Bulletin of the British Museum (Natural History Zoology series Vol 41 1981 British Museum (Natural History) London 1982 Dates of publication of the parts No 1 . . .27 August 1981 No 2 . . 24 September 1981 No 3 . 29 October 1981 No 4 26 November 1981 No 5 17 December 1981 ISSN 0007-1498 Printed in Great Britain by Henry Ling Ltd, at the Dorset Press, Dorchester, Dorset Contents Zoology Volume 41 Page No 1 Anatomy and phylogeny of the Chinese Major Carps Ctenopharyn- godon Steind., 1866 and Hypophthalmichthys Blkr., 1860 By Gordon Howes ......... 1 No 2 Morphology and systematics of some interior-walled cheilostome Bryozoa By P. L. Cook & P. J. Chimonides 53 No 3 Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae By Ailsa M. Clark 91 No 4 Miscellanea Observations on clonal cultures of Euglyphidae (Rhizopoda, Protozoa) By C. G. Ogden 137 The Nuttall and British Museum (Natural History) tick collections: lectotype designations for ticks (Acarina: Ixodoidea) described by Nuttall, Warburton, Cooper and Robinson By J. E. Keirans & B. E. Brewster ... 153 A revision of the spider genus Hispo (Araneae : Salticidae) By F. R. Wanless 179 A revision of the spider genus Phaeacius (Araneae : Salticidae) By F. R. Wanless 199 The protractor pectoralis muscle and the classification of teleost fishes By P. H. Greenwood & G. V. Lauder Jnr 213 No 5 Miscellanea The ostracod genus Loxoconcha from Abu Dhabi lagoon, Persian Gulf By R. H. Bate & A. Gurney . 235 A revision of the spider genus Cocalus (Araneae : Salticidae) By F. R. Wanless . 253 The Phthiracarus species of C. L. Koch By B. W. Kamill ... ... .263 The status ofLamingtona lophorhina McKean & Calaby, 1968 (Chirop- tera : Vespertilionidae) By J. E. Hill & K. F. Koopman 275 The status of Hipposideros galeritus Carter, 1846 and Hipposideros cervinus (Gould, 1854) (Chiroptera: Hipposideridae) By P. D. Jenkins & J. E. Hill 279 XV™ * Bulletin of the British Museum (Natural History) Anatomy and phylogeny of the Chinese Major Carps Ctenopharyngodon Steind., 1866 and Hypophthalmichthys Blkr., I860 Gordon Howes Zoology series Vol 41 No 1 27 August 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1981 ISSN 0007-1498 Zoology series Vol 41 No 1 pp 1-52 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 27 August 1981 GENERAL Anatomy and phylogeny of the Chinese Major Carps Ctenopharyngodon Steind., 1866 and Hypophthalmichthys Blkr., 1860 Gordon Howes Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Contents Synopsis 1 Introduction 1 Abbreviations used in the figures 2 Study material 3 Comparison of selected characters in Ctenopharyngodon and Hypophthalmichthys 4 Analysis of apomorph characters 4 Phylogenetic relationships 38 Relationships of Ctenopharyngodon 38 Synopsis of taxa assigned to the Squaliobarbine group 40 Relationships of Hypophthalmichthys 41 Synopsis of taxa assigned to the Abramine group 45 Discussion 47 Classification of the Leuciscinae 47 Leuciscine hybrids 48 Classification of the 'true' carps 49 Acknowledgements 49 References 49 Synopsis The Chinese major carps, Ctenopharyngodon and Hypophthalmichthys, long held to belong to the subfamily Leuciscinae are each shown to represent separate lineages of a diphyletic Cyprinidae. This result is based on the analysis of 22 character complexes and the identifica- tion of supposed apomorphies. Ctenopharyngodon forms part of a monophyletic assembly termed the squaliobarbine group and Hypophthalmichthys is identified as the most derived member of the abramine group. The squaliobarbines are considered to be the sister group of all other barbelled Cyprinidae. The incongruity of character sets results in the presentation of two hypotheses for abramine intrarelationships. Hypophthalmichthys possesses a unique form of epibranchial organ; derived epibranchial structures are also identified in Xenocypris and Abramis. In the light of the data presented, the classification of the subfamily Leuciscinae is evaluated and many identifications of 'intergeneric' hybrids are falsified. Some comment is made con- cerning the phylogeny of barbelled cyprinids. Introduction The subjects of this study are the Chinese major carps, Ctenopharyngodon (the grass carp or white-amure) and Hypophthalmichthys (the silver and big-head carps). The object is to put forward a hypothesis of relationships of these taxa based on synapomorphic characters. Bull. Br. Mus. nat. Hist. (Zool.)41 (1) : 1-52 Issued27 August 1981 2 G. J. HOWES The major Chinese carps are of considerable economic importance as sources of protein and, in the case of Ctenopharyngodon, a control agent in restricting pond and canal weed growth. That these carps are primary-consumers and breed well in captivity are additional assets in their culture. Although there is much information on their biology-breeding, physiology, biochemistry and culturing, there are very few papers describing their anatomy. Of these only two or three authors have made statements regarding possible phylogenetic relationships. Ctenopharyngodon is a monotypic genus included by Chu (1935) in the subfamily Leuciscinae. Berg (1949) considered the genus to be closely related to Scardinius and placed it in his all-embracing Cyprinini. Gosline (1978) placed Ctenopharyngodon, Mylo- pharyngodon and Squaliobarbus together, and although he was undecided whether they belonged to the subfamily Cyprininae or Leuciscinae, favoured retention in the latter. Even though the characters Gosline used for uniting these genera are considered by the present author to be a mosaic of plesio- and apomorphies, his hypothesis of relationships is not refuted in this study. Hypophthalmichthys includes three nominal species, H. molitrix (Val.), H. nobilis (Richardson) and H. harmandi Sauvage. Oshima (19 19) established the genus Aristichthys to contain nobilis, believing that differences in gill-raker form, abdominal keel length and pharyngeal dentition from those in molitrix were large enough to signify generic rank. Authors have been at variance in treating Aristichthys as a valid genus and many still regard nobilis as a species of Hypophthalmichthys. I would agree with the latter view. The taxa molitrix and nobilis possess unique synapomorphies and thus belong to the same higher category, ie. the genus Hypophthalmichthys. The species H. harmandi appears to have been overlooked by subsequent authors. To my knowledge it is known only from the type specimen from Hanoi, now in the Paris Museum. Giinther (1868) first introduced the category Hypophthalmichthina to contain Hypo- phthalmichthys which taxon Gill (1893) later recognized as a subfamily, the Hypophthal- michthinae. Regan (1911; 1922) thought Hypophthalmichthys to be related to Barilius — a suggestion refuted by Howes (1980). Kryzanovskij (1947) referred Hypophthalmichthys to the Leuciscinae on evidence derived from ontogenetic development. Ryabov (1973) agreed with this subfamilial placement, using as his evidence the viability of larvae produced by cross-breeding Hypophthalmichthys and Abramis — Abramis being considered a typical leuciscine. Gosline (1978), relying on a tri-lobed swimbladder as an indicator of this relationship, considered that there were 'clear indications of a cultrin derivation for Hypophthalmichthys and Aristichthys'. A tri-lobed swimbladder, however, has such a mosaic distribution throughout the Cyprinidae that it has no value as a character indicating shared common ancestry. Besides which, the third 'lobe' in Hypophthalmichthys is a mere constriction or 'tail' of the swimbladder and is variously developed in individual fish. There has been only one study of the comparative osteology of the major Chinese Carps, that by Watanabe (1951). Although this author pointed out significant differences between Ctenopharyngodon, Hypophthalmichthys and Mylopharyngodon, and noted the possible taxonomic value of certain characters, he made no attempt to utilize them in formulating relationships. Finally, it can be noted that artificial hybridization between Ctenopharyngodon and Hypophthalmichthys has been achieved and the morphology of the hybrids described by Berry & Low (1970). The consensus of ichthyological opinion it would seem, is that the two major carps, Ctenopharyngodon and Hypophthalmichthys belong to the same subfamily, the Leuciscinae. My thesis is that they do not, but that each represents a distinct monophyletic lineage, representative of a basic dichotomus division within the Cyprinidae. Abbreviations used in the figures AA Anguloarticular Aap Adductor arcus palantini muscle AEW Anterior wall of epibranchial organ Ah Adductor hyomandibularu ; muscle CHINESE MAJOR CARPS AMP Anterior medial maxillary process LEN BO Basioccipital LOF BOCB Basioccipital cartilaginous bloc ME BOK Basioccipital keel MET BON Basioccipital notch MF BP Basioccipital pharyngeal process MMP BUG Buccal cavity MP CB Ceratobranchial (numbered) N CF Ceratobranchial filaments NC CFE Carotid fenestra NS CP Coronoid process OC CPE Cartilaginous preethmoid OFT Dahm Dorsal section of adductor OP hyomandibularis OPE DPTC Dermopterotic canal PALH EB Epibranchial (numbered) PAR EBD Epibranchial duct (numbered) PE EBO Epibranchial organ PEW EC Ethmoid cartilage PF ECT Epibranchial connective tissue PFO EOD Epibranchial oesaphagal duct PH EOW Epioccipital wing PHF EP Epioccipital Php EPOP Epioccipital process PO ER Epibranchial rakers POC ES Extrascapular PPSF EXO Exoccipital PRO EXOF Exoccipital flange PS F Frontal PSAP FC Frontal canal PSCB FF Frontal foramen PTE FIC Internal carotid foramen PTES FJ Jugular foramen PTS FII Optic foramen PTT FV Foramen for trigeminal trunk PVP FVII Foramen for facial nerve trunk Q FVIII Foramen for acoustic nerve SE FIX Foramen for glossopharyngeal SEW GR Gill raker (globular) SM HMF Hyomandibular fossa SO HYO Hyomandibular SOCC HVII Hyomandibular branch of VII nerve SOP 1C Intercalar SOR ICF Intercalar fossa SP IF Infrapharyngobranchials (numbered) V IO Infraorbitals (numbered) Vahm IOP Interoperculum LE Lateral ethmoid VPSR LEF Lateral ethmoid/frontal fenestra Supraethmoid notch for palatine head Lateral occipital foramen Mesethmoid Metapterygoid Maxillary foramen Midlateral maxillary process Masticatory plate Nasal Neural complex Neural spine Intramural opercular canal Olfactory foramen Operculum Ossified preethmoid Palatine head Parietal Preethmoid Posterior wall of epibranchial organ Palatal fold Prootic foramen Pharyngeal bone Fossa in pharyngeal bone Pharyngeal pad Preoperculum Preopercular canal Prootic/paraspenoid fenestra Prootic Parasphenoid Parasphenoid ascending process Parasphenoid keel Pterotic Pterotic spine Pterosphenoid Posttemporal Pterosphenoid ventral process Quadrate Supraethmoid Supraethmoid wing Ceratobranchial sifting membrane Supraoccipital Supraoccipital crest Subopercle Supraorbital Sphenotic Vomer Ventral section of adductor hyomandibularis muscle Ventral parasphenoid ridge Study material All specimens used in the preparation of this paper are in the osteological and alcohol collections of the British Museum (Natural History). Lists of comparative material will be 4 G. J. HOWES found in previous publications (Howes, 1978; 1979). The catalogue numbers for specimens used in illustrating this work are given in the figure captions. Comparison of selected characters in Ctenopharyngodon and Hypophthalmichthys The following 22 characters were selected after a preliminary study had eliminated others which appeared to contain minimal information concerning their plesio- or apomorphic state (ie. intra- and interspecifically variable or mosaically distributed), and those which were considered to be symplesiomorphies. Guidance to the identification of plesiomorphic character state came both from previous studies on cyprinid phylogeny (Howes, 1978; 1979; 1980) and wider outgroup comparisons — outgroup in this case being other otophysans. Likewise, these earlier studies had led to the final selection of those characters which had previously been useful as phylogenetic indicators. Reasons for polarity assignment are present in the analysis below. Selected characters: 1. Ethmo-vomerine region; 2. Pterosphenoid; 3. Parasphenoid; 4. Prootic; 5. Subtemporal fossa and associated musculature; 6. Dilatator fossa; 7. Frontal; 8. Parietal; 9. Circumorbitals; 10. Supraoccipital; 11. Exoccipital; 12. Epioccipital; 13. Intercalar; 14. Basioccipital; 15. Posttemporal; 16. Opercular series; 17. Suspensorium; 18. Jaws; 19. Vertebral column; 20. Gill-arches; 21. Adductor hyomandibulae muscle; 22. Pectoral girdle. Analysis of apomorph characters 1 . Ethmo- vomerine region Ctenopharyngodon (Fig. 1); supraethmoid broad and short with a shallow medial anterior notch and two lateral notches; preethmoids paired, largely cartilaginous; kinethmoid dorsally expanded; vomer widely V-shaped anteriorly; lateral ethmoid wall truncate. Hypophthalmichthys (Fig. 1); supraethmoid broad and short with a deep medial notch; preethmoid single, small, completely ossified; kinethmoid rod-shaped; vomer thickened anteriorly and forming the floor of the median ethmoid notch; lateral ethmoid contains entire olfactory foramen in a bony tube; the lateral lamella is prolonged distally and curves ventrally; fenestra between the dorsal surface of the lateral wall and the frontal. Both Ctenopharyngodon and Hypophthalmichthys possess a short, broad supraethmoid and depressed mesethmoid. This is hypothesized as the plesiomorph condition in cyprinids on the grounds of widespread occurrence. Within the Cyprinidae there are two basic forms of supraethmoid; (1) in which the anterior border of the bone is indented medially so as to accommodate the retracted kinethmoid (sometimes there may be indentations lateral to the medial one), and (2) where the supraethmoid is rostrally extended, its lateral border rounded and the tip of the extension bearing the median notch. These two types are illustrated in Fig. 2. The first morphotype is that most commonly encountered in non-barbelled Cyprinidae (ie. Leuciscinae), whilst the second is that found in almost all barbelled Cyprinidae (ie. Cyprininae; Barbus, Labeo etc). In Cyprinus, Carassius and several members of the Gobioninae, although there is a prominent rostral process there is no, or but a slight, medial notch — the kinethmoid being attached to the base of the rostral process by a single ligament (oberes Russelband of Fiebiger, 1931). This is similar to the condition in the Catostomidae where the supraethmoid is elongate and without a medial indentation. A rostrally produced supraethmoid is the common condition amongst characoids and a comparison between the ethmoid region of Brycon and that of Catostomus shows a remarkable similarity. In siluroids the common condition, as in cyprinoids, appears to be one where the anterior ethmoid border bears a slight medial indentation. Within the cyprinids, although the extremes of the two ethmoid morphotypes can be readily appreciated, they are linked by a series of intergrades. Particular examples are in Garra LEN CHINESE MAJOR CARPS PE2 PE1 PALH SE Fig. 1 Ethmo-vomerine regions in dorsal view: (upper) Ctenopharvngodon idellus; (lower) Hypophthalmichthys molitrix. Specimens BMNH 1888.5.15:25; 1895.5.31:22. Scales = 10 mm. where the rostrally produced ethmoid shape is still discernible but the anterior border is deeply indented as in 'leuciscines'. Again, in some species of Barbus and Schizothorax the ethmoid is distinctly leuciscine-like. An extreme morphotype is encountered in Semiplotus and Scaphiodonichthys where the entire supraethmoid is almost vertically directed and it is the anterior border of the frontal which is medially indented. In general, it appears that cyprinids with a terminally placed mouth are those with a deeply notched ethmoid. On the other hand, the rostrally produced ethmoid is more characteristic of bottom feeding types where additional, almost ventral extension of the mouth is functionally desirable — as in Semiplotus mentioned above. Fink & Fink (198 1) propose that a rostrally produced ethmoid is plesiomorphic for otophysans. It would appear that in cyprinids a supraethmoid with a slight, but broad rostral process is the plesiomorph state. This plan would offer the potential of producing either kind of morphotype. The variability of the supraethmoid in species of Barbus is discussed by Banister (in preparation). In that group, at least, the ethmoid appears to be a very labile structure and one responding to environmental influences. In Ctenopharyngodon there are lateral indentations on either side of the medial supraethmoid notch (LEN, Fig. 1). These lateral notches accommodate the enlarged anterior condyles of the palatines. A similar feature occurs in Squaliobarbus and Mylopharyngodon and is considered synapomorphic. G. J. HOWES PE2 PE1 SE Fig. 2. Ethmo-vomerine regions in dorsal view: (upper) Cyprinus carpio; (lower) Leuciscus cephalus. Composite drawing of several specimens. Scale = 5 mm. Concerning the posterior border of the supraethmoid, a transverse suture with the frontal appears to be the plesiomorph state. Only in some chelines is the posterior margin rounded or triangular and overlapped by the frontals (see Ramaswami, 19556, fig. 3, and Howes, 1979). Amongst characoids the posterior ethmoid margin is triangular, that of siluroids is irregular, its medial part usually invaded by the anterior tip of the frontal fontanel. A rod-shaped kinethmoid is considered the plesiomorph type (see Howes, 1978). I am unaware of any detailed study concerning the possible derivation or homology of the bone. Goodrich (1909) thought it to be a specialized part of the ethmoid. The common occurrence of a bifurcated kinethmoid suggests that there has been a phylogenetic ossification into a single unit of two premaxillary-ethmoid ligaments. There is no cyprinoid known to me in which these ligaments remain unossified in the adult, or even in larval stages, and which would represent the most plesiomorph state. In its various derived states the dorsal part of the kinethmoid may be expanded in a transverse plane, or the medial part of the bone may be swollen. In Ctenopharyngodon the kinethmoid is dorsally expanded and thickened, a derived condition encountered in other cyprinid groups (eg. chelines; see Howes, 1979). In Hypo- phthalmichthys the bone remains as a slender rod. Ramaswami (1955a) noted the absence of a preethmoid in Hypophthalmichthys, but this is not the case. A preethmoid cartilage is present, although minute and only fully ossified in the adult. Likewise, in other genera which Ramaswami (1955a & b) cites as lacking a preethmoid, I have found it as a poorly ossified element. Ossification of the preethmoid is variable; specimens ofAlburnus show no sign of ossification until they are at least 56 mm SL; CHINESE MAJOR CARPS EC Fig. 3 Development of the ethmoid bloc in Barbus sp. Anterior view (top) from a specimen 13mm SL; (centre) specimen 20mm SL; (bottom) specimen 35mm SL. Dashed areas = cartilage, stippled areas = ossification. Scales = 0-5 mm. the preethmoid in Oreoleuciscus is entirely cartilaginous in a specimen of 1 14 mm SL. It is still recognizable as a preethmoid, being clearly differentiated from the ossified meseth- moidal process. According to Alexander (1966) the preethmoid is an ossification of the antero-lateral process of the ethmoid cartilage (mesethmoid). He notes that in Leuciscus (=Idus of Alexander), the maxillary ascending process articulates with the preethmoid cartilage, and with the palatine via a cartilaginous insert (submaxillary meniscus of Alexander). Corrobora- tion of Alexander's (1966) statement that the preethmoid is an ossification of the mesethmoid comes from an ontogenetic series of Barbus sp. (Fig. 3). In the smallest specimens, 12-14 mm SL, the preethmoid is not differentiated from the ethmoid cartilage and there is a sub-terminal ossification, in 20 mm specimens there is dorsal ossification of the lateral mesethmoid processes, greater area of terminal ossification and antero-lateral extension of the vomer. In specimens of between 25 mm and 27 mm SL there exists only a small area of cartilage between the mesethmoid, ossified preethmoid and vomer. At the 30-35 mm stage, the antero-lateral mesethmoid processes have contacted the now large and completely ossified preethmoids and the vomer. A lateral mesethmoid-vomerine fossa is formed and the preethmoid contacts the fossa via a thin meniscus of cartilage. In all cyprinids I have examined, other than abramines (see below), the preethmoid is contained in a lateral fossa formed from the mesethmoid and the vomer (see Howes, 1 978). A lateral ethmo-vomerine fossa is lacking in Abramis, Rutilus, Chondrostoma and Xenocypris. In these taxa the vomer does not protrude laterally (see below) but extends forward; the preethmoid is extensive and covers the extended vomerine arm, contacting the mesethmoid only along its antero-ventral rim (see Figs 5A-D). The preethmoid is almost entirely cartilaginous with only the anterior tip being ossified in the case of Abramis, Rutilus G. J. HOWES N OF SE LEF PE Fig. 4 Hypophthalmichthys molitrix, lateral view of ethmo-vomerine region. Specimen 1895.5.31 : 22. Scale =10 mm. and Chondrostoma, but a laterally ossified nub in Xenocypris. Although in Hypo- phthalmichthys there is a distinct reduction of the lateral fossa the preethmoid cartilage still contacts the lateral part of the mesethmoid (Fig. 4). Laterally, the preethmoid articulates directly with the palatine facet; only rarely is such a connection lacking (eg. Ochetobius); there is never direct contact with the maxilla. In some cyprinids, however, the preethmoid is extended anteriorly and may ossify into two parts. Such paired preethmoids were recorded in Cyprinus by Fiebiger (1931) as the lateral 'seitliches Knotchen' articulating with the palatine, and the anterior 'Sesamknoten' articulating with the maxilla. Paired preethmoids also occur in Ctenopharyngodon (Fig. 1), Mylopharyngodon, Squaliobarbus and some Barbus species. In these taxa, the anterior preethmoid is often unossified or only partially so. According to Ramaswami (1955a) it is the cartilage sandwiched between the maxilla and the preethmoid which ossifies to form a second preethmoid. However, in the taxa cited above, the additional, anterior ossification would appear to be within the original preethmoidal cartilage. In gobioinids, cobitids and catostomids, the second preethmoid can be exceptionally well- developed and exists as an ossified strut (the septomaxilla of Regan, 1911 and Starks, 1926). Patterson (1975 : 501) reckons the preethmoids of cyprinoids to be neomorphs. As such they must be regarded as autapomorphic for Cyprinoidei. A single, lateral ossification of the ethmoid cartilage articulating with the palatine is hypothesized as representing the plesiomorph condition (see Howes, 1980). The mesethmoid is usually a triangular bloc, its apex forming the rostrum of the entire ethmoid bloc. In those taxa with a medially indented supraethmoid, the mesethmoid becomes thickened anteriorly and forms a somewhat hour-glass shape. In some leuciscines, the mesethmoid is laterally excavated to such an extent that it is virtually a medial septum. The posterior wall of the mesethmoid cartilage is, in one group of leuciscines and some cyprinines, excavated to form the anterior myodome (see below). The lateral ethmoid in cyprinids most often has a well-produced lateral wall with a broadly-triangular base; antero-medially it meets the mesethmoid and contains part of the olfactory nerve foramen in its leading edge; postero-medially it meets its fellow along the midline. In some cases the lateral ethmoids are indented into the mesethmoid cartilage so as to form an anterior myodome. Even though, as in aspinines, the anterior myodome extends CHINESE MAJOR CARPS V OPE CPE- Fig. 5 Ethmo-vomerine regions in dorsal and lateral views: A. Abramis brama; B. Xenocypris argenteus. The lateral views are semi-diagrammatic and show the anterior edge of the vomer as a dashed line, the cartilaginous preethmoid as cross-hatched and the ossified preethmoid as black. Composites, from skeletal and alizarin specimens; all drawn to same scale. well forward, the lateral ethmoids remain in medial contact. However, in Barbus barbus, Cyprinus and Carassius, the myodome invades the mesethmoid cartilage through a parting between the lateral ethmoids. Patterson (1975) was unable to conclude whether the absence of an anterior myodome was a primitive condition or due to secondary loss (apomorphic). Its presence does not seem necessarily to be correlated with any particular degree of ethmoidal elongation, nor its absence with any shortening. In noemacheilids, gobioinids and catostomids, an anterior myodome is lacking. Widespread, mosaic distribution in cyprinids and almost universal presence in characoids probably indicates that an anterior myodome is a plesiomorph feature. In Ctenopharyngodon and Mylopharyngodon the olfactory foramen is entirely counter- sunk within the lateral ethmoid, whereas in Hypophthalmichthys it projects forward as a bony tube (Fig. 4). In the latter genus there is a tubular opening between the medial dorsal part of the lateral ethmoid wall and the overlying frontal (LEF, Fig. 4). This tube contains a thick ligament which covers the top of the lateral ethmoid wall and the outer part of the supraorbital. A similar feature occurs in Xenocypris (Fig. 5B). In cyprinids the olfactory foramen is most often formed from the anterior margin of the mesethmoid. In some taxa, however, the foramen is confined entirely within the lateral ethmoid. This occurs in the taxa cited above, in alburnine genera (Alburnus, Chalcalburnus) and some Barbus. In Barbus tropidolepis the foramen is also housed in an anteriorly projecting tube (Banister, pers. comm.). On the basis of other sets of synapomorphic characters it is hypothesized that the occurrence of a tubular olfactory foramen in diverse taxa is of no phyletic significance. A short and broad vomer is hypothesized as plesiomorphic eg. as in Ctenopharyngodon and Hypophthalmichthys. However, in Hypophthalmichthys the vomer extends anteriorly to floor the median ethmoid notch, its anterior margin is almost straight and is curved upward so as to face the lower part of the ethmoid bloc. Laterally, the vomer is indented to contain the preethmoid. An anteriorly extended vomer, flooring the median ethmoid notch is a rarely encountered feature amongst cyprinids. In some cultrines (Culler) and in Ochetobius (currently included in the Leuciscinae), the vomer forms the floor of a shallow but wide 10 Fll PVP G. J. HOWES VPSR PTS FJ FIC HMF PTE PRO Fig. 6 Ctenopharyngodon idellus, ventral view of anterior otic region. Specimen BMNH 1888.5.15:25.Scale=10mm. ethmoid notch. In Xenocypris and Abramis, the vomer also floors the ethmoid notch but the mesethmoidal walls are deep, almost meeting one another ventrally and leaving only a narrow area of the vomerine floor visible from above (Figs 5A & B). As mentioned above (p. 7) Xenocypris, Abramis, Chondrostoma and Rutilus differ from other cyprinids in employing the vomer as the platform for the preethmoid and excluding the mesethmoid from taking part in its support. In these genera the vomer is deeply forked, each arm extending some distance anteriorly from the mesethmoidal border; in transverse plane the bone is markedly convex. 2. Pterosphenoid Ctenopharyngodon (Fig. 6); contact between the pterosphenoid and ascending process of the parasphenoid is via a thin strut; medial contact with the prootic is by a long suture. On the lateral face of the pterosphenoid is a small, ventrally directed process (autapomorphic feature; PVP, Fig. 6). Hypophthalmichthys (Fig. 7); contact with the parasphenoid is via a thick, anteriorly directed pillar-like process. The part to the anterior trigemino-facialis foramen, and a medial portion of the bone provides the inner wall of the anterior part of the chamber; this is a unique condition in cyprinids. Howes (1979) presented a classification of the various types of contact between the parasphenoid and pterosphenoid, assuming that this classification reflected a phylogenetic sequence for the Cyprinidae as a whole. Further analysis has shown that similar sequential series can be detected in all groups so far identified as monophyletic, ie. the plesiomorphic condition is one with no contact between para- and pterosphenoid (Type 1 and derived Type 1 in Howes, 1980) and the most apomorphic condition involves extensive lateral contact between the pterosphenoid and parasphenoid (Type 5 in Howes, 1980). A pterosphenoid pedicle is present in various teleosts and may or may not contact the CHINESE MAJOR CARPS 11 PRO PPSF PTS PSAP Fig. 7 Hypophthalmichthys molitrix, lateral aspect of anterior otic region, viewed slightly antero- ventrally. Drawn from specimen BMNH 1895.5.31 : 22 with additions from alizarin prepar- ations and dissections from BMNH 1980.5.21 : 5-14. Scale = 10mm. parasphenoid ascending process. Patterson (1975) regards this feature as a plesiomorphic one. 3. Parasphenoid In both Ctenopharyngodon and Hypophthalmichthys the parasphenoid is anteriorly broad, narrowing posteriorly in Hypophthalmichthys but remaining broad in Ctenopharyngodon in which genus it bears a marked ventral medial range. In Hypophthalmichthys the hypophyseal portion of the parasphenoid is ventrally extended into a long deep process, or keel the posterior edge of which is developed into a transverse ridge. The ventral keel serves both as the anterior point of attachment for a medial cartilaginous insert dividing the forward part of epibranchial diverticula (p. 23) and as lateral sites of attachment for the adductor arcus palatini muscle (p. 37). In the majority of cyprinids the central surface of the parasphenoid is grooved, rarely is there a midline ridge or process, and these features are regarded as derived. The anterior broadening of the parasphenoid, although extreme in the Chinese major carps, is a common feature in cyprinids (see Ramaswami, 19556). Weisel (1960) notes that in Catostomus the anterior part of the parasphenoid is considerably wider than that of cyprinids. However, this is not so when compared with the two taxa discussed here. Ctenopharyngodon shares with Mylopharyngodon the developed ventral midline ridge and deep anterior concavity of the parasphenoid, both features regarded as apomorphic. The hypophyseal keel of Hypophthalmichthys is a feature shared with Xenocypris and Abramis, although in these taxa it is developed to a lesser extent (see below). The internal carotid fenestra in Ctenopharyngodon is small and indents the posterior border of the parasphenoid ascending process and the antero- ventral margin of the prootic. In Hypopohthalmichthys the internal carotid artery passes through a small foramen situated in the basal part of the parasphenoid keel (Fig. 7). Patterson (1975) considered that the plesiomorph condition of the carotid opening in 12 G.J.HOWES teleosts is for it to occur as a foramen within the parasphenoid. In nearly all cyprinids I have examined, and in all characoids, the internal carotid passes through a fenestra between the parasphenoid and the prootic, a condition Patterson (1975) thought derived and due to the eroding of the parasphenoid ascending process. In catostomids (see Weisel, 1960) and siluroids (?all) the foramen is situated entirely within the parasphenoid. Although similarly there is a parasphenoid carotid foramen in Hypophthalmichthys I consider this to be a derived condition in this taxon and for the following reasons. In Alburnus, Rutilus, Chondrostoma and Abramis, the internal carotid fenestra is extensive (Fig. 8); it is covered laterally by a fascia of tissue and plugged by a wedge of fat leaving a small ventral opening to allow passage of the carotid vessel. In Abramis there is a tendency for the fenestra to elongate and in Abramis sapa the 'fenestra' is divided; the lower opening being confined to the parasphenoid and carrying the internal carotid artery, the upper indenting the ventral margin of the prootic and not providing passage for any vessel (Fig. 8). It is hypothesized that the condition of the carotid fenestra in Abramis sapa is a derived one and represents a transitional stage toward that in Hypophthalmichthys. That this is the directional path and not the reverse, viz the incorporation of the foramen into a fenestra, as indicated by the morphology of the parasphenoid. In Abramis there is a ventral extension of the hypophyseal part of the parasphenoid similar to that in Hypophthal- michthys (see above), and this is most highly developed in A. sapa (Fig. 8). Furthermore, the carotid foramen in Hypophthalmichthys is situated in the ventral keel, rather than (plesiomorphically) in the ascending process (Patterson, 1975). Similarly, an internal carotid foramen occurs in the same position in Xenocypris but is a larger hole than in Hypo- phthlamichthys. On the basis of this evidence I suggest that the foramen for the internal carotid artery in Hypophthalmichthys and Xenocypris is secondarily derived in its position in the parasphenoid. This positioning has come about both as a consequence of the ventral prolongation of the parasphenoid and the re-orientation of the first branchial arch (p. 31) whereby the plane of the efferent artery would be sagittal rather than transverse. Thus, this shared feature in Hypophthalmichthys and Xenocypris is considered a synapomorphic character. Modifications to the ventral border of the parasphenoid are present in a unit of leuciscine taxa represented by Abramis, Rutilus and Chondrostoma. In these genera, there is a deep ventral channel in the posterior part of the parasphenoid. The channel is confluent with a similar feature in the basioccipital and the whole is filled with a cartilaginous block (Fig. 27). Discussion of this character is reserved for inclusion with that of the basioccipital (p. 23). Mention should be made here of medial parasphenoid processes. In Labeo and related taxa (see Howes, 1980 : 152, fig. 26 and p. 16 below) a process arises from either side of the medial wall of the parasphenoid just anterior to the optic foramen. Each process contacts the ventral border of each respective pterosphenoid. These medial processes vary from thin struts to thick pillars. Only in Pelecus is there a single, medial parasphenoid pillar. These structures are not to be interpreted as homologues of the basisphenoid. Harrington (1955) reported that Holmgren & Stensio (1936) recognized a basisphenoid in some cyprinids, eg. Abramis. However, the structure reported by Holmgren & Stensio is the lateral pterosphenoid strut that makes contact with the parasphenoid wing. The medial parasphenoid extensions of Pelecus and Labeo are considered as neomorphs. 4. Prootic Ctenopharyngodon (Fig. 6); contact with the parasphenoid ascending process is via a long suture. The posterior opening of the trigemino-facialis chamber is large and lies ventrally. The upper posterior part of the bone is deeply concave and forms the entire inner face and part of the roof of the subtemporal fossa. Hypophthalmichthys (Fig. 7); no contact with the parasphenoid ascending process. The posterior trigemino-facialis foramen is small and situated laterally. The upper posterior half CHINESE MAJOR CARPS 13 CFE PTS PFO Fig. 8 Lateral views of anterior otic regions: (top) Rutilus rutilus; (centre) Abramis brama; (bottom) Abramis sapa. Semi-diagrammatic, compiled from several specimens. of the bone is slightly concave and forms only the lower inner face of the subtemporal fossa. The lower part of the anterior border forms the posterior rim of the trigeminal exit fenestra. The plesiomorphic condition of the prootic in cyprinids is hypothesized as an almost square bone, its lateral face hardly depressed, its anterior border interrupted by the trigeminal exit foramen and with a short lateral commissure (see for example, Opsariichthys in Howes, 1980). Derived conditions are those where the anterior trigeminal foramen pierces the lateral face; the lateral commissure is long, and the prootic contributes a ventral portion to the apophyseal platform. The general depression of the otic part of the cranium in Ctenopharyngodon has resulted in the forepart of the prootic being laterally extended (Fig. 6), which, as a consequence, has shifted the orientation of the posterior trigeminal foramen. This opening faces ventrally 14 PTE PTE Fig. 9 Subtemporal fossae: (left) Hvpophthalmichthvs; (right) Ctenopharyngodon. Specimens BMNH 1888.5.31 : 22 and 1888.5.15 : 25. Scales = 10 mm. instead of laterally, so the nerve trunk and jugular vein enter without passing across the prootic face. This derived condition occurs also in Mylopharyngodon and Squaliobarbus. The most outstanding feature of the prootic in Hypophthalmichthys is the exclusion from the bone of the anterior trigeminal foramen. Only the lower part of the prootic forms the border of the opening, the upper part of the foramen being confined to the pterosphenoid. Jn one specimen, the upper part of the foramen, through which mdV and mxV exit is, on one side of the cranium, entirely enclosed in the pterosphenoid. Reduction of the anterior part of the prootic in Hypophthalmichthys appears to have occurred with the posterior movement of the pterosphenoid, itself associated with hypertrophy of the adductor hyomandibularis muscle. The form of the prootic in Hypophthalmichthys is an autapomorphic feature. 5. Subtemporal fossa Ctenopharyngodon (Fig. 9); extensive, hemiovoid with a deep conical anterior portion, or chamber, that extends into the autosphenotic; the levator posterior muscle arising from that bone. The inner wall of the fossa is formed from the prootic, the roof from the pterotic and epioccipital, and anteriorly from the sphenotic. None of these bones meet dorsally and the ventral surface of the parietal is exposed thus also contributing to the Subtemporal fossa roof. Hypophthalmichthys (Fig. 9); almost pyramidal, lacking an anterior sphenotic extension. The posterior, exoccipital border of the fossa is indented by a notch. The bones forming the inner wall of the fossa, prootic, exoccipital and epioccipital, do not meet along their respec- tive dorsal and anterior borders, the space between being a thin cartilage wall. The most frequently occurring Subtemporal fossa morphology amongst the Cyprinidae is one which is deep and circular or ovate. The fossa is the site of origin for levator posterior muscles running to the pharyngeal bone and its size is correlated with that of the muscles' size and their orientation. It is thus difficult to assess the polarity of some features such as the posterior notch, which occurs not only in Hypophthalmichthys but in a variety of Meuciscine' and 'barbine' taxa. There is generally a space between the bones forming the inner wall of the Subtemporal fossa (viz between the posterior-dorsal edge of the prootic, posterior edge of the sphenotic, and antero-dorsal edge of the epioccipital) across which is a thin cartilage and which forms the apex of the fossa. In this feature Ctenopharyngodon and Hypophthal- michthys are each autapomorphic. In the former taxon the space is covered by the parietal CHINESE MAJOR CARPS 1 5 and in the latter taxon the exoccipital contributes substantially to the inner face and its antero-dorsal edge contacts the cartilaginous sheet. The vast subtemporal fossa of Ctenopharyngodon is a feature shared only with Mylopharyngodon and Squaliobarbus. Likewise, in these two genera the fossa extends anteriorly deep into the sphenotic from which the levator posterior musculature originates. The size, shape and sphenotic contribution of the subtemporal fossa in these three genera is considered synapomorphic. The anterior contribution of the sphenotic to the subtemporal fossa occurs also to a greater or lesser degree in Cyprinus, Carassius and in some Barbus, and in the two former taxa part of the levator posterior musculature originates in that anterior extension. This may thus represent a synapomorphy linking all these genera (see p. 49). The shape and depth of the Hypophthalmichthys subtemporal fossa are shared with Xenocypris. However, these parameters are treated with caution as synapomorphies; variations in subtemporal fossa depth cover a wide range throughout the cyprinids. Abramis has what is possibly the deepest subtemporal fossa and again, its pyramidal shape greatly resembles that of Hypophthalmichthys. Neither in Xenocypris nor Abramis does the exoccipital contribute substantially to the inner wall of the subtemporal fossa. Only in Rutilus does the anterior edge of the exoccipital penetrate forward so as to separate the prootic and epioccipital. 6. Dilatator fossa Ctenopharyngodon (Fig. 10); formed mostly from the sphenotic which extends laterally with a small anterior contribution from the frontal; the frontal border forms a medial roof to the fossa. Hypophthalmichthys (Fig. 10); formed in almost equal parts from the frontal and sphenotic and sloped antero-laterally as a broad shelf. The plesiomorph condition of the dilatator fossa in cyprinids is considered to be one in which the fossa is small and formed from nearly equal portions of the frontal, sphenotic and pterotic, and to be roofed or partially roofed by the frontal. This type of fossa morphology is widespread amongst cyprinids and is represented in Ctenopharyngodon. An almost identical type of dilatator fossa morphology in characoids is regarded by Vari (1979) as plesiomorphic. In siluroids there is no cranial fossa, the dilatator operculi muscle originating from the ventral cranial surface (see Alexander, 1964). Similar ventral muscle origin is to be found in some cyprinids and characoids (see below; Alexander, 1964; Howes, 1978). In most catostomids the dilatator fossa is well developed and involves the sphenotic and the medial part of the frontal. Cobitids (sensu lato) possess a small fossa of the type which involves mostly the sphenotic. It would seem that from the plesiomorphic type of dilatator fossa two avenues of develop- ment were open. One was toward extended coverage of the cranial roof, involving mostly the frontal. The second was for the frontal to become invaginated ventro-medially and be pierced by the dilatator operculi. The most derived state of the first morphotype in cyprinids is that in Hypophthalmichthys where the major contributing elements to the fossa are the frontal and sphenotic, the pterotic being virtually excluded. The sphenotic extends laterally as a broad, ventrally sloping shelf confluent with the frontal. The lateral part of the frontal also slopes and medially bears a ridge which forms the border of the fossa. A similar form of dilatator fossa occurs in Xenocypris and Abramis, but in these genera the frontal and sphenotic have a more acute ventral slope. The most derived state of the second morphotype is that where a divided dilatator operculi muscle passes through a frontal foramen and a fenestra between the frontal and sphenotic (see below). These two morphotypes embrace the four enumerated by Howes (1978). Type 1 is described above as the plesiomorph condition and Types 2 and 3 are morphoclinal stages in the encroachment of the cranial surface. Within the second lineage (Type 4 of Howes, 1978), in which the dilatator operculi 16 G. J. HOWES Fig. 10 Dilatator fossae: (upper) Ctenopharyngodon; (lower) Scales = 5 mm. Specimens as previous figure. Hypophthalmichthys. originates from the ventral surface of the frontal, a possible transitional sequence can be distinguished. An early stage is hypothesized as being represented by Squaliobarbus. Here the frontal is deeply indented anteriorly to the sphenotic and contains a small foramen which enters the orbital cavity. In Barbus and Capoeta the muscle has 'broken through' the frontal to attach to its ventral surface inside the orbital cavity. A previous statement (Howes, 1978) to the effect that the frontal and sphenotic in Barbus contacted each other only along their lateral margins is incorrect. The anterior margin of the sphenotic is bordered by the frontal, albeit in some instances an extremely thin segment (Fig. 1 1). Carassius and Cyprinus appear to mark a further stage in development of the foramen for there is in some specimens an almost complete separation between the frontal and sphenotic, the ventro-lateral part of the frontal extending only part way along the sphenotic margin. Cyprinus is the more derived taxon in this respect since it has an almost complete frontal-sphenotic foramen (Fig. 1 1). Labeo and Garra appear to represent the most derived state, with the dilatator operculi being divided, the upper section passing through a frontal foramen, the lower through a foramen between the frontal and a strut-like sphenotic process (Fig. 1 1 ). The doubly foraminate forms CHINESE MAJOR CARPS 17 SPF PRO Fig. 11 Dilatator fossae: (upper) Cvprinus carpio (uncatalogued); (centre) Capoeta capoeta BMNH 1 879. 11.14:19; (lower) Labeo coubie BMNH 1 907. 1 2.2 : 3744. of dilatator channel appear to be rewarding characters for the classification of those taxa in which they occur and work in preparation discusses the feature in more detail (Banister & Howes, in preparation). 7. Frontal Ctenopharyngodon (Fig. 12); broad and short, forming 50% of the cranial length (anterior tip of the ethmoid to the posterior border of the parietal), 5-6 pores in the embedded sensory canal, lateral part sloping ventrally at 27° to the transverse plane, a triangular depression occurs above the supraorbital. Hypophthalmichthys (Fig. 12); broad anteriorly, narrow posteriorly, forming 75% of the cranial length; anterior lateral part sloped ventrally at 45°; sensory canal is an osseus tube on the surface, with 9-10 pores. In cyprinids, the plesiomorph condition of the frontal is considered to be short and broad (Howes, 1978 : 31). Only in derived members of monophyletic lineages are the frontals elongate, with an increased number of sensory pores. The raised tubular sensory canal, as in Hypophthalmichthys, is a feature Tretiakov (1946) used in marking Tinea as a primitive cyprinid related to catostomids — which also possess this character. However, in catostomids 18 EPO G. J. HOWES socc EPOP EOW PTES N PALH Fig. 12 Dorsocrania: (left) Ctenopharyngodon; (right) Hypophthalmichthys. Scales = 20 mm. Specimens as previous figures. the situation is rather different for here the canal is mostly detached from the frontal and runs at the surface of thick subcutaneous tissue covering the cranium (see comments by Sagemahl, 1891 : 508 and Weisel, 1960 : 122). Although raised frontal canals are an uncommon feature in Cyprinidae they occur in abramine genera (see p. 45 for included genera), some cultrines and some chelines (eg. Oxygaster). This pattern of ossification has possibly been derived independently in these several groups. The frontal morphology in Ctenopharyngodon is generally of the plesiomorphic type but sloped and laterally depressed, extended border of the bone is a derived state and shared with Mylopharyngodon. The markedly sloped frontal in Hypophthalmichthys is shared only with Xenocypris and Abramis. 8. Parietal Ctenopharyngodon (Fig. 12); as broad as the frontal and half its length. Hypophthalmichthys (Fig. 12); half the greatest width of the frontal and 65% of its length. Commonly in cyprinids, the parietal is short, being 30%-50% of the frontal length. A parietal fontanelle is usually absent in Cyprinidae but may be found in specimens of Cyprinus and various gobioines (see Ramaswami, 1955a). This feature, widespread in other otophysans is possibly a plesiomorph character and closure of the fronto-parietal fontanelle is to be regarded as apomorphic for the Cyprinidae (see also Fink & Fink, 1981). 9. Circumorbital bones Ctenopharyngodon (Fig. 13); 1st infraorbital thick and curved outward, 2nd and 3rd CHINESE MAJOR CARPS SOR 19 IO5 Fig. 13 Circumorbital bones: (upper) Ctenopharyngodon; (lower) Hypophthalmichthys. a and b are anterior views of the respective 1st infraorbitals. Scales = 10 mm. infraorbitals elongate, 4th almost perpendicular, 5th long, crossing the sphenotic; supra- orbital extensive. Hypophthalmichthys (Fig. 13); 1st infraorbital thin, vertically aligned; 2nd, 3rd and 4th elongate, 5th, an almost right-angled bone, the vertical portion crossing the frontal, the horizontal part across the sphenotic; supraorbital extensive. In both genera the supraorbital is extensive and has the effect of extending ventrally the lateral border of the cranial roof so that the ventral edge of the supraorbital comes to lie in the same horizontal plane as that of the parasphenoid. In Ctenopharyngodon the supra- orbital slopes laterally at the same angle as the frontal, its medial portion is thickened and anteriorly bears a wide groove. The lower anterior border of the bone contacts the postero- dorsal edge of the 1st infraorbital. In Hypophthalmichthys the supraorbital is directed almost perpendicularly and its anterior part contacts half the dorsal margin of the 1st infraorbital. In small specimens (up to 100 mm SL), only the canal tube of infraorbitals 2-5 is ossified, the upper part of the 4th being fragmented into as many as four portions (Fig. 14). Comments made previously concerning infraorbitals (Howes, 1978; 1980) suggested that both extreme reduction and expansion of the elements could be seen as derived states. Of these the most often encountered condition in cyprinids is for there to be a reduction of infraorbital ossification, particularly that of the 5th bone, and also this bone's dissociation from the supraorbital. Gosline (1974) made much of the type of infraorbital-supraorbital sensory canal connection, dividing cyprinids into two groups on the basis of whether or not there was a continuous connection between the two canal systems. Howes (1978) refuted this suggestion by noting that there was always some kind of connection between the canals. What is 20 G. J. HOWES DPTC IO5 DPTC Fig. 14 POC Circumorbital bones and opercular-pterotic canal system of Hypophthalmichthys molitrix. Drawn from an alizarin specimen 65 mm SL. Scale = 3 mm. possibly of more significance is the form of the 5th infraorbital and its association with the dermosphenotic. Contrary to a previous statement (Howes, 1978) the dermosphenotic is present in many cyprinids (the element labelled as a dermosphenotic in Oreoleuciscus [Howes, 1978, fig. 22] appears to be the fragmented 5th infraorbital, and that in 'Chelaethiops' [Fig. 23] part of the dermopterotic). The distribution of the dermosphenotic amongst cyprinids and the morphology of associated canal systems is discussed in a separate paper (Banister & Howes, in preparation). The thickened 1st infraorbital of Ctenopharyngodon is an autapomorphic feature, although there is a tendency towards thickening of that bone in Mylopharyngodon. Both genera, however, share the derived, massive ridged supraorbital. Similarly, the enlarged supraorbital of Hypophthalmichthys is shared with Xenocypris. But, again the infraorbital pattern of Hypophthalmichthys is autapomorphic. 10. Supraoccipital Ctenopharyngodon (Fig. 12); short with posterior lamellate process. Hypophthalmichthys (Fig. 12); long with dorsally thickened crest. The supraoccipital of Ctenopharyngodon is the type most commonly present in Cyprinidae viz an almost square bone with a low crest. The broad, thick supraoccipital crest of Hypophthalmichthys is not encountered elsewhere in the cyprinids. Another autapo- morphic feature occurring in this genus is that the anterior portion of the supraoccipital is on a raised platform, the sides of which are contributed by the parietals. 1 1 . Exoccipital Ctenopharyngodon (Fig. 1 5); contains a fossa posterior to the subtemporal fossa; the postero- ventral border of the bone is horizontal. Hypophthalmichthys (Fig. 15); without supplementary fossa; postero-ventral border sloping laterally and caudally. An earlier statement (Howes, 1978 : 38) claimed that in cyprinids the exoccipital showed little variability. This is not the case in the two genera now under discussion. In Ctenopharyngodon the posterior border of the exoccipital lateral wall is horizontal — a condition considered plesiomorphic on account of its widespread distribution. The dorsal part of the bone, immediately posterior to the subtemporal fossa, bears a shallow depression (also contributed to by the intercalar — see below). From this supplementary fossa there arises a posterior section of the levator posterior muscle. CHINESE MAJOR CARPS 21 EPO EPO 1C 1C PTE PTE EXO Fig. 15 Posterior cranial region: (left) Ctenopharyngodon; (right) Hypophthalmichthys. Scales = 10 mm. Specimens as previous figures. In Hypophthalmichthys the posterior exoccipital wall slopes ventrally at an angle of 45° to the horizontal, its ventral border is also directed laterally at a similar angle with the result that the entire wall forms a triangular wing. The postero-lateral flaring of the exoccipital wall is so great that its distal edge contacts the distal tip of the pterotic spine — a feature otherwise present only in Xenocypris. In all other cyprinids examined the pterotic spine is separated from the exoccipital margin by a substantial part of the pterotic floor or by the intercalar. The dorsal surface posterior to the subtemporal fossa is flat, the levator posterior taking its origin entirely from the subtemporal fossa. A levator posterior with a section originating from a supplementary subtemporal fossa as in Ctenopharyngodon is known elsewhere only in Mylopharyngodon and Squaliobarbus and is considered synapomorphic. The lateral occipital fenestra in Ctenopharyngodon is narrow and when viewed posteriorly is obscured by the exoccipital medial border; a feature which otherwise occurs only in Mylopharyngodon. In contrast, there is an extensive fenestra in Hypophthalmichthys. In the Cyprinidae there is quite some variation in the size and shape of the fenestra although usually it is small and ovate. Rarely does it have the proportions of that in Hypophthal- michthys, which are approached only by Xenocypris and Abramis. To my knowledge, the fenestra is never reduced to the extent of Ctenopharyngodon and Mylopharyngodon. Both these conditions may be recognized as derived states. 12. Epioccipital Ctenopharyngodon (Fig. 1 5); bullate with a slight posterior ridge; posttemporal fossa exists as a mere depression between the epioccipital and the pterotic. Hypophthalmichthys (Fig. 1 5); triangular with a prominent posterior lip below which is a depression of the epioccipital wall; posttemporal fossa well-developed, extending someway anteriorly between the epioccipital and pterotic. The epioccipital in Ctenopharyngodon is greatly developed and contributes the major part to the formation of the subtemporal fossa. As in all cyprinids the epioccipital roof provides the major site of origin for the levator posterior muscles (see Eastman, 1971). The bone in Hypophthalmichthys is virtually pyramidal and posteriorly has a broadly pointed lip. This lip or shelf serves for the attachment of hypaxial musculature and although variously developed throughout cyprinids in none is it more so than in this taxon. 22 G. J. HOWES PTE Fig. 16 Intercalar: (upper) Squaliobarbus curriculum BMNH 1888.5.15:29; (lower) Mylo- pharyngodon piceus BMNH 1895.5.31 : 40. Scales in mm divisions. 13. Intercalar Ctenopharyngodon (Fig. 16); extensive and forming the posterior corner of the cranium between the basioccipital and pterotic. Its ventral surface contributes to the supplementary subtemporal fossa (see above) and provides the site of origin for the posterior section of the levator posterior muscle. Hypophthalmichthys (Fig. 16); smaller than the former taxon, when viewed posteriorly is seen as a triangular wedge intruding into the lateral border of the exoccipital (Fig. 1 5). Its ventral surface is broad and lies between the exoccipital and medial edge of the pterotic spine (Fig. 9). The intercalar morphology of 'Hypophthalmichthys is autapomorphic. In cyprinids the intercalar is transitory, as for example in the aspinine group (see Howes, 1978) where it is well-developed in most members, but absent in the more derived taxa. Rarely is the intercalar developed to the extent that it is in Ctenopharyngodon and occurs similarly only in Mylopharyngodon and Squaliobarbus (Fig. 16) where it also provides the site of origin for part of the levator posterior muscle. In Capoeta the intercalar also contributes to the area of origin of the levator externus (Fig. 17). However, in this genus the intercalar forms the lateral border to the exoccipital. The posterior wall of the exoccipital is strongly concave and provides the site of origin for the remainder of the posterior segment of the levator muscle. The exoccipital wall is strongly concave and its lateral margin is separated by a wide gap from the pterotic spine. In Cyprinus and Carassius the posterior wall of the exoccipital is virtually identical to that of Capoeta. CHINESE MAJOR CARPS 23 1C Fig. 17 Postero-ventral region of the cranium: (upper) Cyprinus carpio (uncatalogued); (lower) Capoeta capoeta BMNH 1879.11.14: 19. Scales = 5 mm. However, in these two genera, the intercalar is minute (or may even be absent) and is confined to the lateral edge of the pterotic; it plays no part in providing an area of attachment for the levator posterior muscle. 14. Basioccipital Ctenopharyngodon (Fig. 19a); ventral keel shallow, concave anteriorly; masticatory plate large and ovate with a shallow ventral depression; pharyngeal pad thick, lozenge shaped with a longitudinally ridged surface; pharyngeal process broad and thick, its distal end rounded and upwardly curved; basioccipital facet circular. Hypophthalmichthys (Figs 1 8a & 1 9e); ventral keel deep with markedly concave anterior border; masticatory plate cardiform, cancellous with a thick anteriorly directed point; pharyngeal pad thin, keratinized, its surface bearing a median ridge; pharyngeal process elongate, cylindrical and ventrally directed; basioccipital facet ovate. Although there is quite some variation in the morphology of the basioccipital keel and its associated masticatory plate, that of Hypophthalmichthys is outstanding in its length and shape. Only in Xenocypris and Distoechodon is there a similar degree of development of the basioccipital keel (see illustrations in Tomodo, 1979 : 96). The ventral prolongation of the basioccipital in Hypophthalmichthys is associated with the epibranchial organs and serves for the attachment of the epibranchial duct and a medial ligament that stems from the cartilaginous bloc running along the midline. Genera which most closely approach Hypophthalmichthys and Xenocypris (including here Distoechodon) are Rutilus and Chondrostoma (Figs 18d & c). In these genera and Abramis, the ventral border of the basioccipital just anterior to the aortic foramen is formed into a deep channel. The most derived condition is in Abramis where a rounded notch indents the ventral border of the bone. The basioccipital channel is confluent with a similar feature of the parasphenoid (see p. 12) and is filled with cartilage. 24 G. J. HOWES Fig. 18 Basicrania in lateral view: a. Hypophthalmichthys molitrix; b. Xenocypris argenteus; c. Abramis brama; d. Rutilus rutilus; e. Chondrostoma nasus. Abramis (c) is also shown in ventral view. The solid arrows indicate considered apomorphic features of the parasphenoid- basioccipital. Composites, all drawn to same scale. The shape and size of the basioccipital masticatory plate is variable in the Cyprinidae. Being part of the pharyngeal feeding mechanism it would be expected that a particular pharyngeal dentition would be associated with a specific plate morphology. By and large this appears to be the case; the most common type of plate is a small triangular one with a compressed triangularly shaped posterior process. This morphology is associated with omnivorous or piscivorous dental types (simple, often single-rowed, recurved teeth). The pharyngeal plate of Ctenopharyngodon greatly resembles that of Squaliobarbus and Rutilus (Figs 19a, c & d), both regarded as being plesiomorphic. However, the posterior pharyngeal process is of a unique type whereas that in Rutilus and Squaliobarbus is of the plesiomorphic form. Mylopharyngodon exhibits a derived, bowl-shaped plate — associated with molariform CHINESE MAJOR CARPS 25 I Fig. 19 Basioccipital masticatory plates: a. Ctenopharyngodon idellus BMNH 1888.5.15 : 25; b. Mylopharyngodon piceus BMNH 1895.5.31:40; c. Squaliobarbus curriculus BMNH 1888.5.15:29; d. Rutilus rutilus uncatalogued skeleton; e. Hypophthalmichthys molitrix BMNH 1895.5.31 : 22; f. Abramis brama uncatalogued skeleton. Scales = 10mm. dentition — but one obviously little modified from the plesiomorphic pattern and seemingly achieved by lateral and posterior expansion (Fig. 19b). A similarly developed plate type is found in Labeo species but here, the pharyngeal process is differently modified, forming a roof over the aortic channel (Reid, 1978, unpublished thesis). Another derived state occurs in Abramis where the pharyngeal plate has virtually disappeared (Fig. 190- Extreme reduction is also found in Semiplotus, but a characteristic surface and aortic channel morphology is apomorphic for this taxon. It is thus concluded that particular details in the morphology of the basioccipital plate and posterior process can be useful indicators of relationship. 15. Posttemporal Ctenopharyngodon (Fig. 12); broad, straight with long dorsal arm, basal part of the bone applied to the pterotic and intercalar, its upper part bridges the shallow posttemporal fossa to lie against the epioccipital. The anterior margin of the dorsal extension is bordered by a broad extrascapula. Hypophthalmichthys (Fig. 12); elongate with a broad paddle-shaped basal portion, its upper half narrow and pointed distally. The anterior border of the bone curves round the posterior margin of the extrascapula. The extrascapula is elongate with an irregular anterior margin. The Ctenopharyngodon type of posttemporal and extrascapula is widespread throughout the cyprinids and is regarded as plesiomorphic. The curved, spine-like posttemporal and narrow extrascapula of Hypophthalmichthys is otherwise present only in Xenocypris. 