Far a car Rd de Ta te o Mh A wehbe rte ar baren RTE Au te AA о LU KA " к “ach A ONAN FE a ge u аль, 259 Ms AUX A UT ROLE Г Ù ae A A do A; Var x RUNDE мы ; A Bar NOMIC il RNS “hae ‘ea a » A permit dv erry YUN RE act $. Sete AG Hee hide RATA N un urn “uN AMAR N » aga Re a y se и N d ' NEN. eh x CRT CONNUE COCO EEE Se ERS Va Bi SER, EL i E un Г aa Aaa À ho tee fe A AA mi re” uh LE eM NIM PC AER Me ge AA be к, ARS na Ja abe ae dd: AAA NTE BE UE, ? AN seta CL CALE TREE АА En En LEE oe MAN e eee ' We Es NA x pes A “ ya LUN A a wie OS ак PRCA, N HE ae 016 AGA Ms e, Mod a asia de AA MCE In gna CRETE hehe te w an, ПА, ACM NT Er DotA Macon A AS uh wets Se eset A eta A 09 de a MAG Sela AL ps fhe Sind COMA A Ve Mae Sie in Sees 4 My nee una Aim ons вам L к 4 > » ar Se eb or ОСА doit Aa! HORII SIN cad Fer NUS # PR KIN А Я AOL La iy Les LL ANR in y re Testa Thain ie RN air und a E fa ns Rn Ar у des DA ae Fes 027 й » “ 23 del 20, le En DRE A y м 0 May ait Btn rh a De iter My ey to aN ES M AD men tod To hy Hy Meh ay De fa in a CES ROUE ¡0 Be Pa a авы: 4 mos MS he Le e ae Wk ie a0 EN ; cae ata yma rah à A Sun um PTOS ates ELLE sobe RO dla; o: ET om ates Abe ae Dey Nigel de С уме Raa Е att м. \ ewan EX rem! 4 RARE Tec ay Dan a y Dos o DLL EE set et ete à ML cs im Den (2 м А AAA ee “etna mn seh as A A dl D tas Вт a Da oh Ieper acs rd y A Bi м ee LES bn Pa, Bas do e Enz RE ` Pur ort Mn gor e pr ul hr ments tnt! weh is tar Saree de EEE atone io RE ЕО a >: Ане Е ИО ЕЯ ее ни ee eng Pa AJ RTL CP APL REL SEE CEE poe FPS PPLE EPL EE POE | A AAA PEPE AAA a A PARLE PET PINS a allie ee eee #1 pe epee deg AA ре нео Mo Re PILES MERE + Posa? РСЯ Bee pes INN AN rots el ЗЕ ent 87% = SPACE em nner lé CS DT » ied OS »” zn pag fpr La BOP ate EN POD: poi Its Shapes Nae у > ДИН PIDA РАЙ AS > ОО ини i Ak г. $ : À C2 g Ppt ( lee: lA fare es ANA МА » w > pres v peg et ve , pc y rá A ag Rx ne E enw yen) ет ; FRS ИИ УГ $ а ОА y rar ИТ Про мрт урл EN 4 AA ER CORPS ECS FPS ren + rt de tho q nl fd te helt el lee FAI RE CE HE OY RELA Gs pro Set OF OA PE ee DENTS A APN А ap Do яв бр bee Bre wrrur en ‘ры. w . < и 2 , Sere PrP eee PSI PT TR D METRE LE RE Ae fix nels oe LEE PEPE SEPT NDS PER ee ore Y Fe ww Den RES PTT EL rs ETA йа E nea sere on 3’ e A E > etal a Le + # IAEA ’ A NV зу YD A Side hiner ae COCOON HARVARD UNIVERSITY 3 Library of the Museum of Comparative Zoology м 0 as a at k ARE u LA ie k N qe VOL. 25, NO. 1 1984 MALACOLOGIA International Journal of Malacology AMERICAN MALACOLOGICAL UNION SYMPOSIUM PROCEEDINGS FUNCTIONAL MORPHOLOGY AND ONTOGENY OF MOLLUSCA AS RELATED TO HIGHER CATEGORY SYSTEMATICS Organized by Richard S. Houbrick 20-21 July 1981, Fort Lauderdale, Florida MALACOLOGIA Editors-in-Chief: GEORGE M. DAVIS ROBERT ROBERTSON Editorial and Subscription Offices: Department of Malacology The Academy of Natural Sciences of Philadelphia Nineteenth Street and the Parkway Philadelphia, Pennsylvania 19103, U.S.A. Associate Editors: JOHN B. BURCH University of Michigan, Ann Arbor ANNE GISMANN Maadi, A. R. Egypt Editorial Assistants: MARY DUNN DAVID WATT MALACOLOGIA is published by the INSTITUTE OF MALACOLOGY (2415 South Circle Drive, Ann Arbor, Michigan 48103, U.S.A.), the Sponsor Members of which (also serving as editors) are: KENNETH J. BOSS Museum of Comparative Zoology Cambridge, Massachusetts JOHN B. BURCH MELBOURNE R. CARRIKER University of Delaware, Lewes GEORGE M. DAVIS Secretary and Treasurer PETER JUNG, Participating Member Naturhistorisches Museum, Basel, Switzerland OLIVER Е. PAGET, Participating Member Naturhistorisches Museum, Wien, Austria ROBERT ROBERTSON CLYDE F. E. ROPER Smithsonian Institution Washington, D.C. W. D. RUSSELL-HUNTER, President-Elect Syracuse University, New York : NORMAN F. SOHL United States Geological Survey Washington, D.C. SHI-KUEI WU, President University of Colorado Museum, Boulder J FRANCES ALLEN, Emerita Environmental Protection Agency Washington, D.C. ELMER G. BERRY, Emeritus Germantown, Maryland Copyright © 1984 by the Institute of Malacology 1984 EDITORIAL BOARD J. A. ALLEN Marine Biological Station Millport, United Kingdom E. E. BINDER Muséum d'Histoire Naturelle Genève, Switzerland A. J. CAIN University of Liverpool United Kingdom P. CALOW University of Glasgow United Kingdom A. H. CLARKE, Jr. Mattapoisett, Mass., U.S.A. B. C. CLARKE University of Nottingham United Kingdom C. J. DUNCAN University of Liverpool United Kingdom Z. A. FILATOVA Institute of Oceanology Moscow, U.S.S.R. E. FISCHER-PIETTE Muséum National d'Histoire Naturelle Paris, France VZERETTER University of Reading United Kingdom E. GITTENBERGER Rijksmuseum van Машийке Historie Leiden, Netherlands A. N. GOLIKOV Zoological Institute Leningrad, U.S.S.R. S. J. GOULD Harvard University Cambridge, Mass., U.S.A. A. V. GROSSU Universitatea Bucuresti Romania T. HABE Tokai University Shimizu, Japan A. D. HARRISON MU University of Waterloo Ontario, Canada K. HATAI Arn 5 1094 Tohoku University Sendai, Japan B. HUBENDICK Naturhistoriska Museet Göteborg, Sweden S. HUNT University of Lancaster United Kingdom A. M. KEEN Santa Rosa California, U.S.A. R. N. KILBURN Natal Museum Pietermaritzburg, South Africa M. A. КЕАРРЕМВАСН Museo Nacional de Historia Natural Montevideo, Uruguay J. KNUDSEN Zoologisk Institut & Museum Kobenhavn, Denmark A. J. KOHN University of Washington Seattle, U.S.A. Y. KONDO Bernice P. Bishop Museum Honolulu, Hawaii, U.S.A. 4. LEVER Amsterdam, Netherlands A. LUCAS Faculté des Sciences Brest, France C. MEIER-BROOK Tropenmedizinisches Institut Tubingen, Germany (Federal Republic) H. K. MIENIS Hebrew University of Jerusalem Israel J. E. MORTON The University Auckland, New Zealand R. NATARAJAN Marine Biological Station Porto Novo, India LIBRARY | MAR AL UNIVI Fe: IT VW « №. J. OKLAND University of Oslo Norway T. OKUTANI National Science Museum Tokyo, Japan W. L. PARAENSE Instituto Oswaldo Cruz, Rio de Janeiro Brazil J. J. PARODIZ Carnegie Museum Pittsburgh, U.S.A. W. Е. PONDER Australian Museum Sydney A. W. B. POWELL Auckland Institute & Museum New Zealand R. D. PURCHON Chelsea College of Science & Technology London, United Kingdom O. RAVERA Euratom Ispra, Italy М. W. RUNHAM University College of North Wales Bangor, United Kingdom S. G. SEGERSTRÂLE Institute of Marine Research Helsinki, Finiand G. A. SOLEM Field Museum of Natural History Chicago, U.S.A. F. STARMUHLNER Zoologisches Institut der Universitat Wien, Austria Y. |. STAROBOGATOV Zoological Institute Leningrad, U.S.S.R W. STREIFF Université de Caen France J. STUARDO Universidad de Chile Valparaiso T. E. THOMPSON University of Bristol United Kingdom Е. HOFFOLEM® Societa Malacologica Italiana Milano R. D. TURNER Harvard University Cambridge, Mass., U.S.A. W. S. S. VAN BENTHEM JUTTING Domburg, Netherlands J. A. VAN EEDEN Potchefstroom University South Africa J.-J. VAN MOL Université Libre de Bruxelles Belgium N. H. VERDONK Rijksuniversiteit Utrecht, Netherlands B. R. WILSON National Museum of Victoria Melbourne, Australia C. M. YONGE Edinburgh, United Kingdom H. ZEISSLER Leipzig, Germany (Democratic Republic) A. ZILCH Natur-museum und Forschungs-Institut Senckenberg Frankfurt-am-Main, Germany (Federal Republic) MALACOLOGIA, 1984, 25(1): 1 FUNCTIONAL MORPHOLOGY AND ONTOGENY OF MOLLUSCA AS RELATED TO HIGHER CATEGORY SYSTEMATICS: INTRODUCTION Richard S. Houbrick Division of Mollusks, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U.S.A. The fields of evolutionary biology and sys- tematics are timely topics among biologists today. In many journals and books new in- terpretations of the process of evolution and reevaluation of the methodology employed in systematic analyses are discussed. There is, however, a tendency to formulate new hy- potheses and theories that are based on poor data derived from unsound, older taxonomic work. Some workers appear to eschew the frequently long and meticulous descriptive work that is necessary, often at the alpha level, to interpret morphological patterns and derive any meaningful analyses of higher taxa. Sound descriptive work uses anatomical and hard part analyses, and the study of ontogeny and adaptive functional morphol- ogy. These provide the characters forming the basic data necessary for complete an- alysis of taxa. Without these kinds of data innovative systematic approaches are im- possible whether they be phenetic, cladistic or traditional. In even the most common of molluscan groups detailed descriptive work that includes anatomy is surprisingly incomplete or lacking. Many higher taxonomic categories are poorly defined and based solely on variable shell characters. These characters are frequently deficient as discriminant tools even at the specific level; yet, new supraspecific taxa are constantly being proposed and accumulate at a rate which leads one to question their valid- ity and usefulness. Perusal of the latest Zoological Record revealed that 116 Recent molluscan higher taxa were proposed in one year. It is doubtful that many of these taxa have their salient characters well-defined and quantified or represent any major adaptive significance. Clearly, there is a real need for careful, holistic descriptive work. The goal of this symposium was to demonstrate that data derived from the study of the functional morphology and ontogeny of molluscan groups can be applied in the formulation of valid, well-defined, higher taxonomic catego- ries. These, in turn, allow malacologists to make meaningful comparisons between groups. The resulting taxonomic judgements and established polarities can then be de- veloped into testable hypotheses about the systematic relationships of mollusks. This symposium complements the 1977 AMU Symposium on the Evolution and Adap- tive Radiation of Mollusca. At that time, editor Davis remarked that the majority of papers dealt with land snails because most of the recent work on molluscan evolution has in- volved terrestrial snails. In general, an- atomical studies have been more common on land and freshwater taxa than among the marine groups. It was thus gratifying to see that malacologists are now beginning to work on the anatomy and ontogeny of some of the vast number of marine groups which are known only by their shells. In the 1981 sym- posium, fourteen papers were presented covering terrestrial, freshwater, and marine groups. Of these, six were on marine groups, four on freshwater taxa and two on land snails. Two papers given by Linsley and Yochelson, respectively, dealt with Paleozoic taxa, showing what can be done by careful analysis of hard parts. Waller's paper demon- strated what can be derived from an onto- logical study of shell characters in elucidating the function of convergent structures. The McLean and Hickman papers approached the relationships of extinct taxa to Recent ones by using data derived from functional morpho- logical analyses of living species. A paper by F. C. Thompson has been added to the Sym- posium Proceedings because it forms a unit with the papers of Davis and Ponder on ris- soacean systematics. The paper given by Warén was withdrawn from publication at his request because he felt that it was too in- complete to appear in print at this time. D y я | q € \ 4 y ИТУ ERA À вм о Vo. “a = AE he Po u =» rts A EE WI D‘ р у Y Sosy. A ö roy BURN Y к \ Pr ABI РА. Aye oe E Wii Rh i О a 4 re: u MALACOLOGIA, 1984, 25(1): 3-20 A CASE FOR DERIVATION OF THE FISSURELLIDAE FROM THE BELLEROPHONTACEA James H. McLean Los Angeles County Museum of Natural History, Los Angeles, California 90007, U.S.A. ABSTRACT Golikov 8 Starobogatov's (1975) theory that the Paleozoic bellerophontaceans gave rise to the Mesozoic-Recent Fissurellidae is considered. Just as the Haliotidae are limpet descendants of the Pleurotomariidae, the Fissurellidae are the logical limpet derivatives of the Bellerophonta- cea. Both derivations require suppression of coiling, shortening of the mantle cavity, and the addition of afferent ctenidial membranes. Arguments favoring this derivation are treated: 1) Fissurellid anatomy differs in major ways from that of other dibranchiate gastropods, particularly the nearly vestigial left kidney. This anatomy could have been shared with the coiled predecessor, the bellerophontaceans. 2) Shell structure of crossed-lamellar aragonite shared by fissurellids and bellerophontaceans supports common ancestry. 3) Contrary to previous accounts, there is no several-whorled, orthostrophi- cally coiled early phase in fissurellid ontogeny. Suppression of bellerophontacean coiling allowed the inherent asymmetry of torsion to have an immediate effect, producing the varying degrees of postprotoconch asymmetry expressed in the Fissurellidae. 4) The fissurellid postpro- toconch shows whorl overlap (as does the mature bellerophontacean), brought about by delayed development of the columellar lip. 5) The internally directed, hook-shaped process of the fissurellid muscle scar may be homologous with the “oblique transdorsal element” of the bellerophontacean muscle scar. 6) Morphology of apertural slits in fissurellids and bellero- phontaceans is similar. 7) The mantle edge in fissurellids can envelop the shell without obliterating the surface sculpture. The apertural shape of bellerophontaceans suggests that some had partially to completely internal shells; mantle folds like those of fissurellids would have allowed elaborate sculpture to be retained on internal shells. Although intermediate forms have not been found as fossils, the evidence favors this derivation. The origin of fissurellids would have been a rapid, paedomorphic event. There are no alternative hypotheses for the origin of fissurellids. If fissurellids were derived from bellero- phontaceans, the late Paleozoic members, at least, were torted gastropods. INTRODUCTION Although the limpet form is represented in many diverse families of gastropods, all are thought to have been derived from coiled predecessors. | have previously discussed the origin of limpets (McLean, 1981), noting that there are so many diverse anatomies represented in limpet families that it is appar- ent that the form itself imposes few con- straints upon the internal anatomy. The es- sential features of the anatomy of a limpet should therefore be similar to those of its coiled predecessor. The Fissurellidae, which first appeared in the Mesozoic, are dibranchiate limpets in which the adult shell and mantle cavity organs, particularly the ctenidia, are es- sentially bilaterally symmetrical; in contrast, other families with paired gills are markedly asymmetrical. Because there is some indica- tion of asymmetrical coiling in the early stages of fissurellids, such authors as Yonge (1947) and Eales (1950) have assumed that the bi- lateral symmetry of the adult is secondary and that they have therefore been derived from asymmetrically coiled predecessors. Current- ly, the Fissurellidae are placed in the sub- order Pleurotomariina (Cox & Knight, 1960; Knight et al., 1960). These authors offered no theory of fissurellid origin, nor did Batten (1975: 26), who concluded: “Because the Fis- surellidae have few shared derived features with any pleurotomariacean group, no likely ancestors are now known.” Golikov & Starobogatov (1975: 198), offered an intriguing theory, claiming that the Fissurellidae were derived from the Paleozoic Bellerophontacea: 4 McLEAN We think it reasonable to unite in one group the planispiral Bellerophontoidea, which be- came extinct in the Triassic, and the limpet- shaped Fissurelloidea, which appeared at that time, in particular if we take into account that the development of the latter group does not show any trace of a conispiral shell in their ancestors. Besides the presence in Fis- surelloidea of a cap-like shell with a horseshoe-shaped columellar muscle de- veloping from nearly equal rudiments (Crofts, 1955), its usual symmetry, and similarities in the location and size of the mantle complex together with various progressive features in the structure of the nervous system, provide good reason to consider this group as derived from Bellerophontoidea, which had a plani- spiral endogastric shell. Golikov & Starobogatov gave no further dis- cussion; their assignment of the Fissurellacea and the Bellerophontacea to an order Di- cranobranchia Gray, 1821, has been ignored by subsequent authors. The Bellerophontacea (Figs. 1, 2), remark- able among univalves for their isostrophic, planispirally coiled shells, endured for some 300 million years in the Paleozoic, becoming extinct early in the Triassic. No other group of fossil mollusks has been as subject to as many differing interpretations as have the Bellerophontacea. To both de Koninck (1843) and Meek (1866), the bellerophontacean slit suggested a relationship to the living families with slits or holes in the shell: the Pleurotomariidae, Haliotidae and Fissurellidae. For over one hundred years this has been the traditional interpretation, accepted by many paleontolo- gists and neontologists. In this century many authors, including Yonge (1947), Knight (1952), Knight et a/. (1960), Morton & Yonge (1964), Batten et a/. (1967), Rollins & Batten (1968), and Linsley (1978a), have regarded the bellerophontaceans as the most primitive gastropods, ancestral to all living slit-bearing gastropods. Another school, basing their evidence on the multiple pairs of muscle scars in some bellerophontiform genera, considers all such forms to have been nontorted and therefore not gastropods but monoplacophorans. This view originated with Wenz (1940) and has been supported by Termier & Termier (1968), Runnegar & Pojeta (1974), Pojeta & Runne- gar (1976), Runnegar & Jell (1976), Lever (1979), Salvini-Plawen (1980, 1981), and Runnegar (1981). A less radical view has been upheld by other paleontologists, who have accepted some bellerophontiform mollusks as nontorted—the “cyclomyan monoplacophor- ans,” but maintain that the Bellerophontacea proper are still to be considered torted. Au- thors who have argued this interpretation are Knight (1947, 1952), Horny (1963, 1965), Yochelson (1967. 1978, 1979), Rollins (1967, 1969), Rollins & Batten (1968), Starobogatov (1970), Peel (1972, 1974, 1976, 1980), Berg- Madsen & Peel (1978), and Linsley (1978a, 1978b). Varying concepts of the separation between the torted and nontorted members have been discussed by these authors. Moreover, Yochelson has maintained that the Bellerophontacea need not be considered the most primitive gastropods. Most recently, Harper & Rollins (1982) have argued that muscle scar patterns are not reliable for phylogenetic assessment and have interpreted the cyclomyan group, as well as the bellerophontaceans, as gastropods. The controversy surely continues! The possibility that fissurellids were derived from belierophontaceans has great intrinsic interest. It also has bearing on this heated controversy: were the bellerophontaceans torted and therefore gastropods, or, were they nontorted and not gastropods at all? If it can be shown that the fissurellids are derivatives of the bellerophontaceans, then the bellero- phontaceans were gastropods. Background References on Fissurellids and Bellerophontaceans The genera referred to here are, for the most part, diagnosed and illustrated in the archaeogastropod volume of the Treatise on Invertebrate Paleontology (Knight et al., 1960), in which the Paleozoic groups were treated by J. B. Knight, R. L. Batten, and E. L. Yochelson, those of the Mesozoic by L. R. Cox, and those of the Cenozoic by A. M. Keen and В. Robertson. Knight's (1941) Paleozoic Gastropod Genotypes provides photographic illustrations useful for Comparison with the drawings of shells in the Treatise. Because authors, dates, and type-species of genera are readily available in these works, this in- formation is not repeated here. The classification of Fissurellidae followed here is that of Thiele (1929), in which two subfamilies, the Emarginulinae and the Fis- surellinae, are distinguished on the basis of major radular differences. | will further discuss the classification of Fissurellidae in other pa- ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 5 pers now in preparation and will suggest that (1960). Other useful overviews are Lind- a number of other taxa proposed as subfamil- stróm's (1884) beautifully illustrated treatment ies be recognized at the tribe level within the of Silurian gastropods, Horny’s (1963) treat- Emarginulinae. ment of early Paleozoic bellerophontaceans, An overall classification of the Bellero- and Yochelson's (1960) review of the Permian phontacea was provided by Knight et al. bellerophontaceans from the southwestern fi SER ee: oes TAO TILL 18 7 FIG. 1. Bellerophontacean genera typical of the Early Paleozoic. A) Tremanotus alpheus Hall, Middle Silurian, 3 views, maximum dimension 77 mm; slit comprised of a row of open tremata, as п Haliotis. B) Salpingostoma boulli (Whitfield), Middle Ordovician, 2 views, maximum dimension 61 mm; slit closed at margin, as in the fissurellid genus Rimula. C) Tropidodiscus curvilineatus (Conrad), Lower Devonian, 2 views, maximum dimension 28 mm; slit extremely deep. D) Pterotheca transversa (Portlock), Middle Ordovician, 2 views, maximum dimension 44 mm; one of the “crepiduliform” genera with an internal shelf and a terminal apex, resembling the fissurellid genus Zeidora. All after Knight et al. (1960). McLEAN 6 United States. Papers by Peel (1972, 1974, Fossil Record of the Fissurellidae 1975, 1976) are particularly relevant to a functional interpretation of bellerophonta- Two bellerophontacean genera ranged into ceans. Shells of some representative bellero- the Mesozoic: Bellerophon (Knight et al. phontacean genera are illustrated in Figs. 1 1960: 182), which persisted through the Low- and 2. er Triassic, and Retispira, also in the Lower CR В В (( К РУКЕ ig AS a & 3 AA ñ Ñ FIG. 2. Bellerophontacean genera typical of the Late Paleozoic. A) Knightites multicornutus Moore, Upper Pennsylvanian, 3 views, maximum dimension 39 mm; the protrusions on either side of the selenizone are believed to represent abandoned inhalant canals. B) Ptychosphaera constricta Perner, Upper Silurian, maximum dimension 19 mm; one of the genera slightly asymmetrical at later growth stages. C) Euphemites ит (Fleming), Lower Carboniferous, maximum dimension 14 mm; a genus likely to have had an internal shell, judging from the breadth and shallow depth of the slit; inductural deposits cover most of the shell. D) Retispira bellireticulata Knight, Middle Permian, 2 views, maximum dimension 11 mm; a sculptured genus likely to have had an internal shell. E) Bellerophon vasulites Montfort, Middle Devonian, 3 views, maximum dimension 21 mm. All after Knight et al. (1960). ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 7 Triassic (Yochelson, personal communica- tion). The oldest fissurellid genus— Emarginula—is dated by Knight et al. (1960) from the Middle Triassic. Thus both groups are scarce in the Triassic. The Jurassic marks the onset of fissurellid radiation. Illustrations of the fissurellid assemblage described by Huddleston (1887— 1896) from the British “Inferior Oolite,” of the Bajocian Stage, early Middle Jurassic, are reproduced here (Fig. 3). Genera represented are: Emarginula, a genus common in the Re- cent fauna, Rimulopsis, which resembles the living Rimula except that the foramen is placed on a broad anterior rib, and Puncturel- lopsis, in which the foramen is somewhat closer to the apex than in Rimula. Other fis- surellid genera that appeared in the Jurassic were: Balinula, Emersonia, Koniakaua, Lox- otoma, Pseudofissurella, and Tauschia. Of these genera only Emarginula now survives, but none of these taxa is very different from living forms. As noted by Batten (1975: 26), and as is apparent in the Jurassic fissurellids in Fig. 3, re nr Shells with high profiles appeared first. By the Eocene, all the modern shell forms, other than those in the subfamily Fissurellinae, were represented in the Paris Basin. Hypothesis for the Origin of the Fissurellidae My hypothesis is that the Fissurellidae are limpet descendants of the Bellerophontacea, just as the Haliotidae are limpet derivatives of the Pleurotomariacea. The Fissurellidae differ from the Bellerophontacea primarily in lacking coiling and whorl overlap. Such other features of bellerophontacean organization as shell structure and the major features of internal anatomy are shared with the presumed an- cestral group. Changes in anatomy would have involved the gill structure. In fissurellids and haliotids the ctenidia have short afferent membranes; in the pleurotomariids the mantle cavity is deep, extending even deeper than the bases of the ctenidia. Afferent support to the ctenidia _is lacking in pleurotomariids; a deep mantle FIG. 3. Early radiation of the Fissurellidae. Copy of part of pl. 41 from Huddleston (1887-1896), showing the fissurellids of the British “Inferior Oolite,” of the Bajocian Stage, early Middle Jurassic. Included are species of Emarginula (upper left, X4), a genus with a marginal slit that is well represented in the Recent fauna; the extinct Puncturellopsis (lower left center, x 10), a Rimula-like genus with the foramen close to the apex; and three species of the extinct Rimulopsis (upper right and lower right, x4; lower left, x10), in which the foramen is midway on a broad anterior rib. 8 McLEAN cavity requires freedom of access to the en- tire mantle cavity that obtains when ctenidia are attached only ventrally. The gills are close to the columella and therefore all the related pleurotomariid musculature is lateral or ven- tral to the gills. In haliotids and fissurellids, the mantle cavity is necessarily short and there is no shell surface for support ventral to the gill; dorsal support for the gill, provided by the addition of an afferent membrane, is es- sential. К is also probable that bellerophontacean ctenidia lacked afferent membranes. Their coiled shell allowed for a deep mantle cavity in which freedom of access would have re- quired ventral ctenidial attachment only. Ven- trally attached bipectinate ctenidia in a bel- lerophontacean would have been close to the columellar muscles; deployment of these cte- nidia therefore would be easily controlled. Thus, the afferent membrane would be a new development in the derivation of fissurellids. EVIDENCE FOR DERIVATION OF FISSURELLIDS FROM BELLEROPHONTACEANS Unique Aspects of Fissurellid Anatomy Except for the fissurellids, rhipidoglossate families of archaeogastropods have a characteristic left kidney, a prominent papil- lary sac (Fig. 4B). This has been described in the Pleurotomariidae (Woodward, 1901; Bouvier & Fischer, 1902; Fretter, 1964, 1966), the Scissurellidae (Bourne, 1908), the Halio- tidae (Crofts, 1929, 1937), and the Trochacea (Randles, 1905; Frank, 1914; Risbec, 1939; Fretter & Graham, 1962; Graham, 1965). The papillary sac has a thickened region—the nephridial gland—in its wall on the side adja- cent to the pericardial cavity. Renopericardial connections are known in the Trochacea, the Haliotidae, and presumably in the Pleuroto- mariidae. FIG. 4. Diagrammatic representation of the inter- relationships of rectum, gonad, pericardial cavity, kidneys and mantle cavity in: A) the fissurellid Diodora, and B) the Trochacea (as well as the Pleurotomariidae, Haliotidae, and Scissurellidae). The rectum penetrates the pericardial cavity and ventricle in both. The left kidney is shown opening on the left side, nearly vestigial in Diodora, but in the Trochacea it is a large papillary sac with a thickened area of nephridial gland on the side toward the pericardial cavity. In the Trochacea a reno-pericardial duct connects the left kidney and pericardium. In both Diodora and the Trochacea the gonad (dark shaded) is connected to the same duct as the right kidney, with openings to the pericardial and mantle cavities. After Fretter & Graham (1962). The Pleurotomariidae, Haliotidae, and Trochidae have many other features in com- mon, including a nacreous shell interior, large, convoluted hypobranchial glands, and the spiral caecum, an appendage to the stom- ach. The Trochacea share so many features with the Pleurotomariidae that their derivation from that group is readily perceived (Fretter, 1964, 1966). The Haliotacea, in turn, are the limpetlike derivatives of the Pleuro- tomariacea.? In contrast, all authors (Boutan, 1885; Illing- worth, 1902; Tobler, 1902; Ziegenhorn & Thiem, 1926; Odhner, 1932; Fretter & Gra- ham, 1962) have noted that the reduced left kidney of fissurellids (Fig. 4A) is unique among archaeogastropods. As summmarized by Fretter & Graham (1962: 486): “the left ‘Although Yonge (1947) considered the presence of an afferent membrane to be primitive in aspidobranch gastropods, the lack of this membrane in the oldest group, the pleurotomariaceans, provides good reason to regard lack of an afferent membrane as primitive. “The Haliotidae were elevated to the rank of superfamily by Golikov & Starobogatov (1975), a decision accepted here. The mantle cavity in the Pleurotomariidae extends deeper than the bases of the ctenidia; in the Haliotidae the mantle cavity is shorter and the ctenidia have afferent membranes. Radular and epipodial characters are also vastly different. The anterior pedal mucous gland, so prominent in pleurotomariids, scissurellids, and trochids, is absent in the Haliotidae. Differences between the Pleurotomariidae and Haliotidae are greater than those among constituent families of either the Trochacea or Patellacea; consequently, the separation of Pleurotomariidae and Haliotidae at the superfamily level is justified and is consistent with the superfamilial distinction between the Bellerophontacea and Fissurellacea proposed here, as well as that between the hypothesized condition in the extinct Euomphalacea and the living Neomphalacea (McLean, 1981). ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 9 kidney is minute, if not functionless. .. The right is responsible for all the excretory activ- ity and also acts as a conduit for the genital products.” Illingworth (1902: 458) described the left kidney of Megathura crenulata (Sowerby), as “a brownish spot in the anterior wall of the pericardial cavity.” Further reductions of other archaeogastro- pod structures in fissurellids include hypo- branchial glands that are reduced to gland cells within the mantle skirt and on the cteni- dial axes, and the spiral caecum of the sto- mach reduced to a vestige. Owen (1958) found other modifications in the stomach that suggested secondary simplicity, compared to the condition in the Trochacea. Other unique anatomical states in the Fis- surellidae include: the capacity of the middle fold of the mantle to envelop the shell (dis- cussed below), a diverticulum extending ven- trally between the two halves of the buccal mass (Fretter & Graham, 1962: fig. 96), and “the separation of the intestinal groove from the main part of the intestine” (Fretter & Gra- ham, 1962: 615). Fissurellid anatomy so differs from that of the three other living dibranchiate families— the Pleurotomariidae, Haliotidae, and Scissurellidae—that it suggests no close relationship to these families. The Fissurelli- dae can be interpreted as sharing the basic anatomy of a coiled predecessor in which the anatomy also differed from that of any living family. The extinct pleurotomariaceans are sufficiently similar in shell morphology to those now living to suggest that their anato- mies were also like that of living forms. Be- cause shell form in the Bellerophontacea diff- ers from that of the Pleurotomariacea, the Bellerophontacea are prime candidates as ancestors of the fissurellids. If the condition of the left kidney of the fissurellids was shared with the bellero- phontaceans, the latter would have been out- side the main line of gastropod evolution. The vestigial condition of the fissurellid left kidney means that most of the excretory function is provided by the right kidney. Modification of the right kidney system for reproductive specialization would be less likely in a lineage in which it must also assume the entire ex- cretory burden. In prosobranchs, the change from the archaeogastropod to mesogastro- pod grade includes the modification of the right kidney to form a more advanced repro- ductive system with glandular gonoducts (Fretter & Graham, 1962; Fretter, Graham & McLean, 1981). However, if the bellero- phontacean kidney had the same limitation as has the fissurellid kidney, experimentation with the limpet form was still possible. Shell Structure in Fissurellids and Bellerophontaceans Shell structure in the fissurellids is unlike that of the pleurotomariaceans or haliotids in lacking a nacreous inner layer. According to Boggild (1930: 300), fissurellid shell structure may be characterized by having “a crossed lamellar layer with concentrical lamellae and in most instances that layer forms the whole shell.” Boggild's analysis was confirmed by MacClintock (1963). The crossed-lamellar shell structure of Fissurella latimarginata Sowerby is illustrated here (Fig. 5A). The phylogenetic significance of shell structure is difficult to evaluate because the shells of most Paleozoic fossils are altered or replaced, destroying all traces of structure. Nothing is known about shell structure in either the cyclomyan group or the oldest bel- lerophontiform mollusks. Nevertheless, some knowledge of shell structure in bellerophonta- ceans has been contributed by MacClintock (1967), Rollins (1967), and Batten (1982). Batten has recently found considerable di- versity in the group, though MacClintock and Rollins found only crossed lamellar aragonite or complex crossed lamellar aragonite in late Paleozoic bellerophontaceans. MacClintock (1967) concluded that “the bellerophontins are more closely related to the fissurelloids than they are to the pleurotomarioids.” MacClintock’s conclusion about a possible relationship between fissurellids and bel- lerophonts has not been mentioned by sub- sequent authors, including Golikov & Staro- bogatov (1975). Derivation of the fissurellids from a pleuro- tomariacean ancestor would require a basic change in shell structure. This would not be required for derivation of fissurellids from bel- lerophontaceans. Therefore, the par- simonious route is derivation of fissurellids from a late Paleozoic bellerophontacean an- cestor. Coiling and Symmetry in Fissurellids and Bellerophontaceans Most authors have claimed that ontogeny in fissurellids includes a trochospirally coiled 10 McLEAN Mer ME. e Y ud ЗЕ 4 à TH «à AA té FIG. 5. Crossed-lamellar aragonite shell structure in fissurellids and bellerophontaceans. A) Fissurella latimarginata Sowerby, fragment of a shell from Iquique, Chile, SEM micrograph showing the smooth internal surface and a fractured edge, the lamellae clearly visible; x 70, courtesy H. Lowenstam. В) The bellerophontacean Euphemites vittatus (McChesney), a fragment showing the “first-order lamellae trending nearly straight across selenizone midway through inner shell layer”; x30, after MacClintock (1967). phase. Yonge (1947: 458) stated: “The shell is secondarily symmetrical, although spiral twisting is apparent during development.” In support of this statement, Yonge copied a figure from Boutan (1885), identified by that author as a fissurellid developmental stage. However, | identify the figure as that of Scis- surella costata Orbigny (see Fig. 6 and cap- tion for further details). It seems that this misidentified figure has unduly influenced the current understanding of fissurellid develop- ment. It was reproduced by Yonge (1947, fig. 8a); Crofts (1955: 747) referred to it as proof that larval fissurellids have a “well developed helicoid spire.” Batten (1975, fig. 25), follow- ing these authors, copied it, and cited it as evidence that fissurellid ontogeny may in- clude a coiled stage of several whorls, resem- bling the coiling of a mature scissurellid.* Bandel (1982) also dismisses the so-called “Scissurella-stage” in fissurellid ontogeny, but continues to maintain that there is a “trocho- spirally coiled secondary shell.” There is clearly an indication of postprotoconch asym- metry in many fissurellids, as can be seen in many of Bandel's illustrations as well as such previously published illustrations as those of Boutan (1885: pl. 43, 44), Batten (1975: figs. 26-28), Fretter & Graham (1976: figs. 10, 12), . and Boggs (1978). The asymmetrically coiled part of the postprotoconch whorl may amount to one-half whorl, but, in my opinion, this is not to be equated with the regular orthostrophic coiling of Scissurella or other trochospirally coiled groups. | have examined the early whorls of a num- ber of fissurellid species and find that postpro- toconch form varies from nearly symmetrical to amore frequent condition in which the apex is offset toward the right (in dorsal view). Nearly symmetrical shells at all growth stages are characteristic of some species of Emargi- nula (Figs. 7A, B). In Diodora there is a slight to moderate projection toward the right (Figs. 7C, D). Many species of Fissurella are nearly symmetrical (Figs. 7E, F). Bellerophontaceans are unique among gastropods for their nearly perfect symmetry “Batten (1975) theorized that Scissurellidae were neotenously derived from the Fissurellidae, basing his conclusion in part upon the supposed “Scissurella-stage” of fissurellids. Bandel (1982) refuted a close relationship between the Scissurellidae and Fissurellidae on other grounds. The fact that the left kidney of the Scissurellidae is a papillary sac (Fretter & Graham, 1962, 1976) provides reason enough to dissociate the two families. ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? hi FIG. 6. Copy of part of Boutan’s (1885) pl. 42, showing two different juvenile shells purported to represent the “Développement de la Fissurelle.” The fissurellid species studied by Boutan is now known as Diodora apertura (Montagu). Boutan's figs. 1 and 3 show the early sculpture of D. apertura, but | identify the shell in figs. 4 and 5 as Scissurella costata Orbigny, a species that occurs at Banyuls-sur-Mer, the Mediterranean locality at which Boutan worked. Vayssiére (1894) studied Scissurella costata at the same locality. Although Boutan gave no magnification, the axial ribs on his figure may be matched with those in a correctly identified figure of S. costata in Batten (1975: fig. 12). In both the Boutan figure (copied by Batten, fig. 25) and Batten s figure 12, the axial sculpture is present for exactly 1% whorl preceding the first appearance of the slit. Boutan made no mention of the discrepancy among the shells figured on pl. 42, either in his detailed caption or in the text. Perhaps his artist made the error, which he failed to notice. Although | identify the shell in Boutan’s figs. 4 and 5 as that of a Scissurella, the animal depicted in these figures is not Scissurella, but is in agreement with his other illustrations of early stages in Diodora apertura. at all growth stages. However, some genera, such as Ptychosphaera (Fig. 2B), are asym- metrical in the final whorl. Few authors have discussed bellero- phontacean protoconchs. Dzik (1978: 295, fig. 4D) illustrated—without giving magnification—a supposed Ordovician bel- lerophontacean protoconch (and first postpro- toconch whorl) identified as “Modestospira? sp.” Peel (1974) described ontogeny in the Silurian bellerophontacean genera Plectono- tus and Tritonophon, based upon “transverse 12 McLEAN FIG. 7. SEM micrographs showing protoconchs, selenizones, or foramina of fissurellid species (SEM stubs in LACM collection). A & B: Emarginula superba Hedley & Petterd, 183 m off Cape Pillar, Tasmania, shell length 3.2 mm. А) Apical view, showing minimal asymmetry and origin of the selenizone less than one protoconch diameter from the protoconch lip, B) lateral view of protoconch from left side. C & D: Diodora inaequalis (Sowerby), Guaymas, Sonora, Mexico, shell length 1.9 mm. C) Dorsal view, showing the first appearance of the selenizone at two protoconch diameters away from protoconch lip, D) oblique anterior view of entire shell. E & F: Fissurella rubropicta Pilsbry, Cabo San Lucas, Baja California, Mexico, shell length 0.5 mm. E) Dorsal view, showing bilateral symmetry of protoconch and early postprotoconch shell, F) oblique lateral view of entire shell, showing the early foramen two protoconch diameters from the protoconch lip; selenizone lacking at all stages. ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 13 cross-sections of specimens, cut so as to contain the axis of coiling and to be per- pendicular to the plane of symmetry.” Cross- sections of early whorls were “subcircular,” with no indication of asymmetry. More needs to be said about bellerophontacean pro- toconchs; meanwhile, there is no indication that they are atypical of gastropods. Even if bellerophontacean protoconchs are completely symmetrical, this need not pre- clude them from being torted gastropods, for there are coiling variations in gastropods that differ from the usual condition for dextral forms. In some cases the anatomy can be dextral, but the shell is sinistral—this is known as hyperstrophic coiling (Cox, 1960: 110; McLean, 1981: 315). Hyperstrophy in living prosobranchs is known in the ampullariid genus Lanistes (see Cox, 1960: fig. 67) and in the planktotrophic veliger stages of the Architectonicidae (Robertson, 1964). If tor- sion in anatomically dextral forms can pro- duce sinistral shells, then the intermediate planispiral condition should also be possible following torsion. The postprotoconch asymmetry of fis- surellids does not agree with the nearly per- fect symmetry of most bellerophontaceans. My explanation of bellerophont symmetry is that—for reasons unknown—the inherent asymmetry of torsion was masked. However, in the limpet derivative, in which extensive coiling in the adult does not take place, the masking effect is removed in conjunction with the suppression of coiling. The adult sym- metry of fissurellids is shared with the an- cestral group. Whorl Overlap in Fissurellid Ontogeny and in Mature Bellerophontaceans Boutan (1885) emphasized that evolution in the Fissurellidae is revealed by the shell ontogeny of advanced genera, in which the foramen first appears as a marginal slit—the “Emarginula-stage’—followed by the “Rimula-stage,” in which the slit closes at the margin, placing the foramen midway on the anterior slope, and progressing to the “Diodora-stage,” in which the foramen obliter- ates the shell apex. | carry this analysis one step further and note that the early stage of the fissurellid resembles that of a mature bel- lerophontacean: the early postprotoconch shell bulges along the columellar side of the aperture in the first postprotoconch whorl of Diodora (Fig. 8). In a mature bellerophonta- 2 FIG. 8. Whorl overlap in fissurellid ontogeny, result- ing from delayed growth along the columella in the fissurellid postprotoconch. Two views of Diodora apertura (Montagu), “rimuliform larva” in ventral view. After Boutan (1885), no original magnification given. cean the previous whorl obstructs the aperture—the “whorl overlap,” discussed by Linsley (1978b: 199)—which characterizes bellerophontaceans at all growth stages. The whorl overlap in fissurellid ontogeny results from retardation of growth along the columellar lip, as is apparent in illustrations of Boutan (1885) and Boggs (1978). In his de- scription of development in the Hawaiian species Diodora granifera (Pease), Boggs (1978) reported that only at the age of three weeks does the shell begin to grow along the columellar margin, the same stage in de- velopment at which the exhalant notch is en- closed to form a foramen. This is a different course of shell development from that taken by trochids, in which growth along the col- umellar lip begins immediately, as may be seen in views of larval trochid shells illus- trated by Bandel (1975: pl. 1). In my interpretation of fissurellid ontogeny, all but the final whorl of the bellerophontacean is reduced to the first postprotoconch whorl of the fissurellid. | therefore see a bellerophont stage in the ontogeny of the Fissurellidae. Muscles and Muscle Scars in Fissurellids and Bellerophontaceans Fissurellid muscle scars (Figs. 9B, C, D) differ from the generalized horseshoe-shaped muscle scars of most limpets in that both 14 McLEAN С О FIG. 9. Muscle scars in а bellerophontacean (A, after Peel, 1980) and fissurellids (B—D, after Mac- Clintock, 1963; x marking the position of the apex). A) Diagrammatic representation of muscle scar in Bellerophon, lateral view. B) Muscle scar of Emar- ginula candida A. Adams, showing the inwardly directed hook-shaped process; shell length 10 mm. C) Muscle scar of Tugali parmophoidea (Quoy & Gaimard), showing the hook-shaped process; shell length 15 mm. D) Muscle scar of Fissurella volcano Reeve, a member of the Fissurellinae, the most advanced fissurellid group, in which the hook- shaped process is lost; shell length 24 mm. lobes have a posteriorly and inwardly directed hook-shaped appendage (except in the sub- family Fissurellinae), as diagrammed by Mac- Clintock (1963: figs. 21-29). In drawings of fissurellid anatomy (Odhner, 1932; Yonge, 1947: fig. 11; Fretter & Graham, 1962: figs. 254, 257), it is clear that these posteriorly directed processes of the muscle are in close proximity to the ctenidial axes and serve to define the posterior-lateral extent of the man- tle cavity. The hooked processes of the mus- cle may therefore aid in the positioning of the ctenidia. Paired columellar muscle scars in bellero- phontaceans (Figs. 9A, 10) have been documented by Knight (1947) and more re- cently in a number of additional bellero- phontacean genera by Rollins (1967) and Peel (1972, 1974, 1976). Peel (1972: 415) described faintly indicated “oblique trans- dorsal swellings.” Peel's illustration is copied here (Fig. 10). Their significance was not understood: “While the location of the various FIG. 10. Lateral view of the “right muscle scar of Bellerophon specimen В” of Peel (1972, text fig. 1) aperture opening toward the right, x6. $1 and $2 represent the markings of the “oblique transdorsal swellings,” which migrate anteriorly with growth, and are here considered as possibly homologous with the internally directed “hook-shaped process” of the fissurellid muscle scar. See Peel (1972) fora more detailed description of this muscle scar con- figuration. swellings on the moulds demonstrates some connection with the musculature, the nature and purpose of the structure is quite un- known.” Although Peel did not know the significance of the “oblique trans-dorsal swellings” he observed, | suggest that their homology with the posteriorly directed process of the fis- surellid muscle be considered. If these trans- dorsal markings serve to define the posterior extent of a bellerophontacean mantle cavity, they would necessarily be transitory and weakly impressed in an actively growing bel- lerophontacean because with growth, the muscle scar would be migrating continually toward the aperture. In a derivation of fissurellids from bellero- phontaceans, the loss of the columella would force an anterior migration of the paired re- tractor muscles and their posterior union to form the horseshoe-shaped fissurellid mus- cle. The “oblique-transdorsal element” of the bellerophontacean muscle would be retained as the inwardly directed hook-shaped proc- ess of the fissurellid muscle, which defines the posterior extent of the mantle cavity. Apertural Slits in Fissurellids and Bellerophontaceans The Fissurellidae show a gamut of possible expressions of the apertural slit: a weakly indented sinus in Scutus, a slightly raised ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 1 sinus in Hemitoma, a Clearly delineated sele- nizone in Emarginula (Figs. 7A, 7B) and Zeidora, a foramen on the anterior slope in Rimula, and an apical perforation in Diodora and Fissurella. The selenizone is present only in the postprotoconch stage of Diodora (Figs. 7C, D), whereas in Fissurella (Figs. 7E, F), the foramen appears without leaving a seleni- zone at any stage.* Bellerophontaceans also display a wide range of possible expressions of the slit. An extremely deep slit was represented in Tropi- dodiscus (Fig. 1C), which ranged from the Lower Ordovician through Devonian. The Ordovician-Silurian Salpingostoma (Fig. 1B) had a foramen on the anterior slope much like the fissurellid Rimula, and the Silurian Trema- notus (Fig. 1A) had a row of open tremata, a condition parallel to that in the Haliotidae. The Ordovician Pterotheca (Fig. 1D) had a raised U-shaped sinus and the shell interior had a shelf, a feature convergent with the interior shelf of the Recent fissurellid Zeidora. The presence of a slit and a slitband or selenizone in a living gastropod is an indica- tion that the animal is a gastropod having paired ctenidia, with the single exception of the mesogastropod family Siliquariidae (see Gould, 1966). Because slits and slitbands are characteris- tic of virtually all dibranchiate gastropods, the argument that the fissurellids have a slit and a slitband suggests that they could have been derived either from pleurotomariaceans or bellerophontaceans. Envelopment of Shell Margin in Fissurellids and Bellerophontaceans Fissurellids have a capacity unique in the Gastropoda—the mantle can envelop the shell without causing a corresponding obliteration of exterior sculpture. In other families with enveloped shells, the mantle produces a glossy external surface, as in the Cypraeidae and Olividae. As detailed by Stasek & McWilliams (1973), the fissurellid mantle margin has three folds. The outer fold secretes the growing edge of the shell, the inner fold extends down to envelop the foot, and the middle fold is capable of partially enveloping the shell, as in Emarginella (Fig. on FIG. 11. Shell envelopment in the fissurellid Emar- ginella clypeus (A. Adams), dorsal view, anterior at top. The shell is half-way enveloped by the middle fold of the mantle and the mantle margin is split anteriorly corresponding to the depth of the slit. Overall length about 2 cm. After Schepman (1908). FIG. 12. Shell envelopment in the fissurellid Megathura crenulata (Sowerby), the shell com- pletely enveloped by the middle fold of the mantle, except for a narrow exposed portion at the foramen; the inner fold of the mantle lifted to show the paired ctenidia, the head with cephalic tentacles, and the foot. Overall length about 12 cm. After Stasek 8 McWilliams (1973). 11), or completely enveloping it, as in Megathura (Fig. 12). Such fissurellids as Di- odora usually envelop only the shell edge with the middle fold, but can fully envelop the shell in defense against sea stars (Margolin, 1964). “Fissurella lacks a selenizone at all stages (see Bandel, 1982, pl. 12, fig. 8, pl. 12, 10; Batten, 1975, fig. 26; Fig. 8F here), providing a taxonomic character significant at the subfamily level in separating the Fissurellinae from other fissurellids. Bandel's figure (1982: pl. 10, fig. 6), identified as an early shell of Fissurella nimbosa, is evidently a misidentification of a species of Diodora, for it shows a selenizone preceding the foramen. 16 McLEAN Some fissurellid genera have a shell that is greatly reduced compared to the size of the body. Scutus has the shell edge permanently enveloped by the mantle, which can expand to fully cover the shell. Laevinesta atlantica (Pérez Farfante) has a completely internal shell; the body is three times longer than the shell (Pilsbry & McGinty, 1952, and personal observation).° The shell of Laevinesta is cen- trally positioned and the gills extend well for- ward of the shell. In Scutus, Emarginella, and Laevinesta the mantle margin is split an- teriorly to a depth corresponding to the posi- tion of the apertural slit, where it forms an excurrent siphon. The selenizone is much broader in both Emarginella and Laevinesta than in Emarginula, a genus in which the shell and body size are equivalent. In the latter genus the entire apertural margin is contained in the same plane. For the genera with an apical foramen, shell envelopment occurs in Lucapina and Megathura and is especially pronounced in Pupillaea and Fissurellidea. In all these genera the body is much larger than the shell, the foramen is relatively large, the mantle cavity and ctenidia project in front of the shell, and the gills are covered only by the greatly thickened mantle fold. The following features of fissurellid an- atomy can usually be deduced from shells: saddle-shaped shells with raised ends corre- late with greatly reduced or internal shells; broad slit bands or large foramina also corre- late with internal shells; shells with the entire apertural margin in the same plane correlate with minimal shell envelopment by the man- tle. Some bellerophontaceans have recently been considered to have had internal shells, particularly those with radial (rather than oblique or tangential) apertures (Linsley, 1977, 1978b). The inductural (callus) deposits of Euphemites were also stressed by Linsley (1978b) in his argument that Euphemites had an internal shell. He rendered it with an an- imal about four times larger than the fully enveloped shell (Fig. 13). An explanation for the exceptionally broad slit bands of some bellerophontaceans (Figs. FIG. 13. Linsley's (19786) reconstruction of the bellerophontacean Euphemites with an internal shell. Overall length about 6 cm. 2C, D) is that bodies were much larger than the internal shells, as is those fissurellids with broad slit bands or large foramina. As in fis- surellids with apertures in the same plane, bellerophontaceans with tangential apertures are likely to have had the least amount of shell envelopment by the mantle, whereas those in which the aperture does not make a plane are almost certainly those in which the shell margin was fully enveloped, if not com- pletely internal. There is little sculptural difference between fissurellids and many bellerophontaceans; both may have similar cancellate sculpture. An internal shell in a bellerophontacean with intricate reticulate sculpture like that of Reti- spira can be attributed to a mantle edge like that of fissurellids. | reconstruct Retispira, with an internal shell (Figure 14). The radial aper- ture of Retispira also suggests that it had an internal shell. Unlike Linsley's drawing of Eu- phemites, my reconstruction of Retispira shows mantle folds like those of fissurellias. Shell envelopment is also possible in the Haliotidae and Trochidae, groups of pleuro- tomariacean affinity. Haliotis asinina Lin- naeus has an epipodial fold that expands over a smooth shell surface. The trochid Sto- matella (subfamily Stomatellinae) has a smooth, enveloped shell. Because shell envelopment in genera of pleurotomariacean- trochacean affinity involves a loss of surface sculpture, a derivation of Fissurellidae from that source is unlikely. CONCLUSIONS The origin of all limpet groups can be re- garded as paedomorphic—retention of an- “The radula of Laevinesta atlantica, unlike the anatomical characters, is so unusual that Hickman (1983) doubted its fissurellid affinity. She compared the radula to that of Clypidina. However, shell and mantle characters of Laevinesta are more like those of Emarginella (Fig. 11), in that both have darkly pigmented bodies and rather flat, enveloped shells with broad selenizones. Subsequent to her publication, | have examined a light microscope preparation of the radula of an unidentified species of Emarginella (USNM 235814) and have noted that the rachidian and lateral teeth have obtusely pointed overhanging tips on which there are fine denticles, a condition that could be further modified to produce the strongly cusped condition of Laevinesta. Although radular differences are still extreme, | maintain that a derivation of the Laevinesta radula from that of Emarginella is possible. ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 17 FIG. 14. Reconstruction of the bellerophontacean Retispira with an internal shell covered by mantle folds like those of fissurellids. The ctenidia project forward of the shell aperture under cover of the inner fold of the mantle, which would be capable of completely enveloping the head and foot. A cleft in the mantle extends the depth of the apertural slit, marking the position of the exhalant siphon. Epi- podial tentacles like those of fissurellids are shown in a row along the side of the foot and neck. The upper surfaces of the foot and mantle are rendered with a textured surface like that of many living fissurellids. Overal length about 2 cm. Drawing by Mary Butler. cestral juvenile characters by later ontogene- tic stages of descendants (see Gould, 1977)—in the sense that limpets are sexually mature postlarval gastropods because they remain uncoiled. If the Fissurellidae are the evolutionary result of the suppression of bel- lerophontacean coiling, it will be nearly im- possible to find transitional forms in the fossil record. A paedomorphic transition acts upon the developmental process and has to be a rapid event. The potential would have existed in any bellerophontacean stock. In the absence of transitional forms, evi- dence for a derivation of fissurellids from bel- lerophontaceans is indirect: 1) Fissurellid an- atomy, which differs extensively from that of the pleurotomariids and haliotids, could be a reflection of the condition in the bellero- phontaceans. 2) The similarity of fissurellid shell structure to that of the late Paleozoic bellerophontaceans suggests common an- cestry. 3) The postprotoconch asymmetry of fissurellids can be explained as an unmasking of the inherent asymmetry of torsion, follow- ing the loss of coiling. 4) Delayed onset of growth on the postprotoconch columellar lip of the fissurellid causes whorl overlap at this stage, which resembles the whorl overlap of mature bellerophontaceans, suggesting that fissurellid phylogeny is revealed in its ontogeny. 5) The hook-shaped process of the fissurellid muscle scar may be homologous to the “oblique transdorsal element” of the bel- lerophontacean muscle scar. 6) The slit in fissurellids and bellerophontaceans is homologous, although it is also homologous with those of the Pleurotomariacea. 7) The unique mantle edge of fissurellids enables the retention of surface sculpture on enveloped shells; a similar capacity in bellerophonta- ceans would be compatible with their in- tricately sculptured shells, which in many were probably internal. Fissurellids are unique in having pores that extend through the shell, particularly in the early stages of the adult shell, as detailed by Bandel (1982). Bandel attaches little phylogenetic significance to these pores; however, if they were to be detected in bel- lerophontaceans, this would present a power- ful argument for the derivation of fissurellids from the bellerophontaceans. Despite the indirect nature of the evidence, | conclude that it favors a derivation of the Fissurellacea from the Bellerophontacea. | therefore transfer the Fissurellacea from the suborder Pleurotomariina to the suborder Bel- lerophontina Cox & Knight, 1960. It therefore follows, in agreement with Yochelson (1967, 1978, 1979), that the suborder Bellero- phontina is not regarded as comprising primitive gastropods, but is instead an offshoot of the Pleurotomariina in which the internal anatomy differs (as in living fis- surellids) from the primitive pleurotomaria- cean condition. It is beyond the scope of the present paper to consider the arguments of the “be- llerophont controversy,” as that has been thoroughly treated by authors mentioned ear- lier. | find the arguments of paleontologists who regard bellerophontaceans as gastro- pods the most convincing, noting also that no author supporting the non-gastropod in- terpretation of bellerophontaceans has pro- vided a reconstruction of the mantle cavity to show how those genera with deep slits would function with the slit in the posterior position. If the Fissurellidae were derived from late Paleozoic Bellerophontacea, the latter, at least, must have been torted. It also follows that further clues to the form and function of bellerophontaceans can be provided by the living fissurellids. 18 McLEAN ACKNOWLEDGMENTS | particularly thank reviewer Ellis L. Yochel- son of the U.S. Geological Survey, Washing- ton, D.C., who directed me to much of the literature on Paleozoic gastropods and criti- cized several drafts of the manuscript. Vera Fretter of the University of Reading, England, reviewed the discussion pertaining to an- atomy. | am grateful also to Eugene Coan, Carole Hickman, Myra Keen, David R. Lind- berg, and John S. Peel for reading drafts of the manuscript and offering helpful sugges- tions. | thank museum illustrator Mary Butler for drawing Fig. 14, Ken Curry for operating the SEM used for Fig. 7, Heinz Lowenstam for supplying Fig. 5A, museum photographers Dick Meier and John DeLeon for preparation of the figures of previously published illustra- tions, and museum volunteer Jo-Carol Ram- saran for library work. LITERATURE CITED BANDEL, K., 1975, Das Embryonalgehause marin- er Prosobranchier der Region von Banyuls-sur- mer. Vie et Milieu, ser. A, 25: 83-118. BANDEL, K., 1982, Morphologie und Bildung der fruhontogenetischen Gehause bei conchiferen Mollusken. Facies, Universitat Erlangen, 7: 1- 198, pl. 1-22. BATTEN, R. L., 1975, The Scissurellidae—are they neotenously derived fissurellids? American Museum Novitates, 2567: 1-29. BATTEN, R. L., 1982, The origin of gastropod shell structure. Third North American Paleontological Convention, Proceedings, 1: 35-38, 1 pl. BATTEN, R. L., ROLLINS, H. B. & GOULD, S. J., 1967, Comments on “The adaptive significance of gastropod torsion.” Evolution, 21: 405—406. BERG-MADSEN, V. & PEEL, J., 1978, Middle Cambrian monoplacophorans from Bornholm and Australia, and the systematic position of the bellerophontiform molluscs. Lethaia, 11: 113- 125. BOGGILD, O. B., 1930, The shell structure of the mollusks. Det Kongelige Danske Videns- kabernes Selskabs Skrifter, Niende Raekke, Naturvidenskabelig og Mathematisk Afdeling 9, Raekke 2: 231-326, pl. 1-15. BOGGS, C. H., 1978, Development of Hawaiian Diodora. Hawaiian Shell News, 26(10): 3-4. BOURNE, G. C., 1908, On the anatomy and systematic position of /ncisura (Scissurella) lyt- teltonensis. Quarterly Journal of Microscopical Sciences, 55: 1-47, pl. 1-5. BOUTAN, L., 1885, Recherches sur l'anatomie et le development de la Fissurelle. Archives de Zoolo- gie Expérimentale et Générale, ser. 2, tome 3, memoir 4: 1-173, pl. 31-44. BOUVIER, E. L. & FISCHER, H., 1902, L’organiza- tion et les affinités des gastéropodes primitifs d'après l'étude anatomique du Pleurotomaria beyrichii. Journal de Conchyliologie, 50: 117- 272. COX, |. R., 1960, Gastropoda—General Character- istics of Gastropoda. In MOORE, В. C., ed., Treatise on Invertebrate Paleontology, Part. |, Mollusca 2: 84-169. COX, L. R. & KNIGHT, J. B., 1960, Suborders of Archaeogastropoda. Proceedings of the Malaco- logical Society of London, 33: 262-264. CROFTS, D. R., 1929, Haliotis. Proceedings and Transactions of the Liverpool Biological Society, 43, Appendix Il: 1-174, 8 pl. CROFTS, D., 1937, The development of Haliotis tuberculata, with special reference to organo- genesis during torsion. Philosophical Transac- tions of the Royal Society of London, ser. B, 228: 219-268. CROFTS, D., 1955, Muscle morphogenesis in primitive gastropods and its relation to torsion. Proceedings of the Zoological Society of Lon- don, 125: 711-750. DZIK, J., 1978, Larval development of hyolithids. Lethaia, 11: 293-299. EALES, N. B., 1950, Secondary symmetry in gas- tropods. Proceedings of the Malacological Soci- ety of London, 28: 185-196. FRANK, E. J., 1914, Beiträge zur Anatomie der Trochiden. Jenaische Zeitschrift für Naturwis- senschaft, 51: 377—486, pl. 5. FRETTER, V., 1964, Observations on the anatomy of Mikadotrochus amabilis Bayer. Bulletin of Marine Science, 14: 172-184. - FRETTER, V., 1966, Observations on the anatomy of Perotrochus. Bulletin of Marine Science, 16: 603-614. FRETTER, V. & GRAHAM, A., 1962, British pro- sobranch molluscs; their functional anatomy and ecology. Ray Society, London, xiv + 755 p. FRETTER, V. & GRAHAM, A., 1976, The pro- sobranch molluscs of Britain and Denmark. Part 1—Pleurotomariacea, Fissurellacea, and Patel- lacea. The Journal of Molluscan Studies, Sup- plement 1: 1-37. FRETTER, V., GRAHAM, A. & MCLEAN, J. H., 1981, The anatomy of the Galapagos Rift Lim- pet, Neomphalus fretterae. Malacologia, 21: 337-361. GOLIKOV, A. & STAROBOGATOV, Y.l., 1975, Systematics of prosobranch gastropods. Mala- cología, 15: 185-232. GOULD, S. J., 1966, Notes on shell morphology and classification of the Siliquariidae (Gastropo- da), the protoconch and slit of Siliquaria squama- ta Blainville. American Museum Novitates, 2263: 1-13. GOULD, S. J., 1977, Ontogeny and Phylogeny. Harvard University Press, Cambridge, ix + 501 p. GRAHAM, A., 1965, Observations on the anatomy ARE FISSURELLIDS BELLEROPHONT DERIVATIVES? 