CO ' f- if) K'AS>''J m xc^osv^z " m _ to — to — co nry g ii LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IINS S3iavaail LIBRAR z co z x co _ z £ /SSMTO& ^ £ XSM vaan libraries Smithsonian institution NouniiiSNi nvinoshiiiais S3iavaan librar m \»/ ' Wf'" 5 > S9B * ~Z. if) z CO * Z CO z vaan libraries Smithsonian institution NouniiiSNi nvinoshiiiais S3 i ava a ii_ li b rar y if Z ^ v CO irXx cn ^ . co ^ ^ ..mix ^ or < CE q [H.Z _ -v/x~ o pc^z ^ X^OcilSglZ^ o X ® O XS\£Cj TUTIONZNOIinillSNl"JNVINOSHimS S3iavaai1 LI B RAR I ES^SMITHSONIAN^INSTITUTION NOlinil. X) > . X) " m m X^os Zjy 11 m _ co " £ co £ co — vaan libraries Smithsonian institution NouniiiSNi nvinoshiiws saiavaan librar z , co Z CO Z CO z . z \X A i /jAfrfaA z xX — i mm/zA z -h z ° cm § (if i?) ° cm * ° a* i W 1 X XF s . t 5 yfr > XiHi/ s ''W >' 5 XXqs'vXz > CO Z to '" Z CO z CO iUTION NOlinillSNI NVIN0SH1IWS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION NOlinil. co ^ > co — co ~ co O ■ W _ Q z • _l z vaan libraries Smithsonian institution NouniiiSNi nvinoshiiws saiavaan librar z r~ z r- z liiis^z m xcvv pc jy' 11 -z/jrs m CO — co X CO _ TUTION NOlinillSNI NVIN0SH1IWS S3iavaai1 LIBRARIES SMITHSONIAN INSTITUTION NOlinil co z _ _ co z , . w z ,Y- £ s - - , v - ^ x - s -pn f£? \ , / // * ~r v\\vvv\N i /rT^llV'lA -*■ . O * Q #S.X. X , "'V co \ S X^x ~ _ CO _ CO HSONIAN INSTITUTION N0I1D1I1SNI NVIN0SH1IWS S3ldVyan LI B R A R I E S SMITHSONIAN INSTITU z , co 2: CO X , CO z X GO o z * > ' __ _ ^ 2 ’’*• > Z CO * Z cov z co x HSONIAN INSTITUTION NOIlfllllSNI NVIN0SH1IWS S3iyVHan LIBRARIES SMITHSONIAN _ INSTITl CO “ CO -p CO 2! 7 — ■ W aC\ x UJ rn XaVASO^X LU . CO O nLy _ o v X o IOSHIIWS ^ S3 I yvy 8 ll^LI BRAR I ES^SMITHSONIAN-1 INSTITUTION NOlinillSNI ~NVIN0SH1IWS S3 I y V Z r- Z r- Z <\vnr, H /y^CT* V'W m x^osv'ix — m x^vos^ix r; x^^x- m CO _ CO — CO — FHSONIAN INSTITUTION NOlinillSNI NVIN0SH1IWS S3iyvyaiT LIBRARIES SMITHSONIAN INSTITl z ■ * co z: co 2: * co z > _ : > A! i -•■ 2 4 6 DALINGWATER and MILLER, trilobite cuticle 26 PALAEONTOLOGY, VOLUME 20 more or less horizontal fibres are visible in the plates of inter-lamina material just below a lamina (PI. 9, fig. 6). Decalcification of the cuticle. One exceptional preparation examined and photo- graphed by reflected light microscopy (PL 10, fig. 2) deserves mention here. A thoracic pleura was removed from a large specimen of A. raniceps and a clean longitudinal perpendicular break achieved. Decalcification was allowed to proceed for five hours and the process observed at regular intervals. Laminae in the central zone of the cuticle were immediately apparent and some perpendicular ducts, about 5p. wide, could just be seen at this stage. After about five hours, much of the inorganic material of the cuticle had been dissolved inwards for some distance from the break surface leaving a layer of brown jelly-like material, with lighter-coloured material standing up through the gel at the laminae. The 5^ ducts could now be seen in depth piercing the brown jelly, their component material apparently unaffected by the E.D.T.A. Dissolution of inorganic material seemed greatest at the outer and inner zones of the cuticle, despite protection of the edges by adhering matrix. DISCUSSION Laminae and pore canals. Teigler and Towe (1975) gave a number of reasons why the crudely laminate aspect of the calcite in some of the trilobite cuticles they examined should not be considered similar to the laminae in many extant arthropod cuticles: 1. Using the TEM they found no difference in the texture of the calcitic material at the laminae, only a slight furrowing, and decided that the laminae could have resulted from variation in the concentration of organic material within the cuticle rather than from some basic variation in the nature of the calcite crystals. However, laminae are essentially reflections of the organization of organic material within the cuticle; the essence of the lamina seems to be an organic lamina membrane (Dennell 1976). Miller’s (1976) observation of flimsy organic membranes left repre- senting the laminae while etching the cuticle of Phacops rana , as well as Teigler and Towe’s comments on the distribution of organic material in the cuticles they studied, seem to indicate that in this context the laminae in trilobite cuticles can be seen to resemble those of many extant arthropods. The furrowing at the laminae in the cuticle of the fossil decapod crustacean studied (PI. 10, fig. 3) strongly resembles EXPLANATION OF PLATE 10 Figs. 1,2. Asaphus raniceps Dalman. 1 , SEM photo-montage of a transverse perpendicular break of cephalic cuticle. The outer laminate zone can just be seen on the left. The wide laminar units of the central zone are prominent along the length of the preparation, except where the cuticle deflects to form the doublure. Prep. 21375-2. x 35. 2, reflected light micrograph of a longitudinal perpendicular break across a thoracic pleura after about five hours decalcification. Prep. 191274-0. x65. Fig. 3. Hoploparia longimana (Sowerby). Transverse perpendicular break of the cuticle from a thoracic limb podite. Two laminar units in the calcified zone. L— lamina. Prep. FI.s. 141175-la. x2100. Fig. 4. Austropotamobius pallipes (Lereboullet). Transverse perpendicular break of the complete thick- ness of the cuticle from a great chela dactylopodite. Ep— epicuticle; Ex — exocuticle; Cz— calcified zone; En — endocuticle. Prep. A.p. 3376- 2a. x280. PLATE 10 DALINGWATER and MILLER, arthropod cuticle 28 PALAEONTOLOGY, VOLUME 20 that seen in the trilobite cuticle. In both cases the absolute alcohol preparative treatment would remove any concentration of organic material from these furrows, for we have found that the organic residues left after complete decalcification of both trilobite and fossil decapod crustacean cuticles are completely dispersed by absolute alcohol. Incidentally, this might explain why well-preserved brown-coloured trilobite exoskeletons are ‘whitened’ by alcohol treatment. 2. They decided that the absence of helical pore canals and parabolic patterns also rendered difficult direct comparison with the usual extant arthropod cuticular laminae. However, pore canals are not necessarily helical (Travis 1970; Mutvei 1974) , nor need their pattern be related to helicoidal architecture (Kennaugh 1965; Travis 1970; Dalingwater 1975). Even if the pore canals of trilobites were originally helical then this could have been altered in the process of fossilization (see discussion in Osmolska 1975). For example, only a straight contained filament or a straight lining might be preserved, with the canal itself (the lumen) infilled in the latter case. Parabolic patterns are only observed in angled sections (classically 45° sections) of extant cuticles, and only in some angled sections according to Drach (1939) and Dennell (1974). Many of Teigler and Towe’s figured sections are more or less per- pendicular slices; one would not expect parabolic patterns in such sections. Again diagenesis might alter the material and make recognition of parabolae in angled sections difficult. 3. They noted that not all trilobite species, even those from the same locality, show laminae, and concluded that a ‘genetic control’ appears to be the major factor in their distribution. We suggest that the precise history of preservation (see Miller 1975) , and possibly the stage in the moult cycle reached before death in the case of intaglios, may determine whether laminae are prominent or even visible in examples studied. Methods of preparation and examination are clearly important: Dalingwater (1973) made an extensive investigation of the cuticle of A. raniceps using specimens from the same locality and horizon as those described above; he included an SEM examination of break surfaces and etched ground slices, but did not observe the distinct laminae that we have demonstrated by the break + etch technique. We suggest that the evidence favours comparability of the laminae of trilobite cuticles, especially those of A. raniceps, with those of the typical arthropod cuticle. Some of the laminar units in A. raniceps cuticle are extremely wide (up to 70/x) compared with those encountered in most extant cuticles, although in the crayfish Austropotamobius pallipes, laminar units about 20/x wide are present in the calcified zone (Dalingwater 1975). However, the trilobite cuticle is usually relatively very thick for an arthropod often only a few centimetres long. These laminar units may represent material laid down rapidly after ecdysis (although it must be admitted that no evidence seems to have been presented correlating width of laminar unit with rate of cuticle formation). Finer co-planar structures within very wide laminar units have been reported from light microscope studies (e.g. Stormer 1930), and occasionally similar fine structures are encountered in SEM preparations. It is possible that these reflect the organization of inter-lamina material in low-angle arcs between laminae. The ultrastructural detail just recognizable in some prepara- tions and described above tends to add strength to this view. DALINGWATER AND MILLER: TRILOBITE CUTICLE 29 Perpendicular canals. Teigler and Towe (1975) termed all primary duct-like processes in trilobite cuticles pore canals. Whilst we recognize that it is difficult to decide whether canals of relatively narrow dimension (of the order of l^i in diameter) are exactly equivalent to the pore canals of extant arthropod cuticles, or are merely unusually narrow tegumental or setal ducts, and that there is some imprecision in the definition of the pore canal even in extant cuticles, we are unhappy with Teigler and Towe’s terminology. Canals 15p and more wide are clearly not comparable with the pore canals of extant arthropod cuticles. Their proposal would place trilobite terminology in step with ostracod cuticle terminology, but out of step with that employed for all other arthropod cuticles. The fine perpendicular canal-like elements present in the inter-laminae of many preparations we have examined are considered similar to the pore canals of extant arthropod cuticles, in that they are relatively narrow (about Ip in diameter), are very numerous, and have been observed in all regions of the cuticle examined. Organic material. The presence of brown organic material in trilobite cuticles has long been recognized. Davies (1894) considered it possibly to be a replacement product of the original organic material of the cuticle, Dalingwater (1973) isolated organic material from a trilobite cuticle after decalcification with E.D.T.A., and Teigler and Towe (1975) carried this study further by examining sections of the material with the TEM. Their observation of a delicate meshwork does not eliminate the possibility of the presence of laminae comparable with those of most extant arthropod cuticles, for their sections are apparently more or less horizontal and the pattern figured is not incompatible with the architecture observed in horizontal slices or breaks of extant cuticles (see Travis 1970; Dalingwater 1975). Further work is needed to establish the nature of this relict organic material and to determine its precise relationship with the inorganic component of the trilobite cuticle. Subdivisions of the cuticle. In our description of the cuticle of Asaphus raniceps we define three zones based on the appearance and dimensions of their laminar units. The central and innermost zones apparently correspond with the median and inner subdivisions that Stormer (1930) defined in his classic study of the cuticle of Tretaspis kiaeri, although as previously noted (Dalingwater 1973) his pigmented layer is a dense micritic envelope. The term zone is deliberately used in this paper to avoid confusion with the layers of the trilobite cuticle recently defined mainly from light microscope observations (see text-fig. 2). The outer prismatic layer was not always present prior to etching (possibly removed with the overlying matrix), but when it was present at least part of it was etched back very rapidly in E.D.T.A., so the zones defined herein possibly all lie within the principal layer (the central and inner laminate zones certainly do). As the outer layer of the cuticle can show aspects other than prismatic (Teigler and Towe 1975), and as it is quite possible that this part of the cuticle may subsequently be shown also to contain laminae, at present we prefer to reserve judgement on the precise relationship of the outer prismatic layer and our outer laminate zone. Examination of newly formed cuticles may help to solve this problem and work is in progress on the material described by Miller et al. (1974). We are confident that the trilobite cuticle described is similar to those of many extant arthropods in laminar organization and divisibility into zones based on 30 PALAEONTOLOGY, VOLUME 20 . N|I[M|IN 1 11 II I I I 1 I 1 11 1TTTITT [TIT1 ffTTI f I II I II 11 I I Her, prismatic layer outer laminate zone - = : central laminate zone inner laminate zone principal layer text-fig. 2. The major layers and laminate zones of the cuticle of Asaphus raniceps. Right : the previously defined layers (Dalingwater 1973). Left : the laminate zones defined in this paper. laminar disposition. However, it is difficult to make any precise correlation of the trilobite zones and layers with those of any specific group of extant arthropods, for there is some variability in cuticular subdivisions between the different grgups of extant arthropods, and the exact subdivision is based largely on histochemical criteria. Nevertheless, in being impregnated with mineral salts and in its general laminar organization, the trilobite cuticle described here most favourably resembles that of decapod crustaceans (e.g. Austropotamobius pallipes , see PI. 10, fig. 4). We agree with Teigler and Towe (1975) that the trilobite cuticle is unusual in being so heavily impregnated with calcite, but suggest that this enables the production of a relatively very thick and strong cuticle, economic in terms of organic material. However, we would hesitate to attach any phylogenetic significance to this particular feature for other groups of crustaceans as well as the ostracods (e.g. cirripedes) have produced heavily calcitized cuticles in functional response to particular environ- mental demands. We must also add that in comparing the trilobite cuticle favourably with that of decapod crustaceans, we attach no phylogenetic significance to our analogy. The epicuticle of many extant decapod crustaceans is calcified (Digby 1967; Welinder 1975) so it is quite possible that traces of a similar layer might be preserved in trilobites; the outermost layer described by Dalingwater (1973) is positionally and DALINGWATER AND MILLER: TRILOBITE CUTICLE 31 structurally compatible with the extant decapod epicuticle. The prismatic layer and the outer laminate zone can be considered equivalent to the exocuticle, and the central zone to the calcified principal zone, of extant decapods. Comparison of the innermost zone of the trilobite cuticle described herein with the inner region of extant decapod cuticles presents a difficulty, for in the latter this is usually an un- calcified endocuticle refractory to fossilization. However, in some regions of, for example, crayfish cuticle, the uncalcified endocuticle laterally transforms into a region which is calcified, strictly speaking therefore part of the calcified principal zone, but still distinguishable by the relative narrowness of its laminar units (Dr. J. Kennaugh, pers. comm.). A similar situation may exist in the trilobite cuticle. Further investigation of the inner region of the trilobite cuticle is necessary, par- ticularly to determine if there are any differences between the cuticles of intaglios and exuviae. This could reveal whether there is any resorption of material prior to ecdysis. Acknowledgements. We thank Professor R. Dennell and Dr. E. N. K. Clarkson for their critical reading of the manuscript, Mr. B. Atherton for photographic services, and the Textile Technology Department, U.M.I.S.T. for SEM facilities. We are particularly grateful to Drs. K. Towe and D. Teigler for allowing us to read the manuscript of their paper prior to its publication. Although we disagree with their con- clusions on trilobite cuticular laminae, we are in agreement on many other aspects of the trilobite cuticle. REFERENCES angelin, N. p. 1854. Palaeontologia Scandinavica. Part I. Crustacea Formationis Transitionis. Fasc. 2. Leipzig, T. O. Weigel. 92 pp. bate, r. h. and east, b. a. 1972. The structure of the ostracode carapace. Lethaia, 5, 177-194. bohlin, b. 1949. The Asaphus limestone in northernmost Oland. Bull. geol. Instn Unix. Upsala , 33, 529-570. bouligand, Y. 1965. Sur une architecture torsadee repandue dans de nombreuses cuticules d’Arthropodes. C. r. hebd. Seanc. Acad. Sci., Paris , 261, 3665-3668. 1972. Twisted fibrous arrangements in biological materials and cholesteric mesophases. Tissue and Cell, 4, 189-217. dalingwater, j. e. 1973. Trilobite cuticle microstructure and composition. Palaeontology , 16, 827-839. 1975. SEM observations on the cuticles of some decapod crustaceans. Zool. J. Linn. Soc. 56, 327-330. davies, a. m. 1894. On the Minute Structure of the Trilobite-crust. 24 pp. Southern Counties Press, Lewes. dennell, r. 1973. The structure of the cuticle of the shore-crab Carcinus maenas (L.). Zool. J. Linn. Soc. 52, 159-163. 1974. The cuticle of the crabs Cancer pagurus L. and Carcinus maenas (L.). Ibid. 54, 241-245. 1976. The structure and lamination of some arthropod cuticles. Ibid. 58, 159-164. digby, p. s. b. 1967. Calcification and its mechanism in the shore-crab Carcinus maenas (L.). Proc. Linn. Soc. Lond. 178, 129-146. drach, p. 1939. Mue et cycle d’intermue chez les Crustaces Decapodes. Annls Inst. Oceanogr., Monaco , 19, 103-391. kennaugh, j. h. 1965. Pore canals in the cuticle of Hypoderma bovis (Diptera). Nature , Lond. 205, 207. miller, j. 1975. Structure and function of trilobite terrace lines. Fossils and Strata , 4, 155-178. — 1976. The sensory fields and life mode of Phacops rana (Green, 1832) (Trilobita). Trans. R. Soc. Edinb. 69, 337-367. Clarkson, E. N. K and dalingwater, J. E. 1974. A moulted trilobite. Trilobite News , 3, 60-62. mutvei, h. 1974. SEM studies on arthropod exoskeletons. Part 1 : Decapod crustaceans Homarus gammarus L. and Carcinus maenas (L.). Bull. geol. Instn Univ. Upsala , N.s. 4, 73-80. Neville, a. c., thomas, m. g. and zelazny, b. 1969. Pore canal shape related to molecular architecture of arthropod cuticle. Tissue and Cell, 1, 183-200. 32 PALAEONTOLOGY, VOLUME 20 osmolska, h. 1975. Fine morphological characters of some Upper Palaeozoic trilobites. Fossils and Strata , 4, 201-207. Richards, a. g. 1951. The Integument of Arthropods. 411 pp. University of Minnesota Press, Minneapolis. rolfe, w. d. i. 1962. The cuticle of some middle Silurian ceratiocaridid Crustacea from Scotland. Palaeonto- logy, 5, 30-51 . st0rmer, l. 1930. Scandinavian Trinucleidae with special reference to Norwegian species and varieties. Skr. norske Vidensk.-Acad. Mat.-naturv. Kl. 4, 1-111. teigler, d. J. and towe, K. m. 1975. Microstructure and composition of the trilobite exoskeleton. Fossils and Strata, 4, 137-149. travis, d. f. 1970. The comparative ultrastructure and organisation of five calcified tissues. In schraer, h. (ed.). Biological Calcification : Cellular and Molecular Aspects , 203-311. North-Flolland, Amsterdam. welinder, b. s. 1975. The crustacean cuticle— III. Composition of the individual layers in Cancer pagurus cuticle. Comp. Biochem. Physiol. 52A, 659-663. J. E. DALINGWATER Department of Zoology The University Manchester M13 9PL Typescript received 27 August 1975 Revised typescript received 6 April 1976 J.' MILLER Department of Educational Studies The University Edinburgh EH8 9JT CALCIFIED PLECTONEMA (BLUE-GREEN ALGAE), A RECENT EXAMPLE OF GIRVANELLA FROM ALDABRA ATOLL by ROBERT RIDING Abstract. Girvanella, not previously reported from rocks younger than Cretaceous, is described from the Recent. It occurs as the calcified sheath of the filamentous blue-green alga Plectonema gloeophi/um Borzi in small freshwater pools on Aldabra Atoll in the Indian Ocean. The sheaths are heavily impregnated during the life of the alga by micrite grade crystals of magnesian calcite which make them hard and potentially fossilizable and which preserve the filaments as small calcareous tubes 4-10 fim in diameter. This confirms the view that Girvanella represents a filamentous blue-green alga preserved by the calcification of its sheath, and it extends the range of the genus from Cambrian to Recent. The consistent relatively thin-walled morphology of Girvanella- type impregnated sheaths contrasts with the thickly microsparite encrusted sheaths of blue-green algae in environments of CaC03 cementa- tion. This suggests a greater degree of control of the calcification process in Girvanella and supports the concept of specificity for calcification in some blue-greens. P. gloeophilum is believed to be only one of a number of extant filamentous blue-greens whose calcified sheaths are referable to Girvanella. At present Girvanella is still most con- veniently placed in the Porostromata, which should be regarded as a group of tubiform calcareous algae. The size and morphology of the microscopic tubiform fossil Girvanella Nicholson and Etheridge suggest that it is the calcified sheath of a filamentous blue-green alga (Pollock 1918; Riding 1972, 1975). Blue-greens are among the oldest known fossils and the similarities between very ancient and extant forms indicate that they are morphologically conservative organisms (Schopf 1968, p. 653). If Girvanella is indeed a blue-green alga then it is reasonable to expect that it has a modern repre- sentative. Yet although Girvanella is known to range from the Cambrian to Creta- ceous it has not previously been recorded from the Cenozoic (Riding 1975). The purpose of this paper is to draw attention to a Recent example of Girvanella from the Indian Ocean atoll of Aldabra. This living Girvanella is the small filamentous blue-green alga Plectonema gloeo- philum Borzi. It inhabits freshwater pools on Aldabra and its sheath in these habitats sometimes becomes heavily impregnated during life by CaC03 which renders it hard and potentially fossilizable. The resulting simple sinuous calcareous tube is a small form of Girvanella. It seems likely that a variety of filamentous blue-greens from several families can, under suitable conditions, produce calcareous tubes referable to Girvanella or to related fossil genera. Thus, P. gloeophilum is thought to be only one of a number of extant blue-greens whose soft parts show them to be quite distinct taxa, but whose simple calcified sheaths can be bracketed together in the single fossil genus Girvanella. CALCIFIED BLUE-GREEN ALGAE Cyanophytes have a long geological record as calcified and silicified fossils and as algal stromatolites. Stromatolites are organosedimentary structures; they do not [Palaeontology, Vol. 20, Part 1, 1977, pp. 33-46, pi. 11.] C 34 PALAEONTOLOGY, VOLUME 20 readily reflect the types of organisms involved in their construction, which may include coccoid and filamentous bacteria and blue-green algae and also green algae. By contrast silicified fossil blue-greens can show remarkable preservation of the soft tissue, including details of cellular ultrastructure. Thus, stromatolitic and silicified blue-green algae represent extremes of preservation. They also differ markedly in abundance, stromatolites being relatively common whereas silicified microfloras are rather rare. Calcified blue-green algae are intermediate between stromatolitic and silicified forms in the degree of morphological detail which they retain, and probably also in their over-all abundance through geological time. In the discussion of calcification (see below) it will be necessary to distinguish between calcification of the sheath on the one hand and massive encrustation of the filament on the other, both being types of calcification. But in Plectonema and Girvanella we are only dealing with the former. Calcification in this case only affects the mucilaginous sheath enclosing the strands of cells (trichomes). It takes place while the alga is living. Because CaC03 is incorporated into the sheath rather than into the cell walls of the alga only the external morphology of the algal filament is preserved and most of the features which are significant for the taxonomy of Recent blue-greens are lost (text-fig. 1). Filamentous text-fig. 1. Post-mortem loss of morphological detail in calcified filamentous blue-green algae. The alga shown is Tolypothrix , a scytonematacean similar to Plectonema but possessing heterocysts, a , portion of living calcified filament, b, empty, potentially fossilizable, calcified sheath. Features observable in a but not in b include: presence of heterocyst, type of branching (false), number (one) and size of trichomes. The sections are longitudinal. blue-green algae calcified in this way produce microscopic, morphologically simple, non-septate tubiform skeletons which may exhibit branching but which show few other distinctive characters. The relatively large diameters of these calcareous tubes and their lack of septa clearly distinguish them from the rather small trichomes of blue-green algae which are strands of cells. The expectation by some previous workers that fossil calcified blue-greens should reflect the form of the trichomes, and so be very small and septate, has tended to hinder recognition of the cyanophyte origin of the larger, non-septate calcified sheaths. Nevertheless, Pia (1927, pp. 37-40) realized the true affinities of several of these small fossils but, being unable to relate them directly to Recent blue-green taxa, he RIDING: CALCIFIED PLECTONEM A FROM ALDABRA 35 erected the group Porostromata to contain them. Not all of the genera originally included by Pia in the Porostromata appear to be blue-greens, but two of them, Girvanella and Ortonella Garwood (text-fig. 2), are regarded here as definitely having this affinity. Johnson’s (1961, pp. 96-99) removal of several porostromate genera to the green Codiales was unwarranted (Riding 1975) and has subsequently confused the identity of fossil calcified filamentous blue-green algae. Much earlier Pollock (1918) had already made the suggestion, which is confirmed here, that Girvanella is a calcified cyanophyte sheath. text-fig. 2. Morphology and orientation of a , Girvanella , and b, Ortonella, from photomicrographs of the type-specimens in Nicholson and Etheridge 1878, pi. 9, fig. 24; and Garwood 1914, pi. 20, fig. 2. Girvanella is flexuous, prostrate, and generally unbranched. Ortonella is straight, erect, and dichotomously branched. These contrasting features were used by Pia (1927) to separate the Porostromata into the Agathidia and Thamnidia respectively. While it is unlikely that fossil calcified blue-greens can be referred to single extant genera it should be possible to discern closer relationships between fossil and living forms than are expressed by current palaeontological classifications and hence to improve our understanding of the nature and environmental significance of fossil calcified blue-green algae. I have suggested that the existence of Recent Girvanella is to be expected (Riding 1972) and that representatives are most likely to be found among filamentous blue-green families such as the Oscillatoriaceae, Rivulariaceae, Scytonemataceae, and Stigonemataceae (Riding 1975). The small Recent scyto- nematacean P. gloeophilum provides one such example of living Girvanella. CALCIFIED PLECTONEM A FROM ALDABRA Habitat. P. gloeophilum occurs in small freshwater pools on West Island, Aldabra; an atoll in the western Indian Ocean, 1200 km WSW. of the Seychelles. Aldabra receives an annual rainfall of approximately 1000 mm (D. R. Stoddart, pers. comm.). The freshwater pools, classed as ‘temporary rainwater rockholes’ by McKenzie (1971) are mostly very small, with volumes usually less than 3 m3. Their water chemistry and algal flora have been studied by Donaldson and Whitton (1976a, b). The pools provide water for tortoises, birds, crabs, and other animals and tend to become contaminated by excreta which produce 36 PALAEONTOLOGY, VOLUME 20 high concentrations of phosphate and ammonia. There is locally also some sea-water contamination of the pools, but they are essentially fresh. The pools support a varied algal flora with filamentous blue- greens such as Calothrix , Oscillatoria , Phormidium , and Plectonema usually present (Donaldson and Whitton 19766). The locations of Aldabran pools are described by Donaldson and Whitton (unpublished). The specimens of Plectonema described here were collected by Alan Donaldson from adjacent pools W7 and W127 on West Island (text-fig. 3) and he has kindly supplied the following details of their occur- rence. These pools have volumes of approximately 1 8 and 0-3 m3 respectively. Plectonema forms mats on Middle Island text-fig. 3. Location (X) of adjacent P/ectonema-contdining pools W7 and W127 on West Island, Aldabra. lime mud and silt flooring the pools. Collection of gas bubbles within the mats lifts them from the sediment and they rise to the surface to form dense globular floes approximately 1 cm in diameter which are pale pink in colour and heavily calcified. In pool W7 from January to June 1973 Plectonema was associated with a wide variety of unicellular and filamentous green and blue-green algae. The following parameters of the pool-water were recorded by Donaldson during this period : mean Ca concentration: 72T mgU1 mean Mg concentration: 18-8 nrgl 1 pH: 7-5-8-6 temperature: 25-3-36-7 °C. No measurements were made of pool W127 situated some 50 m north of pool W7. In April 1973 pool W127 contained a virtual uniculture of P. gloeophilium with abundant calcified floes, samples of which are described and illustrated here. Methods. The samples of Plectonema were air-dried in the shade following collection. For atomic absorp- tion spectrophotometry small specimens were dissolved in 20% HC1. The insoluble residue constituted less than 5% by weight of the specimen. Mineralogy was determined by X-ray diffractometry. For scanning electron microscopy specimens were fractured and coated with carbon and gold. The photographs were taken on a Cambridge Stereoscan Mark II instrument operating at 20 kV. Description. P. gloeophilum is a small, filamentous blue-green which shows false-branching and is non- heterocystous. The sheath is generally thin and cofitains a single trichome. The entire filament is 4-10 ^m in diameter, and the sheath is up to 3 ^m in thickness. Calcified sheaths of this species from pool W127 are figured in Plate 1 1 . They form dense tangled masses of sinuous macaroni-like tubes. The CaC03 occurs as bladed and acicular micrite crystals 1 -3 ^m in size which form tufted clusters crudely orientated normal to the surface of the filament giving it a test-tube brush appearance. These crystals are calcite containing 4-6 mole % MgC03 and thus are close to the 4 mole % division between low- and high-magnesian calcite. Traces of the sheaths occur as patches and strands on the surfaces of the calcified tubes in the SEM. Biological preparations of the material show the calcite crystals to be contained within or very close to the surface of the sheath. Drying of the sheath in air, and in vacuum during preparation for electron RIDING: CALCIFIED PLECTONEMA FROM ALDABRA 37 microscopy, causes it to shrink and exposes the crystals which then form very irregular inner and outer surfaces to the tube. The calcified tubes are approximately 6 n nr in external diameter and 2 /un in the thickness of their walls. They exceed 140 /urn in length and do not taper. Comparison with Girvanella. Girvanella is a small, flexuous, rarely branching, tubi- form fossil with a microgranular (micritic) calcareous wall (text-fig. 4). Specimens text-fig. 4. Girvanella sp., from the Upper Devonian (Frasnian), Mount Flawk Formation, Mount Flaultain, Alberta. Note constancy of tube diameter and rela- tively thin wall. Thin section courtesy of E. W. Mount- joy (McGill University); x 125. with external tube diameters from 1 to 100 ^m are recorded in the literature. Whether the larger sizes really represent Girvanella is open to question but certainly the genus exhibits a large size range. Most specimens recorded have diameters in the range 8-30 fim. For each specimen the tube diameter is constant and the total wall thick- ness is approximately 50% of the external diameter. Girvanella occurs encrusting grains or as small discrete masses which are themselves grains. Whether the latter are fragments or are original unbroken masses of filaments is not usually clear. The similarity between Girvanella and the calcified Plectonema sheaths described here is striking. The ratio of wall thickness to tube diameter in Plectonema (2:3) 38 PALAEONTOLOGY, VOLUME 20 somewhat exceeds that in Girvanella (approximately 1 : 2) and the surface of the tube is more irregular, but it is likely that both these features would be reduced during diagenesis. Otherwise there are no observable differences between Girvanella and these specimens of Plectonema. OTHER CALCIFIED ALGAL FILAMENTS Although this is the first specific report of Recent Girvanella , calcified algal filaments have previously been noted in several studies of Recent intertidal and supratidal algal mats from the Bahamas. Black (1933, p. 170, pi. 22, fig. 28) first noted that the supratidal algal mats of the interior of Andros Island tend to be calcified. Monty (1967, p. B67, pi. 6) compared calcified filaments of Scytonema myochrous from the supratidal of eastern Andros with Ortonella. These specimens of Scytonema are erect and dichotomously branched and are approximately 20 /x m in external diameter. Calcified Scytonema filaments have also been figured from the supratidal of north- western Andros by Shinn et al. (1969, fig. 12 b) but these specimens are less upright in appearance and resemble Girvanella more than Ortonella. Similar tubules have been described from calcareous crusts in late Pleistocene limestones of Barbados (James 1972, fig. 6 d). They are 15-20 ^m in external diameter with a wall thickness of approximately 5 ^m and are composed mainly of equant calcite crystals T5- 2-5 fx m in size (James 1972, p. 826). James compared them with filaments of the extant blue-green alga Schizothrix and attributed them to ‘blue-green algae or possibly root-hairs’. They closely resemble the calcified Plectonema sheaths from Aldabra in morphology, but they are larger. Calcite encrusted cyanophyte filaments which are rather different from those described here have been reported from freshwater stream tufa in south-western Germany (Irion and Muller 1968). The filaments belong to Recent oscillatoriacean blue-greens ‘with a diameter of 5 ; they are encrusted by isometric crystals with a rounded surface. The average diameter of the crystals is 15 p m’ (Irion and Muller 1968, p. 165). Apparently the crystals form a thick crust on, rather than within , the sheath. The specimens are not figured but it can be inferred that they would be tubes with diameters of at least 35 ^m. Their thick walls (more than 85% of the tube diameter) and microsparite grade crystals distinguish them from Girvanella and Plectonema tubes. EXPLANATION OF PLATE 11 Scanning electron micrographs of calcified sheaths of Plectonema gloeophilum Borzi from freshwater pool W7, West Island, Aldabra. Fig. 1. Tangled 'ball of knitting wool’ appearance of mass of tubes, x 220. Fig. 2. Detail of Fig. 1 showing unbranched, uniform diameter sheaths associated with thin dried strands of mucilage. Large uncalcified filament in the upper part of the photograph is probably the green alga Oedogonium , x 1100. Fig. 3. Surface of calcified filament showing bladed and acicular micrite grade crystals of magnesian calcite. Note general orientation of crystals normal to sheath surface, and traces of dried mucilage, x 5500. Fig. 4. Fracture (or termination) of filament showing relatively thin wall and hollow interior, x 5400. PLATE 11 RIDING, calcified Plectonema 40 PALAEONTOLOGY, VOLUME 20 There are relatively few records of Recent calcified blue-green filaments from subtidal marine environments. Winland and Matthews (1974, fig. 3c) figure calcified sheath material associated with grapestone from Great Bahama Bank. The form of the filament is not shown but it has a diameter of 10-12 and a sheath thickness of approximately 2 ^m. The CaC03 crystals are 1-2 in size and have rhombic and acicular forms. These sizes and shapes are similar to those of the crystals in Plectonema from Aldabra. A circular structure, which may be a cross-section of a similar filament, has been figured from Recent Red Sea reefs (Friedman et al. 1974, fig. 15), although it could also be a pellet. It is 9 /xm in diameter with a wall thickness of 3 ^m and consists of high-magnesian calcite. Calcified filaments of the green alga Ostreobium have been described from Recent Bermudan cup reefs (Schroeder 1972). This endolithic alga has a diameter of 2-20 ^m. The principal morphological difference between Ostreobium and blue-green algal filaments is the local swellings shown by the green alga. Ostreobium is not normally calcified but in small internal cavities in the cup reefs it has become encrusted by equant, bladed, or fibrous crystals of high-magnesian calcite during early subsea cementation. The calcite crystals also encrust adjacent non-algal surfaces. The crystals are relatively large with widths of 5 /xm and lengths of 10-20 jx m being common. Not only are the Ostreobium filaments externally encrusted but, following decay of the alga, the remaining cavity is also infilled by calcite crystals. Hence the filaments are preserved by thick external crusts plus internal cores. The core is up to 20 ^m in diameter and the crust is up to 150 /xm in thickness (Schroeder 1972, p. 20). Thus, the total dia- meter of the calcified filament ranges from 15 to 300 ^m or more, and it is ultimately preserved as a rod rather than as a tube. Heavily encrusted filaments of this type resemble those of blue-green algae from freshwater streams and tufa deposits. D. R. Kobluk and M. J. Risk inform me that they have seen calcification of Ostreobium filaments projecting from experimentally introduced Iceland spar crystals in a shallow subtidal environment at Jamaica. Micrite and microspar low- magnesian calcite crystals coat the inner and outer surfaces of exposed dead filaments and give rise to thickly encrusted rods 8-30 (um in diameter comparable with those described by Schroeder. DISCUSSION There are two distinct types of calcified blue-green algal filament : micrit e-impregnated sheaths and microspar -encrusted sheaths (text-fig. 5). Calcified Aldabran Plectonema , and by analogy fossil Girvanella , represent impregnated sheaths without significant external encrustation. The Bahaman filaments described above from both the supra- tidal (Black 1933; Monty 1967; Shinn et al. 1969) and subtidal (Winland and Matthews 1974) fall within the same category. In contrast blue-green filaments in freshwater tufa (Irion and Muller 1 968) are thickly encrusted, as are subtidal cemented green algal filaments (Schroeder 1972). Any heavily encrusted filaments recorded as fossils should be distinguishable from Girvanella and Ortonella tubes in being wider, having greater wall thickness, being composed of coarser crystals, showing variability in diameter within a single speci- men, and in some cases by being infilled by additional cement to form a rod. These RIDING: CALCIFIED PLECTONEMA FROM ALDABRA 41 text-fig. 5. a , micrite impregnated, and b , microspar encrusted sheaths in diagrammatic cross-section. The consistent thin-walled morphology of Girvanella (type a sheath) throughout its geological history indicates that a and b do not simply represent different stages of the same process but that different mechan- isms of calcification are involved : ‘controlled’ (essentially biochemical) in a , ‘uncontrolled’ (essentially physicochemical) in b. contrasts between thickly encrusted sheaths and impregnated sheaths suggest that at least the rate of the calcification process is operating differently in the two cases, but there is also circumstantial evidence which suggests that the type of process differs as well. Calcification. It has gradually become clear that calcification processes operate quite differently in different algal groups (see Arnott and Pautard 1970) and are to some extent reflected by the site of CaC03 deposition on or within the plant. In the red Corallinaceae the site of calcification is the cell wall; in the green Codiales and Dasycladales it is mainly extracellular but the calcareous skeleton forms an intimate part of the plant body. In Aldabran Plectonema , and apparently in calcified blue-greens generally, calcification is entirely extracellular, being restricted to the mucilaginous external sheath. This suggests a mechanism which is not entirely under the control of the plant’s metabolism and is supported by the variable occur- rence of calcification in the group since no blue-green species is known to be always calcified. It is tempting to go on to hypothesize an essentially physicochemical precipitation mechanism in calcified blue-greens, even though this is not applicable 42 PALAEONTOLOGY, VOLUME 20 to any other calcareous algae, and two supporting lines of evidence can be derived from Aldabran Plectonema. (i) Environment of occurrence. CaC03 precipitation due to photosynthetic utiliza- tion of C02 will be limited in normal marine environments by the buffering capacity of sea-water, but this regulation will not operate in fresh water where removal of C02 species can readily cause precipitation. The relatively high temperature, pH, and calcium concentration of the Aldabran pools containing calcified Plectonema are conditions favouring CaC03 precipitation and removal of C02 from the system during algal photosynthesis could provide the final stimulus for this process. (ii) Nature of the CaC03. Three features of the Plectonema skeletons described here are 1, calcite mineralogy; 2, micrite-grade crystal size; and 3, incorporation of magnesium. Calcite formation is inhibited in sea-water by abundance of magnesium cations but it is the normal phase to be precipitated in fresh water. The Mg content of the calcite indicated by the AAS analyses is probably an over-estimate because it will include magnesium from the organic sheath material (see Gebelein and Hoff- man 1971) due to the acid preparation method used. Even so it is only low in the high-magnesian calcite range and can be seen as a reflection of the magnesium content of the pool-water. The small size of the calcite crystals is probably due to a rapid crystallization rate. Thus, the mineralogy, composition, and texture of the CaC03 in these algal sheaths are consistent with their meteoric, surficial environ- ment of formation (see Folk 1974, fig. 5). This conformity with normal physico- chemical controls of carbonate mineral formation is another indication that biological involvement in the precipitation process could be limited. There are, however, two significant features of blue-green calcification which raise doubts about a simple physicochemical mechanism. (i) Calcification in blue-greens is variable but it does not appear to be wholly controlled by the environment since certain species seem to be more predisposed to calcification than others. This specificity for CaC03 deposition is suggested by studies reported by Golubic (1973, p. 436) and is supported by the mutual occurrence of calcified and uncalcified blue-green species. (ii) If Girvanella were primarily dependent upon cementation processes for its skeletogenesis then its wall thickness could be expected to vary greatly according to the degree of precipitation proceeding in the micro-environment of a particular specimen. It should be capable of greatly exceeding the 50% of the tube diameter characteristic of the genus. Furthermore its tubiform morphology should be modified, in some cases at least, by infilling by a cement core as in Ostreobium reported by Schroeder and others. Yet thick microsparite- walled Girvanella- type fossils and similar rod-shaped forms are conspicuous by their absence from the geological record of the past 600 ma. This morphological discontinuity between Plectonema/ Girvanella tubes (impregnated sheaths) and encrusted sheaths suggests that the former are produced by blue-green algae which have the capacity to control calcifica- tion and are not simply dependent upon an external physicochemical process. Accordingly it is reasonable to postulate a distinction not only between the degree of calcification, i.e. between impregnated and encrusted sheaths, but also a corre- sponding distinction in the control (and timing) of calcification, i.e. between essen- tially controlled and uncontrolled calcification in blue-green algae. In forms such as RIDING: CALCIFIED PLECTONEMA FROM ALDABRA 43 Girvanella and Ortonella and their Recent analogues I suggest that the alga exerts a controlling influence over calcification and promotes crystal deposition in the sheath but not beyond it. The skeleton is thus a relatively thin-walled tube formed during the life of the alga. In other calcified forms known from freshwater tufa deposits and from Recent environments of subsea cementation calcification is beyond (or extends beyond) the control of the plant and results in gratuitous deposi- tion of CaC03 either during life or post-mortem. These skeletons are relatively thick-walled tubes or rods. Thus, while several features of CaC03 deposition in blue-greens are consistent with an essentially physicochemical control photosynthetically triggered by an otherwise passive alga, the morphology of Girvanella and its Recent analogues together with the mutual occurrence of calcified and uncalcified forms support the concept of specificity for, and control of, calcification in some blue-greens. The degree of control is clearly not complete since a species which calcifies in one habitat will not necessarily calcify elsewhere. Yet it is of interest that even the relatively unsophisticated blue-greens could have a control over their calcification which might be as distinctive, if less refined, as those operating in green, yellow-green, and red algae. But the nature of this calcification mechanism in blue-greens has still to be elucidated. Taxonomic significance. The specimens of Plectonema from Aldabra and those of Scytonema described by Monty (1967) provide Recent analogues of Girvanella and Ortonella respectively. Is Girvanella really Plectonema , or the other way round? The answer is neither. The amount of morphological detail preserved by Girvanella and Ortonella is so limited (text-fig. 1 ) that it could relate to a wide variety of extant filamentous blue-greens (Riding 1975). It is likely that both Girvanella and Ortonella represent a number of Recent biological taxa, of which Plectonema and Scytonema are only two. Thus, Plectonema is represented by the entire alga, trichomes as well as sheath. But its calcified sheath when the soft parts have been removed is Girvanella. Girvanella , on the other hand, is a calcified blue-green algal sheath devoid of soft parts. It is only distinguished from other calcified sheaths by its generally unbranched, sinuous morphology, constant diameter, and prostrate habit. In contrast, Ortonella is erect with subparallel filaments and dichotomously branched (text-fig. 2). This distinction between Girvanella and Ortonella is very crude when compared with the probable multiplicity of biological taxa merged within them, but it is possible that some environmental information can be gleaned from their gross differences in external morphology to compensate somewhat for the loss of detail in soft-part structure. Rather than merely representing a variety of taxa they might represent groups of morphologically similar ecophenes (environmental growth forms) of a variety of taxa. The distinction between erect and prostrate form seen in Ortonella and Girvanella extends to a number of porostromate genera and was recognized by Pia (1927, pp. 37-40) when he subdivided the Porostromata into the Agathidia (prostrate forms such as Girvanella and Sphaeroc odium) and Tham- nidia (erect forms such as Hedstromia , Mitcheldeania , and Ortonella). The erect, tufted habit of the Thamnidia suggests phototropic growth possibly induced by physical confinement, while the prostrate habit of the Agathidia suggests higher 44 PALAEONTOLOGY, VOLUME 20 light availability. But caution should be exercised in these interpretations and a genetic, as opposed to environmental, control of these forms cannot be excluded (B. A. Whitton, pers. comm.). Pia intended the Porostromata to contain calcareous tubiform fossils which he believed to be blue-green algae but which he was unable to place in extant groups. In practice the affinities of genera normally included in the Porostromata have been in doubt and attempts to use the group as a ‘Section’ of the Schizophyta (bacteria and blue-green algae) have not been very successful (see Johnson 1961, p. 194). Instead the Porostromata have been most useful not as a group of fossil calcareous blue-green algae but as a group of small tubiform fossils. In my view this is still the case and will remain so until the affinities of more members of the group can be clarified. It would be premature to assume that because there is now good evidence that two genera, Girvanella and Ortonel/a , are the calcified sheaths of filamentous blue-greens that this confirms Pia’s view of the group as a whole. It is still possible that calcified green algae are represented by some of the porostromate genera, especially those with relatively large tubes, and an apparent transition from Poro- stromata with closely appressed tubes to members of the Solenoporaceae remains to be clarified. Neither would it be advantageous at present to remove Girvanella and Ortone/la from the Porostromata and place them in extant blue-green families because of the probability that they represent forms from several families. The members of the Porostromata form a reasonably homogeneous morphological group characterized by their microscopic, calcareous, non-septate, tubiform skele- tons. It should be retained as a group of algae rather than blue-green algae. The following genera can be included: Bevocastria Garwood, Cayeuxia Frollo, Girva- nella, Hedstromia Rothpletz, Mitcheldeania Wethered /Garwoodia Wood, Ortonel/a, Spliaeroc odium Rothpletz. Epiphyton Bornemann, originally included by Pia, is not tubiform, and Ottonosia Twenhofel and Somphospongia Beede, included in the group by Johnson (1961, pp. 195-196) are aggregates of porostromate tubes, not tubiform fossils themselves. CONCLUSIONS 1. Calcified P. gloeophilum in freshwater pools on Aldabra is a Recent analogue of Girvanella. It is a ‘living fossil’ in the sense that Girvanella , while known to range from Cambrian to Cretaceous, was previously unrecorded from Cenozoic rocks. This discovery confirms previous suggestions that Girvanella represents the calcified sheaths of filamentous blue-green algae (Pollock 1918; Riding 1972, 1975) rather than being a calcified green alga (Fremy and Dangeard 1935; Johnson 1963, p. 26). 2. In fresh water and marine environments where CaC03 cementation is common, small algal filaments can become thickly encrusted by microspar. Limitation of calcification to the sheath in Plectonema and Girvanella produces a quite different, consistently thin-walled, tubiform micritic skeleton suggesting that precipitation of the crystals does not proceed uncontrolled in this case. This supports the idea that certain blue-greens have a specific capacity for calcification. However, the fact that Plectonema is not always calcified indicates that control of calcification by the alga RIDING: CALCIFIED PLECTONEMA FROM ALDABRA 45 is partial, rather than complete, and is presumably dependent on some overriding environmental factors. 3. The morphological discontinuity between calcified sheaths and encrusted filaments, together with differences in their environment of occurrence, support a distinction between relatively controlled and relatively uncontrolled calcification in blue-greens. Controlled calcification can proceed in a variety of carbonate environ- ments, not only those where cementation and tufa-formation are general, and occurs in living algae. Micrite crystals impregnate the sheath but do not significantly encrust its external surface. In this way the size and form of the sheath are preserved. Uncon- trolled calcification only takes place in environments of cementation and tufa- formation and can occur during life or post-mortem. Filaments are encrusted by microspar or larger crystals to form a thick-walled tube which only grossly reflects the original morphology of the alga. The interior of the tube may be infilled to produce a solid rod of CaC03. Controlled and uncontrolled calcification seem to reflect contrasting mechanisms of CaCO, deposition, the former essentially biochemical, the latter essentially physicochemical. 4. Girvanella probably includes various filamentous blue-green algae and not solely fossil Plectonema. Consequently, the Porostromata remains the most suitable higher taxon in which Girvanella can be placed at present. The Porostromata includes some blue-green genera, but also others whose precise affinities remain uncertain. It should be regarded as primarily a morphological group of algae. Acknowledgements. I am very grateful to Alan Donaldson and Brian Whitton, University of Durham, for providing the specimens of Plectonema and reading the manuscript. Their work at the Royal Society Aldabra Research Station was sponsored by the Society and the Natural Environment Research Council. Analytical facilities together with funds for electron microscopy were provided by the Department of Geology, University of Newcastle upon Tyne. P. Oakley and L. Rhodes performed the AAS and XRD analyses respectively. Brenda Arnold operated the SEM. The Royal Society Aldabra Publications Group made helpful comments. REFERENCES arnott, H. J. and pautard, F. G. E. 1970. Calcification in plants. In schraer, h. (ed.). Biological calcifica- tion: cellular and molecular aspects. Appleton-Century-Crofts, New York, pp. 375-446. black, M. 1933. The algal sediments of Andros Island, Bahamas. Phil. Trans. R. Soc. (B), 222, 165-192. donaldson, A. and whitton, B. a. 1976u. Chemistry of freshwater pools on Aldabra. Atoll Research Bull. In press. — 19766. The algal flora of freshwater pools on Aldabra. Ibid. In press. folk, r. l. 1974. The natural history of crystalline calcium carbonate: effect of magnesium content and salinity. J. sed. Petrol. 44, 40-53. fremy, p. and dangeard, L. 1935. Sur la position systematique des Girvanelles. Soc. Linn. Normandie , 8, 101 111. friedman, g. m., amiel, a. j. and schneidermann , N. 1974. Submarine cementation in reefs: examples from the Red Sea. J. sed. Petrol. 44, 816-825. garwood, e. j. 1914. Some new rock-building organisms from the Lower Carboniferous beds of West- morland. Geol. Mag. Dec. 6, 1, 265-271. gebelein, c. d. and hoffman, p. 1971. Algal origin of dolomite in interlaminated limestone-dolomite sedimentary rocks. In bricker, o. p. (ed.). Carbonate cements. Johns Hopkins Univ. Studies in Geology, 19, 319-326. 46 PALAEONTOLOGY, VOLUME 20 golubic, s. 1973. The relationship between blue-green algae and carbonate deposits. In carr, n. g. and whitton, B. A. (eds.). The biology of blue-green algae. Botan. Mono. 9, 434-472. Blackwell, Oxford. irion, g. and muller, G. 1968. Mineralogy, petrology and chemical composition of some calcareous tufa from the Schwabische Alb, Germany. In muller, g. and friedman, g. m. (eds.). Recent developments in carbonate sedimentology in central Europe. Springer-Yerlag, Berlin, pp. 157-171. james, N. p. 1972. Holocene and Pleistocene calcareous crust (caliche) profiles: criteria for subaerial exposure. J. sed. Petrol. 42, 817-836. Johnson, j. h. 1961. Limestone-building algae and algal limestones. Colorado School of Mines, Boulder. 297 pp., 139 pis. — 1963. Pennsylvanian and Permian algae. Colo. Sch. Min. Quart. 58, 211 pp., 81 pis. mckenzie, K. G. 1971. Entomostraca of Aldabra, with special reference to the genus Heterocypris (Crustacea, Ostracoda). Phil. Trans. R. Soc. (B), 260, 257-298. monty, c. l. v. 1967. Distribution and structure of Recent stromatolitic algal mats, eastern Andros Island, Bahamas. Ann. Soc. Geol. Belg. 90 (3), 55-100. NICHOLSON, H. a. and etheridge, J. R. 1878. A monograph of the Silurian fossils of the Girvan district in Ayrshire with special reference to those contained in the ‘ Gray collection', I, 1, 1-135. Blackwood, Edinburgh. pia, j. 1927. Thallophyta. /uhirmer, m. (ed.). Handbuch der Palaobotanik, 1, 31-136. Oldenbourg, Miinchen. pollock, J. b. 1918. Blue-green algae as agents in the deposition of marl in Michigan lakes. Ann. Rep. Michigan Acad. Sci. 20, 247-260. riding, R. 1972. Calcareous algae and some associated microfossils from Ancient Wall reef complex (Upper Devonian), Alberta. (Abs.). Bull. Am. Assoc. Petrol. Geol. 56, 648. 1975. Girvanella and other algae as depth indicators. Lethaia, 8, 173 179. schopf, j. w. 1968. Microflora of the Bitter Springs Formation, late Precambrian, central Australia. J. Paleont. 42, 651-688. schroeder, J. h. 1972. Calcified filaments of an endolithic alga in Recent Bermuda reefs. Neues Jb. Geol. Paldont. Mh. 1, 16-33. shinn, e. a., lloyd, r. m. and ginsburg, r. N. 1969. Anatomy of a modern carbonate tidal-flat, Andros Island, Bahamas. J. sed. Petrol. 39, 1202-1228. winland, h. d. and Matthews, R. K. 1974. Origin and significance of grapestone, Bahama islands. Ibid. 44, 921-927. Original typescript received 23 December 1975 Revised typescript received 5 April 1976 R. RIDING Department of Geology University College Cardiff CF1 1XL SIGNIFICANCE OF COILED PROTOCORALLA IN SOME MISSISSIPPI AN HORN CORALS by WILLIAM J. SANDO Abstract. Planispirally coiled protocoralla are described in Cyathaxonia tantilla (Miller) from the Lower Missis- sippi (Lower Carboniferous) of the western United States, the first record of this phenomenon in corals. Coiling is interpreted as a mode of attachment of young coralla to planktonic algae. The postulated pseudoplanktonic growth habit may be a significant factor in the widespread distribution of this species and other species of Cyathaxonia , which are generally found in rocks that record a bottom environment considered unfavourable for optimum coral growth. Specimens of horn corals with perfectly preserved tips are rarely observed in compact Palaeozoic limestones, even though corals are abundant in these rocks. Freeing of silicified specimens from limestone matrix by dissolving the matrix in acid often provides unusual opportunities to observe the earliest growth stages of these corals. This report presents the results of a study of approximately 500 etched silicified specimens of Cyathaxonia tantilla (Miller) from the Paine Member of the Lodgepole Limestone (Lower Mississippian, Kinderhookian, lower Tournaisian) collected from seven localities in the western United States (Utah and Montana). At Brazer Canyon, Utah, samples collected from closely spaced intervals through the lower part of the Lodgepole Limestone provided an unusual opportunity to study morphological variation in the earliest stage of these small horn corals. The Brazer Canyon material revealed a planispirally coiled protocorallum never before reported in corals. Discovery of the same phenomenon in specimens from four other localities suggests that a coiled brephic stage was a common feature of C. tantilla. NATURE AND SIGNIFICANCE OF THE BREPHIC STAGE Nature of samples. The material from Brazer Canyon was collected from 0 3 m intervals through the lower 14-4 m of the Lodgepole Limestone and from 15-6 m and 18-7 m above the base. Three hundred and twenty specimens of C. tantilla in which all or part of the brephic stage (0 6 septa) is preserved were isolated for analysis from the Brazer Canyon samples (Table 1 ). In the overwhelming majority of these specimens (288), no evidence of attachment was observed. In most specimens, the tip was broken off (PI. 12, fig. 7); in some, the tip is preserved, but abraded (PI. 12, fig. 4). Perfect specimens are in two categories: (1) coiled tip (25 specimens) (PI. 12, figs. 11-18), and (2) tip with lateral or basal attachment scar (7 specimens) (PI. 12, figs. 9, 10). Coiled tips are present throughout almost the full stratigraphic range of the species (Table 1). Perfect tips are similarly rare in samples of C. tantilla from other localities in Montana and Utah (Table 2). Thus, out of approximately 500 specimens studied, all or part of the brephic stage is preserved in 348, but [Palaeontology, Vol. 20, Part I, 1977, pp. 47-58, pi. 12.] 48 PALAEONTOLOGY, VOLUME 20 breakage or abrasion of the tip prevents determination of the nature of attachment in all but 38 specimens, of which 30 have coiled tips and 8 have lateral or basal scars of attachment. table 1. Nature of protocorallum in 320 specimens of Cyathaxonia tantilla (Miller) from Brazer Canyon section, Utah. Number of specimens with preserved brephic stage Metres above base of USGS Upper Palaeozoic Lateral No evidence Lodgepole Limestone locality number Coiled tips attachment of attachment Total 0-0-6 16801 -PC. 16802-PC 8 3 59 70 0-6- 1-2 16803-PC, 16804-PC 9 2 119 130 1-2-1-8 16805-PC 2 0 18 20 1 -8-2-4 16806-PC, 16807-PC 0 0 0 0 2-4-30 16808-PC, 16809-PC 1 0 11 12 30-3-6 16810-PC 0 0 2 2 3-6-4-2 1681 1-PC 1 0 2 3 4-2-4-8 16812-PC 0 0 1 1 4-8-5-4 16813-PC 0 0 2 2 5-4-60 16814-PC 0 0 0 0 60-6-6 16815-PC 0 0 5 5 6-6-7-2 16816-PC 1 0 1 2 7-2-7-8 16817-PC 0 0 1 1 7-8-8-4 16818-PC 0 0 5 5 8-4-90 16819-PC 0 0 0 0 90-9-6 16820- PC 0 0 3 3 9 6-10-2 16821-PC, 16822-PC 0 0 3 3 10 2-10 8 16823-PC, 16824-PC 0 0 1 1 10-8-11-4 16825-PC 0 1 4 5 1 1 4-12-0 16826- PC 1 0 3 4 12-0 12-6 16827-PC 0 1 27 28 12-6-13-2 16828-PC 2 0 14 16 13-2-13-8 16829-PC 0 0 6 6 13-8-14-4 16830-PC 0 0 0 0 15 6 1683 1-PC 0 0 1 1 18-7 16832-PC 0 0 0 0 Totals 25 7 288 320 table 2. Nature of protocorallum in specimens of Cyathaxonia tantilla (Miller) from localities in Montana and Utah other than Brazer Canyon. Number of specimens with preserved brephic stage USGS Upper Palaeozoic Lateral No evidence Locality locality number Coiled tips attachment of attachment Total Logan. Montana 17356-PC 1 0 1 2 17357-PC 0 0 4 4 White Peak, Montana 20164-PC 1 0 5 6 Squaw Creek, Montana 20600-PC 0 1 2 3 Sacajawea Peak, Montana 2064 1-PC 2 0 7 9 Emma Canyon, Utah 1691 1-PC 0 0 1 1 Baldy Mountain, Montana 1791 1-PC 1 0 1 2 17913-PC 0 0 0 0 179 14- PC 0 0 0 0 17915-PC 0 0 1 1 Totals 5 1 22 28 SANDO: PROTOCORALLA IN CORALS 49 Description of coiled tips. The brephic stage (protocorallum) is defined as the part of the corallum in which the first six septa (protosepta) appear. It is distinguished externally by the absence of the longitudinal septal grooves and interseptal ridges present in later growth stages (PI. 12, figs. 12, 13). In the thirty specimens with coiled tips, the protocorallum ranges from 0-9 to 2 0 mm in length and has a mean length of 1-3 mm. In the lower half of the protocorallum, the corallum is evolutely coiled, ordinarily planispirally, to form a single 360 volution. The diameter of the coil ranges from 0-5 to TO mm and has a mean value of 0-7 mm. The diameter of the corallum at the top of the volution ranges from 0-4 to 10 mm and has a mean value of 0-6 mm. The corallum is coiled about an open umbilicus which ranges from 0 05 to 0-30 mm in diameter and has a mean diameter of 0T3 mm. Above the coil, the protocorallum may be erect (no change in direction of the corallum axis), geniculate (one or more changes in direction of the corallum axis in one plane), or vermiform (one or more changes in direction of the corallum axis in more than one plane). In the specimens studied, 47% are erect, 23% are geniculate, and 30% are vermiform. One or more protosepta were observed within the coiled part of the corallum in seven specimens. Coiled tips are ordinarily observed in individuals that died in the neanic stage. The length of the corallum in specimens studied ranges from 1-9 to 14-2 mm, but the mean length is only 4-2 mm, and all but two specimens are 6-5 mm long or less. The maximum corallum diameter in these specimens ranges from T2 to 4-5 mm, but the mean diameter is 2-4 mm, and all but two specimens have a maximum dia- meter of 3-5 mm or less. The number of septa in the calice ranges from 6 to 34, but the mean is 22, and all but one specimen has 30 septa or less. Interpretation. Most small, solitary, curved-conical rugose corals observed by the writer show basal or lateral attachment scars in young specimens in which the protocorallum is preserved, indicating that these corals ordinarily were fixed to some stable object on the substrate immediately after settling of the larvae. Attach- ment scars are not ordinarily seen on adult specimens of these corals; this suggests that the corallum subsequently broke loose from its attachment when its weight became too great for the cemented tip to support, and then lay on its side, changing its growth direction in order to maintain a position conducive to optimum feeding and growth. Evidence and theories bearing on the growth habits of such corals have been summarized by Sando (1961, pp. 75-79). Lateral or basal attachment scars were observed in 8 of the 38 specimens of C. tantilla having perfectly preserved tips, indicating that some of these corals followed the common growth pattern. The presence of a coiled tip and the lack of attachment scars in the other thirty specimens suggests that coiling about some linear object was the principal mode of fixation of the protocorallum. In order to coil about an object lying on the substrate, the polyp would have had to exert a lifting force or to submerge its feeding mechanism in the substrate; the object of attachment was therefore probably above the sub- strate. The only linear objects of diameter comparable with that of the umbilicus of the coiled tip found in the coral samples are productoid brachiopod spines (0T- 0-2 mm diameter) and the crossbars in fenestellid bryozoan zoaria (010-2 mm dia- meter). These are rejected as objects of attachment because none of the coralla were D 50 PALAEONTOLOGY, VOLUME 20 found attached to them and because many of the coralla have an umbilicus of smaller diameter (0 05 mm) than the diameter of the postulated attachment objects. The absence of attachment objects within the umbilici of the coiled tips suggests that the objects were composed of material that was not preserved. The filaments of modern marine algae are in the size range required (W. H. Adey, pers. comm. 1975) , and marine algae are widely known from rocks of Mississippian age. How- ever, the only possible algal remains known from the Paine Member of the Lodge- pole Limestone are forms identified as Calcisphaera, Vicinisphaera , Bisphaera , and ‘ Radio sphaercC (Mamet in Sando et al. 1969, p. E-13), which are thought to be planktonic algal cysts (B. L. Mamet, pers. comm. 1975). The associated benthonic fauna of foraminifers, corals, brachiopods, gastropods, bryozoans, and pelmatozoans consists of sparse, poorly diversified and depauperate forms adapted to an impover- ished environment. The dark, thin-bedded, cherty, fine-grained, argillaceous, and silty beds of the Paine Member have been interpreted as sediments deposited in relatively deep, turbid water (Wilson 1969, pp. 15-16; Smith 1972, p. 31; Sando 1976) , and as much as 50 m of relief has been measured on associated bioherms (Cotter 1965, p. 888). Thus, the weight of evidence favours an aphotic bottom environment for the Paine Member, which would preclude attachment of Cyathaxonia tantilla to benthonic algae. EXPLANATION OF PLATE 12 All specimens are Cyathaxonia tantilla (Miller) from the Paine Member of the Lodgepole Limestone. All but fig. 5 are from Brazer Canyon, Utah. Fig. 1. Transverse thin section, ephebic stage, USNM 222532b, USGS loc. 16803-PC, x 12. Fig. 2. Longitudinal thin section, ephebic stage, USNM 222530, USGS loc. 16803-PC, x7. Figs. 3, 4. Counter and alar views, respectively, of etched ephebic corallum having abraded tip, USNM 222529, USGS loc. 16803-PC, x2. Fig. 5. Alar view, etched ephebic corallum having coiled tip, USNM 222591, USGS loc. 20641-PC, Sacajawea Peak, Montana, x 2. Fig. 6. Broken tip of etched brephic corallum showing 6 protosepta, USNM 222544, USGS loc. 16803-PC, x 10. Fig. 7. Broken tip of etched brephic corallum showing initial axial septum, USNM 222520, USGS loc. 16801-PC, x 10. Fig. 8. Broken etched ephebic corallum showing tuberculated septa, USNM 222545, USGS loc. 16804-PC, x 5. Figs. 9, 10. Alar views of etched neanic coralla showing lateral and basal attachment scars. 9, USNM 222524, USGS loc. 16802-PC, x 10. 10, USNM 222518, USGS loc. 16801-PC, x 10. Fig. 1 1 . Alar view of etched neanic corallum having coiled tip, USNM 222516, USGS loc. 16801-PC, x 10. Figs. 12, 13. Alar and cardinal views, respectively, of etched neanic corallum having coiled tip, USNM 222514, USGS loc. 16801-PC, x 10. Fig. 14. Enlarged alar view of protocorallum of specimen shown in figs. 12 and 13, USNM 222514, USGS loc. 16801-PC, x 20. Fig. 15. Cardinal view of etched neanic corallum having coiled tip, USNM 222517, USGS loc. 16801-PC, x 10. Figs. 16, 17. Alar and tip views, respectively, of etched neanic corallum having broken coiled tip showing axial septum, USNM 222536, USGS loc. 16803-PC, x 10. Figs. 18, 19. Alar and cardinal views, respectively, of etched neanic corallum having coiled tip, USNM 222515, USGS loc. 16801-PC, x 10. PLATE 12 * 7 - ■v> SANDO, Mississippian corals 52 PALAEONTOLOGY, VOLUME 20 The corals were probably attached to algae that were floating in the photic zone of the sea-water. The small size and consequent low mass of the young corallum is compatible with a planktonic habit. Prior to maturity, increased mass of the corallum may have caused most individuals to descend to the sea-floor, where they spent the remainder of their lives lying on their sides. These forms produced a curved corallum in response to the need for maintaining their oral surfaces in an optimum feeding position (PI. 12, fig. 14). Ordinarily the delicate coiled protocorallum was destroyed on the sea-floor either by other organisms (see Sando 1961, p. 79) or by mechanical abrasion while the corallum was still occupied by the living polyp or after death. Some individuals (PI. 1 2, fig. 5) either survived such destructive influences or remained attached to floating aggregates of algae throughout growth. The straight-conical rather than curved-conical form of the few mature specimens that show coiled tips suggests a pseudoplanktonic habit throughout growth for these individuals. Other individuals having basal or lateral attachment scars evidently were attached to objects on the substrate and were sessile throughout post-larval growth. Significance. The postulated pseudoplanktonic growth habit of the specimens studied has an important bearing on the migration potential of Cyathaxonia. Inasmuch as specimens with perfectly preserved tips are extremely rare, this growth habit may have been a common feature of C. tantilla and perhaps other species of the genus. C. tantilla occurs in strata of middle and late Kinderhookian and earliest Osagean age in both the western and eastern United States (text-fig. 1) despite a • Occurrence of Cythaxonia tantilla (Miller) text-fig. 1. Palaeogeographical map of the United States and adjacent areas during Kinderhookian time, showing occurrences of Cyathaxonia tantilla (Miller). SANDO: PROTOCORALLA IN CORALS 53 bottom environment unfavourable to optimum coral growth and the presence of the Transcontinental Arch which acted as a barrier between the two areas. Meagre data on the migration time of modern coral larvae suggests that most larvae become attached during the first 2 days of their existence, although some remain swimming for as long as 2 months (Connell 1 973, p. 209). A pseudoplanktonic brephic-neanic growth habit would increase the time interval available for migration and thus provide a better explanation of the distribution pattern of C. tantilla. Species of Cyathaxonia have a long stratigraphical range and world-wide distribu- tion despite their almost universal occurrence in dark, silty, argillaceous limestone and calcareous shale, lithofacies that reflect bottom conditions generally unfavour- able to the growth of corals. The prevalence of this association prompted Hill (1938, pp. 5-9) to name the biofacies ‘ Cyathaxonia Fauna’. A pseudoplanktonic growth habit in other species of Cyathaxonia might be a factor in the widespread distribution of the genus. SYSTEMATIC PALAEONTOLOGY Morphological terminology generally follows Hill (1935, 1956). The following abbreviations are used in the text: U.S. National Museum of Natural History, Washington, D.C., U.S.A. (USNM); U.S. Geo- logical Survey, Washington, D.C., U.S.A. (USGS). Family metriophyllidae Hill, 1939, emend. Rozkowska, 1969 Subfamily cyathaxoniinae Milne-Edwards and Haime, 1850 Genus cyathaxonia Michelin, 1847 1847 Cyathaxonia Michelin, p. 257. 1928 Cyathocarinia Soshkina, p. 376. Type species. Cyathaxonia cornu Michelin, 1 847, by subsequent designation of Milne-Edwards and Haime 1850. Lower Carboniferous (Tournaisian), Tournai, Belgium. Diagnosis. Corallum small, ceratoid-cylindrical. Long, contratingent minor septa inserted alternately with metasepta according to metriophyllid septal plan. Columella tall, developed independently of major septa but in contact with them (pseudoseptal columella of Schouppe and Stacul 1961). Tabulae inclined downward towards epitheca. Dissepiments absent. Sides of septa may have rows of tubercles steeply inclined from the horizontal. Discussion. Corals similar in most morphological aspects to the type species of Cyathaxonia have been found in strata ranging in age from Late Devonian (early Famennian) into Early Permian. Such morphologic similarity has resulted in recognition of the type species throughout the entire stratigraphic range of the genus, although subspecies and other species have been proposed for variations in the size and shape of the columella, septal number, size of the corallum, and length of the minor septa. Modifications of the septa have been variously called carinae, tubercles, or spines. Descriptions and illustrations of these modifications indicate that they consist of rows of rounded or pointed protuberances on the sides of the septa and not continuous ridges or flanges. Smith (1931, p. 8) pointed out that these are not true carinae. 54 PALAEONTOLOGY, VOLUME 20 Differences of opinion about the taxonomic significance of tuberculate v. non- tuberculate septa have resulted in different generic concepts. Carruthers (1913, p. 56) noted tubercles in the type species and in C. rushiana Vaughan from the Visean, but did not find them in topotype specimens of the type species (Tournai- sian) and did not consider the tubercles of diagnostic value. Schindewolf (1951, p. 99) found tuberculate septa in specimens of the type species from the Tournaisian of France and Visean of Ireland but not in topotypes of C. rushiana (Visean). He also noted all variations between forms having smooth septa and forms having well-developed tubercles and concluded that the presence or absence of tubercles has no taxonomic or chronologic significance. On the other hand, Soshkina (1928, p. 376) proposed the subgenus Cyathocarinia for Permian species of Cyathaxonia having tuberculate septa. Some subsequent authors (Soshkina et al. 1941, p. 41; Wang 1950, p. 205; Lecompte 1952, p. 482; Rozkowska 1969, p. 55) have continued to recognize Cyathocarinia as a subgenus of Cyathaxonia , whereas others (Soshkina 1939, p. 51; Kostic-Podgorska 1955, p. 170; Hill 1956, p. 264; Soshkina et al. 1962, p. 333; Ivanovskiy 1967, p. 23) regarded Cyathocarinia as a distinct genus. Extensive published records of Cyathaxonia indicate that both tuberculate and nontuberculate forms are present throughout the long stratigraphic range of the genus. Much variation has been noted in the strength and abundance of the tubercles in forms that are very similar or identical in other respects. In the Mississippian species described herein from the United States, tubercles are not present in the higher parts of the calice or in the earlier parts of the corallum, making recognition difficult in many specimens without serial sectioning. Available data on North American Mississippian species suggest that the presence or absence of tubercles is one of the morphologic characters useful for discrimination of species. However, use of this character for generic or subgeneric discrimination does not result in taxonomic groups that are meaningful phylogenetically. Accordingly, I include both tuberculate and nontuberculate forms in Cyathaxonia and regard Cyatho- carinia as a junior synonym. Cyathaxonia tantilla (Miller) Plate 12, figs. 1-18 1891 Zaphrentis tantilla Miller, p. 11, pi. 1, figs. 23, 24; Miller 1892, p. 621, pi. 1, figs. 23, 24; Keyes 1894, p. Ill; [not] Girty 1903, p. 269. 1909 Cyathaxonia tantilla (Miller), Weller, p. 270; Grove 1935, p. 367, pi. 9, figs. 15-17; Easton 1944, p. 30, pi. 6, figs. 7, 8; pi. 16, figs. 16, 17; Conkin and Conkin 1954, p. 214, fig. 1a-d. 1960 Cyathocarinia tantilla (Miller), Sando, p. 168. 1909 Cyathaxonia minor Weller, p. 270, pi. 10, figs. 14-17; [?] Girty 1926, p. 24; [not] Davis 1956, p. 29, pi. 4, fig. 3. 1960 Cyathaxonia arcuata Weller?, Sando, p. 168, pi. 16, figs. 25-27 . 1958 Cyathaxonia cordillerensis Easton, p. 13, pi. 1, figs. 14, 17, 18. Distribution. Chouteau Limestone (Missouri), Fern Glen Limestone (Missouri, Illinois), Springville Shale (Illinois), Hannibal Shale (Illinois), Compton Limestone (Missouri), Sedalia Limestone (Missouri), St. Joe Limestone Member of Boone Formation (Missouri, Arkansas), Reeds Spring Limestone (Mis- souri), Shale below Rockford Limestone (Indiana), [?] limestone of Boone age (Texas), Paine Member of Lodgepole Limestone (Montana, Utah), Represo Formation (Sonora). SANDO: PROTOCORALLA IN CORALS 55 Diagnosis. Cyathaxonia having 28-36 (ordinarily 32) septa in ephebic stage. Coral- lum attaining a maximum length of about 20 mm and a maximum diameter of about 4-5 mm. Septa tuberculate in late neanic and ephebic stages. Description of Utah and Montana specimens. Corallum ceratoid in brephic and neanic stages to cylindrical in ephebic stage (PI. 12, figs. 3-5); cardinal side convex; commonly geniculate, vermiform, or plani- spirally coiled in brephic stage (PI. 12, figs. 9-18); attaining a maximum length of 17 mm and a maximum diameter of 4-8 mm. Epitheca marked by strong longitudinal septal grooves and interseptal ridges, trans- verse rugae, and fine transverse striations (PI. 12, figs. 3, 4, 9, 14) except in brephic stage, where only transverse striations are present (PI. 12, figs. 12, 13). Calice as much as 6 mm deep at maturity. Septa number 32-36 at maturity (corallum diameter 4-5-4-8 mm). Ephebic septal plan (PI. 12, fig. 1) characteristic of the genus, consisting of cardinal septum in poorly defined fossula bounded by minor septa, major septa that reach columella but do not participate in its formation, counter septum on concave side of corallum, and minor septa that are contratingent on cardinal side of all major septa and fall slightly short of columella. Sides of septa usually have rows of tubercles that slope downward at an angle of 60°-80° from horizontal towards corallum axis (PI. 12, fig. 8); tubercles present only in late neanic and ephebic stages (below floor of calice and immediately above floor of calice) ; septa ordinarily smooth in other parts of corallum but may have low ridges corre- sponding to rows of tubercles in calice. Columella smooth sided; ovate in cross-section with long diameter in cardinal-counter plane; 0-3-0-8 mm in short diameter and 0-9- 1 - 5 mm in long diameter at base of mature calice; tapering to a blade-like point just below top of calice (PI. 12, figs. 5, 8). Tabulae (PI. 12, fig. 2) about 0 05 mm thick, concave upward, sloping upward from theca to columella; about five tabulae in vertical distance of 2 mm ; not ordinarily seen in etched specimens. Ontogeny. The earliest growth phase observed (PI. 12, figs. 7, 16) is in the broken tip of several etched specimens (corallum diameter 0-4 mm), which have a single axial septum, to be differentiated later into cardinal and counter septa. In the next observed growth phase (PI. 12, fig. 6), 6 protosepta (cardinal, counter, 2 alar, 2 counter lateral) are present at a corallum diameter of 0-7 mm. Development up to and including formation of the 6 protosepta is included in the brephic growth stage, which occupies the first 2 mm of corallum length to a corallum diameter of about 1 mm. The exterior of the corallum is without longitudinal ribbing in the brephic stage. The neanic stage is recognized for the part of the corallum that includes a complement of 7-31 septa to a corallum diameter of about 4 mm. During this stage, the corallum is characterized by maximum curvature and rapid expansion of its diameter. Forma- tion of the columella as a discrete structure is initiated and continues into maturity. Rapid insertion of major and minor septa takes place according to the ‘cyathaxoniid’ septal plan described by Faurot (1909, pp. 75-80), Duerden (1906, p. 236, figs. 9-12), and Hill (1940, p. 194). In the ephebic stage, the corallum is essentially cylindrical and straight, showing very little expansion in diameter (to a maximum of 4-8 mm) and the insertion of few septa (maximum 32-36). Discussion. North American species of Cyathaxonia have been differentiated mainly on the maximum size of the corallum, the maximum number of septa, the presence or absence of tubercles on the septa, and, to a lesser extent, on external ornamenta- tion. Although statistical analysis of larger samples may ultimately prove otherwise, available evidence does not support separation of the Utah and Montana samples from samples of C. tantilla from the Mississippi Valley area. Easton (1944, p. 30) found only 28 septa at a corallum diameter of 3-5 mm in the ‘most advanced stage observed’ in the cotype lot, but Miller (1891, p. 12) recorded a septal number of 56 PALAEONTOLOGY, VOLUME 20 30 at a diameter larger than 3 mm in a cotype, Keyes (1894, p. Ill) recorded a range of 20-32 septa in specimens from Missouri, and Conkin and Conkin (1954, p. 214) stated that the largest topotype specimens studied by them ‘constantly have 32 septa’ at corallum diameters of as much as 3-6 mm. The Utah and Montana speci- mens show a range of 30-36 septa at corallum diameters ranging from 3-5 to 4-8 mm in the largest coralla observed, but almost all the larger specimens have only 32 septa and a maximum diameter ranging from 3-5 to 4-5 mm (one specimen has 36 septa and one specimen has 34 septa). C. minor Weller (1909, p. 270) is within the range of measurable characters of C. tantilla , as noted by Grove (1935, p. 368) and affirmed by Easton (1944, p. 30). C. cordillerensis Easton (1958, p. 13), characterized by 32-35 septa and a maximum corallum diameter of 4 mm, also cannot be separated from C. tantilla. Sando (1960, p. 168) incorrectly referred specimens of C. tantilla questionably to C. arcuata Weller. A summary of the distinguishing characteristics of the described North American species of Cyathaxonia (Table 3) indicates that C. tantilla differs from all others in having tuberculate septa. Most of the other species have a considerably larger corallum than C. tantilla, with the exception of C. venusta, distinguished by its cuneate corallum, and C. winchelli , a poorly known form that may be a synonym of C. tantilla. Material studied. USNM 222513-222598. See list of localities for geographic locations and stratigraphic positions of collections. table 3. Summary of North American species of Cyathaxonia. Species Formation Locality Maximum septal number Maximum length (mm) Maximum diameter (mm) Other distinguishing features C. tantilla (Miller, 1891) See text See text 36 c. 20 c. 4-5 Tuberculate septa C. arcuata Weller, 1909 Fern Glen Ls. Missouri 36 36 7 C. winchelli Rowley, 1900 Lower Burlington Ls. ,, 36 16? 5? C. cynodon (Rafinesque and Clifford, 1820) New Providence Sh. Kentucky 36 25 7 Epithecal spines C. bordeni Greene, 1900 ,, ,, ,, Indiana 36 25 8 C. parva Greene, 1900 ,, ,, ,, ,, 50 30 10 C. venusta Greene, 1904 Salem Ls. 34 15 5 Cuneate corallum C.? iovaensis Worthern, 1890 Ste. Genevieve Ls. Iowa c. 30 32 13 LIST OF LOCALITIES 1. Brazer Canyon section, NWTSE.J Sec. 20, T. 11 N., R. 8 E., Rich County, Utah (see Sando et al. 1959, fig. 2 for geologic map). Paine Member of Lodgepole Limestone, lower 18-7 m. (USGS Iocs. 16801 -PC- 16832-PC) (see Table 1 for exact positions of individual samples). 2. Emma Canyon section, NETSE.J Sec. 17, T. 11 N., R. 8 E., Rich County, Utah (see Sando et al. 1959, fig. 2 for geologic map). Paine Member of Lodgepole Limestone, 1 -8-2-2 m above base (USGS loc. 16911 -PC). 3. Sacajawea Peak section, south wall of cirque in NE.£NWT Sec. 27, T. 2 N., R. 6 E., Gallatin County, Montana. Paine Member of Lodgepole Limestone, 6-5 m above base (USGS loc. 20641-PC). 4. Logan section, SE.^SW.i Sec. 25, T. 2 N., R. 2 E., Gallatin County, Montana (see Sando and Dutro 1974, pp. 4-8 for geologic map and section description). Paine Member of Lodgepole Limestone, lower I 5 m (USGS loc. 17356-PC), 3 m above base (USGS loc. 17357-PC). 5. Baldy Mountain section, SE.£NWT Sec. 27, T. 7 S., R. 3 W., Madison County, Montana. Paine Member of Lodgepole Limestone, 0-8-L4 m above base (USGS loc. 179 1 1 -PC), 6-2-7 -7 m above base SANDO: PROTOCORALLA IN CORALS 57 (USGS loc. 17913-PC), 9-2 10-7 m above base (USGS loc 17914-PC), 11-6-15-2 m above base (USGS loc. 17915-PC). 6. White Peak section, NE.^ Sec, 2, T. 11 S., R. 4 E., Gallatin County, Montana (see Witkind 1969, pp. 89-93 for section description). Paine Member of Lodgepole Limestone, 6-3—17- 1 m above base (USGS loc. 20164-PC). 7. Squaw Creek section, SE.^NE.^ Sec. 34, T. 4 S., R. 4 E., Gallatin County, Montana. Paine Member of Lodgepole Limestone, 2- 1 -2-6 m above base (USGS loc. 20600-PC). Acknowledgements. I am indebted to W. H. Adey, F. M. Bayer, A. H. Cheetham, Jerzy Fedorowski, and J. W. Wells for helpful discussions about these interesting little corals. I am also grateful to W. A. Oliver, Jun., and E. W. Bamber for critical review of the manuscript. The photographs are the work of R. H. McKinney and El. E. Mochizuki. REFERENCES carruthers, r. G. 1913. Lophophyllum and Cyathaxonia : Revision notes on two genera of Carboniferous corals. Geol. Mag., new ser., dec. 5, 10, 49-56. conkin, j. and conkin, b. 1954. Cyathaxonia from the Fern Glen Formation. Kansas Acad. Sci. Trans. 57,214-217. connell, j. h. 1973. Population ecology of reef-building corals. In jones, e. a. and endean, r. (eds.). Biology and geology of coral reefs , 2, Biology 1, Academic Press, New York and London, 205-245. cotter, e. 1965. Waulsortian-type carbonate banks in the Mississippian Lodgepole Formation of central Montana. J. Geol. 73, 881-888. davis, d. e. 1956. A taxonomic study of the Mississippian corals of central Utah. Brigham Young LJniv. Res. Studies, Geol. Ser. 3, 49 pp. duerden, j. e. 1906. The morphology of the Madreporaria— VIII. The primary septa of the Rugosa. Ann. Mag. Nat. Hist., ser. 7, 18, 226-242. easton, w. h. 1944. Corals from the Chouteau and related formations of the Mississippi Valley region. Illinois Geol. Surv. Rept. Inv. 97, 93 pp. — 1958. Mississippian corals from northwestern Sonora, Mexico. In easton, w. h., Sanders, j. e., knight, J. B. and miller, a. k. Mississippian fauna in northwestern Sonora, Mexico. Smithson. Misc. Colins , 119, 1-40. faurot, l. 1909. Affinites des tetracoralliaires et des hexacoralliaires. Ann. Paleont. 4, 69-108. girty, g. h. 1903. The Carboniferous formations and faunas of Colorado. U.S. Geol. Surv. Prof. Pap. 16, 546 pp. — 1926. The macro-fauna of the limestone of Boone Age. In roundy, p. v., girty, g. h. and Goldman, m. i. Mississippian faunas of San Saba County, Texas. Ibid. 146, 24-43. grove, b. h. 1935. Studies in Paleozoic corals. III. A revision of some Mississippian zaphrentids. Am. Midland Naturalist, 16, 337-378. hill, d. 1935. British terminology for rugose corals. Geol. Mag. 72, 481-519. — 1938. A monograph of the Carboniferous rugose corals ofScotland, pt 1. Palaeontogr. Soc. [Monogr.], 1-78. — 1939. Western Australian Devonian corals in the Wade Collection. J. R. Soc. West. Austral. 25, 141-151. 1940. A monograph of the Carboniferous rugose corals ofScotland, pt. 3. Palaeontogr. Soc. [Monogr], 115-204. — 1956. Rugosa. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology , Pt. F, Coelenterata. Geol. Soc. America and Kansas Univ. Press, F233-F324. Ivanovskiy, a. b. 1967. Etyudy o rannekamennougol’nykh rugozakh [Study of Early Carboniferous Rugosa]. Akad. Nauk SSR, Sibir. Otdei, Inst. Geol. Geofiz., Moskva, 91 pp. [In Russian.] keyes, c. R. 1894. Paleontology of Missouri, Pt. 1 . Missouri Geol. Surv. 4, 271 pp. kostic-podgorska, v. 1955. Donyokarbonski korali iz Paleozoika Reke Sane (Bosna) [Lower Carboni- ferous corals from Paleozoic of Sana River (Bosnia)]. Radova Geol. Inst. Beograd Zborn. 8, 169-179. [In Russian.] 58 PALAEONTOLOGY, VOLUME 20 lecompte, m. 1952. Madreporaires paleozoiques. In piveteau, j. Traite de paleontologie, Paris, 1, 419-538. michelin, H. 1847. Iconographie Zoophytologiques, description par localites et terrains des polypiers fossiles de France et pays environnants. Paris, 249-328. miller, s. a. 1891. Palaeontology. Indiana Geol. Survey Ann. Rept. 17, 103 pp. (advance printing). 1892. Palaeontology. Ibid. 61 1 705. milne-edwards, H. and haime, J. 1850. A monograph of the British fossil corals, pt. 1. Palaeontogr. Soc. [Monogr.], 1-71. rozkowska, m 1969. Famennian tetracoralloid and heterocoralloid fauna from the Holy Cross Moun- tains (Poland). Acta Pal. Pol 14, 6-187. sando, w. j. 1960 [1961]. Corals from well cores of Madison group, Williston basin. U.S. Geol. Surv. Bull. 1071-F, 157-190. — 1961. Morphology and ontogeny of Ankhelasma, a new Mississippian coral genus. J. Paleont., 35, 65-81. — 1976. Mississippian history of the northern Rocky Mountains region. U.S. Geol. Surv. J. Res. 4, 317-338. — and dutro, j. t., jun. 1974. Type sections of the Madison Group (Mississippian) and its subdivisions in Montana. U.S. Geol. Surv. Prof. Pap. 842, 22 pp. — and gere, w. c. 1959. Brazer dolomite (Mississippian), Randolph quadrangle, north-east Utah. Bull. Am. Assoc. Petrol. Geol. 43, 2741-2769. — mamet, b. l. and dutro, j. t., jun. 1969. Carboniferous megafaunal and microfaunal zonation in the northern Cordillera of the United States. U.S. Geol. Surv. Prof. Pap. 613-E, 29 pp. schindewolf, o. h. 1951. Uber ein neues Vorkommen unterkarbonischer Pericyclus- Schichten im Ober- harz. Neues Jb. Geol. Palaont. Abh. 93, 23-1 16. schouppe, a. von and stacul, p. 1961 Die Axialstruktur der Pterocorallia. Ibid. 112, 251 -280. smith, d. l. 1972. Depositional cycles of the Lodgepole Formation (Mississippian) in central Montana. Montana Geol. Soc. 21st Ann. Field Conf. Guidebook, 29-35. smith, s. 1931. Some Upper Carboniferous corals from South Wales. Geol. Survey G.B.: Summ. Progr. [for 1939], pt. 3, 1-13. soshkina, e. 1928. Nizhnepermskie korally zapadnogo sklona Severnogo Urala [Lower Permian corals from west slope of North Urals]. Moskov. Obshch. Ispyt. Prip. Byull., nov. ser., 36, 339-393. [In Russian.] — 1939. Tip Coelenterata [Phylum Coelenterata], In licharev, b. (ed.). Atlas rukovodyashchikh form iskopaemykh faun SSSR [Atlas of leading forms of the fossil fauna of the USSR]. Tsentral. Nauclmo- issled. Geol.-razved. Inst. 6, Permskaya Sistema, 50-58. [In Russian.] — Dobrolyubova, t. a. and kabakovich, N. v. 1962. Podklass Tetracoralla [Subclass Tetracoralla], In sokolov, b. s. (ed.). Osnovy paleontologii ; Gubki, Arkheotsiaty, Kishechnopolostnye, Chervi [ Funda- mentals of paleontology ; Porifera, Archeocyatha , Coelenterata , Vermes ]. Izdatel. Akad. Nauk SSSR, Moskva, 286-356. [In Russian.] — and porfir’ev, g. 1941. Permskie Rugosa Evropeyskoy chasti SSSR [Permian Rugosa of the European part of the USSR], Paleontologiya SSSR, 5, Akad. Nauk SSSR, Pal. Inst. Moskva, 304 pp. [In Russian.] wang, h. c. 1950. A revision of the Zoantharia Rugosa in the light of their minute skeletal structures. Phil. Trans. R. Soc. B„ 234, 175-246. weller, s. 1909. Kinderhook faunal studies V, the fauna of the Fern Glen Formation. Bull. Geol. Soc. Am. 20, 265-332. wilson, j. l. 1969. Microfacies and sedimentary structures in ‘deeper water’ lime mudstones. In friedman, G. m. (ed.). Depositional environments in carbonate rocks. Soc. Econ. Paleont. and Min. Spec. Pub/. 14, 4-19. witkind, i. j. 1969. Geology of the Tepee Creek quadrangle, Mon tana- Wyoming. U.S. Geol. Surv. Prof. Pap. 609, 101 pp. WILLIAM J. SANDO U.S. Geological Survey U.S. National Museum Washington, D.C. 20244 U.S. A. Original typescript received 9 December 1975 Revised typescript received 25 February 1976 COPROLITES CONTAINING PLANT MATERIAL FROM THE CARBONIFEROUS OF BRITAIN by ANDREW C. SCOTT Abstract. Coprolites containing plant material are described from the Middle Coal Measures (Westphalian B) of Swillington, near Leeds, West Yorkshire. Other possible coprolites from the Lower Carboniferous (Calciferous Sandstone Series) of the Loch Humphrey Burn and Glenarbuck localities in the Kilpatrick Hills, Strathclyde are also recorded. The Coal Measure specimens contain either lycopod megaspore fragments, indeterminate plant debris, or a large variety of microspores (attributable to Lycopsida, Sphenopsida, Pteropsida; Filicinae, Gymno- spermae, and Pteridospermae) whereas the Lower Carboniferous specimens consist mainly of rolled plant debris. It is suggested that these coprolites belonged to animal litter feeders and is direct evidence of animals eating vegetation in the Palaeozoic. After a recent symposium on animal/plant interrelationships in the Palaeozoic, Cox (1974) concluded that There seems little chance of a great variety of unequivocal evidence for the interaction between Palaeozoic plants and arthropods’. Although recent studies have shown that some interrelationships between animals and plants may have existed as early as the Devonian (Kevan et al. 1975), it remains true that very little direct evidence of animals feeding on plants has been brought forward. Leaves of Neuropteris with holes, presumably made by an herbivorous animal, have been reported from Coal Measure sediments (Van Ameroin 1966; Van Amerom and Boersma 1971) and similarly eaten leaves of Glossopteris have been described from the Karroo rocks of South Africa (Plumstead 1963). Borings, possibly made by insects, have been reported in Trigonocarpus seeds from the Carboniferous (Barnard, oral comm., see Cox 1974). Animal damage of plant axes (Kevan et al. 1975) and trigonotarbid arachnids within sporangia have been described from the Devonian Rhynie Chert (Rolfe in Kevan et al. 1975, pi. 56). Numerous examples of damaged stems have been reported from the Carboniferous ( Catamites , Seward 1898, Stopes 1907; Myeloxylon , Holden 1910; lepidodendroid axis, Wilkinson 1930) but it is uncertain whether these wounds were caused by natural accidents or whether they were the result of animal damage. A few specimens of bored Carboniferous wood have been figured (i.e. Sigi/laria , Geinitz 1855, pi. 8, fig. 1) and in one specimen borings have been shown to contain minute (< 100 /mn) coprolites attributed to arthropods (Williamson 1880, pi. 20, figs. 65, 66). Other occurrences of damaged wood have been reviewed by Moodie (1923, pp. 99-108). Unfortunately the morphology of Palaeozoic animals has not been of great help in the search for possible herbivores. The large Carboniferous ‘centipede’ Arthro- pleura, for example, was regarded as carnivorous on the basis of the nature of a supposed cephalic limb, but proved to have plant debris (?lycopod) in its gut (Rolfe and Ingham 1967). The record of coprolites containing plant material has been very scanty. The contents of most coprolites are generally unidentifiable (phosphatized) or else they [Palaeontology, Vol. 20, Part 1, 1977, pp. 59-68, pis. 13-14.] 60 PALAEONTOLOGY, VOLUME 20 contain animal remains and are generally attributed to fish (see occasional references in Hantzschel et al. 1968). Coprolites from coal balls have been figured by Mamay and Yochelson (1962) and those consist of partly macerated plant material, mostly epidermal (Mamay, pers. comm. 1976). Brief mention is made of coprolites con- taining plant material by Cox (1974) but no details are given. Other coprolites containing spores and pollen have been reported from post- Palaeozoic rocks. Harris (1957) described, from the Jurassic of Yorkshire, small coprolites containing Caytonia pollen and leaf cuticle of both Sagenopteris and Gingko. Bryant and Williams-Dean (1975) described subfossil human coprolites which contained plant material and pollen which have given information on both early human diet and local vegetation. The occurrence of coprolites containing plant remains from the Coal Measures of Yorkshire and the Calciferous Sandstone of the Kilpatrick Hills is further direct evidence that some animals were exploiting this food source in the Carboniferous. The predominance of plant material in the coprolites strongly suggests that they were produced by herbivores and are not the undigested waste products of plants accidentally eaten by carnivores. DESCRIPTIONS Coal Measures Twelve small coprolites were found in residues obtained by disaggregating shales in hydrofluoric acid. The shales (flood plain deposits) were thin partings in poor coal one metre below and immediately above the Lodgett Coal, Swillington Brick- works, near Leeds, West Yorkshire (SE 384314, Godwin and Calver 1975). These residues also contained abundant megaspores, cuticles, and fusain including an early conifer (Scott 1974). The coprolites are generally cylindrical in shape (up to 3 x 1 mm), although some have been flattened. They are of three main types: one containing mainly lycopod megaspore fragments, one containing a mixture of microspores, and one containing indeterminate plant debris. The megaspore fragments may be identified from the nature of their sculptural elements. One coprolite (PI. 13, figs. 1, 2) contained only fragments of the sigillarian megaspore Tuber culatisporites mammilarius (Bartlett) (PI. 13, figs. 7, 1 1) whereas another (PI. 13, figs. 3, 9, 14) contained mainly fragments of lepidodendroid megaspores such as Lagenicula subpi/osa (Ibrahim) (PI. 13, figs. 8, 12). Occasional Lycospora , other unidentifiable microspores, and cuticle are also found. Those coprolites containing microspores are more irregular in shape, often considerably flattened probably because of the greater compressibility of the more finely divided plant material. This type of coprolite has a wide range of composi- tion, some containing an abundance of two types of spore (Table 1, specimens 10 and 12) and others a large number of types (Table 1, specimens 6 and 8, PI. 1 3, figs. 5 and 6, PI. 14, figs. 1-9). (Numerous sporangia are found in the same residues but are recognized by containing only a single species of spore, layering, and by the undamaged nature of the spores, PI. 13, figs. 4, 10.) Many of the microspores are exceptionally well preserved although many are also broken. The spores which are SCOTT: CARBONIFEROUS COPROLITES 61 table 1. The Composition of Coprolites from the Westphalian B of Yorkshire. The specimens have been lodged in the Hunterian Museum, Glasgow (FSC 2061-2072). Specimen no. LYCOPSID A — MEGASPORES Tuber culatisporites mammilarius (Bartlett) Lagenicula subpilosa (Ibrahim) Indeterminate LYCOPSIDA— MICROSPORES Lycospora (Schopf, Wilson and Bentall) Densosporites (Berry) cf. Crassispora ( Bharadwaj ) ICristatisporites (Potonie and Krenrp) SPHENOPSIDA — MICROSPORES Calamospora Schopf, Wilson and Bentall ILaevigatosporites Ibrahim PTEROPSIDA — MICROSPORES FILICINAE cf. Cyclogranisporites Potonie and Kremp Raistrickia (Schopf, Wilson and Bentall) cf. Verrucosisporites (Ibrahim) cf. Savitrisporites Bharadwaj GYMNOSPERMAE IFlorinites Schopf, Wilson and Bentall PTERIDOSPERMAE Schopfipollenites Potonie and Kremp UNKNOWN AFFINITY cf. Ahrensisporites Potonie and Kremp cf. Pustulatisporites Potonie and Kremp Indeterminate OTHER PLANT MATERIAL Indeterminate cuticle Indeterminate wood Indeterminate plant material 1 2 3 4 5 6 7 8 9 10 11 12 A C c A R R R O O O R A O C ?R O ?A 70 A ?R O R C C ?C O O C ?A O R O O ?A C O AC AOOOARA R = Rare, O = Occasional, C = Common, A = Abundant. recorded represent both a wide variety of morphologies and a number of diverse plant groups including lycopods ( Lycospora and Densosporites , PI. 13, fig. 5), ferns ( Raistrickia , PI. 14, fig. 8), pteridosperms ( Schopfipollenites , PI. 14, fig. 7), and sphenopsids ( Calamospora ) (Potonie 1962, 1965; Potonie and Kremp 1954). It is concluded that these pellets are genuine coprolites, and not burrow fills, peat pellets, or sporangia, because of the repetition of constant shape and the hetero- geneity of plant material. From the variety of plant material in the coprolites it appears that the animals concerned were more likely to have been litter feeders rather than direct cone or leaf feeders. There is no conclusive evidence that these coprolites all belong to the same type of animal, but from the general size and shape this seems a strong possibility. There are a number of different kinds of animal, mainly arthropod, which may have been eating the plant litter and produced these coprolites. Although cuticle of 62 PALAEONTOLOGY, VOLUME 20 arachnid type (PI. 14, fig. 14, L. J. Wills and I. Strachan, pers. comm. 1975) has been found in this deposit there is no evidence to suggest that this was the animal responsible. Lower Carboniferous, Calciferous Sandstone Series (i) Loch Humphrey Burn This material, from the Walton Collection in the Hunterian Museum, Glasgow (prefix Pb to numbers) comes from the Loch Humphrey Burn plant bed (Smith 1964) and is preserved as compression fossils. These ‘coprolites’ (PI. 14, fig. 13) are fairly common and range in length from 20 to 26 mm (e.g. Pb 3313, 3314a, b, 2507, 2516) and when found unsquashed are approximately 7 mm in diameter (PI. 14, fig. 12, Pb 3314c). The flattened cylinders generally show four longitudinal ridges as seen on a cleaved bedding surface (PI. 14, fig. 13). In cross-section they show twelve prominent ridges (PI. 14, fig. 12). It is not fully understood how in the same bed most of these ‘coprolites’ are flattened whilst one remains unsquashed. There is no way to prove that the two types of fossils are related, short of cleaving open the unsquashed specimen to look at the longitudinal view, but from the nature of the shape, a cylinder with prominent ridges, it would seem likely that they are similar objects. It will need, therefore, further specimens before the nature of preservation is fully understood. When macerated these cylinders yield unidentifiable plant debris. They are thought not to be either burrow fills or peat pellets because of their constant shape, but there remains the possibility that they may be some poorly preserved plant organ rather than a coprolite, although the latter is favoured. There are no data concerning the animals that might have produced these ?coprolites. EXPLANATION OF PLATE 13 Westphalian B coprolites and associated plant fragments from shaly partings below and above the Lidgett Coal, Swillmgton, West Yorkshire. All pictures were taken with a Cambridge S600 Scanning Electron Microscope (S.E.M.), the specimens having been coated with gold in a Polaron Sputter-Coating Unit E5000. Figs. 1-3, 9, 14. Coprolites containing megaspore fragments. 1, whole coprolite. Table 1, specimen 1, x 40 (FSC 2061). 2, fragment of megaspore Tuberculatisporites mammilarius (Bartlett) from the same specimen, x 150 (FSC 2061). 3, whole coprolite. Table 1, specimen 7, x40 (FSC 2067). 9, detail of megaspore fragment of Lagenicula subpilosa (Ibrahim) from the same specimen, x250 (FSC 2067). 14, megaspore fragments, microspores, and plant debris from the same specimen, x250 (FSC 2067). Figs. 4, 10. Dispersed sporangium from the same horizon. 4, detail of spores, Calamospora Schopf, Wilson and Bentall, x 160 (FSC 2073). 10, whole sporangium, x 15 (FSC 2073). Figs. 5, 6, 13. Coprolites containing microspores and plant debris. 5, whole doprolite, Table 1, specimen 3, x 15 (FSC 2063). 6, detail of same specimen with ICristatisporites (Potonie and Kremp), x 600 (FSC 2063). 13, detail of coprolite. Table 1, specimen 9, with abundant plant debris, x225 (FSC 2069). Figs. 7, 8, 11, 12. Megaspores from the same horizon. 11, Tuberculatisporites mammilarius , x20 (FSC 2076). 7, detail of distal face, x 150 (FSC 2074). 12, Lagenicula subpilosa, x 25 (FSC 2075). 8, detail of spines, x 275 (FSC 2075). All specimens have been lodged in the Geology Collection of the Hunterian Museum, Glasgow. PLATE 13 SCOTT, plant debris in Carboniferous coprolites 64 PALAEONTOLOGY, VOLUME 20 (ii) Glenarbuck Three slides, in the Walton Collection, Hunterian Museum, Glasgow (FSC 779- 781), show cross-sections of what have been interpreted (on Walton’s slide label) as coprolites. These have a circular, although slightly irregular cross-section (PI. 14, fig. 10) with a diameter of 5 mm (FSC 781 has a diameter of 1 mm but this may be because only the tapered end of the coprolite was sectioned). The three slides are probably peels from the same specimen. The plant material inside the ?coprolite is layered (PI. 14, fig. 11) consisting mainly of stelar and woody elements and other unidentifiable plant material. For these specimens, however, there is less evidence that these are coprolites; they may be simply pellets of petrified peat. No animal remains have been found in association with these ?coprolites. DISCUSSION Which animals produced these coprolites and which Palaeozoic animals were herbi- vores is a matter of speculation. Numerous types of insects were present in the Coal Measures, some of which may have been herbivores (phytophagous). It has been suggested that other arthropod groups such as the Collembola and Arachnida may have been litter feeders, as they are today (Tillyard 1931). The Collembola are also thought to have been spore feeders (Smart and Hughes 1973) as have some of the trigonotarbid arachnids already cited. Kevan et al. (1975) note that spores are often one of the alternative food sources ‘adopted’ by carnivorous (zoophagous) arthropods if animal food is lacking. Of the insects, especially ground-dwelling species, cockroaches (Dictyoptera) were present in the Carboniferous (Muller 1963). I have examined the faecal pellets from an extant genus of the phytophagous cockroach, Blaberus, which is 6-7 cm EXPLANATION OF PLATE 14 Figs. 1-9. Westphalian B coprolites from shaly partings below and above the Lidgett Coal, Swillington, West Yorkshire. (S.E.M. as PI. 13.) These specimens (including fig. 14) have been lodged in the Geology Collection of the Hunterian Museum, Glasgow. 1, whole coprolite. Table 1, specimen 6, x25 (FSC 2066). 2, detail of same specimen with a triangular microspore with conate ornament, ? Pustulatisporites Potonie and Kremp, 600 (FSC 2066). 4, whole coprolite. Table 1, specimen 8, x 15 (FSC 2068). 3, detail of the same specimen with ICyclogranisporites Potonie and Kremp, x 450 (FSC 2068). 5, detail of the same specimen with Lycospora (Schopf, Wilson and Bentall) and IDenosporites (Berry), x 600 (FSC 2068). 6, detail of the same specimen with numerous microspores including ISavitrisporites Bharadwaj and lAhrensisporites Potonie and Kremp, x 300 (FSC 2068). 7, detail of the same specimen with Schopfipollemtes Potonie and Kremp, x250 (FSC 2068). 8, detail of the same specimen with Raistrickia (Schopf, Wilson and Bentall), x 600 (FSC 2068). 9, detail of the same specimen with micro- spores including lAhrensisporites Potonie and Kremp, X 500 (FSC 2068). Specimens shown in figs. 10-13 are from the Walton Collection in the Hunterian Museum, Glasgow. Figs. 10, II. Cross-sections of ?coprolites from the Calciferous Sandstone Series of Glenarbuck. 10, whole cross-section, x 4 (FSC 780). 11, detail showing plant debris and stelar fragments, x 30 (FSC 780). Figs. 12, 13. ?Coprolites from the Calciferous Sandstone Series, Loch Humphrey Burn. 12, cross-section, x 4 (Pb 3314c). 13, longitudinal cleavage compression, x 2 (Pb 2576). Fig. 14. Animal cuticle (?Arachnid) from the shales below the Lidgett Coal, Swillington, West Yorkshire, x8 (A 2618). PLATE 14 SCOTT, plant debris in Carboniferous caprolites 66 PALAEONTOLOGY, VOLUME 20 long. These are small, 3 mm x 1 mm, roughly cylindrical, and occasionally show longitudinal striations. Smaller species would presumably produce smaller faecal pellets (perhaps comparable in size and shape to the Coal Measure specimens). Smart and Hughes (1973) suggest that the ’probosces of Palaeodictyoptera and Megasectoptera could have been used as a probe to work over cones and capsules of the plants of that time (Carboniferous) for spores and pollen and perhaps especi- ally for megaspores . . .’ which Kevan et al. (1975) observe must have represented a particularly nutritious food source for insects. The eating of spores by arthropods, it has been suggested, is likely to have contributed to spore dispersal and even when the spores have passed through the animals’ gut some of them may still have remained viable (Chaloner 1976). Other arthropods which are thought to have been phytophagous include the myriapods (Kevan et al. 1975; Rolfe and Ingham 1967). These include the mille- pedes which have Palaeozoic representatives (Hoffman 1969) and have been shown in some figures to be eating plant material including Lepidodendron (Hoffman 1969). I have examined the faecal pellets of an extant African millepede, 9 cm long. These consist of plant debris and measure 6 mm x 3 mm. The size of the faecal pellets, however, is not directly related to the length of the animal but to a combination of body weight and the quantity of food eaten at any one time (Edwards 1974). Arthro- pleura , a member of an extant group of myriapods, has been shown to be phyto- phagous. A juvenile specimen with lycopod fragments in its gut (Rolfe and Ingham 1967) might have produced faecal pellets about 4 mm in diameter (measured from PI. 13, fig. 8). The adult animal which grew to 1-5 m long must have produced larger faecal pellets, perhaps the size of the Lower Carboniferous specimens, although the genus Arthropleura is only known from Upper Carboniferous deposits (Rolfe 1969). Although most groups of extant arachnids are zoophagous some mites (Acarida) are phytophagous (Wallwork 1967). Most, however, are very small (less than 1 mm, Harding and Stuttard 1974). Some Palaeozoic forms are found up to 8 mm in length (Petrunkevitch 1955) and may have produced faecal pellets in the order of 1 mm long. Few early vertebrates are thought to have been herbivores although the Upper Palaeozoic amphibian Scincosaurus which has spathulate teeth may be such an animal (A. Girvan, pers. comm.). CONCLUSIONS Considering the quantity of plant material available for exploitation as a food source, in the Carboniferous, it is surprising that previously very little direct evidence of herbivores has been found as ultimately all land fauna is dependent on land vegeta- tion, forming the basis of an extensive food chain. • The occurrence of coprolites containing spores and plant tissue is direct evidence that at least some animals were exploiting this food source in the Carboniferous. Acknowledgements. I would like to thank Mr. D. W. Brett for bringing the Lower Carboniferous material to my attention, G. Armitage and Sons for permission to visit their Swillington Quarry, Dr. W. D. I. Rolfe for the loan of specimens in his care, and Dr. P. Banham and Professor W. G. Chaloner for their helpful comments. This work was carried out during the tenure of an N.E.R.C. Studentship which is gratefully acknowledged. SCOTT: CARBONIFEROUS COPROLITES 67 REFERENCES bryant, v. m., jun. and williams-dean, G. 1975. The Coprolites of Man. Scient. Am. 232, 100-109. chaloner, w. G. 1976. The evolution of the adaptive features in fossil exines. In ferguson, i. k. and muller, j. (eds.). Evolutionary Significance of the Exine. Academic Press, London, 1-14. cox, b. 1974. Little evidence for Palaeozoic arthropod and plant interaction. Report of Linn. Soc. meet- ing on the Interrelationships of Palaeozoic terrestrial arthropods and plants. Nature , Lond. 249, 615— 616. edwards, c. a. 1974. Macroarthropods. In dickinson, c. h. and pugh, g. j. f. (eds.). Biology of Plant litter decomposition. Academic Press, London, Vol. 2, 533-554. geinitz, H. B. 1855. Die Versteinerungen der steinkohlenformation sachsen. Leipzig, 61 pp. Godwin, c. G. and calver, m. a. 1975. A review of the Coal Measures (Westphalian) of Leeds. J. Earth Sci. Leeds , 8 (1974), 409-432. hantzschel, w., el-baz, f. and amstutz, g. c. 1968. Coprolites, an annotated bibliography. Mem. Geol. Soc. Am. 108, 1-132. harding, d. j. l. and stuttard, r. a. 1974. Microarthropods. In dickinson, c. h. and pugh, g. j. f. (eds.). Op. cit., Vol. 2, 489-532. Harris, T. M. 1957. How we study fossil plants— Caytonia. New Biol. 22, 24-38. hoffman, R. L. 1969. Myriapoda, exclusive of Insecta. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology. Pt. R, Arthropoda 4. Lawrence, Kansas, 572-606. holden, h. s. 1910. Note on a Wounded Myeloxylon. New Phytol. 9, 253-257. kevan, p. G., chaloner, w. g. and savile, d. B. o. 1975. Interrelationships of early terrestrial arthropods and plants. Palaeontology , 18, 391-417. mamay, s. h. and yochelson, e. l. 1962. Occurrence and significance of marine animal remains in American coal balls. Prof. Pap. U.S. Geol. Surv. 354-1, 193-224. moodie, r. L. 1923. Palaeopathology : An Introduction to the study of ancient evidences of disease. Univ. of Illinois Press, 567 pp. muller, A. H. 1963. Lehrbuch der Palaozoologie. Vol. 3 Arthropoda 2. Veb. Gustav. Fischer Verlag. Jena, 1-257. petrunkevitch, A. 1955. Arachnida. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology. Pt. P. Arthropoda 2. Lawrence, Kansas, 42-162. plumstead, e. p. 1963. The influence of plants and environment on the developing animal life in Karroo times. S. Afr. J. Sci. 59, 147-152. potonie, R. 1962. Synopsis der sporae in situ. Beih. Geol. Jahrb. 52, 204 pp. 1965. Fossile sporae in situ. Forsch. d. land Nordrhein-Westf. 1483, 1-74. — and kremp, g. 1954. Die Gattungen der Palaozoichen sporae dispersae und lhre stratigraphie. Geol. Jb. 69, 111 194. rolfe, w. d. i. 1969. Arthropleurida. In moore, r. c. (ed.). Op. cit., 607-620. — and ingham, J. k. 1967. Limb structure, affinity and diet of the Carboniferous ‘centipede’ Arthro- pleura. Scot. J. Geol. 3, 1 18-124. scott, a. c. 1974. The earliest conifer. Nature, Lond. 251, 707-708. seward, A. c. 1898. Fossil Plants. Cambridge, Vol. I, 452 pp. smart, j. and hughes, n. f. 1973. The insect and the plant: progressive palaeoecological integration. In van emden, F. (ed.). Insect /Plant Relationships, Symp. Roy. Ent. Soc. Lond. 6, 143-155. smith, d. l. 1964. Two Scottish Lower Carboniferous floras. Trans, proc. bot. Soc. Edinb. 39, 460-466. stopes, M. c. 1907. A note on wounded Calamites. Annals Bot. 21, 277-280. tillyard, r. j. 1931. The evolution of the class Insecta. Pap. Proc. R. Soc. Tasm. (1930), 1-89. van amerom, H. w. J. 1966. Phagophytichnus ekowskii nov. Ichnogen & Ichnosp. eine Missbildung infolge von Insecktenfrass, ausdem Spanischen Stephanien (Provinz Leon). Leid. geol. Medecl. 38, 181-184. — and boersma, m. 1971. A new find of the Ichnofossil Phagophytichnus ekowskii Van Amerom. Geologie. Mijnb. 50, 667-670. wallwork, j. a. 1967. Acari. In burges, a. and ran, f. (eds.). Soil Biology. Academic Press, London, 363-395. 68 PALAEONTOLOGY, VOLUME 20 Wilkinson, M. 1930. Note on a wounded lepidodendroid axis. Mem. Proc. Manch. Lit. Phil. Soc. 73, 75-82. Williamson, w. c. 1880. On the organization of the fossil plants of the Coal Measures. X. Including an examination of the supposed radiolarians of Carboniferous rocks. Phil. Trans. Roy. Soc. 171, 493-539. ANDREW C. SCOTT Department of Geology Trinity College Typescript received 16 December 1975 Dublin 2 Revised typescript received 25 February 1976 Eire CLASSIFICATION OF THE TRILOBITE PSEUDAGNOSTUS by JOHN H. SHERGOLD Abstract. Eighty-eight species assigned or assignable to Pseudagnostus sensu lato are divided into two broad divi- sions based on the position of the glabellar node with respect to the anterior or anterolateral glabellar furrows and anterolateral lobes. A spectaculate division, in which the node lies to the rear of the anterior furrow and to the rear of the anterolateral lobes, is divided into nine species groups which are recognized by degree of effacement of external morphology, shield shape, border morphology, glabellar morphology, pygidial spinosity, and metamerism. Three spectaculate species groups, bulgosus , communis , and cyclopvge , are referred to Pseudagnostus Jaekel, 1909, sensu stricto\ one, contractu, to Pseudagnostina Palmer, 1962; and one, securiger , to Sulcatagnostus Kobayashi, 1937, these latter taxa being regarded as subgenera of Pseudagnostus. Four other spectaculate species groups, araneavelatus , bilobus, canadensis, and clavus, are classified with Neoagnostus Kobayashi, 1955, pending clarifica- tion of the taxonomic status and concepts of Euplethagnostus Lermontova, 1940, and Pseudorhaptagnostus Lermon- tova, 1940. Hyperagnostus Kobayashi, 1955 and Machairagnostus Flarrington and Leanza, 1957 are synonymized with Neoagnostus. A papilionate division, in which the axial glabellar node lies between the anterolateral lobes and interrupts the course of the anterior furrow, consists of two species groups, convergens and clarki, which are assigned to Rhaptagnostus Whitehouse, 1936. The generic name Plethagnostus Clark, 1923 is suppressed. The genus Pseudagnostus was erected by Jaekel in 1909 for species of Agnostus possessing a short pygidial axis and an ‘endolobe’, currently termed a deuterolobe (see Opik 1963, p. 31). Jaekel (1909, p. 400) designated as type species Agnostus cyclopvge Tullberg (1880, p. 26; pi. 2, fig. 15a, c, cephalon and pygidium respectively), which occurs in the Olenus Zone and the zone of Parabolina spinulosa with Orusia lenticularis, at Andrarum, Skane, and other localities in Sweden (Tullberg 1880, p. 26; Westergard 1922, p. 1 17; 1944, pp. 32-33; 1947, p. 22). The type specimens, reported by Tullberg (1880, p. 26) to have been deposited in the Geological Museum, University of Lund, have not been identified. According to Dr. Jan Bergstrom (pers. comm. 1973), the specimens were numbered but there is no evidence that they were deposited in the type collections. Although they may not be lost, it might be difficult to differentiate them from other material in the collections from Andrarum. Until the specimens are positively located, the concept of Pseudagnostus cyclopvge is based on specimens collected from the type locality, figured by Westergard (1922, pi. 1, fig. 7, cephalon, Lund University Lo 3066t, and 8, pygidium, Lo 3067t), and reproduced here on Plate 15, figs. 1-2. Westergard’s specimens were obtained from beds containing O. lenticularis. Recent research (Opik 1967; Shergold 1972, 1975) classifies the genus Pseud- agnostus with Pseudagnostinae Whitehouse, 1936, which is regarded as a subfamily of Diplagnostidae Whitehouse, 1936, as emended by Opik (1967). When it was erected, a mere handful of species could be assigned to Pseudagnostus ; but by now it is possible to assign no fewer than eighty-eight species, listed in Appendix B. They are united by the possession of a deuterolobe similar to that seen in the type species. Excluding this characteristic, a wide range of forms has [Palaeontology, Vol. 20, Part 1, 1977, pp. 69-100, pis. 15 16.] 70 PALAEONTOLOGY, VOLUME 20 been included in Pseudagnostus, with the result that any original concept of the genus has become substantially changed, and it has become necessary to discuss the genus in a sensu lato manner. This proliferation of species is directly responsible for the present review, as Pseudagnostus sensu lato is common in the late Cambrian of northern Australia, currently under investigation by the author. There, a range of pseudagnostinid forms spans a wide interval of late Cambrian time, and is poten- tially useful for the close zonation of upper Cambrian strata (Shergold 1975). AGE AND DISTRIBUTION OF SENSU LATO TAXON The time span of Pseudagnostus sensu lato is long : it first appears in the earliest late Cambrian (early Mindyallan Erediaspis eretes Zone in Australia; early Tuorian Agnostus pisiformis/ Homagnostus fecundus Zone in Siberia; and early Dresbachian Cedaria Zone in North America) and continues into the earliest Ordovician (Cana- dian, Symphysurina and Kainella faunas in North America). Pseudagnostoid species are particularly common in correlatives of the late Mindyallan Glyptagnostus stolidotus Zone in Australia, in the late Dresbachian Aphelaspis and Dunderbergia Zones in North America, in the early Franconian and early Shidertan Elvinia and Irvingella Zones in North America and Siberia respectively, and in the late Shidertan Lotagnostus trisectusj Peltura Zone in Siberia and its equivalent post-Idamean/pre- Payntonian interval in Australia. The genus is cosmopolitan, having lived in seas peripheral to Precambrian crustal masses now forming the nuclei of Europe, Eurasia, eastern Asia, North America, South America, Australia, and Antarctica. Within these seas, Pseudagnostus appears to have favoured habitats within ocean-facing environments (Palmer 1972), at the ocean-neritic boundary (Robison 1972) which approximates the boundary between carbonate belt and outer detrital belt as interpreted by Palmer (1960a, 1969, 1972, 1973) and Robison (1960a). Its species are commonly found in the deep subtidal EXPLANATION OF PLATE 15 Figs. 1-2. Pseudagnostus ( Pseudagnostus ) cyclopyge (Tullberg, 1880). 1, LULo 3066, topotype cephalon (Westergard 1922, pi. 1. fig. 7), x7. 2, LULo 3067, topotype pygidium (Westergard 1922, pi. 1, fig. 8), x 7 ; the concept of Pseudagnostus currently rests on these specimens. Figs. 3-4. Pseudagnostus (Pseudagnostus) ampullatus Opik, 1967. 3, CPC 5899, paratype cephalon, x 1 1 • 5. 4, CPC 5896, holotype pygidium, x 1 1 ; an en grande tenue species preserved in limestone. Figs. 5-6. Pseudagnostus ( Pseudagnostus ) bulgosus Opik, 1967. 5, CPC 5904, paratype cephalon, X 12-5. 6, CPC 5901, holotype pygidium, x 12. Figs. 7-8. Pseudagnostus (Pseudagnostus) communis (Hall and Whitfield, 1877). 7, USNM 24557e, silicone replica of hypotype cephalon (Palmer 1955, pi. 19, fig. 20), x8-5. 8, USNM 24557d, replica of hypo- type pygidium (Palmer 1955, pi. 19, fig. 21), x 12. Figs. 9 10. Pseudagnostus (Pseudagnostus) josepha (Hall, 1863). 9, 10, silicone replicas of cotypes, AMNH 311, preserved as sandstone moulds, x 8 ; illustrated for comparative purposes. Figs. 11-12. Pseudagnostus (Pseudagnostina) contracta (Palmer, 1962). 11, USNM 143149b, replica of paratype cephalon, X 16. 12, USNM 143150, replica of holotype pygidium, x 16. Fig. 13. Pseudagnostus (Sulcatagnostus) securiger (Lake, 1906). GSM 57650, replica of holotype, x6. Figs. 14-15. Rhaptagnostus clarki (Kobayashi, 1935). 14, USNM 146887, paratype cephalon, x9-5. 15, USNM 93062, holotype pygidium, x9-5. PLATE 15 SHERGOLD, Pseudagnostus 72 PALAEONTOLOGY, VOLUME 20 (Laporte 1971, for terminology) outer detrital belt dark shale, silt, and finely lami- nated limestone deposited on the oceanic margins of carbonate banks, in deep sub- tidal interbank channels, and other places with current connection to the open ocean. In such environments pseudagnosti presumably contribute to late Cambrian equivalents of Agnostid Community Assemblages 1 and 2 of Jago (1973). Species of Pseudagnostus are also found in coarse calcarenite and debris layers, allochthon- ously deposited on beaches or in channels. Pseudagnosti are less commonly found in the sandstones and dolomites of the inner detrital belt (Palmer’s (1960a, 1969, 1973) terminology). When found in inner detrital depositional environments they appear to retain a more or less constant morphology for an appreciable length of time, e.g. P.josepha (Hall), and are not repeatedly replaced in the biostratigraphical section by rapidly evolving taxa to the same extent as they appear to be in outer detrital and carbonate belt environments. Maximum species diversification is thus observed at the oceanic/neritic boundary, where warm shelf currents and cooler oceanic currents mingle. MORPHOLOGY In assessing morphology it is necessary to know if specimens are testiferous or moulds, as well as the nature of the matrix. Little information, other than the shapes of the cephalic and pygidial shields and a vague lobation and perhaps furrowing, can be gained from sandstone moulds, as is evident from the cotypes of P. josepha (Hall 1863) (PI. 15, figs. 9-10). By way of contrast, the species illustrated in Plate 15, figs. 3-4 and Plate 16, figs. 3-6 demonstrate the range of morphological features shown by both exoskeletons and moulds preserved in limestone. The distinction between morphology exhibited by moulds and by tests is basic to understanding both the reasons for the proliferation of species assigned to Pseud- agnostus, and the difficulties in subdividing them at generic and subgeneric levels. EXPLANATION OF PLATE 16 Figs. 1-2. Rhaptagnostus convergens (Palmer, 1955). 1, USNM 123563, paratype cephalon, X 10. 2, USNM 123562, holotype pygidium, x9-5. Figs. 3-6. Rhaptagnostus bifax (Shergold, 1975). 3, CPC 1 1596, holotype cephalon preserved with exo- skeleton, x 10. 4, CPC 1 1656, paratype pygidial exoskeleton, x 10. 5, CPC 1 1597, replica of parietal surface of paratype cephalon, x 10. 6, CPC 1 1667, parietal surface of paratype pygidium, x7. Figs. 7-8. Neoagnostus bilobus (Shaw, 1951). 7, USNM 124467, replica of paratype cephalon, x 16. 8, USNM 124468, replica of holotype pygidium, x 12-5. Figs. 9-10. Neoagnostus canadensis (Billings. 1860). 9, GSC 858, replica of paratype cephalon, x 10. 10, GSC 858b, replica of lectotype pygidium, x 9. Figs. 1112. Neoagnostus araneavelatus (Shaw, 1951). 11, USNM 124467, replica of paratype cephalon, x20. 12, USNM 124466, replica of holotype pygidium, x22. Figs. 13-15. Neoagnostus clavus (Shergold, 1972). 13, 14, CPC 8454, paratype cephalon, lateral and dorsal views respectively, x9. 15, CPC 8451, paratype pygidium, x 10-5. Fig. 16. Neoagnostus aspidoides Kobayashi, 1955. Replica of holotype cephalon, GSC 12745, x 7. Fig. 17. Neoagnostus [Hyperagnostus] binodosus (Kobayashi, 1955). Replica of holotype cephalon, GSC 12747, x 7. Fig. 18. Neoagnostus [ Trinodus ] priscus (Kobayashi, 1955). Replica of holotype pygidium, GSC 12751, x 7-5. PLATE 16 in SHERGOLD, Pseudagnostus Tfev <> • 74 PALAEONTOLOGY, VOLUME 20 Some species have been described from moulds alone, others from exoskeletons, both generally in an uncritical manner. Reference to Plate 15, figs. 3-4 will show that supposedly diagnostic differences may readily be found if preservation is ignored. When such differences become the basis for generic taxa, obvious problems arise. In an earlier paper (Shergold 1975), and here, a major distinction is made between the morphology of the outer surface of the exoskeleton, and that of its internal surface. The latter can be observed directly on the inner surface of silicified or phosphatized exoskeletons; most commonly, however, it is interpreted from impres- sions on the surface of its internal mould. Such surfaces are here termed ‘parietal surfaces’, and their morphology, which is a negative of the inner surface of the exoskeleton, is termed ‘parietal morphology’. From the classificatory point of view, the morphology of the outer surface of the exoskeleton gives information on lobes and furrows, and the positions of nodes and spines. Parietal morphology, on the other hand, gives basic anatomical informa- tion relating to the internal musculature, and the vascular and caecal systems of the organism, as deduced from the occurrence and distribution of various scrobi- culations, muscle scar impressions, and pits. Morphological terminology applicable to Pseudagnostinae ( inter alia ) has been de- fined exhaustively elsewhere (Opik 1963, 1967; Shergold 1972, 1975). Terminology used in the classification presented here, and not readily evident from these papers or requiring further comment, is summarized in Appendix A. LOCATION OF MATERIAL Observations on personally collected Australian specimens have been supplemented by the acquisition of replicas of the types of most established taxa, and the study of museum materials in Europe, North America, and Japan. Located material is in the collections of repositories abbreviated as follows : AMNH BMNH BYU CPC GMAN GSC GSCh GSM HAN IGAL IGGN IGUT IPP LU LULo MCZ MNHU MSM American Museum of Natural History, New York, U.S.A. British Museum (Natural History), London, U.K. Brigham Young University, Provo, Utah, U.S.A. Commonwealth Palaeontological Collection, Canberra, A.C.T., Australia. Geological Museum, Kazakhstan Academy of Science, Alma Ata, U.S.S.R. Geological Survey of Canada, Ottawa, Canada. Geological Survey of China, Peking, China. Institute of Geological Sciences, London, U.K. Cambrian Catalogue, Geological Survey, Hanover, Germany. Institute of Arctic Geology, Leningrad, U.S.S.R. Siberian Research Institute for Geology, Geophysics, and Mineral Raw Materials (SNIIGGIMS), Novosibirsk, U.S.S.R. Institute of Geology, University of Tokyo, Japan. Institute of Palaeontology, Peking, China. Laval University, Quebec, Canada. Lund University, Lund, Sweden. Museum of Comparative Zoology, University of Harvard, Cambridge, Massachusetts, U.S.A. Museum fur Naturkunde, Humboldt University, East Berlin, D.D.R. Manchurian Science Museum, Mukden, Manchuria, China. SHERGOLD: P S E U D AG NO STU S 75 NHMM Natural History Museum, Mendoza, Argentina. NUP National University, Peking, China. NZAR Athropod Register, New Zealand, Geological Survey, Lower Hutt, New Zealand. OGM Oklahoma Geological Survey, Tulsa, Oklahoma, U.S.A. OUM Oxford University Museum, Oxford, U.K. RMS Riksmuseet, Stockholm, Sweden. SGU Geological Survey of Sweden, Stockholm, Sweden. THU Tsing-hua University, Peking, China. UBA University of Buenos Aires, Argentina. UQ University of Queensland, St. Lucia, Queensland, Australia. USGS United States Geological Survey, Washington, D.C., U.S.A. USNM National Museum of Natural History (formerly United States National Museum), Washing- ton, D.C., U.S.A. UT University of Texas, Austin, Texas, U.S.A. YPM Peabody Museum, Yale University, New Haven, Connecticut, U.S.A. ZSGU Museum of the West Siberian Geological Board, Novokuznetsk, U.S.S.R. CLASSIFICATION For the purposes of this review species assigned to the following closely related or synonymous genera and subgenera were considered: Pseudagnostus Jaekel, 1909; Plethagnostus Clark, 1923; Rhaptagnostus Whitehouse, 1936; Sulcatagnostus Kobayashi, 1937a; Pseudorhaptagnostus Lermontova, 1940; Euplethagnostus Ler- montova, 1940; Neoagnostus Kobayashi, 1955; Hyperagnostus Kobayashi, 1955; Machairagnostus Harrington and Leanza, 1957; and Pseudagnostina Palmer, 1962. These species, together with others previously classified elsewhere, are listed in Appendix B. Three other pseudagnostinid genera were examined but are not con- sidered further in the classification : Litagnostus Rasetti, 1944, because of the extreme difficulty in assessing this very effaced form; Xextagnostus Opik, 1967, because of its uncharacteristically simple articulating device and different pygidial diverticular structure; and Oxyagnostus Opik, 1967, because of its distinct deuterolobe morpho- logy which cannot be confused with that of Pseudagnostus (see Opik 1967, pp. 1 59- 161). For the purposes of classification the most obvious differences between species are carapace shape and degree of effacement. Less obvious, but seemingly more important from the anatomical point of view, is the position of the axial glabellar node with respect to the anterior furrow and anterolateral glabellar furrows. Species fall broadly into two divisions. Spectaculate pseudagnosti are those in which the glabellar node lies to the rear of the anterolateral lobes. Two groups are discernible: those in which the anterior furrow is transverse, straight, or gently curved back- wards and the anterolateral lobes are distinctly separated sagittally, e.g. Pseud- agnostus communis (Hall and Whitfield), as refigured by Palmer (1955) (see PI. 15, figs. 7-8); and those in which the anterior furrow is V-form, and the anterolateral lobes meet, and may fuse, sagittally, e.g. Neoagnostus bi lotus (Shaw, 1951) (PI. 16, fig. 7). Papilionate pseudagnosti are those in which the node interrupts the course of the anterior furrow, dividing it into V-form anterolateral furrows, and lies between the anterolateral lobes as, for example, in Rhaptagnostus clarki (Kobayashi, 1935a) (PI. 15, fig. 14). 