26 G. J. HOWES SOP Fig. 20 Opercular bones: (upper) Ctenopharyngodon idellus BMNH 1861.1.17:6; (lower) Hypophthalmichthys molitrix BMNH 1 895.5.3 1 : 22. Scale = 10 mm. 16. Opercular bones Ctenopharyngodon (Fig. 20); preoperculum with a shallow horizontal limb containing 6 pores of the sensory canal system; operculum large, almost square. Hypophthalmichthys (Fig. 20); preoperculum with deep horizontal limb containing 15 sensory pores; operculum large, deeper than long, its outer surface strongly ridged. In both genera the basal part of the antero-dorsal opercular process (to which is attached the tendon of the dilatator operculi muscle) contains a tubular canal connecting the preopercular canal with the pterotic canal. The connection between the operculum and the pterotic in Hypophthalmichthys is effected through an ossified canal lying vertically above the opercular process. This small canal-carrying bone appears to be part of the dermopterotic. There appears to be some variability in the morphology of the opercular sensory canal. Gosline (1974) noted that the preopercular canal ends 'blindly' or passes across the opercular face. However, where there is a connection between the preopercular and pterotic canals, it can be via a dermal opercular canal or 'suprapreopercula', an opercular tube or an epidermal connection. These varying conditions and their taxonomic significance are discussed in a separate paper (Banister & Howes, in preparation). The opercular series, and particularly the size of the operculum are somewhat variable throughout the Cyprinidae but certain types of opercula can be associated with particular monophyletic groups. Examples are the bariliines where the opercular bones are almost always deep and narrow; the neobolines (Chelaethiops, Neobolus) where the opercula have an attenuated posterior margin; the chelines and aspinines where the opercula possess a long dorsal border. Caution must be exercised, however, in using opercula bone morphology as a phyletic character. Large opercular and expanded preopercular bones are to be found in such widely different trophic specialists as Luciobrama, Barbus, Ctenopharyngodon and CHINESE MAJOR CARPS 27 Hypophthalmichthys. In these taxa the similarity of opercular bone shapes appear to be the result of different functional demands which are at present not understood. 17. Suspensorium Overall, there is a great similarity in the suspensorial elements of Ctenopharyngodon and Hypophthalmichthys. The palatine is broad anteriorly in both genera. However, in Ctenopharyngodon articula- tion of the palatine head is with the lateral supraethmoid notch (see p. 5; Fig. 1). The more usual condition is for the palatine head to be at a lower level, almost in the plane of the preethmoid and to be attached by a ligament to either the anterior rim, the ventral surface of the supraethmoid, or to the upper part of the mesethmoid. Only in Mylopharyngodon and Squaliobarbus is the palatine head elevated as in Ctenopharyngodon and closely articulated with a lateral supraethmoid notch. The posterior part of the palatine in Ctenopharyngodon is rounded and articulates in a deep entopterygoid facet. The posterior region of the palatine in Hypophthalmichthys is laterally compressed and articulates with a shallow, narrow entopterygoid facet. Both types of posterior palatine articulation are encountered widely in cyprinids and it is not possible at present to say which is the more derived. In both genera the hyomandibula is produced laterally and its articulatory dorsal border is steeply angled antero-ventrally. However, in Ctenopharyngodon there is a lateral shelf on the upper face of the hyomandibula and the ventral limb of the bone is short and broad, whereas in Hypophthalmichthys it is elongate. In both genera the entopterygoid and metapterygoid are depressed medially with the metapterygoid having a strongly concave dorsal border and a well-defined lateral ridge; the quadrate bears a wide ventro-lateral shelf, but in Ctenopharyngodon it is a larger bone than in Hypophthalmichthys. There are differences in the form of the symplectic which is elongate in Hypophthal- michthys but short and deep in Ctenopharyngodon. This latter type appears to represent the more derived condition (see Howes, 1980), the elongate symplectic being common in Cyprinidae. In general, there is uniformity of the suspensorial elements in the Cyprinidae; the varia- bility is one of degree, usually involving elongations, deepening or medial depression of the bones. 18. Jaws Ctenopharyngodon (Figs 2 Id, e, 0; maxilla deep with a triangular midlateral ascending process; the anterior medial process is directed almost horizontally to contact its partner via a ligament; maxillary distal portion is broad, medially curved and produced into a ventral flange which contacts the distal part of the premaxilla. Approximately mid-way along the maxilla is a foramen sited below a lateral ridge (which serves as the attachment for the palatine ligament). The premaxilla is deep with only a short anterior ascending process its ventral edge is chisel-like. The dentary has a short, dorsally rounded coronoid process; the anguloarticular, a convex dorsal border. Hypophthalmichthys (Figs 2 la, b, c); maxilla shallow with a high, square midlateral ascending process; the anterior medial process is compressed and directed ventrad toward the midline; maxillary posterior portion is laterally compressed and ventrally extended to articulate with the premaxilla. The premaxilla is exceedingly shallow and thin with a short, broad, anterior ascending process and although its ventral edge is sharp, it is not bevelled as in Ctenopharyngodon. The dentary is shallow at the symphysis but deepens posteriorly and has a high, backwardly sloped coronoid process. The anguloarticular is short and deep with a long, steep dorsal border (60° slope to the horizontal). The plesiomorphic cyprinid maxilla appears to be of the Ctenopharyngodon type, viz deep, with a large mid-lateral, or posteriorly placed ascending process with convex anterior, and concave posterior borders. Apomorphic derivatives involve the formation of a concave anterior border to the mid-lateral process as in Hypophthalmichthys, or complete reduction 28 G. J. HOWES CP MMP AMP Fig. 21 Jaw bones in lateral view: Hypophthalmichthys, a. left dentary, b. right maxilla and c. right premaxilla; Ctenopharyngodon, d. left dentary, e. right maxilla and f. right premaxilla. Scales = 10 mm. of the process as in some bariliines (see Howes, 1980). Another apomorphic feature of the maxilla appears to be the presence of an anterior foramen. This foramen provides the channel for a branch of the facial nerve which usually innervates an associated barbel. Although such a foramen is present in Ctenopharyngodon, there is no associated barbel and it will be argued later that this is due to secondary loss of the barbel (see p. 39). For the most part, however, it is true to say that those cyprinids with a maxillary foramen also have a rostral barbel. One cyprinid group which lack the maxillary foramen but have barbels are some bariliines (see Howes, 1980). However, in these taxa the barbel is not associated with the maxillary bone. The phylogenetic significance of these various conditions is discussed elsewhere (Banister & Howes, in preparation), but here it can be stated that the presence of a maxillary foramen is an apomorphic feature in one group of Cyprinidae (see also Banister & Bunni, 1980: 157). An L-shaped premaxilla without any dorsal or lateral protruberances is thought to repre- sent the plesiomorph condition. Although superficially alike, the lower jaws of Ctenopharyngodon and Hypophthal- michthys differ in the form of the coronoid process and anguloarticular. In the Cyprinidae the most common type of lower jaw morphology is one where the coronoid process occupies a central position and is formed entirely from the dentary. In these cases the anguloarticular has an almost horizontal dorsal border (eg. Barbus, Leuciscus, Aspius, Culler; see fig. 36 in Howes, 1978). Contribution of the anguloarticular to the coronoid may be considered derived and associated with orientation of the jaw and a shift in mandibular muscle insertions. A lower jaw element which may be worthy of more critical attention in cyprinids is the coronomeckelian bone. It commonly occurs as a deep, short bone with a well- developed shelf along its medial face to which attaches the tendon of part of the adductor CHINESE MAJOR CARPS 29 NS4 Fig. 22 Anterior dorsal vertebral elements shown in lateral, dorsal and anterior views of: a. Ctenopharyngodon; b. Hypophthalmichthys', c. Abramis brama. Semi-diagrammatic, compiled from several specimens; all drawn to same scale. mandibulae complex. There is some variation in the shape of this bone and certainly in one group of 'leuciscines' its spear-shaped appearance seems to offer an apomorphic character of phyletic value (Howes, in preparation). The morphology of the lower jaw is possibly the most variable anatomical unit in Cyprinidae, relative growth of its various parts probably being a major contributing factor to this variability. 1 9 . Vertebral column Ctenopharyngodon (Fig. 22a); 1st vertebra compressed and with laterally directed processes, 2nd with an elongate lateral process, 4th with anteriorly sloping pleural rib and backwardly sloping neural spine. Weberian complex: tripus broad, sickle-shaped; ossa suspensoria making extensive medial contact and forming a posteriorly curved transverse septum. Total number of vertebrae 43 (4 + 22 + 1 6 + 1 ). Hypophthalmichthys (Fig. 22b); 1st vertebra with caudally directed lateral processes, 2nd with short, triangular, caudally directed processes, 4th with anteriorly sloping pleural rib and upright neural spine. Weberian complex: tripus elongate with mesially curved tip, ossa 30 G. J. HOWES suspensoria extensive forming an antero-ventrally sloping transverse septum, neural complex bifurcated and articulating with supraoccipital. Total number of vertebrae 39-40 (4+18-19+16+1). The most striking differences in the morphology of the vertebral column in Cteno- pharyngodon and Hypophthalmichthys occur in the anterior vertebrae forming the Weberian complex. The neural complex of Ctenopharyngodon is antero-dorsally expanded and curved forward so that its anterior tip is above the 1 st vertebra (Fig. 22a); there is no contact with the supraoccipital. In Hypophthalmichthys the neural plate is deep and lamellate with a concave anterior border. The antero-dorsal prong is bifurcated and the supraoccipital process con- tained within the fork (Fig. 22b). The dorsal margin of the neural plate is deeply grooved for the midline ligament, as it is in the aspinines (see Howes, 1978 : 19). In the Cyprinidae contact between the supraoccipital crest and the neural complex is uncommon. It occurs in Labeo, Catla, Osteobrama and Semiplotus, but in these taxa there is no deeply grooved dorsal margin to the neural complex, and the anterior bifurcation is either shallow or lacking. Only in Xenocypris does the neural complex have nearly the same morphology as in Hypophthalmichthys. However, a well-developed lamellate neural plate is commonly encountered in the Cyprinidae and may be a plesiomorphic character. In Abramis, Rutilus and some species of Chondrostoma the neural complex has no antero- dorsal extension but is perpendicular or anteriorly inclined and deeply forked; a condition unique in cyprinids (Fig. 22c). In some catostomids (Carpoides) there is contact between the supraoccipital and neural complex. Here the condition is similar to that in Labeo where the neural plate is lamellate and without a dorsal groove or anterior bifurcation. It seems likely that in the latter cases expansion of the neural complex (? fused 3rd and 4th supraneurals) and its contact with the supraoccipital is associated with a deepened body and the forceful forward probing movements of the head when feeding. Reinforcement of the anterior part of the vertebral column and rigidity of the posterior cranium is provided by this arrangement. In Hypophthalmichthys which is essentially a mid-water or surface feeder, this feature must serve a different purpose and may be associated with some elevation of the cranium during feeding. The common (plesiomorph) condition of the first four (Weberian) vertebrae in Cyprinidae is as in Ctenopharyngodon (the typical condition is seen in Opsaridium microcephalus, Howes, 1980, fig. 39D). Similar development of the transverse processes, tripus and ossa suspensoria similar to those of Hypophthalmichthys is encountered elsewhere in members of the 'Leuciscinae' (ie Leuciscus, Abramis) and the aspinine group. However, elongation of these elements is also found in some barbines, chelines and neobolines (eg. Chelaethiops) and its usefulness as an indicator of phyletic relationship is doubtful. 20. Gill- arches Ctenopharyngodon has no particular modifications of the gill arches, which are similar to those of other cyprinids. The gill-rakers are short and spiny, biserially arranged and almost horizontally placed on the arch so that they intermesh with those on the neighbouring arch to form a filter. Hypophthalmichthys displays marked modification of the gill arches and possesses a complex epibranchial organ (Figs 24-26). The gill structures of Hypophthalmichthys have been described in varying detail by a number of authors. Boulenger ( 1 90 1 ) first described the epibranchial organ; Fang ( 1 928) gave a detailed description of the gill arches and epibranchial organ, and speculated on their function. Zambriborshch (1955; 1957) described and figured the gross morphology of the branchial apparatus and histology of the gill-rakers, including comments on function. Verigin (1957) described and figured the epibranchial organ, gill-rakers, upper pharyngeal pad and roof of the pharynx. He also studied the ontogeny of the branchial apparatus. Wilamovski (1972) made a description of the epibranchial organ and gill-rakers, comparing them with the branchial system of Cyprinus. Taken in toto these descriptions give an CHINESE MAJOR CARPS EB4 31 Fig. 23 Posterior gill-arch elements: (left) Hypophthalmichthys molitrix, from dissected specimen BMNH 1895.5.31:22; (right) Xenocypris argenteus from alizarin specimen BMNH 1 889.6.8 : 19. Left scale = 5 mm, right = 2 mm. adequate account of the organ's morphology, and it is necessary here only to emphasize certain of its apomorph characters. The ceratobranchials are elongate, bearing numerous biserially arranged slender gill- rakers. The two rows of rakers are set at a diverging angle to one another thus producing an 'interbranchial space' (Fang, 1928). There are approximately 13-14 rakers per 1 mm of ceratobranchial. Fang (1928) recognized two types of gill-rakers, broad and narrow, the broad type occurring between each 4th and 5th narrow type. The gill-rakers are connected by a reticulate mucuous membrane (sifting membrane of Fang, 1928) which forms a network or filter, whose pores are about 20 // in diameter. The gill-rakers continue onto the epibranchials. The 1st and 4th epibranchials are expanded dorsally into axe-shaped elements. The 1st forms the anterodorsal wall, and the 4th the posteromedial wall of the epibranchial organ (the branchial arches are so orientated that the 4th epibranchial comes to lie in a medial (longitudinal) plane rather than a posterior (transverse) one). Although not entirely ossified, a 5th epibranchial is present as a well-developed, broad, lamellate element, and has a complex articulation with the 4th ceratobranchial (Fig. 23). There are two large well-ossified infrapharyngobranchials recognized as IF2 and IF3; IF4 is represented by an extensive cartilaginous bloc bordering the posterior rim of IF3 and contacting epibranchials 3 and 4 (Fig. 24). In front of IF2 there is another large cartilaginous cap onto which articulates epibranchial 1. This no doubt is IF1. In adult specimens the 5th epibranchial is joined to the capsule of the epibranchial organ via an interposed cartilage which may represent the 5th infrapharyngobranchial. The epibranchial organ is formed by four curled cartilaginous channels or ducts, which are prolongations of the four epibranchials (Fig. 26). Each duct is lined with two rows of biserially arranged gill-rakers, the posterior row of one arch intermeshing with the anterior row of the adjacent arch and in this way running into each duct (Fig. 24). The outer row of the 1st arch enters a small chamber of the first duct and becomes reduced in size and number. The first duct houses the gill-raker rows of the 1st and 2nd epibranchials, the 2nd those of the 32 G. J. HOWES EBD EB1 Fig. 24 Hypophthalmichthys epibranchial organ: a. anterior view of left organ, ossified areas of the infrapharyngobranchials are stippled; b. dorsal view of left organ to show orientation, dashed line represents the ventral midline of the head, and solid arrow indicates the direction of view depicted in a. Scales = 10 mm. 2nd and 3rd epibranchials, the 3rd those of the 3rd and 4th, and the 4th duct those of the 4th and 5th epibranchials. Zambriborshch (1957) referred to the organ ducts as supra- and infrapharyngobranchials. However, the infrapharyngobranchials are here identified as elements distinct from the cartilaginous ducts. The entire epibranchial structure is invested in palatal connective tissue. The tissue below the basal part of the organ is produced into four folds (palatal folds of Fang, 1928) which hang down into the interbranchial channels (Figs 25 & 26). The lateral face of the epibranchial organ, also formed from palatal tissue, is connected to the medial face of the preoperculum via a segment of the adductor hyomandibulae muscle (= pharyngopraeo- percularis muscle of Zambriborshch, 1955); see Fig. 25. The morphology of the epibranchial arches in Hypophthalmichthys is unique. Modified gill-arches and gill-rakers are, however, found elsewhere in cyprinids. In Xenocypris argenteus the epibranchials are rounded over a thickened palatal tissue pad, the gill-rakers are numerous, approximately 25 per 1 mm of gill-arch, and extend to the proximal part of the epibranchial. The infrapharyngobranchials extend into the dorsal surface of the palatal tissue pad. The pads or cushions have a hollow interior, the internal surface of the tissue is highly papillose and convoluted. Similar development of the epibranchials, gill-rakers and palatal cushions are to be found in Plagiognathops, although the gill-rakers on the 1st arch are 'normal' and the palatal cushion is more papillose. An hypothesized plesiomorph condition of the epibranchial organ is encountered in Abramis. Here the epibranchials are markedly rounded and the palatal tissue greatly thickened. This is particularly evident in Abramis sapa where the palatal cushion occupies a large part of the pharyngeal cavity (Fig. 27b). Anteriorly it is rounded and forms a cup- shaped depression in the palate. The palatal cushion resembles that of Xenocypris in having a hollow interior, although not as cavernous. The outer surface of the cushion shows slight ridging between the gill-arches, reminiscent of the palatal folds in Hypophthalmichthys. The cartilaginous proximal tips of the epibranchials are highly developed and lie embedded in a cartilaginous casing which extends along the basioccipital (Figs 27a, b). Again, as in CHINESE MAJOR CARPS HYO 33 EB1 Ah CB1- EOD PF Fig. 25 Hypophthalmichthys, postero-lateral view of epibranchial organ and gill-arch elements. The ceratobranchials have been partly cut away to show the organ's folded tissue surface. Schematic drawing based on two dissected specimens. Xenocypris the cartilaginous tips of the epibranchials extend ventrally into the floor of the palatal cushion roof. Elsewhere in the Cyprinoidei an epibranchial organ is present in Catostomidae. In these taxa as in Hypophthalmichthys and the xenocyprinids, the epibranchials are rounded over a complex pharyngeal pad and are medially prolonged (Fig. 28a). In catostomids the pharyngeal pad more closely resembles that in Xenocypris being a medially undivided structure surrounding the basioccipital process, and not in the form of paired diverticula as in Hypophthalmichthys. The catostomid basioccipital is expanded laterally and is fretted, playing an integral part in the structure of the pharyngeal cushion. This is unlike the cyprinid condition where the basioccipital acts as a support for connective tissue dividing the organ (see Weisel, 1960 for an account of the catostomid pharyngeal organ, and Eastman, 1977 for a description of catostomid pharyngeal anatomy). When compared to the epibranchial organs of other teleosts which possess them, that of Hypophthalmichthys most closely resembles the continuous tube organs of some clupeids (see Miller, 1969). It is possible that a form of epibranchial organ occurs in the south-east Asian cyprinid genus Thynnichthys. This fish shows a remarkable superficial resemblance to Hypophthal- michthys, particularly in its low positioned eye. As in the latter, the gill-arches are elongate but they lack gill-rakers. There does, however, appear to be a swollen area of pharyngeal 34 G. J. HOWES PHF EOD EBD PF Fig. 26 Hypophthalmichthys, ventral dissection of the epibranchial organ, ceratobranchials cut away. Scale=10mm. Drawn from specimens BMNH 1895.5.31 : 36-37 and uncatalogued Swinhoe specimen. tissue with a single posterior median duct leading into the oesophagus. Only a single specimen is available for superficial dissection but those osteological features which are observable are unlike those of Hypophthalmichthys and more closely resemble Ambly- pharyngodon. Although Berry & Low (1970) described certain morphological features in hybrids between Hypophthalmichthys and Ctenopharyngodon, no mention was made of the develop- ment of an epibranchial organ. A specialization of the upper branchial arch shared by Hypophthalmichthys and Xenocypris is hypertrophy of the 5th epibranchial. In Xenocypris this element is broadly lamellate and has a form of articulation with the 4th ceratobranchial similar to that of Hypophthalmichthys (Fig. 28a). In cyprinids the 5th epibranchial is greatly reduced or appears as a small cartilaginous element above the 4th epibranchial (Banister, pers. comm.). The hypertrophy of this element and its articulation with the 4th epibranchial are seen as synapomorphies uniting Hypophthalmichthys and Xenocypris. A superficially similar gill-arch morphology to that in Hypophthalmichthys occurs in Gibelion and Catlacarpio (Fig. 28d). In these taxa also the ceratobranchials are elongate and bear numerous gill-rakers. The epibranchials are sickle-shaped, their medial 'handle' portion being dorso-ventrally expanded and antero-posteriorly compressed. The two infra- pharyngobranchials are thickened, IF2 being virtually reorientated from the usual horizontal plane to a vertical one. The entire upper branchial basket is surrounded by a thick, papillose pharyngeal cushion and this, together with the fretted basioccipital pharyngeal process greatly resembles the upper pharyngeal anatomy of catostomids. It may be that the catline taxa employ similar feeding methods to those of catostomids (see Weisel, 1960). Again, similarly developed epibranchials are present in "Labeo"1 stoliczkae (Fig. 28c), but they are CHINESE MAJOR CARPS 35 EB1 Php BOCB BUG Fig. 27 Upper pharyngeal anatomy ofAbramis. Dorsal branchial elements of: a. Abramis brama, left side BMNH 1 974.9. 5 : l-3l; b. Abramis sapa, right side BMNH 1967.7.17 : 3^1, drawn in situ, scale = 1 mm; c. lateral view of cartilaginous basicranial bloc and pharyngeal pad in A. brama, scale = 10 mm; d. pharyngeal cavity of A. sapa showing 4th ceratobranchial from which most of the gill filaments have been removed, scale = 5 mm. not so acutely curved medially, and neither are the infrapharyngobranchials orientated in the same way. Furthermore, the basioccipital process is thick and broad. Other neurocranial characters suggest that 'Labeo ' stoliczkae has as its closest relative, Cirrhina, whilst Gibelion and Catlacarpio seem closely related to Thynnichthys, Amblypharyngodon and Osteobrama (part). Pharyngeal bones and teeth of both genera have been described and figured by Chu (1935). The pharyngeal bone of Ctenopharyngodon has a greatly attenuated lower limb and there is a prominent process (for the attachment of the posterior portion of the levator posterior) at the curvature of the upper limb. The teeth are in two rows (2, 4-5), greatly compressed and with their grinding surfaces serrate, the serrations being the lateral margins of grooves on the tooth's outer face. 36 G. J. HOWES Fig. 28 Posterior views of the right 4th epibranchial of: a. Xenocypris argenteus BMNH 1889.6.8 : 19-22 (alizarin); b. Abramis ballerus (uncatalogued skeleton); c. Labeo stoliczkae BMNH 1 89 1 . 1 1 .30 : 286 (skeleton); and d. Gibelion catla BMNH 1 908. 1 2.28 : 1 1 22 (skeleton). Scales for a, b, and c = 2 mm, d = 5 mm. According to Chu (1935), the teeth and pharyngeal bone of the Hypophthalmichthinae '. . . are singularly unique, without parallel in other cyprinids'. The anterior limb of the pharyngeal bone is dorsoventrally depressed and the upright posterior limb is antero- posteriorly compressed. There is a large fenestra within the expanded dentigerous part of the bone. In situ this fenestra is covered laterally by a translucent tissue, medially it serves as a cavity into which the expanded edge of the basioccipital masticatory plate inserts (Fig. 26). The 4 teeth of Hypophthalmichthys are in a single row, are compressed, have a concave occlusul surface with a transverse median groove and lateral striation. Chu (1935) is correct in describing the pharyngeal dentition of Hypophthalmichthys as unique. There are, however, similarities between the pharyngeals of this species and those of Xenocypris and Plagiognathops, where there is also an extensive foramen in the pharyngeal bone, and the teeth have similar grinding surfaces (see Chu, 1 935). 2 1 . Adductor hyomandibulae muscle In Ctenopharyngodon the adductor hyomandibulae is developed as in the majority of cyprinids, namely, a thick dorsal element arising from the face of the prootic and inserting, musculously, on the upper medial face of the hyomandibula. In Hypophthalmichthys the muscle is expanded and is continuous (see below) with the adductor arcus palatini, forming a muscular 'wall' which effectively separates the branchial and buccal cavities. The upper part of the muscle (Dahm, Fig. 29) arises from the prootic and inserts on the upper medial face of the hyomandibula as in other cyprinids. However, a posterior segment attaches to the antero-dorsal wall of the epibranchial organ and another part of the (? same segment) of muscle runs from the lateral wall of the epibranchial organ to the preoperculum (pharyngopraeopercularis of Zambriborshch, 1955). Anterior to the dorsal part of the muscle another element can be distinguished originating from the lower part of the prootic and the pterosphenoid spur to insert on the medial section of the hyomandibula. Anteriorly this second portion is continuous with the adductor arcus palatini CHINESE MAJOR CARPS Dahm 37 PRO HVII Vahm HYO PSK MET Fig. 29 Hypophthalmichthvs, ventro-posterior view of adductor hyomandibularis musculature, based on specimens BMNH 1895.5.31 : 36-37. Scale = 10 mm. muscle which originates from the parasphenoid ventral process and inserts along the lower limb of the hyomandibula and, anteriorly, along the dorsal rim of the metapterygoid. Winterbottom (1973) points out that the adductor hyomandibulae and, at least in certain taxa, the adductor arcus palatini are parts of a single muscle. Certainly, in this case it is only the orientation of the fibres that distinguishes the upper medial section of the muscle as part of the adductor hyomandibulae rather than as part of the adductor arcus palatini. Such a highly developed adductor hyomandibularis musculature occurs elsewhere only in Abramis, Rutilus and Chondrostoma. In Abramis the upper medial portion of the muscle is strongly attached to the pharyngeal tissue, and thence to the posterior border of the hyomandibula. Although in Xenocypris the adductor hyomandibulae is not as expansive as in the other genera, the upper medial section is attached, along its entire ventral surface, to the tissue covering the branchial arches, and some fibres appear to extend on to the preoperculum. Morphologically, this can be considered a forerunner of the derived condition realized in Hypophthalmichthys where part of the adductor hyomandibulae becomes attached to the epibranchial organ and preoperculum and possibly acts as the power unit forcing water through the epibranchial ducts. The adductor hyomandibulae morphology in Hypophthalmichthys, Xenocypris, Abramis, Rutilus and Chondrostoma is considered synapomorphic. 38 G. J. HOWES 22. Pectoral girdle The pectoral girdle of Ctenopharyngodon is remarkably similar to that of Barbus and Cyprinus, except that the posterior cleithral lamina is extended posteriorly. The horizontal cleithral limb is broad, with a concave anterior border. The cleithra contact each other only at their anterior tips. The coracoid is shallow and rounded posteriorly; the mesocoracoid is narrow. The pectoral girdle of Hypophthalmichthys displays many modifications relative to the presumed plesiomorph pattern. The horizontal cleithral lamina is greatly expanded and rises in the vertical plane to join its partner for the anterior third of its length. The posterior border of the coracoid is square. The mesocoracoid limb is broad. In both genera the postcleithra are elongate and curved mesially, ie. plesiomorphic state; see Howes (1980). In no other cyprinid is the horizontal cleithral lamina expanded to the same extent as in Hypophthalmichthys. Only in Abramis and Xenocypris is there an approachable condition. In these taxa, however, the medial cleithral border is deeply notched at a point just below the dorsal articulation of the mesocoracoid. The function of this notch is to contain the expanded border of the pharyngeal bone. The mesocoracoid limb is broad in all these taxa. Phylogenetic relationships From the foregoing descriptions and analyses of Ctenopharyngodon and Hypophthal- michthys, a series of apomorph characters can be identified, none of which is shared by both taxa. Instead, these characters serve to relate the respective genera to other taxa and thus Ctenopharyngodon and Hypophthalmichthys must be considered members of two distinct monophyletic assemblages. Relationships of Ctenopharyngodon The following characters are identified as apomorphies. 1 . Subtemporal fossa extending anteriorly to invade the autosphenotic on which part of the pharyngeal musculature originates (p. 1 5) 2. Lateral notch in the supraethmoid with which the palatine head articulates (p. 5) 3. Extensive prootic with ventrally directed trigemino-facialis foramen (p. 1 3) 4. Extensive intercalar forming posterior corner of cranium and contributing to an accessory subtemporal fossa (p. 22) 5. Divided levator posterior muscle, the posterior segment originating from the intercalar (P. 22) 6. Extensive and bullate epioccipital (p. 2 1 ) 7. Reduced lateral occipital fenestra (p. 2 1 ) 8. Olfactory foramen countersunk entirely within the lateral ethmoid (p. 9) 9. Frontal broad and sloping ventrally (p. 18) 1 0. Massive and shelved supraorbital (p. 20) 1 1 . Orbital portion of parasphenoid broadened and with ventral ridge (p. 11) 12. Ventrally directed pterosphenoid process (p. 10) 1 3 . Thickened 1 st infraorbital (p. 20) 14. Distally broadened and upwardly curved basioccipital pharyngeal process (p. 23) Characters 1-6 are synapomorphies with Mylopharyngodon and Squaliobarbus; 7-1 1 with Mylopharyngodon; 12-14 are autapomorphic. From trie distribution of these apomorph characters, Mylopharyngodon is considered to be the sister taxon of Ctenopharyngodon with both forming the derived sister group of Squaliobarbus. These three taxa constitute a monophyletic lineage termed the Squalio- barbine group. Mylopharyngodon (Fig. 30b) is a monotypic genus (type species M. piceus (Richardson, 1845)), characterized by the autapomorphic features of an extensive bowl-like masticatory CHINESE MAJOR CARPS 39 Fig. 30 Outline drawings of members of the squaliobarbine group: a. Ctenopharyngodon idellus; b. Mylopharyngodon piceus; and c. Squaliobarbus curriculus. plate (Fig. 19b); large molariform pharyngeal teeth; small basioccipital facet dorsally over- lapped by an exoccipital ridge (Fig. 31); and narrow, mesially curved posterior cleithral lamina. Squaliobarbus (Fig. 30c) is the plesiomorph member of the lineage differing from its sister taxa in possessing two pairs of small barbels. The anterior pair of barbels are minute and appear superficially as mere appendages of the ventral infraorbital skin fold. However, the barbel is innervated by a sub-branch of the VII facial nerve. The main nerve branch passes through a foramen in the maxilla just below a shelf which forms the attachment point of the palatine ligament. The barbel originates some distance posterior to the maxillary foramen and the main nerve branch supplies the lip fold. It was stated on p. 28 that a maxillary foramen carrying a nerve supply to an anterior barbel is an apomorph character uniting those taxa which possess it. It was also noted (p. 28) that Ctenopharyngodon has a maxillary foramen but both nerve and barbel are lacking. In Mylopharyngodon there is no trace of a maxillary foramen or barbels. Two hypotheses are proposed to account for the distribution of barbels and associated foramina amongst squaliobarbines: 1 . That Ctenopharyngodon and Mylopharyngodon represent a plesiomorph condition in which the foramen and/or barbel has not formed. 2. That the absence of barbels and associated nerve in these genera represents secondary losses. 40 G. J. HOWES BOF Fig. 31 Mylopharyngodon piceus; posterior view of basioccipital from specimen BMNH 1895.5.31 : 40. Scale = 5 mm. It is assumed that the second hypothesis is more likely to be the correct one as the plesio- morph lineage (Squaliobarbus) possesses barbels, and that in this taxon their reduction and that of the associated nerve signifies transition to loss state. Also, reduction of barbels is recorded in other cyprinid taxa (see Banister & Bunni, 1980). The distribution of cyprinid barbels; anterior and posterior, whether they are associated with a foramen, the type of nerve supply and form of proximal attachment is a subject dealt with elsewhere (Banister & Howes, in preparation). For the purposes of this discussion it is considered that the squaliobarbine assemblage form part of that series of taxa which possess a maxillary foramen carrying a nerve supply to an anterior barbel. The suite of apomorph characters defining the squaliobarbines involves, for the most part, the occipital region of the cranium — the expansion of the subtemporal fossa and the inter- calar. These osteological features are concomitant with an increased evolutionary develop- ment of the levator posterior muscle which extended forward into the autosphenotic and was divided posteriorly, the posterior section originating from the intercalar. These specializ- ations could be associated with the essentially phtyophagous habits of the squaliobarbine taxa. The proposed theory of relationships is to regard the squaliobarbine group as a divergently specialized sister lineage to all other barbelled Cyprinidae (i.e. Cyprininae); see Fig. 32. Synopsis of taxa assigned to the Squaliobarbine group Ctenopharyngodon Steindachner, 1866 Pristiodon Dybowskii, 1877 Type and only species: C. idella (Valenciennes, 1844). Distribution: Pacific drainage of the USSR and China, lat. 50°-23°N. Introduced into many tropical and temperate areas including Europe, North America, India, New Zealand, Malaysia. Principal bibliographic sources concerning biology: Lin, 1935; Nikolsky, 1954; Inaba & Nomura, 1956; Hickling, 1967;Okada, 1960; Cross, 1969; Edwards, 1974. CHINESE MAJOR CARPS 41 My Other Cyprininae Fig. 32 Cladogram representing the phyletic relationships of the squaliobarbine group. Ci = Ctenopharyngodon, My = Mylopharyngodon, Sq = Squaliobarbus. Node 1. Possession of maxillary foramen. Node 2. Synapomorphies 1-7 (page 38). Node 3. Synapomorphies 8-1 1 (page 38), also loss of maxillary foramen in Mylopharyngodon. Mylopharyngodon Peters, 1873 Myloleucus Gunther, 1873 Myloleuciscus Garman, 1912 Leucisculus Oshima, 1920 Type and only species: M. piceus (Richardson, 1 845). Distribution: Amur, China, Formosa. Principal bibliographic sources: Nikolsky, 1954 & 1956. Squaliobarbus Gunther, 1868 Type species: S. curriculus (Richardson, 1845). Sauvage (1884) described another species, S. caudalis, from Tonkin. Pellegrin (1934), when commenting on the Hanoi collection of Sauvage, made no mention of this species but included in his list, S. curriculus. Distribution: Amur, China, Western Korea, Hanoi. Principal bibliographic source: Nikolsky, 1956. Relationships of Hypophthalmichthys The following characters are identified as apomorphies: 1 . Medial ethmoid notch floored by the vomer (p. 9) 2. Frontal with an acute lateral slope, providing an extensive cranial dilatator fossa (p. 1 5) 3 . Extensive lateral occipital fenestra (p. 2 1 ) 4. Modification of the adductor hyomandibulae muscle involving partial attachment to the preoperculum and upper pharyngeal tissue (p. 36) 5. Expanded horizontal cleithral lamina (p. 38) 6. Markedly rounded epibranchials (p. 33) 7. Deep basioccipital keel with marked concave anterior border (p. 23) 8. Carotid foramen contained in ventrally produced parasphenoid process (p. 12) 9. Lateral ethmoid-frontal fenestra (p. 9) 1 0. Expanded 1 st and 4th epibranchials, the 5th present and hypertrophied (p. 34) CHINESE MAJOR CARPS 43 1 1 . Extensive fossa in pharyngeal bone (p. 36) 12. Exoccipital wing extended distally to contact pterotic spine (p. 21) 13. Curved posttemporal (p. 25) 14. Pterosphenoid contributing to anterior trigemino-facialis foramen (p. 10) 1 5. Cardiform masticatory plate with tubular posterior extension (p. 23) 16. Intercalar invading border of exoccipital (p. 22) 1 7. Complexly developed epibranchial organ (p. 3 1 ) 18. Neural complex bifurcated and articulating with supraoccipital (p. 30) Characters 1-6 are synapomorphic with Abramis and Xenocypris (1 Plagiognathops and Distoechodon); 7-13 with Xenocypris; 14-18 are autapomorphic. From the distribution of these apomorph characters Xenocypris is regarded as the twin taxon to Hypophthalmichthys. Together they form the sister group to an assemblage including Abramis. Xenocypris (Fig. 33b). The taxonomy of this genus is confused, two other monotypic taxa, Plagiognathops and Distoechodon being regarded by some authors as subgenera (see Banarescu, 1970). All these taxa are presently placed in the subfamily Xenocyprinidinae. Two previous comments regarding Xenocypris are in need of modification. The first (Howes, 1979), was to the effect that the Xenocyprinidinae was polyphyletic, viz that the included genera, Xenocypris, Plagiognathops and Distoechodon shared no derived characters. In fact they do (see below). The second statement (Howes, 1980) was that the closest relatives of Xenocypris are amongst the hemicultrines (Hemiculter, Hemiculterella and Ochetobius). This assumption was based on what are now deemed to be symplesio- morphies, and indeed, the hemicultrines as earlier conceived represent a paraphyletic assemblage. Plagiognathops and Distoechodon share with Xenocypris (as represented by the type species X. argenteus) apomorph characters involving the gill-arches and basioccipital, although they are not so highly derived as in Xenocypris. Lack of osteological material prevents a complex diagnosis of the character complexes and the three genera are considered together as an unresolved trichotomy. Abramis (Fig. 33c), here including the 'genera' Vimba and Blicca\ see p. 46, shares with Xenocypris and Hypophthalmichthys synapomorphic characters 1-6, involving the ethmoid region, cranial roof, pharyngeal roof covering and associated musculature. However, a simple sister-group relationship between Abramis and Xenocypris + Hypophthalmichthys is doubtful in the light of other apomorphies possessed by Abramis and exclusively shared with Chondrostoma and Rutilus (Figs 33d & e). These are: (i) confluent ventral channelling of the parasphenoid and basioccipital; (ii) columnar, dorsally bifurcated neural complex; and (iii) gill-rakers with globular, papillate bases. The underlying synapomorphy (Saether, 1979) uniting Rutilus, Chondrostoma, Abramis, the Xenocypris complex and Hypophthalmichthys is a modification of the basicranium, viz a parasphenoid notch or confluent ventral channelling of the parasphenoid and basioccipital, and deepening of the basioccipital process. From the distribution of synapomorphies within this assemblage — termed the Abramine group — two hypotheses of generic relationships are available and represented as cladograms A & B (Fig. 34). In cladogram A, Abramis is reckoned to be the sister group of Rutilus and Chondrostoma (the similarity between the parasphenoid-basioccipital morphology links these two genera as sister taxa) on the basis of characters (i) and (ii) cited above. In this case, the characters Abramis shares with Hypophthalmichthys and the Xenocypris complex (listed as synapo- morphies 1-6 on p. 41) must be considered as convergent. In cladogram B, Abramis is considered as the sister group to Hypophthalmichthys + the Fig. 33 Outline drawings of members of the abramine group: a. Hypophthalmichthys molitrix; b. Xenocypris argenteus; c. Abramis brama\ d. Chondrostoma nasus; e. Rutilus rutilus. 44 Ru Ch G. J. HOWES Ab Hy X-P-D X-P-D B Fig. 34 Cladograms expressing two schemes of phyletic relationships of the abramine group. Ru = Rutilus, Ch = Chondrostoma, Ab = Abramis, Hy = Hypophthalmichth ys, X-P- D = Xenocypris-Plagiognathops-Distoechodon trichotomy. Cladogram A. Node 1. Modification of basioccipital-parasphenoid; development of pharyngeal pad; transversely convex vomer; hypertrophy of adductor hyomandibulae muscles. Node 2. Basioccipital-parasphenoid channel; columnar neural complex; globular based gill-rakers. (Con- vergence of 6 characters in Abramis with Hypophthalmichthys and Xenocypris.) Node 3. Synapomorphic basioccipital morphology (see Fig. 18). Node 4. Synapomorphic gill-arch struc- ture; development of epibranchial organ and associated basicranial deepening. (Loss of convex vomer and deep ethmoid in Hypophthalmichthys.) Cladogram B. Node 1. Combination of nodes 1 and 2 in cladogram A. Node 2.= node 3 in cladogram A. Node 3. Synapomorphies 1-6 (p. 41). Node 4. Synapomorphies 7-13 (p. 41); also loss of 5 characters through nodes 1-3 (see text, p. 43). Xenocypris complex. In this case, five characters must be recognized as having been secondarily lost in the two latter taxa, viz: parasphenoid-basioccipital channel; columnar neural complex; globular based gill-rakers (lost in both Hypophthalmichthys and Xenocypris); convex and deeply notched ethmoid (lost in Hypophthalmichthys). It may be argued that in their vertebral elements Hypophthalmichthys and Xenocypris are conservative and have retained the plesiomorphic lamellate neural complex. Likewise, Hypophthalmichthys may have retained the plesiomorphic ethmo-vomerine bloc. (No evidence is available to show that apomorphic modifications of the ethmoid region are linked with those of the basicranium.) If this is so, then the more likely hypothesis is that CHINESE MAJOR CARPS 45 Rutilus, Chondrostoma and Abramis form the sister group to Hypophthalmichthys and the Xenocypris complex, as one has only to postulate the loss of the apomorphic ethmo- vomerine assembly in Hypophthalmichthys (cladogram A). This is also the more likely hypothesis of relationships if the autapomorphic parasphenoid-basioccipital and neural complex of Abramis are seen as terminal points of a possible evolutionary sequence represented by the more plesiomorphic types in Rutilus and Chondrostoma. Although it may appear that Abramis displays an incipient form of the epibranchial struc- ture in Xenocypris, this is misleading. In fact, the gill-arch morphology and pharyngeal dentition suggests that the pharyngeal pad of Abramis differs functionally from that of Xenocypris and its relatives. There is indeed, no 'morphoclinal' or 'transitional' series represented in extant abramine genera which might reflect the evolutionary development of an epibranchial organ. However, the basic architectural modification of the basicranium (deepening of the parasphenoid and basioccipital) can be seen as a preadaptation to the formation of an epibranchial structure which in the abramines appears to have followed two evolutionary pathways hypothesized to be most closely represented by the relationships expressed in cladogram A. Falsification of either hypothesis will be forthcoming when more complete anatomical data are available for Plagiognathops and Distoechodon and a functional analysis of pharyngeal/epibranchial mechanisms available for all abramine taxa. Synopsis of taxa assigned to the Abramine group Hypophthalmichthys Bleeker, 1860 Cephalus Basilewsky, 1855 Abramocephalus Steindachner, 1869 Onychodon Dybowski, 1872 Aristichthys Oshima, 1915 Type species: H. molitrix (Valenciennes, 1 844). Two other species, H. nobilis (Richardson, 1844) and H. harmandi Sauvage, 1884 are included. Reasons have already been given for synonymizing Aristichthys (p. 2) Distribution: Amur, China. Introduced into Taiwan, Thailand, Japan, Sri-Lanka, Europe and North America. Principal bibliographic sources concerning biology: Tsui, 1936; Nikolsky, 1956. Xenocypris Giinther, 1868 Type species: X. argentea Giinther, 1868. Another 9 or 10 species are recognized by various authors; see Banarescu (1970) for most recent revision, and Nikolsky ( 1956) for biological data. Plagiognathops Berg, 1907 Type and only species: P. microlepis (Bleeker, 1871). Distribution: Yangtze, middle Amur, Ussuri, Lake Khanka (see Berg, 1949). Distoechodon Peters, 1880 Type and only species: D. tumirostris Peters, 1880. Distribution: Southern China. Rutilus Rafmesque, 1 820 Scardinius Bonaparte, 1837 Hegerius Bonaparte, 1845 Acanthobrama Heckel, 1843 Leucos Heckel, 1843 Pigus Bonaparte, 1845 Cenisophius Bonaparte, 1845 Gar danus Bonaparte, 1845 46 G. J. HOWES Pararutilus ; Bonaparte, 1845 Pachychilon Steindachner, 1 866 Rubellus Fitzinger, 1873 Mirogrex Goren, Fishelson & Trewavas, 1973 Type species: R. rutilus Rafmesque, 1820. Numerous nominal species with a wide distribution throughout Europe, USSR, Israel to the Tigris and Euphrates. The characters used to define Leuciscine genera are criticized below (p. 48). Typical are the use of modal numbers of fin rays, lateral line scales, gill-rakers and plesiomorphically low vertebral counts. No comparisons of the osteology of the various 'genera' given in the above synonymy have been made previously. I have compared the osteology of Rutilus rutilus with a selection of species currently assigned to this genus, to Scardinius, to Acanthobrama and to Mirogrex and am unable to find any noticeable differences. Indeed, all these taxa possess those apomorph characters which identify Rutilus, viz form of the median ethmoid notch; extension of the vomerine arms; preethmoid not, or barely contacting the mesethmoid; columnar neural complex; globular, papillate based gill-rakers. Scardinus is usually separated from Rutilus on the basis of its having the pharyngeal teeth in two rows instead of a single row, and the presence of a ventral body keel. I would consider that the differences in pharyngeal teeth are of a specific nature, as in Abramis (see Shutov, 1967 and Eastman & Underbill, 1973); likewise, the ventral keel is variously developed in Rutilus rutilus. Wheeler (1976) concluded that of the features used to characterize Rutilus and Scardinius the morphology of the pharyngeal teeth was probably the most significant. If Chu's (1935) hypothesis is accepted, namely that three rows of pharyngeal teeth represent the primitive (plesiomorph) condition, then the two rows of Rutilus and the single row of Scardinius are seen as successively derived states. In the majority of Rutilus x 'Scardinius' hybrids a second tooth row is apparent which may indicate a dominant genetical pattern for the plesiomorph condition. Concerning the synonymy of Acanthobrama and Mirogrex; Goren, Fishelson & Trewavas (1973) separate these two genera on modal differences in the number of lateral line scales, gill-raker counts and the number of vertebrae. These authors compare the two 'genera' with Capoetobrama (see below), but not with Rutilus, and they state that the generic position of the presumed ancestral taxon to Acanthobrama, 'Rutilus' tricolor '. . . has not been determined'. I would agree that the species tricolor does not belong to Rutilus and should properly be placed amongst a complex including Leucaspius. However, Acanthobrama and Mirogrex share all the apomorph characters of Rutilus sensu stricto, they are thus to be regarded as belonging to that genus. The genus Capoetobrama (see above) does not belong to this assemblage at all and may be more closely allied to the alburnine group (unpublished obs.). Abramis Cuvier, 1817 Ballerus Heckel, 1843 5//cca Heckel, 1843 Zopa Fitzinger, 1873 Vimba Fitzinger, 1879 Leucabramis Smitt, 1895 Sapa Kazanskii, 1928 Type species: A. brama (Linnaeus, 1 758). Five or six species with wide distribution in Europe, the Baltic, Black, Aral and Caspian seas; Asia Minor, Caucasus. Principal bibliographic source: Backiel & Zawisza, 1968. Shutov (1969) considered Blicca synonymous with Abramis. His analysis of populations of Blicca bjoerkna revealed a morphometric range encompassing that of Abramis brama and a decrease in the numbers of teeth in, and ultimate loss of the second tooth row across the range of the samples. Similarly, I find no differences in the osteology of Abramis and Vimba. CHINESE MAJOR CARPS 47 Both genera possess an identical form of parasphenoid-basioccipital border with ventral channel and notch and highly developed pharyngeal tissue pads. They also share the apomorphic feature of a scaleless pre-dorsal fin ridge. Characters formerly used to separate these 'genera', such as the position of the anal fin origin and the presence of a dorsal ridge on the caudal peduncle are continuously variable and meaningless as taxon defining criteria. All Abramis species display considerable morphometric variability (see Berg, 1949 and Banarescu, 1963). Chondrostoma Agassiz, 1835 Type species C. nasus Agassiz, 1835. Berg (1949) records 18 species, having a wide distribution throughout Europe south to the Tigris and Euphrates. Chondrostoma is characterized by those autapomorphic characters associated with forshortening and deepening of the upper jaw bones and a forwardly inclined coronoid process on the dentary. The posterior extension of the pharyngeal plate is thin and broad with a grooved dorsal surface. Assertions made concerning Chondrostoma in the anatomical descriptions presented here refer only to the type species, C. nasus. Discussion Earlier comments (Howes, 1976) that few, if any, presently recognized cyprinid subfamilies are monophyletic (in the sense of Hennig, 1966) have been amply reinforced by this study. Two genera, Ctenopharyngodon and Hypophthalmichthys previously included in the Leuciscinae are here identified as representative of two monophyletic lineages only distantly related. As noted above (p. 28), the presence of a maxillary foramen (or its suggested past presence) associated with a rostral barbel is considered apomorphic for one branch of a diphyletic division of the Cyprinidae. It is to this assemblage that the monophyletic lineage (squaliobarbine group) containing Ctenopharyngodon belongs. Hypophthalmichthys is contained within the abramine group and belongs to that assemblage of taxa lacking the maxillary foramen. It is the majority of these taxa which are provisionally recognized as the subfamily Leuciscinae. Nikolsky (1954) made a similar dichotomous classification of the Cyprinidae. In one branch he included leuciscines, Aspius '. . . and other fish', the other he represented by Barbus, Cyprinus and Gobio. Nikolsky did not give formal recognition to these divisions and retained the formerly recognized subfamilies without assigning them to one or other lineage. The following discussion is concerned mainly with the classification and intrarelation- ships of the non-barbelled cyprinids (Leuciscinae). Classification of the Leuciscinae The 'subfamily' Leuciscinae (originally a division introduced by Bonaparte, 1846) has long been used as a category for containing 'generalized' cyprinids, including approximately 40% of Old-World genera. Not all authors have accepted the Leuciscinae; Berg (1949) assigned the majority of Old-World genera to the Cyprininae. On the other hand, Hubbs & Lagler (1958) extended the geographical range of the Leuciscinae to embrace New- World genera. Gosline's (1978) concept of the Leuciscinae embraced also the Cultrinae. These decisions demonstrate a lack of methodology in determining cyprinid supraspecific relationships. As it stands, the Leuciscinae includes at least four monophyletic assemblages; aspinine (Howes, 1978), abramine (reported herein), alburnine and phoxinine (unpublished) all of which are interrelated to other non-barbelled cyprinids presently recognized as four mono- phyletic lineages; bariliine (Howes, 1980), cheline (Howes, 1979), hemicultrine (Howes, 1978, part) and cultrine (Howes, 1978, part). It seems, however, that none of these groups is a likely candidate for the status of sister-group to the abramines. 48 G. J. HOWES Identification of the abramine sister-group is hampered by the non-resolution of the monophyly of the nucleus of the Leuciscinae, the genus Leuciscus. It appears that part of Leuciscus is related to the aspinine group and another possibly forms the sister-group to the abramines. Characters supporting an abramine-Leuciscus relationship are the presence in some species of Leuciscus of an incipient ventral basioccipital channel and a similarity in the ethmovomerine region. Abramis has been thought by some authors to be related to the American cyprinid genus Notemigonus (Miller, 1958:203; Banarescu, 1973). Contrary to this idea I find no synapomorphies linking Notemigonus to any member of the abramine group. At this point it must be stated that New- World Cyprinidae have not been surveyed in depth in an attempt to formulate their relationships with any element of the Old-World Leuciscinae. Only Ptychocheilus has been examined in detail and would seem (together possibly with Algansea) to share apomorphies with the Old-World aspinine group (work in preparation). Whilst it may well be that New- World genera such as Notropis and Chrosomus are closely related to the Eurasian Phoxinus (see Mahy, 1975), it is quite wrong to consider them as synonymous (Banarescu, 1973). It must be borne in mind that Phoxinus as it stands is also a paraphyletic group and such is likely to be the status of Chrosomus and Notropis. Christopher & Linder (1979) note the close similarity between Chondrostoma and the American Acrocheilus. Certainly the jaw morphology of both taxa are alike but no other cranial features demonstrate a close relationship. At some level, of course, the New- World cyprinids must form sister groups of the Old- World. Whether they represent a single monophyletic sister group to one or other of the monophyletic lineages contained within the 'Leuciscinae' or comprise several lineages forming twin taxa with other non-barbelled Old-World groups (cultrines, chelines, etc) are theories which remain to be tested. Leuciscine hybrids Futuyama (1979) notes that '. . . typological thinking focuses first on the "pure" forms and then interprets intermediate phenotypes as secondary products. If we focused on the continuum of phenotypes as a unit, we might not worry as much about which are the species and which the hybrids'. Futuyama's statement is supported by a revised concept of leuciscine genera. Former hypotheses for the relationships between leuciscine genera have depended upon degree of genetic compatibility as demonstrated by hybridization. The measure of relationship being determined by terminal development stage of offspring, survival rate and reproductive viability. A contrary opinion regarding the validity of hybridization in a phylogenetic context has been expressed by Rosen (1979). Rosen regards reproductive compatibility as a plesiomorphy marked by mosaic distribution within a group. This hypothesis would certainly account for those conflicting cases amongst cyprinids where supposedly phylo- genetically closely related taxa fail and distantly related ones succeed in being reproductively compatible. Such inconsistencies are usually explained away as misidentification of parental stocks. Suzuki's (1968) explanation for one such conflicting result was that '. . . Some Soviet species of Leuciscinae are probably more closely related to Cyprininae than Japanese species of Leuciscinae'. Such a statement disregards any kind of rigorous methodology in elucidating phylogenetic relationships. A greater number of naturally occurring intergeneric hybrids have been reported for Leuciscine genera than for possibly any other group of teleosts (see Schwartz, 1972). A more rigorous (cladistic) definition of these genera reduces many cases of intergeneric to interspecific hybridization (ie. crosses between 'Vimba' 'Blicca'x Abramis (Berg, 1949), 'Aristichthys'x Hypophthalmichthys (Voropayev, 1969), 'Scardinius'x Rutilus (many reports, see Schwartz, 1972). Naturally occurring intergeneric hybrids are judged to be so on the basis of intermediacy of characters, notably the pharyngeal dentition. A typical example of the use of 'intermediate' characters is given by Berg (1949 : 88-89) for a 'hybrid' popula- CHINESE MAJOR CARPS 49 tion of Leuciscus cephalus x Chalcaburnus chalcoides. As enumerated, these characters could well describe a populational variant of one of the presumed parental species. A similar argument can be applied to many so-called interspecific hybrids. The likelihood is that they are not crosses but clinal or other populational fragments. The numerous sub- species and varieties described in Leuciscine genera by Berg (1949) indicate a considerable degree of populational variability. This of course is not to deny that interspecific hybridiza- tion does occur. Wheeler (1976) notes that high levels of hybridization amongst cyprinid taxa occur in areas where recent introductions have been made. Certainly, nearly all the records I have checked for 'intergeneric' European cyprinid hybrids have been cases involving one or both of the taxa being recently introduced into a 'new' ecological situation. I conclude that 'hybridization' amongst cyprinid taxa must be treated with caution and is a useless character as a measure of phylogenetic affinity. Classification of the 'true' carps The squaliobarbine group belongs to the great assemblage of taxa characterized by the presence of a maxillary foramen and (usually) an associated rostral barbel (see p. 28). As yet no accounts of monophyletic groups, other than the squaliobarbines (dealt with herein), have been published, although Reid (1978, unpublished thesis) has identified a labeoine group. The hypothesis proposed above (p. 40) that the squaliobarbines form the sister group to all other barbelled carps is a temporary expedient. But, with available data I am unable to offer a more worthwhile solution. Such genera as Capoeta, Cyprinus and Carassius appear to be more closely related to the squaliobarbines than do other investigated taxa. Work in prepara- tion (Banister & Howes) seeks to establish monophyletic grouping amongst the barbelled cyprinids. In view of the suspected non-monophyly of so many currently recognized cyprinid assemblages, the reconstruction of ichthyogeographical relationships (eg. Banarescu, 1973) is fruitless. Acknowledgements I am most grateful to my colleagues in the Fish Section of this museum for their help and advice. Particular thanks are due to Drs P. Humphry Greenwood and K. E. Banister for their critical comments of the manuscript and discussions of cyprinid phylogeny; to Alwynne Wheeler for his helpful comments concerning British cyprinids; to Margaret Clarke for her help with synonymies and literature searches; to Robert Travers for preparing alizarin specimens; and to Alice Parshall for providing me with unpublished information on cobitid osteology. Sara and Dr William Fink generously allowed me to see their unpublished manuscript on ostariophysan anatomy. To Dr Chu Xin-Luo of the Kunming Institute, China go my warmest thanks for supplying many small specimens of Hypophthalmichthys. Finally, my thanks to Gina Sandford for typing the manuscript. References Alexander, R. McN. 1964. Adaptation in the skull and cranial muscles of South American characinoid fish. J. Linn. Soc. (Zool.) 45 : 1 69-1 90. 1966. The functions and mechanisms of the protrusile upper jaws of two species of cyprinid fish. J. Zool. Lond. 149 : 188-296. Backiell, T. & Zawisza, J. 1968. Synopsis of biological data on the bream Abramis brama (Linnaeus, 1 758). FAO Fish. Synopsis no. 36 : 1 1 1 pp. Rome. Banarescu, P. 1963. Le genre Vimba (Pisces, Cyprinidae) dans le bassin du Danube. Trav. Mus. hist. nat. 'Grigore Antipa' 4 : 381-400. 50 G. J. HOWES 1970. Remarks on the genus Xenocypris (Pisces, Cyprinidae) with description of a new subspecies. Revue roum. Biol. Zool. 15 (6) : 395^02. 1973. Some reconsiderations of the zoogeography of the Euro-Mediterranean fresh-water fish fauna. Revue roum. Biol. Zool. 18 (4) : 257-264. Banister, K. E. & Bunni, M. K. 1980. A new blind cyprinid fish from Iraq. Bull. Br. Mus. nat. Hist. (Zool.) 38 (3): 151-158. Berg, L. S. 1949. Freshwater fishes of the USSR and adjacent countries (English translation of the 4th ed.) 2: 496 pp. Berry, P. Y. & Low, M. P. 1970. Comparative studies on some aspects of the morphology and histology of Ctenopharyngodon idellus, Aristichthys nobilis and their hybrid (Cyprinidae). Copeia 1970 (4) : 708-726. Bonaparte, C. L. 1 846. Catalogo metodico dei pesci Europei. Alii Soc. ital. Sci. nat., : 1-95. Boulenger, G. A. 1901. On the presence of a suprabranchial organ in the cyprinid fish. Hypophthal- michthys. Ann. Mag. nat. Hist. (7)8 : 186-188. Christopher, W. E. & Linder, A. D. 1979. The osteocranium of the cyprinid fish Acrocheilus alutaceus. Tebiwa(\6): 1-19. Chu, Y. T. 1935. Comparative studies on the scales and pharyngeals and their teeth in Chinese cyprinids, with particular reference to taxonomy and evolution. Biol. Bull. St. John's Univ., Shanghai 2 : 1-225. Cross, D. J. 1969. Aquatic weed control using grass carp. J. Fish Biol. 1 : 27-30. Eastman, J. T. 1971. Pharyngeal bone muscles of the carp, Cyprinus carpio. J. Morph. 