19 of some trochacean gastropods. Bulletin of Marine Science, 15: 202-210. HARPER, J. A. & ROLLINS, H. A., 1982, Recogni- tion of Monoplacophora and Gastropoda in the fossil record: a functional morphological look at the bellerophont controversy. Third North Amer- ican Paleontological Convention, Proceedings, 1: 227-232. HICKMAN, C. S., 1983, Ecological and phylogenetic implications of the unusual radula of Laevinesta atlantica (Mollusca: Gastropoda). Veliger, 25: 323-325, figs. 1-7. HORNY, R. J., 1963, Lower Paleozoic Bellero- phontina (Gastropoda) of Bohemia. Sbornik Geo- logickych Ved Paleontologie, 2: 57-164, 44 pl. HORNY, R. J., 1965, Cyrtolites Conrad, 1838, and its position among the Monoplacophora (Mollus- ca). Sbornik Narodniho Musea Ved Praze, 21: 57-70. HUDDLESTON, W. Н., 1887-1896, A monograph of the British Jurassic Gasteropoda. Part 1. The Inferior Oolite Gasteropoda. Palaeontographical Society, London, 514 pp., 44 pl. ILLINGWORTH, J. F., 1902, The anatomy of Luca- pina crenulata Gray. Zoologische Jahrbuchern, Abtheilung fur Anatomie und Ontogenie der Thiere, 16: 449—480, pl. 31-33. KNIGHT, J. B., 1941, Paleozoic gastropod geno- types. Geological Society of America Special Papers 32, 510 p., 96 pl. KNIGHT, J. B., 1947, Bellerophont muscle scars, Journal of Paleontology, 21: 264—267. KNIGHT, J. B., 1952, Primitive fossil gastropods and their bearing on gastropod classification. Smithsonian Miscellaneous Collections, 117(13): 1-56. KNIGHT, J. B., COX, |. R., KEEN, А. M., BATTEN, В. L., YOCHELSON, Е. |. & ROBERTSON, R., 1960, Systematic descriptions (Archaeogastro- poda), In MOORE, R. C., ed., Treatise on In- vertebrate Paleontology, Part |, Mollusca 1: 169— 310, Geological Society of America and Univer- sity of Kansas Press. KONINCK, L. G. DE, 1843, Description des an- imaux fossiles qui se trouvent dans le terrain carbonifère de Belgique. Liège, 651 p., 69 pl. LEVER, J., 1979, On torsion in gastropods. /n VAN DER SPOEL, S., VAN BRUGGEN, A. C. & LEV- ER, J., eds., Pathways in Malacology, Junk, The Hague, 5-23. LINDSTROM, G., 1884, On the Silurian Gastropo- da and Pteropoda of Gotland. Kongliga Svenska Vetenskaps-Akademiens Handlingar, 19(6): 1— 250, 25 pl. LINSLEY, R. M., 1977, Some “laws” of gastropod shell form. Paleobiology, 3: 196-206. LINSLEY, R. M., 1978a, Shell form and the evolu- tion of gastropods. American Scientist, 66: 432— 441. LiNSLEY, В. M., 1978b, Locomotion rates and shell form in the Gastropoda. Malacologia, 17: 193- 206. MACCLINTOCK, C., 1963, Reclassification of gas- tropod Proscutum Fischer based on muscle scars and shell structure. Journal of Paleontol- ogy, 37: 141-156. MACCLINTOCK, C., 1967, Shell structure of patel- loid and bellerophontid gastropods (Mollusca). Peabody Museum of Natural History, Yale Uni- versity, Bulletin 22: 1-140, 32 pl. MARGOLIN, A. S., 1964, The mantle response of Diodora aspera. Animal Behaviour, 12: 187-194. MCLEAN, J. H., 1981, The Galapagos Rift limpet Neomphalus: relevance to understanding the evolution of a major Paleozoic-Mesozoic radia- tion. Malacologia, 21: 291-336. MEEK, F. B., 1866, Note on the affinities of the Bellerophontidae. Chicago Academy of Scien- ces, Proceedings, 1: 9-11. MORTON, J. E. & YONGE, C. M., 1964, Classifica- tion and structure of the Mollusca. /n WILBUR, K.M. & YONGE, C. M., eds., Physiology of Mollusca, 1: 1-58. ODHNER, М. H., 1932, Zur Morphologie und Syste- matik der Fissurelliden. Jenaische Zeitschrift für Naturwissenschaft, 67: 292-309. OWEN, G., 1958, Observations on the stomach and digestive gland of Scutus breviculus (Blain- ville). Proceedings of the Malacological Society of London, 33: 103-114. PEEL, J. S., 1972, Observations on some Lower Palaeozoic tremanotiform Bellerophontacea (Gastropoda) from North America. Palaeontol- ogy, 15: 412-422, pl. 79. PEEL, J. S., 1974, Systematics, ontogeny and functional morphology of Silurian trilobed bel- lerophontacean gastropods. Bulletin of the Geo- logical Society of Denmark, 23: 231-261. PEEL, J. S., 1975, A new Silurian gastropod from Wisconsin and the ecology of uncoiling in Palaeozoic gastropods. Bulletin of the Geologi- cal Society of Denmark, 24: 211-221. PEEL, J. S., 1976, Musculature and systematic position of Mega!lomphala taenia (Bellero- phontacea; Gastropoda) from the Silurian of Got- land. Bulletin of the Geological Society of De- nmark, 25: 49-55. PEEL, J. S., 1980, A new Silurian retractile mono- placophoran and the origin of the gastropods. Proceedings of the Geological Association, 91: 91-97. PILSBRY, H. А. 8 MCGINTY, T. L., 1952, Notes on Nesta (Laevinesta) atlantica, a Floridan fis- surellid mollusk. Nautilus, 66: 1-3. POJETA, J. & RUNNEGAR, B., 1976, The paleontology of rostroconch mollusks and the early history of the phylum Mollusca. [United States] Geological Survey Professional Paper 968: 1-88. RANDLES, W. В., 1905, Some observations on the anatomy and affinity of the Trochidae. Quarterly Journal of Microscopical Science, 78: 33-78, pl. 4-6. RISBEC, J., 1939, Recherches anatomiques sur les Prosobranches de Nouvelle-Calédonie. An- nales des Sciences Naturelles, Zoologie, ser. 11, 292392299: ROBERTSON, R., 1964, The hyperstrophic larval 20 McLEAN shells of the Architectonicidae. Annual Reports for 1963 of the American Malacological Union: 11-12. ROLLINS, H. B., 1967, The phylogeny and func- tional morphology of the Knightitinae, Carinarop- sinae, and Praematuratropidae (Gastropoda, Bellerophontacea). Columbia University, Ph.D. thesis, 197 p. ROLLINS, H. B., 1969, The taxonomic position of Cyrtonella mitella (Hall) (Mollusca, Monopla- cophora). Journal of Paleontology, 43: 136-140. ROLLINS, H. B. & BATTEN, R., 1968, A sinus- bearing monoplacophoran and its role in the classification of primitive molluscs. Paleontology, 11: 132-140. RUNNEGAR, B., 1981, Muscle scars, shell form and torsion in Cambrian and Ordovician uni- valved molluscs. Lethaia, 14: 311-322. RUNNEGAR, B. & JELL, P. A., 1976, Australian Middle Cambrian molluscs and their bearing on early molluscan evolution. Alcheringa, 1: 109- 138. RUNNEGAR, B. 8 POJETA, J., 1974, Molluscan phylogeny: the paleontological viewpoint. Sci- ence, 186: 311-317. SALVINI-PLAWEN, L. V., 1980, A reconsideration of systematics in the Mollusca (phylogeny and higher classification). Malacologia, 19: 249-278. SALVINI-PLAWEN, L. V., 1981, On the origin and evolution of the Mollusca. In Origine dei Grandi Phyla dei Metazoi, Convegno Internazionale, Roma, 7-9 Maggio 1979, Accademia Nazionale dei Lincei, Atti dei Convegni Lincei, 49: 235-293. SCHEPMAN, M. M., 1908, The Prosobranchia of the Siboga Expedition. Part 1. Rhipidoglossa and Docoglossa. Siboga-Expeditie, 49a: 1-107, pl. 1-9. STASEK, C. R. & MCWILLIAMS, W. R., 1973, The comparative morphology and evolution of the molluscan mantle edge. Veliger, 16: 1-19. STAROBOGATOV, У. |., 1970, Systematics of Ear- ly Paleozoic Monoplacophora. Paleontological Journal, 4: 293-302. [Translated from Russian, Paleont. Zhur. 1970(3): 6-17.] TERMIER, G. & TERMIER, H., 1968, Evolution et paléontologie des Gastéropodes. /n GRASSE, P. P., ed., Traité de Zoologie, Anatomie, Systé- matique, Biologie, V (Ill), 284-923, Masson, Paris. THIELE, J., 1929, Handbuch der systematischen Weichtierkunde. Erster Teil, Loricata, Gastropo- da. |. Prosobranchia (Vorderkiemer). Fischer, Jena, 376 p. TOBLER, M., 1902, Zur Anatomie von Parmophor- us intermedius Reeve. Jenaische Zeitschrift fur Naturwissenschaft, 36: 229—274, pl. 13-15. VAYSSIERE, A., 1894, Etude zoologique de la Scissurella costata, var. laevigata. Journal de Conchyliologie, ser. 3, 42: 18-29, pl. 2. WENZ, W., 1940, Ursprung und frühe Stam- mesgeschichte der Gastropoden. Archiv fur Mol- luskenkunde, 72: 1-10. WOODWARD, М. F., 1901, The anatomy of Pleuro- tomaria beyrichii Hilg. Quarterly Journal of Micro- scopical Science, 44: 215-268, pl. 13-16. YOCHELSON, E. L., 1960, Permian Gastropoda of the southwestern United States. 3. Bel- lerophontacea and Patellacea. Bulletin of the American Museum of Natural History, 119: 205— 294, pl. 46-57. YOCHELSON, Е. LE, 1967, QuosvadısmBer lerophon? In TEICHERT, C. & YOCHELSON, E.L., eds., Essays in Paleontology and Strati- graphy, University of Kansas Press, Lawrence, Kansas, 141-161. YOCHELSON, E. L., 1978, An alternative approach to the interpretation of the phylogeny of ancient mollusks. Malacologia, 17: 165-191. YOCHELSON, E. L., 1979, Early radiation of Mol- lusca and mollusc-like groups. п HOUSE, М. R., ed., The Origin of Major Invertebrate Groups, Systematics Association Special Volume No. 12, Academic Press, London: 323-358, pl. 4-6. . YONGE, C. M., 1947, The pallial organs in the aspidobranch Gastropoda and their evolution throughout the Mollusca. Philosophical Transac- tions of the Royal Society of London, ser. B, 232: 443-517, 1 pl. ZIEGENHORN, A. 8 ТНЕМ, H., 1926, Beitrage zur Systematik und Anatomie der Fissurellen. Jenaische Zeitschrift fur Naturwissenschaft, 62: 1-78, pl. 1-6. MALACOLOGIA, 1984, 25(1): 21-71 A REVIEW OF THE GENERA OF THE IRAVADIIDAE (GASTROPODA: RISSOACEA) WITH AN ASSESSMENT OF THE RELATIONSHIPS OF THE FAMILY W. F. Ponder The Australian Museum, College Street, Sydney, Australia 2000 ABSTRACT The family Iravadiidae is defined and shown to have a close relationship with the Hydrobiidae, not the Rissoidae as previously thought. Nine genera and five subgenera (of /ravadia) are provisionally recognised. Iravadiids live mainly in the Indo-West Pacific region in brackish- waters, or are marine. Seven new species, two new genera and one new subgenus are described. The Iravadiidae have a slit-like ventral opening to the pallial genital duct in the female. The bursal duct and, with one exception, the bursa copulatrix, is entirely in the pallial part of the genital tract, not posterior to the pallial cavity as in the Hydrobiidae and the Rissoidae. It also differs from these families in having a rudimentary oesophageal gland. In most other respects the Iravadiidae resemble the Hydrobiidae. The family is shown to have a wide range of shell, radular and opercular forms. Previously-named genera recognised as valid in the family are /ravadia (with Fluviocingula, Fairbankia and Pseudonoba as subgenera), Rissopsis, Hyala, Ceratia, Nozeba, Chevallieria and Rhombostoma. Key words: Gastropoda; Rissoacea; Iravadiidae; taxonomy; systematics; anatomy; man- groves. INTRODUCTION Several of the genera considered here to be iravadiids have been regarded as mem- bers of the Rissoidae (Cossmann, 1921; Thiele, 1929; Wenz, 1939; Coan, 1964; Pon- der, 1967) by most revisers. This revision has grown out of a survey of the genera of the Rissoidae currently being undertaken by the writer. During this work several genera, which previously had no suspected phyletic rela- tionships, were tentatively grouped together principally because all possessed a peculiar, flat-topped, smooth protoconch with a very small initial whorl. Examination of the radula and anatomy of representative species has confirmed their relationship and suggested their separation into a family within the Ris- soacea. This family grouping shares shell characters with the Rissoidae but in many anatomical characters resembles the Hydro- biidae. In the last major revision of the mesogas- tropods (Wenz, 1939), the genera here re- garded as iravadiids were included in 6 differ- ent families (Iravadiidae, Micromelaniidae, Rissoidae, Aclididae, Thiaridae and Eu- limidae). The purpose of this paper is to out- line the distinctive features of the lravadiidae, define the genera and discuss the rela- tionships of the family. The genus /ravadia Blanford was proposed for a small, spirally-sculptured, brackish- water gastropod from India. The usage of this name has been restricted to a few species of similar appearance, mainly from estuarine habitats. Thiele (1928) proposed new family- group names for /ravadia and for Fairbankia Stoliczka, both genera being based on spe- cies found in mangroves in southern Asia. Brandt (1968) showed that, on the basis of their shells, radulae, opercula and external features of the head-foot, Fairbankia and lra- vadia are closely related. This relationship has been confirmed by anatomical studies during this work. Johansson (1950) described the female re- productive system of a European marine spe- cies, Hyala vitrea (Montagu), and showed it to have an anterior bursa copulatrix and a slit- like opening near the posterior end of the glandular pallial duct. Golikov & Starobogatov 22 PONDER (1975) considered these differences suf- ficiently great to create a new family for Hyala. Anatomical work on species of /ravadia s.l. shows that their anatomy is similar to that of Hyala vitrea and that the /ravadia-Hyala group can be distinguished anatomically from the Hydrobiidae and Rissoidae. MATERIALS AND METHODS The specimens used in this survey are housed in several museums (see list of abbreviations), although the majority of the observations are based on material held in The Australian Museum. Scanning electron micrographs were obtained from material pre- pared as described by Ponder & Yoo (1976). Anatomical work was carried out on material fixed in Bouin's fixative or 5%-10% neutral formalin. This was sectioned at 5-6 um and stained with Mallory’s triple stain. The generic diagnoses are intended to en- compass all species included in the genus but the detail of opercular, radular and head-foot description is based only on those species for which the information is available. These species are cited in the diagnoses except when the genus is monotypic or when only the type-species has been examined. The species listed under “Distribution” are principally those examined and confirmed as belonging to the genus. Whenever possible, names of other species probably belonging to the genus are also given. A list of material examined is given under each genus to indi- cate the range available to the writer and to assist future confirmation. ABBREVIATIONS USED IN FIGURES ag albumen gland agr accessory groove ass anterior sperm sac bc bursa copulatrix bd duct of bursa copulatrix cg capsule gland co coiling part of oviduct df dorsal fold gle glandular epithelium gp gonopore hpg hypobranchial gland ¡bc opening of bursa copulatrix to cap- sule gland If longitudinal fold log lower oviduct gland mf muscular fold obc pallial opening of bursa copulatrix og rudimentary oesophageal gland p prostate gland pd pallial duct from prostate gland pm posterior limit of pallial cavity pvd pallial vas deferens r rectum rgd reno-gonidial duct ro renal oviduct sg sperm groove sr seminal receptacle SV seminal vesicle uog upper oviduct gland МС ventral channel vd vas deferens vf ventral (outer) fold of sperm groove MUSEUM ABBREVIATIONS AIM Auckland Institute and Museum, New Zealand. AMS Australian Museum, Sydney, Au- stralia. ANSP Academy of Natural Sciences of Philadelphia, U.S.A. BMNH British Museum (Natural History), London, U.K. GIT Geological Institute, University of Tokyo, Japan. GNHM Natural History Museum, Genoa, Italy. HUM Humboldt Universitat Museum, E. Berlin, E. Germany IRSNB Institut Royal des Sciences Naturelles de Belgique, Brussels, Belgium. Muséum National d’Histoire Naturelle, Paris, France. NM Natal Museum, Pietermaritzburg, South Africa. National Museum, Wellington, New Zealand. ММУ National Museum of Victoria, Mel- bourne, Australia. NMW National Museum of Wales, Cardiff, U.K. NSMT National Science Museum, Tokyo, Japan. SAM South Australian Museum, Adelaide, Australia. SMF Senckenberg Museum, Frankfurt, W. Germany. TGM Geological Museum, Turin, Italy. United States Bureau of Fisheries. National Museum of Natural History, Washington, D.C., U.S.A. ZMR Zoological Museum, Rome, Italy. IRAVADIID GENERA 23 TAXONOMY Family IRAVADIIDAE Thiele, 1928 Synonyms: Fairbankiinae Thiele, 1928; Hyalidae Golikov & Starobogatov, 1975. Diagnosis. Shell narrowly-conic to ovate- conic, usually solid, lacking inner chitinous shell layer, usually nonumbilicate, smooth or with spiral sculpture predominant, axial sculp- ture present as growth-lines, lamellae or, rarely, ribs. Aperture oval, usually weakly to distinctly angled anteriorly and posteriorly; varix on outer lip broad and strong to absent. Protoconch small, smooth, of about 1%-2% convex whorls, planorbid (coiled in one plane) to depressed dome-shaped; first whorl min- ute, terminated by a distinct varix in some species. Periostracum sometimes well de- veloped, rarely bearing processes. Head-foot often pigmented, cephalic tentacles long, slender, usually with a few stationary cilia, sometimes with colour bands; eyes at their outer bases. Snout of moderate length, usual- ly bilobed. Foot usually with anterior edge indented and expanded laterally. Posterior end of foot pointed, rounded, slightly indented or deeply bifurcate. No posterior pedal mucous gland except for a rich supply of subepithelial glands in some species; anterior pedal gland distinct, confined to pedal haemocoel. Foot lacking a metapodial tenta- cle in most species. No pallial tentacle in most species. Radula taenioglossan, central teeth with zero to four pairs of basal denticles. Operculum with eccentric, submarginal or marginal nucleus. Penis short, thick, bent double or partially coiled when at rest; with a single distal or lateral opening and with accessory, often glandular, processes on a flattened, broad distal end in most species. Female genital duct comprises a short, simple renal oviduct opening directly to albumen gland. One or two seminal receptacles (usual- ly one) open at point where renal oviduct joins albumen gland. Capsule gland with ventral gonopore variable in position and length: usually opening slit-like, either posterior, in middle section, subterminal or (in one spec- ies) most of ventral side of anterior two-thirds of gland open. Bursa copulatrix variable in position, entirely pallial in most species; bur- sal duct, when present, pallial (i.e., opens to capsule gland) or bursa opens directly to pal- lial cavity by vertical opening. A sperm sac anterior to gonopore developed in some species. Long dorsal folds in anterior part of oesophagus; rudimentary oesophageal gland present. Crystalline style present in stomach. Nervous system similar to that of Hydro- biidae. Remarks. The family name lravadiidae was given precedence over Fairbankiidae by Brandt (1968) and this is upheld here as the action of the first reviser. Starobogatov (1970) has used Fairbankiidae to include both Iravadia and Fairbankia. The main characters separating the lrava- diidae from the two closest families, the Ris- soidae and the Hydrobiidae, are given in Table 1. General Anatomical Description No attempt has been made to fully describe the anatomy of any of the species examined. The species examined anatomically are Iravadia ornata (Blanford), /ravadia quadrasi (Boettger), /ravadia (Fairbankia) bombayana (Stoliczka), /. (F.) australis Hedley, /ravadia (Pseudomerelina nov.) mahimensis (Melvill), Nozeba topaziaca (Hedley), and Hyala vitrea (Montagu). The Pallial Cavity. There is a well- developed ctenidium reaching to the posterior end of the pallial cavity and a conspicuous osphradium half to two-thirds the length of the ctenidium. The osphradium consists of a TABLE 1. The main features separating the Iravadiidae, Hydrobiidae and Rissoidae. lravadiidae Shell Aperture With weak to strong anterior channel Sculpture Smooth or spiral pre- dominant Protoconch Planorbid to depressed-dome- shaped, about 2 whorls, smooth Hydrobiidae (s.I.) Rissoidae With or without anteri- or channel Usually without anteri- or channel Smooth, spiral or axial predominant Smooth, spiral or axial predominant Dome-shaped, about 1% whorls, often pitted Dome-shaped to sub- conical, 1% to about 3 whorls, sculpture vari- able, rarely smooth 24 TABLE 1 (Continued) Operculum Radula Central teeth Head-foot Cephalic tentacles Pallial tentacle(s) Metapodial tentacle(s) Posterior pedal mucous gland Reproductive systems Ventral wall of pallial oviduct with fold enclosing ventral channel Ventral wall of pallial oviduct Glandular oviduct Anterior sperm sac Bursa copulatrix Renal oviduct Penis Prostate gland Digestive system Mid-oesophagus Spherules in secretory cells in digestive gland Nervous system Left pleural- and sub- oesophageal ganglia Right pleural-supra- oesophageal ganglia PONDER lravadiidae Hydrobiidae (s.I.) Rissoidae Oval, nucleus eccen- tric or marginal. With- out peg and calcareous material 0—4 pairs of basal cusps, rudimentary in some species Long, ciliated, pig- mented to un- pigmented Rudimentary or absent Usually absent Absent Yes Open, or closed ex- cept for posterior, sub- terminal or median opening Single gland Present or absent Anterior or posterior, within roof of pallial cavity (with one excep- tion) Renal oviduct opens directly to albumen gland With swollen distal end and accessory glandu- lar structures Half within roof of pal- lial cavity With long dorsal folds; with rudimentary oesophageal gland Absent Abutting With long connective Oval to circular, nu- cleus eccentric to cen- tral. Sometimes with calcareous peg or smear O to several pairs of basal cusps Long, ciliated or smooth, pigmented or unpigmented Rudimentary or absent Usually absent Absent Yes Closed except for small terminal or sub- terminal opening Single gland Absent Posterior, within visceral mass Renal oviduct opens to narrow oviducal coil Simple or with acces- sory glandular struc- tures Half within roof of pal- lial cavity or (rarely) entirely behind cavity With long dorsal folds; no oesophageal gland Present Fused or with short connective With long to short con- nective Oval, nucleus eccen- tric. Sometimes with horny peg; without calcareous material 1-3 pairs of basal cusps Long to moderately short, usually ciliated, rarely pigmented Present or absent Often present Present or absent No Closed (open in Mere- lina lredale), except for terminal or subterminal opening Two glands Usually absent Posterior, within visceral mass Renal oviduct opens to upper oviduct gland or (rarely) oviducal coil Usually simple; Ris- soina with swollen dis- tal end and accessory glandular structures None, or half to en- tirely within roof of pal- lial cavity; rarely be- hind cavity With short dorsal folds; no oesophageal gland Present Very short connective to abutting With moderately long connective IRAVADIID GENERA 25 FIG. 1. Head-foot of some species of Iravadiidae drawn from living material. A. /ravadia (Fairbankia) australis (Hedley). Magnetic Island, Queensland, Australia. В. /ravadia (Iravadia) quadrasi (Boettger). Darwin, Northern Territory, Australia. C. /ravadia (Pseudomerelina) mahimensis (Melvill). Maningrida, Northern Territory, Australia. D. /ravadia (Fairbankia) bombayana (Stoliczka). Sembawang, Singapore. E. Iravadia (Iravadia) ornata (Blanford). Deep Bay, New Territories, Hong Kong. Е. /ravadia (Fluviocingula) resima (Laseron). Darwin, Northern Territory, Australia. G. Liroceratia sulcata (Boettger). Taurama, near Port Moresby, Papua New Guinea. H. /ravadia (Pseudonoba) bella (Adams). Magnetic Island, Queensland, Australia. |. Nozeba topaziaca (Hedley). Port Hacking, New South Wales, Australia. Scales = 0.3 mm. 26 PONDER I fee р po ¿SE Pr | e Е a ар SER ae cet » > y Е, lil sé . Me FIG. 2. Transverse sections of the mid-oesophagus of four species of the lravadiidae. A. Nozeba topaziaca (Hedley). В. Hyala vitrea (Montagu). С. Iravadia (‘ravadia) ornata (Blanford). D. Iravadia (Fairbankia) australis (Hedley). Scales = 0.01 mm. broad, central sensory area bordered by nar- row, Ciliated ridges. A hypobranchial gland is also present. Alimentary Canal. Generally similar to that of the Hydrobiidae. The buccal mass is large and there is a pair of jaws composed of chitinous rodlets. The salivary glands are tubular and pass dorsal to the circum- oesophageal nerve ring. The anterior part of the oesophagus is short to rather long, with a pair of long dorsal folds (Fig. 2, df) that coil upwards in the larger species (/ravadia spp., |. (Fairbankia) spp.) (Fig. 2C, D) but are sim- ple in Nozeba topaziaca (Fig. 2A), Hyala vit- rea (Fig. 2B) and /. (Pseudomerelina) mahimensis. All of these species have glandular tissue in the ventral wall of the mid-oesophagus (Fig. 2, og). This glandular epithelium is composed of irregular columnar cells and is interpreted as a remnant of an oesophageal gland. Comparison of the same region in rissoids and hydrobiids shows a simple, ciliated epithelium in which are scat- tered only a few gland cells. The stomach has a crystalline style in a well-developed style sac. The rectum is looped in the dorsal wall of the mantle cavity and typically has the faecal pellets packed sideways. The digestive gland has two types of digestive cells and the ex- cretory cells do not contain large refractive granules. Female Reproductive System. This system was examined in detail, mainly by serial sec- tions. Because of the brittle nature of the oviduct glands and the limited amount of material, some observations could not be confirmed and others are doubtful for particu- lar species. The single female sectioned of /. ornata was not fully mature. The narrow, thin-walled upper oviduct opens into a slightly thicker, ciliated, non- muscular, short renal oviduct. No gonoperi- cardial or reno-gonidial duct was observed in any species except /. ornata, where there is a short, ciliated reno-gonidial duct. The glandu- lar part of the oviduct consists of a posterior albumen gland (Figs. 4-7, ag) which partially displaces the kidney and which is continuous with the capsule gland lying in the roof of the pallial cavity. A seminal receptacle lies along- IRAVADIID GENERA 27 FIG. 3. A-F. Transverse sections through the ventral part of the female pallial genital duct of Нуа/а vitrea from near the anterior end to near the posterior end. All figures to same scale. side the albumen gland and is partly em- bedded in it. Hyala has one (?) or two seminal receptacles and /ravadia (Pseudomerelina) mahimensis appears to have two. The other species appear to have only one (Figs. 4-7, sr). The seminal receptacle(s), together with a renal oviduct, opens into the oviduct just pos- terior to the capsule gland. This gland (Figs. 3—7, cg) opens to the pallial cavity by along to short, slit-like ventral opening (gp) or gono- pore. In Nozeba topaziaca and Hyala vitrea the oviduct aperture is in the posterior part of the capsule gland and the anterior part of the gland is blind (Fig. 21B). The oviduct opening extends along all of the anterior two-thirds of the ventral side of the capsule gland in /. (Fairbankia) bombayana (Fig. 7D) but is short and placed in the anterior part of the gland in I. ornata (Fig. 21G), I. (Pseudomerelina) mahimensis (Fig. 5) and /. (Fairbankia) au- Stralis) (Fig. 6). It is long and in the middle of the gland in /. quadrasi (Fig. 4). There is a deep sperm groove (sg) on the lower, inner part of the inner (left) ventral wall of the cap- sule gland in the vicinity of the pallial opening in I. (F.) australis, |. quadrasi (Fig. 4B—D), I. (P.) mahimensis (Fig. 5B), N. topaziaca and H. vitrea (Fig. 3B-E). This gutter runs forward and closes over to form an anterior sperm sac (Figs. 3-6, ass) which appears to function as a bursa copulatrix. In /. (F.) bombayana there is a sperm sac on the outer (left) side of the inner (left) wall of the posteror third of the capsule gland and it is wholly within the roof of the pallial cavity (Fig. 7, bc). This sac opens by way of a vertical slit (Fig. 7, obc), the outer edge of which represents the outer fold of the sperm groove, and its internal walls are heavily folded. In /. (F.) australis (Fig. 6) there is a similar sperm sac (bc), the opening (obc) of which lies in the anterior second quarter of the capsule gland, but there is also a small anterior sperm sac (ass) anterior to the ovi- ducal opening. To enable these sperm sacs to be distinguished, the structure external to the capsule gland seen in /. (Fairbankia) species is hereafter called the bursa copulatrix and the sperm sac anterior to the oviducal opening is referred to as the anterior sperm sac. Both structures appear to function as a bursa copulatrix but no evidence of sperm ingestion was observed in sections of either structure. They both store sperm, although the bursa copulatrix was empty in most speci- 28 PONDER FIG. 4. Female reproductive system of /ravadia (lravadia) quadrasi. A-E. Sections through the pallial genitalia at the positions marked on Fig. F. F. Lateral view of female genital system (excluding upper oviduct and ovary). Figs. A-E to same scale. IRAVADIID GENERA 29 ass 0-1mm i i is. A-F. Transverse sections of FIG. 5. Female reproductive system of /ravadia (Pseudomerelina) mahimensis. the pallial genitalia at the positions marked on Fig. G. G. Lateral view of female reproductive system, excluding the upper oviduct and ovary. Figs. A-F to same scale. 30 PONDER cg ass ass ass gp A 7 gp bc B C ass 0-1mm ae É ta vc If FIG. 6. Female reproductive system of /ravadia (Fairbankia) australis. A-E. Section through the pallial genitalia at the positions marked in Fig. F. F. Lateral view of female reproductive system, excluding the upper oviduct and ovary. Figs. A-E to same scale. IRAVADIID GENERA 31 I NO ET} N Wiss FIG. 7. Female reproductive system of /ravadia (Fairbankia) bombayana. А-С. Sections through the pallial genitalia at the position marked in Fig. D. D. Lateral view of female genital system, excluding the upper oviduct and ovary. Figs. A-C to same scale. mens sectioned, suggesting a very temporary storage function. In the two species of /ravadia (Fairbankia) the bursa copulatrix is open laterally to the capsule gland for a short distance (Fig. 6, ibc) and is then separated from it and becomes a blind pocket (Fig. 6E, 7C). The seminal receptacle (sr) opens into the posterior end of the ventral channel. The posterior section of the ventral channel has, as in all the species examined, a ciliated fold (If) similar to that seen in the Hydrobiidae (see Discussion). lravadia quadrasi (Fig. 4) has a ventral bursa copulatrix (bc) behind the oviducal opening extending back to the end of the pallial cavity. This structure is formed from the closure of the ventral (outer) fold of the sperm groove (vf), but the pallial opening of the capsule gland is closed slightly in front by the merger of the inner ventral edge of the ventral chan- nel of the capsule gland and the upper edge of an accessory fold on the upper side of the sperm groove (Fig. 4C, D). This accessory fold is the dorsal edge of a glandular groove (Fig. 4, agr) containing distinctive, blue- staining, cuboidal gland cells in its anterior section. Anteriorly, the groove is blind and merges with the capsule gland after forming a tubular structure. Posteriorly, it continues into the bursa copulatrix where its epithelium is similar to that of the remainder of the bursa. The prominent muscular fold (Fig. 4E, mf) separating the sperm groove and the groove dorsal to it, continues through the bursa. It was not possible to determine from the avail- able material if the fold continued the full length of the bursa. In /. (Pseudomerelina) mahimensis (Fig. 5) the duct of the bursa copulatrix (bd) has been formed by the clo- sure of the sperm groove just behind the opening to the capsule gland (Fig. 5, В-С). The tubular bursal duct runs posteriorly and then dorsally into the dorsally placed bursa copulatrix (bc) which lies above the posterior end of the capsule gland (Fig. 5, D-G). The type-species of /ravadia, |. ornata (Fig. 21G), shows a similar arrangement to that seen in the Hydrobiidae in possessing an anterior 32 PONDER oviducal opening, a ciliated fold in the ventral inner wall of the gland to open to the long, channel of the capsule gland and no anterior tubular bursa which lies latero-dorsally on the sperm sac. The bursal duct opens immediately outer (right) side of the capsule gland and the behind the short, slit-like, muscular, sub- anterior half of the albumen gland. Thus the terminal opening and runs dorsally over the bursa extends behind the posterior end of the vd \ ) A IN ; | | ps se Gi T= < FIG. 8. A-G. Penes of some species of Iravadiidae. In Figs. C-F glandular areas are represented by stipple. A. Iravadia (lravadia) ornata (Blanford). Deep Bay, New Territories, Hong Kong. В. /ravadia (Iravadia) quadrasi (Boettger). Hervey Bay, Queensland, Australia. С. /ravadia (Pseudomerelina) mahimensis (Mel- vill). Magnetic Island, Queensland, Australia. D. /ravadia (Fairbankia) australis (Hedley). Magnetic Island, Queensland, Australia. E. Hyala vitrea (Montagu). Ten km W of Strômstad, W Sweden, 35 m. F. Nozeba topaziaca (Hedley). Port Hacking, New South Wales, Australia. G. /ravadia (Fairbankia) bombayana (Stoliczka). Sembawang Estuary, Singapore. H. Diagrammatic representation of the prostate gland and vas deferens of /ravadia (lravadia) ornata (Blanford). Scales = 0.2 mm. IRAVADIID GENERA 33 pallial cavity in this species. А thin-walled vestibular region may function as an anterior sperm sac. Male Reproductive System. Examined in Iravadia ornata, |. quadrasi, |. (Fairbankia) bombayana, I. (F.) australis, I. (Pseudomere- lina) mahimensis and Hyala vitrea. The vas deferens forms a long, coiled semi- nal vesicle (Fig. 8H, sv) and the renal vas deferens is short and ciliated. The bean- shaped prostate gland (p) lies half within the pallial roof and has a median lumen except in a short middle section which is thin-walled ventrally. This section receives the renal vas deferens behind the posterior pallial wall and the pallial vas deferens (pvd) emerges in front of this wall. There is a very short, ciliated duct (pd) in /. (Pseudomerelina) mahimensis and I. ornata which emerges next to the junction of the prostate gland and the pallial vas de- ferens, and opens to the pallial cavity. No similar pallial opening was observed in the other species, although the sections were not sufficiently good to positively exclude the presence of such a duct. The thin-walled, ciliated pallial vas deferens runs along the right side of the body wall and is not contained within it. The penis is large, with an expanded distal portion bearing one or more glandular processes (Fig. 8, А-С). There are both epithelial and internal glands. The single, muscular, ciliated penial duct is completely enclosed in the penis and lacks any epithelial connection to the exterior. Nervous System. This has not been studied in detail but the nervous systems of /ravadia (Fairbankia) bombayana and |. (F.) australis appear to be generally similar to that of most Hydrobiidae. There is a long supra- oesophageal connective and the sub- oesophageal ganglion lies against the left pleural ganglion. The Renal System. The kidney is a simple sac with a thin lining except for a cluster of cells on the outer wall. Key to the genera of the Iravadiidae 1. Shell elongate, sub-cylindrical, translucent, with expanded outer lip.......... Rissopsis Shell elongate to ovate, opaque to semitranslucent, without expanded outer lip ...... 2 2(1) Shell elongate, smooth or with spiral sculpture, outer lip without varix, sinuate, columellasmore-or-less Vertical! ... ....... 20... wie cect 4... Rhombostoma Shell elongate to ovate, variously sculptured or smooth, outer lip with or without varix, more-or-less straight, columella concave in most species ................... 3 3(2) Operculum with submarginal to marginal nucleus and concentric to sub-spiral RN SA ee à 4 Operculum with eccentric nucleus and spiral growth ............................... 5 4(3) Shell with spiral sculpture, well-developed in some species, operculum with mar- incl A e oo epee lravadia Shell with spiral sculpture very weak or absent, operculum with submarginal nu- TELE RL CTI rs ore ee Bios AS SS ee Chevallieria as) ee Onelluwithesthongiy prosocline outer Ир ..........>.. лось Sue o el 6 Shell with approximately orthocline outer lip ..........:............................ 8 Sieissshelliwitnktinexspiralithreads ee ani cca о я Не ВА т. ВЕН Вано Hyala MOI onielll with) penpheral midge: - 22.2... 2.2 oss aes ec so eens ao ce eee ales Acliceratia nov. Shellllackingipenpheralindge 22.20... ооо oso eels esis cme er Ceratia 8(5) Shell with strong spiral lirae and a prominent varix at outer lip ........ Liroceratia nov. Shell with weak spiral threads or smooth, without varix at outer lip ........... Nozeba Genus /ravadia Blanford, 1867: 56-58 Type-species: /ravadia ornata Blanford, 1867 (? = Pyrgula clathrata A. Adams, 1853); monotypy. Recent. India. Synonym: lrawadia, err. auct. Diagnosis. Shell. Elongately ovate-conic to ovate-conic, non-umbilicate to nar- rowly-umbilicate, with predominantly spiral sculpture, a few species with conspicuous axials or with sculpture reduced to spiral rows of pits. Protoconch planorbid to depressed dome-shaped, of about 2-22 smooth, con- vex whorls, usually terminated by a distinct varix. Periostracum often well-developed. Head-foot. Cephalic tentacles often banded 34 PONDER or spotted. Snout of moderate length, bilobed. Foot and head pigmented dorsally in most species. Anterior edge of foot usually in- dented, produced laterally. No pallial tenta- cles and usually no metapodial tentacle. Penis. With broad flattened head and acces- sory glandular areas. Oviduct. With long to short ventral pallial opening; bursa copulatrix present; anterior sperm sac present or possi- bly absent. Operculum. Oval, nucleus on columellar edge, with close, concentric growth lines; some species with weak internal ridge(s). Radula. Central teeth wide, each with a usually wide cutting edge bearing one to several, small, sharp cusps; lateral edges unthickened; zero to four basal denticles, weak in most species. Lateral teeth with nar- row cutting edge, rather short cusps and long outer portion. Marginal teeth long, curved, with several small cusps. Distribution and habitat. Indo-Pacific; in brackish-water or in shallow-water, sheltered marine environments to the continental shelf. Remarks. The genus /ravadia possesses a peculiar operculum with a lateral nucleus and concentric growth rings. Although no satis- factory explanation can be offered for the development of this type of operculum, it is assumed that it arose early in the Tertiary, probably from the Chevallieria lineage. Evi- dence for this assumption is the existence of an operculum intermediate in structure in a Recent species attributed to Chevallieria and described below. Thiele (1928), on the basis of the structure of the operculum, created two new subfamilies, one for Fairbankia in the Micromelaniidae and the other for /ravadia in the Hydrobiidae. His placement of these genera in two different families was based on the presence of lateral (basal) denticles on the central teeth of the radula in /ravadia and the assumed absence of these in Fairbankia. The virtually identical opercular characters and the general similarity of the radula led Brandt (1968, 1974) to include both genera in the lravadiidae. Because a continuum in shell and radular characters can be observed be- tween typical species of /ravadia and Fair- bankia these two groups are regarded here as being congeneric. There are, however several important differences in the female genitalia of the type-species of these two groups. Examination of an additional species of /ravadia s.s., |. quadrasi, indicates that considerable differences in the female genita- lia may exist within the groupings as here recognised. Similarly, differences between /. (Fairbankia) australis and I. (Е.) bombayana are also marked, although not so radically. The taxonomy adopted is conservative in that some traditional characters (especially the operculum) are given considerable weight. There are too many gaps in the avail- able information for genital characters to be used in a primary way at the genus-group level at this time, although, undoubtedly they will ultimately be extremely valuable in refin- ing the classification. Several subgenera are used because these appear to be recognisable groupings, although difficult to define clearly at a higher level. No doubt, when more anatomical in- formation is available, some will be discarded and others may be elevated to generic rank. The five subgenera recognised show a wide diversity of shell sculpture but are gener- ally similar in other shell features and in their radular and opercular characters as well as in the external appearance of the head-foot. /ra- vadia appears to have diversified from a marine group (Subgenus Pseudonoba) which has an ancestry traceable at least to the Miocene, and possibly to the Eocene, of Eu- rope. Several species here included in /rava- dia (Pseudonoba) live in brackish waters or in sheltered, shallow coastal waters; others live on the continental shelf. The incursion into brackish water by members of this genus possibly took place on more than one occasion (Fig: 22). Key to subgenera of /ravadia ile Shell ovate-conic, with heavy spiral cords (two to six on penultimate whorl); oper- culum) with nucleus: in middle of columellar edge ............. 7 2 Shell ovate-conic to narrowly-elongate, smooth or with weak to moderate spiral sculpture (more than four cords on penultimate whorl); operculum with nucleus in middle of columellar edge or displaced from middle ............................. 3 2(1) Shell with distinct axial ribs; operculum without internal ridges . . ..Pseudomerelina nov. Shell with axial ribs or threads; operculum with two internal ridges radiating from nucleus in middle of columellar edge Iravadia S.S. IRAVADIID GENERA 35 3(1) Shell smooth or with weak to moderate spiral cords and fine axial threads or lamel- ее: Sta. Shell with spiral rows of punctures ..... 4(3) Outer lip of aperture evenly prosocline, base of shell without distinct fold о le en a Gr ne ee, ae 4 E Dah AG SIRE chee Fluviocingula .. Fairbankia Outer lip of aperture slightly prosocline, orthocline, or weakly opisthocline, with shallow anterior and posterior channels, base of shell with distinct fold in most SPECIES Mere PR cere re seks à Subgenus /ravadia s.s. Diagnosis. Shell. Of moderate to small size, ovate-conic to elongate-conic, non- umbilicate, with strong spiral cords and weak axial threads to axial ribs. Aperture oval, sub- angled posteriorly and anteriorly, peristome usually thick; outer lip with heavy varix, slight- ly prosocline. Protoconch (Figs. 9F, 10B) as for genus. Periostracum well developed. Figs. 9A, F, H; 10A, B. Head-foot. Cephalic tenta- cles with narrow, black bands (in /. ornata and I. quadrasi) and white spots. Foot with almost straight to indented anterior edge, indented (/. angulata), rounded (/. ornata) or pointed (I. quadrasi) posterior end. No metapodial tenta- cle. Figs. 1B, E. (/. ornata, Hong Kong (New Territories), /. quadrasi, Singapore and Dar- win, /. angulata, Darwin). Figs. 1B, Е. Penis. (Not examined in /. angulata) Bent forwards on itself when at rest, with aperture on inner edge in middle of broad distal half. Outer portion of distal half lamella-like, inner part thick, rugose (/. ornata, Fig. 8A) or both sides rugose (/. quadrasi, Fig. 8B). Oviduct. Known for /. quadrasi (Fig. 4) and /. ornata (Fig. 21G). Anterior sperm sac present and long oviducal opening in middle of capsule gland in /. quad- rasi; bursa copulatrix a posterior, ventral sac. Bursa copulatrix on right side and latero- dorsal to capsule gland and albumen gland in |. ornata and bursal duct opens immediately behind short subterminal opening on left side. A single seminal receptacle in both species. Operculum. As for genus, with nucleus mid- way on columellar edge. Two low radial folds emerge from nucleus and cross about two- thirds of inner surface of operculum (/. ornata, I. angulata, |. quadrasi). Figs. ЭВ, С; 10D, Е. Radula. Central teeth wide, low, each with short, tongue-like projection in middle of ven- tral edge and a single, short, weak denticle just inside each unthickened lateral edge (ab- sent in /. angulata); a second pair of weak denticles on some central teeth of some specimens of /. quadrasi. Lateral and margin- al teeth as for genus; lateral teeth with only a long, primary cusp and a small cusp inside BG Acs ch EAA IE AI CATA oe EA, Pseudonoba this in /. angulata; multicuspate in the other two species examined (/. ornata, I. quadrasi). Figs. 9C-E; 10C. Distribution. Southeast Asia as far N as Hong Kong (New Territories) (/ravadia ornata Blanford, 1867 ? = Pyrgula clathrata A. Adams, 1853, = /ravadia princeps Preston, 1915 ? = I. funera, l.ennurensis and I. an- nandalei Preston, 1916). Central Indo-Pacific (Alvania quadrasi Boettger, 1893 = Rissoa garretti Tate, 1899, nom. nov. pro R. venusta Garrett, 1873, non Philippi, 1844 = Rissoa (Alvania) alveata Melvill & Standen, 1901 = Merelina humera Laseron, 1956 = Merelina goliath Laseron, 1956 = Merelina reversa Laseron, 1956 = Planapexia quadrina Laser- on, 1956 = /ravadia reticulata Brandt, 1968). Northern and NE Australia (Rissoina carpen- tariensis Hedley, 1912 = Pellamora amplexa Laseron, 1956 = Pellamora truncata Laser- on, 1956); Pellamora capitata Laseron, 1956 = Pellamora spiralis Laseron, 1956; Pella- mora angulata Laseron, 1956). Habitat. Iravadia ornata was collected un- der stones in the lower littoral, in an estuarine situation, but not associated with extant man- groves. /ravadia angulata and I. quadrasi have been found in mangroves under objects in shallow pools. Brandt (1974) records /. ornata and I. quadrasi (= reticulata Brandt) as living in brackish water in the drainages of mud flats, nipa palm and mangrove swamps, and in the estuarine area of rivers. They were found partly buried in mud, feeding on decay- ing Organic material. Material Examined. |. ornata. Specimens ex Blanford and several other lots (AMS). НЯ. carpentariensis. Holotype, paratypes and several other lots (AMS). Pellamora species of Laseron, 1956. Types (AMS) and other lots of P. capitata and P. angulata (AMS). A. quadrasi. Lectotype, paralectotypes (SMF) and many other lots (AMS). P. quadrina. Holotype (AMS). M. goliath. Holotype and paratypes (AMS). R. garretti. Holotype (ANSP). R. (A.) alveata. Holotype (BMNH). M. humera. Holotype AMS). M. reversa. Holo- type and paratypes (AMS). 36 PONDER FIG. 9. A-D. /ravadia (Iravadia) ornata (Blanford), type-species of Iravadia. Deep Bay, New Territories, Hong Kong (AMS). A. Shell. В. Operculum, inner side. C, D. Radula. C. Central teeth only. E-H. /ravadia (Iravadia) angulata (Laseron). Norman River, Gulf of Carpentaria, Queensland, Australia (AMS). E. Radula. F. Protoconch. G. Operculum, inner side. H. Shell. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 37 Remarks. The type-species, /. angulata and /. quadrasi are estuarine and, whereas the other two Australian species apparently have not been collected alive, they probably have a similar habitat. lravadia angulata differs from I. ornata in shell characters (compare Figs. 9A and Н), the shell being more like that of species of Iravadia (Fairbankia) in its tall spire, relatively weak spiral cords, thin peristome and weak varix. The angulated whorls are, however, atypical of the subgenus Fairbankia. The operculum (Fig. 9G) is nearly identical to that of /ravadia ornata (Fig. ЭВ) and it is mainly because of the similarity in this structure that |. angulata is tentatively included in /ravadia s.s. The radula of /. angulata has central teeth (Fig. 9E) lacking any basal denticles and the lateral teeth are virtually unicuspid and differ from other species examined in the genus in this respect, except for two new species of /. (Pseudonoba) described in the Appendix and discussed below. /ravadia quadrasi (Fig. 10A) has more pronounced axial sculpture than /. ornata but agrees in other respects. The differences in the female genitalia of /. ornata and I. quadrasi are considerable but other species of /ravadia should be examined to determine the limits of variation of the female genitalia before further subdivision of this group is proposed. Subgenus Pseudomerelina Ponder, n. subgen. Type-species: Alvania mahimensis Melvill, 1893. Recent, Bombay, India. Diagnosis. Shell. Ovate-conic, with oval aperture, anterior subangulation of aperture absent; sculpture of axial ribs and spiral cords, gemmate at points of intersection. Aperture oval, not markedly subangled posteriorly, rounded anteriorly; outer lip prosocline, varix strong. Protoconch (Fig. 10J) as for genus. Periostracum thin. Figs. 101, J. Head-foot. With conspicuous pigmentation on snout and bands on tentacles. Cephalic tentacles with short ‘setae’ distally and active cilia in spiral series along rest of tentacles. Foot weakly cleft anteriorly. Posterior end of foot with a weak indentation and a very short, flattened tentacle bearing short, stationary cilia. (Specimens ex- amined from Darwin and Magnetic Island, Australia and Singapore). Fig. 1C. Penis. Flat- tened, with two protuberances, one small. glandular swelling at about half-length on out- er side and a flattened section near distal end on outer side. Penial duct in rounded, distal lobe on inner side. Fig. 8C. Oviduct. With small anterior bursa copulatrix and pallial, dorsal, posterior bursa copulatrix with narrow, vertical duct. Oviducal opening short, an- teriorly placed. Two seminal receptacles (Fig. 5). Operculum. As in /ravadia s.s. but without internal ridges. An irregular, thickened area inside columella edge in a few specimens. Fig. 10F, G. Radula. As in /ravadia s.s., with one basal denticle on central tooth. Lateral teeth (1-3) + 1 + (3-4). Fig. 10H. Distribution. Southeast Asia, India, and cen- tral Indo-Pacific to tropical Australia. (A/vania mahimensis Melvill, 1893 = Merelina sucina Laseron, 1956 = Merelina solida Laseron, 1956 = /ravadia tuberculata Brandt, 1974). Habitat. Seaward edge of mangroves, es- pecially on the edge of creeks, on weed, etc., and objects in small pools. Usually abundant when present. Material Examined. A. mahimensis. Two syntypes (BMNH) and several other lots (AMS). M. sucina. Holotype and paratypes (AMS). M. solida. Holotype (AMS). Remarks. This subgenus is proposed for a single species, the shell of which is dis- tinguished from those included in /ravadia s.s. by its gemmate sculpture, relatively weak spir- al cords, two purple spiral bands on the body whorl, and more evenly-oval aperture. The radula is almost identical to that of /ravadia ornata but the operculum, although similar in shape, lacks any internal ridges. /ravadia (Pseudomerelina) mahimensis is the only species of the Iravadiidae known to possess a metapodial tentacle (Fig. 1C). It has presum- ably disappeared in the other species of the family examined alive, although there is a low ridge on the posterior end of the foot in /. quadrasi (Fig. 1B). In addition, the penis is of simpler construction than in the other sub- genera of /ravadia; it has a less swollen distal portion and only two small accessory pro- tuberances; and there are two seminal re- ceptacles in the female genital system, not one as in other members of the genus. Iravadia (Pseudomerelina) mahimensis lives in an estuarine habitat and in the shel- tered waters of enclosed bays and appears to have a wide distribution through the central Indo-Pacific. The shells of /ravadia (Pseudomerelina) 38 PONDER FIG. 10. A-E. /ravadia (Iravadia) quadrasi (Boettger). A. Shell. В. Protoconch. С. Radula. D-E. Operculum, inner (D) and outer (E) sides. A, B, D. Proserpine River estuary, Wilson, Queensland, Australia. C. Gatakers Bay, Hervey Bay, Queensland. E. Mouth of Brisbane River, Queensland. All AMS. F-J. /ravadia (Pseudomerelina) mahimensis (Melvill). Type-species of Pseudomerelina nov. F, G. Operculum, inner (F) and outer (G) sides. H. Radula. |. Shell. J. Protoconch. Е. Magnetic Island, Queensland, Australia. G—J. Maningrida, Arnhem Land, Northern Territory, Australia. Both AMS. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 39 mahimensis (Fig. 101) and /ravadia (Iravadia) quadrasi (Fig. 10A) are superficially very sim- ilar to species of Alvania and Merelina (Ris- soidae) but can be distinguished by their pro- socline outer lips and their small, smooth, flattened protoconchs. Subgenus Fairbankia (Blanford MS) Stoliczka, 1868 (July): 274 Type-species: Fairbankia bombayana (Blanford MS) Stoliczka, 1868, original de- signation. Recent, Bombay, India. Synonyms: Pellamora lredale, 1943: 206. Type-species: /ravadia australis Hedley, 1900; original designation. Recent, NE Au- stralia. Wakauraia Kuroda & Habe, 1954: 75. Type-species: Fairbankia (Wakauraia) saka- guchii Kuroda & Habe, 1954. Recent, Japan. Diagnosis. Shell. Elongately-conic, non- umbilicate, solid, with weak to moderate spiral sculpture and weak axial threads or lamellae. Aperture relatively small, weakly to distinctly angled anteriorly, very weakly channelled posteriorly, outer lip slightly to moderately prosocline. Protoconch as for genus. Per- iostracum moderately well-developed, with short processes in some species. Fig. 11A, F. Head-foot. Cephalic tentacles very slightly ‘setose’ distally, with or without a few darkly pigmented bands or black spots. Head and foot pigmented dorsally. Anterior edge of foot indented, posterior end blunt, with shallow indentation. No metapodial tentacle. (1. (F.) australis, |. (F.) bombayana). Fig. 1A, D. I. (F.) cochinchinensis and I. (F.) rohdei appear to be similar according to Brandt’s (1974) descrip- tion. Penis. (/. (F.) australis). Short, com- pressed, with a row of accessory glandular swellings below an expanded distal portion on which penial duct opens subterminally. Fig. 8D., I. (F.) bombayana with broad, simple head; only one glandular area apparent in middle region (Fig. 8G). Oviduct. Anterior sperm sac absent (/. (F.) bombayana) or very reduced (1. (F.) australis); bursa copulatrix with vertical opening separated from oviducal opening. Oviducal opening short and an- teriorly placed (/. (F.) australis) or occupying anterior two-thirds of capsule gland (/. (F.) bombayana). A single seminal receptacle present. Figs. 6, 7. Operculum. Oval, with nucleus approximately two thirds along col- umellar edge, growth-lines concentric, a more-or-less longitudinal, low, internal ridge present (/. (F.) bombayana and 1. (F.) au- Stralis). I. (F.) sakaguchii is similar (Kuroda 8 Habe, 1954, text fig. 12). Fig. 11D, E, G. H. Radula. Similar to that of /ravadia s.s. but central teeth with two or three weak, basal denticles placed against upper lateral mar- gins and just beneath outer, dorsal cutting edge (I. (F.) bombayana and I. (F.) australis). I. (F.) sakaguchii is similar but no basal den- ticles are shown in the illustration of the cen- tral tooth (Kuroda & Habe, 1954, text fig: 110). Fig? BC №. Distribution. South and SE Asia, Sumatra, S China, Philippines (F. bombayana = Fair- bankia quadrasi Boettger, 1893; Fairbankia cochinchinensis Bavay & Dautzenberg, 1940: Fairbankia rohdei Brandt, 1968: Onoba tenuilirata Boettger, 1893). Northern and NE Australia (/ravadia australis). Red Sea (?Ono- ba elongata Hornung & Mermod, 1928). Japan (F. (W.) sakaguchii). Habitat. Mangroves, under objects in shal- low pools (I. (F.) australis) or on the surface of mud in standing water (/. (F.) bombayana). Drainage system of mudflats, nipa palm and mangrove swamps (/. (F.) bombayana, |. (F.) rohdei, Brandt, 1974). Material Examined. F. bombayana. A few lots (AMS, BMNH). F. quadrasi. One lot (AMS). F. cochinchinensis. Three lots, ex Brandt (USNM). /. australis. Holotype, para- types and several other lots (AMS). F. rohdei. Four lots (USNM). О. tenuilirata. Holotype (SMF). O. elongata. Holotype (GNHM). F. (W.) sakaguchii. Paratypes (USNM). Remarks. The species in this subgenus differ from those in /ravadia s.s. in their more elongate shells with weaker spiral sculpture, more oval apertures, and the bursa copulatrix has a vertical opening separate from that of the capsule gland. The central teeth of the radula have only weak, dorso-laterally placed basal denticles and the nucleus of the oper- culum is displaced from the middle of the columellar edge. The species of /ravadia (Fairbankia) live in an estuarine habitat, usually amongst mangroves, like those of /ra- vadia S.S. Brandt (1968) described the radula, oper- culum and animal of his new species F. rohdei from Thailand and Thiele (1928) de- scribed the radula and operculum of F. bom- bayana. These species have radulae and opercula very similar to those of /ravadia ornata; Blanford (1868) and Brandt (1968, 1974) recognised them as closely related genera. Brandt (1974) pointed out the similar- 40 PONDER FIG. 11. A-E. Iravadia (Fairbankia) bombayana (Stoliczka), type-species of Fairbankia. A. Shell. В, С. Radula, central tooth. D-E. Operculum, outer (D) and inner (E) sides. A, D, E. Bombay, India (ex Blanford) (BMNH). В, С. Sembawang Estuary, Singapore (AMS). F-I. /ravadia (Fairbankia) australis (Hedley), type-species of Pellamora. F. Shell. Paratype, Bowen, Queensland, Australia (AMS). G, H. Operculum, outer (С) and inner (H) sides. |, J. Radula. G-J. Magnetic Island, Queensland, Australia (AMS). Scales: shells = 1 mm; opercula = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 41 ity of the shells of Fairbankia bombayana to Mainwaringia Nevill, 1884 (type-species Melania (Mainwaringia) paludomoidea Nevill, 1884) but the radula as figured by Annandale & Prashad (1919) is dissimilar and the oper- culum is described as “horny, extremely thin, paucispiral, with the nucleus eccentric.” Cer- tainly the characters of the radula and op- erculum remove Mainwaringia from any close association with Fairbankia. The shells of the type-species of Pellamora and Fairbankia differ in the relative strength of the spiral sculpture. A gradation is seen, however, in three SE Asian species, from weak spiral sculpture in /. (F.) bombayana to distinct spirals in /. (F.) cochinchinensis to moderately strong spirals in /. (F.) rohdei. The spiral sculpture of this last species approaches that of /. (F.) australis (Fig. 11F). Iravadia angulata is, on the other hand, very similar in shell characters (Fig. 9H) to /. (F.) australis. The type-species of Wakauraia is, as admitted by its authors, very similar to other species of Fairbankia. The main differ- ence appears to be in the degree of angula- tion of the anterior end of the aperture. As this character alone is not considered to be of subgeneric importance, Wakauraia is re- garded here as a synonym of Fairbankia. Blanford (1868) published the description of his new genus (Fairbankia) in December, 1868. In his discussion he cites the Stoliczka (1868) reference in which his genus name was inadvertently introduced in July, 1868. Because Stoliczka gives a full description of the genus and species, he should be re- garded as the author of Fairbankia. Subgenus Fluviocingula Kuroda & Habe, 1954: 73 Type-species: Fluviocingula nipponica Kuroda & Habe, 1954; original designation. Recent, Japan. Synonym: Mesodestea Laseron, 1956: 451. Type-species: Mesodestea resima Laseron, 1956; original designation. Recent, northern Australia. Diagnosis. Shell. Small, ovate-conic, rather thin, usually narrowly-umbilicate with evenly- convex whorls, sculptured with very weak, scarcely-raised spiral cords and axial threads, interspaces forming shallow pits (Fig. 12D). Aperture oval, with very weak anterior and posterior angulations, peristome thin, varix weak, outer lip prosocline. Protoconch as for genus. Periostracum rather thin. Fig. 12A, B, D. Head-foot. Head pigmented, cephalic tentacles with a broad band of pigmentation and stationary ‘setae’ distally. Foot cleft an- teriorly and posteriorly. No metapodial tenta- cle (F. resima, Darwin, N Australia). Fig. 1F. Penis and oviduct not known. Operculum. Similar to that of species of /ravadia (Iravadia) but without internal ridges (/. (F.) nipponica, |. (F.) resima). Fig. 12C, H. Radula. Central teeth similar to those in /ravadia s.s. and |. (Fairbankia), with central teeth showing from none to two laterally placed basal denticles on each side (none or one in /. (F.) resima, two in |. (F.) nipponica). Fig. 12E-G. Distribution. Inland Sea of Japan and Sea of Japan (F. nipponica). Northern Australia (M. resima). Habitat. Associated with mangroves but at their inshore edge in damp areas (1. (F.) re- sima). Saltwater lagoons and estuaries on mud and Zostera, up to 1 m in depth from negative temperatures to 33°C (summer); salinity 4—7% (l. (F.) nipponica, Golikov & Kussakin, 1978). Material Examined. F. nipponica. One lot ex Golikov (AMS); paratypes (USNM). M. re- sima. Holotype and a few other lots (AMS). Remarks. The type-species of Fluviocing- ula and Mesodestea are very similar, although the Australian species has a more elongate shell (Fig. 12A, B) and a slightly different radula (compare Figs. 12E, F with G). This subgenus differs from the others included in /ravadia by its curious shell sculp- ture, a series of minute spiral pits (Fig. 12D) between the weak spiral cords. The narrow umbilicus also sets the shell of this group of species apart from the remainder of the genus, although in an undescribed Australian species the umbilicus is absent. The oper- culum resembles that of species of /ravadia (Pseudomerelina), in lacking any internal ridges. Golikov & Kussakin (1978) report probable viviparity in /. (F.) nipponica. Subgenus Pseudonoba Boettger, 1902: 145 Type-species: Pseudonoba peculiaris Boettger, 1902; original designation. Middle Miocene, Rumania. Synonyms: Sinusicola Kuroda & Habe, 1950 (Jan. 15): 16. Type-species: Turbonilla (Careliopsis) filiola Yokoyama, 1927 (= Ris- soina yendoi Yokoyama, 1927), original de- signation. Upper Pleistocene, Koyasu, Japan. 42 PONDER FIG. 12. A-F. /ravadia (Fluviocingula) resima (Laseron), type-species of Mesodestea. A, B, D: Shell; B, holotype (AMS), D, microsculpture of teleoconch. C. Operculum, outer side. E-F. Radula; F, central teeth. A, C-F. Diana Beach, Darwin, Northern Territory, Australia (AMS). С, Н. Iravadia (Fluviocingula) nipponica (Kuroda & Habe), type-species of Fluviocingula; Posyet Bay, U.S.S.R. (AMS) G. Radula, H. Operculum, inner side. Scales: shells = 1 mm except Fig. D scale = 0.1 mm; opercula = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 43 Paronoba Laseron, 1950 (Jan. 27): 283. Type-species: Paronoba subquadrata Laser- on, 1950; original designation. Recent, SE Australia. Dipsotoma Laseron, 1956: 416. Type- species: Rissoa mercurialis Watson, 1886 (= Rissoa bella A. Adams, 1851, ? = Onoba delicata Philippi, 1849); original designation. Recent, tropical Indo-Pacific. Lucidinella Laseron, 1956: 427. Type- species: Lucidinella conicera Laseron, 1956 (= Rissoa (Amphithalamus) densilabrum Melvill, 1912); original designation. Recent, tropical Indo-Pacific. Iragirissoa Dance & Eames, 1966: 39. Type-species: Rissoa (Amphithalamus) aris- taei Melvill, 1912; original designation. Re- cent, India. Diagnosis. Shell. Elongately-ovate to narrowly-elongate, nonumbilicate or with shallow umbilical chink; most species with distinct basal fold; sculptured with weak to moderate spirals and weak axial threads. Aperture not much expanded, with thick to moderate varix; outer lip slightly to mod- erately prosocline, orthocline or weakly opis- thocline, usually with distinct but shallow an- terior excavation and posterior subangulation. Protoconch as for genus. Periostracum thin, often covered with a reddish brown coating. Figs) 1ЗА-Ю PJ: 14A, В, G; 15А В, Е, Н; 16А, В, G. Head-foot. Cephalic tentacles ип- pigmented, with spiral bands of cilia. Posterior end of foot slightly indented, anterior end indented. No metapodial tentacle. (/. (P.) be- lla, Magnetic Island, Queensland; /. (P.) cf. aristaei, Singapore; /. (P.) sp., cf. bella, Singa- pore). Fig. 1H. Penis and oviduct not known. Operculum. Elongately-oval, usually with columellar margin indented towards lower end, nucleus in middle of columellar margin or displaced slightly. A weak, longitudinal in- ternal ridge sometimes present (Several spe- cies examined). Figs. 14C, D, H; 15C, G; 16C, H, |. Radula. Similar to other members of the genus except central teeth with more pronounced lateral extensions in some spe- cies; one to four small to very large denticles on lateral edges of central teeth. Cutting edge of central teeth broad with numerous den- ticles, to very narrow, with only three. Lateral teeth with one to several cusps (several spe- cies examined—see Remarks). Figs. 14E, F, ВЮ Е: 160, Е. Е. Distribution. Tropical Indo-Pacific (В. (A.) aristaei, Rissoa (Scrobs) ictriella Melvill, 1910, = Rissoa (Amphithalamus) alphesiboei Melvill, 1912; Aclis atemeles Melvill, 1896: ? Onoba delicata Philippi, 1849 ? = Rissoa bella A. Adams, 1851 = Rissoa vitrea Garrett. 