76 PALAEONTOLOGY, VOLUME 20 An intermediate condition, in which the node interrupts the course of the anterior furrow, dividing it into V-form anterolateral furrows, but still lies slightly behind the anterolateral lobes, is included within the spectaculate division of this classifica- tion, but could possibly be regarded as a distinct category. The condition appears to be gradational, but the observed differences could be interpreted as the result of preservation. The spectaculate condition embraces the bulk of species assigned to Pseudagnostus sensu lato. Spectaculate pseudagnosti occur earlier, being first found in the earliest late Cambrian of Australia, North America, and Siberia, and disappear later, in the early Ordovician of North America. The intermediate forms occupy a relatively narrow time interval during the late Cambrian, in the Elvinia and Conaspis Zones and their correlatives in Europe, Siberia, and Australia. Papilionate pseudagnosti arise during this time interval, at least in Australia (Shergold, in prep.), and continue into the earliest Ordovician (in Mexico). Thus there appears to be a time-related forward migration of the axial glabellar node. The function of the axial glabellar node is not known: nor is the reason for its apparent migration. Some speculation may, however, be offered. Parietal surfaces of some pseudagnostinid cephala show the axial and terminal glabellar nodes con- nected by an axial carina which itself bears a longitudinal sulcus (Shergold 1975, pi. 3, fig. 5; text-fig. 15), structures which can be interpreted as having supported the gut and a dorsal tubular heart. The high spot of the axial glabellar node is fre- quently perforated, particularly in limestone moulds. Ruedemann (1916) interpreted the glabellar node of Cryptolithus as a dorsal median eye containing a single fluid lens, and such lenses may have occupied the perforated regions of both cephalic and pygidial axial nodes in agnostids, their function being to assess the intensity and direction of light and orient the animal within the water column. Harris and Mason (1956) have shown that experimentally blinded Daphnia are ‘more sensitive to light than normal ones and show a greater capacity for adaptation to light during the cycle of vertical migration’ (p. 285). Thus the lack of a compound eye, also lacking in agnostids, is no bar to photokinesis. Harris and Mason (1956) have suggested for Daphnia an interrelationship of the photosensitive system, the nervous system, and reflex control of the heart rate. The cephalic and pygidial nodes of pseudagnosti may well, therefore, be connected with photosensitive systems in direct connection with the blood vascular system, the combination facilitating orientation and migration of the animal in the water column. Opik (1961o, p. 417) has considered that the agnostid stomach consists of anterior and posterior sacs connected by a constricted passage beneath the anterior trans- verse glabellar furrow. In pseudagnosti, however, the lateral portions of this furrow merely represent the internally raised anterior margins defining the anterolateral muscle attachment areas which, depending on the length of muscle, would not necessarily constrict the stomach. Any interruption or modification of the course of the anterior furrow, as is apparent, is therefore more likely to reflect a modification of the muscle attachment areas. The positioning of the axial glabellar node between the anterolateral muscle attachment areas in papilionate pseudagnosti may have resulted in the lateral separation of the bases of the appendages attached there, and/or restricted the area of attachment of such appendages, which would presum- SHERGOLD: P S E U D AG N O STU S 77 ably be restricted in size. This situation may be contrasted with that seen in the spectaculate species here assigned to Neoagnostus , in which the attachment areas may meet axially. The anterolateral muscle scars are generally the largest of the pseudagnostinid cephalon, and are thought to represent the areas of attachment for mandibular appendages. Thus modification of the shape and position of the mandibular attachment areas could reflect difference in feeding habit. Hence this classification is based on anatomical features fundamental to the animal. While the position of the axial node and its relationship to the anterolateral muscle scars is regarded here as basic in the classification of Pseudagnostinae, other factors are of considerable value in recognizing species groups. Some can be recog- nized by degree of effacement, whereas in others this characteristic varies presumably in response to environmental conditions. Ideally species groups should comprise taxa in which the condition of effacement is relatively constant, and in practice this is difficult to attain owing to the differing preservation exhibited by established species. Ideally genera should embrace effaced, partially effaced, and en grande tenue species groups. Degree of effacement influences observation on the condition of other character- istics used in this classification. Although all pseudagnosti are deuterolobate, as deduced from parietal morphology, the degree of tumidity of the deuterolabe and degree of incision of its associated bounding accessory furrows vary considerably. Externally effaced and partially effaced pseudagnosti are generally weakly deutero- lobate and non-plethoid parietally. Species externally en grande tenue are usually strongly deuterolobate and plethoid, but there are exceptions, e.g. the bilobus species group described below. In similar manner degree of deliquiation (Shergold 1975) is related to effacement. Externally effaced and partially effaced species are non- deliquiate or subdeliquiate : their moulds are likely to have subdeliquiate and deliquiate marginal furrows. En grande tenue species are deliquiate, and their moulds strongly so: generally cephala are more strongly deliquiate than their assigned pygidia. Carapace shape, which may be subcircular, subovoid, subquadrate, or sub- rectangular, is relatively constant in some, but not all, species groups. Pygidial segmentation, which can be assessed from the number of muscle scars and apodemes on the internal surface of the test, or the number of muscle-scar impressions and notulae on the parietal surface, may be related to shield shape. As a generalization, elongated pygidia have more paired muscle attachment areas, and by inference segments, than subquadrate ones. Similarly, the position of the posterolateral pygi- dial spines may be related to the shield shape: elongate pygidia have anteriorly situated spines with respect to the termination of the deuterolobe, whereas sub- quadrate pygidia have retrally sited spines. Shield shapes, and consequently spine positions, may change during morphogenesis. Most species in which meraspid pygidia have been observed have a subquadrate shield with retrally positioned spines. Change in shape occurs with the development of the deuterolobe during late meraspid morphogenesis (see Palmer 1955, pp. 94-96, pi. 20 for P. communis). Morphological features associated with the anterior portion of the pygidial axis have not been fully exploited in this classification. While the clarity of the transverse furrows, and the axial furrows bounding the pre-deuterolobe axis is dependent on 78 PALAEONTOLOGY, VOLUME 20 the degree of effacement shown by the exoskeleton in general, the degree of anterior divergence of the axial furrows varies considerably, even within species. Among some species assigned here to Neoagnostus, a third pair of muscle scars is consistently enclosed by axial furrows. In others, their visibility is dependent on degree of efface- ment, cf. N. bilobus (Shaw) (PI. 16, fig. 8) and N. araneavelatus (Shaw) (PI. 16, fig. 12), and manner of preservation (exoskeleton or mould). A third pair of muscle-scar impressions is often also visible on exfoliated specimens of Pseudagnostus and Rhaptagnostus, but in these is never bounded by axial furrows (e.g. PI. 16, fig. 2). The characteristic is variable, but may have some use in differentiating isolated pygidia of Neoagnostus from those of Pseudagnostus or Rhaptagnostus. The position of the pygidial axial node is quite variable, but unlike the axial glabellar node this variation is inconsistent. It lies across the sagittal length of the second axial segment, is apparently connected to the axial region between the muscle scars of the first segment, and extends rearwards between the scars of the third segment to varying extents. In Pseudagnostus and its subgenera, it terminates gener- ally at the front of the third intermuscle scar area (PI. 15, figs. 2, 4, 6, 10, 13); in Rhapt- agnostus it may extend considerably further between the third pair of muscle scars (PI. 16, figs. 2, 6); but in Neoagnostus its extent is only consistent within individual species (cf. PI. 16, figs. 8, 10, 12, 15). Size, shape, and transverse separation of the muscle scars evidently influence observation of the extent of this feature. In many pseudagnosti the axial pygidial node is bilobed, having anterior and posterior high spots (see PI. 15, figs. 4, 6; PI. 16, figs. 2, 6), and is bifid, being pos- teriorly cleft (PI. 16, fig. 6). On some specimens the anterior high spot is perforated, like the axial glabellar node. The axial pygidial node is therefore a complicated structure which presumably had multiple functions. DELINEATION OF SPECIES GROUPS Species are grouped here on the combinations of characteristics listed above. Con- siderable variation occurs in some of the groups whereas others are quite homo- geneous. The groups are not species, although synonymous species may exist within them. Species groups may constitute subgenera, but apart from those lying at the limits of variation, e.g. possessing three pygidial spines ( seeuriger group) or are extremely effaced ( contracta ), for which names are already available ( Sulcatagnostus and Pseudagnostina respectively), it is preferred not to name them formally at this time. Not all species have yet been examined at first hand, particularly Russian taxa, and some may further modify the concepts presented herein. Of the species listed those marked with an asterisk signify that type or subsequently figured material has been examined in museum collections, or from silicone replicas. A. SPECTACULATE PSEUDAGNOSTI Araneavelatus group ( PI. 16, figs. 11-12) Nominal species. * Pseudagnostus araneavelatus Shaw, 1951, pi. 24, p. 113, figs. 12-16, holotype pygidium USNM 124466, paratype USNM 124467, remainder untraced, early Ordovician, Missisquoia Zone, Vermont, U.S.A. SHERGOLD: PSEUDAGNOSTUS 79 Diagnosis. Cephalon subcircular, en grande tenue or partially effaced, non-deliquiate or subdeliquiate, spectaculate, tendency to develop prominent furrows to the rear of the anterolateral lobes, anterolateral furrows generally externally effaced, median preglabellar furrow absent. Pygidium subcircular, partially effaced, non-deliquiate, non-plethoid, weakly deuterolobate, spines posteriorly situated close to a transverse line drawn across the rear of the deuterolobe, possibly eight late holaspid metameres. Other species * Pseudagnostus coronatus Shergold, 1975, pp. 85-87, pi. 6, figs. 1-6, holotype cephalon CPC 11692, paratypes CPC 11693 1 1697. Pseudagnostus cyclopygeformis (Sun) sensu Kobayashi 1960, p. 341, pi. XIX, fig. 6, IGUT no number, non fig. 7 ( clarki group), IGUT no number. * Phalacroma cyclostigma Raymond, 1924, p. 397, pi. 12, fig. 4, holotype pygidium YPM 4747. * Pseudagnostus denticulatus Shergold, 1975, pp. 87-89, pi. 8, figs. 1-5, holotype cephalon CPC 11705, paratypes CPC 11706-11709. * Pseudagnostus sp. C Shergold, 1975, pp. 91-92, pi. 7, figs. 5-7, CPC 1 1714- 11716. Age and distribution. Late Cambrian : pre-Payntonian/post-Idamean, Rhaptagnostus clarki patulus with Caznaia squamosa through R. bifax with Neoagnostus denticulatus Assemblage-Zones (Shergold 1975), western Queensland, Australia; late Changshanian ; Kaolishania Zone, Tanggok, South Korea. Early Ordovician; Missisquoia Zone, Vermont, U.S.A. Comments. The Australian species listed here as constituting the araneavelatus group were previously placed in the c/avus group (Shergold 1975). They are nevertheless distinguishable from the c/avus and bilobus groups (below) by their rounded shield shapes and proportions. The arrangement of glabellar lobes and furrows shown is common to all groups. Bilobus group (Shergold 1975, p. 92) (PI. 16, figs. 7—8) Nomina! species. * Pseudagnostus bilobus Shaw, 1951, pp. 112-113, pi. 24, figs. 17-22, holotype pygidium USNM 124468, paratypes USNM 124469-124471, early Ordovician, Missisquoia Zone, Vermont, U.S.A. Diagnosis. Cephalon subquadrate, generally en grande tenue , deliquiate or sub- deliquiate, spectaculate, tendency to over-deepen furrows both in front and behind the anterolateral glabellar lobes; a median preglabellar furrow is generally present. Pygidium subcircular or subquadrate, partially effaced, deliquiate or subdeliquiate, non-plethoid, weakly deuterolobate, spines generally sited on a transverse line across the rear of the deuterolobe or behind it; a third pair of muscle-scar impressions is incorporated into the axis behind and adjacent to the pygidial axial node. Other species * Neoagnostus aspidoides Kobayashi, 1955, pp. 473-474, pi. VII, fig. 5, holotype cephalon GSC 12745, non fig. 4, geragnostoid pygidium GSC 12746, pi. IX, fig. 5 (line-drawing). * Hyperagnostus binodosus Kobayashi, 1955, p. 475, pi. VII, fig. 2, holotype cephalon GSC 12747, non fig. 3, geragnostoid pygidium GSC 12748, pi. IX, fig. 4 (line-drawing). Agnostus cyclopyge Tullberg sensu Sun 1939, p. 30, pi. 1, figs. 1-3, repository and numbers not known. * Pseudagnostus longicollis Kobayashi, 1966, p. 283, fig. 7, IGUT no number. * Trinodus priscus Kobayashi, 1955, p. 476, pi. VII, fig. 6, holotype pygidium GSC 12751. * Pseudagnostus quasibilobus Shergold, 1975, pp. 94-95, pi. 12, figs. 1-7, holotype cephalon CPC 11717, paratypes CPC 11718-11723. Machair agnostus tmetus Harrington and Leanza, 1957, p. 64, figs. 6-7, holotype (complete specimen) UBA 1297, paratypes UBA 1195, 1292-1294, 1298. *Undetermined pseudagnostid, Robison and Pantoja-Alor 1968, p. 780, pi. 97, fig. 23, USNM 158886. 80 PALAEONTOLOGY, VOLUME 20 Age and distribution. Late Cambrian: Payntonian, Neoagnostus quasibilobus with Tsinania nomas Assemblage-Zone (Shergold 1975), western Queensland, Australia; Fengshanian, Yunnan; Wanwanian, Jehol Block, China. Early Ordovician: Oaxaca Province, Mexico; Missisquoia Zone, Vermont, U.S.A.; Tremadocian, Para- bolina argentina Zone, Argentina; Missisquoia, Symphysurina, and Kainella-Evansaspis faunas, British Columbia, Newfoundland, Canada. Comments. Like the clavus group from which it is possibly derived, the bi/obus group represents an assemblage of species linked through the orientation of their glabellar furrows. Cephala are, in general, preserved en grande tenue, whereas pygidia have depressed deuterolobes and effaced accessory furrows. The retral position of the pygidial spines is constant among the species cited. Bulgosus group (PI. 15, figs. 5-6) Nominal species. *Pseudagnostus bulgosus Opik, 1967, pp. 156-158, pi. 38, fig. 8; pi. 62, figs. 1-4, Mindy- allan zones of Erediaspis eretes and Glyptagnostus stolidotus , western Queensland, Australia. Designated holotype pygidium CPC 5901, paratypes CPC 5902-5904, 5656. Diagnosis. Cephalon subovoid, en grande tenue , spectaculate, deliquiate, anteriorly converging acrolobe, median preglabellar furrow absent. Pygidium subovoid, en grande tenue , convergent flanks, subdeliquiate, plethoid, deuterolobate, restricted pleural lobes, very small spines in advance of a transverse line drawn across the rear of the deuterolobe in holaspides. Other species *Oedorhachis boltonensis Resser, 1938, p. 50, pi. 10, figs. 19-20, cotypes USNM 94869. Pseudagnostus levatus E. Romanenko in Romanenko and Romanenko, 1967, pp. 75-76, pi. 1, figs. 18-19, holotype pygidium ZSGU 133/135, paratype ZSGU 133/136. *Oedorhachis mesleri Resser, 1938, p. 50, pi. 10, figs. 13-14, cotypes USNM 94864. * Pseudagnostus nganasanicus Rosova, 1964, pp. 27-28, pi. 16, figs. 3-4, holotype pygidium IGGN 113/875, paratype IGGN 1 13/928. Possibly also included in the bulgosus group is: Pseudagnostus mestus Opik, 1967, pp. 155-156, pi. 62, figs. 5-6, holotype pygidium CPC 5906, paratype cephalon CPC 5905. Age and distribution. Late Cambrian : Mindyallan, Erediaspis eretes and Glyptagnostus stolidotus Zones, western Queensland, Australia; Dresbachian, Cedaria Zone, Virginia and Tennessee, U.S.A.; Tuorian, zones of Agnostus pisiformis with Homagnostus fecundus and Glyptagnostus stolidotus, Kulyumbe River, Katun River, and Lena River, Siberian Platform, U.S.S.R. Comments. This is the earliest species group referable to Pseudagnostus sensu lato , homogeneous in content and character. Canadensis group (PI. 16, figs. 9-10) Nominal species. Agnostus canadensis Billings, 1860, p. 304, fig. 3 a-b, lectotype pygidium (designated Rasetti (1944) GSC 858b, paratype cephalon GSC 858, refigured as Pseudagnostus canadensis (Billings), (Rasetti 1944, p. 234, pi. 36, figs. 8-13, GSC 858, 858b, LU 1 104a-d), late Cambrian, Levis Conglomerate, Quebec, Canada. Diagnosis. Cephalon subquadrate, en grande tenue , wide borders with deliquiate marginal furrows, spectaculate, anterolateral glabellar furrows effaced, but those to SHERGOLD: P S El) D AG N O STU S rear of anterolateral lobes prominent, median preglabellar furrow present parietally. Pygidium subquadrate, en grande tenue , wide borders with deliquiate marginal furrows, strongly deuterolobate, strongly plethoid, restricted pleural areas, retral spines. Other species. Rasetti (1944) has synonymized Agnostus janei Clark, 1923, p. 124, fig. 8; 1924, p. 19, fig. 5, MCZ 1696. Age and distribution. Late Cambrian: Levis Conglomerate, Quebec, Cow Head Group conglomerates, Newfoundland (C. H. Kindle, pers. comm.). Comments. The canadensis group is readily recognized by its strongly en grande tenue condition and fused anterior and anterolateral glabellar lobes. Most probably it has been derived from the same ancestral stock as the araneavelatus group. Clams group (Shergold 1975, p. 82) (PI. 16, figs. 13-15) Nominal species. * Pseudagnostus clams Shergold, 1972, pp. 3 1 -34, pi. 3, figs. 1-8, holotype pygidium CPC 8453, paratypes CPC 8451, 8454-8456; also Shergold, 1975, pp. 84-85, pi. 8, figs. 6-12, CPC 11689- 11704, late Cambrian, Rhaptagnostus bifax with Neoagnostus denticulatus and R. clarki maximus with R. papilio Assemblage-Zones (Shergold 1975), western Queensland, Australia. Diagnosis. Cephalon subquadrate, en grande tenue or partially effaced, wide borders with subdeliquiate or deliquiate marginal furrows, spectaculate, prominent V-form anterolateral glabellar furrows, weakly chevronate furrows to rear of anterolateral lobes, rhomboid anterior lobe, median preglabellar furrow absent, proximally present, or present. Pygidium subquadrate to subovoid, en grande tenue or partially effaced, non-deliquiate or subdeliquiate, non-plethoid or subplethoid, generally weakly deuterolobate, retral pygidial spines, seven to eight metameres in late holaspides. Other species * Rhaptagnostus acutifrons Troedsson, 1937, pp. 22-24, pi. 1, fig. 9, RMS number not known. * Pseudagnostus cavernosus Rosova, 1960, pp. 12-14, pi. 1, figs. 1-4, holotype pygidium IGGN 76/557, paratype 74/556; holotype refigured Rosova in Khalfin 1960, p. 165, pi. Cm-XVIII, fig. 4. Pseudorhaptagnostus punctatus Lermontova, 1940, p. 126, pi. 49, fig. 14, 14a, repository and numbers not known. Pseudorhaptagnostus simplex Lermontova, 1951, pp. 12-13, pi. 2, figs. 11-15, non figs. 16-17, designated holotype is pygidium fig. 1 1, repository and numbers not known; also Pseudagnostus simplex (Lerm.) in Nikitin 1956, pi. XIV, fig. 5, non fig. 4. Euplethagnostus subangulatus Lermontova, 1940, p. 126, pi. 49, fig. 15, 15a, repository and numbers not known. * Pseudagnostus vulgaris Rosova, 1960, pp. 14-16, pi. 1, figs. 5-13, holotype pygidium IGGN 76/645, paratypes IGGN 74/524, 75/582, 76/578, 76/647, 76/653, 76/654, 76/656, 79/633; refigured Rosova in Khalfin 1960, p. 165, pi. Cm-XVIII, fig. 5 a-c. * Pseudagnostus sp. A Shergold, 1975, pp. 89-90, pi. 7, figs. 1-2, CPC 1 1710-1 1711. * Pseudagnostus sp. B Shergold, 1975, pp. 90-91, pi. 7, figs. 3-4, CPC 11712- 11713. Questionably belonging to this group are: * Homagnostus cf. acutus Kobayashi, 1938, pp. 173-174, pi. XV, fig. 4, cephalon GSC 1 1979. Pseudagnostus bituber culatus Ivshin in Khalfin, 1960, p. 165, pi. Cm-XVIII, fig. 6 a-b, repository and numbers not known. Pseudagnostus quadratus Lazarenko, 1966, pp. 46-47, pi. 1, figs. 24-29, holotype cephalon IGAL 36/8907, paratype numbers not known. F 82 PALAEONTOLOGY, VOLUME 20 Age and distribution. Late Cambrian: post-Idamean/pre-Payntonian, Rhaptagnostus clarki patulus with Caznaia squamosa through R. c. maximus with R. papilio Assemblage-Zones (Shergold 1975), western Queensland, Australia; Shidertan, zones of Irvingella to Lotagnostus trisectus/Peltura (Ivshin and Pokrov- skaya 1968), Kazakhstan, Sayan Altai, Olenek River, U.S.S.R.; early Franconian, Elvinia Zone, British Columbia, Canada; Changshanian, Kaolishania Zone, South Korea; late Cambrian, Tienshan, China. Comments. As presently constituted, this is a heterogeneous group that varies considerably in degree of effacement and deliquiation, and somewhat in shield shape. Species are presently linked in possessing a V-form anterior glabellar furrow, rhom- boidal anterior lobe, and seven or eight pygidial metameres. Many Russian species are, however, inadequately known and may be wrongly classified within this group. Possibility exists for further dividing the group on presence or absence of a median preglabellar furrow. As far as is known only Australian representatives lack a well- defined median preglabellar furrow externally. Communis group (Palmer 1968, p. 30) (PI. 15, figs. 7-8) Nominal species. *Agnostus communis Flail and Whitfield, 1877, pp. 228-229, pi. 1, figs. 28-29, late Cam- brian, Dunderberg Shale, locality unknown, Nevada, sensu Palmer (1955, pp. 94-96, pi. 19, figs. 20-21, USNM 24557). Regardless of synonymy, many specimens have been referred to Pseudagnostus communis (Hall and Whitfield), viz: ♦Palmer, 1954, pp. 720-721, pi. 76, figs. 1-3, UT 32205, USNM 123309, UT 32169. ♦Palmer, 19606, p. 61, pi. 4, figs. 3-4, USNM 136832a-b. ♦Robison, 19606, p. 14, pi. 1, figs. 2, 5, BYU 10911 -0-975a-b. ♦Rasetti, 1961, p. 109, pi. 23, figs. 13-17, USNM 143054-143055. Bell and Ellinwood, 1962, p. 389, pi. 51, figs. 7-21, UT 34842-34856. ♦Lochman, 1964, p. 47, pi. 9, figs. 32-36, USNM 140664a-e. ♦Rasetti, 1965, p. 39, pi. 10, figs. 23-25, USNM 144547. ♦Palmer, 1968, pp. 29-30, pi. 7, figs. 5, 10, USNM 136832a-b. Diagnosis. Cephalon subcircular to subovoid, partially effaced, narrow borders with non-deliquiate or subdeliquiate marginal furrows, spectaculate, anterior glabellar furrow usually transverse rectilinear, median preglabellar furrow invariably present. Pygidium subcircular to subovoid, partially effaced, subplethoid, weakly to strongly deuterolobate, narrow borders with subdeliquiate marginal furrows, spines well in advance of a transverse line drawn across the rear of the deuterolobe, eight late holaspid metameres. Other species *Agnostus coloradoensis Shumard, 1861, p. 218, cephalon USNM 26928. * Pseudagnostus convergens Palmer sensu Lochman and Hu 1959, p. 412, pi. 57, figs. 1-6, USNM 137088a-f. *Agnostus josepha Hall, 1863, p. 178, pi. 6, figs. 54-55; 1867, p. 169, pi. 1, figs. 54-55; refigured in Shimer and Shrock, 1944, pi. 251, figs. 5-6, cotypes AMNH 311. * Pseudagnostus Josephus (Hall) [szc] sensu Grant 1965, p. 108, pi. 13, figs. 13-14, USNM 142409-142410. * Pseudagnostus cf. P. laevis Palmer sensu Grant 1965, p. 108, pi. 14, figs. 34-35, USNM 142319. * Pseudagnostus latus Kobayashi, 1938, p. 171, pi. XVI, figs. 23-24, cotypes GSC 11989-11990, Inon figs. 40-41, GSC 11991-11992. * Agnostus neon Hall and Whitfield, 1877, pp. 229-230, pi. 1, figs. 26-27, cotypes USNM 24568; refigured Palmer 1955, pp. 94, 96, pi. 19, figs. 16, 19, and synonymized with communis. * Pseudagnostus orientalis Kobayashi, 1933, pp. 98-99, pi. IX, figs. 20-22, IGUT unnumbered [holotype fig. 22]; 19356, pp. 110-111, pi. Ill, figs. 7-11, 23, IGUT unnumbered. SHERGOLD: P S E U D AG NO STU S 83 *Agnostus prolongus Hall and Whitfield, 1877, pp. 230-231, pi. 1, figs. 30-31, cotypes USNM 24637; refigured Palmer, 1955, pp. 98-99, pi. 19, figs. 17, 22. * Pseudagnostus prolongus (Hall and Whitfield) sensu Lochman and Hu, 1959, pp. 412-413, pi. 57, figs. 7-16, USNM 137089a-k. * Pseudagnostus cf. prolongus (Whitfield) [sic] sensu Lochman and Hu, 1960, p. 812, pi. 95, fig. 36, USNM 138218. Pseudagnostus rotundatus Lermontova, 1940, p. 125, pi. 49, fig. 12, 12a-c, repository and numbers not known. Pseudagnostus rotundatus Lermontova sensu Pokrovskaya in Tchernysheva et at. 1960, p. 464, pi. 2, fig. 7, repository, numbers, and location (apart from Siberia) unknown. * Pseudagnostus sentosus Grant, 1965, pp. 108-109, pi. 9, figs. 2-3, 5, holotype pygidium USNM 142284, paratypes USNM 142283, 142434. * Pseudagnostus vulgaris Rosova sensu Palmer, 1968, p. 30, pi. 12, fig. 5, USNM 146845, Inon fig. 6, USNM 146661. * Pseudagnostus spp.. Palmer, 1962, p. 21, pi. 2, figs. 16, 21, 26, USNM 143147a-b, 143148. * Pseudagnostus sp., Robison and Palmer, 1968, pp. 169-170, photo 3, USNM 158031. Questionably included in the communis group are also: * Pseudagnostus gyps (Clark) sensu Rasetti, 1959, p. 381, pi. 51, figs. 13- 14, USNM 136929. Pseudagnostus cyclopygeformis (Sun) (pars) sensu Endo in Endo and Resser, 1937, p. 316, pi. 65, figs. 19-22, non pi. 68, figs. 8-16, MSM 1080, 1157, 1249, 1260, 2582 (unplaced). Age and distribution. Late Cambrian: late Dresbachian zones of Aphelaspis , Dicanthopyge, and Dunder- bergia ; Franconian zones of Elvinia, Conaspis ( Taenicephalus), Ptychaspis ( Idahoia ), and Saukia (Illaenurus), Nevada, Utah, Texas, Idaho, Montana, Wyoming, Alabama, Wisconsin, Minnesota, Maryland, Ten- nessee, Alaska, U.S.A.; Elvinia Zone, British Columbia, Canada; late Tuorian, zone of Glyptagnostus reticulatus\ Shidertan, zone of Plicatolina perlata, Yakutia, Olenek River, U.S.S.R. (Lazarenko 1966; Ivshin and Pokrovskaya 1968); Paishanian, Chuangia Zone, Liaotung and Taitzuho, Manchuria, and South Korea. Comments. A Pseudagnostus communis species group initially was recognized by Palmer (1968, p. 30), although no unifying or diagnostic characteristics were listed. The assignment of species made here differs from those listed by Palmer. As constituted here, the communis group exhibits variation in shield shape and to some extent the position of the axial glabellar node. Species intermediate between spectaculate and papilionate exist, e.g. P. communis sensu Rasetti, 1961, P. neon (Hall and Whitfield), and the specimens ascribed to P. prolongus by Lochman and Hu (1959), which appear to link the communis group morphologically to that of Rhapt- agnostus clarki (following). Within a single paradigm there is a more or less constant degree of effacement. Contractu group ( PI. 15, figs. 11-12) Nominal species. * Pseudagnostina contracta Palmer, 1962, p. F21, pi. 2, figs. 18-20, 22-25, holotype pygidium USNM 143150, paratypes USNM 143149a-b, 143151-143152, 1 43 1 53a- b, late Cambrian, Dresbachian, Cedaria Zone, Nevada, Alabama, U.S.A. Diagnosis. Cephalon subovoid to subquadrate, effaced or partially effaced, deliquiate and subdeliquiate marginal furrows, strongly spectaculate, rectilinear transverse anterior furrow, median preglabellar furrow absent. Pygidium subovoid to sub- quadrate, effaced or partially effaced, subdeliquiate marginal furrows, non-plethoid, weakly deuterolobate, spines variable but usually posteriorly positioned close to the rear of the deuterolobe. 84 PALAEONTOLOGY, VOLUME 20 Other species * Pseudagnostina vicaria Opik, 1967, pp. 158-159, pi. 55, fig. 4; pi. 63, figs. 8-9, holotype pygidium CPC 5918, paratypes CPC 5816, 5919. * Pseudagnostinal sp. indet. (aff. vicaria sp. nov.) Opik, 1967, p. 159, pi. 63, fig. 10, CPC 5920. Opik (1967, p. 150) has noted that species described as Agnostus douvillei Bergeron by Walcott (1913, p. 100, pi. VII, fig. 3, 3 a-b\ pi. XI, figs. 6-7) and Resser and Endo (in Endo and Resser 1937, p. 162, pi. 49, figs. 25-28) may belong to Pseudagnostina. A. koerferi Monke (1903, pp. 111-114, pi. 3, figs. 1-9; pi. 9; Woodward 1905, pp. 211-215, 251-255, pi. 13, fig. 5), synonymized with douvillei by Walcott ( 1913, p. 100), thus may also belong to Pseudagnostina. Wolfart (1974, p. 90) has synonymized other references to A. dou- villei (Mansuy 1916; Kobayashi 1 9357? ; Lu 1957 ; Lu et at. 1965), with his Pseudagnostus kobayashii, which is tentatively included here in the cyclopyge group. The type cephalon of A. douvillei Bergeron (1899, p. 503, pi. XIII, fig. 3, repository and number unknown), cannot be readily interpreted. The species Oedorhachis boltonensis Resser, which Palmer (1962, p. 21) regarded as belonging to Pseudagnostina, is here placed in the bulgosus species group. Age and distribution. Late Cambrian: Mindyallan, Glyptagnostus stolidotus Zone, western Queensland, Australia; Dresbachian, Cedaria Zone, Alabama, Nevada, U.S.A.; Kushanian, Drepanura-Stephanocare interval, Vietnam, China (Shantung), Manchuria (Liaotung), South Korea. Comments. This group may represent an effaced derivation from the bulgosus group. Cyclopyge group (PI. 15, figs. 1-2) Nominal species. * Agnostus cyclopyge Tullberg, 1880, p. 26, pi. II, fig. 15a, c, as interpreted by Wester- gard 1922, pp. 116-117, pi. 1, figs. 7-8 (LULo 3066t-3067t), late Cambrian, zones of Olenus and Para- bolina spinulosa with Orusia lenticularis, Andrarum, Skane, Sweden (see discussion in introduction to this paper). Of other specimens which have been assigned to this species, the following have been traced: * Agnostus cyclopyge Tullberg sensu Lake 1906, pp. 27-28, pi. II, figs. 21-22, OUM number not known and BMNH 58494 respectively; also * Pseudagnostus cyclopyge (Tullberg) sensu Rushton in Taylor and Rushton 1971, p. 26, pi. VIII, figs. 1-2, GSM Rul202, 1042. Diagnosis. Cephalon subovoid to rounded subquadrate, en grande tenue, wide borders with strongly deliquiate marginal furrows, spectaculate, subovoid to sub- circular acrolobe, transverse rectilinear anterior furrow in early representatives becoming curved backwards or V-form in later ones, median preglabellar furrow present. Pygidium subovoid to rounded subquadrate, en grande tenue , wide borders with strongly deliquiate marginal furrows, plethoid, deuterolobate, spines generally well forwards of a transverse line across the rear of the deuterolobe. Other species * Pseudagnostus ampullatus Opik, 1967, p. 150, pi. 61, figs. 7-1 1, holotype pygidium CPC 5896, paratypes CPC 5897-5900. Pseudagnostus angustilobus Ivshin, 1956, pp. 19-21, pi. 9, figs. 11-15, 18-23, holotype cephalon GMAN 69/926, paratypes GMAN 73/926, 68/926, 72/926, 78/926, 79/926, 85/926, 84/926, 83/926, 123/926, 84/926. Agitostus chinensis Dames, 1883, pp. 27-28, pi. 2. figs. 18-19; Kobayashi 19376, p. 434, pi. 17, fig. 14 a-b, material destroyed; Schrank 1974, pp. 622-623, pi. 1, figs. 1-7, MNHU K302, T893.2, 893.3, 894.1, 895.1 ; non Agnostus chinensis sensu Walcott, 1913, pp. 99- 100, pi. 7, figs. 4-5; non Pseudagnostus chinensis (Dames) sensu Lu et al. 1965, p. 41, pi. 4, figs. 3-5 = Peronopsis rakuroensis (Kobayashi) fide Kobayashi (19376, p. 434). Pseudagnostus communis (Hall and Whitfield) sensu Lu et al. 1965, pp. 41-42, pi. 4, figs. 6-8, repository and numbers not known. Homagnostus convexus Chu, 1959, pp. 88-89, pi. 1, figs. 3-4, non figs. 1-2, 5-7, also in Lu et al. 1965, p. 20, pi. l,fig. 10, non figs. 8-9, IPP 941 1-9412. SHERGOLD: PSEUD AGNOSTUS 85 *Pseudagnostus cf. cyclopyge (Tullberg) sensu Whitehouse, 1936, p. 100, pi. X, fig. 8, UQ F3188; prob- ably equivalent to Pseudagnostus cf. vastulus Whitehouse sensu Opik, 1963, pp. 50-53, text-fig. 13, CPC 4302. * Pseudagnostus idalis Opik, 1967, p. 153, pi. 62, figs. 8-9; pi. 63, figs. 1, 3, holotype pygidium CPC 5908, paratypes CPC 5909-5911, 5913. * Pseudagnostus cf. idalis Opik, 1967, p. 154, pi. 63, fig. 4, CPC 5914. *Plethagnostus jarillensis Rusconi, 1953, p. 4 [nom. nud.\ \ 1954, pp. 19-20, fig. 6, holotype pygidium NHMN 16674. * Pseudagnostus leptoplastorum Westergard, 1944, p. 39, pi. 1, fig. 1, holotype pygidium SGU C459; also P. leptoplastorum Westergard sensu Ivshin, 1962, pp. 16-18, pi. I , figs. 8-18, GMAN 646/105-107, 1 10, 113-117, 120, 124. * Pseudagnostus marginisulcatus Kobayashi, 1962, p. 32, pi. Ill, figs. 10-11, holotype cephalon, paratype pygidium IGUT, unnumbered. * Pseudagnostus nuperus Whitehouse, 1936, p. 100, pi. X, figs. 5-7, holotype cephalon UQ F3199, para- types UQ F3200-3201. *Agnostus obtusus Belt, 1868, pp. 10-11, pi. II, figs. 15-16 [fig. 15, not located, fig. 16 BMNH 58494], * Pseudagnostus primus Kobayashi, 19356, pp. 108-109, pi. XIV, figs. 6-10, IGUT, unnumbered (fig. 6 designated holotype cephalon); also Kobayashi, 1962, pp. 31-32, pi. Ill, figs. 15-17, pi. V, figs. 8-12, IGUT, unnumbered. Pseudagnostus pseudocyclopyge Ivshin, 1956, pp. 17-19, pi. 1, figs. I 8, 10, 16-17, holotype cephalon GMAN 74/926, paratypes GMAN 64/926, 66/926, 121/926, 70/926, 75/926, 83/926, 80/926, 77/926, 87/926, 86/926; Ivshin 1962, p. 18, pi. 1, figs. 19-22, GMAN 646/217, 646/125, 646/123. * Pseudagnostus sericatus Opik, 1967, pp. 152-153, pi. 62, fig. 7, holotype cephalon CPC 5907. * Pseudagnostus sp. undet., Shergold in Shergold et al. 1976, pp. 264-265, pi. 41, figs. 9 II, NZAR 601-603. The cyclopyge group possibly also includes: Pseudagnostus kobayashii Wolfart, 1974, pp. 90-93, pi. 10, fig. 8; pi. 1 1, figs. 3-7, HAN 82/1, 82/2, 83/1, 84/2 (synonymy given). *Oedorhachis tennesseensis Resser, 1938, p. 50, pi. 10, figs. 24-26, cotypes USNM 94871. * Pseudagnostus vastulus Whitehouse, 1936, pp. 99-100, pi. X, figs. 3-4, holotype pygidium UQ F3203, paratype UQ F3202. Pseudagnostus sp., Lu 1956h, pp. 367-368, pi. 1, figs. 1-2, ?IPP 8643-8644; also in Lu et al. 1965, p. 43, pi. 4, figs. 18-19, same repository and numbers. Age and distribution. Late Cambrian : zones of Parabolina spinulosa , P. brevispina, Leptoplastus rhaphi- dophorus , and Peltura scarabaeoides , United Kingdom, Sweden; late Tuorian, Glyptagnostus reticidatus Zone; early Shidertan, Irvingella Zone, central Kazakhstan (Ivshin and Pokrovskaya 1968); Kushanian, Blackwelderia paronai fauna, Manchuria; late Kushanian, Afghanistan; Paishanian, Chuangia Zone, China (Kweichow), Eochuangia Zone, South Korea; late Mindyallan, G. stolidotus Zone, Idamean zones of G. reticulatus with Proceratopyge nectans, Corynexochus plumula , Erixanium sentum, and Irvingella tropica with Agnostotes inconstans (Opik, 1963, 1967), western Queensland, Australia. Species of the group also occur in South America (Argentina), where their age is uncertain; in Antarctica (northern Victoria Land) of probable late Idamean age (Shergold et at. 1976); and possibly in North America (Ten- nessee), early Dresbachian, Blount ia Zone (Resser 1938). Comments. The cyclopyge group is heterogeneous and probably capable of further division. Its members exhibit some variation in shield shape, degree of effacement, orientation of anterior glabellar furrow, extent of pygidial pleural lobes, and strength, shape, and orientation of the pygidial spines. Early species are distinctly spectacu- late, but later Cambrian ones may begin to approach the papilionate condition, in that the axial glabellar node migrates forwards and begins to interfere with the anterior furrow without actually coming to rest between the anterolateral lobes. The cyclopyge group grades with effacement into the communis group. 86 PALAEONTOLOGY, VOLUME 20 Securiger group (PI. 15, fig. 13) Nominal species. *Agnostus securiger Lake, 1906, p. 20, pi. II, fig. 11, GSM 57650, late Cambrian, Olenus Zone, Outwood Shales, Chapel End, near Nuneaton, Warwickshire, U.K. Diagnosis. Cephalon subovoid to subquadrate, en grande tenue, spectaculate, deli- quiate marginal furrows, subovoid acrolobe, transverse rectilinear anterior glabellar furrow, median preglabellar furrow present. Pygidium subovoid to subrounded, en grande tenue , deliquiate marginal furrows, strongly deuterolobate, plethoid, postero- lateral spines in advance of a line drawn across the rear of the deuterolobe plus a third sagittal spine developed from the posterior margin. Other species. No other species are described, although a trispinose pygidium of pseudagnostid type is present in the Elvinia Zone assemblages of Cherry Creek, Egan Range, Nevada, U.S.A., USGS collection C02527 (A. R. Palmer, pers. comm.). Age and distribution. Late Cambrian: late Olenus Zone, Warwickshire, U.K. ; early Franconian, Elvinia Zone, Nevada, U.S.A. Comments. The generic name Sulcatagnostus was erected by Kobayashi (1937a, p. 451) for Agnostus securiger (see below). Apart from its possession of a third, sagittal, pygidial spine, Sulcatagnostus securiger (Lake) compares well with mem- bers of the cyclopvge group such as Pseudagnostus ampullatus Opik and P. idalis Opik, both of which have similar over-all morphology. B. PAPILIONATE SPECIES GROUPS Clarki group (Shergold 1975, p. 60) (PI. 15, figs. 14-15) Nominal species. * Pseudagnostus ( Plethagnostus ) clarki Kobayashi, 1 935a, p. 47, pi. IX, figs. 1-2, holotype pygidium USNM 93062, paratype USNM 146887, late Cambrian ‘ Briscoia ' fauna. Hard Luck Creek, Alaska; refigured Palmer 1968, p. 29, pi. 15, figs. 10, 13-14. Diagnosis. Cephalon subovoid to subcircular, effaced, wide borders with non- deliquiate marginal furrows, papilionate, V-form anterolateral glabellar furrows, median preglabellar furrow externally effaced, present parietally. Pygidium subovoid, effaced, wide borders with non-deliquiate marginal furrows, non-plethoid, weakly deuterolobate, small posterolateral spines sited well in advance of a transverse line drawn across the rear of the deuterolobe, nine to ten axial metameres. Other species * Pseudagnostus clarki maximus Shergold, 1975, pp. 70-71, pi. 5, figs. 1-2, holotype cephalon CPC 11587, paratype CPC 1 1588. * Pseudagnostus clarki patulus Shergold, 1975, pp. 62-64, pi. 1, figs. 1-6, pi. 2, figs. 1-2, holotype cephalon CPC 1 1 524, paratypes CPC 1 1 525- 1 1531. * Pseudagnostus clarki prolatus Shergold, 1975, pp. 64-69, pi. 3, figs. 1 -6, pi. 4, figs. 1 6, holotype cephalon CPC 1 1 532, paratypes CPC 1 1 533- 11542. * Pseudagnostus cyclopygeformis (Sun) sensu Kobayashi, 1933, pp. 97-98, pi. IX, figs. 19, 23-24, pi. X, fig. 7; 19356, pp. 1111 12, pi. Ill, figs. 12-14, IGUT no numbers. * Pseudagnostus elix Shergold, 1975, pp. 71-73, pi. 2, figs. 3-7, holotype pygidium CPC 1 1688, paratypes CPC 11689-11691. * Pseudagnostus laevis Palmer, 1955, pp. 97-98, pi. 19, figs. 8-9, 11-12, holotype pygidium USNM 123559, paratypes USNM 26990, 123560-123561. SHERGOLD: PSEUD AGNOSTUS 87 * Pseudagnostus orbiculatus Shergold, 1975, pp. 73-74, pi. 12, figs. 8-12, holotype cephalon CPC 11591, paratypes CPC 11592-11595. * Pseudagnostus papilio (pars) Shergold, 1972, pp. 28-31, pi. 1, fig. 2, CPC 8443, non pi. 1, figs. 1, 3-8, pi. 2, figs. 1-2 [convergens group]. *Peronopsis planulata Raymond, 1924, p. 395, pi. 12, fig. 9, holotype pygidium MCZ 1729. * Pseudagnostus cf. prolongus (Hall and Whitfield) sensu Kindle and Whittington, 1965, p. 686, pi. 1, figs. 17-20, Kindle Collection. * Pseudagnostus sp. //, Shergold 1972, p. 35, pi. 2, figs. 6-7, CPC 8458-8459. * Pseudagnostus sp. Ill, Shergold 1972, pp. 35-36, pi. 2, fig. 8, CPC 8460. * Pseudagnostus sp., Robison and Pantoja-Alor 1968, p. 780, pi. 97, figs. 17-18, USNM 158884-158885. Possibly also belonging to this species group are : Pseudagnostus a. sp., Kobayashi 1935a, p. 41, plus plate explanation, pi. VIII, fig. 3, repository and number not known. Pseudagnostus (3 sp., Kobayashi 1935a, p. 41, plus plate explanation, pi. VIII, fig. 4, repository and number not known. Age and distribution. Late Cambrian : pre-Payntonian and Payntonian, Rhaptagnostus clarki patulus with Caznaia squamosa through Neoagnostus quasibilobus with Tsinania nomas Assemblage-Zones (Shergold 1975), western Queensland, Australia; Trempealeauan Saukia Zone, Saukiella pyrenel and S. serotina Subzones, Nevada, Montana, Vermont, Alaska, U.S.A.; Trempealeauan?, Oaxaca Province, Mexico; late Changshanian, Kaolishania Zone, Shantung, China; Fengshanian, Tsinania Zone, North and South Korea. Comments. The clarki group is a homogeneous association of taxa exhibiting specific and subspecific variation in shapes and proportions, and border parameters. Convergens group (Shergold 1975, p. 74) (PI. 16, figs. 1-2) Nominal species. * Pseudagnostus convergens Palmer, 1955, pp. 96-97, pi. 19, figs. 14-15, holotype pygidium USNM 123562, paratype USNM 123563, late Cambrian, Trempealeauan, Saukia Zone, Saukiella pyrene Subzone, Nevada, U.S.A. Diagnosis. Cephalon subovoid to subcircular, effaced and partially effaced, narrow borders with non-deliquiate marginal furrows, papilionate, V-form anterolateral furrows, median preglabellar furrow externally effaced, present parietally. Pygidium subovoid, with rearwards converging flanks and acrolobe, narrow borders with non-deliquiate marginal furrows, subplethoid, weakly deuterolobate, very small posterolateral spines well in advance of a transverse line drawn across the rear of the deuterolobe, ten late holaspid metameres. Other species * Pseudagnostus bifax Shergold, 1975, pp. 75-79, pi. 9, figs. 1-7, holotype cephalon CPC 11596, paratypes CPC 11597-11602, 11649, 11656, 11662, 11667 11668. Agnostus cyclopygeformis Sun, 1924, pp. 26-28, pi. II, fig. \a-h, GSCh 501-504, 507-510; 1935, p. 16, pi. Ill, figs. 29-32, NUP 1 194-1 197; also * Pseudagnostus cyclopygeformis (Sun) sensu Endo, 1939, p. 6, pi. 1, figs. 14-15, USNM 96092b-c, non fig. 13, USNM 96092a; P. cyclopygeformis (Sun) sensu Lu et al., 1957, pi. 137, figs. 20-21, repository and numbers not known; 1965, p. 42, pi. 4, figs. 9-12, repository and numbers not known. Pseudagnostus obsoletus Lermontova, 1951, pp. 10-11, pi. 2, figs. 8-10, repository and numbers unknown. Pseudagnostus cf. obsoletus Lerm. (MS), Lermontova, 1940, p. 125, pi. 49, fig. 1 1, repository and number not known. * Pseudagnostus papilio Shergold, 1972, pp. 28-31, pi. 1, figs. 1, 3-8, non fig. 2 [ clarki group], pi. 2, figs. 1-2, holotype cephalon CPC 8442, paratypes CPC 8443-8450; also 1975, pp. 79-82, pi. II, figs. 1 8, CPC 11669-11677. PALAEONTOLOGY, VOLUME 20 Possibly also belonging to this group is: * Pseudagnostus sp. /, Shergold 1972, p. 34, pi. 2, figs. 3-5, CPC 8457. Age and distribution. Late Cambrian : pre-Payntonian, Rhaptagnostus bifax with Neoagnostus denticulatus and R. clarki maximus with R. papilio Assemblage-Zones (Shergold 1975), western Queensland, Australia ; late Changshaman, Kaolishania Zone, Shantung, Hopei, China; Trempealeauan, Saukiella pyrene Sub- zone, Saukia Zone, Nevada, U.S.A. ; Shidertan, Lotagnostus trisectus/Peltura Zone, Kazakhstan, U.S.S.R. Comments. As for clarki group. PSEUDAGNOST1D SPECIES UNASSIGNED * Pseudagnostus cf. P. convergens Palmer sensu Lochman, 1964, p. 51, pi. 13, fig. 10, USNM 140686. Agnostus cyclopyge Tullberg sensu Sun, 1935, pp. 15-16, pi. Ill, figs. 33-36, THU 1198-1201; sensu Lu et al. 1965, pp. 40-41, pi. 4, figs. 1-2 [reproductions of Sun’s figures]. Pseudagnostus cyclopyge (Tullberg) sensu Wilson, 1954, p. 284, pi. 25, fig. 19, pi. 26, fig. 13. *lHypagnostus empozadensis Rusconi, 1953, p. 6 [nom. nud.]\ 1954, pp. 34-35, pi. II, fig. 11, NHMM 16872, an effaced papilionate subquadrate cephalon. *Leiopyge empozadense Rusconi, 1953, pp. 5-6 [nom. nud.\\ 1954, pp. 33-34, pi. II, fig. 10, NHMM 16856, an incomplete pygidium. *Oedorhachis greendalensis Resser, 1938, p. 51, pi. 10, fig. 9, USNM 94861. * Plethagnostus gyps C lark, 1923, p. 124, pi. 1, fig. 9, holotype pygidium MCZ 1700; refigured Clark, 1924, p. 16, pi. 3, fig. 2; sensu Rasetti, 1944, p. 234, pi. 36, figs. 20-22, LU 1105a-c [may represent a distinct en grande tenue papilionate species group], Pseudagnostus impressus Lermontova, 1940, p. 125, pi. 49, fig. 13, 13a, repository and numbers not known. * Pseudagnostus jeholensis Kobayashi, 1951, pp. 76-77 , pi. 7, figs. 13-14, IGUT no numbers; reproduced in Lu et al. 1965, p. 43, pi. 4, figs. 16-17. Pseudagnostus joseplia (Hall) as illustrated by: Frederickson 1949, p. 362, pi. 72, fig. 17, OGM 105-16F-53; Lochman 1950, pp. 329-330, pi. 46, fig. 14; Nelson 1951, p. 776, pi. 107, fig. 5; Bell et al. 1952, pp. 196- 197, pi. 32, fig. 4 a-b, pi. 33, fig. 1 ; Ellinwood 1953, p. 65, pi. 4, figs. 9-10; Wilson 1954, P. 284, pi. 25, figs. 5, 22. * Agnostus maladensis Meek, 1873, p. 464, USNM 24597 [a medley of taxa]. * Pseudagnostus mesleri (Resser) sensu Lochman, 1940, pp. 26-27 , pi. 2, figs. 38-43, USNM 98703. Microdiscus paronai Airaghi, 1902, p. 23, pi. 2, figs. 24-25, repository and numbers not known. * Spinagnostus pedrensis Rusconi, 1951, pi. 8, fig. 9, NHMN 9963 [referred to Leiopyge by Rusconi 1953, p. 6, originally illustrated as a cephalon, this specimen is actually an indeterminate pseudagnostid pygidium], Agnostus pii Airaghi, 1902, p. 19, pi. 1, fig. 28, repository and number not known. * Pseudagnostus prolongus ( Hall and Whitfield) sensu Palmer, 1960, p. 61, pi. 4, figs. 5-6, USNM 1 36833a-b. Pseudagnostus ( Rhaptagnostusl ) semiovalis Kobayashi, 1937a, pp. 452-453, pi. II, figs. 8-9, repository Freiburg, types destroyed (Kobayashi pers. comm., 1972). ? Pseudagnostus sp., Lu 1956a, pp. 282-283, pi. 1, fig. 8; Lu et al. 1965, p. 44, pi. 4. fig. 20, repository and number not known. PSEUDAGNOSTI NOMINA NUDA Pseudagnostus ovatus Rusconi, 1950, p. 94 [plate explanation error, fig. 6, for P. parabolicus], Pseudagnostus huangluoensis Kobayashi, 1951, p. 75. Pseudagnostus minis Pokrovskaya in Vasilenko, 1963, p. 22, Chart 3 [listed], Pseudagnostus solus Endo in Endo and Resser, 1937, p. 304 [listed]. SHERGOLD: PSEUD AGNOSTUS 89 DEFINITION OF GENERIC GROUPS Since Jaekel erected Pseudagnostus in 1909 several attempts have been made to subdivide the genus, e.g. Clark 1923, Whitehouse 1936, Kobayashi 1937c/, 1955, Lermontova 1940, inter alia. Accordingly there is a proliferation of generic and subgeneric names, some of which have doubtful taxonomic validity. These are briefly reviewed below in discussing the generic classification of pseudagnosti. In this classification Pseudagnostus Jaekel, 1909, type species Agnostus cyclopyge Tullberg, 1880 (p. 26, pi. II, fig. 15c/, c), is restricted to include only the cyclopyge , communis , and bulgosus species groups as documented above. Plethagnostus , erected by Clark (1923, p. 124), originally differentially diagnosed from Pseudagnostus by having ‘divergent dorsal [accessory] furrows continued to the border of the pygidium’ as in Plethagnostus gyps Clark (1923, pi. I, fig. 9), its type species, is suppressed. This action is taken because: Many pseudagnostid species are plethoid, the courses of the accessory furrows continuing to the marginal furrows, particularly in en grande tenue and partially effaced species groups, e.g. the cyclopyge and communis group referred above to Pseudagnostus ; the type species is based on a single incomplete pygidium (MCZ 1700) which lacks a posterior margin ; and usage of the name Plethagnostus has been con- fused by Kobayashi (1935//, pp. 46-47), who utilized Clark’s taxon in a subgeneric sense for the species Pseudagnostus clarki , which has a papilionate cephalon and non-plethoid pygidium with effaced accessory furrows! Subsequent usage of Pleth- agnostus appears to have followed Kobayashi’s interpretation. It is recommended that this practice be discontinued because the type specimen of P. gyps quite obviously does not belong to the same species group as Pseudagnostus clarki , and because of its incompleteness cannot be adequately classified anyway. Plethagnostus gyps Clark is left unclassified with respect to species groups and is referred to Pseud- agnostus s.l. Rhaptagnostus Whitehouse (1936, p. 97), based on A. cyclopygeformis Sun, 1924, was differentiated from Pseudagnostus by possessing ‘an elliptical arrangement of foramina [notulae] in the post-axial region’ of the pygidium, and a ‘simple, non- spinose brim’. Sun’s specimens (1924, pi. II, fig. 1 a-h) appear to be parietal surfaces, and such surfaces of many pseudagnosti display notular lines given the right medium of preservation. The absence of pygidial spines in any pseudagnostid species is greatly doubted. Very often these spines are extremely small and readily destroyed by certain modes of preservation, e.g. in arenaceous matrices, and by negligent preparation. Cephala matched by Sun with the pygidia which Whitehouse chose to differentiate from Pseudagnostus, are papilionate and thus different from those of P. cyclopyge (Tullberg), which is spectaculate. P. cyclopygeformis (Sun) is here included within the convergens species group. Rhaptagnostus is used herein to dis- tinguish the two closely related papilionate species groups, clarki and convergens, from spectaculate ones at the generic level. This name is preferred to the previously and erroneously used Plethagnostus Clark sensu Kobayashi 1935a. Sulcatagnostus was erected by Kobayashi (1937a, p. 451) for A. securiger Lake (1906, p. 20, pi. II, fig- 11), and distinguished by possessing ‘irregular divergent furrows on the side lobes’ (Kobayashi, 1937a, pp. 450-451). This was later redefined 90 PALAEONTOLOGY, VOLUME 20 as ‘Pseudagnostinae with reticulated furrows on side lobes’ (Kobayashi 1939, p. 159). The reticulation described (PI. 15, fig. 13) refers to the caecal network of the cephalic acrolobe and pygidial pleural lobes. The presence of a caecal display, common on exfoliated specimens and thin exoskeletons under certain conditions of preservation, is itself insufficient to justify differentiation from Pseudagnostus. Over- looked by Kobayashi, however, is the fact that A. securiger Lake also possesses a trispinose pygidium (Rushton in Taylor and Rushton 1971, p. 20), which certainly is a significant characteristic. In view of the otherwise similar morphology to spec- taculate pseudagnosti, Sulcatagnostus is retained as a subgenus of Pseudagnostus. Euplethagnostus Lermontova (1940, p. 126), based on E. subangulatus Lermontova (1940, pi. XLIX, fig. 15, 15tf), was compared to Plethagnostus Clark, 1923, but distinguished by the possession of posterolateral pygidial spines, sometimes an intranotular axis (lanceolate ridge), and the presence of a posterior pygidial terminal node. Such characteristics are emphatically not diagnostic in generic classification because: the type specimen of P. gyps , type species of Plethagnostus , is damaged and it is not possible to ascertain whether it possessed spines— Clark (1923) assumed that their traces were still apparent, and Rasetti (1944, pi. 36, figs. 20-22) has illus- trated material possessing pygidial spines which can be assigned to P. gyps ; the presence of terminal nodes and intranotular axes is commonly observed on a multi- plicity of pseudagnostid parietal surfaces; and Lermontova’s specimens are poorly illustrated and inadequately described, no type specimen was selected for E. sub- angulatus and the repository of material is unknown. E. subangulatus appears most similar to species classified herein with the clams group. Its type specimens require redescription and re-illustration if the name Euplethagnostus is to be retained. Problems associated with the taxonomic validation of Pseudorhaptagnostus Lermontova (1940, p. 126) have been previously discussed (Shergold 1972, p. 28). These problems remain unresolved. The name Pseudorhaptagnostus was introduced by Lermontova in 1940, P. simplex Lermontova (1951, pp. 12-13, pi. 2, figs. 11-17) being designated as type species. Although a second species, P. punctatus Lermontova, was illustrated in 1940, P. simplex was neither illustrated nor described until 1951. At this time a heterogeneous collection of cephala and pygidia was figured. While the pygidia (Lermontova 1951, pi. 2, figs. 11-14) show morphological consistency, two types of cephalon were figured: one (fig. 15) is probably that associated with the pygidia, whereas the others (figs. 16-17) belong to a species of the convergens group. Nikitin (1956, pi. XIV, figs. 4-5) re-illustrated this combination of convergens group cephalon with Pseudorhaptagnostus pygidium. Such mismatching appears to result directly from Lermontova’s failure to designate a holotype. According to Lermontova (1940), Pseudorhaptagnostus is differentiated from Rluiptagnostus Whitehouse in possessing a thickened pygidial rim and well-developed posterolateral spines. A lanceolate field [intranotular axis] is stated to be present on pygidial internal casts [parietal surfaces]. None of these characteristics is considered to justify separa- tion from Pseudagnostus , either separately or in combination. Other characteristics exhibited by Pseudorhaptagnostus simplex can, however, be utilized if the name is to be retained (Shergold 1972, p. 28). These, which can only be verified from examina- tion of the actual material on which Pseudorhaptagnostus is based, involve the cephalic and pygidial shapes, nature of their borders and pygidial spines, and pygidial seg- SHERGOLD: PSEUD AGNOSTUS 91 mentation. Although it appears that P. simplex has much in common with the clavus species group recognized above, judgement on the validity of the genus must await clarification of the concept of the taxon. It is very likely that Pseudorhaptagnostus and Euplethagnostus Lermontova are synonyms. If so, Pseudorhaptagnostus has priority, being first listed. Neoagnostus Kobayashi (1955, p. 473), type species N. aspidoides Kobayashi (1955, pp. 473-474), possesses a cephalon with trilobed glabella and median preglabellar furrow (PI. 16, fig. 16). The holotype cephalon (Kobayashi 1955, pi. VII, fig. 5) is a parietal mould preserved en grande tenue. The associated pygidium (fig. 4), labelled paratype, is geragnostoid. Hyperagnostus Kobayashi (1955, p. 474), based on H. binodosus Kobayashi (1955, p. 475), also has a trilobed glabella, but is said to differ from Neoagnostus in not possessing a median preglabellar furrow (PI. 16, fig. 17). The holotype cephalon (1955, pi. VII, fig. 2) is partially exfoliated, somewhat distorted, and incomplete, the anterior portion of the cephalon having been lost. A thin veneer of exoskeleton lies across the position in which the median preglabellar furrow would be expected and its presence or absence cannot be absolutely verified ; it appears not to be present on the external surface, but may be weakly present on the parietal surface. The assigned pygidium (1955, fig. 3) is agnostoid rather than pseudagnostoid. Although Neoagnostus and Hyperagnostus occur at different localities and are of slightly different ages, they are temporarily synonymized herein, Neoagnostus taking priority. Machairagnostus Harrington and Leanza (1957, p. 63), based on M. tmetus Harrington and Leanza (1957, p. 64, fig. 7), is represented by parietal surfaces. The glabella is faintly trisegmented, but the cephalic acrolobe is scrobiculate. The pygidium is weakly deuterolobate, and the muscle-scar impressions of the third axial segment are incorporated into the anterior portion of the axis which is delineated by axial furrows. M. tmetus has the glabellar morphology of members of the araneave- latus species group, and pygidial characteristics of the bilobus group with which it is classified here. The scrobiculation of the cephalic acrolobe is non-diagnostic under the conditions of this classification, and so Machairagnostus is synonymized with Neoagnostus. Pseudagnostina Palmer (1962, pp. 20-21), type species P. contracta Palmer (1962, pp. 20-21, pi. 2, figs. 18-20, 22-25), was coined for pseudagnosti having ‘ Peronopsis- like cephalon and Pseudagnostus- like pygidium'. The concept of the taxon is broadened somewhat here by the inclusion of some Asian species previously referred to Agnostus douvillei Bergeron, 1899. Because of over-all basic similarity of the pygidium to that of other pseudagnosti, Pseudagnostina is regarded here as a subgenus of Pseudagnostus effaced and partially effaced, strongly spectaculate, weakly deuterolobate species constituting the contracta species group. SUMMARY Spectaculate species groups fall readily into two larger groupings: 1. The bulgosus-communis-contracta-cyclopyge-securiger grouping whose mor- phology differs by degree and which can largely be encompassed within a concept of Pseudagnostus based on its type species P. cyclopyge (Tullberg 1880). Effaced 92 PALAEONTOLOGY, VOLUME 20 variants are differentiated as the contracta group regarded as a distinct subgenus, Pseudagnostina Palmer, 1962; and the securiger group, with three pygidial spines, likewise is separated at the subgeneric level as Sulcatagnostus Kobayashi, 1937a. 2. The araneavelatus-bilobus-canadensis-clcivus grouping is united by an arrange- ment of glabellar lobation and furrowing different from that of Pseudagnostus. The anterolateral lobes are closer together and may meet adaxially so that the glabellar furrows intersect as a cross. Within this division species groups are differentiated by shield shape, and degree of effacement, particularly by which glabellar furrows are effaced and which are not. Five generic names appear to be available for classifying these groups: Pseudorhaptagnostus Lermontova, 1940; Euplethagnostus Lermontova, 1940; Neoagnostus Kobayashi, 1955; Hyperagnostus Kobayashi, 1955; and Machair- agnostus Harrington and Leanza, 1957. As indicated above, Pseudorhaptagnostus and Euplethagnostus may be synonyms, the former taking priority; and Hyper- agnostus and Machairagnostus are regarded as synonyms of Neoagnostus. Although Pseudorhaptagnostus and Euplethagnostus have priority over Neo- agnostus, their concepts are very poorly understood, and it has been found not possible to obtain adequate information to verify their characteristics. Accordingly, Neoagnostus is temporarily adopted in this classification, in preference to either of the Russian genera, for the araneavelatus-bilobus- canadensis- clavus grouping. The papilionate species groups, clarki and convergens, are placed here in Rhapt- agnostus Whitehouse, 1936, because its interpretable type species is representative of the convergens group. Thus the following classification is adopted: Family Diplagnostidae Whitehouse, 1936, emend. Opik, 1967. Subfamily Pseudagnostinae Whitehouse, 1936. Genus Pseudagnostus Jaekel, 1909. Type species: Agnostus cyclopyge Tullberg, 1880 (designated Jaekel, 1909). Pseudagnostinae with long (sag.) anteriorly rounded or pointed anterior glabellar lobe, axial node lying behind anterolateral lobes and anterior furrow which is straight or curved rearwards sagittally. Shields are subcircular to subovoid, effaced to en grande tenue , non-deliquiate to deliquiate, weakly to strongly deuterolobate. Up to eight late holaspid pygidial metameres. Subgenus Pseudagnostus Jaekel, 1909. Type species and diagnosis as above. Group bulgosus, based on Pseudagnostus bulgosus Opik, 1967. Group communis, based on Agnostus communis Hall and Whitfield, 1877. Group cyclopyge, based on the type species. Subgenus Pseudagnostina Palmer, 1962. Type species: Pseudagnostina contracta Palmer, 1962 (by original designation). Spectaculate Pseud- agnostinae with subquadrate shields and with effaced median preglabellar furrow, accessory lurrows, and deuterolobe. Group contracta, based on the type species. Subgenus Sulcatagnostus Kobayashi, 1937a. Type species: Agnostus securiger Lake, 1906 (designated Kobayashi, 1937a). Spectaculate en grande tenue Pseudagnostinae with trispinose pygidium. Group securiger, based on the type species. SHERGOLD: PSEUDAGNOSTUS 93 Genus Neoagnostus Kobayashi, 1955. Type species: Neoagnostus aspidoides Kobayashi, 1955 (by original designation). Spectaculate Pseud- agnostinae with anterolateral glabellar lobes close together or meeting adaxially, and a tendency to efface or over-deepen either the furrows in front of or behind the anterolateral lobes; anterior lobe small and rhomboid. Pygidium with distinct tendency to incorporate a third segment into that portion of the axis defined by axial furrows ; retral spines. Generally, species have subquadrate shields, whose external morpho- logy may be effaced, partially effaced or en grande tenue , non-deliquiate to deliquiate, weakly to strongly deuterolobate. Up to eight late holaspid pygidial metameres. Group araneavelatus , based on Pseudagnostus araneavelatus Shaw, 1951. Group bilobus, based on Pseudagnostus bilobus Shaw, 1951. Group canadensis, based on Agnostus canadensis Billings, 1860. Group clavus , based on Pseudagnostus clavus Shergold, 1972. Genus Rhaptagnostus Whitehouse, 1936. Type species: Agnostus cyclopygeformis Sun, 1924 (designated Whitehouse, 1936). Papilionate Pseud- agnostinae with subovoid shields, externally generally effaced or partially effaced. Pygidia with ten late holaspid metameres; spines advanced with respect to the rear of the deuterolobe, and minute. Group clarki , based on Pseudagnostus clarki Kobayashi, 1935a. Group convergens, based on Pseudagnostus convergens Palmer, 1955. Non-investigated genera of Pseudagnostinae are Litagnostus Rasetti, 1944, Xestagnostus Opik, 1967, and Oxyagnostus Opik, 1967. Acknowledgements. The author acknowledges all those Collection Managers, Curators, and other persons who gave him information and allowed him to study and replicate museum specimens in Europe, North America, Japan, and Australia. Mr. H. M. Doyle was responsible for the photography. The paper was critically evaluated by Miss Joyce Gilbert-Tomlinson, Dr. A. A. Opik, and Dr. K. S. W. Campbell, and is published with the permission of the Director, Bureau of Mineral Resources, Canberra, Australia. REFERENCES airaghi, c. 1902. Di alcuni trilobiti della Cina. Atti Soc. ital. Sci. nat. 41, 17-27, pi. 1. bell, w. c. and ellinwood, h. l. 1962. Upper Franconian and Lower Trempealeauan Cambrian trilobites and brachiopods, Wilberns Formation, central Texas. J. Paleont. 36, 385-423, pis. 51-64. — feniak, o. w. and kurtz, v. e. 1952. Trilobites of the Franconia Formation, southeast Minnesota. Ibid. 26, 175-198, pis. 29-38. belt, T. 1868. On the 'Lingula Flags’, or ‘Festiniog Group’ of the Dolgelly district. Part III. Geol. Mag. 5, 5-11, pi. II. Bergeron, J. 1899. Etude de quelques trilobites du Chine. Bull. Soc. geol. Fr. 27, 499-519. billings, E. 1860. On some new species of fossils from the limestone near Point Levi opposite Quebec. Proc. Can. Nat. Geol. 5, 301-324, 30 figs. chu chao-ling. 1959. Trilobites from the Kushan Formation of north and northeastern China. Mem. Acad. sin. Inst, palaeont. 2, 44-80 [in Chinese], 81 128 [in English], pis. I-VII. Clark, T. H. 1923. A group of new species of Agnostus from Levis, Quebec. Can. Fid. Nat. 37, 121-125. — 1924. The paleontology of the Beekmantown Series at Levis, Quebec. Bull. Am. Paleont. 10 (41), 1-134, 9 pis. dames, w. 1883. Cambrische Trilobiten von Liautung. In Richthofen, f. f. von, pp. 1-33, pis. I— II. ellinwood, h. l. 1953. Late Upper Cambrian and Lower Ordovician faunas of the Wilberns Formation in central Texas. Unpubl. PhD. Thesis, Univ. of Minnesota. endo, r. 1937. See endo, r. and resser, c. e. 1937. — 1939. Cambrian fossils from Shantung. Jubilee Publ. Comm. Prof. Yabe's 60th Birthday, 1-1 8, pis. 1-2. — and resser, c. e. 1937. The Sinian and Cambrian formations and fossils of southern Manchukuo. Bull. Manchur. Sci. Mus. 1, 1-474, pis. 14-73. frederickson, e. a. 1949. Trilobite faunas of the Upper Cambrian Honey Creek Formation. J. Paleont. 23, 341 363, pis. 68-72. 94 PALAEONTOLOGY, VOLUME 20 grant, r. e. 1965. Faunas and stratigraphy of the Snowy Range Formation (Upper Cambrian) in south- western Montana and northwestern Wyoming. Mem. geol. Soc. Am. 96, 1-171, pis. 5-15. hall, j. 1863. Preliminary notice of the fauna of the Potsdam Sandstone; with remarks upon the pre- viously known species of fossils and descriptions of some new ones, from the sandstone of the Upper Mississippi Valley. 16th A. Rep. N. Y. Cab. Nat. Hist. App. D. Contributions to Palaeontology, 1 19-184, pis. Va-XI. 1867. As above, reprinted in: Trans Albany Inst. 5, 93-195, pis. I-VI. — and whitfield, r. p. 1877. Palaeontology, part II. In king, c., pp. 198-302, pis. I-VIII. hallam, a. (ed.). 1973. Atlas of Palaeobiogeography. xii + 531 pp., Elsevier Scientific Publ. Co., Amsterdam, London, New York. Harrington, H. J. and leanza, a. F. 1957. Ordovician trilobites of Argentina. Spec. Pubis. Dep. Geol. Univ. Kansas (Lawrence), 1, 276 pp., 104 figs. Harris, j. E. and mason, p. 1956. Vertical migration in eyeless Daphnia. Proc. R. Soc. B145, 280-290. ivshin, N. K. 1956. Upper Cambrian trilobites of Kazakhstan, 1. Trudy Inst. geol. Nauk, Alma-Ata , 3-98, pis. I-IX. 1960. See khalfin, l. l. 1960. — 1962. Upper Cambrian trilobites of Kazakhstan, 2. Trudy Inst. geol. Nauk, Alma-Ata, 3-412, pis. I-XXI. and Pokrovskaya, n. v. 1968. Stage and zonal subdivision of the Upper Cambrian. 23rd Int. geol. Congr. Prague, 9, 97-108. jaekel, o. 1909. Uber die Agnostiden. Z. dt. geol. Ges. 61, 380-401. jago, J. b. 1973. Cambrian agnostid communities in Tasmania. Lethaia, 6, 405-421. khalfin, l. l. (ed.). 1960. Palaeozoic biostratigraphy of the Sayan-Altai mining region. Part 1, Lower Palaeozoic, Cambrian System. Trudy SNIIGGIMS. 19, 11-253, 33 pis. [In Russian.] kindle, c. H. and Whittington, h. b. 1965. New Cambrian and Ordovician fossil localities in western Newfoundland. Bull. geol. Soc. Am. 76, 683-688, pis. 1-2. king, c. 1877. Report of the geological exploration of the fortieth parallel. Vol. 4, Prof. Papers Engineer Dept., U.S. Army, 18. kobayashi, t. 1933. Upper Cambrian of the Wuhutsui Basin, Liaotung, with special reference to the limit of the Chaumitien (or Upper Cambrian) of eastern Asia, and its subdivision. J. Jap., Geol. Geogr. 11, 55-155, pis. IX-XV. 1935a. The Briscoia fauna of the late Upper Cambrian in Alaska with descriptions of a few Upper Cambrian trilobites from Montana and Nevada. Ibid. 12, 39-57, pis. 8-— 10. 19356. The Cambro-Ordovician formations and faunas of South Chosen Palaeontology, pt. III. Cambrian faunas of South Chosen with special study on the Cambrian trilobite genera and families. J. Fac. Sci. Tokyo Univ. Ser. 2, 4, 49-344, pis. I-XXIV. — 1937a. The Cambro-Ordovician shelly faunas of South America. Ibid. 369-522, pis. I-VIII. 19376. Restudy on the Dames’ Types of the Cambrian trilobites from Liaotung. Trans Proc. palaeont. Soc. Japan, 12 (7), 70-86, pi. 17, reprinted from J. geol. Soc. Japan, 44 (523-525), 421-437, pi. 6. — 1938. Upper Cambrian fossils from British Columbia with a discussion on the isolated occurrence of the so-called 'Olenus' Beds of Mt. Jubilee. J. Jap. Geol. Geogr. 15, 149 192, pis. XV-XVI. 1939. On the Agnostids (Part 1). J. Fac. Sci. Tokyo Univ. Ser. 2, 5, 69-198. — 1951. Miscellaneous notes on the Cambro-Ordovician geology and palaeontology. No. XXIII. On the late Upper Cambrian (Pengshanian) fauna in eastern Jehol. Trans Proc. palaeont. Soc. Japan, n.s. 3, 75-80, pi. 7. — 1955. The Ordovician fossils of the McKay Group in British Columbia, western Canada, with a note on the early Ordovician palaeogeography. J. Fac. Sci. Tokyo Univ. Ser. 2, 9, 355-493, pis. I-IX. 1960. The Cambro-Ordovician faunas of South Korea, part VII, Palaeontology VI. Supplement to the Cambrian faunas of the Tsuibon Zone with notes on some trilobite genera and families. Ibid. 12, 329-420, pis. XIX-XXI. 1962. The Cambro-Ordovician formations and faunas of South Korea, part IX, Palaeontology VIII. The Machari fauna. Ibid. 13, 1-152, pis. I-VIII. — 1966. The Cambro-Ordovician formations and faunas of South Korea, part X, Stratigraphy of the Chosen Group in Korea and south Manchuria and its relation to the Cambro-Ordovician formations of other areas. Section B. The Chosen Group of North Korea and northeast China. Ibid. 16, 209-31 1. SHERGOLD: P S E U D AG NO STU S 95 lake, p. 1906. Monograph of the British Cambrian trilobites, part 1. Palaeontogr. Soc. [Monogr.], 1-28, pis. 1-2. laporte, L. F. 1971. Paleozoic carbonate facies of the central Appalachian shelf. J. sedim. Petrol. 41, 724-740. Lazarenko, N. p. 1966. Biostratigraphy and some new trilobites from the Upper Cambrian of the Olenek Uplift and Kharaulakh Mountain. Uchen. Zap. nauchno-issled. Inst. geol. Arkt., Paleont. Biostratigr. 2, 33-78, 8 pis. [In Russian.] Lermontova, E. v. 1940. Arthropoda. In vologdin, a. g. (ed. ). Atlas of the leading forms of fossil faunas in the USSR. Vol. 1, Cambrian, pp. 1-194, pis. 1-49. [In Russian.] — 1951. Upper Cambrian trilobites and brachiopods from Boshche-Kul (N.E. Kazakhstan). Trudy vses. nauchno-issetd. geol. Inst. 1-49, pis. I-VI. [In Russian.] lochman, c. 1940. Fauna of the basal Bonneterre Dolomite (Upper Cambrian) of southeastern Missouri. J. Paleont. 14, 1-53, pis. 1-5. — 1950. Upper Cambrian faunas of the Little Rocky Mountains, Montana. Ibid. 24, 322-349, pis. 46-51. — 1964. Upper Cambrian faunas from the subsurface Deadwood Formation, Williston Basin, Montana. Ibid. 38, 33-60, pis. 9-15. and hu chung-hung. 1959. A Ptychaspis faunule from the Bear River Range, southeastern Idaho. Ibid. 33, 404-427, pis. 57-60. - - - — I960. Upper Cambrian faunas from the northwest Wind River Mountains, Wyoming, part 1. Ibid. 34, 793-834, pis. 95-100. lu yen-hao. 1956u. On the occurrence of Lopnorites in northern Anhwei. Acta palaeont. sin. 4, 267-277 [in Chinese], 278-283 [in English], pi. I ■— 19566. An Upper Cambrian trilobite faunule from eastern Kueichou. Ibid. 4, 365-372 [in Chinese], 373-379 [in English], pi. 1. — 1957. Trilobita. In Chung-kuo piao chun hua shih [Index fossils of China, part 3]. Inst. Palaeont ., Acad. Sin. 249-298, pis. 137-155. [In Chinese.] — chang, w. t., chu chao-ling, chien yi-yuan and hsiang lee-wen. 1 965. Chinese fossils of all groups. Trilobita, Vol. 1, 362 pp., pis. 1-66; Vol. 2, pp. 363-766, pis. 67-135. Science Publication Co., Peking. mansuy, h. 1916. Faunes cambriennes de l’extreme-orient meridional. Mem. Serv. geol. Indoch. 5, 1-48, pis. 1-7. meek, f. b. 1873. Preliminary palaeontological report, consisting of lists and descriptions of fossils with remarks on the ages of the rocks in which they were found. In hayden, f. v. (ed.). 6th A. Rep. U.S. geol. Surv. Territ. 431-518. monke, h. 1903. Beitrage zur Geologie von Schantung. 1. Obercambrische Trilobiten von Yen-tsy-yai. Jb. preuss. geol. Landesanst. BergAkad. 23 (1902), 103-151, pis. 3-9. nelson, c. a. 1951. Cambrian trilobites from the St. Croix Valley. J. Paleont. 25, 765-784, pis. 106-110. nikitin, i. f. 1956. Cambrian and Lower Ordovician brachiopods from northeast central Kazakhstan. Trudy Inst. geol. Nauk, Alma-Ata , 3-143, pis. I-XV. [In Russian.] opik, a. a. 1961m Alimentary caeca of agnostids and other trilobites. Palaeontology, 3, 410-438, pis. 68-70. — 19616. The geology and palaeontology of the headwaters of the Burke River, Queensland. Bull. Bur. Miner. Resour. Geol. Geophys. Aust. 53, 5-249, pis. 1-24. 1963. Early Upper Cambrian fossils from Queensland. Ibid. 64, 5- 133, pis. 1-9. — 1967. The Mindyallan fauna of northwestern Queensland. Ibid. 74, Vol. 1, Text, 1-404; Vol. 2, Plates, 1-166, pis. 1-67. palmer, a. r. 1954. The faunas of the Riley Formation in central Texas. J. Paleont. 28, 709-786, pis. 76-92. — 1955. Upper Cambrian Agnostidae of the Eureka District, Nevada. Ibid. 29, 86-101, pis. 1 9 20. — 1960m Some observations on the early Upper Cambrian stratigraphy of White Pine County, Nevada. In Guidebook to geology of central Nevada. Intermount. Ass. Petrol. Geol., 11th Ann. Fid Conf. 53-58. — 19606. Trilobites of the Upper Cambrian Dunderberg Shale, Eureka district, Nevada. Prof. Pap. U.S. geol. Surv. 334C, 1-109, pis. 1-11. — 1962. Glyptagnostus and associated trilobites in the United States. Ibid. 374F, 1-49, pis. 1-6. 1968. Cambrian trilobites of east-central Alaska. Ibid. 559B, 1-13. — 1969. Cambrian trilobite distributions in North America and their bearing on Cambrian palaeo- geography of Newfoundland. In North Atlantic— Geology and continental drift. Mem. Am. Ass. Petrol. Geol. 12, 139-144. 96 PALAEONTOLOGY, VOLUME 20 palmer, a. r. 1972. Problems of Cambrian biogeography. Proc. Intn. geol. Congr. 24th Montreal , 7, 310-315. 1973. Cambrian trilobites. In hallam, a. (ed.), pp. 3-11. POKROVSKAYA, N. V. 1960. See TCHERNYSHEVA, N. E. 1960. 1963. See Vasilenko, v. k. 1963. rasetti, f. 1944. Upper Cambrian trilobites from the Levis conglomerate. J. Paleont. 18, 229-258, pis. 36-39. — 1959. Trempealeauan trilobites from the Conococheague, Frederick and Grove Limestones of the central Appalachians. Ibid. 33, 375-398, pis. 51-55. 1961. Dresbachian and Franconian trilobites of the Conococheague and Frederick Limestones of the central Appalachians. Ibid. 35, 104-124, pis. 21-25. — 1965. Upper Cambrian trilobite faunas of northeastern Tennessee. Smithson, misc. Colins. 148 (3), 1-127, pis. 1-21. Raymond, p. e. 1924. New Upper Cambrian and Lower Ordovician trilobites from Vermont. Proc. Boston Soc. Nat. Hist. 37, 389-466, pis. 12-14. resser, c. E. 1938. Cambrian System (restricted) of the southern Appalachians. Spec. Pap. geol. Soc. Am. 15, 1-140, pis. 2-16. Richthofen, f. f. von. 1883. China. Ergebnisse einiger Reisen und darauf gegrundeter Studien, 4, Palae- ontologischer Teil, vi+288 pp., 54 pis. Verlag von Dietrich Reimer, Berlin. robison, r. a. 1960a. Lower and Middle Cambrian stratigraphy of the eastern Great Basin. Intermount. Ass. Petrol. Geol. 11th Ann. Fid Conf. 43-52. - 19606. Some Dresbachian and Franconian trilobites of western Utah. Res. Stud. Geol. Brigham Young Univ. 7, 58 pp., 4 pis. 1964. Late Middle Cambrian faunas from western Utah. J. Paleont. 38, 510-566, pis. 79-92. — 1972. Mode of life of agnostid trilobites. Proc. Intn. geol. Congr. 24th Montreal , 7, 33-40. — and palmer, A. R. 1968. Revision of Cambrian stratigraphy, Silver Island Mountains, Utah. Bull. Am. Ass. Petrol. Geol. 52, 167-171. — and pantoja-alor, J. 1968. Tremadocian trilobites from the Nochixtlan region, Oaxaca, Mexico. J. Paleont. 42, 767-800, pis. 97-104. romanenko, E. v. and romanenko, M. F. 1967. Some problems of palaeogeography and Cambrian tri- lobites of the Altay Mountains. Izv. vses. geogr. Obshch. altai Otd. 8, 62-93, pis. 1-3. [In Russian.] rosova, a. v. 1960. Upper Cambrian trilobites from Salair (Tolstochikh Suite). Inst. Geol. Geofiz. Sib. Otd., AkadNauk SSSR, 5, 1-116, pis. I- VIII. [In Russian.] 1964. Biostratigraphy and descriptions of Middle and Upper Cambrian trilobites from the north- west Siberian Platform. Ibid. 3-148, pis. I XIX. [In Russian.] ruedemann, r. 1916. The presence of a median eye in trilobites. Bull. N.Y. State Mus. 189, 127-143, pis. 34-36. rusconi, c. 1950. Nuevos trilobitas y otros organismos del Cambrico de Canota. Revta Mus. Hist. nat. Mendoza , 4, 85-94. 1951. Mas trilobitas Cambricos de San Isidro, Cerro Pelado y Canota. Ibid. 5, 3-30. — 1953. Nuevos trilobitas Cambricos de la Quebrada de la Cruz. Boln paleont. B. Aires, 27, 1-8. — 1954. Trilobitas Cambricos de la Quebradita Oblicua, sud del Cerro Aspero. Revta Mus. Hist. nat. Mendoza , 7, 3-59, pis. I-IV. rushton, a. w. a. 1971. See taylor, k. and rushton, a. w. a. 1971. schrank, E. 1974. Kambrische trilobiten der China-Kollektion v. Richthofen. Z. geol. Wiss. Berlin , 2, 617-643, pis. I-V. shaw, a. b. 1951. Palaeontology of northwestern Vermont, 1. New late Cambrian trilobites. J. Paleont. 25, 97-114, pis. 21-24. shergold, j. h. 1972. Late Upper Cambrian trilobites from the Gola Beds, western Queensland. Bull. Bur. Miner. Resour. Geol. Geophys. Aust. 112, 1-126, pis. 1-19. 1975. Late Cambrian and early Ordovician trilobites from the Burke River Structural Belt, western Queensland. Ibid. 153, 251 pp., 58 pis. (2 vols.). — cooper, r. a., mackinnon, d. i. and yochelson, E. l. 1976. Late Cambrian Brachiopoda, Mollusca, and Trilobita from Northern Victoria Land, Antarctica. Palaeontology , 19, 247-291, pis. 38-42. shimer, H. w. and shrock, r. r. 1944. Index fossils of North America, ix + 837 pp., pis. I -303, Technology Press, Mass. Inst. Techn., John Wiley & Sons, inc.. New York. SHERGOLD: PSEUD AGNOSTUS 97 shumard, b. f. 1861. The primordial zone of Texas, with descriptions of new fossils. Am. J. Sci. (2), 32, 213-221. sun yun-chu. 1924. Contribution to the Cambrian faunas of China. Palaeont. sin. Bl, I 109, pis. I-V. — 1935. The Upper Cambrian trilobite faunas of north China. Ibid. B2, 1-69, pis. I-VI. — 1939. On the occurrence of Fengshanian (the late Upper Cambrian) trilobite faunas in W. Yunnan. 40th Anniv. Paps. natn. Unix. Peking 1939, 29-34, pi. 1 ; reprinted June 1947 in Contr. geol. Inst. natn. Unix. Peking , 27, 29-34, pi. I taylor, k. and rushton, a. w. a. 1971. The pre-Westphalian geology of the Warwickshire Coalfield. Bull. geol. Surx. U.K. 35, 152 pp., 12 pis. tchernysheva, N. E. (ed.). 1960. Class Trilobita. In orlov, yu. a. (ed.). Principles of Palaeontology , 8, Arthropoda-Trilobita and Crustacea , 515 pp. Moskva Akad. Nauk SSSR. [In Russian.] troedsson, G. t. 1937. On the Cambro-Ordovician faunas of western Quruq Tagh, eastern Tien-shan. In Report of the scientific expedition to the northwestern provinces of China under the leadership of Dr. Sven Hedin. The Sino-Swedish Expedition Publ. 4. Invertebrate Palaeontology, 1. Palaeont. sin., N.S., B2 (whole ser. 106), 1-74, pis. 1-10. tullberg, s. a. 1880. Agnostus-arierna i de Kambriska aflagringarne vid Andrarum. Sxer. geol. Unders., C42, 1-37, 2 pis. [In Swedish.] Vasilenko, v. K. (ed.). 1963. Decisions of the interdepartmental council for the preparation of the unified stratigraphical schemes for the Yakutian ASSR. Governmental geol. Comm. USSR. [In Russian.] walcott, c. D. 1913. The Cambrian faunas of China. In Research in China. Vol. 3, 3-276, pis. 1-24. Pubis. Carnegie Instn., 54. westergard, a. h. 1922. Sveriges Olenidskiffer. Sxer. geol. Unders., Cal8, 1-188 [in Swedish], 189-205 [in English], 6 pis. 1944. Borrningar genom Skanes Alunskiffer 1941-2. Ibid. C459, Arsb. 38 (1), 3-37 [in Swedish], 38-45 [in English], pis. 1 3. — 1947. Supplementary notes on the Upper Cambrian Trilobites of Sweden. Ibid. C489, Arsb. 41 (8), 3-34, pis. 1 -3. whitehouse, f. w. 1936. The Cambrian faunas of northeastern Australia. Part 1, Stratigraphical outline; part 2, Trilobita (Miomera). Mem. Qd Mus. 11, 59-1 12, pis. 8-10. — 1939. The Cambrian faunas of northeastern Australia. Part 3. The polymerid trilobites. Ibid. 21, 179-282, pis. 19-25. wilson, j. l. 1954. Late Cambrian and early Ordovician trilobites from the Marathon Uplift, Texas. J. Paleont. 28, 249-285, pis. 24-27. wolfart, R. 1974. Die fauna (Brachiopoda, Mollusca, Trilobita) des alteren Ober-Kambrium (Ober- Kushanian) von Dorah Shah Dad, Siidost-Iran, und Surkh Bum, Zentral-Afghanistan. Geol. Jb. B8, 71-184, pis. 10-27. woodward, h. 1905. On a collection of trilobites from the Upper Cambrian of Shantung, North China. Geol. Mag. [n.s. V], II, 21 1-215, 251-255, pi. XIII. J. H. SHERGOLD Bureau of Mineral Resources P.O. Box 378 Canberra A.C.T. 2601 Australia Typescript received 2 September 1975 Revised typescript received 2 February 1976 APPENDIX A Morphological Conditions Applicable to Pseudagnosti Constricted/unconstricted acrolobes. If the lateral margins of the acrolobe are curved slightly inwards, then the condition is known as constricted (Opik 1967). Where a constant curvature of the flanks of the acrolobe is maintained the condition is said to be unconstricted. Most species of Pseudagnostus have constricted pygidial acrolobes, the condition being most readily observed on parietal surfaces. Some species also have constricted cephalic acrolobes. G 98 PALAEONTOLOGY, VOLUME 20 Deliquiate/non-deliquiate marginal furrows. Marginal furrows which are deeply grooved or channel- like are described as deliquiate (Shergold 1975). If the marginal furrow is merely a break in convexity at the junction of the acrolobe and border the condition exhibited is non-deliquiate. Gradations exist, for which the term subdeliquiate is introduced to permit the description of degrees of deepening of furrows. Degree of deliquiation is related directly to degree of effacement. Exoskeleton and mould of the same specimen will have differing degrees of deliquiation, the mould having deeper, wider furrows. Deuterolobate. All pseudagnosti are deuterolobate. Degree of elevation of the deuterolobe, however, varies. En grande tenue species generally have a tumid deuterolobe well defined by accessory furrows. For this condition the term strongly deuterolobate is used here. Species with depressed deuterolobes are described as weakly deuterolobate. Effaced/efifacement/partial effacement. An effaced condition is one in which furrows and lobes with visible convexity are obliterated to give a smooth or nearly smooth surface. In Pseudagnostinae all condi- tions of partial effacement exist, from highly effaced to en grande tenue. En grande tenue. Introduced by Opik ( 1 96 1 A, p. 55), this term was redefined (Opik 1967, p. 56) to categor- ize agnostids having distinct lobes and furrows. Papilionate. Pseudagnostinae in which the axial glabellar node lies between the anterolateral lobes are termed papilionate. The term is derived from the butterfly-like appearance of these lobes and furrows (Shergold 1975, p. 42). Plethoid. Pseudagnostinae with accessory furrows clearly encircling the deuterolobe or continued posteriorly to the marginal furrow exhibit a plethoid condition (Shergold 1972, p. 15). Retral. This term refers to the siting of the posterolateral pygidial spines at the rear of the shield, behind or at the level of the rear of the deuterolobe. Simplicimarginate/zonate borders. Simplicimarginate agnostids have a basic unmodified border. Those having a duplicated posterior margin are said to be zonate (Opik 1967, p. 61). Spectaculate. A term introduced for pseudagnosti in which the axial glabellar node lies to the rear of the anterolateral glabellar lobes, and is therefore also to the rear of the anterior glabellar furrow. The resulting appearance resembles a bespectacled face— hence the term (Shergold 1975, p. 42). APPENDIX B Classification of Species Assignable to pseudagnostus sensu lato Species/Author/Date 1. acutifrons Troedsson, 1937 2. cfr. acutus Kobayashi, 1938 3. ampullatus Opik, 1967 4. angustilobus Ivshin, 1956 5. araneavelatus Shaw, 1951 6. aspidoides Kobayashi, 1955 7. bifax Shergold, 1975 8. bilobus Shaw, 1951 9. binodosus Kobayashi, 1955 10. bituber cuiatus Ivshin, 1960 11. boltonensis Resser, 1938 12. bulgosus Opik, 1967 13. canadensis Billings, 1860 14. cavernosus Rosova, 1960 15. chinensis Dames, 1883 (pars) 16. clarki Kobayashi, 1935a 17. clarki patulus Shergold, 1975 18. clarki prolatus Shergold, 1975 19. clarki maximus Shergold, 1975 20. clavus Shergold, 1972 21. coloradoensis Shumard, 1961 22. communis Hall and Whitfield, 1877 Original Generic Species Assessment Group Rhaptagnostus clavus Homagnostus clavus ? Pseudagnostus cyclopyge Pseudagnostus cyclopyge Pseudagnostus araneavelatus Neoagnostus bilobus Pseudagnostus corner gens Pseudagnostus bilobus Hyperagnostus bilobus Pseudagnostus clavus ? Oedorhachis bulgosus Pseudagnostus bulgosus Agnostus canddensis Pseudagnostus clavus Agnostus cyclopyge Plethagnostus clarki Pseudagnostus clarki Pseudagnostus clarki Pseudagnostus clarki Pseudagnostus clavus Agnostus communis Agnostus communis Revised Generic Assignment Neoagnostus Neoagnostus Pseudagnostus ( Pseudagnostus ) Pseudagnostus ( Pseudagnostus ) Neoagnostus Neoagnostus Rhaptagnostus Neoagnostus Neoagnostus Neoagnostus Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) Neoagnostus Neoagnostus Pseudagnostus ( Pseudagnostus) Rhaptagnostus Rhaptagnostus Rhaptagnostus Rhaptagnostus Neoagnostus Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) SHERGOLD: P S EU D AG N O STU S 99 Species/Author/Date 23. contractu Palmer, 1962 24. convergens Palmer, 1955 25. convexus Chu, 1959 (pars) 26. coronatus Shergold, 1975 27. cyclopvge Tullberg, 1880 28. cyclopygeformis Sun, 1924 29. cyclostigma Raymond, 1924 30. denticulatus Shergold, 1975 31. douvillei Bergeron, 1899 32. elix Shergold. 1975 33. empozadense Rusconi, 1954 34. empozadensis Rusconi, 1954 35. greendalensis Resser, 1938 36. gyps Clark, 1923 37. huangluosensis Kobayashi, 1966 38. idalis Opik, 1967 39. impressus Lermontova, 1 940 40. janei Clark, 1923 41. jarillensis Rusconi, 1953 42. jeholensis Kobayashi, 1951 43. josepha Hall, 1863 44. kobayashii Wolfart, 1974 45. koerferi Monke, 1903 46. laevis Palmer, 1955 47. latus Kobayashi, 1938 48. leptoplastorum Westergard, 1944 49. levatus Romanenko, 1967 50. longicollis Kobayashi, 1966 51. maladensis Meek, 1873 52. marginisulcatus Kobayashi, 1962 53. mesleri Resser, 1938 54. mestus Opik, 1967 55. mirus Pokrovskaya, 1963 56. neon Hall and Whitfield, 1877 57. nganasanicus Rosova, 1964 58. nuperus Whitehouse, 1936 59. obsoletus Lermontova, 1951 60. obtusus Belt, 1868 61. orbiculatus Shergold, 1975 62. orientalis Kobayashi, 1933 63. ovatus Rusconi, 1950 64. papilio Shergold, 1971 65. paronai Airaghi, 1902 66. pedrensis Rusconi, 1951 67. pii Airaghi, 1902 68. planulata Raymond, 1924 69. primus Kobayashi, 1962 70. priscus Kobayashi, 1955 71. prolongus Hall and Whitfield, 1877 72. pseudocyclopyge Ivshin, 1956 73. punctatus Lermontova, 1940 74. quadratus Lazarenko, 1966 75. quasibilobus Shergold, 1975 76. rotundatus Lermontova, 1940 77. securiger Lake, 1906 78. semiovalis Kobayashi, 1937a 79. sentosus Grant, 1965 80. sericatus Opik, 1967 81. simplex Lermontova, 1951 Original Generic Species Assessment Group Pseudagnostina contracta Pseudagnostus convergens Homagnostus cyclopyge Pseudagnostus araneavelatus Agnostus cyclopyge Agnostus convergens Phcilacroma araneavelatus Pseudagnostus araneavelatus Agnostus contracta Pseudagnostus clarki Lejopyge 7 Hypagnostus ? 7 Oedorhachis ? Plethagnostus ? Pseudagnostus Pseudagnostus cyclopyge Pseudagnostus communis Agnostus canadensis Plethagnostus cyclopyge Pseudagnostus ? Agnostus communis Pseudagnostus cy clopy gel Agnostus contracta Pseudagnostus clarki Pseudagnostus communis Pseudagnostus cyclopyge Pseudagnostus bulgosus Pseudagnostus bilobus Agnostus ? Pseudagnostus cyclopyge Oedorhachis bulgosus Pseudagnostus bulgosus Pseudagnostus Agnostus communis Pseudagnostus bulgosus Pseudagnostus cyclopyge Pseudagnostus convergens Agnostus cyclopyge Pseudagnostus clarki Pseudagnostus communis Pseudagnostus Pseudagnostus convergens Microdiscus 7 Spinagnostus 7 Agnostus ? Peronopsis clarki Pseudagnostus cyclopyge Trinodus bilobus Agnostus communis Pseudagnostus cyclopyge Pseudagnostus clavus Pseudagnostus clavusl Pseudagnostus bilobus Pseudagnostus communis Agnostus securiger Rliaptagnostus 7 Pseudagnostus communis Pseudagnostus cyclopyge Pseudorhaptagnostus clavus Revised Generic Assignment Pseudagnostus ( Pseudagnostina) Rliaptagnostus Pseudagnostus ( Pseudagnostus ) Neoagnostus Pseudagnostus ( Pseudagnostus) Rliaptagnostus Neoagnostus Neoagnostus Pseudagnostus ( Pseudagnostina) Rliaptagnostus ? ? ? ? nomen nudum Pseudagnostus (Pseudagnostus) Pseudagnostus ( Pseudagnostus) Neoagnostus y Pseudagnostus ( Pseudagnostus) ? Pseudagnostus ( Pseudagnostus) Pseudagnostus (Pseudagnostus) Pseudagnostus ( Pseudagnostina ) Rliaptagnostus Pseudagnostus ( Pseudagnostus) Pseudagnostus (Pseudagnostus) Pseudagnostus ( Pseudagnostus) Neoagnostus ? Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) nomen nudum Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus ) Rliaptagnostus Pseudagnostus ( Pseudagnostus) Rliaptagnostus Pseudagnostus ( Pseudagnostus) nomen nudum Rliaptagnostus ? 7 ? Rliaptagnostus Pseudagnostus (Pseudagnostus) Neoagnostus Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) Neoagnostus Neoagnostus Neoagnostus Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Sulcatagnostus) ? Pseudagnostus ( Pseudagnostus) Pseudagnostus ( Pseudagnostus) Neoagnostus 100 PALAEONTOLOGY, VOLUME 20 Species/Author/Date 82. solus Endo, 1937 83. subangulatus Lermontova, 1940 84. tennesseensis Resser, 1938 85. tmetus Harrington and Leanza, 86. vastulus Whitehouse, 1936 87. vicaria Opik, 1967 88. vulgaris Rosova, 1960 Original Generic Species Assessment Group Pseudagnostus Euplethagnostus clavus Oedorhachis cy clopy gel Machairagnostus bilobus Pseudagnostus cyclopygel Pseudagnostus con tract a Pseudagnostus clavus Revised Generic Assignment nomen nudum Neoagnostus Pseudagnostus ( Pseudagnostus) Neoagnostus Pseudagnostus ( Pseudagnostus ) Pseudagnostus ( Pseudagnostina ) Neoagnostus TWO NEW BAJOCIAN MICROCONCH OTOITID AMMONITES AND THEIR SIGNIFICANCE by C. F. PARSONS Abstract. Two new Bajocian (Middle Jurassic), microconch species belonging to thp ammonite family Otoitidae, Trilobiticeras ( Trilobiticeras ) cricki nov. and Emileia ( Otoites ) douvillei nov., are described, and are paired with their probable macroconch partners, T. (Emileites) malenotatus (Buckman) and E. (E.) subcadiconica Buckman respec- tively. The stratigraphic distribution of the main members of the subfamily Otoitinae in southern England shows that the two new species fill an important gap in our knowledge of this subfamily. In particular T. ( T .) cricki is of fundamental importance, as it is the undoubted ancestor to both Emileia and Frogdenites and possibly also Pseudo- loites. With their restricted stratigraphic range and wide geographic distribution, these new species are valuable stratigraphic indices for delimiting the base of the ovalis Subzone of the laeviuscula Zone. The most important exposures of the ovalis Subzone are discussed and the famous fauna described from the south of France by Douville is revised. During the course of a recent stratigraphic study of the Bajocian rocks (Middle Jurassic) of southern England (Parsons 1974, pp. 164-171), large ammonite collec- tions have been made from some long-neglected horizons. In particular the ovalis , or so called ‘Lower white iron-shot’ bed of Dundry Hill, near Bristol (Buckman and Wilson 1896, pp. 708-709, table iv), has yielded an interesting assemblage, which may be correlated with the ovalis Subzone of the laeviuscula Zone (Parsons 1974, p. 169)— see Table 1 for the zonal scheme used here. The majority of ammonites table 1. Zones and relevant subzones of the Lower Bajocian substage ( = Middle Bajocian sensu Arkell 1956), after Parsons 1974. LOWER BAJOCIAN ZONES Stephanoceras humphriesianum Emileia (Otoites) sauzei Witchellia ( Witchellia ) laeviuscula Hyperlioceras discites SUBZONES f W. (W.) laeviuscula ( Sonninia ovalis collected from this horizon ( + 56%), belong to the Witchellia I Pelekodites dimorphic group. However, intensive collecting revealed that a small proportion of the total fauna consists of two, as yet undescribed, species belonging to the microconch sub- genera Otoites and Trilobiticeras. A study of the previous literature and pre-existing museum collections showed that these stratigraphically important species had been collected and recognized as distinct for some considerable time; it has merely taken a hundred years for them to be formally described. The small physical size of these two new microconch species, as well as their relative rarity in comparison with the abundant Witchellia , goes some way towards explaining their absence in previous works on the Bajocian Otoitidae (Westermann 1954). [Palaeontology, Vol. 20, Part 1, 1977, pp. 101-118. pi. 17.] 102 PALAEONTOLOGY. VOLUME 20 As already noted, these two new species form only a minor part of the total ammonite fauna of this age in England. However, at comparable horizons in southern Europe the sonninid ammonites are less dominant, whilst the stephanoceratids are relatively more abundant. It is thus not surprising that this area of Europe has pro- duced the majority of previous records of these two new species, particularly the south of France, Portugal, and Sicily. With their wide geographic distribution and relatively restricted stratigraphic range, these ammonites are of great value for corre- lation purposes, particularly for between sonninid-dominated north-west Europe and the Tethyan region. These taxa are also of considerable phylogenetic importance, as they form a critical link between the early stephanoceratids of the concavum/discites Zones and the more abundant forms of the upper laeviuscula/sauzei Zones. As Emileia ( Otoites ) douvillei nov. and Trilobiticeras ( Trilobiticeras ) cricki nov. are characteristic of the ovalis Subzone of the laeviuscula Zone, some discussion of the stratigraphy of this horizon in southern England and elsewhere in Europe is given. Numbers preceded by these abbreviations refer to ammonites in the following collections: BMNH. The British Museum (N.H.), London. Cb. Bristol City Museum. IGS. The Institute of Geological Sciences, London. M. The Ecole des Mines Collection, Universite de Paris-Sud, Paris. OUM. The Oxford University Museum. SM. The Sedgwick Museum, Cambridge. CP. The author’s collection, Liverpool University. Under type species M = macroconch and m = microconch. SYSTEMATIC DESCRIPTIONS Class CEPHALOPODA Subclass AMMONOIDEA Superfamily stephanoceratacea Neumayr, 1875 Family otoitidae Mascke, 1907 The classification of the Otoitidae follows Westermann (1964), who divided this family between two subfamilies, the Otoitinae and Sphaeroceratinae. Subfamily otoitinae Mascke, 1907 The generic classification within this subfamily follows Westermann (1964, pp. 51-54), except for the inclusion of Frogdenites , which on both morphological and stratigraphic criteria has been transferred from the Sphaeroceratinae. Genus Emileia Buckman, 1898 Type species. Emileia (Emileia) brocchii (J. Sowerby, 1818), (M.). Diagnosis. A group of macroconch and microconch, sphaeroconic ammonites, which on the inner whorls bear ‘club’-shaped primary ribs, terminated by blunt nodes, from which divide two or more rounded secondary ribs. The nominate subgenus consists of medium to large-sized macroconch ammonites, with a terminal constric- tion and flared-mouth border, and which vary from involute sphaerocones via cadicones to evolute platycones. PARSONS: BAJOCIAN OTOITID AMMONITES 103 Subgenus Otoites Mascke, 1907 Type species. Emileia ( Otoites ) sauzei (d’Orbigny, 1846), ( m.). Diagnosis. A group of moderately small sized, lappeted microconch ammonites, which are characteristically involute sphaerocones, with a pronounced contraction of the body chamber. The ribbing style on the inner whorls is identical to its macro- conch, Emileia s. str., that is ‘club’-shaped primary ribs, dividing into three to four finer secondaries. On its outer whorls, Otoites tends to develop sharp tubercles or spines at this point of division, whilst on the body chamber the secondary ribs tend to become inflated and coarsened, often with an alternation of strength on the last half whorl, with every second, third, or fourth rib being more inflated than its neighbour. Emileia ( Otoites ) douvillei sp. nov. Plate 17, figs. 6, 7, and 9; text-fig. 1 1885 71921 1925 71929 1960 1960 1961 71967 71974 71975 non 1954 Sphaeroceras sauzei d’Orbigny; Douville non d’Orbigny, p. 41, pi. Ill, fig. 9 a-b. Otoites sauzei ; Riche and Roman non d’Orbigny, p. 138, pi. 6, fig. 8. Sphaeroceras (Otoites) sauzei d’Orbigny; Renz non d’Orbigny, p. 32, pi. II, fig. 8 and 8 a. Emileia sauzei d’Orbigny; Lanquine non d’Orbigny, p. 293, pi. ix, fig. 6. Otoites sp. (O. sauzei Roman non d’Orbigny); Lelievre, p. 17. Otoites sp.; Dubar, p. 52, pi. vii, figs. 25-26 and 724. Otoites cf. sauzei (d’Orbigny) (Douville, 1885, pi. 3, fig. 9); Ruget-Perrot, p. 54. Otoites sp.; Gabilly and Rioult, p. 3. Emileia ( Otoites ) sp. nov.; Parsons, p. 168. Emileia ( Otoites ) sp. nov.; Morton, pp. 84-85, pi. 16, figs. 5-6. Otoites fortis n. sp. ; Westermann, pp. 103-106, pi. 3, figs. 2-4; text-figs. 10 and 21. Material. Two specimens —0-30 m from the top of bed 2, Barns Batch Spinney (National Grid Reference ST 557 659), Dundry Hill, Bristol (Buckman and Wilson 1896, p. 689), BMNH. C79428-79429; five Dundry specimens from old museum collections, BMNH. C75242, IGS. 25189, Cb. 4744, Cb. 4957-4958, and one specimen +0-30 m from the base of bed 4, Bruton Railway quarry (ST 682 345), Bruton, Somerset (Richardson 1916, p. 495), BMNH. C79432, a total of eight specimens. Dimensions. Holotype, Cb. 4744, from Dundry Hill, and by matrix from the ovalis bed. It is complete, with the base of lappets, and just over three-quarters of a whorl of body chamber at a maximum diameter of 2-67 cm. Umbilical Number of Whorl Whorl Diameter diameter primary ribs height breadth Wb (D) (Ud) (Pn) (Wh) (Wb) Wh 2-50 0-83 (33%) c. 20 1 04 (42) 1-36 (54) 1-31 216 0-64(30) — 1 01 (47) 1-50(69) 1 49 First paratype, BMNH. C79429, with a quarter whorl of body chamber. 1-96 0-57 (29) 19 0-89 (45) 1 25 (64) 1-4 1-7 0-46 (27) 17 0-84 (49) 1-13 (67) 1-35 Second paratype, BMNH. C79428. 1-7 0-59 (35) 0-8 (47) 115 (68) 1 44 131 0-39 (30) — - 0-6 (46) 0-92 (70) 1-53 Description. A small (c. 2-3-2-1 cm diameter), sphaeroconic ammonite, with a strongly contracted body chamber, which extends for three-quarters to one whorl. The prorsiradiate primary ribs are short, blunt, and expand into ’club’-shaped nodes at 104 PALAEONTOLOGY, VOLUME 20 the point of division into three to four secondary ribs. These secondaries are relatively coarse, rounded rather than sharp, and bend forward only slightly over the rounded venter. The primary ribs weaken on the body chamber, whilst the secondaries, particularly on the last quarter whorl, are reduced to two per primary, and become inflated and coarsened, often with a tendency to become irregularly swollen. Thus on the last half whorl of the third paratype (1GS. 25189), every second or third rib is more expanded than its neighbour. The relative primary rib density increases rapidly (see text-fig. 1), from 15 to 16 per whorl on the inner whorls, to 20 to 21 on the body chamber. There is very little relative change in whorl cross-section, which stays more or less rounded, but depressed throughout ontogeny. The exception to this is found on the last quarter whorl of the body chamber, where, associated with the rapid un- coiling of the umbilical seam, there is a marked decline in whorl breadth (Wb), relative to whorl height (Wh) — see text-fig. 1. This rapid uncoiling of the body chamber changes the shell shape from a slightly sphaeroconic, to a more evolute form, although it is never as evolute as Trilobiticeras ( T .) crick i nov. The mouth border is characterized by a pair of spatulate lappets, which extend from the mid- whorl position (see PI. 17, fig. 6a). text-fig. 1. A plot of whorl breadth (Wb), whorl height (Wh), and number of primary ribs (Pn), against umbilical diameter (Ud), for Emileia ( Otoites ) douvillei sp. nov. Ht = holotype; P = paratype. Type series (six specimens). The holotype (PI. 17, fig. 6a-b), the best-preserved British specimen, comes from an old collection in the Bristol City Museum (Cb. 4744). The matrix of this specimen is identical to that of bed 2, Barns Batch Spinney, Dundry (Buckman and Wilson 1896, p. 689); that is, the ovalis bed, which is unique to the west side of Dundry Hill. The first paratype (BMNH. C79429), has only a quarter whorl of body chamber, but it is clearly very similar to the holotype in both propor- tions and ornament (see PI. 17, fig. 9 a-b), and like the second paratype (BMNH. C79428), it was collected in situ from bed 2, Barns Batch Spinney, Dundry, thus confirming the type horizon of the holotype. The third paratype (IGS. 25189), has the best-preserved lappets, whilst the fourth is a well-preserved specimen (Cb. 4957), which is recorded on its label as coming from the 'Sherborne district’ of north Dorset. However, the matrix of this specimen, which is totally atypical of the Sherborne area, would suggest a true provenance from the ovalis bed of Dundry. Lastly the fifth paratype (BMNH. C75242), is another well-preserved specimen, with a matrix characteristic of the ovalis bed (PI. 17, fig. la-b). Sexual dimorphism. Sexual dimorphism is well marked in this group. Emileia (Otoites) douvillei nov., like other members of this subgenus, shows typical microconch PARSONS: BAJOCIAN OTOITID AMMONITES 105 features. The mature shell is of relatively small size, and exhibits well-developed lappets. The corresponding macroconch is less well represented in the ovalis bed. A typical specimen from Dundry, ex Etheridge Collection, BMNH. C75279, is an involute, slightly cadiconic form, with bullate primary ribs and a quarter of a whorl of body chamber at a diameter of 4T cm, see Plate 17, fig. 8. The matrix of this speci- men suggests the ovalis bed as its source. This is confirmed by the occurrence of a com- plete specimen, with a plain mouth band, 9-0 cm in diameter, from the ovalis bed of West Dundry (Bristol University, Department of Geology, No. 67). These macro- conchs are best referred to E. ( E .) subcadiconica S. Buckman (1927, in 1909 1930, pi. 711). The type specimen of the latter (IGS. 49304), although fragmentary and wholly septate, exhibits, as noticed by Buckman (loc. cit. ), a typical ovalis bed matrix. Stratigraphic range. E. ( O .) douvillei appears to be characteristic of the ovalis Sub- zone of the laeviuscula Zone. However, this taxon probably ranges up into the base of the laeviuscula subzone, although the specimens referred to it from the basal part of the ‘Sandford Lane fossil bed', near Sherborne, Dorset (ST 628 179), are possibly not conspecific. The latter are slightly larger, with a different primary rib density ( = Otoites sp. nov. ; Parsons 1974, p. 168). By way of comparison, a single specimen from this locality and horizon, in the Reed Collection of the Yorkshire Museum is figured here (PI. 17, fig. 4). Geographic range. Apart from Dundry Hill, and more doubtfully Sandford Lane, the only other British localities to yield this species are Bruton Railway quarry, Bruton, near Castle Cary, Somerset (Richardson 1916, p. 495, bed 4), which has yielded one specimen, BMNH. C79432, and possibly Bearreraig Burn, Isle of Skye, Scotland (= Otoites sp. nov.; Morton 1975, pi. 16, figs. 5 and 6). It is impossible to be certain of the latter, as the specimen is not well preserved and is badly localized. Elsewhere this species has been recorded from the south of Lrance (Douville 1885; Lanquine 1929), Sicily (Renz 1925), Portugal (Ruget-Perrot 1961), and Morocco (Dubar 1960). Discussion. As may be seen from the synonymy list, the presence of this species has been noted by numerous authors, most of whom commented on its distinctive nature. This species has been named in honour of H. Douville, who was the first to figure it (Douville 1885). Dubar (1960, p. 52) has given the most detailed previous descrip- tion of this taxon, and he also pointed out the differences between his specimens and the other species of Otoites which had been figured up to that time. The criteria separating E. (O.) douvillei from E. (O.) fords (Westermann), with which it has been equated (Westermann 1954, p. 103), were given by Dubar (1960) as: its small size, being half to one-third of that of other Otoitids, the very strong contraction of its body chamber, and the modification of its ornament near the mouth border, particu- larly the loss of tubercles or nodes on the last two ribs. These characters appear perfectly valid, although there is some variation in the amount of contraction of the body chamber, since not all specimens show this to the same degree, notably those figured by Douville (1885, pi. Ill, fig. 9) and Lanquine (1929, pi. ix, fig. 6). It is also interesting to note Dubar’s comment (op. cit., p. 52) ‘The exact age of this small form is still to be established . . . perhaps this Otoites will be the most ancient'; 106 PALAEONTOLOGY, VOLUME 20 as it can now be shown that this species is indeed the oldest yet recorded belonging to the subgenus Otoites. There are really no other species of Otoites closely related to E. ( O .) douvillei, since they are all of much larger size and occur at higher stratigraphic horizons. In fact the most closely related forms are the more inflated variants of T. ( T .) cricki nov. which are found together at the same horizon. These two groups are very much alike in size and relative proportions, the main difference being in the style of ribbing. It is thus reasonable to suggest that E. ( E.)/(E .) Otoites evolved from the T. (T.)/T. (Emileites) group at the base of the laeviuscula Zone, especially since no specimens of either Otoites or Emileia have been found below this horizon. Genus Trilobiticeras Buckman, 1919 Type species. Trilobiticeras (Trilobiticeras) trilobitoides Buckman, 1919 (m.). Subgenus Trilobiticeras Buckman, 1919 Diagnosis. A group of very small-sized microconch ammonites, with coronate inner whorls. It has sharp primary ribs, which are terminated by sharp tubercles or spines at the point of division into the fine secondary ribs. The secondaries, of which there are three to five per primary, sweep forward over an arched venter. The body chamber shows an uncoiling of the umbilical seam, a coarsening of the tubercles and secondary ribs, a fading of the primary ribs, and it is terminated by a well-differentiated mouth border, with two spatulate lappets. The stratigraphic range of the subgenus is the same as that of Emileites , its macroconch counterpart, that is, upper concavum Zone to middle laeviuscula Zone. Trilobiticeras (Trilobiticeras) cricki sp. nov. Plate 17, figs. 1-3 and 5; text-fig. 2 1885 Sphaeroceras sauzei (variete ?); Douville non d'Orbigny, p. 41, pi. Ill, fig. 10. 1894 Sphaeroceras sp. nov. ; Crick, p. 436. 1972 Trilobiticeras trilobitoides Buckman; Galacz non Buckman, pp. 43-44, text-fig. 2a- d. 1974 Trilobiticeras sp. nov.; Parsons, p. 169. EXPLANATION OF PLATE 17 (All specimens are coated with ammonium chloride.) Fig. 1 a-b. Trilobiticeras (Trilobiticeras) cricki nov. (m.), holotype, BMNH. C79426, in situ bed 2, Barns Batch Spinney (Buckman and Wilson 1896), Dundry Hill, Bristol; x 1-5. Fig. 2a-b. T. (77) cricki nov., first paratype, BMNH. C79427, in situ bed 2, Barns Batch Spinney; X 1-5. Fig. 3 a-b. T. (T.) cricki nov., third paratype, BMNH. C75287, Dundry; X 1-5. Fig. 4. Emileia ( Otoites ) sp. cf. E. (O.) douvillei nov. (m.), Reed Collection, Yorkshire Museum, YM. 989, the ‘fossil-bed’, Sandford Lane, Dorset; x 10. Fig. 5. T. (T.) cricki nov., second paratype, BMNH. C79425, in situ bed 2, Barns Batch Spinney; X 1-5. Fig. 6 a-b. E. (O.) douvillei nov. (m.), holotype, Cb. 4744, Dundry; x 1-5. Fig. la-b. E. (O.) douvillei nov., fifth paratype, BMNH. C75242; x 1-5. Fig. 8. E. (Emileia) subcadiconica S. Buckman (M.), BMNH. C75279, Dundry; x 10. Fig. 9 a-b. E. (O.) douvillei nov., first paratype, BMNH. C79429, in situ bed 2, Barns Batch Spinney; x L5. Fig. 10. T. (Emileites) malenotatus (S. Buckman) (M.), BMNH. C73979, Dundry; x 10. PLATE 17 PARSONS, Bajocian otoitid ammonites 108 PALAEONTOLOGY, VOLUME 20 Material. Three specimens from the top 10 cm of bed 2, Barns Batch Spinney, Dundry Hill, Bristol, BMNH. C79425-79427; one specimen from bed 4b, Seavington-St.-Mary (ST 398 144), Somerset (Parsons and Torrens in Torrens 1969, p. A27), BMNH. C79433; one specimen from bed 6, Bruton Railway quarry, Bruton, Somerset, BMNH. C79431 ; one specimen from the base of the ‘fossil-bed’, Sandford Lane, Sher- borne, Dorset (Buckman 1893, p. 492, bed 6), BMNH. C79430; one specimen from Cape Mondego, Portugal (Ruget-Perrot 1961, p. 27, bed 5), BMNH. C79435; one specimen from Bradford Abbas, near Sherborne, Dorset, SM. J24550; and three Dundry specimens from old museum collections, BMNH. C75270, C75287, and OUM. .11 163. A total of eleven specimens, plus numerous (+100) other less well- localized specimens in various old museum collections. Dimensions. Holotype, BMNH. C79426, a complete specimen, with lappets and three-quarters of a whorl of body chamber. D Ud Pn Wh Wb Wb Wh 2-82 0-96 (34%) 17 1-19(42) 1-36 (48) 114 2-23 0-75 (34) 16 1-0 (45) 1-32 (59) 1-32 First paratype, BMNH. C79427, a complete specimen, with lappets and three-quarters of a whorl of body chamber. 2-85 1-03 (36) 19 1 11 (39) 1-2 (42) 1 08 2-2 0-78 (36) 17 0-94 (43) 111 (51) 1-18 Second paratype, BMNH. C79425, a complete specimen, with lappets and three-quarters of a whorl of body chamber. 214 0-66 (31) 23 0-95(44) 1-0 (47) 1-05 1-65 0-50 (30) 22 0-78 (47) 1-0 (61) 1 -28 Description. A small microconch ammonite (maximum diameter, modal value from ninety-five specimens == 2-65 cm), with cadicone inner whorls and a rapid contrac- tion of the body chamber, which extends for three-quarters of a whorl. The primary ribs are sharp, prorsiradiate, and approximately one-third of the length of the secondaries. The secondary ribs are extremely fine on the inner whorls, where there are three to five per primary, whilst on the outer whorls, where there are only two to three per primary, they are coarser and more widely spaced. The fine, sharp tubercles at the point of division of the secondary ribs, become stronger on the body chamber, whilst the primary ribs weaken, and in some cases become virtually obso- lescent. The primary rib density is fairly constant on the inner whorls, but shows a marked increase on the penultimate and last whorls, see text-fig. 2. This rib density varies from individual to individual, between 14 and 23 ribs per whorl, the majority of specimens having in the region of 17 to 18 on the last whorl. The contraction of the body chamber results in a relative reduction in whorl breadth, as is clearly shown in text-fig. 2. Since the whorl height continues to increase, although at a reduced rate, right up to the mouth border, there is a pronounced change in the whorl cross- section; the cadicone inner whorls, with an arched venter (Wb/Wh = 1-32), take on a more rounded cross-section (Wb/Wh 114). The mouth border is character- ized by fine spatulate lappets, which develop in the mid-whorl position, see Plate 17, fig. la, and which are usually preceded by a more coarse and prominent secondary rib. Although the inner whorls are cadicone, the contraction of the body chamber gives a more planulate appearance to the mature shell, which is one of the more evolute otoitid microconchs (Ud/Dx 100 = 30-40%). The suture line is difficult to make out on most specimens, but it appears similar to that of the Hungarian specimen (Galacz 1972, fig. 2). PARSONS: BAJOCIAN OTOITID AMMONITES 109 Type series (the six well-localized specimens). The holotype (BMNH. C79426), is one of the dominant morphotypes (PI. 17, fig. 1 a-b), whilst the first paratype (BMNH. C79427) is a slightly less inflated and more finely ribbed variant (PI. 17, fig. 2 a-b). The second paratype (BMNH. C79425), is a small, even more finely ribbed form (PI. 17, fig. 5). All three of the above specimens were collected in situ from bed 2 (top 10 cm). Barns Batch Spinney, Dundry Hill. The third paratype from Dundry (BMNH. C75287, ex Charlesworth Coll.), is a large fine ribbed specimen (PI. 17, fig. 3 a-b), whilst the fourth (BMNH. C75270, ex Pratt Coll.) is one of the more evolute forms. Finally the fifth paratype (OUM. J 1 1 63, ex Goddard Coll.), also from Dundry, is the most coarsely ribbed form I have seen; at a diameter of 2-4 cm, with more than two-thirds of a whorl of body chamber, it has fifteen primary ribs, at an umbilical diameter of 0-8 cm (33%). text-fig. 2. A plot of whorl breadth (Wb), whorl height (Wh), and number of primary ribs (Pn), against umbilical diameter (Ud), for Trilobiticeras ( Trilobiticeras ) cricki sp. nov. Ht = holotype; P = paratype. Sexual dimorphism. This group has well-marked sexual dimorphism and Trilobiti- ceras ( T .) cricki nov. exhibits typical microconch features: small size, with well- developed lappets. The corresponding macroconch, although not quite as abundant as its partner, is well represented in the ovalis bed at Dundry. These specimens are identical in shell form, style of ribbing, and general ornament, to all but the last whorl of T. (T.) cricki nov., but differ by still being wholly septate at a diameter in 110 PALAEONTOLOGY, VOLUME 20 excess of that attained by mature specimens of the latter. A typical ovalis bed macro- conch specimen from Dundry (BMNH. C73979) is figured here (PI. 17, fig. 10). This has one whorl of body chamber at a diameter of 5-05 cm. There are two specific names available for this group. The type specimen of 77 ( Emileites ) malenotatus (Buckman 1927, in 1909-1930, pi. 702, 1GS. 49293), although rather fragmentary and poorly localized, is undoubtedly very similar to many of these macroconchs. As noted by Buckman (loc. cit.), the matrix of this specimen is typical of the ovalis bed of Dundry. The type specimens of 77 ( E .) liebi (Maubeuge 1955, pi. 9, figs. 1 a-d and 2) are even closer in gross morphology to the Dundry forms, and it is likely that this species is a subjective junior synonym of 77 (E.) malenotatus. Stratigraphic range. The majority of members of this species have come from the ovalis bed of Dundry, which is ovalis Subzone, laeviuscula Zone in age (Parsons 1974, p. 169). The lowest occurrence of this species appears to be the discites Zone. A single specimen from Bradford Abbas, near Sherborne, Dorset (SM. J24550), has the iron-shot’, black-stained matrix typical of the top part of the Bradford Abbas Tossil-bed’, which is of this age (Parsons 1974, p. 170). A few specimens have been found as high as the laeviuscula Subzone of the laeviuscula Zone at Sandford Lane quarry, Sherborne (BMNH. C79430; Buckman 1893, p. 492, bed 6c) and at the South Main-road quarry (ST 567 655), Dundry (CP. 2403; Buckman and Wilson 1896, p. 691, bed 5). Geographic range. Outside of the localities in southern England mentioned above, this species has been figured from Hungary (Galacz 1972, fig. 2) and south-east France (Douville 1885, pi. Ill, fig. 10; M .121), whilst I have collected it from Cape Mondego, Portugal (BMNH. C79435; Ruget-Perrot 1961, p. 27, bed 5). Discussion. This species is named in honour of G. C. Crick, who first noticed its presence in the Inferior Oolite of Dundry Hill (Crick 1894, p. 436). T. (T.) cricki nov. is closest in gross morphology to specimens from the upper concavum/ lower discites Zones of southern England, which have been included in T. ( T .) punctum (Vacek); (Westermann 1964, pi. 6, figs. 5-6). It is difficult to establish if the small, fragmentary type specimen of this latter species (ibid., pi. 6, fig. la-b) is truly con- specific with the larger specimens figured by Westermann (loc. cit.), from Seavington- St.-Mary, Somerset. If it is, then this group is consistently more inflated, with a more depressed and coronate whorl section than T. ( T .) cricki. These two also have distinct stratigraphic distributions at Seavington-St.-Mary. The more inflated forms, T. ( T .) punctum , are relatively common in bed 3 (Parsons and Torrens in Torrens 1969, p. A27), which is upper concavum Zone in age, whilst the specimen of T. ( T .) cricki (BMNH. C79433), came from a higher horizon, bed 4b, which is ovalis sub- zone in age. However, the possibility must remain that the lectotype of T. ( T .) punctum is nothing but the poorly preserved inner whorls of a specimen of T. ( T .) cricki , rather than of the other congeneric, but stratigraphically lower, group. Unfortunately there is no definite solution to this problem, as the lectotype of T. ( T .) punctum is virtually uninterpretable. The only other species of Trilobiticeras, T. (77) trilobitoides Buckman and T. (77) platygaster Buckman, both have considerably more coronate inner whorls, a stronger PARSONS: BAJOCIAN OTOITID AMMONITES contraction of the body chamber, and more prominent spines and tubercles. The primary rib density of these two species is also consistently lower at fourteen to eighteen per whorl (average fifteen). As noticed by Crick (1894, p. 436), there is a considerable resemblance between some specimens of T. ( T .) crick i and members of the Pseudotoites group, particularly P. ( Latotoites ) evolution (Tornquist), (Wester- mann 1964, pi. 9, fig. 5). Whilst this resemblance could be nothing but another example of convergent evolution and homeomorphy, taking their relative strati- graphic positions into account, it is possible that these antipodean forms evolved from the Trilobiticeras/ Emileites group at the base of the laeviuscula Zone. PHYLOGENETIC SIGNIFICANCE OF NEW TAX A One of the main requirements for establishing a phylogenetic relationship is a detailed stratigraphic knowledge of the groups concerned. There have been considerable problems in the past in the interpretation of the early history and evolution of the ammonite family Otoitidae. This has largely been due to the inadequate and often erroneous stratigraphic information available to fill the apparent gap between the first well-documented faunas of the discites Zone and the diverse and abundant forms of the sciuzei Zone. One obvious source of confusion has been the error intro- duced by Mascke (1907) and perpetuated by Westermann (1954); that is, the arti- ficial stratigraphic separation of the microconch Otoites above its macroconch counterpart Emileia. The other major stratigraphic error which has compounded the difficulties inherent in interpreting the early history of the Otoitids, has centred on the correlation of the horizons now included within the laeviuscula Zone. Buck- man (1909-1930), when describing several new otoitid species from these horizons, referred them to more than six different hemerae (approximately equal to subzone in present use). When these hemerae were subsequently replaced by Oppel’s re- instated sower by i Zone (Spath 1936), there was little or no evidence available on the correct stratigraphic age of Buckman’s taxa. This is particularly true of the faunas ascribed to the "so-called' trigonalis subzone, which were an artificial combination of species from at least two separate horizons (see Parsons 1974, pp. 162-164, 171, for further details). Because of this confusion it has in the past proved impos- sible to subdivide the highly diverse otoitid faunas found below the sauzei Zone (Arkell 1956, p. 33). This unfortunately led Westermann (1964, text-fig. 14) to use dominantly morphological rather than stratigraphic criteria to establish his otoitid phylogeny. Recent work has revealed that there are numerous inconsistencies present in Westermann’s phylogenetic scheme. Groups such as Labyrinthoceras and Frog- denites, which were linked as macroconch and microconch are now known to occupy different stratigraphic horizons in the sauzei and laeviuscula Zones respectively, whilst other groups, such as E. crater , E. catamorpha , E. brocchii, and E. ( Otoites ) delicata , which were separated by Westermann, are now known to occur together in the lower part of the "fossil-bed’, Sandford Lane, near Sherborne, Dorset (Parsons 1974, p. 168). It has now proved possible to determine the correct stratigraphic position of most members of the Otoitinae, an essential step prior to establishing 112 PALAEONTOLOGY, VOLUME 20 any phylogenetic relationships. The successive otoitid faunas from southern England (after Parsons 1974, pp. 164-171, with additions), are as follows: (i) concavum Zone, rare specimens of Trilobiticeras ( T .) ‘ punctunT, and its un- described macroconch, 77 ( Emileites ) sp. nov., both of which are morphologically very similar to the Abbasites group from the subjacent horizons. (ii) discites Zone, rare specimens of 77 ( T .) trilobitoides , 77 ( T .) platygaster and two undescribed macroconchs, 77 ( Emileites ) spp. nov., along with extremely rare specimens of 77 (77) cricki nov. (iii) ovalis Subzone, relatively common specimens of 77 (77) cricki nov., 77 ( 77 .) malenotatus , 77. (77.) subcadiconica , and E. ( Otoites ) douvillei nov. (iv) laeviuscula Subzone (lower part), a diverse fauna of E. {E.) brocchii, E. (77.) catamorpha Buckman, E. ( E .) contrahens Buckman, E. ( E .) crater Buckman, E. (77.) polyschides (Waagen), E. { Otoites ) delicata (Buckman), E. (O.) douvillei nov., E. ( O .) sauzei (d’Orbigny) group, T. (77) cricki nov., and T. ( Emileites ) sp. (v) laeviuscula Subzone (upper part), E. ( E .) brocchii , E. (77.) bulligera (Buckman), E. (77.) polyschides , E. {Otoites) fortis (Westermann), E. (O.) sauzei group, Frog- denites extension (Buckman), F. gibberulum (Buckman), and F. spiniger Buckman. (vi) sauzei Zone, E. (77.) bulligera , E. (77.) greppini Maubeuge, 77. (77.) multifida Buckman, E. (77.) polyschides, E. (77.) pseudo contrahens Maubeuge, E. {Otoites) sauzei group, and E. {O.) contracta (Buckman non Sowerby). In most of the above faunas, mainly due to lack of material, it is still difficult to specifically link micro- and macroconchs. Thus, except in the case of Frogdenites, where the dimorphism is manifestly intraspecific in character, in the bulk of the Otoitinae, following Callomon (1963), dimorphism is best expressed at the sub- generic level. Since morphological diversity is lower in the microconch groups, it is easier to establish a phylogenetic link between their successive populations. A summary of the stratigraphic distribution and probable phylogeny of the main microconch members of the Otoitinae is thus shown in text-fig. 3. It is clear from this figure that the two new species are of critical phylogenetic importance. E. (O.) douvillei is the earliest member of its subgenus and its close relationship with T. (77) cricki finally establishes the Emileia/ Otoites group as an offshoot of Trilobiticeras. T. (77) cricki is of fundamental importance, as it proves to be the root-stock for several important otoitid groups, as well as a long-ranging taxon useful for demonstrating the continuity between the faunas of the discites and laeviuscula Zones. The genus Frogdenites closely resembles T. (77) cricki , whilst its early members, such as E. extension , also have an overlapping stratigraphic range : both F. extension and T. (77) cricki occur abundantly together in the middle laeviuscula Zone of Cape Mondego, Portugal (pers. obs.; Ruget-Perrot 1961, p. 27, bed 5). Specimens of F. extension from north Dorset (‘green-grained marl’, Oborne Wood, Parsons 1974, p. 167), exhibit relatively coronate inner whorls, with a rapid uncoiling of the umbilical seam, in a similar fashion to T. (77) cricki, to give a more planulate over- all shell form. Only the relatively late forms, such as F. spiniger, have developed the PARSONS: BAJOCIAN OTOITID AMMONITES 113 text-fig. 3. The stratigraphic distribution and probable phylogenetic relationship of the main microconch members of the ammonite subfamily Otoitinae in southern England. 114 PALAEONTOLOGY, VOLUME 20 more inflated, involute, sphaeroconic form, which is transitional to the Sphaero- ceratinae. T. ( T .) cricki thus shows both a similar shell form, with coronate inner whorls and fine sharp tubercles, and a similar style of dimorphism, with a low size ratio between dimorphs, to that exhibited by Frogdenites. Since T. ( T .) cricki is the only group to show these characters, which has a contiguous stratigraphic range with Frogdenites, it thus forms a highly probable ancestor for the latter genus. The connection between the Pseudotoites group and T. ( T .) cricki is the most tenuous of the three phylogenetic relationships suggested. Pseudotoites , particularly in Western Australia, is cryptogenic, as it appears suddenly in a geographically and stratigraphically isolated horizon, the Newmarracarra Limestone (Arkell and Play- ford 1954). The only way to determine the origins of such a group is to attempt to establish its relative stratigraphic position, and its similarities and affinities, if any, with possible related groups. The genus Fontannesia is abundant in Western Australia, where it forms a monospecific swarm at the base of the Newmarracarra Limestone (ibid., text-fig. 2). Fontannesia is found dominantly in the discites Zone in Europe, although it possibly ranges up into the base of the laeviuscula Zone (Pavia and Sturani 1968, p. 31 1). It is significant that this genus has been found in New Guinea associated with " Docidoceras' longalvum (Vacek), another discites Zone form (Westermann and Getty 1970). This would suggest that the succeeding Pseudotoites fauna (Arkell and Playford 1954, text-fig. 2), is just post-discites Zone in age. Since both Fontannesia and Docidoceras are predominantly Tethyan in origin (Arkell 1956, pp. 177 and 209; Westermann and Getty 1970, p. 291), it is logical to seek the ancestors of the Pseudotoites group in the Otoitids of the Tethyan discites Zone. The Pseudotoites group shows a high degree of morphological diversity. The Australian macroconchs ( Pseudotoites s. str .), are medium-sized, relatively inflated forms, with inner whorls exhibiting numerous, fine prorsiradiate secondary ribs (Arkell and Playford 1954, pi. 32, fig. 3), very similar to the T. ( E .) nialenotatus figured here (PI. 17, fig. 10). The outer whorls show a modified ornament, with the development of short, inflated, nodate primary ribs. This is a similar trend to that shown by some undescribed specimens of Emileia from the lowest laeviuscula sub- zone at Dundry. The corresponding microconchs ( Pseudotoites ( Latotoites ) spp.), included by Arkell in Otoites are similar to the latter genus in ribbing style, but differ by being much more evolute. They are thus similar in many characters, except size, to T. ( T .) cricki nov. There are rare specimens in the Australian faunas which show a close similarity with members of the T. (T.) trilobitoides Buckman group; P. ( Lato- toites) depressus (Whitehouse), (Arkell and Playford 1954, pi. 30, fig. 7). Arkell was hesitant about including the latter taxon in Trilobiticeras because of its greater size (c. 4 0 cm), compared to the much smaller European Trilobiticeras known at that time. The description of T. ( T .) cricki, with a maximum diameter in excess of 3-1 cm has thus removed one serious objection to any comparisons between these groups. On both general morphological and stratigraphic grounds, there is thus no really plausible existing alternative to accepting the Trilobiticeras/T. ( Emileites ) group as the root stock of Pseudotoites. If Pseudotoites did not evolve directly from the T. (T.) cricki group, then they both must have had a common ancestor in the discites Zone. An interesting problem is presented by the relationship of the above groups to the genus Docidoceras. As pointed out by Arkell (Arkell and Playford 1954, p. 572), the PARSONS: BAJOCIAN OTOITID AMMONITES 115 type species of this genus, D. cylindroides S. Buckman, bears a close morphological relationship to some members of the Pseudotoites group. The former species is in fact atypical of the rest of the forms included in Docidoceras by Buckman (1909- 1930) (Arkell and Playford 1954, p. 572). There is thus a possibility that this species may have had a common origin, with Trilobiticeras/ Emileites, in the concavum Zone Abbasites , a suggestion already made by Westermann (1969, pp. 129 and 137). The bulk of the other ‘species’ of ‘Docidoceras' , with their more typically stephano- ceratid ribbing style, are undoubtedly closely related to some as yet undescribed species, akin to ‘D.' longalvum (Vacek), from the middle murchisonae Zone of south Dorset (Senior et al. 1970 Docidoceras sp., Horn Park quarry, bed 5a). The genus Docidoceras , as at present defined, could thus be polyphyletic. It seems certain that Docidoceras , taking into account its stratigraphic range and gross morphology, makes a poor dimorphic partner for Trilobiticeras ; on these grounds Emileites is far more satisfactory. text-fig. 4. The geographic distribution of 1, Tri- lobiticeras ( Trilobiticeras) cricki sp. nov., and 2, Emileia ( Otoites ) douvillei sp. nov. STRATIGRAPHIC SIGNIFICANCE The two species described here jointly have a restricted stratigraphic range (text- fig. 3) and a wide geographic distribution (text-fig. 4). They thus form an important part of an ammonite fauna, which over most of Europe is characteristic of the basal ova/is subzone of the laeviuscula Zone (Parsons 1974). Whilst the constituent members of this fauna remain much the same, their relative proportions vary according to geographic position. In southern England, a fairly typical fauna from Barns Batch Spinney, Dundry, has the following proportions: Witchellia 46 specimens, Sonninia 10, Trilobiticeras 17, Emileia 2, Docidoceras 1, Bradfordia 6, Strigoceras 2. 116 PALAEONTOLOGY, VOLUME 20 To the north in Skye (Inner Hebrides, Scotland), the Sonninidae are almost totally dominant, as apart from an isolated specimen of E. ( O .) cf. douvil/ei nov., only specimens of Sonninia s. /at. and Witchellia have been recorded (pers. obs.; Morton 1975). South towards Tethys the Sonninidae become progressively less abundant as there is an increase in diversity linked with a levelling out of the relative proportions of the different ammonite groups present. On the whole in Tethyan and adjoining regions, the Stephanoceratacea and Haploceratacea are the more impor- tant. Of these, the genus Bradfordia is particularly characteristic of the Lower Bajocian in Bulgaria (Sapunov 1971), Sicily (Renz 1925), and Portugal (pers. obs.). The major exception to this trend appears to be south Germany, where probably due to palaeo- geographic isolation, the faunas are largely restricted to Sonninia s. str. The most important areas of preservation of ovalis subzone faunas in Europe are to be found in England, France, and Germany. Southern England The best exposures of rocks assigned to this subzone in England are to be found on Dundry Hill, near Bristol. Here sections at Castle Farm (Buckman and Wilson 1896, p. 676), Barns Batch Spinney (ibid., p. 689), South Main-road (ibid., p. 691), and Rackledown (ibid., p. 692), have all yielded abundant ovalis subzone faunas. The most extensive collection from this horizon (the 'Lower white iron-shot’), has been made at Barns Batch Spinney (ibid., p. 689, bed 2), and this includes: Witchellia (Witchellia) albida (S. Buckman); W. (W.) romanoides (Douville); W. (W.) cf. connata (S. Buckman); W. (W.) cf. sutneri (Branco); W. (Pelekodites) pelekus (S. Buckman); W. ( P .) cf. macra (S. Buckman); Sonninia ovalis (S. Buckman ex Quenstedt); Euhoploceras sp.; Bradfordia cf. inclusa S. Buckman; Strigoceras com- pression (S. Buckman); Toxamblyites sp.; Docidoceras cf. cylindroides S. Buckman; Trilobiticeras ( T.) cricki nov.; T. (Emileites) malenotatus (S. Buckman)-!’. (E.) liebi (Maubeuge) group; Emileia ( Otoites ) douvillei nov. Other areas in England which have produced similar faunas include the Cotswold Hills (Buckman 1895, p. 397, beds 4-5); the Cole ‘syncline’, Bruton, Somerset (Richardson 1916, p. 495, bed 6); the Sherborne district, Dorset (Buckman 1893, p. 493, bed 8), and Seavington-St.-Mary, Somerset (Parsons and Torrens in Torrens 1969, p. A27, bed 4b). France The so-called ‘Witchellia- beds’ of Normandy have produced a fauna of at least the ovalis subzone (Haug 1893; Bigot 1900; Gabilly and Rioult 1974). The most famous exposures of rocks of this age are, however, to be found to the south near Toulon (Douville 1885). Here a thin, condensed limestone has produced one of the most diverse faunas from this subzone, which has been described in detail (ibid.). Whilst the stratigraphy and distribution of this bed is well known (ibid.; Lanquine 1929), the interpretation of its fauna has been difficult, as only a few specimens have been figured by modern methods (Lanquine 1929, pis. IX and X). For the rest only the rather stylized drawings provided by Douville (1885, pis. I I II) have been avail- able. However, Douville’s specimens are still preserved in the Ecole des Mines Collections, now held at the Universite de Paris-Sud. The latter has kindly provided PARSONS: BAJOCIAN OTOITID AMMONITES 117 me with photographs of this material, which have enabled a more accurate inter- pretation of Douville’s figures to be made: Plate I, figure 1, Euhoploceras sp. 2, IE. sp. 3 and 3 a, E. sp. 4, E. cf. palmata S. Buckman. 5, Witchellia (Pelekodites) cf. zurcheri (Douville). 6, W. ( P .) zurcheri, lectotype, designated by Buckman (1909-1930, pi. 399). 7, W. {P.) zurcheri. 8, Zurcheria ubaldi Douville. Plate II, figure 1, Witchellia ( Witchellia ) sayni Haug. 2-5, W. (W.) spp. Plate III, figures 1-2, W. ( W.) sp. 3 and 3 a, W. ( W.) romanoides( Douville). 4, W. ( W.) romanoides. 5, W. ( W.) romanoides. 6 and 6a, Bradfordia praeradiata (Douville), lectotype, designated by Sapunov (1971, p. 79). 7, Bradfordia cf. praeradiata. 8, Emileia (Emileia) a IT. subcadiconica S. Buckman. 9, E. ( Otoites ) douvillei nov. 10, Trilobiticeras ( Trilobiticeras) cricki nov. The two other specimens, which have been figured from this horizon by Lanquine (1929), are Emileia ( Otoites ) cf. douvillei nov. (ibid., pi. IX, fig. 6) and Trilobiticeras ( Emileites ) sp. (ibid., pi. X, fig. 2). This fauna, apart from its slightly higher apparent diversity, is identical to that recorded from Dundry Hill, England; the close simi- larity of the Witchellia groups is particularly striking. Germany and Switzerland The Schwabian Albe of south-west Germany and the adjoining Swiss Jura both possess fossiliferous representatives of the oval is Subzone. The ‘Unterer Wedel- sandstein’ of Schwabia (Parsons 1974, p. 173) is the source of the type specimen of Sonninia ovalis (Buckman ex Quenstedt), and this species is by far the most abun- dant in these beds. Other less-common forms include rare specimens of Witchellia spp. and a solitary specimen of T. {E.) liebi (Maubeuge) (Bayer 1968; Parsons 1974, p. 175). The equivalent beds in the Swiss Jura have apparently yielded the type specimens of T. (E.) liebi (Maubeuge 1955, pi. 9, figs. \a-d and 2), but this needs to be confirmed by the collection of in situ topotypes. Acknowledgements. The majority of the work for this paper was undertaken during the tenure of a Uni- versity of Keele research studentship; this award is gratefully acknowledged. I should like to thank the curators of the following museums for access to material in their care: the British Museum (N.H.) and Institute of Geological Sciences, London; the museum of the Department of Geology, Bristol University; Bristol City Museum; Oxford University Museum; the Sedgwick Museum, Cambridge, and the York- shire Museum, York. I should also like to thank the Universite de Paris-Sud, Centre d’Orsay, for providing photographs of material in the Douville Collection. REFERENCES arkell, w. J. 1956. Jurassic Geology of the World, xiv + 806 pp., Oliver and Boyd. Edinburgh. — and playford, p. E. 1954. The Bajocian Ammonites of Western Australia. Phil. Trans. R. Soc. B237, 547-605, pis. 27-40. bayer, u. 1968. Docidoceras cf. liebi Maubeuge aus dem Unteren Bajocium des Wutachgebietes. Stuttgarter Beitr. z. Naturkde. 183, 1-3. bigot, a. 1900. Normandie: Livret guide des excursions du VIII Congres Geologique International, Paris, 33 pp. buckman, s. s. 1893. The Bajocian of the Sherborne; its relation to subjacent and superjacent strata. Q. Jl geol. Soc. Lond. 49, 479-522. 1895. The Bajocian of the mid-Cotteswolds. Ibid. 51, 388-462. — 1909-1930. Type Ammonites. Wheldon and Wesley, London and Thame, Vols. 1-7, 412 pp., 790 pis. — and wilson, e. 1896. Dundry Hill: its upper portion, or the beds marked as Inferior Oolite in the maps of the Geological Survey. Q. J! geol. Soc. Lond. 52, 669-720. PALAEONTOLOGY, VOLUME 20 callomon, J. h. 1963. Sexual dimorphism in Jurassic ammonites. Trans. Leicester lit. phil. Soc. 57, 21-56. crick, G. c. 1894. On a collection of Jurassic Cephalopoda from Western Australia. Geol. Mag. (4), 1, 395-393, pi. 12, 433-441, pi. 13. douville, e. 1885. Sur quelques fossiles de la zone a Amm. Sowerbyi. Bull. Soc. geol. Fr. 3e ser., 13, 12-44, pis. I-III. dubar, g. 1960. Supplement a l’etude des faunes Aaleniennes de Krendegg et du J. Tratt (Prerif, Maroc). Ann. Soc. geol. du Nord , 80, 50-52. gabii ly, J. and rioult, M. 1974 (preprint, 1967). Le Bajocien et le Toarcien Superieur sur les Bordures du Massif Armoncain. Colloq. Jurassique Luxembourg 1967, Mem. B.R.G.M. Fr. 75, 385-396. galacz, a. 1972. Trilobiticeras (Ammonoidea, Otoitidae) from the Bajocian (Middle Jurassic) of the Bakony Mountains. Ann. Univ. Sci. Budapest Sect. Geol. 15, 39-45. haug, e. 1893. Etude sur les ammonites des etages moyen du System Jurassique. Bui. Soc. geol. Fr. 3e ser., 20, 277-333, 3 pis. lanquine, a. 1929. Le Lias et le Jurassique des Chaines Provengales: I. Le Lias et le Jurassique Inferieur. Bui. Serv. Cart. geol. Fr. 32, 385 pp., 12 pis. lelievre, t. 1960. Etude des Ammonites de l’Aalenien de deux gisements du Nord du Maroc (Prerif). Ann. Soc. geol. du Nord. 80, 15-50, pis. 5-7. mascke, e. 1907. Die Stephanoceras-Verwandten in den Coronatenschichten von Norddeutschland. Inaug. Dissert. Gottingen, 38 pp. maubeuge, p. l. 1955. Les Ammonites aaleniennes, bajociennes et bathoniennes du Jura suisse septentrional. Mem. suisses Paleont. 71, 48 pp., 1 1 pis. morton, n. 1975. Bajocian Sonniniidae and other ammonites from western Scotland. Palaeontology, 18, 41-91, pis. 6-17. parsons, c. f. 1974. The sauzei and ‘so called’ sowerbyi Zones of the Lower Bajocian. Newsl. Stratigr. 3, 153-179. pavia, G. and sturani, c. 1968. Etude biostratigraphique du Bajocien des Chaines Sub-alpines aux environs de Digne (Basses-Alpes). Boll. Soc. geol. It. 87, 305-316. renz, c. 1925. Beitrage zur Cephalopoden-fauna des alteren Doggers am Monte San Giuliano (Monte Erice) bei Trapani in Westsizilien. Abhandl. Schweiz. Palaont. Gese/I. 45, 1-33, 2 pis. richardson, l. 1916. The Inferior Oolite and Contiguous Deposits of the Doulting-Milborne-Port District (Somerset). Q. J! geol. Soc. Lond. 71, 473-520. riche, a. and roman, f. 1921. La Montagne de Crussol. Trav. Lab. geol. Fac. Sci. Lyon , fasc. 1, 196 pp., 8 pis. ruget-perrot, c. 1961. Etudes stratigraphiques sur le Dogger et le Malm Inferieur du Portugal au Nord du Tage. Mem. Serv. geol. Portugal, 7, 197 pp., 1 1 pis. sapunov, i. G. 1971. The Bajocian ammonite genus Bradfordia S. Buckman, 1910 (Oppeliidae) in Bulgaria. Bulg. Acad. Sci. Bui. geol. Inst. ser. Paleont. 20, 73-90, 3 pis. [In Bulgarian, with English summary.] senior, J. r., parsons, c. f. and torrens, h. s. 1970. New sections in the Inferior Oolite of South Dorset. Proc. Dorset Nat. Hist. Archaeol. Soc. 91, 114 -119. spath, l. f. 1936. On Bajocian Ammonites and Belemnites from eastern Persia (Iran). Palaeont. Indica (n.s.), 22, 1-21, 1 pi. torrens, h. s. (Ed.) 1969. International Field symposium on the British Jurassic. Excursion Guide I— Dorset and South Somerset. University of Keele, 71 pp. westermann, g. e. g. 1954. Monographic der Otoitidae (Ammonoidea). Beih. Geol. Jb. 15, 364 pp., 33 pis. — 1964. Sexual-dimorphismus bei Ammonoideen und seine bedeutung fur die taxionomie der Otoitidae. Palaeontographica (A), 124, 33-73, pis. 6-9. — 1969. The ammonite fauna of the Kialagvik Formation at Wide Bay, Alaska Peninsula. Part II. Sonnmia sowerbyi Zone (Bajocian). Bull. Amer. Paleont. 57, 1-226, pis. 1-47. and getty, T. a. 1970. New Middle Jurassic Ammonitina from New Guinea. Ibid. 227-321, pis. 48-62. C. F. PARSONS Department of Geology University of Liverpool Typescript received 27 October 1975 p q gox 147 Revised typescript received 27 January 1976 Liverpool L69 3BX SOME PHACOPINA (TRILOBITA) FROM THE SILURIAN OF SCOTLAND by E. N. K. CLARKSON, N. ELDREDGE, and J.