134(2): 131-140. 1977. The pharyngeal bones and teeth ofcatostomid fishes. Amer. Midland Nat. 97 (1) : 68-88. & Underbill, J. C. 1973. Intraspecific variation in the pharyngeal tooth formulae of some cyprinid fishes. Copeia 1973 (1) : 45-53. Edwards, D. J. 1974. Weed preference and growth of young grass carp in New Zealand. N.Z.J. Mar. Fwtr. Res. 8(2) :34l-350. Fang, P. W. 1928. Notes on the gill-rakers and their related structures of Hypophthalmichthys nobilis and H. molitrix. Contrib. Biol. Lab. Sci. Soc. China 4 (5) : 1-30. Fiebiger, J. 1931. Ober den bau und die Mechanik des Karpfenriissels. Zeit. mikr. anal. Forsch. 27:582-612. Fink, S. V. & Fink, W. L. 198 1 . Interrelationships of the ostariophysan fishes (Teleostei). Zool. J. Linn. Soc. 73 : in press. Futuyama, D. J. 1979. Evolutionary biology. Massachusetts. 565 pp. Gill, T. 1893. Families and subfamilies of fishes. Mem. Nat. Acad. Sci. Wash. 6 : 127-138. Goodrich, E. S. 1909. Vertebrata Craniata Fasc. 1. Cyclostomes and fishes. In: A. treatise on Zoology (ed. R. Lankester)9:518pp. Goren, M., Fishelson, L. & Trewavas, E. 1973. The cyprinid fishes of Acanthobrama Heckel and related genera. Bull. Br. Mus. nat. Hist. (Zool.) 24 (6) : 293-3 15. Gosline, W. A. 1974. Certain lateral-line canals of the head in cyprinid fishes, with particular reference to the derivation of North American forms. Jap. J. Ichth. 21 (1 ) : 9-15. 1978. Unbranched dorsal-fin rays and subfamily classification of the fish family Cyprinidae. Occ. Pap. Mus. Zool. Univ. Mich. 684 : 1-21. Giinther, A. 1 868. Catalogue of fishes in the British Museum 1 London : 5 12 pp. Harrington, R. W. 1955. The osteocranium of the American cyprinid fish, Notropis bifrenatus with an annotated synonymy of teleost skull bones. Copeia 1955 (4) : 267-290. Hennig, W. 1966. Phylogenetic systematics. Urbana. 263 pp. Hickling, C. F. 1967. On the biology of a herbivorous fish, the White Amur or grass carp, Ctenopharyngodon idella Val. Proc. R. Soc. Edinburgh (B) 70 : 62-8 1 . Holmgren, N. & Stcnsio, K. A. 1936. Kranium und Visceralskelett der Akranier Cyclostomem und Fische: 233-500 In: Handbuch der vergleichenden Anatomie der Wirbeltiere 4 (ed. Bolk, L., Goppert, E., Kallius, E. & Lubosch, W.) Howes, G. J. 1978. The anatomy and relationships of the cyprinid fish Luciobrama macrocephalus (Lacepede). Bull. Br. Mus. nat. Hist. (Zool. ) 34 (1) : 1-64. 1979. Notes on the anatomy of Macrochirichthys macrochirus (Valenciennes) 1844 with comments on the Cultrinae (Pisces, Cyprinidae). Bull. Br. Mus. nat. Hist. (Zool.) 36 (3) : 147-200. 1980. The anatomy, phylogeny and classification of the bariliine cyprinid fishes. Bull. Br. Mus. nat. Hist. (Zool.) 37 (3) : 129-198. Hubbs, C. L. & Lagler, K. F. 1958. Fishes of the Great Lakes region. 2 1 3 pp. University of Michigan. CHINESE MAJOR CARPS 5 1 Inaba, D. & Nomura, M. 1956. On the digestive system and feeding habits of young Chinese carps collected in the river Tone. J. Tokyo Univ. Fish. 42 (1 ) : 1 7-24. Kryzanovskij, S. G. 1947. Sistema semeistva karpovych ryb (Cyprinidae). Zoo/. Zh. 26 (1) : 53-64. Lin, S. Y. 1935. Life history of waan we Ctenopharyngodon idellus (C. and V.) Lingnan Sci. J. 14: 129-135. Mahy, G. 1975. Osteologie comparee et phylogenie des poissons cyprinoides. 1. Osteologie cranienne du goujon a fines ecailles, Chrosomm neogaeus (Cope). Naturaliste Can. 102 : 1-31. Miller, R. R. 1958. Origin and affinities of the freshwater fish fauna of Western North America. Publ. Amer. Ass. Advanc. Sci. 51 (1958) : 187-222. Miller, R. V. 1969. Constancy of epibranchial organs and fourth epibranchial bones with species groups ofclupeid fishes. Copeia 1969 (2) : 308-312. Nikolsky, G. V. 1954. Special Ichthyology 2nd ed. Moscow. 458 pp. 1 956. Fishes of the Amur basin. Moscow. 55 1 pp. Okada, Y. 1960. Studies on the freshwater fishes of Japan II. Special Part. J. Fac. pref. Univ. Mie-Tsu 4 (3) : 589-860. Oshima, M. 1919. Contributions to the study of the fresh water fishes of the island of Formosa. Ann. Carnegie Mm. 12 (2-4) : 169-328. Patterson, C. 1975. The braincase of pholidophorid and leptolepid fishes, with a review of the actinopterygian braincase. Phil. Trans. Roy. Soc. B269 (899) : 275-579. Pellegrin, J. 1934. Poissons de la region d'Hanoi (Tonkin). Description d'une variete nouvelle. Bull. Soc. Zoo/. Fr. 59 : 334-337. Ramaswami, L. S. \955a. Skeleton of cyprinoid fishes in relation to phylogenetic studies: 6. The skull and Weberian apparatus of the subfamily Gobioninae (Cyprinidae) Ada Zoo/. 36 (2) : 1 27-1 58. 1955/7. Skeleton of cyprinoid fishes in relation to phylogenetic studies: 7. The skull and Weberian apparatus of the subfamily Cyprininae (Cyprinidae) Ada Zoo/. 36 (3) : 1 99-242. Regan, C. T. 1911. The classification of the teleostean fishes of the order Ostariophysi. 1. Cyprinoidei. Ann. Mag. nat. Hist. (8)8 : 13-32. 1922. The distribution of the fishes of the order Ostariophysi. Bijdr. Dierk. 22 : 203-207. Reid, G. M. 1978. A systematic study of labeine cyprinid fishes with particular reference to the comparative morphology, functional morphology and morphometrics of African Labeo species. Unpublished Ph.D. thesis Univ. London. 1978. Roberts, T. R. 1973. Interrelationships of ostariophysans. In: Interrelationships of fishes. Greenwood, P. H., Miles, R. S. & Patterson, C. (Eds) : 373-395. London and New York. Rosen, D. E. 1979. Fishes from the uplands and intermontane basins of Guatemala: revisionary studies and comparative geography. Bull. Am. Mus. nat. Hist. 162 art. 5 : 267-376. Ryabov, I. N. 1973. Embryonic and larval development characteristics of hybrids of silver carp (Hypophthalmichthys molitrix (Val.)) and eastern bream (Abramis brama orientalis (Berg)). Probl. Ichthy. 13:701-714. Saether, O. A. 1979. Underlying synapomorphies and anagenetic analysis. Zoo/. ScriptaS : 305-312. Sagemahl, M. 1891. Beitrage zur vergleichenden Anatomic der Fische, IV. Das Cranium der Cyprinoiden. Morph. Jahrb. 17 : 489^95. Sauvage, H. E. 1884. Contribution a la faune ichthyologique du Tonkin. Bull. Soc. zoo/. Fr. 9:209-215. Schwartz, F. J. 1972. World literature to fish hybrids, with an analysis by family, species and hybrids. Gulf. Coast Res. Lab. Publ. 328 pp. Shutov, V. A. 1967. Bream with a double row of pharyngeal teeth from Lake Seliger. Dokl. (Proc.) Acad. Sci. USSR 173 : 283-286. 1969. Revision of the genus Blicca and some data concerning phyletic relationships between representatives of the genus Abramis (Pisces, Cyprinidae). Zoo/. Zh. 48 : 1 105-1 107. Starks, E. C. 1926. Bones of the ethmoid region of the fish skull. Stanf. Univ. Public. Biol. Sci. 4(3): 139-338. Suzuki, R. 1968. Hybridization experiments in cyprinid fishes. XI. Survival rate of Fl hybrids with special reference to the closeness of taxonomical position of combined fishes. Bull. Freshwat. Fish. Res. Lab. Tokyo 18 (2) : 1 1 5-1 55. Tomodo, Y. 1979. Discovery of the late Cenozoic Xenocypridinae (Cyprinidae) fishes from Tokai and Kinki Districts. Mem. Nat. Sci. Mus. Tokyo 12 : 93-101 . Tretiakov, D. K. 1946. Systematic groups of Cyprinidae. Zoo/. Zh. 25 (2) : 149-156. Tsui, Y. K. 1936. A preliminary study of the growth rate of grass-carp, big-head, mud carp and the common carp. Fish Newslet. 1(1): 35-5 1 . 52 G. J. HOWES Vari, R. P. 1979. Anatomy, relationships and classification of the families Citharinidae and Distichodontidae (Pisces, Characoidea). Bull. Br. Mus. nat. Hist. (Zool.)36 (5) : 261-344. Verigin, B. V. 1957. Structure of the gill apparatus and the suprabranchial organ of Hypophthal- michthys molitrix Val. Zoo/. Zh. 36 : 595-602. Voropayev, N. V. 1969. Morphological features and some fish cultural characteristics of Aristichthys nobilis, Hvpophthalmichthvs molitrix and of their hybrides (sic.). Trud. vses. nauch. isslcd. Inst. Prud. rybKho: 16: 272-281. Watanabe, M. 1951. The osteological studies of some Chinese cyprinoid fishes introduced into Japan. Misc. Rep. Inst. Nat. Res. Tokyo 19-21 (2) : 14-18. Wheeler, A. C. 1976. On the populations of roach (Rutilus rutilus), rudd (Scardinius erythroph- thalmus), and their hybrid in Esthwaite Water, with notes on the distinctions between them. J. Fish a/o/. 9: 39 1-400. Weisel, G. F. 1960. The osteocranium of the catostomid fish, Catostomus macrocheilus. A study in adaptation and natural relationship. J. Morph. 106 : 109-129. Wilamovski, A. 1972. Structure of the gill apparatus and the suprabranchial organ of Hypoph- thalmichthys molitrix Val. (Silver carp). Bamidgeh 24 (4) : 87-98. Winterbottom, R. 1974. A descriptive synonymy of the striated muscles of the teleostei. Proc. Ac ad. nat. Sci. Philad. 125 : 225-3 1 7. Zambriborshch, F. S. 1955. Adaptation of pharynx in Hypophthalmichthys molitrix to phyto- planktonic food. Dokl. Akad. Nauk SSR 105 : 376-379. 1957. Structure and function of the suprabranchial organ of Hypophthalmichthys molitrix Val. Zool. Zh. 36 : 587-594. Manuscript accepted for publication 28 November 1 980 British Museum (Natural History) 1881-1981 Centenary Publications The Natural History Museum at South Kensington By W. T. Steam This book presents a full history of the Museum, its collections, Directors and eminent members of staff. It provides for the first time an account of the major research undertaken and insights into the personalities of the key people in the Museum's development and evolution. Co-published with William Heinemann. Alfred Waterhouse and the Natural History Museum By Mark Girouard Designed by Alfred Waterhouse in the 1870's the Museum is surely one of London's most outstanding pieces of architecture. This attractively illustrated book describes the development of the design and highlights some of the building's most interesting features. British Museum Natural History By Peter Whitehead & Colin Keates The Museum is really a huge scientific research institution - acquiring, describing and classifying all manner of natural history material both specimens and artworks. It has some of the richest collections of their kind in the world and it is these and how they are obtained and managed that are the subject of this book. The lavish, full colour illustrations and lively text will appeal to everybody interested in natural history. Co-published with Philip Wilson Ltd. Chance, Change and Challenge This multi-author twin volume work is one of the Museum's most ambitious publishing projects. In the first volume The Evolving Earth twenty scientists have been asked to summarise the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume The Evolving Biosphere Museum scientists have chosen an evol- utionary concept - speciation, coevolution, biogeography etc and related this to the group of animals or plants in which they are specialising. Co-published with Cambridge University Press. Animal Identification-A Reference Guide VOLUME 1 : MARINE AND BRACKISH WATER. Edited by R, W. Sims. VOLUME 2 : TERRESTRIAL AND FRESHWATER. Edited by R. W. Sims. VOLUME 3 : INSECTS. Edited by D. Hollis. These guides provide the reader with lits of primary sources of reference that can be used to identify (or lead to the identification of) living animals throughout the world. The references are arranged in systematic and geographical sequence in order to facilitate searching. Co-published with John Wiley & Sons Limited. Nature Stored Nature Studied A short guide complementing the exhibition of that name to be staged in the Museum throughout 1981. It gives a fascinating insight into the history of the Museum collections and the research carried out "behind the scenes". Titles to be published in Volume 41 Anatomy and phylogeny of the Chinese Major Carps Ctenopharyngodon Steind., 1866 and Hypophthalnnchthys Blkr., 1860. By Gordon Howes Morphology and systematics of some interior-walled cheilostome Bryozoa By P. L. Cook and P. J. Chimonides Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae. By Ailsa M.Clark Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural History) Morphology and systematics of some interior- walled cheilostome Bryozoa P. L. Cook & P. J. Chimonides Zoology series Vol 41 No 2 24 September 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull Br. Mus. not. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Dr J. G. Sheals Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISSN 0007-1498 Zoology series Vol 41 No 2 pp 53-89 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 24 September 1981 Morphology and systematics of some interior- walled cheilostome Bryozoa P. L. Cook and P. J. Chimonides Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Contents Synopsis 53 Introduction 53 Material and methods 54 Measurements and abbreviations used 57 Key to species described 58 Systematic section 58 Family Euthyrisellidae Bassler 58 Family Didymosellidae Brown 68 Discussion 81 Summaries in French and German 83 Acknowledgements 85 References 85 Synopsis The colony structure and zoodial morphology found in two cryptocystidean cheilostome bryozoan Families, the Euthyrisellidae and the Didymosellidae are described. All calcified walls are interior, and colony-wide (extrazooidal) basal coeloms are present in nearly all species. In the Family Euthyrisellidae, interzooidal visceral-to-hypostegal communication complements or even replaces the more usual, intrazooidal kind, which is that found in the Family Didymosellidae. In the Euthyrisellidae an extension of the hypostegal coelom on the basal side of the calcified frontal shield is present in most species. One new species, Tropidozoum burrowsi, is introduced. The Family Didymosellidae has a fossil record extending from the Middle Eocene to the Recent, and a wide Tertiary-to-Recent distri- budon. One new Recent species, Didymosella pluma, is introduced, and Tubiporella boninensis Borg is redefined. Tubiporella levinseni Borg is redescribed and assigned to the umbonuloid ascophoran genus Reptadeonella. Introduction The calcified and uncalcified body walls defining the member zooids of bryozoan colonies may conveniently be regarded as belonging to two kinds. The first is the exterior, limiting or boundary wall, which is primarily cuticular and which, if calcified, almost invariably has an outermost layer of cuticle throughout ontogeny. This kind of wall is actually or potentially capable of dividing the internal coelomic tissues of zooids from the environment. Zooids delimited by this kind of wall develop as uncalcified buds by expansion and extension of coelomic tissue through septulae in the calcified wall of an existing zooid. The boundary wall of cuticle and epidermis expands intussusceptively as growth proceeds. The earliest known, Late Jurassic cheilostomes had zooidal walls almost entirely of this kind (see Boardman & Cheetham, 1973), and large numbers of species now living have the same, or similar zooidal structure and budding pattern. Many lateral vertical walls of contiguous Bull. Br. Mus. nat. Hist. (Zool.) 41 (2) 53-89. Issued 24 September 1 98 1 53 54 P. L. COOK & P. J. CHIMONIDES zooids in expanses of more integrated species are also exterior walls, even though they may appear to be "inside' the colony. The second kind of wall is the interior wall, which may or may not possess any cuticular element, and which partitions existing coeloms and is not directly in contact with the environment. Zooids with interior walls are differentiated from a colony-wide expansion of coelom bounded by intussusceptively expanding cuticle and epidermis, by growth of a calcified lamina which, in earlier stages of ontogeny at least, is surrounded by coelom (see Boardman & Cheetham, 1973). Both kinds of wall have arisen several times in the long evolutionary history of the Bryozoa, and both kinds are usually to be found in any cheilostome colony. In some colonies, the distinction between the two kinds of wall is not clear cut. However, the astogenetic and ontogenetic sequences reflected by interior and exterior walls, and the intrazooidal and interzooidal communications associated with them, are important in the study of colony structure, of integration, and of systematic relationships through time. Colonies with a high proportion of their calcified walls of the interior kind are well known among the 'free-living' anascan cheilostomes (Hakansson, 1973, Chimonides & Cook, 1981). Like these, several ascophoran Families, such as the Conescharellinidae, and genera such as Sphaeropora, as well as the Euthyrisellidae and Didymosellidae, have virtually no exterior calcified walls at all (Cook & Lagaaij, 1976, and Cook & Chimonides, 1981). The ascophoran forms show a much larger range of colony morphotypes than the anascan forms. The species of Euthyrisellidae and Didymosellidae have cryptocystidean frontal shields (see Banta, 1970). These frontally placed parts of zooids in which the orifice is situated, are interior calcified walls which partition the visceral from the hypostegal coelom. The exterior frontal wall above the hypostegal coelom is completely cuticular, with underlying epidermal cells. In some species (e.g. Pleurotoichus clathratus) the frontal, cuticular wall is a continuous, colony-wide expanse, connected only intermittently with the vertical walls and the frontal shields. The lateral and distal vertical calcified walls are also interior walls, which partition a colony-wide coelom at the growing edge or tip of an expanse or branch. Basal calcified walls are also interior walls in the majority of species described here. These too, have communication with an extrazooidal basal coelom, which is protected from the environment by a cuticular and epidermal wall. Intrazooidal, interzooidal and extrazooidal communication is an important feature in these colonies (Figs 1A, B) and its skeletal expression, in the form of septulae, and often in subsequent thickening of calcification, enables inferences to be made for fossil and Recent specimens in which the cuticular walls and coelomic tissues are no longer present. Material and methods The possibility of restoring recognizable structures, including muscles and viscera, in previously dried material, and the advantages of study of epoxy-resin mounted thin-sections of 'hard' and 'soft' tissues in situ have been demonstrated (Cook, 1975). Dried specimens were previously soaked in 10% trisodium phosphate solution for 48 hours, gently cleaned in detergent solution, and rinsed in water before staining. The accession of fixed, well-preserved material of Pleurotoichus, Tropidozoum, Neoeuthyris and Didymosella has enabled us to make sections which may be compared with those made from restored material. This com- parison has tested the inferences drawn previously, and will allow further inferences to be made with greater confidence in the future. Specimens for scanning electron microscopy were treated with sodium hypochlorite, rinsed and coated with gold-palladium. Specimens to be thin-sectioned were heavily stained in borax carmine or safranin, impregnated with epoxy-resin, cut and polished attached to 3" x 1 " microscope slides. Early attempts at embedding followed Nye el al (1972) and Sandberg (1971). Three resin types were available: Reichold Epotuf 37 128 with hardener 37 614, Shell Epon 812 and INTERIOR-WALLED CHEILOSTOME BRYOZOA 55 h_ sf Fig. 1 Generalized diagrams to show residual communication after development of interior calcified walls in: (A) Euthyrisellidae; (B) Didymosellidae. Intercoelomic communications arrowed: h hypostegal coelom, sf subfrontal hypostegal coelom, v visceral coelom, b basal coelom. 56 P. L. COOK & P. J. CHIMONIDES Ciba Araldite CY212 with their complementary additives MNA, DDSA, BDMA, DMP 30 etc. The Reichold resin, well tested by Nye et al (1972) was found to be difficult to use. The high viscosity of the components hindered obtaining the intended ratio of 4 parts resin to 1 part hardener and their subsequent mixing. The mix was found to be sensitive to heat and U.V. light (e.g. sunlight) and there was a tendency for thermal runaway to occur. This reaction was more readily provoked if the mix contained slightly excess hardener, even though the plug size was only 30 cm3 and curing was initiated at room temperature, 20°C. It was found that surface temperatures in excess of 80°C were reached, coupled with emission of phenol-smelling fumes, bubbling and volumetric expansion. When curing could be con- trolled by refrigeration a volumetric contraction, estimated as 8% total, was noted. No significant change in behaviour was noted with prolonged storage of the components at 5-10°C. The recommended life of hardener 37 614 is given as 6-12 months in a tightly closed container at room temperature or less. Denaturing is characterized by progressive discolouration to a dark brown colour. Epon 812 was used as described by Sandberg (1971) using only MNA as hardener. DMP 30 and BDMA (Glauert, 1965, 1974) were used as accelerators in parallel series. Araldite CY212 was used in the same way. The preferred ratios of the components were: resin 54%, hardener 44% and accelerator 2%. Both resins were found to have tolerant curing temperatures of 50-80°C. Epon 812, with its low viscosity and production of relatively clear and hard plugs promised to be the ideal medium. It was found, however, that as lapping of thin-sections progressed, small, granular particles of resin (and sometimes of specimen) would break away leaving score marks on the polished surface. This was more noticeable as the thickness of the preparation reached 40 u or less. It was almost impossible to produce an intact section below 20 ji. Freeman & Spurlock (1962), in electron microscopical studies, reported Epon 812 as being inherently granular. This suggests that thin-section disintegration may be a larger scale manifestation of the same characteristic. The disintegration problem was more pronounced at the lower curing temperatures. In addition, it was found that the resin in sections became 'milky' in aqueous solutions (even without the aquon component), or in the lapping lubricant used, Engis Hyprez type W. The resin would also accept stains which were difficult to remove. Further, as the section was thinned, it would distort and lift from the slide. These effects were noted regardless of resin and additive batch and of precautions to ensure uniform mixing of com- ponents and cleanliness of the glass slides. A noticeable volumetric contraction (of about 6-7%) was again found, regardless of recipe and of curing procedure. Araldite CY212, in spite of its high viscosity, was found to be the most suitable resin. Its total curing contraction was estimated as 1-2%, regardless of recipe. Stability during lapping was high, and it was possible to obtain sections in the order of 3 u thickness. One disadvantage was that xylene affected the cured resin and coverslides mounted with Canada balsam were therefore avoided. Where required, coverslides were mounted using more epoxy-resin. In all cases propylene oxide was used as the antemedium. Final embedding was vacuum assisted after first soaking the specimens in an equal mixture of propylene oxide and accelerated resin mix. One well known advantage of the use of propylene oxide is that any residual quantities trapped in the specimen should polymerize with the resin. In all resins used, behaviour appeared equally divided between pre- and post-gel contraction. Reichold Chemicals Inc. (pers. comm. 16.2.1977) suggested that a more typical contraction behaviour of their resin would be approximately 5% total, distributed as 3% pre-gel and 2% post-gel phase. Potter (1970 : 171) identified post-gel shrinkage as the phase at which stresses may be introduced which can lead to damage, although his account relates to electrical component encapsulation and is without reference to scale. It was indicated that post-gel contraction for epoxy resins is normally in the region of 2-3% and that contraction overall is affected primarily by the accelerator used. In making thin-sections the effects of resin contraction were most apparent when attaching preparations to microscope slides, when bowing of the glass, and in many cases, even INTERIOR-WALLED CHEILOSTOME BRYOZOA 57 breaking was noted. The amount of resin was therefore kept to a minimum, and Araldite resin obviously showed to advantage. Within the specimen itself, shrinkage should have little effect providing that complete penetration by the resin has been achieved. Although penetration may appear to be com- plete, it has been found, during cutting and polishing stages, that calcified walls may drop out of the section completely, as opposed to partial fracturing and partial loss. This indicates a lack of specimen support. If a final surface stain was applied, there was a tendency for the stain to delineate the boundary of the calcification, although it should be borne in mind that the micro-relief produced by lapping may have been a contributory factor. In sections of 10 u thickness or more, the stain sometimes soaked into the boundary irregularly. Support failure and staining characteristics suggest an unpredictable penetration of the resin into the calcifi- cation, together with a degree of incomplete adhesion which may be exacerbated by resin shrinkage. The evidence is at present considered to be inconclusive, especially as the effects were observed most often when using the structurally unsatisfactory Epon 812 resin. For these reasons staining before embedding was preferred. Measurements and abbreviations used The value of comparative measurements among, for example fossil and Recent colonies of the same or different species, varies considerably among bryozoans, both with the structures measured and with the species groups being investigated. For example, some species retain recognizable traces in frontal view, of structures such as the tops of vertical walls, through- out ontogeny and in spite of wear. In such species comparisons of zooidal length and width among colonies are easily made and are significant. Primary calcified orifices with distinct boundaries are also capable of yielding reasonably accurate measurements. In the species considered here, one family, the Euthyrisellidae, which has no fossil record, retains throughout ontogeny recognizable features which allow morphometric analysis. The Didymosellidae, on the other hand, includes species in which continuous ontogenetic changes throughout colony life obscure zooidal boundaries and primary orifices, which in themselves are not easily definable. Fossil (and some Recent) specimens, which have been damaged or are worn, are therefore not strictly comparable with well preserved colonies. Other micro-environmental effects, such as those influencing the length of peristomes, may sometimes be recognized, but are difficult to quantify. Abbreviations Measurements in millimetres Length of autozooid Lz Width of autozooid lz Length of brooding zooid Lbz Width of brooding zooid Ibz Length of primary orifice Lo Width of primary orifice lo Length of brooding z. orifice Lbo Width of brooding z. orifice Ibo Length of rostrum Lr Width secondary orifice 12o Length of mandible Lm BMNH British Museum (Natural History) MM Manchester Museum NMV National Museum of Victoria AM Australian Museum, Sydney. ZMA Zoological Museum, Amsterdam SAM South African Museum, Cape Town. 58 P. L. COOK & P. J. CHIMONIDES Key to species described 1 Raised peristomes absent, opercula well-developed, often with two layers and with complex sclerites. Frontal pores and septulae absent or rare; large frontal foramina present or absent. Avicularia and ovicells usually absent. Secondary calcification slight. Brooding zooids strongly dimorphic EUTHYRISELLIDAE 2 Raised peristomes present, opercula simple, not usually visible. Frontal pores and septulae numerous. Avicularia and ovicells present. Secondary calcification extensive. Dimorphic zooids not obvious DIDYMOSELLIDAE 6 2 Colonies encrusting, avicularia and ovicells present, basal coelom absent Neoeuthyris woosteri p. 67 - Colonies erect. Avicularia and ovicells absent, basal coeloms present .... 3 3 Colonies flustriform, unilaminar, with lateral and basal coeloms 4 Colonies cellariiform, with kenozooidal nodes. Internodes with zooids in curved, laterally contiguous series. Basal coelom forming a central column . . . Tropidozoum 5 4 Frontal shield complete, smooth. Lateral coeloms with series of calcified plates Euthyrisella obtecta p. 60 Frontal shield with calcified bars and large, slit-like foramina. Lateral coeloms without series of calcified plates Pleurotoichus dathratus p. 63 5 Autozooidal orifices not sinuate, frontal shield with large foramina Tropidozoum cellariiforme p. 64 - Autozooidal orifices with a deep, triangular sinus, frontal shield with small foramina T. burrowsisp.nov. p. 65 6 Peristomes with large, paired proximal foramina, spiramen absent . Didymosella - Peristomes with one, or no foramen, spiramen present . . . . Tubiporella 13 7 Colonies erect, ligulate, unilaminar. Zooids in 2-5 transverse rows. Basal pores present 8 - Colonies encrusting, semi-encrusting or repent, or bilaminar or cylindrical ... 9 8 Branches flat, 4-5 zooids wide, basal pores numerous Didymosella porosa p. 72 Branches laterally curved towards the basal side, 2-3 zooids wide, basal pores rare D. clypeata p. 72 9 Colonies bilaminar or cylindrical D. /arva/w-complex (fossil populations) p. 73 Colonies unilaminar or semi-encrusting, with large basal foramina, or basal pores and/or calcified processes 10 10 Colonies encrusting or semi-encrusting. No coelom basally, but zooid walls uncalcified or with a large, central foramen. Peristomes deficient distally, raised proximally D. /arva/w-complex (Recent populations) p. 73 Colonies repent, with basal pores and/or long calcified processes. Peristomes tubular . 1 1 1 1 Zooids large Lz>0'70 mm), avicularian rostra with an open channel terminally 12 Zooids smaller (Lz < 0'70 mm), avicularian rostra very small and acute parviporosa p. 75 1 2 Zooids with more than 25 frontal pores. Basal surface with pores (fossil) D. crassa p. 7 1 Zooids with less than 25 frontal pores. Basal surface with long, calcified processes. Avicularian mandible expanded laterally (Recent) . D. pluma sp. nov. p. 76 13 Colonies forming broad, repent expansions. Peristomes with a single foramen early in ontogeny. Avicularia with acute, hooked mandibles Tubiporella magnirostris p. 77 - Colonies ligulate, 4-7 zooids wide, repent. Peristomes entire, very long and tubular. Avicularia elongated, with laterally expanded mandibles . T. boninensis p. 79 Systematic Section EUTHYRISELLIDAE Bassler Euthyrisellidae Bassler, 1953 : G226. TYPE GENUS. Euthyrisella Bassler, 1936. DESCRIPTION. Interior walled cryptocystidean ascophorans, usually with extrazooidal, cuticle-bounded basal coeloms. Colonies encrusting algae, or erect, flexible, rooted. Zooids INTERIOR-WALLED CHEILOSTOME BRYOZOA 59 Table 1 Average measurement (mm) of zooids of Euthyrisellidae. Lz Iz Lbz Ibz Lo lo Lbo Ibo E. obtecta 0-55 0-35 0-73 0-52 0-15 0-19 0-20 0-27 P. clathratus 0-55 0-37 0-60 0-51 0-24 0-24 0-24 0-31 T. cellariiforme 0-65 0-34 0-73 0-38 0-21 0-18 0-24 0-19 T. burrowsi 0-67 0-36 0-67 0-50 0-21 0-15 0-15 0-20 N. woosteri 0-90 0-41 1-60 0-50 0-23 0-17 0-26 0:3 1 with relatively little secondary calcification frontally, frontal septulae restricted or absent. Hypostegal coeloms extensive, often extending basally to the calcified frontal shield, and communicating interzooidally with visceral coeloms of zooids of the previous astogenetic generation. Externally obvious brooding structures and avicularia usually, but not ex- clusively absent. Orifices of brooding zooids enlarged, dimorphic. REMARKS. The family was introduced by Bassler (1953) for the genera Euthyrisella, Pleurotoichus, Neoeuthyris and Urceolipora. The relationships among these genera have been discussed by Levinsen (1909), Harmer (1957) and Hastings (1964). Urceolipora shows many characteristics in common with Euthyrisella and Neoeuthyris on the one hand, and with the genera Calwellia and Icthyaria on the other. These generic groups, together with others which have some similarities in their colony organization, such as Onchoporella and Euthyroides all require detailed analysis, but are beyond the scope of this paper. Urceolipora is therefore excluded here, but the genus Tropidozoum is included in the family. ONTOGENY OF ZOOID WALLS. In all species, the primary ontogenetic changes in calcification of the zooid walls and formation of the orifice, operculum and ascus, appear to be very rapid, as few intermediate stages between undifferentiated buds and complete zooids are present. In addition, growth appears to proceed in distinct episodes, alternating with periods of astogenetic quiescence, as many colonies show little evidence of any zones of developing zooids at the growing edges or tips, and some colonies show distinct 'growth lines' (see below). Where present, the growing zones consist of an undifferentiated area of coelomic tissue bounded by thin cuticle. The calcification of interior walls proceeds distally, and the lateral walls become intermittently attached to the frontal cuticle. At the same time the opercular cuticle becomes thickened and differentiated. Where there is a basal coelom, the calcified lateral walls converge, forming a basal zooid wall which may be linear, or merely a curved continuation of the lateral walls. Contact with the cuticular, basal colony-wide wall occurs later in ontogeny in some species. The development of the lophophore and viscera is com- plete 2-3 zooid generations from the growing zone, just in advance of the final stages of ascus development (see also Harmer, 1902). In Tropidozoum, particularly T. burrowsi (see p. 65, there is some subsequent thickening of the calcification of the frontal shield and basal walls. In Euthyrisella, Pleuro- toichus and Neoeuthyris, thickening is confined to the development of processes attaching the cuticle to the frontal shields, or to the development of calcification within the lateral extrazooidal coeloms, when present. Coelomic systems and communication The intercommunication among the hypostegal, visceral and extrazooidal coeloms is con- siderable, and reflects the varying extent and patterning of these coeloms throughout the colony (Fig. 1A). The virtual isolation of the hypostegal coelom, particularly in Tropidozoum, is a direct result of the rarity of the frontal septulae (areolae) usually found in cryptocystidean ascophorans, and is accompanied in most species by an extension of this coelom beneath the calcified frontal shield (see also Cook, 1975). Transfer of nutrients in 60 P. L. COOK & P. J. CHIMONIDES coelomic fluids is inferred by the presence of communication pores in the distal-terminal or distal-lateral calcified walls, which connect the visceral coelom of one zooid with the hypostegal coelom of the next successive, distal zooid. This system, too, is unusual in cheilostomes, although it probably occurs in the genera Urceolipora, Calwellia and Onchoporella as well (see above). The basal and lateral extrazooidal coeloms are extensive in all the erect genera, but absent in Neoeuthyris, which is encrusting. Where present, they are inferred to function connectively and, as their cuticular walls thicken with ontogeny, to become involved in the anchoring and support of the colony. Opercula The opercula are large and double layered, with strengthening sclerites in most species. The morphology of the complex opercula of Euthyrisella and Pleurotoichus has been fully described by Harmer (1902). Those of Tropidozoum and Neoeuthyris are simpler in con- struction, but also have strong sclerite systems. Brooding zooids All species show dimorphism of brooding zooids, the larger dimorphs having enlarged orifices and, in the erect species, swollen basal walls. Embryos have now been observed in all species and without exception, are associated with the larger dimorphs. They are not readily seen except by transmitted light (cf. Harmer, 1902), but when present, usually fill the central and proximal parts of the cystid and are presumably contained in ovisacs. The ovicells of Neoeuthyris differ considerably in appearance from the brooding zooids of the other genera. However, the brooding cavity may be regarded as a displacement of the swollen basal cavity in a distal and frontal direction. This is correlated with the encrusting mode of life, which does not allow basally directed expansion. EUTHYRISELLA Bassler Euthyrisella Bassler, 1936 : 161 (for Euthyris Hincks preocc.) TYPE SPECIES. Euthyris obtecta Hincks, 1 882. DESCRIPTION. Colony flustriform, branching, multiserial and unilaminar, with basal and marginal extrazooidal coeloms bounded by cuticular walls, which become thickened to form anchoring, rooting structures, with internal calcified plates late in ontogeny. Zooids with depressed, curved, thinly calcified frontal shields, with few pores. Lateral walls converge basally and are attached to the basal cuticular wall by intermittent calcified processes. Dimorphic brooding zooids inflated basally, with enlarged orifices and opercula. Euthyrisella obtecta (Hincks) Euthyris obtecta Hincks, 1882 : 164, pi. 7, fig. 3, 1893 : 177. Harmer, 1902 : 277, pi. 16, figs. 32-37. Levinsen, 1909 : 272, pi. 15, figs 2a-f. Canu & Bassler, 1929 : 464, Figs 199 A-M. Figs 2-7 Scanning electron micrographs of Euthyrisellidae: (2) Euthyrisella obtecta (Hincks) BMNH 1890.3.24.22, Torres Straits, lateral view of zooids showing lateral and oral septulae and basal processes x65; (3) E. obtecta, zooids viewed from distal end, showing distal septulae (arrowed) and frontal processes x!20; (4) Tropidozoum cellariiforme Harmer BMNH 1964.3.10.1, Celebes, frontal view of primary zooids of internode, showing calcified tube at distal end of nodal kenozooid x72; (5) T. burrowsi sp. nov. BMNH 1975.1.4.1, Madagascar, basal view of proximal part of internode, showing autozooids and basal coelomic column, with basal septulae and bars of calcification x30; (6) T. burrowsi, frontal view of autozooids and brooding zooid x53; (7) Neoeuthyris woosteri (MacGillivray) BMNH 1979.2.1.1, Western Australia, frontal view of autozooids and avicularia x30. INTERIOR-WALLED CHEILOSTOME BRYOZOA 61 62 P. L. COOK & P. J. CHIMONIDES SPECIMENS EXAMINED. BMNH. North Australia, 1897.5.1.445 (possibly type material); Broughton Id., New South Wales, 1883.11.29.3; Port Denison, East Queensland, 1883.11.29.43; Murray Id., Torres Straits, 1890.3.24.22, 16-36 metres; 1890.7.23.38, 17-36 metres; 1916.8.23.115, 9-36 metres; Baleine Bank, Northwest Australia, 1892.1.28.34A, 17-22 metres; Holothuria Bank, Northwest Australia, 1892. 1.28.61 A, 22 metres; Long Reef, 5 miles North of Sydney, New South Wales, 1975.1.2.3, 101 metres. DESCRIPTION (Figs 2, 3, 33, 35). Euthyrisella with zooids intercommunicating by 6-8 lateral septulae (8-10 in brooding zooids) and 5-6 distal septulae. Zooids with 12-20 basal septulae (45-50 in brooding zooids) and 2-8 septulae distal to the orifice, which communicate with the hypostegal coelom of the next distal zooid. Frontal shield thinly calcified, with 2-5 pores proximal to the orifice (10-15 in brooding zooids), some of which are septulae communi- cating with the part of the hypostegal coelom basal to the frontal shield. Orifices with small lateral condyles. Dimorphic (brooding) zooids about twice the size of the autozooids. REMARKS. The hypostegal coelom beneath the frontal shield is not nearly as extensive as in Pleurotoichus and Tropidozoum (see below), but forms a small expansion proximally to the orifice (Fig. 35). The dimorphic brooding zooids occur in a proportion of approximately 1 : 20 autozooids. Large embryos are present in one colony from Torres Straits (1890.7.23.38). The colonies are profusely branched and reach a height of 1 50 mm. The branches are 2-3 mm wide and are usually formed by simple bifurcation, but some, astogenetically late, lateral branches occur. Before a bifurcation the width increases to 4 mm, and the number of zooids in a transverse series from 3-4 to 5-6. There is a tendency for one branch of a bifur- cation to include more zooids than its pair. At irregular intervals there is an increase in the width which does not result in a bifurcation, but a sudden decrease in zooid numbers between successive generations. This seems to be associated with a period of cessation of growth, as the basal cuticle is much thicker on the proximal side of the interruption. The early astogeny has been inferred from one colony (1883.11.29.3). A single, erect kenozooid is followed by a distal zooid which itself buds one distal and paired distal-lateral zooids. Thereafter, increase in zooid rows originates from bifurcations of series distal to marginal zooids. At the base of fronds zooids have the series: Lz 0'50, 0'50, 0'67, 0'50, 0'53, 0'58, 0'60 mm. The only other astogenetic changes are confined to brooding zooids which do not appear until the branch is 100 mm in length (300 zooid generations). Brooding zooids are central and often paired in the same transverse series. The extrazooidal marginal coelomic systems are extensive and resemble those formed by lateral kenozooids in several anascan, flustriform genera and the ascophoran species Euthyroides episcopalis and Onchoporella buskii. The changes during colony growth may be regarded as ontogenetic, in that they occur within the bounding walls of an existing structure. The primary changes are found 40 zooid generations back from the growing tip. The lateral cuticle expands intussusceptively and thickens considerably. About 1 50 zooid generations back, calcareous spicules derived from the lateral walls of the marginal zooids expand and form a series of plates within the lateral coeloms. Subsequent series of plates, alternating with cuticle, are formed at the base of the colony, and the investing cuticular wall continues to expand and thicken (see Harmer 1902). At the earliest astogenetic level there are 9-12 series of thickened plates on each side of the branch. The growth of anchoring rootlets is also an ontogenetic rather than an astogenetic change. The marginal elements extend basally and divide into rhizoids which penetrate the substratum, incorporating particles of sediment within a meshwork. Some calcification occurs so that the rhizoids are not only flexible but adherent. It is possible that further extensions are astogenetic and consist of kenozooids budded from the extrazooidal coeloms. It is interesting that an almost identical series of calcareous plates occurs in the kenozooidal, lateral rooting systems of O. buskii. The rooting system of the colony (1883.11.29.3) has incorporated the substratum of shell fragments, sand grains, foraminiferans and dead bryozoans. Although these last are worn, and may have been transported, it is interesting that they comprise INTERIOR-WALLED CHEILOSTOME BRYOZOA 63 Figs 8 and 9 (8) Pleurotoichus clathratus (Harmer) BMNH 1976.1.2.1, Heron Island, sketch of frontal surface of autozooids and brooding zooids, showing opercula and irregular bars of calcification of frontal shield beneath the cuticular frontal wall; (9) Didvmosella larvalis (MacGillivray) BMNH 1890.4.17. l-10pt., Spencer Gulf, sketch of ancestrula. Scale bars = 0-50 mm. rooted 'sand fauna' species (see Cook, 1965, 1966, \919b). They include colonies of Conescharellina sp., the internodes of jointed species of cyclostomes and of the cheilostome anascan, Exostesia didomatia (see Brown, 1948 : 111). Other epizoic forms include a ramify- ing hydroid and numerous egg sacs of some unknown animal. PLEUROTOICHUS Levinsen Pleurotoichus Levinsen, 1909 : 270. TYPE SPECIES. Euthyris clathrata Harmer, 1902. DESCRIPTION. Colony flustriform, branching, multiserial, unilaminar, with marginal and basal extrazooidal coeloms. Zooids with flat, depressed frontal calcified shields consisting of irregular bars and foramina, each bar surrounded by cuticle. Lateral walls curved and converging, basal walls linear, reaching the extrazooidal basal wall in discontinuous series. Dimorphic zooids slightly larger than autozooids, with larger opercula, inflated basally. Lateral coeloms slightly calcified late in ontogeny. Pleurotoichus clathratus (Harmer) Euthvris clathrata Harmer, 1902 : 266, pi. 16, figs 18-31. Pleurotoichus clathratus : Levinsen, 1909 : 270. Canu & Bassler, 1929 : 464, Figs 200 A-N. SPECIMENS EXAMINED. BMNH. Australia, 1899.7.1.2606; Port Jackson, New South Wales, 1903.5.1.2, sublittoral; Swan River, South Australia, 1935.10.18.1; Middle Harbour, Port Jackson, 1975.8.1.7; Watsons Bay, 1975.8.1.8.9, under rock ledges, low tide; Heron Id., Queensland, 1976.1.2.1. DESCRIPTION (Figs 8, 34). Pleurotoichus with zooids communicating by 7-9 lateral and 7-10 distal septulae. Zooids with 12-15 basal septulae, and 3-5 distal to the orifice, 64 P. L. COOK & P. J. CHIMONIDES communicating with the hypostegal coelom of the next distal zooid. The frontal shield is formed by irregular cuticle-covered bars, which anastomose, leaving large slit-like foramina. The hypostegal coelom is extensive both above and below the shield. The dimorphic brooding zooids are only slightly larger than the autozooids; they are inflated basally, and occur in larger numbers than in E. obtecta, in a proportion of 1 : 5 autozooids. REMARKS. The presence of cuticle surrounding the bars of calcification of the frontal shield, and the layers of cuticle on either side of the lateral and distal calcified walls, together with the general delicacy of the interior structures, makes it almost impossible to prepare calcified skeletons of P. clathratus for scanning electron microscopy. Investigation of specimens has therefore been almost entirely confined to examination of thin-sections. The hypostegal coelom, both frontal and basal to the calcified shield is very extensive. The convergent lateral walls reach the basal cuticular extrazooidal wall in one short, linear insertion. The calcified processes are cuticular terminally at the point of contact. Zooids are thus almost suspended from the circum-orificial region in a bag of tough, flexible cuticle. This flexibility is probably correlated with the turbulence of the very shallow waters in which the species lives. The growth form is analogous to that of Chartella papyracea (see Ryland & Hay ward, 1977 : 80), which also inhabits rocky ledges in the sublittoral region. Colonies may reach a height of 185 mm (Harmer, 1902 : 268). The specimens examined have a maximum height of 50 mm, and have very few bifurcations. A few lateral sub- branches occur late in astogeny. The earliest stages are missing, but may be inferred from a colony from Port Jackson (1975.8.1.7), which has 4 zooids in a transverse row at its base. Extrapolating from this, the early astogeny is probably similar to that of E. obtecta, and the zooids show a similar series of increase in size. Interruptions of growth, involving a sudden decrease in the number of zooids in transverse series from 6-7 to 3-4, occur frequently, between 6 and 1 0 zooid generations apart. Dimorphic zooids are frequent (Fig. 2), and are first found 25 generations from the base of the colony. They are placed in the central zooid series, and large embryos are present in the colony from Heron Island (1976. 1.2.1.). The marginal coeloms become inflated with ontogeny, and the cuticular walls become thickened. Small calcified processes develop from the lateral walls of adjacent zooid series, but calcified plates do not occur. The colony from Heron Island has a large number of epizoic forms, particularly colonies of hydroids. On the basal cuticle there are small colonies of jointed cyclostome bryozoans and ofCellaria sp. The lower part of the colony is completely invested by a colonial ascidian. TROPIDOZOUM Harmer Tropidozoum Harmer, 1957 : 1 106. Cook, 1975. TYPE SPECIES. T.cellariiforme Harmer, 1957. DESCRIPTION. Colony cellariiform, internodes connected by complex nodes of partially calcified kenozooids and investing extrazooidal basal coelom. Basal coelomic channel confluent among internodes. Central zooids of internodes with extended basal calcification forming a keel. Zooids with depressed frontal shields, with large foramina. Hypostegal coelom extensive, frontal septulae absent. Brooding zooids large, inflated basally, with dimorphic orifices. Tropidozoum cellariiforme Harmer Tropidozoum cellariiforme Harmer, 1957:1106, pi. 67, figs 15-17, 20. Cook, 1975:162, pi. 2, pi. 3, Figs 2A, 3 (explanation of plates 1 and 3 transposed). Cook, 1979a : 200, pi. l,fig. 1. SPECIMENS EXAMINED. BMNH. Siboga Stn 144, South of Halmaheira, Djilolo, East Celebes, 0-45 metres, coral and lithothamnion bottom, 1964.3.10.1 and 1979.1.4.1. ZMA, as above, TYPE and other material. INTERIOR-WALLED CHEILOSTOME BRYOZOA 65 DESCRIPTION (Figs 4, 1 1, 30). Tropidozoum with internodes 8 mm long and l-6 mm wide, comprising 70 zooids. Longest, usually the central series, of 12 zooids; widest, usually the penultimate transverse series, of 6 zooids. Zooids occupying an arc of 240° in section. The proximal zooid of each internode is wider (Lz 0'37 mm), and the ultimate zooids shorter and narrower (Lz 0'50, Iz 0'30 mm), than other zooids. Orifice rounded distally with large complex condyles delimiting a proximal part which is straight or curved distally. Frontal shield with up to 30 large, irregular foramina. Hypostegal coelom extending beneath the frontal shield. Communication between the visceral coelom of one zooid and the hypostegal coelom of the next distal zooid inferred to be through a pair of minute pores passing through the distal-lateral calcification of the orifice (Fig. 30). Communication with the basal coelom through numerous pores in the confluent lateral and basal walls. Brooding zooids dimorphic, large, inflated basally, placed in the central zooid series, occurring in the proportion of 1 : 10 autozoids. Brooding zooid orifices larger, but similar to those of autozoids. REMARKS. The earliest astogenetic stages are missing from the colonies examined, but it is inferred that they are similar to those of T. burrowsi (see below). The cuticle of the basal coelom becomes increasingly thickened and inflated with ontogeny, and eventually forms a supportive and rooting structure. The septulae connecting the visceral and hypostegal coeloms of sequential zooids are very small, and apparently pass through the thick calcification distal and lateral to the orifice. The tubular communication traces a curved path, at an angle to the longitudinal axes of the zooids, and is impossible to expose in its entirety in either longitudinal-lateral or in tangential thin-section. The pattern of other zooidal and extrazooidal communications has been described by Cook (1975). A few, small embryos have been seen in the brooding zooids. The distal part of the orifice of these zooids appears to be formed by the proximal calcification of the next distal zooid (see Fig. 30). Epizoic forms include other erect, jointed bryozoans such as Cellaria sp., small colonies of a Bowerbankia-like ctenostome, encrusting colonies of Robertsonidra argentea (Hincks), and foraminiferans belonging to the informal 'HomotremcC group. Tropidozoum burrowsi sp. nov. Tropidozoum sp. Cook, 1975 : 165, pi. 1, figs A-C, Fig. 2B (explanation of Plates 1 and 3 transposed). Hayward & Cook (in press). SPECIMENS EXAMINED. BMNH. HOLOTYPE, Tulear, TU 24, Madagascar, exterior Grande Vasque, 23.9.69, 15 metres, 1975.1.4.2. Paratypes, Tulear, Pichon D25, 1975.1.4.1; other material: South of Durban, Umtwalume River mouth North by West 7 miles, 12271, shell fragment bottom, 90 metres, 1949.11.10.184, and 12272, 1949.11.10.250, trans- ported, dead internodes. SAM. 'Meiring Naude\ Southeast Africa, Stn 131, 780 metres, Stn 151, 900 metres, transported, dead internodes. ETYMOLOGY. Named after Mr H. W. Burrows, whose manuscript notes and drawings on the Gilchrist Collection from South Africa are stored at the BMNH. DESCRIPTION (Figs 5, 6, 10). Tropidozoum with internodes 7 mm long and M mm wide, comprising 70 zooids. Longest longitudinal series of 16 zooids; widest, penultimate trans- verse series of 6 zooids. Zooids occupying an arc of 300° in section. Proximal and ultimate zooids wider than other zooids. Central basal keel thickened by base of calcification during ontogeny. Orifices rounded distally, with large condyles delimiting a deep, triangular sinus. Frontal shield with 30 small, irregular foramina, occluded during ontogeny. Dimorphic brooding zooids inflated basally, larger than autozooids, with wide, non-sinuate orifices. REMARKS. The early stages in astogeny of erect branches arise from a confused mass of P. L. COOK & P. J. CHIMONIDES INTERIOR-WALLED CHEILOSTOME BRYOZOA 67 rooting-kenozooids, and it is not known if these represent one, or a large number of colonies. Some kenozooids are rootlets, others are shorter, and from these small erect internodes originate. The maximum number of internodes in an erect branch is 1 5, reaching a height of 130 mm. The primary internodes are nearly all small, including only 3 zooids, and are rounded in shape. Later internodes become progressively more elongated and include more zooids; a typical sequence has 3, 6, 25, and 36 zooids. Occasionally, a small internode occurs later in astogeny, between two longer internodes. Zooid length at the base of an internode forms a series Lz 0*55, 0'47, 0*60, 055, 062, 0'58, 0'60 mm (cf. E. obtecta). Brooding zooids appear by the fifth internode and are placed in the centre of a transverse zooidal series; a maximum of 10 occur in one internode. T. burrowsi is known from direct observation (Dr J.-G. Harmelin, pers. comm. 1974), to be a shallow-shelf species, and all the deep water records given above are of worn, dead internodes found in sediments which had been transported. The colonies from Madagascar were observed alive in situ; they have several epizoic and epiphytic forms, including sponges, small gorgonians, hydroids, serpulids, the bryozoan Cellaria sp. and small algae. The coelomic pigment of living colonies was bright red, and they were found growing together with gorgonians etc., and forming patches 20^40 cm across on calcareous accretions covered by sand. NEOEUTHYRIS Bretnall Neoeuthyris Bretnall, 1921 : 157. TYPE SPECIES. Euthyris woosteri MacGillivray, 1 89 1 . DESCRIPTION. Colonies encrusting, basal walls of zooids uncalcified except at the margins. Frontal shields curved, hyaline. Lateral walls prominent. Orifices with a wide sinus and paired condyles. Avicularia unilateral and oral, mandible hinged on a complete bar. Brooding zooids with large, dimorphic orifices and a brood chamber formed by a large endozooidal ovicell, closed by the operculum. REMARKS. The characters of N. woosteri, the only species, differ from those of other genera of Euthyrisellidae in the encrusting habit, lack of basal coelom, and presence of ovicells and avicularia. The essential structure of the colony, the nature of the frontal shield, the interzooidal communications and the dimorphism of brooding zooids, are, however, all similar to those found in the other genera. Neoeuthyris woosteri (MacGillivray) Euthyris woosteri MacGillivray, 1891 : 77, pi. 9, fig. 2. Neoeuthyris woosteri: Bretnall, 1921 : 158, Fig. 1. Hastings, 1960; 1964 : 245, figs 1,2. SPECIMENS EXAMINED. BMNH. Fremantle, Western Australia, 1948.3.12.1; Western Australia, 1938.8.10.1; Cosy Corner, Torbay Head, Western Australia, 1 979.2. 1.1. AM and NMV: U875 HOLOTYPE (see Hastings, 1964 : 246). DESCRIPTION (Figs 7, 12, 28, 29, 31, 32). Neoeuthyris encrusting algae (usually Metamastophora flabellatd), zooid basal walls uncalcified. Zooids communicating by 3-5 lateral and 2-3 distal septulae placed at the base of the vertical walls. Visceral-to- hypostegal communication by paired groups of 4-5 septulae in the upper part of the distal Figs 10-13 Scanning electron micrographs of brooding zooids of Euthyrisellidae and Didymosellidae: (10) Tropidozoum burrowsi sp. nov. BMNH 1975.1.4.1, Madagascar x!03; (11) T. cellariiforme Harmer BMNH 1964.3.10.1, Celebes x!27; (12) Neoeuthvris woosteri (MacGillivray) BMNH 1979.2.1.1, Western Australia x50; (13) Didymosella larvalis (MacGillivray) BMNH 1979.7.5.2, near Melbourne, showing partially formed ovicells at growing edge x3 5. 68 P. L. COOK & P. J. CHIMONIDES wall behind the orifice (Fig. 32). Frontal shield smooth, very thinly calcified, with small protuberances, but no suboral septulae, and with no extension of hypostegal coelom basally. Two pairs of frontal septulae beside the orifice, one pair giving rise to an avicularium (Fig. 28). Avicularia with acute mandibles, directed proximally. Brooding zooids large, with dimorphic orifices, ovicell apparently formed by an expansion of the lateral and distal walls, protruding into the cavity of the next zooid, which is very short (Lz 0*45 mm). REMARKS. Avicularia on the right hand side of a lobe tend to occur on the right hand side of the zooids (viewed frontally), and vice versa. The algal substratum is influenced during growth by the presence of the bryozoan (see Hastings, 1964). The extreme fragility of the colonies makes investigation very difficult. Scanning electron microscopy of small fragments, and thin sections, have enabled us to reconstruct the interzooidal and intrazooidal communication systems, which are hardly visible in whole preparations using a light microscope. N. woosteri has rarely been reported, and the ancestrula and early astogeny is unknown. Most specimens are known to have encrusted M.flabellata which had been washed up on the shore, and both species probably occur together in deeper water. Nearly all records are from Western Australia, but MacGillivray (1891) reported specimens from Cooktown, Queensland. Table 2 Measurements (mm) of autozooids and avicularia of Didymosellidae Lz Iz 12o (av.) Lr Lm No. of frontal pores Didymosella sp. 0-33-0-50 0-25-0-37 0-15 0-15-0-23 4 (Labracherie) D. irregularis 0-58-0-73 0-34-0-37 0-22 0-14-0-22 10 D. acutirostris 0-50-0-60 0-50-0-60 0-16 - _ D. crassa 0-68-1-00 0-69-0-90 0-27 0-22-0-27 30 D. porosa 0-50-0-75 0-33-0-44 0-15 0-08-0-13 10 D. clypeata 0-88-0-96 0-46-0-57 0-23 0-12-0-15 15 D. 'larvalis'- fossil 0-50-0-60 0-45-0-48 0-16 0-20-0-30 9 (New Zealand) D. 'larvalis '-fossil 0-60-0-65 0-48-0-50 0-19 0-20-0-30 10 (Australia) D. larvalis- Recent 0-55-0-70 0-40-0-50 0-18 0-20-0-30 0-15-0-25 10 D. parviporosa 0-55-0-75 0-35-0-65 0-16 0-13-0-20 20 D. pluma 0-65-0-80 0-64-0-90 0-22 0-28-0-32 0-70-0-80 20 T. magnirostris 0-80-1-00 0-60-0-80 0-15 0-25-0-28 0-23-0-30 20 T. magna 0-65-0-70 0-65-0-70 0-27 0-30-0-35 5 T. boninensis 0-60-0-90 0-50-0-70 0-20 0-15-0-20 0-50-0-75 15 DIDYMOSELLIDAE Brown Didymosellidae Brown, 1952 : 194. TYPE GENUS. Didymosella Canu & Bassler, 1917. DESCRIPTION. Interior-walled cryptocystidean ascophorans, usually with extrazooidal basal coeloms, and considerable basal calcification. Zooid frontal shields with pores and marginal frontal septulae, becoming thickly calcified during ontogeny, obscuring the lateral walls. Peristomes elongated, with or without foramina and/or spiramina frontally. Primary calcified orifice not well defined, operculum a flap-like extension of the distal part of the ascus wall. Septulae in vertical and basal zooid walls numerous. Avicularia adventitious, INTERIOR-WALLED CHEILOSTOME BRYOZOA 69 arising from oral frontal septulae, usually unilateral, with large subrostral chambers and acute rostra. Mandibles orientated laterally, hinged on a complete bar. Brood chambers formed by an expansion of the distal part of the peristome wall, opening above the operculum. Ontogeny of zooid walls As in the Euthyrisellidae, ontogenetic changes in the Didymosellidae are apparently rapid. Few colonies show complete series of frontal shield development, and partially uncalcified buds are contiguous proximally with fully formed zooids in which secondary calcification is well advanced (see Fig. 21). For a brief period, the prominent lateral walls and depressed crypyocystal shield resemble those of Pleurotoichus or Tropidozoum. The presence of numerous frontal marginal septulae, however, means that the frontal calcification soon thickens, obscuring all zooidal boundaries. Funnels in the calcification common to tubular extensions above septulae of adjacent zooids are developed, as in the umbonuloid genus Tremogasterina (see Cook, 1977). The calcification of the primary orifice merges into the development of the peristome and no clear-out boundary is obvious. Avicularia are formed as lateral-oral frontal buds, their subrostral chambers deriving from a flattened plate of the frontal shield, bordered by 3-8 marginal septulae (see Fig. 20). In D. pluma particularly, funnels derived from these septulae, and frontal pores, remain prominent features of the large avicularian chambers quite late in ontogeny (see Fig. 2 1 ). Coelomic systems and communication The presence of frontal septulae is correlated both with the absence of interzooidal visceral-to hypostegal communication (Fig. 1 B), and with the massive calcification of the frontal shields. In nearly all species, basal septulae, which communicate with a thin extrazooidal coelom, have been observed. Basal calcification, too, is considerable, and forms processes or massive pillars, which contact the substratum, allowing a semi-repent mode of life, or strengthening erect branches. The nature of the porous basal calcification has long been noted by authors (Stoliczka, 1865, MacGillivray, 1895, Waters, 1885, Levinsen, 1909, and Brown, 1952), who 'have either compared it to that of Selenaria (which also has an extrazooidal, basal coelom, see Chimonides & Cook, 1981), or have suggested a radicular function for the pores. Opercula, mandibles and peristomes In correlation with the elongated peristomes (see Harmer, 1957 : 652), and in contrast to the Euthyrisellidae, opercula are hardly differentiated, and are flap-like expansions of the ascus wall. Mandibles are acute, and hooked in D. larvalis and T. magnirostris. Those of D. pluma and T. boninensis are elongated and resemble the mandibles of the anascan genus Smittipora (see Harmer, 1926), in having a central rachis, and expanded wings of cuticle. Brooding zooids In some species, brooding zooids and their orifices are larger than those of autozooids, but there seems little evidence of dimorphism in other characters. In some species, the brooding zooids may have smaller avicularia, or none, and there is evidence of patterning of the avicularia of surrounding zooids. Unlike the Euthyrisellidae, the family Didymosellidae has an extensive fossil record (see below). 