1873 = Rissoa (Onoba) mercurialis Watson. 1886 = Onoba philippinica Boettger, 1893 = Rissoina oscitans Preston, 1905 = Amphi- thalamus psomus Melvill, 1918; Chevallieria padangensis Thiele, 1925; Rissoa (Amphi- thalamus) densilabrum Melvill, 1912 = Ono- ba quadrasi Boettger, 1893 (secondary homonym of /ravadia quadrasi (Boettger, 1893)) = Lucidinella conicera Laseron, 1956 = Lucidinella conicera patruelis Laseron, 1956; Lucidinella sublaevis Laseron, 1956). Late Miocene, western Pacific (Eniwetok Atoll) (Cingula (Peringiella) parryensis Ladd, 1966). Southeastern Australia (P. sub- quadrata). Japan (R. yendoi = T. (C.) filiola Yokoyama, 1927). New Zealand (Dipsotoma inflata Ponder, 1968). Middle Miocene, Rumania (P. peculiaris). Eocene, France (? Ceratia (?) allixi Cossmann, 1922). Habitat. Marine to estuarine. /. (P.) bella and /. (P.) sp. cf. aristaei were found living in mangroves under objects on the mud or in shallow pools. /ravadia (P.) padangensis, and the three new species described in the Ap- pendix are fully marine and live in relatively deep water. /ravadia (P.) densilabrum was collected alive under coral blocks at low tide in a fully marine situation. Material Examined. P. peculiaris. Lectotype and many paralectotypes (SMF), one lot, ex Cossmann (NHMP). C. (P.) parryensis. Holotype (USNM). О. delicata. One specimen so named (BMNH). O. bella. Two probable syntypes ex Adams (NMW) and several other lots (AMS). Я. vitrea. Five syntypes (ANSP). В. (O.) mercurialis. Holotype (ВММН). О. phi- lippinica. Lectotype and paralectotype (SMF). R. oscitans. Three syntypes (ВММН); two “paratypes” (ANSP). A. psomus. Holotype (BMNH). P. subquadrata. Six syntypes and several other lots (AMS). C. padangensis. Holotype and 11 paratypes (HUM). O. quad- rasi. Lectotype and one other specimen (SMF). R. (A.) densilabrum. Two syntypes (BMNH) and several other lots (AMS). L. con- icera and L. conicera patruelis. Holotypes and paratypes (AMS). L. sublevis. Holotype (AMS). D. inflata. Holotype (AIM). T. (C.) filiola. Holotype (GIT). Я. yendoi. Holotype (GIT) and one other lot (NSMT). A. (A.) aris- taei. Lectotype and two paralectotypes (ВММН). С. (2) allixi. One lot (topotypes), ex Le Renard. Remarks. At least some of the species of 44 PONDER 7», : \ rt E A E 2 = =D == “A Е Я El A PE = AP} Ike Е D 2 > a y = | e) BN E BZ Ke Е 2Я Я Е =; á = А Е FEN / eI Ç ER a y - A = À 5 Я 2 + H A STA E 4 EE = Е Е Er! — Е A E A J) x À + GES N. 2} AN SX FIG. 13. A-D. /ravadia (Pseudonoba) peculiaris (Boettger), type-species of Pseudonoba; topotype, Kostej, Rumania (ex Cossmann Colin.) (МНМР). A, В. Shell, front and side views. С, D. Protoconch. E-F. Rhombostoma imperforatum (Sacco). Ozciano, Italy, (ZMR). E. Protoconch, F. Shell. G, H. Rhombostoma carmelae (Brugnone), syntype (ZMR); type-species of Rhombostoma; G, H. Shell. H, side view of aperture. |, J. Iravadia (Pseudonoba) subquadrata (Laseron), type-species of Paronoba. |, J. Shells; |. holotype. Port Stephens, New South Wales, Australia (AMS). J. Port Hacking, New South Wales (AMS). Scales: shells = 1 mm; protoconchs = 0.1 mm. IRAVADIID GENERA 45 FIG. 14. A-F. Iravadia (Pseudonoba) densilabrum (Melvill), type-species of Lucidinella. A. Shell. Paratype of Lucidinella conicera Laseron, Whitehaven Beach, near Bowen, Queensland, Australia (AMS). B. Pro- toconch. C, D. Operculum, outer (C) and inner (D) sides. E, F. Radula, Е, central teeth. В, С, E. Lindeman |s., Queensland (AMS); D, Е. Norsup, Malekula, New Hebrides (AMS). G-I. /ravadia (Pseudonoba) profundior sp. nov. G. Shell of holotype. H. Operculum (inner side). I. Radula. H and | from paratype. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. 46 PONDER FIG. 15. A-D. /ravadia (Pseudonoba) bella (Adams), type-species of Dipsotoma Laseron. Magnetic Island, Queensland, Australia (AMS). A. Shell. B. Protoconch, C. Operculum, inner side. D. Radula. E. /ravadia (Pseudonoba) filiola (Yokoyama), type-species of Sinusicola, Tomioka, Amakusa, Kyushu, Japan (NSMT). Shell. F-H. /ravadia (Pseudonoba) sp. Sembawang, Singapore (AMS). F. Radula. G. Operculum, inner side. H. Shell. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 47 FIG. 16. A-E. Iravadia (Pseudonoba) gemmata sp. nov. A. Shell of holotype. B. Protoconch. C. Operculum, D, Е. Radula; E, central teeth. B-E, from paratypes. F-I. /ravadia (Pseudonoba) expansilabrum sp. nov. F. Radula. С. Shell of holotype. H, I. Operculum, inner (H) and outer (I) sides. Е, H, |. from paratype. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm: radulae = 0.01 mm. 48 PONDER this subgenus are fully marine and range from shallow coastal waters to relatively deep water. The type-species of Dipsotoma, Sinu- sicola and Iragirissoa probably normally live in estuarine conditions. The shells of species of this subgenus en- compass a considerable range of form. There appears to be a gradation from very elongate shells (as in the type-species of /ragirissoa and Sinusicola (Fig. 15E)) to others with mod- erate spires, the latter usually having much less-impressed sutures and flatter whorls than the former. The spiral ornament usually consists of rather weak spiral threads but some species are sculptured with close spiral lirae. Species included in Lucidinella by Laseron (1956) have shells with prominent axial threads and a rather strong basal fold. The lower part of the inner lip is separated from this basal fold and an umbilical chink is formed between them. This style of shell is very similar to a relatively weakly and spirally sculptured species of /ravadia s.s., |. capitata (Laseron), and to some species of /ravadia (Fairbankia). Paronoba subquadrata (Figs. 131, J) is somewhat intermediate in the basal features of the shell in sometimes having a basal fold and thus forms a link between Lucidinella conicera (= densilabrum) (Fig. 14A), which always has a basal fold and Dipsotoma mercurialis (= bella) (Fig. 15A), which lacks one. This last species resembles the type-species of Pseudonoba (Fig. 13A, B) in all essential shell features. Habe (1958) described the radula and operculum of Sinusicola endoi (sic!, = yen- doi), which agree closely with those of the type-species of Fairbankia, Pseudomerelina and species of Lucidinella, although the struc- ture of the inner side of the operculum was not noted. The shell of this species (Fig. 15E) is smaller than most of the type-species in- cluded in the synonymy of Pseudonoba but agrees generally with them in shape and sculpture. It is also similar to species in the subgenus Fairbankia in shape but is smaller in size and has an orthocline, not prosocline, outer lip. A very similar species, Rissoa aris- taei Melvill, from Bombay, is the type-species of /ragirissoa. Specimens from Singapore of a possibly undescribed species similar to both R. aristaei and T. filiola (Fig. 15H) differ in the characters of the central teeth of the radula from the figure of the radula of S. yendoi (= T. filiola) (Habe, 1958), but have an operculum like that of species of Fairbankia and Pseudo- noba (Fig. 15G). These species appear to fall between Fairbankia and Dipsotoma in shell features but are here included somewhat ten- tatively in Pseudonoba. The common shell features outlined in the generic diagnosis are a combination of characters which tie the type-species of the genera listed in synonymy together. Further support for their close relationship is gained by the very similar oper- cular characters seen in the species ex- amined, which encompass the majority of the observed shell variations within the sub- genus. Unfortunately, the radular and opercular characters of some of the type-species of the genus-group names included in the syn- опуту are not known, so this grouping must be considered somewhat tentative. The type- species of Pseudonoba is a Miocene fossil but the radula and operculum of a closely similar new species (see Appendix) have been examined (Fig. 14G-I). Paronoba sub- quadrata is also known only from its shell which is, as noted above, similar to P. pecu- liaris and also to Dipsotoma mercurialis in all essential features. The shell of the type- species of /ragirissoa is similar but more elon- gate than the shells of the type-species of Pseudonoba and Dipsotoma. It also has a rather weak basal fold and weak spiral sculp- ture. Pseudonoba is probably more closely re- lated to Fairbankia than to Iravadia, as far as can be judged from the shell, radula and operculum. The probable greater antiquity and mainly marine habitat of species of Pseudonoba suggest that it is the ancestral group within the genus /ravadia. There is very considerable variation in the radular teeth, particularly in the central teeth, in the species examined in this group. This may indicate that the group is polyphyletic but it is also possible that because it is of some antiquity and considerable geographic spread this is divergence within a single phyletic group. The ‘normal’ /ravadia radula is seen in some species (Figs. 14E, Е, I; 15D) but /. (P.) cf. aristaei from Singapore has central teeth with well-defined, long, lateral margins that show some thickening and there is a promi- nent basal denticle on each side of each tooth (Fig. 15F). This type of central tooth is very similar to that seen in the genera Liroceratia nov., Hyala and Nozeba and it is similar to the central teeth of many rissoids and hydrobiids. It is thus probable that this type of tooth can be regarded as “primitive” and that the “typic- al” /ravadia central tooth is derived from it by a IRAVADIID GENERA 49 loss of lateral thickening, a widening of the cutting edge and a reduction in the size of the basal denticles. This condition can be seen in |. (P.) yendoi (see Habe, 1958), a species almost indistinguishable from /. (P.) aristaei, in |. (P.) densilabrum (Fig. 14E, F) and in /. (P.) bella (Fig. 15D). A deep-water species from the Indian Ocean is closer to the type- species of Pseudonoba in shell features than I. (P.) bella and has a radula (Fig. 16F) in which the lateral edges of the central teeth are thickened and the cutting edge is narrow, giving the tooth a sub-triangular outline. There is a prominent basal denticle on the central tooth of this species but in a new species from the Philippines there are three to four large basal denticles (Fig. 16D, E). Both of these species are described in the Appen- dix. /. (Р.) densilabrum has one to three rudimentary denticles, so that multiple basal denticles are not unique in the subgenus. The radula of two of the new species is also unusual in having a very prominent cusp on the lateral teeth and zero or one secondary cusp. In view of the general similarity in the shell and opercular features, and the existence of intermediate radular types, the species are tentatively grouped into a single subgenus, Pseudonoba. It is probable, however, that when additional information is available, fur- ther division will be required. Genus Chevallieria Cossmann, 1888: 244 Type-species: Chevallieria labrosa Coss- mann, 1888; original designation. Eocene, Paris Basin, France. Synonym: Nanadoma Laseron, 1956: 447. Type-species: Nanadoma imitoris Laseron, 1956; original designation. Recent, N Au- Stralia. Diagnosis. Shell. Subcylindrical, thin, non- umbilicate, with convex whorls, basal fold weak to absent, smooth or with extremely fine spiral striae and, in some species, distinct axial growth lines. Aperture pyriform, angled and weakly channelled posteriorly, rounded to weakly angled anteriorly, outer lip ortho- cline to slightly opisthocline, varix strong to absent. Protoconch relatively large, flattened, smooth, of about 1%4-2 whorls. Fig. 17A-D, G, H. Head-foot, penis and oviduct un- known. Operculum. Elongate, with spiral form apparent; columellar edge slightly in- dented, transparent, a weak internal ridge along columellar edge (С. australis sp. nov.). Fig. 17E. Radula. Central teeth rather large, each with four small cusps on either side of a small median cusp; lateral margins un- thickened, with a small denticle; basal margin with tongue-like projection. Lateral teeth with small, sharp cusps 5 + 1 + (?6); marginal teeth with numerous small, sharp cusps (C. australis sp. nov). Fig. 17F. Distribution. Tropical Indo-Pacific (N. im- itoris; Rissoina columen Melvill, 1904 = Ris- soina (Scrobs) elspethae Melvill, 1910). New species from South Australia and the Miocene and Pliocene of Victoria, Australia are described in the Appendix. Eocene, Paris Basin (C. labrosa and Chevallieria cylin- droides Cossmann, 1907). There have been several additional species attributed to Che- vallieria from the Eocene of the Paris Basin, but these species have not been examined. Habitat. The only species found alive (see Appendix) was collected under stones in a marine situation. Chevallieria imitoris also probably lives in a fully marine habitat. Material Examined. C. labrosa. One speci- men ex Cossmann colin. (NHMP), three specimens ex J. Le Renard (AMS). N. im- itoris. Holotype, two paratypes and several other lots (AMS). R. columen. Holotype (ВММН). R. (S.) elspethae. Three syntypes (BMNH). C. cylindroides. One lot (AMS). Remarks. The shell of the type-species of Chevallieria (Fig. 17A) differs from that of the type-species of Nanadoma (Fig. 17G) in hav- ing a weak basal fold and a strong varix on the outer lip. Another congeneric species found in the Eocene of the Paris Basin, C. cylindroides, lacks a basal fold and agrees extremely closely with a Miocene species from Victoria, Australia (Fig. 17B), which is described in the Appendix. This Australian species differs from C. labrosa mainly in hav- ing a weaker varix on the outer lip. The Re- cent C. imitoris has a very weak to absent varix. Chevallieria labrosa closely resembles some species of /ravadia (Pseudonoba) in shell features and this genus is probably the group from which Pseudonoba evolved. Species of Chevallieria differ from those of Pseudonoba by their smaller, thinner shells and extremely delicate spiral sculpture, or smooth surface. The only operculum ex- amined is similar to that of species of /ravadia (Pseudonoba) in shape but is more nearly spiral in construction, the nucleus being placed away from the margin (Fig. 17E). This type of operculum is intermediate between 50 PONDER FIG. 17. A. Chevallieria labrosa Cossmann, type-species of Chevallieria; La Ferme de L'Orme, Yvelines, France (AMS). Shell. B. Chevallieria balcombensis sp. nov. Shell of holotype. C. Chevallieria gippslandica sp. nov. Shell of holotype. D-F Chevallieria australis sp. nov. Holotype. D. Shell. E. Operculum, inner side. F. Radula. G-H. Chevallieria imitoris (Laseron), type-species of Nanadoma, Shoal Bay, Mackay, Queensland, Australia (AMS). G. Shell. H. Protoconch. I-M. Liroceratia sulcata (Boettger), type-species of Liroceratia nov., Taurama, near Port Moresby, Papua, New Guinea (AMS). I. Shell. J. Protoconch. К, L. Operculum, outer side (К) and inner side (|). M. Radula. Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. IRAVADIID GENERA 51 the normal spiral operculum seen in Hyala, Nozeba and Liroceratia and the typical /rava- dia operculum. It is closer in form to the operculum of species of /ravadia (Pseudono- ba) than to any of the other subgenera of lravadia. Genus Rhombostoma Seguenza, 1876: 14 Type-species: Eulima carmelae Brugnone, 1873; subsequent designation Sacco, 1892: 19. Pliocene, Sicily. Synonym: Eulimopsis Brugnone, 1881: 120. Type-species: Eulima carmelae Brug- none, 1873; monotypy. Pliocene, Sicily. Diagnosis. Shell. Small, elongate-conic, smooth or spirally sculptured, with weakly- convex whorls. Aperture elongately-ovate, angled and weakly channelled posteriorly; distinctly and rather deeply-channelled an- teriorly; outer lip sinuate, with middle and uppermost (adapical) section advanced (see Fig. 13H). Inner lip thin, narrow; columella vertical, narrow. No external varix. Pro- toconch relatively small, of about two whorls, the first 1% whorls flat, the last half whorl rapidly descending. Figs. 13E—H. Animal un- known. Distribution. Pliocene and Miocene of Italy (E. carmelae; Ondina imperforata Sacco, 1892 = Ondina pliobliqua Sacco, 1892 = striata auct.). ?Miocene of Austria (Chemnit- па striata Hörnes, 1856). Material Examined. E. carmelae. Syntypes (ZMR). O. pliobliqua. Sacco material (TGM), six specimens (ZMR). C. striata. Photograph of holotype ex A. Warén. Remarks. Pavia (1975) has discussed this genus and describes in some detail one of the fossil species it contains. The shell of these species has a protoconch typical of the Irava- diidae (Fig. 13E) and a broad anterior notch in the aperture similar to that seen in some other genera in the family. The spiral cords of R. imperforata (Fig. 13E) also suggest a relation- ship with /ravadia. The weak varix and rather narrow aperture combined with the solid, tall- spired shell set the species included in this genus somewhat apart from the other genera in the family, although the overall apertural features resemble those of /ravadia (Fair- bankia) sakaguchii. Chemnitzia striata Hórnes possibly belongs to this group although its aperture shows some similarity to that of species of /ravadia (Pseudonoba) in being less distinctly angled anteriorly. Probably this group represents an off-shoot from an early Pseudonoba lineage but, be- cause of its rather distinctive shell characters and the impossibility of confirming a close relationship with /ravadia, it is tentatively sep- arated. Wenz (1940) and Pavia (1975) in- cluded this genus in the Eulimidae. Genus Liroceratia Ponder, gen. nov. Type-species: Cingula sulcata Boettger, 1893. Recent, Philippines. Diagnosis. Shell. Small, solid, elongately- ovate, sculptured with strong spiral cords. Aperture oval, subangled posteriorly, rounded anteriorly, with weak, broad, anterior ex- cavation; outer lip orthocline, with strong, broad varix. Protoconch similar to that of /ra- vadia. Periostracum yellow to brown, con- spicuous. Fig. 171, J. Head-foot. Cephalic tentacles long, strap-like, unpigmented, with stiff ‘setae’ distally; eyes at outer bases in small bulges. Snout of moderate length, bilobed. Foot weakly-cleft anteriorly, very weakly-indented posteriorly, anterior mucous gland indistinct. No posterior pedal mucous gland, no pallial tentacles and no metapodial tentacle. (L. sulcata, Taurama, near Port Moresby, Papua New Guinea). Fig. 1G. Penis and oviduct unknown. Operculum. Oval, coiled, thin, simple, nucleus eccentric, last whorl large. Fig. 17K, L. Radula. Very similar to that of Hyala vitrea, but with cutting edge of each lateral tooth raised on a neck- like extension from the base of the tooth. Fig. 17M. Distribution. Tropical central Indo-Pacific from the Philippines to Fiji and Papua New Guinea (C. sulcata, = Pellamora minatura Laseron, 1956). Rissoa truncata Garrett, 1873 from Fiji is probably related. Habitat. The seaward edge of mangroves under stones in the mid-littoral. Material Examined. C. sulcata. Holotype (SMF), one lot Philippines and several other lots (AMS). P. minatura. Holotype and para- types (AMS). Я. truncata. Three syntypes (ANSP). Remarks. This genus is based on a species which has a shell very like a miniature /rava- dia, particularly in its protoconch characters and in having strong spiral cords. Its radula and operculum, however, resemble those of Hyala vitrea. The shell also superficially re- sembles that of marine species placed in rissoid genera such as Lironoba Iredale, 1915 52 PONDER but is easily distinguished by its protoconch characters (Fig. 17J), the rissoids having dome-shaped, sculptured protoconchs. Species of Chevallieria are very similar to Liroceratia sulcata in their general shell fea- tures. They differ in their weak to absent spiral sculpture and in the details of the radular and opercular features of the one species for which these characters are known. An Eocene species, Ceratia (?) allixi Coss- mann, tentatively included in /ravadia (Pseudonoba) above, has similarities with this genus in most shell characters. Genus Hyala H. & A. Adams, 1852: 359 Type-species: Hyala vitrea (= Turbo vit- reus Montagu, 1803); monotypy. Recent, Eu- rope. Diagnosis. Shell. Small, thin, smooth, or with microscopic spiral threads. Aperture sim- ple, oval, outer lip strongly prosocline, lacking a varix; weak posterior angulation present but no posterior sinus; shallowly but broadly ex- cavated anteriorly. Protoconch smooth, of 25 whorls, lacking a distinct terminal varix, first whorl rising slightly above level of nucleus. Fig. 18E-G. Head-foot. Animal unpigmented. Cephalic tentacles strap-like, long, having six to eight stationary cilia proximally, with eyes in the centre of their bases. Snout rather long, bilobed. Foot indented in front, rounded post- eriorly. No accessory tentacles (H. vitrea, Clark, 1852 and A. Waren in litt., 1981). Fret- ter & Graham (1978), apparently incorrectly, state that the posterior end of the foot is bifid. Penis. U-shaped when at rest, with short filament distally and small glandular bulge about two-thirds of length from base. Fig. 8E. Oviduct. With large anterior sperm sac, no bursa copulatrix and posterior oviducal open- ing. There are either one or two seminal re- ceptacles. Fig. 3 and Johansson (1950). Operculum. Oval, thin, simple, spiral, with eccentric nucleus, last whorl very large (H. vitrea). Fig. 18H. Radula. Lateral and margin- al teeth typical of family, central teeth with cutting edge relatively narrower than in /rava- dia, cusps few, sharp, median cusp rather long; lateral margins spread outwards, slightly thickened; each tooth with a denticle on each side of face (H. vitrea). Fig. 181, J. Distribution. Coast of Europe and Mediter- ranean Sea (H. vitrea = H. mediterranea Nordsieck, 1972). Sea of Japan (? H. adamsi Golikov & Kussakin, 1971). Upper Oligocene, North Sea Basin (Rissoa dissoluta Wiech- mann, 1874). Habitat. Marine, sublittoral and on the con- tinental shelf (15-100 m in deep burrows made by other animals on muddy bottoms (A. Josefson, personal communication, fide A. Waren in litt.) (H. vitrea). Hyala adamsi lives on Zostera in an enclosed bay with fluctuating salinity (Golikov & Kussakin, 1978). Material Examined. H. vitrea. Several lots (BMNH, AMS and other museums). Я. dis- soluta. One lot (NMV). Remarks. Thorson (1946) described a free- swimming larva from Denmark which he iden- tified as “Onoba vitrea.” It had fine spiral striae on the only whorl of the teleoconch that had developed. He surmised that the smooth shell of specimens of Hyala vitrea was due to wear and that his larva belonged to that spe- cies. The shells examined in this study show no trace of spiral striae but Jeffreys (1867) and A. Waren (in litt.) state that very fine microscopic spiral striae are present in per- fect specimens. The type-species of the genus has a thin, smooth shell lacking a varix on the outer lip and has a paucispiral operculum. The some- what flattened protoconch is similar to that of Iravadia and Hyala and also lacks a posteri- or mucous gland and accessory tentacles. In addition, the female reproductive system (Johansson, 1950) has a posterior slit in the glandular pallial duct, an anterior sperm sac, and lacks both a separate upper oviduct gland and a coiled oviduct, features also observed in /ravadia (herein). The apertural features and general texture of the shell recall those of Nozeba and it is probable that these two genera are closely related. Hyala vitrea, however, differs from species of Nozeba in having a narrower shell, a strongly prosocline (not orthocline) outer lip and a weaker posterior angulation of the aperture. Hyala adamsi has a much broader shell than H. vitrea. As | have not been able to examine specimens, the generic location of this species has not been confirmed. If it is a Hyala, a member of this genus can survive in waters of low salinity. Genus Ceratia H. & A. Adams, 1852: 359 Type-species: Rissoa proxima (Alder MS) Forbes & Hanley, 1850; monotypy. Recent, Europe. IRAVADIID GENERA 53 FIG. 18. A-C. Ceratia proxima (Forbes & Hanley), type-species of Ceratia. A. Protoconch. B. Shell. C. Operculum, inner side. A, B. Torbay, Devon, England (BMNH), C. Bay of Biscay, France (AMS). D. cer alta (Gabb), type-species of Hebetaclis; holotype (ANSP). E-J. Hyala vitrea (Montagu), type-species о Hyala. E, F. Shell, Vigo Bay, Spain, 36 m (AMS). G. Protoconch, H. Operculum, outer side. |, J. Radula, J, central teeth. G-J. Horns Reef, Danish North Sea coast, 50 т (ZMC). Scales: shells = 1 mm; opercula and protoconchs = 0.1 mm; radulae = 0.01 mm. 54 PONDER Synonym: Hebetaclis Pilsbry, 1922: 389. Type-species: Auriculina alta Gabb, 1873; original designation. “Miocene,” Dominican Republic. Diagnosis. Shell. Very similar to that of Hyala but with fine spiral striae. Protoconch of about two whorls, first whorl slightly elevated, nucleus depressed, similar to that of Hyala; smooth (except for “some fine spiral lines” according to Fretter & Graham, 1978). Fig. 18A, B, D. Head-foot. Pigmented white. Cephalic tentacles flat, rather short, smooth, gently attenuating and becoming minutely claviform at distal ends, which have a few stationary ‘cilia’; eyes large, on minute swell- ings at outer bases of tentacles. Snout short but extendible, not bilobed. Foot large, fleshy, anterior edge deeply indented and produced into two long, lateral processes. Posterior end of foot divided into two long, wiaely divergent tails. No accessory tentacles (Clark, 1852). Penis and oviduct unknown. Operculum. As in H. vitrea (C. proxima). Fig. 18C. Radula. The only radula of C. proxima available was accidentally mounted upside down but the shape of the central tooth is more nearly square than in that of H. vitrea. Lateral teeth finely cuspate. Distribution. Southern British Isles to the Mediterranean Sea (Rissoa striatula Jeffreys, 1847 (preocc.) = C. proxima.). Ceratia prox- ima has been recorded from the Pliocene of England (Jeffreys, 1867) and the Pliocene and Pleistocene of Italy (Pavia, 1975). Miocene, Dominican Republic (A. alta). Cera- tia minutissima Cossmann, 1888 from the Eocene of the Paris Basin, is possibly con- generic but no specimens have been avail- able for examination. Habitat. Marine, subtidal and on the con- tinental shelf (C. proxima). Material Examined. R. striatula. Neotype (USNM). C. proxima. A few lots (BMNH, USNM). A. alta. Holotype and paratypes (ANSP). Remarks. Ceratia is only tentatively re- garded as a genus distinct from Hyala on the basis of the differences in the shape of the posterior end of the foot. Clark's (1852) de- scription of the animal of C. proxima is based on the careful examination of at least two specimens. He observed this species together with Hyala vitrea and commented at length as to the distinctiveness of the animals. Certainly the deeply cleft posterior end of the foot of C. proxima appears to be unique and has been confirmed by Fretter & Graham (1978). Most species of /ravadia do, however, have the posterior end of the foot weakly- indented. Only one specimen containing a dried animal was available for examination and the mount of the radula was unsuccess- ful. The resuscitated dried remains did show that the posterior end of the foot was deeply cleft and that the operculum is identical to that of H. vitrea. Thus the familial position of this species is based mainly on shell features (which are almost identical to those of Hyala) and on the lack of acccessory tentacles. Dr. А. Waren (in litt.) first examined the type of A. alta (Fig. 18D) and suggested its rela- tionships. An examination of the type-material confirmed his assessment that Hebetaclis is a synonym of Ceratia as far as it is possible to judge from shell characters. Genus Nozeba Iredale, 1915: 453 Type-species: Rissoa emarginata Hutton, 1885; original designation. Plio-Pleistocene and Recent, New Zealand. = Neozeba, err. auct. [? — Pasithea Lea, 1833: 99, 207. Type- species: Pasithea claibornensis Lea, 1833; subsequent designation Gray, 1847: 160. Eocene, Alabama, North America. Synonym: ?Pasitheola Cossmann, 1896: 26. Un- necessary replacement name for Pasithea Lea, 1833, non Pasythea Lamouroux, 1812]. Synonyms: Antinodulus Cossmann, in Cossmann & Peyrot, 1919: 568. Type- species: Bulimus globulus Grateloup, 1827; Original designation. Lower Miocene, France. Syntharella Laseron, 1955: 100. Type- species: Eulima topaziaca Hedley, 1908; ori- ginal designation. Recent, SE Australia. Diagnosis. Shell. Ovate to elongately- conical, surface glossy, non-umbilicate, rather solid, smooth or with spiral threads, sometimes spirals moderately strong on base. Aperture rounded to distinctly-excavat- ed anteriorly, sharply-angled posteriorly but not channelled. Posterior corner of aperture turned forward slightly, rest of outer lip ap- proximately orthocline; inner lip partially dis- connected from parietal wall in some species. No external varix. Protoconch smooth, of 2— 2% convex whorls, flattened on top, last whorl descending, apex partially immersed. Per- iostracum thin. Fig. 19A, C-E, |. Head-foot. Cephalic tentacles long, strap-like, non- pigmented, with the eyes in small bulges at their outer bases; a few stiff “setae” distally. IRAVADIID GENERA 55 FIG. 19. A. Nozeba emarginata (Hutton), type-species of Nozeba. Outer Bay of Islands, New Zealand, 80 m (NMNZ). Shell. B. Pasithea claibornensis Lea, type-species of Pasithea. Shell of holotype. C. Nozeba globulus (Grateloup), type-species of Antinodulus; St. Paul-les-Dax, France (AMS). Shell. D-H. Nozeba topaziaca (Hedley), type-species of Syntharella. D. Shell. E. Protoconch. F, G. Operculum, inner (F) and outer (G) sides. H. Radula. D, F-H, Fisherman’s Bay, Port Hacking, New South Wales, Australia (AMS); Е The Spit, Sydney, New South Wales (AMS). |. Nozeba дийша (Lea). Claiborne Sand Bed, Claiborne, Alabama, U.S.A. (USNM). Shell (subadult). Scales: shells = 1 тт; opercula and protoconch = 0.1 тт; radula = 0.01 mm. 56 PONDER Snout moderately long, extensile, bilobed. Foot slightly pigmented dorsally, with weakly- cleft anterior edge onto which opens a con- spicuous, triangular anterior mucous gland; no posterior mucous gland present. Posterior pallial tentacle very short; no anterior pallial tentacle or metapodial tentacle. Postero- dorsal side of foot simple. (N. topaziaca, Port Hacking, New South Wales). Fig. 11. Penis. When at rest, bent double behind right cephalic tentacle, slightly to right of mid-line of head; wide, rather short, with enclosed duct opening at distal end, with three glandular swellings on inner edge. Fig. 8F. Oviduct. As for Hyala. Operculum. Oval, thin, simple, spi- ral, nucleus eccentric, last whorl large (N. emarginata Ponder, 1967; N. mica, N. topazi- aca). Fig. 19F, G. Radula. Central teeth re- latively large, cutting edge about equal to Ya total width, cusps small; lateral edges at about 45° to cutting edge, thickened, free from rest of base for most of their length. A pair of prominent denticles on face of each central tooth overlap tooth in front; a tongue- like extension of the base between. Lateral teeth elongate, cutting edge rather long, with many small cusps; a small protuberance on face of tooth below cutting edge. Marginal teeth long, curved, with many small cusps (N. mica Ponder, 1967; N. topaziaca). Fig. 19H. Distribution. New Zealand, Recent (Я. emarginata = Rissoina coulthardi Webster, 1908; Nozeba mica Finlay, 1930), Tertiary (Pliocene-Miocene) (four species—see Flem- ing, 1966). Temperate Australia (Eulima to- paziaca Hedley, 1908 = Estea amblycory- mba Cotton, 1944); Lower Miocene, Victoria (Rissoa gatliffiana Chapman & Gabriel, 1914). Eocene of France (Amphimelania luci- da Cossmann, 1886; Balanocochlis eu- limoides Cossmann, 1888). Eocene of Ala- bama, North America (Pasithea guttula Lea, 1833). Miocene of France (B. globulus). A new species from the Philippine Islands is tentatively assigned to this genus and is described in the Appendix. Undescribed species have been seen from the Eocene of New Zealand and Upper Cretaceous (Ripley Formation) of the U.S.A. Habitat. Marine, on the continental shelf (N. mica, N. emarginata) and living on sea grasses in estuaries and embayments in the lower littoral and sublittoral (N. topaziaca). Material Examined. P. claibornensis. Holotype (ANSP). R. emarginata. A few lots (NMNZ, AMS). N. mica. A few lots (NMNZ). Е. topaziaca. Holotype, paratypes and many other lots (AMS). E. amblycorymba. Holotype and other material identified by Cotton (SAM). В. gatliffiana. Three paratypes (NMV). A. /uci- da and B. eulimoides. One lot of each spe- cies, ex J. Le Renard (AMS). P. guttula. Holotype (ANSP) and two other lots (ANSP, USNM). B. globulus. One lot ex J. Le Renard (AMS). Remarks. This genus has a long Tertiary history. The two Recent New Zealand species are marine but N. topaziaca from Australia lives in estuaries. A new species (See Appen- dix) tentatively referred to Nozeba was found in deep water in the Philippines. (Fig. 20G—J). Pasithea may possibly be an earlier name for Nozeba but, unfortunately, that genus name is based on a very poorly preserved specimen which, according to Palmer (1937), may be а gerontic form of P. guttula (Fig. 191). The type specimen (Fig. 19B) is, however, much larger than available material of P. gut- tula and may not be congeneric or even con- familial. Pasithea guttula is, together with some European Eocene species, included in Pasitheola by Cossmann (1921) and these certainly appear to be congeneric with Noze- ba. An undescribed Eocene species from New Zealand is also a Nozeba. Gougerot & Le Renard (1977) have revised the Eocene species from the Paris Basin. Several speci- mens in the USNM from the Upper Cretaceous (Ripley Formation, SE U.S.A., USGS 25923, 28440, 27924) appear to be an undescribed species of Nozeba. Cossmann, in several publications, consid- ered the species he included in Pasitheola to be members of the Thiaridae. Other authors have placed this genus in the Eulimidae. Nozeba topaziaca (Fig. 19E-H), the type- species of Syntharella Laseron, agrees with Nozeba emarginata in essential shell charac- ters and can be included in the genus Nozeba. Syntharella was regarded as a genus in the Eulimidae by Laseron (1955). Genus Rissopsis Garrett, 1873: 228 Type-species: Я. typica Garrett, 1873; monotypy. Recent, Fiji. Diagnosis. Shell. Of moderate size, sub- cylindrical, thin, translucent, surface smooth and glossy, with flat whorls and pyriform aper- ture. Inner lip rather broad, thin; outer lip expanded slightly, orthocline to prosocline, aperture broadly excavated anteriorly; varix absent. Protoconch of about two whorls, very IRAVADIID GENERA 57 FIG. 20. A-D. Acliceratia beddomei (Dautzenberg), type-species of Acliceratia nov.; off Abidjan, Ivory Coast (NHMP). A. Shell. В. Operculum, inner side. С, D. Radula, D, central teeth. E-F. Acliceratia carinata (Smith); off Abidjan, Ivory Coast (NHMP). E. Shell. Е. Operculum, inner side. G-J. Nozeba (2) striata sp. nov. holotype. G. Shell. H-I. Radula, H, central teeth. J. Operculum, inner side. Scales: shells = 1 mm; opercula = 0.1 mm; radulae = 0.01 mm. 58 PONDER small, tightly coiled, planorboid. (See Ponder, 1974, fig. 1). Animal unknown. Distribution. Fiji, Samoa, Marshall Is. (В. typica); Durban, South Africa (Rissopsis tuba Kilburn, 1977; Fusus prolongata Turton, 1932 ? = Rissopsis ligula Kilburn, 1975). Northern Australia (one undescribed species). Philip- pines (one undescribed species). Material Examined. R. typica. Lectotype and paralectotype (ANSP). One other speci- men (SAM). A. tuba. Holotype and six para- types (NM). A. Гоша. Holotype and three paratypes (NM). Habitat. Unknown. Remarks. The type-species and the genus have been discussed by Ponder (1974). Ris- sopsis appears to be related to Chevallieria or, possibly, Hyala, judging from shell char- acters. Species of Rissopsis can be dis- tinguished from those of Chevallieria by their large, smooth, glossy shell, the broadly- excavated anterior edge of the aperture and the subcylindrical spire. Rissopsis species resemble Hyala vitrea in having a smooth shell of similar shape, and, in A. tuba, a strongly prosocline outer lip. A more detailed assessment of the relationships of this genus must, however, await the examination of at least the radula and operculum. Of the two species of Rissopsis from South Africa, one is distinct but the other (Я. ligula) is very similar to R. typica, differing only in its slighty larger size. Single specimens of two undescribed species are known from northern Australia (AMS, C.126209) and the Philip- pines (USNM, 281302). Genus Acliceratia Ponder, gen. nov. Type-species: Aclis beddomei Dautzen- berg, 1912. Recent, West Africa. Diagnosis. Shell. Large for family, elongate-conic, thin, with spiral sculpture and a peripheral angulation or keel. Aperture an- gled posteriorly, rounded anteriorly; outer lip thin, lacking a varix, prosocline, with no pos- terior sinus; inner lip a thin, narrow glaze on parietal wall. Protoconch of about two whorls, typical of family. Fig. 20A, E. Head-foot. Un- known. Faecal pellets in rectum aligned per- pendicular to rectal wall as is typical of family (A. carinata). Penis and oviduct unknown. Operculum. Thin, oval, simple, spiral, with eccentric nucleus, last whorl very large (A. beddomei, A. carinata). Fig. 20B, Е. Radula. (21 FE): — Е primary cusp large, sharp; secondary cusps small; basal cusps prominent, sharp; outer- most cusp on outer edge of tooth. Lateral teeth large, with long bases and prominent, pointed primary cusp; secondary cusps rather small, (3)2 + 1 + 5. Marginal teeth with sharp, rather large cusps, the largest about equal in size to primary cusps of central and lateral teeth (A. beddomei, A. carinata). Fig. 20С, D. Distribution. West Africa (A. beddomei; Aclis carinata Smith, 1871). Paleocene, France (one undescribed species). Habitat. On the continental shelf. Material Examined. A beddomei. One lot (identified Dautzenberg) and one other lot (NHMP). A carinata. Two specimens Daut- zenberg Colln.; IRSNB; one lot (NHMP). A. п. sp. One lot, ex Le Renard. Remarks. Although the two Recent species of Acliceratia agree in opercular and apertural characters with Ceratia and Hyala, they differ in having a much larger shell (9-12 mm in length compared with about 3 mm for Hyala vitrea and Ceratia proxima) which bears a distinct peripheral ridge. The details of the central teeth of the radula of C. proxima are unknown but the very poor available mount shows them to be sub-rectangular, not sub- triangular as they are in Acliceratia species. The central teeth of H. vitrea are more similar in shape to those of Acliceratia but have only one pair of basal denticles. The two species of Acliceratia resemble in size and general shell characters some species included here in /ravadia (Pseudonoba) but they differ from all of the subgenera of /ravadia in having a simple, coiled operculum. The radula charac- ters are similar to at least one species of Iravadia (Pseudonoba) in having multiple basal denticles on the central teeth. This character separates Acliceratia from the other genera (excluding Ceratia in which the details of the central teeth are not known) possess- ing a simple operculum. The shells of the two Recent species (Figs. 20A, E) in this genus differ in size and in details of sculpture. They occur sympatrically and have almost identical radulae. It is possi- ble that they represent morphs (sexual?) of a single species. An undescribed Paleocene species from France came from shallow water deposits at Bachivillers, Oise. It agrees closely with the Recent species in shell characters. Central teeth sub-triangular, IRAVADIID GENERA 59 cg ag sr FIG. 21. Diagrams showing the types of female genitalia in the Iravadiidae and generalized examples of the Hydrobiidae and Rissoidae. The arrows indicate possible directions of evolutionary change in the genitalia and are not intended to indicate phylogenetic relationships. A. Hypothetical ancestral condition. B. Hyala vitrea. С. Iravadia (lravadia) quadrasi. D. Iravadia (Fairbankia) australis. E. Iravadıa (Fairbankia) bom- bayana. Е. Iravadia (Pseudomerelina) mahimensis. G. Iravadia (Iravadia) ornata. H. Generalized Hydro- biidae. |. Generalized Rissoidae. 60 PONDER DISCUSSION Evolution of the lravadiidae Female Genitalia. A hypothetical evolution- ary scheme of the female genital system is schematically shown in Fig. 21. Here it is assumed, following Johansson (1968), that the ancestral condition was an open pallial oviduct (Fig. 21A). Closure of the originally open capsule gland from behind forwards, enclosing the sperm groove within the ventral channel, apparently occurred in the Hydro- biidae (Johansson, 1948) (Fig. 21H). It appears as though, in the Iravadiidae, closure of the ventral opening occurred differently, beginning anteriorly and moving backwards. This resulted in a slit-like opening in the pos- terior half of the capsule gland as seen in Hyala vitrea amd Nozeba topaziaca (Fig. 21B). Further evolution appears to have re- sulted in the opening becoming secondarily anterior (as in /ravadia ornata) (Fig. 21G) or even, apparently secondarily, open over much of its length (as in /ravadia (Fairbankia) bombayana) (Fig. 21E). A sperm sac (ass) developed anterior to the opening by the sperm groove, closing over to become a tubu- lar structure. Posteriorly, in /ravadia quadrasi (Fig. 21C), the deep, muscular sperm groove is Open throughout the posterior part of the oviducal opening and contains unorientated sperm. Just posterior to the oviducal opening the sperm groove closes over, becoming a blind sac (bc) which extends to the posterior end of the pallial cavity. This sac, or bursa copulatrix, develops in a similar way in the other species of /ravadia investigated, although its relative position and its opening vary considerably. In the subgenus Fairbank- га the bursa and its opening have swung dorsally to form a lateral pocket-like structure with a separate pallial opening. /ravadia (Fair- bankia) australis (Fig. 21D) retains a rudimentary anterior sperm sac but this is lost in I. (F.) bombayana (Fig. 21E), in which the anterior two-thirds of the capsule gland is open ventrally. In /ravadia (Pseudomerelina) mahimensis (Fig. 21F), the bursal opening is enclosed immediately behind the oviducal opening. It appears as though the bursa (bc) in this species has migrated dorsally and the opening of the bursal duct has retained its original position. The bursal duct is developed from part of the sperm groove, whereas the bursa itself is probably an outgrowth from this groove. The bursal sac has migrated back behind the pallial cavity and lies on the outer side of the glandular oviduct in /ravadia orna- ta (Fig. 21G). In that species the anterior sperm sac is apparently reduced to a small vestibule-like structure (although the sperm storing function of this area was not соп- firmed) and the bursal duct runs almost ver- tically from just behind the short, subterminal genital opening. Operculum. The peculiar operculum of /ra- vadia (s.l.) is a considerable departure from that of the normal, coiled, paucispiral op- erculum seen in the other genera of the Irava- diidae and in most Rissoidae and Hydro- biidae. The modified operculum may have arisen in a Chevallieria ancestor in the early Tertiary as discussed above (see Remarks under /ravadia and Chevallieria, and Fig. 22). It is assumed that this type of operculum has arisen only once and that /ravadia, as here recognised, is a monophyletic group (Fig. 23). Radula. The assumption that the marine genera are the most primitive is supported by species of Hyala and Nozeba having a single pair of well-developed denticles on the face of the central teeth of their radulae, this being the normal rissoacean condition. This charac- ter is shared by the estuarine genus Lirocera- tia and these three genera are similar in hav- ing a non-pigmented animal and a normal, spirally-coiled operculum. An increase in the number of basal denticles on the central teeth of the radula may have occurred more than once, several pairs being found in Acliceratia and some species of /ravadia (s.l.). The reduction and, sometimes, eventual loss of denticles on the face of the central teeth of the radula, due to their lateral migration, appears to be a common trend in the genus /ravadia ($.1.). Habitat. Species of Chevallieria appear to be marine and the genus, like Nozeba, has an ancestry extending back at least to the Eocene. Chevallieria is probably ancestral to Iravadia (Pseudonoba), which is recorded from the Miocene. Some species of the Pseudonoba group migrated into brackish water and appear to have given rise to the other subgenera of /ravadia, all of which are confined to waters of reduced salinity. Other incursions into brackish water have been made by at least one species in both Nozeba and Liroceratia (Fig. 22) and, possibly, Hyala. This propensity for the iravadiids to move into low salinity areas might suggest that the an- cestral group inhabited sheltered bays and estuaries and that the marine species in the family are derived from these. A detailed an- IRAVADIID GENERA 61 Rissopsis /s pa RECENT É Chevallieria lravadia ee le . . ...... wy Acliceratia she O ICO lO eee в 1 TH LATERAL NUCLEUS 0 00 OIDO OO ооо еее ве ele ele TERTIARY CRETACEOUS Ze Liroceratia Ceratia FIG. 22. A speculative interpretation of the evolution of genera of the lravadiidae through time and their incursions into brackish water. The diagram is intended to indicate possible relationships and display shell diversity and the differences in the structure of the central teeth of the radula. alysis of the habitats of the older fossil spe- cies has not been possible but the indications are that they are found together with typically marine species. The assumed movement into low salinity areas, mainly mangrove habitats, by the ancestral /ravadia stock appears to have been followed by a minor adaptive radiation resulting in a wide diversity of shell form and sculpture (Fig. 23). Species in- cluded here in the subgenus Pseudonoba range from brackish-water mangrove habitats to fully marine, relatively deep-water species. Classification In general, less is known about the genera possessing a spiral operculum than /ravadia. They appear to be generally related and have many of the characters listed in Table 2 in common. There is, however, little justification or advantage, considering the available evi- dence, for merging them to any greater extent than has been done, until more about them is known. It is possible that at least Hya/a and Ceratia may prove to be congeneric but the other groups probably represent lineages that have been long separated and are thus prob- ably anatomically distinct. Fig. 22 is an at- tempt to indicate probable relationships and provide a summary of my concept of the evolution of the group. Two genera (Rissopsis and Rhombostoma) are known only from their shells, so that their detailed relationships are auueny\sy ‘а -- F + - - (+) - — 6, = + + + в = + x - + + + + +6 + - - - - + эциеи “Y yeyIGeH “ZL é = ¿ ¿ ¿ = é 6 6 é + = = > juasqy ‘9 é = é é é = 6 6 6 6 = + = = рэдо|элэр-Ашооз 'g é ote @ 6 é ir é é é é - - + = padojanap-112M “Y oes wads Jouajuy ‘|| ¿ är é é ¿ + é é é é juasqy ‘3 é = é é é = é é ¿ ¿ It = = + ещеа pue jesayeq ‘а ¿ = é é é = é é é é = - + + ¡eyed pue |едиэл “9 ¿ = ¿ ¿ ¿ = ¿ ¿ ¿ ¿ = = sE + Áyneo jeijjed puiyaq ‘jesiog 'g ¿ = ¿ ¿ & = é é de ¿ = + = + jeiyed pue jesiog ‘У xuyejndoo esing ‘OL + (+) = qe SE + = + + + a + 3 Е 9411205044 ‘9 + + - - - - + + + - - -- ~ = BUul|DO4UO ‘g (a) en) (+ aunooyisido “y элпрэде jo dij 1э]по ‘6 + - - + + + + ++ + + + - + sE uoye¡nbue Jolajue yym ¡pays Jo ainuady ‘8 + u - _ - = spealy) jesids pue jeixe Yeam UJIM ‘а + sE Е juaulwold sjeixy ‘9 - - + - + - > - sE - - + среэлц} ¡eds HE9M цим 10 цоошс ‘g oc - — + + - - - = — - sE + + = Jueuluopaid sjeuids “y Lu 21n}din9s ¡Jays ‘/ a ¿ + + + + = = 8 эщиаэээ SN8JONN ‘9 5 6 a ¿ ешблешап$ snajonn 'g а. é é + + + - - + jeulbuew зпаопм “y uninolado ‘9 E = = = € = = é ЕЕ + - - + + auoN ‘9 ¿ = = + de = = ¿ += = + = — SE sured ус 'g é + + = é + + é + + = + = = лед aun y 4198] гепрел |едиээ uo sdsno jeseg ‘$ ¿ > > ¿ — + é é + - u + + + pajuapul Ajyeam 10 papunoy 'g ¿ = = é I = é é PUIG VW 100} JO pue 1012]S0d4 “p é = — é = = é é — + + + os 35 pajuawbid sajoejua] oieyda) “€ ¿ = = é = = a 6 - - - + - - juasaid ajoe]ua] ¡eipodejan ‘с ¿ ля = é = = é é jUasaid ajoejue] ешез ‘+ 8 ® о 9 a y < 3 S 3 o с à à Sr RON RC и - D, © 2 D = о 5 5 = 3 [73 => = D wn о © oa on a a а 5 D = = 2 o Sf ee Westy De Y sat) Si Et] ‘(pedojanap Aood Алэл ‘(+) :padojanap Абиуод$ Алэл + + ‘saloads ¡Je ul juasqe ‘— :591094$ awos ul juasald ‘+ :$э109э4$ 62 |е ul juasaid ‘ +) эерирелел ay) jo exe} dno1B-snuab uaapno; эц} ysinBunsip о} pasn seinjea] yeyIqey pue |P9IBo|oydiou ay) jo эшоз JO 151 Y ‘© ЭЛЯ IRAVADIID GENERA 63 Fairbankia р АХ FIG. 23. Diagram showing diversity of shell morphology and the shape of the central teeth of the radula in the genus /ravadia. In addition, speculative evolutionary relationships and incursions into brackish water are indicated. impossible to assess. Hyala vitrea is known in reasonable detail and only one, possibly aty- pical, species of Nozeba has been in- vestigated anatomically. The head-foot of the type-species of Liroceratia and Ceratia is known and only the radula and operculum of one species of Chevallieria have been avail- able for examination. Thus, the conclusions made about the relationships of these genera must be considered tentative. The subdivision of /ravadia is more con- servative than that of previous classifications because of the considerable similarity of the observed characters, with the exception of the shell characters in some cases. The geni- talia also show differences but there are limitations in the knowledge of the structure of the female genital system, in particular, for some of the groups. The grouping within /ra- vadia assumed to be the most primitive (Pseudonoba), contains species showing a wide diversity of radular structure and habitat and may be subdivisible. Any refinement of the classification, however, seems inadvis- able until more evidence is available. Familial Relationships Anatomically the Iravadiidae are very sim- ilar to the Hydrobiidae (as defined by Davis, 1980) but the shells have more resemblance to those of rissoids. Both the Rissoidae and an expanded concept of the Hydrobiidae (see below) are contrasted in Table 1 with the Iravadiidae. Radoman (1973) has defined nine families in his “superfamily” Hydrobioidea. On the available information these all appear to be 64 PONDER closer to the Hydrobiidae than to the Ris- soidae, lravadiidae or Pomatiopsidae. For this reason the table of characters comprising the Rissoidae, Hydrobiidae and Iravadiidae includes the characters of all Radoman’s Hy- drobioidea, as defined by him, combined with the definition of the Hydrobiidae of Davis (1980). This grouping is loosely referred to as Hydrobiidae sensu lato. The considerable similarity in the structure of the head-foot, the nervous system, the alimentary canal and the male genital system in the Iravadiidae and the Hydrobiidae (s.I.) may be due to convergence but there is also a strong possibility that the two groups may have a common ancestry. The peculiar, flattened, glossy protoconch is the only uniform and distinctive shell fea- ture which characterises the Iravadiidae. The iravadiids investigated apparently lack a dis- tinct posterior pedal gland and, in most spe- cies, a metapodial tentacle and pallial tenta- cles. The peculiar penes seen in this group are reminiscent of some Hydrobiidae in hav- ing accessory, glandular swellings. This fea- ture is also seen in some species of Rissoina (Kosuge, 1965; Ponder, 1968) but is other- wise unknown in the Rissoidae. The long, coiled dorsal folds in the anterior section of the oesophagus are other structures which are shared with the Hydrobiidae. Johansson (1950) described the female re- productive system of Hyala vitrea (Montagu) and showed that it had an anterior sperm sac and that the slit-like opening to the female genital duct was near the posterior end of the glandular pallial duct. Johansson doubted that Hyala vitrea should be included in the Rissoidae but, in the latest revision of the European rissoids (Fretter & Graham, 1978), H. vitrea is, along with Ceratia proxima, in- cluded in Onoba Adams, the type-species of which (O. semicostata (Montagu)) is known to have a normal rissoid female genital system (Fretter & Patil, 1961). The Iravadiidae and Hydrobiidae (s.l.) agree in most anatomical features except for the lack of spherules in the digestive gland of the Iravadiidae and in the structure of the female genital duct. They also differ from each other in their protoconchs; most hydro- biids having dome-shaped protoconchs of about 1% whorls with an irregularly pitted surface. This is in sharp contrast to the small, somewhat planorboid, smooth protoconchs of species of the Iravadiidae. The protoconch differences may, however, be partly due to the assumed different de- velopmental modes in the two families. The veliger larvae of Hyala vitrea are known to be planktonic (Thorson, 1946) and, although there are no direct observations on the other species in the family, the morphology of their protoconchs strongly supports the view that they all have planktotrophic larvae. Hydro- biids, on the other hand, are probably all direct developers. The lack of spherules in the digestive gland is unusual in the Rissoacea and seems to be a constant feature, as determined histologi- cally. These spherules have been con- sistently observed in other rissoacean fami- lies (my own observations, Fretter & Graham, 1962) except the Assimineidae. The female reproductive system of /ravadia ornata superficially resembles that of mem- bers of the Hydrobiidae. Common features include the small, subterminal genital opening and the bursa copulatrix lying, in part, behind the pallial cavity. /ravadia ornata differs from the hydrobiids in having a pallial bursal duct, in much of the bursa lying along the capsule gland, in lacking a narrow, coiled section of the oviduct between the renal oviduct and the oviduct gland, and in having a large seminal receptacle which opens at the junction of the ventral channel and the renal oviduct. It is conceivable that the female genitalia characteristic of the Hydrobiidae evolved from an iravadiid similar to /. ornata. This would have to occur by the posterior migration of the opening to the bursal duct because the bursa copulatrix of the hydrobiids, and its duct, lie behind the pallial cavity where they enter an extension of the ventral channel beneath the albumen gland. This sequence of events, however, is unlikely because of the special- ized opercula in /ravadia species and the lack of an oviducal coil. The longitudinal fold in the ventral channel of the posterior end of the capsule gland and the anterior part of the albumen gland of the iravadiids is probably homologous to the almost identical fold in the ventral channel of the capsule gland of the Hydrobiidae. In that family, however, the albumen gland lacks a ventral channel, this continuing posteriorly from the capsule gland as a closed, separate tube. The ventral channel in the Rissoidae lacks a fold (Johansson, 1948; Fretter & Gra- ham, 1962). The homology of this fold is clear in Hyala vitrea (Fig. 3, №) where the sperm groove (sg) lies behind the fold and continues posteriorly to the opening of the seminal IRAVADIID GENERA 65 receptacle. The situation in the genus /rava- dia (Figs. 4-7) is somewhat complicated by the formation of a secondary fold and groove to a greater or lesser extent within the sperm groove and the derivation of the bursa copu- latrix from the sperm groove. The fold persist- ing to the entrance of the seminal receptacle in /ravadia species appears to be derived from the upper inner edge of a “greater sperm groove” (i.e., the secondary grooves are apparently derivatives of the originally single sperm groove) and can thus be regarded as homologous with the fold in Hyala and the hydrobiids. The bursa copulatrix (as here recognised) of the Iravadiidae is possibly not homologous with that of the Hydrobiidae and Rissoidae. The bursa in these families agrees in position with the iravadiid seminal receptacle and may be homologous with it. The seminal recepta- cle(s), in the Hydrobiidae, in particular, usual- ly opens into the coiled section of the oviduct and may be a new structure. The Pomatiopsidae have been described and defined by Davis (1979, 1980) and show a wide variety of shell characters. All of the members of this family have two openings to the female genital system, a distal oviducal opening and a spermathecal opening. A nar- row spermathecal duct leads to the bursa copulatrix, which lies behind the pallial cavity. The radula also differs; the innermost basal cusps on the central teeth are larger than the outer ones, the opposite condition to that seen in the Iravadiidae. An anterior sperm sac (or bursa copulatrix) is known in two other rissoacean families, the Caecidae (Marcus & Marcus, 1963), which have uncoiled shells, and the Vitrinellidae (Fretter, 1956), which have depressed- helicoid shells and animals with well de- veloped pallial and metapodial tentacles. ACKNOWLEDGEMENTS | thank the curators of the museums listed in ‘Abbreviations’ for making facilities avail- able for the study of material. Several museums kindly loaned material for study, including ANSP, BMNH, NHMP (the De- partments of Marine Invertebrates and Palaeontology), NM, NMNZ, NMV, NMW, NSMT, SAM, USNM and ZMR, and | thank the curators responsible. Specimens were also provided by Dr. A. Waren, Dr. А. М. Golikov and Dr. J. Le Renard. Facilities ena- bling the examination of living material were provided by the Museum of the Northern Territory, Darwin, the University of Hong Kong, the Crocodile Research Institute, Man- ingrida, Northern Territory and Lt. Cmr. D. P. Fairfax (Singapore). Mr. |. Loch, Lt. Cmr. D. P. Fairfax, Dr. B. Morton, Mr. A. Dartnall and W. F. Ponder Jr. assisted with field work. Dr. A. Warén has provided me with valuable in- formation concerning some generic taxa pre- viously Classified in the Eulimidae and Aclidi- dae, groups in which he is specializing. | am particularly grateful to Mr. E. K. Yoo, who mounted and photographed the majority of the material used in this paper during his term of employment as my research assistant. Miss J. Hall and Mrs. D. Hughes were also responsible for some of the S.E.M. work. Mr. Yoo and Miss B. Duckworth produced the drawings of shells. The histological prepara- tion were done by Mrs. G. Serkowski. Mrs. J. Kerslake, Dr. A. Warén, Mr. B. W. Jenkins and Dr. G. Davis made valuable comments on the manuscript. Miss J. Hall prepared the plates for publication and assisted in many other ways. Overseas travel was supported by the Au- stralian Research Grants Committee, the lan Potter Foundation, the Science and Industry Endowment Fund, the British Council and the Australian Museum Trust. The remainder of this work has been supported in large part by an Australian Research Grants Committee grant. REFERENCES CITED ADAMS, H. & ADAMS, A., 1852, On a new arrangement of British Rissoae. Annals & Maga- zine of Natural History, (2)10: 358-359. ANNANDALE, N. & PRASHAD, B., 1919, Some gastropod molluscs from the Gangetic delta. Records of the Indian Museum, 16: 241-257, 1 |. BLANFORD, W. T., 1867, Contributions to Indian malacology No. 8. List of estuary shells collected in the lrawadi in Pegu, with descriptions of new species. Journal of the Asiatic Society of Bengal, 36: 51-72. BLANFORD, W. Т., 1868, Descriptions of Fairbank- ia, a new genus and species of Rissoidae from Western India. Annals & Magazine of Natural History, (4)2: 399—401. BOETTGER, O., 1902, Zur Kenntnis der fauna der mittelmiocanen Schichten von Kostej im Krasso- Szorényer Komitat. Il. Verhandlungen und Mit- teilungen des Siebenburgischen Vereins fur Naturwissenschaften in Hermannstadt, 51: 1- 200. 66 PONDER BRANDT, R. A. M., 1968, Description of new non- marine mollusks from Asia. Archiv ftir Mollus- kenkunde, 98: 213-289. BRANDT, R. A. M., 1974, The non-marine aquatic Mollusca of Thailand. Archiv für Mollusken- kunde, 105: i-iv, 1-423. BRUGNONE, G., 1881, Le conchiglie Plioceniche delle vicinanze di Caltanisetta. Bullettino della Societa Malacologica Italiana, 6: 83-157, 1 pl. CLARK, W., 1852, On some undescribed animals of the British Rissoae. Annals & Magazine of Natural History, 10: 254—264. COAN, E., 1964, A proposed revision of the Ris- soacean families Rissoidae, Rissoinidae, and Cingulopsidae (Mollusca: Gastropoda). Veliger, 6: 164—171. COSSMANN, M., 1888, 1896, Catalogue illustré des Coquilles Fossiles de l'Eocène des environs de Paris. Annales de la Société royale malacolo- gique de Belgique, 23(1888): 1-328, pl. 1-12; 31(1896): 1-94, pl. 1-3. COSSMANN, M., 1921, Essais de Paléoconcholo- gie Comparee. Livr. 12: 1-336, pl. 1-11. Paris. COSSMANN, M. & PEYROT, A., 1916-1919, Con- chologie néogénique de l’Aquitaine. Tome Ш (Gastropodes). Actes de la Société Linnéenne de Bordeaux, 69: 157-365; 70: 5-212. DANCE, S. P. & EAMES, F. E., 1966, New mol- luscs from the Recent Hammar Formation of south-east Iraq. Proceedings of the Malacologi- cal Society of London, 37: 35-43. DAVIS, С. M., 1979, The origin and evolution of the gastropod family Pomatiopsidae, with emphasis on the Mekong River Triculinae. Academy of Natural Sciences of Philadelphia, Monograph 20: 1-120. DAVIS, G. M., 1980, Snail hosts of Asian Schisto- soma infecting man: evolution and coevolution. In The Mekong Schistosome; Malacological Re- view, Supplement 2: 195-238. FLEMING, С. A., 1966, Marwick's illustrations of New Zealand shells with a checklist of New Zealand Cenozoic Mollusca. New Zealand De- partment of Scientific & Industrial Research, Palaeontological Bulletin, 173: 1-456. FRETTER, V., 1956, The anatomy of the pro- sobranch Circulus striatus (Philippi) and a review of its systematic position. Proceedings of the Zoological Society of London, 126: 369-381. FRETTER, V. & GRAHAM, A., 1962, British Pro- sobranch Molluscs, their functional anatomy and ecology. Ray Society, London, xvi + 755 p. FRETTER, V. & GRAHAM, A., 1978, The pro- sobranch molluscs of Britain and Denmark. Part 4—Marine Rissoacea. Journal of Molluscan Studies, Supplement 6: 151-241. FRETTER, V. & PATIL, A. M., 1961, Observations on some British rissoaceans and a record of Setia inflata Monterosato, new to British waters. Proceedings of the Malacological Society of Lon- don, 34: 212-223. GARRETT, A., 1873, Descriptions of new species of marine shells inhabiting the South Sea Is- lands. Proceedings of the Academy of Natural Sciences of Philadelphia, 1873: 209-231, 233- 237. GOLIKOV, А. М. 8 KUSSAKIN, O. G., 1978, Snell bearing littoral gastropods of the seas of the U.S.S.R. Leningrad, 256 p., 33 pl. (in Russian). GOLIKOV, А. М. & STAROBOGATOV, Ya. |., 1975, Systematics of prosobranch gastropods. Mala- cologia, 15: 185—232. GOUGEROT, L. & LE RENARD, J., 1977, Clefs de détermination des petites espèces de gastér- opodes de l'Eocene du Bassin Parisien. VIII. Le genre Graphis; IX. Le genre Pasitheola. Cahiers des Naturalistes. Bulletin des Naturalistes Parisiens, 30: 37-54. GRAY, J. E., 1847, A list of the genera of Recent Mollusca, their synonyma and types. Pro- ceedings of the Zoological Society of London, 15: 129-219. HABE, Т., 1958, On the radulae of Japanese marine gastropods (4). Venus, Japanese Journal of Malacology, 20: 43-60. IREDALE, T., 1915, A commentary on Suter’s “Manual of the New Zealand Mollusca.” Transac- tions & Proceedings of the New Zealand In- stitute, 47: 417—497. IREDALE, T., 1943, A basic list of the fresh water Mollusca of Ausiralia. Australian Zoologist, 10: 188-230. JEFFREYS, J. G., 1867, The Mollusca. Vol. IV. Marine shells, British Conchology, London, p. 1486, pl. 14. y JOHANSSON, J., 1948, Uber die Geschlechtsor- gane der Hydrobiiden und Rissoiden und den ursprunglichen Hermaphroditismus der Pro- sobranchier. Arkiv fór Zoologi, 40A(15): 1-13. JOHANSSON, J., 1950, Uber die weiblichen Ge- schlechtsorgane von Hyala vitrea, einer von dem Rissoa-Typus starck abweichenden Form der Gruppe Rissoacea. Arkiv fór Zoologi, 42A(7): 1-6. KOSUGE, S., 1965, Anatomical study on Japanese Rissoacea (l). On the family Rissoidae (I). Venus, Japanese Journal of Malacology, 24: 130—151, pl. 15. KURODA, T. & HABE, T., 1950, Nomenclatural notes. Illustrated Catalogue of Japanese Shells, 2516: KURODA, T. 8 HABE, T., 1954, New aquatic gas- tropods from Japan. Venus, Japanese Journal of Malacology, 18: 71-79. LASERON, C. F., 1950, Review of the Rissoidae of New South Wales. Records of the Australian Museum, 22: 257-287. LASERON, C. F., 1955, Revision of the New South Wales eulimoid shells. Australian Zoologist, 12: 83-107. LASERON, C. F., 1956, The families Rissoinidae and Rissoidae (Mollusca) from the Solanderian and Dampierian Zoogeographical Provinces. Au- stralian Journal of Marine & Freshwater Re- search, 7: 384—484. LEA, 1., 1833, Contributions to Geology. 