campanian and Maastrichtian rocks of the North Slope, Alaska. 19th Annual Meeting, American Association of Stratigraphic Palynologists, Program and Abstracts, 1 1 . AGER, T. A. and EDWARDS, L. 1986. Comment and reply on ‘Early Tertiary marine fossils from northern Alaska: Implications for Arctic Ocean paleogeography and faunal evolution’. Geology, 14, 802-803. and EDWARDS, l. 1988. Palynology of Maastrichtian and Paleocene rocks, lower Colville River region. North Slope of Alaska. Canadian Journal of Earth Sciences, 25, 512-527. FRiTTS, H. c. 1976. Tree rings and climate. Academic Press, New York. 567 pp. GRYC, G., PATTON, w. w., JR., and PAYNE, T. G. 1951 . Present Cretaceous stratigraphic nomenclature of northern Alaska. Washington Academy of Sciences Journal, 41, 159-167. JEFFERSON, T. H. 1982. Fossil forests from the Lower Cretaceous of Alexander Island, Antarctica. Palaeontology, 25,681-708. KAY, p. A. 1978. Dendroecology in Canada’s forest-tundra transition zone. Arctic and Alpine Research, 10, 133-138. KEMPER, E. 1987. Das Klima der Kreide-Zeit. Geologisches Jahrbuch, Reihe A 96, 5-185. LAMARCHE, V. c., JR. 1974. Palaeoclimatic inferences from long tree-ring records. Science, 183, 1043-1048. MARiNCOViCH, L., JR., BROUWERS, E. M. and CARTER, L. D. 1985. Early Tertiary marine fossils from northern Alaska: implications for Arctic Ocean paleogeography and faunal evolution. Geology, 13, 770-773. MCDOUGALL, K. 1986. Maestrichtian benthic foraminifers from Ocean Point, Alaska. Abstracts and Programs of the Geological Society of America, 18, 688. MOLENAAR, c. M., BIRD, K. J. and KIRK, A. R. 1987. Cretaceous and Tertiary stratigraphy of northeastern Alaska. 513-528. In tailleur, i. and weimer, p. (eds.). Alaskan North Slope geology, v. 1, Society of Economic Palaeontologists and Mineralogists and Alaska Geological Society, Bakersfield and Anchorage. PARRISH, J. M., PARRISH, J. T., HUTCHISON, J. H. and SPICER, R. A. 1987. Late Cretaceous vertebrate fossils from the North Slope of Alaska and implications for dinosaur ecology. Palaois, 2, 1-13. PARRISH, J. T. and SPICER, R. A. 1988«. Middle Cretaceous woods from the Nanushuk Group, central North Slope, Alaska. Palaeontology, 31, 19-34. 19886. Late Cretaceous vegetation of the North Slope of Alaska: a near polar temperature curve. Geology, 16, 22-25. PAUL, G. s. 1988. Physiological, migratorial, climatological, geophysical, survival, and evolutionary implications of Cretaceous polar dinosaurs. Journal of Palaeontology 62, 640-652. 16 PAL .1.1 242 PALAEONTOLOGY, VOLUME 33 PHILLIPS, R. L. 1987. Late Cretaceous to early Tertiary deltaic to marine sedimentation. North Slope, Alaska. American Association of Petroleum Geologists Bulletin, 71, 601-602. SMITH, A. G., HURLEY, A. M. and BRiDEN, j. c. 1981. Plumerozoic palaeocontinental words maps. Cambridge University Press, Cambridge. 102 pp. SPICER, R. A. and PARRISH, J. T. 1986. Paleobotanical evidence for cool North Polar climates in middle Cretaceous (Albian-Cenomanian) time. Geology, 14, 703-706. 1987. Plant megafossils, vertebrate remains, and paleoclimate of the Kogosukruk Tongue (Late Cretaceous), North Slope, Alaska, U.S. Geological Survey Circular, 998, 47-48. GRANT, p. R. 1988. Evolution of the vegetation and coal-forming environments in the Late Cretaceous of the North Slope of Alaska. Abstracts and Programs of the Geological Society of America, 20, A29. WOLFE, J. A. 1979. Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions of the northern hemisphere and Australasia. U.S. Geological Survey Professional Paper, 1106, 37 pp. ZIEGLER, A. M., SCOTESE, c. R. and BARRETT, s. F. 1983. Mesozoic and Cenozoic paleogeographic maps. 240-252. In BROSCHE, p and sundermann, j. (eds.). Tidal friction and the eartCs rotation II, Springer-Verlag, Berlin. ROBERT A. SPICER Goldsmiths’ College University of London Rachel McMillan Building, Creek Road London SE8 3BU, U.K. Present address: Department of Earth Sciences Oxford University Parks Road Oxford 0X1 3PR, UK JUDITH TOTMAN PARRISH Department of Geosciences University of Arizona Tucson, Arizona 85721, USA Typescript received 10 January 1989 Revised typescript received 22 May 1989 ORIENTATION OF CEPHALOPOD SHELLS IN ILLUSTRATIONS by SVEN STRIDSBERG Abstract. Most drawings and photographs of fossil cephalopods show tlie shell upside down in respect to the animal’s living position. As there is no advantage in this way of making illustrations, presumably based on tradition, the author suggests that fossil as well as living cephalopods should be illustrated in life position. This is particularly important today, as functional morphology is of vital interest to cephalopod workers. To facilitate understanding of the behaviour of fossil cephalopods, the first step must be to see them orientated in the same way as they saw each other. Illustrations have always played an important role in palaeontological publications and it is of vital importance that they present material in a proper way. This applies to drawings as well as to photographs. All palaeontologists will agree with the above, but unfortunately we are, in some cases, still trapped in the traditional way of presenting illustrations. In the last century when fossils were scientihcally illustrated for the first time, it seems that aesthetics dictated their orientation. Regarding the cephalopods, evolute and involute specimens were normally illustrated with the body chamber on top of the shell in all lateral views. This might perhaps have been artistically satisfying, but it is definitely misleading for one trying to reconstruct the animal or study its functional morphology. The tradition of presenting illustrations of cephalopods upside down, in respect to the living animal, is firmly established among palaeontologists. In many publications illustrating, for example, various forms of ammonites, it is fairly common for complementary drawings to be included to show the supposed living position of the animal. A very good example of such convention is the number of articles concerning the extant nautiloid. Nautilus, where the complete animal is photographed in living position whilst the cut shell, showing all the chambers, is shown upside down. Even I have been accused by an old friend of having illustrated Nautilus upside down (Stridsberg 1981, fig. 2), after he had studied the literature on the subject. All illustrations of Nautilus he could find showed the shell with the body chamber at the top of the shell. Now I find myself asking the same question (text-fig. 1) as he did: ‘Why do they put it upside down?’. The literature to which my friend referred was not only the popular variety but also palaeontology text-books and the Treatise. In the chapter ‘Living Nautilus' in the latter (Stenzel 1964, pp. K59-K93) Nautilus is nicely illustrated in living position with soft parts (fig. 43) and upside down without soft parts (figs. 54-56). In all it is figured in ten pictures, five in living position and five upside down. I have asked before (Stridsberg 1985, p. 10) and I do so again. Who would dream of illustrating an Australopithecus upside down? It must be in the interests of cephalopod workers to facilitate the understanding of all their readers, laymen as well as professionals, of the results they achieve, and not to use misleading methods. If anyone is in the position to interpret the correct living position it is the palaeontologist, and therefore we have a great responsibility to other readers. Naturally it would be a break with tradition for many palaeontologists to see their material illustrated ‘upside down’ (in the old sense), but I strongly urge cephalopod workers to use common sense rather than to continue to follow old conventions. It might be confusing for those adhering to the ‘old system’ but I consider the present situation to be more confusing with all its possible combinations (text-fig. 2). I Palaeontology, Vol. 33, Part I, 1990, pp. 243-248.| © The Palaeontological Association 244 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 1. A, Nautilus shell in living position. B, Nautilus upside down. What makes b look more attractive than A? c and d, the ammonite Kosmoceras in common publication mode (c) and in living position (d). However, there is a problem in reconstrueting the living position in incomplete involute or evoliite cephalopods as the only indicator of up-and-down is the position of the body chamber. In ammonites where there is good reason to believe that only the body chamber is massing due to the lack of the reinforcements the septa make to the phragmocone, the end of the whorl might as well be orientated at the lower part of the shell as at the upper. This suggestion is based on the fact that the body chamber in ammonites often occupies roughly a whole whorl of the shell. In similarly shaped nautiloids with missing body chamber, variation is great between various taxa. In Opliioceras for example the body chamber will occupy almost a whole whorl, while Nautilus has a body chamber occupying only about a third of a whorl. Nevertheless the seeker of perfection must always try to reconstruct the orientation of the specimen under consideration. Naturally due to lack of information incomplete specimens might be incorrectly orientated in the future but that is not an argument for ignoring the problem. An advanced and accurate method on how to reconstruct the life orientation of fossil cephalopods is demonstrated by Okamoto ( 1988), who investigated some heteromorph ammonoids. STRIDSBERG: CEPHALOPOD SHELL ORIENTATION 245 TEXT-FIG. 2. Illustrations of Ophioceras from six publications demonstrating various orientations: a, Ophioceras simplex Barrande 1865, as figured by Barrande (1865, pi. 97, fig. 2). b, Ophidiocerus reticulatiim Angelin 1880, as figured by Angelin (Angelin and Lindstrom 1880, tab. 16, fig. 1). c, Ophidiocerus reticidatiim Angelin 1880 and Ophidiocerus rotu Lindstrom 1890 as figured by Lindstrom (1890, pi. 7, figs. 29 and 34). d, Ophidiocerus welleri Foerste 1930 and Ophidiocerus wilmingtonense Foerste 1925 as figured by Foerste (1930, pi. 25, figs. 5 and 6). e, Ophiocerus reticulutum (Angelin 1880), the same illustration as in C but figured in the Treatise in the same position as Barrande's O. simplex (A) (Furnish el ul. 1964/i, fig. 270: lb), f, Ophiocerus simplex Barrande 1865 as figured by Turek (1972, fig. 3). Apart from one of Foerste’s figures (D (5)) this is the only illustration figuring a specimen ol' Ophiocerus (Ophidiocerus) in living position. The shells from these animals are extremely difficult to orientate due to their highly irregular shape. Assisted by a computer, Okamoto managed to reconstruct extremely well not only the life orientation of adult animals but also changes in life orientation during their ontogeny (Okamoto 1988, text-fig. 6). Naturally this computer method is also available for other cephalopods and similar investigations with symmetric shells will be less complicated. Flopefully there will soon be more investigations employing computer orientations, and this will increase the demand for standardization of cephalopod illustrations. Unfortunately it is not only evolute or involute cephalopods which are treated in an unfair way, but also many of the orthocones. In a palaeoecological paper. Flower (1957 ; pp. 829-852) discussed the horizontal floating position of various kinds. He (Flower 1957, figs. 2-6) demonstrated from the disposition of internal deposits the resulting floating orientations of the animals. Three of his illustrations (Flower 1957, figs. 4-6) have become classical and have been republished several times as they show very clearly the orthocone floating mechanism (see also Flower 1955, p. 246). The deposits found in orthocones are located in the apical chambers, in the siphuncle, or in both areas. They do not completely fill the apical chambers but are concentrated on the ventral side of the shell to help the animal maintain stability. Strictly speaking they served more or less as ballast, to keep longitudinal as well as rotational stability. In those genera where the deposits were concentrated in the siphuncle, this was not situated in the centre of the shell but was ventral, or ventrally to the centre, to obtain the same result, viz. maintenance of stability. 246 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 3. This text-fig intentionally illustrates four longicones in vertical orientation in a full page figure, as if they were too long to be illustrated horizontally. All specimens have a defined floating position and have their dorsal side towards the left, a, Endoceras with siphuncular deposits, indicating what is dorsal and ventral, b, Orthoceras with ventral deposits in the closed chambers, c, in this nearly mature Glossoceras the dorsally located gas chambers serve to keep the shell in balance, d, the gas-filled apical end of Lituites keeps the dorsal side of the shell upwards. As orthocones with cameral or siphuncular deposits, have a defined living position, there is no reason to illustrate these shells with the ventral side upwards. As soon as we can decide what is dorsal and what is ventral, there should be no hesitation in showing this in illustrations (text-fig. 3). In some groups of more or less straight cephalopods with a horizontal living position, stability was not accomplished by ventral deposits, but by dorsally located gas chambers. In Lituites the coiled apical end of the shell served as a stabilizer (text-fig. 3d) and in Glossoceras the dorsally located gas chambers in the mature animal kept the shell in balance (Furnish et al. 1964u, fig. 190c). Regarding the orthocones and other long shells, it is sometimes not practically possible to print lateral views in a proper way as the length of the shell favours a vertical reproduction. ‘ By tradition ’ the apical end has been located mostly to the top of the page. Again, however, it must be emphasized that all lateral views of orthocones, orientated along the page, should be figured with the dorsal side in the same direction, here suggested to the left, regardless of whether the apical end will be located STRIDSBERG: CEPHALOPOD SHELL ORIENTATION 247 at the top or the bottom of the page. This will facilitate the reader to understand the illustration when turning the figure to place the orthocones horizontally. A good example of illustrating a cephalopod in a proper way is the reconstruction of the ascocerid, Glossoceras Imdstroemi Miller in the Treatise (Furnish et al. 1964u, figs. 190-191). However, in the following figures in the chapter the authors have chosen to place all specimens with the apical end in the same direction, and thus some specimens are illustrated with the dorsal side to the right and some with the dorsal side to the left (e.g. fig. 196: lb and 2b respectively). The authors have been consistent in making their illustrations and all shells have the same orientation, although some shells are not placed correctly based on functional morphology. If all shells had been orientated with the dorsal side up, or at least to the left, it would have facilitated comparison of different specimens. The brevicone nautiloids comprise another group carefully illustrated upside down. As the cameral part of the shell acted as the lifting device and the body chamber the sinking device, I can see no reason for figuring these shells with the apical end downwards. In this case it is probably a heritage from Barrande, who made numerous illustrations of rich material (Barrande 1865, 1866). As some of these groups had interesting apertural openings, the material was reproduced upside down several times, just as in Barrande’s work. I recommend that in the future such specimens ought to be figured with the apical end upwards. I have illustrated brevicone nautiloids (Stridsberg 1985, 1988u and 6), in what I believe is the living position and one comment in a review was ‘it will be normal if you turn the page over’. I believe it would be better for us to overturn the old way of making cephalopod illustrations, rather than to leave it to future readers. In summary I strongly recommend that cephalopods are illustrated according to inferred life position, and I hope that this paper will stimulate fruitful discussion on this topic among cephalopod workers. Acknowledgements. For valuable comments on this controversial subject I am most thankful to Charles H. Holland. Dublin, and Lennart Jeppsson, Lund. Thanks to, or due to, their encouragement, 1 decided to finish this paper. I also thank Ingrid Sawers for improving the English and Claes Bergman for drawing a and d in text-fig. 3. REFERENCES ANGELIN, N. p. and LINDSTROM, G. 1880. Frcigmentci silurica e dono Caroli Henrici Wegelin, 60 pp. Stockholm. BARRANDE, J. 1865. Svsteme sihirien dii centre de la Boheme 2, Cephalopodes\ Pis. 1-107. Prague, Paris. 1866. Svsteme silurien dit centre de la Boheme 2, Cephalopodes, Pis. 108-244. Prague, Paris. FLOWER. R. H. 1955. Saltations in nautiloid coiling. Evolution, 9, 244-260. 1957. Nautiloids of the Paleozoic. Geological Society of America, Memoir 67, 829-852. FOERSTE, A. F. 1925. Notes on cephalopod genera; chiefly coiled Silurian forms. Journal of the scientific laboratories of Denison University. 21, 1-69, pis. 1-24. 1930. Port Byron and other Silurian cephalopods. Journal of the scientific laboratories of Denison University, 25, 1-124, pis. 1-25. FURNISH, w. M. and GLENISTER, B. F. 1964c/. Nautiloidea-Ascocerida. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology, Part K, Mollusca 3, 261-277. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. 19646. Nautiloidea-Tarphycerida. In moore, r. C. (ed.). Treatise on Invertebrate Paleontology, Part K, Mollusca 3, 343-368. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. LINDSTROM, G. 1890. The Ascoceratidae and the Lituitidae of the upper Silurian formation of Gotland. Kungliga Vetenskaps-Akademiens Handlingar, 23-12, 54 pp., 7 pis. OKAMOTO, T. 1988. Changes in life orientation during the ontogeny of some heteromorph ammonoids. Palaeontology, 31, 281-294. STENZEL, H. B. 1964. Living Nautilus. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology, Part K. Mollusca 3, 59-93. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas. 248 PALAEONTOLOGY, VOLUME 33 STRiDSBERG, s. 1981. Apei'tural constrictions in some oncocerid cephalopods. Lethaia, 14, 269-276. 1985. Silurian oncocerid cephalopods from Gotland. Fossils and strata, 18, 65 pp. 1988a. Evolution within the Silurian cephalopod genus Inversoceras. Paldontologische Zeitschrift, 62, 59-69. 1988/5. A Silurian cephalopod genus with a reinforced frilled shell. Palaeontolo^gy, 31, 651-663. TUREK, V. 1972. On the systematic position of the genus Ophioceras Barrande, 1865. Casopis Ndrodmho Muzea -odd. pnrodovedny, 141, 30-33, pi. 1. SVEN STRIDSBERG Institute for Historical Geology and Palaeontology University of Lund Solvegatan 13 S-223 62 Lund Sweden Typescript received 21 September 1988 Revised typescript received 25 April 1989 A NEW EOCRINOID FROM THE LOWER CAMBRIAN OF SPAIN by G. UBAGHS and a. vizcaino Abstract. An eocrinoid similar to Gogia, but having single or partitioned cpispires with dome-shaped calcite cover, is described as Gogia (Alanisicystis) andalusiae siibg. et sp. nov. It is from the Lower Cambrian (Marianian stage) of Alanis (Seville), Andalusia, Spain, and it is the second species of Gogia from the Lower Cambrian. The eocrinoids comprise a heterogeneous assemblage of primitive pelmatozoans, which probably contains the ancestors of all other cystoids s.l. or blastozoan echinoderms (see Smith 1984; Paul 1988). Yet the term ‘eoerinoid’ remains convenient to designate those Cambrian and Ordovician pelmatozoans which, at this time, cannot be referred to any of the currently accepted classes. Only three genera and three species of eocrinoids (including lepidocystoids) are known from the Lower Cambrian, all of them from North America (Sprinkle 1973; Durham 1978). They are Kinzercvstis durhami Sprinkle, Lepidocystis wauneri Foerste and Gogia ojenai Durham. In addition, isolated plates from the Lower Cambrian of California (Sprinkle 1973), England (Donovan and Paul 1982) and south-east Iran (Wolfart 1974) have been described as possible or probable eoerinoid remains. Among the species just referred to, only one has been assigned to the genus Gogia: G. ojenai from the late Lower Cambrian Latham Shale of California (Durham 1978). A second, but somewhat older species, was discovered a few years ago by one of us (D.V.) in south-western Spain (Andalusia) near Alanis (Provinee of Seville). This new species, described below, is of special interest because of its epispires as well as being the oldest recorded representative of the genus. The presence of echinoderm plates in the Lower Cambrian of Alanis has been previously reported by R. and E. Riehter (1940) and G. Henningsmoen (1958). Aceording to the latter, the plates resemble those of Gogia prolifica Walcott from the Mt Whyte Formation, lower Middle Cambrian, British Columbia, Canada. LOCATION AND AGE The fossils dealt with herein were eolleeted near Alanis, about 75 km north-north-east of Seville, in the Sierra Morena Oriental, from an outcrop on a pathway going from Alanis station to Hornillo- Viejo farm, some twenty metres after the crossing of the Benalija River (the position of this outcrop is indicated (2) in fig. 1 of Gil Cid 1972). The layer that yielded the eehinoderm remains belongs to the upper part of the Marianian stage (Lower Cambrian; Sdzuy 1971), eorrelated with beds of Botomian age of the Siberian sequence (Sdzuy 1972). The trilobite fauna of Alanis, as revised by Gil Cid (1975), comprises the following species; Saiikianda andalusiae R. and E. Riehter, Perrector perrectus R. and E. Richter, Eops eo R. and E. Riehter, Strenueva sampelayoi R. and E. Richter, Stremieva melendezi Gil Cid and Alanisia guillernwi R. and E. Richter. (Palaeontology, Vol. 33, Part I, 1990, pp. 249-256, I pl.| © The Palaeontological Association 250 PALAEONTOLOGY, VOLUME 33 SYSTEMATIC PALAEONTOLOGY ‘Class’ Eocrinoidea Jaekel, 1918 Family Eocrinidae Jaekel, 1918 Genus gogia Walcott, 1917 Subgenus alanisicystis subg. nov. Etymology. From Alanis (Seville), Spain, type locality. Type species. Gogia (Alanisicystis) andalusiae sp. nov. Diagnosis A subgenus of Gogia, characterized by single or partitioned epispires provided with external dome-like stereomic cover. Discussion. Alanisicystis conforms with the diagnoses proposed by Sprinkle (1973) for the family Eocrinidae and the genus Gogia, except for the peculiar morphology of its epispires. While those of Gogia are single pore-like sutural openings surrounded by a prominent raised rim, those of Alanisicystis are commonly divided into two funnel-shaped hollows leading to a single or to paired internal pores and protected by an external dome-like stereomic cover or a pair of such covers. Epithecal covering on epispires is known in several eocrinoids but none appears to be the same as that observed in Alanisicystis. In the type specimen of the Middle Cambrian Acanthocystites briareus, a thin lid - possibly on the plate interior (Sprinkle 1973) -with pores at opposite ends closes off the central part of each epispire (Ubaghs 1967), but other specimens assigned to this species by Fatka and Kordule (1984) do not show this feature (Fatka, personal communication). In the Lower Ordovician, Rhopalocystis destomhesi, the existence of an external cover of minute plates over epispire pores has been reported (Ubaghs 1963). In other eocrinoids such as the Ordovician Palaeocystites, it is the plate epistereom itself that covers the epispires, which are therefore exposed only where the external plate surface has been eroded. But it is probably the calcitic cupolas protecting the humatipores of the diploporites Holocystites s.s. and Pustidocystis that the epispire covering of Alanisicystis most resembles, at least superficially. Alanisicystis has an irregularly multiplated calyx and holdfast, more or less numerous epispires, spiralled brachioles (a feature found in some other species of Gogia but in no other echinoderms) attached separately or in groups to spout-like projections of modified calyx plates on the edge of the oral area, as well as an anal pyramid laterally located near the calyx summit. All these characters fit those of Gogia, indicating a close relationship with this genus. Still, with its peculiar and complex epispires, Alanisicystis stands apart from all known representatives of Gogia. It is unlikely that it could have evolved into one of them and probably represents an early offshoot from the main stock of Gogia. It is to mark at one and the same time its distinctiveness from, and its similarity to, Gogia that it is here considered as a subgenus of the latter. Gogia (Alanisicystis) andcdusiae subg. et sp. nov. Etynwlogy. From Andalusia, a region of southern Spain, where this new species was discovered. Holotype. Specimen VCE 24 (PI. 1, fig. 7). Diagnosis. A species of Alanisicystis with calyx probably globose; thecal plates relatively large and thick; showing pustulose and low ridged exterior ornament, epispires all over calyx or part of it, up to 12 per plate in large specimens; holdfast relatively short and distally inflated, composed of numerous small, unornamented plates; transition from calyx to holdfast abrupt; at least 6 or 7 brachioles, spiralled in right hand direction. Material. Seven specimens (three with both part and counterpart) and isolated plates, all preserved as external moulds in a greenish shale and more or less distorted. Brachioles and holdfast still attached to the theca in three UBAGHS AND VIZCAINO: SPANISH LOWER CAMBRIAN EOCRINOID 251 F I H TEXT-FIG. 1. Epispires of Gogia {Alcmisicystis) andalusiae subg. et sp. nov., camera lucida drawings. A, VCE 1 1,7, thecal plates of a small specimen showing simple epispires (one cover missing). B-1, VCE 24; B, thecal plate with partitioned epispires, covers not preserved; C, paired covers and small simple (?incipient) epispire with cover on the right side; D, paired covers slightly displaced; E, small simple epispire with cover between two large plates; F and G, oblique view of epispires with partial and complete partition; H, smooth internal face of a plate, with slight markings associated with sutural openings (note the small size and feeble differentiation of these openings); 1 portion of internal edge of a plate, with simple and partitioned sutural openings. All figures approximately x 17-5. specimens, suggesting rapid burial at or near place of life. The specimens are numbered VCE 1 1,7-i, 23, 24, 25 and 26; they will be deposited in the collection of the Faculty of Sciences of the University of Madrid, Spain. Description. Differences worthy of note may be observed between the smaller and the larger specimens. They mainly concern the shape of the theca, the relative size, complexity and distribution of the epispires, the ornamentation of the calyx plates and the aspect of the anal pyramid. They suggest the possibility of the existence of more than one species. Nevertheless, as the available material is small and variously preserved, it is dealt with below as representing a single taxon. The specimens show a great diversity in size and thecal shape - the latter partly at least as a result of rock deformation. The calyces of the two smallest specimens (together on one slab and having the same orientation) are approximately twice as wide as high; VCE 1 1,7 (PI. I, fig. I) is 6-6 mm wide and 3-9 mm high; VCE 1 1,2 (PI. 1, fig. 4) is 61 mm wide and 2-9 mm high. The largest calyx (VCE 24; PI. I, fig. 7) is higher ( I L2 mm) than wide (9-5 mm), but it is distorted. Its strong convexity suggests, however, that, like the calyx of the two smallest specimens, it was initially globose. 252 PALAEONTOLOGY, VOLUME 33 The calyx plates are relatively thick (up to 0-3 mm in larger specimens), polygonal, tesseiated and irregularity arranged (PI. 1, figs. 1, 4, 5). Their number per side is 18-25, suggesting an average of 35-50 plates for the whole calyx. The larger ones are located in the lower and middle portions of the theca (PI. 1, figs. 1, 4, 7). In the smaller specimens, the calyx plates are slightly to moderately convex; in the larger ones, they have a slightly domed centre, but a few of the smaller plates are flat or even concave. They are rather coarsely ornamented. Those of the smaller specimens have irregular pustules, which become fainter towards the plate edges (PI. 1, figs. 2 and 5), while those of the largest individual have vermiculate pustules concentrated on the upraised central area and passing to the periphery into low ridges which extend to the plate margins or meet the thickened apical portion of the prominent rim of the epispires (PI. 1, fig. 9). Unlike the outer surface, the inner side of the calyx plates is smooth, but for faint markings associated with the sutural pores (text-fig. 1h). The epispires are the most distinctive feature of the species. In the two smaller individuals, they are nearly or entirely lacking from one side of the theca (PI. 1, figs. 4 and 5), while present on most plates of the other side, ranging from none to five per plate (PI. 1, fig. 2; text-fig. I a); most of them are ‘U’ shaped, small (0-2 mm wide), simple, with a raised rim thickened at the apex and provided with a hemispherical cap-like stereomic cover (not always preserved; text-fig. 1a). In the largest available specimen (VCE 24), the epispires seem to occur over the entire calyx, up to 12 per plate. They appear externally as conspicuous funnel-shaped sutural hollows, more than twice as large as those of the smaller specimens. They are surrounded by a powerful rim. with generally a prominent buttress-like thickening at the apex (text-fig. Ib-d). Their floor is rounded and smooth or it may comprise several shallow depressions (text-fig. 1b), but more commonly it is divided into two similar hollows by an internal partition which may be partial or complete. In the former case the epispire has a single internal pore (text-fig. If), in the latter case it has two contiguous internal openings (text-fig. 1g, i). As stated above, each epispire has a conical or rounded cap-like cover, or a pair of such covers when the epispire comprises two separate compartments (PI. I, figs. 2, 7, 9; text-fig. Ic, d). These covers do not seem to be a mere continuation of the calyx plates, but rather distinct skeletal elements, apparently loosely connected to the calyx, for not infrequently they are displaced or altogether missing and, when missing, no trace of their attachment can be observed. They most certainly served to protect the presumed respiratory evaginations of the body wall that the epispires are said to have accommodated during life, but the question arises whether they could open and close, or whether they were permanently closed. In that case, gaseous exchange must have been effected through them, as it was through the calcified external surface of the tubercular humatipores of some diploporite cystoids (Holocystites s.s., Pustulocystis), which structures they somewhat resemble. The partitioning of the epispires suggest that the soft organs they contained could branch, as do the papulae of some Recent asteroids. As to the internal pores, they are small and simple compared to the extended and complex external openings, though they generally show faint markings along their edges (text-fig. 1h). The occasional presence of small simple (?incipient) epispires in the largest plates, is noteworthy, and suggests that new epispires could still be added at a relatively late stage of plate development (text-fig. Ic, e). The transition from the calyx to the holdfast is abrupt. The best preserved holdfast (VCE 1 1,2) seems to be composed of a short (0-5 mm) and wide (T6 mm) cylindrical proximal part, with about 10 plates across the width, and a distal expanded portion (2-9 mm largest diameter; PI. I, figs. 4 and 5). The holdfast is made of tiny unornamented, rounded plates (OT 5-0-2 mm wide), slightly imbricating towards the distal end of the organ (PI. 1, fig. 2). The observed number of brachioles ranges from 5 (VCE 11,7) to 7 (VCE 24) None of these appendages is complete. The longest preserved one (VCE 24) is 27-5 mm, suggesting a brachiole: calyx length ratio equal or greater than to 2-5 to 3. The brachioles are spiralled in a clockwise direction: there are about 3 spirals in the longest observed brachiole; the spiralling in this, as in other brachioles of the same specimen, starts within EXPLANATION OF PLATE 1 Figs. 1-9. Gogia (Akmisicystis) andalusiae subg. et sp. nov. 1, 3 and 6, VCE 11,7, 1, calyx and part of brachioles of a small specimen, x6; 3, anal pyramid (see text-fig. 3A), x 10; 6, proximal portion of two brachioles, X 10. 2, 4 and 5, VCE 1 1,2; 2, calyx and inflated distal portion of the holdfast (note presence of epispires on calyx plates) x 8; 4 and 5, counterpart of the same small specimen and detail of calyx plates and holdfast (note absence of epispires on this calyx face), x 6 and x 12. 7, 8 and 9, VCE 24, holotype; 7, calyx (somewhat distorted) and brachioles (note spiralling of brachioles in clockwise direction; the arrow indicates the anal pyramid), x4; 8, portion of two brachioles (see text-fig. 2b), x 10; 9, detail of calyx plates, x8. All photographs are of latex casts whitened with ammonium chloride sublimate. PLATE 1 9 UBAGHS AND VIZCAINO: Gogia (AUmicystis) amlahisiae nov. 254 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. Gogia (Alanisicystis) andalusiae subg. et sp. nov., camera lucida drawings. A, VCE 11,7, portion of a brachiole, x 17-5. B, VCE 24, portion of two brachioles (see PI. 1, fig. 8) x 17-5. C, VCE 1 1,7, calyx summit showing oral area, insertion and proximal portions of five brachioles (br) and anal pyramid (ap), x 12-5. D, reverse view of the pair of brachioles on the left side of the preceding figure, x 12-5. 2-3 mm of the calyx summit (PI. 1, figs. 1, 6, 7). The brachioles taper very gradually. They are composed of unornamented brachiolar plates, about 15 on a side in 5 mm proximally and about 20 more distally (specimen VCE 24). Their ventral groove, as observed on the proximal portion of a brachiole, is wide and deep; it shows on its lateral walls an abrupt change of surface, possibly serving as the mounting area for cover plates. The latter are a little more numerous than the brachiolar plates, in a ratio of 7 or 8 to 5. There are apparently two sets of them, smaller ones alternating with larger ones (PI. 1, figs. 6 and 8; text-fig. 2a, b). The latter, which slightly imbricate in distal direction, are as wide as the brachiolar groove, while the smaller ones cover only a part of it. The brachioles are attached to unornamented spout-like calyx plates edging the relatively small oral area, either singly or in groups of 2 or 3 (PI. I , figs. 1 and 6; text-fig. 2c, d). Each brachiole is inserted on two thecal plates (text-fig. 2d), except possibly in the A ray (opposite the anus) where an apparently single large thecal plate bears only one attachment facet (text-fig. 2c). Neither the mouth nor the ambulacra! grooves have been observed on the thecal summit, but the arrangement of brachioles around the oral area in specimen VCE 1 1,7 - the only one to show this part of the calyx - suggests the existence of a single ambulacral groove in the A ray and of two lateral ones leading to a pair of brachioles on either side; such triradiate symmetry of the ambulacral system would certainly represent a very primitive condition. The oral surface is paved with small (0T2-0-20 mm wide), unornamented irregular plates. On its very edge or slightly below it rises a relatively conspicuous anal pyramid. In the two smaller specimens (VCE 1 1,7 and 2), the anal pyramid appears as a pointed structure (2-3 mm wide at base, 1-7 mm high in specimen VCE 11,7) made of rows of elongated convex plates of decreasing size towards the summit and ending in needle-shaped platelets (PI. 1, fig. 3; text-fig. 3a). In the largest specimen (holotype), it looks like a truncated cone (L7 mm wide at base, 1-5 mm high) composed of rows (6-7 on the exposed face) of subquadrate, convex plates converging towards the apex (possibly missing; PI. 1, fig. 7: text-fig. 3b). UBAGHS AND VIZCAINO: SPANISH LOWER CAMBRIAN EOCRINOID 255 THXT-FIG. 3. Gogia (Alanisicvstis) andalitsiae subg. et sp, nov., camera lucida drawings, anal pyramid. A, VCE 11,7, (see PI. I, fig. 3), x 17-5. B, VCE 24 (see PI. I, fig. 7), x 17. Comparison. Gogia (Alanisicvstis) andalusiae shows the same general organization as all known species of Gogia. In addition, it shares; I . spiralling of brachioles (though in opposite direction) with G. ojenai, G. spiralis, G. granulosa and G. giintheri', 2. relative shortness and distal expansion of holdfast with G. ojenai and G. goncli ', 3. abrupt transition of calyx to holdfast with G. spiralis and G. radiata', 4. coarseness of calyx plate ornamentation with G. ojenai, G. granulosa and some Poleta plates figured by Sprinkle (1973, pi. 25, figs. 9 and 19). On the whole, it is G. ojenai of late Lower Cambrian (Durham 1978), that it most resembles, for it has in common with this species characters 1, 2 and 4 just mentioned. It differs from G. ojenai by its complex and covered epispires, less sharply ridged calyx plate ornamentation, less numerous and clockwise (instead of anticlockwise) spiralled brachioles, and abrupt rather than gradational transition from calyx to holdfast. While assuredly a primitive echinoderm, G. (Alanisicvstis) andalusiae has surprisingly sophisticated respiratory structures, suggesting an early appearance of advanced features among the cystoid pelmatozoans. Acknowledgements . We thank Dr J. Sprinkle for his helpful comments. Dr O. Fatka, Dr D. Gil Cid and Dr K. Sdzuy for various information. Dr C. R. C. Paul for his editorial advice, and Dr E. Poty and Mr W. Strouvens for the photographic illustrations of this paper. REFERENCES DONOVAN, s. K. and PAUL, c. R. c. 1982. Lower Cambrian echinoderm plates from Comley, Shropshire, England. Geological Magazine, 119, 611 614. DURHAM, j. w. 1978. A lower Cambrian eocrinoid. Journal of Paleontology, 52, 195-199. FATKA, o. and KORDULE, V. 1984. Acanthocystites Barrande, 1887 (Eocrinoidea) from the Jince Formation (Middle Cambrian) of the Barrandian area. Vestnik Ustredm'ho ustavii geologickcdw, 59, 299-302. GIL CID, D. 1972. Strenueva melendezi nov. sp. del Cambrico inferior de Alanis (Sevilla). Estudios geologicos, 28, 463-467. 1975. Los trilobites del Cambrico inferior de Alanis (Sevilla). Boletin Geologico y Minero, 86, 365-378. HENNINGSMOEN, G. 1958. Los trilobites de las capas de Saukianda. Cambrico inferior, en Andalusia. Estudios geologicos, 14 (1957), 251-271. JAEKEL, o. 1918. Phylogenie und System der Pelmatozoen. Paldontologisclie Zeitschriji. 3, 1-128. PAUL, c. R. c. 1988. The phytogeny of the cystoids. Pp. 199-213 in Paul, c. r. c. and smith, a. b. (eds.). Echinoderm phylogeny and evolutionary biology. Clarendon Press, Oxford, x + 373 pp. RICHTER, R. and RICHTER. E. 1940. Die SaukiandaSGdQ von Andalusien, eine fremde Fauna im europaischen Ober-Kambrium. Ahhandlungen der Senckenhergischen Naturforschenden Gesellschaji, 450, 1 -88. sdzuy, k. 1971. Acerca de la correlacion del Cambrico inferior on la Peninsula Iberica. I. Congresso Hispanico Luso Americano geologica economico seccion I, 753-768. 256 PALAEONTOLOGY, VOLUME 33 1972. Das Kambrium der acadobaltischen Faunenprovinz. Gegenwartiger Kenntnisstand und Probleme. Zentralblatt fiir Geologic und Paldontologie, Teil II, Jahrg. 1972, 1-91. SMITH, A. B. 1984. Classification of the Echinodermata. Palaeontology, 27, 431-459. SPRINKLE, J. 1973. Morphology and evolution of blastozoan echinoderms. Special Publication Museum of Comparative Zoology Harvard University, 284 pp. UBAGHS, G. 1963. Rhopalocystis destombesi n.g.n.sp., eocrinoi'de de I’Ordovicien inferieur (Tremadocien superieur) du sud marocain. Notes du Service geologique du Maroc, 23, 25^4. 1967. Eocrinoidea. Pp. S455-495. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part S: Echinodermata 1, 2 vols.. Geological Society of America and University of Kansas Press, Lawrence, Kansas. WALCOTT, c. D. 1917. Cambrian geology and paleontology. IV, Fauna of the Mount Whyte Formation. Smithsonian Miscellaneous Collections, 67, 61-114. WOLFART R. 1974. Die Fauna (Brachiopoda, Mollusca, Trilobita) aus dem Unter-Kambrium von Kerman, Sudost-Iran. Geologisches Jahrbuch, 8, 5-70. G. UBAGHS Laboratoire de Paleontologie Universite de Liege 7, place du Vingt-Aout 4000 Liege BELGIUM Typescript received 10 February 1989 Revised typescript received 5 April 1989 D. VIZCAINO 22, rue Basse 1 1000 Carcassonne France 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 Geology, University of Wales College of Cardiff CFl 3YE, 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 1990 is £45 00 (U.S. $66) for Institutional Members, and £20 00 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available on a non-subscription basis to Ordinary and Student Members only, for their personal use, at a discount of 25 % below the listed prices: contact the Marketing Manager. Non-members may obtain Nos 35^0 (at cover price) from Basil Blackwell Ltd., Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford 0X2 ODT, England, and older issues from the Marketing Manager. For all orders of Special Papers through the Marketing Manager, please add £1-50 (U.S. $3) per item for postage and packing. PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS Special Papers in Palaeontology For full catalogue and price list, send a self-addressed, stamped A4 envelope to the Marketing Manager. Numbers 2-1 1 and 13-30 are still in print and are available together with those listed below: 31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, hy 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, hy o. i. Nikiforova, T. L. modzalevskaya and m. g. bassett. 66 pp., 6 text-figs., 16 plates. Price £10 (U.S. $15). 35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and d. e. g. BRIGGS. 178 pp., 29 plates. Price £30 (U.S. $50). 36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. Kennedy. 145 pp., 43 text-figs., 23 plates. Price £20 (U.S. $35). 37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, hy rachel wood. 89 pp., 31 text- figs., 1 plates. Price £20 (U.S. $35). 38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, hy p. r. BOWN. 118 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50). 39. (for 1988): Late Cenomanian and Turonian ammonite faunas from north-east and central Texas, hy w. j. Kennedy, 131 pp., 39 text-figs., 24 plates. Price £30 (U.S. $50). 40. (for 1988): The use and conservation of palaeontological sites. Edited by p. r. crowther and w. a. wimbledon. 200 pp., 31 text-figs. Price £30 (U.S. $50). 41. (for 1989): Late Jurassic-early Cretaceous cephalopods of eastern Alexander Island, Antarctica, hy p. j. howlett. 72 pp., 9 text-figs., 10 plates. Price £20 (U.S. $35). 42. (for 1989): The Palaeocene flora of the Isle of Mull, hy m. c. boulter and z. kvacek. 149 pp., 23 text-figs., 23 plates. Price £40 (U.S. $70). Field Guides to Fossils and Other Publications These are available only from the Marketing Manager. Please add £100 (U.S. $2) per book for postage and packing plus £1-50 (U.S. $3) for airmail. Payments should be in Sterling or in U.S. dollars, with all exchange charges prepaid. Cheques should be made payable to the Palaeontological Association. 1. (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12) (Members £6 or U.S. $9). 2. (1987): Fossils of the Chalk, compiled by e. owen; edited hy a. b, smith. 306 pp., 59 plates. Price £11-50 (U.S. $18) (Members £9-90 or U.S. $15). 3. (1988): Zechstein Reef fossils and their palaeoecology, by n. hollingworth andi. pettigrew. iv-i-75 pp. Price £4-95 (U.S. $8) (Members £3.75 or U.S. $6). 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 -1-241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24-95. © The Palaeontological Association, 1990 Palaeontology VOLUME 33 • PART 1 CONTENTS Preservation of soft-bodied and other organisms by bioimmuration - a review P. D. TAYLOR 1 Bioimmured ctenostomes from the Jurassic and the origin of the cheilostome Bryozoa P. D. TAYLOR 19 Palaeoperidinium cretaceum ; a brackish-water peridiniinean dino- flagellate from the early Cretaceous I. C. HARDING 35 New Permian crinoids from Australia G. D. WEBSTER 49 Cenomanian ammonite faunas from the Woodbine Formation and lower part of the Eagle Ford Group, Texas W. J. KENNEDY and W. A. COBBAN 75 The actinopterygian fish Prohalecites Uom the Triassic of northern Italy A. TINTORI 155 The classification, origin and phylogeny of thecideidine brachiopods P. G. BAKER 175 Teuthid cephalopods from the Lower Jurassic of Yorkshire P. DOYLE 193 Cluster analysis of previously described communities from the Ludlow of the Welsh Borderland P. J. LESPERANCE 209 Latest Cretaceous woods of the central North Slope, Alaska R. A. SPICER and J. T. parrish 225 Orientation of cephalopod shells in illustrations S. STRIDSBERG 243 A new eocrinoid from the Lower Cambrian of Spain G. UBAGHS D. VIZCAINO 249 Printed in Great Britain at the University Press, Cambridge ISSN 0031-0239 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to promote research in palaeontology and its allied sciences. COUNCIL 1990-1991 President '. Professor J. W. Murray. Department of Geology, The University, Southampton S09 5NH Vice-Presidents'. Dr M. Romano, Department of Geology, University of Sheffield, Sheffield S3 7HF Dr P. R. Crowther, City of Bristol Museum and Art Gallery, Queen’s Road, Bristol BS8 IRL Treasurer: Dr M. E. Collinson. Department of Biology, King’s College, London W8 7AH Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne NEl 7RU Institutional Membership Trea.uirer : Dr A. W. Owen, Department of Geology and Applied Geology, University of Glasgow, Glasgow G12 8QQ Acting Secretary : Dr J. A, Crame, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET Circular Reporter : Dr D. Palmer, Department of Geology, Trinity College, Dublin 2 Marketing Manager: Dr C. R. Hill, Department of Palaeontology, British Museum (Natural History), London SW7 5BD Public Relations Officer: Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 IRJ Editors Dr M. J. Benton, Department of Geology, University of Bristol, Bristol BS8 IRJ Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester, Manchester M13 9PL Dr D. Edwards, Department of Geology, University of Wales College of Cardiff, Cardiff CFl 3YE Dr P. D. Lane, Department of Geology, University of Keele, Keele, Staffordshire STS 5BG (co-opted) Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL Dr P. D. Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD Other Members Dr E. A. Jarzembowski, Brighton Dr E. A. Kennedy, Oxford Dr D. M. Martill, Milton Keynes Dr A. R. Milner, London Dr R. A. Spicer, Oxford Overseas Representatives Argentina: Dr M. O. Mancenido, Division Paleozoologia invertebrados, Facultad de Ciencias Naturales y Museo, Paseo del Bosque, 1900 La Plata. Australia: Dr K. J. McNamara, Western Australian Museum, Francis Street, Perth, Western Australia 6000. Canada: Professor S. H. Williams, Department of Earth Sciences, Memorial University, St John’s, Newfoundland AIB 3X5. China: Dr Chang Mee-mann, Institute of Vertebrate Palaeontology and Paleoanthropology, Academia Sinica, P.O. Box 643, Beijing. Dr Rong Jia-yu, Nanjing Institute of Geology and Palaeontology, Chi-Ming-Ssu, Nanjing. France: Dr J.-L. Henry, Institut de Geologie, Universite de Rennes, Campus de Beaulieu, Avenue du General Leclerc, 35042 Rennes Cedex. Iberia : Prof. F. Alvarez, Departamento de Geologia, Universidad de Oviedo, C/. Jesus Arias de Velasco, s/n. 33005 Oviedo, Spain. Japan : Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Tokyo. New Zealand: Dr R. A. Cooper, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt. Scandinavia: Dr R. Bromley, Fredskovvej 4, 2840 Holte, Denmark. U.S.A. : Prof. A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66044. Prof. N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403. Prof M. A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44961. W. Germany : Prof F. T. Fursich, Institut fiir Palaontologie, Universitat, D8700 Wiirzburg, Pliecherwall 1 MEMBERSHIP Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1990 are: Institutional membership . . £60-00 (U.S. $108) Student membership . £1 1-50 (U.S. $20) Ordinary membership . . £28 00 (U.S. $50) Retired membership .... £14-00 (U.S. $25) There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen, Department of Geology and Applied Geology, The University, Glasgow G12 8QQ. 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 H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne NEl 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer. All members who join for 1990 will receive Palaeontology, Volume 33, Parts I 4. Enquiries concerning back numbers should be directed to the Marketing Manager. Non-members may subscribe, and also obtain back issues up to 3 years old, at cover price through Basil Blackwell Ltd, Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford 0X2 ODT, England. For issues before Vol. 29 contact the Marketing Manager. Cover: Framboidal pyrite with lumen of a tracheid of the early land plant, Grosslingia breconensis, from the Lower Old Red Sandstone of South Wales, x 700. See P. Kenrick and D. Edwards. Botanical Journal of the Linnean Society, 97, 95-123. LOWER CRETACEOUS SPIDERS FROM THE SIERRA DE MONTSECH, NORTH-EAST SPAIN ■h by PAUL A. SELDEN 4A / O 'o Abstract. Four new specimens of spiders (Chelicerata : Araneae), from Lower Cretaceous TBerriasian-Valan- ginian) lithographic limestones of the Sierra de Montsech, Lerida Province, north-east Spain, are described, as Cretaraneus vilaltae gen. et sp. nov., Macryphantes cowdeni gen. et sp. nov. (two specimens), and Palaeoulohoms lacasae gen. et sp. nov. All belong to the infraorder Araneomorphae. Palaeoulohoms is the oldest representative of the superfamily Deinopoidea, Cretaraneus is referred to the superfamily Araneoidea, and Macryphantes is the oldest record of the superfamily Araneoidea, family Tetragnathidae (metine-tetragnathine-nephiline group). All three spiders were web weavers; Macryphantes and Palaeouloborus wove orb webs, and may have used a wrap attack to prey on the abundant contemporaneous insect life preserved in the Montsech deposit. Spiders are rare in rocks of Mesozoic age. Only two specimens, Jiiraraneiis rasnitsyni Eskov, 1984, and Jurarchaea zherikhini Eskov, 1987, from the Jurassic of the USSR, have been adequately described. Three specimens from the Lower Cretaceous of New South Wales (Jell and Duncan 1986) and twelve from the Trias of France (Gall 1971 ) were not identified to a taxonomic rank below that of order. Spiders reported from Canadian Cretaceous amber (McAlpine and Martin 1969), the Jurassic and Cretaceous of Siberia and Mongolia (reported in Eskov 1984), and the Cretaceous of Botswana (Waters 1989) await description. The four specimens described here are sufficiently well preserved for their taxonomic affinities to be determined with some precision, and thus they represent only the third find of Mesozoic spiders to be described and named. The fossil spiders described here are placed in extant superfamilies or families, but closer assignation is considered unwise pending the outcome of current studies on living and fossil members of these groups. GEOLOGICAL SETTING Locality. The Sierra de Montsech lies in the foothills of the eastern Pyrenees, between Balaguer and Tremp in Lerida Province, north-east Spain (see Schairer and Janicke ( 1 970) for details and location map). Three quarries in the vicinity of the abandoned village of Rubies, in the eastern part of the sierra, yield remarkable fossils. ‘ La Pedrera de Meia’ and ‘ La Cabnia’ have been worked for many years, and ‘El Reguer’ is currently under investigation. Stratigraphy. The 50 m succession of limestones exposed in the quarries, the ‘Calcaires lithographiques a Plantes et Vertebres de la Pedrera de Rubies’, is a facies development of the 100 m ‘Calcaires a Charophytes du Montsech’ (Peybernes 1976). The deposit has been determined as late Berriasian to early Valanginian in age on the evidence of ostracodes (Peybernes and Oertli 1972; Brenner, Goldmacher and Schroeder 1974; Whalley and Jarzembowski 1985), a conclusion consistent with evidence from palynology (Barale et al. 1984; R. Porter, personal communication 1988). Sedimentology. The Calcaires a Charophytes du Montsech is a sequence of mostly pale, fine- grained, thinly-bedded limestones. Sedimentary structures include laminations, fine grading, minor deformation, and roll marks; trace fossils include arthropod trackways (Schairer and Janicke 1970). The environment of deposition deduced from the sedimentological and palaeontological evidence I Palaeontology, Vol. 33, Part 2, 1990, pp. 257-285, 4 pls.| © The Palaeontological Association 258 PALAEONTOLOGY, VOLUME 33 (see below) is that of coastal lagoons within a large area of algal flats separating marine conditions to the north from the Ebro continent to the south (Barale et al. 1984). The spider-bearing sequence, the Lithographic Limestones, represents a particularly tranquil lacustrine depositional episode (Lacasa and Martinez 1986). History of the fauna and flora. The exceptionally preserved biota of the Montsech lithographic limestone has received attention from palaeontologists since the beginning of this century. In the last few years, renewed interest in the deposit has brought new material to light, including the specimens described here, and avian fossils of phylogenetic importance from a locality of similar age and lithology in the neighbouring province of Cuenca (Sanz et al. 1988). Lists of the fauna and flora were given by Barale et al. (1984), with the most recent summary, especially of the insects, provided by Lacasa and Martinez (1986). All the indigenous microflora is of terrestrial or non-marine aquatic provenance; the macroflora includes a wide range of gymnosperms, progymnosperms, a few ferns and horsetails, and other, unclassified, plants. Animals include ostracodes, few decapod crustaceans, larval and mature insects belonging to eight extant orders, a wide variety of fish, a few frogs and reptiles, and some bird remains. The spiders were first mentioned by Lacasa (1985, p. 228), and a preliminary report of the results presented here was given by Selden (1989). MATERIAL AND METHODS Preservation. The spiders are preserved on thin slabs of pale buff-grey limestone. Grains are not visible in the rock, and the hackly fracture and vitreous appearance under high-power microscopy suggest crystallization from a lime mud. Calcite-filled cracks cross some specimens. The spiders are preserved as pieces of cuticle on the bedding surface. The cuticle is brittle and brown : thicker parts are deep brown and the thinnest cuticle pale buff. The cuticle has not been chemically analysed; it is presumed still to be organic, but probably not of the original composition. The best-preserved parts are visible through a thin layer of translucent limestone, but their morphological details are hazy due to the presence of the overlying matrix. In such instances, 2-4% hydrochloric acid was used, sparingly and with care, to remove the matrix and thus to reveal fine structural details. The spiders are in varying states of original decay: for example in LC 1753 AP the podomeres, of leg 1 in particular, are crowded with subcircular objects along the central parts of the shaft. These objects are interpreted as the decayed remnants of muscles. LC 1754 AP is very poorly preserved: mainly a mould remains, and this is interpreted as a specimen in which decay has progressed further. Both part and counterpart of specimens LC 1753 AP and LC 1754 AP are preserved, but only the part of specimens LC 1150 lEI and LP 1755 AP was collected. Specimens LC 1753 AP B, LC 1754 AP A, LC 1 150 lEI, and LP 1755 AP represent lower slabs preserving mainly ventral features of the specimens; LC 1753 AP A and LC 1754 AP B are upper slabs with dorsal features. However, splitting of the rock has not resulted in perfect separation of dorsal and ventral, and due to the mode of preservation within the limestone (described above), most of both dorsal and ventral parts are preserved on LC 1753 AP B, and LC 1 150 ILL LC 1754 AP is mainly an external mould with little cuticle remaining, but on LP 1755 AP ventral parts of the body and both dorsal and ventral sides of the distal parts of the legs are preserved. On all specimens, the legs and abdomen (when present) are crushed flat. The carapace has suflficient convexity to produce relief in the fossils, so that on LP 1755 AP, for example, the carapace shape can be determined by the relief of the fossil, and the shapes of the sternum and coxae are outlined by setae and cuticle. The male palps of LC 1753 AP, LC 1754 AP, and LC 1 1 50 lEI appear to occupy depressions in the matrix. This is probably because they were bulbous structures in life and therefore created a concavity in the sediment into which the palp collapsed during burial. The spiders were studied under a Wild M7S stereomicroscope, with the specimens immersed in ethanol or glycerine to enhance their contrast against the pale rock background. Camera lucida drawings were made and photographs were taken under the same conditions. In addition, a Zeiss photautomat microscope was used, in reflected light mode with oil immersion objectives, to view SELDEN; SPANISH CRETACEOUS SPIDERS 259 and photograph details at higher magnifications. The computer program MacClade, version 2.1 (Maddison and Maddison 1987) proved very useful for exploring relationships. Terminology. Setal terminology is somewhat problematical, since different authors have used the same terms in different ways. Furthermore, there is a complete gradation in size between setae (small, and may be short or long, thick or thin), bristles, and spines (large). Macroseta is a term used, for example by Opell (1979), to describe a large seta which could equally be called a small spine or bristle. The common hair-covering of Cretaraneus is the serrate seta, which is smooth apart from a few minute accessory spines which are no longer than the mean thickness of the seta (see Lehtinen 1967, fig. lOu; Kullmann 1972, fig. 7). Two fairly distinct setal types are commonly called plumose (e.g. by Forster and Wilton 1973; Forster and Blest 1979). The first, which is here called plumose, is generally thicker than the serrate seta, and has helical ridges bearing small accessory spines which are much shorter than the width of the seta (see Lehtinen 1967, fig. 8; Kullmann 1972, fig. 8). The second, which is here called feathery, is fine, smooth, and has long accessory branches which are much longer than the width of the seta (see Lehtinen 1967, fig. 9). Many other types of seta and spine exist; there may be complete gradations between them, and the extent of their phylogenetic importance is unknown. Terminology of the sclerites of male palps differs between workers in different groups of spiders due to a lack of understanding of the homologies between the sclerites. Thus the task of recognizing palpal sclerites in fossil spiders is problematical. Useful descriptions of the constituent parts of male palpal organs are found in Comstock (1948), Levi (1961), Merrett (1963), and Millidge (1977). In leg formulae (e.g. 1243), the leg lengths are ranked in order longest (first) to shortest (last). Abbreviations used in the text and text-figures are as follows: ab abdomen 1 labium s serrula bo book-lung operculum Ip left palp St sternum ca cephalic area m maxilla t tegulum cal calamistrum m a median apophysis ta tarsus ch chelicera mt metatarsus ti tibia cx coxa pa patella ti a tibial apophysis e embolus pc paracymbium tr trochanter f fovea pe pedicel fe femur rp right palp Provenance and depository. Three of the fossil specimens, LC 1 150 lEI, LC 1753 AP, and LC 1754 AP, come from the quarry of La Cabnia, the fourth, LP 1755 AP, is from the locality of La Pedrera de Meia. Exact stratigraphical provenance is not known, but both of these localities are in the same 50 m sequence of lithographic limestones, the Calcaires lithographiques a Plantes et Vertebres de la Pedrera de Rubies, described above. The specimens are deposited in the Institut d’Estudis Ilerdencs, Lerida. Preserved specimens of extant spiders were studied for comparative purposes, and in addition to those in the author’s collection of mainly British species, the following specimens were examined. Uloboridae: Hyptiotes flavidus, female, Funchal, Madeira, M. J. Jones Collection No. 119, Manchester Museum; Ulohorus walckenaerius, male and female, Chobham, Surrey, D. W. Mackie Collection No. G4999, Manchester Museum; Philoponella sp., male and female. Lake Naivasha, mature and immature males, Nairobi, Kenya, J. Murphy Collection Nos 1302, 1363. Deinopidae: Deinopis guianensis, female, British Guiana, British Museum (Natural History) (BM(NH)) No. 1939.3.24.228; Deinopis staimtoni, female, Durban, South Africa, BM(NH) No. 1903.8.20.1; Deinopis sp., female and immature, Kilifi, Kenya, J. Murphy Collection; Mennens canielus, females (types), Durban, South Africa, BM(NH) No. 1903.7.10.22; Mennens sp., male, Kitale, and female, Nairobi, Kenya, J. Murphy Collection; Avella angulata, female, Gayndah, Australia, BM(NH) No. 1919.9.18.5732; Avella de.spiciens, female, Sydney, Australia, BM(NH) No. 1919.9.18.5733. 260 PALAEONTOLOGY, VOLUME 33 Dictynidae: Aebutina binotata, immatures, Aguas Negras, near Tarapuy, Napo, Ecuador, British Museum (Natural History) Arachnid Collection. SYSTEMATIC PALAEONTOLOGY Order araneae Clerck, 1757 Suborder opisthothelae Pocock, 1892 Infraorder araneomorphae Smith, 1902 Remarks. The phylogenetic scheme for the higher classification of spiders which is currently widely accepted originated with Pocock (1892). In this scheme, the Liphistiomorphae (Mesothelae) is the sister group to the Opisthothelae (MygalomorphaeT Araneomorphae). The morphology of liphistiomorphs is distinctive (Platnick and Gertsch 1976, fig. 5). The fossil spiders described here differ markedly from this pattern, lacking one or more of the following liphistiomorph synapomorphies : posteromedian invaginations of the fourth coxae, seven or eight forwardly situated spinnerets, sclerite of the first abdominal segment (Platnick and Gertsch 1976), and labium as wide as or wider than the sternum (Raven 1985). Thus they are referred to the Opisthothelae. Mygalomorph synapomorphies include the following external morphological features: cheliceral fang which operates in the sagittal plane (orthognathous), complete loss of expression of the anterior median spinnerets, three or four articles in the posterior lateral spinnerets (Platnick and Gertsch 1976), labial and maxillary cuspules, and reduced male palpal sclerites (Raven 1985). The alternative character states typify the Araneomorphae, and it is to this latter group that the Montsech spiders are referred. In particular, the chelicerae of Cretaraneus gen. nov. are labidognathous (i.e. they operate transversely); the male palp of Macryphantes gen. nov. is more complex than the mygalomorph palp; and Palaeouloborus gen. nov. possesses a calamistrum and a cribellum; the latter is a homologue of the anterior median spinnerets. Neither labial or maxillary cuspules nor three- or four-segmented posterior lateral spinnerets are known in any of the Montsech spiders, and each of them shows greater affinity to araneomorph families than to mygalomorphs. Assignment of the Cretaceous spiders to lower taxonomic categories presents greater difficulty. A competent araneologist can place most living spiders seen in the field into a family with a high degree of certainty. Even in the absence of field characters, such as web type and habitat, spider familial keys are relatively straightforward, only requiring the use of a binocular microscope and normally no dissection. Spider families are diagnosed on unique combinations of morphological characters such as number and position of eyes, spinneret pattern, and number of tarsal claws. However, research has shown that many of the characters on which the families are based are plesiomorphic at higher taxonomic levels, convergent with other groups, or are ‘loss’ characters (Lehtinen 1978). A classic example of this is the recognition of the cribellum as a synapomorphy for Araneomorphae : its presence in lower categories is merely the expression of the plesiomorphic state, and its absence illustrates convergence of the apomorphy in many araneomorph families. Whilst unique combinations of plesiomorphies may work as familial descriptions in practice, they cannot reflect true relationships between families. Of necessity, therefore, recent cladistic analyses have sought new or previously overlooked characters as apomorphies. These characters are commonly behavioural or anatomical, and may be poorly known in groups outside those under particular study. More important in the present work, such characters are most unlikely to be preserved in the fossil record. An additional concern encountered when working with rare fossils which have Recent relatives arises from the possibility that the fossils may be ancestors of Recent species. Should the fossils be classified in the same scheme as their Recent relatives, or in a separate scheme? Such problems have been discussed by Crowson (1970), Farris (1976), Wiley (1981) and others. A useful device for classifying fossils with their Recent relatives is the plesion (Patterson and Rosen 1977); this presupposes, however, that the Recent classification is soundly based on shared derived characters. SELDEN: SPANISH CRETACEOUS SPIDERS 261 In spider systematics this is not yet true. In this paper, the new taxa are classified as close as possible to their presumed position in the Recent classification. Inevitably this entails placement within taxa yet to be defined by synapomorphies or defined on shared derived characters which cannot be seen easily in fossils. Because these are among the first Mesozoic spiders to be described, and because of the paucity of diagnostic features in the fossils, and the present lack of knowledge of synapomorphies in living families, the fossil spiders are assigned to superfamilies, and in one case to a family, but to no extant lower taxa. (Note that among Cretaceous insects, classification to modern genera is not unusual.) More specimens of Mesozoic spiders which are becoming available for study may help to elucidate further the taxonomic positions of the specimens described here, and assist in unravelling the complexities of spider phylogeny and evolution. The ‘consensus' spider classification scheme given in Shear (1986/?) is followed here in general, but with discussion concerning the status of groups assigned to the Araneidae following recent work by Coddington (1989r/, /?, 1990). Superfamily deinopoidea Koch, 1851 Remarks. This superfamily consists of the two cribellate families Uloboridae and Deinopidae, which weave orb webs (most Uloboridae) or spin modified orb webs which are then thrown, retiarius-\\V.Q, at their prey (Deinopidae). Much more is known about uloborids than deinopids, though work on the latter is currently in progress. Although these two families have been considered closely related for nearly a century (Simon 1892), Coddington (1986, p. 359), with reference to his cladogram (p. 358), remarked that; ‘the monophyly of the uloborid-deinopid lineage is based on only three characters, primarily because so little is known about deinopids. ’ The three characters Coddington accepted as deinopoid synapomorphies are; puffed cribellate silk, pseudoserrate plumose hairs, and fourth-tarsal macrosetae; however, he qualified this by suggesting that the first two charaeters may actually be synapomorphies for all orb-web weavers (Deinopoidea and Araneoidea) (Coddington 1986, pp. 327, 359). Recent work by Coddington (1989a, h, 1990) has resolved the apparent trichotomy between Araneoidea, Uloboridae, and Deinopidae (reported in Platnick 1986); Deinopoidea and Araneoidea being seen as sister groups in a monophyletie group of orb-web weavers; the Orbiculariae Walckenaer, 1802. Palaeoulohorus gen. nov., described below, possesses three tarsal claws with accessory claws typical of web weavers and lacks the tarsal adaptations of the superfamilies Thomisoidea, Philodromoidea, Lycosoidea, Clubionoidea, Salticoidea, and Pholcoidea. The fossil genus also lacks the synapomorphies of Scytodoidea (Lehtinen 1986), Dysderoidea (Forster and Platnick 1985), Palpimanoidea (Forster and Platnick 1984), Hersilioidea, Hypochiloidea (Platnick 1977), Eresoidea, and Agelenoidea. Dictynoidea are cribellates, and the deinopoid families have, at one time or another, been referred to this superfamily, and to the family Dictynidae in particular. However, dictynids are generally small spiders, with short legs of approximately equal length, and lack femoral trichobothria (see below); thus they are quite unlike Palaeoulohorus. The charaeters whieh are most useful in placing Palaeoulohorus are; femoral trichobothria on legs 2, 3, and 4, leg 1 more than five times the length of the carapace and more than twice the length of leg 3, all leg tarsi with large accessory claws and apparently non-pectinate paired claws, presence of calamistrum and cribellum, and presence of plumose hairs. Femoral trichobothria occur in only two groups of araneomorph spiders; the metine-tetrag- nathine lineage of the superfamily Araneoidea, and the family Uloboridae (Opell 1979; Lehtinen 1980; Levi 1980, 1981). In both of these groups there is great disparity in length between the elongate anterior legs (1 and 2) and the short third pair (PI. 2, fig. 9), a feature also found in some other Araneoidea (Argiopinae, Levi 1983). Palaeoulohorus is cribellate, and since only the Deinopoidea, but not the Araneoidea, are cribellate, this genus must be referred to the former superfamily. The calamistrum of Palaeoulohorus is situated in a curved depression on the superior surface of the fourth metatarsus. Such a curvature occurs in uloborids (PI. 2, fig. 10), to a much lesser degree in deinopids (Shear 1986a), and also in Aehutina, a poorly known genus tentatively referred to the Dictynidae (only females and immatures of this genus are known). Additionally, 262 PALAEONTOLOGY, VOLUME 33 Palaeoulohorus bears plumose setae, which are found in Deinopoidea but not Araneoidea. The characteristic tarsal macrosetae of deinopoids (see below) appear to be absent from Palaeoulohorus. Palaeoulohorus is referred to the superfamily Deinopoidea, on the evidence given above. The fossil clearly resembles members of the family Uloboridae more closely than the Deinopidae; the latter family has many autapomorphies (e.g. forwardly directed, enlarged, posterior median eyes, elongate legs all of a similar length, web-throwing) and lacks the femoral trichobothria typical of the Uloboridae and the fossil genus. The position of Palaeoulohorus within the Deinopoidea is now discussed. In a recent revision of the Uloboridae, Opell (1979) gave the following characters as diagnostic of the family: (1) lack of poison glands, (2) cribellate orb-web weavers, (3) femoral trichobothria, and (4) row of macrosetae (short spines) on metatarsus and tarsus of leg 4. The tarsal spines cannot be a synapomorphy for the Uloboridae since they also occur in Deinopidae (see below). Coddington (1986) added characters of the silk-glands described by Kovoor (1977), and some behavioural traits, to the list of uloborid synapomorphies, but omitted femoral trichobothria. Coddington’s cladograms (1986, p. 358; 19896, fig. 108) show the metine-tetragnathines, which also have femoral trichobothria, far from the dichotomy of Araneoidea with Deinopoidea, which presumes that femoral trichobothria are a convergent phenomenon in uloborids and metines-tetragnathines. The alternative hypothesis (that they are a synapomorphy for all orb-web weavers) would require their loss in many separate lines. The row of tarsal and metatarsal macrosetae of uloborids was considered to be a synapomorphy of the family by Opell (1979), but in a later paper, Opell (1982) mentioned finding a poorly developed row on leg 4 of deinopids as well. My own observations confirm that macrosetae are present on the inferior surfaces of the distal half of the metatarsus and the tarsus of leg 4, and to a lesser extent on leg 3, in the deinopids Deinopis, Menneus, and Avella. These macrosetae are rather similar to the comb of serrate bristles seen in the Theridiidae in a similar position on the legs. However, they differ from theridiid bristles in being plumose, not serrate. As in the theridiids, they may need to be searched for, since they blend into the general hirsuteness of the tarsus. The macrosetae are not greatly different from the curved bristles normally present on the inferior surface of the distal end of the tarsus. Also, as in theridiids, they are not strictly confined to mt4 and ta4, also being present on leg 3, and they vary from species to species. In all the deinopids I studied they were quite unlike the comb of short, upstanding spines of the uloborid Zosis geniculatus illustrated by Opell (1979, pi. 1, figs. A and C). Rather, they resemble the row of macrosetae of the uloborid Hyptiotes cavatus figured by Opell (1982, pi. 1, fig. C). The similarity between theridiids and deinopids in this feature is presumably due to convergence in their prey-wrapping strategies rather than synapomorphy. Clearly, a comb of macrosetae on the fourth leg is a derived character of uloborids and deinopids which is not present in Palaeoulohorus. The legs of uloborids bear fine feathery setae amongst the normal setae; neither deinopids nor Palaeoulohorus bear them. Palaeoulohorus cannot be included in the family Uloboridae because it has neither feathery setae nor fourth tarsal macrosetae. The fossil lacks the many specializations of the deinopids, and in addition the fourth tarsal macrosetal comb is absent. Rather than redefining the family Uloboridae to accommodate the fossil genus, it is left here within the superfamily Deinopoidea, closer to the Uloboridae than the Deinopidae, but not placed in either family. This placement indicates that the loss of a fourth tarsal macrosetal comb is autapomorphic for the fossil genus. Genus palaeouloborus gen. nov. Derivation of name. Greek, palaios, old, and the living genus, Uloborus, which the fossil genus resembles. Type and only known species. Palaeouloborus lacasae sp. nov. SELDEN: SPANISH CRETACEOUS SPIDERS 263 Diagnosis. Deinopoid with ovate carapace bearing marked break of slope separating anterior of carapace from sloping posterior area; leg 1 more than five times length of carapace and more than twice length of leg 3; many trichobothria on superior ?retrolateral surface of femur of leg 2 and superior ?prolateral surfaces of femora of legs 3 and 4; paired tarsal claws small, without teeth, median claw long, without teeth, pair of large accessory claws; superior surface of metatarsus of leg 4 in gentle S-shape, proximally convex and then concave, straightening out about half-way along podomere, bearing calamistrum which becomes row of curved bristles towards distal end of podomere; plumose setae present. Row of macrosetae absent from fourth tarsus. Palaeouloborus lacasae sp. nov. Plate 1 ; Plate 2, figs. 1-5, 7, 8, 10; text-fig. 1 Derivation oj name. After Antonio Lacasa-Ruiz, palaeontologist at the Institut d’Estudis llerdencs, Lerida, Spain. Type specimen. Holotype LP 1755 AP, from the quarry of La Pedrera de Meia, Sierra de Montsech, north-east Spain, and held in the collections of the Institut d’Estudis llerdencs, Lerida, Spain. Diagnosis. As for the genus. Description. The carapace shape is determined from fragments of cuticle and from the relief of the matrix. The scraps of cuticle preserved in the carapace region mainly represent coxae and sternum. However, on the left side between legs 2 and 3, cuticle of the carapace edge can be seen lying in a depression caused by pressure of the edge into the matrix. Eaint traces of cuticle from the posterior edge of the carapace can also be discerned between the tarsi of legs 3. These cuticle remnants and the faint depression formed by the carapace margin suggests an ovate carapace with a truncated anterior margin. The carapace is T73 mm long, and 1-50 mm wide. It is widest just posterior to its midpoint. There does not appear to be a well defined cephalic area, nor an obvious fovea. There is a marked transverse break of slope just posterior to the widest part of the carapace and separating the raised foveal region from the backwardly sloping posterior part of the carapace. This slope is found in some living spiders (e.g. the uloborid Philoponella, PI. 2, fig. 8), and accommodates an abdomen which extends forwards beyond the pedicel. Eyes not seen, no obvious tubercles. The chelicerae are large, 0 67 mm long, and 0-33 mm wide, and somewhat forwardly directed, with parallel sides. The fangs cannot be seen, since the anterior edges of the chelicerae are obscured by overlying matrix. The palps are clothed with setae, and short spines occur on the superior surface of the tarsus. The tarsus is oval, indicating that it was tumid in life. The distal end is not seen, so the presence of a claw cannot be confirmed. The specimen is therefore not a mature male, and could be an immature or a female. Very little of the basal parts of the palp can be seen, but superimposed on the anterior part of the carapace area is a dark line which, under high magnification, is seen to be serrate (PI. 2, fig. 1 ). This is interpreted as the serrula of the left maxilla; cuticle is absent where the serrula of the right maxilla would have been preserved. The leg formula is 1243. The coxae measure approximately 0 58 mm long, and the trochanters 018 mm. Lengths of the more distal podomeres, in mm, are as follows: leg 1 : fe 3-27, pa 0-77, ti 1-64, mt 2-50, ta 0-96, total 9T4; leg 2: fe 21 1, pa 0-48, ti 1-35, mt 1-44, ta 0-77, total 615; leg 3: fe L35, pa 048, ti 106, mt 048, ta 0-48, total 3-85; leg 4: fe L64, pa 0 48, ti 0-87, mt L25, ta 0 52, total 4-76. The legs are clothed with setae of the plumose type. Under high magnification, they are seen to bear a sculpture of striations arranged in a helical pattern, with abundant, minute serrae (PI. 2, fig. 7). No feathery setae can be seen on the legs. The femora are only sparsely setose, setae and bristles becoming more abundant on distal parts of the legs. Groups of trichobothria are certainly present on the femora of legs 2-A, and on the tibia of leg 4. These trichobothria are not feathered (PI. 2, fig. 2). Isolated trichobothria are more difficult to see, and their presence elsewhere on the legs cannot be confirmed. The leg spines are not large, and because only part of the specimen is preserved, the numbers of spines given below are not the maximum number which may be present on the legs. All tarsi are spineless, and bear two small, non-pectinate paired claws, a large median claw which appears to be non-pectinate or if pectinate then with only minute teeth, and two large accessory claws (PI. 2, fig. 3). Eel bears few setae, mainly on the inferior surface (where they are curved) and on the superior surface, especially proximally and distally. Trichobothria may be present on fel (two possible trichobothrial bases can be seen on fel on the right side) but cannot be confirmed. Eel and pal are spineless. TEXT-FIG. 1. Palaeouloborus lacasae gen. et sp. nov., holotype, LP 1755 AP. Explanatory drawing for PI. 1. See Terminology for explanation of abbreviations. EXPLANATION OF PLATE 1 Palaeouloborus lacasae gen. et sp. nov., holotype; Lithographic Limestone, Lower Cretaceous, Sierra de Montsech, Lerida Province, Spain. LP 1755 AP, whole specimen, lower slab, under ethanol, x 21 . See text-fig. 1 for explanation. PLATE I SELDEN, Palaeouloborus 266 PALAEONTOLOGY, VOLUME 33 Til bears four inferior spines and three superior spines along the shaft of the podomere, one proximal lateral spine, one lateral subdistal spine, and one superior subdistal spine. Mtl bears three inferior and two lateral spines down the shaft of the podomere. Proximally, mtl bears one superior, one lateral, and one inferior spine, and two inferior distal spines are present. Fe2 is spineless and bears few setae, principally on the inferior and superior surfaces. Both second femora are poorly preserved, but a few trichobothrial bases can be seen proximally, apparently on the retrolateral surface. Pa2 is spineless. Ti2 bears bristles proximally and along the inferior surface. Ti2 bears two inferior distal spines, one lateral subdistal spine, one superior subdistal spine, and one superior distal spine. More spines are probably present on ti2 but on both sides of the specimen this podomere is crossed by the large femur of leg 1, which obscures the mid-sections of ti2. Mt2 is very setose, and bears bristles, especially along the inferior surface. Mt2 bears superior and inferior proximal spines, two median inferior spines, and one lateral and two inferior distal spines. Fe3 bears few setae, one tiny spine laterally, and many trichobothria (probably 20-40 in life) over a large area of the ?prolateral surface. Pa3 is spineless. Ti3 is poorly preserved on both sides of the specimen, but bears many bristles. Mt3 bears two short inferior spines and one lateral distal spine. Fe4 is spineless, bears strong, curved bristles distally, and many trichobothria (as many as on fe3) over a large area of the superior, ?prolateral surface. Pa4 bears large bristles laterally. Ti4 bears one superior proximal spine, and prolateral, retrolateral and superior spines subdistally. Ti4 bears about four trichobothria on the proximal superior surface, and long, curved bristles distally. The superior surface of mt4 follows a gentle S-shape, proximally convex and then concave, before straightening out about half-way along the podomere. The superior surface of mt4 bears a calamistrum composed of curved setae, apparently in one row, running from the proximal end of the podomere for about two-thirds of its length, where it passes indistinctly into a row of curved bristles which continues to the distal end of the podomere (PI. 2, figs. 5 and 8). Mt4 bears two short inferior median spines, at least two very short spines inferodistally, and large bristles distally. The abdomen measures 3-67 mm long, and 2-89 mm wide. It is ovate, wrinkled posteriorly, and compressed to the right, indicating that it was globose in life. The abdomen is sparsely setose, the setae becoming thicker, but not longer, posteriorly, where they show their plumose structure under high magnification (PI. 2, fig. 7). A pair of subtriangular areas of darker cuticle, each with a small dark patch anteriorly, is situated at the anterior end of the abdomen. These are interpreted as book-lung opercula. Three pairs of spinnerets are visible in the posterior half of the abdomen. They are compressed to the right. Their position indicates that they were not terminal, but ventral, in position in life. A recurved line immediately anterior to the anterior pair of spinnerets represents the cribellum. Little detail can be discerned because the cribellar plate is not preserved (this is presumably on the counterpart), only the fold of cuticle anterior to the plate. (In living spiders the cribellum is commonly invaginated into a fold in front of the spinnerets when not in use, see PI. 2, fig. 6.) Along this fold there are numerous short, blunt setae of a type not seen elsewhere on the spider (PI. 2, fig. 4). Superfamily araneoidea Latreille, 1806 Remarks. Cretaraneus gen. nov. and Macryphantes gen. nov., described below, are assigned to this superfamily. Both genera lack a calamistrum and cribellum. This does not, by itself, exclude them EXPLANATION OF PLATE 2 Figs. 1-5, 7, 8. Palaeouloborus lacasae gen. et sp. nov., holotype; Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lerida Province, Spain, LP 1755 AP, oil immersion. 1, Serrula of left maxilla, bristle at left end, x 240. 2, Femoral trichobothria, x 150. 3, Claws on distal end of tarsus of left leg 4; note long median claw and accessory claws, x 100. 4, Spinnerets : anterior pair (at top), median pair, and posterior pair (part); recurved line of short setae (seen at left) in front of anterior spinnerets mark approximate position of cribellum; compare with fig. 6, x 85. 5, Proximal part of metatarsus of right leg 4, showing curvature of superior surface with calamistrum; note long, curved bristles at distal end of tibia (bottom left); compare with fig. 10, X 60. 7, Plumose seta, x625. 8, Distal end of metatarsus of left leg 4, showing calamistrum (overlying tarsus of right leg 4, on right) terminating in row of curved bristles to left, x 85. Figs. 6, 9, 10. Philoponetla sp.. Lake Naivasha, Kenya; J. Murphy Coll. No. 1363, under ethanol. 6, Ventral view of posterior end of abdomen of immature male, showing spinnerets and cribellum in front, x 36. 9, Left lateral view of immature male, x 13. 10, Metatarsus of right leg 4 of mature female, showing curvature of superior surface with calamistrum, x 50. PLATE 2 SELDEN, Palaeouloborus, Philoponella 268 PALAEONTOLOGY, VOLUME 33 from a cribellate superfamily because mature males of many living cribellate genera commonly abandon web weaving with the concomitant degeneration of the cribellum and calamistrum, and the preserved specimens of both genera are mature males. However, evidence is presented below that Cretaraneus and Macrypliantes show greater resemblance to araneoids than to any cribellate group. Tarsal adaptations (two claws, scopulae, tarsomeres, etc.) for cursorial, saltatorial, and other specialized locomotory habits, as found in the superfamilies Lycosoidea, Clubionoidea, Philo- dromoidea, Salticoidea, Thomisoidea, and Pholcoidea, are not found in Cretaraneus or Macrypliantes, so referral of the fossil genera to any of these superfamilies is rejected. Cretaraneus and Macrypliantes also lack the synapomorphies of Scytodoidea (Lehtinen 1986), Dysderoidea (Forster and Platnick 1985), Palpimanoidea (Forster and Platnick 1984; but see Archaeidae, below), and the superfamilies Hersilioidea, Dictynoidea, Hypochiloidea, and Agelenoidea. Until recently, few strong synapomorphies could be mustered to support the monophyly of the superfamily Araneoidea (Coddington 1986). Those characters suggested by Coddington (1986) as araneoid synapomorphies (aggregate and flagelliform glands, paracymbium, serrate hairs, web construction technique) are not only difficult or impossible to see in fossils but also have yet to be thoroughly checked in many extant spiders both within and outside the superfamily. The problem of defining the Araneoidea was highlighted by Millidge (1988), in a discussion of the position of the Linyphiidae. He pointed out that the term ‘paracymbium’ covers a number of different morphological features on male palps in araneoid and some non-araneoid families, so this character should not be treated as a synapomorphy for the Araneoidea. He also argued that the sticky microdroplets present on the webs of some linyphiids are not produced by the same type of gland (aggregate glands) as in other araneoids, but may be more closely related to the sticky microdroplets found on agelenid webs. Millidge (1988) concluded (on the basis of other evidence as well as that reported here) that the Linyphiidae should be removed from the Araneoidea, and that the superfamily itself may be an unnatural grouping of families which merely share a lack of synapomorphies of other superfamilies. Millidge’s arguments have not been accepted by Coddington (1990). Recent work by Coddington (1989a, h, 1990) has amassed a great deal more evidence supporting the monophyly of Araneoidea, including characters of the labium and the spinnerets. Cretaraneus and Macrypliantes are assigned to the superfamily Araneoidea. Cladistic analyses of the families within the Araneoidea were attempted by Heimer and Nentwig (1982) and Coddington (1986, 19896, 1990). Shear (19866) included the following major families in the Araneoidea: Theridiidae, Nesticidae, Linyphiidae, Araneidae (including Nephilinae, Metinae and Tetragna- thinae), Theridiosomatidae, Symphytognathidae, Mysmenidae, and Anapidae. The latter four have synapomorphies (Forster and Platnick 1977; Platnick and Shadab 1978a, 6; Coddington 1986) which are seen in neither Cretaraneus nor Macrypliantes, so these families can be discounted. The familial status of Tetragnathidae is discussed below, under that family. The placement of Cretaraneus will be discussed first, followed by that of Macrypliantes. The following characters of Cretaraneus suggest the superfamily Araneoidea : broad, pyriform carapace lacking a distinct fovea, presence of a raised cephalic area, globose abdomen, three foot- claws with associated serrate bristles, serrate hairs, lack or paucity of trichobothria, paracymbium on the male palp, labium wider than long, and spinnerets in a compact group. Members of the families Theridiidae and Nesticidae possess a comb of serrate setae on the inferior side of the fourth tarsus. Such a feature cannot be seen on Cretaraneus, but since the serrate setae are not always present, or not obviously serrate, in smaller species of living theridiids, the lack of this feature in the fossil genera does not necessarily exclude them from the Theridiidae. Levi and Levi (1962) gave as a diagnostic character for the Theridiidae, chelicerae with up to three teeth on the outer margin and rarely one to three teeth or denticles on the inner margin ; Cretaraneus has more cheliceral teeth than this. One feature of Cretaraneus which suggests a link with the Theridiidae (but not the Nesticidae) is the labium which appears not to be rebordered. Palpal characters in Cretaraneus are not sufficiently distinct to suggest any particular araneoid family; although the simplest palps in the Araneoidea occur in the Theridiidae (Levi 1961), and the SELDEN: SPANISH CRETACEOUS SPIDERS 269 Cretaraneus palp appears to be relatively simple for an araneoid, palps are variable within spider families. Therefore Cretaraneus, though apparently lacking a comb of serrate setae on the fourth tarsus, cannot be excluded with confidence form the Theridiidae. Stridulatory ridges are commonly present on the chelicerae of male linyphiids, but may be obscure in some species, and the labium is usually rebordered. Neither stridulatory ridges nor a rebordered labium is seen in Cretaraneus. A link between Cretaraneus and linyphiids is provided by the raised cephalic area, a feature common in linyphiid males. However, a presumed lateral condyle on the chelicera and accessory tarsal claws (serrate bristles) are features not found in the Linyphiidae, but characteristic of the Araneidae. The rotation of the male palpal sclerites in some Araneidae mentioned by Levi (1983) cannot be confirmed in Cretaraneus because of the rotation produced during fossilization, although the general appearance of the palp in the fossil genus is not reminiscent of the araneid palp. The apparent presence of a median apophysis in the palp of the fossil may provide information on its placement, but at present this character is unresolved (Coddington 1989/?). The wide labium, presumed cheliceral condyle, and accessory tarsal claws of Cretaraneus are characteristic of most araneoids (Levi 1980). The rather large, forwardly directed chelicerae, and relatively simple palp of Cretaraneus, suggest the metine-tetragnathine lineage, but the fossil lacks other characteristics of this group, and mature males of some linyphiids also possess enlarged chelicerae. Male palps with a superficially simple appearance are characteristic of the genus Nephila (Schult 1983), but Cretaraneus bears few other similarities to the genus. The Jurassic spider Juraraneus rasnitsyni Eskov, 1984 was placed in a monospecific family, the Juraraneidae Eskov, 1984. Eskov (1984) defined the family on a unique combination of araneoid characters and could find no apomorphies for the family. Cretaraneus resembles Juraraneus in many ways, including: overall size and shape; leg (but not tarsal) lengths; approximate shapes of sternum, labium, and maxillae; general shape of chelicerae. Cretaraneus differs from Juraraneus in lacking the irregular group of denticles on the inner margin of the chelicera, and the palpal characters interpreted by Eskov (1984) as a large, hook-like paracymbium, a large, pointed median apophysis, and a long, straight conductor (= embolus?). The family Archaeidae is included in the Araneoidea by many arachnologists but, in a radical revision of archaeids and some other small families (for example Mimetidae, previously always placed in Araneoidea), Forster and Platnick (1984) removed them to the Palpimanoidea. They also created the monogeneric families Pararchaeidae and Holarchaeidae for genera previously included in the Archaeidae. Forster and Platnick (1984, p. 99) proposed two synapomorphies for the superfamily Palpimanoidea: cheliceral peg-teeth (modified setae as opposed to cuticular teeth), and an elevated cheliceral gland mound. They also mentioned that peg-teeth are found in some unrelated thomisoid and scytodoid genera, as convergent phenomena, and have been secondarily lost in members of six families assigned by them to the palpimanoids. Cretaraneus possesses true teeth on the chelicerae, but is mentioned here because of some similarities with the Pararchaeidae. The Pararchaeidae differ from the other palpimanoids in having serrate, rather than plumose, hairs, and show similarity with Cretaraneus in the enlarged chelicerae with a prominent keel and large bristles, the pectinate paired foot-claws, uncinate median claw, serrate bristles and lack of an onychium on the tarsus. A cheliceral keel is also found in other groups, for example the Leptonetidae (Gertsch 1974). The male palp of pararchaeids has a strongly developed embolus and a large tegular plate (Forster and Platnick 1984, p. 70), features also apparent in the palp of Cretaraneus. It is possible, therefore, that some relationship exists between Cretaraneus and the Pararchaeidae, which may or may not be an araneoid family. An archaeid spider has been described from the Jurassic (Eskov 1987). Since it seems impossible to refer Cretaraneus to an araneoid family, there are two available options. First, a new, monospecific family could be defined to accommodate Cretaraneus. This course of action would be difficult, given the lack of specialized features displayed by the fossil, and would not provide any additional phylogenetic information. The second option, and the one chosen here, is to leave the genus unplaced within the superfamily Araneoidea. It is possible that future 270 PALAEONTOLOGY, VOLUME 33 discoveries will reveal that Juraraneus and Cretaraneus belong to the same group of Mesozoic araneoids. Genus cretaraneus gen. nov. Derivation of name. Latin creta, chalk, and araneus, a spider. Type and only known species. Cretaraneus vilaltae sp. nov. Diagnosis. Araneoid spider with subelliptieal carapace bearing raised cephalie area and no fovea; subtriangular sternum; small, subtriangular labium; serrate setae covering all parts of body. Chelicerae relatively large (04 x length of carapace), forwardly directed (at least in adult male), with inner and outer row of denticles (not peg-teeth), and mesal ridge; male palp with long embolus, and small, proximal ?paracymbium ; legs relatively equal in length, about three times the length of carapaee; femora, tibiae and metatarsi with spines; tarsi with pectinate paired claws, small median claw, and associated serrate bristles; no true trichobothria ; globose abdomen. Cretaraneus vilaltae sp. nov. Text-figs. 2-4 1986 Araneae: Lacasa and Martinez, p. 218; pi. 2, fig. 1. TEXT-FIG. 2. Cretaraneus vilaltae gen. et sp. nov., holotype, LC 1 150 lEI. Explanatory drawing for text-fig. 3. See Terminology for explanation of abbreviations. SELDEN; SPANISH CRETACEOUS SPIDERS 271 Derivation of name. After Sr Ramon Vilalta-Oliva, President of the Institut d'Estudis Ilerdencs, Lerida. Type specimen. Holotype and only known specimen, LC 1150 lEI, complete specimen on single piece of limestone from quarry of La Cabrua, Sierra de Montsech, north-cast Spain; held in collections of Institut d'Estudis Ilerdencs, Lerida, Spain. Diagnosis. As for the genus. Description. A well-preserved spider, and the smallest of the specimens known from Montsech. The carapace cuticle is preserved and is golden brown in colour. The carapace is 1-73 mm long and E37 mm wide, and pyriform; its greatest width occurs at four-fifths of the length behind the anterior margin. Erom greatest width THXT-FiG. 3. Cretaraneas vilaltae gen. et sp. nov., holotype. Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lerida Province. Spain, LC 1 150 ILL See text-fig. 2 for explanation, x 13. 272 PALAEONTOLOGY. VOLUME 33 TEXT-FIG. 4. Cretanineits vilaltae gen. et sp. nov., holotype. Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lerida Province, Spain, LC 1 150 lEI. a. tarsal claws on left leg 2; note large, pectinate, paired claws, small, uncinate, median claw (bottom centre), and curved, serrate bristles, especially inferiorly, x220. h, typical spine, on tibia of left leg 1, showing striations x400. c, right palpal tibia and tarsus; see g for explanation, x88. r/, left palpal tarsus; see / for explanation, x88. c, tarsal claws on right leg 4, x 220. /i explanatory drawing of g, explanatory drawing of e. See Terminology for explanation of abbreviations. the carapace edges converge in very gentle convex curves to the rounded anterior margin, and in similar shaped curves which converge more rapidly to the posterior margin. The posterior margin is broad and has a distinct median embayment. There is no prominent foveal depression, from which it can be concluded that a fovea was probably absent. About one-quarter of the way back from the anterior margin of the carapace, on each lateral margin, is a slight embayment and a dark patch of cuticle which forms a linear feature extending radially SELDEN; SPANISH CRETACEOUS SPIDERS 273 inwards. This is interpreted as a shallow sulcus which reflects a condyle for articulation of the chelicera on the inner surface of the carapace. A radial crack in the posterior right-hand quadrant indicates some crushing of the carapace. Dark areas in the anterior half form a V-shape pointing forwards. The V-shape is asymmetrical with respect to the midline, the left limb being more parallel to the midline than the right, and the whole shape is left of the midline. Some folding is associated with the limbs of the V-shape. This shape is interpreted as a sulcus or break in slope separating the peripheral parts of the carapace from a raised cephalic area in the anterior half of the carapace. The asymmetry of the V-shape, in contrast to the symmetry of the rest of the carapace, indicates left-lateral compression of the cephalic area during compaction of the sediment. A small, semicircular, dark area of cuticle approximately centrally placed near the anterior border of the carapace is interpreted as the posterior border of a right eye, the anterior half of which is missing. Other eyes, if similar in size, are most likely to be obscured by the dark lines of crumpled cuticle around the anterior edge of the cephalic lobe. The whole carapace is separated from the ventral prosoma and moved slightly to the left, exposing the right coxae. A pair of dark lines in the posterior half of the carapace, subparallel to the midline but diverging towards the anterior, are interpreted as left and right edges of the sternum. The anterior border of the sternum is seen as a transverse, recurved line just anterior to the midpoint of the carapace. Due to the left-lateral movement of the carapace, the sternum appears mainly to the right of the midline. The sternum is widest anteriorly, and narrows gradually to a blunt point situated between the coxae of the fourth pair of legs. In front of the anterior margin of the sternum are some rounded dark areas. The somewhat triangular dark area anterior to the midline of the sternum is interpreted as the labium. The labium is widest posteriorly, where it is distinctly separate from the sternum. The lateral edges converge to a rounded anterior margin. The two areas to the right of the sternum represent the right maxilla with the anterolateral carapace sulcus superimposed. The posterior part of the left maxilla can be seen to the left of the labium, but its anterior part is obscured by the right limb of the V-shape surrounding the carapace cephalic area. No serrulae can be seen on the maxillae; there are many setae visible in the intermaxillary area. The chelicerae are about 0-69 mm long, and project forwards in front of the carapace. Their lateral sides are straight and parallel to each other, their inner sides are convex and partly overlap in the fossil. Their dorsal (superior) surfaces bear numerous short, stiff setae; laterally, curved bristles are present. Each chelicera bears two rows of denticles extending from the mesal side of the anterior border to about half-way down the inner edge, the outermost row bearing at least three and probably five denticles, the inner row with at least two denticles. The fangs are not preserved, and were presumably on the counterpart which was not collected. A prominent ridge, or keel, runs from the end of the tooth row (which is ralatively short, about one quarter the length of the chelicera) along the mesal surface. No stridulatory ridges can be seen. The presence of a thickened sulcus on the anterolateral side of the carapace (see above), which probably reflects an internal apophysis for articulation of the chelicera, suggests the presence of a condyle on the chelicera; the condyle itself is not preserved on the specimen. From the morphology of the preserved cheliceral parts, it is apparent that the fangs worked transversely (labidognathous). This specimen is a mature male because the palps are modified for the transmission of sperm (text-fig. 4c, f/, /, g). Both palps are bent over to the right due to the left oblique compression of the specimen. The tarsus of the left palp now appears to the right of the right chelicera, and the right palpal tarsus lies beneath the femur of right leg 1. The appearances of the sclerites on each palp differ because the palps are compressed in different ways. The left palp presents a mesal view, and the right an ectal view. The total length of the palp, from the maxilla to the tip of the bulb (i.e. excluding the embolus, see below) is approximately 2-20 mm. The palpal femur is about equal in length to the adjacent chelicera. The patella and tibia are covered with long setae. The tibia is a distally expanded, triangular podomere, bearing long bristles which radiate distally to partly cover the tarsus. Distal to the tibia is an ovoid body with numerous sclerites superimposed on it. The ovoid body is interpreted as the superimposed bulb and cymbium (modified tarsus). The cymbium is not separately recognizable from the bulb, and is therefore presumed to be no longer than the bulb and related parts. The right palp seems to present an approximately ectal view, and the left palp an approximately superomesal view. Immediately distal to the tibia, an elliptical dark area may represent a small, separate paracymbium. On the right bulb, a curved, lath-like sclerite extends from the superoproximal edge to the inferior side of the bulb. This may be the tegulum. On the left palp, the different direction of compression has caused this structure to appear curving from the apparent inferior edge towards the superodistal direction. On the right palp below the supposed tegulum is a rather complex, hooked structure, also visible on the left palp. This may represent a median apophysis. The interpretations of both tegulum and median apophysis are uncertain. The gently helical, acuminate structure, as long as the bulb itself, extending distally from the distal end of the bulb is interpreted as the embolus. It is easily seen on the left palp, but on the right palp only its basal part is visible, the remainder 274 PALAEONTOLOGY, VOLUME 33 disappearing into the matrix beneath. It is possible that this structure represents the conductor (see, for example, the helical conductors of some tetragnathines, Levi 1981 ) in which case the embolus cannot be seen. The legs are short and nearly equal in length. The leg length formula is 1243. The coxae are visible on the right side. Each is about 0 39 mm long, and slightly longer than broad. The trochanters are very short (about half as long as broad), and each is about 019 mm in length. Approximate lengths of more distal podomeres and total leg lengths, in mm, are as follows. Leg 1 : fe L73, pa 0-46, ti L64, mt 1-08, ta 0-58; total 6-07. Leg 2: fe L54, pa 0 46, ti L50, mt 117, ta 0-69; total 5-94. Leg 3: fe L25, pa 0-46, ti 104, mt 0-87, ta 0 52; total 4-72. Leg 4: fe 1-31, pa 0-46, ti 1-25, mt 0-92, ta 0-52; total 5-04. All legs are thickly clothed in long, fine setae, and the femora, tibiae, and metatarsi bear spines. The setae are not plumose, under high magnihcation, but some appear to be sparsely serrate, and on the larger ones a rectilinear pattern, as observed on the spines, can be seen. Among the leg setae, especially on the distal podomeres, there are a few short and thin, gently S-shaped, rather crumpled, setae which extend outwards from the podomeres at high angles, and are set in follicles which appear rather too large for the thickness of the seta. Some of these hairs are interpreted as chemosensory setae, as described by Foelix (19706). Others may have had a function akin to that of trichobothria; no undoubted trichobothria can be seen. The spines have a rectilinear sculpture (text-fig, 46). The numbers of leg spines given here are the maximum number which can be seen on the specimen. This must be considered to be either the actual or the minimum number present in life, since in this specimen the counterpart was not collected, and some spines may be preserved on the counterpart only. However, since by far the greater part of both dorsal and ventral sides of the animal is preserved on the slab, it is considered likely that few spines are unaccounted for. The femora bear stout, curved spines and some curved bristles in the distal area. Four spines are present on fel, all apparently on the superior side of the podomere, of which two are subdistal and two distal in position; there are three on fe2 and fe3, superior in position ; there is at least one on the posterior side of fe4. No other spines are present on the femora. The patellae are spineless. The tibiae and metatarsi bear spines in three areas: proximal, median, and distal. The distal spines are stout, curved, and inferoanterior and inferoposterior in position; the others are longer, straighter, and occur on all sides of the podomeres. In addition, stiff bristles occur in the median and distal areas of these podomeres. Til bears five spines in a whorl on the proximal area, four (supero- and inferoanterior and supero- and inferoposterior) in the median area, and two distally. Ti2 bears four spines (one superior and three inferior) in the proximal area, two superior median, and two in the distal area. Ti3 has no proximal spines, five in a whorl in the median area, and two distally. Ti4 bears at least five spines in various positions in the proximal and median areas, and two distal spines. Mtl bars five proximal spines, two pairs in the median area, and two distally. Mt2 has four proximal spines, two in the median area, and two distally. Mt3 has three spines in each of the three areas. Mt4 bears three proximal spines, probably four in the median area, and at least three distal spines (one superior in addition to the usual anterior and posterior). All tarsi are spineless. Distally, pectinate paired claws are present, each with one row of nine teeth, and a small, unciform median claw, but no onychium (text-fig. 4u, e). Surrounding the claws are a number of serrate bristles; these are gently S-shaped at the end with the serrations on one side. There is no comb of serrate bristles along the shaft of this podomere. The abdomen is 217 mm long and elliptical to subcircular in outline. Very little cuticle is preserved, so it is presumed to have been thin in life, but the abdomen outline is clearly seen due to its covering of setae. Cuticle between the setae can only be seen in ill-defined areas mainly in the left half of the abdomen; these are interpreted as ?dorsal pigmented areas in life. Generally darker coloration and greater density of setae indicate position of spinnerets which, however, are not separately discernible. The spinneret region can be seen on the right side of the abdomen (due to the left oblique compression) and appears to have been subterminal in position. A longitudinal dark streak left of the midline between the carapace and abdomen is presumed to represent the remains of the pedicel. Family tetragnathidae Menge, 1866 The familial status of Tetragnathidae has been supported by many authors (for example, Kaston 1948, 1978; Locket and Millidge 1953), but resisted by others (Levi 1980; Roberts 1985, p. 198) in the past. The situation is further complicated by the position of the Metinae, placed by some in Tetragnathidae and by others in Araneidae. The most recent opinions of Levi (1986), Coddington (19896), and Platnick (1989) are that tetragnathines, metines, and nephilines should be placed together in the family Tetragnathidae, separate from the Araneidae. In the past, these subfamilies SELDEN: SPANISH CRETACEOUS SPIDERS 275 have been placed in the family Araneidae, but Coddington (1989/), fig. 108) considered the clade Nephilinae + (Tetragnathinae + Metinae) as the sister group to Araneidae + Linyphiidae. First impressions of Macryphantes suggested the ‘crab-spiders’ (superfamilies Thomisoidea and Philodromoidea), on account of the subcircular carapace, prominent eyes, subcircular palp, and long, spinose anterior legs. However, these features are not confined to crab-spiders, and the characteristic features of crab-spiders (two claws, tarsal scopulae, clavate setae, etc.) are lacking in Macryphantes. Furthermore, in thomisoids and philodromoids the legs are laterigrade, not densely setose, and the spines on the legs are concentrated on the mesal surfaces of the tibiae and metatarsi of the anterior prey-capturing legs. This is not the case in Macryphantes. The presence of femoral trichobothria in Macryphantes points to the Deinopoidea or Araneoidea. As discussed above with regard to Palaeontohoriis, only the Uloboridae (Deinopoidea) and the metines-tetragnathines in the Araneoidea bear femoral trichobothria. Since Macryphantes is an adult male, it could lack a calamistrum and cribellum, and correlated with the loss of cribellum and calamistrum in adult male uloborids appears to be the loss of the comb of macrosetae on the fourth tarsus (personal observation from Ulohorus walckenaerius). Arguing against its inclusion in the Uloboridae are: the presence of serrate setae, the absence of plumose setae, and the absence of feathery setae. Serrate setae are characteristic of araneoids (Coddington 1986) and members of the superfamily lack plumose hairs which are found in deinopoids. Furthermore, the large, pectinate, paired tarsal claws of Macryphantes resemble those of araneoids more than the uloborid claw pattern of relatively small, sparsely toothed or non-pectinate paired claws. The presence of femoral trichobothria in Macryphantes places it among the tetragnathines within the Araneoidea; this character has been used to distinguish tetragnathids in familial keys (Kaston 1948; 1972; Locket and Millidge 1953). As mentioned above, Coddington (1989/>) has argued that the tetragnathines are closely related to the metines and nephilines. Whilst a number of other features, such as leg length and possible paracymbium, add weight to this assignation, some characters of Macryphantes are unusual for this group, including: subcircular or broadly pyriform carapace, planospiral embolus, and prominent, dorsally directed, posterior median eyes. Therefore, whilst the presence of femoral trichobothria appear to ally Macryphantes most closely with the tetragnathines, rather than the metines and nephilines which lack this feature, these other characters suggest that inclusion of the fossil in the Tetragnathinae is unwise. Genus macryphantes gen. nov. Derivation of name. Greek makros, long, large, and yphantes. a weaver. Type and only known species. Macryphantes cowdeni sp. nov. Diagnosis. Tetragnathid spider with subcircular, or broadly pyriform, foveate carapace; leg 1 six times the length of carapace and more than twice the length of leg 3; double row of prolateral trichobothria on femur of leg 3, single row of prolateral trichobothria on femur of leg 4; paired tarsal claws pectinate with six teeth, median claw long, curved, not pectinate, serrate bristles (accessory claws) present; male palp with planospirally coiled embolus; serrate, but not plumose, setae present. Macryphantes cowdeni sp. nov. Plates 3 and 4; text-figs. 5 and 6 Derivation of name. In remembrance of a friend and a fellow arachnologist. Dr Douglas Cowden of Worcester. Type specimens. Holotype, LC 1753 AP A (part) and LC 1753 AP B (counterpart). Paratype, LC 1754 AP A (part) and LC 1754 AP B (counterpart). Both are from the quarry of La Cabrua, Sierra de Montsech, north- east Spain, and are held in the collections of the Institut d’Estudis Ilcrdcncs, Lerida, Spain. 276 PALAEONTOLOGY, VOLUME 33 Diagnosis. As for the genus. Description. This is the largest and one of the best preserved of the spiders from Montsech, with long, outstretched legs 1. The description is based on specimen LC 1753 AP (PI. 3; text-fig. 5), which is better preserved (though slightly smaller) than LC 1754 AP (PI. 4, figs. 2 and 4; text-fig. 6); the latter is referred to for confirmation of details. Both specimens are mature males. The carapace is slightly wider (2-83 mm) than long (2 65 mm), and is widest at about midlength. The carapace outline is subcircular, but may be somewhat produced anteriorly where the edge is not preserved. The TEXT-FIG. 5. Macryphantes cowdeni gen. et sp. nov., holotype, LC 1753 AP B. Explanatory drawing for PI. 3, fig. 1. See Terminology for explanation of abbreviations. EXPLANATION OF PLATE 3 Figs. 1 and 2. Macryphantes cowdeni gen. et sp. nov., holotype; Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lerida Province, Spain; under ethanol. 1, LC 1753 AP B, lower slab, x 7; see text-fig. 5 for explanation. 2, LC 1753 AP A, upper slab, x 7. PLATE 3 SELDEN, Macryphantes 278 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 6. Macryplumtes cowdeni gen. et sp. nov., paratype, LC 1754 AP A. Explanatory drawing for PI. 4, fig. 4. See Terminology for explanation of abbreviations. carapace is covered in setae. At the anterior side of the carapace, there are forwardly directed setae and long, curved bristles. At the anterior margin of the carapace two circular structures are interpreted as median eyes. Just posterior to the centre of the carapace, a deep, drop-shaped depression, with its blunt end anteriormost, marks the fovea. The carapace shape is confirmed by specimen LC 1754 AP, in which the carapace is 3-25 mm wide. A pair of deep depressions just posterior to the anterior median eyes are surrounded posteriorly by dark EXPLANATION OF PLATE 4 Figs. 1-7. Macryplumtes cowdeni gen. el sp. nov., holotype and paratype; Lithographic Limestone, Lower Cretaceous; Sierra de Montsech, Lerida Province, Spain. I, 3, 5-7, LC 1753 AP B, holotype, oil immersion. 1, Trichobothria, setae, and spines on prolateral surface of femur of right leg 3, x 85. 3, Right palp, showing apophysis on right side, x 47. 5, Distal end of tarsus of right leg 2, showing paired pectinate claws, long median claw, and curved, serrate bristles (accessory claws), x 320. 6, Spines and setae on shaft of metatarsus of left leg 4, superior to top, x 85. 7, Distal half of tarsus of right leg 2, showing terminal claws and short spine on inferior surface of podomere, x 130. 2, LC 1754 AP B, paratype, upper slab, x4-5. 4, LC 1754 AP A, paratype, lower slab, x 3 6; see text-fig. 6 for explanation. Both under ethanol. PLATE 4 SELDEN, Macryphantes 280 PALAEONTOLOGY, VOLUME 33 cuticle. These structures are interpreted as the proximal ends of the chelicerae which are directed ventrally ; the chelicerae presumably disappear into the matrix beneath the specimen. The sternum appears to be circular; its shape is suggested by the arrangement of the coxae, as seen on both LC 1753 AP and LC 1754 AP. The coxae of legs 4 meet in the midline, and on both the holotype and paratype some remnants of the anterior edge of the sternum can be discerned. The palp on the right side of LC 1753 AP B (the left side of LC 1753 AP A) is preserved (PI. 4, frg. 3), and confirms that the specimen is an adult male. The palpal tarsus is subcircular in outline with a mesal bulge. This bulge represents either the mesal part of a circular cymbium or a mesally situated crescentic cymbium. The bulge extends beyond the oval dark band which is interpreted as the embolus (this structure could be the conductor or some other sclerite functioning as a guide for the embolus). The embolus is coiled in a horizontal planospiral, which in LC 1754 AP A can be seen to be sinistral on the left palp, and dextral on the right, as viewed from above. On the proximolateral side, a thick, reniform lobe is interpreted as a median apophysis. A dark area can be seen on the proximal edge of the cymbium within the coiled embolus; this is inferior in position; this dark area belongs to a sclerite of the palp. A smaller dark patch occurs just anterior to the midpoint of the palp, on the inferior surface; this represents part of another sclerite. Numerous long bristles run right laterally from the base of the tarsus, and some curved bristles are present on its superior surface. Larger, curved spines are present around the base of the tarsus; these appear to originate proximal to the tarsus, that is, on the tibia, which is otherwise poorly preserved. The leg length formula is 1243. The legs are very unequal in length, leg 1 being more than twice the length of leg 3. The coxae are approximately equidimensional, each about 0-58 mm long and wide. The trochanters are not easily visible, but each measures about 0-25 mm in length. Lengths of the more distal podomeres and total leg lengths, in mm, are as follows: leg I : fe 4-33, pa I -35, ti 3-00, mt 4-55, ta 1-70, total 15-76; leg 2: fe 3-17, pa 1-00, ti 1-64, mt 3-00, ta 115, total 10-79; leg 3: fe 2-01, pa 0-39, ti 115, mt 1-64, ta 0-85, total 6-87; leg 4: fe 3-00, pa 0-81, ti 2-04, mt 1-98, ta not preserved (presumed to be c. 0-95 by comparison with ratio of mt4 to ta4 seen in LC 1754 AP, see below), total c. 9-61. LC 1754 AP is poorly preserved, and slightly larger than LC 1753 AP, and its leg measurements are as follows (in mm but with a lesser degree of certainty than in LC 1753 AP): coxae 0-9, trochanters 0-3; leg 1 : fe 4-7, pa 1-4, ti 3-1, mt 5-0, ta 1-8, total 17-2; leg 2: fe 3-8; leg 3: fe 2-3, pa 0-6, ti 1-4, mt L3, ta 0-9, total 7-7; leg 4: fe 2-8, pa 0-9, ti 1-9, mt 2-1, ta 1-0, total 9-9. All legs are thickly clothed with long setae. These setae appear smooth, but high magnification reveals minute accessory spines, especially distally; thus they are the serrate type. No plumose or feathery hairs can be seen. Spines are mostly large and numerous, occurring on all podomeres except the coxae and trochanters. Some spines show a helical pattern of longitudinal lines. Most spine are quite large (mean length 0-48 mm) and straight (PI. 4, fig. 6), but spines at the distal end of podomeres are commonly curved and/or short. Pel bears five pairs of spines along the superior surface and an inferoanterior distal spine (this may be one of a pair). Pel also has many very short, fine hairs on the inferior surface of the distal half of the podoniere. Pal bears a prominent posterior spine, and antero- and posteroinferior distal spines. Til bears a whorl of about five spines proximally, two pairs of spines inferiorly and two single spines superiorly along the podomere, and postero- and anteroinferior distal spines. Mtl bears eight pairs of spines along the inferior surface, including and distal pair. Pe2 bears four pairs of superior spines along the shaft, and postero- and anteroinferior distal spines. Pa2 bears one lateral and antero- and posteroinferior distal spines. Ti2 bears ten long spines along the shaft of the podomere, and two short distal spines. Mt2 bears a pair of spines proximally, a whorl of four spines and then five more along the shaft, with a whorl of five spines distally. Ta2 on the right side of LC 1753 AP B is particularly well preserved (PI. 4, figs. 5 and 7), and shows curved paired claws each with six teeth, an equally long curved median claw, and numerous serrate accessory claws (i.e. hypertrophied bristles, as seen in living Araneidae, see Foelix 1970u). Ta2 bears two small spines on is inferior surface. Along the superior prolateral side of the proximal three-quarters of fe3 are about 24 trichobothria arranged mainly in two rows (PI. 4, fig. 1). The trichobothrial hairs are not feathered. This podomere bears many fine, curved hairs inferiorly, two median superior spines, and two posterior distal spines. Pa3 appears to be spineless. Ti3 bears long, stiff setae, and one median and two subdistal spines. Mt3 has three inferior and one superior proximal spines, one superior, antero- and posterolateral, and two short inferior median spines, followed by one lateral and three curved inferior spines and a whorl of five distal spines. Ta3 bears two small spines on its inferior surface, like those which occur on Ta2. Fe4 bears at least one superior prolateral row of about sixteen trichobothria, fine, curved hairs inferiorly, three superior median spines, and three superior subdistal spines. Pa4 bears one lateral spine. Ti4 has three superior, two inferior, and two lateral spines along the shaft, and apparently no distal spines. Mt4 bears antero- and posteroinferior and posterolateral spines proximally, two inferior median spines, one small curved superior median spine, two small curved distal spines, and two small and one large inferior distal spines. The numerous setae on the superior surface of mt4 are gently curved, giving SELDEN: SPANISH CRETACEOUS SPIDERS 281 the appearance of a weak calamistrum (PI. 4, fig. 6). However, the high density of setae may be an artefact of compression, since a similar density of curved setae is observed on the opposite side of the podomere, and high magnification reveals that the setae are no different in structure from any others. Thus mt4 does not bear a calamistrum. No trace of the abdomen is preserved on LC 1753 AP, but it is preserved on LC 1754 AP; it is oval, and measures 4-80 mm long and 4-20 mm wide. The greatest width is in the anterior half. The abdomen is compressed to the right in LC 1754 AP A, and was probably quite bulbous in life. The spinnerets arc not elongated, and form a compact group in a subterminal position on the abdomen. MODES OF LIFE Palaeoulohorus can be compared most closely with the Uloboridae. Uloborids are orb-web weavers which use a characteristic ‘wrap attack ’ to subdue prey (Robinson 1975). In Ulohonis and Hyptiotes (Nielsen 1932) the median tarsal claw is relatively large, the paired claws are fine and bear few (Hyptiotes) or no teeth (Ulohonis), and in both genera there are large, serrate accessory claws. A similar pattern of tarsal claws occurs in Palaeoulohorus, and it is unlike that found in araneoids, in which the paired claws are large and pectinate (see, for example, Levi 1978). Nielsen (1932, pp. 26-28) described the method of silk handling by Hyptiotes using this claw pattern. Whilst accessory claws are widespread among web-spinning spiders, and are used for silk handling, they are particularly well developed in orb-web weavers; the pattern in the living uloborids studied seems distinctive, and these genera are orb-web weavers. Trichobothria occur on the femora Palaeoulohorus and uloborids. The function of these is not known, but it is interesting that among living araneomorph spiders they are found only in tetragnathines and uloborids, both of which are orb-web weavers. Femoral trichobothria are absent from some adult Pachyguatha, tetragnathines which make no web when adult (Levi 1980). Many orb-web weavers have no femoral trichobothria, but their presence in tetragnathines and uloborids appears to be linked with the habit. Another behavioural similarity between tetragnathines and uloborids is in resting postures (Levi 1980). Tetragnathines and many metines rest with their long legs 1 and 2 stretched out forwards, the fourth legs outstretched behind, and the short third legs pointing backwards and where necessary gripping the twig on which the spider is resting. Opell and Eberhard (1983) distinguished four types of resting posture in uloborids; in three, legs 1 and 2 are stretched forwards in some manner, whereas in the fourth, legs 1 and 2 are held folded with the femora projecting at right angles to the long axis of the body. Opell and Eberhard (1983) remarked on the close similarities of resting postures between uloborids and araneids. The femoral trichobothria are generally on the superior or retrolateral sides of femora 1 and 2, and on the prolateral sides of femora 3 and 4 (Opell 1979). Thus they point laterally when the animal is in the normal resting posture. It is likely that the presence of femoral trichobothria is linked with the uloborid and tetragnathine-metine resting postures. These behavioural characters may be due to convergence, but could conceivably be synapomorphies for all orb-web weavers (Shear 1986u). The great similarity in leg lengths, femoral trichobothrial pattern, and tarsal claws between Palaeoulohorus and the Uloboridae suggests that the web-building and resting behaviour of the fossil genus resembled that of typical members of the living family. The wrap-attack prey capture in uloborids was described by Opell (1979) and Lubin (1986). In it, the spider hangs by the first and second pairs of legs whilst throwing silk over the prey using the fourth leg-pair. After further entanglement of the prey in these threads, the spider approaches closer to the prey and, holding it now with the second and third leg-pairs, wraps it more tightly with silk combed by the row of macrosetae on the fourth legs. A wrap attack is found in a number of spider groups, such as the Theridiidae and Nesticidae (both of which have a comb of setae on the fourth legs), Metinae, Tetragnathinae, Araneidae, Oecobiidae, Hersiliidae, and Pholcidae (Coddington 1986). The wrap attack of uloborids differs from that of other spider families because uloborids lack poison glands and the prey is killed by digestive enzymes during feeding (Opell 1979). Since a wrap attack is found in families which do not have a comb of setae on the fourth legs, such a comb is 282 PALAEONTOLOGY. VOLUME 33 not essential for this method of prey capture. Wrap attack is therefore a possible method of prey capture in Palaeoulohorus. Macryphantes compares most closely in general appearance with large, long-legged araneoids such as the argiopine araneids, tetragnathines, and nephilines. Argiopines are similar in general appearance and habits to uloborids (both are orb-web weaving wrap-attack predators) but they lack femoral trichobothria. Macryphantes has an araneoid pattern of tarsal claws, and bears femoral trichobothria. Therefore, it is suggested that Macryphantes wove an orb web, rested in a posture like that of uloborids or metines- tetragnathines, and may have used a wrap-attack method of prey capture. Cretaraneus has few positive features which would indicate its possible mode of life. The fossil genus is small, short-legged, and bears pectinate, paired claws and serrate accessory claws, which indicate that it is a web-weaving spider. Small, short-legged araneoids, such as Cretaraneus and most theridiids and linyphiids, are weavers of sheet webs in litter, undergrowth, or bushes; such webs catch pedestrian or small flying prey. It is likely that Cretaraneus occupied a similar ecological niche. A great variety of orb webs are woven by uloborid and araneoid spiders, each designed to capture a specific type of prey. They are put up for short periods or longer, day or night, and in open or secluded situations (Riechert and Gillespie 1986; Stowe 1986; Lubin 1986). They vary from massive, collective structures to minimalist devices hardly recognizable as orb web derivatives. It is impossible to suggest what type of prey Macryphantes and Palaeoulohorus captured with their orb webs. There was a wide diversity of insect life in the Montsech area during the early Cretaceous, which suggests that prey specialization may have been practised by orb-web weavers at that time. CONCLUSIONS Described here are the oldest known representatives of the spider superfamily Deinopoidea, the family Tetragnathidae, and the second oldest record of the superfamily Araneoidea in the fossil record. The Deinopoidea and the Araneoidea both contain weavers of orb webs of remarkable similarity. Indeed, there is continuing debate about whether the orb web evolved only once, in the common ancestor of the Deinopoidea and Araneoidea, or is a convergent phenomenon in these two groups. Shear (1986u) comprehensively reviewed the evidence for and against these conflicting hypotheses, and further discussion is not attempted here. However, the presence of well-defined deinopoids and araneoids in the Lower Cretaceous indicates that, whichever hypothesis is favoured, both groups of orb-web weavers were in existence at that time, and suggests that the orb web originated earlier in the Mesozoic, if not before. Acknowledgements . I am very grateful to Antonio Lacasa-Ruiz of the Institut d’Estudis llerdencs for bringing these specimens to my attention, for the loan of the fossils, and for hospitality in Lerida. The specimen of Palaeoulohorus lacasae was found by J. Gonzalez-Redondo, and the hololype of Macryphantes cowdeni is from the collection of Xavier Martinez- Delclds; I am grateful to them for generously donating their specimens for study. Specimens of extant spiders were loaned by the British Museum (Natural History) Arachnid Section (courtesy of Paul Hillyard), the Manchester Museum (courtesy of Charles Pettit), and by John Murphy. Xavier Martinez-Delclos gave important information on the stratigraphy, Richard Porter made useful comments on palynostratigraphy, and for use of the reflected-light microscope 1 am grateful to Richard Pattrick. I thank Peter Gabbutt and John Crocker (Honorary Librarian of the British Arachnological Society) for the loan of publications. John Dalingwater helped with photography, and read and commented on parts of the manuscript. I thank Fred Wanless for some initial ideas, and Bill Shear, Brent Opell, and Jon Coddington for kindly suggesting some lines of inquiry. A Royal Society Scientific Investigations Grant is gratefully acknowledged. SELDEN: SPANISH CRETACEOUS SPIDERS 283 REFERENCES BARALE, G., BLANC-LOUVEL, C., BUFFETAUT, E., COURTINAT, B., PEYBERNES, B., VIA BOADA, L. and WENZ, S. 1984. Les gisements de calcaires lithographiques du Cretace inferieur dii Montsech (Province de Lerida, Espagne), Considerations paleoecologiques. Geobios, Memoires, no. special 8, 275-283. BRENNER, p., GOLDMACHER, w. and SCHROEDER, R. 1974. Ostrakodcn und Alter der Plattenkalk von Rubies (Sierra de Montsech, Prov. Lerida, NE-Spanien). Neues Jiihrbuch fur Geologie und Paldontologie, M omit she fte, 1974, 513-524. CLERCK, c. 1757. Svenska spindlar (Araneae suecici). L. Salvii, Stockholm, 154 pp. [In Swedish and Latin]. CODDINGTON, J. 1986. The monophyletic origin of the orb-web. 319-363. In shear, w. a. (ed.). Spiders - n ebs, behavior, and evolution. Stanford Press, Stanford, California, viii-F492 pp. 1989fl. Spinneret spigot morphology: evidence for the monophyly of orb-weaving spiders, Cyrtophorinae (Araneidae), and the group Theridiidae plus Nesticidae. Journal of Arachnology, 17, 71-95. 19896. Ontogeny and homology in the male palpus of orb-weaving spiders and their potential outgroups, with comments on phylogeny (Araneoclada : Araneoidea, Deinopoidea). Smithsonian Contributions to Zoology, in press. 1990. Cladistics and spider classification: araneomorph phylogeny and the monophyly of orb-weavers. Annales Zoologici Fennici, in press. COMSTOCK, J. J. 1948. The spider hook. Revised and edited by w. j. gertsch. Comstock Publishing Company, Ithaca, New York, xii-l-729 pp. CROWSON, R. A. 1970. Classification and biology. Heinemann, London, x-l-350 pp. ESKOV, K. 1984. A new fossil spider family from the Jurassic of Transbaikalia (Araneae: Chelicerata). Neues Jahrbueh fiir Geologie und Paldontologie, Monatshefte, 1984, 645-653. 1987. A new archaeid spider (Chelicerata: Araneae) from the Jurassic of Kazakhstan, with notes on the so-called ‘ Gondwanan ' ranges of recent taxa. Neues Jahrbueh fiir Geologie und Paldontologie, Ahhandlungen, 175, 81-106. FARRIS, J. s. 1976. Phylogenetic classification of fossils with Recent species. Systematic Zoology, 25, 271-282. FOELix, R. F. 1970a. Structure and function of tarsal sensilla in the spider Araneus diadematus. Journal of E.xperimental Zoology, 175, 99-124. 19706. Chemosensitive hairs in spiders. Journal of Morphology, 132, 313-334. FORSTER, R. R. and BLEST, A. D. 1979. The spiders of New Zealand. Part V. Otago Museum Bulletin, 5, 1-173. and PLATNiCK, N. I. 1977. A review of the spider family Symphytognathidae (Arachnida, Araneae). American Museum Novitates, 2619, 1-29. 1984. A review of the archaeid spiders and their relatives, with notes on the limits of the superfamily Palpimanoidea (Arachnida, Araneae). Bulletin of the .American Museum of Natural History, 178, 1 106. 1985. A review of the austral spider family Orsolobidae (Arachnida, Araneae), with notes on the superfamily Dysderoidea. Bulletin of the American Museum of Natural History, 181, 1-229. and WILTON, c. l. 1973. The spiders of New Zealand. Part IV. Otago Museum Bulletin, 4, 1-309. gall, j.-c. 1971. Faunes et paysages du Gres a Voltzia du nord des Vosges. Essai paleoecologique sur le Buntsandstein superieur. Memoires du Service de la Carte Geologique d' Alsace et de Lorraine, 34, 1-318. gertsch, w. j. 1974. The spider family Leptonetidae in North America. Journal of Arachnology, 1, 145-203. HEiMER, s. and NENTWiG, w. 1982. Thoughts on the phylogeny of the Araneoidea Latreille, 1806 (Arachnida, Araneae). Zeitschrift fiir Zoologische Systematik und Evolutionsforschung, 20, 284—295. JELL, p. A. and DUNCAN, p. M. 1986. Invertebrates, mainly insects, from the freshwater. Lower Cretaceous, Koonwarra Fossil Bed (Korumburra Group), South Gippsland, Victoria. Memoirs of the Association of Australasian Palaeontologists, 3, 1 1 1-205. KASTON, B. J. 1948. Spiders of Connecticut. Connecticut State Geological and Natural History Survey Bulletin, 70, 1-874. 1978. How to know the spiders. 3rd edn. William C. Brown, Dubuque, Iowa, vii-l-272 pp. KOCH, c. L. 1851. Ubersicht des Arachnidensystems Heft 5. C. H. Zeh, Nurnberg, 104 pp. KOVOOR, J. 1977. L'appareil sericigenes dans le genre Uloborus Latr. (Araneae: Uloboridae), 1: Anatomic. Revue Arachnologicjue, 1, 89-102. KULLMAN, E. 1972. The convergent development of orb-webs in cribcllate and ecribellate spiders. American Zoologist, 12, 395-405. LACASA, A. 1985. Nota sobre las plumas fosiles del yacimento eocretacico de ‘La Pedrera - La Cabriia’ cn la Sierra del Montsec (Prov. Lleida, Espafia). Ilerda, 46, 227-238. 284 PALAEONTOLOGY, VOLUME 33 LACASA, A. and martinez, X. 1986. Fauna y flora de los yacimentos Neoconiienses del Montsec (Prov. Lerida). Paleontologia i Evoliaidn, 20, 215-223. LATREILLE, P.-A. 1806. Genera criistaceorwn et insectorum (4 vols, 1806-9.) Araneides. Vol. 1, 82-127. A. Koenig, Paris. LEHTiNEN, p. T. 1 967. Classification of the cribellate spiders and some allied families, with notes on the evolution of the suborder Araneomorpha. Awiales Zoologici Fe?inici, 4, 199^68. 1978. Definition and limitation of supraspecific taxa in spiders. 255-271. In merrett, p. (ed.). Aracimology. (Proceedings of the 7th International Congress of Arachnology.) Symposia of the Zoological Society of London. 42, 1-530. 1980. Trichobothrial patterns in high-level taxonomy of spiders. Proceedings of the 8th International Arachnological Congress. Vienna. 1980. 493^98. 1986. Evolution of the Scytodoidea. 149-157. In eberhard, w. g., lubin, y. d. and robinson, b. c. (eds.). Proceedings of the Ninth International Congress of Arachnology. Panama. 1983. Smithsonian Institution Press, Washington DC and London, 333 pp. LEVI, H. w. 1961. Evolutionary trends in the development of palpal sclerites in the spider family Theridiidae. Journal of Morphology. 108, 1-9. 1968. The spider genera Gea and Argiope in America (Araneae, Araneidae). Bulletin of the Museum of Comparative Zoology. Harvard University. 136, 319-352. 1978. Orb-webs and the phytogeny of orb-weavers. 1-15. In merrett, p. (ed.). Arachnology. (Proceedings of the 7th International Congress of Arachnology.) Symposia of the Zoological Society of London. 42, 1-530. 1980. The orb-weaver genus Mecynogea. the subfamily Metinae, and the genera Pachygnatha. Glenognatha. and Azilia of the subfamily Tetragnathinae north of Mexico (Araneae: Araneidae). Bulletin of the Museum of Comparative Zoology. Harvard University. 149, 1-74. 1981. The American orb- weaver genera Dolichognatha and Tetragnatha north of Mexico (Araneae: Araneidae, Tetragnathinae). Bulletin of the Museum of Comparative Zoology. Harvard University. 149, 271-318. 1983. The orb-weaver genera Argiope. Gea. and Neogea from the western Pacific region (Araneae: Araneidae: Argiopinae). Bulletin of the Museum of Comparative Zoology. Harvard University. 150, 247-338. 1986. The neotropical orb-weaver genera Chrysometa and Homalometa (Araneae: Tetragnathidae). Bulletin of the Museum of Comparative Zoology. Harvard University. 151, 91-215. and LEVI, L. R. 1962. The genera of the spider family Theridiidae. Bulletin of the Museum of Comparative Zoology. Harvard University. 127, 1-71. LOCKET, G. H. and MiLLiDGE, A. F. 1953. British spiders. Vol. II. The Ray Society, London, vii-E449 pp. LUBIN, Y. D. 1986. Web building and prey capture in the Uloboridae. 132-171. In shear, w. a. (ed.). Spiders - webs, behavior, and evolution. Stanford Press, Stanford, California, viii-l-492 pp. McALPiNE, J. F. and martin, j. e. h. 1969. Canadian amber -a paleontological treasure-chest. The Canadian Entomologist. 101, 819-838. MADDisoN, w. p. and MADDisoN, D. R. 1987. MacClcule. Version 2.1. An interactive, graphic program for analyzing phytogenies and studying character evolution; for the Apple Macintosh™ computer. (Available from the authors at : Museum of Comparative Zoology, Harvard University, Cambridge, MA 02138, USA). MENGE, A. 1986. Preussischen Spinnen. Schriften der Naturforschenden Gesellschaft in Danzig. N.E.. 2, 1-152. MERRETT, p. 1963. The palpus of male spiders of the family Linyphiidae. Proceedings of the Zoological Society of London. \4d. 347^67. MILLIDGE, A. F. 1977. The conformation of the male palpal organs of linyphiid spiders, and its application to the taxonomic and phylogenetic analysis of the family (Araneae: Linyphiidae). Bulletin of the British Arachnological Society. 4. 1-60. 1988. The relatives of the Linyphiidae: phylogenetic problems at the family level (Araneae). Bulletin of the British Arachnological Society. 1. 253-268. NIELSEN, E. 1932. The biology of spiders. Vol. I [in English] 248 pp.4-32 pis. Vol. II [in Danish] 723 pp.4-5 pis. Levin and Munksgaard, Copenhagen. OPELL, B. D. 1979. Revision of the genera and tropical American species of the spider family Uloboridae. Bulletin of the Museum of Comparative Zoology. Harvard University. 148, 443-549. 1982. Cribellum, calamistrum, and ventral-comb ontogeny in Hyptiotes cavatus (Hentz) (Araneae: Uloboridae). Bulletin of the British Arachnological Society. 5, 338-343. and EBERHARD, w. G. 1983. Resting postures of orb-weaving uloborid spiders. Bulletin of the British Arachnological Society. 11, 369-376. SELDEN: SPANISH CRETACEOUS SPIDERS 285 PATTERSON, c. and ROSEN, D. E. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the theory and practice of classifying fossils. Bulletin of the American Museum of Natural History, 158, 81-172. PEYBERNES, B. 1976. Lc Jurassique ct le Cretace inferieur des Pyrenees Franco-Espagnoles entre la Garonne et la Mediterranee. These Doctoral Sci. Nat. No. 646, Universite Paul-Sabatier, Toulouse. and OERTLi, H. 1972. La serie de passage du Jurassique au Cretace dans le Bassin sud-pyreneen (Espagne). Compte Rendu Hebdomadaire des Seances de 1’ Academic des Sciences, Paris, serie D, 274, 3348-3351. PLATNiCK., N. I. 1977. The hypochiloid spiders: a cladistic analysis, with notes on the Atypoidea (Arachnida, Araneae). American Museum Novitates, 2627, 1-23. 1986. Seminar: Establishment of phyletic seriations. 55-57. In barrifintos, j. a. (ed.). Actas X Congreso Internacional de Aracnologia, Jaca (Espaha), 1986. Vol 2. Institute Pirenaico de Ecologi'a, Jaca, 73 pp. 1989. Advances in spider taxonomy 1981-1987 : a supplement to Brignoli's A catalogue of the Araneae described between 1941) and 1981. Manchester University Press, Manchester, vii + 673 pp. and GERTSCH. w. j. 1976. The suborders of spiders: a cladistic analysis (Arachnida. Araneae). American Museum Novitates, 2607, 1-15. and SHADAB, M. u. 1978a. A review of the spider genus Mysmenopsis (Araneae, Mysmenidae). American Museum Novitates, 2661, 1-22. 1978/f. a review of the spider genus Anapis (Araneae, Anapidae), with a dual cladistic analysis. American Museum Novitates, 2663, 1-23. POCOCK, R. I. 1892. Liphistius and its bearing upon the classification of spiders. Annals and Magazine of Natural History, (6), 10, 306-314. RAVEN, R. J. 1985. The spider infraorder Mygalomorphae (Araneae): cladistics and systematics. Bulletin oj the American Museum of Natural History, 182, 1-180. RiECHERT, s. E. and GILLESPIE, R. G. 1986. Habitat choice and utilization in web-building spiders. 23^8. In SHEAR, w. A. (ed.). Spiders - webs, behavior, and evolution. Stanford Press, Stanford, California, viii + 492 pp. ROBERTS, M. J. 1985. The spiders of Great Britain and Ireland. Vol. I. Harley Books, Colchester, 229 pp. ROBINSON, M. H. 1975. The evolution of predatory behaviour in araneid spiders. 293-312. In baerends, g., beer, c. and manning, a. (eds.). Function and evolution in behaviour. Clarendon Press, Oxford, xxxi-l-393 pp. SANZ. J. L., BONAPARTE, J. F. and LACASA, A. 1988. Unusual Early Cretaceous birds from Spain. Nature, London, 331, 433^35. SCHAIRER, g. and janicke, v. 1970. Sedimentologisch-palaontologische Untersuchungen an den Plattenkalken der Sierra de Monsech (Prov. Lerida, NE-Spanien). Neues Jahrbuch fur Geologic und Paldontologie, Abhandlimgen, 135. 171-189. SCHULT, J. 1983. Simple bulbs in male spiders - simple or derived? (Arachnida: Araneae). VerhamUung der Naturwissenschaftlichen Verein in Hamburg, 26, 155-160. SELDEN, p. A. 1989. Orb-web weaving spiders in the early Cretaceous. Nature, London, 340, 71 1-713. shear, w. a. 1986a. The evolution of web-building behavior in spiders: a third generation of hypotheses. 364^00. In SHEAR, w. a. (ed.). Spiders - webs, behavior, and evolution. Stanford Press, Stanford, California, viii -1-492 pp. 1986/i. Taxonomic glossary. 401^32. In shear, w. a. (ed.). Spiders - webs, behavior, and evolution. Stanford Press, Stanford. California, viii -f 492 pp. SIMON, E. 1 892. Histoire naturelle des araignees, 2nd edn, vol. I , part 1 . Libraire Encyclopedique de Roret, Paris, viii -I- 256 pp. SMITH, F. p. 1902. The spiders of Epping Forest. Essex Naturalist, 12, 181-201. STOWE, M. K. 1986. Prey specialization in the Araneidae. 101-131. In shear, w. a. (ed.). Spiders - webs, behavior, and evolution. Stanford Press, Stanford, California, viii 4-492 pp. walckenaer, c. a. 1802. Faune parisienne, insectes. Dentu, Paris, 250 pp. WATERS, s. B. 1989. A new hybotine dipteran from the Cretaceous of Botswana. Palaeontology. 32, 657-667. WHALLEY, p. e. s. and JARZEMBOWSKI, E. A. 1985. Fossil insects from the Lithographic Limestone of Montsech (late Jurassic-early Cretaceous), Lerida Province, Spain. Bulletin of the British Museum [Natural History) (Geology), 38, 381-412. WILEY, E. o. 1981. Phylogenetics. Wiley, New York, Chichester, Brisbane, and Toronto, xvi-l-439 pp. PAUL A. SELDEN Department of Extra-Mural Studies University of Manchester Manchester M13 9PL Typescript received 1 March 1989 Revised typescript received 19 April 1989 ! i'" a- b r- P.;. f. r j... . - b' GIANT ACANTHOMORPH ACRITARCHS FROM THE UPPER PROTEROZOIC IN SOUTHERN NORWAY by G. VIDAL Abstract. Large microfossils from a single phosphatic pebble in the late Riphean-early Vendian ( > 612+ 18 — 665 + 10 Ma) Biskopas Conglomerate in the Hedmark Group in southern Norway were described by Spjeldnass (1963) under the name Papillomemhrana compta and interpreted as a possible dasycladaccan alga. The microfossils are evidently organic-walled and here regarded as giant acanthomorph acritarchs. The diagnosis of Papillomemhrana compta is emended. Possibly related Proterozoic taxa are discussed. A new acanthomorph acritarch, Ericiasphaera spjeldnaesii gen. et sp. nov., found with P. compta within the same petrographic thin-section is described. Contrary to former conceptions, recent data seem to indicate that large complex acritarchs were common in late Proterozoic times. Recent studies of Precambrian fossils reveal a formerly unsuspected degree of complexity in the late Proterozoic biota (Knoll and Calder 1983; Vidal and Knoll 1983; Knoll 1985; Butterfield etui. 1988; Zhang, 1989; Zang and Walter 1989). At the same time, the emerging picture sheds additional light on the importance of microfossils of planktonic primary producers for the understanding of the metazoan radiation at, or near the Proterozoic-Cambrian boundary (Germs et al. 1986; Moczydlowska and Vidal 1986; Knoll and Swett 1987). In this light a proper understanding of the taxonomic affinity of Proterozoic fossils seems of prime importance as it aids the acquisition of a more precise picture of the complexity of existing food webs (Knoll and Calder 1983). In this paper, as a result of an ongoing study of the micropalaeontology of Proterozoic-Cambrian sections in southern Norway (Vidal and Nystuen, unpublished data), previously reported microfossils from the Upper Proterozoic Hedmark Group (Spjeldn;es 1963, 1967) are reinterpreted as acritarchs and a new acritarch taxon is described. GEOLOGICAL SETTING AND AGE The Hedmark Basin is located within the southern limit of the Scandinavian Caledonides. The basin contains the Upper Proterozoic Hedmark Group (text-fig. 1), a sequence that has been comprehensively studied, particularly with respect to its general lithostratigraphy, structural geology, sedimentology and depositional history (Bjorlykke £+ «/. 1967; Nystuen 1981 ; Kumpulainen and Nystuen 1985; Nystuen 1987). The Hedmark Group (Bjorlykke et ah 1967) is a wedge of predominantly detrital rocks with subordinate carbonate and igneous rocks which until recently were widely referred to as ‘sparagmites’. The group has a variable thickness of 1 500^000 m (Bjorlykke et al. 1976) and consists of eight formal rock units in part overlain by epicontinental Cambrian-Silurian deposits which attain a thickness of approximately 1000 m. Both the Hedmark Group and the overlying Cambrian-Silurian strata underwent folding and thrusting during the Caledonian Orogeny. The Hedmark Group is considered to have been deposited in connection with fault-bounded basins (Kumpulainen and Nystuen 1985; Nystuen 1987); either within intra-cratonic rift-valleys (Bjorlykke et al. 1976; Bjorlykke 1978) or an aulacogen (Roberts and Gale 1978; Kumpulainen and Nystuen 1985; Nystuen 1987). The deposition history and facies associations were related to a model of development of the Hedmark Basin which involves succeeding episodes of crustal stretching, deep crustal fracture and ensuing block faulting and a final phase of thermal cooling and slow subsidence (Kumpulainen and Nystuen 1985; Nystuen 1987). The Biskopas Conglomerate (Bjorlykke et al. 1967) is a clastic wedge within the Biri Eormation (Bjorlykke et al. 1976). It occurs in several distinctive fan-shaped bodies in the southern and western part of the Sparagmite Basin (Bjorlykke et at. 1976), and the conglomerates interfinger basinward into the Brottum turbidite sandstones or shales of the Biri Formation (Nystuen 1982). Its thickness varies between 200 and 15 m. I Palaeontology, Vol. .33, Part 2, 1990, pp. 287-298, 2 pls.j © The Palaeontological Association 288 PALAEONTOLOGY, VOLUME 33 i A A A AAA/ PEOfEEOWIC/PiMEmiC | SEDIMENTAKY EOCES ^ TURBIDITES 100 m CmOMM COVER V V V ^ V V BASEMENT ROCKS HIM TILLITE ^ CONGLOMEEATE m TEXT-FIG. 1 . Sketch map showing location of Sparagmite Basins in southern Norway. Composite section of the Hedmark Group in the Lake Mjosa region (after Nystuen 1987). The fossiliferous sample with Papillomembrana derives from the basal sedimentary breccia in the Biskopas Conglomerate (Spjeldnaes 1963). The formation consists of well-rounded pebbles and cobbles and occasional boulders in a coarse-grained, slightly clayey arkose matrix (Bjorlykke et al. 1976) interbedded with massive sandstone layers. The conglomerate contains carbonate and phosphate clasts in its lower part which are believed to have been eroded from the early transgressive part of carbonate deposits in the Biri Formation which accumulated in shelf area adjacent to the basin (Bjorlykke 1966; Spjeldnaes 1967). Absence of sedimentary structures appears to indicate deposition in a submarine environment. The conglomerate beds contain clasts, matrix-supported textures and massive and inversely graded bedding. Accordingly, the conglomerates were interpreted as VIDAL: PROTEROZOIC ACANTHOMORPH ACRITARCHS 289 subaqueous fans laid down in front deltas by gravity sediment flow processes (R. Otter, pers. comm, in Nystiien 1982). Bjorlykke et al. (1976) favoured deposition in connection with floods as coarse sheet flows being deposited directly into the basin, or as a fluvial delta becoming reworked into a submarine environment. They also considered it unlikely that the depositional depth could have been greater than 200-300 m. Bjorlykke et at. (1976) pointed out that the conglomerate resembles Quaternary glaciofluvial conglomerates, and that pebbles are often faceted and similar to Quaternary glaciofluvial pebbles. However, they clarified that there is no evidence of ice-contact and that sorting is generally better than in glacial outwash deposits. Roundness and pebble contents suggest long transport (Bjorlykke et al. 1976). The Biri Formation containing the fossiliferous Biskopas Conglomerate is 50-100 m thick and comprises a variety of lithofacies, including subtidal micritic limestone, intertidal to supratidal carbonates and shales with some dolomite displaying mud cracks and intraformational conglomerates, carbonate platform margin oolithic limestones, platform-slope calcareous shales and sandstones and intra-basin dark shale (Bjorlykke et al. 1976). Rocks of the lower member of the Biri Formation are locally missing, probably as a result of erosion preceding the deposition of the Biskopas Conglomerate. Both the upper and lower junctions of the formation are evidently diachronous. Where the Ring Formation is missing, rocks of the Biri Formation are overlain by the Moelv Tillite, while in areas where the Biskopas Conglomerate is absent the Biri Formation is in contact with the Brottum Formation (Bjorlykke et al. 1976). Normal salinity and open marine deposition conditions were suggested by Bjorlykke et al. (1976). Phosphate pebbles in the Biskopas Conglomerate yielded scattered microfossils (Manum 1967) and specimens of the problematic fossil Papillomembrana eompta (Spjeldnass 1963, 1967) and Erieiasphaera spjeldnaesii sp. nov., and a number of quartz replaced, circular structures in the fossiliferous thin-section may represent sections of microfossils as indicated by Spjeldnaes (1967). In fact, some resemble transverse sections of siliceous casts of vase-shaped microfossils (perhaps similar to Melanocyrillhim Bloeser, in Bloeser 1985) as recorded in siliceous phosphates from the late Riphean Visingsd Group in Sweden (Knoll and Vidal 1980). Acritarchs and cyanobacteria! microfossils (Vidal and Siedlecka 1983; Vidal and Nystuen unpublished data) occur in detrital rocks of several units of the Hedmark Group and in the overlying Lower Cambrian units (Downie 1982; Moczydlowska and Vidal 1986). Age data from rocks of the Hedmark Group are restricted to one single, Rb/Sr whole rock age (Welin unpublished; Rankama 1973) of 612+18 Ma on the Ekre Shale. An indirect estimation of the age of the Hedmark Group is offered by Rb/Sr whole rock dates of the Ottfjallet dolerite dike swarm (Claesson 1976, 1977; Claesson and Roddick 1983) which yielded ages of 720 + 260 and 665±10 Ma. Acritarchs from the Biskopas Conglomerate and the Biri Formation indicate a late Riphean to early Vendian (settsu Vidal and Siedlecka 1983) age for units underlying the Moelv Tillite. DISCUSSIQN Microfossils of organic-walled microorganisms are generally abundant in Proterozoic rocks. Their proper taxonomic affiliation is poorly understood and they are therefore treated among the acritarchs (Downie el al. 1963; Evitt 1963; Downie 1973), although morphologically simple forms are occasionally considered among the cryptarchs (Diver and Peat 1979). The probable algal affinity of Proterozoic and early Palaeozoic acritarchs is generally accepted, as it is the idea that they may represent the abandoned organic envelopes of encysted and/or motile stages of prasinophycean green algae, dinoflagellates, or similar groups now extinct (Dale 1977; Tappan 1980). The diagnostic features of acritarchs are the general shape of the vesicle, its surface ornamentation and (if present) the shape and ornamentation of processes and excystment mechanism (Tappan 1980). Being restricted to these few taxonomically diagnostic attributes, descriptions are often imprecise and accompanied by poor illustrations. These are features which have probably contributed to the erroneous generic attribution of some Proterozoic acritarchs to early Palaeozoic genera. Irrespective of this, the taxonomy of acritarchs is meaningful only at the species level, because acritarch genera are simple groupings of form-species sharing superhcially similar morphological features which do not necessarily imply close biological affinity. Wide-ranging dimensional variability is a feature observed among modern eukaryotic plant protists. This same feature appears to apply to Phanerozoic acritarchs, which generally display a broad size range (5-500 /^m; Tappan 1980). Despite this, the significance of large acritarchs in the total picture of the late Proterozoic biota has been strongly overemphasized in the past. Certain 290 PALAEONTOLOGY, VOLUME 33 Proterozoic acritarchs have been compared to the phycoma stages of prasinophycean green algae (Knoll and Calder 1983; Vidal and Knoll 1983), which may be the living counterparts of several acritarch form-genera such as Leiosphaeridia, Tasmanites and Cymatiosphaera. Prasinophycean phycomata are initially about 10 pm in diameter, but the mature cyst may be as much as 100-175 //m in diameter (Tappan 1980). This attribute may be particularly significant to the interpretation of the taxonomic affinity of certain Proterozoic acritarchs which display polymodal, often megascopic or nearly megascopic dimensions (Vidal 1974, 1976; Vidal and Ford 1985). Until quite recently the generally accepted view was that Precambrian acritarchs are mainly spheroidal, morphologically simple forms lacking diagnostic features (Downie 1973). Timofeev el al. (1976) and Vidal (1976) first reported Proterozoic acanthomorphs ornamented with simple, conical and complex membrane-bounded processes, and polygonomorphic acritarchs, a degree of morphological complexity once thought to be restricted to Phanerozoic fossils (Brasier 1977). However, recent data (Vidal 1981; Knoll and Calder 1983; Knoll 1984; Awramik et al. 1985; Pjatiletov and Rudavskaya 1985; Vidal and Ford 1 985 ; Yin 1 985 ; Butterfield et al. 1 988 ; Knoll and Ohta 1988; Zang 1988; Zang and Walter 1989) seem to demonstrate that complex acritarchs were relatively common in the late Proterozoic. It now appears that Papillomembrana compta Spjeldnges (1963) (and an additional acanthomorph acritarch; see below) from the late Proterozoic Biskopas Conglomerate in the southern Norway Sparagmite Basins (Spjeldnaes 1963) may in fact be the earliest report of giant Proterozoic acanthomorph acritarchs (Zhang 1989). Most occurrences of late Proterozoic acanthomorph acritarchs are in probably early diagenetic cherts within subtidal shallow shelf carbonates (Knoll and Calder 1983; Knoll 1984; Awramik et al. 1985; Yin 1985, 1 987 ; Knoll and Ohta 1 988), and subtidal shallow marine shales and carbonates (Timofeev et al. 1976; Vidal 1976, 1981 ; Vidal and Ford 1985; Pjatiletov and Rudavskaya 1985; Butterfield et al. 1988). Zang and Walter (1989) suggested that a rather diverse acritarch assemblage from the late Proterozoic Pertatataka Formation accumulated in distal turbidites and explained the apparent rarity of late Proterozoic ‘giant’ acanthomorphs as perhaps depending on insufficient sampling of deep-marine settings. Lack of convincing sedimentological evidence to support hypothetical turbidite deposition in the Pertatataka shales leaves the occurrence of Proterozoic ‘giant’ acanthomorphs open to alternative explanations (Knoll and Butterfield 1989). Late Proterozoic (Riphean and Vendian) turbidites have been extensively studied, with results diflfering much from those reported by Zang (1988) and by Zang and Walter (1989). Thus, laminated hemipelagic mud deposits are usually rich in detrital organic matter (Palacios-Medrano 1986) and yielded chiefly planktonic microfossils interpreted as chroococcalean cyanobacteria (Mansuy and Vidal 1983; Vidal and Siedlecka 1983; Palacios-Medrano 1986). The scarcity or absence of identifiable non-cyanobacterial acritarchs in turbidite sequences does not imply total absence of eukaryotic plankton in the overlying water column. In fact, cysts may simply have not been produced or alternatively, if produced, had low fossilization potential or accumulated in insignificant numbers (Vidal and Siedlecka 1983). As for any other microfossil group, the concentration of acritarchs in submarine fan deposits would be expected to be small compared with that of adjacent shallow shelves. Thus, while early Cambrian-age platform siliciclastic and carbonate rocks yield rich acritarch assemblages, time-equivalent turbidites in southeastern Poland yield extremely rare acritarchs (Pozaryski et al. 1981). Little is known about the preservation potential of motile stages of microscopic algae or of reproductive structures of thallophytes (e.g. cysts, aplanospores or zygotes), although some inferences are perhaps possible from dinoflagellate data (Evitt 1985). However, routine microscopic examination of thin-sections of numerous rock samples failed to reveal the existence of such hypothetical stages. In fact, it is possible that, like the geologically more recent dinoflagellates, Proterozoic acritarchs may have alternately produced preservable and non-preservable cysts (Evitt 1985). In the above context environmental information concerning the Papillomembrana-hcar'mg Biskopas Conglomerate is of some interest. Available models favour deposition either as subaqueous fans laid down in front deltas (R. Otter, pers. comm, in Nysteun 1982), coarse sheet VIDAL: PROTEROZOIC ACANTHOMORPH ACRITARCHS 291 flows being deposited directly into the basin, or a fluvial delta becoming reworked into a submarine environment Bjorlykke et al. (1976). In any event it appears that numerous features apparently indicate deposition in a submarine environment. Carbonate and phosphate clasts in the lower part of the formation are believed to have been eroded from early transgressive shallow shelf carbonate deposits in the Biri Formation (Bjorlykke 1966; Spjeldnses 1967). Complex acritarchs occur sporadically in the Biskopas Conglomerate and in the Biri Formation (Vidal and Nystuen, unpublished data). On the other hand, as in many previous reported occurrences (see above), greywackes and hemipelagic mudstones from the turbidite-dominated Brottum Formation yield a monotonous assemblage consisting of benthic (possibly reworked) and planktonic cyanobacterial microfossils. SYSTEMATIC PALAEONTOLOGY The symbol v* means that the type specimen of the species has been examined. All specimens come from a single petrographic thin-section of a phosphorite pebble from the basal part of the Biskopas Conglomerate (for details see Spjeldnjes 1963, 1967) with specimen number PMO 73 1 73. The number refers to the collections of the Palaeontological Museum, Oslo. England Finder coordinates are herein provided for figured microfossil specimens (thin-section label orientated to left side of the microscope stage). Group ACRITARCHA Evitt, 1963 Genus ericiasphaera gen. nov. Type species. Ericiasphaera spjeldnaesii gen. et sp. nov. Diagnosis. As for the type and only species of the genus, Ericiasphaera spjeldnaesii. Derivation. From the Latin ericius, meaning hedgehog, referring to the spiny appearance, and sphaera meaning ‘sphere’, a spiny, hedgehog-like sphere. Ericiasphaera spjeldnaesii sp. nov. Plate 1, fig. I V* 1985 Comasphaeridiwni sp. Yin, p. 239, pi. 2, figs. 1 and 2; text-fig. 12. V* 1985 Baltisphaeridium sp. Yin, p. 239, pi. 4, figs. 5-8. V* 1987 Baltisphaeridium maximum sp. nov., Yin, p. 439, pi. 14, figs. 14 and 15. Diagnosis. Large, spherical or spheroidal vesicle which bears numerous evenly scattered, simple, conical processes. Processes are solid and lack communication with the vesicle cavity. Processes are closely arranged, with conical-shaped bases which taper into cilia-shaped distal portions. Diameter of the vesicle cavity is about 280 //m. Diameter of the periphery including the processes is about 300 //m. Space between conical shaped bases is 2-4 /m. Length of processes is 6-10 //m. Holotype. Specimen PMO 73173, England Finder coordinate H/35. PI. I, fig. 1. Derivation. Named in honour of Professor Nils Spjeldnaes. Type locality. Lower part of Biskopas Conglomerate (Spjeldnees 1963, p. 65) at locality Hjellund. Topographic Map Sheet Gjovik, UTM coordinates 585400/676900. De.scription. The holotype and only specimen in the present material is strongly deformed, thus providing an irregular section of the vesicle. The digitalized periphery was converted into a circular projection, which indicates that the minimum diameter of the vesicle cavity could have been at least 277 //m. The restored circular 292 PALAEONTOLOGY, VOLUME 33 periphery of the distal end of processes indicates a diameter of about 292 //m. The space between the conical- shaped bases is 2^ //m. Digitalized measurements of the processes indicate lengths of 6-10 / 131 on the holotype) evenly spaced, tightly arranged processes. The processes are hollow, cylindrical, with angular proximal contacts and bulbous or bifurcated distal ends. Connection between vesicle and process cavity is not evident. Diameter of the inner cavity is c. 518 pm, while the diameter of the total periphery including processes is c. 768 ^m. Length of processes is 30^2 //m (x 36-5 pm, rr = 3T //m, n = 30), with bases 9-13 pm in width (x = 11-6 pm, cr = L6 //m, n = 9), while the width of the top of the processes is 7-15 //m {x = 9 6 pm, a= \-8 pm, n = 20). Wall thickness not measurable. Holotype. PMO 73173, England Finder coordinates R/31. PI. 1. fig. 3. Type locality. Lower part of Biskopas Conglomerate (Spjeldnaes 1963, p. 65) at locality Hjellund. Coordinates on the Topographic Map Sheet Gjovik, UTM coordinates 585400/676900. Description. The holotype of P. compta is deformed, a feature which results in an irregular section of the vesicle. The compressed vesicle has vertically standing processes, a feature which can be observed at the right corner of the micrograph on Pl.l, fig. 3 (holotype). This indicates that the whole surface of the vesicle was probably covered with cylindrical, bulbous processes (text-fig. 3). Digitalized added images produced from optical sections of the rather thick petrographic thin-section appear to support this conclusion. The digitalized peripheries of the three best-preserved specimens (PI. 1, figs. 2 and 3, PI. 2, figs. 1 and 2) were converted into circular projections which indicate the minimum diameters of the sectioned specimens. In the case of the holotype (PI. 1, fig. 3), the measured inner diameter is 517-8 /rm, while the diameter of the outer periphery including the processes is 767-7 /m. The inner diameter of the specimen in PI. 2, fig. 2 is 331-5 pm. The outer diameter is c. 381 pm. These dimensions are probably not more than rough estimates since the specimen is completely compressed and the inner cavity is not evident. The same appears true of the specimen in PI. 1, fig. 2. The holotype (PI. 1, fig. 3) displays irregular convolute organic strands, most likely irregularly folded, condensed cellular remains. However, it cannot be excluded that they may constitute a deformed inner layer of the vesicle wall. Discussion. Spjeldntes (1963, p. 63) pointed out some superficial resemblance to dasycladacean green algae. He also indicated that the fossil is not carbonate encrusted, a feature which, although common, is not general among dasycladaceans. The general features of the dasycladacean thallus, which has an undivided erect axis bearing whorls of simple or bifurcated lateral branches, are in any event missing in Papillomembrana. The taxonomic affiliation of the fossil remains unknown and is here treated among the acritarchs. Nevertheless, this does not exclude the possibility that the fossil(s) could be reproductive structures of thallophytes (e.g. cysts, aplanospores or zygotes). There is compelling evidence indicating that metaphytic green and red algae were extant in late Proterozoic times (Hofmann 1985; Butterfield et al. 1988; Zhang 1989; Vidal 1989). It is evident EXPLANATION OF PLATE 2 Specimen number refers to the collections of the Palaeontological Museum, Oslo. England Finder coordinates are given for each specimen (label orientated on left side of microscope stage). Figs. 1-3. Papillomembrana compta emend. Vidal. 1, low magnification view showing surrounding phosphate matrix, specimen PMO 73173, England Finder Coordinate V/29, x67. 2, poorly preserved specimen, specimen PMO 73173, England Finder coordinates K/41, x 200. 3, detail view of specimen in 1, note the hollow processes, England Finder coordinate as for specimen in fig. 1. Transmitted light micrograph, x 200. PLATE 2 VIDAL^ Papillomemhrana 296 PALAEONTOLOGY, VOLUME 33 that P. compta resembles a number of comparatively large pre-Phanerozoic acritarchs, and another ‘acritarch-like’ form (E. spjeldncpsii) was found in the same sample. Acknowledgements. Grant support from the Natural Science Research Council (NFR) and the Knut and Alice Wallenberg Stiftelse is acknowledged. Drs David Brutton and Niels Spjeldnaes (Oslo) were most cooperative in making possible the loan of the type specimen of Papillomenihrana for which I express my sincere thanks. Professor Spjeldnas also made available additional thin-sections of phosphates from the Biskopas Conglomerate from his own collections. REFERENCES AWRAMIK, S. M., MCMENAMIN, D. S., YIN CHONGYU, ZHAO ZlQIANG, DING QIXIU and ZHANG SHUSEN 1985. Prokaryotic and eukaryotic microfossils from a Proterozoic/Phanerozoic transition in China. Nature, London, 315, 655-658. BJORLYKKE, K. 1966. Studies on the latest Precambrian and Eocambrian in Norway. No. I. Sedimentarv petrology of the Sparagmite of the Rena district, S. Norway. Norges Geologiske undersokelse, 238, 5-53. 1978. The eastern marginal zone of the Caledonide orogen in Norway. In Caledonian-Appalachian orogen of the North Atlantic Region, IGCP-Project 27. Geological Survey of Canada Paper, 78-13, 49-55. ELVSBORG, A. and H0Y, L. 1967. Late Precambrian sedimentation in the central sparagmite basin of South Norway. Norsk Geologisk Tidsskrift, 56, 233-290. ENGLUND, J. o. and KIRK.HUSMO, L. 1967. Latest Precambrian and Eocambrian stratigraphy of Norway. Norges Geologiske Undersokelse, 251, 5-17. BRASiER, M. D. 1977. An early Cambrian chert biota and its implications. Nature, London 286, 719-720. BUTTERFIELD, N. J., KNOLL, A. H. and swETT, K. 1988. Exceptional preservation of fossils in an Upper Proterozoic shale. Nature, London, 334, 424-427. CLAESSON, s. 1976. The age of the Ottfjallet dolerites of Sarv Nappe, Swedish Caledonides. Geologiska Fdreningen i Stockholm Fdrhandlingar, 98, 370-374. 1977. The age of the Ottfjallet dolerites of Sarv Nappe, Swedish Caledonides: a reply. Geologiska Fdreningen i Stockholm Fdrhandlingar, 99, 405-408. and RODDICK, j. c. 1983. ^“Ar/®“Ar data on the age and metamorphism of the Ottfjallet Dolerites, Sarv Nappe, Swedish Caledonides. Lithos, 16, 61-73. DALE, B. 1977. New observations on Pterdinium faeroensis Paulsen (1905), and classification of small orthoperidinioid dinoflagellates. British Phycological Journal, 12, 241-253. DIVER, w. L. and peat, c. j. 1979. On the interpretation and classification of Precambrian organic-walled microfossils. Geology, 1, 401-404. downie, c. 1973. Observations on the nature of the acritarchs. Palaeontology, 16, 239-259. 1982. Lower Cambrian acritarchs from Scotland, Norway, Greenland and Canada. Transactions of the Royal Society of Edinburgh, Earth Sciences, 72, 257-285. evitt, w. r. and sarjeant, w. a. s. 1963. Dinoflagellates, hystrichospheres, and the classification of acritarchs. Stanford University Publications in the Geological Sciences, 7, 3-16. EVITT, w. R. 1963. A discussion and proposals concerning fossil dinoflagellates, hystrichospheres, and acritarchs, II. Proceedings of the National Academy of Sciences of the United States of America, Washington, 49, 298-302. 1985. Sporopollenin dinoflagellate cysts: their morphology and interpretation. American Association of Stratigraphic Palynologists Eoundation, Austin, Texas, 333 pp. GERMS, G. J. B., KNOLL, A. H. and VIDAL, G. 1986. Latest Proterozoic microfossils from the Nama Group, Namibia (South West Africa). Precambrian Research, 32, 45-62. HOFMANN, H. J. 1985. The mid-Protcrozoic Little Dal microbiota, Mackenzie Mountains, north-west Canada. Palaeontology, 28, 331-335. KNOLL, A. H. 1984. Microbiotas of the Precambrian Hunnberg Formation, Nordaustlandet, Svalbard. Journal of Paleontology, 58, 131 162. 1985. The distribution and evolution of microbial life in the Late Proterozoic Era. Annual Review of Microbiology, 39, 391^17. and BUTTERFIELD, N. J. 1989. New window on Proterozoic life. Nature, London, 337, 602-603. VIDAL: PROTEROZOIC ACANTHOMORPH ACRITARCHS 297 and CALDER, s. 1983. Microbiotas of the Late Precambrian Rysso Formation, Nordaustlandet. Svalbard. Palaeontology, 26, 467^96. and OHTA, Y. 1988. Microfossils in metasediments from Prins Karls Forland, western Svalbard. Polar Research, 6, 59-67. and SWETT, k. 1987. Micropaleontology across the Precambrian-Cambrian boundary in Spitsbergen. Journal of Paleontology, 61, 898-926. and VIDAL, G. 1980. Late Proterozoic vase-shaped microfossils from the Visingso Beds, Sweden. Geologiska Foreningen i Stockholm Fdrhandlingar, 102, 207-211. KUMPULAiNEN, R. and NYSTUEN, J. p. 1985. Late Proterozoic basin evolution and sedimentation in the westernmost part of Baltoscandia, 213-232. In gee, d. g. and sturt, b. a. (eds.). The CaledonUle Orogen - Scandinavia and related areas. John Wiley and Sons, Chichester. MANSUY, c. and vidal, g. 1983. Late Proterozoic Brioverian microfossils from France; taxonomic affinity and implications of plankton productivity. Nature, London, 302, 606-607. MANUM, s. 1967. Microfossils from the Late Precambrian sediments around Lake Mjosa, southern Norway. Norges Geologiske Undersokelse, 251, 45-52. MOCZYDLOWSKA, M. and VIDAL, G. 1986. Lower Cambrian acritarch zonation in southern Scandinavia and southeastern Poland. Geologiska Foreningen i Stockholm Fdrhandlingar, 108, 201 -223. NYSTUEN, J. p. 1981. The late Precambrian ‘Sparagmites’ of southern Norway: A major Caledonian allochthon - the Osen-Roa Nappe Complex. American Journal of Science, 281, 69-94. 1982. Late Proterozoic basin evolution on the Baltoscandian craton: the Hedmark Group, southern Norway. Norges Geologiske Undersokelse, 375, 74 pp. 1987. Synthesis of the tectonic and sedimentological evolution of the Proterozoic-early Cambrian Hedmark Basin, the Caledonian Thrust Belt, southern Norway. Norsk Geologisk Tidsskrift, 67, 395—418. PALACIOS-MEDRANO, T. 1986. MicrofosHes de pared organica del Proterozoico Superior (region central de la Peninsula Iherica). Unpublished Ph.D. thesis, Universidad de Zaragoza (Spain). 131 pp. PJATILETOV, V. G. and RUDAVSKAYA, V. G. 1985. Acritarchs from the Yudomian Complex, 151-158. In sokolov, B. s. and IVANOVSKY, A. B. (eds.). Vendskaya Sistema, [Vendian Sv5tp/n], Vol. I. Moscow, Nauka [In Russian]. POZARYSKi, w., VIDAL, G. and BROCHWicz-LEWiNSKi, w. 1981. New data on the Lower Cambrian at the southern margin of the Holy Cross Mountains (SE Poland). Bulletin de I'Academie Polonaise des Sciences de la Terre, 29, 167-174. RANKAMA, K. 1973. The Late Precambrian glaciation, with particular reference to the Southern Hemisphere. Proceedings of the Royal Society of New South Wales, 106, 89-97. ROBERTS, D. and GALE, G. H. 1978. The Calcdonian-Appalachian lapctus Ocean, 255-342. In Tailing, D. H. (ed.). Evolution of the Earth's Crust. Academic Press, London and New York. SPJELDN/ES, N. 1963. A new fossil (Papillomemhrana sp.) from the Upper Precambrian of Norway. Nature, London, 200, 63-64. — - 1967. Fossils from pebbles in the Biskopasen Formation in southern Norway. Norges Geologiske Undersokelse, 251, 53-82. TAPPAN, H. 1980. The paleobiology of plant protists. Freeman and Co., San Francisco, 1028 pp. TIMOFEEV, B. V., GERMAN, T. N. and MIKHAILOVA, N. s. 1976. MicrofitofossHH dokemhriya, kemhriya i ordovika. (Microphytofossils from the Precambrian, Cambrian and Ordovician). Academy of Sciences, USSR. Institute of Precambrian Geology and Geochronology, Leningrad, Nauka, 106 pp. [In Russian]. VIDAL, G. 1974. Late Precambrian microfossils from the basal sandstone unit of the Visingso beds. South Sweden. Geologica et Palaeontologica, 8, 1-14. 1976. Late Precambrian microfossils from the Visingso Beds in southern Sweden. Fossils and Strata, 9, 57 pp. 1981. Micropalaeontology and biostratigaphy of the Upper Proterozoic and Lower Cambrian sequences in East Finnmark, northern Norway. Norges Geologiske Undersokelse, Bulletin, 362, 53 pp. 1989. Are late Proterozoic carbonaceous megafossils metaphytic algae or bacteria? Lethaia, 22, 375-379. and FORD, T. D. 1985. Microbiotas from the Late Proterozoic Chuar Group (Northern Arizona) and Uinta Mountain Group (Utah) and their chronostratigraphic implications. Precanibrian Research, 28, 349^89. and KNOLL, A. H. 1983. Proterozoic plankton. Geological Society of America, 161, 265-277. and siEDLECKA, A. 1983. Planktonic, acid-resistant microfossils from the Upper Proterozoic strata of the Barents Sea region of Varanger Peninsula, East Finnmark, Northern Norway. Norges Geologiske Undersokelse, Bulletin. 382, 45-79. 298 PALAEONTOLOGY, VOLUME 33 YIN LEiMiNG 1985. Microfossils of the Doushantuo Formation in the Yangtze Gorge District, western Hubei. Palaeontologia Cathayana, 2, 229-249. 1987. Microbiotas of latest Precambrian sequences in China, 417^94. In Stratigraphy and Palaeontology of Systemic Boundaries in China. Precamhrian-Cambrian Boundary, 1. ZANG WENLONG 1988. Ediacaran plankton. 7th International Palynologiccd Congress, Brisbane 1988, Abstracts, and WALTER, M. R. 1989. Latest Proterozoic plankton from the Amadeus Basin in central Australia. Nature, London, 337, 642-645. ZHANG YUN 1989. Multicellular thallophytes with differentiated tissues from Late Proterozoic phosphate rocks of South China. Lethaia, 22, 113-132. 184. GONZALO VIDAL Micropalaeontological Laboratory Geological Institute University of Lund Box 124 Typescript received 27 February 1989 Revised typescript received 19 June 1989 S-221 00 Lund Sweden A DISCOGLOSSID FROG FROM THE MIDDLE JURASSIC OF ENGLAND by SUSAN E. EVANS, ANDREW R. MILNER FRANCES MUSSETT Abstract. A discoglossid frog, Eodiscoglossus oxoniensis sp. nov. is described from the Upper Bathonian Forest Marble of Oxfordshire. It closely resembles Eodiscoglossus santonjcie from the Jurassic-Cretaceous boundary of Spain but can be distinguished by characteristics of the ilium and premaxillary. The E. oxoniensis specimens represent the earliest European material critically identifiable as a frog and the earliest discoglossid yet recognised. An association of Eodiscoglossus with Alhanerpeton and a Marniorerpeton-Wkc salamander may have characterized certain freshwater ecosystems in Europe for about 50 million years from the Bathonian to the Barremian-Aptian. The fossil record of frogs prior to the Cretaceous is poor. The single specimen of Triadohatrachus from the Lower Triassic of Madagascar demonstrates that stem-anurans with just a few antiran skeletal characteristics had evolved by the beginning of the Mesozoic (Rage and Rocek 1986, 1989; Milner 1988). However, no other Triassic anurans are known and few frogs have been described from Jurassic rocks, although these are all crown-group representatives with the full suite of anuran skeletal characteristics. They are known from eight localities and are reviewed in the discussion (below). The Middle Jurassic frog material described here was obtained from the microvertebrate assemblage in the Kirtlington Mammal Bed at Kirtlington in Oxfordshire from which Freeman (1979) first recorded frog material. It represents the first discoglossid frog to be reported from pre- Upper Jurassic rocks and also the earliest known critically determinable frog material from Europe. The specimens described and figured here were collected either by Professor K. A. Kermack and colleagues or by Mr E. F. Freeman and have been donated to the Department of Palaeontology, British Museum (Natural History) (BMNH). Mr Freeman is undertaking palaeoecological work with his collections and the specimens collected by him retain his catalogue number (prefixed by EF). Comparative study was also made of Eodiscoglossus material at the Museum National d’Histoire Naturelle, Paris (MNHN). LOCALITY AND HORIZON The new material was collected from various parts of the Old Cement Works Quarry, near Kirtlington in Oxfordshire, (Ordnance Survey Grid Reference SP 495200; Freeman 1976, 1979; Kermack et cd. 1987). The techniques of collection and preparation of microvertebrates from this locality were described by Kermack et al. (1987). The productive horizon is the Kirtlington Mammal Bed, near the base of the Forest Marble, which is of Upper Bathonian age (approximately 170 Ma; Harland et al. 1982). A full account of the local stratigraphy is given by Freeman (1979). The palaeoenvironment appears to have been a shallow non-stagnant water body, with occasional influxes of poorly sorted sediment (Freeman 1979). The Mammal Bed has produced a rich microvertebrate fauna of which only some of the mammals (Freeman 1976, 1979; Kermack et cd. 1987) and salamanders (Evans, Milner and Mussett 1988) have been described so far. I Palaeontology, Vol. 33, Part 2, 1990, pp. 299-31 1.| © The Palaeontological Association 300 PALAEONTOLOGY, VOLUME 33 TEXT-FIG I. Eodiscoglossus oxoniensis n. sp. a, BMNH R. 11700, holotype right ilium in lateral aspect; />, BMNH R. 11720, right ilium in lateral aspect; c, BMNH R. 11703, right premaxillary in lingual aspect; d, BMNH R. 11704, right maxillary in lingual aspect; e, BMNH R. 11707, atlas centrum in dorsal aspect. SYSTEMATIC PALAEONTOLOGY Class AMPHIBIA Order anura Suborder discoglossoidei Sokol, 1977 Family discoglossidae Guenther, 1859 Genus eodiscoglossus Villalta, 1957 Type species. Eodiscoglossus santonjae Villalta, 1957. Range. Bathonian to Barremian/Aptian ; Spain and Great Britain. Diagnosis. Discoglossid frog resembling Discoglossus in one derived character: ilium with dorsal crest and dorsal tubercle; and several primitive characters: 15-18 premaxillary teeth, about 50 maxillary teeth, coronoid process smooth and convex with no notches, anterior vertebrae bearing free ribs and posterior vertebrae bearing no ribs, iliac synchondrosis absent. Discussion. Eodiscoglossus has no apomorphic characters, but is more plesiomorphic than Discoglossus in at least three features: pterygoid process of maxillary poorly developed, elongate flattened atlantal cotyles, neural arches lacking upturned flared posterior margins. It is more EVANS ET At.: MIDDLE JURASSIC FROG 301 plesiomorphic than the similar Wealdenhatrachus from the Barremian/Aptian of Una, in that the ilium bears a single dorsal tubercle whereas Wealdenhatrachus has a pair of tubercles in this position. The diagnosis is based on characters, most of which can be seen in both the Montsech and the British material. It is not entirely satisfactory in that there are no apomorphic characters for the genus, and it is probable that Eodiscoglossus, as defined here, is a primitive grade of Discoglossiis- group discoglossid. However, as E. oxoniensis lacks obvious autapomorphic characters and clearly has a close phenetic resemblance to E. santonjae, a new genus cannot be justified and it is most practical to place both species in one genus. Eodiscoglossus oxoniensis sp. nov. Text-figures 5, 6a-c. Derivation of name. From the county of Oxfordshire. Holotype. British Museum (Natural History) Department of Palaeontology R. 1 1700, a right ilium lacking the end of the iliac blade and part of the acetabulum (text-figs. It/, 6t/, 6h). Paratypes. BMNH R. 11701 (EF 75:10:1:6), R. 11720, 2 right ilia with fully preserved acetabular regions (text-figs. \h and 6c); R. 11702, 11703, 2 right premaxillaries (text-figs. Ic, 2t/-c); R. 11704, 11705, 2 right maxillaries (text-figs. It/, 2e, f); R. 11707, R. 11708, 2 broken atlas centra (text-figs. \e and a-e). Referred material. 5 premaxillaries, 33 maxillaries, 9 angulosplenials (including R. 1 1706), a broken atlas (R. 1 1721), 30 isolated neural arches (including R. 11709-R. 11712), 7 broken scapulae (including R. 11722, R. 1 1723), 2 right humerus distal heads (R. 1 1713, R. 1 1714), a radioulna (R. 1 1715), 51 ilia, 2 ischia (including R. 11716 (EF 76:13/14:36:4)) and 10 tibiofibulae (including R. 1 1718 (EF 75:3: 1 : 10) and R. 11719 (EF 76:4:1:2)). Locality. Old Cement Works Quarry, Kirtlington, Oxfordshire, England, Ordnance Survey Grid Reference SP 495200. Horizon. Kirtlington Mammal Bed, near base of the Forest Marble, aspidioides Zone, Upper Bathonian, Middle Jurassic. Diagnosis. Species of Eodiscoglossus in which the ilium shows the following features in contrast to that of E. santonjae: iliac shaft flattened and broad but narrow in cross-section with lateral ridge; little waisting at the junction of the shaft and the acetabular region; dorsal tubercle poorly developed, shallow and flush with the surface of the iliac shaft: supraacetabular fossa deep. The premaxillary of E. oxoniensis apparently has a low alary process in contrast to the elongate process of E. santonjae. Other bones appear to be indistinguishable in the two species. DESCRIPTION General features The present material comprises about 160 elements as listed above. Scaled against skeletons of Rana temporaria, the larger elements belong to medium-sized frogs of 80 mm snout-vent length, although many of the bones derive from smaller animals. There is no more than one morphological type of any given bone and this, coupled with the numbers of ilia (53) and maxillaries (35), strongly suggests that only a single form is present. The bones are either diagnostically discoglossid or consistent with attribution to the Discoglossidae and so unity of the material is assumed. 