70 P. L. COOK & P. J. CHIMONIDES INTERIOR-WALLED CHEILOSTOME BRYOZOA 71 DWYMOSELLA Canu and Bassler Didymosella Canu & Bassler, 19 1 7 : 43. TYPE SPECIES. Lepralia larvalis MacGillivray, 1869. DESCRIPTION. Zooidal peristomes with large, paired foramina, spiramina absent. REMARKS. Didymosella was included in the family Escharellidae by Canu & Bassler (1929 : 33), and in the Exochellidae by Bassler (1953 : G205). The distinctive characters of the genus were discussed by Brown (1952 : 194), who noted that the Italian fossil species Porina bioculata Waters (1891 : 26, pi. 3, fig. 15), the type species of Bimicroporella Canu (1904 : 12), was not closely related to Didymosella. B. bioculata (Waters), B. ventricosa Canu (1904) and B. watersi Brown (1958) all have some superficial similarity with Didymosella, but have no suggestion of a basal coelom, and have hyperstomial ovicells. Didymosella has an extensive fossil record. The earliest known species is an undescribed form listed by Labracherie (1972 : 44), from the Middle Eocene of North Aquitain, France. Specimens of this species have been examined (Bazas-rural, 200 metres, Labracherie Coll.). The material is fragmentary and comprises only 30 zooids. Most of the fragments appear to have been encrusting, but some have basal grooves and a few have minute pores, which together with the thick, basal calcification are interpreted to indicate the former presence of a basal coelom. The frontal shields have 3-6 pores and 7-9 funnels indicating frontal septulae. Zooids communicate through 1-2 distal, and 2-3 lateral septulae. The paired peristomial foramina are distinct, as are the avicularian rostra, some of which have a complete bar. The rostra are open terminally which, by analogy with D. pluma (see p. 76) probably indicates that the mandibles were elongated. Another fossil species, D. irregularis, was described from the Upper Eocene (Lower Jacksonian) of north-western Florida by Cheetham 1963:66, pi. 2, fig. 13). This had unilaminar colonies and larger zooids with up to 20 frontal pores and very small, acute avicularia. The orifices of the brooding zooids; which had broken peristomes exposing the ovicells, were slightly dimorphic. In view of Labracherie's (1972) records, it appears that D. acutirostris Faura y Sans & Canu (1917 : 90, pi. 7, figs 3-5), from the Upper Eocene of Monresa, Barcelona, is also attributable to Didymosella. The illustration is not conclusive, but the species was described with paired, peristomial foramina and peristomial ovicells. The zooids were larger than those of Labracherie's species. The wide Eocene distribution of Didymosella continued during the Oligocene and Miocene (see below), and is maintained in Recent seas. Didymosella crassa Canu and Bassler Didymosella crassa Canu & Bassler, 1920 : 1 16, pi. 88, figs 1-7. SPECIMENS EXAMINED. BMNH. Oligocene (Vicksburgian), west bank of Conecuh River, Escambia Co., Alabama. Canu & Bassler Coll. Figs 14-20 Scanning electron micrographs of fossil and Recent Didymosellidae: (14) Didymosella clypeata Canu & Bassler BMNH Lower Miocene, Victoria, basal side of branch, showing septulae x31; (15) D. clypeata, frontal side of branch x34; (16) D. porosa (Stoliczka) BMNH D 36784-92, Middle Oligocene, New Zealand, part of frontal margin of branch x21; (17) D. 'larvalis' BMNH D36783-7, Miocene, New Zealand, x40; (18) Tubiporella boninensis Borg BMNH 1889.8.21.105, Recent, China Sea, lateral view of young colony, ancestrula at right, showing basal extrazooidal calcification x25; (19) T. boninensis zooids at growing edge, showing spiramina, frontal pores and avicularian rostra x57; (20) T. magnirostris (MacGillivray) BMNH 1899.7.1.2686, Recent, Victoria, zooids viewed from distal end, showing spiramen (arrowed) and base of avicularian subrostral chamber (bottom right) x57. 72 P. L. COOK & P. J. CHIMONIDES DESCRIPTION (Fig. 23). Didymosella with large zooids and tubular peristomes. Zooids with 25-35 frontal pores and 6-8 marginal septulae. Avicularia large, with channelled rostra open terminally. Basal surface thickened, with grooves and pores. REMARKS. In all respects D. crassa and D. pluma (see p. 76) are so alike that it is possible to postulate a direct genetic link between the Oligocene and Recent populations from the central West Atlantic. The skeletal characters differ from those of D. pluma only in the larger number of frontal pores and in the absence of elongated calcified basal processes, which last may be due to wear during preservation. The open channel at the terminal end of the avicularian rostrum allows the inference that the mandible was elongated, and may have even been expanded laterally like that of/), pluma (see p. 78). Didymosella porosa (Stoliczka) Semiescharipora porosa Stoliczka, 1865 : 128, pi. 19, figs 10-13. Didymosella porosa : Canu & Bassler, 1929 : 328. Didymosella aft. porosa : Brown, 1952 : 198, Figs 139, 140. SPECIMENS EXAMINED. BMNH. Middle Oligocene, Pareora, Tarahoke Quarry, N.W. tip of South Island, New Zealand, D36788-92. DESCRIPTION (Figs 16, 27). Didymosella with erect, ligulate branches, with 4-5 transverse zooid series. Branches flat and bifurcated. Lateral marginal zooids curved outward, their peristomes forming serrations. Secondary orifices with paired spine bases. Marginal avicularia large. Basal calcification with numerous pores. REMARKS. Branches are composed of 20 sequentially distal zooid series. The marginal zooids are curved so that their orifices are at an angle of over 45° to those of the central series. At the same time, the orifices are curved towards the frontal side of the branch. The paired peristomial foramina are separated from the proximal edge of the peristome by a distance greater than 0*20 mm. Stoliczka (1865) mentioned the thickening of the basal calcification and the porous surface, which was also noted by Brown (1952), who suggested that the pores might have a radicular function. In spite of Brown's doubtful assignment of his material to D. porosa, it is so close in general characters, age and locality to Stoliczka's specimens (which were from the Lower Miocene of Orakei (Hauraki) Bay, North Island), that it may be confidently included in the species. Didymosella clypeata Canu & Bassler Didymosella clypeata Canu & Bassler, 1935 : 32, pi. 9, figs 7, 8. SPECIMENS EXAMINED. BMNH. Lower Miocene, Mount Martha, Victoria. Anticline Creek (Janjukian), Dartmoor, Victoria D34200-01. DESCRIPTION (Figs 14, 15). Didymosella with erect ligulate branches, with 2-3 transverse series of zooids. Branches curved towards the basal side so that orifices of the marginal series are at an angle of 70° to those of the central series. The marginal zooids also curve outwards slightly at about 1 5° to the central axis. Secondary orifices with paired spine bases. Frontal pores are stellate (in basal view) and the paired peristomial foramina are close to the proximal edge of the peristome, at a distance of 0- 10-0- 1 5 mm. Basal calcification thickened, with a few irregularly spaced pores. REMARKS. The marginal avicularia are large, and occur about halfway up the zooids. Although very similar to D. porosa, D. clypeata differs in the number of zooids in a transverse series, the curvature of the branches and the placing of the peristomial foramina. INTERIOR-WALLED CHEILOSTOME BRYOZOA 73 The combination Porina clypeata Waters (1881 : 332, pi. 17, fig. 67, 1882a : 268) was used for an entirely different species from the Miocene of Mount Gambier, Victoria. The Didymosella /ai*va/is-complex The Recent species D. larvalis is here confined to algal-associated, encrusting colonies, with no evidence of basal coeloms. Two fossil populations, one from the Australian Miocene, the other from the Oligocene to Miocene of New Zealand, have been ascribed by several authors to D. larvalis. Neither of these populations is assignable to D. clypeata or D. porosa, and they differ considerably from Recent D. larvalis in their colony forms. A. Didymosella aff. larvalis (fossil specimens) Porina larvalis: MacGillivray, 1895:104, pi. 14, fig. 26, Waters, 1882a:269, pi. 8, fig. 19; 1882b: 509; 1887: 189, pi. 6, fig. 8. Didymosella larvalis: Brown, 1952, part: 195, Figs 135-138; 1958 : 56. SPECIMENS EXAMINED. BMNH. Oligocene, Tarakohe Quarry, Waitapu, New Zealand, D36787; Miocene, Weka Pass, New Zealand, D36783-6, D36938, D36961-2; Miocene Muddy Creek, Victoria, Australia, D34255; Mount Gambier, Victoria, D32981-3 (Waters Coll.). MM. Bairnsdale, Victoria, T42 (figured Waters, 1882a, pi. 8, fig. 19). DESCRIPTION (Fig. 17). Didymosella with erect colonies with bilaminar expansions or cylindrical branches. Peristome slightly raised, thickened proximally forming a transverse ridge which overhangs the paired, peristomial foramina. Frontal pores stellate. REMARKS. The bilaminar colonies (from Weka Pass, New Zealand), have basal walls forming a double layer. No pores connecting the layers can be seen. The erect fragments also have complete basal walls, and have a central cavity. One fragment (D36786) has a very long avicularium with a channelled rostrum which is open terminally. In general, the specimens are poorly preserved and worn. The calcification is, however, considerably thicker than that of Recent specimens. The stellate frontal pores can be seen from the inner surface of broken zooids. B. Didymosella larvalis (MacGillivray) Lepralia larvalis MacGillivray, 1869 : 134. 1879 : 30, pi. 37, fig. 5. Escharoides larvalis: Levinsen, 1909 : 318. Livingstone, 1926 : 170, pi. 11. Didymosella larvalis Brown, 1952, part: 195. SPECIMENS EXAMINED. BMNH. Spencer Gulf, 1890. 4. 17.1 Opt. Australia, 1897.5.1.883, 1963.4.18.16; Port Phillip Heads, 1897.5.1.881.882; Adelaide, 1899.7.1.5036A; Cowes Phillipps Id., 20 miles S.E. Melbourne, beach sand, 1979.7.5.2; No locality, 1963.2.12.63. DESCRIPTION (Figs 9, 13, 36). Didymosella with colonies encrusting algae, sometimes rising into small, unilaminar expansions. Zooids communicating by 2 distal and 2-3 lateral multiporous septulae, surrounded by chambers of calcification. Basal wall uncalcified in encrusting zooids, peripherally calcified, or with a large, uncalcified, cuticle covered window in erect zooids. Proximal side of peristome elongated, calcification thin and hyaline. Distal side of peristome deficient. Lateral peristome with spinous processes. Frontal pores stellate, frontal septulae numerous. Avicularia large, with acute mandibles, hooked terminally. Ancestrula with a complex frontal shield of spinous processes. REMARKS. The thin frontal calcification and lack of basal calcification are similar to that of 74 P. L. COOK & P. J. CHIMONIDES INTERIOR-WALLED CHEILOSTOME BRYOZOA 75 the algal-associated Neoeuthyris woosteri (see p. 67), and are also usually present in numerous other cheilostome species which are specific to this type of substratum. In erect zooids, the uncalcified basal area is progressively decreased with astogeny, eventually becoming confined to a window in the basal wall. No evidence of a basal coelom or of radicular expansions has been seen. Specimens are small, but a few parts of colonies (1897.5.1.882) comprising more than 100 zooids, on narrow fronds of algae, show a distinct central zone of brooding zooids which have minute, paired avicularia. They are flanked by lateral series of autozooids in which most of the avicularia are placed with the closed mandible directed medially (cf. T. magnirostris, p. 77). In other colonies, there is no obvious patterning of brooding zooids or of avicularia. Ancestrulae (1890.4. 17. 1. -10 pt., 1963.2.12.63) are large (La 0'52 mm, la 0'40 mm), and distinctly 'cribrimorph' in character. There is a well developed gymnocyst and large, oval opesia, overarched by 4 robust lateral spines. The oral shield is formed by a pair of large, spinous processes, which are fused centrally, leaving an open foramen. The shield terminates distally in 5 spines, and a further, laterally placed pair of bifurcated spines is closely apposed to the lateral walls of the ancestrula and the distal-lateral primary zooids of the post- ancestular triad (see Fig. 9). The systematic significance, if any, of 'cribrimorph' ancestrulae in ascophoran species is unknown. The ancestrulae of other species of Didymosella have not been found, and that of Tubiporella boninensis (see below) is not 'cribrimorph', but resembles subsequent zooids. Most cribrimorph species have anasciform, 'tata' ancestrulae, but 'cribrimorph' ancestrulae are known in the umbonuloid ascophorans Triporula bidenticulata (Canu & Bassler), see Cook (1967:341) and Romancheina asymmetrica Moyano (1975 : 63, Figs 1-8). Temachia opulenta Jullien (1883 : 509, pi. 14, figs 26-29) also has a 'cribrimorph' ancestrula and is so similar to R. asymmetrica that it too may be postulated to have umbonuloid ontogeny. Apart from D. larvalis, the only cryptocystidean species with 'cribrimorph' ancestrulae known at present is Schizoporella kiiensis (Okada & Mawatari), see Mawatari (1952 : 279, fig. 13). Didymosella parviporosa Canu & Bassler Didymosella parviporosa Canu & Bassler, 1929 : 327, pi. 39, fig. 1. Didymosella costulata Canu & Bassler, 1929 : 327, pi. 39, fig. 2. SPECIMENS EXAMINED. BMNH. Philippines, Albatross Stn D5 145, 42 metres, 1931.12.30.119 (D. parviporosa}; Stn D55141, 53 metres, 1931.12.30.120. (D. costulata}. DESCRIPTION. Didymosella with small, narrow zooids, each with an average of 20 frontal pores. Secondary thickening forming interzooidal funnels, with bars of calcification. Basal calcified wall with furrows and a few scattered pores. Avicularia, small, with acute rostra. REMARKS. The specimens include one fragment of D. costulata and three of D. parviporosa, together comprising only 40 zooids. Both species were described from the same region and D. costulata appears to be a slightly worn, ontogenetically thickened stage of development of D. parviporosa. This conclusion is strengthened by the locality data of the BMNH specimens, which were labelled by Bassler. One, labelled D. parviporosa, is from Stn 5145, the original locality for D. costulata; the other labelled D. costulata, is from Stn 5141, one of the original localities for D. parviporosa. Figs 21-24 Scanning electron micrographs of fossil and Recent Didymosellidae: (21) Didymosella pluma sp. nov. BMNH 1899.7.1.5324, Recent, off Brazil, zooids at growing edge, showing raised lateral walls, frontal pores and channelled avicularian rostra x71 ; (22) D. pluma, lateral view of zooids showing lateral septulae and basal calcified processes x62; (23) D. crassa Canu & Bassler BMNH Oligocene, Alabama, showing channelled avicularian rostra x59; (24) Tubiporella magnirostris (MacGillivray) BMNH 1899.7.1.2686, Recent, Victoria, growing edge viewed from distal end of zooids, showing distal and basal septulae and basal calcified processes x7 5. 76 P. L. COOK & P. J. CHIMONIDES D. parviporosa differs from D. larvalis in the presence of a basal coelom, and its small, narrow zooids. It differs from D. pluma in its much smaller zooid size, and in its avicularian mandibles, which may be inferred to have been acute. Didymosella pluma sp. nov. Didymosella sp. Brown, 1952 : 198. Cook, 1981 (in press). SPECIMENS EXAMINED. BMNH. HOLOTYPE, John Adams Bank (=Victoria Bank, off Brazil, Approx. 21°S, 37°W, depth less than 183 metres), 1899.7.1.5322. Paratypes, 1899.7.1.5323, 5324,4476,4477, 176; Havana, 1911.10.1.1702, 146 metres. ETYMOLOGY. Pluma (L) — a feather, referring to the expanded avicularian mandible. The name was given as a manuscript label by Busk. DESCRIPTION (Figs 21, 22, 25). Didymosella with semi-repent colonies forming irregular expansions. Frontal shield with 15-23 pores and 6-8 marginal septulae, the distal pair enlarged. Zooids communicating by 2-4 distal and lateral septulae. Peristomes raised, tubular, with paired, marginal spinous processes. Basal calcified wall becoming extrazooidal with septulae and long, calcified extensions. Avicularian subrostral chambers very large, rostra with a channel open terminally. Mandible very long, expanded and fimbriated laterally. REMARKS. The thickening of the frontal calcification above and around the distal pair of frontal septulae forms large, slit like foramina, one either side of the orifice. Two large lobes of about 300 zooids each, show areas in which groups of zooids with 'left-handed' avicularia are opposed to groups with 'right-handed' avicularia. Unlike the colonies of T. magnirostris (see p. 79), these groups are not obviously associated with the presence of brooding zooids. The avicularian rostra have pores which are visible late in ontogeny. The remarkable mandibles resemble those of Smittipora (see Cook, 1964), but are fimbriated marginally. When open, each mandible covers the frontal foramina of the zooid; when closed, it usually covers the foramina of the adjacent, lateral zooid (see Fig. 25). Presumably the mandibles ensure that the foramina do not become blocked by detritus, which would restrict water flow into the ascus, and the laterally directed water currents produced by mandibular movement may assist in clearing deposits from the frontal surface of the zooids (see Cook, 1981). Expanded mandibles are also present in Tubiporella boninensis (see below). D. pluma is so similar to D. crassa that it is highly probable that Miocene-to-Pliocene specimens, linking the two forms may eventually be found from sediments in the Gulf of Mexico region. TUBIPORELLA Levinsen Tubiporella Levinsen, 1909 : 304. TYPE SPECIES. Lepralia magnirostris MacGillivray, 1883. DESCRIPTION. Colonies semi-encrusting, with free expansions. Zooids with long, tubular peristomes and a frontal spiramen. Frontal and basal septulae numerous, and extrazooidal calcification considerable. Avicularia lateral, arising from several frontal septulae, subrostral chambers large. Rostra acute, directed laterally, mandible slung on a complete bar. Brooding zooids with a slightly enlarged orifice. REMARKS. Tubiporella was introduced with no formal description. Levinsen (1909 : 304) listed it as 'n.g.' in a key to the family Tubucellariidae; T. magnirostris was the only species included. Species of Tubucellaria (or Margaretta, see Harmer, 1957 : 824), have an ascopore, not a peristomial spiramen, and have erect, jointed colonies. Tubiporella was assigned to the Tubucellariidae by Bassler (1953 : G210), and to the Porinidae by Harmer INTERIOR-WALLED CHEILOSTOME BRYOZOA 77 (1957 : 844). Finally, Brown (1952 : 194) recognized its close affinities with Didymosella and included it in the family Didymosellidae. Tubiporella, like Didymosella, has an Australian Tertiary record, although this extends only to Miocene deposits. The characters of fossil specimens are very similar to those of Recent T. magnirostris, but there are some doubts in accepting the identity of these forms without examination of type specimens. Brown (1952) noted that Recent colonies had the larger zooidal dimensions (see Table 2), and in view of the degree of difference between Recent and fossil populations of the D. larvalis-complex, it seems advisable at present to regard the Tertiary forms as distinct taxa. Several fossil forms have been described. Waters (1882# : 268, pi. 9, figs 33, 34) introduced Micropella introversa from the Miocene of Mount Gambier, Australia with two drawings which he stated were magnified at x!2 and x25 respectively. The (secondary) orifice measurements he gave on p. 268 were: Lo 0- 16 mm, lo 0'20 mm. Using the magnifications given, the average dimension of zooids for Fig. 33 would be: Lz l-75 mm, Iz I'OO, Lo 0'25 mm, lo 0'33 mm. These are obviously too large; very few cheilostome zooids are more that 1'5 mm in length, and Waters did not mention that the zooids of M. introversa were particularly large. A magnification of x20 would give Lz O90 mm, Iz O60 mm, Lo 0'15 mm, lo 0*20 mm, which is not only in accordance with the given orifice measurements, but is more feasible. The spiramen, which Waters regarded as 'microporellid' ascopore, was described and figured as being very close to the proximal border of the secondary orifice. When describing further fossil specimens from the River Murray Cliffs, Waters (1885 : 296) called this a 'central pore'. In a still later paper, Waters (1887 : 55) included M. introversa in the synonymy of T. magnirostris although the species had originally been described by him a year before MacGillivray's species. Waters (1885:295, pi. 7, fig. 7) also described, as Microporella magna, fossil specimens from Aldinga and Mount Gambier, and identified them with Lunulites magna Tenison Woods (1880), which he stated was 'very closely allied' to T. magnirostris. Tenison Woods (1880 : 7, pi. 1, figs 6 a-d), had described an irregularly domed, hollow, almost circular colony. His figure (6c) shows zooids with raised peristomes, numerous frontal pores and acute, laterally directed avicularia with a complete bar. This species is obviously ascophoran, and is in no way referable to the free-living, anascan genus Lunulites. Two specimens in the British Museum Collections confirm the existence of this species in the Australian Tertiary. The first (labelled 'Lunulites magna' . D33050, Janjukian, Aldinga) has a domed colony 25 mm in diameter and 10 mm high, somewhat larger than Tension Woods' colonies. The zooids have raised peristomes, extrazooidal frontal thickening with funnels and pores, and a distinct proximal pore which may be a spiramen. The avicularia are directed almost distally, and are much larger than those figured by Tenison Woods (1880, pi. 1, fig. 6c). They have a complete bar, and the rostra are acute, with a terminal channel which may indicate that the mandibles were elongated. The basal surface of the colony is obscured by matrix. The second colony (D31940, River Murray Cliffs) forms an almost truncated cone 20 mm in diameter and 1 3 mm high, but the zooids are not arranged in regular radial series. They are very worn, but the peristomial orifices and avicularia can be seen. The hollow basal side of the colony has numerous pores and irregular masses of calcification and may be inferred to have possessed an extrazooidal coelom in life. Pachystomaria parvipuncta MacGillivray (1895:97, pi. 8, fig. 24), another Miocene Australian form, appears to be a further species belonging to this complex. MacGillivray noted that the porous basal surface was similar to that ofSelenaria, an observation also made by Waters (1885) when describing M. magnirostris. At present it is not known whether any or all of these nominal species are synonymous, or if they are conspecific with Recent T. magnirostris. Tubiporella magnirostris (MacGillivray) Lepralia magnirostris MacGillivray, 1883 : 134, pi. 1, fig. 6. 78 P. L. COOK & P. J. CHIMONIDES 25 INTERIOR -WALLED CHEILOSTOME BRYOZOA 79 Tessaradoma magnirostris: MacGillivray, 1889 : 21 1, pi. 175, fig. 1. Porina magnirostris: Hincks, 1884 : 279 (sep. p. 129), pi. 9, fig. 6. Tubiporella magnirostris: Levinsen, 1909 : 307, pi. 16, figs 5a-d. Canu & Bassler, 1920 : 549, Figs 161 A-D. Tubiporella magnirostris part, 'f. \ictoriensis": Borg, 1940 : 416, Figs 1-3. SPECIMENS EXAMINED. BMNH. Port Phillip Heads, Victoria, Australia, 1887.6.27.1, 1887.12:10.120, 1893.8.11.17, 46 metres; 1927.8.4.24, 1979.1.12.3. Port Phillip, 1899.5.1.1214, 1899.7.1.5409, 5410, 5411, 2333, 2338, 2686, 2687, Victoria, 1979.7.2.1. Amirante Islands, 64 metres, 1 936. 1 2.30. 1 99. NMV. Port Phillip, 63876. DESCRIPTION (Figs 20, 24). Tubiporella with colonies forming broad, foliaceous expansions attached above the substratum by thick, extrazooidal, columns of porous calcification. Zooidal peristomes tubular, but not very prominent, with a small proximal, cuticle-covered foramen early in ontogeny. Spiramen originating on the proximal edge of the base of the peristome, apparently migrating proximally with thickening calcification. Avicularian subrostral chambers without pores, mandible hooked but not greatly elongated. Lateral frontal septulae surrounded by large uncalcified areas, which increase in size with calcifi- cation. Brooding zooids with larger secondary orifices than autozooids, with no avicularia, or paired, small avicularia. Avicularian mandibles of autozooids surrounding brooding zooids are directed away from brooding zooid when closed. REMARKS. The primary orifice can be seen early in ontogeny, and the developing peristome has a small foramen which rapidly becomes obscured. The 'migration' of the spiramen with thickening of the frontal shield is similar to that described in the umbonuloid family Adeonidae (see Cook, 1973). The patterning of brooding zooids and avicularia is similar to that found in D. pluma. The peristomial orifices of the brooding zooids are slightly larger than those of the autozooids, and the distal expansion of the ovicell can be seen. The lateral frontal septulae of the brooding zooids remain visible at the base of the funnels in the calcification, resulting in wide lacunae beneath the investing cuticle (cf. D. pluma}. Levinsen (1909 : 308) mentioned that the 'very thick basal wall of the colony ... is ... perforated by pore canals' which he noted were derived from multiporous basal septulae. Harmer (1957:1012) described a colony of Emballotheca subimmersa (MacGillivray) which encrusted the basal wall of a large specimen of T. magnirostris. Tubiporella boninesis Borg Tubiporella magnirostris part, 'f. boninensis': Borg, 1940 : 420, Fig. 4, (operculum). SPECIMENS EXAMINED. BMNH. Tizard Reef, 64 metres, 1889.21.27; 68 metres, 1892.8.8.26; Macclesfield Bank, China Sea, 66 metres, 1889.8.21.95,105; 64-79 metres, 1934.1 1.6.23; 46 metres, 1 893. 8. 1 1 . 1 7; China Sea, 66 metres, 1979. 1 .6.2; S. of Bua Bua passage, Falafuti, 264 metres, 1903.1.29.60. Fuafatu, 109 metres, 1903.1.29.61. DESCRIPTION (Figs. 18, 19, 26). Tubiporella with narrow, ligulate, semi-erect colonies, with zooids in 4-7 transverse series. Zooids with long, tubular peristomes, imperforate early in astogeny. Spiramen at the base of the peristome, migrating only slightly with ontogeny. Basal calcification produced into a massive, central keel in erect specimens, which produce branched, supporting pillars of extrazooidal, porous calcification 5-7 mm deep. Avicularian subrostral chambers large, imperforate. Rostra acute, with an open channel terminally. Figs 25 and 26 (25) Didymosella pluma sp. nov. BMNH 1899.7.1.5322, off Brazil, sketch of four zooids and two avicularia with open (upper) and closed (lower) mandibles. Scale bar = 0-50 mm; (26) Tubiporella boninensis Borg BMNH 1893.8.11.17, China Sea, sketch of three zooids and two avicularia with open (upper) and closed (lower) mandibles Scale bar = 0- 50 mm. 80 P. L. COOK & P. J. CHIMONIDES Mandibles elongated, directed laterally towards the margins of the colony; expanded laterally, with notched, fimbriated edges. REMARKS. Although Borg (1940) did not consider that his specimens were specifically distinct from T. magnirostris and did not formally introduce the name 'boninensis' for them, the name is available and is therefore used here. Borg noted that the operculum was hardly differentiated from the ascus wall, and that the avicularian mandible was 'distinctly broader than that of the Australian colonies and nearly straight, but with its tip rather sharply bent'. He also described the cuticular expansions and noted that 'this border is often broader on one side'. The mandibles are similar to those of D. pluma; when open they cover the area above the frontal spiramen. (see Fig. 26). The ancestrula resembles later-budded zooids and is unusual in having a lateral avicularium. Generally, T. boninensis differs from T. magnirostris in its colony form, shorter, more erect zooids, longer peristomes, and avicularian mandibles. Note on Tubiporella levinseni Borg, 1 940 In his description of the genus Tubiporella, Borg (1940) described a new species, T. levinseni, which had large, encrusting zooids with long, tubular peristomes, multiserial marginal pores and a proximal spiramen. The centrally placed avicularium had an elongated, almost setiform mandible. Borg mentioned that he had shown his specimens to Sir Sidney Harmer, who had recognized them as belonging to the same species as specimens from the Paternoster Islands which were present in the Siboga Expedition Collections. Harmer was not aware of Borg's description, which was subsequently published in Sweden during the 1939-1946 war (see Explanatory Note by Dr A. B. Hastings in Harmer, 1957 : xiv). The specimens from the Paternoster Islands were described as Reptadeonella flagellifera by Harmer (1957), who, however, had provisionally labelled them as "TubiporellcC. The structure and relationships of T. levinseni differ from those of Tubiporella as represented by the type species, T. magnirostris, and by T. boninensis. The specimens examined by Harmer from the Siboga Collections, together with one large, additional colony without locality, but comprising more than 2000 zooids (which was found fortuitously in the unnamed collections of the BMNH, and was not seen by Harmer), have been examined. A few zooids show a cuticular frontal wall with operculum, which is overarched by calcifi- cation, indicating that the ontogeny of the frontal shield is umbonuloid. There are no frontal pores, and the frontal septulae are multiserial and extend round the distal side of the orifice (see also Borg, 1940). The long tubular peristome is thus subterminal, and the primary orifice is well defined, wih minute condyles and distinct operculum (see also Borg, 1940 : 422, Fig. 8). The avicularium is central in position and is derived from one lateral- oral septula. The subrostral chamber is small, and the rostrum is acute and directed distally and laterally. The mandible is elongated, setiform and slung between condyles. Zooids communicate through septulae at the base of the vertical walls, but there is no basal coelom and the basal wall is completely calcified. The zooids are nearly all encrusting, but where they stretch across irregularities in the substratum, have produced small, peg-like expansions. There are however, no extrazooidal pillars as in Tubiporella. All the character correlations are similar to those found in the umbonuloid family Adeonidae (see Cook, 1973), and they confirm that Harmer (1957) was correct in attributing his specimens to the genus Reptadeonella, which has encrusting colonies. Both Borg (1940) and Harmer (1957) mentioned that some zooids with short peristomes may have been brooding zooids. A few zooids have wide, rather slit-like orifices, but their peristomes do not differ from those of other zooids. Generally, brooding zooids in the Adeonidae are strongly dimorphic, and it is probable that as yet, no brooding zooids have INTERIOR-WALLED CHEILOSTOME BRYOZOA 81 been found in this species. The synonymy, distribution and measurements of R, levinseni (Borg) are as follows: Reptadeonella levinseni (Borg) Tubiporella levinseni Borg, 1940 : 420, pi. 1 and Figs 5-8. Reptadeonella flagellifera Harmer, 1957 : 817, pi. 54, figs 4-7. SPECIMENS EXAMINED. BMNH. Siboga Stn 315, N. of Sumbawa, Paternoster Islands, 0-36 metres, 1979. 1.2.1; no locality, 1979.1.2.2. ZMA. Stn 3 1 5 and Stn 60, 303, Hangsisi, Samau Islands, W. Timor, 0-36 metres. DISTRIBUTION. Jaluit, Marshall Islands, 2 metres (Borg), Paternoster Islands and West Timor. MEASUREMENTS (in mm). Lz 0-80-1-30 Iz 0-50-0-90 Lo 0-13 lo 0-16-0-20 Lr 0-16-0-25 Lm 0-30-0-50. Discussion Study of Bryozoa emphasizing the concept of colonies as entities composed of cooperative members has increased during recent years as a result of observation of living colonies and the application of new techniques. Species show many kinds and degrees of structural and functional integration. Colonies may be composed of almost autonomous member zooids, or have various correlations of morphology and behaviour which in some cases amount to virtual colony control (see Cook, \919a, and Ryland 1979). The relationships of zooid body walls, intercommunications, zooidal and extrazooidal coeloms, astogenetic and ontogenetic changes, polymorphism and the patterning of polymorphs, all express the degree of such colony control and may be analysed and arranged in series of increasing integration (see Boardman & Cheetham, 1973 and Cook, 19790). Expressed in terms of such series, the degree of integration of the families Euthyrisellidae and Didymosellidae is considerable. It is reflected in the zonation and polymorphism of brooding and avicularian zooids, and in the supporting and attachment structures of colonies. Most significantly, it is expressed in the preponderance of interior calcified walls, and in the various combinations of intrazooidal, interzooidal and extrazooidal coelomic communications. Relationships both between the Euthyrisellidae and the Didymosellidae, and between them and other ascophoran groups are hard to evaluate. Although they share crypto- cystidean frontal shield ontogeny, the differences between the two kinds of shield suggest that these structures have arisen more than once in cheilostome evolutionary history. The appearance through time of more than one kind of frontal shield in the 'catenicellid' genera (see Banta and Wass, 1979), indicates that many more cryptocystidean families will require detailed investigation before the degrees and kinds of relationships among apparently similar groups are established. The presence of extrazooidal basal coelom and of interior basal zooid walls is another shared character. However, these structures are also known in such diverse systematic groups as the lunulitiform, free-living anascan Selenariidae (see Chimonides & Cook, 1981) and the cryptocystidean Petraliidae. Petralia undata has erect, reteporiform, rooted colonies with a high degree of organization and an extensive basal coelom. The lateral and basal rooting coleomic systems of the Euthyrisellidae have very close parallels among the many erect, flustriform genera, especially Onchoporella (see p. 59). Peristomal spiramina, too, are found in a diversity of Recent and fossil ascophoran genera; for example, Adeonella (see Cook, 1973), Gigantopora (see Harmer, 1957), and compare Pachydera rarepunctata Voigt (1967 : 67, pi. 24, figs 1, 2, from the Upper Cretaceous of West Kasachstan, USSR). Within the Euthyrisellidae, there is great diversity in frontal shield calcification and in 82 P. L. COOK & P. J. CHIMONIDES INTERIOR-WALLED CHEILOSTOME BRYOZOA 83 orifice shape which is not strongly correlated with colony form. This diversity suggests that the family has a long evolutionary history, but unfortunately, no fossil species are known. Internodes of Tropidozoum are fairly robust, and have been found in transported deposits; they might be expected to be recognizable in Tertiary sediments. Even the thinly calcified zooids of Euthyrisella could be preserved, by analogy with the equally fragile fragments of Selenariopsis, which have been found, albeit rarely, in the fine fractions of Australian Miocene sediments (see Cook & Chimonides, 1981). Lagaaij (1968, 1973) has docu- mented records of similar fragile species, several of which were not previously known as fossils. It is perhaps possible that the ancestral forms of Euthyrisellidae had more heavily calcified colonies, which have not yet been recognized as being systematically related to the Recent species. In contrast, the long fossil record of the Didymosellidae might be expected to provide some evidence of evolutionary trends and relationships. Diversity among the known species, both in zooidaP morphology and colony form is, however, far lower than that found in the Euthyrisellidae. Generally, the species have very similar patterning of avicularia and autozooids, and the repent or semi-encrusting colony form is know in both Eocene and Recent species. The erect, branching colonies of D. porosa and D. clypeata are not found among Recent species, and the algal association of D. larvalis (like that of Neoeuthyris woosteri), may have been evolved fairly recently. To a certain extent, the distribution of species in time and space shows strong collection bias, and reflects the large number of faunal descriptions of the abundant bryozoan sediments of the Tertiary-to-Recent of the Gulf of Mexico and Australasian regions. The Euthryisellidae seems to have a purely Indo-West Pacific and Australian range. Most species are from shallow shelf or sublittoral waters, and several are rooted in sediments. The Didymosellidae are also shallow shelf species, although D. pluma and T. boninensis are found from the deeper end of the range. All species are from warm waters. Labracherie (1972) noted that the occurrence of Didymosella (together with other species) allowed the inference of warm marine conditions in the Middle Eocene of south Western France. The general pattern of distribution of the family in time and space follows that illustrated by Lagaaij & Cook (1973) for other warm water, shallow shelf species, and even for some deeper water forms (see Cook & Lagaaij, 1976). This comprises a very wide (sometimes circumtropical) distribution during the Eocene and Oligocene followed by a progressive restriction of range in a south-easterly direction through time. The occurrence of Recent D. pluma off Brazil and in the Gulf of Mexico is of importance, as it indicates that, like the genus Adeonellopsis (see Cook, 1973), groups which are supposed to be 'extinct' in the western Atlantic may be found to be still living in this area. Summaries in French and German Nous decrivons la structure de la colonie et la morphologic zooidale trouvees dans les deux families cheilostome-cryptocystidiennes des Bryozoaires, la Euthyrisellidae et la Didymosellidae. Tous les murs calcifies sont internes, at les coelomes basals (extra- zooidales) sont presents dans presque toutes les especes. Dans la famille Euthyrisellidae, la communication interzooidale viscerale- a hypostigienne complemente ou meme remplace Figs 27 -30 Scanning electron micrographs of communication septulae: (27) Didymosella porosa (Stoliczka) BMNH D36788-92, Middle Oligocene, New Zealand, basal septulae (left) xlOO (right) x600; (28) Neoeuthyris woosteri (MacGillivray) BMNH 1979.2.1.1, Recent, Western Australia, lateral frontal septulae at base of developing avicularian subrostral chamber x315; (29) N. woosteri, visceral-to-hypostegal septulae (arrowed) viewed from distal side x!97; (30) Tropidozoum cellariiforme Harmer BMNH 1964.3.10.1, Recent, Celebes, visceral-to- hypostegal septula in distal-lateral wall of brooding zooid, viewed from proximal side. Note suture in wall showing contribution of distal zooid calcification to brooding zooid orifice x368. 84 P. L. COOK & P. J. CHIMON1DES as a s 33 --, INTERIOR-WALLED CHEILOSTOME BRYOZOA 85 la sorte intrazooidale la plus commune, cells qu'on trouve dans la Didymosellidae. Dans la Euthyrisellidae, une extension du coelome hypostigienne au-dessous du cote basal de la paroi frontale cacifiee, est presente dans la majorite des especes. Une espece nouvelle est introduite, Tropidozoum burrowsi. La famille Didymosellidae possede un record fossil qui s'etend d'Eocene Moyen et a une distribution tres grande du Tertiare au Recent. Nous introduisons une nouvelle espece Recente, Didymosella pluma. Tubiporella boninensis Borg a ete redefmie. Tubiporella levinseni Borg a ete redecrite et attribue au genre umbonuloide Reptadeonella. Die Koloniestruktur und Morphologic der Zooide der Euthyrisellidae und Didy- mosellidae, zweier cryptocystider Familien der Bryozoa Cheilostomata, werden beschrieben. Alle verkalkten Wande sind intern, und kolonieweite (extrazooidale) Basalcoelome fmden sich bei fast alien Arten. Bei den Euthyrisellidae erganzt oder sogar ersetzt die interzooidale (viscerale bis hypostegale) Kommunikation die gewohnlichere intrazooidale, die sich bei den Didymosellidae findet. Bei den meisten Arten der Euthyrisellidae ist eine Verlangerung des hypostegalen Coeloms an der Basalseite des verkalkten Frontalschildes zu beobachten. Eine neue Art, Tropidozoum burrowsi, wird eingefuhrt. Die Familie Didymosellidae ist fossil seit dem mittleren Eozan nachgewiesen und ist vom Tertiar bis zur Gegenwart weit verbreitet. Eine neue recente Art, Didymosella pluma, wird eingefuhrt und Tubiporella boninensis Borg neu defmiert. Tubiporella levinseni Borg wird wiederbeschrieben und in die umbonuloide Gattung Reptadeonella venviesen. Acknowledgements These investigations would not have been possible without the collection and donation to the British Museum (Natural History) of well-preserved specimens of 'rare' species. We are therefore particularly grateful to Dr P. Bock (Royal Melbourne Institute of Technology, Victoria), Dr J.-G. Harmelin (Station Marine d'Endoume, Marseille), Dr J. S. Ryland (University College of Swansea), Dr R. E. Wass (University of Sydney, New South Wales) and Dr W. J. Woelkerling (La Trobe University, Victoria) for their help. We should also like to thank Dr M. Labracherie (Universite de Bordeaux) and Dr P. Taylor (British Museum, Natural History) for the loan of fossil specimens. References Banta, W. C. 1970. The body wall of cheilostome Bryozoa. Ill, The frontal wall of Watersipora arcuata Banta with a revision of the cryptocystidea. J. morph. 131 : 37-56. Banta, W. C. & Wass, R. E. 1979. Catenicellid cheilostome Bryozoa: 1. Frontal walls. Aust. J. Zool. Suppl. Ser. 68 : 1-70. Bassler, R. S. 1936. Nomenclatorial notes on fossil and Recent Bryozoa. /. Wash. Acad. Sci. 24(4): 156-162. Boardman, R. S. & Cheetham. A. H. 1973. Degrees of colony dominance in stenolaemate and gymnolaemate Bryozoa. In R. S. Boardman, A. H. Cheetham & W. A. Oliver, Animal Colonies. Stroudsburg. Borg, F. 1940. On the genus Tubiporella and on a new boring Bryozoan. Zool. Bidrag. Upps. 18:415-437. Bretnall, R. W. 192 1 . Neoeuthyris: A new genus to accommodate Euthyris woosteri MacGillivray. Rec. Aust. Mm. 13(4): 157-162. Figs 31-33 Micrographs of longitudinal thin-sections of epoxy resin embedded specimens of Euthyrisellidae: (31) Neoeuthyris woosteri (MacGillivray) BMNH 1979.2.1.1, Western Australia, zooids with cuticular basal walls encrusting algae, autozooid (left) and brooding zooid with large ovicell (centre), arrow indicates area enlarged in Fig. 32 x30; (32) N. woosteri visceral- to-hypostegal communications (arrowed) x!65; (33) Euthyrisella obtecta (Hincks) BMNH 1890.7.23.38, Torres Straits, brooding zooids (without ova), cuticular basal wall detached at right x70. a ascus, o ovicell, op operculum, /frontal process, / tentacles, g gut, h hypostegal coelom, v visceral coelom, b basal coelom, as algal substratum. 86 P. L. COOK & P. J. CHIMONIDES •34 op ,36 INTERIOR-WALLED CHEILOSTOME BRYOZOA 87 Brown, D. A. 1948. Six new Recent and Tertiary Genera of cheilostomatous Polyzoa from New Zealand. Ann. Mag. nat. Hist. 12 (1) : 108-122. 1952. The Tertiary Cheilostomatous Polyzoa of New Zealand, i-xii, \-405. London. 1958. Fossil cheilostomatous Polyzoa from South West Victoria. Mem. geol. Surv. Viet. 20 : 1-83. (ami, F. 1904. Les Bryozoaires du Patagonien. Mem. Soc. geol. Fr. Pal. 12 (3) : 1-30. ('ami, F. & Bassler, R, S. 1 9 1 7. A synopsis of American Early Tertiary cheilostome Bryozoa. Bull. U.S. natn.Mus.96: 1-87. 1920. North American Early Tertiary Bryozoa. Bull. U.S. natn. Mus. 106 : i-xx, 1-879. 1929. Bryozoa of the Philippine region. Bull. U.S. natn. Mus. 100 (9) : 1-685. 1935. New species of Tertiary cheilostome Bryozoa from Victoria, Australia. Smithson. misc. Collns93(9): 1-54. Cheetham, A. H. 1963. Late Eocene zoogeography of the eastern Gulf Coast region. Mem. geol. Soc. Am.9l : 1-113. Chimonides, P. J. & Cook, P. L. 1981. Observations on living colonies of Selenaria (Bryozoa, Cheilostomata). 2 Cah. Biol. mar. (in press). Cook, P. L. 1964. Polyzoa from west Africa. 1. Notes on the Steganoporellidae, Thalamoporellidae and Onychocellidae (Anasca, Coilostega). Annls Inst. oceanogr., Monaco ('Calypso' 6) 41 : 43-78. 1965. Notes on some Polyzoa with conical zoaria. Cah. Biol. mar. 6 : 435^54. 1966. Some 'sand fauna' Polyzoa (Bryozoa) from Eastern Africa and the Northern Indian Ocean. Cah. Biol. mar. 7 : 207-223. 1967. Polyzoa (Bryozoa) from West Africa. The Pseudostega, the Cribrimorpha and some Ascophora Imperfecta. Bull. Br. Mus. nat. Hist. (Zool.) 15 (7) : 32 1-35 1 . 1973. Preliminary notes on the ontogeny of the frontal body wall in the Adeonidae and Adeonellidae (Bryozoa, Cheilostomata). Bull. Br. Mus. nat. Hist. (Zool.) 25 (6) : 243-263. 1975. The genus Tropidozoum Harmer. Docums Lab. Geol. Fac. Sci. Lyon H.S. 3, fasc. 1 : 161-168. 1977. The genus Tremogasterina Canu (Bryozoa, Cheilostomata). Bull. Br. Mus. nat. Hist. (Zool.) 32(5): 103-165. 1979a. Some problems in interpretation of heteromorphy and colony integration in Bryozoa. In G. Larwood & B. R. Rosen, Biology and Systematics of colonial organisms. Sytematics Association Special Vol. 1 1 . London & New York. 19796. Mode of life of small, rooted 'sand fauna' colonies of Bryozoa. In G. Larwood & M. B. Abbott, Advances in Bryozoology. Systematics Association Special Vol. 13. London & New York. 198 1 . Bryozoa from Ghana — a preliminary survey. Annls Mus. r. Afr. cent, (in press). Cook, P. L. & Chimonides, P. J. 1981. Morphology and Systematics of some rooted cheilostome Bryozoa. /. nat. Hist. 15 : 97-134. Cook, P. L. & Lagaaij, R. 1976. Some Tertiary and Recent conescharelliniform Bryozoa. Bull. Br. Mus. nat. Hist. (Zool.) 29 (6) : 3 1 7-376. Faura y Sans, M. & Canu, F. 1917. Sur les Bryozoaires des terrains Tertiares de la Catalogne. Treb. Inst. catal. Hist. nat. 1916 : 59-193 (sep. pp. 1-137). Freeman, J. A. & Spurlock, B. O. 1962. A new epoxy embedment for electron microscopy. J. Cell. Biol. 13,3:437-443. Glauert, A. M. 1965. The fixation and embedding of biological specimens. Chap. 7. In D. H. Kay, Techniques for electron microscopy. Oxford. 1974. Fixation, dehydration and embedding of biological specimens. Pt. 1. In A. M. Glauert, Practical methods in electron microscopy. Amsterdam, Oxford & New York. Hakansson, E. 1973. Mode of growth of Cupuladriidae (Bryozoa, Chielostomata). In G. Larwood, Living and Fossil Bryozoa. London & New York. Figs 34-36 Micrographs of longitudinal thin-sections of epoxy resin embedded specimens of Euthyrisellidae and Didymosellidae: (34) Pleurotoichus clathratus (Harmer) BMNH 1976. 1 .2. 1 , Queensland, showing large subfrontal hypostegal coeloms x60; (35) Euthyrisell obtecta (Hiricks) BMNH 1890.7.23.38, Torres Straits, visceral-to-hypostegal communication (arrowed) and minute subfrontal hypostegal coelom x!70; (36) Didymosella larvalis (MacGillivray) BMNH 1 890.4. 1 7. 1-1 Opt., Spencer Gulf, zooids with cuticular basal walls encrusting algae x40. a ascus, o ovicell, op operculum, / frontal process, p primary orifice, 5 secondary orifice, h hypostegal coelom, s/subfrontal hypostegal coelom, v visceral coelom, b basal coelom, as algal substratum. 88 P. L. COOK & P. J. CHIMONIDES Harmer, S. F. 1902. On the morphology of the Cheilostomata. Q.Jl microsc. Sci. 46 : 263-350. 1926. The Polyzoa of the Siboga Expedition, Pt. 2. Cheilostomata, Anasca. Siboga Exped. 28b: 181-501. 1957. The Polyzoa of the Siboga Expedition, Pt. 4. Cheilostomata Ascophora 11. Siboga Exped. 28d: 641-1 147. Hastings, A. B. 1960. Proposal to place the generic name Neoeuthyris Bretnall, 1921 (Phylum Polyzoa) on the Official List of Generic Names in Zoology. Z.N. (S) 1314. Bull. zool. Norn. 17 (6-8) : 244-245. 1964. The cheilostomatous Polyzoa Neoeuthrvis woosteri (MacGillivray) and Reginella doliaris (Maplestone). Bull. Br. Mus. nat. Hist. (Zool.) 11 (3) : 243-262. Hayward, P. J. & Cook, P. L. (in press). The South African Museum's Meiring Naude Cruises, Pryozoa. Ann. S. Afr. Mus. Hincks, T. 1882. Contributions towards a general history of the marine Polyzoa. 10. Foreign Cheilo- stomata. Ann. Mag. nat. Hist. (5) 10 : 160-1 70. 1884. Contributions towards a general history of the marine Polyzoa. 13. Polyzoa from Victoria. Ann. Mag. nat. Hist. (5) 14 : 276-284. 1893. Contributions towards a general history of the marine Polyzoa, 1880-91, Appendix. Ann. Mag. nat. Hist. (6) 11 : 175-182. Jullien, J. 1883. Dragages du Travailleur'. Bryozoaires. Especes draguees dans 1'Ocean Atlantique en 1881. Bull. Soc. zool. Fr. 7 (5) : 497-52 1 . Labracherie, M. 1972. Les Bryozoaires dans 1'appreciation des climats du Tertiaire. Bull. Inst. Geol. Bassin Aquitaine 12 : 43-46. Lagaaij, R. 1968. First fossil finds of six genera of Bryozoa Cheilostomata. Atti Soc. ital. Sci. nat. 108 : 345-359. 1973. Shallow-water Bryozoa from deep-sea sands of the Principe channel, Gulf of Guinea. In G. Larwood, Living and Fossil Bryozoa. London & New York. Lagaaij, R. & Cook, P. L. 1973. Some Tertiary to Recent Bryozoa, In A. Hallam, Atlas of Palaeobiogeography. Amsterdam, London & New York. Levinsen, G. M. R. 1909. Morphological and Systematic studies on the Cheilostomatous Bryozoa. Copenhagen. Livingstone, A. A. 1926. Studies on Australian Bryozoa, No. 4. Rec. Aust. Mus. 15 (2) : 167-176. MacGillivray, P. H. 1869. Descriptions of some new genera and species of Australian Polyzoa. Trans. Proc. R. Soc. Viet. 9 (2) : 126-148. 1879. Polyzoa. In F. McCoy, Prodromus of the Zoology of Victoria, Melbourne, 1879-1890. Decade IV: 21-40. 1889. Polyzoa. In F. McCoy, Prodromus of the Zoology of Victoria, Melbourne, 1879-1890. Decade XVIII: 21 1-291. 1891. Descriptions of new or little known Polyzoa, Pt. 14. Trans. Proc. R. Soc. Viet. n.s. 3 : 77-83. 1895. A monograph of the Tertiary Polyzoa of Victoria. Trans. R. Soc. Viet. 4 : 1-166. Mawatari, S. 1952. Bryozoa of Kii Peninsula. Publs Seto mar. biol. Lab. 2 (2) : 261-288. Moyano, H. 1975. Romancheina asymmetrica sp. nov., nueva especie da la Antartica y clave para las especies antarticas y magallanicas de Exochellidae (Bryozoa, Cheilostomata). Boln Soc. Biol. Concepcion 49 : 6 1-69. Nye, O. B., Dean, D. A. & Hinds, R. 1972. Improved thin section techniques for fossil and Recent organisms. J. Paleont. 46 : 271-275. Potter, W. G. 1970. Epoxide Resins. Plastics Institute. London. Ryland, J. S. 1979. Structural and physiological aspects of coloniality in Bryozoa. In G. Larwood & B. R. Rosen, Biology and Systematics of colonial organisms. Systematics Association Special Vol. 1 1. London & New York. Ryland, J. S. & Hayward, P. J. 1977. A synopsis of the British anascan bryozoans. Syn. Br. Fauna n.s. 10: 1-188. Sandberg, P. 1971. Scanning electron microscopy of cheilostome bryozoan skeletons; techniques and preliminary observations. Micropaleontology 17 : 129-1 5 1 . Stoliczka, F. 1865. Fossile Bryozoen aus dem Tertiaren Griinsandsteine der Orakei-Bay bei Auckland. Reise Novara, Geol. Theil 1, 2 : 87—1 58. Tenison Woods, J. E. 1880. On some Recent and fossil species of Australian Selenariadae (Polyzoa). Trans, phil. Soc. Adelaide 1880 : 1-12. Voigt, E. 1967. Oberkreide-Bryozoen aus den asiatischen Gebieten der UdSSR. Mitt. geol. Stlnst. Hamb. 36 : 5-95. INTERIOR-WALLED CHEILOSTOME BRYOZOA 89 Waters, A. W. On fossil chilostomatous Bryozoa from South-West Victoria, Australia. Q.Jl geol. Soc Lond. 37 : 309-347. 1882a. On fossil chilostomatous Bryozoa from Mount Gambier, South Australia. Q.Jl geol. Soc. Lond. 38 : 257-276. 1882&. On chilostomatous Bryozoa from Bairnsdale (Gippsland). Q.Jl geol. Soc. Lond. 38:501-513. — 1885. Chilostomatous Bryozoa from Aldinga and the River-Murray Cliffs, South Australia. Q.Jl geol. Soc. Lond. 41 : 279-3 10. — 1887. Bryozoa from New South Wales, North Australia, etc. Part 2. Ann. Mag. nat. Hist. (5)20: 181-203. — 1891. North-Italian Bryozoa. Q.Jl geol. Soc. Lond. 47 : 1-34. Manuscript accepted for publication 1 5 January 1981. British Museum (Natural History) 1881-1981 Centenary Publications The Natural History Museum at South Kensington By W. T, Steam This book presents a full history of the Museum, its collections, Directors and eminent members of staff. It provides for the first time an account of the major research undertaken and insights into the personalities of the key people in the Museum's development and evolution. Co-published with William Heinemann. Alfred Waterhouse and the Natural History Museum By Mark Girouard Designed by Alfred Waterhouse in the 1870's the Museum is surely one of London's most outstanding pieces of architecture. This attractively illustrated book describes the development of the design and highlights some of the building's most interesting features. British Museum Natural History By Peter Whitehead & Colin Keates The Museum is really a huge scientific research institution - acquiring, describing and classifying all manner of natural history material, both specimens and artworks. It has some of the richest collections of their kind in the world and it is these and how they are obtained and managed that are the subject of this book. The lavish, full colour illustrations and lively text will appeal to everybody interested in natural history. Co-published with Philip Wilson Ltd. Chance, Change and Challenge This multi-author twin volume work is one of the Museum's most ambitious publishing projects. In the first volume The Evolving Earth twenty scientists have been asked to summarise the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume The Evolving Biosphere Museum scientists have chosen an evol- utionary concept - speciation, coevolution, biogeography etc and related this to the group of animals or plants in which they are specialising. Co-published with Cambridge University Press. Animal Identification-A Reference Guide VOLUME 1 : MARINE AND BRACKISH WATER. Edited by R. W. Sims. VOLUME 2: TERRESTRIAL AND FRESHWATER. Edited by R. W. Sims. VOLUME 3 : INSECTS. Edited by D. Hollis. These guides provide the reader with lists of primary sources of reference that can be used to identify (or lead to the identification of) living animals throughout the world. The references are arranged in systematic and geographical sequence in order to facilitate searching. Co-published with John Wiley & Sons Limited. Nature Stored Nature Studied A short guide complementing the exhibition of that name to be staged in the Museum throughout 1981. It gives a fascinating insight into the history of the Museum collections and the research carried out "behind the scenes". Titles to be published in Volume 41 Anatomy and phylogeny of the Chinese Major Carps Ctenopharyngodon SteincL, 1866 and Hypophthalmichthys Blkr., 1860. By Gordon Howes Morphology and systematics of some interior-walled cheilostome Bryozoa By P. L. Cook and P. J. Chimonides Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae. By Ailsa M. Clark Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the V British Museum (Natural History) Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae Ailsa M. Clark Zoology series Vol 41 No 3 29 October 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to : Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Dr J. G. Sheals Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISSN 0007-1498 Zoology series Vol 41 No 3 pp 91-135 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 29 October 1981 Notes on Atlantic and other Asteroidea. 1. Benthopectinidae Ailsa M. Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Contents Synopsis 91 Introduction 91 Systematic discussion 91 Family Benthopectinidae 91 Genus Cheiraster 107 Genus Pectinaster 118 Genus Benthopecten 122 Acknowledgements 134 References 134 Synopsis Various taxa of the family Benthopectinidae, mainly from the Atlantic, are reviewed. The genera Pectinaster and Cheiraster are redefined, Pectinaster being restored to Perrier's concept of 1894 by transfer of several nominal species without specialized papularia to Cheiraster, which is enlarged by inclusion in it of Luidiaster Studer as a subgenus. Two further subgenera of Cheiraster are described, Barbadosaster for Cheiraster echinulatus (Perrier) and Christopher aster for C. mirabilis (Perrier), also including C. blakei sp. nov. and two other species. On the basis of older collections and new material, mainly from the Pillsbury and Gerda collections of the University of Miami, the morphological and geographical ranges of certain species of these two genera and of Benthopecten are extended and several names are put in synonymy, tabular keys to the Atlantic species are also included. Introduction In the course of a major review of the Atlantic species of Asteroidea, assessment of the neglected deep-sea family Benthopectinidae proved so complicated as to necessitate a preliminary paper to dispose of some of the more complicated taxonomic problems. The ordinal position of the Benthopectinidae will be discussed in the main study on Atlantic asteroids (Clark & Downey). The family was removed from the Paxillosida to the revived order Notomyotida Ludwig, 1 9 1 0 by McKnight (1975). Systematic discussion Family BENTHOPECTINIDAE Verrill, 1894 Archasteridae Benthopectininae Verrill, 1894 : 245-246, 268. Archasteridae Pontasterinae Verrill, 1894 : 246-247, 268. Benthopectinidae Verrill, 1899:200, 217; Ludwig, 1910:442, 458-461; Fisher, 1911:120-122; Verrill, 1914 : 310-31 1; Spencer & Wright, 1966 : U48. Plutonasteridae Pontasterinae Verrill, 1899 : 200. Cheirasteridae Ludwig, 1910 : 442, 444-447. Bull. Br. Mus. nat. Hist. (Zool.) 41 (3) : 9 1-1 35 Issued 29 October 1 98 1 91 > 0 ,-^ c g 1 a o ^> oo ^~ "S -2 1 6 S ~5 -o^S ^- w '^2 § .^ a -S i vt ~ 2 ^ 1 5 . i _o "5 ° o a "° t«-i r~ '— — D. w3 Ch c/i a> o. -3 0 e X) 0 c ^ C a —« TO r . > 00 e E S^^ E o O4 O <*"> 0 r* in "T • ^ c ot ESo— '~ ^loaS^oi^^i^2 1 g"l ^^ in ^ £^ — ("^ 00 £ m r-- o V mtiasler bandanus C \thopecten Verrill, U icanthonotus Fisher, ON ON r-22oO « y — "" fe ^ "" ' 00~ — T'tS -2 tT.S,5o~S --f C^S.Sjfe-g./i £S ££;S-g.£ff~i; Ilils'lllll''?^!^^!!! is|Iilll|l*Stll22.tlll Q t\A C* • ^* "^ ^ --^ "^ *»*«-•*»«» B^* ^J ^> ^J ^ **-X J^ ^ implex (Perrier, 188 oinosissimus (Slader vinosus Verrill, 188^ vinuliger (Ludwig, 1' 'yracius Fisher, 1912 iolaceus (Alcock, 1 8( iraster Studer, 1883 oronatus (Perrier, 1 8 'iomedeae Fisher, 1 9 chinulatus (Perrier, noplusVemll, 1915 azellae Studer, 1883 ranulatus Ludwig, 1 lops Fisher, 1906 -. £ 2 ^ «j 5j ^ ^ -C -*:•—•— s: c s ^^.^,ci,ci.^v c^ c^ «•» *» *» ^ ^ •*• .5- OC OO- — Z •^ cc QQ QQOqQQQgQgDqQcJQgQcJQcJQ^QQQcJQgCQQdDciQaQciQaQd 22 ^ ^ to to (^ Cj Cj t) Cj cj cj Cj Cj 93 -s: §• -s: o 5 00 GO 5 *o c GO 3 S O 0 HH 5 O E o c ° E ON 8. a 11 UJ . ^^ C E^ S c ,^o 4 rN E o 0 £ c" c U C ""* ,-i, •o o £ 6 £ o T^" ^ < O O f*^ °" ^ c I 6 <*-. J 3 o o S -S c \o o Tj- '^- O csj o Jr «a .5 « .2 GO^ t t «5 NO -3 $ C «> oo« g &0 C j= S^-e-c co firsfi c O O4S §" ^ « rt"r^ GOOQ>V5GOl^GOW^U fi *-> . «^U 3 g W^ o C^GJ^Z r* ON oo r^ -'S 2^ ;^o3S 5 O — £ _ : s 2 S *.- • to - C « !'C.2P^-§ :^ > ^ 3 ON 00 oo -i> '5 o 2 »~^?" •S ~ — • -~ _j "5 £! £•« 1 S2 8^ *>- || |^ s: ^i bo^s o a Q ^s c> •••> -CI ^1 s: s: o EX •^ 5 S pilosus (Alcock, 189 richardsoni Fell, 195 snyderi Fisher, 1 906 subtuberculatus (Sla ON ^. — ON fe oo JS — Is * "2 S rt "S^, §^ g& ~-s ~— > .cx^? c OJ •o ca •> • 5 > t T3 « 3 cu i— l 5^ fc'Kj C3 g ^ ? •C5 60 ^j Q 000 oo m ON o"l — c 28^ tT^-c PO o> 2 ON •s ^°° rT gt^«n rllfSS «j ^^» o ^^ *~^ -5 a o_p „ B S? "^ — I-" 0-§<-C.^ >. v r ri ^ $Q S? 1> •~ -^. ft i> - M N S "^ CU e 50 tO s, E^ IS ^^s^st a; a,' a; ex; a; I d *s 1:1 o o d d C/5 C/3 a a -2 -^ a a U U ft, s U _ _ •__ ~-, •_ o o o^ o d d d O d ?u -s: .1 «g ,8 t O d on o o r^ 4 g .2 2iil2||2^| < j cu w cu ^ ^ D U > c — — <£ — C • W 1) - -C U c« •o :5 i 03 z FAMILY BENTHOPECTINIDAE 95 This family of primarily deep-water Asteroidea is in need of review, the last comprehensive survey being by Ludwig (191 0), intended only as a preliminary treatment, modified by Fisher (191 1), while Macan (1938) simply listed the species of Pectinaster and Benthopecten (the genera represented in the Murray Expedition collections), summarizing earlier conclusions about these two genera. Since the present study concentrates on the Atlantic fauna, the genera Acontiaster, Myonotus and Nearchaster — known only from the Indo-Pacific — have not merited attention here but are listed below with the other taxa currently recognized together with nomen- clatural modifications now proposed. Saraster A. H. Clark, 1916, type and only species, S. insignis A. H. Clark, is now synonymized with Nearchaster. It was described without comparative remarks or figures and the holotype (from Lower California) has disintegrated but the description leaves no doubt in my mind that it was conspecific with N. aciculosus (Fisher, 1910), taken at 14 Albatross stations off California (and further north), any differences being accountable to the smaller size of the type. Additionally, the following names formerly reduced to synonyms are now otherwise disposed: Benthopecten armatus (Sladen, 1889), NE America — Portugal, 2290-2470 m., synonymized with B. spinosus Verrill by Macan, 1938, following Verrill, 1899 and Grieg, 1921, treated as valid by Farran, 1913 and Mortensen, 1927 but now synonymized with B. simplex (Perrier), following Ludwig, 1910. B. fischeri Perrier, 1894. NW Africa, 1060-1430 m, synonymized with B. spinosus by Macan, 1938, following Grieg, 1921, but now provisionally treated as valid or a possible synonym of B. simplex. Cheiraster planus Verrill, 1915, West Indies (no details), synonymized with C. mirabilis (Perrier) by Downey, 1 973, but now considered as valid. Pontaster perplexus Perrier, 1894, NW Africa, 2320-2330 m., synonymized with P. tenuispinus (Diiben & Koren) by Ludwig, 1910, but now synonymized with Pontaster sepitus Verrill, which is itself now revived from the synonymy of (C.) echinulatus (Perrier), to which it was referred by Ludwig, 1910, and newly referred to Cheiraster. Pontaster venustus Sladen, 1889, Azores — Portugal, 1250-4000 m., with var. robusta Sladen, 1889, Cape Verde Is area, depth ?, synonymized with Pectinaster filholi Perrier by Macan, 1938, following Ludwig, 1910, but now synonymized with Cheiraster sepitus. B. semisquamatus (Sladen, 1889) was treated by Goto, 1914, without explanation as a synonym of B. spinosus Verrill, apparently extending Verrill's synonymy of B. semi- squamatus var. occidentals to the Pacific semisquamatus itself, despite Ludwig's retention in 1910 of this as valid, followed by Fisher (1919 : 210) and Doderlein (1921). Okutani (1969) has followed Goto. The genera of Benthopectinidae can be distinguished by the tabular key (Table 1). Some of the generic limits within the Benthopectinidae are still ill-defined. In Ludwig's review (1910), on the strength of the unusual dorsal arm muscles, the present family was interpreted as a suborder Notomyota including two families: Cheirasteridae (for Cheiraster Studer, 1883,LwzW/asterStuder, 1 884, Pontaster Sladen, 1885 restricted, Pectinaster Perrier, 1885, Marcelaster Koehler, 1907 and Gaussaster nov., of which Marcelaster was synonymized with Luidiaster by Fisher, 1940) and Benthopectinidae (for Benthopecten Verrill, 1884 and Pararchaster Sladen, 1885). The latter group was distinguished by the presence of odd interradial marginal plates, the superomarginal ones with unusually large spines. Fisher (1911) described two new genera, Nearchaster and Myonotus, both with the 96 A.M.CLARK Table 1 Tabular key to the genera of Benthopectinidae. Entries in brackets indicate a slight development or tendency. 