227 p., 6 pl. MARCUS, E. & MARCUS, E. du B.-R., 1963. Meso- IRAVADIID GENERA 67 gastropoden von der Küste Sao Paulos. Akademie der Wissenschaften und der Literatur, Abhandlungen der Mathematisch- Naturwissenschaftlichen Klasse, 1: 1-105. NEVILL, G., 1884, Hand list of Mollusca in the Indian Museum, Calcutta. Part II. Gastropoda. Prosobranchia-Neurobranchia (contd.). Govern- ment Printer, Calcutta. x + 306 p. PALMER, K. v. W., 1937, The Claibornian Scapho- poda, Gastropoda and Dibranchiate Cephalopo- da of the southern United States. Bulletins of American Paleontology, 7(32)1: 1-548; 2: 549- 730. 90 pl. PAVIA, G., 1975, | Molluschi del Pliocene inferiore di Monteu Roero (Alba, Italia NW). Bollettino della Societa Paleontologica Italiana, 14: 99- 175, 14 pl. PILSBRY, H. А., 1922, Revision of W. M. Gabb's Tertiary Mollusca of Santo Domingo. Pro- ceedings of the Academy of Natural Sciences of Philadelphia, 73: 305—435, pl. 16—47. PONDER, W. F., 1967, The classification of the Rissoidae and Orbitestellidae with descriptions of some new taxa. Transactions of the Royal Society of New Zealand, Zoology, 9: 193-224. PONDER, W. F., 1968, The morphology of some small New Zealand Prosobranchs. Records of the Dominion Museum, 6: 61-95. PONDER, W. F., 1974, A review of the Australian species assigned to Rissopsis Garrett with a description of a new species of Rissopsetia (Mol- lusca: Gastropoda). Journal of the Malacological Society of Australia, 3: 25-35. PONDER, W. F. & YOO, E. K., 1976, A revision of the Australian and tropical Indo-Pacific Tertiary and Recent species of Pisinna (= Estea) (Mol- lusca: Gastropoda: Rissoidae). Records of the Australian Museum, 30: 150-247. RADOMAN, P., 1973, New classification of brack- ish water Prosobranchia from the Balkans and Asia Minor. Prosebna Izdanja. Museum d'His- toire Naturelle de Beograd. Edition hors serie, 32: 1-30. SACCO, F., 1892, | molluschi dei terreni terziarii del Piemonte e della Liguria, parte 11. Eulimidae e Pyramidellidae (parte). Memorie della R. Accademia delle Scienze di Torino, ser. 2, 42: 1-100, 2 pl. SEGUENZA, G., 1876, Studii stratigraphici sulla formazione pliocenica dell' Italia meridionale. Bollettino del R. Comitato geologica d'Italia, 7: 7-15, 91-103, 179-189, 260-271, 355-359 (part of series 1873-77). STAROBOGATOV, Ya. |., 1970, Molluscan fauna and zoogeographical partitioning of continental water reservoirs of the world. Akademiya Nauk SSSR. Zoologicheskii Institut Nauka, Leningrad, 372 p. (in Russian). STOLICZKA, F., 1867-1868, Cretaceous fauna of Southern India. 2. The Gastropoda. Memoirs of the Geological Survey of India Palaeontologia indica, Series V, 2, xi + 500 p., 28 pl. THIELE, J., 1928, Revision des Systems der Hyd- robiiden und Melaniiden. Zoologische Jahr- bücher. Systematik, Okologie und Geographie der Tiere, 55: 351—402. THIELE, J., 1929, Handbuch der systematischen Weichtierkunde. Jena. Part 1: 1-376. THORSON, G., 1946, Reproduction and larval de- velopment of Danish marine bottom in- vertebrates. Meddelelser fra Kommissionen for Danmarks Fiskeri- og Havundersogelser, serie: Plankton, 4(1): 1-523, figs. 1-199. WENZ, W., 1938-1944, Gastropoda. Handbuch der Paläozoologie. 6(1): Lief. 1-7. 1639 р. Berlin. APPENDIX. DESCRIPTIONS OF NEW SPECIES OF IRAVADIIDAE Genus /ravadia Subgenus Pseudonoba Iravadia (Pseudonoba) profundior Ponder, sp. nov. Fig. 14G-I Description. Shell. Small, elongately pupoid, rather solid, non-umbilicate. Pro- toconch small, smooth, glossy, of about two whorls, flat-topped, sharply-terminated. Tele- conch of four convex whorls, sutures im- pressed. Fifteen spiral threads on penultimate whorl, 18 primary spirals on body whorl together with a few, weak secondary spiral threads and several weak spirals on basal fold. Axial sculpture of indistinct growth lines. Lower base with broad, rounded basal fold. Aperture rather large, oval, very weakly- excavated posteriorly, not angled, broadly and shallowly-excavated anteriorly. Outer lip slightly prosocline with sharp edge, thickened within, with broad, prominent varix. Inner lip thickened, narrow, attached to parietal area. Colour yellowish-white, periostracum yellow to brown, with ferruginous deposit in one of the two paratypes (Fig. 14G). Dimensions. length diameter Holotype 3.86 mm 1.47 mm Paratype 4.24 ot 3.86 1.48 Operculum. Oval, columellar edge indented just beyond nucleus, which is displaced from middle of edge. A low, broad, internal fold parallel to columellar edge (Fig. 14H). Radula. Central teeth broadly-triangular, me- dian cusp short, with several (six to nine) minute cusps on either side. A pair of weak denticles near lateral margins of central teeth. Lateral teeth with prominent primary cusp and 68 PONDER several (three to five) small cusps on outer side. Marginal teeth multicuspate (Fig. 141). Type-locality. Off Lusaran Light, Guimaras, Philippines, USBF Stn. 5183, in 96 fathoms (175 m), in soft mud. Types. Holotype and two paratypes (USNM, 281302). All specimens collected alive. Other material examined. Off Adyagan Is., E. Masbate, Philippines, USBF Stn. 5392, 135 fathoms (247 т): two shells (USNM, 281897). Remarks. This new species is superficially similar to Paronoba subquadrata Laseron, 1950 (= /ravadia (Pseudonoba) subquadrata herein) in shell features but differs in being larger, in having a relatively thicker peristome with a much stronger varix, and in having stronger spiral sculpture. /ravadia (Pseudo- noba) bella also has a similar shell but this differs in being larger, more inflated, with rela- tively weaker spiral sculpture and no basal fold. The two additional specimens (USNM, 281897) referred to this species have smaller shells (maximum length 3.19 mm) than the type-series, the outer lip is orthocline, the aperture is weakly subangled posteriorly and the teleoconch consists of only 3% whorls. They agree in most other details and con- sequently are regarded as a variety of /. (P.) profundior. Two lots in the USNM from Daram Channel, W. Samar, 32 fathoms (59 m) (USNM, 280751) and N of Marinduque, 50 fathoms (95 m) (USNM, 276116), have still smaller shells (maximum length 3.01 mm), with relatively stronger spiral sculpture and an orthocline to opisthocline outer lip. They are more similar to the two atypical specimens of I. (P.) profoundior (USNM, 281897) than to the type series and are not considered to be the same species. These specimens also have some similarity to /. (P.) densilabrum. Iravadia (Pseudonoba) expansilabrum Ponder, sp. nov. Fig. 16F-I Description. Shell. Small, elongately- pupoid, thin, non-umbilicate. Protoconch small, smooth, glossy, with two whorls, flat- topped, sharply-terminated. Teleoconch of about 4% convex whorls, with moderately- impressed sutures. Sculpture of close, fine spiral threads (approximately 26 on penulti- mate whorl) and inconspicuous axial growth lines. Lower base with prominent fold sculp- tured with a few weak, spiral threads. Aper- ture oval, very weakly-subangled and ex- cavated posteriorly; convex and broadly and shallowly excavated anteriorly. Outer lip slightly opisthocline, thickened within, with sharp edge, a thick varix behind. Inner lip broadly-expanded over parietal area as a thin callus, not expanded over basal fold. Colour yellowish-white, periostracum yellowish but with a thick, ferruginous coating in the two specimens collected alive (Fig. 16G). Dimensions. length diameter Holotype 3.18mm 1.18 тт Paratypes 2.95 1.10 3.21 1.24 3.03 1.14 3.06 1.16 Operculum. Oval, columellar edge flattened, twisted outwards with a weak, internal, longi- tudinal fold parallel to it. Nucleus marginal, slightly displaced from middle of edge (Fig. 16H, |). Radula. Central teeth broadly- triangular, with strong cusps on narrow ip seed eo 1 prominent denticles on face of each tooth. Lateral teeth with large primary cusps and possibly a small cusp on inside edge. Margin- al teeth multicuspate (Fig. 16F). Type-locality. c. 40 miles (64 km) WSW off Tulear, Malagasy Republic, 23°19'S, 43°36'E, 82 т. 6 August, 1964. A. Bruun Sti. 363W. Types. Holotype and five paratypes (USNM, 717549). Holotype coated with gold. Remarks. This species is closest to /. (P.) subquadrata but its shell differs from that species and all others in the genus by its broadly expanded inner lip. The radular fea- tures, especially those of the central teeth, are unusual for /ravadia, and are discussed in the remarks on Pseudonoba above. cutting edge , a single pair of Iravadia (Pseudonoba) gemmata Ponder, Sp. nov. Fig. 16А-Е Description. Shell. Small, very elongately- conic, solid, non-umbilicate. Protoconch small, smooth, glossy, of 2-25 whorls, flat- topped, sharply-terminated (Fig. 16B). Teleo- conch of about six convex whorls, with deeply-impressed sutures. Sculpture of about 17 nodulous spiral ribs on body whorl, 12-14 IRAVADIID GENERA 69 on penultimate whorl, crossed by weak, close, prosocline axial riblets with small, rounded gemmules at points of intersection. Lower base with.prominent, sharp fold, smooth or with weak spiral striae, partly over- laid by inner lip. Aperture rather large, oval, with weak angulation and excavation posteri- orly, shallowly and broadly excavated an- teriorly; inner lip broad, thickened, attached to parietal area, its outer edge convex; outer lip thickened at edge and behind, slightly pro- socline, with a wide, flat-topped varix im- mediately behind. Colour white, periostracum yellow to brown, thin. Some specimens coat- ed with a ferruginous deposit which stains shell orange-brown to nearly black (Fig. 16A). Dimensions. length diameter Holotype 5.52 тт 1.85 mm Paratypes (277683) 5.33 1.78 4.95 1.62 5.27 1.80 512 1:65 (283016) 4.73 1.65 Operculum. Oval, columellar edge convex, nucleus near middle of edge. No internal ridges (Fig. 16C). Radula. Central teeth broad, dorsal margin convex, cutting edge narrow, with one or two weak lateral cusps and a long, sharp median cusp; lateral mar- gins widely flared, thickened. Three pairs of strong, closely-spaced denticles arise from face of tooth just below lateral margin; fourth pair of weak denticles below outer part of lateral margin. Middle of face of central tooth with a vertical, thickened band; ventral edge of tooth convex. Lateral teeth elongate, with a single, long, sharp cusp at inner edge and very weak cusp outside this cusp; otherwise simple. Marginal teeth with numerous sharp cusps (Fig. 16D, E). Type-locality. Off Limbancauayan 1$., W. Samar, Philippines; 50 fathoms (91.4 m), fine, grey sand, USBF Stn. 5210. Types. Holotype (USNM, 283016a) and three paratypes (two juveniles) (USNM, 283016b); 16 paratypes (one juvenile) (USNM, 177683). Holotype coated with gold. Other material examined. Buton Strait, Celebes (Sulawesi), 37 fathoms (68 т), USBF Stn. 5642: ten specimens (USNM, 279749), one specimen (USNM, 290829). Philippines: S of Corregidor Lt., 35 fathoms (64 т), USBF Stn. 5100: two specimens (USNM, 257398), two specimens (USNM 257397). North off Marinduque, 50 fathoms (91 т), USBF Stn. 5220: 17 specimens (USNM, 276115). North off Marinduque, 193 fathoms (353 т), USBF Stn. 5221: one specimen (USNM, 277065). Lagonoy, East Luzon, 47 fathoms (86 m), USBF Stn. 5448: one specimen (USNM, 289018). Off Tacbuc Point, East Leyte, 62 fathoms (113 m), USBF Stn. 5479: four specimens (USNM, 283016). Off Magabao Is., East Mindanao, 494 fathoms (903 m), USBF Stn. 5236: one speci- men (USNM, 276875). Remarks. The radula of this species differs from all others known in the genus by the characters of the central teeth, as discussed in the remarks on the subgenus Pseudonoba above. The shell is readily separated from other species of Pseudonoba by its gemmate sculpture. The operculum is closest to that of Iravadia (Pseudomerelina) mahimensis in morphology. This species probably represents a sepa- rate grouping within, or close to, /ravadia $.1. but, until the radulae of more species of /rava- dia (Pseudonoba) are known, allowing its re- lationships to be more accurately assessed, it can be tentatively retained in Pseudonoba. The denticles on the face of the central teeth appear to have increased in size and number to take over the scraping function of the lateral cusps, the original cutting edge being reduced in size. Species of /ravadia, typically have several small cusps and weak- er (i.e., less thickened) teeth. Genus Chevallieria Chevallieria australis Ponder, sp. nov. Fig. 17D-F Description. Shell. Minute, pupoid, thin, non-umbilicate. Protoconch relatively large, of 13, whorls, flat on top, terminated abruptly, smooth. Teleoconch of about 2% convex whorls, sutures lightly impressed. Surface smooth except for indistinct axial growth lines. Aperture rather large, oval, subangled and very slightly channelled posteriorly, broadly and very shallowly excavated anteriorly. Out- er lip orthocline, slightly thickened within, no external varix. Inner lip thickened, narrow up- per half attached to, or free from, parietal wall, lower half free. Periostracum thin, yellowish- brown (Fig. 17D). 70 PONDER Dimensions. length diameter Holotype 1.50mm 0.68 mm Paratype 1.64 0.75 Operculum. (Fig. 17E) and radula (Fig. 17F) as described under generic diagnosis for Chevallieria. Type-locality. Emu Вау, М.Е. Kangaroo ls., South Australia, on sheltered, rocky shore under rocks at low water neap; 7 March 1978. Collected by I. Loch, E. К. Yoo and К. Hand- ley. Types. Holotype (C.126181) and paratype (C.126182), AMS. Other material examined. Point Sinclair, S. Australia, two shells (AMS, C.126183); Largs Bay, S. Australia, one shell (SAM); Tumby Bay, S. Australia, one shell (AMS, C.126184); Bathurst Point, Rottnest Is., Western Aus- tralia, one shell (AMS, C.126185). Remarks. This species is closest to Cheval- lieria columen (Melvill) but the shell of that species is larger (greater than 2.5mm in length), relatively narrower and the teleo- conch consists of about 3. whorls. The new species and С. columen differ from the other Recent species described from Australia, C. imitoris (Laseron), in lacking fine spiral striae on the surface of the shell. Chevallieria im- itoris also has a relatively narrower shell than С. australis. lt is known from Darwin, N Aus- tralia (the type-locality) and Queensland. Other specimens of Chevallieria from the tropical Indo-Pacific appear to include un- named species, but insufficient material is available to describe these. Chevallieria balcombensis Ponder, sp. nov. Fig. 17B Description. Shell. Minute, pupoid, thin, non-umbilicate. Protoconch relatively large, of two whorls, flat on top, terminated abruptly. Teleoconch of about three convex whorls, sutures slightly impressed. Surface smooth, shining, with weak axial growth lines the only sculpture. Aperture rather large, oval, weakly- subangled and very shallowly channelled posteriorly, broadly and shallowly-channelled anteriorly. Outer lip very slightly opisthocline to very slightly prosocline, thickened slightly within, with sharp edge and prominent, round- ed external varix. Inner lip narrow, thickened, attached or partially free from parietal wall (Fig. 17B). Dimensions. length diameter Holotype 1.98mm 0.79 тт Paratypes 2.24 0.83 1.98 0.79 1.62 0.73 1.90 0.79 Bird Rock Cliffs 2.20 0.92 Type-localty. Fossil Beach, Balcombe Bay, Mornington, Victoria, Australia. Middle Miocene (Balcombian), J. Voorwinde Colin. Types. Holotype (C.126186) and eight pa- ratypes (C.126187), AMS; two paratypes NMV, P.62046. Other material examined. Topotypes, W. J. Parr Colln., two specimens (ММУ, P.62044, 62047); Bird Rock Cliffs, Torquay, Victoria, Lower Miocene-Upper Oligocene (Janjukian) one specimen (NMV, P.62045). Remarks. Of the Australian species, the shell of C. balcombensis is closest to C. australis in lacking spiral striae, but it differs in its slightly larger size and prominent apertural varix. Chevallieria gippslandica Ponder, sp. nov. Fig. 17C Description. Shell. Small, elongately- conical, rather solid, non-umbilicate. Pro- toconch small, smooth, of two whorls, flat- topped, sharply-terminated. Teleoconch of 4/4 convex whorls; sutures impressed. Sculp- ture of weak, axial growth lines only. Aperture oval, subangled and with shallow channel posteriorly, broadly and slightly excavated an- teriorly. Outer lip slightly opisthocline, thick- ened within, with a sharp edge and flat, weak to prominent external varix. Inner lip narrow, attached to, or separated from, parietal wall (Figs 17€): Dimensions. length diameter Holotype 3.82 mm 1.44 mm Paratypes 3.58 1.41 3.57 1.32 3.18 1.18 Type-locality. Roadcutting, right bank of Meringa Creek, 100m S of Kalimna- Nungerner Road, Gippsland, E. Victoria, Australia. In bed below Lower Jemmys Point shell bed, Jemmys Point Formation, Kalim- nan, Lower Pliocene. Collected by W. F. Pon- der, T. A. Darragh and P. H. Colman, 11 Jan. 1970. IRAVADIID GENERA 71 Types. Holotype (C.126207) and three paratypes (C. 126208), AMS; two paratypes, NMV, P.62048 from type-locality, collected Dec. 1966. Other material examined. Kalimnan: Ditch on E side of Meringa Creek, in lowest shell band four m above creek (NMV, P.62048, two paratypes); SW side of Bunga Creek, road cutting Princes Highway, Gippsland, E. Victor- ia, one specimen (NMV, P.62049); zero to two т above beach in cliff 50-100 m E of Kalimna Jetty, in lower shell bed, Gippsland, E. Victoria (NMV, P.62051). Cheltenhamian: NE side Bunga Creek, road cutting, Princes Highway, Gippsland, E. Victoria; one specimen (NMV, P.62050). Remarks. This species is relatively large for the genus and resembles species of /ravadia (Pseudonoba) in general shell characters. It lacks, however, any trace of spiral sculpture and does not have a basal fold. It appears to be derived from Chevallieria balcombensis, which it closely resembles, except in size and number of whorls. Genus Nozeba Nozeba (?) striata Ponder, sp. nov. Fig. 20G-J Description. Shell. Small, elongately-ovate, rather thin, with dull surface. Protoconch smooth, of moderate size, of 1%4 whorls, nu- cleus very small, depressed. Teleoconch of 2% convex whorls, sutures impressed. Sculp- ture of fine, close, raised spiral lines with linear interspaces and weak axial growth lines. Aperture sub-oval with broad anterior notch and posterior subangulation. Outer lip orthocline, lacking external varix; inner lip nar- row, slightly thickened, free in anterior half. Umbilical chink small in holotype, more pro- nounced in paratype (Fig. 20G). Dimensions. length diameter Holotype 184 mm 1.05 mm Paratype 1.69 1.03 Operculum. Thin, simple, spiral, nucleus eccentric (Fig. 20J). Radula. Central teeth broad, cutting edge concave, lateral wings only slightly thickened, at about 45° to ver- tical; median cusp very small, blunt, six to seven lateral cusps on each side, narrow, sharp, inner ones about twice length of me- dian cusp; no basal denticles, lateral margins slightly thickened. Lateral teeth with narrow cutting edge, primary cusp moderately long, sharp, secondary cusps sharp, с. 3 + 1 + с. 6. Marginal teeth with broad, almost straight cutting edges and long bases, with numerous small, sharp cusps almost equal in length to secondary cusps on lateral and central teeth, but narrower (Fig. 20H, |). Type-locality. Pujada Bay, E Mindanao, Philippines, 218 fathoms (399 m) USBF Stn. 5243. Types. Holotype and paratype (USNM, 311059). Holotype coated with gold. Remarks. This species differs from the two Australasian species of Nozeba, for which the radula is known, in not having a pair of basal denticles on the central tooth of the radula, and in the median cusp of the central teeth being relatively very small. The shell, howev- er, has all the essential characters of Nozeba species as diagnosed herein, except that it lacks a glossy surface. It differs from the described species by the combination of the following characters: the relatively large aper- ture, the spiral sculpture, the dull surface, the rather solid shell and the umbilical chink. Only two other described species have distinct spiral sculpture: N. emarginata and N. gatlif- fiana. MALACOLOGIA, 1984, 25(1): 73-108 POTAMOLITHUS: MORPHOLOGY, CONVERGENCE, AND RELATIONSHIPS AMONG HYDROBIOID SNAILS George M. Davis'* & Maria Cristina Pons da Silva! ABSTRACT Heuristic evaluations of the origin, evolution and deployment of faunas depend on thorough systematic studies. Systematic relationships among rissoacean taxa cannot be adequately assessed without comparative anatomy of all organ systems and ontogenetic analyses of some of them (Davis, 1979). The cohesiveness of the morphological groundplan seen in an endemic adaptive radiation may assure the anatomist that the taxa of that radiation are monophyletic. To state that geographically separated radiations are monophyletic and to show their relationships cladistically is more complex. The problems and constraints have been discussed lucidly (Cain & Harrison, 1960; Cain, 1964). The major problems are convergence and convincing assessments of homologies. The Hydrobiidae sensu lato were defined by similarities of the shell, radula, operculum, and penis. The inadequacy of describing higher taxa by such characters has already been shown (Davis, 1979, 1980). At least 30% to 40% of the characters used to discriminate rissoacean taxa are from the female reproductive system. Shell characters are the least reliable because of convergence (Davis, 1979, 1980, 1981). In the late nineteenth century snails in various places of the world were placed in Lithoglyphus because of a globose to cap-shaped shell, simple penis, and presence of three or four pairs of basal cusps on the rachidian radular tooth. Lithoglyphus s.s. is European and the type-species, L. naticoides Pfeiffer, is the standard for the hydrobiid subfamily Lithoglyphinae (Davis et al., 1976, Appendix). Species of so-called Lithoglyphus of western Yunnan, China are in the Pomatiopsidae: Triculinae (Davis, 1979). Lithoglyphus from southern Brasil (von Ihering, 1885) and Uruguay was monographed by Pilsbry (1911), who placed some thirty species in a new genus Potamolithus in the Amnicolidae (= Hydrobiidae). While several species of Potamolithus clearly resemble triculine Lacunopsis of the Mekong River, an analysis of the female reproduc- tive system of Potamolithus was needed to confirm the suspected relationship. The possibility existed that Potamolithus converges on taxa of the Triculinae because it lives on rocks in high energy environments (Davis, 1979). Anatomical data for Potamolithus ribeirensis Pilsbry show that this species is not a pomatiop- sid but a hydrobiid with closest relationships to Lithoglyphus s.s. Some differences clearly show divergence between these genera while other differences may or may not be shown to be real, pending a more detailed analysis of Lithoglyphus s.s. These findings tend to negate a Gondwanaland origin for Potamolithus. As no Triculinae are found in South Africa (Davis, 1981) and Potamolithus is not a triculine, the Triculinae more probably originated on the Indian Plate section on Gondwanaland. Potamolithus may be closely related to North American genera such as Somatogyrus and Fluminicola, as evidenced by similar shells, radulae, opercula, and penises. However, with convergence in mind, evidence for this must come from detailed anatomical studies of these North American genera. Key words: Potamolithus; Hydrobiidae; morphology; systematics; convergent evolution; South America; Brasil. INTRODUCTION There is considerable current interest in coupling phylogenetic and zoogeographic an- alyses to understand the faunistic patterns seen in the world today. There are serious constraints in assessing phylogeny and in many cases it may not be possible to do so (Cain & Harrison, 1960; Cain, 1964). Major problems include lack of sufficient characters and qualitatively different character-states, convergence, and a correct evaluation of “Supported in part by United States National Institutes of Health Grant No. TMP 11373 and National Science Foundation U.S.A.-Brasil Cooperative Science Program No. DEB-8205827 awarded to G. M. Davis, and CNPq No. 01 .10.004-82 to M. C. Silva. 2Academy of Natural Sciences of Philadelphia, Nineteenth and the Parkway, Philadelphia, Pa. 19103, U.S.A. ЗМизеи de Ciéncias Naturais, Fundacáo Zoobotanica do Rio Grande do Sul, Porto Alegre, RS, Brasil. (73) 74 DAVIS & PONS DA SILVA homologies. It is also important to realize that faunistic spatial patterns result from the sum of vicariant events, dispersal events, and vari- ance of ecological factors through time (Davis, 1981). Adaptive radiations in different parts of the world and at different times, in which the species resemble each other phenetically, are of considerable interest to the evolution- ary biologist and zoogeographer. Do the re- semblances reflect phylogenetic relationships or convergences? In attempting to understand something of the origin and evolution of the rissoacean family Pomatiopsidae, it was noted that var- ious species from around the world were placed in the genus Lithoglyphus by early authors (Davis, 1979). Species were placed in Lithoglyphus because of the resemblances in their shells and opercula (as well as radu- lae and penises in some cases) to those in Lithoglyphus naticoides, the type-species. These features are a large globose shell (4 to 6mm length), two or more pairs of basal cusps on the central tooth; one or more pairs arise from the face of the tooth, not the lateral angle. Penis is without appendages. The operculum is corneous and paucispiral. It was shown that Lithoglyphus s.s. of Europe con- verges on species of so-called Lithoglyphus and other similar looking species of Lacunop- sis from the Yangtze and Mekong rivers of Southeast Asia (Davis, 1979). European Lithoglyphus belongs in Hydrobiidae: Litho- glyphinae while the Asian and Southeast Asian taxa in Pomatiopsidae: Triculinae. Lithoglyphus of Lake Tanganyika, Africa, later considered Lacunopsis, is now known as Spekia, an endemic monospecific cerithia- cean genus that converges on Lacunopsis harmandi of the Mekong River (Davis, 1979). The head and penis of Lithoglyphus lapi- = ARGENTINA 10° a a ( BRASILIA = e les NE J AN i » SN > y 4 TRS ES qa an 222 » a MATO GROSSO VF \ ry DO-SUL AO A A Gu oAvanıa À (ae) о ( A 20 CN 1 EN ES d 1 N ! We \ Cri E LO SE. ® Fe RA RIO DE JANEIRO AO PAULO R. Ribeira Se 30° FIG. 1. Map showing the most important localities for Potamolithus along the Uruguay and Parana rivers. Paysandu in Uruguay is the type-locality of many species of Potamolithus. The type-locality of P. ribeirensis is shown by the arrow. Our study area is shown by the box of dashed lines around Porto Alegre. The region where the Pomatiopsidae: Pomatiopsinae Aquidauania comes from is shown in Mato Grosso do Sul State of Brasil. POTAMOLITHUS SYSTEMATICS 75 dum Orbigny from southern Brasil were illus- trated by von Ihering (1885). The animal, shell and radula so much resembled certain spe- cies of Lacunopsis (a genus endemic to the Mekong River) that the question arose as to whether South American Lithoglyphus was in the Triculinae or phenetically related to the Triculinae; or was the resemblance due to convergence caused by life on rocks and stones in a fluviatile environment (Davis, 1979, pp. 30, 33)? Pilsbry (1896, 1911) cre- ated the genus Potamolithus for the South American radiation that includes L. /apidum. The type-species is P. rushii Pilsbry, 1896 by original designation. The Potamolithus radia- tion comprises some thirty species that par- ticularly abound in the Plata-Uruguay River drainages. We present the anatomy of P. ribeirensis to answer questions raised about phylogenetic relationships and convergence. P. ribeirensis is a species discussed by von Ihering (1885) as Lithoglyphus lapidum. We show that Pota- ) » $ SPORTS X ALEGR Li molithus converges on certain Triculinae and is a member of the Hydrobiidae: Lithoglyphi- nae if we have correctly interpreted the pub- lished anatomy of Lithoglyphus naticoides (Krull, 1935; Krause, 1939; Radoman, 1966). This study underscores once again, that con- vergence must not be underestimated when attempting phylogenetic and zoogeographic analyses. MATERIALS AND METHODS Localities Feitoria River, 100-200 m upstream from bridge crossing on federal road BR116 from Novo Hamburgo to Dois Irmaos, Rio Grande do Sul, Brasil 29'35'5, 510 19,21 March 1981 (Figs. 1, 2). Academy of Natural Sciences (ANSP) catalog nos.: 353484; A8727; 353485; A8728. Type-locality for Potamolithus ribeirensis FIG. 2. Map showing details of our collecting site for Potamolithus ribeirensis N of Porto Alegre, Brasil. The arrow marks the locality. DAVIS & PONS DA SILVA FIG. 3. Shells of Potamolithus ribeirensis. A. Lectotype, ANSP 103076. Shell length is 3.68 mm. Other shells printed at the same magnification. Shells В-Е are from our study site (ANSP 353484). Shell Е has a pronounced keel and excavation of the umbilical region. POTAMOLITHUS SYSTEMATICS 77 Pilsbry. Rio Ribeira, Yporanga (or Iporanga), Sao Paulo, Brasil; 24°35'S, 48°35’W. Lectoty- pe, ANSP 103076 (Fig. 3A). See Appendix. Habitat Mountain stream of clear, clean water flow- ing over stones and boulders; intermittent pools with sandy bottom. P. ribeirensis crowded on all sides of stones and boulders at stream margin where there was continuous current. Snails absent, however, in swift cur- rent and white-water rapids. Associated fauna: Chilina sp. (Chilinidae) in micro- sympatry with P. ribeirensis, but fewer in numbers, approximately 500 to 1. In deposi- tional quiet-water areas, Littoridina sp. sparse; some Pisidium (Pisidiidae) also. Collection and methods Methods are those presented in detail by Davis & Carney (1973), Davis et al. (1976), and Davis (1979). Characters that are shared with other rissoaceans or that are standard in the Hydrobiidae and Pomatiopsidae are not discussed here. Examples of such characters are the style sac, fecal pellets, structure of the ctenidium, etc. Statistical analyses are 1) the standard “t” test for congruency of means, 2) X? analysis, and 3) multivariate morphometric analyses of shell parameters using NT-SYS (Rohlf et al., 1972) with details given in the Appendix. We took a random sample of the population to determine whether there was sexual di- morphism. The sample consisted of 184 in- dividuals drawn from an initial population of over 1000 individuals, which were all the in- dividuals of all size classes at a single site in the river. We determined the number of in- dividuals and their sex for each whorl number. We determined the shell length and length of body whorl for males and females of 4.0 and 4.5 whorls. Multivariate analysis The data matrix consisted of 8 OTUs and 32 characters. Hydrobia truncata represents the genus Hydrobia as typical of the Hydro- biidae: Hydrobiinae. All species of Lacunop- sis have the character states as scored. Lacunopsis is the sole genus of the tribe Lacunopsini of the Pomatiopsidae: Triculinae (Davis, 1979). Pomatiopsis lapidaria and Tomichia ventricosa are representatives of the Pomatiopsidae: Pomatiopsinae. Tricula aperta and T. bollingi are representatives of the generalized Pomatiopsidae: Triculinae: Triculini. A more comprehensive cladistic an- alysis of genera and species of the Pomatiop- sidae: Triculinae: Triculini has been pre- sented elsewhere (Davis & Greer, 1980). Lithoglyphus naticoides represents the Hyd- robiidae: Lithoglyphinae. Of the characters used, 3 (9%) are from the shell; 7 (22%) involve head-foot-mantle: 1 (3%) mantle cavity; 2 (6%) radula; 5 (16%) male reproductive system; 11 (34%) female reproductive system; 3 (9%) involve one each for stomach, nervous system, and eggs. Computations were made using the June 1974 version of NT-SYS (Rohlf et al., 1972). О and В mode analyses were performed as in Hoagland & Davis (1979). Non-metric multi- dimensional scaling (MDS) was emphasized. Limited reliance can be placed on cluster analyses and phenograms to illustrate rela- tionships. Ordination and MDS are freed from the constraints of phenogram construction. Ordination diagrams based on three- dimensional scaling are presented with sub- sets. The subsets algorithm groups OTUs so that the largest distance between OTUs with- in a group is less than the distance between any group member and any candidate for membership. We present a table for character loading on each principal component. We discuss char- acter correlations as pertinent to later dis- cussions of relationships among taxa and problems caused by convergence. Cladistic analysis We present a cladistic analysis based on a set-theory solution involving unique and un- reversed characters sensu Wilson (1965), and Davis & Greer (1980). The cladistic and multivariate analyses complement each other in our assessment of relationships. RESULTS: MORPHOLOGY Shells of the Feitoria population Shells of the Feitoria population are glo- bose, imperforate, and solid, but have a weak outer lip (Figs. 3, 4). Shells with heavy peri- ostracum. The spire is low; there are 4.0 to 4.5 whorls. The body whorl is regularly con- vex in most shells but slightly flattened in a few. The aperture is oblique, inclined about DAVIS & PONS DA SILVA FIG. 4. Shells of Potamolithus ribeirensis. Shell A was figured by Pilsbry (see Appendix). ANSP 103068; length is 4.04 mm. Other shells are printed at the same magnification. Shells B-E are from our study site showing variation in shape. POTAMOLITHUS SYSTEMATICS 79 55° to 40° toward the axis of coiling. The aperture flares at the adapical end, rounded to angular where the columella meets the outer lip. The columella is heavy and cal- loused; it is wide, flat or slightly concave. The parietal callus is weak at the edge, so that the peristome is not complete or barely complete in 95% of all individuals; it is heavy within. A thick, complete parietal callus is only seen on old individuals. The umbilical area is extremely narrow. In some individuals a sharp keel circles from the umbilical area to the adapical end of the inner lip (Figs. 3D, E). In older individuals the um- bilical area can be eroded and excavated to expose an umbilical opening. The Feitoria population mostly consisted of males and females of 4.0 whorls (72.6%; Table 1). Only 8.6% of the population attained 4.5 whorls. Individuals of 3.5 whorls and larger were sexually mature. There was no significant difference in numbers of males TABLE 1. Random sample of 186 individuals from a universe of >1000 individuals, showing the frequencies of males and females at each whorl stage. N = 186. Whorls 3 ? % of population eroded 0 2 He 3.0 2 3 Prd 3:5 11 17 15.1 4.0 66 69 72.6 4.5 7 9 8.6 = 86 100 and females (Table 1). Shell dimensions are given in Tables 2-4. There was no sexual dimorphism in shell size (Table 2). Most snails had 4.0 whorls. Size-frequency histograms of males and females with 4.0 whorls show a normal distribution if both males and females are considered together (Fig. 5). However, considering males and females separately, in the center of the dis- tribution (3.2 to 3.4 mm) there is a deficiency of males. A Х? analysis of numbers in this size class versus numbers in all other size classes combined suggests a significant difference between sexes (Р < .10) in this medium Class. Shell dimensions of the syntypic series are given in Table 3. The syntypes only attained 3.5 whorls; the parietal callus is more pro- nounced in the specimens compared with the condition seen in Feitoria specimens. Shell morphometric analyses of individuals from the two populations are given in the Appendix. Shell dimensions of those individuals used for the anatomical analyses are given in Table 4. Head-foot The head is broad and squat with thick stubby tentacles (Fig. 6). The snout is rel- atively short and narrow compared to the width of the neck. The foot is wide and power- ful, with the usual anterior mucous groove. There is no pronounced omniphoric groove, suprapedal fold, or pedal crease, as in the Pomatiopsinae. The operculum is corneous, paucispiral, with the position of the nucleus as figured for Somatogyrus or Fluminicola by TABLE 2. Shell length (L) and length of body whorl (LBW) (mm) for males and females of 4.0 and 4.5 whorls. N = number measured; X + Sd; (range). SD = significant difference between males and females; X = mean; Sd = standard deviation. SD 3 9 PEO 4.0 3.41 + 0.44 3.46 + 0.43 No (2.40 — 4.32) № = 50 (2.48 — 4.56) N = 50 L 4.5 4.09 + 0.44 4.31 + 0.24 № (3.28 — 4.72)N=7 (3.88 — 4.52) N = 9 4.0 2.98 + 0.48 3.02 + 0.40 No (2.00 — 3.96) М = 50 (2.32 — 3.76) М = 50 LBW 45 3.64 + 0.46 3.85 + 0.18 No (2.72 — 4.12)N=7 (3.52 — 4.44)N=7 CS DAVIS & PONS DA SILVA 80 ll nn (25`0- 770) Wrı-zı) (pre — 967) (rye — l'E) (vp - ge) (wer - 88€) (00'S - 09°) 50`0 + 840 OO EN ChO+FS0E 910 +656 620-107 SHO Oy — 950 + 99% Gp y (25`0 - ro) (l'E - 01) ge — ze) (91€- 99%) (@Ge-962) (9ee-962) (eve - 95) 70`0 = 970 — 90`0 + 101 4110 = 562 O2OF IBS 610+ 526 ECO+ CCE 8IO + EVE Ov 9 (yee — 95'1) (pee — 78") = 7> = — [70 + 60°¢ zu Gt'O + Spa GE g-99e1 :J8AIH BONO y dSNV (09:0 — 90) (701 - +90) (ere - 91) (962-881) (258 - voz) (098 - ве) (e9'e - вес) ge sadA} 600 + 6t 0 810 + 880 9€ 0 = 0,7 57'0 + 6vc eg0 + c8c Gt 0 + 06% 650 + SIE 0} 0'€ G -0}98[B1Ed 8t 0 Ol br? c6c Ove 95'5 89'€ ge | dSNV adÁAjo]997 пе ¡JOYM ainuede aınuade oym Ароа UIPIM uj6ue7 ‘Ou ‘ON Jejlawın]oO ajewnnuad JO ШРМ jo yyBua7] jo y¡Bua7] НОЧМ JO UJPIM JO UJPIM Е) \/-ээел :Áneso¡-adÁ | e eee ee ке ‘иоцеп4о4 euoyay ayy шод suawidads pue sauas-adÁ] ay} jo ззиэшелпзеэш |jays jo чозиедшоэ y ‘€ FIGVL POTAMOLITHUS SYSTEMATICS 81 of Individuals Number Number of Individuals N 2.2 E a 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 Length of Body Whorl (mm) FIG. 5. Histogram of snail sizes (males and females) with four whorls, based on length of the body whorl. From a random sample. 82 DAVIS & PONS DA SILVA Apo FIG. 6. Head and mantle cavity of a female. Note the 1) elongate osphradium, 2) thickened wide mantle border (Tmb), 3) stubby tentacles (TN), 4) nuchal node (Nn), and 5) position and prominence of the common genital aperture (Cga). An, anus; Apo, anterior pallial oviduct = capsule gland; Cga, common genital aperture; Ct, ctenidium; Es, esophagus; Ey, eye; In, intestine; Ma, edge of the mantle; Ne, neck; Nn, nuchal node; Os, osphradium; Sn, snout; Tmb, thick, wide mantle border; Tn, tentacle. Walker (1918). The entire head is black with melanin. On the neck of females, to the right of the mid-line, is a pronounced, fleshy pro- tuberance here called the nuchal node (Nn, Fig. 6). It is situated where the base of the penis is located in males. Mantle cavity The mantle cavity is the standard hydro- bioid type. Organ dimensions are given in Table 5. Three features are distinctive. The mantle collar (Ma, Fig. 6) has a very wide and thickened mantle border (Tmb). The osphra- dium is extremely long, 74 to 76% the length of the ctenidium’s base. The osphradium is set back from the anteriormost part of the mantle cavity because it does not overlap the wide mantle border (Tmb). In the female, the common genital opening (Cga) opens upward from the floor of the mantle cavity next to the neck, as a wide, prominent slit. There is no sexual dimorphism in number of gill filaments; POTAMOLITHUS SYSTEMATICS 83 they average 26.5 and 27.8 for males and females respectively (Table 5). Female reproductive system The uncoiled female without head and with kidney tissue removed is shown in Fig. 7. Organ dimensions are given in Table 5. The distinctive features are: 1) The pallial oviduct complex is squeezed over along the col- umellar muscle (in contrast to the usual posi- tion along the left ventrolateral aspect usually seen—Fig. 6, Davis & Greer, 1980, Pomati- opsidae: Triculinae); 2) the anterior end of the pallial oviduct (Apo) is far removed from the mantle collar (Ma); 3) the digestive gland (Di) covers the posterior part of the stomach (in contrast to being posterior to the stomach as in the Pomatiopsidae). The gonad is the generalized type (Fig. 11, Davis, 1980). The interrelationships of the bursa copulat- rix (Bu), seminal receptacle (Sr) and pallial oviduct (Ppo and Apo) are shown in Fig. 8. The orientation of these organs is the same shown in Fig. 7 except that, as the arrow TABLE 4. Shell measurements (mm) for five males and females of 4.0 to 4.5 whorls from which organ measurements were made. X + standard deviation (range). rr 3 Y Length 4.26 + 0.24 4.63 + 0.44 (3.92 — 4.48) (4.04 — 5.00) Width 3.69 + 0.17 4.08 + 0.30 (3.44 — 3.88) (3.72 — 4.4) Length of body 3.84 + 0.18 4.25 + 0.34 whorl (3.56 — 4.00) (3.76 — 4.48) Length of 3.0 + 0.14 3:13) == 0.14 aperture (2.80 — 3.20) (3.08 — 3.28) Width of 2.84 + 0.5 2.98 + 0.25 aperture (2.36 — 3.52) (2.64 — 3.24) TABLE 5. Length (mm) or number of non-neural structures of Potamolithus ribeirensis. X + standard deviation (range). N = number of specimens. * = significant difference between sexes, P < .05. N 3 N 2 Body 5 7.24 + 0.36 5 7.44 + 0.77 (6.80 — 7.80) (6.6 — 8.4) Buccal mass 5 1.08 + 0.07 5 121014 (0.96 — 1.14) (1.06 — 1.4) Gonad 5 SO En ES) 5 151 031107 (3.10 — 4.00) (1.4 — 1.6) Mantle cavity 6 2.34 + 0.15 5 2.44 + 0.22 (2.20 — 2.60) (2.20 — 2.80) Ctenidium 6 ea) Es ORS 5 2.14 + 0.23 (1.94 — 2.36) (2.00 — 2.56) Gill filaments: number 6 26.50 + 2.5 5 27.8 = 5:6 (23 — 29) (20:32) Penis 5 1.96 + 0.23 = (1.70 — 2.18) Prostate 5 1233720513 = (1.20 — 1.50) Pallial oviduct — 5 2.38 + 0.36 (21—30) Bursa copulatrix — 6 101-1015 (0.86 — 1.3) Digestive gland — 5 3.78 + 0.44 (3.2 — 4.4) Seminal receptacle — 6 0.52 = 0.09 (0.4 — 0.62) a __$_$_$_$$_$_$_$_$_$_$_$_$_$$$$$$$$$$$$$$$$$$$$$$$$LLKKKKK———————— 84 DAVIS & PONS DA SILVA FIG. 7. Uncoiled female without head and with only kidney tissue removed. Note the pallial oviduct (Apo-Ppo) moved over onto the columellar muscle. Note the digestive gland (Di) covering the stomach up to the anterior chamber (Ast). Apo, anterior pallial oviduct = capsule gland; Ast, anterior chamber of stomach; Bu, bursa copulatrix; Cl, columellar muscle; Di, digestive gland; Es, esophagus; Go, gonad; In,, posterior intestine; Ins, anterior intestine; Ma, mantle edge; Ov, oviduct; Ppo, posterior pallial oviduct = albumen gland. indicates, the complex has been lifted up and to the left, away from the columellar muscle, to reveal the pericardium (Pe). A portion of the ventral half of the pallial oviduct is re- moved to expose the interior of the albumen gland (Ppo) and capsule gland (Apo). The bursa copulatrix (Bu) and posterior half of the albumen gland (Ppo) are ventral to the style sac (Sts, Fig. 9) and loop of intestine (Fig. 7, п.) encircling the anterior end of the mantle cavity (Emc, Fig. 8A). The bursa copu- latrix complex has a common opening with the albumen gland and ventral channel (Vc) at the posterior end of the Vc. This common Opening is just posterior to the posterior end of the mantle cavity (Fig. 8A). The ventral channel is open to the pallial oviduct but it is not a simple ciliated gutter or channel directly opening into the cavity of the pallial oviduct. Rather, it is folded around towards the left dorsal aspect of the pallial oviduct as shown (Fig. 8A-C). The dashed lines across the pallial oviduct (Fig. 8A) indi- cate where slices were made through the anterior pallial oviduct. The corresponding cross sections are shown in Figs. 8B, C re- spectively. There is a thickening along the left ventro- lateral edge of the capsule gland lumen where the gland wall folds under and merges into the wall of the ventral channel. This elon- gate rod of tissue has been described in the literature as a longitudinal fold. In Hydrobia truncata studied by one of us (Davis) this fold is best seen by histological section as it is not much thickened and the laterally displaced ventral channel is not prominent. The ventral channel is tightly pressed against the wall of the pallial oviduct. In Potamolithus ribeirensis the fold is much thickened and rounded. It is POTAMOLITHUS SYSTEMATICS 85 / 1.0 mm FIG. 8. Female reproductive system in same orientation as Fig. 7. As indicated by the arrow, the pallial oviduct is lifted up and away from the edge of the columellar muscle (dashed line). A section of the albumen gland (Ppo) is cut away to reveal that part of the bursa copulatrix (Bu) covered by the Ppo. A section of the ventral wall of the pallial oviduct is removed to show the interior; in particular, the depressed oval cavity of the capsule gland (Apo). The thickened core of tissue called the bolster of the ventral channel (Vc) results from the fold of the pallial oviduct wall merging with the inner wall of the ventral channel (Vc). The dashed lines numbered 1 and 2 in Fig. A show where the pallial oviduct was cross sectioned (Figs. B and C). In section B one is looking posteriorly; in section C one is looking anteriorly. The white part of the sections is glandular tissue of the posterior-capsule gland and anterior section of capsule gland respectively. The lined portion of the section indicates the epithelial tissue of the ventral channel (Vc). Note that the Vc in C has separated from the wall of the Apo. Apo, anterior pallial oviduct = capsule gland; Bvc, bolster of ventral channel; Bu, bursa copulatrix; Cga, common genital aperture; Dbu, duct of bursa copulatrix; Dsr, duct of seminal receptacle; Emc, posterior end of mantle cavity; Gpd, gonopericardial duct; Ov, oviduct; Pe, pericardium; Ppo, posterior pallial oviduct; Sr, seminal receptacle; Vc, ventral channel. prominent in gross dissection when the pallial oviduct is cut in cross section with iridectomy scissors and examined (under dilute Bouins solution) under 50 x magnification. This much thickened longitudinal fold is here called the bolster of the ventral channel (Bvc, Fig. 8A). At the anterior end of the ventral channel, the walls merge to form an enclosed tube. The enclosed tube and common genital aperture are on the floor of the mantle cavity beside the neck (Fig. 6). The anterior end of the capsule gland can be slightly posterior to the genital aperture, abreast of it, or developed to extend beyond it to the interior edge of the wide border (Fig. 6). There is no anterior opening of the capsule gland. When the capsule gland was opened a smooth hollow depression could be seen (Fig. 8A) contrasting with the folded 86 DAVIS & PONS DA SILVA Pst Di Es Ast 1.0 mm FIG. 9. Ventral surface of the stomach. Ast, anterior chamber; Di, opening to digestive gland; Es, esophagus; In, intestine; Pst, posterior chamber; Sts, style sac. glandular elements crowding the albumen gland (Ppo). The bursa complex is understood by com- paring Figs. 7, 8A, and 10A, B. The bursa copulatrix (Bu) is % covered ventrally by the Ov Gpd albumen gland. The albumen gland is shown cut away in Fig. 8A. The bursa is a relatively large, sac-like organ, 42.4% the length of the pallial oviduct. The seminal receptacle (Sr) is pressed against the bursa, as shown in Fig. 8A, in 90% of the individuals; in 10%, it is dorsal to the bursa and thus out-of-sight, given the orientation of organs in Fig. 8A. The duct of the seminal receptacle (Dsr) is rel- atively elongate, connecting to the oviduct before the oviduct joins the duct of the bursa (Dbu) at the opening to the common genital groove (Vc). The seminal receptacle is pulled away from the bursa (Fig. 10A) to show the wide coil of the oviduct. The oviduct coil is not pigmented. A section of the coil directly posterior to the duct opening of the seminal receptacle glis- tens under direct illumination. The extent of —— FIG. 10. The bursa copulatrix complex in the same orientation as Figs. 7, 8. A. The seminal receptacle (Sr) has been pulled away from the normal position snug against the left ventrolateral, or dorsolateral edge of the bursa copulatrix (Bu) to show the oviduct coil circling dorsal to the bursa and duct of the bursa (Dbu). В. The duct of the bursa was cut at point A to remove the bursa from obstructing a view of the oviduct coil. The seminal receptacle was removed for the same purpose. The differentially- staining area called a false seminal receptacle is shown (Fsr). A, B, points for orientation, comparing Figs. 10A and 10B; Bu, bursa copulatrix; Dbu, duct bursa copulatrix; Dsr, duct seminal receptacle; Fsr, false seminal receptacle; Gpd, gonopericardial duct; Ov, oviduct; Sr, Seminal receptacle. POTAMOLITHUS SYSTEMATICS 87 FIG. 11. Male reproductive system: A. Head with typically coiled penis. B. Enlarged view of penis anterior. C. Dorsal view of penis showing typical arc of glands (С!) and notched appearance over the papilla (Pa) caused by the configuration of the preputial wings (Pw). D. Ventral view of tip of penis showing thickened collar called the preputial collar (Ppc) surrounding the papilla (Pa). E. The penis in Fig. A folded to the left to show the two massive columns of muscle rooting the penis into the columellar muscle. Note the concavity caused by the thickened rolls of muscles and the shaft of the penis. Ey, eye; Gl, glands; Pa, papilla (eversible); Pen, penis; Ppc, preputial collar; Pw, preputial wings; Sn, snout; T, tentacle. 88 DAVIS & PONS DA SILVA this glistening region is revealed by staining with methylene blue; it takes on a darker blue than the oviduct itself. This distinct region of the oviduct is clearly shown by cutting away the bursa and Sr (Fig. 10B). At first this par- ticular region of the oviduct appeared as a secondary seminal receptacle within the ovi- duct coil. With histological examination, however, no such Sr, was discerned. In this cavernous U-shaped part of the oviduct a Gpw GI section of wall with cuboidal non-ciliated cells contrasted with the remaining wall of col- umnar ciliated cells. While there were oocytes in the ovary, no sperm were seen in the seminal receptacle, bursa copulatrix, or coil of the oviduct of the eight females sectioned. This was apparently not the reproductive season. There is a discrete gonopericardial duct (Gpd, Figs. 8A, 10A, B). Vd Ejc FIG. 12. Penis on a slide under a coverslip viewed by compound microscope. A. Papilla out; B. Papilla withdrawn. White dots along penis (in A) are inclusions appearing white under direct illumination and black with transmitted light. Ci, cilia; Ejc, ejaculatory duct (under the epithelium of the neck); Gpw, greater preputial wing; Gl, glandular patch; Gle, glandular edge of the penis; Lpw, lesser preputial wing; Pa, papilla; Vd, vas deferens; Vd,, vas deferens from prostate through the penis. POTAMOLITHUS SYSTEMATICS 89 Male reproductive system Organ dimensions are given in Table 5. The head-neck region of a male is shown (Fig. 11), displaying the prominent penis coiled in typical fashion. In the living snail the unique feature is the wing-like preputial struc- tures here called preputial wings (Pw). These project as shown with the penis in the coiled position. The penial papilla is relatively long and can be retracted within the penis sheath. The base of the penis differs from that of any other hydrobioid we have examined. Lift- ing the blade of the penis, as seen in Fig. 11A, and pulling it to the left exposes the base as seen in Fig. 11E. The base is rooted by a horseshoe shaped muscle, i.e. there is a shal- low concave hollow within the arc of muscle as shown. Depending on how the tip of the penis is observed in the living animal, the preputium takes on different aspects. Looking down on the preputial wings, the preputium looks as if there were a U-shaped hollow at the end of the penis sheath (Fig. 11C). Turning the penis over 180°, the preputium looks like a thickened collar through which the papilla is extended (Figs MD); When the penis is examined with a com- pound microscope (on a slide under a cover- slip), the result is seen in Fig. 12A. The pre- putial wings are flattened and appear as areas of extended columnar cells. There are irregular tufts of stiff vibratile cilia arising from the preputial wings. The tip of the penis is shown with papilla extended (Fig. 12A) and withdrawn (Fig. 12B). Near the tip of the penis, behind the preputial ring, is a concen- tration of glands (С!) beneath the surface of the epithelium. These are spherules that look black under transmitted light. The vas defer- ens undulates through the penis near the convex edge (Fig. 12A). The thickened ejac- ulatory duct (Ejc) is not embedded in the base of the penis, but in the neck. The uncoiled male without head and kidney tissue is shown in Fig. 13A. The prostate is squeezed against the columellar muscle. Dis- tinctive features are: 1) the gonadal lobes are massive; the gonad does not have a slender, delicate branching system; 2) the anterior gonadal lobes are ventral to the posterior chamber of the stomach, i.e. the gonad is not restricted behind the stomach; 3) the gonad- vas deferens configuration is otherwise of the general type (Davis, 1980); 4) the seminal vesicle (Sv) was hardly developed in 90% of the males dissected. There was hardly a sin- gle fold or loop of the vas deferens (Vd,) near the gonad signifying the seminal receptacle (Fig. 13A). In 10% (older individuals as evi- denced by shell parietal callus development) the seminal vesicle was a pronounced coil of tubes dorsal and obliquely angled to the gonad (Fig. 13B); 5) the prostate overlies the posterior end of the mantle cavity; 6) the posterior prostate is ventral to the entire style sac, thus hiding it from view in Fig. 13A; 7) the posterior prostate is flattened and fan- shaped, contrasting with the inflated bean- shape of the rest of the prostate; 8) the an- terior vas deferens (Vd,) leaves the prostate as shown (Fig. 13A, B); the generalized condi- tion (Davis, 1980). Digestive system Of importance here are the salivary glands, stomach, and radula. The salivary glands are paired, each gland a simple elongate tube that lies dorsal to the nerve ring. The stomach is shown in ventral view (Fig. 9) with a charac- teristically raised section of the anterior cham- ber (Ast), and a distinctive shape of the poste- rior chamber (Pst). The posterior chamber has a smooth, rounded posterior contour, i.e. there are no folds or distended projections of this chamber (contrast the so-called appendix of the stomach—e.g. Radoman, 1977, fig. 2A, Hydrobia acuta). One can see, inside the ventral surface of the posterior chamber, raised ridges shown as dark lines in Pst (Fig. 9). The radula is typically taenioglossate (Fig. 14). Radular statistics are given in Table 6 and cusp formulae in Table 7. Note the con- cave hollow ventral to the basal cusps of the central tooth (Fig. 14B); the lateral angle of the central tooth is widely spread; the in- nermost pair of basal cusps arise from the face of the tooth. There is a pronounced posterior projection of the face of the lateral tooth (Fig. 14C, D). Nervous system The nervous system is standard risso- acean. Statistics on neural structures are given in Table 8. Of note are: 1) the elongate cerebral commissure; 2) the elongate pleuro- supraesophageal connective yielding an RPG ratio of 0.58; 3) the relatively long pleuro- subesophageal connective. This is an open nervous sytem, i.e. the ganglia are separated by elongate connectives and commissures. 90 DAVIS & PONS DA SILVA Glo Sv FIG. 13. Male reproductive system. A. Ventral view of uncoiled snail without head and with kidney tissue removed. B. Some lobes of the gonad are removed to show the position and nature of coils of the seminal vesicle (Sv). Ast, anterior chamber of the stomach: Emc, posterior end of the mantle cavity; Es, esophagus; Cl, columellar muscle; Glo, gonadal Sperm-producing lobe; Go, gonad; In, intestine; Ma, mantle edge; Pr, prostate; Sv, seminal vesicle; Vd,, posterior vas deferens; Vd,, anterior vas deferens. POTAMOLITHUS SYSTEMATICS 91 FIG. 14. Radula. A, portion of radular ribbon; B, central tooth; C, left lateral tooth; D, right lateral tooth; E, outer marginal. Е, inner marginal. C-F are at the same magnification. 92 DAVIS & PONS DA SILVA TABLE 6. Radular characteristics of Potamolithus ribeirensis. Measurements in mm. X, mean; + standard deviation; ( ), range; N = number studied. Radular character 3 2 э a 8 Length 13010110 1.48 + 0.11 1410514 (1.13 — 1.37) N = 5 (1.27 — 1.59) N = 8 (aloe) = 1.58) Width 0.20 + 0.01 0.22 + 0.02 0.20 + 0.02 (0.18 — 0.22)N=5 (0.20 — 0.25) М = 8 (0.18 — 0.25) Rows (no.) 9322182 96:5 == 6.1 95:21 ==6:5 (80 — 102) М = 5 (87 — 106) М = 8 (80 — 106) Rows forming (по.) 64255145 85 ss 240) To EE РА (5-9N=5 (1 SINS (54173) Central tooth (width) 0.0398 + 0.003 0.0460 + 0.002 0.0429 + 0.004 (0.0370 — 0.0440) М = 5 (0.0440 — 0.0480) N = 5 (0.0370 — 0.0480) 0.48 e Lacunopsis 8 T. bollingi L. naticoides P. lapidaria Po. ribeirensis To. ventricosa -0.26 -0.50 1 Н. truncata e -0.74 ea a A es a aS A A -0.71 -0.58 -0.45 -0.32 -0,19 -0.06 0.07 0.20 0.33 0.46 FIG. 15. Ordination diagram following three-dimensional scaling, showing the distribution of eight taxa along axes 1 and 2. Taxa are grouped into sets (solid lines) and sub-sets (dashed lines). See text for explanation. POTAMOLITHUS SYSTEMATICS 93 TABLE 7. Formulae for the most common cusp arrangements of the four radular teeth types of Potamolithus ribeirensis. Ten radulae were examined. Tooth Formula 5-1-5 2(3)-(3)2* 4-1-4 one side 5-1-5 other side Central Lateral Inner Marginals 25 Outer Marginals 30 % teeth with formula Other types 100% — 5-1-5 one side 30% 4-1-5 other side 40% 4-1-5 21% 22, 28-32 18% 42% 20% 24-27, 31 18% 34-36 12% 12% 11% “The third basal cusp is vague and insignificant in a few specimens, a raised ridge in some specimens, and was not discerned in most specimens. TABLE 8. Length (mm) or ratio of neural structures of Potamolithus ribeirensis. N = number. X, mean. + standard deviation (range). N Cerebral ganglion 5 0.34 + 0.09 (0.28 — 0.50) Cerebral commissure 5 0.17 + 0.06 (0.08 — 0.24) Pleural ganglion: right (1) 5 0.14 + 0.02 (0420.16) Pleural ganglion: left 5 0.15 + 0.01 (0.14 — 0.16) Pleuro-supraesophageal 5 0.46 + 0.09 connective (2) (0.36 — 0.6) Supraesophageal 5 0.18 + 0.02 ganglion (3) (0.16 — 0.20) Osphradiomantle nerve 5 0.07 + 0.07 (0202) Pleuro-subesophageal 5 0.04 + 0.03 connective (0 — 0.08) Subesophageal ganglion 5 0.18 + 0.02 (0.16 — 0.20) Pedal ganglion 5 0.25 + 0.04 (0.22 — 0.28) Pedal commissure 4 0.04 + 0.03 (0 — 0.08) Statocyst diameter 1 0.12 Osphradium 4 0.87 + 0.09 (0.80 — 1.0) RPG ratio* 5 0.58 + 0.04 (0.53 — 0.63) “2/1 + 2 +3 RESULTS: ANALYSIS OF RELATIONSHIPS Multivariate analysis The data matrix is shown in Table 9. Three ordination diagrams compare OTU place- ments when principal components (PC) 1 x 2, 1 X 3, and 2 X 3 are compared (Figs. 15-17). PC one (axis 1) has 55.08% of the variance; PC two has 32.70%; PC three has 12.22%. The stress after fifty iterations was 0.008. The matrix correlation was 0.941. In the principal components analysis the first five PCs involved 94.19% of the variance as follows: 1) 35.51%; 2) 23.95%; 3) 15.65%; 4) 11.90%; 5) 7.18%. Character loading on the first four PCs is given in Table 10. PC one: Some fifteen characters load on factor one with scores of 0.482 or more; ten characters load with scores >0.70. This axis is structured with snails having globose shells (character 1), wide parietal callus (2), keel formation (3), a powerful foot (6), gonad over- lapping the stomach (17), pallial oviduct with a ventral channel (18), with opening of pallial oviduct lateral, not at the anterior end (19), eggs without sand covering (30)—to the far left. Snails to the far right (OTUs 4-7) have ovate-conic to turreted shells, without keels, without squat heads or powerful feet, with gonads posterior to the stomach, with sper- mathecal duct separate from the pallial ovi- duct, with oviduct opening at the anterior end, and with eggs covered by sand. Lacunopsis (OTU no. 8) shares shell and 0 с с 0 с | “0 0 (<) sıuad jo aseq ui *(1) 49euU ui ‘(0) auou :}эпр Auoyeinselg “91 | 0 0 0 0 | Е 0 ajosnu 1е|эшп|оэ UO Jano pazaanbs jonplao feed ‘aje]S014 ‘< | 0 0 0 0 0 0 | 0 (1'0) зима “epim siuag pl 0 0 L 0 0 | 0 0 (1'0) ended э|а!лэлэ цим siuag “El L | | | | | | 0 (1'0) 4100} jo ade, Woy sdsno ¡eseq sed элош 10 | :ц}00} ¡e1ua) “21 | | | | | | | 0 (1‘0) sdsno jeseq jo sured Элош JO $ :YJOO} [EAU “LL | 0 | 0 0 | | | (1‘0) ayebuoja :wnipesydso ‘01 0 0 0 0 0 0 0 | (10) eno = paiydoupedAy sey ya] :э|эезиэ1 ‘6 = 0 0 0 0 0 0 0 | (1'0) эрезиз} еше :эвиеи ‘8 < 0 0 0 | | 0 0 0 (1'0) э$еэ.о jepad :1004 ‘7 Q ‘ ‘ . . т | 0 0 0 0 | | 0 (10) Inuamod ‘apim 4004 ‘9 5 0 | | | | 0 | | (L‘0) A Japua¡s-ajeBuoja :sajoejua | ‘G o) | 0 0 0 0 | | 0 (1'0) yenbs ‘реола :реэн ‘+ un ‘ = Г 0 0 0 0 | | 0 (10) 198% = e шло} 0} Spud} :||aUS “€ | 0 0 0 0 | | 0 (1'0) seus /Sn¡eo jeyaued apim 114$ ‘с с | | 0 0 с с | (<) esoqolB “(1) эшоэ-эуело ‘(0) pajauny :119ys “1 ‘dds je ıBunjog euade ESO911JUDA enepide] SISUAJIAQU зарюэцеи ejesuna] 1э}оелецэ sisdoun2e7 BIND] BIND PIYOIWO | sısdonewog Ssnyy/OWEJOS snydA¡Boyy7] PIGOIDAH 8 д 9 G v € с | exe] ‘91n}219]I| BU} Ul вер Jo YOR] JO вер UBAIB э}е}$ JajoeJeyo pawnsaud , ‘sey = | ‘элец jou зэор = 0 ‘(L‘0) U] 'siajoeJeyo pue exe} биоше $Аащзиоце|эл ssasse о} sisAjeue э]еиелципше 104 siajoeJeyo ze Аа (SN1O) Exe} 8 JO XUJEW ‘6 3719VL 94 95 POTAMOLITHUS SYSTEMATICS xt — (10) (xipuadde, =) aouauIuoid J01193sod pue pjo} UJIM YIBUIO]S JO лэашецо JO119ISOY (1'0) зшелб pues Аа рэлэлоэ $663 (L‘o) juasald apou ¡eyann (10) Áuneo эдиеш jo pus 1ouajue ye à siajua uuads (L‘o) pajebuoja esing (10) wnipseoued sJajua JONP ¡eoayyewads (10) ]siM] JONPIAO adA} sisdoun2e7 (10) 1109 yONPIAQ (1) эюе}9эээл ¡eunuas Алериоээ$ Aq paorjda Jo (0) Juaseid a19e1d89a1 ¡euas (1) jonp wuads ем Jo (0) ÁpoaJp ulof esinq jo опр pue зэпрмо (10) 181S0q цим JONPIAO (2) JONPIAO ¡ented о} рипоа Авубц jou pue jonpino je] -Jed jo azis 0} элце|эл эбле| ‘(L) yONpIAO jeijjed о эбра |е1э}е!| -одиэл ¡eipau о} рипоа ÁnyBn ‘JONPIAO jeijjed JO 2ZIS 0} элце|эл ¡¡eus ‘(0) auou :¡auueyo |едиэл (10) diy 1ouaque je jonpino jeijjed jo Buiusdo (10) JONPIAO ¡er ped wos} payesedas jonp BIA WOISÄS siajua WISS :ajeuway (1'0) yoeulo]s jo Jaquieyo 109}504 sdejiano peuob 2 jo ed (1'0) разедиээиоэ wWajsks зполлэм “ZE "LE ‘0€ "6° ‘8c ‘Le ‘9c ‘Ge vd ‘ec ‘CS ‘be "05 61 "81 Alt 96 -0.71 DAVIS & PONS DA SILVA 0.29 e Lacunopsis H. truncata e L. naticoides T. bollingi To. ventricosa , P. lapidaria Po. ribeirensis -0.58 -0.45 -0.32 = 0.5119 -0.06 0.07 0.20 0.33 0.46 FIG. 16. As in Fig. 15, except axes are 1 x 3. TABLE 10. Factor loading of 32 characters for each of the first four principal components that collectively account for 87.01% of the variation in this study. Principal components Character 1 2 3 4 1 0.827 — 0.397 0.354 — 0.106 2 0.840 — 0.488 0.051 0.169 3 0.576 —0.417 —0.194 0.363 4 0.840 — 0.488 0.051 0.169 5 — 0.520 0.601 —0.035 0.353 6 0.840 — 0.488 0.051 0.169 7 — 0:551 0.170 — 0.582 0.380 8 0.111 0.806 0.475 — 0.269 9 0.111 0.806 0.475 — 0.269 10 0.740 — 0.008 0.330 — 0.284 11 OA — 0.806 — 0.475 0.269 12 —0.111 — 0.806 — 0.475 0.269 13 0.265 SOA — 0.431 —0.700 14 0.549 0.073 0.029 0.709 15 0.840 — 0.488 0.051 0.169 16 — 0.482 0227 — 0.289 — 0.310 17 0.864 0.481 0.004 — 0.056 1 e POTAMOLITHUS SYSTEMATICS s 8 Lacunopsis e H. truncata L. naticoides To. ventricosa -0.60 -0.45 -0'. 