-L. HENRY Abstract. Acernaspis ( Eskaspis ), a new subgenus, and Podowrinella , a new genus of Phacopina from Silurian inliers of the Midland Valley of Scotland, are here proposed. A. (Eskaspis) is restricted to the Telychian (upper Llandovery) and is referred to the subfamily Phacopinae. A. (E.) sufferta (Lamont) from the M. crenulata Zone of the Pentland Hills and A. ( E .) woodburnensis from the upper M. sedgwickii Zone of Girvan are described. Podowrinella , which is found in the lower Wenlock of Girvan, and in the Telychian (upper Llandovery) of the Hagshaw and Pentland Hills, can probably be assigned to the Pterygometopidae, but possesses some character states normally considered representative of Phacopidae. Coaptative structures on the ventral doublures of A. ( Eskaspis ) and Podowrinella are described in detail, and information is given on the auxiliary impression system on the glabella of A. (E.) sufferta. Silurian rocks in the Midland Valley of Scotland lie in a chain of inliers extending from Girvan to within a few kilometres of Edinburgh (text-fig. 1). At Girvan, in the Hagshaw Hills, and in the North Esk Inlier of the Pentland Hills, the Llandovery succession begins with turbidites or subturbidites, and passing upwards through shallow-water marine horizons (with shelly faunas well developed at certain horizons at Girvan and in the North Esk Inlier) the sequence changes to thin brackish or freshwater deposits and finally into redbeds, early in the Wenlock. The large inlier of Lesmahagow, lying further to the north, has dominantly fresh or brackish water beds followed by redbeds. In Telychian (upper Llandovery) mudstones and siltstones of the North Esk Inlier there occur abundant faunas, especially in the beds which have been referred to as the ‘ P/ectodonta mudstones’ by Lamont (1947) and others. Fossils were first discovered here in 1838 by Charles Maclaren of the Geological Survey, and Howell and Geikie listed them in 1861. The fine preservation and rich- ness of the faunas led to much palaeontological activity in the next few decades, mainly by members of the Edinburgh Geological Society and associates (Haswell 1865; Henderson 1867; Brown and Henderson 1867; Henderson and Brown 1869; Etheridge 1874; Davidson 1874). But although extensive faunal lists were compiled, only Haswell (1865) and Davidson (1874) described any of the faunal elements. Haswell’s descriptions and plates were very sketchy, but Davidson’s monograph of the brachiopods was a valuable and enduring contribution to the palaeontology of the region. The U.K. Geological Survey’s faunal notes were updated by Peach and Horne (1899) and Mykura and Smith (1962), who described the area in detail. Meanwhile, Lamont (1947, 1948, 1949, 1952) discussed the fauna at length pri- marily in a stratigraphical context, recognizing for the first time that the faunas were pre-Wenlock. New bivalves and chelicerates were later described (Lamont 1954, 1955). Though the starfish from well-known horizons in the Gutterford Burn were formally described by Spencer (1914-1940), most of the other elements in the [Palaeontology, Vol. 20, Part 1, 1977, pp. 119-142, pis. 18-20.] 120 PALAEONTOLOGY, VOLUME 20 rich and diverse fauna of the Pentland Hills remained imperfectly known or un- described until very recently. Interest in the fauna has revived of late, however, and various elements have been revised or described for the first time by various authors. In recent works there have been described a new echinoid (Kier 1973), crinoids (Brower 1975), and the trilobite Scotoharpes (Norford 1973). Tipper (1976) has remapped the region, revised the stratigraphy, erected new formation names, and distinguished three successive faunal assemblages from the upper Deerhope and lower Wether Law Linn Formation (Tipper 1975). In the present paper Tipper’s formation names have been used throughout, replacing the older stratigraphical divisions (text-fig. 1). Trilobites, amongst many other fossils, especially brachiopods, are abundant at certain horizons within the upper Llandovery, and though the fauna is restricted to a comparatively few genera and species, individuals of these may be frequently encountered. In the Reservoir Formation these include occasional phacopids, prob- ably Acernaspis ( Eskaspis ), n. sg. otarionids and odontopleurids, whereas Podow- rinella straitonensis , Encrinurus , and Hemiarges together with less-common elements are present at the top of the Deerhope Formation. In the immediately overlying Wether Law Linn Formation (which includes the P/ectodonta Mudstones), there is a rich fauna with A. (E.) suffer ta (Lamont) ( 1 50 examples), Encrinorus sp. ( c . 150 ex.), Proetus latifrons (M’Coy) (25 ex.), Cyphoproetus depressus (Barrande) (16 ex.), Scotoharpes domina (Lamont) (2 ex.), Youngia (6 ex.), Calymene sp. (6 ex.), Cheirurus sp. (1 ex.), and possibly other genera and species listed by Lamont (1948), though the validity of these has not yet been ascertained. In the Wether Law Linn Formation the most abundant trilobites are found only at the base (in units A and C of Tipper) becoming less common towards the top of unit C of the succession. Encrinurus occurs quite abundantly, however, in the upper- most part of this formation, in unit E, in which the only other trilobite to have been collected is Scotoharpes. Tipper (1975) has shown that trilobites occur independently of the brachiopod communities which he described in this sequence. In this paper the Phacopina of the Pentland Hills are described, these being A. (E.) sufferta (Lamont, 1947) and Podowrinella straitonensis (Lamont, 1965). A. (E.) sufferta is known from the Pentland Hills alone, but an earlier species, A. (E.) woodburnensis, n. sp. from the older Wood Burn Formation at Girvan can also be referred to that genus, and is the only other species known at present. P. straitonensis is known from the Ree Burn Formation in the Hagshaw Hills, from which it was first described from specimens collected by Rolfe, and also from the lower Wenlock Knockgardner Formation at Girvan. In the Pentland Hills it occurs in a coarse siltstone, whereas both in the Hagshaw Hills and at Knockgardner, speci- mens are found at the base of coarse turbidite flows where they occur with brachio- pod faunas (Rolfe 1973; Cocks and Toghill 1973). In all these cases Encrinurus and Hemiarges occur in the same fauna, which, judging by its anomalous stratigraphic position (younger in the west), strongly indicates either a facies fauna (perhaps con- trolled in some way by the regressive sequence), or else inaccuracies in correlation. The present work is the first of a series of papers in which the trilobite fauna of the Pentland Hills is described. The rest of the fauna will be described in later papers, some by other authors. GRIESTONENSIS TURRICULATUS ► >r MAXIMUS ► SEDGWICKI I * CONVOLUTUS ^ Sfraiton grits Knock - gardner Fm. Blair Shale Drumyork Flags Protovir -gularia grits Penkill Fm Pencleuch Shale Tralorg Glenbuck Group - Podowrinel la straitonensis Hemiarges rolfei Encrinurus sp. Parishholm Congl. Smithy Fm. - Podowrinella straitonensis Hemiarges rolfei Encrinurus sp. \HAGSHAW '■ HILLS - Lauchlan Fm. Upper Camregan Maxwellston mdst. Wood Burn Fm A.( Eskaspis) woodburnensis Lr. Camregan grit Wether Lav _ Linn Fm. Deerhope Fm. Reservoir Fm. Encrinurus sp.: Scotoharpes domina A. (Eskaspis) sufferta Encrinurus sp. Leonaspis sp Proetus latifrons Cheirurus sp. Cyphoproetus depressus Calymene sp. Youngia sp. Scotoharpes Podowrinella straitonensis Hemiarges rolfei Encrinurus sp. Eskaspis sp NORTH ESK INLIER PENTLAND HILLS GIRVAN 300- metres text-fig. I . Stratigraphical sections in the Main Outcrop, Girvan, Hagshaw Hills, and North Esk Inker, Pentland Hills, showing distribution of trilobites. Based on Walton (1965) with newer data from Cocks and Toghill (1973), Rolfe (1973), and Tipper (1976), respectively. Lithologies other than conglomeratic horizons omitted for clarity. Arrow heads indicate graptolitic horizons, large asterisks abundant shelly fossils, small asterisks rare shelly fossils. Divisions within the Wether Law Linn Lormation are units A-E of Tipper (1976). 122 PALAEONTOLOGY, VOLUME 20 TAXONOMY Nomenclature Lamont (1965) described Phacops straitonensis from the Hagshaw Hills and Knockgardner. The discovery of this species in the Pentland Hills, and the collec- tion of more material from all three localities gave the incentive for a more detailed study. No question arises over the validity of the specific name, though the holotype has been lost (Lamont, pers. comm.). Podowrinella is here described as a new genus. It seemed to us initially that the poorly known, but suggestively named, genus Pterygometopidella Wedekind, 1912 (based on Phacops quadrilineata Angelin) might be available as a generic taxon for P. straitonensis. Wedekind (1912) described Pterygometopidella very inadequately from only two specimens collected from the c-marl (Silurian) of Gotland, and figured only one of them in a poor and tiny photograph which allowed few details to be seen but which seemed to indicate a close resemblance to Eophacops musheni (Salter). The two specimens, deposited in the collections of the Geologisch-Palaontologisches Institut der Universitat Gottingen, are no longer in the Institute (Jahnke, pers. comm. 1974), and may have been destroyed. Struve (pers. comm. 1974) suggests that Wedekind may have erected the subgenus on an erroneously determined species and Schrank (1972) regards Wedekind’s specimens of P. quadrilineata as incorrectly determined individuals of E. musheni (Salter), indicating that Eophacops is a junior synonym of Pterygometopidella. Though the taxonomic problem of Pterygometopidella may not be entirely resolved, it is clear that this generic name is not available for Phacops straitonensis. A. (E.) suffer ta, on the other hand, was figured, though not described, by Lamont (1948) as Eophacops sufferta , in a paper in which many other trilobites from the Pent- land Hills were also figured but for which no formal descriptions were given. There thus arises the question of the validity of the nomenclature, a matter which Norford (1973) has also referred to in dealing with Scotoharpes. S. domina Lamont was also figured in the same paper by Lamont (1948), and Whittington (1950) rejected the name as being invalid. Norford (p. 12), however, revived the name on the basis of Lamont’s recognizable photograph and few words of description. For the same reasons the specific name sufferta is here considered to be valid, since the photographs of Lamont’s (1948, pi. 1, figs. 21, 22) which show a complete internal mould are clearly recognizable, and Lamont (pers. comm. 1974) has con- firmed that the common phacopid trilobite at Wether Law Linn is, in fact his ‘ Eophacops ’ sufferta , though the type, formerly in his possession, has now been lost. Though Lamont did not name the species in his 1947 paper on the stratigraphy and faunas of the region he states (p. 290) that ‘a large Eophacops is very common at Wetherlaw Linn. It has more than 6 facets per radius in the eye, which rules out E. elliptifrons (Esmark). . . . The rather long pygidium with numerous axial segments, however, points to comparison with E. elliptifrons var glaber Marr & Nicholson . . .’. Lamont’s notes which annotate the Grant Institute Library copy of Haswell’s (1865) guide to the geology and faunas of the Pentland Hills refer to Haswell’s figures (p. 6) of "Phacops stokesii ’ as ‘ E . sufferta ’ and state, ‘It should have shown 7-8 rings in the pygidial axis’ and, ‘Eye usually has 7 facets per column’. CLARKSON ET AL.\ SILURIAN PHACOPINA 123 There thus seems to be no question that the common and only species of phacopid trilobite from the Wether Law Linn Formation is what Lamont recognized as a new species and called E. suffer t a , and it is here redescribed, under the generic name Acernaspis (Eskaspis). SYSTEMATIC DESCRIPTIONS Order phacopida Salter, 1864 Suborder phacopina Struve, 1959 Superfamily phacopacea Hawle and Corda, 1847 Family phacopidae Hawle and Corda, 1847 Subfamily phacopinae Hawle and Corda, 1847 Genus acernaspis Campbell, 1967 Subgenus eskaspis n. sg. Type species. Eophacops sufferta Lamont, 1947. Upper Llandovery (Telychian). Wether Law Linn Lorma- tion. North Esk Inlier, Pentland Hills, Peeblesshire, Scotland. Diagnosis. Relatively small Phacopinae. Cephalon roughly 1-8 times as wide as long; genae somewhat truncate postero-distally (in dorsal view) so that the cephalon is widest just anterior to the posterior branch of the facial suture. Anterior glabella lobe moderately inflated, rather flat on top and shelving more steeply anteriorly. Axial furrows deeply emplaced. Glabellar furrows 3p and 2p lightly impressed; 3p in two distinct, unjoined parts with distal ramus straight, running at an exsagittal angle of 45°, not confluent with axial furrow; proximal ramus of 3p furrows convex anteriorly, inclined slightly antero-laterally. 2p furrows likewise convex and not confluent with axial furrows. Glabellar furrows lp confluent with axial furrow, concave anteriorly and not coalesced mesially, with stout apodemes developed in distal portion of lp external surface. Intercalating ring thus confluent mesially with anterior glabellar lobe; intercalating ring depressed, with distal nodes. Occipital furrow deeply incised laterally. Occipital ring longest (sag.) at the midline, depressed laterally with nodes partially set off by exsag. furrows, not reaching height of anterior glabellar lobe in lateral view. Posterior border furrow lightly impressed, becoming nearly obsolescent laterally but merging with lateral border furrow. Eyes large, nearly reaching, or some distance from, posterior border furrow, and, anteriorly, lateral border furrow, with eye socle developed only as small, depressed area beneath visual surface. Visual surface with sixteen to nineteen dorsoventral files of lenses, all protruding well beyond bounding sclera. Palpebral lobe flat, rather narrow, set off from palpebral area by faint palpebral furrow. Anterior cephalic margin distinct laterally and smoothly continuous with anterior glabellar lobe, disappearing antero-medially. No anterior border furrow developed. Anterior doublure fairly long (sag.), with inflated central lobe sometimes present. Vincular furrow present anteriorly only as paired subfrontal depressions meeting at midline, deeply impressed postero-laterally and bearing nine notches for pleural 124 PALAEONTOLOGY, VOLUME 20 tips. Anterior and posterior regions of vincular furrow unconnected, or connected by a faint furrow. Hypostoma unknown. All parts of exoskeleton minutely granular; auxiliary impression system ovate, developed as depressions devoid of granules on anterior glabellar lobe, and as scars viscerally. Thorax with distinct axial nodes set off by short non-communicating furrows inclined postero-proximally on both the anterior and posterior sides of axial ring. Pleural tips gently rounded. Pygidium with well-defined axial furrows, with about six axial rings and terminal piece. Axis rounded posteriorly, not meeting posterior margin of pygidium. Five pairs of pleurae present; pleural furrows becoming pro- gressively fainter posteriorly; inter-pleural furrows present, but weakly impressed. Remarks. Our diagnosis of Acernaspis ( Eskaspis ) includes character states inclusive of some other, particularly Silurian, Phacopinae. All characters included in the diagnosis do, in fact, vary in general amongst Phacopinae and are of potential value as diagnostic elements. Furthermore, inclusion of ‘primitive’ as opposed to strictly autapomorphous (‘derived’) characters, in a diagnosis, is of value in recognizing distinct genera, for only one or at most a few subordinate taxa in a dehned group may in fact retain a particular primitive feature, hence its valid inclusion in a diag- nosis. The minute granulation of A. (Eskaspis), for instance, is close to the primitive condition for all Phacopinae, yet is typical only of species of A. (Eskaspis) and A. (Acernaspis) among all known Phacopinae. As written, the diagnosis of A. (Eskaspis) embraces some features of only three additional genera known to us: Murphy cops Lesperance, A. (Acernaspis) Campbell, and in ventral morphology Ananaspis Campbell. We emphasize here the distinctly diagnostic traits of Acernaspis (Eskaspis) which serve to distinguish it from these other genera. A. (Eskaspis) shares with Murphy cops some similarity in the coapta- tive device of the cephalic doublure, but otherwise differs from Murphycops in the cuticular granulation, development of auxiliary impression system, presence of a distinct palpebral furrow separating the palpebral lobe from the palpebral area, and the presence of nodes developed on the thoracic axial rings just proximal to the axial furrows. A. (Acernaspis) and A. (Eskaspis) are dorsally very similar to one another, their over-all shape, degree of glabellar inflation, and auxiliary glabellar impressions, for instance are almost identical, and perhaps the only significant difference is the reduction of the anterior border in A. (Eskaspis), so that it is not visible in dorsal view. But we would here draw attention to the ventral morphology of the cephalon, as did Campbell (1967), and because of the coaptative enrolment mechanisms, to the pygidial doublure. Coaptative morphology has not yet been fully explored in Phacopina, but is clearly very important in taxonomy. The ventral cephalic morphology of A. (Acernaspis), here represented by the Idwian A. (A.) elliptifrons (Esmark) from Girvan, Ayrshire, Scotland, is substantially different from that of A. (Eskaspis) suffer la. There are some variations within A. (Acernaspis) judging by the photographs of the doublures of various species given by Campbell (1967), Mannil (1970o, b), and Sherwin (1972), but nowhere is there the CLARKSON ET AL SILURIAN PHACOPINA 125 kind of conformation illustrated by A. ( E .) sufferta. Both elliptifrons and sufferta have nine vincular notches, the last one being indistinct in elliptifrons (text-fig. 3 a-c). In elliptifrons the vincular notches are narrow and all about the same width forming a more or less parallel-sided row deeply indenting the flat surface of the doublure and subparallel with its edges. The notches are all centrally excavated to about the same depth and are joined by shallower passages. The anterior notch runs into a deep vincular furrow behind which the doublure forms a pronounced flat shelf. Sufferta , on the other hand, has a much wider series of obliquely set vincular notches, becoming broader and deeper posteriorly and evidently adapted for receiving a flattened pleural end rather than a more pointed tip. The two ridges separated by the vincular notches are oblique to one another when seen in lateral view, and the outer one is indented by continuations of vincular notches. A. (E.) sufferta , the type species, has no trace of a vincular furrow, but has a pair of shallow subfrontal depressions meeting on the midline. A. ( E .) woodburnensis has these subfrontal depressions as well, but also has a shallow vincular furrow joining each of them to the vincular notches. It is not a deeply incised furrow like that of elliptifrons, however, but a lightly impressed indentation. The vincular morphology of woodburnensis is in many ways intermediate between that of elliptifrons and that of sufferta, being anteriorly more similar to the former, and posteriorly resembling the latter; the three are illustrated here as a morphological series (text-fig. 3). In many ways, however, there is a fair resemblance between the vincular morphology of wood- burnensis, to that of Ananaspis species (Campbell 1967, pi. 14, figs. 9, 13-15), though the presence of the subfrontal depressions links it with Acernaspis ( Eskaspis ) to which it is, with slight reservations, assigned here. A. ( Eskaspis ) seems to have been a localized late Llandovery derivative of the A. ( Acernaspis ) stock, retaining most of its plesiomorphic (primitive) features of dorsal cephalic morphology, but having more advanced ventral morphology. This in woodburnensis is rather like that of the ‘derived’ genus Ananaspis, which it does not closely resemble in dorsal morphology, though it also has features pointing towards the more extreme, and indeed unique construction of the doublure of Acernaspis (E.) sufferta, which seems to have been a terminal endpoint. Acernaspis ( Eskaspis ) sufferta (Lamont, 1947) Plate 18, figs. 1-9; Plate 19, figs. 1-7, 10; text-figs. 2a, f, 3 a, b, 4a, c 1861 Phacops Stokesii M.-Edw. ; Howell and Geikie, p. 134. 1865 Phacops Stokesii Haswell, p. 37, pi. 4, figs. 6, 7. 1867 Phacops Stokesii Henderson, pp. 22-23. 1869 Phacops Stokesii Brown and Henderson, p. 3 1 . 1899 Phacops Stokesii (M.-Edw.); Peach and Horne, p. 597. 1947a Eophacops cf. elliptifrons var. glaber. Marr and Nicholson; Lamont, p. 290. 19476 Eophacops sufferta n. sp. Lamont, p. 6, pi. 1, figs. 21, 22. 1962 Phacops aft. stokesii (Milne-Edwards); Mykura and Smith, p. 138. 1975 Acernaspis sp. Tipper, p. 297 . Material. Neotype, IGS 1034. PI. 18, fig. 1, is here selected since Lamont’s holotype is lost. Other figured material. Gr. E 40282, PI. 18, fig. 2; Gr. E 40279, PI. 18, figs. 3, 7; IGS 5783, PI. 18, fig. 4; Gr. E 40269, PI. 18, fig. 5; Gr. I. 40270, PI. 18, fig. 6; Gr. I. 40262, PI. 18, fig. 8; Gr. I. 40281, PI. 18, fig. 9; 126 PALAEONTOLOGY, VOLUME 20 Gr. I. 40257, PI. 19, fig. 1 ; Gr. I. 40243, PI. 19, fig. 2; Gr. I. 40281, PI. 19, fig. 3; Gr. I. 40260, PI. 19, fig. 4; Gr. I. 40280, PI. 19, fig. 5; Gr. 1. 44093, PI. 19, fig. 6; Gr. I. 40261, PI. 19, fig. 7; Gr. I. 40275, PI. 19, fig. 8; Gr. I. 44094, PI. 19, fig. 10. Additional material examined. Approximately 150 specimens from the Wether Law Linn Formation (Tipper’s units A and C) from the Deerhope Burn and Wether Law Linn localities from the collections of the Grant Institute (Gr. I.) (mainly Tipper and Clarkson Collections), the Royal Scottish Museum (RSM), and the Institute of Geological Sciences (IGS) of Edinburgh. Distribution. This species is confined to the lower Wether Law Linn Formation in the North Esk Inlier, Pentland Hills. It occurs quite abundantly and in good preservation as internal and external moulds at all levels, and complete specimens are not infrequently found. Fragmentary specimens recently discovered by S. Wood in the Reservoir Formation may belong to this species. Diagnosis. A species of Eskaspis lacking connection between the subfrontal depres- sions of the anterior region of the doublure and the posterior, notched portion of the vincular furrow. Eyes large, with nineteen files (rarely eighteen in adults) with a general lens formula of 565 787 877 765 543 2 and about 116 lenses in all. Smaller specimens (approximately half adult size), with eighteen files of no more than six lenses. Lowermost lenses very small, elliptical in some specimens. Dimensions. Holotype: total length (sag.) 20 0 mm, cephalic length (sag.) 5-7 mm, width 13 0 mm. Range in cephalic length 2-2-9 0 mm, width 5-0 17-5 mm. Description. Cephalon relatively broad, width/length ratio being approximately 2:1, and with an almost rounded (subangular) anterior margin (text-fig. 2a). Glabella half the cephalic width at its widest, broadly heptagonal, expanding anteriorly, widest in front of eyes where it is about as broad as its length. Glabellar crown (in profile) slightly higher than palpebral lobes (text-fig. 2d). In lateral profile the glabella is rather flattened on top, but slopes with increasing steepness to meet the anterior margin at about 80° with no real preglabellar furrow or preglabellar field, so that the glabella is directly in contact with the cephalic doublure. Axial furrows moder- ately deeply impressed, running forwards from lobe lp at about 45° to an exsagittal line and changing direction where the palpebral furrow contacts glabellar lobe 3p, thence diverging from the exsagittal line only by some 10°, its course being indented slightly by the anterior edge of the eye. Axial nodes small, usually round but some- times elongate. All glabellar furrows faintly impressed and approximately equidistant from one another. Ip is the deepest; it is discontinuous and anteromedially directed at 45°. EXPLANATION OF PLATE 18 Figs. 1-9. Acernaspis ( Eskaspis ) sufferta (Famont). Wether Faw Linn Formation (Telychian), North Esk Inlier, Pentland Hills. 1, neotype. Wether Law Linn (WL), IGS 1035,* x3-5. 2, complete specimen. Deerhope Burn (DB), Gr. I. 40282, f x2-3. 3, almost complete, somewhat disarticulated specimen (DB), Gr. I. 40279, t x2-5. 4, partially enrolled specimen (WL), IGS 5783,* x 3-5. 5, anterior doublure, showing subfrontal depressions, antero-ventral view (WL), Gr. I. 40269,* x4-5. 6, eye and vincular notches, ventro-lateral view, Gr. 1. 40270,* x 5. 7, right eye, Gr. I. 40279, f x 10. 8, doublure (cf. text-fig. 3) (DB), Gr. I. 40262,* x5. 9, almost complete specimen in lateral view (DB), Gr. I. 40281, t x 4. * Latex replica of external mould, f Internal mould. PLATE 18 CLARKSON, ELDREDGE and HENRY, Acernaspis 128 PALAEONTOLOGY, VOLUME 20 text-fig. 2. Acemaspis ( Eskaspis ) sufferta (Lamont). Reconstructions of a, whole animal in dorsal view. b-d , cephalon in ventral, frontal, and lateral view, e, thoracic segment showing facet in lateral view. f pygidium in lateral view, g , cephalon in oblique lateral view, showing subfrontal depressions (s). a-f. , x 3-5; g, x 8-5. CLARKSON ET AL.: SILURIAN PHACOPINA 129 Furrow 2p nearly transverse with the outer tip recurved; 3p in two distinct unjoined parts, inner part directed slightly forwards, outer part at 45° and not reaching the axial furrow. Intercalating ring not sharply demarcated anteromedially, but con- fluent with composite glabellar lobes. Anterior lobe half the total length of the glabella. Auxiliary glabellar impressions form a circular pattern, some are quite deeply impressed (text-fig. 5) (cf. Acernaspis figured by Eldredge 1971, text-fig. 2j). Occipital furrow straight and shallow medially, more deeply impressed laterally, at the apodemal pit, occipital ring very large (as wide as the posterior part of the glabella including lateral nodes) and with prominent lateral lobes directed forwards and outwards at 45°. Genal region with a definite lateral border which is widest postero-laterally, and delimited by a broad, shallow, lateral border furrow, less distinct below the eye, but becoming more incised anteriorly and nearly reaching the glabella. Posterior border furrow straight and deeply incised. Genal angles truncated, projecting somewhat posteriorly and subangular. Palpebral lobes large, slightly inflated, and with a faint external rim extending from furrow 3p to the occipital furrow. Palpebral furrow usually distinct though may be rather vaguely defined in the central part. Palpebral area quite broad, tri- angular, mainly flat or slightly inflated. Eyes very large, nearly half the total height of the cephalon, the posterior edge slightly further from the midline than the anterior (PI. 18, fig. 7; text-fig. 1 g). Lenses projecting beyond the sclera, not sunken, arranged in nineteen dorso- ventral files (commonly eighteen in juveniles but rarely in adults) with a general lens formula: 567 787 877 765 543 2. Maximum per file eight. Total 1 16. In some specimens the lowermost lenses are very small and elliptical. Eye socle very reduced. Facial suture apparently functional in holaspids, the anterior branch cutting antero-lateral corners of the anterior face of the glabella, meeting at a distinct point on the midline. Posterior branch directed at first anteriorly, then curving out- wards and backwards again to reach the border opposite the posterior edge of the eye. The suture appears again on the postero-lateral part of the doublure making a posteriorly convex, symmetrical U-shaped lobe behind the course of the dorsal part of the suture (text-fig. 3). Cephalic doublure nearly a third the length of the glabella sagittally its rear margin convex posteriorly, the central part convex down- wards. Vincular furrow marked anteriorly only as a pair of elongate, contiguous, and shallow subfrontal depressions on either side of the midline (text-fig. 1 b, g (marked ‘s’)). Where the depressions meet, there is a tiny forwardly projecting point. Subfrontal depressions separated from the notched lateral vincular furrow which extends from below the anterior edge of the eye to nearly below its posterior edge. The inner lamella of the lateral doublure is less steeply inclined in lateral aspect than the outer lamella which forms the postero-lateral margin of the cephalon. Posteriorly, the inner lamella ends abruptly below the dorsal facial suture in front of the U-shaped lobe; the outer lamella extends further back (text-fig. 3). The most anterior notches are less deeply impressed than those which lie posteriorly; nine notches are evident in all, and are arranged radially from a point posterior to the U-shaped lobe defined by the doublural suture. The doublure turns abruptly at the i 130 PALAEONTOLOGY, VOLUME 20 text-fig. 3. The lateral part of the ventral doublure, showing vincular structures and coaptative morphology in Acernaspis ( Acernaspis ) and A. (Eskaspis). a , A. ( Acernaspis ) elliptifrons (Esmark), from the Newlands Formation (Idwian), Girvan, Ayrshire. BM It 9121. fi, A. ( Eskaspis ) woodbumensis n. sp. from Wood Burn Formation (lowermost Telychian), Lauchlan Burn, Girvan, Ayrshire. IGS 5777. c, A. ( Eskaspis ) sufferta (Lamont), Wether Law Linn Formation (upper Telychian), Pentland Hills, near Edinburgh. Gr. I. 40262 (see PI. 18, fig. 8), all x 14. All these illustrations are camera-lucida drawings made from latex replicas of external moulds. genal angle and tails off to terminate just behind the posterior edge of the eye. Occipital and lp apodemes distinct on ventral surface of the cephalon. Thorax of eleven segments tapering backwards slightly, having quite strongly arched rings. Axis is quarter the total width. Each axial ring is indented anteriorly by deep paired incisions, which together with shallower posterior incisions placed closer to the axial furrows delimit chevron-shaped axial lobes, set further forward than the axial part of the ring but are not swollen as nodes (text-fig. 2a , e). Posterior border of each axial ring high and sharp. A distinct furrow divides the articulating half-ring from the axial ring. Pleural regions flat but strongly bent down distally. Pleural furrows straight or slightly curving, narrow, of moderate depth, and extend as far as the articulating facets. Facets very pronounced anteriorly while the pleural ribs are very narrow and curved, but diminishing in size and curvature posteriorly. The rear edge of the flat top of each pleuron has a narrow flange articulating with a corresponding anterior groove on the next segment. CLARKSON ET A L. : SILURIAN PHACOPINA 131 Pygidium with rounded or almost imperceptibly pointed margin with nine to ten rings diminishing in size posteriorly (text-fig. 2 a, /). The last two axial rings are indistinct, the first three or four have a pseudo-articulating half-ring. There are four or five pleural furrows whose obliquity increases backwards, the first being quite deep, the others increasingly shallow, interpleural furrows weakly impressed, and disjunct. All pleural and interpleural furrows fade out some distance before the pygidial margin, leaving a broad smooth border zone. Pygidial doublure broad, with two lateral parts forming a small notch in the sagittal line and shaped to fit the two vincular depressions on the cephalon. All parts of the exoskeleton are covered with dense microgranular sculpture except for the subocular groove which is smooth. On the doublure, glabella, and axial rings and pleural ribs this sculpture is especially dense, though it is sparser on the flat tops of the pleurae. f text-fig. 4. Acernaspis ( Eskaspis ) sufferta (Lamont). Camera-lucida drawings showing auxiliary impres- sion patterns on the anterior glabellar lobe. All x5. a , Gr. I. 40271; b , Gr. I. 40263; c, Gr. I. 40281; d, Gr. I. 40265; e, Gr. I. 40264;/, Gr. I. 40257. 132 PALAEONTOLOGY, VOLUME 20 Acernaspis ( Eskaspis ) woodburnensis , n. sp. Plate 19, figs. 9, 11-13; text-fig. 5a, b 1899 Phacops stokesi (M.-Edw.). Peach and Horne, p. 538. 1906 Phacops elegans Sars and Boeck. Reed, p. 155. Locus tvpicus. Lauchlan Burn (Bargany Pond Burn, Girvan, Ayrshire). Stratum typicum (Wood Burn Formation, lowermost Telychian) (Cocks and Toghill 1973, p. 225) Diagnosis. Axial furrows straight, diverging forwards to a point in front of the eye. Eyes fairly large, not particularly high, posterior edge more distant than the anterior edge from the sagittal line. Lenses distributed in sixteen files as follows: 566 677 776 766 534 3. Maximum per file seven. Total 90. Posterior section of doublure has ?nine vincular notches on each side connected by a shallow vincular furrow to the subfrontal depressions. Types. Holotype IGS 5777. Paratypes IGS 5778-5779, 5780-5781 (parts and counterparts). Other material. Ten specimens in IGS collections. Remarks. A. ( E .) woodburnensis is the oldest member of this subgenus known at present but is stratigraphically younger than the middle Llandovery A. (A.) elliptifrons of the Newlands Formation. It is particularly interesting that here a text-fig. 5. Acernaspis (Esk- aspis) woodburnensis, n. sp. a, b, cephalon reconstructed in dorsal and ventral view, x 2-25. EXPLANATION OF PLATE 19 Figs. 1-7, 10. Acernaspis (Eskaspis) sufferta (Lamont). Wether Law Linn Formation (Telychian), North Esk Inlier, Pentland Hills. 1, cephalon with auxiliary impression patterns, Gr. I. 40257,* x 3-5. 2, cepha- lon and part thorax, oblique anterolateral view, Gr. I. 40243,* x 3-75. 3, almost complete specimen, oblique antero-lateral view, Gr. I. 4028 1 , f x5-25. 4, pygidium, Gr. I. 40260,* x6. 5, almost complete enrolled specimen, Gr. I. 40280,f x2-75. 6, gentd point, visible only on internal mould, Gr. I. 44093, f x 15. 7, small cephalon, Gr. I. 40261, f x4. 8, almost complete specimen, Gr. I. 40275,* x3-75. 10, doublure of small specimen, Gr. I. 44094,| x 5-5. Figs. 9, 11-13. A. (Eskaspis) woodburnensis, n. sp. Wood Burn Formation (lowermost Telychian). ‘Bargany Pond Burn' = Lauchlan Burn, Maxwellton Hill, Girvan. 9, ventral view of holotype cephalon showing doublure with shallow furrow connecting vincular notches to subfrontal depressions, IGS 5777, x3-5. 11, holotype, IGS 5777 x2-5. 12, 13, anterior and dorsal views of complete specimen retaining its cuticle though altered and exfoliated in places, IGS 5780, x 2. * Latex replica of external mould, f Internal mould. PLATE 19 CLARKSON, ELDREDGE and HENRY, Acernaspis 134 PALAEONTOLOGY, VOLUME 20 shallow vincular furrow is present, though absent in the otherwise quite similar A. (E.) suffer ta. Dimensions. Holotype: total length (sag.) 28 0 mm, cephalic length (sag.) 10 0 mm, width 18-5 mm. Range in cephalic length 9-5-1 1-0 mm, width 17 0-20-0 mm. Superfamily dalmanitacea Vogdes, 1890 Family pterygometopidae Reed, 1905 Genus podowrinella gen. nov. Type species. Podowrinella straitonensis (Lamont, 1965) by original designation and monotypy, from the Knockgardner Formation (?Wenlock) of the Blair-Straiton district near Girvan, Ayrshire, and the Ree Burn Formation (?Telychian, upper Llandovery), Hagshaw Hills, and the Deerhope Formation (Tely- chian), North Esk Inlier, Pentland Hills. Etymology. From the Podowrin Burn in the Hagshaw Hills where the Ree Burn Formation outcrops and where fossils are common in the turbidites. Diagnosis. Small Phacopina combining features of both phacopid and pterygometo- pid organization. Cephalon roughly triangular but with truncated genal angles. Glabella pear-shaped with furrow lp connected to 3p as in Pterygometopidae but 3p in two distinct parts as in Phacopidae, and confluent with axial furrows. Eyes with about eighteen files with up to seven lenses each, about 100 lenses altogether. Cephalic doublure traversed by deep, broad coaptative grooves, continuous laterally, becoming dorso-medial just before they meet. Thorax with axial nodes and strongly downturned outer parts of the pleurae. Pygidium somewhat triangular with a rounded or occasionally slightly pointed margin. Podowrinella straitonensis (Lamont, 1965) Plate 20, figs. 1-14; text-fig. 6 a-h 1962 Phacops sp. nov. Rolfe, p. 252. 1965 Phacops straitonensis sp. nov. Lamont, p. 39, pi. V, fig. 5. 1973 Phacops straitonensis. Rolfe, p. 109. EXPLANATION OF PLATE 20 Podowrinella straitonensis (Lamont). 1, neotype cranidium, figured by Lamont (1965, pi. V, fig. 5), almost undistorted. Ree Burn Formation, Hagshaw Hills (Rolfe’s loc. 31), BU 1900,* x6. 2, cranidium in W-mode deformation. Top of Deerhope Formation, Deerhope Burn, North Esk Inlier, Pentland Hills (locality B), Gr. I. 40340,* x5-5. 3, thoracic segment in lateral view. Pentland Hills (loc. B), Gr. I. 40348,1 x 3-5. 4, 5, anterolateral and ventrolateral view of detached librigena showing vincular furrows and eye, small quarry in the Ree Burn Formation, nearly opposite junction of Shiel Burn with Monk’s Water, Hagshaw Hills (loc. A), RSM 1975-43-2,f x6-5. 6, ventral surface showing vincular furrows, RSM 1975-43-3, f x 6. 7, undistorted cranidium. Knockgardner Formation, Knockgardner, IGS M 1 32 1 9,f x3-5. 8, cranidium. Pentland Hills (loc. B), Gr. I. 40310,* x5-55. 9, lateral view of com- pressed pygidium. Pentland Hills (loc. B), Gr. I. 40345,* x 7-5. 10, ventral surface W-mode deforma- tion). Pentland Hills (loc. B), Gr. I. 40358, f x 6. 11, left eye and subjacent librigena. Hagshaw Hills (loc. A), RSM 1975-43-4,* x7-5. 12, pygidium. Hagshaw Hills (loc. A), RSM 1975-43-5, x5-5. 13, pygidium (L-mode deformation). Pentland Hills (loc. B), Gr. I. 40349,* x5. 14, pygidium show- ing accessory half-ring on segment 1. Pentland Hills (loc. B), Gr. I. 40307,* x 6-5. * Latex replica of external mould, f Internal mould. PLATE 20 CLARKSON, ELDREDGE and HENRY, Podowrinella 136 PALAEONTOLOGY, VOLUME 20 Material (a) Hagshaw Hills. ( 1 ) Twenty-five specimens from the Rolfe Collection, University of Birmingham, including BU 1900a, b figured by Lamont (1965, pi. V, fig. 5), which is herein designated as neotype, since the unfigured holotype from Knockgardner is now lost (Lamont pers. comm.). Neotype. BU 1900a, b (from Rolfe’s locality 31/48 in unpublished thesis; University of Birmingham 1960). Other material. Thirty other specimens in the collection of the Department of Geological Sciences, University of Birmingham. (2) One hundred specimens from the Ritchie Collection, Royal Scottish Museum, collected from a small quarry in the Ree Burn Formation on the east bank of the Monk’s Water, opposite the mouth of the Shiel Burn, Hagshaw Hills. These include the paratypes RSM 1975-43-1, RSM 1975-43-2 (PI. 20, figs. 4, 5), RSM 1975-43-3 (PI. 20, fig. 6), RSM 1975-43-4 (PI. 20, fig. 1 1), RSM 1975-43-5 (PI. 20, fig. 12). Also twenty-five RSM specimens from the foot of the Podowrin Burn, east of Monk’s Water, Hagshaw Hills. In addition, fifty-five other specimens from the same localities in the Grant Institute Collections (Gr. I.). ( b ) North Esk Inlier. Specimens from Tipper and Clarkson Collections. Top of Deerhope Formation, Deerhope Burn. Paratypes Gr. I. 40340, PI. 20, fig. 2; Gr. I. 40348, PI. 20, fig. 3; Gr. I. 40310, PI. 20, fig. 8; Gr. I. 40348, PI. 20, fig. 9; Gr. I. 40358, PI. 20, fig. 10; Gr. I. 40349, PI. 20, fig. 13; Gr. I. 40307, PI. 20, fig. 14. (c) Knockgardner. Blair-Straiton Inlier, east of Girvan. Maconochie Collection. IGS M 13219 (PI. 20, fig. 7) and twenty-five other IGS specimens. Also thirty specimens in Grant Institute of Geology. Range. Telychian (North Esk, Hagshaws)— ?lower Wenlock (Knockgardner). Remarks. The material available from Knockgardner and the Hagshaw Hills is all preserved as disarticulated fragments collected from the base of turbidite flows. Specimens are found in all orientations and most individuals have suffered some distortion. The notable variation in width and length between specimens resembling the L and W forms of Henningsmoen (1960, p. 207) and Sadler (1974, pp. 81, 85) is attributed to stress due to crushing pressure of the mass of unconsolidated turbidite after deposition. Some sheared specimens occur also. The reconstructions were made from the less distorted specimens. Dimensions. Holotype: cephalic length (sag.) 5 mm, width 10 mm. Range in cephalic length 4-6 mm, width 8-12 mm. Description. Small Phacopidae with roughly triangular cephalic outline (in dorsal view), modified by truncated genal angles. Cephalon widest at distal librigenal/ fixigenal junction. In lateral view (text-fig. 6c) (in standard orientation), glabella depressed anteriorly, sloping regularly postero-dorsally, with central region of glabella extending above level of palpebral lobes. Large anterior arch developed in frontal view. Axial furrows moderately deeply implaced, diverging at an angle of approximately 65 then abruptly reflected directly in front of the anterior edge of the palpebral lobe, and curving smoothly around anterior glabellar lobe, becoming confluent with anterior and lateral cephalic border furrows. Glabella pear-shaped, with anterior, 3p, 2p, and medial portion of lp lobes united to form composite glabellar lobe (text-fig. 6a). Anterior glabellar lobe sub- ovate, continuous with medial portion of remainder of glabella. Anterior margin of anterior glabellar lobe rounded, bounded by furrow bearing anterior ramus of CLARKSON ET AL.\ SILURIAN PHACOPINA 137 facial suture antero-laterally. Facial suture continuous with anterior border furrow medially; anterior border, border furrow, and facial suture thus meeting in a distinct point medially, delimiting a small, triangular pre-glabellar field proximal to facial suture. Anterior border thin, though distinct in front of this field; concentric with facial suture, but disappearing marginally. Glabellar furrow 3p in two parts, with sharp bend between distal and proximal branches; branches generally connected at bend by shallow furrow, occasionally text-fig. 6. Podowrinella straitonensis (Lamont). a-c, reconstructions of cephalon in a, dorsal, b , frontal, c, lateral aspects, d , cephalic doublure in ventral view, e, f, thoracic segment in dorsal and lateral view, g, h , pygidium in dorsal and lateral view. All 4. wholly disconnected on internal and external moulds. Distal branch of 3p furrow deepest proximally, confluent with axial furrow just opposite anterior margin of visual surface, slightly convex anteriorly, running at an angle of approximately 55e from the midline. Proximal branch of 3p furrow comma-shaped, with proximal terminus slightly posterior to distal terminus. Proximal terminus continuous with a broad, shallow, variably developed longitudinal glabellar furrow running pos- teriorly and communicating smoothly with lp glabellar furrows. Glabellar furrow 2p short, comma-shaped (convex anteriorly) with distal and proximal tips in trans- verse line, and not communicating with axial furrow distally, but confluent with 138 PALAEONTOLOGY, VOLUME 20 longitudinal glabellar furrow medially. Glabellar furrow lp deeply incised distally. confluent with axial furrow, and recurved antero-proximally. Glabellar furrow lp apparently continuous medially as broad, shallow furrow, delineating an intercalating ring. Glabellar lobe 3p hypertrophic distally; ratio of glabellar lobe lengths along axial furrow roughly: 4-7:2: 1 for 3p:2p; lp. 2p lobe small, ovate, lp lobe longest medially, reduced to nodes distally, and depressed below concordant surface of anterior glabellar region and occipital ring. Occipital furrow shallow, straight; occipital ring of approximately equal length (sag.) throughout. Posterior border furrow shallow, communicating with axial furrow, running nearly transversely, distally becoming more shallow and recurved antero-distally, confluent with lateral border furrow of librigena. Posterior border distinct, roughly transverse proximally, becoming thickened distally near genal margin, thereby causing slight deflection to posterior margin of cephalon. Posterior border sharply reflected antero-distally, forming a bluntly pointed genal angle of approximately 100°. Lateral cephalic border reflexed antero-proximally at point where crossed by posterior ramus of facial suture. Lateral border continuous with -anterior border; anterior border becoming progressively thinner approaching midline. Lateral border of librigena steeply inclined, capped by sharp ridge; lateral border furrow broad, smoothly confluent with concave distal portion of librigena (text- fig. 6c). Librigena becoming steeply inclined proximally, merging with eye socle (not distinctly defined because of lack of ornamental features due to poor preserva- tion). Visual surface nearly vertical, with eighteen d-v files with up to seven lenses per file, about 100 lenses in all. Eye moderately long, not reaching posterior border furrow. Palpebral lobe quite narrow, inflated above palpebral area and set off from it by a sharply implaced palpebral furrow. Palpebral area relatively flat, sloping postero-proximally and smoothly confluent with fixigenal area posterior to eye. Cephalic doublure produced anteriorly into flat trapezoidal spatulate process, inclined at an angle of approximately 25° (in lateral view) and visible in frontal view. Spatulate process bounded distally by deep, broad grooves traversing doublure, and inclined antero-proximally at an exsagittal angle of approximately 35°. Grooves continuous laterally, becoming dorsal where converging at midline (text-fig. 6b, d). Anterior-most medial portion of cephalon a thin ridge where anterior border, vincular groove, and spatulate process coterminate. Cephalic doublure developed as thin ridge postero-distally to vincular grooves; inner wall nearly vertical. Hypo- stoma unknown. Auxiliary impression pattern and details of cephalic ornament unknown; cephalon apparently smooth (based on study of external moulds). Thorax incompletely known; most of the thoracic segments examined (probably the anterior ones) with the axis one-quarter the width of the whole, axial nodes set off by short, deep anterior furrows (text-fig. 6e,f). Pleural furrows relatively shallow, of moderate depth, and straight. Inner part of pleura flat, outer third sharply turned down; pleural tips pointed. Pleural facets present but not very distinct. Pygidium (text-fig. 6 g, h) equilaterally triangular with, however, its posterior apex truncated, having approximately eight axial rings and terminal piece; two or more accessory half-rings present. Axis not reaching posterior pygidial margin, acutely CLARKSON ET AL.: SILURIAN PHACOPINA 139 rounded posteriorly. At least four pairs of pleurae present; inter-pleural furrows obscure. Pleural furrows becoming fainter posteriorly, and nowhere reaching pygidial border. Most specimens have a rounded pygidium, but in some examples the posterior border is slightly pointed, especially in those from the Pentland Hills. Remarks. This genus cannot be readily confused with any other taxon known to us. To clarify its diagnosis and to aid in recognition of related forms we briefly discuss which features are peculiar to the taxon alone, and which are ‘phacopid-like’ and ‘pterygometopid-like’. There is no question that Podowrinella is, in the great majority of its features, pterygometopid in aspect. In fact, Podowrinella is very similar indeed to the Ordo- vician pterygometopids Calyptaulax and Eomonarachus, in particular to those species of Calyptaulax from Europe which, as Shaw (1974, p. 42) has noted, have a ‘noticeable bend’ along glabellar furrow 3p. Pterygometopid-like features of Podowrinella include over-all conformation of cephalon; possession of a distinct lateral and anterior border (reduced anteriorly in Podowrinella ); glabellar segmenta- tion, generally, including size and shape of glabellar lobes and presence of longitudinal glabellar furrows; narrow palpebral lobes; anterior branch of facial suture meeting in a distinct point; presence of a spatulate medial projection of the cephalic doublure; and a somewhat triangular pygidium, pointed in some examples. Podowrinella is phacopid-like in the following features: presence of a 3p glabellar furrow divided into two parts, discontinuous in some specimens of P. straitonensis , with the proximal moiety nearly transversely oriented (the distal ramus remains pterygometopid-like in conformation); a smoothly rounded anterior glabellar margin; truncated genal angles, without genal spines; nearly complete loss of inter- pleural furrows, and sharp reduction in the number of pleural furrows, on the pygidium. We hasten to add that some species of Calyptaulax , particularly C. ealli- cephala , are very much like Podowrinella in having a smoothly rounded anterior glabellar margin and truncated genal angles; supposed ‘phacopid-like’ characters. Finally, Podowrinella appears to be rather unique in the coaptative doublure morpho- logy, though some pterygometopid species, again, particularly C. callicephala , are somewhat similar in possessing a vincular furrow traversing the doublure and converging antero-dorsally at the midline. Though we have been unable to reach a final agreement on the taxonomic position of Podowrinella , the unusual combination of characters which it displays does reinforce Eldredge’s (1971) suggestions as to a close relationship between Phaco- pidae and Pterygometopidae. There are some interesting resemblances between A. ( Eskaspis ) and Podowrinella , in terms of glabellar furrowing, eye size and posi- tion, and genal angles. Sometimes specimens of A. (E.) sufferta show traces of extended glabellar furrows, so that lp and 3p are faintly connected (e.g. the slightly crushed specimen in PI. 1 9, fig. 1 ). There seem to be possible grounds for restructuring the current classification of Phacopina and the curious morphology of Podowrinella could be critical. 140 PALAEONTOLOGY, VOLUME 20 Coaptative structures in A. (Acernaspis), A. (Eskaspis), and Podowrinella : some remarks Within trilobites of the Phacopina and other suborders there have been described a substantial array of structures on the ventral doublure of the cephalon which inter- lock during enrolment with mirror-image structures on the thorax and the pygidium. Examples have been figured by Campbell (1967), and others have been discussed in detail by Henry and Nion (1970), Clarkson and Henry (1973), and Henry and Clark- son (1975). These have arisen independently in a number of trilobite groups and have been defined (Hupe 1953; Clarkson and Henry 1973) as coaptative mechanisms. They have, in our estimation, much potential value in taxonomy, as well as their intrinsic interest from the functional and evolutionary point of view, but before this can be effectively realized, it is essential to illustrate and document the various types present in different trilobite groups. This study gives information of the nature of coaptative mechanisms of A. (Acern- aspis), A. (Eskaspis), and Podowrinella. In A. (Eskaspis) (text-fig. 3c) there are nine vincular notches on each side which receive the tips of the thoracic pleurae. Though no complete enrolled specimens have been found which might enable further details of the mechanism to be elucidated, it appears that these notches are for the reception of the last nine pleural tips; the first two pleurae are shorter and have wider facets. During enrolment, they probably locked in against the end of the inner ridge of the doublure (text-fig. 3) close together with the tip of the third pleuron. The more posterior pleurae were fitted into depressions which became progressively more spaced out towards the anterior of the doublure, whilst becoming less deeply impressed. By contrast with A. (Acernaspis) which has a well-marked semicircular vincular furrow connecting the two lateral sets of vincular notches (text-fig. 3a), A. (E.) woodburnensis has only a very shallow furrow, connecting with the two subfrontal depressions, whilst A. (E.) suffer ta only has the two depressions, and no furrow at all. In both species of A. (Eskaspis) two slightly swollen areas on the pygidial doublure fit into the subfrontal depressions; otherwise the pygidium comes to rest freely against the flat (or only slightly indented in the case of A. (E.) woodburnensis surface of the anterior part of the cephalic doublure. This difference between the coaptative morphology of the doublure of A. (Acernaspis) and A. (Eskaspis) is coupled with the fact that in the former subgenus the surfaces of the parallel vincular ridges lie in the same plane, whereas in the latter they are somewhat oblique to one another, so that the inner ridge as seen in lateral profile having the higher angle to the hori- zontal; there are other differences also as previously mentioned (p. 125). It is not certain whether the morphological sequence illustrated in text-fig. 3, and previously referred to is a true evolutionary sequence. Had the specimens all been collected from a single vertical section it would be almost unequivocal. Even so, the morpho- logy, confined distribution, and stratigraphically sequential changes suggest a defini- tive evolution towards the grade of organization in A. (E.) suffer ta within a fairly localized area. Podowrinella has no separate notches for the reception of individual pleural tips, but only a pair of elongated furrows, deepening posteriorly and becoming slightly twisted in the process. With this type of structure the pleural tips come to rest in the groove, as does the doublure of the pygidium, and presumably the triangular point CLARKSON ET AL.\ SILURIAN PHACOPINA 141 at the extreme anterior of the cephalon forms a seal with the similarly angular sagittal edge of the said pygidial doublure. Such coaptative morphology resembles that of some Pterygometopidae though in a more extreme form. The range in morphology of coaptative structures amongst Phacopina is con- siderable; compare, for example, the ventral morphology of the doublure in Kloucekia , Morgatia, and Crozonaspis species (Henry and Nion 1970; Clarkson and Henry 1973), with that of the trilobites discussed in this paper. The value of these in taxonomy is probably very great, but there is, as present, a need for detailed study covering a wider spectrum than is presently known, and only then will their actual value be able to be established. Acknowledgements. We thank Dr. John Tipper who collected many of the specimens, discussed many points with us, and read critically the final manuscript. Dr. Charles Waterston and Mr. William Baird of the Royal Scottish Museum (RSM), Edinburgh, loaned specimens and discussed localities, as did Peter Brand of the Institute of Geological Sciences, Edinburgh (IGS), who first drew the specimens of A. ( E .) woodburnensis to our attention. Dr. Isles Strachan, University of Birmingham (BU) loaned specimens from the Rolfe Collection. Mr. B. Johnstone prepared most of text-fig. 1. Finally, Dr. Archie Lamont of Carlops, Peeblesshire, who encouraged us to work on these trilobites, corresponded with us throughout, and helped with many queries. A substantial grant towards costs of illustrations from the Carnegie Trust for the Universities of Scotland is gratefully acknowledged. REFERENCES brower, j. c. 1975. Silurian crinoids from the Pentland Hills, Scotland. Palaeontology , 18, 631-656, pis. 73-75. brown, d. j. and Henderson, j. 1867. On the Silurian rocks of the Pentland Hills, with notes upon the Brachiopods by Thos. Davidson. Trans. Edinb. Geol. Soc. 1, 23-33. Campbell, K. s. w. 1967. Henryhouse Trilobites. Bull. Oklahoma Geol. Surv. 115, 1-68, pi 1-19. Clarkson, e. n. k. and henry, J.-L. 1973. Structures coaptatives et enroulement chez quelques Trilobites ordoviciens et siluriens. Lethaia, 6, 105-132. cocks, l. r. m. and toghill, p. 1973. The Biostratigraphy of the Silurian rocks of the Girvan District, Scotland. JIGeol. Soc. 129, 209-243. davidson, t. 1874. The Silurian Brachiopods of the Pentland Hills. Trans. Geol. Soc. Glasgow. Palaeont. Ser. 1, 1-24, pis. 1-3. eldredge, n. 1971. Patterns of cephalic musculature in the Phacopina (Trilobita) and their phylogenetic significance. J. Paleont. 45, 52-67. etheridge, R. 1874. Notice of additional Fossils from the upper Silurian of the Pentland Hills. Trans. Edinb. Geol. Soc. 2, 309-314. haswell, c. c. 1865. On the Silurian formation in the Pentland Hills. Willison P. Nimmo, Edinburgh, I -48, pis. 1-4. Henderson, j. 1867. Short notice of three species of trilobites from the Silurian beds of the Pentland Hills. Trans. Edinb. Geol. Soc. 1, 21. 1874. On some Silurian fossils found in the Pentland Hills. Ibid. 2, 373. and brown, d. j. 1869. On the Silurian rocks of the Pentland Hills. Ibid. 1, 266-272. henningsmoen, G. 1960. The Middle Ordovician of the Oslo Region 13. Trilobites of the Family Asaphidae. Norsk Geol. Tidskr. 40, 203-257. henry, j.-l. and nion, j. 1970. Nouvelles observations sur quelques Zeliszkellinae et Phacopidellinae de l’ordovicien de Bretagne. Lethaia , 3, 213-224. — and clarkson, e. n. k. 1975. Enrollment and coaptations in some species of the Ordovician trilobite genus Placoparia. Fossils and Strata , 4, 87-95, pis. 1-3. 142 PALAEONTOLOGY, VOLUME 20 howell, h. h, and geikie, a. 1861. The Geology of the Neighbourhood of Edinburgh. Mem. Geol. Surv. U.K. 1-151. hupe, p. 1953. Classe des Trilobites (Trilobita Walch 1771). In Traite de Paleontologie (dir. J. Piveteau), III, 44-246. Paris, Masson et Cie. kier, p. M. 1973. A new Silurian echinoid genus from Scotland. Palaeontology , 16, 651-663, pis. 80-83. lamont, a. 1947. Gala-Tarannon Beds in the Pentland Hills, near Edinburgh. Geol. Mag. 84, 193-208, 289-303. 1948. Scottish Dragons. Quarry Managers' Journal , 31, 531-535, pi. 1. 1949. New species of Calymenidae from Scotland and Ireland. Geol. Mag. 86, 313-323. 1952. Ecology and correlation of the Pentlandian— a new division of the Silurian system in Scotland. Rep. 18th Geol. Congr. London , 10, 27-32. — 1954. New Lamellibranchs from the Gutterford Burn (Gala-Tarannon) of the Pentland Hills near Edinburgh. Proc. Roy. Soc. Edin. B65, 271-284. 1955. Scottish Silurian Chelicerata. Trans. Edin. Geol. Soc. 16, 200-216. — 1965. Gala-Tarannon Trilobites and an Ostracod from the Hagshaw Hills, Lanarkshire. Scott. Journal Science, 1, 33-46 (privately published). lesperance, p. j. 1968. Ordovician and Silurian trilobite faunas of the White Head Formation. Perce Region, Quebec. J. Paleont. 42, 81 1-826, pi. 106. mannil, R. 1970o. Estonian Lower and Middle Llandovery trilobites of the genus Acernaspis. Eesti. N.S. V. Tead. Akade. Toimet. Keemia. Geol. 19, 156-165, pis. 1-2. [In Russian with English summary.] — 19706. Phacopid trilobites of the Upper Llandoverian of Estonia. Ibid. 342-349. mykura, w. and smith, j. d. d. 1962. In mitchell, g. h. and mykura, w. The Geology of the Neighbour- hood of Edinburgh. Chapter II, Ordovician and Silurian. Mem. Geol. Surv. Scotland, 10-22. norford, b. s. 1973. Lower Silurian species of the trilobite Scotoharpes from Canada and Northern Green- land. Bull. Geol. Surv. Canada , 222, 3-32, pis. 1 -4. peach, b. n. and horne, j. 1899. The Silurian Rocks of Britain. Vol. 1 : Scotland. Mem. Geol. Surv. U.K. 1-749. reed, f. r. c. 1903-1906. The Lower Palaeozoic Trilobites of the Girvan District, Ayrshire. Palaeontogr. Soc. [Monogr.] 1-183, pis. i-xx. rolfe, w. D. I. 1961. The geology of the Hagshaw Hills Silurian Inlier, Lanarkshire. Trans. Edinh. geol. Soc. 18, 240-269. — 1973. The Hagshaw Hills Silurian Inlier. In bluck, b. j. (ed.). Excursion Guide to the Geology of the Glasgow District. Glasgow Geol. Soc. pp. 105-119. Sadler, p. M. 1974. Trilobites from the Gorran Quartzites, Ordovician of south Cornwall. Palaeontology, 17,71-93. schrank, e. 1972. Proetacea, Encrinuridae und Phacopina (Trilobita) aus silurischen Geschieben. Geologie , 76, 1-117, pis. i-xxi. shaw, F. c. 1974. Simpson Group (middle Ordovician) Trilobites of Oklahoma. Paleont. Soc. Mem. 6 (J. Paleont. 48 suppl.), 1-54. sherwin, l. 1972. Trilobites of the subfamily Phacopinae from New South Wales. Rec. geol. Surv. N.S. W. 13, 83-99, pis. 1-8. spencer, w. k. 1914-1940. British Palaeozoic Asterozoa. Palaeontogr. Soc. [Monogr.] Parts 1-10. tipper, j. c. 1975. Lower Silurian animal communities— three case histories. Lethaia, 8, 287-299. — 1976. The stratigraphy of the North Esk Inlier, Midlothian, Scotland. Scott. J. Geol. 12, 15-22. wedekind, R. 1912. Klassifikation der Phacopiden 1. Deutsche Geol. Gesellsch. Zeitschr. 63, 317-336, pi. 15. Whittington, h. b. 1950. A monograph of the British trilobites of the family Harpidae. Palaeontogr. Soc. [Monogr.] 1-55, pis. 1-7. E. N. K. CLARKSON N. eldredge Grant Institute of Geology West Mains Road Edinburgh EH9 3JW Department of Invertebrate Paleontology American Museum of Natural History Central Park West New York, NY 10024, U.S.A. Typescript received 22 September 1975 Revised typescript received 26 April 1976 j.-l. henry Institut de Geologie Universite de Rennes Avenue du General-Leclerc 35031 Rennes-Cedex, France EVOLUTION OF THE CHAROPHYTE FLORAS IN THE UPPER EOCENE AND LOWER OLIGOCENE OF THE ISLE OF WIGHT by MONIQUE FEIST-C ASTEL Abstract. The distribution of charophyte gyrogonites in the Headon, Osborne, and Bembridge Beds is analysed. The flora, including a new species, Nilellopsis ( Tectochara ) latispira, is described. It is considered that transitional forms in the genera Psilochara and Harrisichara constitute examples of evolution in Palaeogene genera. The strati- graphical importance of Sphaerochara subglobosa (Groves) Horn af Rantzien, which extends from the Upper Headon to the Lower Hamstead Beds, is discussed. It is confirmed that the position of the Eocene-Oligocene boundary that most closely agrees with the charophyte distribution is that situated at the Middle Headon Beds. The Hampshire basin Palaeogene charophytes, which were described by Reid and Groves (1921), Groves (1926), and Grambast (1958), are of considerable stratigraphic importance, as the Hordle and the Bembridge floras form the basis of two charo- phyte zones (Castel 1968; Grambast 1972). The charophyte Verzenay (Hordle) Zone, which includes the mammal Euzet Zone (Feist-Castel 1971), is the highest presently defined in the Eocene. The Bembridge Zone, whose type-locality is the same age as Montmartre in the Paris basin (Stehlin 1909), corresponds to the Early Oligocene (Lattorfian) and to the base of the Middle Oligocene. The Eocene-Oligocene boundary is taken as the base of the Middle Headon Beds, as suggested by Curry (1958)- In order to establish the floristic evolution at the Eocene-Oligocene boundary, specimens were collected from the transi- tional series between the Lower Headon and Bembridge Beds in the Isle of Wight. Graduated samplings from the Middle and Upper Headon Beds, as well as from the Osborne Beds, proved to be very rich in specimens. The floras studied here came primarily from personal collecting. Samples found by Professor D. Curry were also studied. LOCATION OF SAMPLES AND CHAROPHYTE DISTRIBUTION The localities are well-documented cliff-sections on the western and eastern coasts of the Isle of Wight. Their position in the series is defined by the stratigraphic succes- sion generally accepted since the work of Forbes (1856) and described in greater detail later (Bristow 1862; White 1921; Curry 1968). Table 1 showing the sampling localities, is followed by a description of that part of the succession from which samples were taken, and which includes lists of charophyte species found. [Palaeontology, Vol. 20, Part 1, 1977, pp. 143-157, pis. 21-22.] 144 PALAEONTOLOGY, VOLUME 20 table I. Distribution of charophyte localities in the Upper Eocene and Lower Oligocene succession of the Isle of Wight. n. Localities Beds^^^^^ Headon Hill Cliff End Horestone Point Whitecliff Bay BEMBRIDGE • • • OSBORNE • • • UPPER HEADON • • • MIDDLE HEADON • LOWER HEADON • headon hill (National Grid reference SZ 305 860) ( а ) Lower Headon Beds. Only one sample was collected, 1 m above the white Barton sands. Cream-coloured calcareous marls with shells (40 cm): Psilochara polita , P. bi- truncata (scarce), Stephanochara edwardsi, Grovesichara distorta , Harrisichara vasiformis, H. vasiformis-tuberculata. All these species occur within the Lower Headon Beds at Hordle, the flora of which was described by Reid and Groves (1921). (б) Middle Headon Beds. Lower part ; brackish clays with lignite (‘ Neritina beds’) : Gyrogona wrighti , Grambastichara tornata , Char a antennata (scarce). The upper part, the marine ‘Venus beds’ and blue clays, did not yield any charo- phytes. (c) Upper Headon Beds. Samples were collected from two stratigraphical positions in these beds : (i) Approximately 3 m above the Middle Headon blue clays, cream-coloured marls with shells: Gyrogona wrighti , Harrisichara vasiformis-tuberculata , Grovesi- chara distorta. (ii) 3 m above the latter, cream-coloured calcareous clays with shells: Gyrogona wrighti , Sphaerochara subglobosa , Harrisichara vasiformis-tuberculata. In these Upper Headon Beds, Professor D. Curry collected charophyte gyro- gonites which have been determined by the writer to be: Gyrogona wrighti, Grovesi- chara distorta , Chara antennata , Psilochara sp., Harrisichara sp. (scarce). FEIST-CASTEL: CHAROPHYTES FROM THE ISLE OF WIGHT 145 cliff end (National Grid reference SZ 331 891) (a) Upper Headon Beds. Cream-coloured marls lying above the Middle Headon Beds: Grovesichara distorta , Harrisichara vasiformis-tuberculata , Psdochara aff. bitruncata , Sphaerochara subglobosa. This locality also revealed an insectivore mammal tooth, cf. Spalacodon , identified by Dr. B. Sige. (b) Osborne Beds. Samples were collected from two stratigraphical positions in these beds: (i) About 30 m above the Upper Headon Beds, calcareous marls with ferruginous concretions : Nitel/opsis ( Tectochara ) aff. aemula , Harrisichara vasiformis-tuberculata , Psdochara aff. bitruncata , Sphaerochara subglobosa. With this flora was found a rodent tooth identified by Dr. M. Vianey-Liaud as Theridomys ( Theridomys ) pseudo skier olithicus De Bonis, -closely resembling that from La Debruge. (ii) Approximately 10 m above the ironstone band, multicoloured clays with concretions of argillaceous limestone: Gyrogona wrighti, Harrisichara vasiformis- tuberculata, Psdochara aff. bitruncata. horestone point (National Grid reference SZ 634 907) Bembridge Limestone. About 1 m above the beach level: brown, then black, marls, topped by a limestone. (i) Brown marls: Harrisichara tuberculata, Nitel/opsis ( Tectochara ) latispira , Gyrogona wrighti , Grovesichara distorta , Rhabdochara stockmansi (scarce). (ii) Black marls : Harrisichara tuberculata. whitecliff bay (National Grid reference SZ 642 863) {a) Osborne Beds. Yellow limestone with charophytes, topped by green marls containing: Gyrogona wrighti , Harrisichara vasiformis-tuberculata , Sphaerochara sp. (scarce), Chara sp. (b) Bembridge Limestone — lower part. (i) Pale-brown clayey limestone : Harrisichara tuberculata. (ii) Superposed to the latter, black marls : Harrisichara tuberculata , Sphaerochara subglobosa. Material found by Professor D. Curry in the Bembridge Limestone has been found to be: Harrisichara tuberculata , Rhabdochara stockmansi , Gyrogona wrighti , G. caelata. SYSTEMATIC DESCRIPTIONS Family characeae L. Cl. Richard, 1815 Genus gyrogona Lamarck, 1904 ex Lamarck, 1822 emend. Grambast, 1956 Gyrogona wrighti (Salter ex Reid and Groves) Pia, 1927 1856 Chara wrighti Salter in Forbes, p. 160, pi. 7, figs. 15-21. 1921 Chara wrighti Salter ex Reid and Groves, p. 183, pi. 4, fig. 1 . 146 PALAEONTOLOGY, VOLUME 20 1927 Gyrogonites wrighti Pia, p. 90. 1954 Aclistochara wrighti , L. and N. Grambast, p. 59, fig. 1. 1954 Br achy char a wrighti , L. and N. Grambast, p. 667. 1956 Brevichara hordlensis Horn af Rantzien, p. 245. 1957 Gyrogona hordlensis, Grambast, p. 280. 1958 Gyrogona wrighti , Grambast, p. 145. Syntypes. Specimen nos. 74500, 76499, Geological Survey, London. G. wrighti occurs in most of our localities (see Table 2). It is particularly abundant in the limestones, and in the brackish facies such as the Middle Headon Neritina Bed at Headon Hill. Gyrogonites from the same locality present quite notable morphological differences: the shape may be more or less subglobular (length/width ratio varying from 0-9 to 1 0), the spiral cells and apical nodules are faintly prominent, the size varying between broad limits (length 750-1100 |um, width 750-1125 ^m). These differences, however, are within the limits of the species. G. wrighti , known from the Bartonian (Calcaire de Saint-Ouen, Grambast 1972) to the Lower Oligocene, is of no stratigraphic use in this study. Genus grovesichara Horn af Rantzien, 1959 Grovesichara distorta (Reid and Groves) Horn af Rantzien Plate 21, fig- 7 1921 Chara distorta Reid and Groves, p. 186, pi. 5, fig. 6. 1959 Grovesichara distorta, Horn af Rantzien, p. 125, pi. 15, figs. 1-7. 1959 Grovesichara distorta, Grambast, p. 8, fig. 2. Lectotype. Reid and Groves 1921, pi. 5, fig. 6. Lowermost right-hand specimen (designated by Horn af Rantzien 1959). The type-material was found by Grambast (1958) to be at the British Museum and not in the collections of the Geological Survey. This well-known species is easily recognizable by its massive and somewhat irregular shape and by its apex, where the thick and straightened cellular ends form a very convex cap. The basal plug is visible somewhat below the pore. Besides Hamp- shire, G. distorta has been reported from the Auversian and Bartonian by Grambast (1958, 1962) as well as from the Lower Oligocene by Riveline (1973). In the Isle of Wight, it occurs in the Upper and Lower Headon Beds, the latter being its type- stratum. Depending on the localities, variations occur in the length/width ratio of the gyrogonites. Thus, in the Lower Headon Beds at Headon Hill, this ratio is between IT and 1-3, whereas in the Upper Headon Beds at Cliff End it is not far from 1 -0, the general shape being more globular. The specimens gathered in the Upper Headon Beds at Headon Hill were not sufficiently abundant for population studies. Genus nitellopsis Hy, 1889 Sub-genus tectochara L. and N. Grambast, 1954 Nitellopsis ( Tectochara ) aff. aemula Grambast, 1972 Plate 22, fig. 4a-c 1972 Tectochara meriani L. and N. Grambast, 1954 ssp. aemula Grambast, p. 23, fig. 10, pi. 8, figs. 1-6. 1972 Nitellopsis (Tectochara) aemula (Grambast) Grambast and Soulie, p. 11. FEIST-CASTEL: CHAROPHYTES FROM THE ISLE OF WIGHT 147 Holotype. Specimen no. V 41 126, British Museum (Natural History), London. This form occurs in the ironstone band of the Osborne Beds, at Cliff End. It is here considered as allied to N. (77) aemula of the Lower Hamstead Beds, but it is not impossible that it might be hereafter attributed to a new and distinct species. Charac- ters common to both are the ovoid shape with a thin basal region, the similar dimen- sions, and the basal pore of small diameter. The form from Cliff End differs from N. ( T .) aemula by its somewhat longer general shape, the less prominent base, the less marked basal funnel, and the basal plug which is relatively thick for a Tectochara. text-fig. 1. Nitellopsis ( Tectochara ) latispira n. sp. Histo- grams showing the variation in gyrogonite length (L), width (1), and number of convolutions (N). One hundred specimens measured. Nitellopsis ( Tectochara ) latispira n. sp. Plate 22, fig. 3 a-d Holotype. Specimen no. C.F. 1584-1 ; Universite des Sciences et Techniques, Montpellier, France. Paratypes. Specimen nos. C.F. 1 584-2 to 1 584-5 ; Universite des Sciences et Techniques, Montpellier, France. Type-locality and horizon. Horestone Point, Isle of Wight (SZ 634 907); brown marls at the base of Bem- bridge Limestone. Material. About 500 specimens. Diagnosis. Gyrogonite subglobular to ovoid, apex truncated or slightly convex, basal region thin, protruding, and truncated. Cells somewhat thinner and narrower at periphery of apex ; apical ends of cells gradually widening, usually with only slightly prominent nodules. Basal ends entirely calcified and delimiting a wide basal funnel. Pore small, measuring 50-80 ^m; relatively thick plug of 170-280 pm in diameter and 130-180 pm in thickness. 148 PALAEONTOLOGY, VOLUME 20 Dimensions. 1160-1560 (im, often 1360-1440 ,u.m long; 960-1420 ^m, often 1140-1280 wide; length/ width ratio: 0-98- L30; 7-10 convolutions. Cells smooth, flat to slightly convex, 160-260 p. m high. Affinity. This new species agrees with the genus Nitellopsis in the general shape of its gyrogonites, more often higher than wide, in the shape of the periapical region where the cells become thinner and narrower, and in the apical nodules which, however, are not very prominent. The thin base and well-developed basal funnel are characteristic of the subgenus Tectochara. Among the known species, the sub- globular shape and narrow base of N. ( T .) latispira recalls N. (T.) supraplana (Peck and Recker) Grambast and Soulie from the Eocene of Peru, but differs by its greater dimensions, especially in length. The shape and the base of this new species is also suggestive of N. ( T .) aemula (Grambast) Grambast and Soulie, from the Lower Hamstead Beds of Isle of Wight, but differs by the less protruding base, which in the latter is emphasized by depressions in front of the basal ends of the cells. The gyrogonite is also more globular in N. ( T .) latispira, the cell width and height of the plug are greater, and there is one less convolution of the spiral cells. Genus sphaerochara Madler, 1952 emend. Horn af Rantzien and Grambast, 1962 Sphaerochara subglobosa (Groves) Horn af Rantzien, 1959 Plate 21, figs. 8, 9 1926 Chara subglobosa Groves, p. 172, pi. 12, fig. 3. 1959 Sphaerochara subglobosa , Horn af Rantzien, p. 129. Syntype. Specimen no. V. 18331. British Museum (Natural History), London. Identification. Grove’s species had not, until now, been rediscovered, or perhaps not recognized, because of the inadequacy of the original illustration. Despite Grove’s very complete diagnosis, only a thorough examination of the type-material allowed us to ascribe to this species some Sphaerochara from the Isle of Wight, as well as some other specimens previously collected in the Lower Oligocene of the south of France. Description. Gyrogonite subglobular to broadly ellipsoid; apex convex, base trun- cated. Spiral cells with granulose surface, ornamented with massive nodules, as high as the cells, generally laterally elongated, disappearing at the periphery of the apex. Apical ends of the cells bulging, forming a rosette where the thickness of the EXPLANATION OF PLATE 21 Fig. 1 . Harrisichara vasiformis (Reid and Groves) Grambast, lateral view, x 60; from Lower Headon Beds, Headon Hill. Figs. 2-3. Harrisichara v asij ormis- tuber culata, lateral view, x60; from Upper Headon Beds; 2, from Headon Hill; 3, from Cliff End. Figs. 4-5. Harrisichara tuberculata (Lyell) Grambast, lateral view, x 50; from Bembridge Beds; 4, from Horestone Point; 5, neotype, from Whitecliff Bay. Fig. 6. Psilochara aff. conspicua Grambast, lateral view, x 50; from Upper Headon Beds, Cliff End. Fig. 7. Grovesicliara distorta ( Reid and Groves) Horn af Rantzien, lateral view, x 45 ; from Lower Headon Beds, Headon Hill. Figs. 8-9. Sphaerochara subglobosa (Groves) Horn af Rantzien, lateral views, x 90; 8, from Upper Headon Beds, Cliff End; 9, from Osborne Beds, Whitecliff Bay. PLATE 21 FEIST-CASTEL, Charophytes from the Isle of Wight 150 PALAEONTOLOGY, VOLUME 20 cells increases again, thus forming nodules similar to the lateral ones; apical nodules sometimes undeveloped. Basal plug level with the pore, surrounded by a crown of nodules. Dimensions. 350-500 long, 350-475 jum broad, spiral cells 50-65 wide, showing 6-9 convolutions. Relationships. This species clearly corresponds with the genus Sphaerochara in its general globose shape, its small size, the prominent apical rosette, and the basal plug level with the pore. It differs from other known species of the genus in its ornamentation. Stratigraphic significance. In the Isle of Wight, S. subglobosa occurs in several locali- ties in the Upper Headon, Osborne, and Bembridge Beds. According to Groves, it persists into the Lower Hamstead Beds; it does not seem to occur below the Upper Headon Beds. S. subglobosa is thus the only species which makes its appearance between the Lower Headon and Bembridge Beds, other taxa present being persistent Eocene or intermediate Eocene-Oligocene forms. Genus harrisichara Grambast, 1957 Harrisichara vasiformis (Reid and Groves) Grambast, 1957 Plate 21, fig. 1 1921 Chara vasiformis Reid and Groves, p. 185, pi. 4, figs. 12-15. 1927 Kosmogyra vasiformis, Pia, p. 90. 1957 Harrisichara vasiformis , Grambast, p. 347, pi. 6, fig. 4. Lectotype. Specimen no. 76528, Geological Survey, London. Designated and refigured by Grambast (1957). The Lower Headon Beds (Upper Bartonian) are characterized by typical specimens of H. vasiformis recognizable by their narrow base and rounded apex. The same typical aspect is shown by populations of H. vasiformis in beds of the same age in the south of France, in the Ales basin for example (Feist-Castel 1971). However, at Headon Hill and at the type-locality at Hordle (Hampshire mainland) specimens are found of similar size but which have a more rounded base and more flattened apex, a shape which is suggestive of H. tuberculata. Harrisichara tuberculata (Lyell) Grambast, 1957 Plate 21, figs. 4, 5 1826 Chara tuberculata Lyell, p. 94, pi. 13, figs. 7, 8. 1828 Chara tuberculosa. Ad. Brongniart, p. 72. 1850 Chara tuberculosa , Unger, p. 33. EXPLANATION OF PLATE 22 Fig. 1. Chara antennata Grambast, lateral view, x 110; from Upper Headon Beds, Headon Hill. Fig. 2. Grambastichara tornata (Reid and Groves) Horn af Rantzien, lateral view, x70; from Middle Headon Beds, Headon Hill. Fig. 3. Nitellopsis ( Tectochara ) latispira n. sp., x 35; 3 a, paratype, lateral view; 3b, holotype, lateral view; 3c, paratype, apical view ; 3d, paratype, basal view. Fig. 4. Nitellopsis (Tectochara) aff. aemula Grambast, x35; 4a. lateral view; 4b, apical view; 4c, basal view; from Osborne Beds, Cliff End. PLATE 22 FEIST-CASTEL, Charophytes from the Isle of Wight 152 PALAEONTOLOGY, VOLUME 20 1920 Chara archiaci Watelet var. tuberculata , Dollfus and Fritel, p. 252. 1927 Kosmogyra tuberculata, Pia, p. 90. 1957 Harrisichara tuberculata, Grambast, pi. 6, figs. 1-3, 8-10. Neotype (the type-material has not been located). Specimen no. C.F. 1 585b- 1 , Universite des Sciences et Techniques, Montpellier, France. It is figured here (PI. 21, fig. 5). Locality. Whitecliff Bay, Isle of Wight; marls in lower part of Bembridge Limestone. H. tuberculata is abundant from the base of the Bembridge Beds upwards (H ore- stone Point, Whitecliff Bay) and extends into the Lower Hamstead Beds. It is a sure indicator of the Lower (but not lowermost) Oligocene and of the base of the Middle Oligocene. Morphologically it is characterized by its broad ellipsoidal shape, its well-defined but short basal stalk, and by its dimensions. The variability of the ornamentation is another feature of the species; the tubercles are either connected by a thin median line, or merged into a nodular ridge which may or may not extend to the apex. More rarely, isolated tubercles, separated from one another by one or two smaller tubercles, may be observed. Harrisichara vasiformis-tuberculata Plate 21, figs. 2, 3 Harrisichara vasiformis and H. tuberculata are two distinct and easily recognizable species. However, in Hampshire, from the Lower Headon Beds upwards and mainly in the series occurring between these and the Bembridge Beds, specimens of Harrisi- chara are found which are morphologically intermediate between these two species. Table 2 summarizes the features peculiar to each population. The forms intermediate between H. vasiformis and H. tuberculata are not restricted to Hampshire; they are also found in the Lowermost Oligocene in southern France (Triat and True 1972; Feist-Castel 1975). Due to their wide geographical distribu- tion and short vertical extension, they are of real stratigraphic value. Genus psilochara Grambast, 1959 Psi/ochara bitruncata (Reid and Groves) Feist-Castel, 1971 1921 Chara strobilocarpa var. bitruncata Reid and Groves, p. 188, pi. 5, fig. 13. 1959 Charites bitruncata, Horn af Rantzien, p. 67, pi. 3, figs. 1-4. 1971 Psilochara bitruncata, Feist-Castel, p. 166. Lectotvpe. Specimen no. V 24122, British Museum (Natural History), London, designated here from the type material figured by Reid and Groves (1921). Description. Gyrogonite elongated, apex truncated, base narrowed. Spiral cells concave to flat, smooth. Apical ends of the cell's pointing upwards; junction line very short. Dimensions. 800-900 long, 660-760 broad. Spiral cells 110-120 p.m wide, showing 7-8 convolutions. Distribution. P. bitruncata, whose type-locality is at Hordle, occurs infrequently in the Lower Headon Beds at Headon Hill. Besides Hampshire, it has been reported from the Upper Bartonian of southern France. FEIST-CASTEL: CHAROPHYTES FROM THE ISLE OF WIGHT 153 table 2. Comparison of Harrisichara from the Lower Headon to Hamstead Beds, Isle of Wight. j H. vasiformis \ H. vasif-tubercul. j 1 H. tuberculata General shape 1 long, ovoid tapering base | | 1 ovoid to ellipsoid 1 1 | wide ellipsoid Dimensions lengths widths cell widths 1 1 1 1 1 720-950 pm 500-760 pm 60-100 pm 1 1 680-920 pm 600-760 pm 50-90 pm 820-1160 pm 760-1000 pm 100-125 pm Ornamentation often isolated tu- 1 bercles, separated by smaller ones j 1 1 1 1 tubercles often connected by a median line 1 1 tubercles connected by a median line or merged, forming a nodular ridge Stratigraphic distribution 1 1 Lower Headon Beds 1 1 J L Lower & Upper, Headon Beds, Osborne Beds 1 _L Bembridge & Hamstead Beds Closely related forms. In the Upper Headon Beds at Headon Hill and Cliff End, as well as in the Osborne Beds at Cliff End, Psilochara is very abundant and appears to be similar to P. bitruncata , without being definitely assignable to it. In these popula- tions, three different forms are distinguishable: 1. Specimens similar to P. bitruncata in having an elongated shape, usually with concave cells but with their dimensions slightly greater. 2. Specimens of the same shape and dimensions as the former, but with convex cells (PI. 21, fig. 6) recalling P. conspicua Grambast. 3. Specimens of ovoid shape, with flat or convex (rarely concave) cells, also characterized by a clear apical zone of large diameter (480-580 /Tin) and by wavy sutures in the lower region of the gyrogonites. Apart from its longer shape, this form is very similar to P. repanda Grambast. Psilochara polita (Reid and Groves) Grambast, 1959 1921 Chara polita Reid and Groves, p. 187, pi. 5, figs. 9, 12. 1927 Gyrogonites politus, Pia, p. 90. 1959 Peckichara polita , Horn af Rantzien, p. 116, pi. 13, figs. 1-3. 1959 Psilochara polita , Grambast, p. 1 1 . Lectotype. Reid and Groves 1921, pi. 5, fig. 12; designated by Grambast 1958, p. 179. Slide 76521, Geo- logical Survey, London. P. polita , which is associated with P. bitruncata in the Lower Headon Beds at 154 PALAEONTOLOGY, VOLUME 20 Headon Hill, differs from the latter by its ovoid shape, its dimensions (680-880 ^m x 600-720 jum), the number of convolutions (8-10), and the undulations of the sutures. Genus stephanochara Grambast, 1959 Stephanochara edwardsi Grambast, 1958, p. 168 Holotype. Specimen no. 9-6-19, British Museum (Natural History), London. This species is abundant in the Lower Headon Beds at Headon Hill, but was not found in younger beds. From a morphological point of view, S. edwardsi is very close to S. grambasti Feist-Castel, a species of nearly the same age from southern France. The latter is, however, easily distinguishable by its narrower basal region and the less well-marked, or even non-existent, periapical constrictions. Genus rhabdochara Madler, 1955 emend. Grambast, 1962 Rhabdochara stockmansi Grambast, 1957 Holotype. Specimen no. C. 140-1, Grambast Collection, Universite des Sciences et Techniques, Mont- pellier, France. R. stockmansi occurs infrequently in the Bembridge Beds at Horestone Point and at Whitecliff Bay. This species, which extends into the Lower Hamstead Beds, is common in other regions of Europe particularly at the base of the Middle Oligocene (e.g. Hoogbutsel, Belgium; Ronzon, France). Genus grambastichara Horn af Rantzien, 1959 Grambastichara tornata (Reid and Groves) Horn af Rantzien, 1959 Plate 22, fig. 2 1921 Chara tornata Reid and Groves, p. 187, pi. 5, figs. 1-3. 1927 Gyrogonites tornatus , Pia, p. 90. non Tectochara tornata , Madler, 1955, p. 296, taf. 26, figs. 19-22. 1959 Grambastichara tornata , Horn af Rantzien, p. 70, pi. 4, figs. 1-6. Holotype. None designated. Lectotype. Reid and Groves 1921, pi. 5, fig. 3 (designation by Madler 1955). Specimen no. 16519, Geo- logical Survey, London. Although the type-locality is the equivalent of that at Hordle, C. tornata was not found in the Lower Headon Beds of the Isle of Wight (from which I collected only very little material). This species is abundant in the Middle Headon Neritina Beds, at Headon Hill. The specimens are very similar to those described by Reid and Groves; only the dimensions are slightly different, the upper and lower limits of variation of length and width being 100 p.m greater. In both localities, the apical rosette which characterizes the genus of Horn af Rantzien is not in every case obvious in G. tornata , its type-species. On the contrary, nearly half of the specimens present a pattern quite similar to that of Chara. It differs from Chara only in the cellular relief, concave in the latter, convex or flat in G. tornata. It is thus obvious that the separation of Grambastichara from Chara remains problematical, a question already raised by Grambast (1962). FEIST-CASTEL: CHAROPHYTES FROM THE ISLE OF WIGHT 155 Char ci antennata Grambast, 1958 Plate 22, fig. 1 1958 Chara antennata, Grambast, p. 188. Holotype. Specimen no. C. 244-6, Grambast Collection, Universite des Sciences et Techniques, Mont- pellier, France. Description. General shape cylindrical, apex convex, base tapering. Spiral cells flat to slightly concave, ornamented with very prominent cylindrical tubercles. Apical ends of the cells concave. Basal pore pentagonal, without a peripheric funnel. Dimensions. 450-620 ^m long, 350-400 p.m wide, 8- 1 1 convolutions. Distribution. In the Isle of Wight, C. antennata occurs in the Upper Headon Beds and, very rarely, in the Middle Headon Beds. Besides these localities, it has been reported from ‘Marnes a Pholadomyes’ (Upper Bartonian) of the Paris basin by Grambast (1958) and Riveline (1973), and also from a locality of the Ales basin by the author (1971). STRATIGRAPHIC IMPLICATIONS OF THE CHAROPHYTE DISTRIBUTION In Table 3, the above data are added to those from Reid and Groves (1921), Groves (1926), and Grambast (1958, summarized by Curry 1966), concerning the floras of the Lower Headon (Hordle), Bembridge, and Hamstead Beds. This study revealed the floras, which were previously unknown, in the levels between the Lower Headon and the Bembridge Beds. From the established succes- sion, one can see that the change in the floras from the Verzenay (Hordle) Zone to the Bembridge Zone is progressive, whereas the transition previously appeared to be more abrupt. Concerning the position of the presently controversial Eocene-Oligocene boun- dary, charophytes provide no decisive information, since the stratigraphic divisions are essentially based upon marine faunas. We have, therefore, indicated only the divisions which seem most consistent with their distribution. In the Palaeogene of the Isle of Wight and of Hampshire in general, the most obvious floristic changes occur at two levels. First, in the Upper Headon Beds where, after the disappearance of numerous species from the Lower Headon Beds, there appears Sphaerochara subglobosa which extends into the Lower Hamstead Beds. Although the floras of the Lower and Upper Headon Beds remain quite distinct (those of the Middle Headon Beds being of no stratigraphic value), it cannot be called a real renewal, since several Bartonian species extend into the Upper Headon Beds, and since transitional forms are observed there as well. The second change occurs at the level of the Bembridge Beds, where there appears typical H. tuberculata as well as the genus Rhabdochara , represented by R. stockmansi. Here the break appears more distinct than previously; however, in other regions (notably the south of France), the transitional flora, which in the Isle of Wight figures only in the Upper Headon and Osborne Beds, may extend into levels equivalent to the Bembridge Beds (work on this subject is now in progress). It is thus noted that the distribution of the charophytes would not usually lend itself to a break at this level. Likewise, the floristic composition does not vary noticeably between the Bembridge and Lower 156 PALAEONTOLOGY, VOLUME 20 table 3. Vertical extension of Charophyte floras in the Upper Eocene and Lower Oligocene of Hampshire. EOCENE OLIGOCENE Age Verzenay Bembridge Charophyte zones HEADON ' HEADON LOWER HEADON MIDDLE C C g 3 c W 2 50 S \ | 5 * 5 C 1 C fr HAMSTEAD o Beds -n 73 Species — c/ ^ r i it c i c/c / i it t u m c uct i// i c ^ r j r j yj t Ui 1 1 uud i tu f in i u L Ur ? lu lu LstMlru unltrlrlCllu -fc, Hamsichara vasifotviis n urrisiL riciru vusij omi is-iu oercu lu lu — , ■ Nitellopsis (Tec tochacuj aetnula D L L i/i ru f It/v f IU f La r C C I Lf (Jt J gfpph(jnnrh(]f(t pitiful* Hamstead Beds; although they may be distinguished by the presence of characteristic species, these two levels do in fact have a number of species in common : H. tuber- cu/ata, R. stockmcmsi , S. subglobosa , and G. wrighti. The position of the Eocene-Oligocene boundary which best agrees with the charo- phyte distribution is thus the one presently accepted by the Stratigraphical Lexicon (Denizot 1957; Curry 1958). Acknowledgements. The material found by Professor D. Curry (University College, London) was sent by him to Professor L. Grambast (Universite de Montpellier-2) who loaned it for the present study. I wish to express my thanks to both. I am also indebted to Dr. G. F. Elliott and Mr. C. H. Shute (Palaeontology Department, British Museum (Natural History)), for permitting me to examine specimens from the Groves Collection. FEIST-CASTEL: CHAROPHYTES FROM THE ISLE OF WIGHT 157 REFERENCES bristow, H. w. 1862. The geology of the Isle of Wight. Mem. geol. Surv., U.K. , 1 vol., 138 pp., London. castel, M. 1968. Zones de Charophytes pour l’Oligocene d’Europe occidentale. Compt. Rend. somm. Soc. geol. Fr. 4,121 122. curry, D. 1958. Lexique stratigraphique international, vol. I, part 3a XII, Palaeogene. Edit. Cent. nat. Rech. scientif. 82 pp., Paris. — 1966. Problems of correlation in the Anglo-Paris-Belgian Basin. Proc. Geol. Ays. 77, 437-467. — 1968. Excursion en Angleterre, Colloque sur l’Eocene. Mem. Bur. Rech. geol. min. 58, 1-24. denizot, G. 1957. Lexique stratigraphique international, vol. I, fasc. 4a VII, Tertiaire. Edit. Cen. nat. Rech. scientif. 217 pp., Paris. feist-castel, m. 1971. Sur les charophytes fossiles du bassin Tertiaire d’Ales. Geobios (4), 3, 157-172, pis. 10-12. Lyon. — 1975. L’evolution floristique des Charophytes a la limite de l’Eocene et de l’Oligocene. Reun. ann. Sci. Terre , p. 149. Soc. geol. Fr., Paris. forbes, E. 1856. On the tertiary fluvio-marine Formation of the Isle of Wight. Mem. geol. Surv. U.K. 1 vol., 162 pp., London. grambast, l. 1958. Etude sur les charophytes Tertiaires d’Europe occidentale et leurs rapports avec les formes actuelles. These, 258 pp., Paris. 1962. Classification de l’embranchement des charophytes. Natur. Monspeliensia Bot. 14, 63-86. — 1972. Principes de l’utilisation stratigraphique des charophytes. Application au Paleogene d’Europe occidentale. Mem. Bur. Rech. geol. min. 77, 319-328, Paris. groves, j. 1926. Charophyta. In reid, e. m. and chandler, m. e. j. The Bembridge Flora. Brit. Mus. Cat. Cainoz. Plants, I, 165-173, London. reid, c. and groves, j. 1921. On the Charophyta of the Lower Headon Beds of Elordle Cliff (Hampshire). Q. J l geol. Soc. Loud. 77, 175-192. riveline, j. 1973. Repartition des Characees dans des calcaires lacustres de l’Eocene terminal et de l’Oligo- cene basal du Sud-Est du Bassin de Paris. Compt. Rend. hebd. Seanc. Acad. Sci. Paris (D), 277, 2641 -2643. stehlin, h. G. 1909. Remarques sur les faunules de Mammiferes des couches Eocenes et Oligocenes du Bassin de Paris. Bull. Soc. geol. Fr. 4eme ser., 9, 488-520. triat, j. m. and truc, g. 1972. L’Oligocene du bassin de Mormoiron (Vaucluse). Etude paleontologique et sedimentologique. Docum. Lab. geol. Univ. Lyon Fac. Sci. 49, 27-52. white, h. j. o. 1921. A short account of the Geology of the Isle of Wight. Mem. geol. Surv. U.K. 219 pp., London. M. FEIST-CASTEL Laboratoire de Paleobotanique et Evolution des Vegetaux E.R.A. 1 14, Universite des Sciences et Techniques du Languedoc Place E. Bataillon Typescript received 5 August 1975 34060 Montpellier Revised typescript received 26 January 1976 France CLASSIFICATION AND PHYLOGENY OF HOMALONOTID TRILOBITES by A. T. THOMAS Abstract. The composition, classification, and phylogeny of the Homalonotidae are discussed, and new family and subfamily diagnoses given. Plaesiacomia hughesi sp. nov. and P. vacuvertis sp. nov. from the Llanvirn of Wales and Saudi Arabia respectively, and Platycoryphe dyaulax sp. nov. from the Llandovery of Saudi Arabia, are described. Some of the problems involved in classifying homalonotid trilobites have arisen because the Silurian and Devonian genera are known principally from old, inadequate descriptions. The difficulties are increased by the rarity of Ashgill and Llandovery forms which has emphasized the apparent morphological gap between post- Ordovician and older genera. Study of the species described here, together with work on middle and upper Silurian homalonotines to be published elsewhere, has shed new light on the classification and phylogeny of the family. Repositories. The following abbreviations are used herein: BM — British Museum (Natural History), London; GSM —The Geological Museum, Institute of Geological Sciences, London; SM— Sedgwick Museum, Cambridge. SYSTEMATIC PALAEONTOLOGY Family homalonotidae Chapman, 1890 [= Portaginidae Lesperance, 1968] Diagnosis. The following diagnosis is based on that of Sdzuy (in Moore 1959, pp. 0454-0455) and incorporates the correction and emendation noted by Whit- tington (1965, p. 488). Cephalic border rarely well defined, often absent. Glabella narrowing forwards, with four or fewer pairs of lateral glabellar furrows. Paraglabellar area usually present. Genal angle generally rounded with facial suture displaying gonatoparian condition; opisthoparian or proparian sutures and genal spine rarely occur. Palpe- bral lobe small, eye ridge present in some early forms. Rostral plate narrows pos- teriorly. Rostral suture ventral or marginal in early genera, dorsal in later forms. Hypostome subquadrate, with small anterior wing and weakly developed lateral notch; posterior margin rarely rounded, usually bifurcate. Thorax of thirteen seg- ments; axis moderately wide in early genera, indistinct and very broad in later ones. Pleural tips blunt. Pygidium subsemicircular to triangular in outline. Discussion. Brongniartella Reed, 1918; Bunneistere/la Reed, 1918; Burmeisteria Salter, 1865; Calymenella Bergeron, 1890; Colpocoryphe Novak in Perner 1918; [Palaeontology, Vol. 20, Part 1, 1977, pp. 159-178, pis. 23-24.] 160 PALAEONTOLOGY, VOLUME 20 Digonus Giirich, 1909; Dipleura Green, 1832; Eohomalonotus Reed, 1918; Para- homalonotus Reed, 1918; Plaesiacomia Hawle and Corda, 1847; and Trimerus Green, 1832 are sufficiently well known for no doubt to arise concerning their membership of the family. Henningsmoen (in Moore 1959, p. 0524) regarded Platycoryphe Foerste, 1919 as of uncertain affinities but I agree with Whittington (1965) that it is a homalonotid. Pribyl (1953, p. 16) placed Bavarilla Barrande, 1868 in the Olenidae but it was excluded therefrom by Henningsmoen (1957, p. 20). Bavarilla is accepted as a homalonotid and its systematic position is discussed further below. No good illustrations are available of Leiostegina Kobayashi, 1937 but the anteriorly nar- rowing glabella and transversely straight, apparently dorsal, rostral suture (see Kobayashi 1937, pi. 6, fig. 18) suggest that it is a homalonotid. Professor S. Kuss (Palaontologisches Institut, Freiburg) informed me (pers. comm., 8 October 1975) that the type material of L. inexpectans , the only known species, seems to be lost. A more detailed appraisal of Leiostegina must therefore await the availability of topotype material. The available illustrations of Pamir otellus Balashova, 1968 ( Pamirites Bala- shova, 1966; non Toumansky, 1938) are poor but what can be seen suggests that it is a synonym of Brongniartella. In particular, the pygidium is longer than wide, well segmented, and obtusely pointed posteriorly (Balashova 1966, pi. 2, fig. 6; 1968, pi. 52, fig. 5c). The glabella (Balashova 1966, pi. 2, fig. 7; 1968, pi. 52, fig. 5 d) is constricted (tr.) anteriorly, a feature typical of many Brongniartella species. Bala- shova gives the horizon of Pamirotellus as Wenlock whereas Brongniartella is not known elsewhere after the lower Llandovery. Possibly a relic species is represented or the horizon given may be incorrect. Since no other faunal elements are listed from this horizon the reasons for suggesting a Wenlock age are unclear. Portaginus Lesperance, 1968, from the upper Ashgill of Quebec, is also synonymous with Brongniartella (see Dean 1976, pp. 238-239). Liangshanaspis Chang, 1974 shows no major differences from Platycoryphe. Apollonov (1974, pp. 59-61, pi. 11, figs. 1-11; pi. 12, figs. 1-6, 8) recently redescribed the type species, and referred it to that genus. Neseuretus Hicks, 1873 (see Whittard 1960, p. 139 for use of this name rather than Synhomalonotus Novak in Perner 1918) has been regarded as a homalonotid by various authors but I agree with Whittington (1966a, pp. 499-500) that it is a caly- menid. Lakaspis Kobayashi, 1937 was provisionally assigned to the Homalonotidae by both Kobayashi and Hupe (1953). The available illustrations (Kobayashi 1937, pi. 2, figs. 27, 28; Lake 1906, pi. 11, figs. 2, 3) are poor but it is clear that the glabella expands (tr.) in its anterior half. This indicates that Lakaspis is not a homalonotid and Jaanusson (in Moore 1959, p. 0358) questionably assigned it to the Nileidae. Clarke (1913) erected Homalonotus ( Schizopyge ) parana for a pygidium from the lower Devonian of Brazil. The specimen displays no homalonotid features and Struve (1958) erected the calmoniid genus Tibagya to receive it (see also Struve in Moore 1959, p. 0486). Classification of the Homalonotidae. Hupe (1953, 1955) placed the Homalonotidae in the superfamily Calymenoidea (which he attributed to Swinnerton 1915). He THOMAS: HOMALONOTID TRILOBITES 161 proposed three subfamilies based on cephalic and pygidial shape, form of the rostral suture, and distinctness of trilobation: eohomalonotinae Eohomalonotus, Calymenella , Plaesiacomia homalonotinae Homalonotus, Burmeisterella , Brongniartella , Parahomalonotus, Digonus, ILakaspis, ILeiostegina trimerinae Trimerus, Burmeisteria , Dipleura Hupe (1953, p. 232) included Neseuretus in the Calymenidae but in 1955 erected the calymenid subfamily Colpocoryphinae to include Neseuretus , Colpocoryphe , and the non-papillate calymenids. In Sdzuy’s (1957; in Moore 1959, pp. 0454-0461) revision, Bavarilla was accepted as a homalonotid and placed in its own subfamily. Sdzuy’s other subfamilies are as follows: eohomalonotinae Calymenella {Calymenella), C. ( Eohomalonotus ), Brongniartella , Neseuretus (this was simultaneously classified as a calymenid, p. 0453; and listed as incertae sedis , p. 0523) colpocoryphinae Colpocoryplie, Plaesiacomia , ILeiostegina homalonotinae Homalonotus , Burmeisterella, Burmeisteria {Burmeisteria), B. {Digonus), Parahomalonotus, Trimerus {Trimerus), T. {Dipleura) Hupe’s Trimerinae was not recognized. Pribyl (1957, p. 92) had pointed to the close similarities between trimerines and homalonotines— especially the tendency to ill- defined trilobation, broad thoracic axis, and similar length to breadth proportions of the cephalon and pygidium. Kobayashi (1960) erected the Synhomalonotinae to include Neseuretus and the calymenid Vietnamia. Having accepted Whittard’s (1960) arguments for regarding Synhomalonotus as a junior synonym of Neseuretus, Kobayashi (1969, p. 243) pro- posed ‘Neseuretinae’ as a replacement subfamily name. This procedure contravenes Article 40 of the ICZN and ‘Synhomalonotinae’ must be retained. As Neseuretus is regarded as a calymenid, this subfamily is referred to the Calymenidae (see also Whittington 1971). Subsequent workers have viewed the suprageneric classification of calymenaceans in different ways. Some authors (Dean 1 966c/ ; Hughes 1969; Henry 1971; Sadler 1974) favour several different families, others (Whittington 1966u, 1971; Dean 1971) utilize a number of different subfamilies assigned to either the Calymenidae or Homalonotidae. A number of genera are present in the Tremadoc and Arenig from which younger calymenids and homalonotids arose (see Whittington 1966u, p. 500). While accept- ing that such early genera may not be easily accommodated in either group, I con- sider that the familial classification should reflect this essentially bipartite pattern. Within the Homalonotidae, three subfamilies are recognized for post-Tremadoc forms. The Cambrian ancestors of homalonotids are unknown but Bavarilla seems to be related. Hence the Bavarillinae Sdzuy, 1957 (as diagnosed by Sdzuy in Moore 1959, p. 0455) is retained. L 162 PALAEONTOLOGY, VOLUME 20 Subfamily eohomalonotinae Hupe, 1953 Diagnosis. Cephalon subtriangular, rounded anteriorly; with anteriorly widened border. Glabella trapezoidal to parabolic; three or fewer pairs of lateral glabellar furrows, lp sigmoidal. Paraglabellar area defined. Rostral suture is marginal or nearly so and describes a wide, forwardly convex curve. Facial suture gonatoparian. Thorax with distinct, moderately wide axis. Pygidium triangular, wider than long, with distinct trilobation. Genera. Eohomalonotus , Calymenella. Remarks. Reed (1918, p. 321) erected Eohomalonotus {pro Brongniartia Salter, 1865 ( pars ); non Leach, 1824, nec Eaton, 1832) as one of his subgenera of Homalonotus. Reed included a number of forms now referable to Platycoryphe as well as species of Eohomalonotus as presently understood, and distinguished it from Calymenella by the lack of a projecting frontal area. Sdzuy (in Moore 1959, pp. 0455-0456) regarded Eohomalonotus as a subgenus of Calymenella , the two being distinguished by the position of the rostral suture and degree of development of pygidial inter- pleural furrows, as well as the structure of the frontal area. Little recent information is available and whether two distinct genera are represented is a matter which must await further research. Many authors have placed Platycoryphe and Brongniartella in the Eohomalo- notinae. I regard them as homalonotines and reasons for doing so are given in the discussion of that subfamily. Subfamily colpocoryphinae Hupe, 1955 Diagnosis. Cephalon subcircular; convex (tr. and sag.), sometimes highly so. Frontal area short (sag.), no anterior or lateral border. Rostral suture marginal. Glabella clearly defined. Palpebral lobe anteriorly placed. Cheek strongly declined laterally; genal angle well rounded, facial suture gonatoparian. Thoracic axis well defined. Pygidial axis wide and well defined. Border furrow deep; border flexed strongly downwards. Genera. Colpocoryphe, Plaesiacomia. Remarks. Vanek (1965), like Hupe (1955) and Dean (1971), excluded the Colpo- coryphinae from the Homalonotidae and considered them to be calymenids close to Neseuretus. The cephalic and glabellar shape, presence of paraglabellar areas (in many species), lack of anterior and lateral cephalic border, marginal rostral suture, moderately wide thoracic axis, thoracic structure, and the poorly segmented pygidium (which lacks interpleural furrows) suggest that colpocoryphines are homalonotids. Sdzuy (in Moore 1959) questionably assigned Leiostegina to this subfamily. The genus is poorly known but shows no obvious features suggestive of a colpocoryphine relationship and 1 regard it as of uncertain subfamilial position. THOMAS: HOMALONOTID TRILOBITES 163 Genus plaesiacomia Hawle and Corda, 1847 Type species. P. rara Hawle and Corda, 1847, p. 55, pi. 3, fig. 30; from the Caradoc of Bohemia; by monotypy. Other species. P. hughesi sp. nov.; P. oehlerti (Kerforne, 1900); P. vacuvertis sp. nov.; IP. brevicaudata (Deslongchamps, 1825). Diagnosis. Colpocoryphine with weakly impressed lateral glabellar furrows, glabella tapering rapidly forwards. Paraglabellar areas usually present; eye ridges absent. Pygidial axis with up to four (?five) weakly defined rings. Discussion. Vanek’s (1965) view that Colpocoryphe and Plaesiacomia were synony- mous was questioned by Dean (19666, pp. 140-141) who listed nine distinguishing characters. Five of these— cephalic convexity, depth of anterior arch, and height of the pygidial axis (all greater in Colpocoryphe ), position of pygidial border furrows, and degree of divergence of the pygidial margins— are interdependent characters connected with enrolment (see Clarkson and Henry 1973). Two more, the develop- ment of glabellar furrows and eye ridges, are too variable to be reliable discrimina- tory characters. Glabellar furrows are usually strongly developed in Colpocoryphe , but in a few cases are as weak as in Plaesiacomia (see Whittington 1953). Paraglabellar areas occur in all Plaesiacomia species known to Dean, but are absent in P. hughesi. Two of the features listed by Dean can be reliably used to distinguish between the genera: the glabella in Colpocoryphe species is bell-shaped, slightly tapering, with an inflated frontal lobe, while that of Plaesiacomia tapers more rapidly to a low frontal lobe and the pygidial axis lacks the clear segmentation of Colpocoryphe. Colpocoryphe species are poorly known and a revision is necessary before a decision concerning its status can be reached. Plaesiacomia hughesi sp. nov. Plate 23, figs. 1-5, 7 v. 1969 Plaesiacomia sp.; Hughes, pp. 95-96, pi. 13, figs. 4-8; pi. 14, figs. 1-2. . 1972 Plaesiacomia sp. ; Robardet et al., p. 122. Name. After Dr. C. P. Hughes, who first described this species and made additional material available for study. Holotype. BM It3024, incomplete internal mould of cranidium, from Ordovician, lower Llanvirn, exposure on the left bank in the ravine in the upper reaches of Camnant Brook immediately north of the S bend 230 yd (207 m) S. 13° W. of the fence crossing the stream source, Builth district, Powys (Radnorshire) (SO 088 575). Paratvpes. BM It 1 38 1 5ab, incomplete cephalon; It 1 38 1 6 1 38 1 9, Itl3821ab, It 1 3823ab, cranidia; It3025, external mould of cranidium; Itl3822ab, Itl3824ab, Itl3825ab, IU3826-13827, It3023, lt3026ab-3027, pygidia; It 13820, external mould of pygidium. Diagnosis. Plaesiacomia species with palpebral lobe placed two-thirds the way along the cephalon from the posterior margin. Glabellar furrows absent or very weak; paraglabellar areas absent. Pygidial axis with four (?five) ring furrows on the internal mould. Description. The following points supplement the description given by Hughes (1969, pp. 95-96). Three pairs of very faint glabellar furrows are developed in the holotype 164 PALAEONTOLOGY, VOLUME 20 (PI- 23, fig. 3 a-b), but these are not clearly seen in other specimens. The pygidial border is broad, almost vertically inclined, and narrows somewhat beneath the posterior tip of the axis. Discussion. Hughes (1969, p. 96) referred this species to Plaesiacomia because of the glabellar outline, poor development of lateral glabellar furrows, lack of prominent eye ridges, and shallow anterior arch. I accept this assignation but in the lack of paraglabellar areas, P. hughesi is reminiscent of Colpocoryphe species. The pygidial axis has more ring furrows than any other Plaesiacomia species but, unlike the situation in Colpocoryphe , these are shallow and faint. P. hughesi differs from other Plaesiacomia species in lacking paraglabellar areas and having the glabellar furrows very faint or absent, as well as by the relatively large number of pygidial axial rings. Plaesiacomia vacuvertis sp. nov. Plate 23, figs. 6, 8-11; Plate 24, figs. 1 -6 Name. From Latin vacuus, empty and vertis, top of the head ; alluding to the absence of lateral glabellar furrows in the later growth stages. Holotype. SM A87322, internal mould of complete specimen, from Ordovician, Llanvirn, type locality of the Hanadir Shale Member of the Tabuk Formation, Jebel Hanadir, Saudi Arabia (25° 28' 00" N. Lat., 43° 27' 30" E. Long.). Paratypes. SM A87353, external mould of complete specimen; A87243ab, A87349, A87355, cephala; A87244, A87321, A87350, partial thoraces and cephala; A87163ab-A87178ab, A87237ab, A87365, A87294, A87306, A87352, cranidia; A87351, internal mould of incomplete thorax; A87162ab, thoracic segment; A87354, internal mould of pygidium and part of thorax; A87179ab, A87295, pygidia. All from locality and horizon of holotype. Diagnosis. Plaesiacomia species with palpebral lobe placed about half-way along the cephalon. Paraglabellar areas well developed; lateral glabellar furrows weak in smaller individuals, absent in larger ones. Pygidial axis with up to three ring furrows. EXPLANATION OF PLATE 23 Figs. 1-5, 7. Plaesiacomia hughesi sp. nov. Lower Llanvirn, exposure on the left bank in the ravine in the upper reaches of Camnant Brook immediately north of the S bend 230 yd (207 m) S. 13° W. of the fence crossing the stream source, Builth district, Powys (SO 088 575). 1 a~b, BM It 1 381 5a, internal mould of cephalon, oblique anterior and dorsal views, x 6. 2 a-b, BM It 1 3821a, internal mould of cranidium, dorsal and right lateral views, x8. 3 a-b, BM It3024, holotype, internal mould of cranidium, oblique anterolateral and dorsal views, x 8. 4, BM It 1 3820, latex cast from external mould of pygidium, dorsal view, x 6. 5, BM It3026b, latex cast from external mould of pygidium, dorsal view, X 8. la-b , BM It 13824a, internal mould of pygidium, right lateral and dorsal views, x8. Figs. 6, 8-11. Plaesiacomia vacuvertis sp. nov. Llanvirn, type locality of the Hanadir Shale Member of the Tabuk Formation, Jebel Hanadir, Saudi Arabia (25° 28' 00" N. Lat., 43° 27' 30" E. Long.). 6, SM A87350, internal mould of incomplete cephalon and thorax, oblique lateral view, x 6. 8, SM A87353, external mould of degree 9? meraspid, ventral view, x20. 9, SM A87243, internal mould of cephalon, oblique anterolateral view, x 8. 10, SM A87295, internal mould of pygidium, posterior view, x6. 11, SM A87322, holotype, internal mould of complete specimen, dorsal view, x4. PLATE 23 9 1] THOMAS, homalonotid trilobites 166 PALAEONTOLOGY, VOLUME 20 Description. Dorsal exoskeleton more than one and one-half times as long as wide, the maximum width (tr.) situated slightly less than one-third the cephalic length forwards from the posterior cephalic margin. Cephalon semicircular, steeply de- clined anteriorly and laterally. Occipital ring wide (tr.), trapezoidal, clearly demar- cated anteriorly by the occipital furrow. Median occipital tubercle present in some meraspid specimens (PI. 24, figs. 1, 2, 5). Axial furrows deep and well defined, shallowing posteriorly. Initially they converge at 45° then, anterior to the occipital ring, converge more gently. Preglabellar furrow similar in form to axial furrow and is almost straight (tr.). Glabella tapering forwards to about one-third the width of the occipital ring in larger specimens (PI. 23, fig. 11). In small meraspids the glabella is subparallel sided, the glabella widening posteriorly during growth (PI. 24, figs. 5, 2, 1, 4). Up to three pairs of short, faint lateral glabellar furrows are seen in some meraspids (e.g. PI. 24, fig. 3) but these become effaced in later growth stages. Para- glabellar areas oval, not deeply impressed, their length (exsag.) about one-quarter that of the glabella. Paraglabellar areas are not clearly seen in meraspids but are marked in larger specimens by a distinct flexure in the axial furrow (PI. 23, fig. 11; PI. 24, fig. 4). Anterior arch shallow, frontal area narrow (sag.), broadening (exsag.) slightly abaxially. Immediately adjacent to the axial furrow the posterior border is the same width (exsag.) as the occipital ring but it expands somewhat abaxially. Posterior border deep on adaxial two-thirds of fixed cheek, becoming fainter abaxially and terminating near the lateral margin. Posterior branch of facial suture strongly curved, normal to axial furrow at first, then backwardly flexed. Anterior branch of facial suture straight. Palpebral lobe placed near anterior end of glabella in smallest specimens (e.g. PI. 24, fig. 5), migrating backwards through ontogeny to assume its adult position. Rostral suture straight (tr.). Hypostome and rostral plate unknown. Thorax with well-defined trilobation. Axis comprising two-thirds the anterior width (in large specimens), tapering to about half posteriorly. Axis rather narrower EXPLANATION OF PLATE 24 Figs. 1-6. Plaesiacomia vacuvertis sp. nov. Llanvirn, type locality of the Hanadir Shale Member of the Tabuk Formation, Jebel Hanadir, Saudi Arabia (25° 28' 00" N. Lat., 43° 21' 30" E. Long.). 1, SM A87166a, internal mould of cranidium, dorsal view, x20. 2, SM A87167a and A87168a, internal moulds of cranidia, dorsal views, x20. 3, SM A87163a, internal mould of cranidium, dorsal view, x20. 4, SM A87 172a, internal mould of cranidium, dorsal view, x20. 5, SM A87170a, internal mould of cranidium, dorsal view, x20. 6, SM A87179a, internal mould of pygidium, dorsal view, X 30. Figs. 7-9, 11. Platycoryphe dyaulax sp. nov. 7-9, from convolutus Zone, Idwian, Llandovery, Qusayba Shale, Qusayba district, Saudi Arabia (26° 53' 54" N. Lat., 43° 32' 15" E. Long.). 11, from Qusayba Shale, Qusayba district (exact locality unknown). 7, SM A87542, external mould of cephalon and internal mould of incomplete thorax, ventral and dorsal views, x 2. 8, SM A8 1 820, holotype, cranidium, dorsal view, x2. 9, SM A87541, pygidium, slightly oblique posterodorsal view, x3. 11,SMA89859, internal mould of thorax and pygidium, dorsal view, X 3. Fig. 10. Brongniartella sp. Dufton Shale, lower Longvillian, south side of valley 780 yd (702 m) at 142° from Dufton Pike, British Lake District (NY 7042 2605). GSM PJ4549, cranidium, oblique lateral view, x4. Fig. 12 a-b. Trimerus cylindricus Salter, 1865. Woolhope Limestone (lower Wenlock), Little Hope, Wool- hope district, Hereford and Worcester (Herefordshire). BM 15308, incomplete enrolled thorax, dorsal and right lateral views, x 1 . PLATE 24 THOMAS, homalonotid trilobites 12a 168 PALAEONTOLOGY, VOLUME 20 in meraspids (compare PI. 23, figs. 8 and 11). Pleural furrow directed slightly pos- teriorly and becomes progressively fainter abaxially. The pleura is thus divided into two portions of subequal size. Pleural furrow becomes forwardly directed at the fulcrum, portion of furrow on articulating facet shallower and narrower than that on the more proximal part of the pleura (PI. 23, fig. 6). Holaspid pygidium about one-third as long (sag.) as the cranidium. Axis consti- tuting almost half the pygidial width, defined by posteriorly converging, deep axial furrows. Ring furrows faint, barely extending to axial furrow in holaspids; much more distinct in meraspids (compare PI. 23, fig. 11 with PI. 24, fig. 6). Larger speci- mens have unfurrowed pleural areas but at least four pairs of pleural ribs are present in meraspids (PI. 24, fig. 6). Pleural areas small and triangular in holaspids, with deep border furrow. Latter widest and deepest anteriorly, becoming narrower and shallower as it nears the axis (PI. 23, fig. 10). Border not developed in meraspid individuals. Discussion. P. vacuvertis is most similar to P. oehlerti and P. rara but the palpebral lobe is placed further forwards than in the former species and further backwards than in the latter. The lateral glabellar furrows are weaker than in either species. The only previous descriptions of juvenile homalonotids have been of homalono- tines (Cooper 1935; Morzadec 1969; Saul 1967; Sheng 1964). It is interesting to note that no border is developed in P. vacuvertis meraspides. The deep border furrow in colpocoryphines is associated with enrolment and overlap of the pygidium by the cephalon in holaspids; no such overlap is possible in meraspids. Subfamily homalonotinae [= Portagininae Lesperance, 1968 (nom. transl. Bergstrom, 1973; ex Portaginidae Lesperance, 1968)] Diagnosis. Cephalon subtriangular or semicircular, usually without border. Glabella trapezoidal, often ill defined. Glabellar furrows indistinct, may be absent. Para- glabellar areas usually defined. Posterior branch of facial suture intersecting margin at, or anterior of, rounded genal angle; genal spine very rarely developed. Triloba- tion often indistinct with very wide axis, though axis may be only moderately wide in Ordovician representatives. Pygidium triangular to semicircular, with tendency to reduced distinctness of trilobation. Remarks. Genal spines in members of this subfamily are known only in two speci- mens of Brongniartella (e.g. PI. 24, fig. 10). In contrast to the condition in Bavarilla , the facial sutures are proparian. Some Devonian homalonotines may also have proparian sutures. Genera. Homalonotus Konig, 1825, Brongniartella , Burmeisteria, Burmeistere/la, Digonus , Dipleura , Parahomalonotus, P/atycoryphe, Trimerus. Discussion. P/atycoryphe was placed in the Eohomalonotinae by various authors (e.g. Hughes 1969; Henry 1971; Apollonov 1974) and Brongniartella in the same subfamily by Sdzuy (in Moore 1959). These genera show few close similarities to Calymenella or Eohomalonotus but differ from younger genera only in the more THOMAS: HOMALONOTID TRILOBITES 169 pronounced trilobation and presence of distinct lateral glabellar furrows in some species. Consequently 1 regard them as homalonotines. Genus platycoryphe Foerste, 1919 Type species. Calymene platycephala Foerste, 1910, pp. 81-82, pi. 2, fig. 7; pi. 3, fig. 21 ; from the Trenton (middle Ordovician) of Clifton, Tennessee; by monotypy. Other species. P. barroisi (Lebesconte, 1887); P. bohemicus (Barrande, 1852); P. christyi (Hall, 1860); P. dangeardi Henry, 1971 ; P. dentatus Dean, 1961 ; P. dubius Savage, 1913; P. dyaulax sp. nov. ; P. herberti (Lebesconte, 1887); P. sinensis (Lu, 1963); P. vulcani (Murchison, 1839); IP. convergens Dean, 1966a. Diagnosis. Glabella elongate, trapezoidal in outline, with two to three pairs of glabellar furrows. Frontal area relatively short (sag. and exsag.). Cephalic axial furrows straight or with slightly convex-outwards curvature. Thoracic axis wide or moderately wide. Pygidium wider than long; rounded posteriorly; with eight or fewer ring furrows and six or fewer pleural furrows which become fainter posteriorly. Platycoryphe dyaulax sp. nov. Plate 24, figs. 7-9, 1 1 Name. From Greek dyo, two and aulax , furrow; alluding to the number of lateral glabellar furrows. Holotype. SM A81820, cephalon, from convolutus Zone, Idwian Stage, Llandovery Series, Qusayba Shale Member of the Tabuk Formation, Qusayba district, Saudi Arabia (26° 53' 54" N. Lat., 43° 32' 15" E. Long.). Paratypes. From the type locality: SM A87542, internal mould of incomplete thorax and external mould of cephalon; A87541, A87543ab, A87754ab, pygidia. From Qusayba district, 22 m below top of Qusayba Shale (26° 53' 30" N. Lat., 43° 32' 30" E. Long.): A87117ab, cephalon. From Qusayba district, 8 m below top of Qusayba Shale (26° 50' 30" N. Lat., 43° 35' 15" E. Long.): A87772, incomplete cephalon; A87920, cranidium. From Qusayba Shale, Qusayba district (exact locality unknown): A87859, internal mould of thorax and pygidium. Diagnosis. Platycoryphe species with two pairs of broad, shallow glabellar furrows. Thoracic axis broad (tr.). Pygidium with six axial rings and five pairs of pleural ribs. Description. Cephalon subtriangular, slightly less than twice as wide as long, rounded anteriorly. Breadth (tr.) of occipital ring slightly less than half the maximum cephalic width and slightly greater than the glabellar length. Occipital ring trapezoidal, narrowing forwards, bounded anteriorly by well-defined occipital furrow. Occipital furrow narrow abaxially but broader and shallow along the sagittal line. Axial furrows broad and shallow, strongly convergent at first (at approximately 45° to the posterior margin), then converging more gently. Glabella narrows anteriorly to slightly more than half its posterior width, median raised area present along the sagittal line. Ip broad and shallow, posteriorly directed, not confluent with axial furrow. 2p arises opposite the anterior end of the palpebral lobe, is broad and shallow and orientated parallel to lp. Adaxially posterior border is the same width (exsag.) as the occipital ring, broadening slightly abaxially. Posterior border furrow broad, shallow, terminating near the rounded genal angle. Paraglabellar area very faint, oval, about one-quarter the glabellar length. Frontal area flat and wide (tr.). Anterior branches of facial suture diverge backwards at first then, behind the frontal area, lie subparallel to an exsagittal line. Posterior branch of facial suture strongly curved at 170 PALAEONTOLOGY, VOLUME 20 first, running normal to the sagittal line; then sharply recurved. Free cheek tri- angular in plan view, steeply sloping to lateral border. Lateral border narrows posteriorly, defined by broad, shallow lateral border furrow. Hypostome unknown. Rostral plate only known from narrow (sag. and exsag.) dorsal portion (PI. 24, fig. 8). Thorax with indistinct trilobation. Axis broad, about half the thoracic width. Pleural furrow deep and narrow; posteriorly directed adaxially, becoming parallel to the segment margins abaxially, then directed forwards on the expanded pleural tip. Pygidium slightly wider than long, rounded posteriorly. Axis infundibular, poorly defined. Ring furrows become shallower on margins of the axis. Pleural furrows directed only slightly backwards, terminating close to the pygidial margin. Discussion. Most Platycoryphe species have three pairs of distinct lateral glabellar furrows— but these are variably developed in some species (see Whittington 1965, p. 488; Henry 1971, p. 70), especially P. vulcani (see Whittard 1961, pi. 22, figs. 12, 16; Hughes 1969, p. 97, pi. 14, figs. 3, 6). P. dyaulax is most similar to the latter species but differs in having a much wider thoracic axis and less deeply impressed axial furrows. The pygidium of dyaulax is more rounded in outline and the axial, ring, and pleural furrows are much shallower than those of vulcani (see Whittard 1961, pi. 22, figs. 8-19). P. dyaulax is closely comparable with Brongniartella and Trimerus species, as well as with other species of Platycoryphe. From species of Brongniartella the cephalon differs principally in lacking indented axial furrows. The Brongniartella pygidium is distinguished by being longer than wide with a well-defined axis and nine to twelve axial rings. With other species of Platycoryphe, P. dyaulax shares the straight cephalic axial furrows, posteriorly rounded pygidium (similar to that of Brongniartella but with fewer segments) and a similar occipital ring. The cephalon of P. dyaulax is similar to that of some Trimerus species, the principal differences being in the pygidium— that of Trimerus being characteristically longer than wide and pointed posteriorly. With Trimerus , and Silurian-Devonian homalonotids in general, P. dyaulax shares the wide thoracic axis. The mosaic of characters seen in P. dyaulax is taken to imply a phyletic relationship between Brongniartella! Platycoryphe and later homalonotines. Because of this mosaic of characters, reference of dyaulax to an existing genus is to some extent both arbitrary and subjective. It has been placed in Platycoryphe rather than either Brongniartella or Trimerus because in this way the minimum disruption is caused to the present concept of these genera. It is likely that the previously apparently well-defined morphological discontinuity between Ordovician and Silurian-Devonian homalonotids was due to the rarity of Llandovery representatives. Apart from P. dyaulax I am aware of only seven records of possible Llandovery homalonotids. Brongniartella is known from the lower Llandovery of South Wales (Temple 1975) and from beds of questionable Llandovery age in southern Jordan (Wolfart et al. 1968). Wolfart (1961) described a Llandovery Trimerus species from Paraguay. The specimens (SM A65437-65438) of ‘ Homa - lonotus' recorded by Jones (1925, p. 370) from the Fronian and Telychian of the Llandovery area are poorly preserved but probably represent a Trimerus species. THOMAS: HOMALONOTID TRILOBITES 171 Yin (1966, p. 297) listed ‘ Homalonotus sp." from the persculptus Zone of the Upper Yangtze Province — the genus Liangshanaspis was later proposed for this material (see discussion p. 160). The specimen (SM A38734ab) from the British Lake District, listed as ‘ HomalonotusT by Marr and Nicholson (1888, p. 664), is not homalonotid. The record of ‘ Homalonotus sp.’ by Pocock et al. (1938, p. 269) rests on a fragment of a thoracic segment (GSM D34473) which is not homalonotid. PHYLOGENY OF THE HOMALONOTIDAE The Cambrian ancestors of homalonotids are unknown but a number of Cambrian genera occur in which the over-all morphology, especially the forwardly narrowing glabella and opisthoparian sutures, is similar to that of early calymenids and homa- lonotids. Warburg (1925, p. 69) noted the similarity of the Calymenacea to Ptycho- paria and its allies, while Shirley (1931, pp. 8-9) pointed to the similarities shared with ptychopariids in general and Alokistocare and Acrocephalites in particular. The resemblance of Lobocephalina (Conocephaloninae; see Moore 1959, pp. 0237- 0238, fig. 176.2c) and Modocia (Marjumiidae; see Moore 1959, p. 0306, fig. 228.2 a-d) to early homalonotids, especially Bavarilla , is also striking. Apart from the glabellar shape and sutural morphology these forms display the tendency to reduced distinct- ness of glabellar furrows, relatively short genal spines, small eyes (in Modocia ), thorax of fourteen segments with rounded pleural tips, and a posteriorly rounded pygidium which possesses an ill-defined border. By the Tremadoc/Arenig a number of genera occur from which evolutionary lines lead to younger calymenids and homalonotids. The difficulty of classifying these early genera has been discussed and is indicative of their common origin. In the Arenig representatives of the Colpocoryphinae, Eohomalonotinae, and Homalonotinae are present, the colpocoryphines and eohomalonotines becoming extinct in the Caradoc and Ashgill respectively. Homalonotines persist until the middle Devonian though they are rare in the Ashgill and even more so in the Llan- dovery (see pp. 170-171). Trimerus first appears in the upper Llandovery and seems to have given rise to younger genera. A new radiation began in the upper Silurian (three genera) and continued in the lower Devonian (six genera). The stratigraphical ranges of the Devonian genera are imprecisely known but on present evidence Trimerus disappears in the Siegenian while Parahomalonotus , Burmeisterella , Burmeisteria , and Digonus are restricted to the interval Gedinnian to Emsian. Only Dipleura is certainly known from the middle Devonian. These genera are in need of revision but show many similarities to Trimerus and Dipleura and were probably derived from one or both of these genera. The cephalon of Burmeisteria , for instance, shares many features with juvenile specimens of Dipleura (compare Morzadec 1969, pi. 2, fig. 6 a with Moore 1959, p. 0460, fig. 361. lu). The form of the glabella and lateral glabellar furrows are exceedingly similar and Burmeisteria may be a neotenous descendant of Dipleura. As far as Ordovician forms are concerned, Colpocoryphe and Plaesiacomia share many common characters, some of which are due to the presence of coaptative structures associated with enrolment. Many of the features of Colpocoryphe are present in a more extreme form in Plaesiacomia— the tendency to efface glabellar 172 PALAEONTOLOGY, VOLUME 20 furrows and loss of distinct pygidial segmentation, for instance— and I regard Plaesiacomia as being derived from early Colpocoryphe species. Colpocoryphe dis- plays some ‘primitive’ features concordant with this view in that strong eye ridges are often developed and paraglabellar areas absent (Dean 1966/), p. 141). The evidence is inconclusive, however, since the oldest Plaesiacomia species have smooth cephala without eye ridges and resemble Colpocoryphe mainly in having a relatively large number of pygidial axial rings. Fortey (1974, pp. 61-65, pi. 20, figs. 1-12) described the unusual olenid Svalhard- ites hamus , from the Arenig of Spitsbergen, and noted (p. 64) the similarity of its pygidium to those of Plaesiacomia species, although he doubted a direct relation- ship. The form of the colpocoryphine pygidium is largely determined by its role in enrolment, and the lack of a deep border furrow in Svalbardites suggests that the resemblance is superficial. There is also some cephalic similarity, notably in eye position and general effacement, but the Svalbardites glabella is completely unlike the forwardly tapering form characteristic of homalonotids. Homalonotids have a well-developed rostral plate, olenids commonly have no rostral plate and the free cheeks are united by a narrow strip of doublure. I consider an olenid-homalonotid relationship unlikely and the similarities between Plaesiacomia and Svalbardites to be superficial. Eohomalonotus and Calymenella form a distinct group. Some early species of Platycoryphe resemble Calymenella in having distinct eye ridges and a sigmoidal lp furrow (see Henry 1971). In other respects (no distinct cephalic border; pygidial form; no prominent pygidial interpleural furrows), however, they are closer to later species of Platycoryphe. Such forms seem to be close to the point of divergence of homalonotines and eohomalonotines. Platycoryphe and Brongniartella are similar genera and the difficulty of discrimi- nating between them when the glabellar furrows are weak has often been noted (Dean 1961, p. 344; Whittington 1965, p. 488; Hughes 1969, p. 98). The resemblances are unsurprising since Brongniartella may have been derived from Platycoryphe (Whittington 1966 b, p. 723, text-fig. 13). The Llandovery homalonotine, P. dyaulax , shows features in common with both younger homalonotids and with Platycoryphe and Brongniartella. I consider that post-Ordovician homalonotines were derived from this stock. Sdzuy (1957, p. 284) suggested that the pygidial morphology of Brongniartella indicated that it did not give rise to younger genera. He did not develop this theme, and Platycoryphe was not discussed. Sdzuy (p. 286, fig. 3) derived Trimerus from Eohomalonotus but, because of the morphological differences dis- cussed above, this view is considered untenable. Silurian-Devonian homalonotines comprise a closely related, morphologically homogeneous group (for a summary of their relationships see Tomczykowa 1975). The Ludlow genus Homalonotus may have evolved from an upper Wenlock form like T. johannis by the loss of glabellar lobation together with a reduction in length, and folding, of the frontal area. Other features of the exoskeleton are very similar (compare Salter 1865, pi. 12, figs. 2-10 and pi. 13, fig. 8 with pi. 12, fig. 1 1 and pi. 13, figs. 1-7). Tomczykowa (1 975, p. 14) suggested that T. johannis might also be ancestral to Digonus. Such a relationship is plausible but a substantial stratigraphical gap exists between the occurrence of T. johannis and the earliest record of Digonus in the THOMAS: HOMALONOTID TRILOBITES 173 Gedinnian. Dipleura first appears in the Ludlow and may have been derived from Trimerus through a form similar to T. lobatus Tomczykowa, 1975. Throughout the Ordovician, and in the early Silurian, homalonotines display a progressive increase in the width (tr.) of the thoracic axis and a tendency to reduced distinctness of trilobation. By the middle Silurian the thoracic axis is very wide and trilobation quite indistinct. The thoracic segments of homalonotines and colpo- coryphines differ from those of all other trilobites in that the articulating and pleural furrows merge and the articulating half ring becomes fused with the anterior pleural band. Hence the articulating half ring becomes greatly enlarged (tr.) and comes to occupy much of the width of the segment (see PL 24, fig. 12 a-b). Material of Caly- menella loaned to me by Professor H. B. Whittington indicates that eohomalonotines differ from other homalonotids in this respect. They have a more ‘normal’ type of articulation, the articulating half rings not extending outside the axis (see also Sdzuy 1957, pi. 1, fig. 8). This phenomenon is associated with the retention of a relatively narrow thoracic axis in eohomalonotines. The selective forces responsible for the evolutionary changes outlined above are unknown. The adoption of a rather specialized mode of life may explain the morpho- logical contrasts between the Colpocoryphinae/Homalonotinae and their inferred ancestor, Bavarilla. The similarity of Silurian-Devonian genera suggests that the post-Ordovician homalonotine radiation was connected with the exploitation of a number of closely related niches, the general mode of life being broadly similar. Homalonotus is the most unusual post-Ordovician homalonotine. The complex structure of the anterior part of the cephalon (see Salter 1865, pi. 12, fig. 2) may indicate that this genus had somewhat different life habits. Eohomalonotines change little throughout their history and show few major differences from Bavarilla. They appear to represent a group whose way of life changed little from that of their ancestors. The retention of several ‘primitive’ characters (e.g. eye ridges, anteriorly widened border, sigmoidal lp furrow, relatively small rostral plate, thoracic morphology) accords with this view. A case could there- fore be argued for placing the colpocoryphines and homalonotines in a separate family from the Eohomalonotinae and Bavarilla. This reflects the inferred phylogeny in which the eohomalonotines represent a relatively unchanged conservative stock while the Colpocoryphinae and Homalonotinae are ‘progressive’ groups united, in particular, by their thoracic structure. While such an arrangement is in some ways appealing, its adoption would be premature. The pattern outlined may need modifica- tion when relationships between homalonotids and calymenids in general, and between Bavarilla and older genera, are better understood. What is known of homalonotid distribution supports the proposed phylogeny. Whittington (19666, 1972) and Whittington and Hughes (1972, 1973) have demon- strated the existence of discrete faunal provinces in which different trilobite groups evolved in relative isolation through much of Ordovician time. In the Arenig- Llandeilo Colpocoryphe , Plaesiaeomia , Eohomalonotus , Calymenella, and Platy- coryphe are restricted to the Florida/west-central Europe/North Africa region (the Selenopeltis province), a part of which Bavarilla inhabited in the Tremadoc. The upper Ordovician decay of faunal provincialism is reflected in homalonotid Leiostegina SUBFAM. UNCERTAIN HOMALONOTINAE EOHOMALON- OTINAE COLPOCORYPH- INAE a £ Devonian post-Ludlow pre-Gedinnian Ludlow Wenlock Llandovery Ashgill Caradoc Llandeilo Llanvirn Arenig Tremadoc text-fig. I . Phylogeny of the Homalonotidae. Solid lines represent known ranges ol genera, broken ones their inferred relationship. THOMAS: HOMALONOTID TRILOBITES 175 distribution with Brongniartella and Platycoryphe extending their range into North America and Kazakhstan during the Caradoc and Ashgill. In the Ashgill Brongniar- tella is widespread in northern Europe, Calymenella being present in central China (Whittington 1966 h, p. 728) and possibly Argentina (Baldis and Blasco 1975). Some of the unusual features of Leiostegina , from the Caradoc of Bolivia, may be due to its having evolved in relative isolation. According to Whittington and Hughes’s (1972, p. 257, fig. 9) palaeogeographical map Bolivia lay on the opposite side of Gondwanaland to the Selenopeltis fauna, although at a similar latitude. No marked provincialism is developed in the Silurian. Platycoryphe occurs in the Llandovery of Saudi Arabia and China and Brongniartella in the low Llandovery of South Wales. Trimerus first occurs in the upper Llandovery of South Wales and Paraguay and extends its range in the later Silurian and lower Devonian, being known from North America, Europe, Mongolia, and Australia. Homalonotus first appears in the lower Ludlow of South Wales and throughout its short range was restricted to Nova Scotia, Britain, Scandinavia, and Poland. On a pre-drift map these areas lie close together some 10°-20 south of the equator. Dipleura is first known from the upper Ludlow of Britain and Poland and was probably derived from one of the Trimerus species found in this area. Homalonotids are widespread in the Devonian but it is not possible to discuss the distribution of individual genera since the taxonomy is poorly known. Acknowledgements. I thank the Arabian American Oil Company and Dr. C. P. Hughes for passing on the material described herein. Professor H. B. Whittington, F.R.S. and Dr. R. M. Owens critically read the manuscript and made useful suggestions for its improvement. Dr. R. A. Fortey and Mr. S. F. Morris (BM), Dr. A. W. A. Rushton (GSM), and Dr. R. B. Rickards (SM) kindly gave access to specimens in their care. My wife carefully typed and checked the manuscript. A N.E.R.C. Research Studentship is gratefully acknowledged. Note on orientation. Dorsal view refers to the orientation in which cephala are placed with the occipital ring in a vertical plane and pygidia placed with the anterior margin of the axis vertical. REFERENCES Apollonov, M. K. 1974. Ashgill trilobites from Kazakhstan. Acad. Nauk. Kazakh. S.S.S.R. 11 36, 21 pis. [In Russian.] Balashova, E. A. 1966. Trilobites from the Ordovician and Silurian Beds of Pamir. Trudy uprav. geol. sov. Minist. Tadzhik, 2, 191-262, 4 pis. [In Russian.] 