12 3 4 5 6 7 89 Benthopecten A w/b L f M/R + 2(3) P F Myonotus A w/b L P M' + 2 P F Nearchaster I w L P M + 2(3) P F Gaussaster I b L f R(M) + 3,2 P F Acontiaster R w D P M (+) 2 u F Cheiraster (Luidiaster) R b(w) L P M -(+) 2 P F Cheiraster (Barbadosasler) R b D f(p) M 2(3) P F Cheiraster (Christopher aster) R d D P(0 M + 1(2) P F Cheiraster (Cheiraster) R b D P M 1 P F Pectinaster R o L P M/R - 1 P F Pontaster R o L(D) P M 2,3 P B 1 . Interradial suture between the marginal plates: A-absent, an odd plate present in each series, the superomarginal one inset on to the disc I-often irregular, plates asymmetric or odd plates in some or all interradii but the superomarginal one on the periphery of the disc R-regular, no odd interr'adial plates (though the sutures are occasionally asymmetric in one or two interradii) 2. Papular areas: b-bilobed distally (except in small specimens, R < 30 mm), limited to the arm bases, initiated from a single median proximal primary pore d-initially double, two lateral areas on each arm base merging proximally in large specimens, R > 80 mm, to form a bilobed area, pores relatively small and interstitial o-oval and swollen with deepened plates forming well-defined median papularia w-widespread over most of disc and at least the proximal parts of the arms 3. Superomarginal plate alignment: D-partly or mainly dorsal, framing the paxillar area in dorsal view L-mainly lateral, inconspicuous dorsally 4. Abactinal plates of disc and arm bases: f-flat, usually scale like, with only a small median elevation, if anything p-(para-)paxilliform with a more or less well defined median column or convexity 5. Armament of most abactinal plates: M-multiple, with a cluster of spinules, often surrounding a larger median spinelet or spine, sometimes totally lacking R-reduced, often to a single spinule or spinelet throughout 6. Central armament of primary and some other proximal plates +-large spines 2+ mm long present — only 1 (rarely 2,3) central spinelets of larger plates enlarged up to c. 1'5 mm 7. Subambulacral spine number 8. Dorsal arm muscles: p-paired, attached proximally to one or two ambulacral plates and the adjacent marginals, or to the abactinal body wall, or both u-united in a single mass, attached proximally only to the body wall 9. Pedicellariae (if present): B-bivalved with only two elongated valves, on adambulacral plates only F-fasciculate with multiple narrow tapering valves, usually in two, occasionally three, combs on adjacent plates (pectinate), usually present on actinal plates, sometimes on abactinal and marginal plates, occasionally lacking. FAMILY BENTHOPECTINIDAE 97 interradial marginals more or less irregular in position, so that one plate of each series is often more or less mid-interradial but the armament of such odd superomarginals is not conspicuous. Inevitably, such intermediate taxa prompted him to amalgamate the Cheirasteridae with the Benthopectinidae. Apart from synonymizing Pararchaster with Benthopecten, he accepted the genera cited by Ludwig. Among the genera with no odd interradial marginals, Fisher distinguished Pontaster and Pectinaster with 'entire and swollen' papularia from Cheiraster and Luidiaster with 'flat and two-lobed' papular areas, separating Pontaster from Pectinaster on the presence of bivalved rather than fasciculate pedicellariae and Cheiraster from Luidiaster on the single rather than multiple subambulacral spines — the only difference evident from a comparison of Ludwig's diagnoses. To supplement these criteria, Fisher proposed to use the proximal attachment of the dorsal arm muscles — to one or two ambulacral plates and the two adjacent marginals in Luidiaster but to the dorsal body wall and the superomarginals in Cheiraster. However, in 1919 he had to modify this after finding that the type-species of Cheiraster, C. gazellae Studer, does have tendons to the twelfth and thirteenth ambulacral plates, though inconspicuous ones, and these plates lack the enlarged crests of the comparable plates of L. hirsutus, the type species of Luidiaster, and even more L. dawsoni. I find that L. hirsutus additionally has the attenuated proximal ends of the muscles attached to the dorsal body wall just distal to the V-shaped area defined by the two lobes of the papular area, as well as having a tendon to a laterally swollen crest of an ambulacral plate (the sixth plate in a specimen with R 85 mm). Fisher himself (1940) found a much stouter link between the main part of the muscle and the dorsal body wall in the Discovery specimen of Pontaster planeta Sladen, a species later referred to Luidiaster (A.M.C., 1962) on account of the V-shaped flat papular areas and multiple subambulacral spines, and closely related to L. hirsutus. The muscle appears to me to extend nearly to the seventh ambulacral, ending opposite the fourth superomarginal, the distalmost papular pores being just proximal to this (as in L. hirsutus), although Fisher described it as ending level with the ninth ambulacral. He also noted a 'rudimentary' tendon to the eighth ambulacral in planeta. In Cheiraster mirabilis (Perrier), which has abruptly enlarged fourth superomarginal plates, I find that there is a stout tendon to this marginal and the adjoining seventh ambulacral but no connection to the dorsal body wall. Clearly, the muscle attachment shows too great a range in these various species for use as a character of generic weight. This leaves only the number of subambulacral spines to distinguish between Cheiraster and Luidiaster but I think that the alignment of the superomarginal plates is also of more than specific significance. Hitherto, this alignment has not been regarded as important, possibly because of the considerable range of form in the very variable species Pontaster tenuispinus, so that Ludwig suggested synonymy of Pontaster venustus with dorsally broad superomarginals with Pectinaster filholi where the marginals are almost entirely lateral. However, limitation of Pectinaster to species with laterally aligned superomarginals results in a much more natural grouping, since such species also have specialized compact papularia, those with dorsally broad superomarginals having much more diffuse pore areas potentially or actually bilobed in form with the abactinal plates within the areas little modified (compare Figs la & b). Verrill (1915 : 137) discounts the importance of the subambulacral spine number in ascribing the specimen he named enoplus to Cheiraster despite the two large subambulacral spines of the large holotype (R 185 mm). Study of a good size range of C. mirabilis now shows that doubling of the subambulacral spine number is common in very large specimens, R > 120 mm, and that C. enoplus Verrill is a synonym, the holotype sharing the very distinctive abrupt enlargement of the fourth superomarginal plates and spines. With regard to the Atlantic species of Benthopectinidae, no less than 18 other nominal species which have been included in Pontaster, Pectinaster, Cheiraster and Luidiaster need to be taken into consideration, namely (in chronological order): 98 A. M.CLARK a Fig. 1 Half sections through the base of a ray bisecting the papularium or papular area, showing the armament of the second supero- and inferomarginal plates with their proximal and internal (hatched) faces, the fully developed spines of the third plates shown by discontinuous lines, (a) Pectinaster filholi Perrier, syntype of Pontaster forcipatus Sladen, Challenger st. 44, 45 or 50, BMNH no. 90.5.7.50(pt), R c.75 mm; (b) Cheiraster (Christopheraster) blakei sp. nov. Oregon st. 4294. R. c. 70 mm. [In (b) the vertical faces of the two marginals are not in the same plane; the position of the lobe of the papular area is indicated by the arrows.] FAMILY BENTHOPECTINIDAE 99 Pontaster tenuispinus (Diiben & Koren, 1 846) NE Atlantic, Arctic Cheiraster echinulatus (Perrier, 1875) West Indies C. mirabilis (Perrier, 1881) West Indies Luidiaster hirsutus Studer, 1 883 Kerguelen, South Africa C. coronatus (Perrier, 1 884) West Indies Pontaster sepitus (Verrill, 1 885) NW Atlantic Pectinaster filholi Perrier, 1885 NE Atlantic Pontaster venustus Sladen, 1 889 Azores, Cape Verde area Pectinaster pristinus (Sladen, 1 889) SW Atlantic Luidiaster planeta (Sladen, 1 889) E & W Patagonia, Falklands Pectinaster vincenti (Perrier, 1 894) West Indies Pontaster perp lexus Perrier, 1894 West Indies Pectinaster oligoporus (Perrier, 1 894) West Indies P. dispar Verrill, 1915 West Indies P. gracilis Verrill, 1915 West Indies P. mixtus Verrill, 1915 West Indies C. planus Verrill, 1915 West Indies Luidiaster dubius H. L. Clark, 1 94 1 West Indies Of these, Pontaster venustus was synonymized with Pectinaster ftlholi by Ludwig (1910), followed by Macan (1938) but is now found to be referable to Cheiraster because of its irregular bilobed papular areas and dorsally broad superomarginal plates. The oval, well- defined papularia ofP.filholi are constructed of much deeper plates than the other abactinal ones, laterally grooved to accommodate the papulae (Fig. la). The same combination of less compact actually or potentially bilobed papular areas and relatively broad supero- marginals as in P. venustus is also shown by the types of Pontaster sepitus, Pectinaster vincenti, Pontaster perplexus, Pectinaster oligoporus, Pectinaster dispar, Pectinaster gracilis and Pectinaster mixtus, all of which have been examined. These encompass a size range of R 26-c. 55 mm. Pectinaster pristinus is known only from a single small specimen, R c. 18 mm, from off the River Plate. This was in poor condition, the abactinal plates obscured by mud; when this was removed, no significant difference could be seen in the armament from specimens ofP.filholi of similar small size from the North Atlantic. Accordingly, both Pontaster and Pectinaster are reduced in the Atlantic to single species (indeed Pontaster is a monotypic genus). Following study of many benthopectinids from the U.S. National Museum and M.C.Z., Harvard collections, together with the Pillsbury and Gerda material of the University of Miami from the tropical Atlantic, I consider that most of the 1 6 remaining names listed above (which fall within the scope of Cheiraster and Luidiaster) are synonyms, only five being for valid species. This superfluity of names arose mainly from Perrier's confused treatment of the Blake collections (1881 and 1884 with afterthoughts in 1894), which was partly resolved but further complicated by Verrill's failure to take growth changes into account in his review of 1 9 1 5. One character of limited use is the R/r ratio, which tends to increase with the absolute size, so that species such as C. mirabilis, which may exceed R 120 mm, may achieve an R/r ratio of as high as 9/1, from usually 7-8/1 at Re. 100 mm or only c. 5/1 at R 20 mm. However, in C. sepitus with a maximum R of probably 70 mm, the ratio stays near 5/1 in specimens larger than R c. 30 mm. Since individual papulae do not appear to increase disproportionately in size with growth of the whole animal, their number must increase in order to maintain adequate respiratory function. In most of the species under discussion the first (primary) pore usually appears at R 10-15 mm near the base of each arm distal to the primary radial plate, which at this small size is usually still distinguishable among the other disc plates; further pores, usually smaller, are added progressively around it, especially distally and laterally. Only Cheiraster mirabilis and C. coronatus (sensu H. L. Clark, 1941) differ from the rest in developing not one primary a e Fig. 2 (a and b) Cheiraster (Christophemster) mirabilis (Perrier). (a) Atlantis st. 3444, 3445, 3447 or 3449, MCZ 3912(pt), R 25 mm, to show disc and proximal marginal spines; (b) Holotype, Blake st. 148, MCZ 10, R 87 mm, papular area (the proximal end lowermost); (c-f) Cheiraster (Christopheraster) blakei sp. nov. (c) Atlantis st. 3444, 3445, 3447 or 3449, MCZ 3912(pt), R 21 mm, for comparison with (a); (d) Atlantis st. 2963F, MCZ 3915, R 112 mm, part of intact papular area (most of the spines unnaturally appressed); (e) Blake st. 19, MCZ 2179, R c. 40 mm, second inferomarginal plate, adjoining actinal area and two adambulacral plates; (0 Blake st. 295, MCZ 216, R c. 45 mm, denuded papular area showing the spine sockets on two midradial plates, the upper one probably the primary radial; (g, h) Cheiraster (Christopheraster) horridus Fisher, holotype, Albatross st. 4079 (Hawaiian Is), USNM 21 156, R 35 mm; (g) part of denuded papular area showing varied plate size and relatively small pores; (h) armanent of one large disc plate. FAMILY BENTHOPECTINIDAE 101 pore but a pair of smaller pores offset one each side of the primary radial plate; additional similarly small pores appear laterally and distally but few develop in the proximal mid- radial area, even at a large size, so that the two lateral pore areas only become loosely linked (Fig. 2b & f)- At R > 100 mm, the total pore number in these two species exceeds 100 on each ray, far more than in the other species of Cheiraster with their larger papulae, so that for instance C. planus at R 100 mm has only c. 50 pores in each area. From the median primary pore these develop fairly regularly in this species in a curved double row on each side, forming a horseshoe-shape; only at R > 70 mm do the added pores form a more irregular and elongated area (Fig. 5). The early stages of this sequence are shown by the illustrated holotype and paratype of Pectinaster gracilis (Verrill, 1915, pi. 6, fig. 1 & pi. 15, fig. Ib); the smaller paratype (R probably c. 20 mm) has only three pores in a transverse line, the median one enlarged, while the holotype (pi. 6) at R 30 mm usually has just two additional pores placed distal to the lateral ones forming a very shallow U. Although not stated by Verrill, the holotype has most superomarginal spines distinctly longer and stouter than the corres- ponding inferomarginal spines — a characteristic of C. planus — as well as compact parapaxillae, uniformly short inferomarginal spinules giving a smooth contour to the plate (apart from the main and accessory spines), large single subambulacral spines and relatively blunt-angled furrow margins to the adambulacral plates. Consequently, P. gracilis should be treated as a synonym of C. planus. A similar sequence in the early development of pores is shown in Cheiraster echinulatus, Pectinaster vincenti and P. mixtus, also in Pontaster sepitus, P. venustus, P. perplexus, Pectinaster oligoporus and P. dispar (and almost certainly occurred in the holotype and only recorded specimen of Luidiaster dubius before it reached its final size of R c. 60 mm). In these, however, the later developed pores are more irregular in position than those of C. planus, forming at first a square or transverse patch (Fig. 4b) and then an elongated bilobed area, often asymmetrically so or with the distal pores of the two lobes approximating medially (as in the holotype of P. vincenti, Fig. 4g) to form an almost oval area. The maximum R in these nominal species does not appear to exceed 70 mm and the pore number 36. Most specimens taken are in the R 20-40 mm range. (Genital pores first appear in Cheiraster at R 10-15 mm close to the first superomarginal plates.) The holotypes of C. echinulatus, mixtus and vincenti have R 16 mm, 26 mm and 42 mm and the pore number in each area is 1-3, 3-5 and c. 12 respectively, an increase compatible with the size differences. All show the following characters: the abactinal plates are very mixed in size, polygonal, abutting fairly closely together, distinctly flattened in the two smaller specimens but somewhat convex in the larger one, especially on the papular areas, armed with numerous (up to c. 30) short 'spinules' of superficially granuliform shape (though with height/breadth >2/l in the largest) arising from most of the surface of the plate and encircling in a double ring on many plates a slender needle-like spinelet; the supero- marginal plates are broad in dorsal view, the two opposite ones together taking up c. 50% of the total arm breadth; the marginal spines are relatively slender, most inferomarginal ones longer than the corresponding superomarginal spines, while several of the inferomarginal spinules below the spine are more or less enlarged into accessory spinelets so as to give a rough profile to the body; the furrow margins of the proximal adambulacral plates form an acute angle prolonged into the furrow (the distal plates being truncated) and the sub- ambulacral spines are multiple (except perhaps in the small holotype of C. echinulatus where Perrier described a complete circle of 10-12 peripheral spines (including those projecting over the furrow, around a larger central spine but all the subambulacral spines are now [1980] lost or displaced). The coincidence of all these characters is convincing evidence that the three are conspecific, so that Cheiraster vincenti and Pectinaster mixtus become synonyms of C. echinulatus. The type localities of C. echinulatus and vincenti are near the neighbouring Lesser Antillean islands of Barbados and St Vincent; that of mixtus is off Havana, Cuba (Verrill's 'West of Florida' deriving from a misreading of the coordinates for Albatross st. 2341). Many other specimens from intermediate and western Caribbean localities to a confirmed depth of 567 metres show variation not only in the convexity of the 102 A. M.CLARK fOOO eooo doooo FAMILY BENTHOPECTINIDAE 103 abactinal plates and arrangement of the pores but also in the size of the pores, which reach a maximum number of 36 in a specimen with R 65 mm. The frequency of pedicellariae, as usual, is rather variable; there are some pectinate actinal ones is c. 50% of specimens but only c. 5% (including the holotypes of C. echinulatus and mixtus) have a few fasciculate inferomarginal pedicellariae adjoining the adambulacrals. The holotype of P. oligoporus from the Leeward Is. in 275 m is in very poor condition but its flattened abactinal plates and the prolonged furrow angles of the adambulacrals show that it too is conspecific with and a synonym of C. echinulatus. For the rest, the type localities of C. sepitus, venustus (with var. robusta), perplexus, dispar and dubius are scattered all over the North Atlantic, respectively: S of Cape Sable, Nova Scotia, 1570 metres, the Azores, 1650 metres (Cape Verde Islands, depth unknown), off Cap Blanc, W. Africa, c. 2325 metres, the Leeward Is, 1260 metres and S of Cuba, 485 metres. (For Pectinaster dispar Verrill cited the source as an Albatross station, locality unknown, but the labels now with the specimen indicate st. 2751, which is 16°54'N, 63° 12'W.) The relevant type specimens have R measurements and pore numbers as shown in table 2. Table 2 Name sepitus venustus v. robusta perplexus dispar dubius R(mm) 37 37 46 52 35+ (745-50) c.63 Pores 15 7-9 8 21 10-13 11-14 The figures for C. sepitus are from what may be called a 'topotype' in the USNM, no. 24525, since the two syntypes (from Albatross st. 2072) are now in a state of disintegration and were never properly described by Verrill. This specimen was named P. sepitus in Verrill's hand and was from a station (2075) with almost identical coordinates and depth. Although sharing with C. echinulatus the dorsally broad superomarginal plates taking up 50% or more of the arm breadth, these all differ from that species in several features. The abactinal plates are more loosely joined and consistently markedly convex and parapaxilli- form, with more elongate and somewhat pointed spinules usually numbering only 5-15, clustered together on the central (highest) part of the plate, though again often surrounding a central spinelet several times longer; the superomarginal plates, though with a similar mean breadth to those of C. echinulatus, are somewhat longer and often more convex, numbering only c. 1 6 at R 20 mm or c. 27 at R 50 mm, compared with c. 20 and c. 30 at the same sizes in echinulatus; the marginal spines are shorter and much stouter basally, being conical in shape, though again the inferomarginal spines tend to be longer than the superomarginal ones and the upper inferomarginal spinules are likewise more or less elongated; the subambulacral spines are usually single, stout-based and conical (as usual in all these species similar in form to the marginal spines) and pedicellariae are rarely present at all, even on the actinal plates they were only found in less than 10% of specimens. Possibly in larger specimens, R > 50 mm, the arms remain relatively shorter than in C. echinulatus, R/r not much exceeding 5-0/1 where echinulatus may reach 6/1, but this remains to be seen from Fig. 3 Cheiraster (Barbadosaster) echinulatus (Perrier) (a) Holotype, Hassler, Barbados, Paris Museum (PM) 3481, R 16mm, disc and arm bases from above, the primary radial and interradial plates stippled, crystal bodies indicated on a few larger abactinal and marginal plates (the centre of the disc rubbed almost bare of spinules); (b, c) the same, part of a ventral interradius and the fourth and fifth adambulacrals with adjacent inferomarginal plates, most armament lost, the subambulacral spine drawn on the fifth plate appressed; (d-f) abactinal spinules of: (d) the holotype of C. echinulatus, (e) holotype of Pectinaster mixtus Verrill, R 26 mm and (0 specimen from Gerda st. 628, R 66 mm. The scale equals 2 mm for (a), 1 mm for (b) and (c) and c. 0-5 mm for (dHO- 104 A.M.CLARK further intact specimens. (Perrier's implication that there is more than one subambulacral spine in the holotype of perplexus is not borne out by re-examination. It may also be noted that he gave the pore number as 'une quinzaine' but removal of the spinelets in one area revealed the higher number of 2 1 . Yen-ill's estimate for the type of dispar was also too low). With all these features in common, I do not think that the wide spread of pore numbers from 2 1 at R 52 in the type of perplexus to only 8 in that of the geographically close venustus var. robusta at R 46 mm is significant. The 'topotype' of seprtus has as many as 1 5 at R <40 mm. Other similar specimens, with R 19-46 mm collected at five Pillsbury stations off Venezuela and two Discovery ones from the Azores, all of which I consider are conspecific with C. sepitus, have 0-12 pores in individual areas, often with a range of three or more in different radii of the same specimen, one with R 41 mm having from 6 to 12 pores in the five radii. Consequently, P. venustus, perplexus, dispar and L. dubius are now referred to the synonymy of C. sepitus, the oldest name. All the West Indian Pillsbury specimens now referred to C. sepitus are from depths greater than 1000 metres, as with the type locality of P. dispar, which casts doubt on the depth record of only 485 metres for the holotype of L. dubius. A further nomenclatural problem remains to be resolved concerning the two species generally known as C. mirabilis and C. coronatus, at least sensu H. L. Clark, 1941 . In his preliminary account of the Blake asteroids, Perrier (1881) described Archaster mirabilis (stations unspecified) firstly (paragraphs 1-4) from a specimen with R/r given as 87/12 mm, only 6-10 spines around the anus and otherwise granulous abactinal plates; he then commented on another specimen with R/r 45/12 mm. In the full report (1884) the first specimen is redescribed at the beginning of the account of A. mirabilis, the station given as 148, St Kitts, and the second specimen is separated off as holotype of Archaster insignis (since referred to Dytaster in the Astropectinidae), as noted by Verrill (1915). Other specimens from 46 different stations were also included under the name mirabilis but the one with R 87 mm from st. 148 must be the holotype. Also in 1884, Perrier described Archaster coronatus from a specimen with R 65 mm, R/r 8/1, citing stations 2 and 19. The sample from st. 19 is in the MCZ, Harvard, labelled as syntypes but the largest specimen is badly broken and its R probably did not exceed 40 mm, so, as Verrill (1915) assumed, the holotype must be from st. 2, off Havana, Cuba, and in the Paris Museum. In 1894, Perrier inadmissably redescribed the holotype of mirabilis (R/r slightly modified to 85/10) as 'C. coronatus premier type', as Verrill (1915) noted when he repeated Perrier's 1884 description for coronatus, ranking it as a subspecies of mirabilis. H. L. Clark (1941) satisfactorily distinguished two West Indian species ofCheiraster with spines several mm long on the disc in the following terms: mirabilis coronatus 1. Paxillae very low and closely covered 1. Paxillae more convex with fewer and by a group of minute spiniform more unequal granules, one (or granules, none of which are enlarged or sometimes more) of them elongated lengthened into spinules [spinelets in into a spinule. my terminology.] 2. 3-10 relatively conspicuous spines 2. Spines up to 3-4 mm long but never 3-8 mm long at the centre of the disc. forming a conspicuous isolated group. 3. Fourth superomarginal spine up to 3. Fourth spine up to 5 mm long. 10-12 mm long. After examining the Blake and Atlantis specimens in the MCZ and the Gerda and Pillsbury material I can add the following: 1 a. Abactinal plates too low to be called 1 a. Armament of abactinal plates 'paxilliform'; armed with up to consisting of up to c. 25 (mean c. 10) c. 35 (mean number c. \ 5) short distinctly elongated radiating spinules, pointed spinules in compact groups. the longer central spinelets graduated to the disc spines. FAMILY BENTHOPECTINIDAE 105 2a. The big disc spines closely clustered 2a. in the centre, not extending to the primary radial plate on the arm base; usually 5-15 spines, rarely >20. 3a. Fourth (rarely the fifth) superomarginal 3a. plate and spine abruptly enlarged, the corresponding inferomarginal plate reduced and often spineless. 4. Marginal plates covered with fairly 4. uniform short pointed spinules, only a few on the lower face of the inferomarginals enlarged to spinelets; profile superficially smooth. 5. Furrow margins of adambulacral plates 5. rounded or with obtuse angles. Big spines spaced, extending on to the papular areas in larger specimens and graduating to the central paxillar spinelets, usually > 10 even at R only 30mm, up toe. 75atR 100mm. In larger specimens the fifth or sixth marginal spines the largest and graduated to the adjacent spines. Several more or less elongated spinules/spinelets at the upper end of the inferomarginals giving a shaggy appearance and rough profile. Furrow margins 90° or acute, angles prolonged into the furrow. Unfortunately, Perrier's 1884 description of C. coronatus described the abactinal armament as including c. 15 spines around the anus and the fourth superomarginal spine as much larger than the others (my italics). Although 1 5 disc spines is a high number for C. mirabilis at R 65 mm it is not unprecedented (see Table 4). Accordingly, it seems unavoidable that the name coronatus should be referred to the synonymy of mirabilis, as was done by Downey (1973), following Ludwig (1910), though evidently without appreciation of the existence of the two species recognized by H. L. Clark. A new name is therefore needed for the species with less conspicuous, more numerous and graduated disc spines and graduated superomarginal spines. Although H. L. Clark has the credit for distinguishing the species, knowing his antipathy to the use of personal names I propose the name C. blakei after the ship which first collected it. As the MCZ specimens from Blake station 19 are in poor condition, a large specimen from Gerda st. 918 (between Florida and the Bahamas), described on p. 113 has been selected as holotype. Turning to the species of Luidiaster (other than L. dubius disposed of above), the smallest specimen available of L. hirsutus from South Africa has R c. 35 mm (R/r 4*5/1) and c. 30 pores/area but a primary median pore is distinguishable, as in Cheiraster sensu stricto. The pore areas are V-shaped with mostly two rows of pores in each limb whereas larger specimens, R c.80 mm, have 1 10-160 pores in a more irregular but still bilobed area. The type series of L. planeta from Chile, R 26-37 mm, shows 9-14 pores and the primary ones are distinct. These two southern species, L. hirsutus and L. planeta, both have the superomarginals inconspicuous in dorsal view, aligned mainly on the sides of the arms, unlike the species of Cheiraster. They also have two or more large subambulacral spines (though at R 35 mm in L. hirsutus the two spines are more or less markedly unequal) and H. L. Clark (1926) found single spines in his smallest specimen, R 16 mm). However, the West Indian Cheiraster echinulatus has two or even three subambulacral spines from a small size, R. c. 20 mm, while large specimens of C. mirabilis and C. blakei (R > 100 mm) often develop a second spine, though most individuals have only one on most plates. (Second spines usually appear first on the distal half of the arm and on the very first plate.) Although large specimens of C. mirabilis and blakei may have the superomarginal plates relatively narrow dorsally, they still form a small border to the paxillar area and in C. echinulatus they are consistently broad. I doubt whether the alignment of the superomarginal plates unsupported is of sufficient weight to justify a generic distinction for Luidiaster hirsutus so that Luidiaster is now treated as a subgenus of Cheiraster. Both names were introduced by Studer, 1883, who separated them, and even referred Luidiaster to the Astropectinidae, because the specimens of Cheiraster gazellae had the dorsal body wall sufficiently transparent to show the underlying muscles, which he took for gonads. However, he failed to name a type species for Luidiaster until 1 884, from which date only the name is valid. 106 A.M.CLARK The main characters by which the seven Atlantic species of Cheiraster can be distin- guished are summarized in Table 3. (The number of adambulacral plates corresponding to the first ten inferomarginals does not seem to me to be significant in Cheiraster, though it may be useful for certain species of Benthopecten.) Synonymies and ranges for five of these species are given on pp. 1 12-8. Table 3 Tabular key to the Atlantic species of Cheiraster sensu lato. Brackets signify occasional or partial conditions. 1 2 3 4 5 6 7 8 9 10 11 12 blakei 6-0-8-5/1 20/27/33 Me D e I g (R) 1(2) AM(D) c. 130 mirabilis 7-5-8-5 20/30/37 L (0 D a I u R 1(2) AM(D) 185 hirsutus 5-5-7-0 20/27/c. 35 A c PC e I g (P) 2,3(1) (A) 92 planeta c. 5 20/35/742 A c PC e I g R,* P 2,3 (1,4) (A,D) 100 planus 6-0-7-0 18/25/30 A c PC e S u R 1 A(M) c. 100 sepitus 4-5-5-5 16/25/30 A c PB e I g (P) 1 ((A))t c. 60 echinulatus 4-8-6-0 19/27/34 A f PB e I g P 2(1-3) A(M) c. 70 *C. planeta. In the type series from S Chile, the furrow margins are acute but not in the Falkland Is specimen. fC. sepitus. Only five out of 60+ specimens examined have any actinal pedicellariae. 1. R/ratR50+mm. 2. Number of superomarginal plates at R 20, 40 and 60 mm (some obtained by extrapolation) 3. Disc spines: L-few and relatively very large, rarely > 15, restricted to centre of disc; no central paxillar spinelets graduating to the spines: M-moderately large centrally but not more than ^ r, spaced and often extending on to the papular areas, graduating in size down to the central paxillar spinelets A-absent, only central paxillar spinelets on many disc and mid-radial arm plates, not > 1 • 5 mm long 4. Proximal abactinal plates: c-convex, parapaxilliform, butting loosely together, spinules clustered on the raised area, mean no. 10-12 or less f-flat or low convex, polygonal, butting more or less closely together, spinules covering most of the surface and more numerous, mean no. 1 5-20 5. Papular areas: PB-initiating from a single large median primary pore, becoming irregularly bilobed, sometimes indistinctly so PC-initiating from a single large median primary pore, becoming crescentic (horseshoe-shaped), V-shaped or bilobed with long lobes D-double, initiating from a pair of pores, one each side of the mid-line, forming two diffuse lobes, merging proximally only in large specimens, R > 70 mm 6. Largest marginal spines: a-Fourth superomarginal spine abruptly and conspicuously enlarged, fourth inferomarginal plate reduced and often lacking a spine e-Marginal spines fairly even in size or the fourth to seventh (according to size, only the fourth in smaller specimens) enlarged but graduated 7. Marginal spines, relative length: I-most inferomarginal spines longer than the corresponding superomarginal ones S-most superomarginal spines longer (at least on the proximal half of the arm) 8. Inferomarginal spinules at the upper end of the plate: g-graduated, several becoming more or less elongated, graduating to the accessory spine(s) or spinelets, profile rough u-fairly uniform, only some enlarged spinelets on the lower face of the plates, profile fairly smooth between the spines FAMILY BENTHOPECTINIDAE 107 9. Furrow margins of adambulacral plates: P-with acute angles prolonged into the furrow, < 90° R-rounded or with angles obtuse or right-angled 10. Subambulacral spine number 11. Pedicellariae: A-actinal M-marginal D-dorsal 12. Maximum recorded R As a result of these comparisons between the Atlantic species, two further supraspecific groups can be recognized within the genus Cheiraster. Firstly, the essentially dual nature and different ontogeny of the papular areas in Cheiraster mirabilis and C. blakei compared with the other species of Cheiraster, in my opinion justifies at least a subgeneric distinction for these two species. The relatively small size of the papulae themselves, the marked size difference of some of the plates (including the primary ones) on the disc and arm bases and the development of spines several millimetres long on some of these plates provide supporting characters. At least two of the non- Atlantic species currently listed as Luidiaster can be placed in this subgenus, for which I propose the name Christopher aster. These are: Pontaster oxyacanthus Sladen, 1889, from Japan, which is very similar in abactinal armament to C. blakei and likewise has the small papular pores almost entirely offset from the midradii, and Cheiraster horridus Fisher, 1906, from the Hawaiian Islands, also with very small lateral papulae, a marked size difference of some of the abactinal plates and more or less enlarged spines on the larger plates (Fig. 2g & h). The holotype of oxyacanthus at R 73 mm has two large subambulacral spines on most plates, while that of horridus at R 35 mm has two subequal spines, so that the tendency to develop a second subambulacral spine, more or less accelerated in different species, may also be used to distinguish the subgenus. Secondly, the unusual combination in C. echinulatus of rather low, polygonal, closely- fitting abactinal plates armed over most of their surface w'th relatively numerous short spinules, dorsally broad superomarginals, unlike Cheiraster (Luidiaster), usually slender marginal spines, unlike the conical ones of Cheiraster (Cheiraster), and multiple sub- ambulacral spines in all but the smallest specimens (R <20 mm), also justifies a subgeneric distinction I believe. The name Barbadosaster (again reflecting the type locality of the type species) seems appropriate. In this case the subgenus appears to be monotypic. One non- Atlantic species of Cheiraster — Pontaster teres Sladen, 1889, from the Banda Sea, East Indies, superficially resembles C. (Barbadosaster) echinulatus in the slender marginal spines and multiple subambulacrals but differs markedly in the pore areas. CHEIRASTER Studer Archaster(pt.) Perrier, 1875 : 348 [1876 : 268] (A. echinulatus); 1884 : 256; Verrill, 1885a : 151. Cheiraster Studer, 1883 : 129; 1884 : 49-50; Perrier, 1885 (pt.) : 70; 1894 : 275; Fisher, 1906 : 1040; Ludwig, 1910:440, 454^57; Fisher, 1911:120, 123; Verrill, 1915:123-124; Fisher, 1919: 190-191; H. L. Clark, 1941 : 26; Downey, 1973:40^1; Jangoux, 1978:95, Walenkamp, 1979:23. Luidiaster Studer, 1883 : 131 [nom. nud., no species named]; 1884 : 46^7; Ludwig, 1910 : 440^*41, 451-454; Fisher, 1911 : 127; Verrill, 1915 : 149-150; Fisher, 1940 : 96; H. L. Clark, 1941 : 28. [Type species: Luidiaster hirsutus Studer, 1884.] Pontaster (pt.) Sladen, 1885 : 610; 1889 : 23-27; Verrill, 1894 : 247; 1895 : 131. Acantharchaster Verrill 1894 : 268-269. [Type species: Archaster dawsoni Verrill, 1880.] Pectinaster: Verrill, 1915:137-139; Downey, 1973:43; Walenkamp, 1979:26. [Non Pectinaster Perrier, 1885.] TYPE SPECIES. Cheiraster gazellae Studer, 1 883, by subsequent designation by Ludwig, 1910. 108 a g • • h-*« • • ..• • • *o- e" •.Q. - ••ff ** CF • • • • • •Q- q . m n • • • • P • • • • • • • V*. Fig. 4 Pore arrangements, showing presumed primary radial plates, (a-h) Cheiraster (Barbadosaster) echinulatus (Perrier). (a) Blake st. 45, MCZ 2181, R 27 mm; (b) Albatross st. 2322, USNM 10109, R c. 30 mm; (c) Blake st. 36, MCZ 24, R c. 20 mm (with proximal superomarginals); (d-f) Atlantis st. 298 1C or 2982C, MCZ 3917, R 48, 47 and 51 mm; (g) holotype of C. vincenti Perrier, Blake st. 231, R c. 42 mm; (h) Atlantis st. 3480 or 3482, MCZ 3924, R c. 55 mm. (i-r) Cheiraster (Cheiraster) sepitus (Verrill), (i) 'topotype', Albatross st. 2075, USNM 24525, R c. 37 mm; 0) holotype of Pectinaster dispar Verrill, Albatross st. 2751, USNM 18468, R c. 50mm; (k) Albatross st. 2531, USNM 12062, R 35-40 mm (with proximal superomarginals); (1) holotype of Pontaster perplexus Perrier, Talisman 1883, st. 96, PM 3378, R 52 mm; (m, n) holotype of Luidiaster dubius H. L. Clark, Atlantis st. 3326, MCZ 38 1 5, R c. 60 mm; (o, p) Princesse Alice st. 743 (Azores), MCZ 2689, R 22 mm; (q) same sample (named Pontaster venustus by Koehler, 1909), R c. 33 mm; (r) Talisman 1883, st. 93, MCZ 30, R c. 45 mm (named P. venustus by Perrier, 1894). FAMILY BENTHOPECTINIDAE 109 (In fact, the spelling of the specific name in 1883 was Gazella, emended in 1884 to the genitive.) A genus of Benthopectinidae without odd interradial marginal plates; abactinal plates usually parapaxilliform, more or less convex but flatter on the arms and sometimes all low, armed with several spinules, in most species many midradial and disc plates with a larger central spinelet, sometimes a spine > 2 mm long; papular pores limited to a bilobed area at the base of each arm, or sometimes a pair of lateral areas loosely integrated at their proximal ends medially, rarely the two lobes not well defined or asymmetrical, undeveloped in younger animals, R 50 mm) by its larger size and position, which is at first distal to the primary radial plate though may become dissociated from it as other plates develop, the later developed pores at first forming a median patch and then extending into a distal lobe on each side; superomarginal plates forming a more or less broad frame to the paxillar area in dorsal view; marginal spines conspicuous, conical or spike-like, some species with a second but smaller inferomarginal spine; adambulacral plates predominantly with single subambulacral spines, similar in shape to the marginal ones, though the first plate and some distal plates may develop a second large spine. TYPE SPECIES. Cheiraster gazellae Studer, 1883. Philippines — NW Australia: 360-470 metres. OTHER SPECIES INCLUDED. In chronological order: Archaster sepitus Verrill, 1885 Pontaster pilosus Alcock, 1893 C. inops Fisher, 1906 C. snyderi Fisher, 1 906 C. niasicus Ludwig, 1910 C. ludwigi Fisher, 1913 C. triplacanthus Fisher, 1913 C.planus Verrill, 1915 C. diomedeae Fisher, 1917 C. weberi Doderlein, 1 92 1 C. otagoensis McKnight, 1973 North and Central Atlantic South of India Hawaiian Is to Sri Lanka Hawaiian Is-Nicobar Is Philippines-Celebes Indonesia Celebes & South-east Africa Gulf of Mexico-French Guiana Philippines Indonesia South New Zealand 485-3700 metres 1080-1 3 10 metres 430- 1250 metres 420- 16 70 metres 485-930 metres 12 70 metres 1050- 1280 metres 550-715 metres 770 metres 470 metres 1000- 12 80 metres (A specimen from offNatal with R c. 40 mm referred to C. triplacanthus (A. M. Clark, 1977), 110 A.M.CLARK has distinctly enlarged primary papular pores; Fisher (1919) noted only that the pore areas in the holotype (R 49 mm) are 'double and joined only at the adcentral end'.) Subgenus LUIDI ASTER Studer, 1884 A subgenus of Cheiraster with abactinal plates parapaxilliform, even those on the disc rounded and abutting loosely together or more or less spaced, most of the larger plates, even laterally on the arms, armed with a central spinelet among the smaller spinules; papulae including a median proximal primary one but this probably soon became indistinguishable as the number of pores increases and the areas spread, though are still bilobed; supero- marginal plates aligned mainly laterally, inconspicuous dorsally; marginal spines con- spicuous, conical or spike-like, usually at least two on most inferomarginal plates; adambulacral plates with two or more large subambulacral spines, the furrow angle variable. TYPE SPECIES. Luidiaster hirsutus Studer, 1884. Kerguelen, South Africa & Bouvet I.; 240-550 metres. OTHER SPECIES INCLUDED. In chronological order: Archaster dawsoni Verrill, 1 880 Bering Sea-British Columbia 1 00-290 metres PontasterplanetaSladen, 1889 S. Chile, Argentina-Falklands 350-450 metres Cheiraster gerlachei Ludwig, 1 903 Antarctic 1 60-8 1 0 metres Pectinaster robustus A. H. Clark, 1917f Chile 2410 metres Luidiaster tuberculatus Djakonov, 1 929 Japan Sea 1 40-850 metres tSeep. 121. Subgenus BARBADOSASTER nov.* A subgenus of Cheiraster with most proximal abactinal plates polygonal in outline, fitting closely together and not markedly convex, sometimes almost flat, armed with relatively numerous short spinules, on the larger plates often forming two rings around a central spinelet; papulae including a median proximal primary one, usually distinct from the rest by its size and position, the later developed pores forming at first a median patch, then extending distally into a lobe on each side, though the two lobes may be unequal or ill- defined, hardly separated medially; superomarginal plates forming a broad frame to the paxillar area in dorsal view; marginal spines slender, not very stout basally, single, though some enlarged spinelets develop below the inferomarginal spine; most adambulacral plates with more than one subambulacral spine, except in very small specimens, R < 20 mm, the furrow angle of the proximal plates markedly acute. TYPE SPECIES. Archaster echinulatus Perrier, 1875. South Florida — Caribbean; 150-570 metres. Monotypic. Subgenus CHR1STOPHERASTER nov.* A subgenus of Cheiraster with the abactinal plates parapaxilliform (but in C. mirabilis only slightly convex), distal plates more or less spaced, the proximal plates mostly fitting fairly closely, very variable in size, the primary central, radial and interradial plates particularly enlarged, some or most of the larger plates armed with a central spine >2 mm long, often conspicuous, surrounded by small spinules, some other plates may have a central spinelet *Named after the type localities of the relevant type species: Barbados for C. (B.) echinulatus and St. Christopher's (Kitts) for C. (C.) mirabilis. FAMILY BENTHOPECTINIDAE 1 1 1 but not in C. mirabilis; no primary midradial papular pore, the first-formed papulae a pair, each latero-distal to the primary radial plate, later pores developing laterally and distally, the two pore areas on each ray only loosely meeting proximally in larger specimens to form a single diffuse bilobed area with few, if any, midradial pores, the pores themselves relatively small and hardly visible with the naked eye, relatively numerous (c. 100 at R 100 mm on each ray); superomarginal plates partly aligned dorsally, forming a narrow frame to the abactinal area but becoming inconspicuous in the largest specimens (R > 120 mm); marginal spines large and spike-like; adambulacral plates with one or two large subambulacral spines, sometimes two only in the largest specimens, furrow angles variable in prominence. TYPE SPECIES. Archaster mirabilis Perrier, 1881. Florida — Windward Is; 380-730 metres. OTHER SPECIES INCLUDED. In chronological order: Cheiraster blakei sp. nov. Florida-Northern Brazil 250-1030 metres Pontaster oxyacanthus Sladen, 1 889 Southern Japan 630 metres C. horridus Fisher, 1906 Hawaiian Is 260-330 metres This leaves a residue of the following species of uncertain affinities which have been referred to Cheiraster or Luidiaster. Cheiraster pedicellaris Studer, 1883. NE of Australia; 1000 metres. The holotype has R c. 30 mm and is probably a young specimen of C. gazellae according to Ludwig (1 9 10). Pontaster subtuberculatus Sladen, 1889. SE of Australia; 1740 metres. The holotype has R only c. 24 mm. Like Christopher aster it has no mid-radial papulae, the 17-21 pores forming two irregular longitudinal rows on each side, but the pores are relatively large and there is no sign of any enlarged spinelet or spine on any abactinal plates. The arms are unusually short for a Cheiraster and the single marginal spines are remarkably short and blunt. Pontaster teres Sladen, 1889. Indonesia; 260 metres. The holotype has R 42mm and superficially resembles C. (Barbadosaster) echinulatus, the superomarginals being relatively broad and the marginal spines unusually slender. However, denuding part of the paxillar area shows an unusual combination of larger convex plates with flattened inter- stitial ones between, fitting closely together except where the relatively large papular pores are developed, at the sides of the arm bases. The subambulacral spines are multiple. Pontaster trullipes Sladen, 1889. Philippines; 1920 metres. The holotype has R only 27 mm. The primary radial and interradial plates are enlarged, as in Christopheraster, but no large spines occur and the papulae form a median group, in which no primary one is distinct. Pontaster cribellum Alcock, 1893. Laccadive Sea; 2200 metres. The holotype has R 45 mm. It is said to resemble P. subtuberculatus in lacking spinopaxillae and having relatively broad superomarginals but the papularia are small and compact, with only 5 or 6 papulae in each and the marginal spines are sharp. Marcelaster antarcticus Koehler, 1907 (referred to Luidiaster by Fisher, 1940). South Orkney Is.; 3250 metres. R in the type series is up to only 33 mm. The proximal plates appear rather variable in size in the poor drawing and there are some relatively large disc spines but the papulae were not described at all so the position must remain uncertain. Cheiraster granulatus Ludwig, 1910. 'Indian Ocean'; depth not given. Ludwig's Valdivia report was never completed and the only information on this barely established species is in the key distinguishing it from C. subtuberculatus by the numbers of inferomarginal and furrow spines and the occurrence of actinal pedicellariae, which could all be correlated with a size difference, if it was appreciably larger than the small type of C. subtuberculatus. 112 A.M.CLARK Cheiraster yodomiensis Goto, 1914 (referred to Luidiaster by Hayashi, 1973). Japan; 160 metres. The holotype has R 1 70 mm and closely resembles a Nearchaster in the many disc spines, laterally aligned superomarginals and great extent of the papular pores along the sides of the paxillar areas. Also the interradial superomarginal spines are relatively large and possibly asymmetrical in some interradii, judging from the photograph. Cheiraster richardsoni Fell, 1958. Cook Strait, New Zealand; 730 metres. The holotype has R 3 1 mm. The superomarginals are broad dorsally and the subambulacral spines single. The papulae were undescribed. Cheiraster monopedicellaris McKnight, 1973. Northern New Zealand; 720-770 metres. The holotype has R 39 mm. The superomarginal plates are very broad but some abactinal plates have enlarged spinelets, unlike C. subtuberculatus. The 3-5 papulae are said to be in a more or less V-shaped arrangement. The complex synonymies of the North and Tropical Atlantic species are as follows: Cheiraster (Barbadosaster) echinulatus (Perrier) Figs 3a-f, 4a-h Archaster echinulatus Perrier, 1875 : 348 [1876 : 269]; 1884 : 263-264, pi. 10, fig. 4. Archaster mirabilis (pt) Perrier, 1 884 : 258-259 (sts 1 43 & 23 1 ). ICheiraster mirabilis: Perrier, 1 894 : 276-278, pi. 20, fig. 4. Cheiraster vincenti Perrier,.! 894 : 275-276. Cheiraster echinulatus: Perrier, 1894 : 278; Verrill, 1915 : 129-130 (non pp. 131-133, pi. 14, fig. 3, pi. 19, fig. 2, pi. 25, fig. 1 ? = C. blakei); Downey, 1973 : 42, pi. 13, figsC, D. Pontaster oligoporus Perrier, 1894 : 293. Luidiaster vincenti: Ludwig, 1910: 452, 453. Pectinaster echinulatus: Ludwig, 1910 : 449. Pectinaster mixtus Verrill, 1915 : 140-145, pi. 6, fig. 2, pi. 15, fig. 2, pi. 17, fig. 1; Macan, 1938 : 349 (listed); Downey, 1973 : 43^4, pi. 14, figsC, D. Pectinaster vincenti: Verrill, 1915 : 139-140; Macan, 1938 : 349 (listed). Pectinaster oligoporus: Verrill, 1915 : 147-148; Macan, 1938 : 349 (listed). Luidiaster mixtus: H. L. Clark, 1941 : 29-30; A. H. Clark, 1954 : 375 (listed). TYPE LOCALITY. Barbados, 183 metres. RANGE. Gulf of Mexico, Florida Strait, south to Nicaragua and Venezuela, east to the Leeward Islands; 150-570 (7586) metres. (The Atlantis specimen from 1000 fathoms E of Cuba said by H. L. Clark (1941) to show some anomalous features, is referable to C. sepitus, having convex spaced abactinal plates with relatively few spinules, stout conical marginal spines, single conical subambulacral spines and relatively few, c. 6, furrow spines in obtusely angled series.) Cheiraster (Christopheraster) mirabilis (Perrier) Fig. 2a, b Archaster mirabilis Perrier, 1881 : 27 (paragraphs 1-4 only); 1884:256-258, pi. 9, fig. 4 (non pp. 258-259, sts 231 & 143, pi. 10, fig. 3 = C. echinulatus)^. 10, fig. 2 (or = C. blakei). Archaster coronatus Perrier, 1884 : 262-263 (non st. 19 = C. blakei). Cheiraster coronatus: Perrier, 1894 : 271-275; Ludwig, 1910 : 455,456. Cheiraster mirabilis (pt) Verrill, 1915: 124-127 (non pi. 14, fig. 5 = C. blakei; non C. mirabilis: Downey, 1973, Walenkamp, 1979 = C.planus). Cheiraster enoplus Verrill, 1915 : 135-137, pi. 18, fig. 1 (non C. enoplus: Downey, 1973 nee Luidiaster enoplus: H. L. Clark, 1941 = C blakei) TYPE LOCALITY. St. Kitts, Leeward Islands, 380 metres. RANGE. SE Florida to the Windward Islands (records from the Gulf of Mexico probably mistakes for other species); 380 (7265) — 1470 (usually 400-700) metres; deeper records, e.g. FAMILY BENTHOPECTINIDAE 1 1 3 Blake st. 196, \030fathoms, may be for C. sepitus. Uncertain identity precludes inclusion of lesser Blake station depths. Cheimster (Christopheraster) blakei sp. nov. Figs Ib, 2c-f Archaster coronatus (pt) Perrier, 1884 : 262 [st. 19 only]. lArchaster mirabilis (pt) Perrier, 1 884, pi. 10, fig. 2 (non A. mirabilis Perrier, 1 88 1 ). Cheiraster mirabilis (pt) Verrill, 1915, pi. 14, figs 5, 5a. Cheiraster echinulatus(pt)Verri\l 1915 : 131-133, pi. 14, fig. 3, pi. 19, fig. 2 [?pl. 25, fig. 1]. Cheiraster coronatus: H. L. Clark, 1941 : 26-27; A. H. Clark, 1954 : 375 (listed) (non C. coronatus: Perrier, 1894, Ludwig, 1910 = C. mirabilis). Luidiaster enoplus: H. L. Clark, 1 94 1 : 28-29 (non C. enoplus Verrill, 1 9 1 5 = C. mirabilis). Cheiraster enoplus: Downey, 1973 : 42-43, pi. 14, figs A, B. DESCRIPTION OF HOLOTYPE. R of the only entire original arm is 1 1 8 mm (two arms with regenerating tips are 95 mm or less); r is 12 mm; R/r = 9-8/1. Arm breadth at the fifth supero- marginals is 9'2 mm and the paxillar breadth here 6*0 mm, or 65%. The arms taper hardly at all for the first third of their length. There are about 35 spines more than 2 mm long on the disc based on enlarged abactinal plates including the primary radial and interradial plates and others. An additional 10-12 smaller spines/spinelets arise in the papular area on the base of each ray, mostly on the row of plates between the two main irregular lines of pores each side of the midradial line. The longest central spines are 3 '6 mm long. Only a few other abactinal plates have an enlarged central spinelet among the small slender tapering spinules, length : basal breadth c. 2-5 : 1. These number usually 6-14, according to the plate size, mean c. 10, but the enlarged disc plates have up to 20 spinules around the central spine. All the abactinal plates are markedly convex and may be called parapaxilliform; they are rounded or slightly polygonal and only butt loosely together over most of the disc, while those on the arms are spaced from each other. The papulae number c. 75 on each ray; only one or two proximal ones (if any) are midradial, the rest being arranged in two distally diverging bands extending from alongside the primary radial plate to level with the distal ends of the third superomarginal plates. There are 55 superomarginals on the intact original arm, framing the paxillar area narrowly, each armed with a single spine of moderate size, the longest on the fifth and sixth plates up to 4-0 mm long, basal breadth c. 0*7 mm. The corresponding inferomarginal spines are similar or slightly longer, with an accessory spine about two-thirds as long below it, often also a third still smaller spine on the proximal plates besides several slender spinelets on the ventral face of the plate among the fine spinules which are of various sizes; the whole armament gives a hirsute or shaggy appearance to the plate as viewed from below. The actinal areas are small and triangular with four or five plates along the midinterradial line on each side of which the irregular adradial series numbers only about seven plates, those bearing the pedicellariae enlarged. The actinal armament is very sparse and mixed in size with only about two spinelets/spinules on most plates. Only 13^-15 adambulacral plates correspond to the first 10 inferomarginals. On each adambulacral no more than eight spines project over the furrow, their bases usually forming an angle of c. 90° but this may be blunted. The subambulacral spines are single, much larger and accompanied by only a few small spinules on the ventral face of the plate. The oral plates have 10 or 11 furrow spines, the apical one much enlarged and the second one intermediate. Probably only one suboral spine is enlarged but the plates are badly rubbed and most spines lost. Pectinate pedicellariae are developed on the adradial actinal plates (usually four in each interradius), on the ventral sides of the inferomarginals adjacent to the adambulacrals (for about three-quarters of the arm length) and abactinally (but only two are present near the centre of the disc). VARIATIONS. Some indications of variations and growth changes can be derived from Table 114 A. M.CLARK u 'o 00 c 8* £ 6 C u go. E-o U C 3« C/3 — 03 u_ j^ £ 03 O 0) 1-1 •^ C/3 C '5. 5 6 c )C o VI •5 O OJ • C 2 O l*> i! O o 2 °- •^too O O oo — o .oooooor^t^ 8 5 r,°\ r-'cxT o O >/^ >o o — 'fNOONOooor~-dor-~- O — ^ o t — M r*^ I — 0-CvOO ON f 116 A.M.CLARK 5. The disc is usually relatively larger than in the holotype and the arms more often taper evenly from the base, with the superomarginals forming a somewhat wider frame to the paxillar area, though never making up as much as 50% of the arm breadth. Often many abactinal paxillae have an enlarged central spinelet, graduating in size to the small spines on the papular areas. The relative number of papulae also varies; one specimen has well over 100/ray even though R does not exceed 100 mm. In small specimens, the first one or two pores appear at R c. 16 mm (at the same time as the genital pores which are adjacent to the first superomarginal plates). 20 specimens with R 14-19 mm show a range in pore/ray number of 0-1 1 but with a mean of only 3. The marginal spines may be relatively larger than in the holotype, up to c. 6 mm long at R > 100 mm. In small specimens the fourth superomarginal spine is the largest but the fifth and sixth ones grow allometrically to equal and then exceed it. The number of adambulacral plates corresponding to the first ten inferomarginals is usually more than in the holotype, often 16 or up to 18 being noted. The bases of the furrow spines usually form a more pronounced angle projecting over the furrow. Finally, the occurrence of pedicellariae varies; some specimens may have more numerous abactinal ones; also at least one was observed with inter-superomarginal pedicellariae inter- radially. AFFINITIES. Cheiraster (Christopher aster) blakei is closely related to C. (C.) mirabilis. Most of the differences are expressed in the tabular key (Table 3), the numerical tables here (Tables 4 and 5) and the comparison on pp. 104-5. These can be summed up as differences in armament, C. mirabilis having the enlarged abactinal and marginal spines relatively larger but fewer than C. blakei and the abactinal and inferomarginal spinules short and more nearly uniform. Additionally, the furrow spines tend to be more numerous in C. mirabilis, their arrangement on the furrow margin less angular and the contours of the abactinal plates lower. TYPE LOCALITY. Gerda st. 918, 26°03'N, 78°05'W, between Florida and the Bahamas, 804-822 metres. To be deposited in the U.S. National Museum. RANGE. Recorded from SE of Florida and the coasts of Cuba east and south to the Lesser Antilles and northern Brazil at least as far east as 44° W; 250 (?225)-1033 (but mostly 500-800) metres. Cheiraster (Cheiraster) planus Verrill Fig. 5a, b Cheiraster planus Verrill, 1915 : 133-1 35, pi. 18, fig. 2. Pectinaster gracilis Verrill, 1915 : 145-147, pi. 6, fig. 1, pi. 14, fig. 4, pi. 15, fig. 1; Macan, 1938 : 349 (listed); Jangoux, 1978 : 95-96, fig. 2 A, B. (IP. gracilis: Walenkamp, 1979 : 26-30, figs 9, 10, pi. 5, figs 1-4, pi. 6, figs 1-4). Cheiraster mirabilis: Downey, 1973 : 41-42, pi. 13, figs A, B; Walenkamp, 1979 : 23-26, figs 8, 10, pi. 4, figs 2-4 (non C. mirabilis (Perrier, 1 88 1 )). TYPE LOCALITY. Unknown to Verrill but the holotype (USNM 18469) may have been from Albatross st. 2396 (Gulf of Mexico between the Mississippi delta and Cedar Keys) from which there is a sample (USNM 10564) including another specimen of C. planus (named mirabilis by Verrill but probably at an earlier date) and a note saying '17 specimens lent to A. E. Verrill'. RANGE. Western Gulf of Mexico and SW Caribbean, eastwards south of Jamaica to the Lesser Antilles and along the north coast of South America to French Guyana; 550 (?408)-715 (7824) metres. A record from SW of Florida, Gerda st. 439, c. 24° N, 82+° W, needs confirmation since even the larger of the two specimens has R only 16 mm; the superomarginal spines appear to be larger than the inferomarginal ones, the abactinal plates are convex, the inferomarginals have fairly uniform spinules and the furrow angles of the adambulacrals are relatively shallow, as in C. planus as opposed to C. echinulatus. However, most of the distinctive subambulacral spines are lost. FAMILY BENTHOPECTINIDAE 117 Fig. 5 (a, b) Pore areas of Cheiraster (Cheiraster) planus Verrill; (a) holotype, Albatross (st. ?), USNM 18469, R 92 mm; (b) Albatross st. 2396, USNM 10564, R c. 60 mm (same sample as holotype of Pectinaster gracilis Verrill); (c) Cheiraster (Cheiraster) sepitus (Verrill), ventral interradius partly denuded, same specimen as in fig. 4(k), the sockets of the missing subambulacral and suboral spines dotted; (d-g) Cheiraster (Luidiaster) robustus (A. H. Clark), holotype, Albatross st. 2789 (Chile), USNM 36945; (d) papular area and adjacent second and third superomarginals; (e) armament of four proximal abactinal plates; (f) ventral interradius, armament partly reconstructed, mostly lost; (g) fourth inferomarginal and adjacent adam- bulacral plates. Cheiraster (Cheiraster) sepitus (Verrill) Figs 4i-r, 5C Archaster sepitus Verrill, 1 885a : 151. Pontaster venustus Sladen, 1889:52-55, pi. 8, figs 5, 6, pi. 12, figs 5, 6; Perrier, 1894:287-288; 1896:47;Koehler, 1909: 14-15, pi. l,fig. 13; Mortensen, 1927:71. Pontaster venustus var. robusta Sladen, 1 889 : 55. Pontaster sepitus: Verrill, 1894:247; 1895 : 137; 1915 : 1 30 (footnote). Pontaster perplexus Perrier, 1894 : 288-290, pi. 21, fig. 1, pi. 22, fig. 1; Mortensen, 1927 : 72. Pectinaster echinulatus (pt) Ludwig, 1910: 449 (sepitus a synonym). Pectinaster dispar Verrill, 1915 : 148, pi. 14, fig. 1; Macan, 1938 : 349 (listed). Pectinaster venustus: Koehler, 1924 : 186. Luidiaster dubius H. L. Clark, 1941 : 30, pi. 2, fig. 2. Luidiaster mixtus (pt) H. L. Clark, 1941 : 29-30 (specimen from 1000 fathoms). 