31 -0.16 -0.02 0.13 0.28 0.42 FIG. 17. As in Fig. 15, except axes are 2 x 3. TABLE 10 (Continued) Principal components Character 1 2 3 4 19 — 0.864 — 0.481 — 0.004 0.056 20 0.875 0.248 — 0.230 — 0.282 21 0.605 — 0.174 — 0.498 — 0.522 22 — 0.551 0.170 — 0.582 0.380 23 0.076 — 0.614 0.544 0.060 24 0.371 0.634 — 0.516 0.285 25 — 0.218 — 0.630 0.628 —0.017 26 —0.218 — 0.630 0.628 —0.017 27 —0.011 —0.094 — 0.654 — 0.492 28 0.371 0.634 — 0.516 0.285 29 0.605 — 0.174 — 0.498 — 0.522 30 — 0.881 0.077 0.358 — 0.143 31 0.504 0.671 0.385 0.337 32 0.549 0.073 0.029 0.709 98 DAVIS & PONS DA SILVA head-foot character states with OTUs 2 and 3 at the far left yet shares five character states (17—20, 30) loading heavily on this axis with OTUs 4-7 at the far right. Accordingly, this taxon is to the left of OTUs 4-7. Hydrobia (OTU no. 1) shares shell and head-foot character-states with OTUs 4-7, and covers eggs with sand as do OTUs 4-8, but shares character states 17-19, and 22 with OTU nos. 2, 3. Accordingly, this taxon, as with Lacunop- sis, is to the left of OTUs 4-8. Character states that relegate Potamolithus and Lithoglyphus to the far left are those shared or unique to one of them, i.e. the condensed nervous system (32), nuchal node (29), bolster of the pallial oviduct (21), eja- culatory duct in the neck (16), and blunt penis (14). A similar arrangement of ducts involving the pallial oviduct, sperm groove, lateral op- ening of the pallial oviduct, and duct of the bursa joining the oviduct (22) place OTUs 1-3 to the left of the second axis. PC two: Eleven characters load on this axis with a score >0.60; four with a score >0.80. At the bottom (Hydrobia, OTU no. 1) the character-states are: the mantle has a tenta- cle (8); the left tentacle has hypertrophied cilia (9); the central tooth has fewer than three basal cusps (usually only one); the stomach has a posterior prominence; the basal cusp arises from the lateral angle, not the face of the tooth (12). Towards the top of axis two snails have no mantle tentacle, are without hypertrophied cilia, the central tooth has three or more basal cusps with the innermost pair arising from the face of the tooth; the stomach does not have a posterior prominence. Lacu- nopsis (OTU no. 8) is at the top of the second factor axis as it alone has accessory seminal receptacles. OTUs 7 and 8 are at the top of the second factor axis, as they alone have a distinctive oviduct twist (contrasted with ovi- duct coil typical of the Hydrobiinae, Pomatiop- sinae, and Lithoglyphinae); they have a sper- mathecal duct that enters the pericardium, a condition not occurring in the other taxa. PC three: Eight characters load on this axis with a score >0.50; no score is >0.654. Only four characters score higher on this axis than axes one or two (nos. 7, 13, 22, 27). In- terpretation of this axis is weak. Those two taxa (OTUs 3 and 8) with the unique char- acter states are separated at either end of the axis. Lacunopsis (OTU no. 8) has the secon- dary seminal receptacles (23) while Potamo- lithus (OTU no. 3) has the nuchal node (29) and bolster of the pallial oviduct (21). Taxa with the pedal crease (7) are low on the axis (OTU nos. 4 and 5). Otherwise, given the character loading score for this axis, we see no other clear biological interpretation. Subsets: Hydrobia and Lacunopsis are not included in subsets. Potamolithus and Lithog- lyphus are in a set clearly separated along axis one from the sets and subsets that group Pomatiopsis and Tomichia (one subset) and the two species of Tricula (other subset). Character correlations: A number of characters are highly correlated and cluster together at >0.90. The most prominent group involves shell shape (1), degree of parietal callus (2), squat head (4), wide and powerful foot (6), and prostate or pallial oviduct squeezed over towards the columellar muscle (15). The concentrated nervous system (32) and blunt penis (14) are thus correlated. The pallial tentacle (8) is highly correlated with hypertrophied cilia on the left tentacle (9). The pedal crease (7) correlates thus with the sperm duct (22). The central tooth characters are highly correlated (11, 12), as are the Lacunopsis-type oviduct twist (25) with the spermathecal duct entering the pericardium (26). Cladistic analysis A set-theory nesting of the eight taxa in question on the basis of unique and presum- ably unreversed characters sensu Wilson (1965) is given in Fig. 18. The data for Lithog- lyphus came from Krull (1935), Krause (1949), and Radoman (1966). The data for the Pomatiopsinae and Triculinae came from Davis, 1979, 1980; Davis & Greer, 1980. The data for Hydrobia were summarized in Her- shler & Davis (1980), with additional data from Radoman (1973, 1976). The largest set A is defined by the most generalized character-states, i.e. shared by all taxa in the smaller sets. Character-states that serve to define sets are given in Table 11, where the letters correspond to the lettered sets of Fig. 19. A cladogram was structured from the sets by rooting the tree at A (Fig. 19), a taxon presumably having the generalized charac- ters listed (Tables 11, 12). Derived character- states are gains or losses of certain states numbered 2 to 8 on Fig. 20 and in Table 12. The question marks indicate that we do not have sufficient data to link lineages B, and Bo. К has been argued that these lineages do not share an immediate common ancestor A POTAMOLITHUS SYSTEMATICS 99 TABLE 11. Character-states defining sets shown in Fig. 18. eee A. Rissoacean grade of organization, including species without pedal tentacles, with closed pallial oviducts, with pallial oviduct comprised of albumen and capsule glands merging directly into each other, with basal cusps on central tooth, with seminal receptacle posterior to bursa copulatrix, where the penis has one duct, where the operculum lacks internal pegs, projections, and calcium smears. B,. 1) Sperm carried to bursa copulatrix in a grooved channel or duct open to the pallial oviduct: 2) female genital aperture of posterolateral to anterior end of pallial oviduct. C,. 1) One or two pairs basal cusps on rachidian tooth; 2) basal cusps arise from lateral angle; 3) hypertrophied cilia on left tentacle; 4) posterior chamber of stomach with posterior protuberance and Ole ES) GONNA СО wise cites Re Hydrobia s.s. Cz. 1) Three or more pairs of basal cusps on rachidian tooth; 2) some of basal cusps arise from face of tooth, not lateral angle; 3) no hypertrophied cilia on left tentacle; 4) stomach without protuberance or Os ©) NU RER PR ro oe: Lithoglyphinae D,. 1) Nuchal node on neck of female; 2) pleuro-supraesophageal connective elongate; 3) penis with protractile papilla; 4) penis slender; 5) sperm groove offset, folded around ventral pallial ОО носовое ee aliens RU lo cede see a Aa Potamolithus Dz. 1) No nuchal node; 2) pleuro-supraesophageal connective short or absent; 3) penis without papilla; 4) penis wide and blunt; 5) sperm groove opens directly into lumen of pallial М oh o qe iron, ee ee ee и Lithoglyphus Bz. 1) Sperm carried in enclosed spermathecal duct not connected to the pallial oviduct; 2) opening of female system at anterior end of pallial oviduct. C3. 1) Spermathecal duct opens at anterior end of mantle cavity; 2) oviduct coil retained; 3) pedal creases4) maintains gonopericardialliduet 2... coins ot eee ......0. Pomatiopsinae D3. 1) Short bursa copulatrix; 2) sperm duct arises from duct of bursa............. Pomatiopsis D4. 1) Long bursa copulatrix; 2) sperm duct arises from bursa some distance from anterior end of SET 23 a A Me eR tr aren a ie Tomichia C,. 1) Sperm enters spermathecal duct from posterior of mantle cavity; 2) no oviduct coil; no pedal lease 24) mnosgonopentCandial AUCtr ее. Triculinae 05. Spermathecal duct does not enter pericardium.................................. T. aperta Оз. Spermathecal duct enters pericardium. E,. 1) Seminal receptacle standard, joins oviduct; 2) no accessory seminal receptacles....... оао nee die ee I EEE Triculinae: “Tricula” bollingi E:. 1) Standard seminal receptacle lost; 2) with several secondary seminal receptacles....... оо обо о D AN Ch due ele ml rio occ woe tlds ВС Lacunopsis FIG. 18. Set theory solution sensu Wilson (1965) to nest taxa on the basis of unique characters given in Table 11. 100 DAVIS & PONS DA SILVA LITHOGLYPHINAE POMATIOPSINAE TRICULINAE HYDROBIA LITHOGLYPHUS POTAMOLITHUS POMATIOPSIS TOMICHIA TRICULA LACUNOPSIS SS NATICOIDES RIBEIRENSIS LAPIDARIA VENTRICOSA T.APERTA T. BOLLINGI Cy D, Do D3 D4 Ds Es E> FIG. 19. Cladogram based on rooting taxa of the sets shown in Fig. 18, based on shared derived character-states of Tables 11, 12. The question marks indicate uncertainty about the two lineages sharing an immediate common ancestor (see text). FIG. 20. Shells of species converging on Potamolithus in shell shape. A, Somatogyrus aureus (ANSP 69284); shell length 6.31 mm; other shells printed at the same magnification; B, Fluminicola seminalis (ANSP 175283); C, Lithoglyphus naticoides (ANSP 345100); D-G, Lacunopsis sphaerica (ANSP 345100). POTAMOLITHUS SYSTEMATICS 101 TABLE 12. Definition of numbered sets of character-states shown in Fig. 19. Numbers 1-2 indicate the maintenance of generalized characters, while 3-9 are sets of shared derived characters. Use with Table 11. he Character-states that are generalized and presumed primitive are: shell small (<5.0 mm), ovate-conic: animal with oviduct coil, with gonopericardial duct, stomach without posterior folds or protuberances of the posterior chamber, central tooth with one pair of basal cusps off lateral angle, seminal receptacle joins the oviduct, pallial oviduct with ciliated gutter to transport sperm, hypertrophied tentacle cilia; with pallial tentacle. 2. Same as 1 above. . Development of fold and protuberance (= appendix of Radoman) of posterior stomach chamber. . Loss of hypertrophied tentacle cilia, loss of pallial tentacle, increased complexity of basal cusps on central tooth, specialized shell shape. . Spermathecal duct separated from pallial oviduct, thus replacing ciliated gutter of pallial oviduct; increased complexity of basal cusps on central tooth of radula (parallel with same trend at 4 above). . Loss of gonopericardial duct, loss of oviduct coil, sperm enter female reproductive system at posterior end of mantle cavity (possibly a primitive character found in ancestral taxa at point 5). . Step-like mode of progression, pedal crease, suprapedal fold, eyes in enlarged bulges at base of tentacles; spermathecal duct extends to anterior end of mantle cavity. . Seminal receptacle joins common sperm duct (illustrated in Davis, 1980, and Davis & Greer, 1980). . Specialized shell shape, specialized central tooth, unique female reproductive system (see Davis, 1979; Davis, 1980; Davis & Greer, 1980). (Davis, 1979, 1980). It is possible that lineages В, and Bos derived from different marine ancestors of rissoacean grade organ- ization. DISCUSSION Higher category relationships There are still insufficient data for hydro- bioids world-wide to attempt a definitive treat- ment such as that by Radoman (1973), who created the superfamily Hydrobioidea with nine families and eleven subfamilies while considering only eastern European taxa. Hy- drobioids are defined as an artificial grouping of fresh and brackish water snails that resem- ble each other in generalized shell, opercular, radular, and penial characters seen in taxa considered Hydrobiidae pre-1979 (Davis, 1979, 1980). We need to assess the cladistic rela- tionships among hydrobioids through a care- ful analysis of characters and their states, geological and paleontological records, zoogeographic patterns, and ecological fac- tors affecting morphological character-states. In this paper we restrict higher category rela- tionships to discussing the Hydrobiidae with subfamilies Hydrobiinae, Lithoglyphinae, and Nymphophilinae. The Pomatiopsidae with its subfamilies, Pomatiopsinae and Triculinae, have been fully discussed elsewhere (Davis, 1979, 1980). Potamolithus does not have the morphol- ogy of the Pomatiopsidae. Further, no hydro- bioid from Africa, India, Southeast Asia (in- cluding Southwest China) has, to our knowl- edge, Potamolithus-like morphology. Pota- molithus shares more presumably derived character-states with Lithoglyphus of Europe than with any other taxon for which we have anatomical data and is therefore classified as Hydrobiidae: Lithoglyphinae. We consider the character-state differences separating Litho- glyphus naticoides and P. ribeirensis to justify generic status but realize that more species of Potamolithus must be studied to determine the character-states shared among them. For example, is the nuchal node common to all species of Potamolithus? Is the pleuro- supraesophageal connective relatively long in all species, etc.? The concentrated nervous system of Lithoglyphus, based on data for L. naticoides is not sufficient reason, of itself, to place Lithoglyphus in a separate subfamily or family as has been done (Radoman, 1973). It is also necessary to study L. apertus Kuster and L. fuscus Pfeiffer in detail over the broad suite of characters presented here, to better define the genus Lithoglyphus. 102 DAVIS & PONS DA SILVA Convergence and defining higher taxa Convergence is probably the most un- derestimated phenomenon in molluscan sys- tematics (Davis, 1979). Unless it is detected and isolated, assessments of relationships will be in error (Cain & Harrison, 1960). Some thirty species and subspecies of Potamolithus have been monographed (Pilsbry, 1911). The shells of these taxa are of a size and shape variation such that many of the species resemble species of the pomatiopsid genera Lacunopsis and Jullienia of the Mekong River in Southeast Asia (Brandt, 1974; Davis, 1979). The squat head and penis morphology of Lithoglyphus lapidum (= P. ribeirensis) (von Ihering, 1895) likewise indicate a possi- ble close cladistic relationship between Pota- molithus and the Mekong River taxa. It is these similarities that prompted this study. It is clear that these hydrobioid radiations converge. With the aid of multivariate an- alysis, several character-states correlate with snails having a globose to cap-shaped shell, i.e. squat head, powerful wide foot, etc., characters presented earlier. The shell size, shape and correlated character-states are associated with ecological factors, namely liv- ing in a high energy environment, clinging to rocks and boulders. Accordingly, such con- vergent characters should not be used to define clades, or if used, it should be with caution. It should be noted that there are species with globose shells living in other microhabitats such as on aquatic vegetation or plowing along on a sandy mud substrate. Many of these are small Hydrobiidae: Amni- colinae with shell lengths <4.0 mm; others are Hydrobiidae: Nymphophilinae. The ovate- conic shell form is generalized while the glo- bose shape is specialized (Davis, 1979, 1980). As with the X? analysis of the relation- ship of specialized character-states with de- grees of current for species in the Mekong River (Davis, 1979, p. 73), it appears that there is a lack of species with generalized states living in swift current and not the ab- sence of species with specialized states from slow water environments. It is evident from the ordination diagrams that although Potamolithus, Lithoglyphus, and Lacunopsis converge in seven character- states, they are highly divergent overall. It is also evident that Potamolithus and Litho- glyphus sharing a set quite removed from Lacunopsis, are more closely associated to each other than Lacunopsis is to other taxa of the Pomatiopsidae. Lacunopsis, of all the taxa represented has the most specialized character-states (Davis, 1979; 1980; Davis & Greer, 1980). The Pomatiopsidae and their subfamilies Pomatiopsinae and Triculinae have been de- fined (Davis, 1979, 1980). The study of Tricu- linae indicates a considerable range of shell shape and sculpture patterns within a sub- family, tribe and genus. For example, in the genus Lacunopsis, shapes range from ovate- globose, to globose, neritiform, and Calyptraea-form (Davis, 1979). We have dis- cussed above how certain suites of char- acters can correlate with shell shape. Osphra- dial length in some instances appears associ- ated with ecological factors (Davis, 1979). It follows, therefore, that one must beware of defining higher taxa by scoring taxa on the basis of similarity of character-states without consideration of their adaptive significance. It has been argued that genera are not simply artificial groupings of species (Davis, 1981). Adaptive radiation studies indicate that one or more novel innovations, morphological or physiological, may be associated with a new adaptive zone. Success in that zone may result in speciation with species occupying different niches. Morphological or physiolog- ical changes associated with different niche dimensions serve to define the species while the features, common to all species of the radiation, that are associated with the new adaptive zone serve for recognition of the genus. Thus defined, a genus is a first order radiation of an adaptive radiation (Davis, 1981). In this regard, the concentrated nervous system and blunt penis type without eversible papilla, together with a generalized type of ciliated sperm groove of the pallial oviduct, apparently serve to define Lithoglyphus. The generalized, more open nervous system, more slender penis with eversible papilla, in- novative folded ciliated sperm groove and bolster, plus nuchal node may (if more spe- cies have these features) serve to define Potamolithus. The character-states that com- bine these genera in the same subfamily are: 1) the lack of a fold and protuberance of the stomach's posterior chamber; 2) there are two or more pairs of basal cusps on the central tooth; one or more innermost pairs of basal cusps arise from the face of the central tooth, not the lateral angle; 3) the penis is simple, i.e. with one duct and without lobes or special complex glandular protuberances; 4) POTAMOLITHUS SYSTEMATICS 103 eggs are not covered in a case covered with sand grains; 5) the shell is globose, with correlated morphological character-states. The last character-state is, in light of the preceding discussion, the weakest because we have emphasized the potential for shell shape variation within a genus, tribe, and subfamily. Thus far, the globose shell is associated with the other four character- states serving to define the Lithoglyphinae. We will maintain this character-state for the subfamily until such time that a hydrobioid species with turreted or ovate-conic shell is found that also has the other four character- states of the Lithoglyphinae. The Hydrobiinae include those genera that have: 1) a fold and protuberance of the stom- ach's posterior chamber; 2) one or more pairs of basal cusps arising from the lateral angle, but none from the face of the tooth (Davis, 1979); 3) penis simple, with one duct, with or without small glandular appendage on the concave curvature; 4) eggs with or without sand covering; 5) shell shape variable. Stom- ach morphology is not known for numerous key taxa, so the use of the fold and pro- tuberance (so-called appendix) character- state introduced by Radoman (reviewed by Radoman, 1973) has yet to be fully evaluated. The presence of one or two seminal recepta- cles, the loss of seminal receptacles with sperm stored in the coil of the oviduct, or the storage of sperm in secondary seminal re- ceptacles are considered possible within a single subfamily (see Davis, 1979; Thomp- son, 1979). Some of the character-states listed in Table 9 are diagnostic for the genus Hydrobia and do not serve to define the Hydrobiinae. These are the pallial tentacle, hypertrophied cilia on the left tentacle, a single pair of basal cusps on the central tooth of the radula, and a simple glandular lobe of the penis in most species. The first two character-states are most probably associated with life in marine or brackish water. Given these genera- specific diagnostic features, there is much less divergence in ground-plan between the Hydrobiinae and Lithoglyphinae of the Hydro- biidae than between the Pomatiopsinae and Triculinae of the Pomatiopsidae. Thompson (1977, 1979) presented data for Marstonia of the United States and Nym- phophilus of Mexico, respectively. He argued that these and allied genera were monophy- letic and belonged to the Hydrobiidae: Nym- phophilinae, of which Radoman’s (1973) Orientaliidae were synonymous. Of the nine character-states presented by him, all but two are those pertaining to the family diagnosis or diagnosis for the Hydrobiinae given above. The two truly diagnostic character-states are 1) the bilobed nature of the penis and, 2) various complex raised glandular structures called apocrine glands. Taxa with these nine character-states are found in Europe as well as North America. In addition to the nine listed character-states, the genera of the Nym- phophilinae have similar embryonic shell mi- crosculpture different from that of other North American hydrobioid taxa (Thompson, 1979). If the embryonic shell sculpture character- state is consistently valid, and if the stomachs of these genera lack the fold and pro- tuberance seen in the Hydrobiinae, then there would perhaps be sufficient character-state divergence from the Hydrobiinae to justify subfamilial status. The Orientaliidae of Rado- man lack the stomach fold and protuberance, as do the Lithoglyphinae. Zoogeographic relationships Parodiz (1969) stated that there were three possible hypotheses to account for the origin of Potamolithus: 1) the group is ancient, cos- mopolitan, derived from Gondwanaland; 2) the genus dispersed to South America from ancestors of North American origin; 3) the genus is modern, derived from marine an- cestors. Parodiz discounted the Gondwana- land origin because of the scant fossil record, i.e. a few Paleocene and lower Eocene occurrences in Chile and Argentina with no other records until the Recent. Parodiz fa- vored the third hypothesis. Pilsbry (1911) wrote a definitive essay on the origins of non-marine molluscan faunas of South America. He considered Potamolithus to have an Austral-South American origin and stated that nothing in the distribution of mol- lusks would lead to the hypothesis that South Africa had ever been connected with Ant- arctica and thereby indirectly with southern South America. He noted that Fluminicola of western North America, Lithoglyphus of Eu- rope, Pachydrobia [sensu lato, non Davis, 1979] Lacunopsis and Jullienia of Indo-China had historically been grouped together in the Lithoglyphinae because of similarities of shell shape. However, he considered Potamolithus to be most closely related to Petterdiana within the Amnicolidae: Amnicolinae, the lat- ter genus from Tasmania-Australia. The other 104 DAVIS & PONS DA SILVA above-mentioned genera belonged to the Amnicolidae: Lithoglyphinae [Amnicolidae = Hydrobiidae of present usage]. The character used to differentiate the two subfamilies was the number of cusps on the marginal teeth: few and large = Lithoglyphinae; many (>20) and small = Amnicolinae. Pilsbry (1911) stated that, as he found the number of basal cusps so variable in many genera, he did not attach much importance to that character. Pilsbry (1911) found the Austral-South American generalized track in part from the individual tracks of Potamolithus, the Potamolithus-Petterdiana connection, and the freshwater clam Diplodon. He noted that Potamolithus was lacking from India, South- east Asia, or the Far East. Pilsbry (1911) was too hasty in discounting an African connection for faunal elements in common between South America and Aus- tralia or between Africa and southern South America. Of particular importance are the corneous operculated Ampulariidae, Planor- bidae, and Mutelidae that have closely related taxa in South America and Africa with those elements reaching North America (Mexico, Florida) by dispersal from the south. The pomatiopsid Aquidauania* of south- western Brasil is closest in relationships with South African Tomichia and Australia Coxiella (Davis, 1979, 1980). Potamopyrgus has been reported from Africa (Brown, 1980), but this African taxon must be studied an- atomically; it may not be Potamopyrgus. Addi- tionally, the African genera Lobogenes and Soapitia that are considered Hydrobiidae (Brown, 1980) must be examined as certain shell features suggest a relationship to Pota- molithus. In summary, there is a Gondwana- land freshwater component shared between Africa and South America. While no snail with a Lithoglyphinae-type morphology has been found in Africa, one must have data for Lobogenes and Soapitia to be sure that no named African taxa have such morphology. There are possible relationships with North American taxa made more probable because of the morphological similarity between Lithoglyphus (Fig. 20C) and Potamolithus. Strong candidates for cladistic association are Fluminicola (Fig. 20B), Somatogyrus (Fig. 20A) from southeastern U.S.A., and Gillia from eastern U.S.A. The remarkable similarity among shell types of Potamolithus, Lacunop- sis, Lithoglyphus, Fluminicola, and Somato- gyrus is seen by comparing Figs. 3, 4, and 21. Of the North American genera the radula and penis type shown or discussed for Gillia (Wal- ker, 1918; Stimpson, 1865; and Somatogyrus (based on S. tenax, Thompson, 1969) come the closest to Potamolithus. Presence of snails with the female reproductive system morphology of the Hydrobiidae, exclusively in freshwater systems of North America, has been demonstrated, i.e. Marstonia and Nym- phophylus (Thompson, 1977, 1979). We pre- dict that the placement of Fluminicola, Gillia, and Somatogyrus depressa (Tryon) (the type- species) in the Hydrobiidae based on total morphology, will be justified. Stimpson (1865) followed by Walker (1918) illustrated and/or discussed the penis of Somatogyrus to be broad and bifid. This description was based on Melania isogona Say. Baker (1926, 1928) pointed out that the penis of Somatogyrus is simple and not bifid. He erected the genus Birgella for taxa with the penis and radula of Melania isogona. The type-species is B. sub- globosa (Say) and M. isogona was relegated to B. subglobosa isogona. Birgella is included in the Hydrobiidae Nymphophilinae (Thomp- son, 1979). In summary of zoogeographic relationships this study proves conclusively the presence of genuine Hydrobiidae in South America. Mor- phological evidence conclusively proving the presence of the Hydrobiidae in southern con- tinents had previously been lacking (Davis, 1979, 1980). We reject Parodiz’s third hypoth- esis. Given the antiquity of hydrobioids (Per- mian of Africa, Knight et al., 1960), their wide- spread distribution, and considering the Paleocene-Eocene records from South Amer- ica, Potamolithus most likely did not evolve from marine ancestors in the late Tertiary, and certainly not in the Recent. The Potamolithus-Petterdiana hypothesis of rela- tionship must be proven, but if correct, it would demonstrate an old Austral-South American relationship. The close morpholo- gical relationship to Lithoglyphus of Europe, the character of marginal cusp number not- withstanding, probably indicates divergence of the Lithoglyphinae before the breakup of Pangaea. The vicariance-dispersal patterns affecting the modern distribution of the Litho- “Malek (1983) has shown that Aquidauania Davis, 1979 is a synonym of /diopyrgus Pilsbry, 1911 by presenting anatomical data for the type-species of Idiopyrgus, |.e. |. souleyetianus Pilsbry, 1911. This clarification was made after this paper was in press. POTAMOLITHUS SYSTEMATICS 105 glyphinae cannot, however, be ascertained without detailed morphological data for the American, African, and Australia-Tasmanian taxa discussed above, as well as for numer- ous other taxa in Mexico, Central and South America about which we know nothing except for shell and radula data. American taxa for which we have few data are among those listed in Taylor (1966) and Thompson (1968). ACKNOWLEDGEMENTS The support of the Zoobotanical Founda- tion of Rio Grande do Sul, Brasil and its president, Professor W. Thome is gratefully acknowledged. We thank Dr. Arno Lise of the same institution for his help in collecting this species. The SEM work, preparation of radu- lar data, histology and photography of shells were carried out by Mrs. Lynn Weidensaul Monarch. The illustrations were drawn by G. M. Davis; illustration shadings and maps were done by Ms. Mary Fuges. The manu- script was read and criticized by Drs. Winston Ponder and Fred Thompson. LITERATURE CITED BAKER, F. C., 1926, Nomenclatorial notes on American fresh water Mollusca. Transactions of the Wisconsin Academy of Science, Arts, and Letters, 22: 193-205. BAKER, F. C., 1928, The Freshwater Mollusca of Wisconsin pt. 1. Gastropoda. Bulletin of the Wis- consin Geological and Natural History Survey No. 70, Madison, IL, xx + 507 p., 28 pl. BRANDT, R. A. M., 1974, The non-marine aquatic Mollusca of Thailand. Archiv für Mollusken- kunde, 105: 1-423. BROWN, D. S., 1980, Freshwater Snails of Africa and Their Medical Importance. Taylor & Francis, London, x, 1-487 р. CAIN, A. J., 1964, The perfection of animals. In: CARTHY, J. E. & DUDDINGTON, C. L., eds., Viewpoints in Biology No. 3, p. 36-63, Butter- worths, London. CAIN, A. J. & HARRISON, G. A., 1960, Phyletic weighting. Proceedings of the Zoological Society of London, 136: 1-31. DAVIS, С. M., 1979, The origin and evolution of the Pomatiopsidae, with emphasis on the Mekong River Triculinae. Monograph of the Academy of Natural Sciences of Philadelphia, No. 20: i-vili, 1-120. DAVIS, G. M., 1980, Snail hosts of Asian Schisto- soma infecting man: evolution and co-evolution. In: BRUCE, J. & SORNMANI, S., eds., The Mekong Schistosome, Malacological Review Supplement No. 2, р. 195-328. DAVIS, G. M., 1981, Different modes of evolution and adaptive radiation in the Pomatiopsidae (Prosobranchia: Mesogastropoda). Malacologia, 21: 209-262. DAVIS, С. M. & CARNEY, W. P., 1973, Description of Oncomelania hupensis lindoensis, first in- termediate host of Schistosoma japonicum in Sulawesi (Celebes). Proceedings of the Academy of Natural Sciences of Philadelphia, 125: 1-34. DAVIS, G. M. & GREER, F. J., 1980, A new genus and two new species of Triculinae (Gastropoda: Prosobranchia) and the transmission of a Malay- sian mammalian Schistosoma sp. Proceedings of the Academy of Natural Sciences of Phil- adelphia, 132: 245-276. DAVIS, G. M., KITIKOON, V. & TEMCHAROEN, P., 1976, “Lithoglyphopsis” aperta (Gastropoda): morphology, evolutionary relationships, and the transmission of Mekong schistosomiasis. Mala- cologia, 15: 241-287. HERSHLER, R. & DAVIS, G. M., 1980, The morphology of Hydrobia truncata (Gastropoda: Hydrobiidae): relevance to systematics of Hy- drobia. Biological Bulletin, 158: 195-219. HOAGLAND, K. E. & DAVIS, G. M., 1979, The stenothyrid radiation of the Mekong River. 1. The Stenothyra mcmulleni complex (Gastropoda: Prosobranchia). Proceedings of the Academy of Natural Sciences of Philadelphia, 131: 191-230. IHERING, H. VON, 1885, Zur Kenntniss der Gat- tung Lithoglyphus. Malakozoologische Blatter, new ser., 7: 96-99, 1 pl. KNIGHT, J. B., BATTEN, В. L., YOCHELSON, Е. L. & COX, L. R., 1960, in MOORE, R. C., ed., Treatise on Invertebrate Paleontology. Part |. Mollusca I: 310-324. University of Kansas Press, Lawrence. KRAUSE, H., 1949, Untersuchungen zur Anatomie and Okologie von Lithoglyphus naticoides (C. Pfeiffer). Archiv für Molluskenkunde, 78: 103- 148. KRULL, A., 1935, Anatomische Untersuchungen an einheimischen Prosobranchiern und Beiträge zur Phylogenie der Gastropoden. Zoologische Jahrbucher (Anatomie), 60: 399-464. MALEK, E. A., 1983, The South American hydro- bioid genus /diopyrgus Pilsbry, 1911. Nautilus, 97(1): 16-20. PARODIZ, J., 1969, The Tertiary non-marine Mol- lusca of South America. Annals of Carnegie Museum, 40: 1-242. PILSBRY, H. A., 1896, Notes on new species of Amnicolidae collected by Dr. Rush in Uruguay. Nautilus, 10: 86-90. PILSBRY, H. A., 1911, Non-marine Mollusca of Patagonia. In: SCOTT, W. B., ed., Reports of the Princeton University Expeditions to Patagonia, 1896-1899, 3(2), pt. 5: 513-633. RADOMAN, P., 1966, The zoogeographical and phylogenetic interrelationships of the genera Lithoglyphus and Emmericia. Bulletin du Museum d'Histoire Naturelle, Belgrade, ser. B, 21: 43—49. 106 DAVIS & PONS DA SILVA RADOMAN, P., 1973, New classification of fresh and bra[c]kish water Prosobranchia from the Bal- kans and Asia Minor. Prirodnjacki Muzej u Be- ogradu. Posebna Izdanja. Editions hors série Knjiga, 32: 1-30. RADOMAN, P., 1976, Hydrobiidae auf der Bal- kanhalbinsel und Kleinasien. Archiv fur Mollus- kenkunde, 107: 203-223. ROLF, F. J., KISHPAUGH, J. & KIRK, D., 1972, NT-SYS: Numerical taxonomy system of multi- variate statistical programs. Technical Report, State University of New York, Stony Brook, N.Y. STIMPSON, W., 1865, Researches upon the Hy- drobiinae and allied forms; chiefly made upon materials of the Smithsonian Institution. Smithsonian Miscellaneous Collections, 201: 1— 59. THOMPSON, F. G., 1968, The aquatic snails of the family Hydrobiidae of Peninsular Florida. Univer- sity of Florida Press, Gainesville, 268 p. THOMPSON, F. G., 1977, The hydrobiid snail genus Marstonia. Bulletin of the Florida State Museum, Biological Sciences, 21: 113-158. THOMPSON, F. G., 1979, The systematic rela- tionships of the hydrobioid snail Nymphophilus Taylor 1966 and the status of the subfamily Nymphophilinae. Malacological Review, 12: 41— 49. WALKER, B., 1918, A synopsis of the classification of the fresh-water Mollusca of North America, North of Mexico. Miscellaneous Publications of the University of Michigan Museum of Zoology, No. 6: 1-213. WILSON, E. O., 1965, A consistency test for phylogenies based on contemporaneous spe- cies. Systematic Zoology, 14: 214-220. APPENDIX: The Identity of Potamolithus ribeirensis Pilsbry The description of Potamolithus ribeirensis Pilsbry, 1911 was based on six shells from Brasil, Sao Paulo State, Ribeira River, Ypor- anga (= Iporanga) collected by von Ihering in 1908 (Fig. 1). The syntypes had ANSP cata- log number 103076. One of us (Davis) selected a lectotype (Fig. 3A; ANSP 103076) and recataloged the remainder as paralec- totypes (ANSP 353441). The species was described as having 3.5 convex whorls, the last globose; length 3.5 mm, width 3.4 mm, and length of aperture 2.7 mm. The species differed from P. lapidum by its broad col- umella and small size. The shells from the Feitoria river some 800 km SSW of the Ribeira River closely resem- ble those of the type-series except that they have as adults, 4.0 to 4.5 whorls and are larger. Statistics of shell parameters between the populations can be compared by examin- ing Tables 2 and 3. The populations were compared using the multivariate analysis program NT-SYS (Rohlf et al., 1972). The six syntypes and ten in- dividuals from the Feitoria River (chosen at random from the most common size classes of 4.0 and 4.5 whorls) were compared over eight shell characters (Table 13). Distance coefficients were generated using UPGMA. We used the minimum spanning tree (MST) and subsets components. Character correla- tions were subjected to Principal Component Analysis (PCA) with components extracted until eigenvalues became less than 1.0. A transposed matrix of the first three principal components was used to produce a matrix of OTU projections in the principal component space. The resulting PCA-based configura- tion was used as the initial configuration for nonmetric multidimensional scaling (MDS). A cophenetic correlation was calculated by comparing the distances between OTUs in the PCA- and MDS-spaces and the Q-mode taxonomic distance matrix. The ordination diagram based on MDS is freed from the constraints of phenogram construction. The results are shown in Figs. 21 and 22. The cophenetic correlation for the phenogram based on three-dimensional scaling was 0.81; the cophenetic correlation for MDS x the Q-mode taxonomic distance matrix was 0.99. In PCA the first component accounted for 83.7% of the variance; the second component 11.95%; the third, 1.74%. Eigenvalues were: first component, 6.69; second component, 0.96. All but character 7 loaded highly on the first axis (values > 0.879). It is evident, as discussed below, that character 7 (loading 0.439 on axis 1) loaded highly on axis 2 of Fig. 22. The stress for three-dimensional scal- ing was 0.009. It is clear from the phenogram and ordination diagram that the individuals of the two populations do not separate into dis- crete clusters. As expected, individuals sepa- rate along the first axis on the basis of size, largest to the far left, smallest to the far right. Only the extreme whorl numbers follow this pattern, i.e. 4.5 whorls to the far left; 3.0 whorls, far right (no. 3). Individuals with 3.5 or 4.0 whorls are not separated in order along axis 1. Separation along axis 2 is clearly based on width or columellar callus, the more slender below axis 1; the wider, above axis 1. The lectotype (no. 1) is close to the center of the diagram. Minimum spanning tree (Prim Network) and subsets in relationship to the POTAMOLITHUS SYSTEMATICS 107 2.190 1.890 1.590 1.290 0.990 0.690 0.390 0 FIG. 21. Phenogram showing phenetic relationships among 16 individuals from the type-series and from our study site (see Table 13 and Appendix). showing position of 16 individual shells FIG. 22. Ordination diagram following three-dimensional scaling, | along axes 1 and 2. Dotted lines are sets. See Table 13 and Appendix. 108 DAVIS & PONS DA SILVA TABLE 13. Matrix of sixteen OTUs and eight shell characters to facilitate a comparison of the type- population with the Feitoria River population using multivariate analyses. Measurements in mm. Shell character 1 2 3 4 5 6 7 8 Length Width Width Whorl Shell body Length Width columellar penultimate OTU no. length whorl Width aperture aperture callus whorl Ribeira River 1 Lectotype 3.5 3.68 3.40 3.36 2.92 2.44 0.48 1.0 2 Paralectotype 3.5 3.68 3.32 3.32 2.96 2.48 0.60 1.0 3 Paralectotype 3.0 2.28 2.04 2.28 1.88 1.60 0.36 0.64 4 Paralectotype 3.5 3.32 2.96 3.04 2.68 2.24 0.52 1.0 5 Paralectotype 3.5 3.64 3.24 3.20 2.68 2.32 0.52 1.04 6 Paralectotype 3.5 2.84 2.56 2.60 2.24 1.88 0.44 0.76 Feitoria River 7 4.5 4.64 4.08 4.04 3.44 2.96 0.52 1.28 8 4.0 3.68 3.28 3.08 2.80 2.20 0.40 1.08 9 4.5 5.00 4.40 4.24 3.12 3.24 0.48 1.44 10 4.0 3.60 3.20 3.20 2.84 2.20 0.48 1.0 11 4.0 3.40 2.96 2.96 2.56 2.24 0.44 1512 12 4.0 3.92 3.52 3.60 3.16 2.60 0.48 1.08 13 4.0 3.52 3.12 3.12 2172 2.28 0.44 1.0 14 4.5 4.60 4.00 4.00 3.40 3.00 0.48 1.28 5 4.5 4.36 3.80 3.88 3.20 3.00 0.44 1.20 16 4.0 3.68 3.24 3.36 2.80 2.48 0.52 1.12 distribution of OTUs in the diagram indicate that a single species is involved where dif- ferences among individuals involve size dif- ferences and width of columellar callus. In- dividuals of the type-population apparently reach maximum size at 3.5 whorls and have an average length at that whorl stage of 3.44 + 0.37 mm (n = 5). One third (33%) of the individuals have a relatively slender col- umellar callus. Individuals of the other popula- tion reach 4.5 whorls with an average length of 4.65 тт + 0.26 тт (п = 4) but have predominantly 4.0 whorls with an average length of 3.46 + 0.43 (n = 50) and about 70% have a relatively slender columellar callus. Because the Feitoria population matures at 4.0 to 4.5 whorls and has a generally more slender columellar callus we distinguish this population as race-B contrasted with the type- population as race-A. Pilsbry (1911) figured one shell (plate XLlb, fig. 4) from a series collected by von Ihering from the Ribeira Riv- er but from a locality called Hiririea; this shell (Fig. 4A) has a “wide, lunate, concave umbi- lical area, defined by an acute black keel, the columella as wide, as in P. ribeirensis” (ANSP 103068). This specimen is eroded but un- doubtedly of 4.0 to 4.5 whorls. The shell is 4.2 mm long. The periostracum has a green- ish color. The large size indicates that some populations in the Ribeira River may reach maturity at 4.0 to 4.5 whorls and have a size similar to that of race-B. This must be de- termined. The strong keel seen in some in- dividuals is not unusual, as discussed in the body of this paper. MALACOLOGIA, 1984, 25(1): 109-141 NORTH AMERICAN FRESHWATER SNAIL GENERA OF THE HYDROBIID SUBFAMILY LITHOGLYPHINAE Fred G. Thompson Florida State Museum, University of Florida, Gainesville, Florida 32611, U.S.A. ABSTRACT The classification of the North American genera of Lithoglyphinae is reviewed, based on anatomical and conchological characters. Five genera are recognized: Gillia Stimpson, 1865, Fluminicola Stimpson, 1865, Somatogyrus Gill, 1863, Clappia Walker, 1909, and Lepyrium Dall, 1896. The North American genera are conservative in their anatomies. Primary morphological differentiations involve radular and shell characters. The North American genera are demon- strated to have evolved through trophic specializations and microhabitat specializations. Lepyrium Dall, 1896, formerly considered a monotypic family, is closely related to Clappia and Somatogyrus. Birgella Baker, 1926, is removed from the Lithoglyphinae, where it has been placed by authors, and is demonstrated to be in the Nymphophilinae. The following species are described in detail: Lepyrium showalteri (Lea, 1861), Somatogyrus rheophilus n. sp., Gillia altilis (Lea, 1841). A neotype is designated for the latter species. Key words: Gastropoda, snails, Hydrobiidae, Lepyriidae, Lithoglyphinae, Lepyrium, Clappia, Somatogyrus, Gillia, Fluminicola, systematics, evolution. INTRODUCTION This paper discusses the systematic rela- tionships between the North American litho- glyphine genera: Lepyrium, Somatogyrus, Clappia, Gillia, and Fluminicola (Class GAS- TROPODA, Subclass PROSOBRANCHIA, Order MESOGASTROPODA, Family HYDROBIIDAE, Subfamily LITHOGLYPHINAE). A sixth genus, Antrobia Hubricht, 1972, is placed by Burch & Tottenham (1980: 100) in the Lithoglyphinae. It is a monotypic genus from a cave in Mis- souri; the anatomy remains undescribed and Antrobia is omitted from further discussion in this paper. Another genus, Cochliopina Morri- son, 1946, traditionally has been associated with the Lithoglyphinae. Hershler (in press) shows that it is a genus of the Littoridininae. Birgella Baker, 1926, is another genus that has been confused with this subfamily, even as recently as 1981 (Clarke). It is in the NYM- PHOPHILINAE, as is discussed in Appendix B. This study stems from two independent investigations. The first was an attempt to determine species-group characteristics with- in Somatogyrus, a genus containing many species (Burch & Tottenham, 1980: 104— 106). The study was tabled temporarily be- cause very little anatomical diversity was dis- covered among the species examined. In- dependently | examined the anatomy of Lepyrium showalteri (Lea), a snail previously placed in a monotypic family of uncertain affinity within the MESOGASTROPODA (Pilsbry & Olsson, 1951). Its soft anatomy was found to be hardly distinguishable from that of Somatogyrus. These two genera have very dissimilar shells, but they have in common similar habitats. Lepyrium and most Somato- gyrus live on rocks and boulders in high- energy rivers. The habitat deployments among these two genera focus on the adap- tive radiation of the Lithoglyphinae in eastern North America. In order to clarify the limits of this basic radiation other relevant genera were examined. The results of these studies are presented herein. MATERIAL AND METHODS Anatomical descriptions and illustrations in this paper are based upon the following speci- mens: Lepyrium showalteri (Lea). Two lots of about 100 specimens each, collected June 21, 1978 (UF 31343) and June 22, 1978 (UF 31342) in the Little Cahaba River, 2.4 km upstream from the Cahaba River, Bibb Co., Alabama by F.G.T. Relaxed with menthol crystals, fixed in Bouin’s solution and pre- served in 70% ethanol. Somatogyrus rheophilus n. sp. (described below). One series of thousands of speci- (109) 110 THOMPSON mens collected October 21, 1973, in the Flint River, 9.7 km SW of Lincoln Park, Upson Co., Georgia by F.G.T. (UF 40511). Relaxed with menthol crystals, fixed in Bouin's solution and preserved in 70% ethanol. Gillia altilis (Lea). One series of about 400 specimens collected June 1, 1980 in Lake Waccamaw, Columbus Co., North Carolina by Hugh J. Porter (UF 27550). Fixed un- relaxed and preserved in 70% ethanol. Fluminicola nuttalliana (Lea). One series of 23 specimen collected July 20, 1974 in the Satsop River at Satsop, Grays Harbour Co., Washington by Dennis R. Paulson (UF 34813). Fixed unrelaxed and preserved in 70% ethanol. Dissections were made in 70% ethanol un- der a WILD M-2 dissecting microscope. Serial sections were made at 10 um and stained in 10% Harris’ hematoxylin stain. Radulae were cleaned in a saturated solution of potassium hydroxide and examined with a HITACHI S- 415A scanning electron microscope at the Department of Zoology, University of Florida. Museum abbreviations are given in Appen- dix C. ANATOMY OF THE GENERA The genera Lepyrium, Clappia, Somatogy- rus, Gillia and Fluminicola are placed in the subfamily LITHOGLYPHINAE. The conical, or depressed-conical, or globose-conical, or neritoid-shaped shells have in common spiral sculpture on the protoconch (Figs. 39-42). Other protoconch sculpture may also be present. Other aspects of the shell are not distinctive at the subfamily levels among North American hydrobiid genera, although the lithoglyphines tend to be stocky species with relatively heavy shells. A high degree of anatomical uniformity ex- ists among the North American LITHOGLYPHI- NAE. Principal morphological differences occur in trophic structures and in the shell. Minor diversity also occurs in the male and female reproductive systems. Because of this basic uniformity it is convenient to describe the anatomy of Lepyrium and compare other genera with it. Lepyrium Dall, 1896 Lepyrium is a monotypic genus endemic to the Coosa and Cahaba Rivers in Alabama where it lives on boulders in high energy shoals. Lepyrium showalteri (Lea) is de- scribed in detail because its systematic affini- ties have been in question since it was dis- covered. Its shell is described in Appendix A. The shell (Figs. 59—62) is peculiar because it has a depressed spire and an expanded, flattened body whorl, giving it a neritid appearance. The first whorl of the protoconch is smooth, except for a few short spiral fur- rows below the periphery (Fig. 39). The neomelanian type operculum (Fig. 7) is mod- ified in shape to conform to the enlarged aperture. This is accomplished by rapid ex- pansion of the last whorl from the paucispiral nucleus, a minor elaboration of the basic con- dition that also exists in Somatogyrus and Gillia (Fig. 52). Its radular teeth are modified for feeding on small-sized food particles. Lepyrium is the most divergent of the North American LITHOGLYPHINAE because of these specializations. Other aspects of its anatomy are conservative and indicate a close relation- ship to Clappia and Somatogyrus. External morphology and color. Foot broad and rounded (Figs. 1-3). Operculigerous lobe overlaps on each side, and edge of oper- culum overlaps extended foot on left side and posteriorly. Mucous groove present along an- terior edge of foot. Food grooves and epitae- nial folds absent on body and snout. Col- umellar muscle extending into shell for about a quarter whorl. Columella muscle insertion short but wide, extending transversely nearly the complete width of body whorl (in other genera the insertion is narrower). Mantle col- lar complete around body, without tentacles or papillae. Mantle cavity of males semicircu- lar in saggital section; triangular in females, bounded posteriorly by pallial gonoduct, per- icardium and stomach. Gill (gf) consisting of 19-20 lamellae that are arranged in an oblique series on mantle wall. Lamellae triangular in shape (Fig. 8), with greatest height along intestine. Osphradium (og) long, narrow, L-shaped. Excurrent and incurrent siphons absent. Kidney small, broadly quad- rangular, overlying posterior left corner of mantle cavity. Tentacles long and slender in life, and actively beat substrate in alternate strokes as animal moves about. Snout highly extendable and constantly sampling substrate in moving animal. Mantle jet black (Fig. 1) on all surfaces, completely opaque to all internal morphology. Snout and tentacles dark gray, fading posteri- orly and laterally to light gray on sides of body NORTH AMERICAN LITHOGLYPHINAE 111 FIGS. 1-3. Lepyrium showalteri (Lea). Fig. 1. Specimen denuded of shell showing pigmentation of mantle. Fig. 2. Male with melanin removed showing internal anatomy. Fig. 3. Female with melanin removed showing internal anatomy. Legend: ast = anterior chambers of stomach; di = digestive gland; gf = gill filaments; goo = ovary; got = testis; in = intestine; ki = kidney; og = osphradium; op = operculum; pe = penis; pov = pallial oviduct; pr = prostate; sty = style sac. and nape. Tentacles also with diffuse, sca- tered xanthophores. Sole very light gray. Center of muzzle white. Penis white. Digestive system. Typically hydrobioid in its configuration (Fig. 4). Two elongate claviform Salivary glands (sg) enter posterior buccal mass along dorso-lateral edges, and extend posteriorly over top of nerve ring. Oesopha- gus (es) entering stomach on left side of posterior chamber (Fig. 5, ast). Stomach with a single opening into digestive gland (odi) posterior to gastric shield (gs). Style sac (sty) slender, at anterior end of stomach on left dorso-lateral surface. Caecae absent. In- testine (in) leaving anterior chamber on left side, passing beneath style sac and up to mantle where it continues diagonally forward to right corner of mantle collar. Anterior end of stomach abutting against kidney, pericardium and pallial gonoduct. In females it is covered by the digestive gland (di) on all sides except anterior end and anterior half of outer wall. In males digestive gland is more restricted due to distribution of testes. Fecal pellets (Fig. 6) cylindrical, tapered at both ends, and spirally coiled with 4-10 whorls. Pellets about 0.40-0.57 mm long and 0.15-0.20 mm wide. Pellets oblique in in- testine at about 45° to longitudinal axis of body. Radula. Taenioglossate (Figs. 15-18, about 1.8mm long, containing 140-149 transverse rows of teeth (7 specimens ex- amined). Cusps rapidly worn, barely distin- guishable on distal third of ribbon. Central tooth (Fig. 15) about 130 jm wide, broadly trapezoidal in shape with a long mid-basal projection; lateral angles with a low ridge bearing 9-11 small, nearly uniform, acumi- nate basocones on each side (Fig. 16), in contrast to allied genera in which the baso- cones are enlarged toward the top of the series; dorsal margin weakly reflected with a slightly enlarged mesocone, and 22-24 ecto- cones on each side. Lateral tooth (Fig. 17) with a long, narrow, laterally projecting, flexed shaft and a strong slender cusp-like basal projection on face of tooth; laterals with 18-22 nearly uniform acuminate cusps. Inner mar- ginals weakly sigmoid with about 50 small cusps. Outer marginals long, slender, with about 35 very small cusps (Fig. 18). Female reproductive system. (Figs. 3, 11). Ovary (goo) consisting of 3—4 large subequal lobes, each of which is partially divided into smaller lobules. Ovary lying against posterior surface of pallial oviduct (bursa copulatrix) and along right side of stomach, but sepa- rated from latter by a thin zone of digestive gland tissue (di); ovary not occupying apical 112 THOMPSON 0 47 0 | 2 тт ML e A AP А FIGS. 4-8. Lepyrium showalteri (Lea). Fig. 4. Lateral view of digestive system excluding digestive gland. Fig. 5. Ventral view of stomach interior. Fig. 6. Fecal pellet. Fig. 7. Operculum. Fig. 8. Single lamella of gill Legend: ast = anterior chamber of stomach; es = esophagus; gs = gastric shield; in = intestine; odi = opening to digestive gland; sg = salivary glands; sty = style sac. NORTH AMERICAN LITHOGLYPHINAE 113 whorl of digestive gland. Primary oviduct (ov) pallial oviduct. A very short gonopericardial relatively stout, passing diagonally forward ductis present. Primary oviduct enlarging and along ventral side of digestive gland, then forming a loop along mesad side of albumen mesad to right of oesophagus and style sac gland; top of loop folded down between rest of and along pericardium at junction of latter with loop and albumen gland (Fig. 11). A short 2 mm oa | 14 FIGS. 9-14. Lepyrium showalteri (Lea). Fig. 9. Male reproductive system. Fig. 10. Diagrammatic cross- section of penis at arrow in 9. Fig. 11. Female reproductive system. Fig. 12. Right side of coil of primary oviduct. Fig. 13. Coil of primary oviduct unfolded to show seminal receptacle. Fig. 14. Cross-section of pallial oviduct at arrow in 11. Legend: ag = albumen gland; bu = bursa copulatrix; cg = capsule gland; етс = posterior wall of mantle cavity; goo — ovary, got = testis; obu = bursa copulatrix duct; ov = oviduct; pe = penis; pr = prostate; sr = seminal receptacle; sv = seminal vesicle; vc = ventral channel; vd, = proximal vas deferens; vd, = distal vas deferens. 114 THOMPSON narrow seminal receptacle (sr) projects up- ward on albumen gland side of descending limb of loop, and is partially covered by the loop so that it can be viewed only by serial section or by teasing the loop free from the albumen gland and unfolding the loop (Figs. 12-13). Seminal receptacle about 140 um long. Descending segment of loop entering albumen gland (ag) where it forks to form ventral canal (vc) of pallial oviduct (pov) lead- ing forward to vagina, and a short broad duct leading posteriorly to bursa copulatrix (bu). Bursa copulatrix large and saccate, lying against posterior end of albumen gland. Ven- tral canal spiral in cross-section (Fig. 14), and continuous with lumen of capsule gland (cg). Albumen gland and capsule gland intricately coalesced to ventral canal, and together form the pallial oviduct. Pallial oviduct confined postero-laterally by posterior wall of mantle cavity (emc); irregularly pyriform in shape when viewed from above; bounded along its left side by intestine. Pallial oviduct terminat- ing within mantle cavity just posterior to man- tle collar and anus. Capsule gland and albu- men gland not clearly differentiated super- ficially. Albumen gland occupying posterior third of pallial oviduct and consisting of tightly coalesced large glandular cells. Capsule gland consisting of smaller, and more com- pact cells. Ventral canal extending about a fourth of its length beyond anterior end of capsule gland. Eggs laid singly in capsules on hard sur- faces. Capsules hemispherical and attached to substrate by a narrow hyaline collar. Width of hemispherical capsule about 0.7 mm; width of collar about 0.05 mm; total width of capsule and collar about 0.8 mm. Male reproductive system (Figs. 2, 9). Tes- tis (got) very large and completely covering posterior 1.5 whorls; overlying entire digestive gland, stomach and posterior third of pros- tate; testis consisting of numerous lobes that subdivide into small lobules that form a mar- bled pattern on outer surface. Lobes dis- charging into the primary sperm duct which forms a highly convoluted seminal vesicle (sv) on ventral surface of testis; this continues into posterior vas deferens (vd,). The latter bearing 3—4 sigmoid coils near its middle. Posterior vas deferens passing mesad across ventral surface of body beneath junction of stomach and prostate, and then upward along right margin of pericardium, oesophagus and style sac to enter ventral side of prostate (pr) near its middle. Prostate with a relatively deep intestinal groove along its mesad curvature. Prostate completely posterior to transverse wall of mantle cavity (emc), but partially over- lapping cavity. Anterior vas deferens (vdi) leaving ventrocolumellar side of anterior edge of prostate, passing vertically down side of mantle cavity, and following an irregular sig- moid course across side of nape to base of penis (pe). Penis originating on right side of nape beneath mantle collar; long, flattened, blade-like, biconvex in cross section (Fig. 10) and unpigmented. When contracted, the penis is folded posteriorly and to the left within the mantle cavity. It bears a small patch of minute glands along its distal left margin on its ventral surface. Clappia Walker, 1909 The genus was founded on a single spec- ies and was distinguished from Somatogyrus because of its open umbilicus (Fig. 68), its large opercular nucleus, and some character- istics of the radula. The sculpture of the pro- toconch remains unknown. The type-species, C. clappi Walker, 1909 (=C. umbilicata [Wal- ker], 1904) from the Coosa River in Alabama apparently is extinct. A second species, C. cahabensis Clench, 1965, from the Cahaba River may also be extinct due to pollution from coal strip-mining in the area. Both spec- ies lived on high energy shoals. External morphology and color. | have ex- amined the dried bodies of 24 paratypes of C. clappi (ANSP 95037). The mantle is uniform black, similar to Lepyrium and most Somato- gyrus. Walker (1909: 90) states that the an- imal is black. Presumably this refers to the foot, snout and nape, as well as the mantle. Dried males have a flattened, blade-like, un- pigmented penis. Radulae were extracted from two specimens. Radula. There are about 56—57 transverse tooth rows (Fig. 19). All teeth are character- ized by having more numerous and smaller cusps than do those of Somatogyrus, but not to the extent that occurs in Lepyrium. Central tooth (Fig. 20) with moderately extended lateral angles; mesocone small, slender, acu- minate, bordered on each side by 6-7 ecto- cones on a low ridge near outer edge of lateral angles; basocones subequal, gradu- ally increasing in size dorsally. Lateral tooth (Fig. 21) with a strongly flexed, slender shaft; mesocone reduced in size, slender, bordered by 7-9 small entocones and 11-12 slender, subequal ectocones; basal projection long NORTH AMERICAN LITHOGLYPHINAE 113 AY sn AAA FIGS. 15-26. SEM photographs of radulae. Figs. 15-18. Lepyrium showalteri (Lea). Figs. 19-22. Clappia umbilicata (Walker). Figs. 23-26. Somatogyrus depressus (Tryon). Enlargements: Figs. 15, 17, 20, 21, 24, 25, 26 x356; Figs. 19, 23 x95; Figs. 18, 22 х475; Fig. 16 х950. Legends: im = inner marginal tooth; om = outer marginal tooth. and slender. Inner marginal tooth (Fig. 22, im) with about 50 very small slender cusps. Outer marginal tooth (Fig. 22, om) with about 35 small slender cusps. Walker (1909: 89) stated that there are about 50, but the cusps are so small his count may have been only a rough approximation. Somatogyrus Gill, 1863 The genus contains numerous species in eastern North America (Burch & Tottenham, 1980: 104—110). They have in common con- ical or ovate-conical shells 3-6 mm high. The protoconch has numerous fine spiral threads. Punctate sculpture may occur in addition (Fig. 40). Walker (1915) stated that some species have pitted sculpture only. | have not been able to confirm this among the material | examined. The columella generally is thick- ened. The thickness and structure of the col- umella, and the nature of the umbilicus pro- vide useful characteristics for grouping spe- cies. The soft anatomy of the type-species, S. depressus Tryon, remains undescribed. The only preserved specimens that | examined were infected with trematode sporocysts, 116 THOMPSON FIGS. 27-38. SEM photographs of radulae. Figs. 27-30. Gillia altilis (Lea). Figs. 31-34. Fluminicola nuttalliana (Lea). Figs. 35-38. Somatogyrus rheophilus п. sp. Enlargements: Fig. 27 х143; Figs. 31, 35 x95; Figs. 28-30, 32-34; 36-37 356; Fig. 38 х475. which grossly distorted the reproductive sys- tem. The radula of S. depressus was de- scribed by Stimpson (1865b: 21-22) who erred in reporting a perforation as occurring on the face of the lateral tooth. Walker (1909) used that as a characteristic to separate Somatogyrus from Clappia. Baker (1928: 148—149) correctly described and illustrated the radula of S. depressus. Its shell is illus- trated in this paper (Figs. 65, 66) and its distribution is mapped (Fig. 71). | have examined the anatomy of several species of Somatogyrus from Georgia, Ala- bama and Florida, and they are virtually iden- tical in most features. | assume that S. de- pressus is not significantly different. Because of similarities in the radula between S. de- pressus (Figs. 23—26) and other species, e.g. $. rheophilus п. sp. (Figs. 35-38), aspects of the anatomy of S. rheophilus are described as representative for Somatogyrus. Some data for other species also are provided. The shell of S. rheophilus is described in Appendix A. External morphology. Body, top of snout, sides of foot and dorsal surface of tentacles grayish black. Slight grayish patch present on each side of posterior base of tentacles. Un- derside of tentacles, muzzle, and sole grayish white. Mantle collar golden-flecked. Penis un- pigmented. Outer surface of mantle cavity NORTH AMERICAN LITHOGLYPHINAE Lay FIGS. 39-42. SEM photographs of protoconchs, showing embryonic sculpture. Fig. 39. Lepyrium showalteri (Lea), x77. Fig. 40. Somatogyrus rheophilus п. sp., «116. Fig. 41. Gillia altilis (Lea), х116. Fig. 42. Fluminicola nuttalliana (Lea), X 77. 118 THOMPSON got sv ASS Ay) AA DE A ЕН 2mm 47 FIGS. 43—47. Somatogyrus rheophilus п. sp. Fig. 43. Male denuded of shell and partially uncoiled. Fig. 44. Female denuded of shell and partially uncoiled. Figs. 45—46. Female reproductive system. Figs. 47. Male reproductive system. Legend: ag = albumen gland; ast = anterior chamber of stomach; bu = bursa copulatrix; cg = capsule gland; di = digestive gland; emc = posterior wall of mantle cavity; gf = gill filaments; goo = ovary; got = testis; in = intestine; ki = kidney; op = operculum; ov = oviduct; pe = penis; pov = pallial oviduct; pr = prostate; sr = seminal receptacle; sv = seminal vesicle; vc = ventral channel; vd, = proximal vas deferens; vd, = distal vas deferens. NORTH AMERICAN LITHOGLYPHINAE 119 with a large dark gray patch that is bounded along right side of pallial oviduct and posteri- orly by hypobranchial gland. Patch variable from light gray to nearly black. Most species of Somatogyrus that | examined are similarly colored; a few species (undescribed) have black blotches and spots on the mantle. Gill lamellae about 26-34. Osphradium elongate, nearly as long as gill. Operculum ovate, paucispiral, with a subcentral nucleus. Radula. Data for S. rheophilus are taken from SEM photos comprising Figs. 35-38 and from prepared slides. Data for S. depressus are presented in Table 1. Radula with about 35—45 transverse rows of teeth. Central tooth trapezoidal in shape with a mid-basal pro- jection; mesocone enlarged, blunt, bordered by 3-4 blunt ectocones on each side; 3-4 basocones in a low ridge in middle of lateral angles; basocones increasing in size dorsally. Lateral teeth (Figs. 35-37 with 4 entocones, an enlarged mesocone and 5-6 ectocones; shaft elongate, slender, and flexed, though not as much as in Lepyrium and Clappia; basal projection stout, pointed. Inner marginal teeth (Fig. 37) stocky, with about 30 fine, acuminate cusps. Outer marginal teeth more slender, with about 30 fine cusps (Fig. 38). Female reproductive system (Figs. 44-46). Ovary (goo) confined to upper whorl of vis- cera along columellar wall and completely imbedded in digestive gland (di) (Fig. 44); consisting of 3-4 lobes, each of which is subdivided into several smaller lobules. Pri- mary oviduct (ov) passing along ventral- mesad side of digestive gland almost to post- erior wall of mantle cavity (emc), and then passing posteriorly to form an open loop along mesad side of albumen gland (ag) (Fig. 45). Gono-pericardial duct short but stout. Seminal receptacle (Figs. 45—46, sr) small, saccate, located on descending arm of ovi- duct loop; appressed against ventral side of bursa copulatrix (bu), but visible externally. Bursa copulatrix large, saccate, overlapping posterior end of albumen gland. Duct from albumen gland joining primary oviduct at posterior partition of mantle cavity to form ventral canal (vc) of pallial oviduct (pov). Cap- sule gland (cg) extending to end of ventral canal. Pallial oviduct usually more slender than in other genera. Eggs deposited in single capsules on hard substrate. Egg capsule low, dome-shaped, 1.20-1.25 mm wide with a flat collar 0.20-0.25 mm wide. Male reproductive system (Fig. 47). Similar in most aspects to Lepyrium except that the penis lacks small dermal glands and is re- latively more slender and blade-like. Testis (got) very large, occupying upper two whorls of viscera, where it overlies dorsal surface of digestive gland, stomach, and posterior edge of prostate (pr, Fig. 43). Testes consisting of many large lobes that fork into 2-4 lobules each. Primary sperm duct lying along mid- ventral side of testis (Fig. 47). Its apical end very slender; in second visceral whorl primary sperm duct becomes greatly enlarged and convoluted forming a seminal vesicle (sv), and then narrowing to a thin delicate vas deferens (vd,) above prostate. Prostate (pr) ovate in shape, imbedded in body wall just behind posterior wall of mantle cavity. Vas deferens (vd,) imbedded in body wall for first half of its length, and then enters body cavity to base of penis (pe), where it courses through left side of penis and discharges at its tip. Penis originating on right side of nape behind right eye tentacle, sickle-shaped, dorso-ventrally flattened and recurved coun- terclockwise into the mantle cavity when con- tracted. TABLE 1. Radular characteristics of some North American lithoglyphines and Birgella. a AS A Е И A A A РЕНИ Inner Outer Tooth Central Central Lateral marginal marginal: rows basocones ectocones cusps cusps cusps ______ 222 m [re SEE er EEE SE NEE EEE A A Е Gillia altilis (7) 51-55 2 34 8-9 ca. 30 6-9 Fluminicola nuttalliana (5)* — 23 4-5 7-8 ca. 16 12-13 Somatogyrus depressus (6)* — 4 3— 8—10 са. 30 са. 25 Somatogyrus rheophilus (7) 35—45 3— 3— 10—11 са. 30 са. 30 Clappia clappi (2) 56-59 6-7 6-7 18-21 ca. 50 ca. 35 Lepyrium showalteri (7) 140-149 SU 22-24 18-22 ca. 50 ca. 50 Birgella subglobosa (2) 48-49 2 34 10-11 12 9 AA AAA KA — _—— *SEM preparations; counts not taken. 