302 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. Eodiscoglossus oxoniensis n. sp. a, BMNH R. 11702, right premaxillary in lingual view; b, c, BMNH R. 11703, right premaxillary in h, lingual and c, labial views; c/, reconstruction of right maxillary in lingual view; e, BMNH R. 1 1704, anterior region of right maxillary in lingual view;yi BMNH R. 1 1705, detail of partial right maxillary showing pedicels and developing crowns of teeth; g, BMNH R. 11706, left angulosplenial in dorsal view. Scale bars = 1 mm. Abbreviation: co. pr. coronoid process. Skull Premaxillary (text-figs Ic and 2a-c). Seven specimens were collected, none of which is complete. The premaxillary has a broad pars dentalis with at least 15 tooth positions (about 18 in £. santonjae, Vergnaud- Grazzini and Wenz, 1975, p. 22). The lateral region of the pars dentalis is long and the medial region is short, as in other discoglossids including E. santonjae. The alary process is low and of moderate width, quite unlike that of other discoglossids including E. santonjae in which this process is as tall as the bone is wide (e.g. Vergnaud-Grazzini and Wenz 1975, fig. 1). It is convex anteriorly and concave posteriorly, with a deep medial excavation which probably received a peg of cartilage from the nasal capsule (text-figs. Ic and 2b). At the anteromedial junction of the alary process and the pars dentalis, there is an excavation showing that the alary processes were separated in the midline. The pars palatina (palatal shelO is narrow laterally and wider medially. At the medial end, it curves sharply backwards into a medial expansion. The end of this expansion is incomplete, so it is not possible to determine whether it was pointed or blunt posteriorly. Maxillary (text-figs. 1 Ic/and 2d-f). About 35 maxillaries were recovered, none of which is complete. The pars dentalis is long and bears approximately 50 tooth positions, a similar number to that in E. santonjae. The pars facialis is long and divided into three regions: (i) a straight, narrow anterior process with a medial, slightly concave overlap surface for the premaxillary, anterior to the tooth-row (text-fig. 2e); (ii) behind the anterior process, the bone expands dorsally for a short distance back to the leading edge of the orbit; (iii) further back, the bone levels off and runs back as a low wall below the orbit. EVANS ET AL.: MIDDLE JURASSIC FROG 303 Medially, at the level of the anterior orbital border, the maxillary bears a depression which opens into a foramen for the superior alveolar nerve. Laterally, the maxillary bears a shallow longitudinal groove at the level of the base of the tooth row. Otherwise the lateral face of the bone is featureless except for several small sensory nerve foramina. The pars palatina, or palatal shelf, begins just behind the anterior process of the bone as a low rounded ridge. Posteriorly, however, this expands into a small narrow shelf - the pterygoid process - where the lateral part of the pterygoid abuts against the maxillary. This process ends at the same level as the tooth row, but the bone continues for a short distance behind the tooth row although the posterior tip is not preserved in our material. The shape of the maxillary bears a close resemblance to that of Wealdenhatrachus (Fey 1988, fig. 22). Dentition (text-fig. 2f). The premaxillary and maxillary teeth are slender and pedicellate. The crowns are always lost on fully erupted teeth but several specimens show isolated crowns either at the tooth bases or moving into position on broken teeth. The crowns are small and bicuspid, and show no other obvious specialization. Hecht (1970) regarded the maxillary of Eodiscoglossus as toothless, but the material described by Vergnaud-Grazzini and Wenz (1975, fig. 1) shows that E. santonjae has toothed premaxillaries and maxillaries. Angulosplenial (text-fig. 2g). The angulosplenial is represented by nine specimens. It bears a coronoid process which is a long low convex bulge with no anterior or posterior notches. TEXT-FIGURE 3. Eodiscoglossus oxoniensis n. sp. a-t/, BMNH R. 1 1707, atlas vertebra in a, anterior, h, posterior, c, ventral and d, dorsal views; e, BMNH R. 1 1708, atlas vertebra in left lateral view; f\ g, BMNH R. 1 1709, anterior trunk vertebral arch in /, anterior view and g. lateral view of rib facet; h, BMNH R. 11711, posterior trunk vertebral arch in dorsal view; i, J, BMNH R. 11712, posterior trunk vertebral arch in /, dorsal and /, lateral views. Scale bar = I mm. 304 PALAEONTOLOGY, VOLUME 33 Axial skeleton Atlas vertebra (text-figs, le and 3a-e). Three incomplete atlantal centra were recovered. The atlantal centrum is dorsoventrally flattened. The anterior cotyles are of depressed oval shape. Medially, they are moderately separated by an intercotylar region bearing a small notochordal pit (text-fig, 3a), but with no medial notch as seen from above or below. The long axes of the anterior cotyles are not horizontal but are orientated at a shallow obtuse angle. Posteriorly, there is a small circular cotyle which may be imperforate (R. 1 1707, 1 1708) or perforate (R. 11721). The presence of this cotyle implies that the following trunk vertebra is opisthocoelous with an anterior condyle. The atlantal centrum is anteroposteriorly short but broad. The ventral surface is smooth except for a few small pits on either side of the midline. The dorsal surface is concave, with weak grooves on either side of a small rounded central ridge. The neural arch pedicel is broad-based but narrows dorsally, leaving an anterolateral notch for the exit of the first spinal nerve, and a long sloping posterolateral border. This atlas with such flattened, slightly separated anterior cotyles corresponds to the type II atlas of Lynch (1971). The distinction between this and the Lynch type III atlas in which the cotyles are confluent, is not always clear. Trueb (1973) identified the atlas of the extant leiopelmatids as type III but now considers them to be type II (pers. comm, in Clarke, 1988). Estes and Sanchiz (1982b) identified the atlas of the Galve material of E. santonjae as type III, but it appears to be very similar to the atlantal centrum described here. The degree of separation of the cotyles is apparently variable and not always clear in imperfect material. The apparent difference between the type II atlas of E. oxoniensis and the type III atlas of E. santonjae described by Estes and Sanchiz may not be of great significance. Trunk vertebrae (text-fig. if-i). No complete trunk vertebrae were collected, although over 30 broken neural arches were recovered. The vertebrae have very narrow pedicels and lightly built arches which are apparently TEXT-FIG. 4. Eodiscoglossus oxoniensis n. sp. a-c, BMNH R. 1 1722, broken left scapula in a, lateral, b, posterior and c, medial views; cL BMNH R. 1 1723, broken left scapula in posterior view; e-g, Rana tetnporaria, left scapula in c, lateral, j\ posterior and g, medial views. Scale bars = I mm (a-d), 5 mm (c-g). EVANS ET AL.. MIDDLE JURASSIC FROG 305 easily broken. The neural arches are flattened, with almost horizontal pre- and postzygapophyses. The pedicel is narrow and there is a small horizontal backwardly-directed neural spine between the postzygapophyses (e.g. text-fig. 3/;, /). Between the neural spine and the postzygapophyses, the posterior surface of the arch is excavated into deep pockets for intervertebral ligaments. The neural arches were probably imbricate but there is no suggestion of flaring of the posterior margin as in some discoglossids. Each pedicel bears a slender transverse process, usually broken at the tip. In a few isolated arches, however, the lateral process is expanded distally and bears a pitted terminal surface for the attachment of a free rib (text-fig. 3/ g). Other lateral processes were not terminally expanded and presumably bore no rib, not even a rudiment (text-fig. 3/). In E. semtonjae and Wealdenhatrachus, free ribs are present on the anterior presacrals only and this appears to have been the condition in E. oxoniensis. Although no trunk centra are known, the presence of a posterior cotyle on the atlantal centrum means that the first trunk vertebra must have had an anterior condyle and have been opisthocoelous. In the absence of other evidence, all the presacral trunk vertebrae are assumed to have been opisthocoeloLis. Only three frog families have such presacral vertebrae, namely the Discoglossidae and the pipoid families Rhinophrynidae and Pipidae. A ppeiulicular skeleton Scapula (text-fig. Aa-d). Of the pectoral girdle elements only 7 broken scapulae have been recovered. They are all too incomplete for the general shape to be determined and a scapula of Rana teniporaria is figured comparatively (text-fig. 4c-g) to clarify the orientation of the fragment figured in text-fig. Aa-c. The scapulae appear to have been bicapitate, i.e. with separate articulations for the clavicle (pars acromialis) and coracoid (pars glenoidalis). Although the pars acromialis is not visible on any specimen, it is clear that there is a distinct pars glenoidalis demarcated ventrally by a deep pocket (text-fig. Ad) which must have separated the ventral region of the scapula into two heads. TEXT-FIG. 5. Eodiscoglossus oxoniensis n. sp. n, BMNH R. 1 1713, distal head of right humerus in ventral view; /), c, BMNH R. 11715, left radioulna in fi, ventral and c, cross-sectional views; d, BMNH R. 11718, incomplete tibiofibula, midshaft region together with proximal and midshaft cross-sections; c, BMNH R. 11719, incomplete tibiofibula, distal shaft region, / g, BMNH R. 1 1716, ischial plate in /, right lateral and g, ventral views. Scale bars = I mm. Abbreviations: co. articular condyle; f.c.v. fossa cubitus ventralis; r. radius; t. tibia; u. ulna. 306 PALAEONTOLOGY, VOLUME 33 Fore-limb (text-fig. 5a-c). Of the fore-limb bones, only the distal heads of two humeri (R. 11713 and R. 1 1714) and a left radioulna (R. 11715) have been recovered. The shape of the humeral heads suggests that both derive from right humeri, but there is little preserved except the hemispherical articular condyle and a deep fossa cubitus ventralis (Estes and Sanchiz I982fi). The radioulna is distinguished from the tibiofibulae by its shorter broader shape (text-fig. 5b, c). There is a prominent radioulnar groove along the visible length of the specimen, a very primitive feature, but as the specimen derives from a very small individual, this may simply reflect immaturity. I r TEXT-FIG. 6. Eodiscoglossiis oxoniensis n. sp. a, b, BMNH R. 11700, holotype right ilium in a, lateral and b, medial views; c, BMNH R. 1 1701, right ilium in lateral and cross-sectional views; d, Eodiscoglossus santonjae MNHN MSE.5, right ilium. Scale bar = 1 mm. Abbreviations; ac. acetabulum; d.c. dorsal crest; d.tb. dorsal tubercle; l.r. lateral ridge; sa.fs. supraacetabular fossa; sa.pr. supraacetabular process; sba. pr. subacetabular process. Uium (text-figs, la, b and 6a-c). More than 50 ilia were collected, most showing only the thickened acetabular region. Three specimens, R. 1 1700 (the holotype), R. 11701 and R. 1 1720 show the important features. The following description uses the terminology of Vergnaud-Grazzini (1966). The iliac shaft is broad, anteriorly recurved and mediolaterally flattened (text-flg. 6c). The dorsal crest is moderately developed and separated from the smooth convex lateral surface by a narrow groove. Posteriorly, at the base of the shaft, the ilium bears an elongate but shallow tubercle (for the musadus iliofemoralis), separated from the remainder of the bone by dorsal and ventral pits. The acetabular region is thickened, but its ventral articular surfaces are completely preserved only in two specimens, R. 1 1701 and R. 1 1720. The acetabulum is prominent and roughly oval, lying towards the anterior edge of the bone (text-figs. \b and c). Anterodorsally, a small pit (for the muscidiis iliacus internus Estes and Sanchiz 1982 a) separates the acetabular surface from the shaft. Anteroventrally, the bone is drawn out into a small triangular subacetabular process. Posteroventrally, there is a larger supraacetabular process, separated from the acetabulum by a long supraacetabular fossa (text-fig. 16). The ventral border of the bone is lightly forked, with pitted surfaces for the pubis and the ischium. Medially, the acetabular region shows little detail. There is no evidence of an iliac synchondrosis, although the medial edge of the pubic facet is roughened, suggesting a ligamentous attachment. In comparison with the ilium of Eodiscoglossus santonjae (MNHN MSE. 5) (text-fig. 6d), that of E. oxoniensis is similar in general structure but several consistent dilTerences can be itemized. (i) In cross-section, the iliac shaft of E. santonjae is triangular, one corner of the triangle being formed by the prominent lateral ridge which extends up the lateral face of the shaft (text-fig. 6d). In E. oxoniensis, there is no lateral ridge and the shaft is narrow in cross-section (text-fig. 6c). (ii) The shaft of E. santonjae is narrow in lateral view whereas that of E. oxoniensis is flattened and broad. (iii) At the junction of the shaft and the acetabular region, the ilium of E. santonjae is sharply waisted; this is less marked in E. oxoniensis. (iv) The dorsal tubercle of E. santonjae is more prominent than that of E. oxoniensis. That of E. santonjae EVANS ET A L. \ MIDDLE JURASSIC FROG 307 extends well above the surface of the iliac shaft, while the dorsal tubercle of E. oxoniensis is shallow and flush with the surface. (v) The supraacetabular fossa is deeper and more marked in E. oxoniensis than in E. santonjae. These features are constant on all the ilia of E. oxoniensis recovered and the differentiating characters were constant on those ilia of E. santonjae which were examined and serve to distinguish the species. The ilium of the recently described Cretaceous discoglossid Wealdenbatrachus jiicarensis (Fey 1988, figs 32-35) is similar in general shape to those of both Eodiscoglossus species. In the holotype specimen (Fey 1988, figs. 34 and 35), there is a dorsal tubercle which is prominent like that of E. santonjae but there is also an accessory tubercle. This tubercle is not found in either Eodiscoglossus species and appears to be the most diagnostic character of Wealdenbatrachus. The ilia of the paratype specimens of Wealdenbatrachus (Fey 1988, figs. 32 and 33) differ from that of E. oxoniensis in that they show greater development of the supra- and subacetabular processes and greater differentiation of the shaft and crest. Ischium (text-fig. 5/i g). Two specimens of fused ischia were recovered (R. 11716, R. 1 1717). The compound bone is semicircular with a pitted margin and a pronounced posteroventral ridge radiating out from the acetabular region as in modern Rana. The posterior region is not preserved and it is not clear whether there was a posterodorsal expansion or not. Tihiofibula (text-fig. 5d, e). Ten tibiofibular shafts were recovered (e.g. R. 1 1718, R. 1 1719). The larger tibia and smaller fibula are firmly fused, being barely distinct in the central shaft but partly separated by deep grooves towards the proximal and distal ends. The tibiofibulae were long and gracile resembling those of jumping anurans such as Rana rather than walking anurans such as Bufo. DISCUSSION Interrelationships of primitive frogs and the systematic position ofE. oxoniensis The Leiopelmatidae {Leiopehna, Ascaphus) and Discoglossidae (Discoglossiis, Alytes. Barhouriila, Bomhina) are widely perceived as the most primitive families of living frogs. Clarke (1988) has recently completed a 95-character analysis of the osteology of all but one of the living species in these two families (Bomhina fortinnptialis was not available for study), and has concluded that each family is monophyletic but that their interrelationships are uncertain. The Leiopelmatidae and Discoglossidae are frequently grouped together as the Discoglossoidei or Discoglossoidea, but it is not clear whether this group is a monophyletic sister-clade to the remaining frogs or a primitive grade of frog with the Discoglossidae closer to the higher frogs, Sokol (1975, 1977) has argued that the Discoglossoidei are a clade and that the Discoglossidae and Leiopelmatidae share derived characters of the tadpole branchial system, namely (i) absence of the interbranchialis III muscle and (ii) extensive fusions between the copula II and the hypobranchials. No characters to support this relationship have been found in the adults however and in Lynch’s cladogram (1973, fig. 3.6), the Discoglossidae share two characters with the higher frogs, namely: (i) presacral column reduced to eight vertebrae or fewer and (ii) muscuhis caudaiiopnhoischiotihialis lost. Both sets of characters are small and the interrelationship of leiopelmatids, discoglossids, and higher frogs is effectively an unresolved trichotomy. Clarke (1988) has used his osteological data to analyse the internal relationships of the extant genera and species of the Discoglossidae. He concluded that Alytes is the sister-taxon to the other genera and that, within the remaining forms, Discoglossus is the sister-taxon to Barhouriila and Bomhina. The following discussion of the characters of E. oxoniensis is based, where possible, on the derived characters supporting this hypothesis of relationships. Eodiscoglossus oxoniensis can be placed within the family Discoglossidae on the basis of two derived characters. Neither is unique to the Discoglossidae, but the combination characterizes only this family and one pipid genus, namely Hymenochirus. (i) Opisthocoelous vertebrae. These only occur in three anuran families, the Discoglossidae, Pipidae and Rhinophrynidae (Trueb 1973). This Kirtlington material shows no other general features of pipids or rhinophrynids. 308 PALAEONTOLOGY, VOLUME 33 (ii) Ilium bearing a dorsal crest. Such ilia are found in the Discoglossidae, Leptodactylidae, Ranidae, and the pipid genus Hynienoclririis, but the Kirtlington material shows no other characteristics of the latter three taxa. In Clarke’s (1988) hypothesis of relationships based on osteology, the subfamily Discoglossinae (Discoglossiis, Bomhina and Barhoiirula) is defined by six osteological characters. One is found in E. oxoniensis: (iii) Occipital condyles with major axes at shallow or steep angle. The occiput of E. oxoniensis is unknown but the atlantal cotyles slope upwards at a shallow angle and so presumably the occipital condyles must have as well. Clarke also defines the Discoglossinae by the presence of neural arches with flared posterior margins but these are not present in E. oxoniensis. Other derived features, which occur in some but not all discoglossines, are also found in E. oxoniensis. These characters are not distributed congruently with each other, or with other characters within the Discoglossidae (Clarke 1988), but broadly support a relationship between E. oxoniensis and the Discoglossinae : (iv) Imbricated vertebrae. These are found in most but not all discoglossine species. (v) Groove at the base of the tooth row on the lateral face of the maxillary. This occurs in Discoglossiis and one species each of Bomhina and Barhoiirula. (vi) Neural spines moderately developed. This occurs in Bomhina and some Discoglossiis species. The Kirtlington material can be associated with the genera Discoglossiis, Wealdenhatracliiis, and Eodiscoglossiis on the basis of one derived character: (vii) The ilium has a distinct dorsal tubercle, not as well-developed as in E. santonjae or the other two genera but more so than in other discoglossids (Estes and Sanchiz 1982r/; Fey 1988; Clarke 1988). There are no derived characters which suggest immediate relationship to Alytes, Barhoiirula or Bomhina. The Kirtlington material can be associated with Eodiscoglossiis on the basis of one character of uncertain polarity: (viii) The atlas has extremely flattened anterior cotyles (Estes and Sanchiz 1982/)). As noted under the generic diagnosis, there is no certain derived character shared by E. santonjae and E. oxoniensis, but the phenetic resemblance is sufficiently great that a new genus cannot be justified. Finally, the Kirtlington material merits a separate species because of the differences in ilium and premaxillary construction listed under the species diagnosis and the descriptions of those elements. Eodiscoglossiis oxoniensis does not provide major new insights into the anatomy of early discoglossid frogs. However, it does combine a derived character of Discoglossiis (ilium with dorsal tubercle) with a primitive feature that places it outside the Discoglossiis- Barhoiirula- Bomhina group (neural arches lack flared posterior margin). This suggests that the characters used to define the taxa, based on modern material, have either been subject to convergence or reversal, or their polarity is not fully understood. The new material extends the stratigraphical range of the genus Eodiscoglossiis and the family Discoglossidae down to the Bathonian, and the geographical range of both taxa to include the British Isles. Finally, although the evidence is not strong, such osteological characters as are present suggest that the ^/vto-group and Discoglossiis- Barhourida- Bomhina group had already differentiated by 170 Ma. The fossil record of Jurassic frogs Jurassic frogs have now been described from six localities and reported from a further two. The sole described Fower Jurassic frog is the type and only specimen of VieraeUa lierhstii from the Roca Blanca Formation of Argentina (Reig 1961; Casamiquela 1965; Estes and Reig 1973). Estes and Reig assigned this genus to the Feiopelmatidae (referred to as the Ascaphidae in that work). However, they noted that the leiopelmatid characters of the specimen were all primitive anuran characters and VieraeUa could equally be a stem-frog with no immediate relationship to any modern EVANS ET AL.\ MIDDLE JURASSIC FROG 309 family. It does not prove the existence of a cladistically defined Leiopelmatidae in the Lower Jurassic. The only Middle Jurassic specimens described in the literature are a possible anuran omosternum from the Bajocian of Aveyron (Seiffert 1969; but see Estes and Reig 1973 for a critical reappraisal) and the Kirtlington material first reported by Freeman (1979) and described in this paper. Upper Jurassic frogs have been described or reported from five areas as follows. (i) In the nineteenth century, the Morrison Formation of Como Bluff, Wyoming, U.S.A. produced a few fragments of frog skeleton including two humeri which have been named: Eohatrachus cigilis Marsh and Comohatruchus aeuignuitis Hecht and Estes (reviewed by Hecht and Estes 1960; Estes and Reig 1973). The latter authors concluded that the Eohatrachus humerus might belong to a pipoid but could not be determined further, while Comohalrachus was indeterminate. Estes and Sanchiz (1982«) noted that the Coinohatrachus humerus had some resemblances to those of discoglossids. Further frog material was collected from the Morrison Formation between 1968 and 1970 but has not yet been described (Prothero and Estes 1980, p. 484). (ii) The Matildense Formation of Argentina has produced several specimens of a frog named Notobatracluis degiustoi. Principal descriptions are by Reig (1957), Casamiquela (1961) and Estes and Reig (1973). Estes and Reig assigned Notohatrachus to the Leiopelmatidae (as the Ascaphidae). (iii) The lithographic limestones of the Sierra del Montsech, Lerida, Spain, dated as uppermost Jurassic or basal Cretaceous, have produced several specimens of at least two types of frog. Most are of the discoglossid frog, Eodiscoglossus saiitonjae, which has most recently been described or discussed by Hecht (1963, 1970), Estes and Reig (1973), and Vergnaud-Grazzini and Wenz (1975). A single specimen of a second frog, Neusihatrachus wdferti, was described by Seiffert (1972) and it can be assigned to the Palaeobatrachidae (Estes and Reig 1973; Vergnaud-Grazzini and Wenz 1975). A third named form, also based on a single specimen, is Moutsechohatrachus gaudryi (Vidal 1902). This poor specimen is generally agreed to be indeterminate, although some features suggest that it may be a palaeobatrachid (Estes and Reig 1973; Vergnaud-Grazzini and Wenz 1975). (iv) Anuran material has been reported, but not described, from the Lower Kimmeridgian lignites of Guimarota, Portugal (Seiffert 1973). (v) New localities in the Purbeck Formation of Dorset, England have recently produced fragments of an anuran which have not yet been determined (Ensom 1988). This material is currently being studied by two of the authors (S.E.E. and A.R.M.). The fossil record as yet permits us to make very few testable statements about the evolution and diversification of frogs in the Jurassic. It is clear that true frogs were present in the Lower Jurassic, but there is no evidence for differentiation into recognizable modern families at that time. By the Bathonian, discoglossids were not only present but may have begun to dift'erentiate as discussed above. However, because of the uncertainty of the interrelationships of primitive frog families to higher frogs and to each other, we cannot yet say which other frog families might be expected also to be present. By the Jurassic-Cretaceous boundary, differentiation at least into leiopelmatids, discoglossids, and palaeobatrachids had taken place. Ecology and chronology The Kirtlington assemblage is incompletely described at present, but preliminary quantification of the amphibian material suggests that it may in future be possible to recognize the associations or communities which included Eodiscoglossus. In the samples studied, the Eodiscoglossus material could not have come from fewer than 28 specimens (right ilia). The five other amphibians recognized, together with the minimum numbers of individuals represented, are; Mannorerpeton kermacki (19 atlantes), Mannorerpeton freemani (1 atlas), a third small salamander (4 atlantes) a primitive salamander (340 atlantes) and an albanerpetontid (1 atlas). It appears that the assemblage incorporated a major association of Eodiscoglossus, Mannorerpeton kermacki and the primitive salamander, with the other forms as exotic elements in the fauna. This association may have been long-lived, at least at the family level. Estes and Sanchiz (19826) 310 PALAEONTOLOGY. VOLUME 33 described similar material from Galve in Spain, including several specimens each of Eodiscoglossiis, Albanerpetoiu and an unnamed Marmorerpeton-\\kt salamander. The small salamander Galverpeton was represented by only a single specimen. The Galve assemblage is Barremian-Aptian and hence 125-1 13 Ma in age (Harland et al. 1982), so it is possible that an amphibian faunal association of Eodiscoglossiis, albanerpetontid, and Marmorerpeton (or similar forms) may have characterized certain freshwater ecosystems in Europe for over 50 million years from the Middle Jurassic to the late Lower Cretaceous. Testing this association against other faunas might eventually be possible but at present most of the assemblages of Mesozoic lissamphibians from Spain and Portugal (e.g. Guimarota, Una) are still undescribed. Acknowledgements. Our thanks go to Professor Kenneth A. Kerinack who organized the collection and preparation of most of the material, and to Mr Eric F. Freeman who first recognized the potential of the locality, and who generously made significant specimens available to us for study. We also thank Dr Barry T. Clarke for reading the manuscript and permitting us to use information in his unpublished thesis; Mr Chris Syms for preparing text-fig. 1 ; Dr Sylvie Wenz for enabling S.E.E. to examine material of E. santonjae\ and two anonymous referees for constructive criticism. S.E.E.’s study in Paris was supported by the University of Fondon Central Research Fund. REFERENCES CASAMiQUELA, R. M. 1961. Nucvos matcrialcs de Notohatraclnis deginstoi Reig. Revista del Miiseo de La Plata, Paleontologia {Nueva Serie), 4, 35-69. 1965. Nuevo material de Vieraella herhstii Reig. Revista del Miiseo de La Plata, Paleontologia (Nueva Serie), 4, 265-317. CLARKE, B. T. 1988. Evolutionary relationships of the discoglossoid frogs - osteological evidence. Unpublished Ph.D. thesis. City of London Polytechnic. ENSOM. p. c. 1988. Excavations at Sunnydown Farm, Langton Matravers, Dorset: amphibians discovered in the Purbeck Limestone Formation. Proceedings of the Dorset Natural History and Archaeological Society, 109. 148-150. ESTES, R. and reig, o. a. 1973. The early fossil record of frogs: a review of the evidence. 1-63. In vial, j. l. (ed.). Evolutionary biology of the anurans : contemporary research on major problems. University of Missouri Press, Columbia. and SANCHiz, b. 1982«. New discoglossid and palaeobatrachid frogs from the late Cretaceous of Wyoming and Montana, and a review of other frogs from the Lance and Hell Creek Formations. Journal of Vertebrate Paleontology, 2. 9-20. 19826. Early Cretaceous lower vertebrates from Galve (Teruel), Spain. Journal of Vertebrate Paleontology, 2, 21-39. EVANS, s. E., MILNER, A. R. and MUSSETT, F. 1 988. The earliest known salamanders (Amphibia : Caudata) : record from the Middle Jurassic of England. Geobios, 21, 539-552. FEY, B. 1988. Die Anurenfauna aus der Unterkreide von Una (Ostspanien). Berliner geowissenschaftliche Abhandlung, 103, 1-99. FREEMAN, E. F. 1976. Mammal teeth from the Forest Marble (Middle Jurassic) of Oxfordshire, England. Science (New York), 194. 1053-1055. 1979. A Middle Jurassic mammal bed from Oxfordshire. Palaeontology, 22, 135 166. HARLAND, W. B., COX, A. V., LLEWELLYN, P. G., PICKTON, C. A. G., SMITH, A. G. and WALTERS, R. 1982. A geologic time scale. Cambridge University Press, 131 pp. HECHT, M. K, 1963. A reevaluation of the early history of frogs. Part II. Systematic Zoology, 12, 20-35. 1970. The morphology of Eodiscoglossiis, a complete Jurassic frog. American Museum Novitates, 2424, 1-17. and ESTES, R. 1960. Fossil amphibians from Quarry Nine. Postilla, 46, 1 19. KERMACK, K. A., LEE, A. J., LEES, P. M. and MUSSETT, F. 1987. A new docodoiit from the Forest Marble. Zoological Journal of the Linnean Society, 89. 1-39. LYNCH, J. D. 1971. Evolutionary relationships, osteology, and zoogeography of leptodactylid frogs. University of Kansas Museum of Natural History, Miscellaneous Publications, 53, 1-238. 1973. The transition from archaic to advanced frogs. 133-182. In vial, j. l. (ed.). Evolutionary biology of the anurans: contemporary research on major problems. University of Missouri Press, Columbia. EVANS ET AL.\ MIDDLE JURASSIC FROG 311 MILNER, A. R. 1988. The relationships and origin of living amphibians. 59-102. In benton, m. j. (ed.). The phylogeny ami classification of the Tetrapods, Vol. 1 : Amphibians, Reptiles, Birds, Syslematics Association Special Volume 35A. Clarendon Press, Oxford. PROTHERO, D. R. and ESTES, R. 1980. Late Jurassic lizards from Como Bluff, Wyoming and their palaeobiogeographic significance. Nature (London), 286, 484-486. RAGE, J.-c. and ROGER, z. 1986. Triadohatrachus revisited. 255-258. In ROGER, z. (ed.). Studies in Herpetology. Charles University, Prague. 1989. Redescription of Triadobatraehus niassinoti (Piveteau, 1936) an anuran amphibian from the early Triassic. Palaeontographica, A, 206, 1-16. REiG, o. A. 1957. Los anuros del Matildense. In stipanigig, p. n. and reig, o. a. El complejo porfirico de la Patagonia extraandina y su fauna de anuros. Acta Geologia Lilloana, 2, 231-297. 1961. Noticia sobre un nuevo anuro fosil del Jurasico de Santa Cruz (Patagonia). Ameghiniana, 2, 73-78. SEiFFERT, J. 1969. Sternalelement (Omosternum) eines mitteljurassischen Anuren von SE-Aveyron/ Siidfrankreich. Zeitschrift fur zoologische Systenuitik und Evolutionsforschung, 2, 145-153. 1972. Ein Vorlaufer der Froschfamilicn Palaeobatrachidac und Ranidae im Grenzbereich Jura-Kreide. Neues Jahrbuch fur Geologic und Paldontologie, Monatshefte, 1972, 120-131. 1973. Upper Jurassic lizards from Central Portugal. In Contribuigao para o conhecimento da Fauna do Kimeridgiano da Mina de Lignito Guimarota (Leiria Portugal). Ill parte. Servicos Geologicos de Portugal, Menioria, 22 (Nova Series), 1-88. SOROL, o. M. 1975. The phylogeny of anuran larvae: a new look. Copeia, 1975, 1-23. 1977. A subordinal classification of frogs (Amphibia: Anura). Journal of Zoology, 182, 505-508. TRUER, L. 1973. Bones, frogs and evolution. In vial, j. l. (ed.). 65-132. Evolutionary biology of the anurans: contemporary research on major problems. University of Missouri Press, Columbia. VERGNAUD-GRAZZiNi, G. 1966. Ees amphibiciis de Miocene de Beni-Mellal. Notes et Memoires du Service geologicjue du Maroc, 27, 43-74. and WENZ, s. 1975. Les discoglossides du Jurassique superieur du Montsech (Province de Lerida, Espagne). Annales de Paleontologie (Vertebres), 61, 19-36. VIDAL, L. 1902. Nota sobre la presencia del tramo Kimeridgense en el Montsech (Lerida) y hallazgo de un batracio en sus hiladas. Memorias de hi Reed Academia de Ciencias v Artes de Barcelona, 4, 263. SUSAN E. EVANS Department of Anatomy and Developmental Biology University College London Rockefeller Building, Llniversity Street London WCIE 6JJ ANDREW R. MILNER Department of Biology Birkbeck College Malet Street London WCIE 7HX FRANCES MUSSETT Department of Biology University College London Medawar Building, Gower Street London WCIE 6BT NOTE ADDED IN PROOF Since completion of the manuscript, vertebral material, including a sacrum and urostyle, has been recovered. The trunk centra and sacrum are identical to those figured for E. santonjae (Estes & Sancjiz 1982f?). The urostyle has small anterior transverse processes, as in many recent frogs, and is consistent with attribution to the Discoglossidae. Typescript received 7 February 1989 Revised typescript received 19 May 1989 I f r, H: 1 LATE CAINOZOIC BRACHIOPODS FROM THE COAST OF NAMAQUALAND, SOUTH AFRICA by C. H. C. BRUNTON Cllld N. HILLER Abstract. An unusual late Tertiary - early Quaternary brachiopod assemblage from shallow water shoreline deposits on the Namaqualand coastal plain of South Africa is described. New species described are Kraussiuu roliimhita, K. laevicostata, K. cimeatu and Cuncellothyris platys, with subspecies C. platys platys and C. platys petalos. In situ specimens, shell growth, abrasion and epizoans all indicate crowded living conditions, commonly on bedrock. Diversity, shell size and shell thickness are consistent with waters having been warmer than in the region today; Atlantic cooling took place from the late Tertiary. Tertiary and Quaternary brachiopod faunas are poorly known throughout the world, with a few notable exceptions, such as those from New Zealand or the Mediterranean. Their presence, commonly in shallow marine sediments, can provide important evidence about the marine conditions of their locations during this period of major temperature and sea-level fluctuations. We were delighted, therefore, to be presented with a well-preserved late Cainozoic brachiopod fauna from Namaqualand, South Africa, particularly as it contains representatives of genera much larger than any seen hitherto, and the in situ relationship with the bedrock of some species allows the formulation of well-founded palaeoecological conclusions. Faunas such as this help our understanding of the origins of Recent brachiopods and we are able to suggest possible connections between this fauna and the Recent brachiopods of southern Africa. Since the discovery of diamonds on the west coast of southern Africa (Wagner and Merensky 1928), considerable attention has been given to the nearshore marine sediments of late Tertiary to early Quaternary age, including papers on their contained fossils. Haughton (1932) presented an overview of the west-coast deposits, and described a number of mollusc species and the brachiopod Kraiissina lata. Since then, papers by Carrington and Kensley (1969), Kilburn and Tankard (1975), Kensley (1972, 1977) and Kensley and Pether (1986) have added to the knowledge of the molluscan fauna. This project started in the mid 1970s when Dr A. J. Carrington presented one of us (C.H.C. B.) with a small collection of brachiopods, thought to be of Pliocene to Pleistocene age, collected during diamond exploration on the Namaqualand coast. Because data on their provenance were not available, work on the specimens ceased until Dr B. Kensley and Mr J. Pether presented us with well-documented comparable material from the same region. This paper describes seven brachiopod species belonging to three genera which were recovered during diamond-mining activities on the three properties Koingnaas, Hondeklip and Avontuur A in the Hondeklip Bay area of Namaqualand coast (text-hg. 1). Geological Setting. Over the years a number of authors have described and interpreted the Cainozoic coastal stratigraphy of the South African west coast. In particular, papers by Carrington and Kensley (1969), Tankard (1975) and Hendey (1981«, h) have helped to elucidate the succession and determine the depositional environments. In this paper we follow the scheme of Pether ( 1986c/), who has summarized and revised the previous work. Pether’s brachiopods were recovered from two regressive sedimentary units that are at present included in the Alexander Bay Formation (South African Committee for Stratigraphy 1980; Pether I986«). Within each unit, named after the immediately preceding transgressive altimetric maximum, Pether (1986r/) has recognized lower IPalaeonlology, Vol. 33, Part 2, 1990, pp. 313-342, 4 pls.| © The Palaeontological Association 314 PALAEONTOLOGY. VOLUME 33 TEXT-FIG. 1. Locality map with inset showing the positions of the main collecting localities (stars). — 1 km I 1 TEXT-FIG. 2. Schematic cross-section showing the stratigraphic relationships of the two regressive sedimentary units (after Pether 1986a)- T = younger terrestrial deposits; FS = foreshore facies; USH = upper shoreface; LSH = lower shoreface; NS = nearshore shelf; BB = back-barrier; BC = barrier complex. Vertical lines indicate an erosional remnant of a nearshore shelf deposit belonging to an earlier (mid-Pliocene) 90 m Unit. shoreface, upper shoreface and foreshore facies (text-fig. 2). The older 50 m Unit, which rests on an eroded Precambrian gneissic basement, also contains sediments interpreted as belonging to a barrier complex, including tidal inlet deposits, and back-barrier environments. Pether (pers. comm. 1986) has reported that brachiopods are found scattered throughout the sands belonging to both units, but those in the 50 m Unit are better preserved than those in the 30 m Unit, having suffered less breakage and abrasion before burial. This may reflect the higher energy open coast conditions prevailing during deposition of the younger unit compared to the more sheltered, quieter conditions of the partially barred coast during deposition of the 50 m Unit (see Pether 1986t/ and text-fig. 2). BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 315 The only information with the Carrington specimens are markings indicating the mid or upper E stage and the Koingnaas locality, where mining exploration was proceeding at that time. The E Stage terminology was introduced by mining geologists and has been variously correlated during subsequent years, but here we follow Pether (1986r/; see our text-hg. 3). The Carrington collection is less well preserved than the more recently collected material from Pether, and some was probably reworked in the Upper E Stage. Unless specifically mentioned, the descriptions of species are based upon the Pether collections. AGE Ma EPOCH Pether (1986) terminology ‘ E ’ stage terminology 0.4 0.6 0.8 1.0 1.2 1.4 1.6 18 2.0 3.0 4.0 5.0 LU Z HI O o I- W LLI UJ z ytj O o oc < UJ LU I- < LU 30m U 50mU 90mU FS USH LSH FS USH LSH NS NS upper E middle E lower E TEXT-FIG. 3. Correlations of terminology and ages of the sedimentary units recognised in Namaqualand, based on Pether 1986«. Age of the fauna. The precise age of the Cainozoic deposits of Namaqualand has been the subject of debate for some time. In the earlier literature a broad late Tertiary - Recent age was considered good enough, but Carrington and Kensley ( 1969) concluded, on the basis of their mollusc studies, that both of the regressive units described above were of Pleistocene age. Later work by Tankard (1975) and Hendey (1981a, h) suggested that these deposits were of Pliocene age. Kensley and Pether (1986) and Pether (1986a) have re-examined the evidence and on the basis of the percentage of extant species in the molluscan fauna, along with comparisons with the Neogene and Quaternary sea-level curves of Vail and Hardenbol (1979) and Beard et al. ( 1982), they concluded that the 50 m Unit is of late Pliocene age and the 30 m Unit can be dated as early Pleistocene (text-fig. 3). More recently, the recovery of a vertebrate fossil renders this interpretation problematic and Pliocene ages for both units must be considered (Pether 19866). THE BRACHIOPODS The late Pliocene - early Pleistocene brachiopod fauna of the Namaqualand coastal plain is a very conservative one, but even so the diversity appears to be a little greater than that of the brachiopod fauna occupying a similar position offshore at the present time. The late Tertiary - early 316 PALAEONTOLOGY, VOLUME 33 Quaternary fauna comprises seven species compared to only four found in water less than 20 m deep around the southern African coast today, namely Terebratulina ahyssicola (Adams and Reeve, 1850), Megerlina pisum (Lamarck, 1819), Discinisca tenuis (Sowerby, 1847) and Kraussina rubra (Pallas, 1766). Of these four, the first two are apparently confined to the warmer waters of the east coast, D. tenuis is found only in the cold waters along the Namibian Coast, and K. rubra ranges from west to east coasts (Penrith and Kensley 1970). This last-named species seems to be the only element of the fossil fauna to have survived to the present, perhaps as a result of its greater range of temperature tolerance. Phyletic relationships of the various members of the fauna are diflicult to establish. The Kraussinidae are regarded as rapidly evolved neotonous forms that appeared with apparent suddenness in the Upper Tertiary (Asgaard 1986). Recently Collins et al. (1988) have used immunological techniques on skeletal macromolecules from Recent brachiopods to suggest affinities within the terebratulids. Their study indicated a close relationship between Kraussina and Megerlia, followed by a relationship with Argyrotheca, as might be expected from current classification, all three being terebratellaceans. The earliest Megerlia specimens are Miocene with Kraussina, here described, already established by late Pliocene. Elliott (1949) for Recent Megerlia, and Ruggiero (1985) for Pleistocene Megerlia from southern Italy, have shown that the earliest brachidium in brachial valves up to about 5 mm wide is a simple V-shaped structure similar to that of adult Kraussina, hence the suggestion of neotenous development. Collins et al. ( 1988) further suggested that the next level of relationship for these terebratellaceans is with terebratulaceans, such as the living Gryplius and Liothyrella. This break from the usual classification is challenging. However, the immunological work is in its infancy and this result may prove false; certainly their study, if accepting the above derivation of Kraussina from a Megerlia- like species, has to explain the derivation of a terebratellacean loop in Megerlia from the short- looped terebratulaceans, within which are few known wide hinged, strongly costate Mesozoic shells looking at all as if they could have given rise to the kraussinids. Possibilities would seem to be limited to Meonia and similar cancellothyrids. The alternative opinion of derivation from the long- looped terebratellaceans might lead to investigation of species like the late Cretaceous (?dallinid) Gemmarcula. The South African Cretaceous and Palaeogene brachiopod faunas, best known from the eastern coastal regions, are not known to contain any terebratellaceans. K. lata Haughton, 1932 and K. laevicostata sp. nov. are recorded from only the 50 m Unit, so they do not seem to extend into the Pleistocene. K. rotundata sp. nov. is tentatively recorded from the 30 m Unit, as well as the 50 m Unit and thus ranges from the Pliocene into the Pleistocene. K. rubra can be taken as extending from the Pliocene to the Recent; it occurs in Cape waters along with K. crassicostata Jackson, 1952 and the poorly known K. cognata (Sowerby, 1847), a species almost devoid of ribs. Thus there is, at present, a gap in the geological record of Kraussina in South Africa equivalent to the middle and late Pleistocene, and it is during this period that any possible connections between the fossil and extant forms must exist. On the basis of its fairly coast ribbing and quite convex shape, one might suggest that K. crassicostata is derived from K. lata, but evidence for this connection is lacking. Palaeoecology Most of the specimens were recovered from sands deposited in an upper-lower shoreface setting which Pether ( 1986rt) interprets as representing water depths of 1-5 m to 10 m. The associated biota includes crustaceans (crabs and barnacles), gastropods, bivalves, scaphopods, polyplacophorans, an occulinid coral, bryozoans, foraminifera, and possible sedentary polychaetes. Rare occurrences of specimens in life position are known (text-fig. 4). Kraussina rotundata sp. nov. and Cancellothyris platys platys subsp. nov. have been found in situ in crevices in the Precambrian bedrock where they tend to form species clusters, although the two occur in close association. Evidence from the collected and studied specimens indicates that K. rotundata lived more commonly within these crevices than C. platys, especially at the Koingnaas locality. In this area almost all specimens of K. rotundata have abraded posterior regions; in some it is so severe as BRUNTON AND HILLER; LATE CAINOZOIC BRACIIIOPODS 317 TEXT-FIG. 4. Photograph showing specimens of Kraiis- sina rotimdata sp. nov. in life position in crevices in Precambrian bedrock exhumed as a result of diamond mining activity. The locality, on Avontuur A, repre- sents a palaeodepth of 4-5 m near the upper shore- face/lower shoreface boundary. The pen is approxi- mately 1 50 mm long. to have removed part of the cardinal process and part of the sockets and teeth. This has led to the enlargement of the pedicle aperture, but in life much of this opening would have been closed by a tough ‘skin’ surrounding the actual pad-like pedicle. This ‘skin’ incorporated muscles which attached to the pedicle collar and these, along with the extensive adjustor muscles (see K. rotimdata description) moved the shell around the pedicle and, by contraction, held the shell tightly to the substrate in a closed condition. Many of the shells display abraded areas on their flanks where the valves moved against rock, while opening and closing. Others display growth distortion as a result of growth around either neighbouring shells or protruding rocks (text-fig. 6a-c; PI. 3, figs. 12-16). The intense nature and local distributions of these abrasions and distortions shows that they occurred during the life of the brachiopod, rather than after death, and this is further supported by the good preservation of other areas of valve surface. Similar, closely clumped associations of living K. rubra are to be found (text-fig. 5g, h) in which younger shells are attached to older individuals which attached to rocks or pebbles. In these situations the individual shells display a little distortion and abrasion of the ventral umbo, but in comparison to the crevice dwelling K. rotimdata, they are free to move in open water. Cancellothyris specimens display less abrasion and distortion. This may be because of their different pedicles. Unlike Kraussimi, the pedicle of living Cancellothyris is long and narrow, furnished with contractile muscles and the ability of adjusting the growing shell to constraints in its immediate environment. The shell of Kraussina could not lift much above its original attachment point, whereas Cancellothyris could, thus enabling its valves greater freedom of growth and movement. Presumably the slight abrasion of the pedicle aperture seen on some Cancellothyris occurred while the shells were held down onto the substrate when closed and ‘avoiding danger’. On other specimens there are patches of abrasion on the flanks posteriorly, commonly affecting both valves, indicating that the umbos were probably attached within a rock crevice. More rarely, in some specimens short lengths of their lateral commissures have growth distortions which have produced cavities (PI. 4, fig. 11). Within these the growth lines developed normally, although somewhat accentuated, but the normal geometry of shell secretion was interrupted by some hard object impinging on the edges of the shell. The internal surfaces of the valves in these areas, apart from being internally convex, appear normal. From the dispositions of the growth lines in these cavities it would seem that biological interference with the valve margins did not take place, but that 318 PALAEONTOLOGY. VOLUME 33 as the shell grew, a short length of its margins was increasingly prevented from normal growth. A protrusion from the edge of the rock crevice in which the specimen grew seems the most likely cause. Another feature, seemingly related to the close hold-fast nature of the Kraiissina pedicle and the abrasion of shell affecting the articulation, is the anterior zig-zag commissure. A commissure of this sort provides great stability to the positioning of the valves when closed and resists torsion between the valves. Normally, terebratulid articulation is sufficiently strong to resist any torsion between the valves, but in those specimens that have weakened articulation, through shell abrasion, the zig-zag commissure must have been helpful in aiding stability. In two examples, severe posterior abrasion has led to a slight dislocation of one valve relative to the other so that the anterior commissure apparently no longer fitted exactly, although the ribs of one valve still partially interlock with the interrib spaces of the other valve. Some species range down into the distal lower shoreface and nearshore shelf environment, and Kraiissina laevicostata sp. nov. and Cancellolhyris plalys petalos sp. et subsp. nov. appear to be confined to this deeper water niche. Table 1 shows the distribution of the fossil brachiopods between the two sedimentary units. This indicates that all seven species are found in the older 50 m Unit, whereas only three are known from the 30 m Unit. Commenting on the diversity contrast between the two units, Kensley and Pether (1986) state that the higher diversity of the 50 m Unit, also TABLE I. Species of brachiopods from three localities in the Hondeklip Bay area of the Namaqualand coast, showing the regressive unit and depositional facies from which they were obtained. USH = upper shoreface; LSH = lower shoreface; NS = nearshore shelf 50 m Unit 30 m Unit Species USH LSH NS USH LSH Cancellothyris platys platys sp. et subsp. nov. X X X X X Cancellothyris platys petalos sp. et subsp. nov. X Kraiissina rubra X X X X X Kraiissina laevicostata sp. nov. X Kraiissina lata X X Kraiissina rotimdata sp. nov. X X X(?) Kraiissina ciineata sp. nov. X(?) Pelagodisciis (?) sp. X X EXPLANATION OF PLATE 1 Scanning electron micrographs of borings and encrusters. Figs. I and 2. Cancellothyris platys platys sp. nov. brachial valve interior. BD6757. 1, part of the posteromedian region, with part of the cardinal process on the right and one crus extending to the bottom edge, showing the blistered appearance of the valve floor resulting from intense boring from the outer surface, x 12. 2, an enlargement of the valve floor showing shell mosaic, endopuncta and the irregular, blistered shell growth, x 80. Figs. 3-8 are all examples of Kraiissina rotimdata sp. nov. with various encrusters and borers. 3, an example of the foraminifera Cibicides lobatidiis (Walker and Jacob) adhering immediately anterior to a growth line. BD67I9. X 64. 4, a cluster of the bryozoan Hippollioa in an interrib space. BD6720. x 55. 5, part of a colony of the sheet-like bryozoan Ccdleporella growing away from the valve margin (bottom left). BD67I6. x 24. 6, part of the bryozoan Tidndipora, associated with Hippollioa, on the mid region of a brachial valve. BD6715. x 24. 7, the mid-lateral region of a brachial valve from which much of the primary shell layer has been removed by ‘grazing’, perhaps with the destruction of the edge of the Celleporeila colony. Hippollioa has grown onto the ‘grazed’ area. BD6716. x 23. 8, the thick-shelled umbonal area of the brachial valve with excavated pits showing some signs of marginal chipping or scratching. The abraded umbo is to the top right. BD67I6. x22. PLATE 1 BRUNTON and HILLER, Namaqualand brachiopods 320 PALAEONTOLOGY, VOLUME 33 reflected in the molluscan fauna, is either real or is possibly an unavoidable bias due to the nature of the available exposures. The reduction in diversity from the 50 m Unit to the 30 m Unit through to the present time possibly reflects a gradual reduction in water temperatures. Kensley and Pether (1986) document the presence of warmer water West African and east coast forms among the fossil molluscs from Hondeklip Bay. On this basis they concluded that the two sedimentary units were deposited in waters significantly warmer than at present and, although the evidence is not clear-cut, they inferred a cooling trend over the period of deposition of the two units. Dingle et al. (1983) have summarized the available data on South African Neogene and Pleistocene palaeoclimates. These point to a general cooling from the time the upwelling of cold central Atlantic water in the Benguela Current system was established in Upper Miocene times (Siesser 1978). The most intense upwelling occurred during the Pleistocene and surface water temperatures reached a minimum during the last glacial maximum (Embley and Morley 1980). Detailed localized palaeotemperature data from oxygen isotope analysis of fossil oyster shells from the Hondeklip Bay area are still awaited. All the species in this fauna attached themselves to the substrate by means of a strong functional pedicle and were thus well adapted for living in high energy shallow water conditions. All the species other than K. ciineata are larger than other known representatives of these genera, and the shell material in K. lata, K. laevicostata and to a lesser extent in K. rotundata is much thicker than met with elsewhere. The secretion of abundant calcium carbonate in shell material is associated with waters warmer than now found off Namaqualand. Externa! encnisters, grazers and borers Many of the well preserved specimens, especially K. rotundata from Avontuur, have bryozoans, foraminifera and rare arenaceous tubiculous polychaetes encrusting their surfaces. Less commonly, specimens have been abraded or bored by organisms. The commonest encrusting bryozoan is the ascophoran cheilostome, Hippothoa, a chain-like and branching form having a distal orifice with a sinus (PI. 1, figs. 4, 7). Less commonly the ascophoran sheet-like genus Celleporeila (PI. 1, fig. 5) and stem-like hexagonally patterned Tubidipora (PI. 1, fig. 6) are preserved, but abraded. Living species of these genera are characteristically epiphytes, living on algae, but they also encrust stones and shells. Hippothoa commonly can be seen to have originated posteriorly, on either valve, and grew essentially anteriorly, as the brachiopod grew. In strongly ribbed areas the bryozoan grew mostly in the interrib spaces (PI. 1, fig. 4) and, less commonly, transversely within the ‘step’ of major growth lines. Overgrowth is quite common and where a valve is shared also by Tubidipora they overgrew each other, although anteriorly the latter overgrew Hippothoa more frequently, as well as individual Cibicides foraminifera. Only one shell has an extensive colony of Celleporeila (PI. 1, fig. 5). There is no bryozoan encrustation on any of the areas of physical abrasion on K. rotundata, indicating that the bryozoans grew on the brachiopods while living in their rock crevices. There is, however, some sign (PI. 1, fig. 7) of Hippothoa growing on areas which appear to have been abraded by grazing benthos (see below). The flattened, normally attached foraminifera Cibicides lobatulus (Walker and Jacob 1789) is found both adherent to outer surfaces of these brachiopods (PI. 1, fig. 3) as well as loose within the sands filling some shells. This species is known on Recent algae, but on Kraussina lived between ribs or below the ‘steps’ of major growth lines near the margins of shells. Kraussina and Cancellothyris specimens display scratch marks resembling those made by the radula of chitons or by echinoids while grazing. The normal valve exterior probably was not attractive to grazers, but if acting as the substrate for algae, sponges, or bryozoa, these surfaces may have provided valuable nutrient (PI. 1, fig. 7). None of the above encrusters, nor signs of grazing abrasion, have been found on inner valve surfaces, so most, if not all, the associations occurred during the life of the brachiopods. Most species, but especially Cancellothyris, display borings from the valve exteriors which, in BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 321 some specimens (text-fig. IOe), has resulted in the removal of the complete external surface over considerable areas. These areas start as a series of minute holes dispersed over a circular area (text- fig. IOd, e) about 1-5-2 mm in diameter. Progressive removal of shell led to the pitted, corroded surface of extreme examples. Many of these borings appear to have gone sufficiently deeply into the shell material to have affected the mantle epithelia, which reacted by secreting extra shell material over the areas involved. The resultant inner surface is clearly thickened, somewhat irregular and tending towards a blistered appearance (PI. 1, figs. 1 and 2). Such a sequence of events shows that the borings occurred during the life of the brachiopod. Other elongate trough-like borings, at the ends of which are small (c. 0-5 mm) circular holes penetrating the valve thickness, occur rarely (PI. 2, figs. 9 and 10). Several K. rotimdata specimens are excavated posteriorly, where the shell is thick, leaving pits up to 5 mm in diameter penetrating up to 2 mm into the shell substance. The edges and bottoms of these may display marks which appear to be scratches left by a predator (PI. 1, fig. 8). Their disposition makes it unlikely that they were produced by echinoids. Apparently confined to Cancellothyris specimens are microscopic boring ramifications attributed to microscopic marine algae or fungi. The extent of infestation varies from a few patches posteriorly to most of the shell substance. These borings would seem to have invaded the thick shell at the umbos and sockets and to have spread anteriorly from there. Although they extend through the shell thickness they are best seen just below the relatively smooth internal surfaces, and we feel that the endolithic microorganism is more likely to have invaded after the brachiopod’s death. The grazings and some borings removed shell material during the brachiopod’s life. The endolithic ramifications, where intense, weakened the general shell fabric, so that in concert with the loss of organic material (fibre sheaths etc.) from the secondary layer, the shell material was weakened and made vulnerable to physical degradation. Collins’s (1986) study of taphonomy in a moderately-deep brachiopod community showed how the shell of Recent Terehralulimi weakened over a period of about 200 days, becoming increasingly liable to physical breakage. This was attributed to the loss of organic material, perhaps resulting from the action of moulds or bacteria. Thus the occurrence of many of these brachiopods, from a shallow-water environment yet in a good state of preservation, is surprising, and we suggest results only from rapid burial by the regressive sands in which they occur. Biogeography The dominant genus is Kraussina, endemic to African, especially southern African, waters at the present time. The only previous fossil records of the genus are also from South Africa. A rather unexpected connection with the Australian region comes in the form of Cancellothyris which is not known from off South Africa at the present, but occurs in Australian waters. A Miocene species is recorded from New Zealand. Pelagodiscus is a very widespread genus in the modern oceans, although it usually occurs in deep waters. However, possible shallow water fossil species are described from Belgium and England. At the species level, four members of the fauna described herein are named as new species, one cannot be named because of insufficient material, and two previously described species, K. lata and K. rubra, are known only from South Africa. Thus it may be concluded that the Namaqualand brachiopod fauna is endemic, although there is a possible connection with the Australasian region. One of the extinct gastropods from the 50 m Unit, Argohuccinum casus, is also linked to that region (Pether pers. comm. 1986). SYSTEMATIC PALAEONTOLOGY Specimens are housed in the British Museum of Natural History, London (BD and ZB registration numbers), or the South African Museum, Cape Town (SAM numbers). 322 PALAEONTOLOGY, VOLUME 33 Class INARTICULATA Huxley, 1869 Superfamily discinacea Gray, 1840 Family discinidae Gray, 1840 Subfamily disciniscinae Schuchert and LeVene, 1929 Genus pelagodiscus Dali, 1908 Pelagodiscus(!) sp. Text-tig. 5a-f Materiul and horizon. A total of thirty-one brachial valves plus some fragmentary material that may include parts of pedical valves, from the 50 m Unit on the farm Hondeklip. Description. Small subcircular brown phosphatic shells; brachial valves are conical with the apex more or less centrally situated, or slightly posterior of centre. The outline is almost circular, although many specimens display an almost straight posterior margin. Height of the cone is equal to about one-half of the shell diameter. In profile, the posterior slope is slightly shorter and steeper than the anterior slope. Ornament of irregular concentric growth lines, but one specimen displays faint radial ornament developed after the shell attained a diameter of 4-3 mm. The pedicle valve is unknown. Dimensions. Typical dimensions (in mm) of specimens are as follows, where a = anterior to posterior diameter, b = left to right diameter, c = height of cone. a b c SAM PQ HB 913 10-6 10-4 5-8 12-3 11 -9 60 14-8 12-4 6-8 11-3 10-9 5-3 SAM PQ HB 662 8-5 8-6 4-5 6-0 6-9 3-2 Discussion. The specimens described here are tentatively assigned to Pelagodiscus on the grounds that the overall morphology of the brachial valve is virtually identical to that of the type species, the living P. atlanlicus (King 1868). The various extant species of the closely related genus Discinisca, all tend to have the apex of the cone more posteriorly situated or they have a marked radial ornament. However, the main difference between these genera is in the form of the lophophore, with Pelagodiscus possessing a schizolophe while that of Discinisca is spirolophous. Thus, without the soft parts it is impossible to be unequivocal about the current assignment. The earliest-formed shell of some limpets living off South Africa resemble these shells, but we reject them for two reasons. There is no sign of twist to their umbos, as seen in the molluscs, and secondly, the shell material is a mineral virtually unknown amongst molluscs. Infra-red spectrum and X-ray analysis indicates that the shell material is a carbonate-fluorine-hydroxy substitute apatite, similar to Dahllite. The presence of chitin was not clearly demonstrated by the analysis, but may have been screened. Fossil Pelagodiscus are poorly known. Thomson (1927) included in the genus two Tertiary species; Discina suessi Bosquet 1858 from the Lower Miocene of Belgium, which he included tentatively, and Discina fallens Wood 1874 from the Crag deposits of East Anglia, England which are of late Pliocene-early Pleistocene age. If the latter is indeed a Pelagodiscus then it represents another shallow water species. P. atlanticus is perhaps the most widespread of all living brachiopod species, living mostly in the depths of the abyssal and lower bathyal regions. Its overall depth range is given as 366-5530 m, although empty shells have been found as deep as 7600 m in the Romanche Trough in the Central Atlantic (Zezina 1980). The recovery of the specimens described here from shallow water sediments would indicate that they occupied quite a different habitat one to two million years ago than the living species. Another difference between the two is in their size; most known specimens of P. BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 323 K G H TEXT-FIG. 5. A-F, Pelagodiscus sp. from Hondeklip. a-b, a brachial valve exterior and interior, BD6670. x 3. c-D. a brachial valve viewed laterally and dorsally, BD6671. x 3. e, detail of the external ornamentation, near the brachial valve apex, scanning electron micrograph (uncoated), x 23. f, detail of the brachial valve interior, close to the margin, showing the smooth but slightly grooved surface. SEM (uncoated), x 37. G and h, Kraussina rubra (Pallas), from the Agulhas Bank, South Africa, at 22 fathoms. J. W. Jackson collection, ZB2240^3. G, clump of four specimens. The specimen at the top appears to have been attached to sponge, the central specimen is attached to the first and the remaining two are attached to the central one. x 1 . h, detail of the central specimen’s attachment to the first specimen; note the tight fit of the pedicle aperture onto the substrate, here the umbo of another specimen, x 2. i-k, Kraussina rubra (Pallas), from Koingnaas, BD6677, x2; posterodorsal view of the complete shell, the brachial valve interior, and the pedicle valve viewed posterodorsally. atlcmticus have a diameter in the range 3-5 mm, whereas these fossil specimens are larger with some exceeding 13 mm in diameter. Thus it seems that relatively large Cainozoic shallow-water species migrated to deeper water, and became smaller, during the last two million years. 324 PALAEONTOLOGY, VOLUME 33 Class ARTICULATA Huxley, 1869 Superfamily terebratellacea King, 1850 Family kraussinidae Dali, 1870 Genus kraussina Davidson, 1859 Type species. Anomia rubra Pallas 1776, designated by Davidson 1853, for Kraussia Davidson 1852, but renamed in 1859 (see text-fig. 5g, h). Diagnosis. Ventribiconvex, rounded to transverse, broad hinged, multicostate to (rarely) smooth shells. Cardinal process low, broad. Dorsal median septum supporting pair of stout ventrolaterally projecting arms of brachidium. Discussion. Until now, with few exceptions, Kraussina species have only been found from present seas around southern Africa. In that region there are four named species in addition to the type: K. cognata (Sowerby 1847), K. gardineri Dali 1910, K. mercatori Helmcke 1939, and K. crassicostata Jackson 1952. The only named fossil species, K. lata Haughton 1932, also came from South Africa, somewhat south of, but of about the same age as, the present fauna. Thus Kraussina, as presently known, is strongly endemic to seas around southern Africa. It is noteworthy that we propose five species of Kraussina for the Pliocene/ Pleistocene seas off south-west Africa {K. rubra, K. lata, K. rotundata sp. nov., K. laevicostata sp. nov., and K. cuneata sp. nov.), the same number as named in today’s waters. Although the time span of the older 50 m Unit, from which all five species were collected, may be about half a million years, the number of species apparently living in the same region is surprising. We suggest, however, that the late Tertiary was a period of evolutionary radiation for Kraussina, with only K. rubra continuing to the present. K. lata could well have evolved into K. crassicostata, while K. rotundata, with further loss of ribbing, may have become K. cognata. Kraussina rubra (Pallas, 1766) Text-fig. 5i-k 1766 Anomia rubra Pallas, p. 182, pi. 14, figs. 2-11. 1952 Kraussina rubra (Pallas); Jackson, p. 22, pi. 3, figs 1 and 2. 1986 Kraussina rubra (Pallas); Hiller, p. 129, fig. 16. Material and horizon. Five conjoined valves, two pedicle valves and three brachical valves from the 50 m Unit on Hondeklip, Avontuur A and Koingnaas plus six pedicle valves and one brachial valve, as well as fragmentary material, from the 30 m Unit on Hondeklip. Description. Biconvex shells with variable transversely oval outlines. The hinge line is nearly straight, almost nine-tenths as wide as the valve. The anterior commissure is rectimarginate to broadly and very gently sulcate. The beak is suberect and irregularly truncated by a large incomplete submesothyridid foramen. The palintropes are triangular, bounded by the foramen and beak ridges. Ornament consists of concentric growth lines and strong rounded ribs which may increase by branching or intercalation; ribbing density is of 2-6 ribs, most commonly 3, in a 5 mm sector at the 10 mm growth stage. The pedicle valve is gently convex in lateral profile but strongly convex, especially medianly, in anterior profile. Brachial valves are gently convex in lateral profile; in anterior profile they are flat or gently sulcate medianly, with gently convex flanks. Pedicle valve interiors have small teeth, without dental plates; a pedicle collar, where preserved, is very short and sessile. Other details are obscure. Brachial valve interiors have widely divergent socket ridges bounding narrow sockets. A small cardinal process is situated between the poterior ends of socket ridges in small specimens but is reduced by abrasion in large specimens. The notothyrial platform consists of a pair of suboval thickenings between the socket ridges and the posterior end of the median septum; rounded depressions on the platform mark the sites of attachment of pedicle muscles. A low median septum extends anteriorly from the BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 325 notothyrial platform and supports, at its distal end, a pair of ventrally divergent lamellae. The rest of the brachidium is unknown. Small submarginal tubercules are present in some specimens. Dimensions. Dimensions of conjoined valves are (in mm): SAM PQ AV 610 SAM PQ HB 217 Length Width Thickness 14-9 16-9 91 19-6 19-8 10-5 14-9 17-8 9-6 170 17-3 10-8 181 20-4 12-6 Discussion. A small proportion of the kraussinids recovered from the 50 m Unit can probably be assigned to Kraussina rubra (Pallas 1766), and it seems that most of those from the 30 m Unit can be similarly assigned. We describe our material since it differs slightly from the living specimens upon which all previous descriptions have been based. The most tangible difference is that the fossil shells tend to be slightly more coarsely ribbed. A small sample of K. rubra recovered live from off the southern Cape coast gave figures of 3-6 ribs per 5 mm sector at the 10 mm growth stage on one, thirteen, twenty-two and four shells. Comparable figures for the fossil material described here are 2-6 ribs on two, eight, seven, six and one shells. This difference is thought not to be significant. Fossil K. rubra has been found in shallow marine deposits of late Tertiary to early Quaternary age in other places around the South African coast. Haughton (1932) listed the species from limestone quarries at Hoedjies Bay, near Saldanha, about 300 km south of Hondeklip Bay. In the same area Kensley (1972) recorded it from the ‘E’ quarry at Langebaanweg and it is known from the Alexandria Formation, a late Pliocene - Pleistocene shoreline deposit cropping out in the coastal areas in the vicinity of Port Elizabeth. These occurrences suggest that the distribution of K. rubra in late Pliocene - early Pleistocene times is very similar to its present distribution, i.e. in intertidal and shallow subtidal settings from Transkei, in the east, to Liideritz on the Namibian coast, in the west. Fossil shells from these other areas show a similar ribbing density to the specimens described here. Kraussina lata Haughton, 1932 Plate 2, figs. 1-13 1932 Kraussina lata Haughton, p. 34. Diagnosis. Somewhat transverse, ventribiconvex Kraussina with strong coarse ribbing. Type. Haughton’s (1932) original material from the ‘basal grit’ at ‘The Point’, on the north side of the Olifants River, cannot be traced. We select, therefore, a brachial valve from Koingnaas, which appears to conform to Haughton’s brief description, as Neotype (PI. 2, figs. 1 and 2), BD6681. Material. In addition to the neotype, in the Carrington collection are six pedicle valves and fourteen brachial valves. From the Pether collection there are one extra pedicle valve and four brachial valves from Koingnaas, Description. The outline is very broadly obovate and the lateral profile strongly ventribiconvex, the brachial valve convexity largely resulting from a weak median sulcation. Pedicle valves are not folded, but the anterior commissure is weakly sulcate. The hinge line is just posterior to the widest part of the shell, but may form the widest part of the brachial valve. The pedicle aperture and interareas are characteristically wide and subject to abrasion. External ornamentation is of growth lines and strong, coarse costae, costellae being added rarely by branching. The total number of ribs varies from 8 to 16 on brachial valves, with 10 or II being the commonest number of costae. Pedicle valve interiors are scarce. Teeth appear to have been short and stubby. There is a sessile pedicle collar and the muscle scars are essentially as in K. rotundata, but the pedicle adjustor muscle scars tend to be shorter 326 PALAEONTOLOGY, VOLUME 33 and wider, and only narrowly separated medianly. On one specimen secondary shell growth has allowed the merger of these scars (PI. 2, fig. 6). Submarginal tubercules are preserved on some brachial and pedicle valves. Dimensions (in mm) Length Width Neotype b.v. 24-6 30-7 BD6681 b.v. 25-8 C.300 BD6682 b.v. 27-8 35-8 BD6683 p.v. 24-8 29-8 BD6684 p.v. 32-6 33-7 BD6685 p.v. 25-7 32-6 BD6686 Discussion. The species is the commonest representative in the collections originally presented by Carrington. Unfortunately details of locality are unknown, but some specimens are noted as from the mid or upper E Stage and marked ‘A 32’ (PI. 2, figs. 3-6). No complete shell is preserved and all show signs of erosion, in some specimens this is severe and has removed structures. Also in the Carrington collection are several incomplete valves of Kraussina we consider as a variety of K. Icitci. These are unusually wide shells with strongly thickened brachial valves (PI. 2, figs. 14-21) in which the pedicle adjustor muscle scars extend anterolaterally well beyond the anterior ends of the sockets. Unfortunately all these valves are badly eroded so rib counts cannot be made. However, the few remaining ribs indicate that they were more frequent than on K. later, we term these specimens K. cf. lata. As Haughton’s description (1932) was so brief we have provided a full description and comparison with other species. In general outline and profile K. lata resembles K. laevicostata, and both tend to be sulcate anteriorly. They differ markedly in the strong ribbing on K. lata, which manifests itself also on the insides of valves. This costation, with rare added costellae, covers the valves, although their prominence decreases towards the posterior margins. This is in contrast to K. rotundata in which the costae are non-existent on the flanks and posterolateral areas. K. rubra specimens are smaller than K. lata and have much finer ribbing (see Table 2). EXPLANATION OF PLATE 2 Figs. 1-13. Kraussina lata Haughton. 1 and 2, neotype, a brachial valve viewed externally and internally. BD668I. X 1. 3 and 4, external and internal views of a brachial valve with a series of growth distortions affecting the left posterior hinge line, the valve medianly and in its mid-length right side sector; the shell is bored posteromedianly leading to extra thickening on the right side of the median septum. Carrington collection, A32, mid to upper E stage. BD6682. 5 and 6, external and internal views of a large brachial valve with borings from the external surface producing ‘blistering’ of the internal surface; there is a small open canal from the pedicle adjustor muscle field to the posteroventral surface of the median septum (arrowed). Carrington collection, A32. BD6683. 7 and 8, external and internal views of a well preserved pedicle valve showing external ornamentation and the pedicle collar. Koingnaas. BD6684. 9 and 10, external and internal views of a somewhat abraded adult pedicle valve with borings, some of which broke through to the inner surface. The pedicle aperture is abraded ventrally and the muscle scars can be distinguished. Carrington collection, A32. BD6685. 11-13, external, internal and posterior views of a pedicle valve with a strongly abraded umbo and consequentially reduced pedicle collar. Carrington collection, A32. BD6686. All x 1. Figs. 14—21, Kraussina cf. lata Haughton, Carrington collection. 14 and 15, external and internal views of a large and badly eroded brachial valve displaying posterior abrasion and external pitting, the possible result of borings; the ‘ ME’ is for mid E stage. BD6691 . 1 6 and 1 7, external and internal views of a partially eroded brachial valve with the median part of its hinge line removed by abrasion ; note the anterolaterally extended pedicle adjustor scars. Upper E. BD6692. 18 and 19, external and internal views of a partially eroded brachial valve. Anteromedialy the ribbing is well preserved; the pedicle adjustor scars compare with those of fig. 15, but contrast with those of figs. 17 and 21. Mid E. BD6693. 20 and 21, external and internal views of a badly worn brachial valve, having lost all its ribbing; the anterior extension of the pedicle adjustor scars seems characteristic of the Upper E specimens. BD6694. All x 1. PLATE 2 BRLfNTON and HILLER, Namaqualand brachiopods 328 PALAEONTOLOGY, VOLUME 33 TABLE 2. Summary of ribbing densities on species from Hondeklip Bay plus K. rubra and K. crassicostata (Recent) from off the southern Cape coast. Number of ribs in 5 mm sector at 10 mm growth stage Sample Species Range Mode Size Age K. rubra 3-6 5 40 Recent K. crassicostata 3^ 3 5 Recent K. rubra 2-6 3 24 Pliocene-Pleistocene K. lata 2-3 3 8 Pliocene K. rotimdata 1-4 2 18 Pliocene K. cimeata 12-16 15 12 Pliocene K. laevicostata ‘Ribs’ develop only after 30 mm growth at 20 mm growth stage Cancellothyris 12-17 13 26 Pliocene platys Growth distortion is uncommon in K. lata, and although their pedicle apertures are abraded to some extent, it is never as severe as in K. rolundata. This indicates that K. lata lived attached to hard substrate but not in crowded conditions or in rock crevices. Until now K. lata was the only described kraussinid from the late Tertiary to early Quaternary deposits of South Africa. Amongst living Kraussina in South African waters K. crassicostata Jackson has a similar number of costae, but it does not grow to the dimensions of K. lata and tends to be as wide as long. It is, however, possible that a reduction in size and relative width in K. lata could have resulted in the living K. crassicostata. EXPLANATION OF PLATE 3 Figs. 1-16. Kraussina rotimdata sp. nov. 1-5, holotype from Avontuur A-T3. BD6705. x L5. 1 and 2, the complete shell viewed dorsally and ventrally showing the median ribbing. 3, the brachial valve interior with a virtually complete brachial support. 4, the shell viewed posterodorsally showing abrasion around the pedicle opening. 5, the pedicle valve showing the teeth. 6-9, a young shell from Avontuur. BD6706. x 1-5. The shell is viewed dorsally, posteriorly, anteriorly slightly agape to show the brachial supports, and laterally; the ventral umbo is virtually unabraded. 10 and 1 1, a mature shell viewed laterally and anteriorly showing the extent to which the valves can open, and the lophophore supports. Avontuur, BD6707. x L5. 12-14, a young shell displaying considerable growth distortion and abrasion (arrowed). Avontuur, BD6708. X 1-5. 12 and 13, brachial and pedicle valve exteriors. 14, brachial valve interior showing the distorted growth of the brachidium, in which the two arms grew almost in contact medianly. 15 and 16, an almost fully grown but distorted and abraded shell. Avontuur, BD6709. x 1-5. 15, the complete shell viewed dorsally; the ventral umbo had been abraded especially on the left side, as far as the tooth. 16, brachial valve interior showing the confined growth of the hinge line. 17 and 18, an adult, strongly abraded shell. Avontuur, BD6710. x L5. 17, oblique lateral view showing severe abrasion at both umbos and flank of the pedicle valve. 18, dorsal view showing median abrasion to the extent that the dorsal median septum shows through the remaining shell (arrowed). 19 and 20, a brachial valve viewed externally and internally from the Carrington collection. Upper E. BD6711. x 1. 21-24, an abraded (right side of the ventral umbo) and posteriorly bored (arrowed) shell viewed ventrally, dorsally and anteriorly. Avontuur. BD6712. 21-23, x 1-5. 24, detail of the anterior margin showing tubercules through the gape. x4. PLATE 3 BRUNTON and HILLER, Namaqualand brachiopods 330 PALAEONTOLOGY, VOLUME 33 TEXT-HG. 6. Kraiissina ronimlata sp. nov. from Avontuur - T3. a. c, a large shell viewed posteriorly (a) and laterally (c) showing posterior abrasion and lateral growth distortion, BD6713, x 1.5. n, enlarged posterior view of the pedicle aperture using bottom lighting to show the lophophore support within the shell, x 5. d-g, scanning electron micrographs (not coated, environmental chamber) of the interior of a brachial valve, BD67I4. D, tubercules at the posterolateral margin, x 19. i-., part of the cardinalia showing a socket, part of the radially ridged cardinal process and part of the pedicle adjustor muscle scar, x 12. F, tubercules at the anterior margin displaying resorption and regeneration patterns, x 42. G, detail from the centre of F showing the shell mosaic in an area of recent resorption and two current tubercules, x 130. BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 331 Kraussimi rolwulata sp. nov. Plate 3. figs. 1-24; Text-figs. 6-8 Diagnosis. Strongly biconvex Kraussina ornamented by relatively few, medianly developed, coarse costae. Type. Holotype BD6705. Paratypcs SAM PQ AV 609, AV 610; SAM PQ HB 217, HB 348, and eight figured shells, BD6706-BD6714 (excluding BD671 1), from Avontuur A. Name. Rotundas (L.) = round, spherical, referring to the strong biconvexity of the shell. Material and horizon. Thirty-three conjoined valves, nine pedicle valves and eight brachial valves all from the 50 m Unit on Hondeklip and Avontuur A. Three shells and a few incomplete valves in the Carrington collection. Description. Strongly biconvex shells with subcircular outlines; the length usually slightly greater than width, although shape can be variable. The hinge line is nearly straight, about four-fifths as wide as the valve. The anterior commissure is rectimarginate to broadly and gently sulcatc. The beak is suberect with a very large submesothyridid foramen; posterior margins of shells are often abraded away by movement against hard substrates. The ornament consists of concentric growth lines and a few coarse subangular costae developed medianly. Ribs extend from the umbo but are often abraded from posterior portions of the valves; ribbing density of one to four ribs in one, twelve, four and one specimens at 5 mm medianly at the 10 mm growth stage, although a few shells show minimal ribbing; lateral areas are devoid of ribbing. Shell substance is thick, compared to K. ruhra. The pedicle valve is strongly convex in anterior and lateral profiles. Brachial valves are convex in lateral profile, and also in anterior profile but with a slight median flattening or incipient sulcus development. The pedicle valve interior has small, robust and strongly cyrtomatodont teeth. The pedicle collar is short, commonly sessile, but rarely free anteriorly resulting from anterior growth as a consequence of posterior abrasion. The ventral pedicle adjustor muscle scars are prominent, somewhat quadrate areas positioned posterolaterally, below the thickened shell of the teeth supports; they are commonly ridged concentrically (text-fig. 7). Between these scars is a slightly concave ovate area, somewhat thickened anteriorly and laterally, which extends forwards to a position just in front of the anterior edges of the adjustor scars; this is the scar TEXT-FIG. 7. Drawing of the posterior internal region of a pedicle valve of Kraussina rotundata sp. nov. showing the dispositions of the muscle attachment areas. 332 PALAEONTOLOGY, VOLUME 33 of the relatively small adductor muscles which, unlike their dorsal ends, are a closely united pair. The diductor muscle scars are weakly impressed, but positioned between the adductor and adjustor scars. (This interpretation is based upon study of the muscles in Recent K. rubra, with very similar muscle scars to K. rotiimiata.) Within about 0-5 mm of the valve margins are small outwardly directed tubercules with a frequency of about sixteen per 5 mm of valve edge. Their preservation is variable, due largely to the periodic resorption and overgrowth of the tubercules during valve growth and thickening (text-fig. 6d, f, g). The brachial valve interior has short, thick widely divergent socket ridges bounding small narrow sockets. The cardinal process is transversely elliptical, situated between the posterior ends of the socket ridges but often is partially removed by abrasion of the posterior shell margin. A broadly triangular notothyrial platform serves for the attachment of pedicle adjustor muscles, marked by a pair of ovoid to quadrate scars; from the anterior margin the low median septum extends to little more than one-half of the valve length (text-fig. 8). The distal end of the septum supports a pair of ventrally diverging brachial lamellae; the ventral end of each lamella widens as it bends posteriorly then narrows to a prong curving ventromedianly (PI. 3, figs. 3 and 1 1). Elliptical to oblong adductor scars are impressed on the valve floor on either side of the median septum, between the notothyrial platform and the brachidium; small subcircular posterolateral elements of the scars are differentiated and show where the smaller posterior adductor muscles were attached. Submarginal tubercules are more strongly and commonly present than in pedicle valves. Diductor scar / (cardinal process) TEXT-FIG. 8. Drawing of a brachial valve interior of Kraussina rotimdata sp. nov. showing the morphology and main muscle attachment areas. Dimensions. Examples of the dimensions of conjoined valves are (in mm): Length Width Hinge w. Thickness Holotype 22-5 224 191 15-2 BD6705 Paratypes 16-3 17-5 13-6 9-2 BD6706 194 20-7 171 10-8 BD6712 22-5 21-9 161 154 BD6713 230 19-6 C.120 12-5 BD6709 240 23-0 170 17-0 BD6707 BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 333 Discussion. The shells deseribed here are immediately referable to Kraussina on the basis of their internal structures, but they differ from other species of the genus in several important respects. Their coarseness and distribution of ribbing immediately separates them from the living species K. ruhra, K. cognata, K. gardineri and K. crassicostata. K. mercatori from off the west African coast, is a form that does not develop any ribs and, although many of the present specimens are sparsely ribbed, the normal form of the shell is with ribs developed medianly and thus they can be separated from K. mercatori, which is also smaller and with a coarsely tubercular valve interior. Of the fossil species, only K. lata has a similar density of ribbing to K. rotimdata sp. nov. but it has ribs strongly developed over the entire valve surface. In addition, K. lata has a flatter brachial valve and is usually wider than long with maximum width at or near the hinge line and rather acute cardinal angles (Haughton 1932). Kraussina laevicostata sp. nov. Text-fig. 9a-h Diagnosis. Large, transverse, ventribiconvex Kraussina lacking persistent ribbing. Types. Holotype: BD6730, an incomplete shell (text-fig. 9a-c). Paratypes: two pedicle valves and an incomplete brachial valve from Koingnaas. Five incomplete brachial valves and one pedicle valve from the ‘E Stage’, Carrington collection. Name. Laevis (L.) = smooth or bald. Costa (L, 0 = rib, referring to the scant ribbing. Description. The outline is very broadly obovate, with a hinge line almost reaching the maximum width. The lateral profile is ventribiconvex, the brachial valve being almost flat to gently convex, with a sulcus starting about 10 mm from the umbo. There is a complimentary fold on the pedicle valve producing a gently sulcate anterior commissure. The pedicle aperture is wide, abraded ventrally, and bordered by variably developed interareas. External ornamentation is of prominent growth lines, especially anteriorly, and irregular, impersistent occasional median ribs, developed only after about 30 mm in length. Pedicle valve interiors have relatively small teeth, which in older specimens are supported by shell thickening below the interareas. There is a short pedicle collar which, in some specimens, is partially worn away. Muscle scars are typically kraussinid. In large shells they are slightly sunken into the valve floor; the laterally placed adjustor scars are more prominent. Brachial valve interiors are similar to those of K. rotimdata. but in adults the prominent adjustor scars are less widely separated and elevated medianly. Both valves have submarginal tubercules with a frequency of ten to sixteen per 5 mm. Dimensions (in mm) Length Width Hinge u'. Thickness Holotype, shell 37-4 42-5 40-8 2M BD6730 Paratypes b.v. 28-0 33-6 32-6 — BD6732 p.v. 30-5 30-4 26-6 — BD6733 p.v. 35-8 47-4 44-8 — BD6731 p.v. 40-6 38-9 c.37-6 — BD6734 Discussion. This large species is comparable in size to the largest, wide example of K. lata, but the two differ in that the latter is strongly ribbed. K. laevicostata would seem to have had the same high degree of shape variation as is seen in the more numerous K. rotimdata. The sulcate commissure developed early in life and the ventral fold may be accentuated by what resembles a pair of ribs (text-fig. 9d). However, these do not seem to be present on brachial valves. In present-day waters off South Africa there is a ventribiconvex species, virtually lacking ribs, K. cognata (Sowerby), which is smaller and much less transverse than K. laevicostata. It is possible that size and relative width reduction may have transformed K. laevicostata into K. cognata. 334 PALAEONTOLOGY, VOLUME 33 I J K M TEXT-FIG. 9. A-H, Kruussiiui laevicostata sp. nov., a-c, Holotype viewed dorsally, ventrally and posterodorsally, Koingnaas, BD6730. d, e, a pedicle valve viewed externally and obliquely internally showing the pedicle collar and pedicle adjustor muscle scar flanking the diductor and median adductor scars. Koingnass, BD6731. f, g, a smaller brachial valve exterior and interior with a distorted right ear. Carrington collection, E stage, BD6732. H, an incomplete pedicle valve interior. Carrington collection, E stage, BD6733. All x 1. i-m, Kraussina cimeata sp. nov., from Koingnaas. i. j, a brachial valve exterior and interior, BD6738, x L5. k-m, holotype, a complete shell with separated valves, BD3739, x 4-5. k, pedicle valve interior showing the teeth and pedicle collar, l, m, brachial valve exterior and interior, showing the tuberculate interior and small inner socket ridges. BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 335 Kraitssina cimeata sp. nov. Text-fig. 9i-m Diagnosis. Small, transverse, entirely finely ribbed Kniiissina with persistent dorsal sulcus. Types. Holotype BD6739 (text-fig. 9k-m) and ten other incomplete valves from the Carrington collection, E stage. Name. Cuneatus (L.) = wedge-shape, or inverted triangle, referring to the shell’s outline. Description. These small shells (5-10 mm long) are broadly rounded triangular in outline, some reaching twice as wide as long. The greatest width is just anterior to the hinge line. The lateral profile is ventribiconvex. The dorsal sulcus is shallow, but originates close to the umbo. The ventral fold is less clear, but the anterior commissure is sulcate. The pedicle aperture is large and the interareas short but broad, small triangular deltidial plates may be preserved. The external ornamentation is of periodical growth-halts plus fine, regular ribbing over the complete shell. Brachial valve costellae are added by intercalation. There are twelve to sixteen ribs per 5 mm width at 10 mm in three, one, two, five and one specimens. The pedicle valve interior has small cyrtomatodont teeth and a relatively long sessile pedicle collar. Muscle scars are ill-defined on the few available specimens, but appear to be as in K. rotimdata. Brachial valve interiors have small widely divergent sockets, almost confined to the thickened posterior margin of the valve. The cardinal process is very short, but wide and typically radially ridged. The notothyrial platform is short and wide, accommodating transversely elliptical pedicle adjustor muscle scars. The median septum is low posteriorly, reaching about one-half of the valve length, and branches ventrolaterally into the main V-shaped brachidium; the complete structure is unknown. Adductor muscle scars are positioned normally, but poorly differentiated. The interiors of both valves display, to a variable extent, the form of the external ribbing and radial lines of tubercles following the rib interspaces. Close to the valve margins are more prominent tubercules with a frequency of between 9 and 13 per 5 mm length of margin. Dimensions (in mm) Length Width Hinge u’. Holotype, shell 51 6-8 60 BD6739 b.v. 8-0 16-5 151 BD6738 b.v. 8-0 14-2 12-6 BD6740 p.v. 9-9 14-6 13-7 BD6741 p.v. 7-5 1 L8 10-8 BD6771 Discussion. These small shells in some respects resemble Megerlina, but cannot be assigned to that genus because they show no sign of the pair of wing-like lateral extensions from the V-shaped part of the brachidium characteristic of Megerlina. The transverse outline, although much smaller, is comparable to the wide variety of K. lata from the Carrington collection. However, although some Kraitssina specimens display an initial 4 mm of ribbing finer than that covering the adult valves, the ribbing covering the valves of K. cimeata is still finer than any seem on other Kraitssina species in these faunas. The persistent dorsal sulcus and internally developed tubercules also differentiate this species. Superfamily cancellothyridacea Thomson, 1927 Family cancellothyrididae Thomson, 1927 Subfamily cancellothyridinae Thomson, 1927 Genus cancellothyris Thomson, 1927 Cancellothyris platys sp. nov. Plate 4, figs. 1-13; Text-fig. 10a-g Diagnosis. Large, relatively broad, rectimarginate to uniplicate Cancellothyris with thickness just over one-half length of shell. 336 PALAEONTOLOGY, VOLUME 33 Types. Holotype, BD6742, from the 50 m Unit at Avontuur A. Paratypes, figured specimens from Avontuur A and Hondeklip. Name. Platys (Gr.) = broad, wide. Material. In addition to the type specimens, there arc fourteen shells from Avontuur A, plus two pedicle valves and two brachial valves; from Hondeklip there are two pedicle and three brachial valves; and in the Carrington collection there are two small shells, three pedicle and three brachial valves, mostly marked ‘B-1 ’, all of which are somewhat eroded. At the Koingnaas locality were recovered two shells, two pedicle and three brachial valves, all in an incomplete state of preservation, representing the large and wide sub-species, C. p. petalos nov. (See the discussion.) Description. The outline is approximately five-sixths as wide as long, the greatest width being at about two- thirds the total length and the anterior margin being widely rounded. The lateral profile is biconvex, with a prominent ventral umbo truncated by the large pedicle foramen, commonly slightly widened by abrasion. Deltidial plates are medianly joined and short. The lateral commissure rises slightly anteriorly, and the anterior margin is rectimarginate to gently uniplicate in specimens over about 30 mm long. Thus in brachial valves over about 28 mm long a slight marginal fold developed. External ornamentation is of growth lines and a well developed fine ribbing, with 12-17 ribs in a width of 5 mm at 20 mm from the dorsal umbo in one, ten, six, six, two and one specimens. Even the larger uniplicate specimens show only the slightest folding or sulcation on their valves. Internally the pedicle valve has strong cyrtomatodont teeth supported by shell thickening on the flanks of the umbonal cavity. There is a well developed pedicle collar, free anteriorly and capable of growth into a short tube (text-fig. IOf). Muscle scars are ill defined, but the diductors are large and spreading. In the brachial valve umbo there is a narrow, ridged cardinal process from which sockets widen anterolaterally, with strong inner socket ridges overhanging the sockets posteriorly. The floor of the functional anterior part of the socket is supported by thickening from the inner surface of the valve (PI. 4, fig. 9). From the anteromedian corners of the sockets crura extend anteromedianly, as if to meet at about two-thirds of the brachial valve length. They are, however, only a few mm long before branching to form the complete ring of the brachidium (PI. 4, figs. 4 and 5). The anterior transverse band is relatively wide, ventrally arched and convex anteriorly. Muscle scars are not clearly difl'erentiated. The median diductor scars are oblong, with rounded anterior margins, and flanked by widely spreading pedicle adjustor scars. Dimensions (in mm). (All but one being from complete shells) Length Width Thickness Holotype 33-5 29-7 18-9 BD6742 Paratypes 36-2 29-9 20-6 BD6745 34-3 29- 1 18-6 BD6743 37-9 28-3 19-4 BD6744 EXPLANATION OF PLATE 4 Figs. 1-14. Cancellothyris platys platys sp. et subsp. nov. 1-5. holotype, the complete shell viewed ventrally and dorsally. The separate pedicle valve internally and the brachial valve internally and oblique internally. Avontuur. BD6742. x 1. 6-10, a complete shell from Hondeklip. BD6743. 6-8, viewed ventrally, dorsally and laterally, x I. 9, the cardinalia and pedicle valve umbo (above), x 3. 10, the umbos of the shell externally showing the deltidial plates and slightly abraded pedicle aperture, x 6. II, lateral view of a shell with growth distortion (arrowed) at its lateral commissure. Avontuur. BD6744. xl. 12-14, a complete shell with, internally, its loop and some of the associated spiculation preserved and protruding from the sand which filled the shell. Avontuur. BD6745. 12 and 14, the open shell viewed anterolaterally and closed, from the other side, x 1. 13, a detail from fig. 12 showing the anterior portion of the brachial loop (arrowed) and, anterior to that, a large remnant of the spicular skeleton that helped support the lophophore beyond the loop. X 10. PLATE 4 BRUNTON and HILLER, Namaqualand brachiopods 338 PALAEONTOLOGY, VOLUME 33 TEXT-HG. 10. For legend see opposite. BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 339 37-9 304 — BD6746 36-8 31-6 21-5 BD675I 30-5 304 16-7 BD6752 324 27-5 15-3 BD6753 28-8 26-3 13-0 BD6754 15-8 1 3-9 10-7 BD6755 16-0 14-8 8-3 BD6756 Discussion. Living Cancellothyris appears to be restricted to Australian waters and taxonomically the nearest species now living in South African waters are Terehratulina ahyssicola (Adams and Reeve), T. meridionalis Jackson 1952 and Terehratulina species of Cooper (1973/?), Hiller (1986) and Jackson (1952). The main distinction between Cancellothyris and Terehratulina is the conjunct deltidial plates in the former, but Cooper (1973a) has commented that T. ahyssa [error for ahyssicola] from off South Africa may occasionally have united deltidial plates. Cooper’s observation raises a question of validity of these two genera, and also provides a clue that this cancellothyrid morphology of conjunct plates may be retained in the South African populations. C. platys differs from T. ahyssicola, as originally described, by being very much larger. However, as pointed out by Jackson (1952), the original Adams and Reeve (1850) specimen is probably young and those figured by Jackson reached 284 mm long. Apart from the non-conjunct deltidial plates, Jackson’s T. ahyssicola specimens have a prominent dorsal sulcus, producing a sulciplicate anterior commissure. The loops of Jackson’s examples and those of C. platys are very similar. C. platys is most abundant at the Avontuur A locality. Shape variation is not extreme, but a few specimens did grow in confined conditions, leading to growth distortions (text-fig. IOg). Most of these distortions affected relatively short lengths of the valve margins and in one shell (PI. 4, fig. 1 1 ) possibly led to its shift in position away from the confining object and consequential extra abrasion at the left side of the pedicle aperture. We believe it more likely that these distortions resulted from growth against hard substrate rather than against other specimens. An unusual feature in some pedicle valves is the extent of anterior growth of the pedicle collar (text-fig. 10c, F). This does not seem to be associated with particularly severe abrasion of the pedicle aperture, as might be expected. Within the coarse sands and fine shelly debris filling specimens of C. platys we have recovered some small fragments of the original spicular skeleton which supported the lophophore. The complexity and stability of these structures are to be expected in this species, in view of the spicules studied in living Terehratulina by such authors as Deslongchamps (1860), Blochmann (1912) and Schumann (1973). Spicules have been described previously from fossil brachiopods, for example by Steinich (1963) from the Cretaceous and by Rowell and Rundle (1967) in Eocene Terehratulina, but seldom with as good preservation or articulation as here. In one specimen the entire mesodermal spicular skeleton has slipped anterolaterally from its original position on the loop and lies, somewhat crushed, within the sediment; those parts from the loop area and one side arm of the lophophore can be seen (PI. 4, figs. 12 and 13). The features which strikingly differentiate these spicules from those previously figured from Terehratulina are their length and intricate intermeshing, each part of the skeleton being made up of many layers of interlocked spicules (text-fig. 1 1). TEXT-FIG. 10. A-G, Cancellothyris platys platys sp. et subsp. nov. from Avontuur, a, b, a pedicle valve interior showing conjoined deltidial plates and teeth ( x 1) and a detail of the umbo internally showing the pedicle collar ( X 2'5) BD6746. c-f, a pedicle valve, BD6747. c, viewed internally showing the well dift'erentiated muscle field, X 1. D, external view, x 1, e, part of the external view enlarged to show the eft'ect of surface borings, x 5. f, the internal posterior region showing the unusually lengthened pedicle collar, the teeth, and the wide adductor and diductor muscle scars, x 2. h-m, Cancellothyris platys petalos sp. et subsp. nov. from Koingnaas, x I . H, i, an incomplete brachial valve exterior and interior with the tooth from the pedicle valve remaining in position on one side, BD6749a, J, k, holotype of the subspecies, a brachial valve viewed externally and internally, BD6750. L, M, the pedicle valve, viewed externally and internally, belonging to the brachial valve figured h, i. Deltidial plates have mostly been broken away when the shell was slightly crushed within the sediment, BD6749b. 340 PALAEONTOLOGY. VOLUME 33 TEXT-FIG. 1 1 . Scanning electron micrographs of part of the spicular lophophore support recovered from an example of Cancellothyris platys sp. nov. from Avontuur. This area is thought to have been associated with the anterior transverse band of the loop, the long elements on the right being associated with the lophophore canals and the reticulated area on the left being part of the central support between the lophophore arms. BD6767. x 75. At the Koingnaas locality there are several incomplete specimens which are larger (around 50 mm long) and relatively wider (the length and width being approximately equal) than the normal C. platys. Growth lines indicate that this shape was more or less consistent throughout life and that this extra width is not simply a gerontic feature. The rib density is comparable to the other specimens. Since these specimens became broader than true C. platys at an early stage of life it is not surprising to find the relative width of the crural bases (text-fig. lOi) is greater than in C. platys. We treat these specimens as a subspecies, C. platys petalos (text-fig. 10h-m) Dimensions (in mm) Length Width Thickness Holotype b.v. 48-0 53-0 — BD6750 Paratypes p.v. c.53-9 C.5T5 — BD6764 shell 51-5 46-4 c.23-8 BD6765 shell 50-3 c.46-3 28-6 BD6766 Acknowledgments. We thank Dr A. J. Carrington for the early donation of specimens from the area and Mr J. Pether for providing the many more, well-documented specimens which form the basis of this paper, and for helpful discussion. We are grateful for constructive comments by Dr L. R. M. Cocks and Dr P. D. Taylor. We thank the Photographic Unit of the British Museum of Natural History for photographs used in this study and the Department of Mineralogy (BMNH) for Infra-red Spectrum and X-Ray analysis of shell material. REFERENCES ADAMS, A. and REEVE, L. A. 1850. Mollusca. 