1968. New representatives of the Order Polymera from various regions of the USSR. In markovsky, b. p. (ed.). New species of prehistoric plants and invertebrates of the USSR. Ministr. geol. okhran. Nedr SSSR. Vses. nauchnoissled. Geol. Inst. (VSEGEI), 194-210, pis. 51-54. [In Russian ] baldis, B. a. and blasco, g. 1975. Primeros trilobites Ashgillianos del Ordovicico Sudamericano. Adas 1 Congreso Argentino de Paleontologla Biostratigrafia, 1, 33-48, 1 pi. barrande, J. 1852. Systeme Silurien da centre de la Boheme. lere partie. Recherches paleontologiques. Vol. 1. Crustaces, Trilobites. xxx + 935 pp., 51 pis. Prague and Paris. — 1868. Faune silurienne des environs de Hof, en Baviere. Pp. 31-110, 1 pi. Prague and Paris. bergeron, j. 1890. Sur une forme nouvelle de Trilobite de la famille des Calymenidae (genre Calymenella). Bull. Soc. geol. Fr. (3), 18, 365-371, pi. 5. bergstrom, j. 1973. Organisation, life, and systematics of trilobites. Fossils and Strata, 2, 69 pp., 5 pis. chang, w.-T. 1974. In A handbook of the stratigraphy and paleontology in southwest China. Edited by the Nanking Institute of Geology and Paleontology Academia Sinica. Science Press, iii+454 pp., 202 pis. [In Chinese.] 176 PALAEONTOLOGY, VOLUME 20 chapman, e. j. 1890. Some remarks on the classification of the trilobites as influenced by stratigraphical relations: with outline of a new grouping of these forms. Trans. R. Soc. Can. 7 (Section 4), (for 1889), 113-120. Clarke, J. M. 1913. Fosseis Devomanos do Parana. Monografias Div. geol. Bras. 1, 353 pp., 7 pis. clarkson, e. n. k. and henry, j.-l. 1973. Structures coaptatives et enroulement chez quelques Trilobites ordoviciens et siluriens. Lethaia , 6, 105-132. cooper, G. A. 1935. Young stages of the Devonian trilobite Dipleura dekayi Gr. J. Paleont. 9, 3-5, pi. 1. dean, w. t. 1961. The Ordovician trilobite faunas of South Shropshire, II. Bull. Br. Mus. nat. Hist. (Geol.), 5, 311-358, pis. 49-55. — 1966u. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the neighbour- ing district of the Montagne Noire, south-western France. Ibid. 12, 245-353, pis. 1-21. — 1966 b. A revision of the Ordovician trilobite genus Plaesiacomia Hawle and Corda, 1847. Sb. ndr. Mus. Praze , 22B, 133-142, pis. 1-3. — 1971 . The Lower Palaeozoic stratigraphy and faunas of the Taurus Mountains near Bey§ehir, Turkey. II. The trilobites of the Seydi§ehir Formation (Ordovician). Bull. Br. Mus. nat. Hist. (Geol.), 20, 1-24, pis. 1-5. — 1976. Some aspects of Ordovician correlation and trilobite distribution in the Canadian Appalachians. In bassett, M. G. (ed.). The Ordovician System : Proceedings of a Palaeontological Association symposium , Birmingham , September 1974. University of Wales Press and National Museum of Wales, Cardiff. Pp. 227-250. deslongchamps, e. 1825. Memoire sur les corps organises fossiles du gres intermediate du Calvados. Soc. Linn. Calvados , 2, 291, 2 pis. eaton, a. 1832. Trilobites. Am. J. Sci. (1), 22, 165-166. foerste, a. f. 1910. Preliminary notes on Cincinnatian and Lexington fossils of Ohio, Indiana, Kentucky and Tennessee. Bull, scient. Labs Denison Univ. 16, 17-87, pis. 1-6. — 1919. Silurian fossils from Ohio, with notes on related species from other horizons. Ohio J. Sci. 19, 367-404, pis. 16-19. fortey, R. a. 1974. The Ordovician Trilobites of Spitsbergen I. Olemdae. Skr. norsk Polarinst. 160, 129 pp., 24 pis. green, j. 1832. A monograph of the trilobites of North America: with coloured models of the species. 94 pp., 1 pi. Philadelphia. gurich, G. 1909. Leitfossilien : ein Hifsbuch zum Bestimmen von Versteinerungen bei geologischen Arbeiten in der Sammlung und im Felde. Zweite Lieferung. Devon. Pp. 97-199, pis. 29-52. Berlin. hall, j. 1860. New species of fossils from the Hudson River Group of Ohio, and other western states. Rep. N.Y. St. Mus. nat. Hist. 13, 119-121. hawle, I. and corda, A. J. c. 1847. Prodrom einer Monographic der bohmischen Trilobiten. 176 pp., 7 pis. Prague. (Also 1848, Abh. K. bohm. Ges. 1 Viss. 5, 117-292, pis. 1-7.) henningsmoen, g. 1957. The trilobite family Olenidae; with description of Norwegian material and remarks on the Olenid and Tremadocian Series. Skr. Norske Vid.-Akad. Oslo , 1. Mat. -Nat. Kl. 1, 303 pp., 31 pis. henry, j.-l. 1971. Les trilobites Asaphidae et Eohomalonotidae du Gres Amoricain superieur (?Areni- gien) de l’ouest de la France. Mem. Bur. Rech. geol. minier. 73, 65-77, 2 pis. hicks, h. 1873. On the Tremadoc rocks in the neighbourhood of St. David’s, South Wales and their fossil contents. Q. J! geol. Soc. Lond. 29, 39-52, pis. 3-5. hughes, c. p. 1969. The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, central Wales. Part 1. Bull. Br. Mus. nat. Hist. (Geol.), 18, 39-103, pis. 1-14. hupe, p. 1953. Classe de trilobites. In piveteau, j. (ed.). Traite de paleontologie , 3, 44-246, 140 text-figs. Paris. 1955. Classification des trilobites. Annls Paleont. 41, 91-325 (111-345), figs. 93-247. jones, o. t. 1925. The Geology vof the Llandovery District, Part 1. The Southern Area. Q. Jl geol. Soc. Lond. 81, 344-388, pi. 21. kerforne, f. 1900. Description de trois nouveaux Trilobites de l’Ordovicien de Bretagne. Bull. Soc. geol. Fr. (3), 28, 783-791, pi. 13. kobayashi, T. 1937. The Cambro-Ordovician shelly faunas of South America. J. Fac. Sci. Tokyo Univ.. 2 (Geol.), 4, 369-522, pis. 1-8. — 1960. Some Ordovician fossils from East Tonkin, Viet Nam. Jap. J. Geol. Geogr. 31, 39-48, pi. 5. THOMAS: HOMALONOTID TRILOBITES 177 kobayashi, t. 1969. The Cambro-Ordovician formations and faunas of South Korea, Part X. Stratigraphy of the Chosen Group in Korea and South Manchuria and its relation to the Cambro-Ordovician forma- tions of other areas. Section D. The Ordovician of East Asia and other parts of the continent. J. Fac. Sci. Tokyo Univ., 2 (Geol.), 17, 163-316. konig, c. e. 1825. leones fossilium sectiles. 4 pp., 19 pis. London. lake, p. 1906. Trilobites from Bolivia, collected by Dr. J. W. Evans in 1901 1902. Q. Jl geol. Soc. Lond. 62, 425-430, pi. 40. leach, w. E. 1824. Monograph on the Cebrionidae, a Family of Insects. Zool. J. Lond. (publication dis- continued 1835), 1, 33-46. lebesconte, m. p. 1887. Constitution generate du massif breton comparee a celle du Finistere. Bull. Soc. geol. Fr. (3), 14 (for 1886), 776-820, pis. 34-36. lesperance, p. j. 1968. Ordovician and Silurian trilobite faunas of the White Ftead Formation, Perce region, Quebec. J. Paleont. 42, 81 1-826, pi. 106. lu, y.-h. 1963. Qinlin huashi shotice (Handbook of fossils from the Ch’in-lin Mountains). [In Chinese.] marr, j. e. and Nicholson, h. a. 1888. The Stockdale Shales. Q. Jl geol. Soc. Lond. 44, 654-732, pi. 16. moore, R. c. (ed.). 1959. Treatise on Invertebrate Paleontology , Part O. Arthropoda 1. xix + 560 pp., 415 figs. Geol. Soc. Amer. and Univ. of Kansas Press (Lawrence). morzadec, P. 1969. Le Devonien de la rive nord de la riviere de Faou (Finistere). Etude Stratigrapliique — Etude des Trilobites. Bull. Soc. geol. mineral, de Bretagne , 1-58, pis. 1 9. Murchison, R. I. 1839. The Silurian System, founded on geological researches in the counties of Salop, Hereford , Radnor, Montgomery, Caermarthen, Brecon, Pembroke, Monmouth , Gloucester, Worcester and Stafford; with descriptions of the coalfields and overlying formations, xxxii + 768 pp., 37 pis. London. novak, o. In perner, j. 1918. Trilobiti Pasma U-d\y z okoli Prazskeho. Palaeontogr. Bohem. 9, 51 pp., 4 pis. pocock, r. w., whitehead, t. h., wedd, c. b. and robertson, t. 1938. Shrewsbury District, including the Hanwood Coalfield. Mem. geol. Surv. U.K. xii -F 297 pp., 8 pis. pribyl, a. 1953. Index of trilobite genera in Bohemia. Krihova ust. Ust. geol. 25, 80 pp. [Czech text, English and Russian summaries.] — 1957. New supplements to the classification of the Bohemian trilobite genera. Sb. ustred. Ust. geol. (Paleont.), 23, 53-66. [In Czech, English text pp. 81-94.] reed, F. R. c. 1918. Notes on the genus Homalonotus. Geol. Mag. (6), 5, 263-276, 314-327. robardet, m., henry, J.-L., nion, j., Paris, f. and pillet, J. 1972. La Formation du Pont-de-Caen (Cara- docien) dans les synclinaux de Domfront et de Sees (Normandie). Annls Soc. geol. N. 92, 117-137, pis. 18-22. sadler, p. m. 1974. Trilobites from the Gorran Quartzites, Ordovician of south Cornwall. Palaeontology, 17, 71-93, pis. 9-10. salter, J. w. 1865. A monograph of the British trilobites from the Cambrian, Silurian and Devonian forma- tions. Palaeontogr. Soc. [Monogr.], (2), 81-128, pis. 7-14. Saul, J. M. 1967. Burmeisteria ( Digonus ) accraensis, a new homalonotid trilobite from the Devonian of Ghana. J. Paleont. 41, 1126-1136, pis. 143-144. savage, t. e. 1913. Stratigraphy and palaeontology of the Alexandrian Series in Illinois and Missouri, Part 1. Illinois Geol. Surv. Bull. 23, 124 pp., 7 pis. (reprinted in complete Bulletin , 23, 1917, 67-160, pis. 3-9). sdzuy, k. 1957. Bermerkungen zur Familie Homalonotidae (mit der Beschreibung einer neuen Art von Calymenella). Senckenberg. leth. 38, 275-290, 1 pi. sheng, s. f. 1964. Upper Ordovician trilobite faunas of Szechuan-Kweichow with special discussion on the classification and boundaries of the upper Ordovician. Acta palaeont. sinica, 12, 537-571 . [In Chinese, English text. Pp. 553-563.] Shirley, j. 1931. A redescription of the known British Ordovician species of Calymene (s.l.). Mem. Proc. Manchr. lit. phil. Soc. 75, 1-35, pis. 1-2. struve, w. 1958. Uber einige Falle von Homonymie bei Gattung der Phacopacea. Senckenberg. leth. 39, 301-302. swinnerton, h. h. 1915. Suggestions for a Revised Classification of Trilobites. Geol. Mag. (6), 2, 487-496, 538-545. M 178 PALAEONTOLOGY, VOLUME 20 temple, L T. 1975. Early Llandovery trilobites from Wales with notes on British Llandovery calymenids. Palaeontology, 18, 137-159, pis. 25-27. tomczykowa, E. 1975. The trilobite subfamily Homalonotinae from the Upper Silurian and Lower Devon- ian of Poland. Acta palaeont. pol. 20, 3-46, pis. 1-6. toumansky, o. G. 1938. Remarks on the ammonite family Popanoceratidae Hyatt with special mention of two new genera, Pamirites and Propopanoceras. Sov. Geol. 8, 106-108, 6 figs. [In Russian.] vanek, J. 1965. New species of the suborder Calymenina Swinnerton, 1915 (Trilobita) from the Barrandian area. Sb. geol. ved. Praha (Pal.), 6, 21-38, 4 pis. warburg, E. 1925. The trilobites of the Leptaena Limestone in Dalarne with a discussion of the zoological position and the classification of the Trilobita. Bull. geol. Instn Univ. Upsala , 17, vi + 446 pp., 11 pis. whittard, w. F. 1960. The Ordovician trilobites of the Shelve Inlier, Shropshire. Palaeontogr . Soc. [Monogr.], (4), 117-162, pis. 16-21. 1961. Ibid. (5), 163-196, pis. 22-25. Whittington, H. B. 1953. A new Ordovician trilobite from Llorida. Breviora , 17, 6 pp., 1 pi. — 1965. Platycoryphe , an Ordovician homalonotid trilobite. J. Paleont. 39, 487-491, pi. 64. — 1966a. Trilobites of the Henllan Ash, Arenig Series, Merioneth. Bull. Br. Mus. nat. Hist. (Geol.), 11, 491-505, 5 pis. — 19666. Phylogeny and distribution of Ordovician trilobites. J. Paleont. 40, 696-737. — 1971. Silurian calymenid trilobites from United States, Norway, and Sweden. Palaeontology, 14, 455-477, pis. 83-89. ' — 1972. Ordovician trilobites. In hallam, a. (ed.). Atlas of Palaeobiogeography, 13-18. Amsterdam. — and hughes, c. p. 1972. Ordovician geography and faunal provinces deduced from trilobite distribu- tion. Phil. Trans. R. Soc. Lond B263, 235-278. 1973. Ordovician trilobite distribution and geography. In hughes, n. f. (ed.). Organisms and continents through time. Spec. Pap. Palaeont. 12, 235-240. wolf art, r. 1961. Stratigraphie und Launa des alteren Palaozoikums (Silur, Devon) im Paraguay. Geol. Jb. 78, 29-102, pis. 2-7. — bender, f. and stein, v. 1968. Stratigraphie und Launa des Ober-Ordoviziums (Caradoc-Ashgill) und Unter-Silurs (Unter-Llandovery) von Siidjordien. Ibid. 85, 517-564, pis. 45-50. yin, t.-h. 1966. China in the Silurian period. J. geol. Soc. Aust. 13, 277-297 . Typescript received 10 January 1976 Revised typescript received 20 Lebruary 1976 A. T. THOMAS Department of Geology Sedgwick Museum Downing Street Cambridge CB2 3EQ DINOFLAGELLATE CYSTS FROM THE BEARPAW FORMATION (7UPPER CAMPANIAN TO MAASTRICHTIAN) OF MONTANA by REX HARLAND Abstract. A dinoflagellate cyst assemblage is described from the Bearpaw Formation of Montana, U.S.A. Deflan- drea montanaensis sp. nov. is described. Archeopyle formation in Senegalinium tricuspis (O. Wetzel) comb. nov. is demonstrated to be of the large intercalary type, and this, together with its basic cavate morphology, allows for the recombination of the species. Also recombined are S. magnified (Stanley). IS. albertii (Corradini), S. boloni- ensis (Riegel), S. gaditanum (Riegel), IS. kozlowskii (Gorka), S. pannuceum (Stanley), S. pentagonalis (Corradini), and S. subquadratum (Corradini). The Bearpaw Formation in Montana, on radiometric and on dinoflagellate cyst evidence is younger than that seen in southern Alberta, and may include strata encompassing the Campanian- Maastrichtian boundary. The Bearpaw Formation of the northern U.S.A. and Canada was deposited during the last major transgression of the Late Cretaceous sea, and before sedimentation was influenced by the growth of alluvial plains from the newly uplifted Cordillera (Stelck 1967). The base of the formation is known to be of Late Campanian age in both southern Alberta and Saskatchewan, Canada (Caldwell 1968). In Saskatchewan the formation is believed to extend into the Maastrichtian, whereas in southern Alberta it is restricted to the Campanian. A potassium-argon date of 75± 4 million years is available on a bentonite close to the base of the formation at Lethbridge, Alberta (Folinsbee et al. 1960, 1961), and is indicative of a Late Cretaceous age (Casey 1964; Lambert 1971). Norton and Hall (1969) state that the Bearpaw Formation at Hell Creek, Montana, U.S.A., is Late Cretaceous in age and report a date of 70 million years from a posi- tion close to the level of their sample KB- 1, i.e. close to the base of the unit as seen at Hell Creek. This suggests that the section in Montana is somewhat younger than that in southern Alberta and is probably equivalent to the upper part of the section in Saskatchewan close to, if it does not contain, the Campanian-Maastrichtian boundary, i.e. 70 m.y. or 72 m.y. (Casey 1964 and Lambert 1971 respectively). Recently Obradovich and Cobban (1975) in discussing a time-scale for the Late Cretaceous of North America suggest a 70-71 m.y. date for the Campanian Maastrichtian boundary, but admit difficulty in defining the boundary palaeon- tologically. Harland (1973) following Caldwell (1968) accepted the base of the Baculites baculus Zone as the Campanian-Maastrichtian boundary but Obradovich and Cobban (1975) suggest it may fall as low as the base of the B. reesidei Zone, i.e. much of the southern Alberta Bearpaw would then be Maastrichtian. This view is not accepted here. The present study was undertaken to describe and compare the dinoflagellate cyst assemblage from Montana with that published from southern Alberta (Harland 1 973). The southern Alberta outcrop is some 230 miles north-west of the Montana section. [Palaeontology, Vol. 20, Part 1, 1977, pp. 179-193, pi. 25.] 180 PALAEONTOLOGY, VOLUME 20 Miles text-fig. 1. Sketch map of a part of the Fort Peck Reservoir in north-eastern Montana, U.S.A. to show the position of the sample locality (marked with an X). Bold lines indicate roads. text-fig. 2. A stratigraphical section of the Bearpaw Forma- tion at the Hell Creek locality showing the position of the samples. MATERIALS AND METHODS In June 1969 Drs. G. Playford, University of Queensland, Australia, and G. D. Williams, University of Alberta, Canada, collected eight samples of Bearpaw Formation from the banks of the Fort Peck Reser- voir, Montana, U.S.A., i.e. locality 1 of Norton and Hall (1969). This locality (text-fig. 1 ) is situated some 22 miles north of Jordan on the Hell Creek part of the reservoir at Sy, Sect. 31, T22N, R37E, Garfield County, Montana. At the time of collection 70 ft (21-33 m) of Bearpaw was exposed above the water level in the reservoir and beneath the overlying Fox Hills Formation. Norton and Hall (1969) described the distinct lithological change from the dark Bearpaw shale to the light-coloured shales and sandstones of the Fox Hills sandstone but no unconformity was noted. The Bearpaw Formation at this locality can be divided into a lower shale unit and an upper silty shale unit and both appear as very dark rock when freshly exposed (Norton and Hall 1969). The sample distribution through the section is shown in text-fig. 2; the samples have been registered in the Palynological Collections of the Institute of Geological Sciences (I.G.S.) at Leeds as SAL 4379- SAL 4386. All eight samples were processed using standard palynological techniques, but beyond the hydrofluoric acid stage the residues were handled using the filtration system of Neves and Dale (1963). Where dimensions are quoted the figure in parenthesis is the arithmetic mean of the measured morpho- logical parameters. All illustrated material is registered in the MPK series of the palynological collections of the I.G.S. at Leeds. The ranges of species as quoted from Harker and Sarjeant (1975) are compilations from their tables in order to indicate the world-wide range. SYSTEMATICS The dinoflagellate cysts of this paper are placed within either a gonyaulacacean or peridiniacean grouping. It is felt that although the emended supra-generic classifica- tion of Sarjeant and Downie (1974) goes a long way to answering the criticisms of HARLAND: DINOFLAGELLATES FROM MONTANA 181 -i; SAL 2 ■« <■5 o T-r i i .4 c I £ v 3 . I s Cn O 1 1 -I §■ §■ t!> s.^Q90Q'>:6.Q“i<,000QQ 4385 4384 4383 4382 4381 4380 4379 '» )( H <> * i( H <> O o 1( t I I I 1 I 1 » K M * text-fig. 3. The range of species through the Bearpaw Formation at the Hell Creek locality, Fort Peck Reservoir, Montana, showing the possible Campanian-Maastrichtian boundary. Range of species that are known to be greater than that as proved at Hell Creek are shown by a broken line. The large dots represent the presence of a species within the sample indicated, and the crosses indicate an occurrence of greater than 10% of the dinoflagellate cyst population. Wall and Dale (1968) there are too many new forms being described and too much taxonomic revision to use it or any other formal supra-generic classification at the present time. This is particularly true of deflandreoid cysts where detailed knowledge of their morphology, particularly archeopyle formation, is still lacking. There is, however, no question that at least two major groupings can be recognized amongst fossil dinoflagellate cysts and these are used here; the Gonyaulacaceae and the Peridiniaceae. 182 PALAEONTOLOGY, VOLUME 20 Division pyrrhophyta Pascher Class dinophyceae Pascher Order peridiniales Lindemann GONYAULACACEAN GROUP Genus cyclonephelium Deflandre and Cookson emend. Williams and Downie 1966 Type species. Cyclonephelium compactum Deflandre and Cookson, 1955; O.D. Cyclonephelium distinction Deflandre and Cookson, 1955 Plate 25, fig. 14 1955 Cyclonephelium distinctum Deflandre and Cookson, pp. 285-286, pi. 2, fig. 14; text-figs. 47, 48. Figured material. Slide SAL 4381-RH1, specimen MPK 917. Remarks. Only a few specimens were found in the lower part of the sequence and the illustrated form falls at the smaller end of the size range of this species as quoted by Harland (1973). Sarjeant (1967u) gives a Hauterivian to Santonian range for this species but it is now known from the Campanian (Clarke and Verdier 1967; Harland 1973) and from the ?Maastrichtian (McIntyre 1974). Harker and Sarjeant (1975) record a range of Berriasian to ?early Palaeocene. Genus dictyopyxidia Eisenack, 1961 Type species. Dictyopyxidia areolata (Cookson and Eisenack) Eisenack and Kjellstrom, 1971 ; O.D. Dictyopyxidia sp. Plate 25, fig. 15 Figured material. Slide SAL 4380-RH1, specimen MPK 918. Remarks. This appears to be a new species but because only a single specimen was observed its formal description is not attempted. It is closely comparable to D. circu- lata Clarke and Verdier, 1967 but does not appear to have the complexity of fields or the rounded ambitus of that species. It was found towards the base of the section. Genus oligosphaeridium Davey and Williams, 1966 Type species. Oligosphaeridium complex (White) Davey and Williams, 1966; O.D. ? Oligosphaeridium anthophorum (Cookson and Eisenack) Davey, 1969 1958 Hystrichosphaeridium anthophorum Cookson and Eisenack, pp. 43-44, pi. 11, figs. 12, 13; text-figs. 16-18. 1969 Oligosphaeridium anthophorum (Cookson and Eisenack) Davey, pp. 147-148, pi. 5, figs. 1-3. Remarks. This species was observed throughout the section. Its range, noted as ?Oxfordian to Albian by Sarjeant (1967u), is now known to be at least into the ?Maastrichtian (McIntyre 1974), and Harker and Sarjeant (1975) indicate a Ber- riasian to ?early Oligocene range. It is quite possible, however, that the present assignment is incorrect, as the Jurassic forms, probably like the holotype, have HARLAND: DINOFLAGELLATES FROM MONTANA 183 solid processes (R. J. Davey pers. comm. 1975) whereas these and other Upper Cretaceous forms have hollow processes. This species could, therefore, benefit from restudy and it is quite likely that these and other Late Cretaceous forms are a separate, possibly new, species. Oligosphaeridium pulcherrimum (Deflandre and Cookson) Davey and Williams, 1966 1955 Hystrichosphaeridium pulcherrimum Deflandre and Cookson, pp. 270-271, pi. 1, fig. 8; text- figs. 21, 22. 1966 Oligosphaeridium pulcherrimum (Deflandre and Cookson) Davey and Williams, pp. 75-76, pi. 10, fig. 9; pi. 11, fig. 5. Remarks. This form was recorded from the lower part of the sequence but its range (Sarjeant 1967a) is Valanginian to lower Eocene. It was recorded from the Bearpaw of southern Alberta (Harland 1973) and more recently from the Campanian and Maastrichtian of Arctic Canada (McIntyre 1974). Barker and Sarjeant (1975) indicate a ?Jurassic to ?middle Eocene range; middle Albian to late Campanian in North America. Genus spiniferites Mantell emend. Sarjeant 1970 Type species. Spiniferites ramosus (Ehrenberg) Mantell, 1854; S.D. by Loeblich and Loeblich 1966. Spiniferites ramosus (Ehren berg) Mantell, 1854 1838 Xanthidium ramosum Ehrenberg, pi. 1, figs. 1, 2, 5. 1854 Spiniferites ramosus (Ehrenberg) Mantell, p. 239, Lign 77, fig. 4. Remarks. Specimens of this species complex were recorded in all the samples studied. S. ramosus has a Berriasian to Recent stratigraphic range (Harker and Sarjeant 1 975). peridiniaceaen group Genus australiella Vozzhennikova, 1967 Type species. Australiella tripartita (Cookson and Eisenack) Vozzhennikova, 1967 ; O.D. Australiella cooksoniae (Alberti) Vozzhennikova, 1967 Plate 25, fig. 9 1959 Deflandrea cooksoni Alberti, 97-98, pi. 9, figs. 1-6. 1967 Australiella cooksoni (Alberti) Vozzhennikova; 132, pi. LXI, figs. 1-4. Figured material. Slide SAL 4385-RH1, specimen MPK 919. Remarks. Specimens of this species occur throughout the studied sequence. Its recorded range (Sarjeant 1967a) is Santonian to Campanian. McIntyre (1974) recorded a form he referred to as Deflandrea sp. cf. D. cooksoni from the '/Maastricht- ian, and Zaitzeff and Cross (1970) have also recorded it from the Maastrichtian. The form described as D. korojonensis Cookson and Eisenack by Harland (1973) should more correctly be assigned here. Harker and Sarjeant (1975) give an ?early Cenomanian to early Palaeocene age for this species. 184 PALAEONTOLOGY, VOLUME 20 Australiella tripartita (Cookson and Eisenack) Vozzhennikova, 1967 I960 Deflandrea tripartita Cookson and Eisenack, 2, pi. 1, fig. 10. 1967 Australiella tripartita (Cookson and Eisenack) Vozzhennikova, pp. 134-135, pi. LXE fig. 1 ; pi. LXIV, figs. 1-4. Remarks. This species was observed throughout the section. Its previously recorded range (Sarjeant 1967a) was Turonian to Campanian, so the present observation may indicate an extension into the lowermost Maastrichtian in North America. Harker and Sarjeant (1975) recently gave an ?early Cenomenian to ?late Maas- trichtian range. Genus ceratiopsis Vozzhennikova, 1963 Type species. Ceratiopsis leptoderma Vozzhennikova, 1963; O.D. Ceratiopsis diebeli (Alberti) Vozzhennikova, 1967 Plate 25, fig. 16 1959 Deftandrea diebeli Alberti, 99-100, pi. 9, figs. 18-21. 1967 Ceratiopsis diebeli ( Alberti) Vozzhennikova, pp. 159-160, pi. CXIX, fig. 4. Figured material. Slide SAL 4384-RH1, specimen MPK 920. Remarks. This species was found throughout the sequence and had a previously recorded range of Santonian to Campanian (Sarjeant 1967a), it is now known from the Maastrichtian and Danian (Wilson 1971). Harker and Sarjeant (1975) indicate a ?late Coniacian to late Eocene range. Genus deflandrea Eisenack emend. Williams and Downie 1966 Type species. Deflandrea phosphoritica Eisenack, 1938; O.D. Deflandrea macrocysta Cookson and Eisenack, 1960 1960 Deflandrea macrocysta Cookson and Eisenack, 3, pi. 1, figs. 7, 8. Remarks. D. macrocysta is found throughout the section. Sarjeant (1967a) and Harker and Sarjeant (1975) recorded a restricted Campanian range for this species so that the present study may indicate a slight extension into the lowermost Maas- trichtian. Deflandrea montanaensis sp. nov. Plate 25, figs. 4, 6, 7, 10-12 Diagnosis. Cavate cyst, epitract conical, hypotract hemispheroidal with an asymmetrical ‘skirt’ and horn. Endoblast elongated apically and hemispheroidal antapically. Wall layers smooth. Apex surmounted by a bifid tip; the larger antapical horn acuminate. Tabulation discernible but not usually recognizable, delimited by low, smooth, or poorly denticulate ridges. Cingulum planar, may or may not be slightly in- dented; sulcus large and broad, widening towards the antapex. Archeopyle intercalary in periphragm and endophragm; commonly attached Ia/Ia (Evitt 1967), and apically/antapically elongate hexagonal in shape. Figured material. Holotype: Slide SAL 4380-REI1, specimen MPK 921, Bearpaw Formation, ?Cam- panian to Maastrichtian, Montana, U.S.A. Paratypes: Slide SAL 4380-RH2, specimens MPK 922-924. HARLAND: DINOFLAGELLATES FROM MONTANA 185 Dimensions. Holotype: length 35 0 p, breadth 16-25 p. Range: length 18-75 (27-50) 36-25 p, breadth 8-75 ( 17-0) 25 00 p. Twelve specimens were measured from a studied population of twenty-five. Description. A diamond-shaped to elongate fusiform cyst made up of the two wall layers that are only adpressed in the cingular region and on the upper part of the hypotract. The epitractal periphragm is drawn out into an apical horn which carries a dorso-ventrally flattened, bifid tip. The antapical ‘skirt’ and horn sometimes carry small, poorly developed, irregular spines, especially on the margin of the ‘skirt’. The tabulation is variously developed but difficult to decipher, it is probably ?4\ la, ?7,/, 6c, 2"', 12"". The cingulum is divided into six well-defined cingular plates. Archeopyle is formed by a single opening through the periphragm and endophragm and the operculum appears to remain attached (PI. 25, fig. 10), type Ia/Ia of Evitt (1967). Remarks. This cyst is closely comparable to Spinidinium clavum Harland, 1973 (see below for further comments), and it is possible that a full range of variation exists between the two forms. This was not seen to be the case, however, in either the present assemblage or in southern Alberta (Harland 1973), and therefore it is regarded as a distinct and separate species. It occurs throughout the studied section. Comparisons. This cyst is closely comparable to D. minor Alberti, 1959 from which it differs in over-all shape, D. minor being more rhomboidal and having a condensed endoblast, and in possessing a tabulation. It is also comparable with D. balmei Cookson and Eisenack, 1962 which again differs in form, in the amount of endo- blastic ‘contraction’ and in possessing spines. It is closest to S. clavum Harland, 1973 but differs in not possessing high, denticulate, sutural crests and in being much smaller, i.e. nearly half the size. It may be an evolutionary descendant of that species. It is also closely comparable to S. rallum Heisecke, 1970, D. irmoechinata Heisecke, 1970, and D. rhombica Cookson and Eisenack, 1974, all of which, however, differ in possessing numerous well-developed spines. Deflandrea cf. pirnaensis Alberti, 1959 Plate 25, fig. 8 1959 Deflandrea pirnaensis Alberti, p. 100, pi. 8, figs. 1 -5. Figured material. Slide SAL 4381-RH1, specimen MPK 925. Remarks. The single specimen encountered is most closely comparable to D. pirna- ensis■, which had a previously published range of Albian to Coniacian (Sarjeant 1967u). Harker and Sarjeant (1975) indicate a late Hauterivian to late Maastrichtian range. Genus diconodinium Eisenack and Cookson, 1960 Type species. Diconodinium multispinum (Deflandre and Cookson) Eisenack and Cookson, 1960; O.D. Diconodinium arcticum Manum and Cookson, 1964 1964 Diconodinium arcticum Manum and Cookson, pp. 1 8 19, pi. 6, figs. 1-4. Remarks. D. arcticum occurs throughout the sequence. Its previously recorded range 186 PALAEONTOLOGY, VOLUME 20 (Manum and Cookson 1964) was early late Cretaceous which has now been ascer- tained by Felix and Burbridge (in press) to be a late Cenomanian to early Campanian age. The present evidence and also that of McIntyre (1974) suggests an extension of the range into the ?Maastrichtian. Harker and Sarjeant (1975) indicate an early Cenomanian to late Campanian range. Diconodinium firmum Harland, 1973 1973 Diconodinium firmum Harland, pp. 669-670, pi. 84, figs. 8, 9, 15; text-fig. 6. Remarks. This form was found almost throughout the sequence. Its previous range was late Campanian (Harland 1973) so that its range may be increased into the ?earliest Maastrichtian (herein and McIntyre 1974). ZaitzeflT and Cross (1970) recorded Diconodinium sp. 1 which is probably synonymous to D. firmum from the Maastrichtian of Texas. After checking the original and comparable specimens of this species (all specimens illustrated in Harland (1973) are now held by the I.G.S. in Leeds) it is thought to have an archeopyle like that illustrated by McIntyre (1975) for his genus Laciniadinium , but since some doubt remains a formal recombination is not attempted here. Genus palaeocystodinium Alberti, 1961 Type species. Palaeocystodinium golzowense Alberti, 1961 ; O.D. Palaeocystodinium golzowense Alberti, 1961 Plate 25, fig. 13 1961 Palaeocystodinium golzowense Alberti, p. 20, pi. 7, figs. 10-12; pi. 12, fig. 16. Figured material. Slide SAL 4381-RH1, specimen MPK 926. Remarks. P. golzowense is confined to samples above SAL 4380. Its previously recorded range (Sarjeant \961a) was Eocene to Oligocene but it has also been EXPLANATION OF PLATE 25 All figures are at a magnification of x 500 unless otherwise stated and were photographed using phase contrast techniques. Fig. 1. Palaeoperidinium pyrophorum (Ehrenberg) Deflandre, MPK 927, dorsal view showing over-all morphology, faint growth lines and slight rupture of the epitract along the cingulum. Fig. 2. Senegalinium magnificum (Stanley) comb, nov., MPK 928, specimen showing the large, single reflected plate archeopyle and the small pericoels. Figs. 3, 5. Senegalinium tricuspis (O. Wetzel) comb, nov., MPK 929, 930, fig. 3 showing the large intercalary archeopyle. Figs. 4, 6, 7, 10-12. Deflandrea montanaensis sp. nov., MPK 921, 922, 923, 924, figs. 4, 7, holotype, figs. 4, 6, x 1000, specimens showing range of variation and over-all morphology. Fig. 8. Deflandrea cf. pirnaensis Alberti, MPK 925. Fig. 9. Australiella cooksoniae (Alberti) Vozzhennikova, MPK 919. Fig. 13. Palaeocystodinium golzowense Alberti, MPK 926, specimen showing the intercalary archeopyle. Fig. 14. Cyclonephelium distinctum Deflandre and Cookson, MPK 917. Fig. 15. Dictyopyxidia sp., MPK 918. Fig. 16. Ceratiopsis diebeli (Alberti) Vozzhennikova, MPK 920. PLATE 25 HARLAND, Campanian-Maastrichtian dinoflagellates from Montana 16 188 PALAEONTOLOGY, VOLUME 20 recorded from the Maastrichtian to ?Palaeocene (Malloy 1972). Barker and Sarjeant (1975) record an early Maastrichtian to late Oligocene range. Genus palaeoperidinium Deflandre, 1934 Type species. Palaeoperidinium pyrophorum (Ehrenberg) Deflandre, 1935, emend. Sarjeant 19676; S.D. Palaeoperidinium pyrophorum (Ehrenberg) Deflandre, 1935, emend. Sarjeant 1967/? Plate 25, fig. 1 1838 Peridinium pyrophorum Ehrenberg, pi. 1, figs. I, IV. 19676 Palaeoperidinium pyrophorum (Ehrenberg) Sarjeant, pp. 246-247, figs. 1 -6. Figured material. Slide SAL 4379-RH1, specimen MPK 927. Remarks. This cyst is found throughout the studied section. Its previously recorded range was Coniacian to Maastrichtian (Sarjeant 1967u) and now Harker and Sar- jeant (1975) give it an ?early Coniacian to ?early Palaeocene range. Genus senegalinium Jain and Millepied, 1973 Type species. Senegalinium bicavatum Jain and Millepied, 1973; O.D. Remarks. The view of Herngreen (1975) on the status of Senegalinium is not accepted here. Unfortunately Jain and Millepied (1973) did not stress the mode of archeopyle formation in their original description of Senegalinium which is characteristic and unique to this genus, i.e. possessing a large, single plate, elongate hexagonal inter- calary archeopyle which almost gives the impression of being precingular. Senegalinium magnificum (Stanley) comb. nov. Plate 25, fig. 2 1965 Deflandrea magnifica Stanley, pp. 218-219, pi. 20, figs. 1-6. Figured material. Slide SAL 4383-RH1, specimen MPK 928. Remarks. This species is herein recombined into Senegalinium as it possesses the large intercalary single-plate archeopyle and the small pericoels characteristic of the genus. It was recorded from a single sample in the Montana section. It had a Palaeo- cene range (Sarjeant 1967u) but Kjellstrom figured it as Lejeunia hyalina from the Maastrichtian (fig. 1 of Kjellstrom 1972) and this is probably confirmed herein. Zaitzeff and Cross (1970) also have recorded it from the Maastrichtian. Harker and Sarjeant (1975) record the range as early Santonian to early Eocene. Senegalinium tricuspis (O. Wetzel) comb. nov. Plate 25, figs. 3, 5 19336 Peridinium tricuspis O. Wetzel, 166, pi. 2, fig. 14. 1970 Astrocysta tricuspis (O. Wetzel) Davey, p. 360. 1973 Lejeunia tricuspis (O. Wetzel) Harland, p. 673, pi. 84, fig. 4. Figured material. Slides SAL 4380-RH1 and SAL 4381-RH1, specimens MPK 929, 930. Remarks. This species was observed to possess a large intercalary archeopyle (PI. 25, HARLAND: DI NOFL AGELLATES FROM MONTANA 189 fig. 3) and together with the small pericoels developed in the regions of the apical and antapical horns indicate its affinities to this genus and not to the genus Astrocysta Davey whose archeopyle is now known to be transapical (Norris and Hedlund 1972). S. tricuspis occurs almost throughout the studied section and indeed Sarjeant (1967tf) and Harker and Sarjeant (1975) have recorded a Santonian to Maastrichtian range for the species. Other species. The following species also belong to this genus: IS. albertii (Corradini, 1972) comb. nov. - Dejiandrea albertii Corradini, 1972, pp. 174-175, pi. 27, hgs. la , 6, 8; pi. 28, fig. 2. S. boloniensis (Riegel, 1974) comb. nov. D. boloniensis Riegel, 1974, pp. 354-356, pi. 1. figs. 6-10; text-figs. 3, 4. S. gaditanum (Riegel, 1974) comb. nov. = D. gaditana Riegel, 1974, pp. 356-357, pi. 2, figs. 8, 9; pi. 3, figs. 1-2. IS. kozlowskii (Gorka, 1963) comb. nov. = Lejeunia kozlowskii Gorka, 1963, p. 41, pi. 5, fig. 4. S. pannuceum (Stanley, 1965) comb. nov. - Z). pannucea Stanley, 1965, p. 220, pi. 22, figs. 1-4, 8-10. S. pentagonalis (Corradini, 1972) comb, nov. — D. pentagonalis Corradini, 1972, p. 175, pi. 28, fig. 3. IS. subquadratum (Corradini, 1972) comb. nov. = D. subquadra Corradini, 1972, pp. 175-176, pi. 28, fig. 1. GROUP UNCERTAIN Genus odontochitina Deflandre emend. Davey 1970 Type species. Odontochitina operculata (O. Wetzel) Deflandre and Cookson, 1955; O.D. Odontochitina operculata (O. Wetzel) Deflandre and Cookson, 1955 1933u Ceratium ( Euceratium ) operculatum O. Wetzel, p. 170, pi. 2, figs. 21, 22; text-fig. 2. 1955 Odontochitina operculata (O. Wetzel) Deflandre and Cookson, pp. 291-292, pi. 3, figs. 5, 6. Remarks. This species was found in the lowermost part of the section. Its previously recorded range was Hauterivian to Campanian (Sarjeant 1967n), but McIntyre (1974) recorded it from the ?Maastrichtian, as did Zaitzeff and Cross (1970) but as O. striatoperforata (see Williams 1974). Harker and Sarjeant (1975) give this species a range of early Valanginian to late Maastrichtian. COMPARISON AND INTERPRETATION The dinoflagellate cyst assemblage recovered from the Bearpaw Formation of Hell Creek, Montana contains a number of forms in common with the same formation in southern Alberta (Harland 1973). These are Cyclonephelium distinction Deflandre and Cookson, Oligosphaeridium anthophorum (Cookson and Eisenack) Davey, O. pulcherrimum (Deflandre and Cookson) Davey and Williams, Spiniferites ramosus (Ehrenberg) Mantell, AustralieUa cooksoniae (Alberti) Vozzhennikova = Deflandrea korojonensis Cookson and Eisenack of Harland (1973), A. tripartita (Cookson and Eisenack) Vozzhennikova, D. macrocysta Cookson and Eisenack, Diconodinium arcticum Manum and Cookson, D. firmum Harland, Senegalinium tricuspis (O. Wetzel) comb, nov., and Odontochitina operculata (O. Wetzel) Deflandre and Cookson. Species that are exclusive to the Montana section are Dictyopyxidia sp., Ceratiopsis diebeli (Alberti) Vozzhennikova, Deflandrea montanaensis sp. nov., D. cf. pirnaensis Alberti, Palaeocystodinium golzowense Alberti, Palaeoperidinium pyrophorum (Ehrenberg) Sarjeant, and S', magnificum (Stanley) comb. nov. The Montana section contains seventeen species of dinoflagellate cysts as compared to the fifty-three recorded from southern Alberta; eleven species are in common. In 190 PALAEONTOLOGY, VOLUME 20 the terms of the categories of relative proportions as given by Harland (1973), Oligosphaeridium anthophorum , A. cooksoniae, Ceratiopsis diebeli , D. macrocysta, D. montanaensis, Palaeocystodinium golzowense, and 5. tricuspis are ‘occasion- ally common’, i.e. making up greater than 10% of the dinoflagellate cyst population on occasions, with the remainder being ‘rare’. No species was ‘common’ throughout the Montana section. It is now possible to compare the described assemblage with those of Zaitzeff and Cross (1970), Jain and Millepied (1973), Riegel (1974), and McIntyre (1974), especi- ally with regard to Australiella , Diconodinium , and Senegalinium. It is also possible to comment that the present assemblages, together with Harland (1973) and the publications mentioned above, differ from those of Clarke and Verdier (1967) and Wilson (1971), especially with regard to the presence of species belonging to Dico- nodinium and Senegalinium. They are comparable, however, in the presence of Australiella species and of Odontochitina operculata. Some provincialism may be indicated or differences may be caused by local facies or palaeoenvironments. Certainly the northern United States and Canadian assemblages, of this age indicate a single water body during the Late Cretaceous. A major reason for the differences between the sections and assemblages from southern Alberta and Montana is age. The three species C. diebeli , P. golzowense, and S. magnificum are all much better known from the Maastrichtian than the Campanian, and on plotting the recovered species from the Montana section and including their known stratigraphical ranges in North America (see text-fig. 3) an apparent change in the assemblage occurs at about the level of sample SAL 4380 with no apparent change in the lithology. Here an assemblage with O. operculata, Oligosphaeridium pulcherrimum , and Cyclonephelium distinctum gives way to one containing Ceratiopsis diebeli, P. golzowense, and S. magnificum. Can this be con- sidered as the Campanian-Maastrichtian boundary? Unfortunately the full results of Wilson’s study on the European type Campanian and Maastrichtian stages, preliminarily reported upon in 1971, are not yet published. It would appear, however, that Odontochitina operculata has a top at the Campanian- Maastrichtian boundary or just within the earliest Maastrichtian, and that C. diebeli is commonly first found in the Maastrichtian (Wilson 1971). There is therefore some evidence for placing the Campanian-Maastrichtian boundary at the level of SAL 4380 and also evidence for regarding the whole section as being Maastrichtian in age. An age assignment of very latest Campanian to Maastrichtian or entirely Maastrichtian may, therefore, be given to the section at Hell Creek in Montana. Evidence from the radiometric data available and the radiometric time scale, as understood at present, appears to support the dinoflagellate biostratigraphy. How- ever, errors are inherent in the construction of such a time scale and the delimitation of absolute time for stage boundaries (see Obradovich and Cobban 1975); but it is interesting that there is some correspondence. The Bearpaw sea in Montana at this time was probably shallow and under a terrigenous influence, because there is a low proportion of dinoflagellate cysts in the total palynomorph content (only between 1-15% throughout the section), a low species diversity, and a high proportion of peridiniacean to gonyaulacacean cysts (see Harland 1973). This is in contrast to the southern Alberta sections where there HARLAND: DINOFLAGELLATES FROM MONTANA 191 were larger and more diverse populations of dinoflagellate cysts, probably reflecting better palaeoenvironmental conditions. Since the formation in Montana is younger than that in Alberta it is likely to be reflecting the growing influence of continental sedimentation. Acknowledgements. I thank Drs. G. Playford and G. D. Williams for access to their material, and especially to Dr. Playford for supplying additional stratigraphical information on the section and for his encourage- ment. The bulk of this work was done at the University of Alberta, Edmonton, whilst the author was in receipt of an Intersession Bursary. I also thank Dr. Roger J. Davey for his advice and comments on the manuscript and to my wife, Patricia, for all her help and encouragement. This paper is published with the approval of the Director, Institute of Geological Sciences. REFERENCES alberti, G. 1959. Zur Kenntnis der Gattung Deflandrea Eisenack (Dinoflag.) in der Kreide und im Alt- tertiar Nord- und Mitteldeutschlands. Mitt. Geol. Staatsinst. Hamburg , 28, 93-105. — 1961. Zur Kenntnis Mesozoischer und Alttertiaren Dinoflagellaten und Hystrichosphaerideen von Nord- und Mitteldeutschlands sowie eimgen Anderen Europaischen Gebieten. Palaeontographica , Abt. A, 116, 1-58. caldwell, w. g. e. 1968. The late Cretaceous Bearpaw Formation in the South Saskatchewan River valley. Sask. Res. Council , Geology Div. Rept. 8, 1-89. casey, r. 1964. The Cretaceous Period. Q. J! geol. Soc. Lond. 120S, 193-202. clarke, r. f. a. and verdier, j. p. 1967. An investigation of microplankton assemblages from the Chalk of the Isle of Wight, England. Verb. K. Ned. Akad. Wet. 24, I 96. cookson, i. c. and eisenack, a. 1958. Microplankton from Australian and New Guinea Upper Mesozoic sediments. Proc. Roy. Soc. Victoria , 70, 19-79. — 1960. Microplankton from Australian Cretaceous sediments. Micropaleontology, 6, 1-18. — 1962. Additional microplankton from Australian Cretaceous sediments. Ibid. 8, 485-507. — 1974. Mikroplankton aus Australischen Mesozoischen und Tertiaren Sedimenten. Palaeonto- graphica, Abt. B, 148, 44-93. corradini, d. 1972. Non-calcareous microplankton from the Upper Cretaceous of the northern Appen- nines. Bull. Soc. Palaeont. Ital. 11, 1 19-197. davey, r. j. 1969. Non-calcareous microplankton from the Cenomanian of England, northern France and North America. Part I. Bulk Br. Mas. nat. Hist. (Geol.), 17, 103-180. — 1970. Non-calcareous microplankton from the Cenomanian of England, northern France and North America. Part II. Ibid. 18, 333-397. — and williams, G. l. 1966. The genus Hystrichosphaeridium and its allies. In davey, r. j. et al. Studies on Mesozoic and Cainozoic dinoflagellate cysts. Ibid. Suppl. 3, 53-105. deflandre, g. and cookson, i. c. 1955. Fossil microplankton from Australian late Mesozoic and Tertiary sediments. Aust. J. Mar. Freshw. Res. 6, 242-313. ehrenberg, c. g. 1838. Uber das Massenverhaltnis der jetzt labenden Kieselinfusorien und liber ein neues Infusorien-Conglomerat als Polierschiefer von Jastraba in Ungarn. Akad. Wiss. Berlin, Abh. [1836], 1, 109-135. eisenack, a. 1961. Einige Erorterungen iiber fossile Dinoflagellaten nebst Ubersicht fiber die zur Zeit bekannten Gattungen. Neues Jb. Geol. Palaont., Abh. 1 12, 281-324. — and kjellstrom, G. 1971. Katalog der fossilen Dinoflagellaten, Hystrichosplidren und verwandten Mikrofossilien. Band. II. Dinoflagellaten. E. Schweizerbart’sche, Stuttgart, 1-1130. evitt, w. r. 1967. Dinoflagellate studies II. The archeopyle. Stanford, Univ. Publ., Geol. Sci. 10, No. 3, 1-82. felix, c. j. and burbridge, p. p. (in press). Age of microplankton studied by Manum and Cookson from Graham and Ellef Ringnes Islands. Geoscience and Man. folinsbee, r. e., baadsg aard , h. and lipson, j. 1960. Potassium-argon time-scale. Internal. Geol. Congress, XXI, Session Rept. 7-17. 192 PALAEONTOLOGY, VOLUME 20 folinsbee, r. e., baadsgaard, h. and lipson, j. 1961. Potassium-argon dates of Upper Cretaceous ash falls, Alberta, Canada. Ann. N.Y. Acad. Sci. 91, 352-359. GORKA, H. 1963. Coccolithophorides, Dinoflagelles, Hystrichosphaerides et microfossiles incertae sedis du Cretace Superieur de Pologne. Acta Pal. Polon. 8, 3-90. harker, s. d. and sarjeant, w. a. s. 1975. The stratigraphic distribution of organic-walled dinoflagellate cysts in the Cretaceous and Tertiary. Rev. Palaeobot. Palynol. 20, 217-315. harland, r. 1973. Dinoflagellate cysts and acritarchs from the Bearpaw Formation (Upper Campanian) of southern Alberta, Canada. Palaeontology, 16, 665-706. heisecke, A. M. 1970. Microplankton de la Formacion Roca de la Provincia de Neuquen. Ameghiniana, 7, 225-263. herngreen, g. f. w. 1975. Palynology of Middle and Upper Cretaceous strata in Brazil. Meded. Rijks. Geol. Dienst., N.S. 26, 39-91. jain, k. p. and millepied, p. 1973. Cretaceous microplankton from Senegal Basin, N.W. Africa. 1. Some new genera, species and combinations of dinoflagellates. Palaeobotanist . 20, 22-32. kjellstrom, G. 1972. Archaeopyle formation in the genus Lejeunia Gerlach, 1961 emend. Geol. For. Stockh. Fork. 94, 467-469. lambert, r. st. J. 1971 The pre-Pleistocene Phanerozoic time scale— a review. In Part I of The Phanerozoic Time Scale — a supplement. Spec. Publ. Geol. Soc. Lond. 5, 9-31. loeblich, a. r., Jun. and loeblich, a. r. III. 1966. Index to the genera, subgenera and sections of the Pyrrhophyta. Stud. trop. Oceanogr. Miami, 3, 1-94. malloy, r. e. 1972. An Upper Cretaceous dinoflagellate cyst lineage from Gabon, West Africa. Geoscience and Man, 4, 57-65. mantell, G. a. 1854. The Medals of Creation', or First Lessons in Geology and the study of Organic Remains. 2nd edn., Bohn, London, 1-930. manum, s. and cookson, i. c. 1964. Cretaceous microplankton in a sample from Graham Island, Arctic Canada, collected during the second ‘Fram’ Expedition (1898-1902) with notes on microplankton from the Hassel Formation, Ellef Ringnes Island. Skr. Norska Vid-Akad. Oslo, Mat.-Naturv. kl. (n.s.), 17, 1-36. mcintyre, d. j. 1974. Palynology of an Upper Cretaceous section, Horton River, District of Mackenzie, N.W.T. Geol. Surv. Can. Pap. 74-14, 1-57. — 1975. Morphologic changes in Deflandrea from a Campanian section. District of Mackenzie, N.W.T. , Canada. Geoscience and Man, 11, 61-76. neves, r. and dale, b. 1963. A modified filtration system for palynological preparations. Nature, 198, 775-776. norris, G. and hedlund, R. w. 1972. Transapical sutures in dinoflagellate cysts. Geoscience and Man, 4, 49-56. Norton, n. j. and hall, j. w. 1969. Palynology of the Upper Cretaceous and Lower Tertiary in the type locality of the Hell Creek Formation, Montana, U.S.A. Palaeontographica, Abt. B, 125, 1-64. obradovich, j. d. and cobban, w. a. 1975. A time-scale for the Late Cretaceous of the western interior of North America. Geol. Assoc. Can., Spec. Paper, 13, 31-54. riegel, w. 1974. New forms of organic-walled microplankton from an Upper Cretaceous assemblage in southern Spain. Rev. Espanola de Micropal. 6, 347-366. sarjeant, w. a. s. 1967a. The stratigraphical distribution of fossil dinoflagellates. Rev. Palaeobot. Palvnol. 1, 323-343. — 19676. The genus Palaeoperidinium Deflandre (Dinophyceae). Grana palynol. 7, 243-258. — and downie, c. 1974. The classification of dinoflagellate cysts above generic level: a discussion and revisions. Birbal Salmi Institute of Palaeobotany, Spec. Pub. 3, 9-32. Stanley, e. a. 1965. Upper Cretaceous and Paleocene plant microfossils and Paleocene dinoflagellates and hystrichosphaerids from northwestern South Dakota. Bull. Am. Paleont. 49, 179-384. stelck, c. R. 1967. The record of the rocks. In hardy, w. g. (ed.-in-chief). Alberta A Natural History. Hurtig, Edmonton, 21-51. vozzhennikova, r. f. 1963. Pirrofitovye Vodorosli. In orlov, y. a. (ed.). Osnovy Paleontologii , 14, 179-185. — 1967. Iskopaemye peridinei yurskikh, melovykh i paleogenovyh otlozhency SSSR. Akad. Nauk. SSSR. Sibirskoe Otledeinie, Inst. Geol. Geofiz. 1-347. HARLAND: DINOFLAGELLATES FROM MONTANA 193 wall, D. and dale, D. 1968. Modern dinoflagellate cysts and evolution of the Peridiniales. Micro] paleonto- logy, 14, 265-304. wetzel, o. 1933a. Die in organischer Substanz erhaltenen Mikrofossilien des baltischen Kreide- Feuersteins mit einem sediment-petrographischen und stratigraphischen Anhang. Palaeontographica . Abt. A, 77, 141-188. 19336. Die in organischer Substanz erhaltenen Mikrofossilien des baltischen Kreide-Feuersteins mit einem sediment-petrographischen und stratigraphischen Anhang. Ibid. 78, 1-110. williams, G. l. 1974. Dinoflagellate and spore stratigraphy of the Mesozoic-Cenozoic, offshore Eastern Canada. In Offshore Geology of Eastern Canada. Geol. Surv. Can. Pap. 74-30, 107-161. — ■ and downie, c. 1966. Further dinoflagellate cysts from the London Clay. In davey, r. j. et al. Studies on Mesozoic and Cainozoic dinoflagellate cysts. Bull. Br. Mus. nat. Hist. (Geol ). Supplement. 3, 215-236. wilson, g. J. 1971. Observations on European Late Cretaceous dinoflagellate cysts. Proc. II Planktonic Conference. Rome 1970, 1259 1275. zaitzeff, j. b. and cross, a. t. 1970. The use of dinoflagellates and acritarchs for zonation and correlation of the Navarro Group (Maestrichtian) of Texas. In kosanke, r. m. and cross, a. t. (eds.). Symposium on Palynology of the Late Cretaceous and Early Tertiary. Geol. Soc. Am. Spec. Pap. 127, 341-377. REX HARLAND Original typescript received 22 October 1975 Revised typescript received 22 January 1976 Institute of Geological Sciences Ring Road, Halton Leeds LSI 5 8TQ N A NEW EOCENE SHARK FROM THE LONDON CLAY OF ESSEX by h. cappetta and d. j. ward Abstract. Teeth of a new scyliorhinid shark Megascyliorhinus cooperi gen. nov., sp. nov. from the London Clay are described and compared with M. miocaenicus (Antunes and Jonet) a Mio-Phocene species from France, Portugal, and Tunisia, previously referred to Rhincodon. It is suggested that the genus Megascyliorhinus inhabited fairly deep water. Recent collecting at the London Clay (Eocene) locality of Burnham-on-Crouch, by the junior author (D. J. W.), has produced teeth of a new species of scylio- rhinid shark. Independent collecting by the senior author (H. C.) at the Neogene localities of La Motte d’Aigues, France and Nabeul, Tunisia has yielded represen- tatives of the same genus, a species previously only recorded from the Miocene of Portugal. LOCALITIES 1. Burnham-on-Crouch, Essex. This locality is situated about 3 km WNW. of the town of Burnham-on-Crouch. The London Clay is exposed on the north bank of the River Crouch in a foreshore section from map references TQ 920968 to TQ 922 966. The formation is a stiff blue-grey clay, weathering to brown, and containing occa- sional mudstones and septaria. From Whitaker and Thresh (1916, pp. 86 111), it is estimated that the base of the clay is 85 m below high-water mark. The invertebrates are characteristic of Wrigley’s divisions 3, 4, and 5 of the London Clay (Wrigley 1924 and 1940). The selachian fauna is identical to that listed from Sheppey (Casier 1966) except for the absence of a few species: Heterodontus xvardenensis Casier, Myliobatis latidens Woodward, and Aetobatis irregularis Agassiz. There are in addition several species present at Burnham which are so far unknown from Sheppey: Triakis sp., Mustelus sp., and Scyliorhinus spp. 2. La Motte d’Aigues, France. At this locality, situated at the foot of the south face of Grand Luberon (Vaucluse), there are three superimposed formations each of which yields a rich selachian fauna. The lower is a marl, the middle a sandstone, and the upper a limestone; the latter yielded a few teeth of Megascyliorhinus miocaenicus. These formations are considered as Helvetian s.l. in age, probably Serravallian. 3. Nabeul, Tunisia. This locality, about 70 km south-east of Tunis, is a large pit where the blue marls have long been worked by local potters. The selachian teeth come from gypsiferous marls interbedded with the blue marls at the base of the outcrop. These marls are considered to be Early Pliocene in age. [Palaeontology, Vol. 20, Part 1, 1977, pp. 195-202, pis. 26-27.] 196 PALAEONTOLOGY, VOLUME 20 SYSTEMATIC PALAEONTOLOGY Family scyliorhinidae Gill, 1862 This family is represented in the London Clay by seven species. The teeth of ‘ Scylio - rhinus biauriculatus Casier and ‘ST minutissimus (Winkler) are large, and their inclusion in Scyliorhinus s.s. is rather doubtful; they somewhat resemble the car- charhinids. S. gilberti Casier and three small undescribed species of Scyliorhinus closely resemble Recent Scyliorhinus teeth (Cappetta 1976). The new species described below is the largest, the anterior teeth reaching 7-5 mm in height. Megascyliorhimis gen. nov. Type species. M. cooperi sp. nov. Referred species. M. miocaenicus (Antunes and Jonet). Diagnosis. Scyliorhinid known only by isolated teeth; anterior teeth large, curving strongly inwards, lingual face of the crown sometimes striated, labial face particularly convex, smooth or finely striated. The cutting edge, where present, generally does not reach the base of the crown. Root high and broad with flat base and deep median groove; well-developed pair of latero-internal foramina. The lateral teeth asym- metrical, very strongly striated, and may have a pair of lateral denticles. Megascyliorhimis cooperi gen. nov., sp. nov. Plate 26, figs. 1-3; Plate 27, fig. 1 Derivation of name. This species is dedicated to Mr. J. Cooper in recognition of his work on the London Clay. Material. Six complete and three fragmentary teeth from Burnham-on-Crouch. Three worn teeth from Sheppey, previously unnoticed among some upper anterior hexanchid teeth in the collection of the B.M. (N.H.). Figured specimens are housed in the B.M. (N.H.), whose registration numbers are quoted. Holotype. Anterior tooth, P. 57621. Plate 26, fig. 1 a-d. Type locality. Burnham-on-Crouch, Essex, England. Age. Eocene; Ypresian (London Clay). Diagnosis. The labial surface of the anterior tooth is inflated towards the bottom of the crown and has short strong parallel striae at its base. The anterior lobes of the root are rounded and outwardly divergent. The root markedly protrudes over the base of the crown in oral view. Extreme lateral teeth have a partially closed root groove. EXPLANATION OF PLATE 26 Figs. 1-3. Megascyliorhimis cooperi gen. nov., sp. nov. x 10. Eocene. Burnham-on-Crouch. In each figure, a. lingual face; b. labial face; c, profile; d , base. The registration numbers are of the Department of Palaeontology, British Museum (Natural History). 1, anterior tooth. P. 57621. Holotype. 2, anterior tooth. P. 57624. 3, lower lateral tooth. P. 57622. PLATE 26 CAPPETTA and WARD, new Eocene shark 198 PALAEONTOLOGY, VOLUME 20 Description. The holotype (PI. 26, fig. 1) is an anterior tooth with a pointed crown and rather stout base. The crown strongly curves towards the inside of the mouth and is oblique to the root. The inner surface is strongly convex and is ornamented with numerous fine parallel close-spaced striations on its lower quarter. The outer surface is less convex but has a basal swelling protruding over the anterior surface of the root; the area below this swelling is concave and bears short vertical striations, some of which are fairly well spaced. The junction of the root and crown is constricted. The root is particularly broad and high; the base is flat and the lobes are separated by a deep groove. The internal protuberance is well developed and separated from the general contour by two notches on the prolongation of the latero-internal foramina. These are well developed and occur singly or paired on the lateral face. The inner surface is rather high; the outer surface is very oblique in profile, follow- ing the spread of the root towards the anterior lobes, which are rounded and separated by a deep groove. When seen in profile, the enamel junction is straight and slopes up posteriorly. The cutting edge is restricted to the upper third of the crown. On a slightly larger anterior tooth, the crown is a little less slanting and the base protrudes less over the anterior surface of the root; the cutting edge occupies half the length of the crown. The lateral teeth are asymmetrical, their crowns being inclined towards the sym- physis. On one tooth (PI. 26, fig. 3), the striations on the outer surface are restricted to the base of the crown, whereas those on the inner surface cover one-half of its height. On the other hand, another tooth (PI. 27, fig. 1), of similar size and presumed jaw position, shows striations up to two-thirds of the way up the crown, and in addition bears a single well-developed lateral denticle, the other represented by a small protuberance. By comparison with the modern scyliorhinids, this more striated tooth probably belongs to the upper dentition. On both these teeth, the roots are asymmetrical with a more developed lingual lobe; the root groove is oblique, in line with the inclination of the crown. One very small lateral tooth shows the same characteristics, with less striation, although more strongly defined and covering almost the whole crown; the root groove is partly covered over. Megascyliorhinus miocaenicus (Antunes and Jonet), 1970 Plate 27, figs. 2-4 1970 Rhincodon miocaenicus , Antunes and Jonet, pp. 152-153, fig. 5; pi. 9, figs. 42-44. Material. Ten teeth from the Helvetian of La Motte d'Aigues (France) and one tooth from the Pliocene of Nabeul (Tunisia). EXPLANATION OF PLATE 27 In each figure, a, lingual face; b, labial face; c, profile; d, base. Fig. 1. Megascyliorhinus cooperi gen. nov., sp. nov. x 10. Upper lateral tooth. B.M. (N.H.) P. 57623. Eocene. Burnham-on-Crouch. Figs. 2-4. Megascyliorhinus miocaenicus (Antunes and Jonet). x 4. Stereophotographs. The registration numbers are those of the Laboratoire de Paleontologie, Universite Montpellier II. 2, anterior tooth. LMA. 1. Helvetian. La Motte d’Aigues. 3, anterior tooth. NAB. 1. Late Pliocene. Nabeul. 4, lateral tooth. LMA. 2. Helvetian. La Motte d’Aigues. PLATE 27 CAPPETTA and WARD, new Eocene shark 200 PALAEONTOLOGY, VOLUME 20 Description, (a) Miocene specimens. An anterior tooth (PI. 27, fig. 2 a-c) has a stout crown and no cutting edge. The labial surface of the crown has a few strong stria- tions below its inflated base. The root is damaged, so that only an indication of its size can be given. A lateral tooth (PI. 27, fig. 4 a c), lacking a fragment of root, is asymmetrical, its crown being inclined towards the commissure. The lingual surface is very convex; the labial is less so, tending to flatten at the base of the crown. The lingual surface has numerous small wavy striations along its entire height; those of the labial surface are less in number, short in the middle, and longer at the sides. On one side there is a sharp, striated lateral denticle well separated from the main crown; the other is broken. The root is rather high with a strong internal protuberance; its base and outer face are flat, the base is divided by a deep groove. This lateral tooth is rather different from the anterior teeth, but its general appear- ance, and in particular its striation and root morphology, favour its inclusion in this species; the presence of denticles on lateral teeth has already been noted in M. cooperi. For a more detailed description of the Miocene species, see Antunes and Jonet (1970). ( b ) The Pliocene specimen. This anterior tooth (PI. 27, fig. 3) is slightly smaller than the Miocene specimens but larger than those from the Ypresian. The crown is stout, and leans less towards the inside of the mouth. There is a single lateral denticle, broad at its base but with a sharp backward-pointing tip. The lower third of the inner surface is striated; the outer surface is smooth. The root is robust, rather high, and has a deep groove at its base. SYSTEMATIC COMPARISON One can immediately see the similarities between the Eocene and Mio-Pliocene species, and there can be little doubt that they are directly related. Their appearance is similar, but there are sufficient small differences to separate them. The teeth of the Eocene species have a longer crown, which is more inclined towards the inside of the mouth, with many well-defined striations at the base of the outer surface. The root is lower and slightly different in shape: the front is rather oblique and the lobes are quite large. This Ypresian species is quite easy to separate from the other Eocene scyliorhinids, in particular by its combination of large size, striations, and the absence of lateral denticles on the anterior teeth. The teeth of M. miocaenicus are stouter. Their crowns lean less towards the inside of the mouth, the striations at the base of the outer surface are less developed, the root is higher, and the outer surface is subvertical. The Neogene species cannot be mistaken for any other scyliorhinid. Only the lateral teeth approach those of Scyliorhinus distans (Probst), and they can be easily separated by their stouter crown, finer and more numerous striations, the shape of the root, and the more jagged outline of crown and denticles. The Mio-Pliocene teeth agree with those from the Tortonian of Portugal figured by Antunes and Jonet (1970, p. 153, fig. 5 and pi. 9, figs. 42-44) under the name Rhincodon miocaenicus; they are anterior teeth without lateral denticles. Antunes and Jonet place in the CAPPETTA AND WARD: NEW EOCENE SHARK 201 synonymy of R. miocaenicus the tooth figured in Cappetta (1970) under the name Rhincodon sp. from the Helvetian of Loupian, Languedoc, France. Antunes and Jonet state: les caracteres morphologiques de la racine de ces dents (i.e. R. mio- caenicus) semblent identiques a ceux de la dent decrite par Cappetta (i.e. Rhincodon sp.) mais il y a quelques differences en ce qui concerne la couronne.1 However, these differences are extremely important, particularly in the region of the root. We have examined the teeth of a specimen of the modern whale shark, R. typus Smith, from the Seychelles: these teeth are characterized by their small size, by a smooth erect crown with a distinct cutting edge, and by a wide rounded enamel lingual apron, which in profile reaches the base of the root. The root is high, elongated labio-lingually, with a large internal protuberance and a deep groove; the sides are very swollen and the latero-internal foramina are particularly close to the lingual surface. The Loupian tooth corresponds completely with the modern teeth and certainly represents a true Rhincodon and not an upper tooth of M. mio- caenicus as supposed by Antunes and Jonet (1970, p. 154). These authors were probably misled by the figure of a tooth of R. typus in Bigelow and Schroeder (1948, p. 190). This illustration is inaccurate because it shows a tooth without its lingual apron, an important feature present in all the teeth we have been able to study, but absent in M. miocaenicus. In 1930 White had already, correctly, figured a Rhincodon tooth (p. 143, fig. 9). As the species miocaenicus does not belong to the genus Rhin- codon, but instead closely resembles the species M. cooperi , it is therefore placed in that new genus. Megascyliorhinus is included in the Scyliorhinidae because of its broad similarity with other members of the family, particularly Scyliorhinus s.s.; nevertheless, this genus differs from all the genera of the family that we have examined. STRATIGRAPHIC DISTRIBUTION Megascyliorhinus first appears in Divisions 3-5 of the London Clay (Ypresian) at Burnham-on-Crouch and Sheppey, England. The tooth described by Brzobohaty and Kalabis (1970, pi. 2, fig. 2a- b) from the Oligocene of Pouzdrany, Czechoslovakia, and referred by those authors to Odontaspisl sp., is very suggestive of this genus. This anterior tooth shows all the characteristics of the genus. Its crown is typically stocky and inclined, resembling the Eocene species more closely than the Mio- Pliocene one. It can be distinguished from those of the Eocene species by the less prominent anterior lobes of the root and by the presence of a pair of lateral denticles; this feature occasionally appears on anterior teeth, as seen in the Pliocene specimen. The genus is not known again until the Helvetian sd. of La Motte d’Aigues, France where it is represented by M. miocaenicus , and again in the Tortonian VII at Mutela in Portugal. The most recent specimen is from the Early Pliocene of Nabeul in Tunisia. This genus probably became extinct in the Pliocene or Pleistocene. PALAEOECOLOGY The base of Wrigley’s London Clay Division 2 and the base of the Aldwick Beds at Bognor Regis, Sussex mark the influx of a number of shark species which make their first appearance in the English Eocene. These include a few species rare or absent 202 PALAEONTOLOGY, VOLUME 20 in contemporary shallow-water deposits in the Paris-Belgian Basin, for example Notorhynchus serratissimus (Agassiz), Xenodolamia eocaena (Woodward), and Anomotodon sheppeyensis (Casier) [ = Oxyrhina sheppeyensis Casier], all presumed to be deep-water species. In the Miocene, Megascyliorhinus is associated with Deania , Centrophorus , Pristiophorus , and Raja. Living species of these genera, with perhaps the exception of Raja , prefer deep water. At La Motte d’Aigues, where there is a series of closely similar faunas in the Helvetian, Megascyliorhinus is restricted to limestone facies at the top, and is absent from the lower marly and sandy facies. In Tunisia, the sole tooth was collected from gypsiferous marls at the base of blue marls rich in otoliths of Myctophidae, which are deep-water fishes. The nature of these deposits and their faunal associations suggest that Megascyliorhinus inhabited depths of around 150-200 m. Acknowledgements. We would like to thank Dr. C. Patterson, Messrs. J. Cooper and J. Hooker of the B.M. (N.H.) for reading and criticizing the manuscript. The photographs of M. cooperi were by Mr. P. York and those of M. miocaenicus by Monsieur J. Martin. Mrs. J. Powell assisted in the translation of the French draft and the English manuscript was typed by Mrs. D. Ward. REFERENCES antunes, m. t. and jonet, s. 1970. Requins de LHelvetien superieur et du Tortonien de Lisbonne. Revta Fac. Cienc. Univ. Lisb. 1, 119-280. bigelow, h. and schroeder, w. c. 1948. Fishes of the Western North Atlantic. Mem. Sears Fdn mar. Res. 1, 59-576. brzobohaty, r. and KALABis, v. 1970. Die Fischzahne aus Pouzdrany (Pouzdranyschichten, Oligozan). Cas morav. Mus. Brno , 55, 41-47. cappetta, h. 1970. Les Selaciens du Miocene de la region de Montpellier. Palaeovertebrata , Mem. Extr., 1-139. - 1975. Les Selaciens miocenes du Midi de la France. Repartitions stratigraphique et bathymetrique. 3e Reunion Sci. de la Terre , Montpellier, p. 90. - 1976. Selaciens nouveaux du London Clay de l’Essex (Ypresien du Bassin de Londres). Geobios , 9(5). casier, E. 1966. Faune ichthyologique du London Clay. British Museum (Natural History), London. 496 pp. whitaker, w. and thresh, j. c. 1916. The water supply of Essex. Mem. geol. Surv. U.K. 510 pp. white, e. G. 1930. The whale shark, Rhineodon typus. Description of the skeletal parts and classification based on the Marathon specimen captured in 1923. Bull. Am. Mus. nat. Flist. 61, 129-160. wrigley, a. 1924. Faunal divisions of the London Clay. Proc. Geol. Ass. 35, 245-259. - 1940. The faunal succession in the London Clay illustrated in some new exposures near London. Proc. Geol. Ass. 51, 230-245. H. CAPPETTA Laboratoire de Paleontologie E R A. No. 261 Evolution des Vertebres Universite des Sciences et Techniques du Languedoc Montpellier Cedex 34060 France D. J. WARD 35 Addington Road West Wickham Kent Typescript received 31 October 1975 England THE GIANT CROCODILIAN SARCOSUCHUS IN THE EARLY CRETACEOUS OF BRAZIL AND NIGER by e. buffetaut and P. TAQUET Abstract. Re-examination of crocodilian remains previously referred to Goniopholis , from the lower Cretaceous of the Bahia basin (Brazil), shows they really belong to the genus Sarcosuchus (Mesosuchia, Pholidosauridae), formerly known from the Aptian of the Sahara. Resemblances between early Cretaceous African and South American vertebrate faunas anterior to the opening of the South Atlantic Ocean are stressed. In 1966 de Broin and Taquet announced the discovery of a new giant crocodilian, Sarcosuchus imperator (suborder Mesosuchia, family Pholidosauridae), in early Cretaceous continental strata of Algeria and Niger. Since then, abundant material has been collected at the Gadoufaoua locality, in the Tegama basin of Niger (Taquet 1970); it includes complete skulls and skeletons, and provides much new information on the anatomy of this African form. On the other hand, a re-examination, by E. Buf- fetaut, of crocodilian remains found at the turn of the century in the early Cretaceous Bahia series of the Reconcavo (or Bahia) basin, on the north-eastern coast of Brazil, shows that the successive attributions given to these fossils by various authors are erroneous. A comparison between the specimens from Brazil and those from Niger leads us to attribute the remains from Bahia to the genus Sarcosuchus , and to stress the similarities between the early Cretaceous African and South American continental faunas. As early as 1860, vertebrate remains from the Reconcavo basin, among which were two different types of crocodilian teeth, were described by Allport. In 1869 Marsh erected two new species for a few isolated teeth found by Hartt in the same area : Crocodilus hartti , with large, finely wrinkled teeth, and Thoracosaurus bcihiensis , with smaller, coarsely striated teeth. Later, Woodward (1888) referred other scanty remains to Hyposaurus der bianus , a species described by Cope (1886) from the Late Cretaceous of Pernambuco. More important material, discovered by Mawson, was also studied by Woodward (Mawson and Woodward 1907), who then attributed all crocodilian remains from the Bahia formation to the genus Goniopholis , with two species, G. hartti (Marsh, 1869)— which included all the newly discovered fossils— and G. bcihiensis { Marsh, 1869). The remains of 'Goniopholis hartti' , now in the British Museum (Natural History), which have been subjected to a re-examination, comprise the anterior part of a large lower jaw (R 3423), a dorsal scute (R 3224), and two teeth (R 2983, R 3079). Similarities between these fossils and S. imperator are numerous (PI. 28). In both instances, the very long mandibular symphysis is indicative of a long-snouted animal (PI. 28, figs. 1, 2), while in the short-snouted genus Goniopholis the symphysis is always short. Other remarkable resemblances between the Brazilian form and [Palaeontology, Vol. 20, Part 1, 1977, pp. 203-208, pi. 28.] 204 PALAEONTOLOGY, VOLUME 20 S. imperator include the spatulate shape of the anterior extremity of the mandible, the relative size of the alveoli (with small first and second, and enlarged third and fourth alveoli), and the coarse ornamentation of the ventral surface of the bone. The scute found in Brazil (PI. 28, fig. 5) is very similar to some dorsal plates of S. imperator (PI. 28, fig. 4), which have the same general outline, likewise bear an antero-external ‘peg’, and possess a smooth anterior rim (which was overlapped by the preceding scute). The presence of a peg on the dorsal scutes is by no means restricted to Gonio- pholis, but occurs in several other crocodilian genera, such as the Jurassic teleosaurid Steneosaurus , as already mentioned by Woodward in a footnote to the 1907 article. The large teeth of the animal from Bahia are also very reminiscent of Sarcosuchus imperator , in their stout general shape and in the ornamentation of their enamel, consisting of fine sinuous wrinkles. Lastly, both the African and the South American forms are very large. S. imperator is one of the largest known crocodilians, with a skull up to 170 cm in length, and an estimated over-all length of 11 m. The jaw fragment from the Reconcavo basin is 43 cm long, although it is only the anterior part of the symphysis. Thus, Woodward’s attribution of the Brazilian fossils to Goniopholis, which was based mainly on the presence of a peg on the dorsal scute, cannot be considered as valid. In our opinion, this South American crocodilian belongs to the genus Sarco- suchus. In the absence of more complete material from South America, it seems acceptable to retain different specific names for the African and the South American forms, and to call the latter S'. hartti( Marsh, 1869). However, the differences between the two forms (in the spacing of the alveoli, for instance) seem rather unimportant, so that the possibility that they may belong to the same species cannot be excluded. There may well be crocodilians other than S. hartii in the Bahia series, but they are represented by very scanty remains, which do not allow a precise identification. Although Roxo (1936) tried to revalidate T. bahiensis on the basis of a tooth and a procoelous vertebra of uncertain origin, the occurrence of the essentially marine late Cretaceous genus Thoracosaurus in the early Cretaceous freshwater Bahia series is most unlikely, as already pointed out by Antunes (1964). There is also no serious reason to assume that H. derbianus occurs in the Early Cretaceous of the Reconcavo basin. Both identifications were made in the erroneous belief that the Bahia series was of late Cretaceous age. Besides crocodilian remains, the fossil material from the Reconcavo basin includes the centrum of a dorsal vertebra of a carnosaur (Allport 1860), too incomplete to be compared with the vertebrae of the several carnosaurs known from Niger. Wood- ward (Mawson and Woodward 1907) also mentions vertebral centra which ‘seem EXPLANATION OF PLATE 28 Sarcosuchus remains from Brazil and Niger, x0-25. Fig. 1. Sarcosuchus hartti, from Brazil, B.M.(N.H.) Specimen No. R 3423; anterior extremity of lower jaw (from Mawson and Woodward 1907); n, in dorsal view; b , in ventral view. Figs. 2-4. Sarcosuchus imperator , from Niger. 2, anterior extremity of lower jaw of young individual; a , in ventral view; b, in dorsal view. The very spatulate shape is due to the youth of the animal. 3, anterior extremity of right dentary of adult specimen, in dorsal view. 4, dorsal scute. Fig. 5. Sarcosuchus hartti , from Brazil: dorsal scute (from Mawson and Woodward 1907). PLATE 28 BUFFETAUT and TAQUET, Giant crocodilian from Brazil and Niger 206 PALAEONTOLOGY, VOLUME 20 to agree closely with the corresponding bones of Iguanodonts’; this information, if true, is interesting, since several iguanodontids are known from the Early Creta- ceous of Niger (Taquet 1970). The vertebrate fauna from Bahia also includes scales of the holostean fish Lepidotus (with three species) and remains of two species of the coelacanth Mawsonia (Patterson 1975). The genus Lepidotus is also represented in the Tegama basin of Niger by numerous skull remains and large scales (now being studied), while a skull and several post-cranial fragments of a Mawsonia from the same area will soon be described by Wenz. Patterson (1975) has stressed the resem- blances between the early Cretaceous fish faunas from the Reconcavo and Gabon basins. We can now parallelize the freshwater faunas from these regions with the one from the Tegama basin (Table 1), and include Sarcosuchus in the list of early Cretaceous genera common to South America and to Africa. table 1 . Comparative list of early Cretaceous vertebrate genera common to South America and Africa. Reconcavo basin (Bahia series) Mawsonia major M. minor Lepidotus mawsoni L. souzai L. roxoi Sarcosuchus hartti Gabon basin Tegama basin (Cocobeach series) (Aptian) Coelacanth Mawsonia (new species) Lepidotus sp. Lepidotus sp. Sarcosuchus imperator The Reconcavo basin is a ‘semi-graben’ (Fonseca 1966) which was filled, during the early Cretaceous, by the several thousand metres of freshwater sediments of the Bahia series, consisting mainly in conglomerates, shales, and sandstones. Both the lithology and the fauna of this so-called ‘Gondwana Wealden’ (Krommelbein 1966) are indicative of a ‘shallow lacustrine environment’ (Fonseca 1966). As pointed out by Beurlen (1961), these sediments can hardly have accumulated in geographical conditions similar to the present ones, i.e. a basin open on the Atlantic Ocean. The eastern part of the graben is to be found on the West coast of Africa (text-fig. 1), text-fig. I Map showing the three basins men- tioned in the text: 1, Reconcavo (or Bahia) basin, Brazil. 2, Gabon basin, Gabon. 3, Tegama basin, Niger. (After Martin (1968), modified.) BUFFETAUT AND TAQUET : CROCODILIAN FROM BRAZIL AND NIGER 207 where the Gabon basin exhibits a succession of freshwater deposits (the Cocobeach series) very similar to the Bahia series (de Klasz 1965). The study of ostracod faunas from both areas shows that, according to Krommelbein (1966), it seems that these ostracods lived in one single basin with completely free interchange’; this has allowed a detailed stratigraphical correspondence to be established (Krommelbein 1971) . The end of lacustrine sedimentation in the Reconcavo-Gabon graben is indicated by salt deposits of Aptian age (Reyment 1973). According to Mawson’s (Mawson and Woodward 1907) and Fonseca’s (1966) maps, the remains of S. hartti were collected in the Ilhas Formation, which belongs to the upper part of the Bahia series; they are thus probably a little older than the Aptian. S. imperator , from the Tegama basin, is of Aptian age. Therefore, the chronological gap between the two species is probably not very great. The formation of the Reconcavo-Gabon graben, and of others of the same kind (Potiguar, Sergipe-Alagoas) at the beginning of the Cretaceous, is considered as an early stage in the separation between Africa and South America (Reyment and Tait 1972) . The presence of the non-marine crocodilian Sarcosuchus in both Brazil and Niger affords additional evidence of the faunal continuity between these continents during the early Cretaceous, before the complete opening of the South Atlantic Ocean. Acknowledgements. We thank Dr. A. J. Charig, of the British Museum (Natural History), for permission to examine the crocodilian remains from Bahia. The fossils from Niger were collected with the aid of a grant from the National Geographic Society. REFERENCES allport, s. 1860. On the discovery of some fossil remains near Bahia in South America. Q. Jl geol. Soc. London , 16, 263-268, 3 figs., 4 pis. antunes, M. T. 1964. Les Tomistoma (reptiles) et leur evolution. Inst. ‘ Lucas Mallada, Cursillos y Con- fer encias, 9,171 173. beurlen, k. 1961. Die palaogeographische Entwicklung des siidatlantischen Ozeans. Nova Acta Leo- poldina, 24 (154), 1-36, 8 figs. broin, f. de and taquet, p. 1966. Decouverte d’un Crocodilien nouveau dans le Cretace inferieur du Sahara. C.r. hebd. Seanc. Acad. Sci. Paris, 262 (D), 2326-2329, 1 fig. cope, e. d. 1886. A contribution to the vertebrate paleontology of Brazil. Proc. Am. pint. Soc. 23, 1-21 fonseca, J. t. 1966. Geological outline of the lower Cretaceous Bahia supergroup, Brazil. Proceedings of the 2nd West African Micropaleontological Colloquium (Ibadan 1965). Leiden. Pp. 49-71, 1 1 figs. klasz, i. de. 1965. Biostratigraphie du bassin gabonais. Mem. Bur. Rech. Geol. & Minieres ( Fr .), 32, 277- 303, 2 figs. krommelbein, k. 1966. On ‘Gondwana Wealden’ Ostracoda from NE Brazil and West Africa. Proceedings of the 2nd West African Micropaleontological Colloquium (Ibadan 1965). Leiden. Pp. 113-118, 1 table. — 1971. Non-marine Cretaceous Ostracods and their importance for the hypothesis of ‘Gondwanaland’. Second Gondwana Symposium, South Africa 1970, Proceedings and Papers. Pretoria. Pp. 617-619, 1 table. marsh, o. c. 1869. Notice of some new reptilian remains from the Cretaceous of Brazil. Am. J. Sci. 47, 390-392. martin, h. 1968. A critical review of the evidence for a former direct connection of South America with Africa. Biogeography and ecology in South America (fittkau, e. j. et al ., ed.). The Hague, 1, 25-53, 6 figs. mawson, J. and woodward, a. s. 1907. On the Cretaceous formation of Bahia (Brazil) and on vertebrate fossils collected therein. Q. Jl geol. Soc. London, 63, 128-139, 3 pis. 208 PALAEONTOLOGY, VOLUME 20 Patterson, c. 1975. The distribution of Mesozoic freshwater fishes. Mem. Mus. natn Hist, nat., Paris, 88, 156-174, 7 figs. reyment, r. a. 1973. Cretaceous history of the South Atlantic Ocean. Implications of Continental Drift to the Earth Sciences (tarling, d. h. and runcorn, s. k., eds.). London and New York, 2, 805-814. — and tait, e. a. 1972. Biostratigraphical dating of the early history of the South Atlantic Ocean. Phil. Trans. Roy. Soc. (B). 264, 55-95, 29 figs. roxo, a. o. 1936. Revalidaqao do Thoracosaurus Bahiaensis e consideragoes sobre a edade da serie da Bahia. Bohn Mus. nac. Rio de J. 12, 59-72, 1 pi., 1 table. taquet, p. 1970. Sur le gisement de Dinosauriens et de Crocodiliens de Gadoufaoua (Republique du Niger). C.r. hebd. Seanc. Acad. Sci., Paris, 271 (D), 38-40. woodward, a. s. 1888. Notes on some vertebrate fossils from the province of Bahia, Brazil, collected by Joseph Mawson, Esq., F.G.S. Ann. Mag. nat. Hist. 2, 132-136. E. BUFFETAUT Laboratoire de Paleontologie des Vertebres et de Paleontologie humaine Universite Paris VI 4 place Jussieu 75230 Paris Cedex 05 France p. TAQUET Institut de Paleontologie 8 rue de Buffon 75005 Paris Typescript received 19 December 1975 France NEW CARBONIFEROUS STENOSCI SM ATACE AN BRACHIOPODS FROM OVIEDO AND LEON, SPAIN by M. L. MARTINEZ-CHACON Abstract. Four new species of Stenoscismatacea, Camerisma (Callaiapsida) alcaldei , C. (Callaiapsida) paucicostala , Cyrolexis granti, and Stenoscisma winkleri are described from the Cantabrian Mountains. The two species of Cal- laiapsida, from the lower and upper Bashkirian, are the oldest attributed to the subgenus, and they show transitional characters between C. ( Camerisma ) and typical C. ( Callaiapsida ); moreover, the distribution of the subgenus is now extended from the Arctic. Cyrolexis granti from the Kashirian is also the oldest species of the genus, previously known only from the Permian. The Carboniferous Rhynchonellida from the Cantabrian Mountains are little known; they are infrequently cited and only two authors, Mallada (1875) and Delepine (1943), have described Carboniferous species belonging to the order. Moreover, the majority of these citations and descriptions refer to the Rhynchonel- lacea and until now only two forms of the superfamily Stenoscismatacea are known in the region: Camarophoria crumena (Martin) described by Mallada (1875) from different points in Palencia and Coledium sp. quoted by Wagner (19716, Winkler Prins’s determination) also from Palencia province. The present paper constitutes a first attempt to attain a regional knowledge of the representatives of this interesting group. Four new species are described, belonging to Camerisma , Cyrolexis , and Stenoscisma. They come from Oviedo and Leon, provinces very different from Palencia, despite all three being in the Cantabrian Mountains. Text-fig. 1 shows N Km.O 20 30 40 50 text-fig. 1. Map showing the situation of the fossiliferous localities. .... provincial boundary. [Palaeontology, Vol. 20, Part 1, 1977, pp. 209-223, pis. 29-30.] O 210 PALAEONTOLOGY, VOLUME 20 U. S. S. R. W EUROPE E ntrago - Pinos a rea Latores area C de Caso-Coballes area Hontoria ! area z PODOLSK1AN C Escalada Fm. Escalada Fm. < > KASHIRIAN o o z < B _i Beleho Fm. Shales and sandstones co < ~r ° VEREYAN CL •“ A San Emiliano z UPPER < co LU Fm. ~ MIDDLE $ < LOWER GO NAMURIAN C Valdeteja Valdeteja Valdeteja NAMURIAN Fm. Fm. Fm. text-fig. 2. Chart of the Carboniferous formations in the area, showing correlation with the Western European and Russian standard sections. the situation of the fossiliferous localities and text-fig. 2 shows the litostratigraphical units in the area and their age. All the specimens are housed in the Departamento de Paleontologia de la Universidad de Oviedo. SYSTEMATIC DESCRIPTIONS Superfamily stenoscismatacea Oehlert, 1887 Family atriboniidae Grant, 1965 Subfamily atriboniinae Grant, 1965 Genus camerisma Grant, 1965 1965a Camerisma Grant, p. 63. 1971 Camerisma Grant; Grant, p. 323. Type species. C. prava Grant, 1965. Subgenus callaiapsida Grant, 1971 1971 Camerisma ( Callaiapsida ) Grant, p. 323. Type species. C. ( Callaiapsida ) kekuensis Grant, 1971. Discussion. The Cantabrian specimens are smaller-sized than typical C. ( Cal- laiapsida) ; they show the characteristic peripheral grooves of the subgenus but with a much smaller development than those of the type species and of C. (C.) arctica (Floltedahl, 1924) in Grant’s (1971) figures; in addition, one of the forms here assigned to the subgenus bears ribs on the anterior part of the flanks. The internal characters coincide with those of Callaiapsida since the space between the camaro- phorium and the hinge plate is filled by secondary shell covering the intercamaro- phorial plate. On the basis of the peripheral grooves and internal characters these MARTINEZ-CHACON: CARBONIFEROUS BRACHIOPODS FROM SPAIN 211 species of Camerisma are assigned to the subgenus Callaiapsida , in spite of the short development of the grooves and small size of both forms, and the ribbed flanks of one of them. Camerisma (Callaiapsida) has been recorded from Podolskian and Permian rocks. The species here described, of lower and upper Bashkirian age, are the oldest yet recorded. It could be supposed that C. ( Callaiapsida ) would evolve in time from C. ( Camerisma ), with larger size and greater development of peripheral grooves; perhaps the Cantabrian species are the link between both subgenera. On the other hand. Grant (1971) suggests that C. ( Callaiapsida ) is exclusively Arctic in its distribution; this may be correct for the Permian, but during the Carboniferous the subgenus was more widely distributed, since it is unlikely that Northern Spain was at such high latitudes in the Carboniferous. Camerisma ( Callaiapsida ) alcalde i sp. nov. Plate 29, figs. 1-7; text-fig. 3 Derivation of name. Dedicated to Dr. J. L. Garcia-Alcalde. Material. Holotype (DPO 6915, PI. 29, tigs. 1-5), 13 paratypes (DPO 6916-6928) and other 42 specimens (DPO 6929-6970) from the type locality and 1 1 specimens (DPO 6971-6981) from Entrago (Oviedo). Type horizon. A light grey limestone, containing many brachiopods, crinoids, bryozoans, and molluscs, from the upper part of the Valdeteja Formation, of Lower Bashkirian age, from Latores village (6 km SW. of Oviedo), at the top of a small hill, 222 m high, near the km 7 of the Vasco-Asturiano railroad. Latitude 43° 19' 55", longitude 2° 12' 15". text-fig. 3. Camerisma (Callaiapsida) alcalde i sp. nov. Transverse sections of specimen DPO 6928 showing internal structures of both valves, especially the development of spondylium and camaro- phorium. Distances from the ventral beak in mm, x4. 212 PALAEONTOLOGY, VOLUME 20 Diagnosis. Species small for the subgenus, with very high and keel-shaped dorsal median fold, incipient peripheral grooves and smooth flanks. Description. Small shell of trigonal outline and keel-shaped profile; slightly trans- verse (L/W average of 0,84), although elongated in young stages; with anteriorly spreading flanks; rather thick (L/T average of 1,31); dorsibiconvex; anterior com- missure strongly uniplicate with inverted V-shape; shell frequently somewhat asym- metrical. Ventral valve gently convex. Beak pointed and strongly curved over dorsal umbo, without pedicle foramen. Sulcus beginning near middle of valve as a shallow and wide depression with rounded floor, widening and deepening rapidly toward the front, where it strongly curves dorsally to join the opposite fold through a high tongue; sulcus with median groove becoming stronger anteriorly. Ventral flanges appear anterior to articulation zone, overlapped by dorsal ones and giving way anteriorly to narrow and shallow peripheral grooves, covered by a very thin exten- sion of valve edge (see PI. 29, figs. 4, 6, 7). Dorsal valve strongly convex, with a beak pointed, directed towards ventral valve and covered by it; umbonal region very convex and high. A sharp and keel-shaped fold starts in posterior half of valve; separated from flanks by two more or less pronounced grooves ending at boundary between anterior and lateral commissures; flanks descending also abruptly (although to smaller degree than fold) from such grooves to the commissure. Dorsal flanges over- lapping ventral ones all around commissure, with greatest overlap in posterior part. Ornament only of growth lines, more visible on ventral valve, especially on the ventral flanks. Dimensions of some specimens in mm DPO L W T DPO L W T 6915 15 20,8 14,8 6923 16,3 17,5 12,6 6916 15,8 19,3 14,4 6924 15,4 19,1 13,3 6917 15,4 20,5 12,5 6925 19 20,5 11,7 6918 13,1 15,6 10,5 6926 9,9 9,2 5 6919 15,9 17,1 9,6 6927 14,6 15,6 9,3 6920 16,6 19,5 14,8 6971 12,4 18,2 11,2 6921 13,1 15 10 6972 13,9 18 13,6 6922 12,1 14,5 8,8 EX PLAN ATION OF PLATE 29 Figs. 1-7. Camerisma (Cal/aiapsida) alcaldei sp. nov., 2. 1-5, holotype, DPO 6915. 1-3, dorsal, ventral, and lateral views. 4, anterior view showing incipient peripheral groove on tongue covered by very thin extension of ventral valve margin. 5, posterior view showing peripheral groove 'on the left lower side and spondylium. 6, 7, anterior view of paratypes, DPO 6916, 6917, both showing incipient peripheral grooves on tongue. Figs. 8-12. Camerisma ( Callaiapsida ) paucicostata sp. nov., x2. Holotype, DPO 6982, dorsal, ventral, lateral, anterior, and posterior views showing pronounced asymmetry; note poorly developed peri- pheral groove on tongue in fig. 11. Figs. 13, 14. Stenoscisma winkleri sp. nov., x 3. DPO 7273, 7272, posterior view of two internal moulds of both valves showing spondylium and camarophorium. Fig. 15. Cyrolexis granti sp. nov., x4. Holotype, DPO 7283, posterior view of a specimen with valves partially destroyed, showing part of spondylium and camarophorium. PLATE 29 MARTINEZ-CHACON, Spanish Carboniferous Stenoscismatacea 214 PALAEONTOLOGY, VOLUME 20 Ventral interior posteriorly very thickened, spondylium large (see text-fig. 3), with thick walls, supported by low and broad median septum duplex, buried by secondary shell in posterior part (and further forward) so as to appear sessile in its beginning; septum continuing slightly forward of the spondylium. Teeth strong, short, rounded, and each one presenting a lateral exterior crenulation. Dorsal interior also posteriorly thickened. Camarophorium with high and slender median septum duplex, supporting narrow trough with semicircular transverse section. Thick and duplex intercamarophorial plate buried posteriorly in secondary shell, free anteriorly when it ceases to support the hinge plate. Short hinge plate finishing behind camarophorium; inner hinge plates extending more anteriorly than outer hinge plates. Big, rounded cardinal process, with myophore of high lamellae. Crura diverging from hinge plate anteroventrally. One specimen, with part of its anterior flank decorticated, shows deep pallial impressions which branch dichotomously, especially close to the anterior margin. Discussion. C. (C.) alcaldei is distinguished from typical Callaiapsida by its smaller size and poorly developed peripheral grooves. Its general form looks like Camerisma ( Corner isma ) sella (Kutorga, 1844) and C. ( Callaiapsida ) paucicostata sp. nov., but the difference between them is the presence of a higher and narrower fold and the absence of ribs on the flanks in the former. Camerisma ( Callaiapsida ) paucicostata sp. nov. Plate 29, figs. 8-12; text-figs. 4, 5 Material. Holotype (DPO 6982, PI. 29, figs. 8 12), 6 paratypes (DPO 6983-6988) and 12 other specimens (DPO 6989-7000) from the type locality (locality 30 of Winkler Prins, 1968) and 2 other specimens (DPO 7001, 7002) from a nearby locality (locality 29 of Winkler Prins, 1968). Type horizon. A limestone in the upper part of the San Emiliano Formation, Upper Bashkirian age, north- east of Pinos village (Leon), slope of the mount on the left from Pinos following the Alcantarilla stream. Latitude 42° 59' 11", longitude 2° 16' 56". Diagnosis. Small for the subgenus, with high and keel-shaped fold, flanks ornamented by one to three ribs, peripheral grooves with slight development and rather pro- nounced asymmetry. Description. Small shell of trigonal outline and keel-shaped profile, transverse (L/W approximately 0,75), rather thick (L/T approximately 1,15), with anteriorly spread- ing flanks; dorsibiconvex; anterior commissure strongly uniplicate and inverted V-shaped; pronounced asymmetry. Ventral valve gently convex. Beak pointed and strongly incurved over dorsal umbo, without pedicle opening. Sulcus beginning near posterior third of valve as gentle and rounded depression deepening anteriorly, where it curves dorsally into the high V-shaped tongue; median groove running along floor and commonly asymmetrical anteriorly; sulcus separated from each flank by one rib. Flank towards which median groove slopes is less well developed. Ventral flanges posteriorly overlapped by dorsal ones, anteriorly giving rise to shallower peripheral grooves, which are covered by a laminar expansion of valve edge (see text-fig. 5 and PI. 29, fig. 11). Dorsal valve very convex in transverse section. Beak ventrally directed, sinking in other valve and covered by opposite beak. Umbo very MARTINEZ-CHACON: CARBONIFEROUS BRACHIOPODS FROM SPAIN 215 convex and high. Median fold starting in posterior third of valve, keel-shaped, posteriorly narrow but widening anteriorly, bordered by grooves separating it from steep flanks. As in the ventral valve, one flank, commonly better developed. Dorsal flanges with greatest overlap posteriorly. Flanks ornamented by one to three strong, angular ribs on each one, those bordering the fold and sulcus developed first, the lateral ribs only formed near the anterior margins. Ribs more prominent on ventral valves and more numerous on the larger side of shells. Growth lines faint except on the tongue where they may be almost rugose. Ventral interior with spondylium supported by low and thick median septum duplex (see text-fig. 4), buried by secondary shell posteriorly so as to appear sessile; when septum is conspicuous, its base is very thickened, thickness decreasing anter- iorly and dorsally; spondylium rather narrow ventrally, widening dorsally. Teeth text-fig. 4. Camerisma ( Callaiapsida ) paucicostata sp. nov. Transverse sections of specimen DPO 6988 showing internal structures. Distances from the ventral beak in mm, x 4. 216 PALAEONTOLOGY, VOLUME 20 are short, rounded, and weak. Dorsal interior posteriorly thickened to bury the high but short median septum duplex supporting a short narrow trough of the camarophorium. Intercamarophorial plate buried posteriorly. Hinge plate with very wide and short outer hinge plates and very narrow and rather long inner ones, which cover the camarophorium only posteriorly. Cardinal process large and rounded, with its myophore of short lamellae. Crura project forward from the hinge plate ventrolaterally. Dimensions of some specimens in mm DPO L W T 6982 15,3 19 13,2 6984 14,1 19,2 — 6985 16 21,7 14 6986 — 21,2 16 6987 17 21,8 14,6 text-fig. 5. Camerisma (Callaiapsida) paucicoslala sp. nov. a, transverse section near the front of de- formed specimen DPO 6983 showing incipient peripheral groove on the right, x 3. b, detail of the same section, with poorly developed groove, x 13. Discussion. C. ( Callaiapsida ) paucicostata and C. ( Callaiapsida ) alcaldei are smaller and have less well developed peripheral grooves than the type species of Callaiapsida. Camerisma (C.) paucicostata is distinguished from the other Cantabrian species by its ribbed flanks, lesser thickness (owing to a lower fold) and greater degree of asymmetry. It resembles C. ( Camerisma ) sella (Kutorga), as figured by Tschernyschew (1902, pi. 23, fig. 4), in its general form and ribbed flanks; however, on the figures the peripheral grooves are not apparent and are not mentioned in the description. Grant (1971) assigned the species to C. ( Camerisma ), so he too must have considered the peripheral grooves to be lacking. On the other hand, paucicostata seems to have more and better-developed ribs, in contrast to C. (C.) sella where they are more evident on the dorsal valve than on the ventral. C. (C.) sella is a little larger than C. ( Callaiapsida ) paucicostata , with the maximum width more posteriorly located and it presents a lower asymmetry level. MARTINEZ-CHACON: CARBONIFEROUS BRACHIOPODS FROM SPAIN 217 Subfamily psilocamarinae Grant, 1965 Genus cyrolexis Grant, 1965 1965a Cyrolexis Grant, p. 88. Type species. Cyrolexis haquei Grant. 1965. Cyrolexis granti sp. nov. Plate 29, fig. 15; Plate 30, figs. 23-26 Derivation of name. Dedicated to Dr. R. E. Grant. Material. Holotype (DPO 7283, PI. 29, fig. 15), specimen with the two valves partially destroyed showing part of the internal structures, and twenty paratypes (DPO 7284-7302, 8522) from the type locality and six specimens (DPO 7303-7307) from three nearby localities. All the specimens are preserved as moulds or casts, except for the holotype in which some shell remains. Type locality. Basal part of a blackish shale, thickness of 25-30 m, in the upper part of the Beleno Forma- tion, about 80 m below the Escalada Formation (Lower Moscovian, probably Kashirian, according to Winkler Prins, 1968), between Campo de Caso and Coballes (Oviedo), on the right of the road from Riano to Oviedo, coming from km 53 to 54. Longitude 1° 39' 50", latitude 43 10' 47". Diagnosis. Small, globose and transverse shell with a fold weakly differentiated from the flanks, and beginning in posterior half of valve; bearing two strong ribs. Sulcus shallower than fold, beginning generally anteriorly to it and having a median rib, flanks smooth or with a rib. Interior with no intercamarophorial plate or developed only as a low elevation on the posterior floor of the camarophorium; hinge plate supported by sides of posterior part of camarophorium. Description. Small (length between 5 and 7 mm), transverse (width between 7 and 9 mm), globose and dorsibiconvex shell. Anterior commissure sulciplicate. Ventral valve with greatest convexity in umbonal region; beak short and slightly incurved over dorsal one. Sulcus shallow, with rounded floor and greatly widening anteriorly; at about 3 mm or more forward from the umbo a wide but low and rounded rib developed. In some specimens a pair of ribs separate the sulcus from the flanks. Posteriorly flat flanges are overlapped by those of the dorsal valve. Dorsal valve more strongly convex; beak short, ventrally directed and covered by that of opposite valve. Fold rather high but not well distinguished from flanks; beginning slightly further back than sulcus and bearing two strong, wide, angulate or rounded ribs, which leave an intermediate groove of approximately equal width. Flanks smooth or bearing one rib on each. Ventral interior with rather large spondylium, sessile in posterior extreme and anteriorly elevated on median septum extending forward of it (see PI. 30, figs. 23-26); spondylium occupying about one-third of valve length, with muscle tracks apparent on it. Both sides of median septum of gerontic speci- mens bearing thick ridges, becoming rather high, supporting sides of spondylium and continuing anteriorly to septum. Teeth small. Dorsal interior with rather long and narrow camarophorial trough on median septum, which increases in height forward to become very high and after that fades sharply (see PI. 30, figs. 23, 24, 26); walls of camarophorium joining underside of hinge plate, camarophorium con- tinuing free in anteroventral direction and extending further forward than spondylium of opposite valve. Hinge plate short, concave, except in its apex, where it bears a big 218 PALAEONTOLOGY, VOLUME 20 cardinal process of trigonal section and with lamellar myophore. Sockets long, narrow, located between lateral extreme of hinge plate and very narrow and low ridge separating them from valve wall. Crura initially following direction of camaro- phorium and later directing more ventrally. Intercamarophorial plate seen only in one specimen and reduced to very low and narrow elevation, located on posterior part of floor of camarophorium. Discussion. The species is assigned to Cyrolexis because of its internal characteristics, although externally it resembles more closely some species of Coledium Grant, 1965. Cyrolexis granti is distinguished from C. haquei and other species of Cyrolexis by its transverse form, less numerous ribs on fold and sulcus, normally smooth flanks, and fold that is little differentiated from them. Exteriorly it looks like Coledium explanatum (McChesney, 1860) and particularly C. globulinum (Phillips, 1834) of Grant (1965u). It is distinguished from the former by the absence of the intercamaro- phorial plate, smaller size, more transverse form, and weaker ribs. In C. globulinum the presence of the intercamarophorial plate has not been firmly established; Grant (1965fl, p. 120) says: intercamarophorial plate not observed, probably present’, and Tschernyschew (1902) does not refer to it. If such a plate does not exist, C. globulinum could belong to Cyrolexis and it would be close to C. granti. Both differ externally by the more transverse form and even smaller size of the Cantabrian species. C. granti is now the oldest species attributed to the genus, which was considered restricted to the Permian. Family stenoscismatidae Oehlert, 1887 Subfamily stenoscismatinae Oehlert, 1887 Genus stenoscisma Conrad, 1839 1965a Stenoscisma Conrad; Grant, p. 138. Type species. Terebratula schlotheimi von Buch, 1835. Stenoscisma winkleri sp. nov. Plate 29, figs. 13, 14; Plate 30, figs. 1-22; text-figs. 6, 7 Derivation of name. Dedicated to Dr. C. F. Winkler Prins. EXPLANATION OF PLATE 30 Figs. 1-22. Stenoscisma winkleri sp. nov. 1-5, x2. Holotype, DPO 7005, dorsal, ventral, lateral, anterior, and posterior views showing the remains of the stolidium. 6-9, x 2. Paratype, DPO 7007, dorsal, ventral, anterior, and posterior views. 10-13, x2. Paratype, DPO 7008, dorsal, ventral, anterior, and posterior views. 14-17, x2. Paratype, DPO 7009, dorsal, ventral, anterior, and posterior views. 18-21, x2. Para- type, DPO 7010, dorsal, ventral, anterior, and posterior views. 22, x 4. Paratype, DPO 7206, fragmentary specimen showing skeletal internal structures of the cardinal region; dorsal valve above. Figs. 23-26. Cyrolexis granti sp. nov. 23, 24, x 5. Paratype, DPO 7286, internal mould of both valves and latex cast of posterior region showing spondylium and teeth in ventral valve and hinge plate and camaro- phorium without intercamarophorial plate in dorsal valve. 25, X 5. Paratype, DPO 7284, posterior view of an internal mould of both valves showing spondylium and camarophorium. 26, x 9. Inclined view of the same specimen to show mould of the cardinal process. PLATE 30 MARTINEZ-CHACON, Spanish Carboniferous Stenoscismatacea 220 PALAEONTOLOGY, VOLUME 20 Material. Holotype (DPO 7005, PI. 30, figs. 1-5), 75 paratypes (DPO 7006-7076, 7142-7145) and 135 other specimens (DPO 7077-7141, 7146-7205) from the type locality; 5 paratypes (DPO 7206-7210) and 42 other specimens (DPO 7211-7252) from a locality NE. of Pinos (province of Leon = locality 29 of Winkler Prins, 1968), and 30 specimens (DPO 7253-7282) from other occurrences. Type locality. Marly limestone of the Escalada Formation, at about 30 m above its base, of upper Kashir- ian age (according to van Ginkel, 1965), from a small cove on the coast north-east of Hontoria (province of Oviedo). Longitude 1° 41' 43", latitude 43° 27' 8". Diagnosis. Stenoscisma with trigonal outline, very expanded anteriorly, transverse. Fold normally with four to five strong ribs and sulcus with one fewer. Flanks smooth or with one, or exceptionally two, ribs on each. Ribs beginning rather far anteriorly to beak. Short stolidium, little developed and inappreciable for most of the specimens. text-fig. 6. Stenoscisma winkleri sp. nov. Transverse sections of specimen DPO 7207 showing internal shell structures. Distances from the ventral beak in mm, x4. Description. Small to medium-sized shell, outline trigonal, widely spreading near anterior margin, transverse although juvenile specimens slightly elongated, dorsi- biconvex; anterior commissure uniplicate. Stolidium in some specimens (see PI. 30, figs. 1-5) extending to anterolateral and, sometimes, anterior margins of both valves, but always only very narrow, reaching a few millimetres in width. Ventral valve gently convex, beak pointed, and strongly incurved dorsally. Small and triangular delthyrium, constricted by narrow and disjunct deltidial plates, leaving triangular foramen. Sulcus beginning from one-third to one-half of valve length, shallow and flat-bottomed but widening to one-third of valve width anteriorly where it deepens as it curves dorsally into the tongue; sulcus is bounded at both sides by wider and more rounded rib than those of its floor. Posterolateral flanges strongly developed, high, and overlapped by those of opposite valve. Dorsal valve more strongly convex. Beak pointed, ventrally directed into ventral valve. Fold beginning at about mid- length, low except anteriorly where the flanks fall away steeply, top flat or gently arched. Median ribs wider than interspaces, slightly diverging forward, angular in most specimens; lateral ribs developed later in life, sometimes by bifurcation. Number of ribs variable (see PI. 30, figs. 4, 8, 12, 16, 20), most frequently 4-5 on fold (some specimens bearing 3 or 6, very few 2, and only one with 8); sulcus one less than fold; flanks usually smooth or with one or, more rarely, 2 ribs each, generally MARTINEZ-CHACON: CARBONIFEROUS BRACHIOPODS FROM SPAIN 221 beginning further anteriorly; weaker than those on sulcus and fold. Growth lines usually weak, but occasionally stronger. Dimensions of the holotype in mm: L 12,7 ; W 1 5 ; T 7,6. It bears 5 ribs on the fold. The dimensions of other specimens are in text-fig. 7. ,7i 15- 13- • . L ' ■ 11- ' ' . 9- 7- . 6- 5-1 , , T t , 4 5 6 7 9 11 13 15 17 W 17- 15- - 13 ' L ' # • ■' 11- . 9- 7 , • 6 . _ 5j , , , , , 19 3 4 6 8 10 12 T text-fig. 7. Dimensions of Stenoscisma winkleri sp. nov. in mm. • specimens from the type locality: x specimens from locality 29 of Winkler Prins (1968), and from other localities nearby. Ventral interior with wide spondylium supported by very low median septum duplex (see text-fig. 6 and PI. 30, fig. 22), posterior part buried by secondary shell, so spondylium appears apically sessile; septum seen anteriorly with a height just over 1 mm, extending slightly anterior to spondylium; length of spondylium about one-third of valve length. Umbonal chambers can be full of secondary shell. Teeth strong, short, upper part rounded and each one having a lateral exterior crenulation. Dorsal interior with low and narrow camarophorial trough, anteroventrally directed and extending slightly in front of spondylium; supported by thin median septum 222 PALAEONTOLOGY, VOLUME 20 duplex, with height increasing towards anterior edge of trough, becoming very high, and from there decreasing parabolically, disappearing 1 or 2 mm anteriorly. Intercamarophorial plate duplex, high and long, occupying all the length of trough, posterior part supporting hinge plate. Posterior part of camarophorium can be buried by secondary shell and intercamarophorial plate not discernible to its anterior edge, in front of hinge plate. Hinge plate undivided, supporting on its apex a big and wide cardinal process, with myophore of high and thin lamellae; inner hinge plates extending rather more anteriorly than outer ones. Crura curving in ventral direction. Sockets narrow, short, and located at anterolateral extremities of outer hinge plates. Discussion. S. winkleri is about the same size as S', schlotheimi , but it differs from the latter by its more trigonal outline with the maximum width nearly at the front, lower fold, generally stronger and more numerous ribs on fold and sulcus, and smaller developed stolidium. It also resembles S. crumenum (Martin, 1809) and S. mutabile (Tschernyschew, 1 902), but is smaller, less thick, and its stolidium is less well developed than in both of these species; in addition S. crumenum has more strongly ribbed flanks and S. mutabile in adulthood is more fully ribbed than S. winkleri. The species is known from San Emiliano, Beleno, and Escalada Formations, which are upper Bashkirian to upper Kashirian age. Acknowledgements. I am very grateful to Dr. R. E. Grant of the Smithsonian Institution, Washington for kindly reviewing the manuscript. I also thank Miss I. Mendez Bedia for her aid in the English translation, Mr. M. Arbizu who has taken the photographs (both are in the Departamento de Paleontologia, Oviedo), and Mrs. A. Borraz for making the drawings. REFERENCES delepine, g. 1943. Les faunes marines du Carbonifere des Asturies (Espagne). Mem. Acad. Sci. Inst. France, 66 (3), 1-122, pis. 1-6. - and llopis llado, n. 1956. Nouvelle faune Carbonifere a Latores (Asturies-Espagne). C.R. somm. Seances Soc. Geol. France , 106 108. ginkel, A. c. van. 1965. Carboniferous fusulinids from the Cantabrian Mountains (Spain). Leidse Geol. Med. 34, 1-225, 13 hgs. grant, r. e. 1965u. The Brachiopod superfamily Stenoscismatacea. Smiths. Misc. Coll. 148, 192 pp., pis. 1-24. - 19656. Superfamily Stenoscismatacea. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology, Pt. H , Brachiopoda, (2), 625-632, figs. 511-516. Kansas. 1971. Taxonomy and Autecology of two Arctic Permian Rhynchonellid Brachiopods. Smiths. Contrib. Paleobiol. 3, 313-335, pis. 1-3. mallada, L. 1875. Sinopsis de las especies fosiles que se han encontrado en Espana. Bol. Com. Map. Geol. Espaha, 2, 1-160, pis. 1-7 (Siluriano), 1, 4-8, 13, 14 (Devoniano), 1-3, 6-1 1 (Carbonifero). Martinez alvarez, J. a. 1962. Estudio geologico del reborde oriental de la Cuenca Carbonifera Central de Asturias. Inst. Est. Asturianos, Dip. Prov. Oviedo, (I) Text, 1-229, (II) Map. tschernyschew, t. 1902. Die obercarbonischen Brachiopoden des Ural und des Timan. Mem. Com. Geol. 16, (I) Text, 1 749, (II) Atlas, pis. 1-63. wagner, r. h. 1971«. Account of the International Field Meeting on the Carboniferous of the Cordillera Cantabrica, 19-26 Sept. 1970. With contributions by a. garcia-loygorri and j. a. knight. Trab. Geol. Fac. Ci. Univ. Oviedo, 3, 1-39, pis. 1-5. - 19716. Carboniferous nappe structures in northeastern Palencia (Spain). Ibid. 4, 431-459, pis. 1-7. M ARTINEZ-CHACON: CARBONIFEROUS BRACHIOPODS FROM SPAIN 223 wagner, R. h., winkler prins, C. F. and riding, R. E. 1971. Lithostratigraphic units in the lower part of the Carboniferous in northern Leon, Spain. With a ‘Note on some goniatite faunas’ by wagner-gentis, c. h. t. Ibid. 4, 603-663, pis. 1-3. williams, a. and rowell, a. j. 1965. Morphology and Morphological terms applied to brachiopods. In MOORE, R. c. (ed.). Treatise on Invertebrate Paleontology, Pt. H, Brachiopoda, 57-155, figs. 59-138. Kansas. winkler prins, c. F. 1968. Carboniferous Productidina and Chonetidina of the Cantabrian Mountains (NW. Spain): Systematics, Stratigraphy and Palaeoecology. Leidse Geol. Med. 43, 41-126, pis. 1-9. Original typescript received 10 November 1975 Revised typescript received 10 March 1976 M. L. MARTINEZ-CHACON Departamento de Paleontologia Umversidad de Oviedo Spain TWO CARBONIFEROUS BLASTOIDS FROM SCOTLAND by D. B. MACURDA, JUN. Abstract. The two fissiculate blastoids, Astrocrinus letragonus (Austin and Austin, 1843) and Hadroblastus(l) benrtiei (Etheridge and Carpenter, 1886), occur together in Early Carboniferous Visean (D) sediments of Scotland. Astrocrinus is a small, free-living eleutherozoic blastoid characterized by well-developed surface ornament. The nodes are interpreted as spine bases which, together with some brachioles, probably stabilized the animal on its substrate in periods of higher energy. //.(?) benniei , a stemmed blastoid formerly identified as Phaenoschisma ? benniei, extends the known geographic range of Hadroblastus from North America to the British Isles, and the stratigraphic range into the Upper Visean, if the species assignment is correct. In their definitive study of the blastoids in 1886, Etheridge and Carpenter listed three Lower Carboniferous blastoids from Scotland. The fragmentary nature of many of these has made their interpretation difficult. Astrocrinus tetragonus (Austin and Austin, 1843) is one of the very few free-living (eleutherozoic) blastoids; it is also found in England and Ireland. A second species, which is known only from Scotland, was originally described as Phaenoschisma benniei Etheridge and Car- penter, 1886. The third species was represented only by isolated ambulacra and thought to be a spiraculate blastoid (Etheridge and Carpenter 1886, p. 279; pi. II, figs. 38-42). Breimer and Macurda (1972) reviewed the first two species in a study of the phylogeny of the fissiculate blastoids. Because of its unique morphology, Austin and Austin (1843) erected the family Astrocrinidae for Astrocrinus , and it remains the only genus in this family. Breimer and Macurda (1972) described the internal structure of A. tetragonus and briefly discussed its growth and geographic occurrence. Etheridge and Carpenter (1886) discussed the nodes on the plates of A. tetragonus and concluded that they were perforate and probably bore a spine which articulated on them. Study of these nodes by scanning electron microscopy has shown them to be imperforate and of importance in deducing the life mode of the animal. Further preparation of specimens of P. benniei has shown that this species almost certainly belongs to the North American Mississippian genus Hadro- blastus and thus extends its geographic range. DIAGENESIS, BIOSTRATINOM Y, AND PAL AEOECOLOG Y The Scottish blastoids discussed herein have been collected from a sequence of Visean D Zone limestones and calcareous shales. Etheridge and Carpenter (1886) listed their specimens as coming from the shales above No. 1 and No. 2 Limestones of the Lower Carboniferous Limestone group near Midlothian and Fife (Astro- crinus) and the Shale above the No. 2 Limestone in the East Salton and Kidlaw Quarries near Gifford, Haddingtonshire (. Hadroblastus (?) benniei). Subsequent collections were made during the first part of the twentieth century by James Wright, [Palaeontology, Vol. 20, Part 1, 1977, pp. 225-236, pis. 31-32.] P 226 PALAEONTOLOGY, VOLUME 20 whose collection is now in the Royal Scottish Museum. His material bears the locality labels ‘Carlops, Peeblesshire’ and ’No. 1 Bed, Invertiel, Fife’. These materials were apparently obtained as a by-product of Wright’s study of Allagecrinus from the localities in the Lower Limestone Group (Wright 1941). George (1971) reviewed the stratigraphic complexities and biostratigraphy of the Lower Limestone Group. The Visean-Namurian boundary occurs somewhere near the top. The blastoids from the Lower Limestone Group are almost invariably crushed. This is not unexpected, as they are preserved in shales which underwent compaction. The thecae of Astrocrinus are compressed vertically and adjacent plates are strongly displaced. One frequent zone of failure is through the base of a theca along a line connecting the C and E ambulacra. This particularly obscures the relationships of the basal plates. //.(?) benniei is crushed flat and usually only partial thecae are recovered. Preservation in the shale, however, has had one beneficial effect. The skeletal micro structure (stereom) of echinoderms is an open lattice with numerous pores. Diagenetic calcite cements crystallize in optical continuity with the stereom and overgrow it externally. This most always obliterates surface detail, and the internal fabric is often disrupted. Preservation in fine-grained sediments instead of skeletal limestones may reduce the severity of the cementation process, resulting in excellent external and internal preservation of the blastoid stereom (e.g. Macurda 1973). The stereom is clearly evident within the plates of the specimen of //.(?) benniei in Plate 32, figs. 5, 8, when it is immersed in xylene. It is also evident on the surface of the plates of some specimens of A. tetragonus (PI. 31, figs. 3, 6). The preservation of the surface detail on the plates of these blastoids is very good; this would suggest they are essentially preserved in the environment in which they lived. After death, they remained unburied long enough for the tissue binding the brachioles, ambulacral covering plates, and the stem in //.(?) benniei , to decompose, permitting disarticula- tion. //.(?) benniei is a blastoid that is conventional in appearance and had a stem cicatrix. There is no morphologic evidence to indicate an unusual mode of life; it was probably a current-seeking (rheophilic) blastoid (Type I of Breimer and Macurda 1972). A. benniei is highly unusual. It is as though the lower half of the blastoid were stunted and grew to the side. It has a quadrate outline (PI. 32, figs. 4, 7, 9), has no stem cicatrix, lacks any attachment scars, and thus was free living. The theca is always small (less than 10 mm) and even moderate oceanic swells or waves or tidal currents would exert a lifting force which would continually reorientate the animal, perhaps detrimentally. Thus it would appear to have lived in quiet or deep-water environments or might have had some special adaptations permitting it to cope with periods of higher environmental energy. The stratigraphic relationships of the Lower Limestone Group are not indicative of deeper water; the geographic extent of Astrocrinus and its local abundance suggest it was a functionally successful design. The ornament found on the surface of the plates is probably important in under- standing the life mode of Astrocrinus. Most blastoid plates were secreted by deposi- tion of calcite laterally along the edges of the plates, and growth lines are evident on the external surface. In Astrocrinus these are only occasionally evident on the sloping walls of the ambulacral sinus formed by part of the deltoid body and RD sector of the radial. Nodes occur on the upper part of the deltoid body along the D. B. MACURDA, JUN.: CARBONIFEROUS BLASTOIDS FROM SCOTLAND 227 crest and its sloping sides (PI. 31, figs. 1, 3). They can have a row-by-row arrangement parallel to the radiodeltoid suture, suggesting formation as the plate grew aborally. Since their growth was upward, they represent deposition on the external surface, either contemporaneously with the growth-line formation or just subsequent to it. Those adjacent to sutures are of equal magnitude to those formed earlier and there is no gap as one approaches the radiodeltoid suture. The maximum diameter of the nodes is 0-15 mm and their height is about the same. There are approximately twelve radially disposed grooves on the side of each node; these terminate below the tip of each node. Nodes are also present on the upper half of the regular radials (above the aboral tip of the ambulacrum, external to the ambulacral sinus). Ornament on the lower half of the radials assumes a more linear appearance, the nodes assuming a laterally directed ridge-like aspect (PI. 32, fig. 3). On the attenuated D radial and the bordering limbs of the C and E radials, ornament consists of sharp, linear ridges (PI. 32, figs. 1, 2). These are nearly perpendicular to their respective sutures. Articular surfaces between larger plates in echinoderms are usually characterized by non-porous stereom, as are the fulcral ridges of articulations in crinoids (Macurda and Meyer 1975) or the mammelons of echinoids (Jensen 1972). The surfaces on which small spines articulate in echinoderms are less obviously expressed in the stereom. Comparative study of the morphology of these surfaces should be helpful in determining the presence of spines in fossil echinoderms. The arms of the ophiuroid Astrophyton bear prominent spines used to snare prey, but the surface expression of the articulation of the spine with the arm is not well defined (Macurda 1976). The Devonian crinoid Arthroacantha bears spines on the surfaces of the basals and radials (e.g. Kesling and Chilman 1975, pi. 29, figs. 1-3). These spines may be at least 5 0 mm long and the diameter of the raised area where they articulate is only 0-3 mm. Without preserved spines, the raised areas would probably be merely inter- preted as ‘surface ornament’. It is tempting to suggest that each node of Astrocrinus tetragonus bore some type of spine and that the grooves on the side of each node represent points of insertion for muscles or ligaments to articulate the spine (PI. 31, fig. 7). Etheridge (1876) reported and illustrated a microscopic spine adhering to one of his specimens by some particles of matrix, but it was not attached in place. No spines were observed during this study. The stereom of each node is apparently solid. Initially there are approximately six grooves (PI. 31, fig. 2) but, as the node becomes larger, new grooves are inserted in intervening spaces. The tip of a node is a blunt, low hemispherical cap (PI. 31, fig. 4). There is no differentiation of the stereom around the base of a node. In the absence of the direct preservation of articulated spines, their presence in Astrocrinus remains speculative, but the morphology of the nodes is suggestive of their presence. The following discussion attempts to interpret the palaeoecology of Astrocrinus assuming their presence. The purpose of the spines was protective and they projected into the surrounding water. The animal sat on the substrate with the area beneath the D ambulacrum in contact with the substrate (see PI. 32, figs. 4, 7, 9). The brachioles from the D sides of the C and E and the attenuated D ambulacra were splayed on to the substrate to help stabilize the organism in this position; the AB interarea projected uppermost. The linear ornament below the D ambulacrum also served to resist displacement due to waves or currents. (The D sides of the C 228 PALAEONTOLOGY, VOLUME 20 and E ambulacra are wider than their opposite sides and less steep; this would have allowed the brachioles to project more nearly parallel to the substrate.) Spines borne in the EA, AB, and BC interareas on the body of the radials immediately below these interareas projected outward to form a protective forest of spines; some of those on the radial bodies may also have projected into the substrate. Thus, in spite of its small size, Astrocrinus could stabilize and position itself on the substrate through a combination of linear ornament, spines, and brachioles. The brachioles of the A and B ambulacra functioned normally but those of the D and parts of C and E were utilized to provide fixity for this free-living blastoid. SYSTEMATIC PALAEONTOLOGY Class blastoidea Say, 1825 Order fissiculata Jaekel, 1918 Family astrocrinidae Austin and Austin, 1843 Genus astrocrinus Morris, 1843 1843 Astracrinites Austin and Austin, p. 1 10, 1843 Astrocrinus Morris, p. 49. 1848 Zygocrinus Bronn, p. 1381. Type species. Astracrinites tetragonus Austin and Austin, 1843. Astrocrinus tetragonus (Austin and Austin) Plate 31, figs. 1-7; Plate 32, figs. 1-4, 7, 9; text-fig. 1 1843 Astracrinites tetragonus Austin and Austin, p. 1 10. 1843 Astrocrinus tetragonus Morris, p. 49. 1876 Astracrinites benniei Etheridge, p. 103. Scottish material. Royal Scottish Museum specimen numbers 1958.1.2355, 2360-2364, 2367, 2376, 2378- 2380. Description. Theca squat, ovoid in lateral view, pentagonal, asymmetric in oral view. Stem cicatrix lacking, animal being eleutherozoic with development of bilateral symmetry along plane of AB interarea and D ambulacrum (PI. 32, figs. 4, 7, 9). AB, BC, and EA interareas protuberant, while CD and DE form a continuous, slightly convex arc in oral view due to shortened D ambulacrum. A and B ambulacra EXPLANATION OF PLATE 31 Astrocrinus tetragonus (Austin and Austin, 1843); Lower Limestone Group, Lower Carboniferous; Carlops, Peeblesshire, Scotland. Scanning electron micrographs. Pigs. 1-7. 1, inclined view of nodes on deltoid, x 115; R.S.M. 1958.1.2363. 2, incipient node on deltoid, <455; R.S.M. 1958.1.2363. 3, plan view of deltoid, oral direction at top, x35; R.S.M. 1958.1.2380. 4, fully developed node on deltoid, x 350; R.S.M. 1958.1.2363. 5, brachiolar facets on edge of ambu- lacrum, oral direction to left, x 115; R.S.M. 1958.1.2363. 6, detail of stereom on wall of ambulacral sinus, x685; R.S.M. 1958.1.2380. 7, fully developed node on deltoid with top broken off, x455; R.S.M. 1958.1.2363. (R.S.M., Royal Scottish Museum.) PLATE 31 MACURDA, blastoids from Scotland 230 PALAEONTOLOGY, VOLUME 20 convex, set within broad shallow ambulacral sinus and extending to near base of theca; C and E ambulacra slightly longer, extending to base of outline in lateral view and visible in aboral view, recurving inward. D ambulacrum short, flat, confined to upper surface of theca. Length 2-32 mm; width, A-CD, 3-65 mm; AB-D, 4-48 mm. Basalia two, extending from centre of aboral surface half-way up lateral surface beneath D ambulacrum (text-fig. 1). Outline elongate, narrow, pentagonal. No stem text-fig. 1. Astrocrinus tetragonus , aboral view. Royal Scottish Museum specimen number 1958.1.2376, x 12. Compare with Plate 32, fig. 7. Radials A-E labelled. Drawn with camera lucida. cicatrix. Basal on aboral surface flat, has two equal short edges against narrow bases of A and B radials (0-56 mm each); basal extending toward and narrowing slightly toward D ray. Edge against C radial of about equal length to A and B (0-48 mm), E about twice as long (0-88 mm); interbasal suture extends diagonally from a point near a centre line of E ambulacrum diagonally towards B side of C radial. Interbasal suture very slightly sinuous. Other basal convex in lateral view, quadrate in plan view, elongate (1-20 mm), with long edges against C and E radials and narrow edge against D radial half-way up lateral edge of theca; maximum width 0-40 mm. Lower basal ornamented with pustulose ornament, upper basal with a few irregular ridges parallel to long axis of plate (PI. 32, fig. 1). EXPLANATION OF PLATE 32 Figs. 1-4, 7, 9. Astrocrinus tetragonus (Austin and Austin, 1843). 1-3, Lower Limestone Group, Lower Carboniferous; Carlops, Peeblesshire, Scotland. 4, Di beds, Visean, Lower Carboniferous; foreshore Oyster Bay, Fenit, Co. Kerry, Ireland. 7, 9, Lower Limestone Group, Lower Carboniferous, Invertiel, Fife, Scotland. 1, D deltoid (right side) and bordering C and E radials and basal (left centre), X30; R.S.M. 1958.1.2361. 2, detail of ornament of D deltoid of fig. 1, x285. 3, radial bodies, A right, B left, x50; R.S.M. 1958.1.2363. 1-3, scanning electron micrographs. 4, oral view of specimen pre- served on limestone surface, A ambulacrum at 9 o’clock, x6; N.M.I. G. 40. 1965. 7, 9, aboral and oral views with A and B ambulacra at 12 o’clock respectively, x7-5; R.S.M. 1958.1.2376. Specimens in 4, 7, and 9 coated with sublimate of ammonium chloride. Figs. 5, 6, 8. Hadroblastus(l) benniei (Etheridge and Carpenter, 1886); Lower Limestone Group, Lower Carboniferous; Carlops, Peeblesshire, Scotland. 5, 8, inclined oral (centred on anal interarea) and lateral views of crushed specimen in xylene, x 4-8; R.S.M. 1958.1.2461. 6, left side of specimen in fig. 5 coated with sublimate of ammonium chloride, x 7-5. (R.S.M., Royal Scottish Museum; N.M.I., National Museum of Ireland.) PLATE 32 MACURDA, blastoids from Scotland 232 PALAEONTOLOGY, VOLUME 20 Radials five, A and B of similar shape; D limbs of C and E truncated, and D reduced in size due to bilateral symmetry. A and B radials pentagonal in plan view with narrow base (0-52 mm), lateral sides diverge rapidly outward to maximum width at aboral tip of deltoids (2-92 mm), upper edges extend into ambulacral sinus at a right angle to axis of sinus. Radials A and B triangular in lateral view, being slightly convex from origin of radial outward in both an aboral and oral direction; lower edge slightly concave. RB sector narrow, straight both parallel and perpendi- cular to RB axis; RB growth front straight (RB 1-64 mm; RBF 0-26 mm). RB sector merges smoothly with RR sector. Latter is very slightly convex parallel to and strongly convex perpendicular to RR axis; RR growth front straight (RR 1-20 mm; RRF 2-84 mm). RB and RR sectors ornamented with nodose ornament or discontinuous linear ridges. No growth lines visible. RD sector at sharp angle to RR sector, forming part of ambulacral sinus, slightly convex parallel to and straight perpendicular to RD axis. RD growth front very slightly convex (RD 1-60 mm; RDF 1-28 mm). Outer part of ambulacral sinus ornamented with nodose ornament; inner part smooth. AB interarea slightly larger than AE or BC. C and E radials similar except RB sectors much shorter with much longer radial-basal suture, and D side of each is about one-half the size of the other half due to the truncate D ambulacrum. No ambulacral sinuses along D sides of C and E radials. D limb of E radial in contact with DE deltoid along narrow suture bordering ambula- crum; D limb of C radial in contact with hypodeltoid. D radial small, relatively flat. RB sectors (RB 1-44 mm; RBF 0 04 mm) and lower half of RR sectors similar in size to those of A and E radials but upper part of RR sectors and RD sectors much shortened and reduced, bordering reduced D ambulacrum (RR 1-08 mm; RRF 1-36 mm; RD 0-72 mm; RDF 0-60 mm). D radial beneath ambulacrum and border- ing limbs of C and E radials; discontinuous linear ridges perpendicular to radial and radial-basal sutures (PI. 32, figs. 1, 2). Deltoids four; three of equal size, DE reduced. Regular deltoid rhombic in plan view, convex in lateral view. Deltoid lip small, hexagonal with edge bordering oral opening and interdeltoid sutures bearing ambulacral tract, sloping downward towards plate edges. Deltoid lip constricts aborally along deltoid-ambulacral suture to adoral end of hydrospire clefts. Central portion of deltoid lip flat, being higher than ambulacral tracts and hydrospire clefts. Deltoid body rhombic in outline, expanding aborally from narrow constriction of deltoid at adoral ends of hydrospire clefts. Body bordered laterally by hydrospire clefts and aborally by radiodeltoid sutures; all of these edges straight (Body L. 1-80 mm). Plane of radiodeltoid sutures nearly horizontal. Deltoid body convex lengthwise, with greatest height being in median position along a rounded crest. Axis of crest may be ornamented by nodes along centre of crest (PI. 31, fig. 1) or these may merge to be a semicontinuous ridge. Sides of deltoid body slope steeply downward into ambulacral sinus, ornamented by nodes which may form a row paralleling the crest. Broad growth lines may be visible parallel to radiodeltoid suture. Body smooth near ambulacra. As on radial, a node has a blunt top but the base is fluted by parallel grooves. (Del. L. 2-40 mm; Gr. Ad. W. 0-56 mm; Min. W. 040 mm; Gr. Ab. W. 1-28 mm; DR 1-40 mm.) DE deltoid smaller, asymmetrical. Deltoid lip as for other deltoids, body much reduced because of asymmetry. Form of DE deltoid body similar to that of regular D. B MACURDA, JUN.: CARBONIFEROUS BLASTOI DS FROM SCOTLAND 233 body but about one-third its size; radiodeltoid suture with D radial long whereas that with E radial short. Furthest aboral extension of deltoid thus near to ambulacrum (Body L. 0-64 mm; Gr. Ab. W. 0-52 mm). Anal deltoids four: a superdeltoid bordering oral opening, two cryptodeltoids (internal) and a hypodeltoid. Size of superdeltoid similar to deltoid lip of regular deltoid. C cryptodeltoid large, D cryptodeltoid small. Hypodeltoid a relatively large plate, pentagonal, with concave adoral edge bordering anus, two straight lateral edges bordering C and D ambulacra, and slightly convex radial-hypodeltoid sutures; D side of plate larger. Plate relatively flat, slopes downward from D to C ambula- crum. Some nodal ornament (HD L. 048 mm; W. 088 mm). Ambulacra five; D much shortened and reduced. Normal ambulacra linear in plan view, convex in lateral view. A and B ambulacra (L. 2 0 mm) curve through 90° while C and E ambulacra (L. 2-4 mm) curve through 120 and are visible in aboral view. Aboral tips of C and E ambulacra more closely approach each other (1-40 mm) than do aboral tips of A and B ambulacra (2-20 mm; PI. 32, fig. 7). A and B ambulacra in ambulacral sinus. B side of C ambulacrum and A side of E ambula- crum also have sloping wall of sinus but D sides flush with bordering radials. Ambu- lacra convex in cross-section (W. 0-68 mm) with median depression along ambulacral tract. Lancet exposed along median adoral two-thirds of ambulacrum. Brachiolar facets along lateral sloping sides of ambulacrum (PI. 31, fig. 5). Side plates seven per mm, quadrate, with slightly convex admedial edge and straight aboral edge. Adoral abmedial edge of plate embayed by triangular outer side plate. Edge of side plate against next adoral side plate straight as is adoral suture against outer side plate. Laterally side plate tapers to a point or narrow rounded tip. Outer side plates widest along lateral edge of ambulacrum. Brachiolar facet ovoid, formed equally from side plate (aboral half) and outer side plate (adoral half) (L. 0-14 mm; W. 0-12 mm). D ambulacrum short (L. 1-2 mm), lanceolate in plan view (W. 0-48 mm), flat in lateral view, convex in cross-section. Less than half of lancet exposed in adoral half of ambulacrum. Ambulacra separated from adjacent radials and deltoids by hydrospire clefts which extend full length of the ambulacrum and provide entrance to the ten hydro- spire groups. Two hydrospires per group except in those of anal interarea and E side of D ambulacrum, where there is only one. Oral opening pentagonal, bordered by four deltoids and superdeltoid ( W. 028 mm). Adoral edges of plates bearing minor grooves of ambulacral tract. Remarks. The description of the external morphology is derived from R.S.M. 1958.1.2376 (PI. 32, figs. 7, 9), and measurements cited above pertain to this specimen. A previous repository and locality citation for this specimen (Breimer and Macurda 1972, pi. XI, figs. 4, 7) was incorrect. Details of the super- and cryptodeltoids and hydrospires are summarized from Breimer and Macurda 1972. Abbreviations for plate descriptions and measurements are given in Breimer and Macurda 1972. Micrographs were taken with a Japan Electron Optic-Laboratory Model JSM-U3 scanning electron microscope. Specimens were coated with 500 angstroms of gold and photographed with Polaroid PN type 55 film at 1 5 KV. 234 PALAEONTOLOGY, VOLUME 20 Family neoschismatidae Wanner, 1940 Genus hadroblastus Fay, 1962 Type species. H. convexus Fay, 1962. Hadroblastus(l) benniei (Etheridge and Carpenter) Plate 32, figs. 5, 6, 8 1886 Phaenoschisma benniei Etheridge and Carpenter, p. 278, pi. II, fig. 37; pi. IV, figs. 5, 6. 