120 THOMPSON Gillia Stimpson, 1865 Gillia is a monotypic genus found in streams along the Atlantic coast of eastern North America. Its shell is described and its distribution is discussed in Appendix A. The shell is characterized by its large size, ovate- conical shape, and fine spiral threads on the protoconch (Fig. 41). External morphology. Similar to Somatogy- rus. Mantle uniformly black or pigmentation may be reduced to a large fuscous blotch covering the mantle cavity and the stomach. Nape, top of snout, and top of tentacles black. Under side of tentacles white. Sides of foot and snout light gray. Operculum paucispiral, broadly ovate with a subcentral nucleus (Fig. 52). Baker (1918) described the eggs as laid singly or in small clusters at up to six on leaves and stems of aquatic plants. The cap- sules are hemispherical and 1.25 mm in di- ameter. An attachment collar is not men- tioned or illustrated. Radula (Figs. 27-30). The radula is spe- cialized for feeding on coarser food particles than do the preceding genera. This is in- dicated by modifications of the central tooth (Figs. 27-28) and the lateral teeth (Fig. 29). The cusps on each tooth are modified and aligned to form a single large serrate blade. There is a correspoding loss of small cusps on the outer marginal tooth (Fig. 30) and among the basocones of the central tooth. Central tooth (Figs. 27-28) broad and with greatly extended lateral angles; mesocone large, acuminate, bordered on each side by 3—4 ectocones that serve as serrations on the sides of the mesocone; basocones two on each side, the uppermost greatly enlarged and acuminate. Lateral tooth (Fig. 29) stout; shaft only slightly flexed, nearly aligned ver- tically with face of tooth; mesocone large, acuminate, bordered by 2 entocones and 4-5 sharp ectocones; basal projection stout, pointed. Inner marginal tooth with about 30 cusps. Outer marginal tooth (Fig. 30) slender, with about 6 relatively stout cusps. Female reproductive system. The female system (Figs. 48-50) is similar to that of Somatogyrus and Fluminicola, and differs from Lepyrium in the size and location of the seminal receptacle (sr) and the extent of the capsule gland (cg). The loop of the oviduct may or may not be folded along the side of the albumen gland (ag). The seminal receptacle originates at the base of the oviduct loop and lies along the lower edge of the bursa copula- trix (bu), and like Fluminicola it is completely covered by albumen gland (ag) tissue. The capsule gland (cg) extends to the end of the ventral canal. Male reproductive system. Gillia (Fig. 51) differs from Lepyrium and Somatogyrus, but is like Fluminicola in the extent of the testes and the pigmentation and shape of the penis (pe). The testis (got) extends forward only to the posterior half of the stomach and to the posterior edge of prostate (pr). Anterior vas deferens (vd,) highly convoluted, passing along right side of body wall and then trans- versely across nape to base of penis. Penis flattened, blade-like (contracted in Fig. 51), and terminating in a slender fleshy papilla through which the vas deferens discharges; distal third of penis with an internal patch of melanophores along vas deferens. Fluminicola Stimpson, 1865 The specimens | examined are tentatively identified as F. nuttalliana (Lea), the type- species of Fluminicola. They are typical in all aspects except that they are smaller than average. | have not seen sufficient material to review the systematics of F. nuttalliana nor to discuss its distribution. The shells of Flumini- cola vary greatly in size, up to 12 mm. The shell varies from ovate-conical to globose. The protoconch of F. nuttalliana (Fig. 42) is similar to Gillia by having fine spiral sculpture. External morphology. Animal like Gillia, Mantle uniformly black. Operculum like Gillia. Eggs laid on hard objects, single or in small clusters; hemispherical, 1.25 mm in diameter, with a narrow hyaline collar 0.12-0.15 mm wide. Radula. The radula (Figs. 31-34) is similar to that of Gillia by having greatly enlarged cusps on the central and lateral teeth. It dif- fers from Gillia primarily by the number of cusps on the marginal teeth. Central tooth with extended lateral angles; upper edge (Fig. 32) with an enlarged acuminate mesocone bordered on each side with 4-5 ectocones that combine to form a large, serrated, pro- jecting blade; uppermost basocone greatly enlarged, followed by 1-2 much smaller basocones on each side. Lateral tooth (Fig. 33) with a weakly flexed, stout shaft, and an enlarged mesocone bordered by two smaller entocones and four ectocones; basal projec- tion long, slender, stout. Inner marginal tooth (Fig. 34) with about 16 acuminate cusps. NORTH AMERICAN LITHOGLYPHINAE 121 FIGS. 48-55. Gillia altilis (Lea). Fig. 48. Female reproductive system. Fig. 49. Coil of oviduct freed from albumen gland showing seminal receptacle. Fig. 50. Right side of oviduct coil showing relationship of bursa copulatrix to seminal receptacle. Fig. 51. Male reproductive system. Fig. 52. Operculum. FIGS, 59—55. Fluminicola nuttalliana (Lea). Fig. 53. Female reproductive system. Fig. 54. Coil of oviduct freed from albumen gland. Fig. 55. Right side of oviduct coil showing relationships of seminal receptacle and bursa copulatrix. Legend: ag = albumen gland; bu = bursa copulatrix; сд = capsule gland; етс = posterior wall of mantle cavity; goo = ovary; got = testis; in = intestine; mc = mantle collar; ov = oviduct; pe = penis; pr = prostate; sr = seminal receptacle; sv = seminal vesicle; vc = ventral channel; vd, = vas deferens. 122 THOMPSON Outer marginal tooth (Fig. 34, om) with about 12—13 relatively stout cusps. Female reproductive system. Similar to that of Gillia altilis except that the loop of the primary oviduct is not secondarily folded (Figs. 53-55). In contrast to other genera, the pallial oviduct lies lower along the right side and the intestine occupies a dorsal position over the pallial oviduct. Male reproductive system (not figured). Most similar to Gillia altilis. The testis extends forward to overlap the posterior half of the stomach and the posterior margin of the pros- tate. Penis pigmented internally with two dif- fuse bands of melanophores along the distal third of the vas deferens, and tip of penis ending in a small fleshy papilla. My observations on the penis differ from the description given by Stimpson (1865b: 24-26). He stated that the left base of the penis has a wing-like expansion. However, Stimpson’s material came from an un- specified locality and was poorly preserved. Thus the identity of the species he examined is uncertain. PHYLOGENETIC RELATIONSHIPS The genera discussed above have been classified in two orders and three families within the Subclass PROSOBRANCHIA. Lepyr- ¡um was first considered to be in the Neriti- dae, a family in the Order ARCHAEOGASTRO- PODA, because of its neritid-shaped shell (Lea, 1861; Binney, 1865; Dall, 1896; Walker, 1918; Wenz, 1939). Pilsbry & Olsson (1951) established its affinities to the order MESO- GASTROPODA on the basis of its taenioglos- sate radula. They proposed the monotypic Family Lepyriidae because the combination of its radular and opercular characters was dissimilar to other families of freshwater gas- tropods. The genera Somatogyrus, Clappia, Gillia, and Fluminicola traditionally have been grouped together in the mesogastropod fami- ly HYDROBIIDAE, Subfamily LITHOGLYPHINAE because of their thick, globose, lithoglyphine- type shells (Stimpson, 1865b; Walker, 1918; Burch & Tottenham, 1980). This relationship is correct but for the wrong reason. The globose lithoglyphine-type shell is an adaptation for two very different habitats, and has evolved at least four times in unrelated hydrobioid subfamilies. It has evolved in- dependently in the Pomatiopsidae (Triculi- nae) and the Hydrobiidae (Lithoglyphinae) as an adaptation for existence in high-energy streams. The globose shell accommodates an enlarged foot and muscle system for attachment to rocks in swift currents (Davis, 1979; Davis & da Silva, 1983; this paper). A similar type of shell has evolved twice again within the hydrobiid subfamily Nymphophili- nae as an adaptation to a very different habi- tat. Birgella and Notogillia, two distantly re- lated genera within the subfamily, live in quiet waters on fine-particle substrates (Thomp- son, 1968; this paper). A wide foot is required to support the snail’s weight on a silt sub- strate, and the enlargement of the foot is accommodated by an enlarged globose shell. It is clear that the lithoglyphine-type shell is highly adaptive, and thus is convergent. Con- clusions concerning suprageneric rela- tionships based on this character-state must take convergence into account. During recent years, as knowledge about the anatomy of hydrobioid snails progressed, it has become increasingly difficult to define family units (families and subfamilies) (Taylor, 1966; Davis, 1966; Thompson, 1968; Radoman, 1973; Davis, 1979; Thompson, 1979; Her- shler & Davis, 1980; Davis, et al., 1982; Davis & Pons da Silva, 1983). With increasing knowledge about the anatomy of additional genera the distinctions between family units becomes less clear, and requires redefinition of established units or the designation of new units. The result is that family units are be- coming separable by fewer and fewer charac- teristics, and frequently their definitions in- clude words such as “except,” “as in,” and “shared with.” Such instability is expected be- cause fewer than 20% of the hydrobioid genera have been studied to the extent that the internal morphology is known for a single species. It is clear that the classification of the hydrobioids is in an embryonic state of knowl- edge. A great deal more must be learned before any stability in classification can be achieved. Such a Classification will have to be based on consistent criteria of morphology, biochemistry, genetics, behavior, and ecoi- ogy. At present the hydrobioids are classified only on morphological criteria that are proving not to be consistent. None-the-less the em- ployment of such criteria is useful in deriving phylogenetic concepts. Family relationships. On the basis of an- atomical data it is clear that Lepyrium, Clap- pia, Somatogyrus, Gillia and Fluminicola con- stitute a compact monophyletic group within the Hydrobiidae as defined by Davis (1979). NORTH AMERICAN LITHOGLYPHINAE 123 The family relationship is established by seven morphological characters. The taenioglossate radula has a trapezoidal cen- tral tooth with pronounced lateral angles ex- tending beyond the posterior margin. The central tooth has two or more pairs of ecto- cones on each side of the mesocone. Epi- taenia and associated food grooves for filter feeding are absent. Sperm enters the female reproductive system through the genital pore at the anterior end of the pallial oviduct where it passes posteriorly via a ciliated ventral ca- nal (except Littoridininae; Hershler, in press). The single seminal receptacle originates on the primary oviduct posterior to the bursa copulatrix. The eggs are deposited singly in tough hemispherical capsules and are not coated with sand. The prostate overlaps the posterior edge of the mantle cavity. Two other characters, though not unique to the Hydro- biidae, also serve to remove the group from other superfamilies. The simple, chitinous, paucispiral operculum lacks an internal peg. Sperm transmission from the male is accom- plished through a penis that originates on the nape and is innervated by the pleuropedal connective. Subfamily relationships. Within the Hydro- biidae six subfamilies presently are recog- nized in North America. Lithoglyphinae (this paper; Davis and Pons da Silva, 1983), Nym- phophilinae (Thompson, 1979; Hershler, in press), Littoridininae (Davis et al., 1982; Her- shler, in press), Hydrobiinae (Davis, 1966; Hershler & Davis, 1980), Amnicolinae (Thompson, 1968), and Fontigentinae (Burch, 1982). The last two subfamilies are excluded from further discussion because of the pres- ence of two (Amnicolinae) or three (Fon- tigentinae) ducts within the penis. They are considered remote in their relationships to the other subfamilies though their anatomies are poorly known. Eight character-states com- mon to the Lepyrium-group of genera are useful for establishing relationships within the remaining subfamilies and for defining the subfamily Lithoglyphinae. (1) The shell is glo- bose or conico-globose. (2) The protoconch is sculptured with spiral lirations; spirally ar- ranged series of pits may occur in addition. (3) The mantle is uniform black or dark gray. (4) One or more pairs of basal cusps arise from the face of the radular central tooth, not the lateral angles. (5) The stomach lacks folds or protuberances on its posterior chamber. (6) The fecal pellets are cylindrical and are spiral in structure. (7) The penis is flattened and blade-like, with a simple duct (vas deferens) internally. (8) The penis lacks lobes or com- plex glandular structures on the outer surface. Within the Hydrobiidae characters 4, 5, 6, and 7 are exclusive to the Lithoglyphinae. Character 3 may also be. | am not aware of data to the contrary. The remaining character- states are shared with one or more subfami- lies. The globose or conico-globose lithogly- phine-type shell also occurs in some Nym- phophilinae and Littoridininae. In those sub- families, it is an uncommon character. All of the genera that have been shown to be Lithoglyphinae because of their soft anato- mies have lithoglyphine-type shells. Two con- clusions are suggested by this. The lithogly- phine-type shell in the Lithoglyphinae is fun- damental, and thus is a primitive character- state. Also, the habitat deployment of the Lithoglyphinae in high-energy streams is a basic, and therefore primitive behavioral characteristic of the subfamily. Some species, e.g. Somatogyrus depressus (Tryon) and Gil- lia altilis (Lea) may inhabit quiet-water habi- tats as well as high-energy streams. This can be considered a secondary adaptation be- cause it occurs seldom and sporadically with the subfamily, not among a cluster of closely related species, and it occurs among species that also inhabit fast streams. The spiral sculpture of the protoconch is characteristic of the Lithoglyphinae. A secon- dary reduction may occur (Lepyrium) or secondary additions may occur (Somatogy- rus). Spiral protoconch sculpture in Birgella (Figs. 79—80) is the only recorded occurrence in the Nymphophilinae. Spiral protoconch sculpture also occurs in some southeastern Amnicolinae, e.g. Lyogyrus retromargo (Thompson). This type of sculpture may be a primitive condition within the Hydrobiidae, with other types of sculpture representing de- rived states. The absence of glandular ridges, raised glands, lobes or papillae on the penis is a character state that also occurs in some Lit- toridininae (Hershler, in press) and in some Hydrobiinae (Hershler & Davis, 1980). In those cases it may be a derived condition through the secondary loss of previously ex- isting characters. The total absence of these structures in the Lithoglyphinae suggests that their absence is a generalized, primitive con- dition in the subfamily. The Lithoglyphinae as presently un- derstood includes seven genera. Five in 124 THOMPSON North America, Potamolithus in South Amer- ica (Davis & Pons da Silva, 1983) and Litho- glyphus in Europe (Krause, 1949). In addition to the nine character-states listed above that differentiate North American lithoglyphines from other Hydrobiidae, the North American genera have in common five character-states that differentiate them from other Lithoglyphi- nae. In sequence with the character-states listed above these are as follows: (9) The testis is very large; it covers almost the entire dorsal surface of the digestive gland and par- tially overlaps the stomach and prostate. (10) The vas deferens is not modified into an enlarged ejaculatory duct at the base of the penis. (11) The penis may or may not have a terminal papilla; when present the papilla is non-retractable. (12) The penis lacks a pre- putium. (13) A nuchal node is absent. The size of the testis separates the New World genera from Lithoglyphus. In Lithogly- phus, Krause (1949) describes the testis as overlying the digestive gland in the first and second whorl. The extent of its distribution in these two whorls is not clear. His illustration (p. 135, fig. 22) depicts a testis that is not larger than the prostate. If that is correct, the testis is very small compared to those in other genera. The remaining four character-states differentiate the North American genera from Potamolithus (see Davis & Pons da Silva, 1983). 56 THE NORTH AMERICAN LITHOGLYPHINE GENERA The evolution of North American Litho- glyphinae centered about microhabitat selec- tion and trophic specialization. Microhabitat selection produced variation in shell-form within the constraints imposed by a lentic environment. Trophic specialization is re- flected in variations of the basic structure of the cusps on the radular teeth and the shaft of the lateral tooth. Character-trends are dis- cernible in shell form and radular structures. Shell form. A high degree of diversity exists in the shells of North American Lithoglyphi- nae. The species vary in size from small to large (2-12 mm) and may be obese with a short or depressed spire, ovate-conical with a pronounced spire, or flattened and limpet-like. They include the largest of the American Hy- drobioidea. Most have a voluminous body whorl. Basically, most shells are imperforate or narrowly rimate. When viewed from the front, most have a noticeable spire protruding from the right side (Fig. 57). When the animal is active, the shell is raised and the eyes, tentacles, and muzzle extend considerably beyond the edge of the lip. Two lines of specialization in shell form occur. Clappia diverges from the basic shell form by having a broadly perforate umbilicus (Fig. 68). The adaptive significance of this is 57 FIG. 56-57. Fig. 56. Anterior view of Lepyrium showalteri (Lea); note reduction of spire to produce a limpet-like shell. Fig. 57. Anterior view of Somatogyrus rheophilus п. sp., a lithoglyphine with a normal-spired shell. NORTH AMERICAN LITHOGLYPHINAE 125 not clear because little information is avail- able on the ecology of Clappia. However, this is only a variation in degree from the perforate umbilicus of some Somatogyrus (Fig. 65). Lepyrium diverges from the basic shell form by being limpet-like, with a depressed spire and a greatly enlarged aperture. When the animal is viewed from the front, the shell covers the body like a low shield and the spire is barely evident (Fig. 56). The animal raises its shell only slightly as it moves, and the eyes, tentacles and snout barely protrude be- yond the edge of the lip. The shell form of Lepyrium is a modification for reducing hydrostatic drag on an animal that lives in fast currents on smooth boulders, where charac- teristically it is found. The neomelanian type operculum of Lepyrium is a secondary specialization related to the enlarged aper- ture. Other variations in shell morphology are significant as species-level criteria or species- group criteria. The large size of Gillia altilis readily distinguishes this monotypic genus from other eastern North American genera, but not from Fluminicola, which achieves an even larger size. However, it is not shell size by which Gillia is separated from other genera, but by trophic structures. Further- more, some of the more globose species now placed in Somatogyrus may be found to be- long in Gillia when their radulae are ex- amined. Radula. Three important character-trends occur in trophic structures. They are the pri- mary modification through which adaptive radiation of the Lithoglyphinae has taken place in North America. Each genus is char- acterized more by its radular features than by other structures. These features indicate dif- ferent feeding roles for the various genera. Radular data are summarized in Table 1. The trends are: (1) modifications in the number of transverse tooth rows, (2) modifications of size and numbers of cusps on the radular teeth, and (3) a corresponding modification in the size and orientation of the shaft of the lateral tooth. Most species have a moderate number of transverse tooth rows, about 35-55 (Table 1, Fig. 23). In two groups, Clappia and Lepyr- ium, there is a significant increase in the number of rows. This increase is accommo- dated by a greater degree of overlap between successive rows, not by increasing the rela- tive length of the ribbon (Figs. 15, 19). Two trends can be recognized in mod- ifications of size and numbers of radular cusps. In Clappia and Lepyrium there is a decrease in the relative size of the cusps accompanied by an increase in the number of cusps on each tooth, indicating that these genera are specialized for grazing on finer plant-food particles than does the related genus Somatogyrus. A second trend occurs in Gillia and Fluminicola. In these genera the basocones in the central tooth are decreased in number, but are greatly enlarged, as are other cusps on the central and lateral teeth. There is also a corresponding decrease in the number, but an increase in relative size of the cusps on the inner and outer marginal teeth. These are specializations for grazing on coarser food materials than do related genera. The shaft of the lateral tooth is moderately stout and is flexed laterally at a slight angle to the tooth face in Somatogyrus (Figs. 23, 25). In genera with reduced cusp size (Clappia and Lepyrium) the shaft is slender and is flexed laterally at a greater angle (Figs. 21, 17). This accommodates an increase in trans- verse tooth rows without increasing the rela- tive length of the ribbon. In genera with en- larged cusps (Gillia and Fluminicola) the shaft is stouter and is aligned almost vertically with the face of the tooth (Figs. 29, 33). These changes are related to trophic specializations. Those genera that feed on coarse foods re- quire stout shafts with rectilinear vertical sup- port for the lateral tooth. Those genera that feed on smaller food particles require less support and can function with slender, strong- ly flexed lateral tooth shafts. Relationships. Ten character-states are useful for determining intergeneric rela- tionships. These states are discussed at var- ious points earlier in this paper and are summarized in Table 2. A phenogram based on these character-states is depicted in Fig. 58. Gillia and Fluminicola have eight charac- ter-states in common (1-8). They differ from each other in two character states (9-10). They differ exclusively from the other genera by six character-states (4-9). Somatogyrus, Clappia, and Lepyrium have three character- states in common (4, 5, 10). Somatogyrus differs from the Clappia-Lepyrium lineage by three character-states (7, 8, 9). Clappia and Lepyrium have three character-states In com- mon (7, 8, 9). They differ from each other in four character-states (1, 2, 3, 6). From the data presented in Table 2 it is apparent that modification of feeding struc- 126 THOMPSON Fluminicola Gillia Somatogyrus Clappia Lepyrium FIG. 58. Phenogram depicting intergeneric relationships of North American Lithoglyphinae based on characters listed in Table 2. TABLE 2. Variation in ten character-states among North American lithoglyphine genera. Somatogyrus Clappia Lepyrium Gillia Fluminicola Shell shape 0 0 1 0 0 ovate-conical (0) neritid (1) Umbilicus 0 1 0 0 0 imperforate-rimate (0) open (1) Operculum 0 0 1 0 0 subcentral (0) excentric (1) Size 1 1 1 0 0 large (<5 mm) (0) small (> mm) (1) Penis papilla 1 1 1 0 0 present (0) absent (1) Central tooth cusps 1 1 2 0 0 large (0) medium (1) small (2) Lateral cusps 1 2 2 0 0 large (0) medium (1) smail (2) Lateral shaft 1 2 2 0 0 vertical (0) angular (1) flexed (2) Inner marginal cusps 1 2 2 1 0 large (16) (0) medium (30) (1) minute (50) (2) Outer marginal cusps 2 2 2 0 1 large (6-9) (0) medium (12) (1) small (25+) (2) NORTH AMERICAN LITHOGLYPHINAE 127 tures (trophic specialization) is the fun- damentally most important factor underlying the adaptive radiation of the Lithoglyphinae in North America. This is coupled with a minor degree of variation in the structure of the penis (reproductive specialization). It is also apparent that variation in the shell and oper- culum are significant at lower taxonomic levels, and that they are adaptations reflect- ing microhabitat specialization. The North American lithoglyphine genera are redefined as follows. They have in com- mon the characters discussed earlier in this paper, which differentiates them from Eu- ropean and South American genera. Gillia Stimpson, 1865a Type-species. Melania altilis Lea, 1841 (see Figs. 63, 64; Appendix A). Definition. Shell medium to large (6-8 тт high). Imperforate or rimate. Conico-globose in shape. Protoconch sculptured with spiral striations. Operculum paucispiral. Penis with a small terminal papilla. Shaft of lateral tooth straight. Cusps of radular teeth enlarged; numbers of cusps given in Table 1. Outer marginal tooth with few (6-9) cusps. Distribution. Atlantic drainage systems of eastern North America from South Carolina north to New York and Vermont. Species. Monotypic. Fluminicola Stimpson, 1965a Type-species. Paludina nuttalliana Lea, 1839 (see Burch & Tottenham, 1980: 101, fig. 142, for an excellent illustration of the spec- ies. Definition. Shell medium to large in size (up to 12mm high). Imperforate or rimate. Conico-globose or globose in shape. Pro- toconch sculptured with spiral striations. Operculum paucispiral. Penis with a small terminal papilla. Shaft of lateral tooth straight. Cusps on radular teeth enlarged; numbers of cusps given in Table 1. Inner marginal tooth with few (16) cusps. Distribution. Pacific drainage systems from California north to Washington and interior basin. Species. Indeterminate. Burch & Tot- tenham (1980: 102) list 12 species. The rad- ula and soft anatomy of only the type-species is known. Note. Taylor (1966) synonymizes Flumini- cola with Lithoglyphus. In light of the dif- ferences in the prostate gland and the radular cusps, they are retained as separate genera pending additional anatomical data on other species of Fluminicola. Radoman (1966) fi- gures the radula of L. naticoides (C. Pfeiffer) as having few cusps on the inner marginal tooth (8-9) and on the outer marginal tooth (7). Somatogyrus Gill, 1863 Type-species. Amnicola depressa Tryon, 1862 (see Figs. 65, 66). Definition. Shell small to medium in size (1-6 mm high). Imperforate, rimate or narrow- ly umbilicate. Conico-globose or globose in shape. Protoconch sculptured with spiral threads, and it may also have spirally ar- ranged series of pits. Operculum paucispiral with a sub-lateral nucleus. Penis without dis- tinct terminal papilla. Shaft of lateral tooth weakly angular. Cusps of radular teeth mod- erately large; numbers of cusps given in Table 1. Distribution. Eastern North America throughout the Mississippi drainage system, and from the Potomac River south and west through the Gulf Coast systems. Panuco River system of Mexico. Species. Numerous. Burch & Tottenham (1980) list 35 species in the United States. Pilsbry (1910) describes a species from the Rio Coy in Mexico. Another is described in this paper. Remarks. Species may differ by the num- ber and size of the cusps on the radular teeth and the number of tooth rows. Two sub- genera have been recognized on the basis of cusp development: Somatogyrus s. s. and Walkerilla Thiele, 1928. The degree of cusp development does not seem to be an ade- quate feature for separating subgenera. Con- vergence in this characteristic among species of quite dissimilar shells (S. coosaensis and S. tenax; see Burch & Tottenham, 1980) sug- gests that Walkerilla, as used previously (Thompson, 1969: 260), is polyphyletic and artificial in concept. Clappia Walker, 1909 Type-species. Clappia clappi Walker, 1909 (= Somatogyrus umbilicatus Walker, 1904) (Fig. 68). Definition. Shell small (about 3 mm high). Broadly umbilicate. Conico-globose In shape. Protoconch sculpture unknown. Operculum 128 THOMPSON paucispiral with a large subcentral nucleus. Penis simple, apparently without a terminal papilla. Shaft of lateral tooth strongly flexed. Cusps of lateral and marginal teeth minute; numbers of cusps given in Table 1. Distribution. Confined to the Coosa and Cahaba rivers in central Alabama. Species. Two; both may be extinct. Lepyrium Dall, 1896 Type-species. Neritina showalteri Lea, 1861 (see Figs. 59—62; Appendix А). Definition. Shell medium in size (about 4 тт high). Imperforate. Flattened, neritid in shape, but with a complete internal spire; aperture greatly enlarged for a limpet-like mode of existence. Protoconch smooth with a few low, wide, spiral grooves. Operculum paucispiral with a very excentric nucleus (neomelanian, Fig. 7). Penis simple, without a terminal papilla. Shaft of lateral tooth strongly flexed. Cusps of all radular teeth minute, ac- companied by a large proliferation of trans- verse tooth rows. Numbers of cusps given in Table 1. Distribution. Coosa, Cahaba, and Little Cahaba rivers in central Alabama. Species. Monotypic. ACKNOWLEDGEMENTS Many people have assisted me in this study. | thank George M. Davis for sharing with me prepublished data on Potamolithus. The following individuals loaned specimens in their charge: John B. Burch, Museum of Zool- ogy, University of Michigan; Kenneth J. Boss, Museum of Comparative Zoology, Harvard University; George M. Davis, Academy of Natural Sciences, Philadelphia; Joseph P. E. Morrison, U.S. National Museum of Natural History. The following persons assisted me in field work or sent me specimens of critical species: Dennis R. Paulson, Hugh J. Porter, William Font, Richard Franz, Beverly E. John- son, Robert Hanley, and Gordon Ultch. Draw- ings of the shell of Lepyrium showalteri were done by Lauren A. Keswick. The other shell drawings were done by Wendy Zomlefer. Financial support for field work was provided by the Florida State Museum, University of Florida. LITERATURE CITED BAKER, F. C., 1902, The Mollusca of the Chicago area: the Gastropoda. Bulletin of the Chicago Academy of Sciences, 3: 137—410, pl. 28-35. BAKER, F. C., 1918, Notes on nidification in Gillia and Amnicola. Nautilus, 32: 19-23, pl. 2, figs. 1-10. BAKER, F. C., 1926, Nomenclatural notes on American freshwater Mollusca. Transactions of the Wisconsin Academy of Arts, Sciences, and Letters, 22: 193-205. BAKER, Е. C., 1928, The freshwater Mollusca of Wisconsin: Gastropoda. Bulletin Wisconsin Geological and Natural History Survey, 70: i-xvii, 1-507, 38 pl. BERRY, E. G., 1943, The Amnicolidae of Michigan: distribution, ecology and taxonomy. Mis- cellaneous Publications, Museum of Zoology, University of Michigan, 57: 1-68, 9 pl. BINNEY, W. C., 1865, Land and freshwater shells of North America: Ampullariidae, Valvatidae, Viviparidae, freshwater Rissoidae, Cyclophor- idae, Truncatellidae, freshwater Neritidae, Heli- cinidae. Smithsonian Miscellaneous Collections, 144: i-viii, 1-120. BURCH, J. B., 1982, Freshwater snails (Mollusca, Gastropoda) of North America. Environmental Monitoring and Support Laboratory, Office of Research and Development, U.S. Environmental Protection Agency: i-vi, 1-294. BURCH, J. B. & TOTTENHAM, J. L., 1980, North American freshwater snails: species list, ranges and illustrations. Walkeriana, 3: 81-215. CLARKE, A. H., 1981, The freshwater Mollusca of Canada. National Museum of Natural Sciences, National Museums of Canada: 1—446. DALL, W. H., 1896, Notes on Neritina showalteri. Nautilus, 10: 13-15. DAVIS, G. M., 1966, Notes on Hydrobia totteni. Venus, Japanese Journal of Malacology, 25: 27— 42. DAVIS, G. M., 1979, The origin and evolution of the gastropod family Pomatiopsidae, with emphasis on the Mekong River Triculinae. Monograph of Academy of Natural Sciences, Philadelphia, 20: 1-120. DAVIS, G. M. & PONS DA SILVA, M. C., 1984, Potamolithus: morphology, convergence, and re- lationships among hydrobioid snails. Malaco- logia, 25: 73-108. DAVIS, G. M., MAZURKIEWICZ, M. 8 MAN- DRACCHIA, M., 1982, Spurwinkia: morphology, systematics, and ecology of a new genus of North American marshland Hydrobiidae (Mol- lusca: Gastropoda). Proceedings of the Academy of Natural Sciences of Philadelphia, 134: 143-177. GILL, T., 1863, Systematic arrangement of the mollusks of the family Viviparidae, and others, inhabiting the United States. Proceedings of the NORTH AMERICAN LITHOGLYPHINAE 129 Academy of Natural Sciences of Philadelphia, 15: 3340. GOODRICH, C., 1941, Distribution of the gastro- pods of the Cahaba River, Alabama. Occasional Papers, Museum of Zoology, University of Michi- gan, 428: 1-30. HALDEMAN, 5. S., 1847-1848, G. Leptoxis. Lep- toxe. Rafinesque. т: CHENU, J. C., ed., /llustra- tions Conchyliologiques. Franck, Paris, 3: 1-6, 5 pl. HERSHLER, В., in press, The systematics and evolution of the hydrobiid snails (Gastropoda: Rissoacea) of the Cuatro Cienegas basin, Coahuila, Mexico. Malacologia, 26: 000—000. HERSHLER, R. & DAVIS, G. M., 1980, The morphology of Hydrobia truncata (Gastropoda: Hydrobiidae): relevance to systematics of Hydro- bia. Biological Bulletin, 158: 195-219. KRAUSE, H., 1949, Untersuchungen zur Anatomie and Okologie von Lithoglyphus naticoides (C. Pfeiffer). Archiv für Molluskenkunde, 78: 103- 148. LEA, 1., 1839, Descriptions of new fresh water and land shells. Transactions of the American Philo- sophical Society, 6: 1-111, 24 pl. LEA, 1., 1841, Continuation of Mr. Lea’s paper оп new fresh water and land shells. Proceedings of the American Philosophical Society, 2: 11-15. LEA, |., 1843, Descriptions of new fresh water and land shells. Transactions of the American Philo- sophical Society, 8: 163-250, pl. 5-27. LEA, 1., 1861, Descriptions of a new species of Neritina from Coosa River, Alabama. Pro- ceedings of the Academy of Natural Sciences of Philadelphia, 13: 55. LEA, |., 1863, New Melaniidae of the United States. Journal of the Academy of Natural Sciences of Philadelphia, 5: 217-356, pl. 34-39. PILSBRY, H. A., 1906, Note on Lepyrium. Nautilus, 20351 PILSBRY, H. A., 1910, New Amnicolidae from the Panuco River system, Mexico. Nautilus, 23: 97— 100, pl. 9. PILSBRY, H. A. & OLSSON, A. A., 1951, The Lepyriidae, a new family of fresh-water snails. Notulae Naturae, Academy of Natural Sciences of Philadelphia, 233: 1-5. RADOMAN, P., 1966, The zoogeographic and phylogenetic interrelations of the genera Litho- glyphus and Emmericia. Bulletin du Museum d'Histoire Naturelle du Beograde, 21: 43—49. RADOMAN, P., 1973, New classification of fresh and brackish water Prosobranchia from the Bal- kans and Asia Minor. Prosebna Izdanja, Museum d'Histoire Naturelle de Beograde, 32: 1-30. SAY, T., 1825, Descriptions of some new species of fresh water and land shells of the United States. Journal of the Academy of Natural Scien- ces, Philadelphia, 5: 119-131. SAY, T., 1829, Descriptions of some terrestrial and fluviatile shells of North America. New Harmony Disseminator of Useful Knowledge, 2: 227-230. STEIN, C. B., 1976, Endangered and threatened plants and animals of Alabama: Gastropoda. Bulletin Alabama Museum of Natural History, 2: 2141. STIMPSON, W., 1865a, Diagnoses of newly dis- covered genera of gastropods, belonging to the sub-fam. Hydrobiinae, of the family Rissoidae. American Journal of Conchology, 1: 52-54. STIMPSON, W., 1865b, Researches upon the Hy- drobiinae and allied forms. Smithsonian Mis- cellaneous Collections, 201; 1-59. TAYLOR, D. W., 1966, A remarkable snail fauna from Coahuila, Mexico. Veliger, 9: 152-228. THIELE, J., 1928, Revision des Systems der Hy- drobiiden und Melaniiden. Abteilung fúr Syste- matik, Okologie and Geographie der Tiere, 55: 351—402. pl. 8. THOMPSON, F. G., 1968, The aquatic snails of the Family Hydrobiidae of peninsular Florida. Uni- versity of Florida Press, Gainesville: i-ix, 1-268. THOMPSON, F. G., 1969, Some hydrobiid snails from Georgia and Florida. Quarterly Journal of the Florida Academy of Sciences, 32: 241-265. THOMPSON, F. G., 1977, The hydrobiid snail genus Marstonia. Bulletin of the Florida State Museum, Biological Sciences, 21: 113-158. THOMPSON, F. G., 1979, The systematic rela- tionships of the hydrobioid snail genus Nym- phophilus Taylor, 1966 and the status of the Subfamily Nymphophilinae. Malacological Re- view, 12: 41—49. TRYON, G. W., 1862, Notes on American fresh water shells, with descriptions of two new spec- ies. Proceedings of the Academy of Natural Sci- ences of Philadelphia, 1862: 451452. TRYON, G. W., 1870, A monograph of the fresh- water univalve Mollusca of the United States. Academy of Natural Sciences, Philadelphia: 1— 238. WALKER, B., 1904, New species of Somatogyrus. Nautilus, 17: 133-142, 5 pl. WALKER, B., 1909, New Amnicolidae from Ala- bama. Nautilus, 22: 95-90, pl. 6. WALKER, B., 1915, Apical characters of Somato- gyrus with descriptions of three new species. Nautilus, 29: 37-41, 49-53. WALKER, B., 1918, A synopsis of the classification of the fresh-water Mollusca of North America, north of Mexico, and a catalogue of the more recently described species, with notes. Mis- cellaneous Publications of the Museum of Zoolo- gy, University of Michigan, 6: 1-213. WENZ, W., 1938-1939, Gastropoda, Parts 2and3, Prosobranchia. p. 432 (Lepyrium) and 555-581 (Hydrobiidae) /n SCHINDEWOLF, O. H., ed. Handbuch der Paláozoologie. Borntraeger, Ber- lin. 130 THOMPSON APPENDIX A The following Lithoglyphinae are described. Adequate descriptions of two are not avail- able in contemporary literature. The third species is new, and its anatomy is discussed earlier in this paper. Lepyrium showalteri (Lea) Neritina showalteri Lea, 1861: 55.—Lea, 1363: 267. ply 35, figs: 78, 78a. Neritella showalteri (Lea), Binney, 1865: 106, fig. 212. Lepyrium showalteri (Lea), Dall, 1896: 13- 15.—Walker, 1918: 38, fig. 139.—Wenz, 1938: 432, fig. 1062.—Pilsbry & Olsson, 1951: 1-5, figs. 1-3, 3a—Stein, 1976: 25.—Burch & Tottenham, 1980: 104, figs. 192-193. Lepyrium showalteri cahawbensis Pilsbry, 1906: 51.—Goodrich, 1941: 7, 10.—Pilsbry 8 Olsson, 1951: figs. 4-6.—Stein, 1976: 25: Shell (Figs. 59-62). Ovate in outline, adults 3.5-4.4 mm high and 4.0-5.0 mm wide, Cat. no. Height Width UMMZ 67445 3.84 4.46 UMMZ 67445 4.03 4.65 UMMZ 97448 4.40 4.96 about 0.81-0.94 times as high as wide. Strongly flattened with a strongly excentric apex and neritid-like in appearance; nearly uniformly dome-shaped with the apex hardly protruding when viewed from the rear (Fig. 61) or front (Fig. 56). Umbilical area im- perforate. About 2.3-3.0 whorls, which rapidly expand. Apical whorls usually eroded to the level of the body whorl. Protoconch de- pressed, nearly smooth with a few low wrin- kled depressions along outer surface (Fig. 39). Subsequent whorls rapidly expanding; sculptured with fine growth striations. Per- istome greatly expanded, circular, but vari- able in shape, large specimens tend to have a proportionally higher aperture than do smaller specimens. Columellar callus deeply dished, about a third the width of the aperture area and with a wing-like extension on the upper left corner. Peristome 0.92-1.08 times as high as wide. Width of peristome about 0.79— 0.95 time width of shell. Aperture flattened dorso-ventrally, continuing as a spire into up- per whorls, not shelf-like as in Neritidae. Measurements in mm of three specimens selected to show variation are: Ap. H. Ap.W. Whorls 3.91 3.60 2.5 4.03 4.40 27 4.40 4.65 3.0 Operculum (Fig. 7). Corneous; retractable into aperture for about Ya whorl. Paucispiral with the nucleus located close to the lower left margin (neomelanian); 24 rapidly expanding whorls. Outer surface sculptured with distinct incremental striations, and finer, close spiral striations. Distribution. Recorded from a short seg- ment of the Cahaba River and the Little Caha- ba River in Bibb County, and a short segment of the Coosa River in Shelby County, Ala- bama. Presumably it is extinct in the Coosa because of impoundment of the river. It still exists in the Cahaba and Little Cahaba rivers. Records of specimens examined are listed in Appendix C. Remarks. Pilsbry (1906) recognized two subspecies, the typical subspecies from the Coosa River, and L. s. cahawbensis from the Cahaba River. The latter was based upon immature specimens which he characterized as being smaller, with a straight columellar edge, and without a raised outer margin of the columellar area (causing the columella not to be dished). Later Pilsbry & Olsson (1951: 2) stated that the characteristics were in- consistent because they were based on ju- veniles, and they doubted the validity of cahawbensis as a distinct taxon. Unfortu- nately very little material is available from the Coosa River. However, specimens | have ex- amined demonstrate that only one form is recognizable. Somatogyrus Gill, 1863 The genus Somatogyrus includes 37 de- scribed species in North America. There are perhaps half again as many remaining to be described. Most of the described species are poorly known and inadequately illustrated, and must be restudied before meaningful specific comparisons can be made. Useful specific characteristics occur in the pro- NORTH AMERICAN LITHOGLYPHINAE 131 64 N 5mm —.........мее 88—66 FIGS. 59-64. Figs. 59-62. [ерупит showalteri (Lea), UMMZ 67445: Cahaba River, Guerney, Bibb Co., Alabama. Fig. 63. Gillia altilis (Lea), neotype: UF 40550. Fig. 64. Gillia altilis (Lea), UF 40551: Lake Waccamaw, Washington Co., North Carolina. 132 THOMPSON toconch sculpture, the radula, the pigmenta- tion patterns of the mantle, tentacles and snout, the shape and structure of the col- umellar lip, other aspects of the aperture, size and obesity. Adults of most species are de- collate due to erosion of the apical whorls. Thus, the last whorl and the aperture provide the only measurements of height that are useful for specific comparisons. Most species of Somatogyrus occur on rocks in high- energy rivers. Some occur in low gradient streams on sand and gravel. Most Somatogy- rus are annual species. Ovipositing usually takes place in May and June, whereupon the adults die. Morphological maturity of the new progeny occurs in October and November. Most samples that are collected between June and September contain only immature specimens, which have not yet developed the definitive characteristics essential for correct species identifications. Somatogyrus rheophilus Thompson, new species Diagnosis. A medium-sized species char- acterized by the tendency for its whorls to be weakly rounded above the periphery, its nar- rowly rimate umbilicus, its receded basal lip, its wide, rounded columellar lip, its advanced posterior corner of the aperture, its angular parietal-columellar corner, and its uniform black or grayish-black mantle. It is most sim- ilar to S. alcoviensis Kreiger from the Yellow River, Newton County, Georgia. The latter differs by having the columellar lip and parie- tal callus form a weak, oblique arch. Shell (Fig. 67). Broadly conical or turbinate. Adults decollate with 2-3 whorls remaining. Medium sized for the genus; eroded adults 3.9-4.5 mm high (holotype 4.3mm); body whorl about 3.5-4.0 mm high and 3.4— 4.0 mm wide. Width about equal to height of last whorl (0.97-1.02). Penultimate whorl 0.48—0.54 times width of last whorl. Last whorl nearly flattened above periphery. Su- ture weakly impressed. Umbilicus narrowly rimate, or occasionally imperforate. Per- iostracum yellow-green with oblique, shallow growth striations. Protoconch (Fig. 40) with fine spiral threads at and below periphery; with dimples superimposed on spiral threads above periphery. Aperture broadly ovate, 0.83-0.95 times as high as wide, 0.66-0.73 times height of last whorl. Plane of aperture at 22-30° to axis of shell. Peristome complete across parietal area by a thick callus. Basal lip slightly receded; base-columellar corner pro- jecting forward. Posterior corner of aperture advanced forward and with a shallow angular groove internally. Columellar lip very wide; rounded in cross section; straight or weakly concave in lateral profile; vertical. Columellar lip forming a pronounced angle with parietal callus. Measurements in mm for the holotype and ten paratypes selected to show variation fol- low (holotype in parenthesis). Height, 3.5—4.0 (3.9); width, 3.4—4.0 (3.8); aperture height, 2.4—2.8 (2.8); aperture width, 2.0—2.5 (2.4). Type-locality. Georgia, Upson County, Flint River at Spewrell Bluff, U.S. Army Corps of Engineers river mile 200. Holotype: UF 40500; collected 22 May 1981 by Fred G. Thompson. Paratypes: UF 31244 (246 speci- mens); 25 specimens each deposited in ANSP, FMNH, MCZ, UMMZ, USNM, Rijks- museum van Natuurlijke Historie, Leiden, Netherlands, Senckenbergische Naturfors- chende Gesellschaft, Frankfurt-am-Main, Germany, and Herbert D. Athearn collection; same data as holotype. Distribution. Endemic to the middle section of the Flint River in Georgia from Meriwether- Pike counties southeast to Taylor-Crawford counties. This species has been found only in shoals and rapids where it occurs on granite boulders and gravel in moderate currents. Locality records are given in Appendix C. Gillia altilis (Lea) Melania altilis Lea, 1841: 13.—Lea, 1843: pl. 5, figs. 23. Leptoxis altilis (Lea), Haldeman, 1847: 6, pl. SO 152. Gillia altilis (Lea), Stimpson, 1865a: 53.— Stimpson, 1865b: 51.—Binney, 1865: 74— 75, fig. 146.—Walker, 1918: 32-33, figs. 115-116.—Burch 4 Tottenham, 1980: 104, fig. 191.—Burch, 1982: 23, fig. 191. Somatogyrus altilis (Lea), Tryon, 1870: 60, pl. 17, fig. 9. Leptoxis crenata Haldeman, 1847: 6, pl. 5, fig. 153: Gillia crenata (Haldeman), Binney, 1865: 74— 75, figs. 147-148. Shell (Figs. 63, 64). Conico-globose. Light yellow-green. About 4.5 whorls, but apex usually eroded, leaving 2—4 whorls in adults. Moderately large, eroded adults usually 6— 8mm high (lectotype 6.6 mm). Body whorl NORTH AMERICAN LITHOGLYPHINAE 133 70 FIGS. 65-70. Figs. 65-66. Somatogyrus depressus (Tryon), UF 34969: Mississippi River, Davenport, lowa. Fig. 67. Somatogyrus rheophilus n. sp.; holotype: UF 40500. Fig. 68a. Clappia umbilicata (Walker); ANSP 95037: Coosa River, Alabama. Figs.69-70. Birgella subglobosa (Say); UF 35008: Ohio River at Five Mile Creek, Hamilton Co., Ohio. 134 THOMPSON conspicuously enlarged; adults tending to be shouldered or fluted below suture; height of body whorl 0.95-1.10 times width. Shell usually rimate; some specimens imperforate, or narrowly umbilicate. Apical whorl of pro- toconch elevated, 1.25 mm wide transverse to initial suture. Protoconch sculptured with fine spiral threads that are uniformly dis- persed over surface of first whorl (Fig. 41). Subsequent whorls strongly rounded; sculp- tured with distinct, regularly spaced in- cremental striations; usually with 1—2 dark growth varices. Aperture broadly ovate- auriculate in shape. Plane of aperture lying at an angle of 18-20” to axis of shell. Height of aperture 0.70-0.78 times height of body whorl; about 0.78-0.90 times as wide as high. Peristome in mature specimens complete across parietal wall by a thin callus; incomplete in sub-adults; peristome dark rimmed. Col- umellar lip moderately thickened, rounded. Outer lip and basal lip sharp-edged, evenly curved through columella. Outer lip con- spicuously arched forward in lateral profile (Fig. 64). Operculum (Fig. 52) oval in shape. Chitinous, yellowish-green. Paucispiral, con- sisting of three whorls. Nucleus located in the lower left third. Outer surface sculptured with fine incremental striations. Measurements in mm based on 29 speci- mens selected to show variation are given below. UF 27500—Lake Waccamaw, North Carolina; UF 35027—Potomac River, District of Columbia; UF 35013—Erie Canal, New York. Cat. no. n Total h. Body wh. h. Width Apert. h. Apert. w. UF 27550 10 5.6-5.9 5.4—6.1 5.3—5.8 4.0—4.5 3.5-3.9 UF 35025 8 6.5-8.1 5.5-7.4 5.3-7.6 3.9-5.5 3.54.5 UF 35013 10 5.6-7.0 5.0-6.1 4.9-6.1 3.5—4.6 3.0-3.9 UF 40550 neotype 6.6 5.9 5.6 4.3 3.9 Type-locality. Lea (1841) stated that his specimens of Melania altilis came from the Santee Canal, South Carolina, and the Sus- quehanna River at Havre de Grace, Maryland. Haldeman (1847) stated that the type- specimen of Leptoxis crenata came from the Santee Canal, South Carolina. Type- specimens for neither Melania altilis Lea nor Leptoxis crenata Haldeman can be located. Presumably they are lost. It is clear that they are the same species, and a neotype must be designated. Melania altilis Lea: Neotype UF 40550 (Fig. 63). Leptoxis crenata Haldeman: Neotype UF 40550; same specimen as neotype for Melania altilis Lea. Neotype local- ity: Lake Waccamaw, Columbus County, North Carolina; neotype collected 12 Septem- ber 1980 by Fred G. Thompson. Lake Wacca- maw is selected as the neotype locality for the following three reasons. The species shows little variation throughout its range, and there is no basis to suspect that the Lake Wacca- maw population is different from other pop- ulations in its taxonomic identity. On three occasions in 1980, 1981, and 1982, | was unsuccessful in finding the species in the Santee River. Apparently it no longer occurs there. The anatomical data given in this paper are based on Lake Waccamaw material. Distribution (Fig. 71). The species is widely distributed in rivers draining into the Atlantic Ocean from South Carolina north to New York and Vermont. It has entered the Lake Ontario system via the Erie Canal. The species is found in quiet lakes and rivers, as well as fast gradient streams. Locality records for this species are given in Appendix C. APPENDIX B A North American snail incorrectly associ- ated with the Lithoglyphinae: Birgella Baker, 1926 Birgella Baker, 1926: 196.—Baker, 1928: 154-155.—Wenz, 1939: 575.—Thompson, 1979: 47.—Burch 8 Tottenham, 1980: 110. (Type-species: Paludina subglobosa Say, 1825). The shells of B. subglobosa (Say) (Figs. 69—70) are so similar to those of Somatogy- rus depressus (Figs. 65—66), the type-species of Somatogyrus, that they were considered congeners for more than a century. Baker (1928) separated Birgella from Somatogyrus on the basis of differences in the verge and radula, but placed both genera in the Litho- glyphinae. Most subsequent authors failed to recognize Birgella as a distinct genus due to NORTH AMERICAN LITHOGLYPHINAE 135 Fig. 71 ® 5. depressus в G altilis Fig: 72 e В. subglobosa FIGS. 71-72. Geographic distributions. Fig. 71. Gillia altilis (Lea) and Somatogyrus depressus (Tryon). Fig. 72. Birgella subglobosa (Say). 136 THOMPSON FIGS. 73-80. SEM photographs of radulae and protoconch sculpture. Figs. 73-75. Nymphophilus minkleyi Taylor, UF 34905: creek 10 km SW Cuatro Cienegas, Coahuila, Mexico. Figs.76-78. Birgella subglobosa (Say), UF 35275, Alabama River, Choctaw Bluff, Clarke Co., Alabama. Figs. 79-80. Birgella subglobosa (Say), UF 35277: Ohio River at Five Mile Creek, Hamilton Co., Ohio. Enlargements: Figs. 73-75, 78 х356. Figs. 76, 77 х238. Figs. 78-80 x48. NORTH AMERICAN LITHOGLYPHINAE 137 82 ze Е ee 84 FIGS. 81-84. Birgella subglobosa (Say). Figs. 81-82. Verge. Fig. 83. Female reproductive system without ovary. Fig. 84. Mid-segment of oviduct showing relationships between seminal receptacle, bursa copulatrix, and albumen gland. Legend: ag = albumen gland; bu = bursa copulatrix; cg = capsule gland; emc = posterior wall of mantle cavity; ov = oviduct; sr = seminal receptacle. the limited anatomical data available on the Hydrobiidae. Thompson (1979) placed Birgel- la) in the Nymphophilinae, where it is clearly related because of features of the reproduc- tive system and radula. Clarke (1981) contin- ued to treat Birgella as a synonym of Somato- gyrus. Birgella is differentiated from other nymphophilids because of its protoconch sculpture, its large globose shell and its ponderous penis. It is not closely related to other known genera. The protoconch sculp- ture is vaguely similar to that of Nymphophilus because of its coarse step-like wrinkles (see Thompson, 1979). The morphology of the penis has some similarities to that of Marsto- nia. Both have a single apocrine gland con- fined to the apical lobe (see Thompson, 1977, for a discussion of the morphology of Marsto- nia). The two genera have such dissimilarly shaped penises, the apocrine gland pattern must be considered convergent. Birgella is characterized within the Nymphophilinae as follows. Shell (Figs. 69-70). Large, about 6-9 mm high with about 4.3 whorls; globose, usually about 0.83-0.87 times as wide as high; some specimens as wide as high with a very ample aperture and a depressed spire. Protoconch sculptured with step-like, rugose wrinkles with superimposed spiral threads (Fig. 79-80). The spiral threads are a unique feature within the subfamily. Operculum paucispiral with about 2.5 whorls. Penis (Figs. 81-82). With a large globose apical lobe on the left side and a short stocky tip on the right side. Tip pigmented with mela- nophores. Apical lobe with a large circular apocrine gland on the inner surface (in- correctly reported to be absent in an earlier report (Thompson, 1979)). Female reproductive system (Figs. 83-84). Typical for subfamily Nymphophilinae (see 138 THOMPSON Thompson, 1979). A single seminal recepta- cle present and completely buried in albumen gland. Ventral canal of anterior pallial oviduct spirally offset. Lumen of canal continuous with capsule gland lumen (Fig. 84). Radula (Figs. 76—78). With large acuminate cusps. Central tooth with 2-3 basocones on each side located on a reflected lateral ridge. Lateral teeth as in other Nymphophilinae, with a rounded basal lobe. Radular data for two specimens | examined (UF 35275, 35278) are given in Table 1. Baker (1928) and Berry (1943) gave slightly different cusp counts. The radula of Nymphophilus minkleyi Taylor is figured for comparison (Figs. 73— 75). Birgella contains a single species. Birgella subglobosa (Say) Paludina subglobosa Say, 1825: 125.— Haldeman, 1847: 10-11, pl. 10, fig. 7 (Type-locality: Northwestern Territory). Somatogyrus subglobosus (Say), Tryon, 1870: 60-61, pl. 17, figs. 10-11.—Baker, 1902: 340-341, fig. 123.—Berry, 1943: 49— 52, pl. 2, fig. 1 (shell), pl. 4, fig. 3 (radula), text-fig. 8 (penis). Melania isogona Say, 1829: 277. (Type- locality: Bear Grass Creek, near Louisville, Kentucky). Somatogyrus isogona (Say), Stimpson, 1865b: 22.—Binney, 1865: 77-78, fig. 151. Birgella s. subglobosa (Say), Baker, 1928: 155-158.—Burch 8 Tottenham, 1980: 110, figs. 188, 198, 202. Birgella subglobosa isogona (Say), Baker, 1928: 159-161, pl. 8, figs. 10-12. Paludina pallida Lea, 1839: 22, pl. 23, fig. 104. (Type-locality: near Cincinnati, Ohio). This is a well-known North American spec- ies. Detailed descriptions are given in Baker (1928) and Berry (1943). Some authors rec- ognize two subspecies. The typical subspec- ies is said to be narrowly umbilicate or rimate and to have a thin parietal callus. The other subspecies is said to be imperforate and to have a thicker parietal callus (Baker, 1928). The material | have examined shows that these characters vary throughout the range of the species, and that they do not segregate with any geographic or ecological factor. In- deed, the differences seem to be con- sequences of growth. Old specimens have a thicker callus which tends to cover the umbili- cus. Thus the two forms do not meet the accepted criteria for subspecies. Birgella subglobosa is widely distributed in the central United States from Wisconsin, Michigan, and Ohio south to Arkansas and Alabama (Fig. 72). It also occurs in New York in the Mohawk and Hudson River systems. Usually it is found in large rivers and lakes. It is not confined to deep water, contrary to published statements. | have collected it in bays and sloughs at less than 1 meter depth. The snail is found most commonly in quiet water on a soft silt substrate. Specimens ex- amined are listed in Appendix C. APPENDIX C Specimens of HYDROBIIDAE examined dur- ing this study are from the following museum collections, which are designated as indicated in parenthesis: Academy of Natural Sciences, Philadelphia (ANSP), Carnegie Museum, Pitt- sburgh (CM), Field Museum of Natural His- tory (FMNH), Museum of Comparative Zoolo- gy, Harvard University (MCZ), Florida State Museum, University of Florida (UF), Museum of Zoology, University of Michigan (UMMZ), National Museum of Natural History (USNM). Lepyrium showalteri (Lea) ALABAMA.—Bibb Co. Cahaba River (UMMZ 97447, UMMZ 97445, MCZ 133133); Cahaba River, near Anita (UMMZ 68301); Cahaba River, 1.6 km N. Centerville (USNM 672419, MCZ 252247); Cahaba River, Lilly Shoals (UMMZ 49272); Cahaba River, near Piper (MCZ 99174, UMMZ 69886, UMMZ 65996); Little Cahaba River (MCZ 99175, UMMZ 97444, UMMZ 97448); Little Cahaba River, 4.8 km E Piper (UMMZ 67444). Dallas Co.: Cahaba River Wildcat Island, near Cane Creek (UMMZ 87446). Shelby Co.: Coosa River (USNM 29016); Gurnee (MCZ 299176, USNM 321180, 321181); Coosa River, 16.1 km above Ft. Williams (USNM 102851, lectotype by present designation; USNM 102851a). Fort Williams was at the conflu- ence of the Coosa River and Cedar Creek, W. of Fayetteville, Talladega Co., Alabama. Clappia clappi Walker (= Clappia umbilicata Walker) ALABAMA.—Chilton Co.: Coosa River, Duncans Riffle (ANSP 95307, paratypes). NORTH AMERICAN LITHOGLYPHINAE 139 Somatogyrus depressus (Tryon) ILLINOIS.—DeKalb Co.: Kishwawkee Creek (MCZ 46592). Fulton Co.: Canton (UMMZ 116934, UF 34975). Hardin Co.: Eliz- abethtown, Ohio River (UMMZ 49781, UF 34976). Rock Island Co.: Rock Island (USNM 476878, USNM 512075). Stephenson Co.: 3.2 km S. Freeport (USNM 49781). Washing- ton Co.: Okaw River, Covington (MCZ 68437). IOWA.—Cherokee Co.: Cherokee (USNM 507933). Clinton Co.: Clinton (USNM 539849). Dickinson Co.: NE Okoboji (USNM 667011). Dubuque Co.: Dubuque (UMMZ 143740). Emmet Co.: Estherville (USNM 506130); Des Moines River near Estherville (USNM 526533). Hardin Co.: Eldora (USNM 506379, USNM 519429, USNM 514807): lowa River, Eldora (FMNH 130351). Hum- boldt Co.: Dakota City (USNM 526532). John- son Co.: lowa City (USNM 506380, UMMZ 69940). Muscatine Co.: Muscatine (USNM 508369). Scott Co.: Davenport (USNM 121041, USNM 27904, USNM 38414—SEM radula, UF 34968-34970). MISSOURI.—Benton Co.: Warsaw, Osage River (UMMZ 67435, UF 34980). WISCONSIN.—Brown Co.: De Pere (FMNH 10649). Jefferson Co.: Pipersville Rapids, Bark River (UMMZ 116940, UF 34979); Watertown (UMMZ 143742, UF 34977). Rock Co.: Evansville (CM 62.24127). Sauk Co.: Wisconsin River (UMMZ 143743, UF 34981); Prairie du Sac (UF 34978). Somatogyrus rheophilus Thompson GEORGIA.—Meriwether Co.: Flint River, 2.7km NE Gay (UF 40508); Flint River, 2.1 km E Gay (UF 40502); Flint River, 5.6 km SE Gay (UF 40503, 40506). Ta/bot Co.: Flint River, 5.1 km NW Carsonville (UF 40501); Flint River, 3.5km NNW Fickling Mill (UF 40509). Taylor Co.: Flint River, 5.8 km NW Fickling Mill (UF 40507). Upson Co.: Flint River at Spewrell Bluff (type-series); Flin Riv- er, 11.9km WSW Thomaston (UF 40504); Flint River at Yellowjacket Shoals, 9.7 km SW Thomaston (UF 31241); Flint River, 9.7 km SW Lincoln Park (UF 34902, SEM shell; UF 34903, SEM radula; UF 40511); Flint River, 11.3 km SSW Lincoln Park. Gillia altilis (Lea) DISTRICT OF COLUMBIA.—Arlington Co.: Analston (USNM 336089, USNM 252023); Anacostia River, Buzzard’s Point (USNM 697026); Popular Point (USNM 697025): An- alston Id. (USNM 697015); between P.R.R. Bridge and Pa. Ave. Bridge (USNM 697021): C & O Canal (USNM 335872); near Asylum Wharf, East Branch (USNM 697023); Poto- mac River (CM 62.24224); above Long Bridge (CM 62.25465); Fox Ferry (USNM 251542, USNM 271700, USNM 465805, MCZ 2181, UMMZ 28918); E. Branch, Potomac River (UF 35025); Potomac River (UF 35024, UF 1002). MARYLAND.—Allegany Co.: Cumberland (USNM 149952); Fall of the Potomac, Poto- mac State Forest (MCZ). Cecil Co.: near Charlestown (USNM 521817). Hartford Co: Havre de Grace (USNM 121450). Montgom- ery Co.: Cabin John, C & O Canal (MCZ 2179); Sycamore Island (USNM 521974). Prince Georges Co.: Fort Washington (UMMZ 118414, UMMZ 364722); Fort Washington, Potomac River (USNM 227686); Fort Wash- ington, Piscataway Creek (UF 35021). NEW JERSEY —Burlington Co.: Burlington (USNM 120468); Burlington, Delaware River (MCZ 57089). Essex Co.: Newark, Morris Canal (MCZ 186744). Hunterson Co.: Lam- bertville (USNM 536807). Mercer Co.: Raritan Canal, aqueduct near Princeton (CM 62.5699). Sussex Co.: Flatbrookville (MCZ 75217). Warren Co.: Phillipsburg, Delaware River (FMNH 87953). NEW YORK.—Albany Co.: Albany (CM): Albany, Hudson River (USNM 465755, FMNH 59965, UF 35015). Dutchess Co.: Tivoli, Hud- son River (MCZ). Herkimer Co.: Mohawk, Erie Canal (UMMZ 45998, UF 35013, USNM 697027, USNM 697028, UMMZ). Monroe Co.: Brighton (UMMZ 118415, UF 35014). Niagara Co.: Niagara Falls (USNM 473979). Onondaga Co.: Syracuse, Erie Canal (UMMZ 69880, UF 35019, FMNH 58688, UMMZ 69880). Rensselaer Co.: Troy (UMMZ 118412, UF 35012); Troy, Champlain Canal (MCZ 2178). Ulster Co.: Heath, Hudson River (MCZ 186739). Wayne Co.: Clyde (USNM 597809). NORTH CAROLINA.—Columbus Co.: Lake Waccamaw (UF 28439, UF 29652, UF 28077, UF 29637, UF 29644, UF 27550, UF 35044, UF 34901—SEM shell, UF 34816— SEM radula). Edgecombe Co.: Swift Creek at NC Hwy. 97 (UF). Nash Co.: Tar River, S of Moccasin Creek (UMMZ 197725). New Hanover Co.: Wilmington, Greenfield Pond (UMMZ 69881). Pitt Co.: Little Continea Riv- er, 9.7 km SE Farmville (UMMZ 197724); S of Sandy Cross (UMM2). 140 THOMPSON ONTARIO.—Lincoln Co.: Niagara-on-the- Lake (MCZ 104863). PENNSYLVANIA.—Bucks Co.: Delaware River, New Hope (CM 62.5700). Clinton Co.: Flemington (USNM 28102). Chester Co.: Schuylkill River, Phoenixville (FNMH 87956). Lancaster Co.: Columbia (CM 62.16370, MCZ 2180, MCZ 186745). Lycoming Co: Muncy, canal (MCZ). Northampton Со.: Delaware River, Easton (FMNH 15693, UF 35020). Philadelphia Co.: Philadelphia (FMNH 87950); Schuylkill Canal, Manayunk (CM 62.5701); Philadelphia (MCZ 186743). SOUTH CAROLINA.—Charleston Co.: Charleston (MCZ). Williamsburg Co.: Lynche’s Creek (USNM 63973). VERMONT.—Franklin Co.: St. Albans Bay (CM 62.32653). Grand Isle Co.: Grand Isle, Lake Champlain (UMMZ 118422, UF 35018); Lake Champlain, Chimney Point (USNM 336443, USNM 336442, USNM 336445, USNM 336444, USNM 591730). VIRGINIA.—Alexandria Co.: Potomac Riv- er (MCZ 186741, MCZ 70540). Amherst Co.: James River, Lynchburg (USNM 451904). Cumberland Co.: Cartersville, James River (MCZ 261289). Fairfax Co.: Dyke, near Mt. Vernon (USNM 420546); Mt. Vernon (UF 18435); near Great Falls (numerous lots, USNM, MCZ, UMMZ, UF). Goochland Co.: Columbia, James River (MCZ 261334). Hen- rico Co.: Richmond, James River (CM 62.24126). Loudoun Co.: Potomac River, 6.4km N. Seneca, MD (USNM 697018). Powhatten Co.: James River across from Maidens (MCZ 261307). Prince Co.: Peters- burg (FMNH 87949, USNM 121477). WEST VIRGINIA.—Jefferson Co.: Harper's Ferry (MCZ 136500); Harper’s Ferry, Poto- mac River (MCZ). Morgan Co.: Cherry Run, Potomac River (numerous lots UMMZ, MCZ, FMNH, UF). Fluminicola nuttalliana (Lea) OREGON.—Lane Co.: Willamette River, Eugene (UF 40521). Linn Co.: Willamette Riv- er, Albany (UF 40523, UF 40524, SEM shell). Birgella subglobosa (Say) ALABAMA.—Choctaw Co.: Tombigbee River, Ezell Fish Camp, E. of Lavaca (UF 35093). Clarke Co.: Alabama River, Choctaw Bluff (CM 65-57). Colbert Co.: Tennessee River, Mile 261.0, Union Carbide (UF 34998). Limestone Co.: Tennessee River, Mile 291.76. Brown’s Ferry (UF 35001); Tennes- see River, Mile 288.78, Brown’s Ferry (UF 34999). Monroe Co.: Alabama River, Clair- borne (UF 35099). Sumter Co.: Tombigbee River, Lock #3, ESE of Whitfield (UF 35095). ARKANSAS.—Dallas Co.: Ouachita River, 3.9km W. Sparkman (UF 35002). Jackson Co.: White River, Newport (MCZ 66659). GEORGIA.—Floyd Co.: Silver Creek (UF 40639). ILLINOIS.