1-88. In zoology of the voyage of H.M.S. Samarang -, under the command of Captain Sir Edward Belcher, during the years 1843-1846. Reeve and Benham, London. ASGAARD. u. 1986, Paedomorphosis in brachiopods: response to stressful environments 743^6. In racheboeuf, p. R. and EMiG, c. c. (eds.). Les Brachiopodes fossiles et actuels. Actes du ler Congres international sur les Brachiopodes, Brest, 1986. Biostratigraphie du Paleozique. 4, 1-500. BEARD, J, M., SANGREE, J. B. and SMITH, L. A. 1982. Quaternary chronology, paleoclimate, depositional sequences, and eustatic cycles. Bulletin of the American Association of Petroleum Geologists. 66, 158-169. BLOCHMANN. F. 1912. Die Brachiopoden der schwedischen Siidpolar Expedition. Wissenschaftliche Ergehnisse der Schwedischen Siidpolarexpedition. 1901-1903. 6, np. 7, 1-12. BRUNTON AND HILLER: LATE CAINOZOIC BRACHIOPODS 341 BOSQUET, j. 1859. Monographic des brachiopodes du Systeme Maestrichtien. Verlumdelmgen uitgegeven door de Commissie helcist met het Vervaardigeu eeiier Geologische Beschriving eii Kaart van Nederland te Haarlem. 3, 1-50. CARRINGTON, A. J. and KENSLEY, B. F. 1969. Plcistoccnc molluscs from the Namaqualand coast. Annuls of the South African Museum, 52, 189-223. COLLINS, M. J. 1986. Taphonomic processes in a deep water ATo(//o/w.v-brachiopod assemblage from the west coast of Scotland. Unpublished Ph.D. thesis. University of Glasgow. CURRY, G. B., QUINN, R., MUYZER, G., ZOMERDIJK, T. and WESTBROEK, p. 1988. Scro-taxonomy of skeletal macromolecules in living terebratulid brachiopods. Historical Biology, 1, 207-224. COOPER, G. A. I973fl. Fossil and Recent Cancellothyridacea (Brachiopoda). Science Reports, Tohoku University (Sendai, Japan), (2) Geology, 6 (Hatai Memorial Volume), 371-390. 19736. New Brachiopoda from the Indian Ocean. Smithsonian Contributions to Paleobiology, 16, 1 43. DALL, w. H. 1870. A revision of the Terebratulidae and Lingulidae. American Journal ofConchology 6, 88-168. 1908. Report on the dredging operations off the west coast of Central America. The Mollusca and the Brachiopoda. Harvard University, Museum of Comparative Zoology, Bulletin, 43, 205-487. 1910. Report on the brachiopods obtained from the Indian Ocean by the Sealark Expedition. 1905. Transactions of the Linnean Society, London, Ser. 2, 13, 439-441. DAVIDSON, T. 1859. In SUESS, E. Ueber die Wohnsitze der Brachiopoden. Sitzungsberichte der Akademie der Wissenschaften, Wien, 27, 185-248. DESLONGCHAMPS, M. E. 1860. Sur la fonction des spicules calcaircs renfermes dans le menteau dc certaines brachiopodes. L'Institut, Paris, 28, 421 422. DINGLE, R. V., SIESSER, w. G. and NEWTON, A. R. 1983. Mesozoic and Tertiary geology of Southern Africa. Balkema, Rotterdam, VIII + 375 pp, ELLIOTT, G. F. 1949. The brachial development of Kraussina (Brachiopoda). Annals and Magazine of Natural //A/on-, (12), 2, 538-546. EMBLEY, R. w. and MORLEY, J. J. 1980. Quaternary sedimentation and palaeoenvironmental studies off Namibia (South West Africa). Marine Geology, 36, 183-204. GRAY, J. E. 1840. Synopsis of the contents of the British Museum. 42 ed., London, 370 pp. HAUGHTON, s. A. 1932. The Late Tertiary and Recent deposits of the west coast of South Africa. Transactions of the Geological Society of South A frica, 34, 19-57. HELMCiCE, J. G. 1939. Krausshut mercatori n. sp. und der Verbreitung der Gattung Kraussina, (‘Mercator', 2). Memoires du Musee Royal d'Histoire Naturelle de Belgicjue. Ser. 2, 15, 13 139. HENDEY, Q. B. 1981o. Palaeoecology of the Late Tertiary fossil occurrences in ‘E’ Quarry, Langebaanweg, South Africa, and a reinterpretation of their geological context. Annals of the South African Museum, 84, 1-104. 19816. Geological succession at Langebaanweg, Cape Province, and global events of the Late Tertiary. South A frican Journal of Science, 77, 33-38. HILLER, N. 1986. The South African Museum’s Meiring Naude cruises, part 16: Brachiopoda from the 1975-1979 cruises. Annals of the South A frican Museum, 97, 97-140. HUXLEY, T. H. 1869. An introduction to the classification cf animals. John Churchill and Sons, London, 147 pp. JACKSON, J. w. 1952. A revision of some South African Brachiopoda, with descriptions of new species. Annals of the South A frican Museum. 41, 1-40. KENSLEY, B. 1972. Pliocene marine invertebrates from Langebaanweg, Cape Province. Annals of the South African Museum, 60, 173 190. 1977. A second assemblage of Pliocene invertebrate fossils from Langebaanweg, Cape. Annals of the South African Museum 72, 189-210. and PETHER, J. 1986. Late Tertiary and Early Quaternary fossil Mollusca of the Hondeklip area, Cape Province, South Africa. Annals of the South African Museum, 97, 141-225. KiLBURN, R. and TANKARD, A. J. 1975. Plcistocene molluscs from the west and south coasts of the Cape Province, South Africa. Annals of the South A frican Museum, Cl,, 183-226. KING, w. 1850. A monograph of Permian Lossils of England. Palaeontographical Society, 3, 1-258. 1868. On some palliobranchiate shells from the Irish Atlantic. Proceedings of the Dublin Natural History Society, 5, 170. LAMARCK, J. p. B. A. DE M. 1819. Histoire naturelle des animaux sans vertehres. Brachiopoda, 6, 1-343. PALLAS, p. s. 1766. Miscellanea Zoologica. Petrum van Cleef, The Hague, XII + 224 pp., 14 pis. PENRITH, M.-L. and KENSLEY, B. F. 1970. Tlic Constitution of the intertidal fauna of rocky shores of South West Africa. Part 1. Liideritzbucht. Cimbebasia, Series A, 1, 191-239. 342 PALAEONTOLOGY, VOLUME 33 PETHER, J. 1986rt. Late Tertiary and Early Quaternary marine deposits of the Namaqualand coast. Cape Province: new perspectives. South African Journal of Science, 82, 464-470. 1986/). Fossil molluscs from Hondeklipbaai. Fourth conference of the Palaeontological Society of South Africa, Abstracts. 16. RUGGIERO, E. T. 1985. Paleoecologia e biostratigrafia delle calareniti a brachiopodi di Castra (Lecce). Bollettino della Societd di Naturalisti in Napoli, 92, 347-413. ROWELL, A. J. and rundle, a. j. 1967. Lophophore of the Eocene brachiopod Terehratulina wardenensis Elliott. Paleontological Contributions of the University of Kansas, 15, 1-8. SCHUCHERT, G. and LEVENE, c. M. 1929. FossiUum Catalogus. 1 : Animalia. Pt. 42: Brachiopoda. Junk, Berlin, 140 pp. SCHUMANN, D. 1973. Mesodermal Endoskelette terebratulider Brachiopoden. I. Palaeontologische Zeitschrift, 47, 77-103. SIESSER, w. G. 1978. Aridification of the Namib Desert: evidence from oceanic cores. 105-1 13 hi van zinderen BARKER, E. M. (ed.). Antarctic glacial history and world palaeoenvironnients. Balkema, Rotterdam, 172 pp. SOUTH AFRICAN COMMITTEE FOR STRATIGRAPHY. 1980. Stratigraphy of South Africa. Part I (Compiled l. e. RENT): Lithostratigraphy of the Republic of South Africa, South West Africa/Namibia and the Republics of Bophuthatswana, Transkei and Venda. Handbook of the Geological Survey of South Africa, 8, i-xix+ 1-690. SOWERBY, G. B. 1847. Tliesaurus Conchyliorum. 1, G. B. Sowerby, London, 337-369. STEiNiCH, G. 1963. Drei neue Brachiopodengattungen der Subfamilie Cancellothyrinae Thomson. Geologic, 12, 732-740. TANRARD, A. J. 1975. The late Cenozoic history and palaeoenvironnients of the coastal margin of the south- western Cape Province, South Africa. Unpublished Ph.D. thesis, Rhodes University, Grahamstown. THOMSON, J. A. 1927. Brachiopod morphology and genera (Recent and Tertiary). New Zealand Board of Science and Art, Manned, 7, 1-338. VAIL, p. R. and HARDENBOL, J. 1979. Sea-level changes during the Tertiary. Oceanus, 22, 71-79. WAGNER, p. A. and MERENSRY, H. 1928. The diamond deposits on the coast of Little Namaqualand. Transactions of the Geological Society of South A frica, 31, 1^1. WALRER, G. and JACOB, E. 1789. In Essays on the microscope .... 2nd Edition with considerable additions and improvements by F. Kanmacher. Dillon and Keating, London, 712 pp. WOOD, s. V. 1874. Supplement to the Monograph of the Crag Mollusca with descriptions of shells from the Upper Tertiaries of the East of England. Palaeontographical Society, 27, 99-231. ZEZiNA, o. N. 1980. The composition, distribution and some biological characteristics of the cold water brachiopods of the southern hemisphere. 9-35. In Ecological Investigations of the Shelf. Akademy Nauk SSSR, Institut Okeanologii im P. P. Shirshova, Moscow, 1-98. C. H. C. BRUNTON Department of Palaeontology British Museum of Natural History London, SW7 5BD N. HILLER Manuscript received 23 March 1989 Revised manuscript received 17 July 1989 Department of Geology Rhodes University Grahamstown 6140, South Africa EARLY MISSISSIPPIAN HYOLITHA FROM NORTHERN IOWA by JOHN M. MALINKY CUld SHIRLEY SIXT Abstract. The exceptionally fine preservation and large number of specimens from a single late Palaeozoic locality makes the hyoliths in the Humboldt Oolite (Osagean, lower Mississippian) unique. All specimens are assigned to Gerkella humholdti n. gen., n. sp., family Hyolithidae, order Hyolithida. There is considerable variation in certain morphological features, such as transverse shape, nature of ornament and apical curvature; however, these differences are judged to be gradational. This indicates that certain features may be variable within one species and shows that establishment of hyolith species should be based upon a variety of features. Other North American late Palaeozoic hyoliths include Hyolithes carhonaria Walcott, H. milleri Sinclair, H. parviilus Girty and H. waverliensis Hyde. Their types lack important morphological features, which makes their generic identifications uncertain. Their names should not be used for any further material until better preserved topotypes become available for study. Early Palaeozoic hyoliths are currently under intensive investigation in the Soviet Union, China and North America. However, late Palaeozoic hyohths have received much less attention largely because of rare occurrence. Only six late Palaeozoic species, represented by fewer than ten specimens, have been reported in North America since the mid-nineteenth century (Sinclair 1946). The number of late Palaeozoic specimens drastically increased with the discovery of approximately forty specimens from the Missourian (Pennsylvanian) Eudora Shale of southeastern Kansas (Malinky and Mapes 1983) and more than 1000 specimens from the Pennsylvanian of Kansas, Oklahoma and Texas (Malinky et al. 1986). Unfortunately, only a small number of those specimens can be identified to genus; most are poorly preserved steinkerns and cannot be assigned to genus or species. Therefore the discovery of seventy-six well preserved hyoliths from the lower Mississippian of northern Iowa is unique among late Palaeozoic occurrences because of the large number of exceptionally well preserved specimens from one late Palaeozoic locality. Morphology of these specimens indicates that all represent the same species; they are assigned to Gerkella humholdti n. gen., n. sp., in the family Hyolithidae and order Hyolithida (text-fig. 1a-p). Finding these specimens permits a survey of the range of morphological variation within a hyolith species. This further allows assessment of the taxonomic significance of features such as transverse shape and curvature of the apical end of the shell. Transverse shape in particular was used as the ruling criterion for establishing hyolith species in the nineteenth and early twentieth century (Barrande 1867; Novak 1891; Holm 1893; Walcott 1886, 1889, 1890; Resser 1938). It formerly served as the sole distinguishing feature between Hyolithes Eichwald, 1840, and Orthotheca Novak, 1 886, the only two generic names in use for hyoliths until recently (Syssoiev 1958, and many other Soviet and Chinese publications). These genera now form the basis for the orders Hyolithida Matthew /7r/c Fisher, 1962, and Orthothecida Marek, 1966, within the class Hyolitha Marek, 1963. Recent studies by Marek (1963, 1967) suggested that while transverse shape remains a characteristic to be considered for diagnosis and recognition of hyolith taxa, other features of the conch and operculum must also be used when available. The gradational nature of transverse shape and other morphological features, such as prominence of growth iirae on the shell, indicates that all specimens from the Humboldt Oolite belong to one species. In addition to G. humholdti, other North American late Palaeozoic hyoliths are the Mississippian species Hyolithes aculeatus (Hall, 1860) from Indiana, H. milleri Sinclair, 1946 from Missouri, H. I Palaeontology, Vol. 33, Part 2, 1990, pp. 343-357. | © The Palaeontological Association TEXT-FIG. 1. Gerkella humboldti n. gen., n. sp., Osagean, northern Iowa, USA. a-d paratype USNM 390573, dorsal, left lateral, right lateral and ventral views respectively, x 6-4. e, paratype USNM 390521, dorsum (note apical curvature toward right), x 7-4. f, paratype USNM 390554, dorsum (note apical curvature toward left), x6-3. G, paratype USNM 390543, dorsum (note apical curvature toward left), x7-2. h, paratype 390532, dorsum (note curvature toward right), x 8-3. i, J, n, holotype USNM 390504, left lateral, dorsal and ventral views respectively, x4-5. k, l, paratype USNM 390531, left lateral and ventral views respectively, x 5-7. m, paratype USNM 390552, dorsum, x 7-5. o, p, paratype USNM 390545, dorsal and ventral views respectively, x5-8. PALAEONTOLOGY. VOLUME 33 MALINKY AND SIXT: EARLY M ISSISSIPPI AN HYOLITHA 345 TEXT-FIG. 2. A-D. Hyolitlu's waverliensis Hyde, holotype OSU 19895, ventral, right lateral, left lateral and dorsal views respectively, x 5-3. E, J, l, m, H. carhomiria Walcott, USNM 14426. E, shell at anterior edge of ligula, x 5. j, internal mould of dorsum, x 8. K, external mould of venter, x 10. l, internal mould of venter with some shell near apical end, x 10. F-i, H. parvidus Girty. F, G, lectotype USNM 121 196, dorsum and venter respectively, X 7-2. H, 1, paralectotype USNM 122195, dorsum and venter respectively, x 7-5. k, H. milleri Sinclair, lectotype UCGM 3900A, venter, x 2. parvidus Girty, 1926 from Texas and H. waverliensis Hyde, 1953 from Ohio, and the Pennsylvanian species H. carhonaria Walcott, 1884 from Nevada (text-fig. 2a-m). The types of H. carhonaria, H. milleri, H. parvidus and H. waverliensis are incomplete, hence their generic identifications may be doubted. The names of these species should not be used for any further material until better- 346 PALAEONTOLOGY, VOLUME 33 preserved topotypes become available for study. The type or types of H. aciileatus (Hall, 1860) cannot be located at present; however, the published description of that species indicates that few morphological features are preserved, so that its generic identification is uncertain. Three other Carboniferous species, H. roenieri Koenen, 1879, H. sicala Koninck, 1883 and H. sturi Klebelsberg, 1912 were reported from western Europe. Although these specimens have not been examined, published descriptions and illustrations of the types indicate that preservation is poor, again rendering generic identifications doubtful. Preservation of more recently-discovered specimens considered to be hypotypes of those species (Zakowa 1971) from the Goniatites granosus zone of eastern Europe is not sufficiently good confidently to identify those specimens. PALAEOECOLOGY AND MORPHOLOGICAL VARIATION Stratigraphic setting and palaeoenvironment. G. hiiniboldti was discovered in the Osagean (Early Mississippian) Humboldt Oolite near Humbolt in northern Iowa (text-fig. 3). The palaeo- environment of this unit was studied in detail by Gerk and Levorson (1982) and Glenister and Sixt (1982); detailed stratigraphic sections of that locality showing distribution of fauna were given by Glenister and Sixt (1982) and by Brenkle and Groves (1985). Previous studies of fauna were summarized by Glenister and Sixt (1982). More recent taxonomic studies of the trilobites and foraminiferans are those of Brenkle and Groves (1985) and Brezenski (1988) respectively. The hyoliths described herein were only previously mentioned once in an unpublished study of the gastropods of this unit (Harper 1977). TEXT-FIG. 3. Locality map of Gerkella humholdti n. gen., n. sp. The hyoliths occur within the bioclastic grainstone facies of the Humboldt Oolite (text-fig. 4). That facies includes a series of poorly-indurated, friable, ooid grainstone lenses, 50-200 mm thick which are most common near the middle of the section. The lenses also contain abundant, small, apparently size-sorted gastropods, brachiopods, bivalves, rostroconchs, ostracods and calcareous algae, all of which are exquisitely preserved. According to Gerk and Levorson ( 1 982), the Humboldt Oolite is thought to represent a shallow to marginal marine deposit which originated in an arid climate. More normal marine deposits occur at the base of the section with environmental and facies MALINKY AND SIXT: EARLY M I SSISSI PPl AN HYOLITHA 347 TEXT-FIG. 4. Reconstruction of facies during deposition of Humboldt Oolite. Hyolith shells (c) accumulated in lenses on bars or shoals (bioclastic grainstone) which sheltered a lagoon (peloidal packstone/grainstone). restriction progressively increasing toward the top. The bioclastic grainstone represents a series of bars or shoals that sheltered a lagoon in which the sparcely fossiliferoiis peloidal packstone facies originated. Gerk and Levorson (1982) suggested an analogy between the hyolith-bearing lenses and recent shell beds in protected back-beach areas of the Bahamas. These modern shell deposits resemble the Humboldt lenses in size, lithofacies association and faunal diversity. The Bahamian deposits are thought to have originated from a coincidence of high tide and a south-west wind, rather than from fair weather wave agitation or reworking associated with storms. Movement of water at high tide transports shells to back-beach areas where they are protected from further reworking. Under these circumstances chances of preservation are greater. Conditions resulting in such deposits are unusual and occur only several times a year. A similar origin was suggested by Gerk and Levorson (1982) for the lenses in the Humboldt Oolite. Hyolith palaeoecology and taphonomy. Various aspects of hyolith palaeoecology were summarized by Fisher (1962) and by Marek and Yochelson (1964, 1976) and will not be reviewed in detail here. However, the unusual lithologic and environmental setting of G. humboldti requires additional comment. Because the hyoliths are relatively small they cannot be observed directly on outcrop and their distribution and orientation within the matrix are unknown. None the less, some generalizations on the palaeoecology and taphonomy of this species are still possible. Late Palaeozoic hyoliths in North America occur in a wide variety of facies from shallow normal marine to offshore, oxygen-poor facies, however they only seem to be abundant in facies which originated in stress environments. Hyoliths have been discovered in normal marine facies such as sandstone (H. waverliensis Hyde, 1953 from Ohio), shale (Malinky, unpublished data), and limestone (all remaining Mississippian occurrences and the Pennsylvanian H. carhonaria Walcott, 1884). This indicates that they were widespread in the marine environment despite their relatively low abundance at any single locality. However, the taxa from normal marine environments are represented by only one or two individuals each. In contrast, in late Palaeozoic facies from stress environments assemblages of hundreds of hyoliths are known. More than 1000 specimens from fourteen different shale units have been discovered in offshore marine though oxygen-poor shales (Heckel 1977) from the Pennsylvanian of the southern Midcontinent (Malinky et al. 1986). Kammer et al. (1986) suggested that this type of environment served as a refuge for hyoliths and other ‘archaic’ taxa such as monoplacophorans which were abundant in the lower Palaeozoic but uncommon in middle and upper Palaeozoic strata. In the oxygen-stressed Pennsylvanian marine environment dysaerobic conditions would have excluded all benthos except those forms such as the hyoliths which seem to have been specially adapted to it. The relative scarcity of other organisms owing to oxygen stress would have decreased competition with the mechanically inefficient hyoliths (Yochelson 1984). 348 PALAEONTOLOGY. VOLUME 33 The environment represented by the Humboldt Oolite includes a number of microenvironments, and therefore it is heterogenous or ‘coarse-grained' (Ricklefs 1979) for slightly mobile, benthic organisms with regard to water depth, amount of agitation and suspended sediment in the water, and salinity fluctuation. Because the bioclastic grainstone facies was deposited nearshore, environmental stress during deposition would have been caused by these factors rather than by low oxygen. Most stenotopic organisms would have been excluded from an environment such as this except for those specially adapted to it. Even though G. humholdti may not have lived in the lenses where it was discovered, the fine preservation of specimens indicates that the amount of transport was minimal and that this form was without question a shallow marine, nearshore species. The relative abundance of this species in the Humboldt Oolite (compared to other Mississippian units) suggests that the species was well adapted and even opportunistic in this particular environment. Yet compared to other invertebrates normally seen in late Palaeozoic marine strata, hyoliths constitute only a small portion of the fauna in the Humboldt Oolite. All ontogenetic stages seem to be represented among the Humboldt hyoliths although the larger, and presumably adult, individuals seem to predominate. This may be partly a function of winnowing in which smaller individuals were removed, or it may reflect normal mortality within a breeding population. Individuals dying in old age will normally disarticulate, and the operculum and helens may be lost. Hyoliths buried alive by a sudden influx of sediment might be articulated (Yochelson, pers. comm. 1984). Most of the hyoliths in the Humboldt Oolite may have died from old age, although if any were rapidly buried while alive, winnowing has caused disarticulation of hard parts for those individuals as well. Following burial, the hyolith conchs in the Humboldt Oolite filled with sediment and small skeletal debris. The sediment and the shells neomorphosed into blocky calcite spar causing all traces of original shell structure to be destroyed. The boundary between the inner shell wall and the spar is distinct, but it is impossible to separate the two to search for muscle scars or other features on the interior of the shell. Calcite spar readily separates from the outer shell wall, suggesting some microenvironmental diflferences in diagenetic conditions between the interior and exterior of the shell. Lack of oncolitic coatings on the hyoliths suggests either sufficient turbidity to block out sunlight or rapid burial of shells following winnowing. Morphological variation. Morphological criteria by which species may be recognized among hyoliths were listed by Syssoiev (1958) and Marek (1967). They specifically mentioned: length of shell (L), shell thiekness, width (W) and height (H) of aperture, details of growth lirae and other ornament, apical curvature, and various angular measurements of the aperture and apex. Marek (1967) also listed many features of the operculum, but these are not considered here because the operculum of G. humholdti is unknown at present. Based upon the number of specimens used by these authors in naming new taxa, neither Marek nor Syssoiev had access to as many individuals as are used in this study. Presumably for these authors, morphologieal gaps existed between specimens from different horizons or different localities, thereby suggesting taxonomic status. The hypothesis proposed herein is that all hyoliths from the Humboldt Oolite represent the same speeies. They occur at the same stratigraphic position and in the same facies, but by themselves these criteria are not conclusive. An examination of selected morphological features must be undertaken either to separate specimens into discrete species, or to survey the range of variation among selected features within the same species. The criteria selected to test the species hypothesis are length of shell, apical curvature, and width and height of the aperture ( = transverse shape). These features were chosen specifically because they allow accurate measurement. Other features, such as length of the ventral ligula and nature of growth lirae eannot be used with the same degree of certainty as the features listed above. On many of the Humboldt specimens the ligula is incomplete and variation among growth lirae may reflect partial dissolution of the outer shell wall. Apical and lateral angles are not used because they are directly proportional to the width and height of the aperture, and their distribution will follow that of the apertural parameters. Bivariate plots of the apertural characteristics and a visual comparison of apical curvature MALINKY AND SIXT: EARLY M ISSISSI PPI AN HYOLITHA 349 A B TEXT-FIG. 5. A, apertural width (W) plotted against apertural height (H). b, H plotted against length of shell (L). c, distribution of W/H among Humboldt hyoliths. d, trend in W versus H. e, H plotted against L. f, L plotted against W/H. Absence of clusters among points demonstrates that, based upon these selected characteristics of the shell, all specimens studied herein belong to the same species. strongly support the notion that all Humboldt hyoliths belong to the same species (text-hg. 5). These results suggest that caution should be used when erecting hyolith taxa based upon relatively few characteristics. Even at present, transverse shape remains the major criterion for some workers when naming and recognizing hyolith species (Syssoiev 1962, 1968; Landing 1988; and many others). Had only the specimens from each end of the range of variation been discovered, the case 350 PALAEONTOLOGY, VOLUME 33 for more than one species based upon transverse shape would have been strong because of the large morphological gap between specimens (text-fig. 6). Conversely, dififerent species established on features such as growth lirae or apertural morphology including lateral sinuses or an apertural flare may have similar transverse shapes. In this case, the large population of hyoliths demonstrates that the selected characteristics are gradational among specimens, although the amount of variation seen here may not be typical of all hyolith populations. The adaptive significance in this environment, if any, for the characteristics mentioned above is unknown. ABC TEXT-FIG. 6. Transverse sections of Gerkella hiimholdti n. gen., n. sp. a, holotype USNM 390504. b, paratype USNM 390521. c, paratype USNM 390545. All x 10. SYSTEMATIC PALAEONTOLOGY Class HYOLiTHA Marek, 1963 Order hyolithida Matthew Fisher, 1962 Family hyolithidae Nicholson Fisher, 1962 Genus gerkella gen. nov. Etymology. The genus is named in honour of Arthur J. Gerk, who discovered these specimens. Type species. Gerkella humboldti n. gen., n. sp. Diagnosis. Hyolithid which has rugae on the exterior of the shell. Apertural rim is orthogonal without a flare, and the ligula is short. Included species. G. humboldti n. sp. and possibly //? centetmialis (Barrett, 1876) from the Devonian of New York. Remarks. This genus is distinguished from others included under the Family Hyolithidae (see Malinky 1988 for list of genera) by the presence of rugae on the shell and other details of ornament and apertural characteristics. Hyolithes Eichwald, 1840, Carinolithes Syssoiev, 1958, Sololites Marek, 1967, Maxilites Marek, 1972, Cavernolites Marek, 1974, Nervolltes Marek, 1974, and Dilytes Marek, 1974 possess longitudinal sculpture on the shell which Gerkella lacks. In addition, Eimwrpholites Marek, 1967, Lirotheca Malinky and Mapes, 1983, and Darwinites Malinky, Mapes and Broadhead, 1986 possess a distinct apertural flare which Gerkella lacks. Elegantilites Marek, 1967, and Joachimilites Marek, 1967 have a small apical angle and fine, closely spaced, transverse ribs which anastomoze in some places; Gomplwlites Marek, 1966, and Buchavilites Marek, 1975 have a tubular shell with a rounded transverse section. Nevadotlieca Malinky, 1988 has nearly angular lateral margins and inflated slopes on the dorsum which Gerkella lacks. Hyolithesl centennialis Barrett, 1876, from the Devonian of New York may represent this genus; the types of that species are covered with prominent rugae, making it the only recorded species from the middle Palaeozoic of North America to have rugae. The generic identification of this species is uncertain (Malinky, Linsley and Yochelson 1987) because the venter is unknown. Until complete specimens become available for study, that species is tentatively retained under Hyolithes. Stratigraphic range. ?Middle Devonian; Osagean, Lower Mississippian. MALINKY AND SIXT: EARLY M I SS I SS I PPI AN FIYOLITHA 351 Gerkella humholdti n. sp. Text-fig. Ia-p Etymology. The species is named after the locality where it was discovered. Diagnosis. Gerkella which has a shallow sinus along the apertural rim of the dorsum. Description. The shell of this species has a broad, nearly flat to slightly inflated venter, which grades into narrowly rounded lateral margins. The dorsum is inflated with a narrowly rounded longitudinal axis and the adjacent slopes vary from nearly flat to slightly inflated. The ligula along the ventral apertural margin is short, and the anterior edge is straight. The dorsal apertural rim lacks a flare, but a shallow sinus occurs in the rim along each lateral margin and in the middle of the dorsum. The apertural rim is orthogonal (perpendicular to the venter). The apical angle of the shell is small, and the apical end curves either to the left or right when viewed dorsally, and on some specimens it also curves toward the venter. The transverse section of the shell is subtriangular (text-fig. 6). The exterior of the shell is covered with widely spaced rugae. On the venter, the rugae curve to follow the outline of the anterior edge of the ligula. The rugae curve on the lateral margins to form a shallow sinus, parallel to that in the apertural rim, and on the dorsum the rugae are nearly transverse except for a shallow median sinus which follows that of the apertural rim. The operculum and helens arc unknown. Remarks. This species is currently known from seventy-six specimens; specimen USNM 390504 is selected as holotype because it is the most complete and best preserved. That specimen is 8-8 mm long, and has an apertural width and height of 2-6 mm and 2-3 mm respectively. Neither the holotype nor any paratype is operculate, and no disarticulated opercula or helens have been found for this species. All specimens are free of matrix, although the interiors of the shells contain blocky calcite spar which cannot be removed without destroying the specimens. Details of the interior are unknown. The specimens upon which this species is based were collected over a twenty year period by A. J. Gerk of Mason City, Iowa. These hyoliths were mentioned in a study of the Gastropoda of the Humboldt Oolite by Harper (1977), who regarded them as conspecific with Hyolithes'l waverliensis Hyde, 1953 from the Mississippian of Ohio. That species is based upon a steinkern which lacks most taxonomically important characteristics such as all features of the apertural end. No meaningful comparison between that species and G. humholdti is possible, and the two are herein treated as separate species. Material. Holotype USNM 390504 and seventy-five paratypes under 390505 through 390580. Occurrence. P. & M. Hodges quarry, sec. 32, T92N, R28W, northeast of Humboldt, Humboldt County, north- central Iowa, from the Humboldt Oolite, Osagean (lower Mississippian) (text-fig. 3). Stratigraphic range. Osagean. Class HYOLITHA iticerlae sedis Hyolithes'l aculeatiis (Hall, 1860) I860 Pugiimculusl (Theca) aculeatus Hall, p. 107. 1862 Pugiuncidusl aculeatus Hall; Winchell, p. 423. 1865 Piigiunculusl aculeatus Hall; Winchell, p. 131. 1898 Hyolithes aculeatus (Hall); Weller, p. 311. 1946 Hyolithes aculeatus (Hall); Sinclair, p. 73. 1967 Hyolithes aculeatus (Hall); Yochelson and Saunders, p, 9. Description. 'Elongate, obtusely triangular bodies, having one side nearly flat, and the other two sides meeting at a very obtuse angle, and slightly incurved towards the angle, the flat side being eonvex in the direction of the length. Aperture obtusely triangular, and a little thickened on the straight side of the lateral angles’ (Hall I860, p. 107). 352 PALAEONTOLOGY, VOLUME 33 Remarks. This species was named and described but not illustrated by Hall (1860, p. 107). To date, this species apparently has never been illustrated, yet the name has been used for other specimens by later workers (see below). Hall (1860) did not select a holotype for this species, and the type or types cannot be located at present. Hall’s description, and indeed, the fact that he was uncertain of the generic identification, leaves little doubt that preservation of the type material was poor. Because so few characteristics of this species are known, its name should not be used for any additional specimens. Winchell (1862, 1865) referred hyoliths which he described as casts ‘without external markings’ to P.? aculeatus, but he provided no additional details of those specimens or their occurrences. Weller (1898), Sinclair (1946) and Yochelson and Saunders (1967) listed the species name in their respective compilations of hyolith species without reporting any specific occurrences other than that of the type. Weller (1898) listed the species under Hyolithes because by that time, both Pugiuuculus Barrande (1847), and Theca (Sowerby Morris 1845) were regarded as junior synonyms of Hyolithes. Occitnence. Hall (I860, p. 107) discovered this species in the ‘Goniatite limestone near Rockford, Indiana.’ That unit is now called the Rockford Limestone (lower Mississippian, Osagean; Gray 1979). Winchell’s (1865) specimens were discovered in the Marshall Formation (Osagean; Ellis 1979) in the SE 1/2 SW 1/2 sec. 23, Adam, Hillsdale County, Michigan and in ‘Alan’s and Germain’s quarries, Hillsdale,’ Michigan (Winchell 1862). Hyolithesl carhonaria Walcott, 1884 Text-fig. 2e, j, l, m v* Hyolithes carbotmria Walcott, p. 264, pi. 23, fig. 3. V* 1892 Hyolithes carhonaria Walcott, p. 333. v* \9A6 Hyolithes carhonaria Walcott; Sinclair, p. 74. V* 1967 Hyolithes carhonaria Walcott; Yochelson and Saunders, p. 9. Description. The venter of this species is flat, but curves slightly to grade into narrowly rounded lateral margins. The dorsum is low, and the longitudinal axis is narrowly rounded. The slopes adjacent to the axis are nearly flat. The ligula along the ventral apertural margin seems to be short, and the anterior edge appears to have been flat. The ligula appears to curve slightly toward the venter. The shell on the venter is covered with fine, closely-spaced lirae which follow the outline of the anterior edge of the ligula. Two longitudinal sulci are located at each edge of the venter and they extend for the entire length of the venter. The sulci are prominent near the apertural end, but become shallower toward the apical end. The dorsal internal mould is smooth except for a prominent indentation located near the apertural end on each slope. The operculum, the shell on the dorsum and the complete aperture are unknown. Remarks. The holotype and only known specimen of this species is 9-5 mm long, and has an apertural width of 2-3 mm. The holotype exists as several counterparts; an external mould furnishes details of the venter, a portion of the internal mold provides some detail of both dorsum and venter, and a fragment of internal mould with shell embedded in matrix furnishes further details of the venter and interior of the shell. Unfortunately, the dorsal apertural rim and the shell on the dorsum are unknown. This specimen lacks an operculum, and no disarticulated opercula were included in this species. The indentations on the dorsal internal mould resemble those of Lirotheca wilsoni Malinky and Mapes, 1983, from the Pennsylvanian of Kansas. Maxilites maximus (Barrande, 1867) from the Caradocian of Czechoslovakia also has two crescent-shaped indentations on the dorsum which may represent muscle scars (Marek 1967), but these are smaller and narrower than the indentations on H. carhonaria. Walcott (1884) also compared this species to H. aclis Hall, 1876, from the Devonian of New York, but the apertural rim on that form is flared, whereas the apertural rim of H. carhonaria is unknown. Material. Holotype USNM 14426, National Museum of Natural History. MALINKY AND SIXT: EARLY MISSISSIPPI AN HYOLITHA 353 Occurrence. Walcott ( 1 884, p. 264) reported that this species was discovered in the ‘ Lower portion of the Lower Carboniferous limestone, in canon directly south of a small conical hill on the east side of Secret-canon-road Canon, Eureka District, Nevada. ’ This locality is probably in the NW 1 /4, SE I /4, sec. 36, T 1 9N, R53E, Pinto Summit (15 minute) quadrangle. The ‘Lower Carboniferous’ limestone is probably the Ely Limestone, now recognized as Morrowan (early Pennsylvanian) in age (Larson and Langenheim 1979). Hyolitliesl milleri Sinclair, 1946 Text-fig. 2k V* ]S9A Hyolithes lanceolalus Miller, p. 317, pi. 19, figs. 35, 36. V* 1946 Hyolithes milleri Miller; Sinclair, p. 73. V* 1967 Hyolithes milleri Miller; Yochelson and Saunders, p. 9. non 1845 Hyolithes lanceolatus Morris, p. 289, pi. 18, fig. 8. Description. The conch of this species has a small apical angle, and the apical end appears to be straight. The venter is nearly flat, and that side is smooth without any lirae or other ornament. The ligula is short and the anterior edge appears to be straight. All other features, such as the aperture and operculum are unknown. Remarks. Characteristics attributed to this species by Miller (1894) were derived in part from about two dozen phosphatic tubes which Miller (1894) mistakenly included under this species. H. milleri is unequivocally represented by one specimen (UGGM 3900A) and possibly by a second (UCGM 3900B); the phosphatic tubes were transferred to Enchosloma by Miller and Gurley (1896). Miller’s illustration of H. milleri is a line drawing of a phosphatic tube that bears no resemblance to any authentic hyolith. Until now, these hyoliths have never been illustrated with photographs. Specimen UCGM 3900A is here designated the lectotype; it is 19-5 mm long, and has a width of 4-4 mm. Only the venter is exposed; the dorsum is embedded in matrix from which extraction intact would probably be impossible. Whether that specimen has a shell or is an internal mould is not known with certainty. If shell is preserved, it appears to be smooth and featureless. The presence of an operculum cannot definitely be ascertained because the apertural end is concealed by matrix. Miller (1894) named this species lanceolalus, but Sinclair (1946) renamed it because the name lanceolatus was preoccupied (Morris 1845). Material. Lectotype UCGM 3900A and possible paralectotype under 3900B reposited at University of Cincinnati Geology Museum. Occurrence. Miller (1894) reported that these specimens were discovered in the Chouteau Limestone (Kinderhookian, lower Mississippian, Thompson 1979) ‘near Sedalia, Missouri’. No other details of the occurrence are known. Hyolithesl parvulus Girty, 1926 Text-fig. 2f-i V* 1926 Hyolithes parvulus Girty, p. 38, pi. 6, figs. 18a-18e, 19a-19d. V* 1946 Hyolithes parvulus Girty; Sinclair, p. 79. V* 1967 Hyolithes parvulus Girty; Yochelson and Saunders, p. 9. Description. The venter of this species is nearly flat and grades into narrowly rounded lateral margins. The dorsum is high and the longitudinal axis is narrowly rounded. The apical end of the shell appears to be straight, and the apical angle seems to be small. The conch is covered with fine lirae; they are nearly transverse on the dorsum but curve on the venter to follow the outline of the ligula. The ligula is short and broadly rounded at the anterior edge. The complete aperture and operculum are unknown. Remarks. Two specimens were assigned to this species by Girty (1926); specimen USNM 121 196 is herein designated the lectotype. That specimen is 5-4 mm long, and has an apertural width and height of 2-9 mm and 2-0 mm respectively. The lectotype retains a shell although both ends are broken, so that the complete aperture, as well as the operculum are unknown. Preservation of a 354 PALAEONTOLOGY, VOLUME 33 possible paralectotype (USNM 121 195) of similar size is comparable to the lectotype although that specimen is covered with coarse ribs; whether it represents the same taxon as the lectotype is uncertain. Material. Lectotype USNM 121196 and paralectotype USNM 121195, National Museum of Natural History. Occurrence. Girty (1926, p. 38) reported these specimens from ‘station 2623, about 5-0 km (3-0 miles) east of San Saba County courthouse’ in San Saba County, Texas. The labels associated with the specimens give more information : ‘on the SW side of the road to Chappel, about 3 miles SE of San Saba courthouse, at first sharp turn well uphill.’ The unit which yielded the hyoliths is ‘a thin limestone above the Ellenburger and below the Barnett Shale’: this is the Chappel Limestone of late Kinderhookian to early Osagean (early Mississippian) age (Kier et al. 1979). A recent visit to that locality by J. M. M. showed that the limestone remains exposed on the southwest side of Chappel Road, but no additional hyolith material was discovered. Hyolithesl waverliensis Hyde, 1953 Text-fig. 2a-d v* 1953 Hyolithes waverliensis Hyde, p. 335, pi. 53, figs. 10-15. V* 1967 Hyolithes waverliensis Yochelson and Saunders, p. 9. Description. The conch of this species has a broadly rounded venter which grades into narrowly rounded lateral margins. The dorsum is low and narrowly rounded along the longitudinal axis; the adjacent slopes are slightly inflated. The apical angle is small, and the apical end curves slightly toward the dorsum. The lateral margins of the internal mould are covered with faint transverse growth lirae, and along each lateral margin faint longitudinal lirae are present. The shell, complete aperture and operculum are unknown. Remarks. This species is known from a steinkern and a fragmentary external mould, both of which are counterparts of the holotype. The steinkern is 8 0 mm long, and has an apertural width and height of 5-0 mm and 3-0 mm respectively. Both ends of the steinkern are broken, and no apertural detail is preserved. The steinkern has no operculum, and no disarticulated opercula have been identified for this species. The appearance of the steinkern closely matches the line drawings of figures 13 and 14 on plate 53 (Hyde 1953); figures 10-12 on that plate seem to be restorations for no available specimens of this species match those illustrations, and there is no evidence for the apertural detail shown. Material. Holotype OSU 19895, Orton Museum of Geology, Ohio State University. Occurrence. Hyde (1953, p. 336) discovered this specimen in Bed 1, Byer Member of the Logan Formation (Kinderhookian, lower Mississippian, Collins 1979), near Sciotoville, Scioto County, Ohio. Acknowledgements. J.M.M. thanks the Smithsonian Institution for the Postdoctoral Fellowship which allowed me to undertake this project, and for continued support after my term as a Fellow was completed. R. E. Grant, E. L. Yochelson, T. J. Frest, B. F. Glenister and R. Diecchio kindly read early and later drafts of the manuscript. We are grateful to the Orton Museum of Geology and the University of Cincinnati Geology Museum for their generous loan of specimens. Amoco Research and Production Company defrayed field expenses for S.S. The National Geographic Society kindly provided financial support (Grant 3479-86) which allowed J.M.M. to visit type localities for some of the species described herein. We thank L. Agramonte for her cheerful assistance in the field at Walcott’s Secret Canyon locality in Nevada. We are especially grateful to A. J. Gerk for graciously allowing access to his private collection of hyoliths from the Humbolt Oolite. MALINKY AND SIXT: EARLY M I SS I SSI PPl AN HYOLITHA 355 REFERENCES BARRANDE, J. 1847. PugiwKulus, cin fossiles Pteropoden-Geschlect. Neues Jahrbuch fur Mineralogie, Geognosie, Geologie imd Petrafaktenkunde, 1847, 554-558. 1867. Systeme silwien du centre de la Boheme. Ordres des Pteropodes 3, Prague, 179 pp. BARRETT, T. s. 1876. Notcs OD the lower Elelderburg rocks of Port Jervis, New York, with descriptions of a new pteropod. Annals of the Lycaeum of Natural History, New York, 11, 290-299. BRENCKLE, p. L. and GROVES, J. R. 1985. Calcareous foraminifers from the Humboldt Oolite of Iowa: key to early Osagean (Mississippian) correlation between eastern and western North America. Palaios, 1, 561-581. BREZINSKI, D. K. 1988. Trilobitcs of the Gilmore City Limestone (Mississippian) of Iowa. Journal of Paleontology, 62, 241-245. COLLINS, H. R. 1979. The Mississippian and Pennsylvanian (Carboniferous) Systems in the United States - Ohio. United States Geological Survey Professional Paper, 1110-E, El E26. EICHWALD, E. 1840. Ueber das silurische Schichtensystem in Estland. Zeitschrift fiir Natur- und Heilkunde, St Petersburg, 1 and 2, 210 pp. ELLIS, G. D. 1979. The Mississippian and Pennsylvanian (Carboniferous) Systems in the United States - Michigan. United States Geological Survey Professional Paper, 1110-J, J1-J17. FISHER, D. w. 1962. Small conoidal shells of uncertain affinities, W99-WI43. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part W. Miscellanea. Geological Society of America and University of Kansas Press, New York and Lawrence, Kansas, 259 pp. GERK, A. J. and LEVORSON, c. o. 1982. Unusual beach deposits in oolitic carbonate environments in Mississippian and Recent. Proceedings of the Iowa Academy of Science, 89, 68-70. GiRTY, G. H. 1926. The macro-fauna of the limestone of Boone age. In roundy, p. v., girty, g. h., and GOLDMAN, M. I. Mississippian Formations of San Saba County, Texas. United States Geological Survey Professional Paper, 146, 24-43. glenister, b. f. and sixt, s. 1982. Mississippian biofacies-lithofacies trends, northcentral Iowa. Geological Society of Iowa, Guidebook, 37, 1-21. gray, h. h. 1979. The Mississippian and Pennsylvanian (Carboniferous) formations in the United States - Indiana. United States Geological Survey Professional Paper, 1110-K, K1-K20. HALL, J. I860. Notes and observations on the fossils of the Goniatite limestone in the Marcellus Shale of the Hamilton Group in eastern and central parts of the state of New York, and those of the Goniatite beds of Rockford, Indiana; with some analogous forms from the Hamilton Group proper. Annual Report of the New York State Cabinet of Natural History, 13, 95-1 12. 1876. Illustrations of Devonian fossils: Ptcropoda and Cephalopoda (plates only). New York State Geological Survey, 74 pis. HARPER, J. A. 1977. Gastropods of the Gilmore City Limestone (Lower Mississippian) of north-central Iowa. Unpublished Ph.D. dissertation. University of Pittsburgh. HECKEL, p. H. 1977. Origin of phosphatic shale facies m Pennsylvanian cyclothems of Midcontinent North America. American Association of Petroleum Geologists Bulletin, 61, 1045-1068. HOLM, G. 1893. Sveriges kambrisk-siluriska Hyolithidae och Conulariidae. Sveriges Geologiska Undersorning, Series C, 112, 172 pp. HYDE, J. E. 1953. The Mississippian formations of central and southern Ohio. Bulletin of the Ohio Division of the Geological Survey, 51, 1-355. RAMMER, T, w., BRETT, c. E., BOARDMAN, D. R. and MAPES, R. H. 1986. Ecologic Stability of the dysaerobic biofacies during the Late Paleozoic. Lethcna, 19, 109-121. KiER, R. s., BROWN, L. F. and MCBRIDE, E. F. 1979. The Mississippian and Pennsylvanian Systems (Carboniferous) in the United States - Texas. United States Geological Survey Professional Paper, 1110-S, S1-S45. KLEBELSBERG, R. V. 1912. Die Marine Fauna der Ostrauer Schichten. Jahrbuch der Kaiserlich-kdniglichen Geologischen Reichenstatt, 62, 461-556. KOENEN, A. 1879. Die Kulm-Fauna von Herborn. Neues Jahrbuch fur Mineralogie, Geologie und Palaeontogie, 1879, 309-346. KONiNCK, L. G. DE 1883. Faune du calcaire carbonifere de la Belgique, IV Gastropodes (suite et fin). Annales Musee Royale d 'Histoire Naturelle de Belgique, 1-224. LANDING, E. 1988. Lowci' Cambrian of eastern Massachusetts: stratigraphy and small shelly fossils. Journal of Paleontology, 62, 661-695. LARSON, E. R. and LANGENHEiM, R. L. 1979. The Mississippian and Pennsylvanian Systems (Carboniferous) in the United States - Nevada. United States Geological Survey Professional Paper, 1110-BB, BB1-BB19. 356 PALAEONTOLOGY, VOLUME 33 MALiNKY, j. M. 1988. Early Paleozoic Hyolitha from North America; reexamination of Walcott’s and Resser's type specimens. Journal of Paleontology, 62, 218-232. and MAPES, R. H. 1983. First occurrences of Hyolitha (Mollusca) in the Pennsylvanian of North America. Journal of Paleontology, 57, 347-353. and BROADHEAD, T. w. 1986. New Late Palaeozoic Hyolitha (Mollusca) from Oklahoma and Texas, and their palaeoenvironmental significance. Palaeontology, 29, 303-312. , UNSLEY, R. M. and YOCHELSON, E. L. 1987. Taxonomic revision of Hyolitha from the middle Paleozoic of North America. Journal of Paleontology, 61, 1173-1186. MAREK, L. 1963. New knowledge on the morphology oi' Hrolitlies. Sbornik Geologickych Veci, Paleontologie, 1, 53-72. 1966. New hyolithid genera from the Caradoc of Bohemia. Casopis Narodnilw Muzea, 135, 89-92. 1967. The class Hyolitha in the Caradoc of Bohemia. Sbornik Geologickych Ved, Paleontologie, 9, 51-113. 1972. Middle Cambrian Hyolitha: Maxilites gen. n. Casopis Narodnilw Muzea, 141, 69-72. 1974. New Lower Ordovician hyolithid genera. Vestnik Ustredniho Ustavu Geologickeho, 49, 287-292. 1975. The discovery of a new hyolithid fauna in the Middle Cambrian of Bohemia. Bohemia Centralis, Praha, 4, 64—71. and YOCHELSON, E. L. 1964. Paleozoic mollusk: Hyolithes. Science, New York, 146, 1674-1675. 1976. Aspects of the biology of Hyolitha (Mollusca). Lethaia, 9, 65-82. MILLER, s. A. 1894. Paleontology. Indiana Geological Survey 18th Report, 257-356. and GURLEY, w. f. e. 1896. New species of Palaeozoic invertebrates from Illinois and other states. Illinois State Museum Bulletin, 11, 1-50. MORRIS, J. 1845. Fossil faunas, pp. 261-312. In strzelecki, p. e. Physical description of New South Wales and Van Dieman's Land. Longman, Brown, Green and Longmans, London, 462 pp. NOVAK., o. 1886. Zur Kenntniss der fauna dcr Etage F-fl, in der palaeozischen Schichtengruppe Bohmens. Kdniglichen bohmens Gesselschaft der Wissenschaften Sitzungsberichten, 1-27. 1891. Revision der palaeozischen Hyolithiden Bohmens. Abhandlungen der koniglische Gessellschaft der Wissenschaften 7, 1^8. resser, c. e. 1938. Fourth contribution to nomenclature of Cambrian fossils. Smithsonian Miscellaneous Collections, 97, 1^3. RiCKLEFS, R. E. 1979. Ecology. Chiron Press, New York, 966 pp. SINCLAIR, G. w. 1946. Notes on the nomenclature of Hyolithes. Journal of Paleontology, 20, 72-85. SYSSOiEV, V. A. 1958. The superorder Hyolithoidea. Fundamentals of Paleontology. Akademiya Nauk, S.S.S.R., 2, 184-190. 1962. Cambrian hyolithids from the lower slope of the Aldan shield. Akademiya Nauk, S.S.S.R., 1-63. 1968. Stratigraphy and hyolithids of the oldest Cambrian beds of the Siberian platform. Yakutsk, 1-67. THOMPSON, T. 1979, The Mississippian and Pennsylvanian (Carboniferous) Systems in the United States - Missouri. United States Geological Survey Professional Paper, 1110-N, N1-N22. WALCOTT, c. D. 1884. Paleontology of the Eureka District. United States Geological Survey, Monograph, 8, 1-298. 1886. Second contribution to the studies on the Cambrian faunas of North America. United States Geological Survey Bulletin, 30, 1-369. 1889. Descriptive notes of new genera and species from the Lower Cambrian or Olenellus zone of North America. Proceedings of the United States National Museum, 12, 33-^6. 1890. The fauna of the Cambrian or Olenellus zone. United States Geological Survey Annual Report, 10, 509-760. 1892. Appendix A: Systematic list of the fossils found at Eureka, Nevada. In Hague, a. (ed.). Geology of the Eureka District, Nevada, with an atlas. United States Geological Survey Monograph, 20, 319-333. WELLER, s. 1898. A bibliographic index of North American Carboniferous invertebrates. United States Geological Survey Bulletin, 153, 1-653. wiNCHELL, A. 1862 (1863). Descriptions of fossils from the Marshall and Huron groups of Michigan. Proceedings of the Academy of Natural Science of Philadelphia, 14, 405-430. 1 865 ( 1 866). Descriptions of new species of fossils from the Marshall Group of Michigan and its supposed equivalents in other states; with notes on some fossils of the same age previously described. Proceedings of the Academy of Natural Science of Philadelphia, 17, 109-133. MALINKY AND SIXT: EARLY M I SS I SSI PPI A N HYOLITHA 357 YOCHELSON, E. L. 1984. Speculative functional morphology and morphology that could not function: the example of Hyolithes and Biconulites. Malacologia, 25, 255-264. and SAUNDERS, w. B. 1967. A bibliographic index of North American Late Paleozoic Hyolitha, Amphineura, Scaphopoda and Gastropoda. United States Geological Survey Bulletin, 1210, 1-271. ZAKOWA, H. 1971. Zone of Goniatites granosus in the Galezice Syncline. Instytut Geologiczny, Brace, 60, 1-137. JOHN M. MALINKY Department of Geology George Mason University Fairfax, Virginia 22030, USA SHIRLEY SIXT Department of Geology Typescript received 26 January 1989 The University of Iowa Revised typescript received 20 July 1989 Iowa City, Iowa 52242, USA PSEUDOPLANKTON by PAUL B. wiGNALL and Michael j. simms Abstract. All attached epifaunal species have the potential to colonize floating substrates such as driftwood, externally shelled cephalopods, Sargassum-\\kt algae and marine vertebrates. Such pseudoplankton are preserved in a much wider range of facies than their benthic relatives. However, they are never as abundant as benthos due to the rarity of attachment sites. Pseudoplanktonic species utilize five attachment strategies; cemented, adpressed, pendent, boring and clinging. Overcrowding appears to be a common problem on floats and consequently the pendent strategy, with its limited attachment area relative to the size of the organism, appears to have been favoured by obligate pseudoplankton. However many species are facultatively pseudoplanktonic, making palaeoecological interpretations difficult. Most reported examples of pseudo- plankton, particularly those from black shale facies, are too abundant to be attributed to this group and, in the majority of cases, a benthic mode of life is more plausible. The fossil record of pseudoplankton is thought to be considerably poorer than has hitherto been suggested. Evaluation of the literature reveals a low, although variable diversity of pseudoplanktonic populations through the Phanerozoic. High diversity in the mid- Palaeozoic is due to the presence of large orthoconic nautiloids which provided ideal ffoating substrates for a number of groups. Unexplained diversity minima occurred in the Permian and Cretaceous. In any palaeoecological work it is essential to distinguish between benthic and pelagic elements of the fauna since each provides evidence of their fundamentally different environments. In most instances it is relatively easy to discern the general mode of life of an organism from the morphology of the hardparts or from comparison with extant relatives. In particular, the constraints on morphology of benthic organisms differ greatly from those of free-swimming (nektonic) and drifting (planktonic) organisms. However, a few species have the morphological adaptations of epifaunal benthos yet they pursue a mode of life attached to floating objects (either organic or inorganic) in the water column and hence are effectively planktonic; these are termed pseudoplanktonic forms (alternatively known as epiplanktonic or pseudopelagic forms in some studies). Excluded from this definition are those organisms which secrete their own float, for these are more properly classified with the true plankton. The basic morphology of pseudoplankton means that, should they become detached from their floating substrate before burial, it may be difficult to deduce their original mode of life. As a result, conflicting palaeoecological interpretations have arisen frequently in the literature. The aim of this paper is to provide criteria for the recognition of pseudoplankton in the fossil record and to discuss some of the biological constraints and consequences of this unusual mode of life. A tripartite classification scheme is then proposed for pseudoplanktonic forms. Finally the moderately diverse, though patchy, history of pseudoplankton in the Phanerozoic is reviewed. A large proportion of fossil pseudoplankton described in the literature is recorded from black shales. The depositional conditions of this facies undoubtedly provide excellent potential conditions for the preservation of pseudoplankton, but in many cases the rationale behind such reports lies in the assumption that black shale environments are inimical to benthic life. Therefore, by default, any apparently benthic fossils are considered to have fallen on to the sea floor from floating substrates higher in the water column. Taphonomic and functional morphological evidence have rarely been cited to support such interpretations and, as will be discussed below, many examples of so-called black shale pseudoplankton were probably truly benthic. IPalaeontology, Vol. 33, Part 2, 1990, pp. 359-378.| © The Palaeontological Association 360 PALAEONTOLOGY, VOLUME 33 THE RECOGNITION OE EOSSIL PSEUDOPLANKTON Virtually every aspect of the biology, taphonomy and facies distribution of pseudoplanktonic species differs from that of benthic species. Thus a number of features can be used to identify pseudoplankton in the fossil record. Association with floating objects Pseudoplanktonic forms are most readily identified when they are preserved together with their floating substrate. The frequency with which this occurs is partially dependent on the style of attachment; for example, cemented bivalves are more likely to remain fixed than byssate forms which readily become dislodged after death. The energy of the depositional environment is also important - quiet conditions, such as those leading to organic-rich shale accumulation, are particularly favourable for the preservation of intact pseudoplanktonic colonies. A range of floating attachment sites was and is available in the marine realm. Driftwood. Floating logs have been available for colonization by pseudoplankton since the appearance of trees in the late Devonian. Estimates for the maximum floating duration of driftwood range from less than one year (Kauffman 1981) to ten years or more (Simms 1986). The actual time will depend on a combination of factors, such as the nature of the wood, its size, and the influence of osmosis and bacterial sealing (Simms 1986). In general the carrying capacity of wood appears to be high. However, logs can become colonized both while they are afloat and after sinking to the substrate. Pseudoplanktonic colonies are generally found beneath driftwood or alongside the wood if it sank to the substrate at an oblique angle (e.g. Seilacher et al. 1968; text-fig. 1). Most pseudoplanktonic colonies on driftwood have been recorded from organic-rich shales (e.g. Withers 1928; McIntosh 1978; Simms 1986) but they are also known from more normal marine mudrocks. Thus Davis and Elliot (1958) record colonies from the Eocene London Clay, whilst a log with abundant epibionts on its lower surface has been examined from aerobic biofacies of the Kimmeridge Clay (text-fig. 2). Lepadomorph barnacles are the dominant driftwood colonizers in modern seas (Schafer 1972). TCXT-HG. 1. Section through a nodule from the Obtusum Zone (Sinemurian, Lower .Turassic) of Stonebarrow Cliff', Charmouth, Dorset, illustrating driftwood, partially calcitized, with individuals of Cuneigervillia (outlined) occurring beneath and alongside the wood. Field of view is 100 mm wide. WIGNALL AND SIMMS: PSEUDOPLANKTON 361 UPPER SURFACE TEXT-FIG. 2. Epizoan abundance measured on a 3-5 m long piece of driftwood from the lower Mutabilis Zone (Lower Kimmeridge Clay, Upper Jurassie) at Wyke Regis, Weymouth, Dorset. The majority of enerusters is found beneath the wood whilst the speeimens located towards the edges on the upper surface were probably originally on the flanks of the log. The occurrence of " Rliynchonella" subvwicibilis beneath this log supports Ager’s (1962) contention that this species was, at least occasionally, pseudoplanktonic. Several branches on the log would have inhibited it from rolling on the sea floor, thereby discounting the possibility that the log was overturned to cause the smothering of the epizoans. Externally shelled cephalopods. Nektonic or nektobenthic cephalopods have been available as hosts for pseudoplankton since the Ordovician. The earliest colonizers were bryozoans and inarticulate brachiopods found on large orthoconic nautiloids in the Upper Ordovician (Havhcek 1 972 ; Lockley and Antia 1980; Baird et al. 1989). Goniatites, the dominant Upper Palaeozoic cephalopods, were generally too small to support any significant epifauna although they may also have been able to defend themselves from colonization (Boston et al. 1988). Large ammonoids became common in the Mesozoic and many examples of oyster encrustation are known (e.g. Meischner 1968; Heptonstall 1970; Riccardi 1980; Seilacher 1982a; Tanabe 1983). There has been considerable debate about whether ammonite colonization occurred during life (Seilacher 1982a, 6), after death but whilst the ammonite was drifting (Palmer 1987; Tanabe 1983), or after the ammonite settled to the bottom to form a localized hard substrate (Kauffman 1981). In the last case, ammonites lying on the sea floor should be encrusted on the upper surface only, although this is not always easy to determine if aragonite dissolution has occurred at an early stage of burial. However, if the ammonite was overturned by foraging organisms then it would be possible for both sides to become encrusted while it lay on the sea floor. Other distinguishing criteria have been discussed by Seilacher ( 1982u). In-life colonization is thought to be characterized by orientated growth, commonly towards the aperture of the cephalopod. Individual ammonites may be heavily encrusted whilst other specimens, from the same horizon, totally lack epibionts. Typically the encrustation is host-specific with heavily ribbed ammonite species being preferentially colonized (Seilacher 1982r/; Doyle and Whitham, in press). Such patterns are only likely to occur if infestation occurred in the water column where some ammonites may have defended themselves against infestation while others positively encouraged epibionts as a form of camouflage. This implies that most ammonite colonization occurred during life. Defence against unwanted epizoans may have been through active cleansing using tentacles. Alternatively, the possession of a thick 362 PALAEONTOLOGY, VOLUME 33 periostracum may inhibit boring and encrustation (Bottjer 1981). In rare cases ammonites have been observed to deviate from their normal planispiral growth pattern due to the presence of large oysters on one flank (Merkt 1966; Heptonstall 1970). Similarly, serpulids growing on the ventral margin of ammonites have been distally overgrown by the later whorls of the ammonite, providing unequivocal evidence of in-life colonization (Merkt 1966). Nekroplanktonic colonization of dead ammonites floating at the surface cannot always be distinguished from in-life colonization. Diagnostic features include the presence of epibionts within the body chamber of ammonites. The uppermost flanks would project above the surface of the water after death and so these areas would be free of epizoans. Such distributions have rarely been recorded. Data from Recent Nautilus suggest that hydrostatic pressure rapidly fills the phragmocone with water following death and thus precludes significant nekroplanktonic drifting (Chamberlain et al. 1981). Ammonites were probably less robust than Nautilus and hence even less likely to remain afloat for long after death (Seilacher 1960). In general the majority of pseudoplankton associated with ammonites appears to have colonized the ammonite whilst it was alive. Ammonites with apparently in situ populations of byssally attached bivalves have been encountered in organic-rich facies from the Lower Jurassic. Seilacher (1982a, fig. 9) illustrated a colony of Gervillia lanceolata clustered around the venter of an ammonite close to the aperture. A similar distribution has been noted for specimens from the Sinemurian (Lower Jurassic) of Dorset; Plagiostoma has been found associated with Amioceras from the Turneri Zone, Cuneigervillia with Asteroceras from the Obtusum Zone, and Oxytoma inequivalve attached to Echioceras from the Raricostatum Zone (text- fig. 3). Ti;xT-FiG. 3. Five specimens of Oxytoma (arrowed) close to the apertural margin on the venter of a specimen of Echioceras. Collected from the Raricostatum Zone (Sinemurian, Lower .lurassic) of Charmouth, Dorset, by C. E. Savrda. The bivalves were probably suspended from beneath the ammonite during life. Benthic colonization would have been expected to produce a more random association. Ammonite is 19 mm in diameter. WIGNALL AND SIMMS: PSEUDOPLANKTON 363 In an unusual case of ammonite encrustation, documented by Cope (1968) from the Kimmeridge Clay, oysters were only found cemented to the lower side of ammonites, mainly in the umbilical region. This he attributed to a photonegative response of the oyster larvae settling under ammonite shells lying on the sea floor. However, extensive collecting from the same succession revealed that, of 20 oyster-encrusted ammonites, 14 had oysters on both flanks, 5 had oysters on the upper surface only and only a single specimen was found to have oysters restricted to its lower surface. These data suggest that most oyster-encrustation occurred while the ammonites were still in the water column. The five examples with colonization on the upper surface tended to have oyster nests nucleated anywhere on the shell with little preferred growth orientation. Examples with oysters on both sides of the ammonite tended to show a radial growth orientation centred on the umbilical region. This pattern may have been caused by the oyster spat seeking out the most sheltered region of the ammonite shell although it could also reflect the greater age of the umbilical region compared to the outer whorls. In support of this, all 14 of the Kimmeridge Clay ammonites with oysters on both flanks are large, old specimens greater than 90 mm in diameter (text-fig 4). Similarly, in-life encrustation of ammonites in the Lower Jurassic is mainly restricted to adult examples of large, presumably long-lived individuals. The radial orientation is caused by the competitive growth of the oysters with each individual being forced outwards from the umbilicus by the presence of its neighbours, causing the growth of wedge-shaped morphologies. 40- [ j Ammonites, unencrusted (129) 32- lav cr OJ 16- ■ Ammonites, encrusted by oysters on both sides (12 specimens) 100 ISO Ammonite diameter (millimetres) TEXT-FIG. 4. Size-frequency histogram illustrating the proportion and abundance of oyster-encrusted pectinatitid ammonites from the Upper Kimmeridge Clay (Upper Jurassic) of the Dorest coast. The over- representation of large and oyster-encrusted specimens in this sample is a collecting artifact. All specimens in the P. B. Wignall collection. Vertebrates. For most marine vertebrates the presence of pseudoplankton is undesirable since it causes a drastic increase in drag. Precautions against such colonization include the ability to shed scales and the development of a skin surface unsuited to settlement. Only in large or slow moving vertebrates, such as whales and turtles, is the addition of large species of pseudoplankton unlikely to alter the hydrodynamic properties. Whales are commonly infested with large numbers of barnacles {Xeiiohalaims and Coronula), copepods and a diverse meiofauna including diatoms (e.g. Holmes 1985). Turtles are commonly host to an even greater range of epizoans, including gastropods, bivalves, hydroids, crabs and barnacles (Frazier et al. 1984). Due to the rarity of soft tissue preservation in the fossil record, no examples of pseudoplankton on vertebrates are known, though it is possible that suitably enlightened investigation of the immediate surrounds of large intact vertebrates in anoxic sediments may reveal their presence. 364 PALAEONTOLOGY, VOLUME 33 Comilariids. A diverse pseudoplanktonic fauna, including punctate brachiopods and bryozoans, has been found attached to conulariids in the Carboniferous, Bear Gulch Limestone of Montana (Williams 1983), whilst inarticulate brachiopods are considered to have attached to conulariids in the Ordovician (Havlicek 1972). Such occurrences are surprising considering the small, delicate nature of these organisms. Chondrophorines. The fragile chitinous floats of these jellyfish-like organisms have only recently been identified in the fossil record, many examples previously having been misidentified as patellid gastropods (Stanley and Kanie 1985). As with conulariids, the chondrophorines were probably too fragile to support epizoans commonly, although an exception is the attached fauna found on chondrophorines in the early Devonian Hunsriickschiefer of Germany (Yochelson et at. 1983). Algae. Vesicular algae, such as Sargassum, are one of the most common attachment substrates for pseudoplankton at the present day. Brown algae are commonly invoked as attachment sites for ancient epizoans, but such soft material has a very low preservation potential (see Jerzmanska and Kotlarczyk (1976) for an exception). Ruedemann (1934) has illustrated examples of so-called algae from Palaeozoic black shales. Many of these are probably trace fossils but a number of carbonaceous branching structures (e.g. Ruedemann 1934, pis. 2-3) remain unexplained. Rickards (1975) suggested that they may be the extrathecal tissue of graptolites. Pumice. Pumice, which is probably the only naturally occurring non-organic substrate for pseudoplankton, has been available throughout the Phanerozoic. Its floating duration is dependent on size, initial density, temperature upon entering water and the size distribution and connectivity of the vesicles (Whitham and Sparks 1986). Their experiments showed that some pumice remained afloat for more than 18 months. Extrapolation from their observations suggested that large, low density masses up to 1 m across may remain afloat for more than 10 years. Jokiel (1984) also inferred, from the size of a coral colony attached to floating pumice, that pumice could remain afloat for two to three years at least. Despite such observations, examples of pseudoplankton attached to pumice are very rare in the fossil record, although Doyle and Whitham (in press) have recorded oxytomid bivalves associated with pumice from the Upper Jurassic. Abundance and facies distribution In many instances pseudoplankton may become detached from its floating substrate before burial or, in the case of epizoans attached to floating seaweed, their attachment site may not be preserved. In such situations less direct methods of taphonomic analysis must be used to determine their original mode of life. Pseudoplankton today only constitutes a tiny fraction of the total abundance of epizoans due to the rarity of floating attachment sites compared to the abundant sites available in the benthic environment. Conditions are unlikely to have been significantly different in the past and consequently pseudoplanktonic species should generally be a rare component of fossil assemblages. For examples modern-day lepadomorph barnacles are an important pseudoplanktonic group but their plates are only found in very small numbers, scattered through a wide range of marine sediments (Schafer 1972). Only under slow sedimentation rates, such as those commonly found in the depositional environments of organic-rich shale (Tyson 1987), will pseudoplankton ever occur in anything approaching moderate numbers. Even under such conditions, pseudoplanktonic forms should not occur as more than a few individuals scattered across bedding planes. Exceptionally slow sedimentation rates may lead to greater abundances although in these instances the pseudo- planktonic species should be accompanied by high concentrations of truly pelagic forms such as fish and marine vertebrates. Epizoans attached to brown algae may reach moderate abundances but do not contribute more than a few percent to the total skeletal carbonate content of modern sediments (Pestana 1985). Pseudoplanktonic drifting causes species to be preserved in a wide range of benthic environments. WIGNALL AND SIMMS: PSEUDOPLANKTON 365 Examples of surprising facies distributions include the rare occurrence of sponges, crinoids and corals in organic-rich shales (e.g. Ruedemann 1934; Bulman 1964; Simms 1986, 19886; Baird et al. 1989) where depositional conditions were oxygen-restricted. These groups are known to be particularly intolerant of such conditions at the present day (e.g. Webster 1975) and they are unlikely to have been true benthos in the ancient examples. This is not to say that all apparently benthic species in black shales were pseudoplanktonic, as has been tacitly assumed in many studies, for many groups of organisms, particularly molluscs, are able to live under conditions of very low oxygen (Sageman et al. in press). Rafts of sunken Sargassum and their attached fauna have been recorded from modern benthic environments ranging from the intertidal zone down to abyssal depths (Schoener and Rowe 1970; Pestana 1985). Such a wide ranging facies distribution can also be expected for fossil pseudoplankton; indeed, a facies-crossing pattern is one of the most reliable, and widely used criteria for detecting ancient examples (e.g. Ager 1965; Tchoumatchenco 1972). Similarly, pseudoplanktonic species have a widespread geographic distribution when compared to their benthic relatives giving them a good potential for correlation (Schafer 1972). For example, the pseudoplanktonic genera Seirocriinis and Pentacrinites are the only Lower Jurassic crinoids known from both the Boreal and Tethyan realms of the Lower Jurassic (Klikushin 1982). Attachment Strategies A precondition for all pseudoplanktonic species is the ability to attach to a floating substrate. Five attachment strategies can be recognized. Cemented. Cementation provides one of the most secure means of attachment but it has the disadvantage that it requires a relatively large attachment area on a substrate where there might be intense competition for space. Oysters, bryozoans and serpulids are amongst the most frequently encountered cemented pseudoplankton in the fossil record, whilst acrothoracian barnacles are common cementers today (e.g. Landman et al. 1987); cnidarians and corals are less frequent cementers (Jokiel 1984). Pendent. Pendent forms have a relatively small attachment area and dangle at some distance beneath their float. This strategy has the advantage of minimizing the area required for attachment on a floating substrate where crowding may be a serious problem. Crinoids (Simms 1986), certain lanceolate bivalves (text-fig. 5), lepadomorph barnacles (Moore 1867) and, more rarely, articulate brachiopods (text-fig. 2) all belong in this category. The pseudoplanktonic adaptations of lanceolate bivalves include a short hinge line and weak dentition. Both factors tend to reduce the articulation strength, though this is not detrimental since the ligament does not have to operate against the confining pressure of the sediment such as is experienced by endobyssate bivalves. The byssus emerges ventrally to a sharp anterodorsal angle of an equivalve shell. Consequently the attachment area is effectively reduced to a point. In benthic epibyssate bivalves, such as Mytilus, the byssus emerges more centrally along the ventral margin and the contact area with the substrate is greater. The lanceolate bivalve morphology is also adapted to a reefal environment (Fiirsich and Wendt 1977) where such forms may be able to hang beneath crevices or from branching corals. However, such reef dwellers are typically much thicker-shelled than their thin, fragile pseudoplanktonic relatives. Few generalizations can be made about the morphology of pseudoplanktonic crinoids since very few are considered to have adopted this mode of life. Both the early Jurassic Pentacrinitidae and the late Devonian Melocrinites have an endotomous pattern of arm branching (Simms 1986), though this is by no means unique and is also found in many benthic taxa. The stem of pentacrinitids differs from other articulates in showing an apparent increase in flexibility distally, as might be expected for a pseudoplanktonic crinoid (Seilacher et al. 1968), but this has not been documented for melocrinitids. The dense spacing of cirri on the proximal and distal parts of pentacrinitid stems may also be specifically adaptated for a pseudoplanktonic mode of life, though 366 PALAEONTOLOGY, VOLUME 33 Driftwood-attached bivalves TEXT-FIG. 5. Lanceolate bivalve genera inferred to have been pseudoplanktonic. Gervillia lanceolata is from the Lower Toarcian and is commonly associated with ammonites. The species of Posidonia and Ccmeyella are from the lower Namurian of northern England and are inferred to have attached to driftwood which was fairly common at this time. again this cannot be confirmed from observation of melocrinitids which have a cemented type of attachment at the base of the stem. Adpressed. Adpressed forms attach themselves by organic tissue or threads and hold tight to the substrate. This has the advantage over the pendant strategy of reducing the chances of dislodgement but, as in cementing forms, the attachment area is relatively large. This mode of life is common in many benthic groups such as mytiliform bivalves, including many inoceramids, and the inarticulate brachiopod genera Discinisca and Orhiculoidea, of which some individuals may have been pseudoplanktonic (e.g. Tanabe 1983; Seilacher 1982a, ib). Boring. Tunnelling into the floating substrate is the most secure strategy against dislodgement but has the detrimental effect of reducing its floating properties. Boring bryozoans and acrothoracian barnacles have been reported from belemnites and Nautilus (Landman et al. 1987; Seilacher 1968). Seilacher (1968) argued that the consistent orientation of barnacle borings in the distal portion of a belemnite guard indicated that such infestation occurred during life. The majority of belemnite encrustation and boring is, however, random and probably occurred after death. Wood is commonly bored by isopods and, more importantly for the fossil record, teredinid bivalves. Boring bivalves appear to utilize wood either as a dwelling, from which to filter feed, or as the source of food itself (Kelly 19886). The latter strategy severely reduces floating duration; consequently wood- eating bivalves are unlikely to be pseudoplanktonic for long. Clingers. These are species which are able to move about their float, often in search of prey; WIGNALL AND SIMMS: PSEUDOPLANKTON 367 Friedrich (1969) referred to them as haptic forms. Only Recent examples are known with certainty. They include pycnogonids, flatworms and gastropods attached to Sargusswu (Morris and Mogelberg 1973), gastropods and crabs attached to turtles (Frazier et al. 1984) and an isopod, Idotea nietallica, which clings to blobs of crude oil (Herring 1969). The possibility that ancient, vagrant epifaunal forms, such as gastropods, may represent pseudoplankton does not appear to have been appreciated, although they are never likely to be common. A number of supposed pseudoplanktonic forms from the fossil record cannot be assigned to any of the above five categories. These examples all occur in finely laminated shales of mid-Palaeozoic age and include the praecardioid bivalve genera Mamiliciila, Butovicella, Cardiola, Slava and Dualma (Watkins and Berry 1977; Watkins 1978) and lingulid brachiopods (Barron and Ettensohn 1981). Biitovicella possibly utilized an epibyssate, adpressed strategy (Kriz 1969) but the remaining bivalves are all endobyssate forms (Pojeta et al. 1976) which are unlikely to have been able to attach to floating objects. Lingulids are infaunal benthic forms which could not have led a pseudoplanktonic life. BIOLOGICAL CONSTRAINTS ON A PSEUDOPLANKTONIC LIFESTYLE The biology of pseudoplankton is severely constrained by a number of ecological features unique to this mode of life. As well as an ability to attach to the float, discussed above, the great rarity of floating substrates necessitates a rapid response when such a site is encountered. Species which produce large numbers of planktonic larvae will clearly have the greatest potential for exploiting floating objects. This can be achieved by large adults producing a large number of offspring at one time or by small adults producing fewer larvae but at more frequent intervals (Jablonski and Lutz 1983). Increasing size of individuals in the former case and continuing recruitment onto the original float in both cases results in a population of large biomass which will rapidly overload the floating attachment site, possibly before the epizoans can reach sexual maturity. Even before the float becomes overloaded it may run aground and cause the premature demise of its occupants. Thus two counteracting selective pressures can be seen to operate on pseudoplanktonic species. A solution to this problem includes the rapid attainment of maturity. This may be achieved by maturation at a relatively small size or by accelerating the growth rate to reach a large size in a short time. The former can be attained relatively easily through heterochrony, in particular paedomorphosis (McNamara 1986), although such forms will be restricted by their small size of producing relatively few larvae at a time. Accelerating growth rates is, perhaps, less straightforward since it requires considerable extra expenditure of both energy and materials. Amongst the pseudoplankton only crinoids appear to have adopted this latter strategy of rapid growth to large size but this required specialized adaptations to increase feeding efficiency. In both the Devonian camerate Melocrinites and the early Jurassic articulates Pentacruutes and Seirocriiius, the arms branch endotomously beyond the second division, an unusual arrangement amongst crinoids generally. This is interpreted as the most efficient filtering arrangement possible for the most economical outlay of materials, comparable with the ideal arrangement of roads on a banana plantation (Cowen 1981). The early Jurassic pentacrinitids further enhanced the efficiency of their filtration fan by the suppression of syzygial articulations in the arms, resulting in the attainment of almost complete pinnulation, a feature peculiar to this group (Simms 1986). A further strategy which increases the likelihood of colonizing rare attachment sites includes the delay of larval metamorphosis, thereby prolonging the time spent in the water column and thus the time available for encountering attachment sites. Lockley and Antia (1980) documented a probable example of delayed larval metamorphosis in Schizocrania, an inarticulate brachiopod attached to orthoconic nautiloids in the Ordovician. There is strong evidence to suggest that the presence of adults may also encourage larval settlement, possibly by a chemoautotrophic response (Crisp 1979; Grosberg and Quinn 1986). The ability of adults to attract larvae probably accounts for the often observed pattern of ‘all-or-nothing’ pseudoplanktonic colonization. Thus, once a few epizoans are established, their presence can rapidly induce large numbers of other individuals to colonize. For 368 PALAEONTOLOGY, VOLUME 33 example, the heavily encrusted log in text-figure 2 occurred with several other large logs which were totally devoid of epizoans. Similarly, Baird et al. (1989) noted all-or-nothing encrustation of orthocones. The development of a relatively lightweight body plan and the reduction of any skeletal components is a further strategy which is likely to prolong the floating duration of attachment sites. Thus, the pseudoplanktonic bakevelliid bivalves of the Lower Jurassic are considerably thinner- shelled than their benthic relatives (Seilacher 1984; text-fig. 5). Many of the biological constraints on marine wood-borers are very similar to those acting on pseudoplankton as both exploit generally rare resources. It is thus interesting that wood-boring bivalves exhibit many of the life strategies found in pseudoplanktonic species; indeed many such bivalves are also pseudoplanktonic due to the nature of their substrate! Turner (1973) noted that such bivalves tended to be T’-strategists. THE CLASSIFICATION OF PSEUDOPLANKTON It is clear from an examination of the Recent and ancient record that pseudoplanktonic species can be readily classified on the basis of the relative frequency with which they adopt this mode of life and their dependence upon it. Obligate Pseiidopkmkton This group encompasses forms which have an exclusively pseudoplanktonic adult stage. The biological constraints, discussed above, require highly specialized adaptations which may inhibit them from returning to a benthic existence. Ancient examples include two independent crinoid groups, the Melocrinitidae and Pentacrinitidae (Wells 1941 ; Simms 1986) and probably many of the thin-shelled, lanceolate bivalve genera of the Palaeozoic and Mesozoic (text-fig. 5). Certain genera of lepadomorph barnacles constitute the most important group of obligate pseudoplankton at the present day with cosmopolitan genera such as Conchoderma and Coronida occurring attached to a wide range of substrates ineluding whales, turtles and sea snakes (Friedrich 1969). Other genera, such as Lepas itself, occur on driftwood (Schafer 1972). In the past many apparently obligate pseudoplanktonic species appear to have been substrate-specific; for example in the Posido- nienschiefer Gervillia is restricted to ammonites (Seilacher 1982a). It is noteworthy that many obligate pseudoplankton utilize the pendent strategy which suggests that a limited attachment area is a strong selective advantage. A rather more diverse range of attachment strategies is illustrated by the obligate pseudoplankton associated with Sargussum (Morris and Mogelberg 1973). Facultative Pseudoplankton The morphology and physiology of many benthic epifaunal species are pre-adapted to enable a limited number of these forms to settle successfully on floating objects. This chance colonization, by normally benthic species, considerably confuses their ecology, particularly in the fossil record, for it misleadingly suggests tolerance of a wide range of facies. For example all the pseudoplanktonic species recorded in text-figure 2 were probably also benthic, for other specimens of these species occur in greater abundance in other Kimmeridge Clay horizons where they are not associated with driftwood. Pseudomytiloides duhius is frequently assoeiated with driftwood in the Posidonienschiefer (Seilacher 1982a), but in many horizons this bivalve is too abundant for all specimens to have been derived from the relatively rare examples of driftwood (Kauffman 1981). The discovery of a colony of pseudoplanktonic species, assoeiated with a float such as driftwood, does not indicate that all individuals of that species led sueh an existenee. A ccidentcd Pseudoplank ton This category includes very rare oeeurrences where benthic species are unintentionally cast adrift. Examples include forams attached to blades of seagrass whieh become detached (Brasier 1975), the WIGNALL AND SIMMS: PSEUDOPLANKTON 369 gastropods found on turtles which are thought to have attached whilst the turtles were at rest on the sea floor (Frazier et al. 1984) and possibly an early Jurassic crinoid (Simms 1988/)). Whilst such chance occurrences are highly unlikely to be preserved in the fossil record, they are of great importance in palaeobiogeography as they vastly increase the dispersal ability of a huge range of benthic species across wide oceans (Ekman 1953; Hallam 1973; Jokiel 1984). The classification of pseudoplankton is complicated by those species which are able to survive once the floating object has sunk to the substrate. For example ammonites have been observed encrusted by oysters, nucleated in the umbilical region, which have grown-over the outer whorls and across the substrate (Clausen and Wignall, 1990, plate 6d). In such examples initial colonization probably occurred whilst the ammonite was alive and floating in the water column, but they subsequently became ‘facultative benthos’ after the ammonite had died and settled to the sea floor. The ability to distinguish between obligate and facultative pseudoplankton is of great importance both in evolutionary and ecological studies. The biological constraints, discussed above, only apply to obligate pseudoplankton. For the facultative species the evolutionary pressures are basically those experienced by the benthic population which contains the majority of individuals of that species. In palaeoecological studies the discovery of a colony of facultative pseudoplankton may be incorrectly extrapolated to assume that all populations of the species pursued this mode of life. The distinguishing features between obligate and facultative pseudoplankton are given in table 1 . TABLE I . Criteria used to distinguish obligate from facultative pseudoplankton in the fossil record. Obligate Facultative Abundance and Always rare, found Occur in a wide range facies distribution in a wide range of of facies but they reach facies. a peak abundance in one facies type. Occurrence Nearly always Very rarely associated with associated with a floating object. floating objects. Commonly found with Found with a range of a particular type of float (host specific) floating objects. Morphology Normally thin-shelled. Broad range of morphologies. commonly pendent. e.g. cemented, adpressed. Lifestyle In comparison to their nearest benthic relatives, they may illustrate delayed larval metamorphosis followed by rapid growth rates. No unusual adaptations. EVOLUTION IN OBLIGATE PSEUDOPLANKTON The profound ecological constraints which influence the biology and morphology of obligate pseudoplankton exert a considerable influence on the evolution of pseudoplanktonic taxa. Once the two basic problems of attachment and reproductive success have been overcome by such taxa, further evolution is likely to be restricted largely to ‘fine tuning’ of the original strategy. However, opportunities to investigate this hypothesis are severely limited as relatively few supraspecific taxa have a fossil record that is sufficiently good to document evolutionary lineages. This problem is compounded in obligate pseudoplankton due to the rarity of such groups, both in numbers of individuals and taxonomic diversity. 