1972 ? Phaenoschisma benniei Breimer and Macurda, p. 18, pi. V, figs. 8, 11, 13. Scottish material. British Museum (Natural History) specimen numbers E666, El I 14, El 115; Royal Scottish Museum specimen number 1958.1.2461. Description. Theca small, biconical, vault and pelvis subequal. Pelvis conical, broad; vault low, with break in profile at junction of radial and deltoids. Cross-section pentagonal (at angular tip of ambulacra); greatest width median at aboral tip of ambulacra. Height and width subequal, near 10 mm or less. Basals one-half of conical pelvis, pentagonal in plan view, very slightly concave in lateral profile. Stem cicatrix small, at very base of theca. Azygous basal not preserved. Zygous basal pentagonal in plan view with straight edges. Median and lateral BR sectors very slightly concave parallel to BR axis; median convex, lateral straight normal to BR axis. Adjacent sectors merge smoothly. Growth lines strong. Typical ZBL 1-8 mm; ZBW 2 0 mm; ZBOPt 2T mm; ZBBR T7 mm; ZBBRF 1 0 mm. Radial pentagonal in plan view. Lower edge radial convex, lateral edges slightly convex, top straight, slightly concave inward. In plan view ambulacral sinus occupies one-half of upper part of radial. In lateral profile radial V-shaped with upper two edges subequal, straight profile, lower edge concave. RB sector straight or very slightly convex parallel to RB axis, slightly concave normal to it, merges smoothly with RR sector. Latter straight parallel to RR axis and convex normal to it. RD sector grew mostly within an ambulacral sinus, straight parallel and normal to RD axis. Narrow external RD growth sector borders ambulacral sinus. Striated ornament (growth lines) paralleling sutures in RB, RR, and external part of RD sectors; small double U-shaped lip at aboral end of ambulacrum at origin of radial. Hydrospires occupy full width RD sector until eight or nine hydrospires formed. Deltoid hexagonal in plan view with straight DD, concave DAF, and straight DRF. Deltoid slightly convex to very slightly concave in lateral profile, with crest forming dominant part of plate. Adoral edge of plate (deltoid lip) prominent, with adorally directed V-shaped rim bordering ambulacral tract. Deltoid crest originates just behind small depression paralleling adoral rim. Crest sharp, directed downward from oral opening. Sides of crest slope moderately downward, with hydrospire slits across DRF until eight to nine hydrospires formed. Formation of hydrospires then ceases and deltoid crest bifurcates in median part of deltoid body to form two aborally diverging ridges which border ambulacral sinus. Median aboral part of deltoid body rhombic, concave in cross-section. Number of anal deltoids unknown. Relatively large hypodeltoid with external growth sector present. Hypodeltoid pentagonal, with straight lateral borders and D. B. MACURDA, JUN.: CARBONIFEROUS BLASTOIDS FROM SCOTLAND 235 convex hyporadial sutures, each with an aborally directed angular bend in the middle. Ambulacra five, sublanceolate in plan view, slightly convex in lateral view extend- ing half-way down theca. Ambulacrum convex in cross-section, with median depres- sion along ambulacral tract. Lancet narrowly exposed, apparently over most of the ambulacral length. Three side plates per mm. Aboral lateral edge of each side plate embayed by a large triangular outer side plate. Side plate pentagonal with straight adoral edge which intersects median line of ambulacrum at an angle ; convex admedial edge; straight aboral medial edge which parallels adoral edge; straight aboral lateral edge against outer side plate, and narrow, slightly convex lateral edge which forms part of outer edge of ambulacrum. Large elliptical brachiolar facet on outer sloping side of ambulacrum (PI. 32, fig. 6), formed equally on side plate (aboral half) and outer side plate (adoral half). Arcuate trough borders aboral margin of brachiolar facet. Four or five minor grooves per side plate border main ambulacral groove; present also on adoral margin of ambulacral side grooves. Ten(?) hydrospire groups. Maximum of eight to nine in regular groups, bordering ambulacrum in shallow ambulacral sinus. Some visible on C side of anal interarea, presumed to be present on D side as well (R.S.M. 1958.1.2461). Number apparently reduced in anal interarea. Stereomic microstructure is well preserved in some specimens. Remarks. The above description is a composite taken from specimens in the Royal Scottish Museum and the British Museum (Natural History). The specimens are crushed and fragmentary, precluding accurate measurement, and only a few have been recovered. Etheridge and Carpenter (1886) originally described this species as Phaenoschisma benniei. Breimer and Macurda (1972) listed this assignment with a question mark. Restudy and further preparation of the material from the Royal Scottish Museum revealed agreement on almost all determinable characters with the definition for Hadroblastus given by Breimer and Macurda (1972). The number of anal deltoids is unknown in //.(?) benniei and the form of the aboral part of the deltoid body is different from the North American species of Hadroblastus. Lack of information on the anal deltoids would appear to be the only character precluding definite assign- ment to Hadroblastus. An incomplete blastoid from beds of C2S age near Dublin, Ireland was questionably identified as Hadroblastus sp.? by Breimer and Macurda (1972). If assignable to Hadroblastus , //.(?) benniei would represent the youngest occurrence of the genus, being D2 in age. Acknowledgements. I wish to thank Dr. C. D. Waterston (Royal Scottish Museum, Edinburgh) for the loan of specimens from the Wright Collection and Drs. R. P. S. Jefferies and H. Owen, British Museum (Natural History), London for access to collections in their care. Part of this research was conducted under National Science Foundation Grant GB-5802. Scanning electron microscopy was conducted in the Scanning Electron Microscope Laboratory, University of Michigan, Dr. W. C. Bigelow, Director. 236 PALAEONTOLOGY, VOLUME 20 REFERENCES Austin, t. and Austin, t., Jun. 1843. Descriptions of several new genera and species of Crinoidea. Ann. Mag. nat. Hist. 1st ser., 11, no. 69, 195-207. breimer, a. and macurda, d. b., Jun. 1972. The phylogeny of the fissiculate blastoids. Verli. K. ned. Akad. wet.. Afd. Natuur. Eerste Reeks 26(3), 390 pp., pis. I-XXXIV, 104 figs. bronn, h. G. 1848. Index palaeontologicus. Nomenclator palaeontologicus. Stuttgart, E. Schweizerbart, 1381 pp. etheridge, r., Jun. 1876. On the occurrence of the genus Astrocrinites (Austin) in the Scottish Carboni- ferous Limestone series; with the description of a new species (A. I benniei ) and remarks on the genus. Q. Jlgeol. Soc. Lond. 32, 103-115, pis. 12-13. and carpenter, p. H. 1886. Catalogue of the Blastoidea. London. Brit. Mus. (Nat. Hist.), 322 pp., pis. 1 XX, 8 figs. george, t. n. 1971 . The classification of the Lower Carboniferous rocks. In george, t. n. and black, w. w. Lower Carboniferous (Dinantian). Lex. Strat. Inter. 1, Fasc. 3a VIE PP- 1-16. jensen, M. 1972. The ultrastructure of the echinoid skeleton. Sarsia , 48, 39-47, pis. 1 11. kesling, R. v. and chilman, r. b. 1975. Strata and megafossils of the Middle Devonian Silica Formation. Mus. Paleont. Univ. Mich. Papers on Paleon. 8, 408 pp., pis. 1-141. macurda, D. b., Jun. 1973. The stereomic microstructure of the blastoid endoskeleton. Contr. Mus. Paleont. Univ. Mich. 24, 69-83, pis. 1-8. 1976. Skeletal modifications related to food capture and feeding behavior of the basketstar Astro- phyton. Paleobiol. 2, 1-7. — and meyer, D. L. 1975. The microstructure of the crinoid endoskeleton. Paleont. Contrib. Univ. Kansas, 74, 1-22, pis. 1-30. morris, j. 1843. A catalogue of British fossils comprising all the genera hitherto described; with references to their geological distribution and to the localities in which they have been found. London, Van Voorst, 222 pp. wright, j. 1941. Allagecrinus biplex Wright— a revision of the species, with notes on some other Scottish Allagecrinidae. Geol. Mag. 78, 293-304, pi. VII. DONALD B. MACURDA Jun. Museum of Paleontology The University of Michigan Ann Arbor, Michigan 48109 U.S.A. Typescript received 5 March 1976 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology. Membership is open to individuals and to institutions on payment of the appropriate annual subscription: Institutional membership .... £25-00 (U.S. $50.00) Ordinary membership .... £8 00 (U.S. #16.00) Student membership .... £5-00 (U.S. $10.00) There is no admission fee. Institutional membership is only available by direct application, not through agents. 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. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London NW1 4NS, England. PALAEONTOLOGY All members who join for 1977 will receive Volume 20, Parts 1-4. All back numbers are still in print and may be ordered from B. H. Blackwell, Broad Street, Oxford OX1 3BQ, England, at £9 00 per part (post free). A complete set. Volumes 1-19, consists of 75 parts and costs £675 00. SPECIAL PAPERS IN PALAEONTOLOGY The subscription rate is £15 (U.S. $30.00) for Institutional Members and £7-50 (U.S. $15.00) for Ordinary and Student Members. Individual subscriptions should be placed through the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NW1 4NS, England. Ordinary and Student members only may obtain individual Special Papers from Dr. Rose at reduced rates. Non-members may obtain them at the stated prices from B. H. Blackwell, Broad Street, Oxford OX1 3BQ, England. COUNCIL 1976-1977 President : Professor W. G. Chaloner, Department of Botany, Birkbeck College, London WC1E 7HX Vice-Presidents : Dr. J. D. Hudson, Department of Geology, The University, Leicester LEI 7RH Dr. J. M. Hancock, Department of Geology, King’s College, Strand, London WC2R 2LS Treasurer : Mr. R. P. Tripp, High Wood, West Kingsdown, Sevenoaks, Kent TN15 6BN Membership Treasurer: Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London NW1 4NS Secretary : Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne NE1 7RU Editors Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Dr. C. P. Hughes, Department of Geology, Sedgwick Museum, Cambridge CB2 3EQ Professor J. W. Murray, Department of Geology, The University, Exeter EX4 4QE Professor C. B. Cox, Department of Zoology, King’s College, Strand, London WC2R 2LS Other Members of Council Dr. M. C. Boulter, London Dr. C. H. C. Brunton, London Dr. J. C. W. Cope, Swansea Dr. G. E. Farrow, Glasgow Dr. G. P. Larwood, Durham Dr. C. R. C. Paul, Liverpool Dr. J. E. Pollard, Manchester Dr. R. E. H. Reid, Belfast Dr. R. B. Rickards, Cambridge Dr. A. W. A. Rushton, London Dr. E. B. Selwood, Exeter Dr. G. D. Sevastopulo, Dublin Dr. P. Toghill, Church Stretton Dr. P. G. Wallace, London Overseas Representatives Australia : Professor B. D. Webby, Department of Geology, Sydney University, Sydney, N.S.W., 2006 Canada : Dr. B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW„ Calgary, Alberta India: Professor M. R. Sahni, 98 Mahatma Gandhi Marg, Lucknow (U.P.), India New Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt West Indies and Central America : Mr. John B. SaCnders, Geological Laboratory, Texaco Trinidad, Inc.-, Pointe-a-Pierre, Trinidad, West Indies Western U.S. A.: Professor J. Wyatt Durham, Department of Paleontology, University of California, Berkeley 4, California Eastern U.S. A. : Professor J. W. Wells, Department of Geology, Cornell University, Ithaca, New York South America: Dr. O. A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela Palaeontology VOLUME 20 PART 1 CONTENTS Structure and incremental growth in the ahermatypic coral Desmophyllum crist agalli from the North Atlantic J. E. SORAUF and J. S. JELL 1 The laminae and cuticular organization of the trilobite Asaphus raniceps J. E. DALINGWATER and J. MILLER 21 Calcified Pleclonema (blue-green algae), a Recent example of Girvanella from Aldabra Atoll R. RIDING 33 Significance of coiled protocoralla in some Mississippian horn corals w. J. SANDO 47 Coprolites containing plant material from the Carboniferous of Britain A. C. SCOTT 59 Classification of the trilobite Pseudagnostus J. H. SHERGOLD 69 Two new Bajocian microconch otoitid ammonites and their significance C. F. PARSONS 101 Some Phacopina (Trilobita) from the Silurian of Scotland E. N. K. CLARKSON, N. ELDRIDGE, and J.-L.. HENRY 119 Evolution of the charophyte floras in the upper Eocene and lower Oligocene of the Isle of Wight M. FEIST-C ASTELL 143 Classification and phylogeny of homalonotid trilobites A. T. THOMAS 159 Dinoflagellate cysts from the Bearpaw Formation (?upper Campanian to Maastrichtian) of Montana R. HARLAND ] 79 A new Eocene shark from the London Clay of Essex H. cappetta and d. j. ward 195 The giant crocodilian Sarcosuchus in the early Cretaceous of Brazil and Niger E. BUFFETAUT and P. TAQUET 203 New Carboniferous stenoscismatacean brachiopods from Oviedo and Leon, Spain M. L. MARTINEZ-CHACON 209 Two Carboniferous blastoids from Scotland D, B. MACURDA, JUN. 225 Printed in Great Britain at the University Press , Oxford by Vivian Ridler, Printer to the University Palaeontology VOLUME 20 PART 2 MAY 1977 Published by The Palaeontological Association ■ London Price £9 THE PALAEONTOLOGICAL ASSOCIATION The Association publishes Palaeontology and Special Papers in Palaeontology. Details of member- ship and subscription rates may be found inside the back cover. PALAEONTOLOGY 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 on all aspects of palaeontology and stratigraphical palaeontology are invited. They should conform in style to those already published in this journal, and should be sent to The Secretary, P.A. Publications Committee, Department of Geology, Sedgwick Museum, Downing Street, Cambridge, CB2 3EQ, England, who will supply detailed instructions for authors on request (these were published in Palaeontology , 15, pp. 676-681). SPECIAL PAPERS IN PALAEONTOLOGY This is a series of substantial separate works. Members may subscribe to the Series; alternatively. Ordinary and Student members only may obtain individual copies at reduced rates. The following Special Papers are available : 1. (for 1967): Miospores in the Coal Seams of the Carboniferous of Great Britain, by a. h. v. smith and m. a. butterworth. 324 pp., 72 text-figs., 27 plates. Price £8 (U.S. $16.00), post free. 2. (for 1968) : Evolution of the Shell Structure of Articulate Brachiopods, by a. williams. 55 pp., 27 text- figs., 24 plates. Price £5 (U.S. $10.00). 3. (for 1968): Upper Maestrichtian Radiolaria of California, by Helen p. foreman. 82 pp., 8 plates. Price £3 (U.S. $6.00). 4. (for 1969): Lower Turanian Ammonites from Israel, by R. freund and M. raab. 83 pp., 15 text-figs., 10 plates. Price £3 (U.S. $6.00). 5. (for 1969): Chitinozoa from the Ordovician Viola and Fernvale Limestones of the Arbuckle Moun- tains, Oklahoma, by w. a. m. jenkins. 44 pp., 10 text-figs., 9 plates. Price £2 (U.S. $4.00). 6. (for 1969): Ammonoidea from the Mata Series (Santonian-Maastrichtian) of New Zealand, by r. a. henderson. 82 pp., 13 text-figs., 15 plates. Price £3 (U.S. $6.00). 7. (for 1970): Shell Structure of the Craniacea and other Calcareous Inarticulate Brachiopoda, by a. williams and a. d. wright. 51 pp., 17 text-figs., 15 plates. Price £1-50 (U.S. $3.00). 8. (for 1970): Cenomanian Ammonites from Southern England, by w. j. Kennedy. 272 pp., 5 tables, 64 plates. Price £8 (U.S. $16.00). 9. (for 1971): Fish from the Freshwater Lower Cretaceous of Victoria, Australia, with Comments on the Palaeo-environment, by m. waldman. 130 pp., 37 text-figs.. 18 plates. Price £5 (U.S. $10.00). 10. (for 1971) : Upper Cretaceous Ostracoda from the Carnarvon Basin, Western Australia, by R. h. bate. 148 pp., 43 text-figs., 27 plates. Price £5 (U.S. $10.00). 11. (for 1972): Stromatolites and the Biostratigraphy of the Australian Precambrian and Cambrian, by m. r. Walter. 268 pp., 55 text-figs., 34 plates. Price £10 (U.S. $20.00). 12. (for 1973): Organisms and Continents through Time. A Symposium of 23 papers edited by n. f. hughes. 340 pp., 132 text-figs. Price £10 (U.S. $20.00) (published with the Systematics Association). 13. (for 1974): Graptolite studies in honour of O. M. B. Bulman. Edited by R. b. rickards, d. e. jackson, and c. p. hughes. 261 pp., 26 plates. Price £10 (U.S. $20.00). 14. (for 1974): Palaeogene Foraminiferida and Palaeoecology, Hampshire and Paris Basins and the English Channel, by J. w. Murray and c. a. wright. 171 pp., 45 text-figs., 20 plates. Price £8 (U.S. $16.00). 15. (for 1975): Lower and Middle Devonian Conodonts from the Broken River Embayment, North Queensland, Australia, by p. g. telford. 100 pp., 9 text-figs., 16 plates. Price £5-50 (U.S. $11.00). 16. (for 1975): The Ostracod Fauna from the Santonian Chalk (Upper Cretaceous) of Gingin, Western Australia, by J. w. neale. 131 pp., 40 text-figs., 22 plates. Price £6-50 (U.S. $13.00). 17. (for 1976): Aspects of Ammonite Biology, Biogeography, and Biostratigraphy, by w. j. Kennedy and w. a. cobban. 94 pp., 24 text-figs., 1 1 plates. Price £6 (U.S. $12.00). 18. (for 1976): Ostracoderm Faunas of the Delorme and Associated Siluro-Devonian Formations, North West Territories, Canada, by d. l. dineley and e. j. loeffler. 218 pp., 78 text-figs., 33 plates. Price £20 (U.S. $40.00). © The Palaeontological Association , 1977 Cover: A Middle Jurassic megaspore, Minerisporites richardsoni (Murray) Potonie, from Oxfordshire (x 350, S.E.M. by T. M. Windle). EVOLUTION IN CARNIVOROUS MAMMALS by R. J. G. SAVAGE Nineteenth Annual Address , delivered 17 March 1976 In tauros Libyci ruunt leones; Non sunt papilionibus molesti. (In Africa lions attack bulls; they do not molest butterflies.) Martial Abstract. Carnivorous mammals are identified on the basis of their carnassial dentition and their predatory preference for vertebrates. Jurassic pantotheres possess a dentition from which arose both carnivore and herbivore specializations. The changes towards carnivorous feeding involved emphasis on the cutting functions of the cheek teeth, with corresponding reduction of the crushing activities. A carnivorous mode of living evolved twice among the marsupials (borhyaenids in South America and some dasyurids in Australia) and twice among the placentals (Creodonta and Carnivora). Characteristics of the post-cranial skeleton in carnivores are mentioned. Features of the jaw mechanics, the brain, and the senses (smell, sight, and hearing) are discussed. Specializations of sabre-like canines and crushing premolars are discussed. The form and function of the carnassial dentition and its evolution are analysed. The evolutionary history of elephants and horses is well documented yet, despite a reasonable fossil record, there is surprisingly little information readily available on the evolutionary history of carnivores. Detailed taxonomic accounts of faunas and phylogenetic studies of families abound, but the literature on functional aspects of carnivore evolution is meagre. Three papers are outstanding exceptions to this generalization; Denison (1938) on creodonts, Simpson (1941) on sabre-toothed carnivores, and Crusafont and Truyols (1956) on the evolution of carnassial teeth. THE LIVING CARNIVORE Let us examine a lion as a living representative of one of the peaks of carnivore evolution and see what distinguishes it from the herbivorous gazelles it preys upon and also from its shrew-like ancestors of the earliest Tertiary. What in other words, starting from a shrew, would be the modifications required to produce a lion? To earn a living as a carnivore, the first essential is to locate your victim — before he locates you. This means having very acute sensory receptors (vision, hearing, and smell), and a first-class brain to integrate rapidly and efficiently and to analyse the information. In the lion, as often in carnivores, sight is the most important sense, with the eyes set well forward giving stereoscopy. Hypermetropic vision, and ability to judge distances accurately, enable the lion to locate its victim precisely. The next move is to creep up close stealthily; ability to move silently close to the ground and to take maximum advantage of cover is essential. Then, to make con- tact, a burst of very high speed is needed; long legs with muscles aligned to give maximum mechanical advantages for speed and physiologically adapted for rapid [Palaeontology, Vol. 20, Part 2, 1977, pp. 237-271] A 238 PALAEONTOLOGY, VOLUME 20 contractions, ensure success. The attack on the victim must be swift to be effective; usually the gazelle is thrown off balance with a sideways crash which enables the lion to plunge its canine teeth into the throat region while holding the prey down with the claws of its forepaws. The carnivore will then rip open the carcass, using incisors and canines. The meat is torn off and cut into chunks, with the powerful closing muscles of the jaws operating the scissor-like carnassial blades of the cheek teeth. But even with all these refinements, the lion has on average only about a one in five chance of success; it is clearly not a game for amateurs. There is no sharp distinction between insectivore and carnivore. Neither is a single natural order of mammals; they reflect modes of living. There are some ten orders of mammals that are basically insectivorous, and about twenty orders that are essentially herbivorous. There remain only two that are fundamentally carni- vorous, the Creodonta and the Carnivora, with the Marsupialia evolving two carnivorous stocks from insectivorous ancestors. In this review we shall omit the Cetacea, which would make the fourth carnivorous order; they are so specialized as marine aquatic mammals that the carnivore specializations are relatively secondary and quite unlike those of land mammals. The two prime essentials of a carnivore are that its diet include a substantial pro- portion of vertebrates, and that it possesses a carnassial dentition. By specifying vertebrates, we can eliminate the insectivores, though this also leaves out animals like the mollusc-feeding sea otter Enhydra , and, of course, those members of the order Carnivora which have become secondarily herbivorous, such as the pandas and some of the bears. A few herbivores have produced omnivorous stocks that are partially carnivorous; for example, some pigs and the extinct mesonychids have large incisor or canine tusks and may supplement their diet with animal food, but they never achieve a carnassial dentition which is the hallmark of a truly carnivorous land mammal. When one or more pairs of upper and lower cheek teeth develop a blade structure, such that on closing the jaws these blades pass over each other like scissor blades, this is spoken of as carnassial specialization (text-fig. 1). Slicing teeth develop in some herbivores, e.g. pyrotheres, deinotheres, barytheres, and macro- podids; these are always on transverse lophs, extend along all or most of the cheek dentition, and are not self-sharpening, so that with wear the dentition becomes flattened. The extinct marsupial Thylacoleo is a puzzling exception; it has a "carnas- sial’ specialization, but otherwise retains many phalanger characters, which makes it difficult to interpret it as anything but a specialized herbivore (Gill 1954). THE ANCESTRY OF CARNIVOROUS MAMMALS The mammals of the Jurassic were essentially small insectivorous beasts and probably nocturnal. Multituberculates were the exception in being specialized rodent-like herbivores, but an aberrant sideline destined to leave no descendants. By mid Cretaceous times we can recognize, in the main stream of mammalian evolution, a fork into marsupial and placental stocks. While this cleft is based fundamentally on reproductive differences, some dental characters enable us to detect it ; for example, the presence of certain stylar cusps on the outer margin of the upper molars is a marsupial feature unknown in placentals. Though our knowledge is as yet very SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 239 text-fig. 1. Views of the carnassial dentition of Felis. a, lateral view, P4 and Mx in relief ; b, internal view of P4 and Mx ; c, occlusal view of upper and lower cheek denti- tion, carnassial blades in heavy lines; dx, carnassial teeth at beginning of occlusion; d2, cheek dentition in full occlusion; e, cusps of P4: me metacone, pa paracone, pas parastyle, pr protocone; F, cusps of M1=pad paraconid, prd protoconid; G, position of glenoid jaw articulation with respect to a. slender, there is evidence that during mid Cretaceous times there were considerable floral changes. Table 1 (after Hughes 1976) shows the proportion of flowering plants (angiosperms) to other megaplants. In the earliest Cretaceous the gymnosperms and pteridophytes share the total in equal proportions. By latest Cretaceous the angio- sperms have taken over 90% of the flora, the number of gymnosperms has dropped dramatically, and although the ferns have actually increased in species numbers, their proportion of the total flora is drastically reduced. Angiosperms arose in the early Cretaceous and rapidly expanded ; this must surely have lead to a diversification table 1. Estimated number of megaplant species (after Hughes 1976). Pteridophyte % Gymnosperm % Angiosperm % 10 000 34 2 000 9 1 500 50 640 0-2 500 2 1 500 50 286 000 96 4 20 000 89 0 0 Recent Latest Cretaceous Earliest Cretaceous 240 PALAEONTOLOGY, VOLUME 20 of the insect fauna. The Cretaceous insect record is, however, very poor, and the most that can be said is that when good faunas are next seen, in early Tertiary, they show a major advance on Jurassic faunas, with a much greater diversity (Crowson et al. 1967). Insectivorous mammals usually include fruit, nuts, and berries in their diet; the much greater range of these, and probably also of insects, in the Late Cretaceous is likely to have been a critical factor in mammalian radiation. By Late Cretaceous times the mammals show considerable diversity, though still largely within the same niches of small sized insectivorous stocks. Within these faunas, however, we can detect the beginnings of several orders of Tertiary mammals, in particular the early trends towards herbivore and carnivore lineages. By terminal Cretaceous and earliest Tertiary, these lineages became distinct as the mammals invaded the niches left vacant by the dinosaur extinctions. The detailed stages of the take over are not known, though many speculations have been made. Small mammals could compete successfully with reptiles throughout the Mesozoic only by occupying niches unavailable to reptiles. Mammalian endothermy, even if relatively inefficient, would have given them an advantage over reptiles in temperate climates (and temperate climate microniches) enabling them to feed at night, so long as they could hide during the day from the predatory lizards and dinosaurs. The need to hide imposed size limitations on the mammals ; furthermore, it is not efficient for an insecti- vorous mammal to grow large unless bulk food is available. Insects in bulk only became possible food sources when colonial ants and termites evolved, probably in the Oligocene. Mammalian eyes are primarily designed for nocturnal vision, since the receptor cells are predominantly rods; they secondarily became adapted for diurnal vision by increasing the numbers of cones. Here again, mammals would have the advantage over the smaller dinosaurs, which may have been endothermic but are likely to have had poor night vision. When the diurnal niches became avail- able in the terminal Cretaceous, the mammals were ready to invade them. Gradual improvements in their physiology, in particular temperature control and metabolic rates, enabled them to increase in size and to acquire specialized trophic preferences. Tooth adaptations evolved in the herbivore lineages to crush food, and in the carni- vores to cut it into chunks small enough to swallow. Text-fig. 2 illustrates trends in mammalian molar teeth during the Mesozoic. While named genera are chosen as examples, this does not imply that they are on direct lines of descent, only that we can trace likely lineages from the stocks to which they belong. In the Jurassic, a pantothere such as Peramus has a cheek denti- tion with primitive tribosphenic pattern, that is to say upper and lower wedge-shaped triangular teeth that cut against each other. A detailed and lucid analysis of the functional principles of tribosphenic molars is to be found in Crompton and Sita- Lumsden (1970). Peramus has a good insectivorous dentition, capable of cutting up small and relatively soft food. From the Early Cretaceous of Texas come Holo- clemensia and Pappotherium , which probably respectively belong to primitive marsupial and placental stocks. Their teeth are rather similar and functionally show a considerable change from those of Peramus; they have acquired a protocone cusp on the inner side of the upper molar, and a talonid or basin has developed in the lower molar to receive the impact of the protocone on occlusion. Thus a crushing action is now added to the slicing action of the Jurassic mammals. By the Late text-fig. 2. Evolution of carnassial teeth ; lower molars stippled and showing the overlap of upper and lower teeth in full occlusion. 242 PALAEONTOLOGY, VOLUME 20 Cretaceous, the marsupial Alphadon and the placental insectivore Cimolestes are in their dental pattern almost unchanged descendants; the marsupials are dis- tinguished by having a broad stylar shelf on the labial side. The ratio of crushing/ slicing varies in different taxa. Crompton and Hiiemae (1970) have estimated that in the living opossum Dide/phis , which differs little from the Late Cretaceous mar- supials, 60% of the chewing time is spent on pulping food with their sharp cusps and 40% on cutting. In the early Tertiary, both marsupial and placental stocks evolved carnivorous lineages (see Table 2 and text-fig. 18). From didelphid marsupials in South America table 2. Families of terrestrial carnivorous mammals. Total Stratigraphic Geographic Extinct Living no. of Orders Families range distribution genera genera genera MARSUPIALIA (Didelphidae) L. Cret.-Rec. N. & S. Am., Eu. (1) (1) Borhyaenidae U. Pal.-Plio. S. Am. 25 - 25 (Dasyuridae) U. Oligo.-Rec. Australia 2 2 4 (Phalangeridae) Mio.-Rec. Australia (1) - (1) Total MARSUPIALIA 29 2 31 CREODONTA Oxyaenidae U. Pal. U. Eo. N. Am., Eu., As. 10 - 10 Hyaenodontidae L. Eo.-U. Mio. N. Am., Eu., As., Af. 35 - 35 Total CREODONTA 45 - 45 CARNIVORA Miacidae M. Pal.-L. Oligo. N. Am., Eu., As. 16 _ 16 Amphicyonidae U. Eo.-U. Mio. N. Am., Eu., As., Af. 29 - 29 Canidae M. Oligo.-Rec. N. & S. Am., Eu., As., Af. 22 13 35 Ursidae U. Eo - Rec. N. & S. Am., Eu., As., Af. 22 6 28 Procyonidae L. Mio. Rec. N. & S. Am., Eu.. As., Af. 14 8 22 Mustelidae L. Oligo.-Rec. N. & S. Am., Eu., As., Af. 65 29 94 Viverridae U. Eo.-Rec. Eu., As., Af. 11 36 47 Hyaenidae M. Mio.-Rec. Eu., As., Af. 8 3 11 Felidae U. Eo.-Rec. N. & S. Am., Eu., As., Af. 31 3 34 Total CARNIVORA 218 98 316 Total MARSUPIALIA, CREODONTA and CARNIVORA 292 100 392 Note. Marsupial families in brackets are mainly non-carnivorous, and only the carnivorous genera are counted. arose the borhyaenids, with twelve known genera ranging from Palaeocene to Pliocene times. Within the dasyurid marsupials of Australia arose carnivorous taxa such as Sarcophi/us (native cat) and Thylacinus (native wolf); we know dasyurid history only back to the Miocene, though its origins are certainly earlier. From placental insectivores two carnivorous lineages differentiated during the Palaeocene, the Creodonta and the Carnivora ( sensu stricto). The classification of the higher categories of placental carnivores has undergone major changes over the past decade, and no widely accepted grouping has yet emerged. The degree of change is apparent when the classification of Simpson (1945), which has for long been the accepted classification, is compared with the one used here (Table 3). The proposed grouping is not as radical as might at first sight appear; it is largely a tidying-up operation following detailed studies on some groups. For almost a SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 243 decade it has been accepted that the Arctocyonidae and Mesonychidae are con- dylarths and not creodonts; they lack carnassial specializations and have developed crushing dentitions. The only remaining creodonts are the truly carnivorous oxyaenids and hyaenodontids. The miacids are the accepted ancestral stock of living carnivores and, since both creodonts and miacids can be traced independently back to an insectivore ancestry in the Palaeocene, 1 consider that the Creodonta and Carnivora should have equal rank as orders. Savage (1957) suggested that the pinnipeds were probably diphyletic. This has been followed up by other authors, and new discoveries, especially that of the new sub-family Enaliarctinae (Mitchell and Tedford 1973), have added greatly to our knowledge of aquatic lineages. Unfortunately the phocids have a poor fossil record compared with the otariids; nevertheless it seems likely that the otariids arose via enaliarctines from canoid ancestors, with odobenines as a side line, while the phocids arose from a mustelid ancestor. Acceptance of this diphyletic origin means the disappearance of the pinnipeds, and so also the need for the term fissiped. I have grouped the carnivore families into five superfamilies; this is a fairly orthodox classification more like that of Gregory and Heilman (1939) than that of Simpson ( 1 945). I cannot identify a mustelid ancestor among the canoids, and the family seems more likely to have arisen directly from a miacid ancestor. It appears equally likely that felids and viverrids have independent origins from miacids. The Eocene faunas of carnivores form a complex of taxa which it is some- times difficult to assign with certainty to living families. table 3. Classifications of carnivores. simpson 1945 Superorder ferae Order carnivora Suborder creodonta Superfamily Arctocyonoidea Mesonychoidea Oxyaenoidea Suborder fissiped a Superfamily Miacioidea Canoidea Feloidea Suborder pinnipedia (Arctocyonidae) (Mesonychidae) (Oxyaenidae, Hyaenodontidae) (Miacidae) (Canidae, Ursidae, Procyonidae, Mustelidae) (Viverridae, Hyaenidae, Felidae) (Semantoridae, Otariidae, Odobenidae, Phocidae) Superorder ferae Order creodonta Superfamily Oxyaenoidea Order carnivora Superfamily Miacoidea Canoidea Musteloidea Viverroidea Feloidea savage 1977 (Oxyaenidae, Hyaenodontidae) (Miacidae) (Amphicyonidae, Canidae, Ursidae, Procyonidae, Otariidae) (Mustelidae, Phocidae) (Viverridae, Hyaenidae) (Felidae) 244 PALAEONTOLOGY, VOLUME 20 THE GEOMETRY OF THE FOSSIL RECORD There are almost 1000 genera of living mammals, and in trophic terms they can be grouped thus: carnivores 10%, insectivores 23%, herbivores 14%, omnivores 9%, gnawers 38%, and aquatic taxa 6%. Leaving aside the aquatic forms, carnivores today make up about 11% of all land mammal genera. Taking the fossil record into account, we know in all about 3000 mammalian genera, and of these 14% are carni- vores; this apparent rise is in part due to smaller mammals (rodents and insectivores) being less well represented in the fossil record. In Table 4 we see an analysis of Tertiary mammal faunas in North America, Africa, and South America. The North American record is good at all periods and shows that carnivores have ranged from 16 to 32% (mean 22%). The African record is much less complete and shows carni- vores with a slightly lower range, 12-28% (mean 18%). The South American record yields figures of 3-9% (mean 6%); this much lower range may in part be due to inadequate sampling, but also to the carnivore niche being in part occupied by the large phororhachid birds. If these birds are taken into account, the range for carni- vores becomes 3-13% (mean 9%). table 4. Percentages of carnivores in Cenozoic land mammal faunas. North America Africa Creodonta Carnivora Total Creodonta Carnivora Total Pleistocene 210 210 — 160 160 Pliocene 23-7 23-7 — 280 28-0 Miocene 31 -6 31-6 60 60 12-0 Oligocene 2-6 13-2 15-8 160 — 160 Eocene 140 5-3 19-3 17-3 — 17-3 Palaeocene 18-4 2-6 210 — — — South America Marsupial Placental Phororhachoid Total carnivores carnivores birds carnivores Pleistocene — 91 2-4 11-5 Pliocene 4-5 1-8 3-7 100 Miocene 6-3 — 7-0 13-3 Oligocene 6-3 — 5-2 115 Eocene 3-9 — — 3-9 Palaeocene 2-6 — — 2-6 Table 4 also illustrates how the Carnivora gradually replaced the Creodonta in North America, and similarly in Africa, though here the creodonts persisted until the Late Miocene. In South America the early Tertiary carnivores were marsupials; these were supplemented in the Oligocene by phororhachid birds, and finally the placental carnivores invaded that continent in the Pliocene and displaced both native stocks. Considering now the biomass relationships, the standing crop ratio of predators to prey is around 2-3%; for example, the Ngorongoro crater in Tanzania supports around 25 000 large herbivores and around 500 large carnivores. These ratios are SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 245 fairly constant for endotherms (Bakker 1972); in a Pliocene mammal community in Nebraska analysed by Voorhies (1969), the ratio was around 3%. Reasonably complete skeletal remains of carnivores are rare, for example the skeletons of only three borhyaenids and six creodonts have been reassembled. A fair number of skulls of most families are preserved, but the majority of taxa are known only from mandibular and maxillary fragments with incomplete dentitions. The Palaeocene record is very weak, with Europe and North America the only continents from which adequate faunas are known throughout the Ter- tiary. Carnivores have a greater size range than that found in any other mammalian order; they range from a weasel of around 50 gm to a Kodiak bear of around 800 kg. The weasel is about ten times larger than the smallest shrew; the largest bear is about a tenth the weight of an elephant. All small carnivores have to compete with the predatory birds, and the majority of carnivores are small to medium sized. Two-thirds of all living carnivore genera belong to two families, the Viverridae and Mustelidae. Omitting the aquatic otters, the mean weight for mustelids is 4-3 kg and for viverrids is 3-6 kg; rodents and insectivores are major items in their diets. The weasel can cope with small rodents such as mice and rats, but if it were any smaller the range of prey would presumably be too restricted and the energy expended in obtaining small food packages would exceed the intake. At the other end of the scale, carnivores cannot afford to become too large, as they pay a heavy price for this in loss of speed. In order to tackle prey much larger than themselves, carnivores hunt in groups, making use of their superior intelligence to overcome other short- comings. The largest known fossil carnivore is the hyaenodont Megistotherium. The skull is about twice the size of that of a large bear— and bears are omnivorous rather than truly carnivorous. It is possible that Megistotherium was a carrion feeder; it appears from the dentition to have been capable of crushing bone and cutting flesh (Savage 1973). The weight of Megistotherium is difficult to estimate as hyaenodonts tend to have heads which are large in proportion to their bodies. In 1973 I suggested that it might have weighed around 880 kg; this is little more than a guess, it could have been half or twice that figure. THE POSTCRANIAL SKELETON The postcranial skeleton of carnivores (text-figs. 3 and 4), excluding the fully aquatic pinnipeds, is less modified than that seen in many other mammals. Nevertheless, there are many differences within carnivores which also show many advances from the primitive insectivore skeleton. Stromer (1902), in a large monograph on the carnivore vertebral column, stated that the morphological details of the individual vertebrae exhibit no important features consistently correlated with major phyla and therefore are of little systematic importance; they reflect function and not relationship. Hildebrand (1961 ), in a paper on the body proportions of marsupials, noted that marsupials have a greater varia- tion in postcranial proportions than is found in placental mammals in general and text-fig. 3. Skeletal reconstructions: borhyaenid Prothylacinus (after Sinclair 1906); creodonts Sinopa (after Matthew 1906), Oxyaena (after Osborn 1900), and Hyaenodon (after Scott and Jepsen 1936); amphicyonid Daphoenus (after Scott and Jepsen 1936). text-fig. 4. Skeletal reconstructions: mustelid Potamotherium (after Savage 1957); canid Canis dims (after Merriam 1912); viverrid Viverra and hyaenid Hyaena (after Blainville 1864); felid Dinictis (after Scott and Jepsen 1936). 248 PALAEONTOLOGY, VOLUME 20 in canids in particular. Within the didelphids there is some correlation between body proportions and habit for aquatic and riparian species, but not for arboreal, semi- arboreal, or terrestrial species. In another paper, Hildebrand (1954) remarked that in the ungulate skeleton nearly every bone has been profoundly modified to adapt the animals for running; although some canids can overtake fast ungulates, none of their bones is so strikingly altered in response to cursorial habit. Nevertheless, most skeletal differences among canids can be attributed to variation in the degree of cursorial specialization. In carnivores, the prime essentials of the postcranial skeleton are strength, flexi- bility, and adaptability. Insectivores are scurrying creatures, carnivores are usually cursorial beasts. The carnivores need to move slowly and rapidly, to crouch down and stand up full stretch, to turn rapidly, and sometimes to climb, dig, or swim, and to use the limbs to clasp struggling prey. Most carnivores can achieve a galloping gait, a gait in which all four feet are off the ground simultaneously. As Smith and Savage (1956) have shown, the carnivore gallop is very different from that of herbi- vores. In carnivores the vertebral column and limbs are flexed, then unfolded like a spring to give the animal vertical and horizontal thrust, projecting it through a parabola with a long airborne phase (text-fig. 5a). Small herbivores may adopt the carnivore method of galloping but very large graviportal forms do not gallop, or do so only rarely. The majority of medium-sized herbivores gallop with a stiff vertebral column that scarcely rises, and most of the energy is expended in swinging the limbs (text-fig. 5b). text-fig. 5. Sequence of phases in the gallop of a dog and a horse (after Muybridge 1899). The vertebrae of carnivores, like the limb bones, are robust and well sculpted for the origins and insertions of muscles. The neural spines of the thoracic vertebrae are long, accommodating the origins of the strongly developed neck muscles. There is little variation in the number of presacral vertebrae, usually 27 zb 1 . The neck does not vary greatly in length, but the atlas is often diagnostic to family and sometimes generic level. The position of the vertebrarterial canal on the atlas has characteristic patterns and there is considerable variation in the shape of the wings; these are, for example, narrow and sagittally elongate in sabre-tooth cats. While there are usually 20 thoracic and lumbar vertebrae in carnivores, the proportions vary. The lumbars SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 249 vary from 4 to 7, though 5 or 6 is normal in both creodonts and living carnivores; the arboreal pandas often have only 4 lumbars. Savage (1957) demonstrated that the zygopophysial articulations of the thoracic vertebrae are orientated to allow great freedom of movement, and in particular to allow the anterior part of the body to twist relative to the posterior. This twisting ability is a uniquely mammalian characteristic, but is poorly developed in herbivores as compared to carnivores. The longer the thoracic region and the more posterior the anticlinal vertebra, the greater is the degree of flexibility. Creodonts have in general a less flexible column than living carnivores. The creodont lumbar vertebrae have tight articulations and very little freedom of movement; this is seen most markedly in Patriofelis, which has sigmoidal-shaped zygopophyses (text-fig. 6). Tail lengths are highly variable. text-fig. 6. Posterior view of lumbar vertebrae, showing extent of articulatory surface of posterior zygapophyses in heavy line. a, typical of Carnivora; b, typical of Creodonta; c, as found in Patriofelis (after Denison 1938). even within a species, as is common in mammals, but on the whole they tend to be long. All creodonts and marsupial carnivores have long tails ; among living carnivores only bears, giant pandas, and a few hyaenas and cats have notably short tails — and of these bears are omnivorous and giant pandas herbivorous. The cats with short tails are the lynxes and some sabre-toothed cats. Lynxes tend to be slow movers, often living in forests, where they climb in trees and feed on rabbits and hares. The sabre-toothed cat Smilodon, with its rather short hind legs, may not have been a fast runner. The marsupial sabre-tooth Thylacosmilus , however, appears to have had a reasonably long tail. Tails are used for balance when changing direction at high speed, for swimming in otters, and other occasional uses include its use as a prehensile organ for climbing as in procyonids. In small species, a bushy tail coiled round the body is important in keeping the animal warm. Much more variation is seen in the girdles and limbs than in the vertebral column. The scapula is usually fairly broad, with the supraspinous fossa often equal to or even larger than the infraspinous fossa; in ungulates the supraspinous fossa tends to become much the smaller. The scapular spine, acromion, and metacromion are usually well developed; they serve for the origins of muscles which abduct the limb, and in ungulates, with their specializations towards antero-posterior limb move- ments, these processes are reduced. The clavicle is vestigial or absent, as would be expected. In the pelvis the ilium and ischiopubis are usually of equal length; in 250 PALAEONTOLOGY, VOLUME 20 aquatic carnivores the ilium is much shortened and the ischiopubis lengthened, while the opposite trend is seen in ungulates. Smith and Savage (1956) demonstrated that this difference was explicable as an adaptation to powerful limb movements in carnivores and rapid movements in ungulates. The ilium in creodonts is usually expanded dorsally and has a strong longitudinal keel. Since creodonts have three sacral vertebrae as in living carnivores, the heavier ilium was not for greater pelvic rigidity, but rather for more massive gluteal musculature. This, combined with the development in creodonts of a third trochanter on the femur, is indicative of the strong abduction capabilities of the limb. In carnivores the hind limb is used almost solely for locomotion, but the fore limb has many additional uses— climbing, digging, grooming, tearing, and holding prey. Limbs can never be as specialized as in herbivores, where they are used almost exclusively for one mode of locomotion. The proportions of the limb segments have attracted the attention of many authors, but it is still difficult to provide more than broad generalizations. In order to compare both relative length and proportions of the limbs, I have compiled a graph (Table 5) where the limbs and their segments are expressed as a percentage of the vertebral column length, measured from the atlas to the last lumbar vertebra. In the less-specialized carnivores (e.g. Mustela or Viverra) the three segments of the limb become progressively shorter distally. In cursorial ungulates (e.g. Equus or Connochaetes ) the reverse is true. In cursorial carnivores the foot is always the shortest element and the other two segments approach equality in length. Hildebrand (1961) concluded that only aquatic adapta- tions (all segments short) were clearly differentiated on limb proportions. Arboreal and fossorial forms also tend to have short limbs. Hind limbs are usually about 10% longer than the fore limbs; hyaenas are the exception with fore limbs equal to, or longer than, the hind. We might define short-legged carnivores as those whose fore limbs are 40% or less of the vertebral column, and the hind limbs 50% or less; this would include the otters, some Mustela species, and the borhyaenid Cladosictis. If we define long-legged as 70% or more for the fore limbs and 80% or more for the hind limbs, then the candidates are some species of Fells and Canis, together with Acinonyx and Chrysocyon ; all these are cursorial and mostly high-speed runners. But Chrysocyon is not a particularly fast runner; increase in speed is bought at the expense of power. The very long legs of Chrysocyon are an adaptation to living in the long pampas grass. From our knowledge of locomotion in living carnivores, we can look at the fossils and make comparisons on limb proportions. The borhyaenid C/adosicitis has limb proportions almost identical with those of the living otter Lutra , and both animals are of similar size. Prothylacinus has the size and proportions of the living wolverine Gulo. Oxyaenids and proviverrines all have shortish limbs, especially the fore limbs; it seems probable that these are primitive character traits rather than aquatic specializations. Hyaenodon has longer limbs, though still shorter than those of living cursorial carnivores. The amphicyonids vary from fairly lightly-built forms to very large bear-like varieties. There are all gradations between plantigrade and digitigrade. On the whole the more primitive carnivores, arboreal, and slower-moving animals, have plantigrade feet, while all cursorial types have digitigrade feet, or are capable of achieving SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 251 table 5. Limb proportions of carnivorous mammals, shown as percentages of vertebral column length (atlas to last lumbar vertebra), with those of three ungulates for comparison. First joint (femur or humerus) and pes shown in black. Second joint (tibia-fibula or ulna-radius) stippled. HIND LIMB FORE LIMB I 1 1 1 1 1 1 1 1 I I I I I I I I I I I I I 100 90 80 70 60 50 40 30 20 10 0 0 10 20 30 40 50 60 70 80 90 100 BORHYAENIDAE CANIDAE Cladosictis Prothylacinus Didelphis Sarcophilus Thylacinus Oxyaena Patriofelis Palaeonictis Tritemnodon Sinopa Hyaenodon horridus Hyaenodon mustelinu Vulpavus Amphicyon Pseudocynodictis Daphoenus Spethos Canis familiarus Urocyon Alopex Vulpes Chrysocyon Ursus Potos Gulo Mustela Potamotherium Potamotherium Lutra Lutra Lutra Meles VIVERRIDAE FELIDAE HYAENIDAE Dinictis Smilodon Felis domesticus Felis leopardus Felis leo Felis leo Felis tigris Acinonyx Crocuta Crocuta Equus Cervus Connochaetes 100 90 80 70 60 50 40 30 20 10 0 0 10 20 30 40 50 60 70 80 90 100 i i i i i l I I I 1 I I I I I I I I I 1 1 1 UNGULATES 252 PALAEONTOLOGY, VOLUME 20 digitigrade posture, which considerably lengthens the limb and so adds to the mechanical advantage of the lever system. Borhyaenids and early creodonts were plantigrade, but later hyaenodontids were fully digitigrade (text-fig. 7). Plantigrade taxa have short, broad, and widely spread metapodials; in digitigrade taxa the metapodials are long and slender and held close together with strong tendons to give support when the foot is elevated. The scapho-lunar bones never fuse in creo- donts, as they do in living carnivores; the unfused condition might have given greater freedom of movement in the wrist joint. Living carnivores have lost this text-fig. 7. Right pes of a, plantigrade creodont Patriofelis (after Denison 1938) and b, digitigrade carnivore Felis (after Merriam and Stock 1932); each pes shown with one digit in lateral view, the articular facets emphasized with heavy line. freedom but, in compensation, the digitigrade foot is more firmly supported. The carnivore astragalus always has a good pulley articulation with the tibia, and in creodonts the astragalar foramen is always present. There are never more than four functional digits, with the first variously reduced; in some arboreal species it may be partially opposed. The ungual phalanges of creodonts are fissured; while this serves to distinguish them from living carnivores, it is not a unique character— they also occur in living condylarths, pangolins, and chalicotheres. Most felids and a few viverrids have evolved retractile claws; this serves to prevent the claws from becoming blunt in walking, so that they retain their maximum sharpness for tearing prey. THE BRAIN AND THE SENSES Until about a decade ago, Edinger’s (1929) paper contained almost all that was known of fossil carnivore brains. Piveteau (1961) illustrated thirteen genera, with anatomical notes commenting on the patterns of the sulci and gyri; his conclusion was that larger and more Recent brains tend to be more fissured and more complex. SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 253 More recently Jerison and Radinsky have in different ways made significant contri- butions. Jerison (1961, 1973) took a broad approach to the problems of brain evolu- tion in vertebrates and concentrated mainly on characteristics of brain weight to body weight ratios. In amniotes the brain weight (E) increases as the two-thirds power of the body weight ( P ), that is E = kP0'66 where k is a constant. Jerison also made use of an encephalization quotient (EQ), which is the ratio of the actual measured brain size to the expected brain size in an ‘average' mammal of the same body weight. The data for carnivores in Table 6 is extracted from Jerison (1973) with the addi- tion of some original data. The estimates of brain weight and body weight can rarely table 6. Brain and body size estimates in fossil carnivores (after Jerison 1973). Brain Body EQ weight (E) weight (P) ( E EQ Genus g kg V kpO 66 J X Creodonts Thinocyon 5-7 0-80 0-55 Cynohyaenodon 8-3 30 0 33 Pterodon 62 42 0-43 0-48 Hyaenodon 85 56 0-48 Megistotherium 375 354 0-62 Palaeogene Plesictis 1 1 1-3 0-77 carnivores Potamotherium 50 9-7 0-92 Daphoenus 1 49 26 0-46 2 66 30 0-54 Hesperocyon 15 2-0 0-79 Pachycynodon 39 9 0-75 0-61 Hoplophoneus 1 47 20 0-53 2 52 49 0-32 Eusmdus 38 21 0-42 Herpestes 13 2-1 0-66 Amphicyon 110 49 0-64 Neogene Plesiogulo 140 38 103 carnivores Mesocyon 1 52 10 0-93 2 37 9-5 0-69 Cynodesmus 36 13 0-54 0-76 Tomarctus 58 15 0-80 Pseudaelurus 89 43 0-60 be determined with accuracy, and a good deal of inspired guesswork has gone into the estimates. Nevertheless, the figures are reasonably consistent. Taking living carnivores as having an EQ of 1T0, then the mean (x) for Neogene carnivores is 0-76, that for Palaeogene carnivores 0-61, and that for creodonts is 0-48; that for the marsupial Didelphis on this basis is 0-22. That creodonts are about half as intel- ligent as living carnivores might be a deduction, but the samples are very small and there is considerable overlap in brain size of carnivore families, which is masked by the grouping. When carnivorous mammals are compared with their herbivorous contemporaries, they are seen to be ahead through the Tertiary (Table 7), though the ungulates appear to be steadily narrowing the gap. Radinsky (1975) used a B 254 PALAEONTOLOGY, VOLUME 20 modified EQ relating brain size to the area of the foramen magnum (Table 8). The foramen magnum area is related to body weight (correlation coefficient 065) and is a more readily available parameter. Using brain size-foramen magnum area relationships, Radinsky found that the modified EQ (EQA) was higher for carnivores than for mammals in general. While the families have considerable overlap in range, the mean for canids is higher than that for viverrids, and the latter is higher than that for felids. table 7. Evolution of relative brain size (after Jerison 1973). Means with sample size in brackets. Archaic Palaeogene Neogene Recent Ungulates 0-18 (13) 0-38 (26) 0-63 (13) 0-95 (25) Carnivores 0-48 (5) 0 61 (11) 0- 76 (6) 1- 10(15) Creodonta Carnivora table 8. Relation of brain size (E) to foramen magnum area (A) (after Radinsky 1975). (EQA = E/Ea where Ea - 1-35A148). Range Mean (x) Canids 1-32-1-70 1-51 Viverrids 0-93-1 58 1-23 Felids 0-98-1-51 115 164 species from 5 orders 1 -0 A totally different approach to brain studies is that of mapping the cortical areas and tracing the evolutionary changes in the fissure patterns (Radinsky 1968, 1969, 1971, 1973, 1975). Experimental work on living mammals (e.g. Welker and Campos 1963; Welker et a/. 1964) has shown that the sulci delimit functional areas of the cortex, and to an increasing extent these are becoming identifiable. Radinsky (1968) examined the brain morphology of living otter genera and noted that the coronal gyrus and lateral part of the posterior sigmoid gyrus were usually enlarged. It has been shown that the latter is associated with sensory receptors for the fore limb, while the coronal gyrus is associated with tactile sensitivity in the head region. For otters in water, the efficiency of sight is reduced and that of smell is totally lost; in compensation, otters have increased tactile sensitivity through well-developed vibrissae. The clawless otters feeding on crustaceans and molluscs use their sensitive digits to locate prey. The earliest known lutrine Potamotherium has an enormous coronal gyrus, and the bone around the upper lip is densely pitted, strongly suggesting well-developed vibrissae (Savage 1957). In the evolution of living families of carnivores, Radinsky (1971) has shown that one structure, the cruciate sulcus, has evolved independently at least five times (text-fig. 8); no specific function is yet known for this feature. The brains of living canids are all very similar, except for the prorean gyrus, and canids share with amphicyonids some fissure patterns which suggest a relatively close relationship. text-fig. 8. Endocasts of carnivore brains, a, Daphoenus. b, Megistotherium. c, Thylacinus. d, Hesperocyon. E, Mesocyon. F, Vulpes. G. Potamotherium. H, Promartes. I, Martes. J, Herpestides. K, Ichneumia. L, Eusmilus. m, Proailurus. N, Felis. (b, c, and g original; others after Radinsky 1971 ). c, coronal gyrus; cr, cruciate sulcus. 256 PALAEONTOLOGY, VOLUME 20 With improved techniques of taking endocranial casts, more information is becoming available. Unfortunately there are very few creodont brain casts known, but there appears to be a progressive increase in relative size through time. Megisto- therium has relatively the largest known brain for a creodont. The pattern is also highly complex and the brain has an unusually large cerebellum. From the brain we now turn to the senses— smell, sight, and hearing. The olfactory sense leaves in fossils two traces, the turbinal bones and the olfactory lobes of the brain. Though we have few complete endocranial casts, the trend in carnivores seems to indicate that the relative size of the olfactory lobes has increased with time. Radinsky (1975) has shown that the ratio of the olfactory lobes to total brain volume in living carnivores is 2-9% for felids, 4-7% for viverrids, and 5-0% for canids. Turbinal bones, if preserved, provide a second line of approach. There are two sets of turbinals in the nasal area. An anterior set, the maxilloturbinals, is innervated by the fifth cranial nerve and is concerned with warming and filtering the air. Behind these is a second set, the ethmoturbinals, which is innervated by the olfactory nerve; the air passes over these, and the olfactory sense is located here. The relative size of the turbinals is significant. If the maxilloturbinals are much larger than the ethmo- turbinals, as in otters, then there is reduced sense of olfaction. If the reverse is the case, as in many mustelines, olfaction is a highly important sense (text-fig. 9). text-fig. 9. Transverse section of nasal region of two carnivores; to same scale, c, cribiform plate; e-t, ethmoturbinal bones; m-t, maxilloturbinal bones. Unhappily, few fossil skulls are so preserved that the relative size of the two areas can be determined. In the creodonts Thinocyon and Cynohaenodon the ethmoturbinals appear to have been relatively large and the animals may, like many mustelids, have hunted mainly by scent. Very reduced olfaction can already be seen in the Early Miocene lutrine Potamotherium. Sight is much more difficult to deduce from the fossil evidence. Firstly, from the orbital area of the skull we can learn nothing of the structure of the eyeball, and hence nothing concerning resolution, distance, diurnal, nocturnal, or colour vision. Further, as much of the information is processed in the eye, the visual areas on the cerebral cortex do not enlarge or increase in complexity in proportion to visual SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 257 acuity. But the size of the orbit does provide some clue, and the orientation of the orbits allows us to assess the degree of stereoscopy. Quantitative evaluation of these characters is difficult, but generally in borhyaenids, creodonts, and miacids the eye is proportionately much smaller than in living carnivores. The degree to which the visual fields overlap provides some measure of stereoscopy (text-fig. 10 and Table 9). Borhyaena appears to have had the best stereovision among marsupial carnivores; the eyes of Thylacosmilus were very small and set facing laterally. Among creodonts, Patriofelis had the most forwardly directed eyes, though not as advanced as in living canids, hyaenids, and felids. text-fig. 10. Dorsal and anterior views of miacid Viverravus and felid Felis skulls to illustrate size and orientation of the orbits (in heavy line). table 9. Degree of stereoscopic vision in carnivores. Marsupialia Creodonta Carnivora Low 30°-50° Cladosictis Prothylacinus Thylacosmilus Hyaenodontidae Miacidae Medium 50°-70° Borhyaena Oxyaena Palaeonictis Amphicyonidae High 70°-80° Patriofelis Ursidae Viverridae Mustelidae Very high 80°-120° Canidae Hyaenidae Felidae The auditory sense is both the most important and the least satisfying sense to study. A great deal of information is often preserved, in both the middle- and inner- ear regions; these are frequently described in great detail (e.g. Hough 1948), yet extremely little of functional significance can be deduced. Hearing is certainly impor- tant in most carnivores, and in canids in particular— the bat-eared fox and desert foxes have remarkably large ears and their auditory perception is very acute. The basic taxonomic divisions within the carnivores are based as much on ear structure as on dentitions. For example, the tympanic bulla is never present on any creodont or miacid. An ossified bulla without a true septum demarcates the canoid and musteloid families, and a bulla with a septum distinguishes the viverroid and feloid families. Recently Hunt (1974) has made a detailed study of the auditory bulla in the context of carnivore evolution. 258 PALAEONTOLOGY, VOLUME 20 Tympanic bullae are absent in almost all Palaeocene and Eocene mammals. They are absent in early insectivores, rodents, perissodactyls, and artiodactyls. All condy- larths and creodonts lack ossified bullae, as do miacids (text-fig. 11). The primitive bulla in carnivores is fiat, as seen in the extinct amphicyonids, in ursids, and some mustelids. Inflated bullae develop in canids, procyonids, viverrids, felids, and in some mustelids and hyaenids. The function of the inflated bulla is not entirely clear; many desert mammals, especially rodents, have highly inflated bullae. Sound absorption in air becomes greater with a decrease of humidity and an increase of temperature. So in warm arid environments, selective pressure will favour forms with acute hearing. The increased acuity is achieved by enlarging the middle-ear CANIS LUTRA FELIS text-fig. 11. Transverse section of the ear region of three carnivores to illustrate the structure of the tympanic bulla. Ectotympanic black; entotympanic stippled; tympanic membrane (eardrum) dashed line. BO, basioccipital; CC, carotid canal; EAM, external auditory meatus; P, petrosum; Ps, pseudoseptum; S, septum; Sq, squamosal. cavity; this is the principal trend in the evolution of the carnivore ear region and is seen in members of most carnivore families except ursids. The increased space is acquired mainly by the expanded growth of the caudal entotympanic bone, and may be supplemented by invasion of the mastoid bone (some mustelids and hyaenids) or by expansion into the external auditory meatus (amphicyonids). Hunt has further shown how the carotid circulation provides a counter-current heat exchange mechan- ism to cool the cerebral arterial blood by the development of internal and external carotid retia. JAWS AND DENTITION Considering the mandibles of creodonts and carnivores, there are a number of consistent differences. Both lineages have evolved large and small taxa, short- and long-jawed types, exclusively carnivorous stocks and others only partially carni- vorous. In both lineages the mandibular dentition is narrower than the maxillary, and chewing occurs on one side at a time. The early creodonts (text-fig. 12) with their loose spherical condylar joint and very well fused and long symphysis, could have swung the mandible sideways into action, the transverse ridges of the molars guiding the occlusion of the carnassial teeth. In true carnivores, the tight cylindrical SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 259 OXYAENODON text-fig. 12. Posterior and internal views of left mandible of creodont Oxyaenodon and of canid Canis, to illustrate shape and position of condyle and symphysial area. condylar joint allows for a shorter and looser symphysis and a lateral slide to bring the carnassials into occlusion. In both lineages the temporal is the main muscle closing the jaw, whereas in ungulates it is the masseter (Smith and Savage 1959). The temporal is well aligned to provide fast action and derives strength from its massive size. Incisors and canines. The incisors are never specialized in carnivores and usually all three are present; the upper incisors are reduced in number in some short-faced forms and in those with sabre-like canines. Canines are always present and usually equally well developed in both upper and lower jaws; they are used for biting, piercing, and holding prey. In some lineages they become massive, that is thick in proportion to their length, as, for example, in amphicyonids, ursids, borophagine canids, and some creodonts such as Megisto- therium. It is, however, the evolution of sabre-like canines that provides the really startling innovations. These have evolved independently in the marsupials with Thylacosmilus, in the creodonts with Apataelurus , and several times among the felids (text-fig. 13). Associated with the evolution of sabre teeth are changes in the architecture of the skull and mandible, in the musculature for opening and closing 260 PALAEONTOLOGY, VOLUME 20 text-fig. 13. Skulls and mandibles of sabre-toothed carnivores, showing the arcs and centres of curvature of the canines. Transverse sections of the canines in black, a, Hoplophoneus. B, Apataelurus. c, Nimravus. d, Smilodon. E, Thylacosmilus. (a, c, and D after Matthew 1910,Bafter Denison 1938, and E after Riggs 1934). the jaws, and in the musculature for bringing the head down in a powerful sweep to provide the stabbing action. The face is always short ; the nuchal crests rise high above and sweep back beyond the occipital condyles. The condyles are often elevated high above the level of the dentition and there is a massive mastoidal area for the attach- ment of musculature to pull the head downward. The mandibles are slender with weak coronoid processes and are capable of opening wide to give a gape sufficient to clear the line of action of the sabre canines. A bony flange often develops on the ramus to protect the upper canine when the jaws are closed. In transverse section the sabres are usually ovoid, with the posterior edge more acute. While a circular section would be strongest, it would also offer great resistance to penetration; like- wise a very thin tooth which would penetrate easily would be liable to damage. Serrations often develop, especially on the more proximal edges of the blade. Riggs SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 261 (1934), describing Thylacosmilus, stated that the sabre canines were open rooted and continually growing; their roots extend much further back into the skull than in any other sabre-toothed carnivore. The immediate ancestors of the Pliocene Thylacosmilus are unknown, although in all features other than the canines it is a normal borhyaenid. The creodont Apataelurus is known only from the mandible and is placed with other less fully sabre-toothed taxa in the subfamily Machaeroidinae of the family Hyaenodontidae. Within the Felidae, lineages are not clear; the Oligo- cene Hoploplioneus does not appear to be ancestral to the Plio-Pleistocene machaero- dontines, and so sabre-like canines have evolved at least twice within the felids. The absence of any living sabre-tooth forms makes it impossible to be totally certain of their mode of functioning. Simpson (1941) has very fully analysed the mechanics of sabre-like canines, and only the briefest outline need be given here. He detailed the four theoretical ways in which the sabres could function as stabbing weapons to pierce thick skin and sink deep into the flesh, causing the victim to bleed to death; this he considered the principal function of the teeth. Secondarily they may have acted to slice open the carcass, but neither the outline of the teeth and their leading edges, nor the orientation of the musculature of the head and neck are optimally adapted for this, and slicing without prior stabbing would be virtually impossible. Premolars. In carnivores, premolars have tended with some notable exceptions to remain unspecialized; they are often reduced in number and the anterior ones are lost. The two striking specializations are the development of P4 as a carnassial tooth in the Carnivora (these are considered below with molars) and the develop- ment of very heavy premolars as crushing teeth, seen in hyaenids, borophagines, and some marsupials and creodonts. In stocks with massive premolars, the teeth become enlarged in girth, remain fairly low crowned and develop thick enamel; usually one or two upper and lower teeth, immediately anterior to the carnassials, are thus modified. It could be argued that, since great power is required to crush bone, these teeth are best situated as far back as possible. But the further back they are, the smaller the gape, and the animals also need to retain a carnassial dentition. The hyaenodont Megistotherium had hyaenoid premolars, presumably for bone crushing. In the creodont Quercytherium the crushing specialization extends over four teeth (P1-?4) (text-fig. 14). Among the borhyaenids, the genus Angelocabrerus QUERCYTHERIUM BOROPHAGUS CROCUTA text-fig. 14. Occlusal views of upper dentitions and lateral views of lower dentitions to illustrate specialized premolars in creodont Quercytherium, canid Borophagus and hyaenid Crocuta. Not to scale. 262 PALAEONTOLOGY, VOLUME 20 has a massive P3 and can be regarded as a likely hyaenoid type. The canid subfamily Borophaginae comprises a suite of large hyaenid-like dogs with heavy crushing premolars in the Neogene of North America. Molars. The main trends which can be distinguished in the evolution of carnivore molars are the simplification of the tooth pattern, the loss of crushing function in the more truly carnivorous types, the loss of the post-carnassial teeth and the increas- ing efficiency in carnassialization. The most sophisticated carnassials seen in felids are structurally very simple teeth, and it is in no small part due to this simplicity that the carnassial dentition at its most refined is so successful. This is seen most clearly when one looks at the cheek dentition of a primitive miacid and of a living felid (text-figs. 1 and 15). In insectivorous stocks, shearing and crushing are performed by two parts of the same tooth. On an upper molar, for example, the metacone provides the blade and the protocone the crushing mechanism, while on the lower molar the paraconid- protoconid becomes the blade and the talonid the basin. This arrangement persists in early creodonts, with the blades enlarged at the expense of the crushing function (text-fig. 16). In the Carnivora, however, since the carnassials are only one pair of text-fig. 15. Occlusal view of carnassial dentition in a miacid (Miacis) and a felid (Felis). Carnassial blades in heavy line. Miacis has a diagonal blade and transverse guiding ridge formed by post-carnassial structures; Felis has a longitudinal blade and lacks any guiding ridge. Not to scale. MIACIS FEUS HYAEN0D0N THINOCYON SINOPA PATRIOFELIS OXYAENA THYLACINUS BORHYAENA TEXT-FIG. 16. HYAENA VIVERRA ▼ MUSTELA AILURUS URSUS <53 CANIS OAPHOENUS Occlusal views of upper cheek dentitions of carnivorous mammals. Carnassial blades in heavy line; M1 indicated with black triangle. 264 PALAEONTOLOGY, VOLUME 20 teeth, the more posterior teeth can take on a crushing function for the smaller pieces of food already cut up by the carnassials. Amphicyonids, canids, mustelids, and, to a lesser extent, viverrids exemplify this trend. Ursids and procyonids have secondarily lost the carnassial function and acquired frugivorous, herbivorous, or omnivorous diets, while the insectivorous hyaenid Proteles has only a vestigial dentition. In those stocks that have become more purely carnivorous, such as felids and hyaenids, there is a tendency to lose the post-carnassial teeth, with the result that the shearing blades become more posteriorly placed and so a more powerful bite is achieved. The study of carnassial specialization involves an analysis of the number of teeth involved, the homology of the teeth, their replacement, the posterior migration, the increase in blade size, the angle of shear, the reduction of the non-shearing parts of the teeth, and the self-sharpening devices. The number of carnassials ranges from three pairs (M1“M3/M2-M4) in many marsupials, through two pairs (M1-M2/M3-M4) in many creodonts, to one pair (P4/Mx) in all Carnivora. Usually, when there are several pairs, the carnassial action is not evenly distributed and one pair tends to be dominant, often the more posterior pair. In Carnivora, carnassial shearing is developed in the milk dentition on dp3/dp4. Similarly, in creodonts in which a milk dentition is known, carnassial shearing is developed on the pair of teeth anterior to the permanent successors, usually dp4/Mx. In creodonts Mx is invariably heavily worn, indicating eruption much earlier than the other molars. THE FORM AND FUNCTION OF CARNASSIAL TEETH In all early carnivorous mammals, the shearing blades are orientated obliquely across the jaw; in the more truly carnivorous lineages the blade tends to become longitudinal, aligned parallel to the dental row (text-figs. 15 and 16). Functionally a longitudinal shear is easier to operate with food held against the side of the jaws, but the price paid is a loss of total shear length (the diagonal being equivalent to the hypotenuse of a right-angled triangle). However, by becoming longitudinal, there is no longer a need for the transverse guide lines provided by the protocone- paracone in the upper teeth and the protoconid-metaconid in the lower teeth. Hence the protocone and metaconid can become reduced or lost, so simplifying the teeth (text-fig. 1). As the protocone is reduced, so the talonid function diminishes and it too becomes vestigial or lost. All this allows the protoconid-paraconid blade of the lower molar to extend backwards and make up for the loss of length due to the realignment in a sagittal plane. This sequence of trends evolved independently in the oxyaenids, leading to Palaeonictis as the most felid-like form, and in the hyaenodontids with Hyaenodon. The Carnivora, having only one pair of carnassials, were able to acquire a very large set of shearing teeth and this probably functions better than two pairs, even allowing for replacement, though there is no obvious proof of this. The length of the carnassial blade is proportional to a the size of the individual, b the size of food chunk that can be swallowed, and c the degree to which the diet is carnivorous. Another interesting feature of carnassial blades is their shape; they are concave in a vertical plane and plano-convex in a horizontal plane (text-fig. 1). On closing, SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 265 the ends of the blades meet first, and only when almost fully closed do the centres meet. The double guillotine action produced by the concave edges allows the maxi- mum force to be exerted on a very small area at a time, and so increases the efficiency of the blade. The concave arc also ensures that the food is trapped between the blades and does not slide out. The terminal pillars of the blades are massive, and the blade thickness tapers towards the centre, where a slit separates the two halves. Opposite the slit the teeth are deeply excavated on the outer face of the P4 and on the inner face of M,; this creates space for the food during slicing and, by its semi- conical shape, helps to retain the chunk of food in position during cutting. A slit in this area is more efficient than a very thin band of enamel, which would be easily damaged. The blade of Mj is always convex outwards in a horizontal plane; the blade of P4 is either slightly convex inwards or flat in the same plane. Thus on closing the jaws (text-fig. Id), contact is made between the two blades at two centrally moving points throughout the traverse. The movement is primarily orthal with a slight medial movement; to allow this, the cylindrical glenoid permits the mandible to slide inwards as it closes, and differential stresses are absorbed in the non-fused symphysis. Several authors have attempted a quantitative assessment of carnassiality, most recognizing that the trend involves a migration of the shear plane from the diagonal to the longitudinal position. Denison (1938) gave figures for the "shear angle’ (unidentified but presumed to refer to the angle formed by the plane of the paraconid- protoconid shear with the jaw axis). Only three genera were used : Oxyaena spp. 60°-30°, Protopsalis very low, and Patriofelis 0°, i.e. parallel to the jaw. Butler (1946) measured the anterior and posterior (i.e. lateral) borders of the upper carnassial teeth and gave the maximum value of the ratio of posterior edge/anterior edge; this ratio he divided by the number of carnassial teeth to produce his index of carnassial differentiation. His table comprised nine genera of creodonts, ranging from 0-50 in the primitive genus Sinopa to 0-90 in the advanced Patriofelis. Crusafont and Truyols (1956) have given an excellent and original account of carnassiality in carnivores, restricting themselves, however, to members of living families of carnivores and omitting creodonts and miacids. The analysis is based on the measurement of the angle of the blade of the upper and lower carnassial teeth for six families throughout the Cainozoic. For example, the angle formed by the protocone-metastyle-paracone is analysed with the following results: Ursidae 26°-52°, Canidae 15°-36°, Mustelidae 19°-44°, Viverridae 18°-25°, Hyaenidae 14°-22°, and Felidae 10°-25°. The smaller the angle, the more reduced the proto- cone and the more longitudinal the shear on the metastyle. Ursids are seen to be hypocarnivores, felids and hyaenids hypercarnivores, with viverrids, canids, and mustelids occupying a central position. Similar results are obtained for the lower carnassials. The range of angles is seen to increase with time; in the lower carnassial it is 47° in Eocene Oligocene and 98° in the Late Miocene-Pliocene (text-fig. 17). The obvious extension would be to apply the Crusafont and Truyols techniques to marsupials and creodonts. However, frequent lack of good reference points made this difficult, and it was found impossible to achieve consistent and reliable results. Of prime importance for efficient functioning of the carnassial dentition is the 266 PALAEONTOLOGY, VOLUME 20 ment of angle a. b, family range of angle a. c, range of angle a through time (after Crusafont and Truyols 1956). maintenance of a sharp cutting edge, and there are various ways in which this could be achieved. In rodents the incisors maintain a cutting edge by having enamel restricted to one edge so that the softer dentine wears down more rapidly and leaves a sharp edge; this rapid wear is compensated for by continual growth from open roots. In sabre-like canines a serrated edge of enamel is developed, since there is nothing for the tooth to be sharpened against. Neither of these solutions would be efficient in a carnassial dentition. The alignment of the blades is very precise, and the rodent incisor method is unlikely to be adequate. The apposition of two teeth rules out the serrated-edge method, which works only in isolation. This leaves only a self-sharpening device, a process named ‘thegosis’ by Every (Every and Kiihne 1971). The self-sharpening works well on all save a few specialized hyaenodonts which, according to Dr. J. S. Mellett (pers. comm.), continue to enlarge their skulls throughout life and so alter the spatial relationships of the teeth, which turn inward and wear down to the roots in an attempt to counter the effects of size increase. BIOGEOGRAPHY The marsupials evolved two successful carnivore lineages. The borhyaenids were the only mammalian carnivores in South America from Palaeocene to Pliocene times; twenty-five genera are known and they range from small otter-like forms to true highly specialized sabre-toothed taxa. Most are poorly known; some reached the size of bears, but none appears to have been truly cursorial. They were only ousted in the Plio-Pleistocene by the invasion of placental carnivores from North America; of these the canids, mustelids, and procyonids have been particularly successful. The dasyurid marsupials in Australia are less well known in the fossil record, with only two truly carnivorous genera in recent times, of which the thylacine closely paralleled Canis. The creodonts successfully occupied the Palaeogene carnivore niches; they did not survive beyond the Oligocene in North America, but continued on into the a » Oh 268 PALAEONTOLOGY, VOLUME 20 early Miocene in Europe and into the late Miocene in Africa and Asia. The oxyaenids were particularly successful in the North American Eocene with seven genera, all relatively short legged, but with carnassial specializations near to those of felids (e.g. Patriofelis). The hyaenodonts were more diverse, with long- and short-faced taxa, with a sabre-toothed group, canid-, felid- and hyaenid-like forms. Some of the latter reached the size of very large bears. The reasons for the replacement of the creodonts by true carnivores are far from clear. Both groups originated in the Palaeocene, and the creodonts rapidly achieved a dominance in the carnivore niches. The specialized creodonts display many of the characters usually associated with the Carnivora— some had digitigrade feet, some had good stereoscopic vision, some had highly efficient carnassial dentitions, and yet others had large and complex brains. The history of replacement is paralleled in those of the replacement of the multituberculates by the rodents, the perissodactyls by the artiodactyls, and the archaic ungulates by the modernized ungulates. The most recent example is that of the replacement in Australia of the thylacine by the dingo. No one single factor can be cited to explain the change; in an intensely com- petitive field the combination of slight advantages are favoured by natural selection; only a thorough field study of the ecology and behaviour of each stock would enable these to be quantified. In the early stages of evolution, the advantages appear to have lain with the creodonts, and only after 20 million years did the balance tip in favour of the Carnivora. None the less, the Carnivora have been about four times as successful as the creodonts; creodonts are known from 45 genera spanning about 45 million years, while Carnivora, excluding living and aquatic genera, are known from 218 genera over a span of about 55 million years. The fossil record of early Palaeogene Carnivora is poor, but by late Eocene times nearly all the terrestrial carnivore families had become established in Holarctic realms. The amphicyonids occupied canid, ursid, and hyaenid niches, to be replaced later by stocks of these modern families. The canids are the most ubiquitous of carnivores, found on all continents save Antarctica. They are usually medium sized and never achieve the degree of specialization found in felids and hyaenids, which may in part account for their success. The mustelids, occupying Holarctic realms, and the viverrids, occupying tropical realms, dominate the smaller-sized carnivore stocks, exploiting the vast potential of rodent prey, though usually supplementing it with other foods. The hyaenids are the youngest family and occupy ossifagous niches in the Old World. The felids can truly be regarded as the acme of carnivore evolution; though limited today to a few genera, their many species are very widely distributed. In almost all their skeletal elements, in their senses qnd dental apparatus, they represent the ultimate in carnivore achievement. CONCLUSIONS Carnivores comprise about 10% of all mammalian genera and about 2% of the mammalian biomass. Carnivore specialization is defined as having a carnassial dentition and a sub- stantial proportion of vertebrates in the diet. Carnivore stocks evolved twice among the marsupials (borhyaenids and dasyurids) SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 269 and twice among the placental mammals (Creodonta and Carnivora). Borhyaenids were the only mammalian South American carnivores until the Pliocene, dasyurids filled a similar role in Australia until Recent times. Creodonts were gradually replaced by Carnivora in North America during the Oligocene, and in Africa and Asia during the late Miocene. The vertebral column of carnivores is strong and flexible, and the tail is usually long. In creodonts the vertebral column was less flexible than in living carnivores. The limbs of carnivores are multi-purpose organs and do not become highly specialized, though most acquire some cursorial adaptations. They are relatively shorter in the Creodonta than in the Carnivora. Aquatic carnivores always have short limbs, and cursorial forms have long limbs. The foot is usually the shortest element of the three segments of the limb, whilst in herbivores the foot is the longest. Primitive carnivores have plantigrade feet, but advanced creodonts and all cursorial carnivores have fully digitigrade feet. The brains of creodonts are smaller and simpler than those of living carnivores, but the brains of carnivorous mammals are always relatively larger than those of the contemporary herbivores. The olfactory sense in creodonts was probably at least as good as in living carnivores. Sight, however, was probably not as good, and fully stereoscopic vision developed only rarely in carnivorous marsupials and creodonts, whereas it is common in canids, hyaenids, and felids. Hearing is acute in canids, felids, and viverrids and is particularly acute in carnivores living in warm arid environments. The mandibles in most creodonts were loose fitting with long firm symphyses, and the carnassials required transverse guiding ridges to ensure precise occlusion. The mandibles of the Carnivora are tightly fitting with no need for long symphyses and guiding ridges; their carnassial blades have rotated from an oblique angle to a longitudinal position in the most specialized forms (canids, hyaenids, and felids). The most advanced creodonts were more specialized than the primitive Carnivora. The canine teeth are usually well developed and may become specialized as stabbing sabre-like teeth: this trend evolved independently in three stocks— marsupials, creodonts, and felids. Premolar teeth in some marsupials, creodonts, and Carnivora have become mas- sively large to function as bone-crushing tools, and in a few cases to act as mollusc- shell crushers. The principal trend in the evolution of molar teeth is one of simplification, by the reduction or loss of the crushing function and loss of the post-carnassial dentition. The evolution of the carnassial dentition involves three, two, or one pair of teeth. The most advanced forms have only one pair, which is large and devoted solely to shearing by loss of all other parts of the teeth not concerned with cutting. The cutting edges of the blade are self-sharpening. c 270 PALAEONTOLOGY, VOLUME 20 REFERENCES bakker, r. t. 1972. Anatomical and ecological evidence of endothermy in dinosaurs. Nature, Loud. 238, 81-85, 2 figs., 2 tables. blainville, H. M. D. de. 1864. Osteographie ou description iconographique du squelette et du systeme dentaire des mammiferes recent et fossiles. Paris. Atlas Vol. IV. butler, p. m. 1946. The evolution of carnassial dentitions in the Mammalia. Proc. zool. Soc. Lond. 116, 198-220, 13 figs. crompton, a. w. and hiiemae, K. 1970. Molar occlusion and mandibular movements during occlusion in the American opossum, Didelphis marsupialis L. Zool. J. Linn. Soc. 49, 21-47, 10 figs., 1 pi. and sita-lumsden, a. 1970. Functional significance of the therian molar pattern. Nature, Lond. 227 , 197-199, 5 figs. crowson, r. a. et al. 1967. Arthropoda. In harland, w. b. et a!, (eds.). The Fossil Record. London (Geo- logical Society), pp. 499-534. crusafont-pairo, m. and truyols-santonja, J. 1956. A biometric study of the evolution of fissiped carnivores. Evolution, 10, 314-332, 14 figs., 9 tables. denison, R. h. 1938. The broad-skulled Pseudocreodi. Ann. N.Y. Acad. Sci. 37, 163-256, 32 figs. edinger, T. 1929. Die fossilen Gehirne. Ergebn. Anat. 28, 1-249, 203 figs. every, R. G. and kuhne, w. G. 1971. Bimodal wear of mammalian teeth. In kermack, d. m. and K. A. (eds.). Early Mammals. London: Academic Press, pp. 23-27, 1 fig., 1 pi. gill, E. d. 1954. Ecology and distribution of the extinct giant marsupial 'Thylacoleo' . Victoria Nat. 71, 18-35, 12 pis. Gregory, w. k. and hellman, m. 1939. On the evolution and major classification of the civets (Viverridae) and allied fossil and recent Carnivora, a phylogenetic study of the skull and dentition. Proc. Am. phil. Soc. 81, 309-392, 36 figs., 6 pis. Hildebrand, m. 1954. Comparative morphology of the body skeleton in Recent Canidae. Univ. Calif. Pubis. Zool. 52, 399-470, 30 figs., 4 pis. 1961. Body proportions of didelphid (and some other) marsupials, with emphasis on variability. Am. J. Anat. 109, 239-249, 9 figs. hough, J. r. 1948. The auditory region in some members of the Procyonidae, Canidae and Ursidae. Bull. Am. Mus. nat. Hist. 92, 67-1 18, 1 1 figs., 7 pis. hughes, N. F. 1976. Palaeobiology of angiosperm origins. Cambridge: Univ. Press, vii + 242 pp., 73 figs. hunt, R. M. 1974. The auditory bulla in Carnivora. J. Morph. 143, 21-76, 3 figs., 13 pis. jerison, H. J. 1961. Quantitative analysis of evolution of the brain in mammals. Science, 133, 1012-1014, 1 fig. 1973. Evolution of the brain and intelligence. New York: Academic Press, xiv+482 pp., illustrated. Matthew, w. d. 1906. The osteology of Sinopa, a creodont mammal of the Middle Eocene. Proc. U.S. natn. Mus. 30, 203-233, 20 figs., 1 pi. — — 1910. The phylogeny of the Felidae. Bull. Am. Mus. nat. Hist. 28, 289-316, 15 figs. merriam, J. c. 1912. The fauna of Rancho la Brea Pt. II. Canidae. Mem. Univ. Calif. 1, 215-272, 43 figs., 5 pis. — and stock, c. 1932. The Felidae of Rancho la Brea. Pubis. Carnegie Instn, No. 422, xvi+231 pp., 152 figs., 42 pis. mitchell, e. and tedford, r. h. 1973. The Enaliarctinae a new group of extinct aquatic Carnivora and a consideration of the origin of the Otariidae. Bull. Am. Mus. nat. Hist. 151, 201-284, 27 figs., 4 tables. Muybridge, E. 1899. Animals in motion. London: Chapman and Hall, x + 264 pp.; illustrated. osborne, h. f. 1900. Oxyaena and Patriofelis restudied as terrestrial creodonts. Bull. Am. Mus. nat. Hist. 13, 269-279, 8 figs., 2 pis. piveteau, J. 1961. Encephales de carnivores fossiles. Trade de Paleontologie, 6 (1), 806-820, 19 figs. radinsky, l. b. 1968. Evolution of somatic sensory specialisation in otter brains. J. Comp. Neur. 134, 495- 506, 6 figs. — 1969. Outlines of canid and felid brain evolution. Ann. N.Y. Acad. Sci. 167, 277-288, 6 figs. 1971. An example of parallelism in carnivore brain evolution. Evolution, 25, 518-522, 4 figs. 1973. Evolution of the canid brain. Brain Behav. Evol. 7, 169-202, 13 figs. 1975. Evolution of the felid brain. Ibid. 1 1, 214-254, 12 figs., 2 tables. SAVAGE: EVOLUTION IN CARNIVOROUS MAMMALS 271 riggs, E. s. 1934. A new marsupial saber-tooth from the Pliocene of Argentina and its relationships to other South American predacious marsupials. Trans. Am. phil. Soc. 24, 1-32, 8 pis., 4 figs. savage, r. j. G. 1957. The anatomy of Potamotherium an Ohgocene lutrine. Proc. zool. Soc. Lond. 129, 1 5 1 244, 3 pis., 38 figs., 8 tables. - 1973. Megistotherium , gigantic hyaenodont from Miocene of Gebel Zelten, Libya. Bull. Brit. Mus. nat. Hist. Geol. 22, 483-511, 20 figs., 4 pis. scott, w. b. and jepsen, g. l. 1936. The mammalian fauna of the White River Oligocene. Pt. I. Insectivora and Carnivora. Trans. Am. phil. Soc. 28, 1 153, 7 figs., 22 pis. simpson, G. G. 1941. The function of saber-like canines in carnivorous mammals. Am. Mus. Novit. No. 1130, 1 -12, 4 figs. — 1945. The principles of classification and a classification of mammals. Bull. Am. Mus. nat. Hist. 85, 1-350. Sinclair, w. J. 1906. Mammalia of the Santa Cruz beds: Marsupialia. Rep. Princeton Univ. Exped. Patagonia , 4, 333-482, 9 figs., 16 pis. smith, J. M. and savage, R. J. G. 1956. Some locomotory adaptations in mammals. Zool. J. Linn. Soc. 42, 603-622, 14 figs., 3 tables. — 1959. The mechanics of mammalian jaws. School Sci. Rev. 141, 289-301, 8 figs. stromer, e. 1902. Die Wirbel der Land-Raubtiere, ihre Morphologie und systematische Bedeutung. Zoo/ogica. 36, 276 pp., 5 pis., 23 tables. voorhies, m. r. 1969. Taphonomy and population dynamics of an Early Pliocene vertebrate fauna, Knox County, Nebraska. Contr. Geol. Univ. Wyoming Spec. Paper , 1, 69 pp., 29 figs., 1 1 tables. welker, w. i. and campos, g. b. 1963. Physiological significance of sulci in somatic sensory cerebral cortex in mammals of the family Procyonidae. J. Comp. Neur. 120, 19-36, 10 figs., 6 tables. welker, w. I., Johnson, j. i. and pubols, B. h. 1964. Some morphological and physiological characteristics of the somatic sensory system in raccoons. Am. Zool. 4, 75-94, 23 figs., 1 table. R. J. G. SAVAGE Department of Geology The University Bristol BS8 1TR Typescript received 5 July 1976 EPIDERMAL STUDIES IN THE INTERPRETATION OF LEPIDOPHLOIOS SPECIES by B. A. THOMAS Abstract. Epidermal details are described from the leaf cushions of five species of Lepidophloios including one new species L. grangeri. Such epidermal information allows the species to be distinguished more accurately. Cushion variation within a species is also considered, with epidermal studies verifying the conspecificity of specimens. It is suggested that young Lepidophloios cushions are of the Lepidodendron form, which subsequently enlarged to increase their photosynthetic ability. Lepidophloios is a genus of arborescent lycopod stems very similar to Lepidoden- dron in size, general morphology, and in having a dense covering of spirally arranged leaves. All but the terminal shoots of both genera shed the apical parts of their leaves, retaining only their basal portions as the familiar leaf cushions. The main distinguish- ing feature is that in Lepidophloios most cushions characteristically bulge outwards and downwards in such a manner that they overlap each other and show rounded lateral and basal angles (text-fig. 1). The leaf scars appear at the bases of the visible portions of the leaf cushions and are always broader than long, usually with one vascular print and two parichnos scars. Therefore the only part of the cushion normally visible is the upper half representing the basal portion of the adaxial leaf surface. The lower half, or abaxial surface, is hidden. Confusion has, however, existed about the separation of these two genera and opinions have varied about the determinations of some specimens. This problem text-fig. 1. Drawings to illustrate the main feature of the Lepidophloios leaf cushion. a, leaf cushions, 1— ligule pit aperture. The stippled area of the heavy print cushion repre- sents the area of connection of the cushion to the stem. B and c, diagrammatic sections through uncompressed leaf cushions. The lines of section are shown on fig. a. [Palaeontology, Vol. 20, Part 2, 1977, pp. 273-293, pis. 33-36.] 274 PALAEONTOLOGY, VOLUME 20 is accentuated when specimens are mistakenly inverted. For example, specimens of L. acerosus Lindley and Hutton have been mistaken for species of Lepidodendron and figured inverted, while specimens of Lepidodendron have been figured upside- down and described as species of Lepidophloios. Such determinations have naturally proved troublesome in the literature and the more important are returned to later. Confusion of this kind is best avoided by careful study and recognition of the finer cushion details. General shape with suggestions of cushion overlap is insufficient for generic determination as a few species of Lepidodendron show this feature in their upper angles (e.g. L. mannabachense Presl., Thomas 1970, pi. 32, figs. 3, 4, 6, 7). One such specimen from the Black Bed Coal of Yorkshire clearly illustrates such a possible confusion. Both Kidston (1893, p. 555, pi. 1, fig. 4) and Crookall (1964, p. 307, pi. 75, fig. 6) described it as Lepidophloios laricinus but 1 believe it to be Lepidodendron mannabachense. I have therefore figured some of its leaf cushions correctly and inverted to show how one can be misled by incorrect orientation (PI. 33, figs. 1, 2). Such cushions presumably bulged outwards more above than below the leaf scars which could then have given this marginal overlap during compression. It is far safer to look for the ligule pit apertures, the parichnos, and the foliar prints, as their positioning is only slightly variable. The ligule pit is always above the leaf scar; the parichnos, when present, are invariably below the leaf scar; while the three foliar prints normally form an arc in the lower half of the leaf scar with the vascular print being slightly below the level of the other two. Even surface ornamentation can sometimes be a valuable guide as certain Lepidodendron species have their cushions distinctly striated above the leaf scar which is a feature not shown by Lepidophloios. The cushion cuticles are described here in some detail for the first time, although certain structural details of their guard cells have already been discussed (Thomas 1974). All show the same general arrangements of epidermal cells and stomata as those described from Lepidodendron (Thomas 1970). Cuticle was prepared by macerating portions of compression in Schulze solution (concentrated nitric acid and about 5% potassium chlorate) normally followed by clearing in dilute ammonia solution. Some cuticles were mounted unstained in glycerine jelly and examined by normal transmitted light, supplemented in some instances by phase-contrast illumination. These prepared slides have been deposited with their respective speci- mens. Other cuticles were coated with gold/palladium and studied with a ‘Stereo- scan’ Scanning Electron Microscope. A range of specimens was examined for most species in an attempt to detect any variation, but this was often prevented by a lack of suitable material. Many specimens are preserved as impressions from which all the compression has either disappeared during fossilization or has been removed since collection. Even when some compression remains it is often badly cracked so that it does not yield cuticle in usable-sized fragments. The most important specimens examined are listed for each species discussed. They belong to, or were deposited in, the British Museum of Natural History, London (BMNH); the Czechoslovakian National Museum, Prague (CNM); the Institute of Geological Sciences: Kidston Collection, London (K) and the general collections of London (IGSLond), Leeds (IGSLds), and Edinburgh (IGSE). THOMAS: LE P I DO P H LO IO S EPIDERMIS 275 SYSTEMATIC DESCRIPTIONS Lepidophloios laricinus Sternberg Plate 33, figs. 1-6; Plate 34, figs. 1-3; text-figs. 2, 3 1820 Lepidodendron laricinum Sternberg, p. 21, pi. 11, figs. 2-4. 1875 Lepidophloios laricinum Sternberg; Feistmantel, p. 191, pi. 33, figs. 1-4; pi. 34, figs. I 4. 1904 Lepidophloios laricinus Sternberg; Zalessky, pp. 30, 99, pi. 6, figs. 8, 10; pi. 7, figs. 1, 2; pi. 8, figs. 7, 9. 1964 Lepidophloios laricinus Sternberg; Crookall, pars, p. 307, pi. 74, fig. 5; text-figs. 98, 100c. Material. Type specimen, CNM CGH316, from the Wranowitzer Shaft in the Radnice Coalfield, Central Bohemia (uppermost Westphalian B-C); BMNH V58739 from Kamenny Ujezd, near Nyrany, Plzen district. West Bohemia (Upper Westphalian B-D?); IGSLds V4486b from above the Main Seam, New- biggin Colliery, Northumberland— Westphalian B; IGSLond 15027 from Dysart, near Kirkaldy, Fife— Westphalian A or B. All the specimens accepted here are of the same general cushion morphology in having downward bulging and overlapping leaf cushions. They all appear to be from mature stems, mostly of large diameters, and none are known which might be thought to come from terminal shoots still retaining the distal lamina portions of their leaves. Visible portions of the overlapping cushions are normally much broader than long although in a few specimens these dimensions are nearly equal. Distinct keels are never present on the cushions but some- times their surfaces bulge in the median line giving the appearance of a very indefinite keel. The three foliar prints are usually in the lower half of the leaf scar but occasion- ally they can be found in the upper half. The type specimen was figured by Sternberg without ligule pit apertures on the leaf cushions, but re-examination shows indis- tinct pit scars about 1-5 mm above the leaf scars (text-fig. 2). This species has often been quoted as having its ligule pit scars a little distance above the leaf scars and in general comparison with other species this appears to be correct. Crookall (1964) cites K 6142, from Glenburn and Hilton Colliery, Skelmersdale, Lancashire, as an exception to this. However, as the visible cushion areas are only 2-0 mm long and 3-5 mm broad, one can hardly expect there to be a very large distance between these two scars. Cushion size will naturally determine the length or area of certain individual features and no definite sizes or areas should ever be quoted. The type specimen is an impression and unfortunately the very little remaining compression gave no cuticle. Preparations were, however, made from BMNH V58739, which came from the same area as the type, and from IGSLds V4486b and IGSLond 5027. text-fig. 2. Lepidophloios laricinus Sternberg. Type specimen CNM CGH316, ■ 2. Cuticle description. The epidermal cellular arrangement appears to be virtually identical on both the exposed and hidden leaf cushion surfaces. Most of the cells pit aperture, c-f, cuticle from BMNH V58739. c, cuticle from median area of the cushion, x 400; drawn from the undersurface; slide no. V58739a. Arrow directed parallel to the vertical axis of the cushion. d, cuticle from the non-median area of the cushion, x 400; drawn from the undersurface; slide no. V58739b. E, ligule pit cuticle, x400; slide no. V58739c. f, diagrammatic median transverse section of a stoma, x 500. EXPLANATION OF PLATE 33 Figs. 1 and 2. Lepidodendron mannabachense Presl. K 1404 from above the Black Bed Coal, Low Moor, Yorkshire; both illuminated from the top left, x 3. 1, orientated correctly. 2, inverted showing ‘Lepi- dophloios' characters. Figs. 3-6. Lepidophloios laricinus Sternberg. IGSLds V4466B from above the Main Seam, Newbiggin Colliery, Northumberland, x4. 4-6, cuticle prepared from BMNH V58739; 4, SEM., x200; 5, 6, transmitted light photographs; slide no. V58739a, x 350. PLATE 33 THOMAS, Lepidophloios 278 PALAEONTOLOGY, VOLUME 20 are elongated along the length of the cushion and about 40 /xmx 15 /xm in size, although those near the cushion sides are isodiametric and about 15 /xm large. Their anticlinal walls are straight, smooth, and 2 /xm thick, while the periclinal walls are flat and smooth. Stomata are randomly arranged on the cushion at about 250 per mm2 and orientated roughly parallel to the vertical axis of the cushion. Individual stomata have an average size of 45 /xm x 30 /xm, having an observed range of about 40-60 /xm x 23-40 /xm, with guard cells sunken in pits about 6-10 /xm deep. The ligule pit cuticles are about 0-7 mm broad, but no length measurements were possible due to their fragmentation during preparation. The pit wall cells are rectangular, 50 /xm x 16 /xm large, longitudinally elongated, and in vertical rows. The anticlinal walls are straight, smooth, and about 2 /xm thick; the periclinal walls are flat and smooth. Comparison. L. laricinus has been described from both the Upper and the Lower Carboniferous rocks of Great Britain, but it is here restricted to the Upper Carboni- ferous with a reassignment of the Lower Carboniferous specimens. K 1828, from the Carboniferous Limestone Series of Grange, Linlithgowshire in Scotland, was originally described by Kidston (18936, p. 560) as L. macrolepidotus and by Crookall (1964) as L. laricinus. Here I am redescribing it as L. granger i sp. nov. (p. 286). IGSE 1296, described by Crookall (1939, p. 14, pi. 2, fig. 2) from the Scottish ‘Mill- stone Grit’ at Cambus, near Alloa, Clackmannanshire, has already been referred to Lepidodendron rhodianum Sternberg (Thomas 1970, p. 168, text-fig. 12). The specimens of Lepidophloios laricinus described by Crookall (1964, p. 310) from the Carboniferous Limestone Series of Scotland differ in two aspects of cushion morphology. They have more prominently raised median areas to their cushions and have ligule pit apertures adjacent to the upper edges of the leaf scars. No cuticles could be prepared, but the differences in cushion morphology seem suffi- cient to strongly question their inclusion within this species and it seems more prudent to keep them separate. They are equally unlike all the other described species, but it is not intended at present to give them a new name. This view is then in agreement with that of Kidston who had originally labelled these specimens as Lepidophloios sp. Those specimens (K 4991-4992) quoted by Crookall (1964, p. 312, pi. 74, figs. 3, 4) as examples of small shoots showing upturned and undeflected leaf cushions are also excluded here from this species. They are instead reassigned to Kidston’s original determination of L. acerosus on the basis of both cushion morphology and epidermal characters seen in cuticle preparations. EXPLANATION OF PLATE 34 Figs. 1-3. Lepidophloios laricinus Sternberg. Cuticles prepared from BMNH V58739. 1, cushion cuticle; SEM., x500. 2, stoma; SEM., xl500. 3, ligule pit cuticle; slide no. V58739c, x 350. Figs. 4-7. Lepidophloios acerosus Lindley and Flutton. 4, K 4947 from above the Parkgate Coal (Old Hards), Hartley Bank Colliery, Horbury, Yorkshire, x 1. 5, K 764 from above the Kiltongue Coal, Foxley, near Glasgow, x2. 6, SEM. showing cushion cuticle from above and below the leaf scar of K 764, x200. 7, stoma photographed with transmitted light; slide no. PF 3195, x 1200. PLATE 34 THOMAS, Lepidopliloios 280 PALAEONTOLOGY, VOLUME 20 Lepidophloios acerosus Lindley and Hutton Plate 34, figs. 4-7; Plate 35, figs. 1, 2; text-fig. 4 1831 Lepidodendron acerosum Lindley and Hutton, pi. 7, fig. 1 ; pi. 8. 1837 Lepidostrobus pinaster Lindley and Hutton, pi. 198. 18936 Lepidophloios acerosus Lindley and Hutton; Kidston, p. 558, pi. 1, fig. 1 ; pi. 2, fig. 9. 1901 Lepidophloios acerosus Lindley and Hutton; Kidston, p. 54, text-fig. 7c (inverted). 1914 Lepidophloios acerosus Lindley and Hutton ; Arber, pp. 396, 415, 444, pi. 28, fig. 20 (inverted). 1917 Lepidophloios acerosus Lindley and Hutton; Kidston, pp. 1057, 1080, 1083, pi. 2, fig. 5 (inverted). 1929 Lepidophloios acerosus Lindley and Hutton; Crookall, p. 26, pi. 3, fig. 50; pi. 22, fig. k (both inverted). 1974 Lepidophloios acerosus Lindley and Hutton; Crookall, p. 313, pi. 75, fig. 7; pi. 76, fig. 2; pi. 79, fig. 5 (inverted); text-fig. 1 00c/. Material. K 764 from above the Kiltongue Coal, Foxley, near Glasgow, Lanarkshire— communis Zone, Westphalian A; K 3459 from above the Fenton Coal, Dodworth Colliery, near Barnsley, Yorkshire— communis Zone, Westphalian A; IGSLds RC 1778 from above the Ince Yard Mine, Mains Colliery, near Wigan, Lancashire— modiolaris Zone, Westphalian A; K 4991 and BMNH V33591 from above the Barnsley Red, Monkton Main Colliery, near Barnsley, Yorkshire— similis pulchra Zone, Westphalian B; IGSLond KP 449 from above the No. 5 seam Chislet Colliery, Kent— Transition Coal Measures, West- phalian ?C; K 6284 from above the Hafod Rider Seam, Hill’s Plymouth Colliery, Pentrebach, near Merthyr, Glamorgan- upper similis pulchra Zone, Westphalian B; K 6142 from Glenburn and Hilton Colliery, 1 mile NE. of Skelmersdale, Lancashire— Westphalian?; K 3634, from above the Fulledge or Yard Mine, New Shaft, Bank Hall Colliery, Burnley, Lancashire— Westphalian B; K 4992 from above the Pargate Coal, Church Lane Colliery, Dodworth, near Barnsley, Yorkshire— modiolaris Zone, West- phalian A. The widely recognized and described form of L. acerosus is a typical ‘ Lepidophloios ’ stem with downturned and overlapping leaf cushions. The exposed cushion areas are longer than broad but their relative dimensions vary slightly with size and the amount of interdependent overlap. Definite cushion keels are usually present but in some specimens they are only slightly raised. The three foliar prints are usually in the lower halves of the leaf scars although they are sometimes nearly central. Ligule pit apertures occur just above the leaf scars, although as Crookall (1964, p. 312) has shown they can be sometimes slightly separated. The smallest shoots are rather different in having undeflected smaller leaf cushions which could be referred to the genus Lepidodendron. Specimens K 3634, K 4991- 4992, K 6142, and K 6284 possess cushions of this kind, while BMNH V33591 also possesses more ‘normal’ slightly elongated and downward defected, leaf cushions as its lower end. Kidston (18936, pi. 1, fig. 1) described a small leafy shoot with up- turned leaf cushions as Lepidophloios acerosus, but I would agree with Jongmans EXPLANATION OF PLATE 35 Figs. 1, 2. Lepidophloios acerosus Lindley and Hutton. Cuticle from the cushion surface above the leaf scar of K 764. 1, SEM. photograph showing stomata and epidermal cells, X 500. 2, cushion cuticle with ligule pit cuticle; slide no. PF 3196, x 125. Figs. 3-6. Lepidophloios macrolepidotus Goldenberg, from the roof of the Fenton Coal, Dodworth Col- liery, near Barnsley, Yorkshire. 3, K 3256, x3. 4, K 3257, x 3. 5, 6, cuticle from K 3256; slide no. PF 2895. 5, x 250; 6, x 550. PLATE 35 THOMAS, Lepidophloios 282 PALAEONTOLOGY, VOLUME 20 (1930, p. 40) and Crookall (1964, p. 315)' that, on the basis of the visible cushion morphology, there appears to be no good reason for including it within this species. A comparison of these stems reveals several interesting facts which are relevant to a consideration of their growth patterns. The undeflected cushions of K 4991 and K 4992 are about 5 0 mm long and 5-5 mm broad, therefore being larger than the smallest of the ‘more usual’ deflected cushions of IGSLond 3459. The cushions on K 764 and K 4947 show a progressive increase in size down the stem, while the former also exhibits a corresponding increase in shoot diameter. There are there- fore shoots with either non-deflected or deflected leaf cushions suggesting that this may be an expression of subsequent growth of the cushions; an idea supported by those specimens showing a gradual basipetal increase in cushion size. This is clearly a complex subject which is not yet fully understood but we shall return to it in the general discussion. Cuticle description. Preparations have been made from a selection of specimens with undeflected and deflected cushions and from cushions of varying sizes. Certain general characteristics can be given for the species although certain slight variations have been noticed. The epidermis is different on the two cushion surfaces (PI. 34, fig. 6). The epidermal cells from the exposed adaxial surface are isodiametric and about 15-20 ^.m large over most of the central area, but near the sides they are roughly 30 pmx 15 /xm and elongated towards the cushion margin. The cells from the hidden abaxial surface are about 60 ^mx 10 fim large and elongated along the cushion length. Small fragments of cuticle with cells about 50 pmx 15 pm large were obtained from some of the larger specimens and although their exact origin is unknown it is suggested that they may have come from the intercushion connecting areas of the stem surface. All the cells have straight, smooth anticlinal walls and flat, smooth periclinal walls. Stomata are present almost solely on the exposed cushion surface (about 180 per mm2) although a few have been noted on the hidden surface —near the edges where the cells are also not quite so elongated. The average stomatal size is about 35 pm x 26 with their guard cells superficial or only very slightly sunken. Details of guard cell anatomy have already been given (Thomas 1974, p. 530) and I can add no more to this at present. The ligule pit cuticles have rect- angular cells, 25-30 ^mxlO /xm large, arranged in vertical rows, with straight, smooth anticlinal walls and flat, smooth periclinal walls. Comparison. The leaf cushions of L. acerosus and L. laricinus can be distinguished by a number of characters. The cushions of L. acerosus are relatively longer and possess definite keels which are never present in the other species. Ligule pit positions are also important as they are normally immediately above and adjacent to the leaf scars in L. acerosus but a short distance above them in L. laricinus. The cuticle preparations support the distinction between the two species as the epidermal arrangement is different above and below the leaf scars in L. acerosus but in L. lari- cinus there is no such distinction. Also the guard cells are sunken in deep pits in L. laricinus while they are superficial or only very slightly sunken in L. acerosus. Various emphases have been previously laid on these various morphological characters in an attempt to separate the two species, although no attempts have been made to utilize epidermal characters. Renier (M. S. in Crookall 1964, p. 312) text-fig. 4. Lepidophloios acerosus Lindley and Hutton. A, BMNH V33591 showing undeflected leaf cushions, x 3. B, IGSLds RC 1778 showing deflected leaf cushions, x 3. c, cushion cuticle from above the leaf scar of V33591, x 400; slide no. V33591a. d-g, cuticle from RC 1778, x 400. D, cushion cuticle from below leaf scar; slide no. PL 279. E, cuticle from cushion edge; slide no. PL 282. F, cushion cuticle from below leaf scar; slide no. PL 279. Arrows directed parallel to the vertical axis of the cushion. G, ligule pit cuticle; slide no. PL 280. 284 PALAEONTOLOGY, VOLUME 20 believed the positioning of the ligule pits to be the most important feature while Nemejc (1947, p. 78) stressed the relative lengths of the cushions and the presence or absence of keels. In contrast to these, Jongmans (1930) suggested that L. acerosus might represent young stems of L. laricinus and Stockmans and Williere (1953) similarly believed them to be conspecific. The two species are distinct, but it is certainly better to use more than one cushion character for differentiation, for, as Nemejc (1947) and Crookall (1964) have already shown, individual characters are neither always well marked nor constant. The wide range of cushion characters quoted by Crookall is, however, unacceptable and I have here included many of his examples as other species. L. acerosus has been confused with other species, a problem which has been accentuated by misinterpretation of cushion orientation. Specimens have been figured inverted, as indeed have been specimens of L. laricinus. L. acerosus , if figured upside-down (e.g. Arber 1914, pi. 28, fig. 20; Kidston 1917, pi. 2, fig. 5; Crookall 1929, pi. 22, fig. k; 1964, pi. 79, fig. 5), could be mistaken for a species of Lepi- dodendron. This could then account for Nemejc’s (1934) view that there was no difference between Lepidodendron dichotomum Sternberg and the figures published for Lepidophloios acerosus by West European authors. He thus believed the two species to be closely allied, if not identical, although he did later revert to giving them as distinct species (Nemejc, 1946, 1947). A similar confusion seems to have arisen when Weiss (1871) joined Lepidodendron brevifolium Ettinghausen (1854) with parts of the Lepidophloios laricinus of Goldenberg (1862) and Schimper (1870) under his new name L. carinatus. Kidston (1886) then referred L. acerosus and Lepidostrobus pinaster Lindley and Hutton to Lepidophloios carinatus Weiss although he later included all in his synonymy for L. acerosus (Kidston 18936, 1894). L. pinaster seems to be correctly reidentified as L. acerosus but mistakenly believed to be inverted. Jongmans (1929) and Crookall (1964) also appear correct in maintain- ing that L. brevifolium is a species of Lepidodendron and that it should therefore be excluded from Lepidophloios acerosus. Lepidophloios macrolepidotus Goldenberg Plate 35, figs. 3-6; text-fig. 5 1855 Lomatophloyos macrolepidotum Goldenberg, p. 22. 1862 Lepidophloios macrolepidotum Goldenberg, p. 37, pi. 14, fig. 25 (inverted). 1870 Lepidophloios macrolepidotus Goldenberg; Schimper, p. 52. 1882 Lepidophloios macrolepidotus Goldenberg; Renault, pars, p. 45, fig. 2 (inverted) not fig. 4. 1890 Lepidophloios macrolepidotus Goldenberg; Seward, pi. 3, figs. 1-4. 1899 Lepidophloios macrolepidotus Goldenberg; Potonie, p. 235, fig. 223. 1959 Lepidophloios macrolepidotus Goldenberg; Remy, p. 83, fig. 81. 1964 Lepidophloios laricinus Sternberg; Crookall, pars, p. 307, pi. 78, fig. 1 (inverted). 1967 Lepidophloios macrolepidotus Goldenberg; Chaloner, p. 571, fig. 390. Material. K 3256-3257, and IGSLond RC 2910 from above the Fenton Coal, Dodworth Colliery, near Barnsley, Yorkshire— communis Zone, Westphalian A; K4392 from above the Halifax Hard Bed, Field- house Colliery, Deighton, Yorkshire— lenisulcata Zone, Westphalian A. The largest slab of bark was K 3257, being 180 mm broad, which also had the largest leaf cushions of average visible length 16 mm and breadth 25 mm. All the THOMAS: LEP I DO P H LO I O S EPIDERMIS 285 specimens, however, have leaf cushions of roughly comparable size and shape, being all broader than long with flattened surfaces possessing no keels. The leaf scars have three obvious prints in their lower halves, while the conspicuous ligule pit apertures are triangular in outline and are clearly separated from the leaf scars. The specimens also show a few cushions which appear to be separated from each other, but this is really an illusion caused by narrow strips of adhering shale which can be removed showing the cushions to be continuous and overlapping in the normal manner. Goldenberg (1862) and Renault (1882) figured similar specimens text-fig. 5. Lepidophloios macrolepidotus Goldenberg. K 3256. a, leaf cushions, xl, 1— ligule pit aperture, b, cushion cuticle, x 400. Arrows directed parallel to the vertical axis of the cushion; slide no. PF 2897. with cushions separated by bands of what appear to be bark, but these could again be only overlying strips of rock matrix. Cuticle was prepared from K 3256-3257, and IGSLond RC 2910 but as K 4392 has but a single leaf cushion it was thought unwise to remove any of its compression. Cuticle description. The epidermis is the same on the exposed and hidden cushion surfaces. Epidermal cells are longitudinally elongated, about 30-60 /xmx 10-15 pcm large, often with pointed ends. Anticlinal walls are straight, smooth, and 1 /. mr thick and the periclinal walls are flat and smooth. Stomata are about 50 per mm2, of average size 40 ju m x 30 ^m, and possessed superficial guard cells. No ligule pit cuticles could be prepared. Comparison. Kidston (1893a, p. 80; 1911, p. 151) suggested that L. macrolepidotus was a larger form of L. laricinus and Crookall (1964, p. 310) doubtfully united the A D 286 PALAEONTOLOGY, VOLUME 20 two. All the described specimens of L. macrolepidotus seem to be of large slabs of bark presumably coming from the main trunk or large branches; but while accepting that there were possibly smaller branches bearing smaller leaf cushions it is not yet proven that these were necessarily of the L. laricinus kind. L. macrolepidotus has leaf cushions which are more flattened than those of L. laricinus and they possess ligule pit apertures which are more distinct and relatively more separated from the leaf scars. The epidermal cells are about the same size in the two species but the anticlinal walls are thicker in L. laricinus (2 pm) than in L. macrolepidotus (1 p m). L. laricinus has 250 stomata per mm2 with guard cells in pits (6-10 deep) but L. macrolepidotus has less stomata (50 per mm2) and superficial guard cells. The ‘young branches’ figured by Renault (1882, fig. 4) are excluded from this species in agreement with Kidston (1893/)) and Crookall (1964). They seem to be more like L. acerosus or to a very similar species. The specimen described by Kidston (1893/)) from the lower Carboniferous of Scotland is also excluded and described here as the holotype of L. grangeri sp. nov. Lepidopldoios grangeri sp. nov. Plate 36, figs. 1-4; text-fig. 6 18936 (?) Lepidophloios macrolepidotus Goldenberg; Kidston, p. 560. 1964 Lepidophloios laricinus Sternberg; Crookall, pars, p. 310, pi. 74, fig. 6 (inverted). Material. Holotype, K 1828 from above the Craw Coal, No. 4 Mine, Grange, Boness, Linlithgowshire— within the Limestone Coal group of the Carboniferous Limestone Series of the Namurian (D3 Coral brachiopod Zone and E, age of the goniatite notation, according to Macgregor in Trueman 1954). The type specimen, which is the only known one referable to this species, is a 100 mm broad slab of compressed bark on an ironstone nodule. Cuticle was easily obtained from the exposed upper cushion surfaces but could not be prepared from the overlapped underlying lower surfaces. The ligule pit cuticles were unfortunately very cracked and only small fragments were therefore obtained. Diagnosis. Exposed portions of leaf cushions broader than long. Ligule pit apertures adjacent to upper angles of leaf scars. Cushion surface smooth with no keel. Epider- mal cells from central area of cushion longitudinally elongated, about 60-70 ^m x 15 pm large. Epidermal cells from sides of cushion roughly isodiametric, 15-20 pm large. Anticlinal walls straight, smooth, 1 pm thick. Periclinal walls flat, smooth. Stomata about 150 per mm2 over the whole cushion surface; average size 40 x 25 pm. Guard cells level with epidermal surface. Ligule pit about 230 pm broad; lining cells rectangular, longitudinally elongated, about 45-55 pmx 12 large. Derivation of name. From the type locality. EXPLANATION OF PLATE 36 Figs. 1-4. Lepidophloios grangeri sp. nov. 1, K 1828 from above the Craw Coal, No. 4 Mine, Grange, Boness, Linlithgowshire, x 2. 2, 3, cushion cuticle, slide no. PF 2900. 2, X200; 3, x 600. 4, ligule pit cuticle; slide no. PF 2898, x 100. Figs. 5-7. Lepidophloios acadianus Dawson from Joggin Mine, Nova Scotia, Canada, 5, K 2318, x0-5. 6, K 2323, x 1. 7, cushion cuticle from K 2323; slide no. PF 3131, x200. PLATE 36 THOMAS, Lepidophloios 288 PALAEONTOLOGY, VOLUME 20 Comparison. The type specimen has been included within L. macrolepidotus and L. laricinus , but it can be distinguished from both of these species on cushion morpho- logy and cuticle characters. The ligule pit aperture is adjacent to the leaf scar in L. granger i but is clearly separated from the scar in the other two species. The epidermis has elongated cells in the central area and isodiametric cells on the lateral portions so it is similar to that of L. laricinus, but different to that of L. macro- lepidotus which has only elongated cells. L. laricinus differs in having shorter cells in the central region even though the cells from the lateral areas are of comparable size to those in L. grangeri. The stomata are of comparable size in all three species, but their frequencies differ. They are 1 50 per mm2 in L. grangeri, 50 per mm2 in L. macrolepidotus, and 250 per mm2 in L. laricinus. The guard cells are also sunken in pits in L. laricinus whereas in the other two they are superficial. L. grangeri also differs from the other species of Lepidophloios described here in coming from the Lower and not from the Upper Carboniferous. The other Lower Carboniferous species, L. scoticus Kidston, has been shown to possess a variety of leaf cushion sizes and shapes, but none is really like those of L. grangeri. The largest text-fig. 6. Lepidophloios grangeri sp. nov. K 1828. a, leaf cushions, X 1. 1— ligule pit aperture. B, c, cushion cuticle, x400; drawn from the undersurface, b, cuticle from the non-median areas; slide no. PF 2900. c, cuticle from the median areas; slide no. PF 2899. THOMAS: LEPIDOPHLOIOS EPIDERMIS 289 A B text-fig. 7. Lepidophloios acadianus Dawson, a, K 2324, xl.B, K 2323, x 1 . c, d, cushion cuticle from No. 2323, x 400, st— stomata, c, cuticle from the median area of the cushion; slide no. PF 3131. d, cuticle from the non-median area of the cushions ; slide no. PF 3 1 32. cushions of L. scoticus have exposed areas which are longer than broad in contrast to those of L. grangeri. The lower edges of the leaf scars are also much flatter and the ligule pit apertures are separated from the leaf scars. No cushion cuticles have yet been described from L. scoticus so no comparison can be made of epidermal features, but even without this additional information the two types of cushion appear sufficiently different for species distinction. Lepidophloios acadianus Dawson Plate 36, figs. 5-7; text-fig. 7 1866 Lepidophloios acadianus Dawson, pp. 163, 168, pi. 10, fig. 45. 1868 Lepidophloios acadianus Dawson, p. 489; text-fig. 171. 1888 Lepidophloios acadianus Dawson, p. 166; text-fig. 44. Material. K. 2318-2324 from Joggin, Canada. All seven specimens were identified by Dawson as L. acadianus and K 2324 is labelled as ‘fragment of type’. The exposed areas of the downturned cushions are 290 PALAEONTOLOGY, VOLUME 20 all broader than long, with the leaf scars at their extreme bases except in K 2318 where a small portion of the cushion can be seen below the scars. Cushion surfaces are flat with no keels and the ligule pit apertures are clearly separated from the leaf scars. Cuticle description. The epidermis is the same from the cushion surfaces above and below the leaf scars. The epidermal cells on the central areas of the cushions are elongated longitudinally and are 40-70 ^mx 18-30 large, while the cells on the cushion are roughly isodiametric and about 15 /xm large. Stomata are about 130 per mm2 and of average size 45 ^m x 33 ^m. The guard cells are usually level with the epidermis but are occasionally sunken in 3 deep pits. Comparison. Dawson published three identical accounts of this species but figured the leaf cushions upside down and without ligule pits. He believed L. acadianus to be closely allied to Ulodendron majus Lindley and Hutton and L. laricinus Sternberg; while Kidston (1901, p. 158), Bell (1944, p. 93), and Crookall (1964, p. 31 1) believed it to be conspecific with L. laricinus. U. majus is a completely different type of stem which possess permanently attached leaves and not downwardly directed leaf cushions (Thomas 1 967Z?) so it has little in common with this species. L. laricinus is naturally similar, but can be distinguished by both cushion morphology and cuticle characters. The exposed parts of the cushions are relatively narrower with more rounded lower angles, the foliar prints are relatively higher on the leaf scars and the ligule pits are more distinct. L. laricinus also has larger epidermal cells, more stomata than L. acadianus , and the guard cells are normally sunken in pits. DISCUSSION Lepidophloios, like the other genera of arborescent lycopods, has been interpreted differently by various authors. Confusion and differences of opinion have existed over the range of species variation and over the very number of species which were thought to exist. Previous workers have used just leaf cushion morphology to identify their specimens, but recent work on other genera (Thomas 1967a, b, 1970) has shown how epidermal characters are clearly of immense value for this purpose. Therefore epidermal cell sizes and shapes, stomatal sizes, numbers, and distributions were studied in a range of species to see if they were similarly useful in this genus. L. laricinus and L. acerosus are the two commonest species which are relatively easily distinguished on a combination of cushion characters, but at times they have been thought to be conspecific and have been also linked with other species of Lepidophloios and even Lepidodendron. Clearly the differences in cushion morpho- logy can be thought to be insufficient for species distinction so there is some diver- gence of opinion here. However, if one takes the two ‘recognizable forms’ and looks at their epidermal details there are additional characters available for comparison. In this instance, such extra information clearly points to a continued separation of the two species. Similar epidermal evidence indicate that Lepidophloios macro- lepidotus, L. grangeri , and L. acadianus are recognizable as distinct species and can not be thought of as growth forms of L. laricinus as has been often suggested. Epidermal studies are thus once again of great value and the information gained THOMAS: LEP1DO P H LO IO S EPIDERMIS 291 has helped to crystallize a better idea of the various species. It has hopefully provided a sounder basis for understanding the range of species variation and for identifying new material. The other important aspect of this type of study is that it should allow us to recog- nize growth forms of the various species. Previous work by Walton (1935), Andrews and Murdy (1958), and Eggert (1961) has suggested that Lepidodendralean trees grew by dichotomizing apices which progressively diminished in size. Therefore shoot diameter is no indication of age for the smaller shoots were merely more terminal and were not young shoots which had not yet enlarged by large amounts of secondary thickening, i.e. large shoots were formed from large apices while small shoots were formed from small apices. Some secondary thickening did occur but only in a manner which seems to have accentuated the primary growth form, for apparently the trunk and main branches were thickened much more than the smaller terminal shoots. Leaf growth also varied proportionally to shoot diameter, so in both Lepidodendron and Lepidophloios the narrower branches have smaller leaf cushions while only the very terminal shoots seemed to retain the distal foliage parts of their leaves. There are, however, certain major differences which existed between these two genera regarding the effects of limited secondary growth on the leaf cushions. Shoot expansion in Lepidodendron apparently did not initially affect the actual leaf cushions, but separated them instead by a gradual growth of the inter- cushion areas (Thomas 1966). In Lepidophloios , however, the situation appears to be rather different for the leaf cushions apparently never separated as in Lepidodendron. Those specimens suggesting such a separation (e.g. Lepidophloios macrolepidotus ) have now been shown to be rather different with the intercushion areas being really narrow strips of shale protruding from between the overlapping leaf cushions. Instead of separating, the cushions appear to have enlarged and bulged further outwards and downwards. Indeed, the specimens of L. acerosus discussed above indicate that the cushions were originally of the Lepidodendron type. Then by expansion they would have bulged outwards and downwards synchronously with the enlargement of the shoot. Obviously much more evidence is needed to clarify our ideas of this par- ticular method of growth and it would be much better if such stages could be demon- strated in species other than Lepidophloios acerosus. Unfortunately we are dealing with growth stages and this is always a major problem, because it is the growing points which are the least likely to become fossilized. The other point of interest which centres around such a peculiar type of shoot growth is the usefulness of cushion enlargement to the growing plant. The question of photosynthetic efficiency of the arborescent lycopods has been broached several times but not in direct relationship to Lepidophloios. Andrews and Murdy (1958) and Andrews (1961) thought these plants possessed relatively small amounts of photosynthetic tissue because only the smallest twigs retained their leaves. Then the demonstration of numerous stomata on the leaf cushions suggested that the stems were much more photosynthetic than previously thought (Thomas 1966). While Chaloner and Collinson (1975) have since shown that Sigillaria possessed even more stomata per unit area than Lepidodendron and suggested that the increased amount of potential photosynthetic activity might help to explain their ability to grow with only a crown of leaves. What we may see in Lepidophloios is a further attempt to 292 PALAEONTOLOGY, VOLUME 20 increase the photosynthetic ability of these plants, for such an increase in cushion size would appear to result in the production of more photosynthetic tissue. Perhaps we could take this to be the very reason for leaf cushion enlargement and the evolu- tionary change from the Lepidodendron type of cushion. Acknowledgements. I thank Professor T. M. Harris, F.R.S., for his guidance during the early part of this study; the directors and staff of the museums for allowing my access to their collections; the University of Reading Research Board, the University of Newcastle upon Tyne, the Royal Society of London, and the University of London Research Board for grants enabling me to carry out various parts of this study. The photomicrographs were taken on a Zeiss Photomicroscope III, whose purchase was financed by a grant from the Royal Society. REFERENCES Andrews, h. n. 1961. Studies in Palaeobotany. John Wiley and Sons, New York. 487 pp. and murdy, w. h. 1958. Lepidophloios— an ontogeny in arborescent lycopods. Am. J. Bot. 45, 552-559. arber, e. a. n. 1914. On the fossil floras of the Wyre Forest, with special reference to the geology of the coalfield and its relationships to the neighbouring Coal Measure areas. Phil. Trans. R. Soc. Ser. B , 204, 365-445. bell, w. a. 1944. Carboniferous rocks and fossil floras of Northern Nova Scotia. Mem. geol. Surv. Can. 238, 276 pp. chaloner, w. G. 1967. In boureau, E. (ed.). Traite de Paleobotanique. II. Masson et Cie, Paris, 437-485. and collinson, m. e. 1975. Application of SEM to a sigillarian impression fossil. Rev. Palaeobot. Palynol. 