118 A.M.CLARK TYPE LOCALITY. S of Cape Sable, Nova Scotia; 1 570 metres. RANGE. South of Nova Scotia south to Cuba and the Lesser Antilles, east to the Azores, Bay of Biscay [? no details, Koehler, 1924], Portugal and south to the Cape Verde area; 485-3703 metres (mostly 1000-2000 metres). PECTINASTER Perrier Pectinaster Perrier, 1885:70; 1894:278-279; Ludwig, 1910:448-451 (pt); Fisher, 1911 : 122-123; 1919 : 182; Macan, 1938 : 348-349 (pt). (non Pectinaster. Verrill, 1915, Downey, 1973). Pontaster (pt) Sladen, 1889 : 43-52. TYPE SPECIES. Pectinaster filholi Perrier, 1885, by subsequent designation by Ludwig, 1910. No type species was designated by Perrier in 1885 when he included two species, Archaster insignis Perrier, 1 884 and P. filholi which was evidently new (though not so stated), the genus being diagnosed only by the presence of fasciculate pedicellariae (though these are relatively simple in insignis having up to only four spinose 'valves'). In 1894 Perrier referred P. insignis to Dytaster Sladen, 1885, leaving P. filholi by elimination as the only eligible type species for Pectinaster, the formal designation awaiting Ludwig (1910: 449). DIAGNOSIS. A genus of Benthopectinidae with no odd interradial marginal plates (though the paired first marginals may be a little asymmetric in some interradii of some individuals), the interradial marginal spines relatively small; abactinal plates small paxilliform, at least the proximal ones elevated medially into a more or less well defined column bearing several spinules, on most plates of the disc and midradially on the arms surrounding a larger central spinelet, the spinules tending to form a continuous surface proximally in well-preserved (and probably in live) specimens; papulae restricted to an elongate oval papularium on to each arm base, formed of specialized markedly deepened abactinal plates, convex at the upper end and abutting together laterally with flanges to accommodate the papulae (see Fig. la), the surface of the papularium tending to bulge upwards (at least in preserved specimens when the dorsal body wall is contracted on to the ambulacra), the first-formed pores often forming transversely-aligned pairs, three or more pores developed by a size of R 20 mm; superomarginal plates mainly aligned laterally, only forming an inconspicuous rim to the paxillar area in dorsal view, each armed with an acute conical spine at the upper end; adambulacral plates with a single large conical subambulacral spine, only the first plate sometimes with a second spine; fasciculate pedicellariae usually present, often numerous, occurring on some abactinal, marginal, actinal and even a few adambulacral plates, usually oval in outline and formed of three or more modified spinelets based on either one or (pectinate pedicellariae) on two, rarely three, plates, the valves surrounding a deep hollow or bevelled notch in the plate or plates. Dorsal muscle bands of arms attached proximally to the wings of one or two ambulacral plates laterally and the adjacent marginal plates. REMARKS. This diagnosis emphasizing the highly modified compact oval papularia, paxilliform abactinal plates and laterally-aligned superomarginal plates not forming a wide border to the paxillar area in dorsal view, restores Pectinaster to the restricted concept of Perrier in 1894. As discussed under Cheiraster, the inclusion in Pectinaster of Archaster echinulatus Perrier, 1875, Pontaster cribellum Alcock, 1893, A. sepitus Verrill, 1885 and Pontaster venustus Sladen, 1889 (the last two as synonyms) by Ludwig (1910), followed by Verrill's addition in 1915 of Pectinaster gracilis, disparand mixtus (all new nominal species) and Pontaster oligoporus and Cheir -aster vincenti, both of Perrier, 1 894, broadened the genus to include species with less specialized papular areas and dorsally broad superomarginal plates. All these I consider have closer affinity with the type species ofCheiraster, C. gazellae Studer, 1883. The type material of most of them is relatively small so that the papulae have not yet formed bilobed areas, though the enlarged median proximal pore indicates that they are potentially bilobed. FAMILY BENTHOPECTINIDAE 1 19 The elimination of these leaves Pectinaster in the Atlantic represented only by P. filholi and P. pristinus (Sladen), though the status ofP.forcipatus (Sladen) as distinct from P. filholi has been debated. In 1910, Ludwig referred Pontasterforcipatus Sladen, 1889, from off the eastern U.S.A., to the synonymy of Pectinaster filholi, type locality off Cap Blanc (with a paratype from the Azores area). He was followed in this by H. L. Clark (1923) but in 1962 I challenged this synonymy on the grounds of apparent differences in the numbers of actinal plates and their armament in the type series of P. forcipatus and P. forcipatus var. echinata (the latter from the Crozet Is area of the Southern Ocean), compared with Perrier's description and figures of P. filholi supported by a specimen (R only 33 mm) in the British Museum collections from SW of Ireland. However, after seeing further NE Atlantic specimens collected by the Discovery, in 1976 (Clark & Courtman-Stock) I came to agree with Ludwig and synonymize both forcipatus and echinatus with P. filholi, which name I used for specimens from South Africa, following H. L. Clark. Not only is there considerable variation in the numbers and shapes of the actinal plates in specimens from various parts of the Atlantic but also the records of Sibuet (1975) from the Gulf of Guinea and off Angola provide a zoogeographical link between filholi from NW Africa and echinatus from the Cape and Southern Ocean. Perrier's illustration of P, filholi (1894, pi. 20, fig. 3b) shows oral plates of exaggerated size at the expense of the actinal area, which appears exceptionally small, only c. 35% of the oral plate length, measured interradially. In four Discovery specimens from the Cap Blanc area (st. 9 1 33/5) an actinal area of each measures 6 1 , 68, 74 and 95% of the oral plate length. This compares with three specimens from W of Ireland with values of 48-92% (the minimum predictably in the smallest specimen with R only 40 mm), ten from the Gulf of Guinea off Ghana with 60-90%, six from off the U.S.A. with 65-109% and six from off South Africa and the Crozet Is area with 76-115%. The specimens with the relatively largest actinal areas, interradially longer than the oral plates, are mostly large, R > 70 mm (see Table 6). Sladen distinguished the variety echinata from the type material of P. forcipatus by the stouter marginal spines and larger central paxillar spinelets, the occurrence of which he thought was more widespread on the lateral paxillae. In the two largest syntypes of echinata (R c. 80 mm) the fourth superomarginal spine is 1*25 and 1-15 mm in basal breadth, the latter measurement being equalled in a South African specimen with R 72 mm. In North Atlantic specimens these spines are up to 0'9-1'Omm — a small difference but just appreciable with the naked eye. Similarly, the longest central paxillar spinelets on the disc measure 1 '0 mm or more in the southern specimens as opposed to usually O'S-0'9 mm in the northern ones, though one syntype ofP.forcipatus has spinelets up to !• 1 mm. The extent of the central spinelets on the lateral paxillae varies also in P. mimicus from the Indo-West Pacific and P. agassizi from the East Pacific and I do not think it is a character of specific weight. Nevertheless there may be sufficient justification in the relative spinelet size for treating the southern specimens as subspecifically distinct from P. filholi but this should await comparison with material from the southern part of West Africa. One regional variation which may be significant is the relative number of pores in the papularia. In the available material this reaches a maximum of 20 in the american specimens and 21 in the NE Atlantic ones. The figured syntype(s) of P. filholi (Perrier, 1894, pi. 20, figs 3a, 3d) shows 7 pores (only 6 in the description). A syntype in the MCZ, Harvard, with R only c. 45 mm, also from the Cap Blanc area, has 10 pores in one area examined. However, a Discovery specimen from this vicinity has as many as 2 1 pores in one area and 1 8 in another at R 65-70 mm. An even higher number is shown by some specimens of a large sample taken off Ghana by the Pillsbury (c. 4° N, 2^° W), of which 22 specimens with r > 8 mm (R probably 40+ mm) have 1 1-27 pores, mean 19. This compares with 6-20 pores and a mean of only 10 in 16 american specimens. Unfortunately the Pillsbury material is dried and in poor condition so other characters such as spinelet length cannot be ascertained in many specimens. Even in spirit-preserved specimens of Pectinaster, owing to frequent breakage of the arm tips, few accurate estimates of R/r or counts of marginals can be made. The mean R/r value for 16 of the type series of P. forcipatus is 5'9/l, compared with 5*5/1 in six C 2 f? < 0 1 ^" *n ~^ | r** ON so n fN O — ' O m oo r~ in r-~ fN ro -'a ON OO o ON ON O r- O O ~ fN — ON OO •^t 1/5 1001 I — 1 1 1 O O — O — 1 6 i i ~* ~* 00 o OO . g^ ^^r-j r^ — — •> in — z a ^ ""! oo" oo O ON " so r — o ' ^ so O o' m m o ON — m r- m 7 t^ >n o ON TJ- TI- r- r~ so — o in rf in ON — ON m e£ sb | "n 1 -rr in in in in 1 in | | in 'Jr 1 1 1 in n in O m fN ON ON •^t — • oo oo ~ ^f Tf rr> fsj — CO fN f") O 'fr Tf ^t — ON C^t --*~v f^ m o O fN \O ^f *-O fN ^J OO SO Tf ^t m so so m in in 1 1 1 m r* so r- ^ m § •o ij >/^ *o ^o jz in in r*i fN 30000 r~.n o t/5 t/5 J3 Specimen c^ u* f*^ f*^ r*^ ^^ ^ ^ I , . rt ?TON ON ON to O 7J ^ -^ -^ ^ CQ -5 | | | _| CQ O ••** •-«* -"-^ Q \^ _j« »^j •— ^. •— ^. ^^< W&NWoflrel Discovery 11 1 1/ Discovery 11 1 1/ Discovery 1109/ Discovery 1109/ . *""* CX ^ Q ^ C >> ^ Ci, w •§ c | 'g GulfofGuinea: Pillsbury st. 34 Pillsbury st. 34 Pillsbury st. 34 South Africa: [No data] Gilchi [No data] Gilchi Crozet Is area: SyntypesofPo^ forcipatus var. echinata FAMILY BENTHOPECTINIDAE 121 specimens from the NE Atlantic but this may not be significant since the latter are mostly smaller. As for Pontaster pristinus Sladen, based on a single small specimen from Challenger st. 325, off the River Plate, this is clearly a Pectinaster sensu stricto, having relatively narrow superomarginals in dorsal view and compact, slightly bulging, papularia. Reexamination of the holotype throws considerable doubt on Sladen's arguments about its distinction from P. forcipatus. The specimen was coated with fine mud. When this was removed, the abactinal plates of the disc were found to have clusters of up to 7 spinules, usually 4 or 5, around the fine spinelets, the tips of which were seen by Sladen, the armament being in fact indistin- guishable from that of specimens of P. filholi of similar size (e.g. from Challenger st. 46, R c. 18 mm). (Sladen somehow misjudged R in pristinus as only 11 -5 mm but the complete arm remaining, though curled over, is at least 5 mm longer than this. Only two pores could be found in the papularium cleared and the abactinal spinelets of the disc are up to 0'6 mm long. The condition of the arm plates is poor but does not appear to be significantly different from small specimens of P. filholi and the madreporite is no more conspicuous than can be attributed to adpression of the armament in its vicinity. There are 4 or 5 furrow spines on the proximal adambulacral plates with a few similar spines in series with them proximal and distal to the large conical subambulacral spine. Three North American specimens with R 18-21 mm, have 3, 4 or 6 pores in one papularium but this is probably not significant in view of the variation in pore number shown by other specimens of filholi. Nor do I consider the absence of pedicellariae in the single specimen of P. pristinus justifies a specific distinction, though pedicellariae are more or less numerous in most specimens from the NW and NE Atlantic examined. Outside the Atlantic, Pectinaster includes P. agassizi (Ludwig, 1905) from the East Pacific and P. mimicus (Sladen, 1889) from the Indian Ocean, East Indies and New Zealand. Despite its having enlarged spinelets on the lateral as well as midradial abactinal arm plates and only 12-14, rather than 15-18, proximal adambulacral plates matching the first ten inferomarginals, P. hylacanthus Fisher, 1913, from the Philippines and Indonesia, was reduced to a subspecies of P. mimicus by Doderlein, 192 1 . If these characters are of less than specific weight, then P. hispidus (Woodmason & Alcock, 1891) from the Indian Ocean is also conspecific with P. mimicus. This synonymy is supported by comparison of a paratype of P. hispidus with the holotype of P. mimicus. Further, Macan (1938) identified several specimens from the Gulf of Aden as P. agassizi or P. agassizi granuliferus subsp. nov. rather than P. mimicus. The new subspecies was said to differ in having abactinal central spine(let)s on the arms restricted to the midradial areas and more granuliform spinelets on the actinal plates (besides differences in the distribution of the pedicellariae — not significant in my view), these two characters being also used by Macan to differentiate P. mimicus from the specimens he called P. agassizi. The fact that his largest specimens from the Gulf of Aden have R 1 19 and 158 mm could well account for the very high number of 40-50 pores in the papularia in apparent agreement with P. agassizi compared with up to only 20 in the few specimens attributed to P. mimicus in which pores have been counted (the types of P. mimicus with R 78 mm and of P. mimicus palawanensis, R c. 95 mm), both Fisher and Doderlein having apparently given little weight to relative pore number as taxonomically significant in Pectinaster. East Pacific material of P. agassizi has even more numerous pores — already c. 55 at R only 65 mm from c. 25 at R 45 mm, so that the number is likely to be even greater when (if) R exceeds 100 mm. Of the other nominal species included in Macan's table for Pectinaster, Pontaster cribellum Alcock, 1893, was compared by Alcock rather with P. subtuberculatus Sladen, since referred to Cheiraster. Alcock's holotype had R only 45 mm and this could account for the still rounded rather than bilobed form of the papular areas. Finally, Pectinaster robustus A. H. Clark, 1917, from off Chile, has two large subambulacral spines and reexamination of the large holotype showed the pore areas (Fig. 5d) to be deeply bilobed distally and not restricted to specialized papularia. P. robustus is therefore now referred to the subgenus Luidiaster of Cheiraster. This action brings robustus into conflict as a junior secondary 122 A. M.CLARK Table 7 Data for the Pillsbury specimens of Benthopecten. For the abactinal spinules per proximal plate '!+' signifies an enlarged spinelet or spine on some plates; for disc spines (which tally includes enlarged spinelets more than twice as long as the ordinary spinules), C, R and I in capitals or lower case signify large or small central, primary radial and primary interradial spines, the rest being irregular in position; for pore extent, the number is the superomarginal plate opposite the distalmost pores; the adambulacral count is of the proximal ones opposite the first 10 inferomarginals; brackets around entries signify occasional occurrence. Pillsbury station Abactinal spinules proximal distal Disc spines Pore Adambs: 10 Furrow extent i.m. pis spines 68 1 , Guyana Basin 1+3-10 1-4 1-6 1-3 lC,4R + c. 50 1C 4 19 3 19* 6-8 5 782, Colombia 3-10 - lc,2r + 3 3 c.6 1178,ECuba 2-5 1-3 0 - 6 120, Florida/ Bermuda 3-6 1,2(3) 1-4 1-3 1-4 1 lc,2r,2i + 2 lC,lr 2R + 3 4 16(15) 2 21 1 19 6-8 C.I c.5 3 14, Gulf of Guinea 1+3-10 1-5 1 1-3 lC,2r + 8 lC,2r,li 5 21 4 21* 1 7-6 34, Gulf of Guinea 1 (2,3) 1 (2,3) 1 1(2) 0 lc+1 2-3 1\\ 2 6(5,7) 7(8) 681, Guyana 1(2) 1 !C,lR,2r, 3(7 + 2)1 + 3 3 16* 5,6 (7) homonym with the species-group name robusta Sladen, 1889, originally described as a variety of Pontaster venustus but now referred to the synonymy of Cheiraster (Cheiraster) sepitus (Verrill, 1885). However, robustus (A. H. Clark) may well prove to be conspecific with C. (L.) planeta (Sladen, 1889), type locality off southern Chile, when specimens of intermediate size are available for comparison. Turning to benthopectinids with odd interradial marginal plates, Benthopecten itself is well represented in the Atlantic and also in need of review. BENTHOPECTEN Verrill Benthopecten Verrill, 1884:218 (footnote); 1895:130; Ludwig, 1910:464^66; Fisher, 191 1 : 142-144; Macan, 1938 : 351-352; Downey, 1973 : 40. Pararchaster Sladen, 1885:610-611; 1889:4-7; Ludwig, 1910:462^64. (Type species, by monotypy, P. pedicifer Sladen, 1 885). TYPE SPECIES. Benthopecten spinosus Verrill, 1884, by monotypy. DIAGNOSIS. A genus of Benthopectinidae with an odd interradial marginal plate in each series, usually in all five interradii, the superomarginal one more or less inset on to the disc and bearing a conspicuous spike-like vertical spine (or occasionally a cluster of c. 3 small spines); abactinal plates thin, flat and scale-like, at least on the arms, but sometimes convex medially on the disc, never markedly paxilliform, each bearing one or more spinules, some of which may be elongated into spinelets or even spines more than 1*5 mm long, especially FAMILY BENTHOPECTINIDAE 123 Pillsbury station Subamb. spines Oral f. spines Suboral Pedicellariae spines Abact. Act. i.m. R r 681, Guyana Basin 2 2 5 5,4 4 (+) + + 4 (+) + + 100 + (?c. 150) 40 + (760-70) 20 9-5 782, Colombia 2 6 4 + + + 20 ++(770- 100) 11-5 1178,ECuba (1) 6 2,3 + + 7? v. poor 9-5 120, Florida/ Bermuda 2(1) 1(2) 1(2) 7,8 c.6 5(4) 5,6 + (+) (+) 4,5(6) (+,1) (+,1) + 3(2) 100 50 + (?c. 70) 30 11 9 6 3 14, Gulf of Guinea 2 2 6,7 6 3 + + + 5 + + + 75++(7c. 130) 100 + (?c. 150) 19 20 34, Gulf of Guinea 1(2) 1(2) 6 7(8) 3 + + + 4 + + + 45 + (770-80) 30 + (?c. 70) 9-5 9-0 681, Guyana 4,5 45++(?c. 110) 15 on the primary plates of the disc; papulae scattered on the arm bases, usually extending further distally close to the marginal plates, so the areas become bilobed, sometimes covering much of the disc and a limited distance (usually <^R) along the arms; marginal plates with at least one long spine, the inferomarginals usually with an accessory spine (or two) below the main one but smaller; adambulacral plates with the furrow margin rounded or angular, armed with 3-8, sometimes more, radiating, slender furrow spines backed by one or two (rarely three) larger subambulacral spines usually aligned almost at right angles to the furrow when more than one occurs but in some species nearly parallel to it; some pectinate pedicellariae usually present, especially on the actinal plates, but often also between some abactinal or inferomarginal plates, especially proximally. REMARKS. Yen-ill's extraordinary omission in 1884 (though rectified in 18856:519) to mention the most obvious characteristic of Benthopecten — the odd interradial supero- marginal plates with their conspicuous single spines — led Sladen (1885 & 1889) to refer his several species to a new nominal genus Pararchaster. In 1894 Verrill referred Pararchaster to the synonymy of Benthopecten. Although Ludwig (1910) tried to retain two genera on the basis of the multiple abactinal spinelets of Pararchaster pedicifer Sladen, the type-species of Pararchaster, as opposed to the single ones of Benthopecten spinosus, Fisher (1911) found this distinction untenable at the generic level, following observations on the more abundant North Pacific material available to him. Study of the material of Benthopecten collected in the deep tropical Atlantic by the Pillsbury raises new doubts about even the specific significance of the frequency of the abactinal armament in many cases. These new specimens 124 A.M.CLARK yield the data included in Table 7, from which it is evident that the very small samples collected at stations as wide apart as the Gulf of Guinea off West Africa, the Guyana Basin off northern South America (outside the Antillean chain) and the North American Basin between Florida and Bermuda, show a wide range within each sample in abactinal spinule frequency. The maximum spinule number is found in one large specimen from st. 68 1 off Guyana and another from st. 314 in the Gulf of Guinea, each sample including a second specimen with much less dense armament but otherwise not significantly different. St. 120 also includes one more spinulose specimen out of three, though the difference is less marked. However, a third specimen was collected at st. 68 1 ; this is the single individual, segregated in Table 7, having the minimum of abactinal spinelets — rarely more than one, which may be specifically distinct. Such a conclusion is supported by its primary interradial disc spines appearing to be better developed than the primary radial ones, coupled with there being rarely as many as 7 furrow spines (and that only on odd plates), where 6 or 7 is the usual number in the other specimens, and only a single (actinal) pedicellaria, where all the rest with R > 50 mm have several actinal and inferomarginal pedicellariae and usually also some abactinal ones. A review of the fairly abundant material of Benthopecten spinosus from the eastern continental slope of the U.S.A. has shown a consistent agreement in these same characters, except that the rare pedicellariae found were only on the inferomarginal and not the actinal plates. In comparison with other Atlantic species, a multispinulose condition with up to 10 spinules on some abactinal disc plates is shown by the four syntypes of B. spinosissimus (Sladen) from off Ascension I (R 60-68 mm) and the holotype of B. fischeri (Perrier) from NW Africa (R c. 40 mm), the latter reexamined at the Paris Museum. In B. spinosissimus already at R < 70 mm the papulae extend distally to level with the sixth or seventh superomarginals and the marginals themselves are relatively short and numerous, numbering 47-53 in each series in the syntypes, giving a ratio of 1 -2-1 -4/1 for R/SM number. This compares with only about 30 marginals and a ratio of c. 2-2/1 at this size in B. spinosus and probably a similar number in B. simplex (Perrier), of which a specimen with an intact arm showing R 55 mm has only 25 superomarginals, also giving a ratio of 2-2/l . A further distinction of B. spinosissimus is the small number of furrow spines, only 3 or 4 on each plate, leaving little doubt of its specific distinction. With regard to B. fischeri, I found the holotype and only recorded specimen to have 4 or 5 spinules on most proximal abactinal plates but additional ones on some plates with an enlarged spinelet or spine, apparently an unusually high frequency at this relatively small size. Although the jaws have a relatively rich armament of c. 10 furrow spines and up to 12 suboral spines and spinelets on each oral plate, there are only 6 furrow spines on most adambulacral plates and the other characters agree with the bulk of the Pillsbury material. (The exception is the single Pillsbury specimen allied with B. spinosus, which is quite distinct in its very sparse armament.) B. fischeri was referred to the synonymy of B. spinosus by Grieg (1921 & 1 932) on its superficial resemblance. Grieg also followed Verrill (1895) in synonymizing Pararchaster armatus Sladen, 1889 with B. spinosus, though Farran (1913) and Mortensen (1927) maintained armatus as a distinct species. Ludwig (1910) had instead referred armatus to the synonymy of Archaster simplex Perrier, 1881, the oldest name for any species of Benthopecten. The holotype (from the Gulf of Mexico NW of Cuba) has R only 18 mm and, until the Oregon and Alaminos took some rather larger specimens (R 48+ mm) in the Gulf of Mexico (described by Downey in 1973), little comparative information was available about B. simplex. Reexamination of these specimens and comparison with the type material of B. armatus from S and E of Cape Cod and various specimens from the vicinity of Portugal, SW Ireland and the Rockall Trough in the NE Atlantic reveals no differences which I consider to be significant (see Table 10). The abactinal armament is multiple in all except immature specimens (numbering usually 1-5 spinules on the proximal plates), the disc spines (when developed) tend to enlarge first on the primary radial more than the interradial plates, the number of furrow spines averages 6 or 7 on most of the adambulacral plate series and the number of sub- FAMILY BENTHOPECTINIDAE 125 ambulacral spines tends to increase from one to two with size, with occasional exceptions; lastly there is a relatively high incidence of pedicellariae, often in all three loci, in all but the smaller specimens (R <40 mm), again with some individual variation. The origin of some of the Pillsbury specimens from areas geographically adjacent to the Gulf of Mexico (the nearest being from st. 1178 east of Cuba) and their morphological resemblance to the rather smaller and mostly less well preserved Oregon and Alaminos specimens reinforces belief that all can be referred to B. simplex but for the exceptional specimen from st. 681 which is indistinguishable from B. spinosus. This serves both to extend the known size range of B, simplex to R c. 1 50 mm and also to extend the geographical range (not unexpectedly in view of the great depths) not only to both sides of the North Atlantic but also both sides of the tropical Atlantic, including the Gulf of Guinea, taking in the type locality of Benthopecten chardyi Sibuet, 1975. Despite several peculiarities of Sibuet's unique holotype (discussed further below), it seems to fall within the range of variation of B. simplex. CHARACTERS. Apart from the density of the abactinal armament (now considered to be of only limited use) the following characters have been adopted as of some taxonomic signifi- cance in distinguishing the species of Benthopecten: 1 . The development of enlarged spines on the centre of the disc (by Ludwig) 2. The relative size of any accessory spines on the inferomarginal plates (Sladen, Ludwig and Fisher) 3. The number of paired marginal plates in each series (Sladen) 4. The numbers of furrow and subambulacral spines (Sladen et al.) 5. The number of oral furrow spines (Ludwig) 6. The number of adambulacral plates corresponding to the first ten paired inferomarginal plates (Ludwig and Macan, 1938) 7. The occurrence of pedicellariae (Sladen et al.) 8. The presence of more than one superomarginal spine (Sladen and Macan) 9. The substitution of several small spines for the usual conspicuous odd interradial superomarginal spine (Perrier, 1894, forB.folini) 10. The extent of papulae along the arm (Fisher and Macan) 1 1 . The extent of the inferomarginal pedicellariae along the arm (Fisher and Macan) 12. The number of actinal pedicellariae (Macan) Other characters which can be considered as possibly significant for some of the species but needing further comparisons for appreciation of their wider significance within the whole genus are: 13. The occasional non-bifurcation of the papular areas distally 14. The relative size of the primary radial and interradial disc spines, when developed 15. The number of suboral spines 16. The presence of more than one series of actinal plates and 17. The relative size of the first paired superomarginal spines. Fortunately, since the arms are rarely preserved intact, the R/r ratio does not appear to be taxonomically significant in this genus, the relative disc size being apparently variable over a similar range in all the species. As in many long-armed asteroids, the R/r ratio tends to increase allometrically with absolute size. As with the characters based on the abundance of abactinal spinules, most of the others are subject to some degree of variation. No. 1 , the development of enlarged spines on the disc, depends to some extent on size and is also subject to variation. For instance, in large specimens of B. spinosus, R >90 mm, the disc spines range from few and inconspicuous, < 1'5 mm long, to a cluster of 20 or more, some exceeding 6 mm in length. An elaboration of this is character no. 14, the relative size of the primary radial and interradial disc spines, when developed, studied now only in B. spinosus and B. simplex, where a range of specimens is available. As detailed under the heading of B. simplex, on average the primary radial spines in that species are equal to or 126 A.M.CLARK slightly larger than the primary interradial ones, whereas in B. spinosus the interradial spines are mostly better developed (not counting those of the madreporic interradius where the spine development is generally erratic). This character remains to be evaluated in other species. No. 2, the mean size of the accessory inferomarginal spine (if developed) relative to the main spine above it, appears to be somewhat variable and is of uncertain value because of paucity of information about its range, only holotypes having often been described. In some species such as B. spinosus it seems to be fairly consistently at least half as long as the main spine, whereas in B. simplex the reverse is true and it rarely exceeds half the spine length. (An exception again is the holotype of B. chardyi where the accessory spine is often large and may even equal the main spine on some plates.) Very spinose species such as B. spino- sissimus have several successive accessory spines on each plate, graduating in size. No. 3, the number of marginals relative to R, remains to be investigated for most species, marginal counts having rarely been cited by authors, partly because so few specimens are usually obtained with any arms intact. The development of additional marginal plates adjacent to the terminal does not keep pace with the elongation of the arms so that the ratio of R/SM number tends to increase with R. In 12 specimens of B. spinosus with R 35-155 mm, the ratio ranges from 1-8/1 (in the smallest) to 3-4/1 (in the largest), whereas nine specimens with R 70-110 mm have a range of 2-3-3-0/1 and a mean of 2-6/1. The holotype of the closely related B. semisquamatus (Sladen), from Japan, does not differ significantly, having a ratio of 3-0/1 at R 166 mm. In contrast, as mentioned above, the four syntypes of B. spinosissimus with R 60-68 mm have a ratio of only 1-2-1-4/1, the marginals being relatively short and more numerous, 47-53, as opposed to only 32 in a specimen of B. spinosus with R 73 mm. No. 4, the numbers of furrow and subambulacral spines, may show a wide range along the furrow length of a single specimen, usually tending to increase beyond the proximal plates and to a greater extent in very large specimens. For instance, in the holotype of B. heteracanthus Macan from the Indian Ocean, R 210 mm, the number increases from 7 or 8 proximally to 1 1 on many distal plates, while the holotype of B. polyctenius Fisher from Indonesia, at R 224 mm, has a range from 9-14. Two smaller specimens of B. polyctenius with R 120 and 80 mm have lower maxima of 1 1 and 9 respectively. In species of more restricted size range, the range in furrow spine number is probably restricted to c. 3 or 4. For instance, in B. spinosus, known to reach R 1 55 mm, the range is 4-6, predominantly 5, even in larger specimens, R >90 mm, while B. simplex with a similar size range usually has 6 or 7, occasionally 5 or 8 spines. The holotype of B. chardyi Sibuet (now thought to be less than specifically distinct from B. simplex) has 5 or even 4 spines on about the first ten adambulacral plates but the number increases first to 6 then 7 beyond about the twentieth plate. As for the subambulacral spine number, up to a size of R 30 mm B. simplex has only single spines on nearly all the plates, then in most specimens a second spine becomes enlarged beyond the most proximal plates so that specimens with R >60 mm usually have two subambulacrals predominating throughout the series. (The holotype of B. chardyi with single spines on most plates could be interpreted as neotenous in this respect; two slightly smaller specimens of B. simplex also from the Gulf of Guinea have a small second spine on some plates.) No. 5, the number of oral furrow spines, would be expected to vary through the same sort of range as the mean number of adambulacral furrow spines. Certainly B. polyctenius shows a range of 8-14 according to Fisher (1919). B. spinosus has 5 or 6, occasionally 4, and B. simplex shows a range from 5 to 8, compared with 10 (or even 1 1) oral furrow spines in the holotype and only mature specimen of B. fischeri (Perrier) at R only c. 40 mm, which in other respects resembles the more spinulose specimens of B. simplex and has only c. 6 furrow spines on the adambulacral plates. Until more material from the type locality of B. fischeri is available, the significance of this character remains to be determined. No. 6, the number of adambulacral plates relative to the first ten paired inferomarginals, appears to vary over a range of 4 or 5, sometimes more, in a single species (see Tables 9, 10), FAMILY BENTHOPECTINIDAE 127 judging from the data available. Unfortunately, even Ludwig, who initiated its use, failed to cite it for several species and not all subsequent authors have included counts in their descriptions. The usual number in B. spinosus is 1 7-20, while in B. simplex between 1 8 and 21 is common but as many as 24 (or even 25) are found, for instance in the anomalous holotype of B. chardyi. Also in the holotype of B. violaceus (Alcock) from the Indian Ocean 24 are supposedly found but this number is derived only from the drawing, not Alcock's description, and may not be correct. Nos 7, 1 1 and 12. The occurrence and frequency of pedicellariae are notoriously variable and rarely reliable in diagnosing most asteroid species; this is equally true of Benthopecten. Only a large sample (rarely provided) can give a good idea of the relative frequencies of abactinal, actinal and inferomarginal pedicellariae. Even in B. spinosus, in which pedicellariae have long been thought to be consistently rare or absent, an additional sample of ten specimens from the Atlantis II collections showed four with at least some infero- marginal pedicellariae. No. 8, the presence of accessory superomarginal spines in species such as B. spinosissimus is correlated with a generally higher frequency of spines on all the plates and is not in itself significant. No. 9, the absence of conspicuous spines from the odd interradial superomarginal plates, as described in B.folini (Perrier, 1894), from off NW Africa, is not shown by all specimens of the type series so one must conclude that either two sympatric species were included or that the character is variable. If there are two species in the type series, the one with multiple diminutive spines on the odd plates also usually has two subambulacral spines. The other, with single large spines may have one subambulacral spine and is then indistinguishable from B. simplex chardyi. It is not certain which specimen in the Paris Museum collections Perrier used for his full description; the arms were broken and he could only cite r, which he gave as 12 mm. The largest specimen from off Cap Blanc (Talisman 1883 haul 96, P.M. no. 3304) is almost certainly the one shown in Perrier's pi. 19, figs la and Ib of a general view, above and below; I estimate its r as lO'Q-10'5 mm (some shrinkage during nearly 100 years in spirit is not unlikely); R is 65 + 710-20 mm; all the odd superomarginals have small spines and there are two subequal subambulacral spines on most plates. Of the two specimens from haul 98, slightly further south (P.M. nos 33 1 7 and 33 1 8) one has had the abactinal body wall removed and is the specimen so mutilated shown in Perrier's pi. 20, fig. Ic. The second has R50 + 720-30 mm, r c. 1 1 mm and so could be the specimen from which pi. 20, figs la and Ib of details of parts of the disc were made, though there are some divergencies between the finer details of armament and Perrier's description. Accordingly, the largest specimen from off Cap Blanc is now designated as lectotype; its subambulacral armament was probably misrepresented as single in pi. 19, fig. la by Perrier's artist. However, of the four (not three) smaller and more broken specimens from this same haul 96, one has relatively large single odd superomarginal spines up to 6 mm long, R being 20 + ?c. 30 mm; its subambulacral spines are mostly single. The other three are more damaged interradially, the spines of the odd plates mostly lost; one has a pair of large spines still attached to one plate and signs of paired spines in at least two other interradii. From Talisman haul 97 (no. 3305) come three more damaged paratypes, at least one with three massive single odd superomarginal spines. Possibly these large spines are vulnerable to damage or loss from the attacks of predators and the anomalies in development are due to differential regeneration. In the holotype of B. pikei H. E. S. Clark, 1969, from New Zealand, in four out of five interradii the spines are relatively small and in 1970 she suggested that this might be attributable to injury when she described a variety australis ofB. pikei from two specimens with all the odd spines large. No. 10, the distal extent of the papulae along the arms, progresses during growth, usually from level with the first superomarginal to the third or fourth, while the distalmost pores are usually lateral rather than midradial (character no. 13). However, in B. acanthonotus Fisher, from off California, the distal extremity varies in three large specimens from level with the fifth superomarginal to the twentieth. In this species the papular areas are consistently bifurcating but the holotype and only described specimen of B. indicus Koehler, from off Sri 128 c !_«; O O o 1) 0 O o? '^ ,8 •n «/^ 1 >/^ o m 1 — n-> . r^ O^ ^^ •o a C/5 ^M — ON ^j Cy ^ c C ^^ ^^ ^ U s — £ + + + + A '5. I5 W3 '55 m \ + + + + + O 2 13 "cB c U (N \ + A + 1 1 '5. 01 'S N U e O *S~t + M CO •• — ' -*s C/5 o • ~ Vi«O 1 O^ f 3 (L) c^ ^"^ i/^ ^^T ^^ A O — (U *T^ I^,N— ^ ^ ^^ "~- ~ a .& OO ^^ '•^ir' ^"0 r^ ^j C^"i 01 *J ^ 1 ^^ ll I 1 u - _c o ^4 — •- "p •§ S 2 ^ £?is •O JS 3 g)Jg JS *"" a o C* "O -^ C C + — "r- IS) D 2 •£ in 1) s o • • CC Q £J "S i> Q — "5 Si ^~ M^ C *C/3 ^ OO * ""i ^r^ — ^ •4—t • ••4 3 Cl ro 00 C/5 C/5 C/5 2 2 • • a> C ^ ^ p C "" 00 g 4_> 5 AJ (/) — O ^ o -H C2 00 C/3 Ci, c" >^ f "3 S T> *S 'K S ^ o __ C t^ OJ _— s~ ~5 ^- ^ .5 o •JC • H *O. "^ c co *;{ 3 •— 1 J. 5 + K %i 91 C — — u^.§ "a & |.E o. n — o QQ i^ •3 c c •— o. E ^ ^5 ^ S •r o "O '— O (N 7 i ^ 2 2X < '•§ .2 ^ c^ « 0 ^O ^^ o JS -2 '"S"H in O 9 flH ^ , «v r- t C W C o C^ ^^ « ^s. j? to the specie sf o 7 o o •*^ cd Q 0—0 *" JH c« >, E a ff'SS ^ o c I- I- 00 JS a'*> ral disc spine al ones larger s on abactina "5 C/) •5 0 EA | narginal spin / inferomargi o "5 "a S | 0 C JS 'a 3 &o >ulacral spine irrow spines il spines ictinal pedice inal pedicella sromarginal j >ecimens recc ;>> u -^ IM JS 3 IS 1 J s^ .2 E S I?! S'-e-s U A 9 <•> fc c slit O "5 c superor ccesson •° £ s 1 •sg c« (B c <*- b x> t> E — 2 «a ea CC CO ^2 t^_ t_ X) ir 3 o O Jl's s § If c~ 0 ^E .0 S3 ^ a •c Cx .1 >5 s .5 e I « it! 1 "^ 'o 22 a3 ^ ^ S w || ^ ^ 55 y> c5 E — C -O t_ o "5 JD X) — E E *^ ^ 3 '3 OO J; -= '5 1> s ^ I>J ^J * §• 5 ^•s ^ ? '^ 1* S ^ 2 iii I ll II Jlloo ol - 0. S.>S -S ^ s. ~ . . O — ^T v~> ^O I OO C 73 C C C T3 c O 73 § S ^3 S « S - c« g g C o ^ S sj CJ C o 03 2 O T3 S '" O £ •£ O O 2 '3 O £* 'o '5 "« 1| u ^« c3 ^^ -^ ^^ c^ ^^ • *•* C^ C^ ^^ cd *ij ^^ >±* ^^ C "~" *7? ^ */? " ~* ^ .— ( . ^- . u, ^ *-* . •— « tn ^T? ^ ^ •*— » 4— » ^G OO *|*O ^^ ^O ^^ ^-v ^^ NO ~~ — • NO m <*? r~- CN ^^ ^j ^j ^^ 0s) ^^ -. 0- - 0 fN 2 i + +X+ + + + i X + + , ^p + J 1 a t/5 U en ^ ^- ^ Xx~, 'o ^^ 1 + + + + + + + 1 1 •"--' s— ^ + + + -— ^ + + + 1/3 o g 0 "+1 c: ^ 1 1 t 1 + 1 -t i i i cJ. + , •4- T' 1 1 *+" + c« ~C9 O ' ' * ' ' ^^^ i c 'C E 0 .2 iJ E — *? T ^ ^ "i ^ *? *? « C ° i: c o ^J NL*' C^« v N O" c^> c^« \tJ ^s| C"* 'SI * r*^j "3 U g] m — D. i> o lorpoorpoo § •noo^n V0«n ^ rn °°, '^il «»• ~ 03 ^— ^ ^o m >n ^o ^o t~- NO Tf t — ^sf NO O- l/~> ^ C^- — •" — •< r~- S)^^ •*« o £ S X ^— v — ! 8-°°' r- "o >< ° m oo ON oo ,/-, — ^ m NO t~~ f~ ^- — t/5 ' *" 4 NO r-L 10 iin ri ^ ^s r, ^-^ 4" 1 1 ~— ' | '•P i2 . ^ OO r^- ^* r^-- NO ^O f*^ r^" iy~) i/~) IO OO r- "C S § o i_ ^3 -a 12 O °° ON — o 10 -J c^- i— 1 i— 1 1— 1 H-J ^ J _) _| 1— 1 ^ &"§ -C t/3 O 73 -S* o. U in NO ^ NO OO *o — >, '0 'st >* Tf co ^" «n \o •* r^ c^- •*-* c tj-s'g u « ^~~ I/-} rNi sU Qj 00 0 - - ro — 7 _^ UM •— ' NO i m 00 — "^ ' J2 « 6°""' •S ^00 _O ^ l/^ £ X XT T ^T ^? T ° + + + + + + c« §-~ Tabular to s ^ i*s s -^ si stoni and ^ us and #. j «//zu5 Fell -C '5 ex on 5 -S -c; >•> t-» ^*> 2 ^^ s ^||^ cognatus^ indicus s: CL ^ &I1I 'S c ^ c a 3 o s ~-s ^> -s: s; sj CS w S ci -H bo s: is ^> s T3 c ^, 'S, ^ ^ M., 111 *• *i & ?^ 129 130 A.M.CLARK Lanka (R c. 120 mm) has the papulae extending to superomarginal 12 for the whole width of the arm. In a Pillsbury specimen from NE of Florida (R 100 mm), now referred to B. simplex, the papulae extend to the fourth superomarginals and the areas do not bifurcate. Also in the holotype of B. chardyi the distalmost pores are adradial rather than abradial and extend only to the third superomarginal though R is 90-100 mm. However, two slightly smaller Pillsbury specimens also from the Gulf of Guinea have the areas bifurcating; their papulae extend distally to the third (or even the distal ends of the second) superomarginals. Isolated cases of non-bifurcating areas have been recorded from at least one non-Atlantic species and proper scrutiny of more material may well show occasional instances of this, suggesting that character 1 3 has little taxonomic significance. Since in some benthopectinids the dorsal arm muscles are attached at least partially to the dorsal body wall near the base of the arm, as well as or rather than to the ambulacrals and marginals, this might have a limiting effect on the extent of the pores medially. No. 15, the number of suboral spines, is usually 3 or 4 (sometimes 5) on each plate in most species, accompanying spinelets being rare. Occasionally higher numbers are found. Reexamination of the holotype ofB.fischeri from NW Africa showed a total of c. 12 suboral spines and spinelets on each plate, four of them distinctly larger than the furrow spines. In contrast, the four similarly multispinose syntypes of B. spinosissimus from near Ascension Island have only 4 suboral spines on most plates, though there may be one or two additional spinelets on some plates. The oral furrow spines on the type ofB.fischeri are also unusually numerous, as discussed under character 5. No. 16, the number of series of actinal plates, is correlated with the relative disc size, as well as tending to increase with growth. Until an R of c. 100 mm, rarely more than one series is well developed. More larger specimens are needed to evaluate this character. No. 17, the relative size of the first paired superomarginal spines, shows a considerable range from very reduced in comparison with the fourth (usually the largest) spine to almost as large as the odd interradial spine. It can result in very different appearances of the interradii but preliminary observations suggest that it is too variable (even within one specimen) to be significant. The characters thought to be most useful are utilized in two tabular keys (Tables 8 and 9) to distinguish the Atlantic and non-Atlantic species of Benthopecten. Analysis of these tables shows only very minor differences in variable or largely untested characters between several of the nominal species or differences possibly attributable to growth changes; notably B. spinosus Verrill and B. semisquamatus (Sladen) are separated only by an insignificant difference in the number of adambulacrals relative to the inferomarginals. The occasional presence of inferomarginal pedicellariae in B. spinosus may well prove to be shared by some specimens of B. semisquamatus other than the syntypes, when more material is available for comparison. This close affinity was observed by Sladen, who treated an east American specimen, now regarded as conspecific with B. spinosus, as only a variety occidentalis of B. semisquamatus. Inter-oceanic distributions of abyssal species such as these are very likely. If the present decision to treat Atlantic specimens showing a considerable range of variation in density of the abactinal armament as a single species, B. simplex, is proved correct, then barriers to the conspecificity of many nominal species in the middle and lower parts of Table 9 will be almost completely broken down and there should be a drastic reduction in the number of valid species of Benthopecten recognized. Since Benthopecten simplex is the first-named species of the genus and its limits — both morphological and geographical — are considerably extended by the Pillsbury and other specimens now considered to be referable to it, a full account of this species is given here. Benthopecten simplex (Perrier) Archaster simplex Perrier, 1881 : 28; 1884 : 264, pi. 1, fig. 8. Pararchaster armatus Sladen, 1889 : 19-22, pi. 1 , figs 5, 6, pi. 4, figs 5, 6. Pararchaster simplex: Perrier, 1894 : 253, 254-256. FAMILY BENTHOPECTINIDAE 131 "Pararchasternov. sp. (P. Fischeri Perrieraff.)' Mortensen in Schmidt, 1904 : 24. Benthopecten simplex: Ludwig, 1910:461, 464, 465, 466; Fisher, 1911 : 143; Verrill, 1915: 1-22; Downey, 1 973 : 40, pi. 1 2, figs C, D. Benthopecten armatus: Farran, 1913 : 2-3; Mortensen, 1927 : 74-75, fig. 41. Benthopecten spinosus (pt) Grieg, 1921 (& 1932): 12-13, pi. 4, fig. 1 (Michael Sars sts 95 & 101 at least); H. L. Clark, 1941 : 26. (Non B. spinosus Verrill, 1884). Benthopecten chardyi Sibuet, 1975 : 283, 289-29 1 , fig. 3, pi. 1 , figs B, D R is probably up to c. 150 mm but rarely more than 80 mm; R/r is 7'7-7'9/l and R/no. of superomarginal plates c. 2*25/1 in three specimens with R 50-55 mm. DIAGNOSIS. A species of Benthopecten with multiple spinules on most proximal abactinal plates, often 2 or 3 but sometimes up to c. 10 when they are accompanied by a larger spinelet or spine, especially at R >c. 50 mm, those on the primary radial plates tending to be better developed than those of the primary interradial plates; papular areas usually distally bifurcate, the distalmost pores level with the second or third superomarginals in specimens with R 50-80 mm but extending to the fourth or even the fifth in those with R > 100 mm; odd interradial superomarginal plates with a single very large spine, other superomarginals also with single spines and a few small spinelets or spinules, inferomarginals with an accessory spine below the main one but usually less than half as long; 16-23, usually 18-21 proximal adambulacral plates corresponding to the first ten paired inferomarginals (though up to 25 in the holotype of B. chardyi); actinal plates few, a second series only developed at R > 100 mm; adambulacral plates with 6 or 7 furrow spines (sometimes 5 or 8) and one or two subambulacral spines, usually two at R > 50 mm beyond the first ten or so plates, where there is more often only a single spine; oral plates with 5 or 6 (sometimes 4 or 7) furrow spines and 3 or 4 (5) suboral spines; actinal and some inferomarginal pedicellariae usually present at R > 50 mm and sometimes also some abactinal ones. VARIATIONS. As mentioned under the generic heading, the armament of the abactinal plates is very variable in B. simplex, if I am right to include most of the Pillsbury specimens. At R 50-70 mm there are usually 2 or 3 (1-5) spinules on the proximal plates and single ones on the plates distal to the pore areas but occasional specimens of this size and larger ones may have some proximal plates enlarged with a bigger spinelet and up to 10 spinules. The degree of enlargement of spinelets and spines is also variable. Even at R c. 70 mm there may be only a single central spine appreciably larger than the spinules, though usually several others may exceed 1'5 mm in length at this body size, especially on some of the primary plates. The papular areas normally bifurcate distally but the largest specimen from Pillsbury st. 120, NE of Florida, has the distalmost pores adradial. The odd interradial superomarginal spines are usually very conspicuous but on occasional specimens one or more of them may be reduced in size. The length of the accessory inferomarginal spine is usually less than half that of the main spine but it may be lacking altogether on most plates of a few individuals or conversely be nearly equal in length to the main spine on some plates, e.g. of the paratype of Pararchaster armatus with R 42 mm from Challenger st. 46 — but not in the larger paratype from the same station. In larger specimens the single subambulacral spines of the proximal adambulacral plates often give way to two spines beyond, as in Downey's specimens from the Gulf of Mexico, shown by reexamination. Other variants are mentioned in the diagnosis. SYNONYMY. Until Downey (1973) described some Oregon and Alaminos specimens of B. simplex from the Gulf of Mexico, only the juvenile holotype (R 18 mm) had been described from the West Indian area and it was difficult to establish whether Sladen's type material of Pararchaster armatus from SE and E of Cape Cod (c. 42° N, 63° W and 40° N, 66° W) is synonymous with B. simplex, as proposed by Ludwig (19 10) or with B. spinosus Verrill, also S2 — f ^OOomoO oooo^^o^oo^^rr 0 u u ^. — — ' O ON OO OO OO r-oor-~r-^ovooovo^ovsovor»^ro do E — ^™ ^— 'o £ ? E O O O O O oo oo r- r^ r- m o m ^ — 5J ^J ^ 1 SJ ^ I Vjrt^ra-^^ ^rr>^«->rr> in o3 H E £ \ + \ 1 it II4- + 4-+IIIIIII 1 (/5 U «3 .5 C C ^S *J ,0 — o ^.^ ^_^ 08 D .,< 0 < + + + + + -t- + i + i+^ + T+ii-i i i 1 '£ • •5 £2 03 eu tj .C cio CO "w S X) \ \ \ \ £ \ \ i i liiiiii iAi i 1 .S ° 5/5 r% — O Ji ^— V S-~* U 2 fc •£ in in •- 1 ^^VO^SOW^ 1 ^vo^^^n-^^vo^^-^- I rf — 3 •c o -^ C c/i u < o *— O «3 .S g Jg7 S-ji^incig tx> u > -5 a - 1 - 1 - 1 " ' • ' -• -^ • • „' ' ^^ ' ^~ ' „' 'e^- "ci.^" fc-U C3 c/3 |*l|o S ^ 1 ^ 1 ^* -o E O O O O O "^c T OU'^r^T^TC^Ov5T^toTl •S i: — 3 '5 5, a .ss H C a — CM r^ TI- >n vo I-- >3 -g H .a C/3 ag ^ * FAMILY BENTHOPECTINIDAE 133 from the eastern U.S.A., as proposed by Verrill (1895). Comparison of the larger specimens of Downey and Sladen, together with a few further individuals from the NE Atlantic (off Portugal, the Bay of Biscay, SW Ireland and the Rockall Trough) yields the data shown in Table 10, which also includes for comparison the only available specimen of B. spinosus of comparable size. Table 7 of the Pillsbury specimens, all but one now referred to B. simplex, should also be consulted. Although the positions of the disc spines and enlarged spinelets are often irregular (especially adjoining the madreporic interradius), the contrast in the specimen of B. spinosus in Table 10 of four relatively large primary interradial spines with the three smaller primary radial spines (reflected also by most large specimens of B. spinosus) is very marked, as well as the absence of multiple abactinal spinules. Further, the presence of only 4 furrow spines on most adambulacral plates, as compared with the usual 6 or 7 of B. simplex, and the total absence of pedicellariae, combine to justify treating B. simplex as distinct from B. spinosus in my view and at the same time synonymizing B. armatus with B. simplex. In fact, two smaller american paratypes of P. armatus (R c. 40 and 30 mm) and the Portuguese paratype (R c. 32 mm) do have only 5 or sometimes 4 furrow spines but this is to be expected at this size and is probably only transitory. Also occasional larger specimens of B. simplex lack pedicellariae and conversely a few specimens of B. spinosus do have some inferomarginal ones. With regard to the relationship of B. chardyi Sibuet with B. simplex, Sibuet had only a single specimen, from the Gulf of Guinea off Gabon (Jean Charcot st. CY 19, c. 2£ S, 8° E, 2584m.), R90-100mm, which has been reexamined. It has only a few inconspicuous spinelets enlarged near the anus, no disc spines, the papular areas do not bifurcate distally (not observed by Sibuet), the accessory inferomarginal spines are relatively large, sometimes almost equal in length to the main spine, there are as many as 24 or 25 adambulacrals corresponding to the first ten inferomarginals (not noted by Sibuet) and the subambulacral spines are almost without exception single; the furrow spines number 6 or 7 on most plates — as usual in B. simplex — the 4 or 5 recorded by Sibuet being limited to some of the proximal plates and transitional to the higher numbers. Apart from the last, the combination of these characters has not otherwise been found in B. simplex but each of them individually is liable to vary. Two (poorly preserved) specimens also from the Gulf of Guinea but to the NW off Ghana (Pillsbury st. 34, c. 4° N, 2{° W, c. 1970 m) with R c. 70 mm have a majority of the adambulacral plates with one rather than two subambulacral spines. However, one of them has two small disc spines developed, their papular areas are bifurcating, the accessory inferomarginal spines were probably relatively small and there are only 1\\ adambulacrals matching the first ten inferomarginals on the only arm of which sufficient remains to allow a count. Two much larger specimens again from the Gulf of Guinea but off western Nigeria (Pillsbury st. 314, c. 5° N, 4° E, 2300 m), R 7130-150 mm, both have only a single disc spine and a few enlarged spinelets, the pore areas bifurcate, the accessory inferomarginal spines are less than half as long as the main spine and there are only c. 2 1 adambulacrals corresponding to the first ten inferomarginals but the subambulacral spines are predominantly two. These few specimens indicate that in the Gulf of Guinea at least, the development of disc spines and a second subambulacral spine is relatively retarded. Whether or not this is sufficiently consistent to justify retention of B. chardyi as a distinct taxon remains to be seen from further material. At present it seems best to treat it as a subspecies of B. simplex. Similar doubts about the specific validity of B. folini must exist considering the inclusion in the type material of at least two specimens with conspicuous large odd superomarginal spines interradially, not the small multiple spines supposed by Perrier to be characteristic of the species. (Most of the remaining paratypes have lost these spines in preservation.) The larger of these two probably had R c. 50 mm and the single subambulacral spines would be expected in B. simplex at this size. It should also be noted that the holotype of B. chardyi has multiple small spines on one of its five odd superomarginal plates. Since the recognition of multispinulose specimens as conspecific with B. simplex, only the numerous oral furrow and suboral spines in the holotype of B. fischeri at R only c. 40 mm 134 A.M.CLARK serve to distinguish it. In spite of this and the presence of two subambulacral spines on most proximal plates, I think it not unlikely that further material from the type locality off NW Africa will show the holotype to be only an aberrant B. simplex. RANGE. From the vicinity of Cape Cod (42^0° N, 63-66° W) south to the Gulf of Mexico, the Colombia and Guyana Basins and from south of Iceland (Thor st. 164) and the Rockall Trough (c. 57° N, 10° W) south to the Gulf of Guinea (c. 2{° S, 8° E); 1 1 75-2585 metres. Acknowledgements Most of the observations on material of the U.S. National Museum and the Museum of Comparative Zoology, Harvard, were made during a visit in 1953/54 but I am indebted to Miss M. E. Downey and Professor R. M. Woollacott of those institutions for recent loans of specimens; also to Professor A. Guille, of the Paris Museum. References Alcock, A. 1893. An account of the collection of deep-sea Asteroidea. ('Investigator'). Ann. Mag. nat. Hist. 11: 73-1 21, pis 4-6. Clark, A. H. 1916. Six new starfishes from the Gulf of California and adjacent waters. Proc. biol. Soc. Wash. 29: 51-62. 1917. Three new starfish and one new brittle-star from Chile. Proc. biol. Soc. Wash.SQ : 151-158. 1954. Echinoderms (other than holothurians) of the Gulf of Mexico. Bull. US Fish Commn 55 : 373-379. Clark, A. M. 1962. Asteroidea. Rep. B.A.N.Z. Antarctic Res. Exped. B9 : 1-104, 18 figs, 5 pis. 1977. The South African Museum's Meiring Naude cruises. 4. Echinoderms. Ann. S. Afr. Mus. 73(6): 133-147. & Courtman-Stock, J. 1976. The echinoderms of southern Africa. British Museum (Nat. Hist.) Publ. no. 776, 277 pp., 276 figs. London. Clark, H. E. S. 1969. Two new species of Benthopecten (Asteroidea) from New Zealand. Trans. R. Soc. N.Z. (Biol.) 11 : 83-88, 1 fig., 1 pi. 1970. Sea stars (Echinodermata : Asteroidea) from 'Eltanin' cruise 26, with a review of the New Zealand asteroid fauna. Zool. Publs Viet. Univ. no. 52: 1-34, 3 figs, 3 pis. Clark, H. L. 1923. The echinoderm fauna of South Africa. Annls S. Afr. Mus. 13 : 221^35, 4 figs, pis 8-13. 1926. Echinoderms from the South African Fisheries and Marine Biological Survey. 2. Sea-stars (Asteroidea). Rep. Fish. mar. biol. Surv. Un. S. Afr. no. 4. Spec. Rep. no. 7 : 1-34, 7 pis. 1941. Reports on the scientific results of the Atlantis Expeditions to the West Indies. The echinoderms (other than holothurians). Mem. Soc. cubana Hist. nat. 15 (1) : 1-54, 10 pis. Doderlein, L. 1921. Die Asteriden der Siboga-Expedition. 1. Porcellanasteridae, Astropectinidae, Benthopectinidae. Siboga Exped. 46" : 1-47, 7 figs, 13 pis. Downey, M. E. 1973. Starfishes from the Caribbean and the Gulf of Mexico. Smithson Contr. Zool. no. 126: 1-158, 48 pis. Farran, G. P. 1913. The deep-water Asteroidea, Ophiuroidea and Echinoidea of the west coast of Ireland. Sclent. Invest. Fish. Brch Ire. 1912 (6) : 1-66, 17 figs, 2 pis. Fisher, W. K. 1906. The starfishes of the Hawaiian Islands. Bull. U.S. Fish. Commn 1903 (3): 987-1 130, 49 pis. 1910. New starfishes from the North Pacific. Zool. Anz. 35 : 545-553; 568-574. 1911. Asteroidea of the North Pacific and adjacent waters. 1. Phanerozonia and Spinulosa. Bull. U.S. natn. Mus. 76 (1) : 1-419, 122 pis. 1919. Starfishes of the Philippine Seas and adjacent waters. Bull. U.S. natn. Mus. 100 (3) : 1-546, 156 pis. 1940. Asteroidea. Discovery Rep. 20 : 69-306, 23 pis. Goto, S. 1914. A descriptive monograph of Japanese Asteroidea. J. Coll. Sci. imp. Univ. Tokyo 29: 1-808, 19 pis. Grieg, J. A. 1921 [reprinted 1932]. Echinodermata. Rep. scient. Results Michael Sars N. Atlant. deep Sea Exped. 3 (2) : 1-47, 5 pis. Hayashi, R. 1973. The sea-stars of Sagami Bay: collected by His Majesty the Emperor of Japan. Biological Laboratory, Imperial Household of Japan, xi + 1 14 pp., 13 figs, 18 pis. Jangoux, M. 1978. Les etoiles de mer (Echinodermata : Asteroidea) recoltees par le M/V Calamar au large des Guyanes et du Venezuela. Bull. zool. Mus. Univ. Amsterdam 6(13): 93-101 , 3 figs. FAMILY BENTHOPECTINIDAE 135 Koehler, R. 1907. Asteries, Ophiures et Echinides recueillis dans les mers australes par la Scotia. Zool Anz.32(6): 140-147. 1909. Echinodermes provenant des campagnes du yacht Princesse- Alice. Result. Camp, sclent. Prince Albert 34 : 1-3 1 7, 32 pis. 1924. Les echinodermes de mers d'Europe. 1 .xiii + 360 pp., 9 pis. Paris. Ludwig, H. 1903. Seesterne. Result. Voy. Belgica (Zool.) 72 pp., 7 pis. 1910. Notomyota, eine neue Ordnung des Seesterne. Sber.preuss. Akad. Wiss. 23 : 435^466. Macan, T. T. 1938. Asteroidea. Sclent. Rep. John Murray Exped. 4 (9) : 323^35, 12 figs, 6 pis. McKnight, D. G. 1975. Classification ofsomasteroids and asteroids (Asterozoa: Echinodermata). Jl R. Soc.N.Z.S: 13-19. Mortensen, T. 1904. In Schmidt, J. Fiskeriundersogelser ved Island og Faeroerne i Sommeren 1903. Skr. Kommn Havunders., Kbh. no. 1 : 16-24. 1927. Handbook of the echinoderms of the British Isles, ix + 47 1 pp., 269 figs. London. Okutani, T. 1969. Synopsis of bathyal and abyssal megalo-invertebrates from Sagami Bay and the south off Boso Peninsula trawled by the R/V Soyo-Maru. Bull. Tokai reg. Fish. Res. Lab. no. 57: 1-6 1,22 figs, 8 pis. Perrier, E. 1875. Revision de la collection de Stellerides du Museum d'Histoire Naturelle de Paris. 384pp. Paris. [Also published in: Archs Zool. exp. gen. 4(1875): 265-450; 5 (1876) : 1-104, 209-309.] 1881. Description sommaire des especes nouvelles d'Asteries. Bull. Mus. comp. Zool. Harv. 9: 1-31. 1884. Memoire sur les etoiles de mer recuillies dans la Mer des Antilles et le Golfe de Mexique. Nouv. Archs Mus. Hist. nat. Paris (2) 6 : 127-276, 10 pis. 1885. Premiere note preliminaire sur les Echinodermes recueillis durant les campagnes de dragages sous-marins du Travailleur et du Talisman. Annls Sci. nat. (6) 19 (8) : 1-72. 1 894. Stellerides. Exped. scient. Travailleur- Talisman. 43 1 pp., 26 pis. 1896. Contribution a 1'etude des Stellerides de 1'Atlantique Nord. Result. Camp, scient. Prince Albert II : 1-57, 4 pis. Sibuet, M. 1975. Asterides abyssales de 1'Atlantique sud. Bull. Mus. natn. Hist. nat. Paris (Zool.) no. 1 99: 28 1-296, 3 figs, 1 pi. Sladen, W. P. 1885. The Asteroidea. Rep. scient. Results Voy. Challenger. Narrative. 1 (2) : 607-617, figs 202-206. 1889. Asteroidea. Rep. scient. Results Voy. Challenger. Zool. 30 : 1-935, 1 18 pis. Spencer, W. K. & Wright, C. W. 1966. Asterozoans. In R. C. Moore, Treatise on invertebrate paleontology, Pt U : U4-107, 89 figs. Studer, T. 1883. Uber der Asteriden welche wahrend der Reise S.M.S. Gazelle gesammelt. Sber. Ges. naturforsch. Freunde Berlin 1883 (8) : 128-132. 1884. Verzeichniss der wahrend der Reise S.M.S. Gazelle um die Erd 1874-76 gesammelten Asteriden und Euryaliden. Phys. Abh. Akad. wiss. Bed. 1883, 1884 (3) : 1-64, 5 pis. Verrill, A. E. 1884. Notice of the remarkable marine fauna occupying the outer banks off the south coast of New England. No. 9. Am. J. Sci. 28:21 3-220. 1885a. Notice of the remarkable marine fauna occupying the outer banks off the south coast of New England. No 1 1. Am. J. Sci. 29 : 149-157. 18856. Results of the explorations made by the steamer Albatross off the northern coast of the United States in 1883. Rep. U.S. Fish. Commn 1883 : 503-543, pis 13-19. 1894. Descriptions of new species of starfishes and ophiurans. Proc. U.S. natn. Mus. 17 (1000): 245-297. 1895. Distribution of the echinoderms of north-eastern America. Am. J. Sci. 49: 127-141, 199-212. 1899. Revision of certain genera and species of starfishes with descriptions of new forms. Trans. Conn. Acad. Arts Sci. 1899 : 145-234, pis 24-30. 