370 PALAEONTOLOGY, VOLUME 33 The Mesozoic crinoid family Pentacrinitidae is the only group of obligate pseudoplankton for which an evolutionary lineage has been recognized and documented. The two closely related genera, Pentacrinites and Seirocrinus, were already quite distinct when first recorded in the late Triassic. Thereafter the two lineages underwent very little morphological change through time and individual species showed unusual longevity by comparison with contemporaneous benthic crinoids. They also showed apparent immunity to the benthic hypoxic event which caused a major faunal turnover in the early Toarcian (Hallam 1986; Simms 1986, 1988o), a feature perhaps to be anticipated in pseudoplanktonic taxa. Seirocrinus subangularis, a Carixian to Toarcian (Lower Jurassic) species, differs from the Norian (Upper Triassic) S. klikushini only in having slightly fewer brachials in each brachitaxis and in the development of a slightly more complex pattern of endotomous arm branching than is seen in other pentacrinitids (Simms 1988a, 1990). The two earlier species of Pentacrinites, P. doreckae (Hettangian to lower Sinemurian) and P. fossilis (upper Sinemurian) show an almost parallel change in the number of brachials per brachitaxis (Simms 1988a). These changes relate to an increased food-gathering capability in the later species. In the four described species of Pentacrinites there are several marked changes between the earlier Pentacrinites doreckae-fossilis part of the lineage and the succeeding P. dichotonnis-dargniesi lineage. In the latter group the stem is very much shorter with densely-spaced cirri, while the arms have a more poorly developed pattern of endotomous branching and contain syzygial articulations at one or two points, interrupting the pinnule spacing. Furthermore, although S. subangularis, P. doreckae and P. fossilis are very frequently found attached to driftwood (data on this are not available for S. klikushini), there are no records of the P. dichotomus-dargniesi group having been found in association with driftwood or any other float, yet P. dichotomus, at least, otherwise conforms to all the criteria used to identify pseudoplankton. The obvious assumption here is that a fundamental change in life strategy occurred between P. fossilis and P. dichotomus. The morphological changes suggest that the latter group were no longer subject to the severe selection pressures which operated on the earlier, definitely pseudoplanktonic, taxa. The implication of this is that they either exploited an unusually stable floating substrate, not preserved in the fossil record, or became either truly planktonic, which seems unlikely, or benthic. The latter strategy certainly seems to apply to P. dargniesi, but the mode of life of P. dichotomus remains unclear. Evolutionary case histories are less well documented for other pseudoplanktonic groups. Gervillia lanceolata, a probable obligate pseudoplanktonic bivalve from the Lower Jurassic, has a long fossil record extending from the Hettangian to the Lower Toarcian (Hallam 1976), although this is not exceptional for a bivalve. Unlike the pentacrinitids, G. lanceolata did not survive the early Toarcian hypoxic event (Hallam 1986). The species duration for the mid-Carboniferous homeomorphs (text- fig. 5) is considerably shorter (P. B. Wignall, unpublished data). In all cases the pseudoplanktonic bivalves appear to be species derived from benthic ancestors rather than part of a pseudoplanktonic lineage, implying that this mode of life was, in many cases, an evolutionary 'dead-end’. BLACK SHALE PSEUDOPLANKTON The previous review of pseudoplankton in the fossil record has revealed that the majority of examples are reported from finely laminated black shale facies. This may be due to the favourable preservational conditions that occur in black shale depositional environments. The generally low sedimentation rates allow relatively large numbers of pseudoplankton to accumulate whilst the low energy conditions and lack of scavengers in such oxygen-restricted environments are further factors which increase the preservational potential of commonly fragile pseudoplanktonic species. In the mid and late Palaeozoic the most commonly reported pseudoplankton in black shales are thin-shelled brachiopods belonging to the chonetids, plectambonitids, strophomenids and Leiorhynchus (Havlicek 1967; Bergstrom 1968; Thayer 1974), and praecardioid bivalves (Krebs 1979; Watkins 1978), which are inferred to have attached to algae. These groups are widespread spatially within a deep-water, fine grained facies, but they do not occur in other facies. This argues against a pseudoplanktonic lifestyle. Their distribution within the sediment is similarly suggestive WIGNALL AND SIMMS: PSEUDOPLANKTON 371 of a benthic existence for they are common on certain bedding planes and absent in the intervening strata; a distribution typical of frequent, opportunistic colonization (Wignall and Myers 1988). Floating algal communities in Recent oceans are characterized by high diversity assemblages dominated by bryozoa. However, the Palaeozoic assemblages are generally mono- or paucispecific. The combined evidence of low diversity, facies restriction and large numbers restricted to individual horizons strongly suggests that the brachiopods and praecardioids were benthic forms in black shales (Thompson and Newton 1987; Sheehan 1977). This has important implications for the depositional conditions of this facies as it indicates that benthic oxygen was available for at least short periods of time. An even larger volume of literature on pseudoplankton relates to a diverse group of pterioid bivalves which occur in black shales from the Devonian to the Cretaceous (Hudson and Cotton 1943; Ichikawa 1958; Hayami 1969). The group includes the posidoniids, halobiids and some of the monotids, buchiids and inoceramids. The functional morphology of these bivalves is far from clearly understood and opinions have changed markedly over the past hundred years. Initially they were thought to be benthic, reclining forms but, with the interpretation of black shale as hostile, anoxic depositional environments, inimical to benthic life, they were reinterpreted as pseudo- plankton (e.g. Paul 1939; Hudson and Cotton 1943); a view still widely held today (Hayami 1969; Krebs 1979; Rieber 1982; Campbell 1985; Schumann 1988). Jefferies and Minton (1965) proposed the interesting alternative of a free-swimming lifestyle for some posidoniids. Other recent studies of Carboniferous and Lower Jurassic black shales have concluded that pterioids were probably truly benthic (Antia and Wood 1977; Kauffman 1981; Wignall 1987) with a few, specialized, pendent exceptions (text-fig. 5). The evidence for a pseudoplanktonic existence is slightly more compelling for these pterioids than for the Palaeozoic brachiopods and praecardioids. Many pterioids occur in a greater range of facies than just black shales, although they are nearly always in fine-grained facies. Also, examples of driftwood and other floating objects colonized by such bivalves are relatively common (e.g. Paul 1939; Hauff and Hauff 1981; Tanabe 1983), but these may only be facultative occurrences. However, it is their distribution within the sediment which provides the strongest evidence of a benthic lifestyle. Like the brachiopods discussed above, the pterioids tend to be abundant in thin horizons and absent in the intervening sediment, suggesting brief benthic opportunistic colonization. Thus, it appears that many of the reported occurrences of pseudoplankton may be more realistically interpreted as true benthos. This considerably increases the diversity of benthic life recorded from black shales whilst substantially reducing the diversity of the pseudoplanktonic record. Many of these occurrences owe their interpretation to the tacit assumption that black shale depositional environments are permanently anoxic, and thus fail to appreciate the highly dynamic nature of many such environments (Sageman et al. in press) where even transient improvements in benthic oxygen levels are rapidly exploited by benthic opportunists. THE PHANEROZOIC HISTORY OF PSEUDOPLANKTON Having re-interpreted a large number of ‘pseudoplanktonic’ occurrences as benthic, the remaining record is relatively sparse and weighted towards cementing forms which are the least likely to become detached from their float. The data have been divided into obligate and facultative pseudoplankton (table 1 and text-fig. 6), thus distinguishing between relatively rare, specialized forms specifically adapted to this lifestyle and the more common benthic forms which have occasionally exploited a chance encounter with a floating object. The potential floating attachment sites in the Cambrian were pumice and possibly algae, but no organisms have been inferred to have adopted a pseudoplanktonic lifestyle during this interval. The appearance of large cephalopods, particularly the orthoconic nautiloids, in the Ordovician was exploited by bryozoans and inarticulate brachiopods (text-fig. 6). Nearly all of these appear to have been host-specific suggesting that they were commensal forms and obligate pseudoplankton (Havhcek 1972; Lockley and Antia 1980; Baird et at. 1989). Drifting graptolites also appeared 372 PALAEONTOLOGY, VOLUME 33 OBLIGATE R Ng Hg K J T A V ro a; E — t= o o D O A, V A V FACULTATIVE TEXT-FIG. 6. Changes in the com- position of pseudoplankton through the Phanerozoic with obligate and facultative occurrences distinguished. Variation in attachment strategies is also illustrated. The majority of obligate forms appears to have been pendent. I ATTACHMENT STRATAGIES [M] Pendent nil Adpressed I I Cemented Bored at this time but are considered to be true plankton (Rickards 1975) and are thus beyond the scope of our study. Pseudoplanktonic diversity reached new heights in the Devonian with the first appearance of driftwood which was rapidly exploited by both crinoids (Wells 1941 ; McIntosh 1978) and bivalves (Nye et al. 1975). Curiously, the Carboniferous marked a decline in the fortunes of pseudoplankton despite the increase in driftwood in the world’s oceans. However, a number of lanceolate bivalves may have attached to driftwood (text-fig. 5). The main cause of the low diversity is the rarity of cephalopod encrustation. A review by Boston el al. (1988) revealed that less than 2% of shells were colonized by epizoans in the Carboniferous and some of these examples may have occurred on the sea floor. This low value was attributed to the success of the cephalopods at defending themselves against unwanted infestation either by physical or chemical means. But this apparent ‘success’, on the part of the cephalopods, may also be due to the rarity of large forms at this time as the dominant group, the goniatites, were typically small. Diversity of pseudoplankton appears to have declined to zero in the Permian. The thick sequence of organic-rich shales in the Upper Permian Phosphoria Formation of the United States would have provided good preservational conditions but no pseudoplanktonic forms have been recorded, despite detailed study (e.g. Yochelson 1963). The Triassic was marked by the appearance of several groups able to exploit a pseudoplanktonic WIGNALL AND SIMMS: PSEUDOPLANKTON 373 lifestyle. These include lepadomorph barnacles, cementing bivalves, the reappearance of the lanceolate bivalve morphotype in the bakevelliid lineage, and pseudoplanktonic crinoids (text-fig. 6). Pseudoplanktonic diversity rose to an all time high in the Lower Jurassic with crinoids, inoceramids and lepadomorphs occurring on driftwood (Moore 1867; Tanabe 1983; Simms 1986). Ammonites were infested by an equally diverse range of epizoans including bryozoans, inarticulate brachiopods, oysters and a range of byssate bivalves (Seilacher 1982a). A number of Jurassic articulate brachiopod genera also may have attached to floating algae (Ager 1962, 1965; Tchoumatchenco 1972), although, as in many supposed ancient algal colonies, associated fauna such as bryozoans curiously are absent. Towards the end of the Lower Jurassic the first wood-boring, teredinid bivalves appeared (Kelly 19886). However these Jurassic forms appear to have been exclusively benthic colonizers because, as Kelly (1988 a) noted for the Upper Jurassic, bored wood is only found in shallow marine arenaceous sediments. Driftwood from contemporary deeper water, muddier facies was not bored. Such a strong facies control upon boring could only occur if the bivalves colonized after the wood had reached the sea floor. During the Cretaceous the frequency of bored wood appears to have increased considerably in all facies, suggesting that the bivalves were able to settle on floating driftwood. This may have had serious consequences for other pseudoplanktonic species as the activity of boring bivalves greatly shortens the floating duration of driftwood. Faunal changes amongst driftwood faunas in the Upper Jurassic may be at least partially driven by the rise of wood-borers. Obligate, pendent, pseudoplanktonic crinoids and lanceolate bivalves both occur for the last time in the late Jurassic (text-fig. 6). The pseudoplanktonic driftwood colony recorded from the Upper Jurassic in text-figure 2 contains an unusual fauna dominated by encrusting forms (serpulids and oysters) in contrast to the adpressed and pendent strategies of earlier Jurassic driftwood colonies (Hauff and Hauflf 1981 ) and coeval colonies of Antarctica (Doyle and Whitham, in press). By the Cretaceous driftwood is rarely associated with any epizoans apart from boring bivalves (E. G. Kauffman, pers. comm.). The Cretaceous, like the Carboniferous, marked a low point in pseudoplanktonic diversity primarily due to the virtual absence of cephalopod epizoans, except for some examples from the Maastrichtian (Dunbar 1928; Riccardi 1980). This may have been due to a widespread ability amongst all cephalopod groups to defend themselves against colonization (cf. Boston et al. 1988). It also testifies to the likelihood that such infestation occurred during life, for nekroplanktonic infestation would produce a more uniform record through the Phanerozoic. However, preservational factors may also have a major influence on the record. Cretaceous and Carboniferous ammonoids are commonly collected as composite moulds produced by aragonite dissolution. Hence any encrusters may remain embedded in the matrix following collection of the ammonoid. A few examples of Lower Tertiary pseudoplankton are known (Davis and Elliot 1958; Lindqvist 1986). Associated with many logs of driftwood in the Eocene London Clay of south east England are crinoids (Isselicrinus subbasaltiformis) and pendent bivalves (Pteria papyracea) - an assemblage strongly reminiscent of Lower Jurassic pseudoplanktonic driftwood colonies, although the Eocene examples more probably colonized the wood after it sank to the sea floor (Taylor 1978). The Eocene also marks the first appearance of a genuine Sargasswn-Vike fish fauna (Jerzmanska and Kotlarczyk 1976), but without any associated invertebrate fauna. The Sargassum fauna constitutes the most diverse pseudoplanktonic community in the modern oceans with over 70 species recorded (Fine 1970; Morris and Mogelberg 1973). The majority are small forms, a strategy to reduce weight and the possibility of overloading the brown algae to which they were attached. The bryozoan Membranipora is the dominant form and, along with the co- occurring annelid, Spirorbis, and the gastropod Litiopa, has a fairly high preservation potential. Thus the absence of a pre-Eocene occurrence of these faunas strongly suggests that Sargassum is a relatively recent innovation of the Cenozoic. Diverse pseudoplanktonic communities are also known from recent studies of Nautilus (Landman el al. 1987) and turtles (Frazier et al. 1984). Indeed, pseudoplankton is probably more abundant today than at any time in the past due to the large amounts of man-made flotsam found in the oceans, such as hollow plastic, glass and metal 374 PALAEONTOLOGY, VOLUME 33 containers. In particular, the use of expanded polymers has created flotsam of very extended floating duration thus increasing the time available for colonization. Boats and ships theoretically provide even more ideal attachment sites since, not only do they have very extended floating durations, they are able deliberately to avoid being cast ashore, thus eliminating one of the major hazards of the pseudoplanktonic lifestyle. Consequently, a wide range of organisms attach to ships (Carlton 1985), much to the chagrin of their owners. Lepadomorph barnacles appear, from personal observations, to be the main group exploiting this new diversity of attachment sites. Acknowledgements. We thank John Hudson and Pete Doyle for their comments on an earlier version of the manuscript. We also thank Gordon Baird, Pete Doyle and Andy Whitham for allowing us to see preprints of their work. Erie Kauffman and Brad Sageman provided useful discussions. This work was undertaken whilst both of us were funded by NERC Postdoctoral Fellowships for which we are grateful. REFERENCES ACER, D. V. 1962. The occurrence of pedunculate brachiopods in soft sediments. Geological Magazine, 99, 184-186. 1965. The adaptations of Mesozoic brachiopods to different environments. Palaeogeography, Palaeo- climatology, Palaeoecology 1, 143-172. ANTiA, D. D. J. and WOOD, B. A. 1977. Some aspects of the palaeoecology of the Elb zone of the Edale Shales (Namurian) in the Edale Valley, Derbyshire. Mercian Geologist, 6, 179-196. BAIRD, G. c., BRETT, c. E. and FREY, R. c. 1989. “Hitchhiking” epizoans on orthoconic cephalopods; preliminary review of the evidence and its implications. Senckenhergiana Lethaea, 69, 439^46. BARRON, L. s. and ETTENSOHN, F. R. 1981. Paleoecology of the Devonian-Mississippian black shale sequence in eastern Kentucky, /n Energy resources of Devonian-Mississippian shales of eastern Kentucky, 9th Annual Field Conference of the Geological Society of Kentucky. BERGSTROM, J. 1968. Some Ordovician and Silurian brachiopod assemblages. Lethaia, 1, 230-237. BOSTON, w. B., MCCOMAS, G. A. and MCGHEE, G. R. 1988. The fossil occurrence of epizoans on living coiled cephalopods in the Upper Paleozoic (Carboniferous). Geological Society of America Abstracts with Programs, 20/1, 9. BOTTJER, D. J. 1981. Paleoecological implications of antifouling adaptations in shelled molluscs. Geological Society of America Abstracts with Programs 13, 413-414. BRASiER, M. D. 1975. An outline history of seagrass communities. Palaeontology, 18, 681-702. BULMAN, o. M. B. 1964. Lower Palaeozoic plankton. Quarterly Journal of the Geological Society of London, 120, 455-476. CAMPBELL, H. J. 1985. The stratigraphic significance of the Triassic bivalves Daonella and Halobia in New Zealand and New Caledonia. In Hornibrook Symposium extended abstracts. New Zealand Geological Survey record, 9, 21-23. CARLTON, J. T. 1985. Transoceanic and interoceanic dispersal of coastal marine organisms: the biology of ballast water. Oceanography and Marine Biology Annual Reviews, 23, 313-372. CHAMBERLAIN, L A., WARD, p. D. and WEAVER, J. s. 1981. Post-mortem ascent of Nautilus shells: implications for cephalopod paleobiology. Paleobiology, 7, 494-509. CLAUSEN, c. K. and wiGNALL, p. B. 1990. Early Kimmeridgian bivalves of southern England. Mesozoic Research, 3. COPE, J. c. w. 1968. Epizoic oysters on Kimmeridgian ammonites. Palaeontology, 11, 19-20. COWEN, R. 1981. Crinoid arms and banana plantations: an economic harvesting analogy. Paleobiology, 1, 332-343. CRISP, D. J. 1979. Dispersal and re-aggregation in sessile marine invertebrates, particularly barnacles. Pp. 319-327 In larwood, g. p. and rosen, b. r. (eds.). Biology and systematics of colonial organisms, Systematics Association Special Volume 11. DAVIS, A. G. and elliot, g. f. 1958. The palaeogeography of the London Clay. Proceedings of the Geologists' Association 68, 255-277. DOYLE, p. D. and WHITHAM, A. in press. The Nordenskjold Formation: evidence for Late Jurassic anoxia in the Antarctic Peninsula. In tyson, r. v. and pearson, t. r. (eds.). Continental Shelf anoxia. Geological Society of London Special Publication. WIGNALL AND SIMMS: PSEUDOPLANKTON 375 DUNBAR, c. o. 1928. On an ammonite shell investing commensal bryozoa. American Journal of Science 16, 164-165. EKMAN, s. 1953. Zoogeography of the sea, Sidgwick and Jackson, London. FINE, M. L. 1970. Faunal variation on pelagic Sargassum. Marine Biology 7, 112-122. FRAZIER, J., MARGARITOULIS, D., MULDOON, K., POTTER, C. W., ROSEWATER, J., RUCKDESCHEL, C. and SALAS, S. 1984. Epizoan communities on marine turtles. I. Bivalve and gastropod molluscs. Marine Ecology 6, 127-140. FRIEDRICH, H. 1969. Marine biology. An introduction to its problems and results. Sidgwick and Jackson, London. FURSiCH, F. T. and WENDT, J. 1977. Biostratinomy and palaeoecology of the Cassian Formation (Triassic) of the southern Alps. Palaeogeography, Palaeoclimatology, Palaeoecology, 22, 257-323. GROSBERG, R. K. and QUINN, J. F. 1986. The genetic control and consequences of kin recognition by the larvae of a colonial marine invertebrate. Nature, 322, 456-459. HALLAM, A. 1973. Distributional patterns in contemporary terrestrial and marine animals. In hughes, n. f. (ed.). Organisms and continents through time. Special Papers in Palaeontology 12, 93-106. 1976. Stratigraphic distribution and ecology of European Jurassic bivalves. Lethaia, 9, 245-259. 1986. The Pliensbachian and Tithonian extinction events. Nature, 319, 765-768. HAUFF, B. and HAUFF, R. B. 1981. Das Holzmadenbuch. Felbach. HAYAMi, I. 1969. Notes on Mesozoic “planktonic” bivalves. Journal of the Geological Society of Japan, 75, 375-385. [Japanese with English abstract]. HAVLiCEK, v. 1967. Brachiopoda of the suborder Strophomenida in Czechoslovakia. Rozprawy iistred Ustredni Geologii, 33, 1-235. 1972. Life habit of some Ordovician inarticulate brachiopods. Vestnik iistred Ustredni Geologii, 47, 229-233. HEPTONSTALL, w. B. 1970. Buoyancy control in ammonoids. Lethaia, 3, 317-328. HERRING, p. J. 1969. Pigmentation and carotenoid metabolism of the marine isopod Idotea metallica. Journal of the Marine Biological Association of the United Kingdom, 49, 766-779. HOLMES, R. w. 1985. The morphology of diatoms epizoic on cetaceans and their transfer from Cocconeis to new genera, Bennettella and Epipellis. British Phycological Journal, 20, 43-57. HUDSON, R. G. s. and COTTON, G. 1943. The Namurian of Alport Dale. Proceedings of the Yorkshire Geological Society, 25, 142-173. ICHIKAWA, K. 1958. Zur Taxonomie und Phylogenie der triadischen "Pteriidae” (Lamellibranch.) mit besonder Beriicksichtigung der Gattungen Claraia, Eumorphotis, Oxvtoma und Monotis. Palaeontographica All 1/5, 131-212. JABLONSKi, D. and LUTZ, R. A. 1983. Larval ecology of marine benthic invertebrates: paleobiological implications. Biological Reviews, 58, 21-89. JEFFERIES, R. P. S. and MINTON, p. 1965. The mode of life of two Jurassic species of Posidonia" . Palaeontology, 8, 156-185. JERZMANSKA, A. and KOTLARCZYK, J. 1976. The beginnings of the Sargasso assemblage in the Tethys? Palaeogeography, Palaeoclimatology, Palaeoecology, 20, 297-306. JOKiEL, p. L. 1984. Long distance dispersal of reef corals by rafting. Coral Reefs, 3, 113-116. KAUFFMAN, E.G. 1981. Ecological reappraisal of the German Posidonienschiefer (Toarcian) and the stagnant basin model. Pp. 311-381 In gray, j., boucot, a. j. and berry, w. b. n. (eds.). Communities of the past, Hutchinson Ross Stroudsberg, Pennsylvania. KELLY, s. R. A. 1988fl. TuniusI davidsoni (de Loriol), the earliest British pholadid wood-boring bivalve, from the Late Jurassic of Oxfordshire. Proceedings of the Geologists' Association, 99, 43^8. 19886. Cretaceous wood-boring bivalves from western Antarctica with a review of the Mesozoic Pholadidae. Palaeontology, 31, 341-372. KLiKUSHiN, v. G. 1982. Taxonomic survey of fossil isocrinids with a list of the species found in the USSR. Geobios, 15, 299-325. KREBS, w. 1979. Devonian basinal facies. In house, m. r., scrutton, c. t. and bassett, m. g. (eds.). The Devonian System, 125-139. Special Papers in Palaeontology 23, 125-139. KRiz, J. 1969. Genus Butovicella Khz 1965 in the Silurian of Bohemia (Bivalvia). Sbormk Geologickych Ved. Paleontologie. Rada p. sv 10, 105-139. LANDMAN, N. H., SAUNDERS, w. B., WINSTON, J. E. and HARRIES, p. J. 1987. Incidence and kinds of epizoans on the shells of live Nautilus. Pp. 163-177 In saunders, w. b. and landman, n. h. (ed.). Nautilus, Plenum Publishing Corporation, New York. 376 PALAEONTOLOGY, VOLUME 33 LiNDQViST, j. K. 1986. Tcredinid-bored Araucariaceae logs preserved in shoreface sediments, Wangaloa Formation (Paleocene), Otago, New Zealand. New Zealand Journal of Geology and Geophysics, 29, 253-261 . LOCKLEY, M. G. and ANTiA, D. D. J. 1980. Anomalous occurrences of the Lower Palaeozoic brachiopod Schizocrania. Palaeontology, 23, 707-713. MCINTOSH, G. c. 1978. Pseudoplanktonic crinoid colonies attached to Upper Devonian logs. Geological Society of America Abstracts with Programs, 10/7, 453. MCNAMARA, K. J. 1986. A guide to the nomenclature of heterochrony. Journal of Paleontology, 60, 4-13. MEISCHNER, D. 1968. Pemiciosc epokie von Placunopsis auf Ceratites. Lethaia, 1, 156-174. MERK.T, J. 1966. Uber Austen und Serpeln als Epoken auf Ammonitengehausen. Neues Jahrbuch fiir Geologic, Paldontologie, Abhandhmgen, 125, 467—479. MOORE, c. 1867. On the Middle and Upper Lias of the south-west of England. Proceedings of the Somerset Archaeological and Natural History Society, 13, 119-219. MORRIS, B. F. and mogelberg, d. d. 1973. Identification manual to the pelagic Sargassum fauna. Bermuda Biological Station Special Publication, 11. NYE, o. B. J., BROWER, J. c. and WILSON, s. E. 1975. Hitchhiking clams in the Marcellus Sea. Bulletin of American Paleontology 67, 287-298. PALMER, C. P. 1987. The Kimmeridgian fauna associated with the Portland plesiosaur. Proceedings of the Dorset Natural History and Archaeological Society, 109, 109-1 12. PAUL, H. 1939. Eine festgeheftete Muschel des Kulm-Meeres, Posidonia becheri. Natur und Volk. 69, 402-406. PESTANA, H. 1985. Carbonate sediment production by Sargassum epibionts. Journal of Sedimentary Petrology, 55, 184-186. POJETA, J., KRiz, J. and berdan, l m. 1976. Silurian-Devonian pelecypods and Paleozoic stratigraphy of subsurface rocks in Florida and Georgia and related Silurian pelecypods from Bolivia and Turkey. United States Geological Survey Professional Paper, 879. RICCARDI, A. c. 1980. Presencia de epizoos en un amonoideo heteromorfo del Maastrichtiano de Argentina: una inferencia paleoautoecologica. Ameghiniana, 57, 11-14. RICKARDS, R. B. 1975. Palaeoecology of the Graptolithina, an extinct class of the phylum Hemichordata. Biological Reviews, 50, 397-436. RiEBER, H. 1982. The formation of the bituminous layers of the middle Triassic of Ticino (Switzerland). P. 530 In EINSELE, G. and seilacher, a. (eds.) Cyclic and event stratification, Springer-Verlag, Berlin. [Abstract]. RUEDEMANN, R. 1934. Palacozoic plankton of North America. Geologiccd Society of America Memoir, 2. SAGEMAN, B., wiGNALL, p. B. and KAUFFMAN, E. G. in prcss. Biofacics models for black shales: tool for paleoenviromental analysis. In ricken, w., einsele, g. and seilacher, a. (eds.). Cycles and events in stratigraphy. Springer-Verlag, Berlin. SCHAFER, w. 1972. Ecology and palaeoecology of marine environments. Oliver and Boyd, Edinburgh. SCHOENER, A. and ROWE, G. T. 1970. Pelagic Sargassum and its presence among deep-sea benthos. Deep-Sea Research, 17, 923-925. SCHUMANN, D. 1988. Environment and post-mortem history of Upper Jurassic ammonites in Nuevo Leon, NE Mexico. Pp. 731-736 In wiedmann, j. and kullman, j. (eds.). Cephalopods - past and present, Stuttgart. SEILACHER, A. 1960. Epizoans as a key to ammonoid ecology. Journal of Paleontology, 34, 189-193. 1968. Swimming habits of belemnites - recorded by boring barnacles. Palaeogeography, Palaeo- climatology, Palaeoecology, 4, 279-285. 1982fl. Ammonites as habitats in the Posidonia shales of Holzmaden - floats or benthic islands? Neues Jahrbuch fiir Geologic, Paldontologie, Monatsheft, 98-114. 19826. Posidonia shales (Toarcian, S. Germany) - stagnant basin model revalidated. Pp. 25-55 In GALLITELLI, E. M. (ed.). Palaeontology, essential of historical geology, STEM Mucchi, Modena Press. 1984. Constructional morphology of bivalves: evolutionary pathways in primary versus secondary soft- bottom dwellers. Palaeontology, 27, 207-238. DROZDZEWSKi, G. and HAUDE, R. 1968. Form and function of the stem in a pseudoplanktonic crinoid (Seirocrinus). Palaeontology, 11, 275-282. SHEEHAN, p. M. 1977. Ordovician and Silurian brachiopods from graptolitic shales and related deep-water argillaceous rocks. Lethaia, 10, 201-203. SIMMS, M. J. 1986. Contrasting lifestyles in Lower Jurassic crinoids: a comparison of benthic and pseudopelagic Isocrinida. Palaeontology, 29, 475^93. 1988a. Patterns of evolution among Lower Jurassic crinoids. Historical Biology, 1, 17-44. 19886. An intact comatulid crinoid from the Toarcian of southern Germany. Stuttgarter Beitrage zur Naturkunde, Serie B {Geologic und Paldontologie), 140, 1-7. WIGNALL AND SIMMS: PSEUDOPLANKTON 377 1990. British Lower Jurassic Crinoids. Palaeontographical Society Monograph, 1-103, 15 pis, STANLEY, G. D. and KANiE, Y. 1985. The first Mesozoic chondrophorine (medusoid hydrozoan) from the Lower Cretaceous of Japan. Palaeontology, 28, 101-109. TANABE, K. 1983. Mode of life of an inoceramid bivalve from the Lower Jurassic of West Japan. Neiies Jahrhucli fur Geologie, unci Palaontologie Monatsheft, 419^28. TAYLOR, J. D. 1978. Cenozoic. Pp. 323-351 In mckerrow, w. s. (ed.). The ecology of fossils, Duckworth, London. TCHOUMATCHENCO, p. 1972. Thanatocoenoses and biotopes of Lower Jurassic brachiopods in central and western Bulgaria. Palaeogeography, Palaeoclimatology, Palaeoecology, 12, 227-242. THAYER, c. w. 1974. Marine paleoecology in the Upper Devonian of New York. Lethaia, 1, 121-155. THOMPSON, j. B. and NEWTON, c. R. 1987. Ecological reinterpretation of the dysaerobic Leiorhynchus fauna: Upper Devonian Geneseo black shale, central New York. Palaios, 1, 274—281. TURNER, R. D. 1973. Wood-boring bivalves, opportunistic species in the deep sea. Science, 180, 1377- 1379. TYSON, R. v. 1987. The genesis and palynofacies characteristics of marine petroleum source rocks. In brooks, J. R. V. and FLEET, A. J. (eds.). Marine petroleum source rocks. Special Publication of the Geological Society of London 24, 47-68. WATKINS, R. 1978. Bivalve ecology in a Silurian shelf environment. Lethaia, 11, 41-56. and BERRY, w. B. N. 1977. Ecology of a Late Silurian fauna of graptolites and associated organisms. Lethaia, 10, 267-286. WEBSTER, s. K. 1975. Oxygen consumption in echinoderms from several geographical locations, with particular reference to the Echinoidea. Biological Bulletin, 148, 157 164. WELLS, J, w. 1941. Crinoids and Callixylon. American Journal of Science, 239, 454—456. WHITHAM, A. G. and SPARKS, R. s. J. 1986. Pumice. Bulletin of Volcanology, 48, 209-223. WIGNALL, p. B. 1987. A biofacies analysis of the Gastrioceras cumbriense marine band (Namurian) of the central Pennines. Proceedings of the Yorkshire Geological Society, 46, 111-121. and MYERS, K. j. 1988. Interpreting benthic oxygen levels: a new approach. Geology, 16, 452-455. WILLIAMS, L. A. 1983. Deposition of the Bear Gulch Limestone: a Carboniferous plattenkalk from central Montana. Sedimentology, 30, 843-860. WITHERS, T. H. 1928. Catalogue of fossil Cirripedia in the Department of Geology. Vol. I. Triassic and Jurassic. British Museum (Natural History), Oxford University Press. YOCHELSON, E. L. 1963. Palcoecology of the Permian Phosphoria Formation and related rocks. United States Geological Survey Professional Paper, 475-B, 123-124. STURMER, w. and STANLEY, G. D. JR 1983. PlectocUscus discoideus (RaulT): a redescription of a chondrophorine from the early Devonian Hunsriick Slate, West Germany. Paldontologisch Zeitschrift, 57, 39-68. PAUL B. WIGNALL Department of Earth Sciences, University of Leeds, Leeds Typescript received 16 March 1989 Revised typescript received 24 July 1989 MICHAEL J. SIMMS Department of Geology, Trinity College, Dublin 2 il CENOMANIAN MICROMORPHIC AMMONITES FROM THE WESTERN INTERIOR OF THE USA by W. J. KENNEDY CUld W. A. COBBAN Abstract - Calcareous concretions from middle and upper Cenomanian (Cretaceous) shale sequences in Montana and Wyoming yield, on rare occasions, abundant minute ammonites. Some are juveniles of large species, and give valuable evidence on early ontogeny and evolutionary affinities of these taxa. They provide evidence for the probable evolutionary origins of a series of hitherto undescribed progenic dwarf genera, adult at 4-5-1 6'5 mm diameter that are a remarkable feature of these assemblages. Five new genera, Kastanoceras, AlzacJites, Microsiilcatoceras, Cryptometoicoceras and Biiccinammonites are introduced. 33 species are described from Montana, Wyoming, Utah, New Mexico and Texas; 16 are new, the others were previously unknown in the region or show new details of early ontogeny. New taxa are: Moremanoceras montanaense sp. nov.; Cunningtoniceras sp. juv.; Tarrcmtoceras exile sp. nov.; Kastanoceras spiniger gen. et sp. nov.; Alzudites alzadensis gen. et sp. nov.; A. westonensis gen. et sp. nov.; A. incomptus gen. et sp. nov.; Alzaditesi sp. ; Alzadites sp. A; Microsidcatoceras puzosiiforme gen. et sp. nov. ; M. crassiim gen. et sp. nov. ; M. texcmwn gen. et sp. nov.; Microsidcatoceras sp.V, Cryptometoicoceras mite gen. et sp. nov.; Ncumometoicoceras nemos sp. nov. ; Nannometoicocerasl gkiber sp. nov. ; Biiccinammonites minimus gen. et sp. nov. ; Idiohamites pidchelhis sp. nov.; I. hispinosus sp. nov.; Carthaginites aquilonius sp. nov.; and Scapliites (Scap/iites) sp. Over wide areas of the Western Interior of the United States (text-fig. 1), marine rocks of middle to late Cenomanian age are partly or wholly in a non-calcareous shale facies. Most of the fauna known from this facies comes from calcareous concretions of early diagenetic origin, although in some units, crushed moulds can be obtained by splitting shales. At many levels concretions are only sparingly fossiliferous, and even then yield only large fossils. In the present communication we describe some remarkable faunas collected over the past 60 years from concretions in the Middle to Upper Cenomanian part of the Belle Fourche Shale of the Black Hills in Wyoming and Montana. In this area we estimate that less than 1 % of concretions are fossiliferous, and only six out of thousands examined preserve minute ammonites in abundance (text-fig. 2) although others preserve larger fossils, including both macro- and microconch ammonites. This preservation is all the more remarkable, since adjacent concretions at the same stratigraphic horizon at the same locality lack such assemblages. Occurrence is not simply a matter of concretions preserving a particular level of fossil concentration, the occurrences are areally limited. The palaeogeographical setting of the Black Hills area during late Cenomanian time places it far from shore, and we find it difficult to interpret these fossil occurrences as current accumulations, especially as some ammonites preserve delicate features of ornament, and occur scattered throughout the concretions rather than concentrated on a single plane. It is also difficult to accept the occurrences as faecal concentrations for the fossils are embedded in sediment matrix, and are in this respect unlike the great mass occurrences of ammonites in the Mowry Shale (Reeside and Cobban 1954, 1960), where the fossils are bound in a matrix of fish scales and debris. The occurrences provide a unique opportunity to study elements of the late Cretaceous ammonite fauna of the Western Interior that were previously unrecognized, especially a series of micromorphs, ammonites that are adults at 4-5 to 16-5 mm diameter. Five new micromorph genera and 12 new species are described, including several additional micromorphs from areas in New Mexico, Utah and northeast Texas. Most have adult phragmocone whorls that share common features with the innermost ornamented phragmocone whorls of co-occurring ‘normal’ size dimorphic ammonites. (Palaeontology, Vol. 33, Part 2, 1990, pp. 379-422, 7 pls.| © The Palaeontological Association 380 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 1. Map of outcrop areas of Cretaceous sedimentary rocks in the Western Interior of the United States. The heavy black bounding lines show the maximum known extent of the mid-Cretaceous seaway. Modified from Cobban and Reeside, 1952, fig. 1. We interpret the new genera as progenic dwarfs that evolved through precocious sexual maturation, an evolutionary process previously invoked to explain the origin of certain other Upper Cretaceous micromorphs, notably Protacanthoceras Spath, 1923 from Acanthoceras Neumayr, 1875 (Wright and Kennedy 1980, 1987; Kennedy and Wright 1985); Nawwmetoicoceras Kennedy, 1988, from Metoicoceras Hyatt, 1903 (Kennedy 1988, p. 63); Plesiacanthoceratoides Kennedy and Cobban, 1990, from the Western Interior acanthoceratine lineage. These micromorphs are not, it must be stressed, juveniles of ‘normal’ sized taxa. They show all the features of maturity common to ammonites, including septal crowding, modification of ornament on the body chamber, and development of distinctive apertural processes. Several of the taxa are monotypic, or represented by few specimens. We justify naming them because they are so distinctive and utterly different in most cases from all previously known taxa. We also hope that their description will stimulate other workers to look carefully for such micromorphs amongst apparent juveniles in their own collections, since we consider it unlikely that the Western Interior occurrences are unique to that region. What the life habits of these micromorphs may have been is a matter of speculation. Their concentrated occurrence suggests to us that some at least may have lived close to the bottom where they were preserved, because we cannot easily accept or see evidence for any physical or biological process that led to their concentration. Micromorphs apart, the concretion faunas studied include abundant juvenile individuals that show for the first time the early ontogenetic development of several genera and species, clarifying their affinities and also pointing to the possible ancestors of the progenic dwarfs. Also present are a series of taxa that are either new, or not previously recorded from the area, including first records of Sumitomoceras in the region, and the first Scaphites (Scaphites) from the Cenomanian of the Western Interior. In all, 33 species are documented. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 381 TEXT-FIG. 2. Mass occurrence of small ammonites in concretions, a, b, USNM 423785 and 423784 from USGS Mesozoic locality 22871, Plesiaccmthoceras wyoiningense zone, with Borissiakoceras and hcteromorphs. c, USNM 423665, from USGS Mesozoic locality 528, Frontier Formation, Middle Cenomanian, F5 km (1 mile) northeast of Wilcox, Albany County, Wyoming, with Johnsonites sulcatus Cobban, 1961 . Collected by Alpheus Hyatt and I. C. Russel in 1888. BIOSTRATIGRAPHY The standard ammonite zonation for the Middle and Upper Cenomanian of the Western Interior is shown in text-hg. 3, and is modified after the work of Cobban ( 1 984, 1 9%la). These are assemblage zones, the bases of the zones marked by the first appearance of a variety of ammonites, including the index species, which is usually, but not invariably limited to its zone. Because of the limited 382 PALAEONTOLOGY, VOLUME 33 CRETACEOUS STAGE AMMONOID ZONE USGS MESOZOIC LOCALITY Nigericeras scotti Neocardioceras juddii upper Vascoceras cauvini Sciponoceras gracile — 23042, D12052 (T3 Metoicoceras moshyense — D8314 a ^ 12650, 12740 c Dunveganoceras pondi — D4462, D4466, 2 D5947 s E Plesiacanthoceras wyomingense— 22871 o c Acanthoceras amphiholum o middle Acanthoceras hellense Acanthoceras muldoonense Acanthoceras granerosense Conlinoceras tarrantense TEXT-FIG. 3. Middle and Upper Cenomanian ammonite zones of the US Western Interior, with the levels of some of the more important collections indicated. vertical and horizontal distributions of ammonites within sedimentary sequences which result from both sedimentary and biological controls, these zones are of different scales and may be recognized over quite limited areas (as in the case of the Conlinoceras tarrantense to Acanthoceras bellense zones), while others can be recognized throughout the Western Interior (e.g. the Acanthoceras amphiholum zone). The oldest fauna described here comes from the Plesiacanthoceras wyoniingense zone, which is placed at the top of the Middle Cenomanian. The Dimveganoceras pondi zone has Calycoceras (Proeucalycoceras) canitaurinum Haas, 1949 as an alternative index species in the southern part of the Interior, where D. pondi is absent. Their contemporaneity is established by the co-occurrence of the two index species at their common type locality near Greybull, Wyoming as well as common occurrence of other species. The Metoicoceras moshyense zone is represented by a great thickness of sediment in the northern Western Interior, and may represent a greater time interval than the other Upper Cenomanian zones, although not susceptible to finer division at present. Sciponoceras gracile is retained as a zonal index because of long and widespread usage and because it is by far the commonest ammonite at that level although it ranges up to the Neocardioceras juddii zone. Metoicoceras geslinianum (d’Orbigny, 1850), M. whitei Hyatt, 1903 (a synonym) and Eiiomphaloceras [Kanahiceras] septemseriatum (Cragin, 1893) have also been used as indices for this zone in recent years. Suggestions that the S. gracile zone can be divided into a lower subzone of Vascoceras diartianian (d’Orbigny, 1850) and an upper subzone of E. septemseriatum (Cobban 1984, p. 81) are here abandoned; V. diartianian occurs below the base of the gracile zone in association with Eiiomphaloceras euomphalum (Sharpe, 1855), Eiicalycoceras pentagonum (Jukes-Browne, 1896) and other ammonites in southwestern New Mexico and the Black Hills area at the top of the M. moshyense zone. Vascoceras cauvini Chudeau, 1909, was proposed as a provisional index for a distinctive and as yet undescribed fauna between the S. gracile and N. juddii zones known only from southwest New Mexico. Subsequent work shows V. cauvini to range down into the correlatives of the gracile zone in Israel (Lewy, Kennedy and Chancellor 1984), and up into the Neocardioceras juddii zone at Chispa Summit in Trans-Pecos Texas, and a replacement index for the zone is needed from among the, at present, undescribed Euomphaloceratinae present in the assemblage. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 383 The Neocardioceras juddii zone can be recognized in Trans-Pecos Texas, New Mexico, Colorado, Arizona, Utah, Wyoming and Montana. Apart from the index species (which is the last of a lineage of Neocardioceras that extends down to the D. pondi zone), Pseiidaspidoceras pseudojwdosoides (Choffat, 1898) is locally common in Trans-Pecos Texas and New Mexico and there is also a diverse, but as yet undescribed, vascoceratid and pseudotissotid fauna in these two states. Suggestions that Gauthicericeras aff. hravaisi (d’Orbigny) of Moreman (1927, p. 96, pi. 4, fig. 2) was an N. juddii (Wright and Kennedy 1981, p. 50) and indicated the presence of \\\q juddii zone in northeast Texas must be discounted; subsequent work shows Moreman’s form to be lower Turonian Watinoceras (Kennedy 1988, p. 50). The highest Cenomanian zone recognized here has Nigericeras scotti Cobban, 1972, as index species, the index species itself being limited to southeast Colorado and northeastern and southwestern New Mexico. LOCALITY DETAILS The more important localities of ammonites described below are shown on text-fig. 4 and their stratigraphic positions are shown in text-fig. 3. Where localities have yielded only limited numbers of specimens, details are given at the appropriate point in the text. Nine localities yielded large assemblages, and in the interests of brevity, full details and faunal lists are given here. uses Mesozoic locality 12650. Collected by W. W. Rubey and others, 1924. Sec. 7, T. 48 N., R. 65 W., 3 2 km southeast of Thornton, Weston County, Wyoming. Belle Fourche Shale, 18-3 m (60 feet) beneath highest yellow concretion. Upper Cenomanian Metoicoceras mosbyense zone. Ammonite fauna is: Borissiakoceras sp. juv., Moremanoceras costatum Cobban, Hook and Kennedy, 1989 (common), Cwmingtoniceras sp. juv., Tarrantoceras exile sp. nov., Metoicoceras cf mosbyense Cobban, 1953 (juveniles), Carthaginites aquilonius sp. nov. uses Mesozoic locality 12740. Collected by M. N. Bramlettc and W. W. Rubey, 1924. E| sec. 6, T. 9 S., R. 59 E., Carter County, Montana. Belle Fourche Shale. Upper Cenomanian Metoicoceras mosbyense zone. Ammonite fauna is: Moremanoceras costatum (common), Neocardioceras sp. nov., Nannometoicoceras nanos gen. et sp. nov., Metoicoceras cf. mosbyense (juveniles). uses Mesozoic locality 23062. Collected by J. B. Reeside Jr., W. A. Cobban and H. R. Christner. 0-8 km east of Five Mile Creek, in the SE| Sec. 25, T. 9 S., R. 60 E., Carter County, Montana. Greenhorn Formation. Upper Cenomanian Sciponoceras gracile zone. The ammonite fauna is : Borissiakoceras sp., Sumitomoceras sp. juv., Kanahiceras septemseriatum (Cragin, 1893), Buccinammonites minimus gen. et sp. nov., Microsulcatoceras puzosiiforme gen. et sp. nov., Metoicoceras geslinianum (d’Orbigny, 1850), Cryptometoicoceras mite gen. et sp. nov., Allocrioceras annulatum (Shumard, 1860), Sciponoceras gracile (Shumard, 1860) (common), Yezoites delicatulus (Warren, 1930). uses Mesozoic locality 22871. Collected by W. A. Cobban, 1947. 9-7 km northwest of Alzada in SEj Sec. 6, T. 9 S., R. 59 E., Carter County, Montana. Middle Cenomanian Plesiacanthoceras wyomingense zone. Borissiakoceras orhiculatum Stephenson, 1955 (common), Moremanoceras straini Kennedy, Cobban and Hook, 1988, M. montanense sp. nov., Plesiacanthoceratoides alzadense (Cobban, 19876), Plesiacanthoceras wyomingense (Reagan, 1924), Tarrantoceras cuspidum Stephenson, 1953, Tarrantoceras sp., Kastanoceras spiniger gen. et sp. nov., Alzadites alzadensis gen. et sp. nov., acanthoceratinae indet., Hamites cimarronensis (Kaufmann and Powell, 1977) (common), Idiohamites pulchellus sp. nov., I. bispinosus sp. nov., Anaptychus sp. uses Mesozoic locality D4462. Collected by W. A. Cobban, 1964. NWj NE^NEi Sec. 24, T. 47 N., R. 65 W., Weston County, Wyoming. Belle Fourche Shale. Upper Cenomanian Dunveganoceras pondi zone. Ammonite fauna is: Borissiakoceras orhiculatum. Metoicoceras aff. praecox Haas, 1949, Cryptometoicoceras mite gen. et sp. nov., Hamites cimarronensis. Idiohamites bispinosus. uses Mesozoic locality D4466. Collected by W. A. Cobban, 1964. 4-8 km NW of Alzada in NEj NWj Sec. 14, T. 9 S., R. 59 E., Carter County, Montana. Belle Fourche Shale, near bentonite G of Knechtel and Patterson (1962). Upper Cenomanian, Dunveganoceras pondi zone. Ammonite fauna is: Borissiakoceras sp.. 384 115° PALAEONTOLOGY, VOLUME 33 110° 105° 100° 45‘ 40° 35° ' 30° - 95° CANADA UNITED STATES .Great Falls 7 ) .Butte Helena Montana — r Grand Forks^ North Dakota Bismarck i Minnesota .Billings i 12740 I 22871 • Cody Mineapolis.. . St. Paul Idaho 23062 South Dakota D4466 Pierre 08315^^^^ ^Rapid City * Wyoming d44'6^^i265o D5947I S '"sl^Sioux City I j-ander ^Casper .Rawlins ' ' * I I Cheyejine Nebraska Salt Lake City Utah n Line Denver Colorado Pueblo Iowa icoln 1 • — TopekS^ Kansas K— ■ Dodge City I^Bolse City 1 Arizona .Phoenix Tucson > MEXICO I .Gallup .Santa Fe I .Albuquerque !xD12052 j New Mexico Roswell i^EI Paso X V. Amarillo Oklahoma Oklahoma City V Fort Worth. Texas Dallas .Austin _l_l_ 0 100 200 miles H 0 100 200 300 kilometres TEXT-FIG. 4. Locality map for some of the more important localities mentioned in the text. Morematweeras coslatum, Tarrantoceras cuspidum, Tarrantoceras sp., Dimveganoceras pondi Haas, 1949, Handles cimarronensis, Idiohamites hispinosus sp. nov. uses Mesozoic locality D5947. Collected by W. A. Cobban, 1961. 4-8 km south of Upton in NWj sec. 14, T. 47 N., R. 65 W., Weston County, Wyoming. Belle Fourche Shale, 19-8 m (65 feet) above 0-6 m (2 feet) bentonite. Upper Cenomanian, Dimveganoceras pondi zone. Ammonite fauna is: Borissiakoceras cf. orbicidatuny Morematweeras coslatum, Tarrantoceras sp., Neocardioceras sp. nov., Plesiacanthoceras cf. hellsanum (Stephenson, 1953), P. cf. wyomingense, Alzadites westonense gen. et sp. nov., Alzaditesl sp., Metoicoceras sp. A; Hamites cimarronensis, Idiohamites hispinosus sp. nov. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 385 uses Mesozoic locality D83I4. Collected by E. A. Merewether, 1971. SEj NEj sec. 33, T. 50 N., R. 66 W., Crook County, Wyoming. Belle Fourche Shale, from ferruginous concretions 7-6“10-7 m (25“35 feet) below top. Upper Cenomanian, Metoicoceras mosbyeme zone. Ammonite fauna is: Borissiakoceras sp. juv., Moremanoceras scotti, Euomphaloceras merewetheri Cobban, Hook and Kennedy, 1989, Hamites salehrosus Cobban, Hook and Kennedy, 1989, Metaptychoceras sp. uses Mesozoic locality D12Q52. Collected by S. C. Hook, O. J. Anderson and W. A. Cobban, 1982. Near mouth of Yellowrock Canyon in NW| SW| sec. 1 7, T. 7 N., R. 20 W., Cibola County, New Mexico. Rio Salado Tongue of Mancos Shale, from white concretions 4-6 m (15 feet) above base. Upper Cenomanian Sciponoceras gracile zone. Ammonite fauna is: Pseudocalycoceras angolaense (Spath, 1931 ), Euomphaloceras septemseriatuny Sumitomoceras sp. juv., Metoicoceras gesliuianum, Nannometoicocerasl glabrum gen. et sp. nov., Allocrioceras anmilatum (Shumard, 1860), Sciponoceras gracile, Worthoceras vermiculus (Shumard, 1860). REPOSITORIES OF SPECIMENS The following abbreviations are used to indicate the repositories of collections: TMM : University of Texas Memorial Museum, Austin, Texas. USNM : National Museum of Natural History, Washington, DC. AMNH : American Museum of Natural History, New York. CONVENTIONS All diameters are given in millimetres; D = diameter; Wb = whorl breadth; Wh = whorl height; U = umbilicus; ic = intercostal dimension; c = costal dimension. Figures in parentheses are dimensions as a percentage of the diameter. The term rib index as applied to heteromorphs is the number of ribs in a distance equal to the whorl height at the mid-point of the interval where the count was taken. The suture terminology of Wedekend (1916) as propounded by Kullmann and Wiedmann (1970) is used here with E = external lobe, L = lateral lobe, U = umbilical lobe and I = internal lobe. SYSTEMATIC PALAEONTOLOGY Order ammonoidea Zittel, 1884, pp. 355, 392 Suborder ammonitina Hyatt, 1889, p. 7 Superfamily haplocerataceae Zittel, 1884, p. 463 Family binneyitidae Reeside, 1928, p. 4 Genus borissiakoceras Arkhanguelsky, 1916, p. 55 Type species. By original designation: Borissiakoceras mirabilis Arkhanguelsky, 1916, p. 55, pi. 8, figs. 2, 3. Borissiakoceras orbiculation Stephenson, 1955 Plate 1, figs. 1-39; Plate 4, figs. 78-83 1955 Borissiakoceras orbiculatum Stephenson, p. 64, pi. 6, figs. IM. 1961 Borissiakoceras orbiculatum Stephenson; Cobban, p. 750, pi. 88, figs. 15—41 ; text-figs. 5a-f (with synonymy). 1988 Borissiakoceras orbiculatum Stephenson; Kennedy, p. 18, pi. 1, figs. 23-26 (with synonymy). 1990 Borissiakoceras orbiculatum Stephenson, 1955; Kennedy and Cobban, p. 85, pi. 1, figs. 1-14. Holotype. USNM 108832 from the basal Eagle Ford Group on Walnut Creek, 7 6 km (4-75 miles) northeast of Mansfield, Texas. Middle Cenomanian, Acanthoceras amphibolum zone. Material. More than 100 specimens, including USNM 423646-423665, from USGS Mesozoic locality 22871, Plesiacanthoceras wyomingense zone. Discussion. Cobban (1961) described this species in some detail and indicated how it differed from others referred to the genus. The present collection contains 50 specimens that were suitable for 386 PALAEONTOLOGY, VOLUME 33 measurement. Of these, 35 were microconchs and 13 macroconchs, two being unassigned. Microconchs (PI. 1, figs. 1-8, 26-38) are adult at 5-1 1 mm diameter and have umbilical ratios of 0-2 1-0-32. Macroconchs (PI. 1, figs. 9-25, 39) are adult at 9-7-27-3 mm and have umbilical ratios of 0-13-0-19. Occurrence. Acanthoceras amphibolum zone of Wyoming, Colorado, Kansas and Texas; Plesiacantlioceras wyomingense zone of Montana; Sciponoceras gracile zone of Texas. Superfamily desmocerataceae Zittel, 1895, p. 426 (nom.trans. Wright and Wright, 1951, p. 18; ex Desmoceratidae Zittel, 1895) Family desmoceratidae Zittel, 1895, p. 426 Subfamily desmoceratinae Zittel, 1895, p. 426 Genus moremanoceras Cobban, 1972, p. 465 Type species. Tragodesmoceras scotti Moreman, 1942, p. 208, pi. 33, fig. 8, text-fig. 2d; by original designation. Moremanoceras straini Kennedy, Cobban and Hook, 1988 Plate 1, figs. 40-45, 55, 56, 60-72 1955 Desmocerasl sp. Stephenson, p. 58, pi. 4, figs. 12, 13. 1977a Desmoceras (Pseudouhligella) aff. D. japonicum Yabe; Cobban, p. 22, pi. 11, figs. 1-6, 9, 10. \971b Desmoceras (Pseudouhligella) japonicum Yabe; Cobban, fig. 4a-e. 1988 Moremanoceras straini Kennedy, Cobban and Hook, p. 36, fig. la-g, i-t. Types. Holotype is USNM 416051 by original designation; paratypes USNM 416052-416060, from the base of the Boquillas Formation, Cerro de Cristo Rey, New Mexico, Acanthoceras amphibolum zone. Material. More than 60 specimens, including USNM 423667-423673, from USGS Mesozoic locality 22871, Plesiacantlioceras wyomingense zone. Discussion. Many of the present specimens retain original shell; specimens studied range from 2-5 to 67 mm diameter. The diagnostic features of the species are the compressed to slightly depressed explanation of plate 1 Figs. 1-39. Borissiakoceras orbicidatum Stephenson, 1955. 1, USNM 423646; 2, USNM 423647; 3-5, USNM 423648; 6, 30, USNM 423649; 7, 8, 34, 35, USNM 423656; 9-11, USNM 423650; 12, USNM 423651; 13, 14, USNM 423652; 15-17, USNM 423653; 18, 19, USNM 423654; 20, 21, 39, USNM 423655; 22, 23, USNM 423657; 24, 25, USNM 423658; 26, 27, USNM 423659; 28, USNM 423660; 29, USNM 423661 ; 31-33, USNM 423662; 36-38, USNM 423666, all from USGS Mesozoic locality 22871, Plesiacantlioceras wyomingense zone. Figs. 40-45, 55, 56, 60-72. Moremanoceras straini Kennedy, Cobban and Hook, 1988. 40^2, USNM 423667; 43-45, USNM 423668; 55, 56, 60-62, USNM 423669; 63-65, USNM 423670; 66-68, USNM 423671 ; 69, 70, USNM 423672; 71, 72, USNM 423673, all from USGS Mesozoic locality 22871, Plesiacantlioceras wyomingense zone. Figs. 46-48, 52-54. Moremanoceras costatum Cobban, Hook and Kennedy, 1989. 46-48, USNM 423676; 52-54, USNM 423690, from USGS Mesozoic locality D4466, Dunveganoceras pondi zone. Figs. 49-51, 57-59. Moremanoceras scotti (Moreman, 1942). 49-51, USNM 423674; 57-59, USNM 423675, from USGS Mesozoic locality D8314, Metoicoceras mosbyense zone. Figs. 15-17, 26-35, 39 are x2; Figs. 18, 19, 36-38 are x3; the remainder x 1. PLATE 1 KENNEDY and COBBAN, Borissiakoceras and Moremanoceras 388 PALAEONTOLOGY, VOLUME 33 whorls, biconcave growth lines on the shell surface (PI. 1, figs. 60, 64, 67) and periodic constrictions on the mould, both of which form an acute chevron on the venter (PI. 1, figs. 61, 65, 70). The venter is initially evenly rounded (PI. 1, figs. 40-45) but a blunt, rounded keel develops at maturity (PI. 1, figs. 69-72) as do blunt adapical collars to the constrictions. Moremamceras elgiiii (Young, 1958) (p. 292, pi. 39, figs. 4—20, 24, 25, 30, 31 ; text-fig. la-e) is more compressed when young, develops thickened collar-ribs to the constrictions from 15 mm diameter, has strong ventrolateral flank ribs when mature and never has a siphonal keel or ridge. M. costatum Cobban, Hook and Kennedy (1989) has a sharp keel that is present from a much earlier stage and strong concave ribs on the ventrolateral shoulder. M. montanaense sp. nov., described below, is a large, stout species that has ribs that are straight on the flanks rather than biconcave, and lacks the pronounced ventral chevron and keel of M. straini. M. scotti (Moreman, 1942) (p. 208, pi. 33, fig. 8; text-fig. 20; see Cobban, 1972, p. 6, pi. 2, figs. 1-23; text-figs. 3-5) has distant, flared collar ribs that are transverse over the venter, and never develops a keel (PI. 1, figs. 49-51; 57-59). Occurrence. A. amphibolwn zone of central and Trans-Pecos Texas; Plesiaccmthoceras wyomingense zone of Montana. Morenianoceras costatum Cobban, Hook and Kennedy, 1989 Plate 1, figs. 46-48, 52-54; Plate 2, figs. 1-35; Plate 4, figs. 76 and 77 1989 Morenianoceras costatum Cobban, Hook and Kennedy, p. 19, figs. 19, 64a-k, 65a-d, g, h (with full synonymy). Types. Holotype is USNM 425133, paratypes USNM 425134-425142, from the Metoicoceras mosbyense zone of uses Mesozoic locality D10186 in Luna County, New Mexico. Material. USNM 423677 to 423683 from USGS Mesozoic locality 12740, all M. mosbvense zone. USNM 423684 to 423687 and 423738 from USGS Mesozoic locality D5947; USNM 423676, 423688 and 423690, from USGS Mesozoic locality D4466, are from the Dunveganoceras pondi zone. D Wb Wh Wb: Wh U 423677 19-5 (100) 6-8 (34-9) 7-2 (36-9) 0-94 0-7 (3-6) 423678 18-2 (100) 9-0 (49-4) 10-5 (57-6) 0-86 1-2 (6-6) 423679 22-2 (100) 12-2(55-4) — — 1-2 (5-4) 423680 25-0 (100) 13-0 (52-0) 13-9(55-6) 0-94 2-1 (8-4) 423681 33-0 (100) 16-6(50-3) 18-5 (56-1) 0-90 2-4 (7-3) 423682 36-7 (100) 19-5 (53-1) 20-3 (55-3) 0-96 3-3 (9-0) 423683 57-8 (100) 30-0(51-9) 33-6 (58-1) 0-89 6-2(10-7) Discussion. The present material is much better preserved than the types. Very young specimens with shell preserved (PI. 2, figs. 1-12) show distant feebly flexuous ribs, and moulds bear constrictions that cross the venter in a narrow chevron with an adapical collar-rib. Both ribs and EXPLANATION OF PLATE 2 Figs. 1-35. Morenianoceras costatum Cobban, Hook and Kennedy, 1989. 1-3, USNM 423677; 4—6, USNM 423684; 7-9, USNM 423685; 10-12, USNM 423678; 13, 14, USNM 423679; 15-17, USNM 423680; 18, 19, USNM 423686; 20, 21, USNM 423687; 22-25, USNM 423681; 26-28, USNM 423682; 29-30, USNM 423688; 31, 32, USNM 423689; 33-35, USNM 423683. 1-3, 10-17, 22-28, 33-35 are from USGS Mesozoic locality 12740, Metoicoceras mosbyense zone. 4—9, 18-21 are from USGS Mesozoic locality D5947, Dunveganoceras pondi zont. 29-30 are from USGS Mesozoic locality D4466, Dunveganoceras pondi zom. 31, 32 are from USGS Mesozoic locality 12621, Dunveganoceras pondi zone. Figs. 7-9 are x 2; the remainder are x 1 . PLATE 2 KENNEDY and COBBAN, Moremanoceras costatum 390 PALAEONTOLOGY, VOLUME 33 constrictions strengthen as size increases (PI. 2, figs. 13-35), while a pronounced siphonal ridge is present on the shell from a diameter as small as 20 mm (PI. 2, figs. 18 and 19). This ridge may be markedly crenulate where crossed by the ribs. Both ribs and keel are present, if less prominent, on large moulds. The presence of pronounced falcoid ribs, especially well-developed on the outer flanks and venter plus the siphonal keel at the apex to a narrow ventral chevron distinguish this species from all others, as is apparent from the discussion under M. straini above. Occurrence. Calycoceras canitaurinum zone in New Mexico, Trans-Pecos Texas, western Oklahoma, central Kansas and north-central Colorado, and in the correlative Dunveganoceras pondi zone in Wyoming and Montana. Metoicoceras mosbyense zone in Wyoming and New Mexico. Moremanoceras montanaense sp. nov. Plate 3, figs. 62-64 Types. Holotype is USNM 423691 from USGS Mesozoic locality D12890, in the lower part of the Greenhorn Formation in sec. 5, T. 9 S., R. 59 E., Carter County, Montana. Paratype USNM 423692 is from USGS Mesozoic locality D 10201, Colorado Formation, 3 m (10 feet) to 4-6 m (15 feet) above base, NEj NW| sec. 20, T. 18 S., R. 20 W., Hidalgo County, New Mexico. M. mosbyense zone. Dimensions D Holotype USNM 423691 60-5 (100) at 450(100) Wb Wh Wb: Wh U 31-3 (51-7) 36-0 (59-7) 0-86 3-7 (61) 23-0 (51-1) 24-8 (55-1) 0-93 5-6 (12-4) Description. Holotype is a phragmocone 62 mm in diameter, retaining recrystallized shell, and slightly crushed, making accurate measurements impossible. Coiling is very involute, with a tiny, deep umbilicus with a flattened subvertical wall and narrowly rounded umbilical shoulder. Whorl section compressed, with flattened subparallel flanks and broadly rounded venter. Ornament is not visible on the innermost flank, but the mid to outer flanks and venter bear crowded ribs of variable strength and spacing. They arise as mere striae, are feebly convex at mid flank, concave over the outer flank and ventrolateral shoulder and cross the venter in a broad convexity. Periodic interspaces are deepened and presumably correspond to constrictions on the mould. Sutures not seen. Discussion. Large size and density of ribbing, course of ribs and lack of a siphonal ridge immediately separate M. montanaense from M. etgini, M. straini and M. costatum. The closest similarities are to M. scotti (Moreman, 1942) (p. 208, pi. 33, fig. 8; text-fig. 28; see Cobban, 1972, p. 6, pi. 2, figs. 1-23; text-figs 3-5), but this Sciponoceras gracile zone species has very widely separated flared ribs that extend down to the umbilical shoulder, and are separated by very fine riblets and growth striae only in middle and later growth. M. montanaense sp. nov. probably arose from M. costatum by retention of the ribbed, non-carinate morphology of the juvenile stages of the latter to a large size, plus modification in rib style and elimination of the marked ventral chevron of the later stages of M. costatum, leaving the broad ventral curvature of the juvenile (PI. 2, figs. 1-10). M. montanaense sp. nov. in turn probably gave rise to M. scotti by differentiation of ribbing during later growth. Occurrence. As for types. Superfamily acanthocerataceae de Grossouvre, 1894, p. 22 {nom. correct. Wright and Wright, 1951, p. 24, pro. Acanthoceratida Hyatt, 1900, p. 585; nom. transl. ex Acanthoceratidae Hyatt, 1900, p. 585; nom. correct, ex Acanthoceratides de Grossouvre, 1894). KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 391 Family acanthoceratidae de Grossouvre, 1894, p. 22 {nom. correct. Hyatt, 1900, p. 585; ex Acanthoceratides de Grossouvre, 1894, p. 22). Subfamily acanthoceratinae de Grossouvre, 1894, p. 22 {non}, correct. Hyatt, 1900, p. 585; ex Acanthoceratides de Grossouvre, 1894; nom. transl. Wright and Wright, 1951, p. 28, ex Acanthoceratides de Grossouvre). Genus cunningtoniceras Collignon, 1937, p. 64 (40) (? = Gnerangericeras Thomel, 1972, p. 119) Type species. Ammonites cwmingtoni Sharpe, 1855, p. 35, pi. 15, fig. 2. Cunningtoniceras sp. juv. Plate 3, figs. 1-7 Material. Five specimens, including USNM 423693-423695, from USGS Mesozoic Locality 12650, Metoicoceras mosbyense zone. Description. Specimens range from 7 to 12-5 mm in diameter. In the smaller specimens coiling is very involute with a small, deep umbilicus comprising approximately 19% of the diameter. The whorl section is depressed, reniform in intercostal section and polygonal in costal section, with the greatest breadth at the outer ventrolateral tubercles. There are 6-7 strong umbilical bullae per whorl that give rise to broad, blunt primary ribs. The latter weaken at mid-flank and alternate regularly with single secondary ribs that arise at mid-flank. All bear a strong conical inner ventrolateral tubercle that appears to have been the base for a long spine. These ribs broaden and sweep forwards to a strong conical outer ventrolateral tubercle on the mould of the phragmocone; USNM 423693 shows these tubercles to have been the bases of long septate spines (PI. 3, figs. 1-3). A broad rib passes straight across the venter, and bears a weak siphonal clavus. Shorter, intercalated ribs are also present; some bear only outer ventrolateral and siphonal tubercles, others bear only the siphonal row; there are as many as 18 ventral ribs per whorl. Discussion. Preservation of septate spines on USNM 423693 gives this specimen a quite remarkable appearance (PI. 3, figs. 1-3). The variable nature of the intercalated ventral ribs shows these tiny specimens to be Cunningtoniceras beyond any doubt; see Wright and Kennedy (1987) for a recent review of the genus. The specimens are so tiny that they cannot be usefully compared with any previously described species, although we have seen comparable tiny limonitic nuclei of Cunningtoniceras from Upper Cenomanian pelagic clay facies in Tunisia. Cunningtoniceras is widespread in the US Western Interior and Gulf Coast regions, with, for instance, C. inerme (Pervinquiere, 1907) in the Conlinoceras tarrantense zone, C. lonsdalei (Adkins, 1928) in the Acanthoceras hellense zone and C . johnsonanum (Stephenson, 1955) in the Acanthoceras amphiholum zone of Texas (Kennedy and Cobban, 1990). A range of species in the upper Cenomanian of New Mexico and Arizona includes C. arizonense Kirkland and Cobban, 1986 (p. 2, pis. 1-8), of which the present specimens might conceivably be nuclei. Occurrence. As under Material. Genus tarrantoceras Stephenson, 1955, p. 59 Type species. Tarrantoceras rotatile Stephenson, 1955, p. 59, pi. 5, figs. 1-10; by original designation = Mantelliceras sellarctsi Adkins, 1928, p. 239, pi. 25, fig. 1 ; pi. 26, fig. 1. 392 PALAEONTOLOGY, VOLUME 33 Tarrantoceras cuspidum (Stephenson, 1953) Plate 3, figs. 8-12 1953 Acanthoceras cuspidum Stephenson, p. 202, pi. 50, figs. 1^. 1990 Tarrantoceras cuspidum (Stephenson, 1953); Kennedy and Cobban, p. 134, pi. 14, figs. 21-24, 26-28. Types. Holotype is USNM 105974, by original designation; paratype is USNM 105975, both from gullies south of the old Sherman road, 4-5 km east of Whitesboro, Grayson County, Texas. An unfigured paratype is from USGS Mesozoic locality 14092, a bluff 1-6 km north and 2-9 km east of Sadler, Grayson County, Texas. All are from the Templeton Member of the Woodbine Formation, Plesiacanthoceras wyomingense zone. Material. USNM 423697 as well as other specimens from USGS Mesozoic locality 22871, P. wyomingense zone. USNM 423696 and other specimens from USGS Mesozoic locality D4466, Upper Cenomanian Dunveganoceras pondi zone. Discussion. USNM 423697 is 22 mm in diameter (PI. 3, figs. 10 12) and differs in no significant respects from the types. USNM 423696 is larger than the types, and shows the same strong ornament persisting to a whorl height of 13 mm (PI. 3, figs. 8 and 9). The innermost whorls of USNM 423697 are well-exposed. They show feeble umbilical bullae giving rise to low, broad prorsiradiate ribs that terminate in massive inner ventrolateral spines that are housed in notches in the umbilical wall of preceding whorl, as in the types. I. cuspidum has not been previously recognized outside Texas, and the present occurrence is of some importance in providing a probable date for the Templeton Member. Occurrence. Plesiacanthoceras wyomingense zone of Texas and Montana; Dunveganoceras pondi zone of Montana. EXPLANATION OF PLATE 3 Figs. 1-7. Cunningtoniceras sp. juv. 1-3, USNM 423693; 4, 5, USNM 423694; 6, 7, USNM 423695, from USGS Mesozoic locality 12650, Metoicoceras moshyense zone. Figs. 8-12. Tarrantoceras cuspidum (Stephenson, 1953). 8, 9, USNM 423696; 10-12, USNM 423697, from USGS Mesozoic localities D4466 and 22871, Dunveganoceras pondi and Plesiacanthoceras wyomingense zones. Figs. 13-21, 26-31, 35-44. Kastanoceras spiniger gen. et sp. nov. 13-15, holotype, USNM 423699; 16-18, paratype USNM 423700; 19-21, paratype USNM 423701 ; 26-28; paratype USNM 423702; 29-31, paratype USNM 423703; 35-40, paratype USNM 423704; 41^4, paratype USNM 423705, all from USGS Mesozoic locality 22871, P. wyomingense zone. Figs. 22-25. Tarrantoceras exile sp. nov. Holotype, USNM 423698, from USGS Mesozoic locality 12650, Metoicoceras moshyense zone. Figs. 32-34. Tarrantoceras sellardsi (Adkins, 1928). USNM 400767, from USGS Mesozoic locality D 12626, Acanthoceras amphibolum zone. Figs. 45, 46, 51-56. Sumitomoceras spp. juv. 45, 46, USNM 423706; 51, 52, USNM 423707, both from USGS Mesozoic locality 23062, Sciponoceras gracile zone. 53-56, USNM 423708, from Mesozoic locality D4628, S', gracile zone. Figs. 47-50, 57-61. Sumitomoceras conlini (Wright and Kennedy, 1981). 47-50, USNM 400804, from USGS Mesozoic locality D11529; 57, 58, USNM 400807, from USGS Mesozoic locality D10196; 59-61, USNM 400805, from USGS Mesozoic locality D11529, all S. gracile zone. Figs. 62-64. Moremanoceras montanaense sp. nov. Holotype, USNM 423691, from USGS Mesozoic locality D 12890, M. moshyense zone. Figs. 1-3, 6, 7, 13-34, 38-48 are x 2; figs. 4, 5 are x 3; the remainder are x 1. PLATE 3 KENNEDY and COBBAN, Cenomanian ammonites 394 PALAEONTOLOGY, VOLUME 33 Tarrantoceras exile sp. nov. Plate 3, figs. 22-25 Derivation of name. Exilis (Latin): slender, thin, pertaining to the whorl section of the species. Holotype. USNM 423698 from USGS Mesozoic locality 12650, Metoicoceras moshyense zone. Dimensions D Wb Wh Wh:Wh U USNM 423698 12-61(100) 4-3 (34-1) 5-1 (40-5) 0-84 4-3 (34-1) Description. Coiling evolute, with broad shallow umbilicus comprising 34% of diameter with low, rounded wall that is indented to accommodate the inner ventrolateral tubercles of the preceding whorl (PI. 3, fig. 23). Whorl section of this specimen is that of a Tarrantoceras, as can be seen by comparison with specimens of T. sellardsi illustrated for comparison (PI. 3, figs. 32-34). Coarseness of ornament yet lack of massive inner and outer ventrolateral tubercles immediately distinguish it from T. cuspidum (PI. 3, figs. 8-12) whereas the combination of evolute slender whorls and coarse ornament gives the shell a quite different appearance than any T. sellardsi we have seen. It represents the youngest member of the Tarrantoceras lineage. Occurrence. As for types. Genus kastanoceras nov. Derivation of name. Kastanos (Greek); chestnut, from the common spinosity of the new genus and the seed cases of that tree. Type species. Kastanoceras spiniger gen. et sp. nov., Plesiacanthoceras wyomingense zone of Montana. Diagnosis. Dwarf, presumed microconch adult at 8 mm, largest (incomplete) macroconch is 10 mm in diameter. Coiling evolute, coronate, intercostal section depressed reniform, costal section with flattened sides that converge to an umbilical wall notched to accommodate outer ventrolateral spines of preceding whorl. Flank ribs feeble, prorsiradiate, terminating in large inner ventrolateral spines. Venter with feeble outer ventrolateral and siphonal clavi that decline at smallest diameters visible. Ornament declines markedly on adult body chamber. Suture with broad, little-incised, asymmetrically bifid E/L, narrower L and simple bifid L/Uj. Discussion. Inner whorls are inseparable from those of T. cuspidum, with which K. spiniger gen. et sp. nov., occurs (PI. 3, figs. 13-21). But whereas T. cuspidum grows to a diameter of at least 35 mm and has outer whorls with very strong, close-spaced clavi and strong flank ornament at this size, the present specimens show approximation of sutures and decline of body chamber ornament that indicates them to be adult at phragmocone diameters of as little as 7-5 mm in the microconch holotype. Kastanoceras is thus a progenic dwarf derivative of Tarrantoceras ]wsi as Protacanthoceras Spath, 1923 is a similarly derived dwarf offshoot of Acanthoceras Neumayr, 1875 (Wright and Kennedy, 1980, 1987; Kennedy and Wright, 1985). Occurrence. P. wyomingense zone, Montana. Kastanoceras spiniger gen. et sp. nov. Plate 3, figs. 13-21, 26-31, 35-44 Types. Holotype is USNM 423699, paratypes USNM 423700-423705 from USGS Mesozoic locality 22871, P. wyomingense zone. Diagnosis. With the characters of the genus. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 395 Description. Small, presumed microconch 8 mm diameter, largest, incomplete macroconch has phragmocone diameter of 7-5 mm and total diameter of 10 mm with only 120° of body chamber preserved. Coiling is evolute with the umbilicus comprising 39% of the diameter, deep, with a very low, rounded wall that is notched to accommodate the inner ventrolateral spines of the preceding whorl. Whorl section depressed reniform in intercostal section. In costal section the flanks are concave and diverge outwards, with the greatest breadth at the inner ventrolateral spine; the venter is broadly arched. There are 9-1 1 feeble umbilical bullae that connect by low, broad prorsiradiate ribs to prominent inner ventrolateral spines that are directed outwards and normal to the median plane of the shell. Feeble outer ventrolateral and siphonal clavi are barely visible at the smallest diameter seen, but they strengthen towards the beginning of the adult body chamber and may be linked by low spiral ridges. Ventral ribbing is feeble or absent; interspaces between rows of ventral tubercles may be deepened and constriction-like. On the adult body chamber all ornament weakens markedly. Discussion. Differences from T. cuspidum, from which the species arose paedomorphically, are discussed under the genus. Occurrence. As for types. Genus sumitomoceras Matsumoto, 1969, p. 280. Type species. Sumitomoceras faustum Matsumoto and Muramoto, 1969, p. 283, pi. 283, pi. 38, figs. 1^; text- fig. 8, by original designation. Sumitomoceras spp. juv. Plate 3, figs. 45, 46, 51-56; Plate 4, figs. 57, 58; Plate 6, figs. 32-35 Material. USNM 423706 and 423707 from USGS Mesozoic locality 23062; USNM 423708 from USGS Mesozoic locality D4628; USNM 423731 from USGS Mesozoic locality D5780; USNM 423747, from USGS Mesozoic locality D12052; all Upper Cenomanian Sciponoceras gracile zone. Description. USNM 423706 is only 7-5 mm in diameter. Coiling is evolute, with U = 21% of diameter, with a low, subvertical wall and narrowly rounded umbilical shoulder. The whorl section is compressed with subparallel flanks, broadly rounded ventrolateral shoulders and an arched venter in intercostal section, the costal section is polygonal. Primary ribs arise from feeble bullae perched on the umbilical shoulder. They are narrow and prorsiradiate and separated by 1 or 2 shorter intercalated ribs. All ribs bear small conical inner ventrolateral tubercles. A strong rib sweeps forward across the ventrolateral shoulder to an outer ventrolateral clavus. Scarcely visible at the smallest diameter visible, this rib becomes more prominent as size increases. USNM 423707 shows similar ornament at the smallest diameter visible, but is 18 mm in diameter (PI. 3, figs. 51, 52), and shows both inner and outer ventrolateral tubercles declining. In USNM 423706 some of the interspaces are distinctly deepened and constriction-like. USNM 423708 is 15-7 mm in diameter, has weak or no umbilical bullae, very weak ornament on the inner flanks, but prominent inner ventrolateral tubercles throughout, and no, or incipient outer ventrolaterals. USNM 423747 (PI. 6, figs. 32-35) has a maximum preserved diameter of 22-5 mm. Coiling is very evolute, with a broad shallow umbilicus comprising 23% of diameter. Whorl section is compressed, with flattened subparallel flanks and broadly rounded venter. Flank ornament consists of low, crowded ribs, 18 per half whorl, of which eight arise at incipient umbilical bullae, the remainder intercalating. Ribs are flexuous and prorsiradiate, convex across the inner mid-flank, thereafter concave and strengthening, crossing the venter in a shallow convexity. There are feeble rounded inner ventrolateral and clavate outer ventrolateral and siphonal tubercles at the smallest diameter visible. The inner ventrolateral tubercles efface as size increases but feeble outer ventrolateral and siphonal tubercles persist to the largest diameter seen. Suture with broad, symmetrically bifid E/L, smaller bifid L, little incised L/U,^ and small U. Discussion. The identity of these specimens is demonstrated by comparison with an ontogenetic series of S. conlini (Wright and Kennedy, 1981) from the S. gracile zone in New Mexico (PI. 3, figs. 47-50, 57-61). The smallest, USNM 400804, can be linked to USNM 400805 which shows very early loss of all tuberculation and is presumed to be a microconch (PI. 3, figs. 59-61); in others, the tubercles persist to a much greater size (USNM 400807: PI. 3, figs. 57 and 58). Occurrence. As for material. 396 PALAEONTOLOGY, VOLUME 33 Genus alzadites nov. Derivation of name. From the town of Alzada, Carter County, Montana, 9-7 km southeast of the type locality. Type species. Alzadites alzadensis gen. et sp. nov., Plesiacanthoceras wyomingense zone, Montana. Diagnosis. Small, adult at 16-5 mm or less in diameter. Involute with small umbilicus, whorl section compressed with flattened subparallel flanks and rounded venter. Phragmocone with tiny, distant umbilical bullae, distant prorsiradiate ribs and feeble to strong inner ventrolateral tubercles plus outer ventrolateral and siphonal clavi. Ventral ribbing strengthens on adult body chamber and tubercles decline, leaving strong, coarse, prorsiradiate, concave ribs on outer flank that cross venter in a broad convexity, or a chevron, separated by broad interspaces. Constrictions may develop on internal moulds of phragmocone and body chamber. Suture with simple, little-incised elements; E/L broad, and symmetrically bifid, L narrow, shallow, bifid, saddles on umbilical lobe simple, bifid. Discussion. Alzadites most closely resembles certain Protacanthoceras Spath, 1923, and the type species is homeomorphic with P. asgeirri Wright and Kennedy, 1980 (p. 90, figs. 20-21, 47). This is scarcely surprising inasmuch as both are interpreted as progenic dwarfs and as a result share certain features that are common to most acanthoceratine nuclei. Protacanthoceras derives from Acanthoceras Neumayr, 1875, and the type species. Ammonites bunhurianus Sharpe, 1853 (p. 25, pi. 9, fig. 3; see Wright and Kennedy, 1980, p. 91, figs. 29-33, 41^3, 48; 1987, p. 215, pi. 55, figs. 10-16; text-figs. 83b, c) and many of the other early Protacanthoceras species are easily differentiated from EXPLANATION OF PLATE 4 Figs. 1-10, 14-16, 43. Alzadites alzadensis gen. et sp. nov. 1-3, 43, paratype USNM 423710; 4-7, holotype USNM 423709; 8-10, paratype USNM 423712; 14-16, paratype USNM 42371 1, all from USGS Mesozoic locality 22871, Plesiacanthoceras wyomingense zone. Figs. 11-13, 17-39. Alzadites westonensis gen. et sp. nov. 11-13, paratype USNM 423715; 17, 18, paratype USNM 423716; 19, 20, paratype USNM 423717; 21-23, paratype USNM 423718; 24, 25, paratype USNM 423719; 26-29, paratype USNM 423720; 30-32, paratype USNM 423721 ; 33-36, holotype USNM 423714; 37-39, paratype USNM 423722, all from USGS Mesozoic locality D5947, Dunveganoceras pondi zone. Figs. 40-42, 46^8. Alzadites incomptus gen. et sp. nov. Paratype USNM 423728, from USGS Mesozoic locality D5249, Sciponoceras gracile zone. Figs. 44, 45. Alzaditesl sp. USNM 423730, from USGS Mesozoic locality D5947, D. pondi zone. Figs. 49-52, 65-70. Microsulcatoceras puzosiiforme gen. et sp. nov. 49-52, holotype, USNM 423734; 65-68, paratype USNM 423735; 69, 70, paratype USNM 423736, all from USGS Mesozoic locality 23062, S. gracile zone. Figs. 53-56. Alzadites sp. A. 53-54, USNM 423732; 55-56, USNM 423733; from USGS Mesozoic locality D5780, S', gracile zone. Figs. 57, 58. Sumitomoceras sp. juv. USNM 423731, from USGS Mesozoic locality D5780. Figs. 59-64. Microsulcatoceras texanum gen. et sp. nov. Holotype USNM 423739 from stream bank 2 4 to 2-9 km southwest of Britton, on and east of Rogers Farm, Ellis County, Texas. Britton Formation, S. gracile zone. Figs. 71-73, 84, 85. Microsulcatoceras! sp. USNM 423740, from USGS Mesozoic locality D11514, Neocardioceras juddii zone. Figs. 74, 75, 86, 87. Microsulcatoceras crassum gen. et sp. nov. Holotype, USNM 423737, from USGS Mesozoic locality D4682, S. gracile zone. Figs. 76, 77. Moremanoceras costatum Cobban, Hook and Kennedy, 1989. USNM 423738, from USGS Mesozoic locality D5947, D. pondi zone. Figs. 78-83. Borissiakoceras orhiculatum Stephenson, 1955. 78, 79 are USNM 423663, from USGS Mesozoic locality D5947; 80-83 are USNM 423664 from USGS Mesozoic locality D4462, all from D. pondi zone. Figs. 37^5, 51, 52, 59-61, 67-77 are x2; the remainder are x 1. PLATE 4 KENNEDY and COBBAN, Cenomanian ammonites 398 PALAEONTOLOGY, VOLUME 33 Alzadites by their polygonal costal whorl section, coarse ribbing and tuberculation which persists to the body chamber. Only the later Protacanthoceras species that are progenically derived from other, already diminutive species of the genus come to resemble Alzadites. The evolutionary origin of Alzadites lies in some upper Cenomanian acanthoceratine of the US Western Interior lineages, rather than in Old World Acanthoceras. There is a marked similarity between the smooth, distantly and feebly ribbed and tuberculate phragmocones of Alzadites and the early whorls of certain Tarrantoceras Stephenson, 1955 (e.g. PI. 4, figs. 32 and 33) although these generally have stronger inner ventrolateral tubercles than in the type species of Alzadites (PI. 4, figs. 1-10, 14, 15, 17, 43), more closely resembling the inner whorls of A. westonensis gen. et sp. nov. (PI. 4, figs. 8-39). Occurrence. Plesiacanthoceras wyommgense zone of Montana and Wyoming. Upper Cenomanian Sciponoceras gracile zone of Utah. Alzadites alzadensis gen. et sp. nov. Plate 4, figs. 1-10, 14-16, 43 Types. Holotype is USNM 423709, paratypes USNM 423710 to 423712; three unfigured paratypes USNM 423713, all from USGS Mesozoic locality 22871, P. wyomingense zone. D Wh Wh Wb:Wh U USNM 423710 160(100) 5-6(35-0) 7-4 (46-3) 0-76 2-9(18-1) USNM 423709 16-6(100) 6-5 (39-2) 9-5 (57-2) 0-68 2-4(14-5) at 14-2 (100) 5-7(40-1) 7-2 (50-7) 0-79 1-7 (12-0) USNM 423711 15-6 (100) 6-4 (41-0) 8-3 (53-2) 0-77 2-1 (13-5) USNM 423712 13-4(100) 5-7 (42-5) 6-7 (50-0) 0-85 1-7(12-7) Diagnosis. Alzadites with coarse, blunt ribs on body chamber, where tubercles decline and ultimately disappear. Description. Coiling involute with small, shallow umbilicus. Umbilical wall low, umbilical shoulder narrowly rounded. Whorl section compressed, with flattened, subparallel flanks, ventrolateral shoulders and venter broadly and evenly rounded. Phragmocone very feebly ornamented. In the best preserved specimen, up to 11 feeble umbilical bullae give rise to low, narrow, prorsiradiate, distant ribs that efface at mid-flank (PI. 4, figs. 8-10); feeble intercalated ribs are also present. Most if not all ribs bear a feeble inner ventrolateral tubercle. The ribs efface over the venter, where there are outer ventrolateral and siphonal clavi (PI. 4, figs. 8 and 10). On the adult body chamber, outer flank and ventral ribbing strengthens and coarsens markedly; the ribs are concave on the outer flank and cross the venter in a broad convexity. The ribs bear outer ventrolateral and siphonal clavi at the beginning of the body chamber, but these progressively efface and disappear. The interspaces between ribs are broad and some are deepened into constrictions. There is a great variation in the strength and visibility of ornament, especially on phragmocones. Suture simple, as for genus. Discussion. A. alzadensis gen. et sp. nov., differs from A. westonensis gen. et sp. nov., in the following respects; it is larger, the inner ventrolateral tubercles are much weaker, the body chamber ribbing blunt and restricted to the outer flank whereas that of A. westonensis extends to the umbilical seam, is markedly flexuous and sharper, with a pronounced acute ventral chevron and persistent tubercles. Occurrence. As for types. Alzadites westonensis sp. nov. Plate 4, figs. 11-13, 17-39 Derivation of name. From Weston County, Wyoming, where the types were found. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 399 Types. Holotype is USNM 423714, paratypes are USNM 423715 to 423723, 12 unfigured paratypes are USNM 423724, all from USGS Mesozoic locality D5947, Dimveganoceras pondi zone. Diagnosis. Alzadites with variable, generally strong ribs that arise singly or in pairs from umbilical bullae or intercalate. The ribs are flexuous, with persistent inner ventrolateral tubercles. Outer ventrolateral and siphonal clavi are borne on chevron ribs that persist to the adult body chamber. Description. Small, adults 15 mm or less in diameter. Coiling involute with small shallow umbilicus, umbilical wall low, rounded, umbilical shoulder narrowly rounded. Whorl section compressed with subparallel flanks and rounded venter in intercostal section; costal section polygonal with greatest breadth at inner ventrolateral tubercle; venter fastigiate. Phragmocone ornament varies from weak to strong. There are thus weak to strong umbilical bullae that give rise to pairs of weak to strong ribs, either singly or in pairs, with oeeasional intercalated ribs to give a total rib density of around 1 1 ribs per half whorl in robustly ornamented individuals. The ribs are prorsiradiate and flexuous, and bear weak to strong, conical, inner ventrolateral tubercles that are housed in notches in the umbilical wall of the succeeding whorl (PI. 4, fig. 38). The ribs sweep forward over the ventrolateral shoulders to clavate inner ventrolateral clavi, linked to turn to strong siphonal clavi at the apex of an acute chevron. This ornament persists onto the adult body chamber. Towards the mature aperture umbilieal and ventrolateral tubercles deeline first. The adult aperture is preceded by a few crowded ribs that are restricted to the outer flank and venter and lack tubercles. There is a pronounced ventral lappet (PI. 4, fig. 34). Suture simple, as for genus. Discussion. Diflferences from A. alzadensis gen. et sp. nov., are discussed under that species. Occurrence. As for types. Alzadites incomptus gen. et sp. nov. Plate 4, figs. 40^2, 46-48; Plate 6, figs. 1-22 Derivation of name. Incomptus (Latin): unadorned. Types. Holotype is USNM 423725, from USGS Mesozoic locality D 12052 as are figured paratypes USNM 423726 and 423727; paratype USNM 423728 is from USGS Mesozoic locality D5249, SEj sec. 12, T. 43 S., R. 2 E., Kane County, Utah, Tropic Shale, 91-10-7 m (30-35 feet) above base; unfigured paratype USNM 423729 is from USGS Mesozoic locality D5255, NE| NE| sec. 32, T. 41, S., R. 7 E., Kane County, Utah, Tropic Shale 4'6-9-2 m (15-30 feet) above base. All Sciponoceras gracile zone. Diagnosis. Small, adult at 12 mm diameter. Phragmocone smooth to feebly to strongly ribbed with blunt umbilical bullae and feeble inner and outer ventrolateral tubercles, interspaces sometimes deepened into constrictions. Tubercles decline on adult body chamber, which is ornamented by delicate prorsiradiate ribs and may be constricted. Description. The type specimens are rather variable (PI. 6, figs. 1-22). Coiling is involute, with a small umbilicus, comprising 21-24% of diameter in adults, shallow, with a low, flattened wall and narrowly rounded umbilical shoulder. The whorl section is compressed, with flattened subparallel sides and a rounded venter. Phragmocones vary from smooth (PI. 6, figs. 10-12) to those with weak umbilical bullae, up to nine per whorl. These give rise to low, blunt, prorsiradiate ribs, singly or in pairs, while shorter, intercalated ribs arise around mid-flank. The ribs are feebly flexed, and strengthen across the flank, crossing the venter in a broad convexity. Interspaces are sometimes deepened into constrictions. Tuberculation is poorly developed, but the most coarsely ribbed individuals develop indications of inner and outer ventrolateral and siphonal tubercles on some ribs. This general style of ornament persists on to the beginning of the adult body chamber, the last part of which is characterized by delicate, flexuous flank ribs that strengthen over the ventrolateral shoulder and venter, are concave on the former and cross the latter in a broad convexity. Suture with little-divided elements; E narrow, E/L broad and bifid, L shallow and bifid. 400 PALAEONTOLOGY, VOLUME 33 Discussion. Weakness of ornament, notably tubercles, plus pattern of ribbing immediately distinguish this species from A. westonensis gen. et sp. nov., described above. A. alzadensis gen. et sp. nov., is more similar, but has a broader, larger shell with coarser ribbing on the adult body chamber. Occurrence. As for types. Alzaditesl sp. Plate 4, figs. 44, 45 Material. USNM 423730 from USGS Mesozoic locality D5947, Dunveganoceras pondi zone. Description. Specimen is a phragmocone retaining traces of the original aragonitic shell and is 8-2 mm in diameter. Coiling is involute with U = 22% of diameter, the umbilical wall low, the umbilical shoulder narrowly rounded. The whorl section is compressed, with flattened subparallel flanks and a rounded venter in intercostal section. Ribs, which number eight per half whorl, are weak and prorsiradiate on the flank but strengthen markedly on the venter where they are high and flared with flattened tops. Occasional unflared intercalated ribs are present. Sutures not seen. Discussion. We believe this specimen to be pathological. It is slightly asymmetrical in ventral view (PI. 4, fig. 45), and resembles symmetrical malformed specimens such as Ammonites salteri of Sharpe, 1857 (pi. 23, figs. 3 and 5). Of species present in the same concretion the general shell morphology most closely resembles that of A. westonense gen. et sp. nov. Occurrence. As for material. Alzadites sp. A. Plate 4, figs. 53-56 Material. USNM 423732 and 423733 from USGS Mesozoic locality D5780, NEf SE| sec. 8, T. 5 S., R. 2 E., Socorro County, New Mexico. Lower part of Mancos Shale, Bridge Creek Limestone Beds, second limestone from base. Sciponoceras gracile zone. Description. The largest complete specimen is 18 mm in diameter. All are crushed, with consequent effacement of ornament. In USNM 423732 (PI. 4, figs. 53-56) the phragmocone is smooth, in USNM 423733 (PI. 4, figs. 53 and 54) feeble bullae give rise to prorsiradiate primary ribs with shorter intercalated secondaries between, all ribs strengthening over the venter. All specimens show persistent ribbing on the body chamber, with some interspaces deepened and constriction-like. Discussion. Such of the ornament as is visible recalls that of A. incomptus gen. et sp. nov., but poor preservation precludes positive determination. Occurrence. As for material. Genus microsulcatoceras nov. Derivation of name. Mikros (Greek), small; sulcus (Latin), groove, pertaining to the size and ornament of the shell. Type species. Microsulcatoceras puzosiiforme gen. et sp. nov., Sciponoceras gracile zone of Montana. Diagnosis. Small, adult at 10 mm or less. Compressed, with flattened subparallel sides and rounded venter. On phragmocone distant umbilical bullae give rise to prorsiradiate, straight, primary ribs that terminate in conical inner ventrolateral tubercles. These tubercles link over the venter via a low. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 401 convex rib, followed by a broad, shallow constriction. Tubercles decline on adult body chamber where constrictions strengthen and are flanked by flexuous, prorsiradiate, collar ribs. Sutures unknown. Discussion. This diminutive genus is quite unlike any other described micromorph. At first sight the presence of flexuous constrictions is suggestive of the superfamily Desmocerataceae Zittel, 1895, notably certain Puzosiinae. The presence of tubercles at such a small size is not, however, a feature of this group (although they may develop at maturity). Instead, we believe the affinities of Microsulcatoceras may lie in certain Acanthoceratinae. There is a marked resemblance of the innermost whorls of the Sumitomoceras from the gracile zone described above to the inner whorls of the new genus (compare PI. 3, figs. 43-56 and PI. 4, figs. 59-75) : both have prominent umbilical and inner ventrolateral tubercles, but Microsulcatoceras lacks the outer, having instead a ventral rib that, although accentuated at the outer ventrolateral position, does not differentiate into a distinct tubercle. Both lack a siphonal tubercle, while certain adult Sumitomoceras, including the type, develop deepened, constricted interspaces between the ribs. On the balance of the evidence, we are inclined to regard Microsulcatoceras as a progenic dwarf derivative of Sumitomoceras. Occurrence. Sciponoceras gracile zone of Montana, Texas, and possibly New Mexico. Microsulcatoceras puzosiiforme gen. et sp. nov. Plate 4, figs. 49-52, 65-70 Derivation of name. Puzosiiforme - Puzosia-like, from the superficial resemblance to certain Puzosia Bayle, 1878. Types. Holotype is USNM 423734, paratypes USNM 423735 and 423736, from USGS locality 23062, S. gracile zone. Diagnosis. Microsulcatoceras with delicately ribbed and constricted body chamber. Description. Small, adult at 9-5 mm. Coiling evolute with small, shallow umbilicus; umbilical wall low, flattened, umbilical shoulder narrowly rounded. Whorl section compressed with flattened, subparallel flanks and broadly rounded venter. Phragmocone has tiny distant umbilical bullae, 8 per whorl, that give rise to low, prorsiradiate ribs that terminate in blunt inner ventrolateral tubercles linked over the venter by a broad, convex rib. Intercalated ribs with feeble to obsolete inner ventrolateral tubercles are occasionally present. The ribs are succeeded by broad, shallow constrictions, most obvious on the outer flank and over the venter. Tubercles decline on the adult body chamber and constrictions strengthen, extending down to the umbilical wall. The constrictions are prorsiradiate, markedly concave on the outer flank, and slightly flexuous; they are bordered by collar ribs and cross the venter in a broad convexity. Interspaces bear shorter intercalated ribs, most prominent just before the adult aperture. Sutures not seen. Discussion. M. puzosiiforme gen. et sp. nov., is easily differentiated from M. crassum and M. texanum gen. et spp. nov., by its delicate body chamber ornament and clearly differentiated constrictions, rather than the coarse decoration of the latter, where constrictions are less well differentiated and umbilical bullae persist. Occurrence. As for types. Microsulcatoceras erassum gen. et sp. nov. Plate 4, figs. 74, 75, 86, 87 Derivation of name. Crassus (Latin): thick, referring to the body chamber ornament. 402 PALAEONTOLOGY, VOLUME 33 Types. Holotype is USNM 423737, from USGS Mesozoic locality D4628, NW| N£i sec. 11, T. 43. S., R. 2 E., Kane County, Utah, Tropic Shale, from concretions 3 m (10 feet) above base. Sciponoceras gracile zone. Diagnosis. Microsulcatoceras with coarse body chamber ornament and persistent umbilical bullae. Description. Holotype and only known specimen is 13 mm in diameter. Umbilicus small, 33% of diameter. Whorl section compressed with flattened subparallel flanks and rounded venter. Blunt umbilical bullae give rise to pairs of flexuous prorsiradiate coarse ribs, with occasional shorter, intercalated ribs. All ribs are concave and strengthen markedly on the outer flanks and venter which they cross in a broad convexity. Inner flank ribs decline in strength on the last part of the specimen and ventral ribbing crowds, suggesting it to be adult. Some interspaces are slightly deepened. Discussion. Coarseness of ornament and persistence of bullae onto the body chamber immediately distinguish this species from M. puzosiiforme. Occurrence. As for type. Microsulcatoceras texanum gen. et sp. nov. Plate 4, figs. 59-64 Type. USNM 423739 from stream bank 2-4 to 2-9 km southwest of Britton, on and east of the Rogers Farm, Ellis County, Texas, Eagle Ford Group, Britton Formation, Sciponoceras gracile zone, ex J. P. Conlin collection. Diagnosis. Small, adult at 9-5 mm diameter. Phragmocone and early body chamber with distant, feebly bullate primary ribs with strong ventrolateral tubercles linked over the venter by a strong convex rib, and occasional prominent constrictions. Last part of body chamber loses tubercles and develops strong, crowded ventrolateral and ventral ribs. Discussion. The phragmocone ribbing and tuberculation are much stronger than in M. puzosiiforme, and persist onto the body chamber; the coarsely ribbed venter immediately preceding the adult aperture is equally distinctive. These features of the body chamber also distinguish the species from M. crassum gen. et sp. nov. Occurrence. As for type. Microsulcatoceras sp. ? Plate 4, figs. 71, 72, 73, 84, 85 Material. USNM 423740 from USGS Mesozoic locality D11514, Slate Creek in the NW| SW^ sec. 36, T. 17 S., R. 18 W., Grant County, New Mexico. Colorado Formation, 9-12 m above flaggy member, Neocardioceras juddii zone. Discussion. This badly preserved specimen is 9-5 mm in diameter. Features suggesting it might possibly be a late species of Microsulcatoceras are the presence of periodic constrictions and associated collar ribs on a compressed, flat sided shell, although the ventrolateral tubercles typical of juvenile Microsuleatoceras are lacking. It might also possibly be a poorly preserved Sumitomoceras or Pseudocalyeoceras. Occurrence. As for material. Genus plesiacanthoceras Haas, 1964 (= Paraeanthoceras Haas, 1963, p. 2; non Furon, 1935, p. 59) Type species. By original designation; Metoicoceras wyomingense Reagan, 1924, p. 181, pi. 19, figs. 1 and 2. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 403 Plesiacanthoceras cf. hellsanum (Stephenson, 1953) Plate 6, figs. 26-28 1953 Mammites bellsanus Stephenson, p. 204 (pars.), pi. 49, fig. 3; pi. 51, figs. 8-1 1. 