20, 85-101 crookall, r. 1929. Coal Measure plants. Edward Arnold, London. 77 pp. 1939. The plant ‘break’ in the Carboniferous rocks of Great Britain. Bull. geol. Surv. G.B. 1, 13-24. — 1964. Fossil plants of the Carboniferous rocks of Great Britain. Mem. Geol. Surv. G.B. Palaeontology , IV, pt. 3,217-354. dawson, j. w. 1866. On the Conditions of the Decomposition of Coal, more especially as illustrated by the coal formation of Nova Scotia and New Brunswick. Q. Jl geol. Soc. Lond. 22, 95-169. 1868. Acadian Geology. 2nd edn. Oliver and Boyd, Edinburgh. 694 pp. 1888. Geological Survey of Plants. D. Appleton and Co., New York. 290 pp. eggert, D. a. 1961. Ontogeny of Carboniferous Arborescent Lycopsida. Palaeontographica. 108B, 43-92. ettingshausen, c. von. 1854. Die Stein kohlenflora von Radnitz in Bohmem. Abhandl. k.k. geol. Reichsant. 1(3), No. 3. Vienna. 74 pp. feistmantel, o. 1875. Die Versteinerungen der Bohmischen Kohlenablagerungen. Palaeontographica , 23, 175-236. GOLDENBERG, f. 1855- 1 862. Flora Saraepontana fossils— Die Pflanzenversteinerungen des Steinkohlengebirges von Saarbriicken. I (1885), 38 pp. ; II (1857), 60 pp. ; III (1862), 47 pp. jongmans, w. j. 1929. Fossilium Catalogus II, Plantae Pars 15, Lycopodiales II, 53-523. W. Junk, Berlin. 1930. Ibid., Pars 16, Lycopiodes III, 329-650. kidston, R. 1886. Catalogue of the Palaeozoic Plants in the Department of Geology and Palaeontology , British Museum ( Natural History). London. 288 pp. 1893a. The Yorkshire Carboniferous Flora. Trans. Yorks. Nat. Union , 18, 65-128. — 18936. Lepidophloios , and on the British species of the Genus. Trans. R. Soc. Edinb. 37 (3), 529-563. — 1894. Fossil Flora of the South Wales Coal Field and the Relationship of its strata to the Somerset and Bristol Coal Field. Ibid., 565-614. — 1901. Carboniferous Lycopods and Sphenophylls. Trans, nat. Hist. Soc. Glasg. 6 (n.s.), 25-140. 1902. Flora of the Carboniferous Period. Proc. Yorks, geol. polytech. Soc. 14, 334-370. — 1911. Les vegetaux houilleres recueillis dans le Hainaut Beige. Mem. Mus. r. Hist. nat. Belg. 4, 282 pp. — 1917. The Forest of Wyre and the Titterstone Clee Hill Coal Fields. Trans. R. Soc. Edinb. 51, 999- 1084. THOMAS: LEP I DO P H LO IO S EPIDERMIS 293 lindley, J. and hutton, w. 1831 1837. The Fossil Flora of Great Britain: or Figures and Descriptions of the Vegetable remains found in a Fossil State in this country. 3 vols. London. I (1831-1833), pp. li + 223; II (1834-1835), pp. 206; III ( 1836-1837), pp. 205. NEMEJC, F. 1934. Critical remarks on Sternberg’s Lepidodendron dichotomum. Bull. int. Acad, tcheque Sci. 35 ann., 75-79. 1946. Further critical remarks on Sternberg’s Lepidodendron dichotomum and its relations to the cones of Sporangiostrobus Bode. Ibid. 47 ann. 35-43. - 1947. The Lepidodendraceae of the coal districts of central Bohemia (a preliminary study). Acta Musei Nationalis Prague , 3B, 2, 45-87. potonie, H. 1899. Lehrbuch der PJlanzenpalaeontologie. Ferd, Diimmlers Verlagsbuchhandlung, Berlin. 402 pp. remy, w. and remy, r. 1959. Pflanzenfossilien. Akademie Verlag, Berlin. 285 pp. Renault, B. 1882. Cours de botanique fossile. G. Masson, Paris. 194 pp. schimper, w. p. 1870. Traite de Paleontologie Vegetale, ou la Flore du Monde primitif dans ses reports avec les formations geologiques et la Flore du Monde actuel. Vol. 2. J. Bailliere et Fils, Paris. 522 pp. Seward, A. c. 1890. Notes on Lomatophloios macrolepidotus (Goldg.). Proc. Camb. phil. Soc. 7 (2), 43-47. Sternberg, c. von. 1820-1838. Versuch einer geognotisch-botanischen. Darstellung der Flora der Vorwelt. Leipzig and Prague. Pt. 1 (1820), 24 pp.; pt. 2 (1821), 33 pp.; pt. 3 (1823), 39 pp.; pt. 4 (1824), 42 pp.; pts. 5 and 6 (1833), 80 pp. ; pts. 7 and 8 (1838), 291 pp. stockmans, F. and williere, v. 1953. Vegetaux Namuriens de la Belgique. Assoc. Etude Paleont. Stratigraph. Houilleres, Publ. 13, Brussels. 382 pp. thomas, B. a. 1966. The cuticle of the Lepidodendroid stem. New Phytol. 65, 296-303. 1967a. The cuticle of two species of Bothrodendron (Lycopsida; Lepidodendrales). J. nat. Hist. 1, 53-60. 19676. Ulodendron Lindley and Hutton and its Cuticle. Ann. Bot. N.s. 31, 775-782. 1970. Epidermal studies in the interpretation of Lepidodendron species. Palaeontology , 13, 145 173. 1974. The Lepidodendroid stoma. Ibid. 17, 525-539. walton, j. 1935. Scottish Lower Carboniferous Plants; The Fossil Hollow Trees of Arran and their Branches. Trans. R. Soc. Edinb. 58, 313-337. weiss, c. E. 1871. Fossile Flora jungsten Steinkohlen-formation und des Rothliegenden im Saar-Rhein Gebiete. K. Akad. d. Wissench. Berlin , Teil 2, Heft 2, 141-212. Zalessky, m. d. 1904. Vegetaux fossiles du terrain carbonifere du Donetz, I, Lycopodiales. Trudy geol. Kom. N.s. 13, 126 pp. B. A. THOMAS Department of Biological Sciences University of London, Goldsmiths’ College Typescript received 22 January 1976 jqew (^ross Revised typescript received 29 March 1976 London SE14 6NW ZYGOSPIRA AND SOME RELATED ORDOVICIAN AND SILURIAN ATRYPOID BRACHIOPODS by PAUL COPPER Abstract. The Zygospiridae are redefined to include ‘primitive’ atrypoid brachiopods with dorsally to dorso- medially directed spiralia, normally with fine ribs and a one-piece jugum dorsal to the spiralia. The Ordovician- Silurian family is divided into three: Zygospirinae, Catazyginae (new sub-fam.), and Tuvaellinae. Internal structures of Zygospira, Anazyga, Catazyga , Pentlandella, and Tuvaella are evaluated in terms of their evolutionary significance. A new genus Zygatrypa is erected. Eospirigerina , which was probably derived from zygospirid stock in the late Ordovician, differs in having ventrally located, separated jugal processes and trends towards ‘frilly’ shells: this represented a major jump towards a postulated Zygospiraella-Protatrypa-Gotatrypa lineage. In late Caradoc- Ashgill times Catazyga inhabited deeper water, whereas Zygospira lived in shallow-water communities. The Clinton- ellinae, including Alispira , are tentatively assigned to the Atrypidae, with Silurian taxa such as Nalivkinia and Anabaria. Internal structures of many of the earlier ribbed atrypoids belonging to the family Zygospiridae, which first appeared in Caradoc time, are not generally well known. The pioneering morphological work of Hall (1862, 1893), Davidson (1882), and Schuchert (1893) on the zygospirids has not been corroborated or used in classi- fications, except in the broadest sense. Data is lacking on apical shell structures (deltidial plates, ‘dental plates’, pedicle collars, etc.) and brachidial structures (socket plates, crura, jugal processes). It is the intent of this paper to fill in some of the gaps in our knowledge of the group, so as to determine generic variation and evolutionary trends. The taxa Anazyga , Zygospira , and Catazyga form the centre of this study, but some related ribbed atrypoid genera are also examined. Serial sections demonstrate the affinities of Pentlandella , Tuvaella , Alispira , ‘ Clintonella', Nalivkinia , and Anabaria. The first two are retained within the zygospirids, but the others are removed. Zygospiraella is interpreted as an ancestor of the Protatrypa- ‘ Gotatrypa'-Atrypa lineage: this will be described in a subsequent paper. Smooth Ordovician spire-bearing genera, such as Protozvga , Idiospira , and Cyclospira , are not within the scope of this paper, but are also under revision. EVOLUTION The main bloom of the group took place during the late Caradoc and Ashgill, but by latest Ordovician time the zygospirids in eastern North America and western Europe had been replaced by Eospirigerina. During this interval, and in the early Llandovery in this region, there appears to be no trace of any zygospirids. Never- theless, by Upper Llandovery (Telychian) time, two zygospirids reappear, Pentlan- della in Britain and Estonia, and Zygatrypa , described herein from Anticosti Island. Tuvaella , from the Tuva region of Asian U.S.S.R. and Mongolia, also appeared in [Palaeontology, Vol. 20, Part 2, 1977, pp. 295-335, pis. 37-40.] 296 PALAEONTOLOGY, VOLUME 20 text-fig. 1. The most important evolutionary trends in the Zygo- spiridae lie with the disposition and nature of the spiralia and the jugum. The spiralia tend to produce increased numbers of whorls, a change from medially to dorsally directed coils, and the jugum shifts in posi- tion from an anterior to a posterior location with an eventual position near the pedicle valve. These trends probably reflect more efficient suspension feeding and changes in the location of the digestive tract. late Llandovery or early Wenlock time, but was endemic to that area. Tuvaella ranged into late Silurian time (Pridoli?), and appears to have been the last-surviving zygospirid. The origin of the earliest atrypoids is unclear. If Protozyga is the common ancestor, a smooth impunctate prototype form is indicated, possibly a camerellid or smooth rhynchonellid. However, if Protozyga represents an offshoot from Anazyga, a ribbed rhynchonellid, or even an orthid, may be the root-stock. We are no closer to an answer than at the time of Cooper’s monograph on the Chazyan brachiopods (1956). COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 297 The geological succession in Ontario is for Protozyga to appear first, followed by Anazyga , then Zygospira and Catazyga. A general trend in the Zygospiridae is towards increasing size in all three sub- families, for example, Zygospira itself reached maximum width in Ashgill time and the largest catazyginids are late Ordovician Catazyga. However, Zygatrypa and Pent/andel/a , offshoots of Ordovician genera, are very small by comparison and appear to show a reversed late trend towards decreasing size and finer ribbing. The origin of Tuvaella lies in the Zygospiridae, but perhaps not with Zygatrypa. Tuvaella is the largest known zygospirid, reaching widths of some 60 mm with T. gigantea Cherny- shev in late Silurian time (Vladimirskaya 1972). Even ‘large’ late Ordovician zygo- spirids such as Catazyga and Zygospira rarely exceed 15 mm. Another trend is rib coarsening in the zygospirinids, leading from Anazyga to early Zygospira and late Zygospira in the Ordovician, and also in the Catazyga group. However, in Silurian time, the zygospirids tend to be very fine ribbed, except for Tuvaella. A reverse trend, towards loss of ribs, may have occurred in some catazyginids. Ulrich (1888, pp. 196-197) described a late Ordovician species of catazyginid, which he called Glassia schuchertana , because ribs were barely, if at all, visible on many well-preserved specimens. Foerste (1910) assigned it to Catazyga , as did Meek (1873) who identified it as the same as the type species of Catazyga , C. headi (Billings 1862). Silurian Pentlandella from Estonia also possess such fine ribs that these are frequently not visible on parts of the anterior shell. It is possible that the loss of ribs resulted in the appearance of Glassia in Llandovery time, since the spiralia in Glassia are medially directed, as in Catazyga. Internal trends differ in individual sub-families but are probably more significant. Anazyga has a fused jugum located antero-dorsally to the cone axes, and the spiralia are medially directed (cone axes parallel or near parallel to the commissural plane). Its descendant Zygospira has a dorsal jugum located at the cone axes, or posterior to the cone axes, and spiralia directed dorso-medially (cone axes some 45° from the commissural plane). If the two taxa form a continuous lineage from Caradoc through Ashgill time, then the jugum migrated from an anterior position to a posterior position and the cone axes rotated some 45° from the commissural plane in a dorsal direction. In addition, the same lineage shows an increasing trend from only one or two spiralial whorls in Anazyga to three to five in Zygospira. It is perhaps significant that in the Atrypidae of the Siluro- Devonian, thirty or more spiral whorls are known and the separated jugal processes are located antero-ventrally. Eospirigerina in the late Ordovician Ellis Bay Formation of Anticosti Island has eight to nine spiral whorls and unconnected jugal processes located antero-ventrally. This clearly separates Eospirigerina from the Zygospirinidae, from which it must have evolved. The dorso-medial turning of the cone axes culminated in almost full dorsal directions in the latest Ordovician (text-fig. 1). It is significant that the. location of the jugum in the Zygospiridae is dorsal: the jugum is derived from the initial whorl of the spiralium but is directed dorsally and lies very close to the median septum of the brachial valve, frequently arching over it. This implies that the jugum may have rested or been suspended over the brachial valve floor. The jugal processes in the Atrypidae and Palaferellidae are ventral, i.e. the jugal process is located ventrally and posterior to the spiralia and thus lies closest to the pedicle valve. Thus not only 298 PALAEONTOLOGY, VOLUME 20 do these other families differ in having a divided jugum, or jugal processes (Copper 1967), but the implication is that if the jugum-jugal processes held mouth parts, their feeding processes would be quite different. In the Catazyginae there are different internal trends. Catazyga probably evolved from Anazyga in the late Caradoc. It is uncertain whether to assign Anazyga to the Catazyga group or to the Zygospira group. Externally Anazyga is closer to Zygospira in its wider shell, fold-sulcus, and beak structure. Internally it could belong to either, except that Anazyga has dental cavities, like Zygospira , and Catazyga has none. If Catazyga evolved from Anazyga , as seems most likely, then this involved a loss of dental cavities and great thickening of the pedicle valve anteriorly. Catazyga , and its Silurian descent Pentlandella , both retain spiralia which are strongly directed towards the centre of the shell. The cone axes show a lesser rotation from the com- missural plane than in the zygospirinids. Catazyga has about seven spiral whorls. Rubel (1970) recorded four whorls for Estonian Pentlandella , but an adult specimen sectioned for this paper, from the same Estonian locality, revealed only two (text- fig. 7). Since Silurian Pentlandella are about half the size of Ordovician Catazyga this does not necessarily reflect an independent evolutionary reduction in spiralia, but is probably a size-related factor. The structure of the jugum and spiralia in Silurian Zygatrypa is similar to that of Zygospira , except in the location of the jugum. In Tuvaella rackoviskii from the Tuva region and Mongolia (text-fig. 14) no dental cavities are present, but the jugum is located dorsally and its central point lies close to the cone apices. This is like other zygospirinids. Functionally this suggests that the central part of the jugum held a mouth organ to which food currents were canalized. The cone axes are directed as in Zygospira , and five to eight whorls were present. In later T. gigantea , Vladimirskaya (1972) recorded up to ten whorls, which seems to be a maximum figure for any Zygospiridae. It is possible that the family gave way to the Atrypidae and other groups because these had the ability to use larger numbers of spiral whorls and were thus more efficient filter feeders. THE ORDOVICIAN-SILURIAN BOUNDARY Atrypoid brachiopods of the transitional period from Ordovician to Silurian time are not well known. In the Anticosti section in eastern Canada, which holds a richly fossiliferous suite of rocks straddling the ‘boundary’, the zygospirids were halted in the Vaureal Formation and do not continue into the Ellis Bay Formation, most of which (on the basis of brachiopods, rugose and tabulate corals, and stromato- poroids) is still in the Ordovician. I agree with Bolton (1972) that the base of the Becscie Formation with Zvgospiraella (Atrypidae) corresponds very closely to the boundary. The zygospirids, therefore, are not of great value in eastern North America in determining the boundary. They were in decline some time before the close of the Ordovician, with only two conservative lines maintained as rare elements in the Llandovery {Pentlandella and Zygatrypa ). Amsden (1971) described a transitional fauna from the Edgewood Formation of Illinois and Missouri including ‘ Eospirigerina putilla (Hall and Clarke, 1893), which is said to be present in the latest Ordovician as well as early Llandovery. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 299 This species may be a true Zvgospira , as described by the original authors, and needs redescription. Eospirigerina is so abundant in the Ellis Bay Formation of Anticosti Island that it forms distinctive shelly horizons; several species are present. The species group of ‘ Plectatrypa' , is readily confused with Eospirigerina in the late Ordovician (but not in the Silurian of Western Europe). 1 Eospirigerina ' hibernica Reed (1932) from Northern Ireland is a late Ordovician form first figured by Port- lock (1843, pi. 37, fig. 4). ‘ Plectatrypa ' partita (Sowerby, 1839) from Goleugoed Hill (Lower Llandovery) in Wales is a similar early Silurian form. The Eospirigerina- ‘ Plectatrypa' group is complex but could perhaps shed light on the boundary problem. Zygospiraella at present is the most useful atrypoid indicator of early Llandovery age. shallower text-fig. 2. During Ashgill time, zygospirids were dominant in two assemblages; a shallower-water Zygo- spira assemblage with bryozoans, tabulates, and colonial rugose corals, and a deeper, quiet water assemblage of Catazyga and sowerbyellids (examples drawn from Z. modesta and C. headi; no ribs drawn on latter). Scale 5/3 x . PALAEOECOLOGICAL DISTRIBUTION The earlier zygospirids Protozyga and Anazyga were probably moderately deep- water benthos. Their common occurrence is in dark-grey to black calcareous shales or muddy limestones. Common associates of Anazyga are other small brachiopods, small hemispherical colonies of Prasopora and ostracodes. Anazyga does not occur abundantly with bioherms or biostromes with stromatoporoids, tabulate or rugose corals, or with digitate bryozoans in higher energy, shallow-water zones. By later Ordovician time, zygospirids were ecologically sorted into two communities or habitats: Zygospira with coarser ribs, well-defined fold-sulcus, and planoconvex- ventribiconvex shells, lived in a shallow-water community with bryozoans (Bretsky 1969; Richards 1972), and Catazyga in a deeper-water, muddy environment, frequently nearly barren of anything but crowded catazygid nests (text-fig. 2). The 300 PALAEONTOLOGY, VOLUME 20 two genera rarely occur together except in ratios of 100:1 or more. In the Vaureal Formation on Anticosti Island Zygospira is apparently absent and Catazyga is abundant. Towards Manitoulin Island, some 400 miles westwards, Zygospira is locally extremely abundant but Catazyga occurs only in isolated patches. A similar presence absence occurrence is described from Indiana and Ohio by Richards (1972, p. 402). The two taxa seem to be reliable relative depth indicators, but at present actual depths are problematical. It is possible that Zygospira represented the depth equivalent in the Silurian of the Pentamerus-Eoeoelia community and Catazyga the equivalent of the Clorinda community (Ziegler 1965). The mode of life of many zygospirids is hypothetical. Zygospira has a small, but distinctive, interarea and anacline beak in maturity, with young specimens having a proportionally larger interarea and apsacline-orthocline beak. This suggests a pedicle-attached mode of life, particularly in early growth stages. However, some gerontic specimens have hypercline beaks, which would have made a functioning pedicle difficult in later growth stages. Richards (1972) has shown the clustering of zygospirids around bryozoan finger-colonies and this may have been a common association. In mass occurrences of zygospirids, bryozoans are too rare to have served as more than an occasional anchoring site: the zygospirids probably attached to each other, young to living parent shells or to vacant shells on the sea bottom. Richards noted that Zygospira often appears to have settled on live substrates (other benthos), and may have been a pioneering species in substrate colonization. This is possible; in one outcrop on Manitoulin Island heavy concentrations of Zygospira preceded a biostrome. Life assemblages of Zygospira show that the shell usually attached itself with its lateral commissure at an angle to the host or substrate, and with the brachial valve on the downward or host-ward side. This would mean that the spiralia would be orientated in the same direction (downwards or host-wards and inwards). Food currents created by the lophophore would have been taken in through the sides of the shell and released anteriorly. Catazyga , with a hypercline beak in maturity, biconvex-ventri-biconvex shell and weak fold-sulcus, is rather different in its usual lack of associates of any kind other than some strophomenids. Cluster accumulations, presumably nests, show that in dense clusters most shells are orientated at a relatively high angle to the substrate with the pedicle valve uppermost and the plane of symmetry tilted away from the vertical. Thus the anterior commissure, and presumably the exhalent current were further away from the substrate than with Zygospira. Shell elongation would accentuate this ( Zygospira is wider: Catazyga longer). However, in more moderate densities of 40-60 per 100 square cm, with non-touching shells on the substrate, the Catazyga generally lie flat on the substrate and with the pedicle valve uppermost. What this means is not clear. Possibly with decreased densities, waste clearance away from the substrate was not such a problem, or with shells no longer touching and therefore supporting each other upright, the pedicle was too weak to maintain a near vertical or angled position. Since beak incurvature was severe in maturity, suggesting less-functional pedicles, the latter explanation may be more correct. A reconstruction (text-fig. 2) demonstrates the postulated modes of life of the two dominant late Ordovician taxa. The catazyginids and zygospirinids in the Silurian seem to have maintained essentially the same ecological polarity. Pentlandella COPPER: ZYGOSP1RA AND SOME RELATED BRACHIOPODS 301 normally occurs in shales or mudstones with few other organisms and rarely with coralline skeletons. Zvgatrypa (known only from the Silurian of Anticosti Island) is an inhabitant of the Stricklandia community, and occurs with Atrypopsis , common Gotatrypa and clintonellinids. SYSTEMATIC DESCRIPTIONS Order atrypida Rzhonsnitskaya, 1960 Family zygospiridae Waagen, 1883 (emend.) [ Anazygidae Davidson, 1883] Primitive, smaller, planoconvex-biconvex, non-lamellose, usually tubular-ribbed atrypoids with minute pedicle openings, beaks orthocline to hypercline, no deltidial plates, or small plates. Internally the family is characterized by a fused jugum, located dorsal to the spiralia and moving in position anteriorly to posteriorly during evolution in the Ordovician. Age: Ordovician (Caradoc) to Silurian (Ludlow). Subfamily zygospirinae Waagen, 1883 Here emended to include Protozyga Hall and Clarke, 1893, Anazyga Davidson, 1882, Zygospira Hall, 1862, Hallina Winchell and Schuchert, 1892, and Zygatrypa gen. nov. The internal structure of Protozyga and Hallina was described, with some doubts, by Hall and Clarke (1893, pp. 149-151). Protozyga was figured with an anteriorly located jugum and short, one coil spiralia (ibid., p. 149) which seems correct in view of the internal morphology of Anazyga. Hallina was figured with only an anterior jugum and no spiralia (ibid., p. 151). Cooper (1956, pp. 689-690) did not define the latter genus nor find any species with spiralia. The type species of Hallina , from the Lebanon Formation of Tennessee, indicates a vertical distribution which in part coincides with that of Anazyga. Hallina may be a junior synonym of Anazyga. Externally they are not possible to differentiate; internally they both have well- defined cavities. Internals of Hallina need revision, but topotypic material is in- variably badly preserved (G. A. Cooper, pers. comm.). The common features of the Zygospirinae seem to lie in the possession of finely ribbed planoconvex- ventribiconvex shells with a modest to sharp ventral fold-dorsal sulcus, orthocline- anacline beaks, small but visible pedicle openings, dental cavities or nuclei, thin shell walls, an anterior to posterior jugum, and medially to dorso-medially directed spiralia. Range: Caradoc to Llandovery. Genus zygospira Hall, 1862 Type species. Atrypa modest a (Say in Hall 1847). Say apparently identified this in collections of the Academy of Natural Sciences, Philadelphia as Producta modesta (ibid., p. 141). Range. Late Caradoc (Maysvillian) to Ashgill (Richmondian). Distribution. North America, western Europe, ?Siberian Platform, Kazakhstan. Not yet reported from South America, but should occur in Australia, in view of great similarities in associated benthic corals with North America. E 302 PALAEONTOLOGY, VOLUME 20 Diagnosis. Moderately sized, relatively coarsely ribbed, carinate zygospirids with ventri-carinate, planoconvex shells, small interarea, orthocline-anacline beaks. Internally, dorso-medially directed spiralia with five to eight whorls and dorsal jugum located at or posterior to spiralial apices; dental cavities, small deltidial plates present. For spiralia refer to text-fig. 3. text-fig. 3. Reconstruction of the spiralia in the type species Zygospira modesta based on serial sections shown in text-hg. 4. Note the nature of the jugum. Scale approx, x 7. Species assigned. Zygospira cincinnatiensis James, in Meek 1873 (p. 126, pi. 11, fig. 5a-c), basal Maysville Formation (Foerste 1910), Cincinnati, Ohio, U.S.A. Zygospira richmondensis Caley, 1936 (p. 78, pi. 1, figs. 4, 6). Kagawong Formation, High Falls, Mani- toulin Island, Ontario, Canada (PI. 37, figs. 9-10). Zygospira modesta kagawongensis Caley, 1936 (p. 58), Unit 8. Meaford Formation, Section no. 16, Kagawong Falls, Manitoulin Island, Ontario, Canada. Species becomes a nomen nudum , unless it is described and figured. Zygospira putilla Hall and Clarke, 1893 (p. 157, fig. 150; pi. 54, figs. 35-37). Hudson River Group, Edgewood, Missouri, U.S.A. Amsden (1974) relegates this species to Eospirigerina. Zygospira parva Rukavishnikova, 1956 (pp. 162 163, pi. 5, figs. 14-16). Otarsk and Dulankarin Horizons, southern Kazakhstan, U.S.S.R. Very weakly carinate. Zygospira kentuckiensis Nettleroth, 1889 (pp. 138-139, pi. 34. figs. 21-25). ‘Hudson River or Cincinnati Group’, Taylor’s Station, Kentucky, U.S.A. One of the largest known Zygospira. Zygospira meafordensis Foerste, 1924 (p. 128, pi. 15, fig. 3 a-c). ‘Queenston Member, Richmond.' North-west Meaford, Ontario. Zygospira raymondi Foerste, 1 924 (p. 128), figured in Raymond (1921, p. 28, pi. 8, figs. 1-5). Collingwood Shale, Craigleath, Ontario, Canada. Zygospira meldalensis Reed, 1932 (p. 144, pi. 22, figs. 12, 12a). Kalstad Limestone, Meldalen, Norway. Zygospira resupinata multicostata Howe, 1965 (pp. 653-655, pi. 81, figs. 1-8). Aleman Limestone, Trans- Pecos, Texas, U.S.A. Zygospira sulcata Howe, 1965 (pp. 655-656, pi. 81, figs. 9-12). Uppermost Uphanr Limestone, Lone Mountain, New Mexico, U.S.A. Finely ribbed form. Zygospira concentrica Ulrich, 1888 (pp. 14-15, pi. 7, figs. 10, 10a, b). Lower part of the Hudson River Group, 300-350 ft above low water, Cincinnati, Ohio, U.S.A. Eocoelia hemispherica crassa Rozman, 1970 (pp. 123-125, pi. 15, figs. 5-8). Upper Ordovician, lower beds of Taskansk Suite, Sette-Daban range. North-east U.S.S.R. Zygospira resupinata Wang, 1949 (pp. 18-19, pi. 10a, figs. 1-12). Cornulites Zone, Brainard Member, Maquoketa Formation (Ashgill), Fairfield, Iowa, U.S.A. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 303 Comparisons. Zygospira is distinguished from Anazyga , its forerunner, by generally more coarsely ribbed, more strongly carinate shell, differentiated mid and lateral ribs and more planoconvex-ventribiconvex shells. Internally there are more whorls in the adult spiralia and the jugum is located postero-dorsally. Hall and Clarke (1893, p. 154) note an absence of ‘dental lamellae’, but this is clearly incorrect; dental cavities are prominent in both genera. Zygospira modesta (Say in Hall, 1847) Plate 37, figs. 1-8; text-figs. 3-4 1847 Atrypa modesta Say in Hall, pp. 141 142, pi. 33, fig. 15. Type. Lectotype AMNH 1356A (selected by Foerste 1910), Plate 37, figs. 1- 4, American Museum Nat. Hist., New York. In the original description Hall noted (1847, p. 142) that the species was rare in New York, where late Ordovician sediments are more sparsely fossiliferous, but abundant in Ohio, Indiana, and Kentucky. He selected no type locality except to state ‘It is quite abundant at numerous western localities, particularly Oxford and Cincinnati (Ohio), Madison (Indiana), Frankfort and Maysville (Kentucky)’. When he established the genus in 1862, and also in commenting on the species modesta in 1857, Hall cited no source localities. The type specimen is labelled ‘Cincinnati’. Foerste (1910, pp. 29-30) examined the type specimen, and stated that although it was labelled ‘Cincinnati’, this could have meant derivation of the specimen from some Cincinnati collection. The type specimen was labelled ‘Hudson River Group’. Foerste (1910) observed that specimens similar to the type came from the Fairmount Bed of the Maysville Formation (late Caradoc) and selected that as the type horizon. Since Z. kentuckiensis (Meek) occurs in the Upper Fairmount (according to Foerste), modesta may belong to the lower, unless the two taxa were synchronous and allopatric. There is no new data on this, particularly since Z. modesta has been very broadly interpreted in the literature and seems to have been ubiquitous in the shallower platform carbonate sequence. Richards (1972) reported Z. modesta from the Richmondian Tanner’s Creek Formation, well above the typical horizon, but these specimens are not similar to the type material. I have not collected topotype or conspecific material from the Cincinnati region, but judging from other text-fig. 4. Serial sections of Zygospira modesta (Say in Hall 1847) based on acetate peels. GS 45391 ‘Cincinnati, Ohio’ (material provided by Dr. G. A. Cooper, U.S. National Museum). Scale x5. 304 PALAEONTOLOGY, VOLUME 20 Zygospira in Ontario, abundant in the Georgian Bay Formation, the typical association of Z. modesta is with shallow-water benthos, especially digitate bryozoans. Sediments ranged from clayey shales in sub- or peri-biostromal occurrences to finely comminuted bioclastic substrates. Diagnosis. Moderately sized, wider than long, planoconvex-weakly biconvex zygo- spirids with 15-18 ribs over all, 4 strong mid-ribs on the pedicle valve, and 3 corre- sponding mid-ribs on the brachial valve and the remaining lateral ribs decreasing in size posteriorly. Average width 8-1 mm, length 7-2 mm, depth 3-9 mm (based on 15 syntypes in the AMNH Hall Collection). Internally, shell thin; weak, irregular pedicle collar structures line pedicle cavity; dental cavities distinct; teeth near- ’horizontal’ in sockets (text-fig. 4). Small nodular crural bases on inner socket ridges, crura diverge widely to sides, jugum arising at or near position of cone axes, initially curving postero-medially dorsal to the spiralia, then straightening medially and curving dorso-anteriorly to fuse. Spiralia D -shaped with straight sides forming a weak V (open anteriorly); three to five whorls seen (text-fig. 3). No cardinal process observed; dorsal septum strong. Muscle scars weakly impressed, poorly known. Remarks. It is premature, without mass collecting, to differentiate clearly between the described species of Zygospira , and to delineate an evolutionary succession, but the species group could become a useful tool in zoning late Carodoc-early Ashgill shallow marine sediments. The coarser-ribbed, larger specimens appear to be youngest, but this is not an invariable rule. Some species, such as Z. kentuckiensis, show confusing and greater infra-specific variability than do most atrypoids in younger rocks. Most of the described species are from the marine platform carbonate succession of Ontario, Michigan, Ohio, Indiana, and Kentucky. On the Appalachian side and in Quebec, Zygospira is rare, and where reported is usually not Zygospira but Anazyga or Catazyga. Genus anazyga Davidson, 1882 Type species. Atrypa recurvirostra Hall, 1847, p. 140. Range. Caradoc to ?Ashgill. Most if not all ‘Trenton’ species belong here. This genus is the earliest known wholly ribbed atrypoid. The oldest species appears to be Anazyga matutina (Cooper 1956) from the Little Oak Formation, Alabama, whereas the youngest may overlap with Zygospira in the late Caradoc to early Ashgill. Distribution. Most of the species have been identified from eastern North America; also Britain, Scandi- navia, and Estonia. Diagnosis. Small, about 5 mm wide, ventribiconvex-biconvex, equidimensional to elongate zygospirid shells, usually with strongly incurved, anacline-hypercline beaks, fine ribbing, and weak fold-sulcus. Carination lacking. Internally, small dental cavities, minute deltidial plates, undifferentiated teeth. Crura not strongly diverging, jugum branching off anterior to spiralial apices, frequently at the distal side of the crura, then fused postero-dorsally in front of spiralial apices; one to three whorls (text-fig. 5). Comparisons. Anazyga is a widespread genus present in ’Chazyan’ rocks (Cooper 1956) of Caradoc age in North America (the oldest is in ’Porterfield’ equivalents). The most comparable genus is Hallina Winchell and Schuchert 1892 whose vertical COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 305 range is almost the same. It is possible that Hallina is a junior synonym: its type species must be recollected and internal structures clarified. Winchell and Schuchert (1892, 1893) described and figured the brachidia as consisting of a single fused loop, which looks very much like the jugum, but no spiral whorls were noted. If this is typical of all Hallina, then possibly the genus is distinct from Anazyga. Hall and Clarke (1893, p. 151, figs. 139-141) appear to confirm the lack of a spiralium and presence of a loop, and believed that the internal structure of their genus Protozyga (ibid., p. 151) was nearly identical, except that Protozyga had the beginnings of the first spiral whorl (ibid., p. 149). If Hallina is distinctive in its primitive non-coiled brachidia, then externally it is a near homoeomorph of Anazyga, according to Boucot et al. (1965). This would create severe taxonomic difficulties in classifying poorly preserved material and would mean that some taxa may be erroneously attributed. Cooper (1956, pi. 143) illustrated H. sqffordi, H. lirata , and H. globularis, which indicate that externally Hallina may be distinct in having the posterior shell umbones smooth, as already pointed out by Winchell and Schuchert (1893, p. 473). It is possible that the Clintonellinae (Poulsen 1943) may have arisen independently from the Hallina group. Hallina cannot be distinguished from Anazyga by its ‘dental plates’, which are present in both genera. Davidson (1883, p. 136) established the family Anazygidae, including Anazyga, Dayia, and Hindella, to include brachiopods with ‘loops arising from bottom of spirals’. This was published in the same year as Waagen (1883), who erected the Zygospiridae : thus seniority is not clear. The Anazygidae are left as a junior synonym of the Zygospiridae. Davidson’s view was based on a misinterpretation of the coiling direction of the spiralia; the last two named genera have athyridoid spiralia directed at mirror-image position compared to the atrypoids. Species tentatively assigned. Atrypa deflecta Hall, 1847 (p. 140, pi. 33, fig. 4a, b). ‘Central part of the Trenton Limestone near Martins- burgh.’ Zygospira recurvirostris noquettensis Hussey, 1926 (pp. 162-163, pi. 11, figs. 1-3). Stonington Beds, Ogontz member, Stratton’s Farm, northern Michigan, U.S.A. Possibly an unusual Zygospira. Zygospira recurvirostris turgida Foerste, 1917 (p. 103, pi. 5, fig. 15«-c). ‘Upper part of the argillaceous Richmond’, Little Bay de Noquette, northern Michigan. This may not be true Anazyga. Zygospira orbis Reed, 1917 (pp. 944-945, pi. 24, figs. 24-27). Stinchar Limestone Group, Craighead, Girvan, Scotland (Lower-Middle Caradoc). Possibly a rhynchonellid. Zygospira variabilis Fenton and Fenton, 1924 (pp. 75-76, pi. 2, figs. 7-9). No single locality cited in literature: Plattin-South Becket Hill, Kentucky; Black River-Frankfort and Paris, Kentucky, U.S.A. Zygospira variabilis fountainensis Fenton and Fenton, 1924 (p. 76, pi. 2, figs. 1-3). Decorah Shale, Fountain, Minnesota, U.S.A. Zygospira calhounensis Fenton and Fenton, 1924 (pp. 16-11 , pi. 2, figs. 4-6). No locality and horizon in literature. Zygospira tantilla Bradley, 1921 (p. 525, no figs, but see PI. 38, fig. 4). Lower Maquoketa, Clermont, Iowa, U.S.A. This is one of the last Anazyga, unless a diminutive, unusual Zygospira (the type specimens from Harvard University show an orthocline beak, PI. 37, figs. 16-20). Zygospira gutta Oraspyld, 1956 (pp. 64-65, pi. 4, figs. 14-15). Vazalemmask Horizon (DIU), Saku, Estonia, U.S.S.R. Zygospira circularis Cooper, 1956 (p. 670, pi. 141C, figs. 18-21 ; pi. 142B, figs. 6-10; pi. 142D, fig. 16). Upper Carters Formation, Franklin, Tennessee, U.S.A. Zygospira elongata Cooper, 1956 (pp. 670-671, pi. 268G, figs. 29-32). Lebanon Formation, Readyville, Tennessee, U.S.A. 306 PALAEONTOLOGY, VOLUME 20 Zygospira lebanonensis Cooper, 1956 (pp. 671-672, pi. 142C, figs. 1 1-15). Lebanon Formation, Shelby- ville, Tennessee, U.S.A. Zygospira matutina Cooper, 1956 (p. 672, pi. 1 4 IB, figs. 13-17). Little Oak Formation, Alabama, U.S.A. Zygospira mediocostellata Cooper, 1956 (pp. 672-673, pi. 143D, figs. 13-18). Sevier Formation, Bulls Gap, Tennessee, U.S.A. Zygospira recurvirostris aequivalvis Twenhofel, 1928 (p. 214, pi. 19, figs. 10-12). English Head Forma- tion, Zone 4, Anticosti Island, Quebec, Canada. Needs confirmation; possibly related to noquettensis Hussey, 1926 and tantilla Bradley, 1921. Anazyga recurvirostra (Hall, 1847) Plate 37, figs. 11 15; text-figs. 5-6 1847 Atrypa recurvirostra Hall, p. 140, pi. 33, fig. 5 a~d. Holotype. AMNH 705.3 (by monotypy), Plate 37, figs. 11-15. American Museum Nat. Hist., New York, from 'near Martinsburgh’, New York (Hall 1847, p. 140). The species is abundant in thin-bedded, dark- grey argillaceous limestones (micrites), at outcrops just south of the bridge over Roaring Brook, 1 -3 miles east of Martinsburg. This seems suitable as a restricted type locality. The type horizon is a 'Compact greyish blue bed of limestone near the centre of the Trenton Limestone’ (ibid.: note that this is almost exactly the middle of the Trenton Group on geological maps, e.g. Miller 1910). This horizon is suspected to be the Shore- ham Member of the Sherman Falls Formation, or in other terminology, the Sugar River Limestone (Kay 1938), the zone of Cryptolithus tesselatus in New York. The Ontario lithic equivalent is the basal Verulam Formation. Clarke (1919), who described the Martins- burg section, listed ' Zygospira recurvirostris ’ at the text-fig. 5. Reconstruction of the spiralia and 100, 165-180, 270-280, and 390 It levels in the Trenton jugum in Anazyga recurvirostra , based on serial Limestone. It is not known where the type locality of sections shown in text-fig. 6. Scale approx, x 7. recurvirostra mentioned above fits into this sequence. Possibly it is the member 6 of Miller (1910, p. 29) in the Trenton Limestone, said to be 475 ft thick at Roaring Brook, or the 165-180 ft level of Clarke (1919, p. 7). The associated fauna is common hemispherical Prasopora colonies, some 2-4 cm in diameter. Praso- pora tends to be more abundant in the lower middle part of the old Trenton and stops at 280 ft above the base according to Clarke (1919). Other fossils are rare. Description. Small, 5-7 mm wide, longer than wide, strongly biconvex- ventribiconvex shells with hypercline beaks. Ribs fine, evenly sized throughout, though slightly larger flanking by sulcus, averaging about twenty-four in number. Brachial valve somewhat flattened anteriorly and very faintly sulcate, pedicle valve more arched; weak anterior fold. Internally, pedicle valve thickened apically with faint median septum, no pedicle constrictions, deltidial plates minute, solid, pointed dorsally, dental cavities elongate, teeth simple, dorso-median projections supported termi- nally by dorsal wall (text-fig. 6). Brachial valve with sturdy, short socket plates separated medially by pit without cardinal process. Crura extended from small round bases located on inner side of socket plates, diverging laterally at about a 45° angle to the mid-shell, then curving inwardly, with spiral whorls begin- ning at anterior-most portion. Jugum in broad V-shaped band joined anterior to spiralial apex; two to three spiral whorls with axes directed medially (text- fig. 5). COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 307 Comparisons. It is difficult to compare A. recurvirostra adequately with the many other species described later (see list). With its rotund, longer-than-wide shell it seems distinct from other species, but the latter still need revision to determine variation. At the type locality recurvirostra is quite consistent in shape and small size, but Hall described the species deflecta from what may be the same locality. The type specimens of deflecta are lost. Many authors have confused Z. modest a with A. recurvirostra— the former occurs in much higher strata. In general, Protozyga characterizes Blackriveran or older Caradoc beds, Anazyga the Trentonian, and Zygospira the post-Trentonian. Bretsky (1970) suggested that recurvirostra had a text-fig. 6. Serial sections of Anazyga recurvirostra (Hall, 1847), based on acetate peels. GS 45383, Shore- ham Member, Sherman Falls Formation (Trenton Lst.), Roaring Brook, 1-3 miles east of Martinsburgh. Inset are umbonal views. Scale x 5. functional pedicle, but this seems unlikely in adult stages, which have hypercline beaks developed. He also suggested that the species, which he interpreted more broadly than here, was most common in Virginia and north-west Tennessee. How- ever, it is locally abundant in New York and Ontario (Manitoulin Island area). In Ontario similar species occur in the Bobcaygeon, Verulam, and Lindsay Forma- tions. A statistical comparison of rich material may reveal useful horizon indicators or possibly other geographical or ecological species or sub-species. Genus zygatrypa gen. nov. Type species. Zygospira paupera Billings, 1866, p. 46 (first illustrations, Twenhofel 1928, pi. 21, figs. 21-22). Range. Llandovery (especially late Llandovery). It must occur in earlier Llandovery strata if derived from Zygospira. Upper limit possibly Wenlock. Distribution. Eastern North America. Diagnosis. Small, carinate, ventribiconvex-planoconvex zygospirid shells. Fine, usually consistently sized ribs, postero-lateral shell flanks smooth. Minute pedicle 308 PALAEONTOLOGY, VOLUME 20 opening, anacline-hypercline beaks, small interarea, and deltidial plates. Commis- sure sulcate. Internally, valves thick-shelled near hinge and under muscle scars, pedicle constrictures, minute posterior dental cavity or dental nucleus expanded anteriorly to strong teeth. Sockets in brachial valve firm, crural bases elongate, ventro-vertically directed, then sharply laterally turned. Spiralia three to four whorls, dorso-medially directed. Jugum simple, U-shaped, arising posteriorly and joining posterior to spiral apices (text-fig. 7). Species assigned. Rhynchonella mica Billings, 1866 (p. 445, no figs.). Jupiter Formation, Zone 9 (Upper Llandovery), Anticosti Island, Quebec, Canada (Twenhofel 1928, pi. 21, figs. 21-22). Plate 37, fig. 30. 1A try pa plicatula Hall, 1843 (p. 71, fig. 4, see also Hall 1852, p. 74, pi. 23, fig. 9 a-h). ‘Calcareous Shale at Reynale’s basin. New York', U.S.A. (Hall ibid.). Requires revision. IZygospira minima Hall, 1879 (p. 14, see Hall 1882, p. 305, pi. 27, fig. 7). Waldron Shale (Wenlock), Waldron, Indiana, U.S.A. If the assignment is correct, this would be the youngest Zygatrypa. Comparisons. Zygatrypa is distinguished externally from Zygospira, its closest relative, by its finer, even ribbing and its absence of ribs near the hinge line. Internally it is distinct by its more massive teeth, smaller dental cavities or nuclei, thickening sub-muscle field shell wall, crura emergent from the tips of socket plates (as compared with medially in Zygospira ), but especially by the nature of the spiralia and jugum. The jugum of Zygatrypa is U-shaped, whereas in the older Zygospira , it is W-shaped (compare text-figs. 3 and 5). In Zygatrypa the jugum appears near the hinge line and in Zygospira it appears at mid-shell (Beecher and Schuchert 1893). The Asiatic, later Silurian genus Tuvaella is much larger in size, has a long, straight hinge, prominent cardinal process, and lacks dental cavities. The new genus Zygatrypa spans part of the stratigraphical gap between youngest Zygospira and oldest Tuvaella EXPLANATION OF PLATE 37 All figures x2, except fig. 21, x4. Figs. 1-8. Zygospira modesta (Say in Hall 1847). 1-4, AMNH 1356a lectotype, ventral, dorsal, lateral, and posterior views. 5-8, AMNH 1356c paralectotype, ventral, dorsal, lateral, posterior views. ‘Hudson River Group, Cincinnati, Ohio’, U.S.A. (data on labels of fifteen syntypes). Ashgill. Figs. 9 10. Zygospira richmondensis C aley, 1936. ROM 12448, holotype, ventral and dorsal views. ‘Kaga- wong Fm. Richmond[ian], High Falls, Manitoulin I.’, Canada. Ashgill. Figs. 11-15. Anazyga recurvirostra (Hall, 1847). AMNH 705.3 holotype, ventral, dorsal, lateral, posterior, and anterior views. ‘Trenton Lst., Martinsburgh, New York’, U.S.A. Caradoc. Figs. 16-20. ? Anazyga tantilla (Bradley, 1921). MCZ8547a lectotype (one of twelve syntypes), ventral, dorsal, lateral, posterior, and anterior views. ‘Lower Maquoketa, Clermont, Iowa', U.S.A. Ashgill. Figs. 21-29. Zygatrypa paupera (Billings, 1866). 21-25, GS 2454a lectotype, ventral (enlarged x4), dorsal, lateral, anterior, and posterior views. Note carination and rib absence postero-laterally. 26-29, GB 2454b paralectotype, ventral, dorsal, lateral, and posterior views. Jupiter Formation (zones 6-7, Twenhofel 1928), Anticosti Island, Canada. Llandovery (top). Fig. 30. Zygatrypa mica (Billings, 1866). GS 2517a, lectotype (largest specimen) and three other syntypes on small slab. Jupiter Formation (Zone 9, topmost), Anticosti Island. Llandovery. Distinguished from older Z. paupera by double fold in brachial valve sulcus. Figs. 31-35. Catazyga headi headi (Billings, 1862). Lectotype from Whiteaves Collection, RM 801a, dorsal, ventral, posterior, lateral, and anterior views of the best preserved of four syntypes. ‘Trois Rivieres, Quebec’ (probably Pontgrave River Formation. Ashgill). PLATE 37 COPPER, Ordovician and Silurian Zygospiridae 310 PALAEONTOLOGY, VOLUME 20 and was possibly a transitional species group. In particular the jugum of Zygatrypa is more like that of Tuvaella. Boucot and Johnson (1967) suggested that Zygospirci paupera may have been ancestral to Coelospira (a dayiacean), but this was based on material from California, not Anticosti, and is clearly incorrect. Dayiacean spiralia are like those of the athyridids. Zygatrypa paupera (Billings, 1866) Plate 37, figs. 21-29; text-figs. 7-8 1866 Zygospira paupera Billings, p. 46 (Lectotype GS 2454a). 1928 Zygospira paupera Billings; Twenhofel, p. 214, pi. 21, figs. 18-20. Type locality. ‘Near Jupiter River’ (Billings 1866), Anticosti Island, Quebec, Canada. From this it seems that the Richardson Collection, on which Billings founded the species, was not actually on the river but close to it. 1 have been unable to find it in outcrops along the Jupiter, but it was common on the road east of the Jupiter, at firetower no. 5 (NTS 12E/1 1 W, 75180:87940). The type horizon is ‘Div. 3, A. G. J. Richard- son’ (ibid.). This is the Jupiter Formation. Twenhofel (1928, p. 214) remarked that he found the species in Zones 1, 6, 7, and 9 of the Jupiter Formation (Upper Llandovery), but it probably does not occur in the uppermost zones (8-9), where it is replaced by Zygatrypa mica (Billings 1866). I have not seen it in the basal zones of the Jupiter Formation. This leaves zones 6 and 7 as the restricted type strata. The lecto- type is one of two syntypes in the Geological Survey (Canada) collections, selected by Twenhofel. GB2454a is the larger and better preserved of these two. The genus occurs in what is probably the equivalent of the Stricklandia community, since this pentamerid is abundant in the same few metres of strata. Flowever, Stricklandia occurs here in nests, and where it is abundant, Zygatrypa is very rare. Instead, Zygatrypa occurs in more muddy, yellowish weathering horizons together with other atrypoids such as ‘ Gotatrypa' sp. (most abundant), ‘ Clintonella' anticostiana (next), and Atrypopsis julia (least common). Trilobites and ostracods are very common but stromatoporoids, corals, and bryozoans are absent. Description. Small, 6-10 mm wide (average 7-0 mm), wider than long, nearly plano- convex shells with anacline hypercline beaks, foramen commonly penetrating umbo. text-fig. 7. Reconstruction of the spiralia and jugum in Zygatrypa paupera (Billings, 1866) based on text- fig. 8. Small cardinal process not shown. Scale approx. x7. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 311 Apical angle 105°- 120°, interarea narrow, pedicle valve angular convex to almost carinate, brachial valve with deep sulcus dividing two convex halves. Ribs 20-24 in number, absent postero-laterally, especially on the brachial valve; rib pattern in general reveals thick, rarely bifurcating ventral mid-rib, flanked by 3-4 fine ribs, then 3-4 thick ribs, and 3 very faint ribs fading into smooth shell surface. Muscle scars not deeply incised, but raised above valve floor; triangular ventral adductor pad matched anteriorly by two obscure elongated diductor scars divided by a weak septum. The large brachial lateral adductors are shaped like footprints and flank two narrow, central adductors raised on the median septum (text-fig. 7). In serial sections, the pedicle valve shows small, round dental cavities forming the nucleus of simple, inwardly directed teeth. A small cardinal process caps the hinge plate. Diagonal crura fit into notches on the teeth and feather out to a solid U-shaped jugum and spiralium with about three whorls (text-fig. 8). text-fig. 8. Serial sections of Zygatrypa paupera (Billings, 1866) based on acetate peels. GS 45374, Zones 6-7, Jupiter Formation, Anticosti Island, Canada (Upper Llandovery); NTS Jupiter River 75140:87850. Inset are umbonal views. Scale x 5. Remarks. The species can be distinguished from younger Z. mica , also described by Billings (1866) from Anticosti Island, by the single, angular crest on the pedicle valve in paupera against a double crest in mica. Z. mica appears to be a very scarce element in the uppermost part of the Jupiter Formation on Anticosti Island. Bolton (1972, pi. 8, figs. 15-17, 20) figured a ‘Zygospird’ jupiterensis from the underlying Gun River Formation. This species can now be assigned to Atrypina , and is not a zygospirid. Zygatrypa has not been found in the earlier Llandovery strata on Anticosti Island (Gun River, Becscie Formations). Subfamily catazyginae subfam. nov. The Catazyga group of Ordovician-Silurian atrypoids include two genera with first appearances in late Ordovician time (late Caradoc), and a decline in post- Ordovician time to eventual extinction by the close of the Llandovery. Included in 312 PALAEONTOLOGY, VOLUME 20 the group are the type genus Catazyga and also Pentlandella. Excluded are Alispira Nikiforova, 1961, Clintonello Hall and Clarke, 1893, Nalivkinia Bublichenko, 1928, and Anabaria Lopushinskaya, 1965. The Catazyginae are defined as small to moder- ately sized atrypoids with uninterrupted tubular ribs, lack of carination, and weak anterior folds and generally biconvex-ventribiconvex shells with small interareas hidden by incurved swollen beaks. Internally they possess a posteriorly located single jugum, and have a modest set of spiral whorls directed dorso-medially. Crura tend to be solidly supported on thick socket plates. Dental cavities are normally absent, but a hidden or weak ventral septum is present; the catazyginids are unusual in their relatively massive deposits along the pedicle cavity and hinge plate. Nikiforova (in press) has discovered a large late Ordovician atrypoid in Central Asia (Shakhriomon area) which she believes may be a coarsely ribbed catazyginid. The internal structures are as yet undescribed but material available to me shows an atrypoid with distinctive wide interarea, large foramen, and weak carination that may be a large Zygospira or spirigerininid ; it is not one of the Catazyga group (see PI. 39, figs. 20-23). Genus catazyga Hall and Clarke, 1893 [Orthonomaea Hall, 1893] Type species. Athyris headi Billings, 1862, p. 147. Range. Late Caradoc to Ashgill. Distribution. North America, western Europe; Catazyga salairica jacutensis Rozman, 1968 from the north- east U.S.S.R. may not be Catazyga. C. salairica Severgina, 1960 needs to be investigated. C. homeospiroides Ross and Dutro, 1966 does not superficially resemble known catazygids. C. rara Nikiforova has become the type species of the genus Arabaria Lopushinskaya, 1965, and is not a catazyginid. Thus the genus is not yet confirmed outside western Europe and North America. Diagnosis. Small to moderately sized, elongate, finely ribbed, biconvex to ventri- biconvex, weakly folded zygospirids with anacline-hypercline beaks covering a minute pedicle opening and deltidial plates (normally not visible). Internally, the EXPLANATION OF PLATE 38 All figures x 2, except figs. 12-14, x4. Figs. 1-14. Catazyga headi headi (Billings, 1862). 1-5, GS 45385 hypotype, ventral, dorsal, anterior, posterior, and lateral views of large adult specimen with minor growth deformity along sulcus. 6-10, GS 45392 hypotype, ventral, dorsal, lateral, posterior, and anterior views of immature specimen. 11, GS 45401 hypotype, calcined, decorticated mature specimen to show ventral view of the inner spiralial whorls. 12, GS 45393, latex mould of brachial valve interior showing muscle field. 13, GS 45391, latex mould of pedicle valve interior. 14, GS 45384, latex mould of pedicle valve interior showing irregu- larly emplaced muscle field. All material Pontgrave River Formation, Becancour (restricted type loc.), Quebec, Canada. Ashgill. Figs. 15-19. Catazyga anticostiensis (Billings, 1862). GS 2038j, lectotype (one of fifteen syntypes) from Billings’s Collection. Probably from 'Hudson River Fm., English Head, Anticosti I.’, and mislabelled in present collection. This material is very similar to other Anticosti material from the Vaureal Formation. Figs. 20-21. Catazyga hicksi (Reed, 1905). 20, A30861 paralectotype, internal mould of brachial valve. 21, A30862 lectotype, internal mould of pedicle valve. Cuckoo Grove Lane, Haverfordwest, Wales. Slade and Redhill Formation (Ashgill). PLATE 38 COPPER, Ordovician and Silurian Zygospiridae 314 PALAEONTOLOGY, VOLUME 20 pedicle cavity and muscle field is deeply incised and contains thick irregular calcite deposits; small teeth in socket cavities have a massive centrally directed base, dental cavities absent or elongated horizontally. Ventral adductor muscle field broad, subrectangular, divided by weak median ridge; diductor area narrow, tapering posteriorly, poorly defined; dorsal adductors elongate, narrow. Dorsal valve with small cardinal process on thick hinge plates, strong socket plates support small crura. Posteriorly located jugum bent towards cone apices; spiralia three to ten whorls (text-fig. 9). text-fig. 9. Reconstruction of the spiralia and jugum in Catazyga anticostiensis (Billings, 1862) based on GS 45394, Vaureal Forma- tion; Anticosti Island, Canada, Loc. BF254 MacDonald Rd. Muscle-scar drawings of C. headi (Billings, 1862). Approx, x 7. Species assigned to Catazyga. Athyris headi borealis. Billings, 1862 (p. 147, fig. 126 a-b). 'Hudson River Formation’, Lake St. John, on the river Saguenay, Quebec, Canada. C. borealis is probably a valid species. Athyris headi anticostiensis Billings, 1862 (p. 147, fig. \21a-b). Vaureal Formation, English Head, Anticosti Island, Quebec, Canada, Plate 38, figs. 15-19. I consider C. anticostiensis a valid species (see PI. 38, figs. 15-19). Catazyga headi filistriata Sproule, 1936 (p. 108, pi. 7, figs. 4-7). Upper 30 ft of 'Cobourg strata’ (ibid., Lindsay Formation?), Georgian Bay, Ontario, Canada, Plate 39, figs. 1-2. Zygospira uphami Winchell and Schuchert, 1892 (p. 291, first figs. Winchell and Schuchert, 1895, pi. 34, figs. 45-48). ‘Middle of the Galena horizon at Weisebach’s Dam near Spring Valley' (ibid., p. 469), Minnesota, U.S.A. Ortliis erratica Hall, 1847 (pp. 288-289, pi. 79, fig. 5a-/). 'Central part of the Hudson River Group’ (ibid.), Washingtonville, New York, U.S.A. (Pulaski member of Lorraine, Foerste 1916, p. 37.) Catazyga uphami australis Foerste, 1909 (pp. 31-32, pi. 2, fig. 19 a-b\ pi. 3, fig. 14 a-c). High Bridge Formation, Camp Nelson member. High Bridge, Kentucky, U.S.A. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 315 Glassia schuchertana Ulrich, 1888 (pp. 186-187, for figs, see Foerste 1909, pp. 32-35, pi. 2, fig. 3; pi. 3, fig. llfl-c). Waynesville Bed (upper part), Hitz Road, Madison, Indiana, U.S.A. Catazyga cartieri Cooper and Kindle, 1936 (pp. 359-361, pi. 52, figs. 8-13, 18). Whitehead Formation, north-west Perce, Gaspe Peninsula, Quebec, Canada. Catazyga arcana Williams, 1962 (pp. 247-248, pi. 25, figs. 20-23, 27-28). Kiln Mudstones, Craighead, Girvan, Scotland. Atrypa headi var. anglica Davidson, 1867 (pi. 22, figs. 1-7). Caradoc. Grangegeeth, Ireland. Species requiring confirmation. Catazyga homeospiroides Ross and Dutro, 1966 (pp. 8-9, pi. 1, figs. 7, 9-10, 15, 17-20). Thin bedded, shelly limestone unit, Jones Ridge, Charley River, Alaska, U.S.A. Authors state possibility of a distinct genus. Orthisl sectostriata Ulrich, 1879 (p. 15, pi. 7, figs. 11,1 \a-b). Hudson River Group, 300-375 ft above low-water mark in the Ohio River, Cincinnati, Ohio, U.S.A. Catazyga salairicajacutensis Rozman, 1968 (p. 73, pi. 62, figs. 6-8). Nalchanskaya Suite (late Ordovician), Sakyndani River basin, north-east U.S.S.R. Catazyga salairica Severgina, 1960. I have been unable to see the original reference. Zygospira hicksi Reed, 1905 (p. 452, pi. 23, figs. 17-19). Cuckoo Grove Lane, Haverfordwest, Wales, Slade Beds, Ashgillian (A30861-30862, Sedgwick Museum). Plate 38, figs. 20-21. Comparisons. Catazyga is distinguished from Silurian Pentkmdella externally by its elongate shape, narrower apical angle, and usually larger shells and internally by its muscle field in both valves (raised, arrow-shaped ventral platform in Pentlandelta ), and bulbous socket plates. Some Catazyga , but not all, have dental cavities (e.g. the type species has cavities but C. borealis (Billings, 1862) does not. The broad-crested, narrow-troughed ribs of Catazyga may be a distinct generic feature. For internal distinctions compare text-figs. 9 and 12, 10-1 1, and 13. Catazyga headi (Billings, 1862) Plate 37, figs. 31-35; Plate 38, figs. 1-14; text-figs. 10-1 1 1862 Athyris headi Billings, p. 147, fig. 125. Types. Holotype or syntypes lost, at least since the time of Foerste (1909), who mentioned that ‘specimens collected by Whiteaves from the type locality are at hand, and may be regarded as replacing the types'. Whiteaves donated four specimens to the Redpath Museum (McGill University, Montreal RM801), which are labelled Trois Rivieres, the type locality (PI. 37, figs. 31-35). None of the four is identical to Billings’s figure, which appears to be somewhat larger than life size (width = 19 mm). None of the four is ideal as a neotype, since the exact location of the material is unknown, but Whiteaves probably knew the type locality first hand and therefore RM801a is here selected as neotype. In collections of the Geo- logical Survey, Ottawa, there are fifteen specimens labelled Trois Rivieres, collected by J. Richardson in 1856. These specimens are substantially smaller than C. headi headi from Trois Rivieres and Becancour and similar in shape and size to C. anticostiensis of the Vaureal Formation on Anticosti Island. It is con- cluded that this suite is mislabelled and represents the lost types of anticostiensis described by Billings. The lectotype of the Anticosti species is here selected as GS 2038j, the best preserved of the suite (PI. 38, figs. 15-19). The type locality is ‘On the south shore of the St. Lawrence opposite Three Rivers’ (Billings 1862). The exact location of the Billings locality is not possible to trace. Directly opposite Three Rivers at the present time there are no outcrops, only glacial erratics. At the Nicolet River section, some 14 km south of Trois Rivieres, the ‘gully section’ of Foerste (1916, p. 18) probably contains the stratigraphically equivalent horizons (Pontgrave River Formation) with C. headi headi, but at present the gully section is not exposed. C. headi headi occurs abundantly at Becancour, about 8 km north of Trois Rivieres and a suitable restricted type locality may be NTS Becancour 31I8/W 01450:38880. The type horizon is ‘Hudson River formation’ (Billings 1 862). The Nicolet River section equivalent is probably zones S and T of Foerste 316 PALAEONTOLOGY, VOLUME 20 (1916, pp, 18-19) which are in the Pontgrave River Formation (‘Richmond’). The Becancour material is in the same horizon. Dr. Yvon Globensky, who kindly provided me with a large collection of specimens from Becancour remarks that the Catazyga occur in limestone beds, interstratified with grey, sandy shales. At this locality C. headi occurs near the base of the Pontgrave above the Carmel River member. Associated with the atrypoids are solitary rugose corals, Strophomena and Sowerbyella. The Catazyga substrate during life was probably a soft calcareous mud, with specimens orientated beak-down, more due to thicker calcite text-figs. 10-11. Serial sections of two specimens of Catazyga headi (Billings, 1862) from Becancour, Quebec, Canada (near Trois Rivieres), NTS Becancour 01450:38880. Top GS 45398; bottom GS 45397. Note the variation in pedicle deposits, typical of many later atrypoids. External views of the umbonal region are inset. Scale x 5. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 317 deposition in the umbonal region, than to anchoring. Sectioned specimens frequently show muddy infilling posteriorly and vugs or drusy calcite anteriorly. Diagnosis. Medium sized to large, globose Catazyga with maximum width at mid- length, hypercline beaks, foramen and area obscured at maturity. Pedicle valve more convex than brachial valve, often with weak narrow anterior brachial valve sulcus and broad pedicle valve fold (over all weakly uniplicate). Internally distinctive muscle fields, thick pedicle callosities, small outwardly extended deltidial plates, dental cavities horizontally elongated, teeth massive, lacking accessory lobes, cardinal process irregular, bulky; crura thick. Spiralia-jugum unstudied. Spiralium reconstruction based on C. anticostiensis (text-fig. 9). Description (based on Becancour locality). Shells average width, peaks at 14 mm, maximum 19 mm, depth peaking at 9 mm, mostly longer than wide (text-fig. 6), apical angles 105-110° (average 108°). Pedicle area covered by beak in very early growth stages, before shell is 10 mm wide; sometimes pedicle opening expanded as slit in umbo. Ribs very fine posteriorly (at 5 mm, 25-30 ribs per 5-mm arc) but at 20 mm from umbo coarsening to 8-11 ribs per 5-mm arc; ribs round-broad crested and narrow-troughed. Brachial valve less convex but well-rounded; both valves usually sulcate with the pedicle valve having broad, flat sulcus. The adductor muscle fields are quite variable. The ventral adductor pad occupies about a quarter of the shell length, is wider than long, rectangular to rounded, and frequently skew or irregular; medially it is divided by a broad, low septum and laterally each side has three to four grooved lobes (usually three). The ventral diductors appear to be raised on two rounded irregular lobes, fusing posteriorly and sometimes raised off the shell floor. These structures may represent pedicle muscle callosity, and if that is the case the diductors possibly are located between the pedicle callosity and the adductor field. The dorsal muscle field is difficult to interpret. Adductors are in the form of two pairs, a posterior bean-shaped pair and an antero-medial pear-shaped pair divided by a rounded median septum. Closer to the hinge plate are one to three pairs of small depressions of unknown origin, possibly accessory diductors (text-fig. 9). The serial sections illustrated (text-figs. 10-11) are largely self-explanatory. Most striking are the thick pedicle cavity linings (squared in outline) and massive hinge plates generally lacking in zygospirinids. The axis of the dental cavity is dorso- anteriorly ventral posteriorly and horizontal anteriorly; teeth are simple stumps, free anteriorly. The pedicle opening is either hidden or sometimes expanded into the ventral umbo as a minute, narrow slit; deltidial plates are pointed dorsally rather than medially. A cardinal process is present in the form of thick irregular outgrowths capping the ends of the socket plate, expanding well into the pedicle cavity; the groove between the socket plates is narrow and slit-like. The crura arise from a point source hidden deep within the hinge plate, and are rounded in cross- section; anteriorly they stand free as thin, raised ridges and then direct themselves ventrally and sharply laterally. Further brachidial structures still unknown (see reconstruction of complete C. anticostiensis in text-fig. 9). Remarks. C. headi, like all Catazyga , had a small functioning pedicle or loss of pedicle in maturity. Epifauna was very scarce on the shells examined, and where found, consisted of small serpulids ( Cornulitesl] ) and Hederella normally located F 318 PALAEONTOLOGY, VOLUME 20 on the anterior central part of the pedicle valve. The probable mode of life of Catazyga was with the umbo down and shell vertical or angled with the pedicle valve upper- most, considering the evidence of epifauna, thick shell callosities posteriorly, and the usual posterior mud infilling of the shell after death. Slabs of C. anticostiensis which contain large numbers of presumably death-orientated shells show a prevailing umbo-down position; clusters of the same Catazyga show orientation towards a common point of fixation, and possibly indicate a thin, thread-like pedicle. The species of Catazyga at present are no more useful as stratigraphic indicators than mentioned by Foerste in 1924 (pp. 129-130). According to Foerste (1916) C. erratica precedes C. headi in the Quebec-Ontario sequences. In Ontario, C.fili- striata in turn appears to precede C. erratica and possibly the oldest Catazyga is C. uphami from Minnesota. If this sequence is correct, then no distinctive external trends are present in the Catazyga lineage, except possibly increasing size. Shape, convexity, and structure of the anterior commissure were random developments at different times and in different places. Lineages of Catazyga , which was a deeper quiet-water inhabitant, are rarely continuous in contiguous sections, and thus reflect water-depth variability and fluctuations in different regions. Such changes will be calculated when more data is available on over-all distribution. C. headi is not sufficiently distinct externally from C. erratica to warrant the separate genus Ortho- nomaea Hall, 1893 for the latter. The species of Catazyga also are not sufficiently well known to split off a new genus on the presence or absence of dental cavities (or ‘dental plates’), as with other Palaeozoic atrypoids. EXPLANATION OF PLATE 39 All figures x 2, except fig. 7, x 4. Figs. 1-2. Catazyga filistriata Sproule, 1936. ROM39a lectotype (one of six syntypes), ventral and dorsal views. ‘Upper 30 ft. of Cobourg strata’ (below Collingwood Black Shale). Locality unknown, possibly old quarry at Bowmanville (Sproule 1936, p. 99), Canada. Caradoc. Figs. 3-7. Pentlandella tenuistriata Rubel, 1970. 3-6, GS 45395 hypotype, ventral, dorsal, lateral, and posterior views of adult specimen. 7, GS 45394, interior view of pedicle valve enlarged ( x 4) to show the arrow-shaped muscle field. Latikula, Estonia, U.S.S.R.; Adavere Formation. Upper Llandovery. Figs. 8-12. Pentlandella pentlandica (Haswell, 1865). 8, BMNFI 12630a hypotype, latex mould of external of pedicle valve. 9, BMNH 12730b, hypotype, latex mould, internal view of pedicle valve. 10, BMNH 12730c hypotype, latex mould, internal view of brachial valve. Pentland Hills, Scotland, Bed D, Esk section. Upper Llandovery. 1 1, BMNH 12632a, latex mould, internal view of pedicle valve. 12, BMNH 12632b, latex mould, internal view of brachial valve. Same horizon and locality. Figs. 13-15. Pentlandella haswelli (Reed, 1908). 13, A32775b paralectotype, latex mould of internal of brachial valve. 14, A32271 lectotype, latex mould of internal of pedicle valve showing muscle field in hatchet-shape more comparable to Catazyga. 15, A32275a paralectotype, latex mould, external of pedicle valve. ‘The Frolic, Haverfordwest’, Dyfed, Wales. Lower Llandovery. Figs. 16-19. Tuvaella rackovskii Chernyshev, 1937. GS 45393 hypotype, ventral, lateral, dorsal, and posterior views of adult specimen. Elegest River, Tuva, U.S.S.R. Wenlock. (See Vladimirskaya 1973.) Figs. 20-23. Undescribed genus (Nikiforova, in press). GS 45396, hypotype, lateral, posterior, dorsal, and ventral views of adult specimen. Khrebta Shakhriomon, central Asiatic U.S.S.R., Ordovician 03 (Ashgill). Affinities of this taxon are still unclear— may be ancestral to Tuvaella, or possibly a spiri- gerininid. PLATE 39 COPPER, Ordovician and Silurian Zygospiridae 320 PALAEONTOLOGY, VOLUME 20 Genus pentlandella Boucot, 1964 Type species. Rhynconella pentlandicus Haswell, 1865 [sic], p. 31, pi. 3, figs. 9-10. (See PI. 39, figs. 8-12.) Range. Llandovery. Distribution. Scotland, Wales, and Estonia. Diagnosis. Small, globose, ventribiconvex, finely ribbed shells with hypercline beaks, minute foramina, and weak fold-sulcus. Ribs have micro-growth lines but no major interruptions; apically ribs are very faint to absent. Internally, the pedicle valve is thickened posteriorly; the brachial valve is also strong. Groove for ventral adductor and diverging diductors implanted into raised median septum. Dorsal muscle field flabellate, also divided by septum. No dental cavities or nuclei, teeth have wide bases, pointed extremities. Cardinal process apparently absent. Crura bases ball- like, thin crura rapidly spread laterally from crural bases. Jugum posteriorly located, joined almost at first spiral whorl, shaped like flattened W, or U with square corners. Spiralia with three to four whorls medially and medio-dorsally directed (text-fig. 12). text-fig. 12. Reconstruction of spiralia and jugum in Pentland- ella tenuistriata Rubel, 1970, based on text-fig. 13. Approx, x 7. Species assigned. In addition to the type species, only two other species appear to be known : Zygospira haswelli Reed, 1908 (pp. 434-435, pi. 14, figs. 4-9). Haverford Mudstone Formation (Lower Llandovery), ‘Locality K, below the path SW. of Uzmaston Farm’, Wales (label on lectotype A32771, Sedgwick Museum). Plate 39, figs. 13-15. Pentlandella tenuistriata Rubel, 1970 (pp. 27-28, pi. 17, figs. 1-24). Adavere Horizon, Estonia, U.S.S.R. (Upper Llandovery), figured by Rosenstein 1941, p. 6, fig. 74a incorrectly as ‘Catazyga' fur cat a (Sowerby, 1839); the Sowerby species is not atrypoid. Comparison. Catazyga is the most similar atrypoid genus, but differs in its muscle field and septal structures, simpler jugum and larger spiralia, and in its massive, bulbous tissue surrounding the crural bases. As the last-surviving catazyginid, ranging into late Llandovery time, and maybe the Wenlock, Pentlandella is not a widespread genus. It appears in Estonia to be a deeper-water inhabitant, like Ordovician Catazyga , which may explain its rarity in the Llandovery of Wales and COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 321 the Welsh Borderland. The genus is absent in richly fossiliferous Llandovery rocks of Anticosti Island, and those of the Siberian Platform, suggesting a provincial distribution. It is possible that the raised muscle areas of Pentlandella represent a trend towards a septally elevated muscle field such as in the mid-Devonian Gruene- waldtia. However, there are no connecting taxa, and raised muscle fields were independently developed in other different atrypoid lineages. The older P. haswelli has a hatchet-shaped ventral muscle field, becoming arrow-shaped in P. pentlandica. Pentlandella tenuis triata Rubel, 1970 Plate 39, figs. 3-7; text-figs. 12-13 Remarks. Rubel (1970) has well illustrated the external features of the Estonian species. Detailed internal serial sections, however, have not previously been avail- able. Material for sectioning was made available to me by Dr. Rubel. The sections can now be compared adequately with those of Catazyga and other zygospirids (text-fig. 13). Note especially the thickened posterior shell wall portions, crura, and spiralia. Subfamily tuvaellinae Alikhova, 1960 (emend.) Vladimirskaya (1972) pointed out that Tuvaella had dorsally directed spiralia and would be more correctly placed in the Zygospiridae amongst the atrypoids, instead text-fig. 13. Serial sections of Pentlandella tenuistriata Rubel, 1970 based on acetate peels. GS 45381, Adavere Horizon (Upper Llandovery); Latikula, Estonia, U.S.S.R. The type species of Pentlandella and related British species are preserved as moulds and unsuitable for sectioning. Scale x 5. 322 PALAEONTOLOGY, VOLUME 20 of in the Orthacea. Alikhova (1960) had erected a special family, with a single genus, in the Orthacea to account for its peculiar morphology. Through the kindness of Dr. Vladimirskaya, who presented me with a suite of specimens, I was able to examine well-preserved specimens of Tuvaella from Tuva in the Asiatic U.S.S.R. This has led to a slightly revised diagnosis and some new discoveries. The Tuvaellinae are unique in their stratigraphical distribution in rocks of latest Llandovery through Ludlow age. In addition to their size, and flattened dimensions, making them the largest known zygospirids, the Tuvaellinae are distinctive in several respects: they possess wide, flat interareas, a massive cardinal process and very thick hinge plates, and a dorsal jugum which comes to a blade-like, ventrally directed point between the spiral cones. They also differ from Zygospirinae by their lack of dental cavities or nuclei. They belong clearly to the Zygospiridae on two counts: firstly, their jugum is dorsal to the spiralia (as against ventral in all other atrypoid stocks), and secondly, the jugum is one piece (as opposed to two separated processes; see text-fig. 14). Otherwise, externally Tuvaella itself has a strong resem- blance to the Carinatina group of the Devonian. It may have evolved independently in that direction, but other ancestors for the Carinatininae lie more in the range of Neospirigerina Rzhonsnitskaya, 1975. A possible ancestor to the Tuvaellidae may be an undescribed genus (Nikiforova, in press; see PI. 39, figs. 20-23) from late Ordovician rocks of central Asia (Shakhrio- mon area). Poorly preserved material given to me by Dr. Nikiforova reveals a coarsely ribbed, large, weakly carinate, zygospirid shell, possessing a clear, relatively wide shelf-like interarea, large exposed foramen, and deltidial plates and orthocline- anacline beak. However, internal structures are still unknown, and the poorly preserved material may be a Spirigerina or Eospirigerina. There are probable phylo- genetic ties from Tuvaella to the Devonian sub-family Carinatininae, including Biconostrophia Havlicek, 1956 (and its synonym Davidsoniatrypa Lenz, 1968), Prodavidsonia Havlicek, 1956, Davidsonia Bouchard-Chantereux, 1849, Carinatina Nalivkin, 1930, and Eifelatrypa Copper, 1973. Some, if not all, of these taxa have highly developed cardinal processes and rather similar hinge plates. If this assump- tion proves correct, then the Palaferellidae, in the sense of Struve (1961) and Copper (1973), are polyphyletic, and need restudy. Genus tuvaella Chernyshev, 1937 Type species. Tuvaella rackovskii Chernyshev, 1937, p. 12. Range. Wenlock-Pridoli(?). Distribution. Asiatic U.S.S.R., Mongolia. Diagnosis. Relatively large and wide, coarsely and evenly ribbed zygospirids with somewhat carinate ventral fold and dorsal sulcus. Hinge line long, straight; inter- area wide and extensive; beaks orthocline to partly anacline; foramen covered by thick deltidial plates. Internally solid teeth, a strong cardinal process, dorsally to medio-dorsally directed spiralia, and a jugum arising posteriorly with a central spine- like meeting point are characteristic of at least the type species (internals of other species not known). COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 323 Tuvaella rackovskii Chernyshev, 1937 Plate 39, figs. 16-19; text-figs. 14-15 1937 Tuvaella rackovskii Chernyshev, pp. 12, 64, pi. 1, figs. 8-11. 1972 Tuvaella rackovskii Chernyshev; Vladimirskaya, pp. 39-42, pi. 6, figs. 1-14. Range. Llandovery- ?Wenlock or Ludlow. Distribution. Altai Mountains, Tuva, western Sayan, eastern Transbaikal, and north-east U.S.S.R. (Amur region), Mongolia. Not known outside U.S.S.R. and Mongolia. Remarks. Vladimirskaya (1972) provided the first detailed internal description of the type species of Tuvaella , including serial sections and photographs of the brachidia. Differential, partial recrystallization of the shells permitted Vladimirskaya to etch specimens to bring out specific features. At the same time this recrystallization also text-fig. 14. Reconstruction of the brachidia in Tuvaella rackovskii Chernyshev, 1937 based on text-fig. 15. obscured the original shell and brachidial structure. I was unable to discover the canals in the dorsal median septum in unrecrystallized material: possibly recrystal- lization of the shell can give the impression of the existence of such canals. The enormous cardinal process, projecting well into the pedicle cavity is unusual. The crura do not geniculate at the angle shown by Vladimirskaya in the material sectioned by me. The jugum is broadly convex and not angular and projects medially into a remarkable blade (text-fig. 15), not known in any other atrypoids. This may have held mouth parts, or less possibly muscle structures. Family atrypidae Gill, 1951 Subfamily clintonellinae Poulsen, 1943 Shells characteristic of this subfamily are biconvex, tubular-ribbed, usually non- carinate, non-lamellose, thin-walled, and with a narrow interarea, small beak, and 324 PALAEONTOLOGY, VOLUME 20 text-fig. 15. Serial sections of Tuvaella rackovskii Chernyshev, 1937 based on acetate peels. GS 45377, Wenlock; Elegest River, Tuva, U.S.S.R. Inset shows the pedicle structure. Scale x 5. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 325 small deltidial plates. Internally the group has relatively large dental cavities, thin hinge plates, delicate narrow crura, separated ventrally located jugal processes, and dorsally directed spiralia. This subfamily has characteristics similar to those of Devonian Desquamatia. Nikiforova and Modzalevskaya (1968) placed Nalivkinia ( Anabaria ), alongside Alispira, in the family Atrypidae, without special subfamily designation. Lopushinskaya (1965) assigned it to the Clintonellinae. Kulkov (1967) thought that Nalivkinia was related to the Karpinskiinae via Eokarpinskia Rzhonsnit- skaya, 1964, and later referred it to the Atrypinae (Kulkov 1974). In 1973 I allocated Nalivkinia to the Atrypidae (erratum under Zygospiridae, p. 488) and Alispira and Clintonella to the Zygospiridae. Work on serial section internals convinces me that Lopushinskaya’s interpretation is substantially correct, that is the Clintonellinae form a genus group including Clintonella , Alispira, and Nalivkinia. These are still believed to have evolved independently from the main Atrvpa group as shown in Copper (1973). Because of the lack of brachidial data on the three clintonellinid genera, I am hesitant about suggesting other affinities. Possibly some of the Pala- ferellidae, like Gracianella and Eokarpinskia, are related to Nalivkinia (Anabaria). Alternatively affinities with Devonian Desquamatia or Carinatinella are suggested in morphology of the dental cavities and brachidia. The Clintonellinids lack the thicker, more massive shells of the palaferellids and their reinforced jugal processes, but may have moved in that direction by changing habitats from quieter marine to the higher-energy reef areas favoured by many palaferellids. Genus nalivkinia Bublichenko, 1927 Type species. Atrypa grunewaldtiaeformis Peetz, 1901, pp. 147-148, 376, pi. 4, fig. 2a -c from the Sara- chumysk River, Kuznetsk Basin, U.S.S.R. Range. Upper Llandovery-Wenlock (possibly to the lower Devonian). Distribution. Tuva, Altai Mountains, Salair, Kazakhstan, Central Asia. Remarks. Nalivkinia ranging from Llandovery to the late Silurian and possibly into lower Devonian time, appears at present to have been found only in the U.S.S.R. Bublichenko (1927) recognized a new genus in the Peetz species and, in addition to the type, described a new species, N. sibiriea, apparently from the same locality and horizon. Subsequently, Nikiforova and Andreeva (1961) described a species they called Catazyga rara from the Siberian platform, later identified by Lopushinskaya (1965) as belonging to a new subgenus Anabaria within Nalivkinia. Internally the only difference between them lies in the thicker muscle pads in Nalivkinia and externally in the greater globosity and beak incurvature. Identical features in the two taxa are the large dental cavities, ‘hair-line’ structures of the dental plate, and delicate crural bases and crural supports feathering distally. These similarities at present outweigh the differences on a generic level. The two sub- genera are maintained because of lack of comparative brachidial evidence. The internal structure of these atrypids is more ‘advanced’ than the zygospirids and members of the Atrypa reticularis group in the presence of clearly definable socket plates, more delicate crura, and a striated cardinal process in the noto- thyrial pit. 326 PALAEONTOLOGY, VOLUME 20 Species assigned to Nalivkinia (Nalivkinia). Inadequate data exists on the correct assignment of many Silurian atrypoid species referred in the past to Nalivkinia , Gruenewaldtia , or Atrypa. Any number of these could belong to Nalivkinia ( Nalivkinia ): doubtful species are preceded by a question mark. Nalivkinia sibirica Bublichenko, 1927 (p. 983, fig. 1 ; pp. 990-991, pi. 49, figs. 2, 7-8; pi. 50, figs. 4-10). This species occurs together with the Peetz type species in the same strata. It appears to be indistinguish- able by its sulcus and more globose shell as described, and may be a population variant. ? Nalivkinia linguata Borisyak, 1955 (pp. 65-66, pi. 10, figs. 4-9; pi. 13, figs. 8-10). Llandovery-Wenlock of central Kazakhstan, U.S.S.R. Nalivkinia minuta Menakova, 1964 (pp. 26-27 , pi. 5, figs. 8 10). Daurich Mountains, U.S.S.R., Wenlock, Bed ‘L’. Nalivkinia ( Nalivkinia ) gruenewaldtiaeformis Peetz, 1901 Plate 40, figs. 1-4; text-fig. 16 1901 Atrypa griinewaldtiaeformis Peetz, pp. 147-148, 376, pi. 4, fig. 2a- c. Holotype, Leningrad University Museum, specimen 81/90. 1927 Nalivkinia griinewaldtiaeformis Peetz; Bublichenko, pp. 989-990, 1002-1003, pi. 50, figs. 1-3. 1967 Nalivkinia griinewaldtiaeformis Peetz; Kulkov, pp. 106-108, pi. 18, figs. 1-2. 1974 Nalivkinia griinewaldtiaeformis Peetz; Kulkov, pp. 61-62, pi. 21, figs. 3-6. Remarks. Both Bublichenko (1927) and Kulkov (1974) illustrate a connected jugum for the type species of Nalivkinia. However, serial sections in Kulkov (1974, p. 62) indicate that the jugal processes are only very weakly joined together, and possibly recrystallization has obscured what may actually be disconnected structures. This is similar to many reconstructions for Devonian Atrypidae, which have been shown to be incorrect (Copper 1967). Jugal processes frequently come so close together that they are easily confused for being fused. More work is still needed on the interior of Nalivkinia. Significant apical internal structures of N. gruenewaldtiaeformis are the large dental cavities, the presence of a thin true dental plate (appearing as a ‘hair-line’ structure in serial peels) along the inner margin of the pedicle cavity. EXPLANATION OF PLATE 40 All figures x 2, except figs. 19-20 approx. x40. Figs. 1-4. Nalivkinia ( Nalivkinia ) gruenewaldtiaeformis (Peetz, 1901). GS 45398 hypotype, ventral, dorsal, lateral, and posterior views of adult. Note that ribs have few anterior growth interruptions and lack frilly projections. Sara-chumysk, Salair, U.S.S.R. (locality K6710 Kulkov); marly limestones, Wenlock. Figs. 5-8. Nalivkina (Anabaria) rara (Nikiforova, 1961). GS 45397 hypotype, ventral, dorsal, lateral, and posterior views of adult specimen. Ribs are tubular, lack frills. Omnutakh River, Siberian Platform, U.S.S.R. Wenlock. Figs. 9-13, 20. ‘ Clintonella ’ anticostiana (Twenhofel, 1928). 9-13, GS 45399 hypotype, ventral, dorsal, lateral, posterior, and anterior views. Note lack of carination, accentuated mid-ribs. 20, GS 45392, photomicrograph of hinge plate and tooth structure (serial peel). NTS Jupiter River 75140:87850, Jupiter Formation (Zones 6-7), Anticosti Island, Canada. Upper Llandovery. Figs. 14-18. Alispira gracilis Nikiforova, 1961. GS 45400 hypotype, ventral, dorsal, lateral, posterior, and anterior views of adult specimen. Note the carination, resulting in dorsal sulcus (contrast with figs. 9-13). Podkamennaya River, Tunguska, Siberian Platform, U.S.S.R. Llandovery. Fig. 19. Catazyga headi headi (Billings, 1862). GS 45397 hypotype, photomicrograph of serial peel showing crural base (cb), hinge plate and tooth (t). NTS Becancour 01450:38880, Pontgrave River Formation, Canada. Ashgill. PLATE 40 19 COPPER, Ordovician and Silurian Zygospiridae and Alispira 328 PALAEONTOLOGY, VOLUME 20 vertically aligned teeth, thin, horizontally aligned socket plates, delicate crural bases, and fibrous crural tips (text-fig. 16). Material sectioned was from the syntype collection of Peetz located in the Museum of the Historical Geology Section of Leningrad University, and made accessible by E. S. Poretskaya. The large syntype collection shows that a gradation exists in external morphology between the type species and the syntopic species N. sibirica Bublichenko, 1927. Nalivkinia ( Nalivkinia ) has not been found on the Siberian Platform, in the Baltic region, or in Britain, and seems to be limited to the southern and Asiatic parts of the U.S.S.R. Its restricted distribution is comparable to Tuvaella. An external rib structure like Nalivkinia occurs in some atrypids from Anticosti Island, Canada, but internally these are very different. text-figs. 16-17. Serial sections of Nalivkinia gruenewaldtiaeformis (Peetz, 1901); (top), GS 45376, Wenlock, Sara-chumysk, Salair, U.S.S.R.; and Anabaria rara (Nikiforova, 1961) (bottom), GS 45375, Wenlock, Omnutakh River, Siberian Platform, U.S.S.R. Scale x 5. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 329 Genus nalivkinia Bublichenko, 1927 Subgenus anabaria Lopushinskaya, 1965 Plate 40, figs. 5-8; text-fig. 17 Type species. Cata:yga( ?) rara Nikiforova and Andreeva, 1961 (pp. 248-249, pi. 54, figs. 1-7). Wenlock, south-west Siberian Platform, Omnutakh River, U.S.S.R. Remarks. Lopushinskaya (1965) correctly saw the affinity of Catazyga{l) rara to the genus Nalivkinia. The type species of Nalivkinia , TV. gruenewaldtiaeformis (Peetz, 1902) differs only slightly in its internal structure from the species rara (see this text). The only difference lies in external form, and in shell-wall thickness, which could be an infrageneric character. Externally, Anabaria has a less-incurved beak and a slightly flatter shell with broad sinus. The internal similarities are so striking that it is hard to tell them apart. Except for jugal processes and spiralia, which were not present in the material sectioned, the following items were identical in the two taxa: large dental cavities and straight hair-line structure (dental plate s.s.) on the inner side of the tooth, deltidial plates (small, hollow apically), cardinal process (thin lining in notothyrial pit), distinct thin socket plates, small crural bases expand- ing into thick, curved, feathery crura (text-fig. 17). The internal affinities of both Nalivkinia and Anabaria are with Desquamatia of the Devonian ; internally Nalivkinia and Anabaria are so much alike it is probably better to classify them under the same genus and even subgenus. Beak incurvature is variable in the Sara-chumysk collec- tions of Peetz. Beak incurvature and shell form are also infraspecifically different in Devonian taxa. However, until the jugal processes and spiralia are better known, it may be advisable to follow Lopushinskaya’s separation. In looking at the serial sections particular attention should be paid to comparisons of gruenewaldtiaeformis and rara (text-fig. 17). Species assigned to Nalivkinia (Anabaria). ? Nalivkinia kazachica Borisyak, 1955 (pp. 61-63, pi. 10, fig. 3; pi. 11, figs. 17-19; pi. 13, figs. 1-3). Silurian, Central Kazakhstan, U.S.S.R. Lateral cavities and ‘dental plates’ seem to indicate Anabaria affinities. INalivkinia kassini Borisyak, 1955(pp. 63-64, pi. 10,figs. 1-2; pi. 13, figs. 6-7). Silurian, Central Kazakh- stan, U.S.S.R. Serial sections not definitive. INalivkinia rhomboidalis Borisyak, 1955 (pp. 64-65, pi. 10, figs. 10-13; pi. 13, figs. 4-5). Silurian, Central Kazakhstan, U.S.S.R. More coarsely ribbed, possibly an Alispira. Genus alispira Nikiforova, 1961 Plate 40, figs. 14 18; text-fig. 18 Type species. Zygospira ( Alispira ) gracilis Nikiforova, 1961, pp. 244-247, pi. 53, figs. 1-8; text-fig. 41. Range. Llandovery-Wenlock. Distribution. U.S.S.R. (Siberian Platform, Estonia). Not known from Britain, the Baltic, or other well- known Llandovery-Wenlock sections. Remarks. The atrypoid relationships of this genus are without doubt, as confirmed by Nikiforova (1961) and Nikiforova and Modzalevskaya (1968). However, since the detailed internal structure of the poorly preserved, synchronous North American 330 PALAEONTOLOGY, VOLUME 20 genus Clint onella is yet unknown (see Boucot and Johnson 1970), doubt remains if Alispira is a junior synonym of Clintonella. If both taxa are characterized by carina- tion, and if internal structures are similar, there may be no way of telling them apart. Clinton-age material in eastern North America is generally either silicified or pre- served as moulds and casts, and sheds no light on the significant internal structures. The Jupiter River Formation of Canada yields Homeospira anticostiana which is probably atrypoid (see PI. 40, fig. 20; text-fig. 19); but whether this is synonymous with Clintonella is not clear, since the Anticosti material lacks carination. Serial sections of topotypic Alispira (text-fig. 18) and the Anticosti material of ''Clintonella'’ (text-fig. 19) is compared side by side. Note the disposition of the hinge plate and crura in both species. text-figs. 18-19. Serial sections of Alispira gracilis (Nikiforova, 1961); (left), GS 45382 from the Pod- kamennaya River, Tunguska, Siberian Platform; Llandovery; (right), Clintonella ? anticostiana (Twenhofel, 1928), NTS Jupiter River 75140:87850; Zones 6-7, Jupiter Formation (Upper Llandovery). Compare the apical structures in the two species. Scale X 5. Genus clintonella Hall and Clark, 1893 Type species. Clintonella vagabunda Hall and Clarke, 1893, pp. 159-161, pi. 52, figs. 1-1 1. Original material derived from ‘drifted and decomposed block of sandstone found without label among the collections presented to the New York State Museum’. Probably derived from the Clinton group in Orleans county (ibid., p. 161). Range distribution. Llandovery and possibly Wenlock, North America. Remarks. The affinity of this genus is still unsettled. Hall and Clarke (1893) intimated a close relationship with Zygospira on the basis of the hinge plate and muscle impres- sions. However, they stated that no spiralia had been found in any material examined. Boucot and Johnson (1970), despite new material, came no closer to an answer on affinities, since they, like Hall and Clarke before them, failed to find spiralia. Nevertheless, they were able to establish a more precise age for some Clintonella specimens, namely Llandovery C3 to C4, though leaving open the question of rela- tions with the Zygospiridae or Atrypidae. Clintonella has affinities with the Siberian Llandovery Alispira Nikiforova, 1961 in sharing a carinate pedicle valve and sulcate COPPER: ZYGOSP1RA AND SOME RELATED BRACHIOPODS 331 brachial valve (see PL 40, figs. 14-18 for topotypic material of Alispira). This carina- tion is not always marked in illustrated material of A. gracilis tennicostata. Alispira possesses dental cavities, dorsal atrypoid spiralia, postero-ventral jugal processes (separated), and somewhat unusual hinge plates (see Modzalevskaya’s serial sections in Nikiforova 1961, p. 246; compare with text-fig. 18). Nikiforova pointed out that some North American ‘ Homeospira e.g. "H.' anticostiana Twenhofel, 1928 and H. subcircularis Savage, 1913 were similar to Alispira. This is a problem still under investigation, particularly in regard to Anticosti material (PI. 40, figs. 9-13, 20; text-fig. 19). A number of North American taxa assigned to ‘ Homeospira" may be true atrypoids, and may belong to a new genus of clintonellinid lacking carination of any kind. Provisionally it is better to retain North American atrypoids of this type under Clintonella. Clintonellal anticostiana (Twenhofel, 1928) Plate 40, figs. 9-13, 20; text-fig. 19 1928 Homeospira anticostiana Twenhofel, p. 220, pi. 20, figs. 1-3. Remarks. Text-fig. 19 serializes the apical structures of a typical specimen from Anti- costi Island. Calcified material has not yet revealed spiralia. Dental cavities are distinct, small deltidial plates about the apex of the brachial valve, teeth are dorso- medially directed and simple in design. The hinge plate is somewhat similar to Alispira , but crura develop rather differently from crural bases (see PI. 40, fig. 20; text-fig. 18). The inner part of the hinge plate is unique for atrypoids, in having medially pointed blades; the significance of this is not clear. Some Alispira (e.g. A. rotundata Nikiforova and Modzalevskaya, 1968, pp. 59-61, pi. 2, figs. 1-7) also have a similar hinge plate development. CONCLUSIONS Before atrypoid taxonomy can reach a satisfactory major scheme, a number of problems need to be solved. Most of these solutions rest on a better knowledge of internal structures, especially of the brachidia and pedicle structures. The theme of this paper is that the early atrypoids, the Zygospiridae (mostly of Ordovician age), were all ‘designed’ with a similar dorso-medially pointed spiralia and a dorsal, fused jugum. This implies a common method of solving the problems of food- gathering and waste disposal. In terms of external morphology, the zygospirids lacked many of the refinements of the Devonian atrypoids (the Silurian representing a transitional stage in morphology). Zygospirids did not develop specializations for settling or anchoring on soft muddy bottoms (e.g. frills, spines, planation or elonga- tion of one or both valves, cementation, pedicle peculiarities). The zygospirinids were probably the ancestors of the Atrypidae, the tuvaellinids possibly of the Carinatininae, and the catazyginids were possibly the ancestors of the Palaferellinids (s.s.) and the Karpinskiinids. Knowledge is needed of the interior of different species assigned at various times to transitional generic forms, e.g. ‘ Eospirigerina ’, ‘ Clinton - ellct (or ‘ Homeospira ’), "Alispira', ‘ Zygospira ’, etc. There is ample scope for further work. 332 PALAEONTOLOGY, VOLUME 20 Acknowledgements. I wish to thank especially the following colleagues for providing me with type materials and/or pertinent literature: Dr. G. A. Cooper, Washington, Dr. T. E. Bolton, Ottawa, Dr. Y. Globensky, Quebec, Dr. O. I. Nikiforova, Dr. T. L. Modzalevskaya, Dr. Y. V. Vladimirskaya, and E. S. Poretskaya, Leningrad, Dr. M. Rubel, Tallinn, Dr. K. Rozman, Moscow, Dr. N. Newell, New York, and Dr. V. Jaanusson, Stockholm. Field and museum work and exchange visits were arranged by the National Research Council (Canada) and the Akademiya Nauk (U.S.S.R.). REFERENCES alikhova, t. n. 1960. Family Tuvaellidae, fam. nov. Osnovy paleontologii (ed. t. g. sarycheva), 1 15-343, 75 pis. Moscow. amsden, t. w. 1971. Late Ordovician-early Silurian brachiopods from the central United States. Mem. Bur. Rech. geol. min. 73, 19-25, pis. 1-2. beecher, c. e. and schuchert, c. 1893. Developments of the brachial supports in Dielasma and Zygospira. Proc. biol. Soc. Wash. 8, 71-82, pis. 10-1 1. billings, e. 1862. New species of fossils from different parts of the Lower, Middle and Upper Silurian rocks of Canada. Palaeozoic fossils, 1 (4), 96-168, Geol. Surv. Canada, Montreal. — 1866. Catalogues of the Silurian fossils of the island of Anticosti, with descriptions of some new genera and species, 93 pp., Geol. Surv. Canada, Montreal. bolton, t. e. 1972. Geological map and notes on the Ordovician and Silurian litho- and biostratigraphy, Anticosti Island, Quebec. Geol. Surv. Pap. Can. 71-19, 1-45, pis. 1-12. borisyak, m. a. 1955u. Silurian (Wenlockian) brachiopods from the Karagandinsk district. Mater Vses. Nauchno-issled. Geol. Inst. 3 (1), 69 pp., 10 pis. — 19556. Stratigraphy and brachiopods from Silurian beds of the Chingiz Mountain region. Ibid. 5 (2), 79 pp., 13 pis. boucot, a. j. 1964. A new genus of Silurian atrypacean brachiopod. Quad. Map. Ser. Maine geol. Surv. 2, 105, pi. 2. — and Johnson, j. G. 1967. Species and distribution of Coelospira (Brachiopoda). J. Paleont. 41, 1226- 1241, pis. 163-166. — pitrat, c. w. and staton, r. d. 1965. Spiriferida. In williams, a. et al. Treatise on Invertebrate Paleontology. Vol. H. Brachiopoda. Geol. Soc. Am. and Univ. Kansas Press, pp. H1-H927, figs. 746. and staton, r. d. 1964. On some atrypoid, retzioid and athyridoid Brachiopoda. J. Paleont. 38, 805-822, pis. 125-128. bradley, J. H. 1921. The Brachiopoda of the Maquoketa of Iowa. Bull. Mus. comp. Zool. Harv. 64, 503-525, pis. 1-2. bretsky, p. w. 1969. Ordovician benthic marine communities in the central Appalachians. Bull, geol Soc. Am. 80, 193-212. 1970. Upper Ordovician ecology of the central Appalachians. Bull. Peabody Mus. nat. Hist. 34, 150 pp., 44 pis. bublichenko, N. l. 1927. Die Brachiopodenfauna des unteren Palaeozoicum aus der Umgegend des Dorfes Sara-Tschumysch aus dem Kohlenbassin von Kusnetzk. Izv. geol. Kom. 46, 979-1008, pis. 49-50. caley, j. f. 1936. The Ordovician of Manitoulin Island, Ontario. Mem. geol. Surv. Can. 202 (2), 21-92, pis. 1-6. CHERNYSHEV, V. V. 1937. Silurian brachiopods of Mongolia and Tuva. Trudy, mongol. Kom. 29 (5), 1-94, 5 pis. clarke, t. h. 1919. A section in the Trenton Limestone at Martinsburg, New York. Bull. Mus. comp. Zool. Harv. 63, 3-18, pi. 1. cooper, g. a. 1956. Chazyan and related brachiopods. Smithson misc. Colins, 127, 1024 pp., 269 pis. and kindle, c. H. 1936. New brachiopods and trilobites from the Upper Ordovician of Perce, Quebec. J. Paleont. 10, 348-372, pis. 51-53. copper, p. 1967. Brachidial structures of some Devonian atrypid brachiopods. Ibid. 41, 1176-1183, pis. 155-156. 1973. New Siluro-Devonian atrypoid brachiopods. Ibid. 47, 484-500, pis. 1-3. cumings, E. R. 1908. The stratigraphy and paleontology of the Cincinnati Series of Indiana. 1189 pp., 55 pis. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 333 davidson, T. 1 867. A monograph of the British fossil Brachiopoda : the Silurian Brachiopoda. Palaeontogr. Soc. [Monogr.] 7, part 2, 89-168, pis. 13-22. — 1882-1883. A Monograph of the British fossil Brachiopoda: Supplement to the British Silurian Brachiopoda. Ibid. (1882, 63-134, pis. I 7), (1883, 135-242, pis. 8-17). fenton, c. l. and fenton, m. a. 1924. Some Blackriver brachiopods from the Mississippi Valley. Proc. Iowa Acad. Sci. 29, 67-77, pis. 1 -2. foerste, a. f. 1909. Preliminary notes on Cincinnatian and Lexington fossils. Bull. Scient. Labs. Denison Univ. 14, 289-324, pis. 7-11. — 1910. Preliminary notes on Cincinnatian and Lexington fossils of Ohio, Indiana, Kentucky and Tennessee. Ibid. 16, 17-87, pis. 1-6. — 1916. Upper Ordovician formations in Ontario and Quebec. Mem. geol. Surv. Can. 83, 279 pp. — 1917. The Richmond faunas of Little Bay de Noquette in northern Michigan. Ottawa Nat. 31, 97-103, pis. 4-6. — 1924. Upper Ordovician faunas of Ontario and Quebec. Mem. geol. Surv. Can. 138, 255 pp., 46 pis. hall, j. 1847. Containing descriptions of the organic remains of the lower division of the New York system (equivalent of the lower Silurian rocks of Europe). Palaeontology of New York, 1, 338 pp., 33 pis. Albany, New York. — 1857. Descriptions of Palaeozoic fossils. Rep. N.Y. St. Mus. nat. Hist. 10, 41-186. 1862. Observations upon a new genus of Brachiopoda. Ibid. 15, 154-155. 1867. Note upon the genus Zygospira and its relations to Atrvpa. Ibid. 20, 267-268. 1879. The fauna of the Niagara Group. Ibid. 28, 98-203, pis. 3-34. — 1882. Descriptions of species of fossils found in the Niagara Group at Waldron, Ind. Rep. Indiana Dep. Geol. nat. Resour. 11, 217-345. and clarke, J. m. 1893. An introduction to the study of the genera of Palaeozoic Brachiopoda. Palaeontology of New York , 8 (2), 394 pp., 84 pis. Albany, New York. haswell, G. c. 1865. On the Silurian Formation of the Pentland Hills. 47 pp., 4 pis. Edinburgh. howe, h. j. 1965. Plectambonitacea, Strophomenacea, and Atrypacea from the Montoya Group (Ordo- vician) of Transpecos, Texas. J. Paleont. 39, 647-656, pis. 81-82. hussey, r. c. 1926. The Richmond Formation of Michigan. Contr. Mus. Paleont. Univ. Mich. 2, 113-187, 1 1 pis. kay, m. 1968. Ordovician formations in northwestern New York. Naturaliste can. 95, 1373-1378. kulkov, N. p. 1967. Brachiopods and stratigraphy of the Silurian of the Altai mountains. Akad. Nauk. Sibirsk. Otdel. Inst. Geol. Geof. 148 pp., 21 pis. Moscow. — 1974. In ivanovskii, a. b. and kulkov, n. p. Rugosans, brachiopods and stratigraphy of the Silurian of the Altai-Sayan mountain region. Trudy Inst. Geol. Geofiz. Sibirsk. Otdel. 231-(3), 39-96, pis. 15-25. lopushinskaya, T. v. 1965. On new brachiopods from Silurian beds of the Siberian Platform. Trudy Sib. nauchno-issled. Inst. Geol. Geofiz. -miner. Syr. 34, 23-33, pi. 1. meek, f. b. 1873. Descriptions of invertebrate fossils of the Silurian and Devonian systems. Geol. Surv. Ohio Paleont. 1 (2), 246 pp., 23 pis. menakova, G. n. 1964. Brachiopods from Lower Silurian beds of the Zeravshan-Gissarsk mountain region. Trudy Uprav. Geol. Okhran. Nedr Sov. Min. Tadzhik SSR, Paleont. Stratigr. 1, 3-73, 10 pis. Moscow. miller, w. i. 1910. Geology of the Port Leyden quadrangle, Lewis County, N.Y. Bull. N.Y. St. Mus. 135, 5-58, map. nettleroth, H. 1889. A monograph of the fossil shells of the Silurian and Devonian rocks of Kentucky, 245 pp., 36 pis. Kentucky Geol. Surv., Frankfort. nikiforova, o. i. 1961. In Nikiforova, o. i. and andreeva, o. n. Ordovician and Silurian stratigraphy of the Siberian Platform and its paleontological basis. Trudy Vses. nauchno-issled. geol. Inst. N.S. 56, 68-289, 56 pis. — and modzalevskaya, t. l. 1968. Some Llandoverian and Wenlockian brachiopods from the north- western parts of the Siberian Platform. Uchen. Zap. nauch-issled. Inst. Geol. Arktik. 21, 50-81, 5 pis. oraspyld, A. L. 1956. New brachiopods from the Iykhvisk, Keilask and Vazalemmask Florizons. Tartu Ulik. Geol-Inst. Toim. 1, 41-67, 4 pis. parks, w. a. and dyer, w. s. 1922. The Molluscoidea. Rep. Ont. Dep. Mines , 30 (7), 1-59, 7 pis. G 334 PALAEONTOLOGY, VOLUME 20 peetz, h. von. 1901. Beitrage zur Kenntniss der Fauna aus den devonischen Schichten am Rande des Steinkohlenbassins von Kusnetzk (West-Siberien). Izv. geol. Kom , 4, 393 pp., 6 pis. portlock, J. E. 1843. Report on the geology of the county of Londonderry and of parts of Tyrone and Fer- managh, 454-455, pi. 37, Geol. Surv. G.B., Dublin. poulson, c. 1943. The Silurian faunas of north Greenland, II, The fauna of the Offley Island Formation. Meddr Grc/mland. 72 (3), 60 pp., 6 pis. reed, F. r. c. 1905. Sedgwick Museum Notes. New fossils from the Haverfordwest District, IV. Geol. Mag. dec. 10, 2, 444-454, pi. 23. 1908. Sedgwick Museum notes: new fossils from the Haverfordwest District. VIII. Ibid. dec. 10, 5, 433-436, pi. 14. 1917. The Ordovician and Silurian Brachiopoda of the Girvan district. Trans. R. Soc. Edinburgh , 51, 795-998, pis. 1-24. — 1932. Report on the brachiopods from the Trondheim area. Norsk. Vidensk. Akad , 4, 115-146, pis. 18-22. Richards, R. p. 1972. Autecology of Richmondian brachiopods (Late Ordovician of Indiana and Ohio). J. Paleont. 46, 386-405, pis. 1-5. rosenstein, e. 1941. The Adavere Formation (Silurian, Llandovery) in western Estonia. Tartu Ulik. Geol- Inst. Toim. 63, 1-7, 1 plate. ROSS, R. J. and dutro, J. T. 1966. Silicified Ordovician brachiopods from east-central Alaska. Smithson. Misc. Colins, 149 (7), 22 pp., 3 pis. rozman, K.h. s. 1968. Class Brachiopoda. In balashov, e. g. et al. Guide atlas of the Ordovician faunas of northwestern USSR. Minist. Geol. RSFSR, 285 pp., 71 pis., Magadan. 1970. Stratigraphy and brachiopods of the Middle and Upper Ordovician of the Sette-Daban range and the Upper Ordovician of the Selennyakh range. Trudy Akad. Nauk SSR, Ord. Trud. Krasn. Znam. Geol. Inst. 205, 8-143, pis. 1-18. rubel, m. p. 1970. Pentamerid and spiriferid brachiopods from the Silurian of Estonia. Inst. Geol. Akad. Nauk Est. SSR, 75 pp., 1 1 pis. Tallinn. rukavishnikova, t. b. 1956. Ordovician brachiopods from the Chu-Iliisk mountains. Trudy Geol. Inst. Akad. Nauk SSR, 1, Ordovik Kazakhstana, 2, 105-168, pis. 1-5. rzhonsnitskaya, M. a. 1960. Order Atrypida. Osnovy Paleontologii, 257-264, pis. 53-56. 1964. On Devonian atrypids from the Kuznetsk Basin. Trudy Vses. nauchno-issled. Geol. Inst. 93, 91-112, 2 pis. — 1975. Biostratigraphy of the Devonian on the outskirts of the Kuznetsk Basin. Ibid. 244, 232 pp., 34 pis. sardeson, F. w. 1901. The range and distribution of the Lower Silurian fauna of Minnesota with descrip- tions of some new species. Bull. Minn. Acad. Sci. 3 (3), 326-343, pis. 1-6. SCHUCHERT, c. 1893. On the development of the shell of Zygospira recurvirostra. Proc. biol. Soc. Wash. 8, 79-82, pi. 11. 1894. A revised classification of the spire-bearing Brachiopoda. Amer. Geologist, 13, 102-107. sowerby, j. de c. 1839. Shells of the Lower Silurian rocks. In Murchison, r. i., The Silurian system, 634- 644, pis. 19-22, Murray, London. sproule, J. c. 1936. A study of the Cobourg Formation. Mem. geol. Surv. Can. 202, 93-1 1 1, pis. 7-9. steele, h. m. and Sinclair, G. w. 1971. A Middle Ordovician fauna from Braeside, Ottawa Valley, Ontario. Bull. geol. Surv. Can. 211, 51 pp., 23 pis. twenhofel, w. H. 1928. Geology of Anticosti Island. Mem. geol. Surv. Can. 154, 481 pp., 60 pis. ulrich, e. o. 1879. Descriptions of new genera and species of fossils from the Lower Silurian about Cincinnati. J. Cincinnati Soc. nat. Hist. 2, 8-30. 1888. A correlation of the Lower Silurian horizons of Tennessee and of the Ohio and Mississippi Valleys with those of New York and Canada. Am. Geologist, 1, 100-110, 179-198, 305-315, 2, 39-44. Vladimirskaya, E. v. 1972. On the systematic position and geological distribution of the genus Tuvaella (Brachiopoda). Paleont. Zhurn. 1972 (1), 37-44, pi. 6. wang, Y. 1949. Maquoketa Brachiopoda of Iowa. Mem. geol. Soc. Am. 42, 55 pp., 11 pis. willard, B. 1928. The brachiopods of the Ottosee and Holston formations of Tennessee and Virginia. Bull. Mus. comp. Zool. Harv. 68 (6), 255-309, pis. 1 -3. COPPER: ZYGOSPIRA AND SOME RELATED BRACHIOPODS 335 williams, a. 1962. The Barr and lower Ardmillan Series (Caradoc) of the Girvan district, south-west Ayrshire, with descriptions of the Brachiopoda. Mem. geol. Soc. Lond. 3, 267 pp., 25 pis. wilson, a. E. 1946. Brachiopoda of the Ottawa Formation of the Ottawa-St. Lawrence Lowland. Bull, geol. Surv. Can. 8, 14 pp., 11 pis. winchell, n. h. and schuchert, c. 1891. The Lower Silurian Brachiopoda of Minnesota. Geology of Minnesota, Geol. Nat. Hist. Surv. Minn. 3(1), 333-474, pis. 29-34. ziegler, a. m. 1965. Silurian marine communities and their environmental significance. Nature , Lond. 207, 270-272. Original typescript received 6 January 1976 Revised typescript received 3 June 1976 p. COPPER Department of Geology Laurentian University Sudbury, Ontario Canada CO R ALL! AN (UPPER JURASSIC) MARINE BENTHIC ASSOCIATIONS FROM ENGLAND AND NORMANDY by F. T. FURSICH Abstract. Using the trophic group approach, seventeen macroinvertebrate and one trace-fossil association have been described quantitatively, using 170 bulk collections, from the Middle and Upper Oxfordian of Normandy, Dorset, Oxfordshire, and Yorkshire. Most associations are dominated by bivalves and gastropods, whilst echino- derms and brachiopods are less important. Sedimentary and biostratinomic evidence indicate that, with few excep- tions, the associations are the autochthonous or parautochthonous relics of ancient communities. The taphonomy and environments of each association have been discussed and comparisons have been drawn with other Mesozoic benthic associations. The highly abundant oyster Nanogyra nana is interpreted as an opportunistic species and the abundance of the pectinid Chlamys is regarded as a feature typical of Corallian faunas. The scarcity of brachiopods, which are also uncommon elements in many other Jurassic and Cretaceous benthic faunas, is thought to be related to the short dispersal time of their larvae, combined with competition by bivalves and changing environments. Palaeoecology, and especially studies of ancient communities triggered off by research in modern ecology have increasingly gained popularity amongst palae- ontologists. Besides the classic work of Petersen (1913) on the fauna of the North Sea and that of Thorson (e.g. 1933, 1957) it is on studies of Russian workers like Turpaeva (1948, 1949, 1957; see also Neyman 1967) that methods and approach in palaeosynecological studies are based. Studies of communities include the descrip- tion of their composition and structure and their relationship to the environment. Generally, these aspects can best be studied on Recent communities whilst another aspect, that of community evolution, represents the main contribution of the palaeontologist to the field of synecology. The aim of this paper is to contribute to this last aspect of community studies, although it can serve only as a small contribu- tion to a framework of data which will eventually enable us to test concepts now in vogue. Studies of ancient communities or associations (the latter term is preferred for reasons given below) have been mainly on the Palaeozoic (e.g. Boucot 1975; Bretsky 1969, 1970; Hurst 1975; Walker and Laporte 1970; Ziegler et al. 1968) and on the Cretaceous (e.g. Kauffman 1967; Rhoads et al. 1972; Scott 1974). For the Jurassic, invertebrate associations have been described quantitatively only recently (Wright 1973, 1974; Duff 1975), the bulk of palaeosynecological work being only semi-quantitative (e.g. Hallam 1960, 1967; Sellwood 1972). The purpose of the paper is to give a quantitative description and environmental interpretation of the seventeen macroinvertebrate and one trace-fossil association which have been recognized. Sedimentary and biostratinomic evidence has been used to establish the taphonomy and general environment of each association. Using this basic information, facies and substrate relationships of the fauna have been investigated and the results published separately (Fursich 1976a, b). [Palaeontology, Vol. 20, Part 2, 1977, pp. 337-385.] 338 PALAEONTOLOGY, VOLUME 20 Geological setting. The Corallian (Middle and Upper Oxfordian; text-fig. 1) stretches from Yorkshire down to Dorset in England and also in Normandy, France (text- fig. 2). The main sections under study were provided by cliffs in Yorkshire (Filey Brigg), Dorset (Bowleaze Cove, Osmington Mills to Ringstead Bay, Shortlake to Black Head, Sandsfoot Castle, East Fleet), and Normandy (Houlgate to Villerville). Additional information, especially on the reef facies, was gained in quarries in the Vale of Pickering, Yorkshire, in Oxfordshire, and in Berkshire (for a list of the localities see appendix). In Dorset, the Corallian consists of four limestone-clay-sand cycles (Talbot 1973), separated partly by erosion surfaces. These cycles correspond to regressing sequences, the erosion surfaces marking renewed transgressions. The sediments were laid down in a shallow sea, probably in the upper shelf region, and the main facies types comprise offshore calcareous sands and carbonates, oolite banks, nearshore muds, subtidal sand bars, lagoonal muds, intertidal sands, estuarine sands and muds, and nearshore ferruginous oolitic muds. The sediments of the Corallian of Yorkshire (sands, mixed limestones, oolites, and patch reefs) and Normandy (similar in lithology to Dorset) have received less attention, and one purpose of the study was to apply the palaeoecological evidence, tested in the well-known Dorset section, to these two areas (Fursich 19766). Previous work. There are several prerequisites to a palaeosynecological study: stratigraphical studies to facilitate correlations; sedimentological studies to define environments; and taxonomic studies to interpret the autecology of the faunal elements. For the Corallian, these prerequisites have been largely fulfilled. Blake and Hudleston (1877), Hudleston (1878), Arkell (e.g. 1927, 1933, 1935-1948, 1936), Wilson (1933, 1949), Callomon (1960), and Wright (1972) have dealt with the strati- graphy, whilst Twombley (1964), Wilson (1968a, b ), Fee (1971), Brookfield (1973a), and Talbot (1973, 1974) have developed various sedimentological models for the Yorkshire and Hampshire Basins. The bivalves, the bulk of the fauna, have been described by Arkell (1929-1937), and the trace fossils by Fursich (1974). Microfaunal studies have been carried out by Whatley (1965), Gordon ( 1965), and Guyader (1968). (For an extensive bibliography see Brookfield 1973a.) First attempts to distinguish faunal assemblages have been made by Arkell (1929-1937) who recognized, for example, a bivalve assemblage typical of the reef environment. Arkell (1928, 1935) also dealt with the ecology of the Corallian patch reefs in a more general way. The trace-fossil fauna has been interpreted in terms of distribution patterns and environ- mental significance by Fursich (1975). Otherwise, no ecological studies have been undertaken. text-fig. 1. Composite sections of the Corallian of the main areas under study; adapted from Fursich (1975). C.O.F. = Calcaire a oolithes ferrugineuses; OOL. D. TROUVILLE = Oolithe de Trouville a Nucleolites scutatus', F. 'CORAL RAG’ D. TROUVILLE = Facies 'Coral Rag’ de Trouville; C. HENN. = Calcaire de Flennequeville; A.N. VILL. = Argiles Noires de Villerville; O.C. = Oxford Clay; THB = ‘ Trigonia hudlestoni Bed; BENCL. G. Bencliff Grit; TRIG. CLAV. B. = ' Trigonia' clavel/ata Beds; SANDSF. C.= Sandsfoot Clay; SANDSF. G. = Sandsfoot Grit; R.W.C. = Ringstead Waxy Clay; K.C. = Kimmeridge Clay; L. CALC. GRIT = Lower Calcareous Grit; P.B. = Passage Beds; HAM. O. = Hambleton Oolite; M. CALC. GRIT = Middle Calcareous Grit; M.O. Malton Oolite. IC.QF1 ARGUES A LOPHA GREGAREA IOOL.D. TROUVILLE IF 'CORAL RAG' D. TROUVILLE |C. HENNJ A. N. VI LL. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 339 NORMANDY DORSET | ; I | arg. limestone limestone o.o oolite 1 0 m text-fig. 2. Outcrop of Upper Jurassic rocks in England and Normandy. FURSICH: CORALLIAN BENTHIC ASSOCIATIONS 341 Material and methods. One hundred and seventy bulk collections with more than 30000 fossils were used for this study. The collections were broken up partly in the field, partly in the laboratory and all macrofossils were saved. For bivalved animals, the fauna has been totalled by counting only the most numerous valve together with the articulated specimens. Preservation, fragmentation, encrusters and borers, size sorting, and relationship between right and left valves in bivalved animals were noted. Orientation of shells, especially with respect to life position was recorded as were the trace-fossil fauna, sedimentary structures, and diagenetic features. The diversity of the samples has been measured using Sanders’s (1968) rarefaction method (text-fig. 3). For grouping the collections into associations the trophic group approach text-fig. 3. Diversity of Corallian benthic associations expressed by rarefaction curves (Sanders 1968). 342 PALAEONTOLOGY, VOLUME 20 has been selected. Firstly, for each collection the trophic nucleus consisting of the numerically dominant species which make up 80% of the fauna (Neyman 1967) was constructed. Then collections with an identical or similar trophic nucleus were grouped into associations. Each of these associations is characterized by one or more species which reach here the peak of their distribution (text-fig. 4). In addition, the associations differ in their trophic nucleus, and in most cases exhibit a distinct ecologi- cal composition (e.g. text-fig. 6). The trophic group analysis, developed in Russia by Turpaeva (1948), has been discussed by Walker (1972) and applied to fossil material, for instance by Rhoads et al. (1972) and by Duff (1975). Rhoads et al. (1972) also discuss the drawbacks of this method when applied to fossil faunas, as biomass and original composition of the fauna have been lost. Walker (1972) sug- gested substituting biovolume for biomass when working with fossil samples and for this purpose, size histograms for the members of the trophic nuclei are given (e.g. text-fig. 7). The feeding levels usually recognized in trophic group analysis (Walker and Bambach 1974) have been augmented by the introduction, in inter- mediate position, between low-level and high-level suspension feeders, of medium- level suspension-feeders represented mainly by the bivalves Pinna and Gerville/la. This was thought necessary to account for the dense feeding stratification realized in the associations which are nearly all dominated by suspension-feeders. Pendent bivalves like Meleagrinella , Oxytoma , and probably also Pteroperna pvgmaea were —in contrast to Duff (1975) included into the benthos as high-level suspension- feeders, because their distribution is governed by the presence of seaweed or algae which in turn is influenced by substrate conditions. The Corallian faunas lack (as do most fossil samples) soft-bodied animals. Where possible, trace fossils have been used to compensate for this but they are no real substitute. The term ‘community’ therefore cannot be applied to these samples and the term ‘association’, meaning a non-random, recurrent assemblage of fossils usually representing the remnant of a community, is preferred (see also Fiirsich 1975; Duff 1975). For each association, the degree of distortion from the original community is discussed using biostratinomic and sedimentological data. The autecology of the bivalves was derived using the morphology of the hard parts and comparison with Recent species (based mainly on Stanley 1970, 1972). Hard parts of gastropods are far less indicative of a particular mode of life and often species of the same genus occupy completely different niches. Because of this, the reconstructed life habits of Corallian gastropods are often conjectural and open to criticism. For each associa- tion, the autecology of the members of the trophic nucleus has been presented in an attempted reconstruction of the ancient community (e.g. text-fig. 5). text-fig. 4. Numerical abundance of the most important Corallian body fossils in the various benthic macroinvertebrate associations (A-R). Dark stippling: association for which a species exhibits a pro- nounced preference and thus is highly characteristic; light stippling: associations in which a species preferredly occurs (usually not more than two); these species are less specific but still useful for defining associations. The nearly ubiquitous Nanogyra nana has been regarded as a very characteristic species where it represents more than 80% of the fauna (R) and as an accessory species where it represents more than 50% of the fauna (G, H). Torqu. = Tor quirky nchia inconstans; the other bar in the same column represents Thurmanella. FURSICH: CORALLIAN BENTHIC ASSOCIATIONS 343 344 PALAEONTOLOGY, VOLUME 20 table I . Modiolus bipartitus/ Pleuromya alduini association (A). (Similar tables of data for the remaining sixteen associations have been deposited with the British National Library, deposition SUP 14007.) Eighteen collections. 1798 specimens. Composition of fauna (members of the trophic nucleus are marked with an asterisk): bivalves (98-94%): °/ /o Pres. % % Pres. % *Nanogyra nana (J. Sowerby) 48-66 0-83 Placunopsis duriuscula (Phillips) 0-05 0-05 *Lopha (A.) genuflecta Arkell 8-17 0 55 Nuculana (N.) sp. 0-05 0-05 * Pleuromya alduini (Brongniart) 7-89 0-83 Myopholas sp. 0-05 0-05 *Chlamys ( R.) fibrosa (J. Sowerby) 7-17 0-66 Arcomytilus pectinatus (J. Sowerby) 0-05 0-05 * Modiolus bipartitus J. Sowerby 6-50 0-83 Praeexogyra sp. 0-05 0-05 *Gryphaea ( B .) dilatata (J. Sowerby) 3-50 0-83 Plagiostoma rigidum (J. Sowerby) 0-05 0-05 Lopha (A.) gregarea (J. Sowerby) Pholadomya aequalis J. de C. 3-05 0-78 gastropods (0T6%): Sowerby 2-66 0-50 Bathrotomaria sp. Oil 0-05 Anisocardia isocardioides (Blake Purpurina sp. 0-05 0-05 and Hudleston) 1 39 0-28 brachiopods (0-32%); Pinna lanceolata J. Sowerby 1-22 0-22 Pleuromya uniformis (J. Sowerby) 1-11 0-44 ‘ Rhynchonella' sp. 0 16 0-05 Deltoideum delta (Smith) 1 05 0-11 Thurmanella sp. 0 1 1 0-11 Ho mo my a sp. 0-88 0-33 ' Terebratula ’ sp. 0-05 0-05 Isognomon promytiloides Arkell 0-83 0-17 serpulids (0-55%): Isognomon sp. 0-44 0-39 Limatula elliptica (Whiteaves) 0-44 0-28 Serpula ( Dorsoserpula ) sulcata Myophorella sp. 0-38 0-22 J. Sowerby 0 11 Goniomya literata (J. Sowerby) 0-27 0-28 Serpula ( Dorsoserpula ) sp. 0-44 Oxytoma expansa (Phillips) 0-22 0-55 Serpula ( Cycloserpula ) intestinalis Quenstedtia daviesi Arkell 0-22 0-55 Phillips 0-55 0 16 Protocardia dyonisea (Buvignier) 0-22 0-55 Serpula ( Cycloserpula ) sp. 0 16 ‘ Ostrea' sp. 0-22 0 1 1 Serpula ( Pentaserpula ) sp. Oil Pholadomya prolei (Brongniart) 0 16 0-17 miscellanea : Parallelodon aemulum (Phillips) 0 16 0 11 Meleagrinella ovalis (Phillips) 0 16 0 1 1 Nubeculinella sp. 0-94 Cercomya undulata (J. de C. ‘ Cidaris ’ spines 0-05 Sowerby) 0 16 0-05 trace fossils: Isocyprina sp. 0 16 Oil Cucullaea contracta Phillips Oil 0-11 Planolites sp. 0-50 Plicatula weymouthiana Damon 0-11 0-11 Chondrites sp. 0-55 Mactromya aceste (d’Orbigny) Oil 0-11 pyritic tubes 0 61 Placunopsis radiata (Phillips) Oil 0-11 Talpina sp. 0-1 1 Protocardia intexta (Muenster) Trautscholdia cf. Tr. cordata Oil 0-05 ecological composition : (Trautschold) on 0-11 epifauna: 75-87% Chlamys (Chi.) splendens (Dollfus) Oil 0-11 semi-infauna: 7-73% Thracia depressa (J. de C. Sowerby) Oil 0-11 infauna: 16-34% Camptonectes ( C .) lens (J. Sowerby) 0-05 0-05 biostratinomic data : Sowerbya triangularis (Phillips) 0-05 0-05 uncemented fauna in life position: 4-63% Entolium corneolum (Young and bivalved fauna preserved with both valves: 22-85% Bird) 0-05 0-05 encrusted: 19-35% of the fauna Pteria sp. 0-05 0-05 bored : 1 T 1 % of the fauna Corbulomima sp. 0-05 0-05 fragmentation: 50-95% FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 345 COR ALLIAN MACROINVERTEBRATE ASSOCIATIONS A. Modiolus biparti tus/ Pleuromya alduini association Description. Eighteen collections with altogether 1798 specimens can be attributed to this association (Table 1). Six bivalves (text-fig. 5) form the trophic nucleus with Nanogyra nana making up nearly 50% of the fauna whilst Lopha genuflecta, Pleuromya alduini, Chlamys (R.) fibrosa , Modiolus bipartitus, and Gryphaea dilatata comprise the rest. Three species (P. alduini, L. genuflecta, M. bipartitus) are characteristic of this association (i.e. they reach in it by far the peak of their distribution). Bivalves account for 98-9% 0 2 0 40 60 % 1 Modiolus bipartitus / Pleuromya alduini ass. text-fig. 5. Trophic nucleus and attempted reconstruction of the Modiolus bipartitus/ Pleuromya alduini association (A). Percentage figures give numerical abundances of the various faunal elements. Length of bar: 4 cm. 1, Nanogyra nana', 2, Lopha genuflecta', 3, Pleuromya alduini ; 4, Chlamys fibrosa ; 5, Modiolus bipartitus', 6, Gryphaea dilatata', 1, Chondrites sp.; 8, Planolites sp. ; 9, pyritic tubes; benthic algae hypothetical. 346 PALAEONTOLOGY, VOLUME 20 of the fauna; the remaining groups are gastropods (0-2%), brachiopods (0-3%), and serpulids (0-5%). Echinoderms are only represented by some cidarid spines. Ammonites (mainly Cardioceras s.l. and Aspido- ceras) are common, belemnites rare. Both these groups have not been included in the statistics as they do not belong to the benthos. The trace-fossil fauna (present in fourteen collections) is unvaried, the abundance is medium, the diversity low. Only three ichnospecies were found, all of them having been created by deposit- feeders: Chondrites , Planolites , and pyritic tubes. The over-all diversity of the benthos is medium (text- fig. 3). Text-fig. 6 shows that more than 60% of the body fauna are epifaunal cemented forms, especially Nano- gyra and Lopha. Deep-burrowing bivalves, mainly Pleuromya and Pholadomya , account for 1 3-6% whilst the semi-infaunal Pinna lanceolata and M. bipartitus form 7-7%. The percentage of epifaunal byssally attached forms is relatively low (9-9%) ; its main representative is Chlamys fibrosa. 4-6% of the uncemented fauna, especially the deep-burrowing Pleuromya and Pholadomya , occur in life position. 22-8% of the bivalved fauna are preserved with both valves; 19-3% of the fauna are encrusted. The epizoans are the foraminifera Nubeculinella (very common), the bivalves Plicatula weymouthiana , Nanogyra nana , and L. gregarea , as well as serpulids of the subgenera Cycloserpula , Dorsoserpula, and Pentaserpula. 11% of the fauna are bored by the bivalve Lithophaga inclusa, polychaetes, and a phoronid (represented by the ichnogenus Talpina (see Voigt 1972)). Fragmentation ranges from 50 to 95%, most collections being near the lower limit. 1 1-7% of the fauna are preserved as steinkerns, especially aragonitic forms like shallow and deep burrowers as well as the ammonites. Gryphaea and Deltoideum are occasionally very worn. In some collections, the oyster fauna is covered with a tlun Fe-hydroxide film, an indication of long exposure on the sea floor. Association A occurs in a narrow substrate and facies range, predominantly in the condensed ferrugin- ous facies consisting of argillaceous micrites or calcareous clays with a varying amount of Fe-ooliths. text-fig. 6. Ecological composition of associations A-C. 1 , epifaunal cemented bivalves ; 2, boring bivalves ; 3, epifaunal byssally attached bivalves; 4, semi-infaunal bivalves; 5, free-resting bivalves; 6, shallow burrowing suspension-feeding bivalves ; 7, deep burrowing suspension-feeding bivalves ; 8, infaunal bivalves deposit-feeding at the depositional interface; 9, infaunal deposit-feeding bivalves; 10, epifaunal brachio- pods; 11, infaunal deposit-feeding echinoids; 12, epifaunal gastropods; 13, infaunal and semi-infaunal gastropods; 14, free-living serpulids. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 347 Discussion. The presence of fauna in life position and the preservation of nearly one- fourth of the fauna with both valves indicate that transport played an insignificant role in the formation of this association. Preservational bias can be ruled out as the aragonitic fauna is still preserved as steinkerns, so that association A can be regarded as the autochthonous part of a true community whose soft parts were removed. Facies, high fossil density, and biostratigraphic evidence point to a very slow accumu- lation of the sediment under predominantly quiet conditions. This, however, did not lead to a mixing of different superimposed communities as is the case with some collections (described below) but the same communities must have persisted for long periods of time. Slow sedimentation rate probably resulted in a slightly consolidated substrate; this might have been responsible for the absence of any deposit-feeders except in the trace-fossil fauna. Shells lying on the surface provided suitable substrate for initial settlement of the oysters which occur usually in clusters, often several generations settling on top of each other. Three feeding levels may be recognized : infaunal deposit-feeders, represented by the trace fossils, low-level suspension-feeders, e.g. shallow and deep burrowers, and Chlamys; and medium-level suspension-feeders, e.g. Pinna. B. Pinna association Description. Six collections with 453 specimens represent this association. P. sandsfootensis is by far the dominant species (75-5%) and, together with Chlamys ( R .) midas , forms the trophic nucleus (text-fig. 7). P. sandsfootensis is confined to this association (see also Arkell 1929-1937) and thus serves as its most characteristic species. C. midas is somewhat less specific as it reaches a second peak in its distribution in text-fig. 7. Trophic nucleus and attempted reconstruction of the Pinna association (B). Length of bar, 8 cm. 1, Pinna sandsfootensis ; 2, Chlamys midas ; 3, Planolites sp. 348 PALAEONTOLOGY, VOLUME 20 the Discomiltha association (F). 99-3% of the fauna are bivalves, the rest gastropods. The only, although abundant, trace fossil occurring with the association is Planolites. The over-all diversity is very low (text- fig. 3). Semi-infaunal bivalves (75-9%; i.e. Pinna , Modiolus bipartitus) dominate the ecological spectrum (text-fig. 6) leaving only 16-8% for the byssally attached epifauna ( Chlamys ). Burrowing forms are rare except Pleuromya uniformis (3-9%) which is the only representative of the deep-burrowing bivalves. The virtual absence of the cemented epifauna so abundant in other associations is also striking. 3-3% of the uncemented fauna (mainly Pinna) are in life position. The number of bivalved animals with both valves preserved is very high: 73T%. The fauna is neither bored nor encrusted. Fragmentation varies between 75 0 and 85 0%. Except for most burrowing forms (mainly Pleuromya ) specimens are preserved with their shell. All six collections come from medium-grained sandstones of the Sandsfoot Grit (Dorset coast) thus showing a clear relationship between substrate/facies and association. Discussion. The six collections have been obtained from different horizons within a massive sandstone body. The close similarity of the samples indicate that the Pinna association persisted for some time during the formation of the sandstone body. Lack of encrusters and borers indicates that shells did not remain ex- posed on the sea floor for long and that the rate of sedimentation was thus fairly continuous. The Pinna association is the autochthonous relic of a former community which is illustrated by the high per- centage of still bivalved specimens and the presence of shells in life position. Numerous flat-lying bi- valved Pinna evidence short periods of stronger currents or wave movements which led to their excavation in situ without further transport. Orien- tation studies on these flat-lying Pinna (text-fig. 8) suggest that either tidal currents or oscillatory wave action were responsible for their excavation. The Pinna association contains two trophic groups: suspension-feeders and deposit-feeders. Three feeding-levels can be distinguished: infaunal deposit-feeders (worms?) repre- sented by Planolites , low-level suspension-feeders (e.g. Pleuromya, Ctenostreon ), and medium-level suspension-feeders {Pinna). C. Neocrassina subdepressa association Description. This association has been found in only one collection with ninety specimens. It is dominated by N. ( N .) subdepressa (text-fig. 9), a species which occurs in other associations only sporadically. The bivalves Nanogyra nana , Gervillella aviculoides, and the gastropod Nerinella are the remaining members of the trophic nucleus. The bivalves Perampliata ampliata (2-2%) and Fimbria umbonata (2-2%) occur only rarely in other associations and are thus also characteristic of association C. Again, bivalves dominate the fauna (94-4%) and brachiopods (11%) form the rest. No trace fossils are present. Shallow-burrowing bivalves ( Neocrassina ) are the most important ecological group (79-8% ; text-fig. 6) whereas byssally attached and cemented epifauna, as well as semi-infaunal bivalves and epifaunal gastropods, share the remaining 20%. No faunal elements are preserved in life position and no bivalves occur with both valves. 1 11% of the fauna are encrusted by Nanogyra nana and Cycloserpula\ borers are missing. Fragmentation is 95%, quite a high value. All specimens, even gastropods, are preserved with their shell; a preservational bias of the fauna can thus be excluded. The Neocrassina association occurs in rubbly, somewhat argillaceous oomicrite. text-fig. 8. Orientation of flat-lying bivalved Pinna in the Pinna associa- tion. Bedding plane in the Sandsfoot Grit, Sandsfoot Castle, near Wey- mouth, Dorset. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 349 Neocrassina subdepressa ass. text-fig. 9. Trophic nucleus and attempted reconstruction of the Neocrassina subdepressa association (C). Length of bar, 6 cm. 1 , Neocrassina subdepressa ; 2, Nanogyra nana ; 3, Nerinella sp. ; 4, Gervillella aviculoides. Discussion. There existed some doubts on the validity of this association. For instance, the absence of bivalved fauna or fauna in life position favours an origin through transportation rather than an in situ accumulation. On the other hand, the specimens of N. subdepressa exhibit a considerable size range and no signs of wear could be detected. Thus, an in situ origin of the association is favoured, assuming that N. sub- depressa lived as a very shallow burrower in the muddy oolite in which it is preserved. Disarticulation and encrustation probably took place after currents excavated the infauna and the shells rested on the sea floor for some time. Association C forms a homogeneous trophic group consisting solely of suspension-feeders. Only two feeding levels can be recognized: the dominating low level suspension-feeders (e.g. Neocrassina) and medium level suspension-feeders (e.g. Gervillella). D. Myophorel/a clave llata association Description. The M. clave data association is represented by eleven collections with 2120 species. The trophic nucleus (text-fig. 10) consists of twelve species: Nanogyra nana is the dominant species, followed by the shallow burrower M. clavellata, the epifaunal byssally attached Chlamys (R.) superfibrosa and C. (Chi.) qualicosta. The remainder is taken up by Trautscholdia morini, Discomiltha rotundata , Cucullaea contracta, Gervillella aviculoides, PUcatula weymouthiana, Procerithium sp., Trautscholdia cf. curvirostra, and T. contejeani. The association is characterized by the occurrence of several species of Trautscholdia, the presence of C. contracta and the abundance of M. clavellata and Chlamys (R.) superfibrosa. The latter has a second peak in its distribution in association G, but occurs there with a different set of species. H 20 40 % Myophorella clavellata ass. text-fig. 10. Trophic nucleus and attempted reconstruction of the Myophorella clavellata association (D). Length of bar, 4 cm. 1, Nanogyra naua ; 2, Myophorella clavellata ; 3, Chlamys superfibrosa\ 4, Chlamys qualicosta; 5, Trautscholdia morini ; 6, Discomiltha rotundata ; 7, Cucullaea contracta; 8, Gervillella avicu- loides ; 9, Plicatula weymouthiana; 10, Procerithium sp.; 11, Trautscholdia curvirostra ; 12, Trautscholdia contejeani ; 13, Spongeliomorpha suevica ; 14, Chondrites sp.; 15, Teichichnus rectus ; benthic algae hypothetical. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 351 Bivalves form 96-9% of the fauna, followed by gastropods (2-7%), and echinoderms (0-4%). Trace fossils show a relatively high diversity (six species) and a medium abundance. The dominant forms are decapod crustacean burrows (Spongeliomorplia suevica var. B), and burrows of various worm-shaped animals {Chondrites, Teichichnus rectus, and Planolites). The over-all diversity is high (text-fig. 3). In the M. clavellata association epifaunal cemented bivalves (50-8%; mainly Nanogyra), byssally attached epifaunal bivalves (1 1-4%; e.g. Chlamys), and shallow burrowers (29 0%; e.g. the bivalves Myophorella , Trautscholdia) form the dominant ecological groups, whilst mucus-tube feeders ( Discomiltha ), semi- infaunal bivalves (Gervillella), and deep-burrowing bivalves (Pholadomya, Pleuromya) are of only little importance (text-fig. 1 1 ). No specimens occur in life position, but 6-6% of the bivalved fauna are preserved text-fig. 1 1. Ecological composition of associations D-G. Legend in text-tig. 6. with both valves. 8-5% of the fauna have been encrusted by Nanogyra, Plicatula, Serpula (Cycloserpula), S. (Pentaserpula) and the foraminifera Nubeculinella. 0-8% of the shells have been bored by cirripeds, polychaetes, and phoronids. Fairly common are browsing traces of echinoderms on Chlamys and other shells. Fragmentation ranges from 50 to 95%, with a mean value of 90%. Shell orientation has been measured in seven collections (Table 2). All but one sample showed a clear dominance of shells in the convex-up position. In two cases imbrication could be found. 1 5-4% of the fauna (mainly shallow- and deep-burrowing bivalves as well as many gastropods) are preserved as steinkerns. In some collections a large number of the specimens are worn or exhibit algal envelopes. The M. clavellata association occurs in Normandy and Dorset, especially in condensed sandy and/or intraclastic sideritic limestones as well as in sandy limestones and marls, thus showing a clear substrate preference. Discussion. The M. clavellata association occurs in two modifications: the collec- tions from Normandy are of low diversity, show considerably better preservation, are less encrusted and bored, and less fragmented, whilst the Dorset collections exhibit a high diversity, often a bad preservation (algal envelopes, signs of wear), and a higher percentage of the fauna is fragmented and bored. In Dorset, high density 352 PALAEONTOLOGY, VOLUME 20 table 2. Shell orientation in the Myophorella clavellata association (D). Collection Convex-up Convex-down Oblique BR 44 100 15 19 BR 47 100 33 47 BH 24 100 25 33 BH 28 top 100 38 21 SC 16 100 28 79* HVI 15 18 54 58* HV II 5 71 11 8 * Shells partially imbricated. of the fauna, bad preservation, and the mixed and sideritic substrate point to a slow accumulation resulting in, at least partially, condensed deposits. Extensive lateral transport of the fauna seems unlikely as no size sorting exists and some shells are excellently preserved and sometimes bivalved. A parautochthonous accumulation caused by excavation and short, small-scale transport of the shells by currents seems a more feasible model for the formation of this association. The Myophorella association of Dorset cannot therefore be regarded as the remnant of one ancient community but components from several neighbouring communities probably contributed to its composition. Very likely not only horizontal but also vertical mixing took place: the high diversity of the fauna (indicated especially by the large number of species in the trophic nucleus) suggests that a change in community took place once or several times during the accumulation of the beds. Despite the fact that faunal mixing played a certain role in the collections from Dorset, they never- theless represent a very distinct association. Only the Neocrassina subdepressa association (C) exhibits a similar ecological structure (compare text-figs. 6 and 11), but there the infauna consists of different elements ( Neocrassina instead of Myo- phorella and Trautscholdia) and the cemented epifauna plays a far lesser role. That a Myophorella community existed at some time in the Upper Jurassic of north-west Europe, can also be demonstrated with most of the collections from Normandy which also show signs of short-stage transport but occur in beds which indicate normal or even rapid sedimentation. Two trophic groups are present in the Myophorella association: 96-7% of the fauna are suspension-feeders, whilst 3-3% are deposit-feeders (mucus-tube-feeding Discomiltha and Procerithium). Four feeding levels can be distinguished: the very rare Palaeonueula as well as the trace fossils Teichichnus, Cylindrichnus, and Chon- drites representing the infaunal deposit-feeders, Discomiltha and Procerithium deposit-feeders at the depositional interface, the common Myophorella and Traut- scholdia the low level suspension-feeders, and Gervillella the medium level suspension- feeders. E. Nucleolites scutatus association Description. Seven hundred and eighty-six specimens in six collections represent this association. The trophic nucleus consists of seven species (text-fig. 12), the most important being the echinoid N. scutatus , followed by Nanogyra nana, Pleuromya uniformis , Meleagrinella ovalis , Sowerbya triangularis , Chlamys 0 FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 353 text-fig. 12. Trophic nucleus and attempted reconstruction of the Nucleolites scutatus association (E). Length of bar, 4 cm. 1, Nucleolites scutatus', 2, Nanogyra nana ; 3, Pleuromya uniformis ; 4, Meleagrinella ovalis; 5, Sowerbya triangularis', 6, Chlamys fibrosa', 1, Meleagrinella laevis ; 8, Teichichnus rectus', 9, Spongeliomorpha suevica\ benthic algae and hydrozoan/anthozoan coelenterate hypothetical. (. R •) fibrosa, and M. laevis. The association is characterized by Nucleolites scutatus and M. ovalis ; in addi- tion, S. triangularis reaches here the peak of its distribution. Bivalves (58-9%) no longer form the only important group of the fauna, but echinoderms (38-5%) are of nearly equal importance. Gastropods (2-5%) only play a minor role. The N. scutatus association is dominated by infaunal elements (text-fig. 11); the echinoids Nucleolites, Clypeus, and Holectypus, shallow-burrowing bivalves like Sowerbya and Aniso- cardia, deep-burrowing Pleuromya, and infaunal gastropods like Pseudomelania constitute 62-8% of the fauna. Byssally attached epifaunal elements (34-3%) are not so much represented by the genus Chlamys (as is the case in most other associations) but by the free-swinging pendant Meleagrinella. The number of epifaunal cemented forms is relatively low (14-4%). Trace fossils are of medium abundance and diversity; the ichnospecies present are: Spongeliomorpha suevica var. B, Teichichnus rectus, Skolithos sp., and the resting trace of an anthotoan ( Bergaueria sp.), i.e. a mixture between deposit-feeders, suspension-feeders, domichnia, fodinichnia, and cubichnia. The over-all diversity is medium (text-fig. 3). Only 0-6% of the uncemented fauna occur in life position (mainly Homomya, Pleuromya), but 22-5% of the bivalved fauna are preserved with both valves (amongst them epifaunal forms like Meleagrinella). Dorsoserpula, Cyclo- serpula , Nanogyra, and the bryozoan Berenicea encrust 3-7% of the fauna; boring elements are not present. 354 PALAEONTOLOGY, VOLUME 20 Fragmentation ranges from 0 to 95%, the mean value is around 60%. Except for 9-9% (mainly deep- burrowing bivalves and some gastropods which occur as steinkerns) the specimens are preserved with their shell. In two collections, a part of the epifauna exhibits algal envelopes; otherwise no signs of wear have been found. The Nucleolites scutatus association shows a substrate preference for limestones which may, however, range from pure micritic limestones to oolites. Sandy, somewhat shelly limestones dominate. Discussion. The N. scutatus association can be regarded as the autochthonous relic of an ancient community. This is stressed by the high percentage of bivalves preserved with both valves. That no fauna in life position has been found might be due to the fact that in the case of the burrowing echinoderm Nucleolites life position and stable position are identical (thus quite a few of the Nucleolites might be preserved in life position). The two collections, in which part of the epifauna is encrusted by calcareous algae and other organisms, have been found in shelly limestone and poorly sorted oolite. There, a relatively high energy level probably led to discontinuous sedimenta- tion and in situ reworking which remained restricted to the epifauna. The other collections represent quieter environments with continuous sedimentation which is indicated by the excellent preservation of the fauna (in addition, many Nucleolites are filled with calcite and not with sediment). Both, deposit-feeders ( Nucleolites ) and suspension-feeders (the remaining fauna), are present in the association and lead to an inhomogeneous trophic grouping. Four feeding levels may be recognized: infaunal deposit-feeders ( Nucleolites , Teichichnus ), low-level suspension-feeders (e.g. Pleuromya , Anisocardia , Nanogyra), medium-level suspension-feeders ( Gervillella ), and high-level suspension-feeders (pendent Meleagrinella). The low number of epifaunal encrusting bivalves is clearly related to the low percentage of other epifauna which restricted the substrate suitable for colonization. F. Discomiltha association Description. The Discomiltha association is only represented by two collections with 702 specimens. The trophic nucleus (text-fig. 13) contains two species of Discomiltha, D. lirata, and D. rotundata , the remaining species being Chlamys (R.) midas, Meleagrinella laevis, Anisocardia isocar dioides, Nanogyra nana, and Pleuromya uniformis. Four of these species are characteristic of the association: D. lirata, D. rotundata, A. isocardioides, and M. laevis. C. midas reaches a second peak in its distribution in the Pinna association and is thus less specific. As usual, bivalves dominate the fauna (99-6%) with gastropods forming the rest. Ammonites (not discussed here) are a very common faunal element. Trace fossils are of low diversity but high abundance and consist of the crustacean burrows Spongeliomorpha suevica var. B, Sp. paradoxica, and of Teichichnus. The over-all diversity is relatively low (text-fig. 3). Ecologically, mucus-tube feeders {Discomiltha) and epifaunal byssally attached bivalves (e.g. Chlamys, Meleagrinella) are of equal impor- tance (31 -8% and 310% respectively; see text-fig. 11). Shallow burrowers (81%; e.g. Pleuromya) and epifaunal cemented forms (9-2%; Nanogyra) comprise the rest. No specimens were found in life position, but 25-5% of the bivalves (predominantly Discomiltha and Pleuromya) are preserved with both valves. 1-4% of the fauna are encrusted by Cycloserpula', no specimens are bored. Fragmentation is in both collections about 50%. Nearly four-fifths of the specimens, mainly the burrowing bivalve fauna and the ammonites, are preserved as steinkerns. Both collections are from medium-grained sandstone which indicates the narrow facies and substrate range of association F. Discussion. The Discomiltha association is the relic of an ancient community which is indicated by the high percentage of specimens preserved with both valves and the low percentage of fragmented forms. The extensive burrow systems of crustaceans ( Spongeliomorpha ) point to fairly stable conditions at least for part of the time. FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 355 Their producers (deposit-/suspension-feeders or scavengers?) constituted an impor- tant part of the community. The absence of infaunal deposit-feeding bivalves is probably due to the relatively coarse substrate, whilst the encrusting fauna lacked substrate suitable for colonization and is thus far below its usual significance. No other association possesses a similar ecological structure. The occurrence of two species of Discomiltha among three most common representatives of the trophic nucleus indicates that interspecific competition must have been low, or else, that small differences in the feeding habit existed. Five feeding levels can be distinguished : infaunal deposit-feeders (represented by Teichichnus ), deposit-feeders at the deposi- tional interface ( Discomiltha ), low-level suspension-feeders (e.g. Chlamys , Pleuromya , Anisocardia), medium-level suspension-feeders, represented by some Pinna and text-fig. 13. Trophic nucleus and attempted reconstruction of the Discomiltha association (F). Length of bar, 3 cm. 1, Discomiltha lirata; 2, Chlamys midas; 3, Discomiltha rotundata ; 4, Meleagrinella laevis ; 5, Anisocardia isocardioides ; 6, Nanogyra nana; 7, Pleuromya um for mis ; 8, Spongeliomorpha suevica\ 9, Teichichnus rectus', benthic algae and hydrozoan/anthozoan coelenterate hypothetical. 356 PALAEONTOLOGY, VOLUME 20 Gervillella, and high-level suspension-feeders (e.g. the pendent Meleagrinella). The numerous vagile ammonites (mainly Amoeboceras , less common perisphinctids) most likely formed a third trophic group, namely scavengers (Lehmann 1975) or predators. G. Chlamys/ Nanogyra nana association Description. Eighteen collections with 2908 specimens have been grouped into this association. N. nana is with 521% the dominant species (text-fig. 14) followed by C. (R.) fibrosa , C. ( R .) superfibrosa, and 0 20 4£ §0% text-fig. 14. Trophic nucleus and attempted reconstruction of the Chlamys/ Nanogyra nana association (G). Length of bar, 4 cm. 1, Nanogyra nana; 2, Chlamys fibrosa; 3, Chlamys superfibrosa; 4, Chlamys qualicosta; 5, Nucleolites scutatus; 6, Spongeliomorpha suevica; 7, Cylindrichnus concentricus; 8, Planolites sp.; benthic algae hypothetical. C.(C.) qualicosta. The burrowing echinoid Nucleolites scutatus and the shallow-burrowing bivalve Sowerbya triangularis complete the trophic nucleus. The association is characterized by the dominance of pectinids and of Nanogyra. Bivalves comprise 93-8% of the fauna, gastropods 3-2%, echinoderms T7%, brachiopods 0-3%, and free-living serpulids 0-2%. The epifauna (861%; text-fig. 11) consists mainly of cemented Nano- gyra and byssally attached pectinids, the infauna ( 12-6%) predominantly of shallow burrowers like Sowerbya , Trautscholdia , and Isocyprina. Infaunal gastropods, Nucleolites, and deep-burrowing bivalves are of lesser significance as are semi-infaunal animals ( Gervillella , Pinna). For the first time, deposit- feeding nuculids (11%) occur in an association more than just as scattered individuals. Trace fossils show a very high diversity (eight species) and medium abundance. The dominant ichnospecies are Spongelio- morpha suevica var. B, Planolites sp., and the burrow of a worm-shaped animal, Cylindrichnus concentricus. The over-all diversity is high (text-fig. 3). No species occur in life position except Nanogyra and other FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 357 cemented forms and the percentage of specimens preserved with both valves is low (2-9%). 5% of the fauna are encrusted by various serpulids ( Cycloserpula , Dorsoserpula, Pentaserpula), the bivalves Nanogyra and Placunopsis , bryozoans, and the foraminifera Nubeculinella. Lithophaga inclusa and cirripedes account for the few bored specimens (01%). Fragmentation varies between 30 and 98% and is usually around 90%. Orientation of shells could be measured in one collection ; convex-up shells dominated (convex-up : convex- down = 100:65; oblique: 41). 5-3% of the fauna, predominantly burrowing forms ( Pleuromya , Palaeo- nucula , Protocardia) more rarely epifaunal elements, are preserved as steinkerns. Only rarely do specimens exhibit signs of wear but commonly show algal envelopes when found in oolites or bioclastic limestones. The Chlamys/N. nana association occurs in a wide range of substrates and facies types, in clays as well as in oolites whereby a certain preference for sandstones and impure limestones was noted. In several collections bored and encrusted pebbles are associated with the fauna. Discussion. The Chlamys/N. nana association is one of the commonest associations in the Corallian, found from Normandy all the way to the Yorkshire coast. Although its characteristic members can also be found in nearly all other associations, it is here that they reach the peak of their distribution ( Nanogyra excepted). The environ- ment in which this association accumulated was quite varied, judging from its wide range of facies, but a fairly high energy level is probable for most samples. Reworking was a common feature (encrusted and bored pebbles, high fragmentation rate, algal envelopes) and some species like Lopha genuflecta reacted in developing occasionally extremely thick and heavy shells. The percentage of still bivalved specimens is insignificant and no forms are preserved in life position. The substrate was of little attraction for the infauna as the cemented and byssally attached epifauna is the dominating faunal element. Nanogyra growing on dead shells formed little clusters on the sea floor and the pectinids similarly may have used shell material for fixation of their byssus. Apart from this high-energy variety, a few collections seem to be derived from a low-energy environment represented by clays and silts. There, the fauna is better preserved and the infauna is of greater abundance and diversity with some deposit- feeding Procerithium , Discomiltha , and Palaeonucula occurring besides infaunal suspension-feeders ( Pseudomelania ). The Chlamys/N. nana association has a similar ecological composition to the oyster /Isognomon promytiloides association (compare text-figs. 11 and 21) but there Chlamys is replaced by Isognomon and Nanogyra has to compete with Plicatula and other oysters. Two trophic groups (deposit-feeders, suspension-feeders) are present in association G and three feeding levels can be recognized: infaunal deposit-feeders ( Palaeonucula ), low-level suspension-feeders ( Chlamys , Pleuromya , Sowerbya), and medium-level suspension-feeders ( Gerville/la , Pinna). H. Thurmanella acuticost a/ Nanogyra nana association Description. Three collections with 921 specimens form this association which has been encountered only in Yorkshire. The trophic nucleus (text-fig. 15) consists of two species, the bivalve N. nana and the brachio- pod Thurmanella acuticosta. Gervillella aviculoides and a species of 'Terebr alula are two further common faunal elements. The association is characterized by the high percentage of rhynchonellid or terebratulid brachiopods which otherwise are a rare faunal element. Hardly any infauna is present (0-9%), but epifaunal cemented bivalves (52-5%), pedicle-attached brachiopods (38-5%) and, to lesser extent, epifaunal byssally attached bivalves (2-0% ; mainly Chlamys) prevail (text-fig. 16). With 5-8%, the semi-faunal bivalves ( Gervillella , Pinna) are relatively well represented. Long stems of the crinoid Millericrinus are commonly found on the bedding planes; starfish ossicles abound in the sediment. Trace fossils ( Spongeliomorpha suevica var. B and Skolithos) are of low diversity and medium abundance. The over-all diversity is similarly i ,Ap ,Dl te 4 H,0! ,lc r 358 PALAEONTOLOGY, VOLUME 20 0 20 40 others 60 y. Thurmanelia / Nanogyra nana ass. text-fig. 15. Trophic nucleus and attempted reconstruction of the Thurmanelia acuticosta/ Nanogyra nana association (H). Length of bar, 2 cm. 1, Nanogyra nana ', 2, Thurmanelia sp.; 3, Spongeliomorpha suevica ; 4, Skolithos sp. 60 % % % 1 Thurmanelia/ Nanogyra nana ass. Lopha gregarea ass. Gervillella aviculoides ass. text-fig. 16. Ecological composition of associations H-K. Legend in text-fig. 6. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 359 very low (text-fig. 3). Of the uncemented fauna 57-6%, mainly Thurmanella and ‘ Terebratula occur in life position as do most cemented Nanogyra , and 39% of the specimens are still bivalved. Only a few specimens are encrusted with the bryozoan Berenicea or with Nanogyra , even less are bored by polychaetes. Fragmentation ranges from 65 to 75%. All faunal elements are preserved with their shells. The three collections show a distinct substrate preference occurring in oolitic/sandy limestones and oolitic calcareous sandstones. i Discussion. As only thick-shelled aragonitic burrowers are preserved ( Myophorella ), some preservational bias cannot be excluded, with the thin-shelled species dissolved during diagenesis. Apart from this, the Thurmanella/ Nanogyra association is the remnant of an ancient community. This can be inferred from the high percentage of fossils still bivalved and in life position. Hardly any transport has been involved: Nanogyra and the brachiopods are found as small clusters with calcite or geopetal fills dispersed on the bedding plane as they were originally on the sea floor. Partly articulated Millericrinus also indicate that the fauna is autochthonous. The low percentage of bored and encrusted shells indicates continuous sedimentation; the substrate and faunal composition suggest moderate currents. Deposit-feeders are not represented, but the abundant starfish ossicles indicate that algal grazing, scaveng- ing, or predatory starfish played an important role in the ecosystem. Three feeding levels can be distinguished: infaunal deposit-feeders (crustaceans) low-level suspension-feeders (e.g. the producer (worms, phoronids?) of Skolithos , Myophorella , Nanogyra , brachiopods), medium-level suspension-feeders ( Gervillella , Pinna), and high-level suspension-feeders ( Millericrinus ). J. Lopha gregarea association Description. The L. gregarea association is represented by 616 specimens in six collections. The trophic nucleus (text-fig. 17) consists of two epifaunal cemented bivalves: L. gregarea and N. nana. The association text-fig. 17. Trophic nucleus and attempted reconstruction of the Lopha gregarea association (J). Length of bar, 4 cm. 1, Lopha gregarea', 2, Nanogyra nana\ 3, Planolites sp. ; 4, Chondrites sp. 360 PALAEONTOLOGY, VOLUME 20 is characterized by L. gregarea which reaches here the peak of its distribution. Bivalves (98-5%) dominate the association with serpulids (0-8%) and brachiopods (0-6%) forming the rest, whilst gastropods are missing. Trace fossils are of low abundance and diversity; they are represented by Chondrites and pyritic tubes. The over-all diversity is very low (text-fig. 3). Except for a few epifaunal byssally attached bivalves (1-0%) and deep-burrowing bivalves (1-3%) the fauna is dominated by cemented bivalves (95-3%; see text-hg. 16). No uncemented forms are preserved in life position but the majority of Lopha and Nanogyra occur in life position forming low biostromes. 11-6% of the specimens are still bivalved. Due to the pre- dominance of epifauna, 57-6% of the fauna are encrusted by Nanogyra , the foraminifera Nubeculinella, the bryozoans Berenicea and Stomatopora , and various serpulids ( Cycloserpula , Dorsoserpula , and Penta- serpula). 4% of the fauna are bored ; the animals responsible being the bivalve Lithophaga inclusa and a phoronid. Fragmentation varies between 75 and 90%. Most specimens are preserved with their shell intact except some deep-burrowing Pholadomya which occur in steinkern preservation. Discussion. Only present in Normandy, the Lopha gregarea association exhibits a similar ecological composition as the Nanogyra association (R), but there Lopha is replaced by Nanogyra which is the only dominant faunal element. The environment is very similar to that of association A (see p. 347), i.e. relatively low rates of sedi- mentation and only interludes of high-energy conditions. This led to a stable sub- strate upon which extensive oyster beds grew up to 20 cm high. These beds could be followed laterally for several hundred metres. The dense mass of L. gregarea pro- vided in turn an ecological niche for smaller encrusters like Nanogyra and the numerous serpulids. Whilst some collections represent the in situ L. gregarea com- munity of Upper Jurassic times, in some others the shell beds have been broken up and clusters of individuals scattered on the sea floor, indicating the destruction of the community by strong currents. The fact that a large part of the sea floor was covered with these oyster beds is probably responsible for the extremely low percentage of burrowing fauna. Only two feeding levels can be distinguished ; infaunal deposit-feeders (sipunculoid worms?) represented by Chondrites and pyritic tubes, and low-level suspension- feeders (rare burrowing bivalves, Lopha , Nanogyra). K. Gervillella aviculoides association Description. Three hundred and fifty-two specimens in three collections form the G. aviculoides association. Its trophic nucleus (text-fig. 18) consists of the bivalves N. nana and semi-infaunal G. aviculoides. The characteristic species is Gervillella which reaches here the peak of its distribution. As usual, bivalves domi- nate the fauna (94-6%) followed by gastropods (4-3%), brachiopods (0-6%), and free-living serpulids (0-6%). The fossils are of medium diversity and abundance; the dominant ichnogenera comprise Chondrites , Spongeliomorpha suevica var. B, and Planolites. The over-all diversity is low (text-fig. 3). Semi-infaunal bivalves (41 -8%) and cemented epifauna (46-3%; e.g. Nanogyra ) are of nearly equal importance whilst the byssally attached epifauna (51%; e.g. Chlamys), epifaunal gastropods (2-3% ; e.g. Ampullina), and infaunal gastropods (2-0% ; Pseudomelania) are of lesser significance (text-fig. 16). Infaunal bivalves represented solely by Isocyprina and Cucullaea are very rare. No specimens are preserved in life position, but 27-1% are still bivalved. 7-4% of the fauna is encrusted by Nanogyra and Pentaserpula , bored specimens were not encountered. Fragmentation ranges from 50 to 75%. All specimens are preserved with their shell. The G. aviculoides association occurs in Yorkshire and Oxfordshire and shows a substrate preference for oolites and impure limestones. Discussion. Calcite preservation of the fauna and absence of any aragonitic burrowers point to a preservational bias of the association but as all aragonitic gastropods are preserved with their shell the absence of the burrowing bivalves seems to be genuine. Thus, the coarse substrate might have been unsuitable for burrowers. The three FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 361 text-fig. 18. Trophic nucleus and attempted reconstruction of the Gervillella aviculoides association (K). 1, Nanogyra nana ; 2, Gervillella aviculoides', 3, Chondrites sp. ; 4, Spongeliomorpha suevica ; 5, Planolites sp. collections represent different stages of reworking of an ancient G. aviculoides community, ranging from a sample from the Tabular Hills (Yorkshire) where nearly all Gervillella are preserved with both valves to a sample from Shellingford Cross Roads Quarry, Berkshire, where most specimens are one-valved and show cur- rent orientation (text-fig. 19). But even there, the transport involved must have been minimal as the Gervillella range from 3-5 to 15 0 cm in size and are well pre- served. Thus, the G. aviculoides associa- tion can be regarded as the partially reworked autochthonous relic of an ancient community which seems to have thrived in a relatively high-energy environ- ment. Three feeding levels can be dis- tinguished: infaunal deposit-feeders rep- resented by Planolites and Chondrites , low-level suspension-feeders (e.g. Iso- cyprina , Nanogyra ), and medium-level suspension-feeders ( Gervillella , large Plagio stoma). text-fig. 19. Orientation of flat-lying Gervillella aviculoides in the Gervillella aviculoides associa- tion. Bedding planes of fallen blocks, Osmington Oolite Group, Shellingford Cross Roads Quarry, Berkshire. 362 PALAEONTOLOGY, VOLUME 20 L. Pleuromya uniformis association Description. The P. uniformis association is with twenty-eight collections and 3272 specimens the commonest of the Corallian invertebrate associations, occurring predominantly in Dorset and, to a lesser extent, in Yorkshire. The trophic nucleus (text-fig. 20) consists of eight species of bivalves. P. uniformis occupies the first-rank position being followed by N. nana, "My Ulus' varians, Myophorella clavellata, Chlamys 0 20 40 •/„ text-fig. 20. Trophic nucleus and attempted reconstruction of the Pleuromya uniformis association (L). Length of bar, 5 cm. 1, Pleuromya uniformis ; 2, Nanogyra nana; 3, ‘ Mytilus ' varians; 4, Myophorella clavellata; 5, Chlamys fibrosa; 6, Gryphaea dilatata; 7, Deltoideum delta; 8, Chlamys midas; 9, Sponge- liomorpha suevica; 10, Teichichnus rectus; 1 1, Chondrites sp. ; benthic algae hypothetical. FURSICH : COR ALLI AN BENTHIC ASSOCIATIONS 363 fibrosa , Gryphaea ( B .) dilatata , Deltoideum delta and C. (R.) midas. Two species, P. uniformis and 'M varians, characterize the association reaching in it the peak of their distributions. As usual bivalves domi- nate the fauna with 961%; gastropods (2-6%), echinoderms (10%), brachiopods, serpulids, and sponges form the rest. The trace-fossil fauna is both very diverse and abundant; the more important members are Spongeliomorpha suevica var. B, Chondrites, Teichichnus rectus, and Cylindrichnus concentricus. The over-all diversity is very high (text-fig. 3). Epifauna and infauna are nearly equal in numbers (text- fig. 21), the semi-fauna represented by Modiolus, Pinna, and Gervillella being only of little importance (11%). Deep-burrowing bivalves like Pleuromya and Pholadomya (40-9%) dominate the infauna, shallow- burrowing bivalves like Myophorella, Isocyprina, and astartids account for 10-5% of the fauna; the rest is formed by the echinoid Nucleolites , the gastropod Pseudomelania, and some mucus-tube feeding bivalves (Discomiltha). In the epifauna, cemented bivalves (mainly Nanogyra) and byssally attached pectinids ( Chlamys , Camptonectes) and Isognomon prevail. Less common are free-resting ( Gryphaea ) or swimming forms (Entolium) and some herbivorous or scavenging gastropods. Only 2-8% of the fauna occur in life position but 48-7% of the specimens are still preserved with both valves. The number of encrusted and bored specimens is fairly low. Epizoans are the foraminifera Nubeculinella and the bivalves Nanogyra, Lopha, ‘ Ostrea ’, and Placunopsis. Lithophaga inclusa and phoronids account for the borings. Fragmenta- tion ranges from 5% to more than 95% and is usually at one end of the scale. In two collections a preferred convex-up orientation of shells larger than 2 cm has been noted, the percentage of shells in an oblique position being relatively high. A large part (39-6%) of the fauna is preserved as steinkerns, mainly the burrowing bivalves Pleuromya, Pholadomya, and Myophorella as well as many gastropods (especially Ampullina and pleurotomariids). Only rarely do specimens show signs of wear or exhibit algal envelopes. The Pleuromya uniformis association occurs in a very wide range of facies, from oolites to clays, but shows a preference for intraclastic, sandy limestones and fine- to medium-grained sandstones. Occasionally drift-wood and in one case bored and encrusted pebbles are associated with the fauna. Discussion. Judging from its widespread occurrence the P. uniformis association thrived in a variety of environments, and seems to have dominated the offshore shelf regions where moderate currents and a somewhat reduced rate of sedimentation L M N O text-fig. 21. Ecological composition of associations L-O. Legend in text-fig. 6. 364 PALAEONTOLOGY, VOLUME 20 favoured deep burrowing, as well as the even more rigorous shallow subtidal bar and nearshore environments. In some collections the composition of the fauna indicates a gradation into the condensed version of the Myophorella association, and the high density of the fauna, and shell beds which accumulated rather by non- deposition and only rarely by current action confirm this observation. Other collec- tions represent a fairly true picture of the ancient Pleuromya community: Pleuromya and clusters of ‘My t Hus' varians (which most likely lived like the present-day M. edulis) in life position indicate that hardly any disturbance has taken place. In general the high percentage of bivalved specimens, the concentration of the fauna in small pockets, and the good preservation of most specimens demonstrate that transport played only a very subordinate role in the formation of the association. The low O 20 40 % text-fig. 22. Trophic nucleus and attempted reconstruction of the Pseudomelania heddingtonensis associa- tion (M). Length of bar, 4 cm. 1, Nanogyra nana\ 2, Pseudomelania heddingtonensis ; 3, Nerinella sp.; 4, Chlamys fibrosa ; 5, Nucleolites scutatus ; 6, Pleuromya uniformis; 7, Nerinella cyane ; 8, sponge Rhaxella (hypothetical reconstruction); 9, Spongeliomorpha suevica ; 10, Teichichnus rectus', 11, Cylindrichnus concentricus; 12, Chondrites sp. ; benthic algae hypothetical. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 365 percentage of specimens in life position and, at the same time, the high percentage of bivalved fauna is in many collections probably due to the very intensive bioturba- tion, especially of burrowing decapods. Three or four trophic groups are represented in the association: deposit-feeders, suspension-feeders, herbivores (e.g. Bathrotomaria), and possibly, scavengers (Ampullinal). Two feeding levels can be recognized: infaunal deposit-feeders indi- cated by Planolites , Cylindrichnus, Chondrites, and Teichichnus, deposit-feeders at the depositional interface (the rare Discomi/tha) and low-level suspension-feeders (e.g. Pleuromya, Myophorella, pectinids). M. P seudomelania heddingtonensis association Description. The P. heddingtonensis association is, with sixteen collections and 2194 specimens, fairly common in the Coralhan. The trophic nucleus consists of seven species (text-fig. 22). N. nana occupies the first rank position followed by two gastropod species, P. heddingtonensis and Nerinella sp. The pectinid Ch/amys (R.) fibrosa , the echinoid Nucleoli tes scutatus, the bivalve Pleuromya uniformis and the gastro- pod Nerinella cyane form the rest. Pseudomelania heddingtonensis reaches in this association the peak of its distribution and thus forms its characteristic species. Bivalves account for only 56-6% of the fauna and gastropods (39-7%) are a very important faunal element. Echinoderms (3-5%), brachiopods (01%), and sponges (0-04%) are of lesser significance. Trace fossils are of a high diversity and only medium abun- dance. The dominant forms are Spongeliomorpha suevica var. B, Chondrites sp., Teichichnus rectus, and Cylindrichnus concentricus. The over-all diversity is high (text-fig. 3). Infauna and epifauna are of nearly equal importance, the semi-fauna being hardly represented (0-6%). Amongst the epifauna (text-fig. 21), cemented bivalves form the largest group (35-5%; mainly Nanogyra), followed by gastropods like Nerinella, Bathrotomaria, Coelostylina , Procerithium, and Ampullina (13- 1%), and by byssally attached bivalves (8-4%; mainly pectinids and limids). The infauna is dominated by the gastropod Pseudomelania. The rest consists of deep-burrowing (e.g. Anisocardia, Isocyprina ) bivalves as well as echinoids ( Nucleoli tes , Pygaster ). No elements of the fauna have been found in life position except cemented bivalves; 9% of the specimens are still preserved with both valves. Nanogyra, Dorsoserpula, and the foraminifera Nube- culinella encrust 6% of the fauna; 0-2% are bored by the bivalve Lithophaga inclusa and by polychaetes. Fragmentation ranges from 20 to 98% and is usually around 90%. The thin-shelled organisms are pre- ferredly fragmented whilst the thick shells of, for instance, Gervillella and Pseudomelania have been less affected. In seven collections orientation of shells larger than 2 cm has been measured (Table 3). The table 3. Shell orientation in the Pseudomelania heddingtonensis association. Only shells with a diameter larger than 2 cm have been counted. The asterisk marks collections in which imbrica- tion has been encountered. Collection Convex-up Convex-down Obliq PII lb 100 12 9 *TB II 10 top 100 20 23 TB II 10 base 100 29 19 *TB 11 14 100 19 28 *TB II 17 100 40 30 *TB II 18 50 13 33 overwhelming majority of the shells occurs convex-up and in four collections imbrication has been observed. 7-2% of the fauna occurs in steinkern preservation. Amongst them are many deep-burrowing bivalves, some shallow burrowers, and most gastropods except Pseudomelania. Occasionally shells exhibit algal envelopes, more rarely are they worn. Discussion. The P. heddingtonensis association is widespread in the Malton Oolite of the Vale of Pickering, Yorkshire, but is also found in Normandy and Dorset. i 366 PALAEONTOLOGY, VOLUME 20 Substrate, algal envelopes, low percentage of bivalved specimens, high fragmenta- tion rate, convex-up orientation, and imbrication of shells testify a high-energy environment for most samples. Lateral transport must have played a role, so that sorting of the fauna certainly took place. Nevertheless a large-scale lateral transport can be excluded, as in some samples Pseudomelania are preserved with their calcareous operculum still in place. As many of the oolites are micritic and alternate with thin bands of marly oolite, stable, less rigorous conditions must have prevailed at times which enabled the colonization of the substrate. With return of the high-energy regime, the fauna was then frequently washed out and redeposited under current action. The existence of an ancient Pseudomelania community is further supported by a few samples from Dorset where a finer-grained sediment (probably stabilized by the sponge Rhaxella whose spicules occur abundantly in the substrate) and a higher percentage of bivalved specimens, together with a low fragmentation rate (20%), indicate a low-energy environment with only little disturbance of the fauna. Thin-shelled burrowing bivalves ( Pholadomya , Pleuromya) are quite common in these samples and indicate that in the collections from the oolites these shells were eliminated during reworking. The reconstruction of the life habits of fossil gastro- pods is conjectural. The pleurotomariids Bathrotomaria, Obornella, and Pleuro- tomaria might have been herbivores. Vogel (1968) argued that nerineids belonged to the liberosessile epifauna due to their heavy shell and the manner in which a rudist encrusted the shell of one species. Nerinella and other nerineids are usually found in high-energy environments often in association with reefs (e.g. Janicke 1970; see also association Q) which strengthens Vogel’s argument. Ampullina probably belonged to the epifauna as well feeding on algae or plant detritus, scaveng- ing or even predating. The high-spired, thick-shelled Pseudomelania in turn could well have lived as a very shallow burrower feeding by ciliary activity, whilst the species of Procerithium present was probably a deposit-feeder. The large number and diversity of gastropods in the Pseudomelania association accounts for the large number of trophic groups present; deposit-feeders, suspension-feeders, herbivores ( Bathrotomaria ), and possibly scavengers ( Ampullina ?). Two feeding levels can be distinguished; infaunal deposit-feeders represented by Nucleolites, Teichichnus, Cylindrichnus , Planolites, and Chondrites , and low-level suspension-feeders (e.g. Pseudomelania , Pleuromya). N. Corbulomima association Description. Five collections with 836 specimens constitute the Corbulomima association. The very diverse trophic nucleus (text-fig. 23) contains thirteen species starting with Corbulomima sp. A, Procerithium (R.) struckmanni, C. suprajurensis , Tancredia subplanata , Nanogyra nana , and Palaeonucula menkii. The remain- ing seven species ( Protocardia dyonisea, Thracia depressa , Placunopsis duriuscula, Trautscholdia extensa , Discomiltha rotundata, Cercomyopsis striata , and Mesosacella sp.) are of lesser significance (altogether 15-2% of the fauna). Four bivalve species are characteristic of this association, namely Corbulomima sp. A, C. suprajurensis , Tancredia subplanata, and Palaeonucula menkii, and one gastropod, Procerithium struck- manni. Bivalves (84- 7%) and gastropods (141%) dominate the association; brachiopods (0-6%), echino- derms (0-4%), and decapods (0-2%) making up the remainder. Trace fossils are present in two of the five collections and consist of Chondrites sp., Planolites sp., Teichichnus rectus, and Spongeliomorpha suevica var. B. Abundance and diversity of the traces are both medium. The over-all diversity of the benthos is medium (text-fig. 3). Some pectinids, arcids, free-swinging Pteroperna, and Procerithium compose the epifauna (31-9%, FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 367 text-fig. 21); the association is dominated by the infauna (66-5%), semi-infaunal bivalves {Pinna) being fairly insignificant (T6%). The diverse infauna is represented by shallow-burrowing suspension-feeders (48-8%) like Corbulomima , Tancredia, and astartids, deposit-feeding bivalves (9-6%; Mecosacella , Palaeo- nucula), and less commonly by deep-burrowing suspension-feeders (e.g. Pleuromya, Cercomya), deposit- feeding gastropods (Dicroloma), mucus-tube feeders ( Discomiltha ), echinoids ( Nucleolites ), or brachiopods (e.g. Lingula). No specimens occur in life position and only 6-3% are still bivalved. Encrusted and bored 0 20 7. 1 i I Corbulomima ass. text-fig. 23. Trophic nucleus and attempted reconstruction of the Corbulomima association (N). 1, Corbulomima sp. A; 2 , Procerithium struckmanni ; 3, Corbulomima suprajurensis', 4, Tancredia subplanata; 5, Nanogyra nan a ; 6, Palaeonucula menkii ; 7, Protocardia dyonisea ; 8, Thracia depressa ; 9, Placunopsis duriuscula; 10, Trautscholdia extensa ; 11, Discomiltha rotundata; 12, Ceratomyopsis striata ; 13, Meso- sacella sp. ; 14, Planolites sp. 368 PALAEONTOLOGY, VOLUME 20 specimens are rare (0-6% and 11% respectively); the borers are represented by phoronids and another, unknown organism; the epizoans by Placunopsis, Nanogyra, and Serpula. Fragmentation varies between 50 and 95% and is usually around 80%. Part of the infauna, especially Corbulomima , Pleuromya, and Discomiltha, as well as some gastropods ( Procerithium ) are preserved as steinkerns. The Corbulomima association occurs in a narrow facies range, i.e. in clays, silts, and fine-grained sand- stones. One collection, however, was recovered from oolites. Discussion. The Corbulomima association is the only association in the Corallian where infaunal deposit-feeders are significant. This trophic group favours a fine- grained substrate which is also the facies in which they occur in the Corallian. By churning the sediment for food, deposit-feeders increase the water content of the sediment near the depositional interface (Rhoads 1970) and such thixotropic muds are easily suspended even by weak currents. This may be the reason for the low percentage of epifaunal suspension-feeders in this association. Where larger shells provided more extensive substrate for the cemented and byssally attached epifauna, the latter increased in numbers leading to a change in community: these shell pave- ments are colonized by members of the N. nana/Chlamys or oyster /Isognomon promytiloides associations. The thixotropic muds probably facilitated the winnowing of the mainly thin-shelled infauna even by relatively weak currents. In situ reworking or small-scale lateral transport is indicated by the low percentage of bivalved fauna and by the fact that no fauna is preserved in life position. As hardly any shells are bored or encrusted, the rate of sedimentation seems to have been fairly continuous. One collection from the Osmington Oolite Group of Dorset does not fit into the general picture of the environment of the Corbulomima association. This collection has been found in high-energy, planar cross-bedded oolites with a very high fragmentation rate (95%). Some shells are worn and quite a few bored (this sample accounts for all bored shells of the association). The shells are found in a lumachelle 2-3 cm in thick- ness and exhibit size sorting, which indicates that they have undergone transport. Four feeding levels can be recognized: infaunal deposit-feeders (e.g. Palaeonucula , Mesosacella), deposit-feeders at the depositional interface ( Discomiltha , Pro- cerithium), low-level suspension-feeders (e.g. Pleuromya , Nanogyra), and inter- mediate-level suspension-feeders (Pinna). O. Oyster /Isognomon promytiloides association Description. The oyster//, promytiloides association contains 459 specimens in nine collections. Cemented epifauna dominates the trophic nucleus (text-fig. 24) with N. nana , Plicatula weymouthiana , and I. promy- tiloides on the first-rank positions. These are followed by Deltoideum delta , Lopha gregarea , Gryphaea ( B .) dilatata, ‘ Ostrea' sp., Clilamys (R.) fibrosa, and, finally, Pleuromya uniformis. Other cemented bivalves include L. solitaria , Placunopsis radio ta. P. duriuscula, ' Ostrea ’ cpiadr angular is, and Praeexogyra sp. D. delta, ‘ Ostrea ’ sp., P. weymouthiana , and /. promytiloides reach their distribution maxima and are thus characteristic of this association. Except for one brachiopod, the association consists solely of bivalves. Trace-fossil diversity is high, abundance is low. The more important forms include the deposit-feeders Teichichnus rectus, Cylindrichnus concentricus. Chondrites, and the crustacean-made Spongeliomorpha suevica var. B. The over-all diversity is medium (text-fig. 3). The epifauna (90-4%) dominates the ecological spectrum and is represented by cemented forms (67 1%; see trophic nucleus), byssally attached bivalves (17 0%), and free-living species (e.g. Gryphaea', 61%). Semi-infaunal bivalves (Modiolus, Gervillella) are of lesser significance (3-0%) as are deep burrowers (5-0%; e.g. Pleuromya, Pholadomya). The remaining ecological groups are rare or missing (text-fig. 21). Only 1-3% of the fauna is in life position but 15-03% of the bivalves are preserved with both valves (mainly Pleuromya , Isognomon, and Plicatula ). 26-8% of FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 369 Oyster / Isognomon promytifoides ass. text-fig. 24. Trophic nucleus and attempted reconstruction of the Oyster / Isognomon promytiloides association (O). Length of bar, 6 cm. 1, Nanogyra nana ; 2, Plicatula weymouthiana; 3, Isognomon promyti- loides; 4, Deltoideum delta ; 5, Lopha gregarea\ 6, Gryphaea dilatata; 7, ‘ Ostrea ' sp.; 8, Chlamys fibrosa ; 9, Pleuromya uniformis; benthic algae hypothetical. 370 PALAEONTOLOGY, VOLUME 20 the fauna are encrusted by a diverse fauna of epizoans such as L. gregarea , L. solitaria, Nanogyut. Plicatula , Placunopsis , the foraminifera Nubeculinella, various serpulids ( Dorso -, Cyclo-, and Pentaserpula) as well as the bryozoans Berenicea and Stomatopora. Phoronids, Lithophaga inclusa , and cirripedes bored 9-8% of the fauna. The percentage of fragmentation is high, i.e. between 80 and 98%. Most burrowing species are preserved as steinkerns, the rest with their shell. In some collections part of the epifauna, especially oysters and Isognomon , is very worn and intensively bored and encrusted. The oyster//, promytiloides association exhibits a clear substrate and facies preference occurring mainly in Fe-oolitic argillaceous limestones and marls. Discussion. The environment in which most of the oyster//, promytiloides association accumulated was similar to that of the M. bipartitus/P. alduini and L. gregarea associa- tions, i.e. slow sedimentation, a fairly stable, fine-grained substrate, occasional influx of Fe-ooliths, and temporarily high-energy conditions. The slow sedimentation rate led to accumulation of shell material on the sea floor which was often intensively encrusted and bored as well as corroded. Large ‘ Ostrea ’ commonly served as sub- strate for small clusters of Lopha, Nanogyra , and Plicatula. In some collections, however, excellent preservation of the epifauna, often still bivalved, indicates relatively rapid sedimentation. Transport played only a minor role. The few shell layers present accumulated in situ due to non-deposition, rather than through current action. Thus the oyster//, promytiloides association represents again the remnant of an ancient community. An inhomogeneous trophic grouping is the result of the presence of deposit-feeders and suspension-feeders. Two feeding levels could be recognized: infaunal deposit-feeders represented by trace fossils, and low-level suspension-feeders (e.g. burrowing bivalves, Nanogyra). P. Serpula intestinalis association Description. Four collections with 638 specimens have been grouped into this association which was found only in the uppermost Corallian and lowermost Kimmeridgian of the Dorset coast. The trophic nucleus (text-fig. 25) consists of the ubiquitous N. nana , S. intestinalis, the brachiopod Torquirhynchia inconstans, and the bivalves Pleuromya uniformis , Ctenostreon proboscideum , Deltoideum delta, and Aniso- cardia isocardioides. S. intestinalis and C. proboscideum only rarely occur elsewhere and T. inconstans is confined to this association; these three therefore serve as its characteristic species. Bivalves form 69-6% of the fauna, gastropods 4-5%, brachiopods 4-7%, and free-living serpulids 21-2%. No trace fossils occur except some indistinct tubes most likely left by deposit-feeders. The over-all diversity is medium (text-fig. 3). Semi-infaunal bivalves (T6%) are represented by Modiolus, the infauna (12-4%) by deep-burrowing bivalves ( Goniomya , Pleuromya, Pholadomya), shallow-burrowing bivalves (e.g. Anisocardia, Protocardia, Isocyprina), and some gastropods (text-fig. 26). Most species (86 0%) belong to the epifauna. They are either cemented bivalves (48-4%; e.g. the oysters Nanogyra, Deltoideum), byssally attached bivalves (6-7%; e.g. Isognomon, Chlamys), free-resting animals like the bivalves C. proboscideum, S. intestinalis, and some gastropods (e.g. Bathrotomaria, Pleurotomaria). Another faunal element are small colonies of corals, mainly Thecosmilia annularis and Thamnasteria which occur scattered throughout the samples, especially those from the Ringstead Coral Bed. No specimens of the uncemented fauna occur in life position, but 23-6% are still bivalved. 13-5% of the fauna have been encrusted by the foraminifera Nubeculinella, the bivalve Nanogyra, and the serpulids Dorsoserpula, Cyclo serpula , and Pentaserpula. In addition, specimens of S. intestinalis were commonly found to be infested by the hydroid Protulophila gestroi. (For detailed discussion of this commensalism see Scrutton 1975). 10-5% of the shells are bored by the bivalve Lithophaga inclusa, cirripedes, polychaetes, phoronids, and some unknown organisms. The degree of fragmentation is fairly low and varies between 10 and 50%. Altogether, 15-7% of the fauna are preserved as steinkerns, mainly the infaunal bivalves as well as gastropods and the rare ammonites. The S. intestinalis association exhibits a clear preference for clay with a varying amount of Fe-hydroxide ooliths. In two samples from the basal Kimmeridgian, reworked phosphate pebbles and phosphate stein- kerns, partly encrusted with Nanogyra, are common. FURSICH: CORALLIAN BENTHIC ASSOCIATIONS 371 O 20 40 % intestinalis ass. text-fig. 25. Trophic nucleus and attempted reconstruction of the Serpula intestinalis association (P). 1, Nanogyra nana; 2, Serpula intestinalis ; 3, Torquirhynchia inconstans; 4, Pleuromya uniformis ; 5, Cteno- streon proboscideum\ 6, Deltoideum delta ; 7, Anisocardia isocar dioides. Discussion. The S. intestinalis association is tied to a specific environment in which non-deposition and reworking played a certain role. Two collections come from the uppermost Corallian, the Ringstead Coral Bed, the other two from the basal Kimmeridgian, the Torquirhynchia Bed. Although Brookfield (1973 b) maintains that Torquirhynchia occurs only in the basal Kimmeridgian, it has also been found occasionally in the Ringstead Coral Bed. Corals are only found in the latter. They never formed reef-like structures but small isolated colonies growing on a consoli- dated substrate. In the lower Kimmeridgian, clay was deposited at a very slow rate with occasional influx of Fe-ooliths which are distributed in the form of small seams Ill PALAEONTOLOGY, VOLUME 20 in the sediment. Encrusted pebbles and steinkerns indicate that reworking took place. The same association is present in both uppermost Corallian and lowermost Kimmeridgian which indicates that no change in environment took place across the stage boundary (see also Talbot 1973). S. intestinalis seems adapted to have lived on fine-grained substrate (text-fig. 25). The asymmetry of Torquirhynchia is thought to be an adaptation to life in a tidal environment (Brookfield 1973b), but there are doubts as to the validity of Brookfield’s interpretation. The relatively high percentage of still bivalved fauna (not only burrowers but also epifaunal species!) suggests that reworking was only on a small scale and that lateral transport was of little significance. Therefore, the S. intestinalis association, too, is the autochthonous relic of an ancient community. Only one feeding level can be distinguished: low-level suspension-feeders ( S . intestinalis burrowing fauna, pectinids). Bathrotomaria represent a second trophic group, probably browsing herbivores. text-fig. 26. Ecological composition of associations P-R. Legend in text-fig. 6. Q. Lithophaga inclusa association Description. Represented by 1421 specimens in fifteen collections the L. inclusa association has been found in Yorkshire, Oxfordshire, and Normandy. Ten species form the trophic nucleus (text-fig. 27) in which N. nana takes the first place followed by Metriomphalus muricatulus, L. inclusa , Procerithium sp. 1 , Nerinella sp., Plagiostoma zonatum , and Barbatia (Acar) sp. The remaining three species are Procerithium sp. 2, Ampullina sp., and Chlamys ( Chi. ) nattheimensis. L. inclusa, M. muricatulus, and Procerithium sp. 1 reach in this association the peak of their distribution and are thus characteristic of it. Bivalves (57-8%) and gastropods ( 14-9%) dominate the fauna; brachiopods and echinoids are very rare except for abundant FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 373 spines of Plegiocidaris florigemma. The only trace fossil found, and that only in one collection, is Chondrites. The over-all diversity is high (text-fig. 3). The majority (79-4%) of the specimens belong to the epifauna, the rest is infauna as semi-infaunal organisms are very rare (text-fig. 26). The epifauna is shared by cemented bivalves (21 -8% ; mainly Nanogyra), byssally attached bivalves (16-5%; e.g. Plagiostoma , Barhatia , Chlamys), and free-living gastropods (38-5%; e.g. Metriomphalus, Procerithium). Corals, also members of the epifauna, are very common but have not been included in the statistics. The commonest forms include Montlivaltia sp., Thamnasteria concinna, Thecosmilia annularis, Isastrea explanata , and Rhabdophyllia phillipsi. The infauna (20-4%) is dominated by the rock-boring bivalve Lithophaga. Deep- and shallow-burrowing bivalves are of only little significance. 19-8% of the uncemented fauna is in life position, mainly Lithophaga, Lithophaga inclusa ass. text-fig. 27. Trophic nucleus of the Lithophaga inclusa association (Q). Length of bar, 4 cm. Because of the compli- cated nature of reef environments no reconstruction of the association has been attempted. 1, Nanogyra nana\ 2, Metri- omphalus muricatulus ; 3, Lithophaga inclusa ', 4, Procerithium sp. 1; 5, Nerinella sp.; 6, Plagiostoma zonatum; 7, Barhatia sp. ; 8, Ampullina sp.; 9, Procerithium sp. 2; 10, Chlamys nattheimensis. and 29-9% of the organisms are still bivalved (again mainly Lithophaga). Only \°/0 of the fauna excluding corals are encrusted by the foramimfera Nuheculinella, the bivalves Nanogyra and Plicatula, two species of Serpula ( Dorsoserpula , Cycloserpula), and the bryzoans Berenicea and Stomatopora. 0-9% are bored by polychaetes, phoronids, and Lithophaga. Fragmentation is usually fairly high, between 75 and 98%. Nearly half the fauna (47-2%) is preserved as steinkerns or moulds, mainly gastropods, arcids, and some burrowing bivalves. The L. inclusa association occurs predominantly in biomicrites and biosparites thus showing a marked substrate preference. Discussion. As the reef environment in the Corallian will be the subject of a more detailed later study, the discussion here is fairly short. The L. inclusa association 374 PALAEONTOLOGY, VOLUME 20 has been found in three reefal environments: (a) in the reef itself, (b) in small pockets of lime mud within the reef, and (c) in the reef talus. Autochthonous in the first two environments, it also occurs in form of parautochthonous faunal elements in (c), there usually partly fragmented and worn (e.g. Wheatley Limestone of Oxfordshire and Berkshire). The Corallian reefs, only small structures at the most 6 m thick but usually far less, have been described in detail already by Arkell (1928, 1935) who compared them with Recent reefs in the Red Sea. Arkell also recognized a typical reef fauna which is more or less equivalent to the reef association described here. A reef provides far more niches than the usually fairly uniform sea floor and this is also reflected in the fauna. Herewith a few examples: Barbatia ( Acar ) dwelt as a byssate nestler in crevices amongst corals ; Lithophaga and the numerous epizoans found extensive substrate on corals, burrowers like Protocardia lived in small pockets of lime mud within the reef and Metriomphalus muricatulus as well as Procerithium sp. I were browsers feeding on algae. The description of feeding levels in a reef is a very complex matter but low-level suspension-feeders (burrowing bivalves), and high-level suspension-feeders (e.g. corals) could be distinguished. Herbivores (e.g. Metriomphalus ) and possibly scavengers (Ampullinal) represent two more trophic groups. R. Nanogyra nana association Description. The N. nana association is represented by 3037 specimens in fifteen collections. N. nana is the only member of the trophic nucleus (text-fig. 28). Although occurring in nearly all other associations as well, often even as a significant part of the community, it is here that Nanogyra reaches the peak of its distribution (85-3%) and is thus the characteristic species. 97-0% of the fauna are bivalves; gastropods (0-5%), brachiopods (0-8%), echinoderms (0-7%), sponges (0-2%), and serpulids (0-7%) forming the rest. Trace fossils are very abundant and diverse, the most important ichnospecies being Spongeliomorpha suevica var. B, Cylindrichnus concentricus, and Teichichnus rectus. The over-all diversity is medium (text- [ 0 20 40 60 80 % others [ Nanogyra nana ass. text-fig. 28. Trophic nucleus and attempted reconstruction of the Nanogyra nana association (R). Length of bar, 1 cm. 1, Nanogyra nana ; 2, Spongeliomorpha suevica ; 3, Cylindrichnus concentricus ; 4, Teichichnus rectus. FURSICH: CORALLIAN BENTHIC ASSOCIATIONS 375 fig. 3). Epifaunal cemented bivalves ( Nanogyra and some other oysters) account for most of the fauna (86-8%) and the only other important group are epifaunal byssally attached bivalves (7-2%; e.g. various species of Chlamys). The remaining ecological groups are fairly insignificant, e.g. the infauna totals 3-4% (text-fig. 26). Only 2-9% of the fauna are preserved with both valves, and of the uncemented fauna, only 0-2% are in life position (some deep-burrowing Pleuromya ). Encrusted are 4-7% of the fauna by one or more of the following epizoans: the oysters Nanogyra , Lopha gregarea, the serpulids Cycloserpula , Pentaserpula, Tetraserpula , and Dorsoserpula, the foraminifera Nubeculinella, and thecideacean brachiopods. Only two specimens are bored by the bivalve Lithophaga inclusa and cirripedes. Nearly all faunal elements are pre- served with their shell except some aragonitic forms (some gastropods and thin-shelled burrowing bivalves). In bioclastic limestones specimens are not infrequently coated with algal envelopes; more rarely specimens are worn. The Nanogyra nana association exhibits a medium substrate range occurring predominantly in sand- stones and impure limestones occasionally with shell beds. Discussion. The Lopha gregarea association (J) has a similar ecological composition to the N. nana association (text-figs. 16, 26), but there L. gregarea replaces Nanogyra as the dominant species. Nanogyra must have been adapted to live in a wide range of environments provided that suitable substrate (e.g. shell material) was present. It was probably an opportunistic species, which took over when other species of the same ecological niche (e.g. Lopha ) could no longer compete, due to unfavourable conditions. An unstable environment is indicated by the low diversity. Which factor was responsible for the eurytopic behaviour is not clear. Settling behaviour of the larvae most likely played an important role. The N. nana association probably thrived in a fairly high-energy environment, indicated by the shell beds, specimens with algal envelopes, and coarse sediment. Small-scale transport was probably a frequent event and most collections are the parautochthonous relic of the ancient community. In some cases, however, Nanogyra seems to have been embedded more or less in life position, i.e. forming small clusters on an otherwise fairly barren sea floor. More than 99% of the fauna are suspension-feeders, solely a few Discomiltha , Procerithium , and trace fossils represent deposit-feeders. Two feeding levels can be recognized: infaunal deposit-feeders (trace fossils) and low-level suspension-feeders (e.g. Pleuromya , Anisocardia). Some (?)herbivorous gastropods ( Bathrotomciria ) account for a third trophic group. S. Diplocraterion association Discussion. Fiirsich (1975) recognized three trace-fossil associations in the Corallian. Two of them, the Teichichnus and the Rhizo cor allium associations, occur together with associations of the body fauna, the third, however, the Diplocraterion association, is not found with any of the associations described under A to R. The Diplocraterion association (text-fig. 29) has been found in eleven samples from Oxfordshire, Dorset, and Normandy. It is characterized by burrows of suspension-feeding crustaceans and ‘worms’, i.e. D. parallelling, D. habichi, Spongeliomorpha nodosa , 5. saxonica, and Arenicolites variabilis. Associated with them are Planolites sp., and less com- monly S. suevica var. A and var. B, Chondrites sp., and Cylindrichnus concentricus. Most of the burrows are deep, and in the case of D. parallelum frequently show adjustment to either erosion or sedimentation by vertical shifting of the burrow. Orientation of burrows is encountered occasionally (text-fig. 30). The sediments 376 PALAEONTOLOGY, VOLUME 20 text-fig. 29. Attempted reconstruction of the Diplocraterion association (S). 1, Diplocraterion parallelism ; 2, Diplocraterion habichi ; 3, Arenicolites variabilis; 4, Spongeliomorpha nodosa/ saxonica. in which the burrows occur are predominantly fine- to medium-grained well-sorted sands and oolites. Sedimentary features are commonly associated with the trace fossils, notably oscillation ripples, large-scale planar or trough cross-bedding as well as herringbone cross-bedding and flaser bedding. The Diplocraterion associa- tion occupied a high-energy environment such as very shallow subtidal or intertidal areas (Fiirsich 1975). This environment must have been too rigorous for most of the bivalves and gastropods which domi- nate the remaining Corallian associations. In addition, removal of the body fauna by transport or breakage by frequent rework- ing could account for the non-preservation of any shells. The harsh environment allowed only animals to flourish which could adjust well to the unstable substrate, i.e. deep and rapid burrowers. Many niches occupied in the other associations are therefore here vacant due to the environ- mental stress. text-fig. 30. Orientation of Diplocraterion parallelism. Fallen block from the Bene I ill Grit, Berkshire Oolite Group, Bowleaze Cove, Dorset. Disturbed assemblages Only four of the 170 samples collected in the Corallian could not be grouped with any of the above associations. Their usually fairly large trophic nucleus either contains important faunal elements of several associations or else indicates selective removal or accumulation of certain elements from one association. In both cases, the structure and composition of the original community are distorted. There are two agents which cause disturbed assemblages: (a) currents which cause mixing and/or impoverishment of communities by transportation; ( b ) reduced sedimenta- tion which causes mixing by superimposition of several communities, that is con- FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 377 densation. Both cases are represented, although one or the other assemblage might have been created by a combination of the two agents. In Table 4 some characteristic features of the four collections are given. When comparing them, sample BH4 from the Osmington Oolite Group at Black Head, Dorset Coast, differs in having the highest fragmentation rate, no bivalved specimens and a different trace fossil suit (suspension-feeders); sedimentary structures are preserved and indicate a high-energy environment. Clearly, this sample is the result of transport especially as the shells form a thin lumachelle in the otherwise unfossili- ferous oolite. Sample SPB from the Shell cum Pebble Bed of Beckley, Oxfordshire, in turn, is a typical example of mixing caused mainly by condensation as the low fragmentation rate and the high percentage of still bivalved and encrusted specimens indicate. The remaining two samples occupy an intermediate position and probably both, reduced sedimentation and transport by currents, played a role in their forma- tion. Recognition of disturbed assemblages is of prime importance for environmental interpretation as well as to avoid making mistakes in reconstructing ancient com- munities and their structure. table 4. Some characteristics of mixed assemblages. Collection ns Bi-valved °/ /O Life position 7 /o Fragmen- tation 7 /o BH 4 9 0 0 95 BH 31 6 7-47 0 70 BH 25 13 1 1 64 0 90 SPB 10 20-20 0 50 Associated sedimentary structures Trace fossils Encrusted % Origin by Large-scale planar cross-bedding Skolithos, Arenicolites 9-90 Transport None None Chondrites Plano lites Chondrites , Planolites 904 9-84 Transport and condensation None Planolites 27-77 Condensation ns = number of species in the trophic nucleus. ENVIRONMENTAL DISTRIBUTION OF CORALLI AN BENTHIC ASSOCIATIONS The value of the associations as environmental indicators and their relationship to various facies types has been discussed elsewhere (Fursich 19766) so that only a summary is presented here. When the distribution of the associations was plotted against the sections of the Normandy and Dorset Corallian no cyclic sequence (representing transgressive and regressive phases) could be observed, in contrast to the sedimentary and trace-fossil evidence (Talbot 1973; Fursich 19766). Thus it seems that within shallow-water deposits of only moderate variations in depth the benthic fauna is not a useful tool for interpreting bathymetry. This is partly due to the fact that some abundant epifaunal species (e.g. Chlamys, Nanogyra) occur in a wide variety of substrates and that the others, though substrate-related, are not necessarily depth-related. 378 PALAEONTOLOGY, VOLUME 20 Despite this, several associations proved to be characteristic of certain environ- ments, especially when certain types of substrates were commonly found in a charac- teristic bathymetric position. The distribution of these associations in the Corallian is shown in text-fig. 31. For instance, the Pseudomelania heddingtonensis association (M) is characteristic of lagoons with lime mud or of sheltered lagoon-like environ- ments. The oyster / Isognomon promytiloides association (O) in turn preferred offshore muds and sands. Offshore condensed calcareous sandstones and sandy limestones were populated by members of the Myophorella clavellata (D) and Pleuromya uniformis (L) associations. The latter is also typical of subtidal sands, whilst the Corbulomima association (N) preferred silty muds of low-energy bays or the offshore shelf. High-energy environments such as near intertidal bars or intertidal sands and i rm 2i i 30 4 rsi text-fig. 31. Facies distribution of some Corallian associations. B = Pinna ass.; D = Myophorella clavel- lata ass. ; L = Pleuromya uniformis ass. ; M = Pseudomelania heddingtonensis ass. ; N = Corbulomima ass. ; O = Oyster/ Isognomon promytiloides ass. ; S = Diplocraterion ass. ; 1 , sand ; 2, silt ; 3, clay, argil- laceous/micritic limestones; 4, condensed sandstones and limestones. silts were usually devoid of body fauna but exhibit a distinct trace-fossil suite, the Diplocraterion association (S). The Pinna association (B) is typical of a similar environment, i.e. current-swept near intertidal bars. Finally, the Lithophaga inclusa association (Q) is restricted to the patch reef environment and to coral thickets or low coral biostromes. In contrast, several associations were found to be fairly independent of the environment. They consist mainly of epibenthic bivalves (e.g. in the Cldamys/ Nanogyra nana (G) and N. nana (R) associations) and occur over most of the Corallian. Thus several Corallian benthic associations, although not useful as bathymetric indicators, are indicative of particular facies and thus aid the reconstruction of the palaeogeographic and environmental set-up of the Middle and Upper Oxfordian of north-west Europe. COMPARISONS WITH OTHER MESOZOIC BENTHIC ASSOCIATIONS Benthic community studies have been carried out especially in the Silurian, Creta- ceous, and the Recent. Different approaches and presentation of data make a straight- forward comparison very difficult as does the degree of splitting or lumping of data FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 379 into ‘communities’. Furthermore, often only part of the fauna, for instance, brachio- pods (e.g. Calef and Hancock 1974; Boucot 1975), has been taken into consideration, and other elements like the infauna have been neglected. Some standardization is needed in the approach to associations, as well as an agreement on the definition of benthic associations or ‘communities’. Only then can direct comparison between benthic associations become meaningful and allow sound inferences about com- munity evolution. Thus the following comparisons are of a tentative nature only. The Lower Oxford Clay associations ( Callovian ) of central England. As no other Middle and Upper Oxfordian data are available yet, the nearest comparison is with the Lower Oxford Clay associations of central England (Duff 1975). Duff divided the fauna into ten biofacies associations, some of them showing only slight varia- tions in faunal composition. For instance, the difference between the blocky claystone, Meleagrinella shell bed, and the deposit-feeder bituminous shale biofacies consists in the different order in which faunal elements are arranged in the trophic nucleus. As we know from present-day communities, the faunal composition of samples from the same community can vary to a large extent; thus the three biofacies types might represent only variations of one association. None of Duff’s associations are found in the Corallian but his Grammatodon- rich bituminous shale biofacies resembles to some extent the Corbulomima association (N). Both have, in prominent position, an epifaunal (?) deposit-feeding gastropod ( Procerithium ), the small infaunal suspension-feeding bivalve Corbulomima , and the infaunal deposit-feeder Palaeo- nucula. Other genera of the trophic nucleus common to both associations are Mesosacella and Thracia. Pendent bivalves like Meleagrinella and Bositra in the Callovian are represented by Pteroperna pygmaea in the Corallian but are there of lesser significance. The Corbulomima association also resembles, to some extent, the Callovian Nuculacean shell bed and Grammatodon shell-bed biofacies which also consist of similar faunal elements. The substrate is identical in the Corallian and Callovian (i.e. mud), and so is the bathymetric position, i.e. low energy but well-aerated offshore subtidal. Middle and Upper Jurassic bivalve associations of Wyoming and South Dakota. Wright (1973, 1974) distinguished six bivalve assemblages, mainly from the Middle Jurassic of Wyoming and South Dakota which inhabited open shelf, littoral, and lagoon areas. The Bajocian Pleuromya subcompressa assemblage is similar to the Corallian P. uniformis association, both being relatively diverse and having roughly equal numbers of epifauna and infauna. Genera common to both trophic nuclei are Pleuromya and Gryphaea. Camptonectes in the Bajocian is replaced by Chlamys in the Corallian, Ostrea by Deltoideum. The P. subdepressa assemblage inhabited the open-shelf living on firm lime mud, which is also the environment in which part of the P. uniformis association in the Corallian is found. Coloradoan ( Cretaceous ) macroinvertebrate assemblages from the Central Western Interior. Kauffman (1967) described twenty-six macroinvertebrate assemblages from the Cretaceous of the Central Western Interior of U.S.A. of which at least two have counterparts in the Corallian. His Pinna assemblage (K) is dominated by Pinna and occurs in shallow-water massive sandstones which is exactly the environment of 380 PALAEONTOLOGY, VOLUME 20 the Corallian Pinna association (B). His plicate oyster assemblage (M) consisting mainly of Lopha which formed local clusters and restricted biostromes on many shallow-water nearshore calcarenites may be compared to the Corallian Lopha gregarea association (J) which similarly formed small biostromes but obviously inhabited a somewhat quieter and deeper offshore environment with a mud substrate (Fiirsich 19766). Lower Cretaceous benthic communities from the Southern Western Interior Basin. Scott (1974) recognized an Arenicolites association indicative of the upper shoreface when studying the benthic faunas in the Lower Cretaceous of the Southern Western Interior of the U.S.A. This association can be compared with the Corallian Diplo- craterion association which is also devoid of body fauna and indicative of a high- energy environment. The Corallian Corbulomima association combines most elements of Scott’s (1970, 1974) Corbula-Breviarea and Nucula-Nuculana associations, con- sisting mainly of shallow infaunal suspension-feeders and infaunal deposit-feeders. Both of Scott’s associations and the Corbulomima association occur in the same substrate (mud, clayey silty sandstones) and roughly similar environments (lower shoreface to open sea). CHARACTERISTIC FEATURES OF THE CORALLIAN ASSOCIATIONS 1 . Nanogyra nana. In most associations, the oyster N. nana occupies one of the first ranks of the trophic nucleus, except for the Pinna , Discomiltha, and Corbulomima associations. Usually, Nanogyra forms clusters either on the sediment or on other shells; shells covered with spat are quite frequent. The widespread occurrence and mode of distribution indicate that Nanogyra most likely acted as an opportunistic species (Levinton 1970) being extremely eurytopic and having few requirements besides hard substrate and more or less fully marine conditions. A similar distribu- tion pattern and mode of life is recorded of small Exogyra (e.g. E. columbella ) from the Cretaceous of the Central Western Interior (Kauffman 1967). These species might have filled the niche in the Coloradoan, which was occupied by Nanogyra in the Corallian. 2. Chlamys. Another conspicuous element of the Corallian fauna are several species of Chlamys, being the dominant group amongst the byssally attached epifauna. Not only characteristic of the Chlamys/N. nana association (G), they also occur in most trophic nuclei of the other associations. Although Chlamys ranges from the Triassic to the Recent, it does not play a role in the Cretaceous macroinvertebrate associations so far studied (Kauffman 1967; Scott 1970, 1974) nor in the other Jurassic benthic associations (Wright 1973, 1974; Duff 1975). The absence of Chlamys from the Lower Oxford Clay is most likely the result of a lack of suitable substrate (they are also rare in the Corallian Corbulomima association which occurs in a similar facies). In the Middle and Upper Jurassic of Wyoming and South Dakota the niche seems to have been filled by Camptonectes which are very widespread, and form the dominant element of the C. bellistriatus assemblage (Wright 1974). 3. Scarcity of brachiopods and crinoids. Brachiopods and crinoids are virtually absent from the Corallian of England and Normandy except for two horizons: FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 381 Torquirhynchia inconstans is a common faunal element in the top Corallian and basal Kimmeridgian, and clusters of the rhynchonellid Thurmanella dominate, together with Nanogyra , part of the Hambleton Oolite of Yorkshire. In the latter horizon the crinoid Miller icrinus is also a conspicuous faunal element. The appearance of Torquirhynchia at the Oxfordian/Kimmeridgian boundary in England and Normandy most likely is connected to the transgression which took place at the base of the Ringstead Coral Bed (Talbot 1973) and which enabled the asymmetric and therefore specialized Torquirhynchia to spread over a large area. Species of Torquirhynchia have been found, in similar stratigraphical positions, over much of Central and Western Europe, and possibly on the Russian platform (Childs 1969). The abundance of Thurmanella in certain beds of the Hambleton Oolite forming small clusters suggests that it, too, can be regarded as an opportunistic species especially as the other main faunal component is N. nana. This assumption is also supported by the low diversity of the fauna (see also Ftirsich 1976a). The scarcity of brachiopods and crinoids, both stenohaline groups, seems to point to lowered salinity as the reason for their restricted distribution. However, other stenohaline groups like ammonites, starfish, and especially echinoids, are not uncommon and in the latter case even form distinct associations, so that changes in salinity do not seem to be responsible. Unsuitable (shifting) substrates can also be discounted, as byssally attached bivalves are very widespread. Brachiopods constitute an important part of the fauna of the sponge reefs on the Swabian Alb and also on the high-energy carbonate platform of the Swiss Jura Mountains (author’s field observation), both in age comparable to the Corallian. The deeper bathymetric position of the sponge reefs, where brachiopods dominate the epifauna, suggests that there brachiopods still held the niches which were occupied by bivalves in the shallower parts of the epicontinental sea. However, this niche replacement cannot explain the abundance of brachiopods in the very shallow carbonate platform sedi- ments of the Swiss Jura Mountains during parts of the Middle and Upper Oxfordian. In equivalent environments in Dorset and Normandy (e.g. Osmington Oolite Group, Oolithe de Trouville) brachiopods do not occur. Differences in food supply combined with competition with the more efficiently feeding bivalves cannot explain this distribution pattern either, as near landmasses (in the case of England and Normandy) food supply was surely at least equivalent to that on the Swiss carbonate platform which bordered the Tethys. The least unsatisfactory explanation which can be offered is that the articulate brachiopods with their very short free-swimming larval period (probably a few hours or at the most a few days according to Rudwick 1970) were at a great disadvantage when colonizing large areas. When dispersal took place from a source area near the border Tethys/epicontinental sea and the continual decrease of the abundance of brachiopods from Switzerland through France into England seems to support this— the short free-swimming larval stage, combined with competition from bivalves and unstable environments (due to several transgressions and regressions) might account for the scarcity of brachiopods over most of the Corallian of England and Normandy. Articulate brachiopods are also rare in the Lower Oxford Clay of Central England (Duff 1975), probably due to the unsuitable substrate which presented difficulties in anchorage for most species and could easily clog their filter-feeding system. K 382 PALAEONTOLOGY, VOLUME 20 Except for Kallirhynchia myrina in the late Callovian and early Oxfordian, brachio- pods form no part of the benthic assemblages of the Middle and Upper Jurassic of Wyoming and South Dakota (Wright 1973), perhaps also for similar reasons as in the Corallian. Brachiopods are of no importance in any of the Cretaceous benthic assemblages. Perhaps their role in the Jurassic and Cretaceous epicontinental sea was chiefly confined to their appearance as opportunistic species, or specialists not being able to compete with bivalves under other (‘normal’) circumstances. Acknowledgements. Most of the work has been carried out whilst the author was a postgraduate at the Department of Geology and Mineralogy, Oxford University. I would like to thank Drs. A. Hallam, J. M. Hurst, W. J. Kennedy, T. J. Palmer, and R. Sykes for numerous discussions and Dipl.-Geol. S. S. Chrulew and F. Frolicher for critically reading parts of the manuscript. I am particularly grateful to Mrs. L. Scholz who drew the biotope reconstructions. The work at Oxford was carried out under the tenure of a grant from the Dr. Carl Duisberg-Stiftung fur das Auslandsstudium deutscher Studenten which is acknowledged with gratitude. This is contribution No. 52 of the Projektbereich Fossilvergesell- schaftungen of the Sonderforschungsbereich 53 ‘Palokologie’ at Tubingen. REFERENCES arkell, w. j. 1927. The Corallian rocks of Oxford, Berks., and north Wilts. Phil. Trans. R. Soc. Lond. B216, 67-181. — 1928. Aspects of the ecology of certain fossil coral reefs. J. Ecol. 16, 134-149. — 1929-1937. A monograph of British Corallian Lamelhbranchia. Palaeontogr. Soc. 90, 392+xxxviiipp. 1933. The Jurassic System in Great Britain. Oxford, 681 pp. — 1935. On the nature, origin and climatic significance of the coral reefs in the vicinity of Oxford. Quart. Jl geol. Soc. Lond. 91, 77-100. — 1935-1948. A monograph of the ammonites of the English Corallian Beds. Palaeontogr. Soc. 420+ lxxxiv pp. — 1936. The Corallian Beds of Dorset. Part I. The coast. Proc. Dorset Nat. Hist. Arch. Soc. 57, 59-93. blake, j. f. and hudleston, w. h. 1877. On the Corallian rocks of England. Quart. Jl geol. Soc. Lond. 33, 260-405. boucot, A. 1975. Evolution and extinction rate controls. Developments in Palaeont. and Stratigr. 1, 427 pp. bretsky, p. w. 1969. Central Appalachian Late Ordovician communities. Bull. Geol. Soc. Amer. 80, 193-212. — 1970. Late Ordovician benthic marine communities in north-central New York. N.Y. State Mus. and Sci. Serv. Bull. 414, 34 pp. Brookfield, m. 1973a. Palaeogeography of the Upper Oxfordian and Lower Kimmendgian (Jurassic) in Britain. Palaeogeogr., Palaeoclimatol., Palaeoecol. 14, 137-167. — 19736. The life and death of Torquirhynchia inconstans (Brachiopoda, Upper Jurassic) in England. Ibid. 13, 241-259. calef, c. E. and Hancock, n. j. 1974. Wenlock and Ludlow marine communities in Wales and the Welsh Borderland. Palaeontology, 17, 779-810. callomon, j. h. 1960. New sections in the Corallian Beds around Oxford, and the subzones of the Plicatilis Zone. Proc. Geol. Assoc. 71, 177-207. childs, a. 1969. Upper Jurassic rhynchonellid brachiopods from north-western Europe. Brit. Mus. Nat. Hist. Bull. (Geol.), Suppl. 6, 119 pp. duff, k. l. 1975. Palaeoecology of a bituminous shale— the Lower Oxford Clay of central England. Palaeontology, 18, 443-482. fursich, f. t. 1974. Corallian (Upper Jurassic) trace fossils from England and Normandy. Stuttgarter Beitr. Naturk., Ser. B, 13, 52 pp. 1975. Trace fossils as environmental indicators in the Corallian of England and Normandy. Lethaia, 8, 151-172. — 1976a. Fauna-substrate relationships in the Corallian of England and Normandy. Ibid. 9, 343-356. FURSICH: CORALLI AN BENTHIC ASSOCIATIONS 383 fursich, f. t. 19766. The use of macroinvertebrate associations in interpreting Corallian (Upper Jurassic) environments. Palaeogeogr., Palaeoclimatol., Palaeoecol. 20, 235-256. Gordon, w. a. 1965. Foraminifera from the Corallian Beds, Upper Jurassic, of Dorset, England. J. Paleont. 39, 828-863. guyader, j. 1968. Le Jurassique superieur de la bale de la Seine — Etude stratigraphique et micropaleonto- logique. Theses Univ. Paris, 269 pp. hallam, a. 1960. A sedimentary and faunal study of the Blue Lias of Dorset and Glamorgan. Phil. Trans. R. Soc. Lond. B243, 1-44. — 1967. An environmental study of the Upper Domerian and Lower Toarcian in Great Britain. Ibid. B252, 393-445. hudleston, w. h. 1878. The Yorkshire Oolites, Pt. 2, The Middle Oolites: section 2, the Coralline Oolites, Coral Rag and supracoralline beds. Proc. Geol. Assoc. 5, 407-494. hurst, j. m. 1975. Wenlock carbonate, level bottom, brachiopod-dominated communities from Wales and the Welsh Borderland. Palaeogeogr ., Palaeoclimatol ., Palaeoecol. 17, 227-255. janicke, v. 1970. Gastropodenfauna und Okologie der Riffkalke von Laisacker bei Neuburg a.d. Donau (Unter-Tithon). Palaeontographica Abt. A 135, 60-82. kauffman, e. g. 1967. Coloradoan macroinvertebrate assemblages, central Western Interior, United States. In kauffman, e. g. and kent, h. c. (eds.). Paleoenvironments of the Cretaceous seaway— a sym- posium. Colo. Sch. Mines , Spec. Publ. 67-143. lee, A. K. 1971. A petrographic and palaeoenvironmental study of the Coralline Oolite Formation (Corallian) east of Newtondale, Yorkshire. Unpublished Ph.D. thesis, University of London (Chelsea). lehmann, u. 1975. Uber Nahrung und Ernahrungsweise von Ammoniten. Palaont. Z. 49, 187-195. levinton, j. s. 1970. The paleoecological significance of opportunistic species. Lethaia , 3, 69-78. neyman, a. a. 1967. Limits to the application of the trophic group concept in benthic studies. Oceanology , Acad. Sci. USSR , 7, 149-155. petersen, c. G. j. 1913. Valuation of the sea. II. The animal communities of the sea bottom and their importance for marine zoogeography. Rep. Dan. Biol. St. 21, 44 pp. rhoads, d. c. 1970. Mass properties, stability, and ecology of marine muds related to burrowing activity. In crimes, T. p. and harper, j. c. (eds.). Trace fossils. Geol. J. Spec. Issue , 3, 391-406. — speden, i. G. and waage, K. m. 1972. Trophic group analysis of Upper Cretaceous (Maestrichtian) bivalve assemblages from South Dakota. Bull. Am. Ass. Petrol. Geol. 56, 1100-1113. rudwick, M. J. s. 1970. Living and fossil brachiopods. London, 199 pp. sanders, h. l. 1968. Marine benthic diversity: a comparative study. Am. Nat. 102, 243-282. scott, r. w. 1970. Paleoecology and paleontology of the Lower Cretaceous Kiowa Lormation, Kansas. Univ. Kansas Publ., Paleont. Contr. 52, 94 pp. — 1974. Bay and shoreface benthic communities in the Lower Cretaceous. Lethaia, 7, 315-330. scrutton, c. t. 1975. Hydroid-serpulid symbiosis in the Mesozoic and Tertiary. Palaeontology, 18, 255- 274, pis. 39-42. sellwood, b. w. 1972. Regional environmental changes across a Lower Jurassic stage boundary in Britain. Palaeontology, 15, 125-157. Stanley, s. M. 1970. Relation of shell-form to life habits in the Bivalvia (Mollusca). Mem. geol. Soc. Am. 125, 296 pp. — 1972. Functional morphology and evolution of byssally attached bivalve mollusks. J. Paleont. 46, 165-212. talbot, m. r. 1973. Major sedimentary cycles in the Corallian Beds (Oxfordian) of Southern England. Palaeogeogr., Palaeoclimatol., Palaeoecol. 14, 293-317. — 1974. Ironstones in the Upper Oxfordian of southern England. Sedimentology, 21, 433-450. thorson, g. 1933. Investigations on shallow water animal communities in the Lranz Joseph Ljord (East Greenland) and adjacent waters. Medd. om Gtfmland, 100, 1-68. — 1957. Bottom communities (sublittoral or shallow shelf). In hedgpeth, j. w. (ed.). Treatise on Marine Ecology and Paleoecology, 1, Ecology. Geol. Soc. Am. Mem. 67, 461-534. turpaeva, E. p. 1948. The feeding of some benthic invertebrates of the Barents Sea. Zool. Zhurn. 27, 503-512. [In Russian.] — 1949. Significance of food interrelationships in the structure of marine benthic biocoenoses. Dokl. Akad. Nauk USSR, 15. 384 PALAEONTOLOGY, VOLUME 20 turpaeva, e. p. 1957. Food interrelationships of dominant species in marine benthic biocoenoses. In nikitin, B. N. (ed. ). Transa , Inst. Oceanol., Mar. Biol. USSR Acad. Sci. Press , 20, 137-148. twombley, B. N. 1965. Environmental and diagenetic studies of the Corallian rocks in Yorkshire , west of Thornton Dale. Unpublished Ph.D. thesis, University of Newcastle. walker, k. r. 1972. Trophic analysis : a method for studying the function of ancient communities. J. Paleont. 46, 82-93. — and bambach, r. k. 1974. Feeding by benthic invertebrates: classification and terminology for paleo- ecological analysis. Lethaia , 7, 67-78. — and laporte, l. f. 1970. Congruent fossil communities from the Ordovician and Devonian of New York. J. Paleont. 44, 928-944. whatley, r. c. 1965. Callovian and Oxfordian Ostracoda from England and Scotland. Unpublished Ph.D. thesis, University of Hull. wilson, r. c. l. 1968a. Upper Oxfordian palaeogeography of southern England. Palaeogeogr., Palaeo- climatol., Palaeoecol. 4, 5-28. — 19686. Carbonate facies variation within the Osmington Oolite Series in southern England. Ibid. 89-123. wilson, v. 1933. The Corallian rocks of the Howardian Hills (Yorkshire). Quart. Jl geol. Soc. Lond. 89, 480-509. — 1949. The Lower Corallian rocks of the Yorkshire coast and Hackness Hills. Proc. Geol. Assoc. 60, 235-271. wright, j. k. 1972. The stratigraphy of the Yorkshire Corallian. Proc. Yorks. Geol. Soc. 39, 225-266. wright, r. p. 1973. Marine Jurassic of Wyoming and South Dakota: Its paleoenvironments and paleo- biogeography. Pap. on Paleont. 2, 49 pp. — 1 974. Jurassic bivalves from Wyoming and South Dakota : a study of feeding relationships. J. Paleont. 48, 425-433. ziegler, a. m., cocks, l. r. m. and bambach, r. K. 1968. The composition and structure of Lower Silurian marine communities. Lethaia , 1, 1-27. F. t. fursich Institut fur Palaontologie u. hist. Geologie Universitat Miinchen Richard-Wagner-Str. 10/11 8 Miinchen 2 West Germany Original typescript received 26 January 1976 Revised typescript received 10 June 1976 APPENDIX Localities of collections Yorkshire: FB Cliffs at Filey Brigg, North of Filey (TA 131815) TB I Betton Farm Quarries, Ayton (TA 002 856 and 001 855) TB II Crossgates Quarry, Seamers (TA 028 843) P I Disused quarries on Stonygate Moor (SE 868 849) P II Newbridge Quarry, Pickering (SE 800 860) NG North Grimston Hill Quarry (SE 8467) SQ Spaunton Quarry (SE 722 868) WQ Whitewall Quarry (SE 79 1 697) NRC Nunnington Railway Cutting (SE 649 787) Oxfordshire/Berkshire : SH Shellingford Cross Roads Quarry (SU 326 940) CH Cothill Quarry (SU 467 997) B Beckley Quarry (SP 570 108) CR Cross Roads Quarry, Oxford (SP 550064) FURSICH: COR ALLI AN BENTHIC ASSOCIATIONS 385 Dorset : BC Cliffs east of Bowleaze Cove (SY 708 818) HC Redclitf Point (SY 711 816) RB Ringstead Bay (SY 755 813) OS Cliff east of Osmington Mills (SY 737 816) OSM Cliff at Osmington Mills (SY 735 816) BR Cliff at Bran Point (S' Y 742 813) SC Cliff below Sandsfoot Castle, and cliffs towards Weymouth (SY 675 774) EF Low cliffs at East Fleet, Weymouth (SY 652 772) BH Cliffs between Black Head (SY 729 819) and Shortlake (SY 722 819) Normandy: HLF Cliffs (Falaises des Vaches Noires) between Houlgate and Villers-sur-Mer, 90 m west of the steps leading down to the beach from the camping ‘Les Falaises’. HLC Cliffs (Falaises des Vaches Noires) between Houlgate and Villers-sur-Mer, 500 m west of the steps leading down to the beach from the camping ‘Les Falaises’. V Cliffs (Falaises des Vaches Noires) south-west of Villers-sur-Mer. BV Low cliffs between Blonville-sur-Mer and Benerville-sur-Mer. MC Small quarries on the slopes of Mount Canisy HV Cliffs between Trouville-sur-Mer and Hennequeville THE CONIFERS FRENELOPSIS AND MANIC A IN THE CRETACEOUS OF PORTUGAL by K. L. ALVIN Abstract. Frenelopsis Schenk has been used to accommodate various species of fossil conifers with segmented shoots and small scale-like leaves at the nodes. As a contribution to the revision of conifers of this type, some of which probably have an affinity with Hirmeriella Horhammer, the species from the Cretaceous of Portugal are here redescribed.' Four species are now recognized : three (one tentatively) are retained in Frenelopsis , while the fourth, with spirally instead of cyclically arranged leaves, is placed in Manica Watson. The genus Frenelopsis was erected by Schenk (1869) for some segmented conifer shoots, somewhat resembling those of the modern genus Callitris. Schenk identified his specimens which were from the Cretaceous of the Carpathians as belonging to the same species as those earlier described by Ettingshausen (1852) from the Cretaceous of Austria as Thuites hoheneggeri. Since then F. hoheneggeri has been reported from various parts of the world and some ten other specific names have been erected. Unfortunately, the genus defined only on the basis of the brief descriptions and drawings in these early publications, has been used to accommodate quite a wide range of species. Recent efforts by a number of palaeobotanists to locate Schenk’s and Ettingshausen’s original material have so far been fruitless. An important character of the type material which is uncertain is the leaf arrange- ment. Some parts of the specimens illustrated in Ettingshausen’s paper appear to show leaves in cyclic arrangement with three or four at each node. Schenk describes the leaf arrangement as opposite decussate, but the only illustration that shows it (Schenk, pi. 6, fig. 6) has the appearance of an interpretative diagram and not a drawing of a specimen. Another doubtful character in both Ettingshausen’s and Schenk’s material is the presence or absence of suture lines or grooves running down the internode from the margins of the leaves above such as occur in modern Cupres- saceae. Ettingshausen’s figures show such lines, at least on the main axes. Of Schenk’s figures, while some show somewhat coarse ridges and furrows, in none, except the small specimen shown in plate 6, fig. 5 which may represent a different conifer, do these look like suture lines. Most modern authors have assumed that the absence of suture lines is a generic character. However, Kon’no (1966) has described a species from Malaysia, F. malaiana , with opposite decussate leaves and well-defined suture lines; this I would include in Cupressinocladus. Velenovsky (1888) described F. bohemica from the Upper Cretaceous (Ceno- manian) of Czechoslovakia. He described it as leafless, but Nemejc (1926) redescrib- ing the species showed that the leaves were in whorls of three and that the internodes lacked sutures. Hlustik (1972, 1974) has shown that the correct name of this species is F. alata (K. Feistmantel) Knobloch. Of the several other species which have been described, most have lacked suture lines. Leaf number and arrangement has frequently not been referred to, but in [Palaeontology, Vol. 20, Part 2, 1977, pp. 387-404, pis. 41-45.] 388 PALAEONTOLOGY, VOLUME 20 some it has been shown to be only one per node, e.g. F. parceramosa Fontaine (1889). Dr. Joan Watson, in the course of her revision of the English Wealden flora (unpublished thesis, University of Reading, 1964), encountered a conifer with segmented shoots and small spirally arranged leaves, one per node, which she tentatively identified with Fontaine’s species. She has proposed that a new generic name Manica Watson (1974) should be used for species resembling Frenelopsis but with this different leaf arrangement. Conifers of the Frenelopsis type occur mainly in the Cretaceous, though Barale (1973) has reported F. rubiesenis Barale, a species with leaves in whorls of three and no suture lines, from beds of Upper Jurassic age in Spain. The first record from Portugal of a fossil conifer of the Frenelopsis type was by Heer (1881) who described some shoots from the Lower Cretaceous ( Aptian- Albian) as a new species, F. occidentalis. The same name was used by Saporta (1894) for further specimens from Cereal (Barremian), Nazare (Cenomanian), and Padrao (Cenomanian). He erected another species, F. leptoelada Saporta, for specimens from Caixarias (Aptian) and from San Sebastian, Spain (Valanginian). Lima (1900) recorded specimens which he referred to as F. occidentalis Heer from Esgueira (Senonian) but Teixeira (1946) referred these specimens to iF. aff. hoheneggerV . Romariz (1946) in a revision of the Portuguese Frenelopsis species recognized three species: F. lioheneggeri (Ettingshausen) from the Lower Cretaceous of Almargem and Olhos Amarelos in which specimens believed to be Heer’s (1881) holotype of F. occidentalis were included, F. lusitanica Romariz from Nazare (now regarded as Cenomanian), and F. oligostomata (Romariz) from the Senonian of Esgueira. All the available material of Frenelopsis in the Museum of the Servigos Geologicos de Portugal has been re-examined. This material does not have catalogue numbers. SYSTEMATIC DESCRIPTION Order coniferales Family ?cheirolepidiaceae Both Frenelopsis and Manica are tentatively classified in this family on the basis of Hlustik and Konzalova’s (1976) report of pollen of the Classopollis type in male cones attributed to F. alata and also similar pollen which Dr. Watson and I have found in male cones attributed to a Manica parceramosa (Fontaine) from the English Wealden. Genus frenelopsis Schenk Frenelopsis alata (K. Feistmantel) Knobloch Plate 41, figs. 1-5; Plate 42, figs. 1-6; text-fig. 1a 1881 Sclerophyllum alatum K. Feistmantel, p. 96; pi. 7, fig 1 a~k. 1882 IFrenelopsis hoheneggeri (Ettingshausen): Zeiller, p. 231; pi. 11, figs. 1-10. 1888 Frenelopsis bohemica Velenovsky, p. 590; figs. 1,2, 10. 1894 Frenelopsis occidentalis Heer: Saporta, p. 199; pi. 36, figs. 1, 2. 1921 Frenelopsis bohemica Velenovsky : Bayer, p. 41 ; text-figs. 5, 6. ALVIN: CRETACEOUS CONIFERS 389 1926 Frenelopsis bohemica Velenovsky: Nemejc, p. 133; pi. 1, figs. 1-3; pi. 2, figs. 1-5; pi. 3, figs. 1-4. 1946 Frenelopsis lusitanica Romariz, p. 144; pi. 3, figs. 2, 3; pi. 4, figs. 1, 2. 1948 Frenelopsis lusitanica Romariz: Teixeira, p. 102; pi. 43, fig. 2, 2a. 1974 Frenelopsis alata (K. Feistmantel): Hlustik, pi. 1, figs. 3, 4; pi. 3, figs. 1, 2. 1975 ? Frenelopsis (from le Brouillard): Broutin and Pons, p. 10, pis. 1-4 (specimens from le Brouillard). Locality. Nazare, Estremadura, Portugal. Horizon. Cenomanian. Description. Four good specimens of this species exist in the Lisbon collection. The one figured by Romariz (pi. 4, fig. 2) and also by Teixeira was used in the present study of the cuticle. Some of the specimens are partially pyritized and somewhat decomposed, others are compressed and carbonized. These latter usually show more or less prominent longitudinal furrowing, probably produced by shrinkage at some stage, but the pyritized specimens have a smooth surface. All the specimens show up to three orders of branching. The stoutest axes are up to 5 mm broad and have segments two to three times longer than broad; they are more or less straight and branched at nearly every node. Lateral branches have longer, narrower segments; these penultimate branches are up to 2 mm wide and somewhat zigzag due to an almost pseudodichotoinous mode of branching. The ultimate branchlets, arising singly at almost every node on the penultimate, are about 1 mm wide, with segments many times longer than wide. Branches arise from parent shoots at about 30°, disregarding the zigzag. There is no doubt that at least the ultimate branchlets had leaves in whorls of three. Plate 41, fig. 1 a, b shows two sides of the same node of an ultimate branch. The leaves are triangular, about 0-6 mm wide, 0-3 mm long, and have a fringe of unicellular hairs along the margins (PI. 41, fig. 4). The leaves are joined laterally to form a sheath as much as I mm deep. The form of the leaf-sheath on axes of lower orders is not clear, but its general appearance is similar. The internode cuticle is about 30 p m in total thickness (measured in folds under the light microscope) with the outer periclinal epidermal wall about 8 pm. Sometimes the cuticle on one side of the specimen appears paler and thinner than that on the other; differences are also seen in the SEM view of the outer surface (cf. PI. 42, figs. 1, 2). The darker more robust cuticle (represented in fig. 1) looks smooth at low magnification, but at high magnification shows some patterning (just discernible in the figure) suggestive of surface wax, whereas the cuticle from the opposite side (fig. 2) shows furrows marking the outlines of epidermal cells and is without evidence of wax. Whether these differences reflect a degree of dorsiventrality cannot be decided on present evidence. It is probably that side of the shoot which is uppermost and exposed on the block which has the thinner, paler type, and it is therefore possible that the exposed cuticle has suffered some kind of degradation since the specimens were collected. The very limited amount of material at my disposal has precluded any further investigation of this problem. The epidermal cells are very clearly seen in both the light microscope and the SEM. They are rather uniform, mostly slightly elongated, but never more than twice as long as broad; longitudinal rows are fairly well marked, (20-) 25 (-45) p. m wide. The anticlinal walls are about 4 ^m thick. Though portions of the inner periclinal 390 PALAEONTOLOGY, VOLUME 20 walls of the epidermis may be cutinized (PI. 42, figs. 3, 4), the hypodermis is generally not marked except in the region of the leaf and sheath (PL 42, fig. 6). The stomata are optically denser than the rest of the cuticle and therefore con- spicuous in the light microscope (PI. 41, fig. 2). They are arranged in rather irregular longitudinal rows, the rows being spaced irregularly, with one to four epidermal cells between, and the stomata irregularly spaced (75-150 /xm apart) in the rows (PI. 41, figs. 2, 5). Stomata continue up on to the base of the sheath but fade out on the teeth-like leaves themselves which were therefore presumably non-photosynthetic. Stomata continue abundantly towards the node below on that part covered by the lower sheath, only ceasing very close to the node. The stomata are highly specialized, with four to five (or six) subsidiary cells sur- rounded by a ring of a usually greater number of somewhat specialized epidermal cells. The form of the stomatal pit is shown reconstructed in text-fig. 1a. This is text-fig. 1. a, Frenelopsis alata (K. Feistmantel) Knobloch. Reconstruction of outside view of stoma and section of cuticle, x 500. b, IFrenelopsis occidentalis Heer. Recon- struction of outside view of stoma and section of cuticle, x 500 approx. based on sections viewed in the SEM (PI. 42, fig. 5) and also on surface views (PI. 42, figs. 1, 2) and light microscope observations (PI. 42, figs. 2, 3). The stomatal pit has a polygonal or more or less stellate mouth surrounded by a lobed canopy. The lobes are not always very distinct, but their number equals that of the subsidiary cells rather than that of the surrounding epidermal cells. I conclude, therefore, that they represent parts of the subsidiary cells. In the throat of the pit is a series of papillae again equal in number to the subsidiary cells. The guard-cell cuticles are thin and EXPLANATION OF PLATE 41 Frenelopsis alata (K. Feistmantel) Knobloch, from Nazare (Cenomanian). Fig. la, b. Two sides of the same node of an ultimate branch. Edges of leaves outlined by broken lines, SEM, x 40. Fig. 2. Cuticle of internode, light microscope, X 200. Fig. 3. Stoma, focused through the papillae in the throat of the stomatal pit, light microscope, X 800. Fig. 4. Edge of leaf, showing short hairs extending from the marginal cells, SEM, x400. Fig. 5. Inside of internode cuticle, SEM, x 200. PLATE 41 ALVIN, Frenelopsis 392 PALAEONTOLOGY, VOLUME 20 frequently absent in preparations, so that the inside view shows the papillae (PI. 42, fig. 4). As seen in outer surface view (PI. 42, fig. 1), the canopy is delimited peripherally by a circular groove, and it is the outer side of the groove which is delimited by the ring of surrounding epidermal cells. This interpretation of the stoma is supported by the appearance of specimens such as that shown in Plate 42, fig. 2 where the anti- clinal walls of the epidermis are marked by furrows; clearly here the stoma has five subsidiary cells and six surrounding epidermal cells. No satisfactory preparations were produced of the adaxial leaf and sheath cuticle, which is delicate, but the adaxial surface appears somewhat papillate and without stomata. Discussion and comparison. This Cenomanian species is well distinguished from the other Portuguese species. I identify it with the species described from Czecho- slovakia now known as Frenelopsis alata (for references, see synonymy). Dr. Hlustik has kindly examined photographs of the Portuguese material and agrees that there is a very close resemblance to that recently studied by him (pers. comm.), and with which he has found male cones containing Classopollis in association (Hlustik and Konzalova 1976). It seems to me from the available descriptions (Zeiller 1882; Carpentier 1937) that the specimens described from Bagnols (Turonian) in the south of France may also belong to the same species. These were described by Zeiller as F. hoheneggeri (Ettingshausen) and the cuticle was examined further by Carpentier when he described other specimens, again under the name F. hoheneggeri, from the Campanian of Sainte-Baume, France. These younger specimens, however, I do not believe are the same, but may be much closer to F. oligostomata from Portugal. The material described by Broutin and Pons (1975) from the Cenomanian of Le Brouillard, France, may also represent F. alata, but the form of the stoma is not quite clear from the description so far published. Frenelopsis oligostomata Romariz Plate 43, figs. 1-6; text-fig. 2a-d 1900 Frenelopsis occidentalis Heer: Lima, p. 1 1 (record only). 1937 ? Frenelopsis hoheneggeri (Ettingshausen): Carpentier, p. 5; pi. 1, figs. 1-5; pi. 2, figs. 1-3. 1946 Frenelopsis oligostomata Romariz, p. 144; pi. 5, figs. 1, 2. 1946 Frenelopsis aff. hoheneggeri (Ettingshausen); Teixeira, p. 236; pi. 1, figs. 1-4; pi. 2, figs. 1,2; pi. 3, figs. 1, 2. 1975 Frenelopsis (from Esgueira), Broutin and Pons, p. 10, pis. 1-4 (specimens from Esgueira). EXPLANATION OF PLATE 42 Frenelopsis alata (K. Feistmantel) Knobloch, from Nazare (Cenomanian). Figs. 1, 2. Stomata from opposite sides of the same branch segment, SEM, x950. Figs. 3, 4. Stomata seen from inside cuticle. Guard cells are absent in fig. 4 and the papillae in the throat of the stomatal pit can be seen, SEM, >< 900. Fig. 5. Vertical section through stomatal pit showing, at top, part of the ‘canopy’ and mid-way down, three of the throat papillae, two in section and one at the back, SEM, x 1600. Fig. 6. Abaxial cuticle of leaf showing cutinized hypodermis, SEM, X 900. PLATE 42 ALVIN, Frenelopsis 394 PALAEONTOLOGY, VOLUME 20 Locality. Esgueira, near Aveiro, Beira Littorale, Portugal. Horizon. Senonian. Emended diagnosis. Branching (?penultimate) shoots up to 5 mm wide, somewhat zigzag. Lateral branches (?ultimate) up to 4 mm wide, arising one at each node of the parent shoot at about 30° (disregarding zigzag). Leaves in whorls of three. Cuticle of internode about 30 ^m in total thickness (measured in folds under the light microscope) with the outer periclinal wall of the epidermis about 10 pm. Epidermal cells clearly marked (PI. 43, figs. 3, 5), mostly isodiametric but rather variable in shape and size (20-45 pm), seldom longer than wide, more frequently wider than long; longitudinal rows not well defined, except for short chains of cells. Anticlinal walls about 8 ^m thick. Elypodermis of isodiametric cells sometimes present. Stomata about the same optical density as the rest of the cuticle (PI. 43, fig. 5); arranged in longitudinal rows, but rows somewhat obscure due to rather distant spacing of stomata (75-250 p m apart) within the rows (PI. 43, figs. 3, 5). Subsidiary cells (four) five (or six). Mouth of stomatal pit bounded by a thick, obscurely lobed, wrinkled ring (PI. 43, fig. 1), the number of lobes equalling the number of subsidiary cells. Papillae in the throat of the pit short, wide, and frequently obscured by a deposit of an opaque, granular substance. Subsidiary cells inconspicuous in the light microscope, but seen clearly from the inside by SEM as wide, convex cells showing faint radial striations (PI. 43, fig. 4), the convexity corresponding to a thickening of the cuticle of the surrounding epidermal cells just beyond the outermost limit of the raised external ring; the furrow between the ring and the thickened area showing as a light translucent ring in the light microscope (PI. 43, fig. 6). Guard cells thinly cutinized, only the parts adjacent to the subsidiary cells generally surviving macera- tion (PI. 43, fig. 4). Stomatal orientation most frequently transverse. Stomata present on the lower part of the internode covered by the leaf-sheath of the node below but, at the top of the internode, ceasing rapidly on the sheath; upper part of sheath and leaves therefore probably non-photosynthetic. Sheath and leaf margins fringed with small unicellular hairs extending from the marginal cells (PI. 43, fig. 2). Further description and discussion. Of the three specimens all from the same locality, one is well preserved with a cuticle, the others being only impressions. The good EXPLANATION OF PLATE 43 Frenelopsis oligostomata Romariz, from Esgueira (Senonian). Fig. 1. Stoma from outside, SEM, x 1600. Fig. 2. Edge of leaf sheath (below) overlapping internode (above), SEM, 450. Fig. 3. Inside of internode cuticle, SEM, x 180. Fig. 4. Stoma seen from inside cuticle, SEM, x 800. Fig. 5. Internode cuticle, light microscope, x 200. Fig. 6. Stoma, focused through the stomatal pit, light microscope, x 800. PLATE 43 ALVIN, Frenelopsis 396 PALAEONTOLOGY, VOLUME 20 specimen, used here for cuticle characters and figured in text-fig. 2a-c, was figured by Teixeira (1946) as his plate 1, fig. 1, but not by Romariz, though he probably used this specimen for his cuticle studies. Brontin and Pons (1975) have recently described some new specimens from Esgueira which they have not yet specifically identified. However, it is clear from their description that they represent this species. text-fig. 2. Frenelopsis oligostomata Romariz. a, outline drawing of specimen. [Part of that figured by Teixeira (1946, pi. 1, fig. 1).] b, c, two successive nodes from the same specimen, x 25. D, reconstruction of outside view of stoma and section of cuticle, x 500. It was not possible to be quite certain about the number of leaves at the node on the basis of the specimens in Lisbon, but the specimens of Broutin and Pons (1975) make it quite clear that there are three. Text-fig. 2b, c represents two successive nodes. Two short, triangular, striated extensions to the sheath are seen on one side of the specimen at one node (text-fig. 2b), but only one at the next (text-fig. 2c). The relationship of the branch emission to the leaves and sheath is not clear. As text- fig. 2b, c shows, the branch joins the stem at about 1 mm below the level of the top of the sheath. If the branch is axillary, then the sheath must either be attached to the main stem at the level of the branch base and then split by the growing branch, or else it may be attached rather higher but along a line which dips strongly at the subtending leaf. ALVIN: CRETACEOUS CONIFERS 397 The stomata, though different in a number of details from those of F. alata , are basically comparable (cf. text-figs. 1a and 2d). Their comparative obscurity in the light microscope (PI. 43, fig. 5) contrasts markedly with their prominence in F. alata (PI. 41, fig. 2). The granular deposit in the stomatal pit was noted by Romariz; it may represent preserved wax. No satisfactory preparations have been made of the adaxial leaf and sheath cuticles. No specimens of Frenelopsis described from elsewhere can be said to be identical to this species, with the possible exception of those reported from the Campanian of Sainte-Baume, France, by Carpentier (1937). Unfortunately, however, Car- pentier’s specimens are not very fully described. The epidermal cells are similar in size, shape, and arrangement and in the thickness of the anticlinal walls, but the stomata, the pits of which are blackened by a dark deposit, show few details. The specimens tentatively named LF. aff. alata ’ by Hlustik (1974) from the Senonian of Czechoslovakia are much more like F. alata than F. oligostomata. Genus manica Watson Manica parceramosa (Fontaine) Watson Plate 44, figs. 1-8; text-fig. 3 1889 Frenelopsis parceramosa Fontaine, p. 218; pi. Ill, figs. 1-5; pi. 112, figs. 1—5 ; pi. 168, fig. 1. 1946 (pro parte ) Frenelopsis hoheneggeri (Ettingshausen) : Romariz, p. 143; pi. 2, figs. 1, 2; pi. 3, fig. 1. 1948 Frenelopsis hoheneggeri (Ettingshausen); Teixeira, p. 65; pi. 24, figs. 1-3. 1965 ( pro parte) Frenelopsis hoheneggeri (Ettingshausen): Reymanowna, p. 19; pi. 1, figs. 1, 3, 6; text-fig. 2a, c, e-m. 1974 Manica parceramosa (Fontaine) Watson, p. 428. (Name only.) Locality. Olhos Amarelos, near Cabo da Serra, Obidos, Estremadura, Portugal. Horizon. Aptian-Albian. Description. Two good blocks exist in the Lisbon collection; one of them was figured by Teixeira (1946) and it is this which I have used in my study. Both blocks bear several short lengths of broken shoots and numerous isolated segments. The shoot segments are 8-10 mm wide and up to about twice as long as wide. At the node is a single, triangular leaf represented by an extension of about 0-8 mm high from the normally continuous, collar-like sheath (text-fig. 3a). Opposite the leaf (text-fig. 3b) the sheath dips to its lowest point, forming a usually shallow sulcus. Occasionally, however, as in the segment above in this same specimen, the sulcus dips right down to the node below so that the sheath is ‘open’ rather than ‘closed’. The leaves are spirally arranged. Branching has not been clearly observed. The flanks of the leaf and the sheath bear abundant unicellular hairs extending from the marginal cells (PI. 44, fig. 2). The internode cuticle is about 30 p. m in total thickness (measured in folds under the light microscope) with the outer periclinal wall of the epidermis about 13 /xm thick. L 398 PALAEONTOLOGY, VOLUME 20 text-fig. 3. Manica parceramosa Fontaine, a, b, opposite sides of the same shoot fragment showing parts of two internodes, 6. c, reconstruction of outside view of stoma and section of cuticle, x 500 approx. The epidermal cells are clearly seen in the light microscope (PI. 44, figs. 7, 8) but owing to often strong cutinization of the inner periclinal wall and part of the hypo- dermis, they are not always clear in the SEM view of the inner surface (PI. 44, fig. 6). The epidermal cells are more or less isodiametric, but usually elongated (up to twice as long as wide) between the stomatal rows; they are 10-20 ju m wide. The anticlinal walls are about 6 jum thick. Hypodermal cells are most strongly cutinized between the stomal rows; they are elongated, about 10 fx m wide, and marked by thin flanges underlying the epidermal cuticle (PI. 44, fig. 6). EXPLANATION OF PLATE 44 Manica parceramosa (Fontaine) Watson, from Olhos Amarelos (Aptian-Albian). Fig. 1. Internode cuticle from outside showing four stomata, SEM, 420. Fig. 2. Edge of leaf-sheath, showing rather long hairs extending from the marginal cells, SEM, x 800. Fig. 3. Outer surface of adaxial leaf cuticle, SEM, x400. Fig. 4. Inner surface of adaxial leaf cuticle, SEM, x 400. Fig. 5. Internode cuticle from inside, SEM, x 220. Fig. 6. Internode cuticle from inside showing two stomata and cutinized hypodermis. Stomata showing ring of subsidiary cells and guard cells, SEM, x 550. Fig. 7. Internode cuticle, light microscope, x 150. Fig. 8. Stoma, focused through stomatal pit, light microscope, x600. PLATE 44 ALVIN, Manica 400 PALAEONTOLOGY, VOLUME 20 The stomata appear optically less dense than the rest of the epidermis and are therefore conspicuous in the light microscope. They are in well-defined rows separated by two to four epidermal cells. There are (four), five to six (or seven) subsidiary cells forming the stomatal pit. Externally (PI. 44, fig. 1), the pit is seen to be bounded by a thick raised ring which is in turn separated from the rest of the epidermis by a fairly deep furrow; again here it is the outer limit of this furrow which is formed by the edges of the surrounding epidermal cells of which there are usually one or two more than the number of subsidiary cells. In the light microscope (PI. 44, fig. 8) the furrow is indicated by the narrow light band just inside the dense ring formed by the surrounding epidermal cells. The guard cell cuticles are fairly thick and therefore generally present in preparations; orientation is random. Stomata fade out towards the edge of the leaf-sheath and are absent from the leaf itself, the outer surface of which is papillate due to protruding epidermal cells. The adaxial surface of leaf and sheath is strongly papillate (PI. 44, fig. 3). The adaxial cuticle of the leaf consists of almost square cells, the regular arrangement of which is sometimes disturbed by a ring-like group of cells suggestive of an abortive stoma (PI. 44, fig. 3, top left). In the sheath, the adaxial cuticle shows a continuous hypo- dermis. Discussion and comparison. Specimens belonging to this species may possibly have been included by Schenk (1869) in the material described as F. hoheneggeri from the Carpathians. Some of his specimens (e.g. Schenk, pi. 4, figs. 5-7) certainly bear a resemblance to this species. As already stated above and by Reymanowna (1965) and Hlustik (1974), the status of F. hoheneggeri is uncertain. However, most of Schenk’s specimens appear to be of a kind having more elongated segments with more than one leaf at the node, possibly three. The material described by Remanowna was certainly mixed. Dr. Reymanowna and I have recently re-examined this material and find that the shoots with more elongated segments and with three leaves show certain cuticle differences from those with shorter segments and only one leaf. These latter, on which Remanowna’s (1965) description of the cuticle was based, are regarded as identical to the Portuguese specimens described here and also to material from the Wealden of England which Dr. Watson and I are at present investigating. Dr. Watson has recently re-examined Fontaine’s type material of EXPLANATION OF PLATE 45 IFrenelopsis occidentalis Heer, from Almargem (Aptian-Albian). Fig. 1. Internode cuticle from outside, SEM, x 170. Fig. 2. Single stoma from outside, showing polygonal mouth of the stomatal pit and papillae in throat, SEM, x 1100. Fig. 3. Internode cuticle from inside, SEM, x 170. Fig. 4. Stoma from inside, showing subsidiary cells with entrances to the hollow throat papillae, SEM, x 800. Fig. 5. Section of stoma. Cuticle somewhat compressed, SEM, x 1400. Fig. 6. Internode cuticle, light microscope, x 200. Fig. 7. Stoma, focused through stomatal pit showing the prominent throat papillae, light microscope, x 800. PLATE 45 ALVIN, IFrenelopsis 402 PALAEONTOLOGY, VOLUME 20 F. parceramosa in the Smithsonian Institution and has concluded that it is identical to the English material. She will shortly publish an emended diagnosis of the species. Variations, however, do occur. The variation within the Portuguese material in the ‘open’ or ‘closed’ nature of the leaf-sheath has also been noted by Watson (1964) in material from England. No papillae have been observed extending from the raised ring around the mouth of the stomatal pit in the Portuguese specimens (though the cuticles of only two individual segments have been examined). Papillae are gener- ally present in material from elsewhere, though they may be absent from some English material. Watson (1974) has proposed the name Manica for Frenelopsis- like conifers with spirally arranged leaves. Certainly, among modern conifers, such a difference would be regarded as fundamental and meriting generic distinction. Accordingly, I am, for the present, using this name. However, the resemblances in morphology and cuticle characters among these fossil conifers, regardless of phyllotaxy, are strong and when more information concerning the other parts of the plants becomes available, it is possible that the difference in phyllotaxy may no longer provide sufficient grounds for generic separation. Oishi (1940) recorded this species from Korea, but his records cannot be accepted on the basis of his description which gives no anatomical detail. Genus ? Frenelopsis Schenk ? Frenelopsis occidentalis Heer 1881 Frenelopsis occidentalis Heer, p. 21 ; pi. 12, figs. 3 b, 4-7. 1946 ( pro parte) Frenelopsis hoheneggeri (Ettingshausen): Romariz, p. 143; pi. 1, figs. 1-3. 1948 Frenelopsis hoheneggeri (Ettingshausen): Teixeira, p. 55; pi. 14, fig. 4. Locality. Vale de Almargem, between Vale de Figuera and Rio de Sapos, near Almargem, Estremadura, Portugal. Horizon. Aptian-Albian. Emended diagnosis. Branching segmented shoots up to 4 mm broad. Segments up to four times longer than broad. Cuticle of internode up to 60 pm in total thickness with the outer periclinal walls of the epidermis about 25 ^m thick. Inner periclinal walls almost continuously cutinized except for small perforations (PI. 45, figs. 3, 4). Anticlinal walls of hypo- dermis slightly cutinized. Epidermal cells in longitudinal rows, isodiametric or up to twice as long as wide, 15-25 p.m wide with anticlinal walls about 6 pm thick. Stomata arranged in longitudinal rows and optically slightly denser than the rest of the epidermis (PI. 45, fig. 6). Subsidiary cells, five to six. As seen from the outside (PI. 45, fig. 1), mouth of stomatal pits flush with general cuticle surface or slightly raised. Pit mouth irregularly polygonal in shape, the number of sides equalling the number of subsidiary cells. Papillae in the throat of the pit equalling the number of subsidiary cells, massive and obscuring the guard cells from above (PI. 45, figs. 2, 5). Guard-cell cuticles rather thin, but commonly surviving maceration. Stomata irregularly orientated, but with some tendency to be parallel in a given area of cuticle. ALVIN: CRETACEOUS CONIFERS 403 Further description and discussion. Heer (1881) originally described his material as coming from a different locality (Vale de Lobos) but Choffat (1894) gave reason for believing that it was from Almargem. This was accepted by Teixeira (1948) and Romariz (1946) who regarded the material as Beer’s holotype of F. occidentalis. 1 therefore use this name. The single block apparently representing Beer’s holotype bears a number of broken shoots and isolated segments. Much of the compressed material has peeled off leaving only impressions. Unfortunately it has not been possible to determine the leaf arrangement and for this reason the generic name must be regarded as tentative. The specimen, together with those from Olhos Amerelos which 1 have identified as Monica parceramosa, was identified as F. hoheneggeri (Ettingshausen) by Romariz. Bowever, although the material from Almargem is fragmentary and has not allowed a very full description, the cuticle is well preserved and provides evidence that it is probably distinct. The most notable differences are the much greater thickness of the cuticle, the more strongly cutinized hypodermis, and stomata differing in a number of details. 1 know of no other species of Frenelopsis with a similar cuticle but the form described as ‘ Frenelopsis sp.’ by Blustik (1974) from Stramberk (?Bauterivian) shows some resemblance. Other records from Portugal. Teixeira (1948) records Frenelopsis from several other Cretaceous localities. Some of these specimens I was able to examine in Lisbon, but in none is the preservation good enough for cuticle preparations and accurate identification. F. leptoclada Saporta (1894) from the Aptian of Caixaria and also from the Valanginian of San Sebastian in Spain, is based on poor impressions only and cannot therefore be characterized in any worthwhile way. Another species, probably new, having two opposite leaves at each node, from Sesimbra (Bauterivian or Barremian) is currently being investigated by Dr. J. Pais, University of Lisbon. Acknowledgements. I am grateful to Dr. Moitino de Almeida, Director of the Museum of the Servipos Geologicos de Portugal, for the help and facilities he extended to me during my visit to Lisbon. I also gratefully acknowledge the grant from the Central Research Fund of London University which made my visit to Lisbon possible. My thanks are also due to Dr. J. Pais, University of Lisbon, for supplying me with information on the Portuguese Cretaceous. The SEM in the Botany Department, Imperial College, was obtained on a grant from the Science Research Council. REFERENCES barale, g. 1973. Contribution a la connaissance de la flore des calcaires hthographiques de la province de Lerida (Espagne): Frenelopsis rnbiesensis n. sp. Rev. Palaeobot. Palynol. 16, 271-287. broutin, }. and pons, d. 1975. Nouvelles precisions sur la morphologie et la phytodermologie de quelques rameaux du genre Frenelopsis Schenk. C.r. Congr. nat. Soc. Savantes, Paris, 2, 29-46. carpentier, a. 1937. Remarques sur des empreintes de Frenelopsis trouvees dans le Campanien inferieur de la Sainte Baume. Annls Mus. Hist. nat. Marseille , 28, 5-14. choffat, P. 1 894. Notice stratigraphique sur les gisements de vegetauxfossiles dans le Mesozoique du Portugal. In saporta, m. de, Flore fossile du Portugal: Nouvelles contributions a la flore mesozoique. Direction des Travaux geologiques du Portugal. Lisbon, pp. 228-288. 404 PALAEONTOLOGY, VOLUME 20 ettingshausen, c. von. 1852. Beitrag zur naheren Kenntnis der Flora der Wealdenperiode. Abh. geol. Bundesanst., Wien, 1 (3/2), 1 32. feistmantel, K. 1881. Der Hangendflotzung im Schlan-Rakonitzer Steinkohlenbecken. Arch. pHrodov. Prozk. Cech. 4 (6), 96-98. fontaine, w. m. 1889. The Potomac, or younger Mesozoic flora. Monogr. U.S. Geol. Surv. 15, 1-377. heer, o. 1881. Contribution a la flore fossile du Portugal Lisbon, pp. I -51. HLUSTIK, A. 1972. Frenelopsis alata (Cupress. fossil) Taxon, 21, 210. — 1974. New finds of Frenelopsis (Cupressaceae) from the Cretaceous of Czechoslovakia and their problems. Cas. Miner. Geol. 19 (3), 263-268. [In Czech, English summary.] — and konzalova, M. 1976. Polliniferous cones of Frenelopsis alata (K. Feistm.) Knobloch from the Cenomanian of Czechoslovakia. Vest, ustred. Ust. geol., 51, 37-45. knobloch, E. 1971. Neue Pflanzenfunde aus dem bohmischen und mahrischen Cenoman. Neu. Jb. Geol Palaont. Abh. 139 (1), 43-56. kon’no, e. 1966. Some younger Mesozoic plants from Malaya. Contributions to the geology and palae- ontology of southeast Asia, 38. Geol Palaeont. S.E. Asia, 3, 135-164. lima, w. de. 1900. Noticia sobre alguns vegetaes fosseis de Flora Senoniana (sensu lato) do solo portuguez. Commungoes Trab. Serv. geol. Port. 4, 1-12. nemejc, f. 1926. On the identity of Sclerophyllum alatum Feistm. and Frenelopsis bohemica Vel. Sb. st. geol Ust. csl. Repub. 6, 133-142. [In Czech, English summary.] oishi, s. 1940. The Mesozoic floras of Japan. J. Fac. Sci. Hokkaido Imp. Univ. (4), 5 (2-4), 123-480. reymanowna, m. 1965. On Weichselia reticulata and Frenelopsis hoheneggeri from the western Car- pathians. Acta Palaeobot., Cracov. 6 (2), 15-26. romariz, c. 1946. Estudo e revisao das formas portuguesas de Frenelopsis. Bolm Mus. Lab. miner, geol. Univ. Lisb. (4), 14, 135-150. saporta, m. de. 1894. Flore fossile du Portugal. Nouvelles contributions a la flore mesozoique. Dir. Trav. geol Port. Lisbon, 1894. schenk, a. 1869. Beitrage zur Flora der Vorwelt. III. Die fossilen Pflanzen der Wernsdorfer Schischten in den Nordkarpathen. Palaeontographica, 91, 1-34. teixeira, c. 1946. Flora Cretaceo de Esgueira (Aveiro). Port. Acta biol. B 1 (3, 4), 235-241. — 1948. Flora Mesozoica Portuguesa. I Parte. Lisbon, Serviqos Geologicos. 118 pp., 45 pis. velenovsky, j. 1888. Ueber einige neue Pflanzenformen der bohmischen Keideformation. Sber. K. bohm. Ges. ITAs. Math. -not. Kl. 1887, 590-598. watson, J. 1964. Revision of the English Wealden Fossil Flora. Unpublished Thesis, University of Reading, no. R 1 1 46. 1974. Manica: a new fossil conifer genus. Taxon, 23, 428. zeiller, r. 1882. Observations sur quelques culticules fossiles. Annls. Sci. nat. Bot. 13, 217-238. K. L. ALVIN Department of Botany Imperial College of Science and Technology Prince Consort Road London SW7 2BB Typescript received 9 January 1976 Revised typescript received 2 April 1976 OSTRACOD ASSEMBLAGES AND THE DEPOSITIONAL ENVIRONMENTS OF THE HEADON, OSBORNE, AND BEMBRIDGE BEDS (UPPER EOCENE) OF THE HAMPSHIRE BASIN by M. C. KEEN Abstract. Six ostracod assemblages have been recognized from the Headon-Bembridge Beds of the Hampshire Basin by the use of cluster analysis employing Jaccard and Dice Coefficients. They are thought to be salinity con- trolled and are comparable with Recent ostracod assemblages. Two are believed to represent a freshwater environ- ment (0-3°/oo salinity), the first characterized by Candona and Cypridopsis, the second by Moenocypris. Three are believed to represent brackish environments, characterized by Cytheromorpha (3-9°/00), Neocyprideis (9-0 1 6-5%0), and Haplocytheridea (16-5-35-0°/oo). The sixth assemblage is believed to represent a truly marine environment. Evidence for post-mortem movement of valves and for mixed assemblages is examined. The autecology of N. colwell- ensis (Jones) is studied, especially the relationship between salinity and size and the development of nodes; variation in carapace ornament is believed to be related to CaC03 content of the water. Seven new species and two new subspecies are described. The Tertiary Fluvio-marine Formation’ of Hampshire and the Isle of Wight has attracted the attention of many geologists interested in ancient environments. Amongst the early workers the most outstanding was Edward Forbes, who recog- nized the freshwater, estuarine, and marine nature of the deposits in his classic work of 1856. In more recent years palaeoecological aspects have been investigated by Bhatia (1957) and Murray and Wright (1974) with Foraminifera; Daley (1972, 1973) on the molluscan assemblages of the Bembridge Beds; Edwards (1967) using the total fauna; and Haskins (1971c) and Keen (1972/7) with ostracods. The ostracods are particularly suited for an investigation of these deposits because of their wide environmental range. Recent years have seen a vast increase in ecological data on living ostracods, so there is now a firm foundation for palaeoecological studies. The taxonomy of these upper Eocene species is also in a reasonably healthy state due to the work of Jones (1856, 1857), Jones and Sherborn (1889), Haskins (1968fl-19716), and Keen (1972c/, 6, 1973c/, b). Haskins (1971c) used ostracods to determine the depositional environments of the Palaeogene of the Isle of Wight, recognizing freshwater, brackish, and marine assemblages in the Headon- Bembridge Beds. Six assemblages are recognized herein, probably essentially salinity controlled. They are comparable with Recent ostracod assemblages, in particular with those recorded from lagoons bordering the Gulf of Mexico. The horizons studied (text-fig. 1) are the Headon and Osborne Beds, Bembridge Limestone, Bembridge Oyster Bed, and Bembridge Marls. Sampling details for the two main sections, Headon Hill and Whitecliff Bay, are shown in text-fig. 1. The age of the deposits remains controversial. The base of the Middle Headon Beds has traditionally been regarded by British palaeontologists as the base of the [Palaeontology, Vol. 20, Part 2, 1977, pp. 405-445, pis. 46-49.] 