1914. Monograph of the shallow-water starfishes of the North Pacific coast from the Arctic Ocean to California. Harriman Alaska Ser. 14 : 1^08, 1 10 pis. 1915. Report on the starfishes of the West Indies, Florida and Brazil. Bull. Labs nat. Hist. St. Univ. la 7(1): 1-232, 29 pis. Walenkamp, J. H. C. 1979. Asteroidea (Echinodermata) from the Guyana shelf. Zool. Verh. Leiden no. 170: 1-97,31 figs, 20 pis. Manuscript accepted for publication 12 January 198 1 British Museum (Natural History) 1881-1981 Centenary Publications The Natural History Museum at South Kensington By W. T. Steam This book presents a full history of the Museum, its collections, Directors and eminent members of staff. It provides for the first time an account of the major research undertaken and insights into the personalities of the key people in the Museum's development and evolution. Co-published with William Heinemann. Alfred Waterhouse and the Natural History Museum By Mark Girouard Designed by Alfred Waterhouse in the 1 870's the Museum is surely one of London's most outstanding pieces of architecture. This attractively illustrated book describes the development of the design and highlights some of the building's most interesting features. British Museum Natural History By Peter Whitehead & Colin Keates The Museum is really a huge scientific research institution - acquiring, describing and classifying all manner of natural history material, both specimens and artworks. It has some of the richest collections of their kind in the world and it is these and how they are obtained and managed that are the subject of this book. The lavish, full colour illustrations and lively text will appeal to everybody interested in natural history. Co-published with Philip Wilson Ltd. Chance, Change and Challenge This multi-author twin volume work is one of the Museum's most ambitious publishing projects. In the first volume The Evolving Earth twenty scientists have been asked to summarise the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume The Evolving Biosphere Museum scientists have chosen an evol- utionary concept - speciation, convolution, biogeography etc and related this to the group of animals or plants in which they are specialising. Co-published with Cambridge University Press. Animal Identification- A Reference Guide VOLUME 1 : MARINE AND BRACKISH WATER. Edited by R. W. Sims. VOLUME 2: TERRESTRIAL AND FRESHWATER. Edited by R. W. Sims. VOLUME 3 : INSECTS. Edited by D. Hollis. These guides provide the reader with lists of primary sources of reference that can be used to identify (or lead to the identification of) living animals throughout the world. The references are arranged in systematic and geographical sequence in order to facilitate searching. Co-published with John Wiley & Sons Limited. Nature Stored Nature Studied A short guide complementing the exhibition of that name to be staged in the Museum throughout 1981 . It gives a fascinating insight into the history of the Museum collections and the research carried out "behind the scenes". Titles to be published in Volume 41 Anatomy and phylogeny of the Chinese Major Carps Ctenopharyngodon Steind., 1866 and Hypophthalmichthys Blkr., 1860. By Gordon Howes Morphology and systematics of some interior-walled cheilostome Bryozoa By P. L. Cook and P. J. Chimonides Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae. By Ailsa M. Clark Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural History) Miscellanea Zoology series Vol 41 No 4 26 November 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or jore of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Dr J. G. Sheals Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISSN 0007-1498 Zoology series Vol 41 No 4 pp 137-234 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 26 November 1981 Miscellanea Contents Observations on clonal cultures of Euglyphidae (Rhizopoda, Protozoa). By C. G. Ogden 137 The Nuttall and British Museum (Natural History) tick collections: lectotype designations for ticks (Acarina: Ixodoidea) described by Nuttall, Warburton, Cooper and Robinson. By J. E. Keirans&B. E. Brewster 153 A revision of the spider genus Hispo (Araneae: Salticidae). By F. R. Wanless . . 1 79 A revision of the spider genus Phaeacius (Araneae: Saliticidae). By F. R. Wanless . 1 99 The protractor pectoralis muscle and the classification of teleost fishes. By P. H. Greenwood & G. V. Lauder Jnr. 213 Observations of clonal cultures of Euglyphidae (Rhizopoda, Protozoa) Colin G. Ogden Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction The identification of small siliceous testate amoebae is often difficult due to three main factors, observations are made at the limits of the optical microscope, scarcity of specimens and inadequate descriptions. The introduction of scanning-electron microscopy for detailed surface examination has reduced the first of these obstacles, as this facility permits more accurate descriptions to be made of the shells of these animals. However, matching these more detailed descriptions with existing reports is often complicated as slight differences of shell structure have frequently prompted authors to propose new specific names, or new Varieties' or new 'forms'. The second factor can also be reduced by establishing clonal cultures in the laboratory; additionally such cultures enable one to study biology and variation in morphology. Previous reports (Hedley & Ogden, 1973, 1974; Hedley et al., 1974) dealt with the biology of four species, namely Euglypha acanthopora (Ehrenberg, 1841), E. rotunda Wailes, 191 1, E. strigosa (Ehrenberg, 1871) and Trinema lineare Penard, 1890. The present account describes the shell morphology and biology of four further species, based on clonal cultures, together with the redescription of a previous clone under a new specific name. Systematics The genera Euglypha and Assulina belong to the family Euglyphidae — the classification adopted here is that proposed by Levine et al., 1980 and Loeblich and Tappan, 1964: Superclass RHIZOPODA Von Siebold, 1845 Class FILOSEA Leidy, 1879 Order GROMIIDA Claparede and Lachmann, 1859 Superfamily EUGLYPHACEA Loeblich and Tappan, 1 96 1 Family EUGLYPHIDAE Wallich, 1864 shell hyaline, symmetrical, elongate, composed of rounded siliceous plates, aperture rounded or elongate; one nucleus. Euglypha acanthophora — In identifying specimens from a clonal culture as E. acanthophora Hedley et al. (1974) accepted the reported variation attributed to this species, although with some reservation because the extent of the variations was not substantiated by our obser- vations of other Euglypha species in culture. These differences mainly concerned the shape and position of the siliceous spines and the shape of the body plates, which are illustrated in Leidy's superb figures (Leidy, 1879) of E. alveolata — a species accepted by both Cash et al., (1915) and Penard (in Cash et al.) as a synonym of E. acanthophora — showing that the siliceous spines project either distinctly from the sides of the body or discretely from the aboral extremity, and that there are distinct tooth-like projections on the posterior edge of the body plates. Penard (1902) proposed two varieties to accommodate specimens with Bull. Br. Mus. not. Hist. (Zool.) 41(4): 137-151 Issued 26 November 1 98 1 137 138 C. G.OGDEN varying spines, and these were accepted by Cash et al., (1915) who also suggested that intermediate forms existed. The extent of these variations is given by Decloitre (1962) in a review of the genus Euglypha, which lists seven varieties of E. acanthophora, The limitation of our research was such that it was expedient at that time (Hedley et al., 1974) to consider our clonal culture to be representative ofE. acanthophora. The problem of designated names now arises with the establishment of another distinct clonal culture within the heterogenous description of E. acanthophora. It appears that the most recent culture is more closely allied to the description of E. acanthophora, whilst the earlier culture (Hedley et al., 1974) fits the description of E. a. var. brevispina given by Cash et al, (1915). However, this latter description does not agree with the original (Penard, 1902) as discussed on p. 140, so to avoid confusion with these variety names it is proposed that a new specific name, E. cashii, be given to the earlier clonal culture. Materials and methods Information on the source of each species is given with the taxonomic description. The animals were obtained from crude cultures made by isolating small portions of each sample placed in a small plastic container and covered with a shallow layer of the culture liquid. Agnotobiotic cultures were kept at room temperature (18-20°C), in similar containers, on a thin layer of agar agar (1 per cent in distilled water) with a sterilised wheat grain added prior to setting and covered by a shallow layer of culture medium. The culture medium was a 5 per cent solution of soil extract plus added nutrient salts (see Hedley & Ogden, 1973). Clonal cultures were established by isolating single active animals, and one clone has been used subsequently to produce the working cultures of each species. Subcultures made at 2-3 week intervals are adequate to maintain active animals. Live specimens were examined by optical microscopy using both phase-contrast and brightfield illumination. Specimens for scanning electron microscopy were cleaned by transferring them through several changes of distilled water. Then manipulated using a single-hair brush onto small drops of Araldite adhesive on a small cover slip, the cover slip having been previously cleaned and dried. The exception to this procedure were specimens of E. dickensii which proved to be fragile and collapsed using this technique. Alternatively they were fixed in 3 per cent glutaraldehyde in 0 1 M cacodylic acid buffer, rinsed in the buffer solution and several times in distilled water, then specimens in distilled water were micropipetted onto a clean cover slip and allowed to dry. Although some specimens still collapsed, the majority retained their natural shape. The prepared cover slips were fastened to aluminium stereoscan stubs with Araldite, prior to being coated evenly with a thin layer of conducting metal. The stubs were examined on a Cambridge Stereoscan 180 operating at lOkV and the results recorded on Ilford HP5 film. Morphology Euglypha acanthophora (Ehrenberg, 1841) syn. Euglypha alveolata (Dujardin, 1 84 1 )— in Leidy, 1879, pi. 35, figs 3, 7, 15-18 Euglypha brachiata Penard, 1902 (Non Leidy, 1878) Euglypha brachiata \ar.flexulosa Penard, 1902 Euglypha brachiata var. brevispina Penard, 1 902 (Non Cash et al., 1915) Euglypha armata Wailes and Penard, 1911 Euglypha crenulata var. minor Wailes, 1912 Euglypha acanthophora var. dorsalis Schonborn, 1962 This species was isolated from a sample of Sphagnum moss collected at Holmsley, near Burley, New Forest, Hampshire, in March 1979. CLONAL CULTURES OF EUGLYPHIDAE 139 Figs 1-5 Euglypha acanthophora: Figs 1 & 2, lateral views to illustrate the distribution of body and elongated plates x 1250 & x 970; Fig. 3, view of aperture with nine apertural plates x 1 170; Fig. 4, single apertural plate, note the slight thickening on the dentate margin x 7500; Fig. 5, terminal or aboral view showing the distribution of oval body plates x 1200. 140 C. G. OGDEN Table 1 Range of measurements (in length breadth depth diameter of aperture Euglypha acanthophora Euglypha cashii (35) a.* (53) 54-4-68-6 65-1-84 26' 33' 4-33-8 6-42 12' 16 8-16 8-25' 5 2 b. (10) 67-88-9 33' 9-39-8 15 3-19 •2 Euglypha compressa (31) 60- 8-74-7 36 2-46-4 19' 7-25 6 12 1-17 •1 Euglypha dickensii (32) 35' 3-53-5 20 7-34-4 14' 1-18 9 9' 1-14' 1 Assulina muscorum (33) 33' 7-54 25' 2-36-6 15 4-19 •6 7' 7-14' 1 * measurements from Hedley el ai, 1 974; bracketed figures indicate the number of specimens measured, a. Clone from Hertfordshire, b. clone from Wales. DESCRIPTION. The shell is elliptical or ovoid, circular in cross section, composed of about one hundred and thirty body plates and up to six elongated plates or spines (Table 1 , Figs 1 & 2). The aperture is terminal, circular and surrounded by between eight and eleven, usually evenly spaced, denticulate apertural plates (Fig. 3). Each apertural plate is roughly oval, between 8'1-10'4 urn in length and 6'3-7-8 um in width, often both the denticular and the opposite margin are pointed (Fig. 4). The dentate edge is barely thickened on the inner surface around the median tooth and the five smaller lateral teeth positioned on each side (Fig. 4). Denticular margins are also seen on the second circle of plates surrounding the aperture. The body plates are arranged in alternate, longitudinal rows, and range from 7'5-9'8 urn in length and 6'3-8- 1 urn in width. Usually the posterior margin of these roughly oval plates has a median rounded projection, this appears to be more pronounced when the plate has an angular shape (Fig. 6). Body plates appear to vary in shape depending on their position, the pointed posterior margin of the plate becoming more pronounced in progres- sion down the length of the body (Fig. 1), only to be rounded in some instances at the aboral extremity (Fig. 5). The elongated plates are dagger shaped, ranging in length from 18-30 urn, and usually project from the sides of the body at an acute angle (Figs 1 & 2). They may be positioned anywhere in the posterior half of the shell, but are only rarely seen projecting from the extremity. The elongated portion of these plates is thin, hence their brightness when bombarded by electrons in the scanning electron microscope (Figs 7 & 8), and vary markedly in breadth which in turn alters the degree of tapering. It would appear that narrow and long spines are more flexible (Fig. 7), whilst the broader and shorter spines are more robust (Fig. 8). DISCUSSION. Both Cash and Penard described E. acanthophora with long thin spines and E. acanthophora var. brevispina with short stout spines. There is good agreement between the diagrams of E. acanthophora given by Cash (1915: pi. 35, fig. 2), Leidy (1879 pi. 35 figs 3, 7, 15-18) and Penard (1902 p. 505 and in Wailes & Penard, 1911: pi. 3, fig. 16), but not between the diagrams for E. a. var. brevispina, those of Cash and Leidy (1915: pi. 33, figs 3, 5; 1879 pi. 35, figs 1, 2, 4) being quite different from the original Penard figures (1902 p. 505 figs 6-10). Whereas the spines shown by the former authors are pointed, Penard's are blunt and appear to have broken or ragged edges. As Penard's two varieties — brevispina and flexulosa — differ only in having as their names imply either truncate or flexous spines, features which are considered to be artifacts from observations on clonal cultures reported here, they are listed as synonyms of E. acanthophora. Although the apertural plates in the above mentioned descriptions are similar, there are differences in the distribution and shape of the body plates. For example, Leidy (1879) noted that there were two rows of denticulate plates surrounding the aperture and that the body plates had small projections from their posterior margins. Both Cash and Penard agree with CLONAL CULTURES OF EUGLYPHIDAE 141 Figs 6-8 Euglypha acanthophora: Fig. 6, area of shell surface showing typical body plates x 3300; Fig. 7, 'dagger-shaped' elongated plate, a narrow and flexible type x2500; Fig. 8, a stout and robust type of elongated plate x 2700. the two rows of denticulate plates, but neither commented on the body plates-except for a reference in the generic discussion (Cash et al, 1915: p. 5) naming just three species with 'scutiform' plates : Zs. scutigera Penard, 1911; E. aspera Penard, 1899 and E. crenulata Wailes, 1911 — nevertheless they synonymize Leidy's description with E. acanthophora. Wailes (1912) in describing a new variety, E. crenulata var. minor, noted that it was only distinguished from E. armata Wailes & Penard, 1911 (a synonym of E. acanthophora proposed by Wailes in Cash et al., 1 9 1 5) by the shape of the body plates. Of the eleven varieties listed by Decloitre (1962, 1976), two have already been mentioned, E. a. var. brevispina and E. a. \ar.flexulosa; four appear to be similar to E. acanthophora as described here, these are E. a. var. cylindracea (Playfair, 1918), E. a. var. dorsalis Schonborn, 1962, E. a. var. fantastica Decloitre, 1965 and E. a. var. longispina Decloitre, 1969; two, E. a. var. equeis Decloitre, 1956 and E. a. var. heterospina Decloitre, 1949, have discrete arrangements of spines and may represent distinct species; and the remaining three are considered later on p. 143. The four similar varieties differ in size (cylindracea), slight deflection of the aperture (dorsalis), the shape of the spines (fantastica), and the number of spines (longispina). Specimens with misaligned apertures are not uncommon in clonal culture, and E. a. var. dorsalis is therefore considered to be a synonym of E. acanthophora. The differences of the three remaining varieties may also be artificial, but it is not possible to comment further on these. Euglypha cashii nom. nov. syn. Euglypha acanthophora var brevispina — in Cash et al., 1915 Euglypha alveolata — in Leidy, 1879: pi. 35, figs 1, 2, 4 Euglypha alveolata var. cirrata Wailes, 1912 Euglypha acanthophora var. deflanderi Decloitre, 1956 Euglypha acanthophora var. elegans Stepanek, 1963 Euglypha acanthophora — in Medley et al., 1974 Euglypha acanthophora — in Ogden & Hedley, 1980 Two clones of this species have been isolated, the first from sewage sludge from the Maple Lodge Works of the Colne Valley Sewage Board, Hertfordshire, in December, 1972 and the second from a sample of Sphagnum moss from Myndd Hiraethog, Denbigh, Clwyd, North Wales, in August, 1980. 142 C. G. OGDEN Figs 9-13 Euglypha cashii: Fig. 9, lateral view showing the distribution of body and elongated plates x 1250; Fig. 10, aperture with ten evenly distributed apertural plates x 1100; Fig. 11, detail of apertural plates x 4400; Fig. 12, elongated plates projecting from aboral extremity x 1 800; Fig. 13, elongated plates following the curvature of the shell x 2400. CLONAL CULTURES OF EUGLYPHIDAE 143 DESCRIPTION. The shell is elliptical or ovoid, circular in cross section, composed of about two hundred body plates and up to six elongated plates (Table 1 , Fig. 9). The aperture is terminal, circular and surrounded by between ten and thirteen evenly spaced denticulate apertural plates (Fig. 10). Each apertural plate is roughly oval, varying from 8-1-12-5 um in length and 7-2-1 1-5 um in width. The dentate edge is slightly thickened and carries a median tooth with either four or five smaller lateral teeth on each side (Fig. 1 1). Specimens are sometimes seen with the average number of plates surrounding the aperture but with either one or two being dentate body plates, similar to those shown in Fig. la of Hedley et al. (1974), instead of dentate apertural plates. The body plates are arranged in alternate longitudinal rows and range from 10-9-12 um in length and 7-9-9-1 um in width. Each oval body plate has a median rounded projection on the posterior margin, these projections are not as pronounced as in E. acanthophora (compare Figs 1 & 9). The elongated plates project slightly from the aboral extremity, or closely follow the outline of the shell in the aboral region, they range from 19-6-29-2 um in length and 6- 1-7-6 um in width (Figs 12 & 13). DISCUSSION. The description of this species is in good agreement with that given by Cash et al. (1 9 1 5: pi. 33, figs 3 & 5) for E. a. var. brevispina. As there is some disparity between these two descriptions and the original description of the variety brevispina given by Penard (1902) (see also p. 140), in addition to the variety name brevispina being used with other species of Euglypha, it was considered that a new specific name would eliminate future confusion of these species. E. cashii is distinguished from E. acanthophora by size, shape of body and elongate plates, but mainly by the positioning of the elongated plates. The three varieties of E. acanthophora listed by Decloitre (1962, 1976) which are similar to E. cashii are E. a. var. cirrata: E. a. var. deflanderi and E. a. var. elegans, they apparently differ only in the deflection of the posterior spines or elongated plates to the shell body. Such features have been seen in clonal culture as a normal variation, and these varieties are there- fore considered to be synonyms of E. cashii. Another species, E. tiscia Gal, 1969, listed by Decloitre (1976), although similar to E. cashii is considered to represent a valid species because of its larger dimensions. ETYMOLOGY. This species is named after Mr James Cash, who contributed so much to the taxonomy of testate amoebae at the beginning of this century. Euglypha compressa Carter, 1 864 This species was isolated from a sample of Sphagnum moss, collected at Subberthwaite, Broughton in Furness, Cumbria, in July, 1978. DESCRIPTION. The shell is ovoid, laterally compressed and composed of about two hundred body plates and thirty spines (Table 1, Fig. 14). The aperture is terminal, circular or elliptical and surrounded by between eleven and fifteen, evenly spaced, denticulate apertural plates (Fig. 15). Each apertural plate is oval, between 6-2-7-1 um in length and 4-6-5-4 um in width. The dentate edge is distally thickened around the large median tooth, but this thickening tapers evenly outwards to the three smaller teeth positioned on each side and equates with the normal shell thickness close to the last tooth (Fig. 16). The body plates are arranged in sequence and range in size from 7'4-8'2 um in length and 3-4-4-4 um in width. They are roughly oval and often characterised by their hexagonal appearance, this latter feature is variable but the squared nature of the narrow margins is usually distinctive (Fig. 17). The spines are positioned along the lateral margins from about the mid-body position to the aboral extremity (Fig. 15). They are sometimes in pairs and usually alternate when viewed laterally, one pointing upwards and the next downwards along the length of the body (Figs 14 & 15). Each spine has a narrow base from which it tapers to its widest point about a quarter of the total spine length, then it tapers gradually over the remaining threequarters to the terminal point (Fig. 1 8). DISCUSSION. The only recorded variety of this species is E. compressa f. glabra Cash et al., 144 C. G. OGDEN Figs 14-18 Euglypha compressa: Fig. 14, lateral view showing the distribution of body plates and spines x 1250; Fig. 15, latero-apertural view to illustrate the circular aperture and the positioning of the spines x910; Fig. 16, detail of single apertural plate, note the thickening around the dentate margin x9500; Fig. 17, arrangement of typical body plates x2350; Fig. 18, illustration of spine projecting from organic cement matrix between two adjacent body plates x 4400. CLONAL CULTURES OF EUGLYPHIDAE 145 1915, a form without spines. However, variation in the shape and size of the siliceous spines was also reported by Cash et al. for E. compressa, and such differences were considered to be a normal feature. This may not now prove to be the case, if, as the information derived from clonal cultures in the present work shows that differences in spine positioning is specific, then it would appear to follow that structural differences of spine construction may also be specific. The spines described here are similar to those reported by de Graaf (1956) and Brown (1910), but they are different from those described by Ogden & Hedley (1980). In the latter report the spines are stout at the base, concave, and taper evenly to a point (see PI. 78D, Ogden & Hedley, 1980). Both types of spine are here considered to represent E. compressa, which is distinguished by the compressed shell and lateral position of the spines, but further work on clonal cultures may establish spine shape as a specific character. Euglypha dickensii sp. nov. This species was isolated from a sample of damp moss taken from underneath sweet chestnut trees at Cobham woods, Rochester, Kent in February, 1974. DESCRIPTION. The shell is ovoid, laterally compressed and composed of about two hundred and eighty elongate body plates (Table 1, Figs 19 & 20). The aperture is terminal, oval or circular, and surrounded by between eleven and fifteen, evenly spaced, denticulate apertural plates (Fig. 21). In a few specimens additional apertural plates are seen in the second circle of plates around the aperture. Each apertural plate is roughly circular, from 5-4-6-4 um in length and 3'8-4'4 um in width. The dentate edge has a large, thick, distinctly curved, median process with a terminal pointed tooth, this is flanked on each side by a medium outward facing tooth and one or two smaller teeth (Figs 22 & 24). The denticular thickening equates with the normal shell thickness at the position of the small teeth. The body plates are elongate, ranging from 5*1-6-2 um in length and 1-6-2-5 um in width, and are arranged in alternate longitudinal rows. Around the mid-body region there are some randomly distributed pointed body plates (Fig. 23). These pointed plates vary from normal plates with a small sharp spike, about 6*9 um long, to tapered spines about twice the length, 9- 1 um, of a normal body plate. Although there is variation in the dimensions of these pointed body plates, their presence in a mid-body position is a reasonably stable feature. DISCUSSION. The species described here is similar to three species previously reported from soil samples, namely E. capsiosa Couteaux, 1978, E. cuspidata Bonnet & Thomas, 1960 and E. simplex Decloitre, 1965. It has similar dimensions to E. simplex but differs in the shape of the apertural plates, which in the latter species have a distinct diamond-shape (see Couteaux et al, 1 979). Both E. capsiosa and E. cuspidata are smaller species with fewer body and apertural plates, and again differ in the shape of the apertural plates. E. dickensii is distinct in size, dentation of the apertural plates and the presence of pointed body plates in the mid-body region. ETYMOLOGY. As this species was found in the countryside frequented by Charles Dickens, the nineteenth century author, and subsequently featured in many of his novels, it is named in his honour. Assulina muscorum Greef, 1888 This species was isolated from a sample of dry moss on soil, collected at Rolestone Farm, Banwell, Somerset in February, 1974. DESCRIPTION. The shell is ovoid, laterally compressed and composed of about two hundred shell plates (Figs 25 & 26). The aperture is terminal and surrounded by between ten and fourteen shell plates arranged in a rather irregular manner, most with their minor axis bordering the opening but often a few have their major axis (Figs 27 & 28). The opening is edged with a thin band of organic cement, this band is frequently thickened on the tips of some plates to form tooth-like projections (Fig. 28). The shell plates are oval, ranging from 5'8-6'8 um in length and 2-5-3-1 um in width. They are usually arranged in alternate, 146 C. G. OGDEN Figs 19-24 Euglypha dickensii: Fig. 19, lateral view showing the distribution of body plates x 1970; Fig. 20, view to illustrate the lateral compression x 1300; Fig 21, aperture with fourteen apertural plates x 2400; Fig. 22, view showing the overlapping of the apertural plates and the thickness of the median tooth x 7800; Fig. 23, portion of shell surface with two pointed body plates x 4600; Fig. 24, circular apertural plate with typical dentate margin x 1 1000. CLONAL CULTURES OF EUGLYPHIDAE 147 27 Figs 25-28 Assulina muscorum: Fig. 25, lateral view to illustrate the distribution of shell plates x2150; Fig. 26, view to show lateral compression and arrangement of plates x 1400; Fig. 27, apertural view x 2100, Fig. 28, detail of 'tooth-like' projections of organic cement around the aperture x 4900. 148 C. G. OGDEN longitudinal rows, with their major axis parallel to the major axis of the body. However, in some instances the axes of the plates are not parallel with the body and in these cases the general pattern is altered. The aboral extremity is also subject to irregular arrangements of shell plates. DISCUSSION. The shell of A. muscorum is reported as being usually brown but occasionally colourless in the wild, whereas in culture it is mainly colourless. Nevertheless, live animals have a distinct band, probably the 'pigment zone', at the mid-body region which is often so large that it may tend to give colour to the shell. Variation in structure of the shell appears to be limited to the formation of an extra large individual, usually confined to less than three per cent of the population. Such specimens have more than the normal compliment of shell plates but the arrangement is the same. Similar large specimens have been reported (Hedley & Ogden, 1973) in clonal cultures of Euglypha rotunda. Reproduction The formation of a daughter-cell by simple division has been observed for all the described species and follows the same pattern in each, the only variation being the additional arrange- ment of elongated plates or spines in E. acanthophora, E. cashii and E. compressa. The sequence of events in E. compressa are described. Euglypha compressa The onset of division begins with the protrusion of a short thick cytoplasmic extension from the parent aperture. The apertural plates are the first to be passed from the storage area adjacent to the nucleus, via the peripheral cytoplasm to become arranged in a circle around this cytoplasmic extension. The remainder of the plates follow the same route and are arranged in sequence, in distinct rows radiating from the apertural plates (Fig. 29a). Each plate is added on the inside of the previous plate so that there is a considerable overlap. This excessive overlapping of the plates in the region of attachment is very noticeable in the early stages of shell construction. All the body plates are therefore in position, but not correctly spaced, well in advance of the shell attaining its full size. When the daughter shell has attained a size about two-thirds that of the parent, vacuolar activity is seen in the anterior third of the parental cytoplasm. This activity proceeds into the daughter cytoplasm as it increases in volume, at the same time the zone of pigment granules begin to extend towards the aperture of the parent. The spines are the last elements to be passed from the parent to the daughter (Figs 29b & c), where they are arranged centrally in the cytoplasm parallel to the shell walls. The pair of mid-body spines are the first to be pushed by cytoplasmic Fig. 29 Diagram illustrating division in Euglypha compressa: formation of daughter shell (a-e) followed by nuclear division (f-1). a. body plates (bp) pass via the peripheral cytoplasm to extremity of cytoplasmic extension of parent, 10-15 minutes; b. as last plates pass into forming daughter shell, siliceous spines (ss) begin moving from parent, 20 minutes; c. spines move into daughter shell, cytoplasmic vacuoles (v) appear close to parent aperture and granules of pigment zone (pz) begin to move, 25 minutes; d. spines arranged for positioning in daughter, and granules in parent move towards aperture, 27 minutes; e. daughter shell full size with some spines in position and cytoplasm filled with cytoplasmic vacuoles, granules begin to extend from parent into daughter, 30 minutes; f. daughter shell complete, granules in position between the opposed apertures and nucleus (n) in parent has small polar extensions, 40 minutes; g. elongation of parent nucleus, 55 minutes; h. granules mainly in daughter and nucleus now 'diamond-shape', 60 minutes; i. elongated nucleus becoming indistinct behind moving granules, 70 minutes; j. arrows depict pathways of rapidly moving granules, 75 minutes; k. movement of granules has slowed down and zones beginning to reform, faint nucleii appear, 90 minutes; 1. contractile vacuoles (cv) and nucleii apparent, pig- ment zones almost reformed, 100 minutes. The times given are based on an estimated starting point. CLONAL CULTURES OF EUGLYPHIDAE 149 150 C. G.OGDEN movement out between adjacent, lateral body plates, they appear to be positioned on each side simultaneously. This pair are followed in succession by other pairs of spines in a progressive sequence towards the aboral region. The last spines to be positioned are those that project from the terminal region. These spines are manoeuvred in the cytoplasm (Fig. 29e) until they are parallel to the aboral shell surface, they are then pushed through and cytoplasmic movement continues until the spines are at right angles to the shell surface. Throughout the time taken to position the spines those that project from the shell are in continuous movement, backwards and forwards like oarsmen in a rowing-boat except that their movements are not synchronized, and there are smaller movements still visible amongst the body plates. Vacuolar activity seems to be concentrated mainly in the daughter with the pigment granules being pushed from the parent into the peripheral cytoplasm of the daughter. A halo effect is most noticeable around the parent nucleus at this time. A few seconds after the last aboral spines are in position, movement of all the shell elements slows down, until the spines are projecting slightly posteriorly when viewed from above. All movement has stopped after a further two minutes from the final positioning of the last aboral spines, and the daughter shell is complete. At about the same time the parent nucleus begins to elongate, initially there are two small polar extensions (Fig. 290, but this changes into a 'diamond-shape' as most of the granules are passed into the daughter cyto- plasm (Fig. 29g). Slightly later, movement of the granules in the daughter cytoplasm seem to suggest that when they reach the aboral extremity they are reflected back via the peripheral cytoplasm to the parent, meanwhile the nucleus has become more elongate (Fig. 29h). Rapid movement of the granules now obscures the changing nucleus (Fig. 29i), and a regular pathway of constantly moving granules is formed between the parent and daughter cyto- plasm (Fig. 29j). When the rapid movement ceases, after about five minutes it is replaced by a slow regular movement with no apparent directional flow, cytoplasm of both cells looking homogenous. After a further five minutes the apertural region of both are relatively free of granules, contractile vacuoles are apparent in the anterior region of each and there is some movement of granules (Fig. 29k). A little while later nuclei are apparent in the posterior region of both cells as the pigment granules are concentrated into the mid-body region (Fig. 291). Just prior to separation the cytoplasm in each shell is withdrawn slightly so that the cytoplasmic connection is severed, the animals move apart independently shortly after this action. The approximate time taken to produce the daughter shell is forty-five minutes, whilst the total time for division is one hundred minutes. Doubling time Estimations on the length of time required to double the population (doubling time) were made on cultures established and maintained under similar conditions. Growth curves calculated from records of daily counts of individuals were made on three replicate cultures of four species, but only one culture was available for E. acanthophora. The results are given in Table 2. Table 2 Estimates of doubling time (in days). Euglypha acanthophora 2*7 Euglypha cashii 2-3-2-8 Euglypha compressa 4-0-4-4 Euglypha dickensii 2-6-3- 1 Assulina muscorum 2-3-2-9 References Bonnet, L. & Thomas, R. 1960. Faune terrestre et d'eau douce des Pyrenees-Orientales. Thecamoebiens du sol. 103 pp. Herman, Paris. CLONAL CULTURES OF EUGLYPHIDAE 1 5 1 Brown, J. M. 1910. Freshwater rhizopods from the English Lake district. J. Linn. Soc. Zool. 30 : 360-368. Carter, H. J. 1864. On Freshwater Rhizopoda of England and India; with Illustrations. Ann. Mag. nat. Hist. (3)13: 18-39. Cash, J., Wailes, G. H. & Hopkinson, J. 1915. The British Freshwater Rhizopoda and Heliozoa. Vol. 3. Rhizopoda, part 3. 1 56 pp. The Ray Society, London, rouleaux, M-M. 1978. Quelques Thecamoebiens du sol du Japon. Revue Ecol. Biol. Sol. 15: 119-126. Munsch, A. & Ponge, J-F. 1979. Le genre Euglypha:essai de taxinomie numerique. Protistologica 15 : 565-579. Decloitre, L. 1949. Materiaux pour une faune rhizopodique d'A.O.F. Bull. Inst. fr. Afr. noire 11:281-301. 1956. Les Thecamoebiens de I'Eqe Greenland. Hermann, Paris. 1962. Le genre Euglypha Dujardin. Arch. Protistenk. 106 : 5 1-100. 1965. Amoebida testacea (Rhizopoda). Zoology Iceland! (1) : 1-58. 1976. Le genre Euglypha. Complements a jour au 31 decembre 1974 de la Monographie du genre parue en 1962. Arch. Protistenk. 118 : 18-33. Ehrenberg, C. G. 1841. Verbreitung und Einflufs des microskopischen Lebens in Siid-und Nord- Amerika. Phys. Math. Abh. K. Akad. Wiss. Berlin 291-446. Gal, D. 1969. Zooplanktonuntersuchungen im Ostlichen-Hauptkanal. Acta biol. Szeged 15 : 93-100. Graaf, Fr. de. 1956. Studies on Rotatoria and Rhizopoda from the Netherlands. I. Rotatoria and Rhizopoda from the 'Grote Huisven'. Biol. Jaarb. 23 : 145-2 1 7. Greef, R. 1888. Studien iiber Protozoen. Sber. Ges. Befordges. Naturw. Marburg : 90-158. Hedley, R. H. & Ogden, C. G. 1973. Biology and fine structure of Euglypha rotunda (Testacea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 25 : 1 19-137. 1974. Observations on Trinema lineare Penard (Testacea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 26 : 187-199. & Krafft, J. I. 1974. Observations on clonal cultures of Euglypha acanthophora and Euglypha strigosa (Testacea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 27 : 103-1 1 1. Leidy, J. 1879. Freshwater Rhizopods of North America. In Vol. 12, United States Geological Survey of the Territories. 324 pp. Washington. Levine, N. D. et al. 1980. A newly revised classification of the Protozoa. J. Protozool. 27 : 37-58. Loeblich, A. R. & Tappan, H. 1964. Thecamoebians'. In Treatise on Invertebrate Palaeontology, Part C, Protista 2, Vol. 1 , C 1 6-C54. The Geological Society of America. Ogden, C. G. & Hedley, R. H. 1980. An Atlas of Freshwater Testate Amoebae. 222 pp. British Museum (Nat. Hist), London & Oxford University Press, Oxford. Penard, E. 1899. Les Rhizopods de Faune profonde dans le Lac Leman. Revue suisse Zool. 7 : 1-142. 1902. Faune Rhizopodique du Bassin du Leman. 700 pp. Geneva. Schonborn, W. 1962. Neue Testaceen aus dem Groben Stechlinsee und dessen Umgebung. Limnologica 1 : 83-91. StCpanek, M. 1 963. Die Rhizopoden aus Katanga (Kongo-Afrika). Annls. Mus. r. Afr. cent. 1 17 : 9-9 1 . Wailes, G. H. 1912. Freshwater Rhizopoda and Heliozoa from the States of New York, New Jersey and Georgia, U.S.A.; with supplemental note on Seychelles species. J. Linn. Soc. Zool. 32 : 121-161. & Penard, E. 19 1 1 . Clare Island Survey. Part 65. Rhizopoda. Proc. R. Ir. Acad. 31 : 1-64. Manuscript accepted for publication 12 December 1980 The Nuttall and British Museum (Natural History) tick collections: lectotype designations for ticks (Acarina: Ixodoidea) described by Nuttall, Warburton, Cooper and Robinson James E. Keirans Rocky Mountain Laboratories, Hamilton, Montana 59840, U.S.A. Bernice E. Brewster Arachnida Section, Zoology Department, British Museum (Natural History), London SW7 5BD Introduction A monographic revision of the Ixodoidea was begun by Nuttall et al. (1908) with subsequent volumes published by Nuttall & Warburton (1911, 1915) and Robinson (1926). A further volume was produced by Arthur (1960). These works were based largely on material gathered by Nuttall and his associates and the Entomological Research Committee for Tropical Africa, forerunner of the present day Commonwealth Institute of Entomology. During his lifetime, Professor Nuttall presented a considerable number of specimens to his correspondents. In 1939, the remainder of the Nuttall Collection was donated by the Molteno Institute to the British Museum (Natural History) and at that time it was the largest tick collection in the world. Currently, the Nuttall collection is kept separately from the main tick collection of the British Museum (Natural History), and types of Ixodoidea designated by Nuttall et al. may be found in both collections. Prior to the formulation of the International Code of Zoological Nomenclature, type specimens were frequently labelled only as 'TYPE' or 'COTYPE' with the result that many of the taxa described by Nuttall et al. were in need of lectotype designations. Consequently these studies were undertaken and form the basis of the present report. At the same time an attempt was made to trace the type material of all taxa described by Nuttall and his associates; appendices are therefore provided listing holotypes, previously selected lectotypes and other types together with their location, where known. The following list of lectotypes is in alphabetical order by specific name within the genera Argas and Ornithodoros in the family Argasidae and within the genera Amblyomma, Aponomma, Dermacentor, Haemaphysalis, Ixodes, Rhipicentor, and Rhipicephalus in the family Ixodidae. First the species binomen, or in a few cases the trinomen as originally described, is given followed by author(s), date, page, and figure(s). Any currently recognized synonyms and the Nuttall tick catalogue number precede a statement detailing all pertinent information relating to the original collection. Next, the current status of the type material is given, including selection of a lectotype and paralectotypes and an updating of the host and locality data. This is followed by subsequent published information relating to the species in question. Finally, there is a brief statement which summarizes hosts and distribution of each tick species. During the course of the studies reported below, each lectotype and each paralectotypic series was isolated within the vial or jar containing specimens of the taxon; the nomenclative status of this material was clearly indicated on an additional included label. Bull. Br. Mm. not. Hist. (Zool.) 41(4): 1 53-1 78- Issued 26 November 198 1 153 154 J. E. KEIRANS & B. E. BREWSTER ARGASIDAE Argas reflexus indicus Warburton, 1910: 396. NUTTALL 1649. The Indian Museum collection contained three specimens of this form. One was found in a book in the Entomological room, and another on a wall of the Museum building. The third specimen was found in a box sent from the Museum to Kurseong, E. Himalayas. The birds building in the roof of the Museum were identified as Passer domesticus and Cypselus affinis. The type species has not been recorded from India.' TYPES. Lectotype 9 Argas reflexus indicus (N1649) found in a book, Entomological room, Indian Museum, Calcutta (22'32'N, 88°22'E), India, 15.111.1910, C. A. Paiva. [Date and collector information taken from vial label]. REMARKS. A second collection (N1243) consisting of 1 cf, 1 9, 1 N A. reflexus indicus, found in basin of water, Entomological room, Indian Museum, Calcutta, 13. IV. 19 10, F. H. Gravely, has a blue label with the word TYPE. We have not accorded these specimens paralectotype status because the specimen numbers and locality data do not conform to Warburton 's published record. Argas reflexus is not known to occur east of 15°00' longitude and without additional collections, the exact taxonomic status of A. r. indicus remains moot. HOSTS & DISTRIBUTION. Argas reflexus indicus is an avian parasite known only from India. Ornithodoros gurneyi Warburton, 1926 : 5 5, fig. 1 NUTTALL 3562. 'Described from two specimens taken in an old shed at Tibboburra, N.S.W., 24. XI. 1922, by Mr W. B. Gurney and sent with other ticks by the Government Entomologist of N.S.W. in 1924 to the Imperial Bureau of Entomology, where it was numbered 1378.' TYPES. Lectotype cf Ornithodoros gurneyi (N3562) taken in an old shed, Pindera Selection, Tibboburra, New South Wales, Australia; 1 N paralectotype Ornithodoros gurneyi, date as above, deposited in the BM(NH) 97 collection (1925.5.9.4). HOSTS & DISTRIBUTION. Ornithodoros gurneyi is primarily a parasite of large macropodids (kangaroos and wallaroos) but will also attack man and dogs in most areas of Australia. Ornithodoros piriformis Warburton, 1981 : 285, fig. 2 NUTTALL 3335. 'Described from 15 specimens taken at Mahabaleshwar, Satara District, at a height of 4200 ft, (?host)by F. H. G., 13-16.IV.12. Indian Mus. SPA./10.' TYPES. Lectotype 9 and 2 9, 1 cf paralectotypes Ornithodoros piriformis (N3335) host unknown, Robber's Cave, Mahabaleshwar (17°55'N, 73°40'E), Maharashtra, India, at a height of 1280 m, 13-16.IV.1912, F. H. G. REMARKS. The locality Robber's Cave is taken from NuttalPs catalogue and on the vial label with the ticks. Two slides, one with hypostome palp and leg I, the other with hypostome, are both missing. HOSTS & DISTRIBUTION. Ornithodoros piriformis has been collected from Miniopterus schreibersii from the type locality. Robber's Cave contains several other bat species including Rousettus leschenaulti and probably Rhinolophus affinis. This tick is known only from India. IXODIDAE Amblyomma atrogenatum Nuttall & Warburton, 1908 : 409, figs 30-32 Junior subjective synonym of Amblyomma clypeolatum Neumann, 1 899. NUTTALL & BM(NH) TICKS 1 55 NUTTALL 43. 'Description based on 6 d1 from Testudo elegans, received from India at the Zoological Gardens, London, 15. VIII. removed 2. X.I 905. (Preserved in spirit.)' TYPES. Lectotype d1 and 4 d1 paralectotypes Amblyomma atrogenatum (N43) ex Testudo elegans received from India at Zoological Gardens, London, 15.VHI.1905. Ticks removed 2.X.1905. REMARKS. The 4 rf paralectotype total includes 1 d pinned specimen. An empty token tube in the Nuttall collection indicates a slide mount of 1 cf which is now missing. HOSTS & DISTRIBUTION. Amblyomma clypeolatum is a tortoise parasite in India and Sri Lanka. Amblyomma caelatumm Cooper & Robinson, 1908 : 460, figs 6-9 Junior subjective synonym of Amblyomma geoemydae (Cantor, 1 847). NUTTALL 2938. Original description 'based on two d"s and one gorged 9. Collected by Mr F. F. Laidlaw, Skeat Expedition, 1899, Kwala Aring, Java. Host, a tortoise (Geoemyda spinosd). Preserved in alcohol.' TYPES. Lectotype 3 and 1 d1, 1 $ paralectotypes Amblyomma caelaturum (N2938), ex Heosemys spinosa, Kwala Aring, Malaysia (locality not verified) collected by Dr N. Annandale, 1899. Gift of Mr F. F. Laidlaw. Correction of locality and collector: Robinson (1926:230). REMARKS. The name A. caelaturum was published in July, 1908. The name A. maylanum Neumann was published in March of the same year. Anastos (1950 : 107) sank the former as a junior synonym of A. malayanum on the grounds of priority. After examination of all relevant types Kohls (1957 : 87) concluded that both A. caelaturum and A. malayanum were junior synonyms of A. geoemydae (Cantor, 1 847). HOSTS & DISTRIBUTION. Amblyomma geoemydae parasitizes reptiles, especially tortoises and monitor lizards in Indonesia, Malaysia, Thailand, Ryukyu Islands, Japan, Taiwan, and the Philippines. Amblyomma cooperi Nuttall & Warburton, 1908 : 410, figs 33-36 Junior synonyms: Amblyomma lutzi Aragao, 1908:112; Amblyomma ypsilophorum Schulze, 1941 :93. NUTTALL 70. Original description, 'based on 1 d1 and 5 9 collected from "Carpinchio" (Hydrochoerus capybard) at Puerto Cooper, Alto, Paraguay, S. America, by W. F. Cooper in 1904. (Dry specimens.)' TYPES. Lectotype d1 and 1 d, 6 cf paralectotypes Amblyomma cooperi, (N70), ex Hydrochoerus hydrochaeris Puerto Cooper (23°03'S, 57°43'W), Alto, Paraguay, 1904, W. F. Cooper. [Included in the above total is one slide-mounted 9]. REMARKS. Robinson (1926 : 75) indicated that more than 1 rf was present in the original collection but gave no total. HOSTS & DISTRIBUTION. Amblyomma cooperi is a parasite of capybaras and tapirs in Paraguay, Bolivia, Brazil, Uruguay, and Argentina. Amblyomma darlingi Nuttall, 1912 : 50, figs 1^ Junior synonym of Amblyomma oblongoguttatum Koch, 1844. NUTTALL 1285-1287. 'Described from 5 d1, 9 9, 40 N, and 23 larvae found on a deer (Odocoleus sp.), Corozal, Panama Canal Zone, 14.VIII.1910, by A. H. Jennings, of Ancon, and 2 N, found on the nape and head of a turkey buzzard (Catharista atratus), Empire, Canal 1 56 J. E. KEIRANS & B. E. BREWSTER Zone, IV. 191 1, by S. T. Darling, M. D., Chief of Laboratory, Isthmian Canal Commission, Ancon, Canal Zone. Named in honour of Dr Darling in view of his distinguished service to parasitology. The types, for which I am indebted to Dr Darling, are in Cambridge (N.1285-1287).' TYPES. Lectotype rf and 2 t/5 SB° D r- S-S 2 S'* 3 ~ >* .2" |5i 8 ^£ §50 o J2 oj • 0 ^ "3 2^ rj O % t> Cll^ |^| ?! 55S2 5?S'.S5 S rS T. (^> > W (^^ 2 oj J Q3 4> 5 ui C "£.£ § ^2 — o 6 c .2 is 'o !•§=& Ills ^^•E o = 'a7^ oo o *L c ^ a c j= «><£•£ ^^-•2^ as X) «a « - UJ > u- ** . JJ a 3 CA O • ~ <^ Si i -3 G S~C/5 "5 C.o U £ c o - 1.S1S o . o a __, w - •£ o. 3^ 2 " " a 3 3 a 228 P. H. GREENWOOD & G. V. LAUDER cypriniform taxa. It is well-developed in the three members of the anotophysi we dissected, particularly in Chanos chanos where it is a thick, deep muscle proximally, and has a broad, tendinous insertion onto the upper half of the cleithrum; in Gonorynchus gonorynchus it is narrow and strap-like, inserting directly onto the girdle, while in Phractolaemus ansorgii it is broad, thin and partly tendinous. Amongst those otophysans with a protractor pectoralis, the muscle, although intra- specifically constant, shows varying degrees of development; also, in some species it inserts directly onto the cleithrum, whereas in others it inserts onto the membrane extending from the girdle to form the posterior wall of the branchial chamber. When the protractor has a 'membrane insertion' it is closely applied to the anterior face of the body musculature which delimits the posterior boundary of the branchial chamber. In general, the protractor pectoralis is narrow and strap-like, and often has a near-vertical orientation. Vetter (1878) claims that the protractor pectoralis (his trapezius), is absent in Cyprinus carpio and Barbus barbus. The muscle certainly is present in C. carpio (see Winterbottom, 1974, fig. 22; personal observations), but may have been overlooked by Vetter because of its slenderness and its rather tendinous nature. A short, fine, protractor pectoralis is also present in a specimen of Barbus barbus we dissected, although the muscle is absent in another member of that genus, B. intermedius, from Kenya, east Africa. A levator posterior muscle appears to be absent in some anotophysans (Chanos chanos), but is present in Gonorynchus gonorynchus. Its absence in Chanos could be correlated with the development of an expansive and complex suprabranchial organ in that genus. The muscle is seemingly also absent in Phractolaemus ansorgii, but the condition of our material does not permit a definite conclusion on that point. All the otophysan taxa examined have a levator posterior muscle; it is particularly well- developed in members of the Cypriniformes (see also Winterbottom, 1974; 252-253, figs 20-22), but in other groups it is often a slender strap-like muscle. The occurrence of a levator posterior muscle in the Ostariophysi further negates Winterbottom's claim that its presence is a neoteleostean feature (see above [Clupeo- morpha], and also p. 2 1 3 regarding that author's handling of the category Ostariophysi). Brousseau (19760 & b) has described parts of the branchial musculature in six ostariophysan taxa. It is clear from his descriptions and figures that, depending on the species involved, he has either misidentified the protractor pectoralis, or confused and compounded it with the levator posterior muscle. PROTACANTHOPTERYGII The distribution of the protractor pectoralis in this group is exceedingly irregular, with, in addition, the muscle occurring in relatively few taxa. As an example of its irregular occurrence, we may note its presence in some but not in other Galaxias species, and its presence in Dallia pectoralis (Fig. 4) but not in Umbra limi, U. krameri or Novumbra hubbsi. None of the protacanthopterygian species we examined has a levator posterior muscle, and we can find no reference in the literature to its occurrence in these fishes (see also Winterbottom, 1974:252-253). ACANTHOPTERYGII Lauder & Lanyon (1980: fig. 2) identified as a protractor pectoralis in Lepomis macrochirus a muscle inserting on the supracleithrum and posttemporal, and extending anteriorly to originate on the posterodorsal aspect of the skull. This muscle is apparently a derivative of the epaxialis and is not homologous with the protractor pectoralis as defined in this paper. The true protractor pectoralis in Lepomis (Fig. 5) originates dorsally from the pterotic adjacent to the origin of the levator posterior, and extends posteroventrally to fan out and insert in the connective tissue between the last gill arch and the cleithrum. No fibres of the protractor pectoralis contact the cleithrum. •A ,- § s s g f I s £ x o 11 II ou "2 w £ B So d. W O O o ^? & ^ ^^ P •M . . M ^ a o o . • c £ Q •*: — 3 . _ C GO'S co £ a •o S SO 3 ^4 OJD £ C/3 232 P. H. GREENWOOD &G. V. LAUDER the path of the protractor is directed so as to lie above the dorsal extremity of the cleithrum. The origin of the protractor in Monopterus albus lies posterior to and well-separated from that of the levators, and its orientation is such that it inserts onto the dorsal tip of the cleithrum. The dorsoposterior margin of the protractor pectoralis is closely applied to the ventral margin of the supracleithrum, but no fibres from the muscle are inserted onto the bone. The muscle in Lepomis identified by Lauder & Lanyon (1980, fig. 2) as a protractor pectoralis (see above), is of uncertain homology. It appears to be similar to the levator pectoralis of Winterbottom (1974: 270, and fig. 25) which has been found in batrachoids, lophiids, and tetraodontiforms. We have also found a similar muscle to the levator pectoralis in scorpaeni forms and nandids, and it may have a much wider distribution than is presently realized. Summary and conclusions Since our interest in the protractor pectoralis was stimulated initially by irregularities in its presumed phylogenetically based pattern of distribution within the Teleostei (see p. 213), our principal concern is with the effects additional data have on the phylogenetic conclusions reached previously (Rosen, 1973; Winterbottom, 1974). Clearly, because of its occurrence in some or even many members of the Clupeomorpha, Ostariophysi and Protacanthopterygii, the protractor pectoralis is not, as was once thought, a feature restricted to the Neoteleostei (see pp. 217-219). Indeed, its overall pattern of occurrence extends beyond the current limits of the Euteleostei (of which the Neoteleostei are a major subdivision; Rosen, 1973; Patterson & Rosen, 1977) to include the Teleostei (sensu Patterson & Rosen) as a whole. Taking the argument further: if one accepts the supposed homology of the muscle throughout the gnathostomes, the protractor pectoralis occurs in some or all members of the Neopterygii, Palaeopterygii, Dipnoi and, very probably, the Elasmobranchiomorphi as well (see pp. 214; 216-221). Amongst the major groups making up the lower gnathostomes, a protractor pectoralis is absent only in the extant Actinistia, a group represented by the single taxon Latimeria chalumnae. The protractor pectoralis, therefore, would seem to be a synapomorphy of the Gnathostomata and thus, within the various lineages of that group, its presence must be treated as a plesiomorph character and not as a synapomorphy of the Teleostei in part. Its distribution among the lower gnathostomes has a somewhat patchy pattern (see Table 1), a pattern whose patchiness is, in some respects, even more marked and more puzzling within the Teleostei. A protractor pectoralis is not developed in the two lineages which are generally thought to be the most primitive amongst living teleosts, namely the Osteoglossomorpha and the Elopomorpha. Patterson & Rosen (1977) for example, consider the Osteoglossomorpha to be the plesiomorph sister group of all other living teleosts combined (ie the Elopocephala), and the Elopomorpha to be the plesiomorph sister group of the other Elopocephala (ie the Clupeocephala). The muscle is also absent in one 'higher' group, the Stomiatiformes, which, on the basis of its having certain derived characters, Rosen (1973: 505) identified as the plesiomorph sister group of all other lineages he brought together as the Neoteleostei (see also p. 213). Within those other lineages (ie the Aulopiformes, Myctophiformes, Paracanthopterygii and Acanthopterygii) a protractor pectoralis has been found in virtually all taxa placed in the supposedly 'higher' categories Paracanthopterygii and Acanthop- terygii,1 in the majority of Aulopiformes, and in many Myctophiformes as well (See Table I; also Winterbottom, 1974). The exceptional taxon is Stephanoberyx monae; no protractor was found in the two specimens we examined. PROTRACTOR PECTORALIS MUSCLE 233 Among lineages which lie, cladistically speaking, between the plesiomorph Osteoglosso- morpha + Elopomorpha on the one hand and the apomorphic Neoteleostei (sensu Rosen) on the other, the muscle occurs in most members of the Clupeomorpha and in many Ostariophysi (including the plesiomorphic Anotophysi), but is found in only a few members of the Protacanthopterygii. An intriguing aspect of this pattern is the contrast between the absence of a protractor pectoralis (a derived condition) in the two most primitive lineages, and its presence (the primitive condition) in the more derived ones. Equally intriguing is the way the muscle, in taxa belonging to the 'intermediate' groups, may be absent in some species of a genus but not in others (for example in the protacanthopterygian genus Galaxias], or in some but not all members of seemingly closely related taxa (for example, again amongst the protacanthop- terygians, its presence in Dallia, but its absence in Umbra and Novumbrd). Similar patterns of presence or absence can be found within the Clupeomorpha and the Ostariophysi (see Table 1). We can offer no explanation for these patterns, and must conclude that the presence or absence of a protractor pectoralis muscle is a feature of little value as an indicator of phyletic relationship except at a high level of universality. In other words, it is a synapomorphy of the Gnathostomata. Within the Teleostei we are impressed by the constancy of its presence in the Paracanthopterygii and Acanthopterygii, and by the constancy of its absence in some other groups (for example, the Osteoglossomorpha). This pattern would suggest to us that in these lineages the ontogenetic canalization leading either to the development or to the suppression of a protractor pectoralis was fixed very early in the history of each lineage. In contrast, the irregular patterns seen in other lineages (for example the Ostariophysi and Protacanthop- terygii) would seem to indicate the retention of a flexible linkage between the pathway of protractor pectoralis ontogeny and other elements of the total ontogenetic pattern. What significance, if any, this may have in tracing phyletic histories remains obscure. All our conclusions (and speculations) are, of course, dependent on the hypothesis that the muscle is an homologous (ie synapomorphic) feature within the Gnathostomata. Our investigations provide no refutation of that hypothesis, but we are of the opinion that the only critical test will come from comparative embryological and ontogenetic studies on a larger scale than has been carried out so far. Acknowledgements We are most grateful for the assistance given to us by our colleagues in the Museum of Comparative Zoology and the British Museum (Natural History), especially Karel Liem, William and Sara Fink, Keith Banister and Gordon Howes (who also drew the figures). This investigation began when the senior author was a Agassiz Visiting Professor at the Museum of Comparative Zoology. He is much indebted to that Museum for providing financial assistance for his visit, and for providing such congenial working conditions when he was there. The junior author gratefully acknowledges the Society of Fellows (Harvard University) for a Junior Fellowship, and a Penrose Fund award (grant 8636) from the American Philosophical Society. References Allis, E. P. 1897. The cranial muscles and cranial and first spinal nerves in Amia calva. J. Morph. 12 : 487-808. 1917. The homologies of the muscles related to the visceral arches of the gnathostome fishes. Q. Jl. Microsc. Sci. 62 : 303^06. Brousseau, R. A. \916a. The pectoral anatomy of selected Ostariophysi. I. The Characiniformes. J. Morph. 148:89-135. \916b. The pectoral anatomy of selected Ostariophysi. II. The Cypriniformes and Siluriformes. J. Morph. 150:79-115. 234 P. H. GREENWOOD &G. V. LAUDER Dohrn, A. 1884. Studien zur Urgeschichte des Wirbeltierkorpers IV. Mitt. zool. Stn Neapel 5: 102-151. 1885. Studien zur Urgeschichte des Wirbeltierkorpers VII. Mitt. zool. Stn Neapel 6 : Edgeworth, F. H. 1911. On the morphology of the cranial muscles in some vertebrates. Q. Jl. Microsc. Sci.56: 167-316. - 1935. The cranial muscles of vertebrates. Cambridge. Greenwood, P. H., Rosen, D. E., Weitzman, S H. & Myers, G. S. 1966. Phyletic studies of teleostean fishes, with a provisional classification of living forms. Bull. Am. Mus. nat. Hist. 131 : 339-456. Lauder, G. V. & Lanyon, L. E. 1980. Functional anatomy of feeding in the bluegill sunfish, Lepomis macrochirus: in vivo measurement of bone strain. J. exp. Biol. 84 : 33-55. Millot, J. &. Anthony, J. 1 958. Anatomie de Latimeria chalumnae. 1 : 1 1 8 pp. C.N.R.S., Paris. Patterson, C. & Rosen, D. E. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the theory and practice of classifying fossils. Bull. Am. Mus. nat. Hist. 158 : 8 1-1 72. Rosen, D. E. 1973. Interrelationships of higher euteleostean fishes. Zool. J. Linn. Soc. 53, Supp. 1 -.397-513. - & Patterson, C. 1969. The structure and relationships of paracanthopterygian fishes. Bull. Am. Mus. nat. Hist. 141 : 357-474. Sewertzoff, A. N. 1928. The head skeleton and muscles of Acipenser ruthenus. Acta zool, Stockh. 9: 193-319. Vetter, B. 1878. Untersuchungen zur vergleichenden Anatomie der Kiemen-und Kiefermusculatur der Fische, II. Jena Z Naturw. 12 : 431-550. Wiley, E. O. 1979. Ventral gill arch muscles and the interrelationships of gnathostomes, with a new classification of the Vertebrata. Zool. J. Linn. Soc. 67 : 149-179. Winterbottom, R. 1974. A descriptive synonymy of the striated muscles of the Teleostei. Proc. Acad. nat. Sci. Philad. 