1971 Mammitesl bellsanus Stephenson; Kennedy, p. 122. 1990 Plesiacanthoceras bellsanum (Stephenson, 1953); Kennedy and Cobban, p. 135, pi. 2, figs. 5-8; pi. 12, fig. 9; test-fig. 23c. Types. The holotype is USNM 105983, paratypes are USNM 105984-6, from the Templeton Member of the Woodbine Formation on a branch of Cornelius Creek, 4-3 km north 50° east of Bells, Grayson County, Texas. One of the syntypes, USNM 105986, is a Metoicoceras latoventer Stephenson, 1953. Plesiacanthoceras wyomingense zone. Material. USNM 423741 from USGS Mesozoic locality D5947, Dunveganoceras pondi zone. Discussion. Mammites? bellsanus is a Plesiacanthoeeras as discussed by Cobban (1987^) and Kennedy and Cobban (1990). One of the syntypes (Stephenson 1953, pi. 51, fig. 11) has smooth, non-tuberculate innermost whorls. A specimen from USGS Mesozoic locality D5974 also has a smooth nucleus, preceding a strongly ribbed and tuberculate stage that confirm this distinctive ontogenetic development. USNM 423741 has only the faintest traces of ornament to a whorl height of 6 mm. Coiling is very involute with a tiny, deep, conical umbilicus. The whorl section is depressed with flattened, subparallel flanks and a broadly rounded venter; the only decoration is prorsiradiate growth lines. Ornament appears abruptly after this smooth stage. Small umbilical bullae give rise to narrow, straight, prorsiradiate ribs, singly or in pairs, with shorter intercalated ribs. All ribs bear well-developed, conical, inner ventrolateral tubercles, linked by a broad, blunt, prorsiradiate rib to prominent outer ventrolateral clavi. A low, broad, transverse rib links these to a very weak siphonal clavus. When compared to P. wyomingense of comparable size (PI. 6, figs. 36 and 37) the only significant difference is the presence of tubercles in P. wyomingense at a stage where P. hellsanum is still smooth (compare PI. 6, figs. 52-54 and PI. 6, figs. 26-28). Occurrence. Plesiacanthoceras wyomingense zone of north-central Texas and, possibly, Dunveganoceras pondi zone of Wyoming. Genus dunveganoceras Warren and Stelck, 1940 Type species. Acanthoceras albertense Warren, 1930, p. 21, pi. 1, figs. 1, 2; by original designation. Dunveganoceras pondi Haas, 1949 Plate 5, figs. 1-5; Plate 6, figs. 43-51 1949 Dunveganoceras pondi Haas, p. 22, pi. 8, figs. 1-5, 8; pi. 9, figs. 1, 3, 4; pis. 10-14; text-figs. 11-13, 16, 17. 1979 Dunveganoceras pondi Haas; Merewether, Cobban and Cavanaugh, pi. 4. 1983 Dunveganoceras pondi Haas; Cobban, p. 12, pi. 15, fig. 1. Types. Holotype is AMNH 26416, the original of Haas 1949, pi. 8, fig. 1, pi. 9 figs. 1, 4; Haas mentions 28 speeimens that are presumed to be paratypes, all from the basal part of the Cody Shale near Greybull, Wyoming, Dunveganoceras pondi zone. Material. USNM 423742 to 423746 from USGS Mesozoie loeality D4466, Dunveganoceras pondi zone. Description. D. pondi is a very large species reaehing a diameter in exeess of 400 mm. It ditfers from other speeies of the genus when adult by virtue of having ribs that are depressed over the mid-line of the venter on the mature body chamber. The early whorls are rather poorly known from Haas’ original work, but the present series of specimens reveals previously unknown details. The smallest speeimen referred to the species is USNM 404 PALAEONTOLOGY, VOLUME 33 423744 (PI. 6, figs. 43^5), only 5 mm in diameter. The shell is globose and highly involute with a depressed, reniform whorl section. The only ornament is distant radial flank ribs, 4 per half whorl, terminating in strong conical inner ventrolateral tubercles. There is no ventral ornament at this small diameter. USNM 423745 is 15 mm in diameter (PI. 6, figs. 46-48). Coiling is very involute, with a tiny, deep umbilicus and a depressed reniform intercostal whorl section. At the smallest diameter seen the ornament is weak, but strengthens rapidly to give a depressed polygonal costal section. There are 9-10 prorsiradiate primary ribs per whorl that may or may not arise at feeble umbilical bullae and alternate irregularly with shorter intercalated ribs to give a total of 9-10 ribs per half whorl. All ribs bear strong, conical, inner ventrolateral tubercles. A broad blunt rib projects slightly forward to strong, clavate, outer ventrolateral tubercles, linked across the venter by a low, broad, transverse rib. At the smallest diameter visible there is a low siphonal ridge, beyond there is only a faint trace of siphonal clavi. USNM 423746 (PI. 6, figs. 49-51) is 22 mm in diameter, the coastal whorl section polygonal and depressed, with a whorl breadth to height ratio of 0-86. There are approximately 14-15 coarse ribs on the outer whorl, with umbilical, inner, and outer ventrolateral tubercles as already described; feeble siphonal tubercles are present throughout. USNM 423742 (PI. 5, figs. 1-3) is 47 mm in diameter, with the following proportions : Wb : 43-6; Wh : 53 4; Wb;Wh : 0 82; U :14-2. There are 17 ribs on the outer whorl corresponding to 8 umbilical bullae that decline markedly as size increases and from which the ribs arise singly or in pairs, with occasional intercalated ribs. Conical inner and clavate outer ventrolateral tubercles are present, but there is no trace of a siphonal row. USNM 423743 (PI. 5, figs. 4, 5), a fragment with a maximum preserved whorl height of 30 mm, shows a change to clavate inner ventrolateral tubercles and has a pronounced siphonal ridge, accentuated between the outer ventrolateral clavi. Discussion. The style of ribbing and tuberculation, asymmetry of outer ventrolateral clavi, siphonal ridge and transient siphonal tubercles are all features shared by Dunveganoceras pondi and Plesiacanthoceras wyomingense of similar size (compare PI. 6, figs. 36-51 and PI. 6, figs. 52-54), and there can be little doubt that the former genus gave rise to the latter. Occurrence. D. pondi zone of Wyoming, Montana, Iowa and, possibly. South Dakota, Kansas and Colorado. Subfamily mammitinae Hyatt, 1900, p. 588 (= Buchiceratinae Hyatt, 1903, p. 26; Metoicoceratidae Hyatt, 1903, p. 115; Fallotitinae Wiedmann, 1960, p. 741) Genus metoicoceras Hyatt, 1903, p. 115 Type species. By subsequent designation by Shimer and Shrock, 1944, p. 591 : Ammonites swcdlovi Shumard, 1860, p. 591. Metoicoceras sp. A Plate 5, figs. 10-12, 17-22 Types. Figured specimens USNM 423748-423752, from the Belle Fourche Shale at USGS Mesozoic locality D5947 in Weston County, Wyoming. EXPLANATION OF PLATE 5 Figs. 1-5. Dunveganoceras pondi Haas, 1949. 1-3, USNM 423742; 4, 5, USNM 423743, from USGS Mesozoic locality D4466, D. pondi zone. Figs. 6-9. Metoicoceras aflf. praecox Haas, 1949. 6, 7, USNM 423753; 8, 9, USNM 423754, from USGS Mesozoic locality D4462, D. pondi zone. Figs. 10-12, 17-22. Metoicoceras sp. A. 10-12, USNM 423748; 17-19, USNM 423749 ; 20-22, USNM 423750, all from USGS Mesozoic locality D5947, D. pondi zone. Figs. 13-16, 23—38. Metoicoceras mosbyense Cobban, 1953. 13-16, USNM AlMSl 23-25, USNM 423758; 26, 27, USNM 423759; 28-30, USNM 423760; 31, 32, USNM 423761 ; 33, 34, USNM 423762; 35, 36, USNM 423763; 37, 38, USNM 423764, all from USGS Mesozoie locality D8314, M. mosbyense zone. Figs. 16, 28, 29 are x 2; fig. 30 is x 3; the remainder are x 1. y. PLATE 5 KENNEDY and COBBAN, Dunveganoceras and Metoicocerus 406 PALAEONTOLOGY, VOLUME 33 Material. Five well-preserved, uncrushed specimens from a limestone concretion. Much of the shell material is retained. Dimensions (costal) D USNM 423748 24-3 (100) USNM 423750 25-3 (100) USNM 423749 32-7(100) Wb 12-6 (51-9) 12-2 (48-2) 15-7 (48-0) Wh 12-5 (51-4) 14-4 (56-9) 16.3 (49-8) Wb.Wh 1-0 0-85 0-96 U 3-4(14-0) 2-5 (10-0) 6-4(19-6) Description. Coiling involute with small umbilicus of moderate depth. Umbilical wall flattened, umbilical shoulder broadly rounded. Intercostal whorl section oval with greatest breadth low on the flanks. Costal section with greatest breadth at umbilical bullae, whorl breadth to height ratio 0-85- 1-0, with rounded, convergent flanks and venter concave between outer ventrolateral clavi. There are 20-22 ribs per whorl between 14 and 35 mm diameter. Primary ribs arise at the umbilical seam and may or may not develop from umbilical bullae, from which ribs arise singly or in pairs; intercalated ribs arise around mid-flank. Conical inner ventrolateral tubercles are present at the smallest diameters visible but are lost by 18 mm diameter in larger specimens. All ribs bear strong outer ventrolateral clavi, linked across the venter by a strong transverse rib. A weak siphonal tubercle is present to as much as 25 mm diameter. Discussion. These specimens closely resemble inner whorls of Metoicoceras latoventer Stephenson, 1953, (p. 209, pi. 53, figs. 1-9; pi. 54, figs. 9-11) from the Woodbine Formation of north Texas in their whorl inflation, ribbing style and presence of a siphonal clavus. The Wyoming specimens differ, however, in the very early loss of inner ventrolateral tubercles, which persist to the end of the adult phragmocone in the Texas material. The early loss of these tubercles is like that of the early whorls of M. praecox Haas, 1949, (p. 15, pis. 5-7, text-figs. 5-9). The present material probably represents an undescribed form that we are referring to as sp. A. EXPLANATION OF PLATE 6 Figs. 1-22. Alzadites incomptus gen. et sp. nov. 1-8, paratype USNM Alillb', 10-15, paratype USNM 4221121 \ 9, lb-22, holotype USNM 423725. All specimens are from USGS Mesozoic locality D12052, Sciponoceras gracile zone. Figs. 23-25. Scaphites (Scaphites) sp. USNM 423802, from USGS Mesozoic locality 22871, Plesiacanthoceras wyomingense zone. Figs. 26-28. Plesiacanthoceras cf bellsanum (Stephenson, 1953). USNM 423741, from USGS Mesozoic locality D5947, Dunveganoceras pondi zone. Figs. 29-31. Hamites salebrosus Cobban, Hook and Kennedy, 1989. USNM 423786, from USGS Mesozoic locality D8314, Metoicoceras mosbyense zone. Figs. 32-35. Sumitomoceras sp. USNM 423747, from USGS Mesozoic locality D 12052, S. gracile zone. Figs. 36, 37, 52-54. Plesiacanthoceras wyomingense (Reagan, 1924). 36, 37, USNM 388161; 52-54; USNM 388159, from USGS Mesozoic locality 22871, P. wyomingense zone. Figs. 38^2. Metoicoceras sp. A. 38, 39, USNM 423751 ; 40-42, USNM 423752, from USGS Mesozoic locality D5947, D. pondi zone. Figs. 43-51. Dunveganoceras pondi Haas, 1949. 43-45, USNM 423744; 46-48, USNM 423745; 49-51, USNM 423746; all from USGS Mesozoic locality D4466, D. pondi zone. Figs. 55, 63, 64. Idiohamites bispinosus sp. nov. Paratype USNM 423793, from the Bighorn Basin of Wyoming, D. pondi zone. Figs. 56, 57. Carthaginites aquilonius sp. nov. Holotype USNM 423801, from USGS Mesozoic locality 12650, M. mosbyense zone. Figs. 58-62. Metaptychoceras spp. 58, 61, USNM 423787; 59, 62, USNM 423789, both from the Lower Turonian part of the Greenhorn Formation on the northeastern flank of the Black Hills in western South Dakota. 60, USNM 423788, from USGS Mesozoic locality D8314, M. mosbyense zone. Figs. 5-12, 20-22 are x2; figs. 43^5, 58, 59 are x 3; the remainder are x 1. PLATE 6 KENNEDY and COBBAN, Cenomanian ammonites 408 PALAEONTOLOGY, VOLUME 33 Occurrence. Known only from a single concretion at USGS Mesozoic locality D5947 in the NW| sec. 14, T. 47 N., R. 65 W., Weston County, Wyoming. Upper Cenomanian zone of Dunveganoceras pondi. Metoicoceras aff. praecox Haas, 1949 Plate 5, figs. 6-9; Plate 7, figs. 3-5, 14-16 Compare: 1949 Metoicoceras whitei Hyatt praecox Haas, p. 15, pis. 5-7; text-figs. 5-9. 1952 Metoicoceras praecox Haas; Cobban and Reeside, p. 1017. 1970 Metoicoceras praecox Haas; Ilyin, text-fig. 2e. \911a Metoicoceras cf. M. praecox Haas; Cobban, p. 25, pi. 16, fig. 25; pi. 21, figs. 8 and 9. 1981 Metoicoceras praecox Haas; Kennedy, Juignet and Hancock, p. 58. Types. The holotype of M. praecox is AMNH 26415, the original of Haas 1949, pi. 5, figs. 1, 5, 8; there are five paratypes, all from the basal part of the Cody Shale 91 km east and 1 L2 km north of Greybull, Wyoming, in the north-central part of Township 53 N., Range 92 W. Material. Four specimens, USNM 423753 to 423756, from USGS Mesozoic locality D4462, Dunveganoceras pondi zone. Description. The earliest stages are shown by USNM 423755 and 423756, 18 and 19-5 mm in diameter respectively (PI. 7, figs. 3-5, 14-16). Coiling is very involute, with a tiny, near-occluded umbilicus. The whorl section is depressed, polygonal in costal section. At the smallest diameter visible there are no umbilical bullae. Faint, straight prorsiradiate ribs arise low on the flank and terminate in conical inner ventrolateral tubercles; the venter is smooth. As size increases the ribs strengthen and total 13 per half whorl. They are alternately long and short, and by 16 mm diameter, weak umbilical bullae appear. The inner ventrolateral tubercles, which dominated ornament at the smallest diameter visible, decline in importance, outer ventrolateral clavi appear and strengthen, and are linked to the inner ventrolateral tubercle by a blunt rib. A low, broad swelling links the outer ventrolateral clavi and bears a weak siphonal clavus (PI. 7, figs. 5 and 16). Larger specimens show a change to the style of ornament typical of middle growth, with bullate prorsiradiate primary ribs separated by shorter intercalatories to give an estimated 12 ribs per half whorl. The inner ventrolateral clavi are lost by a whorl height of 7 mm, although the outer ventrolateral clavi remain prominent, and the siphonal clavus is present to an estimated 25 mm diameter (PI. 5, figs. 6-9). Discussion. The earliest development of ornament of M. praecox has not been described, but the innermost whorls of a topotype have the same development of ornament as that of the specimens from locality D4462 except that the inner ventrolateral tubercles are lost at a smaller diameter. Ribs on the inner whorls of topotypes are also broader and more rounded than those on similar-sized specimens from locality D4462. Two of the four specimens from this locality have parts of body chambers, and it is possible that we are dealing with some diminutive species closely allied to M. praecox. Until more conclusive material is available, we are referring the specimens from locality D4462 to M. aff. praecox. Occurrence. Known from a single limestone concretion in the Belle Fourche Shale in the NEj sec. 24, T. 47 N., R. 65 W., Weston County, Wyoming. Probably low in the Upper Cenomanian zone of Dunveganoceras pondi. Metoicoceras mosbyense Cobban, 1953 Plate 5, figs. 13-16, 23-38; Plate 7, figs. 1 and 2 1953 Metoicoceras mosbyense Cobban, p. 48, pi. 6, figs. 1-14; pi. 7, figs. 1-3. 1953 Metoicoceras muelleri Cobban, 1953, p. 49, pi. 6, figs. 15, 16; pi. 8, figs. 1-7; pi. 9. 1957 Metoicoceras defordi Young, p. 1169, pi. 149, figs. 1-8; text-fig. 1a, e, g, i. non 1960 Metoicoceras muelleri Cobban; Wiedmann, p. 720. non 1964 Metoicoceras muelleri Cobban; Wiedmann, p. 115. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 409 non 1967 Metoicoceras aff. mosbyense Cobban; Collignon, p. 35, pi. 19, fig. 3. 1970 Metoicoceras mnelleri Cobban; Ilyin, text-fig. 2b. 1973 Metoicoceras cf. M. defordi Young; Cobban and Scott, p. 75. 1977 Metoicoceras mnelleri Cobban; Kauffman, p. 258, pi. 21 ; pi. 22, figs. 17, 18. non 1978 Metoicoceras mnelleri Cobban; Wiedmann and Kauffman, pi. 6, fig. 3. 1979 Metoicoceras defordi Young; Merewether, Cobban and Cavanaugh, pi. 2, figs. 17, 20-22. non 1980 Metoicoceras mnelleri Cobban; Wiedmann, pi. 6, fig. 3. 1989 Metoicoceras mosbyense Cobban; Cobban, Hook and Kennedy, p. 43, figs. 85c-t, 861, m. Type. The holotype is USNM 108315, from the Mosby Sandstone Member of the Belle Fourche Shale of east- central Montana, Metoicoceras mosbyense zone. Material. USNM 423757 to 423765, from USGS Mesozoic locality 8314. M. mosbyense zone. Discussion. The present collections show the early development from 8 mm diameter onwards. At this size the coiling is very involute, with a tiny, near-occluded umbilicus (PI. 5, figs. 13-16, 23-32; PI. 7, figs. 1 and 2) flank ribs, no umbilical bullae and strong, conical inner ventrolateral tubercles, and weaker, clavate outer ventrolaterals. There is no trace of a siphonal row. As size increases the dominance of the inner ventrolateral tubercles declines and the outer ventrolaterals become more important, while shorter ribs with outer ventrolateral tubercles only alternate with the primaries for a short interval, although all ribs have a complete set of inner and outer ventrolateral tubercles beyond 12 mm diameter. The largest specimens in the collection are 54 and 62 mm diameter. The former is a compressed individual that has lost its inner ventrolateral tubercles (PI. 5, figs. 35 and 36); the latter is stouter with more robust ornament and feeble inner ventrolateral tubercles visible to 22 mm whorl height (PI. 5, figs. 37 and 38). The early ontogenetic stages of M. mosbyense are immediately separable from those of M. latoventer and M. praecox, which have siphonal tubercles. They are more like those of M. geslinianum (d’Orbigny, 1850), where shorter intercalated ribs without inner ventrolateral tubercles may also be present below diameters of 8 mm, and there is an earlier growth stage with prominent spinose inner ventrolateral tubercles only (PI. 7, fig. 13). As discussed by Cobban, Hook and Kennedy (1989), mature M. mosbyense are dimorphic with the type material of mosbyense representing the probable microconch and the type material of M. mnelleri which is larger, more involute and weakly ornamented, as the macroconch. M. defordi Young, 1957 (p. 1 169, pi. 149, figs. 1-8; text-fig. 1a, e, g, i) from the Upper Cenomanian of Apache County, Arizona is based on microconchs that we cannot separate from those of M. mosbyense and we regard it as a synonym. Occurrence. Metoicoceras mosbyense zone of Montana and Wyoming, south to southwest New Mexico and Arizona, where it is commonly identified as M. defordi Young, 1957. Wiedmann (1964, 1980) records the species (as M. mnelleri) from northern Spain, but his figured specimen belongs to some other species. Genus cryptometoicoceras nov. Derivation of name. Kryptos (Greek): hidden, referring to the obscure origins of the genus. Type species. Cryptometoicoceras mite gen. et sp. nov., Dnnveganoceras pondi zone, Wyoming. Diagnosis. Very small, adult at 12 mm or less. Very involute, with tiny occluded umbilicus. Whorl section as wide as high with flattened, subparallel flanks and broadly arched venter. Phragmocone ornamented by distant, low, broad ribs that terminate at conical inner ventrolateral tubercles at the ventrolateral shoulder. Inner ventrolateral tubercles decline on adult body chamber, small rounded outer ventrolateral tubercles present on first part, thereafter disappearing. Last part of body chamber before aperture with tiny clavi on sharp ventrolateral angles linked across venter by low. 410 PALAEONTOLOGY, VOLUME 33 blunt, convex ribs. Suture very simple with elements only slightly incised. E is broad, E/L comparable and asymmetrically bifid, L broad and bifid, L/U.^ and auxiliary saddles on umbilical lobe entire. Discussion. This remarkable genus is interpreted as a progenic dwarf derivative of Metoicoceras praecox, described above, with which it co-occurs. The earliest ornamented stages of M. praecox have conical inner ventrolateral tubercles only, after which outer ventrolateral and siphonal clavi appear (PI. 7, figs. 3-5, 14-16). In Cryptometoicoceras all of the phragmocone corresponds to the first stage, the outer ventrolaterals are only transiently present on the first part of the adult body chamber, after which the distinctive mature ornament appears (PI. 7, figs. 21-27). The adult Cryptometoicoceras is only 12 mm in diameter, the largest M. praecox seen are up to 138 mm in diameter. There can be no doubt that these specimens of Cryptometoicoceras are adult, for they show modified body chamber ornament and the holotype has the last few sutures crov,'ded together. There are obvious similarities to Nannometoicoceras Kennedy, 1988 (p. 63, pi. 11, figs. 1-24; text- fig. 8a) with Metoicoceras acceleration Hyatt, 1903 (p. 127, pi. 14, figs. 1-11) from the Upper Cenomanian Sciponoceras gracile zone of north-east Texas as type species, a progenic dwarf derivative of Metoicoceras gesliniamon (d’Orbigny, 1850). Being derived by paedomorphic processes from the same genus they both have adult phragmocones with features of the nuclei of their ancestor. But whereas Cryptometoicoceras has only flank ribs and conical inner ventrolateral tubercles, Nannometoicoceras has weak to strong, flexuous primary ribs with up to three EXPLANATION OF PLATE 7 Figs. 1, 2. Metoicoceras mosbyense Cobban, 1953. USNM 423765, from USGS Mesozoic locality 12650, M. mosbyense zone. Figs. 3-5, \M\6. Metoicoceras afi. praecox \\a.?LS, 1949. 3-5, USNM 423755; 14— 16, USNM 423756, both from USGS Mesozoic locality D4462, Dimveganoceras pondi zone. Figs. 6-9, 13. Metoicoceras gesliniamon (d’Orbigny, 1850). 6-9, USNM 423773, from USGS Mesozoic locality 23062; 13, USNM 423722, from the Britton Formation 2-25-2-1 km (F5-1'8 miles) southeast of Britton, Ellis County, Texas, both Sciponoceras gracile zone. Figs. 10-12. Buccinammonites minimus gen. et sp. nov. Holotype USNM 423770, from USGS Mesozoic locality 23062, 5. gracile zone. Figs. 17-20. Nannometoicoceras nemos sp nov. Holotype USNM 423768, from USGS Mesozoic locality 12740, M. mosbyense zone. Figs. 21-27. Cryptometoicoceras mite gen. et sp. nov. 21-24, holotype, USNM 423766; 25-27, paratype USNM 423767, both from USGS Mesozoic locality D4462, D. pondi zone. Figs. 28-31, 38^0, 48. Nannometoicocerasl glabrum sp. nov. Holotype USNM 423769, from USGS Mesozoic locality D 12052, 5. gracile zone. Figs. 32-37, 54, 55, 59-62. Hamites cimarronensis (Kauffman and Powell, 1977). 32, 33, USNM 423774; 34, USNM 423775; 35, USNM 423776; 36, 37, USNM 423777; 54, 55, USNM 423779; 59, USNM 423780; 60, USNM 423781 ; 61, USNM 423782; all from USGS Mesozoic locality 22871, Plesiacanthoceras wyomingense zone. Figs. 41^7, 49-53, 56-58, 65-68. Idiohamites bispinosus sp. nov. 41, paratype USNM 423794; 43-45, paratype USNM 423795; 46, 52, 53, paratype USNM 423796; 47, paratype USNM 423778; 49-51, paratype USNM 423797; 56, paratype USNM 423798; 57, 58, paratype USNM 423799; 65-67, holotype USNM 423792; 68, paratype USNM 423800. Figs. 41^5, 49-51 are from USGS Mesozoic locality D4466, D. pondi zone. Figs. 46, 52, 53, 56-58 are from USGS Mesozoic locality D5947, D. pondi zone. Figs. 65-67 are from USGS Mesozoic locality 22871, P. wyomingense zone. Fig. 68 is from USGS mesozoic locality D4466, D. pondi zone. Figs. 63, 64, 69-71. Idiohamites pulchellus sp. nov. 63, 64, holotype, USNM 423790; 69-71, paratype USNM 423791, both from USGS Mesozoic locality 22871, P. wyomingense zone. Figs. 1-5, 15-16, 21-25, 38^0, 48-51, 60, 61 are x2; figs. 6-13 are x3; the remainder are x 1. PLATE 7 51 1' I «s,g;K- 14 KENNEDY and COBBAN, Cenomanian ammonites 412 PALAEONTOLOGY, VOLUME 33 intercalatories, concial inner and outer ventrolateral tubercles, the latter projected adaperturally of the former, or tubercles only. Body chambers of Nannometoicoceras have primary ribs that are bullate or not with 2-3 intercalatories between, and conical to clavate inner and clavate outer ventrolateral tubercles that persist to the end of the body chamber. Occurrence. Dunveganoceras pondi zone, Wyoming. Cryptometoicoceras mite gen. et sp. nov. Plate 7, figs. 21-27 Types. Holotype is USNM 423766; paratype USNM 423767, from USGS Mesozoic locality D4462, D. pondi zone. Derivation of name. Mite, small. Diagnosis. With the characters of the genus. Description. The holotype is an incomplete adult lacking the adapical part of the body chamber. Its essential characteristics are incorporated in the generic diagnosis. Paratype USNM 423767 is a body chamber fragment of comparable size to the holotype. It has ribs at the adapical end with both inner and outer ventrolateral tubercles, those at the adapical end have lost the outer ventrolateral tubercles and are markedly strengthened, suggesting that this specimen too is an adult. Occurrence. As for types. Genus nannometoicoceras Kennedy, 1988 Type species. Metoicoceras acceleratum Hyatt, 1903, p. 127, pi. 14, figs. 1-1 1 . Upper Cenomanian Sciponoceras gracile zone of north-east Texas. Nannometoicoceras nanos sp. nov. Plate 7, figs. 17-20 Derivation of name. Nanos (Greek) : a dwarf. Types. Holotype is USNM 423768, from USGS Mesozoic locality 12740, Metoicoceras mosbyense zone. Diagnosis. Adult at 12-13 mm diameter. Late phragmocone and early body chamber with conical inner and outer ventrolateral tubercles. Late body chamber with smooth flanks and strong, nontuberculate ventral ribs. Description. Holotype is a complete adult no more than 13 mm diameter. Coiling very involute with minute, near-occluded umbilicus. Whorl section compressed (whorl breadth to height ratio 0-6 approximately), with flattened, subparallel flanks, narrowly rounded ventrolateral shoulders and a flattened venter. No umbilical bullae on phragmocone or body chamber. Weak, distant, long and short ribs alternate more-or-less regularly. All terminate in a conical inner ventrolateral tubercle, of which there are five or six on the first half of the outer whorl. Corresponding to these are minute, feebly clavate, outer ventrolateral tubercles. Tubercles disappear on last section of body chamber, where there are five broad, blunt ventral ribs preserved just before the adult aperture. Last few sutures are crowded, indicating maturity, and are very simple, with narrow E/L and broad, bifid L. Discussion. Small size, absence of umbilical bullae and of strong flank and ventral ribs on the greater part of the body chamber immediately distinguish N. nanos sp. nov., from the type species N. KENNEDY AND COBBAN; CENOMANIAN MICROMORPHIC AMMONITES 413 acceleratum (Kennedy, 1988, p. 67, pi. 11, figs. 1-24; text-fig. 8a). There are obvious similarities to Cryptometoicoceras mite gen. et sp. nov. (p. 41 1, PI. 7, figs. 21-27) which has the same terminal body chamber ornament, but N. nanos sp. nov. has inner and outer ventrolateral tubercles on the phragmocone whereas C. mite lacks the outer ventrolateral. The adult phragmocone of N. nanos sp. nov., closely resembles that of juvenile Metoicoceras moshyense (PI. 7, figs. 1 and 2) of which it is presumed to be a progenic dwarf derivative. Occurrence. As for type. Nannometoicocerasl glahrum sp. nov. Plate 7, figs. 28-31, 38^0, 48 Derivation of name. Glaber (Latin): smooth. Type. Holotype is USNM 423769 from USGS Mesozoic locality D12052, Sciponoceras gracile zone. Diagnosis. Small, adult at 1 1 mm diameter. Phragmocone and early body chamber with distant, conical, outer ventrolateral tubercles followed by a shallow ventral constriction, four on the first half of the outer whorl. Middle section of body chamber with low folds and constrictions, final part with four coarse ventrolateral and ventral ribs, the venter markedly flattened before the adult aperture. Description. Holotype is adult at 1 1 mm diameter. Coiling is very involute with a tiny umbilicus. Whorl section compressed with a broadly rounded venter on the phragmocone. Phragmocone and early body chamber smooth except for distant, conical, outer ventrolateral tubercles, four on the first half of the outer whorl. Venter of middle part of body chamber with low folds. Last part of body chamber has flattened venter in costal section, with four coarse, ventral and ventrolateral ribs separated by deep, wide interspaces. Ribs are transverse on the venter, concave on the ventrolateral shoulder, and connected to the umbilicus by a delicate lira. Sutures not seen. Discussion. Absence of inner ventrolateral tubercles distinguish this species from all other Nannometoicoceras and from Cryptometoicoceras mite gen. et sp. nov. The absence of inner ventrolaterals suggests that, given additional material it might merit subgeneric status within Nannometoicoceras. There is no clear indication of its evolutionary origins. Occurrence. As for type. Genus buccinammonites nov. Derivation oj name. Buccina (Latin); trumpet, in reference to the trumpet-like flared aperture. Type species. Buccinammonites minimus gen. et sp. nov., Sciponoceras gracile zone, southeastern Montana. Diagnosis. Minute, adult at 4-5 mm diameter. Very involute with tiny umbilicus. Whorl section depressed with flattened subparallel flanks, venter broadly rounded. Five ribs per half whorl on the phragmocone are broad and coarse and terminate in strong conical inner ventrolateral tubercles. This style of ornament persists onto the first part of the adult body chamber. Aperture preceded by narrow crowded ribs with minute ventrolateral tubercles. Mouth border with flare that extends out for 30% of the whorl height in a trumpet-like aperture. Suture with very simple, little-incised bifid elements. Discussion. Minute size, simple ornament and the extraordinary flared aperture distinguish Buccinammonites gen. nov., from all other described taxa. The coiling and proportions of the 414 PALAEONTOLOGY, VOLUME 33 phragmocone whorls recall those of Cryptometoicoceras gen. nov., but the body chamber ornament is utterly distinctive. Like Nannometoicoceras and Cryptometoicoceras, the phragmocone ornament of Buccmammonites indicates that it is a progenic dwarf derivative of some other metoicoceratine, but whether this was Metoicoceras (PI. 7, figs. 1-9, 14-16) where the earliest ornamented stage has only flank ribs and inner ventrolateral tubercles or the already dwarf Nannometoicoceras or Cryptometoicoceras, we cannot say. Occurrence. As for genus. Buccinammonites minimus gen. et sp. nov. Plate 7, figs. 10-12 Derivation of name. Minimus (Latin): least. Types. Holotype USNM 423770, paratype USNM 423771, from USGS Mesozoic locality 23062, Sciponoceras gracile zone. Diagnosis. With the characters of the genus. Discussion. The holotype is a complete adult showing all the diagnostic features of the species. Paratype USNM 423771 is incomplete at 4-8 mm diameter and has shallow constrictions. Occurrence. As for types. Suborder ancyloceratina Wiedmann, 1966, p. 54 Superfamily turrilitaceae Gill, 1871, p. 3 Family hamitidae Gill, 1871, p. 3 Genus hamites Parkinson, 1811, p. 145 (= Torneutoceras Hyatt, 1900, p. 586 (objective synonym); Stomohamites Breistroffer, 1940, p. 85; Hamitella Breistroffer, 1947, p. 100 (84) nom. nov. pro. Helicoceras d’Orbigny, 1842, p. 611, non Koenig, 1825, pi. 19). Type species. Hamites attenuatus J. Sowerby, 1814, p. 137, pi. 61, figs. 4 and 5, by the subsequent designation of Diener 1925, p. 65. Hamites cimarronensis (Kauffman and Powell, 1977) Plate 7, figs. 32-37, 54, 55, 59-62 1953 Hamitesl sp. Stephenson, p. 197. 1977 Stomohamites simplex cimarronensis Kauffman and Powell, p. 97, pi. 9, figs. 1, 3, 4; text-figs. 5 and 6. 1990 Hamites cimarronensis (Kauffman and Powell, 1977); Kennedy and Cobban, p. 140, pi. 15, figs. 11, 13, 15, 17, 19-21. Type. Holotype is USNM 167160, the original of Kauffman and Powell 1977, pi. 9, fig. 1, from USGS Mesozoic locality 30235 in Cimarron County, Oklahoma, and from the Hartland Member of the Greenhorn Limestone, late Cenomanian. Material. More than 100 fragments from USGS Mesozoic localities D5947, D4462, D4466 and D22871, Plesiacanthoceras wyomingense to Dunveganoceras pondi zones. Discussion. Fragments of this species are very common in the present collections. The very earliest developmental stages generally lack the protoconch and consist of a straight, smooth, slowly expanding shaft up to 12 mm long (PI. 7, figs. 59-61). This shaft is terminated by a curved section. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 415 at which point ribbing develops. The largest fragment seen has a whorl height of 10 mm, and is part of a curved portion (PI. 7, fig. 62). It and fragments of intermediate size suggest an elliptical coiling with at least three shafts. Ornament consists of fine, dense, prorsiradiate ribs that are weakest on the dorsum and strongest over the venter, straight to feebly convex, with a rib index of 6-8. Most of the material is much smaller than that of the European H. simplex d’Orbigny, 1842 (p. 550, pi. 134, figs. 12-14) making comparison difficult, but large fragments are always much more densely ribbed than the comparable stage of H. simplex. Occurrence. Widespread in the US Western Interior from Montana and Wyoming south to New Mexico and northeast Texas, ranging from Conlinoceras tarrantense to Dunveganoceras pondi zones. Hamites salebwsus Cobban, Hook and Kennedy, 1989 Plate 6, figs. 29-31 1989 Hamites salehrosus Cobban, Hook and Kennedy, p. 57, fig. 95bb, ee, ii. Type. Holotype is USNM 423786 from USGS Mesozoic locality D 12069 in Apache County, New Mexico, Twowells Sandstone Tongue of Dakota Sandstone, Metoicoceras moshyense zone. Material. USNM 423786 from USGS Mesozoic locality D8314. M. moshyense zone. Description and Discussion. Specimen is a body chamber fragment 46 mm long with a maximum preserved whorl height of 17-5 mm. The whorl section is a compressed oval with a whorl breadth to height ratio of 0-7. The rib index is 5, the ribs weakened somewhat on the dorsum but strengthening across the flanks, where they are straight and prorsiradiate, and passing straight across the venter. Large size, coarse ribbing, compression and low rib density distinguish H. salehrosus from all other described species. Occurrence. Calycoceras canitaurinum and M. moshyense zones. New Mexico, Arizona and Wyoming. Genus metaptychoceras Spath, 1926, p. 80 Type species. Ptychoceras smithi Woods, 1896, p. 74, pi. 2, figs. 1 and 2, by original designation. Metaptychoceras spp. Plate 6, figs. 58-62 Compare : 1977 Hemiptychoceras sp. Kauffman and Powell, p. 99, pi. 9, fig. 5; text-fig. 7. Material. USNM 423788, from USGS Mesozoic locality D8314, M. moshyense zone; USNM 423787 and 423789 from the Greenhorn Formation of western South Dakota. Discussion. USNM 423788 is an external mould of two shafts, with a maximum preserved length of 7-5 mm. The smaller shaft is curved and ornamented by strong, straight, weakly prorsiradiate ribs; the rib index is 4—5. The larger shaft has coarser ribs; the rib index is 5. Metaptychoceras is generally uncommon in the US Western Interior. It occurs in the middle Cenomanian of Oklahoma (Kauffman and Powell 1977) and Wyoming (the present record), C. canitaurinum zone of New Mexico, S. gracile zone of Colorado and northeast Texas, lower Turonian of the Dallas area in Texas and the northeast flank of the Black Hills in western South Dakota (PI. 6, figs. 58, 59, 61, 62) and is locally frequent in the upper Turonian in the Waco area in central Texas and Chispa Summit in Trans-Pecos Texas. 416 PALAEONTOLOGY, VOLUME 33 Family anisoceratidae Hyatt, 1900, p. 587 (= Algeritidae Spath, 1925, p. 190) Genus idiohamites Spath, 1925 Type species. Hamites tuberculatus J. Sowerby, 1818, p. 30, pi. 216, figs 4 and 5, by original designation. Idiohamites pulchellus sp. nov. Plate 7, figs. 63, 64, 69-71 1973 Idiohamites sp. Cobban and Scott, p. 50, pi. 13, figs. 1-4. Derivation of name. Diminutive of pulcher (Latin): beautiful. Types. Holotype is USNM 423790, paratype is USNM 423791, from LfSGS Mesozoic locality 22871, Plesiacanthoceras wyomingense zone. Diagnosis. Compressed Idiohamites with narrow prorsiradiate ribs, rib index 9. Periodically strengthened ribs have sharp lower lateral and ventrolateral tubercles, typically with 3-4 non- tuberculate ribs between. Discussion. The holotype is a slightly curved fragment 25 mm long and shows a transition from an initially bituberculate section. Paratype USNM 423971 is much larger, with a whorl height of 11-5 mm and a rib index of 9, showing the same differentiation into stronger tuberculate ribs separated by up to five non-tuberculate ones. I. pulchellus sp. nov., is easily distinguished from 7. bispinosus sp. nov., to be described below, which is the only other species known from the Western Interior, and which lacks lateral tubercles. Occurrence. Conlinoceras tarrantense and Acanthoceras muldoonense zones of southeastern Colorado, Plesiacanthoceras wyomingense zone of Montana. Idiohamites bispinosus sp. nov. Plate 6, figs. 55, 63, 64; Plate 7, figs. 41^7, 49-53, 56-58, 65-68 Types. Holotype is USNM 423792, from USGS Mesozoic locality 22871, P. wyomingense zone. Paratypes USNM 423796, 423798 and 423799 are from USGS Mesozoic locality 5947 ; paratype USNM 423800 is from USGS Mesozoic locality D4466; paratypes USNM 423794, 423795 and 423797 are from USGS Mesozoic locality D4462; paratype USNM 423793 is from the Bighorn Basin of Wyoming; all from the Dunveganoceras pondi zone. Diagnosis. Planispirally coiled in an open ellipse. Whorl section compressed oval with crowded, prorsiradiate, feebly convex ribs. One, sometimes two, linked ribs bear sharp ventrolateral tubercles on moulds that are the bases of septate spines linked across the venter by a pair of looped ribs. There are 1-3 non-tuberculate ribs between the tuberculate ones. Discussion. Specimens range from 2 to 10 mm whorl height. At the smallest sizes there may be some irregularities in ribbing with up to five non-tuberculate ribs between tuberculate ones, and the very earliest stages may lack tubercles. The septate spines are perfectly preserved in USNM 423798 (PI. 7, fig 56). What may be an adult of the species is represented by USNM 423793, from the Upper Cenomanian Dunveganoceras pondi zone near Greybull, Wyoming (PI. 6, figs. 55, 63, 64). The three fragments illustrated were originally part of a single specimen. The smallest piece closely resembles the type series. The middle piece, from a whorl height of 8 5 to 13 mm has a rib index of 8, the ribs KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 417 flexuous and prorsiradiate, and nearly all with a ventral tubercle. The largest fragment, preserved to a whorl height of 22 mm has a rib index of 16, with tuberculate ribs separated by up to three non- tuberculate ones on the first part, after which the ribs are all non-tuberculate for the final 65 mm. Absence of lateral tubercles easily distinguishes I. bispinosus sp. nov., from /. pulchellus sp. nov., the only other species known from the Western Interior. Occurrence. Plesiacanthoceras wyomingense and Dunveganoceras pondi zones of Wyoming and Montana. Family turrilitidae Gill, 1871, p. 3 (= Pseudhelicoceratinae Breistroffer, 1953, p. 1350) Genus carthaginites Pervinquiere, 1907, p. 96 Type species. Turrilites {Carthaginites) kerimensis Pervinquiere, 1907, p. 101, pi. 4, fig. 18. Carthaginites aquilonius sp. nov. Plate 6, figs. 56 and 57 Derivation of name. Aquilonius (Latin): northerly. Type. Holotype is USNM 423801 from USGS Mesozoic locality 12650, Metoicoceras moshyense zone. Description. Specimen consists of one and a quarter whorls, with a maximum preserved whorl height of 6-9 mm. Apical angle low, with seam between successive whorls only slightly indented. 17-18 low, broad, prorsiradiate ribs arise at the upper edge of the outer whorl face and strengthen into small, sharp tubercles a little above the middle of the outer whorl face. A broad, smooth, depressed zone separates these from a row of small, blunt, aperturally displaced tubercles low on the outer whorl face. These show feeble spiral elongation and lie at a sharp angulation in the whorl profile and pronounced facet that extends to the lower edge of the outer whorl. The sharp edge between outer and lower whorl faces is feebly crenulate, the crenulations corresponding in position and number to the lowest row of tubercles. Discussion. The imperfectly exposed suture shows E/L occupying the upper outer and part of the upper whorl face, confirming this as a Carthaginites rather than Neostlingoceras. The presence of strong tubercles immediately distinguishes it from C. krorzaensis Dubourdieu, 1953 (p. 66, pi. 49, figs. 49-52; text-fig. 20). C. kerimensis Pervinquiere, 1907, (p. 101, pi. 4, fig. 18) is based upon a minute specimen with only 6-7 mid-flank tubercles per whorl, and no lower row. Carthaginites virdense Cobban, Hook and Kennedy (1989) has 12-13 tuberculate ribs in the upper row, and those in the lower row twice as numerous. Superficially similar is Neostlingoceras kottlowski Cobban and Hook, 1981 (p. 26, pi. 4, figs. 1-28), which has a third row of tubercles on the underside of the whorl. Occurrence. As for type. Superfamily scaphitaceae Gill, 1871, p. 3 (nom. transl. Wright and Wright, 1951, p. 13, ex Scaphitidae Gill, 1871, p. 3) Family scaphitidae Gill, 1871, p. 3 Subfamily scaphitinae Gill, 1871, p. 3 {nom. transl. Wright, 1953, p. 473, ex Scaphitidae Gill, 1871, p. 3) Genus and subgenus scaphites Parkinson, 1811, p. 3. Type species. Scaphites equalis J. Sowerby, 1813, p. 53, pi. 18, figs. 1-3. 418 PALAEONTOLOGY, VOLUME 33 Scaphites (Scaphites) sp. Plate 6, figs. 23-25 Material. USNM 423802, from USGS Mesozoic locality 22871, Plesiacanihoceras wyomingense zone. Description. Specimen is a wholly septate fragment of less than half a whorl with a maximum preserved whorl height of 9 mm. Coiling very involute, with a depressed, reniform whorl section. Narrow primary ribs arise at the umbilical seam and secondary ribs are inserted between them, both high and low on the flank. Ribs are narrow, straight and prorsiradiate, and cross the venter nearly straight. The last half of the fragment bears small, conical, ventrolateral tubercles on four out of nine primary ribs. Each tubercle gives rise to a pair of ribs that loop across the venter to the tubercle on the other flank. Discussion. This is the only Scaphites {Scaphites) known from the Western Interior below the zone of Sciponoceras gracile (see Cobban, 1952 for details). Occurrence. As for material. Acknowledgements. R. E. Burkholder of the US Geological Survey, Denver took the photographs. We thank the staff of the Geological Collections, University Museum, Oxford, and the Department of Earth Sciences, Oxford, UK for technical assistance. Kennedy acknowledges the financial support of the Natural Environment Research Council (UK), the Royal Society, and the Astor Fund (Oxford). REFERENCES ADKINS, w. s. 1928. Handbook of Texas Cretaceous fossils. University of Texas Bulletin, 2838, 385 pp. 37 pis. ARKHANGUELSKY, A. D. 1916. [The Upper Cretaceous Molluscs of Turkestan] Part I. Trudy Geologicheskago Komiteta, 152, 1-57, pis. 1-8. [In Russian.] BAYLE, E. 1878. Fossiles principaux des terrains. Explication de la Carte Geologique de France, 4(1) (Atlas), 158 pis. Service de la Carte Geologique detaillee, Paris. BREiSTROFFER, M. 1940. Revision des ammonites du Vraconien de Salazac (Gard) et considerations generates sur ce sous-etage albien. Travaux du Laboratoire de Geologic de la Faculte des Sciences de I 'Universite de Grenoble, 22 (1938-9), 17-171 (1-101 in separates). 1947. Sur les zones d’ammonites dans EAlbien de France et d’Angleterre. Ibid. 26, 17-104 (1-88 in separates). 1953. L’evolution des Turrilitides Albiens et Cenomaniens. Compte Rendu Hebdomadaire des Seances de r Academie des Sciences, Paris (D), 237, 1349-1351. CFiOFFAT, p. 1898. Recueil d’etudes paleontologiques sur la faune cretacique du Portugal. I. Especes nouvelles ou peu connues. Deuxieme serie, Les Ammonees du Bellasien, des couches a Neolobites Vibrayeanus, du Turonien et du Senonien. Commision des Travaux Geologiques du Portugal (1898), 41—86, pis. (cephalopodes) 3-22. CHUDEAU, R. 1909. Ammonites du Damergou (Sahara meridionale). Bulletin de la Societe Geologique de France (4), 9, 67-71, pis. 1-3. COBBAN, w. A. 1952. Scaphitoid cephalopods of the Colorado Group. United States Geological Survey Professional Paper, 239, 39 pp., 21 pis. (1951 imprint). 1953. Cenomanian ammonite fauna from the Mosby Sandstone of central Montana. United States Geological Survey Professional Paper, 243-D, 45-55, pis. 6—12. 1961 . The ammonite family Binneyitidae Reeside in the Western Interior of the United States. Journal of Paleontology, 35, 737-758, pis. 87-89. 1972. New and little-known ammonites from the Upper Cretaceous (Cenomanian and Turonian) of the Western Interior of the United States. United States Geological Survey Professional Paper, 699, 24 pp., 18 pis. (1971 imprint). 1977u. Characteristic marine molluscan fossils from the Dakota Sandstone and intertongued Mancos Shale, west-central New Mexico. United States Geological Survey Professional Paper 1009, 30 pp., 29 pis. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 419 1977^. Fossil mollusks of the Dakota Sandstone and intertongued Mancos Shale of west-central New Mexico, hi San Juan Basin HI. New Mexico Geological Society Guidebook, 28, 213-220. 1983. Molluscan fossil record from the northeastern part of the Upper Cretaceous seaway. Western Interior. United States Geological Survey Professional Paper, 1271, 24 pp., 15 pis. 1984. Mid-Cretaceous ammonite zones. Western Interior, United States. Bulletin of the Geological Society of Denmark, 33, 71-89. 1987n. The Upper Cretaceous (Cenomanian) ammonites Metengonoceras dumhli (Cragin) and M. acutum (Hyatt). Bulletin of the United States Geological Survey, 1690, C1-C7, 3 pis. 1987/). Some middle Cenomanian (Upper Cretaceous) acanthoceratid ammonites from the Western Interior of the United States. United States Geological Survey Professional Paper, 1445, 28 pp„ 13 pis. and HOOK, s. c. 1981. New turrilitid ammonite from mid-Cretaceous (Cenomanian) of southwest New Mexico. New Mexico State Bureau of Mines and Mineral Resources Circular, 180, 22-29, pi. 4. and KENNEDY, w. J. 1989. Upper Cretaceous rocks and ammonite faunas of southwestern New Mexico. New Mexico Bureau of Mines and Mineral Resources Memoir, 45, 137 pp. and REESiDE, j. B. 1952. Correlation of the Cretaceous formations of the Western Interior of the United States. Bulletin of the Geological Society of America, 63, 1011-1044, 1 pi. and SCOTT, G. R. 1973. Stratigraphy and ammonite fauna of the Graneros Shale and Greenhorn Limestone near Pueblo, Colorado. United States Geological Survey Professional Paper, 645, 108 pp., 41 pis. (1972 imprint). COLLiGNON, M. 1937. Ammonites Cenomaniennes du sud-ouest de Madagascar. Annales geologiques du Service des Mines. Madagascar, 8, 29-72, pis. 1-1 1. 1967. Les cephalopodes cretaces du bassin cotier de Tarfaya. Notes et Mthnoires, Service des Mines et de la Carte Geologkpie de Maroc, 175, 7-148, 35 pis. (1966 imprint). CRAGIN, F. w. 1893. A contribution to the invertebrate paleontology of the Texas Cretaceous. Texas Geological Survey, 4th Annual Report (1892), 139-246, pis. 24^6. DiENER, c. 1925. Ammonoidea neocretacea. Fossiliurn Catalogus (I : Animalia), 29, 244 pp. DUBOURDiEU, G. 1953. Ammonites nouvelles des Monts du Mellegue. Bulletin du Service de la Carte Geologicpie de r Algerie. P serie, Paleontologie, 16, 76 pp., 4 pis. FURON, R. 1935. Le Cretace et le Tertiaire du Sahara soudanais (Soudan, Niger, Tchad). Archives du Museum National dHistoire Naturelle, Paris (6) 13, 1-96, pis. 1-7. GILL, T. 1871. Arrangement of the Families of Mollusks. Smithsonian Miscellaneous Collections, 227, xvi-|- 49 pp. GROSSOUVRE, A. DE 1894. Recherches sur la craie superieure. 2. Paleontologie. Les ammonites de la craie superieure. Memoires pour servir d I'explication de la Carte Geologicpie Detaillee de la France, 264 pp., 39 pis. (1893 imprint). HAAS, o. 1949. Acanthoceratid Ammonoidea from near Greybull, Wyoming. Bulletin of the American Museum of Natural History, 93 (1), 1-39, pis. 1-5. 1963. Paracanthoceras wyomingense (Reagan) from the Western Interior of the United States and from Alberta (Ammonoidea). American Museum Novitates, 2151, 1-19. 1964. Plesiacanthoceras, new name for Paracanthoceras Haas, 1963 non Furon, 1935. Journal of Paleontology, 38, 610. HYATT, A. 1889. Genesis of the Arietidae. Smithsonian Contributions to Knowledge, 673, xi-t-238 pp., 14 pis. 1900. Cephalopoda, pp. 502-604. In zittel, k. a. von, 1896-1900, Textbook of Palaeontology, transl. EASTMAN, c. R. Macmillan, London and New York. 1903. Pseudoceratites of the Cretaceous. Monographs of the United States Geological Survev, 44, 351 pp., 47 pis. ILYIN, V. D. 1970. [Top Cenomanian ammonites of S.E. Central Asia] In sheetskoi, e. k. [Stratigraphy and Palaeontology of central Asia]. Trudy Vsesoyuznogo Nettyanogo Nauchno-Issledovaterskogo Geologo- Razvedochnogo Instituta (VNIGRI), 69, 10-23, 196-197, 9 pis. [In Russian]. JUKES-BROWNE, A. J. 1896. VI. Critical remarks on some of the fossils. In jukes-browne, a. j. and hill, w. A delimitation of the Cenomanian : being a comparison of the corresponding beds in southwestern England and northern France. Quarterly Journal of the Geological Society of London, 52, 99-178, pi. 5. KAUFFMAN, E. G. 1977. Illustrated guide to biostratigraphically important Cretaceous macrofossils. Western Interior basin, USA. Mountain Geologist, 14, 225-274, 32 pis. 420 PALAEONTOLOGY, VOLUME 33 HATTiN, D. E. and POWELL, J. D. 1977. Stratigraphic, paleontologic and palaeoenvironmental analysis of Upper Cretaceous rocks of Cimarron County, northwestern Oklahoma. Geological Society of America Memoir, 149, 149 pp., 12 pis. and POWELL, j. d. 1977. In kauffman, e. g., hattin, d. e. and powell, j. d. 1977, q.v. KENNEDY, w. J. 1971. Cenomanian ammonites from southern England. Special Papers in Palaeontology, 8, v+ 133 pp., 64 pis. 1988. Late Cenomanian and Turonian ammonite faunas from north-east and central Texas. Special Papers in Palaeontology, 39, 131 pp., 24 pis. and COBBAN, w. a. 1990. Woodbine and basal Eagle Ford (Cenomanian) ammonite faunas from Texas. Palaeontology, 33, 75-154. and HOOK, s. c. 1988. Middle Cenomanian (Late Cretaceous) molluscan fauna from the base of the Boquillas Formation, Cerro de Muleros, Dona Ana County, New Mexico. New Mexico Bureau of Mines and Mineral Resources Bulletin, 114, 35^4. , JuiGNET, p. and HANCOCK, J. M. 1981. Upper Cenomanian ammonites from Anjou and Vendee, western France. Palaeontology, 24, 25-84, pis. 3-17. and WRIGHT, c. w. 1985. Evolutionary patterns in late Cretaceous ammonites. Special Papers in Palaeontology, 33, 131-143. KIRKLAND, J. I. and COBBAN, w. A. 1986. Cunningtoniceras arizonense n. sp., a large acanthoceratid ammonite from the upper Cenomanian (Cretaceous) of eastern central Arizona. Hunteria, 1, 1-14, 8 pis. KNECHTEL, M. M. and PATTERSON, s. M. 1962. Bentonite deposits of the northern Black Hills district, Wyoming, Montana, and South Dakota. Bulletin of the United States Geological Survey, 1082-M, 893- 1030. KOENIG, c. D. E. 1820-5. Icones Fossilium Sectiles. London, 4 pp., 9 pis. KULLMAN, J. and wiEDMANN, J. 1970. Significance of sutures in phytogeny of Ammonoidea. Paleontological Contributions, University of Kansas. Paper 44, 1-32. LEWY, z., KENNEDY, w. J. and CHANCELLOR, G. R. 1984. Co-occurrence of Metoicoceras geslinianum (d’Orbigny) and Vascoceras cauvini Chudeau (Cretaceous Ammonoidea) in the Southern Negev (Israel) and its stratigraphic implications. Newsletters in Stratigraphy, 13, 67-76. MATSUMOTO, T. 1969. In MATSUMOTO, T., MURAMOTO, T and TAKAHASHi, T. Selected Acanthoceratids from Hokkaido. Memoirs of the Faculty of Science, Kyushu University. (Series D) Geology, 19, 251-296, pis. 25-38. and MURAMOTO, T. 1969. In matsumoto, t., muramoto, t. and takahashi, t. Selected Acanthoceratids from Hokkaido, Memoirs of the Faculty of Science, Kyushu University. (Ferry D) Geology, 19, 251-296, pis. 25-38. MEREWETHER, E. A., COBBAN, w. A. and CAVANAUGH, E. T. 1979. Frontier Formation and equivalent rocks in eastern Wyoming. Mountain Geologist, 16, 67-101, 4 pis. MOREMAN, w. L. 1927. Fossil zones of the Eagle Ford of north Texas. Journal of Paleontology, 1, 89-101, pis. 13-16. 1942. Paleontology of the Eagle Ford group of north and central Texas. Journal of Paleontology 16, 192-220, pis. 31-34. NEUMAYR, M. 1875. Die Ammoniten der Kreide und die systematik der Ammonitiden. Zeitschrift der Deutschen Geologischen Gesellschaft, 27, 854—942. ORBiGNY, A. d’. 1840-2. Paleontologic frangaise : Terrains cretaces. 1. Cephalopodes. Masson, Paris, 1-120 (1840); 121^30 (1841); 431-662 (1842), 148 + 3 pis. 1850. Prodrome de Paleontologic stratigraphique universelle des animaux mollusques et rayonnes, 2. Masson, Paris, 428 pp. PARKINSON, J. 1811. Organic Remains of a Former World, 3. Sherwood, Neily and Jones, London, xvi + 479 pp., 22 pis. PERViNQUiERE, L. 1907. Etudes de paleontologie tunisienne. 1. Cephalopodes des terrains secondaires. Carte Geologique de la Tunisie, v + 438 pp., 27 pis. de Rudeval, Paris. REAGAN, A. B. 1924. Cretacic Mollusca of the Pacific Slope. Pan-American Geologist, 41, 179-190, pis. 18-21. REESiDE, J. B. JR. 1928. Cephalopods from the lower part of the Cody Shale of Oregon Basin, Wyoming. United States fJeological Survey Professional Paper, 150-A, 1-19, pis. 1-8. and COBBAN, w. a. 1954. Ammonite accumulations in the Cretaceous Mowry and Aspen Shales. Science, 119,255. 1960. Studies of the Mowry Shale (Cretaceous) and contemporary formations in the United States and Canada. United States Geological Survey Professional Paper, 355, 1-126, 58 pis. KENNEDY AND COBBAN: CENOMANIAN MICROMORPHIC AMMONITES 421 SHARPE, D. 1853-7. Description of the fossil remains of Mollusca found in the Chalk of England. I. Cephalopoda. Palaeontographical Society Monographs. 68 pp., 27 pis. 1-26, pis. 1-10 (1853); 27-36, pis. 1 1-16 (1855); 37-68, pis. 17-27 (1857). ' SHiMER, H. w. and shrock, r. r. 1944. Index fossils of North America. Wiley, New York, ix + 837 pp. SHUMARD, B. F. 1860. Descriptions of new Cretaceous fossils from Texas. Transactions of the Academy of Sciences of St. Louis, 1, 590-610. SOWERBY, J. 1812-22. The Mineral Conchologv of Great Britain. 1, pis. 1-9 (1812), pis. 10-44 (1813), pis. 45-78 (1814), pis. 79-102 (1815); 2, pis. 103-114 (1815), pis. 115-150 (1816), pis. 151-186 (1817), pis. 187-203 (1818); 3, pis. 204-221 (1818), pis. 222-253 (1819), pis. 254-271 (1820), pis. 272-306 (1821); 4, pis. 307-318 (1821), pis. 319-383 (1822). The author, London. SPATH, L. F. 1923. On the ammonite horizons of the Gault and contiguous deposits. Summary of Progress of the Geological Survey of Great Britain (for 1922), 139-149. 1925. On Upper Albian Ammonoidea from Portuguese East Africa, with an appendix on Upper Cretaceous ammonites from Maputoland. Annals of the Transvaal Museum, 11, 179-200, pis. 28-37. 1926. On new ammonites from the English Chalk. Geological Magazine, 63, 77-83, table. 1931. A monograph of the Ammonoidea of the Gault. Part 10. Palaeontographical Society Monographs, 411^42, pis. 43^8. STEPHENSON, L. w. 1953. Larger invertebrate fossils of the Woodbine Formation (Cenomanian) of Texas. United States Geological Survey Professional Paper, 242, iv + 226 pp., 59 pis. (1952 imprint). 1955. Basal Eagle Ford fauna (Cenomanian) in Johnson and Tarrant counties, Texas. United States Geological Survey Professional Paper, 274-C, 53-67, pis. 4—7. THOMEL, G. 1972. Les Acanthoceratidae Cenomaniens des chaines subalpines meridionales. Memoires de la Societe Geologique de France (N.S.), 116, 204 pp., 88 pis. WARREN, p. s. 1930. Three new ammonites from the Cretaceous of Alberta. Transactions of the Royal Society of Canada (3), 24 (4), 2\-26, ph. \^. and STELCK, c. r. 1940. Cenomanian and Turonian faunas, Pouce-Coupe District, Alberta and British Columbia. Transactions of the Royal Society of Canada (3), 34 (4), 143-152, pis. 1^. WEDEKIND, R. 1916. Uber Lobus, Suturallobus und Inzision. Zentralhlatt fiir Mineralogie, Geologic und Paldontologie (for 1916), 185-195. wiEDMANN, J. 1960. Le Cretace superieur de I’Espagne et du Portugal et ses cephalopodes. Comptes Rendus du Congres des Societes Savantes- Dijon, 1959: Collogue sur le Cretace superieur fran^ais, 709-764, 8 pis. 1964. Le Cretace superieur de I’Espagne et du Portugal et ses cephalopodes. Estudios Geologicos Instituto Lucas Mallada, 20, 107-148. 1966. Stammesgeschichte und system der posttriadischen ammonoideen; ein iiberblick. Neues Jahrbuch fiir Geologic und Paldontologie. Abhandlungen, 125, 49-79, pis. 1-2; 127, 13-81, pis. 3-6. 1980. Mid-Cretaceous Events Iberian Field Conference 77. Guide II. Partie Itineraire geologique a travers le Cretace Moyen des Chames vascogotiques et Celtiberiques (Espagne du Nord). Cuadernos de Geologia Iberica, 5, 127-214, 12 pis. (1979 imprint). and KAUFFMAN, E. G. 1978. Mid-Cretaceous biostratigraphy of northern Spain. Annales du Museum cTHistoire Naturelle de Nice, 5, III, 1-34, 12 pis. (1976 imprint). WOODS, H. 1896. The Mollusca of the Chalk Rock. Part 1. Quarterly Journal of the Geological Society of London, 52, 68-98, pis. 2-4. WRIGHT, c. w. 1953. Notes on Cretaceous ammonites. 1 . Scaphitidae. Annals and Magazine of Natural History, 12 (6), 473-476. and KENNEDY, w. J. 1980. Origin, evolution and systematics of the dwarf acanthoceratid Protacanthoceras Spath, 1923 (Cretaceous Ammonoidea). Bulletin of the British Museum {Natural History). Geology, 34, 65-107. 1981. The Ammonoidea of the Plenus Marls and the Middle Chalk. Palaeontographical Society Monographs, 148 pp., 32 pis. 1987. The Ammonoidea of the Lower Chalk. Palaeontographical Society Monographs, Part 2, pp. 127-218, pis. 41-55. and WRIGHT, E. V. 1951. A survey of the fossil Cephalopoda of the Chalk of Great Britain. Palaeontographical Society Monographs. 41 pp. YOUNG, K. 1957. Cretaceous ammonites from eastern Apache County, Arizona. Journal of Paleontology, 31, 1167-1174, pis. 149-150. 1958. Cenomanian (Cretaceous) ammonites from Trans- Pecos Texas. Journal of Paleontology. 32, 286-294, pis. 39^0. 422 PALAEONTOLOGY, VOLUME 33 ziTTEL, K. A. 1884. Handbuch der Paldontologie ... 1, 2 (Lief 3), Cephalopoda, 329-522. R. Oldenbourg, Munich and Leipzig. 1895. Grundziige der Paldontologie (Paldozoologie). R. Oldenbourg, Munich and Leipzig, vii + 972 pp. W. J. KENNEDY Geological Collections University Museum Parks Road Oxford 0X1 3PW. W. A. COBBAN U.S. Geological Survey Mail Stop 919 Box 25046 Federal Center Denver Colorado 80225 U.S.A. Typescript received 23 November 1988 Revised typescript received 8 August 1989 AN APPLICATION OF CRITICAL POINT DRYING TO THE COMPARISON OF MODERN AND FOSSILIZED SOFT TISSUES OF FISHES DAVID M. MARTI LL and LIZ HARPER Abstract. Critical-point-dried samples of recent biological soft tissues can be used to make accurate comparisons with exceptionally well-preserved fossil material. This technique has distinct advantages over thin sections of biological tissues, as palaeontologists are often more familiar with observing three-dimensionally preserved material. This technique offers important opportunities for comparative taphonomic and anatomical studies, especially for palaeontologists working with exceptionally well-preserved soft tissues such as may be found in early diagenetic concretions. Critical point drying (CPD), the drying of biological tissues at the critical temperature and pressure of carbon dioxide, is a technique which allows the examination in three dimensions of soft tissues under a vacuum, and hence is ideal for scanning electron microscopy (SEM). The technique produces a minimum of artefacts during the drying process compared with freeze drying and drying in air, and the CP-dried tissues may be coated with a variety of electrically conductive materials. Biological specimens dried in air are greatly distorted by surface tensional processes, unless they are composed of particularly rigid biopolymers such as chitin. CPD involves taking the liquid in which a specimen is immersed to its critical point; that is the temperature (T^,) and pressure (P^.) at which the liquid changes imperceptibly from a liquid to a gas (or vice versa). At this point, surface tension is zero, and fluid may be released from the tissue causing a minimal amount of morphological change. The procedure is based on Anderson (1951), who also describes in more detail the principles behind the method. An alternative to CPD is freeze drying (Boyde 1974). Freeze drying offers excellent opportunities for examining fractured surfaces, but it also has a number of serious drawbacks, including severe tissue damage due to ice crystal formation (Hayat 1978), and we do not recommend its use here. We demonstrate an application of CPD to a palaeontological problem, an attempt to resolve an early diagenetic event preserved in pre-compaction concretions. Martill (1988) suggested that the preservation of phosphatized soft tissues in Cretaceous fishes from the Santana Formation of Brazil took place very rapidly, and prior to burial. Since burial rates in the deposit are unknown, however, the time interval in which phosphatization took place is also unknown. The soft tissues discovered by Martill show a number of features related to the decomposition and collapse of the tissues. Thus it was believed that phosphatization accompanied active decay rather than post-dating a physico- chemically induced interruption of the decomposition process, but this supposition remained to be tested. On this basis, it should be possible to mimic decomposition processes taking place on the sea floor under laboratory conditions, to sample the decaying tissue at various intervals, and observe changes in tissue morphology with time. CP-dried samples of partly decomposed tissue could then be examined using the SEM, and direct comparisons made between recent and fossil material. METHOD A rainbow trout (Salmo gairdneri Richardson), 20 cm in length, was killed and placed in a bath of fresh sea water with a specific gravity of 102 (normal salinity). Samples of striated muscle, gill filament with secondary lamellae, and ovaries were removed at death. Further samples were removed at hourly intervals up to 4 hours. I Palaeontology, Vol. 33, Part 2, 1990, pp. 423-428, 1 pl.| © The Palaeontological Association 424 PALAEONTOLOGY, VOLUME 33 and then at 6 hours, 24 hours, 48 hours, and 7 days. The experiment was carried out at room temperature in a fume cupboard. The samples of tissue were immediately fixed for 1-5 hours in 1 % osmium tetroxide buffered in phosphate at pH 7-2 to prevent further post-mortem degradation. After two washes in the buffer the tissue was dehydrated by immersion in a series of acetones (30%, 50% - 10 minutes each, then 70%, 85%, 90%, 95%, and 100% for 15 minutes each). The 100% acetone wash was repeated. At this stage all the water in the tissue had been replaced by acetone (alternative methods of dehydration use ethanol or freons, as described by Cohen 1979). The samples were then placed in porous vessels flooded with acetone, and loaded into the CP-drying bomb (text-fig. 1). This was then filled with liquid carbon dioxide and flushed through three or four times over a 3-5 hour period to expel all the acetone from the samples, replacing it with COj. The half-filled bomb was gently heated to bring the CO,^ to its critical point and then vented (T^ = 31 °C, = 72-9 atmospheres). Biological tissues cannot be CP-dried using water as the ambient fluid as the higher temperature and pressure required is not practical (water = 374 °C, = 217-7 atmospheres). The CP-dried samples were sputter-coated with gold and examined using a Hitachi S-520 SEM at 20 kV. CO2 in Vent air out i t TEXT-FIG. 1. Schematic diagram of the critical point-drying bomb. Samples are placed in porous beakers supported on a boat. The apparatus is heated by a hot water circulatory system. CO2 enters through a regulatory valve. The operation can be observed through a clear Perspex safety shield in front of a silica window. Fossilized soft tissues were obtained from teleost fish in pre-compaction concretions from the Santana Formation (Lower Cretaceous, Aptian/Albian), of the Chapada do Araripe, Ceara, north-east Brazil. They were prepared by partial digestion in 10®/o acetic acid so that the fish skeleton remained partly within the concretion. This allows the soft tissues to be more readily identified since they remain in situ within the fossil. The fossil material was also sputter-coated with gold and examined on the same SEM, but at 1 5 kV to minimize beam damage {note\ high kV at high magnifications have been found to damage some phosphatized soft tissues). RESULTS The morphology of all the recent CPD soft tissues examined changed with time after death. Here we discuss only the gill filaments and associated secondary lamellae (text-fig. 2). In the recent samples, secondary lamellae of the gills at time zero are erect and well spaced (plate 1, fig. 1). At high magnification, detail on the surface of individual cells is seen at time zero, but this begins to break down after only one hour (plate 1, fig. 8). As time of decay progresses, individual secondary MARTILL AND HARPER; CRITICAL POINT DRYING 425 TEXT-FIG. 2. Diagram showing position of secondary lamellae and gill filaments in the rainbow trout (Salmo gairdneri). a. Trout with operculum removed to reveal gill arches, b. Single gill arch with gill filaments, c. Detail of gill arch with three filaments showing position of secondary lamellae. Based loosely on Hughes and Morgan (1973). lamellae collapse, possibly due to lack of blood pressure and also because of gravity (plate 1, fig. 3). This collapse produces a prominent kink at the base of each lamella (plate 1, fig. 9). After two hours, the epithelial cells of the recent secondary lamellae begin to detach, leaving only a connective tissue lining (plate 1, fig. 6). Epithelial tissue covering the gill ray begins to detach from the gill ray after four hours. A similar detachment is also seen in the fossil material (plate 1, fig. 4). Gills sampled after seven days show very little identifiable soft tissue remaining. However, bone surfaces show numerous colonies of microorganisms, including spherical and rod-shaped bacteria (plate 1, fig. 10). A large number of the fossil gills show kinks at the base of each lamella (plate 1, fig. 7a, b), although many of the fossil secondary lamellae are preserved fully erect and well spaced (plate 1, fig. 2a, b). Individual epithelial cells are only rarely preserved in the fossil material, but they can sometimes be seen on the surface of gill rays. They do not show details of cell wall ultrastructure, but this may be because of an inability of this preservational style to resolve these features. 426 PALAEONTOLOGY, VOLUME 33 DISCUSSION Timing of phosphatization Results of the experimental decomposition of the trout show that very rapid morphological change occurs in delicate soft tissues such as gill filaments and secondary lamellae. The exceptional preservation of these tissues in fossil fish from the Santana Formation, together with the preservation of artefacts brought about by decomposition, show that phosphatization was clearly an extremely early and rapid diagenetic event. Temperature and salinity are likely to have been major influences controlling the rate of reactions and hence rate of decay of the fossil fish. The temperature of the Santana sea floor has yet to be detemined isotopically, but its position within the palaeo-tropics, and its generally shallow aspect suggest warm, rather than cold, bottom-water conditions. Present-day salinity has been assumed for the Santana Formation for the purpose of this experiment, although a number of authors have considered the salinity to range from fresh to hypersaline (see Martill 1988 for a review). Comparisons between Recent and fossil material at normal salinity and room temperature suggest that phosphatization must have taken place within the first five hours of death of the fish, although preservation of epithelial cells in place suggests a possibly earlier event, c. 1-2 hours. This indicates that very high concentrations of dissolved phosphate were available for rapid precipitation on to nucleating sites. Although the oxygen level of the Santana sea floor is not known precisely, the presence of arthropods and rare molluscs shows that anoxia had not been reached. However, Allison (1988) has shown that most carcasses in marine environments undergo anoxic decomposition even in well-oxygenated water. We therefore assume that oxygen levels are not as important as hitherto believed. Use of CPD by palaeontologists This example shows how critical point drying can be used to help solve a specific palaeontological problem, in this case, one of taphonomic and diagenetic importance. However, there are numerous other taphonomic, taxonomic, and palaeobiological applications for the technique. Palaeobotanists EXPLANATION OF PLATE 1 Comparison between fossil soft tissues and recent trout gills after various times of decomposition. (Fossil material prefixed DM is currently held by the first author at the Open University, but will be transferred to the University of Leicester, Department of Geology.) Fig. 1. Section of CPD fresh trout gill filament with erect secondary lamellae, x 150. Fig. 2. DM 50. fl. Section of fossil gill filament from Santana Formation, x 150. b. Detail of four erect fossil secondary lamellae showing preservation fabric of phosphatic microspheres, x 700. Fig. 3. Three CPD trout gill filaments after one hour of decomposition. Two of the filaments show detached portions of epithelial cells revealing connective tissue lining of blood vessels, x 70. Fig. 4. DM 63. Fossil gill filaments showing same features observed in fig. 3, x 70. Fig. 5. CPD trout gill filament after one hour of decomposition, with post-mortem break-up following boundary between epithelial cells, x 300. Fig. 6. CPD trout gill filament with secondary lamellae after four hours of decomposition. The epithelial cells have detached from the lamellae leaving only a connective tissue lining. Collapse has produced kinking of the lamellae at their bases. Some shrivelling of the connective tissue lining of the blood vessel has taken place, x200. Fig. 7. DM 101. a. Fossil gill filament showing similar features to those seen in figure 6, x400. b. Detail of secondary lamellae/blood vessel junction, x 1000. Fig. 8. CPD trout gill after one hour of decomposition, showing collapsed epithelial cells, x 2000. Fig. 9. CPD trout gill filament showing detail of secondary lamellae/blood vessel junction, after four hours decomposition, x 400. Fig. 10. CPD trout gill after seven days of decomposition. All gill tissue has decomposed, only isolated colonies of bacteria remaining on exposed bone surfaces. PLATE I MARTILL and HARPER, Critical point drying 428 PALAEONTOLOGY, VOLUME 33 familiar with CPD (e.g. Hill 1987) have used it successfully for comparative anatomical studies. Wherever fossil material requires comparison with recent material, especially at high magnifications, this technique allows direct comparisons of three-dimensional material to be made. Very few artefacts are introduced during preparation, although it should be pointed out that there may be some shrinkage of the tissue, and at very high magnifications it is sometimes possible to see osmium tetroxide crystallized on some surfaces from over-osmication. CP-dried material stored in anhydrous conditions has a long shelf life. Bivalve mollusc material prepared by Harper has persisted for at least 12 months without discernible deterioration. Acknowledgements . We thank the Biology Department of the Open University for use of their CPD facility, and Tony King for helpful instruction. Use of the SEM at the Department of Geology, University of Leicester is gratefully acknowledged, with special thanks to Rod Branson. Roy Clements commented on the manuscript. This work was funded by an Open University Research Grant awarded to D. M. L.H. is funded by NERC. ALLISON, p. A. 1988. The role of anoxia in the decay and mineralization of proteinaceous macrofossils. Paleobiology, 14, 139-154. ANDERSON, T. F. 1951. Techniques for the preservation of three dimensional structure in preparing specimens for the electron microscope. Transactions of the New York Academy of Sciences, Series 2, 131, 130-134. BOYDE, A. 1974. Freezing, freeze-fracturing and freeze-drying in biological specimen preparation for the SEM. Scanning Electron Microscopy, 1043-1046. S.E.M. International. AMF O’Hare, Chicago, 1974. COHEN, A. L. 1979. Critical point drying - principles and procedures. Scanning Electron Microscopy, ii 303-324. S.E.M. International. AMF O’Hare, Chicago. HAYAT, M. A. 1978. Introduction to biological scanning electron microscopy, xvii -I- 323 pp. University Park Press, Baltimore, London, Tokyo. HILL, c. R. 1987, Jurassic Angiopteris (Marattiales) from north Yorkshire. Review of Palaeobotany and Palynology, 51, 65-93. HUGHES, G. M. AND MORGAN, M. 1973. The Structure of fish gills in relation to their respiratory function. Biological Reviews 48, 419^75. MARTiLL, D. M. 1988. Preservation of fishes in the Cretaceous of Brazil. Palaeontology, 31, 1-18. REFERENCES Typescript received 1 March 1989 Revised typescript received 27 May 1989 DAVID M. MARTILL LIZ HARPER Department of Earth Sciences The Open University Walton Hall Milton Keynes MK7 6AA, U.K. COMPUTER-AIDED RESTORATION OF A LATE CAMBRIAN CERATOPYGID TRILOBITE FROM WALES, AND ITS PHYLOGENETIC IMPLICATIONS by NIGEL c. HUGHES and Adrian w. a. rushton Abstract. Tectonic deformation is liable to affect the diagnostic characters of fossils, but its effects can be removed with the help of a computer-graphic technique, which is here applied to trilobites for the first time. Dikelocephalusl discoidalis Salter, 1866, with its putative synonym D.l celticus Salter, 1866, is known only from distorted specimens collected from the upper part of the Paraholina spimdosa Biozone in the Dolgellau Formation (upper Cambrian) of North Wales. It has been reconstructed by removing tectonic deformation. D.l discoidalis is now referred to Cermatops Shergold, a member of the Subfamily Iwayaspidinae; this group is considered to be a paraphyletic subgroup within the Family Ceratopygidae. In north-west Europe the Late Cambrian faunas are dominated by olenid trilobites, a specialized group that was adapted to oxygen-deficient environments (Henningsmoen 1957), whereas such cratonic realms as North America, Australia, northern China and Siberia, each supported a diverse and partly endemic suite of genera (Palmer 1977). Besides the agnostids, one of the most widely distributed trilobite groups is the Family Ceratopygidae, members of which are known from most areas yielding late Middle Cambrian to Tremadoc faunas, though despite much recent work the full biostratigraphical potential of the group has yet to be realized. Pwceratopyge is recorded in the English Midlands (Rushton 1983) and Ceratopyge itself from the Tremadoc of North Wales but until now no ceratopygids have been recorded from the Merioneth Series in Wales. However, we here refer Salter’s Dikelocephaliisl species from the Merioneth Series of North Wales to the ceratopygid genus Cermatops Shergold, 1980. Dikelocephalids are common only in North American Trempealeauan deposits from shallow shelf environments (Taylor 1977). Their supposed presence in black shales from North Wales was remarkable, both on account of the slope setting there and because the European P. spimdosa Biozone is correlated with the lower Franconian Taenicephalus Biozone of North America, well below the Trempealeauan. The new assignment negates a suggestion by Conway Morris and Rushton (1988, fig. 3) that dikelocephalids migrated from outer-shelf environments onto the North American craton, but fits well with the known biogeography of ceratopygid trilobites. OCCURRENCE Salter’s ^ Dikelocephalus' material came from the locality ‘Ogof-ddu’, 1 km east of Criccieth, Gwynedd, North Wales, D.l discoidalis being based on cephala and D.l celticus and D.l sp. on pygidia. Ogof-ddu refers to the cliff-section Rhiw-for-fawr (around National Grid reference SH 5135 3795) that extends from the top of the Ffestiniog Flags Formation, through the whole Dolgellau Formation, to the lower part of the Tremadoc (Fearnsides 1910, p. 153); the section encompasses several trilobite biozones. In 1951 officers of the British Geological Survey (BGS) examined the section bed-by-bed and collected fossils throughout; Stubblefield (1953) reported preliminary results. Review of the BGS collection shows that the lowest 13 m of the Dolgellau Formation is referable to the Paraholina spimdosa Biozone. Material of '' Dikelocephaliis' (now Cermatops) was collected only from 8-5 to 9-0 m above the base, that is, entirely within the upper part of the P. spimdosa Biozone. The Cermatops are associated with the following fossils: I Palaeontology, Vol. 33, Part 2, 1990, pp. 429-445, 2 pls.| © The Palaeontologieal Association 430 PALAEONTOLOGY, VOLUME 33 Homagnostus obesus laevis Westergard, Pseiidagnostus cyclopyge (Tullberg), Paraholiua cf. spimdosa (Wahlenberg), Paraholinitesl sp., Lingulella sp., Orusia lenticularis (Wahlenberg) [abundant] and Stenotheca sp. Fearnsides collected a pygidium referable to Cennatops discoidalis (PI. 2, fig. 9) from the stream section below Penmorfa Church (Salter 1866, p. 250). The only other material known to us was collected by Shackleton (1959, p. 222) from the cliff above Cwm-y-ffynnon (SH 5403 5141), 7-5 km west-south-west from the summit of Snowdon and about 14 km N of Ogof-ddu, where rare Cermatops fragments are associated with abundant O. lenticularis and Homagnostus, Pseiidagnostus, Paraholina aff. mohergi Westergard, Paraholinitesl sp. and Maladoidellal ahdita (Salter). Apart from the abundance of O. lenticularis this fauna does not yield clear evidence of the P. spimdosa Biozone, but fragments of P. aff. mohergi are present at the top of the P. spimdosa Biozone at Ogof-ddu. CORRECTION OF DISTORTION Salter’s (1866) Dikeloceplialusl from the Dolgellau Formation are strongly deformed. To assess the generic position of these forms it was desirable to restore the original shape, and to this end we used a computer-graphic method employed at the British Museum (Natural History). The technique involved the digitization of camera lucida or photographic images of the specimens; the digitized images were then displayed on a monitor and progressively adjusted until bilateral symmetry was achieved (see Jefferies et al. 1987 for details). The validity of this method was tested using slabs in which various specimens were preserved in different orientations; some had the sagittal axis parallel to the principal component of strain (x axis of the strain ellipse) whereas in others it was perpendicular or oblique. Bilateral symmetry was restored to each specimen individually and values of id (the proportionate increase of the y co-ordinate relative to the .v co-ordinate) compared. a b c TEXT-FIG. 1. Examples of computer restorations giving bilateral symmetry. Above - camera lucida sketches; below - after restoration, ri^ represents the factor required to restore symmetry to each drawing. HUGHES AND RUSHTON: COMPUTER-AIDED RESTORATION OF TRILOBITE 431 Within individual slabs the value of n- required for each specimen was identical; this validated the use of bilateral symmetry as a criterion for restoring the shape. The program gave an approximation to the original shape (see text-fig. 1 ) but did not produce perfect restorations; in many cases the .v-v plane of the strain ellipse was not exactly the same as the dorso-ventral plane of the specimens, and in some cases pyrite crystals caused local inhomogeneities of strain. Several specimens had been cracked during compression and prior to tectonic distortion. These factors did not, however, significantly limit the use of the method in determining a generalized representation of the overall shape, and this proved important for taxonomic assessment. The bilaterally symmetrical restorations of cranidia and pygidia provided by the computer method were scaled to a standard size using a Rost planvariograph. These images were superimposed and an overall representation of original shape inferred (see text-fig. 2). Because of the complexity of post-mortem deformation, particular weight was given to those specimens which showed least original distortion (for example the pygidium in PI. 2, fig. 9). The three specimens of free cheeks were not analysed on the computer because two of the specimens were comparatively undistorted and also because their shape could largely be inferred from the cranidial restoration. As the sagittal axis of the single hypostome was parallel to the .v axis of strain, it was not possible to restore its shape; the outline of the hypostome in the reconstruction is therefore dotted. In the absence of complete specimens the relative size of cranidium and pygidium was estimated from the size-ranges of the specimens available. Particular features, such as the sculpture and median tubercle, were included in the reconstruction only if they were recognized in at least two specimens. SYSTEMATIC PALAEONTOLOGY Superfamily asaphacea Burmeister, 1843 Family ceratopygidae Linnarsson, 1869 Subfamily iwayaspidinae Kobayashi, 1962 Genus cermatops Shergold, 1980 Type species. C. vieta Shergold (1980, p. 87, pi. 34, figs. 3-11). [Names with -ops are treated as masculine so the specific name is changed here to vietus.] Diagnosis. Ceratopygid trilobites without macropleural pygidial spines (Subfamily Iwayaspidinae), having a subquadrate glabella showing several pairs of furrows and muscle-scars anterior of SI ; palpebral lobe does not reach axial furrow; pygidium transversely semi-oval, the pleural segments having reduced propleural bands. The generic and subfamilial classifications are discussed below. Cermatops discoidalis (Salter, 1866) Plates 1 and 2; text-figs. 1, 2, 3a, 4 1866 Dikelocephalusl (Centropleural) celticiis, n. sp.; Salter, p. 304, pi. 5, figs. 21 and 22. 1866 Dikelocephalusl (Centropleural) discoidalis, n. sp.; Salter, p. 304, pi. 5, figs. 18, 18a, 19. 1866 Dikeloceplialusl (Centropleural) sp.; Salter, p. 305, pi. 5, fig. 20. 1868 Dikeloceplialusl Celticus Sal., D.l discoidalis Sal.; Belt, p. 6 [gives horizon, but incorrectly]. 1914 Dikeloceplialus celticus Salter, D. discoidalis Salter; Walcott, pp. 350, 366 [mentioned as generically indeterminate]. 1919 Dikeloceplialus discoidalis Salter; Lake, p. 115, pi. 14, figs. 2-5 only [not figs. 6 and 7, ? = Lakella invita (Salter)]. 1919 Dikeloceplialus celticus Salter; Lake, p. 116, pi. 14, figs. 8-10. 1935a Briscoia celticus (Salter), B. discoidalis (Salter); Kobayashi, pp. 51-52 [transferred to Briscoia]. 1946 Dikeloceplialus celticus Salter; Lake, p. 343 [discusses similarity to Briscoia]. 1953 " Dikellocephalus' celticus Salter; Stubblefield, p. 56 [discusses horizon], 1988 Briscoial celtica (Salter, 1866); Morris, p. 38 [listed]. 1988 Briscoial discoidalis (Salter, 1866); Morris, p. 38 [listed]. 432 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. Restoration of Cermatops discoidalis (Salter), about x 3. Dorsal view of cranidium: the dorsal terrace-ridges are shown on the left side of the preglabellar field, the border furrow is indicated on the right. Ventral view of right free cheek and hypostome; terrace-ridges are shown on a representative area; hypostomal outline inferred (dashed), course of dorsal suture indicated (pecked). Dorsal view of thoracic segment. Dorsal view of pygidium. HUGHES AND RUSHTON: COMPUTER-AIDED RESTORATION OF TRILOBITE 433 Type material. The lectotype of Z).? discoidalis, selected by Morris 1988, p. 38, is BGS GSM 10214 (PI. 1, fig. 3); it is the original of Salter’s fig. 18 and Lake’s fig. 2. Paralectotypes include cranidia GSM 10209 (Lake’s fig. 3), 10210 (Salter’s fig. 19 and Lake’s fig. 4) and 10216; and free cheeks GSM 10213, 10213A (Salter’s fig. 18a and Lake’s fig. 5) and 10215. The lectotype of celticus, selected by Morris 1988, p. 38, is GSM 10206A (PI. 2, fig. 3) and 10206 (counterparts), the original of Salter’s fig. 22 and of Lake’s fig. 9. The paralectotypes include the original of Salter’s fig. 21 (GSM 10208), and possibly the unfigured specimens GSM 10212, BGS GSd 4587 and two pygidia collected by Homfray (Sedgwick Museum SM A932). Other material. From Ogof- ddu: GSM 85214 (Wyatt-Edgell Coll.), SM A50349-50354 (Fearnsides Coll.), and about forty specimens and fragments collected by S. W. Flester for the Geological Survey in 1951, numbers prefixed Hr. From Penmorfa, SM A51599 (Fearnsides Coll.). From Cwm-y-iTynnon, five specimens presented to the BGS by Professor R. M. Shackleton. Description. Glabella widest (tr.) at midlength (sag.) of LI (= posterior lateral glabellar lobe). Lateral margins subparallel anterior of SI (posterior lateral glabellar furrow). Glabella with rounded anterolateral corners and straight anterior margin. SI furrows crescentic, strongly curved, half of glabellar width, shallow adaxially, deeply incised in distal third, not connecting with axial furrow, anterior part weaker than posterior. S2 deepest distally, transverse, crossing one-fifth of glabellar width; S3 oblique inwards and forwards, one-fifth of glabellar width, deepest at mid-length (tr.). L3 short (exsag). Shallow intercalated furrows subparallel to SO across distal quarter of LI. Median tubercle within bifurcation of SI. Shallow furrows intercalated within L2 and anterior of S3. Glabella lacks sculpture. Occipital furrow deepest distally, shallow in axial third. A low ridge arches anteriorly from posterior border of occipital ring, occupying medial four-fifths of occipital width and extending three-quarters of occipital length (sag.). Shallow transverse furrows run adaxially, about one- third of occipital width (tr.). Axial furrow shallow in front of glabella. Preglabellar field broad, over half as long as pre-occipital glabella, and over twice as wide as occipital ring. Border low, short (sag.), anterior margin slightly angular axially. Terrace-ridges on border face anteriorly. Anterior border furrow crescentic, weakly defined. Preocular sutures diverge forwards at 45 degrees to the sagittal line, then curve in and extend along anterior margin of cephalon. Paradoublural line runs obliquely backwards from anterolateral corner of glabella, subparallel to weak eye ridge. Fixigena narrowest (tr.) opposite L3. Palpebral lobe arcuate, one quarter the length of cranidium, widest (tr.) opposite L2. Palpebral area slopes steeply into axial furrow. Palpebral furrow weakly developed. Posterolateral border area wider (tr.) than occipital ring, straight, narrow (exsag). Postocular suture transverse, subparallel to posterolateral margin. Posterior marginal furrow deeply incised. A shallow furrow runs outwards parallel to posterior margin for four-fifths of width of posterolateral border. Estimated length of cephalic axis in various specimens is 5-1 5 mm. Free cheeks arcuate, border smooth. Genal spine short, its base much narrower (tr.) than width of doublure. Median suture present. Portion of ocular platform within paradoublural line one-fifth of fixigenal medial width (tr.), bounded adaxially by upraised flange. Adaxial margin of doublure sub-parallel to cephalic margin; a slight flexure in its anterior portion presumably accommodated the anterior edge of the hypostome. Doublure extends about three-quarters of the distance from margin to ocular incisure. About twenty terrace-ridges present on doublure, continuous, gently sinuous, steeper slopes facing abaxially, most closely spaced at inner and outer margins. Associated hypostome elongate. Anterior border and anterolateral wing not preserved. Lateral border narrow (tr.) dipping steeply into border furrow; posterior border flatter, longer (sag.) than wide. Median body ovoid. Anterior lobe inflated, posterior lobe gently convex. Median furrow complete, connecting with border furrow. Maculae prominent, ovoid. Sculpture not observed. Number of thoracic segments unknown. Associated thoracic segment shows articulating furrow deeply incised distally, more shallowly in medial third. An arcuate set of fairly continuous terrace-ridges covers the posterior part of the axial ring. Axial furrows parallel to sagittal axis. Pleura gently curved posteriorly. Pleural furrows deeply incised. Pygidium sub-elliptical, wider than long. Margin entire, with slight post-axial emargination developed in larger specimens. Axis narrow, convex, tapering evenly posteriorly, about three-quarters of pygidial length and about one quarter of maximum pygidial width. Articulating half ring short (sag.), crescentic. Axis generally of six rings and terminal piece, a poorly defined seventh ring present in some specimens. A set of arcuate posteriorly facing terrace-ridges, similar to those on the thoracic segment, run from posterior border of axial rings. Ring furrows deeply incised distally, shallow adaxially. Axial furrow deeply incised. Post-axial ridge narrow (tr.) where present, extends to posterior margin. Interpleural furrows narrow, firmly incised, extending almost to pygidial margin. Pleural furrows broad, shallow. Five to seven pleurae present, sixth and seventh poorly defined. First pleura contains equally divided pro- and opisthopleurae; pleural and interpleural furrows 434 PALAEONTOLOGY, VOLUME 33 geniculate at paradoublural line. Subsequent pleura show relative reduction of length and width of propleura. Propleura absent from fifth (and subsequent) pleura, where pleural furrows are undifferentiated from interpleural furrows. Terrace-ridges weakly developed on propleurae. Doublure wide, extending inwards to posterior end of axis, and inwards from anterolateral pygidial margin for half pleural width (tr.). Terrace-ridges of doublure have high relief, steeper slopes facing outwards, distributed most densely along adaxial portion. Estimated length of various pygidia is 5-25 mm. Interpretative remarks 1. Lateral glabellar furrows. The glabellar furrows of C. discoidalis are difficult to interpret because they are variously altered and masked by tectonic compression. Lake thought that the SI furrow was transcurrent, as in other Dikelocephalus. Such a feature is seen only in the most compressed cranidium (PI. 1, fig. 6), and is contradicted by other specimens and our reconstruction. Salter correctly described the obliquity of the furrows -SI sloping inwards and backwards, S2 transverse, S3 inwards and forwards - though neither his figures nor Lake’s show this clearly. Our interpretation (text-fig. 3) is based especially on the lectotype (PI. 1, fig. 3) and on the new material TEXT-FIG. 3. Suggested interpretation of glabellar furrows in a Cermatops discoidalis and b C. vietus. c is C.? tenacella, based on Xiang and Zhang 1985, pi. 41, fig. 11. (e.g. PI. 1, fig. 4). The strongest furrow, with inner ends opposite the glabellar tubercle, is interpreted as SI, as is typical of the Ceratopygidae. The comparatively well-marked furrows opposite the anterior half of the palpebral lobes are homologized with S2 of the primitive asaphine pattern (Fortey and Chatterton 1988). The furrow interpreted as S3 is weaker and lies anterior to the front of the palpebral lobe; it is seen in several specimens but is obscured by a misleading crease in GSM 10209 (PI. 1, fig. 1). A short furrow close to the axial furrow and just in front of S3 is referred to as S4 - a similar furrow is seen in Guozia crassa (text-fig. 5h). Some specimens show weak transverse furrows on glabellar lobes LI and L2. Those on LI are present in some iwayaspidines, as remarked EXPLANATION OF PLATE 1 Figs. 1-9. Cermatops discoidalis (Salter, 1866), all from the Parabolina spinulosa Biozone of Ogof-ddfl, west of Criccieth, North Wales (National Grid ref. SH 5157 3787 approx.). All these specimens are in the collections of the Biostratigraphy Research Group of the British Geological Survey (BGS), Keyworth, Nottingham. All were whitened before photography and, unless otherwise indicated, are internal moulds. 1-6, 9, cranidia. 1, GSM 10209, x3; an Orusia lenticularis lies to the right of the glabella. 2, Hr 921 Pv, x 3. 3, Lectotype, GSM 10214, x3. 4 and 5, Hr 948, showing glabellar furrows, and latex cast of counterpart Hr 948A, showing palpebral lobes. Both x 4. 6, GSM 10210, x4. 7 and 8, free cheeks; 7, GSM 10213, showing the doublure forward of preocular suture and behind it a pyrite infilling between the doublure and the dorsal surface. 8, latex cast of Hr 937, showing small genal spine (slightly retouched). 9, Hr 927B, x 3-2. On the same block as Fig. 2, but oriented at right angles to it. The originals of Figs. 3, 6, 7 were illustrated by Salter 1866, pi. 5, figs. 18, 19, 18a; those of Figs. 1, 3, 6, 7 were illustrated by Lake 1919, pi. 14, figs. 3, 2, 4, 5 respectively. PLATE 1 HUGHES and RUSHTON, Cermatops discoidalis 436 PALAEONTOLOGY, VOLUME 33 below, and in some olenids (Rushton 1982). The L2 furrows are weak in C. discoidalis but there is a pair of stronger impressions in a similar position in C. vietus (Shergold 1980, pi. 34, figs. 3 and 6). 2. Association of the pygidium. Salter assigned the cephalon and pygidium to different species for reasons of caution. Their association in the newer material and their congruence with other Iwayaspidinae indicates that Lake was right to suppose that they belonged to one species. 3. Pygidial shape. The pygidia that Salter described as Z).? celticus differ from those of D.l sp. because the length/ width ratio is greater, the posterior margin is indented and the pleural furrows are more swept back. Lake thought that these were the same species differently compressed; we believe that he was correct. The differences are readily understood if the pygidium was considerably convex and the posterior margin was arched upwards rather than indented. Viewed from above the pygidium is relatively short, the anterior margin straight, the pleurae direct and the posterior arch nearly invisible (text-fig. 4). This is the ‘D.? sp.’ configuration. Viewed obliquely from above and behind, however, the projected length is greater, the anterior margin and pleurae sweep backwards and the posterior arch is more visible (text-fig. 4) - the "celticus' configuration. One reason why the pygidium should appear in two forms is that moulted pygidia could come to rest either on the doublure or upside-down on the dorsal surface; flattening by compaction would then give projections corresponding to the two appearances described above. TEXT-FIG. 4. Sketches to illustrate the differing appearances of the pygidia of Cennatops discoidalis according to whether they were deposited dorsal side up (above) or inverted (below). EXPLANATION OF PLATE 2 Cermatops discoidalis (Salter, 1866). All are from Ogof-ddu, west of Criccieth, North Wales (National Grid ref. SH 5157 3787), except for Fig. 9 which is from Penmorfa Church 2 km west of Tremadog, North Wales (about SH 5418 4030). Figs. 5, 7 and 9 are in the Sedgwick Museum, Cambridge (SM); all the other specimens are in the collections of the Biostratigraphy Research Group of the British Geological Survey. All were whitened before photography and, unless otherwise indicated, are internal moulds. 1-5, 7, 9, 10, pygidia; 6, hypostome; 8, thoracic segment. 1, Hr 925 (external mould), x 3. 2, GSM 10212, x 3. 3, GSM 10206A, X 3. 4, GSM 1021 1, x4. 5, two pygidia showing terrace-lines on the doublure; SM A. 50349, x 3. 6, fragmentary hypostome, Hr 923, x 6. 7, SM A. 933, x 4. 8, thoracic segment, GSM 10216, x 2. 9, least distorted pygidium, SM A. 51 599, from Penmorfa, x 4. 