—Cass Co.: Beardstown, Illinois River (UMMZ 197758, FMNH 15694). Cook Co.: Chicago, Lake Michigan (FMNH 71888). Fulton Co.: Canton (FMNH 71890). Gallatin Co.: Shawneetown (FMNH 115352). Kank- akee Co.: Kankakee Feeder (FMNH 58685). Madison Co.: Alton, Mississippi River (UMMZ 197760). Mercer Co.: Mississippi River (UMMZ 143748, MCZ 2202); Myers Slough (MCZ 2201). Pope Co.: Golconda, Ohio River (UMMZ 197755). Rock Island Co.: Moline, Mississippi River (USNM 465760). Will Co.: Dupage River, Joliet (FMNH 58686); Joliet (CM 62.25453). Williamson Co.: Blaireville, Big Muddy River (UMMZ 117191). INDIANA.—Dearborn Co.: Ohio River, Lawrenceburg (ЕММН 87928). Floyd Co.: (FMNH 58687). Marshall Co.: Lake Maxink- uckee (USNM 697009). IOWA.—Johnson Co.: (FMNH 58701); lowa City (UMMZ 143749). Lee Co.: Mis- sissippi River, 4.7 km N Keokuk (UF 35009); Keokuk, pool above dam (MCZ 175918); Ft. Madison (UF 34997); Montrose (UF 35006). Muscatine Co.: Keokuk Lake (USNM 600748); Muscatine, Mississippi River (UF 31307). Polk Co.: Des Moines, Des Moines River (MCZ 2196); Des Moines, Bayou at N end Fort Dodge (UF 35011). Scott Co.: Le Clair, Mississippi River (MCZ 2197); Mis- sissippi River, Davenport (UF 35003). KENTUCKY.—Campbell Co.: mouth of Five-Mile Creek (UMMZ 70020). Jefferson Co.: Louisville, Falls of the Ohio (FMNH 87919). MINNESOTA.—Washington Co.: Ft. Snell- ing, Minnesota River (MCZ 2200). MISSOURI.—St. Louis Co.: Mississippi River near White House (UMMZ 177032); Jefferson Barracks (UMMZ 197756); Kirk- wood, Meramec River (UMMZ 197759). NEW YORK.—Herkimer Co.: Mohawk (FMNH 15520); Mohawk, Erie Canal (MCZ 62233); Mohawk, Mohawk River (CM 62.7056). Schenectady Co.: Schenectady (MCZ). OHIO.—Erie Co.: Sandusky, Lake Erie NORTH AMERICAN LITHOGLYPHINAE 141 (CM 62.25454). Franklin Co.: Columbus (USNM 30139); Columbus, Ohio Canal (CM 62.7057). Green Co.: Clifton, Miami Canal (USNM 28515). Hamilton Co.: Cincinnati, (FMNH 115552); Cincinnati, Ohio River (CM 62.25457); Culloms Riffle, Ohio River (FMNH 87924); Five-Mile Creek, Ohio River (FMNH 87922): Mouth of Great Miami River (FMNH 87913); old canal bed near Harrison (FMNH 87912); 9.7 km W. Cincinnati (UMMZ 45448). Scioto Co.: Portsmouth, Ohio River (CM 62.8229). Summit Co.: (UF 35010). Tuscar- awas Co.: Mill Race on Ohio Canal, New Philadelphia (CM 62.26560); Tuscarawas River, New Philadelphia (CM 62.25456). QUEBEC.—Rouville Co.: Richelieu River, 3.2 km S. Iberville (MCZ). TENNESSEE.—Nolachucky River (UF 35007). VERMONT.—Addison Co.: Chimmey Point, Lake Champlain (MCZ 28232); Hospi- tal Creek (USNM 336445). St. Franklin Co.: Lake Champlain, St. Albans Bay (MCZ 142046). Grand Isle Co.: Lake Champlain, Grand Isle, 3.2 km SE Hero (MCZ). VIRGINIA.—Fairfax Co.: Great Falls (USNM 252381) (doubtful record). WISCONSIN.—Milwaukee Co.: Milwaukee (MCZ). MALACOLOGIA, 1984, 25(1): 143-160 IMPLICATIONS OF RADULAR TOOTH-ROW FUNCTIONAL INTEGRATION FOR ARCHAEOGASTROPOD SYSTEMATICS Carole S. Hickman Department of Paleontology, University of California, Berkeley, CA 94720, U.S.A. ABSTRACT Functional analysis of complex morphology can be used to generate new sets of taxonomic characters. It may also suggest ecological or biomechanical relationships among characters that otherwise might appear adaptively unrelated. Analysis of tooth-row functional integration in the rhipidoglossan archaeogastropod radula reveals mechanically and taxonomically distinctive interactions between tooth bases, shafts, and cusps. Many of these interactions follow a complex sequence that is not related to the standard concept of teeth arranged in rows and columns. In some groups, rows are impossible to identify in a functional sense, and the concept of base rows and cusp rows is introduced to help clarify previously unappreciated aspects of mechanical integration. Drawing phylogenetic inferences from functional characters carries the burden of demonstrat- ing that convergence has not occurred. In rhipidoglossan radulae the characters involved in integrating tooth-row function are superimposed on conserved groundplans. Even where two morphological solutions to the same functional problem are extremely similar, traces of underly- ing phylogenetic differences remain. Adaptations of cusps for transmitting forces from one tooth to another are taxonomically distinctive as are the shaft expansion and interleaving patterns that transmit force in different ways in different taxa. Basal interaction patterns are particularly elegant because they perform many mechanical functions simultaneously. The lateromarginal plate, a tooth base that is modified as an articulatory structure, has developed independently in different rhipidoglossan groups and is particularly useful in evaluating phylogenetic relationships. INTRODUCTION Taxonomists traditionally have plied their trade with tools for assessing the numbers, sizes, shapes, and qualities of morphology that are best observed in static, non- functioning (usually dead) organisms. Func- tion and principles of design and construction frequently are tied to character convergences and are such poor indicators of evolutionary relationships at a superficial level that many systematists have avoided functional charac- ters in classification and phylogenetic infer- ence. And yet superficially similar morphologies may have subtle but profoundly different un- derlying designs that reflect quite different phylogenetic heritages. Such differences may be more obvious in living, functioning organ- isms. Functional studies can provide valuable clues for reassessment of morphology and for generating new and more refined sets of taxonomic characters. For the molluscan taxonomist, the radula is an excellent example of a complex morpho- logical apparatus that does not really have much of a story to tell when only numbers, sizes, and shapes are considered. Radulae have been used extensively in molluscan classification for many years; but their use has, on the whole, been uncreative. The vocabulary for describing radulae is appal- lingly depauperate, particularly with respect to shape descriptors, in light of the incredible complexity and array of morphologies that have developed within the phylum. One way to circumvent the vocabulary problem with complex morphology is simply to illustrate it. “Here is the Pleurotomaria radula, and a picture is worth a thousand words.” Other alternatives, which have not yet come of age, are computer-assisted multi- variate morphometric analysis and computer graphic analysis from scanning electron stereo-micrograph pairs, which make it possi- ble for the first time quickly to generate, store, and evaluate vast quantities of data (Schooley, Hickman & Lane, 1982). However, a computer will recognize and analyze only what the taxonomist tells it to (143) 144 HICKMAN recognize and analyze, and new characters are introduced into systematics only through developing new ways of observing and assessing morphology (Hickman, 1977, 1980). In this paper | show how analysis of tooth-row functional integration in the rhipi- doglossan radula has led to recognition of sets of characters that are extremely useful in taxonomy. They are characters that, once they have been pointed out, seem so fun- damental that it is surprising that they could have been overlooked for so many years by taxonomists. Functional integration of movement in the radulae of rhipidoglossate marine ar- chaeogastropods involves complex com- binations of between-row and within-row in- teractions of tooth bases and cusps and within-row interactions of tooth shafts. It may also involve interaction facilitated by highly modified plates that no longer function as teeth. Categories of interactions are con- sistently associated with unique variations on basic morphological themes at higher taxonomic levels. One of the most remarkable features of this integration is revealed in the observation that cusp rows and base rows in the gastropod radula do not necessarily correspond. Integration may, in fact, be so complex as to make it difficult to define rows at all within radulae. These problems and their taxonomic implications are explored be- low. METHODS Uniting taxa on the basis of common func- tional characters makes one major demand: the taxonomist must be able to infer which aspects of morphology are conservative and to demonstrate that convergence has not oc- curred. In a previous paper (Hickman, 1980) | identify seven factors that contribute to form and pattern in gastropod radulae and show that underlying phylogenetic factors (con- served groundplans) are readily identifiable. Even where two morphological solutions to the same functional problem are extremely similar, the underlying phylogenetic dif- ferences are generally not obscured. Asymmetry provides an example of a func- tional character that is very useful in systemat- ics and one that has been overlooked by taxonomists (Hickman, 1981). Asymmetry has developed in radulae with large, strongly- cusped pairs of major food-preparing teeth; it functions to facilitate efficient, alternate fold- ing of these large teeth as the radula is with- drawn from the substrate and into the buccal cavity. It also allows economical, compact storage of teeth in alternate, zipper-fashion within the radula sac. Asymmetry has de- veloped convergently at a superficial level; but when one begins to examine the nature of the asymmetry in detail, one rapidly discovers that it is not simply a presence-absence char- acter. It exists in many different forms or states, and these states are consistently associated in archaeogastropods at the su- perfamily level (Hickman, 1981). In the examples that are developed below, morphological and functional data have been obtained from a combination of observations of gastropods radulating glass surfaces and frame-by-frame analysis of slow-motion cine- matograpy (Morris & Hickman, 1981); dis- sections of radulae and manipulations of ex- cised radulae; flat mounts examined with light microscopy; and a combination of single scanning electron micrographs from a variety of angles and stereo paired micrographs (Hickman, 1977). MORPHOLOGY AND FUNCTION OF THE ROW IN GASTROPOD RADULAE Descriptions of the molluscan radula con- sistently refer to the orderly arrangement of teeth in “rows” (e.g. Fretter & Graham, 1962: 169; Hyman, 1967: 236; Purchon, 1968: 45; Solem, 1974: 139; Yonge & Thompson, 1976: 49). The term “transverse row” has been ap- plied by some authors to refer to the full complement of different kinds of teeth (e.g. rachidian, laterals, marginals) that are se- creted multiple times during ontogeny from a basic set of genetic instructions. Because each transverse row is repeated numerous times, there are also orderly longitudinal series of teeth. The term “longitudinal row” is confusing, however, and | prefer to refer to teeth in longitudinal series as “columns” (Hickman, 1980). Fig. 1A illustrates a very clear example of orderly arrangement of rows and columns in the radula of a siphonariid limpet, where teeth are relatively undif- ferentiated. Fig. 1B illustrates rows and col- umns in a more complex radula in which there is a great deal of differentiation and many different column morphologies. The distinction between rows and columns is fundamental and obvious and requires no ARCHAEOGASTROPOD RADULAE 145 further discussion. However, having defined a tooth-row as the basic unit of radular morphol- ogy, | want to expose a heretofore un- appreciated, but serious, discrepancy be- tween morphological rows and functional rows of teeth. There is no problem defining a morpholog- ical tooth-row if one simply isolates all of the different kinds of teeth within the fundamental FIG. 1. Radulae in which rows and columns are easily distinguished. A. Siphonaria sp., a pulmon- ate limpet, with a pavement of relatively un- differentiated teeth arranged in rows (diagonals from top left to bottom right) and columns (di- agonals from top right to bottom left). Scale bar = 100 um. В. Cantrainea panamense (Dall, 1908), a turbinid with greater differentiation of teeth but retaining distinctive pattern of rows and columns. Scale bar = 40 um. unit. These have been illustrated by taxono- mists in drawings for more than a century and represented by means of general radular for- mulae that specify numbers (and sometimes relative sizes and numbers of cusps) of differ- ent kinds of teeth. The difficulty arises in beginning to look at radular function and the ways that teeth in- teract with one another: at this level we can distinguish “cusp rows” and “base rows.” In most radulae cusp rows and base rows coin- cide; but in others there are remarkable pat- terns of deviation such that a transverse row of interlocked tooth bases gives rise to cusps that occupy and function in two or three differ- ent cusp rows. In some radulae the in- teractions between teeth are so complex that it is difficult, if not impossible, to define rows in non-arbitrary fashion, even though columns are clearly distinguishable. Deviations from the coincidence of base rows and cusp rows follow patterns that are consistently repeated within higher taxa. Fig. 2 illustrates two func- СОР ROW BASE ROW FIG. 2. Two distinct row types in rhipidoglossan radulae. A. The standard condition, exemplified by Trochus intextus Kiener, 1850, in which base rows and cusp rows coincide. B. The fissurelline condi- tion, exemplified by Fissurella nimbosa (Linné, 1758), the type of the genus, in which cusp rows (dashed line) are distinct from base rows (dotted line). 146 HICKMAN tionally and taxonomically distinct kinds of rows in rhipidoglossan archaeogastropod radulae. The standard condition In Fig. 2A the fundamental pattern of coinci- dence is illustrated by the radula of Trochus intextus Kiener, 1850. In the genus Trochus Linné, 1758, and other taxa within the sub- family Trochinae, the bases of the rachidian and lateral teeth are interlocking, with proc- esses that fit in ball-and-socket fashion into depressions on adjacent tooth bases. Be- tween the marginal and lateral teeth there is a specialized marginal tooth with a greatly en- larged and modified base and shaft that serves the function of a lateromarginal plate, facilitating interactions between the two major portions of the row (Hickman, 1980). The most important thing to note is that the cusps in the radula of Trochus also overlap and interact within each row and that interacting cusps correspond with interacting bases across each row (Fig. 2A). In other trochid genera and subfamilies cusps and bases may interact in different ways. For example, in the genus Gaza Wat- son, 1879, lateral tooth bases do not interlock but broadly overlap one another (Fig. 3A), while in the genus Solariella Wood, 1842, the pattern of interaction is much more com- plicated and involves teeth of more highly differentiated morphology (Fig. 3B). But both genera resemble Trochus in that tooth shafts are relatively short and give rise to an arcuate row of interacting cusps that correspond to the row of tooth bases. Likewise, in the family Turbinidae, cusp and base rows correspond and basal interactions involve both overlap and interlock (Fig. 3C). It is reasonable to ask at this point whether cusps, which appear to interact in an excised, flat-mounted radula, actually comprise a func- tional unit in the living gastropod during feed- ing. The answer is partially apparent in Figs. 4A and B, where the radula of the northeast- ern Pacific black turban snail, Tegula funebra- lis (A. Adams, 1854), is illustrated as seen during the feeding stroke. Morris (1980) and Morris & Hickman (1981) have shown that the tooth-rows assume a tight semicircular con- figuration during feeding due to certain mechanical properties of flexible slit cylinders. Rows maintain their integrity throughout the complicated deformation of the slit cylinder during each feeding stroke. However, mi- crocinematographic frame-by-frame analysis of a feeding stroke demonstrates that the sequence of teeth that passes over a fixed point on the substrate corresponds neither to a sequence of teeth within a row nor to a FIG. 3. Basal interactions in trochacean radulae. A. Simple overlap of broadly expanded bases in Gaza superba (Dall, 1881). Scale bar = 100 um. В. Shallow interlock in Microgaza sp. Scale bar = 100 рт. С. Deep interlock in Homalopoma carpen- teri (Pilsbry, 1888). Scale bar = 40 um. ARCHAEOGASTROPOD RADULAE 147 FIG. 4. Operational configuration of tooth-rows in the radula of Tegula funebralis (A. Adams, 1854). A. Optical micrograph of feeding stroke of a living snail. Scale bar = 1mm. B. Scanning electron micrograph of an artificial protrusion of radula into same functional conformation. Scale bar = 400 um. Note striking difference between tight semicircular configuration of functioning row and standard configuration illustrated in flat radula mounts (from Morris & Hickman, 1981). sequence within a column (Morris, 1980; Mor- ris & Hickman, in prep.). It is, nevertheless, an elegant sequence of successively refined morphologies for efficiently gathering food particles. The cocculinid condition The row becomes more difficult to define in small rhipidoglossan limpets of the deep-sea, wood-ingesting genus Cocculina Dall, 1882. Tooth cusps are arranged in orderly arcuate rows, but careful attempts to follow the cusps in any one row back to their bases dem- onstrates that marginal cusps arise from bases in a different row from the lateral cusps. Figs. 5A, 6A and C illustrate the condition in Cocculina. It results from a pronounced elon- gation of marginal tooth shafts. The shafts are laterally flattened, thin, and expanded (Fig. 6C) in contrast to the semicircular or CUSP ROW BASE ROW FIG. 5. Alternative configurations of base rows and cusp rows in the cocculinacean radula. A. Typical pattern in the genus Cocculina Dall, 1882, in which an arcuate base row (dotted line) gives rise to long-shafted marginal teeth that function in a differ- ent cusp row from the massive outer lateral tooth. B. Typical pattern in the genus Pseudococculina Schepman, 1908, showing same distinctive pattern of marginal shaft elongation associated with teeth of different morphology and base row configuration. quadrate cross-sectional shape of the typical rhipidoglossan marginal tooth. The terminal cusp is relatively small and finely serrate. Shafts do not have expanded bases and arise from the radular membrane adjacent to the bases of a very large, robust, and heavily reinforced outer lateral major food-preparing tooth. Shafts extend well beyond (anterior to) the cusp of the large outer lateral, however, terminating in a row of small cusps that lie alongside the outer lateral cusp two rows anterior (Fig. 6A). 148 HICKMAN FIG. 6. Basic features of the cocculinacean radula. A. Row configuration and tooth morphologies in Cocculina sp. Scale bar = 40 рт. В. Row configuration and tooth morphologies in Pseudococculina sp. Scale bar = 20 um. C. Detail of flattened marginal tooth shaft and small hooked cusp in a second species of Cocculina. Scale bar = 40 um. D. Low-angle detail of the oval rachidian and interlock system of left lateral teeth in a second species of Pseudococculina. Scale bar = 20 um. A strikingly similar discrepancy between base row and cusp row appears in radulae in the deep-sea limpet genus Pseudococculina Schepman, 1908 (Figs. 6B and D). This is interesting because teeth in the radula of Pseudococculina are of such a different form as to suggest different familial status. The rachidian teeth are well developed and rel- atively thick, oval, overlapping plates; and the inner laterals are better developed and in- terlock basally when the radula is collapsed (Fig. 6D). The innermost lateral has an un- usually long inner basal leg, and the tooth is attached basally along a broad diagonal area extending from the posterior end of the ra- chidian to a position that is anterior to the anterior end of the rachidian. The outer lateral tooth is also greatly enlarged, but it has a very different pattern of cusps and a depression on its inner face to accommodate the basal leg of the adjacent lateral tooth. The marginal teeth are more nearly circular in cross section and the shafts terminate in large, hooked cusps. However, they are extraordinarily long, as in ARCHAEOGASTROPOD RADULAE 149 Cocculina, and they arise from a lateral base row posterior to the rows into which their cusps hang and interact (Fig. 5B). Elongation of marginal tooth shafts has not been observed to date in other rhipidoglossan archaeogastropods and constitutes a power- ful argument for considering relatively close phylogenetic affinities between the two genera. The Cocculinacea is currently a garbage-basket taxon for a diversity of minute deep-sea limpets of diverse radular morphol- ogy, some lacking marginal teeth. Analysis of row interactions may provide more powerful clues to intrageneric phylogenetic rela- tionships than details of comparative tooth morphology. The fissurelline condition A more complicated form of discrepancy between base rows and cusp rows appears consistently in radulae of keyhole limpets of the genus Fissurella Вгидшеге, 1789 (Figs. 2B and 7). The radula is strikingly asymmetric (Hickman, 1981), to accommodate storage of enlarged, heavily-cusped, outer lateral food- preparing teeth. The rachidian is asymmetric, has a narrow simple cusp, and is flanked by four simply cusped inner lateral teeth. The enlarged, massive, outer lateral is twice the height of the rachidian and inner laterals, so that, while its base is aligned with the bases of one set of lateral teeth, its cusp is aligned with the cusps of the next anterior row. A striking feature of the fissurelline radula is the well- developed, uncusped, lateromarginal plate that lies between the large outer lateral and the marginal teeth as a specialized articula- tory structure (Hickman, 1976, 1977, 1980). The posterior margin of the plate is aligned with the posterior margins of the rachidian and laterals, and thus it is not difficult to assign to a base row. The innermost marginal teeth, however, do not arise adjacent to the lateromarginal plate, but rather from behind or slightly anterior to the plate (see Fig. 2B and the isolated plate and inner marginals in Fig. 7B). From a functional or biomechanical point of view, each lateromarginal plate interacts with two rows of marginal teeth (Hickman, 1980). The shallow pocket on the back of the plate fits over the lower portion of the shafts that arise behind it, so that the plate can assist in pushing those marginals out into feeding po- sition. A similar pocket accommodates the mid-portion of the shafts of the next posterior FIG. 7. Basic features of the fissurelline radula. A. Row configuration and tooth morphologies in Fis- surella nimbosa (Linné, 1758). Scale bar = 400 um. B. Detail of isolated lateromarginal plate from F. volcano Reeve, 1849, and posterior set of marginals with which plate interacts. Scale bar = 100 pm. row of marginals and assists in their collapse and in maintaining their alignment. Thus each row of marginal teeth also interacts with two adjacent lateromarginal plates. The shafts of the marginal teeth are so long (three times the length of the rachidian and inner laterals) that they function in a different cusp row from both the rachidian and inner laterals on one hand and the massive outer lateral on the other (Fig. 2B). The lateromarginal plates also have con- nections along the length of the column to one 150 HICKMAN another, so that forces will be transmitted sequentially along the column during the feeding stroke. This situation is sufficiently complicated that it is difficult, if not impossi- ble, to define a row in a functional sense. It is interesting in this respect that, in the early part of the century, Torr (1914), using light micros- copy, observed that more than one set of marginal teeth seemed to be associated with each of the elongate rhomboidal plates. In his illustrations he shows a single tooth row as consisting of two ranks of marginal teeth attached at their bases along one side of the plate. It is clear from his text descriptions that he believed there had been a duplication of marginals, although this is clearly not the case. The complicated situation described above with respect to interactions between marginal teeth and lateromarginal plates is characteris- FIG. 8. Alternative cusp row configurations in two fissurellid subfamilies. A. Subfamily Fissurellinae: the massive outer lateral tooth with base and cusp situated in different rows. B. Subfamily Diodorinae: the massive outer lateral tooth with its base and cusp situated in the same row. tic not only of the Fissurellinae, but occurs in the other fissurellid subfamilies as well. The lateromarginal plate is not so highly de- veloped in the other subfamilies, however. The taxonomic implications of lateromarginal plate functional morphology are discussed further in a following section. On the other hand, the discrepancy between cusp and base rows created by enlargement of the outer lateral tooth does seem to be restricted to the Fissurellinae. In other fissurellid sub- families the outer lateral tooth, although mas- sive, has not doubled in length; and its base and cusp thus correspond with bases and cusps of the adjacent lateral teeth (Fig. 8). Inferences of an advanced, complicated, and highly derived state of functional in- teractions within the fissurelline radula paral- lel evidence from the geologic record and strengthen phylogenetic inferences. The fis- surelline shell does not appear until the Eocene, while emarginuline and diodorine species appear early in the Mesozoic (Trias- sic and Jurassic) (Knight et al., 1960). The neritacean condition A final example of a unique pattern of tooth- rows is seen in neritacean gastropods. The FIG. 9. The neritacean condition. A. Row configura- tion and basic tooth morphologies based on radula of Nerita plicata Linné, 1758. Note how the flange on the fouth lateral tooth fits between and interacts with two sets of marginal teeth. B. Detail of the simultaneous interaction of the third lateral tooth with the base (anterior) of one fourth lateral and the cusp (posterior) of another (see also Fig. 11). ARCHAEOGASTROPOD RADULAE 151 complications are also related to functional in- teractions of tooth bases and cusps. Some important features of radulae from two nerita- cean families, the Neritidae and Phena- colepadidae, are illustrated in Figs. 9 and 10. There is a broad central region in the typical neritacean radula that is occupied by rows of “teeth” that are modified into complex plates of a variety of shapes. Baker (1924) has described the basic forms of these plates in some detail. The terms “base,” “shaft,” and “cusp” are not readily applicable. They are better described in terms of notches and pro- cesses. Briefly, the small quadrate rachidian is flanked by a broad first lateral and relatively small, but thick and complexly curved, plate- like second and third laterals. There is no problem defining rows in this central region of the radula. A major hallmark of the neritacean radula is the massive, compound, fourth later- FIG. 10. Basic features of the neritacean radula. A. Nerita undata Linné, 1758: row configuration and tooth morphologies. Scale bar = 400 um. В. Phenacolepas osculans (С. В. Adams, 1852), a neritacean limpet: row configuration and tooth morphologies. Scale bar = 40 um. С. Nerita plicata Linné, 1758; detail of interaction of the massive fourth lateral with inner laterals (left) and marginals (right). Note also the alternation of cusp edges of marginals with cusps of fourth laterals. Scale bar = 100 um. D. Phenacolepas sp.: detail of interlock of fourth lateral cusp and third lateral (at arrow). Compare alternation of marginal cusp edges and fourth lateral cusps with Fig. 10C. Scale bar = 20 um. 152 HICKMAN al tooth with its heavy base, broad, spoon- shaped cusp, and lateral flange. The place- ment of the base and cusp make it difficult to decide to which inner plate row it belongs: the base interacts with the anterior end of one of the small third laterals while the large spoon- shaped cusp interacts with the posterior end of the small third lateral in the next anterior plate row. In other words, we are looking at an alternation of cusp-base, base-cusp in- teractions. The ambiguity of the situation can be appreciated by examining the inset in Fig. 9. Baker’s (1924) drawings of neritid radulae imply one interpretation, while Fretter's (1965) drawing implies the alternative; and it probably makes little difference how one chooses to define a row in this instance. From some vantage points and with the radula in certain configurations, one set of interactions may appear more important than another: the interaction of the third lateral plate with the fourth lateral cusp is emphasized in the prep- aration and viewing angle of Fig. 11. The same kind of intimate connection of the third lateral simultaneously to two fourth laterals can be seen in the Phenacolepadidae (Fig. 10B). There is also a unique neritacean pattern of interactions between the massive outer later- al tooth and the marginal tooth-rows. Looking at the unfolded or flattened radula of Nerita #1} EN FIG. 11. Nerita funiculata Menke, 1851: tilted view of a neritacean radula emphasizing interaction of third lateral (3L) with base of one fourth lateral (anterior arrow) and the cusp/shaft (posterior arrow) of another. Scale bar = 100 um. Linné, 1758, or Phenacolepas Pilsbry, 1891 (Figs. 10C and D), there is a distinct alterna- tion or interleaving along the column axis of the broad cutting edges of the outer laterals and the compound edges presented by the cusps of the marginal teeth. This interleaving is interpreted as a functional necessity for efficient tooth storage: an alternative to the pronounced asymmetry that has developed in many other rhipidoglossan groups to help deal with accommodation of enlarged major substrate-excavating teeth (Hickman, 1981). Again, there is a question as to which mar- ginal teeth belong with which outer laterals; and, again, there is no clear-cut answer. From the foregoing discussion of the row concept in the gastropod radula, it can be concluded that there is only one kind of row in a strictly morphological sense and that all gastropod radulae have a fixed set of mor- phological units that is repeated numerous times. However, there are other kinds of rows in a functional sense. Base rows and cusp rows may not coincide, and major patterns by which they are defined and functionally in- teractive are useful as characters in taxon- omy and in drawing phylogenetic inferences. In the most complicated kinds of functional integration it is probably not useful to worry about defining rows, but rather to focus on understanding and characterizing the interac- tions of elements within the radula. In the following sections, | provide some examples of specific interactions that are of interest from a systematic point of view. CUSP INTERACTIONS In the preceding section, cusp rows were distinguished as transverse series of food- preparing and food-gathering units that are situated adjacent to one another regardless of whether their bases are adjacent and interact- ing. Cusps interact with one another in a variety of ways during feeding and during storage of the radula when the animal is not feeding. Many cusp interactions are not reflected in morphology; but, particularly in the major substrate-excavating teeth, cusps may dis- play adaptations from transmitting forces to one another. Although these adaptations may look superficially very similar (convergence), different groups of gastropods have de- Bes different ways of solving similar prob- ems. ARCHAEOGASTROPOD RADULAE 153 For example, in both the genus Haliotis Linné, 1758, and the genus Turbo Linne, 1758, there has been reduction in the prom- inence of the rachidian and lateral teeth and enlargement of the inner marginal tooth cusps, with accompanying development of accom- modational asymmetry to the tooth-rows (although note that the rows slope in opposite directions) (Hickman, 1981). The basic pat- terns are illustrated for comparison in Figs. 12A and B. The outer marginal cusps in both genera overlap and interact without any physical modification of cusps. But in both genera the cusps of the three inner marginal teeth are modified physically to interlock (Figs. 12C and D). In Turbo note at the arrow in Fig. 12D how the secondary cusp is accommodated by a strong depression on the top of the main cusp on the adjacent tooth. In Haliotis note at the arrow in Fig. 12C that the FIG. 12. Asymmetric radular patterns and convergent cusp interaction in the Haliotidae and Turbinidae. A. Haliotis rufescens Swainson, 1822. Scale bar = 400 um. В. Turbo fluctuosus Wood, 1828. Scale bar = 500 um. Note that both radulae are asymmetric but that the direction of slope is reversed. C. Cusp interactions in H. rufescens: note at arrow how the edge of one cusp fits into notch at the back of the adjacent cusp. Scale bar = 100 рт. D. Cusp interactions in 7. fluctuosus: note at arrow how secondary cusp fits into depression on top of the adjacent cusp. Scale bar = 100 pm. 154 HICKMAN main cusps overlap in the same manner that they do in Turbo but that the back of the cusp fits into a notch at the rear of the cusp on the adjacent tooth and that the adjacent tooth cusp bears a process that fits into a pocket beneath the neighboring cusp. Although the fine, serrate, outer-marginal teeth that are so characteristic of the rhipido- glossan radula generally do not develop pro- nounced physical accommodations to one another, there are some interesting instances of marginal tooth interactions. For example, Fig. 13 shows how a long process has de- veloped near the base of the serrate margin of the cusp in an undescribed rhipidoglossan deep-sea limpet. It wraps around and under the adjacent cusp, limiting the degrees of freedom of movement. A series of marginal teeth can thereby become linked and aligned as a unit. Standard methods for preparing radulae generally destroy evidence of the ways that teeth are aligned during feeding. This is par- ticularly true of the delicate marginal teeth, which must be folded into an unnatural con- figuration when an excised radula is mounted flat for either optical or scanning electron microscopic viewing. Morris (1980) has de- veloped a method for producing artificial pro- trusions of gastropod radulae that preserve details of cusp relationships and interaction as they exist during feeding (Morris & Hick- man, 1981). Marginal tooth cusps contact the FIG. 13. Marginal tooth cusp interactions in an undescribed deep-sea limpet of uncertain affinities from the Galapagos Rift. Note at arrows how en- larged processes wrap around cusps of adjacent teeth. Scale bar = 10 um. substrate as a unit. Fig. 14 shows how in Tegula funebralis, marginal tooth cusps in the tight semicircular configuration of the radula during the feeding stroke overlap one another to form a long, continuous, serrate, food- collecting edge along the length of the row. SHAFT INTERACTIONS Interactions of tooth shafts also provide an important means of integrating function and transmitting forces. There is virtually no systematic documentation in the literature of patterns of variation in shaft length, cross- sectional shape, and flexural stiffness in radu- lar teeth. Even less is understood about the biomechanical or functional significance of interactions of length, shape, and stiffness. Rachidian and lateral tooth shafts are gener- ally shorter, reinforced into more complex cross-sectional shapes, and less flexible than the shafts of marginal teeth. It makes intuitive sense that the substrate-preparing teeth should have a different set of properties from the brushing teeth that gather up food parti- cles. One of the most interesting, recurring adaptations in rachidian and lateral teeth is development of thin, laterally-extended flanges of chitin off the backs of tooth shafts. These are frequently joined with tooth cusps and are interleaved in complex ways with similar shaft extensions of adjacent teeth. FIG. 14. Tegula funebralis: artificial protrusion of radula in operational configuration showing how cusps of marginal teeth overlap to form a continuous food-collecting edge. Scale bar = 200 рт. ARCHAEOGASTROPOD RADULAE 155 These shaft modifications are important from a systematic point of view because different distinctive patterns have evolved at higher taxonomic levels and characterize major groups. For example, a complex of closely-related genera of deep-sea deposit-feeding trochid gastropods (Bathybembix Crosse, 1893; Cidarina Dall, 1909; Calliotropis Seguenza, 1903) has developed a particularly elaborate set of expanded or “hooded” rachidian and lateral teeth that must greatly alter the magni- tudes and distributions of stress that would occur in free-standing teeth with simple tooth FIG. 15. Shaft interactions in two trochid gastro- pods. A. Calliotropis regalis (Verrill & Smith, 1880): rachidian and lateral teeth with broad interleaving shaft expansion or “hoods.” Scale bar = 400 um. В. Turcica caffea (Gabb, 1865): shaft expansion and interleaving of a similar type. Scale bar = 40 рт. shafts. Fig. 15A illustrates the pattern of in- terleaving of teeth that results. A somewhat simpler form of shaft expansion and interac- tion occurs in closely-related trochaceans of the genus Turcica A. Adams, 1854 (Fig. 15B). Shaft expansion occurs prominently and consistently at one critical point in the radula of neritacean gastropods: between the mas- sive fourth lateral and the marginal tooth com- “lex. There is a well-developed thin extension of the shaft that is joined to the cusp and passes laterally behind the shafts and cusps of the inner marginal teeth of one row, sepa- rating them from the shafts of the next anteri- or marginal tooth-row (Figs. 9, 10C). Ex- pansions of this sort frequently stain well with protein-specific stains, suggesting tanning of the chitin into a more rigid structure that would function effectively in transmitting stress without excessive deformation. Interactions between marginal tooth shafts do exist, but they are of a different nature from rachidian and lateral tooth interactions. They involve primarily fusion of adjacent shafts or, alternatively, incomplete separation of tooth shafts during radular ontogeny. Particularly in radulae in which the marginal teeth are numerous and of very thin, light construction, the lowermost portions of the shafts will be fused. The fusion in such cases is generally restricted to a region below a pronounced twist in the shafts and where the shafts come to lie parallel to the radular membrane (Fig. 16). FIG. 16. Shaft fusion in Trochus intextus. Note at arrows that outer marginal tooth shafts are united. Scale bar = 100 um. 156 HICKMAN Incomplete separation of marginal teeth is relatively common in rhipidoglossate radulae, and it is not clear to what extent the condition may be adaptive and to what extent it may represent a relaxation of selection for separa- tion, tanning, and hardening of each tooth individually (Hickman, 1980). Incomplete sep- aration of teeth seems to be more frequent in deep-sea gastropods. Fig. 17A illustrates fu- sion of both marginal tooth shafts and cusps in the Galapagos rift limpet, Neomphalus McLean, 1981. Fig. 17B, on the other hand, illustrates what is better termed bifurcation or multifurcation, in which a single shaft clearly gives rise, in more or less orderly pattern, to two or more cusps. The animal in this in- stance is an undescribed trochacean from hydrothermal vents at 21°N on the East Pacific Rise. BASAL INTERACTIONS Patterns of basal interaction between radu- lar teeth can be particularly useful in mollus- can systematics. Basal expansion has led to many unique methods of overlapping and in- terlocking, both within rows and along col- umns, that are conserved at higher taxonomic levels. | pointed out previously (Hickman, 1980) that basal interactions are important biomechanically because they perform multi- ple functions. First, they act in dissipating stress applied to the radular membrane by any one tooth. Second, they provide an effec- tive means of dealing with bending and over- turning movements on individual teeth. Third, they can assist in aligning and orienting teeth during feeding and in activating one another sequentially during the feeding stroke. And finally, they can assist in folding teeth together efficiently and economically to occupy the least possible space while the radula is not in use. The interactive potential of tooth bases was not well understood until the scanning electron microscope became available, and SEM studies of a diversity of molluscan radulae all suggest that basal in- teractions are an important functional feature (Solem, 1972, 1973, 1974; Solem & Richard- son, 1975; Solem & Roper, 1975; Bertsch et al., 1973; Bertsch & Ferreira, 1974; Ferreira & Bertsch, 1975). Basal overlap and interlock were used in the foregoing discussion to define base rows, and several illustrations were provided in the и FIG. 17. Joining of marginal tooth shafts. A. Меот- phalus fretterae McLean, 1981: marginal tooth shafts and cusps showing what is interpreted as irregular and partial separation. Scale bar = 20 um. B. Undescribed deep-sea trochacean microgastro- pod with branching marginal tooth shafts (at arrows). Scale bar = 10 pm. context of those earlier remarks (Figs. ЗА-С). Many rhipidoglossan families, subfamilies, and genera have developed distinctive forms of basal interaction between rachidian and lateral teeth. They are best developed in taxa that feed on hard substrates in the rocky intertidal and shallow subtidal zones. Deposit- feeding taxa may have interlocking teeth (e.g. Fig. 15A), although the processes and corre- sponding sockets are of more delicate con- struction. ARCHAEOGASTROPOD RADULAE 157 A different form of basal interrelationship is associated with filter feeding. Various lines of evidence have led Fretter (1975) and, in- dependently, Hickman, McLean & Ponder (unpublished data) to recognize or suspect filter-feeding in a variety of archaeogastro- pods. In filter-feeding taxa, tooth shafts and cusps tend to be lost; and bases, likewise, may be reduced to thin chitinous vestiges, as in Umbonium Link, 1807, and Bankivia Krauss, 1848 (Figs. 18A and B). The situation is less clear in the genus Lirularia Dall, 1909. Two of the eastern Pacific species, L. suc- cincta (Carpenter, 1864) and L. lirulata (Carpenter, 1864), have what | have identified as a filter-feeding radula; however, McLean (personal communication) notes that the gill, although unusual, is not a typical filter-feeding gill. In the western Pacific species L. iridescens (Schrenck, 1863), bases form a distinctive pavement of interacting teeth sep- arating the marginal tooth complexes (Figs. 18C & D). FIG. 18. Degenerate rachidian and lateral teeth of recognized or suspect filter-feeding trochaceans. A. Umbonium (Ethalia) guamense (Quoy & Gaimard, 1834): central portion of radula reduced to thin, vestigial tooth bases. Scale bar = 100 um. В. Bankivia (Leiopyrga) lineolaris (Gould, 1861): similar reduction of teeth to vestigial bases. Scale Баг = 40 ит. С. Lirularia iridescens (Schrenck, 1863): rachidian and laterals lacking cusps and shafts, but not so reduced as in Umbonium and Bankivia. Scale bar = 100 рт. D. L. iridescens: low angle view of overlapping pavement of basal plates. Scale bar = 20 um. 158 HICKMAN The function of the radula in filter-feeders is not clearly understood, and not all taxa that are able to do so are obligate filter feeders. Although Neomphalus can filter-feed, it re- tains well-developed rachidian and lateral teeth and the capacity to graze (McLean, 1981). In any event, the characteristics of reduced radular teeth should prove helpful in distinguishing the phylogenetic origins of filter feeding in different archaeogastropod taxa. Evolutionary degeneration of rachidian and lateral teeth, accompanied by reduction in the degree of interaction between tooth bases, has occurred in non-filter-feeding taxa as well and has already been illustrated for the Coc- culinacea (Figs. 5 and 6) and Fissurellacea (Figs. 2 and 7). Here it is related to the development of massive outer lateral or inner marginal substrate-preparing teeth that are used on a variety of organic substrates (wood, encrusting sponges, compound as- cidians, bryozans, etc.). Highly developed basal interactions in these taxa are associ- ated with the massive food preparing tooth; and, in the Fissurellacea, with the lateromar- ginal plate. The more highly developed and refined a basal interaction becomes, the more useful it should be in taxonomy if the refinements are taxon-specific. Lateromarginal plates may be considered a special case of extreme mod- ification of tooth bases; and, because mod- ification has proceeded along very different lines in different rhipidoglossan groups, they provide a whole series of potentially useful characters. Hickman (1976, 1977, 1980) has emphasized the functional importance of lateromarginal plates as articulatory struc- tures between substrate-preparing and food- gathering portions of rhipidoglossan radulae (i.e. between lateral and marginal teeth), and the highly developed condition of these plates in the fissurelline radula was discussed earlier in this paper. | am currently developing a set of charac- ters and character states for lateromarginal plates that includes: presence/absence of the plate; condition of the plate with respect to shaft and cusp (short shaft and well- developed cusp present; long shaft and well- developed cusp present; long shaft and rudimentary cusp present; shaft and cusp ab- sent) lateromarginal plates linked/not linked to one another; lateromarginal plates mod- ified bases of lateral teeth/modified bases of marginal teeth; plates with/without interaction with outer lateral tooth; plates with/without interaction with inner marginal teeth. Examples of some character states with respect to shaft and cusp are illustrated in Figs. 19A—D. In the genus Diloma Philippi, 1845, there is a very short shaft and well- developed cusp that resembles the cusps of other marginal teeth (Fig. 19A). This state is also characteristic of genera and species within the subfamily Trochinae. In Solariella Wood, 1842 (Fig. 19B), and Calliotropis Seguenza, 1903 (Fig. 19C), the lateromar- ginal plate has a long but rudimentary shaft and cusp. In Fissurella Bruguiere, 1789, the plate is well developed and completely lacks shaft and cusp, although well-developed shafts and cusps are present in other fis- surellid subfamilies. SUMMARY AND CONCLUSIONS Although functional characters must be used with great care in classification and phylogenetic inference, they can be used to generate new sets of taxonomic characters and can provide important insight into rela- tionships. Functional characters are particu- larly useful in archaeogastropod systematics when applied to rhipidoglossan radulae. Stud- ies of tooth-row functional integration lead not only to new systematic insights, but they are also helpful in understanding how radulae work in preparing and gathering a diversity of kinds of food from a diversity of substrates. Some of the more important conclusions are: 1. Although interactions between radular teeth occur primarily within rows and within columns, tooth interactions with the substrate may follow a more complex sequence that is related neither to rows nor columns. 2. Although cusp rows and base rows coin- cide in many rhipidoglossan gastropods, and although there is only one kind of row in a strictly morphological sense, there are many kinds of rows in a functional sense. Base rows and cusp rows frequently do not coincide, and the major patterns by which they are defined and functionally interactive provide useful characters in systematics. 3. In the most complicated kinds of func- tional integration it is not practically possible to define tooth-rows in a functional sense. 4. Complex patterns of tooth interactions are consistently associated at higher taxon- omic levels. 5. Cusp interactions include adaptations for transmitting forces from one tooth to an- other and frequently involve distinctive mod- ifications of morphology. Even where there ARCHAEOGASTROPOD RADULAE 159 are no strong morphological modifications, many cusps may be presented to the sub- strate as a functional unit. 6. Shaft interactions also integrate function and transmit forces according to characteris- tic patterns in different taxa. Shaft expansion and interleaving has been superimposed on a number of different radular groundplans, and fusion of shafts and incomplete separation of shafts can alter both the appearance and biomechanical properties of teeth. 7. Patterns of basal interaction are func- tionally important because they perform many functions simultaneously. Patterns of overlap and interlock are different in different taxa. Lateromarginal plates are specialized tooth bases that are important as articulatory struc- tures. They have developed between the substrate-preparing and food-gathering por- tions of many rhipidoglossan radulae. There are both morphological and functional characters associated with these plates that are particularly useful in systematic assess- ments. FIG. 19. Examples of lateromarginal plates (LMP). A. Diloma nigerrima (Gmelin, 1791): broad plate with well-developed short shaft and cusp. Scale bar = 100 um. В. Solariella nuda Dall, 1896: thin plate with elongate rudimentary shaft and cusp. Scale bar = 100 рт. C. Calliotropis hataii Rehder & Ladd, 1973: well-developed interactive plate with long rudimentary shaft and cusp. Scale bar = 50 am. D. Fissurella volcano Reeve, 1849: massive interactive plate lacking shaft and cusp. Scale bar = 100 um. 160 HICKMAN ACKNOWLEDGMENTS | thank D. R. Lindberg and J. H. McLean for critical review of the manuscript. | am also grateful to J. H. McLean for providing many of the specimens on which radula illustrations are based. Scanning electron micrographs were taken in the Department of Anatomy, University of California, San Francisco; and the assistance and advice of J. A. Long and J. R. Morgan are gratefully acknowledged. Fig. 4 is the work of T. E. Morris. Drawings are by M. E. Taylor, and the manuscript was typed by D. Mickela. This research was supported by National Science Foundation Grants DEB 77-14519 and DEB 80-20992. LITERATURE CITED BAKER, H. B., 1924 [“1923”], Notes on the radula of the Neritidae. Proceedings of the Academy of Natural Sciences of Philadelphia, 75: 118-178. BERTSCH, H. & FERREIRA, A. J., 1974, Four new species of nudibranchs from tropical west Amer- ica. Veliger, 16: 343-353. BERTSCH, H., FERREIRA, A. J., FARMER, W. M. 8 HAYES, Т. L., 1973, The genera Chromodoris and Felimida (Nudibranchiata: Chromodoridi- dae) in tropical west America: distributional data, description of a new species, and scanning elec- tron microscopic studies of radulae. Veliger, 15: 287-294. FERREIRA, A. J. & BERTSCH, H., 1975, An- atomical and distributional observations of some opisthobranchs from the Panamic faunal prov- ince. Veliger, 17: 323-330. FRETTER, V., 1965, Functional studies of the an- atomy of some neritid prosobranchs. Journal of Zoology, London, 147: 46-74. FRETTER, V., 1975, Umbonium vestiarium, a filter- feeding trochid. Journal of Zoology, London, 177: 541-552. FRETTER, V. & GRAHAM, A., 1962, British pro- sobranch mollusks, their functional anatomy and ecology. Ray Society, London, xvi + 755 p. HICKMAN, C. S., 1976, Form, function, and evolu- tion in the archaeogastropod radula. Geological Society of America, Abstracts with Programs, 8: 917-918. HICKMAN, C. S., 1977, Integration of electron scan and light imagery in study of molluscan radulae. Veliger, 20: 1-8. HICKMAN, C. S., 1980, Gastropod radulae and the assessment of form in evolutionary paleontol- ogy. Paleobiology, 6: 276-294. HICKMAN, C. S., 1981, Evolution and function of asymmetry in the archaeogastropod radula. Veli- ger, 23: 189-194. HYMAN, L. H., 1967, The invertebrates: volume VI, Mollusca |; Aplacophora, Polyplacophora, Mono- placophora, Gastropoda. ... McGraw-Hill, New York, etc., vii + 792 p. KNIGHT, J. B., COX, L. R., KEEN, A. M., BATTEN, В. L., YOCHELSON, Е. |. 8 ROBERTSON, R., 1960, Systematic descriptions (Archaeogastro- poda), In MOORE, В. C., ed., Treatise on In- vertebrate Paleontology, Part |, Mollusca 1: 169- 310, Geological Society of America and Univer- sity of Kansas Press. MCLEAN, J. H., 1981, The Galapagos Rift limpet Neomphalus: relevance to understanding the evolution of a major Paleozoic-Mesozoic radia- tion. Malacologia, 21: 291-336. MORRIS, T. E., 1980, Morphological and functional dynamics of the rhipidoglossan radula of Tegula funebralis (Mollusca: Gastropoda). Masters thesis, University of California, Berkeley, 145 p. MORRIS, Т.Е. & HICKMAN, С. S., 1981, A method for artificially protruding gastropod radulae and a new model of radula function. Veliger, 24: 85-90, 1 pl. PURCHON, R. E., 1968, The biology of the Mol- lusca. Pergamon Press, London, 560 p. SCHOOLEY, C., HICKMAN, C. S. & LANE, W. C., 1982, Computer graphic analysis of stereo micrographs as a taxonomic tool. Veliger, 24: 205—207, 1 pl. SOLEM, A., 1972, Malacological applications of scanning electron microscopy Il. Radula struc- ture and functioning. Veliger, 14: 327-336. SOLEM, A., 1973, Convergence in pulmonate radulae. Veliger, 15: 165-171. SOLEM, A., 1974, Patterns of radular tooth struc- ture in carnivorous land snails. Veliger, 17: 81— 88. SOLEM, A. & RICHARDSON, E. S., Jr., 1975, Paleocadmus, a nautiloid cephalopod radula from the Pennsylvanian Francis Creek Shale of Illinois. Veliger, 17: 233-242. SOLEM, A. 8 ROPER, C. F. E., 1975, Structures of Recent cephalopod radulae. Veliger, 18: 127- 133. TORR, C. M., 1914, Radulae of some South Aus- tralian Gastropoda. Transactions of the Royal Society of South Australia, 38: 362-368. YONGE, C. M. & THOMPSON, T. E., 1976, Living marine molluscs, Collins, London, 288 p. MALACOLOGIA, 1984, 25(1): 161-172 HETEROSEMATISM IN SNAILS A. J. Cain Department of Zoology, Liverpool University, Liverpool L69 3BX, England ABSTRACT The colour pattern of the shell of many land snails is somewhat or markedly different on the upper side and on the under side of equi-dimensional to depressed shells. In both these and tall-shelled species the mouth and throat of the shell, and in tall-shelled ones the umbilical region, are often distinctively coloured or patterned. These and the underside of globular to discoidal shells are normally hidden when the snail is attached to a substrate. It is suggested that the different coloration shown if the shell is dislodged by a predator and falls with the concealed area upwards is used to mislead the predator. Examples are given of such a biparatite patterning. Investigation of the attitude assumed when snails of different shapes are dropped suggests that many have a 1:1 chance of showing the upper or the under side. The possible significance of this ratio for the hypothesis is examined, in relation to a predator bearing two hunting images in its mind at once. A classification of such heterosematic behaviour is suggested. INTRODUCTION In many species of land snails with globu- lar, trochoid or depressed shells, the colour pattern is markedly different above and below the periphery of the shell. No such difference is usually seen in snails with tall shells, but in these the umbilical region is often distinctively patterned. In shells of all shapes the lip (if any) and sometimes the throat is strikingly coloured, and often the mantle-surface that fills the mouth of the shell when the animal is resting is strongly pigmented. All land snails attach themselves to a substrate by means of mucus secreted at the mouth of the shell, and the equidimensional shells, and often the de- pressed ones too, sit on a hard substrate in such a way that much of the shell below the periphery is concealed, as is the oral side of the lip, the throat of the shell, and the animal itself. It is somewhat surprising, therefore, that the subperipheral or circumbilical pattern is often bold and striking, and sometimes at least polymorphic when the supraperipheral is not. In recent years, much work has been done on the visual searching for prey by predators. The predator (usually a bird), after the suc- cessful capture of a prey item, forms a hunt- ing image of that item, and will even pass over prey of different aspect to get at more of that type for which it has formed the image (see e.g. Allen, 1972; Manly, Miller & Cook, 1972 for references). Many animals have therefore taken to unpredictable behaviour, such as erratically zigzagging flight, to confuse, or even, by the arousal of conflict in it, to evoke escape reactions from, a would-be predator (references and discussion in Driver & Humphries, 1970; Humphries & Driver, 1970). Humphries and Driver (1967) adopt the term protean behaviour from Chance & Russell (1959), defining it as “behaviour that is suf- ficiently unsystematic to prevent a reactor from predicting in detail the position or actions (or both) of the actor.” It seems not impossible that some at least of these striking patterns normally concealed on snail-shells may function in a similar way to mislead a predator. This paper describes some of the patterns and considers the be- haviour of a shell knocked off its substrate by a predator. The frequencies of different atti- tudes of the fallen shell are examined for different species, and discussed in the light of this suggestion. DIFFERENTIAL COLOUR AND BANDING PATTERNS Among tropical land snails with shells of equidimensional to depressed shape, con- siderable contrast between the colour pat- terns above and below the periphery is not uncommon. Thus E. von Martens (1867) gives beautiful coloured figures in his account of an expedition to the Orient of such species (161) 162 CAIN as his Nanina citrina L. var. aurantia, with a broad purple brown band on a buff back- ground above, and a warm reddish brown below with a narrow white edge, and var. columellaris with the upper surface warm deep brown, the lower very pale greenish yellow. In Nanina sulfurata v. M. the upper side is boldly banded with black on pale yel- low, the lower is plain pale yellow. In Helix bulbulus Mousson, again the upper side is spirally banded dark brown on pale whitish buff, the lower is plain pale dirty buff except for a dark brown umbilical band and the flared lip is brown and white. In Helix pyrostoma Fer. (now a helicostyline) the whole shell is de- pressed and yellow-brown but the underside is diversified by the brilliant red lip and throat, the aboral side of the lip, which is visible from above being merely brown. In Helix ex- ceptiuncula Fér., the upper side is banded black and white or black, white and buff-grey, while the lower side is plain buff-grey. Many other examples could be given from other tropical faunas. Many shells, of course, are plain brown all over, pale brown, yellow brown or even nearly glassy, as von Martens shows in his plate 12, and not all the brightly coloured shells show in all their varieties a diversity between upper and underside. Exactly the same is true of the European fauna, in which the plain-coloured shells are mainly leaf-litter dwellers or noctur- nal; the patterned ones sit out occasionally or habitually in the daytime (Cain, 1977a). Those exposed to intense sunlight are white all over (e.g. Sphincterochila) or at least white or mostly white above, but the latter often retain bands on the underside of the shell (see Cain, Cameron & Parkin, 1969 for the variation in Britain of Helicella itala (L.)) and these may be polymorphic. Similarly, in the common Mediterranean helicid Eobania ver- miculata (Muller), which often sits out on rocks or vegetation, the upper side is usually banded, the bands being often partially broken up into mottles, so that a somewhat uniform marbled effect is obtained. The un- derside, however, is plain white with distinct narrow black bands (Fig. 1), giving a very different effect, which is emphasized by the visible mantle being plain grey, dark grey, or nearly black. This subperipheral pattern var- ies, with the bands either entire, as just de- scribed, or else broken into brown mottles, or represented only by faint lines, and the varia- tion is genetically controlled (Cain, in prep.). Similarly in /berus marmoratus (Fér.), endem- ic to Gibraltar (Fig. 2), the same patterns and variation are seen, the upper side contrasting even more by its mottling with the emphatic FIG. 1. Shells of Eobania vermiculata, 3.3 and 2.9 cm in maximum diameter, showing at left a common morph pattern on the under side of the shell and at right a very common pattern on the upper side. SNAIL HETEROSEMATISM 163 FIG. 2. Shells of Iberus marmoratus, 2.2 and 2.1 cm in maximum diameter, showing contrasting patterns on the under side (left) and upper side (right). FIG. 3. Under and upper sides of the same shell of Discula polymorpha, 1.0 cm maximum diameter. bands of the underside. In the geomitrine Discula polymorpha (Lowe), from Madeira (Fig. 3), the upper side is dark brown and almost uniform, the underside strikingly light with a strong dark band; | have not seen variation in the pattern of the underside in this species. In the bewilderingly variable snail Theba pisana (Miller), the variation in pattern has usually been taken as continuous. Banding may occur all over the shell but is not in- frequently more emphatic on either the upper or lower surface; Fig. 4 shows a collection to show range in variation made by Dr. G. Lewis. Some banding forms are genetically con- trolled (Cain, in prep.; R. H. Cowie, in prep.), and it is usual for them to differ above and below the periphery. Fig. 5 shows the un- banded morph; a form, very well known among helicellines as well as in Theba pi- 164 —- FIG. 4. A range of banding patterns п Theba pisana, giving only а small selection of variation in the species. sana, with a white shell with a strong black peripheral band above and broken bands be- low; a form with yellowish blotches and black dashes above, darkening on the body whorl, but with a thin black band below, or none; and a more conventionally five-banded form in which the dottings and dashes of the upper bands on the spire contrast with the broad nearly plain umbilical area which is empha- sized by strong bordering bands. In Cepaea nemoralis (L.) and C. hortensis (Muller), an unbanded morph and one with a single peripheral band (00300) are well- known, though the latter is very rare in C. hortensis. In both these morphs, the shell appears plain (yellow, pink or brown) both SNAIL HETEROSEMATISM 165 FIG. 5. Morphs of Theba pisana. above and below, but from below, the white or black-brown lip and callus and the animal’s mantle are conspicuous features. All the pos- sible combinations of presence or absence of the five bands, and fusion of adjacent bands appear to have been found, at least in C. nemoralis, but there is an extreme rarity of forms with bands above but not below the periphery, while those with bands below but not above are quite common (Taylor, 1910). 