406 PALAEONTOLOGY, VOLUME 20 KEEN: EOCENE OSTRACOD ASSEMBLAGES 407 Oligocene, so they have been placed into the Lattorfian, the basal Oligocene stage. Others have placed the boundary at a higher position, often drawing the base of the Oligocene at the base of the Hamstead Beds. This is the view adopted here, so the Headon Bembridge Beds are regarded as being of late Eocene age. Edwards (1971) gives a full bibliography. The area of sedimentation was probably a restricted basin bounded by the Ports- down uplift to the north and the Brixton-Sandown uplift to the south. An eastward- flowing river system emptied into this, with the open sea lying to the east and south-east. The sea sometimes penetrated the embayment, giving rise to lagoonal and marine conditions. Evidence of contemporary earth movements has been deduced by Daley and Edwards (1971). METHODS One hundred and seventy-six samples were examined, of which ninety-one yielded ostracods. It had originally been decided to count the number of specimens of each species per unit volume of unwashed sediment, but this had to be abandoned as impractical. Some samples yielded several hundred specimens within a short period of time, while others needed several hours in order to find a dozen or so specimens. While it is possible to average out the number of specimens present to a standard unit volume, this practice was not followed because of the great discrepancies between the size of the final samples. The problem may not be too important, how- ever, since the rate of sedimentation is unknown, and volume for volume compari- sons may not have much significance; for example, does a high frequency of valves indicate slow deposition of sediment or dense populations? Secondly, the assemblages have been determined by using cluster analysis, which requires the recording of the presence or absence of a species in a particular sample. Thus the main consideration is whether a large enough sample has been examined to record all the species present. Kornicker (1965) has indicated that a sample of 300 specimens yields most species present, while even 150 approaches the actual number. These numbers were readily available for the mesohaline and polyhaline assemblages, but difficult to obtain for the freshwater ostracods, and so this criterion was waived for the latter. Table 1 shows the number of specimens examined in assemblages III, IV, and V. Cluster analysis is now a fairly standard technique, and various correlation coefficients have been used for determining ostracod assemblages (Kaesler 1966; Carbonnel 1969; and Cheetham and Hazel 1969). The computer program used in this study is that given by Bonham-Carter (1967). Clustering is by the unweighted pair-group method. In the analysis of the data only the common species were con- sidered, defined as those present in five or more samples, and representing at least 3% of the fauna in each sample; nineteen species fall into this category. The reason for eliminating the rare species is that their presence or absence in a sample is much more a matter of chance than for the common species. The results are presented in two forms. Text-fig. 2 is an R-mode analysis using the Dice Coefficient, presented as a constellation diagram, and using simple absence/presence data. The five assem- blages are shown as being distinct more graphically than by using a dendo- gram. The species of Assemblage V form a large and well-defined cluster with high table 1. Composition of samples belonging to Assemblages III, IV, and V. N = number of individuals; figures for genera indicate percentage occurrence. Note that, because only the common genera have been used, the percentages do not always total 100. Also note some anomalies that occur due to the clustering method; for example, WB8 with 67% Neocyprideis belongs to Assemblage III rather than IV, while BC5 with only 32% Neocyprideis belongs to Assemblage IV : this is because of the presence of a third genus, Cladarocythere in BC5. Note also the increase in diversity from Assemblage III, through IV, to Assemblage V. SAMPLE N CYTHEROMORPHA NEOCYPRIDEIS CLADAROCYTHERE HAPL0CYTHER1DEA BRADLEYA CLITHROCYTHERIDEA CYAMOCYTHERIDEA CYTHERETTA EOCYTHEROPTERON LEGUM1NOCYTHEREIS PTERYGOCYTHEREIS ASSEMBLAGE HH 22 112 50 50 HH 26 103 99 1 WB 8 108 33 67 WB 27 55 100 WB 29 96 95 3 ill WB 39 31 65 33 WB 53 95 89 11 HHCP 6 306 83 6 HHCP 13 77 50 HH 37 337 26 60 13 1 HH 47 398 35 5 25 HH 48 438 1 77 10 6 HH 50 218 80 8 HH 51 85 69 12 HH 58 287 82 17 CB 6 99 1 58 1 6 2 1 19 1 8 8 1 sv BC 5 96 63 32 4 BC 8 173 100 WB 63 61 2 96 2 WB 64 53 1 97 2 NHH 1 183 56 22 22 NHH 3 66 2 97 HH 38 115 31 64 1 2 2 HH 42 200 2 10 1 55 8 7 4 4 1 2 HH 43 348 8 1 29 3 16 10 16 1 4 6 HH 44 403 2 4 1 62 3 1 16 8 2 1 1 CB 1 111 2 39 3 2 5 30 2 10 CB 2 86 1 54 1 15 8 6 3 2 CB 4 813 57 6 1 24 3 2 2 CB 5 147 1 1 70 1 1 21 MF 1 242 7 1 11 7 7 15 48 1 3 1 V MF 2 173 1 5 13 10 3 10 53 1 3 1 BR 3 113 5 5 29 2 1 40 5 8 WB 18 160 4 55 1 25 5 1 WB 19 175 14 54 3 1 1 13 7 WB 22 120 1 40 10 WB 64a 68 2 2 85 Base VB 132 1 12 1 24 12 2 3 32 1 2 2 Mid VB 91 1 5 31 15 9 6 25 1 1 2 Top VB 549 1 6 31 9 4 5 34 1 3 1 KEEN: EOCENE OSTRACOD ASSEMBLAGES 409 B. forbesi C.headonensls C.porosacosta C.fabo/dcs E. wetherelli V pustu/osa HdebiHs- KEY Value of Dice Coefficient: — >70 60-69 50-59 ___ N.coIwqIIqdsis — C.bontonensis C.hebert/ana C.purii C. bulla III Mocnocypr/s spp. II J- L.deh'rata C.subde/to/deo C.da/eyi C.bulbosa V.ed wards! C. cliff endensis text-fig. 2. Constellation diagram showing the level of association between the nineteen common species. This is an R-mode analysis using the Dice Coefficient of correlation. Roman numerals indicate the five assemblages which can be recognized. correlation values between one another, indicating that they frequently occur together. Only one species, Haplocytheridea debilis , has a significant correlation with a species belonging to another assemblage. Text-fig. 3 is a Q-mode analysis using the Jaccard Coefficient, presented as a dendogram, but the species here are divided into absent (< 3%), present (3-25%), and abundant (> 25%) for analysis. In text-fig. 4 the assemblages are shown, together with rare species assigned to them, and with suggested salinities. OSTRACOD ASSEMBLAGES The assemblages recognized here are biofacies, i.e. groups of fossils frequently occurring together, and assumed to have had similar environmental requirements. The first part of the definition is supported by the statistical results, the second is deduced by comparison with living relatives. The environment is thought to be the primary control, interrelationships between organisms being secondary. The distribution of individual species differ, giving overlap between some of the assem- blages. This is the case with modern marine communities, and is to be expected if 00 JACCARD COEFFICIENT 0 QQOqqo oo -1* 1 I I I 1 1 1 1 r HH 53 WB49 HH 55 HH L. CB 12 HH 56 BC 3 HH 30 HH 24 HH 23 HH 19 WB 54 WB 4 HH 20 HH 25 HH21 HH 60 Base of U.B HH 35 WB 34 WB 42 WB 45 WB 65 WB 35 WB 33 MF 4 MF 3 HH 68 HH 64 HH 62 WB 53 WB 29 WB 27 HH 26 WB 39 WB 8 HH 22 CB 6 HH 47 NHH3-4 BC 8 WB 64 WB 63 HH 58 BC 5 HH 37 N HH 12 HH 48 HH 51 HH 50 WB 19 WB 18 BR 3 HH 42 Top VB MF 2 MF 1 Base V B Mid VB HH 43 CB 5 CB 4 CB 2 CB 1 WB 64 a HH 38 WB 22 HH 44 text-fig. 3. Dendogram of Q-mode analysis using the Jaccard Coefficient of correlation. For locality details see Appendix. KEEN: EOCENE OSTRACOD ASSEMBLAGES 411 the environmental needs of the species within a community are not exactly identical. In this study, overlap is particularly apparent with the brackish water assemblages, where there is almost complete gradation from one assemblage to another, while within assemblages some species are more widespread than others. The trophic role of ostracods has never been studied in any detail, although they are known to be predominantly scavengers, with some herbivores. They do not seem to be the specific prey of any animals, apart perhaps from the young stages of some boring gastropods, being eaten by benthonic suspension feeders and nektonic filter feeders. Shelter is important for many ostracods, and this can be provided by vegeta- tion or animals such as bryozoans; other ostracods live interstitially and so are independent even of this. Therefore it seems reasonable to consider the environment as the primary control, and the presence of other organisms as secondary. Comparison with living ostracods suggests that there was no interdependence between the ostracods themselves. Following from this, it appears justifiable to consider the ostracods on their own, without detailed consideration of the other phyla present. Once the assemblages have been recognized, there are other problems to consider. Within any over-all environment there will be different microenvironments; as mentioned above, some ostracods would have lived interstitially, others crawled over the bottom sediment, some lived amongst the weed, others swam amongst it. On death, all of these will be found in the same sediment, giving a single assemblage. It is possible to recognize some of these life styles amongst the freshwater ostracods by comparison with living relatives (Keen 1975), but generally this is not so. How- ever, even with Recent ostracods it is usually impossible to be more precise. Trans- portation after death is another factor to consider, and is discussed in more detail below. However, the assemblages are thought to represent the fossilized remnants of living communities, and as such, approach the ideal of a palaeontologically defined biocoenosis. This does not mean that no movement occurred after death, but it does imply that such movement was of a limited amount. Individual cases of mixed assemblages are discussed in the relevant sections below. One final problem remains to be considered. As will be seen from the following discussion, many of the deposits are regarded as lagoonal. By comparison with modern lagoons, sedimentation would have been slow, but rapid changes in environ- mental conditions could be expected. It is therefore possible that the samples have mixed more than one environment since they were collected from a vertical thickness of about 6 cm. In order to test this idea, three horizons were resampled in detail. Firstly, a small thickness of sediment was sampled in the field using a trowel; a minimum thickness of 2-3 cm could be collected in this way. A second method was to use an aluminium tray, 60 cm long, 5 cm deep, and 10 cm wide which was pressed into a carefully trimmed section, and a spade used to cut it out. In this way a complete section 60 cm long remained in the tray. Top and bottom were marked on the tray, and then it was wrapped in polythene to retain the moisture of the sediment. The sample was examined in the laboratory. Using this method it was possible to examine 0-5-T0 cm thicknesses of sediment. Thus, in addition to the 176 samples already mentioned, a further fifty-three samples were examined in detail. However, the latter were not used in the statistics because they were from selected horizons and would have heavily weighted the results: e.g. there would have been another thirty 412 PALAEONTOLOGY, VOLUME 20 samples of Assemblage V from only 50 cm of sediment, which is out of all propor- tion to sampling from other horizons. The results obtained from these additional samples are discussed further in the relevant sections below. The most important con- clusion, however, is that the original samples are perfectly adequate for this study. FRESHWATER OSTRACOD ASSEMBLAGES Two distinct freshwater assemblages which can be recognized have been dealt with in greater detail by Keen (1975). Both can be recognized in the Hamstead Beds (Keen 1972a), and are also recorded from the Palaeogene of the Paris Basin, Rhine Valley, Mainz Basin, and Hesse. Assemblage I. The Candona-Cypridopsis Assemblage. This is characterized by Candona spp. and Cypridopsis spp., with Virgatocypris edwardsi sp. nov. and Cypria dorsalta Malz and Moayedpour. Associated molluscs are Galba and Planorbina , with rare Melanopsis. Chara is usually abundant and seeds of waterplants often present. The sediment is a grey-green clay, a black carbonaceous clay, or a buff limestone. V. edwardsi , Candona cliffendensis sp. nov., and Cypridopsis bulbosa (Haskins) have a significant relationship with limestones; using a x2 test p was found to be 0 001, 0 005, and 0 03 respectively (Keen 1975). The first two species are more or less restricted to limestones, but C. bulbosa has a much wider distribution. Only one assemblage is recognized, although it could be divided into two sub-assemblages, i.e. one from limestones, and the other from clays. Assemblage II. The Moenocypris Assemblage. Moenocypris is the only ostracod belonging here. Two species are present, M. sherborni Keen in the Lower Headon Beds, and its descendant M. reidi Keen in the Upper Headon and Osborne Beds. Moenocypris is an extinct genus, but there is little doubt that it inhabited freshwater (Keen 1975). The samples come from grey clays with common Viviparus , Melanopsis, and Unio, rarer Potamaclis (Fam. Melaniidae); serpulids (sometimes present); and Chara (rare). Seeds of waterplants are often abundant. There is little doubt that these two assemblages inhabited freshwater. The difficulty is in accounting for the existence of two distinct assemblages. A certain amount of mixing is indicated by the coefficients of correlation; Moenocypris-Cypridopsis bulbosa is 17, with lesser values between other members of Assemblage I and Moeno- cypris. Such values can be explained by post-mortem movement. The controlling factor is believed to be the depth of water. Daley (1972) has already discussed this with regard to the molluscan assemblages, and Keen (1975) concerning the ostracods, so it is only necessary to give the conclusions here. The Candona-Cypridopsis Assemb- lage is thought to have characterized lake margins or shallow lakes up to about 1 m or so in depth. This would coincide with the upper infra littoral zone, or zone of emergent waterplants. The water would have been still or slowly moving, with an alkaline pH (~ 8) and a positive Eh. The Moenocypris Assemblage is thought to have characterized deeper water (2-15 m?) coinciding with the lower infra littoral zone, or zone of submerged waterplants, or bare muddy bottoms. The water would have been still, with poor circulation, an alkaline pH, and the bottom sediments probably had a negative Eh. KEEN: EOCENE OSTRACOD ASSEMBLAGES 413 BRACKISH WATER ASSEMBLAGES Three brackish water assemblages have been recognized, ranging from a low salinity assemblage, through mesohaline, to polyhaline. Assemblage III. The Cytheromorpha Assemblage. This is characterized by C. bulla Haskins; other ostracods found here are discussed below. This assemblage is more common than indicated on the dendogram, but several samples could not be included in the analysis because they contained too few specimens. It is found in two differing sediments: the first is a grey clay with abundant shells of the bivalve Potamomya plana J. Sowerby, subordinate Melanopsis, Theodoxus, and Corbicula , with rare Viviparus , Galba , and Planorbina; the second is a green clay with few or no molluscs. Cytheromorpha is found at the present day in salinities ranging from 0-35°/oo. It is characteristic of brackish waters. Wagner (1957) records it from lower mesohaline (2-10°/oo); Puri (1968) from an estuarine environment (0-26°/oo); Swain (1955) from typical Texan bays (2- 1 0°/oo) ; Curtis ( 1 960) as being typical of ‘closed shallow lagoons, away from fluvial influence and with little marine influence’. The living molluscs are typical of subtropical rivers and estuaries. The presence of rare Galba , Planorbina , and Viviparus suggest a fluvial influence, although these shells are probably alloch- thonous. It is usually assumed that the poorest brackish water assemblages, in terms of number of species, are found in salinities below 9o/0O. This assemblage has the least number of species, and, together with the known range of living Cythero- morpha, is taken to indicate a salinity of 3-9°/00. Assemblage IV. The Neocyprideis Assemblage. N. colwellensis (Jones) is the dominant ostracod of this assemblage, although the actual composition varies between samples. Cladarocy there spp. are commonly present, Cytheromorpha bulla and Haplocytheridea debilis (Jones) more rarely. Bradley a favosa Haskins is assigned to this assemblage, although as it only occurs four times it was not considered statistically; it is only found in samples dominated by N. colwellensis. It is placed in Bradleya for con- venience, belonging to an undescribed genus; a second species, B. dolabra Keen has been described from Oligocene brackish water deposits (Keen 1972a). The sediment is a clay or sandy clay. The molluscan fauna varies; samples with C. bulla contain juvenile Galba , Planorbina , and Corbicula obovata (J. Sowerby); a distinctive fauna from the Middle Headon Beds consists of Batillaria ventricosa (J. Sowerby), Potamides vagus (Solander), Globular ia grosa (Deshayes), and rare Ostrea : B. favosa and H. debilis are found with this second association. The commonest molluscs, however, are C. obovata and P. vagus. The Foraminifera Quinqueloculina is some- times present. Neocyprideis is an extinct genus, but is closely related to the living Cyprideis , which some believe to be its descendant. The latter is probably the most typical brackish water ostracod; its ecology has been fully described by Sandberg (1964): it is found in freshwater, is common in mesohaline waters, rarer in polyhaline and euhaline waters, but abundant in hypersaline conditions. Its greatest development is in mesohaline salinities. It usually inhabits relatively quiet water, such as creeks bordering estuaries, lagoons, or brackish ‘ponds’ close to the sea (Whittaker, pers. comm. 1975). It is a bottom crawler, or burrower, feeding on organic matter in the M 414 PALAEONTOLOGY, VOLUME 20 sediment. In low salinities, Cyprideis develops nodes on its shell, and may increase in size; these factors are studied below. Assemblage IV is taken to represent salinities of 9-0-16-5°/oo. The other characteristic ostracod of this assemblage is Cladcirocy there hantonensis. This extinct genus is found in Eocene and Oligocene brackish water sediments. Its ecology can only be derived from associated faunal elements, and there is some indication that different species had differing salinity preferences. C. hantonensis is taken to have inhabited upper mesohaline and polyhaline salinities, but C. apostolescui (Margerie) has been taken to belong to lower mesohaline salinities in the Paris Basin (Keen 1971, 1972a). This species is also present in the Hampshire Basin and represents a special assemblage consisting almost entirely of C. apostolescui. It is found in the marl immediately below the topmost Bembridge Limestone at Whitecliff Bay, Thorness Bay, and Bouldnor Cliff. Because it occurred in only four samples it could not be included in the cluster analysis, neither could the samples containing it, because the only other ostracod present, Neocyprideis , represented less than 3% of the fauna and so counted as ‘absent’ under the procedure adopted for this study. Assemblage V. The Haplocytheridea Assemblage. This is characterized by H. debilis. The sediment is a silty or sandy clay, often with large numbers of C. obovata in life positions. Other molluscs include Sinodia suborbicularis (Goldfuss), Ostrea , Globu- lar ia, Nucula , Natica , and a few Galba and Planorbina. The Foraminifera Rot alia and Quinqueloculina are commonly present. Living Haplocytheridea are recorded from estuarine and open lagoonal environ- ments, especially along the coast of Texas, Louisiana, and Florida. They are recorded from: depths up to 21 ft, salinity 16~33%0 (Curtis 1960); the Lower Bay subfacies, salinity 10-1 7°/00 (Swain 1955); rocky tide pools of normal marine salinity (Benson in Moore 1961); mangrove swamps and marine environment, salinity 23-40°/oo (Benda and Puri 1962); shallow marine waters (Hulings and Puri 1964); lagoonal areas, salinity 10-25°/oo (Engel and Swain 1967). Howe (1971) has shown how torose cytherideidae, which includes Haplocytheridea , have been used by petroleum geo- logists to locate ancient shorelines in the Gulf states of America ; thus Haplocytheridea is regarded as being typical of areas adjacent to the shoreline, usually lagoonal or estuarine, and generally indicating reduced salinities. From this it can be seen that Haplocytheridea is typical of polyhaline and marine salinities. The other ostracods present vary from sample to sample, but those with living relatives, i.e. Cushmanidea , Cytheretta , Cytherella , Paijenborchella , Paracytheridea, Pterygocythereis, Ponto- cythere , and Krithe , are essentially marine genera, although most can adapt to poly- haline conditions (Swain 1955; Curtis 1960; and especially Puri 1968). Assemblage V is taken to indicate a polyhaline environment, salinity ranging from 16-5 to 35-0°/oo. The differences between samples is discussed further below. The possibility of hyper- saline conditions must be borne in mind. These would have been very temporary, and due to evaporation. There is no geological evidence for permanent high salinities; there are no evaporites, and there is ample evidence of freshwater conditions with many water plants, and lowland vegetation. This contrasts with conditions in the Paris Basin. KEEN: EOCENE OSTRACOD ASSEMBLAGES 415 THE BROCKENHURST BED Assemblage VI. The Hazelina Assemblage. There were not enough samples from this thin (50-100 cm) horizon to warrant statistical treatment. However, the characteristic fauna suggests that this could be regarded as a separate assemblage. It lacks many of the characteristic polyhaline species of Assemblage V, while the presence of such genera as Bairdia, Hazelina , and Idiocythere suggest truly marine conditions. Molluscs are Natica , Ostrea , Modiola , Cardium , Fusus, and Voluta ; solitary corals are recorded; Foraminifera are Cibieides, Rotalia , Globigerina , and Nonion. The paucity of the fauna suggests near-shore, perhaps even sublittoral waters, but of normal marine salinity. MARSH FACIES Finally, there is a widespread facies which has not yielded ostracods. It consists of finely laminated clay and sand, sometimes with thin lignitic layers. It is considered to represent a marsh facies. It is especially prominent in the Lower and Upper Headon Beds. LIFE AND DEATH ASSEMBLAGES It has been suggested that Assemblages III, IV, and V are salinity controlled and probably formed a gradational sequence. Species which had a broad environmental tolerance probably lived amongst more than one assemblage, while movement of valves after death could lead to mixing of assemblages. Cytheromorpha bulla , Neo- cyprideis colwellensis, and Haplocytheridea debi/is are commonly present in more than one assemblage, so it is worth considering in some detail whether their distribu- tion is due to post-mortem movement or reflects original life habitats. Living members of these genera, as noted above, are adapted to a wide range of salinities, so it is possible that these fossil species might have behaved similarly. Several criteria have been used to indicate an ostracod biocoenosis: presence of several different moult stages, including adults; presence of carapaces rather than single valves; and equal numbers of each valve. The following example illustrates the use of these, together with stratigraphical succession and faunal composition. In text-fig. 5 a detailed section through the Cyrena pulchra Bed (Lower Headon Beds) is shown, with the percentages of Neocyprideis, Cytheromorpha , and fresh- water ostracods from the various samples recorded, and a suggested salinity profile. The section starts in a buff limestone with numerous and well-preserved Galba and Planorbina. The ostracods belong to Assemblage I, contain many moult stages and adults, few specimens are broken, Cypridopsis in particular has many carapaces, and there are approximately equal numbers of each valve of each moult stage. This indicates little post-mortem movement, and so is taken to represent the biocoenosis. Immediately above the limestone, the green clay contains many small fragments of Galba ; the most abundant ostracod is Neocyprideis , with common Cytheromorpha , but very rare freshwater ostracods; the second sample is similar. Both of these 416 PALAEONTOLOGY, VOLUME 20 SPECIES SUGGESTED SALINITY •/..-» ASSEMBLAGE II 0-3 III 3-9 IV 9-16-5 16-5-25 - 33 VI 33 Candona cliffendensis SP.NOV. Cypridopsis hessani hantonensis SUBSP.NOV. Cypria dorsalta MALZ 8- MOAYEDPOUR Cypris sp. Dorwinula sp. Candona sp. , Strandesia cf. spinosa STCHEPINSKY Virgatocypris edwardsi SP.NOV. Virgatocypris sp. Candona daleyi SP.NOV. Cypridopsis bulbosa (HASKINS) Moenocypris sherborni KEEN Moenocypris reidi KEEN^ Cytheromorpha bulla HASKINS^ Cytheromorpha unisulcata (JONESr = =F Cytherura pulchra SP.NOV. Neocyprideis colwellensis (JONES)~* Neocyprideis wi I liamsoniana (BOSQUET)1 2 3 4 5 6 Cladarocythere hantonensis KEENJ Cladarocy there apostolescui (MARGARIE)^ Brad leya favosa HASKINS Paracypris sp.A. Haplocytheridea debilis (JONES) Clithrocytheridea faboides (BOSQUET) Cushmanidea haskinsi SP.NOV. Cushmanidea stintoni SP.NOV. Cushmanidea wightensis SP.NOV. Cyamocytheridea herbertiana (BOSQUET) Cyamocytheridea purii HASKINS Cyamocytheridea subdeltoidea HASKINS Bradleya forbesi (JONES & SHERBORN) Cytherella pustulosa KEIJ Cytherelloidea lacunosa HASKINS Cytheretta headonensis HASKINS Cytheretta porosacosta KEEN = : = :±:z: = Eocytheropteron wetherelli (JONES) Flexus ludensis KEEN Leg urn inocy there is cancellosa HASKINS Leguminocythereis delirata (JONES & SHERBORN) Loxoconcha sp. Paracytheridea gradata (BOSQUET) Pai jenborchella brevicosta HASKINS Pterygocythereis pustulosa HASKINS Schuleridea perforata headonensis SUBSP.NOV. Cytheretta aff. C.stigmosa TRIEBEL Haplocytheridea mantelli KEEN Leguminocythereis cf. L. striatopunctata (ROEMER) Pokornyella osnabrugensis (LIENENK LAUS) Ruggeria semireticulata HASKINS Krithe bartonensis (JONES) Bairdia sp. Cytherella cf. C. compressa (VON MUNSTER) Cytherella sp. Hazelina indigene MOOS ? Idiocythere bartoniana HASKINS Pterygocythereis cf. P. fimbriata (VON MUNSTER) E = = El: ~ 1 Lower Headon Beds only 2 Upper Headon & Osborne Beds 3 Headon Beds 4 Osborne & Bembridge Beds 5 Headon & Osborne Beds 6 Bembridge Beds Always present, usually abundant Usually present Rare Present as thanatocoenosis. text-fig. 4. Ostracod assemblages from the Headon Bembridge Beds, with suggested salinities. KEEN: EOCENE OSTRACOD ASSEMBLAGES 417 ostracods are represented by several different moult stages and contain similar numbers of each valve; these therefore represent the biocoenosis. The freshwater ostracods and molluscs are probably derived. Ascending the succession, Neo- cyprideis declines in numbers and eventually disappears, while Cytheromorpha MACROFAUNA-SECTION — SAMPLES -% COMPOSITION 0 100 C,G,S CJ C, T, M, G G, P G, P I I I I I I I i r~ i KEY: » • • • r~r Limestone Clay Scale of Shell bed section 7 5cm. 1 0 C Corbicu/a G Ga/ba M Melania SALINITY 0°/oo 5%o 10%o 15% P Plan orb ina T Theoc/oxus S Serpu/a * Occurs as shell fragments text-fig. 5. Section through the Cyrena pulchra Bed of Headon Hill, showing composition of ostracod fauna and suggested salinity. becomes dominant. Towards the top the freshwater ostracods become common, forming up to 50% of the samples, but they are mainly represented by broken valves, with very few adults, while the freshwater molluscs are also fragmentary. They therefore represent fluvially transported material and belong to the thanatocoenosis. The section is interpreted as showing a shallow freshwater lake (Assemblage I) 418 PALAEONTOLOGY, VOLUME 20 situated near a coastline; the sea breached a barrier some distance away, leading to a rapid rise in salinity as the lake became converted into a lagoon. This caused the death of the freshwater animals and the introduction of a mesohaline fauna (Assem- blage IV). The sea connection was short lived, the lagoon began to silt up, and fresh- water influence increased (Assemblage III). Towards the top of the section fluviatile influence was strong, introducing transported shells into the lagoon. Near the top, the reappearance of Neocyprideis and disappearance of the freshwater ostracods indicates another invasion of the sea, although this was not very long lived and the salinity was soon reduced. Another method of studying the samples is to regard them in terms of ‘end- members’ of gradational series. Four ostracod groups can be thought of as ‘end members’, Haplocytheridea , Neocyprideis , Cytheromorpha , and freshwater ostracods, so that each sample can be described in terms of the percentage occurrence of each of these. Haplocytheridea and the freshwater ostracods are almost mutually exclusive so we are left with a series of one, two, or three component samples which can be studied visually by using triangular diagrams. These are given in text-figs. 6 and 7. The first thing to notice, as would be expected, is that the samples do not plot in a random fashion, i.e. there is no scatter of points all over the diagram. Secondly, the samples show a continuous gradation from freshwater — > Cytheromorpha — > Neo- cyprideis — > Haplocytheridea much as would be expected if the assemblages were controlled by a gradually changing parameter, in this case thought to be salinity. The species present in each sample have been examined, using the criteria outlined above, to determine whether they were more probably biocoenosis or thanatoco- enosis. It is admitted that the conclusions are relatively subjective, but when plotted on the diagrams they help in giving a consistent interpretation. In text-fig. 6 C. bulla, N. colwellensis , and freshwater ostracods have been plotted from fifty-one samples FRESHWATER OSTRACODS text-fig. 6. Triangular diagram of samples belonging to Assemblages I— IV showing percentage composition of freshwater ostracods, Cytheromorpha and Neocyprideis. KEEN: EOCENE OSTRACOD ASSEMBLAGES 419 excluding samples belonging to Assemblage V. It can be seen that Neocyprideis is rarely found in samples dominated by freshwater ostracods: in fact it only occurs once, and then forms only 1% of the fauna; these probably represent valves washed into a freshwater environment. On the other hand, freshwater ostracods are found in many samples dominated by Cytheromorpha and/or Neocyprideis ; these are interpreted as valves carried by rivers into a lagoonal environment. The plots suggest complete gradation along the Cytheromorpha-fresh'water and Cytheromorpha- Neocyprideis axes, indicating considerable overlap between living populations of these two sets of ostracods. In the case of Cytheromorpha and Neocyprideis , the presence of moults and adults, and equal numbers of each valve support this view. In the case of Cytheromorpha and the freshwater ostracods the same criteria suggest that Cytheromorpha could adapt to virtually freshwater conditions, occasionally forming a minor part of the freshwater biocoenosis; the freshwater ostracods in samples dominated by Cytheromorpha seem to belong to the thanatocoenosis. HAPLOCYTHERIDEA text-fig. 7. Triangular diagram of samples belonging to Assemblages III-V show- ing percentage composition of Haplocytheridea, Neocyprideis , and Cytheromorpha. In text-fig. 7 similar methods have been used to define the three fields indicated on the diagram. Thus Cytheromorpha is regarded as belonging to the thanatocoenosis in samples belonging to Assemblage V. These may represent dead valves carried by currents, or possibly stray animals which endured the increase in salinity, but could not breed. Neocyprideis is regarded as a living, breeding, member of the more brackish parts of Assemblage V. One problem which could distort all of the above conclusions is that of rapidly changing conditions leading to remanie contamination. Thus, Neocyprideis might be inhabiting a mesohaline lagoon; a rapid rise in salinity could kill it and introduce a polyhaline fauna. With very slow sedimentation these two populations would become mixed. The example of the Cyrena pulchra Bed given above shows that 420 PALAEONTOLOGY, VOLUME 20 changing conditions can be recognized, at least in some cases, by detailed sampling. The case of Neocyprideis in Assemblage V needs some examination because it is often one of the most abundant constituents in samples from Headon Hill and Colwell Bay. With this in mind, the top 60 cm of the Venus Bed at Headon Hill were examined by taking thirty-three samples covering varying thicknesses of between 1 and 3 cm. The lowest 9 cm consisted of green clay, apparently without macrofossils, but containing fairly numerous Candona , represented by several different moult stages, some Cypridopsis and a few Virgatocypris, i.e. Assemblage 1. The clay was extensively cut by burrows filled with the overlying sediment, so great care had to be taken to avoid contamination of samples. The overlying bed was a sandy clay, grading upwards into a fine sand, but extremely fossiliferous throughout. The burrows into the freshwater clay indicate that the clay was quite hard before burrowing took place, while the lack of burrows in the overlying sand could be taken as evidence of fairly rapid sedimentation. The burrow infills and the lowest 2 cm of this bed contained ostracods of Assemblage V together with freshwater ostracods, which, because of their preservation, were obviously derived from the underlying clay. The remaining 49 cm contained ostracods of Assemblage V, with Neocyprideis evenly distributed throughout. No unusual concentrations of the latter were found, as might be expected if it had only lived at certain times during this interval, and it was always represented by several different moult stages. It could be argued that the changes were so frequent as to destroy any horizons rich in Neocyprideis. Against this is the fact that layering of shells is present in the deposits, although bedding is poorly developed. Secondly, we can consider the likely biology of Neocyprideis. If it was similar to Cyprideis , it probably had a single breeding season per year. Some support for this is indicated by the moult stages forming discrete groupings (text -fig. 11); ostracods which breed more than once a year do not often show this because those hatched during a warmer period grow faster, but are smaller, giving an almost continuous variation in size. If this was the case, Neocyprideis would have needed at least two to three years to establish the viable population indicated by the fossils. It also follows that its presence cannot be explained by seasonal fluctuations, i.e. Neocyprideis living at a time of the year when the salinity was lower. Migration may have been involved but seems unlikely by comparison with living Cyprideis. The evidence therefore suggests that Neocyprideis was a member of Assemblage V. There is considerable variation between samples belonging to Assemblage V. Certain features are recognizable, however. Nearly all the samples are dominated by H. debilis, and Cytheretta porosacosta is usually next in abundance. There are notable differences between samples from Whitecliff Bay and those from the west of the Isle of Wight; Clithrocytheridea, Cyan ? o cy I her idea , and Schuleridea are rare in the east, but abundant in the west, while Neocyprideis is absent in the east. These genera are rather brackish water ones compared with the faunas found in the east. The differences between samples are probably due to differing salinities within the polyhaline range. KEEN: EOCENE OSTRACOD ASSEMBLAGES 421 MORPHOLOGICAL VARIATION OF NEOCYPRIDEIS COLWELLENSIS Brackish water ostracods show considerable variation in size and ornamentation related to variation in salinity. N. colwellensis was examined to see whether such variation offered any help in interpretation of environmental conditions. The results are given in Table 2. table 2. Distribution of size and nodosity in Neocyprideis colwellensis (Jones). Numbers in brackets are the number of measured female left valves; these numbers are low because adults normally form only about 10% of the preserved population. For discussion see text; for locality details see Appendix. NUMBER OF MEAN NODOSITY SAMPLE LENGTH ASSEMBLAGE SPECIMENS $L WB 63 58 (2) 87 IV HH 58 240 (5) 85 IV CB 1 2 (2) 84 V ADULTS AND BR 1 6 (1) •84 V VB 18 5 (5) 80 V INSTARS SMOOTH HHCP 2 104 (12) 78 IV NHH 4 66 (5) 75 IV HH 44 HH 42 16 (3) 20 (1) 89 87 V V ADULTS SMOOTH, HH 22 56 (3) 85 IV OCCASIONAL MF 1 17 (2) 84 V NODOSE INSTARS NHH 3 65 (3) 77 IV HH 51 57 (5) ■80 IV NHH 2 64 (2) 80 IV OCCASIONAL NODOSE ADULTS, HH 47 140 (10) 79 IV NHH 1 101 (11) •79 IV MOST INSTARS NODOSE HH 48 337 (21) 76 IV HH 37 202 (13) 7 7 IV ALL COMMONLY NODOSE Noding Cvprideis and related genera develop hollow nodes or tubercles as a response to decreasing salinities. A salinity of 5°/00 is often quoted as the figure below which noded forms become dominant. Vesper (1972u) has confirmed this in a detailed study of C. torosa (Jones); he found occasional noded adults in salinities up to 14%0, but between 6%0 and 2°/00 they dramatically increased to form 85% of the population. The nodes are genetically controlled because they develop in specific places; Kilenyi (1972) believes that selection operates in favour of noded forms as the salinity decreases, eventually giving balanced polymorphism. With regard to N. colwellensis , only one sample (HH37) contains a high proportion of noded adults; this sample is intermediate between Assemblages III and IV. Noded adults are rare or absent in all other samples, although some samples of Assemblage IV have 422 PALAEONTOLOGY, VOLUME 20 common noded juveniles. Thus, if N. colwellensis reacted to decreasing salinity in the same way as C. torosa , it was never found in salinities below 5%0, lending support to the salinity suggested for Assemblage IV. Size There is some controversy over the relationship between size and salinity in ostracods; a recent review can be found in Van Harten (1975). Barker (1963) indicated a decrease in size with decreasing salinity in the ostracods Leptocythere and Loxo- concha. These were marine forms adapting to an estuarine environment; with euryhaline species Hartmann (in Keen 1971) stated that size is independent of salinity. Vesper ( 1912b) appeared to confirm this in a study of C. torosa , suggesting a positive relationship between size and food supply. However, Van Harten (1975) found a significant negative correlation between size and salinity in this species, but the change was not gradual because populations from salinities below 5%0 were markedly larger than those from above. In this respect the increase in size appears to behave similarly to the increase in nodosity. N. colwellensis shows no clear relationship between size and suggested salinity. In general, specimens from Assemblage V fall into the upper part of the size range while those from Assemblage IV occur through- out the range. This, as with noding, may be because N. colwellensis never inhabited salinities below 5°/00, in which case the variation in size may be positively related to food supply. Haskins (1969) observed that the largest specimens came from a sample collected immediately above the Bembridge Limestone, and postulated that this may have been related to an abundance of calcium carbonate in the water. Applying this to the present study this hypothesis seems unlikely. The largest specimens do come from lime-rich sediments, but so do relatively small ones; and the only pure limestone (HH58) yields individuals of about average size. Finally, although perhaps not strictly relevant to a brackish water animal, Reyment and Brannstrom (1963) found that in laboratory studies of the freshwater ostracod Cypridopsis those reared in abnormally high lime-rich waters or stagnant waters grew to a smaller size than those from the normal environment. Variation in ornamentation Specimens of N. colwellensis may be smooth or punctate. Five samples contained smooth forms, all coming from calcareous horizons. It is postulated that smooth forms are phenotypes which developed in lime-rich waters. LIVING CONDITIONS OF THE BRACKISH WATER ASSEMBLAGES The salinities indicated above are found in both estuarine and lagoonal areas and it is often difficult to differentiate between these in the geological record. Estuaries are high-energy environments with a great deal of transportation of material both from the sea and from the river, which leads to mixing of faunas. Freshwater faunas give way to brackish, which in turn give way to marine faunas; freshwater shells may be found throughout. This, in fact, has been found to be the case in the Headon Beds. Lagoons and bays have a more restricted access to the sea and are generally KEEN: EOCENE OSTRACOD ASSEMBLAGES 423 bodies of very shallow water with gradual salinity changes between lagoons. The fauna may give some indication as to which of these conditions prevailed. In the polyhaline part of modern estuaries the marine element is often represented by large numbers of species, but with few specimens of each, although it is impossible to generalize because much depends upon the size of the estuary (e.g. Wagner 1957; Kilenyi 1969; Whatley and Kaye 1971). Another imponderable is the tidal range, which affects the amount of transportation from the sea, and is itself determined by the palaeogeography. The polyhaline fauna of bays and lagoons is more restricted in numbers of species, many of which have their main habitat in the lagoons, and in that sense can be regarded as being more specialized than estuarine faunas. The fauna of the Headon Beds appears more similar to a lagoonal fauna than an estuarine one. Estuaries may, of course, enter into lagoons before reaching the sea, and many coastlines are a complex mixture of bays, lagoons, and estuaries. The procedure adopted is to look for a model in the present world, the most suitable being the complex system of bays and lagoons developed along the Texan coast. These have been more intensively studied than other similar coastlines. The Hamp- shire Basin in late Eocene times exhibited features similar to those of Texas. Ladd et al. (1957) considered that the best classification of the lagoons was based on salinity, giving closed bays (0-20%o) and polyhaline bays (> 16-5%0). The latter are ‘front bays’ directly connected with the Gulf of Mexico by openings between sandy barrier islands, while the closed bays have no direct connection. Coastal streams drain into the lagoons, and salinity changes are gradual between the various bays. Short-term fluctuations occur: due to periods of drought, giving higher salinities; periods of high rainfall and floods, giving lower salinities; and periods of hurricanes and storms, when the sea may be swept into the lagoons. These rapid changes lead to the mass mortality of many of the animals inhabiting the waters of the lagoons. The lagoons are very shallow, mostly less than 9 ft deep. Studies in Florida show similar features; at Cape Romano (Benda and Puri 1962) a barrier of mangrove islands separates lagoons from the open sea. The lagoons are about 4 ft deep, with a salinity of 1 3-43°/00 ; the mangrove swamps have numerous shallow baylets and inlets, with a salinity of 23-40°/oo. Ostracod assemblages from these areas have been studied by Swain (1955; San Antonio Bay, a closed lagoon, Texas); Curtis (1960; Mississippi Delta); Benda and Puri (1962; Cape Romano, Florida); Hulings and Puri (1964; Florida); Engel and Swain (1967; Texas Gulf coast). Assemblages III and IV are characteristic of the typical Bay Facies of Swain and the ‘closed shallow lagoon, away from fluvial influence and with little marine influence’ of Curtis. For reasons already given. Assemblage III is believed to represent a lower mesohaline environment (3-9°/00) and Assemblage IV higher mesohaline (9-0- 1 6-5°/00). These would then correspond to the Upper Bay and Mid Bay subfacies respectively of Swain. Assemblage V corresponds well with the Lower Bay subfacies of Swain and the ‘open lagoon, small fluvial influence; marine influence prominent’ of Curtis. This is the polyhaline facies of Ladd et al., where the lagoon has a connection with the open sea. The environment envisaged for these deposits, therefore, is a series of lagoons, some open, i.e. with a direct sea connection, others closed, i.e. no direct connection 424 PALAEONTOLOGY, VOLUME 20 with the sea (text-fig. 8). In the polyhaline parts of the open lagoons there existed a fairly rich ostracod fauna, varying according to proximity to the sea. Oyster banks were also extensively developed. Further away from the sea connections, the water became progressively less saline, so that areas of Assemblages III and IV would be met. The fauna became poorer in number of species, but was very rich in individuals. In the mesohaline parts of the open lagoons it is probable that the ‘banks’ of Bati/laria ventricosa would have been found, with the ostracod Bradleya favosa. Samples with worn molluscan shells and common Haploeytheridea and Neocyprideis might repre- sent a channel connection between a closed and an open lagoon. SALINITY PROFILES Bearing in mind all the foregoing discussion, it is possible to allocate a salinity value to each sample. These have been plotted in text-fig. 9 to give salinity profiles for the major sections. Such profiles could be useful for correlation (see below). DISTRIBUTION OF THE ASSEMBLAGES The difficulty with such variable beds as these is that correlations tend to be based on facies rather than strict chronological units. For example, the Venus Beds of KEEN: EOCENE OSTRACOD ASSEMBLAGES 425 text-fig. 9. Salinity profiles determined from ostracod assemblages. 426 PALAEONTOLOGY, VOLUME 20 Whitecliff Bay and Headon Hill are usually correlated; but in fact it is quite possible that the Venus Bed of Headon Hill is a time equivalent of the Brockenhurst Bed and Psammobia Bed. This would correlate the "most marine’ deposits of the eastern and western parts of the basin, and accepts that the marine influence becomes stronger eastwards. This is particularly apparent in the salinity profiles. At both ends of the island a fairly rapid increase in salinity is indicated, followed by a period of slightly lower salinity before returning to freshwater conditions. In the east the high salinity phase is 35%0, the lower 25%0, in the west 30%o and !6%0. Curry (1965) has in fact suggested that the Venus Bed facies was deposited in lagoons around the Brocken- hurst Sea, and Vella (1969) has correlated the Brockenhurst Bed with the Neritina Bed of Colwell Bay, and the Roydon Zone of Whitecliff Bay with the Venus Bed of Colwell Bay. The distribution of Assemblage VI suggests that true marine conditions were only present at Whitecliff Bay and in the New Forest. Assemblage V on the other hand, is present at these localities and also at Colwell Bay, Milford, and Headon Hill. Mesohaline elements such as Neocyprideis are totally lacking in Assemblage V at Whitecliff Bay, while they are present at the other places, suggesting less marine influence in the west. This is also borne out by the Bembridge Oyster Bed, which contains a weakly developed Assemblage V at Whitecliff Bay, and Assemblages III and IV at Bouldnor Cliff. This bed is probably the nearest there is to a time plane throughout the whole sequence. Assemblage IV is mainly present at the western end of the Isle of Wight. In the east it is only found in the Lower Headon Beds and the Bembridge Oyster Bed. Its principal development is in the upper part of the Middle Headon Beds and the Upper Headon Beds of Headon Hill, and it is possible that these are partly time equivalents of the development of Assemblage V in the east (i.e. the Venus Bed). Assemblage III has its main occurrence in the Upper Headon Beds and Osborne Beds of the east, but it is present in the Lower Headon Beds of the west. Once again, it is quite possible that the development of Assemblage III in the east occurred at the same time as that of Assemblages I and II in the west. Assemblage I is found in the Lower and Upper Headon Beds of the west and the Bembridge Limestone, while Assemblage II is found in the Lower Headon Beds of Paddy’s Gap (Milford), and the Upper Headon Beds. COMPARISON WITH FOR AMINIFERAL STUDIES The most detailed works published to date are those of Bhatia (1957) and Murray and Wright (1974) concerning the foraminiferal faunas. Bhatia regarded the fauna of the Brockenhurst Beds as indicating deeper-water infraneritic conditions shallow- ing in the Psammobia Beds. The Venus Bed was taken to indicate shallow epineritic becoming lagoonal and less saline. Murray and Wright have indicated that more modern data favours depths shallower than these. They took the foraminiferids of the Brockenhurst Beds, dominated by Quinqueloculina , Triloculina , and Cibicides, to indicate the seaward part of an enclosed lagoon with normal or slightly hyper- saline salinities; planktonic foraminiferids present in these beds were thought to have been transported into the lagoon from the open sea. The Venus Bed has a low KEEN: EOCENE OSTRACOD ASSEMBLAGES 427 diversity of foraminiferids, and at Whitecliff Bay the succession was interpreted as showing the inner part of a hyposaline lagoon, giving way to a normal marine lagoon. At Headon Hill the foraminiferids are small, usually juveniles, so an abnormal environment was postulated, with intertidal lagoons, salinity > 32 °/00 and hypo- saline subtidal lagoons or estuaries. The limestones of the Upper Headon Beds at Headon Hill yielded foraminiferids at three horizons which were taken to indicate a lagoonal environment with salinity > 32°/00. The dominant genera are Rosalind and Quinqueloculina. The former was also found near the top of the lowermost limestone of the Bembridge Limestone at Whitecliff Bay, and interpreted as indicat- ing very unfavourable conditions with a salinity < 10°/oo. The lowest part of the Bembridge Marls at Whitecliff Bay were interpreted as hyposaline lagoonal deposits, or an estuary, salinity 10-25°/oo. text-fig. 10. The palaeogeography of the Hampshire Basin during the time of the deposition of the Brockenhurst Beds. This is intended to show the relationship between the main geographical units, and is not an exact geo- graphy. The boundaries between the coastal plain, lagoons, and sea fluctuated considerably, so that the positions of the individual lagoons and their outlets are conjectural. Evidence has been used from many sources, both consciously and subconsciously, incorporating stratigraphy, structure, sedimentation, and palaeontology. In particular, it owes much to the work of Dr. N. Edwards, although of course he may not entirely agree with the end result, and the responsibility is mine. There seems to be close agreement between these results and those of the present study. There is a difference in interpretation of the Brockenhurst Beds, but this may be semantic because their text-fig. 28 indicates a similar interpretation to text-fig. 10 of this study. Within the hyosaline environment the ostracods offer more scope for finer detail than the Foraminifera, so that rather more environmental assemblages can be recognized. A second point of apparent disagreement is with the Upper Headon limestones of Headon Hill; Murray and Wright indicate polyhaline or 428 PALAEONTOLOGY, VOLUME 20 marine salinities, while this study indicates predominantly freshwater conditions. However, it has to be borne in mind that thin beds exist within the limestones that have yielded mesohaline ostracods (see text-fig. 1), while the main mass of limestone, with its abundant freshwater molluscan fauna, has not yielded any Foraminifera. TAXONOMY A study such as this inevitably leads to the discovery of new species. These are described in the following section. The faunal list given by Keen (1968) can now be revised as follows: 1968 name Candona sp. A Candona sp. B Cypridopsis sp. A ‘ Eucypris ’ sp. A Neocyprideis cf. williamsoniana Brachycythere sp. A Cushmanidea sp. A Cushmanidea sp. B Haplocytheridea sp. A Limnocythere sp. A Paracytheridea sp. A Schuleridea sp. A Present name Candona cliffendensis sp. nov. Candona daleyi sp. nov. Cypridopsis hessani hantonensis subsp. nov. Virgatocypris edwardsi sp. nov. Neocyprideis colwellensis (Jones) Ruggieria semireticulata Haskins Cushmanidea haskinsi sp. nov. Cushmanidea wightensis sp. nov. Haplocytheridea mantelli Keen Cladarocy there hantonensis Keen Paracytheridea gradata (Bosquet) Schuleridea perforata headonensis subsp. nov. SYSTEMATIC DESCRIPTIONS The classification is that of the Treatise (Moore 1961). In the descriptions, the heading ‘ Material ’ refers to the total sample examined; the numbers refer to individual specimens deposited at the British Museum (Natural History), prefixed Io. Abbreviations. L = length, H = height, W = width; measurements are in millimetres. Order podocopida Muller, 1894 Suborder podocopina Sars, 1866 Superfamily cypridacea Baird, 1845 Family cyprididae Baird, 1845 Genus cypridopsis Brady, 1868 Cypridopsis hessani Carbonnel and Ritzkowski, 1969 Cypridopsis hessani hantonensis subsp. nov. Plate 46, figs. 9- 1 1 1975 Cypridopsis sp. A, Keen, p. 275. Derivation of name. After Hampshire. Type locality. Upper Headon Beds, Headon Hill (HH53). Distribution. Sample HH53 only. Holotype. Io 6736, female left valve. Material. Over 100 valves. Registered specimens Io 6736-6739. KEEN: EOCENE OSTRACOD ASSEMBLAGES 429 Dimensions. Holotype, female left valve, Io 6736: L, 0-44; H, 0-25; L/H, 1-76; W/2, O il. Female right valve, Io 6737: L, 0-43; H, 0-24. Male left valve, Io 6738: L, 0-44; H, 0-23; L/H, 1-91; W/2, 010. Male right valve, Io 6739: L, 0-44; H, 0-23. Diagnosis. Small, elongate, small denticles along anterior margin of right valve. Description and discussion. The nominate subspecies was described from the Melanienton (Rupelian) of Hesse, Germany; it differs from the present subspecies by its large size (0-53 cf. 0-44) and in details of lateral outline. Carbonnel and Ritz- kowski were unable to recognize sexual dimorphism; the L/H ratio given by them, 1 -88, is closer to that of the male in the Headon samples, but as there are differences in lateral outline, and because it is virtually unknown to have a sample consisting of males only, it is advisable to compare the nominate subspecies with both sexes. The female left valve of C. hessani hantonensis differs in having a well-defined highest point with an evenly curved antero-dorsal margin; the male left valve has a more evenly curved dorsal margin, lacking the cardinal angles of C. hessani hessani. Sexual dimorphism is not as prominent in the right valve; the ventral margin is more concave than in C. h. hessani. Internal details as for the genus; the central muscle scars cannot be clearly seen, but appear to be typical of Cypridopsis ; large anterior vestibule present, with numerous short radial pore canals. Genus virgatocypris Malz and Moayedpour, 1973 Virgatocypris edwardsi sp. nov. Plate 47, figs. 3, 6, 9, 12 19686 IScottia sp. Haskins, p. 5, pi. 1, figs. 7-9. 1975 Eucypris cf. tenuistriata (Dollfus); Keen, p. 272. Derivation of name. In honour of Dr. N. Edwards. Type locality. Lower Headon Beds, Headon Hill (HH23). Distribution. Lower and Upper Headon Beds of Headon Hill and Colwell Bay. Holotype. Io 6740, a left valve. Material. Eight complete adult valves, several fragments, forty-five larval stages. Registered specimens Io 6740-6743. Dimensions. Holotype, left valve, Io 6740: L, I 13; H, 0-65; L/H, 1-72. Right valve, Io 6741: L, 116; H, 0-66; L/H, 1-76. Diagnosis. Dorsal margin with prominent, pointed, highest point; ventral margin convex. Description. See Plate 47, figs, 3, 9 for lateral outline. The valves are swollen ventrally. The surface of the valve has fine longitudinal striations; no unornamented specimens have been observed. Internally features are as for the genus: a conspicuous selvage, large anterior vestibule, numerous straight radial pore canals; the central muscle scars have not been clearly observed. N 430 PALAEONTOLOGY, VOLUME 20 Discussion. This differs from other species principally in lateral outline. V. tenui- striata (Dollfus) (Oligocene, Paris Basin) is more elongate and ovoid; V. straubi (Carbonnel and Ritzkowski) (Oligocene, Hesse) has its highest point nearer anterior and has a pronounced concave ventral margin in the right valve; V. virgata Malz and Moayedpour (Miocene, Germany) is more elongate, has a straighter dorsal margin, and a straight ventral margin; V. grisyensis (Margerie) (Upper Eocene, Paris Basin) is smaller, has a weakly developed selvage, and has its highest point placed medianly giving a symmetrically curved dorsal margin. The exact pattern of striations also appears to differ amongst these species. The specimens of Virgatocypris from the Bembridge Limestone have been left in open nomenclature. They differ from the Headon forms in being slightly more elongate, having a less-pointed dorsal margin, and a slightly more tapered posterior. Jones and Sherborn (1889, p. 43) described a new species from the Bembridge Lime- stone of West Cowes, Pseudocythere bristovi, which might be a species of Virgato- cypris. Two specimens of this species are preserved in the collections of the Institute of Geological Sciences, Mik (T) 712 001 and 712 002, neither being type material. The specimens are incomplete, in limestone, and their identity difficult to determine; they are not the same as the Headon species, however. Lamily candonidae Kaufmann, 1900 Genus candona Baird, 1845 Candona cliffendensis sp. nov. Plate 46, figs. 6-8 71889 Pontocypris (?) sp. Jones and Sherborn, p. 16, pi. 1, fig. 13a. 19686 ICandonopsis sp. Haskins, p. 7, pi. 2, figs. 23-27. 1975 Candona ( Pseudocandona ) sp. B, Keen, p. 272. Derivation of name. After the type locality. Type locality. Upper Headon Beds, Cliff End (Colwell Bay, CB12). Distribution. Lower and Upper Headon Beds of Headon Hill and Colwell Bay; Osborne and Bembridge Beds of Whitecliff Bay. Holotype. Io 6744, a left valve. EXPLANATION OF PLATE 46 Figs. 1-5. Candona daleyi sp. nov. 1, 2, holotype, Io 6747, Lower Headon Beds of Headon Hill. 1, right valve internal view, x70. 2, central muscle scars, x450. 3-5, Upper Headon Beds of Colwell Bay (CB12), x 70. 3, left valve of moult no. 7, Io 6749. 4, right valve of moult no. 7, Io 6750. 5. left valve of moult no. 8, Io 6748. Figs. 6-8. Candona cliffendensis sp. nov. 6, Upper Headon Beds of Headon Hill (HH53), left valve of moult no. 7, Io 6746, x70. 7, 8, Upper Headon Beds of Colwell Bay (CB12). 7, right valve internal view, specimen destroyed, x 70. 8, stereo-pair, right valve of holotype, x 50. Figs. 9-11. Cypridopsis hessani hantonensis subsp. nov., Upper Headon Beds of Headon Hill (HH53), x 140. 9, holotype, female left valve, Io 6736. 10, male left valve, internal view, Io 6738. 11, female right valve, Io 6737. Fig. 12. Cypria dorsalta Malz and Moayedpour, left valve, internal view, Io 6734, Lower Headon Beds of Headon Hill (HH30), x 125 (L = 0-54 mm). PLATE 46 KEEN, Upper Eocene ostracods 432 PALAEONTOLOGY, VOLUME 20 Material. Six adult valves, forty larval stages. Registered specimens Io 6744-6746. Dimensions. Holotype, right valve, Io 6744: L, 0-84; H, 0-41 ; L/H, 205; left valve, Io 6745: L, 0-85; H, 0-42; L/H, 2 02. Larval 8, right valve, Io 6746: L, 0-68; H, 0-34; L/H, 2 00. Larval 7, right valve, Io 6746: L, 0-50; H, 0-24; L/H, 2-05. Diagnosis and description. For lateral outline see Plate 46, figs. 7-9. L/H 2-00; larval stages smooth. Internal details as for the genus. Discussion. This is distinguished from other Tertiary species by its elongate shape and smooth larval stages (see C. daleyi below for discussion). The anterior and posterior cardinal angles suggest it belongs to Candona rather than Candonopsis Vavra, although the elongate shape is more characteristic of the latter. Candona daleyi sp. nov. Plate 46, figs. 1 5 19686 Potamocypris sp. Haskins, p. 6, pi. 1, figs. 17-22. 1975 Candona ( Pseudocandona ) sp. A, Keen, p. 272. Derivation of name. In honour of Dr. B. F. Daley. Type locality. Lower Headon Beds, Headon Hill (HH23). Distribution. Lower and Upper Headon Beds and Osborne Beds of Headon Hill, Colwell Bay, Whitecliff Bay; Middle Headon Beds of Headon Hill; Bembridge Limestone of Whitecliff Bay and Bouldnor Cliff. Holotype. Io 6747, a right valve. Material. Two adult right valves, numerous larval stages. Registered specimens Io 6747-6750. Dimensions. Holotype, right valve, Io 6747; L, 0-87; H, 0-55 ; L/H, 1 - 58. Larval 8, left valve, Io 6748: L, 0-81; H, 0-52. Larval 8, right valve: L, 0-75; H, 0-46. Larval 7, left valve; Io 4749: L, 0-58; H, 0-36. Larval 7, right valve: Io 4750: L, 0-53; H, 0-33. Diagnosis. Highest point near posterior; distinct antero-dorsal notch seen in lateral view; L/H 1-58; larval stages punctate. EXPLANATION OF PLATE 47 Figs. 1, 2. Cushmanidea stintoni sp. nov., x 120, Middle Headon Beds of Headon Hill. 1, left valve of female, holotype, Io 6766. 2, right valve of female, Io 6767. Figs. 3, 6, 9, 12. Virgatocypris edwardsi sp. nov. 3, 6, holotype, Io 6740, Lower Headon Beds of Headon Hill. 3, stereo-pair, left valve, x 48. 6, stereo-pair, enlargement showing simple normal pore canals with rims, x300. 9, 12, Upper Headon Beds of Colwell Bay (CB12). 9, right valve, Io 6741, x48. 12, internal view of right valve, Io 6742, x 55. Fig. 4. Cushmanidea haskinsi sp. nov., x 84, left valve of female, holotype, Io 6761, Middle Headon Beds of Headon Hill. Figs. 5, 7, 10, 13. Cushmanidea wightensis sp. nov., x70. Middle Headon Beds of Headon Hill. 5, left valve of male, Io 6773. 7, left valve of female, holotype, Io 6771. 10, right valve of female, Io 6772. 13, internal view of male left valve, Io 6774. Fig. 8. Loxoconcha sp., L = 0-48, left valve, Io 6783, Middle Headon Beds of Headon Hill. Figs. 11, 14. Cytherura pulchra sp. nov., X 100, Lower Headon Beds of Headon Hill. 11, right valve, Io 6777 . 14, left valve, holotype, Io 6776. Fig. 15. ^.Idiocy there bartoniana Haskins, L = 0-47, right valve, Io 6781, Brockenhurst Beds of Whitecliff Bay. PLATE 47 KEEN, Upper Eocene ostracods 434 PALAEONTOLOGY, VOLUME 20 Description. For lateral outline see Plate 46, fig. 1. Adult right valve has a prominent rim around the anterior, ventral, and posterior margins. Surface is smooth in adult, punctate in larval stages. Internal details as for the genus. The right valve of the penultimate larval stage has a similar outline to the adult, including the antero- dorsal notch; the left valve of the penultimate larval stage may therefore indicate the outline of the left valve of the adult. Discussion. According to Triebel (1963) this species should be placed in the sub- genus Candona {Pseudo candona) Kaufmann, 1900, on account of the punctate larval stages. However, recent opinion (especially Dr. R. Meyrick, pers. comm. 1974) would only separate these on soft-part morphology. Thus, no subgeneric status is designated. C. daleyi differs from C.fertilis Triebel in lateral outline, especially the presence of the antero-dorsal notch, and in the presence of a marginal rim. The larval stages of this species are very common in freshwater sediments of the Headon Bembridge series; they are distinguished from larval C. cliffendensis by being shorter and punctate, and from the undescribed species from the Hamstead Beds (Keen 1972a, p. 283, pi. 47, fig. 8) by being shorter. Superfamily cytheracea Baird, 1850 Family cytherideidae Sars, 1925 Subfamily cytherideinae Sars, 1925 Genus neocyprideis Apostolescu, 1956 Neocyprideis colwellensis (Jones) Plate 48, figs. 5-12 1857 Cytherideis colwellensis Jones (pars), p. 49, pi. 14, fig. 13. 1889 Cytherideis colwellensis Jones; Jones and Sherborn (pars), p. 45. 1960 Neocyprideis aff. williamsoniana (Bosquet); Kollmann, p. 177, pi. 3, fig. 2 a-b\ pi. 12, fig. 8; pi. 20, figs. 8-10. 1969 Cyprideis ( Neocyprideis ) williamsoniana (Bosquet); Haskins, p. 155, pi. 4, figs. 10-20. Type locality and distribution. Colwell Bay; Headon-Osborne Beds. EXPLANATION OF PLATE 48 Fig. 1. Ruggieria semireticulata Haskins, L = 0-88, left valve, Io 6780, Middle Headon Beds of Whitecliff Bay (WB22). Fig. 2. Pokornyella osnabrugensis (Lienenklaus), L = 0-59, right valve, Io 6782, Middle Headon Beds of Headon Hilf (HH43). Figs. 3, 4. Hazelina indigena Moos, Brockenhurst Beds of Whitecliff Bay. 3, right valve of female, Io 6778, L = 0-54. 4, right valve of male, Io 6779, L = 0-57. Figs. 5-12. Neocyprideis colwellensis (Jones), all specimens from Headon Hill, figs. 5, 1 1 from Upper Headon Beds (HH58), remainder from Middle Headon Beds (HH48). 5, left valve of male of a smooth form, Io 6753, x 55. 6, stereo-pair, right valve of female punctate form, Io 6751, X 55. 7, stereo-pair, right valve of male punctate form, Io 6752, x 55. 8, stereo-pair, left valve of moult no. 8 showing develop- ment of nodes, note that each ‘node’ is in fact a cluster of small nodes, Io 6755, x 75. 9, detail of anterior part of hinge of Io 6754 (see also fig. 10), x 350. 10, internal view of female right valve, Io 6754, x 55. 1 1, sieve type normal pore canal of Io 6753 (see also fig. 5), x 1300. 12, central muscle scars of Io 6754 (see also fig. 10), x 300. PLATE 48 KEEN, Upper Eocene ostracods 436 PALAEONTOLOGY, VOLUME 20 Lectotype. BM(NH), I 6431 (13). Jones figured two species under the name Cytherideis colwellensis', one has subsequently been described as Cytheretta rhenana headonensis Haskins, 1968. Keen (1972a) and Malz (1973) independently chose the remaining specimen as the lectotype of Neocyprideis colwellensis. The specimen is a larval stage, but adult topotype material is abundant. Material. Registered specimens Io 6751-6755. Diagnosis. Has similar-sized valves, males and females have similar L/H ratios, female carapace inflated at posterior end; anterior and median hinge elements of equal length; twenty anterior radial pore canals; surface with dorsal sulcus. Discussion. There are nine described species of Neocyprideis ranging from the upper Cretaceous to the Miocene. Many of these form part of a continuous phyletic lineage, which makes specific determination difficult unless there is sufficient material for population studies. The two species most closely resembling N. colwellensis are N. apostolescui (Keij), thought to be its ancestor, and N. williamsoniana (Bosquet), which is probably its descendant. Specimens from the Bembridge Beds have been placed in the latter. N. apostolescui differs in having a more evenly rounded anterior end when seen in lateral view, has a more curved anterior hinge element which is shorter than the median element, and has fewer anterior radial pore canals (16-19 cf. 1 7-24). Keij ( 1 957) used the absence of a punctate surface to differentiate N. aposto- lescui from N. williamsoniana ; as indicated earlier, this character is probably pheno- typic, and all species of Neocyprideis may be smooth or punctate. N. williamsoniana differs from N. colwellensis in having more clearly differentiated sexes using L/H measurements, a more angular postero-dorsal angle, especially when viewed from the inside of the valve, a weaker dorsal sulcus, and the posterior inflation of the female carapace is situated in a more dorsal position. Noding. N. colwellensis has ten node positions (text-fig. 11), although several of these are frequently joined by the development of smaller nodes between them. A node may be a single tubercle, or a cluster of small tubercles (PI. 48, fig. 8). All nodes are present in the adult and instars 7 and 8, but earlier instars have fewer nodes, although there does not seem to be any orderly appearance of nodes during ontogeny. Nodes are present on both valves and both sexes with equal frequency. EXPLANATION OF PLATE 49 Figs. 1-4. Schuleridea (Aequacyt her idea) perforata headonensis subsp. nov., x70, Middle Headon Oyster Bed of Colwell Bay. 1, left valve of female, holotype, Io 6756. 2, right valve of female, Io 6757. 3, left valve of male, Io 6759. 4, internal view of left valve of female, Io 6758. Figs. 5, 6. Eocytheropteron wetherelli (Jones), L = 0-70, Middle Headon Beds of Headon Hill (HH43), stereo-pairs. 5, female right valve. 6, normal pore canal showing a larger central pore within the sieve- type pore canal, x 1500. Figs. 7, 8. Leguminocythereis delirata (Jones and Sherborn), L = 0-82, Middle Headon Beds of Headon Hill (HH43). 7, normal pore canals showing a larger central pore within the sieve-type pore canal, x 400. 8, female left valve. Figs. 9, 10. Cytheromorpha bulla Haskins, L 0-64, Lower Headon Beds of Headon Hill (HH26). 9, female left valve. 10, stereo-pair showing sieve-type normal pore canals, X 170. PLATE 49 KEEN, Upper Eocene ostracods 438 PALAEONTOLOGY, VOLUME 20 Larval stages. Six juvenile stages have been recognized; text-fig. 1 1 shows the changes in lateral outline during ontogeny; note the triangular outline with highest point at anterior in the early stages. A punctate surface has been observed in instars 7 and 8, but earlier instars are smooth. The hinge elements are crenulate for instars 6, 7, and 8 and appear to be so for earlier instars. The median element occupies more than 50% of the hinge length in instars 3-7, with a reduced anterior element ; the latter increases for instar 8 and is equal to the median element in the adult. The posterior hinge element retains a similar proportion throughout ontogeny. The number of radial pore canals increases during ontogeny from an average of six in instar 4 to twenty in the adult. These changes during ontogeny parallel the changes seen in phylogeny. ■20 -30 -40 -50 -60 70 -80 -90 Length » TEXT -FIG. 11. Size distribution of Neocyprideis colwellensis (Jones) from the Cyrenapulchra Bed (Lower Headon Beds) of Headon Hill (HHCP2). Genus schuleridea Swartz and Swain, 1946 Subgenus schuleridea (aequacytheridea) Mandlestam, 1947 Schuleridea ( Aequacytheridea ) perforata (Roemer) Schuleridea ( Aequacytheridea ) perforata headonensis subsp. nov. Plate 49, figs. 1 -4 1857 Cytheridea perforata Roemer; Jones, p. 44 (pars), pi. IV, fig. 14 a-e. 1889 Cytheridea perforata Roemer; Jones and Sherborn, p. 39 (pars). 1969 Schuleridea perforata (Roemer); Haskins, p. 161 (pars). Derivation of name. After the Headon Beds. Type locality. Colwell Bay (Oyster Bed). Distribution. Venus Bed of Headon Hill, Colwell Bay, and Whitecliff Bay. Holotype. Io 6756, a female left valve. Material. Forty-seven valves. Registered specimens Io 6756-6760. Dimensions. Mean of eleven female left valves: L, 0-82; H, 0-53; L/H, 1-56. Mean of three male left valves: L, 0-89; H, 0-52; L/H, 1-70. Diagnosis. Pronounced sexual dimorphism, elongate carapace, prominent posterior cardinal angle, no marginal spines or denticles, surface finely punctate. KEEN: EOCENE OSTRACOD ASSEMBLAGES 439 Discussion. According to the literature S. {A.) perforata is one of the most widely ranging species of ostracods in a stratigraphical sense; it is found from the Palaeo- cene to Miocene. Moos (1970) and Malz (1973) have shown how it is possible to recognize distinct species and subspecies by careful study of the lateral outline and ornamentation. Internal features do not vary greatly, and the new subspecies is very similar to the nominate subspecies in this respect. S. {A.) perforata headonensis differs in being more elongate and showing more distinct sexual dimorphism: L/H female 1-56 cf. 1-47 (5”. (A.) perforata perforata from Damery), male 1-70 cf. 1-57. The dorsal margin has a more pronounced posterior cardinal angle; the nomi- nate subspecies almost lacks this, and so has a more tapered and triangular appear- ance. The posterior margin of the nominate subspecies is pointed, which in S. (A.) perforata headonensis it is not. The ventral margin is straight to convex, while the nominate subspecies is straight to concave. The surface is punctate, but this ornamen- tation does not often show very clearly because of the mode of preservation. Subfamily neocytherideidinae Puri, 1957 Genus cushmanidea Blake, 1933 Cushmanidea haskinsi sp. nov. Plate 47, fig. 4 1857 Cytherideis flavida Jones (non Muller), p. 50 (pars). 1889 Cytherideis sp. Jones and Sherborn, p. 45. 71970 Cushmanidea grosjeani Haskins (non Keij), p. 16 (pars), pi. 1, figs. 16, 18, 19, 21. 71970 Cushmanidea sp. Haskins, p. 16, pi. 1, figs. 22-23. Derivation of name. In honour of Dr. C. W. Haskins. Type locality. Middle Headon Beds (Venus Bed), Headon Hill. Distribution. Venus Bed of Headon Hill, Oyster Bed of Colwell Bay, and ?Middle Headon Beds of White- cliff Bay. Holotype. Io 6761, a female left valve. Material. Seventeen valves and carapaces. Registered specimens Io 6761-6765. Dimensions. Female carapace, Io 6764: L, 0-62; H, 0-27; W, 0-25; L/H, 2-30. Male carapace, Io 6765: L, 0-72; H, 0-28; W, 0-26; L/H, 2-57. Diagnosis. Elongate, eight to nine faint ridges parallel to anterior margin, large anterior and narrow posterior vestibules, radial pore canals simple. Description. For outline and ornamentation see figured specimen. Note the eight to nine weak ridges developed in the anterior and antero-ventral areas, and running parallel to the anterior margin. There are some sixty normal pore canals which widen near the outer surface and appear to be of the sieve-type. Internally, all the hinge elements are smooth; the duplicature has large anterior and narrow posterior vestibules, with 26 anterior, 16 ventral, and 7 posterior radial pore canals; the selvage is parallel to the outer margin. The central muscle scars consist of a row of 4 scars, the dorsal-most of which is usually indistinct; the frontal scar is irregular, but tendency towards a V shape; there are 2 small mandibular scars, and some 7 dorsal muscle scars. 440 PALAEONTOLOGY, VOLUME 20 Discussion. C. lithodomoides (Bosquet) has a similar ornamentation, differing by its stronger development and the ridges running parallel to the outer margin so that they are parallel to the ventral margin in their ventral portion. The closest species is C. therwilensis (Oertli) from the upper Oligocene of Switzerland; the ornamenta- tion and internal features are very similar. The main difference is in lateral outline; the highest point is much nearer the posterior in C. therwilensis , and the posterior margin is more pointed. The material described by Haskins probably belongs here although his specimens are larger. The forms referred to as C. grosjeani are probably females, and Cushmanidea sp. the males. Cushmanidea stintoni sp. nov. Plate 47, figs. 1, 2 1970 Cushmanidea grosjeani Haskins (non Keij), (pars), p. 16, pi. 1, figs. 17, 20. Derivation of name. In honour of Mr. F. C. Stinton, a worker on Tertiary otoliths and guide to the Tertiary of Hampshire. Type locality. Middle Headon Beds (Venus Bed), Headon Hill. Distribution. Venus Bed of Headon Hill, Middle Headon Beds and Bembridge Marls Oyster Bed of Whitecliff Bay. Holotype. Io 6766, a female left valve. Material. Twenty-four valves and carapaces. Registered specimens Io 6766-6770. Dimensions. Female carapace, Headon Hill, Io 6768 : L, 0-53; H, 0-23; W, 0-21 ; L/H, 2-30. Female carapace, Whitecliff Bay, Io 6769: L, 0-65; H, 0-27; W, 0-23; L/H, 2-33. Male carapace, Whitecliff Bay, Io 6770: L, 0-69; H, 0-26; W, 0-26; L/H, 2-62. Diagnosis. Small, prominent pits at posterior, medium-sized anterior vestibule, simple radial pore canals. Description. For lateral outline and ornamentation see figured specimens. Each of the pits seen on the surface of the carapace has a sieve-type normal pore canal opening into it. Viewed with an optical microscope, the pits are seen as being pre- dominantly at the posterior, with a lesser development along the venter and anterior. Internally, all hinge elements are smooth, the anterior and median elements of about equal length. The anterior vestibule is of medium size, ventral and posterior vestibules small. There are 35 anterior, 16 ventral, and 10 posterior simple radial pore canals; the selvage is parallel to outer margin. Central muscle scars a row of 4, dorsal-most is large; a single V-shaped frontal scar; 2 small mandibular scars. Sexual dimorphism is apparent in the specimens from Whitecliff Bay, but not in those from Headon Hill; all the latter are presumed to be females. The Whitecliff Bay specimens are also larger, but otherwise identical. Discussion. The most similar species is C. neauphlensis (Apostolescu) from the Lutetian of the Paris Basin. This differs in lateral outline, with its highest point slightly to the anterior of the mid-point, has strong pitting over the entire surface, and a median dorsal sulcus. Keij (1957) placed C. neauphlensis in the synonymy of C. mayeri (Howe and Garrett) from the Eocene of the U.S.A.; certainly the two species are very similar. KEEN: EOCENE OSTRACOD ASSEMBLAGES 441 Cushmanidea wightensis sp. nov. Plate 47, figs. 5, 7, 10, 13 Derivation of name. From the Isle of Wight. Type locality. Middle Headon Beds (Venus Bed), Headon Hill. Distribution. Venus Bed of Headon Hill, Oyster Bed of Colwell Bay. Holotype. Io 677 1 , a female left valve. Material. Twenty-two valves. Registered specimens Io 6771-6775. Dimensions. Female carapace, Io 6775: L, 0-64; H, 0-33 ; W, 0-29; L/H, 1-94. Male left valve, Io 6773: L, 0-77; H, 0-34; L/H, 2-26. Diagnosis. Not particularly elongate, unornamented, hinge elements crenulate, no vestibules, simple radial pore canals. Description. For lateral outline seen figured specimens. Sexual dimorphism is very pronounced, although males are rare (3 out of 22 specimens). Right valve more pointed dorsally. There are 55-60 normal pore canals. Internally, the hinge is crenulate; the anterior element has some 5 coarse denticles, median element very finely crenulate, posterior element coarse, with 6-7 denticles. Ratio of hinge-element lengths, anterior: median: posterior, 3:3:2. There are no vestibules; 16 anterior, 6 ventral, and 4 posterior simple radial pore canals; there is a suggestion in one specimen of a series of blind ventral radial pore canals. Central muscle scars consist of a vertical row of 4, all in contact, with 2 frontal scars very close together and often appearing as a single V-shaped scar. There is a prominent fulcral point, and 2 large mandibular scars. Discussion. There are no other species with which to compare, and it is doubtful whether this is a true Cushmanidea species, because of the lack of vestibules and crenulate hinge. Family cytheruridae G. W. Muller, 1894 Genus cytherura Sars, 1866 Cytherura pulchra sp. nov. Plate 47, figs. 11,14 Derivation of name. After the Cyrena pulchra Bed. Type locality. The Cyrena pulchra Bed, Headon Hill. Distribution. Only recorded in the type deposit of Headon Hill. Holotype. Io 6776, a left valve. Material. Twenty-four valves and several fragments. Dimensions. Holotype, left valve, Io 6776: L, 038; H, 0T8: right valve, Io 6777: L, 039; H, 0-19. Diagnosis. Small, elongate, strong caudal process, surface reticulate, weak longi- tudinal ridges developed, right valve larger. 442 PALAEONTOLOGY, VOLUME 20 Description. Sexual dimorphism cannot be clearly recognized; occasional shorter specimens (0-35 mm) may be females, in which case most of the material consists of males. Although there are not enough specimens to be sure, the right valve appears to be the larger. The surface ornamentation varies in strength between specimens, and is reticulate; there is a tendency for a weak median and ventral ridge to develop. In juveniles the reticulation is more regular, though much weaker, and the surface appears to be punctate. Internally the hinge is as for the genus, the duplicature has a few wavy radial pore canals; the central muscle scars cannot be seen. Discussion. Externally this is similar to several species now placed in Semicytherura: S. dunkeri Moos (Melanienton, North Germany), S. gracilis (Lienenklaus) ( Rupelian, Paris Basin), and S. oedelemensis (Keij) (Eocene-Oligocene). It differs in details of ornamentation and shape, as well as in internal details. Acknowledgements. I thank Drs. B. Daley and N. Edwards for reading a draft of the manuscript and for valuable discussions. I am indebted to Mr. R. Cumberland (Geology Department, Glasgow University) for running the computer program and to Mr. G. McTurk (Geology Department, Leicester University) for the stereoscan photographs. REFERENCES apostolescu , v. 1956. Contribution a l’etude des Ostracodes de l’Eocene inferieur (s.l.) du Bassin de Paris. Revue Inst. fr. Petrole. 11, 1327-1352, pis. 1-4. baird, w. 1845. Arrangement of the British Entomostraca. Hist. Berwicksh. Nat. Club. 145-148. barker, d. 1963. Size in relation to salinity in fossil and recent euryhaline ostracods. J. mar. biol. Ass. U.K. 43, 785-795. benda, w. K. and puri, h. s. 1962. The distribution of Foraminifera and Ostracoda off the Gulf of the Cape Romano area, Florida. Trans. Gulf-Cst Ass. geol. Soc. 12, 303-341. bhatia, s. b. 1957. The Paleoecology of the Late Paleogene sediments of the Isle of Wight, England. Contr. Cushman Fdn. foramin. Res. 8, 11-28. blake, c. 1933. Order Ostracoda in Biol. Surv. Mount Desert Region, dir. W. Proctor. West. Inst. Anat. Biol., pt. V, 229-241, figs. 39-40. bonham-carter, G. f. 1967. Fortram IV program for Q mode cluster analysis of non-quantitative data using IBM 7090/7094 computers. Kansas Geol. Survey Computer Contr. No. 17, 1-28. brady, g. s. 1868. A monograph of the Recent British Ostracoda. Trans. Linn. Soc. Lond. 26, 353-495, pis. 23-41. carbonnel, g. 1969. Les ostracodes du Miocene Rhodanien. Docum. Lab. geol. Fac. Sci. Lyon , No. 32, 1, 228 pp., 16 pis. — and ritzkowski, s. 1969. Ostracodes lacustres de FOligocene. Arch. Sci. Geneve , 22, 55-82, pis. 1-5. cheetham, a. h. and hazel, j. 1969. Binary (presence-absence) similarity coefficients. J. Paleont. 43, 1130-1136. curry, d. 1965. The Palaeogene Beds of South-east England. Proc. Geol. Ass. 76, 151-174. curtis, d. M. 1960. Relation of environmental energy levels and ostracod biofacies in east Mississippi Delta area. Bull. Am. Ass. Petrol. Geol. 44, 471-494. daley, b. 1972. Macroinvertebrate assemblages from the Bembridge Marls (Oligocene) of the Isle of Wight, England, and their environmental significance. Palaeogeogr ., Palaeoclimatol., Palaeoecol. 11, 11-32. 1973. The palaeoenvironment of the Bembridge Marl (Oligocene) of the Isle of Wight, Hampshire. Proc. Geol. Ass. 84, 83-93. — and edwards, n. 1971. Palaeogene warping in the Isle of Wight. Geol. Mag. 108, 399-405. edwards, n. 1967. Oligocene studies in the Hampshire Basin. Thesis, Univ. of Reading, 170 pp., unpublished. - 1971. Stratigraphy and correlation of the Headon, Osborne, Bembridge, and Hamstead Beds (Palaeo- gene), Hampshire Basin— a bibliography (1914-1970). J. Soc. Biblphynat. Hist. 6, 50-60. KEEN: EOCENE OSTRACOD ASSEMBLAGES 443 engel, p. l. and swain, f. m. 1967. Environmental Relationships of Recent Ostracoda in Mesquite, Aransas and Copano Bays, Texas Gulf Coast. Trans. Gulf-Cst Ass. geol. Soc. 17, 408-427. forbes, e. 1853. On the fluviomarme Tertiaries of the Isle of Wight. Jl geol. Soc. Lond. 9, 259-270. — 1856. On the Tertiary Fluvio-marine Formation of the Isle of Wight. Mem. geol. Surv. U.K. 162 pp. haskins, c. w. 1968a. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part I. Revue Micropaleont. 10, 250-260, 2 pis. — 19686. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part II. Ibid. 11, 3-12, 2 pis. — 1968c. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part III. Ibid. 161 175, 3 pis. — 1969. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part IV. Ibid. 12, 149-170, 4 pis. — 1970. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part V. Ibid. 13, 13-29, 3 pis. 1971a. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part VI. Ibid. 207-221, 3 pis. 19716. Tertiary ostracoda from the Isle of Wight and Barton, Hampshire, England. Part VII. Ibid. 14, 147-156, 2 pis. — 1971c. The stratigraphical distribution and palaeoecological significance of the Ostracoda from the Lower Tertiary Beds of the Hampshire Basin, England. Bull. Cent. Rech. Pau—S.N.P.A. 5, suppl., 545-557. howe, H. v. 1971. Ecology of American torose Cytherideidae. Ibid. 349-359, 1 pi. hulings, n. c. and puri, h. s. 1964. The ecology of shallow water ostracods of the West Coast of Florida. Publ. Staz. zool. Napoli, 33, suppl., 308-344. jones, t. r. 1856. Notes on the Entomostraca of the Headon and Osborne Series. In forbes, e. 1856. On the Tertiary Fluvio-marine Formation of the Isle of Wight. Mem. Geol. Surv. G.B., p. 157. — 1857. A monograph of the Tertiary Entomostraca. Palaeontogr. Soc. [Monogr.], 68 pp., 6 pis. — and sherborn, c. d. 1889. A supplemental monograph of the Tertiary Entomostraca of England. Ibid. 55 pp., 3 pis. kaesler, r. l. 1966. Quantitative re-evaluation of ecology and distribution of Recent Foraminifera and Ostracoda ofTodos Santos Bay, Baja California, Mexico. Paleont. Contr. Univ. Kansas. 10, 1-50. kaufmann, a. 1900. Cypriden und Darwinuliden der Schweiz. Revue Suisse Zool. 8, 209-423, pis. 15-31. keen, m. c. 1968. Ostracodes de l'Eocene superieur et I’Oligocene mferieur dans les Bassins de Paris, du Hampshire et de la Belgique, et leur contribution a l’echelle stratigraphique. In Colloque sur l’Eocene, Paris 1968, 1, Mem. Bur. Rech. geol. min. no. 58, 137-145. — 1971. A palaeoecological study of the ostracod Hemicyprideis montosa (Jones and Sherborn) from the Sannoisian of North-West Europe. Bull. Cent. Rech. Pau—S.N.P.A. 5, suppl., 523-543, 2 pis. — 1972a. The Sannoisian and some other Upper Palaeogene Ostracoda from north-west Europe. Palaeontology, 15, 267-325, pis. 45-56. — 19726. Mid-Tertiary Cytherettinae of north-west Europe. Bull. Br. Mus. nat. Hist. (Geol.), 21, 261-349, 23 pis. 1973a. On Haplocytheridea debilis (Jones). Stereo-Atlas Ostr. shells , 1, 181-188. 19736. On Haplocytheridea mantelli Keen sp. nov. Ibid. 189-192. — 1975. The palaeobiology of some Upper Palaeogene freshwater ostracodes. Bull. Am. Paleont. 65, 271-283. keij, a. j. 1957. Eocene and Oligocene Ostracoda of Belgium. Mem. Inst. R. Sci. nat. Belg. no. 136, 210 pp., 23 pis. kilenyi, t. i. 1969. The problems of ostracod ecology in the Thames Estuary. In neale, j. w. (ed). The taxonomy, morphology, and ecology of Recent Ostracoda. Oliver and Boyd, Edinburgh, 251-267. — 1972. Transient and balanced genetic poly-morphism as an explanation of variable noding in the ostracode Cyprideis torosa. Micropaleontology, 18, 47-64, pi. 1 . kollmann, k. 1960. Cytherideinae and Schulerideinae n. subfam. (Ostracoda) aus dem Neogen des ost- lichen Oesterreich. Mitt. geol. Ges. Wien, 51, 89-195, 21 pis. kornicker, l. s. 1965. Ecology of Ostracoda in the north-western part of the Great Bahama Bank. Publ. Staz. zool. Napoli, 33, suppl., 345-360. 444 PALAEONTOLOGY. VOLUME 20 ladd, H. s., hedgpeth, j. w., and post, r. 1957. Environment and Facies of Existing Bays on the Central Texas Coast. In (ladd, h. s.), Mem. geol Soc. Am. 67 (2), 599-640. malz, h. 1973. Ostracoden aus dem Sannois und jungeren Schichten des Mainzer Beckens, 3. Ehemalige 'Cytheridea' — Arten und— Verwante. Notissbl. Hess. L-Amt. Bodenforsch. 101, 188-201, pis. 19-22. -and moayedpour, e. 1973. Miozane Susswater— Ostracoden aus der Rhon. Senckenberg. leth. 54, 281 -309, 5 pis. mandelstam, m. i. 1947. Ostracods from the Jurassic deposits of the Mangychlak Peninsula. In Micro- fauna of the petroleum-bearing deposits of the Caucasus , Emba, and central Asia, VNIGRI, 239-262, 2 pis. [In Russian.] moore, R. c. (ed.). 1961 . Treatise on Invertebrate Pcdeontology, Part Q. Arthropoda 3. Crustacea, Ostracoda. Geol. Soc. Am. and Umv. Kansas Press, 442 pp. moos, b. 1970. Die Ostracoden-Fauna des Unteroligozans von Brandhurst bei Bunde (Bl. Herford-West, 3817), III. Schulerideinae mandlestam 1959 und Cytherideinae Sars 1925. Geol. Jb. 88, 289-320, 5 pis. muller, G. w. 1894. Ostracoden des Golfes von Neapel und der angrensenden Meeresabschnitte. Fauna u Flora Neapel . 21, 403 pp., 40 pis. Murray, j. w. and wright, c. a. 1974. Palaeogene Foraminiferida and palaeoecology, Hampshire and Paris Basins and the English Channel. Spec. Pap. Palaeont. 14, 129 pp., 20 pis. puri, h. s. 1957. Notes on the ostracode subfamily Cytherideidinae puri 1952. J. Acad. Sc. Washington, 47, 305-308. — 1968. Ecologic distribution of Recent Ostracoda. In Proc. Symp. Crustacea, pt. 1, Mar. biol. Ass. India (1966), 457-495. reyment, r. a. and brannstrom, b. 1963. Certain aspects of the physiology of Cypridopsis (Ostracoda, Crustacea). Stockh. Contr. Geol. 9, 208-242. Sandberg, p. a. 1964. The ostracod genus Cyprideis in the Americas. Ibid. 12, 178, pi. 23. sars, G. o. 1866. Oversigt af Norges marine Ostracoder. Forh. VidenskSelsk. krist. 7, 1-131. 1925. An account of the Crustacea of Norway, 9. Ostracoda (1922-1928). Bergen Mus. 9, 277 pp., 1 19 pis. swain, f. m. 1955. Ostracoda of San Antonio Bay, Texas. J. Paleont. 29, 561-646. swartz, f. m. and swain, f. m. 1946. Ostracoda from the Upper Jurassic Cotton Valley group of Louisiana and Arkansas. J. Paleont. 20, 362-373, pis. 52-53. triebel, e. 1963. Ostracoden aus dem Sannois und jungeren Schichten des Mainzer Beckens: 1, Cyprididae. Senckenberg. leth. 44, 157-207, 12 pis. van harten, d. 1975. Size and environmental salinity in the modern euryhaline ostracod Cyprideis torosa (Jones, 1850), a biometrical study. Palaeogeogr., Palaeoclimatol., Palaeoecol. 17, 35-48. vella, p. 1969. Correlation of base of Middle Headon Beds between Whitecliff Bay and Colwell Bay, Isle of Wight. Geol. Mag. 106, 606-608. vesper, b. 1972a. Zum Problem der Buckelbildung bei Cyprideis torosa (Jones, 1850) (Crustacea, Ostra- coda, Cytheridae). Mitt. Hamburg. Zool. Mus. Inst. 68, 79-94. — 19726. Zur Morphologie und Okologie von Cyprideis torosa (Jones, 1850) (Crustacea, Ostracoda, Cytheridae) unter besonderer Berucksichtingung Seiner Biometrie. Ibid. 21-77. wagner, c. w. 1957. Sur les Ostracodes du Quaternaire Recent des Pays-Bas et leur utilisation dans l’etude geologique des depots Holocenes. Diss. Univ. Paris, 259 pp. whatley, r. c. and kaye, p. 1971 . The palaeoecology of Eemian (Last Interglacial) Ostracoda from Selsey, Sussex. Bull. Cent. Rech. Pau—S.N.P.A. 5, suppl., 311-330. Typescript received 27 February 1976 Revised typescript received 11 June 1976 m. c. keen Department of Geology The University Glasgow, G12 8QQ KEEN: EOCENE OSTRACOD ASSEMBLAGES 445 APPENDIX Details of sampling horizons for Headon Hill (HH) (SZ 305 860) and Whitecliff Bay (WB) (SZ 642 863) are given in text-fig. I. Other samples mentioned are: from Headon Hill, HHCP2, Cyrena pulchra Bed sample no. 2 (see text-fig. 5); Base of UB, freshwater clay immediately below the Venus Bed; Base V.B., V.B. 18, Mid V.B., Top V.B., samples through the Venus Bed; NHH 1-4, samples from the horizon of HH48 collected in more detail. CB is Colwell Bay (SZ 327878-330 890) 1 and 2 from the Oyster Bed, 4-6 from a blue sandy clay with shells 60-200 cm above the Oyster Bed, 12 a thin buff limestone in the Upper Headon Beds near Cliff End; HHL is the Howe Ledge Limestone of Colwell Bay; MF is Milford (SZ 278 917), 1 and 2 from the Unio Bed, 3 and 4 from the Middle Headon Beds; BC is Bouldnor Cliff (SZ 401 920), 3 topmost Bembridge Limestone, 570 cm above the limestone, 8, a further 120 cm higher; BR is Brockenhurst (SU 324 036), 1 and 3 from the Middle Headon Beds; Thorness Bay (SZ 464945). O A REVIEW OF THE ECOLOGY OF UPPER CARBONIFEROUS PLANT ASSEMBLAGES, WITH NEW DATA FROM STRATHCLYDE by A. C. SCOTT Abstract. Previous studies on Upper Carboniferous floral palaeoecology are reviewed and relationships between fossil plant assemblages, depositional environments, and contemporaneous plant communities are discussed. A point-quadrat sampling technique was applied to a quantitative study of plant horizons in a ‘roof shale’ of a thin coal below Skipsey’s Marine Band (Westphalian B) at a locality near Annbank, Strathclyde. This study has shown that the number and percentage cover of ‘drifted’ and in situ species varies up the succession. Changes in the deposi- tional environment from swampy flood plain to near channel alluvial discharge, interpreted from the lithofacies, are thought to account for these differences by affecting the quantity of plant material deposited and by incorporating a variety of ‘Plant Communities’ to yield different ‘Fossil Plant Assemblages’. The latter consist of two main types, one of pteridosperms with some Cordaites and another of sphenopsids and pteridosperms with Cordaites. Only very few lycopod remains were found. This is in contrast to the coal-forming swamp ‘community’ (as seen in the spores contained in the coal) which was dominated by lycopods with some sphenopsids. A number of plant communities existed during the Upper Palaeozoic, for example various authors have recognized ‘upland’ and ‘swamp’ floras (Cridland and Morris 1963; Havlena 1971 ; Leary 1974). The relationships, however, of fossil plant assem- blages to contemporaneous plant communities and their sedimentary history has received less attention, although a few workers have tried to define more clearly the ecology of these ‘floras’ (Havlena 1971). Little attempt has been made to relate fossil plant assemblages to depositional environments, which is necessary in order to understand the nature of the assemblages. A number of sampling techniques have been developed by ecologists studying extant floras (Greig-Smith 1964), some of which may be adapted to the fossil situation. The interpretation of the results obtained in the two different examples are fundamentally different. This paper sets out to review previous work on Upper Carboniferous floral palaeo- ecology and discusses the application of Recent ecological-sampling techniques to fossils. A ‘roof shale flora’ of Westphalian B age from Annbank, Strathclyde, Scotland, is taken to illustrate the use of a half-metre square-quadrat technique on cleared bedding planes. The study of the Annbank section is intended to illustrate how detailed analysis of the sediments and flora, together with palynological data from the underlying coal, can give a better understanding of the Coal Measure environment and the possible plant community structure. PREVIOUS STUDIES ON UPPER CARBONIFEROUS FLORAL PALAEOECOLOGY Most early work on Upper Carboniferous (mainly Coal Measures) plants has been concerned with general description : new taxa, their botanical affinities, lists of floras. [Palaeontology, Vol. 20, Part 2, 1977, pp. 447-473, pis. 50-51.] 448 PALAEONTOLOGY, VOLUME 20 and considerations of their stratigraphic age (Scott 1906; Arber 1906). Most dis- cussions concerning the nature and origin of coal concerned an in situ or drifted theory (Potonie 1899a; Grand ’Eury 1900; Kendall 1922; Bennett 1963). Under- standably, as little work had been done on the ecology of extant floras, few workers were concerned primarily with this aspect although occasional observations were noted in passing (Grand ’Eury 1897; Wiinsch 1865; Walton 1935; MacGreggor and Walton 1948; Jongmans 1955). Most of the reconstructions of Coal Measure swamp vegetation (Westphalian) included known plants together and showed plant recon- structions rather than ecological associations (Geinitz 1855; Potonie 18996; Noe 1932; Dahlgren 1931); although sometimes the species associations were noted (drawings by Nathorst in Magdefrau 1968, by Bertrand in Bertrand and Corsin 1950). Lycopods clearly played an important role in the peat accumulations contributing to formation of coal seams (seen both from the stigmarian seatearths of most West- phalian coals (Logan 1841), and in the abundance of those plants in coal balls (Binney 1868, 1871, 1872, 1875)) but the detailed ecology of Westphalian floras was generally not investigated. Advances in the study of coal petrology (Stopes and Wheeler 1918) gave an impetus to the ecological study of the peat- (coal-) forming floras. The recognition of the major coal constituents (Stopes 1919) and later coal macerals (Stopes 1935) enabled those studying compressions of plants to relate these to their material (Hickling and Marshall 1932, 1933) thus allowing the general type of peat composition of a coal to be identified. Several attempts have been made to interpret the habitat of Upper Carboniferous plants by comparison of the internal structure with that of extant plants (Thomas 1911; Weiss 1925; Noe 1932; Cridland 1964; Wartmann 1969). It was concluded that lycopods and sphenopsids lived in wetter conditions than pteridosperms and some cordaites (interpreted from root, stem, and leaf structures). Cridland showed (1964) that one type of Cordaites had stilt roots which may be compared with those of mangroves. This does not necessarily mean, however, that it occupied the same ecological niche. It would appear that different species of the genus Cordaites could live in widely different habitats, from coastal plain (Cridland 1964) to dryer ‘upland’ (Cridland and Morris 1963; Wartmann 1969). The ecological interpretations of some of these botanical structures are difficult as Chaloner and Collinson (1975) have shown that comparing the structures of Recent and fossil plants is ‘fraught with hazards’. Care must also be taken when arguing detailed ecology between Recent and fossil peat-forming environments (Leclercq 1926; Griffiths 1927; R. Teichmuller 1955) as, although general principles of the environment may be correct, detailed vegeta- tional comparison is not possible (at least with the Upper Carboniferous) (Spackman et al. 1 966 ; Habib et al. 1 966 ; Habib and Groth 1 967 ; papers in Dapples and Hopkins 1969; Cohen and Spackman 1972; Cohen 1975). The evidence concerning the climatic conditions of the Euramerian coal swamps has been the cause of much debate (Schopf 1974). The interpretations differ widely from a tropical to a temperate climate (White 1931; Friedricksen 1972; Schopf 1974) although the absence of growth rings in the wood from the Coal Measures (Westphalian) indicates a uniform climate (Chaloner and Creber 1974), perhaps tropical but at least semitropical and SCOTT: CARBONIFEROUS PLANT ECOLOGY 449 moist (Noe 1932) straddling the equatorial belt (Friedricksen 1972; Chaloner and Lacey 1973). The first attempt at a quantitative palaeoecological study of the Coal Measure (Westphalian) flora was the pioneering work of Davies (1908, 1920, 1921, 1929 and summarized by North 1935). Davies showed that data collected in both a quantitative and qualitative way was necessary in any serious attempt to unravel Upper Carboni- ferous shale floras. Davies also showed that the roof-shale floras associated with coal seams varied in terms of variety of species and total quantity of plant material. He noted the change through time of the floras from ones dominated by sphenopsids (in the lower part of the Coal Measure sequence) to ones dominated by pterido- sperms (in the upper part). These changes Davies interpreted in terms of a wet flora giving way to a dry flora. These roof-shale floras do not necessarily indicate changes in coal-swamp floras, rather than changes in sedimentary environments, which Davies did not study in detail. North (1935) commenting on these results remarked ‘it is necessary to remember that in collecting fossil plants we are not dealing with organisms in situ ; the plant impressions in the Coal Measure shales are almost always derived from drifted fragments, and it does not necessarily follow that they represent plants identical with those that gave rise to the carbonaceous material of the next subjacent seam; they may have been transported for a considerable distance by the water that carried the mud in which they were ultimately buried and preserved’. Seward (1933) was one of the first to realize the importance of environments of deposition when he wrote, ‘we cannot assume that the contemporary vegetation was exceptionally meagre or locally unrepresented merely because no fossils have been found; it may be because the conditions under which barren sediments were formed were not such as were favourable to the preservation of plant frag- ments’. Davies (1929) determined a general pattern of change in the Upper Carboniferous floras of South Wales, a theme taken up by Dix (1934) who set up floral zones, using South Wales as her type area, which were later interpreted by R. Potonie (1951, 1952) as changing due to broad climatic shifts affecting the community structure. Sub- sequently floral zones were established in other areas, such as the United States (Read and Mamay 1964). Patton (1922) showed how, in Recent times, sedimentation, physiographic features, and plant distribution are all interrelated, a theme pursued in a speculative paper by Robertson (1952), who considered the role that plants played in controlling rhythmic sediments. Chaloner (1968) suggested that the two are intimately related; changes in one affecting the other and both being affected by other variables (such as climate). Following the early work on coal petrology and using her work on the German Brown Coals, Teichmiiller (1952; Teichmiiller and Teichmiiller 1968) proposed relationships between plant assemblages and coal lithotypes of the Carboniferous. This work led to a theory concerning the ecology of peat-forming communities (Smith 1 962, 1963). These studies, together with those of Oswald (1937) and Karmasin (1952), led Haquebard and Donaldson (1969) to reach slightly different conclu- sions concerning relations between floral changes and coal petrographic variations. These authors adopted the terms used by Karmasin (1952) of different swamp environments and related forest-moor type peat to the formation of vitrite and 450 PALAEONTOLOGY, VOLUME 20 vitro-clarite with abundant pteridosperms, lepidophytes, and ferns; open moor with subaquatic cannel deposition (and spore-rich durite) and a transitional reed-moor for the deposition of spore-rich clarite formed mainly by sphenopsids. They also considered Oswald’s (1937) relations of preservation of plant debris in relation to ground water and recognized ’a terrestrial zone (above the high water mark)’ where dry fusito-clarite was formed, ‘a telamitic zone (between high and low water marks)’ where bright coals would form, and a ‘limnic (subaqueous) zone’ where dull coals were formed. Again rather different assemblages have been found by Habib (1966, 1968) in the Pennsylvanian of the United States, probably as the general sedimentary environment and ecology is not the same as in Canada or Great Britain. The presence of abundant fusain in coal has caused controversy (White and Thiessen 1913; Stiizer 1929; Crickmay 1935; Marshall 1954; Francis 1961; Komarek 1972). Some believe that fusain represents true charcoal, and is the result of forest fire, because of the similar three-dimensional preservation (unaltered by subsequent compaction and hence a rigidity acquired before or very early in peat formation); a similar tempera- ture of formation as shown by electron spin resonance studies and specific heat determinations; a general resemblance to charcoal, often with fine detail preserved without any sign of fungal or bacterial attack, and finally the fragmentation of the material (Grebe 1953; Terres et al. 1956; Harris 1958; Francis 1961; Austen et al. 1 966 ; Stach 1 968 ; Alvin 1 974 ; Scott 1 974). Others, however, feel that the resemblance to charcoal is only superficial and object to this method of formation because of the abundance of fusain in Palaeozoic coals in relation to those of subsequent eras; the absence of conflagration and drought in tropical swamp areas today and the presence of unaltered resin bodies associated with the fusain, but no alternative satisfactory mechanisms have yet been found (White and Theissen 1913; Stiizer 1929; Terres et al. 1956 for Gondwanan coals; see discussion in Austen et al. 1966; Schopf 1975). This controversy remains unresolved. White (1907, 1908, 1909, 1931) showed that in the topmost Upper Carboniferous and Lower Permian rocks there were representatives of an ’upland’ flora. This idea was expanded by Gothan and Gimm (1930) for the European Permian and again later by Gothan and Remy (1957) who recognized a lowland association of Catamites and Peeopteris and an upland association of Callipteris and Walchia. Subsequently such upland floras have been recognized extensively throughout the Carboniferous. Cridland and Morris (1963) in the Pennsylvanian of Kansas extended the work of Moore et al. (1936) and demonstrated the major differences between Upper Carboni- ferous upland and lowland floras, the former containing Taeniopteris , Walchia , and Dichophyllum. Daber (1955, 1957) recognized different plant associations in the Visean and Westphalian of Germany and Havlena (1961, 1970, 1971), working on Carboniferous successions of Czechoslovakia, described both a coal-forming association of plants (Floznah) and an association drifted into the coal-forming area (Flozfern). Obrhel (1960), studying the Stephanian successions in Bohemia, noted relationships between plant assemblages and facies and discussed allochthonous and autochthonous plant preservation. Josten (1961), studying German Upper Carboniferous stratigraphy, noted the presence of definite and repeated cycles of fossil plant assemblages. Following this work Dragert (1964), in an extensive quantitative study of roof-shale floras from SCOTT: CARBONIFEROUS PLANT ECOLOGY 451 a sequence of coals which could be traced some distance laterally, found a differentia- tion of plant associations which he believed to represent actual plant communities. He argued that where the associations have been transported they have the same composition as those which appear to have been preserved in situ. He recognized three associations or ‘community types’: (1) articulate (sphenopsid), (2) pterido- sperm, and (3) lepidophyte-cordaite. It was found that above most coal seams the lycophyte-cordaite association was laterally replaced by the pteridosperm associa- tion which was itself replaced by the arthrophyte (sphenopsid) association and rarely was the first and last of these associations found in juxtaposition. It is difficult to know to what extent the assemblages were related to the coal-forming swamp itself or whether to other physiographic features of a flood plain or another environment. As well as providing a stimulating study this work contains an important reference source to the German literature. More recent work has not solved the problems raised by this work (Keller 1972), although Pfefferkorn et al. (1975) have recalculated the data and compared it with other sites in Europe and America. Several Russian workers have interpreted the palaeoecology of Upper Carboni- ferous plant assemblages (Radchenko 1964; Fissunenko 1965; Krassilov 1972). These studies have included both the Karaganda Basin (Oshurkova 1967, 1975) and the Donetz Basin (Stschegolev 1965, 1975; Fissunenko 1967; Fissunenko and Stsche- golev 1975). However, this work has been mainly qualitative and only one quantita- tive study has been published (Oshurkova 1974). Nevertheless three types of plant community have been recognized (Oshurkova 1967): (1) foreshore-hydrous silvan vegetation (waterlogged marginal lake forests), accumulating on coastal plains consisting mainly of arthrophytes (sphenopsids); (2) ‘Swampy Silvan’ (swamp forest) vegetation in coal-forming areas characterized by lepidophytes (lycopods); and (3) ‘woody shrub vegetation’, in conditions of well-drained plains consisting mainly of pteridosperms. These broad community-types should be regarded as a hypothesis leading towards the recognition of real communities, rather than as real communities in themselves. This pioneering work will be strengthened as more becomes known of the depositional histories of the rock sequences. Knight (1974) has shown that in the Sabero Coalfield of Spain there were two distinctive floras within the Stephanian A; firstly a flora associated with the coal- forming environment (Floznah) where fern species dominated pteridosperm species and secondly a hillslope flora (of younger aspect) in the northern area associated with sandier rocks (Flozfern) where pteridosperm species dominated fern species. A recent discovery of abundant Sporangiostrobus together with an abundant associ- ated flora preserved in volcanic ash within a coal seam from the Upper Stephanian of Puertollano, Spain (Wagner and Spinner 1976) gives us for the first time a detailed picture of this important coal-forming community. The spores of Sporangiostrobus (micro-, Densosporites and mega-, Zonalesporites ) are known to occur abundantly at many levels in coals of Upper Carboniferous age (Grebe 1966), but until now only a few specimens of the parent plant were known. Associated with this plant in the same ash band, which presumably engulfed and preserved this community in situ , are abundant ferns (mainly pecopterids yielding Torispora ), Scolecopteris , sphenopsids ( Asterophyllites and Macrostachya), and cordaites (mainly Corciai- thanthus with only a few leaves). 452 PALAEONTOLOGY, VOLUME 20 There has been renewed interest in various aspects of Upper Carboniferous floral palaeoecology in the United States (Phillips et al. 1973). Peppers and Pfefferkorn (1970) examined both macro- and micro-plant remains from a stratigraphic interval in the Carbondale formation, and distinguished qualitatively numerous plant ‘associations’, related to different physiographic features: ‘Wet swampy’ with lyco- pods, ferns, and rare cordaites; ‘Dry swampy’ with ferns, lycopods, and sphenopsids; ‘Levees and flood plains’ with pteridosperms, ferns, and sphenopsids; and ‘Upland’ with pteridosperms, cordaites, and rare Noeggerathiales. Their schematic interpreta- tion of floral distribution in sedimentary facies has proved useful as a working model, although the basis for some of the ecological interpretation (such as a wet or dry swampy area) remains uncertain. Phillips et al. \914a and Phillips et al. 19746 described the quantitative analysis of coal-swamp vegetation in relation to coal, using the plants present in coal balls as a guide to the constitution of the coal. They have noted that in the Illinois Basin there is a major change in coal-ball floras from the Westphalian to the Stephanian from one dominated by lycopods to one dominated by tree ferns. This change has been interpreted as indicating a broad climatic shift at that time. Darrah (1941) noted that the coal-ball floras from Iowa are dominated by cordaites, pteridosperms, and ferns and so if these floras really represent those plants which formed the coal, as most workers argue, then it would appear that many groups of plants were capable of coal formation and it may have been environmental factors rather than specific communities which were responsible for coal formation, although some were more suited than others. This work is supported by recent advances in coal chemistry (Niklas and Phillips 1974). The age of the Dunkard Series of the eastern United States (Clendening and Gil- lespie 1972) is a problem showing the importance of ecology in any palaeobotanical stratigraphical work— see also Oshurkova (1975). Pennsylvanian upland floras have been studied by Leary (1974, 1975; Blazey 1974) who showed a rather dif- ferent floral composition from the normal swamp flora, which was rich in Pterido- spermales and Noeggerathiales and deficient in ferns and lycopods. Although upland floras were recognized as macrofossils early in this century it was not until 1958 that Chaloner (reinterpreting data from Neves 1958) showed that such floras could be recognized in the palynological record. Schopf et al. (1944) had commented that some of their conifer pollen might have been produced by upland species (p. 28), but this aspect of their work was not followed up until the stratigraphical distribution of spores had been studied (Smith and Butterworth 1967; Chaloner and Muir 1968; Friedricksen 1972). Many workers across the Euramerian coal-forming belt have recognized, par- ticularly in the Westphalian, three major lowland plant assemblages; one dominated by lycopods, one by sphenopsids, and one by pteridosperms as well as various upland floras. It is now important to relate these assemblages to depositional environments with the hope of unravelling the Upper Carboniferous plant com- munity structure. SCOTT: CARBONIFEROUS PLANT ECOLOGY 453 PROBLEMS IN RELATING FOSSIL PLANT ASSEMBLAGES TO CONTEMPORANEOUS PLANT COMMUNITIES Ecologists studying living plant associations have long argued over the definition of a plant community and how to characterize it statistically (Greig-Smith 1964), but the fossil plants a geologist finds generally represent the fragmented parts of indivi- duals, often derived from plant communities only brought together by transport and sedimentary processes. Further the plants observed may have undergone changes of one sort or another during the fossilization process (Rolfe and Brett 1969). Problems in relating invertebrate fossil assemblages to their living communities have been discussed extensively over the last fifteen years (see references in Walker and Alberstadt 1975), and have included the study of processes that may affect an organism after death (Taphonomy). Little work concerning macrofloras has been published, although these problems have been recently discussed in some unpublished theses (Hill, Leeds, 1974; Spicer, London, 1975). Krassilov (1969, 1972, 1974) emphasized palaeofloristic successions and their causes, rather than relating fossil plant assemblages to contemporaneous plant communities. Once that data has been collected in a repeatable fashion, then hypotheses concerning possible plant com- munities will involve the intermediate steps of an interpretation of transport and depositional histories of the plants. These hypotheses may be constantly modified with the collection of new data and also with changes of interpretation of the LIVING FLORA-PLANT COMMUNITY FOSSILIZATION SEQUENCE DEPOSI TION PRESERVATION- FOSSILIZATION COLLECTION FOSSIL PLANT ASSEMBLAGE INTERPRETATION text-fig. 1 . Problems in relating fossil plant assemblages to contemporaneous plant communities. 1, Initiators of the fossilization sequence (on some or all of the plants in a community): (i) Internal factors; organ abscission; organ shedding; disease; resistance of plant organ to decay, (ii) External factors; animal destruction; storms; floods; land subsidence; climatic change; erosion; other natural catastrophies, e.g. forest fire, proximity to site of transport. 2, Key to factors controlling fossil plant assemblages (on some or all the plants in the fossilization sequence): (i) Destructive mechanisms; (a) decay, ( b ) mechanical break-up, (c) immediate post-depositional rework- ing, (d) diagenesis, (e) weathering, (f) collecting bias, (ii) Interference, (g) sorting, (h) additions from other communities. 454 PALAEONTOLOGY, VOLUME 20 sedimentary environments. The major processes involved in the formation of a fossil assemblage are set out in text-fig. 1 . Initiators of a sequence are listed under internal and external factors and ‘noise’ may be related to destruction of all or some of the plants at any stage, or interference by other factors during the major sequence. Text-fig. 1 shows in the left-hand column, those processes which have gone into forming a fossil plant assemblage and in the right-hand column those steps an investigator must take to retrace these paths. This sequence does not interpret the changes in assemblages up a succession but represents primary taphogenic assemblages (those connected with changes affecting plant remains in the course of their transportation and burial, Krassilov 1969). SAMPLING TECHNIQUES Numerous sampling techniques have been employed in ecological studies on fossil plants. The merits of different techniques have recently been discussed by Hill (unpublished thesis, Leeds, 1974) and Spicer (unpublished thesis, London, 1975) but those considered here have been used in Palaeozoic studies. Coal Measure palaeoecological studies were pioneered in South Wales by Davies (1908, 1920, 1921, 1929) and this work was continued by Dix (1934). Techniques employed by both authors suffer from subjectivity and so are not reproducible. Much of the quantitative data collected by Davies was obtained by counting the number of plant fragments found at a particular level. No standard sample size was used and genuine changes in floral assemblages are hidden by the nature of fragmentation of each species and their depositional histories. Neither the absolute numbers obtained nor the relative proportions of each species can be compared with those from other horizons. Dix (1934) also failed to use an objective sampling technique but used a subjective method of designating ‘dominant’ and ‘subsidiary’ species. Harris (1952), whilst studying the Yorkshire Jurassic, used another type of quantitative method, a measure of frequency (in a sense more or less equivalent to ‘fidelity’ of ecologists studying Recent plant communities), which he defined as the number of localities at which a species has occurred as a percentage of the total number of exposures of that horizon studied. This use of ‘frequency’ may be helpful in recording the per- sistence of species and its use may be extended to the study of Coal Measure plant beds rather than localities or exposures. Dragert (1964) used a half-metre square quadrat technique, and used in essence the presence or absence of species to record their vertical and horizontal frequency in roof shales. A common technique for Recent ecologists is the quadrat, either by an analysis of cover or frequency determination (Greig-Smith 1964). The interpretation and problems of quadrat analyses on modern plant communities have been reviewed by Frenkel and Harrison (1974). Both the size of quadrat and the type of data obtained from the quadrat are important and are directly related to the type of palaeontological section to be examined, to the type of data required, and to the time available for study. The quadrat size may have to be smaller than ideal, due to the numerous problems in excavation and data collection. The basic counts which can be made from a fossil quadrat are (1) the number of fragments of each species (Hill, unpub- lished thesis, Leeds, 1974) and (2) the cover area of each species (this paper). The SCOTT: CARBONIFEROUS PLANT ECOLOGY 455 first method has the problems that the results are greatly affected by the extent to which different species break up during transport and also by the process of sampling itself. However, the cover technique does not take into consideration the degree of fragmentation of the plants, and some plants may be under- or over-represented because of their morphology or because of their uneven scatter on the bedding plane. A cover technique has been used in the example described here, but some assessment of the fragmentation and density of each species was also made. The sampling technique used at Annbank was as follows: At each plant horizon a half-metre square quadrat of one or more closely spaced bedding planes was cleared and a grid with 100 random points was placed over the area. The presence and identity of plant remains was recorded underneath each point to give a numerical assessment of the percentage ‘cover’ of each species on the quadrat; the bare rock was also recorded. When a quadrat could not be conveniently cleared, blocks of one horizon were split and laid out in the quadrat area and then the count was made. Using this method counts of plant, bare rock, and space were made and so extra random points were recorded until the total of plants plus bare rock was 100. There are several factors influencing the results (text-fig. 1). Because of the small extent of exposure the quadrats could not be placed in a random fashion but only where the section allowed. Errors may also occur from the nature of the technique itself and also that only 100 points per quadrat were taken. It is felt that 100 points represents a justifiable compromise between the time taken and the reliability of the results. The way in which a rock splits along the bedding may also influence the results, and it is rarely possible to measure cover on a single bedding plane. This problem is accentuated by the presence of plants such as Cordaites whose typically large strap-like leaves increase the probability of the rock cleaving open wherever they occur. In a given volume of rock, therefore, containing some Cordaites but perhaps a greater quantity of smaller parts of other plants, the rock will cleave preferentially to expose the Cordaites , resulting in some bias favouring such large plant parts. This use of ‘cover’ differs from that used by plant ecologists in a living community. There one is measuring the result of interactions between the plants, e.g. the extent to which a species is successfully using the light for photosynthesis. In the fossil (death) assemblages there is no interaction of assembled species and the percentage cover is merely a measure of area of plants (leaves and stems which were originally upright before transport) on the rock surface exposed by splitting. It is probably the easiest parameter to measure related to the mass of plant material contributed by any constituent of the community(ies) from which the assemblage was derived, as well as being related to the total photosynthetic surface of the plant in life— which 'is related to its status in the living plant community. The ‘cover area’ of each species, therefore, is related in part to its original biomass, in part to its preservability, and in part to its hydrodynamic qualities. Such a cover area is a measure of its ‘thanatomass’. Some plants may have a high thanatomass but may have been small contributors to a community whilst others such as ferns may have had an important role in a plant community and perhaps a large biomass, but because of the relative ease with which they are broken up they may have only a small thanatomass in a fossil plant assemblage. table 1. The distribution and abundance of fossil plants at Annbank. Domin Scale: + = a single individual, 1 1-2 individuals, 2 = less than £ 1%, 3= 1-4%, 4 = 5-10%, 5= 11-25%, 6 = 26-33%, 7 = 34-50%, 8 = 51-75%, 9 = 76-90%, 10 91-100% cover. A representative sample of specimens has been deposited in the Hunterian Museum, Glasgow (Pb). Pteridosperms and ferns Alclliopleris lonchitica auct. Aulacotheca sp. Diplotmema furcatum Brongn. Mariopieris muricaia (Schloth.) M. sauveri (Brongn.) A/, nervosa (Brongn.) M. sp. Neuropleris loslii Brongn. N. pseudogigantea Pot. N. rarinervis Bun. N. sp. Trigonocarpus sp. Splienopleris spini/ormis Kid. S. sp. seeds indet. pteridosperm axes indet. Articulates Annularia radiata Brongn. A. sphenophylloides Zenker Asleropliylliles equisetiformis Schloth. Catamites cisti Brongn. C. undulatus Sternb. C. sp. Pinnularia sp. Sphenophyllum cuneifolium Sternb. inc. var. saxifragaefolium S. myriophyllum Crep. CORDAITES Artisia sp. Samaropsis pyriformis Barker Cordaitanlhus sp. Cordaites principalis Germar Lycopods Lepidoslrobophyllum lanceolatum (Lind and Hut) Lycopodites sp. Stigmaria ficoides Sternb. indeterminate plant fragments Bare rock 0 Bed numbers Unknown or 2a 2b 2c 2d 2e 2f 2g 2h 2j 2k 3a 3c 3d 4b 4d 4f 4g 4h museum colln. 5454335 - 5 - 4 - 3 - 5- - 4 1 — — — — — _ — — 3--- - 3 _ _______ 3 — — — 3--- - 4 ___ 3 — _ 4 - 3 - 3 ------ - 5 — — — 4 — — 3 4 4 4 - - - - 3 5444- - 5-434 455- 54 3 — — — — — _ 3 4 3 3 - 3 - - - - 4 3 3 4__ ______ _ 6 - - - 5 3 3 3 4 4- 4 - 3 - 2 3 3 3 4 1-331531 1 1 1343133433 7 - 77889 10 897 10 78798888 PALAEONTOLOGY, VOLUME 20 table 1. The distribution and abundance of fossil plants at Annbank. Domin Scale: + = a single individual, 1 = 1-2 individuals, 2 — less than 1 %, 3 = 1-4%, 4 = 5-10%, 5 = 1 1-25%, 6 = 26-33%, 7 = 34-50%, 8 = 51-75%, 9 = 76-90%, 10 = 91-100% cover. A representative sample of specimens has been deposited in the Hunterian Museum, Glasgow (Pb). 456 PALAEONTOLOGY, VOLUME 20 0 =3 1 l o B C 3 S' C 3 = I I I I I I I 3 I I r? ^ + S I l 00 “■> IN 1 <— m | rs o ro I (N 3 ^ + 3 I xi -a- | a “a | (N - I I + I ^ ^ I I I I I I I ^ I I ^ | ’T ^ I I I ro Tf rj- I I " fb Tf I m I ’dr | uo i i i i ’*■ i m | Tt l i ^ l i ”• | ^ oa I I I I I I I " I + I I I I I I I I I I I m I I I I I I I I I I I I I I I ' I I 460. 8, 15, left valve, 1126/1-1. 8, internal view, x 150; 15, adductor muscle scar, internal view, x 460. 9, 16, right valve, 1116/124. 9, external view, x85; 16, adductor muscle scar, internal view, x 460. 10, right valve, 1126/9-3, adductor muscle scar, external view, x460. 12, left valve, 1 126/9-2, adductor muscle scar, external view, x 460. 13, left valve, 1 126/1-2, adductor muscle scar, internal view, x 460. 14, left valve, 1 126/3-3, adductor muscle scar, external view, x 460. Figs. 1-5. Scanning electron micrographs (Cambridge Instruments ‘Stereoscan 600', A1 coating); figs. 6-16 transmitted light. PLATE 52 GRAMM, Egorovitina kirssanovi sp. nov. 480 PALAEONTOLOGY, VOLUME 20 posterior margins via obtuse cardinal angles. Ventral margin nearly parallel to dorsal and fluently curving merges with the ends. Surface uniformly covered with shallow, smooth, and barely discernible pits. Fairly broad ridge-like swellings (rims) present along both extremities, limited proximally by gentle crescentic depressions. On right valve, at anterior cardinal angle, the swelling is weakly expressed; gradually becoming stronger and acquiring a rim-like form it reaches maximum in the antero- ventral part, where it forms a prominent elongate knob. On the posterior margin the rim is clearly distinguishable and reaches a maximum at the postero-dorsal cardinal angle as a prominent elongate knob. On the left valve the rims are developed more evenly and the maxima (elongated knobs), situated in antero-dorsal and postero-ventral parts, are not so strongly expressed. Owing to diag- onal and oppositely placed elongate knobs, the carapace in dorsal or ventral views has an obliquely compressed appearance. Hinge adont; hinge-margin straight, thin with a narrow groove in the left valve. Calcified inner lamella thin, narrow, and of uniform width, with a contact groove on the left valve. Walls are pierced with numerous, very thin normal pore canals; the distance between canals 2-3 ^m, diameter less than 1 /xm. The adductor muscle scar, roughly circular in form, is located in the mid-valve or slightly in front and higher. Internally it is situated on a round feebly expressed elevation. The structure is compact, round, resembling a five-toed footprint. In the posterior or postero-ventral part there is a supporting spot, away from which five elongate spots are arranged in a pedate pattern. The uppermost is thin, elongate; the three middle spots are the longest, well developed, and spatulate anteriorly; the lowermost spot is usually short and relatively thick. In juvenile forms the upper and lower spots are often indistinct. Mandibular and frontal spots not observed. text-fig. 2. Egorovitina kirssanovi sp. nov. Right valve, 1 1 16/124: a, external view; b, dor- sal view, x60. Reflected light photographs. Lower Carboniferous, Visean, Alexinsky hori- zon; Borovitchi district, Novgorod region, U.S.S.R. Length Height Width fur n) Cm) Cm) Carapace holotype 1126/12-2 525 310 225 Left valve 1126/3-1 235 150 — Right valve 1126/2-2 275 170 — Right valve 1126/10-1 275 160 — Right valve 1126/9-3 275 — — Left valve 1126/3-2 280 170 — Left valve 1126/2-1 285 170 — Left valve 1126/10-2 285 170 — Left valve 1126/9-2 325 200 — Right valve 1126/1-4 325 — — Right valve 1126/10-3 330 200 — Left valve 1126/10-4 330 200 — GRAMM: PALAEOZOIC OSTRACODS 481 Length Height Width (/um) (jum) (ium) Left valve 1126/1-2 360 210 75 Right valve 1126/7-2 385 225 90 Left valve 1126/3-3 385 225 — Left valve 1126/8-2 390 225 100 Left valve 1126/1-1 400 230 90 Left valve 1126/9-1 400 240 — Right valve 1116/124 475 275 105 Right valve 1126/8-1 475 275 110 Right valve 1 126/13-1 500 290 — Left valve 1126/26 500 300 — Carapace 1116/159 550 315 250 Carapace 1116/125 550 325 225 Carapace 1116/144-1 550 325 250 Carapace 1116/144-2 560 330 205 Ontogeny. In smallest specimens, 235-285 long, the anterior is higher than the posterior and the dorsal margin is slightly concave (PI. 52, figs. 6, 7); the adductor muscle scar is slightly posterior of mid-valve length. In specimens 325-475 ^.m long both extremities are of equal height, dorsal margin straight, adductor muscle scar located in the middle of the valve. In carapaces 550-560 long the posterior extremity becomes the highest and the adductor pattern is slightly anterior of mid- valve length. Thus during the ontogeny the interrelation of the height of the anterior or posterior ends and the position of adductor pattern changes. The adductor muscle scar undergoes the following changes: the smallest specimen studied (235 jum long) has the supporting spot and three anterior spots; this leads to a suggestion that these four spots appear first; the upper and the lower spots emerge later. During the ontogeny an increase in the adductor pattern and separate spots took place, the anterior spots (especially the middle ones) grew and lengthened most. Variability. The ridge-like swellings are differently expressed on different specimens. The adductor pattern is also variable; this is seen on the illustrations and finds expression in the length and shape of separate spots. On the young growth-stages the number and shape of spots is also variable (PI. 52, figs. 10-12). Notable are an adductor pattern with differently placed anterior spots (PI. 52, fig. 14) and a pattern with an irregular posterior supporting spot (PI. 52, fig. 13); apparently they are pathological phenomena. Acknowledgements. I thank Dr. G. I. Egorov (Leningrad) and Mr. V. K. Kirssanov (Borovitchi) who generously provided material for study. Miss T. N. Teriokhina processed the samples and Miss V. A. Iljina prepared the photographs. Mrs. O. G. Gein kindly typed the manuscript. My special thanks are to Dr. Alan Lord (University College, London) who prepared the SEM photographs, and helped greatly with the final preparation of the paper. RELERENCES bradfield, h. h. 1935. Pennsylvanian ostracods of the Ardmore Basin, Oklahoma. Bull. Am. Paleont. 22, 1-173. bushmina, l. s. 1968. Early Carboniferous ostracodes of Kuznetzk Basin. Moscow. 128 pp. [In Russian.] 482 PALAEONTOLOGY, VOLUME 20 coryell, h. n. and billings, G. d. 1932. Pennsylvanian Ostracoda of the Wayland Shale of Texas. Am. Midi. Nat. 13, 170-189. gramm, m. n. 1976. The interrelation between the Palaeozoic ostracodes Roundyella and Scrobicula. Geol. For. Stockh. Fork. 98, 217-226. — and posner, v. m. 1972. Morphology and ontogenesis of adductor muscle scar in the Paleozoic ostracode Scrobicula scrobiculata. Paleont. Zh. 3, 99-105. [In Russian.] jones, t. r. and kirkby, j. w. 1895. Notes on the Palaeozoic Bivalved Entomostraca. No. XXXII. Some Carboniferous Ostracoda from Yorkshire. Ann. Mag. nat. Hist. ser. 6, 16, 452-460. ray, G. m. 1940. Ordovician Mohawkian Ostracoda: Lower Trenton Decorah fauna. J. Paleont. 14, 234-269. kellet, b. 1933. Ostracodes of the upper Pennsylvanian and the lower Permian strata of Kansas: 1. The Aparchitidae, Beyrichiidae, Glyptopleuridae, Kloedenellidae, Kirkbyidae, and Youngiellidae. Ibid. 7, moore, r. c. ( ed . ) . 1961. Treatise on Invertebrate Paleontology. Part Q. Arthropoda 3, Crustacea, Ostracoda. Geol. Soc. America, Lawrence. 1-442. posner, v. m. 1951. Lower Carboniferous ostracodes of the western part of the Moscow depression. Trudy vses. neftegaz. nauchno-issled. geol.-razv. Inst. 56, 5-108. [In Russian.] sars, G. o. 1866. Oversight af Norges marine Ostracoder. Fork. VidenskSelsk. Krist. (for 1865), 1-130. scott, h. w. 1944. Muscle scar patterns on some upper Paleozoic ostracodes. J. Paleont. 18, 162-171. stover, L. e. 1956. Ostracoda from the Windom Shale (Hamilton) of western New York. Ibid. 30, 1092-1142. wilson, c. w. 1933. Fauna of the McAlester Shale, Pennsylvanian, of Muscogee County, Oklahoma. Ibid. 7,412-422. 59-108. M. N. GRAMM Typescript received 4 August 1975 Revised typescript received 7 April 1976 Institute of Biology and Pedology Far-East Scientific Centre U.S.S.R. Academy of Sciences 690022 Vladivostok, U.S.S.R. i THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations as well as publishing Palaeontology and Special Papers in Palaeontology. Membership is open to individuals and to institutions on payment of the appropriate annual subscription : Institutional membership .... £25-00 (U.S. $50.00) Ordinary membership .... £8-00 (U.S. $16.00) Student membership .... £5 00 (U.S. $10.00) There is no admission fee. Institutional membership is only available by direct application, not through agents. 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. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London NW1 4NS, England. PALAEONTOLOGY All members who join for 1977 will receive Volume 20, Parts 1-4. All back numbers are still in print and may be ordered from B. H. Blackwell, Broad Street, Oxford OX1 3BQ, England, at £9 00 per part (post free). A complete set, Volumes 1-19, consists of 75 parts and costs £675. SPECIAL PAPERS IN PALAEONTOLOGY The subscription rate is £15 (U.S. $30.00) for Institutional Members and £7-50 (U.S. $15.00) for Ordinary and Student Members. Individual subscriptions should be placed through the Membership Treasurer, Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London, NW1 4NS, England. Ordinary and Student members only may obtain individual Special Papers from Dr. Rose at reduced rates. Non-members may obtain them at the stated prices from B. H. Blackwell, Broad Street, Oxford OX1 3BQ, England. COUNCIL 1977-1978 President : Professor W. G. Chaloner, Department of Botany, Birkbeck College, London WC1E 7HX Vice-Presidents'. Dr, J. M. Hancock, Department of Geology, King’s College, Strand, London WC2R 2LS Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Treasurer : Mr. R. P. Tripp, High Wood, West Kingsdown, Sevenoaks, Kent TN15 6BN Membership Treasurer : Dr. E. P. F. Rose, Department of Geology, Bedford College, Regent’s Park, London NW1 4NS Secretary : Dr. C. T. Scrutton, Department of Geology, The University, Newcastle upon Tyne NE1 7RU Editors Dr. C. P. Hughes, Department of Geology, Sedgwick Museum, Cambridge CB2 3EQ Professor J. W. Murray, Department of Geology, The University, Exeter EX4 4QE Professor C. B. Cox, Department of Zoology, King’s College, Strand, London WC2R 2LS Dr. M. G. Bassett, Department of Geology, National Museum of Wales, Cardiff CF1 3NP Other Members of Council Dr. M. C. Boulter, London Dr. P. J. Brenchley, Liverpool Dr. C. H. C. Brunton, London Dr. J. C. W. Cope, Swansea Dr. G. E. Farrow, Glasgow Dr. R. A. Fortey, London Dr. G. P. Larwood, Durham Dr. S. C. Matthews, Bristol Dr. I. E. Penn, London Dr. R. E. H. Reid, Belfast Dr. R. B. Rickards, Cambridge Dr. E. B. Selwood, Exeter Dr. G. D. Sevastopulo, Dublin Dr. P. Toghill, Church Stretton Overseas Representatives Australia : Professor B D. Webby, Department of Geology, Sydney University, Sydney, N.S.W., 2006 Canada: Dr. B S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW, Calgary, Alberta India: Professor M. R. Sahni, 98 Mahatma Gandhi Marg, Lucknow (U.P.), India New Zealand: Dr. G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt West Indies and Central America : Mr. John B. Saunders, Geological Laboratory, Texaco Trinidad, Inc., Pointe-a-Pierre, Trinidad, West Indies Western U.S. A. : Professor J. Wyatt Durham, Department of Paleontology, University of California, Berkeley 4, California Eastern U.S. A. : Professor J. W. Wells. Department of Geology, Cornell University, Ithaca, New York South America: Dr. O. A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela Palaeontology VOLUME 20 PART 2 CONTENTS Evolution in carnivorous mammals R. J. G. SAVAGE 237 Epidermal studies in the interpretation of Lepidophloios species B. A. THOMAS 273 Zvgospira and some related Ordovician and Silurian atrypoid brachiopods p. copper 295 Corallian (upper Jurassic) marine benthic associations from England and Normandy f, r. fTrsich 337 The conifers Frenelopxis and Monica in the Cretaceous of Portugal K. L. ALVIN 387 Ostracod assemblages and the deposilional. environments of the Headon, Osborne, and Bembridge Beds (upper Eocene) of the Hamp- shire Basin M. c. keen 405 A review of the ecology of upper Carboni- ferous plant assemblages, with new delta from Strathclyde a. c. SCOTT 447 A new family of Palaeozoic ostracods M. N. GRAMM 475 Printed in Great Britain at the University Press , Oxford by Vivian Ridler, Printer to the University c ij q Z I1SNI NVIN0SH1IIAJS S3IHVaan LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSH1IW z r- z r* z RIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSH1IIMS SBiavaaiT LIBRARIES SMITHSONIA co co v Z . LliSNI NVINOSH1IIAIS SBiavaaiT L — co — 'Tr>s %v — ■iS s 8v l s® ^ - X IX o co z Z -4 . q CO ^ O .< ■'o^y 5 B R A R I E S ^SMITHSON I AN INSTITUTION NOlinillSNI NVINOSH1IW CO ^ co — CO Z *- _l z _j Z _J z cRIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSH1I1AIS SBiavaail LIBRARIES SMITHSONIA Z r~ z Z r- Z LliSNI NVINOSH1HMS SBiavaail LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSH1IIA z co z w z co z < s ?/&/,* <0*. s < ,v 2 < > ' 5 co z co z co '•*’* z co ^ R I E S SMITHSONIAN INSTITUTION NOlinillSNI _ NVINOSH1IWS SBiavaaiT LIBRARIES SMITHSONIA co — co z \ 00 Z co uj co V uj /n >J^X. Li-J //jffijfr 1I1SNI NVINOSH1IWS S3 lava an LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI nvinoshiw DO X > 00 yy m co '^nxnv£/ m Xffc p.c>x ^ m co \ ? CO ± CO LRIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHillNS SBIBVaail LIBRARIES SMITHSONIA co z V. CO z CO z < V 2 S < I ; vi; S/.W O ;g%|c v' co , .r 4// co O z > z 3/ z ytt > ' " 1I1SNI NVIN0SH1IWS"S3 I ava a II- LI B RAR I ES 'SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHilt' co — _ co z co O x^vv pcy ~ O x^n py' x^ui. ls^ q 2 'X _j 2 _j 2 -J 2 ARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHilWS SBIBVaail LIBRARIES SMITHSONIA z Z z z , z 1- Z O ~ zr^o^ox 2 rr, 2 _ Z'ooTSX o 00 Atr . on tr 00 o w _ o vvy ~ o _ 2 _J 2 _| 2 avaan libraries Smithsonian institution NoiiniusNi nvinoshiiws S3iavaan libraf . 2 r- riTUTiON NonniiiSNi NViNOSHims saiavaan libraries Smithsonian institution Nouni co - ^ > % *■- z co 2 co _ avaan libraries smithsonian institution NoiiniusNi nvinoshiiws saiavaan librai ^ 2 „ W - _ 5 vv 00 2 co m .« to ^ 00 ,^!v\ u 00 o 2 _l 2 -J Z _ riTUTION NOlinillSNI NVIN0SH1IWS S3iavaaiT LIBRARIES SMITHSONIAN INSTITUTION NOlini! TV > TO JS/ — m ^ m x^yosv^x ^ m c/> — co ± co avaan libraries smithsonian institution NoiiniusNi nvinoshihais saiavaan libraf 2 ^ 00 2*0 O Z CO 2 < 2 xSSToX < ^ 2 < VW > i =.* - ' i > CO 2 CO *“ 2 CO 2 CO "ITUTION NOlinillSNI NVIN0SH1IWS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION NOlinil co o 2 avaan libraries smithsonian institution NoiiniusNi nvinoshiiws saiavaan libraf r- 2 r- z r* titution NoiiniusNi NviNOSHiiiNS saiavaan libraries smithsonian INSTITUTION NOlini 2 2 — f - •/ / -<*« ^ g O o jE 3» ' 2 V Z CO 2 CO _ _ avaan libraries smithsonian institution NoiiniusNi nvinoshiuns saiavaan librai 5 ^ 00 - ^ co ^ - CO 2 ^Ticrvrrx. I I I /Til I I I ^TiC wTTx .V-V- I . I - — — co o z 1I1SNI NVINOSHIUNS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION NOlini r; 2 1- 2 1 z co 00 /oo *oX cj z: /XtfW