10, latex cast of GSM 10208, x 2. The originals of Figs. 3, 4, 10 were illustrated by Salter 1866, pi. 5, figs. 22, 20, 21, and by Lake 1919, pi. 14, figs. 9, 10, 8 respectively. PLATE 2 HUGHES and RUSHTON, Cermatops discoidalis 438 PALAEONTOLOGY, VOLUME 33 Specific dijferentiatiori. The cephalon of C. discoidalis differs most obviously from C. vietus in having larger eyes, a longer and wider frontal area with more divergent preocular sutures and a smaller genal spine. The same cranidial features distinguish it from C.? tenacella (Xiang and Zhang 1985, pi. 41, fig. 1 1). The pygidium of C. discoidalis differs from those of C. vietus and C. sp. of Shergold (1980, pis. 34 and 35) because it has several clearly defined axial rings (six or more rather than three or four). C. discoidalis also has a slight posterior indentation in the pygidium. Generic position. Shergold referred only C. vietus and some unnamed pygidia to Cermatops. C. discoidalis, as reconstructed here, shows many similarities with C. vietus and these we consider to outweigh the obvious differences. Glabellar structure. In ceratopygids SI has an unusual crescentic or longitudinal form, and most iwayaspidines show this and a conventional S2 and S3. Both C. vietus and C. discoidalis, unlike other iwayaspidines, show four pairs of furrows anterior of SI, though their homologies with S2, S3 etc. are not established with certainty. Shergold remarked that the glabellar furrows of C. vietus, which are weak, cannot be distinguished from faint muscle scars on the glabellar lobes (Shergold 1980, pi. 34, fig. 3). Furrows are present in a corresponding position in C. discoidalis, and an interpretation is given in text-fig. 3. Similar structure is also visible in the holotype of Sayramaspis tenacella Xiang and Zhang, 1985 and this may also be referable to Cermatops, though the pygidium (at present unknown) is needed to provide confirmation. The form of the SI furrows and the presence of a median glabellar node in C. discoidalis indicate that it is not closely related to the Dikelocephalidae, in which SI is commonly transcurrent and there is no preoccipital node. Pygidial structure. The pleural regions of the pygidium are well segmented but behind the anterior segment the propleural band (the anterior part of an individual segment) is reduced, both longitudinally and transversely (PI. 2, fig. 2). This is seen also in Tamdaspis (Ergaliev 1980, pi. 19, fig. 8) and, less distinctly, in Guozia (Xiang and Zhang 1985). Both these genera differ from Cermatops in their glabellar form. A similar pygidial structure was independently derived in the Dikelocephalidae, for example Briscoia septentrionalis Kobayashi, 1935a (Palmer 1968, pi. 15, figs. 3 and 4). Pygidia of remarkably similar form have also been described in the family Aphelaspidinae under the generic names Duihianaspis Lu and Lin (1984, pi. 7, figs. 8, 9, 12, 13) and Pseudaphelaspis (Arrhenaspis) Qian (1985, pi. 6, figs. 4—7). (Note that the name of the type species of the latter genus, P. (A.) latelimhata Qian, 1985, is unavailable, being a primary junior homonym of Pseudaphelaspis latelimhata Lu and Lin, 1984.) In each of these genera the pygidium was associated with a typical aphelaspidid cranidium, and if they are correctly so assigned the genera are synonymous. However, those authors have not considered the possibility that the pygidia belong to Cermatops or Tamdaspis. Compared with the pygidia of Cermatops, that attributed to Duihianaspis typicalis Lu and Lin, 1984 has a blunt axis that is barely half the pygidial length, and that attributed to P. (A.) ‘’latelimhata' Qian, 1985 is proportionally much wider, and recalls Tamdaspis. The pygidium from the Elvinia Zone in a borehole in Montana, figured by Lochman (1964, pi. 1 1, fig. 7) as Pterocephalia sanctishae Roemer, differs from other figured pygidia of that species but bears a great likeness to C. discoidalis, though it differs in having the ventral terrace ridges half as densely spaced. The generic assignment of this pygidium is uncertain because none of the associated cranidium is likely congeneric with it. Family relationships. Cermatops is regarded as a member of the Iwayaspidinae (Shergold 1980). This group shares several characters, for example the form of the glabellar furrows, the presence of a median preoccipital tubercle and a median suture, with primitive Asaphidae (in the sense of Fortey and Chatterton 1988) and the Ceratopygidae, but it lacks any convincing autapomorphy ; it is a paraphyletic group, and we find the usual difficulties in assessing the relationships of the taxa within such a group. The genera referred to the Iwayaspidinae commonly have a narrow cephalic border (compared with typical Asaphidae), a distinct preglabellar field and genal spines that are narrower at their base HUGHES AND RUSHTON; COMPUTER-AIDED RESTORATION OF TRILOBITE 439 than the width of the cephalic doublure; the thorax varies, with 8-10 segments. These features are not seen in the Asaphidae but are met with in the Ceratopygidae. Therefore we agree with Shergold (1980, p. 86) and Fortey and Chatterton (1988, p. 196) that the Iwayaspidinae are better referred to the Ceratopygidae than the Asaphidae. The typical Ceratopygidae (Subfamily Ceratopygidae) are characterized especially by the presence of marginal spines in the pygidium that are derived from macropleural segments. The capacity to develop such spines is taken as an autapomorphy for the subfamily, although it is evident that not all such spines are homologous (for example they are developed from the tenth post-cephalic segment in Proceratopyge but the eighth m Dichelepyge). Fortey and Chatterton (1988) justifiably referred the Macropyginae to the Ceratopygidae; this subfamily is characterized by baccular lobes on the cephalon and an exceptionally elongate pygidium (text-fig. 5/, /). Several genera have been referred to the Iwayaspidinae, as follows: the type genus Iwayaspis Kobayashi, 1962 (type species I. asaphoides Kobayashi) has been regarded as a junior synonym of Pseudoyuepingia Chien, 1961 (type species P. modesta Chien), e.g. by Jago 1987. Other genera are Yuepingia Lu, 1956 (type species Y. niobiformis Lu), Aplotaspis Henderson, 1976 (Charchagia erngata Whitehouse, 1939), Cermatops Shergold, 1980, Guozia Xiang and Zhang, 1985 [G. crassa Xiang and Zhang) and Sayramaspis Xiang and Zhang, 1985 {S. angustaxis Xiang and Zhang). Haniwoides Kobayashi, 19356, based on H. longus Kobayashi, is imperfectly known but is probably an iwayaspidine; it generally resembles Yuepingia apart from the apparent absence of a median glabellar tubercle. Haniwoidesl varius Shergold, 1980 (and probably also HP puteolatus Kobayashi, 1962), though doubtfully referred to the genus, has all the features of an iwayaspidine. Norinia Troedsson, 1937 has the typical arrangement of glabellar furrows but has a relatively short, undifferentiated preglabellar field, and it may be better referred to the Asaphidae. Charchaqia Troedsson, 1937 has the axial features effaced and is therefore difficult to evaluate, as already noted by Troedsson. Tamdaspis Lisogor, 1977 can be interpreted as an advanced Iwayaspidine but, depending on how its special features such as the bacculae are evaluated, it may be regarded as a primitive member of the Macropyginae (text-fig. 5/ and j). Psiloyuepingia Qian and Qiu, 1983 (in Qiu et al. 1983) is doubtless a synonym of Tamdaspis. We exclude Pseudohysterolenus Harrington and Leanza, 1957 because it differs from all iwayaspidine genera in the posterior position of the glabellar node. We also exclude Metayuepingia Liu (in Zhou et al. 1977), Yuepingioides Lu and Lin, 1984 and Parayuepingia Zhou et al., 1982) from the Iwayaspidinae. All have short frontal areas, little or no preglabellar fields, and in the first two forms the genal angle is rounded. Most of the above taxa have been proposed since publication of the Treatise Volume O (Moore 1959), and examples are illustrated here in text-fig. 5. Although many of their features are primitive with respect to the Asaphacea a few characters may be regarded as advanced when considering relationships within the group: 1. Reduction of the interocular cheeks. Early Proceratopyge and Asaphidae have the palpebral lobe separated from the glabella by a distance about equal to the length (sag.) of the occipital ring, and this is taken to be the primitive condition. In several species of Iwayaspidines and Ceratopygines the palpebral lobe is enlarged and approaches the glabella. As the anterior end of the palpebral lobe in all such forms lies near S3 and L4, those forms with a longer palpebral lobe necessarily have a more transverse postocular suture (Jago 1987, p. 227). 2. The development of baccular lobes in the adult (as in Tamdaspis, text-fig. 5/). 3. The reduction of the propleural band in the pygidium (e.g. Cermatops, text-fig. 5g). 4. The development of a large posterior indentation in the pygidium (as in Haniwoidesi varius, text-fig. 5/). 5. Effacement (e.g. in Yuepingia) is also regarded as a progressive feature but is so general as to be without classificatory value. 6. The presence of an auxiliary pair of glabellar furrows intercalated between SO and SI appears to be a specialized character but its distribution is sporadic. It is present in Cermatops discoidalis and in single species of Guozia {G.l duhia), Yuepingial {'’Iwayaspis' caelata, text-fig. 5d) and 440 PALAEONTOLOGY, VOLUME 33 G ^o ^ ,0 OO S T3 — 2 On'o Q bX)Tt S 00 — a G -« O, ao r s 00 •V- ^ w 1^ § T3 5 S3 ^ "S bo . ^ K -- - Q-» m '-S . -— CJD > o c "o a < 's: 00 j2 S o tr; ? ^ 2 bs — G ^ ^ W 60 o (£ "H o-’s c3 2 r- ^ r-> 00 TD . C 2f cd ^ ? X CiX) S 5 GO y:: ^ W) c St 00 2 Tj- T3 a o' fN 43 cd a ^ a r- ^ 1 60 G ^ C ^ « a g:'§ « o 5 -r o-^c: - fco cd r^ 'O ON c o c/5 '' . U 00 (U VO "O On C — ' : o> "t -S' o U (h-h D cd C ' (N a c/5 ^ 60 K ^ >- 'S ■2 u' ^ 3 •fe o FF-c a t3 c g: 'ij Cd D ^ cd c ^ S3 00 2 -"a .— cd Ofi-D ^ ri cd Wc^ rn ^ Ti fT* cin X )£ II tu On II {-. —H a ci) w) a S « 3 ■I a S\n-1 N1 bj 5! ^ , ,Co -~ ~o ’>• . 30 c a bj -Q G o^ Id 'a c: 00 00 00 Cd c 00 .2 2 60X G OG 2 cd HUGHES AND RUSHTON: COMPUTER-AIDED RESTORATION OE TRILOBITE 441 TABLE 1. Coding of characters for selected taxa of the Iwayaspidinae. They are mostly the type species of their respective genera, but Sayramaspis yardanshanensis and Pseudoyuepingia whitei (Webby et al. 1988) were preferred on account of their good preservation. The pygidium of Haniwoides convexus was used, as that of H. longus is not known. Character no. 1 2 3 4 5 6 7 8 9 10 11 12 Eoasaphus superstes 0 0 0 0 0 0 0 0 0 0 0 0 Proceratopyge conifrons 0 0 1 1 0 0 1 0 1 0 0 0 Sayramaspis yardanshanensis 0 0 2 0 0 0 1 1 0 0 0 1 Pseudoyuepingia whitei 0 0 1 0 1 0 1 0 0 0 0 9 ‘ Iwayaspis ' caelata 0 0 1 1 0 0 1 0 1 0 2 1 Yuepingia niobiformis 0 0 2 0 0 0 1 0 0 1 2 0 Aplotaspis erugata 0 0 1 0 0 0 1 1 1 1 1 1 Haniwoides longus 0 1 2 0 0 0 0 1 0 1 1 1 Tamdaspis tamdensis 0 0 3 0 0 0 1 0 1 1 2 1 Haniwoidesl varius 1 1 1 0 0 0 1 1 1 0 2 1 Cermatops discoidalis 1 1 1 1 1 1 1 1 0 1 2 1 Cermatops vietus 1 0 1 0 1 1 1 1 0 1 0 1 Guozia crassa 2 1 1 1 1 0 1 1 0 0 2 1 Character no. 13 14 15 16 17 18 19 20 21 22 23 Eoasaphus superstes 0 1 1 1 1 1 0 0 1 0 0 Proceratopyge conifrons 0 0 1 0 0 0 0 0 0 0 1 Sayramaspis yardanshanensis 1 0 0 0 1 1 0 0 0 0 0 Pseudoyuepingia whitei 1 1 0 0 0 0 0 0 1 9 0 'Iwayaspis' caelata 1 0 9 1 2 0 0 0 1 0 0 Yuepingia niobiformis 0 0 9 2 2 2 0 0 0 0 0 Aplotaspis erugata 1 1 1 1 0 0 0 0 1 0 0 Haniwoides longus 9 0 0 2 2 2 0 0 1 0 0 Tamdaspis tamdensis 9 9 9 2 2 2 1 0 1 1 0 Haniwoidesl varius 1 1 1 1 2 2 0 1 1 0 0 Cermatops discoidalis 1 1 1 1 2 0 0 1 1 1 0 Cermatops vietus 1 0 1 0 2 0 0 0 1 1 0 Guozia crassa 9 1 1 1 2 2 0 0 9 1 0 Characters 1-23, scored as follows. 1. Glabellar front: rounded 0, truncate 1, pointed 2. Glabellar sides: straight 0, concave 1. 3. Glabellar furrows: simple 0, asaphoid 1, effaced 2, only SI developed 3. 4. Auxiliary furrow on LI : absent 0, present 1. 5. S4 furrow : absent 0, present 1 . 6. Muscle-scars on glabella: absent 0, present 1. 7. Median glabellar tuberele: absent 0, present 1. 8. Occipital ring: simple 0, compound 1. 9. Plectral lines: absent 0, present 1. 10. Frontal area: border differentiated 0, not differentiated I . 1 1 . Preocular sutures : diverge at < 30° 0, 30°-60° 1, > 60° 2. 12. Paradoublural line on cranidium: absent 0, present 1. 13. Paradoublural line on free cheek: absent 0, present 1. 14. Length of genal spine: > half of rest of cheek 0, < half 1. 15. Width of genal spine at base: > width of doublure 0, < width of doublure 1. 16. Palpebral lobe length — cephalic axial length: < 0-3 0, 0-3-0-4 1, <0-4 2. 17. Distance of palpebral lobe from glabella: > length of occipital ring (SO) 0, = SO 1, < SO 2. 18. Ocular ridge: present 0, absent 1, palpebral lobe touches glabella 2. 19. Bacculae in adult: absent 0, present 1. 20. Pygidial margin: entire 0, emarginate I. 21. Postaxial ridge: absent 0, present 1. 22. Pygidial pleurae: normal 0, propleurae re- duced I . 23. Pygidial marginal spines: absent 0, present I. (Score 9 where a character cannot be coded.) 442 PALAEONTOLOGY, VOLUME 33 Text-fig. 5e 5d 5i 5a 5c 5f 5h 5g 2 Procevatopyge coni f Tons Pseudoyuepingia white i Eoasaphus supevstes Sayramaspis yardanshanensis ' Iwayaspis ' caelata Tamdaspis tamdensis Yuepingia niobifovmis Haniwoides longus Aplotaspis erugata Haniwoides? varius Guozia crassa Cermatops vietus Cevmatops discoidalis Text-fig. 5e 5d 5i 5a 5c 5f 5h 5g 2 Eoasaphus supevstes Sayvamasp i s yavdanshanens i s - Pseudoyuepingia whitei Pvoceratopyge conifvons- ' Iwayaspis' caelata Tamdaspis tamdensis 23 -E 3. 19, (22) Yuepingia niobifovmis Haniwoides longus -7 (5) 16 (-9) Aplotaspis evugata Haniwoides? vavius Guozia cvassa (-11) (20) Cevmatops vietus (-•l 1) Cevmatops discoidalis — | — (20) -E 3. 7 1 1 1 (5), (22). (-9) TEXT-FIG. 6. Relationships of selected taxa in the Iwayaspidinae, as indicated by the PAUP program (see text). Top, consensus tree with Procevatopyge conifrom for out-group comparison. Below, alternative tree with a hypothetical ancestor for out-group comparison. Eoasaphus is the most primitive actual taxon analysed. Numbers refer to characters in Table 1 ; negative numbers indicate character reversals and numbers in parentheses indicate parallelisms. HUGHES AND RUSHTON: COMPUTER-AIDED RESTORATION OE TRILOBITE 443 Sayramaspis (S. temcella, text-fig. 3c, possibly a species of Cermatopsl). It is of uncertain value in classification. 7. The postaxial ridge is well developed in many Iwayaspidines but is absent in some genera {Sayramaspis and Yuepingia) and doubtfully present in others (Psendoyuepingia and Guozia). The polarity of this feature is not clear. To test the relationships of examples of the Iwayaspidinae twenty-three attributes of thirteen species were analysed using the PAUP (Phylogenetic Analysis Using Parsimony) program, as described by Fortey and Chatterton (1988). Table 1 shows the matrix of characters used. The type species of Proceratopyge, P. conifrons Wallerius (Westergard 1948), was included for out-group comparison because it is regarded as a primitive ceratopygine and is stratigraphically the earliest species. When Proceratopyge conifrons was defined as the sister-taxon of the Iwayaspidinae, the program yielded two equally parsimonious but not very robust trees, differing only in the affiliation of Aplotaspis; the consensus tree is shown in text-fig. 6, top. An alternative analysis that compared the coded species with a hypothetical ancestor (which would score 0 in all columns of Table 1 - the Lundberg option) yielded a slightly different tree with the distal groupings unchanged but the basal dichotomies rearranged (text-fig. 6, below); this seems the more probable arrangement because it places Eoasaphns, which appears to lack the typical asaphine glabellar features, in the most primitive position. P. conifrons, which carries the autapomorphy of the Subfamily Ceratopyginae, branches off the tree above Pseudoyuepingia, in such a position that the Iwayaspidinae has to be regarded as a paraphyletic group. The results are viewed with caution because the attributes were mainly gleaned from descriptions and illustrations in the literature, and these are of uneven quality (a more reliable result could be obtained if the attributes were coded from actual specimens). According to these analyses the Iwayaspidinae is a paraphyletic group of the Ceratopygidae that lacks the pygidial spines of the Ceratopyginae and the bacculae and median pygidial extension of the Macropyginae. The Iwayaspidinae fall into three groups: (1) primitive ^orms - Eoasaphus, Pseudoyuepingia and Sayramaspis', (2) a large-eyed group with Haniwoides, Yuepingia and Tamdaspis', (3) a more specialized group with Cermatops, Guozia and //. ? varius. Aplotaspis appears as the sister taxon of (3) or of (2) 4- (3). We recognize that many of the features analyzed are not very compelling because several of them are known to have arisen independently in other groups. The most parsimonious of our cladograms includes several reversals of character-states; for example in C. vietus the small eyes and subparallel preocular sutures appear primitive in comparison with the rest of the taxa in its clade. The analysis indicates that Sayramaspis is probably a synonym of Pseudoyuepingia, and Yuepingia of Haniwoides ', these genera should be investigated further. If glabellar features are seen as most significant for classification Tamdaspis could be maintained as a distinct genus recognized by its bacculae (as well as its pygidial structure) and Guozia could likewise be separated from Cermatops by its distinctive glabellar shape. Acknowledgements . We thank Dr D. E. G. Briggs and Dr R. A. Fortey for helpful discussion. Dr R. P. S. Jefferies, Mr N. Golding and Miss K. Shaw of the British Museum (Natural History) for facilitating the computer-graphic method, and Mr S. Powell, Mrs P. Baldaro and Mr T. Cullen for assistance with aspects of our illustrations. Dr J. E. Dalingwater and two anonymous referees contributed valuable suggestions. Mr F. C. Collier kindly loaned material from the US National Museum of Natural History. N. C. Hughes gratefully acknowledges the receipt of a NERC Studentship. A. W. A. Rushton publishes with the permission of the Director of the British Geological Survey (NERC). REFERENCES APOLLONOV, M. K., CHUGAEVA, M. N. and DUBININA, c. V. 1984. [TrilohUes and conodonts from the Batyrbay Section {uppermost Cambrian-Lower Ordovician) in Malyi Karatau Range] (atlas of the palaeontological plates). Academy of Sciences of the Kazakh SSR, ‘NAliKA’, Alma-Ata, 48 pp., 32 pis. [In Russian]. 444 PALAEONTOLOGY, VOLUME 33 BELT, T. 1868. On the Lingula Flags or Festiniog Group of the Dolgelly district. Part III. Geological Magazine, 5, 5-11. BURMEISTER, H. 1843. Die Organisation der Trilohiten. Reimer, Berlin, viii+148 pp. CHIEN Yi-YUAN. 1961. Cambrian trilobites from Sandu and Duyan, southern Kweichow. Acta Palaeontologica Sinica, 9, 91-129, 5 pis. CONWAY MORRIS, s. and RUSHTON, A. w. A. 1988. Precambrian to Tremadoc biotas in the Caledonides. 93-109. In HARRIS, A. L. and fettes, d. j. (eds.). The Caledonian - Appalachian Orogen. Geological Society Special Publication No. 38, x + 643 pp. DATSENKO, V. A., ZHURAVLEVA, I. T., LAZARENKO, N. P., POPOV, YU. N. and CHERNYSHEVA, N. E. 1968. [Biostratigraphy and fauna of the Cambrian deposits of the northwest part of the Siberian Platform.] Trudy Nauchno-Issledovateliskii Institut Geologii Arktiki, no. 155, 1-213. [In Russian]. ERGALIEV, G. KH. 1980. [Trilobites of the Middle and Upper Cambrian of the Maly Karatau], Akademia Nauk Kazakhskoi SSR, Alma-Ata, 211 pp. [In Russian]. FEARNSiDES, w. G. 1910. The Tremadoc Slates and associated rocks of south-east Caernarvonshire. Quarterly Journal of the Geological Society of London, 66, 142-188. FORTEY, R. A. and CHATTERTON, B. D. E. 1988. Classification of the trilobite suborder Asaphina. Palaeontology, 31, 165-222. HARRINGTON, H. J. and LEANZA, A. 1957. Ordovician trilobites of Argentina. Special Publication, Department of Geology, University of Kansas, 1, 1-276. HENDERSON, R. A. 1976. Upper Cambrian (Idamean) trilobites from western Queensland, Australia. Palaeontology, 19, 325-364. HENNiNGSMOEN, G. 1957. The trilobite family Olenidae. Skrifter utgitt av Det Norske Videnskaps-Akademi i Oslo. 1. Matematisk-naturvidenskapelig Klasse for 1957, No. 1, 1-303. JAGO, J. B. 1987. Idamean (late Cambrian) trilobites from the Denison Range, south-west Tasmania. Palaeontology, 30, 207-231. JEFFERIES, R. P. s., LEWIS, M. and DONOVAN, s. K. 1987. Protocystites menevensis-a stem-group chordate (Cornuta) from the Middle Cambrian of South Wales. Palaeontology, 30, 429-484. KOBAYASHi, T. 1935fl. The Briseoia fauna of the late Upper Cambrian in Alaska with descriptions of a few Upper Cambrian trilobites from Montana and Nevada. Japanese Journal of Geology and Geography, 12, Nos 3-4, 39-57. 19356. The Cambro-Ordovician formations and faunas of South Chosen. Palaeontology. Part III. Cambrian faunas of South Chosen with a special study of Cambrian trilobite genera and families. Journal of the Faculty of Science, University of Tokyo, sect. 2, 4, 49-344. 1962. The Cambro-Ordovician formations and faunas of South Korea, Part IX. Palaeontology. VIII. The Machari Fauna. Journal of the Faculty of Science, University of Tokyo, sect. 2, 14, 1-152. LAKE, P. 1919. British Cambrian trilobites. Part 5. Palaeontographical Society Monographs, vol. for 1917, 89-120, pis. 11-14. 1946. British Cambrian trilobites. Part 14. Palaeontographical Society Monographs, vol. for 1945, 333-350, pi. 47. LiNNARSSON, J. G. o. 1869. Om Vestergdtlands Cambriska och Siluriska aflagringar. Kungliga Svenska vetenskapsakademiens handlingar, 8 (2), 1-89. LISOGOR, K. A. 1977. [Biostratigraphy of Upper Cambrian and Tremadoc trilobites of the Maly Karatau (southern Kazakhstan)]. 197-265. In Zhuravleva, i. t. and rosova, a. v. (eds.). [Biostratigraphy and fauna of the Upper Cambrian and boundary strata. (New data from the Asiatic part of the U.S.S.R.)]. Trudy Instituta Geologii i Geofiziki Akademiya Nauk SSSR, Sibirskoe Otdelenie, no. 313, 1-355. [In Russian]. LOCHMAN, c. 1964. Upper Cambrian faunas from the subsurface Deadwood Formation, Williston Basin, Montana. Journal of Paleontology, 38, 33-60. LU YAN-HAO. 1956. An Upper Cambrian trilobite faunule from eastern Kueichou. Acta Palaeontologica Sinica, 4, 365-380, 1 pi. and LIN HUAN-LiNG. 1984. Late late Cambrian and earliest Ordovician trilobites of Jiangshan-Changshan area, Zhejiang. 45-144. In Stratigraphy and palaeontology of systemic boundaries in China, 1. Cambrian-Ordovician boundary. Nanjing Institute of Geology and Palaeontology, Academia Siniea, Anhui, 405 pp. [In English]. MOORE, R. c. (ed.). 1959. Treatise on invertebrate paleontology. Part O, Arthropoda 1. Geological Society of America and University of Kansas Press, Lawrence, Kansas, xix + 560 pp. MORRIS, s. F. 1988. A review of British trilobites, ineluding a review of Salter’s Monograph. Palaeontographical Society Monographs, 140, (for 1986), 1-316. HUGHES AND RUSHTON; COMPUTER-AIDED RESTORATION OF TRILOBITE 445 PALMER, A. R. 1968. Cambrian trilobites of east-central Alaska. Professional Paper, US Geological Survev, No. 559-B, 1-115. 1977. Biostratigraphy of the Cambrian System - a progress report. Annual Review of Earth and Planetary Sciences, 5, 13-33. QIAN YiYUAN. 1985. Late late Cambrian trilobites from the Tangcun Formation of Jingxian, southern Anhui. Palaeontologia Cathayana, 1, 137-167. QUi HONGAN, LU YANHAO, ZHU ZHAOLING et al. 1983. Trilobita. 28-254. In [Palaeontological Atlas of Eastern China. I. Early Palaeozoic .]. Geological Press, Beijing, 657 pp. [In Chinese]. RUSHTON, A. w. A. 1982. The biostratigraphy and correlation of the Merioneth-Tremadoc Series boundary in North Wales. 41-69. In bassett, m. g. and dean, w. t. (eds.). The Cambrian-Ordovician Boundary, sections, fossil distributions, and correlations. National Museum of Wales, Geological Series No. 3, Cardiff, 227 pp. 1983. Trilobites from the Upper Cambrian Olenus Zone in central England. Special Papers in Palaeontology, 30, 107-139. SALTER, J. w. 1866. On the fossils of North Wales. 239-363. In ramsay, a. c.. The geology of North Wales. Memoirs of the Geological Survey of Great Britain, 3, i-viii+ 1-381. SHACKLETON, R. M. 1959. The stratigraphy of the Moel Hebog district between Snowdon and Tremadoc. Liverpool and Manchester Geological Journal, 2, 216-252, pis. 15, 16. SHERGOLD, J. H. 1980. Late Cambrian trilobites from the Chatsworth Limestone, western Queensland. Bulletin, Bureau of Mineral Resources, Geology and Geophysics, Australia, No. 186, 1-111. STUBBLEFIELD, c. J. 1953. Palaeontological Department. 56-59. In Summary of progress of the Geological Survey of Great Britain and the Museum of Practical Geology for the year 1951, 1-76. TAYLOR, M. E. 1977. Late Cambrian of western North America: trilobite biofacies, environmental significance and biostratigraphic implications. 397^25. In kauffman, e. g. and hazel, j. e. (eds.). Concepts and methods of biostratigraphy . Dowden, Hutchinson and Ross, Stroudsberg, Pennsylvania, 658 pp. TROEDSSON, G. T. 1937. On the Cambro-Ordovician faunas of western Quruq tagh, eastern T’ien-shan. Palaeontologia Sinica n.s. B, no. 2 (whole series no. 106), 1-74. WALCOTT, c. D. 1914. Cambrian geology and palaeontology. 2, no. 13. Dikelocephalus and other genera of the Dikelocephalinae. Smithsonian Miscellaneous Collections, 57 (13), 345^27. WEBBY, B. D., WANG QiZHENG and MILLS, K. J. 1988. Upper Cambrian and Basal Ordovician trilobites from western New South Wales. Palaeontology, 31, 905-938. WESTERGARD, A. H. 1948. Non-agnostidian trilobites of the Middle Cambrian of Sweden. Sveriges Geologiska Undersokning, Avhandlingar och Uppstatser, Series C, no. 498, 1-32. WHiTEHOUSE, F. w. 1939. The Cambrian faunas of north-eastern Australia. Part 3. Memoirs of the Queensland Museum, 11, 179-282. XIANG LiWEN and ZHANG TAiRONG. 1985. Trilobita. 64-136. In [Stratigraphy and trilobite faunas of the Cambrian in the western part of northern Tianshan, Xinjiang]. People's Republic of China, Ministry of Geology and Mineral Resources. Geological Memoirs, series 2, no. 4, i-ix -I- 1-243. [In Chinese, English summary]. ZHOU TiEN-MEi, LIU Yi-YENG, HONG siEN-suNG and SUN ZENG-HUA. 1977. Trilobita, 104-266. In [Atlas of fossils of central and south China. I. Lower Palaeozoic era]. Geological Publishing House, Beijing, 470 pp. [In Chinese]. ZHOU ZHIQIANG, LI JINSENG and QU XINGUO. 1982. Trilobita, pp. 134-213. In [Palaeontological Atlas of Northwest China. Shaanxi, Gansu, Ningxia. I. Proterozoic - Early Palaeozoic]. Geological Publishing House, Beijing, 480 pp. [In Chinese]. N. C. HUGHES Department of Geology The University Bristol BS8 IRJ, UK A. W. A. RUSHTON British Geological Survey Keyworth Nottingham NGI2 5GG, UK Typescript received 10 April 1989 Revised typescript received 20 July 1989 \ PRESERVATION OF AVIAN COLLAGEN IN AUSTRALIAN QUATERNARY CAVE DEPOSITS by R. F. BAIRD and M. j. rowley Abstract. The small well-preserved bones of ten avian fossils, species of the quail Twnix, from hve Australian caves, ranging in age from 9000 to 38,000 years, were tested in a radioimmunoassay for collagen. Collagen was well preserved in all cave environments studied, whether ‘wet’ or ‘dry’, and nine of ten samples tested contained collagen, from 10% to 92% of the collagen content of fresh bone. The age of the bone was not clearly related to its collagen content, although the amount of collagen detected was signihcantly less in the older of the two samples for three of the five caves. The moisture content of the cave was not shown to affect the preservation of collagen. This study suggests that caves may provide a favourable environment for the preservation of collagen where fossils are physically well preserved. Phylogenetic relationships have traditionally been established using morphological criteria. Recently biochemical and immunological comparisons of proteins or of DNA have also been used in determining relationships (Wilson et al. 1977; O’Brien et al. 1985). DilTerences between morphologically and biochemically derived phylogenies therefore have caused debate on the applicability of each method (Hillis 1987). For example, re-evaluation of the morphological evidence for primate evolution in the light of data from comparisons of proteins or of DNA, suggests that gorillas, chimpanzees and man may have shared a common ancestor as recently as 5 million years ago, as predicted by molecular evidence, and not 15-20 million years ago as first suggested from morphological evidence (Lowenstein and Zihlman 1984). Similarly, phylogenies derived by Sibley and Ahlquist (1983) using DNA hybridization are stimulating a reassessment of the origins of many Australian birds. These techniques have been extended to examine relationships of recently extinct animals, such as the quagga Equus qiiagga (Higuchi et al. 1984) and the thylacine Thylaciims cynocephalus (Lowenstein et al. 1981 ), whose preserved skins have been used as a source of organic material, but for genuine fossils, phylogenies continue to be based on morphology. Proteins may survive for considerable periods in fossils, however: bone collagen is one of the sources of used to date fossils, and the characteristic amino acids of collagen have been detected in dinosaur bones 200 million years old (Wyckoff 1980). Furthermore, material which reacts with specific antibodies to collagen has been detected in mammalian fossils millions of years old (Lowenstein 1980, 1981; Rowley et al. 1986). During a study of the survival of collagen in Australian fossil sites (Rowley et al. 1986; Rowley unpubl. data) we gained the impression that caves provided a poor environment for the preservation of collagen. By contrast, bone morphology may be excellently preserved in most caves (Baird in press). We noted that most of the material examined for collagen had been from caves with abundant moisture. This study was therefore commenced to see whether, when well-preserved bones were selected, the moisture level within the cave could be shown to influence the survival of collagen. Bone samples from two ‘dry’ caves were compared with those from three caves which were considered to be ‘wet’. The ability to predict which samples are most likely to contain collagen would reduce sample destruction and preparation time when collagen is used for phylogenetic studies. IPalaeontology, Vol. 33, Part 2, 1990, pp. 447^51. | © The Palaeontological Association 448 PALAEONTOLOGY, VOLUME 33 MATERIALS AND METHODS Environment The caves from which material was studied include Clogg’s Cave (EB-2: cave numbers from Matthews 1985: 148° ITS, 37° 30' E) and McEachern’s Cave (G-5: 141° 00' S, 37° 59' 30" E), Victoria; Koonalda Cave (N-4: 129° 50' S, 31° 24' E), South Australia; and Madura Cave (N-62: 127° 02' S, 32° 00' 30" E) and Devil’s Lair (WI-61E: 1 15° 03' S, 34° 07' E), Western Australia. Both Clogg’s Cave and Devil’s Lair represent caves with moist environments (‘wet’ caves), and Madura Cave and Koonalda Cave represent caves with dry environments (‘dry’ caves). These are relative terms and imply no absolute definition, but generally wet caves have damp sediments, and may have water dripping or periodic inundations, while dry caves have dust. In many subaerial caves, if high humidity occurs for a long enough period, there is no preservation of bone whatsoever (RFB, pers. obs.), so that the determination of ‘wet’ and ‘dry’ applies only to those caves where bone survives. The main taphonomic accumulator was the Barn Owl (Tyto alba, Tytonidae, Strigiformes) for all caves except G-5, which appears to be of fluvial/pitfall origin. The material from Koonalda Cave has been secondarily sorted by fluvial action (Baird 1986). Bone samples The taxa studied were Twnix varia (Painted Button-quail) for Clogg’s Cave, McEachern’s Cave and Devil’s Lair, and Twnix sp. cf. T. velox (Little Button-quail) for Koonalda and Madura Caves (see Appendix). These species have been chosen because of the abundance of their elements in the deposits and the presence of this genus in a number of deposits. A number of criteria were used in selecting material for study, including : the elements were all from the distal end of the humerus, the elements were complete before processing, the elements lacked damage or alteration to their surfaces, and the elements had a uniform light coloration. In some cases one or two of the criteria were not upheld because of the scarcity of material of appropriate geological age (i.e. different elements [youngest WI-61e = incomplete femur, and oldest = incomplete tarsometatarsus], incomplete elements [oldest N-4 = distal end humerus, youngest N-62 = proximal end humerus, and oldest N-64 = distal end humerus] and discoloured elements [oldest WI-61e = dark brown]). Table I gives the chronological distribution of the elements. Preparation of the fossils Fossils were ground to a fine powder, decalcified with 10 volumes of 0-5 m-EDTA, pH 7-5, then re-extracted with 10 volumes of 0-5 m acetic acid. The remaining bone powder was resuspended in 10 volumes of phosphate- buffered saline (PBS), pH 7-3, for testing. Extractions were carried out in siliconized glassware throughout, to minimize loss of protein on the sides of the tubes. Bones from modern Twnix varia and Dromaius novaehollandiae (Emu) were defatted by sequential extraction using acetone and ethyl ether, then air-dried and treated similarly to the fossils. All extractions were carried out at 4 °C. Radioimmunoassay A solid-phase radioimmunoassay for collagen was carried out on flexible polyvinyl microtitre plates. Wells were coated with 50 p\ of EDTA extract, acetic acid extract, or bone powder suspended in PBS and held overnight at 4 °C in a moist chamber. After coating with antigen, the plates were washed 3 times with PBS containing 1 % skimmed milk powder and 0 05% Tween 20 (Blotto), and then washed 6 times with distilled water. The plates were then exposed to 200 p\ of Blotto for 2 hours at room temperature, to coat residual sites on the plastic, and again washed as above. The assay system was completed by adding 50 //I of antiserum dilution to each well; the plates were kept overnight at 4 °C, and then washed as before. Antibody binding to the plates was detected using protein A from Staphylococcus aureus, labelled with 50 p\ ‘“H, 50,000 cpm, with a specific activity of 40 //Ci///g, which was added to each well. The plates were kept overnight at 4 °C, washed as before, cut, and the activity bound to the wells was counted on a gamma counter. Under the conditions of the assay, the amount of radioactivity was proportional to the amount of collagen bound to the plate. Each sample was tested in quadruplicate, using rabbit anti-collagen antiserum to measure specific binding to collagen, and normal rabbit serum to measure non-specific binding. In addition, each serum was tested on BAIRD AND ROWLEY; QUATERNARY AVIAN COLLAGEN 449 uncoated wells of the plates, to determine the ‘background’ binding observed in the absence of fossils. Each sample was counted for 10 minutes, to increase the sensitivity of the assay. Modern Tumix varia and Droniaius novaehollandiae were included for comparison, and a control of soluble collagen from Droniaius novaehotlaiuliae in PBS, in doubling dilutions from 5 ^/m/ml, was included on each plate. Antibodies to collagen Type I collagens were extracted from bird skins by pepsin digestion and purified by differential salt precipitation (Chung and Miller 1974). Rabbits were immunized subcutaneously with 5 mg of collagen in complete Freund’s adjuvant initially, and in PBS 4 weeks later; antibodies reacted predominantly with native collagen and minimally with denatured collagen, as reported previously (Timpl 1982). The antiserum chosen for this study was a rabbit antibody to chicken collagen, selected because it contained the greatest reactivity to denatured collagen of any antiserum tested. Although it reacted most strongly with chicken collagen, it gave 85% of that reactivity with purified Turnix collagen. RESULTS The results of this study are summarized in Table 1. Samples of bone from modern Tumix varia and Dromahis novaehollandiae were included for comparison. However, the amount of collagen extracted from the sample of Turnix varia was approximately 60% of that extracted from the Droniaius novaehollandiae sample, and was lower than the amount obtained from some of the fossils. This bone was from a specimen that had been cleaned of flesh by exposure to dermestid beetles (Dermestes niaculatus), in a warm moist atmosphere over several weeks. By contrast, the taphonomic accumulator for most fossils studied was the Barn Owl (Tyto alba), which implies that the bones would have been cleaned of flesh within hours, and excreted in a dry pellet. After preliminary drying, collagen is much more resistant to subsequent hydrolytic decomposition than is collagen, which remains moist (Wyckhoflf 1972). The bone from the Droniaius novaehollandiae was from a freshly killed bird, which had been stored at —20 °C since death. This difference in preparation may have affected the preservation of collagen. Therefore in Table 1 the amount of collagen found in the bones is expressed as a percentage of that extracted from the Droniaius novaehollandiae bone. TABLE 1. Results of the analysis on collagen in avian fossils (Turnix spp.) of late Quaternary age from caves distributed across southern Australia. Results are expressed as mean counts per minute of radioactivity bound +1 standard deviation Site Cave environment Age (years) Bone sample Collagen % (% of fresh collagen) Anti-collagen NRS Fresh bone (Droniaius novaehollandiae) 18,100 + 2,300 500+110 No antigen — 39 + 4 42 + 26 Clogg’s Cave Wet 8,720 + 230 11,160 + 3,400 210+120 64 17,720 + 840 5,300 + 940 170 + 60 29 Devil’s Lair Wet 11,960+140 2,400 + 350 85+18 13 32,480+ 1,250 4,500 + 450 160 + 4 25 Madura Cave Dry 18,990 + 220 16,650 + 5,600 360+ 110 92 37,880 + 3,520 67 + 2 45 + 6 Koonalda Cave Dry 13,700 + 270 6,300+1,700 100+10 35 > 20,600- < 21,550 1,800 + 370 68 + 9 10 McEachern’s Cave Wet 9,920 + 270 3,200 + 760 520+ 180 18 14,880 + 240 3,700+ 1,400 450+ 160 21 450 PALAEONTOLOGY, VOLUME 33 Although the amounts of soluble collagen in the EDTA and acetic acid extracts were normally at or below 1 % of the total, and therefore are not considered significant in the current study, those measurements from the McEachern’s Cave elements yielded anomalously high counts (i.e. about 10% and about 5% of the total, respectively). DISCUSSION In previous studies on the preservation of collagen in fossils, samples of fossil bone obtained from cave deposits have contained very little collagen, irrespective of the age of the sample (Rowley et al. 1986; Rowley unpubl. data). This study was commenced to evaluate the effect of a damp cave environment on the preservation of collagen in bones. However, contrary to expectations, samples tested from both ‘wet’ and ‘dry’ caves contained significant residual collagen: one sample aged 18,990 + 220 years contained 92% of the collagen detected in a modern bone sample. Therefore, the humidity of the cave does not appear to be a critical factor in the preservation of collagen. Other factors which we believed might have influenced collagen preservation included the geological age of the sample and its preservation history and size. The age of the bone was not clearly related to its collagen content, although the amount of collagen detected was significantly less in the older of the two samples for three of the five caves (Clogg’s Cave, Madura Cave, and Koonalda Cave), and not significantly different for McEachern’s Cave, the cave for which the age difference between levels was least. The effects of preservation remain untested since we purposely chose samples with similar taphonomic histories. In contrast to our previous studies, however, these avian bones have been extremely well preserved, with an undamaged surface and minimal breakage. Such good preservation of small bones, each weighing less than 1 g, suggests that the microenvironments experienced by these bones may have been unusually favourable. Thus, Murray and Goede (1977) have shown that the preservation of elements in a cave environment is directly related to the weight and volume of the specimen. In our previous studies, the bones tested were much larger and less well preserved. Although good physical preservation is not an absolute indication of good preservation of collagen, and the only sample tested which did not contain measurable collagen was also rated as very well preserved, good structural preservation has previously been noted to be associated with preservation of amino acids of collagen (Wyckoff 1972). Acknowledgements. We extend our thanks to the following people and their respective institutions for the use of material in their care in destructive analysis: Dr T. H. Rich (Museum of Victoria), Mr N. Pledge (South Australian Museum) and Dr K. McNamara (Western Australian Museum). M.J.R. was supported by a National Research Fellowship. REFERENCES BAIRD. R. F. 1986. The avian portions of the Quaternary cave deposits of southern Australia and their biogeographical and palaeoenvironmental interpretations. Unpublished Ph.D. thesis. Monash University. ., in press. The taphonomy of late Quaternary cave localities yielding vertebrate remains in Australia. In RICH, p. V., BAIRD, R. F., MONAGHAN, J. M. and RICH, T. H. (eds.). Vertebrate palaeontology of Australasia. Chapman and Hall, London. CHUNG, E. and miller, e. j. 1974. Collagen polymorphism: characterization of molecules with the chain composition [I(III)] in human tissues. Science, 183, 1200-1201. HiGUCHi, R., bowman, B., FREiBERGER, M., RYDER, o. A. and WILSON, A. c. 1984. DNA Sequences from thequagga, an extinct member of the horse family. Nature, 312, 282-284. HiLLis, D. M. 1987. Molecular versus morphological approaches to systematics. Annual Review of Ecology and Systematics, 18, 23-42. LOWENSTEIN, J. M. 1980. Species-specific proteins in fossils. Naturwissenschaften, 67, 343-346. 1981. Immunological reactions from fossil material. Plulosoplucal Transactions of the Royal Society of London, Series B, 292, 143-149. BAIRD AND ROWLEY: QUATERNARY AVIAN COLLAGEN 451 , SARiCH, V. M. and Richardson, b. j. 1981. Albumin systemalics of the extinct mammoth and Tasmanian wolf. Nature, 291, 409^11. and ziHLMAN, A. L. 1984. Human evolution and molecular biology. Perspectives in Biology ami Medicine, 27,611-622. MATTHEWS, P. G. 1985. Australian karst index. Australian Speleological Federation, Melbourne, 322 pp. MURRAY, p. and GOEDE, A. 1977. Pleisocene vertebrate remains from a cave near Montague, N.W. Tasmania. Records of the Queen Victoria Museum, 60, 1-30. O'BRIEN, s. J., NASH, w. G., wiLDT, D. E., BUSH, M. E. and BENVENiSTE, D. E. 1985. A molecular solutioii to the riddle of the giant panda’s phytogeny. Nature, 317, 140-144. ROWLEY, M. J., RICH, p. v., RICH, T. H. and MACKAY, I. R. 1986. Immunoreacti ve collagen in avian and mammalian fossils. Naturwissensehaften, 13, 620-623. SIBLEY, c. G. and ahlquist, j. e. 1983. Phytogeny and classification of birds based on the data of DNA-DNA hybridization. Current Ornithology, 1, 245-292. TiMPL, R. 1982. Antibodies to collagens and procollagens. Methods in Enzymology, 82, 472^98. WILSON, A. c., CARLSON, s. s. and WHITE, T. J. 1977. Biochemical evolution. Annual Review of Biochemistry, 46, 573-639. WYCKOFF, R. w. G. 1972. The biochemistry of animal fossils. Scientechnica, Bristol, 152 pp. 1980. Collagen in fossil bones. In hare, p. e. (ed.). Biogeochemistry of amino acids. 17-22. John Wiley & Sons, New York. 558 pp. R. F. BAIRD Department of Earth Sciences Monash University, Clayton Victoria 3168, Australia and Department of Ornithology Museum of Victoria, Abbotsford Victoria 3067, Australia M. J. ROWLEY Department of Earth Sciences and Centre for Molecular Biology and Medicine Typescript received 16 August 1988 Monash University, Clayton Revised typescript received 1 August 1989 Victoria 3168, Australia APPENDIX Avian fossil material used during the course of this study, arranged with the youngest material from each locality above the older material. Cave numbers are from Matthews (1985). Abbreviations: com. = complete, dist. = distal, fern. = femur, hum. = humerus, incom. = incomplete, MV = Museum of Victoria, prox. = proximal, SAM = South Australian Museum, WAM = Western Australian Museum. Identification Element Museum number Cave Cave no. Turnix sp. cf. T. velox Com. right hum. SAM P.26117 Koonalda N-4 Turnix sp. cf. T. velox Dist. end right hum. SAM P.261132 Koonalda N-4 Turnix varia Incom. right fern. WAM 73.10. 1451 Devil’s Lair WI-6le Turnix varia Incom. right tmt. WAM 86.7.47 Devil’s Lair WI-61e Turnix varia Com. right hum. MV P. 183347 Clogg’s EB-2 Turnix varia Incom. left hum. MV P. 1834377 Clogg’s EB-2 Turnix sp. cf. T. velox Prox. end right hum. MV P. 184897 Madura N-62 Turnix sp. cf. T. velox Dist. end right hum. MV P. 184902 Madura N-62 Turnix varia Incom. right hum. MV P. 161181 McEachern’s G-5 Turnix varia Com. right hum. MV P. 161131 McEachern’s G-5 ■j> - ’’ u »■ ii-^^- .3. - I ■ f- ^ ■:! '.iV ii EVOLUTION OF GRYPHAEATE OYSTERS IN THE MID-JURASSIC OF WESTERN EUROPE by A. L. A. JOHNSON and C. D. lennon Abstract. European Callovian (and later) forms of Gryphaea (Bilobissa) arose not from earlier representatives of the subgenus but from Catimda, a much smaller, frequently ribbed form, here regarded as a subgenus of Gryphaea. Evolution was essentially gradualistic. G. (Catinula) itself arose from an early G. (Bilobissa) species at the Toarcian/Aalenian boundary. In this case evolution was rapid (and apparently restricted to a small geographical area) but there is little evidence of stasis before and afterwards. The earlier G. (Bilobissa) lineage became extinct in the late Bajocian or early Bathonian. The morphologies of G. (Bilobissa) and G. (Catinula) may represent alternative adaptations for reclining in similar, low-energy environments, respectively favoured under conditions of high and low potential for shell growth. Such potential may have been controlled by ocean temperature and/or salinity. Most of the change between G. (Bilobissa) and G. (Catinula) probably resulted from alteration of growth rates. This almost certainly involved genetic change, although ecophenotypic variation may have been a precursor. The coiled oyster Gryphaea has been dubbed the "Drosophila’ of palaeontology (Gould 1972, p. 91). Certainly its evolution has been the subject of many more papers than most other fossil organisms (Gould 1980). However, while studies of Drosophila revolutionized genetics, the same cannot yet be claimed for evolutionary studies on Gryphaea. This is because the history of Gryphaea research is ‘replete with biometrical errors’ (Gould 1972, p. 91 ), and there is still far from complete agreement on the course of Gryphaea evolution. In his most recent publication on the subject, Hallam (1982) has claimed that Gryphaea is a monophyletic genus that evolved in a step-wise fashion, roughly according to the theory of punctuated equilibrium (Eldredge and Gould 1972). However other authors (e.g. Arkell 1934; Cox 1946, 1952; Sylvester-Bradley 1959, 1977) have regarded Gryphaea as no more than an evolutionary grade, attained polyphyletically, and Sylvester- Bradley considered that the genus provided good evidence of phyletic gradualism as well as of ‘quantum’ evolution. Sadly, Professor Sylvester-Bradley died before he was able to present morphometric data, assembled over an interval of nearly thirty years, which he believed supported his views. It is the principal intention of this paper to appraise Sylvester-Bradley’s views in respect of European mid-Jurassic forms, the main subject of his collection and measurement. He put forward a quite explicit gradualistic evolutionary scheme for some of these forms (1959, 1977). This can be tested both as a putative example of gradualism, and, since evolution to Gryphaea allegedly proceeded from forms referred to a separate genus (Catinula), as a case serving to demonstrate the iterative evolution of Gryphaea. We review other phylogenetic schemes involving Catinula and find no evidence that it is more closely related to other oyster genera than to Gryphaea', Sylvester- Bradley’s scheme is thus shown to be plausible at the outset as an alternative to the view of Gryphaea monophyly. We then identify those areas of species-level phylogeny that are critical to the question of the relationship between Catinula and Gryphaea. These are investigated in depth in order to reach ultimately a decision on the overall course of evolution. We also investigate Sylvester-Bradley’s views concerning other aspects of the phylogeny of mid- Jurassic gryphaeate oysters, partly in conjunction with a critical assessment of whether the morphological changes observed are, in fact, evolutionary. Conclusions in respect of the latter are taken into account in the formulation of a revised scheme of supraspecific classification. Extensive reference is made throughout to the work of Brannan (1983); unfortunately this remains [Palaeontology, Vol. 33, Part 2, 1990, pp. 453-485, 3 pls.j © The Palaeontological Association 454 PALAEONTOLOGY, VOLUME 33 unpublished but as a recent and comprehensive study of Jurassic non-lophate oysters it demands the fullest attention. Through the kindness of Mrs Joan Sylvester- Bradley we had available for study the collection of some 1 5,000 specimens assembled by her late husband. Our investigation of these was supplemented by studies at many of the field localities from which the specimens were derived and by limited recourse to other collections. CONFLICTING HYPOTHESES FOR THE PHYLOGENY OF EUROPEAN MID-JURASSIC GRYPHAEATE OYSTERS Principal current views Most large, gryphaeately-coiled oysters encounterd in mid-Jurassic rocks in Europe bear a marked radial posterior sulcus on the left (coiled) valve distinguishing them from representatives of the weakly-sulcate, and principally Liassic, Gryphaea lineage which has been the subject of so many previous studies (see Gould 1980). The latter lineage, referred by Stenzel (1971) to the subgenus G. (Gryphaea) Lamarck, may extend into the very lowest Middle Jurassic (Hallam 1982; Brannan 1983) but there is general agreement that at higher horizons the sulcate forms, which are known from as early as the Sinemurian (Hallam 1982), are the sole Gryphaea stock represented in Europe. This group of forms was referred to the subgenus Bilohissa by Stenzel (1971). While opinions have changed or differed about phylogeny within Bilohissa, most recent workers (Hallam and Gould 1975; Hallam 1978, 1982; Brannan 1983) have ruled out the involvement of any other taxon of similar or higher rank. However, Sylvester-Bradley (1959, 1977) considered that European Callovian forms of Gryphaea (i.e. Bilohissa) had evolved not from the forms of Bilohissa common early in the mid-Jurassic but from Bathonian forms of the oyster Catinula Rollier, similarly deeply- excavate but distinguished by its very much smaller size and development of radial ribbing on the left valve. This idea had been put forward in its essence by Arkell (1934); Sylvester-Bradley added the claim that the transition involved gradualistic change. Sylvester-Bradley’s hypothesis is presented as part of ‘phylogenetic pathway I’ in text-fig. 1. Included within this latter scheme is the derivation of Catinula from an early Bilohissa species (from the uppermost Lower Jurassic) and a link between early and later forms of Catinula. These latter concepts are not clearly expounded in Sylvester-Bradley’s published writings but manuscript notes demonstrate he realized that uppermost Lower Jurassic "Catinula' pictaviensis (Hebert) of his 1959 paper - the supposed ancestor of Middle Jurassic C. heaumonti (Riviere) - is in fact a representative of Bilohissa, a view adopted by all other recent workers (Hallam 1982; Brannan 1983; Bayer et al. 1985). Manuscript notes also show that Sylvester-Bradley intended to refer such early Catinula species as C. heaumonti to a new genus. However, it is reasonable to assume that Sylvester-Bradley saw the ultimate ancestry of later forms of Catinula as lying within this genus, and thus to present route I in text-fig. 1 as a characterization of his views concerning the phylogenetic pathway between early and later forms of European Bilohissa. Brannan (1983) did not TEXT-FIG. 1. Contrasting proposals for the origin of European Callovian Gryphaea (Bilohissa). Route I - the ‘Sylvester-Bradley’ model: gradual evolution from Catinula, itself derived from an early G. (Bilohissa) species. Route II - Hallam’s (1982) model: direct evolution of Callovian G. (Bilohissa) from earlier members of the subgenus by a process involving punctuational change. The first model implies extinction of an early G. (Bilohissa) lineage before the Callovian; the second implies that Catinula died out without leaving any descendants by the early Callovian. Specimens illustrating gradual evolution are from the series depicted by Sylvester-Bradley (1977, text-fig. 1 1), with the largest (latest) specimen excluded. All specimens are left valves, seen from the exterior. From top, clockwise: Leicester University, Dept, of Geology (LEIUG) 104892, 104893, 104880, 104510, 104450, 104537; British Geological Survey, Keyworth, Nottingham GSM 73019; LEIUG 61452; Office national de gestion des collections paleontologiques, Villeurbanne, Lyon, France (ONCP), EM 35001; Museum National D’Histoire Naturelle, Paris, Prance, B. 48576, B. 48575; all xO.75. LOWER JURASSIC MIDDLE JURASSIC JOHNSON AND LENNON : J U RASSIC OYSTER EVOLUTION 455 STAGES Callovian Bathonian Bajocian Aalenian Toarcian Gryphaea (Bilobissa) TEXT-FIG. 1. For legend see opposite. 456 PALAEONTOLOGY, VOLUME 33 consider it necessary to refer early forms of Catinula (e.g. C. heaumonti) to another genus but endorsed Sylvester- Bradley’s views with respect to their origin from Bilohissa (see also Stenzel 1971, p. Nl 102). He did not, however, agree with the idea that Catinula subsequently evolved back into Bilohissa. In text-figure 1 we have followed Brannan in referring both earlier and later forms of the small, ribbed oyster to Catinula. Route II in text-figure 1 is a characterization of Hallam’s latest views (1982) on the origin of European Callovian Bilohissa and of evolutionary tempo within the subgenus. Hallam considers that the uppermost Lower Jurassic forms of Bilohissa referred to above may not be specifically separable from Aalenian and Bajocian forms referred to G. (B.) hilohata J. de C. Sowerby (recte suhlohata (Deshayes)), to which species he also assigns three early Bathonian specimens recorded by Fischer (1964) from France. These are the only examples of Gryphaea recognized from this stage in Europe. An extended period of stasis (> 14 Myr) is thus recognized, ended by the sudden appearance early in the Callovian of a smaller, morphologically distinct, Bilohissa species which persisted for a further 2 Myr. In view of the extreme rarity of Bathonian Bilohissa in Europe one may wish to question whether the origin of the Callovian species can be said to represent a punctuation event, but the essential fact is that Hallam rules out any involvement of Bathonian Catinula. (He expresses no view on the alleged evolution of Catinula from Bilohissa.) Brannan (1983) also excludes Catinula from the ancestry of Callovian Bilohissa but considers that the phytogeny of European Bilohissa is much more complicated than envisaged by Hallam, with as many as three coexisting species in the early mid-Jurassic, and both the Bathonian and earliest Callovian forms possibly representing a separate lineage from one (for which there is no fossil evidence) linking Bajocian and other early Callovian Bilohissa. While it is of no special relevance to our main concern - the relationships of supraspecific taxa - we would agree with Brannan that at least early in the mid-Jurassic, Bilohissa exhibits considerable morphological variation (see Bayer et al. 1985) such that the existence of stasis must be questioned. In text-figure 1 forms which best evince Hallam’s views have been deliberately selected; much more divergent forms could have been illustrated. Of greater significance is the position of the later forms mentioned above - these will be discussed in due course. Summarizing, Sylvester- Bradley thought that 'Catinula'' evolved from Gryphaea (Bilohissa) in the Aalenian and subsequently evolved back into Bilohissa in the Callovian ; Hallam thinks that these Callovian forms derive from a persistent (albeit in the Bathonian, exceedingly rare) Bilohissa lineage, and that they are unrelated to Catinula. Other phylogenetic hypotheses and the definition of genera Brief mention must also be made of other phylogenetic schemes involving the above forms. Perhaps the most obvious possibility is that European Callovian Bilohissa might have evolved from some lineage of Bilohissa which existed outside Europe during the Bathonian. This would go some way to explaining the embarrassing lack of Gryphaea in Europe stratigraphically intermediate between the abundant Bajocian and Callovian forms (Fischer’s three specimens excluded). Marine clays, seemingly suitable for Gryphaea (although see Hallam and Gould 1975), accumulated widely in Europe in the Bathonian, and indeed the presence of Catinula rather than Gryphaea in these probably gave impetus to Sylvester-Bradley’s investigation of Arkell’s original claim concerning the derivation of Callovian Bilohissa from Catinula. Two species, G. impressimarginata McLearn and G. nehraskensis Meek and Hayden, are known from Bathonian rocks in N. America (J. H. Callomon, pers. comm. 1985), and Hallam (1982) recognizes G. costellata (Douville) in this stage in the Middle East. However, Hallam was sufficiently impressed by the morphological differences exhibited by these species (respectively, absence of a posterior sulcus, presence of an anterior sulcus, presence of very strong ribs) to rule them out as members of a lineage with Bajocian and Callovian representatives in Europe. Our limited experience of extra-European forms supports Hallam’s view and, given the latter’s preference for what might seem a rather contrived explanation (that a Bilohissa lineage emigrated from Europe in the Bathonian leaving, however, no trace elsewhere of its continued existence and despite possibly favourable facies in Europe), we feel justified in JOHNSON AND LENNON; JURASSIC OYSTER EVOLUTION 457 excluding from further consideration the possibility that ‘European’ Bilohissa persisted during the Bathonian in N. America or the Middle East. Brannan (1983) tacitly adopted the same view. Several other phylogenetic schemes have been proposed. Siewert (1972) considered Grypliaea to be monophyletic by virtue of a unique, dominantly prismatic, shell structure; similarities in shell structure were taken to indicate that Catimda had evolved from generally flat oysters referred to Liostrea Douville, and the allegedly invariant position of the attachment area posterior of the umbo was regarded as an indication that Catiniila constituted part of the ancestry of the transversely- coiled genus Exogyra Say. The existence of prismatic structure in the innermost parts of the left valve of Grypliaea is not clearly demonstrated by Siewert; his pi. 2, fig. 4 shows only ‘pigment prisms’. Our own investigations, and those of other recent workers (e.g. Stenzel 1971; Brannan 1983), show that where no diagenetic recrystallization has occurred, the shell structure is dominantly foliated calcite. The ‘subrhomboidal’ structure reported by Pugaczewska (1971, p. 276; after Celcova 1969) may well be a variant of the latter (cf. Carter 1980, p. 81). We were unable to detect any difference in shell structure between Grypliaea (Bilohissa) and Catimila (text-fig. 2). Similarly, no difference exists in the mean position of the attachment area (see text-figs. 8-10; Pis. 1-3): contrary to Siewert’s opinion, the attachment area usually truncates the umbo dorsally in Catimda, and the transverse element of coiling is very much weaker than in Exogyra. TEXT-FIG. 2. Photomicrographs of acetate peels showing the foliated shell structure forming the bulk of the shell in Grypliaea (Bilohissa) and Catimda. A: Grypliaea (Bilohissa) - population CRC (horizon and locality, p. 472); LEIUG 106801 ; x 18. B: Cn?/>m/a - specimen collected from same horizon and locality as PBA population (p. 461); LEIUG 96891; x42 (almost full shell thickness shown). Both Rollier (1911) and Charles and Maubeuge (1953) considered plicate oysters belonging to the genus Lopha Roding (including Rastellum Eaujas-Saint-Fond) to be near relatives of Catimda. However, as pointed out by Brannan (1983), this was because of their inclusion of forms referable to the former genus within the latter. Brannan has listed a number of characters which separate the genera, bearing out an obvious distinction based on the form of the ribbing (produced by local thickening of the shell in Catimda, rather than plication). Pugaczewska (1971) and Arkell (1934) considered, like Siewert (1972), that Catimda had evolved from flat oysters referred to Liostrea. The former offered no particular basis for this claim but amongst more general assertions Arkell claimed specifically that Bathonian "Liostrea' liebridica (Forbes) could be traced into a " Catinula stage’. While ribbed morphs are undoubtedly developed in ‘L. ’ liebridica, Hudson and Palmer (1976) have clearly indicated that the species can be distinguished from both Catimda and Liostrea (with the possible exception of early Jurassic L. Iiisingeri (Nilsson)) by the existence of prismatic structure in the outer layer of the left valve, lensoid cavities elsewhere in the shell, and other features. ‘L. ’ liebridica is referable to Praeexogyra Charles and Maubeuge. True Liostrea is distinguished from Catinula by the lack of any appreciable dorso- ventral incurvature in the left valve (Brannan 1983). Brannan has argued convincingly that the earliest Praeexogyra species evolved from Grypliaea (Bilohissa) so the idea that Praeexogyra descended from Catimda, put forward by Charles and Maubeuge (1953), can be discounted. Most 458 PALAEONTOLOGY, VOLUME 33 forms referred to Praeexogyra are flat but two small, excavate species (commonest in the Bajocian) were included in the genus by Brannan (1983). They can be distinguished from Catinula by a complete absence of ribbing. Cox (1946, 1952) held the tentative view that Catinula was polyphyletic but, unlike Arkell, who claimed that Catinula had evolved repeatedly from ^ Liostrea ’, he doubted whether all Catinula had arisen in this way, and was also sceptical of Arkell’s view (see above) that Callovian Gryphaea had evolved from Catinula. As indicated by Brannan (1983), Cox’s acceptance of Catinula as a possibly polphyletic genus can be explained by his inclusion of forms which in fact belong in quite separate genera. Exclusion of these renders Catinula a probable monophyletic taxon. Definition q/ Catinula. It is evident that much of the speculation over the phylogenetic position of Catinula has resulted from failure to define the taxon adequately at the outset. Catinula can be differentiated from other supraspecific taxa in the following way. From Gryphaea (Bilobissa) it can be distinguished by the ribbing and small size previously mentioned. The development of ribs (i.e. local thickenings of the shell) and strong dorso-ventral incurvature, and the absence of plicae, external prismatic shell structure and cavities in the shell, variously distinguish Catinula from other superficially similar oysters. Thus defined there is no reason for supposing that any other Jurassic oyster is more closely related to Catinula than Gryphaea (Bilobissa) (cf. above discussion). The earliest and latest acceptable occurrences of Catinula are, respectively, Aalenian and early Callovian. Pugaczewska’s (1971, p. 216) record from the Lower Jurassic is unsubstantiated. Forms described by Arkell (1934, pp. 60, 64) from the Lower Cretaceous of Texas have since been referred to Texigryphaea Stenzel (Stenzel 1971). This genus exhibits a vesicular shell structure, unlike Catinula and Gryphaea (Bilobissa), and is placed in a separate subfamily (Pycnodonteinae as opposed to Gryphaeinae). Catinula appears to be restricted to Europe. It is worth noting at this juncture that if Catinula gave rise to forms of Gryphaea, but itself arose from Gryphaea (the ‘ Sy Ivester- Bradley ’ model), it would seem most appropriate to regard it as a subgenus of the latter, rather than a separate genus. Thus if the ‘Sylvester-Bradley ’ model were to be vindicated one could not strictly have the iterative evolution of Gryphaea, only of forms referable to the subgenus Bilobissa. Reference to 'Catinula' through the analytical sections of this paper is in accordance with recent custom (i.e. treatment of the taxon as a genus) and carries no implication of our ultimate conclusion concerning the status of the taxon or the course of gryphaeate oyster evolution. A RATIONALE FOR TESTING PHYLOGENETIC MODELS Thus far we have restricted our discussion of phylogenetic hypotheses mainly to the supraspecific level. Two schemes, the ‘Sylvester-Bradley’ and ‘Hallam’ models, which differ in the role accorded to Catinula in the ancestry of European Callovian Bilobissa, remain as viable hypotheses at this level. As partly indicated already, in addition to this dichotomy there exists a variety of views about species-level phylogeny within European Bilobissa and Catinula: in respect of the number of coexisting species, the evolutionary relationships of species, and the tempo of change. Thus whilst agreeing with Hallam over monophyly, Brannan’s (1983) view of species-level evolution in Bilobissa was quite different. In particular, the three Bathonian specimens previously mentioned, considered by Hallam to represent a direct link between European Bajocian and Callovian species, were regarded by Brannan as of uncertain ancestry, providing in themselves no particular support for Bilobissa monophyly. It should be added that they are derived from a highly atypical coral-rich facies (Fischer 1964). The evidence of these forms, notwithstanding the stratigraphic gaps which would still remain, is clearly tenuous. Consequently, the evidence concerning gradual transition from Catinula to Bilobissa, which Hallam has only given very brief consideration (Hallam and Gould 1975, p. 536), assumes prime importance. Other than Sylvester-Bradley, only Brannan has seriously considered this evidence. His analysis led him to a conclusion opposite to Sylvester-Bradley’s : that a morphological discontinuity exists between European Callovian Bilobissa and the latest representatives of Catinula. JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 459 There is a surprising measure of agreement between Brannan (1983) and Sylvester-Bradley (1958, 1959, 1977, MS notes) over specific divisions and phylogeny within Cutiimla. Differences exist over phylogeny in the early Bathonian - Sylvester-Bradley claiming the existence of a semi-discrete lineage confined to the eastern parts of W. Europe, Brannan denying it - but both authors agree that only one Catmula lineage existed in the later Bathonian, persisting into the early Callovian. Species and subspecies constituting this lineage formed the basis of Sylvester-Bradley’s case for gradual transition (a few early Bathonian forms were included), and of Brannan’s for discontinuity, so there are no grounds for suspecting that their contrasting views might result from analysis of material belonging to different lineages. Correspondingly, a reanalysis of Bathonian-Callovian forms, whatever the actual material used, should constitute a valid test of both hypotheses. Sylvester-Bradley (1977, pp. 59-60) considered that the Bathonian-early Callovian lineage identified above was made up of ‘ a succession of forms in which four [gradual] trends are developed : (1) they increase in size; (2) the ribbing gets coarser and less distinct, and in later forms is restricted to early growth stages, or is absent altogether; (3) the left valve deepens so that there is an increase in the angle between the first growth line and the last; (4) a minor but increasing proportion of specimens develop a posterior radial sulcus’. The trends were said to effect a link with Callovian Gryphaea (i.e. Bilohissa). By contrast, Brannan (1983, p. 292) concluded that ’no strong trends either towards or away from gryphaeate or any other type of morphologies exist in the phylogeny of Catimda'. His investigation was based on a much smaller sample than Sylvester-Bradley’s and, notwithstanding the merits of the multivariate approaches used, cannot be said to constitute an adequate test of the assertions relating to single character evolution. We have therefore undertaken an evaluation of the alleged traits, adopting as rationales : ( 1 ) that any demonstration of gradual transitions would place the iterative interpretation of Bilohissa evolution on at least as credible a footing as hypotheses of monophyly involving unaccountable stratigraphic gaps; (2) that Sylvester-Bradley’s criterion for recognizing gradual evolution (occurrence in single characters) is as valid as Brannan’s (trends in values for canonical discriminant functions). Both points could be argued: the cladistic and stratophenetic schools have debated the first and Cheetham (1987) has recently made observations relating to the second. However, we feel our approach is currently justified and hope that the presentation of more data may help towards a resolution of these philosophical questions. We shall also investigate alleged morphological discontinuities which Brannan uses as supporting evidence to conclusions derived from his analysis of supposed transitional links between Catimila and Callovian Bilohissa. MATERIALS AND METHODS The Sylvester-Bradley oyster collection, which formed the basis of our study, included a large quantity of material loaned from institutions in Britain and abroad. This is now in process of being returned but most of the material, personally collected, remains available for study at the Department of Geology, University of Leicester (abbreviated LEIUG). Also available are notebooks detailing location and stratigraphic horizon of samples, unpublished manuscripts, photographs (largely the work of Derek J. Siveter), and a vast compendium of biometric data relating to all the material originally present in the main collection. Further details are given below. Almost all the material originally assembled was from the mid-Jurassic (Toarcian to Oxfordian) interval and consisted of left (or conjoined) valves of oysters referable to Bilohissa or Catinula. Loose right valves were either not collected or separated out at an early stage and stored unprepared. Some two hundred localities, principally in Britain, Erance, Spain, Switzerland and W. Germany, are represented amongst the personally collected material alone. Sampling covered almost all horizons yielding significant numbers of gryphaeate 460 PALAEONTOLOGY, VOLUME 33 oysters in the mid-Jurassic of western Europe. Left valves were grouped according to a morphotype scheme (based on a standard series of measurements) and not according to sample or to species (as diagnosed by Sylvester-Bradley), although this information was preserved with the specimens. Presumably this was some reflection of the intended use of a morphotype-based system of analysis (Sylvester-Bradley 1958). We found it did not assist our research and therefore regrouped the material into the original samples. It is in this form that the material (including separately-bagged right valves and other unprepared material) has been curated at Leicester. The morphotype information can still be related to individual specimens. An explanation of Sylvester-Bradley’s morphotype coding system, elucidated by David J. Siveter and C. P. Palmer, is available with the material. The principal measurements taken by Sylvester-Bradley were as follows: shell height and length, the angle subtended by lines joining the origin of growth with the ventral edge of the attachment area (‘first growth line’) and the ventral margin of the shell (‘last growth line’), the depth of the posterior radial sulcus, and the persistence, height and separation of radial ribs. These were recorded on an interval scale, presumably to facilitate Sylvester-Bradley’s (1958) morphotypic scheme of analysis. We considered this too inaccurate for our purposes and felt that there were certain inconsistencies in the description of ornamental characters. We therefore remeasured non- ornamental characters using a continuous scale and, in view of the difficulty of obtaining precise values for the ribbing characters, adopted a simple presence/absence definition (together with a measure of persistence through ontogeny - see below) in respect of ornamentation. This slightly compromised our evaluation of the second of Sylvester-Bradley’s trends (see above) but we have attempted to make up for this deficiency with illustrations of ornamental variation amongst representative sets of specimens. These latter (text-figs. 8-10; Pis. 1-3) give an indication of the definitions applied herein for ribbed and smooth morphs. The fact that the boundary is somewhat arbitrary, combined with the similar maximum sizes of ribbed and smooth morphs in a given population, provides very clear evidence that populations indeed consist of one species, rather than a mixture of taxa. The measurements taken by us are illustrated approximately in text-figs. 4 and 6 (see below for precise operational definitions). We did not record the proportion of specimens with a posterior radial sulcus (trend 4 above) because the development of this character is clearly size-related and given an increase in size (trend 1) later populations would inevitably include a higher proportion of sulcate forms. Populations consisting only of small individuals clearly manifest a propensity for the development of a sulcus (text-fig. 9e, j). The relative confinement of ribbing to early growth stages (trend 2) is also a redundant parameter given phyletic size increase and if there is a programmed loss of ribbing at some size in ontogeny. This latter is undoubtedly the case - ribbing never extends beyond a peripheral height (P) of 50 mm but we nevertheless measured the size at which ribbing is lost (RP) to investigate possible trends in this character. In order to increase the data base, in a few cases this character was measured on very weakly ribbed specimens, not otherwise recognized (see above) as ribbed morphs. As our means of estimating relative incurvature (the essential character implied in trend 3) we abandoned Sylvester-Bradley’s angular measurement, which is again size-dependent, and substituted ratios of shell dimensions (H/I, H/P) as used by other workers (e.g. Hallam and Gould 1975;Brannan 1983 ; Bayer e? o/. 1985). We chose peripheral height as our measure of size (trend 1) since it is the largest dimension and, unlike height (H), independent of incurvature. refers to the largest single specimen in a population. In addition to measurements taken to test Sylvester-Bradley’s specific claims we investigated length/periphery (L/P) and height/length (H/L) ratios, the direction of the transverse component of coiling (see text-fig. 6f), and the height (AH) and length (AL) of the attachment area, in order to identify any possible morphological discontinuities or further gradual trends. In common with almost all previous work attention was confined to the left valve - in our case principally because of the availability of material. It should be noted that the names applied to the various dimensions are not entirely concordant with any previous scheme but represent a compromise which we hope will be accepted as standard by future workers. JOHNSON AND LENNON; JURASSIC OYSTER EVOLUTION 461 The various shell dimensions are defined precisely as follows (partly adapted from Stenzel 1971, p. N958). Length (L) is the maximum dimension obtained by projecting the extremities of the shell onto the hinge (anteroposterior) axis. Height (H) is the maximum dimension obtained by projecting the extremities of the shell onto a line (the dorsoventral axis) perpendicular to the hinge axis and lying within the plane of commissure. Inflation (I) is the maximum dimension obtained by projecting the extremities of the shell onto a line perpendicular to both the latter line and the hinge axis. Attachment area length (AL) is the maximum dimension obtained by projecting the extremities of the attachment area onto the hinge axis. Attachment area height (AH) is the maximum dimension obtained by projecting the extremities of the attachment area onto a line perpendicular to the hinge axis and lying within the plane of the attachment area. Peripheral height (P) is the distance between the origin of growth and the ventral margin, measured along an imaginary line running around the shell exterior, perpendicular to the hinge axis. Peripheral height of the ribbed zone (RP) is the distance along this line from the origin of growth to the ventral edge of the ribbed zone. Approximate illustrations of these dimensions are provided in text-figs. 4 and 6. Other morphological terms are explained by Stenzel (1971). Whilst we felt unable to make use of Sylvester-Bradley’s biometric data we would emphasize its availability and suggest that it might facilitate future testing of our conclusions, perhaps through a more sophisticated analysis of ornamentation. Our own raw data and statistics are deposited with the collection at Leicester and also with the British Library, Boston Spa, Wetherby, Yorkshire, LS23 7BQ, U.K. as Supplementary Publication No. SUP 14036 (23 pages). Mean sample size for individual statistics - excluding for which entire ‘populations’ (see below) were samples - averaged just under 37. ANALYSIS OF BATHONIAN AND CALLOVIAN FORMS In order to assess the validity of the Bathonian-Callovian section of route I for text-figure 1 we made use of the following ‘populations’ in the Sylvester-Bradley collection. They constitute the largest and stratigraphically best-defined samples from this interval. All are from western Europe to avoid inclusion of a possibly separate ‘eastern’ lineage (see above). We would point out the existence of further material to investigate both this latter topic and the validity of the conclusions reached below. Our survey of populations allegedly representative of the eastern lineage revealed that they span a much shorter interval than supposed by Sylvester-Bradley (being of early Bathonian age rather than as stated in the 1959 and 1977 papers); thus their incorporation into the present analysis could have little affected the overall results. PBA: Port-en-Bessin, Calvados, France; base of the Marnes de Bessin (temdplicatus zone). Notebook reference: S51 PBA3, ‘Cliff section on the west side of Port-en-Bessin harbour. About 1 5 ft of clays (“ Marnes de Port-en-Bessin”) with harder bands of marlstone, overlying the “Passage Beds” (c. 1 ft)...’. ALAJ visited this section in 1984 and found abundant comparable material in a c. 3 cm shell-bed about 6 m above the base of the ‘Marnes’. Less clearly in situ material, possibly derived from a second bed, was found at a level some 4m below. 61 left valves (also 4 ‘miscellaneous’ specimens); LEIUG 104604-104668. WWA : Withy Wood Lane, W. Cranmore, Somerset ; float almost certainly from Rugitela Beds (hodsoni zone). Notebook reference: S49 WWA, ‘Ploughed fields at top of Combe Bottom’, Grid reference: ST 61 9 All . lA left valves (also 1 ‘miscellaneous’ specimen): LEIUG 104373-104449. CVA: Colleville-sur-Orne, Calvados, Erance; Lower Cornbrash equivalent (discus zone, discus subzone). Notebook reference: S51 CVA, ‘Louis Guillaume collns’. In 1984 ALAJ was unable to find any sections at this horizon around Colleville, but nearby coastal exposures between Lion and Luc yielded abundant comparable material. 36 left valves: LEIUG 104474-104509. EA: Le Eresne d’Argences, Calvados, Erance; Upper Kellaways Clay equivalent (calloviense zone, koenigi subzone). Notebook reference: S57 EA, ‘...oysters (O. alimena) from base of brick pit (along drainage trench) - a thin (1 ft) layer of marly limestone and clay ... ’. This horizon was not exposed when ALAJ visited the now defunct brick pit at Argences in 1984. J. H. Callomon (pers. comm., 1985) has however confirmed the presence 462 PALAEONTOLOGY, VOLUME 33 of comparable oyster material at this level. 55 left valves (also 5 ‘miscellaneous’ specimens): LEIUG 104669-104728. PLl, PL2: Putton Lane, Chickerell, Dorset; Upper Kellaways Clay (PLl : calloviense zone, koenigi subzone) and Kellaways Rock (PL2: calloviense zone, calloviense subzone). Notebook reference: S57 PL, ‘Putton Lane brickyard ... Arkell’s [1947, p. 27] Beds I and 2, with large cementstone concretions 1 ft 6 ins below top of section. Uncontaminated samples of Bed 1 difficult to collect, but a collection was made from clays dug out of the flooded base of the pit to form a retaining wall for a sump. Possibly slightly contaminated’. This pit has been long defunct and Bed 1 obscured. However, there are reports of a restart to working. PLl ; 105 left valves (also 38 ‘miscellaneous’ specimens): LEIUG 104729-104871. PL2; 1179 left valves (also 21 ‘miscellaneous’ specimens): LEIUG 104872-106071. KDl : Material collected by K. L. Duff from the London Brick pit at Stewartby, Bedfordshire; Kellaways Rock/Lower Oxford Clay (calloviense zone, enodatum subzone). Notebook reference: S76 KDl, ‘Stewartby, Bed 4 [of Callomon, 1968, pp. 281-2]’. 143 left valves: LEIUG 69964, 69967, 69970, 69971, 69973-69975, 69978, 69980, 69981, 69988, 69991, 70041-70075, 70077-70085, 70090-70099, 70101-70176, 70936. The relative stratigraphical positions of the samples are shown in text-figure 3. KDl was only investigated at a late stage in our study and results were only derived for size and incidence of ribbing. Zones Subzones Populations z < > o < o Calloviense Enodatum KD1 Calloviense PL2 Koenigi FA PL1 Macrocephalus Kamptus Macrocephalus 0 1 1- < CQ Discus Discus CVA Hollandi Aspidoides Hodsoni WWA Morrisi Subcontractus Progracilis T enuiplicatus PBA TEXT-FIG. 3. Stratigraphic position of Bathonian and early Callovian Cat inula and Gryphaea (Bilohissa) populations in relation to ammo- nite zones and subzones. Bio- stratigraphic scheme is that of Cope, Duff et al. (1980). JOHNSON AND LENNON : J U RASSIC OYSTER EVOLUTION 463 Appraisal of alleged discontinuities In addition to finding no evidence for gradualistic links between Catinula and Callovian Bilohissa, Brannan (1983) claimed the existence of important morphological discontinuities between the taxa; in degree of incurvature, and in two internal features, commissural shelf development and muscle scar shape. The latter claim had been previously advanced by Stenzel (1971). We investigated all three claims in respect of forms from the Bathonian-Callovian interval. Relative incurvature. Brannan (1983, p. 291 ) considered that late Bathonian populations of Catinula, allegedly transitional to Bilohissa, could be distinguished from the latter on the basis of degree of incurvature (implying in this case height/inflation ratio). However, regression lines for H against I (text-fig. 4) do not suggest any fundamental discontinuity between Bathonian Catinula and Bilohissa. Rather, there is a marked trend towards higher H/I values with higher stratigraphic position amongst the studied populations. This is interrupted only by the CVA regression which, being based on the smallest statistical sample (28), may be least representative. Of particular interest is the FA population, which, judging from Brannan’s taxonomic scheme and record of stratigraphic range, might well have been placed by him in Gryphaea (Bilohissa) alimena (see also Sylvester- Bradley’s assignment above). This species was considered by Brannan to be separate from the main (unpreserved) Bilohissa lineage leading to later early Callovian forms. However, the intermediate position of the regression for FA in text-fig. 4 suggests that, rather than being a side-issue to a story of monophyletic Bilohissa evolution, such populations actually provide evidence to support a quite different hypothesis; namely, that Callovian Bilohissa evolved from Catinula (as suggested by Sylvester-Bradley from coiling considerations). This issue is considered in full below in conjunction TEXT-FIG. 4. Least squares v-on-.v re- gressions for height (H) versus inflation (I) of left valves from Catinula and Gryphaea (Bilohissa) populations from the Bathonian and early Callovian. Numbers indicate relative age of populations (5 = youngest). For actual stratigraphic positions of populations see text-fig. 3. Dimensions in mm. 0 10 20 464 PALAEONTOLOGY. VOLUME 33 with a more refined analysis of relative incurvature. For the purposes of diagrammatic representation and argument we henceforth regard FA as a very late Catimda population. Notwithstanding his views in general concerning G. (B.) alimerui, it is only fair to add that Brannan might also have preferred to associate population FA with Catimda (e.g. 1983, p. 105), a taxon in his view unrelated to Bdohissa. Internal features. Brannan (1983) claimed that Catimda and Bdohissa could be distinguished on the basis of two internal features: the presence of a marked commissural shelf (a ledge parallel to and just inside the margin of the left valve) in the former, and of an adductor scar with a strongly convex dorsal margin in the latter. Text-fig. 5 shows internal views of left valves from populations which Brannan would refer to Catimda (CVA) and Bdohissa (PLl). It is evident that the supposedly diagnostic features are highly variable and that a marked commissural shelf may occur in Bdohissa while an adductor scar with a convex dorsal margin may be developed in Catimda. Clearly there is no justification for considering that Catimda and Callovian Bdohissa represent entirely separate lineages on the basis of these characters. It therefore remains to assess whether there is any evidence for intergradation amongst other characters. TEXT-FIG. 5. Internal views of left valves from populations belonging to allegedly discrete Gryphaea (Bdohissa) (PLl ) and Catimda (CVA) lineages, showing the range of variation in muscle scar (ms) shape and commissural shelf (cs) development in each population. See text for further details. PLl, left to right: LEIUG 104753, 104737, 104742. CVA, left to right: LEIUG 104477, 104479, 104480. All x2-3. Intergradation Size and rihhing. In connection with the trends reported by Sylvester- Bradley (1977) in these characters, we investigated size (measured by maximum peripheral height), incidence of ribbing (measured by the proportion of ribbed to non-ribbed morphs), and coarseness and distinctness ol ribbing (assessed visually). JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 465 The proportion of ribbed to non-ribbed morphs shows a clear trend, a gradual decrease in the proportion of the former, through the sequence of populations (text-fig. 6a). A progressive increase in maximum peripheral height is also apparent (text-fig. 6b). However, in this case the rate of change is markedly faster amongst early Callovian Bilohissa populations. This fact, together with the geographical variation in both size and incidence of ribbing evinced by the FA and PLl populations (from the same ammonite subzone), allows of the faint possibility that Catimila and Bilohissa constituted two separate lineages, the latter replacing the former following immigration into Europe in the koenigi subzone. This hypothesis fails, however, to account for the evolution of a more Gryphaea-WkQ form in Catimila. Also, of course, it is not substantiated by the existence of an appropriate Bilohissa lineage outside Europe during the Bathonian. One could make the ad hoc suggestion that the gryphaeate trend in Catimila is the result of introgression of Bilohissa genes (implying the existence of two lineages, reproductively incompletely isolated). However, this would be to place yet further demands on the incompleteness of the fossil record. To be tenable, an introgression model - corresponding to some form of ‘reticulate’ evolution (Sylvester-Bradley 1977) -surely requires support from the occurrence of fairly frequent Bathonian Bilohissa in Europe; at least more than the paltry three specimens known. On the grounds of parsimony, the most reasonable interpretation of the above data is in terms of a single late Bathonian-early Callovian lineage. Text-figure 7 shows that the largest representatives of the EA population (‘latest Catimila' \ text- fig. 6a) are approaching the fully gryphaeate form of examples from the PLl population (‘first Callovian Bilohissa'; text-fig. 6a). The same specimens are seen in lateral view in Plates 1 and 2 (respectively figs. 6 and 4) which also depict ornamental variation (and the general range of shape). It can be seen that there is little difference between populations FA and PLl in respect of style of ribbing, and no other distinguishing features immediately present themselves. Plates 1 and 2 are part of a series (text-figs. 8-10; Pis. 13) intended to facilitate assessment of the supposed overall trends in the coarseness and distinctness of ornamentation. It is possible to perceive a gradual trend towards coarser ribbing (as defined by longer wavelength) through the sequence of populations, although the presence of individuals with relatively few ribs in CVA should be noted. No unidirectional trend in distinctness (amplitude) of ribbing can be recognized - the highest ribs are developed midway through the Catimila series. Nevertheless, it is important to point out that the difference between PBA and CVA (i.e. within Catimila) is probably as large as that between CVA and FA, and certainly larger than that between FA and PLl. Both of these latter ‘discontinuities’ might have been argued to represent displacement of Catimila by a separate Bilohissa lineage. To summarize the results thus far: there is compelling evidence for the evolution of Catimila into Bilohissa through a gradual, unidirectional trend in the proportion of ribbed morphs. Slightly more equivocal trends exist in maximum peripheral height and wavelength of ribbing. Rib amplitude follows an oscillatory pattern but this character, and the last two, gives no suggestion of a real discontinuity between Catimila and European Callovian Bilohissa. Gross shell dimensions. We pointed out above that Sylvester-Bradley’s angular measurement does not allow a true evaluation of relative incurvature (left valve ‘depth’; trend 3). Analysis of height (H) in relation to inflation (I), as carried out in connection with Brannan’s claims (see above), represents a better means of assessment but is subject to the difficulty of measuring inflation accurately in small specimens. We investigated incurvature through an analysis of height in relation to peripheral height (P), calculating log log regressions in accordance with the allometric relationship between these characters. Similar investigations were made of length (L) against peripheral height, and height against length. The results for the three critical earliest Callovian populations are presented in text-figure 6c-e. As expected, the H/P regressions differ but there also exist differences between the populations in respect of L/P and H/L. Points on the regressions corresponding to the largest individuals in each population have been identified. The dashed lines represent ‘secondary’ regressions calculated from the coordinates of these points. An interesting fact emerges from this analysis. In each plot the slope 466 PALAEONTOLOGY, VOLUME 33 Zones CALL. 1 Calloviense Macroceph, 1 BATHONIAN 1 Discus Aspidoides Hodsoni Morris! Subcontrac. Progracilis Tenuiplic. 50 100 10 50 100 10 50 100 Zones TEXT-FIG. 6. Biometric data for left valves from Bathonian and early Callovian Cafinula and Gryphuea (Bilohissa) populations, a, b. Stratigraphic variation in: a, proportion of ribbed (solid) to smooth morphs; b, maximum peripheral height Larger value for of CVA population derived from a possible contaminant (see text-fig. 10a). c-e. Least squares v-on-.v regressions (form log y = a log .v-!-log axes log scale) for shell proportions of three early Callovian Ccitimda and Bilohissa populations (continuous lines). Dashed lines are regressions (slope - a- indicated) calculated from values (solid circles) corresponding to the JOHNSON AND LENNON; JURASSIC OYSTER EVOLUTION 467 TEXT-FIG. 7. Anterior views of left valves from populations FA and PLl, showing the approach toward the fully developed gryphaeate coiling of the latter population within the former. LEIUG 104674 (FA), 104732 (PLl); x F7. {a) of the secondary regression is close to unity, implying that the shape of the largest (‘adult’) individuals in each population is extremely similar (cf. Gould 1977, p. 239). To use Gould’s terms, the larger adults of the PL2 population are ‘proportioned giants’. The relationship is preserved if the earlier populations are included in the analysis {a = 0-88, 0-93, 0-98 for secondary regressions of, respectively, H/P, L/P, H/L) but clear graphical representation of the full data set cannot be easily accomplished. This maintenance of adult geometric similarity can be readily interpreted in the context of dissociated size and shape development, and as such strengthens the case for a direct relationship between Callovian BUohissa and Catimda. However, the observed relationship could result from either retarded shape development and greater longevity in populations reaching a larger size, or from accelerated size increase (and unaltered longevity and rate of shape development) in such ‘giant’ populations. The latter might in turn be no more than an aspect of ecophenotypic variation. Ignoring for the moment the seemingly conflicting evidence of differences in the relative frequency of ribbed morphs, this would mean that Catinula and BUohissa were not simply directly related but, in fact, effectively genetically identical ! The link between size and stratigraphic position perhaps favours an evolutionary (i.e. genetic) interpretation, whether involving retarded shape development or accelerated size development, but the apparent existence of geographical variation within the koenigi subzone lends support to the ecophenotypic view. We shall return to this question in discussion of earlier Catinula and BUohissa from near the Lower- Middle Jurassic boundary. From the analysis of H/P it is evident that adult shells show no change in incurvature, contrary to what is implied in Sylvester-Bradley’s trend 3. However, as we have shown, far from refuting the maximum size (.v dimension) in each population, c, height (H) versus peripheral height (P); d, length (L) versus peripheral height; E, height versus length, f-i. Stratigraphic variation in: f, proso- (left), to ortho-, to opisthogyral (right) morphs; G, attachment area height (AH); h, attachment area length (AL); i, peripheral height of ribbed zone (RP). Bars in cui extend one standard deviation either side of the mean. Dimensions in mm. Divisions of ammonite zones are subzones. Line diagrams illustrating characters are of exteriors of ribbed morphs in B, E, i; remaining line diagrams of smooth morphs. Halftone illustrations - PL2 : LEIUG 104872 (left), 104874; WWA; LEIUG 104379 (left), 104382; PBA; LEIUG 104623; all xO-55. 