125 : 225-317. Manuscript accepted for publication 9 January 198 1 British Museum (Natural History) 1881-1981 Centenary Publications The Natural History Museum at South Kensington By W. T. Steam This book presents a full history of the Museum, its collections, Directors and eminent members of staff. It provides for the first time an account of the major research undertaken and insights into the personalities of the key people in the Museum's development and evolution. Co-published with William Heinemann. Alfred Waterhouse and the Natural History Museum By Mark Girouard Designed by Alfred Waterhouse in the 1870's the Museum is surely one of London's most outstanding pieces of architecture. This attractively illustrated book describes the development of the design and highlights some of the building's most interesting features. British Museum Natural History By Peter Whitehead & Colin Keates The Museum is really a huge scientific research institution - acquiring, describing and classifying all manner of natural history material, both specimens and artworks. It has some of the richest collections of their kind in the world and it is these and how they are obtained and managed that are the subject of this book. The lavish, full colour illustrations and lively text will appeal to everybody interested in natural history. Co-published with Philip Wilson Ltd. Chance, Change and Challenge This multi-author twin volume work is one of the Museum's most ambitious publishing projects. In the first volume The Evolving Earth twenty scientists have been asked to summarise the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume The Evolving Biosphere Museum scientists have chosen an evol- utionary concept - speciation, coevolution, biogeography etc and related this to the group of animals or plants in which they are specialising. Co-published with Cambridge University Press. General Editor P. H. Greenwood. Animal Identification-A Reference Guide VOLUME 1 : MARINE AND BRACKISH WATER. Edited by R. W. Sims. VOLUME 2 : TERRESTRIAL AND FRESHWATER. Edited by R. W. Sims. VOLUME 3 : INSECTS. Edited by D. Hollis. These guides provide the reader with lists of primary sources of reference that can be used to identify (or lead to the identification of) living animals throughout the world. The references are arranged in systematic and geographical sequence in order to facilitate searching. Co-published with John Wiley & Sons Limited. Nature Stored Nature Studied A short guide complementing the exhibition of that name to be staged in the Museum throughout 1981. It gives a fascinating insight into the history of the Museum collections and the research carried out "behind the scenes". Titles to be published in Volume 41 Anatomy and phytogeny of the Chinese Major Carps Ctenopharyngodon Steind., 1866 and Hypophthalmichthys Blkr., 1860. By Gordon Howes Morphology and systematics of some interior-walled cheilostome Bryozoa. By P. L. Cook and P. J. Chimonides Notes on Atlantic and other Asteroidea. 1. Family Benthopectinidae. By Ailsa M. Clark Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural Histo Miscellanea Zoology series Vol 41 No 5 17 December 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Dr J. G. Sheals Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISSN 0007-1498 Zoology series Vol 41 No 5 pp 235-294 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 17 December 1981 Miscellanea Contents The ostracod genus Loxoconcha from Abu Dhabi lagoon, Persian Gulf. By R. H. Bate & A. Gurney A revision of the spider genus Cocalus (Araneae: Salticidae). By F. R. Wanless The Phthiracarus species of C. L. Koch. By B. W. Kamill The status of Lamingtona lophorhina McKean & Calaby, 1968 (Chiroptera: Vespertilionidae). By J. E. Hill & K. F. Koopman . The status of Hipposideros galeritus Cantor, 1 846 and Hipposideros cervinus (Gould, 1 854) (Chiroptera: Hipposideridae). By P. D. Jenkins & J. E. Hill Page 235 253 263 275 279 The ostracod genus Loxoconcha Sars from Abu Dhabi lagoon and the neighbouring near-shore shelf, Persian Gulf Raymond H. Bate & Ann Gurney Departments of Palaeontology and Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction The present study on the genus Loxoconcha from the region of Abu Dhabi lagoon (Lat. 24° 32' N, Long. 54° 21' E), follows a number of ecological and taxonomic papers (see Bate 1971, 1973; Bate & Sheppard, 1980; and Gurney 1979a & b) that have been based on material collected by Dr Graham Evans of Imperial College, London, during the period 1961-65. This material, collected during an investigation of the Recent sediments of the Trucial coast, has been supplemented, for the ostracod studies, with samples collected by Professor John Murray (Exeter) in 1969. Elsewhere in the Gulf, Sabeekah Razzaq (1979) has undertaken a preliminary study of the benthic microfauna of Kuwait while Paik (1976 & 1977) has examined the ostracod faunas on the Iranian side of the Gulf as well as from the Gulf of Oman. Other papers relevant to our study are those of Hartmann 1964, on ostracods from the Red Sea; Jain 1978, on the fauna of Mandvi Beach, west coast of India and Bhatia 1979, on part of the fauna from off Karwar, west coast of India. Four species of Loxoconcha (Loxoconcha) and one of the new subgenus Epakroconcha are described here; it is the purpose of this paper not only to place these on record but to discuss the relevance of the ornamental differences (ranging from coarsely reticulate to finely pitted) that are exhibited between the species. Systematic descriptions Genus LOXOCONCHA Sars, 1866 REMARKS. In Bate et al. (in press), the genus Loxoconcha was divided into two subgenera: Loxoconcha and Loxocorniculum. Here a third subgenus Epakroconcha is recognized. Of the five species described from Abu Dhabi lagoon and the Persian Gulf, four belong to the subgenus Loxoconcha, being forms having a clear cut dimorphism of subquadrate females and subrectangular males. The fifth species differs by having a subtrigonal outline for both males and females and is accordingly placed in the separate subgenus Epakroconcha. Subgenus LOXOCONCHA Sars, 1866 REMARKS. The four species described here differ ornamentally by having either a reticulate or a pitted shell surface. We have grouped the species according to this and, as will be discussed under environment, we show that this has a bearing on their distribution either within the lagoon or outside on the nearshore shelf. Bull. Br. Mus. nat. Hist. (Zool.) 41 (5): 235-25 1 Issued 1 7 December 1 98 1 235 236 R. H. BATE& A. GURNEY Reticulate species Loxoconcha (Loxoconcha) multiornata sp. nov. (FigslA-J,2&ll) Loxoconcha ornatovalvae Hartmann; Bate 1971 : 245, 246, 248, 250, pi. 1, figs. Ik & 2k, pi. 2, fig. 3k, pi. 3, figs. 2k & 3k. Loxoconcha sp. A, Jain 1978 : 126, fig. 5 A. HOLOTYPE. BM(NH) no. 1980.236, cf carapace, sample 65 14, central lagoon terrace. PARATYPES. BM(NH)nos 1980.237-243. TYPE LOCALITY. Sample 6514 — Central lagoon terrace, shallow marine, tidal; salinity 50-75%o. DIAGNOSIS. Species of Loxoconcha having very coarsely reticulate ornamentation (as illustrated) with some reticulae developed into ridges; dorsal ridge of carapace characteristic- ally producing pear-shaped pattern. DESCRIPTION. Species dimorphic, the males more elongate than the females. Ornamentally the reticulae are very coarsely developed and may even be produced as low ridges. The eye node is developed as a swelling situated on a low antero-dorsal ridge. Sieve plate normal pores are situated within the reticulae. Ventrally, the edge of the right valve, where it over- laps the left, is marked with short ridges while the left valve in the posterior part of the carapace possesses four rowlock-type projections (Fig. ID). Hinge typically gongylodont with coarsely dentate/loculate median element. Muscle scars with three oval adductor scars in a vertical row with a fourth situated almost in front of the lowermost scar. Frontal scar V-shaped. Duplicature broad with small antero-ventral and narrow postero-ventral vesti- bule. Marginal canals straight, widely spaced and few in number: 9 anteriorly. Outside the selvage a broad flange extends around the anterior and along the ventral margin. DIMENSIONS. Holotype: 1980.236 cf carapace, length 0'46 mm; height 0'26 mm; width 0-26 mm. Paratypes: 1980.237 d1 RV., length 0'47 mm; height 0*26 mm. 1980.238 9 carapace, length 0'38 mm; height 0'24 mm; width 0'23 mm. 1980.239 9 carapace, length 0-40 mm; height 0'25 mm; width 0'23 mm. 1980.240 cf carapace, length 0*50 mm; height 0-26 mm; width 0'28 mm. 1980.241 9 RV., length 0'40 mm; height 0'24 mm. 1980.242 rf RV., length 0*44 mm; height 0*25 mm. REMARKS. Loxoconcha (Loxoconcha) multiornata sp. nov. is morphologically very close to the Red Sea species, Loxoconcha ornatovalvae Hartmann 1964 and was, in fact, identified as such by one of us (Bate, 1971). There are, however, significant differences that make separation of the two species possible: firstly in dorsal view, the curved dorsal ridge of L. multiornata produces a pear-shaped pattern when both valves are together (Fig. IE, F) whereas in L. ornatovalvae (Hartmann, 1964, pi. 20, fig. 93) there is a looped ridge pattern extending from the eye node back along the dorsal margin. The eye node of the Red Sea species is separate from the antero-dorsal ridge, whereas in L. multiornata it is a swollen part of the ridge. The differences in the lateral ornamentation are best explained by illustration (see Fig. 2). Loxoconcha (L.) multiornata has been recorded by Jain 1978 as Loxoconcha sp. A from the Recent sediments of Mandvi Beach, west coast of India and is thus known to occur within the Indian Ocean area. As yet it has. not been recorded from the African coast. If, as appears possible, the Red Sea Loxoconcha ornatovalvae has developed from the Indian Ocean/ Persian Gulf L. multiornata, then we are going some way to implying that species give rise to similarly ornamented species and this could lead us to a subdivision of the subgenus Loxoconcha on basic external morphology. By grouping the five species described in this paper on similar ornamentation we are partly suggesting this but we would not claim that the groupings here necessarily relate to all the species described here. OSTRACOD GENUS LOXOCONCHA 237 Fig. 1 Loxoconcha (Loxoconcha) multiornata sp. nov.: A, X 125, holotype, cf left side, BM(NH) 1980.236; B, X 125, paratype 9 RV., specimen lost; C, X 125, paratype cf RV., BM(NH) 1980.237; D, G, enlargement of ventral margin X 580 and ventral view of carapace X 125, 9 paratype BM(NH) 1980.239; H, X 1- 12K, sieve plate, specimen lost; E, X 125, paratype 2 is short and barely discernible. Setae (tc) and (w) on this segment, and (?c), (w), (p) and s on tarsi II-IV are ribbon-like, hooked distally and covered with whorls of spicules in the middle third. The four setae on femur I (Fig. 7), d, (v) and /', are all located at about the same level on the segment. Seta d which is the shortest of the four (about two-thirds as long as /'), is thickened, bluntly serrated and somewhat curved. MATERIAL: Three specimens from a sample (no. 59015) of moss and bilberries (Vaccinium myrtillus), Schweighauser Forest, Unterkaulhausen, Regensburg, 19.vii.1959 (M. Sellnick), deposited in the collections of the Zoological Museum of the University of Hamburg. One of these specimens (A30/80) is hereby designated as the neotype. Other material was examined from rotten wood and litter beside a brooklet (sample no. 61 R45), Donaustaufer Forest, Regensburg, 27.vi.1961 (L. van der Hammen) (RNH, Leiden). P. crinitus was not widespread in the Regensburg material, being present only in small numbers in the above two samples. REMARKS: Koch recorded crinita predominantly in mosses, Regensburg. Although the original description of the species was rather incomplete, crinita appeared to be character- ized by unusually long notogastral setae — a feature which was noted by Jacot (1936) in his description of specimens which he considered to be conspecific with H. crinita. Jacot's specimens are for the most part conspecific and in good condition. However, as mentioned above, they have not been considered for neotype designation. A neotype was selected from amongst Sellnick's spirit specimens which were found to be conspecific with the mite described by Jacot. Phthiracarus ferrugineus (C. L. Koch) (Figs 9-1 3) Hoplophoraferruginea Koch, 1841 : Heft. 32, t. 10. Regensburg [type series presumed lost]. NEOTYPE (here designated), Taimering, Regensburg (RNH, Leiden, P2005-7). [Phthiracarus ferrugineus: Jacot, 1936 : 179. Misidentification, see under P. longulus.} ADULT: Rather small and weakly sclerotized. The aspis (Figs 10-12) is about 225 /zm in length with a greatest width of about 160 /zm. The rostrals (ro) are short and do not reach the anterior aspal margin. Setae (it) are twice the length of setae (Id) and extend to the rostral bases. The sensilli are long (70-80 //m), slender and distally serrated. The exobothridial setae (ex) are short. The notogaster (Fig. 9), about 350 /zm in length, has a. maximum depth of about 270 jum and bears moderately long (more or less equal to the distance c, -£/,), simple setae which are erect to recurved. Setae c, and c3 are inserted close to the posterior collar margin and seta c2 submarginally. Vestigial f{ is located one-third of the distance between setae h\ and ps}. Only the fissures ia and im appear to be present. On each anal plate there are three setae (setae adt_2 vestigial) of which an,_2 are the longest. The chelicerae are 135-140 //m long with 9 to 10 sharply pointed spines on the paraxial surface of the principal segment and 6 to 10 conical spines antiaxially. The leg chaetotaxy is of the 'complete type' with rather short and straight solenidia. On femur I (Fig. 1 3) seta d is short (about half as long as /'), thickened and, as in crinitus, this seta is curved and bluntly serrated. In the three available specimens, the setal arrangement on this segment shows considerable variation. In the neotype (Fig. 13) seta d is located on a level with seta /' and anterior to the ventral setae, while in a second specimen d is located anterior to the lateral seta, and in a third specimen, posterior to the lateral seta. MATERIAL: Three specimens from rotten material in a moist hayfield, Taimering, Regens- burg, 19.vi.1961 (L. van der Hammen) (sample no. 61 R34), deposited in the collections of the Rijksmuseum van Natuurlijke Historic, Leiden. One of these specimens (P2005-7) is hereby designated as the neotype. KOCH'S SPECIES OF PHTHIRACARUS 267 \°1 10 13 Figs 9-13 Phthiracarus ferrugineus: (9) notogaster, lateral; (10) aspis, lateral; (1 1) aspis, dorsal; (12) sensillus; (13) leg I, trochanter to tibia, dorsal aspect. 268 B. W. KAMILL REMARKS: Koch recorded ferruginea in mosses on trees. He described the species as being very small with long, fine notogastral setae and from his figure it is evident that the species possessed rather long and prominent sensilli. The smallest of the Regensburg species examined in the present study, P. ferrugineus (notogastral length 350 um) is somewhat unusual in its possession of four setae on femur I and a single seta on genu IV, two features which are associated with larger species (notogastral length 500-1000 um) of the genus. Smaller species with a notogastral length of less than 500 um are generally characterized by a 'reduced' form of leg chaetotaxy. Although rather smaller, P. ferrugineus bears considerable resemblance to P. membranifer Parry (notogastral length 31 0-500 /zm) recorded from the fermentation layer under Sitka spruce, Tintern Forest, Monmouthshire. In comparison with the latter, P. ferrugineus differs only in having a 'complete' form of leg chaetotaxy. Phthiracarus globosus (C. L. Koch) (Figs 8, 14-18) Hoplophora globosa Koch, 1841 : Heft. 32, t. 12. Regensburg [type series presumed lost]. NEOTYPE (here designated), Burgweinting, Regensburg (ZM, Hamburg, A3 1/80). Phthiracarus globosus: Willmann, 1931 : 193; Feider &Suciu, 1957 : 5; Sellnick, 1960: 131. Phthiracarus globus Parry, 1979:341. Holotype, Rydal Water, Westmorland, England (BMNH, London, 1976.2.18.3.) [examined]. Syn. nov. ADULT: Large and strongly sclerotized. The aspis (Figs 15-17) ranges in length from 240-3 10 //m with a maximum width of 1 80-250 um. All the dorsal setae are very long, simple and procumbent. Setae (if) and (la) are more or less equal in length and both pairs of setae reach the rostral bases. The sensilli are 40-55 um long, lanceolate and serrated marginally. The exobothridial setae (ex) are moderately long. The notogaster (Figs 14, 18), 500-640 um long and ^ith a maximum depth of 300-460 um, is rather globular in comparison with other species of the genus. All the setae are long (greater than the distance c,-J,), simple and almost erect. Seta c, is inserted on the posterior margin of the collar and setae c2_3 submarginally. Vestigial /J is located adjacent and just dorsal to the seta /z,, towards the mid-dorsal line. The fissures ip and ips are absent. On each anal plate there are five long setae; ant_2 and ad3 being more or less equal in length and somewhat shorter than setae adt_2. The chelicerae are 170-1 90 um long. The principal segment carries 12-26 sharply pointed spines on the paraxial surface and 22-25 conical spines antiaxially. The leg chaetotaxy is of the 'complete type' and on femur I (Fig. 8) the dorsal seta is long (equal in length to /'), slightly curved and covered in whorls of blunt serrations. MATERIAL: Two specimens from litter under 'broom (Genista? [= Sarothamnus] growing under willows, Burgweinting, Regensburg, 16.viii.1959 (M. Sellnick), deposited in the collections of the Zoological Museum of the University of Hamburg. One of these specimens (A3 1/80) from sample no. 59059 is hereby designated as the neotype. A further ten specimens were examined from rotten wood and litter beside a brooklet (sample no. 6 1 R45), Donaustaufer Forest, Regensburg, 27.vi.1961 (L. van der Hammen) (RNH, Leiden). P. globosus was always recorded in small numbers in the Regensburg samples examined. REMARKS: P. globosus is a very distinctive species being characterized by the 'globular' form of the notogaster. In comparison with German specimens, it was found that the British material examined was considerably larger (notogastral length 570-850 um) and more heavily sclerotized. It is interesting that Jacot did not regard globosus as a Phthiracarus species, but rather as the male of Hoplophora decumana ( = Oribotritia decumana). Phthiracarus laevigatus (C. L. Koch) Hoplophora laevigata Koch, 1844 : Heft. 38, t. 16. Regensburg [type series presumed lost]. KOCH'S SPECIES OF PHTHIRACARUS 269 17 16 18 Figs 14-18 Phthiracarus globosus: (14) notogaster, lateral; (15) aspis, dorsal; (16) sensillus; (17) aspis, lateral; ( 1 8) notogaster, dorsal. 270 B. W. KAMILL Phthiracarus laevigatus: Jacot, 1936: 167; van der Hammen, 1963:704. NEOTYPE (designated), Donaustaufer Forest, Regensburg (RNH, Leiden) [examined]. This species has been adequately redescribed by van der Hammen (1963). Phthiracarus lentulus (C. L. Koch) (Figs 19-24) Hoplophora lentula Koch, 1841 : Heft. 32, t. 16. Regensburg [type series presumed lost]. NEOTYPE (here designated), Taimering, Regensburg (RNH, Leiden, P2001^4). Phthiracarus lentulus: Jacot, 1936 : 175 [in part]. Topotypes, Unter-Isling/Burgweinting, Regensburg (MHN, Geneva, 3135h) [examined]. [Phthiracarus lentulus: Feider & Suciu, 1957 : 5. Misidentifi- cation.j ADULT: Medium-sized. The aspis (Figs 21-23) ranges in length from 270-3 10 //m with a maximum width of 220-240 /urn. Setae (it) and (la) are about equal in length and both pairs of setae extend half the distance il-ro. The rostral setae (ro) reach the anterior aspal margin. The sensilli are 70-90 //m in length and slender; proximally the sensillar margin is smooth while the distal part is serrated. The exobothridial setae (ex) are moderately long. The notogaster (Figs 19-20) ranges in length from 500-580 urn with a greatest depth of 400-440 urn. All the setae are short (less than c,-J,), stout and more or less erect. Seta c3 is located on the posterior collar margin and setae c,_2 submarginally. Vestigial f} is located anterior to the seta h{ while the fissures ip and ips are absent. On each anal plate there are three setae (adt_2 vestigial); setae tf«,_2 being slightly longer than adv The chelicerae are 1 90-200 /zm long with 18-26 sharply pointed spines on the paraxial surface of the principal segment and 17-35 conical spines antiaxially. The leg chaetotaxy is of the 'complete type', the solenidia being long and almost straight. On tarsus I the distal seta coupled with solenidion a)2 is long and divided into two parts by a longitudinal constriction. On femur I (Fig. 24) setae d, (v) and /' are all located at about the same level on the segment. Seta d, which is almost as long as /', is thickened, curved and bluntly serrated. MATERIAL: Four specimens from moist wood, moss and litter, Taimering, Regensburg, 19.vi.1961 (L. van der Hammen), deposited in the collections of the Rijksmuseum van Natuurlijke Historic, Leiden. One of these specimens (P2001-4) is hereby designated as the neotype. REMARKS: Koch recorded lentula predominantly in moss in woods and described the species as being of medium size with short notogastral setae. It is evident that the sensilli were probably either short or very fine, since these were not included by Koch in his original figure. The sensilli are indeed very fine in lentulus (Fig. 23), a feature which was also noted by Jacot in his redescription of this species. Phthiracarus longulus (C. L. Koch) (Figs 25-31) Hoplophora longula Koch, 1841 : Heft. 32, t. 17. Schweighauser Forest, Regensburg [type series presumed lost]. NEOTYPE (here designated), Donaustaufer Forest, Regensburg (RNH, Leiden, P2012-15). Phthiracarus ferrugineus (Koch) sensu Jacot, 1936: 179 [in part]. Topotypes, Dechbetten Forest, Regensburg (MHN, Geneva, 3 1 1 9h) [examined]. Phthiracarus tardus Forsslund, 1956:216. Holotype, Vasterbotten, Sweden (paratype, BMNH, London, 1 964.7. 1 3.78) [examined]. Syn. nov. ADULT: Small and weakly sclerotized. The aspis (Figs 27, 29-30) ranges in length from 220-255 //m with a maximum width of 1 70-190 /^m. The interlamellar setae (it) which are inserted on a level with the bothridia, are twice as long as the lamellars (la) and extend almost to the rostral bases (ro). The sensilli are lanceolate, serrated marginally, and range in length from 30-40 um. The exobothridial setae (ex) are short. The notogaster (Figs 26, 28) is some- what elongate and ranges in length from 430-490 //m with a maximum depth of KOCH'S SPECIES OF PHTHIRACARUS 271 19 20 23 21 Figs 19-23 Phthiracarus lentulus: (19) notogaster, lateral; (20) notogaster, dorsal; (21) aspis, lateral; (22) aspis, dorsal; (23) sensillus. 272 B. W. KAMILL 24 Figs 24-25 Leg I, trochanter to tibia, posterolateral aspect: (24) Phthiracarus lentulus; (25) Phthiracarus longulus. 300-320 //m. All the setae are erect and more or less equal to the distance c,-d,. Setae c,_3 form a row immediately behind the posterior collar margin while vestigial f} is located just dorsal to the seta /?, and towards the mid-dorsal line. The fissures ip and ips are absent. On each anal plate (Fig. 31) there are three rather short setae, ad^ being the shortest. The chelicerae are 1 50-200 /zm long with 4-14 sharply pointed spines on the paraxial surface of the principal segment and 6-10 conical spines antiaxially. The leg chaetotaxy is of the 'reduced type' (see Parry, 1979) with three setae on femur I (V absent) and no setae on genu IV (/' absent). Tarsus II bears 12 setae (subunguinial seta present), a feature which is generally associated with 'complete chaetotaxy' species. All the solenidia are moderately long and straight. Solenidion w2 on tarsus I is coupled with a minute distal seta. On femur I (Fig. 25) the dorsal seta is short, serrated, curved distally and located posterior to the setae /' and v'. MATERIAL: Eight specimens from rotten wood and litter beside a brooklet, Donaustaufer Forest, Regensburg, 27.vi.1961 (L. van der Hammen), deposited in the collections of the Rijksmuseum van Natuurlijke Historic, Leiden. One of these specimens (P2012-15) from sample no. 6 1 R45 is hereby designated as the neotype. REMARKS: Koch recorded longula in moss in woods, a species he described as being small with long, fine notogastral setae and rather short, round-ended sensilli. Of the species here described, P. longulus appears to be the most abundant in all the Regensburg samples. The species has been recorded only rarely in the British Isles but in Sweden the author has found it to be extremely widespread. KOCH S SPECIES OF PHTHIRACARUS 273 26 29 28 Figs 26-31 Phthiracarus longulus: (26) notogaster, lateral; (27) aspis, lateral; (28) notogaster, dorsal; (29) sensillus; (30) aspis, dorsal; (3 1 ) ano-genital region. 274 B. W. KAMILL Survey Of the eight Phthiracarus species described by C. L. Koch, six are capable of being reidentified with reasonable certainty. The two remaining species (lucida and straminea) cannot in the author's opinion be separated by any morphological characters mentioned in the original descriptions, although lucida was collected from a 'swampy' meadow, while straminea was recorded from moss in woods. However, Phthiracarus species are not generally considered to be habitat specific and it seems likely therefore, that Koch was in fact concerned with only one species, different specimens of which exhibited minor differences in colour. Acknowledgements The present study was based largely on specimens kindly sent by Dr L. van der Hammen, Rijksmuseum van Natuurlijke Historic (RNH), Leiden. Specimens from the Jacot and Sellnick Collections were examined through the courtesy of Dr B. Hauser, Museum d'Histoire Naturelle (MHN), Geneva, and Dr G. Rack, Zoologisches Institut und Zoologisches Museum (ZM), Hamburg. References Feider, Z. & Suciu, I. 1957. Contributii la cunoa§terea oribatidelor (Acari) din R.P.R. — familia Phthiracaridae Perty 1 84 1 . Studii Cere, stiint. lasi (Biol.) 8 : 23^6. Forsslund, K-H. 1956. Schwedische Oribatei (Acari). III. Ent. Tidskr. 11 : 210-218. Furnrohr, A. E. 1839. Naturhistorische Topographic von Regensburg, Bd II. Regensburg: Manx. Hammen, L. van der, 1963. The oribatid family Phthiracaridae. II. Redescription of Phthiracarus laevigatus (C. L. Koch). Acarologia 5 : 704-7 1 5. Jacot, A. P. 1936. Les Phthiracaridae de Karl Ludwig Koch. Revue suisse Zool. 42 : 161-187. Koch, C. L. 1835-1844. Deutschlands Crustaceen, Myriapoden und Arachniden. Regensburg. Parry, B. W. 1979. A revision of the British species of the genus Phthiracarus Perty, 1841 (Cryptostigmata: Euptyctima). Bull. Br. Mus. nat. Hist. (Zool.) 35 (5) : 323-363. Sellnick, M. 1960. Formenkreis: Hornmilben, Oribatei. Tierwelt Mitteleur. 3 : 45-132. Sherborn, C. D. 1923. On the dates of C. L. Koch, 'Deutschlands Crustaceen, Myriapoden und Arachniden,' 1835^4. Ann. Mag. nat. Hist. 9(11): 568. Willmann, C. 193 1 . Moosmilben oder Oribatiden (Cryptostigmata). Tierwelt Dtl. 22 : 79-200. Manuscript accepted for publication 28 April 1 98 1 The status of Lamingtona lophorhina McKean & Calaby, 1968 (Chiroptera: Vespertilionidae) J. E. Hill Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD, England K. F. Koopman Department of Mammalogy, The American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024, U.S.A. Introduction The genus Lamingtona and its sole species L. lophorhina were described by McKean & Calaby (1968) on the basis of six specimens from Mount Lamington in Northern Province, eastern Papua New Guinea, at 8° 55' S, 148° 10' E. Although showing Lamingtona to have a clear relationship to Nyctophilus, the original account is insufficiently detailed to permit a proper assessment of its status or of possible relationships of L. lophorhina to any of the several currently recognized species of this predominantly Austro-Papuan genus. The original material of L. lophorhina is housed in the South Australian Museum, Adelaide and in the collections of the Commonwealth Scientific and Industrial Research Organisation (Division of Wildlife Research), Canberra. Through the courtesy of Dr J. H. Calaby we have been able to examine the two paratypes from Canberra and to compare them directly with most of the described forms of Nyctophilus, with especial attention to those already known from New Guinea. Systematic Section McKean & Calaby (1968) compared Lamingtona primarily with Nyctophilus, although only two species (timoriensis and geojfroyi) of the latter are actually mentioned in their account. The principal diagnostic character that they use in separating Lamingtona from Nyctophilus and from its close relative Pharotis is the absence in Lamingtona of any band of integument connecting the ears across the forehead and they also remark that the ears of Lamingtona are smaller and the tragus not broadened basally as it is in these genera. In Nyctophilus it has been possible to examine specimens in alcohol of geojfroyi, timoriensis, gouldi, bifax, walkeri, microdon, arnhemensis and microtis (the last including bicolor), the species that tentatively we recognize in this genus, and in Pharotis similar exampks of its sole represent- ative, P. imogene. The ears are connected by a moderate or high band (which incidentally is not an extension of their inner or medial margins, but a separate structure arising from the outer or medial surface of each conch) in Pharotis imogene and in all of the species of Nyctophilus except arnhemensis and microtis. In arnhemensis the band is low and in microtis only slightly developed and barely discernible or absent in the midline, a point made clear in the original description of this species by Thomas (1 888) who says 'band across the forehead nearly or quite obsolete in the centre'. On occasion, therefore, the ears are not joined at all. In Pharotis imogene and in several of the species of Nyctophilus the ears are large or sometimes very large, but those of N. microtis are relatively small, as its name implies, although Thomas (1915) found them to be rather larger than he thought originally (1888). The ears of walkeri and arnhemensis are of similar size or are only a little larger than Bull. Br. Mus. nat. Hist. (Zool.) 41(5): 275-278 Issued 1 7 December 1 98 1 275 276 J. E. HILL&K. F. KOOPMAN those of microtis, which in turn resemble the ears of Lamingtona in size. The tragus is broadened basally in all except Nyctophilus microtis, although the extent of such widening may be slight. In microtis, however, this broadening of the tragus is often very poorly marked and may be virtually absent. According to McKean & Calaby (1968) Lamingtona lophorhina lacks an ear band. Examination of the two paratypes (99 CM 2090, 2091) shows this to be correct in the sense that the ears are not joined near the bases of their inner or medial margins by an obvious band of integument. However, although in both paratypes the area between and in front of the ears is densely pilose, careful investigation reveals the presence of a low integumentary ridge in the dry skins, extending from a point just behind the inner margin of each ear towards the midline. These ridges do not quite reach the median line but diminish in height towards the centre of the head. The condition exactly resembles that found in Nyctophilus microtis in which the ears are supported medially by a thick, tapered integumentary ridge whose outer and highest parts are integral with the medial faces of the ears. As in the paratypes of Lamingtona lophorhina the integumentary ridges taper to the midline, which on occasion they do not quite reach, and at best the ears are joined medianly only by a low elevation. So far as can be determined from the dry skins the tragus in one paratype (CM 209 1 ) is more or less spatulate and little widened basally but in the other example (CM 2090) the tragus is slightly widened in its lower part: in both the tragus has a bluntly rounded tip, inclined anteriorly and is closely similar to the tragus of Nyctophilus microtis. We can see no reason that supports the recognition of Lamingtona as a genus distinct from Nyctophilus on account of the characters originally proposed. Miller (1907) in defining Nyctophilus noted 'Ears . . . usually joined across forehead by a conspicuous band of membrane, but this obsolete at middle in one species . . .'. [Since Miller stated that he had examined timoriensis, microtis and walkeri, the 'one species' was presumably microtis.} Moreover, the extent of joining of the ears (and the shape of the tragus) is known to vary within other bat genera and scarcely justifies even subgeneric recognition, especially since in Nyctophilus the relatively rudimentary integumentary ridges of microtis and lophorhina are linked by the low band of arnhemensis to the higher, more obvious connecting bands of species of similar size such as bifax, walkeri and microdon. There appear to be no significant cranial features to support the subgeneric separation of microtis and lophorhina (and possibly arnhemensis) from the remaining species currently recognized in Nyctophilus. In discussing the affinities of Lamingtona McKean & Calaby (1968) remarked that it 'probably belongs in the subfamily Nyctophilinae but because of the ear characteristics it falls outside that subfamily as currently defined'. This claim deserves examination, especially in view of the close correspondence between the ear of L. lophorhina and that of Nyctophilus microtis, even to the shape and size of the small internal lobe near the base of its posterior or outer margin. Although Tate (1941) did indeed mention 'large united ears' as one of several features of the subfamily, the original definition by Miller (1907) characterized it externally by the form and structure of the muzzle and noseleaf, without reference to the ears. Miller remarked that the subfamily 'Differs from the Vespertilioninae in the abruptly truncate muzzle, on the anterior face of which the nostrils open forward beneath a distinct horseshoe-shaped ridge or small noseleaf. Incidentally, though the New World genus Antrozous (usually, though perhaps erroneously placed in the Nyctophilinae) has an ear band basically similar to that of Nyctophilus, it is interrupted in the middle, as on occasion is the rudimentary band of N. microtis, and therefore does not (contra Tate) join the two ears together. There are evidently no grounds for excluding lophorhina and microtis from the Nyctophilinae unless it is proposed radically to redefine this alleged subfamily. As pointed out by McKean & Calaby, lophorhina has the small noseleaf and grooved muzzle character- istic of Nyctophilus: in fact, so far as can be determined from the dry skins of the paratypes, it is very similar in rhinarial structure to microtis, with an apparently truncate muzzle. The two paratypes (CM 2090, 2091) are dorsally a rather uniform reddish brown, ventrally a little paler and more ochreous: the wing membranes are blackish brown, the tail membrane brownish. Both have every indication of prolonged immersion in alcohol (from STATUS OF LAMINGTONALOPHORHINA 277 1929 to 1967 according to McKean & Calaby, 1968): the dorsal pelage is blackish brown at the base but otherwise bleached reddish brown, the ventral pelage similarly coloured but with the tips of the hairs more ochreous or buffy brown. In general the pelage is dense and long both above and below, almost shaggy. It is noticeably long and thick on the head as far anteriorly as a line just anterior to the eyes: the muzzle and the anterior part of the underside of the chin are sparsely covered with shorter hairs. A similar distribution of fur on the head is found in Nyctophilus microtis. So far as can be determined, the narial foliations of lophorhina consist of a raised ridge surmounting the nostrils, the second low ridge lying behind it divided medianly. A similar structure (type 2 of Thomas, 1915) characterizes Nyctophilus microtis. The ears closely resemble those of this species, and are moderate and full, with a rounded tip, the posterior margin of the ear evidently concave above the antitragal lobe. The basal third to one half of the medial surface of the conch is sparsely pilose as in N. microtis. As already discussed, the interaural band in the paratypes exactly resembles the rudimentary band found in that species, and both lophorhina and microtis have an essentially similar tragus. The skull of lophorhina closely resembles that of Nyctophilus microtis in both size and structure. The sagittal crest in the two paratypes examined (one, CM 209 1 with slightly worn teeth), however, is less developed frontally than in adults of N. microtis. The upper surface of the rostrum also provides a point of variation: in CM 2090 the rostrum has a very shallow, almost imperceptible median depression, while CM 2091 has a relatively deep, well-defined longitudinal sulcus. Mr Peter Aitken of the South Australian Museum has examined the holotype (M6404) and two of the remaining paratypes (M6402, M6403): the final paratype (M6401) is in alcohol with the skull in situ. He reports that M6402 has a relatively shallower, more saucer-shaped depression on the top of the rostrum than M6403 and M6404, which each have a very distinct, deep longitudinal rostral depression forming a more obvious trough. There is a similar range of variation among the specimens representing Nyctophilus microtis and those referred to this species that we have examined, all from Papua New Guinea. The young adult holotype of bicolor ( = N. microtis bicolor) (BM(NH) 5.1 1.28.2) from an unknown locality near the coast on the Aroa River, c. 9° 05' S, 146° 48' E has a dorsally flattened rostrum with no more than the slightest suggestion of any median depression, while another immature (AMNH 108531) from Sogeri, 9° 30' S, 147° 33' E, the type locality of microtis itself, has such a depression only faintly indicated. The holotype of microtis (BM(NH) 88.4.14.1) has a very shallow rostral depression much like that of CM 2090. A further specimen (BM(NH) 88.4.14.5) from Sogeri has a moderate longitudinal rostral depression similar to that of CM 209 1 but shallower, while an example (BM(NH) 34.1.14.7) from Kokoda, 8° 53' S, 147° 44' E has a large, very deep median sulcus that is more greatly developed than the sulcus of this paratype of lophorhina. This last locality, incidentally, is but little distant from Mount Lamington at 8° 55' S, 148° 10' E, whence the original series of lophorhina was obtained. The collection of the American Museum of Natural History also includes a further six specimens from a variety of localities in Papua New Guinea that reflect this variation in the degree of development of the rostral depression. From the limited material available to us it does not seem to be correlated with geography, sex or with age (judged by tooth wear) once full maturity is achieved: the only specimens in which the rostral depression is virtually absent are clearly immature or young. In size lophorhina almost exactly resembles Nyctophilus microtis. Ranges of selected measurements of lophorhina (from McKean & Calaby, 1968), followed by those of adults of N. microtis in the British Museum (Natural History) and in the American Museum of Natural History (number of specimens examined in parentheses): length of forearm 39-40 (6), 37-5^0-5 (8); length of ear 15-19-5 (6), 15-19 (7); condylobasal length 13-8-14-7 (5), 13-5-14-5 (5); length of maxillary toothrow (c-m3) 5-3-6-0 (5), 5-5-6-4 (8). Although small, the sample of specimens that we have examined leaves little doubt that lophorhina and microtis are conspecific, if indeed not synonymous, and also closely related to bicolor. However, the close proximity in eastern Papua of the type localities of these three 278 J. E. HILL & K. F. KOOPMAN forms and the paucity of material from this area makes any further evaluation of their taxonomic validity difficult if not impossible. The differences between such specimens as we have seen are small and indicate that at best only subspecific separation can be justified. For the present, therefore, we are inclined to accept only a single species, Nyctophilus microtis, with the provisional recognition of three subspecies, N. m. microtis Thomas, 1888, N. m. bicolor Thomas, 1915 and N. m. lophorhina McKean & Calaby, 1968. Summary A review of the characters ascribed to the vespertilionid genus Lamingtona McKean & Calaby, 1968 from New Guinea shows that it cannot be separated generically from the predominantly Austro-Papuan genus Nyctophilus and that there is no justification for its retention as a valid subgenus within Nyctophilus. Examination of paratypes of its type species L. lophorhina McKean & Calaby, 1968 together with further information from the original series indicates that it is conspecific and possibly synonymous with Nyctophilus microtis Thomas, 1888. Three subspecies of N. microtis (N. m. microtis, N. m. bicolor Thomas, 1915, N. m. lophorhina McKean & Calaby, 1968) are tentatively recognized in eastern Papua but lack of adequate material prevents any proper assessment of their valdity. Acknowledgements Our thanks are due to Dr J. H. Calaby, of the Division of Wildlife Research, Commonwealth Scientific and Industrial Research Organisation, Canberra, Australia who made possible an examination of part of the original series of Lamingtona lophorhina, and to Mr Peter Aitken, Senior Curator at the South Australian Museum, Adelaide, Australia who examined the holotype and other relevant paratypes on our behalf. References McKean, J. L. & Calaby, J. H. 1968. A new genus and two new species of bats from New Guinea. Mammalia 32 : 372-378, 3 figs, 2 tabs. Miller, G. S. 1907. The families and genera of bats. Bull. U.S. natn. Mus. 57 : i-xviii, 1-282, 49 figs, 14 pis. Tate, G. H. H. 1941. Results of the Archbold Expeditions. No. 40. Notes on vespertilionid bats. Bull. Am. Mus. nat. Hist. 78 : 567-597, 4 figs. Thomas, O. 1 888. Description of a new bat of the genus Nyctophilus. Ann. Mag. nat. Hist. (6), 2 : 226. 1915. Notes on the genus Nyctophilus. Ann. Mag. nat. Hist. (8), 15 : 494^99. Manuscript accepted for publication 7 May 198 1 The status of Hipposideros galeritus Cantor, 1846 and Hipposideros cervinus (Gould, 1854) (Chiroptera: Hipposideridae). Paulina D. Jenkins & J. £. Hill Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction Considerable taxonomic confusion has surrounded the small leaf-nosed bats of Indo- Australia allied to Hipposideros galeritus Cantor, 1 846 from Penang Island, and numerous names have been proposed since that date for specimens or populations from this extensive region. Suggested classifications have varied from the recognition of three or more species to the definition of no more than one, the grouping of the various named forms changing accordingly. However it has long been apparent that the Bornean population of these bats seemed critical to any consideration of the groups of taxa as a whole, but for many years the relative lack of specimens from a variety of Bornean localities made critical assessment difficult. Moreover, this difficulty is compounded by much confusion in the literature. This study was initiated through the interest of the Earl of Cranbrook (formerly Lord Medway) who obtained specimens in Sarawak while with the Royal Geographical Society Expedition to Mount Mulu, 1977-1978. These although superficially referable to H. galeritus, on close examination seemed to represent two very similar but nevertheless distinct taxa. They have led to a further examination of specimens ascribed to H. galeritus or to its nominal allies in the collections of the British Museum (Natural History) and as a result it has proved possible to attempt some clarification of the status and relationships of the majority of the named forms previously associated in one way or another with this species. Representative material of a minority of described taxa or populations is either lacking or unavailable and to this extent our conclusions must remain provisional, although even so it has been possible to advance tentative opinions. At the same time we have endeavoured to indicate through detailed synonymies the wide variety of conflicting usages and differing opinions that exist in the literature of this relatively small group of bats, of which representatives are likely to be found in the majority of collections from Indo- Australia. Material and Methods One hundred and thirty-three specimens have been examined in the course of this study, all in the British Museum (Natural History) [BM(NH)]. For the most part they are identified by their registration numbers but a minority of specimens collected recently in southeastern Sulawesi has yet to be accessed and will be reported more fully in an account of the collection of which they form a part. By the courtesy of Dr C. Smeenk, we have also been able to borrow a series of critical specimens from the Rijksmuseum van Natuurlijke Historic, Leiden (RNH) which have proved very valuable. In addition, Professor J. D. Smith of California State University has generously provided measurements of a series of specimens that he obtained recently on the islands of the Bismarck Archipelago, off northeastern New Guinea. Qualitative assessments of external, cranial and dental characters are summarized in the systematic accounts. Measurements are in millimetres and were Bull. Br. Mus. nat. Hist. (Zool.) 41(5): 279-294 Issued 1 7 December 1 98 1 279 280 P. D.JENKINS &J. E. HILL made entirely by one of us (P.D.J.) with dial calipers. In the text those relevant in diagnosis appear with the following abbreviations: Length of forearm FA Length of 2nd phalanx of third and fifth digits III2, V2 Length of tail TL Length of skull from canine to occipital condyle (condylocanine length) CCL Greatest width across zygomata ZW Greatest mastoid width MW Greatest diameter of tympanic annulus TD Crown width across third molars M3-M3 Length of upper toothrow from canine to third molar inclusive C-M3 Width of third molar M3 Where series of specimens have been available the range of each measurement is given, preceeded in parentheses by the number examined and followed by the arithmetic mean, also in parentheses. Measurements of holotypes appear separately but are also included in the appropriate series. History Although Hipposideros galeritus Cantor, 1846 and H. cervinus (Gould, 1854) are in general use as the earliest identifiable names in this small complex of bats, it is possible that both may be antedated by Rhinolophus crumeniferus Lesueur & Petit, 1807 (pi. 35), from Timor Island. The status of this name has been reviewed and discussed by Tate (1941 : 367, footnote, 382, 387), Laurie & Hill (1954 : 56), Oey & Feen (1958 : 230), Hill (1963 : 23) and Goodwin (1979 : 79). It is based solely on a coloured plate and no holotype appears to exist. There is very little doubt that the bats depicted belong to the galeritus complex and although the plate is insufficiently detailed, they are probably referable either to galeritus or to cervinus. There are no subsequent records of bats referable to either of these from Timor and for the present the name is best left unallocated. The taxonomic history of this small but involved grouping of bats thus begins with the description of Hipposideros galeritus Cantor, 1846 from Penang Island, Malaya, followed in quick succession by the recognition of cervinus Gould, 1854 from northern Australia, labuanensis Tomes, 1859 from Borneo, longicauda Peters, 1861 from Java and, rather later, of brachyotus Dobson, 1874 from India. Even at this early stage, considerable differences of opinion are apparent, Peters (1871) for example separating galeritus subgenerically from labuanensis, cervinus and longicauda but Dobson (1876, 1878) synonymizing labuanensis, longicauda and brachyotus into galeritus which, like cervinus, he considered a valid species. This author noted variations in the size of the noseleaf, tail length and wing insertion on the hindlimb in H. galeritus as he understood it, considering specimens of labuanensis and brachyotus intermediate between longicauda at one extreme and galeritus at the other. However, this variation is at least partially due to the fact that his H. galeritus is a composite of//, galeritus and H. cervinus as we define them, although not unexpectedly some of the features to which Dobson drew attention retain diagnostic significance. His association of labuanensis with galeritus, rather than with cervinus, has persisted until the present day and has led to many of the difficulties encountered in classifying this group of bats, since the majority of subsequent comparisons have been made against members of the Bornean population that labuanensis represents, rather than galeritus of which few specimens are known. Five further names have been proposed since Dobson's studies: batchianus Matschie, 1901 from Batjan Island in the Molucca Islands, schneideri Thomas, 1904 from Sumatra, the posthumously published misorensis Peters, 1906 probably from Misor Island ( = Schouten Island) off northeast New Guinea, insolens Lyon, 1911 from Borneo and finally celebensis Sody, 1936 from Sulawesi. Among these the account by Lyon of insolens marks an STATUS OF H. GALERITUS & H. CER VINUS 28 1 important point since this author found insolens and 'galeritus' to occur sympatrically and thus demonstrated for the first time the existence of two distinct species in the Bornean population. Most later authors have endeavoured to recognize this in different ways but have been handicapped by the unavailability of type material, by a lack of specimens from critical areas or by reliance on the concept of labuanensis as representative of H. galeritus. On this last point for example, it seems clear from the measurements that they cite, that 'galeritus' as understood by Lyon (191 1), by Chasen (1931) when reporting specimens from Sabah or by Sody (1936) when using Bornean specimens for comparison is in fact the taxon that we recognize as H. cervinus labuanensis. Hipposideros galeritus and its allies have been listed, reviewed and discussed either entirely or in part by Chasen (1940), Tate (1941), Laurie & Hill (1954) and Hill (1963). Chasen (1940) admitted weaknesses in his arrangement of the Malaysian members of the complex but correctly maintained two separate taxa in Borneo. He recognized two species, H. longicauda with insolens as its subspecies and H. galeritus with subspecies labuanensis and schneideri. Although he had at one time examined the holotype of galeritus, Chasen had no access to it when writing and took for comparison 'galeritus' from Pahang which may well have been H. cervinus labuanensis as we understand it, leading him to associate galeritus with labuanensis and schneideri, rather than with insolens and longicauda which we have found to be its allies. The first attempt at a cohesive review of the entire complex was made by Tate (1941) who presented much useful information and many useful characters for distinguishing the taxa. Unfortunately it is difficult to extract a coherent view of their classification from this rather confusing paper. The caption to the distribution map (p. 365) for the galeritus group provides one opinion with the recognition of four species: H. galeritus with subspecies schneideri, H. longicauda, H. labuanensis and H. cervinus with subspecies insolens, celebensis and batchianus. The text however (p. 367) offers a different view, Tate observing that for the sake of consistency with the then prevailing view that galeritus extends through Sumatra, Java and Borneo to Celebes, it must be further extended to include cervinus of New Guinea and Australia. He concludes that brachyotus, labuanensis, insolens, celebensis, batchianus and cervinus would be more or less valid subspecies of galeritus, with schneideri and longicauda as derived species. Later (p. 369) he reaffirms that cervinus, batchianus, celebensis and insolens are no doubt conspecific but remarks that labuanensis and schneideri are good species. Finally (p. 391) labuanensis, celebensis and cervinus (also on p. 392) are considered subspecies of galeritus but insolens and schneideri distinct species. Moreover (p. 391) he reports both galeritus and schneideri from North Pagi Island in the Mentawei Islands. Tate (p. 367) also suggested that the holotype of H. galeritus consisted of a mismatched skin and skull, although apparently he only saw a photograph of the skull and had not examined the skin but instead relied upon Cantor's description. He considered that the skin had distinct affinities with cervinus but that the skull appeared to be that of a bat of the bicolor group. As a result he designated the skin as the type of Cantor's name. This opinion was accepted by Laurie & Hill (1954) and more particularly by Hill (1963) who concluded that the skull agreed closely with that of the cranially similar species H. ater, a member of his bicolor subgroup, otherwise readily distinguished from H. galeritus by the absence of lateral supplementary leaflets. Subsequent examination has shown clear cranial differences between H. ater and H. galeritus and with the advantage of additional specimens from Malaya and Thailand, that agree closely both externally and cranially with the holotype of galeritus, we have been able to establish that these assumptions are erroneous and that it is very unlikely that any mismatch has occurred. Laurie & Hill (1954) endeavoured to reconcile the varying opinions advanced by Tate with the specimens then available in the British Museum (Natural History) although concerned chiefly with New Guinea and Sulawesi. These authors recognized two species, H. galeritus extending eastwards perhaps to Sulawesi, H. cervinus from Borneo to the Solomon Islands, with subspecies insolens, celebensis and batchianus. They considered the eastern 282 P. D. JENKINS & J. E. HILL cervinus to be slightly less specialized than the western galeritus and its allies and, although differing in detail, their concept has elements in common with that advanced in this paper. Hill (1963) in reviewing the complex as a whole, adopted much of the treatment by Laurie & Hill and carried it further in uniting galeritus and cervinus into one species, galeritus, of which all other forms were made subspecies; insolens becoming a synonym of labuanensis, their differences being thought to reflect individual variation. This arrangement has been widely accepted but proves to be incorrect. Systematic Section The review by Hill (1963, p. 19) of the galeritus subgroup and (p. 52) of//, galeritus as this author understood it provides an account of the major external, cranial and dental features of bats here attributed to H. galeritus or H. cervinus but since it is based on a composite of both further definition is needed. Hipposideros galeritus Cantor, 1 846 DIAGNOSIS AND DESCRIPTION. Antitragus subangular, tall, nearly one third of ear length in height; antitragal projection absent (a small antitragal fold towards rear of antitragus may suggest an antitragal projection but disappears when ear is flattened); internarial septum linear, slightly swollen (Fig. 1); upper margins of narial lappets slightly lobed; intermediate leaf simple, slightly inflated medianly, expanded laterally to form a small lappet on either side, nearly as wide or wider than posterior leaf; wing elements relatively long; tail long; generally five caudal vertebrae (exceptionally four or six); tail membrane large and extensive. Skull with long braincase and short toothrow relative to condylocanine length (Fig. 2); rostrum inflated, in profile swollen above interorbital region which is very constricted and shortened anteroposteriorly, with an angular outline when viewed from above; zygomata slender with prominent, short, abruptly rising jugal process; greatest zygomatic width generally less or subequal to mastoid width; anterorbital foramen subtriangular to suboval; (a) (b) Fig. 1 Frontal view of noseleaf of: (a) Hipposideros galeritus insolens; (b) Hipposideros cervinus labuanensis. STATUS OF H. GALERITUS &H. CERVINUS 283 braincase globose, frontal and parietal regions divided by a shallow depression; interparietal moderately swollen; mastoid well developed; post-palatal spicule obsolescent or absent; anterior sphenoidal ridges equal in length to posterior ridges to enclose a diamond-shaped sphenoidal depression; width of cochleae exceeding their distance apart, each one third concealed by a large tympanic annulus whose greatest diameter (2 -70-3 '00 [2 -8 3]) is a little less than or approximately equal to the distance between the annuli across the sphenoidal (a) (b) Fig. 2 Lateral view of the skull of: (a) Hipposideros galeritus galeritus; (b) Hipposideros cervinus labuanensis. 284 P. D. JENKINS & J. E. HILL »»:£* • • $3v-&v.i .-: • >S» * . • '•V . to^:^ ?•.&? P'.Sfe >^?^ :.-^ -«^/iSfl*« ! • ;• * (a) (b) Fig. 3 Ventral view of external structure of right auditory region of skull of: (a) Hipposideros galeritus insolens; (b) Hipposideros cervinus labuanensis. depression (except in H. g. longicaudus see discussion of this subspecies), annulus acutely angled to longitudinal axis of skull (Fig. 3); upper incisors bicuspid, outer lobes sometimes small; M3 with well developed third commissure, fourth commissure small but evident; crown area of inner lower incisors less than crown area of outer pair. DISTRIBUTION. India and Sri Lanka to Borneo and Java. Hipposideros galeritus galeritus Cantor, 1 846 Hipposideros galeritus Cantor, 1846:183 (Pinang [ = Penang I.], Malaya); Lekagul & McNeely, 1977 : 167 (Satun Province, Thailand); Tate, 1941 : 367 (discussion, lectotype designated). Rhinolophus (Phyllorrhina) galeritus, Wagner, 1855 : 659 (Pinang [ = Penang I.], Malaya). Phyllorrhina galerita, Fitzinger, 1870 : 885 (Malay Peninsula; Pinang [ = Penang I.]). Phyllorhina galerita, Peters, 1871 : 316 (Pinang [ = Penang I.]); Dobson, 1876 : 69, 1878 : 141 (in part, Pinang [ = Penang I.]). Hipposideros galeritus galeritus, Chasen, 1940 : 46 (in part, Malay States); Tate, 1941 : 365 (mapped), 367 (type material discussed), 391 (N. Pagi, Mentawei Is.); Laurie & Hill, 1954 : 56 (in part, Penang, Malay States); Hill, 1963 : 53 (in part, Malay Peninsula); Medway, 1969 : 29, 1978 : 29 (in part, Penang I.). Hipposideros galeritus group, Tate, 1941 : 358, 360, 364 (discussed). [Hipposideros] galeritus, Tate, 1941:359 (listed), 369 (discussed), 384 (type material, further discussion). HOLOTYPE. Adult cf BM(NH) 79. 1 1 .2 1 .85. Slightly damaged skin, skull in good condition but left premaxilla and incisor missing, left zygoma broken, occipital region slightly damaged. Penang Island, Malaya. Originally in Indian Museum, Calcutta. STATUS OF H. GALERITUS & H. CER VINUS 285 OTHER MATERIAL. MALAYA: 9 BM(NH) 61.1718 (in alcohol, skull extracted) Bukit Kutu, Selangor, 1061 m; THAILAND c?9 BM(NH) 78.2351-2 (skins and skulls) Wang Bla Chan, Muang, Satun Province. DIAGNOSIS. As species diagnosis and description; relatively small. Measurements of holotype: FA 46'3; III2 19'2; V2 9'5; CCL 14-0; ZW 7'9; MW 8'9; TD 2-9; M3-M3 5-6; C-M3 5-3; M3 1 -2. Measurements of the series examined: FA (4) 39-(M8'l (43*95); III2 (4) 19-0-2O7 (19-80); V2 (4) 9-5-10-7 (10-15); TL (3) 36-0, 40'0, doubtfully 49-0; CCL (4) 14-0-14-5 (14-30); ZW (4) 7-9-8-3 (8-15); MW (4) 8'3-8'9 (8'68); TD (4) 2-80-2-90 (2-88); M3-M3 (4) 5'5-5'7 (5-61); C-M3 (4) 5-3-5-6 (5-46); M3 (4) 1-20-1-35 (1-26). REMARKS. Dobson (1876, 1878) evidently saw the holotype of Hipposideros galeritus, probably at the Indian Museum, Calcutta (it was presented to the British Museum (Natural History) in 1879) although (1876; 198) he does not list it from that collection. Since then it appears to have been commented upon only by Tate ( 1 94 1 ) and Hill ( 1 963). Tate (1941) considered that the skin and skull of the holotype of H. galeritus were mismatched, although he had not examined the skin and apparently only had a photograph of the skull, and stated that the skin alone represented Cantor's type. However, the skull associated with this skin has an inflated rostrum, slender zygomata and U-shaped palation, all features distinguishing it from the bicolor subgroup, to which Tate and subsequently Hill (1963) thought it to belong and it is very similar to the skulls of other specimens of//. galeritus — as understood in this paper — from India, Sri Lanka, Thailand, Malaya and Borneo. In all, the zygomatic width is less than or subequal to the mastoid width, a feature otherwise characteristic of//, bicolor and its close allies. Tate also thought that the skin had affinities with H. cervinus. However, it has been possible to examine the ears and noseleaf in some detail after thorough soaking and there is clearly a tall, angular antitragus lacking any posterior antitragal projection, a prominent internarial septum and well developed, broad intermediate leaf, features which are characteristic of our concept of galeritus but not found in cervinus. Therefore there seems little to substantiate the contention that a mismatch of skin and skull has occurred and Tate's designation (p. 367) of the skin as the lectotype of H. galeritus should be discounted. Lekagul & McNeely (1977) provided useful photographs of the noseleaf and skull, with notes on ecology and behaviour. However, those remarks extrapolated from Medway (1958) probably refer to //. cervinus (H. c. labuanensis) as understood here, rather than to H. galeritus. DISTRIBUTION. MALAYA: Selangor; Penang Island. THAILAND: Satun Province. Hipposideros galeritus brachyotus (Dobson, 1874) Phyllorhina brachyota Dobson, 1874 : 237 (Central India). Phyllorhina galerita, Dobson, 1876 : 69 (in part, Central India; Deccan). Hipposideros galeritus, Blanford, 1888:287 (Indian Peninsula; Ceylon [=Sri Lanka]); Brosset, 1962 : 618 (Western and Central India: Chikalda; Ellora; Kanheri; Bombay; Bedsar Caves; Badomi). Hipposideros brachyotus, Wroughton, 1913 : 36 (Kanara); 1915 : 85, 101 (Bihar; Ceylon [ = Sri Lanka]; Ryley, 1914 : 689 (Gujerat); Phillips, 1923fl : 274, 19236 : 155, 1924 : 25, 1935 : 94 (Ceylon [ = Sri Lanka]). [Hipposideros] brachyotus, Tate, 1941 : 358 (listed), 364 (mentioned), 369 (discussed), 381 (listed, holotype). Hipposideros galeritus brachyotus, Tate, 1941 : 367 (discussion); Hill, 1963 : 54 (Ceylon [ = Sri Lanka]; Mysore; Bombay; Bengal). HOLOTYPE. Adult rf BM(NH) 9.1.4.70. In alcohol in fair condition, wing membranes damaged, skull extracted, in good condition, right upper incisor, left upper canine and left anterior upper premolar (PM2) missing. Central India. OTHER MATERIAL. INDIA: BM(NH) 75.11.3.3 (skin and skull) Lingasoogar, N.N.W. of 286 P. D.JENKINS &J.E. HILL Bellary;