166 CAIN Thus such forms as 00345, 00(34)5, 00(345), 00:45, 00:(45), and 00045 (the parentheses indicate fusion of the contained bands) are far from uncommon, whereas 12300 is extremely rare. It is known that visual predation can exert selection upon the morphs of Cepaea such that heavy banding is at an advantage in very stripy habitats, lack of banding (00000) and effective lack of banding in the normal position of a resting snail (00300, 00345 etc.) in more uniform ones. It is also highly likely that in regions of strong insolation, the upper part of the shell at least must be light-coloured (yellow and effectively unbanded) to reflect back as much incident radiation as possible (see Jones, Leith & Rawlings (1977) for a review). Insolation will affect the underside of the shell less; it therefore can be banded or not. Under visual selection, it may be necessary to be banded all over, and indeed the fully- banded morph, 12345, is common in both species. But it is noticeable even in this morph that the banding is very different on the upper and under sides (Fig. 6), and that each band has its characteristic position and breadth. It was the highly individual charac- ters of the bands of C. nemoralis that led G. von Martens (1832) to the realization that the variation was not merely random, and to the present system of their nomenclature. Bands 1 and 2 are thin and narrow, 3 wider, 4 the widest, and 5 somewhat variable but broad like 3. As seen from above, therefore, the five-banded form shows 3 narrow dark bands (band 3 being foreshortened) on a paler back- ground. From below (apart from the lip, callus and mantle) one sees a large area of pale shell colour emphatically bordered by two large black bands very close to each other (and often fused together). The pattern above is rather light and stripy; below, it is of larger FIG. 6. Variation in banding within a single random sample of a population of Cepaea nemoralis; shells within each principal morph in the sample selected at random to show the upper or the under side. SNAIL HETEROSEMATISM 167 plainer areas, much as in the five-banded morph of Theba pisana (Fig. 5). The visual effect, therefore, although banded in both aspects, is markedly different. It seems worthwhile asking, therefore, whether even when under the constraint of visual selection for general crypsis there is a difference of pattern above and below such that if a snail were to be dislodged by pred- ators and fall base upwards, there might be a chance of it appearing different enough to be overlooked. Certainly, the remarkable dif- ferentiation of the bands needs explaining. When in many species of terrestrial snails (e.g. in the helicostylines and oreohelicids) very numerous bands, not groupable into the standard five of the helicids, can be seen, it can hardly be suggested that there is any inherent necessity for the bands of Cepaea to be so individually characterized. Moreover, the genetic control of the banding is not by individual bands but by whole patterns at a time; the principal banding morphs are not 10000, 02000, etc., but 12345, 00345, 00300 and 00000. Pettitt (1973) in proposing a mode of scoring the morphs of species of Littorina modelled on that for Cepaea has overlooked this important consideration. Similarly, in Theba pisana with such a remarkable variety of bands and bandlets, one needs to find an explanation of why certain banding com- binations giving different appearances above and below the periphery are clipped together genetically. The measurement of actual selection coefficients in the field is often extremely dif- ficult for a number of reasons (Cain, 1977b), and failure to show differential predation may have no significance. One can ask, however, what consequences are likely to follow from the hypothesis outlined above. If snails are seen by a predator crawling or sitting on a substrate, and are recognized, seized and eaten, then the only selection exerted by the predator is likely to be for some sort of cryp- sis, unless the prey is so distasteful that con- spicuousness is advantageous. This has not been suggested, to my knowledge for snails (though it may be true of some slugs). Nor are snails known as mimics of distasteful animals. (Certainly, bird predators generalize their ex- periences sufficiently far for that.) If an attached snail is dislodged and drops to the ground, it has a possibility of misleading the predator if it shows a different pattern on its under side, which was till then concealed. Only discoidal or depressed shells have an obvious upper and under-side; globular ones have no plane of even approximate geomet- rical symmetry, and the distribution of the organs of the body in the shell, with possible differences locally in specific gravity, do not seem to have been investigated. How do snails land when they fall and roll or bounce? Furthermore, it is evident that tall shells, with height much greater than maximum di- ameter, can only lie on their side when they fall, and no distinction between patterns above and below the periphery can be made. Very few shells in this position show any dorso- ventral flattening; the few that do are plain shells, without any pattern or with un- differentiated mottling, probably lurkers in leaf-litter. The vast majority have no flattening and, unlike many of the equidimensional to discoidal shells, they have no differentiation of the pattern which is the same on whatever side they lie, with one important exception. This is the mouth and throat, and not in- frequently the umbilical region. It is a ques- tion, therefore, whether the asymmetry of the mouth does compel the shell to lie in a mouth- up or mouth-down position. POSITIONS OF REST Tests Table 1 shows the species of live snail available for investigating this question. A single individual was used of each, since their shells differed markedly in shape, size and weight, except that three individuals of Helix aspersa Müller differing in spire index were used, and an adult and juvenile of Theba pisana. Each snail was left dry for a day or more so that it would be retracted into the shell but filling it more or less to the mouth. One H. aspersa was tested after several weeks of drought and was so far back in its shell that about a quarter of the body-whorl was empty, but the results obtained with it hardly differed from those after it was fed and watered and brought into the same position as the others. For equidimensional and depressed shells, each individual was held at 20 cm above a flat cloth spread on a hard bench-top, adjusted so that its spire was uppermost, in both cases with the columella vertical, or it was placed in the broad side down position with the col- umelia horizontal and the outer edge of the mouth uppermost. It was then dropped by 168 CAIN TABLE 1. Systematic positions, names, and physical characters of the individuals used in the experiments. rdelicidae Helicinae Helix aspersa (Müller) 1 Helix aspersa 2 Helix aspersa 3 Eobania vermiculata (Müller) Pseudotachea splendida (Draparnaud) Alabastrina alabastrites (Michaud) Levantina caesareana (Parreyss) Theba pisana (Müller) adult Theba pisana juvenile Helicigoninae Dinarica pouzolzi (Ferussac) Leptaxinae Leptaxis undata (Lowe) Geomitrinae Discula polymorpha (Lowe) Helminthoglyptidae Monadenia troglodytes Hanna & Smith Zonitidae Oxychilus cellarius (Muller) Subulinidae Rumina decollata (L.) Clausiliidae Papillifera bidens (L.) Bulimulidae Placostylus porphyrostomus (Pfeiffer) Placostylus fibratus (Martyn) parting the fingers with a jerk; usually, on hitting the cloth it had enough energy to bounce and/or roll. Its final position was re- corded as with apex up or down. For tall shells, only the apex up and apex down posi- tions were used at release; positions of rest were scored as with mouth up (u), mouth down (d), on its side with the umbilical area above the mouth (s), on its side with the outer lip uppermost (s’), and intermediate positions (e.g. s'u, su) as required. The whole pro- cedure for both types of shell was repeated from 10 cm above the cloth. Small and light shells (e.g. Oxychilus, juvenile Theba, Dis- cula) withstood the tests well, as did large and heavy but thick ones (Placostylus). As 20 observations were taken at every position, each shell had a possible total of 360 impacts. It is perhaps not surprising that the large but rather thin-shelled Helix aspersa could not withstand this battering and chipped or cracked before the full set was completed. (An Otala punctata cracked at the first drop from 20 cm.) The Monadenia was given only a partial test. Weight Height Diameter (gm) mm mm h/d 3.80 23.3 27.1 0.86 4.48 26.3 27.3 0.96 8.99 34.4 30.6 1:12 5.73 20.1 27.6 0.73 1.40 10.8 19:3 0.55 2.61 13.4 21.4 0.63 6.60 23.2 30.5 0.76 2.52 15.3 19.4 0.79 0.13 Sal 8.1 0.63 16.73 24.6 40.0 0.62 4.07 15.5 252 0.62 0.25 5:5 10.1 0.54 2.01 14:5 22.6 0.51 0.14 3.0 7.8 0.38 2.27 26.4 12.2 2.16 0.09 13.8 3.8 3.63 26.47 69.4 32.9 2.11 11.28 54.6 25.5 2.14 Results Table 2 gives the actual results for near- equidimensional to depressed shells, Table 3 for tall ones. The results of x? tests for heterogeneity of data within or between runs for each species, and for departure from 1:1 ratio on the data in Table 2 are given in Table 4. For all but three of the individual snails there is no heterogeneity in the results, and for eight no significant departure from a 1:1 ratio. For the Helix aspersa 2 and 3, the shell rested more often with mouth up and apex down than would be expected if the ratio were 1:1, and the same was true of the Oxychilus, but for the Leptaxis undata and for Theba pisana, especially the juvenile, the departure was in the opposite direction, the shell resting more often with the apex up. The peculiar shape of the juvenile Theba pisana, with al- most flat upper surface and strongly rounded underside suggests that the other way up would be most stable, but with the apex up, it rolls around like a top without reversing its position. Of those showing heterogeneity in SNAIL HETEROSEMATISM 169 TABLE 2. Resting positions of individuals dropped from different heights and initial attitudes. In each pair of figures, the first is the number of times the snail came to rest with apex up (and mouth down), the second with apex down (and mouth up). Height and initial attitude 20 cm 10cm Broad Broad Apex Apex side Apex Apex side Species up down down up down down Total Helix aspersa 1. 32/28 31/29 11/9 32/27 30/30 —/- 137/123 Helix aspersa 2. 15/25 18/22 — 3/17 6/14 —/- 42/78 Helix aspersa 3. 16/44 —/- —/- 35/25 24/36 22/38 97/143 Eobania vermiculata 28/32 21/39 33/27 25/35 28/32 27/33 162/198 Levantina caesareana 27/33 33/27 22/38 31/29 36/24 31/29 180/180 Alabastrina alabastrites 38/22 33/27 27/33 31/29 33/27 25/35 187/173 Pseudotachea splendida 32/28 33/27 26/34 23/37 33/27 30/30 177/183 Theba pisana adult 34/26 30/30 34/26 33/27 34/26 34/26 199/161 Theba pisana juvenile 40/20 39/21 34/26 34/26 39/21 37/23 233/137 Dinarica pouzolzi 27/33 32/28 32/28 36/24 47/13 21/39 195/165 Leptaxis undata 36/24 32/28 29/31 42/18 42/18 30/30 211/149 Discula polymorpha 24/36 24/36 33/27 32/28 27/33 35/25 175/185 Monadenia troglodytes 22/38 24/36 =/= 31/29 31/29 —/- 108/132 Oxychilus cellarius 24/36 27/33 22/38 22/38 27/33 31/29 153/207 the results, Helix aspersa 3 showed marked departures when dropped from 20cm and 10 cm with apex up, but the departures are in opposite directions. The significance (if any) is unclear. In the case of Dinarica pouzolzi (Fér.) the reasons are obvious. This is a large heavy broad snail, and when dropped from only 10 cm tended to keep its orientation. (Its’ one-sided results when dropped broad side down may point to some overall asymmetry, but this is the position least easy to standard- ize with accuracy). The Leptaxis showed a strong tendency to rest with apex up when dropped from 10 cm in both the apex-up and apex-down positions, as well as an overall trend in this direction. There may be a real bias here. Of the species investigated, Leptaxis un- data (Lowe) and Oxychilus cellarius (Muller) are fairly uniformly brown; the rest have dif- ferent bands above and below, the difference being most striking in Discula polymorpha (Lowe), Eobania vermiculata (Muller) and (often) Theba pisana. An overall x? test for departure from 1:1 on the totals for each specimen is still highly significant at the 95% level (d.f. 13, x? = 72.9) as is an overall heterogeneity test on all the runs of observa- tions, usually six for each specimen but sometimes less (d.f. 76, x? = 173.8). In Table 3, it can be seen without statistical tests that the tall shells also have a strong bias to resting with the mouth definitely up (u) or nearly so (su) or definitely down (d) but that intermediate positions are not uncommon. It might well be thought that a small light form such as the little clausiliid Papillifera bidens (L.) might be held in any position by the roughness of the soft cloth onto which the snails were allowed to fall, as a larger shell might be on grass, and that the frequency of four positions merely reflects the natural tendency of the observer to overclassify a continuum of position into four quadrants. This, however, is not the case; if the shell is placed on a sheet of plastic which can be vibrated gently with the fingers, to test the stability of any given position, it is found that there are four principal positions, although those with mouth up and mouth down are more readily retained than the other two. In the case of the Placostylus, these heavy forms with somewhat irregular shells can take a greater variety of positions which depend on the exact shape of those parts of the shell resting on the substrate. Even these tend to rest most frequently with mouth down or up, or nearly up (su) in P. porphyrostomus (Pfeif- fer). Of the tall shells, Rumina decollata (L.) and Papillifera bidens are plainly coloured. (P. bidens has a reddish-brown band along the suture ornamented with whitish dots, but this 170 CAIN TABLE 3. Resting positions, after dropping, of tall-shelled snails. For explanation of the positions, see text. Height and initial attitude Rumina decollata 20 cm 10 ст Papillifera bidens 20 cm 10 ст Placostylus porphyrostomus 20 cm 10 cm Placostylus fibratus 20 cm 10 cm apex up apex down apex up apex down apex up apex down apex up apex down apex up apex down apex up apex down apex up apex down apex up apex down u su 23 = 27 = 20 = 20 = 90 == 12 = 16 — 22 = 13 = 63 — 14 28 6 20 9 30 6 27 35. 105 12 2 15 4 10 = 9 1 56 7 | 8 D D © s'u TABLE 4. Tests of heterogeneity between runs for each individual used, and for departure of the ratio of the resting positions from 1:1. Departure from 1:1 Helix aspersa 1 Heterogeneity of: x? 4 0.76 Helix aspersa 2 3 5.60 Helix aspersa 3 3 13.06 Eobania vermiculata 5 5.25 Levantina caesareana 5 8.00 Alabastrina alabastrites 5 Ul Pseudotachea splendida 5 5.70 Theba pisana adult 5 0.87 Theba pisana juv. 5 2.46 Dinarica pouzolzi 5 25.88 Leptaxis undata 5 11.60 Discula polymorpha 5 7.66 Monadenia troglodytes 3 4.44 Oxychilus cellarius 5 4.19 “Indicates significant difference at the 95% level. SNAIL HETEROSEMATISM 171 appears to be a breaking-up of the regular line of the suture—compare Cott, 1940 р. 110—and, being continuous, is not related to attitude.) The two species of Placostylus have dark throats, richly coloured in P. porphyro- stomus, bordered by prominent pale lips, but the body of the shell is fairly uniform in colour. DISCUSSION If the underside of depressed shells or the oral-umbilical area in tall ones were invariably shown when the animal fell and the colora- tion was vivid in comparison with that of the upper side or the main cone of the shell re- spectively, the result would be a sort of flash coloration but with the disadvantage that it could not be concealed again rapidly, the animal returning to a cryptic pattern. The up- per side of such forms as /berus marmoratus or Discula polymorpha is indeed more cryptic than the underside, but in some morphs of Cepaea (e.g. red 00045) and some of the species figured by E. von Martens neither aspect appears particularly cryptic, only noticeably different from the other aspect. In other morphs of Cepaea both aspects may be cryptic (e.g. yellow 12345) but still differ noticeably. It seems reasonable to suggest, therefore, that the difference is used for mis- leading the predators, not as warning colora- tion or a different kind of crypsis. The narrow black bands on a pale ground on the un- derside of Eobania vermiculata, lberus mar- moratus or Discula polymorpha are hardly cryptic yet unlike the heavily banded and mot- tled or dark brown upper surface. It seems unlikely that selection for crypsis could cause such differences; and selection against in- solation should merely pale the upper sur- face, leaving the lower unaffected. None of the shells investigated rested in- variably, or even nearly so, with the normally concealed parts exposed; on the contrary, ratios from about 1:2 to about 2:1 seem nearly always to hold, not 1:0 nor any approxima- tions to it. Equal or subequal tendencies to show the upper and under sides are seen not only in those shells with contrasting colour patterns but also in those which are plain. This property must be a function of the shape of the shell, not at all associated with a par- ticular style of patterning, and there is little indication that the disposition of the body in the shell causes much bias, although this needs further examination. Preliminary runs with an empty shell of Dinarica pouzolzi gave very similar results to those with a full shell, and tests with one of Placostylus porphyro- stomus on a gently agitated tympanum (as described above for Papillifera bidens) gave the same positions of stability as with the live animal. Furthermore, land snail shells do not have projecting wings or vanes, so that orientation by the medium through which they fall (air) will not occur, in contrast to the marine shells investigated by Palmer (1977). Exactly what determines the shape of any terrestrial gastropod’s shell is as yet quite uncertain, and it would not be reasonable to suggest that any of these shells have evolved primarily to fall in one of two opposite posi- tions. But just as haemoglobin was surely evolved for its oxygen-carrying properties, yet its concomitant redness can be used in blush- ing (Goodhart, 1960), so the shape of the shell, dictated by other considerations, may be used in whatever further ways it permits. To take the most frequent case so far, what will be the effect on a predator if the shell is equally likely to fall with the mouth up or the mouth down (in depressed shells with apex down or apex up respectively) and the colour patterns of the two aspects are different? This can be answered only by experimental observation, but it can be suggested that an arrangement which maximises the number of times the same aspect is shown before and after falling will strengthen the primary hunt- ing image of it. One which maximises the number showing the different aspect and thereby misleading the predator will be most strongly selected for, but the success of the device will depend on the strength of reinforcement of the primary hunting image and the unpredictability of what will be shown on each occasion. In fact, a random fall pro- ducing one of two markedly different images in an overall ratio of 1:1 may be the most effective strategy as in the game “which hand is it in?” (Driver & Humphries, 1970). Furthermore, since the underside pattern is normally hidden, it is free to vary, at least as far as visual selection for crypsis and selec- tion against heating by insolation are con- cerned. As already mentioned, there is a series of genetically determined variations of the under-side pattern in Eobania vermiculata and /berus marmoratus which leave the upper-side pattern much the same. Relative variability in under-side patterns needs more investigation in other species, such as Cepaea nemoralis and Theba pisana. 172 CAIN Humphries and Driver (1970) include vis- ible polymorphism in the shell of Cepaea un- der their heading of protean behaviour. Mimicry, however, they include (p. 286) under the heading of regular devices in predator/ prey strategy, as against the irregular devices which are characteristic of protean behaviour, but as an example of mimicry they refer to Sheppard's work which is on polymorphic mimicry. The status of polymorphic pop- ulations in their classification is therefore not clear. Their application of Chance and Rus- sell's term protean to many sorts of highly complex sequences of behaviour seems ex- cellent, but there are two groups of be- haviours which connect protean to regular be- haviour, namely (i) polymorphic mimicry in which each individual has only one constant signal-pattern but this varies from individual to individual, and (ii) such examples as the snails dealt with here are suggested to be. In these, a population may be monomorphic or polymorphic but some, if not all, of the in- dividuals have two distinctly patterned aspects. While recognizing that there is almost certainly a continuum, | would prefer to keep protean for behavior which is indeed protean and use the less specific term heterosematic (signalling different things) for less complex misleading behaviour. ACKNOWLEDGMENTS | am very grateful to Drs. L. M. Cook, J. Heller, C. Inglesfield, P.B. Mordan, C. R. C. Paul, B. Roth and S. Tillier, Dr. R. H. Cowie, and Mr. T. Grier for living material; to Dr. G. M. Davis, Dr. R. Robertson, Dr. S. Tillier and Dr. R. H. Cowie for criticism of this paper; and to Dr. R. H. Cowie for assistance with the work reported, Mr. B. Lewis for photography and Mr. C. West for discussion. LITERATURE CITED ALLEN, J. A., 1972, Evidence for stabilizing and apostatic selection by wild blackbirds. Nature, London, 237: 348—349. CAIN, A. J., 1977a, The uniqueness of the polymorphism of Cepaea (Pulmonata: Helicidae) in western Europe. Journal of Conchology, 29: 129—136. CAIN, А. J., 1977b, The efficacy of natural selection in wild populations. /n GOULDEN, C. (ed.), The changing scenes in natural sciences 1776-1976. Academy of Natural Sciences Special Publica- tion, 12: 111-133. CAIN, A. J., CAMERON, R. A. D. 8 PARKIN, D.T., 1969, Ecology and variation of some helicid snails in northern Scotland. Proceedings of the Malacological Society of London, 38: 269-299. CHANCE, M. R. A. 8 RUSSELL, W. M. S., 1959, Protean displays: a form of allaesthetic be- haviour. Proceedings of the Zoological Society of London, 132: 65—70. COTT, H. B., 1940, Adaptive coloration in animals. Methuen, London, xxxii + 508 p., 48 pl. DRIVER, P. M. & HUMPHRIES, D. A., 1970, Pro- tean displays as inducers of conflict. Nature, London, 226: 968-969. GOODHART, C. B., 1960, The evolutionary signifi- cance of human hair patterns and skin colouring. Advancement of Science, 17: 53—59. HUMPHRIES, D. A. & DRIVER, P. M., 1967, Erratic display as a device against predators. Science, N.Y., 156: 1767-1768. HUMPHRIES, D. A. 8 DRIVER, P. M., 1970, Pro- tean defence by prey animals. Oecologia, Berlin, 5: 285-302. JONES, J. S., LEITH, B. H. & RAWLINGS, P., 1977, Polymorphism in Cepaea: a problem with too many solutions? Annual Review of Ecology and Systematics, 8: 109—143. MANLY, B. F. J., MILLER, P. & COOK, L. M., 1972, Analysis of a selective predation experiment. American Naturalist, 106: 719-736. MARTENS, E. VON, 1867, Die preussische Е tion nach ost-Asien. Zoologischer Theil. Decker, Berlin, 2: i-xii, 1-447, 22 pl. MARTENS, G. VON, 1832, Uber die Ordnung der Bander an den Schalen mehrerer Landsch- necken. Deutsche Akademie der Naturforscher, Nova Acta physicomedica (Nova Acta Leo- poldina) 16: 177-216. PALMER, A. R., 1977, Function of shell sculpture in marine gastropods: hydrodynamic destabiliza- tion in Ceratostoma foliatum. Science, N.Y., 197: 1293-1295. PETTITT, C., 1973, A proposed new method of scoring the colour morphs of Littorina saxatilis (Olivi, 1792) (Gastropoda: Prosobranchia). Pro- ceedings of the Malacological Society of London, 40: 531-538. TAYLOR, J. W., 1910, Helix nemoralis Linné. Monograph of the Land and Freshwater Mol- lusca of the British Isles, part 17: 274-325. MALACOLOGIA, 1984, 25(1): 173-192 PATTERNS OF DIGESTIVE TRACT MORPHOLOGY IN THE LIMACISATION OF HELICARIONID, SUCCINEID AND ATHORACOPHORID SNAILS AND SLUGS (MOLLUSCA: PULMONATA) Simon Tillier Laboratoire de Biologie des Invertébrés marins et de Malacologie, Muséum national d'Histoire naturelle, 55, rue Buffon, 75005 Paris, France ABSTRACT The posterior part of the digestive tract of most Stylommatophora becomes hypertorted some 180° more during ontogeny than is seen in the post-torsional embryo. In a family of helicoid snails such as the Helicarionidae s./., this hypertorsion is retained during the initial steps of limacisation, and may be retained or lost in semislugs and slugs. The process by which the stomach is included in the foot cavity in the last steps of limacisation depends on the position of the upper posterior edge of the foot cavity, which in a semislug may be either behind or in front of the stomach. As a result there are four main processes by which the digestive tract of a helicoid snail may be integrated into the foot cavity. Some slugs originating from helicoid snails have exaggerated hypertorsion of their digestive tract. In succineids, the digestive tract is detorted in the course of limacisation. This detorsion may be related to their bulimoid shell shape. Athoracophorids have more exaggerated hypertorsion of the digestive tract than any other slug. Their peculiar anatomical characters may result from advanced limacisation. In particular, their lung may result from the development of structures that occur as accessory structures in the lung of some unrelated slugs. It is concluded that athoracophorids are more closely related to other Aulacopoda than to succineids. INTRODUCTION Although the digestive system occupies nearly the whole body and visceral cavity of the Stylommatophora, there are few data on this system (cf. Runham, 1975). Its descrip- tion takes only about twenty pages of Simroth & Hoffmann’s monumental 1354-page work (1908-1928). In most of the few general drawings scattered in the literature, the posi- tion of the tract has been modified to show as much as possible of its surface. Furthermore, as the internal morphology generally is not described, it is impossible to propose meaningful comparisons among taxa. An ex- ample is Rigby’s statement (1965) that suc- cineids are terrestrial opisthobranchs, proved erroneous by Solem (1978). As the digestive system of the Stylommatophora is relatively big, its comparative anatomy is very important in the study of /imacisation, the process by which a snail with a helicoid visceral hump evolves into a slug without a prominent visceral hump. The systematic position of two stylomma- tophoran families including slugs or semi- slugs, the Succineidae and Athoracophor- idae, has been controversial for more than a century, aS Summarized by Van Mol (1967) and Solem (1978). Solem places these two families among the Aulacopoda, an infraorder of sigmurethrous Stylommatophora that sur- prisingly includes helicoid snails and slugs, but no bulimoid snails other than succineids. The purpose of this paper, mainly based upon comparative gross anatomy of the digestive tract, is to show: (1), how the digestive tract of the Aulacopoda becomes entirely contained in the foot cavity (example using the helicar- ionids s./. [= most of Ariophantacea auct.)); and (2), new insights into systematic rela- tionships resulting from a comparison of these stages with the situation found in the families Succineidae and Athoracophoridae. To choose the helicarionids as a basis for comparison is justified not only by their un- rivalled wealth in semislugs, but also be- cause, like succineids and athoracophorids, they show nearly complete absence of in- ternal ridges in the stomach. The central ner- vous system and the pallial complex, which have been used to provide major taxonomic characters, have also been employed. (173) 174 TILLIER MATERIALS AND METHODS All animals mentioned in this paper are preserved in alcohol in museum collections (Table 1). To show the gross morphology of the digestive tract, the outline of the animals was first drawn with a camera lucida. The slugs and bulimoid snails were drawn in dor- sal view, the columellar axis parallel to the drawing plane; the helicoid snails were drawn seen from the apex of the shell, the columellar axis perpendicular to the drawing plane. The shell, when present, was removed or dis- solved in acid and the snail or slug pro- gressively dissected. Each time a part of the digestive tract was uncovered by dissection it was drawn under a camera lucida with the animal positioned as described above. Final- ly, when the digestive tract had been entirely drawn in situ, it was opened longitudinally to observe its internal morphology, which was depicted in the same drawing as the external morphology. For all drawings, the rectum was cut off when removing the lung during dis- section; its position along the suture of the body whorl is the same for all snails. No live specimen of any species studied could be obtained, and thus it was impossible to observe ciliary currents and food transport. Histological study of the digestive tract has not been attempted, since it seems more urgent to define the gross morphology first. A few sections of lung were made but not illustrated; they were also prepared from ani- mals preserved in alcohol. The lung was re- fixed in Bouins and the sections stained with Masson's trichrome. TABLE 1. List of specimens cited in this paper (MNHN = Museum national d'Histoire naturelle, Paris; MRAC = Musée Royal de l'Afrique Centrale, Tervuren). Athoracophoridae Athoracophoridae n. sp., Valley of Amoa, New Caledonia, Tillier coll. 18.1.1981, MNHN. Aneitea simrothi Grimpe & Hoffmann, Thiem, New Caledonia, Bouchet coll. 25.12.1978, G. M. Barker det., MNHN. Succineidae Succinea putris (L.), Le Hade (Seine maritime), France, Chevallier coll. 29.9.1967, MNHN. Succinea propinqua Drouët, llet la Mere, French Guyana, Е. Geay coll., MNHN. Hyalimax perlucidus (Quoy & Gaimard), St-Philippe forest, La Réunion, Lantz coll., MNHN. Omalonyx matheroni (Potiez & Michaud), between Cayenne and Kourou, French Guyana, Tillier coll. 1978, MNHN. Helicarionidae Helicarioninae (7?) Kalidos oleatus (Ancey), Marojezy, W from Sambava, alt. 1300 m, Madagascar, 8.12.1972, Fischer- Piette det., MNHN. Helicarioninae sp., Poindimié, New Caledonia, Bouchet coll. 29.12.1978, MNHN. Malagarion paenelimax Tillier, holotype, Marojezy, alt. 600 m., Madagascar, Blanc coll. 12.1972, MNHN. Ariophantinae (Parmarioni) Parmarion sp., Cambodia, Harmand coll. 1898, MNHN. Ariophantinae (Girasii) Mariaella dussumieri Gray, Mahé, probably southern India, Dussumier coll. 1835, MNHN. Gymnarioninae (?) Gymnarion sowerbyanus (Pfeiffer), Assinie, Ivory coast, Chaper coll. 1882, MNHN. Urocyclinae (7?) Granularion lamotiei Van Mol, Mt Nimba, Guinea, Lamotte coll. 1956, MNHN. Estria? sp. A (Van Goethem, 1977), Gopoupleu, Ivory Coast, Condamin and Roy coll. 1959, MNHN. Tresia parva Van Goethem, paratype, Mt Nimba, Guinea, Lamotte et al. coll. 1957, MNHN. Elisolimax madagascariensis (Poirier), Montagne d'Ambre, Madagascar, Blanc & Salvat coll. 1970, MNHN. Atoxon pallens Simroth, Virunga park, Zaïre, Vanschuytbroek coll. 1954, Van Goethem det. et leg., MNHN. Mesafricarion maculifer Pilsbry, Lodjo, Mongbwalu, Zaïre, Lepersonne coll. 7.1939, Van Mol det., MRAC. Parmacellidae: Parmacella sp., Algeria, M. Mares coll. 1876, MNHN. SOME PULMONATE DIGESTIVE TRACTS 175 GENERAL CHARACTERS OF THE DIGESTIVE TRACT OF THE STYLOMMATOPHORA Gross anatomy of the digestive tract In nearly all Stylommatophora, the most posterior region of the digestive tract, equiv- alent to the uppermost region in a coiled visceral mass, is the stomach, which is de- fined as the part of the digestive tract that receives the two ducts of the digestive gland. Between the mouth and the stomach are the buccal mass, the oesophagus and the crop; beween the stomach and the anus is the intestine. The anus opens on the right side of the body, except in sinistral species, and generally close to the pneumostome. The oesophagus always begins at the pos- terior upper aspect of the buccal mass. In carnivorous snails and slugs, the part of the buccal bulb behind the opening of the oesophagus is secondarily elongated back- ward to form an evaginable snout (e.g. Wat- son, 1915), but this is not the case in any species studied here. The oesophagus generally has internal longitudinal ridges. In the families studied here, these ridges, if present, generally are thin. In some species only two ridges are present (Elisolimax, Fig. 12), or ridges are absent (Aneitea simrothi Grimpe & Hoffmann, Fig. 20). The oesopha- gus passes posteriorly to an inflated crop. The term oesophagus is ambiguous for two reasons: (1) In most holopod snails and some Holopodopes (classification of Solem, 1978), the oesophagus itself is inflated into a crop separated from the inflated region im- mediately anterior to the stomach by a section of typically folded oesophagus (Tillier, un- published). This may be an oesophageal crop, and thus not homologous with the crop of such snails as Achatina, which Ghose (1963) claimed is of gastric origin. (2) The part of the digestive tract between oesophagus and stomach is sometimes divided into two inflated pouches separated by a constriction (cf. Gymnarion, Fig. 4, and Malagarion, Fig. 3). They are here called the anterior and posterior crop. It is unknown if the anterior crop is formed by dilatation of a section of oesophagus or if it is homologous with the anterior part of the crop of the snail ancestor, brought to this position by shortening the oesophagus during limacisation. In many species the transition from oesophagus to crop is gradual. It is incorrect to assume the oesophagus is shorter than the crop in all Stylommatophora (Runham, 1975): this paper shows that shortening of the oesopha- gus is the result of limacisation. In families studied here the crop has at most a few internal ridges which are possibly secondary. No species of these families examined has two longitudinal ventral ridges leading to the stomach, as found in Oxychilus (Rigby, 1963). In nearly all snails and slugs, the stomach is the part of the digestive tract extending farthest from the mouth, forming a bend from which the intestine goes forward. In no family studied here is the stomach clearly dif- ferentiated from the crop, inside or outside. It appears to be a simple end of the crop receiv- ing the two ducts of the digestive gland. As will be discussed later, the intestine opens from the lower columellar side of the stomach of most snails, and from the right lower side of the stomach of most slugs: if the digestive tract of snails is unwound, the two positions are the same. Generally, in the families here studied, there are no internal ridges extending from the duct openings of the digestive gland to the intestine, as are found in Oxychilus (Rigby, 1963) or Agriolimax (Runham, 1975). The most important differentiations are found in succineids, where the ducts of the digestive gland are connected by a deep groove (Rig- by, 1965), and Gymnarion, where two short and unequal typhlosoles extend to the be- ginning of the proximal intestine (Fig. 4). In most Stylommatophora the anterior duct of the digestive gland opens into the angle formed by the crop and the intestine either through a slit (Fig. 2) or through an oval or round aperture (Gymnarion, Fig. 4). The pos- terior duct forms an angle of about 90° with the anterior duct and opens through a round aperture into the posterior columellar side of the stomach. In all Stylommatophora but a few carnivorous slugs, the intestine goes forward along the left side of the anterior digestive tract to turn around the aorta, clockwise when seen in dorsal view. When it opens from the right side of the stomach it has to pass under the crop or the oesophagus. From the aorta, the intestine has one bend backward before going forward again to the anus. The two bends of the intestine will be called the peri- aortic bend and the prerectal bend. In some slugs the body cavity is too short to house the two intestinal bends along the crop, and as a result the prerectal bend coils around the crop or even in the posterior part of the foot cavity, 176 TILLIER behind the stomach (Elisolimax, Fig. 12; Milax gracilis [Leydig], Watson, 1930, pl. 2, fig.12). As demonstrated below, this is a secondary coiling, clearly not homologous in any way with the coil of the visceral mass in snails. Except for a few rectal ridges the intestine of helicarionids, succineids and athoracophorids does not show any internal morphological differentiation. Torsion and hypertorsion of the stylommatophoran digestive tract In embryos of stylommatophoran snails, torsion is about 90° counterclockwise in dor- sal view, head anterior, previous to the de- velopment of a helicoidal visceral mass (Helix, Fol, 1880; Achatina, Ghose, 1963; Fretter, 1975). In the anterior digestive tract of most snails and slugs, this 90° torsion opens the oesophagus into the left anterior extremity of the crop, and results in the disposition of the crop floor on the right side of the foot cavity instead of the pedal side (for example, Fig. 8). Just after torsion of the embryo's visceral mass, the intestine opens dorsally from the left side of the stomach and goes directly to the left side of the aorta without having to cross the crop or oesophagus. In this stage the posterior duct of the digestive gland opens into the left side of the stomach and the anterior duct opens into the right side of the stomach (Ghose, 1963). In any adult snail with a coiled visceral mass, the intestine passes under the crop or oesophagus, de- scribing a half circle around them before join- ing the aorta (Fig. 1 and 15; Orthurethura, Steenberg, 1925; also examined but not fig- ured here in about one hundred species). Thus, the posterior part of the digestive tract of Stylommatophora has been hypertorted during ontogeny after torsion of the embryo. The degree of this hypertorsion may be de- termined: there is no hypertorsion when the intestine runs directly from the upper left side of the stomach to the left side of the aorta (observed only in embryos and in some slugs, as Estria, Fig. 10); there is a 360° hypertor- sion when the intestine also issues from the upper left side of the stomach but describes a complete whorl around the right side of the crop before turning backward around the aor- ta (observed in some slugs, as Elisolimax, Fig. 12). In bulimoid snails with an elongated spire it seems that, at least for a few whorls, the degree of hypertorsion is somewhat a function of the number of whorls of the heli- coid visceral mass occupied by the posterior digestive tract: 90° in Succinea propinqua Drouét for half a whorl occupied (Fig. 16); 120° in Succinea putris (L.) for a three-quarter whorl occupied (Fig. 15); 180° in Euglandina carminensis (Morelet) for one and a quarter whorls occupied. This hypertorsion is primar- ily the result of the presence of the posterior digestive tract in the helicoidal visceral mass. A helix is indeed the result of winding plus torsion: the winding corresponds to the pro- jection of the spire on a plane perpendicular to its axis, and torsion is proportional to the height of the spire, which seems to be approx- imately the case for hypertorsion of high- spired snails. But this explanation of hypertor- sion as primarily a mechanical result of the development of the visceral mass and poste- rior digestive tract into a high conical spire requires further discussion; in fact, shell shape varies in Stylommatophora from buli- moid and nearly parallel to foot length to helicoid or even flat, with shell axis per- pendicular to foot length as in Kalidos (Fig. 1). Hypertorsion is at least 90° in all helicoid or planorboid snails that | have dissected, and is often more. The most common case is the opening of the intestine from the columellar lower side of the stomach (= lower right side), as in Kalidos. Furthermore, as shown in this paper, hypertorsion is preserved in most stages of shell reduction and even exagger- ated in some slugs, whereas it is lost in oth- ers. We have no data about hypertorsion in the embryo of flat-shelled snails, but Fol (1880) described a 180° hypertorsion when the visceral mass of the embryo of Limax sinks down into the foot cavity. Hypertorsion may result from the development of an elon- gated helicoid visceral mass in a hypothetical bulimoid stylommatophoran ancestor and may be retained for other reasons in de- scendants having a different shell shape, but this is at the moment purely hypothetical. The problem of variations in torsion of the digestive tract of Stylommatophora may be treated in three ways: as relationships of hypertorsion and shell shape, as variation found in the limacisation of flat-shelled snails, or as variation found in the limacisation of bulimoid snails. This paper analyses mainly the variations in the disposition of the di- gestive tract found in flat-shelled snails, through the example of the helicarionids which show various patterns. The variations found in the limacisation of bulimoid snails is analyzed through the example of the suc- SOME PULMONATE DIGESTIVE TRACTS 177 cineids; as far as | know from dissections in other families (oleacinids and acavids), it is representative of the general pattern found in the limacisation of bulimoid snails. LIMACISATION OF HELICARIONIDS Snails belonging to diverse groups have been used to understand how the digestive tract is modified in limacisation. Knowing the direction of evolutionary change, from snails to slugs, and applying the principle of parsi- mony, it is easy to arrange linearly the differ- ent steps observed in a number of series as little as possible. This does not imply at all monophyly of different steps of each series; it implies that in each series the ancestors of the most advanced slugs probably passed through stages in which they were similar to animals here considered representative of less advanced stages of the same evolution- ary series. This means that evolution from snails to slugs is canalized. There are only a limited number of patterns, and when limacisation advances, the number of possi- ble options for further steps is more and more limited. The suprageneric groups of Asian and Pacific helicarionids used here were defined by Baker (1941) and revised by Solem (1966). The taxonomic basis for African groups is defined by Van Mol (1970) and Van Goethem (1977). The suprageneric groups are not necessarily used here at the same rank as these authors used, which changes nothing in the analysis of evolutionary trends. Here the name Helicarionidae includes the Ariophantinae, Durgellinae, Dyakiinae, Gym- narioninae, Girasiinae, Helicarioninae, Par- marioninae, Sesarinae, Trochozonitinae and Urocyclinae of various authors. As defined, the helicarionids occur from Tasmania to South Africa through the Indo-Australian Archipelago, Southeast Asia, India, Mada- gascar and the Ethiopian region. They are absent from the Middle East and Iran. The first steps of limacisation Helicarionids with a well-developed shell, up to six whorls (more in a few species), are found all over the group’s range. No African or Asian helicarionid snail was examined here. They are probably similar to Madagascan and Melanesian species here examined, but it is not absurd to suppose from slugs described FIG. 1. Kalidos oleatus (Ancey), Madagascar. Rec- tum removed. Scale line 5 mm. AD, anterior duct of digestive gland; BM, buccal mass; C, crop; |, in- testine; OE, oesophagus; PD, posterior duct of digestive gland; S, stomach; SG, salivary glands. further that at least their oesophagus is longi- tudinally ribbed. Madagascan and New Caledonian helicarionids examined have no internal morphological differentiation of their digestive tract; it is possibly a secondary loss, but | have no argument to accept or reject this hypothesis. In a four-whorled Kalidos oleatus (Ancey) from Madagascar (Fig. 1), the lung occupies about one-third of the body whorl and the foot cavity is shorter than the tail. The oesophagus is longer than the crop into which it passes progressively at the level of the upper end of the lung. The intestine opens from the col- umellar upper end of the stomach, and passes under the oesophagus (hypertorsion) to reach the left side of the aorta. The two ducts of the digestive gland are almost circu- lar in section and open into both sides of the entry of the intestine, forming an angle of nearly 180°. The anterior duct opens between the crop and the intestine, and the posterior duct opens into the columellar side of the stomach. A New Caledonian two and three-quarter- whorled helicarionid has the same basic dis- position, but the oesophagus, crop and lung are all shorter than in Kalidos (Fig. 2). As a result, the intestine passes under the crop. The transition from oesophagus to crop is more abrupt and the opening of the anterior 178 TILLIER FIG. 2. Helicarionidae sp., New Caledonia. Scale line 2.5mm. AD, anterior duct of digestive gland; BM, buccal mass; С, crop; I, intestine; OE, oesophagus; PD, posterior duct of digestive gland; RM, retractor muscles. duct of the digestive gland is a transverse slit here, but these two characters probably do not reflect a general trend in the limacisation of helicarionids. Semislugs Semislugs are snails in which shell reduc- tion has proceeded so far, and with such drastic shortening of the oesophagus, that the crop (when it is not separated from the stom- ach by a section of oesophagus) is at least partly contained in the foot cavity and the animal cannot retract inside the shell. How- ever, the posterior edge of the foot cavity is lower and further forward than the most pos- terior part of the digestive tract, and the stom- ach is retained in the upper visceral cavity. In the limacisation of semislugs the crop and stomach are progressively uncoiled and be- come parallel to foot length, but may or may not retain hypertorsion. Plesiomorphic hypertorted semislugs. The reduction in number of visceral mass whorls results in the reduction of the foot and visceral cavities. The lung is reduced and the kidney either is shortened or becomes transverse and folded onto itself, as described by Solem (1966), Van Mol (1970) and Van Goethem (1970) in helicarionids. These modifications of the pallial complex are not sufficient to preserve the same ratio of internal volume/ body size; the digestive system has to be reduced, and/or the foot cavity enlarged. Theoretically, all digestive tract parts could be reduced and retain the same proportions. In fact, at least in helicarionid semislugs, the stomach and crop retain approximately the same size, whereas oesophagus length is largely reduced and intestine length less re- duced. As shown by the Madagascan Malagarion (Fig. 3) and the West African Gymnarion (Fig. 4), the crop partly sinks into FIG. 3. Malagarion paenelimax, Madagascar. Scale line 5mm. AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; SG, salivary glands. SOME PULMONATE DIGESTIVE TRACTS 179 SG AD FIG. 4. Gymnarion sowerbyanus, Ivory Coast. Rec- tum removed. Scale line 5 mm. AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; SG, salivary glands. the foot cavity, which in this stage keeps the same relative length as in snails. When the passage from foot cavity to visceral cavity is necessarily reduced, because of the pres- ence of the lung, if the foot cavity is not enlarged, the crop is divided into two inflated chambers separated by a constriction: the anterior crop, contained in the foot cavity, and the posterior crop, contained in the visceral cavity. The visceral part of the digestive tract retains hypertorsion and the intestine passes under the constriction of the crop (the figure of the digestive tract of Malagarion in Tillier, 1979, is erroneous). The digestive tract of Malagarion has no internal ridges. Gymnarion has oesophagal ridges, a few longitudinal ridges in the constriction of the crop, plus a few small ridges in the stomach region (two thin longitudinally unequal ridges from the openings of the digestive gland, one small transverse ridge). The constriction between foot cavity and visceral cavity, which differentiates an ante- rior and a posterior crop, is clearly related to the anterior position of the visceral cavity edge in the groups studied here. If the vis- ceral mass is further back, as in the family Vitrinidae, the passage from the foot cavity to the visceral cavity is wider and the crop keeps the same diameter from the end of the oesophagus to the stomach (Tillier, un- published). Apomorphic semislugs retaining hypertor- sion. When the shell and visceral cavity are reduced more than in Malagarion and Gymnarion, the part of the digestive tract anterior to the stomach is entirely uncoiled and the whole crop is parallel to body length. Such a disposition is shown by the African Mesafricarion maculifer Pilsbry (Fig. 5; slight- ly contracted). In this species the hermaphro- dite gland is retained on the right side of the visceral mass. The two lobes of the digestive gland and their ducts occupy, as a result, a position different from the one found in Gymnarion: the anterior duct is in a lower left and posterior position, whereas the posterior OE AC AD 5 FIG. 5. Mesafricarion maculifer, Zaire. Rectum ге- moved. Scale line 5mm. A, aorta; AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; RM, retractor muscles; S, stomach. 180 TILLIER FIG. 6. Parmarion sp., Cambodia. Rectum re- moved. Scale line 2.5mm. A, aorta; AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; RM, retractor muscles; S, stomach; SG, salivary glands. duct is in the position where the anterior duct would be expected. The genital apparatus is entirely contained in the lower and right part of the foot cavity, the hermaphroditic gland excepted. Apomorphic semislugs partly losing hypertorsion. In some advanced semislugs, such as the Asiatic Parmarion (Fig. 6) or the West African Granularion (Fig. 7), the poste- rior part of the crop and stomach are not only parallel to the body axis, but also partly de- torted when compared with the Gymnarion- Malagarion stage. The posterior crop is no longer differentiated from the stomach and proximal intestine in diameter, and the in- testine now opens to the left and forward from a stomach. As a result of this 90° detorsion clockwise, the posterior duct of the digestive gland opens upward into the stomach, where- as it would open downward without detorsion. SG FIG. 7. Granularion lamottei, Guinea. Rectum re- moved. Scale line 5 mm. AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; S, stomach; SG, salivary glands. It still forms an angle of about 90° with the anterior duct. The anterior crop is more de- veloped than in former stages of limacisation and is entirely contained in the foot cavity. In Parmarion, only a few thin oesophageal ridges are found, whereas in Granularion these ridges are thicker and extend into the anterior extremity of the crop, where they are wrinkled. In Parmarion, the foot cavity is lon- ger than the tail. In Granularion, it is shorter and its posterior extremity is further back than its posterior upper edge: the tail is partly hollow. It cannot be proven that partial detorsion did not happen sooner in limacisation than this stage, given our present stage of knowl- edge. It is improbable, but not impossible. SOME PULMONATE DIGESTIVE TRACTS 181 Slugs A semislug evolves into a slug when the stomach sinks into the foot cavity, lower than the posterior edge of the foot cavity. When this process occurs, this edge of the foot cavity may be in front or behind the posterior end of the digestive tract and visceral cavity. As an advanced semislug may have retained or lost hypertorsion of the digestive tract, four situations are possible, depending on which type of semislug is ancestral: 1) a partly de- torted semislug having the posterior edge of the foot cavity in front of the stomach; 2) behind it; 3) a hypertorted semislug having the posterior edge of the foot cavity in front of the stomach; 4) behind it. When the posterior edge of the foot cavity is in front of the stom- ach, the tail becomes progressively hollow posteriorly and the digestive tract has to turn around to go down into the enlarged foot cavity. When the posterior edge of the foot cavity is behind the stomach, the foot be- comes more hollow ventrally and the di- gestive tract may sink directly into it without any change in its disposition. Slugs originating from partly detorted semi- slugs. Group |: Posterior upper edge of foot cavity anterior to stomach. This process may be illustrated by the Estria-Rhopalogonium group, West African taxa considered monophyletic by Van Goethem (1977). By its general as well as genital anatomy, Granular- ion is probably a member of this group less advanced in limacisation. As described above, in this genus only the crop is in the foot cavity that enters a little into the tail (Fig. 7). Estria? sp. A (Van Goethem, 1977) may be representative of the next step in limacisation from a similar ancestor (Fig. 8). The foot cavity is longer than in Granularion, although its upper edge occupies nearly the same posi- tion. The stomach sinks into the foot and forms a bend in a plane perpendicular to foot sole and foot length. The intestine is entirely included in the foot cavity on the left side of the crop. In Tresia parva Van Goethem (Fig. 9), the foot cavity extends backward to the middle of the tail; the disposition of the di- gestive tract is similar but here the bend formed by the stomach can spread parallel to foot sole. The loss of hypertorsion is achieved in the stomach region where the intestine opens from the left side of the stomach. The process is completed in Estria (Fig. 10, re- drawn from Poirier, 1888), which has a still longer foot cavity where the oesophagus and crop are straight again, as in the ancestor apomorphic semislug, and parallel to foot length. The loss of hypertorsion appears defi- nitely retained. In Granularion, the oesophagus has straight folds, prolonged inside the ventral side of the crop by a few wrinkled ridges; the sides and upper surface of the crop have internal, thin, scarce transverse ridges. Tresia has only oesophageal folds. No Estria could be directly observed. Group Il: Posterior upper edge of foot cavity shifted backward. The final stage of this proc- ess is illustrated by the Indian Mariaella (Fig. 11). The transformation from a disposition similar to the one found in Parmarion (Fig. 6) FIG. 8. Estria? sp. A (Van Goethem, 1977), Ivory Coast. Rectum removed. Scale line 5 mm. A, aorta; AD, anterior duct of digestive gland; BM, buccal mass; С, crop; |, intestine; OE, oesophagus; PD, posterior duct of digestive gland; S, stomach. 182 TILLIER FIG. 9. Tresia parva Van Goethem, paratype, Guinea. Rectum removed. Scale line 2.5 mm. A, aorta; AD, anterior duct of digestive gland; BM, buccal mass; С, crop; I, intestine; OE, oesophagus; PD, posterior duct of digestive gland; RM, rectrac- tor muscles; S, stomach. is a simple vertical translation ventrally, allowed by the position of the upper posterior edge of the foot cavity further back. Here is a large distance between the two ducts of the digestive gland. The oesophagus has a few longitudinal ridges; the crop has a few trans- verse ridges. The crop forms a kind of rostrum in front and beneath the opening of the oesophagus, an arrangement that is also found more or less developed in Gymnarion, Parmarion, Granularion and Estria? sp.A (Figs. 4, 6, 7, 8). The intestine has a few longitudinal short ridges in the beginning of the periaortic bend. In Mariaella the lung is prolonged downward and backward by air sacs; the frontal air sac separates the prev- ious foot cavity from the previous visceral Cavity. Austenia doisutepensis Solem from Cam- bodia possibly illustrates a process of limacisation by loss of hypertorsion with the foot cavity edge in an intermediate position, FIG. 10. Estria alluaudi, Ivory Coast, reproduced from Poirier (1888). Scale line 5 mm. A, aorta; AD, anterior duct of digestive gland; BM, buccal mass; С, crop; |, intestine; PD, posterior duct of digestive gland; S, stomach; SG, salivary glands. but the stomach is not located in Solem’s figure (1966) and it is difficult to determine... Slugs originating from hypertorted semi- slugs. Group Ш: Posterior edge of foot cavity an- terior to stomach. Slugs possibly illustrating the steps of limacisation from such a semislug were not observed, which does not prove that it did not or does not occur. In such a process, which would start from a semislug similar to Mesafricarion (Fig. 5) but having the visceral hump further in front, the proximal intestine should go down first into the right part of the foot cavity, followed by the stomach and final- ly the posterior crop. Intermediate positions do not seem very functional, mainly because the proximal intestine would have to occupy the space normally filled by the genital appa- ratus. However, hypertorted slugs having the lung in the anterior part of the body, as Arion, could have evolved through such a process; the necessity for housing the proximal in- testine on the right side of the crop in the intermediate steps of limacisation would ex- SOME PULMONATE DIGESTIVE TRACTS 183 FIG. 11. Mariaella dussumieri, probably southern India. Rectum removed. Scale line 5 mm. A, aorta; AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; RM, rectractor muscles; S, stom- ach; SG, salivary glands. plain why the spermoviduct is on the left side in Arion (general anatomy of Arion depicted by Van Mol, 1962, digestive tract slightly de- torted for drawing). Group IV: Posterior upper edge of foot cav- ity behind stomach. In such a case the hyper- torted digestive tract, similar to the one of Mesafricarion (Fig. 5), sinks into the foot cav- ity, enlarged ventrally. The intermediate steps of this process have not yet been observed in helicarionids, but Parmacella illustrates this process in another family (Tillier, un- published). Exaggeration of hypertorsion in urocycline slugs. In Elisolimax the intestine opens from the upper left end of the stomach and makes a complete coil around the crop before turning around the aorta (Fig. 12). This is half a whorl more than in slugs directly derived from a FIG. 12. Elisolimax madagascariensis, Madagas- car. Rectum removed. Scale line 5 mm. À, aorta; AD, anterior duct of digestive gland; BM, buccal mass; C, crop; |, intestine; OE, oesophagus; PD, posterior duct of digestive gland; RM, retractor muscles; S, stomach; SG, salivary glands. hypertorted semislug similar to Mesafricarion. Van Goethem (1977) believes that the Urocy- clini are related to Mesafricarion on the basis of their genital characters. If he is right and the Urocyclini ancestor was similar to Mesafricarion, this implies that exaggeration of hypertorsion occurred either during the in- termediate steps, or after limacisation. Exag- gerated hypertorsion may be related to the maintenance of the whole genital apparatus on the right side of the digestive tract. No observation makes these hypotheses un- likely. In particular, the anterior duct of the digestive gland of Elisolimax is in the position expected if the digestive tract of Mesafricarion is hypertorted 180° (Figs. 5, 12), and the exaggeration of hypertorsion is obviously easier to obtain from a hypertorted semislug than from a detorted one. Atoxon (Fig. 13) is 184 TILLIER FIG. 13. Atoxon pallens, Zaire. Rectum removed. Scale line 5 mm. A, aorta; AC, anterior crop; AD, anterior duct of digestive gland; BM, buccal mass; |, intestine; OE, oesophagus; PC, posterior crop; PD, posterior duct of digestive gland; S, stomach. modified still more than Elisolimax when com- pared with Mesafricarion. The anterior crop is differentiated by internal ridges, and the pre- rectal intestinal bend is longer, which puts the anterior duct of the digestive gland farther to the left. LIMACISATION OF SUCCINEIDS In succineids which are plesiomorphic in limacisation, e.g. Succinea putris (Fig. 15), the shell is bulimoid and only slightly oblique with relation to foot length. The shell has little more than three whorls in Succinea putris and not much more in any succineid snail. When compared to stylommatophoran standards for whorl number, even the most plesiomorphic succineids are advanced in shell reduction. The absence of succineid species or genera having a more developed visceral mass, in- termediate between the stage observed in Succinea and the stage observed in most snails (at least five whorls) probably explains FIG. 14. Possible evolutionary migration of inser- tion of penial retractor along inner wall of body cavity, from hypothetical columellar origin “0.” 01, position found in succineids; 02, position found in other Aulacopoda; PR1 and PR2: corresponding position of penial retractor. Outline of Succinea putris (Fig. 15). Rectum removed. why succineids were considered as a distinct group for such a long time. The relative com- paction of the succineid oviduct when com- pared with the spermoviduct of most Stylom- matophora, and probably the transverse posi- tion of the kidney, may be the result of the first steps in limacisation. The compaction of the spermoviduct is observed in bulimulid semi- slugs (Van Mol, 1971), and the kidney is transverse in Asian advanced helicarionids (although folded and retaining an U-shaped ureter; Solem, 1966). In the oleacinids, an- other family of bulimoid snails with trends in shell reduction, the kidney also has odd arrangements: the kidney is wider than long in the Streptostylini, and the primary ureter is independent of the rectal side of the kidney in the Euglandini. Possibly because all in- termediates between an eight-whorled shell and a semislug occur, these original trends were never used to define a suborder, as in the case of succineids. Even the most plesiomorphic succineids differ basically from all families studied here in SOME PULMONATE DIGESTIVE TRACTS 185 FIG. 15. Succinea putris, France. Rectum removed. Scale line 2.5 mm. A, aorta; AD, anterior duct of digestive gland; BM, buccal mass; C, crop; DC, digestive caecum; |, intestine; OE, oesophagus; PD, posterior duct of digestive gland; PR, penial retractor; SG, salivary glands. the insertion position of their penial retractor, which begins beside the origin of the aorta, inside the periaortic bend of the intestine (01, Fig. 14 and Tillier, 1981, fig. 2). In all Au- lacopoda which | have examined, except in a few endodontids where the penial retractor is a branch of the columellar retractor, it starts from the lung floor, lung border or body wall outside the periaortic bend of the intestine (02, Fig. 14). If we consider the penial retrac- tors of all Stylommatophora as homologous, the position of its insertion implies that from a probable columellar origin it shifted to the left in succineids, whereas it migrated to the right and then to the left around the crop in all other Aulacopoda. This purely theoretical migration is shown in Fig. 14. From a columellar origin 0, the insertion of the penial retractor may have migrated along the palatal wall of the visceral cavity to 01 in succineids, the retrac- tor passing under the crop and above the proximal intestine all along its migration. In other Aulacopoda, it may have migrated from the same origin 0 along the floor of the lung to 02, the retractor passing above the crop all along its migration. In Fig. 14 the outline of Succinea putris is used for clarity although this process, if it ever occurred, probably did so in a stage in which the shell and visceral mass were much more developed than in Succinea. In Succinea putris (Fig. 15), the oesopha- gus is ribbed, short and opens into the left anterior end of the crop (torsional twist of Rigby, 1965). The crop is long and makes nearly one whorl in the visceral mass before ending, without any morphological discontinu- ity, in the stomach. The hypertorsion of the stomach and proximal intestine is about 120°. The anterior duct of the digestive gland opens backward into the angle formed by the crop and proximal intestine. The posterior duct op- ens into the columellar side of the stomach and forms an angle of about 90° with the anterior duct. The posterior duct has a caecum in which the groove that joins the openings of the ducts internally ends. In the Guyanese Succinea propinqua, the visceral mass has about one whorl less than in Succinea putris (Fig. 16). The oesophagus is shorter in the former and has only very thin longitudinal ridges. Hypertorsion and the oesophageal torsional twist are partly lost. As a result, the posterior duct of the digestive gland opens upward and opposite the col- umellar side of the stomach. lts caecum op- ens at the junction of the duct and intestine. In Hyalimax perlucidus (Quoy & Gaimard) from the Mascarenes the reduced shell and visceral mass are completely covered by the mantle (Fig. 17). The foot cavity is not en- larged as compared with the foot cavity of Succinea, whereas the upper visceral cavity is greatly reduced. This situation results in great reduction of all parts of the digestive tract, a unique feature. Furthermore, the di- gestive tract has not only large oesophaageal ridges, but also thin crop ridges that run along the right inner side of the crop. Because of detorsion, these ridges that run to the poste- rior duct of the digestive gland are dorsal, whereas in a similar position they would be ventral in a torted slug. The crop ridges and also sacculations of the intestine, observed only in Hyalimax among succineids ex- amined, may be related to the general relative reduction in size of the digestive tract as they increase the internal digestive surface. The arrangement of the digestive tract is other- wise about the same as in Succinea pro- pinqua. Detorsion is, however, a little more 186 TILLIER DC AD