468 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 8, Catimila Rollier - population PBA (a-t: LEIUG 104604-104623 respectively): variation in external morphology of left valves. All specimens are ribbed morphs; x TV. possibility of a link between Carinula and Bilobissa, the very fact of maintenance of shape in the context of differing adult sizes provides compelling support for a relationship. Additional features. Data for measures of shell obliquity, attachment area size and peripheral height of the ribbed zone are presented in text-figure 6f-i. These characters, supplementary to those considered by Sylvester- Bradley (1977), provide no grounds to support the view that Catinula and Callovian Bilobissa represent anything other than segments of a single lineage. In the case of obliquity (text-fig. 6f), some difference exists between the critical, approximately contemporaneous, FA {Catinula) and PLl (Bilobissa) populations. However, this is comparable to the difference between successive, earlier populations - PBA and WWA - both referable to Catinula. Small differences also exist between FA and PLl in respect of mean dimensions of the attachment area (text-fig. 6g and h) but here again a discontinuity cannot reasonably be inferred in view of the complete overlap of bars representing one standard deviation from the mean. In the case of peripheral height of the ribbed zone (text-fig. 6i), values for the FA and PLl populations are almost identical, and the pattern of stratigraphic change in this character over the complete sequence of populations could be interpreted as a mildly oscillating gradual trend ; positive evidence, under this view, of a link between Catinula and Bilobissa. Summary and conclusions There is no compelling evidence for the existence of a morphological discontinuity between L. Callovian oysters referable to Bilobissa and Bathonian-L. Callovian forms referable to Catinula. The existence of a gradual, unidirectional trend towards reduced frequency of ribbed morphs, together with somewhat less uniform trends in maximum size, coarseness of ribbing and peripheral height of the ribbed zone, provides, in contrast, positive evidence that L. Callovian Bilobissa evolved from Catinula. This conclusion is strongly supported by close similarities in gross adult shell proportions, despite differing adult sizes. JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 469 TEXT-FIG. 9. Catinula Rollier - population WWA (a-v: LEIUG 104373-104394 respectively): variation in external morphology of left valves, a, f, g, k, q-s, ribbed morphs; remainder smooth; all x L7. ANALYSIS OF TOARCIAN AND AALENIAN FORMS We have found in favour of the ‘ Sy Ivester- Bradley ’ model (route I in text-fig. 1) for the immediate (Bathonian) ancestry of Callovian Bilohissa. It remains to be shown whether Toarcian G. (Bilobissa) pictaviensis gave rise to Aalenian Catinula heaumonti and thus whether route I is correct in its entirety. Large amounts of material were available to us in the Sylvester- Bradley collection to test this proposition. The large number of G. (B.) pictaviensis samples provided, moreover, an opportunity to test whether this species, variably-ribbed like the later Bilohissa/ Catinula group analyzed above, exhibited a pattern of within-species geographic variation in morphology analogous to that inferred in the later group of forms. The occurrence of an analogous (well- developed) pattern of variation would provide additional support for the interpretation of Callovian Bilohissa as a descendant of Catinula. We also wished further to investigate the possibility that variation might be ecophenotypic. A similar biometric investigation was made of the following nine ‘populations’; the first, following Sylvester-Bradley’s MS notes, referable to C. heaumonti, and the last eight to G. (B.) pictaviensis. 470 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 10. Catimda Rollier - population CVA (a-s: LEIUG 104474-104492 respectively): variation in external morphology of left valves, a, e, j-l, o-q, smooth morphs; remainder ribbed; all x L7. The exceptional size of the specimen figured as a may indicate that it is a contaminant. AVB: Airvault, Deux Sevres, France: Aalenian, opalimmi zone. Notebook reference: S74 AVB, ‘Old quarries in environs of new cement quarry complex. AVA: Lower beds just above unconformity with large oysters (rare) in Bed no. 13 (MS section. = Bed No. 39 of Gabilly 1973) (inactra horizon). AVB: Higher in section’. ALAJ visited Airvault quarry (worked by Societe des Ciments Franipais) in 1984. Large quantities of comparable material were obtained from c. 1 m of limestones and marls including Bed 77 of Gabilly (1973, Annexe 1, fig. 9). There can be no doubt that these horizons were the source of sample AVB. 216 left valves (also 47 ‘miscellaneous’ specimens): LFIUG 101333-101596. TTB, TTC : Tartareu, Lerida, Spain. Notebook references: S75 TTB, ‘ Barley field about 4 km NF of Tartareu with R. [Rhynchonella] cynocephala (about 500 m N of road)’; S75 TTC, ‘Hillside above fields and in maquis about L5 km north of TTB - especially where dam for small pond has been excavated’. TTB; 138 left valves (also 6 ‘miscellaneous’ specimens): LFIUG 103660-103803. TTC; 175 left valves (also 3 ‘miscellaneous’ specimens): LEIUG 70930-70933, 103482-103655. EXPLANATION OF PLATE 1 Figs. 1-26. Catinula Rollier - population FA (LEIUG 104669-104694 respectively): variation in external morphology of left valves. 3, 4, 7, 8, 12, 13, 16-18 and 24-26, ribbed morphs; remainder smooth; all x 1-7. PLATE 1 JOHNSON and LENNON, Catinula 472 PALAEONTOLOGY, VOLUME 33 CRC: Camarasa, Lerida, Spain. Notebook reference: S76 CRC, ‘Off trail to Ermita S. Jiordi ...Clutson unit D3’. No further details of Michael Clutson’s stratigraphic scheme are available and it is uncertain whether it was published. 418 left valves (also 6 ‘miscellaneous’ specimens): LEIUG 106800-107223. NV : Nevian, Aude, France. Notebook reference: S58 NV, ‘Map 1 : 50000 coloured edit. (Type 1922, published 1955) - Sheet Beziers (XXV - 45). Exposure in vineyard banks on eastern slopes of small hill 1 km S by E of Nevian - [grid reference] 646 0 x 100-5. The oysters are found through vineyards extending a considerable way up the hill. It seemed that those in the higher vineyards (geographically; stratigraphical relationships not decided) were wider and less often had ribbed umbones than those from the lower vineyards ; this suggests that two horizons may be implicated’. 213 left valves: LEIUG 70934, 106582-106793. TZ: Chateau Taziere, Fourchambault, Nievre, France. Notebook reference: S74 TZ, ‘Mapped as a faulted inlier of Aalenian ... .To the east of the chateau field is apparently developed on floor of old clay pit. Fossils come from a slab of limestone weathering out of wall of pit below wall of garden of chateau’. 233 left valves: LEIUG 104012-104245. LBB: La Bonnette valley, St Antonin, Tarn et Garonne, France. Notebook reference: S58 LLB, ‘On east side of valley, natural section at foot of scars, above scree slopes. About 10' yield Grvphaea ... LBB from scree. [Grid reference] 553-5 x 207-5 - Map 206 SE [Type 1889, 1 : 50000 - Cahors]’. 56 left valves: LEIUG 106272-106327. BZ: Bizanet, Aude, France. Notebook reference: S58 BZ, ‘Oysters from outcrop off Ruisseau de la Sauzine NNE of Bizanet. [Grid reference] 643-6 x 97-4 - Map 1 : 20000 Capendu No. 4 (XXIV - 46 - No. 4). A small quarry immediately above vineyard. Oysters collected from both quarry and vineyard. In quarry found at two horizons, about 6ft apart’. 126 left valves (also 1 ‘miscellaneous’ specimen): LEIUG 61200-61214, 106160-106271. AG: Chateau d’ Aguilar, Tuchan, Aude, France. Notebook reference: S58 AG, ‘Oyster beds outcrop between fossiliferous Whitbian (Hildoceras, big pectens, belemnites, terebratulids) shales and unfossiliferous massive Bajocian limestones in col immediately north of chateau. In situ in vineyard banks, and ploughed up in vineyards. (Perhaps mainly upper horizon present?)’. 88 left valves: LEIUG 106072-106159. Sylvester- Bradley (MS notes) apparently considered that all the G. (B.) pictaviensis samples (last eight) were from the levesquei zone of the Toarcian. We have accepted the opinion of J. Gabilly (University of Poitiers; pers. comm., 1984), founded on detailed work in the Poitou region of France, that G. {B.) pictaviensis does not occur outside this zone. C. beaumonti (restricted to Poitou) appears in the succeeding opalimim zone (Gabilly 1973). Size and ribbing. Data for these characters are presented in the form of a bivariate plot (text-fig. 1 1a). From this it is clear that G. (B.) pictaviensis exhibits considerable inter-population variation, of a magnitude rather larger than that observed in the koenigi subzone of the Callovian and, in that case, ascribed to geographic variation within one species. Even given that the G. {B.) pictaviensis samples may not all be from precisely the same horizon (see locality details), it seems highly probable that there was a good deal of geographic variation in this species. By implication therefore, variation in the koenigi subzone may be confidently accepted as intraspecific - developed within a single Catinula/ Bilobissa lineage. A regression calculated from the G. (B.) pictaviensis data passes remarkably close to the value for the C. beaumonti population (AVB). The whole array of data points can thus be interpreted in terms of a simple pattern of covariation. This, together with the near identity of ribbing form (see text- EXPLANATION OF PLATE 2 Figs. 1-24. Grvphaea (Bilobissa) Stenzel - population PLl (LEIUG 104729-104752 respectively): variation in external morphology of left valves. 8, 13, 17 and 18, ribbed morphs; remainder smooth; all x 1-7. PLATE 2 JOHNSON and LENNON, Gryphaea (Bilohissa) 474 PALAEONTOLOGY, VOLUME 33 Opalinum Levesquei Rrmr snnn rrnri^ Mill 1%I 1 1 1 1 1 WY////A nil'll — mum — AH AVB TTC - TTB CRC NV TZ LBB BZ - AG - - o”" AL H RP TEXT-FIG. 11. Biometric data for left valves from Aalenian Catinula beaumonti and Toarcian Gryphaea (Bilobissa) pictaviensis populations, a. Plot of percentage smooth morphs versus maximum peripheral height (P^ax)- Regression (least squares v-on-.v) calculated from data for G. (B.) pictaviensis populations (solid circles). B-D. Least squares v’-on-.x regressions (form log v = a log x + log b, axes log scale) for shell proportions of C. beaumonti (AVB) and four G. (B.) pictaviensis populations. Dashed lines are regressions (slope-u-indicated) calculated from values (solid circles) corresponding to the maximum size (.v dimension) in each population, b, height (H) versus peripheral height (P); c, length (L) versus peripheral height; d, height versus length, e-h. Stratigraphic and geographic (G. (B.) pictaviensis, levescpiei zone) variation in: E, proportion of proso- (left), to ortho-, to opisthogyral (right) morphs; F, attachment area height (AH); G, attachment area length (AL); H, peripheral height of ribbed zone (RP). Populations arranged in order of decreasing maximum peripheral height upwards. Bars in f-h extend one standard deviation either side of the mean. Dimensions in mm. Characters illustrated in text-fig. 6. Halftone illustrations - G. (B.) pictaviensis: ONCP EM 35001 (left), LEIUG 61920; C. beaumonti: LEIUG 61452 (left), 101333; all xO.4. JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 475 fig. 11a), bears out Sylvester-Bradley’s (1959, MS notes) and Brannan’s (1983) conclusions regarding the evolution of C. heaumonti from B. pictaviensis. In that the available data (text-fig. 6a, b) point to an inverse relation between size and the proportion of ribbed morphs, a comparable form of covariation also exists in the koenigi subzone. However, the details of the relationship are different (the two populations from the koenigi subzone would plot well above the levesquei-zor\Q regression in text-fig. 11a), so in respect of the actual morphologies developed, information derived from study of geographical variation in G. (B. ) pictaviensis tends only qualified support to the idea that but one lineage exists in the koenigi subzone. However, the fact that the whole array of Bathonian-Callovian populations displays a quite clear relationship between size and the incidence of ribbing (text-fig. 6a, b) provides further evidence that they form part of an evolutionary continuum, even though the details of the relationship are obviously different from those applying to G. (B.) pictaviensis. We noted above in discussion of gross shell dimensions of Bathonian-Callovian forms the possibility that variation might be of an ecophenotypic rather than genetic nature (cf. Johnson 1981), being related to differences in rate of size increase. Clearly, since the proportion of ribbed morphs in Toarcian-Aalenian populations shows such a close relationship to adult size, we must also consider the possibility that this character is under ecophenotypic control. However, unlike gross shell dimensions, whose allometric growth must inevitably result in different patterns of size/shape development if the relative rate of size increase is altered, there is no obvious reason why a higher incidence of ribbing should result from stunting (or more smooth forms from the reverse). For this reason genetic rather than ecophenotypic explanations for ribbing variation must be preferred. However, it is worth noting that Medcof (1949) documented ecophenotypic development of ribbing (albeit of a rather different form) in conjunction with reduced rates of growth in the living oyster Crassostrea virginica (Gmelin). In this case the determining factor appears to have been exposure to sunlight. Other cases of ecophenotypic variation in rib strength are known (Stenzel 1971). Gross shell dimensions. Data for four of the G. {B.) pictaviensis populations and for the single C. heaumonti population are presented in text-figure 11b-d. Secondary regressions (see discussion of Bathonian-early Callovian forms) calculated from values corresponding to the largest individuals in each population have slopes near to unity in the case of the H/P and H/L plots. In respect of these ratios, therefore, adult shape is almost identical amongst the populations and can be interpreted, as before, as a product of heterochronic change in the rates of either size or shape development. In view of the stratigraphic equivalence of four of the populations, an ecophenotypic control is conceivable. The relatively poor approximation of the L/P secondary regression to a slope of unity might well be a result of the small size range of the AVB sample (P; 14-43 mm), leading to an unrepresentative primary regression and consequently inappropriate coordinates for ‘adult’ L/P (derived by projection of the value for onto the regression) in this sample. There is therefore no special reason to doubt that overall adult shape was much the same amongst Toarcian and Aalenian populations, as amongst Bathonian and Callovian populations. Additional features. Data for shell obliquity, attachment area dimensions and peripheral height of the ribbed zone (text-fig. 11e-h) show no significant difference between the opcdinum zone population and the levescpiei zone samples. There are thus no reasons to doubt that C. heaumonti evolved from G. (B.) pictaviensis on the bases of these data. Summary and Conclusions None of the features considered suggests a fundamental discontinuity between G. (B.) pictaviensis and C. heaumonti, and the patterns of variation in size and the incidence of ribbing, and in gross shell dimensions, provide particularly compelling evidence that these conventionally generically- separated species have an ancestor-descendant relationship. Analogies with the pattern of variation 476 PALAEONTOLOGY, VOLUME 33 in Bathonian-Callovian forms support the conclusion that Callovian Bilohissa evolved from Catimla. The earlier group of Bilohissa/ Cat inula provides particularly strong grounds for suspecting that variation might be, at least partly, under ecophenotypic control. This possibility could be substantiated by demonstration of a facies correlation. ALAJ visited Poitou to examine sites yielding forms referable to G. (B.) pictaviensis and C. heaumonti but could find little significant correlation with sedimentary or biofacies. Both forms occur in marls and marly limestones with an apparently fully marine fauna including belemnites and articulate brachiopods. Ammonites are rare in association with C. heaumonti, and other bivalves with G. (B.) pictaviensis, but it is doubtful whether this has great environmental significance. A slight change in water depth may be implied (see also Gabilly 1973). It is possible that temperature differences might have little impact on sedimentation and the general composition of the fauna, and yet influence form in the Bilohissa/ Catinula group. A salinity effect is an alternative explanation, but rather less plausible given the apparently marine fauna associated with both G. {B.) pictaviensis and C. heaumonti. Certainly mollusc growth rates are affected by both these factors (Vermeij 1980; Tevesj and Carter 1980). Isotopic analysis of shell material would be a way of evaluating these possibilities (Tann and Hudson 1974; Rye and Sommer 1980). A control by the amount and/or intensity of sunlight is worth considering (cf. above) but, other than perhaps variation in aquatic plant growth, it is difficult to conceive a cause for significant sunlight variation, given the geographical proximity and apparently similar palaeoenvironments of the populations. We have attempted to investigate further the possibility of ecophenotypic variation by an analysis of size in relation to age (as determined by ligamental growth bands; cf. Hallam 1982). Thus far the investigation has shown only the difficulty of reliable age determination in relatively small shells. However, this method provides, at least in principle, a means for establishing that differences exist in the rate of size (rather than shape) development; a reasonable basis in our view for considering allometry-related ‘static’ variation to be ecophenotypic. In the absence of such information it is best to assume that all the variation described herein is genetic (see also above). B'rannan (1983) considered that the evolution of C. heaumonti from G. (B.) pictaviensis represented an example of progenesis (cf. Gould 1977). However, we have shown that the transition involves something more than a simple truncation of development because of the increased frequency of ribbed morphs. Moreover, in the case of gross shell dimensions, the relative rates of size and shape development are altered. We would nevertheless agree with Brannan that the origin of Catinula in a small area (Poitou) at the margins of the ancestral species’ range represents a classic case of allopatric speciation. Insofar as the evolution occurs between adjacent ammonite zones and in the time taken for the deposition of only 2-3 m of sediment it also appears to represent a case of punctuational speciation (Eldredge and Gould 1972). However, morphological stasis either side of the evolutionary burst remains to be demonstrated, and in our view does not exist (see also Brannan 1983). EVIDENCE FROM OTHER MIDDLE JURASSIC CATINULA AND BILOBISSA Brannan (1983) has reviewed the morphology and taxonomy of the Aalenian-Bathonian group of Bilohissa and Catinula stratigraphically intermediate between the two groups considered in detail above. He found no difficulty in differentiating Bilohissa and Catinula, a point which we can endorse in the case of this group of forms. There exist populations of medium to large-sized forms which never develop umbonal ribbing on the left valve and populations of small to medium-sized forms EXPLANATION OF PLATE 3 Figs. 1-20. Gryphaea (Bilohissa) Stenzel - population PL2 (LEIUG 104872-104891 respectively): variation in external morphology of left valves. 2, 3 and 11, ribbed morphs; remainder smooth; all x 1-7. PLATE 3 JOHNSON and LENNON, Gryphaea (Bilohissa) 478 PALAEONTOLOGY, VOLUME 33 Stages Zones G. (Bilobissa) Catinula CALLOVIAN Jason Calloviense ... Macrocephalus BATHONIAN Discus 1 Aspidoides Hodsoni 1 Morrisi Subcontractus i • Progracilis Tenuiplicatus 1 1 Zigzag "I BAJOCIAN Parkinsoni • 0 Garantiana 1 I Subfurcatum Humphriesianum 1 Sauzei Laeviuscula Discites AALENIAN Concavum Murchisonae Opalinum TOARCIAN Levesquei Thouarsense Variabilis TEXT-FIG. 12. Zonal occurrence of Gryphaea (Bilohissa) and Catimda in the European mid-Jurassic. Based on material in the Sylvester-Bradley oyster collection and records in Bayer et al. (1985). Gryphaea (Bilohissa) occurrence in the early Bathonian (G. (B.) gallica Fischer, 1964) arbitrarily assigned to the tenuiplicatus zone. Zonal scheme is that of Cope, Duff et al. (1980) and Cope, Getty et al. (1980). which include smooth and ribbed individuals. The first set can be assigned unequivocally to Bilohissa and the second, in view of the lack of any tendency for populations of relatively large individuals to show a reduced incidence of ribbing, to Catimda. It seems therefore that two lineages existed side by side during the Aalenian to early Bathonian interval. The implied genetic distinction supports our conclusion that the transitions from Bilohissa into Catimda, and subsequently of Catinula into Bilohissa, represent genuine evolutionary changes. Nevertheless, in plotting the zone- by-zone occurrences of Catimda and Bilohissa through the mid-Jurassic interval (text-fig. 12) we have noted the interesting fact that distributions are largely mutually exclusive. This again gives pause for thought that the two forms might be ecophenotypes (the products of secular environmental change), although oscillatory evolution is perhaps equally plausible. In the absence of intergradation in Aalenian to early Bathonian forms the most appropriate interpretation remains, however, that two separate lineages existed during this interval. The ‘Sylvester-Bradley’ model for the ancestry of Callovian Bilohissa (route I in text-fig. 1) can therefore be accepted in its entirety. The absence of intergradation in Aalenian to early Bathonian forms also renders implausible any explanation for the morphological trends subsequently occurring in Catinula in terms of the introgression of Bilohissa genes (cf. above). A ‘reticulate evolution’ model (Sylvester-Bradley 1977), involving gene transfer between coexistent Catinula and Bilohissa, cannot therefore be sustained. It remains to be seen whether the reticulate evolution model is applicable in the more restricted sense in which it was actually proposed by Sylvester-Bradley for mid-Jurassic gryphaeate oysters. It was implied that gradual evolution in Catinula was the result of geographical differentiation within the taxon, followed (necessarily, for the applicability of the model) by introgression between demes. In fact, as noted above, those populations allegedly evincing geographical differentiation (supposedly forming a semi-discrete ‘eastern’ lineage) are all of early Bathonian age. Thus only for this interval does it seem possible that evolution in ‘western’ populations (analyzed above) was influenced by introgression. The evolution of the highly variable species Gryphaea (Bilohissa) pictaviensis JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 479 into relatively invariant Gryphaea (Bilohissa) and Catinula lineages could be interpreted as representing the first (‘eruptive’) and last (‘stabilized’) phases of a reticulate evolution cycle. However, it is not possible to say whether there was an intervening (‘reticulate’) phase involving introgression between denies. Following our rejection of Hallam’s model for the origin of European Callovian Bilohissa (route II in text-fig. 1) we must conclude that European Toarcian-early Bathonian Bilohissa represent a lineage that became extinct sometime in the Bathonian. If we follow Brannan’s (1983) proposition that the three early Bathonian Bilohissa specimens (referred to Gryphaea {B.) gallica Fischer) are unrelated to earlier forms, then the last representatives of the lineage may be late Bajocian. A MODEL FOR THE DEVELOPMENT OF RIBBING IN GRYPHAEATE OYSTERS We have noted that in both the Toarcian-Aalenian and Bathonian-early Callovian groups of Catinula and Bilohissa considered above, geometrical similarity is maintained in adult forms despite differences in size. This might result from changes in the rate of shape development proportionate to length of life (and hence time available for size increase). However, a simpler alternative is that it results from a variable rate of size increase and a fixed longevity and rate of shape development. If we assume that Catinula and Bilohissa were characterized by different rates of size development we can develop a conceptual model which accounts for the relationship between ribbing incidence and size. The growth rate model is set out in text-figure 13 and requires the existence of a ‘zone of ribbing’ in size/age space. Its shape is defined by the need to account for: (1) ubiquitous ribbing in PBA (the most ‘catinulate’ population; smallest and with inferred slowest growth), combined with the late ontogenetic development of a smooth shell in the largest individuals in this population (text-fig. 8f); (2) a complete absence of ribbing in KDl, the Bilohissa population attaining the largest size (inferred fastest growth). A form such as Gryphaea (B.) pictaviensis fits satisfactorily into this model because, with its smaller average maximum size, it may be predicted to have had a lower growth rate and would thus pass through the ‘zone of ribbing’ in early ontogeny. That G. (B.) pictaviensis has an intermediate rate of growth could, in theory, be tested by determinations of age in relation to size through analysis of growth-bands (see above). Analysis of other mid-Jurassic forms would provide a further test of the model’s applicability. The Aalenian and’ Bajocian species G. (B.) calceola Quenstedt and G. {B.) suhlohata (Deshayes) both completely lack umbonal ribbing. A specimen (LL 35353) of the latter in the British Museum (Natural History) has a peripheral height of 260 mm, far in excess of the unribbed early Callovian population KDl, so the lack of ribbed morphs in G. (B.) suhlohata, which presumably grew very rapidly, accords with expectation. By contrast the maximum peripheral height of G. (B.) calceola is about 75 mm, equivalent to that of the smallest (frequently ribbed) G. (B.) pictaviensis population. This anomaly might be explained by growth at a rate similar to KDl but death at a relatively young age (text-fig. 13), a proposition which could be tested by growth-line analysis. There appears to be no difference in the size attained by smooth and ribbed morphs in populations including both. This probably implies within-population variation in the shape of the zone of ribbing since a variety of age/size curves, all terminating at the same size, seems a less plausible explanation. This, in turn, is most easily interpreted as an aspect of genetic, rather than ecophenotypic, variation. The existence of genetic variation of this type also helps to explain the non-monotonic relationship between maximum size and proportion of ribbed morphs in Bathonian-early Callovian Catinula and Bilohissa. Therefore, we again reach the conclusion that evolutionary change is involved in this sequence of forms. We may be able to show that there are differences in growth rate amongst the forms under consideration but this does not prove that they cause differences in the development of ribbing - i.e. that a ‘zone of ribbing’ of the shape indicated exists in size/age space. We can however put 480 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 13. Model to explain the occurrence of ribbing in Catinula and Gryphaea (Bilobissa) as a function of growth rate. Curves represent predicted typical growth rates for PBA (a Catinula population entirely composed of ribbed morphs), KDl (a G. (Bilobissa) population entirely lacking ribbed morphs), and an average G. (B.) pictaviensis population (with an admixture of smooth and ribbed morphs). Solid circles represent maximum size (peripheral height: P) in each of these populations, and for the unribbed species G. (B.) calceola which is predicted to have grown rapidly but to have had a short lifespan. See text for further explanation. P in mm. forward functional arguments, relating to habitat and mode of life, for the existence of such a zone. Contrary to the opinion of Brannan (1983), it is our experience that Catinula and Bilobissa occur in remarkably similar sedimentological settings. Both occur most commonly in relatively isolated shell beds within clay-grade sequences. These probably represent winnowed horizons (a slightly coarser grain size and reworked shell material are common associates) which afforded a slightly firmer substrate for colonization (see also Bayer et al. 1985). There is an approximately equal (low) incidence of overturning in Bilobissa shell beds and those containing Catinula, so the energy of the environment cannot have been very different. Occasional occurrences of Catinula in very high- energy oolite deposits (e.g. Brannan 1983, p. 293) may represent transported assemblages. Reclining bivalves, such as gryphaeate oysters, typically display some adaptation towards maintaining stability on the sea floor (Stanley 1970). Two such adaptations, large size and a thick shell, are exhibited by Gryphaea itself ; more particularly the former in the case of Bilobissa. Neither condition can, however, exist without significant growth so these adaptive strategies require rapid growth for maximum efficiency. Given an inability (genetic or otherwise) to grow fast, other strategies for obtaining stability might prove superior. Brannan (1983) has suggested that the ribs of Catinula might function to provide stability by gripping the sediment. While secretion of ribs JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 481 entails a diversion of shell material from pure size increase, it might, at generally small sizes (and under the same hydrodynamic regime), confer a stability greater than that obtained by a slight increase in size. As such it would be a better adaptive strategy in situations where rapid growth to large size is not possible. Obviously, flume tank experiments (cf. Hallam 1968) afford a means of testing this proposition. In conclusion, our suggestion is that the transition from Bilohissa to Catimila and then back to Bilobissa (as manifested in size, ribbing incidence and gross shell dimensions) may represent a response to environmental conditions which favoured, respectively, rapid, then slow, then rapid growth. Of what seem the two most plausible controlling factors, temperature and salinity (see above), the former is more likely to be implicated in view of the occurrence of Catimda in apparently fully marine Aalenian assemblages. However, the notion of salinity control should not be abandoned completely. Bathonian reduced-salinity faunas have been widely documented in northern Europe (Hudson 1980; Hudson and Palmer 1976; Palmer 1979). At this time Catinula was present in huge numbers, to the almost total exclusion of Bilohissa. However, it has to be said that the particular assemblages containing Catinula are south of the main areas of freshwater influence and do not provide any clear indications of abnormal salinity. It is worth noting here that European Bathonian scallops attain only modest dimensions (Johnson 1984), and other bivalves seem to be generally small. TAXONOMY Having shown that Gryphaea (Bilobissa) is diphyletic it would be our preference to separate the earlier and later lineages taxonomically. This would involve creation of a new subgeneric name for the later group of European forms since Bilobissa, whose type species is Bajocian (Duff 1978), is clearly best reserved for the earlier group. In the present state of knowledge we could only offer a ‘stratigraphic’ diagnosis for the new subgenus. This would not appear to satisfy the provisions of Article 13 (a) (i) of the ''International code of zoological nomenclature' (Ride et al. 1985, p. 35) which states that new names must be ‘accompanied by a description or definition that states in words characters that are purported to differentiate the taxon’. European Callovian Gryphaea (and descendant forms) must therefore still be referred to Bilobissa until such time as an apomorphy is discovered which can be used as a basis for the erection of a new subgeneric name. Since Catinula is evidently an integral part of Gryphaea phytogeny it seems appropriate, as indicated earlier, to demote the taxon to subgeneric rank within Gryphaea. The features distinguishing G. (Catinula) from the closely related subgenus G. (Bilobissa) were outlined earlier in discussion of generic differentiation. It remains to give precise definitions of G. (Catinula) and G. (Bilobissa). The following diagnoses are therefore provided, constructed so that the boundaries of the taxa correspond most nearly with previous conceptions of Catinula and Bilohissa: G. (Bilobissa) Stenzel 1971, p. N1099. Medium-sized Gryphaea', adult peripheral height greater than 45 mm. Usually deep radial posterior sulcus in adults with posterior flange well detached. Radial ribbing on umbonal region of left valve present in 0-80% of individuals in populations. Lower-Upper Jurassic; Europe, Asia, N. Africa, N. and S. America. Type species: Gryphaea hilobata J. de C. Sowerby, 1835, p. 244 (= 1840, Alphabetic index, p. 4; = G. dilatata var. f. J. Sowerby, 1816, p. 113, pi. 149, fig. 2), Inferior Oolite (Bajocian), England; original designation by Stenzel (1971, p. N1099). See Duff (1978, pp. 76, 77) for further details. G. (Catinula) Rollier 191 1, p. 272. Small Gryphaea', adult peripheral height less than 45 mm. Radial ribbing present on left (and commonly right) valve of 40-100% of individuals in populations. Middle Jurassic; Europe. Type species: Ostrea knorri Voltz (=0. knorrii Voltz, 1828, p. 60), Bathonian, Schonmatt, near Basel, Switzerland; subsequent designation by Arkell (1932, pp. 149, 180). These diagnoses may be used in conjunction with the extensive generic diagnoses of Stenzel (1971) and Duff (1978), amended slightly to incorporate the findings of this study so that Gryphaea 482 PALAEONTOLOGY, VOLUME 33 includes: a, forms which are small as adults and therefore neither particularly thick-shelled nor enrolled; h, forms with a weakly convex dorsal margin to the adductor scar; and c, prosogyrous forms. CONCLUSIONS (1) Forms of Gryphaea referable to the subgenus Bilohissa have evolved iteratively (at least twice). (2) European Callovian Bilobissa arose from the small, ribbed, gryphaeate oyster Catinula by evolution over an interval of about 6 Myr (the duration of the Bathonian stage according to Harland et al. 1982). Change in single characters followed differing patterns - varying between unidirectional and at a steady rate, and oscillatory - but was never punctuational. Evolution between Catinula and Callovian Bilohissa can therefore best be described as gradualistic (cf. Sheldon 1987). (3) Catinula apparently evolved from Bilobissa at the Toarcian/Aalenian boundary in the Poitou region of France. Evolution occurred rapidly in a peripheral isolate population (allopatric speciation). (4) Catinula is best regarded as a subgenus of Gryphaea. (5) The evidence for morphological stasis in mid- Jurassic Gryphaea (Bilohissa) is suspect and, notwithstanding point 3 above, the occurrence of punctuated equilibrium must be doubted. (6) An early G. (Bilohissa) lineage became extinct in the late Bajocian or early Bathonian, its extinction perhaps being due to the development of cooler or less fully marine conditions in which an adaptive strategy involving rapid growth became inviable. (7) Evolution of G. (Catinula) possibly introduced an adaptive strategy for maintaining shell stability under environmental conditions (? lowered temperatures or salinities) which precluded the acquisition of stability whilst reclining by the development of a large shell. (8) The small size of Gryphaea (i.e. G. (Catinula)), and of other bivalves, during the Bathonian in Europe should serve as an impetus for further geochemical studies to investigate the possibility of large-scale environmental (? climatic) changes during the stage. Investigations should also be made of the palaeoenvironments of variably-ribbed populations of Bathonian Praeexogyra hehridica (see Hudson and Palmer 1976). (9) Both factual and philosophical considerations strongly favour the view that the transitions between G. (Bilohissa) and G. (Catinula) represent evolution and not merely ecophenotypic responses. However, this should not deter further investigation of the involvement of ecophenotypic variation (e.g. by analysis of growth lines to establish the relationship between size, shape and age). The close correspondence of patterns of variation to those which might be expected under circumstances of environmental control raises the possibility that evolution may have involved the ‘genetic assimilation’ of ecophenotypic variation (Waddington 1957; see also Matsuda 1982). Acknowledgements. This work was financed by NERC grant GR3/5081, which is gratefully acknowledged. Our particular thanks go to Drs J. D. Hudson and David J. Siveter (University of Leicester) for obtaining support for the research (principally carried out at Leicester), and for continued encouragement and assistance, and to Mrs Joan Sylvester-Bradley for granting access to the collections at Noon’s Close and for providing the most welcoming of environments there. Others who have supplied notable assistance at various stages of the research are: J. Brannan, U. Bayer, J. H. Callomon, R. C. Clements, J. Gabilly, A. Hallam, C. P. Palmer and Derek J. Siveter. The latter deserves special mention for compiling a large dossier of oyster photographs, some of which have been used herein. J. Brannan kindly gave permission to quote from his unpublished thesis. Technical assistance was provided by R. Branson and C. Cunningham (University of Leicester), and T. Easter (Goldsmiths’ College, University of London). We thank an anonymous reviewer for a number of helpful comments and corrections, but accept full responsibility for the views finally expressed. JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 483 REFERENCES ARKELL, w. J. 1929-37. A monograph of British Corallian lamellibranchs. Palaeontogmphical Society, {Monographs), (1), 1929, 1-72; (2), 1930, 73-104; (3), 1931, 105-132; (4), 1932, 133-180; (5), 1933, 181-228; (6), 1934, 229-276; (7), 1934, 277-324; (8), 1935, 325-350, i-xvi; (9), 1936, 351-376, xvii-xxii; (10), 1937, 377-392, xxiii-xxxviii. 1934. The oysters of the Fuller’s Earth; and on the evolution and nomenclature of the Upper Jurassic Catinulas and Gryphaeas. Proceedings of the Cotteswold Naturalists' Field Club, 25, 21-68. 1947. The geology of the country around Weymouth, Swanage, Corfe and Lulworth. Memoirs of the Geological Survey of the United Kingdom, 386 pp. BAYER, u., JOHNSON, A. L. A. and BRANNAN, J. 1985. Ecological patterns in Middle Jurassic Gryphaea: the relationship between form and environment. In bayer, u. and seilacher, a. (eds.). Sedimentary and evolutionary cycles (Lecture notes in earth sciences, 1), 436-463. Springer, Berlin and Heidelberg, 465 pp. BRANNAN, J. 1983. Taxonomy and evolution of Triassic and Jurassic non-lophate oysters. D. Phil, thesis (unpubl.), 2 vols.. University of Oxford. CALLOMON, J. H. 1968. The Kellaways Beds and Oxford Clay. In sylvester-bradley, p. c. and ford, t. d. (eds.). The geology of the east Midlands, 264-290. University Press, Leicester, 400 pp. CARTER, J. G. 1980. Environmental and biological controls of bivalve shell mineralogy and microstructure. In RHOADS, D. c. and LUTZ, R. A. (eds.). Skeletal growth of aquatic organisms, 69-113. Plenum, New York and London, 750 pp. CELCOVA, N. A. 1969. Znacenie mikrostruktury rakoviny melovych ustric dlja ich sistematiki. Nauka, Moscow, 86 pp. [In Russian.] CHARLES, R.-p. and MAUBEUGE, P.-L. 1953. Les huitres plissees jurassiques de Test du Bassin Parisien (II). Bulletin du Museum d Histoire Naturelle de Marseille, 12 (for 1952), 113-123. CHEETHAM, A. H. 1987. Tempo of evolution in a Neogene bryozoan: are trends in single morphologic characters misleading? Paleobiology, 13, 286-296. COPE, J. C. W., DUFF, K. L., PARSONS, C. F., TORRENS, H. S., WIMBLEDON, W. A. and WRIGHT, J. K. 1980. A correlation of Jurassic rocks in the British Isles. Part 2. Middle and Upper Jurassic. Geological Society of London. Special reports, 15, 109 pp. , GETTY, T. A., HOWARTH, M. K., MORTON, N. and TORRENS, H. s. 1980. A Correlation of Jurassic rocks in the British Isles. Part 1. Introduction and Lower Jurassic. Geological Society of London. Special reports, 14, 73 pp. cox, L. R. 1946. Undescribed lamellibranchs from the English Inferior Oolite. Proceedings of the Malacological Society, London, 27, 22-32. 1952. The Jurassic lamellibranch fauna of Cutch (Kachh). No. 3. Families Pectinidae, Amusiidae, Plicatulidae, Limidae, Ostreidae and Trigoniidae (supplement). Memoirs of the Geological Survey of India. Palaeontologia Indica, series 9, 3 (4), 1-128. DUFF, K. L. 1978. Bivalvia from the English Lower Oxford Clay. Palaeontographical Society (Monographs), 137 pp. ELDREDGE, N. and GOULD, s. J. 1972. Punctuated equilibria: an alternative to phyletic gradualism. In schopf, T. J. M. (ed.). Models in paleobiology, 82-115. Freeman Cooper, San Francisco, 250 pp. FISCHER, j.-c. 1964. Contribution a I’etude de la faune bathonienne dans la vallee de la Creuse (Indre): brachiopodes et mollusques. Annales de Paleontologie (Invertebres), 50, 19-101. GABiLLY, J. 1973. Le Toarcien du Poitou. Doctoral thesis. University of Poitiers, 1 vol. and Annexe 1 (Lithologie et biostratigraphie des ammonites). GOULD, s. J. 1972. Allometric fallacies and the evolution of Gryphaea : a new interpretation based on White’s criterion of geometric similarity. Evolutionary biology, 6, 91-118. 1977. Ontogeny and phylogeny. Harvard University Press, Cambridge, Massachusetts, 501 pp. 1980. (ed.). The evolution of Gryphaea. Arno, New York. (Collected papers, no sequential pagina- tion.) HALLAM, A. 1968. Morphology, palaeoecology and evolution of the genus Gryphaea in the British Lias. Philosophical Transactions of the Royal Society of London, Series B, 254, 91-128. 1978. How rare is phyletic gradualism and what is its evolutionary significance? Evidence from Jurassic bivalves. Paleobiology, 4, 16-25. 1982. Patterns of speciation in Jurassic Gryphaea. Paleobiology, 8, 354-366. and GOULD, s. j. 1975. The evolution of British and American Middle and Upper Jurassic Gryphaea : a biometric study. Proceedings of the Royal Society of London, Series B, 189, 511-542. 484 PALAEONTOLOGY, VOLUME 33 HARLAND, W. B., COX, A. V., LLEWELLYN, P. G., PICKTON, C. A. G., SMITH, A. G. and WALTERS, R. 1982. A geological time scale. Cambridge University Press, Cambridge, 131 pp. HUDSON, j. D. 1980. Aspects of brackish-water facies and faunas from the Jurassic of north-west Scotland. Proceedings of the Geologists' Association, 91, 99-105. and PALMER, T. j. 1976. A euryhaline oyster from the Middle Jurassic, and the origin of the true oysters. Palaeontology, 19, 79-93. JOHNSON, A. L. A. 1981 . Detection of ecophenotypic variation in fossils and its application to a Jurassic scallop. Lethaia, 14, 277-285. 1984. The palaeobiology of the bivalve families Pectinidae and Propeamussiidae in the Jurassic of Europe. Zitteliana, 11, 1-235. MATSUDA, R. 1982. The evolutionary process in talitrid amphipods and salamanders in changing environments, with a discussion of ‘genetic assimilation’ and some other evolutionary concepts. Canadian Journal of Zoology, 60, 733-749. MEDCOF, J. c. 1949. Effects of sunlight exposure in rearing young oysters. Journal of the Fisheries Research Board of Canada, 6, 209-216. PALMER, T. J. 1979. The Hampen Marly and White Limestone formations: Elorida-type carbonate lagoons in the Jurassic of Central England. Pcdaeontology, 22, 189-228. PUGACZEWSKA, H. 1971. Jurassic Ostreidae of Poland. Acta Palaeontologica Polonica, 16, 195-311. RIDE, w. D. L., SABROSKY, c. w., BERNARDi, G. and MELVILLE, R. V. 1985. International code of zoo- logical nomenclature. Third edition. International Trust for Zoological Nomenclature, London, xx-f 338 pp. ROLLIER, L. 1911. Les facies du Dogger on Oolithique dans le Jura et les regions voisines. Fondation Schnyder von Wartensee (Memoire), Zurich, 352 pp. RYE, D. M. and SOMMER, M. A., II. 1980. Reconstructing paleotemperature and salinity regimes with oxygen isotopes. In rhoads, d. c. and lutz, r. a. (eds.). Skeletal growth of aquatic organisms, 169-202. Plenum, New York and London, 750 pp. SHELDON, p. R. 1987. Parallel gradualistic evolution of Ordovician trilobites. Nature, London, 330, 561- 563. SIEWERT, w. 1972. Schalenbau und Stammesgeschichte von Austern. Stuttgarter Beitrdge zur Naturkunde, Series B, 1, 1-57. SOWERBY, J. 1815-18. The mineral conchology of Great Britain, 2, (1), 1815, 1-28; (2), 1816, 29-116; (3), 1817, 117-194; (4), 1818, 195-239. Meredith, London. SOWERBY, J. DE c. 1826^6. The mineral conchology of Great Britain, 6, (1), 1826, 1-86; (2), 1827, 87-156; (3), 1828, 157-200; (4), 1829, 201-235; Preface to the general indexes and systematic index to the six volumes, 1835, 239-250; 7, Alphabetic index to volumes 1-6, 1840, 1-1 1 ; (1), 1840?, 1-8; (2), 1841, 9-16; (3), 1843, 17-24; (4), 1844, 25-56; (5), 1846, 57-80. Meredith, London. STANLEY, s. M. 1970. Relation of shell form to life habits in the Bivalvia (Mollusca). Memoirs. Geological Society of America, 125, 1-296. STENZEL, H. B. 1971. Oysters. N953-N1224. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part N. Bivalvia 3. Geological Society of America and University of Kansas Press, Boulder and Lawrence. SYLVESTER-BRADLEY, p. c. 1958. The description of fossil populations. Journal of Paleontology, 32, 214- 235. 1959. Iterative evolution of fossil oysters. Proceedings of the XVth International Congress of Zoology, London, section 2, 193-197. 1977. Biostratigraphical tests of evolutionary theory. In kauffman, e. g. and hazel, j. e. (eds.). Concepts and methods of biostratigraphy, 41-63. Dowden, Hutchinson and Ross, Stroudsburg, 658 pp. tann, f. c. and Hudson, j. d. 1974. Isotopic studies of the palaeoecology and diagenesis of the Great Estuarine Series (Jurassic) of Scotland. Scottish Journal of Geology, 10, 91-128. TEVESJ, M. J. s. and carter, j. g. 1980. Environmental relationships of shell form and structure of unionacean bivalves. In rhoads, d. c. and lutz, r. a. (eds.). Skeletal growth of aquatic organisms, 295-322. Plenum, New York and London, 750 pp. VERMEiJ, G. J. 1980. Gastropod shell growth rate, allometry, and adult size; environmental implications. 379-394. In rhoads, d. c. and lutz, r. a. (eds.). Skeletal growth of aquatic organisms. Plenum, New York and London, 750 pp. voLTZ, p. L. 1828. Uebersicht der Petrefakten der beiden Rhein-Departemente. In aufschlager, j. f. Das Elsass Oder die Departemente des Ober- und Niederrheins (Supp.), Heitz, Strasbourg, 64 pp. JOHNSON AND LENNON: JURASSIC OYSTER EVOLUTION 485 WADDINGTON, c. H. 1957. The strategy of the genes. George Allen and Unwin, London, 262 pp. ANDREW L. A. JOHNSON Department of Earth Sciences Goldsmiths’ College, University of London Rachel McMillan Building Creek Road London SE8 3BU, UK Present address: Raymont Hall Wickham Road Brockley London SE4 ILX, UK CHRISTOPHER D. LENNON 36 Wodehouse Street Typescript received 5 October 1988 Norwich NR3 4TZ Revised typescript received 24 July 1989 Norfolk, UK ■i«^ ■::4 J ) THE TRILOBITE P RO TO LLO Y DO LITH U S FROM THE MIDDLE ORDOVICIAN OF NORTH PORTUGAL by M. ROMANO Abstract. Protolloydolithus sp. nov., from beds of Llandeilo age in north Portugal, is described and figured. The genus was previously unknown from Iberia, and the present record indicates faunal links with southern Britain during middle Ordovician times. The appearance of the genus is linked to a transgressive event following an early Llandeilo lowering of the sea level. In 1908 Delgado published extensive faunal lists from the Ordovician sequences of Portugal and recognized over thirty trilobite species from his ‘Ordovician moyen’ of the Valongo area of north Portugal. Many of these species have since been revised (see Henry 1980; Hammann 1974, 1983; Hammann et al. 1982; Gutierrez-Marco et al. 1984, and references therein) and the dominantly mudrock sequence comprising Delgado’s ‘Ordovician moyen’ (Valongo Formation of Romano and Diggens 1973^) is now known to range in age from at least early Llanvirn to probably early Caradoc. No trinucleid trilobites were recorded by Delgado from north Portugal, but Onnia and/or Deanaspis occur in lower Caradoc strata in central Portugal and Spain (Hammann 1976; Rabano 1984; Young 1985). In Brittany, Manolithus occurs in the upper part of the Postolonnec Formation (upper Llandeilo age, Henry 1980) but has not been recorded from coeval strata in central Portugal, despite the remarkable similarity of the two sequences (Henry et al. 1973^; Henry and Romano 1978; Paris 1981 ; Romano and Henry 1982; Young 1988). Young (1985) regards their absence in central Portugal as being due to a hiatus on a topographic high. REMARKS ON THE DISTRIBUTION OE PROTOLLOYDOLITHUS A single specimen assigned to Protolloydolithus sp. nov. was collected from the upper part of the Valongo Formation near Covelo (text-fig. 1). The beds are within the top of the Placoparia (Placoparia) tournemini Biozone (rarely, specimens of the younger P. (Coplacoparia) borni also occur) of early Llandeilo age (Romano 1976). The specimen is unusual in that at present it is the oldest and only the second marrolithid known from rocks of middle Ordovician age in the Iberian/ Armorican part of Gondwana, although rare specimens of ‘Hanchungolithinae gen. et sp. inc.’ occur in the P. (C.) borni Biozone of Spain (I. Rabano, pers. comm.). It is also the only record of Protolloydolithus outside Britain. The genus first appears in the lower Llanvirn of the Shelve inlier and South Wales with P. ramsayi (Hicks) and then P. neintianus (Whittard) (Thomas et al. 1984; Kennedy 1988). P. salax (Rushton and Hughes 1981) is known from the Llanvirn of the Great Paxton Borehole, Cambridgeshire, to be replaced in the lower Llandeilo of the Builth area by P. reticulatus (Elies) (Hughes 1971). AFFINITIES OF THE LOWER AND MIDDEE ORDOVICIAN TRILOBITE FAUNAS OF NORTH PORTUGAL The trilobite faunas from the Arenig of southern Britain (Whittard 1966; Thomas et al. 1984; Fortey and Owens 1987) and the lower Llanvirn of North Portugal (Delgado 1908; Romano 1976, 1982(3, 6; Rebelo and Romano 1986) have much in common, namely Neseuretus, Placoparia, Ectillaenus, Selenopeltis, 1 Asaphellus, Colpocoryphe and probably Ogyginus at generic level; I Palaeontology, Vol. 33, Part 2, 1990, pp. 487-493.] © The Palaeontological Association 488 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 1. Locality maps and simplified geological map (after Delgado 1908) of Valongo area, north Portugal. Star indicates locality where Protolloydolithus sp. nov. was collected. Placoparia (P.) cambriensis and 1 Selenopeltis macrophthalma at specific level. The ranges of these genera in the Shelve, Carmarthen/Whitland and north Portugal are shown in text-fig. 2. The Arenig trilobite faunas of Shelve and Carmarthen have many taxa in common but also show important diflferences. Fortey and Owens (1987) pointed out that the varied trinucleid assemblages from the upper part of the Mytton Formation are absent in South Wales, and the Llanfallteg Formation fauna is not represented in the Shelve area. These authors explained the absence of the Llanfallteg fauna by suggesting that since the fauna and sediment of this unit represent a regressive phase, and as the Shelve area occupied an ‘on-shelf’ position, the Llanfallteg faunas did not reach the Shelve area. The importance of transgressive/regressive events in controlling the distribution of shallow-water faunas is also well demonstrated in the composition of the north Portuguese faunas. Although approximately twenty-five and seventy species are present in the Shelve Arenig faunas and deeper- water equivalents in south Wales respectively (Whittard 1966; Fortey and Owens 1987), in north Portugal the shallower-water coeval Armorican Quartzite facies has yielded only Neseuretus (similarly for the Stiperstones Formation in the Shelve), and in Brittany some three species are ROM ANO; ORDOVICIAN TRILOBITE FROM PORTUGAL 489 TEXT-FIG. 2. Simplified Ordovician sequences of Shelve Inlier, south Wales and north Portugal with ranges of selected trilobites (after Whittard 1966, 1979; Fortey and Owens 1987; Kennedy 1988; and author). Abbreviations: MORI. = Moridunian; WHIT. = Whitlandian ; FENN. = Fennian. Thick vertical bars indicate similar species are present in all three areas. recorded (Henry 1971, 1980). Fortey and Morris (1982), Fortey and Owens (1987) and Cocks and Fortey (1988) have previously pointed out the significance of Neseuretus in identifying inshore facies, although species of the genus also occur in deeper- water platform sites (Hammann 1983). Fortey and Owens noted that the diverse Arenig trilobite faunas of south Wales only appeared in the more shelfward sites of Gondwana during the Llanvirn or later. In north Portugal this is well shown by the appearance of a relatively diverse trilobite fauna of early Llanvirn age in mudrocks above the thinly bedded, heterolithic Armorican Quartzite sequences (text-fig. 2). Deeper-water atheloptic assemblages (characterized by ‘blind or nearly blind trilobites’: Fortey and Owens 1987, p. 106), such as those from the Fennian of south Wales are not known in Portugal; these assemblages were considered to imply a water depth of 300 m or more (Fortey and Owens 1987, p. 106). Estimates based on sedimentological studies in central Portugal suggested that water depth there was probably no more than 80 m during the Llandeilo (Brenchley et al. 1986). Assuming a maximum average slope to the north of 01° for the Iberian Platform in Llandeilo times (Brenchley et al. 1986, p. 252), this would suggest that the water depth in north Portugal was in the order of 100-150 m, well above the inferred depth to support atheloptic assemblages. The appearance of Protolloydolitims in the Llandeilo of north Portugal, along with a relatively sudden increase in diversity of the trilobite species, is related to another regressive/transgressive 490 PALAEONTOLOGY, VOLUME 33 couplet. The regressive event is documented in central Portugal (and Crozon Peninsula of north- west France) by the sandstones of the Monte da Sombadeira Formation (and Kerarvail Formation of north-west France), followed by a transgressive phase and the return to mudrock deposition with an increase in diversity and individual abundance of the benthos. This sandstone unit is not present in the deeper- water sequences of north Portugal, but work in progress at Valongo with Dr T. Young has identified slight lithological changes which, together with the faunal changeover, reflect the events recognized in central Portugal. However, the diverse Llandeilo faunas of north Portugal do not show such close affinities with the southern British faunas as was apparent in the early Ordovician. Only four trilobite genera are common to both regions in the Llandeilo; ProtoUoydolithus, Selenopeltis, Ogyginus and Nohiliasaphus. Of these, Protolloydolithus probably spread from the Builth area of Wales, while the other three were already endemic to Iberia (Delgado 1908; Romano 1982fi; Romano et al. 1986). The reduced faunal affinities of southern Britain and north Portugal in the Llandeilo are considered to be the result of an increased separation of the two regions by middle Ordovician times and/or the relatively smaller magnitude of the transgression compared with that during the Llanvirn. It is of interest to note that on a global scale Fortey and Cocks (1988) record an early Llandeilo regression followed by a major early Caradoc transgression. In Iberia there is evidence of two Llandeilo shallowing events: that represented by the Monte da Sombadeira Formation and a younger one which reached maximum shoaling during Cabril Formation times (Young 1988, p. 390). Both these formations (or equivalent units) are also recognized in Brittany and crop out in an area of over 75000 km^. It therefore appears unlikely that the regressive sandstones of these two formations (and the intervening transgressive mudrocks with their relatively enriched benthos) represent a purely local tectonic effect. A more likely scenario is that the Llandeilo regression was composite, prior to the major early Caradoc transgression. Further detailed investigations into other middle Ordovician platform sequences may provide supportive evidence. SYSTEMATIC PALAEONTOLOGY Family trinucleidae Hawle and Corda, 1847 Subfamily marrolithinae Hughes, 1971 Genus protolloydolithus Williams, 1948 Type species. Trinucleus Ramsay i Hicks, 1875. Original designation by Williams 1948, p. 66. Lower Llanvirn, Ramsey Island. Protolloydolithus sp. nov. Text-fig. 3 Material. Single specimen (SG 6717); nearly complete but distorted fringe showing internal mould of part of upper lamella and external mould of part of lower lamella. Specimen housed in the museum of the Portuguese Geologieal Survey, Lisbon. Horizon and locality. Upper part of Valongo Formation, c. 70 m below top of unit; top of Placoparia (P.) tournemini Biozone, lower Llandeilo. Locality 3/41 (see Romano 1976), a small quarry to the east of the road approximately 1 km ESE of Covelo (text-fig. 1). Description. Cephalon approximately 14 mm wide across base of genal spines, margin apparently evenly rounded. Fringe more or less flat, although with slightly upturned outer arc of pits, of broadly constant width, but possibly narrows anteriorly in front of glabella. Preserved length of genal spine is 2-5 mm (but possibly at least 5 mm long), extending posteriorly and slightly outwards; lateral margins of fringe and genal spine are not in a straight line. Dorsal surface of spine is flat with very faint median furrow (possibly due to flattening). Pits (El) of outer arc larger than other pits; at least forty-five are visible but it is not possible to give an accurate fringe formula. Pits are largest anterolaterally and on the whole are fairly regularly disposed. Internal to El there is a well-marked girder, and a girder list which dies out before reaching genal spine. II arc ROMANO: ORDOVICIAN TRILOBITE FROM PORTUGAL 491 TEXT-FIG. 3. Piotolloydolithus sp. nov. (SG 6717). Valongo formation, Placoparia (Pkuoparia) tournemini Biozone, Llandeilo. From small quarry approximately 1 km ESE of Covelo and 1 1 km SSE of Valongo. x 6 5. reasonably distinct, pits smaller and more numerous than in El (where preserved, six or seven El pits lie adjacent to ten II pits). Preservation is generally rather poor for the rest of the fringe, and apart from the distinct innermost arcs appears to consist of randomly arranged pits. However, nine or ten A ‘arcs’ may be present across the fringe between the cheeks and El . In or E (flange) pits (Hughes et al. 1975) cannot be clearly distinguished although on the right-hand side there are possibly a few F pits. Laterally at least two regular inner arcs are present. The pits of these arcs show regular concentric and radial arrangement for at least fourteen radial rows, while two more external arcs are seen (where preservation allows) to continue this regular arrangement for at least six radial rows. Suggestion of very faint facial suture cutting across base of genal spine, running subparallel to distal part of posterior border margin. Glabella and cheeks not preserved, although part of inner flange of fringe is present on right-hand side. Rest of exoskeleton unknown. Discussion. The recognition of a single El arc, prominent girder and strong girder list, together with the generally irregular arrangement of the pits, indicates that the species may be assigned to ProtolloydoUthus. According to the diagnosis of the genus given by Whittard (1955, p. 40 and 1956, p. 41) and Hughes et al. (1975, p. 577), the present species shows some ?minor differences. II pits are not noticeably larger than the other inner pits; I pits are not all ‘irregularly positioned’; F pits do not appear to be well developed. However, in P. ramsayi the 11 pits are not always significantly larger than the other pits (Whittard 1956, pi. V, fig. 1 1) and there is a tendency in P. ramsayi and P. neintianus for some regularity of inner I arcs anteriorly, and laterally in P. sala.x. The specimen is considered to represent a new species since it does not closely resemble any described species of the genus. It is left unnamed at present as only a single specimen is known. The Llanvirn species P. ramsayi, ProtolloydoUthus sp. (Kennedy 1988) and P. neintianus, as well as the Llandeilo form P. reticulatus (Elies) (Hughes 1971) differ mainly from the Portuguese species in not possessing such well-ordered internal I arcs; in addition, the latter species has a much narrower fringe. The well-ordered pits indicate an advanced form (Dr J. K. Ingham, pers. comm.) and are 492 PALAEONTOLOGY, VOLUME 33 considerably better ordered than the arrangement seen in the broadly contemporaneous Builth species. Acknowledgements. I should like to thank Dr J.-L. Henry for encouraging me to 'think again’ about the specimen; Drs R. A. Fortey and J. K. Ingham for helpful discussions; Dr T. P. Young for reading the manuscript and photographing the specimen. Dr I. Rabano sent me information on Spanish middle Ordovician trinucleids. Mr M. Cooper redrew the diagrams, and Miss P. Mellor typed the manuscript. REFERENCES BRENCHLEY, p. J., ROMANO, M. and GUTiERREZ-MARCO, J. c. 1986. Proximal and distal hummocky cross-stratihed facies on a wide Ordovician shelf in Iberia. 241-255. In knight, r. j. and mclean, j. r. (eds.). Shelf sands and sandstones. Canadian Society of Petroleum Geologists, Mem. II, pp. 1-347. COCKS, L. R. M. and fortey, r. a. 1988. Lower Palaeozoic facies and faunas around Gondwanaland. 183-200. In audley-charles, m. g. and hallam, a. (eds.). Gondwana and Tethys. Geological Society Special Publication No. 37, 317 pp. DELGADO, J. F. N. 1908. Systeme silurique du Portugal; etude de stratigraphic paleontologique. Commission du Service geologiqiie du Portugaf 1-245. FORTEY, R. A. and COCKS, L. R. M. 1988. Arenig to Llandovery faunal distributions in the Caledonides. 233-247. In HARRIS, A. L. and fettes, d. j. (eds.). The Caledonian- Appalachian Orogen. Geological Society Special Publication No. 38, x + 643 pp. and MORRIS, s. F. 1982. The Ordovician trilobite Neseuretus from Saudi Arabia, and the palaeogeography of the Neseuretus fauna related to Gondwanaland in the earlier Ordovician. Bulletin of the British Museum (Natural History) (Geology), 36, 63-75. and OWENS, R. m. 1987. The Arenig Series in South Wales. I. Bulletin of the British Museum (Natural History) (Geology), 41, 69-307. GUTIERREZ-MARCO, J. c., RABANO, I., PRIETO, M. and MARTIN, J. 1984. Estudio biocstatrigrafico del Llanvirn y Llandeilo (Dobrotiviense) en la parte meridional de la Zona Centroiberica (Espana). Cuadernos geologia Iberica, 9, 289-219. HAMMANN, w. 1974. Phacopina und Cheirurina (Trilobita) aus dem Ordovizium spaniens. Senckenbergiana lethaea, 55, 1-51. 1976. Trilobiten aus dem oberen Caradoc der ostlichen Sierra Morena (spanien). Senckenbergiana lethaea, 57, 35-85. 1983. Calymenacea (Trilobita) aus dem Ordovizium von Spanien; ihre Biostratigraphie, Okologie und Systematik. Abhandlungen hrsg. von der senckenbergischen naturforschenden Gessellschaft, 542, 1-177. ROBARDET, M. and ROMANO, M. 1982. The Ordovician System in southwestern Europe (France, Spain and Portugal). Correlation chart and explanatory notes. International Union of Geological Sciences, no. 1 1, 1^7. HAWLE, I. and CORDA, A. J. c. 1847. Prodrom einer Monographie der bohmischen Trilobiten. J. G. Calve, London, 176 pp. HENRY, j.-L. 1971. Les trilobites Asaphidae et Eohomalonotidae du Gres Armoricain superieur (Arenigien) de I’ouest de la France. Memoires du Bureau de Recherches Geologiques et Minieres, 73, 65-11 . 1980. Trilobites ordoviciens du Massif Armoricain. Memoires de la Societe geologique et mineralogique de Bretagne, 22, 1-250. NiON, J., PARIS, F. and thadeu, d. 1973^. Chitinozoaires, Ostracodes et Trilobites de I’Ordovicien du Portugal (serra de Buyaco); essai de comparaison et signification paleogeographique. Commission du Service geologique du Portugal, LVII, 303-345. and ROMANO, m. 1978. Le genre Dionide Barrande, 1847 (Trilobite) dans I’Ordovicien du Massif Armoricain et du Portugal. Geobios, No. 1 1, 327-343. HUGHES, c. p. 1971. The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, Central Wales. Part II. Bulletin of the British Museum (Natural History) (Geology), 20, 117-182. INGHAM, J. K. and ADDISON, R. 1975. The morphology, classification and evolution of the Trinucleidae (Trilobita). Philosophical Transactions of the Royal Society of London, Series B, 242, 537-607. KENNEDY, R. J. 1988. Ordovician (Llanvirn) trilobites from SW Wales. Palaeontographical Society [Monograph], (Publ. No. 576, part of Vol. 141 for 1987), 1-55. PARIS, F. 1981. Les chitinozoaires dans le Paleozoique de sud-ouest de I’Europe. Memoires de la Societe geologiques et mineralogique de Bretagne, 26, 1-412. ROMANO: ORDOVICIAN TRILOBITE FROM PORTUGAL 493 RABANO, I. 1984. Trilobites Ordovicicos de Macizo Hesperico Espanol: una vision bioestratigrafica. Cuadernos geologia Iberica, 9, 267-287. REBELO, J. A. and ROMANO, M. 1986. A contribution to the lithostratigraphy and palaeontology of the Lower Palaeozoic rocks of the Moncorvo region, northeast Portugal. Comwiicafoes dos Servifos Geologicos de Portugal, 72, 45-57. ROMANO, M. 1976. The trilobite genus Placoparia from the Ordovician of the Valongo area, north Portugal. Geological Magazine, 113, 11-28. 1982u. The Ordovician biostratigraphy of Portugal - a review with new data and re-appraisal. Geological Journal, 17, 89-1 10. 1982Z). A revision of the Portuguese Ordovician Odontopleuridae (Trilobita): Selenopeltis and Primaspis. Comunicagdes dos Servigos Geologicos de Portugal, 57, 213-223. BRENCHLEY, p. J. and MCDOUGALL, N. D. 1986. Ncw information concerning the age of the beds immediately overlying the Armorican quartzite in central Portugal. Geobios, No. 19, 421-A33. and DiGGENS, J. N. 1973-4. The stratigraphy and structure of Ordovician and associated rocks around Valongo, north Portugal. Comunicagoes dos Servigos Geologicos de Portugal, 57, 23-50. and HENRY, J.-L. 1982. The trilobite genus Eoharpes from the Ordovician of Brittany and Portugal. Palaeontology, 25, 623-633. RUSHTON, A. w. A. and HUGHES, c. p. 1981. The Ordovician trilobite fauna of the Great Paxton Borehole, Cambridgeshire. Geological Magazine, 118, 623-646. THOMAS, A. T., OWENS, R. M. and RUSHTON, A. w. A. 1984. Trilobites in British Stratigraphy. Special Report of the Geological Society of London, 16, 1-78. WHiTTARD, w. F. 1955. The Ordovician trilobites of the Shelve Inlier, west Shropshire. Palaeontographical Society [Monograph], part I, 1^0. 1956. The Ordovician trilobites of the Shelve Inlier, west Shropshire. Palaeontographical Society [Monograph], part II, 41-70. 1966. The Ordovician trilobites of the Shelve Inlier, west Shropshire. Palaeontographical Society [Monograph], part VIII, 265-306. 1979 (compiled by w. T. dean). An account of the Ordovician rocks of the Shelve inlier in west Salop and part of north Powys. Bulletin of the British Museum {Natural History) (Geology), 33, 1-69. WILLIAMS, A. 1948. The Lower Ordovician cryptolithids of the Llandeilo district. Geological Magazine, 85, 65-88. YOUNG, T. p. 1985. The stratigraphy of the upper Ordovician of central Portugal. Unpublished Ph.D. Thesis, University of Sheffield. 1988. The lithostratigraphy of the upper Ordovician of central Portugal. Journal of the Geological Society of London, \45, Ml -392. MICHAEL ROMANO Department of Geology University of Sheffield Beaumont Building Typescript received 23 April 1989 Brookhill Revised typescript received 14 June 1989 Sheffield S3 7HF, UK 1 Fi :4- ■Ji* i i i i i /)» ■'■ : - ' CADDISFLY PUPAE FROM THE MIOCENE INDUSIAL LIMESTONE OF SAINT-GERAND-LE-PUY, FRANCE MARGUERITE HUGUENEY, HENRI T ACHET flWi/ ER AN^OIS ESCUILLIE Abstract. Trichoptera pupae are described for the first time in the Indusial Limestone Formation of Saint- Gerand-le-Puy, Allier, France. The caddisfly pupal cases, named Imiusia tubulosa, constitute this limestone. The relationships of these pupae to extant families are analysed and they are placed in the Limnephilinae. Fossil preservation and palaeoenvironment are discussed. Bosc(1805) was the first author to mention the Indusial Limestone of Saint-Gerand-le-Puy and to describe the tubes of which it is composed as consisting ‘de tres-petits helices fossiles agglutines, tantot en masse irreguliere, tantot en forme de cylindres ouverts par un bout et fermes par I’autre, ou mieux, de cones creux d’environ un pouce et demi de long sur cinq lignes de diametre total, et un peu plus d’une ligne d’epaisseur’. He gave them the name of Indusia tubulosa, but was not able to establish if they were made by polychaete annelids or by caddisflies (Insecta: Trichoptera). Brongniart (1810) demonstrated the freshwater origin of the limestone and assigned a caddisfly larval origin to the tubes. As the knowledge on extant Trichoptera progressed, Oustalet (1870) classified these cases in the genus Phrygauea ; he even recognized two new species ; P. gerandiana for the cases from Saint-Gerand and P. corentiana for those found in Gergovie (Puy-de-D6me). Up to now, no body part of Trichoptera had ever been described from Saint-Gerand, but recently two fossil pupae were found which prove that the tubes are really caddisfly cases. This study presents these new finds and provides evidence for systematic assignment and taphonomy, taking into account that much progress has been made in the taxonomical and biological knowledge of the Trichoptera since the first record. GEOLOGICAL SETTING The fossils were found at Le Vendant limestone quarry, Bouce commune, Allier Department, France, which is situated in the Limagne d’Allier basin in the Massif Central. The Limagne d’Allier cuts a north-south furrow in the Hercynian basement of the Massif Central. The Allier River runs through this basin, which is 180 km long and roughly 35 km wide. Geophysical studies and boreholes demonstrate that, from the Eocene to the beginning of the Miocene, a series of small sedimentary basins, isolated from each other by tectonic or volcanic thresholds, opened from north to south in a plateau area and contained lakes of various sizes. Rapid subsidence led to the deposition of more or less thick sequences of clastic and carbonate sediments (Donsimoni and Giot 1977). Progressive migration from south to north during the Oligocene restricted the upper sequence (Upper Oligocene-Aquitanian) to the Vichy area; here, the fluvio- lacustrine sediments are eharacterized by the widespread development of stromatolitic limestones, which reached a maximum in the Saint-Gerand-le-Puy area. Algal growths in various shapes and sizes, from small ooliths and pisoliths to large spheroidal and columnar concretions, coaleseed to form a limestone reef more than 20 m thiek. This limestone often encloses an aggregate of tubes 3 or 4 cm long and 1 cm wide, generally covered by small mollusc shells and bound together by a IPalaeontolog)', Vol. 33, Part 2, 1990, pp. 495-502, 1 pl.| © The Palaeontological Association 496 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 1. Location map: Le Vendant quarry is indicated by an arrow. calcareous crust: since the tubes resemble caddis larvae cases, the limestone is known as the ‘Calcaire a Phryganes’ or the ‘Indusial Limestone’. The oldest bioherms are contemporaneous with the fossil mammal site of Cournon (Puy-de- Dome), which is dated at about 30 Ma; the youngest correlate with the mammalian fauna of Marcoin (Puy-de-D6me), which is older than the marine Burdigalian (>21-9Ma). Favourable conditions for caddises lasted for 8 million years. Later on, the lake progressively filled and the swamps dried up. Le Vendant quarry belongs to the quarry group known since 1833 as Saint-Gerand-le-Puy, when Geoffroy-Saint-Hilaire recorded fossil vertebrate-bearing deposits there. These quarries have been exploited for more than 150 years, and the large and small mammals recovered are so numerous and important that these localities have been selected as a reference level for the fossil mammal biozonation of Europe (Mein 1976). They have also yielded an exceptional variety and abundance of fossil bird remains, mostly aquatic forms (Cheneval 1984), and are internationally recognized fossil bird localities, having no equivalent within or outside Europe. Nevertheless, caddisflies have never before been described from there. SYSTEMATIC PALAEONTOLOGY Class INSECTA Linnaeus, 1758 Order trichoptera, Kirby, 1813 Family limnephilidae Kolenati, 1848 Subfamily limnephilinae Kolenati, 1859 Genus indusia Bose, 1805 Indusia tuhulosa Bose, 1805 Plate 1, figs. 1^; Text-fig. 2A 1805 Indusia tuhulosa Bose, pp. 397^00, pi. 7, figs A-E. 1810 Indusia tuhulata Bose; Brongniart, p. 357. 1829 Indusia tuhulosa Bose; de Serres, p. 208. 1833 Indusia tuhulata \ Geoffroy-Saint-Hilaire, p. 77. HUGUENEY ET AL. : FRENCH MIOCENE CADDISFEY PUPAE 497 71871 Phryganea Gerandicma Oustalet, p. 102. 71871 Phryganea Corentiana Oustalet, p. 101. 1969 Inditsia tubidosa de Serres, non tidmlata Brongniart; Fischer, p. 326. 1973 Boscindusia tahidata (sic) Bose; Vialov, p. 586. Emended diagnosis. See Vialov and Sukatsheva 1976. Remarks. Bose (1805) always referred to the fossils as ‘Indusie tubuleuse’; the name of the taxon is latinized only in the title of his plate, which is placed much farther on in the volume. This fact led to errors on the valid specific name and the author’s attribution of this binomial. The International Code of Zoological Nomenclature (3rd edition, 1985) makes a clear distinction between the work of an animal (art. 23 f-iii) and its traces (art. 23 g-iii). In our opinion a fossil caddis case corresponds to the work of an animal and is not a trace fossil : normally, trace fossils are features independent of their producers; this is not at all the case for a caddis case, for the larva cannot live without it. So we consider that we have to refer to art. 23 f-iii and that the name Indusia tuhulosa, created for the cases, is valid for the animals which made these cases. Repository. Centre des Sciences de la Terre, Universite Lyon I, France; No. FSL 97698-97701. Description. One of the two specimens is lying in its case and only its back is exposed. The other is free, but fragments of the case remain attached to it and only the ventral face can be observed clearly. Head: eyes, antennae, labial and maxillary palpi and haustellum are well preserved. In front of the maxillary palpi, the head shows a bifid structure interpreted as the mandibulae spread out and joined to the labrum. The dimensions and the relative position of the labial and maxillary palpi are characteristic for pupae of the families Phryganeidae or Limnephilidae. Thorax: coxae and femora of the prothoracic legs and coxae of the mesothoracic ones are easily recognizable. The entire left mesothoracic leg, with all its distinct segments, is lying on the wing pad as in Recent Trichoptera pupae. Dorsal face: similar in shape and ornamentation to those of Phryganeidae and Limnephilidae. Discussion. The pupae and the cases (indusia of the earlier authors) identify this form as a member either of the Phryganeidae or of the Limnephilidae. In these two families there exist genera whose larvae construct large cylindrical cases. In the extant Phryganeidae, the cases are constructed exclusively from plant material arranged in a spiral. The pupae of Indusia tubidosa are closer to those of the family Limnephilidae and even to the subfamily Limnephilinae. In the genus Linmepliilus, the larvae build their cases from more diverse materials : owing to the availability of construction materials and behavioural preferences, they choose vegetable or mineral matter, or mollusc shells, or a mixture of these components. The fossil forms of Saint-Gerand seem to have preferred mollusc shells, ostracod carapaces or ooliths (Donsimoni 1975; Guillot 1979), but we observed also, in some fossil cases, vegetable fragments mixed with mollusc shells exactly as in the extant Limnephilus. In the Limnephilinae, different species can build similar cases but, reciprocally, different cases can be built by the larvae of a single species. So it is impossible to determine if the Saint-Gerand Limnephilinae represent one or more species and, without adult specimens, to decide whether the genus Indusia is a synonym of the genus Limnephilus, the cases of which are very similar to those of Indusia. The author of the taxon Indusia tubidosa is Bose (1805) and not Brongniart (1810) or de Serres ( 1829) as indicated by Fischer (1969) and Vialov and Sukatsheva (1976). The genus Indusia includes several fossil species (often only cases) from the Cretaceous and the Tertiary of Mongolia, Siberia and North America (Fischer 1969, Vialov and Sukatsheva 1976). As far as true Limnephilidae are concerned, the genus name Phryganea given by Oustalet (1871) to his new species is not correct, but as different species of Limnephilidae can build similar cases, it is impossible to decide if I. gerandiana and /. corentiana are synonyms of I. tubidosa. 498 PALAEONTOLOGY, VOLUME 33 TEXT-FIG. 2. A, Indusia tubidosa Bose, 1805. FSL 97699. Ventral aspect of body, a, antenna; c, case; e, eye; h, haustellum; 1-md, mandibulae joined to the labrum; Ip, labial palp; ml, mesothoracic leg; mxp, maxillary palp; pi, prothoracic leg; wp, wing pad. The left maxillary and labial palpi and part of the left prothoracic leg are broken. B, Limnephilus sp. Ventral aspect of the anterior part of an extant pupa showing especially the masticatory apparatus. TAPHONOMY AND PALAEOEN VIRONMENT This find of fossil caddis pupae is quite remarkable, considering this stage lasts only about two weeks in the trichopteran life cycle, and how fragile the animals are at that precise moment when the most intensive histolysis of the larval tissues takes place. When natural mortality of the pupa occurs, the dead tissues decay rapidly (in a few days) and only an empty, floppy pupal cuticle remains in the case. Evidently, the caddisflies were encrusted very rapidly, just before emergence, at the precise moment when the tissues became firm; but the tissues themselves are not preserved and the two specimens are natural moulds of external surfaces of the pupae. On the other hand, since all the cases are the same size, Donsimoni (1975) believed, rightly in our opinion, that they are pupal case concentrations. It is unlikely that these concentrations are due to water currents because, just before the metamorphosis and even in stagnant conditions, the Trichopteran larvae fasten their cases on the substrate (here, rocky substrate or other cases) with silk threads. Under natural conditions, bacterial decay destroys the silk threads that bind the constitutive elements of the cases together in just a few months. In Saint-Gerand, most of the cases EXPLANATION OF PLATE 1 Eigs. 1^. Indusia tubulosa Bose, 1805. Indusial limestone of Saint-Gerand-le-Puy (Le Vendant quarry), Aquitanian. la, 6, FSL 97699, isolated pupa, x 3-6; la, stereo-pair of ventral side; 16, stereo-pair of lateral side. 2, FSL 97698, section of an indusial bioherm showing the caddisfly tubes, transversely or longitudinally cut, encrusted by repeated algal laminations, x 1. 3, ESL 97700, isolated and non-incrusted tube from another more sandy part of the same quarry ; this exceptional tube differs from the others by the helicoidal and regular arrangement of the Pseudamnicola gerannensis (Rey, 1974) shells, x 1-6. 4, ESL 97701, pupa preserved in its case, stereo-pair of dorsal side, x 2-5. Fig. 5. Limnephilus sp., extant larval case made from mollusc shells, x 2-2. PLATE 1 HUGUENEY et al., Indusia tuhulosa 500 PALAEONTOLOGY, VOLUME 33 are empty or filled with marly sediments. The rapidity of fossilization explains why they are preserved in such large quantities. Virtually all the older records of Trichoptera, from Upper Cretaceous to Tertiary (the last partly contemporary with the Indusial Limestone fauna), are based on small adult remains from worldwide amber (Ulmer 1912; Botosaneanu and Wichard 1983), while Indusia are large. In amber, the case-makers are generally represented by extant genera of the families Phryganeidae, Calamoceratidae, Molannidae and Sericostomatidae. But the family Limnephilidae has not yet been reported. In Laurentiaux’s opinion (1953) families recovered from amber could indicate a tropical or subtropical climate. Even if the lack of Limnephilidae in amber is due more to the adult way of life than to a real absence of the family at that time, we can suppose that the exceptional accumulation of the Limagne Indusial Limestone could be related to the Tertiary climatic changes (in particular progressive cooling). Indeed, whilst tropical or subtropical forms are represented in the rich vertebrate fauna of Saint-Gerand (crocodile, anhinga, secretary bird, parrot, pangolin, tapir), palynology suggests the temperate deciduous forest region {Ulmus, Alnus, etc.: Gorin 1975). At present the Limnephilinae are characteristic of holarctic temperate or cold habitats (Schmid 1955). Most of the species recovered from amber, certainly adapted to higher temperatures, must have disappeared from the holarctic region and have been partly replaced by the Limnephilidae, more flexible in behaviour and with a greater ecological plasticity. Only the extant subfamily Limnephilinae includes species able to survive in temporary pools (Wichard and Reichel 1970; Wiggins 1973) due to an imaginal diapause. Moreover, the larvae of some species of this subfamily are the only Trichoptera to occur in brackish water (Sutcliff 1960; Leader 1971; Malicky 1974; Colburn 1983). In Saint-Gerand, skeletal elements of fossil flamingoes (Phoenicopterus croizeti, not very different from the extant Phoenicopterus ruber) are commonly found, and modern flamingoes are known to exploit food sources available only in brackish water, therefore salinity tolerance is also suggested by Donsimoni (1975) for the fossil Indusia. The fossil assemblage of Saint-Gerand, and especially the flamingoes, remind us of the modern community of the brackish Natron Lake in Africa, since one or two million Phoeniconaias minor and several thousand Phoenicopterus ruber roseus live there. Sedimentary conditions are strikingly similar (Hillaire-Marcel and Casanova 1987) but, as the distributional range of the Limnephilinae does not reach Africa, there is no opportunity for the development of trichopteran bioherms. Due to the drying up of the lakes, the one or more Limnephilinae species of Saint-Gerand, with most probably a high degree of dependence on particular habitat features (salinity, volcanic particles supply, high algal productivity), have certainly disappeared, and this explains why we have no modern equivalent of the exceptional Indusial Limestone of Limagne. CONCLUSION Several taxonomic problems remain to be settled, since the two specimens lack crucial features. As the extant adult Limnephilinae, which show considerable flying ability, leave the aquatic environment just after emergence and return to it only to lay their eggs, it is very unlikely that we would find adult Indusia that would permit a more accurate determination; the discovery of prepupae is more likely and should complete the systematic information provided by the pupae. Nevertheless, this discovery emphasizes the remarkable expansion of a trichopteran family which (certainly in response to Tertiary climatic changes) invaded the whole Northern Hemisphere, where it grew more and more numerous and diversified. Acknowledgements. The authors would like to thank Dr Robert Hugueney for photographs 1^, and anonymous reviewers for constructive suggestions. HUGUENEY ET AL. : FRENCH MIOCENE CADDISFLY PUPAE 501 REFERENCES Bosc, L. 1805. Note sur un fossile remarquable de la montagne de Saint-Gerand-le-Puy, enlrc Moulins et Roane, Departcment de I’AlIier, appele I’Indusie tubuleuse. Journal des Mines, 17, 397-400, pi. 7. BOTOSANEANU, L. and wiCHARD, w. 1983. Upper Cretaceous Siberian and Canadian amber caddisflies (Insecta: Trichoptera). Bijdragen tot de Dierkunde, 53. 187-217. BRONGNiART, A. 1810. Sur des terrains qui paraissent avoir ete formes sous Feau douce. Annates du Museum d'Histoire naturelle de Paris, 15, 357^05. CHENEVAL, J. 1984. Les oiseaux aquatiques (Gruit'ormes a Anseriformes) du gisement aquitanien de Saint- Gerand-le-Puy (Allier, France): revision systematique. Paiaeovertehrata, 14, 33-115. COLBURN, E. A. 1983. Effect of elevated temperature on osmotic and ionic regulation in a salt-tolerant caddisfly from Death Valley, California. Journal of Insect Physiology, 29, 363-369. DONSiMONi, M. 1975. Etude des calcaires concretionnes lacustres de FOligocene superieur et de FAquitanien du bassin de Limagne (Massif Central, France). These 3e cycle (unpublished), Universite de Paris VI. and GiOT, D. 1977. Les calcaires concretionnes lacustres de FOligocene superieur et de FAquitanien de Limagne (Massif Central). Bulletin du Bureau de Recherches geologic/ues et minieres, 2, 131 169. FISCHER, F. c. J. 1969. Trichopterorwn Catalogus. Pars 3. Limneplnlidae, X, Nederlandsche entomologische Vereeniging, Amsterdam, 332 pp. GEOFFROY-SAiNT-HiLAiRE, E. 1833. Consideration sur des ossemens fossiles la plupart inconnus, trouves et observes dans les bassins d’Auvergne. Revue encyclopedique, 59, 76-95. GORIN, G. 1975. Etude palynostratigraphique des sediments paleogenes de la Grande Limagne (Massif Central). Bulletin du Bureau de Recherches geologiques et minieres, 1, 147-182. GuiLLOT, L. 1979. Recherches sur les Phryganes de la Limagne tertiaire. Revue scientifique du Bourbonnais, 18-31. HiLLAiRE-MARCEL, c. and CASANOVA, J. 1987. Isotopic hydrology and palaeohydrology of the Magadi ( Kenya )-Natron (Tanzania) basin during the Late Quaternary. Palaeogeography, Palaeoclinuitology, Palaeoecology, 58, 155- 181. INTERNATIONAL CODE OF ZOOLOGICAL NOMENCLATURE. 1985. International trust for zoological nomenclature and British Museum (Natural History), London, 338 pp. KIRBY, w. 1813. Strepsiptera, a new order of insects proposed; and the characters of the order with those of its genera laid down. Transactions of the Linnean Society of London, 11, 86-122. KOLENATi, F. 1848. Genera et species Trichopterorwn, 1, Moscou, 108 pp. 1859. Genera et species Trichopterorwn, 2, Moscou, 143-296. LAURENTIAUX, D. 1953. Classe des Insects. In Piveteau, J. (ed.), Traite de paleontologie. III, 397-527. Masson, Paris, 1063 pp. LEADER, J. p. 1971. Effect of temperature, salinity, and dissolved oxygen concentration upon respiratory activity of the larva of Philanisus plebejus (Trichoptera). Journal of Insect Physiology, 17, 1917-1924. LINNAEUS, c. 1758. Systema naturae, sive regna tria naturae systematice proposita per classes, ordinis, genera et species. I : Regnum animale, 824 pp. Laurentii Salvii, Holmiae. MALiCKY, H. 1974. Fine im marinen Gezeitenbereich lebende europaische Trichopterenlarve. Archiv fiir Hydrobiologie, 73, 266-269. MEIN, p. 1976. Biozonation du Neogene mediterraneen a partir des Mammiferes. Proceedings of the Vlth Congress of the Regional Committee on Mediterranean Neogene Stratigraphy (Bratislava 1975), 2, 18. OUSTALET, E. 1870. Recherches sur les Insectes fossiles des terrains tertiaires de la France. Annales des Sciences geologiques, II, 1-178. SCHMID, F. 1955. Contribution a I'etude des Limnophilidae (Trichoptera). La Concorde, Lausanne, 245 pp. SERRES, M. DE 1829. Geognosie des terrains tertiaires, Pomathio-Durville, Montpellier and Paris, 277 pp. SUTCLIFF, D. w. 1960. Observation on the salinity tolerance and habits of a euryhaline caddis larva, Limnephilus affinis Curtis (Trichoptera: Limnephilidac). Proceedings of the Royal Entomological Society of London, 35, 156-162. ULMER, G. 1912. Die Trichopteren des baltischen Bernsteins. Beitrdge zur Naturkunde Preussens, 10, 1-380, VIALOV, o. s. 1973. Classification of fossil larval tubes (Trichoptera). Academy of Sciences of the Ukrainian Soviet Socialist Republic, 7, 585-588. [In Russian]. and SUKATSHEVA, I. D. 1976. Fossil caddisfly cases (Insecta, Trichoptera) and their stratigraphical value. In Paleontology and Biostratigraphy of Mongolia. The Joint Soviet- Mongolian Paleontological Expedition (Transactions). Nauka, Moscow, 3, 169-232. [In Russian]. 502 PALAEONTOLOGY, VOLUME 33 wiCHARD, w. and reichel, h. 1970. Zur Trichopterenfauna periodischer Gewiisser; Trichopterenstudien zur Duisburger Sechs-Seen-Platte, II. Nachrichtenblatt der Bayerischen Entomologen, 18, 57-58. WIGGINS, G. B. 1973. A contribution to the biology of caddisflies (Trichoptera) in temporary pools. Life Sciences Contributions of the Royal Ontario Museum, 88, 1-28. MARGUERITE HUGUENEY, FRANCOIS ESCUILLIE Centre de Paleontologie stratigraphique et de Paleo- ecologie, Universite Claude-Bernard, Lyon 1, associe an CNRS (URA 11), 27^3 bvd du 11 novembre F-69622 Villeurbanne Cedex, France HENRI TACHET URA CNRS 367 ‘Ecologie des eaux douces’ 43 Typescript received 26 January 1989 bvd du 11 novembre, F-69622 Villeurbanne Cedex Revised typescript received 20 July 1989 France f i 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 Geology, University of Wales College of Cardiff CFl 3YE, 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 1990 is £45-00 (U.S. $66) for Institutional Members, and £20-00 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available on a non-subscription basis to Ordinary and Student Members only, for their personal use, at a discount of 25 % below the listed prices: contact the Marketing Manager. Non-members may obtain Nos 35^0 (at cover price) from Basil Blackwell Ltd., Journal Subscription Department, Marston Book Services, P.O. Box 87, Oxford 0X2 ODT, England, and older issues from the Marketing Manager. For all orders of Special Papers through the Marketing Manager, please add £1-50 (U.S. $3) per item for postage and packing. PALAEONTOLOGICAL ASSOCIATION PUBLICATIONS Special Papers in Palaeontology For full catalogue and price list, send a self-addressed, stamped A4 envelope to the Marketing Manager. Numbers 2-1 1 and 13-30 are still in print and are available together with those listed below; 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. (for 1984): Autecology of Silurian organisms. Edited by M. G. BASSETT anr/ 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., 16 plates. Price £10 (U.S. $15). 35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and d. e. g. BRIGGS. 178 pp., 29 plates. Price £30 (U.S. $50). 36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. KENNEDY. 145 pp., 43 text-figs., 23 plates. Price £20 (U.S. $35). 37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, by rachel wood. 89 pp., 31 text- figs., 1 plates. Price £20 (U.S. $35). 38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, by p. r. BOWN. 118 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50). 39. (for 1988): Late Cenomanian and Turonian ammonite faunas from north-east and central Texas, by w. j. Kennedy, 131 pp., 39 text-figs., 24 plates. Price £30 (U.S. $50). 40. (for 1988): The use and conservation of palaeontological sites. Edited by p. r. crowther and w. a. wimbledon. 200 pp., 31 text-figs. Price £30 (U.S. $50). 41. (for 1989): Late Jurassic-early Cretaceous cephalopods of eastern Alexander Island, Antarctica, by p. j. howlett. 72 pp., 9 text-figs., 10 plates. Price £20 (U.S. $35). 42. (for 1989): The Palaeocene flora of the Isle of Mull, hv m. c. boulter and z. kvacek. 149 pp., 23 text-figs., 23 plates. Price £40 (U.S. $70). Field Guides to Fossils and Other Publications These are available only from the Marketing Manager. Please add £1-00 (U.S. $2) per book for postage and packing plus £1-50 (U.S. $3) for airmail. Payments should be in Sterling or in U.S. dollars, with all exchange charges prepaid. Cheques should be made payable to the Palaeontological Association. 1. (1983): Fossil Plants of the London Clay, bv m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12) (Members £6 or U.S. $9). 2. (1987): Fossils of the Chalk, compiled by E. Owen; edited by a. b. smith. 306 pp., 59 plates. Price £11-50 (U.S. $18) (Members £9-90 or U.S. $15). 3. (1988): Zechstein Reef fossils and their palaeoecology, by n. hollingworth and T. Pettigrew. iv-|-75 pp. Price £4-95 (U.S. $8) (Members £3.75 or U.S. $6). 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 -1-241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24-95. © The Palaeontological Association, 1990 Palaeontology VOLUME 33 • PART 2 CONTENTS Lower Cretaceous spiders from the Sierra de Montsech, north-east Spain P. A. SELDEN 257 Giant acanthomorph acritarchs from the Upper Proterozoic in southern Norway G. VIDAL 287 A discoglossid frog from the Middle Jurassic of England S. E. EVANS, A. R. MILNER and F. MUSSETT 299 Late Cainozoic brachiopods from the coast of Namaqualand, South Africa C. H. C. BRUNTON N. HILLER 313 Early Mississippian Hyolitha from northern Iowa J. M. MALINKY S. SIXT 343 Pseudoplankton P. B. WIGNALL M. J. SIMMS 359 Cenomanian micromorphic ammonites from the Western Interior of the United States W. J. KENNEDY and W. A. COBBAN 379 An application of critical point drying to the comparison of modern and fossilized soft tissues of fishes D. M. MARTILL 0/7(7 L. HARPER 423 Computer-aided restoration of a late Cambrian ceratopygid trilobite from Wales, and its phylogenetic implications N. C. HUGHES and A. W. A. RUSHTON 429 Preservation of avian collagen in Australian Quaternary cave deposits R. F. BAIRD 0/7t/ M. J. ROWLEY 447 Evolution of gryphaeate oysters in the mid-Jurassic of Western Europe A. L. A. JOHNSON (3/7t/ C. D. LENNON 453 The trilobite Protolloydolithus from the Middle Ordovician of north Portugal M. ROMANO 487 Caddisfly pupae from the Miocene Indusial Limestone of Saint-Gerand-le-Puy, France M. HUGUENEY, H. TACHET and F. ESCUILLIE 495 Printed in Great Britain at the University 0031-0239 ■3el C. _ 5 1;^ ^<^inl^ Q L I B R AR I ES^ SMITHSONIAN"" INSTITUTION ^ Ni % “ 5 \ I INSTITUTION NOlinillSNI NVINOSHimS SBIBVaaiT : s3iHvaan CO z ^j) -~ u> ^ in LIBRARIES SMITHSONIAN INSTITUTION NOlifUllSNI NVINOSHillAJS SaiBVMan LI ^ z , in 2 CO 2 INSTITUTION > s NOliniliSNI^NVINOSHill^S^^SB I aVB 3 H^LI B RAR I ES^SMITHSONIAN INSTITUTION Ni — tn ~ ^2/0^ ^ p, 52 m = W “ 0) \ ? CO NOlinillSNI NviNOSHims SBiavHan libraries Smithsonian institution n Z <2 z • ?2 2 vv W < 2 2 O {/> CO X O INSTITUTION 2 .V W CO saiavaan LI B RAR I Es“ SMITHSONIAN‘S INSTITUTI0N'"N0linillSNrNVIN0SHllWs‘^S3 I B VB 8 IT L in ~ (O —: in — 5£ -J • ° MNSTITUTI0N^N0linillSNI^NVIN0SHimS^S3iyViian”^LIBRARIES^SMITHS0NIANj^lNSTITUTI0N^N ” ^ p I i CO 2 m 'sajosjiiX' ™ rn CO ™ c/j ■ — ■ to — CO S3 1 a VM an LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIIAIS SHIMVaaiT LI ” 2 '« to 2 to 2 to > 2 'institution '''N0linillSNI^NVIN0SHims‘^S3 I a VM a n^LI B RAR I Es”sMITHSONIAN INSTITUTION N< CO — €/» n <0 UJ -j sBiavaan libraries smithsonian institution NoiiniiisNi nvinoshiivms saiavaan li 2 z S C 2: I" 2 z < »: O DV^ „ N^illSSX O " — Q X^N PC^ __ -■Sj^r' O Nl“'NVIN0SHilWS^S3 I avy a n"^UBRAR I ES^SMITHSONIAN~'lNSTlTUT!ON^NOIini!iSNl''NVINOSHimS T- -.. 7'. t— ^ r~ -?■ c~ > z “V C) > m , ifl — ujnt/) ±t/> — [S SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHillAIS S3iavaan LIBRARIES SMITHSONIAN rn ' /Ai rs' ffi "5^ •_ (/) 2 < Nl"^NVIN0SHims‘^S3 I aVH8 ll^U BRAR I Es'^SMITHSONIAN^lNSTITUTION '^NOIiniliSNI_NVINOSHilWS “ '•«, tn /’rt /rt —» fn ES SMITHSONIAN INSTITUTION^NOIini!lSNl“*NV!NOSHillAIS^ S3 I B VB 8 L! B RAR I ES SMITHSONIAN o - c w/ (/> :N1 NVINOSHIIINS SBIBVBail LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIINS to ^ . t/i -» tn ^ C/ Es‘”sMITHS0NIAN^INSTITUTI0N‘''N0lini!lSNrNVIN0SHlltMs‘^S3 IBVBaiT^LIBRARI Es'^SMITHSONIAn"' in — if% tn ^ in in INI^NVINOSHlIIAIS^^Sa I B VB a ll^LI BRAR I ES^SMITHSONIAN^INSTITUTION ^NOlinillSNl^NVINOSHlIlAls" — tn “ tn — tn ^ \ o tn ES SMITHSONIAN"*INSTITUTION^NOIinillSNl“‘NVINOSHlllAIS^ S3 I B VB a II"" LI B R AR I ES SMITHSONIAN g C 2 C V . 2 C ^ '■